text
stringlengths 4
2.52M
| meta
dict |
|---|---|
{#sp1 .418}
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
*Candida* vulvovaginal infection is a frequent and common distressing disease affecting 70%--75% of women of childbearing age worldwide at least once during their lifetime. 5%--10% of women with a primary episode of VVC subsequently experience frustrating recurrent infection (RVVC) [@pone.0022770-Fidel1], defined as at least three-four specific episodes within one year [@pone.0022770-Sobel1], [@pone.0022770-Magliani1]. There are two forms of RVVC: primary RVVC is idiopathic with unknown predisposing factors, secondary RVVC is the occurrence of frequent episodes of acute VVC because of certain predisposing factors such as hormone replacement therapy or diabetes mellitus [@pone.0022770-Fidel2].
It has been demonstrated that the vaginal mucosa, its tissue structure and cervicovaginal fluids, contains both humoral and cellular components of innate and acquired immune responses [@pone.0022770-Cassone1]. Animal models are frequently used to evaluate host defense mechanisms against *Candida* vaginitis [@pone.0022770-Naglik1]. *Candida albicans* is not a natural colonizer of the vaginal mucosa but the use of estrogen favors a persistent infection and this can be exploited in order to study the immune response [@pone.0022770-Fidel2], [@pone.0022770-Cassone1]. The mouse infection model has many advantages, such as its similarity with the human infection and the possibility of performing vaginal lavages to quantify the fungal burden. The major disadvantages are the requirement for a condition of pseudo-estrus induced by estrogen, the neutral vaginal pH, and the lack of symptoms [@pone.0022770-Naglik1].
The Th1 response has been shown to be induced in the systemic draining lymph nodes of vaginally infected mice, but does not protect against the infection [@pone.0022770-Fidel3]. Despite the evidence reported for competent vaginal cell-mediated immunity, resident T cells, during infection, appeared to have little or no protective role [@pone.0022770-Fidel4], [@pone.0022770-Nandi1], [@pone.0022770-Fidel5], [@pone.0022770-Taylor1], [@pone.0022770-Ibraghimov1], [@pone.0022770-Saavedra1].
Th17 cells belong to a lineage different from that of Th1 and Th2 cells, and they are characterized by the production of IL-17A, IL-17F and IL-22 [@pone.0022770-Harrington1]. The protective action of IL-17 against extracellular pathogens also involves neutrophil recruitment to the infection sites [@pone.0022770-Matsuzaki1]. IL-17 has a central role in protective immunity against *C. albicans* systemic and oral infections [@pone.0022770-Huang1], [@pone.0022770-Conti1], [@pone.0022770-Pirofski1]. In response to a systemic challenge with *C. albicans*, IL-17AR-deficient mice showed a reduced survival rate and a significant increase of kidney fungal burden. Mobilization and influx of neutrophils to infected organs were also impaired and delayed [@pone.0022770-Huang1]. In another study, the Th17 response also conferred protection against oropharyngeal candidiasis through neutrophil recruitment and antimicrobial factor production [@pone.0022770-Conti1]. In the present work, we focus on the role of IL-17 in protecting against vaginal candidiasis, exploiting a new in vivo imaging technique that we have recently developed [@pone.0022770-Enjalbert1], [@pone.0022770-Pietrella1]. In the in vivo imaging method a *C. albicans* strain is used that expresses a cell surface luciferase constructed by fusing the *Gaussia princeps* luciferase to the *C. albicans* PGA59 glycosylphosphatidylinositol-linked cell wall protein [@pone.0022770-Enjalbert1]. This technique allows continuous, non-invasive monitoring of the spatial and temporal progression of vaginal infection in live mice and has proved useful to evaluate vaccinal and immunotherapeutic approaches to the treatment of vaginal candidiasis [@pone.0022770-Pietrella1].
Results {#s2}
=======
Th17 response in murine vaginal infection by *Candida albicans* {#s2a}
---------------------------------------------------------------
We investigated the role of Th17 during vaginal candidiasis using CD1 female mice maintained under pseudo-estrus conditions and vaginally infected with the pathogenic *C. albicans* gLUC59 strain as previously reported [@pone.0022770-Enjalbert1]. gLUC59 has been used to monitor the course of different types of *C. albicans* infection. Studies performed on mice infected with *C. albicans* strain CA1398 carrying the *ACT1p-gLUC* fusion (gLUC59), or the control strain CA1399 which did not express *gLUC59*, showed that these gLUC59 and control strains were equally pathogenic [@pone.0022770-Enjalbert1]. Data presented in [Figure 1](#pone-0022770-g001){ref-type="fig"} show that significant luminescence signals, obtained following injection of the luciferase substrate coelenterazine in the vagina and animal imaging using the Xenogen IVIS-200™ imaging system, were evident after 4 days of infection and declined after day 18 post-infection ([Fig. 1B](#pone-0022770-g001){ref-type="fig"}). The analysis of the total photon emission showed a significant reduction of the fungal load from day 20 post-infection; a similar trend was observed by CFU counting from vaginal lavages from the same mice ([Fig. 1C](#pone-0022770-g001){ref-type="fig"}). Analysis of the two parameters measured to estimate the fungal burden, photon emission and CFUs, showed a good correlation between the results obtained with the two methods, for all experimental time points ([Fig. 1D](#pone-0022770-g001){ref-type="fig"})
{#pone-0022770-g001}
In selected experiments, a histological analysis of the vagina was performed and the cellular composition of the vaginal fluid was examined. Results reported in [Fig. 2A](#pone-0022770-g002){ref-type="fig"} show that different cell types were present in the vaginal fluid of infected mice, especially neutrophils and epithelial cells. In uninfected mice, mainly epithelial and, rarely, immune cells, were detected. The trend of the cellular influx in the vagina of infected mice was evaluated by cytofluorimetric analysis ([Fig. 2B](#pone-0022770-g002){ref-type="fig"}) showing that, 48 h post-infection, a massive infiltration of neutrophils was evident (75% of total cells). Subsequently, the level of neutrophils decreased to about 60% of the total cells, a level that was maintained until the monitoring of the *Candida* infection finished ([Fig. 2B](#pone-0022770-g002){ref-type="fig"}). The histological analysis performed on day 1 post-infection evidenced that intraepithelial microabscesses consisting of polymorphonuclear cells were only present in the vagina of infected mice ([Fig. 2C](#pone-0022770-g002){ref-type="fig"}). Moreover a massive presence of fungal hyphae with rare blastospores was observed in the vaginal lumen and in the superficial epithelial layer until day 22 post-infection ([Figure 2C](#pone-0022770-g002){ref-type="fig"}).
{#pone-0022770-g002}
Because it has been demonstrated that IL-17 has a critical role in protecting against mucosal, particularly oropharyngeal, candidiasis [@pone.0022770-Conti1], we investigated the presence of this cytokine in our model of vaginal infection. Results reported in [Figure 3A](#pone-0022770-g003){ref-type="fig"} demonstrate that there is an early production of IL-17, starting 48 h after the challenge, reaching a maximum 14 days post-infection, and subsequently decreasing to return to basal levels after 5 weeks of infection. The early production of IL-17 in the vaginal wash could presumably be attributed to PMN and epithelial cells, which are known to be innate system cells capable of producing IL-17 [@pone.0022770-Cua1]. We also performed experiments using lower doses of *C. albicans*. The results showed that inocula of 10^7^, 5×10^6^ or 10^6^ are also able to induce IL-17 production ([Fig. 3B](#pone-0022770-g003){ref-type="fig"}). In a parallel experiment we evaluated the ability of the control strain of *C. albicans* CA1399 to induce IL-17 production. All concentrations of inocula used in the experimental vaginal infection induced cytokine production at day 14 ([Figure 3C](#pone-0022770-g003){ref-type="fig"}), as did the gLUC59 strain. It has been reported that elicitation of the IL-17 response correlated with the ability of *C. albicans* hyphae to stimulate antigen-presenting cells for the priming of Th-17 responses in vitro and for the production of IL-23, but not IL-12 [@pone.0022770-AcostaRodriguez1]. IL-23 induces the differentiation of naive CD4(+) T cells into helper T cells that produce IL-17 [@pone.0022770-Iwakura1]. Moreover, IL-23 appears to induce IL-17, IL-1 and IL-6 production from cells of the innate immune system [@pone.0022770-Bettelli1]. Given this premise, we analyzed IL-23 levels in the intravaginal lumen. Results reported in [Figure 3D](#pone-0022770-g003){ref-type="fig"} show the presence of IL-23 in the vaginal wash of infected mice. An increased production of this cytokine was manifested 24 h after the challenge, reaching a maximum at 48 h, declining from day 3--10 post-infection and then increasing until the end of the monitoring. Importantly, IL-23 was recovered in vaginal fluids of infected mice at all time points except day 14 post-infection ([Fig. 3D](#pone-0022770-g003){ref-type="fig"}).
{#pone-0022770-g003}
Role of vaginal and lymph node cells in Th-17 response {#s2b}
------------------------------------------------------
In selected experiments the vagina and the lumbar lymph nodes from both infected and non infected mice were removed and vaginas were treated with collagenase. Results reported in [Figure 4A](#pone-0022770-g004){ref-type="fig"} demonstrate that IL-17 was produced by vaginal cells recovered from infected mice and that production reached a maximum 15--22 days post-infection, then declining. In a parallel analysis the production of IL-23 was tested in the same supernatants. Kinetic analysis showed a significant increase of IL-23 levels at day 4, which declined in the following days ([Fig. 4B](#pone-0022770-g004){ref-type="fig"}). A similar trend for IL-17 and IL-23 production was observed when vaginal cells were re-stimulated with heat-inactivated yeasts. To study the phenotype of cells producing IL-17, cells cultivated for 72 h were recovered, fixed, and labeled for Gr-1 or CD4 receptors. Intracellular IL-17 expression was evaluated in CD4 T cells and neutrophils, which are the major IL-17 producing cells. In our experimental conditions CD4 positive cells appeared the main source of IL-17, with maximum intracellular expression between days 8 and 21 ([Fig. 4C](#pone-0022770-g004){ref-type="fig"}). Yet some Gr-1 positive cells (neutrophils) also had appreciable levels of IL-17 intracellular expression seven days post infection ([Fig. 4D](#pone-0022770-g004){ref-type="fig"}).
{#pone-0022770-g004}
Draining lymph nodes were also removed, cultivated for 72 h, then either left untreated, or stimulated with heat inactivated *C. albicans*, and the supernatant fluids were tested for the presence of IL-17 and IL-23. The results in [Figure 5A](#pone-0022770-g005){ref-type="fig"} show that IL-17 was produced by lymph nodes 22 days post-infection. After stimulation in vitro lymph node cells produced significant levels of IL-17 even at day 8, suggesting that some cells were already activated. A significant increase in IL-23 production was manifested 7 days after infection reaching a maximum at day 22 ([Figure 5B](#pone-0022770-g005){ref-type="fig"}). A similar trend was observed when cells were re-stimulated with heat inactivated *Candida* cells.
{#pone-0022770-g005}
Role of Th17 in vaginal *Candida* infection {#s2c}
-------------------------------------------
Recently it has been demonstrated that halofuginone is a specific potent inhibitor of mouse and human Th17 differentiation [@pone.0022770-Sundrud1]. To analyze the role of Th17 response in our experimental model of vaginal candidiasis we treated mice with halofuginone every two days, starting two days before the infection ([Fig. 1A](#pone-0022770-g001){ref-type="fig"}). The analysis of IL-17 in the vaginal fluids recovered at different times after infection showed that the halofuginone treatment at a dose of 5 µg/mouse significantly lowered IL-17 production from day 6 to day 21. A dose of 10 µg/mouse completely suppressed IL-17 production ([Fig. 6A](#pone-0022770-g006){ref-type="fig"}). In addition, the fungal load evaluated from the bioluminescence intensity was found to be significantly higher in halofuginone treated mice (5 µg/mouse) with respect to diluent treated mice ([Fig. 6B](#pone-0022770-g006){ref-type="fig"}). This was correlated to a significant increase of CFU (507% fourteen days after infection and 208% twenty-five days after infection) in the vaginal wash of halofuginone treated mice ([Fig. 6C](#pone-0022770-g006){ref-type="fig"}). In further experiments, mice were treated with a higher dose of halofuginone (10 µg/mouse). In this case the increase of CFU was of 608% fourteen days after infection and 462% twenty-five days after infection. In parallel experiments the IL-17 depletion due to halofuginone treatment was restored by intravaginal administration of mouse recombinant IL-17 every 2 days. The fungal load was monitored 4, 8 and 14 days post infection. The bioluminescence and the CFU in rIL-17 treated mice were similar to those observed in the diluent treated mice used as a control group, and not statistically significant ([Fig. 7A](#pone-0022770-g007){ref-type="fig"}).
{#pone-0022770-g006}
{#pone-0022770-g007}
Antimicrobial peptides have an important role in the innate immune response of host cells [@pone.0022770-Cua1] and are secreted particularly by vaginal epithelial cells [@pone.0022770-Ganz1], [@pone.0022770-Ganz2], [@pone.0022770-Liang1]. We analyzed the kinetics of β-defensin 1, 2 and 3 (BD-1, BD-2 and BD-3, respectively) level in the vaginal wash at different days after infection. The results reported in [Figure 7B](#pone-0022770-g007){ref-type="fig"} demonstrate that halofuginone treatment produced a significant down-regulation of the production of BD-2 at day 14 and 21 (*p*\<0.01). In contrast, intravaginal administration of rIL-17 together with halofuginone restored defensin production at day 14 and 21 (*p*\<0.01). BD-1 was not produced. BD-3 was produced 2 and 7 days post infection, and halofuginone was also able to down regulate the production of this peptide. Moreover, the administration of recombinant IL-17 was able to restore the production of the antimicrobial factor. The cytofluorimetric analysis of cells recovered from the vaginal washes showed that the BD2 was produced by epithelial cells ([Figure 7C](#pone-0022770-g007){ref-type="fig"}). The increment of β-defensin 2 and 3 production after rIL-17 treatment of halofuginone treated mice correlated with a more rapid clearance of *C. albicans* (data not shown).
Discussion {#s3}
==========
Mouse models of mucosal candidiasis, including gastrointestinal, oropharingeal and vaginal candidiasis, have provided an invaluable contribution to the understanding of the local immune response to *C. albicans*.
In this study we investigated Th17 responses during vaginal candidiasis and demonstrated that i) vaginal candidiasis leads to a strong influx of neutrophils to the vagina; 75% of the cells present in the vaginal wash soon after the challenge are neutrophils, a percentage which decreased to 60% until the resolution of infection; ii) IL-17 was detected during infection in vaginal fluids, its production reaching a maximum 14 days after the challenge and subsequently decreasing; iii) IL-17 was produced by CD4 T cells in particular, and there was a correlation between the presence of vaginal IL-17 and fungal burden; iv) an appreciable level of IL-23 was observed in the vagina and the amount and kinetics of IL-23 in vaginal fluids were comparable to those in vaginal cells; v) the inhibition of Th17 differentiation leads to significant inhibition of IL-17 production, with consequent exacerbation of infection; and vi) production of beta defensin 2 and beta defensin 3 was manifested in cells of infected mice, this production being strongly reduced when Th17 differentiation was inhibited and stimulated by rIL-17 treatment. Taken together, these results suggest that Th17 responses play a central role in protecting against vaginal candidiasis, possibly through their influence on antimicrobial peptide production by vaginal epithelial cells.
Here we demonstrate that the course and resolution of vaginal infection is strictly correlated to the presence of IL-17 secreted locally by CD4 vaginal T cells. This is consistent with a report showing that vaginal T cells have been characterized and quantified in the lamina propria and the epithelia of the vagina and cervix. It has been estimated that there are about 240 T lymphocytes per mm^2^ of vaginal epithelial tissue [@pone.0022770-Ildgruben1] and it is conceivable that the majority of vaginal T cells migrate to the vaginal epithelium in response to inflammatory chemokines following local antigenic stimulus. This increase of IL-17 in the vaginal wash mirrored the kinetics of IL-17 production by CD4 vaginal cells, however Gr-1 positive cells showed a marginal production of IL-17 too. It is conceivable that an early presence of IL-17 could be due to neutrophils and epithelial cells. This is consistent with previous research showing that IL-17 is also produced by neutrophils [@pone.0022770-Ferretti1] and epithelial cells [@pone.0022770-Cua1]. It is noteworthy that the level of IL-23 was modest with respect to IL-17, in all determinations performed. In addition, the seeming independence of IL-17 from IL-23 could be due to prompt reutilization of IL-23 by T cells. Alternatively, expansion of Th17 cells could occur, even when IL-23 levels are low, as has been described in other experimental models [@pone.0022770-Miyahara1].
Halofuginone, a low molecular weight derivative of the plant alkaloid febrifugine, is able to inhibit collagen α1 gene expression in several animal models of fibrosis and cancer [@pone.0022770-Elkin1], [@pone.0022770-Pines1], [@pone.0022770-Pines2]. Recently, it has been reported that halofuginone inhibits transforming growth factor-β (TGF- β) signaling [@pone.0022770-Pines1] which is known to drive differentiation of Th17 cells in the mouse. It has recently been reported that halofuginone induces selective inhibition of Th17 differentiation and does not influence Th1, Th2, or T~reg~ differentiation [@pone.0022770-Sundrud1]. In our experimental system the infection was exacerbated and IL-17 secretion was inhibited by halofuginone treatment. Compelling evidence revealed a critical role for IL-17 in the induction of natural immune genes, including antimicrobial peptides such as defensins, S100, mucins, etc [@pone.0022770-Fantini1], [@pone.0022770-Raffatellu1], [@pone.0022770-Shen1], [@pone.0022770-Yano1]. In our experimental system beta defensin 1 was not recovered in the vagina, conversely a strong increase of β2 defensin and a significant increase of beta 3 defensin were manifested during the infection. Halofuginone treatment, which strongly impaired IL-17 secretion, also caused a significant decrease of beta defensin 2 and 3. The direct relationship between the presence of IL-17 and beta defensin 2 and 3 was also corroborated by the drastic increase of beta defensin 2 after treatment with rIL-17.
Th17 responses have been shown to be involved in the protective response against fungal and bacterial mucosal infections [@pone.0022770-Happel1], [@pone.0022770-Godinez1], [@pone.0022770-Levitz1]. Current literature attributes a critical role in neutrophil recruitment to IL-17 [@pone.0022770-Aujla1]. In a mouse model of systemic candidiasis a protective role was attributed to IL-17 because of its ability to induce neutrophil recruitment [@pone.0022770-Pirofski1]. The mechanism by which Th17 immunity leads to resistance to OPC involves both IL-17-induced neutrophil recruitment and direct IL-17-induced antimicrobial effects. In our system the increase of IL-17 in the vaginal lumen and its secretion by vaginal cells seems to be independent of the neutrophil influx. As a matter of fact the robust early neutrophil migration observed soon after infection ([Fig. 2](#pone-0022770-g002){ref-type="fig"}) seems mainly attributable to chemotactic molecules, produced by epithelial cells following interaction with *C. albicans* [@pone.0022770-Yano1]. Indeed the level of neutrophils also remained high during the resolution of infection, while the IL-17 production paralleled the course of infection. Given that a correlation between infiltration of polymorphonuclear neutrophils and symptomatic vulvovaginal candidiasis has been observed [@pone.0022770-Fidel6], the lack of correlation between the presence of IL-17 and neutrophil infiltration suggests the role of IL-17 may be to protect from, rather than to participate in, the inflammatory response.
As previously reported, we observed a massive influx of neutrophils to the vagina [@pone.0022770-Fidel5]. These cells were unable to inhibit the growth of *C. albicans* yeast and hyphae formation, despite their primary role as anti-*Candida* effector cells. There are two possible explanations for this inefficiency in the neutrophil performance: one is that the neutrophil influx occurred when the yeast-mycelial transition had already finished, the second is that neutrophils, due to microenvironmental conditions, are unable to inhibit *Candida* growth in the vaginal compartment.
Epithelial cells could indeed strongly contribute to *Candida* clearance [@pone.0022770-Gupta1], [@pone.0022770-Han1] and vaginal epithelial cells are described as having an important role in retarding or arresting *C. albicans* growth in a non inflammatory manner [@pone.0022770-Nomanbhoy1].
Our results show for the first time that a robust Th17 response is found to occur in the vagina during vaginal candidiasis, and IL-17 plays a role in controlling *C. albicans* infection as it induces vaginal epithelial cells to produce antimicrobial peptides.
Materials and Methods {#s4}
=====================
Ethics Statement {#s4a}
----------------
All animal experiments adhered to the EU Directive 86/609. Experiments were performed according to the guidelines of the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes. (ETS No. 123). The protocol was approved by Perugia University Ethics Committee (Comitato Universitario di Bioetica) (permit numbers 41-2005B and 34/2003-A). All efforts were made to minimize suffering during experiments.
Microorganisms {#s4b}
--------------
In this study a novel reporter system for imaging *Candida albicans* infections was used. This luciferase reporter was constructed by fusing a synthetic codon-optimized version of the *Gaussia princeps* luciferase gene to *PGA59* gene of *C. albicans* (strain CA1398), which encodes a glycosylphosphatidylinositol-linked cell wall protein, under the control of ACT1 promoter [@pone.0022770-Enjalbert1]. For experimental infections, cells from stock cultures in YPD agar with chloramphenicol (1% yeast extract, 2% peptone, 2% glucose, all w/v) were grown in YPD medium for 24 h, then harvested by centrifugation, washed and counted as previously described [@pone.0022770-Pietrella1] in an hemocytometer, and resuspended to the desired concentration in sterile physiological saline. *Candida albicans* cells were inactivated by heating at 60°C for 30 minutes.
Mice {#s4c}
----
Female CD1 mice obtained from Harlan Italy Laboratories (Udine, Italy) were used at seven weeks of age. Mice were allowed to rest for 1 week before the experiment; by that time the animals were roughly 8 weeks old.
Animal model of vaginal infection {#s4d}
---------------------------------
The mouse model of vaginal infection has been described previously [@pone.0022770-Enjalbert1], [@pone.0022770-Pietrella1]. Six days prior to infection a pseudo-estrus condition was induced in mice by subcutaneous injection of 0.2 mg of estradiol valerate in 100 µl of sesame oil (Sigma-Aldrich); this was repeated weekly until the completion of the study. Mice anesthetized with 2.5 (v/v) isoflurane gas were infected twice at a 24 h interval with 10 µl of a 10^9^/ml, 5×10^8^/ml or 10^7^/ml of *C. albicans* cell suspensions administered by mechanical pipette into the vaginal lumen, close to the cervix. Following infection the fungal burden was monitored by an in vivo imaging system and by analysis of the vaginal wash ([Figure 1A](#pone-0022770-g001){ref-type="fig"}).
Monitoring of mouse vaginal infection {#s4e}
-------------------------------------
To monitor the course of infection, at selected days post-infection (starting 4 days after the challenge), 10 µl (0.5 mg/ml in 1∶10 methanol∶H~2~0) of coelenterazine (Synchem, OHM) were added to the vaginal lumen. Afterwards, mice were imaged in the IVIS-200TM imaging system under anaesthesia using 2.5% isoflurane. Total photon emission from vaginal areas within the images of each mouse was quantified as previously described [@pone.0022770-Enjalbert1], [@pone.0022770-Pietrella1]. The vaginal lavages were conducted using 150 µl of saline with repeated aspiration for 10 to 20 times and the fluid was serially diluted and plated on YPD agar plus chloramphenicol. CFU were then evaluated and expressed as CFU/ml.
Vaginal cell collection and processing {#s4f}
--------------------------------------
The vaginal washes obtained at different days after infection were treated with protease inhibitors (Complete Protease Inhibitor Cocktail, Sigma-Aldrich) and centrifuged. After centrifugation at 600×g the supernatants were recovered and stored at −20°C and the cells of the vaginal wash were fixed with 4% of PFA. To analyze the polymorphonuclear cell number in the vaginal wash the cells were incubated with rat anti-mouse Ly6G FITC conjugate for 45 minutes on ice. Labelled cells were analyzed by a cytofluorimeter. The cells were first analyzed using FSC/SSC parameter and polymorphonuclear cells were gated according to their size and granularity, then gated cells were tested for Ly6G expression.
After vaginal infection, at different time points, the vaginas were aseptically removed from dead mice; the vaginal tissue was cut longitudinally and minced with sterile scalpel in cRPMI medium consisting of RPMI 1640 supplemented with 10% heat-inactivated fetal calf serum and 100 µl/ml penicillin streptomycin (Gibco).The tissues were digested with cRPMI supplemented with sterile 0.25% collagenase from *Clostridium histolyticum* (Sigma-Aldrich), following incubation for 30 minutes at 37°C [@pone.0022770-DeBernardis1]. After digestion, tissues and cells were filtered with a cell strainer 100 µm (BD Falcon) and washed with RPMI 1640 medium. To evaluate the cytokine production, the vaginal cells were counted by a hemocytometer, left untreated or re-stimulated with 4×10^6^ heat inactivated *C. albicans*, and incubated for 72 hours in cRPMI at 37°C. The supernatants from the culture wells were treated with protease inhibitors and stored at −20°C until the ELISA test, and the vaginal cells were fixed with 4% of PFA.
Lymph node collection and processing {#s4g}
------------------------------------
Vaginal draining lumbar lymph nodes were identified on the posterior abdominal wall lateral to the inferior vena cava and abdominal aorta, respectively. These lymph nodes were excised in cRPMI medium, homogenized and counted with a hemocytometer. To evaluate the cytokine production, the lymph node cells (2×10^6^/ml) were left untreated or re-stimulated with 4×10^6^ heat inactivated *C. albicans* for 72 hours in cRPMI at 37°C. The supernatants from the culture wells were treated with protease inhibitors and stored at −20°C until the ELISA test. The lymph nodes cells were fixed with 4% of PFA.
Detection of intracellular and supernatant cytokines and flow cytometry {#s4h}
-----------------------------------------------------------------------
The supernatants of vaginal washes, and vagina and lymph node cell cultures, obtained from infected and non-infected mice were assayed for the presence of IL-17 and IL-23 cytokines using the enzyme-linked immunosorbent ELISA cytoset (eBioscience and Biosource respectively). The analysis of surface molecules, intracellular IL-17 and β-defensin contents of vaginal and lymph node cells was performed by flow cytometry using standard methodology for direct and indirect immunofluorescence. Briefly, 2×10^5^ cells treated with 10 µg/ml of Brefeldin A (Calbiochem) for 3 h, were incubated with PBS-S buffer (PBS with 0.1% of saponin) for 45 minutes at 4°C with combinations of different antibodies. For IL-17 detection, cells were labelled with rat anti mouse IL-17 PE conjugate (BD Pharmingen) and either rat anti mouse CD4-FITC conjugate (Santa Cruz Biotechnology) or rat anti mouse Gr-1 FITC conjugate; for β-defensin detection, rabbit anti mouse β-defensin 1, goat anti mouse β-defensin 2 or goat anti mouse - β defensin 3 (Santa Cruz Biotechnology), with goat anti-rabbit TRIC conjugate or rabbit anti goat PE conjugate (Sigma-Aldrich), were used. For epithelial cell labelling, a FITC conjugate antibody anti-pan cytokeratin (Sigma) was used. Stained cells were washed with PBS-S buffer, resuspended in fluorescent buffer (PBS with 1% FBS and 0.5% NaN~3~) and analyzed using a FACScan cytofluorimeter (Becton Dickinson, BD). The acquired data were analyzed with CELLQuest software (BD).
Histological analysis {#s4i}
---------------------
For histological evaluation, the mice were sacrificed, and the vaginas were removed and immediately fixed in 10% (v/v) neutral buffered formalin for 24 h. They were then dehydrated, embedded in paraffin, sectioned into 3- to 4-µm-thick sections, and stained with periodic acid-Schiff reagent.
Microscopic analysis of vaginal wash cells {#s4j}
------------------------------------------
The vaginal wash cells were cytospun onto a glass slide (400×g for 7 min) and air dried for 1 h. Cells were then stained using Diff-Quick staining and examined under a light microscope at a magnification of ×40. Neutrophils were identified by their characteristic tri-lobar nucleus.
Halofuginone treatment {#s4k}
----------------------
Mice were injected intraperitoneally with 5 µg/100 µl or 10 µg/100 µl of a solution of halofuginone (obtained from Collgard Biopharmaceuticals (Tel Aviv, Israel) or diluent, two days before and every two days after infection. In parallel experiments and at the same time as the halofuginone treatment, mice were treated intravaginally with 10 µl (10 pg/mouse) of recombinant mouse IL-17 (eBioscience).
Statistical analysis {#s4l}
--------------------
Photon Flux emission, CFU counts, PMN count, IL-17, IL-23 and beta-defensin production were compared using the non-parametric Mann--Whitney U-test. The alpha value was set at 0.05. The correlation between the Total Photons emitted and CFU count in the vaginal wash was assessed with the Pearson\'s correlation test.
We thank Catherine Macpherson for editorial and secretarial assistance.
**Competing Interests:**The authors have declared that no competing interests exist.
**Funding:**This study was funded by the European Commission (FINSysB Marie Curie Initial Training 16 Network, PITN-GA-2008-214004) and by a grant from Fondazione Cassa Risparmio Perugia 2009.010.0492. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
[^1]: Conceived and designed the experiments: DP AR AV FB. Performed the experiments: DP AR NP PM. Analyzed the data: DP AR AV PM. Contributed reagents/materials/analysis tools: MP. Wrote the paper: AV DP CD.
| {
"pile_set_name": "PubMed Central"
} |
Data underlying the study cannot be made publicly available due to legal and ethical considerations. European Union (GDPR) and French (Law n°78-17 of 6 January 1978) laws restrict the public sharing of personally identifiable data. Requests for data will be processed according to the French MR-003 Code of conduct by the data controller, ClinSearch, which allows for the use of data for the purpose of reproducing study results. Requests to access the data for this purpose may be sent to the data protection officer of ClinSearch: <[email protected]> and researchers outside the European Union will need to sign a transfer agreement.
Introduction {#sec005}
============
How many healthcare professionals consider all the characteristics of their patients prior to prescribing or delivering a specific preparation of paracetamol (acetaminophen)? Beyond contraindications, such as avoiding sodium loaded effervescent tablets in patients at risk of hypertension \[[@pone.0221261.ref001]\], or specific excipients due to intolerances, dozens of alternatives are available on the market to select the medicinal product best adapted to obtain an adequate acceptability for the concerned patient.
In the context of medicinal product administration, the earliest occurrence of the word "acceptability" that we have found in Pubmed was published in 1965 \[[@pone.0221261.ref002]\]. Since, the number of manuscripts addressing this concept has steadily increased. Acceptability of medicinal products has emerged as a major factor in compliance, and consequently as a key element for treatment effectiveness, especially in vulnerable populations \[[@pone.0221261.ref003]\].
In their role as a regulatory body, the European Medicines Agency (EMA) included a full section dedicated to acceptability in their 2013 guideline on the pharmaceutical development of medicines for paediatric use \[[@pone.0221261.ref004]\]. Therein, applicants were encouraged to integrate the evaluation of patient acceptability as a vital part of pharmaceutical and clinical development. The EMA has defined patient acceptability as, "The overall ability and willingness of the patient to use and its caregiver to administer the medicine as intended" \[[@pone.0221261.ref004]\]. This implies an interrelationship between users (patients and caregivers) and their characteristics (age, frailty, culture...) on one side, and on the other, the medicine and its characteristics (excipients, size, texture, dose, device...). The combination of all of these characteristics must coalesce to result in proper preparation and administration, and in general a good acceptability of the medicinal product for the user.
The importance of acceptability of medicines is not limited to paediatrics, as highlighted by the 2009 ICH Q8 guideline: "in all cases, the product should be designed to meet patients' needs and the intended product performance" \[[@pone.0221261.ref005]\]. Adequate acceptability is thus crucial for all vulnerable patients, including those of the older population (≥ 65 years of age). The EMA called attention to this issue in 2017 through the draft version of their reflection paper on the pharmaceutical development of medicines for use in the older population \[[@pone.0221261.ref006]\]. Manufacturers and clinicians have thus been encouraged to consider the typology of each user individually to reach an adequate acceptability by proposing well suited treatments.
Acceptability is a multidimensional concept simultaneously driven by a composite of factors belonging to both the user and the medicinal product. Despite the numerous tools used to assess acceptability, a recent review was unable to identify a standardized methodology \[[@pone.0221261.ref007]\]. Furthermore, none of the referenced studies simultaneously considered the different dimensions of acceptability as defined by the EMA guideline \[[@pone.0221261.ref004]\]. Statistical processing of acceptability has generally been restricted to univariate data analysis, even in those cases where several measures were collected and a multivariate approach would have been more appropriate \[[@pone.0221261.ref008]\].
In this study, we have adopted such a multivariate approach, concurrently exploring many different facets of acceptability. Focusing on paracetamol, the active principal ingredient (API) most frequently used in France \[[@pone.0221261.ref009]\], which is also widely used in Europe and throughout the world \[[@pone.0221261.ref010], [@pone.0221261.ref011]\], we have observed and analysed the acceptability of many medicinal products with various dosage forms, in patients ranging from infants to centenarians.
Methods {#sec006}
=======
To assess the multi-faceted concept of acceptability ([Fig 1](#pone.0221261.g001){ref-type="fig"}), we used a novel multivariate approach: CAST---ClinSearch Acceptability Score Test \[[@pone.0221261.ref012]--[@pone.0221261.ref014]\].
{#pone.0221261.g001}
Based on real-life, observed data, the questionnaire examines the following observable contents:
- results of intake (the required dose fully, partly or not taken at all),
- patient reaction during the administration (positive, neutral or negative reaction),
- time needed to prepare (from opening any packaging to having a required dose of medication ready to use, including all handling and modifications),
- time needed to administer the required dose of medication (from a required dose of medication ready to use to the end of the intake),
- recourse to methods to ease/achieve administration (dividing the intake of a dose which cannot be taken as a whole, altering the intended use \[modify dosage form such as tablet crushed or capsule opened; use a device not provided; use another route/mode of administration\], use of food/drink to mask a taste or ease swallowing, patient reward, use of restraint).
Based on data distribution and the clinical practice expertise of the authors, the preparation and administration time (10 second accuracy) was classified as short, medium or long. These scores were defined as less than 20 seconds, from 21 seconds to 1 minute, or longer than 1 minute, in the older inpatient population \[[@pone.0221261.ref014]\], and as less than 1 minute, 1 to 2.5 minutes, or longer than 2.5 minutes, in paediatric patients \[[@pone.0221261.ref012]\].
Each evaluation of one medicinal product, taken by one patient, corresponded to a particular combination of observed measures.
A multiple correspondence analysis (MCA) summarised the key information contained in the data on a 3-dimensional acceptability map, then a hierarchical clustering process on principal components (HCPC) gathered the evaluations into clusters. The "positively accepted" evaluations naturally emerged in a different cluster than those which were "negatively accepted", defining two different acceptability profiles. This acceptability reference framework proposed in a targeted population allows for standardised scoring for medicinal products, and the assessment of both user and medicine characteristics impact on acceptability scores.
This observational study used two acceptability reference frameworks, one for older patients and a second for the paediatric population. The data in this study were exempt from review by an institutional review board as confirmed by one ethics committee (Comité de Protection des Personnes Ile de France VII). Approvals were obtained from the Advisory Committee for Data Processing in Health Research (CCTIRS---\"Comité Consultatif sur le Traitement de l'Information en matière de Recherche dans le domaine de la Santé\") and the French Data Protection Authority (CNIL---\"Commission Nationale de l'Informatique et des Libertés\"). Freely given informed consent has been obtained from every subject (or legal representative) prior participation in the study.
These frameworks were based on large sets of medicinal product use evaluations collected between May 2015 and August 2018. Patients were recruited at random, on a voluntary basis, by many doctors and pharmacists from ten hospitals, eight nursing homes, and more than 150 community dispensaries throughout France. Age was the only eligibility criterion. We also excluded any infusions administered when a catheter was already in place, considering that the insertion of the catheter--which belongs to the administration sequence of the medicine--would be lacking from the data collected. Following study inclusion, the caregiver for outpatients, or the healthcare professional for inpatients, completed the questionnaire related to the observed use of the medicinal product. In cases of polymedication, only the first treatment administered was observed.
To explore the factors influencing acceptability we focused in this study on paracetamol medicinal products.
When a specific medicinal product was evaluated by at least 30 patients in the set of evaluations, the barycentre of these evaluations was positioned on the acceptability map. The acceptability score was structured by the acceptability profile of their barycentre and the proportion of confidence ellipses belonging to it, as previously described \[[@pone.0221261.ref012]--[@pone.0221261.ref014]\]. Similarly, the acceptability of these medicinal products in subpopulations of patients were studied.
Statistical tests were used to assess the significance of the differences observed between the variables for each distinct paracetamol medicinal product. Pearson's Chi-squared test was used when there was a minimum expectation of 5 for 80% of cells without any null expectation \[[@pone.0221261.ref015]\]. When there were fewer patients observations, Fisher's exact test provided an alternative to surmount this difficulty.
R and SAS were used. The R packages FactoMineR \[[@pone.0221261.ref016]\] and MissMDA \[[@pone.0221261.ref017]\] were used to perform MCA and HCPC, and to handle missing data. Data analyses in R and results were checked with SAS 9.4 for patients with complete data.
Results {#sec007}
=======
Populations and paracetamol medicinal products {#sec008}
----------------------------------------------
A total of 2304 evaluations describing the use of many different medicinal products support the acceptability reference framework that was built for each population (1016 in paediatrics and 1288 in the elderly). Among these evaluations, there were 502 observer reports on the intake by an individual patient of one paracetamol product. Among the 37 distinct paracetamol medicinal products assessed, five had ≥30 evaluations.
### Paediatric population {#sec009}
Among the 283 medicinal products assessed in the 1016 evaluations supporting the acceptability reference framework in paediatrics ([S1](#pone.0221261.s001){ref-type="supplementary-material"} and [S2](#pone.0221261.s002){ref-type="supplementary-material"} Tables), there were 18 distinct paracetamol products. These included eight different concentrations (2.4g/100ml, 3g/100ml, 100mg, 150mg, 200mg, 300mg, 500mg and 1000mg) and nine dosage forms (oral suspension, oral solution, powder for oral solution, suppository, capsule, tablet, coated tablet, effervescent tablet and oral lyophilizate), produced by four different marketing authorisation holders. [Table 1](#pone.0221261.t001){ref-type="table"} presents those medicinal products that were assessed in more than five patients. All of the products shown below were manufactured by the same marketing authorisation holder.
10.1371/journal.pone.0221261.t001
###### Demographic characteristics of the paediatric population patients administered the studied paracetamol medicinal products with greater than five assessments.
{#pone.0221261.t001g}
Characteristics of patients 2.4% Oral suspension (n = 199) 300mg Powder for oral solution (n = 39) 500mg Powder for oral solution (n = 22) 100mg Suppository (n = 19) 500mg Capsule (n = 14) 500mg Tablet (n = 8) 100mg Powder for oral solution (n = 7) 150mg Suppository (n = 6)
----------------------------- -------------------------------- ----------------------------------------- ----------------------------------------- ---------------------------- ------------------------ ---------------------- ---------------------------------------- --------------------------- ------
**Gender** Girls 51% 45% 50% 61% 57% 62% 57% 17%
*md*: *7* *md*: *1* *md*: *1*
**Age** (years) \[birth; 2\] 64% 8% 0% 100% 0% 0% 43% 100%
\[3; 5\] 25% 26% 0% 0% 0% 0% 29% 0%
\[6; 8\] 11% 38% 32% 0% 0% 12% 29% 0%
\[9; 11\] 0% 26% 50% 0% 57% 75% 0% 0%
\[12; 14\] 0% 3% 18% 0% 36% 12% 0% 0%
\[15; 17\] 0% 0% 0% 0% 7% 0% 0% 0%
*md*: *6* *md*: *1*
**Place** Hospital 45% 59% 68% 0% 86% 38% 0% 0%
Ambulatory 55% 41% 32% 100% 14% 62% 100% 100%
*md*: *2* *md*: *2*
**Exposure** First intake 11% 15% 18% 37% 29% 0% 29% 17%
*md*: *4*
md: missing data
### Older population {#sec010}
Among the 315 medicinal products assessed in the 1288 evaluations supporting the acceptability reference framework in the older inpatient population ([S3](#pone.0221261.s003){ref-type="supplementary-material"} and [S4](#pone.0221261.s004){ref-type="supplementary-material"} Tables), there were 19 distinct paracetamol products. There were three strengths (10mg/1ml, 500mg and 1000mg), seven formulations (capsule, tablet, coated tablet, effervescent tablet, orally disintegrating tablet (ODT), powder for oral solution and solution for injection), and eight marketing authorisation holders. [Table 2](#pone.0221261.t002){ref-type="table"} presents those paracetamol products assessed in more than five patients, these three paracetamol products all had the same marketing authorisation holder.
10.1371/journal.pone.0221261.t002
###### Demographic characteristics of the older inpatients administered the studied paracetamol medicinal products with greater than five assessments.
{#pone.0221261.t002g}
Characteristics of patients 500mg Capsule (n = 76) 500mg Orally disintegrating tablet (n = 34) 500mg Powder for oral solution (n = 30)
-------------------------------------------------------- ------------------------ --------------------------------------------- ----------------------------------------- -----------
**Gender** Women 71% 74% 63%
**Age** (years) \[65; 75\] 4% 0% 0%
\[75; 85\] 42% 41% 23%
\[85; 95\] 39% 53% 53%
\[95; 110\] 15% 6% 23%
**Place** Hospital 82% 100% 100%
Nursing home 18% 0% 0%
**Exposure** First intake 3% 79% 3%
*md*: *4* *md*: *1*
**Disorders** Swallowing disorder 4% 18% 30%
*md*: *2*
Memory disorder 59% 24% 70%
*md*: *2*
Muscular or rheumatologic disorders of the upper limbs 35% 15% 43%
*md*: *2*
md: missing data
Paracetamol medicinal products acceptability scores {#sec011}
---------------------------------------------------
[Fig 2](#pone.0221261.g002){ref-type="fig"} presents the acceptability scores of the two paracetamol medicinal products that were assessed more than 30 times in the paediatric population: 2.4% oral suspension and 300mg powder for oral solution.
{#pone.0221261.g002}
The greatest number of evaluations was obtained for the 2.4% oral suspension, 77% were gathered into the "positively accepted" profile zone compared to 74% for the 300mg powder for oral solution. The barycentres of their evaluations have been assigned to the \"positively accepted" profile as well as 100% of the confidence ellipses.
[Fig 3](#pone.0221261.g003){ref-type="fig"} presents the acceptability scores of the three paracetamol medicinal products, in the elderly population, that were assessed more than 30 times. These were 500mg capsules, 500mg orally disintegrating tablets and the 500mg powder for oral solution.
{#pone.0221261.g003}
The greatest number of evaluations was obtained for the capsule, 89% were gathered into the "positively accepted" profile zone compared to 88% for the ODT and 73% for the powder for oral solution. The barycentres of all of their evaluations have been assigned to the \"positively accepted" profile, together with 100% of their confidence ellipses, with the exception of a small part of the ellipse, in the third dimension, for the powder for oral solution.
The paracetamol medicinal products assessed in this study appeared to be accepted in both of the populations studied, as the barycentre, along with the entire confidence ellipses, with the exception of the powder for oral solution in the elderly, belonged to the \"positively accepted\" profile. These dosage forms seem appropriate for the patients of these two populations.
Objective comparisons between these different "positively accepted" products are possible; their barycentre positions on the map permit ranking. In this study, although the powder for oral solution was classified as positively accepted by the paediatric and almost similar in the older populations, it was not the best accepted dosage format in either. Its barycentre fell into the furthest limit of the zone of positively connoted categories on the map. For the older inpatients, ODT tended to be more positively accepted than the powder for oral solution, but ODT presented a lower acceptability profile than capsules in this population. These results reflect differences in the observational measures.
Other formulations have not been presented on the map, but even in these cases with fewer than 30 evaluations some tendencies could be observed. The powder for oral solution with a higher strength (500mg), taken by older children, was positioned closer on the map to positively connoted categories, while a lower strength (100mg), taken by younger children, was positioned more distally from them. For those patients under 3, even though the 100mg and 150mg suppository had a better score than the 100mg powder for oral solution, the 2.4% oral suspension of paracetamol was located closest to the ideal position on the map.
Constituting variables {#sec012}
----------------------
Each evaluation reflects the relation of a specific treatment with a specific user. The barycentre of evaluations summarises in one point the variability of uses; it takes into consideration combinations of evaluations which themselves are composed of combinations of measures.
[Table 3](#pone.0221261.t003){ref-type="table"} presents the distribution, per treatment, of the related observational measures.
10.1371/journal.pone.0221261.t003
###### Observational measures per variables.
{#pone.0221261.t003g}
Paediatric population Older population
-------------------------- ----------------------- ------------------ ----------- ----------- ---------
**Result of the intake**
Fully taken 176 (88) 31 (82) 75 (100) 34 (100) 24 (80)
Partly taken 21 (11) 7 (18) 0 (0) 0 (0) 6 (20)
Not taken 2 (1) 0 (0) 0 (0) 0 (0) 0 (0)
*md*: *1* *md*: *1*
**Patient's reaction**
Positive reaction 105 (53) 5 (14) 6 (8) 4 (12) 1 (3)
Neutral reaction 49 (25) 15 (40) 65 (87) 23 (70) 23 (77)
Negative reaction 45 (22) 17 (46) 4 (5) 6 (18) 6 (20)
*md*: *2* *md*: *1* *md*: *1*
**Prep.-admin. Time**
Short time 116 (59) 22 (58) 33 (43) 2 (6) 3 (10)
Medium time 48 (24) 10 (26) 34 (45) 2 (6) 13 (43)
Long time 33 (17) 6 (16) 9 (12) 30 (88) 14 (47)
*md*: *2* *md*: *1*
**Divided dose** 140 (70) 12 (31) 27 (36) 2 (6) 6 (20)
**Alteration** 2 (1) 0 (0) 5 (7) 1 (3) 0 (0)
**Food/drink** 21 (11) 7 (18) 8 (11) 2 (6) 1 (3)
**Reward** 14 (7) 0 (0) **⧫** **⧫** **⧫**
**Restraint** 39 (20) 7 (18) 1 (1) 1 (3) 5 (17)
number and percentages: n(%)---md: missing data -**⧫** Data not collected in older population
As such, observations must take into account those differences in each population, as shown in Tables [1](#pone.0221261.t001){ref-type="table"} and [2](#pone.0221261.t002){ref-type="table"}. Children administered the 2.4% oral suspension dosage form were significantly younger than those administered the powder for oral solution (p\<0.001). In the oldest population, those patients that were administered each of the three paracetamol medicinal products differed for several characteristics. Patients suffering from swallowing disorders were underrepresented in the capsule group and preferentially treated with the powder for oral solution (p\<0.001). Moreover, in nursing homes capsules were the sole dosage form observed on five or more occasions. Those patients treated with ODT had fewer cases of muscular (p = 0.03) or memory disorders (p\<0.001) than patients treated with the other two dosage forms, and they were more often naïve to previous exposure (p\<0.001). This last finding is likely due to the recent introduction of this dosage form at one of the recruiting sites.
Focusing on observations of medicinal product use, the powder for oral solution presented a lower rate of dose fully taken and a higher rate of negative reactions in both populations. The capsule and the ODT were the two formulations that were systematically fully taken in the elderly, while the powder for oral solution was less often fully taken (p\<0.001). For more than 50% of the evaluations of the oral suspension use, observers reported a positive reaction of the patient. This oral suspension is flavoured with strawberry and used principally in children under 6 years old, 89% of the paediatric patients were treated with this medicinal product.
In the older inpatient population, preparation and administration time were significantly different for each medicinal product. This was highlighted by the median times observed: 30 seconds for the capsule, 55 seconds for the powder for oral solution and 115 seconds for the ODT.
Paradoxically the prescribed dose of the most accepted dosage form in both populations was divided significantly more often than other forms (p\<0.001 in the paediatric population; p = 0.003 in the older population). In the youngest population, this was notably observed in hospital: a single bottle of oral suspension was used for several patients, but the syringe was systematically changed. Among the questionnaire choices for methods used to ease/achieve administration, nurses thus entered that the syringe provided by the manufacturer had not been used. However, during the data review it was decided not to consider this as an alteration of intended use: the procedure had clearly been established by hospital staff for hygienic reasons. Moreover, this will facilitate international comparisons, as in many countries syringes are not provided with oral suspensions. By contrast, in cases where parents reported the use at home of a syringe provided with another medicinal product, this has been coded as an alteration of the intended use.
In the older inpatient population, to avoid an interpretation bias related to differences in the strength prescribed, we verified the absence of significant differences in each group. The ratio of patients with a prescription for 2 or more units (capsule, sachet or tablets) was high for all three groups: 89% with capsule (md: 15), 77% with ODT (md: 8) and 86% with powder for oral solution (md: 2) (p = 0.34). Interpretation of the observed number of doses divided must therefore take into account that the majority of the patients were obliged to take 2 unit doses for each intake.
Alteration of the medicinal product prior to administration was only reported in the older inpatient population. For two patients treated in a nursing home the contents of their capsules were mixed once with coffee served with milk, and in the second case with applesauce. For another patient, treated in a hospital, the content of his capsule was mixed in water. The fourth report of alteration concerned a patient that kept the capsule in their mouth waiting for its dissolution. For one patient the use of another route/mode of administration was reported, indeed the ODT was swallowed. No swallowing disorders were diagnosed for any of these patients.
Food or drink taken just before or after the administration to mask the taste or ease swallowing, were used more often in the paediatric population for the powder for oral solution, but not among older inpatients. Whereas, restraint was significantly over represented with the powder for oral solution in our dataset from the elderly population (p = 0.006), effectively the observers had more often reported that patients had to force themselves to manage to drink the solution.
Discussion {#sec013}
==========
The acceptability profiles of five medicinal products of a single active principal ingredient, paracetamol, in four different dosage forms were analysed in two populations at each extremity of the life-cycle, including patients from infants to centenarians. Data from further medicinal products were also collected, providing preliminary results for certain subgroups of patients. Underlying conditions appear to have been the principal driver for caregivers and healthcare professionals in choosing the most appropriate dosage form for older patients, while age appears to have most influenced the choices made for the paediatric patients. Of the five complete medicinal product profiles analysed, powder for oral solution was the least well-accepted in both populations. In the older inpatient population, the most accepted formulation identified was the capsule; it was even used in those 11 (14%) patients over 95 years of age that were still able to consume 2 units in a single sitting. Among those 13% of older inpatients that had swallowing disorders, powder for oral solution seems to have been preferentially prescribed (50%). But as swallowing disorder may be screened by the water-swallowing test, it was not surprising that individuals unable to drink a half-glass of water with no interruptions or difficulties were less amenable to the administration of 1000mg of paracetamol dissolved in a half-glass of water. A limited set of data regarding the recently introduced ODT dosage form implied that this form of administration could represent an interesting alternative in terms of acceptability for patients with swallowing disorders. In the paediatric population, paracetamol oral suspension had the best profile of acceptability. Nonetheless, children over five were quite often switched to powder for oral solution even though acceptability was not improved, especially in hospital. This suggests that healthcare professionals and caregivers categorised certain dosage forms to specific ages in the paediatric population. In this case, however, the summary of product characteristics (SmPC) for the oral suspension of paracetamol indicates that it is appropriate for children weighing from 3kg to 26kg (approximately birth to nine years old), thus providing a single, well-accepted liquid dosage form that may be used until the patient might be switched directly to capsules.
Using our multivariate approach, we were able to test the acceptability of different medicinal products of the same API while simultaneously accounting for individual patient characteristics. Multivariate exploratory data analysis provided an in-depth, nuanced understanding of acceptability. Given similar sample sizes, the larger confidence ellipse for the powder for oral solution revealed a greater heterogeneity of patient evaluations compared with other paracetamol dosage forms. This variability implied that subgroups of patients might be identified. In this case, swallowing disorders were present in 30% of the older population receiving the powder for oral solution dosage form, and preliminary results from patients administered an ODT dosage form suggest that it may offer a better accepted alternative. As this was an observational study, however, we were unable to influence the use of any particular dosage form, and at present additional data from dysphagic patients in the older population will be required to adequately test this hypothesis. Similarly, the powder for oral solution was not optimal in the paediatric population; again due to the study's non-interventional design and to respect the current practice of healthcare professionals and caregivers, our study lacks the necessary evaluations of alternative dosage forms to compare their acceptability in the targeted age range. It must also be noted that the older population studied here consisted exclusively of patients in institutions with a distribution of treatment per unit dose by the internal pharmacy. Thus, the adequacy of the primary and secondary packaging has not been assessed for these patients; instruction leaflets were not read by patients as the medicinal products were prepared by healthcare professionals. In this controlled environment, some potential misuses which could have occurred in the home may not have been captured. It is important to note that acceptability profiles need to be considered with regards to the studied population, in a contextual environment. As such, these results cannot be directly extrapolated to the older home dwelling population, specific observations related to these patients is ongoing but the recruitement is notably slower.
In light of two 2017 reviews of acceptability testing in the paediatric and geriatric populations \[[@pone.0221261.ref018], [@pone.0221261.ref019]\], to the best of our knowledge this is the first study exploring the acceptability of various formulations of an API in real life conditions for both of these age groups. A contemporary review dedicated to the methodology used to assess acceptability of oral paediatric medicines \[[@pone.0221261.ref007]\] further corroborates this. According to Drumond, only "10 studies had used well-defined protocols and observational endpoints to investigate patient appropriateness". Patient reported outcome tools (PRO), which are not recommended for use with children under 6 and questionable before 12 years of age \[[@pone.0221261.ref020]\], were used in many of the paediatric acceptability studies collected in these reviews. Alternatively, proxy questionnaires, which are not accepted by the Food and Drug Administration (FDA) \[[@pone.0221261.ref021]\], were used in other cited studies. In addition, those studies designed with a sole hedonic criterion were susceptible to generating misinterpretations. For example, in the entire population of older inpatients supporting the acceptability reference framework (n = 1288), positive patient reactions were overrepresented (55%) in only one class of treatments, ATC N02AA or natural opium alkaloids (n = 66). After interviewing the nurses involved, we understood that this positive reaction was generated upon announcing their arrival with the patient's opioid, prior treatment administration. Therefore, the hedonic criteria did not measure the acceptability of the treatment with regards to the formulation, but rather the acceptability of the anticipated effect of the treatment. We also identified in the paediatric population that paracetamol 2.4% oral suspension was the sole medicinal product with a majority of positive reactions. This is certainly related to this population's affinity for the strawberry aroma of this dosage form. As this could lead to unintended administrations, the health authorities prefer neutral reactions. By collecting many different variables for an analysis without weighting, this multivariate approach is able to simultaneously capture the range of factors comprised in the concept of acceptability. Three-dimensional mapping of these results places their barycentre position within a cluster profile that draws upon a reference framework, whereas in the literature acceptability is frequently limited to sequential comparisons of each specific criterion \[[@pone.0221261.ref018]\]. While the EMA guideline on pharmaceutical development of medicines for paediatric use states that \"adequate patient acceptability is not to be understood as 100% acceptability of a medicine\" \[[@pone.0221261.ref004]\], an 80% threshold appears to have been generally employed to date \[[@pone.0221261.ref007]\]. The use of aggregate scoring of acceptability criteria presented in this study better facilitates decisions regarding each medicinal product. If we take into consideration this study, the percentage of multivariate assessments per medicinal product falling within the "positively accepted" cluster failed to meet this 80% threshold in three of the five cases. As the barycentre positions and the confidence ellipses of all five of these dosage formulas belong to the cluster "positively accepted", we consider that these multivariate analyses circumvent the need to assign an artificial threshold for acceptability as would be required when considering unidimensional criteria.
Paracetamol, with 124 marketed medicinal products in France, accentuates these challenges. Healthcare professionals must match the optimal set of characteristics to select a dosage form best adapted to individual patients of different ages, and each with a different set of underlying conditions. With this study, we set out to draw the attention of both clinicians and medical institutions to certain factors of particular interest. At the institutional level, we identified more than 16 paracetamol medicinal products referenced on hospital prescription lists, some of which were poorly adapted for the population targeted by the concerned facility. As discussed above, powder for oral solution was often administered to patients from the older inpatient population with a swallowing disorder, whereas our results indicate that it may not be the most appropriate dosage form. In a pragmatic approach to better understand the inclusion of such dosage forms in hospital dispensaries, working groups including all the actors of the hospital (i.e., nurses, physicians, pharmacists, administration and patient representatives) must be consulted. Observational studies might then be conducted to verify if any related or alternative dosage forms might be more accepted in the targeted population, and confirm their presence on the prescription list. Those dosage forms identified to cover the unmet need of a specific subpopulation, could subsequently be integrated into the institution's standard practice through internal communication and training. A discrepancy in the use of administration devices was also uncovered by our study. For hygienic reason, additional oral administration syringes were purchased separately: in one hospital their volume was inferior to the original device, covering patient weighing from 3 kg to 8 kg instead of 3 kg to 13 kg. Unfortunately, 43% (n = 32/74) of the doses administered in this hospital were prescribed for patients \>8 kg. The selection of this device imposed time-consuming additional handlings on the nursing staff, and more critically brought in a new risk factor for dosage error. We were not able to investigate the purchasing process of this syringe, but prior consultation of a multidisciplinary group would have improved the final process of oral suspension treatment administration in this setting.
A number of observations made in this non-interventional study call for further inquiry. Preliminary observations here revealed that ODT tended to be more suitable than the powder for oral solution in patients with swallowing disorders, a common complaint in older patients \[[@pone.0221261.ref022], [@pone.0221261.ref023]\]. In spite of a longer administration time, this dosage form was mainly related to positively connoted categories, but more evaluations in these patients with swallowing disorders are required. Ideally, this treatment option would spare individuals with dysphagia from the administration of powder for oral liquid solution. As some ODT contain sodium, however, potential cardiovascular risk contraindications among the older population must be considered prior to selecting which paracetamol ODT to stock. In parallel, two-thirds of the older patients treated with powder for oral solution had no swallowing disorder, we would be interested in verifying the acceptability the capsule dosage form in this population. Based on French public prices this would reduce these patients' paracetamol budget by up to 25%. In the paediatric population, the place of the powder for oral solution might also be questioned: in this study, this dosage form was administered in children over nine that should be capable of swallowing capsules, and in children between three to five years eligible for treatment with the better accepted oral suspension. To address these questions, we have planned an interventional study to be conducted in hospitals, for both populations. Administration of a randomized adapted dosage form of paracetamol to the non-contraindicated population would be alternated over different weekdays using a cross-over methodology.
Conclusions {#sec014}
===========
Appropriate prescription of medicines extends beyond API and dosage levels. Although any number of dosage forms of an API may be accepted in the general population, additional care must be taken to integrate individual patient characteristics when prescribing for vulnerable populations. Selecting an appropriate adapted dosage form permits healthcare professionals to improve acceptability. In the case of paracetamol, institutions could be aided to reduce the number of medicinal products stocked in their dispensaries while retaining the necessary alternative dosage forms for vulnerable subpopulations. Further well-designed investigations will be vital to facilitate these processes to better serve these patients.
Supporting information {#sec015}
======================
###### Demographic characteristics of the patients from the older population.
(DOCX)
######
Click here for additional data file.
###### Characteristics of the medicines assessed in the older population.
(DOCX)
######
Click here for additional data file.
###### Demographic characteristics of the patients from the paediatric population.
(DOCX)
######
Click here for additional data file.
###### Characteristics of the medicines assessed in the paediatric population.
(DOCX)
######
Click here for additional data file.
The authors gratefully acknowledge all patients for their participation. We would like to thank all members of the R2G working group (\"Rencontres Galénique & Gériatrie\") for their continuous support and their help in data collection as well as all healthcare professionals working at the hospitals (Hôpital Jean-Verdier---93140 Bondy; Hôpital Rothschild---75012 Paris; Hôpital Joffre Dupuytren---91210 Draveil; Hôpital Broca---75013 Paris; Hôpital Sainte-Périne---75016 Paris; Hôpital Vaugirard---75015 Paris; Hôpital Fernand Widal---75010 Paris; Hôpital René Muret---93270 Sevran; Centre Hospitalier Ouest Vosgien---88300 Neufchâteau; Centre Hospitalier de Versailles André Mignot---78150 Le Chesnay), the nursing homes (Résidence Degommier---91590 Cerny; Résidence Vauban Fondation Claude Pompidou---90000 Belfort; Maison de retraite protestante---92000 Nanterre; Maison du parc---75013 Paris; Résidence John Lennon---95370 Montigny-lès-Cormeilles; Korian Les Arcades---75012 Paris; Korian Nymphéas Bleus---27200 Vernon; Korian Samarobriva---80000 Amiens), and the community dispensaries involved in the study.
In addition, we would like to thank the ClinSearch research associates who collaborated in the data collection, and Christopher Jarvis, medical writer at ClinSearch, for linguistic revision.
[^1]: **Competing Interests:**The authors have read the journal's policy and the authors of this paper have the following competing interests: FR and TV are paid employees of ClinSearch. This alters our adherence to PLOS ONE policies on sharing data and materials. Please see the Data Availability statement for further information. There are no patents, products in development or marketed products to declare.
| {
"pile_set_name": "PubMed Central"
} |
All relevant data are within the paper and its Supporting Information files.
Introduction {#sec001}
============
Gastric cancer is the fourth most common cancer and the second leading cause of cancer related death in the world \[[@pone.0128751.ref001]\]. Its morbidity and mortality is particularly pronounced in Asian countries due to a variety of influences \[[@pone.0128751.ref001]\]. Since its presentation is often associated with advanced disease, there is an urgent need for advances in its detection and ultimately its management. At present, the understanding of microRNAs(miRNAs) on influencing gastric cancer formation is occurring at a rapid pace.
Since the first description of miRNA in the nematode *C*. *elegans* back in 1993, the impact of these small non-coding RNAs has transcended multiple branches of molecular biology \[[@pone.0128751.ref002]\]. MiRNAs are highly tissue specific biomarkers with the potential to alter and transform resident tissue. Because overexpression and under-expression have both been associated with tumorigenesis \[[@pone.0128751.ref003]\], their roles as oncogenes and tumor suppressor genes are both well-established \[[@pone.0128751.ref004], [@pone.0128751.ref005]\]. Over the last several years, their impact on development and detection of solid organ tumors including gastric cancer is slowly being elucidated. There are already several miRNAs identified in the gastric cancer anti-apoptotic mechanism such as miR-21 and miR-148a \[[@pone.0128751.ref006], [@pone.0128751.ref007]\]. Other pathways influenced by miRNAs include cell cycle progression comprising of miR-222/221, miR-106b/93/25 and miR-24 \[[@pone.0128751.ref006], [@pone.0128751.ref007]\].
One of the other promising new miRNAs for solid organ tumors includes miR-206 \[[@pone.0128751.ref008]\]. This particular miRNA belongs to a group of "myomiRs" that is involved in skeletal muscle development \[[@pone.0128751.ref009]\]. Having been associated with numerous other diseases including heart disease, chronic obstructive pulmonary disease and Alzheimer's, its role in oncogenesis received scrutiny more recently including rhabdomyosarcoma, lung cancer, colorectal cancer, schwannoma, and gastric cancer \[[@pone.0128751.ref008], [@pone.0128751.ref009]\]. Although elevated in a few types of cancer including ovarian and Waldenstrom macroglobulinemia, miR-206 is mostly suppressed in solid organ tumors \[[@pone.0128751.ref009]\]. miR-206 has previously been shown to inhibit gastric cancer proliferation in part by suppressing cyclin D2 \[[@pone.0128751.ref010]\]. In this investigation, we concentrated on the role of miR-206 in gastric cancer oncogenesis through the c-Met pathway, which has traditionally been an influential signaling pathway for oncogenesis in a variety of tumors \[[@pone.0128751.ref011]\]. c-Met has been predicted and shown to be the target gene of multiple miRNAs including miR-206 \[[@pone.0128751.ref009], [@pone.0128751.ref012]\].
Results {#sec002}
=======
Suppression of miR-206 led to increased c-Met expression in gastric cancer {#sec003}
--------------------------------------------------------------------------
Real-time RT-PCR analysis was performed to detect the expression of miR-206 in 40 gastric cancer specimens and normal tissues. miR-206 levels in most tissue samples of gastric tumor (34/40) were found to be significantly lower than normal tissues ([Fig 1A](#pone.0128751.g001){ref-type="fig"}). miR-206 expression was inversely related to the level of c-Met observed in tumor samples ([Fig 1B](#pone.0128751.g001){ref-type="fig"}). Most tumor samples, with decreased miR-206 expression, showed high percentage (\>50%) of c-Met staining. Conversely, tumors with normal expression of miR-206 showed very weak or negative c-Met expression.
{#pone.0128751.g001}
miR-206 induced G1 arrest and inhibited cell proliferation, migration and invasion of AGS gastric cancer cells {#sec004}
--------------------------------------------------------------------------------------------------------------
As miR-206 expression was decreased in gastric cancer specimens, we sought to determine whether the introduction of miR-206 had any biological effect on AGS cells. AGS cells transfected with the miR-206 molecule showed inhibition of cell growth as compared to negative control based on the MTS assay ([Fig 2A](#pone.0128751.g002){ref-type="fig"}). FACS analysis of the cells showed G1 cell cycle arrest ([S1 Fig](#pone.0128751.s001){ref-type="supplementary-material"}). The number of colonies was also reduced with transfection of miR-206 ([S2 Fig](#pone.0128751.s002){ref-type="supplementary-material"}).
{#pone.0128751.g002}
miR-206 can inhibit migration and invasion of AGS cells ([Fig 2B](#pone.0128751.g002){ref-type="fig"} and [S3 Fig](#pone.0128751.s003){ref-type="supplementary-material"}). A dramatic reduction of migration towards the lower chambers was observed in miR-206 transfected AGS cells (86 ± 15 vs. 165 ± 16 in AGS cells, *P* \< 0.01, n = 3). In addition, cells transfected with miR-206 showed that HGF-induced invasiveness was also significantly hampered following miR-206 transfection (57 ± 12 vs. 116 ± 14 in AGS cells, *P* \< 0.01, n = 3).
miR-206 downregulated c-Met expression and other cell cycle-related proteins {#sec005}
----------------------------------------------------------------------------
We have previously identified c-Met as a direct target of miR-206 \[[@pone.0128751.ref009]\]. Western blot analysis confirmed that c-Met expression was reduced by miR-206 transfection in AGS cells ([Fig 3](#pone.0128751.g003){ref-type="fig"}). Concurrently, ectopic miR-206 also down-regulated the expression of CDK4, p-Rb, p-Akt and p-ERK.
{#pone.0128751.g003}
Downregulation of c-Met inhibited gastric cancer cell proliferation, migration and invasion {#sec006}
-------------------------------------------------------------------------------------------
Next, c-Met specific siRNA was first used to decrease the expression of c-Met in AGS cells ([S4 Fig](#pone.0128751.s004){ref-type="supplementary-material"}). MTS assays were performed to detect the proliferation of cells. AGS cells transfected with c-Met siRNA showed reduced cell growth as compared to negative control cells ([Fig 4A](#pone.0128751.g004){ref-type="fig"}; 25.10 ± 3.81% decrease). Both HGF-induced migration and invasion were decreased when comparing c-Met siRNA transfected cells to negative control transfected cells. As indicated in [Fig 4B](#pone.0128751.g004){ref-type="fig"} and [S5 Fig](#pone.0128751.s005){ref-type="supplementary-material"}, decrease in migration (88 ± 10 vs. 155 ± 15, P \< 0.01, n = 3) and invasion (72 ± 7 vs. 126 ± 12, P \< 0.01, n = 3) were both statistically significant.
{#pone.0128751.g004}
Introduction of miR-206 suppressed tumor growth *in vivo* {#sec007}
---------------------------------------------------------
We next investigated if overexpression of miR-206 could repress tumor growth *in vivo*. After 8 weeks, the averaged tumor volumes were significantly lower from cells infected with lentivirus expressing miR-206, as compared with control ([Fig 5](#pone.0128751.g005){ref-type="fig"}).
{#pone.0128751.g005}
Discussion {#sec008}
==========
Gastric cancer has remained a significant health care burden worldwide despite years of research \[[@pone.0128751.ref001]\]. According to the WHO, gastric cancer remains one of the top cancer etiologies with a high rate of mortality \[[@pone.0128751.ref001]\]. As with other solid organ tumors, it is increasingly clear that better understanding of tumor oncogenesis is necessary to improve the prognosis of these patients. Being prevalent in Asian countries, data are emerging on the role of miRNAs on the development or suppression of gastric cancers \[[@pone.0128751.ref006]\].
miRNAs have dual functions in terms of gastric cancer development \[[@pone.0128751.ref006], [@pone.0128751.ref013]\]. Some miRNAs are tumor suppressors and others are oncomiRs \[[@pone.0128751.ref006]\]. Ueda et al. found 22 miRNAs upregulated and 13 downregulated in the plasma of patients with gastric cancer \[[@pone.0128751.ref013]\]. Targets that are under investigation in the case of gastric cancer include regulators of p16, via miR-24, and p21 through miR-222/221 and miR-106b/93/25 \[[@pone.0128751.ref006]\]. Others, such as miR-21, can be upregulated in up to 92% of gastric cancer tissue samples \[[@pone.0128751.ref014]\]. Candidates identified in gastric cancer serving as tumor suppressors include miR-181b, miR-101 and miR-486 \[[@pone.0128751.ref006]\]. For instance, miR-101 is thought to target EZH2, Cox-2, Mcl-1 and Fos \[[@pone.0128751.ref015]\]. miR-486 is thought to affect OLFM4, possibly affecting apoptosis downstream \[[@pone.0128751.ref016]\].
Having established that the majority of miRNAs serve as tumor suppressors, we investigated the c-Met pathway in gastric cancer. In studying the c-Met pathway for rhadbomyosarcoma, we found the importance of this pathway in sarcoma \[[@pone.0128751.ref009]\]. c-Met is known to be dysregulated in gastric cancer \[[@pone.0128751.ref011], [@pone.0128751.ref017]--[@pone.0128751.ref019]\]. The MET proto-oncogene encodes a protein known as hepatocyte growth factor (HGF) receptor which possesses tyrosine kinase activity \[[@pone.0128751.ref018]\]. We normally find it expressed only in stem cells but have also seen its dysregulation in oncogenesis \[[@pone.0128751.ref018]\]. In this study, we were able to confirm that c-Met is significantly involved in gastric cancer and its role as a miR-206 target is pivotal in oncogenesis.
Cell cycle progression is dysregulated in gastric cancer. Based on previous reports, cyclin D2 is elevated in gastric cancer when miR-206 is affected \[[@pone.0128751.ref010]\]. miR-206 has been shown to be a potent prognostic marker \[[@pone.0128751.ref010]\]. Its downregulation is associated with shorter overall survival. Restoration of miR-206 leads to G0/G1 cell cycle arrest, confirming its role as a tumor suppressor. Our work also showed cell cycle dysregulation with CDK4 and phosphorylated-Rb being affected. CDK4 is a member of the cycline dependent kinase family with ser/thr protein kinase activity. It leads to G1 phase progression. In addition, the kinase leads to the phosphorylation of Rb, which is a tumor suppressor involved in cell cycle progression.
Although similar findings have been confirmed in gastric, ovarian and breast cancers, the role of miR-206 seems to have an opposite effect in colon cancer \[[@pone.0128751.ref006], [@pone.0128751.ref020]\]. An inverse correlation was noted between miR-206 and KLF4 in a panel of human colon cancers \[[@pone.0128751.ref020]\]. KLF4, which has oncogenic properties in other cancers such as breast, skin and lung, functions as a tumor suppressor in colon cancer \[[@pone.0128751.ref020]\]. Its promoter is hyper-methylated with elevated levels of miR-206 leading to downregulation of KLF4 \[[@pone.0128751.ref020]\]. These findings indicate that miR-206 may not function similarly in all epithelial cells, which negates the one size fits all approach in chemotherapeutics.
In the present study, we identified a mechanism for regulation of c-Met gene expression through miR-206 in gastric cancer. c-Met overexpression following miR-206 downregulation seems to be the common etiology for the pathogenesis of gastric cancer in the majority of samples examined in this study. In summary, we demonstrated that miR-206 negatively modulates the c-Met signaling pathway involved in cell proliferation and migration. Our studies will hopefully have important clinical consequences in the treatment of gastric cancer. miR-206 is a candidate for both immunohistochemical detection of small tumors and possible target for biological therapeutics.
Materials and Methods {#sec009}
=====================
Cell culture {#sec010}
------------
The human gastric cancer cell line, AGS, purchased from ATCC (Manassas, VA), was grown in Ham\'s F-12 Medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS; Hyclone, Logan, UT).
Ethics statement {#sec011}
----------------
This study was carried out in strict accordance with the recommendations and approval of the Wenzhou Medical University Animal Care and Use Committee (Permit Number: WZMCOPT-043011). Forty gastric tumor specimens and normal donor gastric tissues were obtained from the second affiliated hospital of Wenzhou Medical University (Wenzhou, China). Sample collection was approved by the Wenzhou Medical University Ethics Committee on research involving human subjects, and written informed consent was obtained from each case. All experiments were performed in compliance with the Helsinki Declaration and national laws.
Quantitative RT-PCR {#sec012}
-------------------
Total RNA was extracted from human gastric tumor samples and normal controls with Trizol reagent (Invitrogen). 10 ng of total RNA were used for cDNA synthesis by the Taqman MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA), and miR-206 expression level was quantified by the Taqman MicroRNA Assay (Applied Biosystems). Real-time RT-PCR was performed using the Applied Biosystems ViiA 7 Real-Time PCR System (Applied Biosystems).
Immunohistochemical staining of c-Met {#sec013}
-------------------------------------
Sections of formalin-fixed paraffin-embedded human gastric tumor specimens (5 μm) were made and then de-paraffinized with xylene and ethanol. Slides were treated with 0.3% hydrogen peroxide and then incubated overnight at 4°C with c-Met antibody at 1:300 dilution (CST, Beverly, MA). Immunohistochemical staining was performed using the EnVision HRP/DAB detection system (Dako, Glostrup, Denmark).
Cell proliferation assay {#sec014}
------------------------
AGS cells were plated at 2000 cells per well in 96-well plates (Costar, High Wycombe, UK) for each transfection. Transfections were performed with reagent (Lipofectamine RNAiMAX; Invitrogen), in triplicates. For each well, 50 nM miR-206 mimics molecule (Ambion, Austin, TX), or a negative control (Ambion) transfection was employed. After 72 hours of culture, cell proliferation was assessed by MTS assay (CellTiter 96 AQueous; Promega, Madison, WI).
Transwell migration assays {#sec015}
--------------------------
24 hours following transfection, AGS cells were harvested by trypsinization and washed once with D-Hanks solution (Invitrogen). To measure cell migration or invasion, 8-μm pore size culture or Matrigel inserts (Transwell; Costar) were placed into the wells of 24-well culture plates. In the lower chamber, 400 μL F-12 containing 10% FBS and 20 ng/ml of HGF was added. Then, 5x10^4^ cells were added to the upper chamber. After 20 hours of incubation, the cells that had migrated through the pores were stained with crystal violet and observed under the microscope (Zeiss, Oberkochen, Germany) using a 20X objective.
Western blot analysis {#sec016}
---------------------
AGS cells (1x10^5^) were seeded in 6-well plates and grown in F-12 for 24 hours prior to transfection. 72 hours after transfection, the cells were washed with cold PBS and subjected to a lysis buffer (35 mM Tris-Cl \[pH 6.8\], 20 g/L sodium dodecyl sulfate \[SDS\], 100 mM dithiothreitol). Protein lysates (20 μg each) were separated using 8% SDS-polyacrylamide gel electrophoresis, then electro-transferred onto nitrocellulose filter membranes. The membranes were blocked with a buffer containing 5% nonfat milk in PBS with 0.05% Tween-20 for 2 hours and incubated overnight with antibody at 4°C. After a second wash with PBS containing 0.05% Tween-20, the membranes were incubated with peroxidase-conjugated secondary antibodies (Millipore, Darmstadt, Germany) and developed with an enhanced chemiluminescence detection kit (Pierce, Rockford, IL). GAPDH was used as a loading control. Antibodies for CDK4, p-Rb, ERK, Akt, p-ERK, p-Akt, c-Met and GAPDH were from Cell Signaling Technology (Beverly, MA).
SiRNA assays {#sec017}
------------
c-Met-specific siRNA (Ambion) and negative control siRNA (Ambion) were used to downregulate c-Met expression in AGS cells. 50 nM of c-Met-specific siRNA or negative control siRNA was transfected into AGS cells with Lipofectamine RNAiMAX. MTS assay was carried out 72 hours after transfection, whereas Transwell and Matrigel assays were performed 24 hours after transfection, as described above.
*In vivo* tumor growth assay {#sec018}
----------------------------
The pre-microRNA expression constructs lenti-miR-206 and pCDH-CMV-MCS-EF1-copGFP control vector were purchased from System Biosciences (Mountain View, CA). AGS cells were infected with lentivirus expressing miR-206 or negative control. Female nude mice, 6 weeks of age, were inoculated with AGS cells (8x10^6^) expressing miR-206 or negative controls in their flanks, and then sacrificed after 8 weeks. Tumor size was measured and volume was calculated using the formula: (*L* x *W* ^2^) x 0.5, (*L*, length; *W*, width), according to the method previously reported \[[@pone.0128751.ref021]\]. All studies and procedures were approved by the Wenzhou Medical University Animal Care and Use Committee.
Statistical analysis {#sec019}
--------------------
All data were shown as the mean ± SEM. Results shown are expressed as the mean value ± SEM of the results obtained from triplicates in one experiment. Results represent those obtained in three separate experiments. Differences between experimental groups and control groups were analyzed using the Student's *t*-test. Statistical significance was accepted at *P* \< 0.05.
Supporting Information {#sec020}
======================
###### FACS analysis of AGS cells transfected with miR-206.
AGS cells were collected 48 hours after transfection with miR-206 or NC, stained with propidium iodide, and analyzed by flow cytometry. Ten thousand cells were evaluated in each sample. The most representative results in three independent experiments are depicted.
(TIF)
######
Click here for additional data file.
###### Colony formation assay.
AGS cells transfected with miR-206 or NC were seeded at low density. After 7 days, colony formation was determined by staining with crystal violet. Typical results in three independent experiments are shown.
(TIF)
######
Click here for additional data file.
###### The effects of miR-206 on AGS cells were assessed with Transwell and Matrigel assays.
The number of cells that had migrated through the culture insert pores (up) or had invaded through the Matrigel insert pores (down) was photographed using a 20X microscope objective.
(TIF)
######
Click here for additional data file.
###### Downregulation of c-Met by siRNA.
Western blot analysis was performed to confirm suppression of c-Met expression after lipofectamine transfection of AGS cells with either c-Met siRNA or a negative control (NC).
(TIF)
######
Click here for additional data file.
###### The effects of c-Met siRNA on AGS cells were assessed with Transwell and Matrigel assays.
The number of cells that had migrated through the culture insert pores (up) or had invaded through the Matrigel insert pores (down) was photographed using a 20X microscope objective.
(TIF)
######
Click here for additional data file.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: ZZ DY XD. Performed the experiments: DY XC GL. Analyzed the data: ZZ DY XC GL DZ LZ XD. Contributed reagents/materials/analysis tools: ZZ DY HH KC. Wrote the paper: ZZ DY LT XD.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION {#sec1-1}
============
Human embryonic stem cells (hESCs) can be maintained in culture in a self-renewing state and differentiate into all three embryonic germ layers. Although they hold great promise for regenerative medicine, hESC-based therapy faces several challenges. Concerns have been raised regarding their genetic stability. Although not fully understood, challenges also exist over epigenetic stability. Epigenetic changes make nuclear program altered without changing the primary DNA sequence. Because epigenetic changes can substantially modify cellular behavior and are mitotically and meiotically heritable, investigation of the epigenetic properties of human hESC is desirable prior to considering their use in vivo. Epigenetic state is one of the hESCs states that enables stem cells with the unique properties to self renew or differentiate into any cell type in the body. The hESC state may be influenced by the manner in which ESCs are derived and maintained. Recent studies have showed that the efficiency of induced pluripotent stem (iPS) cells formation is enhanced upon addition of valporic acid, an inhibitor of histone deacetylases, to the culture medium. Sodium butyrate, a naturally occurring short-chain fatty acid, supports the extensive self-renewal of mouse embryonic stem cell (hESCs) and[@CIT1]
X-chromosome inactivation (XCI) phenomenon has been used to examine the epigenetic stability of hESC. Because XCI is one of the first measurable epigenetic changes in the early mammalian embryo and is coincident with differentiation, XCI marker serves as an excellent tool to investigate the epigenetic behavior of hESC. XCI is a mechanism to compensate gene load difference between XY males and XX females in mammals. During early embryogenesis, one of two X-chromosomes in every cell is inactivated, and stably inherited through cell division of somatic cells.[@CIT2] XIST makes a noncoding RNA required to initiate silencing during XCI. Before XCI in mESCs, *XIST* is expressed at low level. Upon cell differentiation and the onset of XCI, XIST RNA is transcriptionally induces and forms a cloud around the inactive X (Xi). However, it is known that the two X chromosomes are active in oocytes, indicating that the inactive X chromosome must be reactivated during germ cell development. The reactivation of inactive X chromosome occurs at least twice during mammalian development, once in the epiblast cell lineage at the peri implantation stage and once in the primordial germ cells (PGCs) at the midgestation.[@CIT2] The dynamics of X chromosome activity is tightly correlated with major genomic reprogramming events occurring during mammalian development.
Recently, 11 distinct hESC lines have been studied in order to investigate their epigenetic properties by using XCI markers mainly studying *XIST* expression. Unlike mESCs, hESCs are pre-XCI and there is variability of *XIST* expression among different hESC lines.[@CIT3] These cell lines can be subgrouped into three classes. Class I line has the capacity to recapitulate XCI when induced to differentiate in culture. Class II cells have already undergone XCI. In class III cell lines, despite losing tirmethylation of histone H3-K27 (H3K27me3), there is a tendency to lose XIST RNA expression.
Histone lysine methylation has been shown to index silenced chromatin regions at pericentric heterochromatin or of the inactive X chromosome.[@CIT4] H3K27me3 is a repressive chromatin mark. Recently, 3- deazaneplanocin A (DZNep) was discovered to selectively inhibit H3K27me3. DZNep affects multiple histone methyltransferases and can epigenetically reactivate a different cohort of genes.[@CIT5]
In the current study, our goal was to reprogram class II cell lines to class I by using DZNep as H3K27me3 methyltransferase inhibitor, and sodium butyrate as histone deacetylase inhibitor which enhances self-renewal status of embryonic stem cells and also up-regulates germ-cell specific markers.[@CIT1] We used three Class II hESC lines: HSF6-8, HSF6-10, and HSF6-S9.
METHODS {#sec1-2}
=======
{#sec2-1}
### Cells and Drug Treatment {#sec3-1}
Initial lines of hESC (HSF6-8, HSF6-10, and HSF6-S9) were cultured on a feeder layer of mouse embryonic fibroblasts (MEF). Human ESC culture medium (hESM) was consisted of DMEM/f12 supplemented with 100ml KSR, 2.5 ml L-glutamine, 500 μl BME, and 5 ml nonessential amino acids. Then cells were treated with 0.1 μM DZNep and 0.2 mM sodium butyrate.
### Histone Protein Extraction {#sec3-2}
Human Embryonic stem cells were harvested and washed twice with ice-cold PBS. Cells were re-suspended in Triton Extraction Buffer (TEB: PBS containing 0.5% Triton X 100 (v/v), 2mM phenylmethylsufonyl fluoride (PMSF), 0.02% (w/v) NaN2) with a cell density of 10^7^ cells per ml. Cells were lysed for 10 min using gentle stirring, and centrifuged at 6500× g for 10 minutes at 4 degree C to spin down the nuclei. Pellet was re-suspended in 0.2 N HCl at a density of 4x10^7^ nuclei per ml. The extract was stored overnight at 4 degree C. Samples were centrifuged at 6500xg for 10 minutes to pellet debris. Supernatant was stored and the histone concentration was detected by Commassie Blue.
### Western Blots {#sec3-3}
Cells were harvested by treatment with Try-pLE express. Cells were lysed using 1.0 ml hypotonic lysis buffer (10mM Tris-HCl, 1.5mM MgCl~2~ 10mM KCl, 0.34M sucrose, 10% glycerol,40μl of PI stock solution, and 1.0 μl of DTT). Nuclear histones were separated on 16% SDS-polyacrylamide gels at 200v, 150mA for an hour and stained with commasie blue. Equal amounts of protein were separated in SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membranes. The blots were probed with antibodies against histone H3, H3K27me3, and H3K4me3.
RESULTS {#sec1-3}
=======
{#sec2-2}
### Differential Expression of XCI Marker in Cultures of Female hESC Lines {#sec3-4}
Using immunostaining of H3K27me3, cultures of female HSF6-8, HSF6-10, and HSF6-S9 cells showed 0-50% of XCI marker. Cells were stained with DAPI (a dye binding to DNA), Oct4 (a stem cell marker), and H3K27me3. In a subset of stem cell population, H3K27me3 accompanied Xi, however, some H3K27me3 didn't accompany Xi ([Fig. 1](#F0001){ref-type="fig"}). The absence of H3K27me3 marker is not indicative of X chromosome loss. X chromosome DNA FISH in a subset of cells which didn't have H3K27me3 enrichment along with Xi, showed the presence of two X chromosomes ([Fig. 2](#F0002){ref-type="fig"}).
{#F0001}
{#F0002}
### Optimization of DZNep Concentration and Colony Morphology {#sec3-5}
In cancer treatment, DZNep is used with high concentration (10μM) in order to induce apoptosis in cancerous cells, and reactivate genes which have became silent by cancer development. Therefore, in the current study the DZNep concentration should have been optimized. NaBu was added at 0.2 mM during three passages and then different concentrations of DZNep were added. The optimal concentration for growing colonies of DZNep seemed to be 0.1 μM ([Fig. 3](#F0003){ref-type="fig"}). Healthy, undifferentiated HSF6-8, HSF6-10, and HSF6-S9 colonies treated with 0.2μM NaBu, and 0.1μM DZNep seemed to have well-defined uniform borders and the individual cells within the colony appeared to be similar. More cells per colony were observed compared to the untreated colonies ([Fig. 4](#F0004){ref-type="fig"}).
{#F0003}
{#F0004}
### hESCs Treated With DZNep Show Tendency to Lose XCI H3K27me3 Marker {#sec3-6}
Immunostaining analyses showed that cells tend to lose H3K27me3 enrichment associated with Xi by several passages. This effect was prominent when NaBu was added to cells for 11 passages. Cells lost H3K27me3 marker by 70%. Treating cells with DZNep for five passages caused cells to lose H3K27me3 by 90% ([Fig. 5](#F0005){ref-type="fig"}).
{#F0005}
### X-Chromosomes are associated with Euchromatic Region in Reprogrammed Cells {#sec3-7}
HSF1 male cell line immunostained with H3K4me3 (a euchromatic region marker) showed homogenous red color with no exclusion mark which is indicative of a heterochromatic region. X-chromosome DNA FISH showed one active X chromosome ([Fig. 6A](#F0006){ref-type="fig"}). In a subpopulation of HSF6-10 cell line, there was a homogenous staining for H3K4me3 with no exclusion mark, and DNA FISH showed the presence of two active X chromosomes ([Fig. 6B](#F0006){ref-type="fig"}). In another subset of HSF6 (10) cell line, there was an exclusion mark in H3K4me3 staining, and DNA FISH showed two X chromosomes ([Fig. 6C](#F0006){ref-type="fig"}).
{#F0006}
### Western Blot Analyses Showed no Global Changes in Trimethylation of H3K27 and H3K4 {#sec3-8}
Control cells had no drug added. NaBu and DZNep were added to treated cells with 0.2μM and 0.1μM concentrations respectively. Western analyses of HSF6-8, HSF6-10, and HSF6-S9 control cells, treated cells with NaBu, and treated cells with NaBu and DZNep showed no global changes in trimethylation of H3K27 and H3K4. Histone H3 was the loading control. The trimethylation of H3K27 demonstrated no obvious changes in control and treated cells ([Fig. 7](#F0007){ref-type="fig"}).
{#F0007}
DISCUSSION {#sec1-4}
==========
In order to mimic *in vivo* germ cell development pathway, hESCs should be in their native state, and keep their native state during passages. Three classes of hESC lines derived from ICM of developing blastocyst are similar in their pluripotency potential and forming teratomas when injected to mice. The different point among them is H3K27me3 enrichment deposited in Xi. Also, these three classes have different expression of XIST RNA. Class I cell lines with active X chromosomes are the ones which resembles the in vivo system. However, hESC lines are usually a kind of class II and class III, and even if they are class I, they would transform to class II or III through passaging.
Treated cells with 0.2mM NaBu and 0.1mM DZNep formed colonies with typical stem cell colony appearance which indicates that these two drugs would not interfere with self-renewal state of the cells. Additionally, treated cells had more cells per colony which suggests that the drugs even promote the self-renewal state. Less differentiation in treated cells compared the controls suggests that the differentiation pathway would stop or slow down by NaBu and DZNep.
The loss of H3K27me3 associated with Xi confirmed by immunostaining analyses suggests that NaBu and DZNep have ability to change H3K27 modification. H3K4me3 staining confirmed the existence of two active X chromosomes. In H3K4me3 staining there was no exclusion mark which suggests that both X chromosomes were active. Immunostaining results indicate that 30% of treated class II cells have been reprogrammed to Class I.
H3K27me3 is a suppressing marker. Naturally, many genes should be kept silent, such as oncogenes. Paternally or maternally expressed genes should become silent in the other parent. Therefore, the integrity maintences of necessary silent genes are crucial. Western blot analyses showed that NaBu and DZNep didn't change the global trimethylation of H3K27 and H3K4, and they selectively change the trimethylation of H3K27 associated with Xi. Therefore, using these two drugs with mentioned concentrations is safe, and won't have any global effect.
For future direction, we want to perform differentiation assays to confirm the proper reprogramming process. Also, gene expression assays seem to be informative for investigating gene expression profiles between control and treated cells. In order to rule out any chromosomal abnormality causing by DZNep and NaBu, cytogenetic analysis is necessary. Finally, investigating the effects of NaBu and DZNep in newly developed cell lines will further confirm the ability of the drugs to epigenetically reprogram class II cell lines to the class I.
**Conflict of interest statement:** All authors declare that they have no conflict of interest.
| {
"pile_set_name": "PubMed Central"
} |
Many endocrinologists thrive on the expansive nature of endocrine disease. However, sometimes it feels like clinical endocrinology is dominated by diabetes. Understandably, since diabetes is arguably the most common endocrine disease, certainly time-consuming and expensive, and apparently at no risk of abating. This issue of JCTE is nominally focused on diabetes. However, the work of diabetes is ever-growing. So, clinically and translationally interested endocrinologists will notice that even when "focused" on diabetes, it is hard to avoid a theme of expansion.
There is a broad swath of diabetic issues covered, from basic science and epidemiology to clinical observation. Demonstrating the necessarily expansive nature of our approach to the epidemic of diabetes, articles in this issue address epidemiology, social science, dentistry, nephrology, podiatry, nutrition, and metabolism. Demonstrating the global and universal reach of diabetes, articles in this issue come from the United States -- with high representation in ethnic and racial minorities, and internationally from Finland, Germany, Singapore, and the Pacific Islands.
Across this broad array of studies focused on diabetes, it is interesting that each approach reported in this issue -- epidemiological, social, nutritional, clinical -- come to strong conclusions that should enable further study or change our clinical approach and knowledge base. Epitomizing the international scope and the need to keep an expansive open mind, a paper from Gujral, collaborating with colleagues at Emory and the CDC in Atlanta as well as investigators from Chennai demonstrate that there is certainly more than one "flavor" to prediabetes and type 2 diabetes [@bib0010]. It is highly accepted that the pathophysiologic basis of type 2 diabetes typically requires both insulin resistance and relative beta-cell insufficiency. Using data from American and South Asian databases, they demonstrate that both groups are at very high risk for prediabetes and diabetes, but there is a striking susceptibility to beta-cell dysfunction in the South Asian population compared to American populations across different ethnicities, which may account for the higher rate of diabetes in South Asians. In other words, at least two different pathways to diabetes must be considered -- one epitomized by early insulin resistance and the other by early beta-cell dysfunction, though both requiring some combination of the two to lead to eventual type 2 diabetes. One must keep an expansive open mind to understand there are more than one pathways to type 2 diabetes.
| {
"pile_set_name": "PubMed Central"
} |
All relevant data are within the paper and in supporting information files. DNA sequences have been deposited at DDBJ/EMBL/GenBank under the accession LNFW00000000.
Introduction {#sec001}
============
*Cochliobolus miyabeanus* ((Ito & Kuribayashi) Drechsler ex Datur.) (anamorph = *Bipolaris oryzae* (Breda de Haan) Shoemaker) is a common fungal pathogen worldwide. In the U.S., it has been documented from the north, in North Dakota and Minnesota, to the south, from Florida to Texas, as well as in areas of California \[[@pone.0154122.ref001]--[@pone.0154122.ref004]\]. It is a major pathogen of rice (*Oryza sativa* L.) in all areas of the world where this crop is grown \[[@pone.0154122.ref005]\]. In addition, it has the potential to cause a severe yield limiting leaf spot disease on two North American non-traditional grass crops, switchgrass (*Panicum virgatum* L.), grown for cellulosic biofuel production \[[@pone.0154122.ref004]\], and American wildrice (*Zizania palustris* L.), hypothesized to have originated in North America \[[@pone.0154122.ref006]\] and grown commercially for its gourmet grain \[[@pone.0154122.ref007]\]. *C*. *miyabeanus* causes fungal brown spot (FBS) on wildrice that can lead to economically disastrous losses in paddy-grown crops \[[@pone.0154122.ref008], [@pone.0154122.ref009]\], resulting in a greater reliance on fungicide to bring about profitable yields. In susceptible wildrice, fungal conidia usually germinate by 8 h after deposition on leaves and develop club-shaped appressoria by 18 h. Infection hyphae break through the cuticle, or less frequently through stomata, develop under the cuticle, and later invade inter- and intracellular spaces. Symptoms appear about 18 to 48 h after inoculation as brown-purple to dark spots that enlarge over time into oval lesions with brown to tan necrotic centers, frequently surrounded by chlorotic halos \[[@pone.0154122.ref010]\]. Lesions tend to coalescence, whitening aerial leaves. Stems and sheaths can also be infected and the weakened stems frequently break, causing considerable kernel loss \[[@pone.0154122.ref003]\].
To mitigate grain yield reduction, a few wildrice cultivars have been released with improved genetic resistance to FBS \[[@pone.0154122.ref011]\]; however, the molecular bases of resistance are not known. Further, fungal mechanisms of virulence on wildrice have not been broadly studied in contrast to those of other species of *Cochliobolus*.
Pathogenicity in many *Cochliobolus* species is largely due to host-specific toxins (HSTs). The first HST described was victorin, a nonribosomal peptide (NRP), produced by *C*. *victoriae*, the causal agent of Victoria blight in oat \[[@pone.0154122.ref012]\]. It has been proposed that in susceptible oat genotypes carrying the homozygous dominant *Vb* locus, the fungal toxin binds to a thioredoxin guarded by a NB-LRR protein that in turn triggers apoptosis, facilitating disease for this necrotrophic fungus \[[@pone.0154122.ref013]\]. *C*. *carbonum* race 1 produces HC-toxin, a tetrapeptide that inhibits histone deacetylases involved in DNA repair, modification, and transcription. The locus *TOX2* contains essential genes for toxin synthesis which includes *HTS1*, a nonribosomal peptide synthetase (*NRPS*) \[[@pone.0154122.ref014]\], *TOXA*, a cyclic peptide efflux pump for toxin detoxification \[[@pone.0154122.ref015]\], *TOXC*, a fatty acid synthetase \[[@pone.0154122.ref016]\], *TOXF*, a branched-chain amino acid transaminase \[[@pone.0154122.ref017]\], *TOXG*, an alanine racemase \[[@pone.0154122.ref018]\], and *TOXE*, an atypical regulatory sequence that controls expression of *TOXA* and *TOXC* \[[@pone.0154122.ref019]\]. Race T of *C*. *heterostrophus* produces a linear polyketide HST (T-toxin), which generates pores in the inner mitochondrial membrane and subsequent leakage in maize carrying the Texas male sterile cytoplasm (*T-urf13*) gene. The complex locus *TOX1* includes genes for synthesis of the toxin found in two unlinked loci, *ToxA* and *ToxB* \[[@pone.0154122.ref020]\]. *ToxA* contains two monomodular polyketide synthase (*PKS*) genes required for toxin production and virulence \[[@pone.0154122.ref021], [@pone.0154122.ref022]\]. *ToxB* comprises a decarboxylase (*DEC1*) and three reductases (*RED1*, *RED2*, and *RED3*) \[[@pone.0154122.ref020], [@pone.0154122.ref023]\] of which only *RED2* participates in the toxic peptide synthesis. Lastly, a putative NRPS associated with virulence on barley was recently uncovered in *C*. *sativus* through comparative genomics among *Cochliobolus* species \[[@pone.0154122.ref024]\].
*C*. *miyabeanus* strains that are pathogenic on common rice do not have unique PKSs or NRPSs and are not known to produce an HST \[[@pone.0154122.ref024]\]. *C*. *miyabeanus*, and other *Cochliobolus* species, make non-specific phytotoxic cyclic sesquiterpenes commonly known as ophiobolins \[[@pone.0154122.ref025]\], which are also produced by non-pathogenic fungi \[[@pone.0154122.ref026]\], suggesting that they have functions other than in interactions with plant hosts. Purified ophiobolins have been found to have a number of effects on plants including inhibiting root growth, stimulating electrolyte leakage from roots, and inducing stomatal opening \[[@pone.0154122.ref026]\]. Ophiobolins also have antimicrobial activity and cause hyphal deformation \[[@pone.0154122.ref027]\].
*Cochliobolus* belongs to the class Dothidiomycetes. This class consists of fungi with a wide assortment of life styles that live in ecologically diverse environments. It is thought that members of this class descended from a common ancestor over 280 million years ago, and contemporary species exhibit genomes with macro-, meso-, and microsynteny, variation in genome sizes attributed to the amount of repeated DNA, and yet conserved gene numbers \[[@pone.0154122.ref028]\]. Plant pathogens within the Dothidiomycete*s* contain genes with 10 unique Pfam domains and 69 expanded domains that are not present in other plant pathogens. The proteomes of plant pathogens within the *Pleosporales* are enriched for cysteine-rich small secreted proteins (SSPs) ≤ 200 amino acids (aa) in length, some of which are thought to be involved in plant-fungus interactions. Some SSPs are common to all *Cochliobolus* species while others are unique to single species \[[@pone.0154122.ref024]\]. The Dothideomycetes have a vast number of genes for production of secondary metabolites, including PKSs, NPSs, and terpene synthases \[[@pone.0154122.ref028]\]. A comparative study of *Cochliobolus* plant pathogens showed that some of the secondary metabolites synthesized by *NRPS* and *PKS* genes were conserved among *Cochliobolus* species, while others were unique to a single species. A few of them (NPS1, NPS3, and NPS13) with discontinuous distribution among species have complicated patterns of evolution that include expansion, loss, and recombination of adenylation (AMP) domains that may generate novel plant toxic peptides \[[@pone.0154122.ref024]\].
Here, we report the draft genome assembly and catalog of genes of a *C*. *miyabeanus* strain originally isolated from infected wildrice as well as the transcriptome of the pathogen during an early time point of wildrice colonization. In our analysis we focus on potential effectors, including small secreted proteins and genes potentially involved in pathogenicity on wildrice, with the objective of gaining a better understanding of the mechanisms of pathogenesis to assist in enhancing genetic resistance in wildrice breeding lines.
Materials and Methods {#sec002}
=====================
Fungal strain and DNA extraction {#sec003}
--------------------------------
The *C*. *miyabeanus* strain TG12bL2 (hereafter referred to as *Cm*TG12bL2) was isolated from a wildrice leaf with FBS symptoms collected from a paddy in Aitkin, Minnesota, USA as previously described \[[@pone.0154122.ref029]\]. For DNA extraction, the fungus was grown in 2% (w/v) water agar (Bacto Agar, DIFCO) for approximately two weeks until spores were produced. Spores collected from four Petri dishes were added to two sterilized 500 ml glass flasks containing 200 ml of liquid minimal medium \[[@pone.0154122.ref030]\]. The flasks were shaken at 150 rpm for 6 days at room temperature in ambient light, breaking the mycelium apart every 48 h to promote new growth, and placing in fresh medium. Mycelium was harvested by filtration, freeze dried, and DNA was extracted using the protocol of Raeder and Broda \[[@pone.0154122.ref031]\]. DNA concentration and quality were measured using the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA).
Fungal genome sequencing and *de novo* assembly {#sec004}
-----------------------------------------------
The *Cm*TG12bL2 genome was shotgun sequenced using paired-end reads with 101 cycles with Illumina HiSeq 2000 technology at the Mayo Clinic, Rochester, Minnesota, USA. A paired-end library was prepared according to the Illumina Preparation Guide (<http://www.illumina.com>). Briefly, 2 to 5 μg of DNA in 100 μl of 10 mM Tris and 0.1 mM EDTA, pH 8, was fragmented using a Covaris E210 sonicator (Woburn, MA) generating double-stranded DNA fragments with blunt or sticky ends with a fragment size mode between 400 to 500 bp. The ends were repaired using Klenow DNA polymerase and T4 DNA polymerase, phosphorylated using T4 polynucleotide kinase, after which an adenine was added to the 3' ends of double-stranded DNA using Exo-Minus Klenow DNA polymerase. Paired-end DNA adaptors with a single thymine base overhang at the 3' end were ligated to DNA fragments and the resulting constructs were separated on a 2% agarose gel. DNA fragments of approximately 500 bp were excised from the gel using GeneCatcher tips and purified using Qiagen Gel Extraction Kits (Qiagen, Valencia, CA). The adapter-modified DNA fragments were enriched by 12 cycles of PCR using primers PE 1.0 and PE 2.0. The concentration and size distribution of the library was determined with a DNA 1000 chip on the Agilent 2100 Bioanalyzer (Agilent Technologies). The library was loaded onto an indexed lane of paired-end flow cell. The reads in the flow cell were sequenced as paired-end indexed reads on an Illumina HiSeq 2000 using TruSeq SBS sequencing kit version 1 and HiSeq data collection v. 1.1.37.0 software. Base-calling was performed using Illumina's RTA v. 1.7.45.0. Reads quality control was performed in Galaxy (<https://galaxyproject.org/>).
The fungal genome was *de novo* assembled using Velvet v.1.2.10 \[[@pone.0154122.ref032]\]. Hash length (*k*-mer) ranged from 31 to 91 by two nucleotide increments. Contigs were merged into scaffolds using paired-end information and a coverage cutoff and expected coverage set to 'auto' and an average insert length of 385 bp. The optimal hash length for the assembly was selected based on the maximum N~50~ length, large *k*-mer specificity, and high coverage, without losing overall genome or scaffold length. Parameter optimization for the selected assembly was further refined using the expected coverage (76.56×) and cutoff coverage (38.28×), following software recommendations \[[@pone.0154122.ref033]\], to discriminate unique genomic regions from repeats, and to further eliminate errors due to low coverage, respectively.
Protein-coding sequence prediction, annotation, and validation {#sec005}
--------------------------------------------------------------
Protein-coding sequences were predicted using GeneMark_ES v.2.3c \[[@pone.0154122.ref034]\], which employs an intrinsic method with unsupervised training based on an *ab initio* algorithm. The algorithm features an enhanced intron sub-model to accommodate sequences with and without branch point sites found in several fungi (Ascomycota, Basidiomycota, and Zygomycota). Proteins were annotated against NCBI non-redundant (NCBI_nr) (<http://www/ncbi.nlm.nih.gov>) and HMMER (<http://hmmer.janelia.org>) databases. The latter was also used to identify protein profiles (Pfam). For overall analyses with blastp an *e*-value cut off of 1E-5 was used, except for detecting candidate effectors where a higher *e*-value cut off was permitted (1E-2). Proteins with a signal peptide or anchor peptide were predicted by SignalP 4.1 \[[@pone.0154122.ref035]\]. Small secreted proteins (SSPs) were considered those of less than 200 aa without transmembrane domains. Cysteine-rich SSPs in the *Cm*TG12bL2 genome were identified as previously described \[[@pone.0154122.ref028]\]. HCSSPs were resolved as in \[[@pone.0154122.ref028]\] with at least twice the average of cysteines in the *Cm*TG12bL2 proteome and as in \[[@pone.0154122.ref024]\] as those SSPs containing more than 2% cysteines.
The *Cm*TG12bL2 predicted genes were validated using three strategies. First, the protein set was compared to a core of 458 highly conserved proteins \[[@pone.0154122.ref036]\]. This core consisted of highly homologous proteins that are present in six eukaryote genomes (*Homo sapiens*, *Drosophila melanogaster*, *Caenorhabditis elegans*, *Arabidopsis thaliana*, *Saccharomyces cerevisiae*, and *Schizosaccharomyces pombe*), and was selected from the KOG (clusters of euKaryotic Orthologous Groups) database. Second, the genomes of *Cm*TG12bL2 and *C*. *miyabeanus* WK-1C (ATCC 44560 v1.0) (<http://genome.jgi.doe.gov>) were compared using Gepard \[[@pone.0154122.ref037]\]. Third, by comparing the genes to the fungal transcriptome at 48 h post-inoculation that was assembled and identified as described below. We adopted the following nomenclature throughout the manuscript: "Cm" refers to gene; "CM" to protein and "t_Cm" to transcript. The number that follows each prefix is the same for gene, protein, and transcript.
Identifying carbohydrate-active enzymes (CAZymes) {#sec006}
-------------------------------------------------
For the detection of encoded CAZymes, each *Cm*TG12bL2 protein model was compared (blastp) to proteins listed in the Carbohydrate-Active Enzymes database ([www.cazy.org](http://www.cazy.org)) \[[@pone.0154122.ref038]\]. Model *Cm*TG12bL2 proteins with a length of over 50% identity to those in CAZy database were directly assigned to the same family (or subfamily when relevant). Proteins with less than 50% identity to a protein in CAZy were all manually inspected for conserved features such as the catalytic residues. Sequence alignments with isolated functional domains were performed in the case of multimodular CAZymes \[[@pone.0154122.ref039]\]. The same methods were used for all fungi that were compared to *Cm*TG12bL2 proteins. Putative roles of the enzymes were inferred from annotations of proteins returned from searches against the profile-HMMER database (hmmscan) and protein sequences database (phmmer) using an *e*-value cut off of 1E-33.
Nonribosomal peptide synthetase and polyketide synthase identification and phylogenetic analysis {#sec007}
------------------------------------------------------------------------------------------------
NRPSs and PKSs were identified by comparing the *Cm*TG12bL2 proteins to the well-curated dataset of *C*. *heterostrophus* (Ch_NPSs) \[[@pone.0154122.ref040], [@pone.0154122.ref041]\] using an *in-house* script and to other NRPSs and PKSs from NCBI_nr (blastp, 1E-5). Additional motifs in proteins with AMP domains and ketoacyl synthase (KS) domains were identified using HMMER hmmscan (with default settings unless otherwise specified). A degenerate AMP domain in the predicted *Cm*TG12bL2 protein CM_1661, homologous with Ch_NPS6 was identified through manual inspection \[[@pone.0154122.ref042]\]. AMP domains of CM_3163 were compared to the corresponding *C*. *heterostrophus* homologous domains NPS1, NPS3, and NPS13. Finally, putative NRPSs and PKSs from *Cm*TG12bL2 were blasted to proteins of *C*. *miyabeanus* WK-1C. A phylogenetic analysis with the *Cm*TG12bL2 AMP domains was done using methods described by Bushley and Turgeon \[[@pone.0154122.ref040]\] to compare conservation and relationships of AMP domains with those identified in the reference species *C*. *heterostrophus* C5, *C*. *heterostrophus* C4, and the rice isolate *C*. *miyabeanus* WK-1C \[[@pone.0154122.ref024]\]. Several AMP domains from NRPSs in other Ascomycetes species that encode known chemical products were included as well. Acyl-adenylating enzymes (acyl-CoA-synthetases, *CPS1*, long chain fatty acids, acyl-CoA-ligases, and ochratoxins) from *C*. *heterostrophus* and other ascomycetes were used as outgroups. Protein sequences were aligned using MAFFT and a maximum likelihood phylogenetic tree was constructed in RAxML using the best protein model determined by ProTest (RT-REV-F) as in \[[@pone.0154122.ref024], [@pone.0154122.ref040]\].
Predicted transporter and detoxification systems {#sec008}
------------------------------------------------
P450 monooxygenases were obtained by comparing the *Cm*TG12bL2 protein set (blastp, *e*-value cut off 1E-5) to the Fungal Cytochrome P450 Database (FCPD; <http://p450.riceblast.snu.ac.kr/download.php>). The MFS and ABC transporters were predicted by comparison to the transporter classification database ([www.tcdb.org](http://www.tcdb.org)) with an *e*-value cut off of 1E-05.
Wildrice inoculation with *C*. *miyabeanus* {#sec009}
===========================================
Leaves of wildrice plants of the variety Itasca C-12 with improved resistance to FBS \[[@pone.0154122.ref011]\] were inoculated with 20,000 conidia/ml of the isolate *Cm*TG12bL2, in four biological replicates. Plants were kept in a mist chamber for 16 h then moved to a greenhouse and kept at 24°C with 75% relative humidity. Samples of the flag leaf and flag leaf-1 were collected at 24 h and 48 h after inoculation (hai), immediately frozen in liquid nitrogen, and kept at -80°C until used for RNA extraction.
*C*. *miyabeanus* transcriptome assembly and analysis {#sec010}
-----------------------------------------------------
Total RNA was extracted with the RNeasy Mini Kit (Qiagen Inc, Valencia, CA) according to manufacturer's instructions and treated with Ambion DNAse I (Life Technologies, Carlsbad, CA) to eliminate traces of genomic DNA. Only samples collected at 48 hai were sequenced for the transcriptome analysis. The Illumina TruSeq RNA Sample Preparation Guide was followed to prepare the samples for sequencing according to manufacturer's recommendations (<http://www.illumina.com>). Illumina library preparation was performed as previously described \[[@pone.0154122.ref043]\]. The samples were barcoded, multiplexed, and sequenced using a single-end read with 50 cycles using an Illumina HiSeq2000 machine at the Biomedical Genomic Center at the University of Minnesota. Read quality control was carried out in Galaxy using the Tuxedo suite. Only reads with an average Q score ≥30 were aligned onto the *Cm*TG12bL2 draft genome for transcriptome assembly using TopHat v1.4.1 \[[@pone.0154122.ref044]\] implemented in Galaxy.
The CuffLinks suite of tools \[[@pone.0154122.ref045]\] was used for transcript assembly and quantification. Transcript abundance was estimated using RPKM (Reads Per Kilobase of transcript per Million mapped reads) \[[@pone.0154122.ref046]\]. RPKM values were log~2~ transformed and data distribution was visualized using JMPin (SAS Institute Inc., Cary, NC, USA). Experimental variability of the log~2~ of RPKM between biological replicates was estimated using Pearson's correlation coefficient for the 9,960 transcripts that were present in all replicates. For subsequent data analysis only transcripts found in at least three of four replicates were used. Transcripts were functionally categorized using Blast2GO ([www.blast2go.com](http://www.blast2go.com)) against the NCBI non-redundant (NCBI_nr) and SwissProt/InterPro databases. An enrichment analysis was performed with the 10% most abundant transcripts using a Fisher exact test with a false discovery rate (FDR) of 0.05 implemented within Blast2GO \[[@pone.0154122.ref047]\] to identify overrepresented or enriched GO terms using the whole set of transcripts as reference.
Quantitative RT-PCR (qRT-PCR) experiments {#sec011}
-----------------------------------------
qRT-PCR experiments were done to validate expression of the transcriptome analysis at 48 hai and to compare the expression levels to an additional time point at 24 hai *in planta* and to the fungus grown *in vitro*. cDNA synthesis for all the treatments was carried out with the iScript cDNA synthesis kit (Bio-Rad Laboratories Inc., Hercules, CA). *Cm*TG12bL2 specific primers for the selected genes *Ecp6* (Cm_2799), CYP53 (Cm_8068), salicylate hydroxylase (Cm_9653), β-1,4-endoglucanase (Cm_1858), β-1,4-endoxylanase (Cm_8238), and β-1,4-glucosidase (Cm_28)) were designed using the SciTools software at IDT (<http://www.idtdna.com/scitools/Applications/RealTimePCR>). The genes expression levels *in planta* were compared to those of the fungus grown *in vitro* (reference sample). qRT-PCR reactions were done in 96-wells plates using an Applied Biosystems 7500/7500 Fast Real-Time PCR system (Applied Biosystems, Foster City, CA). Each reaction of 20 μL contained 10 μL iTAQ^™^ Universal SYBR Green Supermix (Bio-Rad Laboratories Inc.), 0.1 μL of each primer diluted to 0.1μM, 3 μL of cDNA template and 6.8 μL of nuclease-free water. PCR condition were 95°C for 2 min, 40 cycles of 95°C for 3 s, 60°C for 30 s, and 95°C for 15 s, followed by 60°C for 1 min, 95°C for 15 s, 60°C for 15 s (melt curve generation). The delta-delta Ct method was computed by the Applied Biosystems software and used for relative quantification of gene expression. Expression data were normalized against an endogenous control (glyceraldehyde-3-phosphate dehydrogenase gene). Two to three biological replications with at least two technical replicates were analyzed per gene. Primer efficiencies for the genes analyzed were 100% ±10 (SD). Data are presented as the Log~2~ of relative gene expression (fold change). Expression values in the reference sample were set to 0 (Log~2~ of 1). Significance of gene differential expression between treatments was tested using T-tests with the significance threshold set at 0.05.
Results {#sec012}
=======
*C*. *miyabeanus* fungal genome assembly, gene prediction, and validation {#sec013}
-------------------------------------------------------------------------
Read statistics and quality are summarized in [S1 Table](#pone.0154122.s008){ref-type="supplementary-material"}. High quality reads were assembled with various *k*-mer lengths ([S1 Fig](#pone.0154122.s001){ref-type="supplementary-material"}). The *65*-mer assembly yielded the largest N~50~ (75,371 bp), and contained 3,475 scaffolds, with a maximum scaffold size of 348,044 bp and a total genome length of 31,696,836 bp ([S1 Fig](#pone.0154122.s001){ref-type="supplementary-material"}). After parameter optimization ([S2 Fig](#pone.0154122.s002){ref-type="supplementary-material"}) and removal of contigs shorter than 100 bp, the final draft was 31,788,735 bp long ([Table 1](#pone.0154122.t001){ref-type="table"}) and contained 2,378 scaffolds with a N~50~ value of 74,921 bp. This Whole Genome Shotgun project has been deposited at DDBJ/EMBL/GenBank under the accession LNFW00000000. The version described in this paper is version LNFW01000000.
10.1371/journal.pone.0154122.t001
###### Summary features of the *Cochliobolus miyabeanus Cm*TG12bL2 isolate genome assembly and annotation.
{#pone.0154122.t001g}
Parameter
-------------------------------------------------------------------------------------- ----------------
Genome size (draft) (bp)[^a^](#t001fn001){ref-type="table-fn"} 31,788,735
Number of scaffolds 2,378
Mean/Maximum scaffold length (bp) 13,368/348,108
Scaffold N~50~ length (bp) 74,921
Number of scaffolds with length over N~50~ 131
Mean coverage (X) 229.30
GC genome content (%) 50.50
Total predicted protein-coding genes[^b^](#t001fn002){ref-type="table-fn"} 12,142
Partial genes[^c^](#t001fn003){ref-type="table-fn"} 1,142
Predicted complete set of protein-coding genes[^d^](#t001fn004){ref-type="table-fn"} 11,000
Mean coding gene length (bp)[^e^](#t001fn005){ref-type="table-fn"} 1,588
Mean exon length (bp)[^e^](#t001fn005){ref-type="table-fn"} 529
Mean intron length[^e^](#t001fn005){ref-type="table-fn"} 93
Mean number of exon/gene[^e^](#t001fn005){ref-type="table-fn"} 2.71
Mean gene density[^f^](#t001fn006){ref-type="table-fn"} (genes/Mbp) 346
Mean amino acid number/protein 476
Genes with a signal peptide[^g^](#t001fn007){ref-type="table-fn"} 1,066
^a^ Genome size: total base pairs after optimizing the parameters for cutoff coverage (38.3X), expected coverage (76.56X), and removing contigs smaller than 100 bp.
^b^ Total number of protein-coding genes as predicted by GeneMark-ES v.2.3c.
^c^ Predicted gene sequences lacking part of the coding region (i.e. start or stop codons) or with stop codon(s) in the sequences.
^d^ Total number of predicted complete genes used in further analyses.
^e^ Mean values based on 11,000 genes.
^f^ Calculated as total protein coding gene sequences/ (protein-coding gene sequences + total intergenic areas).
^g^ Predicted using SignalP4.1 server and based on 11,000 genes.
The total number of protein-coding sequences initially predicted by GeneMark-ES was 12,142. After discarding partial genes (1,142), due to the presence of internal stop codons or lack of either initiation or stop codons, 11,000 coding sequences remained for further analyses. The average C+G content was 52.63% for predicted genes, 52.74% for exons, and 45.69% for introns, *versus* 50.5% for the entire genome. Additional genome assembly features are shown in [Table 1](#pone.0154122.t001){ref-type="table"}. Interrogation (blastp, 1E-05) of the 11,000 predicted proteins with the NCBI_nr database resulted in 10,395 matches (94.5%), of which 92.3% were annotated as hypothetical proteins, mostly belonging to other *Cochliobolus* species (92.2%). A total of 7,567 motifs were identified when comparing (blastp, 1E-05) all predicted proteins to collections of protein profiles (<http://hmmer.janelia.org/>).
*Cm*TG12bL2 gene models were also validated (blastp, 1E-05) against a core dataset of 458 highly conserved eukaryote proteins \[[@pone.0154122.ref036]\]. All but two of the 458 core proteins had at least one predicted *Cm*TG12bL2 homolog ([S2 Table](#pone.0154122.s009){ref-type="supplementary-material"}). Of the two core proteins that could not be directly identified: KOG1291 (Mn^2+^ and Fe^2+^ transporters of the NRAMP family) and KOG2749 (mRNA cleavage and polyadenylation factor IA/II complex), the latter matched a partial protein. Additionally, two *Cm*TG12bL2 proteins, CM_8436 and CM_2839, each matched two core proteins, but only one of each *Cm*TG12bL2 protein (highest percentage of identity) is presented. Thus, 99.1% of the core dataset could be identified in our predicted protein set. The *Cm*TG12bL2 genome assembly was compared to the *C*. *miyabeanus* WK-1C genome (<http://genome.jgi.doe.gov>). The ordered scaffolds showed good pair-wise alignment ([S3 Fig](#pone.0154122.s003){ref-type="supplementary-material"}) validating the assembly over relatively long distances. An additional validation was done by mapping fungal transcripts against the draft genome. Overall, 97.6% of the predicted genes that were expressed at 48 hai in at least three biological replicates were mapped to the genome sequence.
Genes involved in plant-pathogen interactions {#sec014}
---------------------------------------------
Comparative genome analyses identified a number of genes in *Cm*TG12bL2 previously shown to be required for pathogenicity in other *Cochliobolus* species ([Table 2](#pone.0154122.t002){ref-type="table"}). For instance, the *C*. *heterostrophus CGBI* gene encodes a signaling G-protein ß-subunit involved in conidia production and female fertility as well as appressorium and mycelium pigmentation, mycelium morphogenesis, and virulence \[[@pone.0154122.ref048]\]. The *Cm*TG12bL2 putative homolog has eight exons and seven introns and encodes a polypeptide with two conserved domains, a WD40 domain and a phosducin-like domain that is a cytosolic regulator of G-proteins from the thioredoxin-like superfamily. The *C*. *heterostrophus CPS1* gene encodes an acyl CoA ligase-like protein, which is required for normal virulence in maize. Because it is present in several pathogenic and saprophytic Ascomycetes *CPS1* may also have a role in stress tolerance \[[@pone.0154122.ref049]\]. The *Cm*TG12bL2 homolog contains two large AMP-binding domains, and a DMAP1-binding domain at its amino terminal end. The *HDC1* to *HDC4* genes of *C*. *carbonum* encode histone deacetylases \[[@pone.0154122.ref050], [@pone.0154122.ref051]\]. *HDC1* is required for virulence (penetration efficiency) in maize, normal conidia size, and growth in complex polysaccharides \[[@pone.0154122.ref050]\]. The *Cm*TG12bL2 homolog is identical to the 505 amino acid (aa) sequence of *HDC1* but is longer at the carboxyl terminus. An 880 aa protein of *Cm*TG12bL2 was 99% identical to the protein kinase ccSNF1 from *C*. *carbonum*, including the predicted nuclear localization signal \[[@pone.0154122.ref052]\]. ccSNF1 regulates expression of extracellular fungal enzymes for degradation and uptake of carbohydrates and is required for full virulence in maize. A homolog of the DNA-binding transcriptional repressor CreA \[[@pone.0154122.ref053]\] for fungal degrading enzymes of plant cell walls, and putatively regulated by ccSNF1 in *C*. *carbonum*, is also present in the *Cm*TG12bL2 genome. The *BKM1* gene of a rice isolate of *C*. *miyabeanus* has a homolog in *Cm*TG12bL2. Disruption of *BKM1* causes defective mycelia and colony growth, loss of conidiation and pathogenicity on rice leaves \[[@pone.0154122.ref054]\]. A homolog of the Ch_NPS6, required for virulence and insensitivity to hydrogen peroxide in rice \[[@pone.0154122.ref055]\] is also present in the *Cm*TG12bL2 genome.
10.1371/journal.pone.0154122.t002
###### *Cochliobolus miyabeanus* TG12bL2 genes with similarity to genes involved plant-pathogen interactions in other *Cochliobolus* species.
{#pone.0154122.t002g}
*C*. *miyabeanus* gene ID Gene name in other *Cochliobolus* species *e*-value (Percent amino acid identity; percent of sequence coverage) Function of homolog in *Cochliobolus* sp. Reference
--------------------------- ------------------------------------------- ----------------------------------------------------------------------- ------------------------------------------- ----------------------------
Cm_7805 SNF1 (AF159253) 0.0 (99; 100) Catabolic-repression and uptake \[[@pone.0154122.ref052]\]
Cm_6457 CGB1 (AAO25585) 0.0 (100; 66) Signaling G-protein β-subunit \[[@pone.0154122.ref048]\]
Cm_6993 BMK1 (BAD42855) 0.0 (100; 95) Development pathways \[[@pone.0154122.ref054]\]
Cm_5509 CreA (Q9HFS2) 0.0 (98; 100) DNA binding transcriptional repressor \[[@pone.0154122.ref053]\]
Cm_10788 CPS1 (AAF53991) 0.0 (96; 89) Acyl-CoA ligase/stress response \[[@pone.0154122.ref049]\]
Cm_6585 HDC1 (AF306507) 0.0 (100; 68) Histone deacetylase \[[@pone.0154122.ref050]\]
Cm_3336 HDC2 (AF349677) 0.0 (99; 100) Histone deacetylase \[[@pone.0154122.ref050]\]
Cm_3871 HDC3 (AF307341) 0.0 (98; 100) Histone deacetylase \[[@pone.0154122.ref051]\]
Cm_10259 HDC4 (AF537126) 0.0 (95; 99) Histone deacetylase \[[@pone.0154122.ref051]\]
Cm_1661 NPS6 (ABI51982) 0.0 (99; 91) Iron extracellular siderophore \[[@pone.0154122.ref055]\]
Cm_1096 CHAP1 (AAS64313) 0.0 (91; 100) Redox-regulated transcription factor \[[@pone.0154122.ref056]\]
Cm_3707 PGN1 (P26215) 0.0 (95; 100) Endopolygalacturonase \[[@pone.0154122.ref057]\]
Cm_30 PGX1 (Q00359) 0.0 (96; 100) Exopolygalacturonase \[[@pone.0154122.ref058]\]
Cm_3810 XYL2 (AAT49296) 0.0 (97; 100) Endo-beta-1,4 xylanase \[[@pone.0154122.ref059]\]
Cm_10775 XYL3 (AAC62816) 9E^-160^ (96; 100) Endo-beta-1,4 xylanase \[[@pone.0154122.ref059]\]
Cm_1858 CEL2 (AF336799) 0.0 (95, 100) Cellulase \[[@pone.0154122.ref060]\]
Cm_8632 ALP1 (AAB03851) 3E^-167^ (87; 100) Trypsin-like serine protease \[[@pone.0154122.ref061]\]
The *Cm*TG12bL2 genome has genes that encode homologs of proteins involved in causing disease but are not required for pathogenicity ([Table 2](#pone.0154122.t002){ref-type="table"}). An example is a predicted homolog of *CHAP1* that in *C*. *heterostrophus* encodes a redox-regulated transcription factor necessary for gene activation for resistance to oxidative stress, although it is not involved in virulence in maize \[[@pone.0154122.ref056]\]. The *Cm*TG12bL2 genome has homolog proteins to five *C*. *carbonum* cell wall degrading enzymes, endopolygalacturonase (PGN1), exo-α-1,4-polygalacturonase (PGX1), two ß-1,4-xylanases (XYL2 and XYL3), and glucohydrolase (cellulase; CEL2), which were proven to be dispensable because strains with mutations in those genes were still pathogenic in maize \[[@pone.0154122.ref057], [@pone.0154122.ref058], [@pone.0154122.ref059], [@pone.0154122.ref060]\]. Additionally a *Cm*TG12bL2 protein was similar to *C*. *carbonum ALP1*, a trypsin-like serine protease that is not required for pathogenicity in maize \[[@pone.0154122.ref061]\]. Several of the *Cm*TG12bL2 genes are expected to have similar functions in the *C*. *miyabeanus*-wildrice interaction, particularly genes such as the iron extracellular siderophore *NPS6*, and *BKM1* that are involved in virulence/pathogenicity in the rice isolate of *C*. *miyabeanus* \[[@pone.0154122.ref054], [@pone.0154122.ref055]\].
Small secreted proteins (SSPs) {#sec015}
------------------------------
The fungal secretome is essential for interactions with the surrounding environment and potential hosts. Within the SSP group are cysteine-rich secreted proteins (candidate effectors) that could interfere with plant defense mechanisms. The *Cm*TG12bL2 proteome includes a total of 187 proteins less than 200 aa long, with signal peptides and without transmembrane domains. The number of SSP proteins is comparable to the number found in other Dothidiomycetes, although less numerous than those found in other *Cochliobolus* species \[[@pone.0154122.ref028]\] ([S3 Table](#pone.0154122.s010){ref-type="supplementary-material"}). Most of the *Cm*TG12bL2 SSPs (95.7%) have similarities to hypothetical proteins from other *Cochliobolus* species in the NCBI_nr database including *C*. *miyabeanus* WK-1C. SSP features ([S3 Table](#pone.0154122.s010){ref-type="supplementary-material"}) are within or close to the range limits found in *Cochliobolus* species or other Dothidiomycetes \[[@pone.0154122.ref024], [@pone.0154122.ref028]\]. The percentage of SSPs with Pfam domains in *Cm*TG12bL2 (13.9%) is higher, albeit close to that found in other Dothidiomycetes. Examples of proteins with motifs identified in *Cm*TG12bL2 SSPs are hydrophobins, hydrophobic peptides with predicted functions as surfactants \[[@pone.0154122.ref062]\][,]{.smallcaps} CAZymes, essential to the hydrolysis of carbohydrates of plant cells walls, and proteins with carbohydrate-binding modules (described elsewhere in the manuscript). Other proteins included a fungal-specific CFEM protein with eight conserved cysteines, some of which have been implicated in fungal pathogenicity \[[@pone.0154122.ref063]\], a protein with an Hce2 domain that constitutes the mature part of the *Ecp*2 effector protein from the tomato pathogen *Cladosporium fulvum* (detailed in the next section), and a protein similar to Asp f 13-like from *C*. *lunatus* that contains a cerato-platanin domain. Proteins with this domain are phytotoxic and cause cell necrosis and induce plant defenses \[[@pone.0154122.ref064]\] but also have structural functions in the fungal cell wall. Additional SSPs identified included a clathrin adaptor complex small chain protein involved in the clathrin-mediated pathway for endocytosis of small molecules and a peptidase inhibitor I9 ([S4 Table](#pone.0154122.s011){ref-type="supplementary-material"}).
Effector-like secreted proteins {#sec016}
-------------------------------
Some of the effector-like proteins present in *Cm*TG12bL2 are similar to proteins known to be involved in pathogenicity, or to interfere with plant defense mechanisms. One example is CM_6804 with a pathogen effector Hce2 motif, a putative necrosis-inducing factor. This protein is similar to the extracellular protein 2 (Ecp2) of the biotrophic pathogen *Cladosporium fulvum* ([S4 Fig](#pone.0154122.s004){ref-type="supplementary-material"}), that is secreted into the apoplast during colonization of tomato leaves and is considered to have a role in virulence \[[@pone.0154122.ref065]\]. *Cm_6804* is 488 bp long, with two exons and one intron, and encodes a 142 aa peptide with a predicted signal peptide of 19 aa. Similar proteins have been found in *Mycosphaerella graminicola* and *M*. *fijiensis* \[[@pone.0154122.ref066]\], *Gibberella zeae*, *C*. *sativus* ND90Pr, and *C*. *heterostrophus* ([S4 Fig](#pone.0154122.s004){ref-type="supplementary-material"}) and they could play an important role in pathogenesis.
The CM_2799 protein is similar to *C*. *fulvum* Ecp6, which is secreted into the apoplast of tomato leaves during colonization and is involved in virulence \[[@pone.0154122.ref067]\]. Ecp6 is found in many other plant pathogenic and saprophyte fungi \[[@pone.0154122.ref067], [@pone.0154122.ref068]\]. The three Epc6 LysM domains act as carbohydrate binding modules with affinity to chitin tri-, penta-, and hexa-oligosaccharides and compete with chitin-binding plant receptors, thus preventing initiation of defense responses by the plant immune system \[[@pone.0154122.ref069]\]. CM_2799 has three typical LysM domains (pfam01476) with one cysteine at the beginning and end of each domain ([S4 Fig](#pone.0154122.s004){ref-type="supplementary-material"}), which could form intramolecular disulfide bonds. In addition, the protein has a cysteine at the C-terminus. Similar peptides are found in other fungi \[[@pone.0154122.ref069]\] including the hemibiotrophs *C*. *sativus* NDPr90, *Collectotrichum truncatum*, *Setosphaeria turcica*, and *Mycosphaerella graminicola* and the necrotroph *Cochliobolus heterostrophus* \[[@pone.0154122.ref070], [@pone.0154122.ref071]\].
Two proteins from *C*. *miyabeanus*, CM_2749 and CM_6024, have similarity to the Nep1-like (necrosis and ethylene inducing peptide-like or NLPs) family of proteins. In some fungal pathogens NLPs are expressed *in planta* between the biotrophic and necrotrophic phase in dicot hosts \[[@pone.0154122.ref072], [@pone.0154122.ref073]\], or at the end of the symptomless phase in some monocot hosts \[[@pone.0154122.ref074]\]. NPLs elicit a hypersensitive-like response in dicots \[[@pone.0154122.ref075]\]. In monocots, however, NLPs do not appear to induce similar responses as the disruption of *NLP* does not affect pathogenicity or virulence; instead, a role in microbe inhibition has been suggested. CM_6024 homologs are present in *C*. *sativus*, *C*. *heterostrophus* and also in *Collectrichum higginsianum*, where a necrosis-inducing protein (ChNLP1) is expressed only during the switch to the necrotrophic phase in *Nicotiana benthamiana* \[[@pone.0154122.ref071]\].
Other larger secreted proteins with potential roles in pathogenicity included predicted cutinases, peptidases, fungal transporters belonging to the major facilitator superfamily (MFS), and ATP-binding cassette (ABC) group proteins, in addition of CAZymes. Some of these are described in further detail below.
Carbohydrate-active enzymes (CAZymes) {#sec017}
-------------------------------------
Fungal plant pathogens need access to nutrients located mostly in host cell protoplasts. Thus, they secrete a wide array of proteins to breach plant barriers (i.e. cuticle and cell walls) and degrade complex carbohydrates for carbon acquisition. The *Cm*TG12bL2 genome encodes 604 carbohydrate binding- and catalytic protein-modules distributed in 530 protein-coding genes. They are involved in binding, modification and breakdown of plant carbohydrates, fungal cell wall biosynthesis, remodeling and turn over, as well as *N*- and *O*-glycoprotein synthesis and processing. The number of CAZymes identified was in the upper range of those found in other Ascomycetes ([Table 3](#pone.0154122.t003){ref-type="table"}) and substantially higher than in *Saccharomyces cerevisiae*, the obligate biotroph *Blumeria graminicola*, the hemibiotroph *Mycosphaerella graminicola*, and the necrotrophs *Pyrenophora tritici-repentis*, *P*. *teres* f. *teres*, *Leptosphaeria maculans*, and *Alternaria brassicicola*, but close to the number found in the hemibiotrophs *Gibberella zeae* and *Magnaporthe oryzae*, and the necrotrophs *Phaeosphaeria nodorum* SN15 and *Cochliobolus heterostrophus*. The number and variety of *Cm*TG12bL2 CAZymes are within the range of those found in other *Cochliobolus* species without considering CAZyme auxiliary activity modules \[[@pone.0154122.ref028]\].
10.1371/journal.pone.0154122.t003
###### Comparison of predicted carbohydrate protein modules present in 12 Ascomycete genomes.
{#pone.0154122.t003g}
Species[^a^](#t003fn001){ref-type="table-fn"} Glycoside hydrolases Glycosyl-transferases Polysaccharide lyases Carbohydrate esterases Auxiliary activities Carbohydrate-binding modules Total[^b^](#t003fn002){ref-type="table-fn"}
----------------------------------------------- ---------------------- ----------------------- ----------------------- ------------------------ ---------------------- ------------------------------ ---------------------------------------------
*S*. *cerevisiae* 52 67 0 2 7 15 143
*B*. *graminicola* 66 56 0 9 7 14 152
*M*. *graminicola* 195 105 3 16 47 27 393
*P*. *tritici-repentis* 233 100 10 39 92 52 526
*L*. *maculans* 222 99 19 36 83 64 523
*A*. *brassicicola* 243 93 24 43 83 63 549
*P*. *teres* f. *teres* 236 99 10 39 69 61 514
*P*. *nodorum* 253 90 10 49 100 61 563
*G*. *zeae* 245 107 20 45 72 89 578
*C*. *miyabeanus* 253 101 14 50 99 87 604
*M*. *oryzae* 259 102 5 51 93 119 629
*C*. *heterostrophus* 274 104 15 49 89 102 633
^a^*S*. *cerevisiae*: S*accharomyces cerevisiae* S288c, *B*. *graminicola*: *Blumeria graminicola* f. sp. *hordei*, *M*. *graminicola*: *Mycosphaerella graminicola*, *P*. *tritici-repentis*: *Pyrenophora tritici-repentis*, *L*. *maculans*: *Leptosphaeria maculans*, *A*. *brassicicola*: *Alternaria brassicicola*, *P*. *teres* f. *teres*: *Pyrenophora teres* f. *teres*, *P*. *nodorum*: *Phaeosphaeria nodorum* SN15, *G*. *zeae*: *Gibberella zeae*, *C*. *miyabeanus*: *Cochliobolus miyabeanus* TG12Lb2, *M*. *oryzae*: *Magnaporthe oryzae* 70--15, *C*. *heterostrophus*: *Cochliobolus heterostrophus* C5.
^b^Total is the sum of all CAZYmes (binding and catalytic protein-modules).
We identified 69, 29, three, and nine families of glycoside hydrolases (GH), glycosyltransferases (GT), polysaccharide lyases (PL), and carbohydrate esterases (CE), respectively, in addition to 12 families of auxiliary activity (AA) enzymes \[[@pone.0154122.ref076]\] in the *Cm*TG12bL2 genome. There are 13 families of non-catalytic carbohydrate binding modules (CBMs) ([S5 Table](#pone.0154122.s012){ref-type="supplementary-material"}) in addition to five proteins with distant similarity to plant expansins. More than 30 *Cm*TG12bL2 genes had combined catalytic and non-catalytic CAZy modules.
A wealth of auxiliary activities (AA) enzymes (99) was identified with diverse catalytic functions that facilitate plant lignocellulose degradation. The most prominent families were AA1, AA3, and AA9. All three subfamilies of AA1 were found containing laccases, ferroxidases, and laccase-like multicopper oxidases (LMCOs), respectively. Additional details of *Cm*TG12bL2 CAZymes are provided in [S1 Text](#pone.0154122.s015){ref-type="supplementary-material"} \[[@pone.0154122.ref077], [@pone.0154122.ref078], [@pone.0154122.ref079], [@pone.0154122.ref080], [@pone.0154122.ref081], [@pone.0154122.ref082], [@pone.0154122.ref083], [@pone.0154122.ref084]\].
To infect and obtain nutrients from wildrice tissues, *C*. *miyabeanus* needs to breach the cuticle and cell walls composed of cellulose, cross-linking glycans (hemicellulose), and pectin. Numerous CAZymes capable of degrading the backbone chains of those layers as well as removal of substitution residues for complete and efficient degradation, were predicted through electronic annotation (blastp, 1E-22) ([S6 Table](#pone.0154122.s013){ref-type="supplementary-material"}). A schematic representation of the possible function of some of the *Cm*TG12bL2 CAZymes that potentially allows the fungus to breach the structural barriers of wildrice tissues and acquire nutrients to sustain plant colonization is presented in [Fig 1](#pone.0154122.g001){ref-type="fig"} \[[@pone.0154122.ref085]\].
![Schematic representation of hypothetical degradation of wildrice cell wall polysaccharides by CAZymes.\
A. Degradation of cellulose by cellulolytic enzymes: endo-β-1,4-glucanase, cellobiohydrolase, β-glucosidase, and lytic polysaccharide monooxygenase; B. Degradation of hemicellulose polysaccharides and substituted residues by hemicellulolytic enzymes. *Xylan*: endoxylanase, β-xylosidase, acetylxylan esterase, feruloyl esterase, α-L-arabinofuranosidase, α-glucuronidase; *Xyloglucan*: endo-β-1,4-endoglucanase, α-L-fucosidase, β-galactosidase, α-xylosidase; *Galactomannan*: endo-β-1,4-mannanase, β-mannoside, α-galactosidase. C. Degradation of pectin polysaccharides and substituted moieties by pectinolytic enzymes. *Homogalacturonan*: endo- and exopolygalacturonase, pectin and pectate lyase, pectin acetylesterase, pectin methylesterase; *Xylogalacturonan*: endo- and exogalacturonase, β-xylosidase; *Rhamnogalacturonan*: rhamnogalacturonase, rhamnogalacturonan lyase, rhamnogalacturonan acetylesterase, α-L-rhamnosidase, β-1,4-galactosidase, β-1,4-galactanase, β-glucuronidase, α-L-arabinofuranosidase, α-L-arabinanase, β-1,6-galactosidase, β-1,3-galactanase α-L-fucosidase. The glycan symbol nomenclature used was according to \[[@pone.0154122.ref085]\].](pone.0154122.g001){#pone.0154122.g001}
Nonribosomal peptide synthetases and polyketide synthases {#sec018}
---------------------------------------------------------
Fungi produce a plethora of natural products that include NRPs, polyketides, terpenoids, and alkaloids, among others. They play important roles in growth and development, reproduction, response to oxidative stress, and pathogenicity toward plants and microorganisms \[[@pone.0154122.ref086]\]. NRPs are synthesized through a ribosome-independent pathway by mono- or multimodular NRPSs where each module is composed of a set of core domains. The minimal module consists of an adenylation (AMP) domain, a peptidyl (P) carrier protein or thiolation (T) domain and a condensation (C) domain whose biochemical functions are described elsewhere \[[@pone.0154122.ref086]\]. In some enzymes the C domain is replaced by a thioesterase NADP(H) dependant reductase domain \[[@pone.0154122.ref040]\]. Additional domains for epimerization and methylation may be present in the enzyme, adding diversity to the final polypeptide chain \[[@pone.0154122.ref087]\]. Previous phylogenetic analysis indicated that mono- and bimodular subfamilies contain more conserved domain structures and may have originated earlier than enzymes with more modular composition \[[@pone.0154122.ref040]\].
In the *Cm*TG12bL2 genome we identified one α-aminoadipate reductase (AAR) and twelve putative NRPSs ([Table 4](#pone.0154122.t004){ref-type="table"}), including proteins with homology to Ch_NPS2, Ch_NPS4, Ch_NPS6, and Ch_NPS10. The last four are the most conserved among Dothidiomycetes and are involved in virulence, morphology, cell surface hydrophobicity, sensitivity to stresses, and fitness \[[@pone.0154122.ref088]\]. CM_4358 has a monomodular structure similar to Ch_NPS10, with A, T, and NAD binding domains and a short-chain dehydrogenase, which confers substrate specificity and provides a catalytic site. Mutations in the latter gene affect colony morphology and tolerance to oxidative stress \[[@pone.0154122.ref086]\]. *Ch_NPS10* has the highest rate of conservation among 18 Dothidiomycetes \[[@pone.0154122.ref028]\]. CM_8122 is a multimodular enzyme with a similar structural organization to Ch_NPS2. The latter catalyzes the synthesis of an intracellular siderophore (ferricrocin) involved in iron storage and essential for development of asci and ascospores in *C*. *heterostrophus* \[[@pone.0154122.ref089]\] but not required for virulence in maize. CM_10529 is highly similar to Ch_NPS4, which is also well conserved among filamentous ascomycetes and is involved in cell surface hydrophobicity in *C*. *heterostrophus*, *Alternaria brassicicola*, and *Giberella zeae*, and also plays a role in conidial cell wall development and rate of germination in *A*. *brassicicola* \[[@pone.0154122.ref090]\]. CM_1661 has high similarity to Ch_NPS6 ([Table 4](#pone.0154122.t004){ref-type="table"}), an extracellular siderophore that in *C*. *heterostrophus*, as well as its homologs in a rice-infecting strain of *C*. *miyabeanus*, and other filamentous ascomycetes are involved in virulence. In *C*. *heterostrophus*, NPS6 also confers tolerance to oxidative stress \[[@pone.0154122.ref042]\], and it is thought to supply iron to the fungus during host colonization \[[@pone.0154122.ref055]\]. CM_194 and CM_5231 are similar to Ch_NPS12, but the latter *Cm*TG12bL2 protein has a lower percent identity and lacks the ferric reductase-like membrane component (Pfam01794) and instead has a spore coat protein U domain (pfam05229) usually found in bacterial NRPSs. CM_10197 has homology to the hybrid NRPS/PKS protein Ch_NPS7/PKS24 ([Table 5](#pone.0154122.t005){ref-type="table"}). CM_10647 is like an AAR, proteins that are phylogenetically close to multimodular NRPSs and involved in lysine biosynthesis in fungi \[[@pone.0154122.ref040]\]. CM_8006 has similarity to a less conserved NRPS in *Cochliobolus* species, the multimodular Ch_NPS3. Three additional NRPS-like proteins (CM_9450, CM_4038, and CM_6570) were found that are similar to hypothetical proteins of *C*. *heterostrophus* C5 and *C*. *miyabeanus* WK-1C.
10.1371/journal.pone.0154122.t004
###### Identification and structure of nonribosomal peptide synthetases in *Cochliobolus miyabeanus* TG12bL2.
{#pone.0154122.t004g}
CM ID number[^a^](#t004fn001){ref-type="table-fn"} (amino acids) NRPS homolog (organism; accession number) Percent protein coverage[^b^](#t004fn002){ref-type="table-fn"} *e*-value[^b^](#t004fn002){ref-type="table-fn"} Percent identity[^b^](#t004fn002){ref-type="table-fn"} Predicted modular organization of *C*. *miyabeanus* proteins[^c^](#t004fn003){ref-type="table-fn"}
------------------------------------------------------------------ ------------------------------------------- ---------------------------------------------------------------- ------------------------------------------------- -------------------------------------------------------- ----------------------------------------------------------------------------------------------------
CM_8122 (5394) NPS2 (Ch; AAX09984) 99 0 86 A\[T\]C-ATC-ATC-ATC TCTC
CM_10529 (7224) NPS4 (Ch; AAX09986) 100 0 93 TC-C-ATC-ATC-C-ATC-ATC-CT-CT
CM_1661 (1939) NPS6 (Ch; AAX09988) 90 0 94 ATC-(dA)TTC
CM_4358 (1282) NPS10 (Ch; AAX09992) 100 0 97 ATN D
CM_8006 (5142) NPS3 (Ch; AAX09985) 100 0 92 ATC-AMTC-ATC-A\[M\]TC
CM_10197 (2542) NPS7 (Ch; AAX09989) 99 0 88 [^d^](#t004fn005){ref-type="table-fn"}
CM_194 (1006) NPS12 (Ch; AAX09994) 100 0 94 A \[R\]
CM_5231 (1017) NPS12 (Ch; AAX09994) 100 0 55 A \[SC\]
CM_3163 (3908) NPS1/NPS3/NPS13 expanded NPS-like \- \- \- C-ATC-A\[M\]TC-ATC-d(A)
CM_10647 (1179) AAR-like (Cm; XP007682966) 100 0 100 ATN
CM_9450 (1051) NPS-like (Ch; EMD86802) 98 0 89 A\[T\]N
CM_4038 (1218) NPS-like (Cm; XP007686974) 100 0 98 ATN-AAA
CM_6570 (1063) NPS-like (Cm; XP007692684) 100 0 100 ATN
^a^ *Cochliobolus miyabeanus* TG12bL2 protein identification number.
^b^ Percent of protein coverage in pairwise alignment, *e*-value and percent of identity determined using BLAST.
^c^ Predicted domain structural organization based on similarity to the HMMER database (<http://hmmer.janelia.org/>) with a cut-off value = 0.01 and hit = 0.03.
Domains within brackets were detected with cut off \> 0.01 and ≤ 1, parentheses indicate domains detected after manual inspection based on conserved amino acids. A = AMP-binding (pfam00501); T = PP (pfam00550), phosphopantetheine attachment site or thiolation domain (or peptidyl carrier protein); C = condensation domain (pfam00668), N = NAD_binding_4 (pfam07993), D = short chain dehydrogenase (pfam00106), M = methyltransferase domain (pfam13847), R = Ferric reductase-like membrane component (pfam01794), SC = spore coat protein U domain (pfam05229), dA = putative degenerated A2 domain, AAA = ATPases associated with diverse cellular activities (pfam13671). AAR = α-aminoadipate reductase. Ch = *C*. *heterostrophus*; Cm = *C*. *miyabeanus* ATCC 44560.
^d^ Structural domain organization of NPS7/PKS24 hybrid protein CM_10197 described in [Table 5](#pone.0154122.t005){ref-type="table"}.
10.1371/journal.pone.0154122.t005
###### Identification and structure of polyketide synthases in *Cochliobolus miyabeanus Cm*TG12bL2.
{#pone.0154122.t005g}
CM ID number[^a^](#t005fn001){ref-type="table-fn"} (amino acids) PKS homolog (organism; accession number) Percent protein coverage[^b^](#t005fn002){ref-type="table-fn"} *e*-value[^b^](#t005fn002){ref-type="table-fn"} Percent identity[^b^](#t005fn002){ref-type="table-fn"} Predicted modular organization of *C*. *miyabeanus* proteins[^c^](#t005fn003){ref-type="table-fn"}
------------------------------------------------------------------ ------------------------------------------ ---------------------------------------------------------------- ------------------------------------------------- -------------------------------------------------------- ----------------------------------------------------------------------------------------------------
CM_4294 (2487) PKS3 (Ch; AAR90258) 100 0 96 KS-AT-DH-ME-ER-KR-PP
CM_1415 (2639) PKS5 (Ch; AAR90260) 100 0 94 KS-AT-DH-ME-ER-KR-PP
CM_1147 (2550) PKS6 (Ch; AAR90261) 95 0 93 KS-AT-DH-ME-ER-KR-PP
CM_2466 (2335) PKS7 (Ch; AAR90262) 99 0 32 KS-AT-DH-ER-KR-PP
CM_416 (2621) PKS8 (Ch; AAR90263) 97 0 96 KS-AT-DH-ME-ER-KR-PP
CM_9255 (2274) PKS9 (Ch; AAR90264) 100 0 91 KS-AT-DH-ER-KR-PP
CM_6592 (2468) PKS12 (Ch; AAR90267) 100 0 84 KS-AT-DH-ME-ER-KR-PP
CM_466 (2155) PKS14 (Ch; AAR90268 98 0 91 KS-AT-DH-ME-ER-KR-PP
CM_10206 (2441) PKS15 (Ch; AAR90269) 93 0 91 KS-AT-DH-ME-ER-KR-PP
CM_5798 (2155)[^d^](#t005fn004){ref-type="table-fn"} PKS18 (Ch; AAR90272) 99 0 96 KS-AT-DH-PP-PP-Th
PKS1 (Cm, BAD22832) 100 0 99
CM_8343 (1784) PKS19 (Ch; AAR90273) 100 0 94 KS-AT-DH-PP
CM_4293 (2583) PKS21 (Ch; AAR90275) 84 0 97 KS-AT-PP-ME-N
CM_6273 (2715) PKS22 (Ch; AAR90276) 86 0 88 KS-AT-DH-PP-PP-PP-ME-abH-Pep
CM_10241 (2232) PKS23 (Ch; AAR90277) 92 0 85 KS-AT-DH-PP-PP-ME
CM_10197 (2542) PKS24 (Ch; AAR90278) 99 0 89 AMP-PP-KS-AT-KR-PP-N
CM_11065 (2249) PKS14 (Cs, EMD68395) 100 0 89 KS-AT-DH-PP-PP-N
CM_1155 (2465) PKS6 (Cs; EMD66380) 100 0 93 KS-AT-DH-ME-ER-KR-PP
CM_1154 (1987) PKS16 (Cs, EMD66379) 88 0 100 KS-AT-DH-PP-TE-UbiA
CM_803 (1802) PKS13 (Cs; EMD66882) 100 0 92 KS-AT-PP
^a^ *Cochliobolus miyabeanus* protein identification number.
^b^ Percent of protein coverage in the pairwise alignment.
^c^ Predicted domain structural organization based on similarity to the CDD (<http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml>) with a cut-off *e*-value = 0.01, and/or HMMER database (<http://hmmer.janelia.org/>) with a gathering cut-off value Model = 0.01 and hit = 0.03.
^d^ PKS18 (AAR90272) and PKS1 (BAD22832) from *C*. *heterostrophus* and *C*. *miyabeanus*, respectively are both homologs proteins of CM_5798.
KS = beta-ketoacyl-synthase domain (pfam00109); AT = acyl transferase domain (pfam00698); PP = phosphopantetheine attachment site (pfam00550 or smart00823); DH = polyketide synthase dehydratase domain (pfam14765); ME = methyltransferase domain (pfam08242); ER = enoyl reductase domain (cd05195) or ADH_N, alcohol dehydrogenase (pfam08240) and ADH_Zinc_N, zinc-binding dehydrogenase (pfam00107), TE = thioesterase (pfam00975), N = NAD_binding_4 (pfam07993), abH = alpha/beta hydrolase (pfam07859), Pep = Peptidase (pfam00326); A = AMP-binding (pfam00501). PKS = polyketide synthase, Hp = hypothetical protein. Ch = *Cochliobolus heterostrophus*, Cs = *Cochliobolus sativus*, Cm *= Cochliobolus miyabeanus*.
A phylogenetic tree was constructed to analyze the conservation and relationships among AMP domains present in the *Cm*TG12bL2 NRPS proteins with those of other *Cochliobolus* species ([S5 Fig](#pone.0154122.s005){ref-type="supplementary-material"}). Within the NRPS conserved group, the single AMP domain of each *Cm*TG12bL2 NPS6, NPS7/PKS24, NPS10, the two NPS12, and the related adenylating enzyme AAR, clustered with their *Cochliobolus* counterpart with well supported bootstrap values ([S5 Fig](#pone.0154122.s005){ref-type="supplementary-material"}). Each of the four AMP domains of *Cm*TG12bL2 for NPS2 and NPS4 proteins grouped with the corresponding domain in the other three *Cochliobolus* isolates. This was also the case for the AMP domains of the *Cm*TG12bL2 homolog of less conserved multimodular NPS3. The monomodular NPS11 was not present in *Cm*TG12bL2, nor was it found in *C*. *miyabeanus* WK-1C \[[@pone.0154122.ref024]\]. However, the single NPS11 AMP domain of both *C*. *heterostrophus* strains grouped with the first AMP domain (AMP1_2) of the two bimodular NRPSs: GliP, responsible for gliotoxin production in *A*. *fumigatus* (EAL88817), and SirP, responsible for sirodesmin biosynthesis in *Leptosphaeria maculans* (AAS92545). Interestingly, both strains of *C*. *miyabeanus* each contain a NRPS-like protein, CM_2059 in *Cm*TG12bL2 and ID 107726 in *C*. *miyabeanus* WK-1C, with a single AMP domain that groups instead with the second module (AMP2_2) of both GliP and SirP proteins ([S5 Fig](#pone.0154122.s005){ref-type="supplementary-material"}).
In *Cochliobolus* species, NPS1, NPS3, and NPS13 proteins are discontinuously distributed. Their AMP domains display duplications/deletions, fusions, and/or recombinations that are thought to give rise to novel NRPSs within these expanded clades \[[@pone.0154122.ref024]\]. CM_3163 is a tetra-modular protein that has a homolog in *C*. *miyabeanus* WK-1C (W6YWH5) and belongs to this expanded group. We analyzed the placement of the AMP domains with regard to those in *C*. *miyabeanus* WK-1C, and the two *C*. *heterostrophus* strains (C4 and C5) ([S5 Fig](#pone.0154122.s005){ref-type="supplementary-material"}). Each AMP domain of the NPS1/NPS3/NPS13 NRPS-like proteins from both *C*. *miyabeanus* isolates always clustered together. Additionally, the AMP1_4 of CM_3163 grouped with maximum bootstrap support (100%) to AMP1_1 of the pseudogene *Ch_NPS13* while AMP2_4 and AMP3_4 of CM_3163 clustered with AMP2_4 of Ch_NPS3. The fourth AMP domain of CM_3163 was weakly related to AMP3_4 of Ch_NPS3 ([S5 Fig](#pone.0154122.s005){ref-type="supplementary-material"}). Homologous AMP domains from Ch_NPS1 are absent in both *C*. *miyabeanus* isolates. The AMP1_1 domains of NPS-like proteins (CM_9450, CM_4038, CM_6570) were located with the outgroup set, and grouped within a strongly supported subset of AMP domains of *Fusarium graminearum* and *Aspergillus* species involved in ochratoxin biosynthesis ([S5 Fig](#pone.0154122.s005){ref-type="supplementary-material"}).
Most fungal polyketides are synthesized by type I PKSs and are monomodular enzymes with multiple catalytic domains carrying out repeated biosynthetic reactions. These rapidly evolving proteins catalyze the condensation of Co-enzyme A, either from CoA thioesterified carboxylic acids in reducing PKSs, or from acetyl- or malonyl-CoA in nonreducing PKSs \[[@pone.0154122.ref041]\] to form carbon chains or cyclic forms of varying lengths. A minimal module of PKS consists of a β-ketosynthase (KS), acyltransferase (AT), and acyl carrier protein (ACT) or PPT attachment site (PP) whose roles in the enzyme have been previously described \[[@pone.0154122.ref041]\]. Additional modules with specific functions, β-ketoreductase (KR), dehydrogenase (DH), enoyl reductase (ER), methyltransferase (ME), and thioesterase (TE) add modifications or cyclize the polyketide compounds. Based on the modules present, the PKSs can be subdivided into reducing or non-reducing enzymes, the latter lacking some or all of the reducing domains (KR, DH, and ER) \[[@pone.0154122.ref041]\].
Nineteen PKSs were identified in the *Cm*TG12bL2 genome. Overall, 15 proteins matched the *C*. *heterostrophus* C4 PKS set (Ch_PKS). Those included ten (Ch_PKS3, Ch_PKS5, Ch_PKS9, Ch_PKS12, Ch_PKS14, Ch_PKS15, Ch_PKS18, Ch_PKS19, Ch_PKS23, and Ch_PKS24) that are common to all *Cochliobolus* species and additional five (Ch_PKS6, Ch_PKS7, Ch_PKS8, Ch_PKS21, and Ch_PKS22) with a discontinuous distribution among the species of the genus. Four others are homologs of PKS6, PKS13, PKS14, and PKS16 of *C*. *sativus* ([Table 5](#pone.0154122.t005){ref-type="table"}). Predictions of protein domains, excluding PKS24, indicate that eight PKSs have the seven ancestral modules of the fully reduced type I PKSs, nine are partially reducing (lacking one or more DH, KR, or ER domains), and one is a non-reducing type. In the *Cm*TG12bL2 genome, two of the 21 PKSs reported for *C*. *miyabeanus* WK-1C were not detected, including the duplicated and expanded PKS14. Overall, the number of *PKSs* and *NRPSs* found in the *Cm*TG12bL2 genome was similar to the number in other members of the Pezizomycotina, and the organization of structural domains are identical or similar to those found in other *Cochliobolus* species.
Predicted transporter and detoxification systems {#sec019}
------------------------------------------------
Cytochrome P450 monooxygenases (CYPs) are heme-thiolate enzymes involved in degradation and detoxification of xenobiotics \[[@pone.0154122.ref091]\]. CYPs catalyze chemical modifications in lipophilic compounds from primary and secondary metabolism to create more hydrophilic derivatives. We identified 113 CYPs in the *Cm*TG12bL2 genome ([S7 Table](#pone.0154122.s014){ref-type="supplementary-material"}), similar to the number found in other members of the Pezizomycotina \[[@pone.0154122.ref092]\]. Examples of enzymes predicted to participate in primary metabolism include eburicol 14-α demethylase (CM_837) that is a member of the CYP51 family, a sterol C-22 saturase from the fungal-specific CYP61 family (CM_7374) involved in membrane ergosterol biosynthesis, members of the CYP52 family (CM_10978, CM_255, and CM_1621) that catalyze initial degradation of n-alkanes and fatty acids, and a CYP56 family member (CM_6988) that participates in formation of the N, N'-bisformyl dityrosine spore cell wall component \[[@pone.0154122.ref091]\]. The *Cm*TG12bL2 genome encodes CYPs from families known to be involved in biosynthesis of aflatoxins (CYP59 and CYP62), fumonisins (CYP65 and CYP505), tricothecenes (CYP65, CYP68, and CYP526), and gibberellin (CYP68) ([S7 Table](#pone.0154122.s014){ref-type="supplementary-material"}). Additionally, other CYPs could be involved in detoxification and degradation of xenobiotics. For example, CM_10633 is similar to one of the pisatin demethylase genes, PDA6-1 from *Nectria hematococca* (CYP57) involved in degrading the pterocarpan phytoalexin produced by *Pisum sativum* \[[@pone.0154122.ref093]\]. CM_8068 is a likely ortholog of CYP53A15 from *C*. *lunatu*s, a benzoate para-hydroxylase with O-demethylase activity that catalyzes degradation of benzoic acid and derivatives as well as natural toxins \[[@pone.0154122.ref094]\]. CM_6446, a CYP504A member, and CM_5407 from family CYP504B have similarity to phacA, a phenylacetate 2-hydrolase from *Aspergillus nidulans* \[[@pone.0154122.ref095]\] and to phacB, a 3-hydroxyphenylacetate 6-hydrolase \[[@pone.0154122.ref096]\], respectively. Both participate in phenylacetate degradation to Krebs cycle intermediate compounds.
Fungal transporters provide resistance to a variety of drugs, exogenous mycotoxins and fungicides, and contribute to pathogenicity by delivering mycotoxins outside fungal cells \[[@pone.0154122.ref097]\]. In the *Cm*TG12bL2 genome, as in other Ascomycetes, the major facilitator superfamily (MFS) and the ATP-binding cassette (ABC) superfamily, with 279 and 46 genes respectively, account for most of the transporters identified. The *Cm*TG12bL2 MFS transporters are distributed in 22 families ([Fig 2](#pone.0154122.g002){ref-type="fig"}). Five families have the majority of the predicted MFS proteins: the sugar porter (SP), the Drug:H^+^ antiporter-1 (12 Spanner) (DHA1), the drug:H^+^ antiporter-2 (14 Spanner) (DHA2), the monocarboxylate porter (MCP) and the anion:cation symporter (ACS). Examples of MFS found include a monosaccharide transporter similar to MstC, the low affinity glucose:H^+^ symporter from *A*. *niger* \[[@pone.0154122.ref098]\], and a high-affinity xylose--proton symporter GXS1 \[[@pone.0154122.ref099]\], a protein with similarity to Flr1 of *Saccharomyces cerevisiae*, a MFS involved in resistance to the antifungal drug fluconazole and several chemically unrelated drugs \[[@pone.0154122.ref100]\], a nicotin acid permease, and H^+^:biotin symporter involved in uptake of vitamins \[[@pone.0154122.ref101]\].
{#pone.0154122.g002}
Transporters within the ABC superfamily typically have two homologous components each containing a nucleotide-binding domain (NBD) and six transmembrane-spanning helices that form the transmembrane domain (TMD~6~); however, half-size transporters are also found \[[@pone.0154122.ref097], [@pone.0154122.ref102]\]. Most of the *Cm*TG12bL2 ABC transporters are in the ABC1 and ABC2 superfamilies ([Fig 3](#pone.0154122.g003){ref-type="fig"}), with the majority within the families that participate in efflux of toxins \[[@pone.0154122.ref103]\]. Predicted *Cm*TG12Lb2 proteins are similar to those in the multidrug resistance exporter family, ABCB (i.e., Mdr1 from *A*. *fumigatus* that confers resistance to antifungal compounds \[[@pone.0154122.ref097]\]), and the pleiotropic drug resistance family ABCG (i.e., AtrB transporter of *A*. *nidulans* associated with resistance to cycloheximide, a variety of fungicides, and toxins including antifungal plant compounds \[[@pone.0154122.ref102]\]). Additional *Cm*TG12Lb2 proteins within the ABCG family have similarity to CDR1 and CDR2 of *Candida albicans*, that play a role in resistance to antifungal azoles; and to STS1, the suppressor of sporidesmin toxicity, of *Saccharomyces cerevisiae* that confers resistance to sporidesmin and others drugs such as cycloheximide \[[@pone.0154122.ref104]\]. Putative homologs of the metal resistance protein YCF1 from *S*. *cerevisiae* in the ABCC family are also present.
{#pone.0154122.g003}
*C*. *miyabeanus* transcriptome *in planta* {#sec020}
-------------------------------------------
As expected, the vast majority of the reads generated were of plant origin. The totality of reads mapped to the *Cm*TG12bL2 genome yielded, depending on the particular replicate, 1.6% to 3.0% fungal reads. Pairwise correlation coefficient for log~2~ transformed RPKM between biological replicates ranged from r = 0.87 to r = 0.93, indicating low variation between samples. The average number of *Cm*TG12bL2 transcripts at 48 hai across replications was 10,787 ± 70, of which 10,674 that were present in at least three out of the four replicates, were selected for further analyses. Average log~2~ transformations of the transcripts showed a double exponential distribution (data not shown). For the enrichment analysis (Fisher test), we only investigated transcripts (1,039) located in the three upper quartiles of the distribution.
Functional categorization of fungal transcripts {#sec021}
-----------------------------------------------
Over 5,700 gene ontology (GO) terms were assigned to the 10,674 transcripts. Annotations were found for 17 categories within the *biological process* domain and the most populated were metabolic processes, cellular processes, single-organism processes and localization. The highest number of transcripts for the *molecular function* domain was in binding and catalytic activities; and within the *cellular component* domain was in cell, organelle, and membrane functions ([S6 Fig](#pone.0154122.s006){ref-type="supplementary-material"}).
Gene expression and enrichment analysis {#sec022}
---------------------------------------
We further investigated the highest expressed genes in the upper 10% quartile. Those with predicted functions are involved in protein synthesis (40S and 60S ribosomal proteins) and protein modification (heat shock proteins), translation (eukaryotic translation initiation factor 2), signaling (14-3-3 proteins, calmodulin, MAPK), and transport of secretory proteins (SEC14 cytosolic factor). Transcripts present in high numbers or in the enriched or over-represented pool are associated with reactive oxygen species scavenging, such as superoxide dismutases (t_Cm_4737, t_Cm_8015), glutathione peroxidase (t\_ Cm_5661), peroxiredoxins (t_Cm_5333, t_Cm_1899), and catalases (t\_ Cm_1931, t_Cm_4481, t_Cm_5915). Transcripts involved in signaling through regulation of oxidative stress systems such as mitogen-activated protein kinase HOG1 (t_Cm_6356) and putative homologs of ChAp1 transcription factors (t_Cm_3085) and Skn7 (t_Cm_9152) were also found. Other known stress-related genes highly expressed in *Cm*TG12bL2 included cytochrome c (t_Cm_10384), ubiquitin-conjugating enzyme E2 N (t_Cm_2552), peptidyl-prolyl cis-trans isomerase (t_Cm_4487, t_Cm_6343), ribosomal protein S5 (t_Cm_2988), and protein phosphatase PP2A (t_Cm_2839). A salicylate hydroxylase (t_Cm_9653) was also over-represented.
Seventy-four CAZymes genes were highly expressed, of which 45 were over-represented transcripts. The vast majority were GH (51%) with similarity to cellulase (GH7: t_Cm_1858), mixed-linkage glucanase (GH12: t_Cm_6551), exoglucanases/cellobiohydrolases (GH6: t_Cm_2649; GH7: t_Cm_2863), and putative β-glucosidases (GH3: t_Cm_28, t_Cm_9000). Within the auxiliary group, five AA9 transcripts, with two of them appended to a single CBM1 module were enriched as well. CAZymes that depolymerize other plant cell wall glycans were also over-represented. These included a β-1,4-endoxylanase precursor (GH11: t_Cm_8238), α- and β-galactosidases (GH36: t_Cm_3708; GH35: t_Cm_7042), acetylxylan esterase (CE1: t_Cm_7354), and feruloyl esterase (CE1: t_Cm_6253). Other transcripts encoded enzymes that could degrade pectin such as rhamnogalacturonan acetyl esterase (CE12: t_Cm_4971), pectate lyase (PL3: t_Cm_3311), and pectin esterase (CE8: t_Cm_2172). Additionally, transcripts containing CBM domains like CBM1, CBM35, CBM48 and CBM50 in single or repeated units, present alone and in association with catalytic domains were in the enriched pool. Two putative cutinases (CE5: t_Cm_5545 and t_Cm_6538) were in the enriched set, which could contribute to initiating invasion of wildrice tissues.
We also found 32 SSP-encoding genes with high expression including five that were over-represented and belonged to the GO categories of "extracellular region", "structural molecular activity", "heterocyclic compound binding", and "catalytic activity". Within the group of effector homologs, only t_Cm_2799 a putative Ecp6 homolog, was over-represented.
No *NRPS* or *PKS* genes were over-represented in the transcriptome at 48 hai, and only PKS7 was within the most expressed genes.
Among CYPs, seven over-represented transcripts were associated with metabolic processes, cell and heterocyclic compound binding. These included monooxygenases from the CYP504 family (t_Cm_6446), putatively involved in phenylacetate degradation, and CYP53 (t_Cm_8068), a benzoate para-hydroxylase, and transcripts of the CYP532 family associated with xenobiotic metabolism, as well as members of families CYP552, CYP68, and CYP52.
Within the overrepresented transporters were 11 predicted MFS members including a glucose/xylose symporter1 (GXS1) (t_Cm_11802), a sugar and polyol (SPT1) and monosaccharide transporters (MstC) (t_Cm_1809 and t_Cm_6329), a STL1 permease (t_Cm_4510), a boron exporter (Atr1) (t_Cm_1340), and nicotinic acid permease (t_Cm_571, t_Cm_8891, t_Cm_9423, t_Cm_7821), among others. Four predicted ABC transporters in the *Cm*TG12bL2 genome were in the enriched pool and had similarity to the multidrug resistance protein CDR1 from *C*. *albicans* (t_Cm_6382) and bacterial-like transporters (t_Cm\_ 3444, t_Cm_5374, t_Cm_3853).
qRT-PCR experiments {#sec023}
-------------------
Transcript accumulation at 48 hai of the six genes selected from the enriched pool, Ecp6, CYP53, salicylate hydroxylase, β-1,4-endoglucanase, β-1,4-glucosidase, and β-1,4- endoxylanase were validated by qRT-PCR experiments. The relative quantification of the expression of those genes *in planta* at 24 hai and 48 hai was always significantly higher than that *in vitro*. In addition, levels of expression of CYP53, β-1,4- endoxylanase and β-1,4-glucosidase were similar at the two time points considered, while *Ecp6* and salicylate hydrolase were higher at 48 hai compared to 24 hai ([S7 Fig](#pone.0154122.s007){ref-type="supplementary-material"}), but only the latter was significant. β-1,4-endoglucanase, involved in the degradation of cellulose, was higher at 24 hai. Overall, qRT-PCR expression values were similar to their computed digital counterparts ([S7 Fig](#pone.0154122.s007){ref-type="supplementary-material"}), thus validating the transcript expressions in the transcriptome analysis at 48 hai.
Discussion {#sec024}
==========
Here, we report the *de novo* genome assembly of a strain of *C*. *miyabeanus*, *Cm*TG12bL2, isolated from American wildrice. We also assembled and studied the fungal transcriptome during the initial phases of plant infection and colonization. *C*. *miyabeanus* is found infecting *Oryza* species all over the world \[[@pone.0154122.ref005]\], and in North America it has the potential to cause yield limiting disease on American wildrice \[[@pone.0154122.ref003]\] and switchgrass \[[@pone.0154122.ref004]\]. Despite its importance, relatively little is known about the mechanisms of pathogenicity, or host-pathogen interactions \[[@pone.0154122.ref105]\]. Unlike other *Cochliobolus* pathogens, *C*. *miyabeanus* does not appear to utilize HSTs during plant colonization \[[@pone.0154122.ref024], [@pone.0154122.ref106]\]. In our transcriptome analysis we focused on validating the gene repertoire of the genome analysis and identifying transcripts potentially involved in pathogenicity and others produced in response to host defenses.
The *Cm*TG12bL2 genome is similar to sizes reported for other *Cochliobolus* species. For instance, the draft genome of *C*. *heterostrophus* C4, assembled with Illumina technology was 32.93 Mb, while *C*. *heterostrophus* isolate C5 using Sanger sequencing was 36.46 Mb, and the *C*. *sativus* ND90Pr v.1 genome assembled with a combination of Sanger, 454, and Illumina sequencing yielded 34.42 Mb \[[@pone.0154122.ref024],[@pone.0154122.ref028]\]. The length of our assembly is similar to the *C*. *miyabeanus* WK-1C reference genome (31.4 Mb) downloaded from the JGI website (January, 2014; <http://genome.jgi.doe.gov/Cocmi1/Cocmi1.download.html>) and reported by Condon et al. \[[@pone.0154122.ref106]\]. Importantly, 92% of the *Cm*TG12bL2 predicted proteins are putative homologous to those in WK-1C, including NPRS, PKS, and most of the SSPs. The majority of the proteins in the set that did not match those in the rice isolate were hypothetical proteins, while the rest (1%) carried known domains and could be associated to DNA replication (DNA polymerases, exonucleases) and repair (Rad51 proteins, endonucleases), retrotransposon activities (retrotransposon gag proteins, integrases, and reverse transcriptase), transcription factors (Zinc-finger, CCHC- and C2H2-types) and proteins of unknown functions (DUF domains). The *Cm*TG12bL2 genome shares a set of proteins with other *Cochliobolus* species that are involved in plant-pathogen interaction, including pathogenicity and virulence. Even though most of them were expressed, and some of them at high levels such as the trypsin-like serine protease Alp1 \[[@pone.0154122.ref061]\], they were not found statistically over-represented in the transcriptome at 48 hai.
Overall, a relatively large number of the *Cm*TG12bL2 genes (\~10%) were annotated as potentially being secreted. Many of those genes have roles in host recognition and/or pathogenicity, and were highly expressed or statistically overrepresented in the transcriptome analysis. Some examples include hydrophobins involved in masking spore recognition by hosts and in sensing the host surface \[[@pone.0154122.ref062]\], and predicted cutinases that catalyze the disruption of the fatty acid cutin, the main constituent of plant cuticle. The *Cm*TG12bL2 genome harbored a substantial number of gene-encoding CAZymes with more than 8% of their transcripts in the enriched transcriptome pool that could facilitate ingress and colonization of wildrice tissues. Those included enzymes catalyzing reactions for degradation of cellulose, such as endoglucanases, cellobiohydrolases, and β-glucosidases, together with lytic-polysaccharide monooxygenases of family AA9 that are responsible for oxidative cleavage of cellulose. A few AA9 CAZymes are bound to non-catalytic CBM1, credited with providing a more efficient binding to cellulose \[[@pone.0154122.ref083]\]. Others were enzymes able to depolymerize other plant cell wall glycans, particularly xylan, such as xylanases and β-galactosidases, as well as acetylxylan and feruloyl esterases. The upregulation of some of these genes *in planta* was validated by qRT-PCR analyses. Additionally, a transcript that could be involved in sucrose degradation was highly expressed. Plant sucrose derivatives are used as carbon or energy sources by fungi. Transcripts of predicted MFS transporters putatively involved in sugar uptake and transfer were overrepresented, including a putative homolog of the glucose/xylose symporter1 (GXS1) \[[@pone.0154122.ref099]\], a low affinity glucose monosaccharide transporter (MstC), \[[@pone.0154122.ref098]\], and a sugar and polyol transporter (SPT1) that could facilitate growth and adaptation to a variety of nutrient conditions.
A reduced percent of the fungal secretome (17.5%) consisted of small cysteine-rich proteins of less than 200 aa. SSPs are important for pathogenic actions since they can play roles as effectors promoting disease and/or altering host defense mechanisms \[[@pone.0154122.ref107]\]. Over 96% of all SSPs were proteins that matched those of the rice *C*. *miyabeanus* WK-1C isolate. At 48 hai, 17% of SSPs were highly expressed but only a few were over-represented in the transcriptome enrichment analysis. Within those few having known domains are SSP predicted to be involved in ribosomal activities (guanyl-specific ribonuclease Pb1; translation initiation factor SUI1), assistance in protein folding (peptidyl-prolyl cis-trans isomerase FKBP10), and cell wall stability and resistance to antifungal agents (cell wall mannoprotein PIR3). One particular SSP contained a cerato-platinin (CP) domain. Proteins with CP domains are preferentially located in fungal cell walls, and participate in growth and development in many fungi \[[@pone.0154122.ref108], [@pone.0154122.ref109], [@pone.0154122.ref110]\]. Other roles for CP proteins are in plant-fungal interactions not only as elicitors of plant defense mechanisms but also as effectors because they are able to bind chitin and its oligomers, thus, avoiding PAMP-triggered immunity detection during plant invasion \[[@pone.0154122.ref110]\]. A putative homolog to the extracellular effector Cf_Ecp6 of *Cladosporium fulvum* \[[@pone.0154122.ref067]\] is overrepresented at 48 hai. The higher expression of this gene *in planta* compared to that *in vitro* was confirmed in our qRT-PCR analysis. Cf_Ecp6 has high affinity for short chitin oligosaccharides and competes efficiently with extracellular plant receptors for degraded chitin fragments preventing activation of PTI in plant hosts \[[@pone.0154122.ref069]\]. In our transcriptome analysis at 48 hai we identified plant transcripts (data not shown) with very high homology to OsCERK1, the rice chitin elicitor receptor kinase 1. OsCERK1 forms a hetero-oligomer receptor complex with a glycoprotein, chitin-binding receptor CEBiP that upon recognition of chitin fragments triggers production of reactive oxygen species, diterpenoid phytoalexins, and expression of basal defense genes in rice \[[@pone.0154122.ref111], [@pone.0154122.ref112]\]. Even though a transcript similar to CEBiP was not identified at this time point, four other wildrice LysM containing proteins (receptor-like proteins) were expressed. Thus, *C*. *miyabeanus* might avoid elicitation of wildrice defense mechanisms by sequestering chitin fragments.
Proteins synthesized by *NRPS* and *PKS* genes are important in fungal fitness, response to environmental cues, and interaction with other organisms \[[@pone.0154122.ref086]\]. In other *Cochliobolus* species a great number of these metabolites are HSTs \[[@pone.0154122.ref012],[@pone.0154122.ref014],[@pone.0154122.ref021],[@pone.0154122.ref022],[@pone.0154122.ref024]\]. A subset of *NRPS* genes are discontinuously dispersed among *Cochliobolus* species, and could evolve rapidly by recombination, rearrangement and gain/loss of their AMP domains \[[@pone.0154122.ref040]\] to produce new phytotoxins. For instance, a protein of this class found in *C*. *sativus* pathotype 2, is thought to play a major role in causing virulence on barley cultivar Bowman \[[@pone.0154122.ref024]\]. One *Cm*TG12bL2 protein (CM_3163) belongs to the rapidly evolving and expandable NPS1/NPS3/NPS13 group and has a homolog in the WK-1C strain. In the phylogenetic analysis, adenylation domains of CM_3163 are closely related to AMP1_1 of Ch_NPS13 (*Cm*TG12bL2 AMP1_4), AMP2_4 of Ch_NPS3 (*Cm*TG12bL2 AMP2_4 and AMP3_4) and AMP3_4 of Ch_NPS3 (*Cm*TG12bL2 AMP4_4). Both, duplication of portions of *NRPS* genes (i.e. *NRPS3*), and either recombination or gene fusion (i.e. *NRPS13*) may have generated the two genes unique to *C*. *miyabeanus* strains. The fact that AMP domains of the mono-modular NPS11 from *C*. *heterostrophus* and a mono-modular NRPS-like gene from *C*. *miyabeanus* clustered separately, the first one with AMP1_2 and the second with AMP2_2 of bi-modular NRPSs GliP of *A*. *fumigatus* and SirP of *Leptosphaeria maculans*, suggests differential retention of AMP domains from an ancestral bi-modular *NRPS* gene.
Overall, 19 PKSs were detected within the *Cm*TG12bL2 predicted set of proteins, including the 10 PKSs common to all *Cochliobolus* species. However, *Cm*TG12bL2 isolate has only one of the two *PKS14* genes found in *C*. *miyabeanus* WK-1C \[[@pone.0154122.ref024]\]. Overall the number of NRPS and PKS found in our assembly is within the ranges of these proteins reported for *Cochliobolus* species \[[@pone.0154122.ref024]\]. The wildrice isolate appears to contain the same suite of NPSs and a similar number of PKSs as the rice *C*. *miyabeanus* WK-1C isolate \[[@pone.0154122.ref024]\] with no additional unique genes.
Cytochrome P450 monooxygenases have been associated with adaptation of fungi to new niches because they catalyze the degradation of chemical substances, and might facilitate pathogenesis \[[@pone.0154122.ref092]\]. Reduction in phenolic compounds content has been reported in rice leaves at 48 hai with a pathogenic strain of *C*. *miyabeanus* \[[@pone.0154122.ref113]\]. Three enzymes with homology to benzoate 4-monooxygenases (CYP548, CYP552, and CYP583) and another similar to a benzoate para-hydrolase of family CYP53 were highly expressed during *Cm*TG12bL2 colonization of a wildrice cultivar with improved resistance to FBS. A putative ortholog, *BPH*, from *C*. *heterostrophus* was upregulated during maize infection, suggesting that maize defenses could involve benzoate biosynthesis in \[[@pone.0154122.ref114]\]. In *C*. *lunatus*, a similar protein (bph) \[[@pone.0154122.ref094]\], was shown to be a key enzyme in benzoate detoxification in a dose- and time-dependent manner. Thus, some *Cm*TG12bL2 cytochrome P450 monooxygenases could participate in degrading phenolic compounds synthesized by wildrice during fungal infection. A putative salicylate hydroxylase was upregulated at 24 hai and at 48 hai. Similar proteins in *Fusarium graminearum* degrade the signaling molecule salicylic acid, necessary for plant defense \[[@pone.0154122.ref115]\]. Salicylic acid-induced transcripts were found in the wildrice transcriptome (data not shown) and could be related to defense against the fungus. Identification of fungal genes implicated in the detoxification or inhibition of host compounds is important for identifying the plant defense mechanisms used for counterattack of fungal colonization.
Our results are in agreement with a study of proteomics of *C*. *miyabeanus* during infection on rice \[[@pone.0154122.ref105]\] and the *Bipolaris sorghicola* transcriptome during sorghum infection \[[@pone.0154122.ref116]\]. Particularly the findings of fungal oxidative stress activity, expression of CAZymes such as α-L-arabinofuranosidase, xylanase and glucanase, acetylxylan esterase, and LysM domain containing proteins, as well as expression of cutinases, and Alp1 transcripts, suggesting commonalities of virulence mechanisms among *Cochliobolus*/*Bipolaris* species infecting monocots. The *Cm*TG12bL2 genome and transcriptome assemblies and analyses contribute to a better understanding of fungal pathogenicity and open new avenues for targeted mutagenesis in this pathosystem (i.e. cysteine-rich SSP, putative effectors molecules, and CYPs). Further, it offers genomic resources for the interpretation of the *C*. *miyabeanus* infected wildrice transcriptome.
Conclusions {#sec025}
===========
The genome of *Cochliobolus miyabeanus* isolate *Cm*TG12bL2 pathogenic on wildrice was sequenced using Illumina short-read technology, together with the transcriptome *in planta* after infection. Proteins involved in ROS scavenging, plant tissue degradation and carbohydrate binding, SSPs, a known effector, and detoxification systems were expressed in the infected leaves. This study advances our understanding of fungal pathogenicity on a distant relative of rice, American wildrice. The predicted gene and protein sets will facilitate targeted mutagenesis to infer the functions of pathogenicity and effector genes, as well as comparative transcriptomics of *Cochliobolus* pathogens when colonizing different hosts (i.e. wildrice, switchgrass, and common rice). Further, it may help in refining host breeding strategies for developing more effective genetic resistance against *C*. *miyabeanus*.
Supporting Information {#sec026}
======================
###### Length of N~50~, total draft genome, and maximum scaffold with varying *k*-mer values before parameters optimization.
A. Length of N~50~ values at *k*-mers varying from 31 to 91 nucleotides increasing by 2 nucleotides. Black column indicates the longest N~50~ length value (75,371bp) for 65 *k*-mer. B. Total length of the draft *C*. *miyabeanus* genome (Mbp) produced by assemblies with increased *k*-mer lengths from 31 to 91 nucleotides. Black column indicates initial genome draft size (31,696,836 bp) for 65 *k*-mer. C. Black column indicates maximum scaffold size (348,044 bp) for *k*-mer 65.
(DOCX)
######
Click here for additional data file.
###### Distribution of contig-weighted coverage of *Cochliobolus miyabeanus Cm*TG12bL2 assembly using Velvet.
The white arrow indicates average weighted coverage (76.56X) of the assembly. The black arrow indicates the cut-off value (38.28X). These two values were used for optimizing the final *Cm*TG12bL2 assembly.
(DOCX)
######
Click here for additional data file.
###### Genome-wide DotPlot of *Cochliobolus miyabeanus* TG12bL2 and *C*. *miyabeanus* WK-1C.
Ordered scaffolds of both species were used to build a dotplot matrix in Gepard software that uses suffix array data for heuristic dotplot computation. Minimum word length = 10.
(DOCX)
######
Click here for additional data file.
###### Alignment of effectors from *Cladosporium fulvum* with putative homologous fungal proteins.
A. Multiple alignment of proteins with homology to Cf_Ecp2 = *Cladosporium fulvum*, Ch = *C*. *heterostrophus*; COCHEDRAFT_1130950\_(gi\|452001939), Cs = *C*. *sativus*; COCSADRAFT_196914\_(gi\|451853714), Cm = *C*. *miyabeanus* TG12bL2; CM_6804, and Fg = *Fusarium graminearum* (XP391494). B. Pairwise alignment of the Cf_Ecp6 and CM_2799 protein from *C*. *miyabeanus* TG12Lb2. Alignments were done with ClustalW (<http://www.genome.jp/tools/clustalw/>) with a dynamic programming method and using a default Blosum (for proteins) matrix, with gap penalties of 10 and gap extension penalties of 0.05. Identical amino acid residues are indicated with asterisks and highlighted in light gray, cysteines are highlighted in dark gray and LysM domains according to Cf_Ecp6 are indicated by continuous lines.
(DOCX)
######
Click here for additional data file.
###### Phylogenetic tree of adenylation binding domains of NRPS proteins of *Cochliobolus* species and other Dothediomycetes.
Neighbor joining analysis of adenylation domains (AMPs) sequences of NRPS proteins (Left): I. AMPs of conserved NRPS proteins across *Cochliobolus* species (Red); II. AMPs of NRPS proteins with less degree of conservation in *Cochliobolus* species (Blue). III. AMPs of NRPS proteins with discontinuous distribution in *Cochliobolus* species (NRPS expanded group). IV. Outgroup set: Related adenylation modules: Long Chain Fatty Acid ligases (LCFA), Acyl-CoA synthetases (Acyl-CoA-Synth), Ochratoxin synthetases (ochratoxins), Acyl-CoA ligase (CPS1). ChetC4: *Cochliobolus heterostrophus* C4, ChetC5v2.0: *C*. *heterostrophus* C5, Cmiya: *C*. *miyabeanus* WK-1C, and Cmwr: *C*. *miyabeanus* TG12bL2. Other sequences are described in \[[@pone.0154122.ref040]\]. Blow-up clusters (Center): Main clusters harboring AMP domains related to Ch_NPS11 and GliP (gliotoxin) and SirP (sirodesmin) proteins sequences and of *Cm*TG12bL2 NPS1/NPS3/NPS13 expanded protein. Inset (Right): Graphic representation of AMP domains belonging to *C*. *heterostrophus* NPS3 and NPS13 (blue) and to *C*. *miyabeanus* NPS1/NPS3/NPS13 expanded protein (green).
(PDF)
######
Click here for additional data file.
###### Functional annotation of *Cochliobolus miyabeanus* TG12bL2 transcripts at 48 hours after inoculation of wildrice leaves.
The sequence distributions by GO terms (level 2) of the fungal transcripts were done with Blast2Go software. A. *Biological processes*: 1. immune system process (GO:0002376), 2.single-organism process (GO:0044699), 3. response to stimulus (GO:0050896), 4. biological adhesion (GO:0022610), 5. cellular process (GO:0009987), 6. metabolic process (GO:0008152), 7. rhythmic process (GO:0048511), 8. cellular component organization or biogenesis (GO:0071840), 9. developmental process (GO:0032502), 10. reproduction (GO:0000003), 11. biological regulation (GO:0065007), 12. growth (GO:0040007), 13. locomotion (GO:0040011), 14. multi-organism process (GO:0051704), 15. localization (GO:0051179), 16. multicellular organismal process (GO:0032501), 17. signaling (GO:0023052). B. *Cellular component*: 1. extracellular region (GO:0005576), 2. nucleoid (GO:0009295), 3. symplast (GO:0055044), 4. synapse (GO:0045202), 5. virion (GO:0019012), 6. membrane (GO:0016020), 7. cell (GO:0005623), 8. organelle (GO:0043226), 9. macromolecular complex (GO:0032991), 10. membrane-enclosed lumen (GO:0031974), 11. cell junction (GO:0030054). C. *Molecular function*: 1. receptor activity (GO:0004872), 2. molecular transducer activity (GO:0060089), 3. protein binding transcription factor activity (GO:0000988), 4. nucleic acid binding transcription factor activity (GO:0001071), 5. structural molecule activity (GO:0005198), 6. electron carrier activity (GO:0009055), 7. antioxidant activity (GO:0016209), 8. enzyme regulator activity (GO:0030234), 9. binding (GO:0005488), 10. nutrient reservoir activity (GO:0045735), 11. catalytic activity (GO:0003824), 12. transporter activity (GO:0005215).
(TIF)
######
Click here for additional data file.
###### Relative expression of *Cm*TG12Lb2 selected genes by qRT-PCR at 24 h and 48 h after inoculation and transcriptome expression levels at 48 h after inoculation.
A. The vertical axis shows the log ~2~ of relative quantification (fold change) for each of the selected genes. The horizontal axis indicates two time points of sample collection after *Cm*TG12Lb2 inoculation (24 h and 48 h). The reference gene expression was set to 0 (Log~2~ (1). Gene expression was normalized with glyceraldehyde-3-phosphate dehydrogenase expression. B. Transcript abundance (RPKM Log~2~ transformed values) of six genes at 48 h after inoculation. Abreviations: *Ecp*6 = *Cladosporium fulvum Ecp*6 homolog, CYP53 = Cytochrome P450 CYP53, SH = salicylate hydroxylate, β-1,4-EG = β-1,4-endoglucanase, β-1,4-EX = β-1,4-endoxylanase, β-1,4-EG = β-1,4-glucosidase.
(TIF)
######
Click here for additional data file.
###### Basic statistics and quality measures of the *Cochliobolus miyabeanus* TG12bL2 sequencing process.
^a^ The overall % GC of all bases in the sequences followed a normal distribution overlapping the theoretical distribution. Consistently, there were no overrepresented sequences or *k*-mers. ^b^ Quality scores across all bases sequenced. ^c^ Percentage of duplicated sequences relative to unique sequences indicating that some sequences had 10 or more duplicates.
(DOCX)
######
Click here for additional data file.
###### Best hits of *C*. *miyabeanus* TG12bL2 genes to the CEGMA (Core Eukaryotic Gene Mapping Approach) gene list from six eukaryotic genomes: *Homo sapiens* (Hs), *Drosophila melanogaster*, *Caenorhabditis elegans* (CE), *Arabidopsis thaliana* (At), *Saccharomyces cerevisiae* (Y), and *Schizosaccharomyces pombe* (SP).
Function. *Cellular processes and signaling*: M: Cell wall/membrane/envelope biogenesis, O Posttranslational modification, protein turnover, chaperones, T: Signal transduction mechanisms, U: Intracellular trafficking, secretion, and vesicular transport, V: Defense mechanisms, Y: Nuclear structure, Z: Cytoskeleton. *Information storage and processing*: A: RNA processing and modification, B: Chromatin structure and dynamics, J: Translation, ribosomal structure and biogenesis, K: Transcription, L: Replication, recombination and repair. *Metabolism*: C: Energy, D: Cell cycle control, cell division, chromosome partitioning, E: Amino acid transport and metabolism, G: Carbohydrate transport and metabolism, H: Coenzyme transport and metabolism, I: Lipid transport and metabolism, P: Inorganic ion transport and metabolism, Q: Secondary metabolites biosynthesis, transport and catabolism. *Poorly characterized*: R: General function prediction only, S: Function unknown.
(XLSX)
######
Click here for additional data file.
###### Comparison of small secreted proteins in *C*. *miyabeanus Cm*TG12bL2 to those reported for eighteen Dothidiomycete genomes.
\(a\) HMMER using hmmscan algorithm to compare protein sequences against collections of profiles (only Pfam protein family was used). Proteins with transmembrane regions as detected by *Phobius* were removed from the analysis. (b) Values of HCSSPs calculated as in \[[@pone.0154122.ref028]\] (as least twice the general average of cysteines in the *Cm*TG12bL2 proteome), and (c) as in \[[@pone.0154122.ref024]\] (over 2% of cysteines). All SSPs were under 200 amino acids in length.
(XLSX)
######
Click here for additional data file.
###### Pfam domains identified in *C*. *miyabeanus* TG12bL2 small secreted proteins.
Pfam domains were identified using HMMER software (<http://hmmer.janelia.org>) with the algorithm/program hmmscan and a gathering cut off to compare protein sequences against collections of profiles (only Pfam protein family was used). Proteins with transmembrane regions as detected by *Phobius* were removed from the analysis.
(XLSX)
######
Click here for additional data file.
###### Distribution of carbohydrate active enzyme families in 12 Ascomycete genomes.
GH: Glycosyl hydrolases; ^(a)^: Indicates GH5 Sub family; GT: glycosyl transferases, PL: polysaccharide lyases, and CE: carbohydrate esterases, CBM: carbohydrate binding modules.
(XLSX)
######
Click here for additional data file.
###### Identification of CAZyme in *Cm*TG12bL2 genome associated to degradation of cellulose, hemicellulose, and pectin.
(XLSX)
######
Click here for additional data file.
###### *Cochliobolus miyabeanus* TG12bL2 predicted P450 monooxygenases.
Fungal Cytochrome P450 Database (FCPD; <http://p450.riceblast.snu.ac.kr/download.php>).
(XLSX)
######
Click here for additional data file.
###### CAZymes supplementary information.
(DOCX)
######
Click here for additional data file.
We would like to thank Bruce Elckloff from the Mayo Clinic, Rochester, MN for sequencing the *C*. *miyabeanus* isolate, and Dr. B. Gillian Turgeon and Bradford Condon for NRPS sequence data from the *C*. *miyabeanus* WK-1C and *C*. *heterostrophus* C4 and C5 strains, and other ascomycetes used in the analyses, and to Dr. B. Gillian Turgeon and the U.S. Department of Energy Joint Genome Institute (<http://genome.jgi.doe.gov>) for sharing the *C*. *miyabeanus* WK-1C DNA sequences and proteins. The genome and transcriptome of *C*. *miyabeanus* work were sequenced and annotated by the U.S. Department of Energy Joint Genome Institute, a DOE Office of Science User Facility, supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.
Mention of a trademark, proprietary product or vendor does not constitute a guarantee or warranty of the product by the USDA, and does not imply its approval to the exclusion of other products and vendors that might also be suitable.
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: CVC-M DAS. Performed the experiments: CVC-M. Analyzed the data: CVC-M JJG-G KEB MH BH. Contributed reagents/materials/analysis tools: ZJT. Wrote the paper: CVC-M. Edited the manuscript CVC-M DAS JJG-G ZJT KEB MH BH. Submitted the sequence data to GenBank: JJG-G CVC-M.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION {#sec1-1}
============
Ethylene glycol monomethyl ether (2-methoxyethanol \[2-ME\]) regarded as a potent occupational toxin and its toxicological manifestations are well known in humans. The main sources of 2-ME toxicity are of industrial products, including paints, inks, hydraulic fluids, and jet fuels.\[[@ref1]\] It has been reported that ME toxicity can interfere the physiological, biochemical, and behavioral in human and animal, which, including disorders of cardiovascular, brain, hematopoietic, renal, reproductive and hepatic system producing serious toxicity,\[[@ref2][@ref3][@ref4]\] while the likely effect of toxicity on the cardiovascular system hasn\'t been precisely proven.
Creatinine kinase-MB (CK-MB) and lactate dehydrogenase (LDH) are the indicative serum marker enzymes of cardiotoxicity. This enzymes escape via the injured tissues, and they are the most effective marker of cardiotoxicity because of their tissue selectivity and serum catalytic activity.\[[@ref5][@ref6]\]
The 2-ME mechanism induced toxicity and this occurs once there is unevenness between the formation of reactive oxygen species (ROS) and the glutathione (GSH), Catalase, superoxide dismutase (SOD), and glutathione peroxidase (GPx) in biological organelles.\[[@ref7][@ref8][@ref9]\]
A current research has proven that the ROS like superoxide ion (O~2~^−^), hydrogen peroxide (H~2~O~2~), and hydroxyl radical (OH ^−^) play an essential function in 2-ME-induced cellular toxicity.\[[@ref10][@ref11]\] Furthermore, ROS are very reactive to membrane lipids, protein, and are primarily responsible for causing stress injuries and rapid cellular damage. Malondialdehyde (MDA) is a trivial product of lipid peroxidation and can be used as a marker for cell injury. A rise in MDA levels depicts antioxidant failure and the development of free radicals improves lipid peroxidation thereby resulting in cellular toxicity.\[[@ref5][@ref6][@ref12]\] The MDA level is the direct evidence of tissue injury processes caused by free radicals. Several researchers have proven that antioxidants can prevent or inhibit the oxidation of molecules by the ROS in tissue or cellular organisms, scavenge free radicals which can weaken harmful effects of ROS.\[[@ref13][@ref14]\]
The role of antioxidant activity or inhibition of the production of free radicals plays an important role in the prevention of such oxidative damage. Hence, it has been claimed that protective agents against free radicals like antioxidants, might be effective prevention for oxidative stress induce tissue damage.\[[@ref8][@ref10][@ref15]\] Thus, it is believed that antioxidant should be one of the important components for treatment of ME poisoning. There is an increasing interest toward the use of naturally occurring phytochemicals with cardioprotective and antioxidant activity in 2-ME intoxication therapy.
Based on the aforementioned statement, numerous researches have been conducted with different natural product that contain antioxidant properties to examine their probable protective effects in 2-ME caused tissue damage. Among those *Mucuna pruriens*, *Withania somnifera*, and *Garcinia mangostana* have been proven to have protective tasks against 2-ME intoxication.\[[@ref7][@ref16][@ref17]\] *G. mangostana* L is one of the medicinal plant which also shown antioxidant activity. Phytochemical studies of *G. mangostana* show that this plant contains xanthones.\[[@ref18]\] The ethnopharmacological views of xanthone suggest remarkable properties such as antioxidants, antitumor, anti-inflammatory, analgesic, antiviral activities, cardioprotective effects, antifungal, antiallergy, antibacterial, antituberculosis, and immunomodulation.\[[@ref18][@ref19][@ref20]\] Xanthone has been shown to have strong antioxidant activity.\[[@ref21]\] Therefore, the aim of this study was to prove that the SOD and GPx have role important on xanthone in protected 2-ME-induced cardiac damage in mice.
MATERIALS AND METHODS {#sec1-2}
=====================
Experimental animal {#sec2-1}
-------------------
The experimental animals used were male mice, and they each weighed between 25 and 30 g (2.5--3 months). They were gotten from the Veterinary Farm, Surabaya, Indonesia for experimental use. The mice were kept in plastic cages with a constant temperature of 26°C ± 2°C and 12 h rotation of light and dark cycles. The mice drank tap water that contained *ad libitum*, and they ate economical rat food. This study was reviewed by the Ethical Clearance Committee for preclinical research, Faculty of Medicine, Airlangga University and obtained ethical clearance under No. 178/FK/1/2018. Date: March 14, 2019.
Experimental design {#sec2-2}
-------------------
The sample used 40 male mice were divided into five groups: control group (mice were given daily with distilled water); 2-ME group (mice were given with 2-ME, orally, at a dose of 200 mg/kg BW for the 35 consecutive days), and the treatment group (mice were given the xanthone 60 mg; 120 mg; 240 mg/kg BW orally once in a day for 40 days and on 5^th^ day, were given with 2-ME, orally, at a dose of 200 mg/kg BW 1 h after the xanthone). On day 40, the mice blood samples were taken by cardiac puncture to be measured levels of LDH and CK-MB. Furthermore, mice were sacrificed, and cardiac tissues were homogenized in ice-cold 50 mM sodium phosphate buffer (pH 7.4) containing 0.1 mM ethylenediamine tetraacetic acid. The supernatant was separated by centrifugation at 1000 g for 20 min at 4°C. The supernatant was used for the analyses of MDA. The cardiac tissues were also fixed in a 10% neutral-buffered formalin solution for immunohistochemical evaluation of the expression of SOD and GPx and histopathological evaluation of the cardiac damage.
Measurement of lactate dehydrogenase and creatine kinase-MB fraction {#sec2-3}
--------------------------------------------------------------------
The serum was examined for the existence of various enzymes linked to myocardial infarction like LDH and CK-MB fraction.\[[@ref5]\] All analyses were conducted with commercially available kits based on the references via the experimenter.
Measurement of malondialdehyde {#sec2-4}
------------------------------
MDA was observed in the supernatant of homogenate cardiac tissue using the thiobarbituric acid (TBA) technique. The concentration of MDA was acquired to be 532 nm by the absorbance coefficient of MDA-TBA complex.\[[@ref5]\] MDA is expressed as nanomoles MDA/g tissue.
Histopathological examination of heart damage {#sec2-5}
---------------------------------------------
Tissues extracted from the heart was placed in 10% neutral-buffered formalin solution, inserted in paraffin, and utilized for histopathological analyzes with hematoxylin and eosin stain.
Immunohistochemical examination of the expression of superoxide dismutase and glutathione peroxidase {#sec2-6}
----------------------------------------------------------------------------------------------------
The paraffin-inserted heart was sliced into 4 μm parts and firmly fixed on positively charged slides for the expression of caspase-3 immunohistochemistry. This was carried out using peroxidase/anti-peroxidase (PAP) procedure. Methanol consisting of 0.1% H~2~O~2~ was used for the extensive peroxidase reaction. The four parts were nurtured with natural goat serum to prevent extensive reactions, with some particular antibodies. The tissue parts were then cleaned with phosphate buffer and nurtured with secondary antibodies (1:2000; Sigma, USA). It was cleaned with phosphate buffer and nurtured with the PAP procedure (dilution, 1:200). The peroxidase reaction was carried out using a solution of 3, 3\'-diaminobenzidine tetrahydrochloride containing 0.01% H~2~O~2~ in Tris-HCl buffer (0.05 M, pH 7.6).
Statistical analysis {#sec2-7}
--------------------
Data were presented as means ± standard deviation. Oneway ANOVA has carried *post hoc* test and the statistical comparisons among the groups were performed with an LSD test using a statistical package program SPSS version 17.0 (SPSS Inc, Chicago, USA).
RESULTS {#sec1-3}
=======
Effects of xanthone on 2-methoxyethanol-induced changes in creatinine kinase-MB, lactate dehydrogenase and malondialdehyde {#sec2-8}
--------------------------------------------------------------------------------------------------------------------------
The results of the CK-MB, LDH, and MDA levels in each group are shown in [Table 1](#T1){ref-type="table"}. The administration of 2-ME on mice caused a significant increase of CK-MB, LDH, and MDA (*P* \< 0.05) when compared with the control group. However, the treatment with xanthone in mice only at dose 240 mg/kg BW but not at dose 60 mg/kg and 120 mg/kg BW showed a significant decrease (*P* \< 0.05) in CK-MB, LDH, and MDA level as compared to the 2-ME group.
######
Cardioprotective effect of xanthone on CK-MB, LDH and MDA against 2-ME induce cardiotoxicity
Groups Means±Standard Deviation
----------------------- -------------------------- ---------------- ---------------
Control group 66.7^a^±6.35 101.6^a^±8.41 52.72^a^±4.18
2-ME group 104.5^b^±8.53 154.7^b^±13.42 76.23^b^±6.16
Xanthone 60 mg/kg BW 98.2^b^±7.82 148.4^b^±16.72 73.42^b^±5.32
Xanthone 120 mg/kg BW 91.7^b^±6.35 141.7^b^±9.37 67.79^b^±4.36
Xanthone 240 mg/kg BW 76.1^c^±5.78 126.3^c^±9.63 59.36^c^±4.48
^a,b,c^Different superscript within each column indicate significant difference between the means (*P*\<0.05)
Effects of xanthone on the superoxide dismutase expression in against 2-methoxyethanol-induced cardiotoxicity {#sec2-9}
-------------------------------------------------------------------------------------------------------------
[Table 2](#T2){ref-type="table"} and [Figure 1](#F1){ref-type="fig"} showed the results of xanthone on the expression of SOD in against 2-ME-induced cardiac cell toxicity. The administration of 2-ME on mice caused a significant decrease in the expression of SOD of cardiac tissue compared to the control group (*P* \< 0.05). The treatment xanthone group at dose 240 mg/kg BW but not at dose 60 mg/kg and 120 mg/kg BW increase cardiac tissue SOD expression in mice induced-2-ME which significantly was different to the 2-ME group (*P* \< 0.05).
######
Cardioprotective effect of xanthone on SOD and GPx expression in against 2-ME induce cardiotoxicity
Group Means±Standard deviation
----------------------- -------------------------- -------------
Control group 9.8^a^±1.14 8.7^a^±0.94
2-ME group 3.2^b^±0.52 2.6^b^±0.52
Xanthone 60 mg/kg BW 2.9^b^±0.51 3.3^b^±0.71
Xanthone 120 mg/kg BW 4.8^b^±0.62 3.9^b^±0.87
Xanthone 240 mg/kg BW 7.2^c^±0.89 6.2^c^±0.69
^a,b,c^Different superscript within each column indicate significant difference between the means (*P*\<0.05)
{#F1}
Effects of xanthone on the glutathione peroxidase expression in against 2-methoxyethanol induce cardiotoxicity {#sec2-10}
--------------------------------------------------------------------------------------------------------------
[Table 2](#T2){ref-type="table"} and [Figure 2](#F2){ref-type="fig"} showed the results of xanthone on the expression of GPx in against 2-ME induce cardiac cell toxicity. The administration of 2-ME on mice caused a significant decrease in the expression of GPx of cardiac tissue was compared to the control group (*P* \< 0.05). Treatment with xanthone at dose 240 mg/kg BW but not at dose 60 mg/kg and 120 mg/kg BW increase cardiac tissue GPx expression which significantly was different to the 2-ME group (*P* \< 0.05).
{#F2}
Effects of xanthone on 2-methoxyethanol induce cardiac cell damage {#sec2-11}
------------------------------------------------------------------
Histopathological study was conducted using light microscopy. Histological investigation on the control group showed that in a cardiac cell have a normal structure. In the administration of 2-ME in mice showed cardiac cell damage (necrosis). In the treatment with xanthone, the number and morphological integrity of cardiac cells are being maintained. The results show that the cardiotoxic effects of 2-ME were inhibited by xanthone \[[Figure 3](#F3){ref-type="fig"}\].
{#F3}
DISCUSSION {#sec1-4}
==========
The CK-MB and LDH are biomarkers measured to evaluate heart function. They can be useful in the early prediction of cardiotoxicity. The serum CK-MB and LDH level are the best markers of cardiotoxicity due to cardiac tissue damage.\[[@ref5][@ref6]\] In our results showed that the administration of 2-ME resulted in a significant increase in the level of the LDH and CK-MB was compared with the control group. This suggests that the 2-ME might cause cardiac lipid peroxidation leading to cardiac cell damage followed by the secretion of CK-MB and LDH into the serum.
Toxic effects mechanism of 2-ME are via oxidative damage on the cardiac cell by increasing the generation of ROS which consist mainly of O~2~, H~2~O~2~ and OH ^-^. Reactions of these ROS with cellular biomolecules have been shown to lead to lipid peroxidation, membrane protein damage, altered antioxidant system, DNA damage, altered gene expression, and apoptosis.\[[@ref13][@ref15]\] In addition, 2-ME also decreasing endogenous antioxidants which cause significantly to the development of toxic oxidative stress.\[[@ref8]\] If these ROS-mediated oxidative stress are not balanced by repair processes, affected cells undergo apoptosis or necrosis.\[[@ref11]\]
When the antioxidant failure to inhibit the formation of ROS (free radicals) so can enhance lipid peroxidation and causing cellular toxicity, which produces MDA. In our investigation showed that the administration of 2-ME on mice can cause a significant increase in cardiac tissue MDA. This result suggests that 2-ME-induced cardiotoxicity on mice can be caused by increasing free radicals.
ROS are efficiently detoxified by antioxidant enzymes such as SOD and GPx in normal healthy conditions. It has been reported that antioxidant activity or inhibition of generation of free radicals plays a crucial role in protection against 2-ME-induced toxicity. In the administration of 2-ME, the expression of SOD and GPx of cardiac tissue was significantly decreased compared to the control group. The activities of SOD and GPx have been used to determine oxidative stress in cells. This treatment with 2-ME inhibited SOD, and GPx activities are in agreement with previous studies. This results suggested that administration of 2-ME can cause oxidative stress by decreasing the activity of antioxidant enzyme SOD and GPx. The decrease in the functions of SOD and GPx in 2-ME-induced cardiotoxicity might be as a result of an increased production of reactive oxygen radicals like superoxide and hydrogen peroxide, which in turn causes the prohibition of the functions of these enzymes. The outcomes are the same with the discoveries of Adedara and Farombi (2010) that recorded a huge decrease in SOD, Catalase, and GPx activities and a notable increase in the MDA level in the testis of rats susceptible to 2-ME treatment. The administration of herbal medicine that contains antioxidant properties such as *M. pruriens*, *W. somnifera*, and *G. mangostana* have been proven can be used as a protector on 2-ME intoxication through the increasing SOD and GPx expression, and also inhibiting LDH, CK-MB and MDA.\[[@ref7]\]
It has been reported that protective agents against free radicals, such as antioxidants, might be effective preventives for 2-ME toxicity. The natural product can be a better alternative as the antioxidant as a result of their economic costs, availability, and deficiency of undesirable adverse effects.\[[@ref13][@ref15]\]
This research concentrated on the protective effect of herbal medicine having antioxidant properties of xanthone for protective in decreasing free radical-induced cardiac cell damage. Phytochemical studies of *G. mangostana* prove that this plant contains xanthones which have very strong antioxidant activity. The administration of the xanthone, only doses 240 mg/kg BW significantly decreased LDH, CK-MB, MDA, and increase expression of SOD and GPx on cardiac cell damage induced by 2-ME. The xanthone 240 mg/kg BW also demonstrated significantly improved heart cell damage. It has been reported that xanthone is a well-known scavenger of free radicals and can greatly prohibit the production of ROS like superoxide anions, H~2~O~2~, and OH-radical generation.\[[@ref21]\] This effect has played an important role in decreased MDA, CK-MB, LDH and increased SOD, GPx expression in against 2-ME-induced cardiotoxicity. It also improved heart cell damage. These result indicated that the cardioprotective effects of xanthone are greatly related to play an essential role in reducing LDH, CK-MB, MDA and raising SOD, GPx expression against 2-ME-induced cardiotoxicity. It also can inhibit cardiac cell injury. The results prove that the cardioprotective consequences of xanthone are hugely linked to their antioxidative activities. Devi and Vijayaraghavan (2007) also reported that xanthone decreased MDA, CK-MB, LDH, and increased SOD, GPx level on isoproterenol-induced myocardial infarct.
CONCLUSION {#sec1-5}
==========
The outcome of the current research showed that the 2-ME-induced cardiotoxicity can be associated with oxidative stress. The treatment with xanthone can prevent 2-ME-induced cardiac cell damage through increasing antioxidant enzyme (SOD and GPx) and inhibiting MDA, CK-MB, and LDH. Furthermore, protective effects of xanthone can be developed to treat patients with 2-ME-induced testicular toxicity.
Financial support and sponsorship {#sec2-12}
---------------------------------
This work was supported in part by grant from the Educational Fund Management Institution (LPDP), Ministry of Finance, Republic of Indonesia, through Doctoral Program, Faculty of Medicine, Airlangga University, Surabaya, Indonesia.
Conflicts of interest {#sec2-13}
---------------------
There are no conflicts of interest.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#sec0005}
============
Chronic inflammatory disorders, including psoriasis, are characterized by enhanced atherosclerosis and consequently higher cardiovascular morbidity and mortality rates compared with the general population.[@bib0005], [@bib0010], [@bib0015], [@bib0020]
Therefore, particular attention should be paid to conventional cardiovascular risk factor treatment, including dyslipidemia, in these patients. Statins are effective in reducing disease activity, and in *post hoc* analysis of randomized clinical trials, statins improved lipid levels and cardiovascular outcomes in patients with and without psoriasis, supporting statin use in patients with psoriasis.[@bib0025] However, there is no certain indication for the use of lipid-lowering therapy only on the basis of the presence of the disease.
The traditional risk scales used to estimate cardiovascular risk have great limitations; due to the fact that they were not developed specifically for psoriasis, they have a tendency to underestimate the risk.[@bib0030], [@bib0035] The decrease in the indication of interventions in cardiovascular prevention, such as statins, may be the result of this deficient evaluation.
Two strategies have been proposed for the cardiovascular risk management of these patients. The first, recommended by the European Society of Cardiology (ESC), the European Society of Atherosclerosis (EAS), and the European League Against Rheumatism (EULAR), is to adjust the risk calculated by a multiplying factor (1.5×) and follow the recommendations for statin therapy of the general population.[@bib0040], [@bib0045]
The second strategy places psoriasis as a clinical situation that increases cardiovascular risk and consequently, it favors the indication of statins at least in subjects with intermediate risk. In this case, no adjustment factor is suggested. This strategy is recommended by the new American College of Cardiology/American Heart Association (ACC/AHA) guidelines for cholesterol management introduced at the end of 2018.[@bib0050]
Taking into account the previously mentioned considerations, the objectives of this study were as follows: (1) to estimate the cardiovascular risk measured by two different scores in patients with psoriasis without previous cardiovascular disease (CVD); (2) to evaluate two cardiovascular prevention strategies in patients with psoriasis, analyzing which proportion of patients would be candidates to receive statin therapy; (3) to establish the reasons that justify this indication.
Methods {#sec0010}
=======
A retrospective cohort was selected from a secondary database (electronic medical records). The sample was obtained from a private health system constituted by two university hospitals and a network of 21 associated peripheral centers distributed in the city of Buenos Aires and in the province of Buenos Aires, Argentina. All patients older than 18 years with a diagnosis of psoriasis were included. Patients with previous CVD, chronic renal insufficiency, or concomitant lipid-lowering treatment were excluded.
The atherosclerotic cardiovascular disease (ASCVD) calculator used by the 2018 ACC/AHA guidelines for cholesterol management[@bib0050] and the Systematic Coronary Risk Evaluation (SCORE) estimate the ten-year risk of ASCVD fatal events, corresponding to low-risk countries used by the 2016 ESC/EAS guidelines on CVD prevention in clinical practice,[@bib0040] were calculated. The choice of SCORE, corresponding to low risk countries, was arbitrary, based on the fact that most of the Argentine immigrant population comes from those countries. The risk value calculated by the SCORE was adjusted by a multiplication factor of 1.5, following the latest European recommendations.[@bib0040], [@bib0045] Following the recommendations of the mentioned guidelines, indications for statins with a level of recommendation I or IIa were selected for this analysis.
Applying the 2018 ACC/AHA guidelines, the following recommendations were taken into account for patients in primary prevention: (a) in patients 40--75 years of age with diabetes mellitus and LDL-C ≥70 mg/dL, start moderate-intensity statin therapy without estimating the ten-year ASCVD risk calculator; (b) in patients 20--75 years of age with an LDL-C level of 190 mg/dL or higher, high-intensity statin therapy is recommended without estimating the ten-year ASCVD risk calculator; (c) in adults 40--75 years of age without diabetes mellitus and ten-year risk ≥20%, start high-intensity statin therapy; (d) in adults 40--75 years of age without diabetes mellitus and ten-year risk of 7.5--19.9% (intermediate risk), risk-enhancing factors favor initiation of moderate-intensity statin therapy. Risk-enhancing factors include psoriasis.
Applying the 2016 ESC/EAS guidelines, the following recommendations were taken into account for patients in primary prevention: (a) in patients with diabetes \>40 years of age, start moderate/high-intensity statin therapy without estimating SCORE risk calculator; (b) in patients with an LDL-C level of 190 mg/dL or higher, statin therapy is recommended; (c) in adults \>40 years of age without evidence of CVD or diabetes with a calculated SCORE ≥1% and \<5% for ten-year fatal CVD and LDL-C ≥100 mg/dL, start statin therapy; (d) in adults \>40 years of age without evidence of CVD or diabetes with a calculated SCORE ≥5% and \<10% for ten-year fatal CVD and LDL-C ≥70 mg/dL, start statin therapy; (e) in adults \>40 years of age without evidence of CVD or diabetes with a calculated SCORE ≥10% for ten-year risk of fatal CVD, start statin therapy.
Continuous data between two groups were analyzed using Student\'s *t*-test if the variables were normally distributed or with the Wilcoxon--Mann--Whitney test otherwise. Categorical data analysis was performed using the chi-squared test. Continuous variables are given as mean ± standard deviation, while categorical variables are given as percentages. The agreement between both strategies in selecting patients with statin indication was analyzed, using the Fleiss kappa index. Mild or poor, acceptable or discrete, moderate, significant, or almost perfect agreement was defined if the kappa value was \<0.20, between 0.21 and 0.40, 0.41 and 0.60, 0.61 and 0.80, or 0.81 and 1, respectively. The chi-squared test for homogeneity was performed to compare between kappa values. A value of *p* \< 0.05 was considered statistically significant. STATA v. 13.0 and EPIDAT v. 3.1 software packages were used for statistical analysis.
The study was conducted in compliance with the recommendations for medical research contained in the Declaration of Helsinki, the Good Clinical Practice standards, and the applicable ethical regulations.
Results {#sec0015}
=======
A total of 892 patients (mean age 59.9 ± 16.5 years, 54.5% women) were included in the study. The average body mass index was 28.2 ± 5.9 and the mean HDL-C, triglyceride, and total cholesterol values were 51.5 ± 14.5 mg/dL, 119.6 ± 70.5 mg/dL, and 198.3 ± 42.2 mg/dL, respectively. Importantly, 54.3% of patients were hypertensive and 24.8% were active smokers. The baseline characteristics of the population are described in [Table 1](#tbl0005){ref-type="table"}.Table 1Characteristics of the population.Table 1*Continuous variables, mean (SD)n* = 892 Age, years59.9 (16.5) Systolic blood pressure, mmHg127.5 (15.9) Diastolic blood pressure, mmHg77.9 (9.7) Body mass index, kg/m^2^28.2 (5.9) Total cholesterol, mg/dL198.3 (42.2) LDL-C, mg/dL123.6 (34.2) HDL-C, mg/dL51.5 (14.5) Triglycerides, mg/dL119.6 (70.5) Non HDL-C, mg/dL146.9 (40.5) Blood glucose, mg/dL100.8 (25.8) HbA1c, % (patients with diabetes)6.3 (1.0) Creatinine, mg/dL0.87 (0.22)
*Categorical variables (%)* Male gender45.5 Hypertension54.3 Smoking24.8 Diabetes13.9 Obesity30.9 Psoriatic arthritis7.0
The median 2018 ACC/AHA score and SCORE risk calculator values were 13.4% (IQR 6.1--27.0%) and 1.9% (IQR 0.4--5.2), respectively. According to the 2018 ACC/AHA score, 20.1%, 11.0%, 32.9%, and 36.4% of the population was classified at low, borderline, moderate, or high risk, respectively. Applying the SCORE risk calculator, 26.5%, 42.9%, 20.8%, and 9.8% of patients were stratified as having low, moderate, high, or very high risk, respectively.
Overall, the proportion of subjects with statin indication was similar using both strategies (*p* = 0.91). According to the 2018 ACC/AHA guidelines and based on the ASCVD calculator, the use of statins was recommended in 60.1% of cases. When the 2016 ESC/EAS guidelines were applied using the SCORE risk calculator, statins were recommended in 60.9% of cases. However, the concordance between both strategies in selecting patients with statin indication was moderate (*κ* = 0.46). The reasons why statin therapy was indicated using both strategies are shown in [Figure 1](#fig0005){ref-type="fig"}, [Figure 2](#fig0010){ref-type="fig"}.Figure 1The reasons why statin therapy was indicated using the 2016 European Society of Cardiology/European Society of Atherosclerosis (ESC/EAS) guidelines.Figure 2The reasons why statin therapy was indicated using 2018 American College of Cardiology/American Heart Association (ACC/AHA) guidelines.
The patients with statin therapy indication, by both guidelines, showed more cardiovascular risk factors and a higher prevalence of psoriatic arthritis than subjects without this pharmacological indication. Characteristics of the population according to the indication of statin therapy recommended by the 2018 ACC/AHA and 2016 ESC/EAS guidelines can be seen in [Table 2](#tbl0010){ref-type="table"}.Table 2Characteristics of the population according to the indication of statin therapy recommended by the 2018 ACC/AHA and the 2016 ESC/EAS guidelines.Table 22018 ACC/AHA guidelinesWith statin-indication (*n* = 536)Without statin-indication (*n* = 356)*pContinuous variables, mean (SD)* Age, years63.7 (11.6)54.3 (12.7)\<0.001 Systolic blood pressure, mmHg131.4 (16.2)121.7 (13.7)\<0.001 Diastolic blood pressure, mmHg79.8 (9.7)75.0 (9.1)\<0.001 Body mass index, kg/m^2^28.9 (6.2)27.1 (5.6)0.002 Total cholesterol, mg/dL202.1 (44.1)192.7 (38.4)0.001 LDL-C, mg/dL128.1 (35.9)117.3 (30.6)\<0.001 HDL-C, mg/dL49.8 (13.8)54.1 (15.1)\<0.001 Triglycerides, mg/dL125.7 (67.9)110.4 (73.4)0.002 Non HDL-C, mg/dL152.4 (42.1)138.7 (36.3)\<0.001 Blood glucose, mg/dL103.9 (29.4)96.2 (18.2)\<0.001 Creatinine, mg/dL0.91 (0.26)0.88 (0.26)0.362
*Categorical variables (%)* Male gender57.527.5\<0.001 Hypertension66.935.1\<0.001 Smoking29.517.7\<0.001 Diabetes14.912.40.278 Obesity37.421.4\<0.001 Psoriatic arthritis8.64.50.0192016 ESC/EAS guidelinesWith statin-indication (*n* = 536)Without statin-indication (*n* = 356)*pContinuous variables, mean (SD)* Age, years67.4 (12.7)47.9 (14.7)\<0.001 Systolic blood pressure, mmHg130.7 (16.3)122.5 (13.9)\<0.001 Diastolic blood pressure, mmHg78.7 (9.6)76.6 (9.7)0.002 Body mass index, kg/m^2^28.6 (6.1)27.5 (5.8)0.074 Total cholesterol, mg/dL203.2 (42.6)190.5 (40.2)0.001 LDL-C, mg/dL129.4 (34.3)115.5 (32.4)\<0.001 HDL-C, mg/dL51.5 (14.9)51.3 (13.9)0.865 Triglycerides, mg/dL124.2 (70.8)112.2 (69.5)0.014 Non HDL-C, mg/dL151.7 (40.7)139.2 (38.8)\<0.001 Blood glucose, mg/dL105.4 (30.6)93.5 (12.2)\<0.001 Creatinine, mg/dL0.91 (0.27)0.87 (0.64)0.341
*Categorical variables (%)* Male gender50.637.4\<0.001 Hypertension66.235.1\<0.001 Smoking26.721.60.08 Diabetes21.91.4\<0.001 Obesity34.325.4\<0.001 Psoriasic arthritis7.95.40.196
In the analysis according to sex, men had a greater indication for statin therapy compared to women (75.9% *vs*. 46.9%, *p* \< 0.001) according to the 2018 ACC/AHA guidelines. Similar findings were found when applying 2016 ESC/ESC guidelines (men 67.7% *vs*. women 55.1%, *p* \< 0.001). The application of the 2018 ACC/AHA guidelines selected a higher proportion of men with statin indication in comparison with the 2016 ESC/ESC guidelines (75.9% *vs*. 67.7%, *p* = 0.01). Contrarily, the use of the 2016 ESC/ESC guidelines selected a higher proportion of women with statin indication in comparison with the 2018 ACC/AHA guidelines (55.1% *vs*. 46.9%, *p* = 0.005).
The concordance between two strategies in selecting patients with statin indication was moderate in both sexes (men: *κ* = 0.46, women: *κ* = 0.42; *p* = 0.29). The reasons why statin therapy was indicated using both strategies in the analysis according to sex showed in [Figure 3](#fig0015){ref-type="fig"}, [Figure 4](#fig0020){ref-type="fig"}.Figure 3The reasons why statin therapy was indicated using the 2016 European Society of Cardiology/European Society of Atherosclerosis (ESC/EAS) guidelines according to sex.Figure 4The reasons why statin therapy was indicated using the 2018 American College of Cardiology/American Heart Association (ACC/AHA) guidelines according to sex.
Discussion {#sec0020}
==========
Psoriasis is a chronic inflammatory skin disease associated with increased cardiovascular morbidity and mortality.[@bib0055], [@bib0060] Several mechanisms have been proposed to explain the relationship between psoriasis and cardiovascular risk. Indeed, patients with psoriasis have an increased prevalence of classic cardiovascular risk factors, including obesity, hypertension, diabetes, dyslipidemia, metabolic syndrome, and nonalcoholic fatty liver disease.[@bib0065], [@bib0070], [@bib0075], [@bib0080] However, psoriasis may provide an additional and independent cardiovascular risk factor, most likely because several cytokines (tumor necrosis factor-alpha, interferon, interleukin-17, interleukin-6) released by skin lesions can directly favor the development and progression of atherosclerosis.[@bib0085] Likewise, the risk of CVDs is increased in chronic inflammatory disorders, with evidence that risk is associated with severity of inflammation.[@bib0090]
The SCORE risk calculator is recommended for CVD risk prediction in the general population by the ESC/EAS guidelines. However, CVD risk prediction models developed for the general population do not include non-traditional CVD risk factors. If these models are applied in patients with psoriasis, there is a possibility of underestimating the CVD risk.
In addition to appropriate cardiovascular risk management, some authors believe that the risk score models should be adapted for patients with psoriasis by introducing a multiplication factor that takes into account the presence of psoriasis.[@bib0045]
Other authors establish that psoriasis is a "risk enhancer" that can be used to favor initiation or intensification of statin therapy, particularly in stratified patients with intermediate risk. This strategy, based on the ASCVD calculator, is recommended by the new 2018 ACC/AHA guidelines for cholesterol management and does not contemplate any multiplication or adjustment factor.[@bib0050]
In the present study, when it was applied the 2018 ACC/AHA guidelines, the majority of patients were classified as intermediate or high risk. However, when it was applied the 2016 ECS/EAS guidelines, the majority of patients with psoriasis were stratified as intermediate risk. These finding coincides with a cross-sectional study of 234 patients with psoriasis that showed that the cardiovascular risk estimated by the Framingham score was on average 11.2% (intermediate risk).[@bib0095] Similarly, a study that analyzed 395 patients with psoriasis showed that the proportion of patients at intermediate and high risk of suffering a major cardiovascular event in the next ten years was 30.5% and 11.4%, respectively, based on the Framingham risk score.[@bib0035] Another study conducted in Brazil involving the assessment of 190 patients with psoriasis showed that 47% had moderate or high risk of fatal and non-fatal coronary events in ten years.[@bib0105]
The main finding of the study was that by using two different strategies of cardiovascular prevention for the management of patients with psoriasis, the proportion of eligible patients for statin therapy was similar (close to 60%). However, the concordance between both strategies in selecting patients with statin indication was moderate, indicating that individually, some subjects had different indications according to the guideline used.
These findings differ from another study that evaluated patients with rheumatoid arthritis. Tournadre et al. calculated the proportion of patients eligible for statins according to ESC guidelines, the Adult Treatment Panel III, and the ACC/AHA in a French cohort of statin-naïve rheumatoid arthritis patients at least 40 years of age. A marked discordance in risk assessment and cholesterol treatment was observed between the three sets of guidelines.[@bib0110] The difference with the present work could be explained by the populations studied, plus the fact that a multiplication factor was used in this study.
Several publications showed that men more frequently receive an indication of statin therapy than women when analyzing the general population.[@bib0115], [@bib0120] Regarding this topic, but analyzing only patients with psoriasis, the present study showed that the proportion of subjects with an indication for statin therapy was higher in men than women, regardless of the chosen strategy.
To determine whether a patient is a candidate for statin therapy, clinicians must first determine the patient\'s risk of having a future CVD event. However, clinicians' ability to accurately identify a patient\'s true risk is imperfect, because the currently available risk estimation tools have been shown to underestimate the risk in patients with chronic inflammatory diseases.[@bib0030], [@bib0125], [@bib0130] Consequently, the current methods of cardiovascular risk assessment in the course of chronic inflammatory diseases are a subject of considerable controversy.
In the present study, the main reason for the indication of statin treatment was the intermediate cardiovascular risk calculated with the scoring method. In addition, this finding was observed when using both strategies and in both sexes. These findings reinforce the importance of using tools for the stratification of cardiovascular risk, despite its limitations.
According to the results of this study, the authors believe that not all patients with psoriasis should receive statins. European and North American strategies, analyzed in this study, agree that patients in primary prevention with diabetes and/or a level of LDL-C \>190 mg/dL should receive statins, whether they have psoriasis or not. The other group of patients, who are candidates for statins, should be defined according to the estimated cardiovascular risk. Europeans guidelines use SCORE and adjust the value by a correction factor. Americans guidelines use the ASCVD calculator and do not use an adjustment factor. Both strategies will recommend giving statins to patients at intermediate or high risk.
In synthesis, every psoriatic patient with a very high cholesterol value, diabetes, or an intermediate/high risk should receive statins. Taking into account that patients with psoriasis have a higher cardiovascular risk independently of the presence of conventional risk factors, it would be advisable to apply a correction factor to the cardiovascular risk score calculated as the Europeans recommendations do.
This study had some limitations. It was a secondary database study (electronic medical records); consequently, there could be information bias. Furthermore, the data on the severity of psoriasis and pharmacological treatments could not be reliably obtained retrospectively; therefore, this data could not be included in the analysis. Despite its limitations, this study represents a valuable contribution, as a large group of patients with psoriasis but without CVD was examined. Research for risk factors and proper risk stratification are rare in patients with psoriasis. Knowledge of the application of different strategies in cardiovascular prevention could favor the difficult task of estimating cardiovascular risk in this particular group of patients.
Conclusion {#sec0025}
==========
This study\'s findings showed that not all patients with psoriasis should receive statins. The population with psoriasis without CVD was mostly classified at moderate--high risk, and the statin therapy indication was similar when applying the two strategies evaluated.
Financial support {#sec0030}
=================
None declared.
Author\'s contribution {#sec0035}
======================
Walter Masson: Statistical analysis; approval of the final version of the manuscript; conception and planning of the study; elaboration and writing of the manuscript; obtaining, analyzing and interpreting the data; critical review of the literature; critical review of the manuscript.
Martín Lobo: Statistical analysis; approval of the final version of the manuscript; obtaining, analyzing and interpreting the data.
Graciela Molinero: Approval of the final version of the manuscript; conception and planning of the study; elaboration and writing of the manuscript; critical review of the manuscript.
Emiliano Rossi: Statistical analysis; approval of the final version of the manuscript; elaboration and writing of the manuscript; obtaining, analyzing and interpreting the data; critical review of the literature.
Conflicts of interest {#sec0040}
=====================
None declared.
How to cite this article: Massom W, Lobo M, Molinero G, Rossi E. Should all patients with psoriasis receive statins? Analysis according to different strategies. An Bras Dermatol. 2019;94:691--7.
Study conducted at the Hospital Italiano de Buenos Aires, Buenos Aires, Argentina.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#sec1-1}
============
Rheumatoid arthritis (RA) is a chronic systemic autoimmune disease\[[@ref1]\]. RA can cause joint pain, deformation, and even disability, having a serious impact on the quality of life of patients. At present, it is still not clear what the etiology of RA is, but most scholars think it is caused by geographical environment and genetic factors\[[@ref2],[@ref3]\]. A study showed\[[@ref4]\] that approximately 20 million people would have been influenced by RA by 2015 worldwide, and the majority of RA patients are in 30\~50 years old, the morbidity rate of which in women is three times higher than that in men. The study of Abubakar II et al. showed\[[@ref5]\] that the death rate of RA patients increased by nearly 1.3 times from 28000 in 1990 to 38000 in 2013 in the past 20 years.
The continuous development of molecular biology provides new conditions for studying the mechanism of RA at the gene level. The main manifestation of RA is the abnormal proliferation of synovial tissue contributing to progressive tissue damage, and important biological changes take place in synovial cells in the process of proliferation\[[@ref6]\]. In addition to binding to elastase and collagenase, synovial cells are also involved in the recruitment of lymphocytes and macrophages, activating cytokines and adhesion molecules\[[@ref7]\]. Any block in these links has an impact on the occurrence and development of RA, and these links are potential therapeutic targets for RA. In the present study, it is shown\[[@ref8]\] that the pathological feature of RA was mainly the immune inflammation mediated by the activation of macrophages, plasma cells and T cells caused by synovial infiltration and the release of a large number of inflammatory cytokines after the activation of inflammatory cells. Transforming growth factor-β1 (TGF-β1), which is a multifunctional cytokine, is an important member of the transforming growth factor β superfamily regulating the cell response and involved in bone metabolism through joint soft tissue cells or synovial tissues\[[@ref9]\]. At present, it is found that there is a gene polymorphism in TGF-β1. Some of the locus gene polymorphisms were associated with the expression level of TGF-β1 and the gene polymorphism of TGF-β1 was associated with the existence of RA.
Therefore, in this study, relationships between the polymorphism of the gene first exon +869T/C and rheumatoid arthritis were examined.
Materials and methods {#sec1-2}
=====================
In this study, one hundred and fifty patients who were diagnosed with RA at the Department of Rheumatology between March 2014 and May 2017 were selected as subjects (inpatient group), including 116 female and 34 male, aged from 33 to 62 years with an average age of 48.58±10.55 years. All subjects in the inpatient group were in accordance with the diagnostic criteria of rheumatoid arthritis established by the American Society of Rheumatology\[[@ref10]\]. In addition, the volunteers for physical examination (control group) were selected at the same time in the physical examination center of our hospital to carry out the comparison, including 100 female and 50 male, aged from 30 to 60 years old with an average age of 47.94±11.05 years. This study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Chengdu Medical College. All the patients and their family members were informed and signed the consent.
Inclusion and exclusion criteria {#sec2-1}
--------------------------------
Inclusion criteria: all the patients were more than 18 years old; no major operation had been performed before examination; no blood transfusion treatment had been performed one month before examination; patients had no depression or memory impairment; patients had complete clinical data, cooperated with treatment and were followed up.
Exclusion criteria: patients with less than half a year course of disease; patients with other hereditary diseases; patients with congenital defects; patients with mental illness or disability; patients who have received radiotherapy and chemotherapy within half a year; patients with cancer and immune diseases; pregnant and lactating women.
DNA extraction and RFLP-PCR detection {#sec2-2}
-------------------------------------
The peripheral blood (10 mL) was extracted from the venous blood of the subjects and the collected samples were labeled with patients'information. 2 mL was extracted using EDTA-k anticoagulant tube and the other blood samples were packed with different blood vessels and sent to the laboratory of our hospital for the detection of ESR, rheumatoid factor, C reactive protein, acute phase reaction protein and blood routine tests.
The blood samples in EDTA-k anticoagulant tube were stored in -80°C refrigerator and Genomic DNA was extracted from EDTA whole blood using a spin column method according to the protocol (QIAamp Blood Kit; Qiangen, Germany). DNA purity was detected by UV spectrophotometer.
The design and synthesis of +869T/C primers were carried out by Shanghai Bioengineering Co., Ltd and the primer sequence used for +869T/C in upstream was 5'-CGGCACCTCCCCCTGGCTCG-3' and that in downstream was 5'\--CCTCCCCACCACACCAG-3'. Reaction system: 2.5 µL for 10XBuffer, 2.5 µL for 25 mmol/L Mg^2+^, 0.5 µL for upstream primer, 0.5 µL for downstream primer, 2.0 µL for 1 µg/µL DNA template, 1U/µL for Taq enzyme (ABI, United States) and double distilled water added to 25 µL. Reaction conditions: 10 min after pre-denaturation at 94°C, 30 s for denaturation at 94°C, 30 s for annealing at 60°C, 1 min for an extension at 72°C, 40 cycles, and 10 min for an extension at 72°C after the completion of the cycle. Gel Imaging Analysis was obtained from Beijing Maxwell High Technology Co., Ltd.
The PCR products were preserved at 4°C, restriction endonuclease was purchased from Baosheng Biological Co., Ltd and digested with 1U restriction endonuclease Bsu36 I for 2.5 h at 37°C. The genotypic identification of enzyme-digested products identification was performed by 2.5% agarose gel electrophoresis stained by EB, and three genotypes of enzyme-digested products were wild homozygous (CC)(190 bp, 229 bp), mutant heterozygote (CT)(419 bp, 229 bp, 190 bp) and mutant homozygous (TT)(419 bp).
Statistical analysis {#sec2-3}
--------------------
The data was examined by Hardy-Weinberg balance using R language genetics package and *P*≥0.05 suggested that the gene frequency was in accordance with Hardy-Weinberg's law. SPSS 22.0 (Beijing Situangweida Technology Co., Ltd.) was used for statistical analysis, and all data were expressed by Means ± SD. Comparing the data of the two groups, t-test was used to examine the inter-group measurement data, x\[[@ref2]\] was used to examine the inter-group counting data, and relationships between RA patients and genotypes were analyzed using logistic regression. *P*\<0.05 suggested that the difference was statistically significant.
Results {#sec1-3}
=======
The contrast of clinical data between the two groups of patients {#sec2-4}
----------------------------------------------------------------
Comparison of clinical data and clinical test results between the inpatient group and the control group indicated that there was no statistical difference in the two groups in age, gender, exercise status, residence, smoking history, white blood cell, red blood cell, platelet, hemoglobin and acute phase reaction protein (*P*\>0.05), while the rheumatoid factor, ESR and C reactive protein of patients significantly increased, and the difference was significant (*P*\<0.01). The details were shown in \[[Table I](#T1){ref-type="table"}\].
######
Contrast of Clinical Data and Clinical Detection Index between the two groups of patients \[n(%)\].
Inpatient Group (n=150) Control Group (n=150) x^2^/t *P* Value
-------------------------------------- ------------- ------------------------- ----------------------- -------- -----------
Gender 0.053
Male 34 (22.67) 50 (33.33)
Female 116 (77.33) 100 (66.67)
Age (Years) 0.482 0.487
≥45 83 (55.33) 77 (51.33)
\<45 67 (44.67) 73 (48.67)
Exercise Status 0.400
Yes 50 (33.33) 58 (38.67)
No 100 (66.67) 92 (61.33)
Smoking History 0.257
Yes 40 (26.67) 50 (33.33) 0.053
No 110 (73.33) 100 (66.67)
White Blood Cell (x10^9^/L) 6.84±2.58 6.65±1.92 0.725 0.470
Red Blood Cell (x10^9^/L) 4.68±0.35 4.60±0.58 1.446 0.149
Platelet (x10^9^/L) 294.50±75.35 285.64±62.31 1.110 0.268
Hemoglobin (g/L) 135.54±11.33 133.61±10.83 1.508 0.133
Rheumatoid Factor (IU/mL) 45.88±4.25 13.07±1.58 88.624 0.000
ESR (mm/h) 37.84±3.66 15.32±1.36 70.639 0.000
C Reactive Protein (mg/L) 14.84±9.36 2.54±0.33 16.084 0.001
Acute Phase Reaction Protein (IU/mL) 46.38±4.55 45.76±8.32 0.878 0.380
Analysis of genotype and gene frequency of +869T/C locus {#sec2-5}
--------------------------------------------------------
The two groups were examined by Hardy-Weinberg balance using R language genetics package, and it is found that all groups were in accordance with Hardy-Weinberg's law suggesting that the study group was representative. The details were shown in \[[Table II](#T2){ref-type="table"}\]. By comparing the genotype frequency of the two groups, we found that the composition ratio was significantly different (x^2^=9.972, *P*=0.007) in \[[Table III](#T3){ref-type="table"}\], and compared to control group, the genotype frequency of CC inpatient group significantly decreased, while that of CT and TT increased. The C allele frequency the inpatient group showed a downward trend, but the T allele frequency increased, and the difference was statistical (*P*\<0.05).
######
Hardy-weinberg balance test.
Gene Locus Inpatient Group Control Group
------------ ----------------- --------------- ------- -------
+869T/C 0.747 0.382 0.443 0.506
######
Distribution of gene and genotype frequency.
SNP Inpatient Group Control Group X^2^ *P*
--------- ------------- ----------------- --------------- ------- -----
+869T/C Genotype
CC 26 (17.3%) 49 (32.7%) 0.003
CT 80 (53.3%) 70 (46.7%) 1.333 0.248
TT 44 (29.3%) 31 (20.7%) 0.109
Allele 8.640 0.003
C 132 (44.00) 168 (56.00)
T 168 (56.00) 132 (44.00)
Relationships between RA patients and genotypes {#sec2-6}
-----------------------------------------------
Analysis of logistic regression showed that there was a statistical difference suggesting that patients with CC genotype were more likely to develop RA than those with non-CC genotype, and the risk of patients with CC genotype was about 2.8 times higher than that of those with non-CC genotype. The details were shown in \[[Table IV](#T4){ref-type="table"}\].
######
Logistic regression analysis of the relationships between rheumatoid arthritis and genotypes.
Genotype β Wald X^2^ *P* OR 95%CI
---------- ------- ----------- ------- ------- --------------
Sex 0.884 0.384 0.825 1.084 0.358\~1.786
CC 0.984 5.758 0.015 2.834 1.384\~6.547
CT 0.745 0.358 0.873 1.025 0.325\~1.635
TT 0.835 0.425 0.892 1.063 0.348\~1.684
Notes: β means regression coefficient and intercept; OR means hazard ratio; Wald X^2^ means Chi-square value; 95%CI means confidence interval.
Discussion {#sec1-4}
==========
Rheumatoid arthritis is a chronic systemic inflammatory disease of unclear etiology at present. The main clinical manifestations of RA are symmetrical multiple joint recurrent attacks and cartilage and bone injury caused by the progressive development of synovium and hyperplastic tissue, thus contributing to dysfunction of joint function inpatients\[[@ref11]\]. Studies have shown\[[@ref12]\] that RA patients without effective treatment have a disability rate of more than 60% within 2\~3 years. However, at present, there is no satisfactory specific medicine for the treatment of RA, and normal routine treatment can only alleviate the inflammation of the patients and relieve the pain, so as to improve the mobility and quality of life of patients, but it cannot cure the disease fundamentally. In recent years, the development of many types of drugs has alleviated the course of RA, but it is only a kind of exploratory control therapy and still unable to cure the disease fundamentally\[[@ref13]-[@ref15]\]. Therefore, it is imperative to seek new treatment methods and corresponding nursing methods to delay the conditions of patients and improve their quality of life.
TGF-β1, which is a polypeptide isolated from the cultivation process of mouse T cells transformed by Moloney tumor virus, can induce non-tumor cells to lose their inhibitory proliferation characteristics and further obtain growth function\[[@ref16]\]. Studies have shown that TGF-β is expressed by three subtypes in mammalian tissues. TGF-β1 is expressed in connective tissues, and TGF-β2 and TGF-β3 are expressed in epithelial neuron cells and in interstitial cells respectively\[[@ref17]\]. A study has shown\[[@ref18]\] that a high concentration of TGF-β1 has a function in the promotion of multiple growth factors synthesis and in the protection of newly formed fibroblasts. However, it can also destroy macrophages to produce H~2~O~2~ destructive to the newly formed fibroblasts. Study of Cheon H reported that the high expression of TGF-β1 in fibroblast-like synovial cells could promote the expression of various inflammatory cytokines (IL-1β, TNF-α, IL-8) in RA patients, thus promoting the occurrence of inflammatory reaction in RA patients. It also showed that the high expression of TGF-β1 in fibroblast-like synovial cells of RA patients was specific indicating that TGF-β1 was involved in the occurrence and development of RA patients' conditions\[[@ref19]\].
At present, the gene polymorphism of TGF-β1 has been studied in various diseases. Diseases, related to the polymorphism of TGF-β1 such as myocardial infarction, cancer, hypertension, diabetes, organ fibrosis, and autoimmune diseases, have been reported, and different expression levels of TGF-β1 in blood plasma have been found among individuals\[[@ref20]\]. In this study, by detecting the gene polymorphism of RA patients and normal patients, we found that the genotype frequency of CC in +869T/C locus of RA patients in inpatient group significantly decreased when compared to control group, while that of CT genotype and TT genotype increased. The C allele frequency the inpatient group showed a downward trend, but the T allele frequency increased, and there was a difference. We speculated that patients with CC genotype were more likely to develop RA than those with non-CC genotype, and we confirmed our point of view using the following logistic regression analysis. We also cannot explain the reason why these patients lack or carry CC genotype, but we speculate that this might be associated with the living environment of patients.
However, there are a few limitations in this study. First of all, the data of only two groups of patients were contrasted and analyzed, three genotypes of patients were not compared, and it was still not clear whether there was any difference among them. Second, the small number of subject may have an impact on our research and we cannot be certain whether the accuracy of the outcomes was biased. Therefore, we hope that the sample size and content of analysis will be increased to better verify the accuracy of our results in future studies.
In conclusion, the polymorphism of the gene first exon +869T/C in TGF-β1 significantly correlates with rheumatoid arthritis (RA) and CC genotype may be a susceptible gene of RA.
Funding {#sec2-7}
-------
*This work was supported by the Natural Science Foundation of Chengdu Medical College (CYZ15-10), Natural Science Foundation of China (Grant No. 31701104), Scientific Research Fund of the Sichuan Provincial Education Department (16ZA0289) and National Undergraduates Innovating Experimentation Project (201513705007)*.
Authors' contributions {#sec2-8}
----------------------
*WS and MY collected and analyzed the clinical data of patients. YB, LW, JC, and YR analyzed Genotype and Gene Frequency of +869T/C Locus. XL, HW, and YM were responsible for DNA Extraction. QZ performed PCR. All authors read and approved the final manuscript*.
Ethics approval and consent to participate {#sec2-9}
------------------------------------------
*The study was approved by the Ethics Committee of the First Affiliated Hospital of Chengdu Medical College. Patients who participated in this research signed the informed consent and had complete clinical data. Signed written informed consents were obtained from the patients and/or guardians*.
The authors have no conflict of interest.
Edited by: G. Lyritis
| {
"pile_set_name": "PubMed Central"
} |
1.. Introduction {#sec1}
================
*Streptococcus pneumoniae* is a major human pathogen that is responsible for respiratory-tract infections, septicaemia, otitis media and meningitis. Current broad-spectrum antibiotic treatments for *S. pneumoniae* are increasingly unsuccessful owing to the emergence of drug-resistant strains (Thornsberry *et al.*, 1999[@bb20]). There are multivalent capsular polysaccharide vaccines available for pneumococcal disease, but their efficacy in certain high-risk groups has been questioned (Siber, 1994[@bb18]). More recently, a conjugate polysaccharide vaccine has been introduced successfully (Prevenar, Wyeth), but there are concerns about how readily it can be introduced globally and about its continuing efficacy. A search is therefore on for new drug and vaccine candidates for pneumococcal diseases. Several virulence factors may contribute to colonization and early infection processes (Jedrzejas, 2001[@bb8]). Sialidases are one key virulence factor as they remove sialic acid from host cell-surface glycans, probably unmasking certain receptors to facilitate bacterial adherence and colonization (Paton *et al.*, 1993[@bb16]). To date, all *S. pneumoniae* clinical isolates investigated have had prominent sialidase activities. Up to three distinct sialidases, NanA (Camara *et al.*, 1994[@bb4]), NanB (Berry *et al.*, 1996[@bb3]) and NanC, are encoded in *S. pneumoniae* genomes, with a recent study revealing NanA to be present in all clinical strains (Pettigrew *et al.*, 2006[@bb17]). Gene-knockout studies in mouse models have shown that NanA and NanB are essential for *S. pneumoniae* infection (Manco *et al.*, 2006[@bb12]).
In this paper, we report the 2.5 Å resolution X-ray crystallographic structure of the catalytic domain of *S. pneumoniae* NanA and its complex with the inhibitor 2-deoxy-2,3-dehydro-*N*-acetyl neuraminic acid (Neu5Ac2en). This provides a framework for the structure-based design of specific inhibitors of pneumococcal sialidases as potential therapeutic agents.
2.. Experimental {#sec2}
================
Full-length NanA could be recombinantly expressed and purified, but failed to crystallize. Limited proteoloysis using trypsin, followed by mass spectrometry of the cleavage products, identified a stable subdomain, which we designate CNanA, that spans residues 319--822 and encompasses a domain that retains the enzyme activity of the full-length NanA. The *S. pneumoniae nanA* gene in a pQE30 vector was used as a template in polymerase chain reaction (PCR) with the following primers: 5′-ACCT**CCATGG**AAGGAGCGGCTTTAACAGAGA-3′ and 5′-GGGC**CTCGAG**TTAGACCAATACTTCTGAGTCG-3′ (*Nco*I and *Xho*I restriction sites in bold). The PCR product was then ligated into the pHISTEV vector, containing six histidines and a tobacco etch virus (TEV) cleavage peptide at the N-terminus, and plasmid DNA was extracted using a Mini-Prep Kit (Promega). The plasmid was transformed into *Escherichia coli* BL21 (DE3) expression strain (Novagen) for protein expression. The transformed *E. coli* was inoculated into Luria--Bertani (LB) medium with 100 µg ml^−1^ kanamycin at 310 K. 0.5 m*M* isopropyl β-[d]{.smallcaps}-thiogalactopyranoside (IPTG) was added to induce CNanA expression when the optical density at 600 nm (OD~600~) of the cultures reached 0.6. Cell culture continued at 310 K for 3 h before harvesting by centrifugation at 4500*g* for 30 min at 277 K. The harvested cell pellets were resuspended in 0.1 *M* phosphate pH 7.4, 10 m*M* imidazole and sonicated with 5 × 30 s bursts. Protease-inhibitor cocktail tablets (one tablet per 25 ml extract; Roche Diagnostics) and DNAase (Sigma; final concentration 20 µg ml^−1^) were then added. The crude cell extract was centrifuged at 43 000*g* for 20 min at 277 K to remove the cell debris and the supernatant was filtered with a syringe-driven filter (0.45 µm) before starting protein purification. Soluble cell extract was loaded onto a 5 ml nickel column (GE Healthcare) and the bound protein was eluted with 300 m*M* imidazole in 0.1 *M* phosphate buffer pH 7.4. Protein purity was assessed by sodium dodecyl sulfate--polyacrylamide gel electrophoresis (SDS--PAGE) and matrix-assisted laser desorption time-of-flight mass spectrometry (MALDI--TOF). Relatively high-purity target protein was pooled for gel filtration using a 120 ml Sephacryl-200 column (GE Healthcare). The purified CNanA was dialysed against 0.1 *M* Tris--HCl pH 8.0, 150 m*M* NaCl overnight before concentration and storage.
Purified protein was concentrated to 10.9 mg ml^−1^ for crystallization experiments using the sitting-drop vapour-diffusion method at 290 K with the commercial kits Classics (Jena Bioscience), JCSG, Nextal PEGs and Nextal pH Clear (Qiagen). Crystalline materials were observed after 3 d from condition No. 32 of Nextal PEGs \[0.1 *M* MES pH 6.5, 25%(*w*/*v*) PEG 4000\]. This condition was selected for crystallization optimization with crystallization drops made up of equal amounts of protein solution and reservoir solution (2 µl each). The best condition for CNanA crystallization was 0.1 *M* MES pH 6.5, 30%(*w*/*v*) PEG 4000. Crystals appeared in about 2 d and reached their maximum size after one week. The crystals have the remarkable habit of a square hollow tube (Fig. 1[▶](#fig1){ref-type="fig"}). Crystals of the Neu5Ac2en complex structure were grown in the presence of 10 m*M* Neu5Ac2en.
Crystals were cryoprotected by transfer for a few minutes into a solution of the crystallization buffer with 20%(*v*/*v*) glycerol before data collection at 100 K. Data were collected in-house (Rigaku-MSC MicroMax-007 HF X-ray generator and Saturn 944+ CCD detector). *HKL*-2000 (Otwinowski & Minor, 1997[@bb15]) was used for data processing and scaling and data-collection statistics are shown in Table 1[▶](#table1){ref-type="table"}. The catalytic domain of *Clostridium perfringens* sialidase NanI (Newstead *et al.*, 2008[@bb14]; PDB code [2bf6](http://scripts.iucr.org/cgi-bin/cr.cgi?rm=pdb&pdbId=2bf6)) was used to solve the CNanA structure by molecular replacement using *Phaser* (McCoy *et al.*, 2007[@bb11]) in the *CCP*4 suite (Collaborative Computational Project, Number 4, 1994[@bb5]). The asymmetric unit contains two CNanA molecules. The automated model-building wizard of the *Phenix* package (Adams *et al.*, 2002[@bb1]) was used to build the initial structure using the 2.5 Å resolution Neu5Ac2en complex data. This procedure built 85% of the residues with an *R* and *R* ~free~ of 0.29 and 0.35, respectively. *Coot* (Emsley & Cowtan, 2004[@bb7]) and *REFMAC*5 (Murshudov *et al.*, 1997[@bb13]) were used to refine and build the final model, which was validated with *MolProbity* (Lovell *et al.*, 2003[@bb9]). Refinement statistics are summarized in Table 1[▶](#table1){ref-type="table"}. The first 20 amino acids, including the six-histidine tag and TEV cleavage peptide, and the last 20 amino acids are not visible in the electron-density maps; the final model consists of residues 322--791. Molecule *A* is generally well ordered, whereas molecule *B* shows disorder in its N- and C-terminal regions. Both monomers have Neu5Ac2en bound.
3.. Results and discussion {#sec3}
==========================
The structure of CNanA shows the canonical six-bladed β-propeller fold common to all sialidases (Fig. 2[▶](#fig2){ref-type="fig"}). In common with the catalytic domains of *C. perfringens* NanI (Newstead *et al.*, 2008[@bb14]) and leech sialidase (PDB code [1sli](http://scripts.iucr.org/cgi-bin/cr.cgi?rm=pdb&pdbId=1sli); Luo *et al.*, 1998[@bb10]), CNanA has a small domain, residues 436--535, inserted between the second and third strands of the second sheet. CNanA and NanI share 41% sequence identity and the r.m.s.d. between 413 C^α^ atoms is 1.26 Å. The two monomers in the asymmetric unit are related by a noncrystallographic twofold axis, superimpose with an r.m.s.d. of 0.39 Å over 470 C^α^ atoms and bury ∼1200 Å^2^ of surface at their dimer interface. The calibrated gel filtration suggests that CNanA is predominantly a monomer in solution; however, it is possible that full-length NanA may be a dimer on the bacterial surface. The active site contains the usual catalytic residues common to all sialidases (Taylor, 1996[@bb19]): three arginines (Arg347, Arg663, Arg721) that interact with the carboxylate group of sialic acid, a nucleophilic tyrosine (Tyr752) that is proposed to form a covalent intermediate (Watts *et al.*, 2006[@bb21]; Newstead *et al.*, 2008[@bb14]) and its associated glutamic acid (Glu647), an aspartic acid (Asp372) that acts as an acid/base and a hydrophobic pocket that accommodates the acetamidomethyl group of sialic acid. In common with other bacterial sialidases, the hydroxyl group at C4 on Neu5Ac2en interacts with an arginine (Arg366) and an aspartic acid (Asp417). The O8 and O9 hydroxyls of the ligand's glycerol group interact with Tyr590 and Gln602, respectively. The topology of the surface of CNanA around the location of the aglycon is flat and open, in line with the promiscuity shown by NanA towards α(2,3), α(2,6) and α(2,8) linkages (unpublished data) and in contrast to leech sialidase and *Trypanosoma cruzi trans*-sialidase, which show specificity for α(2,3)-linked sialic acids (Amaya *et al.*, 2003[@bb2]). Two inhibitors of the influenza virus neuraminidase, zanamivir and oseltamivir, are currently licensed as treatments for influenza and both were developed based on the framework of Neu5Ac2en. The study reported here lays the groundwork for the potential elaboration of the core of Neu5Ac2en, particularly around the acetamido and glycerol moieties, in order to develop specific inhibitors of the pneumococcal sialidases.
Supplementary Material
======================
PDB reference: [NanA sialidase, 2vvz, r2vvzsf](2vvz)
GX was supported by Biocryst Pharmaceuticals Inc, Birmingham, Alabama. Resources of the St Andrews-based Scottish Structural Proteomics Facility, funded by the Scottish Funding Council, the Biotechnology and Biological Sciences Research Council (BBSRC) and the University of St Andrews, were used in this project.
{#fig1}
![Crystal structure of CNanA. Orthogonal views of the CNanA dimer are shown in (*a*) and (*b*), where (*b*) is related to (*a*) by a 90° rotation about a horizontal axis. The N- and C-termini are indicated by blue and red spheres, respectively. Molecule *A* (green) and molecule *B* (cyan) are drawn with the inserted domains (residues 436--535) drawn in lighter shades. The inhibitor Neu5Ac2en is shown in each monomer and is drawn in space-filling mode. (*c*) Stereoview of the active site of monomer *A* showing the hydrogen-bond interactions made between Neu5Ac2en and CNanA, with only key amino acids drawn for clarity. The 2*F* ~o~ − *F* ~c~ electron-density map contoured at 1σ is only drawn around the inhibitor for clarity. These figures were drawn using *PyMOL* (DeLano, 2007[@bb6]).](f-64-00772-fig2){#fig2}
###### Crystallographic summary
Values in parentheses are for the highest resolution shell.
--------------------------------------------- -------------------------------------
Space group *P*2~1~2~1~2~1~
Unit-cell parameters (Å) *a* = 49.2, *b* = 95.6, *c* = 226.6
Maximum resolution (Å) 2.5 (2.54--2.50)
Unique reflections 36773
Completeness 95.2 (71.2)
*I*/σ(*I*) 27.3 (4.6)
Mosaicity (°) 0.64
Redundancy 2.9 (2.5)
*R*~merge~[†](#tfn1){ref-type="table-fn"} 0.063 (0.248)
*V*~M~ (Å^3^ Da^−1^) 2.36
Refinement
Protein atoms 7442
Other atoms 40 (Neu5Ac2en), 124 waters, 1 Cl^−^
Resolution range (Å) 30--2.5
*R*~cryst~[‡](#tfn2){ref-type="table-fn"} 0.246
*R*~free~[‡](#tfn2){ref-type="table-fn"} 0.298
Mean temperature factor (Å^2^)
Protein, monomer *A*/*B* 29/45
Neu5Ac2en, monomer *A*/*B* 25/45
Waters 25
R.m.s.d. bond lengths (Å) 0.007
R.m.s.d. bond angles (°) 1.145
Ramachandran favoured/outliers (%) 90.5/3.7
--------------------------------------------- -------------------------------------
*R* ~merge~ = .
*R* ~cryst~ and *R* ~free~ = − ; *R* ~free~ was calculated for a 5% set of reflections excluded from the refinement.
| {
"pile_set_name": "PubMed Central"
} |
###### Key messages
What is already known about this subject?
=========================================
- Previous studies suggest that the early stages of rheumatoid arthritis (RA) may offer a therapeutic window of opportunity during which it may be possible to prevent joint damage.
- In methotrexate (MTX)-naive patients with RA, tofacitinib monotherapy is associated with inhibition of structural damage, reductions in clinical signs and symptoms of RA and improvements in physical functioning, over a period of 24 months.
What does this study add?
=========================
- Tofacitinib 5 mg two times a day appeared more effective in achieving clinical outcomes in early (disease duration \<1 year) versus established RA (disease duration ≥1 year).
- We observed that compared with MTX, tofacitinib monotherapy (5 or 10 mg two times a day) significantly improved signs and symptoms and physical function, and inhibited the progression of structural damage in MTX-naive patients regardless of disease duration.
How might this impact on clinical practice?
===========================================
- The diagnosis of RA should be made, and treatment started, as early as possible after the development of symptoms.
Introduction {#s1}
============
Rheumatoid arthritis (RA) is a chronic autoimmune disease characterised by synovial inflammation and joint damage that may lead to significant disability.[@R1] The goal of treatment is disease remission or, if remission is unattainable, the lowest disease activity possible, to improve the signs and symptoms, reverse functional disability and halt radiographic progression.[@R2]
It has been reported that disease duration has an impact on the effectiveness of treatment, and that comprehensive early treatment may offer an opportunity to preserve physical function and prevent disability.[@R3] It has been postulated that the early stages of RA may, therefore, offer a therapeutic window of opportunity in which to prevent joint damage from occurring;[@R7] [@R8] it has been suggested that this window may exist from 3 months to 2 years following the onset of symptoms.[@R9] [@R10] Much of the evidence for this has been suggested by trials of conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) and tumour necrosis factor inhibitors (TNFi), which have demonstrated that their use in patients with early RA is associated with better outcomes than in patients with established RA.[@R11]
Tofacitinib is an oral Janus kinase inhibitor for the treatment of RA. The efficacy and safety of tofacitinib 5 and 10 mg two times a day administered as monotherapy or in combination with csDMARDs, mainly methotrexate (MTX), in patients with active RA, has been demonstrated in Phase 2[@R15] and Phase 3[@R20] trials of up to 24 months\' duration, and in long-term extension studies with up to 84 months of observation.[@R26] [@R27]
The ORAL Start study (NCT01039688)[@R23] was a 24-month, Phase 3, randomised, double-blind, parallel-group, multicentre study, which compared tofacitinib monotherapy (5 or 10 mg two times a day) with MTX monotherapy in MTX-naive adult patients with RA.[@R23] The results of the primary analysis are reported elsewhere.[@R23] In brief, tofacitinib monotherapy was associated with statistically significant reductions in radiographic progression (measured using van der Heijde modified total Sharp score (vdHmTSS)), reductions in clinical signs and symptoms of RA, and improvements in physical functioning versus MTX monotherapy at month 6, with similar results at months 12 and 24, thus demonstrating durability of response.[@R23] Safety observations were consistent with prior Phase 2 and Phase 3 clinical trials of tofacitinib in patients with RA, and included infections (eg, herpes zoster), gastrointestinal disorders, lymphoma and other malignancies; no new safety signals were identified.[@R23]
The objective of this post hoc analysis was to compare the efficacy and safety of tofacitinib monotherapy in MTX-naive patients with early or established RA in the Phase 3 ORAL Start study. For the purposes of this analysis, early RA was defined as disease duration \<1 year, and established RA as disease duration ≥1 year from disease diagnosis. These boundaries were chosen to be within the window suggested by the literature, and allowed for an even split of patient data between the two groups.
The ultimate goal of this analysis was to provide insight into the optimal therapeutic window of opportunity during which tofacitinib may be used, thereby improving clinical and radiographic outcomes for patients with RA.
Methods {#s2}
=======
Study design {#s2a}
------------
This was a 24-month, randomised, Phase 3, double-blind, parallel-group trial conducted in 151 centres worldwide. The full methodology used in this study, with inclusion and exclusion criteria, has been reported elsewhere.[@R23]
The study was conducted in accordance with applicable legal and regulatory requirements, as well as the general principles set forth in the International Ethical Guidelines for Biomedical Research Involving Human Subjects, International Conference on Harmonization Guidelines for Good Clinical Practice, and the Declaration of Helsinki. The study protocol and informed consent documentation were approved by the institutional review board or independent ethics committee at each investigational centre. All patients provided written, informed consent.
Patients {#s2b}
--------
Patients were aged ≥18 years, with a diagnosis of RA as defined by American College of Rheumatology (ACR) 1987 criteria,[@R28] and who had ≥3 distinct joint erosions on posteroanterior hand and wrist or foot radiographs (locally read), or a positive test for rheumatoid factor, or antibodies to cyclic citrullinated peptide. Eligible patients also had ≥6 painful or tender joints, ≥6 swollen joints and an erythrocyte sedimentation rate (ESR) \>28 mm/hour by Westergren method or high sensitivity C reactive protein (hsCRP) levels \>7 mg/L. Patients were excluded if they received \>3 weekly doses of MTX, or if ≤3 weekly doses had been received but MTX had been stopped due to an MTX-related adverse event (AE). Patients were randomised in a 2:2:1 ratio to one of three parallel treatment arms: tofacitinib 5 mg two times a day, tofacitinib 10 mg two times a day or MTX weekly. MTX was initiated at a dose of 10 mg/week and, if well tolerated, increased by 5 mg/week every 4 weeks up to 20 mg/week by week 8 for the duration of the study.
We performed a post hoc subgroup analysis using the final 24-month data in patients with early and established RA.
Efficacy of treatment {#s2c}
---------------------
Effects on signs and symptoms of RA were evaluated using ACR response rates (ACR20, ACR50 and ACR70) and Disease Activity Score in 28 joints (DAS28-4(ESR) \<2.6 (remission) and ≤3.2 (low disease activity)). Physical function was assessed using the Health Assessment Questionnaire-Disability Index (HAQ-DI), in which a score \<0.5 is considered a normal value, 0.5 to ≤1 mild-to-moderate disability, \>1 to ≤2 moderate-to-severe disability and \>2 to 3 severe/very severe disability.
Joint structural preservation was assessed using the vdHmTSS[@R29] (range 0--448), with higher scores indicating greater joint structural damage, and an evaluation of the proportion of patients with no radiographic progression, defined as change from baseline in vdHmTSS≤0.5.
In order to assess the sensitivity of findings on the selection of the time points for defining early RA, key efficacy parameters were also calculated for patients with disease duration \<6 and ≥6 months.
Safety {#s2d}
------
The incidence and severity of all AEs was recorded, as previously described.[@R23] The incidence of treatment-emergent AEs (TEAEs), discontinuations due to AEs, AEs of special interest (SIEs), serious AEs (SAEs) and deaths were compared in patients with early and established RA.
Statistical analyses {#s2e}
--------------------
Unless otherwise stated, efficacy analyses were based on the Full Analysis Set, which included all patients who received at least one dose of study drug and for whom data were available from at least one postbaseline assessment. Binary end points (eg, achievement of ACR20 response) were compared between tofacitinib (5 or 10 mg two times a day) and MTX in a post hoc analysis using the normal approximation to the binomial distribution. Missing values were computed using the non-responder imputation method. Continuous end points, such as change from baseline in HAQ-DI scores, were analysed using a linear mixed-effect model for repeated measures. Treatment, visit, treatment-by-visit interaction and baseline were included as fixed effects, and with patients as a random effect. Estimates of mean changes from baseline for each treatment, as well as mean differences versus MTX were obtained from the model as least squares (LS) means, with corresponding SEs. For vdHmTSS, missing values due to patient discontinuation were linearly extrapolated based on the baseline value and postbaseline value prior to discontinuation. The binary variable of progression/non-progression was obtained from linearly extrapolated imputation data. Observed case data were used in any descriptive analyses. ORs with 95% CI, comparing early versus established disease for each treatment group, were provided for each end point of interest. If the 95% CI did not contain the value 1.0, the difference between the two groups was considered statistically significant at the 5% significance level. No preservation of type I error or multiple-comparisons correction was applied to p values, as statistical significance defined as p\<0.05 was exploratory in nature.
Analyses of safety data were based on observed cases. Incidence rates (IRs; unique patients with events per 100 patient-years of observation) for SIEs were based on the number of patients with an event, and the total exposure time censored at time of event, death or discontinuation from the study, and compared between treatment groups. Exact Poisson 95% CIs adjusted for exposure time were calculated for IRs.
Results {#s3}
=======
Demographics {#s3a}
------------
Baseline patient demographics and disease characteristics are summarised in [table 1](#RMDOPEN2016000262TB1){ref-type="table"}. Approximately half the 956 patients treated (54% (515/956)) had a disease duration of \<1 year; the proportions of patients with RA of duration \<1 and ≥1 year were similar between treatment groups ([table 1](#RMDOPEN2016000262TB1){ref-type="table"}). The majority of patients in the study were women and Caucasian ([table 1](#RMDOPEN2016000262TB1){ref-type="table"}). Baseline disease activity (DAS28-4(ESR), tender and swollen joint counts) and functional disability (HAQ-DI) in each treatment group were similar in patients with RA duration of \<1 and ≥1 year, whereas baseline vdHmTSS and erosion scores, as expected, were higher in patients with longer duration of disease ([table 1](#RMDOPEN2016000262TB1){ref-type="table"}).
######
Baseline patient demographics and disease characteristics by disease duration
Duration of RA Tofacitinib 5 mg two times a day Tofacitinib 10 mg two times a day MTX
------------------------------- ---------------------------------- ----------------------------------- ------------ ------------ ----------- -----------
Demographics
Female, n (%) 147 (73.1) 139 (80.8) 162 (78.3) 165 (86.8) 87 (81.3) 58 (73.4)
Caucasian, n (%) 135 (67.2) 104 (60.5) 131 (63.3) 135 (71.1) 77 (72.0) 50 (63.3)
Mean age, years 50.2 50.5 48.1 50.7 47.9 50.1
Mean disease duration, years 0.32 6.01 0.31 6.71 0.35 5.98
Disease characteristics\*
TJC 26.0 25.2 25.2 25.0 24.7 26.4
SJC 15.9 16.7 16.0 15.1 17.4 16.0
vdHmTSS 8.6 31.6 5.0 31.4 7.8 27.8
Erosion score 4.5 14.4 2.8 15.5 4.4 13.9
HAQ-DI 1.47 1.62 1.49 1.51 1.49 1.57
DAS28-4(ESR) 6.53 6.72 6.52 6.57 6.53 6.70
hsCRP, mg/dL 22.7 22.7 21.0 19.5 24.9 27.3
RF positive, % 82.6 82.0 80.2 83.2 83.2 86.1
ACPA positive, % 85.1 84.9 82.6 79.5 79.4 96.2
Early RA was defined as disease duration \<1 year, and established RA was defined as disease duration ≥1 year.
\*Mean values, unless otherwise specified.
ACPA, anticyclic citrullinated peptide antibodies; DAS28-4(ESR), Disease Activity Score in 28 joints (erythrocyte sedimentation rate); HAQ-DI, Health Assessment Questionnaire-Disability Index; hsCRP, high sensitivity C reactive protein; MTX, methotrexate; n, number of patients; RA, rheumatoid arthritis; RF, rheumatoid factor; SJC, swollen joint count; TJC, tender joint count; vdHmTSS, van der Heijde modification of the total Sharp score.
Clinical responses {#s3b}
------------------
ACR response rates, DAS28-4(ESR) \<2.6 and DAS28-4(ESR) ≤3.2 responses at month 24 were generally numerically higher in tofacitinib-treated patients, both 5 and 10 mg two times a day, with early RA versus those with established disease ([figure 1](#RMDOPEN2016000262F1){ref-type="fig"}). ACR response rates, DAS28-4(ESR) \<2.6 and DAS28-4(ESR) ≤3.2 responses were generally statistically significantly higher in patients treated with tofacitinib 5 or 10 mg two times a day compared with MTX, irrespective of disease duration stratification ([figure 1](#RMDOPEN2016000262F1){ref-type="fig"}). Similar trends in clinical responses were observed at month 12, with higher response rates in tofacitinib-treated patients with early versus established disease, and higher response rates in patients treated with tofacitinib 5 or 10 mg two times a day compared with MTX, irrespective of disease duration stratification (see online [supplementary figure](#SM1){ref-type="supplementary-material"} S1). Mean DAS28-4(ESR) levels at months 12 and 24 are presented in online [supplementary table](#SM1){ref-type="supplementary-material"} S1.
10.1136/rmdopen-2016-000262.supp1
{#RMDOPEN2016000262F1}
The mean improvement from baseline in HAQ-DI at month 24 was numerically greater for tofacitinib-treated patients (5 or 10 mg two times a day) with early RA versus those with established disease ([figure 1](#RMDOPEN2016000262F1){ref-type="fig"}). The mean improvement from baseline in HAQ-DI at month 24 was significantly greater for tofacitinib 10 mg two times a day compared with MTX for both disease duration subgroups, and for tofacitinib 5 mg two times a day for patients with early RA ([figure 1](#RMDOPEN2016000262F1){ref-type="fig"}). Mean improvements from baseline in HAQ-DI were also numerically greater for patients treated with tofacitinib 5 or 10 mg two times a day in early RA compared with established disease, at month 12 (see online [supplementary figure](#SM1){ref-type="supplementary-material"} S1). Mean improvement from baseline in HAQ-DI was greater at month 12 for both tofacitinib doses compared with MTX, but achieved statistical significance only for patients with early RA (p\<0.001 for both tofacitinib 5 and 10 mg two times a day). Mean HAQ-DI scores at months 12 and 24 are presented in online [supplementary table](#SM1){ref-type="supplementary-material"} S1.
Radiographic assessments {#s3c}
------------------------
There was significantly greater inhibition of radiographic progression in patients with early RA who received tofacitinib 5 or 10 mg two times a day compared with MTX at month 24; for patients with established RA, there was numerically greater inhibition of radiographic progression in both tofacitinib groups versus MTX at month 24, but the difference was statistically significant in the 10 mg two times a day group only ([figure 2](#RMDOPEN2016000262F2){ref-type="fig"}A).Within treatment groups, changes from baseline in vdHmTSS scores at month 24 were similar between patients with early RA and established RA ([figure 2](#RMDOPEN2016000262F2){ref-type="fig"}A). The proportion of patients with no radiographic progression at month 24 was greater versus MTX in those with early RA who received tofacitinib 5 mg two times a day, and in those with early and established RA who received tofacitinib 10 mg two times a day ([figure 2](#RMDOPEN2016000262F2){ref-type="fig"}B). Within treatment groups, the proportion of patients with no radiographic progression at month 24 was similar between patients with early RA and patients with established RA ([figure 2](#RMDOPEN2016000262F2){ref-type="fig"}B). Findings at month 12 were similar (see online [supplementary figure](#SM1){ref-type="supplementary-material"}s S2A, B). Mean vdHmTSS scores at months 12 and 24 are presented in online [supplementary table](#SM1){ref-type="supplementary-material"} S1.
{#RMDOPEN2016000262F2}
Treatment differences with tofacitinib versus MTX {#s3d}
-------------------------------------------------
Treatment differences (tofacitinib vs MTX) with 95% CIs for all clinical and radiographic assessments at month 24 are shown in [figures 1](#RMDOPEN2016000262F1){ref-type="fig"}[](#RMDOPEN2016000262F2){ref-type="fig"}--[3](#RMDOPEN2016000262F3){ref-type="fig"}. Across all end points, treatment differences versus MTX were generally larger for patients with early RA who received tofacitinib (5 or 10 mg two times a day) than for corresponding patients with established RA, although CIs overlapped ([figure 3](#RMDOPEN2016000262F3){ref-type="fig"}).
{#RMDOPEN2016000262F3}
In patients with early RA, differences in clinical end points with tofacitinib (5 or 10 mg two times a day) were statistically significant versus MTX (all p\<0.01, except for DAS28-4(ESR) \<2.6 in patients who received tofacitinib 10 mg two times a day, p\<0.05). For patients with established RA, statistically significant treatment differences were more frequently observed in the tofacitinib 10 mg two times a day group (all comparisons statistically significant vs MTX at the p\<0.05 level), compared with the tofacitinib 5 mg two times a day group (four out of eight comparisons were statistically significant vs MTX at the p\<0.05 level). For ACR responses, DAS28-4(ESR) \<2.6 and DAS28-4(ESR) ≤3.2, larger treatment differences for tofacitinib 5 mg two times a day versus MTX were observed in patients with early RA versus established RA ([figure 3](#RMDOPEN2016000262F3){ref-type="fig"}). Corresponding treatment differences at month 12 are presented in online [supplementary figure](#SM1){ref-type="supplementary-material"} S3.
For patients with early RA, differences in LS mean change from baseline in vdHmTSS, and the proportion of patients without radiographic progression, observed with tofacitinib (5 or 10 mg two times a day) versus MTX at month 24 were significant at the p\<0.01 level ([figure 3](#RMDOPEN2016000262F3){ref-type="fig"}). In patients with established RA, the LS mean change from baseline in vdHmTSS with tofacitinib 10 mg two times a day was significant versus MTX at the p\<0.05 level, and the proportion of patients with no radiographic progression was significant versus MTX at the p\<0.05 level with tofacitinib 5 and 10 mg two times a day (p=0.0199 and p=0.0039, respectively). Treatment differences for radiographic assessments at month 12 are presented in online [supplementary figure](#SM1){ref-type="supplementary-material"} S3.
Odds ratios for clinical and radiographic responses in early versus established RA {#s3e}
----------------------------------------------------------------------------------
At month 24, patients with early RA had numerically greater odds of achieving ACR20, ACR50, ACR70, DAS28-4(ESR) \<2.6, DAS28-4(ESR) ≤3.2 and no radiographic progression compared with those who had established RA, regardless of the study treatment received, with the exception of DAS28-4(ESR) \<2.6 for tofacitinib 5 mg two times a day and ACR70 for MTX ([figure 4](#RMDOPEN2016000262F4){ref-type="fig"}). These differences were significant in patients who received tofacitinib 5 mg two times a day for ACR20, ACR50, ACR70, DAS28-4(ESR) \<2.6 and DAS28-4(ESR) ≤3.2 ([figure 4](#RMDOPEN2016000262F4){ref-type="fig"}A). The corresponding 95% CI of the OR for no progression in vdHmTSS with tofacitinib 5 mg two times a day included 1 ([figure 4](#RMDOPEN2016000262F4){ref-type="fig"}A). For patients receiving tofacitinib 10 mg two times a day or MTX, 95% CI for ORs included 1.0 for all clinical and radiographic parameters ([figure 4](#RMDOPEN2016000262F4){ref-type="fig"}B, C). At month 12, similar trends were observed in the tofacitinib 5 mg two times a day treatment group; however, the 95% CI did not include 1.0 for ACR70 and DAS28-4(ESR) \<2.6, indicating statistical significance. ORs for clinical and radiographic assessments at month 12 are shown in online [supplementary figure](#SM1){ref-type="supplementary-material"} S4.
{#RMDOPEN2016000262F4}
Sensitivity analysis {#s3f}
---------------------
At month 24, patients with disease duration of \<6 months who received tofacitinib (5 or 10 mg) had higher ACR response rates (with the exception of ACR20) and rates of DAS28-4(ESR) \<2.6 and DAS28-4(ESR) ≤3.2 than those who had a disease duration of ≥6 months. In addition, ACR response rates, DAS28-4(ESR) \<2.6 and DAS28-4(ESR) ≤3.2 responses were higher in patients treated with tofacitinib 5 or 10 mg two times a day compared with MTX, irrespective of disease duration stratification (data not shown).
Safety analysis {#s3g}
---------------
A comprehensive analysis of safety data from this study has been reported elsewhere.[@R23] [Table 2](#RMDOPEN2016000262TB2){ref-type="table"} presents TEAEs, discontinuations due to AEs, SIEs, SAEs and deaths by disease duration. No meaningful differences were observed between patients with early and established RA with respect to all the safety parameters assessed.
######
Safety data: number of patients with AEs
Duration of RA Tofacitinib 5 mg two times a day Tofacitinib 10 mg two times a day MTX
------------------------------------ ---------------------------------- ----------------------------------- ------------ ------------ ----------- -----------
TEAEs, n (%) 165 (82.1) 132 (76.7) 175 (84.5) 159 (83.7) 83 (77.6) 64 (81.0)
Discontinuations due to AEs, n (%) 22 (10.9) 18 (10.5) 21 (10.1) 20 (10.5) 14 (13.1) 11 (13.9)
Serious infections, n (%) 3 (1.5) 8 (4.7) 4 (1.9) 4 (2.1) 3 (2.8) 2 (2.5)
SAEs, n (%)\* 17 (8.5) 22 (12.8) 15 (7.2) 26 (13.7) 11 (10.3) 11 (13.9)
Deaths, n (%)\*
Within 30-day rule 0 (0.0) 1 (0.6) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Outside 30-day rule 2 (1.0) 1 (0.6) 1 (0.5) 0 (0.0) 0 (0.0) 0 (0.0)
Early RA was defined as disease duration \<1 year, and established RA was defined as disease duration ≥1 year.
\*Two patients had SAEs and one patient died before being randomised to treatment. Therefore, these events are not included in the above table.
AE, adverse event; MTX, methotrexate; n, number of patients; RA, rheumatoid arthritis; SAE, serious adverse event; TEAE, treatment-emergent adverse event.
Discussion {#s4}
==========
This post hoc analysis, comparing treatment response in MTX-naive patients receiving tofacitinib 5 or 10 mg two times a day monotherapy or MTX monotherapy, has shown that for signs and symptoms, responses at month 24 were statistically significantly greater in patients with early RA versus established RA for patients receiving tofacitinib 5 mg two times a day. In general, numerically similar clinical and functional responses were observed in patients with early RA compared with patients with established RA for those treated with tofacitinib 10 mg two times a day and MTX. Consistent with the results of the primary analysis of all randomised and treated patients,[@R23] we observed that MTX-naive patients with early or established RA showed significant reductions in the signs and symptoms of RA, improvements in physical functioning and inhibition of radiographic progression when treated with tofacitinib monotherapy (5 or 10 mg two times a day) compared with MTX monotherapy. Additionally, there were no meaningful differences in safety parameters seen between patients with early and established RA.
In subgroup analyses of clinical trials of TNFi therapy in patients with RA, better outcomes have generally been demonstrated in patients with early RA compared with those with established RA.[@R12] [@R30] In the current post hoc analysis, patients with early RA generally showed similar improvements in clinical and radiographic efficacy end points compared with patients with established RA at baseline; however, there were significant improvements in ACR20/50/70 responses, and DAS remission and low disease activity in patients with early RA versus established RA in the tofacitinib 5 mg two times a day group at month 24, consistent with the data generated in clinical trials of TNFi therapy. Patients who received tofacitinib monotherapy had a statistically greater likelihood of achieving clinical response compared with those who received MTX, whichever tofacitinib dose or disease duration was compared. Baseline mean vdHmTSS and erosion scores were observed to be higher in this study in patients with longer compared with shorter disease duration, as expected, with increased joint damage in established RA that has not been treated effectively.
It has been postulated that there is a window of opportunity during which it may be possible to prevent joint damage with early and aggressive treatment of RA.[@R7] [@R9] The American College of Rheumatology now defines early disease as symptoms of \<6 months;[@R31] however, it should be noted that definitions of early RA vary across clinical trials from \<3 months to \<3 years.[@R12] [@R14] Furthermore, duration of RA may be challenging to assess in clinical trials, and may not be consistent across trials; time may be determined from onset of symptoms, onset of swelling or fulfilment of other classification criteria.[@R10] Additionally, consideration must be given to the duration of this window of opportunity,[@R32] and when it may close.[@R33]
Given the varying disease durations used in defining early RA in other studies, results for key efficacy parameters were reviewed for patients with disease duration \<6 months compared with those with disease duration ≥6 months, to assess the sensitivity of the results to the selection of the time points for defining early RA. Although this analysis demonstrated that treatment response at month 24 in patients treated with tofacitinib 5 mg two times a day was higher in patients with shorter disease duration, no significant difference between early and established RA was seen in the tofacitinib 10 mg two times a day or MTX treatment groups. Thus, our post hoc analysis did not provide any definitive support for the concept of a window of opportunity. This study was not designed to address differing responses dependent on disease duration, and further investigation of the effect of tofacitinib in early versus established RA is warranted.
In considering these results in terms of clinical practice, it could be concluded, in line with the principle of treating to target, the best results with tofacitinib or MTX are obtained in patients with early RA. Therefore, the diagnosis of RA should be made, and treatment started, as early as possible after the development of symptoms. If after 3 months of aggressive MTX therapy (ie, at least 20--25 mg/week) an adequate response is not observed, or if MTX is not tolerated, then, if possible, the patient could be switched to tofacitinib monotherapy. If a patient has some response to MTX, but not a significant response, it would be reasonable to add tofacitinib to MTX. An ongoing Phase 4 study will evaluate an add-on versus switch strategy in patients with RA with an inadequate response or intolerance to MTX.[@R34]
One potential limitation of this post hoc analysis is that disease duration was defined by the reporting of time since diagnosis at baseline by the investigator, rather than any objective criteria. Diagnosis of RA will vary from clinic to clinic, depending on local practice; the fact that the mean duration of disease for patients in the established RA group was around 6 years, but that these patients were still MTX-naive, highlights the potential uncertainty around disease duration for some patients. A further possible limitation is that this analysis was not prespecified; as such, comparison of tofacitinib effects with those of MTX within disease duration groups was not planned. Also, randomisation into the clinical trial was therefore not stratified by disease duration; the distribution of patients between groups is therefore determined by the decision to use disease duration of \<1 and ≥1 year to categorise early versus established RA. However, the results of the sensitivity analysis suggest that the selection of 1 year as a cut-off to define early RA was not critical.
Conclusions {#s5}
===========
Within treatment groups, improvements in signs and symptoms, physical function and inhibition of progression of structural damage were generally similar between patients with early RA and patients with established RA, although statistically significantly greater improvements in signs and symptoms of disease were observed at month 24 with tofacitinib 5 mg two times a day monotherapy in patients with early RA versus those with established RA.
Tofacitinib 5 and 10 mg two times a day monotherapy significantly improved signs and symptoms and physical function, and inhibited the progression of structural damage, in MTX-naive patients with early or established RA, as compared with MTX monotherapy. No meaningful differences in safety end points were observed between early and established RA subpopulations.
Editorial assistance was provided by Claire Cridland of Complete Medical Communications, and statistical support was provided by Ermeg Akylbekova of Quintiles; and both were funded by Pfizer.
**Contributors:** RMF, BW, RD and KK planned the study, RMF, AFK, TWJH, RFvV, KK, RD and BW analysed the data, KK performed the analysis and RMF, AFK, TWJH, RFvV, KK, RD and BW contributed to the writing of the manuscript. All authors read and approved the final manuscript.
**Funding:** This study and post hoc analysis were supported by Pfizer.
**Competing interests:** RMF has received research grants and consultancy fees from Pfizer. TWJH reports that the department of rheumatology has mutual projects with a number of EFPIA members (see <http://www.BTCURE.com>), plus grants from the Dutch arthritis foundation, the Dutch NIH and the EU. This author also reports lecture fees/consultancy fees from Merck, UCB, Bristol-Myers Squibb, Biotest AG, Pfizer, Novartis, Roche, Sanofi-Aventis, Abbott, Crescendo Bioscience, Nycomed, Boehringer, Takeda and Eli Lilly, is a paid instructor for Roche and Pfizer, and is on the Speakers Bureau of Roche, Pfizer, Eli Lilly and Bristol-Myers Squibb. A Kavanaugh has received research support from Pfizer. BW, KK and RD are employees and stockholders of Pfizer. RFvV has received grants and/or research support from Pfizer, Abbvie, BMS, GSK, UCB and has acted as a consultant for Pfizer, Abbvie, BiogenIDEC, Biotest, BMS, Celgene, Crescendo, GSK, Jansen, Lilly, MSD, Pfizer, UCB and Vertex.
**Ethics approval:** The study was approved by the institutional review board or independent ethics committee at each investigational centre.
**Provenance and peer review:** Not commissioned; externally peer reviewed.
**Data sharing statement:** No additional data are available.
| {
"pile_set_name": "PubMed Central"
} |
**CDC Adoption of ACIP Recommendations**
Recommendations for the routine use of vaccines in children, adolescents, and adults are developed by the Advisory Committee on Immunization Practices (ACIP). ACIP is chartered as a federal advisory committee to provide expert external advice and guidance to the Director of CDC on use of vaccines and related agents for the control of vaccine-preventable diseases in the civilian population of the United States. Clinical recommendations for routine use of vaccines are harmonized to the greatest extent possible with recommendations made by others (e.g., the American Academy of Pediatrics, the American Academy of Family Physicians, the American College of Obstetricians and Gynecologists, and the American College of Physicians).
ACIP recommendations adopted by the CDC Director become agency guidelines on the date published in MMWR. The accompanying recommendations that summarize the ACIP findings and conclusions were drafted based on the recommendations and revised based on feedback from ACIP voting members. The CDC Director approved these recommendations prior to publication. Opinions of individual members of ACIP might differ to some extent from the recommendations in this report as these recommendations are the position of CDC based on the ACIP recommendations to the CDC Director. Additional information regarding ACIP is available at <https://www.cdc.gov/vaccines/acip>.
Introduction
============
Influenza viruses typically circulate widely in the United States annually, from the late fall through the early spring. Although most persons with influenza will recover without sequelae, influenza can cause serious illness and death, particularly among older adults, very young children, pregnant women, and those with certain chronic medical conditions ([@R1]--[@R6]).
Routine annual influenza vaccination for all persons aged ≥6 months who do not have contraindications has been recommended by CDC and CDC's Advisory Committee on Immunization Practices (ACIP) since 2010 ([@R7]). This report updates the 2016--17 ACIP recommendations regarding the use of seasonal influenza vaccines ([@R8]) and provides recommendations and guidance for vaccine providers regarding the use of influenza vaccines for the 2017--18 season. A variety of different formulations of influenza vaccine are available ([Table 1](#T1){ref-type="table"}). Contraindications and precautions to the use of influenza vaccines are summarized ([Table 2](#T2){ref-type="table"}). Abbreviations are used in this report to denote the various types of vaccines ([Box](#B1){ref-type="boxed-text"}).
###### Influenza vaccines --- United States, 2017--18 influenza season[\*](#FN1){ref-type="fn"}
Trade name Manufacturer Presentation Age indication Mercury (from thimerosal, *µ*g/0.5 mL) Latex Route
------------------------------------------------------------------------------------------------------------------------ --------------------------------------------------------------------- ---------------------------------------------------- --------------------- ---------------------------------------- ------------- --------
**Inactivated influenza vaccines, quadrivalent (IIV4s), standard-dose^†^**
Afluria Quadrivalent Seqirus 0.5 mL prefilled syringe ≥18 years NR No IM^§^
5.0 mL multidose vial ≥18 years (by needle/syringe)
18 through 64 years (by jet injector) 24.5 No IM
Fluarix Quadrivalent GlaxoSmithKline 0.5 mL prefilled syringe ≥3 years NR No IM
FluLaval Quadrivalent ID Biomedical Corp. of Quebec (distributed by GlaxoSmithKline) 0.5 mL prefilled syringe ≥6 months NR No IM
5.0 mL multidose vial ≥6 months \<25 No IM
Fluzone Quadrivalent Sanofi Pasteur 0.25 mL prefilled syringe 6 through 35 months NR No IM
0.5 mL prefilled syringe ≥3 years NR No IM
0.5 mL single-dose vial ≥3 years NR No IM
5.0 mL multidose vial ≥6 months 25 No IM
**Inactivated influenza vaccine, quadrivalent (ccIIV4), standard-dose,^†^ cell culture-based**
Flucelvax Quadrivalent Seqirus 0.5 mL prefilled syringe ≥4 years NR No IM
5.0 mL multidose vial ≥4 years 25 No IM
**Inactivated influenza vaccine, quadrivalent (IIV4), standard-dose, intradermal^¶^**
Fluzone Intradermal Quadrivalent Sanofi Pasteur 0.1 mL single-dose prefilled microinjection system 18 through 64 years NR No ID\*\*
**Inactivated Influenza Vaccines, trivalent (IIV3s), standard-dose^†^**
Afluria Seqirus 0.5 mL prefilled syringe ≥5 years NR No IM
5.0 mL multidose vial ≥5 years (by needle/syringe)
18 through 64 years (by jet injector) 24.5 No IM
Fluvirin Seqirus 0.5 mL prefilled syringe ≥4 years ≤1 Yes**^††^** IM
5.0 mL multidose vial ≥4 years 25 No IM
**Adjuvanted inactivated influenza vaccine, trivalent (aIIV3),^†^ standard-dose**
Fluad Seqirus 0.5 mL prefilled syringe ≥65 years NR Yes**^††^** IM
**Inactivated Influenza Vaccine, trivalent (IIV3), high-dose**^§§^
Fluzone High-Dose Sanofi Pasteur 0.5 mL prefilled syringe ≥65 years NR No IM
**Recombinant Influenza Vaccine, quadrivalent (RIV4)**^¶¶^
Flublok Quadrivalent Protein Sciences 0.5 mL prefilled syringe ≥18 years NR No IM
**Recombinant Influenza Vaccine, trivalent (RIV3)**^¶¶^
Flublok Protein Sciences 0.5 mL single-dose vial ≥18 years NR No IM
**Live Attenuated Influenza Vaccine, quadrivalent (LAIV4)\*\*\* (not recommended for use during the 2017--18 season)**
FluMist Quadrivalent MedImmune 0.2 mL single-dose prefilled intranasal sprayer 2 through 49 years NR No NAS
**Abbreviations**: ACIP = Advisory Committee on Immunization Practices; ID = intradermal; IM = intramuscular; NAS = intranasal; NR = not relevant (does not contain thimerosal).
\* Immunization providers should check Food and Drug Administration--approved prescribing information for 2017--18 influenza vaccines for the most complete and updated information, including (but not limited to) indications, contraindications, warnings, and precautions. Package inserts for U.S.-licensed vaccines are available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm093833.htm>. Availability of specific products and presentations might change and differ from what is described in this table and in the text of this report.
**^†^** Standard dose intramuscular IIVs contain 15 *µ*g of each vaccine HA antigen (45 *µ*g total for trivalents and 60 *µ*g total for quadrivalents) per 0.5 mL dose.
^§^ For adults and older children, the recommended site for intramuscular influenza vaccination is the deltoid muscle. The preferred site for infants and young children is the anterolateral aspect of the thigh. Specific guidance regarding site and needle length for intramuscular administration is available in the ACIP General Best Practice Guidelines for Immunization, available at <https://www.cdc.gov/vaccines/hcp/acip-recs/general-recs/index.html>.
^¶^ Quadrivalent inactivated influenza vaccine, intradermal: a 0.1-mL dose contains 9 *µ*g of each vaccine HA antigen (36 *μ*g total).
\*\* The preferred injection site is over the deltoid muscle. Fluzone Intradermal Quadrivalent is administered per manufacturer's instructions using the delivery system included with the vaccine.
**^††^** Syringe tip cap might contain natural rubber latex.
^§§^ High-dose IIV3 contains 60 μg of each vaccine antigen (180 *μ*g total) per 0.5 mL dose.
^¶¶^ RIV contains 45 μg of each vaccine HA antigen (135 *μ*g total for trivalent 180 *μ*g total for quadrivalent) per 0.5 mL dose.
\*\*\*ACIP recommends that FluMist Quadrivalent (LAIV4) not be used during the 2017--18 season.
###### Contraindications and precautions to the use of influenza vaccines --- United States, 2017--18 influenza season[\*](#FN1){ref-type="fn"}
Vaccine type Contraindications Precautions
--------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
IIV History of severe allergic reaction to any component of the vaccine^†^ or after previous dose of any influenza vaccine Moderate-to-severe acute illness with or without fever
History of Guillain-Barré syndrome within 6 weeks of receipt of influenza vaccine
RIV History of severe allergic reaction to any component of the vaccine Moderate-to-severe acute illness with or without fever
History of Guillain-Barré syndrome within 6 weeks of receipt of influenza vaccine
LAIV
For the 2017--18 season, ACIP recommends that LAIV not be used. Content is provided for information**.** History of severe allergic reaction to any component of the vaccine^†^ or after a previous dose of any influenza vaccine
Concomitant aspirin or salicylate-containing therapy in children and adolescents
Children aged 2 through 4 years who have received a diagnosis of asthma or whose parents or caregivers report that a health care provider has told them during the preceding 12 months that their child had wheezing or asthma or whose medical record indicates a wheezing episode has occurred during the preceding 12 months
Children and adults who are immunocompromised due to any cause (including immunosuppression caused by medications or by HIV infection)
Close contacts and caregivers of severely immunosuppressed persons who require a protected environment
Pregnancy
Receipt of influenza antiviral medication within the previous 48 hours Moderate-to-severe acute illness with or without fever
History of Guillain-Barré syndrome within 6 weeks of receipt of influenza vaccine
Asthma in persons aged ≥5 years
Other underlying medical conditions that might predispose to complications after wild-type influenza infection (e.g., chronic pulmonary, cardiovascular \[except isolated hypertension\], renal, hepatic, neurologic, hematologic, or metabolic disorders \[including diabetes mellitus\])
**Abbreviations:** ACIP = Advisory Committee on Immunization Practices; IIV = Inactivated Influenza Vaccine; LAIV = Live-Attenuated Influenza Vaccine; RIV = Recombinant Influenza Vaccine.
\* Immunization providers should check Food and Drug Administration--approved prescribing information for 2017--18 influenza vaccines for the most complete and updated information, including (but not limited to) indications, contraindications, and precautions. Package inserts for US-licensed vaccines are available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm093833.htm>.
^†^ History of severe allergic reaction (e.g., anaphylaxis) to egg is a labeled contraindication to the use of IIV and LAIV. However, ACIP recommends that any licensed, recommended, and appropriate IIV or RIV may be administered to persons with egg allergy of any severity (see Persons with a History of Egg Allergy).
###### Abbreviation conventions for influenza vaccines used in this report
- Inactivated influenza vaccines are abbreviated IIV. For the 2017--18 season, IIVs as a class will include:
- egg-based, unadjuvanted, and adjuvanted trivalent influenza vaccines (IIV3s); and
- egg-based or cell culture-based unadjuvanted quadrivalent influenza vaccines (IIV4s).
- RIV refers to recombinant hemagglutinin influenza vaccine, available in trivalent (RIV3) and quadrivalent (RIV4) formulations for the 2017--18 season.
- LAIV refers to live-attenuated influenza vaccine, available as a quadrivalent formulation (LAIV4) since the 2013--14 season.
- IIV, RIV, and LAIV denote vaccine categories; numeric suffix specifies the number of hemagglutinin (HA) antigens in the vaccine.
- When necessary to refer specifically to cell culture-based vaccine, the prefix "cc" is used (e.g., ccIIV4).
- When necessary to refer specifically to adjuvanted vaccine, the prefix "a" is used (e.g., aIIV3).
- When necessary to refer specifically to standard-dose or high-dose vaccines, the prefixes "SD-" or "HD-" are used (e.g., SD-IIV3 and HD-IIV3).
This report focuses on the recommendations for use of influenza vaccines for the prevention and control of influenza during the 2017--18 season in the United States. A summary of these recommendations and a Background Document containing additional information on influenza-associated illnesses and influenza vaccines are available at <https://www.cdc.gov/vaccines/hcp/acip-recs/vacc-specific/flu.html>.
Methods
=======
[ACIP]{.smallcaps} provides annual recommendations for the use of influenza vaccines for the prevention and control of influenza. The ACIP Influenza Work Group meets by teleconference once to twice per month throughout the year. Work Group membership includes several voting members of ACIP and representatives of ACIP Liaison Organizations.[^\*^](#FN1){ref-type="fn"} Discussions include topics such as influenza surveillance, vaccine effectiveness and safety, vaccine coverage, program feasibility, cost-effectiveness, and vaccine supply. Presentations are requested from invited experts, and published and unpublished data are discussed.
In general, the Background Document is updated to reflect recent additions to the literature related to the following: 1) recommendations that were made in previous seasons, 2) changes in the viral antigen composition of seasonal influenza vaccines, and 3) minor changes in guidance for the use of influenza vaccines (e.g., guidance for timing of vaccination and other programmatic issues, guidance for dosage in specific populations, guidance for selection of vaccines for specific populations that are already recommended for vaccination, and changes that reflect use consistent with Food and Drug Administration \[FDA\]--licensed indications and prescribing information). The summary included in the Background Document for such topics is not a systematic review, but is intended to provide a broad overview of current literature. In general, systematic review and evaluation of the evidence using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach is performed for new recommendations or substantial changes in the recommendations (e.g., expansion of the recommendation for influenza vaccination to new populations not previously recommended for vaccination or potential preferential recommendations for specific vaccines).
Updates and changes to the recommendations described in this report are of five types: 1) the vaccine virus composition for 2017--18 U.S. seasonal influenza vaccines; 2) recent regulatory actions, including new vaccine licensures and labeling changes for previously licensed vaccines; 3) updated recommendations for the use of influenza vaccines in pregnancy, including a recommendation that pregnant women may receive any licensed, recommended, age-appropriate influenza vaccine; 4) a recommendation that the trivalent inactivated influenza vaccine (IIV3) Afluria (Seqirus, Parkville, Victoria, Australia) may be used for persons aged ≥5 years, consistent with FDA-approved labeling; and 5) a recommendation (continued from the 2016--17 season) that LAIV4 not be used during the 2017--18 season. Systematic review and GRADE were not performed for these updates and changes. Information relevant to these changes includes the following:
- Recommendations for composition of Northern Hemisphere influenza vaccines are made by the World Health Organization (WHO), which organizes a consultation, generally in February of each year. Surveillance data are reviewed and candidate vaccine viruses are discussed. A summary of the WHO meeting for selection of the 2017--18 Northern Hemisphere vaccine viruses is available at <http://www.who.int/influenza/vaccines/virus/recommendations/201703_recommendation.pdf>. Subsequently, FDA, which has regulatory authority over vaccines in the United States, convenes a meeting of its Vaccines and Related Biological Products Advisory Committee (VRBPAC). This committee considers the recommendations of WHO, reviews and discusses similar data, and makes a final decision regarding vaccine virus composition for influenza vaccines licensed and marketed in the United States. A summary of the FDA VRBPAC meeting of March 9, 2017, at which the composition of the 2017--18 U.S. influenza vaccines was discussed, is available at <https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/BloodVaccinesandOtherBiologics/VaccinesandRelatedBiologicalProductsAdvisoryCommittee/UCM552054.pdf>.
- With regard to recommendations for newly licensed influenza vaccines and changes to the licensed indications for existing vaccines, ACIP relies on FDA for review of safety, immunogenicity, and effectiveness data pertaining to the licensure of influenza vaccines. Regulatory information pertinent to the two recently licensed products and one labelling change discussed in this report is available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm518291.htm> (for Afluria Quadrivalent; Seqirus, Parkville, Victoria, Australia), <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm524660.htm> (for Flublok Quadrivalent; Protein Sciences, Meriden, Connecticut), and <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm366061.htm> (for FluLaval Quadrivalent; ID Biomedical Corporation of Quebec, Quebec City, Quebec, Canada).
- For the recommendation that pregnant women may receive any licensed, recommended, and age-appropriate influenza vaccine, because there have been no studies evaluating use of RIV3 or RIV4 in pregnancy, ACIP reviewed available information presented by the manufacturer and summarized in public documents by FDA concerning pregnancies that occurred among women participating in studies of RIV and pregnancy registry data. RIV reports to the Vaccine Adverse Event Reporting System (VAERS) also were presented. Regulatory information pertinent to this discussion is available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm524660.htm>[.]{.ul}
- For the recommendation that Afluria (IIV3; Seqirus, Parkville, Victoria, Australia) may be used for persons aged ≥5 years, ACIP reviewed data from studies performed by the manufacturer concerning the cause of an increase in the rate of febrile seizures that occurred in association with the 2010 Southern Hemisphere formulation of this product, and resulting changes in the vaccine manufacturing process. This change makes the ACIP recommendation consistent with FDA-approved labelling for this product. Minutes of the ACIP presentation of these data are available at <https://www.cdc.gov/vaccines/acip/meetings/downloads/min-archive/min-2017-02.pdf>. Regulatory information is available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm094043.htm>.
- The recommendation that LAIV4 not be used due to concerns regarding its effectiveness against influenza A(H1N1)pdm09 viruses during the 2013--14 and 2015--16 U.S. seasons was initially made for the 2016--17 season ([@R9]). This recommendation continues to be made for the 2017--18 season. ACIP will continue to review data concerning LAIV4 as they become available.
Primary Changes and Updates in the Recommendations
==================================================
Routine annual influenza vaccination of all persons aged ≥6 months without contraindications continues to be recommended. No preferential recommendation is made for one influenza vaccine product over another for persons for whom more than one licensed, recommended product is available. Updated information and guidance in this report includes the following:
- 2017--18 U.S. trivalent influenza vaccines will contain an A/Michigan/45/2015 (H1N1)pdm09--like virus, an A/Hong Kong/4801/2014 (H3N2)--like virus and a B/Brisbane/60/2008--like virus (Victoria lineage). Quadrivalent vaccines will include an additional vaccine virus strain, a B/Phuket/3073/2013--like virus (Yamagata lineage). This represents a change in the influenza A(H1N1)pdm09 virus component from the previous season.
- Recent regulatory actions, including two new licensures and one labelling change are described:
- Afluria Quadrivalent (IIV4; Seqirus, Parkville, Victoria, Australia) was licensed by FDA in August, 2016 for persons aged ≥18 years. Regulatory information is available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm518291.htm>.
- Flublok Quadrivalent (RIV4; Protein Sciences, Meriden, Connecticut) was licensed by FDA in October 2016, for persons aged ≥18 years. Regulatory information is available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm524660.htm>.
- The age indication for FluLaval Quadrivalent (IIV4; ID Biomedical Corporation of Quebec, Quebec City, Quebec, Canada) was extended from ≥3 years to ≥6 months in November 2016. Regulatory information is available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm366061.htm>. Children aged 6 through 35 months may receive FluLaval Quadrivalent at the same 0.5 mL per dose (containing 15 *µ*g of hemagglutinin \[HA\] per vaccine virus) as is used for older children and adults. This licensure creates an additional option for vaccination of children aged 6 through 35 months, in addition to the previously available 0.25 mL per dose presentation (containing 7.5 *µ*g of HA per vaccine virus) of Fluzone Quadrivalent (IIV4; Sanofi Pasteur, Swiftwater, Pennsylvania).
- Pregnant women may receive any licensed, recommended, age-appropriate influenza vaccine.
- Afluria (IIV3; Seqirus, Parkville, Victoria, Australia) is now recommended for persons aged ≥5 years, consistent with FDA-approved labelling.
- In light of its low effectiveness against influenza A(H1N1)pdm09 in the United States during the 2013--14 and 2015--16 seasons, for the 2017--18 season, ACIP continues the recommendation that LAIV4 should not be used. Because LAIV4 is still a licensed vaccine that might be available and that some providers might elect to use, for informational purposes only, reference is made in this report to previous recommendations for its use.
Recommendations for the Use of Influenza Vaccines, 2017--18 Season
==================================================================
Groups Recommended for Vaccination
----------------------------------
Routine annual influenza vaccination is recommended for all persons aged ≥6 months who do not have contraindications. Recommendations regarding timing of vaccination, considerations for specific populations, the use of specific vaccines, and contraindications and precautions are summarized in the sections that follow.
Timing of Vaccination
---------------------
Optimally, vaccination should occur before onset of influenza activity in the community. Health care providers should offer vaccination by the end of October, if possible. Children aged 6 months through 8 years who require 2 doses (see Children Aged 6 Months through 8 Years) should receive their first dose as soon as possible after vaccine becomes available, to allow the second dose (which must be administered ≥4 weeks later) to be received by the end of October. Although some available data indicate that early vaccination (e.g., in July and August) might be associated with suboptimal immunity before the end of the influenza season, particularly among older adults, the relative contribution of potential waning of immunity compared with those of other determinants of the impact of vaccination (e.g., timing and severity of the influenza season, the impact of missed opportunities when individuals delay vaccination and fail to return later in the season, and programmatic constraints) is unknown. Although delaying vaccination might result in greater immunity later in the season, deferral also might result in missed opportunities to vaccinate, as well as difficulties in vaccinating a population within a more constrained time period. Community vaccination programs should balance maximizing likelihood of persistence of vaccine-induced protection through the season with avoiding missed opportunities to vaccinate or vaccinating after onset of influenza circulation occurs. Revaccination later in the season of persons who have already been fully vaccinated is not recommended.
Vaccination should continue to be offered as long as influenza viruses are circulating and unexpired vaccine is available. To avoid missed opportunities for vaccination, providers should offer vaccination during routine health care visits and hospitalizations when vaccine is available. Vaccination efforts should be structured to ensure the vaccination of as many persons as possible before influenza activity in the community begins.
In any given season, the optimal time to vaccinate cannot be predicted precisely because influenza seasons vary in timing and duration. Moreover, more than one outbreak might occur in a given community in a single year. In the United States, localized outbreaks that indicate the start of seasonal influenza activity can occur as early as October. However, in 74% of influenza seasons from 1982--83 through 2015--16, peak influenza activity (which often is close to the midpoint of influenza activity for the season) has not occurred until January or later, and in 59% of seasons, the peak was in February or later ([@R10]).
In recent seasons, initial shipments of influenza vaccine have arrived to some vaccine providers as early as July. Very early availability of vaccine as compared with typical onset and peak of influenza activity raises questions related to the ideal time to begin vaccination. Several observational studies of influenza vaccine effectiveness have reported decreased vaccine protection within a single season, particularly against influenza A(H3N2) ([@R11]--[@R14]). In some of these studies decline in VE was particularly pronounced among older adults ([@R12],[@R13]). Some studies have documented decline in protective antibodies over the course of one season ([@R15]--[@R17]), with antibody levels decreasing with greater time elapsed postvaccination. However, the rate and degree of decline observed has varied. Among adults in one study, HA and neuraminidase antibody levels declined slowly, with a two-fold decrease in titer estimated to take \>600 days ([@R18]). A review of studies reporting postvaccination seroprotection rates among adults aged ≥60 years noted that seroprotection levels meeting Committee of Proprietary Medicinal Products standards were maintained for ≥4 months for the H3N2 component in all 8 studies and for the H1N1 and B components in five of seven studies ([@R19]). A recent multiseason analysis from the U.S. Influenza Vaccine Effectiveness (U.S. Flu VE) Network found that VE declined by about 7% per month for H3N2 and influenza B, and 6%--11% per month for H1N1pdm09 ([@R20]). VE remained greater than zero for at least 5 to 6 months after vaccination. Similar waning effects have not been observed consistently across age groups and virus subtypes in different populations, and the observed decline in protection could be attributable to bias, unmeasured confounding, or the late season emergence of antigenic drift variants that are less well-matched to the vaccine strain.
Vaccination efforts should continue throughout the season because the duration of the influenza season varies and influenza activity might not occur in certain communities until February or March. Providers should offer influenza vaccine routinely, and organized vaccination campaigns should continue throughout the influenza season, including after influenza activity has begun in the community. Although vaccination by the end of October is recommended, vaccine administered in December or later, even if influenza activity has already begun, is likely to be beneficial in the majority of influenza seasons.
Guidance for Use in Specific Populations and Situations
-------------------------------------------------------
### Populations at Higher Risk for Medical Complications Attributable to Severe Influenza
All persons aged ≥6 months without contraindications should be vaccinated annually. However, vaccination to prevent influenza is particularly important for persons who are at increased risk for severe complications from influenza and for influenza-related outpatient, ED, or hospital visits. When vaccine supply is limited, vaccination efforts should focus on delivering vaccination to the following persons at increased risk for medical complications attributable to severe influenza who do not have contraindications (no hierarchy is implied by order of listing):
- all children aged 6 through 59 months;
- all persons aged ≥50 years;
- adults and children who have chronic pulmonary (including asthma) or cardiovascular (except isolated hypertension), renal, hepatic, neurologic, hematologic, or metabolic disorders (including diabetes mellitus);
- persons who are immunocompromised due to any cause (including immunosuppression caused by medications or by HIV infection);
- women who are or will be pregnant during the influenza season;
- children and adolescents (aged 6 months through 18 years) who are receiving aspirin- or salicylate-containing medications and who might be at risk for experiencing Reye syndrome after influenza virus infection;
- residents of nursing homes and other long-term care facilities;
- American Indians/Alaska Natives; and
- persons who are extremely obese (BMI ≥40).
ACIP recommends that LAIV4 not be used during the 2017--18 season for any population. Providers who elect to use it should consider previous guidance for use of LAIV4 for high-risk populations ([Table 2](#T2){ref-type="table"}).
### Persons Who Live With or Care for Persons at Higher Risk for Influenza-Related Complications
All persons aged ≥6 months without contraindications should be vaccinated annually; however, continued emphasis should be placed on vaccination of persons who live with or care for persons at higher risk for influenza-related complications. When vaccine supply is limited, vaccination efforts should focus on delivering vaccination to persons at higher risk for influenza-related complications listed above, as well as these persons:
- health care personnel, including physicians, nurses, and other workers in inpatient and outpatient-care settings, medical emergency-response workers (e.g., paramedics and emergency medical technicians), and employees of nursing home and long-term care facilities who have contact with patients or residents, and students in these professions who will have contact with patients. ACIP guidance for immunization of health care personnel has been published previously ([@R21]);
- household contacts (including children) and caregivers of children aged ≤59 months (i.e., aged \<5 years) and adults aged ≥50 years, particularly contacts of children aged \<6 months; and
- household contacts (including children) and caregivers of persons with medical conditions that put them at high risk for severe complications from influenza.
ACIP recommends that LAIV4 not be used during the 2017--18 season for any population. Providers who elect to use it should consider previous guidance for use of LAIV4 for persons who care for or have contact with immunocompromised persons ([@R22]). Health care personnel and persons who are contacts of persons in these groups and who are contacts of severely immunocompromised persons (those living in a protected environment) may receive any IIV or RIV that is otherwise indicated. ACIP and HICPAC have previously recommended that health care personnel who receive LAIV should avoid providing care for severely immunosuppressed patients requiring a protected environment for 7 days after vaccination, and that hospital visitors who have received LAIV4 should avoid contact with severely immunosuppressed persons (i.e., persons requiring a protected environment) for 7 days after vaccination. However, such persons should not be restricted from visiting less severely immunosuppressed patients.
### Children Aged 6 Months Through 8 Years
**Dose volume for children aged 6 through 35 months:** Children aged 6 through 35 months may receive one of two products at the appropriate volume for each dose needed: 0.5 mL FluLaval Quadrivalent (containing 15 *µ*g of HA per vaccine virus) or 0.25 mL Fluzone Quadrivalent (containing 7.5 *µ*g of HA per vaccine virus). These are the only two influenza vaccine products licensed for this age group. Care should be taken to administer the appropriate volume for each needed dose of either product. In either instance, the needed volume may be administered from an appropriate prefilled syringe, a single dose vial, or multidose vial, as supplied by the manufacturer. Note, however, that if a 0.5 mL single-use vial of Fluzone Quadivalent is used for a child aged 6 through 35 months, only half the volume should be administered and the other half should be discarded.
Before November 2016, the only influenza vaccine formulations licensed for children aged 6 through 35 months were the 0.25 mL (containing 7.5 *µ*g of HA per vaccine virus) dose formulations of Fluzone and Fluzone Quadrivalent. The recommendation for use of a reduced dose volume for children in this age group (half that recommended for persons aged ≥3 years) was based on increased reactogenicity noted among children (particularly younger children) following receipt of influenza vaccines in trials conducted during the 1970s. This increased reactogenicity was primarily observed with whole-virus inactivated vaccines ([@R23]--[@R27]). Studies with vaccines more similar to currently available split-virus inactivated products demonstrated less reactogenicity ([@R27]). Recent comparative studies of 0.5 mL vs. 0.25 mL doses of IIV3 conducted among children aged 6 through 23 months ([@R28]) and 6 through 35 months ([@R29]) noted no significant difference in reactogenicity at the higher dose. In a randomized trial comparing immunogenicity and safety of 0.5 mL FluLaval Quadrivalent with 0.25 mL Fluzone Quadrivalent, safety and reactogenicity were similar between the two vaccines. In a post-hoc analysis, superior immunogenicity was noted for the B components of FluLaval Quadrivalent among infants aged 6 through 17 months and for unprimed children (those who had not previously received at least 2 doses of influenza vaccine) aged 6 through 35 months ([@R30]).
**Number of doses for children aged 6 months through 8 years:** Evidence from several studies indicates that children aged 6 months through 8 years require 2 doses of influenza vaccine (administered a minimum of 4 weeks apart) during their first season of vaccination for optimal protection ([@R31]--[@R34]). Children aged 6 months through 8 years who have previously received ≥2 total doses of trivalent or quadrivalent influenza vaccine before July 1, 2017 require only 1 dose for 2017--18. The 2 doses of influenza vaccine do not have to have been administered in the same season or consecutive seasons. Children in this age group who have not previously received ≥2 doses of trivalent or quadrivalent influenza vaccine before July 1, 2017 require 2 doses for the 2017--18 season. The interval between the 2 doses should be at least 4 weeks ([Figure](#F1){ref-type="fig"}).
{#F1}
### Pregnant Women
Because pregnant and postpartum women are at higher risk for severe illness and complications from influenza than women who are not pregnant, ACIP recommends that all women who are pregnant or who might be pregnant in the influenza season receive influenza vaccine. Any licensed, recommended, and age-appropriate influenza vaccine may be used. Influenza vaccine can be administered at any time during pregnancy, before and during the influenza season. ACIP recommends that LAIV4 not be used in any population for the 2017--18 season. Providers should note that, as a live virus vaccine, LAIV4 should not be used during pregnancy.
Although experience with the use of IIVs is substantial, and data from observational studies are available to support the safety of these vaccines in pregnancy, data are more limited for vaccination during the first trimester (see Safety of Influenza Vaccines: Pregnant Women and Neonates in the Background Document). Moreover, there is substantially less experience with more recently licensed IIV products (e.g., quadrivalent, cell culture-based, and adjuvanted vaccines) during pregnancy in general. For RIV (available as RIV3 since the 2013--14 influenza season, and as RIV3 and RIV4 for 2017--18), data are limited to reports of pregnancies occurring incidentally during clinical trials, VAERS reports, and pregnancy registry reports. Pregnancy registries and surveillance studies exist for some products; information may be found in package inserts ([@R35]--[@R42]), available at <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm094045.htm> for trivalent vaccines and <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm295057.htm> for quadrivalent vaccines.
### Older Adults
For persons aged ≥65 years, any age-appropriate IIV formulation (standard-dose or high-dose, trivalent or quadrivalent, unadjuvanted or adjuvanted) or RIV are acceptable options. Fluzone High-Dose (HD-IIV3; Sanofi Pasteur, Swiftwater, Pennsylvania) met prespecified criteria for superior efficacy to that of SD-IIV3 in a randomized trial conducted over two seasons among 31,989 persons aged ≥65 years, and might provide better protection than SD-IIV3 for this age group ([@R43]--[@R45]). In an exploratory analysis of data from a single-season randomized trial conducted among 8,604 adults aged ≥50 years, Flublok Quadrivalent (RIV4; Protein Sciences, Meriden, Connecticut) was more efficacious than SD-IIV4 ([@R46],[@R47]); however, no claim of superiority was approved for the package insert ([@R47]). Fluad (aIIV3; Seqirus, Holly Springs, North Carolina) was more effective against laboratory-confirmed influenza than unadjuvanted SD-IIV3 among adults aged ≥65 years (N = 227) in an analysis from a small observational study ([@R48]). No preferential recommendation is made for any specific vaccine product. Vaccination should not be delayed if a specific product is not readily available.
Because of the vulnerability of this population to severe influenza illness, hospitalization, and death, efficacy and effectiveness of influenza vaccines among older adults is an area of active research (see Immunogenicity, Efficacy, and Effectiveness of Influenza Vaccines: HD-IIV3, aIIV3, and RIV4 for Older Adults in the Background Document). Recent comparative studies of efficacy/effectiveness against laboratory-confirmed influenza outcomes among older adults have focused on HD-IIV3 (Fluzone High Dose; Sanofi Pasteur, Swiftwater, Pennsylvania) ([@R43],[@R49]--[@R51]), aIIV3 (Fluad, Seqirus, Holly Springs, North Carolina) ([@R48]), and RIV4 (Flublok Quadrivalent; Protein Sciences, Meriden, Connecticut) ([@R46]). Characteristics of these studies are summarized ([Table 3](#T3){ref-type="table"}). In each instance, the comparator vaccines have been standard dose, inactivated vaccines (SD-IIV3 as the comparator for HD-IIV3 and aIIV3; SD-IIV4 as the comparator for RIV4). No data are yet available from studies comparing the efficacy or effectiveness of HD-IIV3, aIIV3, and RIV4 with one another among older adults. This lack of comparative data prevents recommending one of these three vaccines over another for this population. HD-IIV3 exhibited superior efficacy over a comparator standard-dose IIV3 for adults aged ≥65 years in a large (N = 31,989), two-season randomized, controlled, double-blind trial ([@R43],[@R44]), and might provide better protection than SD-IIV3s for this age group. Additional data concerning relative efficacy of HD-IIV3 for other clinical outcomes, as well as cost-effectiveness analyses and observational studies, are summarized in the Background Document. In a single-season randomized, controlled, double-blind trial comparing RIV4 with a standard-dose unadjuvanted IIV4 among adults aged ≥50 years (N = 8,604), RIV4 was more effective; however, approval for a claim of superiority was not made as a result of this exploratory analysis ([@R46],[@R47]). Additional data, including discussion of immunogenicity studies, are described in the Background Document. Fluad (aIIV3; Seqirus, Holly Springs, North Carolina) was more effective against laboratory-confirmed influenza than unadjuvanted SD-IIV3 among adults aged ≥65 years (N = 227) in an analysis from a small observational study ([@R48]); no data are yet available concerning efficacy of Fluad compared with nonadjuvanted IIV3 against laboratory-confirmed influenza outcomes from a randomized trial in this population. Additional data concerning aIIV3, from studies examining immunogenicity and non-laboratory confirmed influenza outcomes, are discussed in the Background Document. ACIP will continue to review data concerning the efficacy and effectiveness of these vaccines as more information emerges.
###### Summary of studies of efficacy/effectiveness of HD-IIV3, aIIV3, and RIV4 compared with unadjuvanted SD-IIVs against laboratory-confirmed influenza among older adults[\*](#FN1){ref-type="fn"}
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Comparator (proprietary name) Study design Ages enrolled No. participants Season(s) (predominant viruses)^†^ Location Primary outcome Main efficacy/effectiveness findings
--------------------------------------------------- ---------------------------------------------------------------------- -------------------------------------------- ----------------------------------- --------------------------------------------------------------- ------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**HD-IIV3 (Fluzone High-Dose, Sanofi Pasteur)**
SD-IIV3 (Fluzone)^§^ RCT, double-blind ≥65 years 6,107 HD-IIV3
3,051 SD-IIV3 2009--10 (\[H1N1\]pdm09; not contained in HD-IIV3 or SD-IIV3) U.S. (99 sites) Culture- and/or RT-PCR-confirmed ILI, caused by types/subtypes similar to those contained in the vaccine Primary outcome not evaluable due to emergence of (H1N1)pdm09 pandemic (no cases meeting primary endpoint of laboratory-confirmed influenza caused by viral types/subtypes similar to those in vaccine were observed; all confirmed cases were due to \[H1N1\]pdm09)
SD-IIV3 (Fluzone)^¶^ RCT, double-blind ≥65 years 15,990 HD-IIV3
15,993 SD-IIV3 2011--12 (H3N2) and 2012--13 (\[H3N2\], mismatch) U.S., Canada (126 sites) Laboratory confirmed (culture- and/or RT-PCR) influenza caused by any influenza viral types or subtypes, in association with protocol-defined ILI **Culture- and/or RT-PCR-confirmed influenza; any type or subtype, with protocol-defined ILI:**
All influenza: RE 24.2% (95% CI = 9.7--36.5)
Influenza A: RE 24.0% (95% CI = 7.8--37.4)
Influenza B: RE 25.5% (95% CI =-15.7--52.4)
**Culture- and/or RT-PCR-confirmed influenza; strains similar to vaccine, with protocol-defined ILI:**
All influenza: RE 35.4% (95% CI = 12.5--52.5)
Influenza A: RE 31.7% (95% CI = 2.9--52.3)
Influenza B: RE 45.2% (95% CI = -2.2--71.5)
**Culture-confirmed influenza; any type or subtype, with protocol-defined ILI:**
All influenza: RE 23.1% (95% CI = 7.5--36.2)
Influenza A: RE 23.4% (95% CI = 6.0--37.6)
Influenza B: RE 21.7% (95% CI =-23.8--50.8)
**Culture-confirmed influenza; strains similar to vaccine. with protocol-defined ILI:**\
All influenza: RE 31.5% (95% CI = 4.6--51.1)\
Influenza A: RE 27.0% (95% CI = -8.5--51.2)
Influenza B: RE 41.4% (95% CI =-10.3--69.8)
**RIV4 (Flublok Quadrivalent, Protein Sciences)**
SD-IIV4 (Fluarix Quadrivalent)\*\* RCT, double-blind ≥50 years 4,303 RIV4
4,301 IIV4 2014--15 (H3N2, mismatch) U.S. (40 sites) RT-PCR-confirmed ILI caused by any type or subtype. **RT-PCR-positive protocol-defined ILI;**
**Aged ≥50 years:**
All influenza: RE 30% (95% CI = 10--47)
Influenza A: RE 36% (95%CI = 14--53)
Influenza B: RE 4% (95% CI =: -72--46)
**RT-PCR-positive protocol-defined ILI;
Aged 50 through 64 years:**
All influenza: RE 42% (95% CI = 15--61)
**Aged ≥65 years:**
All influenza: RE 17% (95% CI = -20--43)
**Culture-positive protocol-defined ILI:
Aged ≥50 years:**
All influenza: RE 43% (95% CI = 21--59)
Influenza A: RE 44% (95% CI = 22--61)
Influenza B: RE 25% (95% CI =-121--75)
**Culture-positive protocol-defined ILI:**\
**Aged 50--64 years:**
All influenza: RE 44% (95% CI = 10--65)\
**Aged ≥65 years:**
All influenza: RE 42% (95% CI = 9--65)
**aIIV3 (Fluad, Seqirus)**
SD-IIV3;
unvaccinated^††^ Nonrandomized, observational, prospective test negative case-control ≥65 years, presenting with symptoms of ILI 165 aIIV3
62 IIV3
55 unvaccinated 2011--12
(H3N2) Canada (3 health authorities) RT-PCR- confirmed ILI **Effectiveness of aIIV3 vs. unvaccinated:**\
58% (95% CI = 5--82)
**Effectiveness of IIV3 vs. unvaccinated:**\
-2% (95% CI = -139--57)
**Relative effectiveness of aIIV3 vs. IIV3:**\
63% (95% CI = 4--86)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
**Abbreviations:** aIIV3=trivalent inactivated influenza vaccine, adjuvanted; CI = confidence interval; ILI=influenza-like illness; HD-IIV3 = High-Dose Inactivated Influenza Vaccine, trivalent ; RE=relative efficacy (compared to active comparator vaccine); RIV4 = Recombinant Influenza Vaccine, quadrivalent; SD-IIV3 = Standard-Dose Inactivated Influenza Vaccine, trivalent.
\* Studies conducted among human participants of laboratory-confirmed (by viral culture and/or RT-PCR) influenza outcomes. Modeling and medical record database studies are not included in this Table, but are discussed in the Background Document.
^†^ Information on predominant viruses is from U.S. national surveillance data (CDC, FluView, available at <https://www.cdc.gov/flu/weekly/index.htm>).
^§^ **Source:** DiazGranados CA, Dunning AJ, Jordanov E, Landolfi V, Denis M, Talbot HK. High-dose trivalent influenza vaccine compared to standard dose vaccine in elderly adults: safety, immunogenicity and relative efficacy during the 2009--2010 season. Vaccine 2013;31:861--6.
^¶^ **Source:** DiazGranados CA, Dunning AJ, Kimmel M, et al. Efficacy of high-dose versus standard-dose influenza vaccine in older adults. N Engl J Med 2014;371:635--45.
\*\* **Source:** Dunkle LM, Izikson R, Patriarca P, et al.; PSC12 Study Team. Efficacy of recombinant influenza vaccine in adults 50 years of age or older. N Engl J Med 2017;376:2427--36.
^††^ **Source:** Van Buynder PG, Konrad S, Van Buynder JL, et al. The comparative effectiveness of adjuvanted and unadjuvanted trivalent inactivated influenza vaccine (TIV) in the elderly. Vaccine 2013;31:6122--8.
### Immunocompromised Persons
Immunocompromised states comprise a heterogeneous range of conditions. In many instances, limited data are available regarding the use of influenza vaccines in the setting of specific immunocompromised states. ACIP recommends that LAIV4 not be used in any population for the 2017--18 season; providers considering its use should note that live virus vaccines should not be used for persons with most forms of altered immunocompetence ([@R52]), given the uncertain but biologically plausible risk for disease attributable to the vaccine virus. In addition to potential safety issues, immune response to live or inactivated vaccines might be blunted in some clinical situations, such as for persons with congenital immune deficiencies, persons receiving cancer chemotherapy, and persons receiving immunosuppressive medications. For this reason, timing of vaccination might be a consideration (e.g., vaccinating during some period either before or after an immunocompromising intervention).
The Infectious Diseases Society of America (IDSA) has published detailed guidance for the selection and timing of vaccines for persons with specific immunocompromising conditions, including congenital immune disorders, stem cell and solid organ transplant, anatomic and functional asplenia, and therapeutic drug-induced immunosuppression, as well as for persons with cochlear implants or other conditions leading to persistent cerebrospinal fluid-oropharyngeal communication ([@R53]). ACIP will continue to review accumulating data on use of influenza vaccines in these contexts.
### Persons with a History of Guillain-Barré Syndrome Following Influenza Vaccination
A history of Guillain-Barré Syndrome (GBS) within 6 weeks following a previous dose of any type of influenza vaccine is considered a precaution to vaccination ([Table 2](#T2){ref-type="table"}). Persons who are not at high risk for severe influenza complications (see Populations at Higher Risk for Medical Complications Attributable to Severe Influenza) and who are known to have experienced GBS within 6 weeks of a previous influenza vaccination generally should not be vaccinated. As an alternative to vaccination, physicians might consider using influenza antiviral chemoprophylaxis for these persons ([@R54]). However, the benefits of influenza vaccination might outweigh the risks for certain persons who have a history of GBS and who also are at high risk for severe complications from influenza.
### Persons with a History of Egg Allergy
As is the case for other vaccines, influenza vaccines contain various different components that might cause allergic and anaphylactic reactions. Not all such reactions are related to egg proteins; however, the possibility of reactions to influenza vaccines in egg-allergic persons might be of concern to these persons and vaccine providers. Currently available influenza vaccines, with the exceptions of RIV3, RIV4 and ccIIV4, are prepared by propagation of virus in embryonated eggs. Only RIV3 and RIV4 are considered egg-free. For ccIIV4 (Flucelvax Quadrivalent; Seqirus, Holly Springs, North Carolina), ovalbumin is not directly measured. During manufacture of ccIIV4, viruses are propagated in mammalian cells rather than in eggs; however, some of the viruses provided to the manufacturer are egg-derived, and therefore egg proteins may potentially be introduced at the start of the manufacturing process. Once these viruses are received by the manufacturer, no eggs are used, and dilutions at various steps during the manufacturing process result in a theoretical maximum of 5x10^-8^ *μ*g/0.5 mL dose of total egg protein (Seqirus, unpublished data, 2016).
Severe allergic reactions to vaccines, although rare, can occur at any time, despite a recipient's allergy history. Therefore, all vaccine providers should be familiar with the office emergency plan, and be certified in cardiopulmonary resuscitation ([@R52]). For persons who report a history of egg allergy, ACIP recommends the following (based upon the recipient's previous symptoms after exposure to egg):
- Persons with a history of egg allergy who have experienced only urticaria (hives) after exposure to egg should receive influenza vaccine. Any licensed and recommended influenza vaccine (i.e., any IIV or RIV) that is otherwise appropriate for the recipient's age and health status may be used.
- Persons who report having had reactions to egg involving symptoms other than urticaria (hives), such as angioedema, respiratory distress, lightheadedness, or recurrent emesis; or who required epinephrine or another emergency medical intervention, may similarly receive any licensed and recommended influenza vaccine (i.e., any IIV or RIV) that is otherwise appropriate for the recipient's age and health status. The selected vaccine should be administered in an inpatient or outpatient medical setting (including, but not necessarily limited to, hospitals, clinics, health departments, and physician offices). Vaccine administration should be supervised by a health care provider who is able to recognize and manage severe allergic conditions.
- A previous severe allergic reaction to influenza vaccine, regardless of the component suspected of being responsible for the reaction, is a contraindication to future receipt of the vaccine.
No period of postvaccination observation period is recommended specifically for egg-allergic persons. However, ACIP recommends that vaccine providers consider observing patients for 15 minutes following administration of any vaccine to decrease the risk for injury should syncope occur ([@R52]).
Persons who are able to eat lightly cooked egg (e.g., scrambled egg) without reaction are unlikely to be allergic. Egg-allergic persons might tolerate egg in baked products (e.g., bread or cake). Tolerance to egg-containing foods does not exclude the possibility of egg allergy. Egg allergy can be confirmed by a consistent medical history of adverse reactions to eggs and egg-containing foods, plus skin and/or blood testing for immunoglobulin E directed against egg proteins ([@R55]).
Occasional cases of anaphylaxis in egg-allergic persons have been reported to VAERS after administration of influenza vaccines ([@R56],[@R57]). ACIP will continue to review available data regarding anaphylaxis cases following influenza vaccines.
### Vaccination Issues for Travelers
Travelers who want to reduce the risk for influenza infection should consider influenza vaccination, preferably at least 2 weeks before departure. In particular, persons residing in the United States who are at high risk for complications of influenza and who were not vaccinated with influenza vaccine during the preceding Northern Hemisphere fall or winter should consider receiving influenza vaccine before departure if they plan to travel:
- to the tropics,
- with organized tourist groups or on cruise ships, or
- to the Southern Hemisphere during the Southern Hemisphere influenza season (April--September).
No information is available indicating a benefit to revaccinating persons before summer travel who already were vaccinated during the preceding fall. In many cases, revaccination will not be feasible as Southern Hemisphere formulations of influenza vaccine are not generally available in the United States. Persons at high risk who receive the previous season's vaccine before travel should receive the current vaccine the following fall or winter. Persons at higher risk for influenza complications should consult with their health care practitioner to discuss the risk for influenza or other travel-related diseases before embarking on travel during the summer.
In temperate climate regions of the Northern and Southern hemispheres, influenza activity is seasonal, occurring approximately from October through May in the Northern Hemisphere and April through September in the Southern Hemisphere. In the tropics, influenza occurs throughout the year. Travelers can be exposed to influenza when travelling to an area where influenza is circulating, or when traveling as part of large tourist groups (e.g., on cruise ships) that include persons from areas of the world in which influenza viruses are circulating ([@R58],[@R59]). In a survey of Swiss travelers to tropical and subtropical countries, among 211 who reported febrile illness during or after travelling abroad and who provided paired serum samples, 40 demonstrated serologic evidence of influenza infection ([@R60]). Among 109 travelers returning to Australia from travel in Asia who reported acute respiratory infection symptoms, four (3.7%) had evidence of influenza A infection (evidenced by fourfold rise in antibody titer) ([@R61]).
Influenza vaccine formulated for the Southern Hemisphere might differ in viral composition from the Northern Hemisphere vaccine. However, with the exception of the Southern Hemisphere formulation of Fluzone Quadrivalent (IIV4; Sanofi Pasteur, Swiftwater, Pennsylvania), Southern Hemisphere formulation seasonal influenza vaccines are not licensed in the United States, and Southern Hemisphere formulations generally are not commercially available in the United States. More information on influenza vaccines and travel is available at <https://www.cdc.gov/flu/travelers/travelersfacts.htm>***.***
### Use of Influenza Antiviral Medications
Administration of IIV or RIV to persons receiving influenza antiviral medications for treatment or chemoprophylaxis is acceptable. ACIP recommends that LAIV4 should not be used during the 2017--18 season. If used, providers should note that influenza antiviral medications may reduce the effectiveness of LAIV4 if given within 48 hours before to 14 days after LAIV4 ([@R62]). Persons who receive influenza antiviral medications during this period surrounding receipt of LAIV4 may be revaccinated with another appropriate vaccine formulation (e.g., IIV or RIV).
### Concurrent Administration of Influenza Vaccine with Other Vaccines
Data regarding potential interference following simultaneous or sequential administration for the many potential combinations of vaccines are limited. Therefore, following the ACIP General Best Practice Guidelines for Immunization is prudent ([@R52]). IIVs and RIV may be administered concurrently or sequentially with other inactivated vaccines or with live vaccines. LAIV4 is not recommended for use in 2017--18. Providers considering its use should note that although inactivated or live vaccines can be administered simultaneously with LAIV4, after administration of a live vaccine (such as LAIV4), at least 4 weeks should pass before another live vaccine is administered.
Relatively limited data are available on the concurrent administration of influenza vaccines with other vaccines. In a study comparing the immunogenicity of IIV and zoster vaccine given either concurrently or separated by a 4-week interval to adults aged ≥50 years, antibody responses were similar for both schedules ([@R63]). In some studies, reduced responses have been noted to PCV13 ([@R64],[@R65]), tetanus antigens ([@R66]), and pertussis antigens ([@R66]) when co-administered with IIV; in most instances the clinical significance of this is uncertain. Reassuring safety profiles have been noted for simultaneous administration of zoster vaccine ([@R63]), PCV13 ([@R64],[@R65]), PPSV23 ([@R67]) and Tdap ([@R66]) among adults and of Tdap among pregnant women ([@R68]). Increased prevalence of local and/or systemic adverse reactions have been noted with concurrent administration in some of these studies, but these symptoms have generally been reported to be mild or moderate. Among children, co-administration of IIV and PCV13 was associated with increased risk of fever on the day of vaccination and the day following (i.e., days 0--1 postvaccination) after vaccination in children aged 6 through 23 months in a study conducted during the 2011--12 season ([@R69]). Increased risk of febrile seizure in this age group has been noted within days 0--1 following coadministration of IIV with PCV7, PCV13, or DTaP-containing vaccines during the 2006--07 through 2010--11 seasons ([@R70]), and with PCV13 during the 2014--15 season ([@R71]). No changes in the recommendations for administration of these vaccines were made, and these vaccines may be given concomitantly. Surveillance of febrile seizures is ongoing through VAERS, and the Vaccine Safety Datalink annual influenza vaccine safety surveillance includes monitoring for seizures following vaccination.
Concurrent administration to children of LAIV3 with MMR and varicella vaccine was not associated with diminished immunogenicity to antigens in any of the vaccines in one study ([@R72]); diminished response to rubella was observed in another examining coadministraion of LAIV3 and MMR ([@R73]). Administration of OPV was not associated with interference when administered with LAIV ([@R74]). No safety concerns were revealed in these studies.
Influenza Vaccine Composition and Available Products
====================================================
Influenza Vaccine Composition for the 2017--18 Season
-----------------------------------------------------
All influenza vaccines licensed in the United States will contain components derived from influenza viruses antigenically similar to those recommended by FDA ([@R75]). Both trivalent and quadrivalent influenza vaccines will be available in the United States. The 2017--18 U.S. influenza vaccines will contain the following components:
- an A/Michigan/45/2015 (H1N1)pdm09--like virus,
- an A/Hong Kong/4801/2014 (H3N2)--like virus, and
- a B/Brisbane/60/2008--like virus (Victoria lineage).
The 2017--18 U.S. quadrivalent vaccines will contain the same three antigens and an additional influenza B virus component, a B/Phuket/3073/2013--like virus (Yamagata lineage). Compared with 2016--17, the composition for 2017--18 represents a change in the influenza A(H1N1)pdm09--like virus.
Vaccine Products for the 2017--18 Season
----------------------------------------
A variety of influenza vaccine products are licensed and available from several different manufacturers ([Table 1](#T1){ref-type="table"}). For many vaccine recipients, more than one type or brand of vaccine might be appropriate within approved indications and ACIP recommendations. A licensed, age-appropriate influenza vaccine product should be used. Not all products are likely to be uniformly available in any practice setting or locality. Vaccination should not be delayed in order to obtain a specific product when an appropriate one is already available. Within these guidelines and approved indications, where more than one type of vaccine is appropriate and available, no preferential recommendation is made for use of any influenza vaccine product over another.
Since the publication of the previous season's guidelines, two new influenza vaccine products have been licensed; Afluria Quadrivalent (IIV4; Seqirus, Parkville, Victoria, Australia) and Flublok Quadrivalent (RIV4; Protein Sciences, Meriden, Connecticut). In addition, a labeling change has been approved for a previously-licensed product: FluLaval Quadrivalent (IIV4, ID Biomedical Corporation of Quebec, Quebec City, Quebec, Canada) is now licensed for children aged 6 months and older. These are described in the New Influenza Vaccine Product Approvals section. New licensures and changes to FDA-approved labeling might occur subsequent to this report. These recommendations apply to all licensed influenza vaccines used within FDA--licensed indications, including changes in FDA-approved labeling that might occur after publication of this report. As these changes occur, they will be reflected in the online version of [Table 1](#T1){ref-type="table"}, available at <https://www.cdc.gov/flu/protect/vaccine/vaccines.htm>.
Dosage, Administration, Contraindications, and Precautions
----------------------------------------------------------
### Inactivated Influenza Vaccines (IIVs)
**Available products:** IIVs comprise multiple products ([Table 1](#T1){ref-type="table"}). Both quadrivalent and trivalent formulations are available.
With one exception, U.S.-licensed IIVs are manufactured through propagation of virus in eggs. The exception, the cell culture-based vaccine Flucelvax Quadrivalent (ccIIV4; Seqirus, Holly Springs, North Carolina), contains vaccine viruses propagated in Madin-Darby canine kidney cells. Flucelvax Quadrivalent is not considered egg-free, as some of the initial vaccine viruses provided to the manufacturer by WHO are egg-derived. For the 2017--18 season, the influenza A (H1N1) and both influenza B components will be egg-derived; the influenza A (H3N2) component will be cell-derived.
With one exception, IIVs licensed in the United States contain no adjuvant. The exception, Fluad (aIIV3; Seqirus, Holly Springs, North Carolina) contains the adjuvant MF59.
There are IIVs that are licensed for persons as young as age 6 months. However, age indications for the various individual IIVs differ ([Table 1](#T1){ref-type="table"}). Only age-appropriate products should be administered. Afluria (IIV3, Seqirus, Parkville, Victoria, Australia), which was previously recommended for persons aged ≥9 years, is now recommended for persons aged ≥5 years. Providers should consult package inserts and updated CDC/ACIP guidance for current information.
**Dosage and administration:** All IIV preparations contain 15 *µ*g of HA per vaccine virus strain (45 *µ*g total for IIV3s and 60 *µ*g total for IIV4s) per 0.5 mL dose, with two exceptions. Fluzone High Dose (HD-IIV3; Sanofi Pasteur, Swiftwater, Pennsylvania), an IIV3 licensed for persons aged ≥65 years, contains 60 *µ*g of each HA per vaccine virus strain (180 *µ*g total) ([@R44]). Fluzone Intradermal Quadrivalent (intradermal IIV4; Sanofi Pasteur, Swiftwater, Pennsylvania), an intradermally administered IIV4 licensed for persons aged 18 through 64 years, contains 9 *µ*g of each HA per vaccine virus strain (36 *µ*g total) ([@R36]).
For children aged 6 through 35 months, two IIV products are licensed by FDA. The approved dose volumes differ for these two products. Children in this age group may receive either 1) 0.5 mL of FluLaval Quadrivalent (ID Biomedical Corporation of Quebec, Quebec City, Quebec, Canada) ([@R37]), which contains 15 *µ*g of HA per virus, or 2) 0.25 mL dose Fluzone Quadrivalent (Sanofi Pasteur, Swiftwater, Pennsylvania) ([@R35]), which contains 7.5 *µ*g of HA per virus. Care must be taken to administer each at the appropriate dose for each product in this age group. If prefilled syringes are not available, the appropriate dose can be taken from a single-use or multidose vial, at the appropriate volume for the given product.
Children aged 36 months through 17 years (for whom only intramuscular IIVs are licensed) and adults aged ≥18 years who are receiving intramuscular preparations of IIV should receive 0.5 mL per dose. If a smaller intramuscular vaccine dose (e.g., 0.25 mL) is administered inadvertently to an adult, an additional 0.25 mL dose should be administered to provide a full dose (0.5 mL). If the error is discovered later (after the patient has left the vaccination setting), a full 0.5 mL dose should be administered as soon as the patient can return. Vaccination with a formulation approved for adult use should be counted as a dose if inadvertently administered to a child.
With the exception of Fluzone Intradermal Quadrivalent (Sanofi Pasteur, Swiftwater, Pennsylvania), IIVs are administered intramuscularly. For adults and older children, the deltoid is the preferred site. Infants and younger children should be vaccinated in the anterolateral thigh. Additional specific guidance regarding site selection and needle length for intramuscular administration are provided in the ACIP General Best Practice Guidelines for Immunization ([@R52]). Fluzone Intradermal Quadrivalent is administered intradermally, preferably over the deltoid muscle, using the included delivery system ([@R36]). Two IIVs, Afluria and Afluria Quadrivalent (Seqirus, Parkville, Victoria, Australia), are licensed for intramuscular administration via jet injector (Stratis; Pharmajet, Golden, Colorado) for persons aged 18 through 64 years ([@R39],[@R76]).
**Trivalent versus quadrivalent IIVs:** Both trivalent and quadrivalent IIVs will be available during the 2017--18 season. Quadrivalent vaccines contain one virus from each of the two influenza B lineages (one B/Victoria virus and one B/Yamagata virus), whereas trivalent vaccines contain one influenza B virus from one lineage. Quadrivalent vaccines are thus designed to provide broader protection against circulating influenza B viruses. However, no preference is expressed for either IIV3 or IIV4.
**Contraindications and precautions for the use of IIVs:** Manufacturer package inserts and updated CDC/ACIP guidance should be consulted for current information on contraindications and precautions for individual vaccine products. In general, history of severe allergic reaction to the vaccine or any of its components (including egg) is a labeled contraindication to the receipt of IIVs ([Table 2](#T2){ref-type="table"}). However, ACIP makes specific recommendations for the use of influenza vaccine for persons with egg allergy (see Persons with a History of Egg Allergy). Influenza vaccine is not recommended for persons with a history of severe allergic reaction to the vaccine or to components other than egg. Information about vaccine components is located in package inserts from each manufacturer. Prophylactic use of antiviral agents is an option for preventing influenza among persons who cannot receive vaccine ([@R54]).
Moderate or severe acute illness with or without fever is a general precaution for vaccination ([@R52]). GBS within 6 weeks following a previous dose of influenza vaccine is considered a precaution for use of influenza vaccines ([Table 2](#T2){ref-type="table"}).
### Recombinant Influenza Vaccine (RIVs)
**Available products:** Two RIV products, Flublok (RIV3) and Flublok Quadrivalent (RIV4), are expected to be available for the 2017--18 influenza season. RIV3 and RIV4 are indicated for persons aged ≥18 years. RIVs are manufactured without the use of influenza viruses; therefore, similarly to IIVs, no shedding of vaccine virus will occur. These vaccines are produced without the use of eggs, and are egg-free. No preference is expressed for RIVs versus IIVs within specified indications.
**Dosage and administration:** RIVs are administered by intramuscular injection. A 0.5 mL dose contains 45 *µ*g of HA derived from each vaccine virus (135 *µ*g total for RIV3 and 180 *µ*g total for RIV4).
**Trivalent versus quadrivalent RIV:** Both trivalent and quadrivalent RIV will be available during the 2017--18 season. Quadrivalent vaccines contain one virus from each of the two influenza B lineages (one B/Victoria virus and one B/Yamagata virus), whereas trivalent vaccines contain one influenza B virus from one lineage. Quadrivalent vaccines are thus designed to provide broader protection against circulating influenza B viruses. However, no preference is expressed for either RIV3 or RIV4.
**Contraindications and precautions for use of RIV:** RIVs are contraindicated in persons who have had a severe allergic reaction to any component of the vaccine. Moderate or severe acute illness with or without fever is a general precaution for vaccination ([@R52]). GBS within 6 weeks following a previous dose of influenza vaccine is considered a precaution for use of influenza vaccines ([Table 2](#T2){ref-type="table"}). Flublok is not licensed for use in children aged \<18 years.
### Live Attenuated Influenza Vaccine (LAIV4)
Pending further data, for the 2017--18 season, ACIP recommends that LAIV4 not be used because of concerns regarding its effectiveness against influenza A(H1N1)pdm09 viruses in the United States during the 2013--14 and 2015--16 seasons. As it is a licensed vaccine and might be available during 2017--18, the material in this section is provided for information.
**Dosage and administration:** LAIV4 is administered intranasally using the supplied prefilled, single-use sprayer containing 0.2 mL of vaccine. Approximately 0.1 mL (i.e., half of the total sprayer contents) is sprayed into the first nostril while the recipient is in the upright position. An attached dose-divider clip is removed from the sprayer to administer the second half of the dose into the other nostril. If the vaccine recipient sneezes immediately after administration, the dose should not be repeated. However, if nasal congestion is present that might impede delivery of the vaccine to the nasopharyngeal mucosa, deferral of administration should be considered until resolution of the illness, or another appropriate vaccine should be administered instead.
**Contraindications and precautions:** ACIP recommends that LAIV4 not be used during the 2017--18 season. Previously issued guidance regarding contraindications and precautions is provided for informational purposes only ([Table 2](#T2){ref-type="table"}).
New Influenza Vaccine Product Approvals
---------------------------------------
Since the publication of the previous season's guidance, there have been two new product approvals (Afluria Quadrivalent and Flublok Quadrivalent) and one change to the approved indication for an existing product (expansion of the age indication for FluLaval Quadrivalent from ≥3 years to ≥6 months).
### Afluria Quadrivalent (IIV4)
Afluria Quadrivalent (IIV4, Seqirus, Parkville, Victoria, Australia) was licensed by FDA in August 2016, for persons aged ≥18 years, and was available during the 2016--17 season alongside the trivalent formulation of Afluria. In a prelicensure study of the safety and immunogenicity of Afluria Quadrivalent compared with two formulations of Afluria (each containing one of the two influenza B viruses contained in the quadrivalent) among persons aged ≥18 years, Afluria Quadrivalent met pre-specified criteria for immunologic noninferiority for all four vaccine viruses, and criteria for immunologic superiority for each B virus as compared to the trivalent formulation containing the alternate B virus.
Some local injection site reactions were more common among those who received Afluria Quadrivalent, including an imbalance of Grade 3 injection site induration/swelling at 0.3% in the Afluria Quadrivalent group and 0.06% in the pooled Afluria trivalent groups), but rates of these reactions were low overall ([@R77],[@R78]).
### Flublok Quadrivalent (RIV4)
Flublok Quadrivalent (RIV4; Protein Sciences, Meriden, Connecticut) was licensed by FDA in October 2016, for persons aged ≥18 years. It is anticipated that Flublok Quadrivalent will be available for the 2017--18 season, alongside the trivalent formulation of Flublok. In a prelicensure analysis of immunogenicity data from a subset of participants enrolled in a randomized relative efficacy trial comparing Flublok Quadrivalent with a licensed comparator standard-dose IIV4 among persons aged ≥50 years during the 2014--15 season, Flublok Quadrivalent met criteria for noninferiority to the comparator IIV4 for the A(H3N2) and B/Yamagata antigens, but not for the A(H1N1) or B/Victoria antigens ([@R47]). In an exploratory analysis of data from this trial (N = 8,604), Flublok Quadrivalent demonstrated 30% greater relative efficacy (95% confidence interval \[CI\] = 10--47) over IIV4 ([@R46],[@R47]). In a second prelicensure study, evaluating safety, reactogenicity, and immunogenicity compared with a licensed IIV4 among persons aged 18 through 49 years during the 2014--15 season, Flublok Quadrivalent met criteria for noninferiority to the comparator IIV4 for the A(H1N1), A(H3N2) and B/Yamagata antigens, but not for the B/Victoria antigen. Safety data from both studies suggested comparable safety to the comparator IIV4 for persons aged ≥18 years ([@R47]).
### FluLaval Quadrivalent (IIV4)
In November 2016, FDA approved expansion of the licensed age indication for FluLaval Quadrivalent (IIV4; ID Biomedical Corporation of Quebec, Quebec City, Quebec, Canada). Previously licensed for persons aged ≥3 years, FluLaval Quadrivalent is now licensed for persons aged ≥6 months. The approved dose volume is 0.5 mL for all ages. This represents a new option for vaccination of children aged 6 through 35 months, for whom previously the only approved influenza vaccine formulation was the 0.25 mL dose volume of Fluzone Quadrivalent. With this approval, children in this age group may receive either 0.5 mL of FluLaval Quadrivalent or 0.25 mL of Fluzone Quadrivalent for each dose needed.
In a prelicensure study comparing the immunogenicity and safety of 0.5 mL of FluLaval Quadrivalent to that of 0.25 mL of Fluzone Quadrivalent among children aged 6 through 35 months, FluLaval Quadrivalent met criteria for immunogenic noninferiority for all four vaccine strains. Safety and reactogenicity were similar between the two vaccines ([@R30],[@R79]).
Storage and Handling of Influenza Vaccines
------------------------------------------
In all instances, approved manufacturer packaging information should be consulted for authoritative guidance concerning storage and handling of all influenza vaccines. Vaccines should be protected from light and stored at recommended temperatures. In general, influenza vaccines are recommended to be stored refrigerated between 2° to 8°C (36° to 46°F) and should not be frozen. Vaccine that has frozen should be discarded. In addition, the cold chain must be maintained when LAIV4 is transported. Single-dose vials should not be accessed for more than one dose. Multiple-dose vials should be returned to recommended storage conditions between uses, and once first accessed should not be kept beyond the recommended period of time. For information on permissible temperature excursions and other departures from recommended storage conditions that are not discussed in the package labelling, contact the manufacturer. Vaccines should not be used after the expiration date on the label.
Additional Sources for Information Regarding Influenza and Vaccines
===================================================================
Influenza Surveillance, Prevention, and Control
-----------------------------------------------
Updated information regarding influenza surveillance, detection, prevention, and control is available at <https://www.cdc.gov/flu>. U.S surveillance data are updated weekly during October--May on FluView (<https://www.cdc.gov/flu/weekly>). In addition, periodic updates regarding influenza are published in *MMWR* (<https://www.cdc.gov/mmwr>). Additional information regarding influenza vaccine can be obtained from CDC by calling 1-800-232-4636. State and local health departments should be consulted about availability of influenza vaccine, access to vaccination programs, information related to state or local influenza activity, reporting of influenza outbreaks and influenza-related pediatric deaths, and advice concerning outbreak control.
Vaccine Adverse Event Reporting System
--------------------------------------
The National Childhood Vaccine Injury Act of 1986 requires health care providers to report any adverse event listed by the vaccine manufacturer as a contraindication to further doses of the vaccine, or any adverse event listed in the VAERS Table of Reportable Events Following Vaccination (<https://vaers.hhs.gov/docs/VAERS_Table_of_Reportable_Events_Following_Vaccination.pdf>) that occurs within the specified time period after vaccination. In addition to mandated reporting, health care providers are encouraged to report any clinically significant adverse event following vaccination to VAERS. Information on how to report a vaccine adverse event is available at <https://vaers.hhs.gov/index.html>[.]{.ul} Additional information on VAERS and vaccine safety is available by emailing <[email protected]> or by calling 1-800-822-7967.
National Vaccine Injury Compensation Program
--------------------------------------------
The National Vaccine Injury Compensation Program (VICP), established by the National Childhood Vaccine Injury Act of 1986, as amended, provides a mechanism through which compensation can be paid on behalf of a person determined to have been injured or to have died as a result of receiving a vaccine covered by VICP. The Vaccine Injury Table (<https://www.hrsa.gov/vaccinecompensation/vaccineinjurytable.pdf>) lists the vaccines covered by VICP and the associated injuries and conditions (including death) that might receive a legal presumption of causation. If the injury or condition is not on the Table, or does not occur within the specified time period on the Table, persons must prove that the vaccine caused the injury or condition. Eligibility for compensation is not affected by whether a covered vaccine is used off-label or inconsistently with recommendations.
To be eligible for compensation under VICP, a claim must be filed within 3 years after the first symptom of the vaccine injury. Death claims must be filed within 2 years of the vaccine-related death and not more than 4 years after the start of the first symptom of the vaccine-related injury from which the death occurred. When a new vaccine is covered by VICP or when a new injury/condition is added to the Table, claims that do not meet the general filing guidelines must be filed within 2 years from the date the vaccine or injury/condition is added to the Table for injuries or deaths that occurred up to 8 years before the Table change ([@R80]). Persons of all ages who receive a VICP-covered vaccine might be eligible to file a claim. Additional information is available at <https://www.hrsa.gov/vaccinecompensation> or by calling 1-800-338-2382.
Additional Resources
--------------------
### ACIP Statements
- General Best Practice Guidelines for Immunization: Best Practices Guidance of the Advisory Committee on Immunization Practices (ACIP). <https://www.cdc.gov/vaccines/hcp/acip-recs/general-recs/index.html>
- Immunization of Healthcare Personnel: Recommendations of the Advisory Committee on Immunization Practices (ACIP), 2011. MMWR Recomm Rep 2011;60(No. RR-7). <https://www.cdc.gov/mmwr/preview/mmwrhtml/rr6007a1.htm>
- Recommended Immunization Schedules for Adults, 2017: <https://www.cdc.gov/vaccines/schedules/hcp/adult.html>
- Recommended Immunization Schedule for Children and Adolescents Aged 18 years or Younger, 2017: <https://www.cdc.gov/vaccines/schedules/hcp/child-adolescent.html>
### Vaccine Information Sheets (VISs)
- VIS for IIV and RIV: <https://www.cdc.gov/vaccines/hcp/vis/vis-statements/flu.pdf>
- VIS for LAIV: <https://www.cdc.gov/vaccines/hcp/vis/vis-statements/flulive.pdf>
### Influenza Vaccine Package Inserts
- Trivalent Vaccines: <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm094045.htm>
- Quadrivalent Vaccines: <https://www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm295057.htm>
### CDC Influenza Antiviral Guidance
- Antiviral Drugs: Information for Healthcare Professionals: <https://www.cdc.gov/flu/professionals/antivirals/index.htm>
### American Academy of Pediatrics (AAP) Guidance
- AAP Recommendations for Prevention and Control of Influenza in Children: <https://redbook.solutions.aap.org/ss/influenza-resources.aspx>
### Infectious Diseases Society of America (IDSA) Guidance
- 2013 IDSA Clinical Practice Guideline for Vaccination of the Immunocompromised Host: [https://academic.oup.com/cid/article/58/3/e44/336537/2013--IDSA-Clinical-Practice-Guideline-for](https://academic.oup.com/cid/article/58/3/e44/336537/2013-IDSA-Clinical-Practice-Guideline-for)[^1]
ACIP Members, July 1, 2016--June 30, 2017
=========================================
**Chair:** Nancy Bennett, MD, University of Rochester School of Medicine and Dentistry, Rochester, New York.
**Executive Secretary:** Amanda Cohn, MD, National Center for Immunization and Respiratory Diseases, CDC, Atlanta, Georgia.
**Members:** Robert L. Atmar, MD, Baylor College of Medicine, Houston, Texas; Edward Belongia, MD, Marshfield Clinic Research Foundation, Marshfield, Wisconsin; Echezona Ezeanolue, MD, University of Nevada, Las Vegas, Nevada; Paul Hunter, MD, University of Wisconsin School of Medicine and Public Health, Milwaukee, Wisconsin; Allison Kempe, MD, University of Colorado School of Medicine, Denver, Colorado; Grace M. Lee, MD, Harvard Pilgrim Health Care Institute and Harvard Medical School, Boston, Massachusetts; Kelly Moore, MD, Vanderbilt University School of Medicine, Nashville, Tennessee; Cynthia Pellegrini, March of Dimes, District of Columbia; Arthur L. Reingold, MD, University of California, Berkeley, Berkeley, California; Laura E. Riley, MD, Harvard Medical School, Boston, Massachusetts; José R. Romero, MD, University of Arkansas for Medical Sciences, Little Rock, Arkansas; David Stephens, MD, Emory University School of Medicine, Atlanta, Georgia; Peter Szilagyi, MD, University of California, Los Angeles, Los Angeles, California; Emmanuel (Chip) Walter, MD, Duke University School of Medicine, Durham, North Carolina.
**Ex Officio Members:** Centers for Medicare and Medicaid Services, Mary Beth Hance, Baltimore, Maryland; Department of Defense, Eric Deussing, MD, Atlanta, Georgia; Department of Veterans Affairs, Jane A. Kim, MD, Durham, North Carolina; Food and Drug Administration, Wellington Sun, MD, Rockville, Maryland; Health Resources and Services Administration, Narayan Nair, MD, Rockville, Maryland; Indian Health Service, Amy Groom, MPH, Albuquerque, New Mexico; National Vaccine Program Office, Bruce Gellin, MD, District of Columbia; National Institutes of Health, Richard L. Gorman, MD, Bethesda, Maryland.
**Liaison Representatives:** American Academy of Family Physicians (AAFP), Margot Savoy, MD, Wilmington, Delaware; American Academy of Pediatrics (AAP), Committee on Infectious Diseases (COID), Carrie L. Byington, MD, Salt Lake City, Utah; American Academy of Pediatrics (AAP); Red Book Editor, David Kimberlin, MD, Birmingham, Alabama; American Academy of Physician Assistants (AAPA), Marie-Michèle Léger, MPH, Alexandria, Virginia; American College Health Association (ACHA), Susan Even, MD, Columbia, Missouri; American College of Nurse Midwives (ACNM), Carol E. Hayes, MN, MPH, Atlanta, Georgia; American College of Nurse Midwives (ACNM), (alternate) Pamela M. Meharry, PhD, Middletown, Connecticut; American College of Obstetricians and Gynecologists (ACOG), Kevin A. Ault, MD, Kansas City, Kansas; American College of Physicians (ACP), Sandra Adamson Fryhofer, MD, Atlanta, Georgia; American College of Physicians (ACP) (alternate), Gregory A. Poland, MD, Rochester, Minnesota; American Geriatrics Society (AGS), Kenneth Schmader, MD, Durham, North Carolina; America's Health Insurance Plans (AHIP), Mark J. Netoskie, MD, MBA, Houston, Texas; American Medical Association (AMA), Sandra Adamson Fryhofer, MD, Atlanta, Georgia; American Nurses Association (ANA), Charles (Chad) Rittle, DNP, Pittsburgh, Pennsylvania; American Osteopathic Association (AOA), Stanley E. Grogg, DO, Tulsa, Oklahoma; American Pharmacists Association (APhA), Stephan L. Foster, PharmD, Memphis, Tennessee; Association of Immunization Managers (AIM), Christine Finley, MPH, Burlington, Vermont; Association for Prevention Teaching and Research (APTR), W. Paul McKinney, MD, Louisville, Kentucky; Association of State and Territorial Health Officials (ASTHO); Terry L. Dwelle, MD, Bismarck, North Dakota; Biotechnology Industry Organization (BIO) Phyllis A. Arthur, MBA, District of Columbia; Council of State and Territorial Epidemiologists (CSTE), Christine Hahn, MD, Boise, Idaho; Canadian National Advisory Committee on Immunization (NACI), Ian MacDonald Gemmill, MD, Kingston, Ontario, Canada; Infectious Diseases Society of America (IDSA), Kathleen M. Neuzil, MD, Baltimore, Maryland; Infectious Diseases Society of America (IDSA) (alternate), Carol J. Baker, MD, Houston, Texas; National Association of County and City Health Officials (NACCHO), Matthew Zahn, MD, Santa Ana, California; National Association of County and City Health Officials (NACCHO), (alternate) Jeffrey Duchin, MD, Seattle, Washington; National Association of Pediatric Nurse Practitioners (NAPNAP), Patricia A. Stinchfield, MS, St. Paul, Minnesota; National Foundation for Infectious Diseases (NFID), William Schaffner, MD, Nashville, Tennessee; National Immunization Council and Child Health Program, Mexico, Ignacio Villaseñor Ruiz, MD, Mexico City, Federal District, Mexico; National Medical Association (NMA), Patricia Whitley-Williams, MD, New Brunswick, New Jersey; National Vaccine Advisory Committee (NVAC), Kimberly Thompson, ScD, Orlando, Florida; Pediatric Infectious Diseases Society (PIDS), Sean O'Leary, MD, Denver, Colorado; Pediatric Infectious Diseases Society (PIDS) (alternate), Mark H. Sawyer, MD, San Diego, California; Pharmaceutical Research and Manufacturers of America (PhRMA); David R. Johnson, MD, Swiftwater, Pennsylvania; Society for Adolescent Health and Medicine (SAHM), Amy B. Middleman, MD, Oklahoma City, Oklahoma; Society for Healthcare Epidemiology of America (SHEA), David Weber, MD, Chapel Hill, North Carolina.
ACIP Influenza Vaccine Work Group
=================================
**Chair:** Emmanuel (Chip) Walter, MD, Durham, North Carolina.
**Members:** Kevin Ault, MD, Kansas, City, Kansas; Edward Belongia, MD, Marshfield, Wisconsin; Henry Bernstein, DO, Hempstead, New York; Karin Bok, MS, Bethesda, Maryland; PhD; Jeff Duchin, MD, Seattle, Washington; David Cho, PhD, Silver Spring, Maryland; Janet Englund, MD, Seattle, Washington; Sandra Fryhofer, MD, Atlanta, Georgia; Bruce Gellin, MD, District of Columbia; Ruth Karron, MD, Baltimore, Maryland; Wendy Keitel, MD, Houston, Texas; Lucia Lee, MD, Silver Spring, Maryland; Marie‐Michèle Léger, MPH, Alexandria, Virginia; Susan Lett, MD, Jamaica Plain, Massachusetts; Jamie Loehr, MD, Ithaca, New York; Flor M. Munoz, MD, Houston, Texas; Kathleen M. Neuzil, MD, Baltimore, Maryland; Cynthia Nolletti, MD, Silver Spring, Maryland; Roshan Ramanathan, MD, Silver Spring, Maryland; Chris Roberts, PhD, Rockville, Maryland; William Schaffner, MD, Nashville, Tennessee; Robert Schechter, MD, Richmond, California; Kenneth Schmader, MD, Durham, North Carolina; Tamara Sheffield, MD, Salt Lake City, Utah; Patricia Stinchfield, MS, St. Paul, Minnesota; Rob Stirling, MD, MSc, MHSc, Toronto, Ontario, Canada; Peter Szilagyi, MD, Los Angeles, California; Wendy Vaudry, MD, Edmonton, Alberta, Canada; Matthew Zahn, MD, Santa Ana, California.
[^1]: A list of Work Group members may be found on page 20 of this report.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
A correct interpretation of the direction of evolution in basal parts of the mammalian tree has important implications for different aspects of biology and also for medicine (e.g., the selection of appropriate model organisms). However, some basal relationships may still need further examination before being considered as conclusively and finally settled. Paleontological data show a sudden radiation of mammals in the late Cretaceous \[[@pcbi-0030002-b001]\]. Molecular data might resolve the succession of the early diversification events of placental mammals, but molecular analyses in general suggest an earlier timeframe \[[@pcbi-0030002-b002],[@pcbi-0030002-b003]\]. In particular, the phylogenetic positions of rodents, primates, and carnivores are still contentious, with traditional morphology supporting a primate--rodent clade \[[@pcbi-0030002-b004]\] (called Supraprimates or Euarchontoglires) and molecular studies showing support for either a primate--rodent clade \[[@pcbi-0030002-b005]--[@pcbi-0030002-b009]\] or a primate--carnivore clade \[[@pcbi-0030002-b010]--[@pcbi-0030002-b012]\]. The results of Jorgensen et al. \[[@pcbi-0030002-b013]\] support a rodent outgroup to a primate--artiodactyl clade based on full genome analyses. Lin et al. \[[@pcbi-0030002-b014]\] report a primate--rodent clade but only after constraining the rodents to be strictly monophyletic. Mitogenomic studies almost invariably support the primate--carnivore clade including that of Janke et al. \[[@pcbi-0030002-b015]\], who presented the first marsupial rooting of the eutherian tree. This topology has also been confirmed in subsequent studies using mixed data \[[@pcbi-0030002-b016],[@pcbi-0030002-b017]\]. The molecular studies differed in the type (nuclear, mitochondrial, or both) and in the amount of genomic data (more species versus more genes) as well as in the tree reconstruction methods.
The inconsistency among these results underlines the difficulty in resolving the three-taxon relationship involving rodents, primates, and carnivores. The short branches separating these groups reside deep within the mammalian phylogenetic tree, thereby enhancing the effects of any reconstruction artifacts. These can be related to data quality or any failure to accurately model particular aspects of evolution such as parallel evolution, lineage specific mutation rates, or other changes in the evolutionary process \[[@pcbi-0030002-b014]\].
Long branch attraction (LBA) may occur when an ingroup has a faster rate of evolution, thereby promoting migration of the long branch with accelerated evolution toward the long branch of the outgroup. This phenomenon was first examined by Felsenstein \[[@pcbi-0030002-b018]\], who showed that trees with long branches could be positively misleading when reconstructed under the parsimony criterium. Parsimony, which computes the minimum number of evolutionary steps required to explain the observed sequences, however, does not have the properties of distance and is not additive. Additive means that for a lineage *A* → *B* → C, the equation *d~AB~* + *d~BC~* = *d~AC~* is satisfied in the expected value. In case any particular taxon (e.g., mouse \[[@pcbi-0030002-b019]\]) evolves faster than the other two in our three-taxon analysis, LBA could possibly affect the outcome for parsimony or nonadditive distance measures. Additive distance estimates such as those produced by the Markov model of evolution used in this study should not be affected by LBA. However, systematic biases such as those produced by parallel evolution or other deviations from the model can affect any evolutionary distance measure.
Parallel morphological or molecular evolution can occur when two species develop similar characteristics because of adaptation to similar environments or life strategies. A coupling between molecular and morphological evolution among mammals is highly speculative, however. To counter any systematic biases, we have made the precaution of using different phylogenetic methods based on different evolutionary phenomena because it is unlikely that all methods will be affected by systematic biases in the same way.
Taxon sampling may affect the accuracy of phylogenetic reconstruction \[[@pcbi-0030002-b020]--[@pcbi-0030002-b023]\]. Some authors argue that increasing the number of characters sampled per taxon improves the accuracy, while others state that accuracy is better improved by subdividing long branches by including more taxa, resulting in fewer characters overall. In any case, the choice of more characters versus more taxa depends on the phylogeny under consideration. In the problem under investigation here, a very short branch separates two possible phylogenies, and the comparison is between the number of mutations that occurred on the short common branch and the number of homoplasies that occur on the longer branches. With increased character sampling, we increase the chance of detecting the relatively few changes that occur on the common branch. In certain cases, increasing the number of taxa is useful to divide long branches to help to identify homoplasies \[[@pcbi-0030002-b024]\]. Therefore, in an extended analysis, we also included all available mammalian genomes.
To combat the problems inherent in elucidating this difficult topology, we used a wide range of methods that measure different aspects of molecular evolution, with the view that it is very unlikely that a specific change in the evolutionary process (e.g., different DNA repair mechanisms in murid rodents \[[@pcbi-0030002-b014]\]) would affect *all* of the measures used in this study.
Results/Discussion {#s2}
==================
Phylogenetic Analysis {#s2a}
---------------------
The evolutionary relationship of dog (*Canis familiaris,* order Carnivora), mouse (*Mus musculus,* order Rodentia), and human (*Homo sapiens,* order Primates) was examined applying a wide variety of distance measures and using the opossum (Monodelphis domestica) as the outgroup because, as a marsupial, it is the closest relative to the eutherian dataset. Thus, solving the problem of the phylogeny of the mouse, human, and dog requires finding the root of the tree. This can be achieved by placing the outgroup on one of the three branches leading to the orders. [Figure 1](#pcbi-0030002-g001){ref-type="fig"} shows the three possible positions of the outgroup as well as the resulting rooted trees for the opossum and the three eutheria. We endeavored to answer the question of which of these three hypotheses---a primate--carnivore clade, a primate--rodent clade, or a rodent--carnivore clade---represents the true phylogeny.
{#pcbi-0030002-g001}
### Measures of genomic distance. {#s2a1}
To counteract the possible influence of reconstruction artifacts, we applied a collection of methods for tree reconstruction and included a collection of additional complete mammalian genomes. Distance trees were constructed based on a variety of genomic distance measures, parsimony trees were evaluated with and without gapped positions, and likelihood trees were computed on multiple sequence alignment (MSA) columns containing no gaps.
The three types of genomic distance measures used---amino acid replacement, synonymous nucleotide substitutions, and gene reordering---measure different aspects of molecular evolution. All measures correlate with time but are measuring independent properties in that a distance of 0 in one measure does not necessarily affect the distance computed from another measure. For example, synonymous substitutions are independent of amino acid replacement while both are independent from gene reordering. [Table 1](#pcbi-0030002-t001){ref-type="table"} and the following subsections summarize and discuss the results from the various methods used.
######
Summary of the Methods Used and Their Results
### Distance trees built using mean PAM, CodonPAM, and SynPAM between orthologs. {#s2a2}
Four methods were employed to measure the evolutionary change between two proteins or their coding DNA sequence: PAM, CodonPAM, SynPAM, and dS. PAM distance is a long-used measure of evolutionary distance of protein sequences, which estimates the distance using empirically determined amino acid mutation matrices \[[@pcbi-0030002-b025]\]. SynPAM \[[@pcbi-0030002-b026]\] and CodonPAM \[[@pcbi-0030002-b027]\] are recent extensions of the empirical model to coding DNA. Both methods are based on a 64 × 64 mutation matrix describing substitution probabilities between any two codons. The SynPAM method estimates the distance only on positions with conserved amino acids, while CodonPAM uses all aligned codons. Therefore, CodonPAM combines the synonymous substitutions with amino acid replacing changes and has been shown to improve the accuracy of distance estimates \[[@pcbi-0030002-b028]\].
Synonymous mutations in coding DNA do not alter the encoded protein. Thus, they are under no strong selection pressure and are less constrained by functional changes. Because of these properties they are particularly robust against the effects of parallel evolution. Therefore we employed a second method measuring synonymous changes, the number of synonymous substitutions per synonymous site, dS.
Distance trees using PAM, CodonPAM, SynPAM, and dS distances were created from the complete set of orthologous groups from eight mammals with completely sequenced genomes. Adding more in-group genomes reduces the number of complete groups of orthologous sequences, but adds more intergenome distances, making the trees more robust and reducing possible biases from particular genomes. All of the trees constructed using any of these four distance measures supported the primate--carnivore clade as shown in [Figure 2](#pcbi-0030002-g002){ref-type="fig"}A. To assess the reliability of the resulting trees, bootstrapping was performed by sampling orthologous groups. Bootstrapping is an empirical way of estimating the variance of a result without knowledge of the underlying distribution. Using very large amounts of data, as in this study, leads to a very low variance, and therefore the results from the bootstrapping are always 100%. For this reason, they are not reported in [Figure 2](#pcbi-0030002-g002){ref-type="fig"}. The fit of the distances to each of the three topologies in [Figure 1](#pcbi-0030002-g001){ref-type="fig"} was computed for each method via least squares, and the residuals (normalized by degrees of freedom) are reported in [Table 1](#pcbi-0030002-t001){ref-type="table"}.
{#pcbi-0030002-g002}
### Distance trees obtained using reversal distances. {#s2a3}
Several mechanisms can alter the ordering of genes on and across chromosomes. As changes of this kind accumulate over time and eventually become fixed in a population, the number of such changes between two genomes can serve as an evolutionary distance. One of the simplest operations to model genome reordering is an inversion or reversal, where a part of the DNA is removed from the strand and reinserted in the reverse direction. The minimal number of such inversion operations that are needed to transform the gene order in one genome to the order in another genome is called the reversal distance. Both signed and unsigned reversal distance (if the direction of the gene is considered or not) were used to compute distance trees for human, chimpanzee, dog, mouse, rat, and chicken. The percentage of inversions observed for the most distant pair (chicken versus rat), were 27% for the signed and 23% for the unsigned versions of the algorithm. Tests on simulated data revealed that for these distances the exact number of reversals was found more than 99% of the time (unpublished data). This means that for this distance range, multiple inversions almost never occur in a way that could be explained by fewer reversals, which would cause an underestimation of the distance. The tree obtained by this method, again supporting a rodent outgroup to primates/carnivores, is shown in [Figure 2](#pcbi-0030002-g002){ref-type="fig"}B. The normalized residuals from the least squares fitting for each topology are given in [Table 1](#pcbi-0030002-t001){ref-type="table"}.
Parsimony Analysis of Characters from Multiple Sequence Alignments {#s2b}
------------------------------------------------------------------
MSAs for parsimony analysis were created for the three eutheria plus opossum with aligned amino acid positions considered as characters. The numbers of positions supporting each topology were counted using either all informative positions or excluding positions with gaps and are summarized in [Table 1](#pcbi-0030002-t001){ref-type="table"}. Because sequence evolution is a stochastic process and the branch separating the conflicting hypotheses is very short, the absolute difference between the numbers of characters supporting each hypothesis can be small. However, with the large number of characters available, the variance of the estimation is small compared with the large numbers of supporting positions. In [Table 2](#pcbi-0030002-t002){ref-type="table"} the significance of the number of positions supporting the primate--carnivore clade compared with the primate--rodent clade are reported in standard deviations and are very significant.
######
Number of MSA Positions Supporting the Three Topologies under the Parsimony Criterium and the Significance of the Difference as a Function of the Percentage of Allowable Gaps
Some genes from the complete genomes are only fragments, and this may result in the MSAs being of reduced quality. Therefore, in a second analysis, alignments with a high frequency of gapped positions were excluded. The results of analysis with and without gaps are shown in [Table 2](#pcbi-0030002-t002){ref-type="table"}, in which the numbers of characters supporting each topology are shown as a function of the allowable percentage of gapped positions in the alignment. It is interesting that for the analysis including gapped positions (columns 4--7 in [Table 2](#pcbi-0030002-t002){ref-type="table"}), as the percentage of gaps allowed in the groups increases from 1% to 10%, the support for the mouse outgroup hypothesis becomes stronger because of the increase in data occurring with the addition of groups. As the allowed percentage of gaps exceeds 10%, the quality of the data deteriorates and the significance of the support for the mouse outgroup hypothesis decreases. When all positions containing gaps are excluded (columns 8--11 in [Table 2](#pcbi-0030002-t002){ref-type="table"}), the support for the primate--carnivore clade continually increases with increasing amounts of data.
In an extended analysis, the groups of genes were broken into thirds---short, medium, and fast-evolving---based on the sum of all pairwise distances. The results for those with ≤5% gapped positions are reported in [Table 2](#pcbi-0030002-t002){ref-type="table"}. Each of the thirds supports the same primate--carnivore grouping. The decrease in significance with decreasing evolutionary distance can be attributed to the decreasing amount of informative positions as the sequences become more similar.
To assess the influence of the choice of the outgroup, all genes represented by the three eutheria and the chicken and the opossum and containing at most 5% gaps were analyzed. Both outgroups support a primate--carnivore clade, although the significance decreases when the opossum is used as the outgroup.
Likelihood Analysis {#s2c}
-------------------
The same MSAs as for the parsimony analyses were used for a likelihood analysis of quartets using either the chicken or the opossum as an outgroup. First, all genes in the orthologous groups were concatenated to form one supergene. The log-likelihoods of the data given each of the three topologies in [Figure 1](#pcbi-0030002-g001){ref-type="fig"} were computed and are shown in [Table 1](#pcbi-0030002-t001){ref-type="table"}. For both outgroups, the likelihood is orders of magnitude greater for the topology supporting a primate--carnivore clade than for the alternatives. Because different orthologous groups were included for the analysis of each outgroup, the likelihoods between the outgroups are not comparable. A second analysis was performed by taking all orthologous groups containing the opossum, creating a gene tree for the four sequences in each group, and computing the likelihood of the data for each topology. The number of gene trees supporting each topology was counted, resulting in a clear majority supporting the primate--carnivore clade.
Conclusions {#s2d}
-----------
The analysis of the three-taxon relationship (mouse/human/dog) based on data from complete nuclear genomes strongly suggested a sister relationship between human (primates) and dog (carnivores) to the exclusion of mouse (rodents). The limited length of the branch separating the topologies may make analysis of the tree sensitive to the choice of evolutionary model and data. Therefore, the analyses were conducted applying a number of independent methods to the genome-sized datasets, all of which supported this relationship. The effects of adding more taxa versus sampling more characters were also investigated. Inclusion of additional genomes (rat, chimp, macaque, cow) did not change the topology of the tree. However, sampling many characters yielded very significant support for the short internal branch. Therefore, we suggest that this difficult phylogenetic problem can only be solved using thousands of genes, which are only available from complete genomes. When the critical branch is so small, the use of a large number of genes is the only way to reduce the variance enough to get statistically significant results.
Materials and Methods {#s3}
=====================
Data and implementation. {#s3a}
------------------------
All analyses were performed on fully sequenced genomes from human, dog, mouse, and opossum. As an extension, up to four other mammalian genomes were included in some analyses as was the complete genome of the chicken, which was used as an alternative outgroup. The genomes downloaded from ENSEMBL \[[@pcbi-0030002-b029]\] had version numbers NCBI 35 for human \[[@pcbi-0030002-b030]\], BROADD1 for dog, NCBI m36 for mouse \[[@pcbi-0030002-b019]\] and 0.5 for opossum. The other genome databases (Bos taurus (Btau_2.0), Gallus gallus (WASHUC1), Pan troglodytes (CHIMP1), Rattus norvegicus (RGSC3.4), and Macaca mulatta (Mmul 0.1) ) are also from ENSEMBL. All databases were converted to the *Darwin* database format for further computations. The implementation of the methods was entirely done in the *Darwin* programming language \[[@pcbi-0030002-b031]\] with the exception of the computation of dS.
Orthologs. {#s3b}
----------
Groups of orthologous proteins from the orthologous matrix (OMA) project \[[@pcbi-0030002-b032]\] constituted the basis for building the trees. The OMA project is a large-scale effort to identify groups of orthologous sequences in a fully automated manner. This is achieved by computing all-against-all sequence alignments between all sequenced genomes from all kingdoms of life (297 genomes completed at the time of writing). The OMA groups are conservative in that a careful search for possible paralogy discards all suspicious matches. Every candidate pair of sequences is verified against all other genomes to identify gene loss that could lead to inclusion of paralogs. The orthology assignments are done without assuming an underlying species tree, which would cause a bias for the inference of a phylogeny. In the latest OMA release, we find 11,022 groups having a representative in each of the four primary species under study.
Phylogenetic methods. {#s3c}
---------------------
Distance trees were calculated using the PhylogeneticTree package in *Darwin*. For distance trees, all pairwise distances and variances are estimated, and a tree is sought that best approximates the distance information via weighted least squares. Finding the best-fitting tree is an NP-complete problem \[[@pcbi-0030002-b033]\]. The polyalgorithm implemented in *Darwin* solves the problem in the following way: one neighbor-joining tree and 29 trees with random topologies are created as starting points. All trees are then optimized using branch-swapping heuristics, and the best-fitting tree is retained. When considering only four or five species, the exact computation of the best tree could be done manually (three or 15 topologies to analyze, respectively). For relatively small problems such as those encountered here (at most eight leaves), the algorithm almost certainly finds the optimal topology. This was verified by simulation studies (unpublished data).
Evolutionary distance. {#s3d}
----------------------
The PAM, CodonPAM, and SynPAM methods were implemented in *Darwin*. Proteins and coding sequences were aligned using dynamic programming \[[@pcbi-0030002-b034],[@pcbi-0030002-b035]\], and the distances were estimated by maximum likelihood. The likelihood method implemented in the *codeml* program from the *PAML* software package \[[@pcbi-0030002-b036]\] was used to compute dS from pairwise alignments of coding DNA.
Only complete groups of orthologs (groups with one member in all genomes under consideration) were used to compute the average distances between two species. The distances from all pairwise alignments of orthologous sequences are averaged for each pair of species, resulting in a distance and a variance matrix from which trees are built. For DNA-based methods, groups had to be excluded when the coding DNA for at least one of the member proteins was not or was only partially available. Also excluded were alignments with fewer than 100 synonymous positions because distance estimates based on short alignments suffer from statistical biases. Additionally, groups were excluded from the SynPAM analysis if at least one distance estimate was higher than 1,000 (corresponding to approximately ten substitutions per synonymous codon). For the dS analysis, groups with one value higher than ten were discarded. Such high values indicate that the synonymous substitutions between two genes are saturated and thus have a very high variance.
Gene reordering (reversal distance). {#s3e}
------------------------------------
The gene orders of two genomes can be formulated as a permutation of a list of integers, referring to the order of orthologous genes in the genomes. As an example, we consider two genomes A and B with only three genes, where the first gene in A is orthologous to the second gene in B, and vice versa, while the third genes in both genomes are orthologous. Stated as a permutation, the genes \[[@pcbi-0030002-b001], [@pcbi-0030002-b002], [@pcbi-0030002-b003]\] in A are transformed to \[[@pcbi-0030002-b002], [@pcbi-0030002-b001], [@pcbi-0030002-b003]\] in B. A reversal operation inverts a subsequence of the gene order. Applying this operation to the first two genes will transform the gene order in A to the one in B. Therefore, the reversal distance between A and B is one. If the direction of the genes is known and used for the computation, this is called "signed" (because the numbers in the permutations are labeled with a minus sign if the genes are found on the complement strand of the DNA) and can be computed in linear time \[[@pcbi-0030002-b037],[@pcbi-0030002-b038]\]. If the direction of the gene is not known, it is called "unsigned." In this case, the problem of finding the optimal sequence of reversals is NP-hard \[[@pcbi-0030002-b039]\]. We implemented a k-greedy algorithm in *Darwin* to solve the unsigned problem.
Computing the reversal distance can only be done reliably if large stretches of the genome are assembled. Unfortunately, our version of the opossum genome was only assembled into scaffolds, but not yet into complete chromosomes. Therefore, we used the chicken as the outgroup because the genes are already placed on chromosomes. For the same reason, macaque and cow could not be used for this method. Because only a very small number of reversals decide the phylogeny in this study, we filtered the orthologous groups as much as possible to reduce noise. Groups were excluded from the analysis if at least one gene was placed on a scaffold instead of a chromosome or if two genes had overlapping coding regions.
Parsimony over multiple-sequence alignments. {#s3f}
--------------------------------------------
MSAs of orthologs from four species---human, mouse, dog, and opossum---were built using two methods available in *Darwin,* probabilistic and circular tours \[[@pcbi-0030002-b040]\]. In *Darwin,* MSAs can be improved using gap-rearrangement heuristics. All the MSAs are scored and the best-scoring one selected. 102 alignments were eliminated from the analysis for having fewer than 100 positions, leaving 10,920 groups. To eliminate the influence of gene fragments and misfound start and stop positions, the alignments were truncated at both ends---only characters between the first and last position containing no gap in any sequence were counted. To decide which of the three quartets is the most parsimonious one, only those alignment positions at which two species share one character and the other two share another character (2--2 cases) are informative. (Positions where all species share the same characters or have all four different, as well as 3--1 splits, are uninformative. The 2-1-1 splits have parsimony costs of 2 for all three topologies and are also uninformative. Thus, only the characters that have a 2--2 split are of interest.) The standard deviations separating the two topologies in [Table 2](#pcbi-0030002-t002){ref-type="table"} are computed under the assumption of a binomial distribution of the counts for the primate--carnivore clade (*n* ~1~) and the primate--rodent clade (*n* ~2~ **)** as:
Likelihood analysis. {#s3g}
--------------------
Likelihood trees were also implemented in *Darwin*. Since optimizing topology and branch lengths of likelihood trees is very time-consuming, only trees with four leaves (human, mouse, dog, and one of opossum and chicken) were computed. Likelihood trees were constructed for each group (MSA) individually and for the concatenated alignments. Positions containing a gap in any of the four sequences were completely ignored. Optimizing the lengths of the five branches for a given quartet topology was initialized with the branch length of the least squares tree, and then numerically improved first by steepest-ascent, and then by multidimensional Newton. The optimization of the likelihood for one topology over approximately 5 million characters was computed in about one hour on a desktop Linux machine.
We would like to acknowledge Úlfur Árnason, Svenja Bossard, Christophe Dessimoz, Eric Gaucher, Manuel Gil, Dan Graur, David Liberles, and Alexander Roth for their help.
**Competing interests.** The authors have declared that no competing interests exist.
A previous version of this article appeared as an Early Online Release on November 20, 2006 (doi:[10.1371/journal.pcbi.0030002.eor](10.1371/journal.pcbi.0030002.eor)).
**Author contributions.** GC, AS, and GG conceived and designed the experiments, performed the experiments, analyzed the data, and contributed analysis tools. GC and AS wrote the paper.
**Funding.** The authors received no specific funding for this study.
LBA
: long branch attraction
MSA
: multiple sequence alignment
OMA
: orthologous matrix
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Almost half of the patients suffering from a major depressive episode (MDE) fail to achieve a response, despite sequential combination or augmentation treatment strategies, irrespective of the operational definition adopted for treatment-resistant depression (TRD) ([@B1], [@B2]). The rates of treatment-resistant bipolar depression (TRBD) within the course of bipolar disorder (BD) can even exceed those documented for major depressive disorder (MDD) and present a differential pattern of response to different medications, indeed ([@B3]). Both TRD ([@B4]) and TRBD ([@B5], [@B6]) pose a significant clinical and societal burden, soliciting the assessment of putative predictors using clinical, neuropsychological, neuroimaging, and genetic measures ([@B7], [@B8]). However, the general failure to find sensitive and specific predictors of treatment response in MDD and BD, either current- of future-state markers ([@B9]), is in part, due to the heterogeneity of these conditions, which range from biologically determined to event-dependent conditions ([@B10]). The initial hype about digital phenotyping has been tempered by increasing awareness of the significant technical and analytical difficulties hindering translational change ([@B11], [@B12]) most likely due to differential neurobiology underpinning MDD ([@B13], [@B14]) and BD ([@B15]--[@B17]). To date, no single biological marker stood-out in selecting the best initial antidepressant treatment for a given individual, despite promising evidence coming from recent pharmacogenomics studies ([@B18]). Clinical risk stratification models for TRD, incorporating baseline sociodemographic and clinical features ([@B19]), preceded algorithm-based approaches proposed for TRBD ([@B20]). For MDD, a cluster of variables including marital status, insomnia, psychosocial impact, trauma, education, energy, disorder recurrence, comorbidity, race, and severity of current MDE resulted in a positive predictive value of TRD =.61 and a negative predictive value =.69, consistently across males and females and primary- vs. specialty care settings ([@B19]). Over the past decade, findings from the Group for the Study of Resistant Depression (GSRD), a multinational European research consortium, pointed toward strong effects of clinical variables on treatment outcome of MDD ([@B21]), as well as specific genetic signature ([@B22]). This is particularly relevant considering that, when considered separately, the potential predictors of TRD usually showed odd ratios (ORs) around 1.5, and thus, were not applicable as univocal hints in the clinical setting ([@B23], [@B24]). Nonetheless, four potent clinical predictors of TRD emerged in a validation study that showed a collective predictive accuracy of 87%. These latter include symptom severity, suicidal risk, comorbid anxiety disorder, and the lifetime number of MDE episodes ([@B25]--[@B27]). Overall, symptom severity, suicidal risk, higher number of lifetime of MDEs, and comorbid anxiety disorder were consistently replicated as the most prominent risk factors for different stages of TRD in large-sampled studies ([@B28]).
Machine learning approaches may, however, further increase the detection of TRD ([@B29], [@B30]), as reinforced by the 2019 report by the GSRD study group for TRD ([@B22]).
Concerning BD, clinical predictors of "full clinical recovery" include outpatient treatment and more extended periods of lifetime mania ([@B31]). In contrast, higher baseline severity scores, presence of MDEs in the previous year, poor social functioning predict failure to achieve remission or recovery in BD ([@B32]). Among other clinical hallmarks, earlier (hypo-)manic or mixed recurrence, including sub-threshold manic symptoms and the proportion of days spent with an elevated mood in the preceding year, predict TRBD ([@B33]). The conceptualization of TRBD is more articulated than one of TRD, due to the differential natural and pharmacological course of BD and the controversial role of the antidepressants in BD ([@B2]). While drug resistance or worsening in MDD has been associated with mixed features and other bipolarity hints ([@B34]), relatively little is known about the impact of mixed features in bipolar depression and subsequence failure to respond. However, side-by-side comparisons of unipolar vs. bipolar MDEs relying on a relatively limited-range of retrospective data, suggesting that TDR represents a distinct psychopathological condition and not necessarily a prodromal state of BD ([@B35]).
The aim of the present exploratory-hypothesis generating study is to assess multiple clinical predictors of treatment resistance comparing unipolar vs. bipolar patients across different clinical subtypes of current MDE. To the best of our knowledge, the present data-driven study is the first of its kind to a side-by-side comparison of a range of putative clinical predictors of treatment resistance across different clinical subtypes of depression, either unipolar or bipolar major depression.
Methods {#s2}
=======
Participants {#s2_1}
------------
A hundred and thirty-one outpatients, both sexes, 18--65 years old, consecutively enrolled in a non-interventional study at a tertiary care unit of the Department of Psychiatry of the Federico II University of Naples, Italy, between May 3, 2017, and June 17, 2019. All patients met the Diagnostic and Statistical Manual for Mental Disorders, Fifth Edition (DSM-5) ([@B36]) criteria for a current MDE, either in the course of MDD or BD. Six experienced psychiatrists (MF, AF, SN, AA, ABl, PM) made the clinical diagnoses. Eligible patients signed a valid written informed consent after the study procedures had fully explained by the appointed principal investigator (MF). The protocol of the study was approved by the local I.R.B. on May 2^nd^, 2017, according to the 1964 Declaration of Helsinki and its subsequent amendments ([@B37]). Exclusion criteria were the lack of valid informed consent, untreated thyroid disease, schizophrenia, severe brain syndrome, or other organic conditions, potentially hampering the validity of the consent.
Evaluation Procedure and Operational Definitions {#s2_2}
------------------------------------------------
The participating outpatients were rated according to the Hamilton Rating Scale for Depression, 17-item version scale (HAM-D-17) ([@B38]), Young Mania Rating Scale (YMRS) ([@B39]), and Temperament Evaluation of Memphis, Pisa, Paris, and San Diego, 110-item version (TEMPS-110) ([@B40]). The following clinical variables were recorded: age, sex, duration of the untreated illness (DUI), family history for header MDD or BD, lifetime suicidal behavior, significant psychiatric and medical comorbidities, and essential current and lifetime psychiatric treatment. The DUI was defined as the "time interval between the first lifetime onset of symptoms of depression, not necessarily reaching the diagnostic threshed set by the DSM-5: at least five symptoms of depression, and the first adequate treatment" ([@B41]). TRD was defined as "failure of two or more consecutive trials with antidepressants" ([@B2]). TRBD was defined as a lack of response to treatment to at least two previously established treatments for bipolar depression (e.g., olanzapine/fluoxetine combination, brexpiprazole, cariprazine, quetiapine, aripiprazole). TRBD could, however, have been exposed to at least one mood stabilizer drug plus an antidepressant in case of the previous failure to "adequate" trials for acute bipolar depression, especially whenever such therapy had been started elsewhere before seeking our consultation or in the presence of significant comorbidities (e.g., anxiety or obsessive--compulsive disorders).
Finally, the mixed features of the current MDE reflected the DSM-5 specifier. Besides, the "overlapping" symptoms otherwise excluded by the DSM-5 for a DSM-5-defined MDE with mixed features, i.e., "irritability," "distractibility," "impulsivity," "mood lability," and "psychomotor agitation," were nonetheless recorded given their clinical relevance ([@B42]).
Statistical Analysis {#s2_3}
--------------------
The statistical analysis was performed using IBM™ SPSS™ Statistics™ v.25.0.0 for Windows™ ([@B43]), adopting alpha.05, two-tailed. Non-parametric distributions were transformed as necessary. The interrater reliability for the analysis using Cohen\'s k statistics was ascertained for consistency issues. Descriptive statistics followed by principal component analysis (PCA) and by two-step cluster analysis were carried-out for MDD and BD patients, separately. Patients recorded as TRD or TRBD were considered as "cases"; those without a history of treatment resistance were deemed as "controls." Oblique rotation solution (Promax) was performed by calculating the coefficient for each pair of variables, and multi-collinearity was excluded; coefficient ≤.3 were suppressed. We assumed the conservative loadings \>.3 to be significant. The Kaiser--Meyer--Olkin measure of sampling adequacy resulted in.65, indicating an acceptable sample size, with Bartlett\'s test of sphericity scoring lower than.05. To determinate the component number, we considered only eigenvalues \> 1. For readers\' convenience, an eigenvalue of 1 would explain as much variance as the one explained by a single variable.
In plain, the PCA aimed to extract the psychopathological components captured by the HAM-D scale, while the cluster analysis carried aimed at discriminate the cases at study based on the PCA-fetched components. The generated clusters of depressions were compared against each other using descriptive statistics and binary logistic regression analysis.
Results {#s3}
=======
Demographic and Clinical Variables Across Different Clinical Subtypes of Depression {#s3_1}
-----------------------------------------------------------------------------------
A summary of the essential clinical and demographic features of the included sample has been reported in [**Table 1**](#T1){ref-type="table"}. Essential current and lifetime pharmacological history are presented in [**Table 2**](#T2){ref-type="table"}, while [**Table 3**](#T3){ref-type="table"} provides the core rating tools and predominant affective temperament. Among other clinical variables, mixed, anxious, and psychotic features of current MDE were statistically more frequent among BD patients compared to the MDD counterpart.
######
Essential clinical and demographics across unipolar and bipolar cases.
Study subjects (n = 131) MDD (n = 78) BD (n = 53, of whom BD-I = 24; BD-II = 29) *t* or χ^2^ (df) *p*
------------------------------------------------------------------ --------------------- -------------------------------------------- ------------------ ------------
Essential demographics
Age, (mean ± sd) 49.41 ± 13.90 50.30 ± 12.14 .379 (129) .705
Sex F/M, n (%) 49 (37.4)/29 (22.1) 35 (26.7)/18 (13.7) .142 (1) .706
Ethnicity (Caucasian, other), n (%) 78 (59.5) 53 (40.5) -- --
Primary school, n (%) 14 (10.7) 11 (8.4) 4.948 (3) .176
Secondary school, n (%) 29 (22.1) 13 (9.9)
High school, n (%) 29 (22.1) 19 (14.5)
College, university, or higher education, n (%) 6 (4.6) 10 (7.6)
Unmarried, n (%) 14 (10.7) 15 (11.5) 13.462 (2) **.001**
Married/cohabiting, n (%) 61 (46.6) 27 (20.6)
Widow/divorced, n (%) 3 (2.3) 11 (8.4)
Essential lifetime comorbidities and clinical features
Lifetime substance use disorder, n (%) 4 (3.1) 6 (4.6) 1.716 (1) .190
Lifetime panic disorder, n (%) 28 (21.4) 17 (13) .204 (1) .651
Lifetime generalized anxiety disorder, n (%) 58 (44.3) 33 (25.2) 2.176 (1) .140
Lifetime anorexia nervosa, n (%) 15 (11.5) 6 (4.6) 1.467 (1) .226
Lifetime bulimia nervosa, n (%) 5 (3.8%) 4 (3.1) .064 (1) .801
Lifetime binge eating disorder, n (%) 8 (6.1%) 10 (7.6) 1.974 (1) .160
Lifetime obsessive--compulsive disorder, n (%) 13 (9.9) 9 (6.9) .002 (1) .962
Lifetime specific phobia, n (%) 17 (13) 7 (5.3) 1.555 (1) .212
Lifetime impulse control disorder, n (%) 5 (3.8) 20 (15.3) 20.054 (1) **\<.001**
Borderline personality disorder DSM-5, n (%) 3 (2.3) 5 (3.8) 1.718 (1) 0.190
History of previous psychiatric hospitalization, n (%) 25 (19.1) 29 (22.1) 6.691 (1) **0.01**
Lifetime history suicidal behavior, n (%) 28 (21.4) 23 (17.6) .746 (1) .388
Duration of untreated illness, in years, (mean ± sd) 3.14 (3.77) 2.81 (5.61) 2.412 (129) .81
Duration of current MDE, in weeks, (mean ± sd) 24.96 (40.34) 40.13 (84.06) 1.301 (114) .196
Age at onset of depression (self-report), in years, (mean ± sd) 34.09 (14.32) 32.85 (14.49) −.483 (129) .630
Essential specifies
DSM-5 MDE with melancholic features, n (%) 11 (8.4) 4 (3.1) 1.337 (1) .247
DSM-5 MDE with mixed features, n (%) 3 (2.3) 7 (5.3) 3.922 (1) **.048**
DSM-5 MDE with anxious distress, n (%) 33 (25.2) 13 (9.9) 4.378 (1) **.036**
DSM-5 lifetime MDE with psychotic features, n (%) 3 (2.3) 9 (6.9) 6.543 (1) **.001**
DSM-5 MDE with atypical features, n (%) 17 (13) 11 (8.4) .020 (1) .887
DSM-5 MDE with catatonic features, n (%) 0 0 -- --
Rapid cycling, course of depression, n (%) 1 (0.8) 5 (3.8) 4.798 (1) .028
DSM-5 MDE with post-partum onset, n (%) 13 (14) 11 (11.8) .331 (1) .565
MDD, major depressive disorder; BD, bipolar disorder; DSM-5, Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition; MDE, major depressive episode.
Note: Statistically significant difference in bold or p ≤.05.
######
Essential pharmacological treatment history across the included cases.
Current MDE
------------------------------------------------------------------------------ ------------- ------------ ------------ ---------
Data expressed as "n (%)"
Currently on antidepressant(s) 31 (81.59) 13 (86.67) 37 (63.79) 18 (90)
Currently on SGA(s) 21 (55.26) 6 (40) 1 (1.72) 12 (60)
Currently on lithium 13 (34.21) 11 (73.33) 4 (6.89) 10 (50)
Currently on FGA(s) 6 (15.79) 4 (26.67) 6 (10.34) 3 (15)
Currently on mood stabilizer(s) other than lithium 22 (57.89) 8 (53.33) 10 (17.24) 5 (25)
Currently on aripiprazole 3 (7.89) 4 (26.67) 0 2 (10)
Currently on asenapine 4 (10.53) 0 0 0
Currently on brexpiprazole 0 0 0 0
Currently on cariprazine 0 0 0 0
Currently on clozapine 0 0 0 0
Currently on lurasidone 0 0 0 0
Currently on olanzapine 4 (10.53) 2 (13.33) 0 2 (10)
Currently on paliperidone 0 0 0 0
Currently on quetiapine 9 (23.68) 0 1 (1.72) 7 (35)
Currently on risperidone 0 0 0 0
Currently on ziprasidone 1 (2.63) 0 0 0
Lifetime exposure to MOAI(s) 0 0 0 1 (5)
Lifetime exposure to carbamazepine 9 (23.68) 3 (20) 0 4 (20)
Lifetime exposure to SGA 26 (68.42) 12 (80) 7 (12.07)) 16 (80)
Lifetime exposure to TCA 17 (44.74) 8 (53.33) 10 (17.24) 8 (40)
Lifetime exposure to valproate 25 (65.79) 10 (66.67) 12 (20.69) 9 (45)
Lifetime exposure to lithium 11 (28.95) 9 (60) 4 (6.89) 12 (60)
Lifetime psychotropic polypharmacy (four or more psychotropic drugs at once) 23 (60.53) 7 (46.67) 4 (6.89) 13 (65)
MDE, major depressive episode; MDD, major depressive disorder; BD, bipolar disorder; SGA, second-generation antipsychotic; FGA, first-generation antipsychotic; TCA, tricyclic antidepressant; MAOI, monoamine oxidase inhibitor; TRD, treatment-resistant depression; TRBD, treatment-resistant bipolar depression.
TRD = treatment resistant depression was defined as failure to respond to at least two adequate trials with antidepressant drugs. TRBD = treatment resistant bipolar depression was defined as lack of response to treatment to at least two previous established treatments for bipolar depression (e.g. olanzapine/fluoxetine combination, brexpiprazole, cariprazine, quetiapine, aripiprazole). Most of the bipolar depressed patients were currently in receipt of at least one mood stabilizer drug plus an antidepressant due to the high frequency of comorbid conditions, tough the antidepressant would not represent the established treatment for bipolar depression.
######
Rating and temperamental profile of MDD and BD patients.
---------------------------------------------------------------------------------------------------------------------------------------------------------
Study subjects (n = 131) MDD (n = 78) BD (n = 53, of whom BD-I = 24; BD-II = 29) *t* or χ^2^ (df) *p*
--------------------------------------------------------------- -------------- -------------------------------------------- ------------------ ----------
Essential ratings
HAM-D-17 total score, (mean ± sd) 19.63 ± 6.19 19.55 ± 5.33 −.078 (129) .334
YMRS item 2, "increased psychomotor activity" n (%) 3 (3.85) 42 (79.25) 10.935 (3) **.012**
YMRS item 3, "hypersexuality" n (%) 1 (1.28) 7 (13.21) 7.827 (1) **.005**
YMRS item 5, "irritable mood" n (%) 29 (31.18) 30 (56.6) 8.082 (3) **.044**
YMRS item 7, "distractibility" n (%) 10 (12.82) 14 (26.42) 11.020 (2) **.004**
TEMPS-A-110 domains for affective temperaments
Depressive temperament domain score, range 1--21 (mean ± sd) 10.14 ± 4.04 10.58 ± 3.78 .568 (102) .571
Cyclothymic temperament score, range 1--21 (mean ± sd) 6.64 ± 4.41 8.91 ± 4.44 2.590 (102) **.011**
Hyperthymic temperament, score 1--21\ 3.69 ± 3.21 5 ± 4.83 1.653 (102) .101
(mean ± sd)
Irritable temperament score, range 1--21 (mean ± sd) 4.08 ± 3.29 6.47 ± 4.37 3.170 (102) **.002**
Anxious temperament score, range 1--21 (mean ± sd) 12.22 ± 6.2 12.22 ± 5.16 .002 (102) .999
---------------------------------------------------------------------------------------------------------------------------------------------------------
Statistically significant difference in bold or p ≤.05.
MDD, major depressive disorder; BD, bipolar disorder; HAM-D, Hamilton Depression Scale; YMRS, Young Mania Rating Scale; TEMPS, Temperament Evaluation of Memphis, Pisa, Paris, and San Diego-auto-questionnaire 110-item version.
Selected items of the YMRS emphasized herein as they represent the "overlapping" mixed features not accounted by the DSM-5 for the diagnosis of the "with mixed features" specifier of MDE.
Principal Component and Cluster Analysis Statistics {#s3_2}
---------------------------------------------------
PCA was performed on the items of the HAM-D-17 to characterize the sample at the report better, discriminating between unipolar and bipolar cases of depression. Two-step cluster analysis of cases based on the three components generated by the PCA led to two clinically interpretable clusters, both in the MDD ([**Table 4**](#T4){ref-type="table"}) and the BD ([**Table 5**](#T5){ref-type="table"}) group. Notably, the fetched clusters comprised different component-items from the original HAM-D scale across the BD and the MDD groups, as detailed in [**Tables 4**](#T4){ref-type="table"} and [**5**](#T5){ref-type="table"}, also reporting the proportion of resistant cases across different clusters, respectively.
######
Clusters and their characterization generated by two-step cluster analysis based on the three components obtained upon principal component analysis (PCA) on MDD patients.
Components generated by the PCA Loading HAM-D-17 items Interpretation of clusters Patients, n (%)
----------- --------------------------------- ------------------------------------------------------- ------------------------------------------------------------- -----------------------------------
Cluster 1 1 1, 2, 3, 4, 5, 6, 7, 8, 10, 15 Psychic symptoms, depressed mood, suicide, guilty, insomnia 22 (28.2), of whom 11 (14.1%) TRD
2 12, 14, 16, 17 Genitourinary, gastrointestinal, weight loss, insight
Cluster 2 3 9, 11, 13 Somatic anxiety and agitation 56 (71.8), of whom 9 (11.5%) TRD
HAM-D-17, Hamilton Depression Rating Scale, 17-item version.
######
Clusters and their characterization generated by two-step cluster analysis based on the three components obtained upon Principal Component Analysis (PCA) on BD patient.
Components generated by the PCA Loading HAM-D-17 items Interpretation of clusters Patients, n (%)
----------- --------------------------------- ----------------------------------------------------- ------------------------------------------------------------------------------------- ------------------------------------
Cluster 2 1 1, 2, 3, 5, 6, 7, 8, 10, 14 Psychic symptoms, depressed mood, suicide, guilty, insomnia, genitourinary symptoms 16 (30.2), of whom 5 (9.4%) TRBD
2 11, 13, 15 Hypochondria, somatic anxiety and fatigue, headache
Cluster 1 3 9, 12, 16, 17 Agitation, gastro-intestinal, weight loss, insight 37 (69.8), of whom 10 (18.9%) TRBD
HAM-D-17, Hamilton Depression Rating Scale, 17-item version.
Prediction of Treatment Resistance Based on Clinical Subtyping of Depression and Mixed Features Across Unipolar and Bipolar Clusters {#s3_3}
------------------------------------------------------------------------------------------------------------------------------------
Overall, the clinical subtypes of depression fetched by cluster analysis failed to allow a conclusive distinction between resistant vs. non-resistant cases (please refer to [**Tables 4**](#T4){ref-type="table"} and [**5**](#T5){ref-type="table"} for details).
The presence of DSM-5-defined mixed features among MDD and BD cases was inversely related to the resistance outcome. In contrast, the presence of the "overlapping" symptoms of depression among MDD and BD---excluded by the DSM-5 specifiers of mixed depression---such as "distractibility," "impulsivity," "increased psychomotor activity," and "hypersexuality," predicted higher rates of treatment resistance, consistently in the MDD and the BD groups. Such "permissive" criteria for mixed depression, already outlined by the 1978 Research Diagnostic Criteria ([@B44]), consistently predicted treatment resistance both in the MDD and the BD groups even in the presence of different operational definitions and pharmacological courses differed between TRD and TRBD cases. Please refer to [**Table 6**](#T6){ref-type="table"} for details. Finally, in the MDD group, the DSM-5 mixed features and the broadly defined mixed features were endorsed by 1.5% and 6.1% of cluster 1 patients, and 0.8% and 6.1% of cluster 2 patients, respectively. In the BD group, group the DSM-5 mixed features and the broadly defined mixed features were disclosed by 7.6% and 12.2% of cluster 1 patients, and 0.8% and 1.5% of cluster two patients, respectively.
######
Binary logistic regression, significant predictors only.
B (S.E.) Exp(B) 95% C.I. for EXP(B)
-------------------------------------- -------------------------------------------------------------------------------------------------------------------------------------------- -------------- --------------------- -------- --------
Major depression Cluster 1 membership: Psychic symptoms, depressed mood, suicide, guilty, insomnia & genitourinary, gastro-intestinal, weight loss, insight 1.447 (.643) **4.251** 1.206 14.983
Mixed features (DSM-5 criteria) −.281 (1.340) .764\* .59 9.879
Mixed features (permissive criteria) 2.409 (.683) **11.124** 2.916 42.435
Bipolar depression Mixed features (DSM-5 criteria) −0.88 (9.29) .916\* .148 5.658
Mixed features (permissive criteria) 1.964 (.801) **7.130** 1.485 34.247
In the MDD group, cases correctly classified = 80.8%, overall; Nagelkerke\'s R^2^ =.371. Cluster 2 membership among MDD cases (somatic anxiety and agitation) failed to reach the statistical significance threshold in the present model. In the BD group, cases correctly classified = 71.7%, overall; Nagelkerke\'s R^2^ =.194. "Permissive" refers to the standard DSM-5 specifier for mixed features of a major depressive episode broadened to include also the "overlapping" symptoms (e.g., irritability, impulsivity, distractibility) otherwise accounted by the Research Diagnostic Criteria (RDC).
\*Note: For reader convenience, an odd ratio, OR---namely an Exp(B) in this instance---below one means that the given predictor is inversely related to the outcome at study ("protective" factor), whereas a value approximating 1 or equal to one means equal chances of the outcome at study occurring or not. All the bolded text indicate p \<. 001.
Discussion {#s4}
==========
Consistent with our hypothesis, the propensity toward TRD may vary across different clinical subtypes of depression. Cluster 1 membership among MDD cases, including the core depressive symptoms, psychomotor retardation, suicidal behavior (ideation or attempt), and somatic-anxiety (namely, gastrointestinal, indigestion, cardiovascular, palpitations, headaches, respiratory, and genitourinary symptoms), predicted TRD in half of such cases (14.1% of 28.2% of cluster 1 MDD patients). This is in line with previous evidence which, albeit failing to discriminate across different clinical subtypes of depression, highlighted the role of suicidal behavior---even those of mild to moderate intensity ([@B27])---anxiety ([@B25]), and severity of the current MDE in the prediction of the treatment outcome ([@B26], [@B28]), as well as the impact of psychotic features of depression in the prediction of subsequent TRD ([@B45]). This is also in line with previous evidence involving patients diagnosed using the codes provided by the Fourth Edition of the Manual (DSM-IV) ([@B46]), yet failing to adopt a data-generated clustering of psychopathology ([@B47]). However, it must be noticed that the clustering had a poor predictive value overall in the MDD group and no significant predictive value for TRBD cases. Similarly, the affective temperaments measured by means of the TEMPS had no discriminative value for treatment resistance, neither in the MDD, neither in the BD subgroups, in line with previous reports comparing unipolar and bipolar cases ([@B48]).
On the other side, the present study highlights the impact of mixed features of depression, broadly defined to encompass also the not overlapping symptoms otherwise excluded by the DSM-5. Mixed features of depression were somewhat similar in terms of frequency among different data-generated subtypes of depression, in the MDD and the BD groups. Overall, mixed features consistently predicted treatment resistance among BD and MDD cases across different clinical subtypes of depression, especially when these latter included "excitement" symptoms among MDD patients. Indeed, mixed features of depression have been suggested to represent hallmarks of bipolarity among DSM-defined cases of BD, which may, on turn, be prone to treatment failure with standard MDD medications ([@B49]). In particular, MDD patients may require strategies other than the two "adequate" trials with antidepressants before receiving a diagnosis of treatment resistance in the presence of mixed features, especially when these latter encompass the "excitement" overlapping symptoms. In the MDD group, patients without broadly-defined mixed features, or DSM-5 defined mixed features of depression neither, may be more likely to respond to standard antidepressant care, thus endorsing lower rates of TRD. Stating the categorical nature of the operational definitions adopted by the present study for TRD or TRBD, it cannot be excluded, however, that the propensity for treatment resistance may span across a linear gradient ranging from almost neutral risk (OR =.916; C.I. =.148--5.658 in BD patients with narrowly-defined mixed features of depression) toward a 7-fold increased odd (OR = 7.130; C.I. = 1.485--34.247) in the BD group endorsing broadly-defined mixed features. Moreover, since the DSM-5-defined mixed features had negligible predictive value for treatment resistance, the use of the sole boundary-defined mixed features proposed by the DSM-5 itself may lead to improper pharmacological management of depression ([@B50]).
The categories proposed the DSM consist of "ideal types," abstractions from reality, which we can use to assess empirical observations but which are not themselves, empirical facts. This latter concept, which is central to Jaspers\' thinking in his textbook "General Psychopathology" ([@B51]), was derived from the work of Max Weber ([@B52]). While the ideal types may aid in discriminating depression from, say mania, the most common clinical presentations are an admixture of both conditions ([@B53]). From this perspective, we remark the relevance of the present study, although its preliminary nature solicits for additional research on the matter. However, more sensible operational definitions for TRD, especially in BD, are warranted (Fornaro M. et al., submitted for publication).
There several limitations of the present study, which was conducted in a tertiary care outpatient setting that needs to be accounted for in the interpretation of the results. The sample size is a relatively small one, with no longitudinal follow-up or systematic stratification of pharmacological history across different classes of antidepressants, titration modality. No stratification according to the predominant mood polarity or polarity index of BD was reported. Recall bias may have hampered the validity of some diagnoses, including general medical comorbidities (which may have nonetheless affected the propensity of treatment response). Specifically, those BD patients who were also in receipt of antidepressant(s) drugs beyond the mood stabilizer(s) need to be further stratified by upcoming studies in order to better account for several comorbidities (e.g., obsessive--compulsive disorder) that may have led to pseudo-resistance in some cases. However, per study-protocol, even those BD patients who were currently in receipt of antidepressant(s) at the time of the assessment had already failed at least two adequate courses with established treatments for acute bipolar depression (e.g., the atypical antipsychotic quetiapine, or the olanzapine/fluoxetine combination). Besides, there was no virtually no record of the psychosocial or neuro-modulatory interventions, any correlation measures with neither neuroimaging, molecular nor genetic marker. Yet, neuro-modulatory interventions such as electroconvulsive therapy represent consolidated treatment avenues for TRD ([@B54]). Besides, the psychopathological clustering of depression based entirely on the sole HAM-D, which may not be sufficient to catch the complexity of depression exhaustively.
However, stating the data-generating hypothesis of the present preliminary report, we further reinforce the need for upcoming studies to replicate the present results across different clinical subtypes of depression and a large number of putative biological and neuropsychological measures.
Data Availability Statement {#s5}
===========================
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.
Ethics Statement {#s6}
================
The studies involving human participants were reviewed and approved by Federico II University of Naples. The patients/participants provided their written informed consent to participate in this study.
Author Contributions {#s7}
====================
MF and AF conceived the study and extensively edited the main-text and its attachments together with SN, ABl, AA, PM, FI, and ABa. Finally, all the co-authors contributed substantially to the present piece of work.
Conflict of Interest {#s8}
====================
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.
[^1]: Edited by: Brisa Fernandes, University of Texas Health Science Center at Houston, United States
[^2]: Reviewed by: Massimo Pasquini, Sapienza University of Rome, Italy; Lucie Bartova, Medical University of Vienna, Austria
[^3]: This article was submitted to Mood and Anxiety Disorders, a section of the journal Frontiers in Psychiatry
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
Graphene is a monolayer of sp^2^-bonded carbon atoms, and this sp^2^ bond makes the graphene structure look like honeycomb crystal, as shown in Figure [1](#F1){ref-type="fig"}\[[@B1]\]. Graphene is called the mother of graphite (many layers of graphene) because it can act as the basic building block of these allotropes \[[@B2],[@B3]\]. Graphene was theoretically discovered back in the 1940s, but at that time, graphene (a 2D layer crystal) was believed to be too thermodynamically unstable to be produced in the real world \[[@B4]\].
{#F1}
After Andre Geim and Konstanstin, Novoselov successfully produced graphene from Scotch tape in 2004, research attention has moved forward rapidly on graphene.
Graphene has attractive electrical properties such as the ability to continuously tune the charge carriers from holes to electrons, high mobility, and high-carrier velocity \[[@B5]-[@B8]\]. The charge carrier (electron) in graphene can be explained by electron propagation through the honeycomb lattice of graphene that develops after the electrons lose their effective mass, which yields quasi-particles called 'Dirac fermions' \[[@B9]\]. These Dirac fermion particles are hard to imagine because they have no known analogies \[[@B9]\]. They can be illustrated by a combination of both Dirac and Schrödinger equations. In addition, graphene requires current to be effective, precise, and faster than any other metal on biosensors, in the same way as a biomimetic membrane-coated graphene biosensor \[[@B10]\]. Several types of animal and plant cells are surrounded with a two-layer covering, which is called the phospholipid bilayer \[[@B11]\]. As shown in Figure [2](#F2){ref-type="fig"}, the molecules that make up the phospholipid bilayer, called phospholipids, organize themselves into two corresponding layers, shaping a covering that can only be infiltrated by certain kinds of substances \[[@B11]\]. This gives the cell an apparent barrier and keeps useless materials out \[[@B12]\].
{#F2}
Although the phospholipid bilayer frequently works well, it can be damaged, and some superfluous materials can penetrate it. Phospholipids have two ends; the first is hydrophilic and attracts water; and the second is hydrophobic and resists water \[[@B12]\]. As the inside of the cells is typically water and the region outside the cells is generally water, these molecules organize themselves into two sheets, with the hydrophilic ends of each layer pointing outwards and the hydrophobic parts pointing inwards \[[@B1]\]. While they are fats or lipids, they are not crushed by the water and are firm enough to prevent large molecules passing through without the assistance of some other material \[[@B1]\].
Some smaller molecules, such as carbon dioxide and oxygen, can pass through without difficulty on their own, but larger molecules such as water, sodium, or magnesium cannot pass easily \[[@B13]\]. The interior of the membrane is also liquid, and this lets proteins, cholesterol, sphingolipids, or sterols converge in it. The role of sphingolipids is to protect the outside of the cell, and the role of the sterols and cholesterols is to stabilize the phospholipid bilayer in plant and animal cells, respectively \[[@B13]\]. Although this is critical for cells to have enough constancy, a large amount of cholesterol can make them inflexible, which is hazardous especially if they are part of a vein that must be flexible to allow blood flow \[[@B10]\]. The proteins are used to transfer materials in or out of the cell throughout the bilayer and to provide places for certain materials to attach to the exterior of the cell \[[@B10]\]. Providing the structure to a cell is one of the major roles of the phospholipid bilayer, which it performs because of the natural arrangement of the hydrophilic and hydrophobic ends of the phospholipids and, once stable, the sterols and cholesterol \[[@B10]\]. Its other role is to control the kinds of materials that can go into the cell or attach to it, which it does in a number of ways using proteins \[[@B4]\]. The kinds of protein that expand from the top of the membrane can be used to recognize the cell or to make a place for specific materials to attach to it \[[@B1]\]. Also, some types of proteins can shape tunnels or channels to allow certain substances to go through. Some channels are always open for certain types of molecules, while others need energy to open and close like gates \[[@B14]\]. This kind of transportation is active transport and can work in both ways, to bring substances in and out of the cell. It is generally used with materials like calcium, potassium, and sodium \[[@B15]\]. A charged lipid bilayer adsorbing on the surface can adopt the electronic properties of graphene. An electrolyte-gated biomimetic membrane-graphene transistor can be used to monitor electrically the bio-recognition events that lead to changes in the membrane\'s uprightness. Graphene can sense electrically the bactericidal motion of antimicrobial peptides based on a multipart interaction of an ionic screening effect and biomolecular doping \[[@B15]\]. The graphene-based FET structure can be used in the sensing of biological events when there is variation of electrical parameters. The observed transfers of the Dirac point, along with the indication of lipid charges, is an indicator of the charge-impurity potential made by the lipid membranes and shows clearly that the exciting lipid membranes adapt the electronic properties of graphene considerably. Assuming an equivalent division of exciting lipids in the two leaflets, since graphene is an electrically neutral substrate, the concentration of charged pollutants in the lipid membranes can be approximated from the surface area connected to a lipid head group. Also, an analytical modeling for electrolyte-gated biomimetic membrane-graphene biosensor is essential to improve and more recognize the impact of both thickness and electrical charge on the biomimetice membrane. By means of the charged lipid bilayer\'s adsorption on the membrane surface, the conductance of graphene can be adapted and replicated. Biorecognition actions which cause modifications to the membrane integrity can be considered electrically using an electrolyte-gated biomimetic membrane-graphene biosensor (GFET). In the current paper, a monolayer graphene-based GFET with a focus on the conductance variation caused by membrane electric charges and thickness is studied. Monolayer graphene conductance as an electrical detection platform is suggested for neutral, negative, and positive electric membrane. In addition, the effect of charged lipid membranes on the conductance of graphene-based GFET is estimated regarding the significant shift in the Dirac point in the *G*-*V*~g~ characteristic of the graphene-based biosensor. A monolayer graphene-based GFET with a focus on the conductance variation caused by membrane electric charges and thickness is studied. Monolayer graphene conductance as an electrical detection platform is suggested for neutral, negative, and positive electric membrane. The electric charge and thickness of the lipid bilayer (*Q*~LP~ and *L*~LP~) as a function of carrier density are proposed and the control parameters are defined.
Proposed model
--------------
The monolayer graphene in an electrolyte-gated biomimetic membrane graphene transistor with a ballistic channel is assumed to monitor the changes in membrane integrity. High-carrier mobility is reported in experiments on the graphene, which is thought to be due to the totally ballistic carrier transportation in the graphene, which leads to a higher transmission probability. By applying the Taylor expansion on graphene band energy near the Fermi point, the *E* (*k*) relation of the GNR is obtained as \[[@B17]\].
$$E\left( \overset{\rightharpoonup}{k} \right) = \pm \frac{3\mathit{ta}}{2}\sqrt{k_{x}^{2} + \beta^{2}}$$
where *k*~*x*~ is the wave vector along the length of the nanoribbon and *β* is quantized wave vector given by \[[@B18]\]. Based on this wave vector, number of actual modes *M*(*E*) at a given energy which is dependent on the sub bands location can be calculated. By taking the derivatives of wave vector *k* over the energy *E* (dk/dE), the number of the mode *M*(*E*) is written as
$$M\left( E \right) = \frac{\mathit{\Delta E}}{\mathit{\Delta k}.L} = \frac{3\mathit{at}}{2L}\left( {\frac{4E}{3\mathit{at}} - 2\beta^{2}} \right)^{\frac{1}{2}}$$
where *L* is the length of the nanoribbon. A higher transmission probability causes a higher carrier conductance from source to drain, as provided by the Boltzmann transport equation \[[@B2],[@B3]\]:
$$G = \frac{2q^{2}}{h}\int\limits_{- \infty}^{+ \infty}{dEM}\left( E \right)T\left( E \right)\left( {- \frac{df}{dE}} \right)$$
where *q* is the electron charge, Planck\'s constant is shown by *h*, *E* is the energy band structure, *M*(*E*) is the number of modes, *f* is the Fermi-Dirac distribution function and *T*(*E*) is the transmission probability. On the other hand, because of the ballistic transport *T*, the possibility of one inserted electron at one end that can be conveyed to other end is considered equivalent to one (*T*(*E*) = 1). The number of modes in accordance with the Landauer formula with respect to the conductance of monolayer graphene can be written as
$$M\left( E \right) = \frac{\propto - 3\beta k^{2}}{l}$$
where the length of the graphene channel is shown with parameter *l*, *k* is the wave vector, and $\propto = \frac{\Delta v_{F}}{\left. \sqrt{}2t \right.},$$\beta = \frac{v_{F}^{3}}{\Delta\left. \sqrt{}2t \right.}$. It can be affirmed that the length of the channel has a strong influence on the conductivity function. Taking into consideration the effect of temperature on graphene conductance, the boundary of the integral is changed. This equation can be numerically solved by employing the partial integration method:
$$\begin{array}{l}
{G = \frac{3q^{2}\left( {3\pi a^{3}t^{3}k_{B}T} \right)^{\frac{1}{2}}}{\mathit{hl}}} \\
{\mspace{54mu} \times \left\lbrack {{\int_{0}^{+ \infty}{\frac{X^{- 1/2}}{\left( 1/\left( {1 + e^{X - \eta}} \right) \right)}\mathit{dX} + {\int_{0}^{+ \infty}\frac{X^{- 1/2}}{\left( 1/\left( {1 + e^{X + \eta}} \right) \right)}}}}\mathit{dX}} \right\rbrack} \\
\end{array}$$
where *x* = (*E* - *E*~g~)/*k*~B~*T* and the normalized Fermi energy is *η* = (*E*~F~ - *E*~g~)/*k*~B~*T*. Thus, the general conductance model of single-layer graphene obtained is similar to that of silicon reported by Gunlycke \[[@B16]\]. According to the conductance-gate voltage characteristic of graphene-based electrolyte-gated graphene field effect transistor (GFET) devices, the performance of biomimetic membrane-coated graphene biosensors can be estimated through this equation. By assuming that the source and substrate terminals are detained in ground potential, the channel region has the characteristics of the resistor in the small voltage between the source and drain (*V*~DS~). As shown in Figure [3](#F3){ref-type="fig"}, the performance of a lipid bilayer-based sensor based on graphene nanostructure is assessed by the conductance characteristic. Before the electrolyte solution has been added, pure water as a water-gated ambipolar GFET was added into the membrane to measure the transfer curve. There is substantial agreement between the proposed model of the lipid bilayer-based biosensor and the experimental result which is extracted from the reference \[[@B10]\].
{#F3}
As depicted in Figure [4](#F4){ref-type="fig"}, by applying the gate voltage to the biomimetic membrane, it is clearly seen that the conductance of GFET-based graphene shows ambipolar behavior. The doping states of graphene are monitored by the *V*~g,min~ to measure the smallest conductance of the graphene layer, which is identified from the transfer characteristic curve. In total, the *V*~g,min~ shift (at the Dirac point) can be considered as a good indicator for lipid bilayer modulation and measurement. Nevertheless, the magnitude of the voltage shift from both positive and negative lipids is comparable when this shift is measured from the position of the minimum conductivity of bare graphene. As shown in Figure [4](#F4){ref-type="fig"}, the changes in the membrane\'s electric charge can be detected electrically. The conductivity graph is changed when the electric charges are changing for biomimetic membrane-coated graphene biosensor. So, more electrically charged molecules will be adsorbed and the sensor will be capable of attracting more molecules, which leads to a change in the *V*~g,min~ on the device, and the hole density value can be estimated as decreasing. A negatively exciting membrane demonstrates a very small enhancement in conductivity and a positive change in the Dirac point compared with that of exposed graphene.This is because of an enhancement in the remaining pollution charges caused by the negatively charged membrane. A detection-charged lipid bilayer can be obtained based on a detectable Dirac point shift. In light of this fact, the main objective of the current paper is to present a new model for biomimetic membrane-coated graphene biosensors. In this model, the thickness and the type of coated charge as a function of gate voltage is simulated and control parameters are suggested. Subsequently, to obtain a greater insight into the role of both the thickness and the type of lipid bilayer, GFET modeling is employed to identify the relationship between the conductance and the voltage of the liquid gate, where two electrodes of the sensor, as shown in Figure [5](#F5){ref-type="fig"}, are considered as the source and drain contacts.The conductance of the GFET channel is dependent not only on the operating voltage on the source-drain channel and graphene organization, but also on the biomimetic membranes of diverse surface charges that accumulate on the graphene and charged lipid bilayer, absorbed by the graphene surface. The conductivity of the graphene-based GFET device is influenced by the charge carrier density changing in the channel. As shown in Figure [6](#F6){ref-type="fig"}, because of the membrane thinning effect, the conductance of the channel is altered.
![**Comparison between GFET-conductance model and extracted experimental data**\[[@B10]\]**.** For graphene coated with negatively charged, positively charged and neutral POPC membranes.](1556-276X-9-371-4){#F4}
![**Schematics of the structure and the electrical circuit of the electrolyte-gated graphene-FET for charged lipid bilayer detection**\[[@B10]\]**.**](1556-276X-9-371-5){#F5}
{#F6}
Different ions can be adsorbed by changes in the membrane\'s electric charge and thickness, and subsequently, the sensor will be capable of attracting the ions in the solution which have caused a transformation in the conductance of the graphene-based biosensor. Dependent upon the channel conductance in the biomimetic membrane-coated graphene biosensor, it is concluded that GLP is a function of electric charge and thickness, where GLP is the channel conductance after adding the lipid bilayer. The focus of the present paper is to demonstrate a new model for GFET to measure changes in the membrane\'s electric charge and thickness. In other words, the conductance of the GFET device as a function of different electric charges and thicknesses is simulated and an electric charge factor (*α*) and thickness factor (*β*) are suggested. Subsequently, for better understanding of the role of the lipid bilayer, FET modeling is employed to obtain an equation describing the conductance, electric charge, and thickness, where the suggested structure of the GFET is shown in Figure [5](#F5){ref-type="fig"}. This means that *G*~LP~ is considered to be a function of electric charge (*Q*~LP~) as follows. G~*LP*~ = *G*~*Neutral*~ + *αQ*~*LP*~ where electric charge factor (*α*) is assumed, *G*~LP~ is the channel conductance of graphene with biomimetic membranes of different surface charges, and *Q*~LP~ is the electrical charge of the membrane. Consequently, the supposed conductance model of the graphene-based GFET channel can be written as.
$$\begin{array}{l}
{G_{\mathit{LP}} = \left( {\frac{3q^{2}\left( {3\pi a^{3}t^{3}k_{\mathit{BT}}} \right)^{\frac{1}{2}}}{\mathit{hl}}\left\lbrack {\Im_{\frac{- 1}{2}}\left( \eta \right) + \Im_{\frac{- 1}{2}}\left( {- \eta} \right)} \right\rbrack} \right)} \\
{\mspace{54mu} + \left( {\alpha Q_{\mathit{LP}}} \right)} \\
\end{array}$$
In Figure [7](#F7){ref-type="fig"}a,b, each diagram clearly depicts the specific electric charge. For example, when graphene is coated with a negative charge, it is noteworthy that the model is closer to the experimental data; in the same manner, we can compare graphene coated with the positive charge as well. It is clearly shown that, by varying the electric charge through the electric charge factor, the *G*-*V*~g~ characteristic curve can be controlled.
![**Comparison between graphene conductance model and extracted experimental data**\[[@B10]\]. **(a)** For negatively electric charges. **(b)** For positively electric charges.](1556-276X-9-371-7){#F7}
Furthermore, the proposed model is strongly supported by the experimental data. According to Figure [7](#F7){ref-type="fig"}, the amount of *V*~g,min~ shift is changed by changing the electric charge from a negatively charged membrane to a positively charged membrane, probably due to the fact that the changing electric charge has an effect on the Dirac point. It is evident that the graphene channel will be doped to an n-type region with a negatively charged membrane, whereas it changes to hole doping under a positively charged membrane. By increasing the membrane thickness on the graphene surface, the *V*~g,min~ is dramatically left-shifted. It can therefore be concluded that *V*~g,min~ is very sensitive to the electric charge and the thickness of the membrane. To support this, the gate voltage shifted leftwards owing to the fact that the graphene will be n-doped by the high membrane thickness. On the other hand, the conductivity of the graphene-based FET device is influenced by the increased number of carriers in the channel. In other words, the *V*~g,min~ will be shifted leftwards and the extent of the shift increases with the increasing thickness of the membrane from 0.01 nM to 10 μM. In order to verify the proposed model, the effect of membrane thickness will be assumed and *G*~LP~ is modified as a function of electric charge (*Q*~LP)~ and membrane thickness as follows:
$$G_{\mathit{LP}} = G_{\mathit{Neutral}} + \left( {\alpha Q_{\mathit{LP}} + \beta L_{\mathit{LP}}} \right)$$
where (*β*) and *L*~LP~ are the thickness parameter and thickness of the adsorbed lipid bilayer, respectively. In the non-saturation region, the GFET conductance model is involved as a result of gate electrical energy and the perfect conductance-voltage related to the graphene channel of the GFET device, which leads to the modified conductance as:
$$\begin{array}{l}
{G_{\mathit{LP}} = \left( {\frac{3q^{2}\left( {3\pi a^{3}t^{3}k_{\mathit{BT}}} \right)^{\frac{1}{2}}}{\mathit{hl}}\left\lbrack {\Im_{\frac{- 1}{2}}\left( \eta \right) + \Im_{\frac{- 1}{2}}\left( {- \eta} \right)} \right\rbrack} \right)} \\
{\mspace{54mu} + \left( {\alpha Q_{\mathit{LP}} + \beta L_{\mathit{LP}}} \right)} \\
\end{array}$$
In Figure [8](#F8){ref-type="fig"}b, all the theoretical *G*~LP~-*V*~g~ characteristics of graphene-based GFET with *L*~LP~ = 10 μM are plotted. Comparing Figures [8](#F8){ref-type="fig"}a and b, it can be seen that the biomimetic membrane-coated graphene biosensor model according to the suggested parameters (*α* and *β*) indicates the same trends as those reported by \[[@B10]\]. In both the experimental and theoretical data, there is a clear shift in *V*~g,min~ with increasing membrane thickness. Comparison of the experimental data depicted with the theoretical data in Figure [8](#F8){ref-type="fig"} shows that a 10 μM membrane thickness caused a 10-meV shift in *V*~g,min~.
![**Extracted experimental data for membrane thickness effect and*G*-*V***~**g**~**characteristic of proposed conductance model. (a)** Extracted experimental data for membrane thickness effect of biomimetic membrane-coated graphene biosensor. **(b)***G*-*V*~g~ characteristic of proposed conductance model with experimental data \[[@B10]\] for 10-μM membrane thickness.](1556-276X-9-371-8){#F8}
In the suggested model, differently charged lipid bilayers and membrane thicknesses are demonstrated in the form of *G*~LP~ and *L*~LP~ parameters, respectively, in agreement with the reported data which is shown in Table [1](#T1){ref-type="table"}. The *V*~g,min~ did not shift further at greater membrane thicknesses due to the saturation current density of the injected carrier concentration by the charged lipid bilayer.
######
**Different*Q***~**LP**~**and*L***~**LP**~**values with*V***~**g,min**~**changes**
**V**~**g,min**~**(V)**
------------- -------------------------
Q~LP~
Neutral 0.11
Negatively 0.29
Positively -1.1
L~LP~
10 nm 0.24
0.1 μm 0.135
1 μm 0.09
10 μm 0.045
According to the saturation region of the presented conductance model and given that g~m,min~ belongs to the graphene-based biosensor, the control parameter with respect to the iteration method is suggested as:
$$\beta = l_{1}e^{- l_{2}L_{\mathit{LP}}}$$
where *l*~1~ = 0.4157 and *l*~2~ = -0.543. In addition, *α* for the neutrally, negatively, and positively charged membrane is assumed to be 0, 1, and -1, respectively. Consequently, the justified model for the interaction of charged impurity and the consequence of charged lipid membranes in a biomimetic membrane-coated graphene biosensor is proposed as
$$\begin{array}{l}
{G_{\mathit{LP}} = \left( {\alpha Q_{\mathit{LP}} + L_{\mathit{LP}}l_{1}e^{- l_{2}L_{\mathit{LP}}}} \right)} \\
{\mspace{54mu} + \left( {\frac{3q^{2}\left( {3\pi a^{3}t^{3}k_{\mathit{BT}}} \right)^{\frac{1}{2}}}{\mathit{hl}}\left\lbrack {\Im_{\frac{- 1}{2}}\left( \eta \right) + \Im_{\frac{- 1}{2}}\left( {- \eta} \right)} \right\rbrack} \right)} \\
\end{array}$$
The proposed model, coupled with the experimental data, is shown in this work to confirm that the conductivity of the graphene-based biosensor is changed by the electric charge and membrane thickness of the lipid bilayer. In a nutshell, electrolyte-gated graphene field-effect transistor structure was used after chemical vapor deposition (CVD) as the electrical transduction stage because of its high electrical conductivity, optical transparency, and large area, given the likelihood of manufacturing a dual-mode optical and electrical detection system for detecting the changes of membrane properties. Based on what has been discussed, one could firmly claim that, in response to changes of the charged lipid membranes and charges of biomimetic membranes of different thicknesses, a significant shift in *V*~g,min~ of the ambipolar FET occurs due to the electronic devices on both the n-doping and p-doping materials.
Conclusion
==========
The emerging potential of nanostructured graphene-based biosensors in the highly sensitive and effective detection of single-base polymorphism or mutation, which is thought to be the key to diagnosis of genetic diseases and the realization of personalized medicine, has been demonstrated. In a lipid bilayer-based biosensor, the graphene carrier concentration as a function of the lipid bilayer can be modeled. In this research, the total conductance of graphene as a function of the electric charge (*Q*~LP~) and thickness of the adsorbed lipid bilayer (*L*~LP~) is presented. A dramatic decrease in the minimum conductance related to the gate voltage (*V*~g,min~) by both changing the electrical charge from negative to positive and decreasing the lipid thickness has been reported. In the presented model, the *V*~g,\ min~ variation based on the adopted experimental data as an electrical detection platform is considered and the sensor control parameters are defined. The presented model confirms the reported experimental data and in addition facilitates the employment of alpha and beta as biosensor control parameters to predict the behavior of graphene in graphene-based biosensors.
Competing interests
===================
The authors declare that they have no competing interests.
Authors' contributions
======================
MJK wrote the manuscript and contributed to the analytical modelling of the presented FET via MATLAB software. Dr. FKCh and Dr. MTA revised the manuscript and coordinated between all the contributors. HKFA, MR, and AH organized the final version of the manuscript. All authors read and approved the final manuscript.
Acknowledgment
==============
The authors would like to acknowledge the financial support from the Fundamental Research Grant Scheme for research grant 'Novel hybrid nanocomposite large sensor array for future nose on a chip' of the Ministry of Higher Education (MOHE), Malaysia. The authors also thank the Research Management Center (RMC) of University Technology Malaysia (UTM) for providing an excellent research environment in which to complete this work.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-ijms-18-02624}
===============
Nitrogen (N) is a fundamental constituent of many cell components such as amino acids, proteins, cell walls, membranes and nucleic acids. Nitrogen deficiency reduces plant growth and development, photosynthesis, leaf area, and ultimately limits plant productivity \[[@B1-ijms-18-02624],[@B2-ijms-18-02624],[@B3-ijms-18-02624],[@B4-ijms-18-02624]\]. For a sustainable crop production system, there is a requirement to reduce nitrogen fertilizer input, and increase nitrogen use efficiency. This may be achieved by understanding the relationship between N nutrition and the photosynthetic rate in the leaf \[[@B5-ijms-18-02624],[@B6-ijms-18-02624]\].
Photosynthesis is a biological process whereby the sun's energy is captured and stored in a series of events that convert the pure energy of light into the biochemical energy needed to power life \[[@B7-ijms-18-02624]\]. Photosynthesis includes two reactions: light reactions and carbon reduction reactions \[[@B7-ijms-18-02624]\]. Light reactions take place on thylakoid membranes. Thylakoid membranes mainly include four membrane protein complexes: photosystem I (PSI), photosystem II (PSII), cytochrome *b*~6~*f* complex (Cyt *b*~6~*f*), and adenosine triphosphate (ATP) synthase. Carbon reduction reactions take place in the stroma of chloroplast. Ribulose-1,5-bisphosphate carboxylase (Rubisco), phosphoenolpyruvate carboxylase (PEPc) and pyruvate orthophosphate dikinase (PPDK) are the key enzymes for C~4~ plants. Light is absorbed by pigments (mainly chlorophylls) associated with two photosystems: PSI and PSII. For light energy to be stored by photosynthesis, it is channeled by a number of different processes including photochemistry, heat dissipation (including photo-protective heat dissipation and other heat dissipation), and chlorophyll fluorescence \[[@B8-ijms-18-02624]\]. Chlorophyll fluorescence is an important parameter in representing changes in the growth environment and function of the photosynthetic process \[[@B8-ijms-18-02624]\]. For example, in rice, nitrogen deficiency decreases the actual quantum yield of PSII photochemistry (Φ~PSII~), the maximal efficiency of PSII photochemistry (*F*v/*F*m), excitation energy capture efficiency of PSII (*F*v'*/F*m'), as well as the electron transport rate (ETR) \[[@B9-ijms-18-02624]\].
Photosynthesis depends on many physiological and biochemical processes such as stomatal conductance, intercellular CO~2~ concentration, photochemical capacity of PSII, and contents and activities of carbon fixation enzymes \[[@B10-ijms-18-02624]\]. Photosynthesis has a positive relationship with leaf nitrogen \[[@B3-ijms-18-02624],[@B5-ijms-18-02624],[@B11-ijms-18-02624],[@B12-ijms-18-02624]\]. This is because about 70% of leaf nitrogen is located in the chloroplast \[[@B11-ijms-18-02624],[@B13-ijms-18-02624]\]. Nitrogen deficiency reduces the content of chlorophyll, Cyt *f*, coupling factor, N content of thylakoid in light reactions, as well as the ETR \[[@B5-ijms-18-02624],[@B14-ijms-18-02624]\]. Nitrogen deficiency also decreased the content and/or activity of Rubisco, PEPc and PPDK \[[@B3-ijms-18-02624],[@B5-ijms-18-02624],[@B11-ijms-18-02624],[@B12-ijms-18-02624]\]. The decreased photosynthesis in nitrogen deficiency may be caused by different reasons in different plants under different conditions. For example, decreased photosynthesis in low nitrogen was mainly associated with lower stomatal conductance in sorghum under outdoor pot-culture conditions \[[@B10-ijms-18-02624]\]. At light saturation, the decline in photosynthesis in nitrogen deficiency was mainly caused by the limitation of reduced mesophyllic activity, rather than by stomatal limitation in sunflowers \[[@B15-ijms-18-02624]\]. Nitrogen deficiency repressed ΦPSII, causing decreased ETR to match the decreased requirements for ATP and triphosphopyridine nucleotide (NADPH) in the case of decreased CO~2~ assimilation capacity in maize grown outdoors \[[@B16-ijms-18-02624]\].
Maize is an important crop worldwide with multiple purposes such as food, fodder, and bioenergy. Although the relationship between nitrogen deficiency and photosynthesis has been widely studied in maize, less is known about the molecular mechanism underlying the response of the photosynthetic system in maize to nitrogen deficiency. Such information is important not only for improving photosynthesis performance under insufficient nitrogen supply via gene manipulation, but also for the development of potential molecular tools for the diagnosis of plant nitrogen nutrition status. This could be used in both the selection of nitrogen-efficient genotypes and precision nitrogen fertilizer management. In the present study, using an integrated approach including physiological analysis and RNA sequencing (RNA-Seq), we aimed to explore the genes related to photosynthesis under nitrogen deficiency, especially those related to light reactions.
2. Results {#sec2-ijms-18-02624}
==========
2.1. Effect of Nitrogen Supply on Biomass {#sec2dot1-ijms-18-02624}
-----------------------------------------
Nitrogen supply has a significant effect on biomass and nitrogen accumulation. Low nitrogen treatment reduced total biomass by 15%, and shoot biomass by 25% ([Table 1](#ijms-18-02624-t001){ref-type="table"}). However, low nitrogen treatment increased the root--shoot ratio by about two-thirds compared with the high nitrogen treatment, indicating that nitrogen-starved plants allocated more biomass to the root. Total nitrogen accumulation, as well as shoot nitrogen accumulation, in low nitrogen plants were reduced by 74% and 77%, respectively ([Table 1](#ijms-18-02624-t001){ref-type="table"}).
2.2. Effect of Nitrogen Supply on Leaf Photosynthesis {#sec2dot2-ijms-18-02624}
-----------------------------------------------------
As expected, low nitrogen significantly reduced the leaf area ([Figure 1](#ijms-18-02624-f001){ref-type="fig"}a). Nitrogen supply has a significant effect on specific leaf nitrogen (SLN) and the photosynthetic rate ([Figure 1](#ijms-18-02624-f001){ref-type="fig"}b). SLN is lower (82%) in low nitrogen treatment compared with high nitrogen treatment. The photosynthetic rate in low nitrogen stress is reduced by 83% compared with high nitrogen treatment. The chlorophyll (Chl), chlorophyll a, and chlorophyll b concentration are lower (67%, 69% and 59% respectively) in low nitrogen plants compared with high nitrogen plants ([Figure 2](#ijms-18-02624-f002){ref-type="fig"}). The ratio of Chl to N was 31% higher in low nitrogen treatment compared with high nitrogen treatment.
2.3. Low Nitrogen-Induced Gene Expression in the Leaf {#sec2dot3-ijms-18-02624}
-----------------------------------------------------
We used RNA-Seq analysis of the mature tissue of an expanding leaf in order to unravel the molecular mechanisms underlying the response of the photosynthetic system to low nitrogen supply. A total of more than 33 million high quality 125-bp paired-end reads were generated by RNA-Seq for each treatment. Comparisons of two biological replicates showed that expression values were highly correlated (average *R*^2^ = 0.9288, [Figure S1](#app1-ijms-18-02624){ref-type="app"}).
Setting thresholds of a False Discovery Rate of less than 0.05 and a two-fold change in expression, we found 1625 differentially expressed genes (DEGs) in low nitrogen compared with high nitrogen treatment. Of these genes, 928 were downregulated whilst 697 upregulated ([Supplemental Table S1](#app1-ijms-18-02624){ref-type="app"}). The identified genes were subjected to gene ontology term (GO) enrichment analysis. The downregulated DEGs were mainly involved in photosynthesis, homeostatic processes, regulation of nitrogen metabolism, redox activity, and response to abiotic stimulus ([Figure 3](#ijms-18-02624-f003){ref-type="fig"}a; [Supplemental Table S2](#app1-ijms-18-02624){ref-type="app"}). The upregulated DEGs were mainly involved in cellular polysaccharide processes, carbohydrate biosynthesis processes and cellular metabolic processes ([Figure 3](#ijms-18-02624-f003){ref-type="fig"}b; [Supplemental Table S2](#app1-ijms-18-02624){ref-type="app"}). Using MapMan analysis, it is further identified that most downregulated DEGs were involved in photosynthesis processes, TCA cycling, and nitrogen assimilation processes ([Figure 4](#ijms-18-02624-f004){ref-type="fig"}; [Supplemental Table S3](#app1-ijms-18-02624){ref-type="app"}). The upregulated DEGs were mainly associated with lipids, the cell wall, and secondary metabolism processes ([Figure 4](#ijms-18-02624-f004){ref-type="fig"}; [Supplemental Table S3](#app1-ijms-18-02624){ref-type="app"}).
The DEGs involved in photosynthesis were analyzed ([Figure 5](#ijms-18-02624-f005){ref-type="fig"}; [Supplemental Table S4](#app1-ijms-18-02624){ref-type="app"}). Under low nitrogen supply, 11 DEGs encoding light harvesting complex II (LHCII) and three DEGs encoding LHCI were downregulated, 16 DEGs encoding PSII polypeptide subunits and 10 DEGs encoding PSI polypeptide subunits were downregulated, 3 DEGs encoding ATP synthase were downregulated, 1 DEG encoding ferredoxin reductase was downregulated, and 10 DEGs involved in Calvin cycle were downregulated. In those DEGs, GRMZM2G351977, GRMZM2G414192, GRMZM2G149428 and GRMZM2G092427 encoded components of PSII: *Lhcb1*, *Lhcb2*, *Lhcb5*, respectively. The genes GRMZM2G038519 and GRMZM2G072280 encoded components of PSI: *Lhca1* and *Lhca2*, respectively.
2.4. Effect of Low Nitrogen Supply on Leaf Chlorophyll Fluorescence {#sec2dot4-ijms-18-02624}
-------------------------------------------------------------------
Leaf chlorophyll fluorescence was analyzed to confirm the effect of low nitrogen supply on the light reaction system. Nitrogen supply has a great effect on chlorophyll fluorescence ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}). The *F*v/*F*m and *F*v'/*F*m' were 6.6% and 20.4% lower in low-nitrogen plants compared with high nitrogen plants, respectively ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}a). The actual quantum yield of PSII photochemistry (Φ~PSII~) was reduced by 67.4% in low nitrogen treatment compared to high nitrogen treatment ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}b). The electron transport rate (ETR) was 68.3% lower in low nitrogen plants than in high nitrogen plants ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}b). The non-photochemical quenching (*qN*) was 4.6% higher in low nitrogen treatment than in high nitrogen treatment ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}c). The photochemical quenching (*qP*) was lower (49.6%) in low nitrogen treatment than in high nitrogen treatment. The 1 − *F*v'/*F*m' and *F*v'/*F*m'·\* (1*qP*) were higher (18.3% and 10.3% respectively) in low nitrogen plants compared to high nitrogen plants ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}d).
3. Discussion {#sec3-ijms-18-02624}
=============
Increased photosynthesis with less input of land, water, nutrients, etc., is essential to sustainably meet global food and bioenergy demands \[[@B17-ijms-18-02624],[@B18-ijms-18-02624]\]. New models have been proposed to increase the efficiency of light capture, light energy conversion, and carbon capture and conversion, possibly by rapidly developing genetic engineering technologies \[[@B18-ijms-18-02624]\]. Photosynthesis has a close relationship with leaf nitrogen. In our study, leaf area was reduced by 22% in low nitrogen treatment. The photosynthetic rate in low nitrogen plants was 83% lower than in high nitrogen plants. Thus, the decrease in photosynthesis is the main reason for the decreasing biomass.
Nitrogen is a constituent of chlorophyll, photosynthetic enzymes (included Rubisco, PEPc and PPDK), and thylakoid membranes. These cellular features are located in chloroplast. About three-quarters of total nitrogen is found in chloroplast \[[@B11-ijms-18-02624],[@B13-ijms-18-02624]\]. The optimal SLN for the maximum photosynthetic rate (*p*~n~) is reported to be approximately 1.5 g·m^−2^ in field-grown maize \[[@B4-ijms-18-02624],[@B19-ijms-18-02624],[@B20-ijms-18-02624],[@B21-ijms-18-02624]\]. In this research, the specific leaf nitrogen is only 0.25 g·m^−2^ in low nitrogen plants, 82% lower than that of the control treatment. This correlates well with the 83% reduction in the photosynthetic rate. This reduction in SLN is the result of the contribution of the different photosynthetic components mentioned above \[[@B3-ijms-18-02624],[@B5-ijms-18-02624],[@B11-ijms-18-02624]\]. Here, we focused on the effect of nitrogen deficiency on the light reaction in photosynthesis.
Complexes PSII and PSI take part in light absorption, transport, and conversion. Photosystems are composed of two sections: (1) a reaction center devoted to the conversion of light energy into chemical energy; and (2) an antenna system that increases the capacity of light absorption and contributes to photoprotection \[[@B22-ijms-18-02624]\]. The antenna system consists of many light-harvesting complexes (*Lhc*). Different members are associated with PSI (*Lhca* proteins) and PSII (*Lhcb* proteins) \[[@B23-ijms-18-02624]\]. Over 60% of all the chlorophyll in plants is bound to light-harvesting complexes. Nitrogen deficiency leads to reduced total chlorophyll, chlorophyll a, and chlorophyll b contents ([Figure 2](#ijms-18-02624-f002){ref-type="fig"}) \[[@B3-ijms-18-02624],[@B5-ijms-18-02624],[@B11-ijms-18-02624]\]. The ratio of chlorophyll to leaf N could represent N allocation into chlorophyll \[[@B24-ijms-18-02624]\]. The ratio of chlorophyll to leaf N was higher in low nitrogen treatment ([Figure 2](#ijms-18-02624-f002){ref-type="fig"}). This suggests that maize tends to invest relatively more N into light harvesting complexes (LHCs), PSI and PSII under nitrogen stress. Chlorophyll fluorescence is frequently used to monitor the responses of the photosynthetic apparatus to environmental stressors \[[@B8-ijms-18-02624],[@B25-ijms-18-02624]\] including temperature \[[@B26-ijms-18-02624]\], nitrogen \[[@B16-ijms-18-02624]\] and drought \[[@B27-ijms-18-02624]\]. In our study, all of the chlorophyll fluorescence parameters were affected by nitrogen supply. The ETR was 68% lower in low nitrogen treatment compared with high nitrogen treatment ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}b), suggesting that low nitrogen impairs electron transport systems. Nitrogen stress reduced by 6.6% in *F*v/*F*m and 20.4% in *F*v'/*F*m' ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}a), suggesting that an energy dissipation exists in low nitrogen plants \[[@B28-ijms-18-02624]\]. Light energy absorbed by chlorophyll in a leaf has three fates: it can be used to drive photosynthesis (photochemistry), excess energy can be dissipated as heat, or it can be re-emitted as light-chlorophyll fluorescence \[[@B8-ijms-18-02624]\]. The fraction of light absorbed in PSII that is dissipated thermally (D = 1 − *F*v'/*F*m') was 18.3% higher in low nitrogen plants than in high nitrogen plants ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}d). Accordingly, nitrogen-deficient plants had a higher non-photochemical quenching (*qN*) in low nitrogen treatment ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}c) \[[@B9-ijms-18-02624],[@B16-ijms-18-02624]\]. The fraction of light absorbed in PSII that is utilized in PSII photochemistry could be represented by Φ~PSII~. Φ~PSII~ was 67.4% lower in low nitrogen plants compared with in high nitrogen plants. Nitrogen-deficient plants had a lower photochemical quenching (*qP*) in low nitrogen treatment ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}c). Lower *qP* means a lower degree of open PSII, resulting in lower Φ~PSII~. The fraction of light absorbed in PSII that is dissipated in chlorophyll fluorescence (*F*v'/*F*m · \* (1 − *qP*)) is increased by 10.3% in low nitrogen plants compared with high nitrogen plants ([Figure 6](#ijms-18-02624-f006){ref-type="fig"}d). Thus, it is supposed that low nitrogen impaired PSII, as well as electron transport. Excess energy can be dissipated thermally or re-emitted as light-chlorophyll fluorescence. In accordance, RNA-Seq analysis found that about 27 DEGs encoding PSII systems, and 13 DEGs encoding PSI systems, were downregulated by low nitrogen supply ([Figure 5](#ijms-18-02624-f005){ref-type="fig"}). Among these downregulated DEGs, GRMZM2G351977, GRMZM2G414192, GRMZM2G149428 and GRMZM2G092427 encoded *Lhcb1*, *Lhcb2*, *Lhcb5*, and *Lhcb6*, respectively. Six different proteins comprise the antenna system of PSII: three minor proteins (CP29, CP26 and CP24) encoded by *Lhcb4*, *Lhcb5* and *Lhcb6*; and the major complex LHCII, which consists of three proteins encoded by *Lhcb1*, *Lhcb2* and *Lhcb3*, of which the *Lhcb1* and *Lhcb2* proteins are by far the most abundant \[[@B29-ijms-18-02624]\]. In antisense plants of *Arabidopsis thaliana*, there is an absence of the *Lhcb1* and *Lhcb2* proteins, accompanied by reduced light absorption \[[@B29-ijms-18-02624]\]. In *Arabidopsis thaliana*, mutant lines of *Lhcb5* and *Lhcb6* have a decreased efficiency of energy transfer from LHCII to the reaction center of PSII \[[@B30-ijms-18-02624]\]. In addition, *Lhcb6* absence limits plastoquinone diffusion, and thus inhibits the electron transport rate \[[@B30-ijms-18-02624]\]. Thus, lower expression of GRMZM2G351977, GRMZM2G414192, GRMZM2G149428 and GRMZM2G092427 resulted in lower synthesis of *Lhcb1*, *Lhcb2*, *Lhcb5*, and *Lhcb6*, respectively. Reduced expression of these genes also resulted in limited light absorption, light transport and electron transport. This is consistent with the results of chlorophyll fluorescence.
GRMZM2G038519 and GRMZM2G072280 encoded *Lhca1* and *Lhca2*, respectively. The antenna systems of PSI exist as *Lhca1*/*4* and *Lhca2*/*3* dimers. The absence of one antenna complex leaves a "hole" in the structure that cannot be filled by other *Lhca* proteins \[[@B31-ijms-18-02624]\]. Therefore, lower expression of GRMZM2G038519 and GRMZM2G072280 resulted in lower synthesis of antenna systems of PSI, thus reducing light absorption in low nitrogen stress.
Overapplication of nitrogen has been a common problem in China, resulting in low N use efficiency (NUE) and environmental pollution \[[@B32-ijms-18-02624]\]. NUE can be improved by precision nitrogen management and nitrogen-efficient cultivars. For both approaches, it is necessary to diagnose plant nitrogen nutrition status timely and precisely. Chlorophyll fluorescence has been used to estimate plant nitrogen status \[[@B33-ijms-18-02624]\]. The findings of this research suggest that chlorophyll fluorescence and the related genes could potentially be explored for developing new tools for the diagnosis of plant nutrition status. Based on the chlorophyll fluorescence parameters, and the differentially expressed genes in response to nitrogen deficiency, nitrogen-efficient genotypes could be discerned at the early seedling stage. Nitrogen deficiency could be detected at the early seedling stage so that nitrogen application recommendation can be made in a timely manner. Indeed, in potatoes, it was found that chlorophyll fluorescence can discriminate between genotypes, predict plant age, and yield performance under field conditions \[[@B34-ijms-18-02624]\].
In conclusion, our results show that low nitrogen supply results in the downregulation of key genes involved in light reactions of photosynthesis. Of particular importance are PSI and PSII genes including GRMZM2G351977, GRMZM2G414192, GRMZM2G149428, GRMZM2G092427, GRMZM2G038519 and GRMZM2G072280. Downregulated expression of those genes resulted in reduced light absorption and light transport. As a result, the parameters relating to chlorophyll fluorescence (Φ~PSII~, electron transport rate, *qP*) were lower, which contributed to a lower photosynthetic rate. These findings suggest the potential to utilize chlorophyll fluorescence parameters, as well as the related genes, as indicators of plant nitrogen nutrition. In addition, this could be used to develop new tools to make precise nitrogen fertilizer recommendations, and select nitrogen-efficient genotypes.
4. Materials and Methods {#sec4-ijms-18-02624}
========================
4.1. Plant Material and Growth Condition {#sec4dot1-ijms-18-02624}
----------------------------------------
Seeds of maize (*Zea mays* L.) inbred line B73 were sterilized in 10% (*v*/*v*) H~2~O~2~ for 30 min, washed with distilled water, and then soaked in saturated CaSO~4~ solution for 8 h. Then, the seeds were washed and placed between sheets of filter paper and germinated in the dark at room temperature. When the roots were approximately 1 cm long, uniform seedlings were wrapped in filter paper and transferred into a plastic container filled with distilled water. The seedlings with two visible leaves were then transferred into porcelain pots (4 seedlings per pot) containing 2 L of nutrient solution after the endosperm was removed. The plants were grown in a growth chamber at 28/22 °C with a 14/10 h light/dark cycle. During the light cycle, the photosynthetic photon flux density was 250--300 µmol·m^−2^·s^−1^ at canopy height.
The nutrient solution contained 0.75 mM K~2~SO~4~, 0.1 mM KCl, 0.25 mM KH~2~PO~4~, 0.65 mM MgSO~4~, 0.13 mM Ethylenediaminetetraacetic acid (EDTA)-Fe, 1.0 µM MnSO~4~, 1.0 µM ZnSO~4~, 0.1 µM CuSO~4~ and 0.005 µM (NH~4~)~6~Mo~7~O~24~ \[[@B35-ijms-18-02624]\]. Plants were supplied with half-strength nutrient solution containing 4.0 mM NO~3~ (provided as Ca(NO~3~)~2~) for 2 days, and then transferred into full-strength solution with 4.0 mM NO~3~ (high nitrogen). When the third leaf was fully expanded (approximately 6 days later), half of the plants were moved into a solution with 40 µM NO~3~ (low nitrogen), with CaCl~2~ added to equalize calcium concentration between the treatments. The pH of the solution was adjusted to 5.8--6.0 with KOH. The solution was renewed every other day and was aerated continuously by a pump. The pots were arbitrarily placed and rotated when the nutrient solution was renewed.
4.2. Plant Weight, Leaf Area, and Photosynthesis {#sec4dot2-ijms-18-02624}
------------------------------------------------
When the sixth leaf was fully expanded, plants were harvested and separated into root, leaf 6, and shoot (other leaves plus stems). The length and width of the leaves were measured with a ruler. Half of sixth leaves were stored at −80 °C. Other samples were dried in an oven at 70 °C. Dried samples were weighed and ground to a powder and the nitrogen concentration was determined by an elemental analyzer (vario MACRO, Elementar, Langenselbold, Germany).
The sixth expanded leaf from 16 plants of each nitrogen treatment was used to measure the net photosynthetic rate by LI6400 (LI-COR, Lincoln, NE, USA) at a light density of 800 µmol·m^−2^·s^−1^ \[[@B5-ijms-18-02624]\]. The CO~2~ concentration inside the chamber was controlled at 400 ± 1 µmol CO~2~ (mol air)^−1^. Chlorophyll (Chl) was extracted using acetone and ethanol, and the absorbance of extracts was spectrophotometrically measured at 645 and 663 nm \[[@B36-ijms-18-02624]\].
4.3. Chlorophyll Fluorescence {#sec4dot3-ijms-18-02624}
-----------------------------
Chlorophyll fluorescence was measured with an integrating fluorescence fluorometer (LI-6400, Leaf chamber fluorometer, LI-COR, Lincoln, NE, USA). The fluorescence instable state (*F*s), maximum fluorescence under light (*F*m') and minimum fluorescence (*F*o') were measured under light density of 800 µmol·m^−2^·s^−1^ and 400 ± 1 µmol CO~2~ (mol air)^−1^. After dark-adaptation of samples for 1 h, the minimal fluorescence (*F*o) and maximum fluorescence (*F*m) in darkness were measured. Other fluorescent parameters were calculated as follows \[[@B37-ijms-18-02624]\]: maximal photochemical efficiency (*F*v/*F*m):*F*v/*F*m = (*F*m − *F*o)/*F*m; excitation energy capture efficiency of PSII reaction centers (*F*v'/*F*m'):*F*v'/*F*m' = (*F*m' − *F*o')/*F*m'; Φ~PSII~ = (*F*m' − *F*s)/*F*m'; ETR = PPFD × Φ~PSII~ × 0.85 × 0.5 (where PPFD is photosynthetic photon flux density); Photochemical quenching (*qP*): *qP* = (*F*m' − *F*s)/(*F*m' − *F*o'); Non-photochemical quenching (*qN*): *qN* = (*F*m − *F*m')/(*F*m − *F*o'). The following three derived chlorophyll fluorescence parameters were employed to analyze the allocation of fraction of excitation energy: D = 1 − *F*v'/*F*m' is the fraction of photon energy absorbed in PSII and dissipated via thermal energy in the antenna; Φ~PSII~ represents the fraction of photon energy absorbed in PSII utilized for photosynthetic electron transport; Ex = *F*v'/*F*m' ·\* (1 − *qP*) is the estimate of the fraction of excess excitation energy re-emitted as light-chlorophyll fluorescence.
4.4. RNA Library Construction and Illumina Sequencing {#sec4dot4-ijms-18-02624}
-----------------------------------------------------
The mature tissue of an expanding leaf was sampled. The samples were frozen in liquid nitrogen and stored at −80 °C. Total RNA was extracted \[[@B38-ijms-18-02624]\]. Two replicates were used to RNA-Seq. RNA-Seq libraries were prepared according the manufacturer's protocol of the Illumina Standard mRNA-Seq library preparation kit (Illumina, San Diego, CA, USA) and were sequenced to generate 125-nucleotide paired-end reads on an Illumina HiSeq platform (Illumina HisSeq 2500, San Diego, CA, USA).
4.5. Bioinformatics Analysis of RNA-Seq Data {#sec4dot5-ijms-18-02624}
--------------------------------------------
Raw reads were pre-processed to remove low quality regions and adapter sequences. At least 30 million clean reads were obtained. Clean reads from each sample were aligned to the maize reference genome (B73 RefGen_v3, available online: <http://www.maizegdb.org/assembly/>) using TopHat2 \[[@B39-ijms-18-02624]\]. Aligned reads from TopHat2 mapping were subjected to String Tie for DeNovo transcript assembly \[[@B39-ijms-18-02624],[@B40-ijms-18-02624]\]. The R package edgeR was used to identify the differentially expressed genes \[[@B41-ijms-18-02624]\]. The expression of each gene was normalized to fragments per kilobase of transcript per million reads (FPKM) to compare different samples. Low-level expressed genes were removed and only genes with an expression level of at least 1 FPKM in at least two samples were kept for further analysis. A gene was regarded as differentially expressed if the false discovery rate (FDR) was less than 0.05, and had a log2 fold change higher than 1.
The Gene Ontology (GO) term enrichment of differential expressed genes was conducted using the web-based agriGO software (Available online: <http://bioinfo.cau.edu.cn/agriGO/analysis.php>). Singular enrichment analysis (SEA) was used to compute enriched categories by comparing a list of differentially expressed genes to all expressed genes. GO terms of gene sets of interest compared with the genome-wide background with an adjusted *p* value (FDR) cutoff of 0.01. MapMan was used to show the functional categorization of differentially expressed genes in different cellular and metabolic processes \[[@B42-ijms-18-02624]\].
Raw sequencing data are stored at the Sequence Read Archive (Available online: <http://www.ncbi.nlm.nih.gov/sra>) under accession number GSE107562.
4.6. Statistical Analysis {#sec4dot6-ijms-18-02624}
-------------------------
Data were subjected to variance analysis using ANOVA procedure implemented in SPSS Statistics 19.0 (SPSS, Inc., Chicago, IL, USA). Differences were compared using the least significant difference test at 0.05 level of probability.
We gratefully acknowledge the financial support of the National Science Foundation of China (No. 21022182).
######
Click here for additional data file.
Supplementary materials can be found at [www.mdpi.com/1422-0067/18/12/2624/s1](www.mdpi.com/1422-0067/18/12/2624/s1).
Xiaohuan Mu performed the experiments, analyzed the samples, and drafted the manuscript; Qinwu Chen helped in performing experiments and analyzing data; Fanjun Chen provided material and contributed to the experimental design; Lixing Yuan supervised and assisted with writing; Guohua Mi conceived of the original screening and research plans, and revised the manuscript.
The authors declare no conflict of interest.
{#ijms-18-02624-f001}
{#ijms-18-02624-f002}
{#ijms-18-02624-f003}
{#ijms-18-02624-f004}
{#ijms-18-02624-f005}
{#ijms-18-02624-f006}
ijms-18-02624-t001_Table 1
######
Effect of nitrogen supply on biomass and nitrogen accumulation in maize plants at the V6 stage.
N Treatment Total Biomass Shoot Biomass Root-to-Shoot Ratio Total N Content Shoot N Content
--------------- --------------- --------------- --------------------- ----------------- -----------------
High Nitrogen 2.74 a 2.05 a 0.34 b 132.26 a 101.33 a
Low Nitrogen 2.32 b 1.54 b 0.51 a 33.89 b 23.48 b
Note: Different letters in the same column indicate significant differences between nitrogen treatments (*p* \< 0.05).
| {
"pile_set_name": "PubMed Central"
} |
Photosynthesis is a vital process for plant development and survival which is precisely regulated at several genetic, molecular and physiological levels during the plant growth cycle[@b1]. Photosynthetic pigments (chlorophylls and carotenoids) are a group of molecules that belong to a large class of compounds of different biochemical nature, the isoprenoids. Isoprenoids comprise a chemically and functionally wide heterogeneous group including more than 30,000 molecules[@b2]. In addition to photosynthesis, plants produce many isoprenoids whose functions are essential in developmental processes such as growth regulation (gibberellins, cytokinins, brassinosteroids and abscisic acid), defense mechanism (phytoalexins), membrane structure (sterols) and redox chemistry (plastoquinone, ubiquinone). Despite their diversity, all isoprenoids are formed from two common precursors, isopentenyl diphosphate (IPP) and its isomer dimethylallyl diphosphate (DMAPP)[@b3]. In higher plants, there are two distinct cellular compartmentalized pathways for the biosynthesis of IPP and DMAPP[@b4][@b5][@b6][@b7]: the cytosolic mevalonate (MVA) pathway, which is shared with fungi, yeast, and animals[@b8][@b9], and the non-mevalonate pathway designated as the 2-*C*-methyl-D-erythritol-4-Phosphate (MEP) pathway, located in plastids and also occurring in eubacteria, green algae and protozoa[@b10][@b11][@b12][@b13][@b14]. The initial step of the MEP pathway, a transketolase-type condensation of pyruvate with D-glyceraldhyde-3-phosphate to form 1-deoxy-D-xylulose-5-phosphate (DXP), is catalyzed by DXP synthase (DXS)[@b12]. The cloning and characterization of the *DXS* gene was first described in *Escherichia coli*[@b15]. *DXS* homologous genes were subsequently cloned from several plant species, including peppermint[@b16], pepper[@b17], Arabidopsis[@b18] and tomato[@b19]. Since first reported, the *DXS* gene has been studied in diverse plant species in order to gain deeper knowledge of the DXS family, and functional diversification has been suggested in this family due to the existence of at least three functionally specialized types of DXS, i.e. DXS1, DXS2 and DXS3[@b20][@b21][@b22]. Despite high amino acid sequence homology among DXS proteins, mainly between DXS1 and DXS2, not all *DXS*-like genes have the same expression pattern during plant development. DXS1 has been proposed to be specifically involved in the synthesis of essential isoprenoids (such as photosynthetic pigments and phytohormones) in chloroplasts[@b18][@b19]. In turn, DXS2 is thought to be located in non-photosynthetic plastids, involved in the synthesis of specific isoprenoids as secondary metabolites related to mycorrhiza[@b23]. In addition, it has been suggested that DXS3 might participate in the synthesis of essential isoprenoids required at low levels, such as phytohormones[@b20].
Many isoprenoids, carotenoids among them, have gained importance as biotechnological tools as they show beneficial effects on human health[@b24][@b25]. Since DXS1 is the key enzyme that limits output from the MEP pathway in plants[@b19][@b26][@b27], changes in *DXS1* expression may directly influence the content of photosynthetic pigments. Many experiments have therefore been carried out to enhance the level of some isoprenoids in model and crop species, mainly by the generation of *DXS* overexpression transgenic lines[@b28][@b29]. Indeed, constitutive expression of a bacterial *DXS* gene under the control of the CaMV 35S promoter in transgenic tomato lines did alter the isoprenoid end products in fruits, but not in leaves[@b28]. Similarly, the overexpression of *DXS1* in Arabidopsis[@b26] and the constitutive expression of the soybean *DXS1* gene in tobacco transgenic plants[@b29] produced increases in various isoprenoids, including total chlorophylls and carotenoids. However, results from *DXS1* overexpression experiments differed among plant species, including tomato, indicating a complex regulation of DXS activity[@b28]. Additionally, characterization of *DXS* mutants has shed light on the key function of DXS proteins in plant development[@b26][@b30]. In Arabidopsis, DXS1 plays an essential role in chloroplast development during leaf cell maturation[@b30][@b31], while DXS2 seems to have no DXS activity, but it has acquired another unknown biochemical function[@b32]. In tomato, the DXS1 function is required for carotenoid biosynthesis during fruit ripening[@b19], while tomato DXS2 participates in the biosynthesis of isoprenoids in trichomes[@b16][@b22][@b33] and of secondary metabolites[@b16]. Such results suggest that differences in the biochemical roles of DXS proteins seem to be related with variable functions in plant development and growth.
In this work, we report the identification and molecular characterization of the tomato *white lethal seedling*-2297 (*wls*-2297) mutant, which was isolated from the screening of a T-DNA mutant collection generated by an enhancer trapping gene construct. Seedlings of the *wls-2297* mutant showed albino phenotype and were unable to develop any true leaves from the shoot apical meristem; as a consequence, albino seedlings died after reaching the fully expanded cotyledon stage. Expression and functional analyses demonstrated that these phenothypic alterations were caused by the loss-of-function of the *DXS1* gene, indicating that, besides its role in fruit carotenoid biosynthesis, DXS1 activity is required at early stages of plant development and determines plant survival.
Results
=======
The *wls-2297* mutant showed significant alterations in pigment content and plant development
---------------------------------------------------------------------------------------------
A tomato T-DNA mutant collection generated using an enhancer-trapping construct was screened for developmental traits in TG2 plants grown in soil. As a result, the *wls-2297* mutant was selected due to the noteworthy albino phenotype of its cotyledons. After seed germination, mutant seedlings expanded cotyledons, but they did not fully elongate and never showed the characteristic green color associated with chlorophyll biosynthesis ([Fig. 1a,b](#f1){ref-type="fig"}). Development of the shoot apical meristem was also arrested in the *wls-2297* mutant, as a few days after the cotyledons emerged seedlings experienced premature senescence and died without developing any true leaves. In general, albino phenotypes commonly result from alterations in the photosynthetic process, either directly affecting chloroplast development and functionality, or indirectly, as a consequence of mutations that promote significant defects in plant development or metabolism[@b31]. In a first step toward determining the metabolic pathway affected in the *wls-2297* mutant, seedlings were grown *in vitro* using a standard Murashige and Skoog (MS) culture medium supplemented with a high sugar concentration. Under these conditions, *wls-2297* mutant seedlings survived 3--5 days longer than in soil but finally died; during this period, they developed slower than wild type (WT) seedlings and developed no true leaves ([Fig. 1c,d](#f1){ref-type="fig"}).
With a view to verifying whether the *wls-2297* mutation is involved in chloroplast development, fresh cotyledon tissues were observed under bright field microscopy, which revealed that cells of *wls-2297* cotyledons had an apparent lower density of chloroplasts than WT ones. This suggested decreased chlorophyll content of mutant cotyledons, in line with the pale green pigmentation of *wls-2297* chloroplasts as compared with WT ones ([Fig. 1e,f](#f1){ref-type="fig"}). Therefore, chlorophyll and carotenoid contents were measured in WT and mutant cotyledons. As expected, results showed that levels of both chlorophylls and carotenoids were significantly lower in *wls-2297* mutant cotyledons ([Fig. 1g](#f1){ref-type="fig"}), proving that the albino mutant phenotype was caused by a severe deficiency in photosynthetic activity.
Molecular cloning and characterization of the *wls-2297* mutation
-----------------------------------------------------------------
Genetic analysis carried out on a segregating TG2 progeny of 121 plants indicated that the *wls-2297* mutant phenotype was inherited as a monogenic recessive trait (3:1 ratio, 84 WT:37 mutant; χ^2^ = 2.01; *P* = 0.16). This result agreed with that of Southern blot hybridization, which showed that a single T-DNA copy segregated in the *wls-2297* TG2 family here analyzed ([Fig. 2a](#f2){ref-type="fig"}). Moreover, 88 TG2 seedling plants were screened *in vitro* for kanamycin resistance, and 22 of them displayed a *wls-2297* mutant phenotype. A complete co-segregation was found between *wls-2297* mutant and kanamycin resistance phenotypes in this segregating population (1:2:1 ratio; 22 mutant and kanamycin resistant:41 WT and kanamycin resistant:25 WT and kanamycin susceptible; χ^2^ = 0.61; *P* = 0.74), indicating that a single T-DNA insertion was responsible for the albino phenotype showed by the *wls-2297* mutant.
The genomic region flanking the T-DNA insertion was cloned and the sequences obtained were compared with the reference sequence of tomato genome (<https://solgenomics.net>). Results showed that the T-DNA is located on chromosome 1, the right border (RB) of the T-DNA being located at position 76834.3 kb, 1528 bp downstream of a gene encoding a putative PEROXIDASE enzyme (*POX, Solyc01g067850*). In addition, the left border (LB) of the T-DNA is placed at position 76872.9 kb, 217 bp upstream from the *DXS1* gene (*Solyc01g067890*) encoding for 1-deoxy-D-xylulose-5-phosphate synthase. According to this *in silico* analysis, the T-DNA insertion causes a 38.6 kb-deletion in the genome of the *wls-2297* mutant, which affects four genes, three putative genes encoding for POX enzymes (*POX24-Solyc01g067860, POX24-Solyc01g067870* and *POX27-Solyc01g067880*) and the *DXS1* gene ([Fig. 2b](#f2){ref-type="fig"}).
In order to demonstrate that the 38.6 kb-deletion was the only mutation promoted by the T-DNA insertion and that no other chromosome rearrangements (e.g. translocation) occurred during mutagenesis, specific primers of the four genes located on the 38.6 kb genomic region were designed and used in PCR experiments to confirm their absence in *wls-2297* mutant plants. PCR products specific to the four reported genes from WT had the expected size and their sequences coincided with those reported in the tomato genome database (<https://solgenomics.net>), whereas no amplification products were obtained from the *wls-2297* mutant DNA ([Fig. 2c](#f2){ref-type="fig"}). On the other hand, a long-PCR assay was performed to amplify the region between the two previously cloned T-DNA flanking regions. The long-PCR product obtained from *wls-2297* genomic DNA had the expected size of 5.2 kb ([Fig. 2d](#f2){ref-type="fig"}) and its sequence confirmed the presence of a single T-DNA insertion causing a 38.6 kb-deletion in the genome of the *wls-2297* mutant. A co-segregation analysis performed using allele-specific primers designed from both the T-DNA and the genomic flanking sequences tagged in the *wls-2297* mutant revealed that the T-DNA insertion co-segregated with the *wls-2297* mutant phenotype ([Fig. 2e](#f2){ref-type="fig"}). Taken together, these results provide strong evidence as to the genetic and molecular nature of the *wls-2297* mutant phenotype, which may be due to a 38.6 kb-deletion promoted by a single copy T-DNA insertion that affects three *POX* genes and the *DXS1* gene.
Gene silencing experiments
--------------------------
Given the role of the DXS1 enzyme in isoprenoid biosynthesis through the MEP pathway[@b18] and the pigment measurements in the present work, the *DXS1* gene was considered the best candidate gene to be responsible for the *wls-2297* mutant phenotype. In order to demonstrate this hypothesis, we generated *DXS1* silencing lines in the cultivar Moneymaker using an RNA interference strategy (RNAi). After ploidy level determination, nine RNAi *DXS1* diploid lines showing a significant reduction of *DXS1* transcript levels ([Fig. 3a](#f3){ref-type="fig"}) were further characterized under *in vitro* conditions. Interestingly, all *DXS1* repressed lines showed an albino phenotype, and even developed young leaves when sucrose was added to the MS medium, allowing explant development ([Fig. 3b,c](#f3){ref-type="fig"}). Moreover, neither RNAi *DXS1* explants nor seedlings could be propagated since they died at early developmental stages. Therefore, the silencing of *DXS1* leads to similar developmental changes to *wls-2297* mutation, which provides strong evidence that the *wls-2297* mutant phenotype was due to the lack of *DXS1* gene function.
In addition, RNAi lines simultaneously silencing the three *POX* genes located in the deleted region were generated by using a gene construct that included a consensus sequence shared by the three *POX* messengers ([Fig. S2b](#S1){ref-type="supplementary-material"}). Fifteen *POX* silencing diploid lines were selected for further analysis as they showed a significant decrease of *POX* transcripts ranging from 63 to 82% with respect to the control background ([Fig. 4a](#f4){ref-type="fig"}). All of the RNAi *POX* lines developed normally, although plant vigor was slightly affected most likely due to peroxidase deficiency ([Fig. 4b](#f4){ref-type="fig"}). Most importantly, none of these RNAi lines showed phenotype changes resembling those of the *wls-2297* mutant ([Fig. 1a--d](#f1){ref-type="fig"}), which indicates that *POX* genes affected by the T-DNA insertion are not involved in the mutant phenotype here reported.
Expression patterns of the *DXS* gene family
--------------------------------------------
Tomato *DXS1* gene, which is deleted in the *wls-2297* mutant, is located on chromosome 1; according to the tomato genome database (<https://solgenomics.net>), it has a genome size of 4.2 kb and includes 10 exons and 9 introns. In addition to *DXS1*, two other tomato genes belonging to the *DXS* family have been identified, i.e. *DXS2*, located on chromosome 11 (*Solyc11g010850*) and *DXS3* located on chromosome 8 (*Solyc08g066950*), both also composed of 10 exons and 9 introns. The ORFs from the three genes, *DXS1, DXS2*, and *DXS3* code for proteins of 719, 714 and 709 amino acid residues, respectively ([Fig. S3](#S1){ref-type="supplementary-material"}). Phylogenetic analysis performed with the three tomato DXS proteins as well as representative DXS protein of diverse plant species showed three independent groups of DXS proteins conserved among plants, i.e. DXS1, DXS2 and DXS3 clades ([Fig. S4](#S1){ref-type="supplementary-material"}), a result which agrees with those previously described by Hofberger *et al*.[@b34]. In addition, the overall similarity among the three DXS tomato proteins ranges between 59 and 72%, DXS3 being the most divergent. The three proteins have conserved the three characteristic domains of the DXS family, these are an N-terminal thiamine pyrophosphate (TPP)-binding module, a pyrimidine binding domain and a transketolase C-terminal domain ([Fig. S3](#S1){ref-type="supplementary-material"}).
With a view to analyzing the spatial expression pattern of the *DXS1* tomato gene, as well as the other two members of the *DXS* gene family, quantitative RT-PCR experiments have been carried out ([Fig. 5a](#f5){ref-type="fig"}). A preliminary picture showed that all three *DXS* genes are expressed in a wide range of tissues, but a very low level of *DXS3* transcripts was found in comparison to *DXS1* and *DXS2* transcripts. *DXS1* expression was detected in both vegetative, i.e. root and leaf, and reproductive organs, including floral buds, flowers and fruits at several stages of development ([Fig. 5a](#f5){ref-type="fig"}). *DXS1* transcript accumulation was high during flower development and later decreased sharply in fully developed green fruits (MG stage; [Fig. 5a](#f5){ref-type="fig"}). Expression of *DXS2* was also found in leaf and flowers at several developmental stages, and particularly in flowers at anthesis; however, no *DXS2* transcripts were detected in roots, and only a very low number in MG fruits ([Fig. 5a](#f5){ref-type="fig"}). These last two features, i.e. lack of expression in roots and a significant reduction of transcript level in MG fruits, also characterized the expression pattern shown by the *DXS3* gene ([Fig. 5a](#f5){ref-type="fig"}).
The analysis of *DXS* transcript levels in the seedlings of the *wls-2297* mutant and in the *DXS1* RNAi silencing lines showed that *DXS1* expression was undetectable in the former, while it was silenced in the two RNAi lines analyzed ([Fig. 5b](#f5){ref-type="fig"}). Though *DXS2* expression was not significantly affected in the *wls-2297* mutant, it decreased in the RNAi *DXS1* lines ([Fig. 5c](#f5){ref-type="fig"}), despite the fact that the RNAi construct was specifically designed for silencing the *DXS1* gene ([Fig. S2a](#S1){ref-type="supplementary-material"}). Finally, *DXS3* transcript levels were not significantly altered in the RNAi *DXS1* lines, but were slightly increased in *wls-2297* mutant seedlings ([Fig. 5d](#f5){ref-type="fig"}).
*In vivo* and molecular complementation of the *wls-2297* mutant phenotype
--------------------------------------------------------------------------
To corroborate the functional role of *DXS1* gene in determining plant development and survival, two complementation experiments were performed. Thus, *in vivo* complementation of the *wls-2297* mutant phenotype was attempted by treatment of seedling plants grown under *in vitro* conditions with 1-Deoxy-D-xylulose-5-phosphate (DXP), the product of the DXS enzyme activity. When DXP was added to the *in vitro* culture media, mutant seedlings not only did not recover the wild type phenotype, rather they died even faster than *wls-2297* mutant ones ([Fig. 6a,b](#f6){ref-type="fig"}). On the contrary, daily treatment with DXP on the shoot apical meristem led the mutant seedlings to develop green live tissues from those cells which incorporated DXP ([Fig. 6c](#f6){ref-type="fig"}), and they were even able to develop small green leaves ([Fig. 6e,f](#f6){ref-type="fig"}), suggesting that DXP is essential for photosynthetic activity of plant tissues. Nevertheless, it is noteworthy that plant development ceased and seedlings died a few days after DXP treatment ceased ([Fig. 6d](#f6){ref-type="fig"}).
In addition, the *DXS1* gene was constitutively expressed under a 35S promoter in *wls-2297* mutant seedlings to ascertain whether increased *DXS1* transcript levels were able to rescue the wild type phenotype. Results showed that transgenic explants develop green tissues from albino tissue sections ([Fig. 6g](#f6){ref-type="fig"}), confirming that the lack of photosynthetic activity displayed by the *wls-2297* albino mutant can be firstly reverted by *DXS1* expression. However, as the overexpressing *DXS1* explants grew, tissues turned yellow and finally acquired an albino phenotype and died.
Effects of *DXS1* overexpression on tomato fruit pigmentation
-------------------------------------------------------------
It has been proposed that the protein coded by *DXS1*, the plastidic form of 1-Deoxy-d-xylulose-5-phosphate synthase, is the limiting enzyme for isoprenoid biosynthesis[@b26]. In order to determine whether the *DXS1* gene is a limiting step in the MEP pathway of tomato plants, nine diploid transgenic lines overexpressing *DXS1* cDNA (OX lines) were generated in the cultivar Moneymaker using a constitutive 35S promoter gene construct, and subsequently characterized in greenhouse conditions ([Fig. 7](#f7){ref-type="fig"}). With the exception of the OX7 line, no changes in the colour and lycopene concentration were observed in the fruits of 35S::*DXS1* lines even though they expressed the *DXS1* gene at a higher level than control plants ([Fig. 7a](#f7){ref-type="fig"}). Contrary to expectations, tomato fruits yielded by the OX7 line showed a pale colour during the whole ripening process ([Fig. 7b](#f7){ref-type="fig"}). The lack of red colour in mature fruits was most probably due to the low level of lycopene pigment detected in the OX7 transgenic line ([Fig. 7c](#f7){ref-type="fig"}). Analysis of *DXS1* relative expression carried out in tomato fruits at breaker (BR) and red ripe (RR) stages of development showed that *DXS1* was not overexpressed, but rather it was repressed in these OX7 fruits, suggesting that the pale colour and low lycopene concentration displayed by this line was caused by the post-transcriptional gene silencing (PTGS) of the *DXS1* gene ([Fig. 7d](#f7){ref-type="fig"}). Indeed, PTGS triggered by the 35S promoter have widely been reported in transgenic plants[@b35]. However, the fact that T-DNA integration site in the OX7 line might be responsible for the altered phenotype cannot be definitively ruled out.
Influence of *DXS1* overexpression and silencing on the MEP pathway
-------------------------------------------------------------------
The MEP pathway operates through the participation of eight consecutive enzymes to generate IPP and DMAPP, which serve as the basis for the biosynthesis of all isoprenoid compounds[@b36][@b37]. To investigate the effect of *DXS1* overexpression and silencing on the MEP pathway, the expression of the genes participating in this pathway was analyzed by qRT-PCR in the three transgenic lines showing either the highest or lowest *DXS1* expression levels ([Fig. S5a](#S1){ref-type="supplementary-material"}).
Within the MEP pathway, DXS generates DXP by the transketolase-type condensation of pyruvate and D-glyceraldehyde 3-phosphate, which is converted to MEP by the enzyme coded by the *DXP* reductoisomerase gene (DXR, *Solyc03g114340*), whose expression profile remained quite similar in wild type and *DXS1* overexpression lines ([Fig S4b](#S1){ref-type="supplementary-material"}). However, *DXS1* RNAi lines showed a decreased tendency in *DXR* transcript level compared to wild type plants ([Fig S4b](#S1){ref-type="supplementary-material"}). Subsequently, MEP is converted into IPP and DMAPP by the consecutive action of five independent enzymes: i.e. 2-C-methyl-D-erythritol 4-phosphate cytidyltransferase (MCT, *Solyc01g102820*), 4-diphosphocytidyl-2-C-methyl-D-erythritol kinase (CMK, *Solyc01g009010*), 2-C-methyl-D-erythritol 2,4-cyclodiphosphate synthase (MDS, *Solyc08g081570*), 4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase (HDS, *Solyc11g069380*), and 4-hydroxy-3-methylbut-2-enyl diphosphate reductase (HDR, *Solyc01g109300*). Expression analysis carried out in *DXS1* overexpression lines showed similar expression profiles between wild type and transgenic lines in the transcript levels of the genes coding for these five enzymes ([Fig. S5c--g](#S1){ref-type="supplementary-material"}). On the other hand, the transcript levels of *MCT, CMK* and *MDS* genes were remarkably reduced in *DXS1* RNAi lines ([Fig. S5c--g](#S1){ref-type="supplementary-material"}). However, *HDS* and *HDR* genes showed certain variability in expression levels across these silencing lines [Fig. S5c--g](#S1){ref-type="supplementary-material"}).
In the final step of the MEP pathway, geranylgeranyl diphosphate (GGPP), a common precursor for the synthesis of phyllochinones, tocopherols, plastoquinones, chlorophylls, gibberellins and carotenoids, is generated by geranylgeranyl diphosphate synthase (GGPS, *Solyc11g011240*) that catalyzes the condensation of three IPP and one DMAPP units. Expression levels of the *GGPS* gene in *DXS1* RNAi lines showed variable levels of expression ([Fig. S5h](#S1){ref-type="supplementary-material"}), similar to those observed for *HDS* and *HDR* genes; whereas slight differences were found in *DXS1* overexpression lines compared to wild type plants ([Fig. S5h](#S1){ref-type="supplementary-material"}).
Discussion
==========
In recent decades, insertional mutant collections have been particularly useful for identifying and characterizing genes of interest in different model species such as Arabidopsis, rice and tomato[@b38][@b39][@b40]. This work reports the isolation and functional analysis of the tomato *wls-2297* T-DNA mutant, whose characterization led us to demonstrate once again the value of T-DNA insertional mutant libraries for gene discovery in a species of agronomic interest such as tomato. The *wls-2297* mutant was selected based on its albino phenotype and the arrested development at expanded cotyledon stage. Molecular analyses performed on the *wls-2297* mutant revealed that the observed albino phenotype is due to a single T-DNA insertion, whose integration in the tomato genome caused a 38.6 kb-deletion at chromosome 1 that resulted in the loss of *DXS1* and three *POX* genes. Phenotypic and expression analyses of RNAi silencing lines led us to conclude that *wls-2297* is a knock-out mutant affected in the *DXS1* gene, and that the loss of DXS1 activity is responsible for the *wls-2297* mutant phenotype. Indeed, this result was further confirmed by means of both *in vivo* complementation assays with DXP and *DXS1* overexpression on the *wls-2297* mutant, which partially rescued the albino phenotype. At early stages of plant development, the *wls-2297* phenotype was completely rescued; however, over time the leaves of the *wls-2297* mutant plants overexpressing the *DXS1* gene turned yellow, became albino and eventually died prematurely. Functional assays based on *DXS1* overexpression showed dissimilar results in different plant species, ranging from no changes in isoprenoid content[@b28][@b41] to unequal increases in various isoprenoid compounds including total chlorophylls and carotenoids[@b26][@b29]. Overall, these previous works have proved that constitutive *DXS1* overexpression does not necessarily mean high levels of isoprenoid end products. Indeed, it has been found that *DXS1* transcript levels in overexpression lines, and therefore isoprenoid levels, are usually not so high as might be expected[@b26][@b28][@b29][@b41]. In this study, expression analysis of the MEP pathway genes suggested that *DXS1* overexpression did not significantly alter transcript profiles of the MEP pathway since gene expression levels in *DXS1* overexpressing lines were quite similar to those found in wild type plants ([Fig. S5](#S1){ref-type="supplementary-material"}). However, different expression profiles of the MEP pathway genes were found in the *DXS1* silencing lines. Thus, the expression levels of *DXR, MCT, CMK* and *MDS* genes were significantly reduced in albino RNAi transformants compared to wild type non-transformed plants ([Fig. S5b--e](#S1){ref-type="supplementary-material"}), suggesting a role for DXS in transcriptional regulation of the first steps of the MEP pathway. On the other hand, expression profiles of *HDS, HDR* and *GGPS* genes showed an evident variation among different *DXS1* RNAi lines ([Fig. S5f--h](#S1){ref-type="supplementary-material"}). Most likely, the explanation for these results relies on the complex regulation of the MEP pathway, which occurs at both transcriptional and post-transcriptional levels[@b42] to ensure its proper functioning. Due to this pathway complexity, it has been suggested that elevated levels of *DXS1* transcripts can only exert an influence on the MEP pathway when adequate precursors from intermediary metabolism are accessible[@b28]. Given that levels of the different isoprenoids do not change equally in *DXS1* overexpressing lines[@b26][@b28][@b29][@b41], the fact that overexpression of *DXS1* did not completely rescue the albino phenotype of the *wls-2297* mutant might be due to disproportionately altered isoprenoid levels. However, the possibility that some *wls-2297 DXS1* overexpressing lines here reported may undergo co-suppression effects of *DXS1* expression cannot be discarded, as it is a commonly reported phenomenon in genetic transformation experiments using strong constitutive promoters[@b35]. Unfortunately, we were not able to perform expression studies to corroborate this hypothesis, as 35S::*DXS1* plants died prematurely.
In Arabidopsis, the *DXS1* knock-out mutants *cla1-1*[@b31] and temperature-sensitive allele *chs5*[@b30] also showed an albino phenotype as well as altered plant development. When both mutants were grown in a nutrient medium supplemented with glucose, they were able to sustain their development while maintaining their albino phenotype. Thus, *chs5* plants developed slowly but reached a similar size to wild type plants[@b30], whereas the *cla1-1* mutant even generated floral buds, which did not however manage to produce normal flowers[@b31]. In this study, tomato RNAi *DXS1* lines developed three-four albino leaves, probably as a consequence of residual *DXS1* expression; however, the *wls-2297* mutant did not develop true leaves, but rather its growth was arrested at the fully expanded cotyledon stage or slightly later when grown in a nutrient medium supplemented with a high sugar concentration. These results suggest a significant difference between DXS1 action mechanisms in Arabidopsis and tomato. Araki *et al*.[@b30] suggested that a plant with a loss of DXS activity did not survive past the seedling stage due to the lack of thiamine, one of the first branches of the MEP pathway[@b43]. They therefore proposed that the partial development of albino *chs5* plants was facilitated by other *DXS* genes[@b30]. Arabidopsis *DXS1* and *DXS2* genes show high homology and are both included in the subfamily DXS1[@b34], suggesting that they would have redundant functions. However, the Arabidopsis *DXS2* gene has recently been renamed as *DXL1*, which might not have DXS activity, but rather has acquired a novel, as yet unknown biochemical function[@b32]. In contrast, the Arabidopsis *DXS3* gene has been grouped within a clade composed of highly divergent protein sequences[@b20], which lack many of the highly conserved residues known to be essential for DXS activity[@b44]. Nonetheless, marginal, though significant, DXS activity has recently been reported for the corresponding ortholog DXS3 in maize[@b20]. Hence, if the DXS3 function was conserved in other species, the survival of Arabidopsis *chs5* and *cla1-1* mutants might be explained by the DXS activity of the DXS3 enzyme. As in Arabidopsis, the tomato DXS family also consists of three genes, but in contrast to Arabidopsis, tomato *DXS* genes code for proteins belonging to each one of the proposed DXS clades ([Fig. S4](#S1){ref-type="supplementary-material"}), and in addition they have different expression patterns. The tomato *DXS1* gene is ubiquitously expressed and its highest transcript levels are observed during fruit ripening[@b34][@b45]. However, *DXS2* transcripts are abundant in only few tissues, including young leaves and flowers, as well as isolated trichomes[@b34][@b45]. On the other hand, *DXS3* transcript levels are extremely low compared to the other two tomato *DXS* genes according to previous results[@b34]. The roles of tomato DXS1 and DXS2 have been described without overlapping function in modulating isoprenoid metabolism[@b16][@b22][@b33][@b45]. Our results are in agreement with this hypothesis, since DXS2 and DXS3 activities did not compensate for the loss of DXS1 function in the *wls-2297* mutant, indicating that DXS1 plays a vital role in the development of tomato plants.
Since the non-mevalonate pathway of isoprenoid biosynthesis was described, the exchange of isoprenoid intermediate compounds between MVA and MEP pathways has been a matter of some dispute. Certain evidence of cross-talk between MVA pathway in the cytosol and MEP pathway in the plastid has been reported in different plant species, including Arabidopsis and a solanaceous species like tobacco[@b46][@b47][@b48]. However, it has also been proposed that a tightly regulated cellular compartmentalization of MVA and MEP pathways occurs during isoprenoid biosynthesis[@b28]. In addition, a tissue-dependent regulation of tomato MEP genes has recently been reported supporting the complexity of this pathway[@b49]. Characterization of the *wls-2297* mutant, which displays a complete inhibition of DXS activity due to a loss-of-function of the *DXS* gene, showed a drastic reduction in chlorophyll and carotenoid biosynthesis ([Fig. 1g](#f1){ref-type="fig"}), a feature that should not be detected if there were an active exchange of isoprenoid intermediates between both pathways. Therefore, our results not only support that the MEP pathway is essential for photosynthetic pigment biosynthesis, but also that functional cross-talk with the MVA pathway may not occur at least during the first steps of isoprenoid biosynthesis. Although periodic supplementation of *wsl-2297* mutant seedlings with DXP during several days promoted minor development, it did not avoid death of the seedlings once the treatment ceased. This result suggests that if an exchange of isoprenoid intermediates such as IPP and DMAPP occurs in tomato chloroplasts, it must be at a rate that is insufficient to allow plant growth. In addition, this hypothesis agrees with the blocking of plastidial isoprenoid biosynthesis observed after treatment with fosmidomycin, a specific inhibitor of DXP reductoisomerase, the enzyme catalyzing the previous step to that of DXS in the isoprenoid biosynthetic pathway[@b50]. Taken as a whole, these results regarding the lethality of the *wls-2297* mutation, also found in *DXS1* silencing lines, prove that the *DXS1* gene is a crucial genetic factor for plant organ growth and survival.
Methods
=======
Plant material and growth conditions
------------------------------------
The *wls-2297* mutant was isolated from a T-DNA insertional collection, which was generated from the tomato (*Solanum lycopersicum* L.) cultivar Moneymaker through infection with *Agrobacterium* strain LBA4404 containing the enhancer trap vector pD991[@b51] (kindly supplied by Dr. Thomas Jack; Department of Biological Sciences, Dartmouth College, USA). The TG1 line was self-pollinated to obtain the TG2 progeny, which was grown in soil conditions (a mixture of peat and coconut fiber (1:1) and, after sowing, a cover of vermiculite). Due to the fact that the *wls-2297* mutant homozygous plants did not survive enough to achieve fruit set, *wls-2297* line were perpetuated by self-pollination of wild type heterozygous plants growth in soil under greenhouse conditions using standard practices with regular addition of fertilizers. Seedlings analysed in this project were grown under standard conditions (16 h light/8 h darkness, ≈ 25 °C, 60% RH) in a growth chamber. Co-segregation studies of the mutant phenotype with kanamycin resistance conferred by the *NEOMYCIN PHOSPHOTRANSFERASE III (NPTII*) marker gene was carried out by sowing seeds on Murashige and Skoog[@b52] (MS) agar medium supplemented with sucrose (10 g l^−1^) and kanamycin (100 mg l^−1^). The sodium salt of 1-Deoxy-D-xylulose-5-phosphate (DXP, Sigma-Aldrich) resuspended in water (1 mg/mL) was used to perform *in vivo* complementation of the albino phenotype in the *wls-2297* mutant.
Microscopy analysis of chloroplasts and pigment measurement
-----------------------------------------------------------
For observation of chloroplasts/plastids in living leaf tissues, a section of tomato cotyledon was excised and its cuticle removed manually. Then, these cotyledons were examined using a bright field microscope Nikon Optiphot-2 equipped with Nikon Digital Camera. The chlorophylls and carotenoids concentrations were determined by UV-VIS Spectroscopy following the method described by Lichtenthaler and Buschmann[@b53].
DNA isolation and DNA-blot hybridization
----------------------------------------
Genomic DNA was isolated from 100 mg of young leaves using Plant DNAzol Reagent (Invitrogen) following the instructions provided by the manufacturer. In order to determine the number of T-DNA insertions introduced in the *wls*-*2297* mutant, a DNA-blot hybridization was carried out from 15 μg of genomic DNA digested by restriction enzymes *Eco*RI and *Hin*dIII, electrophoresed through-out a 0.8% agarose-TBE gel, blotted onto Hybond N+ membrane (Amersham) and hybridized with a *FA*-*NPTII* radioactive probe, which was labelled with ^32^P by random priming with the Rediprime II Random Prime Labelling System (Amersham). Nylon membranes hybridized were exposed to Hyperfilms (Amersham). The *FA-NPTII* probe was synthetized including the complete coding sequence of the tomato *FALSIFLORA (FA*) gene, which was employed as hybridization positive control, and the *NPTII* gene used as a selective marker in the T-DNA.
Cloning of T-DNA flanking sequences and PCR genotyping
------------------------------------------------------
Sequences flanking the *wls-2997* T-DNA insertion were cloned by amplification with a single-side specificity (anchor-PCR) assay described by Loh[@b54]. With this end, genomic DNA was digested with restriction endonucleases that generate blunt ends, i.e. *Alu*I, *Dra*I, *Eco*RV, *Hin*cII, *Pvu*II, *Sca*I, *Stu*I and *Sma*I (Takara Bio company). After digestion, adaptors were linked to the digested genomic DNA restriction fragments. The fragments linked to adaptors were then amplified by three assembled PCR with specific primers for the Right Border sequence (Anchor Right-1, Anchor Right-2 and Anchor Right-3; [Table S1](#S1){ref-type="supplementary-material"}) or for the Left Border (Anchor Left-1, Anchor Left-1 and Anchor Left-1; [Table S1](#S1){ref-type="supplementary-material"}) from the T-DNA, combined with adaptor specific primers Adaptor-1, Adaptor-2 and Adaptor-3 ([Table S1](#S1){ref-type="supplementary-material"}). PCR amplification products were purified using GenElute PCR Clean-up Kit (SIGMA Aldrich) and sequenced in an Applied Biosystems 3500 Genetic Analyzer. The cloned sequences were compared with SGN Database (<https://solgenomics.net>) to assign the T-DNA insertion site on tomato genome. Co-segregation of the T-DNA insertion site with the *wls-2997* phenotype was evaluated by PCR using i) the specific genomic forward and reverse primers to amplify the wild type allele (without T-DNA insertion) and ii) one specific genomic primer and the specific T-DNA primer to amplify the mutant allele (carrying the T-DNA insertion). The sequence of genotyping primers used is listed in [Table S1](#S1){ref-type="supplementary-material"}.
Specific primers for *DXS1 (Solyc01g067890*) and the three *POX (Solyc01g067860, Solyc01g067870, Solyc01g067880*) genes were designed ([Table S1](#S1){ref-type="supplementary-material"}) and used in PCR experiment to demonstrate that a 38.6 kb-deletion is on *wls-2297* mutant plants. Amplification was performed using in a volume of 30 μl using 10 ng of total DNA, 50 ng of each primer, 0.25 mM dNTPs, 2.5 mM MgCl~2~, and 1 U BIOTAQ^TM^ DNA Polymerase (Bioline) in 1X Taq buffer. DNA was amplified under the following thermal cycling conditions: 94 °C for 5 min, followed by 35 cycles at 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 2 min, and a final extension of 5 min at 72 °C. PCR products were analysed in 1% agarose gels in 1X SB buffer (200 mM NaOH, 750 mM boric acid, pH 8.3) and visualized with ethidium bromide. In addition, to amplify the genomic region located between the T-DNA flanking regions a Long-PCR assay was carried out using Elongase^®^ Enzyme Mix (Invitrogen) following the instructions given by the manufacturer. The primers used in this experiment were Genotyping-F1 and Genotpyping-R2 ([Table S1](#S1){ref-type="supplementary-material"}). Genomic DNA from wild type (cv. Moneymaker) and the *wls-2297* mutant plants was used as template.
RNA preparations and gene expression analysis
---------------------------------------------
Total RNA was purified from using Trizol reagent (Invitrogen) following the manufactures instructions. Contamination of genomic DNA was removed by using the DNA-free^TM^ kit (Ambion), and cDNA was synthetized from the DNA-free total RNA by means of a M-MuLV reverse transcriptase (Invitrogen) using as primers a mixture of random hexamer and 18-mer oligo(dT).
Expression analysis were performed by real-time PCR assays using the SYBR Green PCR Master Mix kit in a 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and a pair of specific primer pairs for each gene analysed ([Table S1](#S1){ref-type="supplementary-material"}). Data obtained were collected and analysed using the System Sequence Detection Software v1.2. Results were calculated using ΔΔCt method by comparison to a data point from the wild type samples and are expressed in arbitrary units. The housekeeping *UBIQUITINE3* gene was used as an internal reference in the gene expression analyses performed. Absence of genomic DNA contaminating in the RNA sample analysed by RT-PCR was demonstrated as previously described[@b55]. Mean comparison (Fisher's Least Significant Difference test, LSD) was used to determine significant differences among gene expression levels. Analyses were performed using the Statgraphics Centurion XVI software package and data presented as means ± standard error.
Generation of DXS transgenic tomato plants
------------------------------------------
The *DXS1 (Solyc01g067890*) complete open reading frame (ORF) was amplified from tomato cv. Moneymaker cDNA using a pair of specific primers 35SDXS1F and 35SDXSR to introduce a *Bam*HI restriction site 127 bp upstream the initiation codon and a *Kpn*I restriction sites downstream the stop codon ([Table S1](#S1){ref-type="supplementary-material"}). The PCR product was cloned into a p-GEM^®^-T-Easy plasmid and sequenced in order to confirm the absence of mutations in the amplicon. The selected plasmid was digested with *Bam*HI and *Kpn*I, and the *DXS1* cDNA was subcloned under the control of the constitutive CaMV promoter together a plant kanamycin resistant selection marker into the binary vector pROKII to generate an overexpression gene construct used to transform plants of the cv. Moneymaker. The same *DXS1* cDNA was subcloned under the control of the constitutive CaMV promoter together with a plant hygromycin resistant selection marker into the binary vector pCAMBIA1302 to generate an overexpression gene construct used to complement the *wls-2297* mutant seedlings.
In order to silencing the *DXS1* gene, an interference RNA (RNAi) approach was followed. A 330 bp specific fragment of the *DXS1* cDNA was amplified, using primers RNAiDXS1F, to introduce a *Xba*I and a *Xho*I restriction sites, and RNAiDXS1R, to introduce a *Cla*I and a *Kpn*I restriction sites ([Fig. S2a](#S1){ref-type="supplementary-material"}; [Table S1](#S1){ref-type="supplementary-material"}), and cloned into p-GEM^®^-T-Easy. The insert of RNAiDXS1 was liberated from the p-GEM^®^-T-Easy twice, firstly by *Xho*I and *Kpn*I digestion and secondly by *Xba*I and *Cla*I digestion. Both inserts liberated were cloned as inverted repeats into vector pKannibal. The resulting plasmid was digested with *Not*I and fragment isolated was cloned into the binary vector pART27 to express inverted repeat sequences of *DXS1* separated by intronic sequences under the control of the constitutive promoter 35S. In order to generate RNAi lines to inhibit the three *POX* genes (*Solyc01g067860, Solyc01g067870, Solyc01g067880*) simultaneously, a pair of primers annealing in a conserved domain for the three POX, RNAiPOXsF and RNAiPOXsR ([Fig. S2b](#S1){ref-type="supplementary-material"}; [Table S1](#S1){ref-type="supplementary-material"}), were used to generate the silencing construct obtained as described for the *DXS1* gene.
Overexpression and silencing binary vectors were electroporated into *Agrobacterium tumefaciens* and Agrobacterium-mediated transformation of Moneymaker tomato cultivar cotyledons was performed following the protocols previously described[@b56]. The ploidy level in transgenic plants was evaluated by flow cytometry according to the protocol described by Atarés *et al*.[@b57] and diploid transgenic lines were selected for further phenotypic and expression analyses.
Additional Information
======================
**How to cite this article:** García-Alcázar, M. *et al*. Albino T-DNA tomato mutant reveals a key function of 1-deoxy-D-xylulose-5-phosphate synthase (DXS1) in plant development and survival. *Sci. Rep.* **7**, 45333; doi: 10.1038/srep45333 (2017).
**Publisher\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Material {#S1}
======================
###### Supplementary Material
This work was supported by research grants from the Spanish Ministry of Economy and Competitiveness and the UE-European Regional Development Fund (AGL2015-64991-C3-1-R, and AGL2015-64991-C3-3-R), and Junta de Andalucia (P12-AGR-1482). PhD fellowship to M.G.-A. was funded by the FPU Programme of the Spanish Ministry of Science and Innovation. The authors thank research facilities provided by the Campus de Excelencia Internacional Agroalimentario (CeiA3).
The authors declare no competing financial interests.
**Author Contributions** M.G.-A., B.P., C.C., B.G.-S., S.S. and T.A. performed field and laboratory experiments. M.G.-A., E.G., F.Y.-L., J.C. and R.L. wrote the manuscript. J.C., V.M., T.A., and R.L. conceived the research plans, supervised the study and collaborated in data analysis. All authors revised the draft manuscript and read and approved the final manuscript.
{#f1}
{ref-type="supplementary-material"}.](srep45333-f2){#f2}
{#f3}
{#f4}
{#f5}
{#f6}
{#f7}
[^1]: These authors contributed equally to this work.
| {
"pile_set_name": "PubMed Central"
} |
Jean-Paul Iyombe-Engembe, Dominique L Ouellet, Xavier Barbeau, Joël Rousseau, Pierre Chapdelaine, Patrick Lagüe and Jacques P Tremblay
*Molecular Therapy---Nucleic Acids* (2016) **5**: e283; doi:[10.1038/mtna.2015.58](/doifinder/10.1038/mtna.2015.58)
Following publication, the authors noted that in several instances "gRNA 4/54" should have been "gRNA 5/54." A revised article showing the correct number is available as Supplementary Material to this correction.
Supplementary Material {#sup1}
======================
######
Click here for additional data file.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#S0001}
============
Colorectal cancer (CRC) is the second most common cancer in women and the most common cancer in men worldwide.[@CIT0001] With changes in life styles and nutritional habits, the incidence and mortality rates of CRC in China have increased gradually in recent years.[@CIT0002] The effective treatments for CRC are surgery, radiation, and chemotherapy in various combinations, but patients still typically have a poor prognosis, resulting in severe side effects, such as nausea, vomiting, weight loss, and anemia.[@CIT0003] Therefore, there is an urgent need to explore novel therapeutic agents for the treatment of CRC patients.
Increasing evidence has indicated that fungi polysaccharides possess various biological activities, including anti-tumor,[@CIT0004] anti-inflammatory,[@CIT0005] and immunopotentiation[@CIT0006] functions. Fungi polysaccharides, with no or poor toxicity, have attracted much attention in medical research. *R. nigricans*, a zygomycete filamentous fungus, is widely used in the pharmaceutical industry for processes such as biotransformation and production of organic acids.[@CIT0007] We previously reported that the exopolysaccharide (EPS1-1), isolated from the fermentation broth of *R. nigricans*, is a 9682 Da consisting of glucose, mannose, galactose, and fructose in β-anomeric configuration in the molar ratio of 5.89:3.64:3.20:1.00.[@CIT0008] Previous studies showed that EPS1-1 could induce apoptosis in HCT-116 colon cancer cells through the mitochondrial pathway[@CIT0008] and significantly inhibit the occurrence and development of AOM/DSS-induced colorectal cancer in mice.[@CIT0009] In addition, EPS1-1 also possesses immuno-enhancing activities.[@CIT0010] However, mechanistic studies aiming to understand EPS1-1-mediated metabolic alterations in colon cancer leading to cancer prevention have not been reported till date.
In recent years, metabolomics has emerged as a powerful tool for exploring metabolites (with molecular weight \<1000 Da) and to characterize the metabolic phenotype of a biological system. It presents great potential for the investigation of perturbed metabolic profiles in tumor studies, prognostic or predicative interpretation of cancer status, and exploring the mechanisms of cancer drug intervention.[@CIT0011],[@CIT0012] With the emergence of liquid chromatography-mass spectrometry (LC-MS/MS)-based metabolomics, it is possible to profile and even quantify the metabolites found in a pathway. Metabolites have been applied widely in studies of CRC,[@CIT0013] breast cancer,[@CIT0014] lung cancer,[@CIT0015] bladder cancer,[@CIT0016] oral cancer,[@CIT0017] prostate cancer,[@CIT0018] pancreatic cancer,[@CIT0019] and so on. In this study, to investigate the metabolic profiles and potential biomarkers in a mouse model of AOM/DSS-induced CRC, LC-MS/MS-based metabolomics was used. Moreover, the effects of EPS1-1 against carcinogenesis and on the metabolic profile were also detected by this strategy. Our findings provide new insights into the diagnosis of cancer at the metabolic level and help to identify the underlying molecular mechanisms of EPS1-1 treatment.
Materials and Methods {#S0002}
=====================
Chemicals and Reagents {#S0002-S2001}
----------------------
Azoxymethane (AOM) was purchased from Sigma Aldrich (St. Louis, MO, USA). Dextran sulfate sodium salt (DSS) was obtained from Yuanye Tech. Co. Ltd (Shanghai, China). DEAE Sepharose Fast Flow, Sephadex G-100, and Sephadex G-75 columns were purchased from GE Healthcare Life Sciences (Sweden). Methanol, fatty acids, ammonium acetate, cellulose acetate nitrate, and l-2-chlorophenylalanine were supplied from Dingguo Corp (Beijing, China).
Preparation of Exopolysaccharide (EPS1-1) {#S0002-S2002}
-----------------------------------------
*R. nigricans* was isolated from straw and preserved in the Laboratory of Biomass Resources, Shandong University (Qingdao, China). EPS1-1 was obtained according to a previously reported method.[@CIT0008] Briefly, *R. nigricans* was cultured in potato dextrose broth at 28°C and 130 rpm for 10 days. The fermentation liquor was precipitated with three volumes of 95% ethanol overnight at 4°C. The precipitate was dissolved in distilled water for deproteinization by the savage method and decoloration by D301R resin, followed by dialysis and lyophilization. The colorless powder was further purified using DEAE-Sepharose column (1.6 cm × 20 cm), Sephadex G-100 column (1.6 cm × 60 cm), and Sephadex G-50 column (1.6 cm × 60 cm). The exopolysaccharide obtained was named EPS1-1.
Animal Experiments {#S0002-S2003}
------------------
The animal studies were performed in accordance with the regulations of the Affairs Concerning Experimental Animals of China and the Ethics Committee of Shandong University with the approval number SYDWLL-2018-12. Mouse-colitis-associated colorectal cancer was induced according to a previously described method.[@CIT0009] Thirty healthy male BALB/c mice (4--6 weeks old, body weight 18--22 g) were obtained from Beijing Vital River Laboratories Co. (Beijing, China). The mice were adapted to a new environment (25 ± 2°C, 40 ± 10% relative humidity, and a 12:12 light dark cycle) for a week and had free access to water and food. As shown in [Figure 1A](#F0001){ref-type="fig"}, mice were randomized into three groups: Model group (n = 10), intraperitoneally injected with AOM (10 mg/kg) and fed 2.5% DSS after a week for seven days and then normal water for 2 weeks for four cycles; EPS1-1 group (n = 10), orally treated with EPS1-1 (180 mg/kg) daily from the first day of the first cycle till the end; Control group (n = 10), free access to regular water and food. At the end of the experimental procedure, all mice were sacrificed and intestinal tissue samples (from cecum to the anal verge) were collected for metabolomic data analysis. Sera samples were used for the measurement of blood lipids. All samples were stored at −80°C until additional analysis.Figure 1Effect of EPS1-1 on AOM/DSS-induced colorectal cancer. (**A**) Experimental procedure of a murine model of colorectal cancer induced by AOM/DSS. (**B**) Macroscopic photos of colon tissues in the Control, Model, and EPS1-1 groups. (**C**) Number of tumors in the Control, Model, and EPS1-1 groups was counted. Compared to the Model group, \*\**p* \< 0.01.
Extraction of Metabolites {#S0002-S2004}
-------------------------
Frozen intestinal tissues (100 mg) were individually ground with liquid nitrogen, following which 200 μL of L-2-chlorophenylalanine was added to the powder. 400 μL of precooled 80% methanol was added to 100 μL of each homogenate, followed by vortexing. Samples were incubated at −20°C for 60 mins and then centrifuged at 14,000 g at 4°C for 15 mins. The supernatants were transferred to a fresh microcentrifuge tube and dried under vacuum in a centrifugal evaporator. Dried metabolite pellets were redissolved in 80% methanol and analyzed by LC-MS/MS.
LC-MS/MS Analysis {#S0002-S2005}
-----------------
LC-MS/MS analysis was performed using a Vanquish UHPLC system (Thermo Fisher) coupled with an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher) operating in the data-dependent acquisition (DDA) mode. Samples were injected onto an Accucore HILIC column (100 × 2.1 mm, 2.6 μm) using a 20-mins linear gradient at a flow rate of 0.3 mL/min. The eluents of the positive polarity mode were eluent A (0.1% FA in 95% CAN, 10 mM ammonium acetate) and eluent B (0.1% FA in 95% CAN, 10 mM ammonium acetate). The eluents of the negative polarity mode were eluent A (95% CAN, 10 mM ammonium acetate, pH 9.0) and eluent B (50% CAN, 10 mM ammonium acetate, pH 9.0). The solvent gradient was set as follows: 2% B, 1 min; 2--50% B, 16.5 mins; 50--2% B, 2.5 mins. Q-Exactive Hf-X mass spectrometer was operated in positive/negative polarity mode with spray voltage of 3.2 kV, capillary temperature of 320°C, sheath gas flow rate of 35 arb, and aux gas flow rate of 10 arb.
Data Processing and Multivariate Data Analysis {#S0002-S2006}
----------------------------------------------
The acquired raw data from LC-MS/MS analysis were imported to the CD search library. Data pretreatment procedures such as peak detection, deconvolution, normalization, alignment, denoising, and data reduction were performed to obtain a list of mass and retention times corresponding to all the detected peaks from each data file in the data set. SIMCA_P software was used for principal component analysis (PCA) for showing the distribution of origin data, and PLS-DA was used to check the model's validity. Based on variable importance in the projection (VIP) threshold of 1 from the PLS-DA model, metabolites responsible for the difference between the AOM/DSS-induced CRC mice and the other two groups could be obtained. In parallel, the metabolites identified by the OPLS-DA model were validated at a univariate level using the S-plots. Metabolic pathways were identified using Kyoto Encyclopedia of Genes and Genomes (KEGG) and MetPA. Statistical significance analysis was performed using SPSS 17.0 software (SPSS 17.0 (SPSS Inc., Chicago, IL)). Multiple group comparisons were tested using one-way analysis of variance (ANOVA) followed by Dunnett's test. *p*-value \<0.05 was considered to be statistically significant.
Results {#S0003}
=======
EPS1-1 Inhibits Tumor Growth in AOM/DSS-Induced Colorectal Cancer Tissues {#S0003-S2001}
-------------------------------------------------------------------------
To demonstrate the anti-CRC properties of EPS1-1, a mouse model induced by AOM/DSS was generated as described previously ([Figure 1A](#F0001){ref-type="fig"}). Thirty male BALB/c mice were sacrificed at the end of the experiment and colon tissues were collected. As shown in [Figure 1B](#F0001){ref-type="fig"} and [C](#F0001){ref-type="fig"}, mice in EPS1-1 group demonstrated a significant decrease in the number of colon tumors as compared with the Model group, which are consistent with our previous reports.[@CIT0009] This suggests that EPS1-1 can effectively suppress AOM/DSS-induced colorectal cancer tumorigenesis.
LC-MS/MS Metabolite Analysis {#S0003-S2002}
----------------------------
Typical LC-MS/MS total ion current (TIC) chromatograms of colonic tissue samples from Control, Model, and EPS1-1 groups are shown in [Figure 2](#F0002){ref-type="fig"}. Majority of the peaks in the TIC profiles were identified by their retention times, mass spectra characteristics, and the LECO Fiehn Metabolomics Library. As shown in [Figure 2](#F0002){ref-type="fig"}, peaks among the three groups presented various differences, suggesting that the TIC chromatograms could mirror the distinction among the three groups.Figure 2Representative LC-MS/MS total ion current (TIC) chromatograms of the three groups. (**A**) Control group; (**B**) Model group; (**C**) EPS1-1 group.
Statistical Comparison of Metabolites in the Three Groups {#S0003-S2003}
---------------------------------------------------------
To evaluate the effect of EPS1-1 on induced colorectal cancer mice, metabolomic data analysis was performed. PLS-DA was used to obtain a higher level of group separation, and hence, get a better understanding of variables responsible for classification. As shown in [Figure 3](#F0003){ref-type="fig"}, the PLS-DA score parameters were as follows: R2Y = 0.98, Q2Y = 0.2 for Model vs Control; R2Y = 0.98, Q2Y = 0.24 for EPS1-1 vs Control; and R2Y = 0.98, Q2Y = −0.04 for EPS1-1 vs Model ([Figure 3A](#F0003){ref-type="fig"}--[C](#F0003){ref-type="fig"}). These data indicate that the model was stable. To further validate our model, a signal 10-fold cross-validation was performed to evaluate the robustness and the predictive ability of the model.[@CIT0020] The R2 and Q2 intercept values were 0.95 and −0.78 ([Figure 3D](#F0003){ref-type="fig"}), 0.96 and −0.7 ([Figure 3E](#F0003){ref-type="fig"}), and 0.94 and −0.59 ([Figure 3F](#F0003){ref-type="fig"}) after 200 permutations for Model vs Control, EPS1-1 vs Control, and EPS1-1 vs Model, respectively. The low values of Q2 intercept highlighted the robustness of the models, and thus demonstrated a lower risk of over fitting.[@CIT0021] Therefore, these data indicate that the PLS-DA model could be used to identify differences among the three groups.Figure 3PLS-DA score plots (**A--C**), validation plots of PLS-DA valid (**D--F**) derived from LC-MS/MS metabolite profiles of Model vs Control (**A** and **D**), EPS1-1 vs Control (**B** and **E**), EPS1-1 vs Model (**C** and **F**).
Metabolite Analysis Among the Three Groups {#S0003-S2004}
------------------------------------------
In this study, we detected 1257, 1257, and 1288 altered compounds for Model vs Control, EPS1-1 vs Control, and Model vs EPS1-1, respectively. Among these, the number of identified metabolites were 74, 45, and 13 for Model vs Control, EPS1-1 vs Control, and Model vs EPS1-1, respectively. A total of 19 significantly altered metabolites were identified based on *p*-values from the two-tailed student's *t*-test (*p* \< 0.05) and VIP values from the OPLS-DA models (VIP ≥ 1). However, 6 metabolites were ultimately identified as common to all three groups ([Table 1](#T0001){ref-type="table"}). These significantly altered potential metabolites were cytosine, deoxyuridine, 2-dehydro-3-deoxy-6-phospho-Dgluconic acid, 20-hydroxy-leukotriene E4, hematoporphyrin, and L-homocysteic acid. Compared with the Control group, the altered metabolites included increased cytosine, deoxyuridine, 20-hydroxy-leukotriene E4, hematoporphyrin, and L-homocysteic acid. 2-dehydro-3-deoxy-6-phospho-D-gluconic acid was decreased when compared with Model group. Furthermore, L-homocysteic acid levels were lower, whereas 2-dehydro-3-deoxy-6-phospho-D-gluconic acid levels were higher in EPS1-1 group than the Model group, indicating that significant metabolic profiling changes were induced by EPS1-1. Heatmap in [Figure 4](#F0004){ref-type="fig"} shows that the significantly altered metabolites in Model group were reversed in EPS1-1 group. The representative metabolites are displayed in box-and-whisker plots ([Figure 5](#F0005){ref-type="fig"}), along with the concentration range, median quartiles, and extremes. The change in metabolite expression upon EPS1-1 treatment is in accordance with its anti-tumor effects as demonstrated in our previous report.[@CIT0009] Table 1List of Differential Metabolites in Colonic Tissues in Model Group Relative to Control Group, and EPS1-1 Group Relative to Control GroupMetabolitesRt^a^ (Min)Model vs ControlEPS1-1 vs ControlEPS1-1 vs ModelVIP^b^p-Value^c^FC^d^VIP^b^p-Value^c^FC^d^VIP^b^p-Value^c^FC^d^Cytosine1.071.390.042.892.430.011.944.020.0020.682ʹ-Deoxyuridine2.011.690.0093.562.620.0042.703.920.0040.922-dehydro-3-deoxy-6-phospho- D-gluconic acid4.394.520.0040.052.020.011.091.090.012.1920-hydroxyleukotriene E45.601.160.022.902.880.00031.634.180.00010.74Hematoporphyrin2.251.760.0024.092.280.0016.211.430.024.72L-Homocysteic acid0.481.340.00052.791.340.021.464.050.0050.93Orotidylic acid2.031.050.020.40Docosatetraenoic acid10.751.070.00020.401.090.0030.45N-Formyl-Dkynurenine8.181.040.020.46Isobutyryl CoA3.471.500.042.81Prostaglandin H25.021.300.033.52Palmitoylcarnitine9.891.290.0060.436-Hydroxymelatonin2.111.140.042.04Dihydrofolic acid1.451.330.012.18L-Cysteinylglycine2.071.450.042.40Deoxythymidine2.081.420.012.55N-Carbamoyl- L-aspartate3.441.330.0050.42Hypoxanthine0.641.550.042.32Bis(5ʹ-guanosyl) tetraphosphate2.561.950.022.89[^1] Figure 4Heatmap indicates metabolic changes in the three groups (Control group, Model group, EPS1-1 group). Shades of red and blue represent fold increase and fold decrease of a metabolite, respectively. (**A**) Control group; (**B**) Model group: (**C**) EPS1-1 group.Figure 5Box-and-whisker plots illustrating discrimination among the Control group, Model group, and EPS1-1 group. Horizontal line in the middle portion of the box, median; bottom and top boundaries of boxes, 25th and 75th percentiles, respectively; lower and upper whiskers, the minimum and maximum value, respectively. (**A**) 2ʹ-Deoxyuridine; (**B**) Cytosine; (**C**) 20-Hydroxy-leukotriene E4; (**D**) Hematoporphyrin; (**E**) L-Homocysteic acid; (**F**) 2-dehydro-3-phospho-D-gluconic acid (DKHP).
Analyses of Perturbed Metabolic Pathways {#S0003-S2005}
----------------------------------------
Metabolomics analysis demonstrates not only individual metabolite variations, but also provides a holistic view of metabolic variations induced by biologically functional substances.[@CIT0022] In the present study, MetPA was used to identify the pathways involved in the anti-tumor effects of EPS1-1 on AOM/DSS-induced colorectal cancer. Enriched pathways related to the altered metabolites were identified through KEGG and literature review, and included biosynthesis of unsaturated fatty acids, pyrimidine metabolism, phenylalanine metabolism, tryptophan metabolism, inositol phosphate metabolism, arginine and proline metabolism, phenylpropanoid biosynthesis, fatty acid metabolism, and folate biosynthesis. Next, we listed six altered metabolic pathways with *p*-value \<0.1, including biosynthesis of unsaturated fatty acids, pyrimidine metabolism, phenylalanine metabolism, fatty acid metabolism, folate biosynthesis, and inositol phosphate metabolism ([Table 2](#T0002){ref-type="table"} and [Figure 6](#F0006){ref-type="fig"}). Our results suggest that these six metabolic pathways represent the potential mechanisms underlying the anti-tumor effects of EPS1-1 on AOM/DSS-induced colorectal cancer in mice. Table 2List of Differential Metabolic Pathways in Colonic Tissues in Three GroupsMap IDMap TitleP-valueMap01040Biosynthesis of unsaturated fatty acids0.01Map00240Pyrimidine metabolism0.06Map00360Phenylalanine metabolism0.001Map01212Fatty acid metabolism\<0.001Map00790Folate biosynthesis0.06Map00562Inositol phosphate metabolism0.09 Figure 6Potential metabolic pathways related to AOM/DSS-induced colorectal cancer. (**A**) Model group vs Control group; (**B**) EPS1-1 group vs Control group.
Discussion {#S0004}
==========
Numerous studies in recent years have revealed that malignant tumors are closely related to altered metabolism, including pathways such as cell signaling, glycolysis, and respiration (known as the Warburg effect).[@CIT0023]--[@CIT0025] Previous studies showed that EPS1-1 could significantly inhibit tumor growth in AOM/DSS-induced colorectal cancer in mice.[@CIT0009] In the past, urine and blood are the most frequently studied samples among a diverse range of types involved in human metabolomics studies.[@CIT0026] In this study, we aimed to provide a better understanding of the underlying mechanisms and to identify the differential metabolites via metabolomics profiling of mice colon tissues with AOM/DSS-induced colorectal cancer. Using LC/MS-based metabolomics, metabolome profiles were generated that showed significant variations in the Model group. The PLS-DA models derived from our metabolomic data were able to differentiate the Model group and EPS1-1 group from the Control group, highlighting the diagnostic potential of this noninvasive analytical approach. LC-MS/MS-based metabonomic study identified significant variations in the AOM group, which were characterized by 19 differentially expressed metabolites in mice colonic tissue. Differentially expressed metabolites are listed in [Table 1](#T0001){ref-type="table"}. Heatmap ([Figure 4](#F0004){ref-type="fig"}) and box-and-whisker plots ([Figure 5](#F0005){ref-type="fig"}) show less significant fluctuations in metabolite levels in the EPS1-1 group, indicating that EPS1-1 is able to alleviate the AOM/DSS-induced metabolic perturbation in mice. The differential metabolites identified by MetPA mainly include docosatetraenoic acid, 2ʹ-deoxyuridine, cytosine, palmitoylcarnitine, 7,8-dihydrofolic acid (DHF), and 2-dehydro-3-deoxy-6-phospho-D-gluconic acid, which are involved in the biosynthesis of unsaturated fatty acids, pyrimidine metabolism, phenylalanine metabolism, fatty acid metabolism, folate biosynthesis, and inositol phosphate metabolism ([Table 2](#T0002){ref-type="table"} and [Figure 6](#F0006){ref-type="fig"}).
2ʹ-deoxyuridine, an important intermediate product of pyrimidine metabolism, is the downstream product of dUMP. It has been reported that iodine-labeled 2ʹ-deoxythymidine is highly expressed in small intestine and liver in tumor-bearing mice using a tracking technique.[@CIT0027] Compared with the Control group, 2ʹ-deoxyuridine was increased in the Model group, and serve as a tumor marker.
7,8-dihydrofolic acid (DHF) is a metabolite of the folic acid pathway. It was previously reported that DHF, like folate, could induce apoptosis in colon cancer cells by decreasing the protein expression of IGF-IR. This suggested that DHF is a potential anti-cancer substance.[@CIT0028] In our experiments, treatment with EPS1-1 significantly induced the expression of DHF, indicating that EPS1-1 could induce DHF production to suppress tumor growth, thus showing the anti-tumor activity.
In recent years, studies have indicated that tumor growth is closely related to tryptophan metabolism.[@CIT0029] Tryptophan affects tumor growth in vivo via two metabolic pathways: serotonin metabolism and kynurenine metabolism. Kynurenine metabolism is the main metabolic pathway associated with tryptophan and plays an important role in immune system function. N-formylkynurenine is the precursor substance of kynurenine.[@CIT0029] The level of N-formyl-kynurenine was significantly reduced in the Model group compared with Control group, suggesting that tryptophan metabolism was significantly disrupted and immune system functions were affected in AOM/DSS-induced colorectal cancer mice. In addition, serotonin metabolism is closely related to the nervous system and serves as an important neurotransmitter, 5-HT.[@CIT0030] 6-hydroxymelatonin is a metabolite of the serotonin metabolic pathway. Compared with Control group, we found that the level of 6-hydroxymelatonin was significantly increased in the EPS1-1 group, indicating that EPS1-1 could alleviate AOM/DSS-induced colorectal cancer through the serotonin metabolic pathway.
Cyclooxygenases (COXs) are key enzymes mediating the conversion of free arachidonic acid into prostaglandins H2 (PGH2), which are the precursors of molecules such as prostaglandins, prostacyclin, and thromboxanes.[@CIT0031] COX-2 is upregulated in inflammatory and neoplastic states and catalyzes the synthesis of PGH2.[@CIT0032],[@CIT0033] High levels of the PGH2 were seen in the Model group compared with Control group. Proglumide, a gastrin and cholecystokinin receptor antagonist, has been shown to inhibit growth of transplantable murine colon carcinoma and to prolong survival in tumor-bearing mice.[@CIT0034] The level of proglumide was significantly decreased in AOM/DSS-induced mice compared with Control mice. Hematoporphyrin is an endogenous porphyrin produced by hydrolysis of hemoglobin. It has been reported that hematoporphyrin could inhibit the proliferation of tumor cells.[@CIT0035]--[@CIT0037] The level of hematoporphyrin in EPS1-1 group was higher than the Model group, showing that EPS1-1 could induce the production of hematoporphyrin to suppress the growth of colorectal cancer cells.
A notable feature of the metabolic profile of the Model group was disrupted nucleic acid metabolism, as characterized by the increased levels of deoxyuridine and cytosine in colonic tissue.
In summary, we performed global tissue profiling and identified altered metabolites in mice with AOM/DSS-induced colorectal cancer. EPS1-1 was able to attenuate AOM/DSS-induced metabolic perturbations. It also appears that EPS1-1 induced significant metabolic alterations independent of the AOM/DSS-induced metabolic changes. Collectively, our results indicated that EPS1-1 can serve as a potential therapeutic agent for the treatment of colorectal cancer.
Strength and Limitation of This Study {#S0005}
=====================================
The strength of this study lies in the use of two analytical techniques -- LC and MS/MS. LC-MS/MS possesses high sensitivity and high unit mass accuracy to identify a number of metabolites. It can also conduct sophisticated library searches and/or sum formula calculations. A limitation of our study was the small sample size. For follow-up studies, we would expand the number of induced mice to identify additional cancer biomarkers.
This study is financially supported by the National Key Research and Development Program of China (No. 2018YFA0902000), Major Program of Shandong Province (the Key Technology) (No. 2015ZDJS04002), the Major State Basic Research Development Program of China (973 Program) (No. 2012CB822102), the High Technology Research and Development Program of China (863 Program) (No. 2012AA021501), and the Natural Science Foundation of Education Department of Anhui Province (No. KJ2018ZD025).
Disclosure {#S0006}
==========
The authors report no conflicts of interest in this work.
[^1]: **Notes:** ^a^Rt: Retention time. ^b^VIP, also named variable importance in the projection, was obtained from PLS-DA model with a threshold of 1.0. ^c^p-Value was calculated from Wilcoxon-Mann--Whitney test. ^d^FC, also named fold change, was obtained by comparing those metabolites in two groups.
| {
"pile_set_name": "PubMed Central"
} |
Case Report
===========
Numerous cases of sudden deaths caused by intentional propane and/or butane inhalation have been described in the existing literature,[@b1-ndt-11-1157],[@b2-ndt-11-1157] but cases of survivors, with relative descriptions of the associated consequences, are very rare. We report a patient who survived after intentional inhalation of liquefied petroleum gas (LPG), and who developed progressive ataxia, Parkinsonism, and dystonia.
A 38-year-old prisoner with an unremarkable medical history was found unconscious inside his cell. He was lying on the bed with a large cellophane bag near his head. An empty gas cylinder containing LPG, often used by prisoners to cook in their cells, was found close to him.
The man was brought to the emergency room in a comatose state and with shortness of breath. He was promptly intubated and admitted to the intensive care unit. Analysis of arterial blood gases showed hypoxia, hypercapnia, and mixed acidosis. Intravenous administration of naloxone had no effect. The patient's blood pressure and electrocardiography were normal. Serum basic chemistry was unremarkable and the alcohol level was zero. Urine toxicology was negative for amphetamines, cocaine, barbiturates, benzodiazepines, and opiates. He had no pulmonary edema.
Urgent computed tomography showed slight hypodensities in both cerebellar hemispheres. In the first 48 hours, the patient rapidly improved. He was able to open his eyes on command and began to move his arms and legs. Brain MRI showed isolated bilateral lesions involving the caudate nucleus and globus pallidus, as well as diffuse signal alterations in both cerebellar hemispheres ([Figure 1](#f1-ndt-11-1157){ref-type="fig"}). The imaging findings were consistent with a metabolic or toxic etiology.
In the subsequent week, the clinical picture continued to improve gradually; he was awake and alert, and he was able to stand up, but unable to walk because of impaired balance. Despite significant motor impairment, his sensory and cognitive functions were normal (Mini-Mental State Examination score =30/30).
Two months after discharge, stiffness, bradykinesia, and hypomimia were more evident. He also developed a dystonic posture of his right foot and left hand.
^123^I-ioflupane single-photon emission computed tomography showed bilaterally decreased uptake in the caudate nucleus and putamen. He was treated with levodopa and dopamine agonists for the akinetic--rigid syndrome without any benefit.
At present, 1 year after the event, the patient is not able to walk and he is in a wheelchair. The main symptoms are severe ataxia, widespread plastic rigidity, bradykinesia, and dystonia. He is moderately dysarthric and his speech is very slow with frequent interruptions. He works as an art critic, and has written and published poetry. The current brain MRI scan has not shown significant changes when compared with the previous one.
The practice of inhaling hydrocarbons, despite being uncommon, has increased greatly in the last few years, especially among adolescents in search of getting high and among the prison population where drug addicts, who are unable to obtain their habitually abused substances, resort to this alternative practice.[@b3-ndt-11-1157] LPG is a flammable mixture of hydrocarbon gases and is used as a fuel, which contains propane as a major component and, in addition, isobutane. These substances are lipophilic, so that after being inhaled and after being taken up from the lungs into blood, they are distributed at high concentrations in lipid-rich tissues, especially in the brain and fat tissues.[@b2-ndt-11-1157]
Pathologically, anoxic injuries can be classified as hypoxic hypoxia, from decreases in the partial pressure of blood oxygen (as in patients with hanging or drowning); histotoxic hypoxia, from the tissue's inability to utilize oxygen (as in the case of exposure to mitochondrial toxins like carbon monoxide and cyanide); anemic hypoxia, from decreased hemoglobin content or function (as in cases of severe anemia); or hypoxic--ischemic encephalopathy, typically seen after cardiac arrest.[@b4-ndt-11-1157] The clinical evolution and the MRI abnormalities in our patient are similar to those described in cases of poisoning by CO, where the mechanism of anoxic brain injury is attributable to histotoxic hypoxia.[@b5-ndt-11-1157] The brain areas affected in CO toxicity are the ones with a high metabolic rate and oxygen demand, and include the basal ganglia, and particularly the globus pallidus, substantia nigra, and hippocampus. Laminar necrosis of the cerebellar cortex and Purkinje cell loss are often seen.[@b5-ndt-11-1157] Similar lesions have been reported following global cerebral hypoxia, but frequently with analogous alterations in the cerebral cortex, lacking in this case. Bilateral basal ganglia lesions can also be observed in patients with illegal drug abuse, such as those using heroin, cocaine, and ecstasy. In these cases, the pathomechanism is usually vascular related (vasospasm, ischemia, hemorrhage, and vasculitis), and the white matter is frequently involved.[@b6-ndt-11-1157] The progressive ataxia and the akinetic--rigid syndrome, which appeared a few days after the event, were the prominent clinical manifestation in our patient. Before brain insult, the patient had no neurological symptoms. He had a normal birth and development, and there was no relevant family history. Ataxia was caused by bilateral and symmetrical lesions of the cerebellar hemispheres, compatible with toxic or metabolic etiology. Akinetic--rigid syndrome and dystonia are instead clear signs of damage at the level of the basal ganglia, which are very sensible to hypoxic/anoxic damage, especially at the globi pallidi. Dystonia could be also explained by aberrant functional output from the cerebellum,[@b7-ndt-11-1157] which was extensively damaged in our patient.
Conclusion
==========
In conclusion, we believe that LPG, considered until now a mix of gas with low neurotoxic power, may have caused direct toxic damage to the brain, mediated by a mechanism of histotoxic hypoxia, such as in CO intoxication. To our knowledge, this is the first reported case featuring a neurological syndrome that presented after LPG intoxication. It would be useful to know if LPG is really able to cause or amplify hypoxic/anoxic damage in humans. For this purpose, laboratory analysis of blood and urine samples up to 24 hours postexposure should be used to confirm the diagnosis of volatile substance intoxication and to demonstrate not only the concentration of inhaled gases, but also the duration of LPG exposure.[@b8-ndt-11-1157] Although blood butane levels can be measured by headspace gas chromatography--mass spectroscopy, the majority of clinical laboratories do not run these tests routinely, and their availability is extremely limited. Therefore, we recommend greater diffusion of this technique, especially in emergency departments.
**Disclosure**
The authors report no conflicts of interest in this work.
{#f1-ndt-11-1157}
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION
============
Compared to ST elevation myocardial infarction (STEMI), unstable angina and non-ST elevation myocardial infarction (UA/NSTEMI) show lower early mortality, but the long-term prognosis of these conditions is known to be worse ([@B1]-[@B5]). As the patients with UA/NSTEMI tend to be older and have more extensive coronary disease and more co-morbidities compared with those with STEMI ([@B1]), their health status would be so worse to be related to adverse long-term clinical outcome.
Health-related quality of life (HRQOL) scoring systems have been used as the objective measurement of performance status in various cardiovascular diseases ([@B6]-[@B9]). Although the role of percutaneous coronary intervention (PCI) in chronic stable angina is well established as it alleviates ischemic symptoms and improves quality of life ([@B10], [@B11]), quality-of-life has been relatively ignored in the management of acute coronary syndrome (ACS) because mortality reduction is the primary goal of treatment of ACS.
Our hypothesis is that HRQOL status would be worse in patients with UA/NSTEMI compared with STEMI. To elucidate the relationship between the change of health status and clinical factors, we investigated the HRQOL of survivors of ACS who had been treated with PCI.
MATERIALS AND METHODS
=====================
Study design and population
---------------------------
We designed the Korean multicenter survey of health status outcomes after percutaneous coronary angioplasty (MUSTANG) study to invest HRQOL status for ACS patients treated with optimal PCI ([@B12]). From September 2009 to July 2010, consecutive patients were enrolled prospectively from 48 hospitals. Patients were considered eligible if they presented with UA/NSTEMI or STEMI and underwent timely PCI. Exclusion criteria were failure to achieve optimal reperfusion, delayed treatment in patients with STEMI (PCI done ≥ 12 hr after the onset of symptoms), inability to answer the questionnaires to assess health status or unwillingness to provide written informed consent.
Patients were diagnosed and managed following the current practice guidelines at the discretion of the attending physicians ([@B13]-[@B15]). Decision to refer to PCI was also at the discretion of the attending physician. A previous cardiovascular event was defined as a past history of myocardial infarction, PCI, coronary artery bypass graft surgery, or stroke. Patients were followed up for 30 days after the index PCI. Major adverse clinical events were defined as one or more of following: cardiovascular or noncardiovascular related death, non-fatal myocardial infarction, non-fatal stroke, recurrent significant angina requiring hospital admission and target vessel revascularization.
Health status data
------------------
Baseline health status data were acquired directly from the patients within 24 hr after PCI through an in-depth individual interview conducted by trained personnel. The second measures were acquired at 30 days after the index PCI during the scheduled visits or by telephone interview. Two types of standardized written questionnaires were used: the Seattle Angina Questionnaire (SAQ) and the Euro Quality of Life 5-Dimensional Classification (EQ-5D) ([@B16], [@B17]). The original version SAQ tool is a 19-item questionnaire that is composed of five subscales of disease-specific health status for patients with coronary artery disease. We used physical limitation, angina frequency and quality-of-life (SAQ QOL) subscales. The scores of each subscale ranged from 0 to 100, where higher scores indicated better health status. The EQ-5D is a 5-item questionnaire that is composed of 5 subscales of general health status, that is, mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Scores of each subscale were converted to a summary index (EQ-5D index) that ranged from 0 (worst possible health state or death) to 1 (best possible health state).
The primary endpoint was the score on the SAQ QOL subscale or the EQ-5D index at 30 days after index PCI. Significant improvement in HRQOL was defined as an increase of ≥ 10 points on the SAQ QOL and an increase of ≥ 0.2 points on the EQ-5D index ([@B18]). The secondary endpoints were the score on each of the three subscales of SAQ and the EQ-5D index, and the interval changes of scores (ΔHRQOL) between each measurement.
Statistical analysis
--------------------
Baseline clinical and demographic characteristics and HRQOL outcomes were analyzed according to the clinical diagnosis of UA/NSTEMI and STEMI. Continuous data were presented as the mean ± SD. Frequency data were presented as the number and percent. Differences between the clinical diagnosis groups were compared using Student\'s t-test for continuous data and the chi-square test for frequency data. Independent factors of STEMI were determined by using logistic regression analysis. Models included baseline characteristics and medications, and the results of stepwise variable selection were presented. Interval changes of each subscale of HRQOL from baseline to 30 days after PCI (ΔHRQOL) were compared using the paired t-test. Independent predictors of the 30-day HRQOL and ΔHRQOL were determined by using multiple linear regression analysis. Baseline characteristics and medications were included in multivariate model, and the results were presented as the beta coefficient (β), standard error (SE) and their 95% confidence interval (CI). To adjust for the effect of various clinical factors on the 30-day EQ-5D index scores between the patients with UA/NSTEMI and STEMI, general linear model was constructed. Clinical and demographic variables were added sequentially to the model. Every time each variable was added to the model, the mean EQ-5D index scores of both diagnosis groups were compared, and the contribution of the variable was assessed by the change in the coefficient of determination (R^2^). Missing values were imputed from the median. There was no substantial difference between analyses with a complete data set and those with a data set that consisted wholly of imputed values. Two-sided *P* values \< 0.05 were considered statistically significant in all tests. All analyses were performed with the use of SPSS for Windows software, version 17.0 (SPSS Inc, Chicago, IL, USA).
Ethics statement
----------------
The local institutional review board of each individual center approved the study protocol. All patients enrolled in the study offered written informed consent. The data were collected by the investigators at each hospital and analyzed by independent researchers. All authors had approved and vouched for the accuracy and completeness of the contents.
RESULTS
=======
Patient population
------------------
Among 3,577 patients recruited, data from 3,362 patients were analyzed. Incomplete data from 215 patients were excluded. Major adverse clinical events occurred in 52 patients during the observation period. The incidence of ACS reached a peak in the sixth decades in males, and the eighth decades in females. In STEMI, the proportion of males below 70 yr old was high compared with the females.
Baseline characteristics according to the clinical diagnosis are listed in [Table 1](#T1){ref-type="table"}. The patients with UA/NSTEMI were older and showed higher frequencies in female sex, diabetes and hypertension compared with those with STEMI. Independent variables associated with STEMI were younger age (OR, 0.992; 95% CI, 0.984-1.000, *P* = 0.048), male sex (OR, 1.260; 95% CI, 1.024-1.550, *P* = 0.029), current smoking status (OR, 1.951; 95% CI, 1.619-2.351, *P* \< 0.001), absence of diabetes (OR, 0.800; 95% CI, 0.664-0.965, *P* = 0.020), hypertension (OR, 0.593; 95% CI, 0.500-0.703, *P* \< 0.001) and previous history of major cardiovascular events (OR, 0.571; 95% CI, 0.458-0.713, *P* \< 0.001). The medications at 30 days according to the clinical diagnostic groups are shown in [Table 2](#T2){ref-type="table"}. Beta blockers, and angiotensin-converting enzyme inhibitors or angiotensin receptor blockers were administered more frequently in patients with STEMI than those with UA/NSTEMI (*P* \< 0.001).
General outcomes about HRQOL
----------------------------
The rate of response for the 30-day HRQOL data was 94%. The unadjusted scores of HRQOLs acquired at the baseline, and at the 30th day after PCI, and the interval change (Δ) between the two observation periods are shown in [Fig. 1](#F1){ref-type="fig"}. Both of the anginaspecific and general HRQOL were improved significantly at 30 days after PCI.
Angina-specific health status measured by SAQ subscales
-------------------------------------------------------
For angina-specific HRQOL status, the baseline scores of all three SAQ subscales were reported as significantly lower in patients with UA/NSTEMI than in those with STEMI. Mean ± SD were 75.2 ± 22.8 and 83.1 ± 22.4 (*P* \< 0.001) for the SAQ physical-limitation subscale, 67.4 ± 26.2 and 81.0 ± 22.1 (*P* \< 0.001) for the SAQ angina-frequency subscale, and 45.0 ± 19.0 and 50.1 ± 20.5 (*P* \< 0.001) for the SAQ QOL subscale in the UA/NSTEMI and STEMI groups, respectively. The degree of improvement (Δ) was significantly higher in the UA/NSTEMI group: mean ± SD change were 13.0 ± 21.5 and 6.9 ± 21.3 (*P* \< 0.001) for the ΔSAQ physical-limitation subscale, 23.0 ± 28.4 and 11.6 ± 24.1 (*P* \< 0.001) for the ΔSAQ angina-frequency subscale, and 11.2 ± 22.7 and 6.5 ± 23.1 (*P* \< 0.001) for the ΔSAQ QOL subscale in the UA/NSTEMI and STEMI groups, respectively. At day 30, SAQ QOL was comparable in two groups (56.1 ± 18.6 vs 56.6 ± 18.7, *P* = 0.521).
The number of patients in the UA/NSTEMI group who showed significant improvement (≥ 10 points) in the SAQ QOL subscale (ΔSAQ QOL) was higher (36.8% vs 44.2%, *P* \< 0.001, [Fig. 2](#F2){ref-type="fig"}) than the STEMI group. In addition to the diagnosis of UA/NSTEMI (β = 3.60, SE 0.98, *P* \< 0.001), factors predicting significant improvement in the ΔSAQ QOL were current smoking status (β = 3.14, SE 0.98, *P* = 0.001), and the usage of beta blockers (β = -2.17, SE 0.93, *P* = 0.020) and calcium channel blockers (β = 3.11, SE 1.04, *P* = 0.003).
In the subgroup analysis for the patients with UA/NSTEMI, independent factors predicting high 30-day QOL were age (β = 0.101, SE 0.039, *P* = 0.010), male sex (β = 2.421, SE 0.864, *P* = 0.005), and no usage of beta blocker (β = -2.863, SE 0.850, *P* = 0.001). For the STEMI group, no specific predictor was identified.
General health status measured by the EQ-5D index
-------------------------------------------------
At the baseline, there was no significant difference in the unadjusted EQ-5D index between the UA/NSTEMI and STEMI groups. The mean ± SD unadjusted EQ-5D indices were 0.77 ± 0.27 and 0.78 ± 0.29, respectively (*P* = 0.207, [Fig. 1](#F1){ref-type="fig"}). The baseline EQ-5D index showed independent correlation with age (β = -0.001, SE 0.001, *P* = 0.002), female sex (β = -0.071, SE 0.011, *P* \< 0.001) and the presence of hypercholesterolemia (β = 0.028, SE 0.010, *P* = 0.014). At 30 days after PCI, the EQ-5D index were lower in patients with UA/NSTEMI than in STEMI; the mean ± SD EQ-5D indices were 0.86 ± 0.21 and 0.89 ± 0.17, respectively (*P* = 0.001). In patients with UA/NSTEMI, independent predictors for the 30-day EQ-5D index were age (β = -0.002, SE 0.001, *P* \< 0.001), male sex (β = 0.051, SE 0.010, *P* \< 0.001), a history of cardiovascular events (β = -0.022, SE 0.010, *P* = 0.030) and diabetes (β = -0.022, SE 0.009, *P* = 0.021).
Compared with the SAQ QOL, the number of patients who showed significant improvement over 30 days in the EQ-5D index was fewer, and was not different between the two groups ([Fig. 2](#F2){ref-type="fig"}).
Regarding the 30-day EQ-5D index between the UA/NSTEMI and STEMI groups, significant difference between the two groups was maintained in the process of sequential adjustment for each clinical factor and baseline HRQOL status (*P* \< 0.001, [Fig. 3](#F3){ref-type="fig"}). The addition of age in the model explained a 2.6% of the variance. The addition of the baseline SAQ QOL explained a further 7.4% of the variance.
DISCUSSION
==========
HRQOL status in patients with ACS after PCI for resolving critical ischemia was was the primary concern of this study. The patients with UA/NSTEMI showed a lower general HRQOL status than those with STEMI even after optimal PCI. Unfavorable baseline clinical profiles of the patient with UA/NSTEMI are likely to be the cause, which is consistent with what was previously known ([@B1]-[@B5]). Poor general health status of patients with UA/NSTEMI might be related to their adverse long-term clinical outcomes.
Although most patients with ACS do not experience angina before index events, advanced atherosclerosis is often found. This is particularly true for patients with UA/NSTEMI, who usually have more extensive systemic atherosclerosis than those with STEMI ([@B1]). STEMI is the most severe form among ACS events, in which a gross rupture of plaque is accompanied by a storm of inflammatory and coagulative responses. Complete occlusion of a single vessel causing localized transmural infarction presents a dramatic change in the electrocardiogram, which is frequently observed in patients with STEMI. In UA/NSTEMI, which is often caused by small plaque erosion, irreversible myocardial necrosis per se tends to be less than STEMI due to several mechanisms including flow in the collateral circulation despite the much more extensive endangered. Even after optimal PCI, the improvement of performance status would be limited. Patients with UA/NSTEMI should be managed in the context of chronic heart failure as well as the ischemia-oriented acute stage care.
Advanced atherosclerosis is often associated with increased cardiovascular mortality in the setting of various non-cardiac chronic illnesses such as end-stage renal disease even in patients with considerable malnutrition ([@B19]). The heavier burden of systemic atherosclerosis is often associated with lower health status and poor long-term outcomes in such chronic diseases. Although there is controversy about the causal relationship, the poor general health status of patients with UA/NSTEMI may be attributable to the advanced atherosclerosis that is often found in chronic illnesses.
The significant improvement was evident in the angina-specific HRQOL, not in the general HRQOL. The degree of improvement in angina-related health status would be greater in patients with more extensive disease. Patients with UA/NSTEMI who have multiple coronary lesions would receive more benefit from PCI treatment than patients with an isolated lesion. As the severity of myocardial damage is less in UA/NSTEMI than in STEMI, one would expect the health status of patients with UA/NSTEMI to improve much more after optimal revascularization therapy compared with that of patients with STEMI. General health status, however, still remained significantly low in patients with UA/NSTEMI even after optimal PCI. Such findings may be the result of the limited health reserve of patients with UA/NSTEMI due to their advanced age, co-morbidities and heavier burden of atherosclerosis. Previous studies showed that the general health status as indicated by the EQ-5D score is a meaningful prognostic index ([@B11], [@B20]). Additional efforts such as rehabilitation and psychological support should be given to patients with UA/NSTEMI to improve their long-term outcome, particularly for those with low general health status.
Among medications, calcium channel blockers had a favorable effect on the improvement of angina-specific health status after PCI. Beta blockers, however, showed a negative impact. Although beta blockers are well-known to relieve the myocardial workload and reduce mortality in various cardiovascular diseases ([@B21]), the benefits of beta blockers might be limited in terms of the HRQOL in patients with UA/NSTEMI who underwent optimal PCI. Another explanation would be that beta blockers were intolerable because they were administrated without careful titration. Current treatment guidelines for heart failure recommend that beta blockers should be initiated with a low dose and titrated slowly over a sufficient period of time ([@B21]). The results might be different if our data had been obtained after a longer follow-up period. Further study is needed to clarify the effects of beta blockers on the QOL of patients who have received optimal revascularization.
The smoking paradox was observed in the present study. Current smoking status was an independent predictor of significant improvement in angina-related QOL in patients with UA/NSTEMI. Despite a controversy, apparent smoking paradox is now attributed to lower risk profile and more aggressive treatment ([@B22], [@B23]). We suppose the reason why our result is a matter of quality of life, not a matter of such \"hard\" outcomes. Smoking cessation after PCI could contribute to the improvement in health status.
The present study has several limitations. First of all, our results of patients with STEMI showing a better HRQOL outcome after PCI were based on the premise that patients with STEMI were treated optimally including timely PCI during the acute presentation. Our results cannot be applicable to patients with STEMI who did not receive appropriate treatment. In addition, as the study population consisted of patients who could answer questionnaires about HRQOL, our results cannot be extended to all ACS patients. The lower rates of adverse clinical events could be explained by such a context. The second is about the baseline health status. As ACS is a rapidly progressive disease, it is difficult to evaluate the past health status before the index event. Our results of the baseline HRQOL particularly for STEMI should be interpreted carefully. In addition, as HRQOL outcomes even after critical treatment are inherently dependent on baseline health status, most of the patients with a high HRQOL score at 30 days also had a higher baseline HRQOL. The 30-day health status would be even more greatly dependent on the baseline health status and clinical characteristics rather than on revascularization therapy, therefore the HRQOL results should be interpreted as a complex of results of treatments and clinical factors. Thirdly, 30 days might be short for patients with ACS to attain full recovery of their general health. Re-evaluation after a sufficient period of time would provide more valuable information. Lastly, the missing of some important clinical information such as angiographic data is a critical limitation. As the primary goal of this study is to investigate HRQOL status, we wholly focused to HRQOL data. That is why the minimum number of variables other than HRQOL was included at the stage of the study planning.
In conclusion, compared with the patients with STEMI, the general health status of those with UA/NSTEMI was not good even after optimal PCI. Although PCI is an effective treatment for acute myocardial ischemia, its benefit in general health status would not be sufficient. In addition to angina-specific therapy, comprehensive supportive care would be needed to improve the general health status of ACS survivors, which might improve long-term clinical outcome especially in patients with UA/NSTEMI.
This study was financially supported by Pfizer Pharmaceuticals Korea Ltd (grant number 2010, NRA2580172).
Pfizer Pharmaceuticals Korea Ltd supported this study but played no role in the study design, data collection, analysis, interpretation, manuscript writing, or decision to proceed with publication.
{#F1}
{#F2}
{#F3}
######
Baseline characteristics of the study population
Values are mean ± SD. CABG, coronary artery bypass graft; MI, myocardial infarction; NSTEMI, non-ST elevation myocardial infarction; PCI, percutaneous coronary intervention; STEMI, ST elevation myocardial infarction; UA, unstable angina.
######
Medication at 30 days after PCI according to the clinical diagnosis
ACEI, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; NSTEMI, non-ST elevation myocardial infarction; STEMI, ST elevation myocardial infarction; UA, unstable angina.
| {
"pile_set_name": "PubMed Central"
} |
{.330}
{.331}
| {
"pile_set_name": "PubMed Central"
} |
{#Fa}
Preventing Chronic Disease (PCD) recognizes that public health and clinical interventions are often collaborative, multifaceted, multicomponent, and multisite with diverse participants, stakeholders, and partnerships ([@R1]). As such, evaluation of these efforts cannot rely solely on linear approaches to assess the complex mix of individual, familial, organizational, economic, environmental, and other contextual factors that contribute to the success of interventions. In light of that complexity, it is critically important that researchers, evaluators, and program implementers not focus solely on program outcomes but also spend time to rigorously examine and describe how the program's components produced the reported outcome ([@R2]). It is important that they faithfully execute the implementation plan, success being contingent on the "degree to which a program is delivered as originally designed" ([@R3]) with consideration to local context to improve adoptability and sustainability ([@R2]).
In early 2018, PCD addressed these important considerations by introducing Implementation Evaluation, a new article type that provides the journal's readers (program planners, policy makers, evaluators, researchers, and diverse stakeholders) with information on how to refine evaluation methods, make health system improvements, strengthen collaborations and partnerships, build organizational infrastructure, measure return on investments, and enhance data collection approaches ([www.cdc.gov/pcd/for_authors/types_of_articles.htm](http://www.cdc.gov/pcd/for_authors/types_of_articles.htm)). Implementation Evaluation articles provide insights into factors that affect the ability of public health practice to successfully package and disseminate effective interventions implemented and evaluated in real-world settings. PCD's interest in this area extends to research that examines which factors positively or negatively impact the diffusion of proven interventions and the degree of integrity needed to generate success. Specific program elements such as "adherence to intervention, exposure, or dose, quality of delivery, participant responsiveness and program differentiation" are all factors associated with implementation fidelity ([@R4]). Implementation Evaluation articles published by PCD offer readers timely research that examines in comprehensive ways how evidence-based interventions are implemented in comparable real-world settings.
PCD was fortunate in its inaugural year of introducing this new article type to receive many outstanding submissions. The journal is excited to present this collection of 5 articles that highlight research findings from implementation evaluation efforts that address a variety of topics:
1. a call to action for public health professionals to advance dissemination and implementation science;
2. use of an alcohol surveillance system to assess quality, usefulness, and timeliness of data;
3. the application of a pragmatic framework to guide health care systems in assessing implementation and impact of an evidence-based physical activity program;
4. an assessment of the effectiveness and cost benefit of a program for weight loss and diabetes prevention in a rural setting; and
5. an evaluation of activities to reduce the intake of sodium in community settings.
As part of its effort to provide more research on topics related to implementation evaluation, PCD has recruited associate editors and editorial board members with considerable experience and expertise in implementation dissemination, implementation science, and implementation evaluation. The evolution of work occurring in these areas has expanded over the past 25 years, with the fundamental goal of better identifying program components in public health that contribute to achieving success in population health outcomes. A major component of this goal is to find cost-effective ways to disseminate effective interventions in alignment with local context and real-world settings. An essay from authors Estabrooks, Brownson, and Pronk of our editorial board and associate editor teams provides an overview of dissemination and implementation science, including a review of frameworks, models, theories, concepts, and principles over the past 25 years ([@R5]). These authors discuss the importance of developing individual and team-based skills and abilities among public health professionals that increase adoption and scalability of evidence-based interventions.
Public health surveillance systems are an important aspect of implementation evaluation in collecting and analyzing timely data and disseminating findings that guide public health response to pressing public health issues ([@R5]). Public health surveillance systems, when developed in with input from stakeholders, can be implemented and sustained on an ongoing basis ([@R6]). Hagemery and colleagues conducted an assessment of the alcohol surveillance system to assess quality usefulness and timeliness of data ([@R7]). Researchers completed this assessment through data collection, systematic literature searches, and an interview with the New Mexico Department of Health's alcohol epidemiologist. Authors assessed that the alcohol surveillance system in New Mexico was a useful, stable, and acceptable system capable of monitoring trends and identifying interventions to reduce the prevalence of alcohol-attributable morbidity and mortality in New Mexico ([@R7]). Authors discuss how findings from the assessment were used to enhance the state's alcohol-related surveillance efforts. The evaluation process used by researchers may be useful to others interested in assessing strengths and areas for improvement regarding alcohol-related surveillance at the state level.
In addition to public health surveillance systems, other systems-based approaches must strike a balance between rigor and relevance in considering ways to evaluate the adoption, scalability, and sustainability of interventions ([@R8]). Hence, implementation science research, evaluation, and practice should use tailored evaluation designs that carefully align with the components of the intervention ([@R9]). Stoutenberg and coauthors applied the RE-AIM framework, an approach to planning and evaluating factors related to internal and external validity, to guide health care systems in assessing the implementation and impact of the Exercise is Medicine (EIM) program ([@R10]). EIM is an initiative that integrates physical activity assessment, prescription, and patient referrals as a standard of care ([@R10]). Authors provide recommendations and insights into ways the EIM in health systems can be effectively implemented and evaluated.
Economic evaluations are another aspect of implementation evaluation that is becoming increasingly helpful in informing decision-making to operationalize and sustain implementation strategies and best practices ([@R11]). Economic evaluations are critical to public health professionals, health care organizations, and funders interested in deciding how to maximize use of limited fiscal and human resources ([@R11]). McKnight and associates assessed the effectiveness and cost benefit of replicating a 12-week wellness program targeting adults in 4 rural locations ([@R12]). Researchers reported information on participation, completion, and changes in several health outcomes and discussed how a combination of factors influenced researchers' ability to achieve results similar to those derived in the original wellness program.
Finally, the collection includes research on reducing intake of sodium in community settings, which has remained a national public health issue ([@R13]). This public health goal is particularly important given that diets high in salt are linked to high blood pressure, which is a major risk factor for stroke among adults ([@R14]). Community-based salt reduction programs may be effective in a range of settings, but more robust evaluation methods are needed. Scaling up these efforts in coordination with national initiatives could provide the most effective and sustainable approach to reducing population salt intake ([@R15],[@R16]). In 2016, the Centers for Disease Control and Prevention (CDC) launched the Sodium Reduction in Communities Program (SRCP) to help increase consumers' options for lower-sodium foods and create healthier food environments in communities ([@R17]). CDC's SRCP funded and provided technical assistance to 8 recipients to increase the availability and purchase of lower-sodium food options by implementing 1) food service guidelines and nutrition standards, 2) procurement practices, 3) meal and/or menu modifications, and 4) environmental strategies and behavioral economics approaches to increase consumers' options of lower-sodium foods ([@R17]). Long and coauthors present findings generated from baseline and 1-year follow-up from the SRCP implemented in Arkansas ([@R18]). Researchers describe how program staff worked closely with personnel in a school district and in a community meal program to implement intervention activities to reduce dietary sodium among the food options available and served. Researchers reported that mean sodium content of meals was reduced among participants in both the schools and the community meal program.
This collection of articles from PCD's first year of Implementation Evaluation articles represents an exciting new area of focus for the journal. PCD will continue to identify and publish cutting-edge implementation evaluation research that helps all populations benefit from the dissemination of new and proven discoveries. Toward that end, the journal seeks to gain a deeper understanding of how factors like staffing capacity, economics, leadership support, and intervention fidelity influence scaling up and sustaining proven, culturally appropriate, and setting-relevant interventions. PCD is also committed to publishing articles that use implementation evaluation findings to identify circumstances under which intervention activities should be reduced or discontinued because of factors such as premature adoption (implementing intervention activities before or without having proven evidence of effectiveness), harmful effects, or wasteful use of fiscal or human resources ([@R19]). PCD encourages authors to visit the Author's Corner section of the journal's website at [www.cdc.gov/pcd/for_authors/index.htm](http://www.cdc.gov/pcd/for_authors/index.htm) to learn more about requirements for submitting an Implementation Evaluation manuscript for consideration.
The opinions expressed by authors contributing to this journal do not necessarily reflect the opinions of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors\' affiliated institutions.
*Suggested citation for this article:* Jack L Jr. Promoting the Science and Practice of Implementation Evaluation in Public Health. Prev Chronic Dis 2018;15:180645. DOI: <https://doi.org/10.5888/pcd15.180645>.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-ijerph-17-05944}
===============
Adolescent mental health problems are common and linked to premature death and serious dysfunction in adulthood \[[@B1-ijerph-17-05944]\]. Approximately 10% to 20% of adolescents worldwide suffer from mental health problems, and such problems have become the leading cause of the subsequent development of psychological disorders, such as eating disorders, risk-taking behaviors, and self-harm or even suicide \[[@B1-ijerph-17-05944],[@B2-ijerph-17-05944],[@B3-ijerph-17-05944]\]. These problems have the potential to disrupt education and employment. Previous study has indicated that the incidence and severity of mental health problems have increased more in students than nonstudent populations \[[@B4-ijerph-17-05944]\]. High school students are at a stage of life during which they undergo rapid biological and behavioral changes \[[@B5-ijerph-17-05944],[@B6-ijerph-17-05944]\], and these physical, cognitive, social, and psychological changes increase the risk of anxiety, depression, and other mental health problems \[[@B7-ijerph-17-05944],[@B8-ijerph-17-05944],[@B9-ijerph-17-05944]\]. Anxiety (with prevalence rate up to 31.9%) and behavior disorders (with prevalence rate ranging from 16.3% to 19.1%) were found to be the most frequent mental health problems in high school students, and those dealing with these issues may feel unable to pursue their studies \[[@B7-ijerph-17-05944],[@B10-ijerph-17-05944],[@B11-ijerph-17-05944]\]. Students with mental health problems do not entirely integrate into peer social circles and fear bullying or harassment as a result of their illness, leading to negative consequences in interpersonal communications \[[@B1-ijerph-17-05944],[@B12-ijerph-17-05944]\].
Given that adolescence is a significant period when the foundations for psychological well-being are laid, adolescents who experience mental health problems enter adulthood with disadvantages and will likely continue to experience mental health problems as adults. Indeed, half of all lifetime cases of mental health problems begin by the age of 14 years, and three-quarters begin before the age of 25 \[[@B2-ijerph-17-05944],[@B3-ijerph-17-05944],[@B13-ijerph-17-05944]\]. The prevention of mental health problems in adolescence may substantially decrease mental health problems in adults, as well as benefit broader society by supporting the development of more productive citizens \[[@B14-ijerph-17-05944],[@B15-ijerph-17-05944]\]. However, even when effective treatments for adolescent mental health problems are developed, many adolescents in need do not benefit. Estimates indicate that less than one-third of adolescents have sought treatment due to feelings of stigma, limited mental health infrastructure and policies, funding, and a scarcity of mental health professionals in low- and middle-income countries \[[@B16-ijerph-17-05944],[@B17-ijerph-17-05944]\].
Risk reduction remains a fundamental and cost-effective strategy in reducing the number of individuals affected by and living with mental health problems. As a first step, it requires the identification of modifiable risk factors in order to develop effective prevention strategies. Recognized risk factors for mental health problems among adolescents can be categorized into demographic risk factors (e.g., advancing age, perceived obesity), physical status (e.g., physical disease, teenage pregnancy), unhealthy lifestyle behaviors (such as poor diet, gaming addiction, smoking, alcohol or substance abuse), and familial factors (e.g., living apart from family, parenting styles, family functioning, family income, and parental education) \[[@B1-ijerph-17-05944],[@B9-ijerph-17-05944],[@B18-ijerph-17-05944],[@B19-ijerph-17-05944],[@B20-ijerph-17-05944],[@B21-ijerph-17-05944]\]. For example, rapid changes in family structure, such as divorce, often lead to mental health problems in adolescents \[[@B18-ijerph-17-05944]\].
Relatively little is known about the mental health problems of high school students in mainland China. Previous studies combined middle and high school student samples, making it challenging to create a mental health picture of high school students alone, even though they are an important cohort \[[@B8-ijerph-17-05944],[@B22-ijerph-17-05944],[@B23-ijerph-17-05944]\]. Chinese college admission policies strongly affect the country's entire education system, because success in life and careers can almost be determined by National College Entrance Examination (NCEE) scores. Despite young people's fears that they will fail the NCEE, their parents expect them to go to university, which is a significant source of stress for them \[[@B24-ijerph-17-05944]\]. Moreover, many risk behaviors, such as smoking, binge drinking, and early engagement in sexual activity, begin in or are observed throughout adolescent development, especially middle adolescence (14--16 years) \[[@B5-ijerph-17-05944]\]. In the Youth Risk Behavior Surveys, which were conducted among American high school students, 8.8% of respondents were smokers, 16.5% endorsed binge drinking, and 39.5% had sexual intercourse \[[@B25-ijerph-17-05944]\]. These behaviors have been found to be significantly associated with the development of mental health problems in previous studies \[[@B26-ijerph-17-05944],[@B27-ijerph-17-05944],[@B28-ijerph-17-05944]\].
Henan province is in central China and is the most populous of the 31 mainland Chinese provinces and municipalities. According to 2018 data from the Ministry of Education of the People's Republic of China, there were 23.67 million high school students in China, and more than 2 million resided in Henan province; it has more high school students than the other hinterland provinces, and students are currently under a high level of stress due to the very competitive NCEE. The aim of the present study was to describe the positive rate of mental health problems and their associated factors among high school students in Henan province, China.
2. Materials and Methods {#sec2-ijerph-17-05944}
========================
2.1. Sample and Setting {#sec2dot1-ijerph-17-05944}
-----------------------
From June 2018 to August 2018, a school-based epidemiological survey was performed to explore the positive rate of mental health problems and to analyze their associated factors among high school students in Henan province.
When this survey began, more than two million adolescents studied at 852 high schools in Henan province, with an average of 2465 students per school. As the main aim of this study was to evaluate mental health problems in high school students, the sample size was calculated as N = $Z_{\mathsf{\alpha}/2}^{2}$P$\left( {1 - P} \right)/d^{2}$. With the maximum value of 0.25 for P(1-P), α = 0.05 (Z~α/2~ = 1.96), and the absolute error (d) at 1%, a sample size of 9604 was determined to be sufficient. Predicting a 30% nonrespondent rate, the sample size was estimated to be 13,720 participants. It was further assumed that 300 students (100 from each grade) would be selected from each high school; therefore, 46 high schools would be needed.
A stratified two-stage cluster sampling procedure was used, and participants were selected from 46 high schools located in all 17 provincial cities in Henan province. In Stage 1 of the study, 46 high schools were randomly selected based on probability proportional to enrollment size. In Stage 2, grades were considered as strata, and 2--3 classes in each grade of the 46 selected high schools were randomly selected as clusters based on class sizes. All students in the selected classes were invited to participate in the study by their head teachers, and individuals willing to take part were provided written detailed instructions and information about data collection procedures. All participants completed their surveys during self-study classes in the presence of the principal investigator.
2.2. Ethical Considerations {#sec2dot2-ijerph-17-05944}
---------------------------
Ethical approval was granted by the Ethics Committee of Health Science Center, Xi'an Jiaotong University, and the local schools that participated in this study (project number: 2018-296). Detailed information about the study, its potential risks and benefits, and participants' roles were provided to both participants and their parents, and informed consent was obtained from all participants and their parents before data collection. Participation was voluntary and anonymity was guaranteed.
2.3. Instrumentation {#sec2dot3-ijerph-17-05944}
--------------------
### 2.3.1. Demographic Information {#sec2dot3dot1-ijerph-17-05944}
A demographic information form was developed drawing from the literature on factors that influence high school students' mental health, and it included four key domains: demographic information (age, sex, height, weight, grade, ethnicity, and residence on campus or not), physical illnesses (physical disease and chronic constipation as defined by the Rome IV criteria \[[@B29-ijerph-17-05944]\]), unhealthy behaviors (smoking, alcohol consumption, sexual behavior, and gaming addiction), and family factors (inhabitation, family type, and parents' education level). Based on developmental features, age was subdivided into two groups: middle adolescence (14--16 years) and late adolescence (17--21 years) \[[@B5-ijerph-17-05944],[@B6-ijerph-17-05944]\]. Body mass index was calculated and categorized into three groups: underweight, normal weight, and overweight \[[@B30-ijerph-17-05944]\].
### 2.3.2. Mental Health Inventory of Middle School Students {#sec2dot3dot2-ijerph-17-05944}
The Mental Health Inventory of Middle School Students (MMHI-60) is a self-administered screening test designed to assess students' general mental health. It was developed by Wang et al. \[[@B31-ijerph-17-05944]\] after two years of follow-up investigation into the mental health problems of middle and high school students in more than 100 schools and has been successfully applied to middle and high school students in China \[[@B32-ijerph-17-05944],[@B33-ijerph-17-05944],[@B34-ijerph-17-05944]\]. The MMHI-60 consists of 60 items, each requiring respondents to indicate on a five-point scale whether they have recently experienced a particular symptom or type of behavior. Examples of the questions include "Have you recently felt nervous and strung out?", "Do you frequently think of committing suicide?", and "Do you feel unable to solve problems?". The MMHI-60 includes 10 subscales expressing the most relevant aspects of mental health: obsessive--compulsive tendencies, paranoid ideation, hostility, interpersonal sensitivity, depression, anxiety, academic stress, maladaptation, emotional disturbance, and psychological imbalance. MMHI-60 scores are calculated by summing the 60 items and dividing the sum by 60, and each subscale score is calculated by adding the six items and dividing the sum by six, yielding subscale scores and total scores ranging from 1 to 5. The higher the score, the worse the mental health condition. According to Wang et al. \[[@B31-ijerph-17-05944]\], a cutoff score of 2 is associated with having a mental health problem, and that cutoff score has shown good sensitivity and specificity in previous studies \[[@B32-ijerph-17-05944],[@B33-ijerph-17-05944],[@B34-ijerph-17-05944]\]. The internal consistency reported by Wang et al. was found to be sufficient (test--retest reliability: 0.716--0.873; split-half reliability: 0.634--0.873). The Cronbach's alpha of the total scale was 0.967 and ranged from 0.684 to 0.862 between subscales in this study.
2.4. Data Analysis {#sec2dot4-ijerph-17-05944}
------------------
All the data collected from the study were entered in duplicate using EpiData 3.1 software. All statistical calculations were performed using IBM SPSS 22.0. Continuous variables were expressed as the mean "±SD". Categorical or ordinal data were expressed as number frequencies (%), and between-group differences were tested using the Pearson's χ^2^-test. Estimates of the positive rate of mental health problems with 95% confidence intervals were calculated separately for the overall population. A univariate logistic regression was performed to select possible associated factors for mental health problems with *p* values of less than 0.05. Variables significantly associated with mental health problems in the univariate analysis and reported by previous studies were entered as independent variables in a multivariate regression analysis. A two-sided *p* value less than 0.05 was considered statistically significant.
3. Results {#sec3-ijerph-17-05944}
==========
A total of 15,732 high school students from 46 high schools consented to participate in the survey, yielding a response rate of 91%. After excluding 677 ineligible questionnaires (more than 15% of responses missing), 15,055 students (7514 males and 7541 females) were recruited (see [Figure 1](#ijerph-17-05944-f001){ref-type="fig"}).
3.1. Demographic Information {#sec3dot1-ijerph-17-05944}
----------------------------
Participants' characteristics are shown in [Table 1](#ijerph-17-05944-t001){ref-type="table"}. Their mean age was 16.73 years old (SD = 0.88, ranging from 14 to 20). Approximately half were female (50.1%) and came from rural areas (55.9%). The majority was Han Chinese (97.5%), lived on campus (73.1%), and came from two-parent families (93.4%). In total, 4.8% of participants consumed cigarettes, 13% consumed alcohol, 4.5% had experienced sexual activity, and 6.4% had physical diseases. The education level of most of the students' fathers (10,790; 71.7%) and mothers (11,436; 76.0%) was less than junior college.
3.2. Positive Rate of Mental Health Problems {#sec3dot2-ijerph-17-05944}
--------------------------------------------
The overall positive rate of mental health problems was 41.8% (95%CI = 41--42.6; [Table 2](#ijerph-17-05944-t002){ref-type="table"}). More than half of participants had symptoms of academic stress (58.9%, 95%CI = 58.1--59.8), emotional disturbance (55.5%, 95%CI = 54.7--56.3), obsessive--compulsive tendencies (53.2%, 95%CI = 52.4--54), and anxiety (52.8%, 95%CI = 52--53.6). More than two-fifths of participants reported maladaptation (47.5%, 95%CI = 46.8--48.3), interpersonal sensitivity (47.2%, 95%CI = 46.4--48), paranoid ideation (41.4%, 95%CI = 40.7--42.2), and depression (40.9%, 95%CI = 40.1--41.6). The positive rate of hostility and psychological imbalance was 33.9% (95%CI = 33.2--34.7) and 30.4% (95%CI = 29.8--31.2) respectively.
Mental health problems were found to occur more frequently in males than females (43.3% versus 40.2%, *p* \< 0.01; [Table 3](#ijerph-17-05944-t003){ref-type="table"}). The positive rate of mental health problems increased with age, ranging from 40.6% at 14--16 years old to 42.5% at 17--20 years old (*p* \< 0.05; [Table 3](#ijerph-17-05944-t003){ref-type="table"}). The positive rate of mental health problems in first-year high school students was significantly lower than students in second year (39.9% versus 41.9%, *p* \< 0.05) and third year (39.9% versus 43.2%, *p* \< 0.05). Other information on the distribution of mental health problems by population characteristics is summarized in [Table 3](#ijerph-17-05944-t003){ref-type="table"}.
3.3. Factors Associated with Mental Health Problems {#sec3dot3-ijerph-17-05944}
---------------------------------------------------
The univariate analysis of demographic variables is presented in [Table 3](#ijerph-17-05944-t003){ref-type="table"}. Mental health problems were more likely to be reported among students who were older, male, in higher grades, had physical disease, had chronic constipation, consumed cigarettes or alcohol, engaged in sexual behavior, lived on campus, reported gaming addiction, and had a single-parent family or a mother with a master's degree (*p* \< 0.05). Students who lived in urban areas or had a mother who held a bachelor's degree were less likely to have mental health problems (*p* \< 0.05).
The results of the multivariate analysis are presented in [Table 4](#ijerph-17-05944-t004){ref-type="table"}. Higher grades, residence on campus, engagement in sexual behavior, alcohol consumption, gaming addiction, physical disease, chronic constipation, and having a mother with a master's degree were significantly associated with higher odds of having mental health problems, while living in an urban area or having a mother with a bachelor's degree was associated with lower odds (*p* \< 0.05). The strongest associated factor for mental health problems was physical disease (OR = 2.04, 95%CI = 1.77--2.35), followed by chronic constipation (OR = 2.02, 95%CI = 1.8--2.25), and having a mother with a master's degree (OR = 1.87, 95%CI = 1.17--2.99).
4. Discussion {#sec4-ijerph-17-05944}
=============
This study produced several important findings. First, we found that mental health problems are serious public health issues among Chinese high school students, with 41.8% of students reporting mental health problems, suggesting that it is critical for Chinese health authorities and policy makers to strengthen public prevention measures. Second, the most common dimensional symptoms were academic stress (58.9%), emotional disturbance (55.5%), and obsessive--compulsive tendencies (53.2%). Finally, the odds of having mental health problems were strongly associated with higher grades, physical illnesses, unhealthy behaviors (such as alcohol consumption, gaming addiction, and sexual behavior), and familial factors.
This study found that mental health problems were common in Chinese high school students, which is consistent with previous studies conducted in low- and middle-income countries, such as Tanzania (41%) \[[@B35-ijerph-17-05944]\] and Iran (40%) \[[@B36-ijerph-17-05944]\]. One possible reason for this high positive rate might be a low rate of the use of mental health services in China due to pervasive stigma, human resource shortages, and fragmented service delivery models \[[@B37-ijerph-17-05944],[@B38-ijerph-17-05944]\]. Our study implies a need to focus on the mental health problems of Chinese high school students. Its findings also revealed that 58.9% of students had different sources of academic stress, which was in line with previous studies \[[@B39-ijerph-17-05944],[@B40-ijerph-17-05944]\]. Academic stress is a psychological stress that is caused by the expectations of parents and teachers \[[@B41-ijerph-17-05944]\]. In Iran, the most important academic stress was caused by taking university entrance exams \[[@B42-ijerph-17-05944]\]. Academic matters have been found to be the most important sources of chronic and sporadic stress for young people in both Western and Asian countries, and they have significant associations with mental health problems, such as depression, anxiety, and suicidal ideation \[[@B39-ijerph-17-05944],[@B41-ijerph-17-05944]\]. Research found that adolescents with severe academic stress need to be identified early, as interventions to reduce academic stress affect the occurrence and severity of depression \[[@B40-ijerph-17-05944]\]. This study suggests there is a need to examine the demands placed on students in Chinese schools and parents that might contribute to students' academic stress.
Although a comparison between the sexes revealed that sex is not an independent factor for mental health problems, the overall positive rate of mental health problems was higher in males than females, which is similar to a Japanese nationwide survey suggesting that more male high school students feel unhappy than females \[[@B43-ijerph-17-05944]\]. This discrepancy may be due to sex differences in interpersonal relationships and friendships, as males were usually less likely to consult others about their struggles than females \[[@B44-ijerph-17-05944]\]. In our study, age was also not an independent factor for students' mental health problems, while students in higher grades were significantly associated with higher odds of having mental health problems than first-year students, which is consistent with other studies \[[@B45-ijerph-17-05944],[@B46-ijerph-17-05944]\]. One possible reason may be that students in higher grades face more stress because they are nearing their entrance to university \[[@B45-ijerph-17-05944],[@B46-ijerph-17-05944]\].
In this study, high school students' self-reported mental health problems were significantly associated with physical disease and chronic constipation. These findings supported the claim that the odds of having mental health problems were twice as high in individuals with physical illnesses, such as obesity, metabolic syndrome, diabetes mellitus, cardiovascular disease, and respiratory disease, suggesting that attention should be paid to mental health problems when caring for students with physical illnesses \[[@B26-ijerph-17-05944]\]. Research has found that nearly half of individuals with chronic constipation suffer from sleep disorders, which were some of the most common risk factors for developing emotional disorders, such as anxiety and depression \[[@B47-ijerph-17-05944],[@B48-ijerph-17-05944]\]. The association between chronic constipation and mental health problems provides a new path for the development of schools' health services to improve adolescent mental health.
The present study indicated that unhealthy behaviors, such as smoking, alcohol consumption, sexual activity, and gaming addiction were common in individuals with mental health problems, supporting a literature review identifying a relationship between mental health problems and unhealthy lifestyle behaviors \[[@B26-ijerph-17-05944]\]. These researchers claimed that high levels of mental health problems were associated with an increased onset of unhealthy lifestyle behaviors, leading to poor physical health and chronic disease, which can further exacerbate mental health problems \[[@B26-ijerph-17-05944]\]. Another study suggested that adolescents with symptoms of mental health problems, such as depression and anxiety, were more likely to consume alcohol \[[@B49-ijerph-17-05944]\]. Research also found that young people who reported electronic gaming were significantly associated with higher odds of having mental health problems \[[@B50-ijerph-17-05944]\]. Prophylactic programs, which include healthy lifestyle elements, should be established to address mental health problems among adolescents. Our study encourages schools and policy makers to focus more on the mental well-being of high school students, particularly those who exhibit unhealthy behaviors.
Our findings were consistent with previous literature indicating the inverse associations of adolescent mental health problems and familial factors. Familial factors that jeopardized the mental health of high school students in this study included living on campus, having a family living in a nonurban area, having a single-parent family, and parental education. This correlates with a Swedish study that reported that high school students living away from home experience poor mental health \[[@B51-ijerph-17-05944]\]. Furthermore, we found that high school students living in urban areas have lower instances of self-reported mental health problems than their counterparts who do not live in urban areas, which is consistent with previous studies \[[@B23-ijerph-17-05944],[@B52-ijerph-17-05944]\]. One possible explanation is that nonurban students have less access to numerous medical resources and are therefore unable to receive psychological support from health-care institutions when they have mental health problems \[[@B23-ijerph-17-05944],[@B52-ijerph-17-05944],[@B53-ijerph-17-05944]\]. Additionally, students living under the care of a single parent tend to be less proactive in health-seeking behavior and face a higher prevalence of mental health problems \[[@B54-ijerph-17-05944]\]. Our study consistently reported inequalities in adolescent mental health problems in relation to the degree of parental education. Students of parents with low education (a maximum of the lower-secondary level) had more mental health problems compared to students of parents with bachelor's degrees \[[@B55-ijerph-17-05944],[@B56-ijerph-17-05944]\]. This relationship of parental education and children's mental health might be related to socioeconomic factors or the different roles of parents in nurturing and educating their kids \[[@B55-ijerph-17-05944],[@B56-ijerph-17-05944]\]. However, we found an interesting differential association of maternal educational levels and adolescents' mental health problems in which students of mothers with master's degrees showed higher odds of having mental health problems compared to students of mothers with lower levels of education (a maximum of the lower-secondary level). It remains unclear how maternal education influences the individual development of mental health in adolescents over time. Further research is warranted to identify the proxy mechanisms between parental education and students' mental health problems. The strong connection between mental health and familial factors suggests that family members, especially parents, should pay closer attention to the psychological health of their adolescent children and relatives.
5. Conclusions {#sec5-ijerph-17-05944}
==============
In this study, we found a relatively high proportion of mental health problems in high school students. High school students were capable of identifying significant differences in their mental health problems, suggesting that universal mental health risk screening via student self-reporting is potentially useful in identifying aggregated community risk in a given school that may warrant differential deployment of mental health prevention and intervention strategies. We suggest that the prevention of mental health problems in high school students be strengthened, especially in those with physical illnesses, unhealthy behaviors, and single-parent families. This cross-sectional study has significant advantages over other studies in terms of the stratified cluster sampling process and its large sample. However, several limitations must also be recognized. Despite the large sample size, participants were recruited from one province and the findings cannot be generalized to the whole adolescent population in China. In addition, due to the cross-sectional design of our study, no causal relationships can be addressed. More longitudinal, school-based cohort studies and randomized controlled trials on the effectiveness of specific interventions addressing modifiable risk factors are clearly needed in the future.
The authors are grateful to all the high school students who participated in this study and all the head teachers for collection of the data. We also thank all the researchers from Xi'an Jiaotong University for data collection, including Saiyu Wang, Yashuang Guo, Jingzhe Guo, Yichen Yang, Mengshu Li, Menghao Teng, Jia Zhang, Xinrui Han, Zhen Nie, Qianrun Mu, Hang Zhang, Qianyao Zhang, Meiyu Li, Yulin Zhou, Fan Zhang, Yingtong Liu, Ge Wu, Sheng Zhang, Xin Chen, Shengao Yao, Xinxin Feng, Shiwen Chen, Xuyang Sun, Yige Liu, Menghua Xie, Zhihui Chen, Jinmin Li, Yunfei Zhang, Weifei Wang, Yunfei Li, Benchi Ma, Jiabin Liu, Jiahao Sun, Zhimin Wang, Shuai Zhang, Yifeng Guo, Chengyou Jia, and Jing Liu.
Conceptualization, Z.C. and Y.L.; methodology, Z.C. and Y.L.; formal analysis, Y.L., Z.C., and K.W.; investigation, Z.C., Z.L., and Y.L.; writing---original draft preparation, Y.L., and Z.C.; writing---review and editing, Y.L., P.Z., J.W., and J.H.; supervision, Y.L.; project administration, Y.L. and Z.C.; funding acquisition, Y.L. All authors have read and agreed to the published version of the manuscript.
This research was funded by the Soft Science Research Program of Shaanxi Province, grant number 2020KRM169, and the Fundamental Research Funds for the Central Universities, grant number xjj2018153.
The authors declare no conflict of interest. The sponsors had no role in the design, execution, interpretation, or writing of the study.
{#ijerph-17-05944-f001}
ijerph-17-05944-t001_Table 1
######
Major characteristics of study population.
---------------------------------------------------------------------------------------------------------------------------------------
Variables Total\ No Mental Health Problems\ With Mental Health Problems\ *p* Value
(n = 15,058)\ (n = 8768)\ (n = 6287)\
No. (%) No. (%) No. (%)
---------------------- ------------------------ --------------- ---------------------------- ------------------------------ -----------
Age (years) 14--16 5881 (39.1) 3493 (39.8) 2388 (38.0) 0.011
17--20 9174 (60.9) 5275 (60.2) 3899 (62.0)
Sex Male 7514 (49.9) 4259 (48.6) 3255 (51.8) \<0.001
Female 7541 (50.1) 4509 (51.4) 3032 (50.2)
BMI Underweight 1346 (9.1) 782(9.0) 564 (9.1) 0.235
Normal weight 12,294 (81.7) 7197(83.0) 5097 (82.2)
Overweight 1226 (8.2) 687(7.9) 539 (8.7)
Ethnicity Han 14,679 (97.5) 8562 (97.7) 6117 (97.3) 0.093
Minority 376 (2.5) 206 (2.3) 170 (2.7)
Inhabitation Urban 5984 (39.7) 3781 (43.1) 2203 (35.0) \<0.001
Rural 8411 (55.9) 4671 (53.3) 3740 (59.5)
Rural--urban continuum 660 (4.4) 316(3.6) 344 (5.5)
Grade First year 5279 (35.1) 3173 (36.2) 2106 (33.5) 0.002
Second year 4009 (26.6) 2328 (26.6) 1681 (26.7)
Third year 5620 (37.3) 3190 (36.4) 2430 (38.7)
Fourth year ^§^ 147 (1.0) 77 (0.9) 70 (1.1)
Residence on campus No 4047 (26.9) 2522 (28.8) 1525 (24.3) \<0.001
Yes 11,008 (73.1) 6246(71.2) 4762 (75.7)
Smoking No 14,333 (95.2) 8464 (96.5) 5689 (93.4) \<0.001
Yes 722 (4.8) 304 (3.5) 418 (6.6)
Alcohol consumption No 13,091 (87.0) 7895 (90.0) 5196 (82.6) \<0.001
Yes 1964 (13.0) 873 (10.0) 1091 (17.4)
Sexual behavior Never active 14,381 (95.5) 8457 (96.5) 5924 (94.2) \<0.001
Ever active 674 (4.5) 311 (3.5) 363 (5.8)
Gaming addiction No 13,982 (92.9) 8331 (95.0) 5651 (89.9) \<0.001
Yes 1073 (7.1) 437 (5.0) 636 (10.1)
Physical disease No 14,086 (93.6) 8395 (95.7) 5691 (90.5) \<0.001
Yes 969 (6.4) 373 (4.3) 596 (9.5)
Chronic constipation No 13,471 (89.5) 8142 (92.9) 5329 (84.8) \<0.001
Yes 1584 (10.5) 626 (7.1) 958 (15.2)
Single parent No 14061 (93.4) 8265 (94.3) 5796 (92.2) \<0.001
Yes 994 (6.6) 503 (5.7) 491 (7.8)
Paternal education High school and below 10,790 (71.7) 6061 (69.1) 4729 (75.2) \<0.001
Junior college 2208 (14.7) 1439 (16.4) 769 (12.2)
Bachelor 1700 (11.3) 1064 (12.1) 636 (10.1)
Master 140 (0.9) 85 (1.0) 55 (0.9)
PhD 217 (1.4) 119 (1.4) 98 (1.6)
Maternal education High school and below 11,436 (76.0) 6482 (73.9) 4954 (78.8) \<0.001
Junior college 1748 (11.6) 1099 (12.5) 649 (10.3)
Bachelor 1569 (10.4) 1037 (11.8) 532 (8.5)
Master 101 (0.7) 45 (0.5) 56 (0.9)
PhD 201 (1.3) 105 (1.2) 96 (1.5)
---------------------------------------------------------------------------------------------------------------------------------------
Abbreviations: BMI, body mass index. ^§^ students need to retake the National College Entrance Examination, and they take another year of high school with third-year students.
ijerph-17-05944-t002_Table 2
######
Symptoms associated with mental health problems in high school students.
Characteristics Participants Positive Rate (95%CI), % Sort Number
---------------------------------- -------------- -------------------------- -------------
Total scale 6287 41.8 (41.0--42.6) \-
Academic stress 8872 58.9 (58.1--59.8) 1
Emotional disturbance 8355 55.5 (54.7--56.3) 2
Obsessive--compulsive tendencies 8012 53.2 (52.4--54.0) 3
Anxiety 7943 52.8 (52.0--53.6) 4
Maladaptation 7158 47.5 (46.8--48.3) 5
Interpersonal sensitivity 7103 47.2 (46.4--48.0) 6
Paranoid ideation 6239 41.4 (40.7--42.2) 7
Depression 6151 40.9 (40.1--41.6) 8
Hostility 5108 33.9 (33.2--34.7) 9
Psychological imbalance 4579 30.4 (29.8--31.2) 10
ijerph-17-05944-t003_Table 3
######
Positive rate of mental health problems and univariate analysis.
Variables Positive Rate (95%CI) OR (95%CI)
---------------------- ------------------------------ ----------------------- -------------------
Age (years) 14--16 (ref.) 40.6 (39.4--41.9)
17--20 42.5 (41.6--43.5) 1.08 (1.01--1.16)
Sex Male (ref.) 43.3 (42.2--44.4)
Female 40.2 (39.1--41.2) 0.88 (0.83--0.94)
BMI Underweight (ref.) 41.9 (38.1--53.4)
Normal weight 41.5 (40.6--42.2) 0.98 (0.88--1.10)
Overweight 44.0 (41.1--46.8) 1.09 (0.93--1.27)
Ethnicity Han (ref.) 41.7 (40.8--42.5)
Minority 45.2 (39.6--50.5) 1.16 (0.94--1.42)
Inhabitation Urban (ref.) 36.8 (35.6--38.1)
Rural 44.5 (43.4--45.6) 1.37 (1.28--1.47)
Rural--urban continuum 52.1 (48.3--56.2) 1.87 (1.59--2.20)
Grade First year (ref.) 39.9 (38.7--41.1)
Second year 41.9 (40.5--43.5) 1.09 (1.00--1.18)
Third year 43.2 (41.9--44.4) 1.15 (1.06--1.24)
Fourth year 47.6 (39.5--55.8) 1.37 (0.97--1.90)
Residence on campus No (ref.) 37.7 (36.1--39.2)
Yes 43.3 (42.3--44.2) 1.26 (1.17--1.36)
Smoking No (ref.) 40.9 (40.1--41.7)
Yes 57.9 (54.6--61.8) 1.98 (1.70--2.31)
Alcohol consumption No (ref.) 39.7 (38.9--40.5)
Yes 55.5 (53.4--57.6) 1.90 (1.73--2.09)
Sexual behavior Never active (ref.) 41.2 (40.4--42.0)
Ever active 53.9 (50.0--57.6) 1.67 (1.43--1.95)
Gaming addiction No (ref.) 40.4 (39.6--41.1)
Yes 59.3 (56.2--62.2) 2.15 (1.89--2.44)
Physical disease No (ref.) 40.4 (39.6--41.2)
Yes 61.5 (58.4--64.4) 2.36 (2.06--2.69)
Chronic constipation No (ref.) 39.6 (38.7--40.4)
Yes 60.5 (58.0--62.9) 2.34 (2.10--2.60)
Single parent No (ref.) 41.2 (40.4--42.0)
Yes 49.4 (46.4--52.4) 1.39 (1.22--1.58)
Paternal education High school and below (ref.) 43.8 (42.9--44.8)
Junior college 34.8 (35.8--36.8) 0.69 (0.62--0.75)
Bachelor 37.4 (35.1--39.9) 0.77 (0.69--0.85)
Master 39.3 (32.1--47.1) 0.83 (0.59--1.17)
PhD 45.2 (38.2--51.6) 1.06 (0.81--1.38)
Maternal education High school and below (ref.) 43.3 (42.4--44.2)
Junior college 37.1 (34.9--39.2) 0.77 (0.70--0.86)
Bachelor 33.9 (31.5--36.3) 0.67 (0.60--0.75)
Master 55.4 (45.5--65.3) 1.63 (1.10--2.42)
PhD 47.8 (41.3--54.7) 1.20 (0.91--1.58)
Abbreviations: BMI, body mass index; CI, confidence interval.
ijerph-17-05944-t004_Table 4
######
Multivariable analysis of factors associated with mental health problems.
Variables OR (95%CI)
---------------------- ------------------------------ -------------------
Age (years) 14--16 (ref.)
17--20 0.97 (0.89--1.05)
Sex Male (ref.)
Female 0.96 (0.90--1.03)
BMI Underweight (ref.)
Normal weight 1.03 (0.91--1.16)
Overweight 1.10 (0.94--1.29)
Inhabitation Urban (ref.)
Rural 1.26 (1.16--1.36)
Rural--urban continuum 1.68 (1.42--1.99)
Grade First year (ref.)
Second year 1.15 (1.05--1.26)
Third year 1.13 (1.03--1.24)
Fourth year 1.11 (0.78--1.58)
Residence on campus No (ref.)
Yes 1.11 (1.02--1.21)
Smoking No (ref.)
Yes 1.12 (0.94--1.34)
Alcohol consumption No (ref.)
Yes 1.62 (1.45--1.81)
Sexual behavior Never active (ref.)
Ever active 1.31 (1.10--1.55)
Gaming addiction No (ref.)
Yes 1.79 (1.56--2.04)
Physical disease No (ref.)
Yes 2.04 (1.77--2.35)
Chronic constipation No (ref.)
Yes 2.02 (1.80--2.25)
Single parent No (ref.)
Yes 1.19 (1.04--1.37)
Paternal education High school and below (ref.)
Junior college 0.76 (0.68--086)
Bachelor 0.93 (0.80--1.07)
Master 0.73 (0.48--1.09)
PhD 0.61 (0.34--1.08)
Maternal education High school and below (ref.)
Junior college 0.91 (0.80--1.03)
Bachelor 0.80 (0.68--0.93)
Master 1.87 (1.17--2.99)
PhD 1.43 (0.79--2.57)
Abbreviations: BMI, body mass index; CI, confidence interval.
[^1]: Contributed equally.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec0005}
===============
Neurocysticercosis (NCC) caused by the larval stage of the pig tapeworm *Taenia solium* is a serious neurological disease ([@bib0130]). In humans, epilepsy, headache and impaired vision are common clinical presentations of NCC and leading causes of morbidity. Dementia, learning difficulties and changes in cognition are often secondary sequelae of humans with NCC ([@bib0045]). In endemic areas NCC is estimated to be responsible for one third of all late onset epilepsy cases ([@bib0080]) and causes substantial health and economic burdens in the affected populations ([@bib0015], [@bib0090], [@bib0120], [@bib0125]).
Whilst in humans these neuropathologic and clinical aspects have been well studied and documented, there is limited information on the symptomatology the parasite causes in pigs. A number of researchers have previously reported the rarity of signs developed in association with porcine cysticercosis and a limited number of studies exploring the latter ([@bib0045], [@bib0105]). However, in a study conducted in India, eye blinking, tearing and excessive salivation were reported as clinical signs suggestive of porcine NCC. To confirm presence of cysticerci in the brain and correlate the signs to NCC, the pigs underwent magnetic resonance imaging (MRI) examination ([@bib0095]). Since the animals included in the study of [@bib0095] were chosen based on the signs observed by animal caretakers and as the pigs had to be sacrificed for gross pathology and histopathology examination at the end of the MRI scans, the signs could not be reconfirmed. Furthermore, as the study design did not include control pigs, it is not known if the reported signs occurred due to the health conditions of the animals or their environment during the study. Another study by [@bib0075] reported clinical manifestations such as dullness, sluggishness, somnolence, apathy and loss of consciousness in pigs naturally infected with *T. solium* cysticerci. The observed clinical manifestations were only considered as suggestive of NCC because no control pigs were available in that study.
In pigs, seizures have been observed to be caused by salt poisoning, dehydration or pseudo rabies ([@bib0055]), however, the zoonotic parasite *T. solium* has never been reported as an agent causing seizures in this species.
The aim of the present study was to describe possible clinical manifestations associated with NCC and correlate the manifestations to the number, location and distribution of cysticerci in brains of naturally infected pigs in Tanzania.
2. Materials and methods {#sec0010}
========================
2.1. Animals {#sec0015}
------------
The study was carried out at the experimental animal facilities at Sokoine University of Agriculture (SUA), Morogoro, Tanzania. Infected pigs were purchased from farmers in the rural area of Kongwa district, Dodoma region, an area where *T. solium* is known to be highly prevalent ([@bib0070]). To ascertain presence of infection, the pigs were diagnosed by tongue examination ([@bib0035]). Pigs with more than three cysticerci under the tongue were included in the study. As the sensitivity of the diagnostic method is low, non-infected pigs were purchased from smallholder farmers in villages of Morogoro rural district, Morogoro region, where the prevalence of porcine cysticercosis is known to be low (Ngowi, personal communication).
Pigs not sexually mature (maturity determined by age (younger than 6 months) and size (lower than 50 cm in height) or in poor condition (under two thirds of the average weight of 40 kg for a healthy adult pig and visibly ill (covered with ectoparasites and/or with injuries)) were excluded. Infected pigs were purchased first and non-infected pigs were matched with infected ones a posteriori by applying the same exclusion criteria. On arrival of the pigs at SUA, all animals were checked for hard ticks and lice. To eliminate possible confounders (endoparasites and ectoparasites) pigs were treated twice, at an interval of two weeks, with a subcutaneous injection of 0.3 mg/kg of ivermectin (ivermectin 1.0%) ([@bib0010]). Ivermectin was considered safe, as no adverse effects were observed in a study on ivermectin in pigs with cysticercosis ([@bib0070]).
2.2. Study design {#sec0020}
-----------------
An observational study was carried out on 16 pigs naturally infected with *T. solium* and 15 non-infected pigs. A parallel group design was formed. As the disease cannot spread between pigs, three to four infected pigs and their respective controls were randomly chosen and housed together. All pigs were housed under equal conditions in pens (4 × 3 m) with a cement floor and walls. The pens were cleaned every day and the pigs were fed twice a day with commercial dry pig feed. Water was provided *ad libitum* and after morning feeding the animals were provided with forage (*Leucaena leucocephala* or *Amaranthus spinosus*). The mean room temperature of the stable was of 25 °C. Light was present from 6:00 am until 7:00 pm.
Pigs were kept at the experimental animal facilities for one month. During the first two weeks of acclimatisation, the pigs were observed during the day (7:30 am until 6:00 pm). Abnormal occurrences such as: trembling, twitching, uncontrolled movements of rostrum, mouth rigidity and ear paralysis, eye blinking, dribbling, salivating, body stiffening, ataxia, tonic/clonic contractions, panting, collapse of the animal and stereotypic walking in circles were recorded when observed.
After two weeks of acclimatisation, the pigs were continuously videotaped for 14 consecutive days using close circuit television cameras (Velleman^®−^ CCTVPROM16). One camera was mounted above each pen in a central position, permitting a top down view of the whole pen. Each pig was colour-marked using coloured stock markers on its back and on its sides to allow individual identification on the video recordings. The videos of the animals where abnormal occurrences were observed during the acclimatisation period were scanned for abnormal behaviour using fast forward video visualization.
At the end of the video recording period, all 31 pigs (infected and non-infected) were slaughtered by a local butcher. After bleeding, the carcasses of the pigs were transported to the pathology laboratory of the Department of Veterinary Pathology, Faculty of Veterinary Medicine at SUA. There the animals were decapitated at the atlanto-occipital joint ([@bib0005]). To remove the brain from the cranium the procedure described by [@bib0005] was adopted.
To evaluate the distribution of the cysticerci, the cerebral part of the brain was divided into its left and right hemisphere. Thereafter each hemisphere was divided into frontal, temporal, parietal and occipital lobe and cerebellum using sulci as landmarks. Each lobe was carefully sliced to examine and enumerate cysticerci. The cysticerci were evaluated to determine whether they were located in the extra-parenchymal part (parts not completely surrounded by brain tissue) i.e. subarachnoid dorsal and subarachnoid base or in the parenchymal part of the frontal, temporal, parietal and occipital lobe and cerebellum. Cysticerci that could not be located were counted as cysticerci found on the cutting board. Furthermore, the cysticercus developmental stage (vesicular, colloidal or calcified) was recorded ([@bib0040]).
2.3. Seizure classification {#sec0025}
---------------------------
Seizures were classified according to a modified version of the International League Against Epilepsy (ILAE) seizure system developed for dogs ([@bib0060]). The seizure classification summary (Table S1) used is available in Supplementary material. Seizures were divided into four categories: Partial, complex partial, partial with secondary generalization and generalized ([@bib0060]).
2.4. Data analysis {#sec0030}
------------------
Data were entered in Microsoft Excel 2010 and analysed using the statistical software R (R [@bib0025]). Descriptive statistics were performed to compute means and proportions of cysticerci in different parts of the brain. Multiple logistic regression was adopted to assess the effect of age, total number and location of cysticerci on presence of seizures. P-values lower than 0.05 were considered statistically significant.
2.5. Research ethics {#sec0035}
--------------------
Practices employed in the study were approved by SUA, Morogoro, Tanzania (Ref. no. RPGS/R/AS/42/2014) and in accordance with the national guidelines of ethics for health research and to the Animal welfare act (2008) ([@bib0065], [@bib0115]).
3. Results {#sec0040}
==========
3.1. Seizures {#sec0045}
-------------
During the study period two infected pigs were observed having seizures. Seizures occurred recurrently in both pigs. The average seizure duration was 20:51 minutes (Standard Deviation (SD) 14:20 minutes). Video clip 1--7 (available in the electronic version) show the abnormal behaviours observed during a seizure.
### 3.1.1. Description of seizures in pig 1 {#sec0050}
Observed autonomic signs were chewing motions with foamy salivation and dribbling, ear rigidity, eye blinking and panting. Motor signs included tremor, ataxia and tonic/clonic muscle contractions followed by a sudden diminution of all muscle function leading to collapse (Video clip 1---available in the electronic version). Stereotypic walking in circles was observed in several occasions after the occurrence of the autonomic and motor signs (Video clip 2---available in the electronic version). Vomiting also occurred during a seizure (Video clip 3---available in the electronic version). During the post-ictal phase the pig gave the impression of being disoriented as bumping into walls and stumbling over the feeding and water trough was observed (Video clip 4---available in the electronic version). According to the modified version of the ILAE seizure systems, seizures in pig 1 were classified as partial with secondary generalization and generalized.
### 3.1.2. Description of seizures pig 2 {#sec0055}
Observed autonomic signs were uncontrolled movements of mouth and rostrum, dribbling, eyes widening, blinking and ear paralysis. Others signs were brief respiratory arrests, mouth paralysis with continuous swallowing movements of throat and tongue. Also chewing and salivating were common. Observed motor signs included tonic muscular contractions and/or clonic movements like trembling of certain body parts or regions (e.g. head or whole body). Body stiffness and ataxia followed (Video clip 5---available in the electronic version). Other observed motor signs were unwanted repeated lifting of one limb and unbalanced walking. Stereotypic walking in circles was also observed in several occasions during a seizure (Video clip 6 and 7---available in the electronic version). Consciousness seemed to be preserved during partial seizures, while impaired during secondary generalization. According to the modified version of the ILAE seizure systems, seizures in pig 2 were classified as partial seizures, complex partial and partial with secondary generalization.
3.2. Cysticercus distribution and location {#sec0060}
------------------------------------------
Brain cysticerci were found in all 16 pigs classified as *T. solium* infected by tongue examination. No cysticerci were found in any of the 15 control pigs. Cysticerci found in the brains were vesicular and colloidal. Calcified cysticerci were not detected in any of the brains. The number of cysticerci varied from 4 to 418 cysticerci, with a median of 67.5 cysticerci and a mean of 106.7 ± 28.8 standard error of the mean ([Table 1](#tbl0005){ref-type="table"}). [Table 2](#tbl0010){ref-type="table"} describes the distribution and localisation of cysticerci found in the 16 infected pigs. The mean age of the pigs was 18.3 months ± 8.2 standard deviation ([Table 1](#tbl0005){ref-type="table"}). Although only two pigs were observed having seizures, multiple logistic regression analysis found this to be significantly related to age (p \< 0.001) but not to total number, distribution and localisation of cysticerci in the brain.
4. Discussion {#sec0065}
=============
This is the first study that has monitored a group of non-infected pigs and pigs naturally infected with *T. solium* in a controlled environment continuously for one month. Results of this study have shown that pigs with NCC can develop clinical signs and suffer from seizures like humans with NCC associated epilepsy.
In this study, seizures could be detected as the animals were extensively monitored many days in a row. In sub-Saharan Africa pigs are usually left to roam freely and therefore neither controlled nor observed by their owners. Furthermore, pig farmers might not be aware of seizures in pigs or if occurring they might relate it to "possession of bad spirits" as in sub-Saharan Africa, epilepsy is believed to be contagious and caused by witchcraft or evil spirits ([@bib0020], [@bib0135]). People with epilepsy are stigmatized; hence, if farmers observed seizures in pigs, they might fear stigmatization and omit to report it to authorities or researchers. This might potentially explain the lack of previous documentation of seizures in pigs by pig farmers.
The occurrence of seizures and their average duration in this study were in line with studies on epilepsy in other animal species, where they were also observed to vary largely ([@bib0060]).
The autonomic and motor signs observed during seizure manifestations in the two pigs were consistent with signs observed in seizures of other species ([@bib0060]). Two studies had previously reported signs of NCC in pigs ([@bib0075], [@bib0095]). [@bib0095] reported eye blinking, tearing and excessive salivation to be suggestive signs of NCC. In this study eye blinking and salivation were observed, however not to an extent that these signs could be useful to identify pigs with cysticercosis. Abnormal signs observed in this study were more in line with those of [@bib0075]; however dullness, sluggishness and apathy could not be reported, as a clear definition of the terms was not provided by the authors.
In this study, age has shown to be a factor associated with seizures, with older pigs being more prone to express seizures compared to younger ones. This corresponds with knowledge obtained from studies in humans, where seizures were reported to be a frequent manifestation in patients with degenerating cysticerci ([@bib0050]). Moreover the probability of older pigs having degenerating cysticerci is higher compared to younger pigs ([@bib0140], [@bib0105]). Vesicular and colloidal cysticerci were found in both pigs with seizures showing on-going immuno-inflammatory reaction in the brain. This is also in line with studies carried out in humans, where patients suffering from NCC had prominent immuno-inflammatory reactions provoking parasite destruction and leading to clinical manifestations ([@bib0040]).
In the present study, no associations were observed between seizures and number, distribution and localisation of cysticerci in the pig brain. Similar results were also obtained in a study on 29 symptomatic NCC patients, where no significant associations were observed between presence of seizures and number and localisation of the cysticerci in the human brain ([@bib0100]).
In this study, both pigs with seizures had a high number of cysticerci, while in humans also a single cysticercus can lead to seizures ([@bib0085]).
Results of study carried out on 20 symptomatic NCC patients by [@bib0105] showed symptoms to be significantly associated with the presence of cysticercus in the subarachnoid base of the human brain. In the two pigs of this study that showed seizures, the latter was not observed. The latter might be explained by the fact that symptomatic subarachnoid NCC in humans takes years to develop symptoms ([@bib0040]).
Finally, [@bib0105] found clear differences between porcine and human cysticercosis however, our results run contrary to those conclusions. The clinical and biological similarities with human NCC associated epilepsy found in this study together with the functional and anatomical similarities of the pig and human brain make the pig a promising model to further understand the aetiology of epilepsy and the relationship between epilepsy and NCC in both humans and pigs ([@bib0110]).
Animals were treated with ivermectin twice at the beginning of the study. An inflammatory response around cysts which may lead to seizures cannot be ruled out, however, in a study by [@bib0070] no adverse effects were observed when ivermectin was used in pigs with cysticercosis ([@bib0070]). A study on humans, reported that ivermectin may have an effect on human cysticercosis, however the authors concluded that ivermectin did not give rise to any side effects in treated patients ([@bib0030]).
5. Conclusion {#sec0070}
=============
Results of this study have shown that pigs with NCC can develop clinical signs and suffer from seizures like humans with NCC associated epilepsy. Results of this study could potentially open up a new experimental pathway to explore the aetiology of neurological symptoms in humans. However, further studies are warranted to deepen our knowledge on why some pigs with NCC suffer from seizures, while others are asymptomatic. Moreover pigs with NCC and especially those clinically affected should be further studied, as the animal welfare aspect should not be neglected.
Appendix A. Supplementary data {#sec0080}
==============================
The following are Supplementary data to this article:
The study was partly supported by: the Center for Research in Pig Production and Health (CPH Pig) in Copenhagen, Denmark; Bill and Melinda Gates foundation under the ICTC-project (Integrated Control of Taeniosis/Cysticercosis in sub-Saharan Africa) and the Danish International Development Agency (DANIDA) with its project: SLIPP-project (Securing rural Livelihoods through Improved smallholder Pig Production in Mozambique and Tanzania), file no. 09-007LIFE.
The authors would like to acknowledge Mr M. Asheri and Mr F. Mafuru of the Department of Animal Science and Production at SUA for their assistance in taking care of the animals and Mr. M. Mlekwanyuma of the Department of Veterinary Pathology at SUA for the technical assistance during post-mortem examination of the pigs.
Supplementary data associated with this article can be found, in the online version, at [http://dx.doi.org/10.1016/j.vetpar.2016.02.025](10.1016/j.vetpar.2016.02.025){#intr0005}.
######
Number of brain cysticerci, pig age in months and presence or absence of seizures in 16 *T. solium* naturally infected pigs.
Table 1
-------------------------------------------
Pig\ No of brain\ Pig age\ Seizures
ID cysticerci (months)
------ -------------- ---------- ----------
1 247 36 1
2 241 36 1
3 47 18 0
4 418 18 0
5 4 24 0
6 178 12 0
7 4 24 0
8 88 12 0
9 27 12 0
10 21 18 0
11 36 8 0
12 90 12 0
13 10 8 0
14 117 18 0
15 41 18 0
16 136 18 0
-------------------------------------------
######
Distribution and localisation of brain cysticerci found in 16 *T*. *solium* naturally infected pigs.
Table 2
Right hemisphere Left hemisphere Both hemispheres
---------------------- ------------------ ----------------- ------------------ ------ ----- ------
Distribution
Frontal 309 34.0 281 37.6 590 36.3
Temporal 200 18.0 149 24.3 349 21.5
Parietal 114 19.8 164 13.9 278 17.1
Occipital 179 26.8 222 21.8 401 24.7
Cerebellum 3 0.4 3 0.4 6 0.4
Ventricles 17 1.0 8 2.1 25 1.5
Location
Subarachnoid dorsal 356 44.2 347 42.4 703 43.3
Subarachnoid base 41 5.1 35 4.3 76 4.7
Parenchyma 408 50.7 437 53.3 845 52.0
Cutting board 58
\*No: number.
| {
"pile_set_name": "PubMed Central"
} |
Citation {#SECID0EOGAC}
========
Costa DNR, Mathis WN, Marinoni L (2016) A revision of the shore-fly genus *Lamproclasiopa* Hendel (Diptera, Ephydridae). ZooKeys 631: 1--99. doi: [10.3897/zookeys.631.10718](10.3897/zookeys.631.10718)
Introduction {#SECID0E2GAC}
============
The need for revision of *Lamproclasiopa* [@B15] is abundantly apparent. Over half of the included species (13 of 24 species) were undescribed previous to this paper, and the genus has never been treated comprehensively. We also document in this revision how the nomenclatural history of the genus is reflective of changing concepts in classification that resulted from an improved understanding of the tribal phylogeny ([@B36], [@B37]).
While change is seemingly inevitable in science, including the classification of shore flies, an objective of taxonomy/systematics is a stable classification and nomenclature. These objectives are best achieved with discovery of accurate and well-documented phylogenetic relationships of the included taxa, as well as their accurate and detailed descriptions. Although we focus primarily on descriptive details in this revision, including keys, illustrations and photographs, we also provide a brief phylogenetic framework for the genus and to a lesser degree for species groups within the genus.
[@B15] first described *Lamproclasiopa* as a subgenus within the genus *Discocerina* Macquart and included *Discocerina facialis* Hendel, which he described in the same paper, as the type species. Hendel (p. 80) also included *Discocerina chalybea* Hendel in *Lamproclasiopa*. Hendel's name for *Discocerina facialis*, however, was preoccupied ([@B31]) and was corrected when [@B34]: 7) proposed *Lamproclasiopa hendeli* as a replacement new name. [@B10], [@B11], [@B12]) continued usage of *Lamproclasiopa* as a subgenus but changed its concept to include those species of *Discocerina* with a bare parafacial and three facial setae. The type species of *Lamproclasiopa* (*Lamproclasiopa hendeli*), however, only has two facial setae. Toward the end of Cresson's illustrious career ([@B10]: 116), he described yet another subgenus within *Discocerina*, *Basila* (type species: *Ditrichophora nadineae* Cresson) for those species with two facial setae and a bare parafacial. Cresson's precedent was adopted by Wirth in his catalogs for the shore-fly fauna of the New World ([@B33], [@B34]) and by [@B24] in their world catalog. [@B36] recharacterized *Lamproclasiopa* as part of their phylogenetic revision of the tribe Discocerinini, and the former subgenus was accorded generic status. Their recharacterization essentially reverted back to Hendel's original diagnosis of two facial setae, but they also included a bare parafacial. Their revised concept of *Lamproclasiopa* included 10 New World species and one species, *Lamproclasiopa laevior* (Cresson), from the Indian Subcontinent.
[@B12] published the first synopsis of Neotropical Discocerinini, which was the last in a series of synopses for the region that he produced. Cresson included 28 species in four genera, and some of these species were based on tentative identifications. Over twenty years later, [@B34] produced the first catalog of shore flies from the Neotropical Region and listed 30 species in the same four genera. In recent revisions ([@B25], [@B26]; [@B20]), herein, and in a forthcoming revision of the Neotropical species of *Discocerina* (Costa et al. in prep), we treat 38 species so far in seven genera from the fauna of Brazil alone. The purpose of this paper is to revise species of the genus *Lamproclasiopa*, including description of thirteen undescribed species.
This revision of *Lamproclasiopa* Hendel directly results from recent field work in southern Brazil that is part of an overall survey of the shore flies of this biologically diverse country. Although it is recognized that field and laboratory work are complimentary, even synergistic, the balance too often favors laboratory work. Herein we emphasize how extensive field work is a necessary and desirable compliment to comprehensive research at the desk level and eventually to publication. An objective of the field work in 2009-2010 and 2015 was the shore-fly fauna from the state of Paraná and to a lesser degree from Amazonas, Santa Catarina and São Paulo and resulted in numerous specimens of Discocerinini. Our sampling from Brazil (specimens recently collected and those from museums), however, is mostly from the southern states, and given this incomplete sampling, we anticipate additional species will yet be collected and eventually added to this diverse fauna. The same patchiness or in some cases the total lack of sampling applies to the Neotropical fauna in general. Responsible collecting of the Neotropical shore-fly fauna needs to be encouraged and to receive sustained support.
Methods and materials {#SECID0ENPAC}
=====================
The descriptive terminology, with the exceptions noted in [@B19] and [@B22], follows [@B28]. Because specimens are small, usually less than 2.60 mm in length, study and illustration of the male terminalia required use of a compound microscope. We have followed the terminology for most structures of the male terminalia that other workers in Ephydridae have used (references in [@B19]; [@B22], [@B23]), such as surstylus. [@B35] suggested that the pre- and postsurstylus correspond with the pre- and postgonostylus and that the subepandrial sclerite is the same as the medandrium. The terminology for structures of the male terminalia is provided directly on Figs [3--6](#F2){ref-type="fig"}. We use the term basal flagellomere for the large antennomere beyond the pedicel. We prefer this term over "first flagellomere" as there may be more than one flagellomere involved, and basal does not imply a number or numbers. We likewise do not use "postpedicel" ([@B30]) for this antennomere because at least the multisegmented arista is beyond the pedicel in addition to the large antennomere, and postpedicel is thus ambiguous and lacking precision.
Dissections of male terminalia were performed following [@B1] and [@B14]. Abdomens were removed with microforceps and macerated in a sodium hydroxide solution. Cleared genitalia were then transferred to glycerin for observation, description, and illustration. The dissected abdomen was placed in a plastic microvial filled with glycerin and attached to the pin supporting the remainder of the insect from which it was removed. These structures for species of *Lamproclasiopa* are minute, and for accurate determinations using them, we often had to use a compound microscope to see them clearly.
The species descriptions are composite and not based solely on holotypes. One head and two venational ratios used in the descriptions are based on three specimens (largest, smallest, and one other): gena-to-eye ratio -- genal height (immediately below maximum eye height)/eye height; costal vein ratio -- the straight line distance between the apices of R~2+3~ and R~4+5~/distance between the apices of R~1~ and R~2+3~; M vein ratio -- the straight line distance along vein M between crossveins dm-cu and r-m/distance apicad of dm-cu.
Distribution maps were made using ESRI ArcView GIS 3.2. Longitude and latitude coordinates were obtained for the locality where each specimen was collected and entered into a Microsoft Excel spreadsheet. If unavailable directly from specimen labels, longitude and latitude were estimated using gazetteers and maps to determine the geographical coordinates. Localities of specimens were plotted on a world land projection, presented within ESRI ArcView layouts and exported as encapsulated postscript(EPS) files.
The habitus illustrations are digital photographs taken with a Visionary Digital System. The images series obtained were combined by Zerene Stacker and Photoshop CS5 was used to adjust the color and make minor corrections (e.g., remove debris). Illustrations of male terminalia were made in Adobe Illustrator CS5.
Many specimens examined for this study are in the National Museum of Natural History, Smithsonian Institution, Washington, D.C.(USNM) and in the Universidade Federal do Paraná, Coleção Entomológica Padre Jesus Santiago Moure, Departamento de Zoologia, Curitiba, Paraná, Brazil(DZUP). We also borrowed and studied numerous specimens, especially primary types from the following museums:
AMNH American Museum of Natural History, New York, New York (David A. Grimaldi)
ANSP Academy of Natural Sciences of Philadelphia, Pennsylvania (Jon K. Gelhaus and Jason D. Weintraub)
BMNH The Natural History Museum (formerly the British Museum (Natural History)), London, England, United Kingdom (Kim Goodger)
DEBU Department of Environmental Biology, University of Guelph, Guelph, Ontario, Canada (Stephen A. Marshall)
INPA Instituto Nacional de Pesquisas da Amazônia, Manaus, Amazonas, Brazil (Márcio Oliveira, José Albertino Rafael and Rosaly Ale-Rocha)
IOC Instituto Oswaldo Cruz, Rio de Janeiro, Brazil (Jane Costa.)
MNCR-A Colección de Entomología del Museo Nacional de Costa Rica, San José, Costa Rica (including collections of former INBio) (Silvia Lobo C.)
MNRJ Museu Nacional do Rio de Janeiro, Rio de Janeiro, Brazil (Márcia Souto Couri)
MZLU Museum of Zoology, Lund University (Roy Danielsson)
MZUSP Museu de Zoologia da Universidade de São Paulo, São Paulo, Brazil (Carlos José Einicker Lamas)
NMW Naturhistorisches Museum, Wien, Austria (Peter Sehnal)
UMCE Instituto de Entomología, Universidad Metropolitana de Ciencias de la Educación, Santiago, Chile (Patricia Estrada M.)
Taxonomy {#SECID0EGPAE}
========
Tribe. Discocerinini
--------------------
Animalia
Diptera
Ephydridae
Cresson
1. Discocerinini[@B3]: 228 \[as Discocerini\]. Type genus: Discocerina[@B18]. [@B10]: 104 \[correct spelling, as a "new tribe" in key\]. [@B27]: 435 \[diagnosis, monophyly\]. [@B24]: 163-186 \[world catalog\]. [@B36]: 5-51 \[tribal revision\]. [@B37]: 1-34 \[phylogenetic review of tribe\].
### Diagnosis.
A tribe of Gymnomyzinae that is distinguished from other tribes of the subfamily by the following combination of characters:
*Head*: Frontal vitta (or ocellar triangle) mostly bare of setulae, not conspicuously setulose; ocellar setae well developed, inserted anterolaterad of anterior ocellus; reclinate fronto-orbital seta inserted anteromediad of proclinate fronto-orbital (if 2 proclinate fronto-orbital setae, reclinate seta inserted anteromediad of larger, posterior, proclinate seta); pseudopostocellar setae well developed, proclinate, slightly divergent, usually at least half length of ocellar setae. Pedicel bearing a large seta anterodorsally; arista bearing 4-6 dorsal rays, inserted along length of arista; conical process of basal flagellomere in lateral view finger-like. Face generally shallowly arched, frequently more prominent at level of dorsal facial setae, not conspicuously pitted, rugose, tuberculate, or carinate. Gena generally short (secondarily high in some species), bearing setulae (including midportion) and 1 large seta, its posterior (postgenal) margin rounded, not sharp. Oral opening and clypeus narrow; mouthparts generally dark colored; proboscis with number of pseudotracheae quite variable; lacinia Y-shaped with narrow posteromedial arm, dorsal arm spatulate; 2 different kinds of cibarium: (1) primitive type with dispersed medial sensillae arranged sparsely in a horizontal line; (2) advanced type with medial sensillae arranged densely in a sinuous line.
*Thorax*: Mesonotum generally microtomentose, frequently densely so, although variable; mesonotal setae weakly developed, only posteriormost pair of dorsocentral and acrostichal conspicuous; postsutural supra-alar seta usually evident although sometimes reduced or absent; prescutellar acrostichal setae inserted approximate and posterior of alignment of posteriormost dorsocentral setae; scutellar disc usually densely setulose; scutellum bearing 2 large, marginal setae: notopleural setae 2, inserted at same level near ventral margin, in some genera notopleuron bears setulae in addition to the two large notopleural setae (Figs [2](#F1){ref-type="fig"}, [9](#F4){ref-type="fig"}); anepisternum with 2 subequal setae inserted along posterior margin. Wing with vein R~2+3~ moderately long. Foreleg normally developed, not raptorial with greatly enlarged femur.
{#F1}
![*Lamproclasiopa laevior* (Cresson). (Sri Lanka. N. E. District: Horton Plains) **3** epandrium and cerci, posterior view **4** same, lateral view **5** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **6** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g002){#F2}
*Abdomen*: Five tergites visible, usually not densely covered with microtomentum. Male terminalia: Epandrium as inverted U, encircling cerci, anterior margin rounded, in lateral view with setae mainly on dorsum and along anteroventral margin; cerci paired, hemispherical, setose; presurstylus lacking or fused indistinguishably with ventral margin of epandrium; anterolateral arms of epandrium attached with ventral apex of gonites, middle of posterior margin a base for phallapodeme; phallapodeme situated under aedeagus, associated with hypandrium and with ventral part of base of aedeagus, ventral margin with lobate appendix providing attachment for genital muscles that move aedeagus; gonite paired, connecting sides of base of aedeagus and laterodorsal margin of epandrium, bearing 1 or some setulae; aedeagus tubular, tapered anteriorly; ejaculatory apodeme usually lacking, if present as a spatula ([@B37], Figs [99](#F34){ref-type="fig"}--[100](#F35){ref-type="fig"}).
### Discussion.
Several of the characters noted in the diagnosis are synapomorphies and establish the tribe's monophyly ([@B37]). These are as follows: (1) ocellar setae inserted slightly in front of alignment of anterior ocellus; (2) reclinate fronto-orbital seta inserted in front of proclinate fronto-orbital seta; (3) conical process of basal flagellomere in lateral view finger-like; (4) prescutellar acrostichal setae small and inserted close together and behind the transverse alignment of the posteriormost dorsocentral setae (secondarily lacking in some species); and (5) presurstylus of the male terminalia either lacking or fused indistinguishably with the ventral margin of the epandrium. Larvae are microphagous and in other aspects are similar to those of Hyadinini (Ilytheinae).
As currently characterized, the tribe Discocerinini is one of the richest tribes within the family Ephydridae (225 species), and numerous additional species, especially from tropical zones, remain to be described. Many of the undescribed species are already in collections, and undoubtedly numerous others await collection. With the recent phylogenetic review of the tribe ([@B37]) and description of additional genera and subgenera, there are now 13 genera and two subgenera. Two genera are monotypic and have relatively localized distributions: *Galaterina* in the Solomon and Andaman Islands and *Pectinifer* limited to the Neotropics ([@B20]). Other genera are more speciose and widespread. *Aquachasma* (24 species), *Facitrichophora* (4 species), *Hydrochasma* (10 species), and *Polytrichophora* (nominate subgenus) (22 species) are found in the New World. The distributions of *Lamproclasiopa* (24 species) and *Orasiopa* (15 species) extend from the New World into the Australasian and Oriental Regions. *Diclasiopa* (4 species), *Gymnoclasiopa* (25 species), *Hecamedoides* (26 species) and *Ditrichophora* (39 species) have been recorded from all Regions except the Neotropics. Two genera, *Discocerina* (20 species) and *Polytrichophora* (subgenus *Sklodowskopa*) (10 species), are essentially cosmopolitan.
### Phylogenetic considerations.
[@B37] proposed division of Discocerinini into four groups of genera (their proposed synapomorphies are provided in parentheses):
1. The *Gymnoclasiopa* group with *Gymnoclasiopa* (aedeagus with lateromedial appendices and facial setae arranged close to eye margin);
2. The *Diclasiopa* group with *Diclasiopa*, *Ditrichophora*, *Hecamedoides* and *Pectinifer* (gonite elongated, that is tapered apically);
3. The *Lamproclasiopa* group with *Galaterina*, *Lamproclasiopa*, and *Orasiopa* (subgenera *Orasiopa* and *Reymontopa*) (palpal setae with papilla-like bases); and
4. The *Discocerina* group with *Aquachasma*, *Discocerina*, *Facitrichophora*, *Hydrochasma* and *Polytrichophora* (nominate subgenus and subgenus *Sklodowskopa*) (reduced number of pseudotracheae, modified cibarium and the ventral receptacle bearing anterodorsal projection).
[@B37] acknowledged that groups three and four together (*Aquachasma*, *Discocerina*, *Galaterina*, *Hydrochasma*, *Lamproclasiopa*, *Orasiopa*, and *Polytrichophora*) form a clade that is the best supported lineage within the tribe, being based on (1) notopleuron setulose and (2) gonites elongated and bar-like without an anterior projection or the gonite is fused with the hypandrium (character 32.1-2). As such, we prefer the continued recognition of these seven genera as a single group, the *Discocerina* group, and use subgroups for further division of this group (the *Lamproclasiopa* and the *Discocerina* groups of [@B37]).
In the classification that [@B37] proposed (their character numbers are in parentheses), the *Lamproclasiopa*-subgroup has palpal setae with papilla-like bases (character 13). Within the *Lamproclasiopa* subgroup, the monophyly of the genus *Lamproclasiopa* is established by two characters (autapomorphies): (1) postsutural supra-alar (character 11 in [@B37]) and (2) prescutellar acrostichal setae greatly reduced or lacking (character 22 in [@B37]). The monophyly of its sister group, the combined *Galaterina* + *Orasiopa*, is confirmed by an increased number of pseudotracheae (convergent with *Pectinifer*). Thus, in the most recent classification, *Lamproclasiopa* is the sister-group of the combined lineage of *Galaterina* + *Orasiopa*, and these three genera together form an assemblage that is now the *Lamproclasiopa* subgroup.
### Key to genera of Discocerinini
---- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ----------------------------------------------
1 Notopleuron bare of setulae **2**
-- Notopleuron setulose in addition to 2 large setae **6**
2 Forefemur bearing distinct row of stout, short setae along apical half of posteroventral surface **3**
-- Forefemur lacking row of short, stout setae along posteroventral surface **4**
3 Face metallic shiny bearing white microtomentose spots laterally; forefemur slightly enlarged ***Pectinifer* Cresson**
-- Whole face shiny or completely covered with microtomentum; forefemur normally developed ***Ditrichophora* Cresson**
4 Postsutural supra-alar seta strong, distinct, longer than posterior notopleural seta. Face with dorsoclinate seta at lower lateral extremity ***Diclasiopa* Hendel**
-- Postsutural supra-alar seta very short or absent, if distinguishable distinctly shorter than posterior notopleural seta. Face without dorsoclinate seta at lower lateral extremity **5**
5 Hindtibia with a preapical, ventral, spur-like seta; face rather prominent at level of dorsal facial setae, sometimes transversely carinate; facial series of setae inserted in some distance to parafacial, comprising 2-3 large setae; dorsal seta inserted slightly medially from other setae and arising from distinct, shiny papilla, with a small, slightly dorsoclinate seta laterad of dorsal seta ***Hecamedoides* Hendel**
-- Hindtibia lacking a preapical, ventral spur-like seta; face rather flattened, antennal grooves not always sharply defined ventrally; facial series of setae inserted very close to parafacial, comprised of 2 large setae; dorsal seta not arising from a shiny papilla and lacking a smaller seta laterad of dorsal seta ***Gymnoclasiopa* Hendel**
6 Mesonotum bearing numerous, long setulae ***Galaterina* Zatwarnicki & Mathis**
-- Mesonotum lacking numerous, long setulae 7
7 Face with 2 or more conspicuous rows of setae/setulae on each side, parallel to facial suture setal row medial, with a row(s) of setulae between setal row and parafacial **8**
-- Face with a single row of setae laterally **10**
8 Face with setae and setulae of rows inclinate or ventroinclinate ***Facitrichophora* Mathis & Zatwarnicki**
-- Face with secondary series of dorsolaterally inclined setae laterad to primary series (*Polytrichophora* Cresson) **9**
9 Parafacials becoming 3-4 times wider ventrally; gena high, at least 1/4 eye height subgenus ***Polytrichophora* Cresson**
-- Parafacials 2-3 times wider ventrally; gena narrow, less than 1/4 eye height subgenus ***Sklodowskopa* Zatwarnicki**
10 Gena and lower part of parafacial broad; lateral margin of abdomen usually with gray to whitish microtomentose areas, these usually wedge shaped **11**
-- Gena and parafacial rather narrow; abdomen lacking wedge-shaped, light-colored areas laterally **12**
11 Head subglobose, oral opening comparatively large; dorsum of tergites darker dorsomedially than on lateral margins, but without contrasting areas ***Hydrochasma* Hendel**
-- Head not subglobose, oral opening comparatively small; dorsum of tergites 2-4 extensively dark gray to black with sharply contrasted gray lateral margin or with wedge-shaped silvery-gray areas ***Aquachasma* Zatwarnicki**
12 Parafacial bearing setulae ***Discocerina* Macquart**
-- Parafacial lacking setulae **13**
13 Facial series of setae 2, these well separated, distance between subequal to length of basal flagellomere; parafacial very narrow at anteroventral margin of eye; postsutural supra-alar and prescutellar acrostichal setae greatly reduced or lacking ***Lamproclasiopa* Hendel**
-- Facial series of setae 3-4, distance between setae conspicuously less than length of basal flagellomere; parafacial evenly wide throughout length; postsutural supra-alar and prescutellar acrostichal setae present (*Orasiopa* Zatwarnicki & Mathis) **14**
14 Species slender; antenna largely yellow; arista bearing 5 dorsal rays; palpus yellow; knob of halter dark; thorax and abdomen gray microtomentose; legs mostly yellow (sometimes midfemur dark) subgenus ***Orasiopa* Zatwarnicki & Mathis**
-- Species compact; antenna dark brown or black; arista bearing usually 7--11 dorsal rays; palpus brownish or black; knob of halter white; coloration of thorax and abdomen dark brown or black; legs mostly dark brown to black subgenus ***Reymontopa* Zatwarnicki**
---- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ----------------------------------------------
Genus. Lamproclasiopa
---------------------
Animalia
Diptera
Ephydridae
Hendel
1. Lamproclasiopa[@B15]: 79 \[as a subgenus of Discocerina\]. Type species: Lamproclasiopa facialis[@B15], original designation. [@B36]: 36 \[status as a genus; generic diagnosis\]. [@B34]: 7 \[Neotropical catalog\]. [@B24]: 168-169 \[world catalog\]. [@B37]: 16-19 \[recharacterization in phylogenetic review of tribe\].
2. Basila[@B10]: 116. Type species: Ditrichophora nadinae[@B3], original designation. [@B36]: 36 \[synonymy\].
### Diagnosis.
*Lamproclasiopa* is distinguished from other genera of Discocerinini by the following combination of characters: Small to medium-sized shore flies, body length 1.7--3.0 mm; generally sparsely to densely microtomentose, subshiny to dull species (Figs [2](#F1){ref-type="fig"}, [9](#F4){ref-type="fig"}, [25](#F10){ref-type="fig"}, [57](#F20){ref-type="fig"}, [69](#F24){ref-type="fig"}). *Head*: One proclinate and one reclinate pair of fronto-orbital setae. Arista usually bearing 5 dorsal rays, rarely 6. Face moderately prominent at level of dorsal facial seta; antennal grooves generally distinctly defined ventrally; face lacking secondary series of setae; facial setae 2, dorsal setae not arising from shiny papilla, lacking a dorsoclinate seta at lower lateral extremity; parafacial narrow to moderately wide throughout length, lacking ventroclinate setulae; gena generally short but very high in the *polita* group. Eye generally oval, moderately microsetulose, bearing interfacetal setulae (sometimes not discernible by light stereomicroscope). Proboscis with 7 pseudotracheae; cibarium of primitive type with 4 medial sensillae arranged in a horizontal row and 4 moderate posterior sensillae. *Thorax*: Anterior notopleural seta inserted near middle toward ventral margin, distance between anterior and posterior setae slightly less than half distance between postpronotal seta and anterior notopleural seta; notopleuron bearing several setulae in addition to 2 larger setae; presutural supra-alar seta usually present, well developed; postsutural supra-alar seta lacking; acrostichal setae, including prescutellar pair, lacking, only tiny setulae present. Wing variable, mostly to completely hyaline in most species but some with maculation pattern; costa bearing 5--6 long, dorsal setae between humeral and subcostal breaks; costal vein ratio varying between 0.40--0.90. Forefemur normally developed, lacking row of short, stout setae along posteroventral surface; hindtibia lacking a preapical, ventral, spur-like seta. Stem of halter blackish brown, knob white to whitish yellow. *Abdomen*: Tergites usually unicolorous, lacking pale-colored areas laterally; male tergite 4 longer than tergite 3. Male terminalia: Epandrium as inverted U in posterior view, dorsal arch complete; arms separate ventrally beyond cerci, surface covered with setae; cercus not fused with epandrium, in posterior view semicircular or crescent-shaped; gonites variously shaped, usually symmetrical, separate from hypandrium, in lateral view generally lunate without setulae; aedeagus longer than wide, mostly tubular, in ventral view navicular, without projections, in lateral view cigar-shaped or tapered toward apex; phallapodeme separate from aedeagus, in ventral view variously shaped; in lateral view irregularly triangular with distinct ventral projection; hypandrium in ventral view U- or Y-shaped with long posterolateral arms (incision reachs to 1/3--1/2 hypandrial length, in lateral view flat, sometimes slightly arched; ejaculatory apodeme absent. Female terminalia: Ventral receptacle without operculum, C-shaped stalk with broader head.
### Distribution
(Figs [7](#F3){ref-type="fig"}, [14](#F6){ref-type="fig"}, [36](#F13){ref-type="fig"}, [59](#F21){ref-type="fig"}, [81](#F28){ref-type="fig"}, [104](#F36){ref-type="fig"}, [111](#F39){ref-type="fig"}, [139](#F49){ref-type="fig"}). Oriental, Nearctic and Neotropical Regions.
### Discussion.
With the exception of *Lamproclasiopa laevior* (Cresson), which is a very disjunct species, occurring only on the Indian Subcontinent, the other congeners are found thus far only in the New World and there primarily in the Neotropics. In the older literature, including catalogs, this genus was frequently treated as a subgenus of *Discocerina* (see generic and species' synonymies).
We have arranged all recognized species into species groups based primarily on similarity, both external features and structures of the male terminalia. These groups are not necessarily monophyletic, although some are. Within a species group, the species are treated in alphabetical order.
### Key to species of *Lamproclasiopa*
------ -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------
1 Gena high to very high (gena-to-eye ratio between 0.30--0.64) and with an acutely sharp genal/postgenal margin (Figs [26](#F10){ref-type="fig"}, [38](#F14){ref-type="fig"}) **2**
-- Gena relatively short (gena-to-eye ratio less than 0.20) and with genal/postgenal margin rounded **4**
2 Frons nearly bare, shiny (Figs [37](#F14){ref-type="fig"}, [38](#F14){ref-type="fig"}). Distal 3 tarsomeres black; male tergite 5 narrowly rounded posteriorly ***Lamproclasiopa polita* (Edwards)**
-- Frons with anterior half densely microtomentose (Fig. [24](#F10){ref-type="fig"}). Distal 2--3 tarsomeres yellow; male tergite 5 truncate posteriorly **3**
3 Female frons with broad, transverse stripe on anterior half; male mesonotum with microtomentum on anterior third ***Lamproclasiopa lapaz* sp. n.**
-- Female frons mostly bare, shiny, at most with an anteromedial spot and at base of ocellar setae and on some parts of ocellar triangle (Fig. [25](#F10){ref-type="fig"}); male mesonotum with broad stripe of microtomentum, stronger anteriorly, becoming weaker posterior (Fig. [24](#F10){ref-type="fig"}) ***Lamproclasiopa auritunica* sp. n.**
4 Wing maculate, at least over crossveins or generally conspicuously infuscation (Figs [58](#F20){ref-type="fig"}, [62](#F22){ref-type="fig"}, [70](#F24){ref-type="fig"}) **5**
-- Wing generally hyaline, lacking a maculation pattern or general infuscate **8**
5 Wing generally infuscate with blackish veins and crossveins (Fig. [132](#F46){ref-type="fig"}) ***Lamproclasiopa fumipennis* (Wirth)**
-- Wing with pattern of spots or with a spot over crossveins but not generally infuscate **6**
6 Only crossveins r-m and dm-cu with darkened cloud; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein (Fig. [62](#F22){ref-type="fig"}) ***Lamproclasiopa mancha* sp. n.**
-- Wing with numerous dark spots; vein R~2+3~ angulate subapically and bearing a stump vein with a posteroapical orientation, a second stump vein near middle (Figs [58](#F20){ref-type="fig"}, [70](#F24){ref-type="fig"}) **7**
7 Mesonotum with 4 brown interrupted vittae, each with elongate, mostly separate spots, none in acrostichal row (Fig. [69](#F24){ref-type="fig"}) ***Lamproclasiopa painteri* (Cresson)**
-- Mesonotum with 7 brown, mostly entire vittae, including a medial vitta in acrostichal area (Fig. [57](#F20){ref-type="fig"}) ***Lamproclasiopa balsamae* (Cresson)**
8 Head, thorax, and abdomen generally shiny black, only tarsi and antennal grooves yellow or cinereous (Figs [1--2](#F1){ref-type="fig"}, [8--9](#F4){ref-type="fig"}, [98--99](#F34){ref-type="fig"}) **9**
-- Body with extensive surfaces sparsely to densely microtomentose **11**
9 Face completely shiny black (Fig. [8](#F4){ref-type="fig"}) ***Lamproclasiopa brunnea* sp. n.**
-- Face microtomentose, silver white to golden **10**
10 Frons generally shiny black; coxae blackish brown ***Lamproclasiopa hendeli* (Wirth)**
-- Frons microtomentose, silvery white or golden; forecoxae light gray (Fig. [98--99](#F34){ref-type="fig"}) ***Lamproclasiopa argentipicta* sp. n.**
11 Forebasitarsomere white, contrasted with black apical tarsomeres (Fig. [2](#F1){ref-type="fig"}) ***Lamproclasiopa laevior* (Cresson)**
-- Forebasitarsomere yellow to slightly blackish yellow, not distinctly contrasted with coloration of apical tarsomeres **12**
12 Gena relatively high, height subequal to height of basal flagellomere **13**
-- Gena relatively short, height about ½ height of basal flagellomere **18**
13 Presutural supra-alar seta lacking; katepisternum, especially anterior half, and anteroventral portion of anepisternum shiny black; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin **14**
-- Presutural supra-alar seta well developed; katepisternum and anepisternum thinly microtomentose, generally appearing dull, not shiny; forefemur with posteroventral setae slender, not stout and peg-like **15**
14 Pocket between epandrial arms of male uniformly U-shaped ***Lamproclasiopa nadineae* (Cresson)**
-- Pocket between epandrial arms of male bottle-shaped, with basal half as a narrower neck and apical half wider (Fig. [94](#F33){ref-type="fig"}) ***Lamproclasiopa aliceae* sp. n.**
15 Eyes covered with tiny, dense setulae; facial microtomentum gray; frons concolorous with mesonotum **16**
-- Eyes covered with very sparse setulae or bare; facial microtomentum gray or dark gray; anterior portion of frons usually yellowish orange to some degree **17**
16 Epandrium higher than wide; aedeagus thin, narrowly funnel-like, straight (Figs [135--138](#F48){ref-type="fig"}) ***Lamproclasiopa puella* (Cresson)**
-- Epandrium as high as wide; aedeagus wide, with apex acutely pointed and curved (Figs [128--131](#F45){ref-type="fig"}) ***Lamproclasiopa caligosa* sp. n.**
17 Aedeagus thin, narrowly funnel-like, straight; gonites without laterodorsal extensions (Fig. [116--117](#F41){ref-type="fig"}) ***Lamproclasiopa aracataca* (Cresson)**
-- Aedeagus wider, curved laterally; gonites with laterodorsal extensions (Fig. [130--131](#F45){ref-type="fig"}) ***Lamproclasiopa curva* sp. n.**
18 Antenna yellow (Figs [75](#F26){ref-type="fig"}, [86](#F30){ref-type="fig"}) **19**
-- Antenna extensively darkened dorsally, only basoventral portion of basal flagellomere and pedicel partially orange to yellowish **21**
19 Tibiae entirely black; presutural supra-alar seta well developed; frons and face distinctly two-toned ***Lamproclasiopa nana* (Williston)**
-- Tibiae partially or entirely yellow; frons and face generally unicolorous; presutural supra-alar lacking **20**
20 Tibiae black brown with the distal third yellow; frons with 2 small shiny black areas lateroanteriorly ***Lamproclasiopa furvitibia* sp. n.**
-- Tibiae entirely yellow; frons without shiny black areas (Figs [86--87](#F30){ref-type="fig"}) ***Lamproclasiopa xanthocera* sp. n.**
21 Face with a mediovertical, narrowly triangular, sparsely microtomentose spot (Fig. [18](#F8){ref-type="fig"}) ***Lamproclasiopa triangularis* sp. n.**
-- Face mostly shiny black, especially medial portion, this area lacking a microtomentose, triangular pattern **22**
22\. Foretarsus blackish yellow, apical tarsomeres becoming darker ***Lamproclasiopa zerafael* sp. n.**
-- Foretarsus yellow, apical tarsomeres yellow or becoming darker **23**
23 Foretarsus yellowish, apical 1-2 tarsomeres darkened ***Lamproclasiopa bisetulosa* (Cresson)**
-- Foretarsus completely yellow ***Lamproclasiopa ecuadoriensis* sp. n.**
------ -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------
The *laevior* group (*Lamproclasiopa laevior*) {#SECID0E11BG}
----------------------------------------------
**Diagnosis.** Body generally subshiny to shiny black. *Head*: Frons and face generally unicolorous; gena moderately high (gena-to-eye ratio 0.16-0.22); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta well developed; pleural areas generally shiny black. Wing hyaline to very faintly infuscate, lacking pattern of spots; vein R~4+5~ extended gradually toward costa, curved gently subapically, not angulate or bearing a stump vein. Femora and tibiae black; foretarsus with basal 2 tarsomeres white, tarsomere 3 darkened, apical 2 tarsomeres black; mid- and hindtarsus with basal 2 tarsomeres yellow, apical 3 tarsomeres blackish; forefemur lacking 4-5 stout, peg-like setae on apical third along posteroventral margin. *Abdomen*: Male terminalia: Epandrial sides in posterior view uniformly thin, thinner than width of cerci; cerci almost same height as epandrium; hypandrium generally U-shaped, narrow, without a wide projection ventrally.
**Remarks.** For the present, *Lamproclasiopa laevior* is the only included species in this species group, and among all known congeners, this is the only species known to occur only in the Old World. Its anomalous and disjunct distribution (see "Remarks" for this species) is perhaps partially explained by the proposed sister group of *Lamproclasiopa*, which is the combined lineage of *Galaterina + Orasiopa* ([@B37]). The latter two genera are found primarily in the Old World (*Orasiopa mera* (Cresson) occurs also in the New World, probably as an introduction), and there are many species of *Orasiopa* that occur in the Oriental Region ([@B24]). Thus far, however, no species of either *Galaterina* or *Orasiopa* are known from the Indian Subcontinent. We suggest that this may also well represent sampling error rather than actual distributions of all included species whether described or not. The shore-fly fauna of the Oriental Region has not been well sampled.
### Lamproclasiopa laevior
Animalia
Diptera
Ephydridae
(Cresson)
[Figs 1--2](#F1){ref-type="fig"} [, 3--6](#F2){ref-type="fig"} [, 7](#F3){ref-type="fig"}
1. Ditrichophora laevior[@B7]: 200 \[India. Darjeeling, Behar; HT ♂, ANSP (6509)\].
2. Discocerina (Lamproclasiopa) laevior. [@B11]: 59 \[generic combination\]. [@B24]: 168 \[world catalog\].
3. Lamproclasiopa laevior. [@B36]: 39 \[generic combination\].
#### Diagnosis.
This species is distinguished from other congeners by the following combination of characters: Small to moderately small shore-flies, body length 1.85--2.65 mm, generally black, subshiny to shiny species. *Head*: Frons black, very sparsely microtomentose, anterior half shiny, posterior subshiny, transition from shiny to subshiny gradual. Antenna black, densely microtomentose, appearing velvety black. Face black, sparsely microtomentose, antennal grooves and lateral areas shiny, otherwise subshiny, lacking prominent, vertical stripes; face bearing 2--3 larger facial setae, dorsal seta at about midfacial height, dorsomesoclinate; ventral seta just dorsad of epistomal margin, slightly dorsoclinate; parafacial blackish yellow; gena moderately high, gena-to-eye ratio 0.16--0.22. *Thorax*: Mesonotum uniformly sparsely microtomentose, black, subshiny; presutural supra-alar seta well developed; pleural area concolorous with mesonotum, subshiny black. Wing hyaline to very faintly infuscate, lacking pattern of spots; vein R~4+5~ extended at gradual to costa, not angulate subapically or bearing a stump vein; costal vein ratio 0.47--0.55; M vein ratio 0.60--0.62. Femora and tibiae black; Forefemur with posteroventral setae slender, not stout and peg-like; foretarsus with basal 2 tarsomeres white, tarsomere 3 darkened, apical 2 black; mid- and hindtarsus with basal 2 tarsomeres yellow, apical 3 tarsomeres blackish. *Abdomen*: Generally black, mostly subshiny to shiny, dorsum of tergites very sparsely and finely microtomentose. Male terminalia (Figs [3--6](#F2){ref-type="fig"}): Epandrium in posterior view (Fig. [3](#F2){ref-type="fig"}) generally vertically oval, each lateral arm narrow, almost parallel sided, acutely pointed ventrally, in lateral view (Fig. [4](#F2){ref-type="fig"}) with dorsal 2/3 rectangular, basal 1/3 almost twice width as dorsal portion, widest subventrally, apex broadly rounded, posteroventral portion bearing several larger setulae; cerci in posterior view (Fig. [3](#F2){ref-type="fig"}) elongate, narrowly semicircular, medial margin nearly straight, gradually tapered toward ventral apex, this apex acutely pointed, dorsal apex with medial short, digitiform extension, in lateral view (Fig. [4](#F2){ref-type="fig"}) semicircular; gonite in lateral view (Fig. [6](#F2){ref-type="fig"}) narrowly elongate, somewhat rod-like, ventral 1/3 narrower than dorsal 2/3, in ventral view (Fig. [5](#F2){ref-type="fig"}) very robustly C-shaped with medial concavity, wider than high; aedeagus in lateral view (Fig. [6](#F2){ref-type="fig"}) tubular, gradually tapered to apex, apex generally broadly rounded with very apex narrowed, curved anteriorly at nearly right angle; phallapodeme in lateral view (Fig. [6](#F2){ref-type="fig"}) very narrow, L-shaped, apex toward base of aedeagus acutely pointed, apex toward hypandrium slightly flared and truncate, in ventral view (Fig. [5](#F2){ref-type="fig"}) as an elongate hourglass, expanded at each apex and truncate; hypandrium in lateral view (Fig. [6](#F2){ref-type="fig"}) narrowed posteriorly, then abruptly expanded to widest point subanteriorly, anterior margin abruptly narrowed, digitiform, in ventral view (Fig. [5](#F2){ref-type="fig"}) generally U-shaped, lateral arms widest at midlength, anterior margin broadly rounded, deep, posterior emargination narrowed on anterior 1/3, thereafter posterior more than twice anterior width.
#### Type material.
The holotype male of *Ditrichophora laevior* Cresson is labeled "Darjeeling Fruhstorfer/9233/TYPE Ditrichophora LAEVIOR E. T. Cresson, Jr. \[maroon-red; "Ditrichophora LAEVIOR" handwritten\]." The holotype is double mounted (minuten pin in a rectangular block of foam), is in good condition (abdomen removed, dissected, and in an attached microvial), and is deposited in the ANSP (6509). A female paratype (ANSP) bears the same locality label as the holotype.
#### Type locality.
India. West Bengal: Darjeeling, Cooch Behar (26°24.7\'N, 89°23.1\'E). The holotype was apparently collected in the foothills of the Himalayas in the state of West Bengal between Darjeeling and Cooch Behar.
#### Other specimens examined.
INDIA. Meghalaya: Shillong (Botanical Gardens; 25°34.6\'N, 91°53.2\'E), 20 Apr 1980, A Freidberg (9♂, 2♀; USNM).
SRI LANKA. Central Province: Horton Plains (6°48.7\'N, 80°47.3\'E), 23 Apr 1980, W. Mathis, T. Wijesinhe & L. Jayawickrema (9♂, 4♀; USNM).
#### Distribution
(Fig. [7](#F3){ref-type="fig"}). Oriental: India (Meghalaya, West Bengal), Sri Lanka.
{#F3}
#### Remarks.
Although similar and perhaps related to *Lamproclasiopa ecuadoriensis*, this species is distinguished from congeners by the white forebasitarsomere, which is contrasted with the black, apical tarsomeres; the short gena (gena-to-eye ratio 0.10--0.14); and the shape of structures of the male terminalia (Figs [3--6](#F2){ref-type="fig"}).
The distribution of this species is anomalously disjunct. Except for this species, which is found on the Indian Subcontinent, all other congeners occur in the New World, especially in tropical zones, which are many thousands of kilometers from India and Sri Lanka. This distributional anomaly prompts a number of questions. For example, is this species indeed a valid congener? If so, what is it related to? How did it come to be found on the Indian Subcontinent when other congeners occur in the New World? Although our responses to these and other questions are not wholly satisfactory, often being based on tenuous evidence, we offer some thoughts and observations.
Based on morphological evidence, both external and from structures of the male terminalia, we affirm that this is a congener within the genus *Lamproclasiopa*. Within *Lamproclasiopa*, we suggest that this species, being generally characterized by homoplasious characters, would probably be near the base of the evolutionary stem or node giving rise to all other included congeners. See also our remarks under the species group.
As we noted in the introduction, however, characterization of *Lamproclasiopa* has only become reasonably well resolved in recent decades, resulting in earlier recognized species, including this species, being first described in another genus. [@B7] first described this species in *Ditrichophora* then transferred it to the subgenus *Lamproclasiopa* within *Discocerina* ([@B11]). Over 50 years later, [@B36] accorded generic status to *Lamproclasiopa* and included this species along with other congeners in this genus.
The *hendeli* group (*Lamproclasiopa brunnea*, *Lamproclasiopa hendeli*) {#SECID0EWKAI}
------------------------------------------------------------------------
**Diagnosis.** Body generally subshiny to shiny black, only tarsi and antennal grooves yellow or cinereous. *Head*: Frons and face generally unicolorous; frons sparsely microtomentose; genal height variable, moderately high to high (gena-to-eye ratio 0.12--0.25); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta lacking or indistinguishable from surrounding setae; katepisternum and anepisternum thinly microtomentose, generally appearing dull, not shiny. Wing generally hyaline to very faintly infumate; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with posteroventral setae slender, not stout and peg-like; tarsi yellowish. *Abdomen*: Male terminalia: Cerci around 2/5 height of epandrium; hypandrium generally wide, with arms long or short dorsally, not U-shaped.
**Remarks.** This species group is mostly based on homoplasious characters, and we cannot confirm its monophyly. The two included species are similar to each other and the species group can be diagnosed. These are the bases for recognition of this species group.
### Lamproclasiopa brunnea sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/2875BDA6-0054-4F85-AC3A-8B75173BBA37
[Figs 8--9](#F4){ref-type="fig"} [, 10--13](#F5){ref-type="fig"} [, 14](#F6){ref-type="fig"}
#### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Moderately small shore flies, body length 2.60 mm. *Head*: Frons shiny black brown, with ocellar triangle and fronto-orbital plate densely setulose. Antenna blackish brown, slightly lighter than head, posterior margin of flagellomere light brown; face blackish brown, shiny; parafacial yellowish brown. Gena moderately high, gena-to-eye ratio 0.12--0.15. *Thorax*: Mesonotum shiny black, covered with microtomentum; presutural supra-alar seta lacking or indistinguishable from surrounding setae; pleural region less microtomentose, anepisternum and katepisternum almost bare, shiny black. Wing hyaline, lacking any pattern or markings; costal vein ratio 0.52--0.64; M vein ratio 0.59--0.65; Legs blackish brown except yellowish tarsi; forefemur with posteroventral setae slender, not stout and peg-like. *Abdomen*: Generally shiny black, bare of microtomentum; tergites 3--5 equal in length and larger than tergites 1--2. Male terminalia (Figs [10--13](#F5){ref-type="fig"}): Epandrium in posterior view (Fig. [10](#F5){ref-type="fig"}) generally oval, higher than wide, dorsal portion very thin, each lateral arm gradually becoming wider ventrally, widest on apical third, apex rounded, oriented medially, ventral half with slightly increased number of setulae, in lateral view (Fig. [11](#F5){ref-type="fig"}) as roughly rectangular, slightly wider ventrally with rounded posteroventral portion, pointed anteroventrally; cerci in posterior view (Fig. [10](#F5){ref-type="fig"}) elongate, thin, generally shallowly arched, ventral and dorsal apices acutely pointed, setulose evenly along length, in lateral view (Fig. [11](#F5){ref-type="fig"}) as an inverted drop; gonite in lateral view (Fig. [13](#F5){ref-type="fig"}) rod-like, dorsal fourth curved basally and pointed, in ventral view (Fig. [12](#F5){ref-type="fig"}) irregularly bar-like, curved, apex toward aedeagal base tapered, pointed apically, apex toward hypandrium bluntly rounded; aedeagus in lateral view (Fig. [13](#F5){ref-type="fig"}) more or less rectangular, basal third wider than apical portion, narrowest medially, apex angulate, in ventral view (Fig. [12](#F5){ref-type="fig"}) elongate, thin, after slightly bulbous base parallel sided, apex tapered, pointed; phallapodeme in lateral view (Fig. [13](#F5){ref-type="fig"}) as a dissected triangle, with an obvious keel, vertex toward hypandrium sharply tapered, acutely pointed, vertex toward aedeagal base thumb-like, keel narrow, moderately elongate, pointed, in ventral view (Fig. [12](#F5){ref-type="fig"}) as an asymmetrical spool, wider toward aedeagal base, apex toward hypandrium shallowly trilobed; hypandrium in lateral view (Fig. [13](#F5){ref-type="fig"}) elongate, thin, irregularly rod-like, shallowly sinuous, both apices narrowly rounded; in ventral view (Fig. [12](#F5){ref-type="fig"}) robustly Y-shaped, with base wide, lateral margins almost serrate, arms of Y posterior, flared posterolateral, each arm thin, digitiform, forming posterior, evenly rounded, moderately deep emargination.
{#F4}
![*Lamproclasiopa brunnea* sp. n., male paratype (Costa Rica. San José. Moravia) **10** epandrium and cerci, posterior view **11** same, lateral view **12** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **13** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g005){#F5}
#### Type material.
The holotype male of *Lamproclasiopa brunnea* is labeled "COSTA RICA. Prov. San José. Moravia. Zurquí de Moravia, Tower path. 1600m. 2--9 AGO 2013. Proyeto ZADBI. Mix methods, ZADBI-1076/HOLOTYPE ♂ *Lamproclasiopa brunnea* Costa, Mathis & Marinoni USNM \[red\]." The holotype is double mounted (glued to a paper triangle) and is in very good condition, and is deposited in MNCR-A. Thirty-eight paratypes (20♂, 18♀; MNCR-A, USNM) bear the same label data as the holotype. Other paratypes are as follows: COSTA RICA. **Cartago.** Paraíso, Parque Nacional Tapantí (09°43.3\'N, 83°46.5\'W; 1600 m), 4--11 Ago 2013, Proyeto ZADBI (1♀; MNCR-A). **Guanacaste.** Macizo Miravalles, Cabro Muco Station (10°43.1\'N, 84° 51.3\'W; 1100 m), 15 Mar--2 Abr 2003, J. Azoifeifa. (1♀; MNCR-A).
#### Type locality.
Costa Rica. San José. Zurquí de Moravia (10°02.8\'N, 84°0.6\'W; 1588 m).
#### Distribution
(Fig. [14](#F6){ref-type="fig"}). Neotropical: Costa Rica (Cartago, Guanacaste, Puntarenas, San José).
{#F6}
#### Etymology.
The species epithet, *brunnea*, is of Latin derivation and means brown, referring to the dark brown color of this species, especially its head.
#### Remarks.
This species is very similar and superficially appear to be closely related to *Lamproclasiopa ecuadoriensis*, as evidenced by the generally dark brown body color and yellowish parafacies of both species. The thoracic pleural area of *Lamproclasiopa brunnea*, however, is more sparsely microtomentose. Moreover, the frontal microtomentum of *Lamproclasiopa brunnea* covers only the fronto-orbital plates and the ocellar triangle, which easily distinguishes this species from *Lamproclasiopa ecuadoriensis*.
### Lamproclasiopa hendeli
Animalia
Diptera
Ephydridae
(Wirth)
[Figs 15--17](#F7){ref-type="fig"} [, 36](#F13){ref-type="fig"}
1. Discocerina (Lamproclasiopa) facialis[@B15]: 79.
2. Discocerina (Lamproclasiopa) hendeli[@B34]: 7 \[replacement name for Discocerina facialis[@B15], not [@B31]\]. [@B24]: 168 \[world catalog\].
3. Lamproclasiopa hendeli. [@B36]: 39 \[generic combination\].
#### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Moderately small shore flies, body length 2.50 mm. *Head*: Frons broader than long, 1.5 times as wide as an eye; glossy black with anterior margin reddish yellow, between ocelli and fronto-orbits somewhat dull by very fine reddish brown microtomentose. Pedicel black, slightly whitish dusted dorsally; basal flagellomere red yellow, darkened along outer margin; arista with 5 long rays dorsally. Face protruded in lateral view; dorsal half of face with distinct antennal grooves, these separated by a vertical ridge, just ventrad of ventral margin of antennal grooves a transverse ridge. Ventral portion of face flat, receded towards oral margin; face silvery white, microtomentose with 2 vertical stripes toward middle. Gena moderately high, ¼ height of eye; silvery white at edge of eye; with 1 strong seta. *Thorax*: Shiny black, smooth; central portion and scutellar disc covered with very short and fine reddish brown microtomentum that reduces the shine; within this microtomentum on mesonotum, shiny black dots at bases of short setae. No prescutellar pair of setae or supra-alar seta. Legs shiny black; tarsi reddish yellow. Wing grayish hyaline, with yellow veins; costal section II about 1.5 times as long as costal section I; veins R~4+5~ and M~1~ parallel, last section of M~1~ twice as long as penultimate section. Halters with black stem and light yellow knob. Calypteres with dark brown margins and hairy. *Abdomen*: Concolorous with mesonotum, with dense, short, black setulae; tergites 3--5 almost equally long. Male terminalia (Figs [15--17](#F7){ref-type="fig"}): Epandrium in posterior view (Fig. [15](#F7){ref-type="fig"}) roundly U-shaped, except for ventral gap, oval, only slightly narrower dorsally and ventrally, widest at midheight, dorsal arch relatively narrow, each lateral arm widest ventrally, ventral margin evenly rounded, lacking medial or ventral extensions, ventral portion bearing numerous, loosely clustered, long setulae; cercus hemispherical, tapered ventrally to pointed apex, more setulose dorsally, medial margin straight; gonite in lateral view rod-like, shallowly curved, banana-like, very slightly wider toward hypandrium than toward aedeagal base, in ventral view shallowly curved, distinctly expanded on portion toward hypandrium with extension toward aedeagal base tapered to a narrow apex, lateral margin with a wide, short irregularly shaped keel; aedeagus in lateral view (Fig. [17](#F7){ref-type="fig"}) very elongate, narrowly triangular, almost parallel sided, tapered evenly to moderately narrow, rounded apex, in ventral view as an elongate, very narrow, parallel-sided, rod-like structure, basal end shallowly and bluntly rounded, apical 1/5 tapered toward narrow point; phallapodeme in lateral view (Fig. [17](#F7){ref-type="fig"}) more or less irregularly triangular, with moderately long, narrow extensions toward aedeagal base and hypandrium, keel distinct, relatively narrow, irregular, somewhat pointed apically; hypandrium in lateral view (Fig. [17](#F7){ref-type="fig"}) generally narrow, rod-like, very shallowly sinuous, in ventral view almost rectangular, wider than long, anterior margin shallowly emarginate, posterior margin more deeply emarginate, moderately deeply and broadly U-shaped, depth of emargination about half length of anterior portion.
![*Lamproclasiopa hendeli* (Wirth) (redraw from [@B36]) **15** epandrium and cerci, posterior view **16** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **17** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g007){#F7}
#### Type material.
The holotype male of Discocerina (Lamproclasiopa) facialis Hendel was published as "1 ♂, Cuesta von Cillutiucara, Bolivien, 3200 m (Fassl) ." The holotype male is housed in the NMW.
#### Type locality.
Bolivia. La Paz: Sillutincara (= Cilluntincara) (16°17\'S, 67°54\'W; 3200 m).
#### Distribution
(Fig. [36](#F13){ref-type="fig"}). Neotropical: Bolivia (La Paz).
#### Remarks.
Although similar and apparently closely related to *Lamproclasiopa argentipicta*, this species is distinguished from this congener by the shiny black frons and blackish brown coxae.
[@B12]: 148) suggested that this species could be *Discocerina nitida* Cresson. We confirm that this is an included species in *Lamproclasiopa*, and further, that it is not related to *Discocerina nitida*. The illustrations of this species in [@B36] are of the holotype and are further evidence that it is a congener within *Lamproclasiopa*.
Our description of external features was extracted and interpreted from M. Kotrba's English translation of Hendel's original description, as we did not have access to the holotype. Although we have repeatedly inquired about and asked to examine the holotype male of this species, our requests were ignored. Fortunately, however, we have the illustrations of structures of the male terminalia that Zatwarnicki ([@B36]) produced of the holotype male, and these are the basis for much of our diagnosis of this species.
The *triangularis* group (*Lamproclasiopa triangularis*) {#SECID0EGCBI}
--------------------------------------------------------
**Diagnosis.** Body generally subshiny to shiny black. *Head*: Frons mostly brownish black to black; frons sparsely microtomentose; ventral half of face with a microtomentose triangle, sometimes dorsal angle of triangle extended dorsally to level of base of antennae, base of triangle sometimes partially bare, otherwise face largely bare, black except for yellow to yellowish orange lateral margins adjacent to parafacial and antennal grooves; gena relatively short (gena-to-eye ratio 0.06--0.10); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta lacking or indistinguishable from surrounding setae; katepisternum and anepisternum thinly microtomentose, generally appearing dull, not shiny. Wing hyaline, lacking any pattern or markings; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with posteroventral setae slender, not stout and peg-like; tarsi yellowish. *Abdomen*: Male terminalia: Cerci slightly wider dorsally than ventrally; aedeagus in lateral view robust, almost rectangular, only slightly tapered on apical half.
**Remarks.** This species group currently includes only *Lamproclasiopa triangularis*, which exhibits unique character states in the triangular microtomentose area on the face and the robust, rectangular aedeagus in lateral view.
### Lamproclasiopa triangularis sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/1A4E189A-B737-4B14-8591-12D1102AE698
[Figs 18--19](#F8){ref-type="fig"} [, 20--23](#F9){ref-type="fig"} [, 81](#F28){ref-type="fig"}
#### Diagnosis.
This species is distinguished from other congeners by the following combination of characters: Small to moderately small shore-fly species, body length 1.65--2.10 mm; generally black, subshiny to shiny. *Head*: Frons generally mostly brownish black to black, moderately microtomentose, subshiny; mesofrons more microtomentose, tan to brown; some specimens with 2 gray spots along ventral margin just dorsad of antennal bases. Antenna mostly black, especially scape and pedicel, only basal flagellomere with ventrobasal area with some yellow to yellowish orange coloration. Ventral half of face with a microtomentose triangle (Fig. [18](#F8){ref-type="fig"}), sometimes dorsal angle of triangle extended dorsally to level of base of antennae, base of triangle sometimes partially bare, otherwise face largely bare, black except for yellow to yellowish orange lateral margins adjacent to parafacial and antennal grooves; bearing 2 larger facial setae, dorsal seta at about midfacial height, dorsomesoclinate; ventral seta just dorsad of epistomal margin, slightly dorsoclinate; parafacial silvery white. Gena relatively short, gena-to-eye ratio 0.06--0.10. *Thorax*: Mesonotum uniformly whitish gray microtomentose; pleural area very sparsely microtomentose, mostly dark brown, partially subshiny; presutural supra-alar seta lacking or indistinguishable from surrounding setae. Wing hyaline, lacking any pattern or markings; costal vein ratio 0.74--0.79; M vein ratio 0.55--0.59. Legs, except tarsi, black; forefemur with posteroventral setae slender, not stout and peg-like; tarsi yellow; apical tarsomere slightly darker than other tarsomeres. *Abdomen*: Generally black, subshiny to mostly shiny, dorsum of tergites very sparsely and finely microtomentose, faintly whitish gray; sternite 3 of male rectangular, parallel sided, length twice width; sternite 4 of male rectangular, length almost twice width; sternite 5 of male a single, deeply U-shaped plate, length about twice width, opening of U posterior. Male terminalia (Figs [20--23](#F9){ref-type="fig"}): Epandrium in posterior view (Fig. [20](#F9){ref-type="fig"}) almost as wide as high, as an inverted U, dorsal arch very thin, verticolateral arms gradually becoming wider, width wider than width of cercus, in lateral view (Fig. [21](#F9){ref-type="fig"}) narrow, elongate, overall as a robust, irregular tear drop with an anteroventral, short, shallowly pointed projection; cercus in posterior view (Fig. [20](#F9){ref-type="fig"}) bar-like, elongate, narrow, with dorsal half wider than ventral half, slightly tapered from dorsum to ventral margin, not fused with ventral margin of cercal cavity, in lateral view (Fig. [21](#F9){ref-type="fig"}) elongate, dorsal half slightly wider than ventral half; gonite in ventral view (Fig. [22](#F9){ref-type="fig"}) as an inverted, robust comma, in lateral view (Fig. [23](#F9){ref-type="fig"}) bar-like, shallowly arched; aedeagus in lateral view (Fig. [23](#F9){ref-type="fig"}) robust, narrowly and irregularly rectangular, widest basally, thereafter slightly tapered to truncate apex, in ventral view (Fig. [22](#F9){ref-type="fig"}) elongate, narrow, narrowly ovate, acutely pointed apically; phallapodeme in lateral view (Fig. [23](#F9){ref-type="fig"}) as a deeply dissected triangle, extended keel rounded apically, each extended arm narrow, in ventral view (Fig. [22](#F9){ref-type="fig"}) narrow spindle shaped with a medial bulge, basal and apical widths subequal; hypandrium in ventral view (Fig. [22](#F9){ref-type="fig"}) as a robust V-shaped structure, vertex especially robust, in lateral view (Fig. [23](#F9){ref-type="fig"}) narrow, elongate, shallowly arched.
{#F8}
![*Lamproclasiopa triangularis* sp. n., male paratype (Peru. Madre de Dios: Manu) **20** epandrium and cerci, posterior view **21** same, lateral view **22** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **23** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g009){#F9}
#### Type material.
The holotype male of *Lamproclasiopa triangularis* is labeled "PERU. Madre de Dios: Manu, Rio Manu, 250 m\[,\] Pakitza,12°7\'S, 70°58\'W \[11°56.6\'S, 71°16.9\'W\], 9--23 Sep 1988\[,\] Amnon Freidberg/USNM ENT 00118309 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa triangularis* Costa, Mathis & Marinoni, USNM \[red\]." The holotype is double mounted (minuten pin in a block of plastic), is in excellent condition, and is deposited in the USNM. Forty three paratypes (21♂, 22♀; DZUP, INPA, USNM) bear the same label data as the holotype but with W. N. Mathis as the collector.
#### Type locality.
Peru. Madre de Dios: Río Manu, Pakitza (11°56.6\'S, 71°16.9\'W; 250 m).
#### Other specimens examined.
BRAZIL. **Amazonas**: Manaus, Universidade Federal do Amazonas (03°05.9\'S, 59°58.2\'W; 50 m), 7 May 2010, D. and W. N. Mathis (3♂, 6♀; INPA, USNM). **Paraná**: Antonina, Reserva Natural Rio Cachoeira (25°19\'S, 48°41,6\'W), 8 Feb 2010, D. Negoseki (1♂; DZUP). **São Paulo**: Ubatuba, Cachoeira da Lage (23°17.6\'S, 44°52.1\'W; 100 m), 30 Mar 2010, D. and W. N. Mathis (1♂, 3♀; DZUP, USNM).
ECUADOR. **Orellana**: Rio Tiputini (0°38.2\'S, 76°8.9\'W), 12--26 Aug 1999, W. N. Mathis, A. Batista, M. Kotrba (7♂, 2♀; USNM).
GUYANA. Kaieteur Falls (5°10.7\'N, 59°29.2\'W; 570 m), 7 Apr 1994, W. N. Mathis (1♂, 2♀; USNM). Kanuku Mountains, Kumu River and Falls (3°15.9\'N, 59°43.5\'W), 28--30 Apr 1995, W. N. Mathis (1♂; USNM). Kanuku Mountains, Moco Moco River (3°18.2\'N, 59°38.9\'W), 29 Apr 1995, W. N. Mathis (2♂, 1♀; USNM). Lethem (25 km SE; 3°18.2\'N, 59°38.9\'W), 4--5 Apr 1994, W. N. Mathis (2♂; USNM).
PERU. **Madre de Dios**: Río Manu, Pakitza (11°56.6\'S, 71°16.9\'W; 250 m), 9--23 Sep 1988, A. Freidberg, W. N. Mathis (21♂, 23♀; USNM).
#### Distribution
(Fig. [81](#F28){ref-type="fig"}). Neotropical: Brazil (Amazonas, Paraná, São Paulo), Ecuador (Orellana), Guyana, Peru (Madre de Dios).
#### Etymology.
The species epithet, *triangularis*, is of Latin derivation, meaning triangular, and refers to the small triangular microtomentose area on the face of this species.
#### Remarks.
This species is distinguished from congeners by the triangle-shaped facial spot that is sparsely microtomentose. The triangle is situated medially on the ventral half of the face and is sometimes slightly elongated. Also distinguishing this species are the robust (thick) aedeagus in lateral view with its truncate apex and the narrow and elongated keel of the phallapodeme.
The *polita* group (*Lamproclasiopa auritunica*, *Lamproclasiopa lapaz*, *Lamproclasiopa polita*) {#SECID0EKSBI}
-------------------------------------------------------------------------------------------------
**Diagnosis.** Body generally shiny black. *Head*: Gena high to very high, gena-to-eye ratio 0.30--0.64; merger of posterior genal margin and lateral postgenal margins forming a sharply angulate, joint margin (convergently similar to *Athyroglossa*). *Thorax*: Presutural supra-alar seta well developed; katepisternum and anepisternum mostly to entirely bare, shiny black. Wing generally hyaline to faintly infumate; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin. *Abdomen*: Male terminalia: Keel of phallapodeme short and sometimes difficult to discern.
**Remarks.** This species group appears to be monophyletic based on the following two synapomorphies: (1) gena high to very high, gena-to-eye ratio 0.30--0.64; (2) merger of posterior genal margin and lateral postgenal margins forming a sharply angulate, joint margin (convergently similar to *Athyroglossa*).
Both of the new species included in the species group, *Lamproclasiopa auritunica* and *Lamproclasiopa lapaz*, are very similar to each other, and together, form a separate lineage that is distinguished by synapomorphies, such as the broad, truncate dorsal, epandrial margin in posterior view, the anterior extension of the epandrium, and the division of the aedeagus into a basiphallus and distiphallus. Certainly these characters are unique within *Lamproclasiopa*, and are the basis for the monophyly of these two species as a separate and distinct lineage.
### Lamproclasiopa auritunica sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/1854E006-D51C-4FA1-B50D-DA92F39754E7
[Figs 24--26](#F10){ref-type="fig"} [, 27--30](#F11){ref-type="fig"} [, 36](#F13){ref-type="fig"}
#### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Moderately small shore flies, body length 2.30--2.80 mm; generally a shiny black species. *Head*: Frontal microtomentum sexually dimorphic; male with dense and extensive microtomentum over slightly more than anterior half of frons, also within ocellar triangle (Fig. [24](#F10){ref-type="fig"}), female with microtomentum only around bases of fronto-orbital setae and ocellar setae, thereafter as a thin stripe within ocellar triangle extended posteromedially, convergent within ocellar triangle, and a small medial spot just before anterior margin (Fig. [25](#F10){ref-type="fig"}). Antenna black, with dense microtomentum especially evident on basal flagellomere laterally; arista bearing 3--4 dorsal rays (usually 4). Face with moderately deep antennal grooves on dorsal half, shallowly angulate in lateral view, vortex of angle at midheight near dorsal facial seta, ventral half of face receded, facial microtomentum in both sexes generally dense, golden brown dorsally, becoming more silvery ventrally, female with some bare areas, especially at base of facial setae and adjacent to parafacial; parafacial and anterior half of gena densely microtomentose in male, in female with thin area microtomentose at anterior and ventral margins of eye, otherwise bare, shiny; gena very high, gena-to-eye ratio 0.42--0.64; posterior margin of gena at merger with lateral margin of postgenal sharply angulate. *Thorax*: Mesonotum shiny black, pattern of microtomentum evident as a broad band, much denser anteriorly, becoming sparse posteriorly, microtomentum extended onto scutellar disc; lateral to microtomentose band mostly bare, shiny except for microtomentose anterior surface of postpronotum and ventral margin of notopleuron; presutural supra-alar seta well developed; pleural region generally bare, shiny black. Wing hyaline to faintly infumate, faintly tannish, lacking any pattern or markings. Costal vein ratio 0.50--0.58; M vein ratio 0.59--0.78. Coxae black, shiny; forecoxa with vertical microstriae; femora and tibiae black; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; basal 2--3 tarsomeres yellow, apical 2--3 brownish black to dark brown. Halter with base black, knob whitish yellow. *Abdomen*: Generally shiny black; male tergite 5 truncate apically. Male Terminalia (Figs [27--30](#F11){ref-type="fig"}): Epandrium in posterior view (Fig. [27](#F11){ref-type="fig"}) irregularly hexagonal with dorsal 2/3 quadrate, as wide as high, corners rounded, ventral third with lateral margin slanted medially ventrally and ventral margin shallowly concave, dorsal portion thinly developed, lateral portions wide, each subequal to width of cercal cavity, setulae more or less evenly distributed laterally, thereafter with a gap, then clumped ventrolaterally, in lateral view (Fig. [28](#F11){ref-type="fig"}) more or less and irregularly L-shaped, thin dorsally, with an obtusely angulate ventral portion and a moderately narrow anterior extension with a flared, somewhat truncate anterior margin; cerci in posterior view (Fig. [27](#F11){ref-type="fig"}) elongate, moderately thin, generally shallowly arched, lunate, ventral and dorsal apices tapered, in lateral view elongate, narrow, elliptical; aedeagus in lateral view (Fig. [30](#F11){ref-type="fig"}) as 2 structures, basiphallus L-shaped with a digitiform process from one arm, distiphallus shallowly arched, wider basally, with ribbon-like extension, in ventral view (Fig. [29](#F11){ref-type="fig"}) with basiphallus spindle-like, elongate, distiphallus rectangularly ovate; phallapodeme in lateral view (Fig. [30](#F11){ref-type="fig"}) L-shaped, each arm narrow and of equal length, in ventral view Y-shaped with base shorter than either arm; gonite in lateral view irregularly pear-like, in ventral view (Fig. [29](#F11){ref-type="fig"}) rod-like; hypandrium in lateral view (Fig. [30](#F11){ref-type="fig"}) thin, elongate, width irregular and with a short, thin process near middle, in ventral view (Fig. [29](#F11){ref-type="fig"}) robust, with anterior 2/3 diamond-shaped, posterior third widely and shallowly U-shaped.
{#F10}
![*Lamproclasiopa auritunica* sp. n. (Bolivia. Oruro: Paznã) **27** epandrium and cerci, posterior view **28** same, lateral view **29** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **30** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g011){#F11}
#### Type material.
The holotype male of *Lamproclasiopa auritunica* is labeled "**BOLIVIA. Oruro**: Paznã (S. of the town; 18°36.2\'S, 66°54.7\'W, 3750 m), 22 Mar 2001\[,\] Wayne N. Mathis/USNM ENT 00119995 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa auritunica* Costa, Mathis & Marinoni, USNM \[red\]." The holotype is double mounted (minuten pin in a plastic block) and is in very good condition, and is deposited in USNM. Three paratypes (1♂, 2♀; USNM) bear the following label data: Bolivia. Oruro: Challapata (45 km S; 19°12.9\'S, 66°47.7\'W, 3690 m), 22 Mar 2001, A. Freidberg, W. N. Mathis (1♂, 2♀; USNM). Bolivia. La. Paz: Tiahuanaco Ruins (16°33.7\'S, 68°40.7\'W; 3870m), 28 Mar 2001, W. N. Mathis (1♀;USNM); Patacayama (7 km NE; 17°9.5\'S, 67°56.7\'W; 3800m), 21 Mar 2001, W. N. Mathis (1♀; USNM).
#### Type locality.
Bolívia. Oruro: Paznã (S. of the town; 18°36.2\'S, 66°54.7\'W, 3750 m).
#### Distribution
(Fig. [36](#F13){ref-type="fig"}). *Neotropical*: Bolivia (La Paz, Oruro).
#### Etymology.
The species epithet, *auritunica*, is of Latin derivation, meaning coat of gold, and refers to the golden microtomentum that covers much of the head of this species.
#### Remarks.
This species is very similar and closely related to *Lamproclasiopa lapaz* and to a lesser degree *Lamproclasiopa polita* but is distinguished from these two species as follows: Female frons mostly bare, shiny black, lacking a broad, transverse stripe as in *Lamproclasiopa lapaz*; male mesonotum with a broad longitudinal band over entire length, although it is weaker posteriorly, not on anterior third only. Structures of the male terminalia are also diagnostic.
### Lamproclasiopa lapaz sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/6F5AC7A7-4A10-4FC3-BC73-7C97C2C5621F
[Figs 31--35](#F12){ref-type="fig"} [, 36](#F13){ref-type="fig"}
#### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Moderately small shore flies, body length 2.40--2.97 mm; generally a shiny black species. *Head*: Frontal and facial microtomentum sexually dimorphic; male with dense and extensive microtomentum on the frons, also within ocellar triangle, anterior laterals of frons bare, shiny black, Female frons with broad, transverse stripe of microtomentum on the center; male mesonotum with microtomentum on anterior third. Antenna black, with dense microtomentum especially evident on basal flagellomere laterally; arista bearing 3--4 dorsal rays (usually 4). Face with moderately deep antennal grooves on dorsal half, shallowly angulate in lateral view, vortex of angle at midheight near dorsal facial seta, ventral half of face receded, male facial microtomentum generally dense, golden brown dorsally, becoming more silvery ventrally, female face most bare, with silvery microtometum at the ventral portion of face and at base of facial setae and adjacent to parafacial; parafacial and anterior half of gena densely microtomentose in male, in female with thin area microtomentose at anterior and ventral margins of eye, otherwise bare, shiny; gena very high, gena-to-eye ratio 0.42--0.54; posterior margin of gena at merger with lateral margin of postgenal sharply angulate. *Thorax*: Mesonotum shiny black, male mesonotum with microtomentum on anterior third, with a thin lateral extension at level of suture and extended along posterior margin of notopleuron; lateral to microtomentose band mostly bare, shiny except for microtomentose anterior surface of postpronotum and ventral margin of notopleuron; presutural supra-alar seta well developed; pleural region generally bare, shiny black. Wing hyaline to faintly infumate, faintly tannish, lacking any pattern or markings. Costal vein ratio 0.40-0.51; M vein ratio 0.69--0.83. Coxae black, shiny; forecoxa with some lateral areas microtomentose but lacking vertical microstriae; femora and tibiae black; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; basal 2--3 tarsomeres yellow, apical 2--3 brownish black to dark brown. Halter with base black, knob whitish yellow. *Abdomen*: Generally shiny black; male tergite 5 truncate apically. Male Terminalia (Figs [31--35](#F12){ref-type="fig"}): Epandrium in posterior view (Fig. [31](#F12){ref-type="fig"}) with dorsal half transversely rectangular, lateral margin shallowly convex, dorsal margin broadly truncate, very thin above cercal cavity, ventral half thinner than dorsal half, demarcation sharply angulate, thereafter ventral extensions almost parallel sided, ventral margin broadly bilobed with moderately deep, thin, incision, setulae clumped, at ventral margin, at beginning of ventral half and at 2 sites along dorsal margin, in lateral view (Fig. [32](#F12){ref-type="fig"}) with posterior portion linear, thinnest dorsally and subventrally, thereafter ventrally enlarged, clavate, with well-developed anterior, hook-like extension, hook angulate rather than rounded, bearing setulae at vortices of angles; cerci in posterior view (Fig. [32](#F12){ref-type="fig"}) narrow, elongate, rod-like, slightly wider dorsally, apparently fused ventrally with ventral margin of cercal cavity, in lateral view (Fig. [33](#F12){ref-type="fig"}) narrowly lunate; aedeagus in lateral view (Fig. [35](#F12){ref-type="fig"}) as 2 structures of differing lengths, elongate basiphallus curved, irregularly tapered, base of basiphallus T-shaped, bar formed by pointed, lateral projections, apex of basiphallus acutely pointed and more curved, distiphallus less than half length of basiphallus, shallowly curved, otherwise rod-like, in ventral view (Fig. [34](#F12){ref-type="fig"}) with basiphallus as a thick, inverted Y, incised gap narrowly and deeply U-shaped with a heart-shaped extension at base, distiphallus with base within apical gap of basiphallus, narrow, straight, rod-like; phallapodeme in lateral view (Fig. [35](#F12){ref-type="fig"}) C-shaped, each arm expanded apically, in ventral view as 2 stacked, moderately broad, short arrowheads; gonite in lateral view irregularly clavate, narrow, elongate, straight, rod-like, in ventral view (Fig. [34](#F12){ref-type="fig"}) shorter than gonite in lateral view, rod-like; hypandrium in lateral view (Fig. [35](#F12){ref-type="fig"}) thin, elongate, irregularly clavate basally with midlength, short projections, in ventral view (Fig. [34](#F12){ref-type="fig"}) as 2 irregular, almost parallel, rectangular sclerites, wider anteriorly than posteriorly, with a W-shaped base with narrow arms extended posteriorly and slightly laterally, and 2 short medial bumps along base.
![*Lamproclasiopa lapaz* sp. n. (Bolivia. La Paz: La Paz) **31** epandrium and cerci, posterior view **32** same, lateral view **33** hypandrium **34** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **35** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g012){#F12}
#### Type material.
The holotype male of *Lamproclasiopa lapaz* is labeled "**BOLIVIA. La Paz**: La Paz (6 km NE; 16°25.7\'S, 68°04.3\'W; 4130m), 19 Mar 2001\[,\] Wayne N. Mathis/USNM ENT 00119994 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa lapaz* Costa, Mathis & Marinoni USNM \[red\]." The holotype is double mounted (minuten pin in a block of plastic), is in good condition (abdomen removed, dissected, parts in an attached microvial), and is deposited in the USNM. Two female paratypes bear the following label data: Bolivia. La Paz (NE; 16°27.4\'S, 68°06\'W; 3940m), 19 Mar 2001, W. N. Mathis (2♀; USNM).
#### Type locality.
BOLIVIA. La Paz: La Paz (6 km NE; 16°25.7\'S, 68°04.3\'W; 4130m).
#### Distribution
(Fig. [36](#F13){ref-type="fig"}). *Neotropical*: Bolivia (La Paz).
{#F13}
#### Etymology.
The species epithet, *lapaz*, refers to the capital of Bolivia, La Paz, where the type series was collected. La Paz is Spanish for peace, which we embrace and recommend to all.
#### Remarks.
This species is very similar both of the other species of the *polita* group, especially *Lamproclasiopa auritunica*, but is distinguished from these two species as follows: Female frons with broad, transverse stripe on anterior half (female frons in *Lamproclasiopa auritunica* is mostly bare, shiny black); male mesonotum with microtomentum on anterior third only (male mesonotum in *Lamproclasiopa auritunica* has a broad longitudinal band over entire mesonotal length, although it is weaker posteriorly). Shapes of structures of the male terminalia are also diagnostic.
### Lamproclasiopa polita
Animalia
Diptera
Ephydridae
(Edwards)
[Figs 37--38](#F14){ref-type="fig"} [, 39--42](#F15){ref-type="fig"} [, 104](#F36){ref-type="fig"}
1. Ditrichophora polita[@B13]: 117.
2. Discocerina (Basila) polita. [@B12]: 149 \[generic combination\]. [@B34]: 7 \[Neotropical catalog\]. [@B16]: 24 \[list, Argentina\]. [@B17]: 13 \[Argentina catalog\]. [@B24]: 165 \[world catalog\].
3. Lamproclasiopa polita. [@B36]: 39 \[generic combination\].
#### Diagnosis.
This species is distinguished from other congeners by the following combination of characters: Small to moderately small shore flies, body length 1.60--2.70 mm; generally a shiny black species. *Head*: Frons shiny black. Antenna black except for basoventral yellowish orange to orange of basal flagellomere. Face black except for silvery gray, microtomentose antennal grooves, microtomentum sometimes extended ventrally onto ventral portion of face, in lateral view rounded, obtusely angulate, greatest extension at midheight. Antenna black. Gena high; gena-to-eye ratio 0.30--0.34. *Thorax*: Mesonotum and pleural areas shiny black; presutural supra-alar seta well developed. Wing hyaline, immaculate; costal vein ration 0.43--0.45; M vein ratio 0.52--0.57. Legs black, mostly shiny except for yellow basal 2 tarsomeres; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; *Abdomen*: Tergites shiny black, almost completely bare of microtomentum; male tergite 5 more or less triangular, posterior margin narrowly rounded. Male terminalia (Figs [39--42](#F15){ref-type="fig"}): Epandrium in posterior view (Fig. [39](#F15){ref-type="fig"}) more or less oval, flattened dorsally, narrowed ventrally, setulae more evident ventrally, in lateral view (Fig. [40](#F15){ref-type="fig"}) longer than wide, ventral half robust, widest just ventrad of midheight, narrowly rounded at apex; cerci in posterior view (Fig. [39](#F15){ref-type="fig"}) narrow, elongate, slightly curved, ventral apex narrowly pointed, in lateral view (Fig. [40](#F15){ref-type="fig"}) as an elongated teardrop, shallowly curved, wider dorsally, ventral portion becoming narrower ventrally; gonite in lateral view (Fig. [42](#F15){ref-type="fig"}) elongate, posterior margin more or less evenly developed, anterior margin with angular protuberance, ventral apex shallowly bifurcate, in ventral view (Fig. [41](#F15){ref-type="fig"}) robustly developed medially, apices thin, angulate laterally; aedeagus in lateral view (Fig. [42](#F15){ref-type="fig"}) clavate, base narrower than globular apex, rounded apically, in ventral view (Fig. [41](#F15){ref-type="fig"}) with base quadrate with short triangular extension; phallapodeme in lateral view (Fig. [42](#F15){ref-type="fig"}) robustly L-shaped, in ventral view (Fig. [41](#F15){ref-type="fig"}) dome-like; hypandrium in lateral view L-shaped, anterior portion longer and more robustly developed than narrow, posterior portion, in ventral view (Fig. [41](#F15){ref-type="fig"}), slightly more than semicircular, broadly and evenly rounded.
{#F14}
![*Lamproclasiopa polita* (Edwards). (Chile. Osorno: Anticura) **39** epandrium and cerci, posterior view **40** same, lateral view **41** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **42** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g015){#F15}
#### Type material.
The holotype female of *Ditrichophora polita* Edwards is labeled "Holotype/Type/Argentina: Terr. Río Negro. F.&M. Edwards. B.M. 1927--63./Lake Gutiérrez 3--14.xi.1926./*Ditrichophora polita* Edw. F. W. Edwards det. 1932/HOLOTYPE *Ditrichophora polita* Edwards det. J.E. Chainey, 1995/NHMUK010240990. The holotype is double mounted (glued to a plastic triangle), is in good condition, and is deposited in the BMNH.
#### Type locality.
Argentina. Río Negro: Lake Gutiérrez (41°11.5\'S, 71°23.7\'W).
#### Other specimens examined.
CHILE. **Atacama**: Huasco (28°28\'S, 71°13.1\'W), 21 Oct 1957, L. E. Peña (1♂, 1♀; USNM). **Cautin**: Temuco (20 km E; 38°44\'S, 72°35\'W), 7 Jan 1951, A. E. Michelbacher, E. S. Ross (1♀; USNM). **Coquimbo**: Incahuasi (27°02\'S, 68°18\'W), 30 Sep 1952, P. G. Kuschel (13♂, 7♀; USNM); Ovalle (32 km SE; 30°36\'S, 71°11\'W), 12 Dec 1950, A. E. Michelbacher, E. S. Ross (3♂, 2♀; USNM). **Lanquihue**: Peulla (41°28\'S, 72°57.7\'W) (1?; BMNH); Puerto Varas (41°18.6\'S, 72°59.6\'W) (1♂; BMNH). **O'Higgins**: Río Claro (5 km N Rengo; 34°24\'S, 70°52\'W; 300 m), 23 Jan 1978, W. N. Mathis (1♀; USNM). **Osorno**: Anticura (4 km W; 37°40\'S, 72°01\'W; 400 m), 3 Feb 1978, W. N. Mathis (3♂, 4♀; USNM); Anticura (1 km W; 40°39\'S, 72°10\'W; 430 m), 5--12 Feb 1978, W. N. Mathis (1♂; USNM); Lago Puyehue (SE shore; 40°45\'S, 72°25.2\'W), 6--10 Feb 1978, W. N. Mathis (4♂, 1♀; USNM); Lago Rupanco, El Encanto (40°49\'S, 72°28\'W), 6 Feb 1978, W. N. Mathis (1♂, 1♀; USNM); Laguna El Pato (41°10\'S, 73°40\'W; 1100 m), 13 Feb 1978, W. N. Mathis (2♂, 2♀; USNM); Termas de Aguas Calientes (1 km SE; 40°41\'S, 72°21\'W; 530 m), 7--8 Feb 1978, W. N. Mathis (2♂, 2♀; USNM). **Santiago**: El Alfalfal (33°30\'S, 70°11\'W; 1320 m), 22 Jan 1978, W. N. Mathis (1♂, 2♀; USNM); Quebrada de la Plata (near Maipú; 33°30\'S, 70°55\'W; 550 m; Malaise trap), 12 Mar 1986, M. E. Irwin (1♂; USNM). **Talca**: Río Lircay (11 km N Talca; 35°23\'S, 71°39\'W; 85 m), 23 Jan 1978, W. N. Mathis (2♂, 7♀; USNM).
#### Distribution
(Fig. [104](#F36){ref-type="fig"}). Neotropical: Argentina (Río Negro), Chile (Atacama, Cautin, Coquimbo, Lanquihue, Malleco, O'Higgins, Osorno, Santiago, Talca).
#### Remarks.
Although similar to *Lamproclasiopa auritunica* and *Lamproclasiopa lapaz* in having a high gena (gena-to-eye ratio 0.30--0.34), this species is unlike these two species by having a nearly bare and shiny male frons; a triangular-shaped male tergite five, which is narrowly rounded posteriorly; a rounded epandrium, and an aedeagus without any kind of division.
The *ecuadoriensis* group (*Lamproclasiopa ecuadoriensis*, *Lamproclasiopa zerafael*) {#SECID0EE3CI}
-------------------------------------------------------------------------------------
**Diagnosis.** Body with extensive surfaces sparsely to densely microtomentose. *Head*: Frons and face generally unicolorous; gena relatively short (gena-to-eye ratio 0.05--0.12); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta variable, well developed in *Lamproclasiopa ecuadoriensis*, lacking in *Lamproclasiopa zerafael*; katepisternum and anepisternum thinly microtomentose, generally appearing dull, not shiny. Wing generally hyaline to very faintly infumate (*Lamproclasiopa mancha* with crossveins r-m and dm-cu with darkened cloud); vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin. *Abdomen*: Male terminalia: Keel of phallapodeme short and sometimes difficult to discern.
**Remarks.** This species group, like the *hendeli* group, is mostly based on homoplasious characters, and we cannot confirm its monophyly. The two included species are similar to each other and the species group can be diagnosed. These are the bases for recognition of this species group. Structures of the male terminalia of *Lamproclasiopa zerafael* are quite different from all congeners, especially the very robust aedeagus that is slightly asymmetrical, and the very wide and dissected hypandrium in ventral view.
### Lamproclasiopa ecuadoriensis sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/148D944E-D393-46B5-A9F5-1E9031EA655E
[Figs 43--44](#F16){ref-type="fig"} [, 45--48](#F17){ref-type="fig"} [, 104](#F36){ref-type="fig"}
#### Diagnosis.
This species is distinguished from other congeners by the following combination of characters: Small shore-fly species, body length 1.55--1.80 mm; generally black, subshiny to shiny. *Head*: Frons mostly brownish black to black, sparsely brownish microtomentose, more so on anterior portion, subshiny, ocellar triangle extended to anterior margin of frons, some specimens with grayish red areas along anterior margin just dorsad of antennal bases, parafrons with narrowly oval, densely microtomentose areas at anterolateral corner. Antenna mostly black, only basal flagellomere with ventrobasal area with some yellowish to yellowish orange coloration. Face narrow, mostly shiny black, especially over greater medial portion, extreme lateral margin adjacent to parafacial yellowish, narrow whitish gray, transverse band just ventrad of antennal base and through dorsal portion of antennal grooves, ventral half of face slightly receded; bearing 2 larger facial setae, dorsal seta at about midfacial height, dorsomesoclinate; ventral seta just dorsad of epistomal margin, slightly dorsoclinate; parafacial thin, yellow dorsally, adjacent to eye, black ventrally and extended to gena. Gena short, gena-to-eye ratio 0.04--0.06. *Thorax*: Mesonotum uniformly sparsely microtomentose, brownish black to black, subshiny; presutural supra-alar seta well developed; pleural region black; dorsal 2/3 of anepisternum finely granulose, subshiny, anteroventral portion smooth, shiny. Wing hyaline, lacking any pattern or markings; costal vein ratio 0.80--0.81; M vein ratio 0.61--0.64. Legs, except tarsi, black; tarsi yellow; apical 1--2 tarsomeres darker, tan to brown; forefemur with sparse row of 4--5, stouter, spine-like setae along apical half of posteroventral surface. *Abdomen*: Generally black, subshiny to mostly shiny, dorsum of tergites very sparsely and finely microtomentose, faintly whitish gray; sternite 3 of male rectangular, parallel sided, length twice width; sternite 4 of male rectangular, length almost twice width; sternite 5 of male as 2 sternites, length nearly twice greatest width, anterior margin narrow, becoming slightly wider on anterior 1/3, thereafter tapered to a posterolateral point, lateral margin straight, medial margin angulate. Male terminalia (Figs [45--48](#F17){ref-type="fig"}): Epandrium in posterior view (Fig. [45](#F17){ref-type="fig"}) robustly oval, arched and thin dorsally, gradually becoming wider ventrally than narrowed on apical 1/3, in lateral view (Fig. [46](#F17){ref-type="fig"}) with dorsal 2/3 thirds narrow, strap-like, thereafter ventrally abruptly widened with anterior, pointed extension, ventral margin broadly rounded; cerci in posterior view (Fig. [45](#F17){ref-type="fig"}) elongate, narrowly semicircular, ventral apex more acutely pointed than more widely produced dorsal margin, in lateral view (Fig. [46](#F17){ref-type="fig"}) irregularly, narrowly semihemispherical, wider subdorsally than ventrally; gonite in lateral view narrowly rod-like, arched, only slightly wider toward aedeagal base than toward hypandrium, in ventral view (Fig. [47](#F17){ref-type="fig"}) robustly hook-like with shank of hook narrow and rounded portion very robustly developed; phallapodeme in lateral view (Fig. [48](#F17){ref-type="fig"}) L-shaped, arm extended to aedeagal base slightly more robust, length of both arms about equal, in ventral view (Fig. [47](#F17){ref-type="fig"}) as a dog bone, expanded at each apex; hypandrium in lateral view (Fig. [48](#F17){ref-type="fig"}) elongate, robust, sinuous, more or less parallel sided, in ventral view (Fig. [47](#F17){ref-type="fig"}) as a very robust H with long posterior arms, lateral margins conspicuously sinuous, anterior emargination shallowly concave, posterior emargination deep, broadly U-shaped.
{#F16}
![*Lamproclasiopa ecuadoriensis* sp. n. (Ecuador. Orellana: Rio Tiputini) **45** epandrium and cerci, posterior view **46** same, lateral view **47** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **48** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g017){#F17}
#### Type material.
The holotype male of *Lamproclasiopa ecuadoriensis* is labeled "**ECUADOR.** Prt. Or\[e\]\[l\]lana: RioTiputini (0°38.2\'S, 76°8.9\'W), 12--26 Aug 1999,W.N.Mathis, A. Baptista, M. Kotrba/USNM ENT 00118307 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa ecuadoriensis* Costa, Mathis & Marinoni USNM \[red\]." The holotype is double mounted (minuten pin in a plastic block), is in excellent condition, and is deposited in the USNM. Four paratypes (3♂, 1♀; DZUP, USNM) bear the same label data as the holotype.
#### Type locality.
Ecuador. Orellana: Río Tiputini Biodiversity Station (0°38.2\'S, 76°8.9\'W).
#### Distribution
(Fig. [104](#F36){ref-type="fig"}). *Neotropical*: Ecuador (Orellana).
#### Etymology.
The species epithet, *ecuadoriensis*, refers to the country of Ecuador, where this species was collected.
#### Remarks.
This species is similar to *Lamproclasiopa laevior* and *Lamproclasiopa polita*, although it can be distinguished from congeners by the densely microtomentose anterolateral, narrowly oval black velvet spots on the frons; the narrow, shiny black face; the comparatively elongate costal section III (section III slightly less than section II); and the shape of structures of the male terminalia (Figs [45--48](#F17){ref-type="fig"}).
### Lamproclasiopa zerafael sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/3BA512CE-7CF2-4236-868D-0833AC936736
[Figs 49--50](#F18){ref-type="fig"} [, 51--54](#F19){ref-type="fig"} [, 104](#F36){ref-type="fig"}
#### Diagnosis.
This species is distinguished from other congeners by the following combination of characters: Small to moderately small shore-fly species, body length 1.50--2.05 mm; generally black, subshiny to shiny. *Head*: Frons mostly brownish black to black, sparsely brownish microtomentose, more so on anterior portion, subshiny, some specimens with 2 gray spots along ventral margin just dorsad of antennal bases. Antenna mostly black, only basal flagellomere with ventrobasal area with some yellowish to yellowish orange coloration. Face mostly shiny black, especially medially and laterally, between with some areas sparsely microtomentose and in antennal grooves, dorsal half; antennal grooves evident , dorsad of dorsoclinate facial pair of setae; ventral half of face slightly receded; bearing 2 larger facial setae, dorsal seta at about midfacial height, dorsomesoclinate; ventral seta just dorsad of epistomal margin, slightly dorsoclinate; parafacial thin, black. Gena relatively short, gena-to-eye ratio 0.06--0.07. *Thorax*: Mesonotum uniformly sparsely microtomentose, brownish black to black; presutural supra-alar seta lacking or indistinguishable from surrounding setae; pleural region black; dorsal 2/3 of anepisternum finely granulose, subshiny, anteroventral portion smooth, shiny. Wing hyaline, lacking any pattern or markings; costal vein ratio 0.75--0.89; M vein ratio 0.58--0.61. Legs, except tarsi, black; tarsi yellow; apical 1--2 tarsomeres darker, tan to brown; forefemur with sparse row of 4--5, stouter, spine-like setae along apical half of posteroventral surface. *Abdomen*: Generally black, subshiny to mostly shiny, dorsum of tergites very sparsely and finely microtomentose, faintly whitish gray. Male terminalia (Figs [51--54](#F19){ref-type="fig"}): Epandrium in posterior view (Fig. [51](#F19){ref-type="fig"}) almost as wide as high, as an inverted U, dorsal arch very thin, vertical, lateral arms essentially parallel sided, wider than width of cercus, in lateral view (Fig. [52](#F19){ref-type="fig"}) widest at ventral 1/3, ventral margin step-wise rounded, overall as a robust tear drop with an anterior, short, shallowly pointed projection at widest width; cercus in posterior view (Fig. [51](#F19){ref-type="fig"}) bar-like, elongate, narrow, parallel sided, not fused with ventral margin of cercal cavity, in lateral view (Fig. [52](#F19){ref-type="fig"}) elongate, dorsal half slightly wider than ventral half; gonite in ventral view (Fig. [53](#F19){ref-type="fig"}) triangular, with basal angle projected into narrow process, in lateral view (Fig. [54](#F19){ref-type="fig"}) obtusely angulate, extension toward aedeagal base slightly thinner; aedeagus in lateral view (Fig. [54](#F19){ref-type="fig"}) irregular, wider apically, margin irregular, in ventral view (Fig. [53](#F19){ref-type="fig"}) slightly wider subapically, irregularly rounded apically; phallapodeme in lateral view (Fig. [54](#F19){ref-type="fig"}) angulate, L-shaped, extension toward hypandrium shallowly angulate subapically, in ventral view (Fig. [53](#F19){ref-type="fig"}) spindle shaped, with basal portion much wider than apical portion; hypandrium in ventral view (Fig. [53](#F19){ref-type="fig"}) as a very wide and short structure, anterior and posterior emarginations shallow, anterior arms with oblique crossbar, in lateral view (Fig. [54](#F19){ref-type="fig"}) as an irregular H, with lateral portions of H irregular.
{#F18}
![*Lamproclasiopa zerafael* sp. n. (Brazil. Amazonas: Manaus) **51** epandrium and cerci, posterior view **52** same, lateral view **53** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **54** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g019){#F19}
#### Type material.
The holotype male of *Lamproclasiopa zerafael* is labeled "**BRAZIL.** Amazonas: Reserva Ducke (02°55.8\'S, 59°58.5\'W; 40 m), 5 May 2010, D. & W. N. Mathis/USNM ENT 00118311 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa zerafael* Costa, Mathis & Marinoni, INPA \[red\]." The holotype is double mounted (minuten pin in a plastic block), is in excellent condition, and is deposited in INPA. Fourteen paratypes (11♂, 2♀; DZUP, INPA, USNM) bear the same label data as the holotype. Other paratypes are as follows: BRAZIL. **Amazonas**: Manaus, Universidade Federal do Amazonas (03°05.9\'S, 59°58.2\'W; 50 m), 7 May 2010, D. and W. N. Mathis (5♂, 11♀; DZUP, INPA, USNM); Reserva Cuieiras (02°35.2\'S, 60°07.2\'W; 110 m), 8 May 2010, D. and W. N. Mathis (3♂; INPA, USNM).
#### Type locality.
Brazil. Amazonas: Reserva Ducke (02°55.8\'S, 59°58.5\'W; 40 m).
#### Distribution
(Fig. [104](#F36){ref-type="fig"}). *Neotropical*: Brazil (Amazonas).
#### Etymology.
The species epithet, *zerafael*, refers to José (Zé) Albertino Rafael, student of Diptera and Zoraptera (especially the Amazonian fauna) and who kindly hosted and guided us while in Manaus, Amazonas. The name is a noun in apposition.
#### Remarks.
This species is distinguished from congeners, especially *Lamproclasiopa triangularis*, by having a sparsely microtomentose body generally, a mostly shiny black face; a short gena (height about half height of basal flagellomere), a hyaline wing, a blackish yellow foretarsus. The shape of structures of the male terminalia also distinguishes this species, especially the relatively gross, thickened aedeagus that is slightly asymmetrical, the wide and thinly dissected hypandrium, and the funnel-shaped gonites in ventral view.
The *painteri* group (*Lamproclasiopa balsamae*, *Lamproclasiopa mancha*, *Lamproclasiopa painteri*) {#SECID0EZUDI}
----------------------------------------------------------------------------------------------------
**Diagnosis.** Body with extensive surfaces sparsely to densely microtomentose. *Head*: Frons and face generally unicolorous; gena relatively short (gena-to-eye ratio between 0.05--0.10); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta well developed; katepisternum and anepisternum thinly microtomentose, generally appearing dull, not shiny. Wing with numerous dark spots (*Lamproclasiopa balsamae*, *Lamproclasiopa painteri*) or with darkened clouds over crossveins r-m and especially over dm-cu (*Lamproclasiopa mancha*); vein R~2+3~ either angulate subapically and bearing a stump vein with a posteroapical orientation, a second stump vein near middle (*Lamproclasiopa balsamae*, *Lamproclasiopa painteri*) or vein R~2+3~ with apex more abruptly curved toward costa (*Lamproclasiopa mancha*). Forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin.
**Remarks.** This species group comprises species with some pattern in the wing and is thus distinctive from all others, which have mostly hyaline or very faintly infumate wings. The pattern, however, differs. For *Lamproclasiopa balsamae* and *Lamproclasiopa painteri* the pattern comprises numerous distinctive brown spots, and vein R~2+3~ is distinctly angulate subapically with the apices abruptly angled subapically toward costa. At the vertex of the abrupt, subapical angle there is also a stump vein and often another stump vein near middle of this vein. Certainly these two species form a monophyletic lineage. The inclusion of *Lamproclasiopa mancha* in this species group may be artificial, as the pattern in the wing is quite different (see species description of *Lamproclasiopa mancha*).
### Lamproclasiopa balsamae
Animalia
Diptera
Ephydridae
(Cresson)
[Figs 55--58](#F20){ref-type="fig"} [, 59](#F21){ref-type="fig"}
1. Ditrichophora balsamae[@B5]: 77.
2. Discocerina (Basila) balsamae. [@B12]: 149 \[generic combination, review\]. [@B34]: 7 \[Neotropical catalog\]. [@B24]: 165 \[world catalog\].
3. Lamproclasiopa balsamae. [@B36]: 39 \[generic combination\].
#### Diagnosis.
This species is easily distinguished from congeners by the following combination of characters: Small shore flies, body length 1.65--1.85 mm. *Head*: Frons bi- or tricolored, lacking iridescent microtomentum, ocellar triangle largely and fronto-orbits whitish tan to tan, ocellar triangle with anteromedial, narrow, slightly oval darkened area, triangle broadly extended to anterior margin, parafrons grayish charcoal. Antenna largely yellow, only dorsum of basal flagellomere slightly darkened. Face narrowed at midheight, mostly unicolorous, whitish gray to blackish gray except for mediovertical brown vitta; parafacial creamy white. Gena relatively short, gena-to-eye ratio 0.10. *Thorax*: Mesonotum with 7 brown vittae, including a medial vitta along acrostichal area (Fig. [57](#F20){ref-type="fig"}); presutural supra-alar seta well developed. Wing conspicuously patterned with distinct brown spots (Fig. [58](#F20){ref-type="fig"}); vein R~2+3~ distinctly angulate subapically, apices angled toward costa; at vertex of angle also bearing a stump vein, another stump vein near middle; costal vein ratio 0.67--0.68; M vein ratio 0.66--0.71. Femora brownish black; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; tibiae largely brownish black, apices yellow; tarsi yellow.
{#F20}
#### Type material.
The holotype female of *Ditrichophora balsamae* Cresson is labeled "Puerto Castilla B. F. Hond. 6-V-26. R. H. Painter, Co \["B. F." handwritten /TYPE No. 6365 Ditrichophora BALSAMAE E T Cresson, Jr. \[red; "6365 Ditrichophora BALSAMAE" handwritten\]/1182." The holotype is double mounted (minuten pin in a block of fine foam), is in excellent condition, and is deposited in the ANSP (6365).
#### Type locality.
Honduras. Colón: Puerto Castilla (16°0.5\'N, 85°57.7\'W).
#### Other specimens examined.
BRAZIL. **Rio de Janeiro**: Gavea (22°58.6\'S, 43°13.7\'W), Mar 1929, H. Souza Lopes (6♀; IOC).
COSTA RICA. **San José**: Pedregoso (9° 22.45\'N, 83° 43.2\'W), D. L. Rounds (1♀; USNM).
EL SALVADOR. **La Liberdad**: Santa Tecla (12 km NW; 13°45.1\'N, 89°22.1\'W), Oct 1953, W. B. Heed (1♀; USNM).
#### Distribution
(Fig. [59](#F21){ref-type="fig"}). *Neotropical*: Brazil (Rio de Janeiro), Costa Rica (San José), El Salvador (La Liberdad), Honduras (Colón).
{#F21}
#### Remarks.
Although very similar and apparently closely related to *Lamproclasiopa painteri*, this species is distinguished from congeners by the number and entirety of the mesonotal stripes. There are seven longitudinal vittae, including a medial vitta in the acrostichal area.
A male of this species is unavailable, and thus, our diagnosis is incomplete for structures of the male terminalia.
### Lamproclasiopa mancha sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/A9F90349-9066-4D6B-8D7B-7AF6C13B5DEF
[Figs 59](#F21){ref-type="fig"} [, 60--62](#F22){ref-type="fig"} [, 63--66](#F23){ref-type="fig"}
#### Diagnosis.
This species is distinguished from other congeners by the following combination of characters: Small to moderately small shore-flies, body length 1.55--2.10 mm. *Head*: Frons mostly yellowish to golden tan, especially outline of mesofrons and fronto-orbits, anterior portion immediately dorsad of antennae yellowish orange, microtomentum denser and whiter at base of fronto-orbital setae. Antenna yellow to yellowish orange; basal flagellomere slightly brownish dorsally. Face black but completely to mostly silvery white microtomentose, often with medial, darker stripe with microtomentum thinner, otherwise lacking prominent, vertical stripes; bearing 2 larger facial setae, dorsal seta at about midfacial height, dorsomesoclinate; ventral seta just dorsad of epistomal margin, slightly dorsoclinate; parafacial silvery to creamy white; gena very short. Gena-to-eye ratio 0.05--0.07. *Thorax*: Mesonotum uniformly tannish to golden tan microtomentose; presutural supra-alar seta well developed; pleural area very sparsely microtomentose, mostly dark brown, partially subshiny, contrasted with densely microtomentose mesonotum. Wing hyaline except for conspicuous darkened clouds over crossveins r-m and especially over dm-cu (Fig. [62](#F22){ref-type="fig"}); vein R~2+3~ with apex more abruptly curved toward costa; costal vein ratio 0.51--0.60; M vein ratio 0.59--0.65. Femora mostly black; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; tibiae mostly brownish black, apices yellow; tarsi yellow. *Abdomen*: Generally black, mostly subshiny to shiny, dorsum of tergites very sparsely and finely microtomentose, faintly whitish gray. Male terminalia (Figs [63--66](#F23){ref-type="fig"}): Epandrium in posterior view (Fig. [63](#F23){ref-type="fig"}) elongate, inverted U-shaped, dorsal arch narrow, becoming wider ventrally, ventral margin bearing loose cluster of longer setulae, in lateral view widest subventrally, ventral margin more narrowly rounded; cercus in posterior hemispherical, not fused with ventral margin of cercal cavity, with somewhat evenly scattered setulae, those toward ventral margin longer; gonite in lateral view (Fig. [66](#F23){ref-type="fig"}) somewhat rod-like, shallowly curved, both ends tapered, in ventral view (Fig. [65](#F23){ref-type="fig"}) knife-like, medial end blade-like, lateral extension narrow with apical portion curved and pointed; aedeagus in lateral view (Fig. [66](#F23){ref-type="fig"}) as an angulate funnel, comparatively wide basally, thereafter apically angles posteroventrally, tapered to narrowly pointed apex, in ventral view (Fig. [65](#F23){ref-type="fig"}) an elongate, narrow funnel; phallapodeme in lateral view (Fig. [66](#F23){ref-type="fig"}) irregularly Y-shaped, keel long, narrow, linear; hypandrium in ventral view (Fig. [65](#F23){ref-type="fig"}) as 2 narrow, parallel sided, thin sclerites, with posterior connection, in lateral view (Fig. [66](#F23){ref-type="fig"}) an elongate, slender, rod-like structure, bulbous posteriorly, shallowly bifurcate anteriorly.
{#F22}
![*Lamproclasiopa mancha* sp. n. (Brazil. Paraná: Curitiba) **63** epandrium and cerci, posterior view **64** same, lateral view **65** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **66** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g023){#F23}
#### Type material.
The holotype male of *Lamproclasiopa mancha* is labeled "**BRAZIL.** Paraná: Curitiba, UFPR \[Universidade Federal do Paraná, Reserva Biológica\] (25°26.9\'S, 49°14\'W; 915 m),6 Feb2010\[,\] D. & W. N. Mathis/USNM ENT 00118310 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa mancha* Costa, Mathis & Marinoni DZUP \[red\]." The holotype is double mounted (minuten pin in a block of plastic), is in excellent condition, and is deposited in DZUP. Twenty-nine paratypes (24♂, 5♀; DZUP, USNM) bear the same locality data as the holotype, with dates from 9 Dec 2009--13 Feb 2010.
#### Type locality.
Brazil. Paraná: Curitiba, Universidade Federal do Paraná, Reserva Biológica (25°26.9\'S, 49°14\'W; 915 m).
#### Other specimens examined.
ARGENTINA. **Misiones**: Puerto Bemberg (25°55\'S, 54°37\'W), 13 Oct 1950, L. C. Shum (1♀; USNM).
BRAZIL. **Paraná**: Colombo (Santa Monica tennis club; 25°23.1\'S, 49°08.8\'W; 860 m), 18 Mar 2015, Daniel N. R. Costa (1♂; DZUP); Curitiba, Universidade Federal do Paraná, Reserva Biológica (25°26.9\'S, 49°14\'W; 915 m), 1--4 Feb 2010, 13--17 Out 2014, Daniel N. R. Costa (13♂, 4♀; DZUP); Morro do Araçatuba (Município de Tijucas do Sul; 25°53.8\'S, 49°01.2\'W; 910 m), 27 Feb 2015, W. N. Mathis (1♀; DZUP). **Santa Catarina**: Nova Teutônia (27°11\'S, 52°23\'W; 3--500 m), Jul-Nov 1970, 1971, F. Plaumann (9♂, 4♀; MZUSP). **São Paulo**: Itú (23°15.9\'S, 47°17.9\'W), Sep 1960, M. A. V. D'Andretta (8♂, 2♀; MZUSP).
#### Distribution
(Fig. [59](#F21){ref-type="fig"}). Neotropical: Argentina (Misiones), Brazil (Paraná, Santa Catarina, São Paulo).
#### Etymology.
The species epithet, *mancha*, is the Portuguese word for a stain and refers to the darkened clouds over crossveins r-m and dm-cu, diagnostic of this species.
#### Remarks.
This species is easily distinguished from congeners by the wing pattern. The wing is mostly hyaline except for darkened clouds over crossveins r-m and especially over dm-cu; and vein R~2+3~ has the apex more abruptly curved toward the costa. Sometimes the darkened spots over the crossveins are slightly faded.
### Lamproclasiopa painteri
Animalia
Diptera
Ephydridae
(Cresson)
[Figs 59](#F21){ref-type="fig"} [, 67--70](#F24){ref-type="fig"} [, 71--74](#F25){ref-type="fig"}
1. Ditrichophora painteri[@B5]: 76.
2. Discocerina (Basila) painteri. [@B12]: 149 \[generic combination\]. [@B34]: 7 \[Neotropical catalog\]. [@B24]: 165 \[world catalog\].
3. Lamproclasiopa balsamae, of authors, not Cresson \[misidentification\]. [@B36]: 36 \[generic combination\].
4. Lamproclasiopa painteri. [@B36]: 39 \[generic combination\].
#### Diagnosis.
This species is easily distinguished from congeners by the following combination of characters: Small shore flies, body length 1.15--1.80 mm. *Head*: Frons bi- or tricolored, lacking iridescent microtomentum, ocellar triangle largely and fronto-orbits whitish tan to tan, ocellar triangle with anteromedial, narrow, slightly oval darkened area, triangle broadly extended to anterior margin, parafrons grayish charcoal. Antenna largely yellow, only dorsum of basal flagellomere slightly darkened. Face narrowed at midheight, mostly unicolorous, whitish gray to blackish gray except for mediovertical brown vitta; parafacial creamy white. Gena relatively short, gena-to-eye ratio 0.08--0.09. *Thorax*: Mesonotum with 4 elongate, mostly separated spots (Fig. [69](#F24){ref-type="fig"}); presutural supra-alar seta well developed. Wing conspicuously patterned with distinct brown spots (Fig. [70](#F24){ref-type="fig"}); vein R~2+3~ distinctly angulate subapically, apices angled toward costa; at vertex of angle also bearing a stump vein, another stump vein near middle; costal vein ratio 0.76--0.87; M vein ratio 0.66--0.75. Femora brownish black; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; tibiae largely brownish black, apices yellow; tarsi yellow. *Abdomen*: Tergite 5 of male truncate posteriorly. Male terminalia (Figs [71--74](#F25){ref-type="fig"}): Epandrium in posterior (Fig. [71](#F25){ref-type="fig"}) view roundly U-shaped, bluntly oval, narrower dorsally and ventrally, slightly wider at midheight, lateral arm becoming wider ventrally, curved medially ventral margin ventromedial gap V-shaped, ventral angle bearing loosely clustered setulae; cercus hemispherical, pointed dorsomedially, more setulose dorsally, medial margin straight; gonite in lateral view (Fig. [74](#F25){ref-type="fig"}) robustly rod-like, shallowly curved toward aedeagal base, shaped like a banana, in ventral view shallowly curved with extension toward aedeagal base narrow, thumb-like, thereafter moderately wide, widest subapically; aedeagus in lateral view (Fig. [74](#F25){ref-type="fig"}) comparatively narrowly truncate basally, thereafter expanded, widest sub-basally, thereafter tapered to rounded apex, apex with short, recurved anterior point, in ventral view (Fig. [73](#F25){ref-type="fig"}) as an elongate, shallowly rounded, narrowed medially, basal margin somewhat truncate with shallow, medial emargination, apical margin tapered to angulate, rounded apex; phallapodeme in lateral view (Fig. [74](#F25){ref-type="fig"}) Y-shaped with one arm of Y a short, irregularly narrow keel, keel irregularly tapered, pointed apically; hypandrium in lateral view (Fig. [74](#F25){ref-type="fig"}) narrowed basally, apical ½-2/3 wider, narrowly rectangular, rounded anteriorly, in ventral view as a very broad, short H with short arms, anterior emargination broadly V-shaped, posterior emargination very broadly and shallowly U-shaped.
{#F24}
![*Lamproclasiopa painteri* (Cresson). (Peru. Madre de Dios: Río Manu, Pakitza) **71** epandrium and cerci, posterior view **72** same, lateral view **73** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **74** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g025){#F25}
#### Type material.
The holotype female of *Ditrichophora painteri* Cresson is labeled "Puerto Castilla B. F. Hond. 6-V-26 \[6 May 1926\] R. H. Painter, Co \["B. F." handwritten\]/TYPE No. 6366 Ditrichophora PAINTERI E.T.Cresson, Jr, \[red; "6366 Ditrichophora PAINTERI" handwritten\]/1182." The holotype is double mounted (minuten pin in a block of fine foam), is in excellent condition, and is deposited in the ANSP (6366). Three paratypes (2♂, 1♀; ANSP) bear the same locality label as the holotype.
#### Type locality.
Honduras. Colón: Puerto Castilla (16°0.5\'N, 85°57.7\'W).
#### Other specimens examined.
BELIZE. **Stann Creek**: Cockscomb Basin Wildlife Sanctuary (16°47\'N, 88'30\'W), 5--6 Apr 1993, W. N. Mathis (9♂, 8♀; USNM); Maya Center: Cabbage Haul Creek (16°48\'N, 88°23\'W), 3 Apr 1993, W. N. Mathis (1♂, 1♀; USNM).
BRAZIL. **Amazonas**: Marco (near Leticia=Tabatinga; 04°13.9\'S, 69°56\'W), Aug 1960, H. L. Carson, W. B. Heed (1♂; USNM). **Pará**: Oriximiná-Obidos, estrada (01°47.7\'S, 55°36\'W), Nov 1969 (1♂; MZUSP); Tucuruí, Morro do Senador (03°59.4\'S, 49°44.8\'W), Dec 2001, J. A. Rafael, J. Vidal (1♀; INPA). **Paraná**: Antonina, Reserva Natural Rio Cachoeira (25°19\'S, 48°41,6\'W), 8 Feb 2010, D. N. R. Costa (3♂; DZUP). **Rio de Janeiro**: Gávea (22°59.2\'S, 43°14.7\'W), 31 Mar 1937, H. de Souza Lopes (5♂, 1♀; ANSP, MNRJ + 2♂; NMNH).
ECUADOR. **Orellana**: Río Tiputini (0°38.2\'S, 76°8.9\'W), 12--26 Aug 1999, W. N. Mathis, A. Baptista, M. Kotrba (5♀; USNM).
GUYANA. Moco-Moco (30 km E Lethem in Kanuku Mountains; 3°18.2\'N, 59°39.0\'W), 29 Apr 1995, W. N. Mathis (9♂, 3♀; USNM).
PANAMA. **Colon**: Juan Gallegos (9°37\'N, 79°34\'W), 10 Jul 1982, R. B. Kimsey (1♀; USNM).
PERU. **Madre de Dios**: Río Manu, Pakitza (11°56.6\'S, 71°16.9\'W; 250 m), 9--23 Sep 1988, W. N. Mathis (27♂, 20♀; USNM).
TRINIDAD and TOBAGO. Tobago. **St. John**: Charlotteville (5 km S; 11°18.9\'N, 60°34.5\'W), Hermitage River and beach, 22 Apr-11 Jun 1993, 1994, D. and W. N. Mathis (5♂, 5♀; USNM); Parlatuvier (creek; 11°17.9\'N, 60°35\'W), 20 Apr-14 Jun 1993, 1994, W. N. Mathis (2♂, 1♀; USNM). **St. Paul**: Argyle Falls (11°15\'N, 60°35\'W), 21 Apr 1994, W. N. Mathis (1♂; USNM); Roxborough (6 km NNW; 11°16\'N, 60°35.4\'W), 20 Apr 1994, W. N. Mathis (13♂, 14♀; USNM). Trinidad. **Caroni**: Talparo (2 km N, 10°31\'N, 61°17\'W), 22 Jun 1993, W. N. Mathis (1♂; USNM). **St. George**: Filette (1 km SE; 10°47\'N, 61°21\'W), Yarra River, 25 Jun 1993; W. N. Mathis (2♀; USNM). **Victoria**: Basse Terre (7 km E; 10°07\'N, 61°14\'W), 27 Jun 1993, W. N. Mathis (3♂, 3♀; USNM).
VENEZUELA. Caife, Jan 1943, P. J. Anduze (2♀; USNM).
#### Distribution
(Fig. [59](#F21){ref-type="fig"}). Neotropical: Belize (Stann Creek), Brazil (Amazonas, Pará, Paraná, Rio de Janeiro), Ecuador (Orellana), Guyana, Honduras (Colón), Peru (Madre de Dios), Trinidad and Tobago, Venezuela.
#### Remarks.
Although similar to *Lamproclasiopa balsamae*, this species is distinguished from that species and all other congeners by having a conspicuously spotted wing, as in *Lamproclasiopa balsamae*, and by having the mesonotum with four stripes, each as a short series of two to four more or less elongated spots. In *Lamproclasiopa balsamae* there are seven distinct and complete or nearly complete stripes. Structures of the male terminalia also distinguish this species.
The illustration of *Lamproclasiopa balsamae* that Mathis and Zatwarnicki (2001) published is actually that of *Lamproclasiopa painteri*.
The locality of the specimen from Venezuela is a mystery to us. We have checked and rechecked the spelling on the label, "Caife," but have been unable to locate this name on maps or gazetteers. Perhaps it is a misspelling.
The *nana* group (*Lamproclasiopa nana*) {#SECID0E4GBK}
----------------------------------------
**Diagnosis.** Body with extensive surfaces sparsely to densely microtomentose. *Head*: Frons and face distinctly two-toned; antenna yellow; gena relatively short (gena-to-eye ratio 0.06--0.10); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta well developed. Wing generally hyaline to faintly infumate; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; tibiae entirely black. *Abdomen*: Male terminalia: Dorsal epandrial margin interrupted, each lateral arm of the epandrium robustly developed; hypandrium in ventral view rectangular.
**Remarks.** This species group, comprising a single species, is very distinctive among all congeners and is easily recognized. Although distinctive, it is apparently related to the *furvitibia* group.
### Lamproclasiopa nana
Animalia
Diptera
Ephydridae
(Williston)
[Figs 75--76](#F26){ref-type="fig"} [, 77--80](#F27){ref-type="fig"} [, 81](#F28){ref-type="fig"}
1. Discocerina nana[@B31]: 396.
2. Ditrichophora nana. [@B2]: 159 \[generic combination\].
3. Discocerina (Basila) nana. [@B12]: 149 \[review\]. [@B34]: 7 \[Neotropical catalog\]. [@B21]: 824 \[review of Williston's St. Vincent species\]. [@B24]: 165 \[world catalog\].
4. Lamproclasiopa nana. [@B36]: 39 \[generic combination\]. [@B29]: 177 \[list, Hispaniola\].
#### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Small to moderately small shore-flies, body length 1.60-2.40 mm. *Head*: Frons distinctly two toned, fronto-orbits and narrow, medial triangular area densely microtomentose, velvety black; mesofrons other than narrow, medial triangle, seriaceus, bronzish to copperish gray to blue. Antenna yellow, some specimens slightly black anterodorsally. Facial series with 2 setae on each side; face also distinctly two toned, a narrow, bare, shiny, vertical stripe that is bordered laterally by dense, palely golden-white microtomentum; parafacial very narrow, densely silvery white microtomentose. Gena relatively short, gena-to-eye ratio 0.06--0.10. *Thorax*: Mesonotum almost uniformly colored and invested with light dusting of microtomentum, lacking distinct stripes or isolated spots; presutural supra-alar seta well developed. Wing hyaline; costal vein ratio 0.71--0.90; M vein ratio 0.57--0.62. Coxae, femora and tibiae black, tarsi yellow; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin. *Abdomen*: Tergites shiny black. Male terminalia (Figs [77--80](#F27){ref-type="fig"}): Epandrium in posterior view (Fig. [77](#F27){ref-type="fig"}) disconnected dorsally, each lateral portion vertically elongate and robustly developed, height nearly 3× width, ventral portion on each side tapered laterally to medial, pointed apex, sloping ventral margin more conspicuously setulose, setulae mostly straight, also setulose dorsally, these setulae shallowly curved; cerci in posterior view (Fig. [77](#F27){ref-type="fig"}) very narrowly developed, elongate, bearing more setulae dorsally, linear; gonite in lateral view (Fig. [80](#F27){ref-type="fig"}) shallowly and broadly zigzagged, each apex narrowed, in ventral (Fig. [79](#F27){ref-type="fig"}) view linear, rod-like; aedeagus in lateral view (Fig. [80](#F27){ref-type="fig"}) generally rectangular, more broadly developed apically, apical margin slightly emarginate, with an elongate, narrow membranous extension, in ventral view an elongate, narrow, rod-like structure, narrowly truncate basally, apex roundly tapered to narrow point; phallapodeme in lateral view (Fig. [80](#F27){ref-type="fig"}) with distinctive, relatively broad keel, each extended process abruptly narrowed, in ventral view elongate with each end wider and with short, midheight papilla-like extensions laterally; hypandrium in lateral view (Fig. [80](#F27){ref-type="fig"}) mostly rectangular with posterior, narrow process angled toward aedeagus, in ventral view (Fig. [79](#F27){ref-type="fig"}) rectangular, wider than long, anterior margin produced to form a short, medial point, posterior margin very shallowly concave.
{#F26}
![*Lamproclasiopa nana* (Williston). (Brazil. Paraná: Curitiba) **77** epandrium and cerci, posterior view **78** same, lateral view **79** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **80** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g027){#F27}
#### Type material.
The lectotype male of *Discocerina nana* Williston (designated by [@B21]: 824) is labeled "Co-type \[circular label with a yellow border\]/1000 feet/\[a black, square label\]/St. Vincent, W.I. H. H. Smith./W.Indies. 1907--66./Discocerina nana Will. \[handwritten, two red submarginal borders\]/LECTOTYPE Discocerina nana Will. ♂ By W.N.Mathis \[handwritten except for "LECTOTYPE" and "By", black sub-border\]." the lectotype is double mounted (pin in a rectangular piece of cardboard), is in good condition, and is deposited in the BMNH. There are also eight paralectotypes as follows: BMNH (5♂, 1♀), AMNH (1♂, 1♀). Williston, in the original description, noted that the type series included "Numerous specimens."
#### Type locality.
West Indies. St. Vincent (13°15\'N, 61°11\'W).
#### Other specimens examined.
BOLIVIA. **La Paz**: Guanay (22 km SE; 15°17.8\'S, 68°15.6\'W; 540 m), 17 Mar 2001, W. N. Mathis (1♂; USNM).
BRAZIL. **Paraná**: Colombo (Santa Monica tennis club; 25°23.1\'S, 49°08.8\'W; 860 m), 18 Mar 2015, Daniel N. R. Costa (1♂; DZUP); Curitiba, Parque Iguaçú (25°33.4\'S, 49°13.6\'W; 880 m), 20--31 Aug 2012, Daniel. N. R. Costa (5♂; DZUP); Curitiba, Universidade Federal do Paraná, Reserva Biológica (25°26.9\'S, 49°14\'W; 915 m), 6--13 Feb 2010, D. and W. N. Mathis (2♂, 1♀; DZUP, USNM).
COLOMBIA. **Antioquia**: Medellín (6°14.1\'N, 75°34.5\'W; coffee finca; 1525 m), Nov 1955, W. B. Heed (2♂; USNM). **Cauca**: Popayán (30 km N; 2°36\'N, 76°31.6\'W), M. R. Wheeler (1♂; USNM). **Valle de Cauca**: Palmira (3°32.4\'N, 76°18.4\'W), Mar 1958, M. R. Wheeler (1♂, 1♀; USNM). Cali (10 km W; 3°26.7\'N, 76°37.3\'W; 1640 m), 20 Mar 1955, E. I. Schlinger, E. S. Ross (1♂; USNM).
COSTA RICA. **Cartago**: Juan Viñas (09°53.6\'N, 83°15.3\'W), 28 Apr 1910, P. P. Calvert (1♀; ANSP); La Suiza (9°51.5\'N, 83°37.5\'W), 6 May 1926, P. Schild (3♀; ANSP, USNM). **Puntarenas**: Pedregosa (09°08.4\'N, 83°43.5\'W), D. L. Rounds (1♀; ANSP). San José: Moravia, Zurquí de Moravia (Tower path; 10°2.8\'N, 84°0.6\'W, 1600 m), 6 Sep-29 Dec 2012, 2013, Project ZADBI (2♂, 5♀; MNCR-A).
CUBA. **Sancti Spiritus**: Topes de Collantes (21°55.2\'N, 80°02\'W; 350 m), 10 Dec 1994, W. N. Mathis (1♂; USNM).
DOMINICA. Antrim Valley (15°20.7\'N, 61°22.2\'W; 305 m), 17 Mar 1956, J. F. G. Clarke (1♂, 3♀; USNM). Clarke Hall (15°24.5\'N, 61°23.7\'W), 16 Jan-21--31 Mar 1965, W. W. Wirth (1♂, 1♀; USNM). Fond Figues River (15°24\'N, 61°18\'W; 122 m), 30 Jan-22 Apr 1965, 1966, R. J. Gagné, W. W. Wirth (9♀, 2♀; USNM). G'leau Gommier near Belles (15°25.4\'N, 61°20.4\'W), 17 Mar 1956, J. F. G. Clarke (2♂, 6♀; USNM). Layou River (15°25.8\'N, 61°21.4\'W; 5 km E), 23 Mar 1989, W. N. Mathis (1♂; USNM). Pont Cassé (15°22.7\'N, 61°20.6\'W), 12 Feb-18 Jun 1965, 1991, D. and W. N. Mathis, W. W. Wirth (11♂, 2♀; USNM). South Chiltern Estate (15°14.8\'N, 61°21.8\'W), 2 Feb 1965, W. W. Wirth (1♀; USNM). Toucari (2 km S; 15°36.6\'N, 61°27.8\'W), 21 Mar 1989, W. N. Mathis (2♂, 1♀; USNM). Trafalgar Falls (15°19\'N, 61°21\'W; 365 m), 6 Mar-19 Jun 1965, 1991, R. J. Gagné, D. and W. N. Mathis, W. W. Wirth (4♂, 4♀; USNM).
DOMINICAN REPUBLIC. **Barahona**: San Rafael (18°01.9\'N, 71°08.4\'W), 22 Mar 1999, W. N. Mathis (3♂, 3♀; USNM). **La Vega**: Jarabacoa (5 km S; 19°05.8\'N, 70°36.5\'W; 640 m), 8-20 May 1995, W. N. Mathis (9♂, 2♀; USNM); Paso de la Vaca, road from Monseñor Nouel to Constanza (19°04\'N, 70°16.9\'W; 1500 m), 27 Dec 1955, J. Maldonado Capriles (1♂, 1♀; USNM). **Pedernales**: Rio Mulito (21 km N Pedernales; 18°09.3\'N, 71°45.6\'W; 270 m), 18-20 Mar 1999, W. N. Mathis (1♀; USNM). **San Jose de Ocoa**: San Jose de Ocoa (30.6 km N; 18°49.5\'N, 70°30\'W; 914 m), 30 Jul 1991, D. Grimaldi, J. Stark (2♂; AMNH).
JAMAICA. **Portland**: Reach Falls (18°01.8\'N, 76°18.7\'W), 15 May 1996, D. and W. N. Mathis, H. Williams (1♂; USNM); Reach Falls, Drivers River (18°01.9\'N, 76°18.7\'W; 70 m), 25 Apr 2000, W. N. Mathis (1♀; USNM); Section (0.5 km E; 18°05.2\'N, 76°43.9\'W; 1020 m), 28 Apr 2000, W. N. Mathis (9♂, 4♀; USNM). **St. Andrew**: Cinchona (18°04.4\'N, 76°39.3\'W; 1400 m), 29 Apr 2000, W. N. Mathis (3♀; USNM); Hardwar Gap (18°04.2\'N, 76°44\'W), 17 May 1996, D. and W. N. Mathis, H. Williams (2♂; USNM); Hardwar Gap (18°04.2\'N, 76°44\'W; 1170 m), 27 Apr 2000, W. N. Mathis (4♂, 1♀; USNM); Hollywell (18°05.2\'N, 76°43.9\'W; 1100 m); 28 Apr 2000, W. N. Mathis (7♂; USNM); Mavis Bank (near coffee factory; 18°01.4\'N, 76°39.7\'W; waterfall), 21-23 Apr 2000, W. N. Mathis (1♂; USNM); Mavis Bank (1.5 km W; 18°01.4\'N, 76°39.9\'W), 22 Apr 2000, W. N. Mathis (1♂, 2♀; USNM); Mavis Bank (1.7 km E; 18°02.4\'N, 77°39.5\'W; 575 m), Yallahs River, 21-22 Apr-1 May 2000, W. N. Mathis (1♂; USNM); Newcastle (6 km S; 18°04.3\'N, 76°42.6\'W; 950 m; waterfall), 30 Apr 2000, W. N. Mathis (14♂; USNM).
PERU. **Cuzco**: Paucartambo, Puente San Pedro (ca 50 km NW Pilcopata; 13°03.3\'S, 71°32.8\'W; 1600 m), 3 Sep 1988, W. N. Mathis (1♂, 1♀; USNM). **Madre de Dios**: Río Manu, Pakitza (11°56.6\'S, 71°16.9\'W; 250 m), 9-23 Sep 1988, W. N. Mathis (1♀; USNM).
PUERTO RICO. Adjuntas (near; 18°09.7\'N, 66°46.6\'W), 22 Sep 1995, D. and W. N. Mathis (25♂, 4♀; USNM). El Verde (18°13.5\'N, 66°0-3.2\'W; near biological station road), 3 Feb 1989, S.A. Marshall (1♂, 6♀; GUE). El Yungue (18°18.4\'N, 65°45.6\'W), 20-22 Mar 1954, J. Maldonado Capriles (1♂; USNM). Maricao, Los Viveros (18°10.5\'N, 66°59.2\'W), 21 Sep 1995, D. and W. N. Mathis (1♂, 1♀; USNM). Maricao (4 km WNW; 18°10.7\'N, 66°59.6\'W), 21 Sep 1995, D. and W. N. Mathis (1♂; USNM).
ST. LUCIA. Fond St. Jacques (13°50\'N, 61°02\'W), 13-14 Jun 1991, D. and W. N. Mathis (16♂, 8♀; USNM). Soufrière Botanical Garden (13°51\'N, 61°04\'W), 12 Jun 1991, D. and W. N. Mathis (1♂, 1♀; USNM).
ST. VINCENT. **St. Patrick**: Hermitage (6 km E Spring Village at Cumberland River; 13°14\'N, 61°13.2\'W; 550 m), 9 Jul 1989, M. Sorensson, B. Mårtensson (8♂, 2♀; MZLU); Hermitage (13°15\'N, 61°12.9\'W), 9 Sep 1997, W. N. Mathis (1♂; USNM).
TRINIDAD AND TOBAGO. Tobago. **St. John**: Charlotteville (2 km S; 11°19\'N, 60°33\'W), 10 Jun 1993, 1994, W. N. Mathis (1♂, 1♀; USNM); Parlatuvier (creek; 11°17.9\'N, 60°35\'W), 20 Apr 1994, W. N. Mathis (1♂, 1♀; USNM). **St. Paul**: Roxborough (6 km NNW; 11°16\'N, 60°35.4\'W), 20 Apr 1994, W. N. Mathis (8♂, 6♀; USNM); Roxborough (6.5 km N; 11°17\'N, 60°35\'W), 14 Jun 1993, W. N. Mathis (1♂; USNM). Trinidad. **St. George**: Marianne River (9 km S; 10°46\'N, 61°18\'W), 25 Jun 1993, W. N. Mathis (1♀; USNM); Mount St. Benedict (10°39\'N, 61°24\'W), 18-21 Jun 1993, W. N. Mathis (1♂; USNM); Mount St. Benedict (10°39\'N, 61°24\'W; creek near base), 19 Jun 1993, W. N. Mathis (1♂; USNM).
#### Distribution
(Fig. [81](#F28){ref-type="fig"}). Neotropical: Bolivia (La Paz), Brazil (Paraná), Costa Rica (Cartago), Colombia (Antioquia, Cauca, Valle de Cauca), Peru (Cuzco, Madre de Dios), Trinidad and Tobago, West Indies (Cuba, Dominica, Dominican Republic, Grenada, Jamaica, Puerto Rico, St. Lucia, St. Vincent).
{#F28}
#### Remarks.
This species is beautifully and strikingly colored, especially the head, and this color pattern distinguishes it from congeners. These characters, as noted in this species' diagnosis, are as follows: Frons distinctly two toned; fronto-orbits narrow; medial triangular area densely microtomentose, velvety black; mesofrons, other than narrow, medial triangle, seriaceus, bronzish to copperish gray to blue; antenna yellow, some specimens slightly black anterodorsally. Facial series with two setae on each side; face also distinctly two toned, a narrow, bare, shiny, vertical stripe that is bordered laterally by dense, palely golden-white microtomentum; parafacial very narrow, densely silvery white microtomentose; wing hyaline.
Just as the color pattern of the head is unique among congeners, the structures of the male terminalia likewise represent a departure from the more typical pattern. This is especially evident in the shape of the epandrium, which exhibits a dorsal gap and the lateral arms are greatly thickened. Like many other species of *Lamproclasiopa*, there are loosely clustered setulae along the ventromedial margin. The cerci are slender and comparatively elongated. The internal structures are more typical, although the aedeagus in lateral view is more rectangular than being narrowly triangular and a slender, membranous distiphallus is often exposed.
### The *furvitibia* group (*Lamproclasiopa furvitibia*, *Lamproclasiopa xanthocera*) {#SECID0EEOCK}
**Diagnosis.** Body with extensive surfaces sparsely to densely microtomentose. *Head*: Frons and face generally unicolorous; antenna yellow; gena relatively short (gena-to-eye ratio less than 0.06-0.09); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta lacking; katepisternum and anepisternum thinly microtomentose, generally appearing dull, not shiny. Wing generally hyaline to very faintly infumate; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with 4-5 stout, peg-like setae on apical third along posteroventral margin; tibiae variable, mostly black with apex yellow or entirely yellow. *Abdomen*: Male terminalia: Ventral epandrial margin bearing a cluster of setulae; aedeagus in lateral view with margins irregular.
**Remarks.** This species group is partially based on homoplasious characters, and we cannot confirm its monophyly. The two included species are similar to each other and the species group can be diagnosed. Moreover, the ventral epandrial margins bear a cluster of closely set setulae (also expressed in a few other congeners) and the aedeagus in lateral view is irregular, sinuous. These are the bases for recognition of this species group.
#### Lamproclasiopa furvitibia sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/9B556FB7-2108-4B56-8999-3D9AC58D4D03
[Figs 14](#F6){ref-type="fig"} [, 82--85](#F29){ref-type="fig"}
##### Diagnosis.
This species can be distinguished from congeners by the following combination of characters: Small to moderately small shore flies, body length 2.00-2.24 mm. *Head*: Frons with yellowish tan to golden tan microtomentum, with slightly silver white microtomentum anteriorly and two small areas shiny black, without microtomentum; mesofrons evident by slight lateral lines. Antenna yellow; basal flagellomere with slightly darker dorsal margin. Face completely and more or less uniformly silvery white microtomentose, lacking vertical stripes; 2 prominent facial setae, dorsal seta at midheight, other seta near epistomal margin; parafacial thin, more densely silvery white microtomentose than face. Gena relatively short, gena-to-eye ratio 0.07--0.09. *Thorax*: Generally black. Mesonotum black with thin, golden brown microtomentum, subshiny, slightly less dense than microtomentum of frons; presutural supra-alar seta lacking; pleural areas more sparsely microtomentose than mesonotum, blackish brown to black, becoming less microtomentose ventrally and posteriorly, subshiny to shiny. Wing completely hyaline, lacking darkened areas; costal vein ratio 0.59--0.60; M vein ratio 0.70--72. Femora grayish to blackish brown, subshiny; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; tibiae blackish brown with distal third yellow; tarsi yellow. *Abdomen*: Tergites shiny black, with little or very sparse microtomentum. Male terminalia (Figs [82--85](#F29){ref-type="fig"}): Epandrium in posterior view (Fig. [82](#F29){ref-type="fig"}) nearly as wide as high, as in inverted, bow-shaped U, dorsal arch narrow, becoming wider ventrally, in lateral view (Fig. [83](#F29){ref-type="fig"}) tear-drop shaped, widest ventrally, ventral margin broadly rounded and bearing distinct row or setulae; cercus in posterior view (Fig. [82](#F29){ref-type="fig"}) hemispherical, not fused with ventral margin of cercal cavity, ventral portion diffuse, membranous, generally uniformly setulose, in lateral view (Fig. [83](#F29){ref-type="fig"}) narrowly and irregularly semicircular, slightly wider subdorsally than ventrally; gonite in ventral view (Fig. [84](#F29){ref-type="fig"}) more or less irregularly triangular, narrowed toward aedeagal base, wider toward hypandrium, in lateral view (Fig. [85](#F29){ref-type="fig"}) moderately elongate, rod-like, shallowly curved, end toward hypandrium narrower than opposite end; aedeagus in ventral view (Fig. [84](#F29){ref-type="fig"}) narrowly elongate, 6× longer than wide, basal half wider than narrow apical half, tapered, base narrowly bifurcate, apical half digitiform, apex narrowly rounded; phallapodeme in lateral view (Fig. [85](#F29){ref-type="fig"}) triangular, angular extensions toward aedeagal base and toward hypandrium digitiform, extension toward aedeagal base longer than extension toward hypandrium, keel robust, tapered, apex narrowly rounded, in ventral view (Fig. [84](#F29){ref-type="fig"}) elongate, bone-like, with apical and basal cross, both apices truncate; hypandrium in ventral view (Fig. [84](#F29){ref-type="fig"}) with anterior 2/3 more or less rectangular, lateral margin scalloped, posterior 1/3 slightly flared laterally, thumb-like extensions, posterior margin deeply emarginate, broadly V-shaped, in lateral view (Fig. [85](#F29){ref-type="fig"}) elongate, shallowly angularly rod-like, an elongate and shallow Z.
![*Lamproclasiopa furvitibia* sp. n. (Costa Rica. San José: Moravia) **82** epandrium and cerci, posterior view **83** same, lateral view **84** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **85** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g029){#F29}
##### Type material.
The holotype male of *Lamproclasiopa furvitibia* is labeled "COSTA RICA. Prov. San José. Moravia. Zurquí de Moravia, Tower path. 1600m. 6--13 SEP 2013. Proyeto ZADBI. Malaise trap \#1, 0m, ZADBI-1136. -84:00:57 10:02:58 \#107741/INB0004403109 INBIOCRI COSTA RICA \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa furvitibia* Costa, Mathis & Marinoni MNCR-A \[red\]." The holotype is double mounted (glued to a paper triangle) and is in very good condition, and is deposited in MNCR-A. Thirty-four paratypes (13♂, 21♀; MNCR-A, USNM) bear the same label data as the holotype. Other paratypes are as follows: COSTA RICA. **Cartago.** Paraíso, Parque Nacional Tapantí (09°43.3\'N, 83°46.5\'W; 1600 m), 9 Mar -- 29 Sep 2013, Proyeto ZADBI (7♂, 14♀; MNCR-A). **Puntaneras.** Coto Brus, Las Alturas Biological Station (08°56.7\'N, 82°50\'W; 1500--1600 m), 7--13 May 2013, Proyeto ZADBI (1♀; MNCR-A). **Guanacaste.** Macizo Miravalles, Cabro Muco Station (10°43.1\'N, 84° 51.3\'W; 1100 m), 22 Jun--9 Jul 2003, J. Azoifeifa. (2♂; MNCR-A).
##### Type locality.
Costa Rica. San José. Zurquí de Moravia (10°2.8\'N, 84°0.6\'W; 1588 m).
##### Distribution
(Fig. [14](#F6){ref-type="fig"}). Neotropical: Costa Rica (San José, Puntaneras).
##### Etymology.
The species epithet, *furvitibia*, is of Latin derivation, meaning darkened tibia and refers to the partially darkened tibiae, one of the distinguish features of this species.
##### Remarks.
This species is closely related to *Lamproclasiopa xanthocera* but can be distinguished from it by having two small shiny black areas on the anterolateral portion of the frons and by having mostly blackish brown tibiae with the distal third being yellow. The shape of structures of the male terminalia also distinguishes this species from congeners.
#### Lamproclasiopa xanthocera sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/B35ED0F5-9221-4E9B-A7E5-0E66ED7FADB1
[Figs 86--87](#F30){ref-type="fig"} [, 88--91](#F31){ref-type="fig"} [, 104](#F36){ref-type="fig"}
##### Diagnosis.
This species can be distinguished from congeners by the following combination of characters: Small to moderately small shore flies, body length 1.73--2.18 mm. *Head*: Frons with yellowish tan to golden tan microtomentum, some areas slightly darker; parafrons with slightly thinner investment of microtomentum; mesofrons evident by slight lateral lines. Antenna yellow; basal flagellomere with slightly darker dorsal margin. Face completely and more or less uniformly silvery white microtomentose, more thinly microtomentose ventrally except for extreme ventral margin, lacking vertical stripes; 2 prominent facial setae, dorsal seta at midheight, other seta near epistomal margin; parafacial thin, more densely silvery white microtomentose than face. Gena relatively short, gena-to-eye ratio 0.06--0.08. *Thorax*: Generally black. Mesonotum black with thin, golden brown microtomentum, subshiny, although less dense than microtomentum of frons; presutural supra-alar seta lacking or indistinguishable from surrounding setae pleural areas more sparsely microtomentose than mesonotum, blackish brown to black, becoming less microtomentose ventrally and posteriorly, subshiny to shiny. Wing completely hyaline, lacking darkened areas; costal vein ratio 0.59--0.60; M vein ratio 0.57--0.65. Femora grayish to blackish brown, subshiny; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; tibiae and tarsi yellow. *Abdomen*: Tergites shiny black, with little or very sparse microtomentum. Male terminalia (Figs [88--91](#F31){ref-type="fig"}): Epandrium in posterior view (Fig. [88](#F31){ref-type="fig"}) nearly as wide as long, as in inverted U, dorsal arch narrow, becoming wider ventrally, in lateral view (Fig. [89](#F31){ref-type="fig"}) narrowly triangular, widest ventrally, ventral margin broadly rounded; cercus in posterior view (Fig. [88](#F31){ref-type="fig"}) hemispherical, not fused with ventral margin of cercal cavity, uniformly setulose, in lateral view (Fig. [89](#F31){ref-type="fig"}) narrowly semicircular, slightly wider subdorsally than ventrally; gonite in ventral view (Fig. [90](#F31){ref-type="fig"}) more or less triangular, narrowed toward aedeagal base, wider toward hypandrium, in lateral view (Fig. [91](#F31){ref-type="fig"}) elongate, rod-like, end toward hypandrium narrower than opposite end; aedeagus in ventral view (Fig. [90](#F31){ref-type="fig"}) narrowly elongate, 6× longer than wide, nearly parallel sided, apex pointed, in lateral view (Fig. [91](#F31){ref-type="fig"}) elongate, L-shaped, short arm basally, wider, thereafter parallel sided, membranous on apical ¼; phallapodeme in lateral view (Fig. [91](#F31){ref-type="fig"}) triangular, angle toward aedeagal base digitiform, longer than extension toward hypandrium, keel tapered, apex rounded, in ventral view (Fig. [90](#F31){ref-type="fig"}) rectangular, apical 1/3 to hypandrium slightly tapered, both apices truncate; hypandrium in ventral view (Fig. [90](#F31){ref-type="fig"}) as a broad, short H with posterior arms flaring posterolaterally, posterior margin broadly emarginate, anterior margin shallowly concave; in lateral view (Fig. [91](#F31){ref-type="fig"}) elongate, rod-like, anterior apex irregular.
{#F30}
![*Lamproclasiopa xanthocera* sp. n., male holotype (Brazil. Paraná: Curitiba) **88** epandrium and cerci, posterior view **89** same, lateral view **90** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **91** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g031){#F31}
##### Type material.
The holotype male of *Lamproclasiopa xanthocera* is labeled "**BRAZIL.** Paraná: Curitiba, UFPR \[Universidade Federal do Paraná, Reserva Biológica\] (25°26.9\'S, 49°14\'W; 915 m),1--2Feb2010\[,\] D. & W. N. Mathis/USNM ENT 00118308 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa xanthocera* Costa, Mathis & Marinoni DZUP \[red\]." The holotype is double mounted (minuten pin in a block of plastic), is in excellent condition, and is deposited in DZUP. Paratypes are as follows: BRAZIL. **Paraná.** Morro do Araçatuba (Município de Tijucas do Sul; 25°53.8\'S, 49°01.2\'W; 910 m), 27 Feb 2015, W. N. Mathis (2♀; DZUP, USNM). **São Paulo.** Estação Biológica de Boracéia, Salesópolis (23°32\'S, 45°50.8\'W), Aug 1969, N. Papavero (1♀; MZUSP).
##### Type locality.
Brazil. Paraná. Curitiba, Universidade Federal do Paraná, Reserva Biológica (25°26.9\'S, 49°14\'W; 915 m).
##### Distribution
(Fig. [104](#F36){ref-type="fig"}). Neotropical: Argentina, Brazil (Paraná, São Paulo).
##### Etymology.
The species epithet, *xanthocera*, is of Latin derivation, meaning yellow horn and refers to the yellow antenna, one of the distinguishing features of this species.
##### Remarks.
Although similar to *Lamproclasiopa bisetulosa*, this species is distinguished from it and other congeners by having a generally microtomentose body, yellow antenna with little or no darkening along dorsal surfaces, a hyaline wing, and a blackish yellow foretarsus. The shape of structures of the male terminalia also distinguishes this species, especially the elongate, thick, and conspicuously sinuous aedeagus with an apical papilla-like apex.
### The *nadineae* group (*Lamproclasiopa aliceae*, *Lamproclasiopa argentipicta*, *Lamproclasiopa nadineae*) {#SECID0EAEDK}
**Diagnosis.** Body generally shiny black. *Head*: Frons and face generally unicolorous; gena moderately high (gena-to-eye ratio 0.13--0.22); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta lacking; katepisternum, especially anterior half, and anteroventral portion of anepisternum shiny black. Wing generally hyaline to faintly infumate; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin.
**Remarks.** Two of the species in this species group, *Lamproclasiopa aliceae* and *Lamproclasiopa nadineae*, form a monophyletic lineage that is characterized by synapomorphies (presutural supra-alar seta lacking; katepisternum, especially anterior half, and anteroventral portion of anepisternum shiny black; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin). These two species are likewise unique in occurring only in the Nearctic Region. The inclusion of *Lamproclasiopa argentipicta* in this group may be artificial, having a homoplasious basis. We have not discovered a synapomorphy that is unique to these three species.
#### Lamproclasiopa aliceae sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/FE0C666F-5C59-416E-A3DE-E66AB7937A60
[Figs 92--93](#F32){ref-type="fig"} [, 94--97](#F33){ref-type="fig"} [, 111](#F39){ref-type="fig"}
##### Diagnosis.
This species is easily distinguished from congeners by the following combination of characters: Small to moderately small shore-flies, body length 1.80--2.10 mm. *Head*: Frons with dorsal 2/3 bearing brown, moderately sparse microtomentum, thereafter ventrally with a transverse bear, shiny black band, then a more grayish microtomentose, transverse band just before margin and dorsad of antennal bases. Ventral portion of face mostly unicolorous, moderately grayish brown microtomentose, antennal grooves more densely, whitish gray microtomentose; parafacial grayish white to creamy white. Gena moderately high, height subequal to height of basal flagellomere. Gena-to-eye ratio 0.14--0.17. *Thorax*: Mesonotum moderately sparsely brown microtomentose, mostly appearing subshiny black, lacking elongate spots; presutural supra-alar seta lacking. Katepisternum, especially anterior half, and anteroventral portion of anepisternum shiny black. Wing hyaline and lacking stump veins; costal vein ratio 0.55--0.63; M vein ratio 0.51--0.55. Forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; femora black; tibiae with basal 2/3-¾ black, apical ¼-1/3 yellowish; tarsi yellow. *Abdomen*: Tergites very sparsely microtomentose medially to complete bare laterally, mostly shiny black. Male terminalia (Figs [94--97](#F33){ref-type="fig"}): Epandrium in posterior view (Fig. [94](#F33){ref-type="fig"}) roundly U-shaped except for incised ventromedial opening or pocket, widest at midheight, each lateral arm distinctly wider basally, ventromedial pocket between epandrial arms bottle-shaped, with dorsal half as a narrowed neck and ventral half wider than high, in lateral view (Fig. [95](#F33){ref-type="fig"}) narrow, angulate, widest just ventrad of cercal cavity, angulate, epandrial arm expanded at apex; cercus hemispherical, gradually tapered ventrally to broadly rounded apex, evenly setulose over length; gonite in lateral view (Fig. [97](#F33){ref-type="fig"}) robustly rod-like, essentially straight, narrower and abruptly curved toward aedeagal base, in ventral view (Fig. [96](#F33){ref-type="fig"}) clavate, wider toward hypandrium, much narrower, digitiform on extension toward aedeagal base; aedeagus in lateral view (Fig. [97](#F33){ref-type="fig"}) more or less funnel-like, curved subapically, widest basally, thereafter apically curved to acute point, in ventral view (Fig. [96](#F33){ref-type="fig"}) longer than wide, straight, base expanded, lateral margins of base rounded, thereafter apically evenly tapered to point; phallapodeme in lateral view (Fig. [97](#F33){ref-type="fig"}) as an inverted Y with an elongate, narrow, extension toward aedeagal base, keel elongate, narrow, parallel sided, apex irregularly rounded, in ventral view (Fig. [96](#F33){ref-type="fig"}) hourglass-like, sub-rectangular at apex toward aedeagal base, broadly expanded with lateral phalanges toward hypandrium, this apex truncate; hypandrium in lateral view (Fig. [97](#F33){ref-type="fig"}) elongate, narrow, bar-like, narrowed anteriorly, in ventral view (Fig. [96](#F33){ref-type="fig"}) robustly X-shaped to quadrate with lateral and poster margins concave.
{#F32}
![*Lamproclasiopa aliceae* sp. n. (USA. New Mexico: Silver City) **94** epandrium and cerci, posterior view **95** same, lateral view **96** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **97** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g033){#F33}
##### Type material.
The holotype male of *Lamproclasiopa aliceae* is labeled "**U\[nited\]S\[tates of\]A\[merica\]. N\[ew\]M\[exico\].** Grant: Silver City (32°46.4\'N, 108°16.5\'W; 1790 m), 14 Aug 2007,D.&W.N.Mathis/USNM ENT 00118306 \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa aliceae* Costa, Mathis & Marinoni USNM \[red\]." The holotype is double mounted (minuten pin in a block of plastic), is in excellent condition, and is deposited in the USNM. Eight paratypes (7♂, 1♀; DZUP, USNM) bare the same label data as the holotype.
##### Type locality.
United States. New Mexico. Grant: Silver City (Big Ditch; 32°46.4\'N, 108°16.5\'W; 1790 m). The "Big Ditch" is a large, canal-sized ditch that traverses Silver City, and during dry-weather seasons, the ditch has a small stream running through it. The ditch is frequently scoured out when heavy rains occur, sometimes resulting in flash floods in the "Big Ditch." The type series was collected when dry weather prevailed.
##### Distribution
(Fig. [111](#F39){ref-type="fig"}). Nearctic: United States (New Mexico).
##### Etymology.
The species epithet, *aliceae*, is a Latin genitive patronym to honor Alice Joy Brown† (nee Peacock, 1931--2016), a wonderful friend and supporter.
##### Remarks.
Although similar and apparently closely related to *Lamproclasiopa nadineae*, this species is distinguished from that species by the shape of the ventral portion of the epandrium, which has a bottle-shaped gap ventrally between the lateral arms. From other congeners, this species is distinguished by the shiny black katepisternum, especially its anterior half, and the anteroventral portion of the anepisternum. The forefemur also bears 4--5 stout, peg-like setae on the apical half of the posteroventral margin.
#### Lamproclasiopa argentipicta sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/D26C06A0-302A-4EDE-B2E5-70D6C9DB50E7
[Figs 98--99](#F34){ref-type="fig"} [, 100--103](#F35){ref-type="fig"} [, 104](#F36){ref-type="fig"}
##### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Moderately small to medium-sized shore flies, body length 2.90--3.20 mm. *Head*: Frons with two longitudinal, grayish microtomentose stripes; fronto-orbits and narrow, medial triangular area shiny black. Antenna blackish brown. Face with light silver microtomentum, except for shiny black lateral margins; parafacials white, microtomentose. Gena moderately high, gena-to-eye ratio 0.16--0.22. *Thorax*: Mesonotum shiny black, covered with brownish microtomentum; presutural supra-alar seta lacking or indistinguishable from surrounding setae; pleural region less microtomentose, anepisternum and katepisternum almost bare, concolorous with mesofrons. Wing hyaline, lacking any pattern or markings. Costal vein ratio 0.45--0.50; M vein ratio 0.57. Forecoxae light gray, mid and hind coxae blackish brown; forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; femora and tibiae blackish brown, except for distalmost part of tibiae, yellowish; tarsi yellow. *Abdomen*: Generally shiny blackish brown, sparsely microtomentose; tergites 5 larger than previous tergites. Male terminalia (Figs [100--103](#F35){ref-type="fig"}): Epandrium in posterior view (Fig. [100](#F35){ref-type="fig"}) with dorsal 2/3 quadrate, as wide as high, corners rounded, ventral third as 2 thumb-like projections, dorsal portion thickly developed, as wide or wider than width of lateral structure, setulae evenly distributed dorsally and laterally, becoming very sparse ventrally, ventral extensions bearing tiny setulae in verticomedial alignment, apically with cluster of small setulae, in lateral view (Fig. [101](#F35){ref-type="fig"}) as 2 right angles, dorsal portion more robust, thick, then a right anterior angle, then a ventral right angle to form digitiform extension that bears closely set setulae along anterior margin; cerci in posterior view (Fig. [102](#F35){ref-type="fig"}) elongate, moderately thin, generally shallowly arched, slightly ventrally than dorsally, dorsal angle with vertex narrowly rounded, ventral apex acutely pointed, in lateral view elongate, narrow, ventral portion tapered; aedeagus in lateral view (Fig. [103](#F35){ref-type="fig"}) as an irregular funnel, tapered from thick base to pointed apex, narrowed more abruptly on apical 1/8 then narrowly pointed at right angle, in ventral view (Fig. [102](#F35){ref-type="fig"}) generally clavate, gradually becoming wider from truncate base toward apex, widest subapically, thereafter abruptly narrowed to slender, digitiform apex; phallapodeme in lateral view (Fig. [103](#F35){ref-type="fig"}) narrowly triangular with vertex toward aedeagal base elongate and narrow, keel narrow, subequal to process extended toward hypandrium, in ventral view (Fig. [102](#F35){ref-type="fig"}) narrow, elongate with a sub-basal cross-piece and narrow, lateral extensions, thereafter almost parallel sided; gonite in lateral view irregularly rod-like, in ventral view (Fig. [102](#F35){ref-type="fig"}) wider than high and with a mediobasal, short, digitiform projection; hypandrium in lateral view (Fig. [103](#F35){ref-type="fig"}) thin, elongate, rod-like, shallowly curved, slightly wider anteriorly than posteriorly, in ventral view (Fig. [102](#F35){ref-type="fig"}) robustly U-shaped with thickened base, broadly rounded anterior margin and deeply U-shaped posterior emargination.
{#F34}
![*Lamproclasiopa argentipicta* sp. n., male paratype (Costa Rica. San José: Moravia) **100** epandrium and cerci, posterior view **101** same, lateral view **102** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **103** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g035){#F35}
##### Type material.
The holotype male of *Lamproclasiopa argentipicta* is labeled "COSTA RICA. Prov. San José. Moravia. Zurquí de Moravia, Tower path. 1600m. 9--16 AGO 2013. Proyeto ZADBI. Malaise trap \#1, 0m, ZADBI-1018. -84:00:57 10:02:58 \#107537/INB0004433874 INBIOCRI COSTA RICA \[plastic bar code label\]/HOLOTYPE ♂ *Lamproclasiopa argentipicta* Costa, Mathis & Marinoni MNCR-A \[red\]." The holotype is double mounted (glued to a paper triangle) and is in very good condition, and is deposited in MNCR-A. Three paratypes (1♂, 2♀; MNCR-A, USNM) bear the same label data as the holotype.
##### Type locality.
Costa Rica. San José. Zurquí de Moravia (10°2.8\'N, 84°0.6\'W; 1588 m).
##### Distribution
(Fig. [104](#F36){ref-type="fig"}), Neotropical: Costa Rica (San José).
{#F36}
##### Etymology.
The species epithet, *argentipicta*, is of Latin derivation and means painted with silver, referring to the silver microtomentose areas of this species, especially its face.
##### Remarks.
This species is apparently closely related to *Lamproclasiopa hendeli*, based on external features, such as the shiny black body and the face that is covered with silver gray microtomentum. The male terminalia, however, are unique within *Lamproclasiopa*, with two setulose projections ventrally and with the ventral projection of the phallapodeme being very thin, appearing to be almost fused with the hypandrium.
#### Lamproclasiopa nadineae
Animalia
Diptera
Ephydridae
(Cresson)
[Figs 105--106](#F37){ref-type="fig"} [, 107--110](#F38){ref-type="fig"} [, 111](#F39){ref-type="fig"}
1. Ditrichophora nadineae[@B4]: 166.
2. Discocerina (Basila) nadineae. [@B10]: 116 \[generic combination\]. [@B33]: 738 \[Nearctic catalog\]. [@B24]: 165 \[world catalog\].
3. Lamproclasiopa nadineae. [@B36]: 39 \[generic combination\].
##### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Small to moderately small shore-flies, body length 1.65--2.50 mm. *Head*: Frons generally microtomentose but unevenly, microtomentum on ocellar triangle and especially along anterior margin just dorsad of antennal bases gray and denser, otherwise sparse and grayish brown to brown, areas toward anterior margin of frons yellowish orange to red; ocellar triangle extended to anterior margin. Antenna dark brown dorsally, extensively yellow to yellowish orange ventrobasally. Face generally microtomentose, becoming bare laterally, most prominent anteriorly in lateral view at ventral margin of antennal grooves; parafacial bare of ventroclinate setulae. Gena moderately high, gena-to-eye ratio 0.13--0.17. *Thorax*: Mesonotum black with uniform, fine, thin investment of gray to brown microtomentum, lacking pattern of spots; presutural supra-alar seta lacking or indistinguishable from surrounding setae. Katepisternum, especially anterior half, and anteroventral portion of anepisternum shiny black. Wing hyaline, lacking any maculation pattern or stump veins; costal vein ratio 0.50--0.52; M vein ratio 0.55--0.61. Forefemur with 4--5 stout, peg-like setae on apical third along posteroventral margin; legs generally black except for yellowish apices, yellowish apices of tibiae more extensive; tarsi yellow. *Abdomen*: Tergites subshiny to shiny black, generally lacking microtomentum or very sparse. Male terminalia (Figs [107--110](#F38){ref-type="fig"}): Epandrium in posterior view (Fig. [107](#F38){ref-type="fig"}) roundly U-shaped, except for ventral opening, oval, narrower dorsally, broadly rounded, widest at midheight, each lateral arm distinctly wider, ventromedial margin nearly straight, bearing cluster of setae ventrally; cercus hemispherical, gradually tapered ventrally to broadly rounded apex, evenly setulose over length; gonite in lateral view (Fig. [108](#F38){ref-type="fig"}) robustly rod-like, slightly curved, narrower toward hypandrium, in ventral view clavate, wider toward hypandrium, much narrower, digitiform on extension toward aedeagal base; aedeagus in lateral view (Fig. [110](#F38){ref-type="fig"}) more or less rectangular, widest apically and anteroventral corner more narrowly produced than posterodorsal corner, base rounded with slight, dorsal extension, in ventral view (Fig. [109](#F38){ref-type="fig"}) as a relatively short, moderately narrow structure, base rounded and with an medioapical papilla, apex moderately broadly rounded; phallapodeme in lateral view (Fig. [110](#F38){ref-type="fig"}) as a spike-heeled shoe with an elongate, narrow, extension toward aedeagal base, keel moderately long, narrow, irregularly rounded, in ventral view (Fig. [109](#F38){ref-type="fig"}) with thumb-like narrow process toward aedeagal base, thereafter toward hypandrium with right angle laterally, hypandrial 2/3 widest sub-basally thereafter concave to truncate apex; hypandrium in lateral view (Fig. [110](#F38){ref-type="fig"}) conspicuously sinuous, narrow, elongate, in ventral view robustly X-shaped with truncate anterior margin, lateral margins concave, posterior margin deeply V-shaped.
{#F37}
![*Lamproclasiopa nadineae* (Cresson). (USA. California. Jasper Ridge Biological Preserve) **107** epandrium and cerci, posterior view **108** same, lateral view **109** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **110** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g038){#F38}
##### Type material.
The holotype male of *Ditrichophora nadineae* Cresson is labeled "Berkeley Hills Alameda Co. IV.20.'08. Cal./♂/TYPE Ditrichophora NADINEAE E. T. Cresson. Jr. \[maroon red; "Ditrichophora NAEINEAE" handwritten." The holotype is double mounted (minuten pin in a thin, rectangular piece of card), is in excellent condition, and is deposited in the ANSP (6348). Four paratypes (1♂, 3♀; ANSP) bear the same locality label as the holotype.
##### Type locality.
United States. California. Alameda: Berkeley Hills (37°53.5N, 122°16.1\'W).
##### Other specimens examined.
UNITED STATES. Arizona. **Cochise**: Portal, Southwestern Research Station (31°53\'N, 109°12.4\'W), 5--9 Jun 1972, W. W. Wirth (1♂; USNM). California. **Alameda**: Berkeley (37°52.3\'N, 122°16.4\'W), 23 Mar 1919, B. Brookman, T. Aitken (1♀; USNM). Berkeley Hills (37°53.5\'N, 122°16.1\'W), 8 Mar-11 May 1908 (2♂, 3♀; ANSP). **Humboldt**: Willow Creek (40°56.4\'N, 123°37.9\'W), 26 Jul 1951, A. H. Sturtevant (1♂; USNM). **Kern**: Kern Canyon (35°41.8\'N, 118°40.1\'W), 1 Apr 1950, A. H. Sturtevant (2♂, 1♀; USNM). **Los Angeles**: Arcadia, Arboretum (34°8.6\'N, 118°3.2\'W), 14 Set 1949, A. H. Sturtevant (2♂, 5♀; USNM); Pasadena (34°9.4\'N, 118°7.9\'W), 15, 17 Feb 1950, A. H. Sturtevant (3♂, 3♀; USNM); Rio Hondo (33°55.9\'N, 118°10.5\'W), 15 Feb 1950, A. H. Sturtevant (2♂, 2♀; USNM). **Marin**: Lagunitas Creek (38°4.8\'N, 122°49.6\'W), 19 Oct 1947, W. W. Wirth (1♂; USNM); Muir Woods (37°53.6\'N, 122°34.4\'W), 7 Aug 1915, A. L. Melander (1♀; ANSP, USNM). **Riverside**: Riverside (33°57.2\'N, 117°23.8\'W), 3 Feb-5 May 1935, A. L. Melander (2♂, 1♀; ANSP, 5♂, 3♀; USNM). **San Mateo**: Corte Madera Creek (37°24.1\'N, 122°14.3\'W), 28 Oct 1953, P. H. Arnaud (1♂, 1♀; USNM); Stanford University, Jasper Ridge Biological Preserve (37°24\'N, 122°14.5\'W; 110 m), 14-17 Feb 2006, 2007, P. H. Arnaud, Jr. & M. M. Arnaud (29♂, 20♀; USNM); Searsville Lake (37°24.3\'N, 122°14.3\'W), 5 May 1953, P. H. Arnaud (1♂, 3♀; USNM). **Santa Clara**: Stevens Creek County Park (37°17.1\'N, 122°4.6\'W), 13 Mar 1976, L. Bezark (1♂; USNM). **Siskiyou**: Gasquet (41°50.7\'N, 123°58.2\'W), 18 Sep 1934, A. L. Melander (1♀; ANSP). Oregon. **Benton**: Cary's Grove (44°22.6\'N, 123°36.1\'W), 2 Sep 1974, W. N. Mathis (2♂, 3♀; USNM); Corvallis (44°33.9\'N, 123°15.7\'W), 1 Aug 1935, K. Gray, (1♂; USNM); Philomath (1.6 km SW; 44°31.8\'N, 123°22.9\'W), 29 May 1972, W. N. Mathis (1♀; USNM); Rock Creek (6.4 km SW Philomath; 44°30.1\'N, 123°26.2\'W), 29 May 1972, W. N. Mathis (2♀; USNM). **Douglas**: Elkton (43°38.25\'N, 123°34.1\'W), 28 Jul 1951, A. H. Sturtevant (1♀; USNM). **Lane**: Burp Hollow at Long Tom River (44°09.5\'N, 123°25.4\'W), 30 Jun 1988, R. Danielsson (3♂; MZLU). **Lincoln**: Waldport (10.4 km E; 44°25.6\'N, 123°56.25\'W), 27 May 1972, W. N. Mathis (4♂, 5♀; USNM). **Polk**: Helmick State Park (44°46.9\'N, 123°14.2\'W), 20 March 1972, W. N. Mathis (1♂; USNM). **Washington**: Hillsboro (45°31.3\'N, 122°59.3\'W), 9 Apr 1936, K. Gray, J. Shuh (1♀; USNM). **Yamhill**: Carlton (45°17.6\'N, 123°10.65\'W), 31 Jul 1951, A. H. Sturtevant (1♂, 2♀; USNM).
Washington. **Pierce**: Dupont (5 km WSW; 47°03.8\'N, 122°41.7\'W), 13 Apr 1971, W. N. Mathis (3♀; USNM); Gig Harbor (47°19.8\'N, 122°34.8\'W), 17 May-9 Jun 1971, W. N. Mathis (18♂, 3♀; USNM).
MEXICO. **Puebla**: Puebla (9.6 km SW; 18°58.4\'N, 98°16.9\'W), 2 Jul 1953, University of Kansas Mexican Expedition (1♂; USNM).
##### Distribution
(Fig. [111](#F39){ref-type="fig"}). Nearctic: United States (California, New Mexico, Oregon, Washington). Neotropical: Mexico (Puebla).
{#F39}
##### Remarks.
Although similar and apparently closely related to *Lamproclasiopa aliceae*, this species is distinguished from that species and other congeners by the absence of a presutural supra-alar seta; the shiny black katepisternum, especially the anterior half; the shiny black anteroventral portion of the anepisternum; and the presence of four to five stout, peg-like setae on the apical third of the forefemur along the posteroventral margin. The shape of structures of the male terminalia also distinguish this species from *Lamproclasiopa aliceae*. So far as we know, however, the distribution of this species and of *Lamproclasiopa aliceae* do not overlap, with this species only known from the west coast of North America and Puebla in Mexico. The non-overlapping distributions of these two species may be a function of sample error, however.
### The *puella* group (*Lamproclasiopa aracataca*, *Lamproclasiopa bisetulosa*, *Lamproclasiopa caligosa*, *Lamproclasiopa curva*, *Lamproclasiopa fumipennis*, *Lamproclasiopa puella*) {#SECID0E1UEK}
**Diagnosis.** Head, thorax, and abdomen variable, either generally shiny black or with extensive surfaces sparsely to densely microtomentose. *Head*: Frons and face generally unicolorous; gena moderately high (gena-to-eye ratio less than 0.11--0.19); genal/postgenal margin rounded. *Thorax*: Presutural supra-alar seta well developed; katepisternum either thinly microtomentose, generally appearing dull, not shiny, or especially anterior half, and anteroventral portion of anepisternum shiny black. Wing generally hyaline to very faintly infumate; vein R~2+3~ curved gently apically, not angulate subapically nor bearing a subapical stump vein. Forefemur with posteroventral setae slender, not stout and peg-like.
**Remarks.** This is the largest species group with six included species and its recognition is based on homoplasious characters. Thus, the group may be artificial. The included species are quite similar, however, and are the bases for our recognition and diagnosis of the group.
#### Lamproclasiopa aracataca
Animalia
Diptera
Ephydridae
(Cresson)
[Figs 81](#F28){ref-type="fig"} [, 112--113](#F40){ref-type="fig"} [, 114--117](#F41){ref-type="fig"}
1. Discocerina aracataca[@B9]: 5.
2. Discocerina (Lamproclasiopa) aracataca. [@B12]: 149 \[review\]. [@B34]: 7 \[Neotropical catalog\]. [@B24]: 168 \[world catalog\].
3. Lamproclasiopa aracataca. [@B36]: 36 \[generic combination\].
##### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Small to moderately small shore flies, body length 1.80-2.35 mm. *Head*: Frons dull, anterior margin yellowish orange in some degree, posterior portion grayish black, concolorous with mesonotum, some specimens with frons entirely grayish black, without distinctly marked iridescent microtomentose stripes. Antenna mostly grayish black to black, only ventral margin of segments yellowish orange. Face nearly unicolorous, blackish gray, not distinctively marked; parafacial bare of ventroclinate setulae, generally dull, creamy white, contrasted with face. Gena moderately high, gena-to-eye ratio 0.17. *Thorax*: Mesonotum uniformly faintly grayish black, finely microtomentose, lacking stripes; presutural supra-alar seta well developed. Scutellum dorsally covered with fine, sparse setulae, sometimes almost bare. Wing completely hyaline, lacking pattern of spots; vein R~2+3~ with apical portion a continued extension of angle at merger with costa; costal vein ratio 0.50-0.53; M vein ratio 0.57-0.63. Forefemur with posteroventral setae slender, not stout and peg-like; femora and tibiae grayish black to black, apices of tibiae yellowish; tarsi entirely yellowish or with apical 1-2 tarsomeres darkened. *Abdomen*: Tergites more sparsely microtomentose than mesonotum, shinier black or brown, especially laterally and mostly of tergites 4 and 5. Male terminalia (Figs [114--117](#F41){ref-type="fig"}): Epandrium in posterior view (Fig. [114](#F41){ref-type="fig"}) roundly U-shaped, except for ventral gap, oval, widest a midheight, dorsal arch very narrow, gap at ventral margin widely and shallowly U-shaped with lateral margins becoming wider ventrally, each lateral arm widest ventrally with short, medial extension, almost touching opposite medial extension, ventral extension bearing numerous setulae loosely organized as a group; cercus hemispherical, tapered ventrally to narrowly rounded apex, more setulose dorsally; gonite in lateral view (Fig. [117](#F41){ref-type="fig"}) robustly rod-like, almost straight, wider toward hypandrium, in ventral view (Fig. [116](#F41){ref-type="fig"}) shallowly curved with extension toward aedeagal base tapered to a narrow apex, apex toward hypandrium widest, with a medial, blunt, short extension; aedeagus in lateral view (Fig. [117](#F41){ref-type="fig"}) elongate, narrowly triangular, tapered evenly to narrowly rounded apex, in ventral view (Fig. [116](#F41){ref-type="fig"}) as an elongate, very narrow, rod-like structure, widest sub-basally, thereafter tapered to apex, apex with a short nipple; phallapodeme in lateral view (Fig. [117](#F41){ref-type="fig"}) more or less triangular, with moderately long, narrow extensions toward aedeagal base and hypandrium, keel distinct, relatively narrow, somewhat blunt apically; hypandrium in lateral view (Fig. [117](#F41){ref-type="fig"}) generally narrow, rod-like, basal third obtusely angulate, narrowed, digitiform, apical 2/3 narrowly rectangular, in ventral view (Fig. [116](#F41){ref-type="fig"}) as a very broad, robust H with short arms, emarginate anteriorly and posteriorly, anterior emargination shallow, posterior emargination more deeply excavate, broadly and rounded U to V-shaped.
{#F40}
![*Lamproclasiopa aracataca* (Cresson). (Chile. Bío Bío: Santa Barbara) **114** epandrium and cerci, posterior view **115** same, lateral view **116** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **117** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g041){#F41}
##### Type material.
The holotype female of *Discocerina aracataca* Cresson is labeled "Colombia Ujhelyi/Aracataca 1912. II./825/Holo-TYPE Discocerina ARACATACA E. T. Cresson Jr \[red; "Discocerina ARACATACA" handwritten\]." The holotype is double mounted (minuten pin in a thin rectangular piece of foam), is in excellent condition, and is deposited in the ANSP (6541).
##### Type locality.
Colombia. Magdalena: Aracataca (10°35.6\'N, 74°12\'W).
##### Other specimens examined.
ARGENTINA. **Chubut**: El Hoyo (42°3.8\'S, 71°31.6\'W), 21 Jan 1965, A. Kovacks (1♀; AMNH). **Río Negro**: Correntoso (41°5.94\'S, 71°26.65\'W), Nov 1926, R. C. & E. Shannon (1♂, 1♀; USNM); Lago Nahuel Huapi (east end; 41°06.5\'S, 71°12.4\'W) (1♂; BMNH); Puerto Blest (41°2.6\'S, 71°49.6\'W), 2 Dec 1926, R. C. & E. Shannon (1♂, 1♀; USNM); San Carlos de Bariloche (41°06.5\'S, 71°12.4\'W) (1♂; BMNH); Nov 1926, R. C. & E. Shannon (4♂, 2♀; USNM).
CHILE. **Araucaína**: Angol (37°48\'S, 72°43\'W), 15 Set 1931, D. S. Bullock (1♂; USNM). **Aysen**: Puerto Puyuguapi (44°19.5\'S, 72°33.5\'W), Feb, 13 Out 1939, G. H. Schwable (1♀; USNM). **Bío Bío**: El Abanico (37°20\'S, 71°31W), 31 Dec 1950, A. E. Michelbacher, E. S. Ross (1♀; USNM); Nuble (40 km E San Carlos; 36°24.5\'S, 71°31.5\'W), 23 Dec 1950, A. E. Michelbacher, E. S. Ross (4♂, 2♀; USNM); Santa Barbara (25 km E; 37°29.3\'S, 72°4.1\'W; 350 m), 24 Jan 1978, W. N. Mathis (12♂, 1♀; USNM). **Coquimbo**: Incahuasi (29°13.6\'S, 71°0.7\'W), 30 Sep 1952, P. G. Kuschel (3♂; USNM); La Serena (50 km S; 30°21.25\'S, 71°15.1\'W), 1 Dec 1950, A. E. Michelbacher, E. S. Ross (1♀; USNM). **Lanquihue**: Casa Pangue (41°03\'S, 71° 52\'W), Dec 1926, R. C. & E. Shannon (1♂, 3♀; USNM); Castro (41°27.7\'S, 72°56.1\'W) (ex.; BMNH); Ensenada (41°12.6\'S, 72°32.3\'W) (ex.; BMNH); Los Riscos (41°13.7\'S, 72°44.7\'W), 14 Sep 1954, P. G. Kuschel (2♂, 1♀; USNM); Peulla (41°28\'S, 72°57.7\'W) (1♂; BMNH); Puerto Montt (41°28\'S, 72°56\'W), Dec 1926, R. C. & E. Shannon (2♂, 1♀; USNM); Puerto Varas (41°18.6\'S, 72°59.6\'W) (2♂; BMNH); Dec 1926, R. C. & E. Shannon (2♂, 1♀; USNM). **Los Lagos**: Chiloé Island, Chepu (on seashore; 42°5\'S, 73°59.65\'W), Oct 1958, G. Kuschel (11♂, 3♀; USNM); Chiloé Island, Ancud (41°52\'S, 73°50\'W), 20--28 Jan, 1952, G. Kuschel. (1♀; AMNH). **Los Rios**: Valdivia (25 km. N; 39°35.56\'S, 73°14.55\'W), 26 Jan 1978, W. N. Mathis (2♀; USNM). **Maule**: Cajon de Rio Claro, (S. E. Los Quenes; 34°59.9\'S, 70°49\'W; 1000--1200m), 9 Oct 1966, E. I. Schlinger (1♂; USNM). **O'Higgins**: Río Claro (5 km N Rengo; 34°24\'S, 70°52\'W; 300 m), 23 Jan 1978, W. N. Mathis (16♂, 2♀; USNM). **Osorno**: Anticura (1 km W; 40°39\'S, 72°10\'W; 430 m), 1--6 Feb 1978, W. N. Mathis (26♂, 23♀; USNM); Anticura (4 km W; 37°40\'S, 72°01\'W; 400 m), 3 Feb 1978, W. N. Mathis (1♀; USNM); Anticura (6 km W; 37°40\'S, 72°01\'W; 400 m), 3 Feb 1978, W. N. Mathis (1♂, 1♀; USNM); Río El Gringo (40°41\'S, 71°59\'W; 1015 m), 13 Feb 1978, W. N. Mathis (1♂; USNM); Lago Puyehue (SE shore; 40°45\'S, 72°25.2\'W), 6--10 Feb 1978, W. N. Mathis (9♂, 6♀; USNM); Lago Rupanco, El Encanto (40°49\'S, 72°28\'W), 6 Feb 1978, W. N. Mathis (1♀; USNM); Laguna El Espejo (40°44.5\'S, 72°19.67\'W), 7 Feb 1978, W. N. Mathis (1♀; USNM); Laguna El Pato (40°40.6\'S, 72°0.2\'W; 1100 m), 13 Feb 1978, W. N. Mathis (3♀; USNM); Laguna El Toro (40°45.2\'S, 72°18.5\'W; 780 m), 8 Feb 1978, W. N. Mathis (3♂, 3♀; USNM); Pucatrihue (40°32.6\'S, 73°43.1\'W), 27--30 Jan 1978, W. N. Mathis (7♂, 9♀; USNM); Termas de Aguas Calientes (1 km SE; 40°41\'S, 72°21\'W; 530 m), 7--8 Feb 1978, W. N. Mathis (5♂, 5♀; USNM). **Santiago**: El Alfalfal (33°30.1\'S, 70°11.7\'W; 1320 m), 22 Jan 1978, W. N. Mathis (13♂, 4♀; USNM); Quebrada de la Plata, Rinconada, Maipu (33°31\'S, 70°47\'W; 510 m), 16 Aug 1966, M. E. Irwin (2♀; USNM); Río Maipo (7 km E San José de Maipo; 33°35.8\'S, 70°22.1\'W; 1065 m), 22 Jan 1978, W. N. Mathis (1♂; USNM). **Talca**: Río Lircay (11 km N Talca; 35°23\'S, 71°39\'W; 85 m), 23 Jan 1978, W. N. Mathis (1♀; USNM). **Valparaíso**: Marga Marga (road to Colliguay; 33° 5.6\'S, 71°12.8\'W), 14--15 Mar 1964, L. E. Peña (2♂, 2♀; USNM).
PERU. **Junin**: Tarma (11°25\'S, 75°41.2\'W; 3000 m). 11 Jul 1965, P. & B. Wygodzinsky (2♀; AMNH). **Lima**: San Jeronimo de Surco (11°53.1\'S, 76°26.4\'W; 1900--2100 m), 17 Aug 1965, P. & B. Wygodzinsky (2♂, 3♀; AMNH).
##### Distribution
(Fig. [81](#F28){ref-type="fig"}). Neotropical: Argentina (Chubut, Río Negro), Chile (Aracunaína; Aysen, Bío Bío, Coquimbo, Lanquihue, Los Lagos, Los Rios, Maule, O'Higgins, Osorno, Santiago, Talca, Valparaíso), Colombia (Magdalena), Peru (Junin, Lima).
##### Remarks.
This species is very similar and apparently closely related to *Lamproclasiopa puella* and is difficult to distinguish from that species using external characters. The diagnostic characters presented in original descriptions (anterior margin of frons yellowish orange, antenna mostly yellowish than grayish black, in opposition to *Lamproclasiopa puella*) are inconsistent, and specimens of *Lamproclasiopa puella* could be identified as *Lamproclasiopa aracataca* and vice versa using them. We propose a more reliable character: Scutellar disc covered with fine, sparse setulae, sometimes appearing almost bare. The more definitive diagnostic characters are the shapes of structures of the male terminalia, especially the wide ventral apices of the epandrium, the phallapodeme that has an extended keel, and the gonite in ventral view that is nearly truncate basally. The shape of the hypandrium in ventral view is very similar to that of *Lamproclasiopa puella* with posterior arms that are less flared.
#### Lamproclasiopa bisetulosa
Animalia
Diptera
Ephydridae
(Cresson)
[Figs 118--119](#F42){ref-type="fig"} [, 120--123](#F43){ref-type="fig"} [, 139](#F49){ref-type="fig"}
1. Ditrichophora bisetulosa[@B8]: 7.
2. Discocerina (Basila) bisetulosa. [@B12]: 148 \[generic combination\]. [@B34]: 7 \[Neotropical catalog\]. [@B16]: 24 \[list, Argentina\]. -[@B17]: 13 \[Argentina catalog\]. [@B24]: 165 \[world catalog\].
3. Lamproclasiopa bisetulosa. [@B36]: 39 \[generic combination\].
##### Diagnosis.
This species can be distinguished from congeners by the following combination of characters: Small to moderately small shore flies, body length 1.45--2.10 mm. *Head*: Frons with golden tan to slightly darker microtomentum, parafrons with slightly thinner investment of microtomentum; mesofrons evident by slight lateral lines. Antenna yellow; basal flagellomere with darker dorsal margin. Face completely and more or less uniformly silvery white microtomentose, more thinly microtomentose ventrally except for extreme ventral margin, vertical lacking stripes; 2 prominent facial setae, dorsal seta at midheight, other seta near epistomal margin; parafacial thin, more densely silvery white microtomentose than face. Gena moderately high, gena-to-eye ratio 0.12. *Thorax*: Mesonotum with golden brown microtomentum, subshiny, although less dense than microtomentum of frons; presutural supra-alar seta well developed; pleural areas more sparsely microtomentose than mesonotum, blackish brown to black, becoming less microtomentose ventrally and posteriorly, subshiny to shiny. Wing completely hyaline, lacking darkened areas; costal vein ratio 0.47--0.60; M vein ratio 0.55--0.75. Forefemur with posteroventral setae slender, not stout and peg-like; femora and tibiae grayish black to black, apical 1/4 of tibiae yellowish; tarsi yellowish, apical 1--2 tarsomeres darkened. *Abdomen*: Tergites more sparsely microtomentose than mesonotum, shinier black, especially laterally and mostly of tergites 4 and 5. Male terminalia (Figs [120--123](#F43){ref-type="fig"}): Epandrium in posterior view (Fig. [120](#F43){ref-type="fig"}) roundly U-shaped, except for ventral gap, oval, widest a midheight, dorsal arch very narrow, gap at ventral margin widely and shallowly U-shaped with lateral margins becoming wider ventrally, each lateral arm widest ventrally with short, medial extension, almost touching opposite medial extension, ventral extension bearing numerous setulae loosely organized as a group; cercus hemispherical, tapered ventrally to narrowly rounded apex, more setulose dorsally; gonite in lateral view (Fig. [123](#F43){ref-type="fig"}) robustly rod-like, almost straight, wider toward hypandrium, in ventral view (Fig. [122](#F43){ref-type="fig"}) shallowly curved with extension toward aedeagal base tapered to a narrow apex, apex toward hypandrium widest, with a medial, blunt, short extension; aedeagus in lateral view (Fig. [123](#F43){ref-type="fig"}) elongate, narrowly triangular, tapered evenly to narrowly rounded apex, in ventral view (Fig. [122](#F43){ref-type="fig"}) as an elongate, very narrow, rod-like structure, widest sub-basally, thereafter tapered to apex, apex with a short nipple; phallapodeme in lateral view (Fig. [123](#F43){ref-type="fig"}) as an inverted Y, each arm digitiform, process toward aedeagal base longer than other 2, in ventral view (Fig. [122](#F43){ref-type="fig"}) narrowly rectangular, robustly rod-like with shallow indentations toward hypandrium, keel digitiform; hypandrium in lateral view (Fig. [123](#F43){ref-type="fig"}) generally narrow, rod-like, essentially straight, basal third more thinly developed than anterior half, not obtusely angulate, in ventral view (Fig. [122](#F43){ref-type="fig"}) as a very broad, robust H with long posterior arms, lateral margins conspicuously sinuous, anterior emargination V-shaped, posterior emargination deep, broadly U-shaped.
{#F42}
![*Lamproclasiopa bisetulosa* (Cresson). (Argentina. Buenos Aires: José C. Paz) **120** epandrium and cerci, posterior view **121** same, lateral view **122** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **123** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g043){#F43}
##### Type material.
The holotype male of *Ditrichophora bisetulosa* Cresson is labeled "Paraguay Friebrig/S\[an\].Bernardino 1907. XI-/TYPE Ditrichophora BISETULOSA E. T. Cresson, Jr. \[red; "Ditrichophora BISETULOSA" handwritten\]." The holotype is double mounted (minuten pin in a thin rectangular piece of fine foam), is in good condition (some setulae missing or displaced), and is deposited in the ANSP (6574)\].
##### Type locality.
Paraguay. Cordillera: San Bernardino (25°18.8\'S, 57°18\'W).
##### Other specimens examined.
ARGENTINA. **Buenos Aires**: Buenos Aires (34°36\'S, 58°22.9\'W), 21 Oct 1926, F. & M. Edwards (1♂; USNM); San Isidro (34°29.6\'S, 58°32.6\'W), 2 Sep 1927, R. C. Shannon (1♀; ANSP. 1♂, 1♀; USNM); José C. Paz (34°31\'N, 58° 46\'W), 24 Aug 1939, A. Ogloblin (2♂; USNM). **Misiones**: Santa Ana (27°22.1\'S, 55°34.9\'W), 9 Dec 1949, H. Aesel (1♂; USNM).
URUGUAY. **Montevideo**: Montevideo (34°53.3\'S, 56°11\'W), 15 Jan 1965, E. F. Legnef (2♂; USNM).
##### Distribution
(Fig. [139](#F49){ref-type="fig"}). Neotropical: Argentina (Buenos Aires, Misiones), Paraguay (Cordillera), Uruguay (Montevideo).
##### Remarks.
This species is very similar to *Lamproclasiopa aracataca* externally and in the shape of structures of the male terminalia. These similarities indicate that these two species are closely related. The differences, although seemingly slight, are consistent, and are the basis for our continued recognition of this species. This species is distinguished from *Lamproclasiopa aracataca* by being slightly shinier externally and by the shape of structures of the male terminalia: the hypandrium has a less well-developed base, and the phallapodeme has a narrow keel.
#### Lamproclasiopa caligosa sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/EB6EA466-B799-4FC2-A080-25CFD27494EF
[Figs 124--127](#F44){ref-type="fig"} [, 139](#F49){ref-type="fig"}
1. Lamproclasiopa puellaof authors, not Cresson (misidentification). [@B36]: 41 \[illustration of male terminalia\].
##### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Small to moderately small shore flies, body length 1.80-2.80 mm. *Head*: Frons dull, uniformly grayish black concolorous with mesonotum, some specimens with anterior margin yellowish orange, except for ocellar triangle and fronto-orbital stripe slightly grayer, without distinctly marked iridescent microtomentose stripes, some specimens with anterior margin faintly reddish orange. Antenna mostly grayish black to black, only ventral margin of segments yellowish orange. Face nearly unicolorous, grayish black, not distinctively marked; parafacial bare of ventroclinate setulae, generally dull, creamy white anteriorly, grayish black ventrally, similar to facial color. Gena moderately high, gena-to-eye ratio 0.15-0.18. *Thorax*: Mesonotum uniformly faintly grayish to brownish black, finely microtomentose, faintly subshiny, lacking stripes; presutural supra-alar seta well developed. Scutellum dorsally covered with strong setulae. Wing completely hyaline to faintly infuscate, lacking pattern of spots; vein R~2+3~ with apical portion a continued extension of angle at merger with costa; costal vein ratio 0.45-0.55; M vein ratio 0.54-0.59. Forefemur with posteroventral setae slender, not stout and peg-like; femora and tibiae grayish black to black, apices of tibiae yellowish; tarsi entirely yellowish or with apical 1-2 tarsomeres darkened. *Abdomen*: Tergites more sparsely microtomentose than mesonotum, shinier black or brown, especially laterally and mostly of tergites 4 and 5. Male terminalia (Figs [124--127](#F44){ref-type="fig"}): Epandrium in posterior view (Fig. [124](#F44){ref-type="fig"}) generally oval, as high as wide, dorsal portion thin, each lateral arm gradually becoming wider ventrally, widest subapically, apex tapered, shallowly recurved, rounded pointed, ventral half with slightly increased number of setulae; cerci in posterior view (Fig. [124](#F44){ref-type="fig"}) elongate, thin, generally shallowly arched, ventral half generally tapered to acute point, slightly curved, setulose on dorsal 2/3; gonite in lateral view (Fig. [127](#F44){ref-type="fig"}) rod-like, shallowly arched, apical half toward aedeagal base digitiform, almost parallel sided, basal half wider than apical half, apex truncate with tiny, narrow emargination, basolateral margin irregularly serrate, in ventral view (Fig. [126](#F44){ref-type="fig"}) irregularly clavate apical half narrow, parallel sided, basal half becoming wider with subapical notch, thereafter thumb-like; aedeagus in lateral view (Fig. [127](#F44){ref-type="fig"}) narrowly funnel-like, wider basally, apical 2/3 tapered to acutely pointed and curved apex, in ventral view (Fig. [126](#F44){ref-type="fig"}) elongate, thin, tapered very gradually from base to subapex, apical portion more abruptly tapered to pointed apex; phallapodeme in lateral view (Fig. [127](#F44){ref-type="fig"}) shallowly L shaped, each arm tapered to narrowed apex, short arm toward hypandrium 1/3 length of longer and wider than dorsal arm toward aedeagal base, shallow keel at vertex, slightly extended, in ventral view (Fig. [126](#F44){ref-type="fig"}) as an inverted bottle with basal 2/3 very shallowly arched, nearly parallel sided, neck robust, slightly flared apically, apex with medial, short, narrow emargination; hypandrium in lateral view (Fig. [127](#F44){ref-type="fig"}) irregularly rod-like, sinuous, both apices tapered, posterior apex narrowly digitiform, anterior apex tapered to acute point, in ventral view (Fig. [126](#F44){ref-type="fig"}) with anterior half robustly developed, more or less quadrate, anterolateral corners slightly extended anteriorly, anterior margin shallowly emarginate, posterior extensions elongate, slightly tapered, directed posterolaterally, posterior margin deeply emarginate, widely U-shaped.
![*Lamproclasiopa caligosa* sp. n. (Chile. Osorno: Anticura) **124** epandrium and cerci, posterior view **125** same, lateral view **126** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **127** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g044){#F44}
##### Type material.
The holotype male of *Lamproclasiopa caligosa* is labeled "CHILE. Osorno: Anticura (1 km W; 40°39\'S, 72°10\'W; 430 m), 5 Feb 1978\[,\], W. N. Mathis/HOLOTYPE ♂ *Lamproclasiopa caligosa* Costa, Mathis & Marinoni USNM \[red\]." The holotype is double mounted (minuten pin in a block of plastic), is in excellent condition, and is deposited in USNM. Five paratypes (5♂, DZUP, USNM) bear the same label data as the holotype. A male paratype (USNM) is as follows: CHILE. Bío Bío: Santa Bárbara (25 km E; 37°44.4\'S, 71°47.9\'W; 350 m), 24 Jan 1978, W. N. Mathis.
##### Other specimens examined.
CHILE. **Bío Bío**: Santa Barbara (25 km E; 37°29.3\'S, 72°4.1\'W; 350 m), 24 Jan 1978, W. N. Mathis (1♀; USNM). **Osorno**: Anticura (1 km W; 40°39\'S, 72°10\'W; 430 m), 1-6 Feb 1978, W. N. Mathis (19♂, 26♀; USNM); Laguna El Toro (40°45.2\'S, 72°18.5\'W; 780 m), 8 Feb 1978, W. N. Mathis (2♂; USNM); Termas de Aguas Calientes (1 km SE; 40°41\'S, 72°21\'W; 530 m), 7-8 Feb 1978, W. N. Mathis (1♂; USNM).
##### Type locality.
Chile. Osorno: Anticura (1 km W; 40°39\'S, 72°10\'W; 430 m).
##### Distribution
(Fig. [139](#F49){ref-type="fig"}). Neotropical: Chile (Bío Bío, Osorno).
##### Etymology.
The species epithet, *caligosa*, is of Latin derivation and means misty, obscure or uncertain, referring to the difficulty in distinguishing this species from congeners, especially *Lamproclasiopa puella*.
##### Remarks.
Externally, this species is very similar to *Lamproclasiopa puella*, leading to the confusion and misidentification of this species with *Lamproclasiopa puella* ([@B36]). The structures that [@B36] illustrated of the so-called *Lamproclasiopa puella* are actually of this species. This species is distinguished from *Lamproclasiopa puella* by the posterior hypandrial arms being more widely separated and more flared laterally, and the aedeagal base in lateral view is wider, almost bulbous, and has a more abrupt taper after the basal one-third, and the apex is curved anteriorly and acutely pointed.
#### Lamproclasiopa curva sp. n.
Animalia
Diptera
Ephydridae
http://www.zoobank.org/AA7B9380-0C01-4C56-92D0-7B7EBB231538
[Figs 128--131](#F45){ref-type="fig"} [, 139](#F49){ref-type="fig"}
##### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Small to moderately small shore flies, body length 1.80--2.35 mm. *Head*: Frons dull, anterior margin yellowish orange in some degree, posterior portion grayish black, concolorous with mesonotum, some specimens with frons entirely grayish black, without distinctly marked iridescent microtomentose stripes. Antenna mostly grayish black to black, only ventral margin of segments yellowish orange. Face nearly unicolorous, blackish gray, not distinctively marked; parafacial bare of ventroclinate setulae, generally dull, creamy white, contrasted with face. Gena moderately high, gena-to-eye ratio 0.11--0.18. *Thorax*: Mesonotum uniformly faintly grayish black, finely microtomentose, lacking stripes; presutural supra-alar seta well developed. Scutellum dorsally covered with fine, sparse setulae, sometimes almost bare. Wing completely hyaline, lacking pattern of spots; vein R~2+3~ with apical portion a continued extension of angle at merger with costa; costal vein ratio 0.43--0.45; M vein ratio 0.56--0.58. Forefemur with posteroventral setae slender, not stout and peg-like; femora and tibiae grayish black to black, apices of tibiae yellowish; tarsi entirely yellowish or with apical 1--2 tarsomeres darkened. *Abdomen*: Tergites more sparsely microtomentose than mesonotum, shinier black or brown, especially laterally and mostly of tergites 4 and 5. Male terminalia (Figs [128--131](#F45){ref-type="fig"}): Epandrium in posterior view (Fig. [128](#F45){ref-type="fig"}) generally oval, almost as wide as high, dorsal portion thinner, each lateral arm gradually becoming wider ventrally, ventral half with sides almost parallel sided, apex almost truncate, shallow arched, ventral 1/4 with slightly increased number of setulae, in lateral view (Fig. [129](#F45){ref-type="fig"}) higher than wide, dorsal half thinner and tapered to acute apex, ventral half becoming expanded, ventral margin shallowly rounded with short, anterior extension, extension tapered to anterior point; cerci in posterior view (Fig. [128](#F45){ref-type="fig"}) elongate, moderately thin, curvature very slight, almost straight, ventral half generally tapered to acute point, more setulose on dorsal half, in lateral view (Fig. [129](#F45){ref-type="fig"}) rounded posteriorly, tapered from dorsum to ventral point; gonite in lateral view (Fig. [131](#F45){ref-type="fig"}) robust developed, higher than wide, with posterior margin extended, keel-like, margin serrate, anterior margin straight, in ventral view (Fig. [130](#F45){ref-type="fig"}) with irregularly rectangular base, dorsal extension narrow, parallel sided, curved medially; aedeagus in lateral view (Fig. [131](#F45){ref-type="fig"}) narrowly funnel-like, wider basally, apical 3/4 tapered to acutely pointed and straight apex, in ventral view (Fig. [130](#F45){ref-type="fig"}) elongate, thin, uniquely and asymmetrically curved laterally, forming banana-like structure, apical half more curved, tapered to bluntly rounded apex, basal margin bilobed with deep and rather narrow medial incision; phallapodeme in lateral view (Fig. [131](#F45){ref-type="fig"}) shallowly L shaped, each arm tapered to narrowed, short arm toward hypandrium 1/3 length of longer, digitiform process, keel at vertex of L, robustly developed, elongate, slightly longer than dorsal arm, in ventral view (Fig. [130](#F45){ref-type="fig"}) as an robustly rod-like with shallowly sinuous sides, ventral apex slightly flared and shallowly bilobed, dorsal apex truncate; hypandrium in lateral view (Fig. [131](#F45){ref-type="fig"}) irregularly rod-like, conspicuously sinuous, both apices tapered, in ventral view (Fig. [130](#F45){ref-type="fig"}) asymmetrical with one lateral half shorted than opposite lateral half, anterior margin truncate, posterior margin deeply emarginate, U-shaped with elongate posterolateral extensions.
![*Lamproclasiopa curva* sp. n. (Chile. Lanquihue: Casa Pangue) **128** epandrium and cerci, posterior view **129** same, lateral view **130** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **131** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g045){#F45}
##### Type material.
The holotype male of *Lamproclasiopa curva* is labeled "Casa Pangue (41°03\'S, 71° 52\'W), Llanquihue, Chile Dec1926, R&EShannon//HOLOTYPE ♂ *Lamproclasiopa curva* Costa, Mathis & Marinoni, USNM \[red\]." The holotype is double mounted (glued to a paper triangle) and is in good condition (abdomen removed, dissected, and in an attached microvial) and deposited in the USNM. One paratype (1♂; USNM) bears the same label data as the holotype. Other paratypes are as follows: CHILE: **Los Lagos**: Chiloé Island, Chepu (on seashore; 42°5\'S, 73°59.65\'W), Oct 1958, G. Kuschel (4♂; USNM).
**Other specimen examined.** CHILE. **Malleco**: Angol (37°48\'S, 72°43\'W), 18 Oct 1931, D. S. Bullock (1♂; USNM).
##### Type locality.
Chile. Lanquihue: Casa Pangue (41°03\'S, 71° 52\'W).
##### Distribution
(Fig. [139](#F49){ref-type="fig"}). Neotropical: Chile (Lanquihue, Los Lagos, Malleco).
##### Etymology.
The species epithet, *curva*, is of Latin derivation and means curved, bent, or arched, referring to the curved aedeagus of this species.
##### Remarks.
Externally, this species is very similar to *Lamproclasiopa aracataca* and *Lamproclasiopa puella*, and we primarily rely on structures of the male terminalia to distinguish between these three species. The most obvious distinguishing characters are the asymmetry of the aedeagus and hypandrium in ventral view. The curved aedeagus is the basis for this species name. Other distinguishing characters of this species are the extended, narrowly rectangular keel of the phallapodeme and the gonal width with serrations along some of its posterior margin.
#### Lamproclasiopa fumipennis
Animalia
Diptera
Ephydridae
(Wirth)
[Figs 132](#F46){ref-type="fig"} [, 139](#F49){ref-type="fig"}
1. Discocerina (Basila) fumipennis[@B32]: 53; [@B34]: 7 \[Neotropical catalog\]. [@B24]: 165 \[world catalog\].
2. Lamproclasiopa fumipennis. [@B36]: 39 \[generic combination\].
##### Diagnosis
(based on Wirth's original description). This species is distinguished from other congeners by the following combination of characters: Moderately small shore flies, body length about 3.00 mm; generally subshiny, blackish; sides of body with more or less dull brown microtomentum. *Head*: 1.2× broader than high. Frons 1.4× as broad as long; frons and occiput densely brown microtomentose; 1 pair of proclinate and 1 pair of reclinate fronto-orbitals, ocellar setae more widely separated than posterior ocelli, situated at a level about midway between bases of fronto-orbitals and anterior ocellus; pseudopostocellar setae 1/2 length of ocellar setae; medial and lateral vertical seta well developed. Basal flagellomere and palpus yellowish brown; arista with 5 dorsal rays. Face, parafacials, and gena gray, microtomentose; 2 pairs of strong facial setae; a row of very fine setulae at each parafacial suture; parafacial narrow, bare. *Thorax*: Mesonotal and discal setulae numerous and unordered; notopleuron and anepisternum with sparse setulae. A pair of strong humeral setae; notopleural setae strong, both pairs located near notopleural suture, anterior pair slightly closer to posterior pair than to humeral setae; presutural setae, supra-alar setae and prescutellar setae well developed; a somewhat weaker pair of postalar setae; lateral and apical pairs of scutellar setae each about as long as prescutellar setae. Wing densely brown infuscate (Fig. [132](#F46){ref-type="fig"}), veins blackish; costal section II 2.2× as long as section III; apex of vein R~3+4~ not noticeably curved into costa. Halter with yellow knob. Knees narrowly pale brownish; basal 2 tarsomeres yellowish, apical 3 brown; setae and setulae of legs and abdomen rather strong; no flexor armature on femora. *Abdomen*: Tergites more sparsely microtomentose than mesonotum, shinier black or brown, especially laterally and mostly of tergites 4 and 5. Male terminalia: Epandrium in posterior view generally oval, higher than wide, dorsal portion thin, gradually becoming wider ventrally, widest subapically, apex tapered, rounded pointed, apex and dorsal half bearing more setulae, in lateral view with dorsal half almost parallel sided, ventral portion expanded, with rounded ventral margin and shallow, anterior point subapically; cerci in posterior view elongate, thin, ventral half tapered to acute point, slightly curved, setulose on dorsal half, in lateral view elongate, thin, dorsal half wider than ventral portion, tapered toward ventral apex; in about as wide as long, narrower dorsally and ventrally, widest at midheight, each lateral arm widest ventrally, ventral margin mostly evenly rounded, with a shallow medioventral extension, more or less evenly setulose along length; cercus narrowly hemispherical, slightly curved, gradually tapered toward ventral apex; gonite in lateral view rod-like, shallowly arched, apices tapered, with broad, short process beyond midlength, in ventral view irregularly V-shaped, arm toward aedeagal base much longer, tapered, more basal arm short, digitiform; aedeagus in lateral view narrowly funnel-like, wider basally, apical half tapered to narrow, parallel-sided extension, in ventral view elongate, thin, tapered very gradually from base to apex, narrow apex rounded, base arched; phallapodeme in lateral view L shaped, each arm tapered to narrowed apex, keel at angle, slightly extended, in ventral view narrowly spindle shaped, apices expanded, end toward hypandrium bifurcate; hypandrium in lateral view rod-like, shallowly arched, in ventral view with anterior half robustly developed, more or less quadrate, anterolateral corners rounded, anterior margin shallowly emarginated, posterior extensions elongate, tapered, posterior margin deeply emarginate, V-shaped.
{#F46}
##### Type material.
The holotype female of Discocerina (Basila) fumipennis Wirth is labeled "Chile. Juan Fernández Islands, Masatierra, Plazoleta del Yunque (33°38.8\'S, 78°50.1\'W); HT ♀, UMCE\]." Holotype female, allotype male, Masatierra, Plazoleta del Yunque, 200 meters, 9 January 1952, (in dense forest).
##### Other specimens examined.
CHILE. **Valparaíso**: Juan Fernández Islands, Robinson Crusoe Island (Plazoleta, trail sweep; 33°38.8\'S, 78°50.3\'W), 1--8 Jan 1993, S. A. Marshall (17♂, 2♀; DEBU, USNM); (quebrada S side Mirador, fern forest, 20 pans; 33°38.7\'S, 78°51.1\'W), 1--10 Jan 1993, S. A. Marshall (1♂; DEBU); (El Yunque Trail; 33°39\'S, 78°50.6\'W), 9 Jan 1993, S. A. Marshall (4♂; USNM); (English Bay; 33°37.2\'S 78°50.9\'W), 5 Jan 1993, S. A. Marshall (1♂, 1♀; DEBU).
##### Type locality.
Chile. Valparaíso: Juan Fernández Islands, Masatierra, Plazoleta del Yunque (33°38.8\'S, 78°50.1\'W; 200 m; dense forest).
##### Distribution
(Fig. [139](#F49){ref-type="fig"}). Neotropical: Chile (Valparaíso: Juan Fernández Islands.)
##### Remarks.
Our diagnosis of this species is partially based on Wirth's original description (1955), as we have not been given access to the holotype, which is a female, or to the male paratype, which Wirth designated as the allotype.
When [@B32] described this species, he wrote the following as a comparative diagnosis. This species is (p. 54) "Most closely related to Discocerina (Basila) puella (Cresson) from Chile, but that species has the wings hyaline, the body much duller, microtomentose above and the tarsi entirely yellow. Discocerina (Basila) polita (Edwards) from Chile is a polished, metallic black species with hyaline wings."
We concur with Wirth that this species is indeed closely related to *Lamproclasiopa puella* and suggest, further, that these two "species" may be conspecific. Externally there are some differences, as Wirth noted and as we have confirmed herein (see key and respective diagnoses). Moreover, we have observed that these external differences, although slight, are consistent. The populations are separable. The shapes of structures of the male terminalia, however, are essentially the same for *Lamproclasiopa puella* and the darkened specimens from the Juan Fernández Islands. Thus, while we have observed that a level of genetic diversification has occurred in the island populations, the question of whether it is sufficient to represent speciation remains an open question---the dilemma of diversified, allopatric populations. For the present, we are continuing to recognize the populations from the islands as a separate species, especially as we have not been able to study the type series.
Our records indicate that this is the only congener known to occur on the Juan Fernández Islands.
#### Lamproclasiopa puella
Animalia
Diptera
Ephydridae
(Cresson)
[Figs 133--134](#F47){ref-type="fig"} [, 135--138](#F48){ref-type="fig"} [, 139](#F49){ref-type="fig"}
1. Ditrichophora puella[@B6]: 91.
2. Discocerina (Basila) puella. [@B12]: 148 \[generic combination\]. [@B34]: 7 \[Neotropical catalog\]. [@B16]: 24 \[list, Argentina\]. --[@B17]: 13 \[Argentina catalog\]. [@B24]: 165 \[world catalog\].
3. Lamproclasiopa puella. [@B36]: 39 \[generic combination\].
##### Diagnosis.
This species is distinguished from congeners by the following combination of characters: Small to moderately small shore flies, body length 1.80--2.80 mm. *Head*: Frons dull, uniformly grayish black concolorous with mesonotum, some specimens with anterior margin yellowish orange, except for ocellar triangle and fronto-orbital stripe slightly grayer, without distinctly marked iridescent microtomentose stripes, some specimens with anterior margin faintly reddish orange. Antenna mostly grayish black to black, only ventral margin of segments yellowish orange. Face nearly unicolorous, grayish black, not distinctively marked; parafacial bare of ventroclinate setulae, generally dull, creamy white anteriorly, grayish black ventrally, similar to facial color. Gena moderately high, gena-to-eye ratio 0.16--0.19. *Thorax*: Mesonotum uniformly faintly grayish to brownish black, finely microtomentose, faintly subshiny, lacking stripes; presutural supra-alar seta well developed. Scutellum dorsally covered with strong setulae. Wing completely hyaline to faintly infuscate, lacking pattern of spots; vein R~2+3~ with apical portion extended at same angle to costa; costal vein ratio 0.42--0.46; M vein ratio 0.55--0.60. Forefemur with posteroventral setae slender, not stout and peg-like; femora and tibiae grayish black to black, apices of tibiae yellowish; tarsi entirely yellowish or with apical 1--2 tarsomeres darkened. *Abdomen*: Tergites more sparsely microtomentose than mesonotum, shinier black or brown, especially laterally and mostly of tergites 4 and 5. Male terminalia (Figs [135--138](#F48){ref-type="fig"}): Epandrium in posterior view (Fig. [135](#F48){ref-type="fig"}) generally oval, higher than wide, dorsal portion thin, gradually becoming wider ventrally, widest subapically, apex tapered, rounded pointed, apex and dorsal half bearing more setulae, in lateral view (Fig. [136](#F48){ref-type="fig"}) with dorsal half almost parallel sided, ventral portion expanded, with rounded ventral margin and shallow, anterior point subapically; cerci in posterior view (Fig. [135](#F48){ref-type="fig"}) elongate, thin, ventral half tapered to acute point, slightly curved, setulose on dorsal half, in lateral view (Fig. [136](#F48){ref-type="fig"}) elongate, thin, dorsal half wider than ventral portion, tapered toward ventral apex; in about as wide as long, narrower dorsally and ventrally, widest at midheight, each lateral arm widest ventrally, ventral margin mostly evenly rounded, with a shallow medioventral extension, more or less evenly setulose along length; cercus narrowly hemispherical, slightly curved, gradually tapered toward ventral apex; gonite in lateral view (Fig. [138](#F48){ref-type="fig"}) rod-like, shallowly arched, apices tapered, with broad, short process beyond midlength, in ventral view (Fig. [137](#F48){ref-type="fig"}) irregularly V-shaped, arm toward aedeagal base much longer, tapered, more basal arm short, digitiform; aedeagus in lateral view (Fig. [138](#F48){ref-type="fig"}) narrowly funnel-like, wider basally, apical half tapered to narrow, parallel-sided extension, in ventral view (Fig. [137](#F48){ref-type="fig"}) elongate, thin, tapered very gradually from base to apex, narrow apex rounded, base arched; phallapodeme in lateral view (Fig. [138](#F48){ref-type="fig"}) L shaped, each arm tapered to narrowed apex, keel at angle, slightly extended, in ventral view (Fig. [137](#F48){ref-type="fig"}) narrowly spindle shaped, apices expanded, end toward hypandrium bifurcate; hypandrium in lateral view (Fig. [138](#F48){ref-type="fig"}) rod-like, shallowly arched, in ventral view (Fig. [138](#F48){ref-type="fig"}) with anterior half robustly developed, more or less quadrate, anterolateral corners rounded, anterior margin shallowly emarginated, posterior extensions elongate, tapered, posterior margin deeply emarginate, V-shaped.
{#F47}
![*Lamproclasiopa puella* (Cresson). (Chile. Osorno: Anticura) **135** epandrium and cerci, posterior view **136** same, lateral view **137** internal structures of male terminalia (aedeagus \[shaded\], phallapodeme, gonite, hypandrium), ventral view **138** same, lateral view. Scale bar = 0.1 mm.](zookeys-631-001-g048){#F48}
##### Type material.
The holotype male of *Ditrichophora puella* Cresson is labeled "HOLOTYPE/Casa Pangue 4--10.xii.1926./S.Chile: Llanquihue Prov F.&M. Edwards. B.M.1927--63./Holo-TYPE *Ditrichophora puella* E. T. Cresson Jr./NHMUK010240992". The holotype is double mounted (glued to a plastic triangle) and is in good condition (head missing), and is deposited in BMNH.
##### Type locality.
Chile. Lanquihue: Casa Pangue (41°03\'S, 71°52\'W; 779 m).
##### Other specimens examined.
CHILE. **Aysen**: Puerto Puyuguapi (44°19.5\'S, 72°33.5\'W), Feb, 13 Out 1939, G. H. Schwable (1♀; USNM). **Lanquihue**: Los Riscos (41°13.7\'S, 72°44.7\'W), 14 Sep 1954, P. G. Kuschel (1♂; USNM); **Osorno**: Anticura (1 km W; 40°39\'S, 72°10\'W; 430 m), 1--6 Feb 1978, W. N. Mathis (11♂, 4♀; USNM); Lago Puyehue (SE shore; 40°45\'S, 72°25.2\'W), 6--10 Feb 1978, W. N. Mathis (3♂; USNM); Pucatrihue (40°32.6\'S, 73°43.1\'W), 27--30 Jan 1978, W. N. Mathis (4♂; USNM); Puyehue (20 km E.; 40°38.8\'S, 72°5.1\'W), 25 Jan 1951, A. E. Michelbacher, E. S. Ross (1♂; USNM); Termas de Aguas Calientes (1 km SE; 40°41\'S, 72°21\'W; 530 m), 7--8 Feb 1978, W. N. Mathis (13♂, 8♀; USNM); Volcan Puyehue (40°36.7\'S, 72°8.4\'W; 1400 m), 4 Feb 1978, W. N. Mathis (1♂; USNM).
##### Distribution
(Fig. [139](#F49){ref-type="fig"}). Neotropical: Chile (Aysen, Lanquihue, Osorno).
{#F49}
##### Remarks.
This species is challengingly similar to *Lamproclasiopa aracataca* and distinguishing between them is difficult. The diagnostic characters presented in the original descriptions (frons entirely grayish black, antenna mostly grayish black than orange, in opposition to *Lamproclasiopa aracataca*) are inconsistent, and specimens of *Lamproclasiopa puella* could easily be identified as *Lamproclasiopa aracataca* and vice versa. We dissected the male holotype to confirm the identity of *Lamproclasiopa puella*, and based on these characters we propose the more reliable, external character: scutellum covered with strong setulae. The shape of structures of the male terminalia also distinguish this species, especially the narrow aedeagus that is straight in ventral view and the less flared posterior hypandrial arms. We have studied specimens from Juan Fernández Islands and these specimens have wings slightly darker than specimens from the continent (Fig. [132](#F46){ref-type="fig"}). This corresponds to Wirth's description of *Lamproclasiopa fumipennis*, but terminalia structures clearly correspond to *Lamproclasiopa puella*. As we have not been given access to the holotype of *Lamproclasiopa fumipennis*, we decided not to propose this synonymy as yet.
Supplementary Material
======================
###### XML Treatment for Discocerinini
###### XML Treatment for Lamproclasiopa
###### XML Treatment for Lamproclasiopa laevior
###### XML Treatment for Lamproclasiopa brunnea
###### XML Treatment for Lamproclasiopa hendeli
###### XML Treatment for Lamproclasiopa triangularis
###### XML Treatment for Lamproclasiopa auritunica
###### XML Treatment for Lamproclasiopa lapaz
###### XML Treatment for Lamproclasiopa polita
###### XML Treatment for Lamproclasiopa ecuadoriensis
###### XML Treatment for Lamproclasiopa zerafael
###### XML Treatment for Lamproclasiopa balsamae
###### XML Treatment for Lamproclasiopa mancha
###### XML Treatment for Lamproclasiopa painteri
###### XML Treatment for Lamproclasiopa nana
###### XML Treatment for Lamproclasiopa furvitibia
###### XML Treatment for Lamproclasiopa xanthocera
###### XML Treatment for Lamproclasiopa aliceae
###### XML Treatment for Lamproclasiopa argentipicta
###### XML Treatment for Lamproclasiopa nadineae
###### XML Treatment for Lamproclasiopa aracataca
###### XML Treatment for Lamproclasiopa bisetulosa
###### XML Treatment for Lamproclasiopa caligosa
###### XML Treatment for Lamproclasiopa curva
###### XML Treatment for Lamproclasiopa fumipennis
###### XML Treatment for Lamproclasiopa puella
We gratefully acknowledge the assistance and cooperation of many organizations and individuals who contributed to the field work and production of this paper. Photographs of the specimens, especially the faces, were expertly taken with a Visionary Digital System. For reviewing an early draft of this paper we thank Anthony "Tony" G. Irwin and Tadeusz Zatwarnicki, and Jens-Hermann Stuke reviewed the final draft. We thank T. Zatwarnicki in particular for allowing us access to an early version of an important paper on the phylogeny of the tribe Discocerinini and usage of some of his illustrations of structures of the male terminalia ([@B36]). Marion Kotrba very kindly translated Hendel's description of *Lamproclasiopa facialis* and we gratefully thank her and acknowledge her time and efforts.
We thank the curators and collections managers who loaned collections or facilitated work in their museums: David A. Grimaldi (AMNH), Jon K. Gelhaus and Jason D. Weintraub (ANSP), José Albertino Rafael and Rosaly Ale-Rocha (INPA), and Carlos José Einicker Lamas (MZUSP). This study was supported by grants from CNPq, Conselho Nacional de Desenvolvimento Científico e Tecnológico -- Brazil (Process number 234167/2014--9) and CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior -- Brazil), including a recent field work in Brazil (December 2009-June 2010) that resulted in the vast majority of specimens from Brazil that were studied in this paper (CNPq, Visiting Researcher/Process number 401609/2009--0), which we gratefully acknowledge. The ZADBI project was funded by U.S. National Science Foundation grant DEB 1145890 to B. Brown and A. Borkent. We thank Dianne Mathis for helping with all aspects of the production of this paper, especially the field work in Brazil. We also thank A. Bernardo Carvalho and his lab (Elisa Carvalho, Monica Carvalho, Susana Vaz) for hosting us while conducting field work along the coast of São Paulo, and José Albertino Rafael and Rosaly Ale-Rocha (INPA) for hosting us while working at INPA and conducting field work in the general environs of Manaus.
01\. *Lamproclasiopa aliceae* sp. n. (United States. New Mexico. Grant: Silver City (Big Ditch; 32°46.4\'N, 108°16.5\'W; 1790 m)),
02\. *Lamproclasiopa aracataca* (Cresson)
03\. *Lamproclasiopa argentipicta* sp. n. (Costa Rica. San José. Zurquí de Moravia (10°2.8\'N, 84°0.6\'W))
04\. *Lamproclasiopa auritunica* sp. n. (Bolívia. Oruro: Paznã (S. of the town; 18°36.2\'S, 66°54.7\'W, 3750 m).)
05\. *Lamproclasiopa balsamae* (Cresson)
06\. *Lamproclasiopa bisetulosa* (Cresson)
07\. *Lamproclasiopa brunnea* sp. n. (Costa Rica. San José. Zurquí de Moravia (10°2.8\'N, 84°0.6\'W))
08\. *Lamproclasiopa caligosa* sp. n. (Chile. Osorno: Anticura (1 km W; 40°39\'S, 72°10\'W; 430 m))
09\. *Lamproclasiopa curva* (Chile. Los Lagos: Chiloé Island, Chepu (on seashore; 42°5\'S, 73°59.65\'W))
10\. *Lamproclasiopa ecuadoriensis* sp. n. (Ecuador. Orellana: Río Tiputini Biodiversity Station (0°38.2\'S, 76°8.9\'W))
11\. *Lamproclasiopa fumipennis* (Wirth)
12\. *Lamproclasiopa furvitibia* sp. n. (Costa Rica. San José. Zurquí de Moravia (10°2.8\'N, 84°0.6\'W))
13\. *Lamproclasiopa hendeli* (Wirth)
14\. *Lamproclasiopa laevior* (Cresson)
15\. *Lamproclasiopa lapaz* sp. n. (Bolívia. La Paz: La Paz (6 km NE; 16°25.7\'S, 68°04.3\'W; 4130m))
16\. *Lamproclasiopa mancha* sp. n. (Brazil. Paraná: Curitiba, Universidade Federal do Paraná, Reserva Biológica (25°26.9\'S, 49°14\'W; 915 m))
17\. *Lamproclasiopa nadineae* (Cresson)
18\. *Lamproclasiopa nana* (Williston)
19\. *Lamproclasiopa painteri* (Cresson)
20\. *Lamproclasiopa polita* (Edwards)
21\. *Lamproclasiopa puella* (Cresson)
22\. *Lamproclasiopa triangularis* sp. n. (Peru. Madre de Dios: Río Manu, Pakitza (11°56.6\'S, 71°16.9\'W; 250 m))
23\. *Lamproclasiopa xanthocera* sp. n. (Brazil. Paraná. Curitiba, Universidade Federal do Paraná, Reserva Biológica (25°26.9\'S, 49°14\'W; 915 m))
24\. *Lamproclasiopa zerafael* sp. n. (Brazil. Amazonas: Reserva Ducke (02°55.8\'S, 59°58.5\'W; 40 m))
[^1]: Academic editor: R. Meier
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
As an integrated component of computerized and flexible manufacturing systems, the Job-Shop Scheduling Problem (JSP) is encountered in many industrial contexts. The importance of this problem is two-fold. First, it has a wide-spread applicability in manufacturing, and second, despite its easy-to-state description, it is a notoriously difficult and intractable problem which provides an ideal framework to evaluate innovative algorithmic approaches. Successful approaches for this easy-to-state problem can be later modified to cope with hard-to-state scheduling circumstances.
Among many different procedures developed to cope with the JSP, those which employ local searches have the most robust and effective aspects. A local search differs from a systematic tree search in that systematic tree search expands a graph of partial solutions, whereas a local search explores a virtual graph connecting each complete solution to its neighboring complete solutions. The number of arcs in this virtual graph is affected by the neighborhood scheme employed by the local search, with the larger size of neighborhood leading to higher number of arcs and consequently larger or even impractical required computational times. That is why the endeavor of defining a proper neighborhood scheme highly determines the success of any local search algorithm.
Defining a proper neighborhood scheme for a local search is, however, involved with highly conflicting factors, in the sense that despite the fact that many neighborhood schemes seem to be only superficial variation of one another, they can easily demonstrate entirely different results. The reason of this phenomenon has been partly described by the notion of fitness landscape (Forrest and Mitchell [@CR19]), and it seems that successful neighborhood schemes have the capability of effectively managing a trade-off between computational time and the number of arcs in their virtual graphs.
In tackling the JSP, this paper presents a procedure that combines small and large neighborhood structures. The procedure, called SLENP (Small-Large Embedded Neighborhood Search), has four synergetic characteristics of (i) making use of a forward-backward construction method for generating initial solutions, (ii) employing a small neighborhood search, (iii) using a feedback-based mechanism in generating perturbation for improving the result of the small neighborhood search, and (iv) using a large-neighborhood search for improving the overall result of the combination of the small-neighborhood search and the perturbation mechanism employed.
The feedback process employed for igniting solution perturbation is based on memorizing the values of solutions generated, and is aimed at minimizing the chance of existing any redundancy in the search. The rationale behind the use of this feedback process in performing perturbation is that any perturbation contributes to exploration but spoils any exploitation aspect of the search and hence it should be performed by extreme care. Memory undeniably is a vital constituent of any successful search, and here it has been used for igniting perturbation to avoid the same area of the virtual graph to be visited repeatedly.
Mixing different neighborhood structures is one of the building blocks of the variable neighborhood search (Hansen and Mladenović [@CR24]), and it seems that combining small and large neighborhood schemes can have dual benefits. On the one hand, the poor decision made in a small neighborhood, which is the natural consequent of its limited scope, may be rectified, and on the other hand, because of the comparatively high quality of its initial solution, the employed large neighborhood search may require less computational effort in producing its final result.
The SLENP performs its large neighborhood search through a variant of the shifting bottleneck procedure which works both in the forwards and backward directions. By performing in the backward direction, the ordinary operations of the shifting bottleneck procedure are executed on an inverted network, called mirror network. The term mirror best reflects how, by reversing the precedence relations of the initial network, this modified network is created and why the overall solution based on this modified network can be mirrored to show a solution to the original network.
The outline of the paper is as follows. The next section starts with presenting the formulation of the job shop problem and providing a brief literature survey on the problem. Section Related works presents the related work. In Section The SLENP, the SLENP is discussed and a stepwise description is provided that describes how the procedure operates and clarifies how its different components interact with one another. Section Computational experiments presents the results of computational experiments. A summary of the results as well as the suggestions for future work are discussed in Section Concluding remarks.
Problem formulation {#Sec2}
===================
The JSP consists of *n* jobs and *m* machines, with each job having a specific processing order on the machines. A typical schedule for the JSP is the allocation of jobs to the time slots of the machines to minimize the makespan. In other words, each job is comprised of a sequence of *m* operations, each to be processed on a specified machine within a particular time. The goal is to minimize the finish time of the last activity completed subject to the constraint that, once started, an operation cannot be interrupted and should continue until it has been completed. It is worth noting that makespan is a regular criterion and any method capable of handing this criterion has the potential of being modified to handle other regular criteria like total weighted flow time, weighted tardiness, weighted sum of tardy jobs, and maximum tardiness (Mati et al. [@CR33]). The reverse is also true with the procedures handling other regular criteria, like weighted tardiness (Bülbül [@CR9]; Kreipl [@CR27]).
As an NP-hard problem (Lawler et al. [@CR28]), the JSP is a notoriously difficult and an intractable combinatorial optimization problem. An evidence for its intractability is that finding the optimal solution of a relatively small problem instance presented in (Fisher and Thompson [@CR18]), with the dimension of 10\*10, despite the focus of intensive research on it, remained unsolved for 26 years until it was solved by the exact procedure developed in (Carlier and Pinson [@CR12]). This celebrated instance, which in the literature is called ft10, is still used by many researchers to test their algorithms. The other exact procedures that have been successfully applied to small instances have been developed in (Applegate and Cook [@CR2]) and (Brucker et al. [@CR8]). By exploring specific knowledge about the longest path in the disjunctive graph and employing the cutting-plane method for obtaining lower bounds, these methods are aimed at applying sophisticated inference rules to cut the enumeration tree in its early expansion phase, with both of the methods able to solve the ft10 to optimality within several minutes.
One of the most effective formulations of the JSP is performed by the use of disjunctive graphs (Brucker et al. [@CR8]). Figure [1](#Fig1){ref-type="fig"} shows a sample problem and Figures [2](#Fig2){ref-type="fig"} and [3](#Fig3){ref-type="fig"} show two different fixation settings for the disjunctive arcs of the sample problem. As is seen, the makespan of the problem, as the longest path from the starting to the ending node, is different for Figures [2](#Fig2){ref-type="fig"} and [3](#Fig3){ref-type="fig"}. In effect, Figure [3](#Fig3){ref-type="fig"} shows an optimal fixation setting for the disjunctive arcs, which has led to the optimum makespan of 22.Figure 1**A sample 4-machine 4-job JSP problem in which the number of the required machine has been written in each circle (operation).**Figure 2**Fixing the disjunctive arcs of the sample problem leading to the makespan of 33.**Figure 3**Fixing the disjunctive arcs of the sample problem leading to the makespan of 22 (optimal solution).**
Since all operations executed on the same machine require a given order, the notion of feasible order plays a key role in such graphs, with machine *i* being associated with the order π~i~, which shows the permutation of jobs on that machine. An order, Π, which consists of {π~1~, π~2~,\....,π~m~} is feasible if it does not introduce any loop in the graph. Figure [4](#Fig4){ref-type="fig"} shows an infeasible fixation of disjunctive arcs and the resultant loop produced.Figure 4**An infeasible fixation of disjunctive arcs leading to a loop of 1-2-9-10-1.**
Denoting the completion time of the last completed job with *makespan*, the JSP can be simply formulated as follows:
Related works {#Sec3}
=============
Solution strategies presented for the JSP range from artificial bee colony optimization and hybrid genetic tabu searches (Banharnsakun et al. [@CR5]; Meeran and Morshed [@CR35]; Zhang et al. [@CR52]; Zhang et al. [@CR51]) through dynamic and linear programming (Gromicho et al. [@CR23]; Bülbül and Kaminsky [@CR10]) to path relinking and particle swarm optimization (Pongchairerks [@CR40]; Nasiri and Kianfar [@CR36]). In an overview of scheduling models presented in (Framinan et al. [@CR20]) several of these strategies have been examined. Non-exact solution strategies for the JSP, to which the method presented in this paper belongs, can be categorized into six different categories, namely (i) construction methods, (ii) local searches, (iii) metaheuristics, (iv) evolutionary algorithms, and (iv) hybrids. Interestingly, nearly all successful techniques in these categories model the JSP as a disjunctive graph. In each category, only those works have been discussed which have affected the SLENP.
Construction methods build a solution progressively, starting with a null schedule and expanding it gradually until a full schedule is obtained. In the process of creating a full schedule, a sequence of intermediate partial schedules is created, with each partial schedule expanding the previous partial schedule. In general, myopic decisions are the backbone of such methods in the progressive expansion of the intermediate partial schedules. Priority-based techniques are the oldest techniques classified as construction methods.
The method presented in (Giffler and Thompson [@CR21]) is one of the most effective methods in the category of construction methods. This method is able to accept a set of priorities and create either an active or a non-delay schedule through a dispatching mechanism which schedules eligible operations one at a time based on their priorities.
The second major construction method is the shifting bottleneck procedure (SBP) (Adams et al. [@CR1]), which decomposes the JSP into several, *m*, one machine problems and solves each problem to optimality with the Carlier's method. As an intricate algorithm, the SBP repeatedly redirects the search towards scheduling the machine which imposes the most severe constraint in the sense of increasing the objective function.
As the counterpart of gradient optimization in continuous spaces, local searches probe discrete spaces through the fundamental notion of *move* to find local optimal solutions. In effect, modern local searches are now the leading procedures in solving the JSP, and this is mainly due to effective neighborhood schemes developed in the last three decades. In general, local searches convert a complete solution to another complete solution through local changes.
A major point with local searches is that they cannot be effective unless they exploit the structure of the problem through a proper definition of a neighborhood structure and an effective mechanism for the manipulation of such a structure. Nearly all effective neighborhood structures for the JSP are based on the basic notion of critical path in a resolved disjunctive graph. In effect, each critical path represents the longest route through operations and its length is equal to the makespan. Major neighborhood structures for the JSP are as follows.
N1 neighborhood has been proposed in (Van Laarhoven et al. [@CR47]) and defines a move by interchanging two successive operations of the same machine on a critical path. The design of N1 has been made based on two principles: (i) changing the order of two non-critical operations cannot improve the solution and may only create a cycle in the disjunctive graph, and (ii) changing the order of two adjacent operations cannot create a cycle.
N2 neighborhood (Dell\'Amico and Trubian [@CR14]) can reverse more than one arc on the critical path. Assuming *i* and *j* are two consecutive operations of the same block and one of them is at an extreme point of the block, the predecessor of *i* and the successor of *j* can also be subject to reversal with their predecessors and successors, respectively.
In the N3 neighborhood (Dell\'Amico and Trubian [@CR14]), a sequence of three operations on the critical path can be reversed subject to the condition that such a reversal does not lead to any loop. As an extension of N1, N3 is not limited to the reversal of triplets and includes the interchange of a pair of operations as well.
In N4 neighborhood (Dell\'Amico and Trubian [@CR14]), each operation of a block can move to any location of the block, subject to creating no cycle. Unlike in the other three neighborhood schemes, which are based on adjacent interchanges of operations, in this neighborhood a shift is performed. Based on this shift, an operation jumps over several other operations in its corresponding block to the left or right. In effect, this neighborhood can be considered as an expansion of all other previous ones.
As a restricted version of both N1 and N4, N5 is a neighborhood scheme, developed in (Nowicki and Smutnicki [@CR37]), in which the first two or the last two operations of each block are interchanged. The only exceptions are the first and last blocks, in which only their last and first two operations are interchanged, respectively.
The rationale behind the development of N5 is that the size of N1 is large and includes a large percentage of moves that cannot lead to any improvement. These unfruitful moves are those which are involved with two internal operations in the corresponding block. The removal of these unfruitful moves out of N1 leads to the creation of N5, which includes a restricted collection of highly effective moves.
However, despite using such highly effective moves, a drawback with N5 neighborhood is that its corresponding search space is disconnected. This disconnection removes any guarantee for the existence of a path between an optimal solution and an arbitrary seed. In comparison with N4, N5 is involved with the reversal of only one disjunctive arc, and this makes its corresponding neighborhood considerably smaller.
In N6 neighborhood (Balas and Vazacopoulos [@CR4]), each operation of a block can move precisely after the last or before the first operation of the black, subject to creating no cycle. N6 is very close to N4, and the major difference existing between these two neighborhood schemes is that N4 allows each operation of the block to move to any other location of the block, subject to creating no cycle. This makes the size of N6 slightly smaller than that of N4. N6 has been extended to N6' in (Zhang et al. [@CR50]) by being allowed to move the first or the last operation of the block into the interior operation inside the block.
It is worth mentioning that in the literature these notations are not unique and different authors have used different notations. For instance, whereas in (Blazewicz et al. [@CR7]), N5 is referred to the same neighborhood we mentioned, in (Vaessens et al. [@CR46]) N5 is used to refer to the neighborhood scheme developed in (Adams et al. [@CR1]), which, as a very large neighborhood, can completely change the order of operations on one machine. That is why in this paper, we refer to that neighborhood schemes as N5'.
Local searches, in general, and neighborhood schemes, in particular, are mainly used in the context of metaheuhurtics. Tabu searches, as part of metaheuristic category, select improving moves, and are aimed at avoiding to return to the solutions they have visited recently. In the cases where no improving move exists, simply the least disapproving move is chosen. The tabu search presented in (Nowicki and Smutnicki [@CR37]) has been one of the most effective searches presented for the JSP. Although it owes its effectiveness to both the employed neighborhood structure and the balance it maintains between diversification and intensification, restarting the search form elite solutions plays a crucial role in its success. In this metaheuristic, the search is controlled through a backtracking scheme, and the seed are provided through generating active schedules. The backtracking process embedded in the search causes the search to restart from various high quality solutions encountered from the beginning of the search. In other words, it recovers elite solutions with which to restart the search.
In the JSP whose fitness landscape has big valley structure (Pardalos et al. [@CR39]), the recovery of elite solutions and then restating the search with these solutions using different parameters has proved to be very effective. Because of a random setting, even with using the same high quality solution as a seed, each restarting initiates finding a new promising trajectory. It is through these trajectories that the big valley structure of the JSP is exploited and high quality solutions are obtained.
As two other effective tabu searches, we can name those presented in (Taillard [@CR45]) and (Barnes and Chambers [@CR6]), with both using N1 as their neighborhood structure. Whereas the first procedure changes the size of tabu iteratively and calculates the objective function value of each neighbor approximately, the other procedure has a fixed-length tabu list and computes the objective function value of each neighbor exactly. For generating a seed, several non-delay schedules are generated and the best one is selected.
As another metaheurstic, the guided local search presented in (Balas and Vazacopoulos [@CR3]), uses a variable neighborhood search in escaping local optimality. The main difference between this search and a typical variable neighborhood search is that it uses a tree whose nodes correspond to the orientations, with each descendant node being a neighbor of its parent node. In (Mattfeld [@CR34]), this search has cleverly been classified as a variable depth search due to the famous algorithm of (Lin and Kernighan [@CR30]), which has first been applied to the Travelling Salesman Problem (TSP). As another variable neighborhood search for the JSP we can mention the VNS presented in (Wang and Zhang [@CR48]).
The variable depth search starts with an initial solution, and in each iteration of the search, it starts with the best solution found in the previous iteration. Unlike in tabu search, each iteration performs not one but a number of potential profitable moves, which in general may be worsening moves, and during the corresponding iteration never reverses any of those moves.
Despite the fact that in the next iteration, previously prohibited moves are allowed to be performed again, the list of forbidden moves in the variable depth search grows much faster than that in tabu search. The other variable depth search procedure developed for the JSP, is the procedure presented in (Dorndorf and Pesch [@CR15]). Unlike in the TSP, this search has not been extremely successful for the JSP. It seems that this is partly due to the complicated structure of the JSP, which makes potentially profitable moves unrecognizable.
In (Lourenco [@CR31]; Lourenço and Adviser-Shmoys [@CR32]), a combination of N1 and a large step optimization search has been employed to tackle the problem. In the corresponding large step optimization, randomly two machines are selected and all disjunctive arcs related to these two machines are removed. Then, a new order is found for each of these machines through using the Carlier's method (Carlier [@CR11]), which solves one machine problem to optimality.
The fourth category, genetic algorithms, is not much related to our work, and the most successful genetic algorithms provided for the JSP include those presented in (Yamada and Nakano [@CR49]), (Dorndorf and Pesch [@CR16]), (Falkenauer and Bouffouix [@CR17]), and (Gonçalves et al. [@CR22]). Finally, we briefly survey hybrids as the last category. They comprise a variety of algorithms ranging from genetic (Qing-dao-er-ji and Wang [@CR42]) through ant bee-based (Zhang et al. [@CR52]) to differential evolution hybrids (Ponsich and Coello Coello [@CR41]).
The first related hybrid discussed is the algorithm based on global equilibrium presented in (Pardalos and Shylo [@CR38]). This algorithm, which is called Global Equilibrium Search (GES), has some common features with simulated annealing algorithm. In each stage of the search, GES collects information about the solution space for its next stages, and similar to the SA, the GES performs the search as a chain of temperature rounds.
The procedure also employs a local search which uses two neighborhood structures. The first neighborhood structure is N1 and the second structure manages to move each operation on the block either to the beginning or to the end of its corresponding block, similar to N4. The authors have modified an accelerating method in literature for the evaluation of moves in their second neighborhood structure. This method, which instead of computing the exact value associated with a move, calculates its tight lower bound at the cost of negligible computational effort, has significantly contributed to the effectiveness of the procedure.
The next related work in the hybrids is a filter-and-fan approach presented in (Rego and Duarte [@CR43]). The authors have presented a filter and fan approach for solving the JSP. The SBP (Shifting bottleneck procedure) has been used both for generating initial solutions and enhancing the final solutions as a post optimization procedure. It incorporates a tree search for restricting the solution space and works similar to the beam search.
Whereas beam search works in construction heuristics, filter and fan search, as its natural generalization, can work both in construction heuristics and local searches. When used in local searches, it places a local optimal solution at the root of the search tree, and the best *m* solutions obtained in this process are located at level 1, among the neighbors of these *m* solutions, the best *m* neighbors are selected and are placed in level 2. In effect, each level consists of *m* nodes that are selected among the best neighbors of nodes existing in its previous level. By using a hash mechanism, any repeat in this search tree is prevented. After generating *k* solutions and selecting the best solution obtained in the process, it becomes the root of the tree, and the search restarts.
The employed filter and fan procedure can work based on the first-improvement strategy as well. This means that whenever in the process, the root is improved, the enhanced solution becomes the root, and the best *m* solutions obtained in the previous search tree are placed in the first level of the next search tree.
The next related work in hybrids is a TS/SA algorithm presented in (Zhang et al. [@CR50]). The procedure developed by these authors is based on their conjecture that the quality of solutions obtained by the tabu search is determined based on the quality of the initial solutions. That is why they use a simulated annealing algorithm to generate high quality initial solutions for a tabu search. In other words, the main principle guiding this search is that SA generates elite solutions and TS improves the solutions generated by SA.
These kinds of integration are usually very effective. For instance, in (Huang and Liao [@CR25]) ant systems are combined with tabu searches and provide assistance for a decomposition method inspired by the shifting bottleneck procedure, used as the construction method of the procedure. Having briefly reviewed the related work, we can now describe the SLENP.
The SLENP {#Sec4}
=========
The SLENP combines small and large neighborhood schemes in coping with the job shop problem. Four modules have been combined to create this procedure. The first module is aimed at generating semi-active schedules by a forward-backward method called Semi-Active Schedule Generator (Forward-Backward-SASG). The second module is based on a local search that manipulates a small neighborhood structure guided by a tabu list, and the third module includes a feedback-based mechanism for perturbing the solutions generated. A very large-neighborhood local search, which is guided by a forward-backward shifting bottleneck (Forward-Backward-SBP) method, comprises the fourth module. Whereas the shifting bottleneck mechanism and tabu list are used as a machinery to manipulate the neighborhood structures, the perturbation mechanism diversifies the examined solutions, and the semi-active generator generates initial schedules.
The employed small neighborhood structure is highly restrictive, and the reason we have used semi-active, instead of active schedules, is that in (Jain et al. [@CR26]), through computational experiments, it has been shown that semi active schedules better match with restrictive neighborhood structures. The conclusion made in (Jain et al. [@CR26]) is partly based on semi-active schedules employed in (Nowicki and Smutnicki [@CR37]). It should be noticed that when the employed neighborhood structure is not restrictive, active schedules perform better than semi-active schedules. The reason is twofold. First, makespan is a regular criterion and the optimum belongs to the set of active schedules. Second, active schedules are a subset of semi-active schedules, and, on average, are of higher quality than semi-active schedules.
In the procedure, the repeats are detected by a feedback mechanism called repeat-check, which causes perturbation to occur whenever the total number of successive repeats for two identical values of the makespan reaches a given threshold. The shifting bottleneck module manipulates a large-neighborhood and is aimed at enhancing the solution obtained by the other three modules. To improve the results obtained by the shifting bottleneck method, a forward/backward mechanism has been added to it.
The pseudocode of the SLENP has been represented in Figure [5](#Fig5){ref-type="fig"}. The integrating part of the pseudocode is *Elite Heap,* which is a priority queue for keeping high quality solutions for possible improvement and releasing them based on their quality. First, at line 3, the *EliteHeap* is initially filled by repeatedly calling the Forward-BackwardSASG procedure. This is performed through generating *n* solutions by the Forward-BackwardSASG procedure, and selecting the best *m* solutions among them. The Forward-BackwardSASG procedure has its own pseudocode and will be discussed in detail.Figure 5**The c-type pseudocode of the SLENP.**
After the filling of the *EliteHeap*, the main loop in pseudocode starts at line 5. The main goal of this loop is to make possible improvement in the solutions located in the *EliteHeap*. Line 7 removes a solution from the *EliteHeap* and line 9 performs a limited tabu-search on this solution. In this tabu search, the tabu list includes elements which show the sequence of operations on a particular machine. In this tabu search, both N5 and N6′ neighborhoods are used. As discussed, N6′ is a version of N6 in which the possibility of moving the starting and ending operations of the block to the interior positions of the block have been considered.
The employed tabu list is aimed at determining whether, within a particular short-term period, a potential solution has been visited and decreases the possibility of repeatedly visiting the same sequence of solutions. Since this particular short-term memory cannot exclude the large sequences of repeats, lines 13 and 14 record the occurrence of every two consecutive solution values, and line 15 prevents any such possible repeat through making a perturbation in the current solution, aiming at further decrease in the possibility of repeats.
The performance of such perturbation depends on detecting a large sequence of repeats. For this purpose, if *x* has been followed by *x*' previously, C\[*x*,*x*'\] \> 1, the value of *totalRepeats* is incremented at line 14, and as soon as *totalRepeats* exceeds a certain threshold, the current solution is perturbed using the N1 neighborhood. As lines 17 and 18 indicate, with each perturbation, the history of the recorded solution values is discarded. The reason for discarding this history is that any perturbation changes the course of possible repeats. Line 22 ensures that the procedure is repeated until there is no improvement for *MaxIter* iterations. The loop terminates at line 23.
The best solution obtained in the loop undergoes an iterated forward-backward, one-machine post-optimization process for possible further improvement. The loop terminates at line 23, and line 24 applies a large-neighborhood search to the best solution obtained. This is performed by calling *ForwardBackwardSBP*, whose complete pseudocode has been given in Figure [6](#Fig6){ref-type="fig"}.Figure 6**The c-type pseudocode of the ForwardBackwardSBP module.**
This complete pseudocode describes a modified version of the post-optimization phase of the shifting bottleneck procedure (SBP) (Adams et al. [@CR1]). The modification performed includes adding a backward process to the procedure. The pseudocode starts with initializing a list of machines, L*,* at line 5.Then a loop starts at line 9. In each iteration of this loop, at line 12 the status of a machine is set to *open* (i.e. "not scheduled"), and at line 14 the sub-problem corresponding to that machine is given to the Carlier procedure to be solved.
Based on line 11, the procedure randomly switches from the forward case to the backward case and vice versa. After solving the corresponding one machine problem with the Carlier method at line 14, the output sequence is applied to the solution and the makespan is calculated at line 15. Then, line 16 replaces the previous sequence with the new one if the new sequence is of higher quality than the previous one. Lines 7, 9, 19, 21 ensure that the procedure is terminated if there is no improvement for 2*m* consecutive iterations. As lines 23 through 27 show, after each full optimization cycle for all *m* machines, the machine numbers in the list *L* are rearranged based on their solution cost to the one-machine problem, with the machines producing the longest makespan sitting in the top of the list and those producing the smallest makespan sitting in the bottom of the list. The pseudocode terminates at line 29.
As mentioned, the *EliteHeap* is filled by repeatedly calling the Forward-BackwardSASG procedure. The pseudocode of this procedure has been presented in Figure [7](#Fig7){ref-type="fig"}. The pseudocode, which generates semi-active schedules, starts with lines 4 and 5 and initializes a forward and backward binary heap, respectively. Then with a loop starting at line 8, forward and backward processes are used alternatively, starting with the forward process, which is set at line 7. Depending on whether the forward or backward process is in place, an element is taken from the forward or backward heap in lines 11 and 13, respectively.Figure 7**The c-type pseudocode of the ForwardBackwardSASG.**
The pseudocode constructs a solution by iteratively sequencing jobs on machines. This is done in the two directions of forward and backward alternatively, in the sense that in each iteration based on the value of the *direction* variable, an operation is scheduled on the beginning or the end of the schedule, respectively. Since in lines 11 and 13 from the top *k* elements of forward_BinaryHeap or backward_BinaryHeap, an element is chosen, and larger values of *k* can cause further diversification, the parameter *k* plays a key role in the quality of solutions generated.
The smaller values of *k* lead to generating a limited number of solutions, all in the same high quality region. On the other hand, when the value of *k* is increased, say to 3 or 5, the diversification is increased at the cost of decreasing the quality of regions. Since for filling the *EliteHeap*, this module is called for *n* times and among the *n* solutions generated the best *m* solutions are selected, the values of each of the three parameters *k, m*, and *n* should be selected based on the other two parameters.
The other issue contributing to the effectiveness of the procedure is the forward-backward mechanism embedded in the Carlier's method as the key component of the shifting bottleneck procedure. Both in the forward and backward processes, it is vital that the sub-solution provided by the Carlier's procedure matches with the current solution, in the sense that the Carlier' procedure should not introduce any loop in the current disjunctive graph. As it has been stated in (Adams et al. [@CR1]), introducing loops through the Carlier's procedure is not a common occasion and rarely can happen in practice. In our implementation, both in the forward and backward processes, we have prevented loops as follows.
Suppose, that the Carlier's procedure suggests 71, 78, 30, 1, 42, and 35 as the sequence of operations that should be processed, one after another, on the corresponding machine. Now in the disjunctive graph, we find all successors of the operation 35 and make sure that none of the operations before the operation 35 in the proposed sequence by the Carlier's procedure is among these successors. Then we find all successors of the operation 42, and make sure that none of the operations before the operation 42 is among them. Checking for the existence of any violation continues until we find that the operation 71 is not among the successors of the operation 78.
In case of encountering any violations, they are recorded, and the Carlier's procedure is called again, albeit with the set of recorded violations as a constraint for being avoided. This process is repeated until the sub-solution provided by the Carlier's procedure matches with the current solution. The employed tabu search selects improving moves and avoids returning to the solutions it has visited recently.
Computational experiments {#Sec5}
=========================
The SLENP has been implemented in C++ and compiled via GNU GCC compiler on a DELL PC with 2.2 Ghz speed. The benchmark problems to which the procedure has been applied include 43 instances extracted from ORLIB site managed by Brunel University, UK. The selected instances comprise a combination of representative problems collected from the literature. They range from 6 × 6 to 20 × 20 in size, with the first number showing the number of jobs and the second number showing the number of machines. They include 3 instances, *ftxx*, from (Fisher and Thompson [@CR18]), 11 instances, *laxx*, from (Lawrence [@CR29]), 5 instances, *abzx*, from (Adams et al. [@CR1]), 10 instances, *orbxx*, from (Applegate and Cook [@CR2]), 4 instances, *ynx*, from (Yamada and Nakano [@CR49]), and 10 instances, *swvxx*, from (Storer et al. [@CR44]).
The procedure has 11 parameters, and most of these parameters have been set in terms of the number of jobs, *n*, and the number of machines, *m*. In setting the parameters, care has been taken to increase the exploration power of the procedure with respect to the increase in *m*, and ⌊*n*/*m*⌋. Table [1](#Tab1){ref-type="table"} shows how these parameters have been set. A brief description of these parameters is as follows; (i) *TotSol* denotes the number of total initial solutions generated, (ii) *k* represents diversification parameter of the Forward-BackwardSASG module which controls the diversity of initial solutions, (iii) *EliteHeapSize* indicates the size of EliteHeap, (iv) *PerturbProb* represents the chance by which the solution is perturbed after being removed from eliteheap, (v) *MaxIterNonImprov* denotes the number of iterations after which the tabu search stops if no improvement occurs in makespan, (vi-vii) *TabuSize*~*min*~ and *TabuSize*~*max*~ show the extremes of the bound in which the size of the tabu list as a uniform random variable can change in each run, (viii-ix) *RepeatTolerance*~*min*~, and *RepeatTolerance*~*max*~ represent the extremes of the bound in which the tolerance for accepting consecutive makespan repeats is changed randomly, (x) *N5N6′Prob* denotes the chance of selecting N5 neighborhood in each iteration and consequently the chance of selecting N6′ neighborhood is *1-N5N6′Prob*, (xi) *TripleMovesProb* represents the probability by which at the start and the end of the critical block a triple move is performed.Table 1The value of each parameter as either a constant or a function of the number of jobs (n) and the number of machines (m)ParameterValue*TotSoln.m*+1000*k*10*EliteHeapSize*500*PerturbProbMaxIterNonImprovTabuSize* ~*min*~*TabuSize* ~*max*~*RepeatTolerance* ~*min*~*RepeatTolerance* ~*max*~*N5N6'Prob*0.1*TripleMovesProb*0.6
As stated in the previous section, the SLENP uses both N5 and N6′ as its small neighborhood. That is why among the parameters described above, the parameter *N5N6′Prob* has been used to determine the chance of selecting each of the two neighborhoods. Using one of the two neighborhoods randomly causes that starting with the same initial starting point leads to different courses of actions and consequently to different solutions, improving the diversity of solutions generated and increasing the chance of escaping local optimality.
Before presenting the performance of the procedure for the benchmark instances and comparing the obtained results with the best available solutions, we first show the results of its operations on the *ft10*, which is the most famous instance. Figure [8](#Fig8){ref-type="fig"} shows the solution value (makespan) per each schedule evaluation. As is seen, the makespan has converged towards 930, which is the optimal solution of the problem. It is worth mentioning that value of 930 has been found after applying the shifting bottleneck improvement heuristic in the final stage of the procedure. The shifting bottleneck improvement heuristic has been executed after the procedure has evaluated 272850 schedules, and has been performed on a solution whose makespan was 934.Figure 8**Solution value (makespan) per each schedule evaluation for the instance ft10.**
Figure [9](#Fig9){ref-type="fig"} shows the best makespan obtained for each element removed from the elite heap. In effect, for 90 different elements taken from the elite heap, this figure has presented the trend in which the corresponding element has been improved in the search. The 90 different peaks in the figure correspond to the 90 solutions taken out from the heap and each shows the makespan of the corresponding element. Associated with each peak is a dip which shows the makespan of the best solution obtained.Figure 9**Best makespan achieved per each element removal from the elite heap.**
As is seen in Figure [9](#Fig9){ref-type="fig"}, each of the elements taken from the heap has led to a different solution, and sometimes high quality elements taken from the heap have produced solutions which cannot compete with solutions produced by low quality elements. The trend, however, is towards the improvement of the final solution. Figure [10](#Fig10){ref-type="fig"} shows such a trend by depicting the changes in the global best solution per the removal of each element from the elite heap.Figure 10**The changes of global best solution per each removal from the elite heap.**
Now we present the results of applying the procedure to the entire benchmark instances selected. To remove the effect of the random seed, in line with other procedures, for each instance, the SLENP has been run for 10 times each with a different random seed. The time allowed for each run is *n*(9*n*-60)/*m* seconds for instances with *n* ≥ 10 (jobs) and 1 second for the instances with *n* \< 10. Moreover, since in the case of availability of the optimal makespan it is given to the procedure as an input, the procedure can stop as soon as a solution with such quality is achieved.
With respect to performance, Table [2](#Tab2){ref-type="table"} compares the procedure with one of the fastest available procedures for the JSP, namely TSSA (Tabu Search Simulating Annealing) (Zhang et al. [@CR50]). In this table, %DEV~best~ represents the deviation percentage of the obtained solution from the best available solution in the literature, BKS, and has been obtained based on the formula of (*s*-*BKS*)/*BKS*, with *s* being defined as the best solution returned by procedure. The running times of the TSSA have been reported on a Pentium IV 3.0 Ghz CPU.Table 2**Comparing the performance of the SLENP with that of TSSA**SLENPTSSA**InstanceSizeLBBKSBest%DEV** ~best~**T** ~best~**AvgT** ~avg~(s)**Best%DEV** ~best~**AvgT** ~avg~ft066 × 65555**55**0.0000550.005--------ft1010 × 10930930**930**0.0000.4559309.2249300.0009303.8ft2020 × 511651165**1165**0.0000.4811652.727--------la1910 × 10842842**842**0.0000.138420.7768420.0008420.5la2115 × 1010461046**1046**0.0005.31046.715.21610460.000104615.2la2415 × 10935935**935**0.00010.29936.520.3609350.000936.219.8la2520 × 10977977**977**0.0006.297713.6999770.000977.113.8la2720 × 1012351235**1235**0.0009.08123531.98012350.000123511.7la2920 × 101152115211620.86886.641163.540.02411530.0871159.263.9la3615 × 1512681268**1268**0.0003.41268.312.93712680.00012689.9la3715 × 1513971397**1397**0.0001.9713979.18113970.0001402.542.1la3815 × 1511961196**1196**0.0004.351198.614.83611960.0001199.647.8la3915 × 1512331233**1233**0.0001.311233.619.09912330.0001233.828.6la4015 × 151222122212240.1646.571226.615.92612240.1641224.552.1abz510 × 1012341234**1234**0.0000.5812343.545--------abz610 × 10943943**943**0.0000.129430.151--------abz720 × 156566566620.91539.0466464.8226580.305661.885.9abz820 × 156456656680.451106.52672.455.9356670.301670.390.7abz920 × 156616786881.47598.7689.535.8206780.000684.890.2orb0110×1010591059**1059**0.0000.71059.65.96110590.00010593.5orb0210×10888888**888**0.0000.148880.4758880.000888.16.4orb0310×1010051005**1005**0.0000.36510057.41510050.0001012.513.8orb0410×1010051005**1005**0.0000.1551006.27.58010050.0001008.314.3orb0510×10887887**887**0.0001.2388712.0938870.000888.66.6orb0610×1010101010**1010**0.0000.231010.99.16510100.00010108.5orb0710×10397397**397**0.0000.143970.2843970.0003970.5orb0810×10899899**899**0.0002.268996.0208990.000902.57.2orb0910×10934934**934**0.0000.189340.5099340.0009340.4orb1010×10944944**944**0.0000.159440.1769440.0009440.3yn120×208268848920.90566.63897.740.0408840.000891.3106.3yn220×208619079110.4411.78913.462.3129070.000911.2110.4yn320×208278929000.89759.85903.142.1788920.000895.5110.8yn420×209189689821.44649.81986.856.0479690.103972.6108.7swv0120×101407140714372.132115.331458.560.64214120.3551423.7142.1swv0220×101475147515052.03492.27152064.43114750.0001480.3119.7swv0320×101369139814262.00382.68143448.66513980.0001417.5139.1swv0420×101450147015112.78966.471517.857.66414700.0001483.7143.9swv0520×101424142414753.58145.4611492.358.09214250.0701443.8146.7swv0620×151591167817303.09943.941738.281.89216790.0601700.1192.5swv0720×151446160016322.00088.79164868.54616030.1881631.3190.2swv0820×151640175618072.904147.391814.171.76517560.0001786.9190swv0920×151604166117012.408126.121707.568.66116610.0001689.2193.8swv1020×151631175418123.307122.661820.685.66617540.0001783.7184.6
Table [2](#Tab2){ref-type="table"} shows that, in 25 out of 43 cases, the SLENP has been able to find the best known solution for the corresponding benchmark instance. As is seen in Table [2](#Tab2){ref-type="table"}, for four of the benchmark instances, the TSSA has no corresponding output. Removing these four rows out of consideration, the following conclusions can be drawn. In 53.8%, 21/39, of cases, the SLENP has generated solutions with the same quality as those generated by the TSSA and in general the solutions generated by the TSSA are on average around 1.05% better than those generated by the SLENP. However, the solutions produced by the SLENP have been obtained on average 112.85% faster than those generated by the TSSA. Since both procedures have used only a single processor, taking the difference between the clocks pulses on which these two procedures have been run, 2.2 versus 3.0 Ghz, implies the chance that this speed percentage may be larger than the value presented.
Also since in (Chassaing et al. [@CR13]), a comparison has been made among several procedures based on their performance on solving la01 to la40 instances, we have tested our procedure on these instances as well. It is worth noting that despite the fact some of these instances were included in our first experiments, we have solved them again, with new initial random solutions. The results have been shown in Table [3](#Tab3){ref-type="table"}. As is seen, except for three instances, the SLENP has been able to find the optimal solutions of all instances. One of these instances belongs to what those authors have classified as strongly large instances, and the other two instances belong to what they have classified as large instances.Table 3**The performance result on la01 to la40**SLENP**InstanceSizeBKSBest%DEV** ~best~**T** ~best~(s)**AvgT** ~avg~(s)la0110 × 56666660.0000.006660.00la0210 × 56556550.0000.066550.07la0310 × 55975970.0000.065970.09la0410 × 55905900.0000.055900.06la0510 × 55935930.0000.005930.00la0615 × 59269260.0000.009260.00la0715 × 58908900.0000.008900.01la0815 × 58638630.0000.008630.00la0915 × 59519510.0000.009510.00la1015 × 59589580.0000.009580.00la1120 × 5122212220.0000.0012220.00la1220 × 5103910390.0000.0010390.00la1320 × 5115011500.0000.0011500.00la1420 × 5129212920.0000.0012920.00la1520 × 5120712070.0000.0212070.15la1610 × 109459450.0000.139450.69la1710 × 107847840.0000.117840.13la1810 × 108488480.0000.128480.16la1910 × 108428420.0000.168420.51la2010 × 109029020.0000.149020.20la2115 × 10104610460.0002.681046.814.58la2215 × 109279270.0000.729275.50la2315 × 10103210320.0000.2310320.24la2415 × 109359350.0004.04935.222.75la2515 × 109779770.0001.51977.210.97la2620 × 10121812180.0000.4112180.44la2720 × 10123512350.0002.06123527.79la2820 × 10121112160.4130.4512160.79la2920 × 10115211630.95525.801164.332.50la3020 × 10135513550.0000.3813550.40la3130 × 10178417840.0000.6817840.70la3230 × 10185018500.0000.0218500.04la3330 × 10171917190.0000.2317190.70la3430 × 10172117210.0000.6617210.67la3530 × 10188818880.0000.7418880.77la3615 × 15126812680.0005.281268.416.01la3715 × 15139713970.0001.1213973.89la3815 × 15119611960.0001.751199.214.18la3915 × 15123312330.0006.541235.421.19la4015 × 15122212250.24525.121227.117.95
Concluding remarks {#Sec6}
==================
Towards generating both fast and high-quality solutions to the JSP, the SLENP has synergistically combined a construction method, a local search, and a large-neighborhood technique as a post-optimization component. Its construction technique constructs feasible schedules iteratively, one element at a time. The innovative feature of this component is the use of a forward-backward mechanism in scheduling activities.
The second component has been a local search, which starts with the initial schedules generated by the construction method. Consistent with all current effective neighborhood structures for the JSP, the employed local search has been founded on the concept of the critical block, which guides the construction neighbors. The innovative feature of this component is the use of two different neighborhoods that a parameter decides which to work in each round.
The third major component of the SLENP is the post-optimization method whose development has been inspired by the combination of the forward-backward shifting bottleneck procedure and the biased randomizing search. The reason for selecting this procedure as the post-optimization component for the SLENP has been twofold.
First, the shifting bottleneck heuristic is one of the most effective heuristics for the JSP which by sequencing the bottleneck machine successively, can provide high quality solutions for the JSP. Second, equipping this mechanism with a forward-backward process can further improve the accuracy of this highly effective mechanism.
By using these components, the SLENP can find solutions with high quality in a matter of seconds. This indicates that the components of the procedure act synergistically. Towards its enhancement, three major directions can be envisaged for sketching the procedure.
First, since a large portion of execution time is spent on calculating the values of the makespan, a faster evaluation technique, which without explicit calculation can estimate the makespan, can lead to producing solutions with higher quality through increasing search efficiency.
Second, in a parallel environment, various local searches can simultaneously operate to cooperatively locally optimize various parts of the same encoding. In this parallel environment, the local searches can communicate with one another so that each local search can ignore those parts of the encoding fixed by other local searches and concentrate only on manipulating its own part.
Third, since in the employed tabu search, the management of tabu list, which, as a short term memory, keeps the forbidden moves, has played a critical role in the quality of the overall result, other possible mechanisms in managing the tabu list can be tested. One promising mechanism is an adaptive tabu list, which based on a feedback received from the corresponding fitness landscape, can alternate between a fixed list and a random-sized list. Such a flexible tabu list, whose size is determined adaptively, can properly adjust the short term memory of the search and can possibly lead to higher quality solutions.
**Competing interests**
This research has been done with the full support of Wollongong University in which Mr Mehrdad Amirghasemi is a PhD candidate and Dr Reza Zamani is his supervisor. The authors declare that they have no competing interests.
**Authors\' contributions**
MA carried out design and programming tasks, Also all computational experiments were performed by MA. RZ helped in debugging and improving the performance of the computer code. Both authors read and approved the final manuscript.
| {
"pile_set_name": "PubMed Central"
} |
Data are available from Oregon State University <ScholarsArchive@OSU> (<https://ir.library.oregonstate.edu/concern/datasets/8623j475z>).
Introduction {#sec001}
============
The genus *Alkanna* belongs to family Boraginaceae, subfamily Boraginoidae, tribe Lithospermeae \[[@pone.0233516.ref001]\]. The *Alkanna* species show the greatest diversity in the southern part of the Balkan Peninsula, in the Mediterranean and Irano-Turanian region, and subtropical areas of the world \[[@pone.0233516.ref002]\]. Generally, genus *Alkanna* is represented by a large number of local and regional endemic species \[[@pone.0233516.ref003]\]. Six *Alkanna* species are found in the Bulgarian flora and five of them are either Balkan or Bulgarian endemics. *Alkanna primuliflora* Griseb., *A*. *stribrnyi* Velen. and *A*. *graeca* Boiss. & Spruner are Balkan endemics, while *A*. *stojanovii* Kožuharov and *A*. *jordanovii* Kožuharov are two Bulgarian endemics \[[@pone.0233516.ref004]\]. The three studied species have very limited distribution in Bulgaria as follow: *A*. *graeca*---in two floristic regions, Pirin Mt (South), Thracian Lowland; *A*. *stribrnyi*--in three floristic regions, Slavyanka Mt, Thracian Lowland, Pirin Mt (South); and *A*. *primuliflora*---mainly in four floristic regions, Thracian Lowland, Pirin Mt (South), Valley of Struma river, and Slavyanka Mt. The populations of *A*. *primuliflora*, *A*. *stribrnyi*, and *A*. *graeca* are included in the European ecological network EU NATURA 2000 under Directive 92/43 / EEC on Habitats. In Bulgaria, the populations of *Alkanna* species are distributed on dry, rocky places. Generally, the populations of three *Alkanna* species are small, fragmented and represented by a very limited number of individual plants.
Overall, there is a large variation in the morphological characteristics between the *Alkanna* species but also within a species. Some variations include variously colored corolla in yellow (*A*. *primuliflora)*, orange (*A*. *stribrnyi*) or yellow or orange (*A*. *graeca*), with more or less cylindrical tube and infundibuliform limb, with a ring of hairs \[[@pone.0233516.ref004]\]. *Alkanna primuliflora*, *A*. *stribrnyi* and *A*. *graeca* are perennial, pubescent herbaceous plants. They are insect-pollinated, they flower in April-May, and are seed propagating plants \[[@pone.0233516.ref004]\]. Morphological traits of taxonomic importance include color of the corolla and nutlets, shape and size of calyx, leaf form and trichomes \[[@pone.0233516.ref004],[@pone.0233516.ref005]\]. The trichomes in the leaf and other parts of *Alkanna* species may be glandular or non-glandular \[[@pone.0233516.ref004],[@pone.0233516.ref006]\]. According to Kožuharov \[[@pone.0233516.ref004]\], the three *Alkanna* species prefer dry and hot places, and they are found on rocky slopes of up to 900 m asl. *Alkanna stribrnyi* and *A*. *graeca* have transient morphological characteristics with those of *A*. *primuliflora* \[[@pone.0233516.ref004]\].
There are no comprehensive phylogenetic studies on the genus *Alkanna*, although several molecular studies have been published \[[@pone.0233516.ref007]--[@pone.0233516.ref010]\]. However, the latter reports described studies on the evolutionary history and phylogenetic relationships in the Boraginaceae family, especially on taxa belonging to Lithospermeae and Boragineae \[[@pone.0233516.ref007]--[@pone.0233516.ref011]\].
The DNA and ISSR analysis can be additional keys to describe and differentiate the *Alkanna* species. Also, the content of alkaloids in various *Alkanna* species can be used as a chemotaxonomic marker. The value of chemotaxonomic markers in the systematics of the family Boraginaceae has been elaborated previously \[[@pone.0233516.ref012],[@pone.0233516.ref013]\]. Plants of the family Boraginaceae are known for the synthesis and accumulation of pyrrolizidine alkaloids (PAs) \[[@pone.0233516.ref014]\]. This class of alkaloids received considerable attention because of their hepatotoxic, mutagenic and cancerogenic activities and this class of alkaloids has been widely used as a chemotaxonomic marker \[[@pone.0233516.ref012], [@pone.0233516.ref015]--[@pone.0233516.ref019]\]. The PAs have been previously documented only in two *Alkanna* species; *A*. *orientalis* and *A*. *tinctoria* \[[@pone.0233516.ref020]--[@pone.0233516.ref025]\].
*Alkanna primuliflora*, *A*. *stribrnyi* and *A*. *graeca* have limited distribution in the flora of the Balkans. Generally, there are several major reasons for the limited distribution of plant species, and these include; (1) habitat preferences; (2) poor competitiveness; (3) problems in the reproductive potential of the species. Investigation of the peculiarities in the generative sphere in a particular plant species is a key factor in assessing the species reproductive potential. The features of reproductive biology in studied species (processes of development of gametes, pollination, the formation of embryo and endosperm, pollen and seed vitality) determining their reproductive potential, can provide important information on the factors limiting or preventing the reproduction of a specific species. The study of the biology of reproduction together with the determination of the genetics of their populations will help to define the causes for the limited distribution of the three *Alkanna* species. Such studies in combination with pollination studies, would facilitate defining strategies for plant species conservation (e.g. *in vitro* propagation, ex-situ collection). Due to the limited distribution of endemic *Alkanna* species, there are no reports on pyrrolizidine alkaloids content in *A*. *primuliflora*, *A*. *stribrnyi*, and *A*. *graeca*. Embryological studies also were not previously conducted on the three targeted species. In addition, no comparative study on the three *Alkanna* species embryology was published. Therefore, the purpose of this study was to: (1) distinguish and detect molecular genetic markers between *A*. *primuliflora*, *A*. *stribrnyi*, *A*. *graeca*; (2) reveal the alkaloids composition; (3) reveal embryological features that may be contributing to the limited distribution of these species.
The hypothesis was that the limited distribution of *A*. *primuliflora*, *A*. *stribrnyi*, and *A*. *graeca* was due to their reproductive capacity and genetic diversity. Furthermore, we hypothesized that the three species will contain pyrrolizidine alkaloids (PAs), as other species of the genus *Alkanna* (Boraginaceae), but they would have differential alkaloids composition.
The alkaloid profile of the investigated species could also provide supplementary information for the species diversity. However, the application of alkaloids for this purpose has to be cautiously used, because the alkaloid pattern is influenced by the environmental factors. Therefore, the alkaloid profile of the species was compared at two levels, between different species and within the same species in different natural localities or stage of development. The Ministry of Environment and Water of Bulgaria could use the result from this study and consider these endemic species for a protection status.
2. Materials and methods {#sec002}
========================
2.1. Plants material {#sec003}
--------------------
The plant materials for this study were *Alkanna primuliflora* Griseb., *A*. *stribrnyi* Velen., and *A*. *graeca* Boiss. & Spruner. The five collection locations of *Alkanna* species with the exact coordinates, altitude, and habitats are presented in [Table 1](#pone.0233516.t001){ref-type="table"}. Voucher specimens of *A*. *primuliflora*, *A*. *stribrnyi*, *A*. *graeca* were deposited at the Herbarium of the Agricultural University, Plovdiv, Bulgaria (SOA) \[[@pone.0233516.ref026]\].
10.1371/journal.pone.0233516.t001
###### Collection sites, location coordinates, altitude (m.a.s.l.) and habitats of *Alkanna primuliflora*, *A*. *stribrnyi* and *A*. *graeca* in Bulgaria.
{#pone.0233516.t001g}
Species Location and habitat GPS coordinates masl
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ---------------------------------------------------------------------------------------------------------------------------- ------------------------------------------
*A*. *primuliflora* Thracian lowland, Plovdiv, Sienite rocks, dry places (Aprm1, in flowering); (Aprm2, after flowering) 42°08\'16.9\"N 024°43\'63.8\"E 107masl
*À*. *stribrnyi* Thracian lowland above Asenovgrad town, Sv. Petka, dry stone places, metamorphic rocks, The soil cover is very thin (Ast1) 41°59\'35.8\" N 024°52\'20.1\" E 310masl
Thracian lowland, above Asenovgrad town, Asenova kr, dry stone places, metamorphic rocks, The soil cover is very thin. (Ast2) 41°59\'51.8\" N 024°52\'53.1\" E 418masl
Thracian lowland, Novo selo villige, karst marbles and gabbro-diorite covered with thin lake Pliocene sediments. The soil cover is very thin. Almost completely deforested. (Ast3) 42°05\'39.9\"N 024°28\'01.1\"E 328masl
*A*.*graeca* Above Asenovgrad town, dry stone places, metamorphic rocks, The soil cover is very thin. (Agr1) 41°59\'13.8\" N 024°52\'17.1\"388masl
*A*. *primuliflora*---Aprm1 (population 1, in flowering); Aprm2 (population 1, after flowering); *A*. *stribrnyi* -Ast1 (population 1, in flowering); Ast2 (population 2, in flowering), Ast3 (population 3, in flowering); *A*.*graeca*--Agr 1 (population 1, in flowering).
2.2. Methods {#sec004}
------------
Plant samples were taken separately for DNA, phytochemical, and embryological studies. Plants were sampled for DNA, alkaloid, and some embryological analyses in May. Only plant leaves were collected for the DNA analyses, while whole plants were collected for the alkaloids testing. For the embryology analyses, flower buds and open flowers were collected in May and then seeds were collected in June-July.
### 2.2.1. DNA extraction {#sec005}
This study attested that the individual plants within populations of the three *Alkanna* species in Bulgaria had a very limited number. Therefore, leaf samples were collected carefully from a few plants to minimize the potential damage to the populations. We performed the sampling depending on the density of each population; plants were randomly chosen for sampling. For genetic analysis, we used the leaves of the three species at flowering (in May) as follow: *A*. *primuliflora*--ten samples; *A*. *graeca*--eight samples; *A*. *stribrnyi*--six samples. Тhese are the total number of samples per species per population as shown in [Table 1](#pone.0233516.t001){ref-type="table"}.
The DNA from the three endemic species was extracted using the standard CTAB extraction procedure \[[@pone.0233516.ref027]\] with minor modifications. The modified method involves isolation of gDNA by a modified-CTAB method employing a high concentration of PVP at around 2.5% w/v and 2-BME up to 5% v/v. DNA quality and concentrations were measured using a NanoDrop ND-1000 spectrophotometer (Saveen and Werner AB, Limhamn, Sweden). DNA samples were diluted to 25 ng/μl and stored at -20°C before amplification.
### 2.2.2. Inter-simple sequences repeat (ISSR) amplification product analysis {#sec006}
First, the polymorphism of 40 markers was tested and then 10 polymorphic and reproducible ISSR primers (Microsynth, Balgach, Switzerland) were selected and used ([Table 2](#pone.0233516.t002){ref-type="table"}). Polymerase chain reactions (PCR) were performed according to the method described previously \[[@pone.0233516.ref028]\]. PCR reactions were carried out in a Techne TC-5000 gradient thermal circler (Techne, Staffordshire, UK). The reproducibility of the technique was tested by replicating each amplification reaction twice. Amplification products were separated on 2% agarose gels stained by incorporating 1% GelRed (Biotium Inc., USA) at 1.5h, 135V along with 100 bp Plus DNA Ladder (Thermoscientific, Vilnius, Lithuania). The DNA fragments were visualised under UV light and further analysed with a video image analyser (BioImaging Systems, Cambridge, UK).
10.1371/journal.pone.0233516.t002
###### ISSR primers used for analysis of genetic diversity of *Alkanna primuliflora*, *A*. *stribrnyi* and *A*. *graeca* in Bulgaria.
{#pone.0233516.t002g}
Primer Sequence 5\'- 3\' Total number of bands Number ofpolymorphic bands Polymorphism (%) Annealing temperature(°C)
-------- ------------------- ----------------------- ---------------------------- ------------------ ---------------------------
ISSR1 (CA)~8~G 12 7 58 55
ISSR2 (AC)~8~C 14 14 100 55
ISSR3 (AC)~8~G 13 10 77 55
ISSR4 (AG)~8~YC 10 7 70 60
ISSR5 (AC)~8~YT 14 14 100 60
ISSR6 (AC)~8~YG 9 4 44 60
ISSR7 (AG)~10~C 10 8 80 60
ISSR8 (AG)~8~YT 12 8 67 60
ISSR9 (AC)~8~YA 12 5 42 60
ISSR10 (AC)~8~YG 9 7 78 60
Total 115 85 72
P: % of polymorphic loci; *He*: expected heterozygosity, *uHe*: unbiased expected heterozygosity; SI: Shannon's information index; Standard error is shown in parentheses.
### 2.2.3. ISSR data analysis {#sec007}
To construct binary matrix the ISSR bands were scored as presence (1) or absence (0) of binary characters. The matrix was used for calculation of population genetic variation indices and assessment of the genetic diversity and structure of the *Alkanna* populations. The following genetic diversity indexes were calculated to determine the level of genetic variation within populations: the percentage of polymorphic loci (*P*), expected heterozygosity (*He*), unbiased expected heterozygosity (*uHe*) and Shannon's information index (*SI*) using GenAlEx *ver*. 6.5 \[[@pone.0233516.ref029]\]. Also, Principal Coordinate Analysis (PCoA) and Hierarchical Analysis of Molecular Variance (AMOVA) in GenALEx software were conducted to show the genetic relationships among investigated *Alkanna* populations based on Nei's genetic distances. Bayesian clustering-based structure analysis was performed on the entire data set in STRUCTURE *ver*. 2.3.4), which uses a Markov Chain Monte Carlo (MCMC) algorithm to detect population structure \[[@pone.0233516.ref030]\]. To estimate the admixture of individuals into groups (K) and the best K value (number of groups), the Evanno test was performed on STRUCTURE results using STRUCTURE HARVESTER *v*.0.6.8 \[[@pone.0233516.ref031],[@pone.0233516.ref032]\]. Finally, genetic distances between all *Alkanna* individual plants were calculated based on the results of Nei's genetic distance \[[@pone.0233516.ref033]\].
### 2.2.4. Extraction of plant material and qualitative analysis of alkaloids {#sec008}
The plant materials used included above ground plant parts of at least 10--15 randomly selected plants (per species) of *A*. *primuliflora* (Aprm1- 24.17g), *A*. *stribrnyi*, (Ast1- 14.29g), (Ast2- 6.89 g) (Ast3- 11.46g) and *A*. *graeca* (Agr1- 14.41g) in flowering period and *A*. *primuliflora* (Aprm2- 5.08g) in after flowering period from the five collection locations ([Table 1](#pone.0233516.t001){ref-type="table"}). The plant material of each sample was extracted exhaustively with CH~3~OH in a Soxhlet apparatus. The combined CH~3~OH extracts after evaporation to dryness were acidified with 5% HCl, filtered and extracted with CHCl~3~. The aqueous acid solutions were stirred with zinc (Zn) dust (24 h) then filtered and made alkaline with 25% NH~4~OH to pH 9. The alkaline solutions were extracted with CHCl~3~ to give crude alkaloid mixtures (CAM (Apr1)---15.61 mg; CAM (Aprm2)---4.67 mg; CAM (Ast1)---10.76 mg; CAM (Ast2)---6.1 mg; CAM (Ast3)---9.15 mg and CAM (Agr1)---12.27 mg). The crude mixtures of alkaloids were analyzed by Gas chromatography-mass spectrometry (GC/MS). The alkaloids 1 to 8 were identified by comparison of their mass spectral fragmentation with standard reference spectra from database Wiley 275 and the literature data \[[@pone.0233516.ref013],[@pone.0233516.ref034]\]. The GC/MS analysis was performed with GC Hewlett Packard 6890 (Palo Alto, California, U.S.A.) using MS detector Hewlett Packard 5973 fitted with an HP-5 MS column (30 m x 0.25 mm x 0.25 μm). The program was as follows: injector T 250°C; the temperature program was 100°C (2 min) to 280°C, 5°C/min, isothermal at 280°C for 20 min. Split ratio 1: 30, carrier gas (He), constant flow of 0.8 mL/min.
### 2.2.5. Embryological analyses {#sec009}
The major parameters of the reproductive biology of the studied species (structures and processes in the generative sphere, pollen and seed viability) were investigated to estimate their reproductive capacity. For the embryological studies we used flowers, buds and open flowers in different developmental stages collected from 30 individuals of each population. The material (flower buds and open flowers) was fixed in a mixture FAA (formalin: glacial acetic acid: 70% ethanol in a ratio of 5:5:90 parts, respectively). Then the fixed plant material was treated according to the classical paraffin methods \[[@pone.0233516.ref035]\] to obtain permanent microscopic slides. Heidenhain's haematoxylin as colouring agent \[[@pone.0233516.ref036]\] and Entellan Merck (rapid non-aqueous mounting medium, containing xylene) as a slide mounting medium were used for the permanent slides. The main embryological structures and processes in the male and female generative spheres were established after observations on Olympus Light CX2 microscope (Olympus Corporation, Shinjuku, Tokyo, Japan). The microphotographs were prepared using "Infinity lite" digital camera 1.4 Mpx.
*2*.*2*.*5*.*1*. *Polen viability*. Assessment of the pollen viability was conducted using Acetocarmine test \[[@pone.0233516.ref037]\]. According to the method used, temporary slides were prepared after staining the pollen grains. The pollen grains from 30 anthers from different individual plants per species were counted. Before that, the pollen grains were coloured with a solution containing 1% acetocarmine and were stained in red (viable, fertile) and unstained (nonviable, sterile). The mature pollen grains were counted (in the visual field at 100x magnification) using a light microscope.
*2*.*2*.*5*.*2*. *Seed viability testing*. The tetrazolium test was used for estimation of the viability of seeds. One hundred seeds per population of *A*. *primuliflora*, *A*. *stribrnyi*, and *A*. *graeca* were used. The seeds from different individual plants per tested species were sampled at maturity of the three species, which occurred in June--July. Before treatment with 1% solution of the tetrazolium chloride, the seeds were kept for 24 hours in Petri dishes on wet filter paper at 25°C temperature. The seeds were incubated in 1% solution of 2, 3, 5-triphenyltetrazolium chloride according to the methods outlined in AOSA (Association of Official Seed Analysts) \[[@pone.0233516.ref038]\]. Initially, the tetrazolium solution is colourless, but later on, it changes from dark pink to red as a result of the action of hydrogen ions coming from the respiration process of the seeds. Embryos that show a physiological activity (active respiration) turn red and are therefore considered viable. The darker the red colour, the greater the respiratory activity of the seeds.
3. Results {#sec010}
==========
3.1. Population genetic diversity study {#sec011}
---------------------------------------
Of the 40 ISSR primers used for polymorphism validation, 10 primers were polymorphic and up to 115 reproducible bands (all species together) were generated, and 85 (73.9%) of those, were polymorphic ([Table 2](#pone.0233516.t002){ref-type="table"}). The size of all bands ranged from 100 to 3000 bp and the number of bands by each primer was from 9 to 14 with an average of 11.5 per primer ([Table 3](#pone.0233516.t003){ref-type="table"}). The percentage of polymorphic loci (*P*) varied from 41.7% in *A*. *stribrnyi* to 64.6% in *A*. *graeca*. *Alkanna graeca* exhibited the highest values of genetic indexes expected heterozygosity *(He)*, unbiased expected heterozygosity *(uHe)* and Shannon's information index *(SI)* of 0.157, 0.167 and 0.256, respectively. The investigated *Alkanna* species had a total of 20 private alleles, nine of *A*. *primuliflora* and *A*. *graeca* and two of *A*. *stribrnyi* ([Table 3](#pone.0233516.t003){ref-type="table"}).
10.1371/journal.pone.0233516.t003
###### Genetic diversity indices of investigated *A*. *primuliflora*, *A*. *stribrnyi* and *A*. *graeca*.
{#pone.0233516.t003g}
Population *P* (%) *He* *uHe* SI Private bands
------------------------ ------------ -------------- -------------- -------------- ---------------
*Alkanna primuliflora* 62.5 0.137(0.022) 0.144(0.023) 0.227(0.032) 9
*Alkanna graeca* 64.6 0.157(0.022) 0.167(0.024) 0.256(0.033) 9
*Alkanna stribrnyi* 41.7 0.147(0.028) 0.161(0.031) 0.220(0.041) 2
Mean (SE) 56.3(7.32) 0.147(0.014) 0.157(0.015) 0.234(0.020) 6.67
### 3.1.1. Population genetic structure and differentiation {#sec012}
In the two-dimensional PCoA, the results showed a clear differentiation between the species; the first two principal components explained 18.7% and 32.2% of the total variance, respectively ([Fig 1](#pone.0233516.g001){ref-type="fig"}). Here, the individuals clustered strongly according to their species assignation and form separate groups. For example, individuals from *A*. *primuliflora* and *A*. *graeca* form two separate clusters that combined overlap only a few individuals. *Alkanna stribrnyi* clustered separately and some individuals fell quite isolated from each other and the other populations.
{#pone.0233516.g001}
The individuals of populations were able to form clear clusters, according to their geographical location, which suggests a strong genetic structure of populations studied herein. Only a few *A*. *graeca* individuals fell into *A*. *primuliflora* population cluster. This result could be explained with the presence of overlapping areal between natural *Alkanna* populations.
Population genetic structure was further studied using a Bayesian clustering algorithm, perform in STRUCTURE ([Fig 2](#pone.0233516.g002){ref-type="fig"}).
![(A) Plot of K clusters versus mean (±SD) likelihoods (L\[K\]): (B). K plotted against the second order rate of change of the likelihoods (ΔK); (C). Genetic relationships among the populations of *Alkanna* estimated using STRUCTURE program based on ISSR data. STRUCTURE clustering results for K = 3 to as suggested in (B). Estimated genetic clustering obtained with the STRUCTURE for all investigated *Alkanna* individuals.](pone.0233516.g002){#pone.0233516.g002}
According to the method described by Evanno et al. \[[@pone.0233516.ref031]\], the optimal cluster number was K = 3 ([Fig 2A and 2B](#pone.0233516.g002){ref-type="fig"}). The genetic differentiation was determined by hierarchical AMOVA and showed significant genetic differences among the populations of the three species; 71% of the variation resided within populations and showed that genetic differentiation was predominantly between individuals of each population ([Table 4](#pone.0233516.t004){ref-type="table"}). Besides, we performed genetic relationship analysis among populations with neighbor-joining criteria based on the results of Nei's genetic distance. The three *Alkanna* populations were divided into three main clusters, which corresponded to the geographic population areas ([Fig 3](#pone.0233516.g003){ref-type="fig"}).
{#pone.0233516.g003}
10.1371/journal.pone.0233516.t004
###### Analysis of molecular variance based on the ten ISSR markers for the *Alkanna* populations.
{#pone.0233516.t004g}
Source of variance df *SS* MS Variance component Percentage total (%) *F* statistic *P*
------------------------ -------- ------------- -------- -------------------- ---------------------- ----------------- -----------
**Among populations** 2 44.025 22.013 2.138 29 *F*~PT~ = 0.203 **0.001**
**Within populations** 21 110.58 5.265 5.265 71 *F*~PT~ = 0.289 **0.001**
**Total** **23** **154.583** **7.403** **100**
\* df-Degree of freedom, SS-sum of squares, MS-expected mean squares; P-value denotes the probability of null hypothesis.
3.2. Determination of alkaloid content {#sec013}
--------------------------------------
This study of the three *Alkanna* species determined eight alkaloids of pyrrolizidine type ([Table 5](#pone.0233516.t005){ref-type="table"}). Eight alkaloids were determined in *A*. *primuliflora* and *A*. *graeca*, and six alkaloids in *A*. *stribrnyi*. The main alkaloid in *Alkanna* species was triangularine (5). The *A*. *primuliflora* plants at both development stages (Aprm 1-in flowering and Aprm 2 --after flowering) and *A*. *graeca* had more complicated alkaloid profile than *A*. *stribrnyi* due to the presence of a larger number of alkaloids. The plants of *A*. *stribrnyi* in all three populations (Ast1, Ast2, Ast3) had a similar alkaloid composition. The differences were only in the population (Ast3), where alkaloid dihydroxytriangularine (7) was absent and 9-tigloylretronecine (4) was present in the alkaloid mixture ([Table 5](#pone.0233516.t005){ref-type="table"}).
10.1371/journal.pone.0233516.t005
###### Alkaloids identified from *A*. *primuliflora*, *A*. *graeca*, *and A*. *stribrnyi* from different natural populations.
{#pone.0233516.t005g}
Alkaloids/species *A*. *primuliflora* *A*. *graeca* *A*. *stribrnyi*
---------------------------------- --------------------- --------------- ------------------ ---- ---- ----
7-Angeloylretronecine (**1**) \+ \+ \+ \+ \+ \+
9-Angeloylretronecine (**2**) \+ \+ \+ \+ \+ \+
7-Tigloylretronecine (**3**) \+ \+ \+ \+ \+ \+
9-Tigloylretronecine (**4**) \+ \+ \+ \+
Triangularine (**5**) ++ ++ ++ ++ ++ ++
Triangularicine (**6**) \+ \+ \+
Dihydroxytriangularine (**7**) \+ \+ \+ \+ \+
Dihydroxytriangularicine (**8**) \+ \+ \+
+presence of alkaloid; ++ dominant alkaloid; *A*. *primuliflora*---Aprm1 (population 1, in flowering); Aprm2 (population 1, after flowering); *À*. *stribrnyi*--Ast1 (population 1, in flowering); Ast2 (population 2, in flowering), Ast3 (population 3, in flowering); *A*.*graeca*---Agr 1 (in flowering).
3.3. Embryological research {#sec014}
---------------------------
This embryological study showed the main characteristics of the structures and processes in the male and female generative sphere and revealed similarities between the three endemic *Alkanna* species.
### 3.3.1. Male generative sphere {#sec015}
The characteristics of the male generative sphere were: tetrasporangiate anthers ([Fig 4A](#pone.0233516.g004){ref-type="fig"}), four-layered anther wall developing after Dicotyledonous-type \[[@pone.0233516.ref039]\] and multilayered sporogenous tissue. The anther wall consisted of the epidermis, fibrous endothecium, one ephemeral middle layer and glandular tapetum. At the beginning of anther ontogenesis, the sporogenous cells were polygonal and fit closely with each other. Lather on, they elongate, rounded up and differentiated into microspore mother cells (MMCs) from which after simultaneous microsporogenesis predominantly tetrahedral microspore tetrads were formed. The mature pollen was 2-celled ([Fig 4B](#pone.0233516.g004){ref-type="fig"}).
{#pone.0233516.g004}
### 3.3.2. Female generative sphere {#sec016}
The female generative sphere consisted of a two-locular ovary with one anatropous tenuinnucelate ovule in each locule ([Fig 4C](#pone.0233516.g004){ref-type="fig"}). *A*fter meiosis in the ovule, one-cellulate archesporium formed hypodermal layer ([Fig 4D](#pone.0233516.g004){ref-type="fig"}). The embryo sac developed according to Polygonum---(monosporic)---type. From the specialized ovule structures in the three *Alkanna* species were established integumental tapetum and postament ([Fig 4E](#pone.0233516.g004){ref-type="fig"}). The embryo development follows the Asterad-type. The endosperm formation passes a nuclear stage and transforms into cellular at the globular embryo stage ([Fig 4F](#pone.0233516.g004){ref-type="fig"}). The mature embryo was straight and occupied ¾ parts of the seed ([Fig 4G](#pone.0233516.g004){ref-type="fig"}). A proterandry was established, when in the anthers one-celled pollen grains were formed, in the ovule macrospore mother cell was observed ([Fig 4H](#pone.0233516.g004){ref-type="fig"}). During the study, apomixis was not registered.
### 3.3.4. Pollen and seed viability {#sec017}
After the application of acetocarmine stain technique for estimation of the pollen viability, cytoplasm and nuclei of the viable pollen grains were stained in red ([Fig 4B](#pone.0233516.g004){ref-type="fig"}). The results of this study showed high viability of the mature pollen in the populations of the three *Alkanna* species: 96.4% in the population of *A*. *primuliflora*, 89.5% in the population of *A*. *stribrnyi* and 88.9% in the population of *A*. *graeca*. The tetrazolium test showed viable seeds (embryos) ([Fig 4G](#pone.0233516.g004){ref-type="fig"}) that was estimated as follows: 50% in *A*. *primuliflora* 54.2% in *A*. *stribrnyi* and 56.1% in *A*. *graeca*.
4. Discussion {#sec018}
=============
Information on the current status and the conservation of plant habitats, as well as the limited distribution of endemic species, is important for ecologists, conservationists and public institutions. In this respect, the main objective of the European ecological network European Union (EU) NATURA 2000 and Directive 92/43 / EEC on Habitats is the conservation of wild flora and fauna (<https://ec.europa.eu/environment/index_en.htm>). The initiative covers around 18% of the EU's land area, as well as 6% of the EU marine territory. Therefore, Natura 2000 is the largest coordinated network of protected areas in the world. This document and the initiatives require special measures to be taken to ensure the sustainability of the populations and conservation of species. As a basis for determining biodiversity conservation measures and strategies, it is necessary to know: (1) the biology and ecology of the species; (2) the species reproductive capacity; and (3) the species genetic diversity. The study of reproductive characteristics and abilities is essential in the case of endemic and restricted species \[[@pone.0233516.ref040]\].
Although the observed habitats of three *Alkanna* species in Bulgaria were small, fragmented, and with a limited number of individual plants, this had no effect on the reproductive structures and capacity of the plants. The peculiarities of the reproductive biology of the three species revealed in the study, characterized them as strictly sexually reproducing species (no evidence of apomictic development was observed). This study also established the stability of the processes in the male and female generative spheres such as the normal formation of male and female gametophytes without deviations and degenerative processes. Because, from the tree analysis *A*. *graeca* and *A*. *stribryni* seem closer, since they share a branch in contrast to the PCoA where *A*. *graeca* and *A*. *primuliflora* are closer, we can assume the existence of hybrid individuals. In general, we hypothesize that the opportunity for hybridization exists between the *Alkanna* species studied here. The sexual reproduction by hybridization could further contribute to the genetic diversity of the three *Alkanna* species. It also increases the species' ability to adapt to climate change. In agreement with this, there is an overlap of blooming period and pollinators in sympatric areas of studied *Alkanna* species, which potentially favour pollen exchange between species. In comparison with other *Alkanna* species e.g. *A*. *orientalis* \[[@pone.0233516.ref041]\], this study revealed average levels of genetic diversity among the natural populations of the three *Alkanna* species. The observed levels of genetic diversity between *Alkanna* individuals in this study can be considered a positive indicator for the population viability. Furthermore, the estimated high pollen and embryo viability play an important role in maintaining the size and resilience of populations of the three *Alkanna* species. The proterandry in the studied species revealed in this work is considered as one of the mechanisms for self-incompatibility \[[@pone.0233516.ref040]\]. The self- incompatibility is one of the most important means for avoiding self-pollination and self-fertilization \[[@pone.0233516.ref042]--[@pone.0233516.ref044]\] as well means for preventing inbreeding, and for promoting the generation of new genotypes in plants. It is one of the dominant reasons for the spread and success of the angiosperms on earth \[[@pone.0233516.ref042]--[@pone.0233516.ref044]\]. The revealed features of the male and female generative sphere, as tenuinucellate ovule with one-celled archesporium, differentiation of specialized structures enabling the trophic of the embryo sac (integumentary tapetum, postament) are signs of high specialization of the generative sphere \[[@pone.0233516.ref040]\]. The lack of apomixis shows an adaption to the environmental conditions that are specific for each plant species. The ability of species to reproduce asexually along with the sexually one (facultative apomicts) increases its adaptability and expands its range of distribution. Hojsgaard at al. \[[@pone.0233516.ref045]\] pointed out that the facultative apomixis in perennials provides long-term reproductive stability for the colonization of large areas. In contrast, the lack of apomixis (when the species is strictly sexually reproductive) diminishes its adaptability and makes it dependent on environmental conditions. Therefore, we consider that the lack of apomixis is one of the probable causes for adherence to particular habitats and the limited distribution of the studied species. We can assume that the observed genetic variation and the established stability of the processes in the male and female generative spheres indicate good gene flow between individuals. Despite the theoretical prediction, that fragmenting small populations of species may exhibit low levels of genetic diversity, due to fragmentations and its a negative effect on pollination and on plant reproduction \[[@pone.0233516.ref046]\]. The established moderate level of genetic variation, unusual for an endemic, indicates a lack of genetic bottleneck in this species. This result was also supported by AMOVA where a high level of intrapopulation variation was observed (Figs [1](#pone.0233516.g001){ref-type="fig"} and [2](#pone.0233516.g002){ref-type="fig"}). The genetic values found in our study might represent historic genetic diversity, derive from an ancestral population. In order of this, the size of *Alkanna* populations has been dramatically reduced in the last years, and the present habitats are highly fragmented. In all existed localities, population size is limited to several plants and the exchanges of pollen between individuals is highly restricted. This process will reduce the long term populations adaptability to present changing environmental conditions. This basic knowledge will help to determine the vulnerability and the resilience of important Bulgarian terrestrial ecosystems, with respect to global change factors such as potential increased temperatures and drought periods. More specifically, the results from this study may contribute to the development of management strategies for the protection and conservation of natural habitats such as the ecosystem community dominated by *Alkanna* species.
Pyrrolizidine alkaloids (PAs) are important secondary metabolites synthesized and accumulated in Boraginaceae, including plants in genus *Alkanna* \[[@pone.0233516.ref013],[@pone.0233516.ref047]\]. The PAs are known to be toxic compounds that can alkylate DNA and thus cause mutations and even cancer in humans and in grazing animals \[[@pone.0233516.ref014]\]. Despite the toxicity of these alkaloids, plants synthesizing and accumulating them are of industrial interest for the pharmaceutical industry for potential new drug development \[[@pone.0233516.ref048]\]. *Alkanna tinctoria* has been the most extensively studied of the species in genus *Alkanna* and it was previously reported to contain PAs (7-*A*ngeloylretronecine, 9-angeloylretronecine, 7-tigloylretronecine, 9-tigloylretronecine, 7-seneioylretronecine, 9-senecioylretronecine, triangularine, dihydroxytriangularine, triangularicine, dihydroxytriangularicine, 7-angeloyl-9-(hydroxypropenoyl) retronecine, 7-tigloyl-9-(hydroxypropenoyl) retronecine, 7-angeloyl-9-(2,3-dihydroxypropanoyl) retronecine, 7-tigloyl-9-(2,3-dihydroxypropanoyl) retronecine) \[[@pone.0233516.ref013],[@pone.0233516.ref034]\]; some of these alkaloids were found in the three endemic *Alkanna* species subject to this study. This study found eight alkaloids in *A*. *primuliflora* and *A*. *graeca* and six alkaloids in *A*. *stribrnyi* ([Table 5](#pone.0233516.t005){ref-type="table"}). The main alkaloid in all investigated species was triangularine (5). *Alkanna primuliflora* and *A*. *graeca* showed similar alkaloid composition, whereas in *A*. *stribrnyi* the alkaloids triangularicine (**6**) and dihydroxytriangularicine (**8)** were not found ([Table 5](#pone.0233516.t005){ref-type="table"}).
Because the metabolite patterns of the genus are influenced by the environmental conditions, the alkaloid profile of *A*. *stribrnyi* was investigated from three various natural localities ([Table 1](#pone.0233516.t001){ref-type="table"}). The observed differences on the alkaloid content of *A*. *stribrnyi* (Ast1, Ast2 and Ast3) are minimal, mainly concerning the alkaloids 9-tigloylretronecine (**4**) and dihydroxytriangularine (**7**) ([Table 5](#pone.0233516.t005){ref-type="table"}). There were no differences in alkaloid profile in the two stages of development of *A*. *primuliflora* (Aprm1, Aprm2) ([Table 5](#pone.0233516.t005){ref-type="table"}). The environment has a little influence on the alkaloid biosynthesis in the genus.
5. Conclusions {#sec019}
==============
Conservation of biodiversity is closely linked to the genetic diversity and reproductive capacity of species. The population genetic structure of *A*. *primuliflora*, *A*. *stribrnyi* and *A*. *graeca* showed a clear differentiation between the three species. Most probably, the genetic diversification of the populations are very high and has likely been limited by the ecological factors and habitat specialization. However, these factors have little influence on the alkaloid biosynthesis in the genus, and the three *Alkanna* species showed similar alkaloid patterns. The pyrrolizidine alkaloids were established for the first time in *A*. *primuliflora*, *A*. *graeca* and *A*. *stribrnyi*. The peculiarities of the reproductive biology of the three *Alkanna* species revealed them as sexually reproducing species and showed similarities between them. The successful development of the reproductive sphere (male and female) of studied species is the guarantee for the formation of seeds. The seed formation is very important for the dispersal of the species and the population size, and to carry out both *in-situ* and *ex-situ* conservation for these species. This study also observed the limited number of the three *Alkanna* species in their natural populations. The *in-situ* and *ex-situ* activity are important actions for the conservation of the endangered plant species. The authors will make the results available to the Ministry of Environment and Water of Bulgaria. These endemic species and their populations need to be considered for protection status.
The authors are grateful for the in-kind support provided by the Bulgarian Academy of Sciences in Sofia and by the Agricultural University in Plovdiv, Bulgaria.
Aprm1
: *A*. *primuliflora* in flowering
Aprm2
: *A*. *primuliflora* after flowering
Ast1
: *A*. *stribrnyi*
GC/MS
: Gas chromatography--mass spectrometry
ISSR
: Inter-simple sequence repeats
PAs
: pyrrolizidine alkaloids
TLC
: thin layer chromatography
[^1]: **Competing Interests:**NO authors have competing interests.
| {
"pile_set_name": "PubMed Central"
} |
1.. Introduction
================
Traumatic brain injury (TBI) triggers a cascade of changes inducing the development of secondary brain damage with nitric oxide (NO) playing an important role. NO is a strong vasodilator and regulates cerebrovascular tone and perfusion \[[@b1-ijms-15-04088]\]. Therefore, it is not surprising that NO is one of the key players in the development of cerebral vasospasm after traumatic and aneurysmal subarachnoid hemorrhage (SAH) as well as in the development of secondary brain damage after severe TBI \[[@b2-ijms-15-04088],[@b3-ijms-15-04088]\].
NO is synthesized from [l]{.smallcaps}-arginine by a family of NO-synthases (NOS): Two constitutive forms (cNOS) have been described, the endothelial NOS (eNOS) and the neuronal NOS (nNOS), as well as an inducible form (iNOS) \[[@b4-ijms-15-04088]\]. The synthesis of NO via NOS can be inhibited by the endogenous competitive NOS-inhibitor asymmetric methylated [l]{.smallcaps}-arginine (ADMA), which is synthesized by protein-arginine methyltransferase 1 (PRMT1) and hydrolyzed by dimethylarginine dimethylaminohydrolase (DDAH1 and DDAH2) \[[@b5-ijms-15-04088],[@b6-ijms-15-04088]\]. After TBI, an early decrease in intracerebral NO concentration occurs, which lasts from 30 min up to 6 h after TBI \[[@b2-ijms-15-04088]\]. During this time, eNOS and nNOS activity appear normal and unchanged after TBI in humans \[[@b7-ijms-15-04088]\]. The administration of exogenous NOS inhibitors leads to controversial effects \[[@b2-ijms-15-04088],[@b8-ijms-15-04088]--[@b10-ijms-15-04088]\]. Furthermore, administration of [l]{.smallcaps}-arginine resulted in reduced contusion volumes and had beneficial effects on cerebral blood flow \[[@b11-ijms-15-04088]--[@b15-ijms-15-04088]\]. Martens-Lobenhoffer *et al.* hinted at the endogenous NOS-inhibitor ADMA after TBI. They described an increase of ADMA levels in the cerebro-spinal fluid of patients with traumatic subarachnoid hemorrhage \[[@b16-ijms-15-04088]\] parallel to the time-course of blood-brain barrier breakdown after TBI \[[@b17-ijms-15-04088]\]. Variations in NO availability and controversial effects of exogenously administered NOS-inhibitors after TBI may be evoked by the competitive endogenous NOS inhibitor ADMA.
We, therefore, hypothesized that ADMA and the enzymes involved in ADMA-metabolism might be involved in TBI. Therefore, our goals were (1) to determine whether ADMA is expressed in brain tissue after TBI, focusing on the damaged brain site (TBI lesion) and the potentially influenceable penumbra (perilesional zone); (2) to establish whether time-dependent changes of ADMA expression exist; and (3) to investigate if the expression of enzymes involved in NO- and ADMA-synthesis might also be affected.
2.. Results and Discussion
==========================
2.1.. Neuroscore
----------------
Motor function performance after controlled cortical impact injury (CCII) is known to drop immediately after injury. In previous studies using a CCII rat model the impairment of motor function was found to be maximal during the first two days after injury \[[@b18-ijms-15-04088]--[@b20-ijms-15-04088]\]. We examined Garcia neuroscore performance \[[@b21-ijms-15-04088]\] of control animals (*n* = 5) and of TBI animals at different time-points after CCII (*n* = 6; at each time point). Due to the experimental set-up no neurological testing could be performed within 30 min after CCII. All control animals reached the maximum attainable score of 18 points (100% level). In accordance with previous reports \[[@b18-ijms-15-04088]--[@b20-ijms-15-04088]\] we find a significant (*p* \< 0.01) and immediate drop of 33% in performance over the first 2 days after TBI (13 ± 2 Garcia points; *n* = 24) compared to controls (18 Garcia points, *n* = 5).
2.2.. ADMA, PRMT, DDAH and NOS Expression
-----------------------------------------
In control/sham animals which did not suffer from TBI, ADMA is strongly and evenly expressed in the supratentorial brain regions (35 ± 7 positively stained cells/FOV). After TBI ADMA expression decreases in all TBI animals compared to controls. However, significant differences can be found within the different brain regions (contusion/TBI lesion and penumbra/perilesional zone) ([Figures 1](#f1-ijms-15-04088){ref-type="fig"} and [2](#f2-ijms-15-04088){ref-type="fig"}). Thirty minutes after TBI, ADMA expression strongly decreases within the TBI lesion (*p* \< 0.001) and remains low during the subsequent course while perilesional ADMA expression shows a decelerated decline ([Figure 1](#f1-ijms-15-04088){ref-type="fig"}). Early after TBI (30 min) the perilesional ADMA expression remains at 24 ± 8 positively stained cells/FOV. Then 3 to 8 h after TBI perilesional ADMA expression significantly decreases (*p* \< 0.01) compared to the early phase (30 min), followed by an increase, reaching 24 h after TBI a second peak (*p* \< 0.01) compared to the earlier values 3 and 8 h after TBI ([Figures 1](#f1-ijms-15-04088){ref-type="fig"} and [2](#f2-ijms-15-04088){ref-type="fig"}). Analyzing the time course of ADMA expression with respect to the Garcia neuroscore performance, we found a significant correlation in this perilesional area (cc = 0.46; *p* = 0.01; 95% CI: 0.12--0.7).
PRMT1, the enzyme which synthesizes ADMA by post-translational methylation of protein-bound [L]{.smallcaps}-arginine \[[@b22-ijms-15-04088]\] can be detected in control animals ([Figure 3](#f3-ijms-15-04088){ref-type="fig"}). Control animals showed higher PRMT1 expression (20 ± 5 positively stained cells/FOV) than the animals submitted to TBI (range: 0--8 positively stained cells/FOV depending on the time-point after TBI) (*p* \< 0.01). Early after TBI (30 min) PRMT1 expression disappears within the lesion. Then, three to eight hours later, some isolated scattered cells express PRMT1 ([Figures 3](#f3-ijms-15-04088){ref-type="fig"} and [4](#f4-ijms-15-04088){ref-type="fig"}). Forty-eight hours after TBI, PRMT1 expression increases especially within the lesion but also in the perilesional area compared to earlier time-points after TBI (*p* \< 0.001) ([Figure 3](#f3-ijms-15-04088){ref-type="fig"}).
DDAH1 and DDAH2 are expressed in control and TBI animals. Especially in control animals and early after TBI, DDAH 1 expression is scattered and low (0.7 ± 1, range: 0--3 positively stained cells/FOV). Within 24 h after TBI, DDAH1 expression increases significantly inside the lesion and perilesionally to a peak (5.5 ± 3.1 and 2.8 ± 2.5 positively stained cells/FOV, *n* = 6, *p* \< 0.01), which is also significantly higher than in the controls (*p* \< 0.01) ([Figures 5](#f5-ijms-15-04088){ref-type="fig"} and [6](#f6-ijms-15-04088){ref-type="fig"}). DDAH2 expression after TBI is increased and stays elevated compared with controls ([Figure 7](#f7-ijms-15-04088){ref-type="fig"}). However, no significant changes can be observed within the lesion or perilesional area at the different time points (*p* \> 0.05). DDAH expression is low (0 to 8 cells/FOV) and remarkably individual and variable resulting in strong standard-deviations at the different time-points after TBI ([Figure 8](#f8-ijms-15-04088){ref-type="fig"}). This might be related to the individual development of secondary brain injury and neuroscore performance. Corresponding to this, analyzing the individual DDAH expression of each animal at the different time-points with respect to the Garcia neuroscore performance ([Figure 9](#f9-ijms-15-04088){ref-type="fig"}), we found that in the perilesional area DDAH1 and DDAH2 expression were negatively correlated (cc = −0.42; *p* = 0.02; 95% CI: (−0.67)--(−0.8) and cc = −0.41; *p* = 0.02; 95% CI: (−0.67)--(−0.07), respectively).
In this study, a significant increase in iNOS expression, which starts 8 h after TBI can be observed ([Figure 10](#f10-ijms-15-04088){ref-type="fig"}). Comparable time-dependent and delayed increases in iNOS expression have been described previously \[[@b23-ijms-15-04088]--[@b25-ijms-15-04088]\]. eNOS expression significantly decreases 30 min to 3 h after TBI and subsequently regains control/sham levels 8 h after TBI ([Figure 10](#f10-ijms-15-04088){ref-type="fig"}). Comparable with these findings, Wada *et al.* reported after an immediate increase a sustained reduction in cNOS activity \[[@b24-ijms-15-04088]\]. In humans, Gahm *et al.* could detect an iNOS peak 8 to 23 h after TBI, while eNOS and nNOS remained unchanged compared to controls \[[@b7-ijms-15-04088]\]. Furthermore, Gahm *et al.* reported decreased nNOS expression after CCII in rats \[[@b23-ijms-15-04088]\]. As iNOS generates and emits more NO than cNOS \[[@b26-ijms-15-04088]\], iNOS derived NO has been suggested as a potential mediator of secondary brain injury \[[@b2-ijms-15-04088],[@b25-ijms-15-04088]\]. NO by eNOS plays a major role in maintaining cerebral blood flow in the perilesional zone, which is on the one hand susceptible to secondary brain injury and on the other a potential target for treatment. Therefore, we focused on eNOS and iNOS expression.
As NO was not directly measured in these CCII animals, we calculated a ratio between NOS and ADMA, as indirect measure for NO availability: (eNOS + iNOS expression)/ADMA expression. In the perilesional area NOS/ADMA ratio significantly increases three hours after TBI (*p* \< 0.01), which indicates, in accordance with previous reports \[[@b2-ijms-15-04088],[@b15-ijms-15-04088],[@b23-ijms-15-04088],[@b24-ijms-15-04088]\], a delayed increase of NO in this region, which is likewise caused by an increase in iNOS expression \[[@b2-ijms-15-04088],[@b15-ijms-15-04088],[@b23-ijms-15-04088],[@b24-ijms-15-04088]\] ([Figure 10](#f10-ijms-15-04088){ref-type="fig"}).
Reduced NO levels, as found early after TBI, in spite of unchanged or normal NOS activity can be explained if endogenous NOS inhibitors as ADMA play a role. Conversely, increased NO levels after TBI may be the result of increased NOS expression and reduced endogenous NOS inhibition, comparable to the observed delayed iNOS increase which was accompanied by an increase in DDAH and an overall decrease in ADMA compared to controls.
ADMA expression is detected in brain tissue of both controls and of animals which suffer TBI. Although ADMA expression strongly decreases within the traumatic lesion, it remains unchanged perilesionally during the early stage after TBI, followed by a decrease 3 h after TBI and a subsequent increase about 24 h after TBI. ADMA protein levels were not measured in this study. Martens-Lobenhoffer *et al.* described in a previous study increased ADMA protein levels in the cerebro-spinal fluid (CSF) and plasma of patients with traumatic subarachnoid hemorrhage after TBI \[[@b16-ijms-15-04088]\]. Similarly increased ADMA levels accompanied by decreased NO levels in CSF and NOS inhibition have been demonstrated in cases of brain injury caused by aneurysmal subarachnoid hemorrhage \[[@b27-ijms-15-04088],[@b28-ijms-15-04088]\]. Furthermore, Thampatty *et al.* reported elevated CSF ADMA level within 3 days after TBI in children \[[@b29-ijms-15-04088]\]. The delayed detection of elevated ADMA levels in CSF and serum parallels the timescale for apoptosis and neuronal cell-death as well as the time-course of blood brain-barrier breakdown after TBI \[[@b17-ijms-15-04088],[@b30-ijms-15-04088]\]. Therefore, the observed overall decrease in tissue ADMA expression, especially within the necrotic TBI lesion, compared to controls as well as the increase of ADMA expression within the perilesional zone 24 h after TBI in this study, which are also parallel to the described time-course of blood-brain barrier breakdown \[[@b17-ijms-15-04088]\] do not seem to be conflicting. On the contrary, it underlines once more the variety of pathological and time-dependent changes observed after TBI. Furthermore, former experimental TBI studies in which NOS inhibitors were administered lead to controversial effects: In one study, which used the fluid percussion injury model, administration of [l]{.smallcaps}-NAME 5 min prior to injury resulted in an increased mortality rate \[[@b8-ijms-15-04088]\]. In another study, administration of [l]{.smallcaps}-NAME pre- and post injury showed no adverse effects on outcome \[[@b31-ijms-15-04088]\]. While some studies showed no effect, others demonstrated a reduction in edema, contusion volume and improvement in outcome \[[@b9-ijms-15-04088],[@b10-ijms-15-04088],[@b31-ijms-15-04088]\]. The differing benefits achieved with administered NOS inhibitors become comprehensible when one considers the different time-points of administration together with the observed undulated course of ADMA and DDAH after TBI in our study. In addition, [l]{.smallcaps}-arginine, the substrate for NO production, also represents an ADMA competitor for NOS. In accordance with treatment successes reported on cerebral vasospasm after SAH \[[@b32-ijms-15-04088],[@b33-ijms-15-04088]\], the administration of [l]{.smallcaps}-arginine resulted in a reduction of contusion volume after TBI \[[@b11-ijms-15-04088]--[@b15-ijms-15-04088]\]. [l]{.smallcaps}-arginine administration increased NO production \[[@b34-ijms-15-04088]\] and, furthermore, it restored cerebral blood flow in various cortical contusion models of traumatic brain injury \[[@b12-ijms-15-04088],[@b13-ijms-15-04088]\]. These observations suggest that [l]{.smallcaps}-arginine may overcome local competitive inhibition of NOS by ADMA after TBI and supports our hypothesis that ADMA is involved in the development of secondary brain damage, especially within the perilesional area surrounding the TBI lesion.
PRMT1, the enzyme which synthesizes ADMA by post-translational methylation of protein-bound [l]{.smallcaps}-arginine \[[@b22-ijms-15-04088]\] can be detected in control animals, where it is evenly expressed in cortical neurons comparable with previous reports on mouse brain sections \[[@b35-ijms-15-04088]\]. PRMT1 is significantly less expressed after TBI than in controls: Early after TBI (30 min) PRMT1 expression disappears within the lesion. Then, three to eight hours later, some isolated scattered cells express PRMT, reaching a maximum 48 h after TBI within the lesion but also in the perilesional area.
DDAH1 and DDAH2 hydrolyze ADMA. DDAH1 is mainly expressed in neuronal tissue \[[@b36-ijms-15-04088],[@b37-ijms-15-04088]\] while DDAH2 is found perivascular similar to the eNOS expression \[[@b38-ijms-15-04088]\]. In this study, however, no co-immunoprecipitation studies were performed. Both enzymes are expressed in control and TBI animals. Results on DDAH1 and DDAH2 in control brains are comparable to previous reports \[[@b39-ijms-15-04088]\]. Within 24 h after TBI, DDAH1 expression increases significantly inside the lesion and perilesionally to a peak and is significantly higher than in the controls. DDAH2 expression after TBI is increased and stays elevated after TBI compared with controls. Furthermore, a negative correlation between DDAH1 and DDAH2 expression and neuroscore performance could be detected in the perilesional area.
Two different mechanisms may raise ADMA concentrations: (1) increased methylation of [l]{.smallcaps}-arginine by up-regulation of PRMT1 and (2) decreased hydrolysis of ADMA by DDAH. In hypercholesterolemia an up-regulation of PRMT mRNA expression has been reported in human endothelial cells by low-density lipoproteins (LDL) and suggested to be responsible for ADMA elevation \[[@b40-ijms-15-04088]\]. While control animals showed strong PRMT1 expression, TBI resulted in reduced PRMT1 expression. We observed an early and strong decrease followed by an increase in PRMT1 expression within the injured hemisphere peaking 48 h after TBI. This was predominant within the lesion but also detectable in the perilesional area. The second mechanism, a decline in DDAH activity, occurs under pathological conditions such as cerebral vasospasm after subarachnoid hemorrhage and hypercholesterolemia \[[@b27-ijms-15-04088],[@b41-ijms-15-04088]\]. After experimental subarachnoid hemorrhage DDAH2 expression was attenuated in cerebral vessels being affected by vasospasm \[[@b27-ijms-15-04088]\]. In this traumatic brain injury study DDAH1 and DDAH2 expression significantly increased within the lesion and perilesionally, accompanied by a decrease in ADMA which could not be overcome by elevated PRMT1 expression 48 h after TBI. Therefore, both mechanisms may contribute to the overall effect.
Although we did detect a general decrease of ADMA after TBI, we cannot exclude local changes in ADMA expression with increased local levels of ADMA especially in cases of traumatic vasospasm after severe TBI. This is important, since distinct local and temporal profiles of cerebral blood volume may also contribute to secondary changes and functional outcome after TBI \[[@b2-ijms-15-04088]\]. In the vulnerable perilesional area they seem to be potentially associated with the time course of sensory-motor deficit \[[@b42-ijms-15-04088]\]. Similar, we found in our study a significant association between neuroscore and ADMA expression and a negative correlation of the neuroscore with DDAH1 and DDAH2 in the perilesional area.
DDAH inhibition and activation have been suggested as useful molecular probes to better understand cellular regulation of nitric oxide. Several potent DDAH inhibitors have recently been published \[[@b43-ijms-15-04088]\] and suggested as potential therapeutic agents for treatment of pharmacological states associated with inappropriate under- and over-production of nitric oxide, such as septic shock \[[@b44-ijms-15-04088]\]. Therefore, ADMA and especially DDAH changes in the perilesional area might offer potential new therapeutic targets in treatment of TBI and secondary brain damage development.
3.. Experimental Section
========================
3.1.. Animals
-------------
Male Sprague-Dawley rats (*n* = 35) (Charles River, Germany) weighing 260 to 370 g were used in the present investigation. All experimental procedures were reviewed by the institutional committee for animal care and approved by the local veterinary authority (Regierungspräsidium Karlsruhe, Germany, G147/09).
3.2.. Anesthesia and Brain Injury Model
---------------------------------------
Rats were anaesthetized in an isoflurane chamber (5%). Thereafter, anesthesia was maintained with a face mask using 1.5%--1.8% isoflurane, 30% O~2~, and 68% N~2~O. Via a catheter in the tail artery systemic blood pressure and blood gases were monitored. Body temperature was kept constant at 37.0 °C by a feedback controlled heating pad. After induction of anesthesia, the head was fixed in a stereotactic frame and a right-sided craniotomy was performed between the coronal, sagittal and lambdoid sutures with a micro-drill and under permanent cooling with saline. Specific attention was paid to leave the dura mater intact. Controlled cortical impact injury (CCII) was performed perpendicular to the surface of the brain as previously described \[[@b45-ijms-15-04088]\]. The diameter of the impactor was 5 mm, the velocity 7.5 m/s, 1.5 mm depth of impression and the impact duration 300 ms. Craniotomy was closed immediately after CCII by replacing the bone-flap and fixation with dental cement.
3.3.. Experimental Groups
-------------------------
One experimental group (*n* = 30) and one control group (*n* = 5) were investigated. All animals (experimental and control group) received craniotomy while rats of the control group were randomized to a sham group without receiving trauma/CCII. Animals in the experimental group were subjected to CCII and sacrificed after perfusion 30 min, 3, 8, 24 and 48 h after CCII (*n* = 6, at each time-point). Control animals were sacrificed directly after sham surgery.
3.4.. Immunohistochemistry
--------------------------
After perfusion animal supratentorial brains intended for immunohistochemical staining were dissected out of the skull and snap frozen in 2-methylbutan (AppliChem, Darmstadt, Germany) and stored at −80°C for prompt cryosectioning. For immunostaining of different antibodies adjacent sections of 6 μm were used. On one object slide 2 brain sections with a minimum distance of 48 μm were mounted and both were stained with one of the following antibodies: Rabbit polyclonal antibody against endothelial NOS (eNOS), the inducible NOS (iNOS), ADMA and PRMT1 as well as goat polyclonal antibody against DDAH1 and DDAH2 were used with a 1:50 dilution for eNOS, iNOS, PRMT1, DDAH1 and DDAH2 (Abcam, Cambridge, UK). The antibody against ADMA was diluted 1:100. Sections without primary antibodies were used as negative controls in each staining set. Binding was visualized by a biotin conjugated secondary donkey immunoglobulin G antibody *versus* rabbit/goat immunoglobulin G antibody (Abcam, Cambridge, UK), adding Avidin/Biotin-complex (ABC) as well as 3,3′-Diaminobenzidine (DAB) as chromogen. Counterstaining was performed with hematoxylin. Distribution of immunohistologically stained cells within different brain areas (lesion, and perilesional zone) (e.g., [Figure 2A](#f2-ijms-15-04088){ref-type="fig"}) were analyzed at different time-points following CCII. In control animals the corresponding homotopic regions below the craniectomy were examined. As no difference could be found for any antibody staining among the different control regions their results were combined. Immunohistochemical staining was analyzed quantitatively by two independent observers who were blinded to the previous treatment protocols of animals: In two sections and in six fields of view (FOV) at a high-power magnification of 400×, the number of positively stained cells was counted and a mean value was calculated for each group and for the different brain areas at different time points (30 min, 3, 8, 24 and 48 h) following CCII. Furthermore, a ratio between expression of NO synthases and the NOS inhibitor ADMA was calculated as an indirect measure for NO availability, called in the following "NOS/ADMA ratio": (eNOS + iNOS expression)/ADMA expression. For interobserver agreement, κ-statistics were performed, revealing excellent agreement (κ = 0.95).
3.5.. Neurological Assessment
-----------------------------
The neurological examination was performed referring to the modified grading system by Garcia *et al*. \[[@b21-ijms-15-04088]\]. The Garcia test is a composite neurological test in which the rats are evaluated for various sensorimotor deficits: (1) spontaneous activity; (2) symmetry in four limb movement; (3) forepaw outstretching; (4) climbing; (5) body proprioception; and (6) response to vibrissal touch. For each mode of motor or behavioral deficit between 0 and 3 points were allocated, leading to a cumulative maximum score of 18 points as found in normal, healthy animals. Neurological deficits lead to lower scores ([Figure 3](#f3-ijms-15-04088){ref-type="fig"}) \[[@b21-ijms-15-04088]\]. Garcia neuroscore performance was tested in all control animals (*n* = 5). TBI animals were examined before they were sacrificed at 3, 8, 24 and 48 h after CCII (*n* = 24; 6 animals at each time point). Due to the experimental set-up no neurological testing could be performed within 30 min after CCII (*n* = 6).
3.6.. Statistics
----------------
Data are presented as mean value ± standard deviation (SD). A threshold value of significance (*p*-value) of less than 0.05 was applied. For statistical comparisons between groups, two-tailed Student's *t*-tests for normally distributed data were used. Analysis of the normally distributed variables at the different time points was performed using one-way analysis of variance test (ANOVA) followed by multiple comparison post test (Tukey's test and Dunnett's test for comparison of multiple groups with one control group). Pearson's correlation coefficient served to assess correlation. All calculations were performed with a standard statistical software package (GraphPad InStat, Version 3.05, GraphPad Software, Inc., La Jolla, CA, USA).
4.. Conclusions
===============
This is, to the best of our knowledge, the first study that describes temporal profiles of NOS, ADMA, PRMT1, DDAH1 and DDAH2 in the brain after experimental contusion/TBI. The main findings of our experiments are: (1) ADMA is expressed in brain sections of animals after suffering TBI as well as in controls; (2) TBI changes ADMA expression strongly: ADMA expression decreased in all time-groups after TBI compared to the control group. However, within the lesion ADMA expression drops immediately and stays low, while the level in the perilesional area decreases 3 h after TBI and increases 24 h after TBI compared to previous time-groups, but remains at all times lower than in controls; (3) PRMT1 also decreased in all animals after TBI compared to the control group, while; (4) DDAH1, and DDAH2 increased in comparison to controls; (5) In the perilesional area ADMA is positively correlated with neuroscore performance; and (6) DDAH1 and DDAH2 are negatively correlated with neuroscore performance.
Although, ADMA and the enzymes involved in ADMA-metabolism show significant changes after TBI, future experiments have to prove if the competitive endogenous NOS inhibitor ADMA effects variations in NO availability and previously described controversial effects of exogenously administered NOS-inhibitors after TBI. Nevertheless, ADMA and its metabolizing enzymes might be new targets in TBI treatment.
This work was supported by a stipend of the German society of neurological surgeons (Stipendium Neurochirurgische Forschung der Deutschen Gesellschaft für Neurochirurgie (DGNC)) awarded to C.S. Jung. We thank Rolf Warta for technical assistance in performing photography of immunostaining.
The authors declare no conflict of interest.
{#f1-ijms-15-04088}
{#f2-ijms-15-04088}
{#f3-ijms-15-04088}
{#f4-ijms-15-04088}
{#f5-ijms-15-04088}
{#f6-ijms-15-04088}
{#f7-ijms-15-04088}
{#f8-ijms-15-04088}
{#f9-ijms-15-04088}
{#f10-ijms-15-04088}
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#s0005}
===============
The development of appropriate patient selection criteria for tumor specific treatment including chemo- and radiotherapy (RT) is the cornerstone of modern precision oncology. With regard to life-threatening conditions such as brain-disseminated cancer it is widely believed and anticipated that aggressive antitumor treatment should be started as soon as possible to prolong survival and maintain quality of life. However, it has long been recognized that some patients will not derive any profit from active treatment while other will do. In order to discriminate those patients potentially benefitting from antitumor therapy, several prognostic scores such as the RTOG Recursive Partitioning Analysis (RPA) [@b0005] score or the Graded Prognostic Assessment (GPA) score [@b0010] were developed and validated [@b0015], [@b0020]. Although offering a possibility to separate the survival curves, i.e., identifying patient groups with better and worse prognosis, these models have not sufficient predictive ability for deciding which patients will not derive sufficient benefit from brain RT, which is important within the context of counselling patients about their prognosis and treatment options. It is known that several widely available and cheap blood tests such as albumin and lactate dehydrogenase (LDH) can be used as surrogate parameters in survival prediction [@b0025], [@b0030]. The number of extracranial organ systems involved has also been found to be an independent and highly significant predictor of overall survival [@b0035], [@b0040]. The combination of those three parameters was used by Nieder et al. to develop a simple extracranial score (EC-S) as a reasonable addition to the aforementioned intracranial prognostic factors helping to identify the patients with brain metastases with very bad prognosis in whom best supportive care could be the best choice [@b0045]. The main goal of this study was an independent validation of the EC-S as a possible tool for predicting very limited survival and to compare it with the RPA and disease-specific GPA (DS-GPA) scores.
2. Materials and methods {#s0010}
========================
We extracted all available patient records referring to a diagnosis of brain metastases from the clinical RT software Mosaiq© from the Department of Radiation Oncology at the University hospital Wuerzburg. Overall survival (OS) data from 524 patients treated between 04.02.2008 and 08.11.2017 were available. Brain MRI was obligatory as a part of staging for patients with primarily non-metastasized lung cancer and malignant melanoma. In other cases the clinical symptoms of brain metastases prompted cranial imaging. In our clinic the vast majority of tumor patients receive their treatment recommendation after discussion in interdisciplinary tumor boards. Further diagnostic work-up was performed according to the advice of the tumor board and based basically on national guidelines. Prognostic scores for each patient were determined as originally described [@b0005], [@b0010]. For calculating the EC-S one point was counted for each elevated LDH, decreased albumin and more than one extracranial site of metastatic involvement, so that the final score ranged from 0 to 3 (3 indicating the worst prognosis) [@b0030].
Elevated LDH was defined as above 250 U/l, and decreased albumin was defined as below 3.5 g/l according to the normal levels of the local laboratory. LDH and albumin measurements were only considered if taken within 2 weeks before the first fraction of RT. Since both blood tests are not mandatory in our radiotherapy department, only 157 out of 524 extracted patient records contained information on all 3 extracranial prognostic parameters. In contrast, the RPA score could be computed for all, and DS-GPA score for 466 patients. Only 134 cases had both DS-GPA and EC-S available. An overview of all relevant patient characteristics is given in [Table 1](#t0005){ref-type="table"}.Table 1Patient characteristics (n = 524). Abbreviations used in the table: renal cell carcinoma (RCC), Karnofsky performance score (KPS), whole brain radiotherapy (WBRT), radiotherapy (RT).ParameterN%(whole)EC-S available(whole)EC-S availableGender Female242644641 Male282935459
Age (years)Median 63Median 63Range (20--92)Range (21--86)
Primary Lung270875155 Breast667135 RCC22745 Melanoma74341422 Head and neck6010 GI37775 CUP19644 other30966
KPS KPS \>70%228704344 KPS ≤70%297875756
RPA class 1133342522 2262865055 3129372524
DS-GPA class 1190614143 2154383327 392262018 42816611
EC-S 01510 16944 26038 3138
Molecular target with therapeutic relevance present Yes97301919 No4241268181
Primary controlled Controlled305835853 Not controlled219744247
Number of brain metastases One143402726 Two or three107342022 Multiple274835253
Extracranial metastases No101231915 Single organ152402926 Multiple271945260
Treatment modality Surgery + adjuvant RT or stereotactic radiosurgery103512032 WBRT and others4211069368
Actuarial survival from the first day of whole-brain (WB) or other RT was calculated using the Kaplan-Meier method and compared between different groups with the Log-rank test. Discrimination of each score was assessed by Gönen & Heller's concordance probability estimate (CPE) for the Cox model [@b0050]. The concordance probability is a general measure of discriminatory power of a nonlinear statistical model, with a probability of 0.5 indicating random discrimination and 1 perfect discrimination. The calibration (external validity) was checked by comparing median survival estimates of each risk group with the corresponding values of the datasets from which the scores were derived [@b0055]. To determine the most important prognostic factors in multivariable analysis, Cox regression was used. The following covariates from [Table 1](#t0005){ref-type="table"} were judged as putatively important prognostic factors: treatment modality (surgery + adjuvant RT or stereotactic radiosurgery/WBRT and others), age (\<65/≥65 years as used in the RPA score), gender, baseline Karnofsky performance score (KPS; ≤70/\>70), primary tumor type (8 strata in total), presence of a molecular target with therapeutic relevance (yes/no), the number of extracranial organs involved (0/1/\>1), number of brain metastases (1/2 or 3/\>3), serum albumin (normal/decreased), LDH (normal/elevated), RPA score, DS-GPA score and EC-S. Those variables were available for 140 patients of which 119 had died. Given the large number of variables compared to the number of events, we conducted variable selection using the LASSO method which shrinks regression coefficients of less important variables to 0 and typically yields lower estimation variance than stepwise selection methods [@b0060]. The optimal penalty parameter λ was determined based on 10-fold cross validation and used for determining the most important predictor variables. These selected variables were then used to build a new predictive model. We adhered to the TRIPOD criteria to assure the transparence of our data presentation and analysis [@b0065]. Statistical analysis was performed with IBM-SPSS-25© and R version 3.5.0.
3. Results {#s0015}
==========
Most patients in the whole cohort had multiple brain metastases (52%) and multiple extracranial metastases (52%) ([Table 1](#t0005){ref-type="table"}). The median KPS was 80, range 30--100. The most frequent primary tumor was lung cancer (51.3%), followed by malignant melanoma (14.3%) and breast cancer (12.7%). Albumin and LDH measurements prior to RT were available for 165 and 260 patients respectively, and 157 patients had both proteins measured. 132 events were registered in this latter group and 25 cases were censored. We used the DS-GPA (available for 464 patients with 410 events and 54 censored cases) and RPA scores (available for 524 patients with 464 events and 60 censored cases) as established reference to compare with EC-S. The results are summarized in [Table 2](#t0010){ref-type="table"}, and [Fig. 1](#f0005){ref-type="fig"} and [Fig. 2](#f0010){ref-type="fig"} show the Kaplan-Meier survival curves for patients stratified according to the DS-GPA score and EC-S, respectively. Significant survival differences between all groups within each prognostic score were seen except for groups 0 and 1 of the EC-S (p = 0.974) which had similar median survival estimates ([Table 2](#t0010){ref-type="table"} and [Fig. 2](#f0010){ref-type="fig"}). However, the EC-S had the best discriminatory power as judged by Gönen & Heller's CPE. The EC-s was also the best score for discriminating patients with a particularly poor prognosis, since its worst prognostic class possessed the largest hazard ratio (6.36) compared to the most favorable class. A comparison between the median survival estimates of our cohort and those of the datasets from which the different scores had been derived revealed differences between the survival predictions for the derivation datasets and our data. In this respect, the predictions for classes 0, 2 and 4 of the EC-S appeared to be the best calibrated.Table 2Results concerning the calibration and discrimination of the three scores applied to our dataset. Gönen & Heller's CPE is an estimate of the concordance probability of the Cox models. Abbreviations used in the table:Confidence interval (CI), hazard ratio (HR), overall survival (OS).RPADS-GPAEC-S1233.5--4.02.5--3.01.5--2.00--1.00123N133262129289215419015696013\# events10823212417751351839565413Median OS8.385.161.7712.917.857.362.147.98.82.00.6Median OS 95% CI6.80--10.413.91--6.701.38--2.279.56-NA6.74--14.265.16--9.361.81--2.863.9--125.7--10.91.4--3.90.1--1.1HR11.472.8611.642.264.1211.052.316.36HR SE0.121.140.270.260.260.360.360.44HR 95% CI1.16--1.852.19--3.730.97--2.781.36--3.752.50--6.810.52--2.141.14--4.692.67--15.14Median OS in derivation dataset7.14.22.316.7 (14.7--18.8)9.6 (8.7--10.6)5.4 (4.9--5.9)3.1 (2.8--3.5)9.03.52.30.7
Gönen & Heller's CPE0.5938 ± 0.01160.6110 ± 0.01230.6258 ± 0.0220Fig. 1Actuarial Kaplan-Meier survival plot for patients with available DS-GPA-class (n = 466). Differences between groups significant in pairwise comparison (Log rank, p \< 0.05).Fig. 2Actuarial Kaplan-Meier survival plot for patients with available EC-S (n = 157). Difference between all but groups 0 and 1 were significant in pairwise comparison (Log rank, p \< 0.05).
Using the LASSO method to build a multivariable prognostic Cox model from our own data (140 patients and 119 events), the following variables were selected: Treatment modality, Albumin, LDH, DS-GPA score and EC-S. Using 5- or 20-fold instead of 10-fold cross validation for finding the optimal LASSO penalty parameter did not change this variable selection result. The regression coefficients of the final model fitted with these variables are given in [Table 3](#t0015){ref-type="table"}. The CPE of the final Cox model was 0.7230 ± 0.0212, indicating a significant increase in discriminatory power compared to every score on its own ([Table 2](#t0010){ref-type="table"}).Table 3Prognostic factors and regression coefficients in the final Cox model obtained after LASSO variable selection.CoefficientHazard ratiop-ValueAlbumin \<3.5 g/l0.677 ± 0.4031.970.093LDH \<250 U/l−0.072 ± 0.5700.930.900Treatment: Whole brain RT and no (radio-)surgery0.448 ± 0.2461.570.068DS-GPA: 2.5--3.01.295 ± 0.4883.650.008DS-GPA: 1.5--2.01.590 ± 0.4834.900.001DS-GPA: 0.5--1.02.181 ± 0.5098.861.8 × 10^−5^EC-S: 1−0.733 ± 0.4490.480.1979EC-S: 20.438 ± 0.7241.550.545EC-S: 30.297 ± 0.9961.350.766
Given the selection of both DS-GPA and EC-S into the final Cox model, we heuristically combined both scores into a new "EC-GPA" score with 4 categories ([Table 4](#t0020){ref-type="table"}). Stratification of the 134 patients for which the new EC-GPA combination could be calculated resulted in a clear separation of the survival curves ([Fig. 3](#f0015){ref-type="fig"}). The hazard ratios of EC-GPA classes 1, 2 and 3 compared to class 0 were 4.84 (95% CI 1.42--16.46), 9.26 (2.83--29.97) and 31.64 (8.36--119.76). Furthermore, with a CPE = 0.6355 ± 0.0230 the discrimination between the four prognostic groups was slightly increased compared to the EC-S alone (CPE = 0.6258 ± 0.0220).Table 4Calculation of the combined EC-DS-GPA score.DS-GPAEC-S012312223222233111340013Fig. 3Actuarial Kaplan-Meier survival plot for patients with ES-GPA-score (n = 134). Difference between all groups significant in pairwise comparison (Log rank, p \< 0.05).
4. Discussion {#s0020}
=============
A problem of proper patient and/or therapy selection has not lost its importance since the beginning of local therapy of brain metastasis. Selecting patients who have a chance to derive any benefit from antitumor treatment is a prerequisite for omitting unneeded treatment of those who have very limited survival prognosis.
Some discouraging results of more aggressive local treatment of brain metastases [@b0070], [@b0075] challenged clinicians and statisticians. Several methodologies on the way to prognostic tools for discrimination of patient populations benefiting from more aggressive treatment were developed [@b0010], [@b0080] and validated [@b0085], [@b0090], [@b0095].
Despite the validation of the RPA score we believe that it is not helpful in answering any of the above questions. Sperduto et al. modified the original GPA score obviously due to the need to better stratify patients with brain metastases. The DS-GPA and finally Lung-molGPA [@b0100] and Melanoma-molGPA [@b0105] scores were developed and externally validated [@b0095]. Nevertheless, even these most recent scores have limited ability of reliably predicting individual patient prognosis, sometimes classifying several long term survivors into the group with the worst prognosis and vice versa [@b0105].
The role of extracranial factors such as performance status, extent of extracranial metastases or control of the primary tumor provided some additional valuable information about an individual's prognosis [@b0110]. A further refinement of the individual prognosis within the group of patients with very limited survival was possible after inclusion of such widely available and cheap biochemical surrogate parameters such as LDH and albumin [@b0030].
The application of the EC-S to our data yielded similar results as in the original derivation study by Nieder et al. [@b0030]. First, with the exception of the fairly good prognosis class 1, the EC-S appeared well calibrated with median OS differences between our data and the derivation data not larger than 1.1 months ([Table 2](#t0010){ref-type="table"}). Second, the EC-S had the highest discriminatory power as judged by Gönen & Heller's CPE. Third, the EC-S performed better than the RPA or DS-GPA score in separating the group of patients with very poor prognosis. However, in contrast to the DS-GPA score, the EC-S was not able to separate the two groups of patients with a good and fairly good prognosis in our data. This indicates some miscalibration of the model underlying the EC-S for patients with more favorable prognoses, so that survival predictions for new patients are not necessarily reliable.
In building a multivariable prognostic Cox regression model on our dataset, the RPA score was not selected as a prognostic factor, while both the DS-GPA and EC-S were. Due to these findings we pursued the idea of combining these two scores. We built 4 classes in the collective of 134 patients with both known DS-GPA and EC-S as displayed in [Table 4](#t0020){ref-type="table"}. Application of the new combined score (EC-GPA) resulted in a clear separation of the survival curves ([Fig. 3](#f0015){ref-type="fig"}) and yielded slightly better discrimination between prognostic groups than either the DS-GPA or EC-S alone, although the CPEs of the EC-S and EC-GPA scores overlap within their uncertainties due to the small sample size. We believe that, although acquired heuristically, the EC-GPA score combination could be a good prognostic tool which should be evaluated in future studies using larger combined or independent datasets.
It is obvious that the paradigm of reserving stereotactic radiosurgery for treating patients with the most favorable prognoses changed over time. Low toxicity and wide availability of stereotactic radiosurgery led to its more frequent application and omission of WBRT [@b0115] despite still controversial evidence for a clinical benefit and clear concerns from detailed analysis of available randomized trials [@b0120], [@b0125]. Prognostic scores were not analyzed in the EORTC 22952--26001 trial, and only the RPA-score was used in the trial of Yamamoto et al. The stratification according to GPA score in the secondary analysis of Aoyama et al. demonstrated clear benefit of WBRT in combination with stereotactic radiosurgery in the group with best prognosis also in terms of OS.
Since the publication of the QUARTZ-trial there is a good level of evidence that in preselected patients the use of very hypofractionated WBRT has a limited effect on OS and quality of life [@b0130]. Due to several limitations of the trial such as intention-to-treat analysis (ca. 20% of patients in the WBRT group did not receive WBRT), a low treatment dose of 20 Gy, an obvious negative selection of patients in both groups and more aggressive antitumor therapy in the control arm there are still some questions about its practice changing role. Despite the main conclusion of this trial, younger patients did derive a clear benefit from WBRT in terms of overall survival. GPA was not a significant variable in the survival analysis, probably because of the limited number of patients with a high score and limited statistical power to address this research question.
Our study shows that incorporation of extracranial factors into a prognostic model significantly improves discriminatory power. Nevertheless, our study has some limitations due to its retrospective nature and limited sample size, especially for the subset of patients for which the EC-S could be calculated. Furthermore, it was not planned a priori to develop a new prognostic score -- the combination of DS-GPA and EC-S was rather a data-driven heuristic approach for improving the discrimination of various patient groups with distinctly different prognoses. Due to the limited data quality associated with the retrospective nature of this study, the ad-hoc definition of the combined EC-GPA score should be seen as a limitation. We prefer to consider the EC-GPA score as a hypothetical possible way for further refinement of both the DS-GPA and EC-S that should however be evaluated in future studies.
5. Conclusions {#s0025}
==============
In conclusion, our study provides an independent validation of the prognostic EC-S developed by Nieder et al. [@b0030]. Their prognostic model which is based solely on extracranial factors appeared to have a higher external validity than both the RPA and DS-GPA scores when applied to the unselected patients from our hospital. It was also the best prognostic model for defining the patients who obviously did not benefit from RT of brain metastases at least in terms of OS. The combination of the EC-S with the established DS-GPA score resulted in a slight gain of discriminatory ability. Further validation of the EC-S and the new EC-GPA score will be pursued in an ongoing analysis of combined patient cohorts from different tertiary care centers.
Conflict of interest {#s0030}
====================
None.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
Chronic obstructive pulmonary disease (COPD) and tuberculosis (TB) are both diseases that mainly affect the lungs and are major causes of morbidity and mortality worldwide.
One third of the world population is infected with *Mycobacterium tuberculosis* and eight million new cases of TB are reported every year [@pone.0010138-Dye1]. In Sweden, as in many industrialised nations, the incidence of TB has declined over the past decades, presently being a low-incidence region with an annual incidence of around 5.5 cases/100,000 inhabitants, with a total of around 500 new cases annually. General vaccination against tuberculosis of newborns was performed from the 1940s until 1975 and thereafter selective vaccination has only been offered to children and young adults considered to run a high risk of exposure to TB. The majority of individuals who develop active TB in Sweden today are either relatively young immigrants from areas with a high incidence of TB or elderly native Swedes who were most probably infected at a younger age and suffer from reactivated TB [@pone.0010138-Romanus1].
The development of COPD results from a combination of polygenic inflammatory vulnerability and environmental factors, mainly tobacco smoke, leading to lung tissue remodelling and a non-reversible airflow limitation [@pone.0010138-Barnes1]. The prevalence of COPD is increasing worldwide, and it is estimated that COPD will become the third-leading cause of death by 2020[@pone.0010138-Calverley1]. In Sweden, approximately 500,000 individuals out of a population of 9 million suffer from COPD[@pone.0010138-Lindberg1], and around 2,500 die from COPD annually [@pone.0010138-1].
TB and COPD have common risk factors such as smoking and low socioeconomic status [@pone.0010138-Calverley1], [@pone.0010138-Lonnroth1]. In addition, both diseases have a significant genetic vulnerability component, although little is known regarding the extent to which these traits are shared [@pone.0010138-Maartens1], [@pone.0010138-Silverman1]. In the present case-control study, 231,734 COPD patients and randomly-selected control subjects from the general population were compared to investigate whether an underlying diagnosis of COPD affects the risk of active TB and the impact of this co-morbidity on TB mortality.
Methods {#s2}
=======
Ethics {#s2a}
------
The study was approved by the Lund University Research Ethics Committee (590/2004 and 294/2007). Informed consent was not obtained since all information regarding individuals was made anonymous to investigators prior to analysis in accordance to Swedish regulations. [@pone.0010138-Ludvigsson1].
Setting {#s2b}
-------
Swedish health care is publicly financed and all inpatient care is provided independently of health insurance and the patient\'s financial status. A unique, lifelong ten-digit personal identity number assigned to each person living in Sweden provides the possibility of linking records in databases administrated by the following federal institutes: Centre for Epidemiology (Swedish National Board of Health and Welfare), Statistics Sweden, and The Swedish Institute for Infectious Disease Control.
**The Swedish National Inpatient Register**, which was applied nationwide in 1987, includes individual information on all hospital discharges since 1964 with diagnosis coded according to The International Classification of Disease (ICD), Coding is performed by physicians at discharge and contra-signed by board-certified specialists. All records are controlled for general inadequacies and specific validation studies indicate that coverage is above 98% and that almost 90% of the diagnoses reported to the inpatient register are correct when compared with the original medical files[@pone.0010138-2].
**The Total Population Register** includes individual information on vital status, dates of immigration and emigration, county of residence, and country of birth.
**The Swedish National Population Censuses** performed in 1980, 1985 and 1990 provide information on housing conditions and socioeconomic status (SES). Participation was mandatory and response rates were 99.0%, 99.2% and 97.5%, respectively. Individuals were classified into different socioeconomic index groups (SEIs) based on type of occupation, educational background and responsibility levels according to specific criteria defined by Statistics Sweden. Subjects with missing information were mainly old age pensioners, unemployed, disability pensioners, homemakers and students[@pone.0010138-3].
**The Tuberculosis Register**: The Swedish Institute for Infectious Disease Control (SMI) monitors the epidemiology of infectious diseases in Sweden. All forms of TB have been notifiable according to law in Sweden since 1939. In 1969 a national tuberculosis register based on individual reports according to the communicable disease act was set up; since 1989 this register is administered by SMI. Reporting incident TB cases to the TB register is done in parallel by both the microbiological laboratories and clinicians when a patient is culture-positive. Clinically diagnosed TB is reported by clinicians only. The summarised statistics for the Tuberculosis Register during 1989--2007 is shown in [TABLE 1](#pone-0010138-t001){ref-type="table"}.
10.1371/journal.pone.0010138.t001
###### Summary statistics from Swedish Tuberculosis Register 1989--2007.
{#pone-0010138-t001-1}
Tuberculosis in Sweden 1989--2007[1](#nt101){ref-type="table-fn"}
------------------------------------------------------------------- --------------
No of TB cases 9,586 (100%)
Culture confirmed 7,893 (82%)
Pulmonary TB 6,234 (65%)
*Sex*
Male 5,044 (53%)
Female 4,542 (47%)
*Immigration Status*
Born in Sweden 3,761 (39%)
Born abroad 5,825 (61%)
*Age at TB-diagnosis*
\<40 years of age 4,310 (45%)
40--59 years of age 1,670 (17%)
60--79 years of age 2,367 (25%)
≥80 years of age 1,239 (13%)
Sources: The National Bacteriological Laboratory/Swedish Institute for Disease Control and Swedish Heart and Lung foundation. The Swedish Tuberculosis Index 1989--2002, ISSN 1103--4955. Tuberkulos I Sverige 2003--2008, in Swedish Only, in print.
Population {#s2c}
----------
The Centre for Epidemiology identified all individuals 40 years of age or older with a hospital discharge diagnosis of COPD between 1987--2003 in the Inpatient Register according to the International Classification of Diseases, Ninth and Tenth revisions (ICD-9: 491--492, 496; ICD-10: J41--J44), either as a main diagnosis (hospitalised *because of* COPD) or as secondary diagnosis (hospitalised *with* COPD). A total of 151,364 patients were identified, born 1880--1963, mean year of birth 1922. For each individual with a diagnosis of COPD, Statistics of Sweden randomly selected one control subject out of the general population from the Total Population Register matched for sex, year of birth, and county of residence during the year of first hospital discharge listing a COPD diagnosis.
Follow Up {#s2d}
---------
All patients and control subjects were linked to the Total Population Register to obtain information on country of birth, vital status and date of emigration. Each individual was also linked to The Swedish National Population Censuses of 1980, 1985 and 1990 to obtain information on SEI. The SEI groups were combined into three broad categories: manual, non-manual, and other types of occupation. Through linkage with the Inpatient Register we obtained information on all hospital discharges and all diagnoses between 1987 and 2003. To identify individuals with incident TB The Swedish Institute for Infectious Disease Control linked all individuals to the Tuberculosis Register from 1989 to 2007.
The primary outcome was defined as the first episode of any form of TB notified in the Tuberculosis Register in the overall analysis of TB, and the first episode of TB per localisation (pulmonary and extra-pulmonary) in the analysis by localisation. Individuals with more than one episode were counted only once. Follow up started one year after the first hospitalisation with COPD and on the corresponding date for control subjects to account for the fact that early TB symptoms may mimic COPD. Follow up ended with the first episode of TB, date of emigration, date of death or 31 December, 2007, whichever came first. A flow chart showing the principles of the study design is shown in [Figure 1](#pone-0010138-g001){ref-type="fig"}.
{#pone-0010138-g001}
Exclusion {#s2e}
---------
Among the 151,364 COPD patients, 101 (0.07%) individuals were excluded because of data-irregularities. Another 27,357 (18.1%) COPD patients, who died during the first year after COPD diagnosis, were excluded since they had a follow-up time of less than a year. 749 (0.49%) COPD patients were excluded from the analysis because of TB or TB-sequelae prior to or during their first year of hospitalisation with COPD identified in either the Tuberculosis Register (n = 212) or the Inpatient Register (n = 537) 1987--2003 (ICD-9: 010--019, 131; ICD-10: A15--A19, B90). 7,216 (4.8%) control subjects had a follow up time of less than a year, 162 (0.11%) control subjects were excluded because of TB or TB-sequelae prior to inclusion identified in either the Tuberculosis Register (n = 68) or the Inpatient Register (n = 94) 1987--2003.
As controls were randomly selected from the background population, some of the COPD patients also appeared as control subjects with time. We excluded 2,528 (1.6%) who had been hospitalised with a COPD-diagnosis prior to inclusion. Another 3,935 (2.7%) control subjects were initially included in the analysis but censored by the date of the first hospitalisation listing COPD. Because of the matched design 7,290 (4.8%) COPD-patients lacking controls due to exclusions and 25,591 (16.9%) odd controls were excluded. Analysis was based on the remaining 115,867 case-control pairs. Baseline characteristics of the study populations are shown in [TABLE 2](#pone-0010138-t002){ref-type="table"}.
10.1371/journal.pone.0010138.t002
###### Demographic characteristics of the study populations.
{#pone-0010138-t002-2}
COPD; n = 115,867 (100%) Controls; n = 115,867 (100%)
---------------------------------------------------- -------------------------- ------------------------------
*COPD Classification*
Primary diagnosis 70,404 (60.8%)
Additional diagnosis 45,463 (39.2%)
*Sex*
Male 62,226 (53.7%)
Female 53,641 (46.3%)
*Year of birth*
1888--1919 44,560 (38,5%)
1920--1939 59,821 (51,6%)
1940--1963 11,486 (9,9%)
*Age Distribution[1](#nt102){ref-type="table-fn"}*
(at COPD diagnosis)
40--59 years of age 15,756 (13.6%)
60--79 years of age 72,658 (62,7%)
≥80 years of age 27,453 (23.7%)
*Immigration status*
Born in Sweden 103,045 (88.9%) 105,950 (91.4%)
Born abroad 12,822 (11.1%) 9,916 (8.6%)
*Socioeconomic Status*
Non-manual 21,141 (18.3%) 30,092 (26.0%)
Manual 33,197 (28.6%) 31,385 (27.1%)
Other 11,569 (10.0%) 12,695 (11.0%)
Outside workforce 49,960 (43.1%) 41,695 (35.9%)
Control subjects were matched for year of birth, sex and county of living during the year of first hospital discharge listing COPD.
Statistical analysis {#s2f}
--------------------
Cox proportional hazards models were used to estimate hazard ratios (HR) of subsequent TB, comparing patients with COPD and control subjects. HRs were calculated separately for pulmonary, extra-pulmonary and pooled TB infections. All models were internally stratified by year of birth, sex and county of residence and subsequently adjusted for SES, co-morbidity and immigration status, dichotomised by birth in Sweden (yes/no). Co-morbidity was defined as prior hospital discharge diagnosis of diabetes mellitus, cardiac disease, alcohol liver disease or alcohol abuse, or silicosis, according to ICD-9 or ICD-10, as well as total number of hospitalizations and total duration of hospital stay. Interactions between COPD-status and age at inclusion, year of birth, sex, SES, immigration status, the level of diagnosis and prior hospitalization on outcome were tested by entering interaction terms to the fully adjusted COX model. Baseline predictors of TB among COPD patients were estimated using Cox regression stratified by year of birth among those born after 1915 who were not old age pensioners when the population censuses were performed. In the sub-population of patients with TB, logistic regression adjusted for sex and age at TB-diagnosis was used to estimate the OR of one-year all-cause mortality between COPD patients and control subjects. Proportionalities of hazards were assessed graphically and by testing for a non-zero slope in a generalised linear regression of the scaled Schoenfeld residuals as a function of time. All analyses were performed using STATA/SE (version 10.1 for Windows; StataCorp LP, College Station, TX).
Results {#s3}
=======
Relative risk of TB in COPD patients versus the general Population {#s3a}
------------------------------------------------------------------
Among the 231,734 cases and controls, a total of 291 first episodes of TB occurred - 201 among COPD patients and 90 among the control subjects during 680,818 and 986,119 person-years of follow-up, respectively, as detailed in [TABLES 3](#pone-0010138-t003){ref-type="table"} and [4](#pone-0010138-t004){ref-type="table"}. The resulting crude incidence rates of TB were 3.0 per 10,000 person-years of follow-up (95% confidence interval \[95% CI\] 2.6 to 3.4) in COPD patients and 0.9 (95% CI 0.7 to 1.1) in control subjects. In absolute terms this corresponds to one excess TB case among COPD patients per 4,900 person-years. The age-, sex- and county-stratified hazard ratio (HR) estimated by Cox regression was 3.3 (95% CI 2.5 to 4.2). Adjustment for immigration status, SES and co-morbidity resulted in a HR ratio of 3.1 (95% CI 2.4 to 4.1); see [TABLE 5](#pone-0010138-t005){ref-type="table"} for details.
10.1371/journal.pone.0010138.t003
###### In-patient care and co-morbidity at inclusion.
{#pone-0010138-t003-3}
COPD; n = 115,867 (100%) (n = 115,867; 100%) Controls; n = 115,867 (100%)
----------------------------------------------------- ---------------------------------------------- ------------------------------
Number of hospital discharges
Mean/median 3.8/2 1.8/1
Number of days spent in hospital
Mean/median 32.4/8 19.1/1
*Co-morbidity*
Alcohol consumption[1](#nt103){ref-type="table-fn"} 4,671 (4.0%) 1,070 (0.9%)
Cardiac disease[2](#nt104){ref-type="table-fn"} 25,509 (22.0%) 12,463 (10.8%)
Diabetes mellitus[3](#nt105){ref-type="table-fn"} 6,814 (5.9%) 4,092 (3.5%)
HIV 119 (0.10%) 9 (0.01%)
Silicosis 44 (0.04%) 9 (0.01%)
Renal failure 730 (0.6%) 259 (0.2%)
Including alcohol abuse and/or alcohol-induced liver disease.
Including ischemic heart disease and cardiac failure.
Including diabetes mellitus type 1 and 2.
10.1371/journal.pone.0010138.t004
###### Characteristics of patients suffering from tuberculosis in the presence and absence of COPD as a pre-morbid condition.
{#pone-0010138-t004-4}
COPD Controls
------------------------------------------------ ------------- ------------
No of TB cases 201 (100%) 90 (100%)
Positive culture 158 (78.6%) 79 (87.8%)
*Localization*
Pulmonary 157 (78.1%) 51 (56.7%)
Extra-pulmonary 26 (12.9%) 28 (31.1%)
Both 18 (9.0%) 11 (12.2%)
*Sex*
Male 127 (63.2%) 59 (65.6%)
Female 74 (36.8%) 31 (34.4%)
*Immigration status*
Born in Sweden 149 74
Born abroad 52 16
Age at TB-diagnosis in years
mean/median 75.2/77 77.8/78
40--59 14 5
60--79 117 47
≥80 70 38
No of deaths 93 (46.2%) 29 (32.2%)
(within first calendar year after diagnosis)
*Age at death*
*(within first calendar year after diagnosis)*
mean/median 77.8/79 81.9/82
40--59 4 0
60--79 46 10
≥80 43 19
10.1371/journal.pone.0010138.t005
###### Hazard ratios for TB among COPD-patients (n = 115,867) compared to general population control subjects (n = 115,867).[1](#nt106){ref-type="table-fn"}
{#pone-0010138-t005-5}
Univariate HR (95%CI) Multivariate HR (95%CI)[2](#nt107){ref-type="table-fn"}
--------------------------------------------------------------------- ----------------------- ---------------------------------------------------------
Control subjects 1 1
COPD-patients 3.26 (2.53--4.20) 3.14 (2.42--4.08)
*Socioeconomic Status*
Non-manual occupation 1 1
Manual occupation 0.99 (0.68--1.44) 0.86 (0.59--1.24)
Other occupation 1.01 (0.63--1.63) 0.98 (0.61--1.58)
Missing information 2.04 (1.40--2.94) 1.40 (0.97--2.04)
*Immigration Status*
Born in Sweden 1 1
Born abroad 3.55 (2.65--4.74) 3.08 (2.29--4.15)
*Co-morbidity*
Alcohol over-consumptionconsumption[3](#nt108){ref-type="table-fn"} 1.38 (0.66--2.88) 1.06 (0.50--2.26)
Cardiac disease[4](#nt109){ref-type="table-fn"} 1.00 (0.71--1.42) 0.85 (0.57--1.26)
Diabetes mellitus[5](#nt110){ref-type="table-fn"} 1.13 (0.62--2.09) 1.12 (0.60--2.10)
*Inpatient care* [6](#nt111){ref-type="table-fn"}
No of days spent in hospital
0 1 1
1--29 1.04 (0.80--1.34) 1.41 (0.19--10.38)
≥30 1.09 (0.76--1.55) 1.62 (0.21--12.36)
No of hospital discharges
0 1 1
1--4 1.02 (0.79--1.32) 0.61(0.08--4.46)
≥5 1.18 (0.84--1.67) 0.52 (0.07--4.06)
Estimated by COX regression stratified for year of birth, sex and county of residence, using calendar time as time scale.
Adjusted for all parameters in the table.
Including alcohol abuse and alcohol-induced liver disease.
Including ischemic heart disease and cardiac failure.
Including diabetes mellitus type 1 and type 2.
Prior to inclusion.
There was no evidence of effect modification by age at inclusion (p = 0.26), year of birth (p = 0.89) sex (p = 0.50), SES (p = 0.19), immigration status (p = 0.77) or the level of diagnosis (i.e. whether the COPD diagnosis was primary or secondary) (p = 0.13). A stratified analysis, restricted to those pairs of COPD-patients/controls where the control subject had been hospitalized prior to inclusion (n = 118,570), thus using an in-patient control group yielded similar results, with an HR of 3.06 (95% CI 1.98--4.73). A subgroup analysis between this subgroup and the subgroup of pairs where the control had no prior hospitalization showed no significant difference between the groups, (p = 0.93 from likelihood ratio test for interaction). A sensitivity analysis was performed wherein control subjects who were later hospitalised for COPD were not censored, thus analysing them as both COPD patients and controls. This analysis yielded a similar result after stratification for age, sex, and county of residence, adjusted for co-morbidity, SES and immigration status, resulting in a HR of 3.0 (95% CI 2.3--3.9). An additional analysis was performed in the 35,497 excluded case-control pairs, in whom 412 TB cases occurred during 1989--2003 (310 in COPD-patients and 102 in controls). Conditional logistic regression yielded OR 2.9 (2.3--3.7) adjusted for immigration status.
Localisation of TB: Comparing COPD patients and Controls {#s3b}
--------------------------------------------------------
Pulmonary TB was seen in 175 COPD patients and 64 control subjects, corresponding to an HR of 3.7 (95% CI 2.9 to 5.1) when stratified by age, sex and county and adjusted for immigration status, co-morbidity and SES. Extra-pulmonary TB was seen in 44 COPD patients and 39 control subjects resulting in an HR of 1.7 (95% CI 1.0 to 2.6) stratified by age, sex and county and fully adjusted.
Mortality after TB within first year {#s3c}
------------------------------------
Of the 291 patients who developed TB, 122 died within the first calendar year after the TB-diagnosis; 93 of the 201 COPD patients and 29 of the 90 control subjects. Median age at death was 79 years of age among COPD patients and 82 years of age among control subjects. COPD patients with TB had an odds ratio (OR) with respect to death of 2.2 (95% CI 1.3 to 3.9) adjusted for age at TB-diagnosis, compared to general population controls with TB. Cardiac disease and SES were the only covariates significantly associated with death in univariate analysis and none of the investigated co-morbidities, immigration status, sex nor SES covariates reached significance in multivariate analysis (data not shown). Estimation of the OR, with respect to death within two years after TB-diagnosis, yielded similar results.
Predictors of TB among COPD patients {#s3d}
------------------------------------
217 TB cases occurred among the 86,920 COPD patients born after 1915 who were not old age pensioners when the population censuses were performed. Hence missing information on SES was equal to being outside the workforce mainly as disability pensioner, unemployed, homemaker or student. Immigration status, sex and SES, all predicted the risk of TB in univariate analysis stratified by year of birth. In the multivariate model stratified for year of birth and adjusted for all of the above parameters, immigration status, sex and SES remained significant; see [TABLE 6](#pone-0010138-t006){ref-type="table"} for details. The younger birth cohorts had a lower absolute incidence of active TB compared to the older birth cohorts but there was no effect modification by birth cohort.
10.1371/journal.pone.0010138.t006
###### Predictors of TB among individuals discharged with a prior diagnosis of COPD (n = 86,920).[1](#nt112){ref-type="table-fn"}
{#pone-0010138-t006-6}
Univariate HR (95% CI) Multivariate HR (95% CI)[2](#nt113){ref-type="table-fn"}
---------------------------------------------------------- ------------------------ ----------------------------------------------------------
*Immigration status*
Born in Sweden 1 1
Born abroad 3.19 (2.25--4.53) 3.11 (2.18--4.44)
*Sex*
Male 1 1
Female 0.72 (0.52--1.00) 0.66 (0.47--0.91)
*Socioeconomic Status*
Non-manual occupation 1 1
Manual occupation 0.71 (0.46--1.10) 0.64 (0.42--0.99)
Other 0.65 (0.35--1.22) 0.61 (0.32--1.14)
Outside workforce 1.24 (0.81--1.89) 1.10 (0.72--1.68)
*Co-morbidity*
Alcohol over-Consumption[3](#nt114){ref-type="table-fn"} 1.08 (0.47--2.47) 0.96 (0.42--2.22)
Cardiac disease[4](#nt115){ref-type="table-fn"} 0.91 (0.58--1.41) 0.86 (0.54--1.35)
Diabetes mellitus[5](#nt116){ref-type="table-fn"} 0.87 (0.38--1.97) 0.85 (0.37--1.96)
*Age at inclusion*
*(first hospitalisation listing COPD)*
40--59 1 1
60--79 1.58 (0.59--4.21) 1.57 (0.59--4.18)
≥80 1.06 (0.28--3.95) 1.08 (0.29--4.05)
Estimated by Cox regression, stratified for year of birth, using calendar time as time scale.
Adjusted for all parameters in the table.
Including alcohol abuse and alcohol-induced liver disease.
Including ischemic heart disease and cardiac failure.
Including diabetes mellitus type 1 and 2.
Discussion {#s4}
==========
Previous studies have shown that COPD is a frequent co-morbid condition in TB patients [@pone.0010138-Wang1], [@pone.0010138-Liu1], [@pone.0010138-Aktogu1]. However, to the best of our knowledge there is no study that has specifically studied the association between COPD and subsequent TB. In this population-based study comprised of 115,000 COPD patients, we show that individuals with a hospital discharge diagnosis of COPD have a three-fold increased risk of developing active TB compared to the general population, mainly due to an excess risk of pulmonary TB. Moreover, our study shows that TB-patients with COPD have a two-fold increased risk of death within first year after TB-diagnosis when compared to general population control subjects suffering from TB.
The national, publicly-financed health care system ensures that the present study covers all patients hospitalised with COPD in Sweden during the observation period. However, COPD patients who were only treated as outpatients were not included in the analysis. The prevalence of physician-diagnosed COPD in Sweden is estimated to be around 5% [@pone.0010138-Hasselgren1], [@pone.0010138-Lundback1], [@pone.0010138-Montnemery1]. Of the randomly selected general population control subjects, 4.3% did have a hospital discharge diagnosis of COPD during the observation period which implies that a large proportion of patients with COPD will eventually be hospitalised.
Validation of the COPD diagnosis was not possible due to anonymisation of data prior to analysis, but there was no evidence for an effect modification by the level of diagnosis, suggesting that the diagnosis of COPD, whether primary or secondary, was equally accurate. Swedish validation studies of hospital discharge coding suggest a slight overlap between COPD and asthma [@pone.0010138-4]. Previous studies have associated a history of asthma with a decreased risk of TB [@pone.0010138-Schatz1], [@pone.0010138-Shirakawa1], [@pone.0010138-Lienhardt1], [@pone.0010138-Bahceciler1]. In the present study, a misclassification of the COPD-diagnosis, in addition to the inclusion of COPD patients receiving only outpatient care as control subjects, would presumably bias the estimates downwards.
A limitation of the present study is the inability to estimate the risk of exposure to TB, previous or current. In addition, it was not possible to identify individuals who suffered from latent tuberculosis infection or who were recently exposed to TB at the time of the study entry. Individuals identified as having had TB or TB-sequelae were excluded from analysis since pulmonary TB itself can lead to COPD [@pone.0010138-Salvi1]. Comparison of the Inpatient register and the TB register for the overlapping time periods (1989--2003) yielded around 70% excess TB-diagnoses in the Inpatient register. The TB register contains valid incident cases whereas the Inpatient register probably also contains cases of former TB and cases examined for TB but, but this incongruence needs to be further evaluated. We used the TB register as the golden standard of incident TB but we still used a diagnosis of TB in the Inpatient Register prior to inclusion as an exclusion criterion to be certain that former TB cases were properly excluded.
Models were adjusted for the use of inpatient care at baseline and inpatient co-morbid conditions associated with COPD and previously known to increase the risk of active TB [@pone.0010138-5], [@pone.0010138-Barnes2]. Residual confounding from alcohol intake or co-morbid conditions only requiring outpatient care cannot be ruled out. HIV, renal failure or silicosis were not included in the models since these were not listed in the hospital discharge diagnoses of TB-cases. Globally, HIV infection is an important modifier of the risk of active TB but the impact in the Swedish setting is probably only of minor importance since the prevalence of HIV-infection is low (around 0.05%) [@pone.0010138-6].
There is growing consensus that smoking increases the risk active TB-disease[@pone.0010138-Lin1]; three different meta-analyses essentially covering the same studies found a 1.5--2 fold increased risk of TB among current smokers [@pone.0010138-Bates1], [@pone.0010138-Lin2], [@pone.0010138-Slama1]. A limitation of the present study is the inability to control for smoking, which was not possible since none of the registers with national coverage contain smoking data. In this study, although on a group level, COPD-patients with an ever-smoking rate of around 75% [@pone.0010138-Montnemery2], [@pone.0010138-Lindberg2] are compared with general population control subjects with an ever smoking rate of around 50% [@pone.0010138-7].
Smoking prevalence in Sweden as reflected in the control group has declined during the past decades[@pone.0010138-7]. Smoking rates have gone from 35% and 28% in men and women in 1980 to 14% and 18% in 2004, respectively. The birth cohorts of the 1940s and 1950s have the highest smoking prevalence and smoking habits have been shown to be strongly associated with SES and immigration status[@pone.0010138-7]. Prevalence studies of COPD in Sweden have found that current smoking rates range between 31% to 47% of people with physician-diagnosed COPD (not necessarily hospitalised) [@pone.0010138-Montnemery2], [@pone.0010138-Lindberg2], whereas in a study of patients hospitalised for severe COPD (GOLD-stage IV) only 9% were current smokers [@pone.0010138-Stenfors1]. The smoking rate by the first hospitalisation for COPD has not been studied.
Despite the inability to control for smoking, our results are in line with previous studies. A case-control study of from England [@pone.0010138-Jick1] found that patients with emphysema and bronchitis had a two to three-fold increased risk of developing TB, although the results were based on a small sub-group of individuals suffering from emphysema and bronchitis (n = 264). A cohort-study from Denmark with 14,000 participants [@pone.0010138-Benfield1] also found a two to three-fold increased RR for hospitalisation with TB for participants with moderate to severe lung function impairment compared to participants with normal lung function.
Apart from smoking, COPD patients suffer from conditions that could potentially increase the risk of active TB such as low body mass index [@pone.0010138-Cegielski1] and impaired muco-ciliary clearance. Several bacterial species are important both as colonisers and as inducers of COPD exacerbations [@pone.0010138-Sethi1]. Factors promoting the survival of these bacteria in the lower airways are likely to promote the survival of *M Tuberculosis* as well. In addition, patients suffering from COPD-exacerbations are often treated with oral corticosteroids, a well known risk factor for TB [@pone.0010138-Jick1].
There are some conflicting findings in the literature as to whether COPD is a risk factor for death from TB [@pone.0010138-Ormerod1], [@pone.0010138-Fielder1]. We found a two-fold increased risk of all cause mortality for COPD patients within first year after TB the diagnosis. This could be due to death from other causes related to COPD, caused by worsened outcome because of delayed diagnosis or speculatively, altered host defence functions in the lower airways.
This population-based study, comprising a large number of COPD patients, shows that COPD patients have an increased risk of developing active TB compared to the general population. These results raise concern that the increasing global burden of COPD will further enhance the incidence of active TB especially in settings with a high burden of latent TB. The underlying contributory factors such as smoking, low body mass index, impaired muco-ciliary clearance, treatment with corticosteroids or impaired host defense functions need to be disentangled and evaluated in further studies.
We would like to express our gratitude to Lena Brandt for providing statistical expertise.
**Competing Interests:**Claes-Göran Löfdahl has been paid for lectures by AstraZeneca, Boehringer-Ingelheim, GlaxoSmithKline and Pfizer; taken part and been paid for ad hoc advisory boards for the same companies; and has had institutional support as unrestricted grants from AstraZeneca, Boehringer-Ingelheim, and GlaxoSmithKline. This study has in some parts been funded by a commercial funder, but that does not alter the authors\' adherence to all the PLoS ONE policies on sharing data and materials.
**Funding:**This study was supported by grants from the Swedish Heart and Lung Foundation (project 20080246), an unrestricted grant from GlaxoSmithKline, the Swedish Research Council (project K2008-57X-20674), the Swedish Government Funds for Clinical Research, i.e. Avtal om Läkares Forskning (ALF), the University Hospital in Lund, and the Foundations of Th C Bergh, Crafoord, Hedberg, Bengt Ihre, Greta & Johan Kock, Alfred Osterlund, and Marianne & Marcus Wallenberg. M.I. is the recipient of a fellowship from ALF. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
[^1]: Conceived and designed the experiments: MI AE GE BL VR CGL AE. Analyzed the data: MI GE. Contributed reagents/materials/analysis tools: MI GE BL VR CGL AE. Wrote the paper: MI AE GE BL VR CGL AE. Data collection: MI.
| {
"pile_set_name": "PubMed Central"
} |
All relevant data are within the paper and its Supporting Information files.
Introduction {#sec001}
============
*Brassica oleracea* is an important vegetable crop and is genetically diverse, having various subspecies, such as cabbage, cauliflower, broccoli, kale and wild-type, and having many known useful traits, such as its strong resistance against *Sclerotinia* incorporated from wild subspecies of *B*. *incana* \[[@pone.0193548.ref001], [@pone.0193548.ref002]\]. *B*. *napus* is an important oilseed crop in the world, originating from a natural interspecific hybridization between *B*. *rapa* and *B*. *oleracea* \~6000 years ago \[[@pone.0193548.ref003], [@pone.0193548.ref004]\]. This crop's genetic basis was narrower than the parental species due to its short history and domestication through modern breeding methods \[[@pone.0193548.ref005]\]. Introgression of genetic resources from parental species into *B*. *napus* is necessary to broaden and improve its genetic basis \[[@pone.0193548.ref006]--[@pone.0193548.ref009]\].
To utilize the genetic resources of parental species, the strategy of crossing *B*. *napus* and its parental species is commonly used. In the strategy, it is easy to select new types of *B*. *napus*, either gaining useful traits from parental species \[[@pone.0193548.ref010], [@pone.0193548.ref011]\] or having the potential to broaden the genetic basis of natural *B*. *napus* \[[@pone.0193548.ref008], [@pone.0193548.ref012], [@pone.0193548.ref013]\]. However, it is difficult to select new types of *B*. *oleracea/B*. *rapa* individuals, which might due to higher competition of *B*. *napus* gametes than *B*. *oleracea/B*. *rapa* gametes. To verify this hypothesis, the meiotic behavior of interspecific hybrid between *B*. *napus* and *B*. *oleracea* and its microspore-derived (MD) individuals were analyzed.
Microspore culture is widely applied in *Brassica* species to produce double haploid (DH) individuals in germplasm collection, QTL mapping, genetic engineering and crop improvement \[[@pone.0193548.ref014]--[@pone.0193548.ref017]\]. This method is less commonly used in interspecific hybrids between *Brassica* species due to the difficulty in obtaining embryoids \[[@pone.0193548.ref018]--[@pone.0193548.ref020]\]. However, scientists have used the technique in interspecific hybrids to induce microspore-derived lines, aiming to study male meiotic behavior, since there is no selection pressure from females compared with self-pollination and backcrossing \[[@pone.0193548.ref021], [@pone.0193548.ref022]\]. In the present study, an interspecific hybrid between *B*. *napus* and *B*. *oleracea* was developed, and its meiotic behavior and gamete behavior in microspore-derived individuals were analyzed, showing that the *B*. *napus*-like gamete had high fitness and competition in the hybrid. This suggested that high viability of *B*. *napus*-like gametes might make it easy to select new types of *B*. *napus* from the interspecific hybrid between *B*. *napus* and its parental species by self-pollination.
Materials and methods {#sec002}
=====================
Plant materials {#sec003}
---------------
The interspecific hybrid ACC was developed from hybridization between *B*. *napus* 'Zhongshuang 9' and *B*. *oleracea* '6m08' via embryo rescue and propagated on MS regeneration medium (MS + 3 mg/L 6-BA + 0.02 mg/L NAA) via tissue culture for microspore culture \[[@pone.0193548.ref023]\]. Morphology, fertility, chromosome number and genetic components of MD lines were evaluated, and their genetic diversity was compared with 34 natural *B*. *napus* and 42 *B*. *oleracea* ([S1 Table](#pone.0193548.s001){ref-type="supplementary-material"}).
Cytological observations {#sec004}
------------------------
### Chromosome number at mitotic metaphase {#sec005}
To check the chromosome numbers of the ACC hybrid and MD progenies, the young ovaries were collected and pretreated with 2 mmol/L 8-hydroxyquinoline for three to four hours at room temperature and later fixed in Carnoy's solution (V~ethanol~: V~acetic\ acid~ = 3:1) and stored at 4 °C. Mitotic observations were made according to the methods as described by Li et al. \[[@pone.0193548.ref024]\]. The ovaries at mitosis were hydrolyzed in 1 M HCl at 60 °C for 8 min and stained with 10% modified carbol fuchsin and observed under microscope.
### Chromosome pairing and segregation at meiosis {#sec006}
For meiotic analysis, buds were fixed in Carnoy's solution for 24 h and then transferred into fresh mixture and stored at --20 °C for future use. Meiotic observations of pollen mother cells (PMCs) were made according to the methods of Li et al. \[[@pone.0193548.ref024]\]. The anthers at meiosis stage were hydrolyzed in 1 M HCl at 60 °C for 2 min, stained with 10% modified carbol fuchsin and observed under microscope. The chromosome pairing at metaphase I and chromosome segregation at anaphase I in PMCs were recorded.
Pollen fertility {#sec007}
----------------
Pollen fertility was determined by the percentage of pollen grains stained with 1% acetocarmine according to the method of Li \[[@pone.0193548.ref024]\]. Three flowers were counted from ACC hybrid and MD lines. More than 300 pollen grains were recorded for each line. Grains that were round and stained red were considered normal, whereas small and non-stained ones were considered dead pollen.
Microspore isolation {#sec008}
--------------------
Microspore culture was performed by the method described by Lichter \[[@pone.0193548.ref025]\], with minor modifications. A total of 30 flower buds ranging in length from 2.5 to 3.5 mm from the ACC hybrid were selected and sterilized in 10% sodium hypochlorite solution for 15 min. The sterilized buds were then released with B5-13 medium. The solution along with the microspores were filtered through a 48-μm filter and transferred into a sterile 10 mL centrifuge tube, and the volume was adjusted to 8 mL with B5-13 media. The microspores were then centrifuged for 3 min at 1200 rpm, and the supernatant was discarded. B5-13 media was added to mix the microspores, and then they were centrifuged for 3 min at 1200 rpm again. The supernatant was discarded and microspores were re-suspended in 8 mL NLN-13 solution (NLN medium plus 13% sucrose in Millipore water, pH to 5.8).
The microspore suspension was divided into 4 Petri dishes with diameter of 70 mm, and 4 mL NLN-13 and 1 mL 10% activated charcoal were added into each Petri dish, which were later sealed with parafilm. The isolated microspores were incubated at 32 °C for 48 hours and then transferred into a 24 °C incubator for 20 days \[[@pone.0193548.ref026]\]. The plates were then put on a shaker (60 rpm) for embryo development. Three weeks later, young embryos were transferred into ½ MS medium for plant induction.
SSR marker analysis {#sec009}
-------------------
Genomic DNA was isolated from young leaves using the CTAB method \[[@pone.0193548.ref027]\]. 30 MD lines randomly selected were genotyped with 34 natural *B*. *napus* and 42 *B*. *oleracea* using 35 sets of SSR primers ([S2 Table](#pone.0193548.s002){ref-type="supplementary-material"}). The SSR results were described by the absence (0) or presence (1) of a band.
The genetic distance (GD) between accessions *X* and *Y* was calculated using the formula, GD*xy* = 1 --N*xy* / (N*x* + N*y*), where N*xy* is the number of common bands shared by accession *X* and *Y*, and N*x* and N*y* are the total number of bands in accession *X* and *Y*, respectively \[[@pone.0193548.ref028]\]. The phylogenetic tree was constructed using the neighbor-joining method implemented by MEGA version 6 \[[@pone.0193548.ref029]\].
Statistical analysis {#sec010}
--------------------
Analysis of variance (ANOVA), Pearson's simple correlation coefficient and *X*^*2*^ test were calculated using the statistical package SAS version 8.0 \[[@pone.0193548.ref030]\].
Results {#sec011}
=======
Development of interspecific hybrid between *B*. *napus* and *B*. *oleracea* {#sec012}
----------------------------------------------------------------------------
Immature embryos 7\~10 days after pollinating with *B*. *oleracea* (6m08) pollen on the stigma of *B*. *napus* (Zhongshuang 9) were cultured on ½ MS medium via embryo rescue. Three weeks later, an interspecific hybrid was developed, sharing intermediate morphology between two parents and having lighter green leaf color than both parents ([Fig 1A--1C](#pone.0193548.g001){ref-type="fig"}). Its pollen fertility was 34.82%, which was lower than the parental species (Zhongshuang 9: 99.5%, 6m08: 96.4%), and its chromosome number was 28 in meiotic and mitotic cells ([Fig 2A](#pone.0193548.g002){ref-type="fig"}).
{#pone.0193548.g001}
{#pone.0193548.g002}
Meiotic behavior of an interspecific hybrid between *B*. *napus* and *B*. *oleracea* {#sec013}
------------------------------------------------------------------------------------
Different chromosome conformations, such as univalents, bivalents, trivalents and quatrivalents, were observed in pollen mother cells (PMCs) at metaphase I (MI) of the hybrid. The average chromosome association was 9.66I + 9.12II + 0.01III + 0.02IV in 170 PMCs at MI. In certain cases, the frequency of the pattern of 9II + 10I (84.71%) was higher than the pattern of 10II + 8I (12.84%) ([Fig 2B and 2C](#pone.0193548.g002){ref-type="fig"}). Despite the high frequency of chromosome segregations of 13:15 (36.81%), 12:16 (30.77%) and 14:14 (15.93%) patterns were observed in 182 PMCs at anaphase I (AI), the fitness of the 9:19 (1.10%) pattern was the highest among the 6 patterns observed with a 5.49-fold higher fitness than the theoretical expectation ([Fig 2D and 2E](#pone.0193548.g002){ref-type="fig"}). This finding suggested that there would be a high probability of producing *B*. *oleracea*/*B*. *napus*-like gametes (C/AC = 9:19). In certain cases, meiotic irregularities, such as chromosome bridges and lagging chromosomes, were observed during the first and second divisions in the hybrid ([Fig 2F](#pone.0193548.g002){ref-type="fig"}).
Development of microspore derived plants {#sec014}
----------------------------------------
In total, 115 embryoids (18.55%) were induced from 620 flower buds between late uninucleate stage and early binucleate stage (2.5\~3.5mm flower bud) of the hybrid via microspore culture. Only 43 MD lines were obtained after transplanting these embryoids on ½ MS medium for plant-induction. All of these lines shared light green leaf color with the hybrid but had different number of leaf auricles, for example, the number of the leaf auricles ranged from 0 to 5 ([Fig 1D--1I](#pone.0193548.g001){ref-type="fig"}).
Pollen fertility and chromosome number of microspore-derived plants {#sec015}
-------------------------------------------------------------------
Pollen fertility of the MD lines ranged from 0 to 98.89%, with an average of 49.42%. Fertility was significantly positively correlated with chromosome number (*P* = 0.0027, *r* = 0.70; [S3 Table](#pone.0193548.s003){ref-type="supplementary-material"}), suggesting that lines with more chromosomes had higher pollen fertility.
In the 43 MD individuals, diverse chromosome numbers were observed. With the exception of one individual having 66 chromosomes and one having more than 80 chromosomes, the chromosome number of the other 41 individuals ranged from 15 to 56. Of these individuals, 14 were haploid, and 29 were polyploid by natural chromosome doubling. In detail, five of them had 17 chromosomes, five had 38 chromosomes, four had 30 chromosomes, four had 56 chromosomes and three lines had 19 chromosomes. Overall, twelve patterns of gametes were found. The frequency of actual gametes was significantly different from the theoretical gametes via *X*^*2*^ test (*P* \< 0.0001). This analysis showed that all the individuals had more than 14 chromosomes, suggesting that gametes having more chromosomes might survive, whereas the ones with less might die during the meiosis stage in the interspecific hybrid between *B*. *napus* and *B*. *oleracea*.
Although 27 (65.85%) individuals were aneuploid (n ≠ 19), five (12.20%) individuals were unreduced gametes (n = 28), individuals having gametes with 19 chromosomes (19.51%, 8/41) were the most common of all the patterns ([Fig 2G--2I](#pone.0193548.g002){ref-type="fig"}). This indicated that *B*. *napus*-like individuals having gametes with 19 chromsomes were more competitive than others in the hybrid between *B*. *napus* and *B*. *oleracea*.
Genetic diversity of microspore derived plants {#sec016}
----------------------------------------------
To verify the genetic diversity of the MD population, 115 polymorphic loci were amplified by genotyping 30 MD individuals with 35 combinations of SSR primers. Compared to the parental species (Zhongshuang 9 and 6m08), the MD population shared on average \~53 loci (45.71 ± 1.11%) with both parental species, \~37 loci (32.26 ± 1.60%) with the single parent *B*. *napus* (Zhongshuang 9) and \~14 (12.09 ± 1.60%) with the single parent *B*. *oleracea* (6m08). However, these plants also had \~11 unique loci (9.94 ± 1.11%) distinct from both parents. The average genetic distance between the MD population and the *B*. *oleracea* parent (0.91 ± 0.07) was significantly farther than the *B*. *napus* parent (0.34 ± 0.05, *P* \< 0.0001). Compared to Zhongshuang 9, the genetic distance of *B*. *napus*-like individuals (0.39 ± 0.07, *P* = 0.046) was significantly more distant than aneuploid (0.33 ± 0.04) and unreduced gametes (0.32 ± 0.02). This finding was similar to the distance between MD lines and 6m08 (*B*. *napus*-like individuals: 0.94 ± 0.09; aneuploid: 0.91 ± 0.07; unreduced gamete: 0.88 ± 0.05; [Table 1](#pone.0193548.t001){ref-type="table"}), suggesting more genetic components from the *B*. *napus* parent than the *B*. *oleracea* parent were inherited by the MD individuals.
10.1371/journal.pone.0193548.t001
###### Comparing the genetic distance of *B*. *napus*-like individuals, aneuploid and unreduced gametes of microspore-derived lines to the parental species, natural *B*. *napus* and *B*. *oleracea*.
{#pone.0193548.t001g}
*B*. *napus*-like[\*](#t001fn001){ref-type="table-fn"} Aneuploid Unreduced gamete
--------------- -------------------------------------------------------- ------------- ------------------
Zhongshuang 9 0.39 ± 0.07 0.33 ± 0.04 0.32 ± 0.02
6m08 0.94 ± 0.09 0.91 ± 0.07 0.88 ± 0.05
AACC group 0.43 ± 0.04 0.39 ± 0.03 0.38 ± 0.02
CC group 0.94 ± 0.05 0.94 ± 0.04 0.93 ± 0.04
\*: *B*. *napus*-like individuals with n = 19 chromosomes.
This finding was in accordance with the distance among MD population, natural *B*. *napus* and *B*. *oleracea*. In comparison with 34 *B*. *napus* and 42 *B*. *oleracea* subspecies, the average genetic distance between MD population and *B*. *oleracea* population (0.97 ± 0.37) was similar to that between *B*. *napus* and *B*. *oleracea* population (0.97 ± 0.33), but it was further than that between MD population and *B*. *napus* population (0.42 ± 0.17), suggesting the MD population is different from natural *B*. *napus* and *B*. *oleracea*, but close to *B*. *napus*. The obvious genetic differences among MD lines, *B*. *napus* population and *B*. *oleracea* population were also supported by the phylogenetic tree ([Fig 3](#pone.0193548.g003){ref-type="fig"}). Although the average genetic distance among *B*. *napus*-like individuals, aneuploid and unreduced gametes were similar, the genetic distance of *B*. *napus*-like individuals to the natural *B*. *napus* group (0.43 ± 0.04, *P* = 0.0091) was farther than that of aneuploid (0.39 ± 0.03) and unreduced gametes (0.38 ± 0.02; [Table 1](#pone.0193548.t001){ref-type="table"}). This finding indicated that these *B*. *napus*-like individuals, having gametes with 19 chromosomes, had the potential to widen the genetic basis of *B*. *napus*.
{#pone.0193548.g003}
Discussion {#sec017}
==========
Meiotic behavior of *Brassica* interspecific hybrid revealed by microspore culture {#sec018}
----------------------------------------------------------------------------------
Interspecific hybridization plays an important role in exchanging genetic components, widening and improving genetic resources in *Brassica* species. Although high frequency of euploids (new type *B*. *napus*) was observed in the interspecific hybrid between *B*. *napus* and parental species \[[@pone.0193548.ref008], [@pone.0193548.ref009], [@pone.0193548.ref020], [@pone.0193548.ref031], [@pone.0193548.ref032]\], aneuploid and unreduced gametes occurred frequently due to abnormal meiosis of interspecific hybrids \[[@pone.0193548.ref033]--[@pone.0193548.ref036]\]. In the present study, only 43 individuals were developed from the interspecific hybrid between *B*. *napus* and *B*. *oleracea* due to the difficulty in generating a large number of microspore-derived lines, and these individuals exhibited 19.51% euploid, 65.85% aneuploid and 12.20% unreduced gametes.
The frequency of aneuploid, euploid and unreduced gametes in the interspecific hybrid might be attributable to genotype-specific effects, such as sharing a common subgenome, or environmental factors, such as cold or fluctuating temperatures, plant nutrition, water stress and disease \[[@pone.0193548.ref037]--[@pone.0193548.ref042]\]. In the present study, the interspecific hybrid sharing a common C-subgenome from *B*. *napus* and *B*. *oleracea*, and produced high frequency of euploid (19.51% *B*. *napus*-like gametes), which was similar to the interspecific hybrid between *B*. *napus* and *B*. *rapa* sharing an A-subgenome \[[@pone.0193548.ref020]\]. It is necessary to investigate the genetic or developmental factors that may give rise to this apparent selection for the variation of gametes in the interspecific hybrid between *B*. *napus* and *B*. *oleracea* in the future.
Genetic variance of microspore derived lines {#sec019}
--------------------------------------------
In interspecific hybridization, chromosomes of related species recombine and interact regularly, causing homoeolog expression bias, genomic dominance and genomic imprinting \[[@pone.0193548.ref043]--[@pone.0193548.ref045]\]. In the MD lines, the genetic distance was closer to the *B*. *napus* parent (0.34 ± 0.05) than the *B*. *oleracea* parent (0.91 ± 0.07), suggesting more genetic components of *B*. *napus* than *B*. *oleracea* were inherited into the MD population. This might be due to genomic dominance and genomic imprinting of the *B*. *napus* parent in the MD lines. Subgenome dominance is an important phenomenon in allopolyploids, it was also observed in the interspecific hybrids. For example, in the interspecific hybrids (wheat × *Aegilops*), C-subgenome nucleolar organizing regions loci are dominant \[[@pone.0193548.ref046]\]. In addition, the subgenome dominance occurred instantly following the hybridization \[[@pone.0193548.ref047]\]. This bias in gene expression must be investigated in exploring the mechanism of *B*. *napus* genomic dominance.
In the present study, all of the MD lines were different from the parental species, especially the *B*. *napus*-like individuals, which has the potential to broaden the genetic basis of natural *B*. *napus*. The other lines might be used to produce monosomic alien addition lines and nullisomic lines, which can be used as bridge to transfer desired genes from wild *B*. *oleracea* species into *B*. *napus* \[[@pone.0193548.ref048], [@pone.0193548.ref049]\]. The role of these novel MD plants in *Brassica* species improvement needs to be evaluated in the future.
Supporting information {#sec020}
======================
###### Accessions of 34 natural *B*. *napus* and 42 *B*. *oleracea* used to analyze genetic diversity of microspore-derived lines derived from the hybrid between *B*. *napus* and *B*. *oleracea*.
(XLSX)
######
Click here for additional data file.
###### List of SSR primers.
(XLSX)
######
Click here for additional data file.
###### Data from microspore-derived lines derived from interspecific hybrid between *B*. *napus* and *B*. *oleracea* relating to fertility, chromosome number and genetic variance.
(XLSX)
######
Click here for additional data file.
This study was partly supported by the National Key Research and Development Program of China (2017YFD0101804), the China Postdoctoral Science Foundation (2015M582500), the Fundamental Research Funds for the Central Universities in China (XDJK2016C080), the Chongqing Postdoctoral Science Foundation (Xm2016029), and the Science and Technology Innovation Project of Chongqing (cstc2015shms-ztzx80005, cstc2015shms-ztzx80007 and cstc2015shms-ztzx80009).
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Myofascial Pain Syndrome (MPS) is exceedingly common in Western society \[[@B1]-[@B3]\]. Vernon and Schneider note that MPS is thought to be the leading diagnosis among pain management specialists \[[@B4],[@B5]\] and the leading diagnosis in pain patients reporting to general practitioners \[[@B4],[@B6]\]. MPS can be viewed as a regional pain syndrome such as, neck, low back and upper quadrant pain syndromes. MPS can also be found focally in discrete painful sites known as Myofascial Trigger Point (MTrP). Locations of MTrP\'s have been characterized on clinical grounds throughout the musculoskeletal system, starting with the seminal work of Travell and Simons \[[@B6],[@B7]\]. MTrP\'s have been described as active (clinically active with pain referral upon palpation) or latent (not clinically active, but tender on manual palpation) \[[@B7]-[@B9]\]. The presence of associated features such as \"local twitch sign\" and palpable taut bands is thought to confirm the presence of an active MTrP, while latent TP\'s may not present with the classic features.
Numerous treatments are currently available for MPS and MTrP\'s. Several systematic reviews have recently been published on a variety of treatments \[[@B4],[@B10]-[@B14]\], including pharmacologic agents, physical agents, complementary and alternative therapies. One commonly used therapy is the application of topical agents to relieve pain \[[@B15]-[@B17]\]. These agents can be classified as pharmacologic or non-pharmacologic. The former category includes non-steroidal anti-inflammatory agents (such as ketaprofen), opioid agents, classical analgesic agents (such as lidocaine), novel analgesic agents such as capsaicin and rubefascient agents (containing salicylates or nicotinamides) \[[@B15]-[@B17]\]. Non-pharmacologic agents generally fall into the complementary and alternative medicine (CAM) category for MPS and MTrP treatments. They may contain putative analgesic agents or counterirritant agents and they may exert cold, hot or neutral effects.
Several studies have investigated the effect of such topicals in the treatment of osteoarthritis, particularly of the knee \[[@B18],[@B19]\]. However, there appears to be a lack of studies investigating the effects of topical agents for the treatment of MPS or MTrP. A randomized, placebo-blinded clinical trial of non-pharmacological topical analgesics was conducted comparing leading national and professional brands in the treatment of a myofascial trigger point.
Materials and methods
=====================
Subjects
--------
Subjects were selected from consecutive clinical presentations of patients for treatment in a multidisciplinary health clinic. Subjects aged 18-80, male or female, with or without neck pain were included. Subjects were excluded for the following reasons: acute pain presentation preventing comfortable participation and absence of a palpable tender spot in the right upper trapezius region. Subjects were informed of the nature of the experiment and consented to participate. This protocol was approved by the Research Ethics Board of the Canadian Memorial Chiropractic College.
Randomization
-------------
A random number sequence was generated by an independent party and kept in the care of the clinic administrator until required for subject allocation. The randomization schedule was concealed from all other study staff prior to selection and prior to group allocation.
Outcome measures
----------------
Pre- and post-intervention measures consisted of 1) tenderness at an MTrP in the upper trapezius and, 2) right and left cervical spine lateral flexion. Tenderness was measured using a manual pressure algometer and was reported in kg/cm^2^\[[@B20]\]. Once identified by palpation and marked on the skin, the MTrP was subjected to vertical pressure with the pressure algometer at the rate of 1 kg/cm^2^per second. The subject was instructed to indicate verbally when the first instance of tenderness was felt. A single trial was used. Right and left cervical spine lateral flexions were measured using a cervical rangiometer \[[@B21]\] and were reported in degrees. A single trial was used. Both measures have proven reliable and valid in the assessment of MTrP\'s \[[@B9],[@B21]-[@B27]\]. Following the intervention, subjects were asked to rate their level of satisfaction on a verbal satisfaction scale (P = poor, F = fair, G = good, E = excellent).
Interventions
-------------
Six topical products were tested: three were ointments, two were roll-on gels and one a non-medicinal placebo cream (PLA) which served as the control. The comparative topicals were Professional Therapy MuscleCare Roll-on^a^(PTMC roll-on), Motion Medicine cream^b^(MM cream), Bengay Ultra Strength Muscle Pain ointment^c^(BG), Icy Hot Extra Strength Cream^d^(IH), and Biofreeze roll-on gel^e^(BF). All of these products were placed in identical 0.5 ounce white plastic screw top containers or 3 ounce generic white roll-on bottles. Only a coded letter was applied as a label. The master code for these products was kept with the clinic administrator and was unknown to all study participants and assessors.
Procedures
----------
Subjects were seated erect in a comfortable ergonomic chair in a private room. Subjects wore a nose clip for the entire duration of the study so they could not smell any of the scents from the topical analgesics. Subjects were not allowed to view the application of the topical analgesics and they were not permitted to look at their shoulder during the entire process. Assessor \#1 entered the room and palpated the subject\'s right shoulder in order to determine the presence and location of an MTrP in the upper trapezius muscle adjacent to the 7^th^cervical vertebrae and the 1^st^thoracic vertebrae. This was marked with a black dot. Assessor \#1 exited the room and assessor \#2 entered and performed the baseline testing of the outcome measures (Figure [1](#F1){ref-type="fig"}). The rangiometer was placed on the subject\'s head. From a neutral position, right and left active end-range lateral flexion measurements were recorded. The pressure algometer was applied over the marked trapezius trigger point for the baseline pressure reading. The subject was instructed to indicate when the pressure point was painful.
{#F1}
Following the initial outcome measurements, assessor \#1 re-entered the room and applied one of the six samples according to the randomization schedule which was revealed only at that time in the area of the marked pressure point (Figure [1](#F1){ref-type="fig"}). The subject was instructed to stay seated in the chair with little head movement for five to seven minutes, following which, assessor \#2 reassessed and recorded the pain and range of motion outcome measures. The study was conducted over a 7-day period.
Analyses
--------
Data were first analyzed descriptively\--baseline, post-intervention and change scores. Change scores for each of the three measurements (Pressure pain threshold, Right and Left lateral flexion) were analyzed using a 5-way ANOVA. Statistically significant ANOVA\'s (p \> 0.05) were then subjected to multiple t-tests with Holm\'s adjustment of the p-value. As such, the alpha levels were set at 0.007.
Results
=======
120 subjects were entered into the study, 20 in each group (63 females; ages 18-80). Fifty-six percent (56%) of patients presented with clinical shoulder or neck pain. Group demographics and clinical baseline values can be found in Table [1](#T1){ref-type="table"}. There were no significant differences regarding group demographics and clinical baseline values.
######
Group demographics and clinical baseline values
-------------------------------------------------------------------------------------------------------------------------
Variable Group\ Group\ Group\ Group\ Group\ Group\ Statistic
A B C D E F
------------------------------------ --------- --------- --------- --------- --------- --------- ------------------------
AGE 48.5\ 48.5\ 47.9\ 47.1\ 45.0\ 51.1\ NS\
(17.7) (14.7) (15.1) (14.7) (12.5) (15) F = 0.3\
p = 0.89
GENDER M = 8\ M = 8\ M = 11\ M = 11\ M = 11\ M = 7\ NS CSq = 4.0, p = 0.97
F = 12 F = 12 F = 9 F = 9 F = 9 F = 13
NECK and SHOULDER PAIN Y = 11\ Y = 12\ Y = 11\ Y = 10\ Y = 12\ Y = 11\ NS CSq = 2.2, p = 0.99
N 9 N = 8 N = 9 N = 10 N = 8 N = 9
Pressure Pain Threshold (kg/in^2^) 4.5\ 4.6\ 4.9\ 4.2\ 4.1\ 4.5\ NS\
(2.2) (1.5) (2.4) (1.3) (0.9) (2.4) F = 0.41\
p = 0.83
-------------------------------------------------------------------------------------------------------------------------
The pre- and post-treatment results for pressure threshold are displayed in Table [2](#T2){ref-type="table"}. ANOVA for intergroup differences was significant (F = 5.2, p = 0.002). Significant intra-group increases for BG, PTMC roll-on and MM cream were obtained with an increased threshold of 0.5 kg/cm^**2**^(+/-0.15), 0.72 kg/cm^**2**^(+/-0.17) and 0.47 Kg/cm^**2**^(+/-0.19) respectively.
######
Pressure pain thresholds.
----------------------------------------------------------------------------------------------
Group Pre-Tx\ Post-Tx\ Change\ Intra-group\ Inter-group\
(kg/cm^2^) (kg/cm^2^) (kg/cm^2^) Significance Significance\*
------- --------------- ---------------- ---------------- -------------- ---------------------
**A** 4.5 (+/-0.48) 5.0 (+/-0.56) 0.5 (+/-0.15) p = 0.002 A \> C (p = 0.0003)
**B** 4.6 (+/-0.33) 4.5 (+/-0.35) -0.45 (.12) p = 0.37 NS
**C** 4.9 (+/-0.53) 4.6 (+/-0.54) -0.3 (+/-0.19) p = 0.067 NS
**D** 4.2 (+/-0.28) 4.96 (+/-0.39) 0.72 (+/-0.17) p = 0.0002 D \> C (p = 0.002)\
D \> E (p = 0.006)
**E** 4.1 (+/-0.20) 4.4 (+/-0.32) 0.35 (+/-0.23) p = 0.07 NS
**F** 4.4 (+/-0.55) 4.9 (+/-0.59) 0.47 (+/-0.19) p = 0.01
----------------------------------------------------------------------------------------------
A = Bengay Ultra Strength Muscle Pain ointment, B = Biofreeze roll-on gel, C = Placebo, D = Professional Therapy Muscle Care roll-on, E = Icy Hot Extra Strength Cream, F = Motion Medicine
\*Bonferroni corrected p value = 0.007
With respect to the inter-group comparisons, PTMC roll-on showed significant increases in pressure threshold compared with PLA (p = 0.002) and IH (p = 0.006). In addition, BG demonstrated significant increases in pressure threshold compared with PLA (p = 0.0003).
The pre- and post-treatment results for right and left lateral flexion did not show significant increases in range of motion for any of the topical analgesics (Table [3](#T3){ref-type="table"}).
######
Right Lateral Flexion.
-----------------------------------------------------------------------
Group Pre-Tx\ Post-Tx\ Change\ Significance
(Degrees) (Degrees) (Degrees)
------- --------------- --------------- ---------------- --------------
**A** 39.4 (+/-3.3) 44.9 (+/-2.6) 5.45 (+/-1.7) NS
**B** 45.9 (+/-2.3) 46.7 (+/-2.9) .8 (+/-1.5) NS
**C** 40.6 (+/-2.8) 40.1 (+/-2.4) -0.45 (+/-1.1) NS
**D** 44.4 (+/-2.9) 48.3 (+/-3.0) 3.95 (+/-1.6) NS
**E** 46.9 (+/-2.7) 48.4 (+/-2.7) 1.4 (+/-1.5) NS
**F** 44.5 (+/-1.7) 44.7 (+/-1.6) 0.2 (+/-1.1) NS
-----------------------------------------------------------------------
A = Bengay Ultra Strength Muscle Pain ointment, B = Biofreeze roll-on gel, C = Placebo, D = Professional Therapy Muscle Care roll-on, E = Icy Hot Extra Strength Cream, F = Motion Medicine
Patient satisfaction ratings were high in each group and there were no significant differences between groups. There were no adverse reactions or complaints of pain aggravation reported in this study.
Discussion
==========
This study investigated the immediate effect of six different topical analgesic creams on pressure pain threshold at a single trapezius MTrP and on bilateral cervical lateral flexions. We found that the topical analgesics BG, PTMC roll-on and MM cream demonstrated significant and clinically important increases in pressure threshold when comparing short-term MTrP tenderness. In addition, PTMC roll-on was better than placebo and Icy Hot Extra Strength Cream in reducing short-term MTrP tenderness. BG was better than placebo at reducing short-term MTrP tenderness.
None of the groups demonstrated a statistically significantly increase in right or left lateral flexion.
The effectiveness of the topical analgesics that showed clinically significant improvements in cervical spine pressure threshold may be due to several factors. Eucalyptus oil, that was found in the PTMC solution, has been shown to transport active ingredients deep into the subcutaneous tissues \[[@B28],[@B29]\]. Camphor and menthol, found in both BG and PTMC, have been proven to provide immediate pain relief \[[@B19],[@B30],[@B31]\]. Glucosamine sulfate, chondroitin sulfate, dimethyl sulfoxide and Boswellia serrata extract, found in the PTMC roll-on formulation, have been shown to improve circulation and reduce inflammation, thus reducing pain in the short-term \[[@B19],[@B32],[@B33]\]. In addition, magnesium chloride, which is unique to the PTMC roll-on, has been shown to be effectively absorbed through the dermis into muscle \[[@B34],[@B35]\].
Topical analgesics are typically easy to self-apply. While this is a feature of many non-pharmacologic topicals, evidence of the effectiveness of the topical analgesics tested in this study is significantly lacking in the research literature. The results from the study demonstrate promising results for patients with myofascial pain. Since the topical analgesics are easy to use, they may improve the patients\' self-care through increased compliance and help alleviate the painful symptoms they experience from MTrPs.
Many clinicians use these topical analgesics during multiple sessions or in combination with other proven therapies. Many clinicians theorize that the combination therapy or multiple session uses will result in long-term benefits. The intervention in this study was applied in one single session with only short-term outcomes measured. A similar study has been published on the outcomes from a single manual therapy session for patients with neck pain \[[@B36]\]. A clinical prediction rule has been formulated for single-session response to manual therapy for neck pain. Our findings show similar benefits compared with the manual therapy neck pain study and provides support for further research. Future studies with multiple sessions over several treatment days or in combination with other therapies are needed to validate the effects of the topical analgesics commonly used in clinical settings.
Limitations
-----------
This study used participants with and without cervical spine and trapezius pain. This may have attenuated the outcomes measured in this study. Future research should separately evaluate pain and non-painful groups to eliminate any confounding factors.
Subjects were not excluded if they had prior experience, nor, was prior experience measured as an explanatory variable. However, our efforts to blind subjects to the nature of the topicals used should have reduced the effect of any prior treatment experience.
Only one trial of pressure algometery and cervical ranges of motion was obtained pre-post intervention for each patient. As there is likely to be some variability with examiner testing, multiple recordings should be obtained in future studies.
It is unknown if any of the significant results obtained immediately post-intervention would be sustained over a longer period of time. Future studies should measure longer term outcome measures, such as a few hours after treatment.
Conclusions
===========
This study demonstrated that some topical analgesic products do reduce myofascial pain or tenderness. The results of this study demonstrated that there were clinically important and significant differences between the topical analgesics tested. The BG, MM and the PTMC roll-on ointment demonstrated significant increases in pressure threshold levels. PTMC roll-on and BG were significantly superior to the placebo in the short-term reduction of myofascial tenderness. Furthermore, the PTMC roll-on demonstrated that it was significantly superior to the IH in the short-term reduction of myofascial tenderness.
There were no significant differences in range of motion testing for any of the topical analgesics. Future studies are needed to evaluate the effect of topical analgesics on the duration of effects of topical pain relief.
Endnotes
========
^a^Professional Therapy Muscle Care^®^Extra Strength Pain Relief Roll-on, Muscle Care Products (div. of Active & Innovative Inc.) Toronto, ON M4T 2V7
^b^MM Cream, Motion Medicine Inc., Calgary, AB T2Y 2Z7
^c^Bengay Ultra Strength Muscle Pain non-greasy, non-staining cream, Johnson & Johnson Inc. Markham, ON L3R 5L2, Canada.
^e^Icy Hot Extra Strength Pain Relieving Cream, Chattem Canada, Mississauga, ON L5N 2K7, Canada
^f^Biofreeze Pain Relieving Roll-on Gel with Ilex, Royal Cross, Concord, ON L4K 2L9, Canada
Competing interests
===================
The authors declare that they have no competing interests.
Authors\' contributions
=======================
DA have made substantial contributions to the acquisition of data, analysis and interpretation of data, have been involved in drafting the manuscript or revising it critically for important intellectual content and have given final approval of the version to be published. AH and CH have made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data, have been involved in drafting the manuscript or revising it critically for important intellectual content and have given final approval of the version to be published. HV made substantial contributions to the design of this study as well as to the analyses and interpretations of the findings. He was substantially involved in drafting the manuscript or revising it critically for important intellectual content and has given final approval of the version to be published. All authors read and approved the final manuscript.
Acknowledgements
================
We thank Roslyn Manchee for her help as project administrator.
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
Leptospirosis is a zoonotic infectious disease caused by bacteria of the genus *Leptospira* that affects domestic and wild animals. The disease is distributed worldwide and of great public health importance, especially in warm and humid climates. The bacterium is shed in the urine of infected animals and this is the main transmission route for human infection \[[@CR1]\]. The disease has also been recognized as one of the most important diseases in livestock, particularly in cattle, due to negative impacts on reproduction \[[@CR2],[@CR3]\].
It is well established that *Leptospira* infection in Chile is present both in domestic and wild animals \[[@CR4],[@CR5]\]. The apparent seroprevalences in different domestic animal species are high, ranging from 59 to 91% in cattle, 24% in goats, 7.1% sheep, 49% in equine, 70% in swine and 47% in wild mice \[[@CR5]\]. In a study from Southern Chile 162/361 (45%) serum samples from apparently healthy cattle were seropositive for *Leptospira* using the Microscopic Agglutination Test. The proportion of seropositive samples was highest for *Leptospira* serovar Hardjo (68%), followed by serovars Pomona (11%), Tarassovi (8.6%), Bratislava (1.9%), Canicola (1.9%), Icterohaemorrhagiae (1.9%) and Ballum (1.2%) \[[@CR4]\]. Recently, a study was carried out to determine *Leptospira* seroprevalence and to evaluate risk factors associated with seropositivity at herd level in smallholder bovine dairy herds in southern Chile, and 75% of the included herds (52/69) showed serological titers against one or more *Leptospira* serovar, where *Leptospira borgpetersenii* serovar Hardjo was the serovar most frequently (81%) reported from animals with positive results \[[@CR6]\].
Infection by *L. interrogans* serovar Hardjo (type Hardjoprajitno) in cattle has not been investigated previously in Chile. The importance of serovar Hardjoprajitno infection on the rate of abortion has been estimated to be 30% as opposed to what happens with *Leptospira borgpetersenii* Hardjo bovis where the rate reaches only a 3 to 10%. Additionally, acute infection of dairy cows with Leptospira *interrogans* Hardjoprajitno is associated with a drop in milk production \[[@CR7]\]. Therefore, the aim of the present study was to present both pathological and microbiological evidence of the *Leptospira interrogans* Hardjoprajitno virulence from its isolation and characterization from a calf that died of clinical leptospirosis.
Case presentation {#Sec2}
=================
Study animal {#Sec3}
------------
In a dairy herd in southern Chile, eleven out of 130 calves died after presenting clinical signs such as depression and haematuria. One of these calves, a female of eight months, was submitted to the Department of Animal Pathology at the Faculty of Veterinary Science, Universidad Austral de Chile, Valdivia, Chile for necropsy.
Pathological findings {#Sec4}
---------------------
Necropsy showed a marked yellowing pigmentation in all mucosal body openings and in the subcutaneous tissue, fat and muscles. There were also isolated petechia in the kidneys and the bladder contained approximately two liters of red-tinged urine (Figure [1](#Fig1){ref-type="fig"}).Figure 1**Calf with clinical sign and pathological findings consistent with** ***Leptospira*** **infection.**
A urine sample was collected by puncturing the bladder and sent to the Leptospirosis and Paratuberculosis Laboratory, Department of Biochemistry and Microbiology, at the Faculty of Sciences, Universidad Austral de Chile, Valdivia, Chile. All the procedures were in strict accordance with the recommendations in the Guide of Use of Animals for Research of Universidad Austral de Chile, approved by the Committee on the Ethics of Animals for Research ([www.uach.cl/direccion/ investigacion/uso_animales.htm](http://www.uach.cl/direccion/%20investigacion/uso_animales.htm)).
Bacteriological analysis {#Sec5}
------------------------
The urine sample was investigated using dark field microscopy and bacterial structures consistent with *Leptospira* were found. Thereafter, 200 μl of urine sample with four replicates was cultured in EMJH medium at 29°C \[[@CR8]\]. After a month of incubation, a positive result was reported with a typical Dinger ring growth.
*Leptospira* DNA was extracted from positive cultures and in order to identify *Leptospira* species, primers covering the most common pathogenic *Leptospira* species were used (664--665 *L. kirschneri* fla gene; 1280--1281 *L. interrogans* IS1500; 1805--1809 *L. borgpetersenii* IS1533) \[[@CR9]\]. The total PCR reaction was 50 μl, of which 5 μl was 10× Taq polymerase buffer (Promega, Madison, WI), 2 μl dNTPs (2.5 mM stock containing all four dNTPs) (Promega, Madison, WI), 0.5 U Taq Polymerase (Promega, Madison, WI), 1 μl (each) primer (stock concentration = 100 pmol/μl; final concentration 2 pmol/μl), 35.5 μl dH~2~O and 5 μl template. The PCR reactions considered 40 cycles of 94°C for 15 sec; 60°C for 30 sec and 68°C for 2 min. then 10°C hold. Negative and positive PCR controls were included as well as DNA extraction negative and positive controls.
To refine our understanding of the *Leptospira* specie and serovar associated with this clinical case, a Variable Number Tandem Repeat (VNTR) analysis was done. The VNTR primers were designed exclusively for use with *L. interrogans* \[[@CR10]-[@CR12]\]. PCR products for VNTR loci 4, 7, 10, 23, 27, 29, 30, 31, and 36 were assessed using the same PCR reaction as described above and the primers used were as previously reported \[[@CR12]\]. PCR products were separated by agarose gel electrophoresis and visualized, and their sizes were calculated by comparing with reference standards (100-bp ladder; Invitrogen, Carlsbad, CA) and with the literature \[[@CR10],[@CR11]\]. As a complement, we also amplified the gene *secY*, which is a house keeping gene that consists of alternating conserved and variable regions, making it suitable to deduce primers that generate amplicons with sufficient sequence heterogeneity to enable phylogenetic interpretation for *Leptospira* \[[@CR13]\]. A 202 bp product was amplified by conventional PCR in 25 μl mixture containing 5 μl diluted template (1:100), 0.2 μM each primers SecYIVF (5′-GCGATTCAGTTTAATCCTGC-3′) and SecYIV (5′-GAGTTAGAGCTCAAATCTA-AG-3′), 0.625 U GoTaq Flexi DNA Polymerase in 1X Green Buffer GoTaq (Promega, Madison, WI), 3.0 mM MgCl2, 0.3 mM dNTPs (Promega, Madison, WI), and 400 ng mL-1 bovine serum albumin (BSA; BioLabs, Ipswich, England). Cycle conditions included an initial denaturation step at 95°C for 5 min followed by 40 cycles at 94°C for 1 min, 57°C for 1 min and 72°C for 1 minute and a final elongation step at 72°C for 10 minutes. The PCR products obtained were separated on 1.5% agarose gel, stained with Gel Red (GelRed, Biotium Inc, Hayward, U.S), excised and purified using a commercial kit (E.Z.N.A® Gel Extraction Kit, Omega Bio-Tek, Norcross, U.S). Amplicons were sequenced by Macrogen Inc (Seoul, Korea). The consensus nucleotide sequence obtained in this study was compared with *secY* gene of *Leptospira interrogans* serovar Hardjo prajitno (GenBank accession number EU357983.1). DNA alignments were done using clustalW tools (<http://www.ebi.ac.uk/Tools/msa/clustalw2>).
Results and discussion {#Sec6}
======================
The isolate obtained in this study was confirmed by PCR as *L. interrogans* and its VNTR profile properly matched with *L. interrogans* type Hardjo prajitno (Figure [2](#Fig2){ref-type="fig"}). The *secY* gene alignment done by clustalW did reveal sequence identity strain belonging to the species *L. interrogans* serovar Hardjo prajitno (Figure [3](#Fig3){ref-type="fig"}). This confirms that *L. interrogans* type Hardjo prajitno is associated with acute infection of cattle in Chile. Previous studies have shown that the abortion rate after *Leptospira borgpetersenii* serovar Hardjo bovis infection is 3 to 10% whereas the rate increases up to 30% for *L.* Hardjo prajitno infection \[[@CR3]-[@CR7]\], which underscores the importance of this serovar. The clinical information in the presented study was conveniently complemented with bacteriological findings to describe the isolated strain affecting the clinical case presented.Figure 2**PCR analysis of the polymorphism of nine representative VNTR loci.** Amplification was performed on the VNTR 4, 7, 10, 23, 27, 29, 30, 31, 36 loci of *L. interrogans* strains.Figure 3**ClustalW aligment for 202 bp fragment,** ***SecY*** **gen.**
The typing method used based on VNTR polymorphism provided a rapid characterization together with the highly discriminatory power reported \[[@CR10],[@CR11]\] to identify *L. interrogans* serovars using clinical specimens. Furthermore, the use of *secY* gene in combination with the latter allowed a more robust result due to its great phylogenetic potential \[[@CR13]\].
Conclusions {#Sec7}
===========
The present finding represents the first isolation confirmed as *L. interrogans* serovar Hardjo prajitno from cattle with clinical disease in Chile. The importance of this serovar in Chilean cattle needs to be investigated further. This information add to the value of serologic results commonly reported, which encourage vaccination improvements to match circulating strains. The latter is based on a previous published study on serological cross-reactivity between Hardjo bovis and Hardjoprajitno serovars, which implies similar antigenic determinants; although with substantial genomic differences as well as in their pathogenicity in the bovine specie \[[@CR14]\].
Due to above mentioned, we emphasize the need to isolate, preserve and characterized strains of *Leptospira* in order to improve and standardize currently available diagnostic techniques. This will help to improve our understanding of the epidemiology and impact of this infection as well as to identify optimal option for surveillance and control.
**Competing interests**
The authors declare that they have no competing interests.
**Authors' contributions**
MS: lab work and draft writing; BO lab work and draft writing; MM: pathological study and draft writing: ES: lab work and draft writing; GR: draft writing; SB: draft writing, CE: molecular analysis; CMZ: draft writing. All authors read and approved the final manuscript.
This work was supported by DID GRANT S-2012-19. We also wanted to acknowledge Dr Richard Zuerner for provide technical help, writing assistance, and lab support at the Department of Biomedical Sciences and Veterinary Public Health, Swedish University of Agricultural Sciences.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
Achalasia is a rare esophageal motility disorder involving the smooth muscle layer of the esophagus and the lower esophageal sphincter (LES). It is characterized by incomplete LES relaxation, increased LES tone, and nonperistaltic contractions of the esophageal body. The main symptoms are dysphagia, regurgitation, and vomiting. Chest pain, coughing, and loss of weight can be encountered also.
Achalasia can occur at any age but it usually presents between the third and fifth decade of life with the same rate between men and women \[[@B1]\] and a prevalence of 8 per million population \[[@B4]\]. The onset is insidious and that could be the reason why it rarely coexists with pregnancy. In pregnant women, it has been associated with maternal malnutrition and death, as well as preterm delivery, fetal growth restriction, and fetal demise \[[@B2]\].
Due to the rarity of achalasia in pregnancy, there are no data regarding perinatal management. We present a case of esophageal achalasia in the last trimester of pregnancy, managed with success with total parenteral nutrition (TPN) administered through a peripheral line.
2. Case Presentation {#sec2}
====================
A 38-year-old Caucasian woman, gravida 3 para 2, presented to our hospital with an intrauterine pregnancy at 29 weeks because of occasional vomiting, dysphagia for liquids and solids, anorexia, mild uterine contractions, and a nine kg weight loss over a 4-week period. Her first two children were delivered vaginally at term. Initial physical examination on admission revealed a fatigued woman with a weight of 62 kg, height of 178 cm, blood pressure of 110/75 mmHg, pulse 80 beats per minute, and temperature of 36.5°C. An enlargement of the left neck region was noted on palpation of the thyroid gland, with no enlarged peripheral lymph nodes. The rest of the physical examination was unremarkable. All laboratory investigations, including full blood count, liver function tests, and thyroid function tests were within normal range. Urine test results were unremarkable.
The patient was admitted on the antenatal ward for observation and administration of i.v. fluids, antiemetics, and vitamins. A viable singleton pregnancy was confirmed by transabdominal ultrasonography with umbilical and uterine artery Doppler ultrasound velocimetry within normal limits. An estimated fetal weight (EFW) of 1681 gr was calculated via ultrasonography.
Imaging studies included an ultrasound of the thyroid gland showing a normal-sized gland with the presence of a mass displacing the left thyroid lobe and the vessels of the left neck region. An MRI of the neck revealed a normal sized thyroid, larynx and lymph nodes, and a marked esophageal dilatation with the presence of food remnants along its length with displacement of the trachea to the right (Figures [1](#fig1){ref-type="fig"} and [2](#fig2){ref-type="fig"}).
The patient\'s symptoms did not improve over the next two days and a gastroenterologist was consulted. An upper gastrointestinal (GI) endoscopy showed an enlarged tortuous esophageal lumen, incomplete LES relaxation, esophagitis with difficult passage of the endoscope into the stomach, and no evidence of intraluminal esophageal compression. Examination of the second part of the duodenum was unremarkable. Esophageal manometry was also performed, showing an increased LES pressure of 44 mmHg with incomplete relaxation, consistent with achalasia.
Treatment options to improve the nutritional status of both mother and fetus were discussed with the patient, including calcium channel antagonists, nitrates, pneumatic dilation, botulinum toxin injection, and parenteral nutrition (TPN). The woman denied any intervention during pregnancy and opted for TPN through a peripheral line of 2000 mL/day, with 1215 Kcal/day. During hospitalization, monitoring of the fetal condition was performed with cardiotocography and biophysical profile twice weekly. Fetal growth was within normal range and maternal body weight increased by two kg over the next eight weeks. An active, 2160 g male baby was delivered by a cesarean section at 37 weeks of gestation, with an Apgar score of 9. The baby was not admitted to neonatal intensive care unit and showed normal growing rate. The patient underwent laparoscopic myotomy and fundoplication seven days postpartum. Postoperatively, symptomatology from the digestive tract improved significantly and her appetite increased. She gained five kg of body weight in the next month.
3. Discussion {#sec3}
=============
In pregnancy, the symptomatology of achalasia can be aggravated due to the physiologic changes in the LES and easily confused with gastroesophageal reflux or gravidarum. The diagnosis can be delayed considerably and the resulting maternal malnutrition can lead to adverse perinatal outcomes, such as fetal growth restriction, preterm delivery, and even fetal death \[[@B2]\]. Maternal mortality has also been reported \[[@B3]\]. Due to the rarity of this disorder, specific guidelines of management and treatment in pregnancy do not exist \[[@B4]\]. In addition, the age of onset of achalasia is varied and the time interval between the onset and clinical symptomatology can be unpredictable (from one to five years) \[[@B5]\].
The management options depend on the gestational age and severity of symptomatology. These include diet modification and medical therapy (calcium channel blockers and nitrates) in cases of late onset or mild disease, but they should be used with caution. The severity of achalasia in our patient precluded the use of medical therapy. Pneumatic dilation and injection of botulinum toxin in the LES are considered more effective measures until the delivery of the baby \[[@B6], [@B7]\]. However, they do pose some risks, such as esophageal rupture (3% risk) \[[@B8]\] and miscarriage after the use of botulinum toxin (category c medication) \[[@B9]\]. Wataganara et al. \[[@B10]\] have described the use of intersphincteric injection of botulinum toxin with the delivery of a healthy baby at 37 weeks by cesarean section. Various authors have reported cases of successful treatment of esophageal achalasia with endoscopic pneumatic balloon dilation with good perinatal outcomes \[[@B7], [@B11]--[@B13]\].
Cardiomyotomy (open or endoscopic) is another option but it carries the risk of general anesthesia and surgery and is contraindicated in pregnancy. It can be offered to the patient postpartum as a definite treatment. Palanivelu et al. \[[@B15]\] described the successful management of achalasia with laparoscopic Heller\'s myotomy during the second trimester of pregnancy. Ohno et al. \[[@B2]\] concluded that surgical myotomy can improve pregnancy outcome in women previously suffering from esophageal achalasia. Finally, Roca et al. \[[@B16]\] reported the use of an esophageal self-expanding prosthesis in a 36-year-old woman at 26 weeks gestation with good outcome ([Table 1](#tab1){ref-type="table"}).
In our patient, where severe maternal malnutrition was noted, TPN can be used as a source of caloric source of glucose, lipids, electrolytes, and trace elements \[[@B17], [@B18]\]. In the past, TPN was thought to be the cause of uterine contractions and preterm labour \[[@B19]\]. Recent data suggest that increased preterm delivery rates observed in the past were a reflection of the underlying maternal condition, rather than direct consequence of the lipid component formulations for TPN \[[@B17]\]. Despite the risk of infection and thrombosis, maternal and neonatal outcome are not compromised by the use of TPN \[[@B20], [@B21]\]. Russo-Steiglitz et al. \[[@B19]\] had studied the nutritional status and maternal/neonatal outcome in ten patients with severe hyperemesis treated with TPN. They did not notice an increased incidence of intrauterine growth restriction (IUGR), preterm labor, or placental insufficiency in any of the patients studied. Additionally, recent progress in nutritional supplementation has shown that it can be used to defer definitive treatment after delivery \[[@B14]\]. Our case is supporting the importance of conservative management of esophageal achalasia during pregnancy with TPN, when definite surgical treatment cannot be applied.
4. Conclusion {#sec4}
=============
Achalasia in pregnancy remains a rarely encountered condition. Esophageal achalasia should be suspected in pregnant women presenting with dysphagia for solids and liquids, occasional vomiting, and loss of weight. The management and treatment needs to be individualized and its advantages and disadvantages discussed thoroughly with the patient. Total parenteral nutrition in pregnancy is a safe option of conservative management of achalasia and it has not been associated with adverse perinatal outcomes such as preterm delivery, growth restriction, or increased perinatal mortality. It remains a valid and effective choice of management to any other intervention during pregnancy, especially when the patient does not opt for a surgical treatment.
{#fig1}
{#fig2}
######
Reported cases of achalasia in pregnancy with the type of treatment and pregnancy outcome (from 1969 till 2010).
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Author Year Number of cases Age of pt/weeks gestation Type of treatment Outcome
------------------- ------ ----------------- --------------------------- -------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------------
Paulsen et al. 2010 1 34 yo/? Balloon dilation Uncomplicated birth
Wataganara et al. 2009 1 39 yo/33 w Botulinum toxin 35 w C-section
Diaz Roca et al. 2009 1 36 yo/26 w Self expanding prosthesis Uneventful delivery
Palavinelu et al. 2008 1 24 yo/2nd trimester Laparoscopic Heller\'s myotomy Healthy baby
Pulanic et al. 2008 1 30 yo/26 w Pneumatic dilation Vaginal delivery 38 w
Ohno et al. 2000 1 34 yo/27 w Surgical myotomy postpartum Intrauterine fetal death
Kalish et al. 1999 1 42 yo/31 w Antifungal medication Spontaneous vaginal delivery 38 w
Fassina et al. 1995 1 23 yo/24 w Unexplained sudden maternal death (6-month pregnancy), megaesophagus
Fiest et al. 1993 1 24 yo/8 w Balloon dilation Spontaneous vaginal delivery 35 w (healthy infant)
Satin et al. 1992 1 28 yo/38 w Pneumatic dilation Induced vaginal delivery 38 w (healthy infant)
Mayberry and\ 1987 20 18 yo--45 yo\ Comparison of reproductive histories of women with achalasia with those of a control group No significant difference in the number of live births of patients versus controls 3 miscarriages after diagnosis
Atkinson (mean 32 yo)/?
Clemendor et al. 1969 10 22 yo--37 yo\ 1case Pneumatic dilation (1st report)\ Living offspring in only 5 cases (the rest: 2 terminations, 2 stillbirths, 1 premature delivery).
(mean 29.1 yo)/24 w--30 w 5 cases Bougie dilation\
3 cases medical treatment\
1 case no treatment
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
[^1]: Academic Editors: T. Guvenal, T. Levy, and S. Rasmussen
| {
"pile_set_name": "PubMed Central"
} |
1.. Introduction
================
Human segmentation in uncontrolled environments is a hard task because of the constant changes produced in natural scenes: illumination changes, moving objects, changes in the point of view, occlusions, just to mention a few. Because of the nature of the problem, a common way to proceed is to discard most part of the image so that the analysis can be performed on a reduced set of small candidate regions. In \[[@b1-sensors-12-15376]\], the authors propose a full-body detector based on a cascade of classifiers \[[@b2-sensors-12-15376]\] using HOG features. This methodology is currently being used in several works related to the pedestrian detection problem \[[@b3-sensors-12-15376]\]. GrabCut \[[@b4-sensors-12-15376]\] has also shown high robustness in Computer Vision segmentation problems, defining the pixels of the image as nodes of a graph and extracting foreground pixels via iterated Graph Cut optimization. This methodology has been applied to the problem of human body segmentation with high success \[[@b5-sensors-12-15376],[@b6-sensors-12-15376]\]. In the case of working with sequences of images, this optimization problem can also be considered to have temporal coherence. In the work of \[[@b7-sensors-12-15376]\], the authors extended the Gaussian Mixture Model (GMM) of GrabCut algorithm so that the color space is complemented with the derivative in time of pixel intensities in order to include temporal information in the segmentation optimization process. However, the main problem of that method is that moving pixels corresponds to the boundaries between foreground and background regions, and thus, there is no clear discrimination.
Once a region of interest is determined, pose is often recovered by the determination of the body limbs together with their spatial coherence (also with temporal coherence in case of image sequences). Most of these approaches are probabilistic, and features are usually based on edges or "appearance". In \[[@b8-sensors-12-15376]\], the author propose a probabilistic approach for limb detection based on edge learning complemented with color information. The image of probabilities is then formulated in a Conditional Random Field (CRF) scheme and optimized using belief propagation. This work has obtained robust results and has been extended by other authors including local GrabCut segmentation and temporal refinement of the CRF model \[[@b5-sensors-12-15376],[@b6-sensors-12-15376]\].
In this paper, we propose a full-automatic Spatio-Temporal GrabCut human segmentation methodology, which benefits from the combination of tracking and segmentation. First, subjects are detected by means of a HOG-based cascade of classifiers. Face detection and skin color model are used to define a set of seeds used to initialize GrabCut algorithm. Spatial information is taken into account by means of Mean Shift clustering, whereas temporal information is considered taking into account the pixel probability membership to an historical of Gaussian Mixture Models. Moreover, the methodology is combined with Shape and Active Appearance Models (AAM) to define three different meshes of the face, one near frontal view, and the other ones near lateral views. Temporal coherence and fitting cost are considered in conjunction with GrabCut segmentation to allow a smooth and robust face fitting in video sequences. Finally, the limb detection and a CRF model are applied on the obtained segmentation, showing high robustness capturing body limbs due to the accurate human segmentation. The main limitation of our approach is that it depends on a correct detection of the person and his/her face, in order to get the desired result. In order to test the proposed methodology, we use public datasets and present a new Human Limb dataset useful for human segmentation, limb detection, and pose recovery purposes.
The rest of the paper is organized as follows: Section 2 describes the proposed methodology, presenting the spatio-temporal GrabCut segmentation, the AAM for face fitting, and the pose recovery methodology. Experimental results on public and novel datasets are performed in Section 3. Finally, Section 4 concludes the paper.
2.. Full-Body Pose Recovery
===========================
In this section, we present the Spatio-Temporal GrabCut methodology to deal with the problem of automatic human segmentation in video sequences. Then, we describe the Active Appearance Models used to recover the face, and the body pose recovery methodology based on the approach of \[[@b8-sensors-12-15376]\]. All methods presented in this section are combined to improve final segmentation and pose recovery. [Figure 1](#f1-sensors-12-15376){ref-type="fig"} illustrates the different modules of the project.
2.1.. GrabCut Segmentation
--------------------------
In \[[@b4-sensors-12-15376]\], the authors proposed an approach to find a binary segmentation(background and foreground) of an image by formulating an energy minimization scheme as the one presented in \[[@b9-sensors-12-15376]--[@b11-sensors-12-15376]\], extended using color instead of just gray-scale information. Given a color image *I*, let us consider the array *z* = (*z*~1~, ..., *z~n~*, ..., *z~N~*) of *N* pixels where *z~i~* = (*R~i~*, *G~i~*, *B~i~*), *i* ∈ \[1, ..., *N*\] in RGB space. The segmentation is defined as array ***α*** = (*α*~1~, ...*α~N~*), *α~i~* ∈ {0, 1}, assigning a label to each pixel of the image indicating if it belongs to background or foreground. A trimap *T* is defined by the user---in a semi-automatic way---consisting of three regions: *T~B~*, *T~F~* and *T~U~*, each one containing initial background, foreground, and uncertain pixels, respectively. Pixels belonging to *T~B~* and *T~F~* are clamped as background and foreground respectively---which means GrabCut will not be able to modify these labels, whereas those belonging to *T~U~* are actually the ones the algorithm will be able to label. Color information is introduced by GMMs. A full covariance GMM of *K* components is defined for background pixels (*α~i~* = 0), and another one for foreground pixels (*α~j~* = 1), parametrized as follows $$\mathbf{\theta} = \left\{ \pi(\alpha,k),\mu(\alpha,k),\sum(\alpha,k),\alpha \in \left\{ 0,1 \right\},k = 1‥K \right\},$$being *π* the weights, *μ* the means and Σ the covariance matrices of the model. We also consider the array **k** = {*k*~1~, ..., *k~i~*, ...*k~N~*}, *k~i~* ∈ {1, ...*K*}, *i* ∈ \[1, ..., *N*\] indicating the component of the background or foreground GMM (according to *α~i~*) the pixel *z~i~* belongs to. The energy function for segmentation is then $$\mathbf{E}(\alpha,k,\theta,z) = \mathbf{U}(\alpha,k,\theta,z) + \mathbf{V}(\alpha,z),$$where **U** is the likelihood potential, based on the probability distributions *p*(·) of the GMM: $$\mathbf{U}(\mathbf{\alpha},\mathbf{k},\mathbf{\theta},\mathbf{z}) = \sum\limits_{i}{- \log p(z_{i} \mid \alpha_{i},k_{i},\mathbf{\theta}) - \log\pi(\alpha_{i},k_{i})}$$and *V* is a regularizing prior assuming that segmented regions should be coherent in terms of color, taking into account a neighborhood *C* around each pixel $$\mathbf{V}(\mathbf{\alpha},\mathbf{z}) = \gamma\sum\limits_{\{ m,n\} \in C}{\lbrack\alpha_{n} \neq \alpha_{m}\rbrack\exp( - \beta\left\| {z_{m} - z_{n}} \right\|^{2})}$$
With this energy minimization scheme and given the initial trimap *T*, the final segmentation is performed using a minimum cut algorithm \[[@b9-sensors-12-15376],[@b10-sensors-12-15376],[@b12-sensors-12-15376]\]. The classical semi-automatic GrabCut algorithm is summarized in Algorithm 1. **Algorithm 1 Original GrabCut algorithm.** 1:Trimap *T* initialization with manual annotation. 2:Initialize *α~i~* = 0 for *i* ∈ *T~B~* and *α~i~* = 1 for *i* ∈ *T~U~* ∪ *T~F~*. 3:Initialize Background and Foreground GMMs from sets *α~i~* = 0 and *α*~i~ = 1 respectively, with *k*-means. 4:Assign GMM components to pixels. 5:Learn GMM parameters from data z. 6:Estimate segmentation: Graph-cuts. 7:Repeat from step 4, until convergence.
2.2.. Automatic Initialization
------------------------------
Our proposal is based on the previous GrabCut framework, focusing on human body segmentation, being fully automatic, and extending it by taking into account temporal coherence. We refer to each frame of the video as *f~t~*, *t* ∈ {1, ..., *M*} being *M* the length of the sequence. Given a frame *f~t~*, we first apply a person detector based on a cascade of classifiers using HOG features \[[@b1-sensors-12-15376]\]. Then, we initialize the trimap *T* from the bounding box *B* retuned by the detector: *T~U~* = {*z~i~* ∈ *B*}, *T~B~* = {*z~i~* ∉ *B*}. Furthermore, in order to increase the accuracy of the segmentation algorithm, we include Foreground seeds exploiting spatial and appearance prior information. On one hand, we define a small central rectangular region *R* inside *B*, proportional to *B* in such a way that we are sure it corresponds to the person. Thus, pixels inside *R* are set to foreground. On the other, we apply a face detector based on a cascade of classifiers using Haar-like features \[[@b2-sensors-12-15376]\] over *B*, and learn a skin color model *h~skin~* consisting of a histogram over the *Hue* channel of the *HSV* image representation. All pixels inside *B* fitting in *h~skin~* are also set to foreground. Therefore, we initialize *T~F~* = {*z~i~* ∈ *R*} ∪ {*z~i~* ∈ *δ*(*z~i~*, *h~skin~*)}, where *δ* returns the set of pixels belonging to the color model defined by *h~skin~*. An example of seed initialization is shown in [Figure 2(b)](#f2-sensors-12-15376){ref-type="fig"}.
2.3.. Spatial Extension
-----------------------
Once we have initialized the trimap, we can apply the iterative minimization algorithm shown in steps 4 to 7 of original GrabCut (Algorithm 1). However, instead of applying k-means for the initialization of the GMMs we propose to use Mean-Shift clustering, which also takes into account spatial coherence. Given an initial estimation of the distribution modes *m~h~*(**x**^0^) and a kernel function *g*, Mean-shift iteratively updates the mean-shift vector with the following formula: $$\mathbf{m}_{h}(x) = \frac{\sum_{i = 1}^{n}{x_{i}g\left( \left\| \frac{x - x_{i}}{h} \right\|^{2} \right)}}{\sum_{i = 1}^{n}{g\left( \left\| \frac{x - x_{i}}{h} \right\|^{2} \right)}}$$until it converges, where **x***~i~* contains the value of pixel *z~i~* in CIELuv space and its spatial coordinates, and returns the centers of the clusters (distribution modes) found. After convergence, we obtain a segmentation ***α****^t^* and the updated foreground and background GMMs ***θ****^t^* at frame *f~t~*, which are used for further initialization at frame *f~t~*~+1~. The result of this step is shown in [Figure 2(c)](#f2-sensors-12-15376){ref-type="fig"}. Finally, we refine the segmentation of frame *f~t~* eliminating false positive foreground pixels. By definition of the energy minimization scheme, GrabCut tends to find convex segmentation masks having a lower perimeter, given that each pixel on the boundary of the segmentation mask contributes on the global cost. Therefore, in order to eliminate these background pixels (commonly in concave regions) from the foreground segmentation, we re-initialize the trimap *T* as follows $$\begin{array}{lll}
T_{B} & = & {\{ z_{i} \mid \alpha_{i} = 0\} \cup \left\{ {z_{i} \mid \frac{\underset{k = t - j}{\overset{t}{\mathbf{\sum}}}{p(z_{i} \mid \alpha_{i} = 0,k_{i},\mathbf{\theta}^{k})}}{j} > \frac{\underset{k = t - j}{\overset{t}{\mathbf{\sum}}}{p(z_{i} \mid \alpha_{i} = 1,k_{i},\mathbf{\theta}^{k})}}{j}} \right\}} \\
T_{F} & = & \left\{ z_{i} \in \delta(z_{i},h_{\textit{skin}}) \right\} \\
T_{U} & = & {\left\{ z_{i} \mid \alpha_{i} = 1 \right\}\backslash T_{B}\backslash T_{F}} \\
\end{array}$$where the pixel background probability membership is computed using the GMM models of previous *j* segmentations. This formulation can also be extended to detect false negatives. However, in our case we focus on false positives since they appear frequently in the case of human segmentation. The result of this step is shown in [Figure 2(d)](#f2-sensors-12-15376){ref-type="fig"}. Once the trimap has been redefined, false positive foreground pixels still remain, so the new set of seeds is used to iterate again GrabCut algorithm, resulting in a more accurate segmentation, as we can see in [Figure 2(e)](#f2-sensors-12-15376){ref-type="fig"}.
2.4.. Temporal Extension
------------------------
Considering *A* as the binary image representing ***α*** at *f~t~* (the one obtained before the refinement), we initialize the trimap for *f~t~*~+1~ as follows $$\begin{array}{lll}
T_{F} & = & \left\{ z_{i} \in I \mid z_{i} \in A \ominus ST_{e},\alpha(z_{i}) = 1 \right\} \\
T_{U} & = & {\left\{ z_{i} \in I \mid z_{i} \in A \oplus ST_{d},\alpha(z_{i}) = 1 \right\}\backslash T_{F}} \\
T_{B} & = & {\left\{ z_{i},z_{i} \in I \right\}\backslash(T_{F} \cup T_{U})} \\
\end{array}$$where ⊖ and ⊕ are erosion and dilation operations with their respective structuring elements *ST~e~* and *ST~d~*, *α~i~* := *α*(*z~i~*), and \\ represents the set difference operation. The structuring elements are simple squares of a given size depending on the size of the person and the degree of movement we allow from *f~t~* to *f~t~*~+1~, assuming smoothness in the movement of the person. An example of a morphological mask is shown in [Figure 2(f)](#f2-sensors-12-15376){ref-type="fig"}. Spatial information could be also included in the mean-shift algorithm in conjunction with color and spatial information. However, we included this information explicitly to be anisotropic. The whole segmentation methodology is detailed in the ST-GrabCut Algorithm 2. **Algorithm 2 Spatio-Temporal GrabCut algorithm.** 1:Person detection on *f*~1~. 2:Face detection and skin color model learning. 3:Trimap *T* initialization with detected bounding box and learnt skin color model. 4:Initialize *α~i~* = 0 for *i* ∈ *T~B~* and *α~i~* = 1 for *i* ∈ *T~U~* ∪ *T~F~*. 5:Initialize Background and Foreground GMMs from sets *α~i~* = 0 and *α~i~* = 1 respectively, with Mean-shift. 6:**for** *t* = 1 ... *M* 7: Person detection on *f~t~*. 8: Assign GMM components to pixels of *f~t~*. 9: Learn GMM parameters from data z. 10: Estimate segmentation: Graph-cuts. 11: Repeat from step 8, until convergence. 12: Re-initialize trimap *T* ([Equation (6)](#FD6){ref-type="disp-formula"}). 13: Assign GMM components to pixels. 14: Learn GMM parameters from data z. 15: Estimate segmentation: Graph-cuts. 16: Repeat from step 12, until convergence. 17: Initialize trimap *T* using segmentation obtained in step 11 after convergence ([equation 7](#FD7){ref-type="disp-formula"}) for *f~t~*~+1~. 18:**end for**
2.5.. Face Fitting
------------------
Once we have properly segmented the body region, the next step consists of fitting the face and the body limbs. For the case of face recovery, we base our procedure on mesh fitting using AAM, combining Active Shape Models and color and texture information \[[@b13-sensors-12-15376]\].
AAM is generated by combining a model of shape and texture variation. First, a set of points are marked on the face of the training images that are aligned, and a statistical shape model is build \[[@b14-sensors-12-15376]\]. Each training image is warped so the points match those of the mean shape. This is raster scanned into a texture vector, **g**, which is normalized by applying a linear transformation, **g** ↦ (**g** − *μ~g~***1**)/*σ~g~*, where **1** is a vector of ones, and *μ~g~* and *σ~g~* are the mean and variance of elements of **g**. After normalization, **g***^T^***1** = 0 and \|**g**\| = 1. Then, principal component analysis is applied to build a texture model. Finally, the correlations between shape and texture are learnt to generate a combined appearance model. The appearance model has parameter **c** controlling the shape and texture according to $$x = \overline{x} + \mathbf{Q}_{s}\mathbf{c}$$ $$g = \overline{g} + \mathbf{Q}_{g}\mathbf{c}$$where *x̄* is the mean shape, *ḡ* the mean texture in a mean shaped patch, and **Q***~s~*, **Q***~g~* are matrices designing the modes of variation derived from the training set. A shape **X** in the image frame can be generated by applying a suitable transformation to the points, **x : X** = *S~t~*(**x**). Typically, *S~t~* will be a similarity transformation described by a scaling *s*, an in-plane rotation, *θ*, and a translation (*t~x~*, *t~y~*).
Once constructed the AAM, it is deformed on the image to detect and segment the face appearance as follows. During matching, we sample the pixels in the region of interest ***g****~im~* = *T~u~*(**g**) = (*u*~1~ + 1)**g***~im~* + *u*~2~**1**, where **u** is the vector of transformation parameters, and project into the texture model frame, $g_{s} = T_{u}^{- 1}(g_{im})$. The current model texture is given by **g***~m~* = *ḡ* + **Q***~g~***c**, and the difference between model and image (measured in the normalized texture frame) is as follows $$\mathbf{r}(\mathbf{p}) = \mathbf{g}_{s} - \mathbf{g}_{m}$$
Given the error *E* = \|**r**\|^2^, we compute the predicted displacements *δ***p** = −**Rr**(**p**), where $\mathbf{R} = \left( {\frac{\partial\mathbf{r}^{T}}{\partial\mathbf{p}}\frac{\partial\mathbf{r}}{\partial\mathbf{p}}} \right)^{- 1}\frac{\partial\mathbf{r}^{T}}{\partial\mathbf{p}}$. The model parameters are updated **p** ↦ **p** + *kδ***p**, where initially *k* = 1. The new points **X**′ and model frame texture $\mathbf{g}_{m}^{\prime}$ are estimated, and the image is sampled at the new points to obtain $\mathbf{g}_{mi}^{\prime}$ and the new error vector $\mathbf{r}\prime = T_{u\prime}^{- 1}(g_{im}^{\prime}) - g_{m}^{\prime}$. A final condition guides the end of each iteration: if \|**r**′\|^2^ \< *E*, then we accept the new estimate, otherwise, we set to *k* = 0.5, *k* = 0.25, and so on. The procedure is repeated until no improvement is made to the error.
With the purpose to discretize the head pose between frontal face and profile face, we create three AAM models corresponding to the frontal, right, and left view. Aligning every mesh of the model, we obtain the mean of the model. Finally, to determine the class of a fitted face by AAM models, that is given by its proximity to the closest mean model.
Taking into account the discontinuity that appears when a face moves from frontal to profile view, we use three different AAM corresponding to three meshes of 21 points: frontal view ℑ*~F~*, right lateral view ℑ*~R~*, and left lateral view ℑ*~L~*. In order to include temporal and spatial coherence, meshes at frame *f~t~*~+1~ are initialized by the fitted mesh points at frame *f~t~*. Additionally, we include a temporal change-mesh control procedure, as follows $$\Im^{t + 1} = \min_{\Im^{t + 1}}\left\{ E_{\Im_{F}},E_{\Im R},E_{\Im L} \right\},\Im^{t + 1} \in \nu(\Im^{t})$$where *ν*(ℑ*^t^*) corresponds to the meshes contiguous to the mesh ℑ*^t^* fitted at time *t* (including the same mesh), and *E*~ℑ~*~i~* is the fitting error cost of mesh ℑ*~i~*. This constraint avoids false jumps and imposes smoothness in the temporal face behavior (e.g., a jump from right to left profile view is not allowed).
In order to obtain more accurate pose estimation, after fitting the mesh, we take advantage of its variability to differentiate among a set of head poses. Analyzing the spatial configuration of the 21 landmarks that composes a mesh, we create a new training set divided in five classes. We define five different head poses as follows: right, middle-right, frontal, middle-left, and left. In the training process, every mesh has been aligned, and PCA is applied to save the 20 most representative eigenvectors. Then, a new image is projected to that new space and classified to one of the five different head poses according to a 3-Nearest Neighbor rule.
[Figure 3](#f3-sensors-12-15376){ref-type="fig"} shows examples of the AAM model fitting and pose estimation in images (obtained from \[[@b15-sensors-12-15376]\]) for the five different head poses.
2.6.. Pose Recovery
-------------------
Considering the refined segmented body region obtained using the proposed ST-GrabCut algorithm, we construct a pictorial structure model \[[@b16-sensors-12-15376]\]. We use the method of Ramanan \[[@b6-sensors-12-15376],[@b8-sensors-12-15376]\], which captures the appearance and spatial configuration of body parts. A person\'s body parts are tied together in a tree-structured conditional random field. Parts, *l~i~*, are oriented patches of fixed size, and their position is parameterized by location (*x*, *y*) and orientation *ϕ*. The posterior of a configuration of parts *L* = *l~i~* given a frame *f~t~* is $$P(L \mid f_{t}) \propto \exp\left( {\sum\limits_{(i,j) \in E}{\Psi(l_{i},l_{j})} + \sum\limits_{i}{\Phi(l_{i} \mid f_{t})}} \right)$$
The pair-wise potential Ψ(*l~i~*, *l~j~*) corresponds to a spatial prior on the relative position of parts and embeds the kinematic constraints. The unary potential Φ(*l~i~*∣*I*) corresponds to the local image evidence for a part in a particular position. Inference is performed over tree-structured conditional random field.
Since the appearance of the parts is initially unknown, a first inference uses only edge features in Φ. This delivers soft estimates of body part positions, which are used to build appearance models of the parts and background (color histograms). Inference is then repeated with Φ using both edges and appearance. This parsing technique simultaneously estimates pose and appearance of parts. For each body part, parsing delivers a posterior marginal distribution over location and orientation (*x*, *y*, *ϕ*) \[[@b6-sensors-12-15376],[@b8-sensors-12-15376]\].
3.. Results
===========
Before the presentation of the results, we discuss the data, methods and parameters of the comparative, and validation measurements.
Data
----
We use the public image sequences of the Chroma Video Segmentation Ground Truth (cVSG) \[[@b17-sensors-12-15376]\], a corpus of video sequences and segmentation masks of people. Chroma based techniques have been used to record Foregrounds and Backgrounds separately, being later combined to achieve final video sequences and accurate segmentation masks almost automatically. Some samples of the sequence we have used for testing are shown in [Figure 4(a)](#f4-sensors-12-15376){ref-type="fig"}. The sequence has a total of 307 frames. This image sequence includes several critical factors that make segmentation difficult: object textural complexity, object structure, uncovered extent, object size, Foreground and Background velocity, shadows, background textural complexity, Background multimodality, and small camera motion.
As a second database, we have also used a set of 30 videos corresponding to the defense of undergraduate thesis at the University of Barcelona to test the methodology in a different environment (UBDataset). Some samples of this dataset are shown in [Figure 4(b)](#f4-sensors-12-15376){ref-type="fig"}.
Moreover, we present the Human Limb dataset, a new dataset composed by 227 images from 25 different people. At each image, 14 different limbs are labeled (see [Figure 4(c)](#f4-sensors-12-15376){ref-type="fig"}), including the "do not care" label between adjacent limbs, as described in [Figure 5](#f5-sensors-12-15376){ref-type="fig"}. Backgrounds are from different real environments with different visual complexity. This dataset is useful for human segmentation, limb detection, and pose recovery purposes \[[@b18-sensors-12-15376]\].
Methods
-------
We test the classical semi-automatic GrabCut algorithm for human segmentation comparing with the proposed ST-GrabCut algorithm. In the case of GrabCut, we set the number of GMM components *k* = 5 for both foreground and background models. Furthermore, the already trained models used for person and face detectors have been taken from the OpenCV 2.1.
We also test the mesh fitting and body pose recovery methodologies on the obtained segmentations. The body model used for the pose recovery was taken directly from the work of \[[@b8-sensors-12-15376]\].
Validation measurements
-----------------------
In order to evaluate the robustness of the methodology for human body segmentation, face and pose fitting, we use the ground truth masks of the images to compute the overlapping factor *O* as follows $$O = \frac{\sum{M_{GC} \cap M_{GT}}}{\sum{M_{GC} \cup M_{GT}}}$$where *M~GC~* and *M~GT~* are the binary masks obtained for spatio-temporal GrabCut segmentation and the ground truth mask, respectively.
3.1.. Spatio-Tempral GrabCut Segmentation
-----------------------------------------
First, we test the proposed ST-GrabCut segmentation on the sequence from the public cVSG corpus. The results for the different experiments are shown in [Table 1](#t1-sensors-12-15376){ref-type="table"}. In order to avoid the manual initialization of classical GrabCut algorithm, for all the experiments, seed initialization is performed applying the commented person HOG detection, face detection, and skin color model. First row of [Table 1](#t1-sensors-12-15376){ref-type="table"} shows the overlapping performance of [Equation (13)](#FD13){ref-type="disp-formula"} applying GrabCut segmentation with *k*-means clustering to design the GMM models. Second row shows the overlapping performance considering the spatial extension of the algorithm introduced by using Mean Shift clustering ([Equation (5)](#FD5){ref-type="disp-formula"}) to design the GMM models. One can see a slight improvement when using the second strategy. This is mainly because Mean Shift clustering takes into account spatial information of pixels in clustering time, which better defines contiguous pixels of image to belong to GMM models of foreground and background. Third performance in [Table 1](#t1-sensors-12-15376){ref-type="table"} shows the overlapping results adding the temporal extension to the spatial one, considering the morphology refinement based on previous segmentation ([Equation (7)](#FD7){ref-type="disp-formula"}). In this case, we obtain near 10% of performance improvement respect the previous result. Finally, last result of [Table 1](#t1-sensors-12-15376){ref-type="table"} shows the full-automatic ST-GrabCut segmentation overlapping performance taking into account spatio-temporal coherence, and the segmentation refinement introduced in [Equation (6)](#FD6){ref-type="disp-formula"}. One can see that it achieves about 25% of performance improvement in relation with the previous best performance. Some segmentation results obtained by the GrabCut algorithm for the cVSG corpus are shown in [Figure 6](#f6-sensors-12-15376){ref-type="fig"}. Note that the ST-GrabCut segmentation is able to robustly segment convex regions. We have also applied the ST-GrabCut segmentation methodology on the image sequences of UBDataset. Some segmentations are shown in [Figure 6](#f6-sensors-12-15376){ref-type="fig"}.
3.2.. Face Fitting
------------------
In order to measure the robustness of the spatio-temporal AAM mesh fitting methodology, we performed the overlapping analysis of meshes in both un-segmented and segmented image sequence of the public cVSG corpus. Overlapping results are shown in [Table 2](#t2-sensors-12-15376){ref-type="table"}. One can see that the mesh fitting works fine in unsegmented images, obtaining a final mean overlapping of 89.60%. In this test, we apply HaarCascade face detection implemented and trained by the Open Source Computer Vision library (OpenCv). The face detection method implemented in OpenCV by Rainer Lienhart is very similar to the one published and patented by Paul Viola and Michael Jones, namely called Viola--Jones face detection method \[[@b19-sensors-12-15376]\]. The classifier is trained with a few hundreds of sample views of a frontal face, that are scaled to the same size (20 × 20), and negative examples of the same size. However, note that combining the temporal information of previous fitting and the ST-GrabCut segmentation, the face mesh fitting considerably improves, obtaining a final of 96.36% of overlapping performance. Some example of face fitting using the AAM meshes for different face poses of the cVSG corpus are shown in [Figure 7](#f7-sensors-12-15376){ref-type="fig"}.
To create three AAM models that represent frontal, right and left views, we have created a training set composed by 1,000 images for each view. The images have been extracted from the public database \[[@b15-sensors-12-15376]\]. To build three models we manually put 21 landmarks over 500 images for each view. The landmarks of the remaining 500 images which covers one view, has been placed by a semi-automatic process, applying AAM with the set learnt and manually correcting. Finally, we align every resulting mesh and we obtain the mean for each model. As the head pose classifier, to classify the spatial mesh configuration in 5 head poses, we have labeled manually the class of the mesh obtained applying the closest AAM model. Every spatial mesh configuration is represented by the 20 most representative eigenvectors. The training set is formed by 5,000 images from the public database \[[@b15-sensors-12-15376]\]. Finally, we have tested the classification of the five face poses on the cVSG corpus, obtaining the percentage of frames of the subject at each pose. The obtained percentages are shown in [Table 3](#t3-sensors-12-15376){ref-type="table"}.
3.3.. Body Limbs Recovery
-------------------------
Finally, we combine the previous segmentation and face fitting with a full body pose recovery \[[@b8-sensors-12-15376]\]. In order to show the benefit of applying previous ST-GrabCut segmentation, we perform the overlapping performance of full pose recovery with and without human segmentation, always within the bounding box obtained from HOG person detection. Results are shown in [Table 4](#t4-sensors-12-15376){ref-type="table"}. One can see that pose recovery considerably increases its performance when reducing the region of search based on ST-GrabCut segmentation. Some examples of pose recovery within the human segmentation regions for cVSG corpus and UBdataset are shown in [Figure 8](#f8-sensors-12-15376){ref-type="fig"}. One can see that in most of the cases body limbs are correctly detected. Only in some situations, occlusions or changes in body appearance can produce a wrong limb fitting.
In [Figure 9](#f9-sensors-12-15376){ref-type="fig"} we show the application of the whole framework to perform temporal tracking, segmentation and full face and pose recovery. The colors correspond to the body limbs. The colors increase in intensity based on the instant of time of its detection. One can see the robust detection and temporal coherence based on the smooth displacement of face and limb detections.
3.4.. Human Limb Data Set
-------------------------
In this last experiment, we test our methodology on the presented Human Limb dataset. From the 14 total limb annotations, we grouped them into six categories: trunk, up-arms, up-legs, low-arms, low-legs, and head, and we tested the full pose recovery framework. In this case, we tested the body limb recovery with and without applying the ST-GrabCut segmentation, and computed three different overlapping measures: (1) %, which corresponds to the overlapping percentage defined in [Equation (13)](#FD13){ref-type="disp-formula"}; (2) wins, which corresponds to the number of Limb regions with higher overlapping comparing both strategies; (3) match, which corresponds to the number of limb recoveries with overlapping superior to 0.6. The results are shown in [Table 5](#t5-sensors-12-15376){ref-type="table"}. One can see that because of the reduced region where the subjects appear, in most cases there is no significant difference applying the limb recovery procedure with or without previous segmentation. Moreover, the segmentation algorithm is not working at maximum performance due to the same reason, since very small background regions are present in the images, and thus the background color model is quite poor. Furthermore, in this dataset we are working with images, not videos, and for this reason we cannot include the temporal extension in our ST-GrabCut algotithm for this experiment. On the other hand, looking at the mean average overlapping in the last column of the table, one can see that ST-GrabCut improves for all overlapping measures the final limb overlapping. In particular, in the case of the Low-legs recovery is when a more clear improvement appears using ST-GrabCut segmentation. The part of the image corresponding to Low-legs is where more background influence exists, and thus the limb recovery has the highest confusion. However, as ST-GrabCut is able to properly segment the concave regions of the Low-legs regions, a significant improvement is obtained when applying the limb recovery methodology. Some results are illustrated on the images of [Figure 10](#f10-sensors-12-15376){ref-type="fig"}, where the images on the bottom correspond to the improvements obtained using the ST-GrabCut algorithm. Finally, [Figure 11](#f11-sensors-12-15376){ref-type="fig"} show examples of the face fitting methodology applied on the human body limb dataset.
4.. Conclusions
===============
In this paper, we presented an evolution of the semi-automatic GrabCut algorithm for dealing with the problem of human segmentation in image sequences. The new full-automatic ST-GrabCut algorithm uses a HOG-based person detector, face detection, and skin color model to initialize GrabCut seeds. Spatial coherence is introduced via Mean Shift clustering, and temporal coherence is considered based on the historical of Gaussian Mixture Models. The segmentation procedure is combined with Shape and Active Appearance models to perform full face and pose recovery.
This general and full-automatic human segmentation, pose recovery, and tracking methodology showed higher performance than classical approaches in public image sequences and a novel Human Limb dataset from uncontrolled environments, which makes it useful for general human face and gesture analysis applications.
One of the limitations of the method is that it depends on the initialization of the ST-GrabCut algorithm, which basically depends on the person and face detectors. Initially, we wait until at least one bounding box is returned by the person detector. This is a critical point, since we will trust the first detection and start segmenting with this hypothesis. In contrast, there is no problem if a further detection is missed, since we initialize the mask with the previous detection (temporal extension). Moreover, due to its sequential application, false seed labeling can accumulate segmentation errors along the video sequence. As the next step, we plan to extend the limb recovery approach so that more complex poses and gestures can be recognized, and feed a gesture recognition system \[[@b20-sensors-12-15376]\] with the temporal aggregation of the recovered poses along the sequence in order to look for motion patterns of the limbs.
As a future work, the algorithm could be extended in order to segment sequences with more than one person present in the images, since our current method only segments one subject in the scene.
This work has been supported in part by projects IMSERSO-Ministerio de Sanidad 2011 Ref. MEDIMINDER, RECERCAIXA 2011 Ref. REMEDI, TIN2009-14404-C02 and CONSOLIDER-INGENIO CSD 2007-00018. The work of Antonio is supported by an FPU fellowship from the Spanish government.
{#f1-sensors-12-15376}
{#f2-sensors-12-15376}
{#f3-sensors-12-15376}
{#f4-sensors-12-15376}
{#f5-sensors-12-15376}
{#f6-sensors-12-15376}
{#f7-sensors-12-15376}
{#f8-sensors-12-15376}
{#f9-sensors-12-15376}
{#f10-sensors-12-15376}
{#f11-sensors-12-15376}
######
GrabCut and ST-GrabCut Segmentation results on cVSG corpus.
Approach Mean overlapping
-------------------- ------------------
GrabCut 0.5356
Spatial extension 0.5424
Temporal extension 0.6229
ST-GrabCut 0.8747
######
AAM mesh fitting on original images and segmented images of the cVSG corpus.
Approach Mean overlapping
------------------------------------ ------------------
Mesh fitting without segmentation 0.8960
ST-Grabcut & Temporal mesh fitting 0.9636
######
Face pose percentages on the cVSG corpus.
Face view System classification Real classification
----------------- ----------------------- ---------------------
Left view 0.1300 0.1211
Near Left view 0.1470 0.1347
Frontal view 0.2940 0.3037
Near Right view 0.1650 0.1813
Right view 0.2340 0.2590
######
Overlapping of body limbs based on ground truth masks.
Approach Mean overlapping
------------------------------------ ------------------
Limb recovery without segmentation 0.7919
ST-Grabcut & Limb recovery 0.8760
######
Overlapping percentages between body parts (intersection over union), wins (comparing the highest overlapping with and without segmentation), and matching (considering only overlapping greater than 0.6).
**Trunk** **Up-arms** **Up-legs** **Low-arms** **Low-legs** **Head** **Mean**
---------------------------------------------------------------- --------------------- ----------- ------------- ------------- -------------- -------------- ------------ ----------
\% **No segmentation** 0.58 0.53 0.59 0.50 0.48 0.67 0.56
**STGrabCut**[\*](#tfn1-sensors-12-15376){ref-type="table-fn"} 0.58 0.53 0.58 0.50 0.56 0.67 **0.57**
Wins **No segmentation** 106 104 108 109 68 120 102.5
**STGrabCut**[\*](#tfn1-sensors-12-15376){ref-type="table-fn"} 121 123 119 118 159 107 **124.5**
Match **No segmentation** 133 127 130 121 108 155 129
**STGrabCut[\*](#tfn1-sensors-12-15376){ref-type="table-fn"}** 125 125 128 117 126 157 **129.66**
STGrabCut was used without taking into account temporal information.
| {
"pile_set_name": "PubMed Central"
} |
Approximately 20% of women experience an episode of major depression, a rate that is twice that of men.^[@ref1]^ The period of greatest vulnerability for women appears to be the childbearing years, with the initial onset of depression most likely to occur between the ages of 25 and 44.^[@ref2]^ Several forms of depression are unique to women because of their apparent association with changes in reproductive hormones: premenstrual dysphorias, including premenstrual syndromes (PMSs) and premenstrual dysphoric disorder (PMDD), postpartum depression (PPD), and depression in the perimenopausal period. The link among these depressive disorders appears to be a sensitivity to normal shifts in gonadal hormones, which affect, neuroregulatory systems that play a role in affective disorders.^[@ref3],[@ref4]^ Such shifts occur during the menstrual cycle, in pregnancy and postpartum, and with ovarian aging in the years leading to the menopause.
Historically, depression has been underrecognized and undertreated. Until recently, diagnostic criteria were imprecise, clinical trials of purported treatments for menstrually related depressions were lacking or poorly done, and treatment options were generally unsupported by scientific data. Over the last two decades, considerable scientific research has focused on the depressions unique to women. This review examines the evaluation and treatment of depression that occurs premenstrually, postpartum, and in the perimenopause based on the current clinical literature.
Premenstrual dysphorias
=======================
Of the depressive disorders that affect only women, PMSs are the most extensively studied. Severe PMS is a chronic mood disorder that continues for many years in reproductive-age women.^[@ref5]^ The etiology remains unconfirmed. Moderate-to-severe forms of the syndrome result in diminished functioning and impaired relationships that cannot be dismissed as trivial. The *Diagnostic and Statistical Manual for Mental Disorders*, 4th cd *(DSM-IV)* provides specific diagnostic criteria for severe dysphoric PMS termed PMDD.^[@ref6]^
Prevalence
----------
Survey studies indicate that up to 40% of menstruating experience some difficulty with premenstrual symptoms.^[@ref7],[@ref8]^ When premenstrual distress is dominated by emotional symptoms such as irritability, nervousness, tension, and depressed mood, it is a powerful predictor of treatment-seeking behavior. In a recent community-based study, 22% of menstruating women rated moderate-to-severe premenstrual distress on an analog measure of distress; this subjective distress was highly correlated with each of the impairment variables, occupation, leisure, partner, and friends.^[@ref9]^ Other studies show that approximately 3% to 10% of reproductive-age women met the specific criteria for PMDD.^[@ref10],[@ref11]^
Depression and PMS/PMDD
-----------------------
By definition, PMDD is a severe and dysphoric form of PMS. Symptoms of irritability, emotional hypersensitivity, increased anxiety and food cravings, sleep difficulties, and decreased concentration characterize PMDD as well as depression, particularly atypical depression. A lifetime history of depression ranges from about 20% to 76% in samples of women diagnosed for PMS or PMDD,^[@ref12],[@ref13]^ with the higher rates substantially greater than the lifetime prevalence of about 20% for major depression in the female population of this age-group.^[@ref1]^ A family history of depressive illness is common in women with PMS/PMDD.^[@ref5]^ Women who seek treatment for premenstrual symptoms frequently have other emotional disorders, most commonly depressive disoders, substance abuse, or anxiety. Conversely, women who have mood disorders frequently experience worsening of symptoms premenstrually.^[@ref14]^
In spite of these observable similarities, increasing evidence suggests that PMS/PMDD is not a simple variant of depression, but a distinct disorder. While depressive symptoms characterize a substantial number of women with severe perimenstrual distress, there is also a sizeable group of women who do not suffer from either depression or anxiety symptoms, but experience severe or moderate perimenstrual distress with symptoms of irritability, nervousness, and tension.^[@ref9]^ On the basis of epidemiologic findings, the researchers posited that the features of irritability and tension irrespective of the presence of depressive symptoms may form the core symptoms of the disorder, a premise also advanced by Eriksson, on the basis of results of antidepressant treatment studies.^[@ref15],[@ref16]^ Data from epidemiologically based twin studies indicated that the degree to which premenstrual symptoms shared generic and environmental risk factors with major depression was modest, a possible indication that there is no close etiologic relationship between the two disorders.^[@ref17]^ The evidence that severe PMS/PMDD responds much more robustly to serotonergic antidepressants than to other antidepressants that are clearly effective for depressive disorders suggests differences in underlying mechanisms. There is also empirical evidence that a good response of PMS/PMDD patients to serotonergic antidepressants is not. explained by depressive symptoms or a history of depression.^[@ref18],[@ref19]^
The rapid response of several days rather than several weeks of PMS patients to selective serotonin reuptake inhibitors (SSRIs), the efficacy of SSRIs at low doses, and the efficacy of other serotonin agonists, including fenfluramine^[@ref20]^ and buspirone,^[@ref21]^ which are not effective for depression, all suggest that the underlying mechanisms of severe PMS/PMDD differ from other depressive disorders.
Evaluation
----------
The diagnosis of premenstrual dysphoria, PMS, or PMDD is made primarily on the basis of the symptom pattern and the exclusion of other possible diagnoses. The essential elements are confirmation of the expected relationship of the symptoms to the menstrual cycle, ie, that the symptoms occur premenstrually and remit with menses; that the symptoms are distressing and warrant treatment, ie, that the symptoms impair usual functioning; and that the symptoms are not due to another physical or mental disorder. These elements are incorporated in the diagnostic criteria for PM\'DD, which, in addition, requires at least 5 of 11 specified symptoms that are severe premenstrually and remit with menses.^[@ref6]^ It is underscored that the symptom *pattern,* rather than specific symptoms or the number of symptoms, defines the disorder.
For a PMS/PMDD diagnosis, it is essential to confirm the symptom pattern for two to three menstrual cycles with prospective daily symptom ratings maintained by the patient, especially if the symptoms are mild. Less than half the women who report PM.S provide daily symptom reports that corroborate their retrospective reports,^[@ref22]^ which are less reliable when symptoms are not consistent and severe.^[@ref23]^
The major consideration after identifying the symptom pattern is whether the condition is purely PMS/PMDD or a premenstrual exacerbation of other psychiatric problems or medical conditions. Premenstrual exacerbation of symptoms may occur in other conditions such as asthma,^[@ref24],[@ref25]^ migraine,^[@ref26]^ seizure disorders,^[@ref27]^ alcohol intake,^[@ref28],[@ref29]^ depression,^[@ref14]^ and schizophrenia.^[@ref30]^ There is no laboratory test that identifies PMS/PMDD, and such tests are useful only if there are other questions that might, be answered. PMS and PMDD are based on regular menstrual cycles within the normal range of 22 to 35 days, and patients with irregular cycling should be examined for other conditions. Standard hematology and blood chemistry profiles are conducted to confirm general good health. A thorough examination includes a review of current and past psychiatric status, particularly mood and anxiety disorders that are commonly associated with PMS/PMDD. A gynecologic examination is important to rule out problems such as endometriosis, which might account for the symptoms.
Serotonergic antidepressants
----------------------------
The serotonergic antidepressants, particularly the SSRIs, appear to be the treatment of choice for severe PMS and PMDD at this time. Modulating serotonergic function is consistent with the dominant theoretical view that the normal gonadal steroid fluctuations of the menstrual cycle trigger an abnormal serotonergic response in vulnerable women. Indications of abnormalities in markers of serotonergic transmission in women with severe PMS include evidence of a lowered platelet imipramine binding (a peripheral marker of serotonin \[5-hydroxytryptamine, 5-HT\] function) in the luteal phase,^[@ref31]^ decreased platelet 5-HT content and 5-HT uptake during the luteal phase,^[@ref32],[@ref33]^ and significantly decreased whole blood 5-HT levels premenstrually.^[@ref34]^ PMS patients showed a lower 5-HT response to tryptophan (a 5-HT precursor) during the luteal phase compared with the follicular or midluteal phases.^[@ref35]^ Challenge tests depleting tryptophan provoked PMS symptoms,^[@ref36]^ while tryptophan supplementation relieved PMS symptoms in open-label treatment.^[@ref37]^. Following administration of the serotonin -releasing fenfluramine, the women with PMDD had a significantly blunted prolactin response compared with the normal controls.^[@ref38]^ Fenfluramine administered to PMS subjects improved depressed mood and food cravings.^[@ref20]^
Administration of the serotonin agonist m-chlorophenylpiperazine (*m*-CPP) showed improvement of PMS symptoms in the luteal phase and a blunted response to Cortisol and adrenocorticotropic hormone (ACTH) in both the follicular and luteal phases of the PMS subjects compared with normal controls.^[@ref39]^ Although existing information does not indicate a causal relationship between serotonin and PMS, the data suggest involvement of the serotonergic system in this disorder.
A meta-analysis of randomized controlled trials of SSRIs in treatment of PMS/PMDD concluded that these drugs were an effective first-line therapy, with the overall standard mean difference in favor of SSRIs equivalent to an odds ratio of 6.91.^[@ref40]^ Efficacy has been clearly shown for fluoxetine,^[@ref41]-[@ref46]^ sertraline,^[@ref18],[@ref47]-[@ref50]^ paroxetine,^[@ref51]^ citalopram,^[@ref52]^ venlafaxine,^[@ref53]^ and clomipramine.^[@ref54],[@ref55]^ Open-label studies showed that nefazodone^[@ref56]^ and fluvoxamine^[@ref57]^ also had response frequencies in PMS treatment similar to those in the placebo-controlled studies of SSRIs. The only medication with Food and Drug Administration (FDA) approval for treatment of PMDD are fluoxetine and sertraline.
In all reports of SSRI and other serotonergic antidepressant treatments for PMS/PMDD, the effective doses have remained at the low end of the dose range. An adequate trial of a serotonergic antidepressant is at least two menstrual cycles with a third cycle if there is partial response. If a patient has an insufficient response or continuing side effects with an initial SSRI, another SSRI can be tried.^[@ref58]^ Side effects are common with the onset of treatment, but are usually transient and disappear during the first treatment cycle. The most common side effects include headache, nausea, insomnia, fatigue or lethargy, diarrhea, decreased concentration, and dizziness. Decreased libido is also a common side effect of SSRI treatment, although the few published reports of PMS patients identified a relatively low incidence of decreased sexual interest or reduced orgasm of 10% to 12%.^[@ref18],[@ref53]^ In contrast, the frequency of less sexual arousal reported by women in depression studies ranged from 32% (sertraline) to 40% (paroxetine).^[@ref59]^ Whether there is a true difference between PMS and major depression patients with respect to this side effect, is not known, but the PMS/PMDD reports clearly are from acute treatment trials, do not represent a systematic assessment of sexual function, and may not represent, experience with longer maintenance treatment.
Luteal phase closing
--------------------
The use of medication only in the symptomatic luteal phase of the menstrual cycle is of particular interest in PMS/PMDD because of the cyclic pattern of the symptoms, which includes a clear symptom-free interval each month, and the rapid response of these patients to SSRIs. A number of preliminary studies examined luteal phase dosing regimens of SSRIs and consistently reported efficacy.^[@ref48]-[@ref50],[@ref52],[@ref60],[@ref61]^ Two large multicenter trials reported efficacy of fluoxetine^[@ref62]^ and sertraline^[@ref13]^ administered for the last 2 weeks of the menstrual cycle. As with daily dosing, the mean doses remained low and the medications were well tolerated. No studies to date have reported discontinuation symptoms with luteal phase dosing. Although several reports suggested superiority of luteal phase dosing over daily dosing, none were designed or sufficiently powered to answer this question. Overall, the studies indicate that luteal phase dosing is effective for clearly diagnosed PMS/PMDD; previous daily treatment with an SSRI is not required; response is usually at the low end of the dose range; side effects are similar to those seen in continuous dosing; and discontinuation symptoms do not appear to be a problem in the luteal phase dosing regimens.
Other antidepressants
---------------------
The antidepressant response in PMS/PMDD appears to be associated with potent, serotonergic activity and is not a general antidepressant effect. Other antidepressants, which are clearly effective for major depression, such as desipramine (a tricyclic noradrenergic antidepressant),^[@ref18]^ buproprion (with weak inhibition of both serotonin and norepinephrine reuptake),^[@ref45]^ and maprotoline (a selective noradrenaline reuptake inhibitor),^[@ref51]^ were no more effective than the placebo in PMS treatment.
Long-term treatment of PMS/PMDD
-------------------------------
All published studies of treatment efficacy for PMS/PMDD are based on acute treatment of 2 to 3 months\' duration. Anecdotal reports and several small pilot investigations^[@ref63]-[@ref66]^ suggest that PMS symptoms return within several months if medication is stopped. It also appears that untreated symptoms do not resolve spontaneously, as may occur in depression, but continue for many years, based on information from clinical trials, which report that the duration of the disorder is in the range of 8 to 10 years prior to treatment. Well-designed, long-term maintenance studies have not been conducted for this disorder, but these observations suggest that long-term maintenance treatment may be appropriate for patients with severe PMS/PMDD, particularly if they experience a rapid return of symptoms after responding to medication.
Insufficient response to serotonergic antidepressants
-----------------------------------------------------
The overall response of PMS/PMDD patients to SSRIs is approximately 60% in controlled trials, but up to 40% may not have sufficient response. No strong predictors of response have been identified.^[@ref19]^ An expert consensus group recommended the common clinical practice of shifting to a second SSRI when the patient has an insufficient response or is intolerant to the initial SSRI.^[@ref58]^ Augmenting an SSRI with other medications has not been tested in PMS/PMDD studies. Switching to another class of medication that has shown efficacy for PMS/PMDD, such as anxiolytics or gonadotropin-releasing hormone (GnRH) agonists, is suggested, but there are no data that indicate whether nonresponders to an SSRI will respond to another class of medication. Nonresponse may also be due to other comorbid disorders. A thorough review of the diagnosis and adjustments of the premenstrual doses of medication for the primary disorder should be considered before pursuing other treatments for PMS.
Other treatments
----------------
### *Hormonal treatment*
Hormonal treatments for PMS/PMDD are not supported by consistent scientific information in spite of evidence of hormonal involvement in the disorder.^[@ref67]^ GnRH agonists, such as depot leuprolide^[@ref68],[@ref69]^ and intranasal buserelin,^[@ref70],[@ref71]^ effectively reduce PMS symptoms, but arc of limited use because of the risks associated with low estrogen levels, particularly osteoporosis, and these medications are viewed as appropriate only as a diagnostic tool or for patients who do not respond to other treatments. Results of preliminary investigations of add-back therapy using low-dose estrogen and progesterone in conjunction with a GnRH agonist are inconsistent and do not yet definitively indicate that this is a safe as well as effective approach for long-term treatment.^[@ref72]-[@ref74]^ Limited data indicate that tibolone (a selective estrogen enzyme modulator) administered with a GnRH agonist in PMS treatment protects against the bone loss observed with GnRH agonists and docs not reduce the therapeutic effect of the agonist.^[@ref75]^
There are few randomized, placebo-controlled studies of oral contraceptives (OCs) as a treatment for severe PMS or PMDD, and no consistent scientific evidence of their efficacy for the disorder.^[@ref76],[@ref77]^ A triphasic OC was more effective than placebo only for physical symptoms of breast pain and bloating.^[@ref76]^ A recent trial of an OC containing a new progestin, an analog of spironolactone with antimineralocorticoid and antiandrogenic activity, showed a consistent reduction of both physical and behavioral PMS symptoms including dysphoric mood, but additional studies with sufficient statistical power are needed.^[@ref78]^
From a clinical perspective, OCs are widely viewed as both improving and worsening PMS symptoms. Combination OCs have both estrogenic and progestational effects that vary considerably among the more than 40 compounds available in the USA. Relative absorption of the hormones, peripheral conversion, the degree of follicular development in the placebo interval, individual susceptibility to monophasic or triphasic formulations, and side effects have large variations among women and are not well understood in relation to PMS. Moreover, OCs can have side effects of water retention, bloating, appetite changes, and depressed mood, which are also PMS symptoms. Some studies showed that OC users had fewer PMS symptoms than nonusers overall,^[@ref79]^ but other investigations found few symptom differences between the two groups and no difference with respect to mood changes.^[@ref80]^ In sum, there is little empirical support or guidance for OCs as a treatment for PMS/PMDD,^[@ref81]^ although it is reasonable to try OCs, particularly when contraception is also required. If mood symptoms are predominant and persist, a serotonergic antidepressant is considered the first-line therapy.
Estrogen therapy with dose regimens sufficient to suppress ovulation significantly decreased the dysphoric mood and physical symptoms of PMS.^[@ref82]-[@ref84]^ However, estrogen must be cycled with progesterone to reduce the risk of uterine cancer, and the extent to which exogenous progesterone results in return of PMS symptoms remains unclear. Progesterone treatment, of PMS was advocated for many years, but numerous studies, including three large randomized controlled trials, failed to show improvement significantly greater than placebo for the mood and behavioral symptoms of PMS.^[@ref85]-[@ref87]^
### *Anxiolytics*
Alprazolam and buspirone showed modest efficacy for PMS in some studies,^[@ref87]-[@ref91]^ but not others.^[@ref92],[@ref93]^ The well-known risk of dependence with alprazolam must be considered, and this medication should be tried only when the patient, has symptoms clearly limited to the luteal phase (so that the medication is stopped for at least 2 weeks in each cycle) and no history of substance abuse. These medications offer an alternative to antidepressants, but the extent to which patients who fail to respond to antidepressants respond to these anxiolytics is not known.
### *Nonpharmacologic approaches*
Numerous nonpharmacologic approaches have been advocated for PMS, but few are supported by solid empirical evidence.^[@ref94]^ A large study of calcium supplementation (600 mg twice daily) for PMS reduced premenstrual depression, fatigue, edema, and pain significantly more than the placebo. However, the severity of the dysphoric mood symptoms was not indicated, and further information is required to determine the efficacy of this treatment for premenstrual dysphorias.^[@ref95]^
A meta-analysis showed that vitamin B~6~ was about twice as likely as placebo to improve PMS symptoms overall, with an odds ratio for improvement, in depressive symptoms of 1.69, but the researchers concluded that the quality of the studies was too poor to have confidence in the results.^[@ref96]^ There was no significant dose response, indicating that the amount of vitamin B~6~ did not affect improvement, and reports of peripheral neuropathies with doses exceeding 200 mg preclude the use of megadoses.^[@ref96]^
Several reports of cognitive therapies show improvement of premenstrual symptoms.^[@ref94]^ Other complementary and alternative therapies showed no convincing evidence of efficacy for PMS in a review of randomized controlled trials (dietary supplements, 13 trials; herbal medicines, 7 trials; biofeedback, 2 trials; homeopathy, relaxation, massage, reflexology, and chiropractic, 1 trial each).^[@ref97]^
Emerging from a long history with little understanding and many treatments of doubtful effect, clinically significant PMS is now recognized as a chronic disorder that impairs functioning and personal relationships for a sizeable number of women. Serotonergic antidepressants are the first-line treatment at this time. Using these medications only in the symptomatic luteal phase is effective for women without, other comorbid disorders. Hormonal treatments for PMS are not supported by consistent scientific data on efficacy and safety, in spite of evidence of hormonal involvement in the disorder.
Perimenopause
=============
When women are in their forties, anovulation becomes more frequent and menstrual cycles are altered in length and frequency.^[@ref98]^ This reproductive transition extends for 2 to 8 years before the menopause, which is defined as the point where there was no menstrual bleeding for 12 months.^[@ref99]^ The mean duration of the perimenopausal transition is about 5 years; the onset occurs between the ages of 39 and 51 years for 95% of women.^[@ref100]^ Although cycle irregularity is the traditional clinical marker for perimenopause, it is increasingly clear that hormonal changes and distressing menopausal symptoms such as hot flashes can occur before observed cycle changes, which are an unreliable indicator of the perimenopause.^[@ref101]^
Prevalence of depression
------------------------
Among the most controversial issues in the transition to menopause is its association with depression. Whether there is an association, what the causes are, and how such depression should be treated are questions that continue to have no definitive answers. Epidemiologic studies based on self-report of menopausal status and dysphoric mood have consistently shown that most postmenopausal respondents do not report high rates of depressive symptoms and that reported depressive symptoms were not related to menopause per se but to other health problems,^[@ref102]-[@ref104]^ and suggest that menopause does not cause depressive illness.^[@ref105]^ The National Comorbidity Study reported 30-day estimates of major depression of 5.0% and lifetime estimates of 21.8% for women aged 45 to 54 years, slightly lower than the estimates for the 35- to 44-ycar age-group, which had the highest, rates of depression.^[@ref106]^ However, the rates of recurrent depression were highest, in the 45- to 54-year age-group when compared with older women.^[@ref107]^ The Massachusetts Health Study^[@ref102]^ found that in the cohort of women aged 45 to 55 years at baseline prior depression was the variable most predictive of subsequent depression, based on self-report of depressive symptoms using the Center for Epidemiologic Studies-Depression (CES-D) scale.^[@ref108]^ The chance of a *recurrence* of depression is high (50% after a first episode, 70% after two episodes, and 90% after three episodes)^[@ref109]^ and may coincide with the perimenopausal years.
Most cross-sectional studies suggest that women in the 40- to 55-year age-group are more likely to report depressive symptoms than pre- and postmenopausal women. In the current Study of Women\'s Health Across the Nation (SWAN), 40.5% of the sample of women aged 40 to 55 years reported that they had felt depressed within the past 2 weeks.^[@ref110]^ Dennerstein et al defined perimenopausal status from bleeding patterns and reported that 38% of late perimenopausal women reported depressive symptoms in the previous 2 weeks compared to 26% of the premenopausal and 28% of the postmenopausal women.^[@ref111]^ Bosworth et al reported that 28.9% of women aged 45 to 54 years had a high level of depressive symptoms based on an abbreviated CES-D scale, but found no association between the depressive symptoms and menopausal stage defined by the women\'s perceptions.^[@ref112]^ Soares et al identified 28.7% of women aged 40 to 58 years attending a menopause clinic as meeting *DSM-IV* criteria for depressive disorders.^[@ref113]^ While all these studies suggest an increased prevalence of depressive symptoms and possibly depressive illness in the transition to menopause, whether these depressive symptoms are associated with hormonal fluctuations or changes that characterize the transition to menopause remains unclear.
Estrogen as an antidepressant
-----------------------------
Estrogen treatment is widely believed to improve depressive symptoms in menopausal women,^[@ref114]-[@ref118]^ but study results are inconclusive because of large variations in study design and measures, hormonal status and diagnosis of the subjects, the estrogen compound, dose, and duration of use, and failure to find an effect greater than the placebo response.^[@ref119]-[@ref122]^
Burt et al^[@ref123]^ identified six studies that included perimenopausal women for estrogen treatment of depressive symptoms. Only two studies were placebo-controlled; only one of these showed significant improvement with estradiol compared with placebo after 4 months of treatment, but the treatment advantage over placebo was not sustained after 12 months of treatment.^[@ref124]^ In an uncontrolled study of women judged to be depressed or not depressed on the basis of the Beck Depression Inventory, only the group that was not depressed responded to standard replacement doses (0.3-0.625 mg/day) of conju-gated estrogen.^[@ref125]^ Pharmacologic doses of estradiol (5-25 mg/day) showed improvement greater than placebo in women diagnosed with depressive disorders^[@ref126]^ and in a study of postmenopausal women with scores signifying mental distress (1-4 mg/day).^[@ref127]^ Conclusions cannot be drawn from the conflicting results of these studies, which are limited by designs that do not clearly identify essential variables, such as menopausal status and diagnosis of depression, and also lack comparability in the form and dose of estrogen treatment.
Two recent well-designed studies found 17β-estradiol to be effective for depression in perimenopausal women. Both studies clearly diagnosed depression, endocrinologically defined perimenopausal status and administered transdermal 17β-estradiol (the major circulating estrogen in women) using randomized, placebo-controlled, double-blind designs and showed that estrogen may be an effective treatment for major or minor depression in perimenopausal women. Soares et al^[@ref128]^ reported remission of depression in 68% of the estradiol group compared with 20% of the placebo group after 12 weeks. Schmidt et al^[@ref129]^ showed a full or partial response for 80% of the estradiol group compared with 22% of the placebo group after 6 weeks of estradiol. A progestin added at 7 weeks did not negate improvement with one exception of worsening of early morning waking. Both studies indicated that the effect of estradiol on mood was independent of hot flashes - an important finding that suggests that the improvement of depressed mood with estrogen treatment, was not simply a result of improving hot flashes. Both studies identified a rapid onset of antidepressant response in perimenopausal depression. However, the brief duration of the progestin use may be inadequate to determine whether long-term progesterone use reduces the beneficial estradiol effect on mood.^[@ref116]^ Further studies are needed to confirm these positive findings and determine long-term effects of estradiol treatment.
Estrogen administration throughout the cycle may be more effective than the standard OC regimen for decreasing depressive symptoms in perimenopausal women. Blümel et al compared a standard OC (20 μg ethinyl estradiol and 150 mg desogestrel for 21 days followed by placebo for 7 days) with the same OC followed by only 2 days of placebo and 5 days of 10 μg ethinyl estradiol in a randomized trial.^[@ref130]^ Depressive, vasomotor, and somatic symptoms and sexual function improved significantly more in the group with estrogen continued throughout the cycle. The results were interpreted by the researchers to indicate that increasing the days with estrogen in women using OCs restimulated estrogen receptors and improved cerebral neurochemistry.
Antidepressant medications
--------------------------
The SSRIs (fluoxetine, paroxetine, and sertraline) and other serotonergic antidepressants such as venlafaxine, nefazodone, and clomipramine are currently viewed as the first-line treatment for most depressive disorders because of extensive data supporting their efficacy, the minimal need for dose titration, and generally favorable side-effect profiles.^[@ref58]^ However, there is growing evidence of gender differences and effects of menstrual status in treatment response and tolerability to SSRIs. Women with chronic major depression were more likely to respond to sertraline compared with men, who were more likely to respond to the tricyclic antidepressant, imipramine.^[@ref131]^ Menstrual status affected this response, with premenopausal women significantly more likely to respond to the serotonergic than a tricyclic antidepressant, while the postmenopausal women responded similarly to both medications. The postmenopausal women who were taking imipramine also had significantly lower attrition rates than premenopausal women. Similar results were observed in a comparison of fluoxetine with maprotiline.^[@ref132]^
Other observations of postmenopausal women identified an interaction between estrogen status and antidepressant therapies: women who were using estrogen replacement therapy and received fluoxetine had a greater antidepressant response than the women who received only fluoxetine^[@ref133]^; similarly, older depressed women who received both estrogen and sertraline responded better than those who received only sertraline.^[@ref134]^ Another trial of fluoxetine in depressed menopausal women failed to find any effect of estrogen therapy, but the subjects were younger (≥45 years of age) and menopausal status was determined only by age, limiting interpretation and comparisons of the results.^[@ref135]^ Overall, these preliminary findings suggest that menstrual status is an important consideration in selecting an antidepressant for women, and that the estrogen status (which differs in pre-, peri-, and postmenopausal women) may be associated with the response to antidepressants.
Management of depression in perimenopausal women
------------------------------------------------
Current consensus guidelines for treatment of depression in perimenopausal women recommend an antidepressant for severe depression.^[@ref58]^ Data indicate that an SSRI may be preferred to a tricyclic antidepressant for women who are not postmenopausal. For women with previous episodes of depression, the general guideline is to prescribe the antidepressant used in the previous episode if the patient had a satisfactory response.
Transdermal estradiol (0.05-0.10 mg/day) may be of benefit for perimenopausal women with major or minor depression, based on preliminary but consistent findings of two new studies.^[@ref128],[@ref129]^
Minor mood symptoms associated with the perimenopause are also improved with estrogen therapy.^[@ref116]^ A progestin must also be prescribed for women with a uterus and may reduce the improvement of depressed mood in some women. Estrogen therapy is generally contraindicated for women with breast cancer, any potentially estrogen-dependent malignancy, active liver disease, and active thrombosis. Speroff et al indicate close surveillance for women with seizure disorders, familial hyperlipidemias, and migraine headaches.^[@ref136]^ Other considerations include a history of breast disease, history of stroke, myocardial infarction or thrombosis, and active gall bladder disease or gallstones.
The estradiol dose of hormone replacement therapy (HRT) does not suppress ovulation or provide contraception for perimenopausal women, who continue to be at risk of pregnancy until the menopause.^[@ref137]^ For contraceptive protection and for estrogen-related symptoms such as hot flashes, an OC with estrogen rather than HRT may be preferred for perimenopausal women. However, there is no evidence at this time that OCs effectively treat major or minor depression in perimenopausal women. Recent studies suggested that reducing the placebo interval of OCs and extending estradiol through the cycle improved depressive symptoms, but these findings do not extend to women diagnosed with depressive illness. The association of cardiovascular events with estrogen is dose-related and the current low-dose OCs (\<50 μg ethinyl estradiol) can be used by perimenopausal women with normal blood pressure.^[@ref137]^ Smokers over age 35 should not use OCs.
A frequently asked question is whether estrogen and antidepressant therapies can be combined. The strongest rationale for using both medications is the known benefits of each. In addition to its psychotropic effect, estradiol reduces menopause-related hot flashes, improves vaginal dryness, is protective against bone loss, which may result in osteoporosis, and may be protective against cardiovascular disease.^[@ref137]^ Recent evidence indicates that estrogen can produce an antidepressant response in perimenopausal women with major or minor depression.^[@ref128],[@ref129]^
Antidepressants are clearly effective for dysphoric mood. Although preliminary findings have suggested that the combined use of HRT and serotonergic antidepressants enhanced the antidepressant response, these observations pertained to elderly depressed women and cannot be generalized to perimenopausal women. There are no studies of combined estrogen and antidepressant therapies in women identified as perimenopausal.
To assess perimenopausal depression, the pattern and severity of the depressive symptoms should be determined. Perimenopausal status is suggested by the presence of vasomotor symptoms such as hot flashes or irregular menstrual cycles, although it is entirely possible that a perimenopausal woman has neither, particularly in the early stages of the menopausal transition. Careful physical examination and medical history should be obtained to determine other coexisting conditions, previous experience of any depressive disorder, the onset of the depressive symptoms in conjunction with menstrual cycle changes, and to identify risk factors and contraindications for estrogen therapy. A follicle-stimulating hormone (FSH) level greater than 20 IU/L is a hormonal marker of the perimenopause, but single measures are considered unreliable because of sporadic follicular activity and competence. The decline in mean estrogen levels occurs primarily in the year before menopause,^[@ref138]^ although there is considerable individual variability. Burger et al recently reported that mean estrogen levels started to decrease about 2 years before the final menstrual period, but fell substantially only in the year before the final menses with the most rapid decrease around the time of the final menstrual period.^[@ref139]^ Recognition that the perimenopause can extend over a number of years and that the hormonal shifts that occur in this transition may be associated with depressive symptoms is important for patient care. However, scientific data to guide treatment of depression in the perimenopause are limited and inconclusive. Estrogen therapy may be helpful for major or minor depression as well as for depressive symptoms linked to the menopause. OCs containing estrogen are another possible option for perimenopausal women, who still require contraceptive protection, but there is no consistent evidence at this time of their antidepressant effects. The hormone doses of OCs are considerably higher than those of HRT, but only the HRT dose levels of estrogen have shown antidepressant effects in perimenopausal women at this time. Serotonergic antidepressants are clearly effective for depressive illness, and are now widely considered the first-line treatment for moderate-to-severe depressive illness or a repeated episode of depression in perimenopausal women.
Postpartum depression
=====================
The months following childbirth have been recognized throughout history as a period of increased risk of depression for vulnerable women, although diagnostic criteria have emerged only in recent decades, and there are few well-designed controlled studies of treatment efficacy. Two defining characteristics of PPD are its occurrence at a time of large hormonal shifts and its high likelihood of recurrence with subsequent pregnancies.
Prevalence
----------
PPD is a nonpsychotic depression that meets the diagnostic criteria for major depression and occurs within several months of delivery as defined in the *DSM-IV*.^[@ref6]^ The point-prevalence of PPD within 6 to 9 weeks of delivery is about 12%. ^[@ref140],[@ref141]^ Postpartum psychosis is uncommon, but potentially lethal to the woman or the infant, and occurs in 1 to 2 per 1000 women following childbirth, with onset usually within 2 to 4 weeks of delivery.^[@ref142]^ Mood lability, or "baby blues" within the initial days following delivery is very common, with estimates up to 80% for brief periods of symptoms such as tearfulness, fatigue, and insomnia that occur within the first 2 weeks of childbirth, peaking at about 5 days postpartum.^[@ref143]^
It was long believed that women were at decreased risk of depressive disorders during pregnancy, and few studies examined associations between depression during pregnancy and the postpartum period. However, studies show that depression can increase steadily from the second trimester of pregnancy to 9 weeks postpartum,^[@ref140],[@ref143]^ with little difference in prevalence (9% in the second trimester; 12% postpartum) or even greater prevalence during pregnancy than postpartum.^[@ref144],[@ref145]^ Data also indicate that the depressive symptoms differ when compared during pregnancy and in early and later postpartum periods, corroborating both the occurrence of depressive symptoms during pregnancy and identifying differing vulnerabilities to depression throughout pregnancy and the postpartum period.^[@ref146]^
PPD is strongly associated with previous depressions. A recent review indicated that the increased risk was 25% for women with a history of depression, 50% for women with previous PPD, and 75% for women with depression during the current pregnancy.^[@ref58]^ Twenty-nine percent of women diagnosed with late luteal phase disorder and 43% of women diagnosed with PMS had experienced PPD, suggesting possible association with premenstrual syndromes.^[@ref147],[@ref148]^ Other risk factors for PPD include poor social support and chronic stressors.^[@ref149]^
Treatment of postpartum depression
----------------------------------
Reported treatments for PPD include antidepressants, hormones, and psychotherapy, but there is a paucity of well-designed controlled studies, samples are small and there are no definitive conclusions.
### *Antidepressants*
Sertonergic antidepressants with reported efficacy for PPD include fluoxetine in double -blind study,^[@ref150]^ and sertraline, venlafaxine, and fluvoxamine in open studies.^[@ref151]-[@ref153]^ Results in the sertraline study suggested that response is swift (by 2 weeks) and at low doses, but these findings were not supported in another retrospective record review.^[@ref154]^ The many antidepressants available are clearly effective for depression but are generally unstudied for PPD. Other considerations in the selection of an antidepressant are the patient\'s tolerability of side effects and the response to a previously prescribed antidepressant. For women with previous episodes of depression, the general guideline is to prescribe the antidepressant used in the previous episode if the patient had a satisfactory response.
A major concern about drug therapy for breast-feeding mothers is the effect of medication on the infant.^[@ref155]^ In small studies, amitriptyline, nortriptyline, desipramine, clomipramine, imipramine, sertraline, fluvoxamine, and paroxetine were not detected in quantifiable amounts in infant, plasma and all infants were thriving.^[@ref156]-[@ref161]^ The results are encouraging, but cannot be generalized to all infants exposed to these medications.
Because of the high risk of repeated PPD, the question of prophylactic treatment is important but unanswered. In the only controlled study of an antidepressant administered as a prophylactic, the tricyclic antidepressant, nortriptyline, initiated immediately postpartum in nondepressed women at risk of a subsequent PPD, was not better than placebo; 25% of the women in each group had a recurrence of PPD.^[@ref162]^
### *Hormone treatments*
Sublingual 17β-estradiol (1 mg, 3 to 8 times/day to achieve a serum concentration of 400 pmol/L) in open treatment for 8 weeks resulted in rapid and significant improvement for women with severe PPD.^[@ref163]^ The women had very low serum estradiol concentrations at the pretreatment baseline (mean = 21.7 pg/mL), but whether the low estradiol levels differed from those of asymptomatic postpartum women could not be determined in the absence of a control group. Women with postpartum psychosis also responded to 17β-estradiol treatment in a similar study conducted by the same researchers.^[@ref164]^
Transdermal 17β-estradiol (delivery of 200 μg/day for 6 months) was significantly better than placebo for PPD, meeting criteria for major depressive disorder.^[@ref165]^ The response occurred in the first month of treatment and was sustained for the 6 months of the randomized, double-blind study. The effect on symptoms of a progestin added after 3 months was not reported; endometrial curettage at the end of treatment showed endometrial changes (sic) in three women, which resolved on follow-up.
A very small open pilot study administered estrogen immediately after delivery to prevent recurrent PPD.^[@ref166]^ A much lower relapse rate than expected in the ensuing year (9% versus an expected 35% to 60% without prophylaxis) suggested the utility of estrogen for high-risk women and supported the hypothesis that PPD may be triggered by rapid changes in the levels of estradiol in vulnerable women. However, possible problems with highdose estrogen therapy such as the need to administer coagulants and the interference of estrogen with breast milk are not yet answered.^[@ref167]^
Evidence does not support progesterone treatment for PPD.^[@ref168]^ Nor is there scientific evidence to support the prophylactic use of progesterone to prevent the recurrence of PPD. A randomized controlled trial of norethisterone enanthate given within hours of delivery reported an *increased* risk of developing PPD.^[@ref169]^
### *Psychotherapy*
Although there is increasing evidence that antidepressants are relatively safe, concerns for possible effects of pharmacologic treatments in the infants supports psychotherapy as the treatment for PPD in breast-feeding mothers. Evaluation of women mild-to-moderate major depression, who were randomized to interpersonal psychotherapy (IPT) or wait-list condition for 12 weeks demonstrated significantly greater improvement in PPD and social adjustment for the psychotherapy group.^[@ref170]^ Another study reported that fluoxetine and six sessions of cognitive behavioral therapy were each effective for minor and major depression occurring in the first 6 to 8 weeks postpartum, but also found no advantage to receiving both treatment modalities.^[@ref171]^
### *Management*
The most serious problem in the management of PPD is its underrecognition and undertreatment. Maternal depression can impair mother-infant bonding and affect cognitive and emotional development.^[@ref172]^ Women with a history of PPD or another mood disorder require close observation and prompt treatment of depressive symptoms. However, detection of possible PPD has been poor in routine clinical evaluation.^[@ref173]^ Identification of possible PPD significantly increased when a simple screening scale was used (6% vs 35%).^[@ref173]^ Another study showed that the rate of diagnosis of PPD increased from about. 4% to 11% following the implementation of a universal screening of postpartum women.^[@ref141]^ A brief screening scale (eg, the 10-item Edinburgh Postnatal Depression Scale)^[@ref171]^ appears to be an essential tool for identifying women who may have clinically significant depression in the postpartum period. Consensus guidelines indicate that the first-line treatment of PPD is antidepressants whether or not the mother is breast-feeding.^[@ref58]^ Although case-series comparisons have consistently reported no clinically significant differences in the infants of mothers taking or not taking antidepressant medications, relatively small numbers of women and their infants have been studied, and the findings are not sufficient to generalize to all infants at this time. Thus, it remains important to consider the risk/benefit equation for each woman in selecting treatment for PPD.
Conclusions
===========
Serotonergic antidepressants are generally the first-line treatment for menstrually related depressions - PPD, premenstrual dysphorias, and depression in the perimenopause - using regimens that are proven for major depression. While there is increasing evidence that the reproductive hormones are involved in these disorders, knowledge of their use as effective and safe treatments is still limited. Preliminary studies indicate that estrogen replacement therapy is effective for major and minor depression in perimenopausal women with or without a history of depression. Pilot data suggest that estradiol may be effective for severe PPD. Progesterone is clearly not effective for depressive symptoms in PMS/PMDD, the postpartum or perimenopause. Safety concerns for medications during pregnancy and breast-feeding point to psychotherapy as the treatment for PPD, but the risk calculation of each treatment modality should be made on an individual basis. No one treatment entirely fits each of these complex disorders with their still-heterogeneous populations, and a one-fits-all treatment approach is not possible. Nonetheless, understanding knowledge of the causes and treatments of women\'s depressions is increasing, and many women with these disorders can obtain relief with effective medical treatment.
ACTH
: adrenocorticotropic hormone
GnRH
: gonadotropin-releasing hormone
HRT
: hormone replacement therapy
5-HT
: 5-hydroxytryptamine (serotonin)
m-CPP
: m-chlorophenylpiperazine
OC
: oral contraceptive
PMDD
: premenstrual dysphoric disorder
PMS
: premenstrual syndrome
PPD
: postpartum depression
SSRI
: selective serotonin reuptake inhibitor
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION
============
Primary bone lymphoma (PBL) is defined as lymphoma localized to the bone without evidence of involvement of lymph nodes or other tissues at presentation. It one of the rarest primary bone malignancies, accounting for less than 5% of all primary bone tumors \[[@ref1]\]. PBL constitutes less than 1-2% of all malignant lymphomas in adults \[[@ref2]\]. Most PBLs are primary bone diffuse large B-cell type lymphomas with a rare occurrence of follicular, marginal zone and lymphoplasmacytic types \[[@ref3]\]. The long bones are primarily affected and the femur is the most commonly involved location as a single site \[[@ref2],[@ref4]\]. The common signs and symptoms are local bone pain with or without soft tissue swelling and pathological fracture. Spinal cord compression is reported in 14% of patients with vertebral involvement but the presence of B symptoms is relatively uncommon \[[@ref2],[@ref5]\]. PBL has a better prognosis following radiotherapy and chemotherapy than many other malignant tumors, and therefore early identification allows for appropriate treatments \[[@ref2],[@ref6]\]. In this report, the authors present a 61-year-old patient with a primary vertebra lymphoplasmacytic lymphoma presenting with spinal cord compression.
CASE REPORT
===========
A 61-year-old woman presented to the emergency department with a 3-month history of progressive chest and back pain, 1-month history of numbness and weakness of the lower extremities, and paraplegia for 1 day. Initially, the patient had a paroxysmal pain of the chest and lower back, which spread progressively to the bilateral scapula, oxter, and praecordia. Two months later, she felt numbness in her left lower extremity. After 1 week, she felt weakness in the lower extremities and had difficulty in walking. In three months, the symptoms worsened and hypoesthesia appeared. She became paraplegic the day before admission to hospital. The history revealed no cardiac, bowel, or bladder problems and her pain was not associated with motor or sensory neurological deficits at presentation. On physical examination she had no superficial lymphadenopathy or hepatosplenomegaly. The superficial hypoesthesia was located below the bilateral papilla plane. Tenderness and pain in percussion were positive at the level of vertebra T1-T3, in addition to a mild pain at the level of vertebra L4. Muscular force was normal for upper limbs, and the forces of hip, knee, and ankle joints were Grade III for both extending and flexing. Patellar reflex was strengthened and Achilles tendon reflex was normal. Computerized tomography (CT) scan showed a space-occupying lesion located in and outside the left canalis spinalis and foramen intervertebral levels T1-T3, accompanied with the destruction of the second vertebra. Serum calcium, albumin, and lactate dehydrogenase were within normal range. β2-microglobulin was slightly increased (2.57 mg/L; normal range is 0.7-1.8 mg/L). Blood and urine immunofixation were positive for λ chain. Serum-free lambda light chain was 175.3 mg/L (normal range is 6.72-22.81 mg/L), and κ chain was 10.6 mg/L (normal range is 5.81-21.04 mg/L). The 24-h urine λ chain was 949.2 mg (normal range is \<7.8 mg). Serum IgG was normal, but a slight decrease was seen for IgA and IgM, at 0.616 g/L (normal range is 0.7-4.0 g/L) and 0.192 g/L (normal range is 0.4-2.3 g/L), respectively. The erythrocyte sedimentation rate was 21 mm/h (normal range is 0-20 mm/L for females), and the International Prognostic Index score was 4. Bone marrow examinations with both smear and biopsy were normal. Informed consent was obtained from the patient.
On 30 September 2010, the patient received decompressive laminectomy, and the excision material was evaluated by hematoxylin-eosin (H&E) staining and immunohistochemistry. The results revealed diffuse tumor cell proliferation, which infiltrated and damaged the adjacent bone and soft tissues. The middle-sized tumor cells showed plasma cell features that were characterized by abundant cytoplasm and asymmetrical nuclei. The chromatin was granulated and Russell bodies could be observed. No plasmablasts could be found. The morphology suggested a proliferative disease of plasma cells ([Figures 1A-1D](#f1){ref-type="fig"}). The immunohistochemical staining result was CD20 (+) ([Figure 2A](#f2){ref-type="fig"}), CD138 (±) ([Figure 2B](#f2){ref-type="fig"}), CD56 (+) ([Figure 2C](#f2){ref-type="fig"}), CD79a (+) ([Figure 2D](#f2){ref-type="fig"}), CD38 (+), Lambda (+), NUM-1 (+), Ki-67 (some +), PAX5 (weakly positive), Bcl-2 (+/-), CyclinD1 (-), CD21 (-), CD5 (-), CD3 (-), CK-P (-), CD43 (-), CD23 (-), CD10 (-), Bcl-6 (-), MPO (-), Kappa (-), EMA (-), and EBER (-). The pathological diagnosis was lymphoplasmacytic lymphoma.
Positron emission tomography (PET)-CT was utilized to stage and determine the disease focus. The results revealed that the second vertebra had hypermetabolism accompanied with bone destruction, and the standardized uptake value (SUV) was 3.2 (average: 2.1) ([Figure 3A](#f3){ref-type="fig"}). The fourth lumbar vertebra was also hypermetabolic accompanied with bone destruction, and the SUV was 3.0 ([Figure 3B](#f3){ref-type="fig"}), which suggested infiltration of lymphoma and compression fracture of the lumbar vertebra. There was no significant hypermetabolic focus in other sites.
Local radiation therapy was given to the second thoracic and fourth lumbar vertebras (total: 44 Gy/22 f). The back pain gradually disappeared and muscle strength and feeling gradually recovered during the treatment. The patient then performed functional training for 1 month. Muscular tension and strength recovered to normal and she could move freely. Reexamination of the serum and urine immunofixation showed positivity of the λ chain. The λ chain was 413.1 mg (decrease of \>50%) for 24 h. PET-CT showed significantly decreased focus in the second thoracic and fourth lumbar vertebras ([Figures 3C and 3D](#f3){ref-type="fig"}). After that, the patient received R-CHOP chemotherapy (R: rituximab, 600 mg; C: cyclophosphamide, 1125 mg; H: epirubicin, 112.5 mg; O: vindesine, 4 mg; P: prednisone, 100 mg).
DISCUSSION
==========
PBL is a rare presentation of non-Hodgkin's lymphoma. It was first described as a distinct clinicopathological entity in 1939 by Parker and Jackson \[[@ref7]\]. The real prevalence of PBL is unclear because of the considerable difficulty in distinguishing primary from secondary bone lymphoma \[[@ref2]\]. According to Coley's criteria \[[@ref8]\], primary bone lymphoma should have: 1) a primary focus in a single bone; 2) positive histological diagnosis; and 3) no evidence of distant soft tissue or distant lymph node involvement. Regional lymph node involvement at diagnosis is not considered as an exclusion criterion \[[@ref9],[@ref10]\]. Recently, multiple bone involvement was also recognized as primary bone lymphoma, as long as the other 2 criteria are met \[[@ref9]\]. According to the WHO classification, lymphoma involving the bone can be classified in 4 groups: group 1, lymphoma in a single bone site with or without regional lymph node involvement; group 2, lymphoma in multiple bones, but no visceral or lymph node involvement; group 3, bone tumor with involvement of other visceral sites or lymph nodes at multiple sites; and group 4, lymphoma involving any other site and found by bone biopsy that was done to rule out possible involvement \[[@ref2]\]. In this case, the patient was considered as having PBL involving 2 separate sites of the bone based on the result of accessory examination.
The stage of PBL was established with the Ann Arbor staging system. Single localized bone lesions were classified as Stage IE, and in the case of lymph node involvement on the same side of the diaphragm, patients were considered to have Stage IIE. If lymph nodes in both sides of the diaphragm were involved, the case was considered as Stage IIIE. Stage IVE disease was defined as cases of multiple sites of bone involvement. Hence, the present patient was evaluated as a Stage IVE case.
Lymphoplasmacytic lymphoma (LPL) is one of the rare subtypes accounting for only 1% to 2% of all non-Hodgkin's lymphomas. This lymphoma arises from plasma cells, which are mature B-lymphocytes \[[@ref11]\]. It is most often associated with increased IgM protein produced by the lymphoma cells and is commonly referred to as Waldenström's macroglobulinemia (WM), which is considered an indolent lymphoma; the disease progresses very slowly and patients usually live many years after diagnosis. Recently, there was also a case report of non-secretory immunoglobulin of LPL \[[@ref12]\]. Instead of the signs and symptoms caused by infiltration and the circulating IgM of WM, such as lymphadenopathy, hepatosplenomegaly or hyperviscosity, our patient only presented with increased λ chain and spinal cord compression. The immunohistochemistry revealed positive staining for CD20, CD79a, and Lambda of B cell markers and negative for CD38 and CD138 of plasma cells, and so this patient was diagnosed with lymphoplasmacytic lymphoma. It is difficult to rule out plasmacytoma when the bone is involved, but B cell antigen is negative in these patients.
Several studies have suggested that combined modality (chemotherapy and radiotherapy) was the best treatment for patients with PBL \[[@ref2],[@ref13]\]. Beal et al. concluded that PBL patients treated with combination chemotherapy and irradiation had significantly better survival than patients treated with a single modality (chemotherapy or radiotherapy alone), but the 5-year overall survival rate between the 2 groups was not significantly different \[[@ref2],[@ref14]\]. Ramadan et al. reported patients with advanced-stage disease who received chemotherapy plus irradiation with a poor outcome when compared with those who received chemotherapy alone (10-year overall survival rates were 25% and 56%, respectively) \[[@ref4]\]. However, this difference must be very cautiously interpreted because the decision to use radiotherapy was individualized. It is possible that patients with more biologically aggressive disease were more likely to receive irradiation, obscuring its impact \[[@ref3]\]. Moreover, this result was derived from PBL of diffuse large-cell lymphoma \[[@ref3]\]. However, there was no available information in this report about lymphoplasmacytic lymphoma.
After therapeutic and diagnostic surgery, the patient received radiation therapy for local lesions of the vertebras. As she was in an advanced stage (Stage IVE), the R-CHOP regimen was given for the subsequent treatment. She is currently alive after the treatment and was periodically reexamined for almost 1 year; there was no sign of disease progression or relapse.
CONFLICT OF INTEREST STATEMENT
==============================
The authors of this paper have no conflicts of interest, including specific financial interests, relationships, and/ or affiliations relevant to the subject matter or materials included.
{#f1}
{#f2}
{#f3}
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Colorectal cancer (CRC) was the third most common cancer and the third most common cause of cancer-associated mortality in 2019 in the USA ([@b1-ijmm-46-04-1525]). Patients with stage I and II CRC who undergo colectomy, proctectomy or proctocolectomy often receive neoadjuvant radiation or chemotherapy treatments to lower the risk of recurrence, whereas patients with stage III and IV CRC are typically treated with chemo-therapy or radiotherapy ([@b2-ijmm-46-04-1525]). However, studies have suggested that drug resistance developed by patients with CRC limits the therapeutic efficacy of anticancer agents ([@b3-ijmm-46-04-1525],[@b4-ijmm-46-04-1525]). Therefore, developing novel effective chemotherapy drugs will benefit patients with CRC.
*Centella asiatica*, a herb of the apiaceae family ([@b5-ijmm-46-04-1525]), contains a variety of active ingredients, such as triterpenoids and saponins ([@b6-ijmm-46-04-1525]). AC and madecassoside are the principal monomer components of saponins in *C. asiatica* ([@b7-ijmm-46-04-1525]). Previous studies have revealed that AC possesses a wide range of biological functions, such as facilitating angiogenesis by increasing collagen synthesis ([@b8-ijmm-46-04-1525],[@b9-ijmm-46-04-1525]), inhibiting inflammatory process ([@b10-ijmm-46-04-1525],[@b11-ijmm-46-04-1525]), suppressing scar formation ([@b12-ijmm-46-04-1525],[@b13-ijmm-46-04-1525]) and promoting wound healing ([@b12-ijmm-46-04-1525],[@b13-ijmm-46-04-1525]). Additionally, AC has a crucial function in inflammatory lung diseases ([@b14-ijmm-46-04-1525]), neurological disorders ([@b15-ijmm-46-04-1525]) and osteogenic differentiation of human periodontal ligament cells ([@b16-ijmm-46-04-1525]). The clinical applicability of AC in the treatment of malignant tumors has been extensively studied over the last few years. Studies have confirmed that AC has a particular therapeutic effect on breast cancer ([@b17-ijmm-46-04-1525]), multiple myeloma and other tumors ([@b18-ijmm-46-04-1525]), and can improve the sensitivity of cancer cells to chemotherapy drugs ([@b18-ijmm-46-04-1525],[@b19-ijmm-46-04-1525]). However, to the best of our knowledge, the biological function of AC in CRC cells remains largely unknown.
AC reduces the incidence of DMBA-induced breast cancer in rats by inhibiting expression of TNF-α and IL-1β ([@b20-ijmm-46-04-1525]). AC treatment considerably suppresses the proliferation of human breast cancer MCF-7 cells, and induces cell apoptosis *in vitro* ([@b17-ijmm-46-04-1525]). A previous study on multiple myeloma cancer cells indicated that AC inhibits cellular proliferation by inducing autophagy coupled with elevated expression of LC3-II ([@b18-ijmm-46-04-1525]). Notably, AC increases sensitivity of tumor cells to vincristine by promoting apoptosis and inducing cell cycle arrest ([@b19-ijmm-46-04-1525]).
Therefore, the aim of the present study was to improve understanding of the biological effects of AC on CRC cells. Using several functional experiments *in vitro* and in mice models *in vivo*, the present findings revealed that AC could inhibit activation of the NF-κB signaling pathway by suppressing IκBα phosphorylation, which may be one of the primary mechanisms of AC-induced CRC cell apoptosis and cell cycle arrest. Overall, the present data suggest that AC exhibits antitumor effects in CRC.
Materials and methods
=====================
Cell culture and solution preparation
-------------------------------------
Normal human intestinal FHC cells and human CRC cell lines (HCT116, SW480 and LoVo) were obtained from the Cell Bank of Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. The cells were cultured in Dulbecco\'s modified Eagle\'s medium (Hyclone; GE Healthcare Life Sciences) supplemented with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin-streptomycin (Gibco; Thermo Fisher Scientific, Inc.) in an incubator at 37°C, containing 5% CO~2~. Culture medium was changed every 24 h.
A total of 95.90 mg AC (Santa Cruz Biotechnology, Inc.) was weighed and dissolved in 100 *µ*l dimethyl sulfoxide (DMSO; Santa Cruz Biotechnology, Inc.) to prepare a 1 mM storage solution. Subsequently, different concentrations of the working solutions (0.1, 0.5 and 2 *µ*M) were prepared using culture medium. Cells at the logarithmic growth phase were treated with the working solutions. A quantity of 0.2% DMSO was used as the control.
Cell Counting Kit-8 (CCK-8) assay
---------------------------------
Cell viability was assessed using a CCK-8 assay. CRC cells and FHC cells in the logarithmic growth phase were digested and seeded into 96-well plates at a density of 5×10^3^ cells/well followed by treatment with various concentrations of AC (0.1, 0.5 and 2 *µ*M) at 37°C. Cells treated with 0.3% DMSO were used as the negative controls. Subsequently, 10% CCK-8 reagent (Beijing Zoman Biotechnology Co., Ltd.) was added to each well at different time intervals (24, 48 and 72 h), and cells incubated for 2 h at 37°C. The absorbance was then measured using a microplate reader (Bio-Rad Laboratories, Inc.) at a wavelength of 450 nm.
Colony formation assay
----------------------
CRC cells in the logarithmic growth phase were seeded in 6-well plates (1×10^3^ cells/well) and then treated with different concentrations of AC (0.1, 0.5 and 2 *µ*M) at 37°C. After 7 days, the cell colonies were fixed with 4% paraformaldehyde solution (Beijing Solarbio Biotechnology Co., Ltd.) for 15 min at room temperature. Cells were rinsed three times with PBS and then stained with 0.1% crystal violet solution (Beijing Solarbio Biotechnology Co., Ltd.) according to the manufacturer\'s instructions. The number of colonies and the number of cells within the colonies were counted in five randomly selected fields under an optical microscope (magnification, ×40).
Cell cycle analysis
-------------------
CRC cells in the logarithmic growth phase were seeded into 6-well plates (5×10^5^ cells/well), and then incubated with various concentrations of AC (0.1, 0.5 and 2 *µ*M) for 24 and 72 h. The cells were collected, digested with 0.25% trypsin (Gibco; Thermo Fisher Scientific, Inc.) and centrifuged at 200 × g for 4 min at 4°C. Subsequently, the cells were fixed with 70% pre-cooled ethanol (prepared in PBS) overnight at 4°C, followed by centrifugation at 700 × g for 4 min at 4°C. Cells were rinsed with 1 ml pre-cooled PBS and resuspended in pre-cooled PBS at a density of 1×10^6^ cells/ml. Next, 100 *µ*l cell suspension was treated with 10 *µ*g/ml RNase A. Cells were incubated with 20 *µ*g/ml propidium iodide (PI) solution at room temperature for 10 min and then centrifuged at 700 × g for 4 min at 4°C. The cells were then rinsed with pre-chilled PBS and resuspended in pre-chilled PBS. Flow cytometry was performed to assess the cell cycle using a CytoFLEX flow cytometer (Beckman Coulter, Inc.). Data were analyzed using FlowJo (version 7.6.1; FlowJo LLC).
Cell apoptosis analysis
-----------------------
Cell apoptosis analysis was performed using the Annexin V-Alexa Fluor 488/PI kit (cat. no. CA1040, Beijing Solarbio Biotechnology Co., Ltd.) according to the standard procedures. CRC cells were treated with various concentrations of AC (0.1, 0.5 and 2 *µ*M) for 24 and 72 h. Cells were collected, rinsed twice in PBS and then resuspended in a binding buffer. The cells were subsequently stained with Annexin V-Alexa Fluor 488/PI for 5 min at room temperature following the manufacturer\'s instructions, and cell apoptosis detected using a CytoFLEX flow cytometer (Beckman Coulter, Inc.). Data analysis was performed using FlowJo (version 7.6.1; FlowJo LLC).
Analysis of mitochondrial membrane potential
--------------------------------------------
Mitochondrial Membrane Potential assay kit with JC-1 (cat. no. M8650; Beijing Solarbio Biotechnology Co., Ltd.) was used to evaluate the mitochondrial membrane potential of CRC cells following treatment with AC. CRC cells in the logarithmic growth phase were seeded into 6-well plates (5×10^5^ cells/well), and then treated with various concentrations of AC (0.1, 0.5 and 2 *µ*M) for 72 h. The cells were harvested, rinsed twice with PBS and incubated with 1.2 *µ*g/ml phenol red-free medium containing JC-1 probes for 30 min at 37°C. Subsequently, the cells were collected, rinsed and resuspended in PBS. Fluorescence of the cells was measured using a CytoFLEX flow cytometer (Beckman Coulter, Inc.) at wavelengths of 590 and 530 nm. Data were analyzed using FlowJo (version 7.6.1; FlowJo LLC).
Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis
-------------------------------------------------------------------------------
TRIzol reagent (Thermo Fisher Scientific, Inc.) was used to isolate the total RNA from CRC cells treated with AC. The concentration and purity of RNA was detected with a UV spectrophotometer (A260/A280). A ratio of A260/A280 between 1.8 and 2.0 was regarded as a sufficient quality. A quantity of 500 ng total RNA was reverse transcribed into cDNA using a PrimeScript™ RT Master mix (cat. no. RR036A; Takara Bio, Inc.). The RT conditions were 37°C for 15 min, followed by 85°C for 5 sec, and storage at 4°C. qPCR was performed on a LightCycler^®^ 96 system (Roche Diagnostics) using One Step TB Green^®^ PrimeScript™ RT-PCR kit (cat. no. RR066A; Takara Bio, Inc.). The thermocycling conditions were as follows: Pre-denaturation at 95°C for 2 min; 40 cycles of denaturation at 95°C for 30 sec, annealing at 60°C for 30 sec, and extension at 72°C for 30 sec. The relative expression levels of genes were calculated using the 2^−ΔΔCq^ method ([@b21-ijmm-46-04-1525]), and β-actin was used as an internal control. The experiment was performed in triplicate. All primer sequences used in the present study are listed in [Table I](#tI-ijmm-46-04-1525){ref-type="table"}.
Western blotting
----------------
Total protein was extracted from CRC cells or tumor tissues using RIPA lysis buffer (Beyotime Institute of Biotechnology) containing 1% protease inhibitor (Beyotime Institute of Biotechnology). Protein concentration was deter-mined using a BCA protein assay kit (Nanjing KeyGEN Biotech, Co., Ltd.) according to the manufacturer\'s protocol. Subsequently, proteins (30 *µ*g per lane) were separated by 10% SDS-PAGE (Nanjing KeyGEN Biotech, Co., Ltd.) and then transferred onto polyvinylidene fluoride membranes (EMD Millipore). The membranes were blocked with 5% skimmed milk (prepared in TBST containing 0.1% Tween-20) for 2 h at room temperature and incubated with corresponding primary antibodies overnight at 4°C. The following primary antibodies were used for incubation: Anti-NF-kB P65 (1:1,000; cat. no. ab16502; Abcam), anti-phosphorylated (p)-P65 (1:1,500; cat. no. ab86299; Abcam), anti-caspase-3 (1:1,000; cat. no. ab4051, Abcam), anti-cleaved caspase-3 (1:1,000; cat. no. ab32042; Abcam), anti-caspase-9 (1:1,000; cat. no. ab32539; Abcam), anti-cleaved caspase-9 (1:1,000; cat. no. ab32539; Abcam), anti-Bcl-2 (1:1,000; cat. no. ab32124; Abcam), anti-Bax (1:1,000; cat. no. ab32503; Abcam), anti-CDK4 (1:1,000; cat. no. ab199728, Abcam), anti-Cyclin D1 (1:2,000; cat. no. ab205718; Abcam), anti-IκBα (1:1,000; cat. no. 9242; Cell Signaling Technology, Inc.), anti-p-IκBα (1:1,000; cat. no. 2859; Cell Signaling Technology, Inc.), anti-Histone H3 (1:3,000; cat. no. ab1791; Abcam) and anti-β-actin (1:5,000; cat. no. ab8227; Abcam). Subsequently, the membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG H&L (1:5,000; cat. no. ab205718; Abcam) or HRP-conjugated goat anti-mouse IgG H&L (1:5,000; cat. no. ab205719; Abcam) secondary antibodies for 2 h at room temperature. The immunoreactive proteins were then visualized using immobilon western chemiluminescent HRP substrate (EMD Millipore). The western blots were analyzed with Image Lab 3.0 software (Bio-Rad Laboratories, Inc.).
Nuclear/cytosol fractionation kit assay
---------------------------------------
Nuclear and cytosolic fractions were prepared using the nuclear/cytosol extraction kit (cat. no. K266-25; BioVision, Inc.) according to the manufacturer\'s instructions. All buffers were stored in ice throughout the experimental period. All centrifugation steps were performed at 4°C. CRC cells were treated with AC (2 *µ*M) for 72 h at 37°C, harvested and rinsed twice with pre-cooled PBS. Cells were subsequently centrifuged at 600 × g for 5 min. A total of 1×10^6^ cells were treated with 200 *µ*l pre-chilled CEB-A mix (1 ml CEB-A mix containing 1 *µ*l DTT and 2 *µ*l protease inhibitor), vortexed vigorously for 15 sec to fully resuspend the cell pellets and incubated on ice for 10 min. Subsequently, 11 *µ*l pre-cooled cytosol extraction buffer-B was added, vortexed vigorously for 5 sec, and incubated on ice for 1 min. The tube was continuously vortexed vigorously for 5 sec, followed by centrifugation at 14,000 × g for 5 min. The supernatant (cytoplasmic extract) was immediately transferred to a clean pre-cooled tube and the tube was placed on ice.
Next, 100 *µ*l ice-cold nuclear extraction buffer mix (1 ml NEB containing 2 *µ*l protease inhibitor cocktail and 1 *µ*l DTT) was added to the cell pellets (nuclei), and then vortexed vigorously for 10 sec. The cell pellets were placed on ice for 40 min, vortexed for 15 sec and the samples placed on ice after 10 min. Finally, following centrifugation at 16,000 × g for 5 min, the supernatant (nuclear extract) was immediately transferred into a pre-cooled clean tube and stored at -80°C. The cytosol and nuclear extracts were then used for western blotting to measure P65 expression level.
Animal experiments
------------------
A total of 18 5-week-old female BALB/c-nude mice (weight, 20-25 g) were purchased from the Model Animal Research Center of Nanjing University and maintained according to regulations of the Animal Care Committee of The First Affiliated Hospital of Xiamen University. The housing conditions of mice were as follows: The temperature was 22-24°C, humidity was 45-60%, ventilation was 15 times/h, and 12 h of light and 12 h of dark. Mice were provided food and water *ad libitum*. In the present study, mice were randomly divided into three groups (n=6 per group): i) Control group; ii) 5 mg/kg AC treatment group; and iii) 10 mg/kg AC treatment group. HCT116 cells (5×10^6^) were subcutaneously injected into the right hind legs of the mice. Subsequently, 1 week later, 5 or 10 mg/kg AC or PBS (10 mg/kg) was administered to mice by oral gavage every 2 days for 6 consecutive weeks. Tumor diameters were measured using calipers every week. Mice were euthanized at 42 days after the first administration of AC or PBS by cervical dislocation, and the tumors were carefully excised and weighed. All mice were in good health throughout the experimental period. All *in vivo* experiments were performed at the Animal Center of The First Affiliated Hospital of Xiamen University (Xiamen, China) according to the guidelines of the National Institutes of Health and approved by the Animal Care and Use Committee of The First Affiliated Hospital of Xiamen University (approval no. XMU-AEA-20180137).
Statistical analysis
--------------------
Statistical analyses were performed using GraphPad Prism 6.0 software (GraphPad Software, Inc.) and SPSS software version 21.0 (IBM Corp.). Comparisons between groups were performed using one-way analysis of variance with Bonferroni\'s post hoc analysis. Data from three independent experiments are expressed as the mean ± standard deviation. P\<0.05 was considered to indicate a statistically significant difference.
Results
=======
AC inhibits CRC cell proliferation and induces CRC cell cycle arrest at the G0/G1 phase
---------------------------------------------------------------------------------------
CCK-8 assay and colony formation assay were performed to determine the effect of AC on the viability of CRC cells. Results of the CCK-8 assay demonstrated that AC significantly reduced the viability of HCT116, SW480, and LoVo cells in a time and dose-dependent manner ([Fig. 1A-C](#f1-ijmm-46-04-1525){ref-type="fig"}), whereas AC had no significant effects on normal human intestinal FHC cells at a given range of concentrations (0.1-8 *µ*M) ([Fig. 1F](#f1-ijmm-46-04-1525){ref-type="fig"}). However, the viability of FHC cells was significantly reduced after treating with 16 *µ*M AC for 72 h, indicating that high concentration of AC has an adverse effect on human normal epithelial cells. Additionally, HCT116 cells, with the lowest IC~50~ value, were selected for subsequent analyses. Colony formation assay revealed that AC significantly inhibited the colony formation of HCT116 cells in a dose-dependent manner ([Fig. 1D and E](#f1-ijmm-46-04-1525){ref-type="fig"}). Furthermore, flow cytometry analysis indicated that AC treatment for 24 or 72 h significantly induced CRC cell cycle arrest at the G0/G1 phase. The percentage of cells in the G2/M phase significantly decreased following treatment with AC for 24 or 72 h and no difference was observed for the percentage of cells in the S phase ([Fig. 2](#f2-ijmm-46-04-1525){ref-type="fig"}).
AC suppresses the expression levels of cell cycle-associated genes
------------------------------------------------------------------
RT-qPCR and western blotting assays were performed to further understand the specific mechanism of the effects of AC on the CRC cell cycle. The findings revealed that AC significantly decreased the expression levels of CDK4 and Cyclin D1, whereas the expression levels of P53 and P21 were significantly higher in AC-treated cells compared with the control cells ([Fig. 3](#f3-ijmm-46-04-1525){ref-type="fig"}).
AC decreases mitochondrial membrane potential and promotes CRC cell apoptosis
-----------------------------------------------------------------------------
In the present study, JC-1 probes were used to measure mitochondrial membrane potential, and cell apoptosis was detected using flow cytometry. JC-1, a cyanine dye, promotes discrimination of energized and deenergized mitochondria, where the normally green fluorescent dye forms red fluorescent aggregates when concentrated in energized mitochondria ([@b22-ijmm-46-04-1525]). However, when cell apoptosis occurs, the mitochondrial transmembrane potential is decreased, JC-1 is released from the mitochondria, decreasing in concentration and maintaining the monomeric form with green fluorescence. Therefore, it is widely used to determine the apoptosis of the cell by detecting green (low) and red (high) fluorescence. The present results indicated that AC significantly decreased the mitochondrial membrane potential of HCT116 cells in a dose-dependent manner ([Fig. 4A and B](#f4-ijmm-46-04-1525){ref-type="fig"}). Furthermore, AC significantly increased the apoptosis of CRC cells compared with the untreated cells ([Fig. 4C-F](#f4-ijmm-46-04-1525){ref-type="fig"}).
AC regulates the expression of apoptosis-related genes in CRC cells
-------------------------------------------------------------------
RT-qPCR and western blotting were performed to further elucidate the biological function of AC in CRC cells ([Fig. 5](#f5-ijmm-46-04-1525){ref-type="fig"}). The results demonstrated that AC significantly increased the activation of caspase-9 and caspase-3 in HCT116 cells in a dose-dependent manner ([Fig. 5C](#f5-ijmm-46-04-1525){ref-type="fig"}), but had no effect on the mRNA levels of caspase-9 and caspase-3 ([Fig. 5A](#f5-ijmm-46-04-1525){ref-type="fig"}). In addition, AC significantly downregulated Bcl-2 expression and upregulated Bax expression at both the mRNA and protein levels ([Fig. 5A-C](#f5-ijmm-46-04-1525){ref-type="fig"}. The ratio of Bcl-2/Bax protein expression was significantly decreased following AC treatment compared with the control cells ([Fig. 5C](#f5-ijmm-46-04-1525){ref-type="fig"}).
AC suppresses phosphorylation of IκBα and NF-κB p65, inhibiting the NF-κB signaling pathway
-------------------------------------------------------------------------------------------
The NF-κB signaling pathway is crucial for the regulation of various biological activities of eukaryotic cells, such as cell proliferation, survival and apoptosis ([@b23-ijmm-46-04-1525],[@b24-ijmm-46-04-1525]). Inhibitors of the NF-κB signaling pathway are potential anticancer drug candidates ([@b25-ijmm-46-04-1525]). The present study assessed the activation and expression of molecules associated with the NF-κB pathway in HCT116 cells treated with AC using RT-qPCR and western blotting ([Fig. 6](#f6-ijmm-46-04-1525){ref-type="fig"}). The results revealed that treatment of HCT116 cells with AC significantly decreased phosphorylation of IκBα and P65 ([Fig. 6B and C](#f6-ijmm-46-04-1525){ref-type="fig"}), but had no effect on the mRNA levels of total IκBα and P65 compared with the control cells ([Fig. 6A](#f6-ijmm-46-04-1525){ref-type="fig"}). Furthermore, AC significantly increased the content of P65 in the nucleus, but decreased P65 content in the cytoplasm, suggesting that AC treatment inhibited nuclear translocation of P65 ([Fig. 6D and E](#f6-ijmm-46-04-1525){ref-type="fig"}).
Synergistic inhibitory effects of AC and the NF-κB signaling pathway inhibitor JSH-23 on CRC cells
--------------------------------------------------------------------------------------------------
Administration of 0.5 *µ*M AC and the NF-κB signaling pathway inhibitor JSH-23 (100 nM) significantly suppressed CRC cell viability, significantly induced cell cycle arrest at the G0/G1 phase, significantly promoted cell apoptosis and significantly deceased the phosphorylation of IκBα and P65 in HCT116 cells in comparison to HCT 116 cells treated either with AC or JSH-23 ([Fig. 7](#f7-ijmm-46-04-1525){ref-type="fig"}). A synergistic inhibitory effect of AC and JSH-23 on CRC cells was observed, indicating that AC may be a promising adjuvant agent in treatment of CRC.
AC inhibits CRC tumor cell growth in vivo
-----------------------------------------
The present findings revealed that administration of AC in mice with tumors notably inhibited tumor growth; tumor diameters and tumor volumes were significantly decreased in comparison with the control group ([Fig. 8A and B](#f8-ijmm-46-04-1525){ref-type="fig"}). Furthermore, RT-qPCR and western blotting revealed that the protein expression levels of P53, P21 and Bax were significantly upregulated, whereas expression levels of CDK4, Cyclin D1 and Bcl-2 were significantly downregulated in tumor tissues following treatment with AC compared with untreated controls. Moreover, administration of AC caused a significant decrease in the phosphorylation of IκBα and P65 ([Fig. 8C and D](#f8-ijmm-46-04-1525){ref-type="fig"}). These results were consistent with those of the *in vitro* analysis.
Discussion
==========
The present study revealed that AC, a plant-derived triterpenoid, induced cell cycle arrest at the G0/G1 phase, promoted CRC cell apoptosis and suppressed cell viability. AC treatment contributed to the aberrant expression of the cell cycle-related genes CDK4, Cyclin D1 and P53 and P21. AC increased the activation of caspase-9 and caspase-3 (apoptosis-related genes), upregulated Bax level, and downregulated Bcl-2 expression. Furthermore, AC inhibited the phosphorylation of IκBα and NF-κB p65, resulting in the inhibition of the NF-κB signaling pathway. This consequently affects various biological activities in eukaryotic cells, including cell proliferation, survival and apoptosis. Furthermore, a synergistic inhibitory effect of AC and the NF-κB signaling pathway inhibitor, JSH-23 on CRC was observed. AC suppressed proliferation of CRC cells both *in vitro* and *in vivo*, implying that it can be a potential adjuvant agent for the treatment of CRC.
A number of studies have demonstrated that AC exhibits antitumor activity ([@b17-ijmm-46-04-1525],[@b18-ijmm-46-04-1525],[@b20-ijmm-46-04-1525]). AC increases the sensitivity of the human oral epithelial cancer cells and breast cancer cells to chemotherapy drugs by increasing intracellular reactive oxygen species and activating apoptotic pathways *in vitro* ([@b19-ijmm-46-04-1525]). AC inhibits breast cancer progression in mice by downregulating IL-1β expression and activating downstream signaling pathways *in vivo* ([@b20-ijmm-46-04-1525]). The present study demonstrated that AC significantly suppressed proliferation and colony formation in CRC cells. Decreased mitochondrial membrane potential is a common phenomenon of mitochondrial apoptosis in cancer cells ([@b4-ijmm-46-04-1525]), which results in the release of mitochondrial apoptogens to initiate the caspase cascade, causing programmed cell death ([@b26-ijmm-46-04-1525],[@b27-ijmm-46-04-1525]). The present study identified that AC treatment for 24 h decreased the mitochondrial membrane potential in HCT116 cells and induced CRC cell apoptosis. At this point, the proportion of cell apoptosis in the AC-treated group was not so high, but the mitochondrial membrane potential was significantly altered, indicating that the change of mitochondrial membrane potential occurred earlier than cell apoptosis. When HCT116 cells were treated with AC for 72 h, cell apoptosis was significantly elevated in comparison with the untreated cells. Hence, the mitochondrial membrane potential was not detected.
Previous studies have demonstrated that the decreased ratio of Bcl-2/Bax regulates apoptosis in cells ([@b28-ijmm-46-04-1525],[@b29-ijmm-46-04-1525]). In the present study, AC-induced cell apoptosis was coupled with upregulation of Bax, cleaved caspase-3 and cleaved caspase-9, and the downregulation of Bcl-2. Additionally, the cell cycle transition from G0/G1 to S is effected by multiple genes, such as CDK4/6 and Cyclin D1 ([@b30-ijmm-46-04-1525]). A study has indicated that downregulation of CDK4 expression causes cell cycle arrest at the G0/G1 phase ([@b31-ijmm-46-04-1525]), which may explain the mechanism of antitumor effects of several drugs. The present results indicated that AC significantly inhibited the proliferation of CRC cells by interrupting the cell cycle at the G0/G1 phase. AC treatment substantially reduced the expression of CDK4 and Cyclin D1. However, the specific mechanism involved in the regulation of CDK4 and Cyclin D1 gene expression remains unknown.
AC suppresses the inflammatory response after injury via regulation of the NF-κB signaling pathway ([@b14-ijmm-46-04-1525],[@b32-ijmm-46-04-1525],[@b33-ijmm-46-04-1525]). IκBα is an inhibitory protein of the NF-κB signaling pathway that normally binds to P65 resulting in the inhibition of P65 phosphorylation and its nuclear translocation ([@b34-ijmm-46-04-1525]). Conversely, IκBα phosphorylation promotes its own ubiquitination and degradation, thereby eliminating the inhibitory effect on P65 ([@b35-ijmm-46-04-1525],[@b36-ijmm-46-04-1525]). Once P65 enters the nucleus, it triggers a series of gene expressions such as Bcl-2 and CDK4 ([@b37-ijmm-46-04-1525]). Results of the present study revealed that AC treatment significantly reduced IκBα and P65 phosphorylation, resulting in inhibition of P65 nuclear translocation and interfering with the expression of downstream molecules. Although results of the *in vivo* experiments demonstrated that AC administration significantly inhibited tumor growth without evident adverse effects on mice, the clinical use of AC as an antitumor drug requires further study.
In conclusion, AC inhibited activation of the NF-κB signaling pathway and subsequently modulated the expression of its downstream genes by attenuating IκBα and P65 phosphorylation, which may induce apoptosis and cell cycle arrest in CRC cells. Thus, proliferation of CRC cells *in vitro* and tumor growth *in vivo* were suppressed. Therefore, AC is a promising adjuvant with antitumor effects in CRC.
Not applicable.
Funding
=======
This study was financed by the National Social Science Foundation of China (grant no. 81871877), National Major Scientific and Technological Special Project for \'Significant New Drugs Development\' (grant no. 2020ZX09201005) and Youth Foundation of Xiamen Cancer Center (grant nos. ZLYYA201710 and ZLYYA201704).
Availability of data and materials
==================================
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Authors\' contributions
=======================
YM and CK conceived and designed the study. XZ, YL and YM analyzed the experimental data. YM and CK acquired the funding. XZ, YL, CR, TL and FD performed the experiments. XZ, YL, CR, YM and CK participated in the design of the experimental methods. XZ, YM and CK wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
==========================================
All animal experiments were approved by the Animal Care and Use Committee of The First Affiliated Hospital of Xiamen University (Xiamen, China; approval no. XMU-AEA-20180137) and carried out in accordance with the Guide for the Care and Use of Laboratory Animals.
Patient consent for publication
===============================
Not applicable.
Competing interests
===================
The authors declare that there have no competing interests.
{#f1-ijmm-46-04-1525}
{#f2-ijmm-46-04-1525}
{#f3-ijmm-46-04-1525}
{#f4-ijmm-46-04-1525}
{#f5-ijmm-46-04-1525}
{#f6-ijmm-46-04-1525}
{#f7-ijmm-46-04-1525}
{#f8-ijmm-46-04-1525}
######
Primers for reverse transcription-quantitative PCR.
Gene Forward sequence Reverse sequence
----------- ------------------------------- -------------------------------
β-actin 5′-CCTCGCCTTTGCCGATCC-3′ 5′-GGATCTTCATGAGGTAGTCAGTC-3′
IκBα 5′-CTCCGAGACTTTCGAGGAAATAC-3′ 5′-GCCATTGTAGTTGGTAGCCTTCA-3′
P65 5′-ATGTGGAGATCATTGAGCAGC-3′ 5′-CCTGGTCCTGTGTAGCCATT-3′
Caspase-9 5′-CTCAGACCAGAGATTCGCAAAC-3′ 5′-GCATTTCCCCTCAAACTCTCAA-3′
Caspase-3 5′-AGAGGGGATCGTTGTAGAAGTC-3′ 5′-ACAGTCCAGTTCTGTACCACG-3′
P21 5′-TGTCCGTCAGAACCCATGC-3′ 5′-AAAGTCGAAGTTCCATCGCTC-3′
Bax 5′-CCCAGAGTTTGAGCCGAGTG-3′ 5′-CCCATCCCTTCGTCGTCCT-3′
Cyclin D1 5′-CAATGACCCCGCACGATTTC-3′ 5′-CATGGAGGGCGGATTGGAA-3′
CDK4 5′-CTGGTGTTTGAGCATGTAGACC-3′ 5′-GATCCTTGATCGTTTCGGCTG-3′
Bcl-2 5′-CCAGCGTATATCGGAATGTGG-3′ 5′-CCATGTGATACCTGCTGAGAAG-3′
P53 5′-GTTTCCGTCTGGGCTTCTTG-3′ 5′-CACAACCTCCGTCATGTGCT-3′
[^1]: Contributed equally
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
Left ventricular non-compaction (LVNC) is an unclassified cardiomyopathy and there is no consensus in the diagnosis of LVNC. The aims of this study were to establish quantitative method to diagnose isolated LVNC (INC) using cardiac magnetic resonance (CMR) imaging and to suggest novel qualitative method to diagnose isolated LVNC.
Methods
=======
This retrospective study included 145 subjects with moderate to severe trabeculation of LV myocardium \[24 patients with isolated LVNC, 33 patients with non-isolated LVNC, 30 patients with dilated cardiomyopathy (DCM) with noncompaction, 27 patients with DCM and 31 healthy control subjects\]. LVNC patients had to fulfill Petersen\'s CMR criteria. LV ejection fraction, global LV volume, trabeculated LV volume, and number of segments with late gadolinium enhancement were measured. And most prominent noncompacted (NC), compacted (C), normal mid-septum, normal mid-lateral wall and apical trabeculation thickness on the end-diastolic frames of each long-axis slice was also measured.
Results
=======
*Results* In the patients with isolated LVNC, the percentage of trabeculated LV volume was 1.4 times higher (42.6 ± 13.8) than in DCM (30.3 ± 14.3, p\<0.0001)\[YHC1\], and 1.7 times higher than in controls (24.8 ± 7.1, p\<0.0001). However, there were no significant differences between INC and DCMNC (47.1 ± 17.3, p = 0.210). And a value of percentage of trabeculated LV volume above 32% was predictive of isolated LVNC with a specificity of 90.3% (CI, 74.2-98.0%) and sensitivity of 79.2% (CI, 57.8-92.9%). A value of NC/septum over 1.1 was considered predictive for isolated LVNC with a specificity of 80.6% (CI, 62.5-92.5%) and sensitivity of 95.8% (CI, 78.9-99.9%). And a value of apex/C above 3.1 was considered predictive of isolated LVNC with a specificity of 93.5% (CI, 78.6-99.2%) and sensitivity of 87.5% (CI, 67.6-97.3%).
Conclusions
===========
As a quantitative approach, a trabeculated LV volume above 32% of the total LV volume is diagnostic for isolated LVNC with high sensitivity and specificity. And as a qualitative approach, apex/C and NC/septum ratio could be useful for supplemental diagnostic criteria.
Funding
=======
None.
| {
"pile_set_name": "PubMed Central"
} |
Cholelithiasis is a major public health problem in developed countries, affecting up to 20% of the population. [@JR1800087cr-1] Cholelithiasis is responsible for 90 to 95% of cases of acute cholecystitis, and 2% of patients with nonsevere cholecystitis experience recurrence within 8 to 10 weeks. [@JR1800087cr-2] Cholecystectomy is the first choice for the treatment of symptomatic cholelithiasis, especially for patients with acute cholecystitis. [@JR1800087cr-3] Intraoperative complications of laparoscopic cholecystectomy include: bile duct and organ injury and bleeding due to vascular injury. Uncontrollable hemorrhage during laparoscopic cholecystectomy occurs in 0.1 to 1.9% of all cases, [@JR1800087cr-4] leading to conversion to open surgery in up to 2% of all laparoscopic cholecystectomies. Furthermore, in 88% of these events, bleeding originates from the gallbladder bed. [@JR1800087cr-5] The gallbladder bed is reported as a common vascular injury site, mostly secondary to trauma to the middle hepatic vein. [@JR1800087cr-6] We present a case report of the management of a middle hepatic vein injury during laparoscopic cholecystectomy.
Case Presentation
=================
A 67-year-old male presented to our hospital after experiencing right upper quadrant abdominal pain for the last 2 weeks. Clinical history included treatment for follicular B lymphoma, stage IIA, and during his checkups a computed tomography (CT) scan of the abdomen was performed describing the presence of cholelithiasis. The patient was scheduled for an elective laparoscopic gallbladder removal on an outpatient basis. The day of the surgery, the patient was in good general health with normal vital signs. Abdominal examination revealed mild abdominal pain without signs of an acute abdomen. During the procedure, the hepatocystic triangle was dissected without complications. During dissection of the gallbladder from the cystic plate, a major venous hemorrhage erupted. Direct compression and electrocautery were not successful in controlling the bleeding. [Fig. 1](#FI1800087cr-1){ref-type="fig"} . Due to continuous bleeding, hemodynamic changes and the inability to control bleeding with laparoscopy, the laparoscopy was aborted and a supraumbilical laparotomy incision was performed. The bleeding was controlled with ligation of the vessel and the peritonization of the gallbladder bed. The patient had a successful recovery and was discharged on postoperative day 3. On outpatient follow-up, the patient remained asymptomatic.
{#FI1800087cr-1}
Discussion
==========
Laparoscopic cholecystectomy has been established as the gold standard for the treatment of gallstone disease, but it can be associated with significant morbidity and mortality. [@JR1800087cr-7] Bleeding complications are an important cause of mortality, especially when facing major bleeding during a laparoscopic procedure where the bleeding control can be technically challenging. Between 10 and 15% of patients will display a large branch of the middle hepatic vein adherent to the gallbladder bed, presenting an increased risk of vein injury during cholecystectomy. Excluding major vessels, bleeding can originate from the gallbladder bed itself, and the middle hepatic vein has been described to be a cause of uncontrollable bleeding. [@JR1800087cr-4] [@JR1800087cr-8] Misawa et al reported that the branch of the middle hepatic vein was completely adherent to the gallbladder bed in 5 of the 50 patients, and in one patient the diameter of the branch was 3.5 mm. In three patients, branch diameters were 3.0 to 3.8 mm traversed as close as 1.0 mm from the gallbladder bed. [@JR1800087cr-6] The literature describes varies reflections on how to face this possible complication. Some proposed strategies include delayed cholecystectomy and using low energy cauterization. Other suggestions include a screening method to determine the middle hepatic vein distance from the gallbladder bed before laparoscopy. [@JR1800087cr-6] As reported in this case, middle hepatic vein injury is an uncommon yet eventful situation. It is critical that the surgeon keeps in mind this anatomy, especially during the final steps of gallbladder dissection from the plate during laparoscopic cholecystectomy [Fig. 2](#FI1800087cr-2){ref-type="fig"} .
{#FI1800087cr-2}
**Conflict of Interest** None.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
A unique feature of the space radiation environment is the presence of galactic cosmic rays (GCRs) and solar particle events (SPEs). GCR involves protons and fully ionized atomic nuclei such as ^56^Fe, while SPE includes predominantly low-to-medium energy protons with a small heavy ion component. These exposures pose a significant hazard to space flight crews not only during the mission but also at later times after the mission when slow-developing adverse effects could finally become apparent. The hazards associated with the space environment will likely impact many organs, including the brain.
In people exposed to irradiation, the nature and extent of behavioral and cognitive changes is variable and cannot easily be predicted on the basis of radiation dose and type. This suggests the involvement of genetic factors. Conditions such as early-onset Alzheimer's disease show simple Mendelian patterns of inheritance for which mutations in a single gene are necessary and sufficient to cause the disorder. In contrast, complex behavioral traits and their potential alteration by space irradiation are likely influenced by multiple genes. To date, radiation studies typically involved inbred strains of mice and rats, crosses of inbred mice to generate F1 mice, stocks with very limited genetic heterogeneity, or outbred mice or rats but without analysis of the genetic factors of the individual animals. The advent of advanced heterogeneous stocks, inexpensive high-throughput single-nucleotide polymorphism (SNP) genotyping, and new analytical approaches enables studying gene--environment interactions and genotype--phenotype relationships involved in radiation effects on the brain. Analyzing CNS radiation effects in genetically heterogeneous mouse populations therefore likely offer a better model for genetically diverse human populations. In this study, we utilized animals from the HS/Npt colony, which was developed in 1991 as a tool to investigate complex genetic traits and captures a significant amount of the genetic diversity that is available in *Mus musculus* ([@B35]). These mice are now at G~70~, leading to further expansion of the genetic map.
Based on practical considerations of performing radiation studies at Brookhaven National Laboratories (BNLs) and the number of mice required for studying genotype--phenotype relationships, we searched for cognitive tests that: (1) allow testing relatively large groups of mice; (2) can be performed by mice of various genetic backgrounds, including HS/Npt mice; (3) are sensitive to detect effects of irradiation on the brain; and (4) are relevant with regard to environmental psychological and physical stressors astronauts experience during space missions. Contextual and cued fear learning and memory fit these criteria. They allow testing of relatively large groups of mice and can be used to distinguish hippocampus-dependent function (i.e., contextual fear) from hippocampus-independent, amygdala-dependent functions (i.e., cued fear) ([@B2], [@B3]). Fear learning and memory has been assessed in mice of various genetic backgrounds ([@B26]; [@B10]; [@B25]; [@B45]), including HS/Npt mice ([@B42]). Fear learning and memory is also being used as cognitive test in humans ([@B22]) and is associated with trait vulnerability to anxiety ([@B14]). Including controlled environmental emotional stressors in assessments of effects of space irradiation on brain function is relevant, as the environmental conditions astronauts experience during space missions include not only ionizing radiation but also psychological and physical stressors ([@B41]). Exposure to space radiation might modulate the response to such stressors and this response might depend on the genetic background. Previous studies have shown that fear conditioning is sensitive to effects of gamma rays ([@B37]; [@B24]; [@B17]), ^28^Si ion irradiation ([@B32], [@B31]) and ^56^Fe ion irradiation ([@B43]; [@B29]) on hippocampus-dependent contextual fear memory and effects of gamma rays ([@B24]), ^16^O ion irradiation ([@B30]) on hippocampus-independent and amygdala-dependent cued fear memory, and effects of ^40^Ca ion exposure on locomotor baseline activity (prior to the first tone) and responses to tone and shock without affecting fear learning and memory ([@B33]). In this study, the effects of HZE ion or gamma ray irradiation on contextual and cued fear learning and memory were assessed in the genetically heterogeneous HS/Npt population, by irradiating mice with 0.4°Gy of 240 MeV/n ^28^Si or 600 MeV/n ^56^Fe ions or 3°Gy of ^137^Cs gamma rays, or sham irradiating them.
There are three possibilities arising from cognitive tests following ionizing radiation. Cognitive measurements could be strongly altered resulting in significant group mean differences. Second, cognitive measurements could be affected in more subtle ways mitigated by internal mechanisms. Third, ionizing radiation could have minimal effects on cognition. In the case of compensatory mechanisms, genetic factors are a distinct possibility. Genome-wide association studies have detected numerous genetic variants associated with complex behavioral and cognitive traits. For the most part, the QTL identified so far explain only a small proportion of phenotypic variability ([@B5]). Dissection of genetic architecture of complex traits such as behavioral and cognitive phenotypes has typically pursued the identification of individual Quantitative Trait Loci (QTLs); this approach has the great appeal and promise of identifying Quantitative Trait Genes (QTGs). However, in spite of the increased resolution of advanced crosses such as the HS/Npt, identification of QTGs remains an elusive goal since, to date, most QTLs contain multiple genes and in many cases hundreds of genes. In summary, traditional QTL analysis often identifies numerous QTL regions with small-to-moderate effect size and each genomic location can contain numerous genes although the utilization of advanced heterogeneous stocks such as the HS/Npt partially mitigates this situation.
Given these observations, our strategy for analyzing the HS/Npt-derived genetic, behavioral and cognitive data was twofold. First, we employed a distance/similarity based multi-locus technique that offers high flexibility in the number of loci included in the analysis ([@B11]; [@B46]); this is further enhanced by machine learning techniques. Second, we focused on identifying biological pathways overrepresented in genes within significant loci by incorporating genomic annotation in the form of gene chromosomal locations and participation in functional groups.
Materials and Methods {#s1}
=====================
Animals
-------
HS/Npt is a reverse engineered outbred mouse population derived from the A/J, AKR/J, BALB/cJ, C3H/HeJ, C57BL/6J, CBA/J, DBA/2J, and LP/J strains as progenitors. The breeding colony consist of 48 families that are maintained by a circular breeding scheme that maximizes genetic heterogeneity. Three HS/Npt breeder mice of each family at generation 71, one male and two females, were provided by Dr. Robert Hitzemann, OHSU, and shipped to Colorado State University for breeding the experimental mice of the current study. Because of the limited number of breeders, matings were set up and litter generation was performed as follows. A first mating was set up which generated about 600 pups (cohort 1). When these animals reached 8--12 weeks of age they were shipped to BNL, Upton, NY, United States for irradiation. A second mating was then set up which generated about 1,200 pups (cohort 2) that were also shipped to BNL and similarly irradiated. Consequently, all of the mice in cohort 1 were from first litters, and those from cohort 2 were from second or third litters or first surviving litters from dam that had not previously had a surviving litter. Irradiation was either sham-irradiation or 0.4°Gy ^56^Fe (600 MeV/n), or 0.4°Gy ^28^Si (240 MeV/n), or 3°Gy of ^137^Cs gamma-rays at the NASA Space Research Laboratory (NSRL) building. While the shipping and irradiation conditions were the same between the cohorts, urine was collected from the second cohort 24 h after sham-irradiation or irradiation. Urine collection involved firmly holding the mouse over parafilm. In some cases, it was necessary to gently massage the mouse's abdomen. Restraint time was generally, but not always, less than a minute. Urine was not collected from the first cohort. The week following irradiation, the mice were shipped to Colorado State University. Three months later, the mice were weighed and tested for contextual (hippocampus-dependent) and cued (amygdala-dependent and hippocampus-independent) fear conditioning. Potentially due to the additional stressor of the urine collection in the second cohort, there was a significant cohort effect and the mean and variance of a majority of phenotypes differed in the two cohorts of animals. Therefore, we analyzed the two cohorts separately and subsequently compared and contrasted the resulting findings.
Fear Conditioning
-----------------
The mice were tested for fear conditioning using Med Associates NIR Video and automated analysis (Med Associates, St. Albans, VT, United States) utilizing Med Associates Video Freeze automated scoring system. Pavlovian fear conditioning is a versatile and well-understood method of assessing associative learning and memory. In this task, mice learn to associate a conditioned stimulus (CS, e.g., a tone) with an unconditioned stimulus (US, e.g., a foot shock). CS--US pairings are preceded by a short habituation period, from which a baseline measure of locomotor activity is measured. On day 1, training, the mice were placed inside a white LED lit (100 lux) fear conditioning chamber (Context A). Context A consists of a metal grid floor with gray and white walls. There was a 90-s baseline followed by five CS--US pairings. During acquisition, the 30-s tone (CS) (80 db, 2,800 Hz) co-terminated with 2-s foot shocks (0.7 mA) (US). The inter-tone interval (ITI) was 90 s. Motion during shock (arbitrary units from proprietary index) was measured to explore potential treatment-induced differences in response to the aversive stimulus. Percent time freezing during each subsequent ITI and tone presentation was measured to assess acquisition of the fear response. On day 2, mice were exposed first to Context A for 300 s and to a new environment, Context B, 4 h later. Context B consists of a smooth white plastic floor, with a "tented" black plastic ceiling and scented with a 10% isopropanol solution. There was a 90-s baseline and a 180-s tone. Freezing was defined as the absence of motion with the exception of respiration. The freezing response is a widely used indicator of a conditioned fear response ([@B2], [@B3]). Motion during shock (arbitrary units from proprietary index) was measured to evaluate potential treatment-induced differences in response to the aversive stimulus. Percent time freezing during each subsequent ITI and tone presentation was measured to assess acquisition of the fear response. On day 2, mice were first exposed to the training environment for 300 s and motion and freezing levels were analyzed. Subsequently, the mice were exposed to a new environment, consisting of a smooth white plastic floor, with a "tented" black plastic ceiling and scented with a 10% isopropanol solution. There was a 90-s baseline and a 180-s tone, i.e., a tone (CS), with a foot shock (US), and thereby come to fear the CS. Trained mice display this conditioned fear by ceasing all movement except for respiration in an attitude called "freezing." Training takes place in a light and sound attenuated chamber (termed the "conditioning chamber") equipped with a video camera. The type of training determines the brain regions involved in the learning and memory processes. The mice were tested for learning, memory, and extinction of hippocampus-dependent contextual fear conditioning using Med Associates NIR Video and automated analysis (Med Associates, St. Albans, VT, United States) utilizing Med Associates Video Freeze automated scoring system. Freezing was defined as the absence of motion with the exception of respiration. The freezing response is a widely used indicator of a conditioned fear response. On day 1 (training), each mouse was placed inside the enclosure with a house light (100 lux) on during the training trial. The test involved a 60-s baseline period in which the activity was recorded, followed by four tone--shock pairings. The tones were 80 dB at 2,800 Hz for 30 s, co-terminating with 2 s shocks (0.7 mA). Activity in the novel environment was defined as movement during the 60-s baseline period. The ITI was 90 s. Motion during shock (arbitrary units from proprietary index) was measured to evaluate potential treatment-induced differences in response to the aversive stimulus. Percent time freezing during each subsequent ITI and tone presentation was measured to assess acquisition of the fear response. On days 2 (day 1 of extinction) to 9 (day 8 of extinction), mice were first exposed to the training environment for 300 s and motion and freezing levels were analyzed. Subsequently, the mice were exposed to a new environment, consisting of a smooth white plastic floor, with a "tented" black plastic ceiling and scented with a 10% isopropanol solution. There was a 90-s baseline and a 180-s tone.
Genotyping
----------
DNA was prepared from tail snips (∼2 mm) collected at weaning. Each mouse was genotyped for 77.8 K SNP markers by GeneSeek (Lincoln, NE). We identified ∼40 K markers with sufficient genetic diversity (minor allele frequency above 5%). The mapping resolution was in the 200 kb to 2 Mb range for this HS mapping population.
Multivariate Distance-Based QTL Detection
-----------------------------------------
For the QTL detection, we used a modified version of multivariate distance matrix regression (MDMR), which relates *P* variables to *M* factors collected on *N* individuals, where *P* \>\> *N*. The most important feature of this procedure is the capacity to evaluate individual markers as well as groups of markers. In the present study, this feature was utilized to select groups of markers that collectively display the strongest association with the phenotype. MDMR analysis involves computing the distance between all pairs of individuals with respect to *P* variables of interest and constructing an *N* × *N* matrix whose elements reflect these distances. Permutation tests are used to test hypotheses and derive *p*-values that consider whether or not the *M* genetic factors (loci) can explain variation in the observed distances between and among the *N* individuals as reflected in the matrix of phenotype differences.
An essential step in the MDMR procedure involves selecting appropriate distance measures for each data modality -- SNPs for the genotype and composite measures for the behavioral and cognitive data. For both the genotype data and phenotype data, we utilized the Manhattan distance, assigning equal weight to each genotype and phenotype. The essential steps of this approach are outlined in **Figure [1](#F1){ref-type="fig"}**.
{#F1}
For the purpose of finding optimal groups of genetic markers, we further combined the MDMR procedure with a pattern detection algorithm known as forward selection. Starting with the most significant genetic marker, we included additional markers to the analysis, selecting them in decreasing order of significance as quantified by the Mantel correlation with the phenotype. For each additional marker, the genotype distance matrix was recomputed and correlated once more with the phenotype distance matrix. The typical outcome of this procedure consists of initial increases in the cumulative/group correlation, followed by a peak reaching a plateau, after which including additional markers starts to decrease the correlation (**Figure [3A](#F3){ref-type="fig"}**).
While the forward selection procedure is often very powerful, it has the potential drawback of overfitting the data. To guard against this possibility, we employed two procedures. First, as outlined above, for each marker we computed the individual correlation (with the phenotype), as well as the associated *p-*value which was obtained by a resampling procedure in the Mantel test. Only markers individually significant at a *p*-value \< 0.01 were included. Second, we randomized the data by shuffling the sample labels in the phenotype data and repeated the whole procedure, in effect producing an empirical distribution of results that can be obtained by pure chance. The real cumulative correlation value was combined with the empirical distribution of permutation-derived correlations and *Z-*scores were produced. This procedure facilitated selecting the number of markers that cumulatively produce robust and significant results.
Statistical Analyses
--------------------
Statistical analyses were performed in the R programming and analysis framework ([@B28]).
Results
=======
Summary and Interpretation of Behavioral and Cognitive Measurements
-------------------------------------------------------------------
An initial examination of the behavioral data identified numerous phenotypes that displayed high correlations (Pearson *r* \> 0.8 -- see **Supplementary Table [1](#SM1){ref-type="supplementary-material"}**). Most of the strong correlations were detected among related phenotypes. For example, correlations between the several "context (contextual fear memory)"-related phenotypes are in the range of 0.5--0.9. Similarly, high correlations were found when "cued" (cued fear memory) and "train" (learning day) phenotypes were compared within the same category. In contrast, we only detected small to moderate correlations (\<0.6) among phenotypes that capture distinct behavioral and cognitive measures (e.g., between "cued" and "context" phenotypes). In most cases, we detected very low correlations, as in between "shock" (learning) and "context" phenotypes. These observations justified the grouping of related phenotypes into six main categories: *context_pctfrze* (*percent freezing during the contextual fear memory test*), *context_avgmot* (*average motion during the contextual fear memory test*), *cued_pctfrze* (*percent freezing during the cued fear memory test*), *train_pctfrze* (*percent freezing during acquisition*/*learning of fear memory on the training day*), *shock_avgmot* (*average motion during the shocks*), and *train_avgmo* (*average motion during the training*/*learning session*). To perform subsequent analyses and in particular multi-locus QTL mapping, related individual behavioral and cognitive measures were combined in a vector and, for each vector, we constructed pairwise distance matrices between all individuals.
Cohort Effects in the Behavioral and Cognitive Measures
-------------------------------------------------------
Due to the large number of animals utilized in this study (∼1,800), the shipment to BNL and radiation and sham-irradiation there, and return shipment from BNL to CSU, behavioral and cognitive testing was performed in two cohorts, of ∼600 and ∼1,200 animals, respectively. For the second, but not first, cohort, there was urine collection 24 h after radiation or sham-irradiation. Due to the fact that the urine collection procedure and associated handling has the potential to serve as an additional environmental challenge and alter subsequent behavioral performance, we tested for the presence of differences in behavioral and cognitive measures between the two cohorts. For nearly all behavioral and cognitive measures we found significant differences in mean and variance \[see **Figure [2](#F2){ref-type="fig"}** for an example involving the percent freezing during training in the fear conditioning test (*train_pctfrze* phenotype)\]. This illustrates the sensitivity of the behavioral and cognitive measures assessed to alterations in the environment. Based on this result, subsequent analyses were performed independently for the two cohorts although we compared and contrasted quantitatively and qualitatively the findings from the two cohorts.
{#F2}
Detection of Genomic Locations and Marker Sets Associated With Behavioral and Cognitive Measures
------------------------------------------------------------------------------------------------
Behavioral and cognitive measures generally fall into the category of complex traits, which implies that they do not follow simple Mendelian models of inheritance and their genetic control is dispersed throughout the genome ([@B5]). Our analysis approach was selected based on this observation. Both phenotypes and genotypes were represented as multidimensional vectors; sample pairwise differences were computed between these vectors and correlated utilizing the Mantel procedure ([@B21]); this general procedure has been previously utilized in both genetic ([@B46]) and gene expression analyses ([@B39]). We proceeded in three steps. First, we applied the procedure independently to each marker--phenotype pair and derived Mantel statistics (correlations) and associated *p*-values. Second, we utilized the forward selection machine learning technique and we combined the most significant markers for each phenotype and derive cumulative Mantel statistics. Third, we utilized randomization procedures to evaluate the robustness of this approach.
Performance of the Distance-Based Multi-locus Genetic Association Procedure
---------------------------------------------------------------------------
As outlined above, our procedure starts from the most significant marker and expands the group of markers until either a peak is achieved or the individual marker significance *p*-value falls under 0.01. An example of the performance of this procedure is outlined in **Figure [3A](#F3){ref-type="fig"}** for the average movement during the contextual fear memory test (*context_avgmo*) phenotype. In this case, the performance (magnitude of the Mantel correlation) reached a plateau after 500 markers (thick black line in **Figure [3A](#F3){ref-type="fig"}**). We next repeated this procedure on randomized data (order of the samples in the phenotype datasets was scrambled). Examples of the resulting curves are illustrated in **Figure [3A](#F3){ref-type="fig"}** (thin gray lines). This procedure illustrates that peak performance in the original data (not randomized) was unlikely to be achieved by chance alone. We further quantified these results by constructing *Z*-statistics from the randomized and un-randomized data (**Figure [3B](#F3){ref-type="fig"}**); this revealed that as the number of markers grows beyond 2--300, *Z*-scores increase \>3 (*p* \< 0.0014) and eventually achieve levels \>5 (*p* \< 00001), is highly unlikely due to chance. Qualitatively similar results were obtained for all phenotypes (data not shown). We conclude that our procedure robustly detects groups of genetic markers that achieve a high correlation with the behavioral phenotypes. Importantly, both overall performance (correlation) and robustness are achieved when including larger number of genetic markers. Uncertainty remains about the influence of single markers, even when examining the marker achieving the greatest genome wide significance; this manifests in the data by sharp jumps in the significance and *Z*-score curves (**Figures [3A,B](#F3){ref-type="fig"}**).
{#F3}
Detection and Comparison of Peak Genotype--Phenotype Cumulative Correlations
----------------------------------------------------------------------------
We detected sets of markers that cumulatively raised to as high as Mantel correlation of 0.35 (*p* \< 0.001 based on 1,000 permutation tests) (*context_avgmo* phenotype -- see **Figure [3A](#F3){ref-type="fig"}**). The average motion during the contextual fear memory test (*context_avgmo*) phenotype had the highest correlation with the genotype in both cohorts, while the percent freezing during the contextual fear memory test (*context_pctfrze*) had the lowest correlation (**Figure [3C](#F3){ref-type="fig"}**); the rankings of "genetic correlation" for the six composite phenotypes were identical in the 600 and 1,200 cohorts, as can be seen in **Figure [3C](#F3){ref-type="fig"}**. However, there was a significant shift downward in the magnitude of correlations between the 600 and 1,200 groups, as indicated by the departure from the diagonal of the linear relationships in **Figure [3C](#F3){ref-type="fig"}**. This set of results illustrates that: (1) all behavioral and cognitive measures had a significant genetic component, albeit dispersed among numerous genomic locations; (2) changes in environmental conditions between the two cohorts resulted in differences in genetic components of behavioral and cognitive measures, as quantified by the Mantel correlation; and (3) the ranking of the strength of the genetic control over behavioral and cognitive performance was preserved across the two cohorts.
Environment Effects on the Genotype*--*Phenotype Correlations
-------------------------------------------------------------
Given the fact that the different testing environment in the two cohorts was associated with differences in the strength of genetic control over behavioral and cognitive performance, we inquired whether the identity of the significant genomic locations was altered as well. We utilized the Fisher exact test to quantify the level of overlap between sets of significant markers independently detected in each cohort. We found that for most behavioral and cognitive measures, the genomic locations that showed association with behavioral and cognitive performance shifted, with the average motion during training (*train_avgmo*) measure being the most affected (**Figure [3D](#F3){ref-type="fig"}**). We conclude that a change in the environmental conditions, which in our case was a recent intrusive urine collection, changed the identity of the genomic locations showing an association with certain behavioral and cognitive measures.
Radiation Effects on the Genotype--Phenotype Correlations
---------------------------------------------------------
Given the fact that shifts in environmental setting (urine testing) induced changes in phenotype--genotype correlations, we inquired whether radiation exposure has similar effects. This hypothesis was evaluated independently in the two cohorts. We separated the two main cohorts between the sham versus irradiated animals (combining all radiation types together). We then performed genetic scans independently in the irradiated and sham-irradiated animals. Comparing the identity of the sets of genetic markers detected revealed large shifts in their composition (**Figure [4](#F4){ref-type="fig"}**). Lack of overlap was quantified by the magnitude of the Fisher test odds ratio, with values below one being no different from chance. For nearly all behavioral and cognitive measures, the identity of the significant markers changed dramatically to the point that there was no overlap beyond what can be achieved by chance. It is important to note that the one phenotype that showed concordance between sham-irradiated and irradiated animals, average motion during training (*train_avgmo*) was the same in the two cohorts. We conclude that radiation exposure has the potential to alter phenotype--genotype correlations.
{#F4}
Annotation of Genes Sets Associated With Selected Loci
------------------------------------------------------
We next inquired whether biological pathways or other functional groups are enriched in genes within the significant genomic locations. For each behavioral phenotype, we intersected the significant markers detected in each cohort. Next, we collected genes within 10^6^ base pairs of any of these common markers. We recognize that these gene sets might include some genes without direct association with the phenotype. However, since genes in chromosomal physical proximity often participate in common biological functions, this wide inclusion criteria can help facilitate the detection of relevant functional groups. Gene sets were evaluated for enrichment in gene members in Gene Ontology or KEGG pathways. We detected significant enrichment for a majority of phenotype/irradiation type combinations (**Supplementary Table [2](#SM2){ref-type="supplementary-material"}**). For example, the average motion during training (*train_avgmo*) phenotype following sham-irradiation showed enrichment in "response to pheromone" and "response to chemical stimulus" GO functional categories. For the percent freezing during the contextual fear memory test (*context_pctfrze*) with Gamma radiation, we detected "regulation of proteolysis" and "negative regulation of endopeptidase activity."
The phenotype-irradiation combinations with the most abundant annotations were context_pctfrze under ^56^Fe irradiation and train_avgmo under ^28^Si irradiation (**Supplementary Table [2](#SM2){ref-type="supplementary-material"}**). These two phenotypes also displayed the strongest genetic basis in both batches (**Figure [3C](#F3){ref-type="fig"}**). Importantly, ^56^Fe and ^28^Si are two radiation types uniquely encountered in space flight. Given these findings, an in-depth examination of the specific GO and KEGG pathways uncovered by the genetic analysis is warranted.
Among the annotations associated with train_avgmo under ^28^Si, we have the PPAR_SIGNALING_PATHWAY (*p* \< 2.2 × 10^-9^). This pathway is associated with regulation of peroxisome proliferation and adipocyte differentiation. However, emerging evidence suggests that within the adult mouse brain PPAR is highly expressed primarily not only in the hypothalamus, but also in the neocortex, the olfactory bulb, the organ of the vasculosum of the lamina terminalis (VOLT), and the subfornical organ ([@B20]). This study also found that within the hypothalamus, suprachiasmatic nucleus displays moderate levels of PPAR which is upregulated by fasting. Neurogenetic imaging has also revealed that genetic variability in PPAR is related to cerebral connectivity in preterm infants ([@B16]). Taken together, these findings provide indications of neurobehavioral roles of the PPAR pathway.
Among the annotation categories associated with train_avgmo under ^28^Si, we also find KEGG_MM_FATTY_ACID_METABOLISM (*p* \< 5.7 × 10^-8^). There is extensive literature on the effect of ionizing radiation on fatty acid metabolism ([@B1]; [@B9]; [@B23]; [@B34]). For KEGG_MM_RETINOL_METABOLISM (*p* \< 1.6 × 10^-6^), there is evidence that retinoids reduce cellular apoptosis following irradiation ([@B44]); other studies have also found involvement of retinol in molecular or cellular responses to irradiation ([@B36]).
Among the annotation categories associated with context_pctfrze under ^56^Fe irradiation, we find GO_MF_MM_ATPASE_ACTIVITY (*p* \< 9.7 × 10^-7^). The mitochondrial activity of ATPASE has been shown to be affected by radiation in female mice ([@B15]). Additionally, there have been early studies indicating changes in mouse serum esterase concentrations following ionizing radiation ([@B13]); we also find several esterase-related annotations associated with context_pctfrze under ^56^Fe irradiation (**Supplementary Table [2](#SM2){ref-type="supplementary-material"}**).
Discussion
==========
Our main finding is that phenotype/genotype correlations are plastic and can shift depending on the external environment, and in particular following ionizing radiation. Recent studies have revealed that a significant proportion of QTLs reported in the literature fail to replicate in subsequent studies ([@B38]). Given these findings, we leveraged the presence of two distinct cohorts and exposure to different types of radiation to not only detect significant QTL groups, but more importantly to evaluate their reproducibility across cohorts and radiation types. Our work therefore contributes to the understanding of causes of the varying levels of reproducibility affecting QTL studies.
This observation has important implications regarding potential effects of space radiation on behavioral and cognitive performance. Our results suggest that complex behaviors and their genetic basis have to be studied under realistic conditions, which in our case means irradiated animals. Behavioral QTLs from non-irradiated animals will have limited predictive power once the animals are subjected to irradiation.
We find that alterations in environmental conditions and environmental challenges other than irradiation also have the potential to alter behavioral and cognitive performance and their genetic basis, an observation consistent with previous findings ([@B8]; [@B6]). Stressful events (such as intrusive urine collection) have a strong effect on the actual behavioral and cognitive performance ([@B18]). It is important to note that there is a qualitative distinction between the effects of stress versus the effects of ionizing radiation. Stressful events lead to actual shift in behavioral and cognitive performance, but do not completely alter the genotype/phenotype correlations (odds ratios ≫1 in **Figure [3D](#F3){ref-type="fig"}**). In contrast, ionizing radiation does not directly alter behavioral and cognitive performance but has profound effects on the genotype/phenotype correlations (**Figure [4](#F4){ref-type="fig"}**). It is striking that in the larger cohort of 1,200 mice in which there was urine collection 24 h following radiation exposure, the mean percent freezing during training, a measure of learning, as well as the variability in this measure in individual mice was much higher than that in smaller cohort of 600 mice in which there was no urine collection (**Figure [2A](#F2){ref-type="fig"}**). Striking cohort effects were also seen in **Figure [4](#F4){ref-type="fig"}**. There was more overlap between the genetic effect on activity during the contextual fear memory test. This was detected the day following training and between irradiated and sham-irradiated mice in the cohort of 1,200 mice. As noted above, in the 1,200 mice cohort there was urine collection 24 h following radiation exposure. In contrast, in the cohort of 600 mice cohort there was no urine collection. In other words, there appear long-term effects of urine collection that diminishes the ability to detect genetic effects of irradiation on behavioral performance in a hippocampus-dependent fear memory test. No such difference was seen for hippocampus-independent cued fear memory. These data indicate that the hippocampus might be particularly sensitive to long-term effects of environmental stressors that might affect radiation effects. However, the cohort effects were also seen when activity during fear conditioning training was compared. The odds ratio for activity during the training session was much higher in the cohort of 1,200 than in the cohort of 600 mice. These data suggest that the urine collection causes a dramatic increase in overlap between the genetic effects on activity during the session in which the mice learned this task. As astronauts during space missions encounter environmental challenges in addition to radiation exposures, increased efforts are warranted to determine how environmental challenges other than radiation might modulate long-term effects of space radiation on the brain.
Numerous studies have encountered and commented on the varying levels of reproducibility of genetic association studies ([@B40]; [@B19]; [@B7]; [@B38]). Benefiting from moderate sample sizes over each batch/radiation type combination, our study addresses the issue of reproducibility in a systematic manner. It is important to note that since we do not focus on individual markers or genes but on sets that collectively achieve high correlation with the phenotype, this confers an increased level of robustness to the findings.
Within behavioral genetics there is wide recognition that Mendelian mutations leading to large shifts in behavioral and cognitive performance are extremely rare ([@B12]). Given that many behavioral and cognitive measures are still clearly under genetic control, this observation points to the possibility that combinations of genetic loci of small-to-moderate effects collectively control behavioral and cognitive performance. The multi-locus analysis methodology employed here is well suited to perform analyses under these conditions. We note that we utilized a simple identity by state (IBS) distance matrix ([@B46]); alternatives include weighing each marker by its allele frequency or by functional role.
Our annotation approach closely follows ([@B4]) and is based on formulation of hypotheses that do not focus on single genes but instead consider groups of genes grouped either based on functional roles (pathways and ontologies) or based on location (vicinity to a marker of moderate significance). The critical assumption of this approach is that the two gene groups are constructed independently, an assumption satisfied in our study. As noted in [@B4], when the relevant functional unit is a group of genes (a pathway), one can expect that the genetic markers falling within these genes will have increased individual significance scores, even though these will be within the stochastic noise range and far from achieving the genome-wide significance of top results in a GWAS study. However, *when considered as a group*, their collective significance will clearly point to the implication of the relevant pathway.
We examined closely some of the pathways associated with the two behavioral phenotypes with the strongest genetic component, context_pctfrze under ^56^Fe irradiation and train_avgmo under ^28^Si irradiation. There are several conclusions that we draw from this analysis. First, at the level of annotations and pathways, there is no overlap in the categories detected for different radiation types and behavioral and cognitive phenotypes. This finding suggests that distinct radiation types lead to the "engagement" of different biological mechanisms, pathways, and ontologies. This observation is concordant with previous experimental work showing that irradiation parameters induce different types of DNA damage ([@B27]), and also with modeling work aligning distinct irradiation features with distinct patterns of DNA damage.
At the conceptual level, our findings suggest the following model of interaction between irradiation, genetics, and behavioral and cognitive patterns. Ionizing radiation engenders effects on behavioral and cognitive performance, and internal compensatory mechanisms mitigate these effects. However, genetics plays a part as to which compensatory mechanisms are engaged in each individual animal. Individual genetic variability interacts with variation in the irradiation type, resulting in heterogeneity in the identity of individual genetic loci and of the functional pathways affected.
Ethics Statement
================
The experiments were approved by the IACUC committees from BNL and CSU.
Author Contributions
====================
OI, JB, MW, and JR designed the experiments and wrote the manuscript. SB, RO, MD, BS, ME, TM, CF, and EE acquired and analyzed the data and wrote the manuscript.
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.
**Funding.** This work was supported by NASA NNX12AB54G.
We would like to acknowledge Dr. Robert Hitzemann, Department of Behavioral Neuroscience, OHSU, for providing the HS/Npt mice for breeding the mice used in this study. We would also like to acknowledge the scientists and staff of the BNL Medical Department, and especially wish to thank Peter Guida, Adam Rusek, and other NSRL physics staffs for their invaluable help, and the BNL Laboratory Animal Facility for their invaluable assistance.
Supplementary Material
======================
The Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fgene.2018.00404/full#supplementary-material>
######
Click here for additional data file.
######
Click here for additional data file.
[^1]: Edited by: Dimitrios Avramopoulos, Johns Hopkins University, United States
[^2]: Reviewed by: Alex Hatzimanolis, National and Kapodistrian University of Athens, Greece; Peristera Paschou, Purdue University, United States
[^3]: ^†^Deceased
[^4]: This article was submitted to Behavioral and Psychiatric Genetics, a section of the journal Frontiers in Genetics
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
RNA interference (RNAi) is a highly conserved biological phenomenon in all eukaryotes, including renal cells. In the late 1990s, due to the development of molecular biology and genetics, the biological understanding of RNA evolved from simply an intermediate between DNA and protein to a dynamic and versatile regulator that functions in genes and cells in all living organisms. In 1998, as a milestone, Fire *et al*. injected a few molecules of double-stranded RNA (dsRNA) into *Caenorhabditis elegans* and found that dsRNAs could specifically interfere with the protein expression of an endogenous gene \[[@CR1]\]. This molecule was named small interfering RNA (siRNA) that mediates RNAi \[[@CR2]-[@CR4]\]. siRNA is able to recognize and degrade a homologous host mRNA. Therefore, the gene from which the mRNA is transcribed is silenced, which is referred to as post-transcriptional gene silencing \[[@CR5],[@CR6]\].
Although RNAi naturally exists, synthetic artificial siRNA exerts similar effects as natural endogenous microRNA (miRNA). Both sense and antisense strands of siRNA can be synthesized separately and annealed to form double stranded siRNA duplexes *in vitro.* After the siRNA is delivered into the cytoplasm, the artificial siRNA silences the target gene using similar biological processes as endogenous miRNA. Since the introduction of 21-nucleotide artificial siRNAs that triggered gene silencing in mammalian cells \[[@CR7]\], synthetic siRNA has generated much interest in biomedical research, in which the kidney is one of important key players. siRNA as a strategic molecule has been highly expected in the field of innovative therapy. Because siRNA is highly efficient at gene silencing, it is possible to develop specific siRNA-based drugs that could target any genes, including those that have no known pharmacological antagonists or inhibitors. Different types of synthetic siRNA have been tested for their efficacy in various disease models, including cancer \[[@CR8],[@CR9]\], autoimmune disorders \[[@CR10]\], cardiovascular injuries \[[@CR11],[@CR12]\], and organ transplantation \[[@CR13],[@CR14]\], including native and transplanted kidney injuries \[[@CR15]\].
As siRNA is a posttranscriptional regulator, it must first be absorbed into the target cells. Therefore, the kidney could be an excellent target organ for siRNA therapy because it benefits from rapid, vast blood flow physically and subsequent glomerular filtration and tubular absorption. In fact, systemic administration of siRNA leads to rapid uptake by the kidney, yielding a significant decrease of target protein expression \[[@CR15]\]. Consequently, RNAi by siRNA has advantages for the treatment of renal diseases due to the unique urological system. In addition, the preservation of donor kidneys before transplantation also provides a suitable time window for the intervention of siRNA. Therefore, we performed a series of experiments using naked caspase-3 siRNA to investigate its efficacy, off-target effects and compensative responses in *in vitro*, *ex vivo* and *in vivo* models of transplant-related renal injuries.
In this review, we highlighted the design and delivery of siRNA, its therapeutic effects, off-target responses and systematic compensations, as well as potential challenges, with a focus on kidney diseases, including ongoing clinical trials.
Current principle of siRNA design {#Sec2}
=================================
The design of potent siRNAs has been greatly improved over the past decade. The basic criteria for choosing siRNAs includes the consideration of thermodynamic stability, internal repeats, immunostimulatory motifs, such as GC content, secondary structure, base preference at specific positions in the sense strand, and appropriate length \[[@CR16]\].
Chemical modifications significantly enhance the stability and uptake of naked siRNAs. Importantly, siRNAs can be directly modified without crippling the silencing ability. Chemical modifications have been rigorously investigated for virtually every part of siRNA molecules, from the termini and backbone to the sugars and bases, with the goal of engineering siRNA to prolong half-life and increase cellular uptake. The most common chemical modification involves modifying the sugar moiety. For example, the incorporation of 2'-fluoro (2'-F), −O methyl, −halogen, −amine, or -deoxy can significantly increase the stability of siRNA in serum. Locked nucleic acid (LNA) has been also applied to modify siRNA. The commonly used LNA contains a methylene bridge connecting the 2'-oxygen with the 4'-carbon of the ribose ring. This bridge locks the ribose ring in the 3'-endo conformation characteristic of RNA \[[@CR17]\]. Additionally, recent studies, including ours \[[@CR18]\], have proven the efficacy of LNA modified siRNA in terms of prolonged half-life in serum, but without detectable adverse effects, suggesting that the natural RNAi machinery could accommodate a certain degree of alterations in the chemical structure of siRNAs \[[@CR19]\].
siRNA delivery {#Sec3}
==============
The biggest obstacle faced by siRNA therapies is the *in vivo* delivery of genetic materials. The virus-based delivery system, while efficient, may be fatally flawed due to raised safety concerns, such as inducing mutations and triggering immunogenic and inflammatory responses \[[@CR20]\]. Therefore, extensive research had been performed to develop efficacious non-viral delivery systems, including direct chemical modification of siRNA (as described above) and/or optimization of delivery materials, such as liposome formulation, nanoparticle conjugation and antibodies that target cellular moieties \[[@CR20]\].
To date, studies on synthetic siRNA therapy have been performed in a variety of cell culture and rodent models \[[@CR21]\] that produced exciting results and were cost effective, but failed to faithfully mimic human diseases. Therefore, large animal models, such as porcine models, are indispensable to compensate for the limitations of rodent models due to their greater similarity to human beings. The investigations on siRNA conducted in our laboratory have reflected this trend in the field \[[@CR14],[@CR18],[@CR22]\].
In vitro delivery {#Sec4}
-----------------
Cell culture is an important model for investigating the cellular and molecular mechanisms of diseases. Lentivirus vectors and liposomes are widely applied for the transduction of siRNA into different types of cultured renal cells. For example, silencing *ccr2* using its siRNA delivered by a lentivirus significantly ameliorated MCP-1 induced podocyte apoptosis under diabetic conditions \[[@CR23]\]. Similarly, degradation of FoxO1 by transfecting its siRNA via lentiviral vectors overrode the limited cell cycle and stimulated proliferation of glomerular mesangial cells \[[@CR24]\]. In our previous study, we transfected porcine proximal tubular cells (LLC-PK1) with synthetic caspase-3 siRNA using a cationic lipid-based transfection reagent. The caspase-3 siRNA inhibited apoptosis and inflammation in LLC-PK1 cells that were subjected to hydrogen peroxide stimulation \[[@CR25]\].
Ex vivo delivery {#Sec5}
----------------
In addition to *in vitro* delivery of siRNA, *ex vivo*/*in vivo* siRNA delivery to target organs is an indispensable step before its clinical application. If it were directly delivered into the kidneys, siRNA could obtain higher local concentrations, which would result in improved gene silencing efficacy. During kidney transplantation, *ex vivo* local delivery of siRNA into the donor kidney is feasible because it could be facilitated by the unique structure of the kidney and the characteristics of kidney transplantation. Recently, we utilized an *ex vivo* isolated porcine kidney reperfusion system to assess the efficacy of naked caspase-3 siRNA. The caspase-3 siRNA was directly infused into the renal artery (locally) and autologous blood perfusate (mimic systemic delivery) before 24-hour cold storage (CS), followed by a further reperfusion for 3 hours. The results demonstrated that the caspase-3 siRNA improved ischemic reperfusion (IR) injury with reduced caspase-3 expression and apoptosis, better renal oxygenation and acid--base homeostasis \[[@CR22]\]. These promising proof-of-principle observations provide valuable guidance for further development before siRNA clinical practice.
In vivo local delivery {#Sec6}
----------------------
Based on the anatomical and physiological characteristics of the kidney, local delivery can be achieved through several routes: (1) renal artery: first targeting the glomeruli or tubules \[[@CR26],[@CR27]\]; (2) renal vein: predominately targeting tubulointerstitium \[[@CR28]\]; (3) intra urethral: administered into the renal pelvis and interstitium \[[@CR29]\]; and (4) sub-capsular administration: achieves intraparenchymal silencing \[[@CR30]\]. Due to the rich blood flow through the glomeruli, siRNA injection via the renal artery followed by electroporation could silence specific genes in the glomeruli, such as TGF-β 1, which subsequently ameliorates matrix expansion in an experimental glomerulonephritis model \[[@CR26]\]. However, acute kidney injury (AKI), such as IR injury, is usually characterized by tubular apoptosis and inflammatory infiltration in the tubulointerstitium, which should be suitable for local siRNA delivery via any above mentioned method.
It has been revealed that an injection of a single-dose Fas siRNA through the renal vein post ischemia provided a survival advantage in a murine IR model, which was due to the anti-apoptosis and anti-inflammation effects of the Fas siRNA \[[@CR28]\]. Unilateral ureteral obstruction (UUO) is a well-established model for tubulointerstitial fibrosis. Xia et al. injected the siRNA of heat shock protein 47 once via the ureter at the time of UUO preparation, leading to significantly reduced fibrosis-related protein expression and a remarkable alleviation of the accompanying interstitial fibrosis \[[@CR29]\]. Sub-capsular administration is still used in some experiments due to its unique advantages, although it requires an invasive procedure and has limitations in clinical practice. Cuevas et al. reported that an infusion of DJ-1 (an antioxidant) specific siRNA into the sub-capsule silenced DJ-1 expression in the renal cortex and increased ROS production \[[@CR30]\].
In vivo systemic delivery {#Sec7}
-------------------------
Systemic delivery is a common and convenient clinical practice, although current clinical trials using siRNAs are almost directly administered to the target site, such as the nostril, eye and lung, thereby avoiding the complexity of systemic delivery \[[@CR31]\]. The most common method of systemic siRNA delivery is a hydrodynamic intravenous injection with hydraulic pressure to assist siRNA cell entry. However, the pharmacokinetic metabolism of siRNA is more complicated during systemic delivery because siRNAs can be rapidly degraded by nucleases in the serum and cleared by the kidney and liver. To enhance the *in vivo* efficacy of siRNA treatment, a variety of approaches have been attempted for both siRNA itself and delivery techniques \[[@CR30]-[@CR36]\], as mentioned above and discussed later.
Due to its anatomical and physiological characteristics, the kidney is the most preferable target organ of systemic siRNA administration. siRNA access to the kidney is thought to be dependent on the filtration and reabsorption functions of the kidney. Proximal tubule cells (PTCs) are the primary site for rapid and extensive endocytic uptake of siRNA within the kidney following glomerular filtration. In an AKI model, naked synthetic siRNA targeting p53 that was intravenously injected 4 hours after renal ischemic injury significantly reduced upregulated p53 expression and protected both the PTCs and kidneys \[[@CR37]\]. In another study performed by Zheng et al., siRNA was systematically injected to target complement 3 (C3) and caspase-3 in a murine renal IR injury model. The results showed that the level of serum creatinine and blood urea nitrogen was significantly decreased in the siRNA-treated mice \[[@CR38]\]. These studies highlighted the potential feasibility of systemically delivered siRNA for the treatment of kidney diseases in future clinical practice.
siRNA therapeutic efficacy {#Sec8}
==========================
To date, siRNA therapy has been successfully applied for acute and chronic kidney diseases, as well as renal tumors.
Acute kidney injury {#Sec9}
-------------------
AKI is a common complication of hospitalization that has a mortality rate as high as 30-50% \[[@CR39],[@CR40]\]. IR injury is the primary cause of AKI, particularly during kidney transplantation, in which the kidney is exposed to hypoxia and experiences a series of oxidative, inflammatory and apoptotic responses \[[@CR41],[@CR42]\]. Consequently, specific siRNAs targeting critical molecules that are involved in the processes of oxidation, inflammation and apoptosis have been developed.
Caspase-3, which mediates apoptosis and inflammation, is upregulated by IR injury. Multiple pharmacological interventions against caspase-3, including enzyme inhibitors and genetic modification, have been investigated. In recent years, our group studied the delivery and efficacy of caspase-3 siRNA in *in vitro*, *ex vivo* and *in vivo* porcine models. The synthetic caspase-3 siRNA was initially tested in porcine PTCs, with or without hydrogen peroxide (H~2~O~2~) stimulation. Apoptotic cells and activated IL-1β protein expression were significantly reduced by the caspase-3 siRNA, with improved cell viability \[[@CR25]\]. This outcome led to siRNA application in an isolated organ perfusion system, as described above, and the efficacy of caspase-3 siRNA was further proven \[[@CR22]\].
We then used naked caspase-3 siRNA in a porcine kidney auto-transplant model for the first time. The left kidney was retrieved from mini pigs and was infused with University of Wisconsin solution, with or without 0.3 mg naked caspase-3 siRNA, via the renal artery, which was followed by renal artery and renal vein clamping for 24-hour cold storage (CS, mimicking donor kidney preservation before transportation in clinic). After right nephrectomy, the left kidney was auto-transplanted into the right nephridial pit for 48 hours without systemic siRNA treatment (Figure [1](#Fig1){ref-type="fig"}). The expression of caspase-3 mRNA and active caspase-3 protein, as well as its precursor, was downregulated by siRNA in the post-CS kidney. In the siRNA preserved post-transplant kidney, however, caspase-3 precursor was further decreased while caspase-3 mRNA and its activated subunits were upregulated, which resulted in increased apoptosis and inflammation. This study indicated that the naked caspase-3 siRNA was effective for cold preservation but was not effective at protecting post-transplant kidneys, which may be due to systemic compensative responses overcoming local effects. Therefore, to overcome the systemic response and prolong the therapeutic time window, we subsequently utilized a novel, serum stable caspase-3 siRNA, both locally as before and systemically via a pre-transplantation intravenous injection, and observed the animals for up to 2 weeks post-transplantation. The effectiveness of the novel caspase-3 siRNA was confirmed by downregulated caspase-3 mRNA and protein in the post-CS and/or post-transplant kidneys, as well as reduced apoptosis and inflammation. More importantly, renal function, associated with active caspase-3, HMGB1, apoptosis, inflammation and tubulointerstitial damage, was improved by this novel, serum stable caspase-3 siRNA \[[@CR18]\].Figure 1**Schematic drawing showed a series of studies using caspase-3 siRNA.** The caspase-3 siRNA was first used to protect porcine renal tubular epithelia cells against hydrogen peroxide-induced injury. The renoprotection of naked caspase-3 siRNA with the same sequences was further validated in a porcine *ex vivo* isolated reperfusion model, then shown that the siRNA was effective for cold preservation, but not in auto-transplanted kidneys without systematic siRNA treatment. Finally, the modified siRNA of caspase-3 via locked nucleic acid stabilized the siRNA in serum, and significantly protected auto-transplanted kidneys.
p53, another pivotal protein in the apoptotic pathway, has been identified as a mediator of transcriptional responses to IR injury \[[@CR43]-[@CR45]\]. Molitoris et al. revealed that intravenously injected p53 siRNA attenuated ischemic and cisplatin-induced AKI \[[@CR37]\]. Fujino et al. also tested the efficacy of transarterial administration of siRNA that targeted p53. p53 siRNA injected into the left renal artery immediately after ischemia improved tubular injury and downregulated GSK-3β expression \[[@CR46]\]. In a diabetic mouse model, p53 inhibition by siRNA also reduced ischemic AKI \[[@CR47]\].
Silencing other important transcription factors or immunity related receptors using siRNAs have also been studied. Renal IR injury and inflammation are related to postsurgical healing and both processes can be influenced by Toll-like receptor (TLR) signals. Effective TLR9 silencing by siRNA decreases renal cell apoptosis, mitigates AKI severity, and increases the mice survival \[[@CR48]\]. NF-κB, a pro-inflammatory transcription factor induced by TLR and other signals, plays a key role in AKI. NF-κB activation depends on the activation of the inhibitor of κB kinase β (IKKβ). Wan et al. demonstrated that silencing IKKβ using siRNA diminished inflammation and protected the kidneys against IR injury \[[@CR27]\]*.* These studies clearly demonstrate the therapeutic potential of siRNA-induced silencing of key AKI mediators, which are activated and involved in the pathways of apoptosis, inflammation, immunity, etc.
Chronic renal disease {#Sec10}
---------------------
Glomerulonephritis, resulting from multiple factors, is the most common primary disease leading to chronic kidney dysfunction. Interventions based on siRNA therapy for glomerulonephritis are promising, although a tissue-specific regimen has not been established. Shimizu et al. utilized the electrostatic interactions between positively charged nanocarriers, polyethylene glycol-poly(l-lysine)-polymers, with negatively charged siRNA to facilitate delivery. MAPK1 siRNA/nanocarrier complex used in a model of glomerulonephritis demonstrated that MAPK1 suppression remarkably improved kidney function, reduced proteinuria, and ameliorated glomerular sclerosis. The expression of the profibrotic marker TGF-β was also significantly decreased after MAPK1 siRNA therapy and glomerulonephritis progression was prevented \[[@CR49]\]. However, in contrast to AKI treatment, repeated siRNA administration will be necessary to achieve a therapeutic effect in chronic conditions. This necessity leads to some concerns for siRNA therapy, such as the medical cost and potential long-term side effects, which will be discussed later.
Renal cell carcinoma {#Sec11}
--------------------
Advanced renal cell carcinoma (RCC) is highly resistant to standard chemotherapy regimens, resulting in a 5-year survival rate of only 10% for those with stage IV disease \[[@CR50]\]. Gene therapy provides an efficient method for targeting the specific genes involved in RCC pathogenesis. For example, the HuR protein is a nucleocytoplasmic protein that plays an important role in the regulation of mRNA stability \[[@CR51]\]. A recent study examined the effects of siRNA-mediated HuR gene silencing in multiple RCC cell lines *in vitro* and demonstrated a 60% reduction in tumor cell growth compared with control cells. These findings were then successfully replicated *in vivo* using tumor-bearing mice \[[@CR52]\]. Shang *et, al.* further demonstrated the potential mechanism involved in the proliferative ability of RCC cells using siRNA technology \[[@CR53]\]. Other siRNA experiments have revealed that histone H3 acetylation was responsible for preserving drug sensitivity in RCCs, which indicates the important role of epigenetics in RCC \[[@CR54]\].
Nevertheless, silencing a single gene may cause tumor self-regulation and the development of drug resistance. Cancers, as well as other angiogenesis-related diseases, often result from the overexpression of multiple endogenous and exogenous pathogenic genes. One strategy for overcoming these challenges is using a cocktail therapy that combines multiple siRNAs targeting multiple disease-causing genes. Indeed, siRNA cocktails have demonstrated better anti-tumor potency compared with siRNAs that target a single gene \[[@CR55],[@CR56]\] (Table [1](#Tab1){ref-type="table"}).Table 1**Studies on kidney diseases using siRNAsInvestigatorsModelModificationTransfectionDeliveryTarget genePathwayReference**Yang et al.Porcine LLC-PK1-Cationic lipidIn vitroCaspase-3Apoptosis\[[@CR25]\]Yang et al.Porcine-UW solutionEx vivo, renal artery injection and autologous blood perfusateCaspase-3Apoptosis\[[@CR22]\]Yang et al.Porcine-UW solutionIn vivo, renal artery injectionCaspase-3Apoptosis\[[@CR14]\]Yang et al.PorcineLNAUW solutionIn vivo, renal artery injection and systemically iv.Caspase-3Apoptosis\[[@CR18]\]Liu et al.Mice-PlasmidIn vivo, hydrodynamic injection iv.TLR9Innate immunity\[[@CR48]\]Hamar et al.Mice2'-O-ACE-RNA phosphoramiditesPhosphate buffered salineIn vivo, hydrodynamic injection iv.FasApoptosis\[[@CR28]\]Wan et al.Rats-Phosphate buffered salineIn vivo, renal artery injectionIKKβInflammation\[[@CR27]\]Xia et al.Mice-Cationized gelatin microspheresIn vivo, injected via ureterHSP47collagen-producing, fibrosis\[[@CR29]\]Molitoris et al.Rats2'O-methylationLipofectamine 2000In vivo, iv.p53Apoptosis\[[@CR37]\]Zheng et al.Mice-Lipofectamine 2000In vivo, hydrodynamic injection iv.Complement 3 and caspases-3Complement and apoptosis\[[@CR38]\]Shimizu et al.Mice-Nanocarrier complexationIn vivo, ip.MAPK1Immunity\[[@CR49]\]Shang et al.ACHN, A498 RCC cells-Lipofectamine 2000In vitroHIF-1α, HIF-2αTumorigenesis\[[@CR53]\]Juengel et al.Caki-1 RCC cells-Transfection reagent supplied by QiagenIn vitroHDAC1, HDAC2Acetylation\[[@CR54]\]Fujino et al.Mice-Cationic lipidIn vivo, renal artery injectionp53Apoptosis\[[@CR46]\]UW: University of Wisconsin; TLR: toll-like receptor; siSTABLE: stability-enhanced siRNA; RCC: renal cell carcinoma; MAPK1: mitogen-activated protein kinase 1; HIF: hypoxia induced factor; HDAC: histone deacetylases.
Off-target side effects/toxicities of siRNA {#Sec12}
===========================================
Off-target effects are one of the major obstacles for siRNA therapy. The induction of various side effects may be caused by unexpected perturbations between RNAi molecules and cellular components. The off-target effects of siRNA were first reported by Jackson and colleagues in 2003 \[[@CR57]\]. Broadly speaking, off-target effects can be siRNA specific or non-specific. The former are caused by limited siRNA complementarity to non-targeted mRNAs. The latter, resulting in immune- and toxicity-related responses, are due to the construction of the siRNA sequence, its modification or the delivery vehicle.
The off-target effects associated with siRNA delivery fall into three broad categories: (1) microRNA-like off-target effects, referring to siRNA-induced sequence-dependent regulation of unintended transcripts through partial sequence complementarity to their 3′UTRs; (2) inflammatory responses through the activation of TLR triggered by siRNAs and/or delivery vehicles (such as cationic lipids and viruses); and (3) widespread effects on microRNA processing and function through the saturation of the endogenous RNAi machinery by exogenous siRNAs \[[@CR58],[@CR59]\].
MicroRNA-like off-target effects {#Sec13}
--------------------------------
The siRNAs and microRNAs share similar machinery downstream of their initial processing. Using several different siRNAs targeting the same gene, microarray profiling showed that each siRNA produced a unique, sequence-dependent signature. Sequence analysis of off-target transcripts revealed that the 3' UTR regions of these transcripts were complementary to the 5' end of the transfected siRNA guide strand \[[@CR57]\]. It is now understood that for the off-targeting mechanism to occur, a perfect complementation between the nucleotide positions 2--7 or 2--8 (seed region) of the antisense strand and the 3'UTR of the transcript is necessary \[[@CR58],[@CR60]\]. Base mismatches in the 5' end of an siRNA guide strand reduced silencing of the original set of off-target transcripts, but introduced a new set of off-target transcripts with 3' UTRs that were complementary to the mismatched guide strand \[[@CR58]\].
RNAi regulation by miRNAs involves partial complementarity between the target RNA and miRNA. Because miRNAs cause gene silencing through mRNA degradation and translation inhibition, the siRNA mediated off-target effects may also be acting at two levels. For this reason, there should be greater emphasis on improving siRNA design, as well as monitoring gene and protein levels following RNAi therapy to account for any off-target effects.
Recognition/Stimulation by the innate immune system {#Sec14}
---------------------------------------------------
The recognition and stimulation of the immune system is a nonspecific off-target effect of siRNA therapy. The RNA-sensing pattern recognition receptors (PRRs), localized in endosomes, are the most important components of the innate immune system. The responses of PRRs to siRNAs are either TLR-mediated or non-TLR-mediated. The PRR responses are also associated with siRNA sequence specific side effects and have recently attracted lots of attentions from researchers \[[@CR61]\]. RNA-sensing TLRs (TLR3 and TLR7) are predominantly located intracellularly and recognize nucleic acids released from invading pathogens. The non-TLR-mediated innate immune responses triggered by siRNA binding are linked to RNA-regulated expression of protein kinase (PKR) and retinoic acid inducible gene 1 (RIG1), which further induce caspase-3 and NF-κB expression, respectively. The activation of PRRs generates excessive cytokine release and subsequent inflammation \[[@CR62]\].
Based on this second type of off-target RNAi effects, our group further investigated the mechanism of how short-acting caspase-3 siRNA impaired post-transplanted kidneys. The results suggested that the amplified inflammatory responses in caspase-3 siRNA preserved auto-transplant kidneys were associated with TLR3, TLR7 and PKR activation, which may be due to systemic compensative responses, although persistent actions initiated by short-acting caspase-3 siRNA cannot be completely excluded \[[@CR63]\]. Other studies have also indicated that the horseshoe-like structure of TLR3 facilitates dsRNA recognition \[[@CR64],[@CR65]\]. Interactions between TLR3 and dsRNA were originally reported in 2001 when TLR3-deficient mice exhibited reduced immune responses to dsRNA viruses \[[@CR66]\].
Several studies have demonstrated that the immune response to siRNAs is cell type-dependent, due to the selective expression of TLRs. siRNAs stimulate monocytes and myeloid dendritic cells through TLR8 to produce pro-inflammatory cytokines, or activate plasmacytoid dendritic cells through TLR7 to produce type I interferons \[[@CR67]-[@CR69]\]. In addition, the volume of hydrodynamic naked siRNA delivery influences immune activation. Rácz et al. compared the immune responses induced by 50 μg siRNA dissolved in either low-volume (1 mL/mouse) or high-volume (10% of body weight, 2.5 mL/mouse in average) physiological salt solution delivered *in vivo*. Low-volume hydrodynamic injection induced slight alanine aminotransferase (ALT) elevation and mild hepatocyte injury, whereas high-volume hydrodynamic injection resulted in higher ALT levels and extensive hepatocyte necrosis. High-volume hydrodynamic injection also led to a time-dependent slight increase in IFN-related gene expression \[[@CR70]\]. Collectively, these studies suggest that there is a need for improving siRNA design, establishing experimental controls and carefully interpreting results.
Systemic compensative actions post siRNA treatment {#Sec15}
--------------------------------------------------
Silencing the target mRNA occasionally induces systemic compensative actions *in vivo*. In our previous research, the unmodified naked caspases-3 siRNAs aggravated renal graft injury \[[@CR14]\]. A potential reason for this effect is due to the kidney structure. The pore size of the glomerular filtration barrier is approximately 8 nm and naked siRNA has been observed to pass through this barrier into the urine \[[@CR20]\]. Upon reperfusion, the caspases-3 siRNA in the preserved kidney was flushed, while the local siRNA in the post-CS kidneys may also rapidly disappear due to a surge of blood flow into the kidney, leading to siRNA degradation and elimination. After transplantation, a series of *in vivo* complementary responses to the lower level of caspase-3 mRNA in the post-CS kidneys was initiated, which first led to an increase in caspase-3 mRNA synthesis. As the consequence of siRNA degeneration and caspase-3 mRNA synthesis, the level of caspase-3 mRNA increased in the siRNA preserved post-transplant kidneys. It has been suggested that a post-transplant feedback loop was initiated by the effective delivery of caspase-3 siRNA during CS, which led to increased mRNA synthesis after transplantation. However, with improved nuclease stability, the LNA modified caspases-3 siRNA protected both the post-CS and post-transplant renal grafts. More importantly, the modified siRNAs did not induce systemic compensative actions, which was proven by monitoring systemic inflammation \[[@CR18]\].
From bench to bedside: clinical trials {#Sec16}
======================================
The numbers of RNAi-based preclinical studies and clinical trials have grown over the past several years. To date, there have been 27 registered clinical trials using siRNA worldwide. These studies include retinal degeneration, dominantly inherited brain and skin diseases, viral infections, respiratory disorders, metabolic diseases and, of particular note, kidney diseases. In 2011, Quark Pharmaceuticals completed a phase I, randomized, double-blind, dose escalation, safety and pharmacokinetic study (NCT00554359) on QPI-1002, also designated I5NP, which was a synthetic siRNA that temporarily inhibits p53 expression that is in early development for acute kidney failure therapy. I5NP is the first siRNA to be systemically administered in humans. Based on data from animal studies, the intravenous injection in the human studies was performed within 4 hours of bypass surgery and pharmacokinetic data were obtained during the first 24 hours. Follow-up was conducted for safety and dose-limiting toxicities until hospital discharge and then by phone 6--12 months after surgery. Recently, Quark initiated a subsequent clinical trial to determine whether a single administration of I5NP can prevent delayed graft function in kidney transplant recipients. Data from this study will be used to identify I5NP doses for follow-on efficacy studies (NCT00802347). Another ongoing phase I trial investigating solid tumors, including RCCs, was conducted by Calando Pharmaceuticals. The investigators used CALAA-01, whose active ingredient is a type of siRNA, to inhibit tumor growth and/or reduce tumor size. This siRNA inhibits the expression of the M2 subunit of ribonucleotide reductase and resists nuclease degradation by using a stabilized nanoparticle that targets tumor cells (NCT00689065, the above clinical trials can be found at ClinicalTrials.gov, Table [2](#Tab2){ref-type="table"}).Table 2**Clinical trials of siRNA therapy in kidney diseasesStudyTarget/siRNA drugStatusDisease**NCT00554359I5NPPhase I, completedKidney injury; Acute renal failureNCT00802347I5NPPhase I/II, active, not recruitingDelayed graft function in kidney transplantationNCT00689065M2 subunit of ribonucleotide reductase/CALAA-01Phase I, terminatedSolid tumor cancersNCT02166255siRNA-transfected peripheral blood mononuclear cells APN401Phase I, not yet recruitingMelanoma, kidney cancer, pancreatic cancer, or other solid tumors that are metastatic or cannot be removed by surgery
Perspectives and challenges {#Sec17}
===========================
Despite the enormous potential of siRNA therapy, additional research must be performed before its large-scale clinical application.
Target gene selection {#Sec18}
---------------------
Genome-wide or pathway-specific siRNA libraries have become available using high-throughput screening approaches. Establishing *in vitro* pre-screening leads to signaling pathway prediction and target validation in *in vivo* renal disease. However, choosing one or a set of reasonable target gene(s) is the key for designing specific siRNA treatments. The pathophysiological changes during kidney disease, like any other disease, refer to a complex gene and protein regulation network. For example, the network that exists during kidney transplantation involves the original conditions of the donors and the interactions between the donor kidneys and the recipients, which could direct the progression, as well as the recovery, of the injury. Fortunately, transcriptome measurements of the transplant kidney may provide a comprehensive understanding of gene regulation and would be beneficial for target gene selection.
Mueller et al. analyzed the transcriptome of post-reperfusion implant biopsies in living donors (LD) and deceased donors (DD). Hundreds of mRNAs were identified that predicted delayed graft function \[[@CR71]\]. In a recent prospective study using human post-transplant kidney biopsies, 20 mRNAs and two miRNAs were identified as molecular signatures of AKI. Elevated secretory leukocyte peptidase inhibitor in AKI allografts was validated and miR-182-5p was identified as a molecular regulator \[[@CR72]\]. These genes could be used as potential targets of siRNA therapy. We recently identified 3 times more differentially expressed genes in renal allograft biopsies between living donors and cadaveric donors at 30 min than 3 months post-transplantation. The majority of these differentially expressed genes are responsible for acute responses at 30 min, but involved in inflammation, nephrotoxicity and proliferation at 3 months (Figure [2](#Fig2){ref-type="fig"}). These divergent transcriptome signatures between two types of donors might be linked with not only the initial injury of the donors, but also the immune responses of the recipients.Figure 2**Study design of identifying and validating potential targets for siRNA treatment using human transplant kidney biopsies.**
Another method for selecting target genes is by identifying their translation product proteins. To find a single or a set of crucial protein(s) involved in kidney allograft rejection, we explored potential transcriptional factors and regulation networks in 352 kidney transplant recipients, of which 85 suffered from acute rejection (AR). The results demonstrated that the dominant processes and responses were associated with inflammation and complement activation in AR. A number of transcription factors were identified in AR patients, including NF-κB, signal transducer and activator of transcription (STAT) 1 and STAT3 \[[@CR73]\]. Our recent study further revealed inflammation-derived kidney allograft injury, such as AR, chronic rejection, and impaired renal function without rejection. We identified 12 common proteins and 11 level-specific proteins from the phenotype-related protein--protein interaction networks \[[@CR74]\]. These potential biomarkers also provide valuable targets for transplant-related injury siRNA design.
Timely application {#Sec19}
------------------
Compared with shRNA, an advantage of siRNA for AKI therapy is the time-controlled treatment. Silencing the target gene for a short time or a long time should be addressed before RNAi application. The silenced genes may be multifunctional according to the surrounding milieu. For example, caspase-3, generally considered an executor in cell apoptosis, should be inhibited in tissue injuries. However, it is also a loyal scavenger in malignant transformation cells, which could be an unavoidable side effect in any caspase-3-targeting siRNA therapy. For AKI, siRNA ineffectiveness is needed after the therapeutic time window. Additionally, siRNA application avoids intracellular traffic. In certain circumstances, shRNA delivery could be harmful to the organ or even fatal. A study from Grimm et al. investigated the long-term effects of sustained high-level shRNA expression in the livers of adult mice. An evaluation of 49 distinct adeno-associated virus/shRNA vectors, with unique lengths and sequences that were directed against six targets, showed that 36 vectors resulted in dose-dependent liver injury, with 23 ultimately causing death. The observed morbidity was associated with the downregulation of liver-derived microRNAs (miRNAs), indicating possible competition of the latter with shRNAs (*through saturation of the endogenous RNAi machinery by the exogenous siRNAs*) for the limited cellular factors required for the processing of various small RNAs \[[@CR75]\]. Therefore, controlling intracellular shRNA expression levels will be imperative, but siRNA would not influence the endogenous process of RNA degradation mediated by miRNAs.
Cell-specific delivery {#Sec20}
----------------------
For optimized RNAi therapy, it should be determined whether regional delivery with partial knockdown or systemic delivery with global knockdown is required. However, there is still a lack of target specificity during systemic siRNA delivery.
It is known that p53 in PTCs promotes AKI, whereas p53 in other tubular cells does not \[[@CR40]\]. Therefore, the design of PTC-specific p53 siRNA is necessary. It is also expected that, for example, apoptosis-inducing siRNA should be directly delivered into tumor cells, rather than the surrounding normal cells.
Recently, antibody conjugation technology has made tumor-targeting drug delivery systems available. In general, these systems consist of a tumor recognition moiety and a cytotoxic warhead connected directly or through a suitable linker to form a conjugate. The conjugate can be regarded as a "guided molecular missile" that specifically targets tumor unique antigens \[[@CR76]\]. Inspired by cancer therapy strategies, siRNAs have also been "packed" to be delivered to target organs, even cells. Recently, a type of asymmetric liposome particle (ALP) has been developed, which highly efficiently encapsulates siRNA without nonspecific cell penetration. Two types of lipid inverted micelles have been prepared for the purpose of obtaining asymmetric liposome particles. The ALPs protected siRNA from ribonuclease degradation. ALPs without any surface modification elicited almost no uptake into cells, while the polyarginine peptide surface-modified ALPs induced nonspecific cell penetration. The conjugation of an anti-human epidermal growth factor receptor antibody (anti-EGFR) to the ALPs induced EGFR-mediated uptake into non-small cell lung cancer cell lines, but not into NIH-3 T3 cells that do not have the receptor \[[@CR77]\]. This result represents great progress in siRNA delivery system development. Antibody-mediated specific recognition may be more mainstream than virus-mediated recognition in the future.
Nanoparticulate systems have emerged in last few years as an alternative material for advanced diagnostic and therapeutic applications in medicine. Compared with molecular medicine, nanotechnology offers many advantages that overcome the range of challenges and barriers summarized in the previous section, particularly the bioavailability and biodistribution of therapeutic agents. The first remarkable property of nanoparticles is their superior *in vivo* retention due to decreased enzymatic degradation and sequestration by phagocytes in the reticulo-endothelial system. This property is primarily attributed to their immunochemically inert surfaces that are in contact with the biological environment. Methods of conjugating siRNAs with other inert and biocompatible molecules, such as cholesterol and long-chain fatty acids, have also been reported \[[@CR78],[@CR79]\].
Minimizing side effects {#Sec21}
-----------------------
The knockdown of two or more genes simultaneously using an siRNA cocktail has been recently reported. Many applications of siRNA cocktails have demonstrated significant benefits compared with siRNA targeted to a single gene, particularly in anti-cancer and anti-viral therapy \[[@CR56],[@CR80]\]. A high concentration of individual siRNAs may represent the key off-target effect in terms of competition for endogenous miRNA biogenesis machinery. Therefore, the other advantage of an siRNA cocktail is the relatively low concentration of each siRNA, which may reduce off-target signatures without sacrificing silencing potency \[[@CR55]\].
Conclusions {#Sec22}
===========
The kidney is a comparatively easy target organ for siRNA therapy due to its unique structural and functional characteristics. siRNA intervention is effective, feasible and has great potential for fighting against kidney diseases. The safety of siRNA therapy has been proven by rapidly emerging clinical studies and off-target and compensative responses can be managed using several strategies. We believe that optimized siRNA therapy, in conjunction with advanced genetic screening technologies, could facilitate timely and specific treatment for kidney diseases, as well as other organ diseases in the near future.
**Competing interests**
The authors declare that they have no competing interests.
**Authors' contributions**
Cheng Yang and Bin Yang conceived and prepared the manuscript. Bin Yang contributed throughout the preparation and edited the manuscript. Bin Yang, Tongyu Zhu and Cheng Yang provided foundation support. Chao Zhang and Zitong Zhao revised the manuscript. All authors read and approved the final manuscript.
This study was supported by the UK-China Fellowship for Excellence, Department for Business Innovation and Skills (to BY); University Hospitals of Leicester NHS Trust, UK, and National Natural Science Foundation of China (81170689 to BY, 81270833 to TZ and 81400752 to CY).
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
The malignant nature of circulating tumor cells (CTCs) has been recently demonstrated using xenograft models \[[@CR1], [@CR2]\], and it is widely accepted that these cells are responsible for tumor metastasis. CTCs have been detected in several tumor types, including breast, prostate, colorectal, non-small cell lung cancer (NSCLC), and small cell lung cancer (SCLC) \[[@CR3]--[@CR8]\]. Their detection has been associated with poor patients' clinical outcome, irrespectively of the disease stage. In addition, the presence of CTCs before and after the completion of adjuvant chemotherapy in breast cancer patients is associated with poor clinical outcome \[[@CR6], [@CR9], [@CR10]\].
Several studies have shown that CTCs present phenotypic and molecular differences from the corresponding primary tumor cells \[[@CR11]--[@CR13]\] and their changes in response to treatment could provide a useful alternate to solid biopsy \[[@CR13], [@CR14]\]. In addition, their detection and enumeration can give useful information regarding patient's prognosis providing a surrogate marker for treatment efficacy \[[@CR15]--[@CR17]\]. Nevertheless, it is well known that chemotherapy cannot effectively eliminate all CTCs; therefore, the identification of new molecules on these cells could offer new therapeutic targets.
Furthermore, only a subgroup of patients with detectable CTCs will eventually relapse. This is attributed to the fact that some of the identified CTCs are destined to die \[[@CR18]\]. Therefore, only a proportion of the detected tumor cells have metastatic potential; however, it is currently unknown which is the dangerous phenotype.
CTCs hold epithelial and/or mesenchymal (EMT) characteristics as well as stem cell features \[[@CR19]--[@CR21]\], and the acquisition of an EMT phenotype makes them invisible with common detecting platforms, based on the expression of epithelial markers. Consequently, there is an urgent need for new biomarkers to improve the identification and characterization of these cells.
Bioinformatics analysis is a new tool, based on Gene Expression databases, which can provide information about a number of genes-proteins that could potentially be upregulated in CTCs. Thus, the aim of the current study was the identification of novel biomarker/therapeutic targets in CTCs isolated from metastatic breast cancer patients based on bioinformatics analysis.
According to this analysis, CXCR4 is the most important molecule, potentially overexpressed in CTCs, while JUNB, YWHAB, TYROBP, NFYA, and PRDX1 were also found to be related to CXCR4 in different pathways and they could be potentially upregulated in CTCs \[[@CR22]\].Table 1Patient characteristicsMetastatic breast cancer No. of patients enrolled---100 Age, years Hormone receptor status Median, range 59 (66--81) ER-positive/PR-positive---45 (45%) ECOG performance status ER-positive/PR-negative---16 (16%) ER-negative/PR-positive---5 (5%) 0---66 (66%) ER-negative/PR-negative---18 (18%) 1---23 (23%) Unknown---16 (16%) 2---6 (6%) No. of disease sites Unknown---5 (5%) 1---51 (51%) Histology 2---21 (21%) Ductal---75(75%) 3---11 (11%) Lobular---10 (10%) ≥ 4---2 (2%) Other---9(9%) Unknown---15 (15%) Unknown---6 (6%) Visceral disease Menopausal status Yes---37 (37%) Premenopausal---26 (26%) No---50 (50%) Perimenopausal---12 (12%) Unknown---13 (13%) Postmenopausal---53 (53%) Triple negative (ER negPRnegHER2neg) Unknown---9 (9%) 11 (11%) HER2 status HER2 negative---64 (64%) Her2 positive---18 (18%) Unknown---18 (18%)
It is important that CXCR4 has been found to be a prognostic marker in various tumor types, including BC \[[@CR23]\]. It is upregulated in tumor tissue as compared to normal tissues. It has also been reported that CXCR4 is important for CTC-seeding and metastatic potential in early-stage BC of node-positive patients. Its expression was related to epithelial-mesenchymal transition (EMT), hence, associated with an acutely malignant phenotype \[[@CR24]\].
JUNB has been also associated to invasion/metastasis in solid tumors including BC \[[@CR25]--[@CR27]\] and represented as an important target in diseases associated with EMT, including cancer and fibrosis \[[@CR28], [@CR29]\]. It has been implicated in the earliest events of the development of resistance to kinase inhibitors in BC \[[@CR30]\]; however, the expression and the prognostic values of this transcription factor in CTCs have not been studied. Furthermore, the rest of the examined molecules (YWHAB, TYROBP, NFYA, and PRDX1) have been related to cancer progression and invasion \[[@CR31]--[@CR35]\], but to the best of our knowledge, they have not been studied in CTCs.
Methods {#Sec2}
=======
Dataset integration and bioinformatics analysis {#Sec3}
-----------------------------------------------
We collated an ensemble of nine human genome microarrays (GSE22820; GSE19783; GSE31364; GSE9574; GSE18672; GSE27562; GSE16443; GSE15852; GSE12763), from the Gene Expression Omnibus (GEO) database \[[@CR36]\]. The integration of the datasets followed the process described in the literature \[[@CR22]\], resulting in 498 primary breast cancer tissues and 124 blood samples from breast cancer patients, as well as 104 normal breast tissue samples and 85 peripheral blood samples from healthy donors. Three comparative analyses were performed in the pooled set of samples, using the "Significance Analysis of Microarrays" (SAM) \[[@CR37]\] as the feature selection approach of choice with the siggenes package of R/Bioconductor using the same parameters for all genes entries, in order to reveal comparison-specific differentially expressed genes. The false discovery rate (FDR) \[[@CR38]\] was used as the criterion for determining the set of genes that exhibit differential expression with critical value set to 0.01 for all comparisons.
Patient samples and cytospin preparation {#Sec4}
----------------------------------------
One hundred newly diagnosed and treatment-naïve patients with metastatic breast cancer (mBC) were enrolled in the study. All patients were enrolled in one center (Department of Medical Oncology, University General Hospital of Heraklion, Crete, Greece) and were treated with front-line chemotherapy, according to the national guidelines regardless of the CTCs' results. Patients' characteristics are shown in Table [1](#Tab1){ref-type="table"}. In addition, 10 female normal blood donors were also included as controls. All blood samples were obtained at the middle of vein puncture, after the first 5 ml of blood was discarded. These precautions were undertaken in order to avoid contamination of the blood sample with epithelial cells from the skin during sample collection. All patients gave their informed consent to participate in the study, which has been approved by the Ethics and Scientific Committees of our Institution.Table 2Mean intensity of JUNB, CXCR4, TYROBP, PRDX1, NFYA, and YWHAB in breast cancer cell lines, PBMCs, and CTCsMean intensityNormal PBMCsPatients PBMCsSKBR3MCF7MDA-MB 231HelaInt CTCsJUNB5.2 ± 0.154.4 ± 0.1126.1 ± 1.88.9 ± 0.566.6 ± 0.156.2 ± 0.659.6 ± 6.3CXCR410.62 ± 1.0125.4 ± 1.59168.8 ± 8.2513.86 ± 0.5110.22 ± 0.726.85 ± 2.6933.57 ± 9.25TYROBP5.17 ± 0.674.82 ± 0.213.97 ± 0.133.61 ± 0.15.08 ± 0.093.49 ± 0.248.89 ± 3.53PRDX115.9 ± 3.6234.73 ± 9.693.53 ± 0.226.25 ± 0.1822.01 ± 1.644.83 ± 0.0029.34 ± 24.65NFYA5.91 ± 0.526.08 ± 0.253.95 ± 0.234.93 ± 0.097.26 ± 0.155.41 ± 0.185.20 ± 1.36YWHAB26.13 ± 10.100 ± 04.04 ± 0.157.60 ± 0.1226.45 ± 3.528.20 ± 0.45.70 ± 0.28
Twenty milliliters of blood in EDTA was obtained from all patients, and peripheral blood mononuclear cells (PBMCs) were isolated with Ficoll-Hypaque density gradient (*d* = 1077 g/mol) centrifugation at 1800 rpm for 30 min. PBMCs were washed three times with PBS and centrifuged at 1500 rpm for 10 min. Aliquots of 500,000 cells were cyto-centrifuged at 2000 rpm for 2 min on glass slides. Cytospins were dried up and stored at -- 80 °C. Two slides with 10^6^ cells in total were analyzed from each patient for every molecule and for CK/CD45 staining. This number of PBMCs according to previous studies of our group could be representative of the characterization of CTCs in patients' blood \[[@CR18], [@CR19], [@CR39]\].
Cell cultures {#Sec5}
-------------
All cell lines were obtained from ATCC (American Type Culture Collection, USA) and used for spiking experiments. The MCF7 adenocarcinoma cells were cultured in 1:1 Dulbecco's modified Eagle medium (DMEM Glutamax) (GIBCO-BRL Co, MD USA) supplemented with 10% fetal bovine serum (FBS) (GIBCO-BRL), 16 ng/ml insulin, and 50 mg/ml penicillin/streptomycin (GIBCO-BRL). MDA-MB-231 cells were cultured in DMEM, supplemented with 10% FBS and 50 mg/ml penicillin/streptomycin. SKBR3 breast cancer cells were cultured in RPMI supplemented with 10% FBS. Hela were cultured in 1:1 Dulbecco's modified Eagle medium (DMEM Glutamax) supplemented with 10% FBS and 50 mg/ml penicillin/streptomycin. Cells were maintained in a humidified atmosphere of 5% CO~2~ in the air. Sub-cultivation was performed with 0.25% trypsin and 5 mM EDTA (GIBCO-BRL). All experiments were performed during the logarithmic growth phase.
Double staining experiments and confocal laser scanning microscopy {#Sec6}
------------------------------------------------------------------
The presence of CK-positive cells in PBMCs' cytospin was investigated using the mouse A45-B/B3 (detecting CK8, CK18, and CK19) (Micromet Munich, Germany) antibody and anti-CD45 (common leukocyte antigen) (Santa Cruz, Santa Cruz, CA, USA) in order to exclude possible ectopic expression of cytokeratin on hematopoietic cells. The cyto-morphological criteria proposed by Meng et al. \[[@CR40]\] (i.e., high nuclear/cytoplasmic ratio, larger cells than white blood cells, etc.) were used in order to characterize a CK-positive cell as a CTC.
Fixed cells with acetone/methanol (9:1 *v*/*v*) were stained with A45-B/B3 antibody for 1 h, followed by Alexa488 anti-mouse for 45 min. Afterwards, CD45 anti-rabbit was used for 1 h. Consequently, the samples were stained with the corresponding Alexa555 anti-rabbit fluorochrome. Finally, cells were stained with DAPI (Invitrogen, Carlsbad, CA, USA) conjugated with antifade and analyzed with confocal laser scanning microscopy.
Triple immunofluorescence {#Sec7}
-------------------------
Triple immunofluorescence was performed to cytospins' preparations from all patients harboring CK (+)/CD45 (−) CTCs. For the CK/CXCR4/JUNB staining, cells were fixed with 3% paraformaldehyde (PFA) for 30 min at room temperature (RT). Permeabilization was achieved with 0.5% Triton X-100 for 10 min at RT. After blocking with PBS supplemented with 10% (*v*/*v*) FBS for 1 h, cells were incubated with A45-B/B3 for 1 h and then with Alexa 488 for 45 min. Consequently, slides were incubated with JUNB antibody conjugated with Alexa647 (Santa Cruz, CA, USA) for 1 h. CXCR4 anti-rabbit antibody (ABCAM, Cambridge, MA USA) was added to the samples for 1 h, followed by the corresponding Alexa555 anti-rabbit (Molecular Probes, Invitrogen, Carlsbad, CA, USA) for 45 min. Finally, cells were stained with DAPI conjugated with antifade.
For the triple staining with the CK/PRDX1 (Santa Cruz, USA)/TYROBP (Abcam) or the CK/NFYA (Abcam)**/**YWHAB (Santa Cruz), cells were initially incubated with acetone/methanol (9:1) solution for 10 min and, then, washed 3 times with PBS. Blocking solution with 10% (*v*/*v*) FBS in PBS for 1 h was used to eliminate non-specific binding. Consequently, slides were incubated with the corresponding primary and secondary antibodies. Zenon technology (FITC-conjugated IGg1 antibody) (Molecular Probes, Invitrogen) was used for CK detection with the A45-B/B3 antibody. Zenon antibodies were prepared within 30 min before use. Cells were also stained with DAPI conjugated with antifade.
Positive and negative controls were used in each experiment (Additional file [1](#MOESM1){ref-type="media"}: Figure S1, Additional file [2](#MOESM2){ref-type="media"}: Figure S2), using breast cancer cell lines' cytospins by omitting one of the first antibodies. Therefore, each experiment included three different negative controls and one positive for all the antibodies.
Slides were, then, analyzed with confocal laser scanning microscopy, while the quantification of the protein expression was performed using the software platform of the ARIOL microscope. The platform automatically measured the intensity per pixel of each fluorochrome in every distinct isolated CTC.
Statistical analysis of the clinical data {#Sec8}
-----------------------------------------
All statistical tests were performed at the 5% level of significance. SPSS version 20 (SPSS Inc., Chicago, IL) statistical software was used for the analysis. Overall survival (OS) was defined as the time from treatment initiation, until death from any cause. Progression-free survival (PFS) was defined from the enrolment to the study until disease relapse or death whatever occurred first. Kaplan-Meier curves and Cox regression analysis for PFS and OS were compared using the log-rank test to provide a univariate and multivariate assessment of the prognostic value of selected clinical risk factors.
Immunohistochemistry: JUNB staining in tissue samples {#Sec9}
-----------------------------------------------------
Eleven primary and four metastatic tumors were available for JUNB staining from the same cohort of patients. One patient had available primary breast tumor and pleural metastasis. Three-micrometer-thick formalin-fixed paraffin-embedded (FFPE) tissue sections on charged glass slides were deparaffinized in xylene and rehydrated in a graded series of ethanols. Epitope retrieval was heat-induced in a vegetable steamer treated for 45 min in a solution buffered with citric acid (pH 6). Endogenous peroxidase was blocked by applying UltraVision Hydrogen Peroxide Block (Thermo Scientific, Waltham, MA) for 15 min. Nonspecific protein-binding sites were blocked with UltraVision Protein Block (Thermo Scientific, Waltham, MA) for 5 min. Sections were stained with Jun B mouse monoclonal antibody clone C-11 (1:50) (Santa Cruz, CA, USA) for 1 h at room temperature. Immunodetection was performed using UltraVision Quanto Detection System HRP Polymer DAB (Thermo Scientific, Waltham, MA, USA) according to the manufacturer's instructions. 3,3′Diaminobenzidine Quanto Chromogen (Thermo Scientific, Waltham, MA) was used as chromogen. Slides were counterstained with hematoxylin.
Samples' histoscore (*H*-score) was calculated by a semi-quantitative assessment of both the intensity of staining (graded as 0, no-staining; 1, weak; 2, medium; or 3, strong) and the percentage of positive cells. The percentage of cells at each staining intensity level was calculated, and an *H*-score was assigned summing the individual *H*-scores for each intensity level using the following formula: \[1 × (% cells 1+) + 2 × (% cells 2+) + 3 × (% cells 3+)\]. The range of possible scores was from 0 to 300.
Results {#Sec10}
=======
Bioinformatics results and functional enrichment evaluation {#Sec11}
-----------------------------------------------------------
Based on our previous work \[[@CR22]\], we performed a cross-platform comparative study of integrated public datasets, which revealed gene signatures exhibiting differentially overexpression patterns. In particular, the integrated datasets were of different "origin" (peripheral blood and tissue) and disease status (healthy and cancerous), and three different comparisons were performed (Fig. [1](#Fig1){ref-type="fig"}): cancer versus healthy tissue samples (Comparison 1, C1), cancer versus healthy blood samples (C2), and cancer tissue versus healthy blood samples (C3). These comparative analyses aimed to reveal persisting differences in the expression profile of certain genes in healthy and cancer status and possibly identify a limited set of candidate CTC biomarkers. For the C1 comparison, a total of 3725 genes were identified, exhibiting significant differential overexpression in cancer cells over control tissue samples. In the same way, 79 overexpressed genes were derived from the C2 comparison in cancerous peripheral blood (PB) samples and 245 overexpressed genes were extracted from the C3 comparison in cancer tissue samples over control PB samples. Based on our hypothesis, the peripheral blood carries information regarding both the primary and metastasis, as well as other cancer-induced alterations. We considered the intersection (C1∩C2∩C3) of the three gene signatures as a target of our biological assessment, since it includes only genes overexpressed in PB, related to cancer-associated factors, eliminating genes overexpressed in normal PB. This common set includes 24 genes, and CXCR4 plays a central role. A functional enrichment evaluation was further performed. Functional enrichment evaluation is generally recognized as a secondary analysis on large gene sets, resulted from high-throughput genomic methods. This step allowed the validation of the gene signature status, based on the biological significance of the 24 extracted genes. We further analyzed each of the 24 genes for their "intrinsic" properties and the 24-gene signature as whole according to their direct or indirect link with (i) *CXCR4* pathway and (ii) *CXCR4* biological features. Relevant biological information was extracted from four databases as follows: (a) pathways were acquired from the G2SBC (Genes-to-Systems Breast Cancer) \[[@CR41]\], KEGG (Kyoto Encyclopedia of Genes and Genomes) \[[@CR42]\], Gene Set Enrichment Analysis (GSEA) \[[@CR43]\], and WebGestalt (WEB-based GEne SeT AnaLysis Toolkit) \[[@CR44]\] in the category of "Pathway Commons" (PC); (b) molecular alterations in breast cancer, and the shortest pathways from *CXCR4* to other genes, were delivered from G2SBC; (c) gene ontology (GO) terms were yielded from G2SBC and GSEA; and (d) oncogenic molecular signature overlaps were computed from GSEA. Only genes participating in enriched biological terms at the *P* ≤ 0.05 level, after multiple test correction \[[@CR38]\], were considered as candidate genes and, if they shared multiple biological functions with CXCR4, were prioritized for further analysis in biological samples. Five genes, named JUNB, YWHAB, TYROBP, NFYA, and PRDX1, were evaluated as highly ranked genes, which could be tested in samples from mBC patients and cell lines (Fig. [1](#Fig1){ref-type="fig"}).Fig. 1Bioinformatics and functional enrichment analysis for biomarker discovery. Twenty-four genes were obtained from a data-driven computational procedure (data integration, cross-site comparisons) and were further explored by using functional enrichment analysis. CXCR4 plays a central role in this analysis. Based on four databases (G2SBC, WebGestalt, GSEA, and KEGG), significant biological information was validated and five genes of the 24-signature were prioritized according to their direct or indirect association with the CXCR4 pathway. These six molecules (CXCR4, *JUNB*, *YWHAB*, *TYROBP*, *NFYA*, and *PRDX1*) were experimentally evaluated in biological samples
Expression pattern of CXCR4 and JUNB in patients' CTCs {#Sec12}
------------------------------------------------------
Quantification of the mean intensity per pixel of each fluorochrome by the ARIOL system provided an expression pattern of every investigated protein in all the examined breast cancer cells lines from different subtypes (Luminal: MCF7, HER-positive: SKBR3, basal-like: MDA-MB 231). In addition, the expression of each protein in normal donors' PBMCs was quantified and compared to patients' monocytes. Hela cells were used in the analysis, because they were suggested by the manufacturers as positive controls for some of the antibodies (Fig. [2](#Fig2){ref-type="fig"}).Fig. 2Expression of cytokeratin, CXCR4, and JUNB in CTCs isolated from breast cancer patients. **a** Quantification of CXCR4 expression (intensity per pixel) in breast cancer cell lines, Hela cells, PBMCs, and patients' CTCs, using ARIOL system automate software. **b** Quantification of JUNB expression (intensity per pixel) in breast cancer cell lines, in Hela cells, PBMCs, and patients' CTCs, using ARIOL system automate software. **c** Cytospins obtained from metastatic breast cancer patients were triple stained with CK (green), CXCR4 (orange), and JUNB (red) antibodies plus DAPI (blue) and analyzed with ARIOL system (magnification X40). **d** Representative images from confocal laser scanning microscopy. Patients' samples were triple stained with CK (green), CXCR4 (red), and JUNB (blue) antibodies plus DAPI (not shown due to the absence of corresponding laser) magnification (× 60)
In the breast cancer cell lines MCF7 and SKBR3, the intensity of CXCR4 was statistically increased (*p* = 0.017 and *p* = 0.0001 respectively) compared to normal PBMCs (Fig. [2](#Fig2){ref-type="fig"}a). In addition, the mean intensity of CXCR4 in normal donors' PBMCs was 10.62 ± 1.01, while in patients' blood, the mean intensity was enhanced (25.4 ± 1.59; *p* = 0.0001; Table [2](#Tab2){ref-type="table"}) compared to normal PBMCs. Furthermore, the mean intensity of isolated CTCs in the patients was 33.57 ± 9.25 which was statistically higher compared to normal donors' PBMCs (*p* = 0.05) and to patients' PBMCs (*p* = 0.0001). Moreover, the expression of CXCR4 was higher in CTCs compared to MCF7 (*p* = 0.0027) and to MDA-MB 231 (*p* = 0.011) (Table [2](#Tab2){ref-type="table"}, Fig. [2](#Fig2){ref-type="fig"}a). Setting as the threshold value the mean intensity of CXCR4 in normal donors' PBMCs, 90% (18 out of 20) of the patients with detectable tumor cells in their blood harvested CTCs positive for CXCR4. In addition, 82.61% of the total isolated CTCs have an intensity higher than the threshold and considered positive for CXCR4.
JUNB expression in CTCs {#Sec13}
-----------------------
JUNB intensity was statistically higher in all breast cancer cell lines (SKBR3: *p* = 0.0001, MCF7: *p* = 0.001, MDA-MB231: *p* = 0.0001) compared to normal donors' PBMCs (Table [2](#Tab2){ref-type="table"}, Fig. [2](#Fig2){ref-type="fig"}b). In addition, the intensity of JUNB in isolated patients' CTCs was statistically higher compared to normal (*p* = 0.007) and patients' (*p* = 0.002) PBMCs. As it is shown in Fig. [2](#Fig2){ref-type="fig"}c, d, JUNB expression in CTCs was enhanced compared to patients' PBMCs, providing a potential biomarker to distinct cancer cells in patients' blood.
Furthermore, the intensity of JUNB was increased compared to MCF7 (*p* = 0.009) and MDA-MB231 (*p* = 0.0008) cells. Using as threshold value the mean intensity of JUNB in normal PBMCs, 65% (13 out of 20) of the patients with detectable tumor cells in their blood harbored CTCs positive for JUNB. In addition, 78.57% of the total isolated CTCs were JUNB-positive.
Expression pattern of TYROBP and PRDX1 in breast cancer cell lines and in patients' CTCs {#Sec14}
----------------------------------------------------------------------------------------
Quantification of TYROBP expression in breast cancer cell lines and in PBMCs from normal donors and patients revealed that this protein was downregulated in MCF7 (*p* = 0.008) cells compared to normal PBMCs. In addition, a significant difference was observed between the mean intensity of TYROBP in patients' PBMCs and breast cancer cell lines \[SKBR3 (*p* = 0.022), MCF7 (*p* = 0.007)\] (Table [2](#Tab2){ref-type="table"} and Fig. [3](#Fig3){ref-type="fig"}a).Fig. 3Expression of cytokeratin, TYROBP, and PRDX1 in CTCs isolated from breast cancer patients. **a** Quantification of TYROBP expression (intensity per pixel) in breast cancer cell lines, Hela cells, PBMCs, and patients' CTCs, using ARIOL system automate software. **b** Quantification of PRDX1 expression (intensity per pixel) in breast cancer cell lines, in Hela cells, in PBMCs, and in patients' CTCs, using ARIOL system automate software. **c** Cytospins obtained from metastatic breast cancer patients were triple stained with CK (green), PRDX1(orange), and TYROBP (red) antibodies plus DAPI (blue) and analyzed with ARIOL system (magnification × 40). **d** Representative images from confocal laser scanning microscopy. Patients' samples were triple stained with CK (green), TYROBP (blue), and PRDX1 (red) antibodies plus DAPI (not shown due to the absence of corresponding laser) and analyzed with confocal laser scanning microscopy (magnification × 60)
The mean intensity of TYPORB expression in patients' CTCs (8.89 ± 3.53) was higher than that of breast cancer cell lines \[MCF 7 (3.61 ± 0.1, *p* = 0.05), SKBR3 (3.97 ± 0.13, p = 0.007), MDA-MB231 (5.08 ± 0.09, *p* = 0.11)\]. It was also higher compared to normal (5.17 ± 0.67, *p* = 0.322) and to patients' (4.82 ± 0.21, *p* = 0.03) PBMCs.
Using as threshold value the intensity of TYROBP in normal PBMCs, 64.3% of the isolated CTCs were positive for TYROBP. In addition, 75% (12 out of 16) of the patients with detectable tumor cells in their blood harvested TYROBP-positive CTCs.
PRDX1 expression in CTCs {#Sec15}
------------------------
PRDX1 was also quantified in the same cohort of patients (Table [2](#Tab2){ref-type="table"}). The results revealed that PRDX1 expression was higher in MDA-MB231 cells (22.0 ± 1.64, *p* = 0.12) compared to normal PBMCs (15.9 ± 3.62). In addition, the mean PRDX1 expression in CTCs (29.34 ± 24.6) was numerically, but not statistically higher than that observed in breast cancer cell lines and in the normal PBMCs (15.9 ± 3.62) (Table [2](#Tab2){ref-type="table"}, Fig. [3](#Fig3){ref-type="fig"}b).
Using as threshold value the mean intensity of PRDX1 in normal PBMCs, 20% of the isolated CTCs were positive for PRDX1. In addition, only 12.5% (2 out of 16) of the patients with detectable tumor cells in their blood had detectable PRDX1-positive CTCs.
Expression pattern of NFYA and YWHAB in breast cancer cell lines and in patients' CTCs {#Sec16}
--------------------------------------------------------------------------------------
Evaluation of mean intensity of NFYA revealed that the highest expression among the breast cancer cell lines was observed in MDA-MB 231 \[(Table [2](#Tab2){ref-type="table"}), 7.26 ± 0.15)\] which was statistically different compared to normal donors' PBMCs (5.9 ± 0.52, *p* = 0.016). The intensity of NFYA expression in patients' CTCs (5.20 ± 1.36) was higher compared to MCF7 (4.93 ± 0.016) and SKBR3 (3.95 ± 0.021) cells, but it was lower than in normal PBMCs (5.9 ± 0.52, *p* = 0.005; Fig. [4](#Fig4){ref-type="fig"}a).Fig. 4Expression of cytokeratin, NFYA, and YWHAB in CTCs isolated from breast cancer patients. **a** Quantification of NFYA expression (intensity per pixel) in breast cancer cell lines, Hela cells, PBMCs, and patients' CTCs, using ARIOL system automate software. **b** Quantification of YWHAB expression (intensity per pixel) in breast cancer cell lines, in Hela cells, in PBMCs, and in patients' CTCs, using ARIOL system automate software. **c** Cytospins obtained from metastatic breast cancer patients were triple stained with CK (green), YWHAB (orange), and NFYA (red) antibodies plus DAPI (blue) and analyzed with ARIOL system (magnification × 40). **d** Representative images from confocal laser scanning microscopy. Patients' samples were triple stained with CK (green), NFYA (blue), and YWHAB (red) antibodies plus DAPI (not shown due to the absence of the corresponding laser) and analyzed with confocal laser scanning microscopy (magnification × 60)
Using as a threshold the intensity of NFYA expression in normal donors' PBMCs, only 6.45% of the whole number of isolated CTCs were positive for this protein. In addition, only 14.3% (2 out of 14) of the patients with detectable CK-positive cells were considered positive for NFYA.
YWHAB expression in CTCs {#Sec17}
------------------------
The intensity of YWHAB was also evaluated in breast cancer cell lines and in patients' CTCs. The mean intensity of YWHAB expression in normal donors' PBMCs (26.13 ± 10.10) was higher than that observed in MCF7 (7.60 ± 0.12, *p* = 0.043) and SKBR3 (4.04 ± 0.15, *p* = 0.058) cell lines. In contrast, MDA-MB231 (26.45 ± 3.52, *p* = 0.043) cells revealed significant higher expression compared to normal PBMCs.
In patients' CTCs, the mean intensity of YWHAB expression was 5.70 ± 0.28, which was significantly higher than the intensity observed in SKBR3 (4.04 ± 0.15, *p* = 0.01) cells, whereas it was lower compared to MCF7 (7.6 ± 0.12, *p* = 0.0001) and MDA-MB231 (26.45 ± 3.52, *p* = 0.002) cells. The mean intensity of YWHAB expression in CTCs was numerically but not significantly lower than in normal PBMCs (*p* = 0.07; Fig. [4](#Fig4){ref-type="fig"}b). None of the patients could be considered positive for YWHAB, taking as threshold value the normal donors' PBMCS mean intensity.
Clinical impact of the evaluated molecules {#Sec18}
------------------------------------------
Analysis of the clinical data of all the examined patients (100) revealed that after a median follow-up period of 24 months (range 1--101), 60 patients died due to disease progression. The median overall survival (OS) was significantly lower, in the group of patients harboring JUNB-positive CTCs (HR = 2.308, *p* = 0.026, Cox regression and *p* = 0.02: Kaplan-Meier analysis), compared to patients without JUNB-expression in their CTCs \[(17 months (range 0--38) vs 24.5 months (range 0--101); Fig. [5](#Fig5){ref-type="fig"} (A)\]. Particularly in the group of patients with JUNB-positive CTCs, 90% died during the follow-up period (9 out of 10 with available follow-up data) while in the cohort of patients without JUNB expression, 75% (45 out of 90) died during the follow-up period. Similarly, the median progression-free survival (PFS) was significantly lower in patients harboring JUNB-positive CTCs compared to patients without JUNB-positive CTCs \[3.5 months (range 0--30) vs 9 months (range 0--72), *p* = 0.015, Kaplan-Meier analysis\] (Fig. [5](#Fig5){ref-type="fig"} (B)). Interestingly, the presence of CK/CD45 cells in these patients did not correlate with clinical outcome. Furthermore, statistical analysis of the rest of the examined proteins did not lead to clinical correlation with OS or PFS. Multivariate analysis revealed that JUNB expression in CTCs is an independent prognostic factor (*p* = 0.016, HR 2.2484) for OS in breast cancer patients.Fig. 5**I** (A) Overall survival in patients with JUNB-positive CTCs compared to JUNB-negative patients (*p* = 0.02, Kaplan-Meier analysis). In the group of patients harboring JUNB-positive CTCs, 90% died during follow-up period vs 75% of deaths in the cohort of patients without JUNB expression. **I** (B) Kaplan-Meier analysis revealed that the progression-free survival in patients harboring JUNB-positive CTCs was statistically lower compared to patients without JUNB expression (*p* = 0.015). Patients without JUNB-positive CTCs experienced longer PFS vs JUNB-negative patients \[3.5 (range 0--30) vs 9 (range 0--72) months\]. **II** (A) JUNB staining in primary tumor obtained from BC patient. JUNB-positive tumor cells (brown) were very rare in the sample. Inner frame is shown control sample from a squamous cell carcinoma (magnification × 40). **II** (B) JUNB staining in metastatic tumor obtained from mBC patient. The majority of the tumor cells were positive for JUNB (magnification × 40)
Further analysis of the other clinicopathological characteristics revealed that reduced OS was significantly associated with performance status (PS) (*p* = 0.001, HR = 2.391, Cox regression).
JUNB expression in primary tumors {#Sec19}
---------------------------------
Eleven primary tumors from the same cohort of patients were screened for JUNB expression. Eight of them were negative for JUNB. Mostly all of the blood samples from these patients were also negative for CTCs. Only one patient in this group harbored JUNB-positive CTCs.
Three primary tumor samples revealed low expression of JUNB (*H*-scores 5, 6, and 9). One of them had also JUNB-positive CTCs.
On the other hand, all the examined samples (four) from metastatic tumors were positive for JUNB and the *H*-scores were very high (20, 120, 105, and 140). Three of these patients harbored also JUNB-positive CTCs in their blood.
It is interesting that in a patient with available primary tumor and pleura metastasis, the sample from the primary tumor was completely negative for JUNB expression (*H*-score 0), while in metastasis the *H*-score was extreme (120). This patient harbored also JUNB-positive CTCs (Table [3](#Tab3){ref-type="table"}).Table 3JUNB expression in primary and metastatic tumorsPatientMaterialJUN B IHC EVAL - FFPE tissuesJUNB positive (+) CTCs% (0--100)Intensity (0--3)H-score\* (0--300)1Primary tumor00*0*--2Primary tumor00*0*+3Primary tumor32*6*+4Primary tumor51*5*Negative for CTCs5Primary tumor00*0*Negative for CTCs6Primary tumor00*0*Negative for CTCs7Primary tumor00*0*Negative for CTCs8Primary tumor91*9*--9Primary tumor00*0*Negative for CTCs10Primary tumor00*0*Negative for CTCs11Primary tumor00*0*+Metastasis403*120*12Metastasis102*20*Negative for CTCs13Metastasis353*105*+14Metastasis40 &202&3*140*+\*Histoscore (H-score) was calculated by a semi-quantitative assessment of both the intensity of staining (graded as: 0: no-staining; 1: weak; 2: medium; or 3: strong) and the percentage of positive cells. The percentage of cells at each staining intensity level was calculated, and an H-score was assigned summming the individual H-scores for each intensity level using the following formula: \[1 × (% cells 1+) + 2 × (% cells 2+) + 3 × (% cells 3+)\]. The range of possible scores was from 0 to 300
Discussion {#Sec20}
==========
It is widely accepted that the detection of CTCs is a valuable parameter for the assessment of prognosis in cancer patients. CTCs can also be an important therapeutic target since in a randomized phase II trial it was observed that in patients with HER2-negative early breast cancer targeting HER2-positive CTCs with trastuzumab after the completion of adjuvant chemotherapy resulted in a significant improvement of patients' DFS \[[@CR45]\]. However, just the enumeration of CTCs is not sufficient for defining prognosis in cancer patients. Conversely, their characterization, which could be based on the expression of different proteins, is critical, since it allows the identification of CTCs undergoing EMT or expressing stem cell markers. In addition, it has been recently reported that certain CTCs' phenotypes can be related to metastatic procedure \[[@CR1], [@CR2]\]. Therefore, it is obvious that there is an unmet need for the development of new biomarkers for their detection and characterization. Bioinformatics analysis, based on gene expression databases, could help to this end.
Current study investigated whether bioinformatics analysis using public databases could define putative biomarkers with clinical relevance in CTCs, isolated from metastatic breast cancer patients. The presented bioinformatics data revealed that 6 genes could be potentially overexpressed in CTCs derived from breast cancer patients. All these genes are related to different CXCR4 signal transduction pathways. CXCR4 is a chemokine receptor that binds CXCL12. It has been shown that CXCR4 is upregulated in a number of tumors, such as breast, melanoma, ovarian, and prostate cancer, whereas its expression is low in normal tissues \[[@CR46], [@CR47]\]. It is also interesting to note that CXCR4 is induced by angiogenic factors, such as VEGF, and previous studies from our group had shown that VEGF and VEGFR2 were upregulated in CTCs derived from metastatic breast cancer patients \[[@CR39]\]. Interestingly in the present study, 90% of the patients with detectable CTCs in their blood harbored CTCs with CXCR4-positive phenotype. Furthermore, quantification of this expression (intensity per pixel) revealed that CXCR4 is significantly upregulated in patients' CTCs compared to MCF7 (*p* = 0.027) and MDA-MB231 (*p* = 0.011) cells and in comparison to normal PBMCs (*p* = 0.05). This observation strongly suggests that CXCR4 receptor is implicated to the metastatic process. These results confirm the conclusions of our bioinformatics analysis, and they are in agreement with previous studies regarding the CXCR4 mRNA expression in CTCs from breast cancer patients \[[@CR24], [@CR48], [@CR49]\]. CXCR4 expression in these studies was associated with the presence of \> 3 positive lymph nodes and with EMT markers such as Vimentin and Snail \[[@CR24]\] whereas Mego et al. \[[@CR48]\] reported that CXCR4 expression was related to CTCs exhibiting epithelial markers. In addition, CXCR4 was upregulated in CTCs derived from different tumor types such as small cell lung cancer, and this expression was associated with poorer PFS \[[@CR50]\]. However, in the current study although 82.6% of the whole number of detected CTCs were CXCR4-positive, this detection was not associated with the clinical outcome. Conversely, patients with JUNB expression in their CTCs revealed poorer OS (*p* = 0.002) and PFS (*p* = 0.015) compared to patients without this phenotype. Moreover, multivariate analysis showed that the detection of JUNB in CTCs could be emerged as an independent prognostic factor, associated with poor clinical outcome in terms of OS (*p* = 0.016). This clinical relevance of JUNB-positive CTCs raises the question of its significance as potential therapeutic target for the elimination of these cells. However, analysis of the different breast cancer patients' subtypes (hormone receptor-positive, HER2-positive, triple negative) separately did not reveal statistical differences.
The data of the current study are in line with previous studies reporting the role of JUNB in cancer progression, such as Hodgkin's disease and anaplastic large cell lymphomas \[[@CR51], [@CR52]\]. Moreover, it has been shown that JUNB is induced by ALK-NPM, participating the mTOR pathway \[[@CR53]\] and is required for cell cycle re-entry, after quiescence, and it cooperates with c-jun for the development of fibrosarcoma \[[@CR54]\]. In addition, there are some recent data indicating that JUNB is implicated in the earliest events of development of resistance to kinase inhibitors in breast cancer \[[@CR30]\]. However in the current study, JUNB overexpression could not be attributed to the treatment, because the samples were obtained before the initiation of first-line treatment. Combining all these data with our findings strongly supports that JUNB plays a critical role in cancer metastasis. This assumption was reinforced by the analysis of 11 primary tumors from the same group of patients, showing that only three of them revealed low JUNB expression, while the rest were completely negative. However, all the metastatic tumors (four) were strongly positive for JUNB. Furthermore, in one patient with available primary tumor and metastasis, JUNB was dramatically increased (from *H*-score 0 to *H*-score 120) in metastatic tissue, denoting the critical role of this molecule in cancer progression.
Our data also indicate an increased intensity of PRDX1 expression in patients' CTCs compared to normal PBMCs, but this difference was not statistically significant. This observation is in line with previous studies demonstrating that PRDX is increased in the most aggressive \[triple negative (TN)\] subtype of breast cancer patients \[[@CR55]\]. It is interesting that previous studies have shown that PRDX provide a protective role in cancer cells regarding doxorubicin-induced toxicity and they are associated with overall mortality in breast cancer patients \[[@CR33], [@CR35]\].
The intensity of TYROBP expression was also increased in CTCs compared to normal PBMCs and MCF7 cells, implying an upregulation of this protein in aggressive metastatic cancer cells. The strong statistical difference of TYROBP intensity between MCF7 and MDA-MB 231 (*p* = 0.00001) reinforced this assumption.
Conversely, although bioinformatics analysis revealed that NFYA and YWHAB genes' expression could be upregulated in CTCs, our results did not confirm this assumption. This could be attributed to post-transcriptional or post-translational modifications that prohibit the upregulation of protein levels of NFYA and YWHAB.
Conclusions {#Sec21}
===========
Bioinformatics analysis could be a useful tool for the identification of new biomarkers and therapeutic targets. Experimental approach to this analysis confirmed the overexpression of CXCR4, JUNB, and TYROBP. In addition, the presence of JUNB-positive CTCs emerged as an independent prognostic factor for OS in breast cancer patients.
Additional files
================
{#Sec22}
Additional file 1:**Figure S1.** Positive and negative controls of CK/CD45 and CK/JUNB/CXCR4 stainings. (IA): Positive controls for CK/CD45 staining: Cytospins with SKBR3 cells spiked in normal donor's PBMCs were stained with CK (green) anti-mouse, Alexa 488 anti-mouse, CD45 (blue) anti-rabbit, and Alexa 633 anti-rabbit antibodies. (IB): Negative controls for CK/CD45 staining: Cytospins were stained with all the above antibodies except the primary CK anti-mouse antibody. (IIA) Positive controls for CK/JUNB/CXCR4 staining: Cytospins with SKBR3 cells were stained with CK (green), JUNB (blue), CXCR4 (red) antibodies, and the corresponding fluorochromes. (IB) Negative controls for CK/JUNB/CXCR4 staining: Cytospins were stained with all the corresponding antibodies except the primary CXCR4 antibody. (IIC) Negative controls for CK/JUNB/CXCR4 staining: Cells were stained with all the corresponding antibodies except the primary JUNB antibody. (TIF 4960 kb) Additional file 2:**Figure S2.** (IA) Positive controls for CK/TYROBP/PRDX1 staining: MDA-MB 231 cells were stained with CK (green), TYROBP (blue), PRDX1 (red) antibodies, and the corresponding fluorochromes. (IB) Negative controls for CK/TYROBP/PRDX1 staining: Cells were stained with all the corresponding antibodies except the primary TYROBP antibody. (IC) Negative controls for CK/TYROBP/PRDX1 staining: Cells were stained with all the corresponding antibodies except the primary PRDX1 antibody. (IIA) Positive controls for CK/NFYA/YWHAB staining: MDA-MB 231 cells were stained with CK (green), NFYA (blue), YWHAB (red) antibodies, and the corresponding fluorochromes. (IIB) Negative controls for CK/NFYA/YWHAB staining: Cells were stained with all the corresponding antibodies except the primary NFYA antibody. (IIC) Negative controls for CK/NFYA/YWHAB staining: Cells were stained with all the corresponding antibodies except the primary YWHAB antibody. (TIF 4741 kb)
BC
: Breast cancer
CK
: Cytokeratin
CTCs
: Circulating tumor cells
CXCR4
: C-X-C chemokine receptor type 4
EMT
: Epithelial-mesenchymal transition
NFYA
: Nuclear transcription factor Y subunit alpha
OS
: Overall survival
PBMCs
: Peripheral blood mononuclear cells
PFS
: Progression-free survival
PRDX
: Peroxiredoxin
TYROBP (DAP12)
: Tyrosine kinase-binding protein
**Publisher's Note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
GK participated in the design and coordination of the study. She performed IF experiments. She performed the data interpretation, and she also drafted the manuscript. VT performed most of the immunofluorescence experiments. SA participated to biostatistics analysis and help drafted the manuscript. EB performed the functional enrichment analysis. EL and AK performed the experiments and the evaluation of the primary tumors. NZ and CS provided general support and help to draft the manuscript. SA and SA provided technical and economic support. MZ provided general support and participated in the study design and biostatistics analysis. VG provided general support and participated in the study design. He also participated in the manuscript preparation. All authors read and approved the final manuscript.
This study was supported, regarding consumables and study design, by the OncoSeed Program PN 3175 funded by European Union and the Greek General Secretary of Research and Technology. Cretan Association for Biomedical Research (CABR) provided support in data interpretation. Authors also appreciate funding of this work by the International Scientific Partnership Program ISPP at the King Saud University through ISPP\#009 which provide help in data interpretation.Financial support was also provided by Hellenic Oncology Research Group (HORG). This study was also co-financed by the European Regional Development Fund of the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation under the call RESEARCH-CREATE-INNOVATE (project code: T1EDK-03505, e-MASS)
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
The protocol has been approved by the Ethics and Scientific Committees of our institution, and all patients and healthy blood donors gave their informed consent to participate in the study.
Our study does not content any individual person's data.
The authors declare that they have no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction
===============
The chemoselective reduction of organic compounds is one of the most fundamental reactions in organic chemistry. Among these reductive reactions, the deoxygenation of epoxides into alkenes ([Scheme 1](#molecules-16-08209-f006){ref-type="scheme"}) has attracted much attention in both organic synthesis and biological chemistry where, for example, it occurs in the protection-deprotection cycle of carbon-carbon double bonds \[[@B1-molecules-16-08209],[@B2-molecules-16-08209],[@B3-molecules-16-08209]\] and in the production of vitamin K in the human body \[[@B4-molecules-16-08209],[@B5-molecules-16-08209]\].
{#molecules-16-08209-f006}
Generally, stoichiometric deoxygenations of epoxides have been conducted using expensive and/or toxic reagents such as low valent metals \[[@B6-molecules-16-08209],[@B7-molecules-16-08209],[@B8-molecules-16-08209],[@B9-molecules-16-08209],[@B10-molecules-16-08209],[@B11-molecules-16-08209],[@B12-molecules-16-08209],[@B13-molecules-16-08209]\], iodides \[[@B14-molecules-16-08209],[@B15-molecules-16-08209],[@B16-molecules-16-08209],[@B17-molecules-16-08209]\], phosphines \[[@B18-molecules-16-08209],[@B19-molecules-16-08209],[@B20-molecules-16-08209]\], and silane compounds \[[@B21-molecules-16-08209],[@B22-molecules-16-08209],[@B23-molecules-16-08209]\]. These reagents are often employed in excess, resulting in the production of large amounts of waste. To date, catalytic deoxygenations using Re complexes with triphenylphosphine \[[@B24-molecules-16-08209],[@B25-molecules-16-08209],[@B26-molecules-16-08209]\], a Fe complex with NaBH~4~ \[[@B27-molecules-16-08209]\], and a Co complex with Na \[[@B28-molecules-16-08209]\] have been reported, but these homogeneous catalytic systems have suffered from tedious work-up procedures, air- and moisture-sensitive reaction conditions, low atom efficiency and low catalytic activities. Therefore, the development of highly efficient and environmentally benign catalytic deoxygenation of epoxides is still a challenging issue.
Gold nanoparticles (Au NPs) have received much attention in the catalysis field due to their unique and high oxidation ability in various reactions such as oxidation of CO \[[@B29-molecules-16-08209]\] and alcohols \[[@B30-molecules-16-08209],[@B31-molecules-16-08209]\], and epoxidation of alkenes \[[@B32-molecules-16-08209],[@B33-molecules-16-08209]\]. On the other hand, the reduction ability of the Au NP catalysts has not yet been widely studied despite their remarkable activities \[[@B34-molecules-16-08209],[@B35-molecules-16-08209]\]. We have focused on exploring the novel catalytic reactions of Au NPs, and found that supported Au NPs showed unprecedented reduction activities in the deoxygenation of epoxides \[[@B36-molecules-16-08209],[@B37-molecules-16-08209],[@B38-molecules-16-08209]\], *N*-oxides, sulfoxides, and amides \[[@B39-molecules-16-08209]\].
In this account, we review our recent progress on the unique catalysis of Au NPs supported on hydrotalcite \[HT: Mg~6~Al~2~CO~3~(OH)~16~\] (Au/HT) for deoxygenation of epoxides to the corresponding alkenes using alcohols \[[@B36-molecules-16-08209]\] or CO/H~2~O \[[@B37-molecules-16-08209]\] as reductant ([Scheme 2](#molecules-16-08209-f007){ref-type="scheme"}). A wide range of epoxides were deoxygenated to the corresponding alkenes with over 99% selectivities. The C=C bonds of the products were not hydrogenated at all. After the reaction, the solid Au/HT catalyst could be easily separated and reused without loss of its activity or selectivity. Furthermore, Au/HT was successfully applicable to an ideal deoxygenation, *i.e.*, the H~2~-mediated deoxygenation of epoxides where only water is formed as a by-product \[[@B38-molecules-16-08209]\].
{#molecules-16-08209-f007}
2. Results and Discussion
=========================
2.1. Characterization of Au/HT
------------------------------
Au/HT was prepared by the deposition precipitation method (see Experimental). From atomic force microscopy (AFM) and transmission electron microscopy (TEM) analyses, the mean diameter (*d*) of Au NPs on the surface of HT was 2.7 nm with a standard deviation (*σ*) of 0.7 nm ([Figure 1](#molecules-16-08209-f001){ref-type="fig"}).
{#molecules-16-08209-f001}
2.2. Deoxygenation of Epoxides to Alkenes
-----------------------------------------
### 2.2.1. Au/HT-Catalyzed Deoxygenation of Epoxides Using Alcohols as a Reductant
Recently, we have found that Au/HT could catalyze the highly efficient aerobic oxidation of alcohols \[[@B40-molecules-16-08209]\] and the lactonization of diols \[[@B41-molecules-16-08209]\]. These results allowed us to predict that if epoxides could be employed as hydrogen acceptors in place of molecular oxygen under the alcohol oxidation conditions, a green catalytic deoxygenation of epoxides with alcohols could be developed ([Scheme 3](#molecules-16-08209-f008){ref-type="scheme"}) \[[@B36-molecules-16-08209]\].
{#molecules-16-08209-f008}
To demonstrate the above hypothesis, we carried out the deoxygenation of *trans*-stilbene oxide (**1**) using Au/HT with 2-propanol in toluene as the solvent at 110 °C under Ar atmosphere for 4 h. **1** was successfully deoxygenated to give the corresponding alkene *trans*-stilbene (**2**) in quantitative yield. Notably, no by-products such as 1,2-diphenylethanol or 1,2-diphenylethane resulting from the hydrogenation of **1** or **2** were formed ([Table 1](#molecules-16-08209-t001){ref-type="table"}, Entry 1). During the deoxygenation of **1**, the amounts of acetone and water generated were almost equivalent to that of **2**. Among the alcohols tested, 1-phenylethanol and benzyl alcohol could also function as reductants (Entries 2 and 3), while the use of an aliphatic primary alcohol such as 1-octanol resulted in lower yield (Entry 4). Next, the effects of inorganic supports of Au NPs were investigated. Au NPs on basic supports of Al~2~O~3~ and MgO afforded good to moderate yields of **2** (Entries 5 and 6), whereas non-basic supports like TiO~2~ and SiO~2~ were not effective (Entries 7 and 8). Other Au compounds like HAuCl~4~, Au~2~O~3~ and bulk Au metal did not promote the deoxygenation (Entries 9--11). These results indicate that the combination of Au NPs and a basic support is necessary to achieve the high catalytic activity for the deoxygenation.
molecules-16-08209-t001_Table 1
######
Deoxygenation of *trans*-stilbene oxide ^a^.
Entry Catalyst Alcohol Conv. ^b^ (%) Sel. ^b^ (%)
------- --------------- ----------------- --------------- --------------
1 Au/HT 2-propanol 99 \>99
2 Au/HT 1-phenylethanol 99 \>99
3 Au/HT benzyl alcohol 91 \>99
4 Au/HT 1-octanol 37 \>99
5 Au/Al~2~O~3~ 2-propanol 60 \>99
6 Au/MgO 2-propanol 43 \>99
7 Au/TiO~2~ 2-propanol 19 \>99
8 Au/SiO~2~ 2-propanol 5 \>99
9 HAuCl~4~ 2-propanol \<1 \-
10 Au~2~O~3~ 2-propanol \<1 \-
11 bulk Au metal 2-propanol 0 \-
^a^ Au cat. (0.45 mol%), toluene (5 mL), alcohol (10 mmol); ^b^ Determined by HPLC using an internal standard technique.
Other metal NPs on HT were examined for the deoxygenation of **1** ([Figure 2](#molecules-16-08209-f002){ref-type="fig"}). Among the catalysts tested, Ag/HT also showed excellent catalytic activity for deoxygenation, while other metal NPs did not function as catalysts.
{#molecules-16-08209-f002}
To investigate the possibility of the leaching of active metal species from Au/HT into the reaction mixtures, Au/HT was filtered from the reaction mixture at 50% conversion of 1, and treatment of the filtrate with additional stirring under similar conditions did not give any product. Furthermore, inductively coupled plasma atomic emission spectral (ICP-AES) analysis revealed no Au species in the filtrates (detection limit: 0.1 ppm). These results clearly proved that no leaching occurred and the deoxygenation proceeded on the Au NPs on HT.
The outstanding catalytic activity of Au NPs encouraged us to investigate the scope of epoxides in the deoxygenation ([Table 2](#molecules-16-08209-t002){ref-type="table"}). Various epoxides were efficiently converted into the corresponding alkenes with over 99% selectivity. Both aromatic and aliphatic epoxides could be deoxygenated. Epoxides having ether and hydroxyl groups were also successfully employed as substrates (Entries 9 and 16). Notably, the reducible C=O bonds of epoxyketones were tolerated in the deoxygenation (Entries 14 and 15). *cis*-Stilbene oxide and *cis*-2,3-epoxyoctane gave (*Z*)/(*E*)-alkene stereoisomers. The selectivities for *cis*-alkenes were 60% and 50%, respectively (Entries 4 and 13).
molecules-16-08209-t002_Table 2
######
Deoxygenation of various epoxides using Au/HT ^a^ .
Entry Substrate Product Time (h) Conv. ^b^ (%) Yield ^b^ (%)
-------- ---------------------------------- ---------------------------------- ---------- --------------- -------------------
1   4 99 99
2 ^c^ 4 99 99
3 ^d^ 4 97 97
4   4 99 97 ( *E/Z*) = 2/3
5   4 98 98
6 ^e^   6 \>99 98
7   6 96 92
8   6 94 91
9   6 93 90
10   24 89 89
11   24 87 87
12 ^e^   4 \>99 97
13 ^f^   12 81 77 (E/Z) = 1/1
14 ^f^   24 88 85
15 ^f^   24 99 94
16 ^f^   6 93 91
17   24 72 68
^a^ Reaction conditions: Au/HT (0.1 g), substrate (1 mmol), Ar; ^b^ Determined by GC or HPLC using an internal standard technique; ^c^ Reuse 1; ^d^ Reuse 2; ^e^ 80 °C; ^f^ Substrate (0.3 mmol), catalyst (0.2 g).
After the deoxygenation of **1**, the solid Au/HT catalyst could be easily separated from the reaction mixtures and reused with retention of its performance (Entries 2 and 3). TEM images showed that the Au NPs on HT after reuse were similar to fresh Au/HT in average diameter and size distribution and no aggregation of the used Au NPs was observed ([Figure 3](#molecules-16-08209-f003){ref-type="fig"}).
{#molecules-16-08209-f003}
Atomic-scale analysis using Au L-edge EXAFS of Au/HT showed that the intensity of the FT peak derived from the Au-Au shell at 2.8 Å was not changed, supporting the observation that the Au NPs after reuse were of the same size as the originals. These results are consistent with the high durability of Au/HT in the recycling experiments.
Au/HT was also applicable in a preparative scale reaction ([Scheme 4](#molecules-16-08209-f009){ref-type="scheme"}). Thus, the deoxygenation of 20 mmol of **1** successfully proceeded to afford **2** with 95% isolated yield in 2-propanol as solvent after 72 h, where the TON and TOF reached 20,000 and 270 h^−1^, respectively. These values are three orders of magnitude greater than those of previously reported catalytic systems such as Tp'ReO~3~-PPh~3~ (TON = 19) \[[@B24-molecules-16-08209]\], polystyrene-supported Re-PPh~3~(TON = 18, TOF = 12 h^−1^) \[[@B26-molecules-16-08209]\], CH~3~ReO~3~-PPh~3~ (TON = 8, TOF = 0.4) \[[@B25-molecules-16-08209]\], \[Fe~4~S~4~(SC~6~H~5~)~4~\]^3−^-NaBH~4~ (TON = 4, TOF = 1 h^−1^) \[[@B27-molecules-16-08209]\], and Co complex-Na (Hg) (TON = 5, TOF = 0.8 h^−1^) \[[@B28-molecules-16-08209]\].
{#molecules-16-08209-f009}
In separate experiments, the use of *d*-benzhydrol \[C~6~H~5~CD(OD)C~6~H~5~\] as a reductant for the deoxygenation of **1** afforded **2** and D~2~O with all hydrogen atoms in the alkene product retained. From both these results and the positive effect of basic supports as shown in [Table 1](#molecules-16-08209-t001){ref-type="table"} (Entries 1, 5 and 6), we propose the following mechanism as shown in [Scheme 5](#molecules-16-08209-f010){ref-type="scheme"}.
{#molecules-16-08209-f010}
A basic site (denoted as BS) of HT abstracts the H^+^ from the hydroxyl group of the alcohol to promote the formation of an Au-alcoholate species which subsequently forms an Au-hydride species and the corresponding carbonyl compound through β-hydride elimination \[[@B42-molecules-16-08209]\]. The Au-hydride species and H^+^ attack an epoxide, providing an alkene and H~2~O. The distinguished deoxygenation activities of Au NPs from those of other metal NPs should be attributed to the reactivity toward an epoxides (III→I) because the other metal NPs of Cu, Ru, Pd and Rh, which can form metal-hydride species from the reaction with the alcohol (I→III), do not deoxygenate the epoxide.
### 2.2.2. Deoxygenation of Epoxides with CO/H2O
We have previously reported that Rh carbonyl species can deoxygenate nitro compounds in the presence of amines as bases under water-gas shift reaction conditions (CO + H~2~O → CO~2~ + H~2~) \[[@B43-molecules-16-08209]\]. In this reaction, CO and H~2~O react with the Rh carbonyl species and an amine to form a Rh-hydride species that is active for the reduction of nitro compounds. The formation of the Rh-hydride species in cooperation with amines under water-gas shift reaction conditions inspired us to develop an alternative catalytic deoxygenation system using Au/HT. Namely, we proposed that an active Au-hydride species for the deoxygenation of epoxides can be formed through the cooperative effect of HT as a base under water-gas shift reaction conditions ([Scheme 6](#molecules-16-08209-f011){ref-type="scheme"}). Thus, the attack of H~2~O on the basic sites of HT to CO adsorbed on Au NPs generates \[Au-COOH\]^−^, followed by the elimination of CO~2~ to give the Au-hydride species and H^+^, which then act in concert to deoxygenate epoxides to alkenes \[[@B37-molecules-16-08209]\].
{#molecules-16-08209-f011}
Based on the above hypothesis, we carried out the deoxygenation of styrene oxide (**1a**) under water-gas shift reaction conditions in the presence of Au/HT. Styrene (**2a**) was quantitatively obtained as the sole product under atmospheric pressure CO in water at room temperature ([Table 3](#molecules-16-08209-t003){ref-type="table"}, Entry 1). Various epoxides tested in the Au/HT-2-propanol system were also reactive under water-gas shift conditions ([Table 4](#molecules-16-08209-t004){ref-type="table"}). Compared with the Au/HT-2-propanol system, this Au/HT-CO/H~2~O system can promote the deoxygenation of epoxides under mild and convenient reaction conditions, e.g., at room temperature, 1 atm of CO, and in the absence of organic solvents.
molecules-16-08209-t003_Table 3
######
Deoxygenation of styrene oxide using CO/H~2~O ^a^.
Entry Catalyst Yield ^b^ (%) Sel. ^b^ (%)
------- ------------------------ --------------- --------------
1 Au/HT \>99 \>99
2 ^c^ Au/HT \>99 \>99
3 ^d^ Au/HT 97 \>99
4 Au/Al~2~O~3~ 79 \>99
5 Au/MgO 36 \>99
6 Au/TiO~2~ 18 \>99
7 ^e^ Au/TiO~2~ + Na~2~CO~3~ 57 \>99
8 Au/SiO~2~ 3 \>99
^a^ Reaction conditions: catalyst (M: 0.9 mol%); water; ^b^ Determined by GC using an internal standard technique; ^c^ Reuse 1; ^d^ Reuse 2; ^e^ Na~2~CO~3~ (1.5 mmol) was added.
molecules-16-08209-t004_Table 4
######
Deoxygenation of various epoxides with CO/H~2~O using Au/HT ^a^.
Entry Substrate Product Time (h) Yield ^b^ (%) Sel. ^b^ (%)
------- ---------------------------------- ---------------------------------- ---------- --------------- --------------
1   6 \>99 (94) \>99
2   12 90 (85) \>99
3   12 99 (94) \>99
4   16 90 (85) \>99
5   16 96 (90) 96
6   24 90 (83) \>99
7   24 75 (69) \>99
^a^ Reaction conditions: Substrate (0.5 mmol), Au/HT (0.1 g, Au: 0.9 mol%); ^b^ Determined by GC using an internal standard technique. Values in parentheses are isolated yields; ^c^ Cinnamaldehyde was formed as a byproduct.
{#molecules-16-08209-f012}
The solvent effect on the deoxygenation is shown in [Scheme 7](#molecules-16-08209-f012){ref-type="scheme"}. Notably, water was found to provide the highest yield among all the solvents tested despite the water-insoluble nature of **1a**. In the case of the epoxy alcohol 2,3-epoxy-3-phenyl-1-propanol (**1b**), the highest yield and selectivity of cinnamyl alcohol (**2b**) were obtained in water. After the deoxygenation of **1a**, **2a** was easily extracted from the reaction mixture by *n*-hexane and the recovered aqueous phase containing solid Au/HT could be recycled with no decrease in catalytic activity ([Table 3](#molecules-16-08209-t003){ref-type="table"}, Entries 2 and 3).
{#molecules-16-08209-f004}
To gain more insight into the deoxygenation under the water-gas shift conditions, the following control experments were carried out. When the reaction was conducted in the absence of **1a** under otherwise identical conditions, H~2~was not detected. The use of D~2~O in place of H~2~O significantly affected the reaction rate for the deoxygenation of **1a** with a *k~H~/k~D~* value of 3.9. These results rule out the participation of H~2~ in the Au/HT-catalyzed deoxygenation reaction, while indicating that not only CO functions as a sole reductant, but also water takes part in the deoxygenation. An additional experiment using *trans*-2-octenal in place of **1a** revealed that chemoselective reduction occurred to give *trans*-2-octen-1-ol as the sole product while retaining the C=C double bond of the starting material. From the above results, we believe that an active Au-hydride species is generated in situ from the reaction of H~2~O with CO during the deoxygenation of epoxides. According to the proposed reaction mechanism ([Scheme 6](#molecules-16-08209-f011){ref-type="scheme"}), a basic site of HT facilitates the formation of the Au-hydride species through the nucleophilic attack of OH^−^ on the Au-CO species followed by a decarboxylation, which is well evidenced by the positive effect of the additional base of Na~2~CO~3~ to the Au NPs on the non-basic material of TiO~2~ ([Table 3](#molecules-16-08209-t003){ref-type="table"}, Entry 6 *vs.* 7). Finally, to confirm the generation of the Au-H species, Fourier transform infrared (FT-IR) studies of Au/HT were conducted in the presence of CO and H~2~O. When Au/HT was treated with CO and H~2~O vapor at 298 K, a new band attributed to the generation of Au-H species was detected around 1750 cm^−1^ ([Figure 4](#molecules-16-08209-f004){ref-type="fig"}). Next, this treated Au/HT was exposed to the vapor of **1a**, and the band attributed to the Au-H species gradually disappeared. The detection of the Au-H species agreed with recent IR and theoretical studies on Au--H species that predicted a band around 1800 cm^−1^ \[[@B44-molecules-16-08209],[@B45-molecules-16-08209]\]. These above control experiments are consistent with the proposed reaction mechanism as shown in [Scheme 6](#molecules-16-08209-f011){ref-type="scheme"}. The heterolytic H^+^ and Au-hydride species generated in situ on Au/HT deoxygenate the epoxide to form the corresponding alkene and water.
After the publication of our Au/HT-CO/H~2~O catalyst system, Cao *et al.* reported a deoxygenation method using Au/TiO~2~-VS (very small gold NPs on TiO~2~) under water-gas shift reaction conditions \[[@B46-molecules-16-08209]\]. Au/TiO~2~-VS showed high catalytic activity (TON = 9,600, TOF = 400 h^−1^) in the deoxygenation of styrene oxide in the mixed solvent of acetone with H~2~O under a high pressure CO atmosphere (10 atm).
### 2.2.3. Selective Deoxygenation of Epoxides Using Molecular Hydrogen
The ideal green methodology for the catalytic deoxygenation of epoxides is to utilize molecular hydrogen (H~2~) as a reductant due to the formation of non-toxic water as a by-product. However, the use of H~2~ often causes non-selective reduction of epoxides to yield alcohols and alkanes as byproducts through hydrogenation of the epoxides and overhydrogenation of the desired alkenes, respectively. Although there are a few successful reports on the selective deoxygenation of epoxides using H~2~, high selectivity for alkenes is restricted to low conversion levels \[[@B46-molecules-16-08209]\] or a limited range of substrates \[[@B47-molecules-16-08209]\]. Therefore, the development of an efficient catalytic system for the selective deoxygenation of epoxides to the corresponding alkenes using H~2~ is a challenging task.
With supported Au NPs in hand, the deoxygenation conditions were optimized \[[@B38-molecules-16-08209]\]. When the deoxygenation of **1a** using Au/HT was carried out at 80 °C under 1 atm of H~2~ for 6 h, **1a**was converted to **2a** in 97% yield accompanied by the formation of ethylbenzene (**3a**) as a byproduct through the hydrogenation of the desired product **2a** ([Table 5](#molecules-16-08209-t005){ref-type="table"}, Entry 1). Next, Au NPs on other supports were investigated in the deoxygenation of **1a** under similar reaction conditions. Au/CeO~2~ and Au/Al~2~O~3~ had lower selectivities for **2a**, which caused hydrogenation of **2a** (Entries 4 and 5) \[[@B48-molecules-16-08209]\]. Interestingly, Au/TiO~2~ gave **2a** with over 99% selectivity, though the conversion of **1a** was much lower than that of Au/HT (Entry 6). Au/SiO~2~ did not exhibit any catalytic activity toward this reaction (Entry 7). Remarkably, when the reaction was carried out at 60 °C for 8 h, Au/HT produced **2a** in over 99% yield without formation of any side products (Entry 2). Moreover, the C=C bond of **2a** was completely intact when the reaction time was prolonged (Entry 3). The Au NP catalysis exhibited different activity from other metal NPs. Ag/HT, Rh/HT, Ru/HT, and Cu/HT did not function as catalysts for this reaction (Entries 10-13). On the other hand, Pd/HT and Pt/HT afforded undesired **4a** with over 99% selectivity through the hydrogenation of **1a** (Entries 8 and 9).
molecules-16-08209-t005_Table 5
######
Deoxygenation of styrene oxide using H~2~^a^.
Entry Catalyst Temp. (°C) Time (h) Conv. ^b^ (%) Sel. for 2a ^b^ (%) Sel. for 3a ^b^ (%) Sel. for 4a ^b^ (%)
------- -------------- ------------ ---------- --------------- --------------------- --------------------- ---------------------
1 Au/HT 80 6 97 97 3 0
2 Au/HT 60 8 \>99 \>99 0 0
3 Au/HT 60 24 \>99 \>99 0 0
4 Au/CeO~2~ 80 6 64 81 19 0
5 Au/Al~2~O~3~ 80 6 82 36 64 0
6 Au/TiO~2~ 80 6 26 \>99 \<1 0
7 Au/SiO~2~ 80 6 \<1 \- \- \-
8 Pd/HT 80 6 \>99 0 0 \>99
9 Pt/HT 80 6 87 0 \<1 \>99
10 Ag/HT 80 6 \<1 \- \- \-
11 Rh/HT 80 6 \<1 \- \- \-
12 Ru/HT 80 6 \<1 \- \- \-
13 Cu/HT 80 6 \<1 \- \- \-
^a^ Reaction conditions: Catalyst (Au: 0.9 mol%), toluene (5 mL), 80 °C, 6 h; ^b^ Determined by GC using an internal standard technique.
Next, we conducted further studies on Au/HT and Au/TiO~2~ in the hydrogenation of **2a** in the presence or absence of *p*-methylstyrene oxide (**1c**) ([Scheme 8](#molecules-16-08209-f013){ref-type="scheme"}). In the absence of **1c**, Au/TiO~2~ hydrogenated **2a** into **3a**. Surprisingly, Au/HT did not show any activity toward the hydrogenation of **2a**. Neither Au/HT nor Au/TiO~2~ hydrogenated **2a** in the presence of **1c**. These sharply contrasting results indicate that the hydrogen species generated on Au/HT are active for the deoxygenation of epoxides, but are completely inactive for the hydrogenation of C=C bonds. On the other hand, the high selectivity of Au/TiO~2~ for alkenes in the deoxygenation of epoxides ([Table 5](#molecules-16-08209-t005){ref-type="table"}, Entry 6) is attributable to the preferential adsorption of epoxides over alkenes, which is a similar phenomenon to the previous report that the nitro group in 3-nitrostyrene was reduced by Au/TiO~2~ catalyst while retaining C=C bonds \[[@B49-molecules-16-08209]\].
{#molecules-16-08209-f013}
Various Au/HTs with different mean diameters of Au NPs were tested for the deoxygenation of **1a** with H~2~ ([Figure 5](#molecules-16-08209-f005){ref-type="fig"}). Different sized Au/HTs could be prepared by varying the concentration of HAuCl~4~ solution \[[@B38-molecules-16-08209]\]. Interestingly, larger Au NPs (\>3 nm) showed lower catalytic activity and selectivity for the deoxygenation due to the hydrogenation of the C=C bond. The selectivity and yield of **2a** increased with decreasing mean diameter of supported Au NPs. From these results, it can be said that immobilizing small Au NPs (\<3 nm) is the key to promoting the selective deoxygenation of epoxides to alkenes. The lower selectivity of larger Au NPs indicates that non-polar hydrogen species active for the hydrogenation of C=C bond were easily generated on the surface of larger Au NPs through the homolytic dissociation of H~2~. Au/HT with a mean diameter of 2.7 nm showed high catalytic activity and selectivity for the H~2~-driven deoxygenation of both aromatic and aliphatic epoxides to alkenes ([Table 6](#molecules-16-08209-t006){ref-type="table"}). After the reaction, Au/HT could be reused with no loss of its catalytic efficiency (Entries 2 and 3).
{#molecules-16-08209-f005}
molecules-16-08209-t006_Table 6
######
Deoxygenation of various epoxides using Au/HT ^a^.
Entry Substrate Product Time (h) Conv. ^b^ (%) Sel. ^b^(%)
------- ---------------------------------- ---------------------------------- ---------- --------------- ------------------
1   8 \>99 \>99
2 ^c^ 8 97 \>99
3 ^d^ 8 97 \>99
4   12 81 \>99
5   12 84 \>99
6   24 85 \>99
7 ^e^   12 98 \>99
8 ^e^   12 96 \>99 (E/Z = 2/3)
9   24 84 \>99
10   24 89 \>99
11   24 92 \>99
^a^ Reaction conditions: Au/HT (0.1 g), toluene (5 mL), substrate (0.5 mmol), 60 °C; ^b^ Determined by GC or HPLC using an internal standard technique; ^c^ Reuse 1; ^d^ Reuse 2; ^e^ At 100 °C.
Bearing in mind that basic ligands of transition-metal complexes promote heterolytic cleavage of H~2~ to give metal-hydride species, we propose a concerted effect between the basic sites of HT and Au NPs. Namely, heterolytic cleavage of H~2~ occurs to give \[Au-H\]^−^ and \[HT-H\]^+^ species at the interface between Au NPs and HT, which then react with an epoxide to afford an alkene and H~2~O ([Scheme 9](#molecules-16-08209-f014){ref-type="scheme"}). The formation of \[Au-H\]^−^ and \[HT-H\]^+^ was confirmed by FT-IR studies. When Au/HT was treated with 50 mbar of H~2~, IR bands attributed to \[Au-H\]^−^ and \[HT-H\]^+^ appeared around 1,748 cm^−1^ and 3,200 cm^−1^, respectively. The exclusive formation of the heterolytically cleaved hydrogen species of Au-hydride and H^+^ on HT would selectively deoxygenate epoxides to alkenes.
{#molecules-16-08209-f014}
3. Experimental
===============
3.1. General
------------
All organic reagents were purified before use. The products were identified by GC, HPLC and NMR analysis. Retention times (GC or HPLC) and chemical shifts (^1^H and ^13^C-NMR) of the products were in agreement with those of authentic samples (commercially available). HAuCl~4~·H~2~O was obtained from N. E. CHEMCAT. Co., Ltd. (Tokyo, Japan). MgO (GR for analysis) was purchased from Merck Chemicals Japan Co., Ltd. (Tokyo, Japan). Al~2~O~3~ (JRC-ALO-3), SiO~2~ (JRC-SIO-6) and TiO~2~ (JRC-TIO-4) were supplied by the Catalysis Society of Japan (Tokyo, Japan). Inductively coupled plasma measurements were performed by SII Nano Technology SPS7800. ^1^H and ^13^C-NMR spectra were recorded on JEOL JNM-AL400 and JNM-GSX270 spectrometers, respectively. GC (Shimadzu GC-2014) analysis was carried out with a KOCL-3000T and column. High-performance liquid chromatography (HPLC) was performed on a Shimadzu LC-10ADvp: STR ODS-IV. Atomic force microscopy (AFM) and transmission electron microscopy (TEM) micrographs were obtained with a Shimadzu SPM-9700 and Hitachi HF-2000 type microscope, respectively. Au L-edge X-ray absorption spectra were collected in the quick mode and recorded at room temperature in transmission mode at the facilities installed on the BL-01B1 line attached with a Si (311) monochromator at SPring-8, Japan Atomic Energy Research Institute (JASRI), Harima, Japan. Data analysis was performed using the REX 2000 program, ver. 2.5.7 (Rigaku). Fourier transformation (FT) of k~3~-weighted extended X-ray absorption fine structure (EXAFS) data was performed to obtain the radial structural function. FT-IR data were collected in a JASCO FT-IR 410 spectrometer equipped with a MCT detector. Self-supporting pellets were prepared from the sample powders and treated directly in the IR cell allowing thermal treatments under a controlled atmosphere.
3.2. Preparation of Au/HT
-------------------------
HT was prepared by the previously reported method \[[@B40-molecules-16-08209]\]. Au/HT was synthesized as follows: HT (1.0 g) was added to 50 mL of an aqueous HAuCl~4~ solution (2 mM). After stirring for 2 min, 0.09 mL of aqueous NH~3~ solution (10%) was added. The mixture was further stirred at room temperature for 12 h. The obtained slurry was filtered and washed with deionized water, and dried in vacuo to afford HT-supported Au(III) species \[Au(III)/HT\] as a pale yellow powder. Au(III)/HT (0.9 g) was subsequently stirred in 50 mL of KBH~4~ solution (18 mM) under Ar atmosphere at room temperature for 1 h. The solid was filtered and washed with deionized water to give Au/HT as a reddish purple powder.
4. Conclusions
==============
We discovered that Au/HT catalyzed the highly efficient deoxygenation of epoxides to the corresponding alkenes using various reductants. The obtained TON in the Au/HT-alcohol system was three orders of magnitude greater than that of previous reports. An alternative catalytic deoxygenation system was developed using CO/H~2~O as a reductant. The Au/HT-CO/H~2~O system could promote the deoxygenation of epoxides under mild reaction conditions (water, at room temperature, under 1 atm of CO). Under the water-gas shift reaction conditions, IR experiments revealed the *in situ* generation of the Au-hydride species. Finally, Au/HT realized an ideal green deoxygenation of epoxides using H~2~ as a reductant with water as the sole by-product. During these deoxygenations, no leaching of Au NPs from Au/HT to the reaction mixture occurred and Au/HT could be reused with retention of its high catalytic activity and selectivity. It is notable that the non-polar C=C bonds of products remained intact during these deoxygenations. The key to the above successful deoxygenation is the in situ generated Au-hydride and H^+^ species obtained through the concerted effect of the interface between Au NPs and basic sites of HT. We believe that this Au NP catalysis can be applied to other chemoselective reductions.
This work was supported by a Grant-in-Aid for Young Scientists (A) (23686116) from the Japan Society for the Promotion of Science (JSPS). We thank Uruga, Tanida, Nitta, Taniguchi and Hirayama (SPring-8) for XAFS measurements. The TEM experiments were carried out at a facility of the Research Center for Ultrahigh Voltage Electron Microscopy, Osaka University. One of the authors A.N. thanks the JSPS Research Fellowships for Young Scientists. He also expresses his special thanks to The Global COE (Center of Excellence) Program "Global Education and Research Center for Bio-Environmental Chemistry" of Osaka University.
*Sample Availability:*Not available.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-molecules-25-01922}
===============
Metabolic syndrome is a clustering of known cardiovascular risk factors including obesity, hyperglycemia, insulin resistance, dyslipidemia, and hypertension \[[@B1-molecules-25-01922],[@B2-molecules-25-01922]\]. There is a strong relationship between insulin resistance, visceral adiposity, and endothelial dysfunction. Each factor can contribute to the progression and development of each other \[[@B1-molecules-25-01922]\]. Moreover, the risk factors of the metabolic syndrome can depend on genetic and environmental factors. Family history of type 2 diabetes, hypertension, insulin resistance, and ethnic background have greatly increased the risk of progression of metabolic syndrome \[[@B2-molecules-25-01922]\]. The pathophysiology of metabolic syndrome is very complex and still unclear. Metabolic syndrome is associated with impairment of nitric oxide (NO) signalling and increasing the reactive oxygen species (ROS) generation which is related to the endothelial dysfunction \[[@B1-molecules-25-01922],[@B3-molecules-25-01922]\]. Three approaches have been suggested to correct the balance between increased oxidative stress and decreased NO synthesis. First is reduction of ROS bioavailability via administration of different antioxidant compounds including polyphenol reach substances. Second is elevation of NO level by administration of NO donors such as nitroglycerin, nitrosothiols, or mono/dinitrates. And third, probably the most effective approach, is simultaneous ROS reduction and NO stimulation, e.g., by the treatment with angiotensin-converting enzyme inhibitors, AT1-receptor blockers, β-receptor blockers, or statins with NO-increasing properties \[[@B3-molecules-25-01922],[@B4-molecules-25-01922],[@B5-molecules-25-01922],[@B6-molecules-25-01922],[@B7-molecules-25-01922]\]. Actually, statins belong to the first choice in the treatment of cardiovascular and obesity-associated diseases \[[@B5-molecules-25-01922],[@B7-molecules-25-01922]\]. Under certain conditions, statin therapy may be however associated with residual risk and several side effects, e.g., locomotion disturbances, nonallergic rhinitis, hyperglycemia, or rhabdomyolysis \[[@B7-molecules-25-01922]\]. Therefore, attention is also paid to the research of alternative treatments without harmful or side effects.
Recently, many studies have pointed to the beneficial effects of polyphenolic substances contained in Cornelian cherry (*Cornus mas* L., CC). Cornelian cherry is a member of the *Cornaceae* family and known as a medicinal plant that grows in eastern and southern Europe, southwest Asia and Middle East \[[@B8-molecules-25-01922],[@B9-molecules-25-01922],[@B10-molecules-25-01922]\]. All cultivars of the cornelian cherry have a high biological value, mainly associated with their antioxidant and anti-inflammatory activities which are attributed to a rich polyphenolic composition \[[@B11-molecules-25-01922]\]. CC includes mainly anthocyanins, flavonoids, iridoids, phenolic acids, and tannins \[[@B8-molecules-25-01922],[@B12-molecules-25-01922]\]. Actually, concentration of anthocyanins determines the color of fruits \[[@B8-molecules-25-01922],[@B11-molecules-25-01922]\]. Except for polyphenols, CCs is famous for being a rich source of ascorbic acids, and essential minerals. CCs have a higher level of ascorbic acid than oranges and strawberries \[[@B8-molecules-25-01922],[@B13-molecules-25-01922],[@B14-molecules-25-01922]\]. It also includes potassium and magnesium and in lower amount zinc, iron, copper, manganese, and sodium \[[@B8-molecules-25-01922],[@B12-molecules-25-01922],[@B13-molecules-25-01922]\].
Recently it has been shown that CCs have anti-diabetic, anti-obesity, hypolipidemic and anti-atherosclerotic properties that were attributed to their anti-inflammatory and antioxidant effects \[[@B13-molecules-25-01922],[@B14-molecules-25-01922]\]. In Wistar rats, hydroalcoholic fruits of CC were able to decrease blood glucose in a dose dependent manner \[[@B15-molecules-25-01922]\]. In alloxan-induced diabetic rats, hydroalcoholic fruits of CCs also decreased triglycerides, very low-density lipoprotein (VLDL) and low-density lipoprotein (LDL) levels \[[@B16-molecules-25-01922]\]. In the similar diabetes model, CC fruits effectively prevented the development of diabetes mellitus, increase of triglycerides and LDL, as well as elevation of aspartate, alanine aminotransferase, and alkaline phosphatase activities. Effects of CC fruits were comparable to that of glibenclamide \[[@B17-molecules-25-01922]\]. On the other hand, in streptozotocin-induced diabetic rats, cornelian cherry dried powder was not able to normalise glucose level, however, it decreased cholesterol, LDL and increased high-density lipoprotein (HDL) levels and liver antioxidant capacity comparing the diabetic group. In the same model, cornelian cherry dried powder had a similar inhibitory effect on liver HMG-CoA reductase activity as lovastatin \[[@B18-molecules-25-01922]\]. In New Zealand hypercholesterolemic rabbits, long-term treatment of CC powder decreased fibrinogen level even more significantly than lovastatin \[[@B19-molecules-25-01922]\]. In high-fat diet mice, anthocyanins and ursolic acid extract from CCs improved glucose tolerance and decreased bodyweight gain by decreasing lipid accumulation \[[@B20-molecules-25-01922]\]. Moreover, administration of CC powder in hypercholestrolemic rats were able to decrease triglycerides and had protective effects on atherosclerosis through enhanced PPARα protein expression and regulation of ROS production and inflammatory process \[[@B21-molecules-25-01922]\].
The aim of our study was to investigate the effects of two varieties of CC, namely Koralovij Marka (KM) and Wild Type (WT) on lipid profile, blood pressure, reactive oxygen species (ROS) and nitric oxide (NO) production in obese Zucker rats. Moreover, the effects were compared with effective antioxidant---coenzyme Q10 (CoQ10).
2. Results {#sec2-molecules-25-01922}
==========
2.1. Cornelian Cherry: Preparation and Characterisation {#sec2dot1-molecules-25-01922}
-------------------------------------------------------
Wild Type of Cornelian cherries had about 3× higher content of total polyphenols, 2× higher antioxidant capacity and comparable concentration of total anthocyanidins comparing to Koralovij Marka ([Table 1](#molecules-25-01922-t001){ref-type="table"}).
2.2. Bodyweight, Relative Heart Weight and Blood Pressure {#sec2dot2-molecules-25-01922}
---------------------------------------------------------
After six weeks of treatment, only KM group was able to decrease bodyweight in comparison to the control group. The relative heart weight (HW/TL) was significantly decreased in WT group only ([Table 2](#molecules-25-01922-t002){ref-type="table"}). Treatment with CoQ10, KM, or WT did not affect blood pressure in obese Zucker rats ([Table 2](#molecules-25-01922-t002){ref-type="table"}).
2.3. Plasma Lipid Profile {#sec2dot3-molecules-25-01922}
-------------------------
Plasma concentrations of total cholesterol and LDL were lower in WT group only compared to the control obese Zucker rats. Neither triglycerides nor HDL were changed within all groups ([Table 3](#molecules-25-01922-t003){ref-type="table"}).
2.4. Total NOS Activity {#sec2dot4-molecules-25-01922}
-----------------------
Total NOS activity in the left ventricle (LV) was not changed significantly within all groups ([Figure 1](#molecules-25-01922-f001){ref-type="fig"}A). On the other hand, CC varieties; KM and WT significantly increased NOS activity in the aorta ([Figure 1](#molecules-25-01922-f001){ref-type="fig"}B). CoQ10 did not affect NOS activity in both LV and aorta ([Figure 1](#molecules-25-01922-f001){ref-type="fig"}A,B).
2.5. Protein Expressions of eNOS, NADPH Oxidase, and SOD {#sec2dot5-molecules-25-01922}
--------------------------------------------------------
Western blot analysis was used to determine protein expressions within all groups. KM treatment increased eNOS protein expression ([Figure 2](#molecules-25-01922-f002){ref-type="fig"}A,B) and did not affect SOD ([Figure 3](#molecules-25-01922-f003){ref-type="fig"}A,B) or NADPH oxidase ([Figure 4](#molecules-25-01922-f004){ref-type="fig"}A,B) expressions, while WT treatment increased SOD ([Figure 3](#molecules-25-01922-f003){ref-type="fig"}A,B) and decreased NADPH oxidase ([Figure 4](#molecules-25-01922-f004){ref-type="fig"}A,B) without affecting eNOS ([Figure 2](#molecules-25-01922-f002){ref-type="fig"}A,B) expressions in both left ventricle and aorta. CoQ10 did not affect expressions of any protein studied ([Figure 2](#molecules-25-01922-f002){ref-type="fig"}A,B, [Figure 3](#molecules-25-01922-f003){ref-type="fig"}A,B, and [Figure 4](#molecules-25-01922-f004){ref-type="fig"}A,B).
2.6. Conjugated Diene Concentration {#sec2dot6-molecules-25-01922}
-----------------------------------
All substances studied, namely CoQ10, KM and WT significantly and comparable decreased CD concentrations in the left ventricle ([Figure 5](#molecules-25-01922-f005){ref-type="fig"}).
3. Discussion {#sec3-molecules-25-01922}
=============
The effects of two varieties of CC, namely Koralovij Marka and Wild Type, on lipid profile, blood pressure, reactive oxygen species, and nitric oxide production in obese Zucker rats were studied and compared with the effective antioxidant---coenzyme Q10. According to several studies, deficiency of CoQ10 is related to different diseases such as diabetes mellitus, atherosclerosis, hypertension, dyslipidemia, muscular dystrophy, and others. CoQ10 administration can protect against oxidative stress under conditions of cardiovascular disease, metabolic syndrome and type 2 diabetes \[[@B22-molecules-25-01922]\]. Moreover, CoQ10 may decrease oxidised-LDL induced generation of ROS and improved antioxidant capacity. It also reduced oxidised-LDL-mediated downregulation of eNOS and upregulation of iNOS \[[@B23-molecules-25-01922]\]. In our study, however, CoQ10 did not affect lipid profile and NOS activity either in the left ventricle or aorta. On the other hand, both KM and WT increased NOS activity in the aorta. While increased eNOS expression seems to be responsible for this effect in KM group, WT treatment probably stabilised NOS dimer by decreasing ROS production leading to increased NOS activity.
The characterization of Cornelian Cherrie varieties showed that Wilde Type from Slovakia had about 3× higher content of total polyphenols, 2× higher antioxidant capacity and comparable concentration of total anthocyanidins comparing to Koralovij Marka from Ukraine. Higher number of polyphenols may be responsible for marked antioxidant capacity of Wilde Type. However, the antioxidant capacity may be increased also by the content of vitamin C, which has not been determined in the present study. According to the available literature, vitamin C content ranges from about 35 to 420 mg/100 in different CC and may vary according to the changing conditions \[[@B8-molecules-25-01922],[@B18-molecules-25-01922]\]. A similar range of total polyphenols, as has been measured in our CC varieties, was determined in 10 samples of cornelian cherry fruits collected in Caucasus regions \[[@B24-molecules-25-01922]\]. In the same study, total anthocyanin level varied from 11.2 to 92.2 mg/100g \[[@B18-molecules-25-01922]\] and completely coincides with the level determined in our study (41.9 mg/100g in KM and 50.3 mg/100g in WT group).
Among CC groups, Koralovij Marka decreased bodyweight and WilC Type relative heart weight. Neither CoQ10 nor CCs affected BP. Similarly, supplementation of CC once a day in aged hamsters decreased weight gain and caused considerable hypoglycemic effect \[[@B25-molecules-25-01922]\]. The anthocyanin extract from CC caused a 24% decrease in weight gain in high-fat-fed C57BL/6 mice. These mice also showed decreased lipid accumulation and triacylglycerol concentration in the liver \[[@B20-molecules-25-01922]\]. In alloxan-induced diabetic rats, hydroalcoholic fruits of CCs in the dose 100 mg/kg for 110 days decreased triglycerides, VLDL and LDL levels \[[@B16-molecules-25-01922]\]. In our study, plasma concentrations of total cholesterol and LDL were lower only in the WT group compared to the obese controls. Neither TG nor HDL were changed within all groups.
We assume that decreased oxidative stress achieved by increased SOD and decreased NADPH oxidase protein expressions may be responsible for reduced cholesterol and LDL in WT group in our experiments. Similarly, Sozanski \[[@B21-molecules-25-01922]\] reported protective effects of CC on diet-induced hypertriglyceridemia and atherosclerosis through regulation of oxidative stress in rabbits \[[@B21-molecules-25-01922]\]. In another study, Sozanski et al. \[[@B26-molecules-25-01922]\] documented that iridoids and anthocyanins from cornelian cherry fruits reduced formation of atherosclerotic plaques in the aorta of cholesterol-fed rabbits. Again, decreased oxidative stress determined as an increase of GSH and attenuated lipid peroxidation measured as malondialdehyde concentrations were responsible for this beneficial effect \[[@B26-molecules-25-01922]\]. Similarly, Alavian et al. \[[@B27-molecules-25-01922]\] showed reduced lipid peroxidation after CC treatment in CCl4-induced hepatotoxicity in rats \[[@B27-molecules-25-01922]\]. Using a similar rat model, Es Haghi et al. \[[@B28-molecules-25-01922]\] showed that treatment of rats with different doses of CC fruit extract (300 and 700 mg/kg) improved the alterations developed by CCl4 in lipid peroxidation, antioxidant defences, biochemical and renal lesions. In particular, CC fruit treatment increased serum SOD, catalase and glutathione peroxidase \[[@B28-molecules-25-01922]\]. Our results are in good agreement with these studies since we found decreased concentration of another marker of lipid peroxidation and conjugated dienes after WT treatment in the heart as well. Interestingly, lipid peroxidation in the heart was reduced also after CoQ10 and KM treatments. Probably, another ROS regulating system, except SOD or NADPH oxidase, may be responsible for this effect.
We also hypothesised that decreased oxidative stress after WT treatment stabilised NOS dimer and may be responsible for increased NOS activity in the aorta. KM treatment did not affect SOD or NADPH oxidase protein expression. However, it increased NOS activity as well. In this case, we hypothesised that increased protein expression of eNOS, found in the aorta, may contribute significantly to this effect. Similarly, treatment of rabbits with high-cholesterol diet with loganic acid or anthocyanins extracted from CC led to increased mRNA expression of eNOS in thoracic aortas. \[[@B29-molecules-25-01922]\]. It seems that CC-derived anthocyanins represent promising molecules and dietary bioactive components that can effectively improve different factors of metabolic syndrome \[[@B30-molecules-25-01922]\]. Previously, we have also reported that red wine polyphenolic compounds containing anthocyanins increased eNOS protein expression accompanied by increased NOS activity and improved vasorelaxation in L-NAME-hypertensive rats \[[@B31-molecules-25-01922]\] or protected cardiac function in different models of metabolic syndrome \[[@B3-molecules-25-01922]\].
Taken together, both varieties of CC---Koralovij Marka and Wild Type---increased NOS activity in the aorta of obese Zucker rats. While increased eNOS expression seems to be responsible for this effect in KM group, WT treatment probably stabilised NOS dimer and increased activity by decreasing ROS production via increased SOD and decreased NADPH oxidase protein expressions. It seems that other polyphenols, different from that involved in Koralovij Marka, are mostly responsible for antioxidant effect of Wild Type variety of Cornelian Cherry.
4. Materials and Methods {#sec4-molecules-25-01922}
========================
4.1. Chemicals {#sec4dot1-molecules-25-01922}
--------------
Most of the chemicals and reagents were obtained from Sigma-Aldrich (Saint-Louis, MO, USA); if not, the company is indicated.
4.2. Cornelian Cherry Preparation and Characterisation {#sec4dot2-molecules-25-01922}
------------------------------------------------------
Koralovij Marka were provided by the National Botanical Garden in Kiev, Ukraine. Wild Type originating in the White Carpathians, Slovakia. Both CCs were stored and dried under the same conditions. The fresh fruit was mixed with the standard feed and the addition of water so that the mixture was mouldable into the desired cuboid form 3 × 3 × 3 cm. Subsequently, the blocks were dried for 6 h at 50 °C to 90% dry weight on a tray dryer.
For determination of anthocyanins, total polyphenols and antioxidant activity, the stoned fruit was homogenised, and 15 g of fruit was extracted in 30 mL of acidified 70% ethanol for 30 min, the extraction was repeated until the extractant had decolorised.
The AOAC differential pH method \[[@B32-molecules-25-01922]\] in two buffered solutions (KCL buffer pH 1.0 and sodium acetate buffer pH 4.5) was used to determine anthocyanin levels. The samples were diluted so that the absorbance of the solution was 0.2 to 1.2. After standing for 15 min in the dark, the absorbance at 510 and 700 nm was measured.
The total polyphenols content of the fruit extracts was determined by the Folin--Ciocalteu colorimetric method at 765 nm \[[@B33-molecules-25-01922]\]. The total polyphenol content was calculated by the calibration curve method as an equivalent of gallic acid with a linearity of 100 to 800 mg/L, corresponding to an absorbance of 0.1--0.9 (*R*^2^ = 0.9954). The total polyphenols content was converted to the polyphenols content in fruits in mg/kg.
The antioxidant activity of the extracts was determined by the free radical DPPH method with an absorbance λ 0.8 at 516 nm \[[@B34-molecules-25-01922]\] and calculated as a percentage of inhibition of the free radical DPPH. The EC 50 was calculated from the measured activities as the amount required to inhibit 50% DPPH in the reaction.
4.3. Animals and Treatment {#sec4dot3-molecules-25-01922}
--------------------------
All procedures and experimental protocols were approved by an Ethical committee of the Institute of Normal and Pathological Physiology Slovak Academy of Sciences (Ro-3601/17-221/3) according to the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purpose, Directive 2010/63/EU of the European Parliament.
Obese Zucker (fa^−^/fa^−^) rats were obtained from Charles River, USA. They were housed in groups of 2 animals, under a 12 h light- 12 h dark cycle, at a constant humidity (45--65%) and temperature (20--22 °C).
Twelve-week-old male Zucker (fa^−^/fa^−^) rats were divided into the control group and groups treated with coenzyme Q10 and different varieties of CC namely: Koralovij Marka and Wild Type. Each group consisted of 6 animals. Control and CoQ10 groups were fed with a standard diet ad libitum, CC groups were fed with special diet which contained dry fruit of CC (5 g/kg/day) and mixed with standard diet (30 g/day). CoQ10 (30 mg/kg/day) was administered via the drinking water. The treatment lasted for 6 weeks. Daily water consumption was estimated individually for every animal and adjusted, if necessary. Blood pressure was measured noninvasively, using tail-cuff plethysmography weekly. At the end of the treatment, the animals were sacrificed; heart weight (HW) and tibia length (TL) were determined. Relative heart weight was calculated as a HW/TL ratio. Samples of left ventricle and aorta were used to determine NOS activity and eNOS, SOD and NADPH oxidase protein expressions. Lipid profile was analysed in the plasma and conjugated diene levels in the left ventricle.
4.4. Plasma Lipid Profile {#sec4dot4-molecules-25-01922}
-------------------------
The levels of triglyceride, total cholesterol, HDL and LDL were measured in the plasma by commercially available kits.
4.5. Total NOS Activity {#sec4dot5-molecules-25-01922}
-----------------------
Total NOS activity was determined in crude homogenates of the LV and aorta by measuring the formation of \[^3^H\]-L-citrulline from \[^3^H\]-L-arginine (ARC, Saint Louise, MO, USA). 50 µL of 20% homogenates were incubated in the presence of 0.5 M Tris-HCl, pH 7.4, 10 mM NADPH, 20 mM CaCl~2~, 100 µM \[^3^H\]-L-arginine), 1mg/mL calmodulin, 1:1 FAD/FMN, and 50 mM TH~4~ in a total volume of 100 µL. After 30 min of incubation at 37 °C, the reaction was stopped by the addition of 1 mL of 0.02 M HEPES buffer pH 5.5, containing 2 mM EDTA, 2 mM EGTA, and 1 mM L-citrulline. The samples were then applied to 1 mL Dowex 50WX-8 columns (Na^+^ form). \[^3^H\]-L-citrulline was measured with Quanta Smart TriCarb Liquid Scintillation Analyser (Packard Instrument Company, Meriden, CT, USA).
4.6. Protein Expression Analysis by Western Blot {#sec4dot6-molecules-25-01922}
------------------------------------------------
Protein expressions of eNOS, SOD, and NADPH oxidase were determined in the LV and aorta by Western blot analysis. The samples were probed with polyclonal rabbit anti-eNOS, anti-SOD and anti-NADPH oxidase antibodies and anti-GAPDH and anti β-actin as a loading control (Abcam, Cambridge, UK). The intensity of bands was visualised using the enhanced chemiluminescence system (ECL, Amersham, UK), quantified by using ChemiDocTM Touch Imagine System (Image Lab^TM^ Touch software, version 5.2, BioRad, Hercules, CA, USA) and normalised to GAPDH bands for LV and β-actin bands for aorta.
4.7. Conjugated Dienes Determination {#sec4dot7-molecules-25-01922}
------------------------------------
The concentration of conjugated dienes was measured in lipid extracts of the LV. Samples of the LV were homogenised in 15 mmol/dm^3^ EDTA containing 4% NaCl. Lipids were extracted using a 1:1 chloroform-methanol mixture. Chloroform was evaporated in the N~2~ atmosphere and after the addition of cyclohexane, conjugated diene concentrations were determined spectrophotometrically (λ = 233 nm, NanpDrop 2000c, UV-Vis spectrophotometer). The concentration of CD was expressed as nmol per g tissue.
4.8. Statistics {#sec4dot8-molecules-25-01922}
---------------
The results are expressed as mean ± SEM. One-way analysis of variance (ANOVA) and Bonferroni test were used for statistical analysis. Values were considered significant with probability value *p* \< 0.05 (both for ANOVA and Bonferroni test). The *p* values were multiplicity adjusted.
5. Conclusions {#sec5-molecules-25-01922}
==============
Our study has contributed significantly to the finding that cornelian cherries can be a significant functional food component with beneficial effects on cardiovascular and metabolic disorders. Moreover, for the first time, it has been shown that mechanisms of these beneficial effects may differ depending on the variety of cornelian cherry. We have demonstrated that the variety with higher content of total polyphenols had better antioxidant properties and that both varieties of cornelian cherry were able to increase NOS activity although by different mechanisms. It means that cornelian cherries have the ability to increase the production of endothelial NO and thus contribute to the improvement of cardiovascular function in metabolic syndrome.
Sample availability
===================
Samples of the compounds are not available from the authors.
We are grateful to S.V. Klymenko from M.M. Gryshko National Botanical Garden of National Academy of Sciences of Ukraine, Kiev, Ukraine for providing CC variety Koralovij Marka.
**Sample Availability:** Samples of the compounds are not available from the authors.
Conceptualisation, O.P.; J.L. and M.C.; Formal analysis, O.P.; E.D. and M.C.; Funding acquisition, O.P.; Methodology, E.D.; E.P. and M.C.; Project administration, O.P. and J.L.; Writing--original draft, O.P.; Writing--review & editing, O.P.; E.D. and M.C. All authors have read and agreed to the published version of the manuscript.
This work was supported by national grant agencies APVV 14-0932, and VEGA 2/0112/19; 2/0132/20; 1/0035/19; 2/0151/18.
The authors declare no conflict of interest.
{#molecules-25-01922-f001}
{#molecules-25-01922-f002}
{#molecules-25-01922-f003}
{#molecules-25-01922-f004}
{#molecules-25-01922-f005}
molecules-25-01922-t001_Table 1
######
Determination of anthocyanins, total polyphenols and EC50 DPPH in stoned fruit.
Type Total Polyphenols \[mg/100 g\] Total Anthocyanidins \[mg/100 g\] EC50 \[g/L\]
----------------- -------------------------------- ----------------------------------- --------------
Koralovij Marka 151.2 ± 18.1 41.9 ± 5.0 0.54
Wild type 408.4 ± 49.0 50.3 ± 6.0 0.98
molecules-25-01922-t002_Table 2
######
Bodyweight (BW), heart weight (HW), heart weight (HW)/tibia length (TL) ratio and blood pressure (BP) in the control, coenzyme Q10 (CoQ10), Koralovij Marka (KM) and Wild Type (WT) groups.
BW (g) HW (g) HW/TL (×10^−2^) BP (mm-Hg)
--------- ------------------- ---------------- ----------------- ------------
Control 698.5 ± 20.4 1.33 ± 0.05 3.3 ± 0.1 147 ± 2.5
CoQ10 639.8 ± 42.3 1.29 ± 0.05 3.1 ± 0.1 142 ± 2.3
KM 611.5 ± 15.0 \*\* 1.29 ± 0.02 3.2 ± 0.07 143 ± 5.4
WT 664 ± 10.4 1.15 ± 0.02 \* 2.8 ± 0.03 \*\* 137 ± 3.9
Data are means ± SEM from 6 animals in each group. \*\* *p* \< 0.01 and \* *p* \< 0.05 compared to the control group.
molecules-25-01922-t003_Table 3
######
Lipid profile of control, coenzyme Q10 (CoQ10), Koralovij Marka (KM), and Wild Type (WT) groups.
TG (mmol/L) CHOL (mmol/L) HDL (mmol/L) LDL (mmol/L)
--------- ------------- ---------------- -------------- ---------------
Control 2.87 ± 0.21 7.65 ± 0.18 147.3 ± 10.1 70.9 ± 2.7
CoQ10 2.91 ± 0.48 6.23 ± 0.52 143.2 ± 6.3 49.6 ± 4.1
KM 2.88 ± 0.42 6.17 ± 0.40 125.7 ± 3.9 46.9 ± 3.2
WT 2.77 ± 0.17 4.10 ± 0.33 \* 128.3 ± 4.4 27.0 ± 1.3 \*
TG; triglyceride, CHOL; total cholesterol, HDL; high-density lipoprotein, LDL; low-density lipoprotein. Data are means ± SEM from 6 animals in each group. \* *p* \< 0.001 compared to the control group.
| {
"pile_set_name": "PubMed Central"
} |
Dear Editor
I wish to thank Dr. Enrico Pisoni and Dr. Rita Van Dingenen for their effort to write comments on my recent paper "*Assessing nitrogen dioxide* (*NO* ~*2*~) *levels as a contributing factor to coronavirus* (*COVID-19*) *fatality*" ([@bb0005]). I will answer all their comments as they appear in the letter to the editor.
[Answer to comment 1:]{.ul}
I agree with the comment about the limitation of satellite data. This data provides a spatial assessment of the NO~2~ concentration of a large area. However, even the ground-level NO~2~ measurement has limitation because it is valid only to the place of measurement and it is not possible to conclude about a certain area using this data. Therefore, the satellite, although it has its own limitations, provides continuous spatial NO~2~ data even in remote regions where no in-situ device is found. I should add that the absolute value itself is not important for the frame of my article but the relative value. The main conclusion of my research was that the high fatalities due to COVID-19 were observed in regions with high NO2 concentrations provided by the Sentinel 5.
The figures provided by the authors (Figs. 1, 2) show the annual mean NO~2~ concentrations observed at traffic stations in 2017, while the satellite image shows not only different periods (the two winter months January--February) but also different years (2020). Even if we assume that we can compare the two figures, the points in Fig. 2 are traffic stations which are located close to the emission sources (near roads, highways etc.) and therefore provide high NO~2~ values. These stations are not equally distributed, and their locations were chosen carefully. Therefore, the question is whether we can trust also the in-situ data for assessing the impact of air pollution on COVID-19? From my point of view, we cannot.
The boxplots shown in Fig. 3 are also not comparable. The data used in my article covers 66 regions in four different countries while the boxplots show a summary on the country level and the data is not equally distributed in space.
[Answer to comment 2:]{.ul}
The onset is different for each country. But it is also different for each region and each city. We can always go more into higher and higher resolutions and that would cause the data to be meaningless when the purpose is to publish preliminary results in a short communication manuscript.
[Answer to comment 3:]{.ul}
I do agree. Different countries reported it differently. If so, no research can or should be conducted using this dataset. Regarding to this remark, should the research stop? The data was provided from reliable sources and while conducting the research it also seems trustworthy.
[Answer to comment 4:]{.ul}
I agree that correlation is not causality. See my comments on that in [@bb0010] for more details.
[Answer to comment 5:]{.ul}
The number of deaths is related also to the number of people living in the region and to the number of people that were exposed to the virus. In the UK there are 66.6 million people and in Italy 60.4 million people. Additionally, the lockdown in the UK started several weeks after Italy.
[Concluding remark:]{.ul}
To summarize, my study '*Assessing nitrogen dioxide* (*NO* ~*2*~) *levels as a contributing factor to coronavirus* (*COVID-19*) *fatality*' was one of the first to shift the focus from the pre-existing diseases to the pre-existing environmental condition. During that time, the global community was thirsty for innovative ideas to tackle the global crisis and as a researcher, it was my great privilege to serve that purpose.
Declaration of competing interests
==================================
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction
===============
Parkinson's disease (PD) is a common neurodegenerative disease in the aging population. PD is characterized by the selective death of dopaminergic neurons in the substantia nigra leading to low levels of dopamine in the striatum \[[@B1-ijms-15-13388]\]. The etiology of PD is still elusive. Oxidative damage and mitochondrial dysfunction have been proposed to play a major role in the pathogenesis of PD \[[@B2-ijms-15-13388]\]. Rotenone, a widely used pesticide, has been used for many years as a tool to induce PD model *in vitro* and *in vivo* \[[@B3-ijms-15-13388],[@B4-ijms-15-13388],[@B5-ijms-15-13388]\] and proven to be reproducible. Rotenone inhibits complex I of the mitochondrial electron transporter chain (ETC). Its neurotoxicity may be related to the ability of generating reactive oxygen species (ROS) and disrupting mitochondrial oxidative phosphorylation. Elevated ROS levels cause depolarization of the mitochondrial membrane and release of pro-apoptotic factors such as apoptosis inducing factor (AIF), which eventually causes neuronal death \[[@B6-ijms-15-13388]\].
Coenzyme Q10 (CoQ10), also known as ubiquinone Q10, is an electron transporter that transports electrons from ETC complex I and complex II to complex III. In addition, CoQ10 also possesses antioxidant and anti-apoptotic effects \[[@B7-ijms-15-13388],[@B8-ijms-15-13388],[@B9-ijms-15-13388]\]. CoQ10 has lipid-soluble and water-soluble formulas. Most CoQ10 from commercial sources is lipid-soluble and has low bioavailability. In recent years, water-soluble CoQ10 has been developed. Compared with lipid-soluble CoQ10, the water-soluble formula has a much better bioavailability when administered orally \[[@B10-ijms-15-13388]\]. Rotenone has been shown to activate a mitochondria-initiated, caspase-dependent cell death pathway. It is controversial whether rotenone can also activate a caspase-independent cell death pathway. In this study, we first investigated the effects of water-soluble CoQ10 on rotenone-induced neuronal cell death. We then measured ROS production, mitochondrial membrane potential and AIF nuclear translocation. We finally detected cytochrome *c* release and caspase-9 activation.
2. Results and Discussion
=========================
2.1. Effects of Rotenone on Viability of HT22 Cells
---------------------------------------------------
To define the adequate dosage of rotenone for inducing a targeted \~50% cell death, HT22 cells were incubated with four different concentrations (0.25, 0.5, 1.0, 2.0 µM) of rotenone for 24 h. As shown in [Figure 1](#ijms-15-13388-f001){ref-type="fig"}, cell viability decreased with increasing concentrations of rotenone. Naïve (Co) and vehicle (VC, 0.1% DMSO) treated control cells had a viability close to 100%. Rotenone at 0.25 and 0.5 µM resulted in 40% cell death, while rotenone at 2 µM caused 70% cell death. Viability was about 50% in the 1.0 µM rotenone group. As a result, this 1.0 µM rotenone was selected for subsequence experiments.
{#ijms-15-13388-f001}
2.2. Protective Effects of CoQ10 against Rotenone Toxicity
----------------------------------------------------------
To determine the efficacy of water-soluble CoQ10 on rotenone-induced cell death, three dosages (50, 100, 200 µM) of CoQ10 were tested. The cells were treated with CoQ10 for 3 h prior to 1 µM rotenone exposure. Cell viability was assessed by alamar blue at 24 h after rotenone addition. As shown in [Figure 2](#ijms-15-13388-f002){ref-type="fig"}, water-soluble CoQ10 itself did not induce stress to the cell as the viability was close to 100% when cells were treated with CoQ10 alone with dosages of 50, 100, and 200 µM. The efficacy of CoQ10 against rotenone toxicity is in a dose-dependent manner. At the concentration of 50 µM, CoQ10 significantly increased the percentage of viable cells from 40% in 1.0 µM rotenone incubated cells to 60% in CoQ10 treated cells. When the CoQ10 concentration increased to 200 µM, the percent of viable cells further increased to 80% ([Figure 2](#ijms-15-13388-f002){ref-type="fig"}). Because 200 µM CoQ10 displayed the best efficacy, this concentration was chosen for subsequent CoQ10 experiments.
2.3. Effects of CoQ10 on ROS Production
---------------------------------------
DHE is oxidized by superoxide to form ethidium and precipitated on the sites where superoxide anions were produced. Rotenone incubation for 24 h induced a more than 100% increase in the production of superoxide comparing to DMSO vehicle-incubated cells ([Figure 3](#ijms-15-13388-f003){ref-type="fig"}). CoQ10 200 µM pretreatment reduced the superoxide level back to normal. CoQ10 itself did not increase superoxide formation.
{#ijms-15-13388-f002}
{#ijms-15-13388-f003}
2.4. Effects of CoQ10 on Mitochondrial Membrane Potential
---------------------------------------------------------
Mitochondrial membrane potential decreased significantly in cells incubated with rotenone. Pretreatment with 200 µM CoQ10 prevented the depolarization of mitochondrial potential caused by rotenone ([Figure 4](#ijms-15-13388-f004){ref-type="fig"}).
{#ijms-15-13388-f004}
2.5. Blockade of Rotenone-Induced AIF Nuclear Translocation by CoQ10
--------------------------------------------------------------------
AIF normally localized in the mitochondria. When cells are under stress or injury conditions, AIF relocated to the nuclei where it causes DNA and nuclear envelope damage. Our results showed that rotenone increased AIF level in the nuclear fraction after 24 h of incubation. Treatment of rotenone-incubated cells with 200 µM water soluble CoQ10 significantly reduced the protein level of AIF in the nuclear fraction ([Figure 5](#ijms-15-13388-f005){ref-type="fig"}), suggesting a blockade of AIF nuclear translocation by the CoQ10.
In order to verify whether rotenone also activates a mitochondria-initiated, caspase-dependent cell death pathway, we detected cytochrome *c* in the cytosolic and mitochondrial fractions. As shown in [Figure 6](#ijms-15-13388-f006){ref-type="fig"}A, not even traceable level of cytochrome *c* could be detected in the cytosolic fractions after 24 h rotenone incubation. This was supported by the results showing no decrease of cytochrome *c* in the mitochondrial fraction ([Figure 6](#ijms-15-13388-f006){ref-type="fig"}B,C) and no increase of cleaved caspase-9 in the cytosolic fraction ([Figure 6](#ijms-15-13388-f006){ref-type="fig"}D,E).
{#ijms-15-13388-f005}
{#ijms-15-13388-f006}
2.6. Discussion
---------------
In this study, we have revealed that rotenone induces neuronal death in a dose-dependent manner and water-soluble CoQ10 exerts a potent protection against rotenone-induced cell death. We have further demonstrated that the neuroprotective effect of CoQ10 is associated with its ability in preventing ROS increase, blocking mitochondrial membrane depolarization and inhibiting AIF translocation from the mitochondria to the nuclei.
In present study, ROS levels increased significantly in rotenone incubated HT22 cells, which is consistent with previous reports \[[@B11-ijms-15-13388]\]. ROS are mainly produced in mitochondrial electron transport chain complexes I and III due to incomplete oxygen reduction. Rotenone binds to complex I, impeding the electron transport from complex I to III leading to the generation of superoxide \[[@B12-ijms-15-13388]\]. It has been proposed that rotenone results in mitochondrial dysfunction and apoptosis through enhancing the production of ROS \[[@B13-ijms-15-13388],[@B14-ijms-15-13388]\].
It is well known that increases in ROS production can trigger the formation of the mitochondrial permeability transition pore (MPTP). The MPTP causes the mitochondria to release proapoptotic proteins including cytochrome *c*, Smac/DIBLO, endonuclease G (EndoG) and AIF \[[@B15-ijms-15-13388],[@B16-ijms-15-13388],[@B17-ijms-15-13388],[@B18-ijms-15-13388]\], which eventually result in cell death. Rotenone chronically intoxicated rats show increased release of cytochrome *c* and activation of caspase-3 \[[@B19-ijms-15-13388]\]. Human Parkinson's disease brain samples have increased nuclear localization of EndoG \[[@B20-ijms-15-13388]\]. AIF is normally localized in the mitochondrial intermembrane space \[[@B21-ijms-15-13388]\]. Under pathological situations such as oxidative stress, AIF is released from the mitochondria to the nucleus to activate a mitochondria-initiated, caspase-independent cell death pathway, also called AIF-mediated apoptosis. In a previous study, Li and colleagues observed that rotenone induced cytochrome *c* and AIF release; but it did not induce caspase-9 and -3 activation in undifferentiated human neural stem cells \[[@B22-ijms-15-13388]\]. However, cleaved caspase-9 and -3 were increased by rotenone when differentiated human neural stem cells were used \[[@B22-ijms-15-13388]\]. In this study, we have observed that rotenone incubation causes a drastic increase of AIF in the nuclear fraction, suggesting activation of a mitochondria-initiated, caspase-independent cell death pathway. To further examine whether rotenone also activates a caspase-dependent cell death pathway, protein contents of cytochrome *c* and cleaved caspase-9 were measured in the mitochondrial and/or cytosolic fractions 24 h after rotenone incubation. The results showed that a caspase-dependent pathway was not activated since cytochrome *c* was not released into the cytosol and caspase-9 was not activated. These results further strengthened the conclusion that rotenone activates an AIF-mediated cell death pathway. AIF could be leaked out from mitochondria through the MPTP or a Bax-mediated channel on the mitochondrial membrane. Our results showed a significant depolarization of mitochondrial membrane potential in rotenone-incubated cells, suggesting AIF is probably released through the MPTP.
In this study, we have observed rotenone increases both production of ROS and nuclear translocation of AIF and decreases mitochondrial membrane potential in HT22 cells. Such phenomenon has also been observed in SH-SY5Y cells exposed to oxygen-glucose deprivation \[[@B23-ijms-15-13388]\] and in MRC-5 fibroblasts overloaded with calcium \[[@B24-ijms-15-13388]\]. Although it has been reported that ROS can induce mitochondrial AIF release and nuclear translocation, it is not known whether AIF can increase the production of ROS. A recent study shed light on this issue; using an AIF and complex I deficient mouse model, Chinta and colleagues found AIF did not directly modulate the generation of ROS \[[@B25-ijms-15-13388]\]. It is likely that the AIF translocation observed in rotenone-intoxicated cells resulted from high ROS levels.
CoQ10 localizes in the mitochondrial inner membrane and transports electrons from complexes I and II to III. In addition, CoQ10 possesses antioxidant effects \[[@B26-ijms-15-13388]\]. CoQ10 protects cells against ultraviolet B-induced cell death and glutamate cytotoxicity by ameliorating ROS production and inhibiting the release of cytochrome *c* and activation of capase-3 in cultured cells \[[@B27-ijms-15-13388],[@B28-ijms-15-13388]\] and improves neurological functional performance in a rat model of Parkinson's disease \[[@B29-ijms-15-13388]\]. In present study, we have demonstrated that water-soluble CoQ10 suppresses ROS production, inhibites AIF nuclear translocation, preventes the fall of mitochondrial membrane potential and ameliorates cell death induced by rotenone. CoQ10 may block mitochondrial AIF release through two mechanisms. Firstly, CoQ10 reduces the ROS production, thereby preventing the formation of MPTP. Secondly, CoQ10 may directly bind to the MPTP and inhibit its opening.
3. Experimental Section {#sec3-ijms-15-13388}
=======================
3.1. Chemicals
--------------
Dulbecco's modified Eagle's medium (DMEM), phosphate buffered saline (PBS), Fetal bovine serum (FBS), Trypsin-Versene Mixture, [l]{.smallcaps}-Glutamine and Penicillin-Streptomycin solution were purchased from HyClone laboratories (GE Healthcare Life Sciences, Logan, UT, USA). Water-soluble CoQ10 was purchased from Zymes (Zymes LLC., Hasbrouck Heights, NJ, USA). Rotenone was purchased from Sigma-Aldrich (St. Louis, MO, USA).
3.2. Rotenone and CoQ10 Incubation
----------------------------------
Murine hippocampal HT22 neuronal cells were cultured in DMEM containing 10% FBS, 2 mM glutamine, and 200 mM streptomycin/penicillin and then maintained at 90%--95% relative humidity in 5% CO~2~ at 37 °C. The culture medium was renewed every 2 days. Cells were seeded in 96-well cell culture plates or T-75cm^2^ flasks and incubated in the above medium for at least 24 h in a CO~2~ incubator to reach 80% confluence. Water-soluble CoQ10 (50, 100, 200 μM) were given 3 h prior to rotenone (0.25, 0.5, 1.0, 2.0 μM) incubation and maintained for 24 h post-rotenone treatment. Cells treated with 0.1% DMSO were served as vehicle controls.
3.3. Cell Viability Assay
-------------------------
Alamar blue (Thermo Fisher Scientific, Waltham, MA, USA) was used to determine cell viability. Viable cells take up the dye and the fluorescent intensity could be detected using a plate reader. After treatment, alamar blue reagent (100 µg/mL) was added directly to cell culture medium and incubated for 3 h at 37 °C in a CO~2~ incubator. The plates were read using PHERA Star Microplate Reader (BMG Labtech, Ortenberg, Germany) at the excitation and emission wavelengths of 520 and 570 nm, respectively.
3.4. Measurements of Superoxide
-------------------------------
Intracellular ROS (superoxide anion) production was measured using dihydroethidine (DHE) fluorescent probe (Life Technologies, Grand Island, NY, USA). Briefly, prior to the termination of each experiment, cells (2 × 10^6^/mL) were incubated with DHE (2.5 μM) for 30 min at 37 °C. The cells were washed, resuspended in PBS and fluorescent intensity was read using a Fluoromax-4 spectrofluorometer (HORIBA Jobin Yvon Inc., Edison, NJ, USA) at the excitation and emission wavelengths of 480 and 590 nm, respectively. Data were presented as relative fluorescence intensity (RFI).
3.5. Measurements of Mitochondrial Membrane Potential
-----------------------------------------------------
Mitochondrial membrane potential was measured using the tetramethyl rhodamine methyl ester (TMRM). TMRM is a cell permeable cationic fluorescent dye that is readily sequestered by active mitochondria. Briefly, HT22 cells were grown on 96-well black/clear plate (BD Biosciences, San Jose, CA, USA) at a density of 5000 cells per well. After treatment with rotenone for 16 h, the cells were incubated with 100 nM TMRM (Life Technologies) and 10 µg/mL Hoechst (Life Technologies) at 37 °C for 30 min. Cells were washed with PBS twice and covered with 200 µL PBS. Stained cells were placed on the stage of Cell INSiGHT NXT High Content Screening Platform (Thermo Fisher Scientific). Images were acquired with a 10× objective (Olympus Tokyo, Japan). TMRM was excited at 549 nm with LED excitation only. Up to 25 fields of view per well were captured. Image processing was performed to quantify the total TMRM fluorescent intensity per cell using Thermo Fisher Scientific HSC studio: Cellomics Scan Version 6.4.4. In brief, TMRM fluorescent intensity was counted in TMRM channel and followed by cell detection using the Hoechst/nuclear channel. The TMRM intensity was tabulated for each cell and then a well-level result was calculated by averaging the TMRM intensity for each cell in a particular well.
3.6. Western Blot Analysis
--------------------------
At 24 h following rotenone and CoQ10 treatments, cells were collected and lysed on ice in a lysis buffer containing 20 mM Tris pH 7.4, 10 mM KCl, 3 mM MgCl~2~, 0.5% NP40 and complete protease inhibitors (EMD Millipore, Billerica, MA, USA). The cell lysates were centrifuged at 500× *g* for 10 min and resulted in a supernatant (S1) and a pellet (P1). The S1 fraction was centrifuged at 20,000× *g* for 20 min and the resulting supernatant was used as a cytosolic fraction. The P1 fraction was washed twice with lysis buffer and resuspended in a lysis buffer containing 1% SDS. The resulting lysates were sonicated briefly with a ultrasonic cell disrupter (Misonix, Farmingdale, NY, USA) and then centrifuged at 20,800× *g* for 30 min. The resulting supernatants were designated as nuclear fractions. The purity of different cellular fractions has been previously demonstrated \[[@B30-ijms-15-13388]\] and also demonstrated in [Figure 7](#ijms-15-13388-f007){ref-type="fig"}. The nuclear fractions were used to detect AIF, mitochondrial and cytosolic fractions for cytochrome *c*, and cytosolic fractions for cleaved caspase-9. Protein contents from each sample were measured using Microplate BCA Protein Assay Kit (Thermo Scientific). Equal amount of protein (20 µg) was loaded into each lane, separated in 4%--12% NuPAGE gels (Invitrogen, Carlsbad, CA, USA), transferred to PVDF membranes (Millipore), and probed with antibodies against: AIF (Cell Signaling, 1:500 dilution), cleaved caspase-9 (Cell Signaling, 1:1000 dilution), cytocrome *c* (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), Lamin B~1~ (Life Technologies, 1:1000 dilution), VDAC (Cell Signaling, 1:1000 dilution), and β-Actin (Cell Signaling, 1:1000 dilution), Histone H3 (Cell Signaling, 1:500 dilution), GAPDH (Abcam, 1:10,000 dilution), COX IV (Cell Signaling, 1:800 dilution). Blots were imaged using a LI-COR Biosciences Odyssey Infrared Fluorescent scanner (Lincoln, NE, USA). Both target protein bands and internal loading control protein bands were measured using LI-COR software and presented as a ratio of target protein band fluorescent intensity/loading control protein band intensity.
{ref-type="sec"} and equal amount of proteins (20 µg) were loaded side by side on the same gel. The membrane was probed separately with three primary antibodies against GAPDH, COX IV and Histone H3. A single band for GAPDH at approximately 37 kD in the cytosolic fraction, an obvious band for COX IV at approximately 17 kD in the mitochondrial fraction, and a band for Histone H3 closing to 17 kD in the nuclear fraction was observed. There was no obvious contamination among different fractions.](ijms-15-13388-g007){#ijms-15-13388-f007}
3.7. Statistical Analysis
-------------------------
Data are presented as means ± s.d. One-way ANOVA followed by Newman-Keuls test was used for statistical analysis. A value of *p* \< 0.05 was considered statistically significant.
4. Conclusions
==============
Rotenone is capable of inducing AIF-mediated cell death. Water-soluble CoQ10 pretreatment protects HT22 cells against rotenone-induced apoptosis via decreasing ROS formation, preventing mitochondrial membrane potential depolarization and inhibiting AIF nuclear translocation.
PAL is supported by the National Institute of Health (7R01DK075476). The BRITE is partially funded by the Golden Leaf Foundation.
Conceived and designed the experiments: Haining Li, Guisheng Chen, Ping-An Andy Li. Performed the experiments: Haining Li, Wanrui Ma. Contributed reagents/materials/analysis tools: Ping-An Andy Li. Wrote the manuscript: Haining Li, Ping-An Andy Li.
The authors declare no conflict of interest.
| {
"pile_set_name": "PubMed Central"
} |
Image in medicine {#sec1}
=================
A 40-year-old woman, without medical or surgical history, presented in our department for a gradually enlarged mass exteriorized from vagina (figure A). There was no history of vaginal discharge, lower abdominal pain or urinary complaints. On speculum examination, a pedunculated, non-tender, brownish and elastic tumor measuring 8 x 2 cm arising from the anterior cervix labium was noted (figure B), suggestive of a fibroepithelial polyp or a soft tissue tumors. The cervix was macroscopically normal. Under spinal anesthesia, a circular incision was made around the cervix implantation of the lesion with clear margins (about 1cm of the normal tissue) forcing excision of the mass. Histopathological examination revealed a lobulated tumor containing prominent myxoid stroma and thin walled vessels, with spindle to stellate tumor cells in higher power view; there were no atypical or mitotic figures in the tumor cells. The pathologic report was benign superficial angiomyxoma. The postoperative course was smooth and she was discharged 1 day after surgery. No recurrence was noted during a postoperative follow-up period of 2 years. Superficial Angiomyxoma is a benign mesenchymal tumor which rarely occurs in cervix, but it should be considered in differential diagnosis of cervical lesions, especially in women of reproductive age.
{#f0001}
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
As occurs with other chronic inflammatory conditions, ankylosing spondylitis (AS) is associated with a process of accelerated atherosclerosis \[[@CR1], [@CR2]\] which leads to increased rates of subclinical atherosclerosis \[[@CR3]\] and cardiovascular (CV) events \[[@CR4]\]. A recent meta-analysis of seven longitudinal studies reported an increased frequency of myocardial infarction (odds ratio (OR) 1.60, 95% confidence interval (CI) 1.32--1.93) and stroke (OR 1.50, 95% CI 1.39--1.62) in AS patients when compared to the general population \[[@CR4]\]. In a population-based study which included 21,473 AS patients and 86,606 controls matched for age, sex, and location of residence, Haroon et al. \[[@CR5]\] found a 36% higher risk of vascular mortality in AS.
Primary prevention strategies designed to avoid atherosclerosis-related CV events in the general population are based on the identification of individuals at high CV risk who can benefit from appropriate prevention measures \[[@CR6]\]. In this regard, statin use has proved to be very effective, being able to reduce overall mortality by 15% \[[@CR7]\]. The benefit may be even greater in AS patients according to a recent study which found a 37% lower risk of all-cause mortality associated with statin initiation \[[@CR8]\].
The 2016 "European Guidelines on CV disease prevention in clinical practice" recommended using the total cholesterol systematic coronary risk evaluation (TC-SCORE) to stratify the CV risk and to identify individuals at high CV risk who were candidates for treatment \[[@CR6]\]. The TC-SCORE predicts the individual's absolute risk for fatal CV events considering age, sex, total cholesterol levels, smoking, and blood pressure, stratifying the CV risk into low (\< 1%), moderate (≥ 1 and \< 5%), high (≥ 5 and \< 10%), and very high (≥ 10%).
Unfortunately, predictive models designed for the general population do not consider the inflammatory process as a proatherogenic factor, leading to an underestimation of the CV risk when used in inflammatory diseases like AS. In this regard, a recent study disclosed in AS patients a 10-year cumulative incidence of CV events three times higher than that predicted based on the Framingham Risk Score (FRS) \[[@CR9]\].
In an attempt to improve the TC-SCORE predictions, the 2016 European Society of Cardiology (ESC) guidelines proposed the use of different imaging techniques to identify high-CV risk individuals with subclinical atherosclerosis \[[@CR6]\]. Carotid ultrasonography (US) and multidetector coronary tomography (MDCT) allow one to detect, respectively, carotid plaques and coronary calcifications, which are considered independent predictors of CV events capable of providing additional value to the FRS when estimating CV risk \[[@CR10], [@CR11]\]. According to the 2016 ESC guidelines, the presence of both surrogate markers of atherosclerosis automatically implies a very high CV risk \[[@CR6]\]. Our group has recently demonstrated that carotid US and MDCT are very useful to redefine the CV risk in AS: up to 63% and 30% of our AS patients with moderate CV risk according to the SCORE risk charts had carotid plaques and coronary calcifications, respectively \[[@CR12]\]. However, the restricted availability of MDCT in the daily clinical practice and the considerable amount of radiation associated with the use of this technique may be limitations for to its generalized use.
Abdominal aortic calcium (AAC) constitutes another surrogate marker of atherosclerosis easily detectable in a lateral lumbar spine radiography \[[@CR13]\]. Unlike MDTC, it is available in most AS patients. Aortic calcification has been demonstrated to represent true intimal atherosclerosis in postmortem studies \[[@CR14]\], correlating with the degree of atherosclerosis in the coronary arterial beds \[[@CR15]\]. As occurs with carotid plaques and coronary calcification, AAC deposits constitute an independent predictor of subsequent vascular morbidity and mortality capable of improving the risk prediction based on the FRS \[[@CR16]\].
Taking all these considerations into account, the purpose of the present study is to determine whether, besides carotid US, the presence of ACC deposits detected by plain lumbar spine radiographs is able to improve the identification of those AS patients with high/very high CV risk who, therefore, would be candidates to receive intensive preventive therapy.
Methods {#Sec2}
=======
Patients {#Sec3}
--------
In this cross-sectional study, a set of 125 consecutive AS patients seen over a 5-year period at Hospital Universitario Marqués de Valdecilla and Hospital de Laredo (Cantabria, northern Spain) who fulfilled definitions for AS according to the 1984 modified New York criteria \[[@CR17]\] were recruited. Patients with a history of CV events (ischemic heart disease, cerebrovascular accident, peripheral arterial disease, or heart failure) were excluded. This was also the case for those with type 2 diabetes mellitus or with two fasting plasma glucose levels on different days at the time of disease diagnosis or over the extended follow-up \> 125 mg/dl as well as those with chronic kidney disease (glomerular filtration rate \< 60 ml/min/1.73 m^2^) because they are considered as having high or very high CV risk according to current guidelines.
Two clinical indexes of disease activity (Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and Ankylosing Spondylitis Disease Activity Score (ASDAS)), a functional status index (Bath Ankylosing Spondylitis Functional Index (BASFI)), a metrologic index (Bath Ankylosing Spondylitis Metrology Index (BASMI)), and an enthesitis index (Maastricht Ankylosing Spondylitis Enthesitis Score (MASES)) were evaluated in all patients at the time of the carotid US assessment \[[@CR18]--[@CR22]\].
Information on history of hip involvement, synovitis, enthesitis, extra-articular manifestations (anterior uveitis, psoriasis, and inflammatory bowel disease), syndesmophytes, and HLA-B27 status was also assessed. This was also the case for data on family history of early CV events in first-degree relatives, waist circumference, body mass index, blood pressure at the time of study, and history of traditional CV risk factors (smoking, hypertension, dyslipidemia, and obesity).
Data on C-reactive protein (CRP) and the erythrocyte sedimentation rate (ESR) at the time of recruitment and at disease diagnosis, identification of patients with CRP serum levels higher than 3 mg/L at the time of diagnosis, and total cholesterol, HDL-cholesterol, LDL-cholesterol, and triglycerides at the time of the study were also assessed. Information on therapy including treatment with anti-tumor necrosis factor (anti-TNF) alpha agents from the disease diagnosis was also reviewed.
The TC-SCORE system estimates the 10-year risk of a first fatal atherosclerotic event, whether heart attack, stroke, or other occlusive arterial disease, including sudden cardiac death. Risk estimates have been produced as charts for high and low-risk regions in Europe \[[@CR6]\]. Spain was included in the low-risk region of Europe. The risk factors incorporated in the TC-SCORE are those previously described: age, gender, smoking, total cholesterol levels, and systolic blood pressure. Subjects with TC-SCORE \< 1% are included in the category of low risk. Those with a TC-SCORE ≥ 1% and \< 5% are in the category of moderate risk. When the chart TC-SCORE result is ≥ 5% and \< 10% they are classified as having high risk. Finally, those patients with TC-SCORE results ≥ 10% are included in the category of very high CV risk.
The earliest age at which the CV risk scores should be used in the general population has not been rigorously established. Both European \[[@CR6]\] and North American \[[@CR23]\] guidelines recommend their application in individuals over 40 years old. However, this cutoff point is not uniform and varies across the different CV risk scores: 30 years old for the FRS \[[@CR24]\], 40 years old for the TC-SCORE \[[@CR6]\], and 45 years old for the Reynolds Risk Score \[[@CR25]\], which can only be applied from this age onward. In Spain, the Framingham-based REGICOR adapted function, a CV risk function validated in the Spanish population \[[@CR26]\], established its use at an intermediate point of 35 years old. Since patients with AS have early accelerated atherosclerosis, and we observed carotid plaques in patients under 40 years old, we included in the analysis all patients who were age 35 years and older.
Carotid US examination {#Sec4}
----------------------
A set of 125 AS patients underwent carotid US examination to detect focal plaques in the extracranial carotid tree. According to the Mannheim Carotid Intima-media Thickness and Plaque Consensus, a carotid plaque was defined as a focal protrusion in the lumen at least cIMT \> 1.5 mm, protrusion at least 50% greater than the surrounding cIMT, or arterial lumen encroaching \> 0.5 mm \[[@CR27]\]. Carotid US was performed using a commercially available scanner (Mylab 70; Esaote, Genoa, Italy) equipped with a 7--12 MHz linear transducer and the automated software-guided radiofrequency technique Quality Intima Media Thickness in real time (QIMT; Esaote, Maastricht, Holland). Patients with carotid plaques were considered as having very high CV risk.
Evaluation of abdominal aortic calcium deposits {#Sec5}
-----------------------------------------------
The lateral lumbar spine radiography was performed in the standing position and all X-ray scans were evaluated by an independent reader blinded to the participant's clinical status. We assessed the presence of calcific deposits at each vertebral segment from the first to fourth lumbar vertebrae, both in the posterior and anterior walls of the aorta (Fig. [1](#Fig1){ref-type="fig"}). Calcific deposits were regarded as present if densities were visible in an area parallel to the lumbar spine and anterior to the lower part of the spine. Aortic densities at the upper part of the lumbar spine (L1--L2 region) often overlapped the vertebrae as the abdominal aorta in the L1--L2 region is often lateral to the spine. Densities overlapping the vertebrae were deemed present only if they extended from or formed a clear pattern with those of the lower part of the aorta. Since calcific deposits tend to occur first in the lower aorta \[[@CR28], [@CR29]\], calcific deposits at the upper levels were almost always accompanied by extensive calcifications in the lower part of the aorta. Other calcific deposits visible in lumbar films, such as intestinal calcifications and calcified costal cartilages, were clearly distinguishable from aortic calcifications (Fig. [1](#Fig1){ref-type="fig"}).Fig. 1Abdominal aortic calcifications and calcified costal cartilages. Abdominal aortic calcifications seen as calcium deposits localized in area parallel to lumbar spine and anterior to lower part of spine (long arrow). Calcified costal cartilages have a typical pattern distinguishable from aortic calcifications (arrowhead)
The reproducibility of the evaluation of AAC deposits was also evaluated in 20 patients by two investigators (JR-G and AC). The correlation coefficient for AAC was 0.96.
The subject's written consent was obtained for all of the cases. The study was approved by the local ethical committee.
Statistical analysis {#Sec6}
--------------------
Categorical variables were described as percentages and quantitative variables as mean ± standard deviation (SD) or median (interquartile range (IQR)).
Correlation between the presence of carotid plaques and lumbar aortic calcification was analyzed using the Pearson correlation. For each CV risk model, sensitivity, specificity, percentage of correctly classified patients, and area under the receiver operating characteristic (ROC) curve (with 95% CI) were estimated.
Results {#Sec7}
=======
Characteristics of AS patients {#Sec8}
------------------------------
The main features of the 125 AS patients included in the study are summarized in Table [1](#Tab1){ref-type="table"}.Table 1Features of 125 ankylosing spondylitis (AS) patientsVariableAS (*n* = 125)Men/women, *n*78/47Age at time of study (years), mean ± SD48.3 ± 9.6Age at time of diagnosis (years), mean ± SD39.5 ± 9.6HLA-B27 positive, *n* (%)92 (74.8)Syndesmophytes, *n* (%)51 (41.5)History of synovitis, *n* (%)35 (28.0)History of enthesitis, *n* (%)45 (36.0)Extra-articular manifestations, *n* (%)38 (30.4) Psoriasis12 (9.6) Inflammatory bowel disease10 (8.0) Uveitis22 (17.6)Therapy with TNF inhibitors, *n* (%)49 (39.8)ASDAS, mean ± SD2.4 ± 0.9BASFI, mean ± SD3.9 ± 2.5BASMI, mean ± SD3.2 ± 1.7MASES, median (IQR)1 (0.0--4.0)BASDAI, mean ± SD3.7 ± 2.0CRP (mg/l), median (IQR) At time of study5.2 ± 6.7 At time of disease diagnosis10.8 ± 19.0CRP \> 3 mg/L at time of disease diagnosis, *n* (%)72 (57.6)ESR (mm/1st hour), median (IQR) At time of study12.4 ± 13.4 At time of disease diagnosis16.1 ± 15.9History of classic cardiovascular risk factors, *n* (%) Current smokers37 (29.6) Ex-smokers31 (24.8) Obesity23 (18.4) Dyslipidemia40 (32.0) Hypertension21 (16.8)Blood pressure (mmHg), mean ± SD Systolic130.6 ± 15.3 Diastolic79.5 ± 9.7Cholesterol and triglycerides (mg/dl), mean ± SD Total cholesterol200.7 ± 34.5 HDL cholesterol55.6 ± 14.8 LDL cholesterol124.6 ± 31.7 Triglycerides97.3 ± 49.4Carotid plaques, *n* (%)55 (44.0)Aortic calcification, *n* (%)28 (22.4)TC-SCORE ≥ 5, *n* (%)7 (5.6)Main epidemiologic, clinical, radiographic, and ultrasonography features of a series of 125 AS patients older than 35 years of age without history of cardiovascular events, diabetes mellitus, or chronic kidney disease*ASDAS* Ankylosing Spondylitis Disease Activity Score, *BASDAI* Bath Ankylosing Spondylitis Disease Activity Index, *BASFI* Bath Ankylosing Spondylitis Functional Index, *BASMI* Bath Ankylosing Spondylitis Metrology Index, *CRP* C-reactive protein, *ESR* erythrocyte sedimentation rate, *HDL* high-density lipoprotein, *IQR* interquartile range, *LDL* low-density lipoprotein, *MASES* Maastricht Ankylosing Spondylitis Enthesitis Score, *SD* standard deviation *TC-SCORE* total cholesterol systematic coronary risk evaluation, *TNF* tumor necrosis factor
Men outnumbered women (*n* = 78; 62.4%) and the mean ± SD age at the time of the study was 48.3 ± 9.6 years. HLA-B27 was positive in 92 (74.8%) patients, and syndesmophytes were present in 51 (41.5%) patients. Thirty-eight (30.4%) patients had extra-articular manifestations: uveitis, psoriasis, and inflammatory bowel disease were present in 17.6%, 9.6%, and 8% of patients, respectively. The mean ± SD values of the BASDAI and ASDAS were 3.7 ± 2.0 and 2.4 ± 0.9. Seventy-two (57.6%) patients were found to have CRP \> 3 mg/L at the time of disease diagnosis. TNF inhibitors were used in 39.8% of cases. Regarding findings of subclinical atherosclerosis, carotid plaques and AAC deposits were found in 55 (44.0%) and 28 (22.4%) patients, respectively. Other characteristics of this series of patients with AS are presented in Table [1](#Tab1){ref-type="table"}.
TC-SCORE risk groups and severe atherosclerotic disease using carotid US and lateral spine radiography {#Sec9}
------------------------------------------------------------------------------------------------------
The CV risk was calculated using the TC-SCORE. Based on this algorithm, patients were classified into four different CV risk categories: low, moderate, high, and very high. Then, the frequency of carotid plaques in each category was calculated to assess the ability of the TC-SCORE to correctly classify patients as having high/very high CV risk (Table [2](#Tab2){ref-type="table"}). We also analyzed the prevalence of AAC in the lateral spine radiography in each group of risk (Table [2](#Tab2){ref-type="table"}). Following this approach, 33 of 54 (61.1%) patients classified as having moderate CV according to the TC-SCORE had plaques when the carotid US was performed, whereas 21 of 54 (38.9%) patients with moderate CV risk according to the TC-SCORE also had AAC in the lateral lumbar spine radiography. Moreover, 23.4% and 4.7% of AS patients who fulfilled the category of low CV risk according to the TC-SCORE had carotid plaques and AAC, respectively (Table [2](#Tab2){ref-type="table"}).Table 2Prevalence of carotid plaques and abdominal aortic calcium in the different groups of cardiovascular riskTC-SCORECarotid ultrasonographyLateral lumbar X-rayCarotid plaques\
(*n* = 55, 44%)AAC deposits\
(*n* = 28, 22.4%)Low (\< 1%)*n* = 6415/64 (23.4%)3/64 (4.7%)Moderate (≥ 1% and \< 5%)*n* = 5433/54 (61.1%)21/54 (38.9%)High (≥ 5% and \< 10%)*n* = 77/7 (100%)4/7 (57.1%)Very high (≥ 10%)*n* = 00 (0%)0 (0%)Presence of carotid plaques and AAC deposits in 125 ankylosing spondylitis patients older than 35 years of age without cardiovascular events, diabetes mellitus, or chronic kidney disease, classified according to their cardiovascular risk*AAC* abdominal aortic calcium, *TC-SCORE* total cholesterol systematic coronary risk evaluation
Correlation between AAC and carotid US in AS patients {#Sec10}
-----------------------------------------------------
The validity of lateral lumbar spine radiography to identify AS patients at high CV risk was assessed by analyzing the correlation between AAC and carotid plaques (Table [3](#Tab3){ref-type="table"}). We observed that 25 of 28 (89.3%) patients with AAC deposits also had plaques in the carotid US assessment. Regarding patients without AAC deposits in the lateral lumbar spine radiography, 69.1% of them did not have carotid plaques either (Table [3](#Tab3){ref-type="table"}). We observed a positive correlation between AAC and carotid plaques (*r*^2^ = 0.49, *p* \< 0.001).Table 3Correlation between abdominal aortic calcium (AAC) deposits and carotid plaquesLateral lumbar X-rayCarotid ultrasonographyPresence of carotid plaques\
(*n* = 55/125, 44%)Absence of carotid plaques\
(*n* = 70/125, 56%)Presence of AAC*n* = 28 (22.4%)25/28 (89.3%)3/28 (10.7)Absence of AAC*n* = 97 (77.6%)30/97 (30.9%)67/97 (69.1%)Correlation between the presence of AAC deposits and the presence of carotid plaques in 125 patients with ankylosing spondylitis older than 35 years of age without cardiovascular events, chronic kidney disease, or diabetes mellitus
Model to establish the presence of high/very high CV risk in patients with AS {#Sec11}
-----------------------------------------------------------------------------
Since many patients categorized as having low or moderate CV risk when the TC-SCORE was applied had subclinical atherosclerosis, we set up a predictive model to identify AS patients with high/very high CV risk (Table [4](#Tab4){ref-type="table"}). In this regard, according to the ESC 2016 guidelines, we classified patients as having high/very high CV risk if they had carotid plaques in the carotid US \[[@CR6]\]. Following this approach, 55 of 125 patients fulfilled definitions of high/very high CV risk. However, only 7 (12.7%) of them were detected using the TC-SCORE, without the inclusion in the model of results of imaging techniques (model 1). The detection of aortic atherosclerosis by lateral lumbar spine radiography yielded a high specificity (95.7%) with low sensitivity (45.5%) to identify AS patients with very high CV risk (model 2). A predictive model that included a TC-SCORE ≥ 5% or the presence of AAC in lateral lumbar spine radiography in patients with moderate CV risk according to the SCORE (≥ 1% and \< 5%) (model 3) showed the same results as those found in model 2. A higher sensitivity (50.9%) with the same specificity (95.7%) was achieved when the predictive model included TC-SCORE ≥ 5% or the presence of AAC in the lateral lumbar spine radiography in patients with both moderate and low CV risk according to the TC-SCORE (\< 5%) (model 4).Table 4Diagnostic models designed to identify ankylosing spondylitis patients with very high cardiovascular risk using presence of carotid plaques as the gold standard testSensitivity (%)Specificity (%)Correctly classified (%)ROC (95% CI)Model 1. TC-SCORE ≥ 5%12.710061.60.56 (0.52--0.61)Model 2. Lateral lumbar spine radiography (presence of AAC)45.595.773.60.71 (0.64--0.78)Model 3. TC-SCORE ≥ 5% or TC-SCORE ≥ 1% and \< 5% plus lateral lumbar spine radiography (presence of AAC)45.595.773.60.71 (0.64--0.78)Model 4. TC-SCORE ≥ 5% or TC-SCORE \< 5% plus lateral lumbar spine radiography (presence of AAC)50.995.776.00.73 (0.66--0.80)*AAC* abdominal aortic calcium, *CI* confidence interval, *ROC* receiver operating characteristic, *TC-SCORE* total cholesterol systematic coronary risk evaluation
Discussion {#Sec12}
==========
To the best of our knowledge this is the first study that aimed to assess the ability of the lateral lumbar spine radiography, a diagnostic tool widely available in AS patients, to improve the CV risk stratification in patients with AS. We observed that the presence of AAC deposits detected by a lateral lumbar spine radiography allows one to identify individuals with AS at high risk of CV disease. In this regard, in the present series of 125 AS patients without CV events, chronic kidney disease, or diabetes mellitus, 21 of 54 (38.9%) patients with moderate TC-SCORE and 3 of 64 (4.7%) patients with low TC-SCORE showed AAC in the lateral lumbar spine radiography.
The presence of AAC in a plain radiography has been reported to be associated with an increased risk of CV events, CV mortality \[[@CR13]\], and congestive heart failure \[[@CR30]\] in the Framingham cohort. Overall, AAC adds to the prediction for intermittent claudication (IC), ischemic stroke (IS), and coronary heart disease over and above traditional CV risk factors in patients with low and intermediate risk \[[@CR16]\]. AAC deposits also demonstrated to be a predictor of incident stroke in the Rotterdam study, even stronger than carotid plaques \[[@CR31]\]. Both the Rotterdam and the Framingham studies used a quantitative scale to grade the severity of the aortic calcium deposits. It is worth noting that, even though the increase of CV risk was found to be proportional to the extent of the calcifications, AAC demonstrated to be a predictor of CV events and mortality also in cases of reduced calcifications. In keeping with these findings, in a Netherlands population Witteman et al. \[[@CR32]\] found a strong independent association between CV deaths and aortic calcifications of any size, regardless of the extent of the calcific deposits. A postmortem study that analyzed the accuracy of the radiologic AAC for the diagnosis of true aortic atherosclerosis confirmed that even the smaller densities indicated the presence of advanced atheromatous plaques, which were almost invariably ulcerated \[[@CR14]\]. Witteman et al. \[[@CR32]\] also disclosed a high predictive role of AAC for CV death in young individuals, with a six-fold increased risk in men aged 45 years independently of the major CV risk factors, and no excess risk at age 75 years. These findings are particularly relevant in AS patients since the disease typically starts before the age of 45 years and the TC-SCORE tends to underestimate the CV risk, especially in this age group in which practically all patients are classified as having low or moderate CV risk. Indeed, the mean age in our series was 48 years and 118 of the 125 (94%) AS patients were found to have a low or moderate TC-SCORE.
The 2014 ESC guidelines on the diagnosis and treatment of aortic diseases \[[@CR33]\] recommended adopting general preventive measures to control risk factors in the presence of aortic atherosclerosis. Treatment with statins led to regression of thoracic aortic and retardation of abdominal aortic atheroma burden assessed by magnetic resonance imaging \[[@CR34]\], as well as reduction of the inflammation assessed by PET \[[@CR35]\]. In a retrospective study of 519 patients with severe aortic plaque seen on transesophageal echocardiography, only statin treatment was associated with a 70% lower risk of events \[[@CR36]\]. A correlation between AAC and atherosclerotic findings in the carotid and coronary arterial beds has previously been reported \[[@CR15], [@CR37]\]. This finding constitutes an additional argument to support statin use in AS patients with AAC. The presence of carotid plaques, considered a surrogate marker of subclinical atherosclerosis in the 2016 ESC guidelines, automatically implies a very high CV risk with the consequent indication of statin use. Taking into account these considerations, we can conclude that almost 50% of AS included in the categories of low and moderate CV risk according to the TC-SCORE risk algorithm would benefit from additional primary prevention measures, with special emphasis on the use of statins, if a plain lateral lumbar spine radiography is performed. This is a crucial aspect considering that statins have been shown to decrease mortality by 37% in AS, a figure double that observed in the general population \[[@CR8]\].
The major limitation of our study was the absence of prospective follow-up data for the studied patients. Because of this, we dealt with a surrogate outcome based on extrapolation of the data that might not be applicable to AS.
In an attempt to elucidate the best strategy to identify AS patients at high CV risk, we assessed the ability of lateral lumbar spine radiography to detect severe findings of subclinical atherosclerosis. A predictive model that included a TC-SCORE ≥ 5% or the presence of AAC in lateral lumbar spine radiography in patients with a low or moderate TC-SCORE (TC-SCORE \< 5%) allowed us to detect half of the AS patients with high/very high CV-risk, also demonstrating a high degree of specificity (95.7%).
Conclusions {#Sec13}
===========
When carotid US is not available, plain lateral lumbar spine radiography can be an easy and affordable tool to identify AS patients at high CV risk.
AAC
: Abdominal aortic calcium
AS
: Ankylosing spondylitis
ASDAS
: Ankylosing Spondylitis Disease Activity Score
BASDAI
: Bath Ankylosing Spondylitis Disease Activity Index
BASFI
: Bath Ankylosing Spondylitis Functional Index
BASMI
: Bath Ankylosing Spondylitis Metrology Index
CRP
: C-reactive protein
CV
: Cardiovascular
ESR
: Erythrocyte sedimentation rate
FRS
: Framingham Risk Score
IC
: Intermittent claudication
IQR
: Interquartile range
IS
: Ischemic stroke
MASES
: Maastricht Ankylosing Spondylitis Enthesitis Score
MDCT
: Multidetector coronary tomography
OR
: Odds ratio
ROC
: Receiver operating characteristic
SD
: Standard deviation
TC-SCORE
: Total cholesterol systematic coronary risk evaluation
TNF
: Tumor necrosis factor
US
: Ultrasonography
The datasets generated and/or analyzed during the current study are not publicly available because the individual privacy could be compromised, but are available from the corresponding author on reasonable request.
JR-G made substantial contributions to the conception and design of the manuscript. FG analyzed and interpreted the patient data. AC and RB were involved in acquisition and analysis of data. PF, VP, RE, and CM were involved in acquisition and interpretation of data. TP was involved in acquisition and analysis of data. CG-J revised the manuscript critically for important intellectual content. LR-R analyzed and interpreted the patient data. MAG-G was involved in the conception and drafting and revising the manuscript critically for important intellectual content. All authors read and approved the final manuscript.
Ethics approval and consent to participate {#FPar1}
==========================================
The subject's written consent was obtained in all cases. The study was approved by the local ethical committee.
Consent for publication {#FPar2}
=======================
Not applicable.
Competing interests {#FPar3}
===================
The authors declare that they have no competing interests.
Publisher's Note {#FPar4}
================
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
Nowadays, many of pathogenesis of diseases have been determined \[[@B32]--[@B34]\]. Methylation in the 5′ cytosine in the CpG dinucleotides is crucial a mechanism that regulates gene expression without changing DNA sequence and can be inherited to the offspring \[[@B1]\]. The promoter region contains various transcription factor binding motifs with numerous CpG dinucleotides. Some transcription factors are blocked by methylated CpG island resulting in inhibition of gene expression \[[@B2]\]. Somatic cell nuclear transfer (SCNT) technique is used to generate an identical genetic background offspring \[[@B3], [@B4]\]. However, SCNT cloning animals usually showed low survival rate and impropriate methylation reprogramming process \[[@B5]\]. This dilemma of SCNT animal may be caused by methylation controlled genes, such as imprinting genes \[[@B6]\].
*OPN* is an extracellular matrix protein and hydrophilic glycoprotein identified firstly in the bone as a sialoprotein. It contains a thrombin and transglutaminase cutting site, and the molecular weight is about 25 kDa to 75 kDa; in pig, the molecular is about 67 kDa; it contains numerous isoforms \[[@B7]\].*OPN* has a hydrophobic N terminal; thus, it can be secreted out of cell membrane; the amino sequence of*OPN* is full of Asp, Thr, and Ser that can elevate the binding activity with calcium, glycosylation, and phosphorylation, respectively \[[@B8]\]. Thus,*OPN* plays numerous roles in many aspects, such as bone remodeling, cell migration, iNOS regulation, repairment, and leucocyte recruitment \[[@B9]\]. And acquired*OPN* expression has been found in a variety of cancer cell types, especially in the liver, lung, breast, prostate, colon, brain, and spleen \[[@B10], [@B11]\].*OPN* is cleaved by MMPs protein to generate functional*OPN* that can bind to *α*v*β*3 \[[@B12]\]. This integrin binding with*OPN* has influence on NF*κ*B signaling transduction \[[@B13], [@B14]\]. Therefore, overexpressed*OPN* is associated with tumorigenesis, tumor invasion, and metastasis \[[@B15], [@B16]\]. Previous study suggested that overexpressed*OPN* induces the serious cardiac fibrosis \[[@B17], [@B18]\]. Thus, our cloned pigs were also surrounded by various defects in heart fibrosis and retardation of growth of bones. Therefore, this study focuses on the methylation change of*OPN* promoter that may be disrupted by inappropriate reprogramming process. Consequently, aberrant methylation of promoter could lead to aberrant expression of*OPN*. In the previous studies,*OPN* expression was induced with TSA (trichostatin A) in mouse undifferentiated mesenchymal cell line by AP1 site \[[@B19]\]. The TSA is a histone deacetylase inhibitor. It can lose the chromatin structure in order to let gene restore its expression. 5-aza-dc is also an analog with the same structure of cytosine without methyl group adding in the 5′C end \[[@B20]\].
Thus, 5-aza-dc addition leads to low methylation percentage in the CpG sites rich region. The hypomethylation status in the promoter may contribute its gene transcription activity. Porcine fetal fibroblasts in 5th passage cultures were treated with 0.5, 1.0, 2.0, and 3.0 *μ*M 5-aza-dc for 96 h; 5-aza-dc inhibited the growth of cell at all concentrations. 5-aza-dc induced a reduction of transcripts level in*DNMT1* and increasing expression in imprinted gene,*IGF2*\[[@B21]\]. Furthermore human*OPN* promoter sequence is similar to porcine in the front 400 nt of the porcine promoter. Therefore, we investigated*OPN* RNA and promoter methylation changes in the porcine ear fibroblast cell. Data showed that the elevated*OPN* expression and in 5-aza-dc treated fibroblast cell is due to the decreased methylation of*OPN* promoter. Cloned pigs samples had found extremely methylation changes, especially in the brain (99.75% upregulation), heart (11.50% down-regulation), and ear (18.03% down-regulation). Deletion analysis of the promoter region revealed 5-aza-dc induced luciferase response that was regulated by −2615 to −2239 of the*OPN* promoter. These data suggested that methylation in the*OPN* promoter plays a crucial role in the regulation of*OPN* expression. Methylation of*OPN* promoter may be an epigenetic marker of diagnosis of cancer.
2. Materials and Methods {#sec2}
========================
2.1. CpG Island Prediction {#sec2.1}
--------------------------
The sequence of a putative CpG island in*OPN*promoter was analysed by using MethPrimer software (<http://www.urogene.org/methprimer/index1.html>).
2.2. Cell Culture {#sec2.2}
-----------------
The porcine fibroblast cell line was grown in Dulbecco\'s modified Eagle\'s medium (DMEM; Gibco-BRL, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum (FBS, Gibco BRL) and containing 100 U/mL penicillin and streptomycin. The cells were incubated at 37°C in humidified incubator with 5% CO~2~.
2.3. 5-aza-dc Demethylation Drug Treatment {#sec2.3}
------------------------------------------
For 5-aza-dc treatment, porcine fibroblast cell in 5th passage cultures was treated with 5-aza-dc (sigma) at various concentrations, that is, 0 (control), 0.5, 1.5, and 2.0 *μ*M, for 72 h. Medium was changed every 24 h and then cells were collected for RNA and DNA extraction and stored at −80°C \[[@B21]\].
2.4. Quantitative Real Time-PCR {#sec2.4}
-------------------------------
2 *μ*g RNA of ear fibroblast cell was used to be transformed to cDNA. 0.5 *μ*L of cDNA was performed for quantitative real time-PCR with Rotor-Gene 6000 (Corbett). *β*-actin was the internal control for normalize target gene,*OPN*. The calculated gene expression fold from CT value was according to the previous study. *P* value less than 0.5 exhibited the obviously significant difference.
2.5. Methylation Analysis by Combined Bisulfite Restriction Analysis (COBRA) {#sec2.5}
----------------------------------------------------------------------------
For amplification of porcine*OPN* promoter methylation analysis site, PCR was performed using 2 *μ*L of bisulfite-converted genomic DNA as template. The primer sets of COBRA were*OPN*-C sense 5′-TTTTTTGAGGGAGATTAGTTTTTG-3′ and antisense 5′-ATTCTACTAAAATCCAACCACCC-3′. The COBRA-PCR products were purified by phenol/chloroform, followed by ethanol precipitation. The DNA was resuspended in 8.5 *μ*L of distilled deionized water. Purified PCR products were then digested with 10 U*BstU*I restriction enzyme (New England Biolabs, MA, USA) at 65°C. Products were electrophoresed on 6% native acrylamide gel, stained with 200 g/mL ethidium bromide, and visualized using a Kodak 1D software.
2.6. Methylation Specific-PCR {#sec2.6}
-----------------------------
Genomic DNA (0.5 *μ*g) was treated with sodium bisulfite according to the manufacture\'s recommendations (EZ DNA Methylation Kit; Zymo research, CA, USA) and amplified with specific primers for methylated or unmethylated DNA. The primer sets of MS-PCR were*OPN*-M sense 5′-AAGCGGGGAAGGAGTTATTACGT-3′, antisense 5′-TCCGACAAAACGAAACGATCATACA-3′,*OPN*-U sense 5′-GAAGTGGGGAAGGAGTTTATTATGT-3′, and antisense 5′-CAATAACTCCAACAAAACAAAACAATC-3′. All PCR reactions were performed on PTC 200 thermocyclers (MJ Research, MA, USA) and in 25 *μ*L volume using the PlatiumTaq DNA polymerase system (Invitrogen, CA, USA). PCR products were separated on 1.5% agarose gels. The M-set primers contained at least three CpG sites to distinguish the methylation status of investigated region. And U-set primers overlapping the M-set primers were used to amplify the unmethylated region.
2.7. Plasmid Constructs {#sec2.7}
-----------------------
A full length pig*OPN* promoter (−2615-luc) was amplified from wild-type pig heart tissue cDNA. This fragment was cloned into a luciferase fusion plasmid, pGL3-Enhancer vector (Promega), to generate p*OPN*-full-luc.*Hind*III and*Nco*I cutting sites were used for cloning. Three truncated forms of p*OPN* promoter were prepared by PCR using the p*OPN*-full-luc as a template and using synthesized oligonucleotides as follows: p*OPN*-full-luc: sense, 5′-AAGCTTGAATTCACTCGTCTTTCCTTTGAGA-3′, and antisense, 5′-CCATGGGCTGACAGCCTGGACCTCCCC-3′; −2239-luc: sense, 5′-AAGCTTCCTATAACTGTCTACGTTCATATTAGAC-3′, and antisense, 5′-CCATGGGCTGACAGCCTGGACCTCCCC-3′; −1505-luc: sense, 5′-AAGCTTAATTTTCATTTAAGTAACCAACTTTATATATC-3′, and antisense, 5′-CCATGGGCTGACAGCCTGGACCTCCCC-3′; −495-luc: sense, 5′-AAGCTTGCCTGAACAATATAGCCTTGTCGC-3′, and antisense, 5′-CCATGGGCTGACAGCCTGGACCTCCCC-3′. The sequence of constructs was confirmed by DNA sequencing. There were two-point mutation different from NCBI: one is 287A to T and the other is 957T to A.
2.8. Transient Transfection and Luciferase Assay {#sec2.8}
------------------------------------------------
Pig ear fibroblast cells were transfected using the Lipofectamine 2000 (Invitrogen). Fibroblast cells were incubated at a density of 8 × 10^5^ cells into 35 mm diameter dishes. After 24 h when cell was adherent to the dishes, 3 *μ*g of reporter plasmid DNA was transfected for 6 h in Lipofectamine mixture (Invitrogen). 24 h after the transfection, cell lysates were collected for a luciferase assay. The luciferase activity of the cell lysates was detected by Dual-light system (Applied biosystems). The activity data was measured with PARADIGM Detection Platforms (Beckman Coulter). Luciferase activity was normalized with 1 *μ*g *β*-gal plasmid. All luciferase assays were carried out in triplicate.
2.9. *In Vitro* Methylation of the*OPN* Promoter Region {#sec2.9}
-------------------------------------------------------
The*OPN* reporter construct −495-luc was methylated by incubation with*Sss*I methyltransferase (New England BioLabs). The −2615-luc construct was methylated by*Hha*I and*Hpa*II methyltransferase (New England BioLabs) for 16 h at 37°C. The methylation status was also verified by digested with*Hha*I and*Hpa*II enzyme.
2.10. Electrophoresis Mobile Shift Assays {#sec2.10}
-----------------------------------------
Nuclear extracts were prepared from HEK293T cells. Two probes were designed for methylation binding activity test. Two probes containing either 6--8 CpG sites or 13th CpG site in the −2615 to −2239 of the*OPN* promoter were generated by annealing two complementary oligonucleotides (*OPN 6--8*: 5′-TGCATGATCGTTCCGTCCTG[CCGGAGTCACT]{.ul}GACGGAACCAGACCGAGGT-3′; 5′-ACCTCGGTCTGGTTCCGTC[AGTGACTCCGG]{.ul}CAGGACGGAACGATCATGCA-3′, the predicted core sequence of the AP1 binding site is underlined;*OPN*13th: 5′-[CCTCCG]{.ul}TGTTCCCTGTTAATGTGTAGCGCGTCGTTGTTGGGAAATAGTTC-3′; 5′-GAACTATTTCCCAACAACGACGCGCTACACATTAACAGGGAACA[CGGAGG]{.ul}-3′; the predicted core sequence of the ADR1 binding site is underlined). The transcription factor prediction software is TFSEARCH 3.0 version. The probes were labeled with *γ*-^32^P-ATP by using T4 kinase (Promega). Annealing probes also were methylated with SssI methyltransferase (NEB). Nuclear extracts containing 5.6 *μ*g of the protein were preincubated in 20 *μ*L of binding buffer (50 mM Tris-HCl (pH 8.0), 750 mM KCl, 2.5 mM EDTA, 0.5% Triton-X 100, 62.5% glycerol (v/v), and 1 mM DTT) with or without unlabeled competitor (10-fold molar excess). For supershift assay, antibody of AP1 was added to the preincubation buffer. After 10 min of preincubation on ice, the DNA probe labeled with \[*γ*-^32^P\]-ATP was added, and the mixtures were incubated at room temperature for 30 min. The reaction mixtures were resolved on 6% polyacrylamide gels. The gels were dried and subjected to PhosphorImager analysis using a Typhoon system and ImageQuant TL software (Amersham Biosciences, Sunnyvale, CA, USA).
2.11. Molecular Modelling {#sec2.11}
-------------------------
The molecular docking was then further analyzed for proving the further mechanism of our findings; we further surveyed the interaction of AP1 (c-Jun) and TFIIB by computational biology. The AP1 (c-Jun) is possible higher spot for hypermethylation in*OPN* promoter area and provides the binding domain for RNA-polymerase II initial binding transcription factor (TFIIB) in this study. Therefore, we first utilized the Z-DOCK program to simulate the structures of c-Jun and TFIIB. After that, we further used molecular dynamics (MD) to validate the stability of c-Jun and TFIIB complex under the GROMACS 4.5.5 program \[[@B22]\] with charmm27 force field. The model is set in the TIP3P water modeling in 1.2 nm distance of box for water box setting. Na and Cl ions in the concentration of 0.145 M NaCl model are used for system neutralization. All bonds are fixed by linear constraint solver (LINCS) algorithm to constrain all bonds lengths in the simulation system. Newton\'s Law is utilized for calculating the motion of molecular dynamics as follows: $$\begin{matrix}
{\frac{d^{2}r}{dt} = M^{- 1}F.} \\
\end{matrix}$$
The Particle mesh Ewald (PME) is also used for calculate the coulomb type of electrostatics. The Van der Waals (VDW) interactions are set as 1.4 nm cut-off distance for nonbound interaction. The first step is set on the 5,000 cycle steps performed in the manner of Steepest Descent algorithm for energy minimization. And then, equilibration was performed with a time period of 1 ns for position restraints set under the constant temperature dynamics (NVT type) conditions. The third step is calculating the production run for 5000 ps under constant pressure and temperature dynamics (NPT type). All the MD systems are set by 310 K temperature during all simulation times. MD frames data were saved every 20 ps for all production runs.
2.12. Molecular Dynamics Analysis {#sec2.12}
---------------------------------
First we survey the stability of all atoms performed by using the GROMACS 4.5.5 software though the commands of g_rms and g_gyrate to calculate root mean square deviation (RMSD) and radius of gyration (Rg), respectively. Secondly, we calculate the total energy for all the systems by the command of g_energy. Thirdly, we further calculate root mean squared fluctuation (RMSF) for each protein residue by commands of g_rmsf. Fourthly, the distance between c-Jun and TFIIB and movement analysis are calculated by the g_dist program. Fifthly, the migration of dock protein (c-Jun and TFIIB) is presented by mean square displacement (MSD) under the command of g_msd module in GROMACS during all the simulation times. Sixthly, the g_cluster program is selected for further calculation of the representative structure from all MD frames, and the representative structure is taken for further snapshot analysis. We also employed DSSP analysis and matrices of the smallest distances between each residue to investigate the stability of the protein structure. The principle component analysis (PCA) is then applied to observe the protein motion changes during all the MD frames. Finally, in order to observe the compactness between c-Jun and TFIIB, Caver 3.0 software \[[@B23]\] was used to predicted space in the complex.
3. Results {#sec3}
==========
3.1. Distribution of Porcine*OPN* CpG Island {#sec3.1}
--------------------------------------------
There are dense CpG sites existing in the front of the*OPN* promoter region. One putative CpG island was found (CpG island size \> 100, GC Percent \> 50.0, Obs/Exp \> 0.6) by MethPrimer program.
3.2. Methylation and Expression Analysis of*OPN* in Cloned Pig Various Tissues {#sec3.2}
------------------------------------------------------------------------------
Firstly, the*OPN* mRNA expression was investigated in the WT pig tissues and cloned pig tissues. Data showed the various expression levels in different tissues. Particularly, Copy1 brain overexpressed the*OPN* and Copy2 ear with no expression of*OPN* relative to its wild-type tissue, respectively ([Figure 1](#fig1){ref-type="fig"}). The unique aberrant expression patterns exhibited the different control way of the*OPN* expression. We proposed that*OPN* expression may be a tissue-specific manner. MS-PCR primers were designed to estimate the methylation status of*OPN* promoter. Hypomethylation generally appeared in the various tissues of cloned pigs. However, there were still some tissues that showed the methylated region in*OPN* promoter ([Figure 2](#fig2){ref-type="fig"}).
3.3. 5-aza-dc Increases*OPN* mRNA and Decreases Methylation of*OPN* Promoter in Pig Ear Fibroblast Cell {#sec3.3}
-------------------------------------------------------------------------------------------------------
In order to realize whether the methylated*OPN* promoter affects the activity of*OPN* promoter. The 5-aza-dc treated porcine ear fibroblast cells showed that when the concentration of 5-aza-dc level increased, it will decrease the methylation of*OPN* promoter and restore the*OPN* RNA expression at 0.5 to 2.0 *μ*M ([Figure 3](#fig3){ref-type="fig"}). The results suggested that the activity of*OPN* promoter can be affected by DNA methylation directly or indirectly. COBRA assay was also used to investigate the methylation status of*OPN* promoter in WT tissues. Brain, ear, liver, and lung tissues exhibited little part methylation of the*OPN* promoter (data not show). This is thought that the methylation of*OPN* promoter in the aforementioned tissues may be involved with*OPN* transcript regulation mechanism.
3.4. Bisulfite Sequencing Analysis of the Whole CpG Sites Methylation Profile in Cloned Pigs {#sec3.4}
--------------------------------------------------------------------------------------------
To investigate which region of the*OPN* promoter is affected by methylation in the CpG site, bisulfite sequencing was performed to dissect the methylation status of CpG sites in the*OPN* promoter. In the brain tissue, bisulfite sequencing of Copy5 and Copy1 brain exhibited the saturated status in their methylated region while WT brain exhibited fragmentary methylated CpG sites. The 18.06% methylation percentage of Copy5 brain was more than the WT brain, 14.25% ([Figure 4](#fig4){ref-type="fig"}). In the heart tissues, bisulfite sequencing showed that the Copy2 heart had extremely hypermethylated percentage with 13.33% more than WT heart, 2.94% ([Figure 4](#fig4){ref-type="fig"}). Particularly, the data showed the inhibition of Copy2 heart mRNA ([Figure 1](#fig1){ref-type="fig"}). In the liver, the different methylation pattern also appeared in the Copy4 liver relative to WT liver. Copy2 ear with 18.05% methylation in the*OPN* analyzed region was higher than the WT ear, 11.67%; the result was proved by the experiment of*OPN* mRNA expression ([Figure 1](#fig1){ref-type="fig"}). These bisulfite sequencing results matched our previous hypothesis that methylation in the*OPN* promoter region regulates the activity of*OPN* promoter. Moreover, the hypermethylated*OPN* promoter may directly affect the activity of*OPN* transcription, especially in heart and ear tissues.
3.5. Analysis of Methylation Implication of*OPN* Transcription by Promoter Assay {#sec3.5}
--------------------------------------------------------------------------------
The analyzed*OPN*promoter region may involved with methylated control of gene transcription. Promoter assay was designed to explain the direct inhibition of*OPN* promoter activity by methylation on the*OPN* CpG sites. Four different truncated forms of*OPN* full length (2.6 kb) were used to prove the hypothesis ([Figure 5](#fig5){ref-type="fig"}). Particularly, the truncated form 2.2 kb deleted the 377 bp promoter region (−2615\~ −2239 nt). This region is the analyzed region for bisulfite sequencing profile. And this region is thought to be the most possible element that regulates the*OPN* transcription. Moreover, four CpG sites were*in vitro* added to the methyl group by methyltransferase that can provide important evidence how methylation affect the*OPN* transcription. [Figure 5(b)](#fig5){ref-type="fig"} shows that the*OPN* promoter activity was significantly decreased in methylated −2615-luc plasmid. However, −2239-luc that deleted the 377 bp containing methylated characteristic DNA element leads to less inhibition of promoter activity than the methylated vector −2615 M-luc (Figures [5(a)](#fig5){ref-type="fig"} and [5(b)](#fig5){ref-type="fig"}). The results indicated that methylation in the front*OPN* promoter is not only decreasing the promoter activity to the basal level but also recruiting the inhibition factors to enhance the inhibition ability. In order to avoid the effects of CpG sites in PGL3-enhancer backbone, −495 M-luc that have no methylated CpG sites in the*OPN* promoter but it can be methylated in the vector backbone CpG sites compared to −495-luc. Data showed that there is no difference in the promoter activity between methylation or unmethylation in the PGL3-enhancer backbone CpG sites ([Figure 5(b)](#fig5){ref-type="fig"}). It is suggested that methylation in the critical region, such as*OPN* promoter front end, may lead to the rearrangement of chromatin structure. Otherwise, deletion of the*OPN* promoter to the 495 bp with significant promotion of the promoter activity indicated that in the middle part of promoter DNA element may able to inhibit the activity of*OPN* promoter.
3.6. Methylation in CpG 13th and CpG 1st of*OPN* Promoter Blocks the Binding Access of Transcription Factors {#sec3.6}
------------------------------------------------------------------------------------------------------------
We investigate that the methylated CpG sites in the −2615\~−2239 nt of the*OPN* promoter region affect the transcription factor binding activity. Electrophoresis mobile shift assay was performed with nuclear extracts from human HEK293T and SH-SY5Y cell line. Four EMSA probes that contain the CpG 1, CpG 3-4, CpG 6--8, CpG 11--15, and CpG 19-20 were designed according to the CpG sites in our analyzed region ([Figure 6](#fig6){ref-type="fig"}). EMSA data suggested that CpG 13th and 1st sites showed methylation noncompetition phenomenon which had influence on binding with transcription factor ([Figure 7](#fig7){ref-type="fig"}). 19-20th CpG sites containing EMSA probe showed no competition ability in the methylated or unmethylated status. The premix with antibody and nuclear extracts by EMSA assay indicated that the c-Jun and c-Fos were involved in the binding to CpG sites 13 and 20 ([Figure 8](#fig8){ref-type="fig"}). However, the adding of antibody in the mixture of probe and nuclear extracts showed no significant shift bands. It is indicated that other transcription factors may also participate in the transcription activity*OPN* promoter. Thus, the c-Jun and c-Fos could be involved in the partial*OPN* transcriptional activity in a competition way. In addition, the c-Jun had higher binding affinity than c-Fos in the EMSA probe analysis. We therefore did followed computational survey to see why c-Juc affects the consequent*OPN* transcription.
3.7. The Computational Biology Results {#sec3.7}
--------------------------------------
After surveying the possible zone for*OPN* promoter hypermethylation, we found that the AP1 (c-Jun) sequence frequently appeared in our molecular laboratory study. Then we further validated that the c-Jun methylation will cease the further mRNA production by inhibiting the binding of RNA-polymerase II initiation factor TFIIB. We do the further computational modelling for mechanism survey between the c-Jun and TFIIB. By the Z-DOCK analysis, we found the c-Jun and TFIIB could combine tightly ([Figure 9](#fig9){ref-type="fig"}). We chose the highest docking pose (dock score = 18.04) for further MD analysis.
Then the disorder predication was employed to observe the protein folding analysis and the result is shown in [Figure 10](#fig10){ref-type="fig"}. We found that the c-Jun has relative high disorder in folding than TFIIB; we suppose that the flexibility of c-Jun structure could easier bound to TFIIB. This finding could be an explanation why the c-Jun bound to TFIIB by Z-DOCK program. To confirm the stability of the c-Jun and TFIIB complex, the series of molecular dynamic studies further visualize their interactions.
Protein complex RMSD analysis proved that the c-Jun and TFIIB were stable from 3000 ps to 5000 ps. In addition, we found the TFIIB are easier to be stable during the molecular dynamics ([Figure 11(a)](#fig11){ref-type="fig"}). We also found the radius of gyration tend to be stable for all simulation times with average of 2.05 nm ([Figure 11(b)](#fig11){ref-type="fig"}), suggesting that the two protein structures are compact after binding together. [Figure 11(c)](#fig11){ref-type="fig"} also shows the binding complex in a stable fluctuation and the energy of the binding complex is stable around −9.15 × 10^5^ (kJ/mol).
3.8. Stability Analysis of Residues on the Major Binding Region during MD Simulation {#sec3.8}
------------------------------------------------------------------------------------
To analyze the flexibility of residues on protein structure, the RMSF calculation was used to observe the flexibility of each residue; [Figure 12(a)](#fig12){ref-type="fig"} shows that the chain A of c-Jun had high frequency of fluctuation (binding site 200--210 binding resides). However, the chain B of the c-Jun has relative fewer frequency of fluctuation and the major binding region (from 228 to 240 residues) showed a less fluctuation as shown in [Figure 12(b)](#fig12){ref-type="fig"}. [Figure 12(c)](#fig12){ref-type="fig"} reveals the binding regions (200--210 binding residues) for chain A of c-Jun that is more fluctuated and unstable compared to the binding regions (228--240) for chain B of c-Jun binding site in the TFIIB binding region. [Figure 13](#fig13){ref-type="fig"} is the result for secondary structure variation calculated by DSSP analysis. Most of the main scaffold belong to alpha Helix; there are no significant changes during the whole MD simulation. All helices of the secondary structure for c-Jun and TFIIB binding remained stable during a 5,000 ps simulation time ([Figure 13](#fig13){ref-type="fig"}). We thereafter surveyed the distance between each residue of c-Jun for 5,000 ps. The variation of distances between residues in c-Jun chain is wider than the distances of residues in c-Jun chain B. Therefore, the chain B of c-Jun is more stable for TFIIB binding ([Figure 14](#fig14){ref-type="fig"}).
The hydrophobic area was then calculated by SASA in [Figure 15(a)](#fig15){ref-type="fig"}; the value of hydrophobic area decreased during the last 1000 ps. This indicated the compactness of the c-Jun and TFIIB binding increased by the MD time period in our study. It is worthy to know that the distance between centrals of masses of c-Jun and TFIIB was decreasing more and more after time goes by in the 5000 ps survey ([Figure 15(b)](#fig15){ref-type="fig"}). In the migration analysis of c-Jun and TFIIB, the MSD was employed to count the migration of c-Jun and TFIIB. The c-Jun is more unstable than TFIIB during the binding interaction throughout the whole MD simulation in 5000 ps period ([Figure 15(c)](#fig15){ref-type="fig"}). Besides, we further utilized the principal component analysis (PCA) to measure all MD frames over all simulation times. The first two eigenvectors (PC1 and PC2) were shown in [Figure 18](#fig18){ref-type="fig"}; most of frames are ranged in the short range of eigenvalues −10 and 10 in PC1 ([Figure 16(a)](#fig16){ref-type="fig"}), and arranged in eigenvalues from −5 to 5 in PC2 ([Figure 16(b)](#fig16){ref-type="fig"}). The phase space comparing for PC1 and PC2 was shown in [Figure 17](#fig17){ref-type="fig"}; we found that each frame could be grouped into two clusters. This suggests that the motion of each frame was not changed significantly over all simulation times. In order to select the most representative structure for snapshot investigation, we did cluster analysis ([Figure 18](#fig18){ref-type="fig"}). We found that the last group (cluster 14) is the predominant cluster and is also displayed in the time range from 4000 to 5000 ps; the cluster 14 also appears most predominant in frame numbers ([Figure 18](#fig18){ref-type="fig"}), and the middle structure (4260 ps) of cluster 14 is regarded as representative frame. For snapshot analysis, the comparison of initial and representative frames is shown in [Figure 19](#fig19){ref-type="fig"}; we found that the chain A of c-Jun is more encompassed by TFIIB at 4260 ps through the inward rotation of TFIIB. This made the bindings between chain A of c-Jun and TFIIB more compacted through 0 ps to 4260 ps along with time. The elevated activation between TFIIB and chain A of c-Jun is also confirmed in RMSF analysis. Therefore, we supposed the initiated transcription factor on RNA polymerase II (TFIIB) is closed interaction to the chain A of c-Jun (AP1) from 0 ps to 4260 ps. This hypothesis was also confirmed by [Figure 20](#fig20){ref-type="fig"}. There were more spaces between chain B and TFIIB than chain A. Hence, we could see that the TFIIB acts more close to the chain A of c-Jun ([Figure 20](#fig20){ref-type="fig"}). Overall, we presume that the initiation of*OPN* transcription started from TFIIB binding to chain A of c-Jun.
4. Discussion {#sec4}
=============
Previous study has shown that DNA element (GGGTCATATGGTTCA) located in osteopontin promoter −2245 to −2259 nt can be regulated by vitamin D3 \[[@B24]\]. This DNA regulation region can easily be affected by the change of calcium concentration. The promoter region of porcine*OPN* was analyzed in transcription factor binding sites except the region −2615 to −2239. Interestingly, this region of porcine*OPN* promoter is rich in CpG sites compared to human, mouse, and bovine genome. Sakata also proved that*OPN* promoter transcription activity is regulated by some specific DNA modification mechanism of rearrangement of chromatin structure \[[@B19]\].
In the present study, four cloned pigs were surrounded by many defects. For example, Copy1 pig had a retardation of limb bone growth. Copy2 heart organ showed a pericarditis and copy3 heart had valvular heart disease. This physiology defects appeared aberrant development, especially in bone or heart, may involved in the initially fetus stage with inappropriate organ differentiate. Thus, our data suggested that the consequent result in aberrant*OPN* expression or incompletely epigenetic modification in*OPN* promoter ([Figure 5](#fig5){ref-type="fig"}). These aberrant molecular data of*OPN* are correlated with the defects of bone and heart in cloned pigs. Semiquantitative PCR of*OPN* mRNA showed that discrepant expression pattern was identified in several cloned pig tissues, especially in brain (99.75% up-regulation), heart (11.5% downregulation), and ear (18.03% downregulation) ([Figure 2](#fig2){ref-type="fig"}).*OPN* mRNA has different expression in brain development in different embryonic stages \[[@B25]\]. The overexpression of*OPN* in brain tissue may cause some unexpected brain damage or neuron development.*OPN* can induce myocardial fibrosis and repair tissue after inflammation. Lacking*OPN* will cause faulty wound healing after myocardial infarction \[[@B26], [@B27]\]. Silent expression of*OPN*in cloned pig\' heart tissue may also be the main cause of heart disease.
Recent studies indicated that*OPN* gene expression may be affected by treatment of TSA (a histone deacetylase inhibitor). The results applied that*OPN* promoter could be regulated by epigenetic mechanism \[[@B19]\]. In this study, we investigated*OPN* methylation profile after 5-aza-dc treatment. The results indicated that mRNA expression of*OPN* is directly affected by adding methyltransferase inhibitor 5-aza-dc ([Figure 4](#fig4){ref-type="fig"}). COBRA was performed to study the methylation of*OPN* front end promoter in wild-type and cloned pig different tissues. Sodium bisulfite sequencing analysis also revealed that the methylation of CpG sites concentrated in front of the 20 CpG sites in front of*OPN* promoter ([Figure 5](#fig5){ref-type="fig"}). Discrepancy methylation in this promoter region also happened in brain, heart, ear, and liver tissues between wild-type and cloned pigs ([Figure 5](#fig5){ref-type="fig"}). It revealed that DNA methylation of*OPN* promoter may be involved with regulation of expression of*OPN* mRNA. In order to characterise which promoter DNA element is important, four constructs of*OPN* promoter (−2615-luc, −2239-luc, −1505-luc, and −495-luc) were used for analysis. Obvious downregulation in methylated −2615-luc (*Hpa*II and*Hha*I methyltransferase) was observed. Compared to the deletion of this control region (\~390 nt) in front of*OPN* promoter, the decreasing level of promoter activity is not as obvious as −2615 M-luc construct ([Figure 7](#fig7){ref-type="fig"}). It means that methylation in the front of*OPN* promoter caused some silent mechanism that make chromosome structure more compact or block some promotion transcription factors. The EMSA data indicated that 13th CpG site of our analyzed region could bind to AP1 transcription factor and binding activity is affected by methylation in this CpG site ([Figure 7](#fig7){ref-type="fig"}). Taken together, all these findings correlated with DNA methylation in tissue- or cell-specific gene expression.*OPN* promoter region was densely methylated in some low expression (Figures [2](#fig2){ref-type="fig"}, [4](#fig4){ref-type="fig"}, and [5](#fig5){ref-type="fig"}).
Our data revealed that DNA methylation of CpG sites in*OPN* promoter was the main mechanism through specific transcription factor that makes the tissue-specific expression. In previous study, AP1-like binding site (TGAGCGA) was identified as a methylated insulator region in human blastoma cell line \[[@B28]\]. Analyzed region in front of porcine*OPN* promoter showed that CpG 1st binding site contained the specific binding site sequence. While in our interesting DNA regulation region range from −2615 to −2239 nt of*OPN* promoter also exhibited little block access in the competition of probe. It is suggested that CpG 1st and CpG 13th play an important role in methylation controlled mechanism to regulate gene expression. We finally utilize Z-dock program \[[@B29]\] to analyze the interaction between AP1 (PDB code: 1JNM) and RNA polymerase II initial transcription factor (TFIIB) (PDB code: 1VOL) \[[@B30]\] to see if they had stable binding. From the docking result of Z-dock ([Figure 8](#fig8){ref-type="fig"}), we proved the AP1 is significantly bound to TFIIB. We also found the AP1 can autoregulate the HDAC-1 in promoter region and lead to significant higher degrees of hypermethylation in the*OPN* promoter region and cause AP1 to be hypermethylated consequently ceasing the*OPN* mRNA expression \[[@B31]\].
Further mechanical studies by the computational biology also pointed out that the DNA sequence for hypermethylation of*OPN* promoter binding sites is c-Jun. The chain A of c-Jun could be encompassed more tightly by inward rotational structure change of TFIIB during the MD process ([Figure 19](#fig19){ref-type="fig"}). Therefore, we found c-Jun had crucial role for interaction of initiating transcription by RNA polymerase II. The methylation of c-Jun leads to of hyper-condense helix structural change and makes transcription termination which stops*OPN* mRNA production. Therefore, the MD docking results reconfirm the c-Jun partake the crucial roles in consequent*OPN* transcription that matches our wet laboratory studies. We suppose this will cause the problems in the embryonic development and lead to threatened conditions. Therefore, adjusting*OPN* promoter c-Jun (AP1) methylation will affect transcription binding and could be the treatment for genetic developing errors in the future.
In conclusion, aberrant methylation of porcine*OPN* gene was frequently found in different tissues of somatic nuclear transferred cloning pigs and bisulfite sequence data suggested that the*OPN* promoter region of −2615 to −2239 nt may be a crucial regulation DNA element. In pig ear fibroblast cell culture study, the demethylation of*OPN* promoter was found in dose-dependent response of 5-aza-dc treatment and followed the*OPN* mRNA reexpression. In cloned pig study, discrepant expression pattern was identified in several cloned pig tissues, especially in brain (99.75% up-regulation), heart (11.5% down-regulation), and ear (18.03% downregulation). Promoter assay data revealed that four methylated CpG sites presenting in the −2615 to −2239 nt region cause significant downregulation (approximately 75%) of*OPN* promoter activity (*P* \< 0.001). EMSA data also suggested that CpG 13th and 1st sites showed methylation noncompetition phenomenon which had influence on binding with transcription factor.
This research was supported by Grants NSC-95-2313-B-005-030-MY3, NSC-101-2313-B-005-014-MY3, 101-2314-B-039-013-MY3, NSC102-2325-B039-001, and NSC102-2221-E-468-027- from the National Science Council and was partly supported by the Ministry of Education, Taiwan, under the aiming top university plan (ATU-101-s0508). The authors would like to thank their colleagues (Drs. Tung-Chou Tsai, Cheng-Wei Lai, and Zi-Lun Lai) in the Molecular Embryology & DNA Methylation Laboratory for their help with discussions and technical issues. This study is also supported in part by Taiwan Department of Health Clinical Trial and Research Center of Excellence (DOH102-TD-B-111-004), Taiwan Department of Health Cancer Research Center of Excellence (MOHW103-TD-B-111-03), and CMU under the Aim for Top University Plan of the Ministry of Education, Taiwan.
Conflict of Interests
=====================
The authors declare that there is no conflict of interests regarding the publication of this paper.
Authors\' Contribution
======================
Chih-Jie Shen, Yung-An Tsou, and Hsiao-Ling Chen contributed equally to this work.
{#fig1}
{#fig2}
{#fig3}
{#fig4}
{#fig5}
{#fig6}
{#fig7}
{#fig8}
{#fig9}
{#fig10}
{#fig11}
{#fig12}
{#fig13}
{#fig14}
{#fig15}
{#fig16}
{#fig17}
{#fig18}
{#fig19}
{#fig20}
[^1]: Academic Editor: Chung Y. Hsu
| {
"pile_set_name": "PubMed Central"
} |
Rheumatoid arthritis (RA) is a systemic autoimmune disorder involving chronic inflammation. RA is characterized by an excessive inflammatory response and deforming polyarthritis, activation of T cells. Interleukin (IL)-17 produced by T helper (Th) 17 cells increases in the peripheral blood of patients with RA[@b1]. Additionally, Th17 cells exacerbate tissue destruction by inducing chronic inflammation in patients with RA[@b2]. The expression of proinflammatory cytokines is related to the pathogenesis of RA where several proinflammatory cytokines participate in the augmentation of RA[@b3]. In particular, IL-17 leads to a chronic immune inflammatory response in patients with RA[@b2], and its production is upregulated in patients with RA[@b4].
Signal transducer and activator of transcription (STAT) 3 is a DNA binding transcription factor that controls the production of a number of cytokines and T cell lineage. Activation of STAT3 induces the production of IL-17[@b5][@b6] and inflammatory CD4^+^ T cells, such as Th17[@b6][@b7]. Indeed, the differentiation of Th17 is regulated directly by STAT3[@b8]. This in turn allows for STAT3 being a potential target for RA therapy where inhibiting it reduced Th17 differentiation in an experimental autoimmune arthritis model[@b9][@b10].
A considerable proportion of patients with RA fail to respond to drug treatment. For instance, 30--40% of patients with RA receiving anti-tumor necrosis factor (TNF)-α therapy did not respond[@b11]. Several TNF-α blocker including infliximab do not have therapeutic effects in one-third of patients with RA[@b12]. In addition, these patients have an increased possibility of not responding to other biologics[@b13].
High-temperature requirement protein A (HtrA) 2 is a serine protease that is released into the cytosol during apoptosis and is involved in neurodegeneration[@b14]. There is evidence that HtrA2 deficiency or a point mutation in the gene can result in neurodegeneration[@b14][@b15][@b16]. In RA patients, HtrA2 may be related to nonresponsiveness to RA drug treatment. It is well documented that gene expression of HtrA2 in peripheral blood mononuclear cells of nonresponders to methotrexate (MTX) is significantly decreased compared with that in responders; therefore, HtrA2 is an attractive candidate to test in nonresponders to MTX[@b17].
Motor neuron degeneration (mnd)2 mice lack HtrA2 activity because of the missense mutation Ser276Cys in the protease domain of HtrA2[@b18]. It has been documented that loss of HtrA2 function results in dysfunction including cell death[@b19]. Indeed, mnd2 mice show tendrils in the spleen and thymus, and death occurs 40 days after birth[@b20]. It is also widely believed that this specific gene mutation causes disproportionate inflammation and untimely death[@b21].
We hypothesized that HtrA2 is associated with RA pathogenesis mediated by STAT3 and Th17. The current study was performed to identify whether HtrA2 could be a therapeutic target in RA. First, we investigated the role of HtrA2 on STAT3 expression and Th17 differentiation. Second, we analyzed the therapeutic effect of HtrA2 expression in a model of experimental autoimmune arthritis. Finally, we examined whether HtrA2 inhibits the development of plaque and Th17 differentiation in APOE knockout mice immunized with a proteoglycan to induce a hyperlipidemia-based animal model of RA.
Results
=======
Excessive activation of T cells in mnd2 mutation
------------------------------------------------
HtrA2^mnd2^ is single point mutation of HtrA2 causing an untimely death[@b20] and dysregulated inflammation[@b21]. We hypothesized that HtrA2^mnd2^ may be involved in an excessive immune response. HtrA2^mnd2^ caused an abnormal small body size and a smaller spleen with fewer cells ([Fig. 1A](#f1){ref-type="fig"}). On the other hand, CD4^+^ cells in the spleen of HtrA2^mnd2^ increased significantly compared to that of wild type (WT) ([Fig. 1B](#f1){ref-type="fig"}). We also investigated whether HtrA2 could downregulate the activation of T cells. In the mixed lymphocyte reaction (MLR), HtrA2 overexpression in LBRM cells significantly decreased T cell proliferation compared with that induced by mock vector in LBRM cells ([Fig. 1C](#f1){ref-type="fig"}). We also found that the numbers of naïve T cells (both CD4^+^ and CD8^+^) were reduced but those of memory T cells were increased in spleens from HtrA2mnd2 mice ([Fig. 1D](#f1){ref-type="fig"}). CD4^+^ T cells were also slightly higher in spleen, lung, liver, and gut from HtrA2^mnd2^ mice than in the same tissues from WT mice ([Fig. 1E](#f1){ref-type="fig"} and F).
Regulation of STAT3 production by HtrA2 expression
--------------------------------------------------
To identify whether HtrA2 reduces STAT3 expression associated with inflammation[@b22][@b23], we performed an *in vitro* cleavage assay using GST-fusion proteins of STAT3 and HtrA2. We found that STAT3 was degraded by HtrA2 ([Fig. 2A](#f2){ref-type="fig"}). However, the number of pSTAT3 705^+^ and pSTAT3 727^+^ cells significantly increased in the spleen of HtrA2^mnd2^ mice compared to WT ([Fig. 2B](#f2){ref-type="fig"}). CD4^+^IL-17^+^ cells were also promoted in splenocytes of HtrA2^mnd2^ mice compared to that in WT ([Fig. 2C](#f2){ref-type="fig"}). In addition, the level of pSTAT3 was increased by IL-6 stimulation, but IL-12 failed to increase pSTAT1 expression ([Fig. 2D](#f2){ref-type="fig"}).
Excess IL-17 production induced by HtrA2 deficiency
---------------------------------------------------
As STAT3 directly regulates IL-17 expression[@b6], we analyzed IL-17 levels mediated by HtrA2. Gene expression of IL-17 in LBRM decreased significantly by overexpressing HtrA2 ([Fig. 3A](#f3){ref-type="fig"}). IL-17A promoter activity also decreased significantly by HtrA2 overexpression ([Fig. 3B](#f3){ref-type="fig"}). In contrast, siRNA HtrA2 increased IL-17 mRNA significantly ([Fig. 3C](#f3){ref-type="fig"}). We also found that the HtrA2 inhibitor, UCF101, significantly decreased the level of HtrA2 mRNA but enhanced that of IL-17 ([Fig. 3D and E](#f3){ref-type="fig"}). In addition, CD4^+^IL-17^+^ and CD4^+^IFN-γ^+^ cells were increased in spleen, liver, and lung from HtrA2^mnd2^ compared with WT. In contrast, CD4^+^IL-13^+^ cells were reduced in these tissues from HtrA2^mnd2^ mice compared with WT ([Fig. 3F--H](#f3){ref-type="fig"}).
HtrA2 deficiency and mnd2 mutation results in Th17 differentiation and osteoclastogenesis
-----------------------------------------------------------------------------------------
STAT3 directly controls development of Th17, which induces osteoclastogenesis[@b6][@b24]. We analyzed Th17 differentiation and osteoclastogenesis to determine whether HtrA2 can modulate Th17 differentiation and osteoclastogenesis. CD4^+^IL17^+^ cells of HtrA2^mnd2^ significantly increased under the Th17 condition compared to those in WT ([Fig. 4A](#f4){ref-type="fig"}). UCF101 also significantly increased CD4^+^IL17^+^ cells compared to that in the vehicle ([Fig. 4B](#f4){ref-type="fig"}). HtrA2^mnd2^ significantly increased osteoclastogenesis compared to that of WT ([Fig. 4C](#f4){ref-type="fig"}). mRNA levels of tartrate-resistant acid phosphatase (TRAP), matrix metalloproteinase9, cathepsin-K, and integrin-β3 involved in osteoclastogenesis were promoted in HtrA2^mnd2^ compared to those in WT ([Fig. 4D](#f4){ref-type="fig"}).
HtrA2 overexpression attenuates CIA severity reducing Th17 differentiation
--------------------------------------------------------------------------
To determine whether HtrA2 has therapeutic activity in RA, we measured the development of CIA in mice injected with either the HtrA2 overexpressing or mock vector once weekly from day 7 after the first immunization. HtrA2 overexpression significantly reduced CIA severity compared to that of mock ([Fig. 5A](#f5){ref-type="fig"}). A histological analysis indicated that HtrA2 overexpression attenuated inflammation and bone and cartilage damage ([Fig. 5B](#f5){ref-type="fig"}). The overexpression of HtrA2 decreased Th17 differentiation in the spleens of mice with CIA ([Fig. 5C](#f5){ref-type="fig"}). Regulatory T (Treg) cell differentiation was not changed in CIA-induced mice injected with the HtrA2 overexpressing vector ([Fig. 5D](#f5){ref-type="fig"}).
HtrA2 overexpression decreased atherosclerotic plaque burden and inflammation in hyperlipidemia-based mouse RA
--------------------------------------------------------------------------------------------------------------
Metabolic disorders are correlated with non-response to RA treatment[@b25]. As ApoE^−/−^ mice generate atherosclerotic lesions repeatedly at susceptible locations, particular at branch points of the aorta[@b26][@b27], we examined whether overexpressing HtrA2 is therapeutic and reduces aortic plaque progression in ApoE^−/−^ mice immunized with proteoglycan to induce a hyperlipidemia-based RA animal model. Results showed that HtrA2 overexpression inhibited aortic plaque formation compared to that of mock and enbrel ([Fig. 6A](#f6){ref-type="fig"}). HtrA2 overexpression also decreased the infiltration of immune cells and bone destruction ([Fig. 6B](#f6){ref-type="fig"}). There was a reduction in Th17 differentiation, however, Treg differentiation remained unchanged with HtrA2 overexpression ([Fig. 6D](#f6){ref-type="fig"}).
Discussion
==========
HtrA2 has been investigated primarily as a serine protease with neurodegeneration and apoptosis functions [@b14][@b16]. HtrA2 has been suggested to be a candidate associated with non-response to drug treatments for RA[@b17]. However, minimal information is available regarding the therapeutic effect of HtrA2 in RA. Here, we investigated the therapeutic function of HtrA2 and identified its remedial process in RA.
The most remarkable finding of this study was that HtrA2 ameliorated RA by inhibiting STAT3. As STAT3 enhances IL-17 production and Th17 differentiation[@b5][@b6][@b7], inhibiting STAT3 is a promising strategy for RA therapy. In this study, we found that HtrA2^mnd2^ significantly increased pSTAT3 expression, whereas HtrA2 induced STAT3 degradation *in vitro*. As well, HtrA2 revealed curative activity in a CIA and hyperlipidemia-based RA. To our knowledge, this is the first study suggesting the use of HtrA2 in RA therapy by inhibiting STAT3. Furthermore, results found that immoderate T cell activation and proliferation of CD4^+^ T cells, which can induce uncontrolled inflammation such as an inflammatory storm, and may be involved in early death[@b20].
Th17 release of IL-17 has been shown to be positively correlated with the progression of RA. As IL-17 upregulates the inflammatory response[@b28][@b29][@b30], Th17 and IL-17 exacerbates inflammation in RA[@b2][@b31][@b32]. Th17 results in osteoclastogenesis and bone annihilation as RA progresses[@b24][@b33]. The release of IL-17 from Th17 cells is controlled by IL-13, which can inhibit the expression of proinflammatory cytokines[@b34][@b35]. Our data indicates that HtrA2 overexpression decreased the activation of T cells and IL-17 gene expression and promoter activity, however, the HtrA2 inhibitor and HtrA2^mnd2^ increased IL-17 mRNA levels as well as the proliferation of Th17 cells. Supporting this, CD4^+^IL-17^+^ and CD4^+^IFN-γ^+^ cells were also increased but CD4^+^IL-13^+^ cells were reduced in several tissues from HtrA2^mnd2^mice. Additionally, osteoclastogenesis of HtrA2^mnd2^ increased significantly compared to that of WT. As Th17 and IL-17 cause excessive inflammation during RA development, our results suggest a novel therapeutic function of HtrA2 as a treatment for RA.
CD4^+^ and CD8^+^ T cells are involved in maintaining homeostasis of the immune system[@b36]. Because naïve T cells (CD62L^hi^CD44^lo^) differentiate into memory T cells (CD62L^lo^CD44^hi^) after exposure to antigen, memory T cells have previously encountered and responded to antigen[@b37]. Our observation that naïve T cells were decreased whereas memory T cells were increased in HtrA2^mnd2^ mice suggests that HtrA2 may play an important role in the immune response.
Metabolic disorders are correlated with being non-responsive to RA treatment[@b25] and may indicate similar effects to other RA drugs[@b13]. HtrA2 has been pointed out as a significant target for non-response to MTX[@b17]. Moreover, HtrA2 mRNA levels in peripheral blood mononuclear cells and CD4^+^ T cells of healthy individuals and patients with RA from the National Center for Biotechnology Information Gene Expression Omnibus database (GSE15573 and GSE4588) were analyzed. Here, HtrA2 gene expression increased in patients with RA compared to that in healthy controls ([Supplementary Figure 1A](#S1){ref-type="supplementary-material"}). However, HtrA2 mRNA expression in patients with familial hypercholesterolemia, a genetic disorder of uncontrolled high cholesterol level related to primary hyperlipidemia[@b38], decreased compared to that of controls ([Supplementary Figure 1B](#S1){ref-type="supplementary-material"}). Additionally, HtrA2 gene expression levels in patients with RA within the GSE58795 database inclusive to clinical information was analyzed and found that HtrA2 expression of non-responders and moderate responders was significantly downregulated compared to that of responders ([Supplementary Figure 1B](#S1){ref-type="supplementary-material"}). These results suggest that HtrA2 expression may be increased to protect against RA pathogenesis in a way similar to that of IL-10[@b39]. HtrA2 may have a remedial function of reducing plaque formation in the aorta and differentiation of Th17 in hyperlipidemia-based RA. Thus, an HtrA2 deficiency may have no response to RA associated with a metabolic disorder.
Granulocyte-macrophage colony-stimulating factor (GM-CSF) is an immune modulatory cytokine that plays a key role in immune tolerance. There is evidence that GM-CSF ameliorates autoimmunity[@b40][@b41]. GM-CSF also increases differentiation of Treg cells, thereby reducing the severity of autoimmune disease[@b42]. However, GM-CSF can also mediate excessive inflammation and exacerbate autoimmune diseases including RA[@b34]. Particularly relevant to this study is the demonstration that GM-CSF deficiency protected against CIA progression[@b41].
Eye-mediated immune tolerance may also be a good strategy for CIA treatment. It has been suggested that eye-induced immune tolerance to type II collagen has therapeutic benefit in an arthritis model[@b44]. CD8^+^ Tregs also perform a therapeutic function in CII-induced disease by inducing eye-induced tolerance[@b45]. Therefore, the therapeutic activity of HtrA2 in CIA may be related to induction of immune tolerance. However, further study is required to identify whether GM-CSF expression and eye-mediated immune tolerance are affected in CIA mice treated with HtrA2.
Diminished HtrA2 expression may be involved in the non-responsiveness to drugs for treating RA, however, there are few data available regarding the anti-arthritic effects of HtrA2. The new identified HtrA2 functions demonstrate its large role in ameliorating RA progression, which may shed new light on treating non-responders to medication. We have shown that HtrA2 suppressed Th17 differentiation and IL-17 expression by inhibiting STAT3. These preliminary data suggests that HtrA2 could be a crucial candidate for treating patients with RA including non-responders.
Materials and Methods
=====================
Animals
-------
Male DBA1/J (Orient, Seoul, Korea) and apolipoprotein E deficient (ApoE^−/−^) mice (Jackson Laboratories, Bar Harbor, ME, USA) at 6--8-weeks-old were maintained in groups of five in polycarbonate cages in a specific pathogen-free environment. They were fed standard mouse chow (Ralston Purina, Gray Summit, MO, USA) and water *ad libitum*. Mice harboring the mutant motor neuron degeneration (mnd) 2 on a C57BL/6 mice background were obtained from Professor Hyangshuk Rhim (Department of Medical Life Sciences, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea). Heterozygous mnd2 (HT, mnd2/+) males were crossed with HT females 3--6 months of age to generate littermates of all three genotypes: wild-type (WT, +/+), HT, and homozygous mnd2 (mnd2/mnd2). The genotypes of the mice were analyzed by the polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) method described previously[@b46].
Ethics statement
----------------
The Animal Care Committee of The Catholic University of Korea approved the experimental protocol. All experimental procedures were evaluated and carried out in accordance with the protocols approved by the Animal Research Ethics Committee at the Catholic University of Korea. All procedures performed followed the ethical guidelines for animal studies.
Type II collagen immunization to induce arthritis
-------------------------------------------------
Collagen induced arthritis (CIA) was induced in DBA1/J mice (n = 10/group) by injecting 100 μg bovine type II collagen (Chondrex Inc., Redmond, WA, USA) dissolved overnight in 0.1 N acetic acid (4 mg/ml) in complete or incomplete Freund's adjuvant (Chondrex) via the base of the tail. The immunization was performed intradermally.
Clinical scoring of arthritis
-----------------------------
Peripheral joints were examined for arthritis visually twice weekly. Arthritis was graded based on the method of Williams *et al*.[@b47] with the following five grades: 1) Grade 0: no evidence of erythema or swelling; 2) grade 1: erythema and mild swelling confined to the mid-foot (tarsals) or ankle joint; 3) grade 2: erythema and mild swelling extending from the ankle to the mid-foot; 4) grade 3: erythema and moderate swelling extending from the ankle to the metatarsal joints; and 5) grade 4: erythema and severe swelling encompassing the ankle, foot, and digits. The final value represented the average index from all four legs recorded by two independent observers.
Induction of hyperlipidemia-based RA and preparation of atherosclerotic lesions
-------------------------------------------------------------------------------
ApoE^−/−^ mice were maintained in a temperature controlled (23 ± 2 °C) room. ApoE^−/−^ strain mice were injected with proteoglycan. The ApoE^−/−^ mice were fed a western type diet (Teklad Adjusted Calories Western-type diet), containing 21% fat by weight (0.15% by weight cholesterol and 19.5% by weight casein without sodium cholate) for 16 weeks. The aorta was uncovered longitudinally from the aortic root to the iliac bifurcation. The stained aorta tissues were kept in fixation solution until images were captured.
Injection of agents
-------------------
Eight days after immunization, mice with autoimmune arthritis were injected intravenously with 100 μg HtrA2 or mock vector in 2 ml saline over a 10-sec period. The same mice received an intramuscular injection of 50 μg HtrA2 or mock vector in the leg using electrical stimulation (electroporation). Enbrel (Pfizer, Berlin, Germany) was injected into ApoE^−/−^ mice. The injection was conducted subcutaneously.
HtrA2 siRNA transfection
------------------------
Small interfering RNA (siRNA) constructs for HtrA2 siRNA and non-targeting siRNA (Dharmacon, Lafayette, CO, USA) were obtained using siGENOME SMARTpool reagents (Dharmacon). siRNA was transfected using the Amaxa 4D-nucleofector X unit and the DN-100 program according to the manufacturer's recommendations (Lonza, Cologne, Germany).
Intracellular staining for flow cytometry
-----------------------------------------
Before cell staining, differentiated CD4^+^ T cells were stimulated with 25 ng/ml phorbol myristate acetate and 250 ng/ml ionomycin (both from Sigma, St. Louis, MO, USA) in the presence of GolgiStop (BD Pharmingen, San Diego, CA, USA) for 4 hours. The cells were stained with anti-mouse CD4 peridin chlorophyll protein (PerCP), anti-mouse CD25 allophycocyanin (APC), anti-mouse IL-17 fluorescein isothiocyanate (FITC), and anti-mouse FOXP3 phycoerythrin (PE) (eBiosciences, San Diego, CA, USA) followed by fixation and permeabilization with a Foxp3 staining buffer kit according to the manufacturer's instructions to examine intracellular cytokines. All samples operated on a FACSCalibur (BD Pharmingen) and data were analyzed using FlowJo software (Tree Star, Ashland, OR, USA).
Confocal microscopy
-------------------
Tissue cryosections (7 μm thick) were fixed with acetone and stained with FITC-, PE-, PerCP-Cy5.5-, or APC-conjugated monoclonal antibodies against mouse CD4, pSTAT3 (Tyr 705, Ser 727), IL-17, and FOXP3 (eBioscience). After an overnight incubation at 4 °C, the stained sections were visualized by confocal microscopy (LSM 510 Meta; Zeiss, Göttingen, Germany).
Transfection
------------
The HtrA2 vector was obtained from Professor Hyangshuk Rhim (Department of Medical Life Sciences, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea) was used to overexpress HtrA2. The mock and HtrA2 vector constructs were transfected into LBRM cells using an Amaxa 4D-Nucleofector X unit and the DN- program 100 according to the manufacturer's recommendations (Lonza).
T cell responses *in vitro*
---------------------------
LBRM cells transfected with mock or HtrA2 overexpression vector were cultured in 96-well plates containing 200 μl/well of complete medium, at 37 °C in a humidified 5% (v/v) CO~2~/air atmosphere. Cells were pulsed with 1 μCi of tritiated thymidine (3\[H\]-TdR; NEN Life Science Products Inc., Boston, MA, USA) 18 h before harvesting and were counted with an automated harvester (PHD Cell Harvester; Cambridge Technology, Inc., Cambridge, MA, USA). Results are expressed as the mean c.p.m. of triplicate samples ± SD.
Real-time quantitative polymerase chain reaction (PCR)
------------------------------------------------------
Total RNA was extracted using TRIzol reagent (Molecular Research Center. Cincinnati, OH, USA). The concentration of RNA in each sample was measured using a NanoDrop ND-1000 (Thermo Fisher Scientific, MA, USA). Total RNA (2μg) was reverse transcribed into cDNA using the Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Science). Gene expression were estimated using a LightCycler 2.0 instrument (Roche Diagnostic, Mannheim, Germany) and ver. 4.0 software. All reactions were performed with LightCycler FastStart DNA Master SYBR Green I (Takara, Shiga, Japan) following the manufacturer's instructions. Relative mRNA levels were normalized to that of β-actin. The primer sequences used to amplify the mouse genes are listed in [Supplementary Table 1](#S1){ref-type="supplementary-material"}.
Protein purification
--------------------
The pGST-STAT3 and pGST-HtrA2 plasmids were grown overnight in 20 ml LB containing 100 μg/ml ampicillin (LBAmp). The cultures were transferred to a fresh LBAmp medium and grown for an additional 90 min at 37 °C, followed by inducing the GST-fusion proteins with 0.5 mM isopropyl-β-D-1 thiogalactopyranoside for 2 additional hours. Cells were harvested and resuspended in 1 ml EBC buffer \[50 mM Tris-HCl (pH 8.0), 120 mM NaCl, and 0.5% NP40\] containing 100 μg/ml lysozyme. The cell suspension was lysed by sonication for 30 sec on ice and the lysates were collected by centrifugation at 12,000 rpm for 1 min. The GST-fusion proteins were purified from the cell lysates under non-denaturing conditions by selective binding to glutathione-Sepharose beads for 20 min at room temperature. Protein quantity and purity were analyzed by comparing with bovine serum albumin of known concentrations by sodium dodecyl sulfate- polyacrylamide gel electrophoreses (SDS-PAGE), followed by Coomassie Brilliant Blue staining (0.1% Brilliant Blue R, 45% methanol, and 10% acetic acid).
*In vitro* cleavage assay
-------------------------
GST-HtrA2 (0.2 μM) was incubated with GST-STAT3 (1 μM) in 100 μl of cleavage buffer \[50 mM Tris-HCl (pH 7.5), 1 mM DTT\] for 16 hr at 37 °C for the *in vitro* cleavage reaction. The proteins bound to the beads and the supernatant were separated by centrifugation at 12,000 rpm for 5 min and resolved by 15% SDS-PAGE and analyzed by IB assays.
Statistical analysis
--------------------
All data are expressed as mean ± standard deviation. The statistical analysis was performed using SPSS 10.0 for Windows (IBM Corp., Armonk, NY, USA). Numerical data of the groups were compared using one-way analysis of variance (ANOVA) and the nonparametric Mann--Whitney test. Differences in the mean values of various groups were analyzed by ANOVA with a post-hoc test. P-values \<0.05 were statistically significant.
Additional Information
======================
**How to cite this article**: Lee, S. H. *et al*. HtrA2 suppresses autoimmune arthritis and regulates activation of STAT3. *Sci. Rep.* **6**, 39393; doi: 10.1038/srep39393 (2016).
**Publisher\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Material {#S1}
======================
###### Supplementary Information
###### Supplementary Table 1
This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI15C3062). This study was supported by a grant of the Korean Health Technology R&D Project, Ministry for Health & Welfare, Republic of Korea.(HI14C1851). This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI15C1062).
**Author Contributions** S.H.L., Y.M., H.S. and C.M.L. designed the experiments, analyzed the data. S.H.L. wrote the manuscript along with input from Y.M., H.S., S.K., E.K., J.Y., N.M., J.M., S.P. and H.R. S.H.L., Y.M., H.S., S.K., E.K., J.Y. and N.M. conducted all *in vitro* assays with help from J.M., S.P. and H.R. S.H.L., H.S. and S.K. performed animal experiments. E.K. conducted all immunohistochemistry experiments. S.H.L., Y.M., H.S. and C.M.L. discussed and developed the concept. All authors critically reviewed and approved the final form of the manuscript.
{#f1}
{#f2}
{#f3}
{#f4}
{#f5}
{#f6}
[^1]: These authors contributed equally to this work.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#s1}
============
The potential importance of translational research for autism spectrum disorder (ASD) is increasing given that its biological pathology and fully effective treatments have not yet been determined. Specifically, the acquisition of more knowledge about objective and quantitative neurobiological and behavioral indices may facilitate the development of basic and clinical research as well as lead to possible ASD phenotypes being identified.
In this context, locomotor dynamics may constitute a useful objective and quantifiable measure for translational research, as investigation of this behavioral index is less invasive and can be continued for an extended period of time in both laboratory (animal) and clinical research for the study of psychiatric and developmental disorders (Nakamura et al., [@B14]). Focusing on higher-order statistics such as skewness might be especially useful when characterizing behavioral alterations in psychiatric and developmental disorders. For example, we recently reported that in children with ASD, locomotor activity was significantly more negatively skewed---defined as a left-skewed distribution (or a long left tail relative to the right tail) with extreme values lower than their mean---for all-day activity, and also tended to be more negatively skewed for daytime activity, and, that this pattern of locomotor activity was related to acoustic hyper-reactivity (Takahashi et al., [@B15]). Further extending the use of actigraphy to help determine the clinical relevance of locomotor activity in ASD may help reveal important mechanisms in the underlying neurophysiology of this condition, while using higher-order statistics might further contribute to the realization of this purpose.
The objective of this study therefore was to investigate the relationship between locomotor activity and different clinical characteristics to determine the clinical relevance of locomotor activity in ASD and typically developing (TD) children. More specifically, as the investigation of locomotor activity is more common in attention-deficit hyperactivity disorder (ADHD; Cheung et al., [@B3], [@B4]; De Crescenzo et al., [@B6]) compared to ASD, and ASD is known to have several comorbid psychiatric and developmental problems including attention deficit and hyperactivity (Lai et al., [@B10]), in this study we investigated the relationship between locomotor activity and emotional/behavioral problems as well as autistic traits. We hypothesize that locomotor activity indexes measured in daily life, such as negative skewness---which are related to ASD---will also be related to these clinical characteristics of ASD.
Materials and Methods {#s2}
=====================
Participants {#s2-1}
------------
Data were used from 27 Japanese children aged from 7 to 16 years old. Fourteen had ASD (13 boys) while 13 were TD children (10 boys). All participants were included in our previous study (Takahashi et al., [@B15]). Diagnoses were made by child psychiatrists after medical records were reviewed and a clinical interview had been performed based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (American Psychiatric Association, [@B1]). Diagnoses were confirmed by using the Autism Diagnostic Observation Schedule (Lord et al., [@B12]) and Autism Diagnostic Interview-Revised (Lord et al., [@B11]). There were no significant differences between the groups in terms of sex, age (age in months; ASD 125.6 ± 30.9; TD 138.5 ± 38.2; *U* = 76, *p* = 0.467). The Wechsler Intelligence Scale for Children-Third Revision (Wechsler, [@B19]) showed that, the estimated IQ was above 70 for every child included in the study, and did not differ between groups (IQ: ASD 105.7 ± 23.3; TD 104.7 ± 18.3; *U* = 31, *p* = 0.958). No children were being medicated with psychotropic substances and none of them smoked. Besides having autism, no children had any central nervous system abnormalities. For the TD group having a previous or current psychiatric diagnosis or learning disability served as exclusion criteria.
Ethics Approval and Consent to Participate {#s2-2}
------------------------------------------
This study was conducted in accordance with the Helsinki Declaration with institutional review-board approval being granted by the Research Ethics Committee of the National Center of Neurology and Psychiatry (\#A2013-112) and the research ethics committee of the Graduate School of Education, the University of Tokyo (\#13-119). After the details of the study were explained to them, all participants and their parents provided written informed consent before being included in the study.
Assessment of Autistic Traits and Emotional/Behavioral Problems {#s2-3}
---------------------------------------------------------------
Parents used the Japanese version (Kamio et al., [@B9]) of the Social Responsiveness Scale (SRS; Constantino and Gruber, [@B5]) to assess quantitative autistic traits. This scale consists of five treatment subscales: social awareness, social cognition, social communication, social motivation, and autistic mannerisms. Higher SRS scores indicate increased social impairment.
To assess emotional and behavioral problems, we used the Japanese version (Moriwaki and Kamio, [@B13]) of the Strengths and Difficulties Questionnaire (SDQ; Goodman, [@B7]). This measure also has five subscales: four difficulty subscales (Emotional Symptoms, Conduct Problems, Hyperactivity Inattention, and Peer Problems) and one strengths subscale (Prosocial Behavior). The scores of the four difficulty subscales can be summed to create a total difficulties score. As with the overall SDQ difficulties score, higher scores on the individual difficulty subscales reflect greater difficulties. In contrast, higher scores on the prosocial behavior subscale indicate increased prosociality.
Assessment of Locomotor Dynamics {#s2-4}
--------------------------------
In order to assess locomotor dynamics participants were informed that they should wear the MicroMini Motionlogger actigraph (Ambulatory Monitors Inc., Ardsley, NY, USA; Teicher, [@B16]) on the wrist of their non-dominant hand. Specifically, that they should wear it on their wrist for \> 7 days during the spring, summer, or winter school vacations (TD: 7.7 ± 1.9 days, ASD: 7.8 ± 1.8 days, *U* = 80, *p* = 0.574). Full details of the assessment of locomotor dynamics are provided in our previous study (Takahashi et al., [@B15]). It was expected that children would wear this device continuously throughout the study period. However, they were instructed to remove it when either bathing, or during rigorous exercise, or when performing any activity that might result in the device being damaged. They were also informed that they should just lead their lives normally while wearing the actigraph device. After the recording period had finished, three boys with ASD and two boys with TD had to be excluded from the actigraph behavioral data analysis for not wearing the actigraph for a long enough period of time during the daytime hours.
As described previously (Takahashi et al., [@B15]), sleep--wake cycles were scored using Action W-2 software, and we now further examine the mean and skewness, the first- and third-order statistical moments, of the locomotor distributions, for all-day and daytime activity. Sleep parameters were not examined, since none of them differed significantly between ASD and TD children.
Statistical Analysis {#s2-5}
--------------------
To examine categorical differences in the participants' scores we used Chi-square tests (and Fisher's exact tests when necessary). Due to the non-normal distributions of most of the locomotor activity and clinical characteristic variables nonparametric analyses were performed. The Mann-Whitney U test was used to compare mean parameter values. Spearman's rank order correlation coefficients were used to examine the relationships between variables. The level of statistical significance was set at *p* \< 0.05. As multiple tests were performed a Bonferroni adjustment was subsequently applied to determine statistical significance. SPSS Ver. 22 (IBM Japan, Tokyo, Japan) was used to perform all statistical analyses.
Results {#s3}
=======
Differences in Clinical Characteristics Between Children With Autism Spectrum Disorders and Controls {#s3-1}
----------------------------------------------------------------------------------------------------
All SRS scores were significantly higher in the ASD group than in the controls (Table [1](#T1){ref-type="table"}). Compared with the TD group, the ASD group had a significantly lower SDQ Prosocial Behavior score and significantly higher scores for Total Difficulties and the SDQ Hyperactivity Inattention and Peer Problems subscales (Table [1](#T1){ref-type="table"}).
######
Clinical characteristics of the participants.
Typical development Autism spectrum disorders
------------------------------------------ --------------------- --------------------------- ------ ------ ------ ---------
Social Responsiveness Scale
Total Score 20.5 10.0 70.1 34.4 8 \<0.001
Social Awareness 3.2 2.3 9.9 3.5 2 \<0.001
Social Cognition 4.0 3.2 14.6 5.4 7.5 \<0.001
Social Communication 6.6 3.9 22.2 13.7 16.5 \<0.001
Social Motivation 4.3 3.6 8.7 6.3 49 0.043
Autistic Mannerisms 2.5 1.5 14.7 9.1 12 \<0.001
Strengths and Difficulties Questionnaire
Total Difficulties Score 6.3 3.7 15.8 6.6 17.5 \<0.001
Emotional Symptoms 1.2 1.7 2.9 2.3 52.5 0.610
Conduct Problems 1.6 1.4 2.9 2.0 57 0.105
Hyperactivity Inattention 2.3 0.9 5.9 2.3 19 \<0.001
Peer Problems 1.2 1.2 4.1 2.8 32.5 0.003
Prosocial Behavior 7.7 1.4 4.9 2.9 33 0.004
*SD, standard deviation; Mann-Whitney U test. Number of participants (typical development: autism spectrum disorders) = 13:14*.
Relationship of Locomotor Dynamics to Clinical Characteristics {#s3-2}
--------------------------------------------------------------
We found significant relationships between the locomotor dynamics and clinical characteristics which are presented in Figure [1](#F1){ref-type="fig"}. Specifically, for all children combined there was a significant negative relationship between all-day skewness and the SRS social awareness score (Figure [1A](#F1){ref-type="fig"}). In addition, there was also a significant negative relationship between daytime skewness and the SDQ Hyperactivity Inattention score (Figure [1B](#F1){ref-type="fig"}). No other significant relationships were observed for the clinical characteristics of locomotor dynamics, while the significant relationships described above became non-significant when we divided the children into groups (ASD and TD).
{#F1}
Discussion {#s4}
==========
The results of this study showed that more negatively skewed all-day locomotor activity evaluated with an actigraph was related to the SRS social awareness score when the ASD and TD children were combined in one group. In addition, more negatively skewed locomotor activity during the daytime was related to the SDQ Hyperactivity Inattention score. To the best of our knowledge, this is the first time an association has been reported between locomotor dynamics and clinical characteristics, including autistic traits and emotional/behavioral problems, in children with ASD and TD. Our results suggest that besides using standard descriptive statistics when examining behavioral characteristics in ASD children, it may also be beneficial to utilize third-order statistical moments such as skewness.
For all children combined, the autistic trait of social awareness was related to significantly more negative skewness in all-day locomotor activity (defined as a left-skewed distribution with extreme values lower than their mean, which suggests behavior marked by an increase in large sporadic "troughs" below mean activity levels), with this behavioral characteristic having been previously found in ASD compared to TD children (Takahashi et al., [@B15]). Although the biological background of social awareness is uncertain, it might be associated with von Economo's neurons, which are also known to be involved in motor awareness (Cauda et al., [@B2]). Thus, the use of higher order statistics, such as skewness to examine locomotor activity in animals and humans might help clarify the biological background of autistic traits, including social awareness.
The behavioral problem of hyperactivity/inattention was related to significantly more negative skewness of daytime locomotor activity, which tended to be more negative in children with ASD compared to control children (Takahashi et al., [@B15]). These results support the idea that ADHD may be highly comorbid in ASD (Lai et al., [@B10]), and suggest daytime locomotor activity and its skewness might serve as a potential behavioral phenotype that is connected with the comorbid clinical features of hyperactivity/inattention in ASD children. Several comorbid problems are frequently reported in ASD (Lai et al., [@B10]), including hyperactivity, inattention as well as motor abnormalities, such as motor delay, deficits in coordination and movement planning. Atypical movement in ASD may be regulated by emotion (Trevarthen and Delafield-Butt, [@B17]; Vernazza-Martin et al., [@B18]). Future studies investigating these comorbid features in relation to locomotor activity and emotional regulation in daily life might help elucidate the neurophysiological mechanisms that may underlie these features in ASD.
The small number of children in the ASD and TD groups is a significant study limitation. While it was still possible to detect significant associations between some aspects of locomotor dynamics and clinical characteristics for all children combined, significant relationships were not observed when the children were divided into groups. Further, gender differences exist in many aspects of ASD (Lai et al., [@B10]), however, participants in this study were mainly boys, while the age span was rather large. In addition, although we did not find significant differences in sleep parameters in our previous study (Takahashi et al., [@B15]), sleep problems are frequently observed in ASD children (Lai et al., [@B10]) and might possibly be seasonal (Hayashi, [@B8]), which could have impacted on our analysis of both nighttime and daytime locomotor activity. Given this, research that uses a larger number of children of both sexes with a narrower age range that also controls for season is now warranted to more clearly determine the relationship between locomotor dynamics and different clinical characteristics.
Conclusion {#s5}
==========
Negatively skewed all-day locomotor activity (as seen in activity that was characterized by large sporadic all-day "troughs,") might be a potentially useful quantitative behavioral index related to ASD, especially autistic social awareness. For all children, more negatively skewed daytime locomotor activity was also associated with comorbid hyperactivity/inattention behavioral problems. The results of this study thus build on and extend previous research on locomotor dynamics and further develop understanding of the potential neurophysiological mechanisms that may underlie clinical characteristics in ASD.
Data Availability {#s6}
=================
The datasets generated for this study are available on request to the corresponding author.
Author Contributions {#s7}
====================
HT, ToN, JK, HK, KY, TA, YY, and YK conceived and designed the experiments. HT, ToN, YY, and YK supervised the project. KO, HT and YK confirmed diagnoses. KO, HT, ToN, JK, and TaN performed the experiments. KO, HT, ToN, JK, YY, and YK analyzed the data. KO, HT, ToN, JK, KE, TS, AS, YY, and YK wrote the manuscript. All authors read and approved the final manuscript.
Conflict of Interest Statement {#s8}
==============================
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.
**Funding.** This study was supported by Grants-in-Aid from the Japanese Ministry of Education, Culture, Sports, Science and Technology (23890257 to HT and 24591739 to HT), Intramural Research Grant (23-1 to YK and 26-1 to YK) for Neurological and Psychiatric Disorders of NCNP, Research Grants from the Ministry of Health, Labour and Welfare of Japan (H19-KOKORO-006 to YK and H20-KOKORO-004 to YK), and the Center of Innovation Program from Japan Science and Technology Agency, JST (to YK). This study was partially supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology for the work of YY (\#26242069, \#15K12678). The work of ToN was supported by PRESTO from the Japan Science and Technology Agency.
We wish to thank all the children who participated in this study and their parents.
[^1]: Edited by: Ryouhei Ishii, Osaka University, Japan
[^2]: Reviewed by: Mitsuru Kikuchi, Kanazawa University, Japan; Kuriko Shimono, Osaka University, Japan; Nagy A. Youssef, Augusta University, United States
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-sensors-18-00116}
===============
Surface plasmon polaritons (SPPs) are electromagnetic fields propagating along the interface of metal-insulator, and have been widely discussed for several decades due to the ability to modulate light in nanoscale, as well as break the diffraction limit \[[@B1-sensors-18-00116]\]. Recently, many kinds of plasmonic devices have been investigated, such as filters \[[@B2-sensors-18-00116]\], absorbers \[[@B3-sensors-18-00116]\], splitters \[[@B4-sensors-18-00116]\], and sensors \[[@B5-sensors-18-00116]\]. Among these, plasmonic sensors have drawn more attention, because when compared with traditional optical sensors, such as fiber sensors and silicon-based sensors, plasmonic sensors have much smaller size with comparable sensing performance, which means that they are more suitable for integrating \[[@B6-sensors-18-00116]\]. Various of plasmonic sensors have come out lately such as refractive index sensors \[[@B7-sensors-18-00116]\], temperature sensors \[[@B8-sensors-18-00116]\], phase sensors \[[@B9-sensors-18-00116]\], and gas sensors \[[@B10-sensors-18-00116]\].
The metal-insulator-metal (MIM) waveguide is one of the basic plasmonic waveguides with the capability to confine light within considerable propagating length \[[@B11-sensors-18-00116]\], and many works of sensor are based on this structure \[[@B12-sensors-18-00116],[@B13-sensors-18-00116],[@B14-sensors-18-00116],[@B15-sensors-18-00116],[@B16-sensors-18-00116],[@B17-sensors-18-00116],[@B18-sensors-18-00116],[@B19-sensors-18-00116],[@B20-sensors-18-00116]\]. As a sensor, it requires both high sensitivity (S) and high figure of merit (FOM) to promise an excellent performance with high optical resolution. Great efforts have been made to improve sensitivity on MIM plasmonic sensors, but longer wavelength always suffers from wider FWHM \[[@B12-sensors-18-00116],[@B13-sensors-18-00116],[@B14-sensors-18-00116],[@B15-sensors-18-00116]\], which means lower FOM. Recently, Fano resonance has been universally applied to enhance FOM that results from structure symmetry break or dark-bright resonance interference, and it has an asymmetric spectral line shape with narrower FWHM \[[@B16-sensors-18-00116],[@B17-sensors-18-00116],[@B18-sensors-18-00116],[@B19-sensors-18-00116],[@B20-sensors-18-00116]\]. But such resonance is unstable and can be easily broken due to phase or mode mismatch, which could be caused by localized corrosion of structure, change of localized refractive index, and so on.
In this study, a traditional MIM plasmonic sensor, coupled with simple concentric double rings resonator (CDRR), is proposed. Utilizing the special supermodes, the FOM of that can be significantly improved to as high as 272.3 in the visible-light region without Fano resonance, as well as 203.8 in the near-infrared region. Such value of FOM is quite larger than other recent works \[[@B5-sensors-18-00116],[@B7-sensors-18-00116],[@B12-sensors-18-00116],[@B13-sensors-18-00116],[@B14-sensors-18-00116],[@B15-sensors-18-00116],[@B16-sensors-18-00116],[@B17-sensors-18-00116],[@B18-sensors-18-00116]\]. The peculiar properties of CDRR supermodes and the impact of structure parameters on sensing performance are also discussed. The two-dimensional (2D) Finite-Difference Time-Domain (FDTD) solution is used to simulate this structure with perfectly matched layer (PML) boundary condition.
2. Structural Model and Theory Analysis {#sec2-sensors-18-00116}
=======================================
As shown in [Figure 1](#sensors-18-00116-f001){ref-type="fig"}, a MIM waveguide coupled with a CDRR is proposed. The silicon substrate is plated with silver, which is chosen for the relative low loss \[[@B21-sensors-18-00116]\]. The waveguides are etched on the silver surface. Such a structure can be fabricated by the method of focused ion beam (FIB) \[[@B22-sensors-18-00116]\]. As shown in 2D picture, the grey and white areas represent silver and air, respectively. The permittivity of air is set as $\varepsilon_{i} = 1$, as for silver (Ag) the Drude model is utilized as follows \[[@B2-sensors-18-00116]\]:$$\varepsilon_{m} = \varepsilon_{\infty} - \frac{\omega_{p}^{2}}{\omega(\omega + i\gamma)}$$
Here $\varepsilon_{\infty}$ gives the medium constant for the infinite frequency, $\omega_{p}$ refers to bulk frequency for plasma, $\gamma$ means damping frequency for electron oscillation, and $\omega$ shows incident light angular frequency. The parameters for silver are $\varepsilon_{\infty}$ = 1, $\omega_{p}$ = 1.37 × $10^{16}$ Hz, and $\gamma$ = 3.21 × $10^{13}$ Hz.
In this structure, $R\ = \ 350\ {nm}$ and $r\ = \ 270\ {nm}$ is the radius of the outer and inner ring, respectively. The width of the waveguide is set as $w\ = \ 50\ {nm}$, and the gap between inner and outer ring is $d\ = \ 30\ {nm}$. The gap between CDRR and bus waveguide is $g\ = \ 10\ {nm}$. To collect the incident and transmitted power, two monitors are put, respectively, at $P_{in}$ and $P_{out}$. The transmittance of power can be calculated by $T\ = \ P_{out}/P_{in}.$
TM mode SPPs can propagate at the surface of interface in the waveguide when coupled into the MIM structure. When compared with incident wavelength, the width of the bus waveguide is much smaller, so fundamental TM mode can exist only. The dispersion relation of this fundamental mode is described as follows \[[@B11-sensors-18-00116]\]:$$\frac{\varepsilon_{i}p}{\varepsilon_{m}k} = \frac{1 - e^{kw}}{1 + e^{kw}}$$ $$k = k_{0}\sqrt{{(\frac{\beta_{spp}}{k_{0}})}^{2} - \varepsilon_{i}},p = k_{0}\sqrt{{(\frac{\beta_{spp}}{k_{0}})}^{2} - \varepsilon_{m}}$$ $$\beta_{spp} = n_{eff}k_{0} = n_{eff}\frac{2\pi}{\lambda}$$
Here, $w$ refer to the width of bus waveguide, $\lambda$ shows incident light wavelength in vacuum, $\varepsilon_{i}$ and $\varepsilon_{m}$ give the relative dielectric and metal permittivity, $\beta_{spp}$ and $n_{eff}$ are propagation constant and effective refractive index of SPPs, and $k_{0} = 2\pi/\lambda$ means wave number.
Most of the proposed and designed MIM plasmonic devices have used 2D simulation, which is infinite in one dimension, to test the performance of the devices with much shorter simulation time and lower loss \[[@B23-sensors-18-00116]\], and so is this work. But, the height of the waveguide has a significant effect on the loss of the system, which has to be considered for practical processing. Here, the relationship between the effective refractive index of the SPPs and the height of the waveguide at 850 nm is shown in [Figure 2](#sensors-18-00116-f002){ref-type="fig"}. As the height increases, both the real and imaginary part of the $n_{eff}$ will decrease at start, which is due to the decrease of the fraction corresponding to the modal power in the metal at interfaces. The reduction of the imaginary part of $n_{eff}$ means less loss, which can lead to longer propagation length \[[@B24-sensors-18-00116]\]. However, when greater than a certain height, $n_{eff}$ will remain stable. If continuing improving the height, $n_{eff}$ will approach the value in 2D simulation, which is the $\left. h\rightarrow\infty \right.$ case. The distribution $\left| P_{x} \right|$ of the fundamental mode at 850 nm when *h* = 50 nm is presented in the inset of [Figure 2](#sensors-18-00116-f002){ref-type="fig"}. At other wavelengths, the outcome will be similar. After comparing the values at different wavelengths, the most suitable height can be selected. Our result is agreed with \[[@B24-sensors-18-00116]\].
The transmission spectra of MIM waveguide coupled with single inner (outer) ring or CDRR are shown in [Figure 3](#sensors-18-00116-f003){ref-type="fig"}. In order to illuminate the principle of CDRR better, the gap between bus waveguide and single inner (outer) ring is set as 30 nm (10 nm). [Figure 3](#sensors-18-00116-f003){ref-type="fig"} indicates that the transmission spectrum of CDRR is almost the superposition of the other two spectra with slight shifts, and the FWHM of the mode near 1060 nm becomes narrower. For a single ring resonator, the resonance condition can be described as \[[@B5-sensors-18-00116]\]: $$N\lambda = {Re}\left( n_{eff})L_{eff},N = 1,2,3... \right.$$ where $n_{eff}$ refers to the effective refractive index of the ring resonator, which can be solved by Equations (2)--(4), $L_{eff}$ means effective perimeter, generally refers to the average of the inner and outer perimeters. $N$ refers to mode number which is an integer.
For the CDRR, the resonance condition is the superposition of that for single ring, and the shift of resonant wavelength comes from the change of effective index of each ring resulting from the interaction between each other.
It is found that the modes of CDRR exist in the form of supermodes, which means that the inner and outer ring have resonance simultaneously. The novel supermodes of CDRR are introduced, as shown in [Figure 4](#sensors-18-00116-f004){ref-type="fig"}. However, we notice that there are only two kinds of supermodes: when the energy is mainly trapped in outer ring, the mode in the inner ring will have anti phase; when mainly trapped in inner ring, it is in phase. Refer to the supermodes between two parallel waveguides \[[@B25-sensors-18-00116]\]; the two supermodes could be called odd mode (with anti-phase) and even mode (with in-phase). The mechanism of such supermodes can be explained by the coupled-mode theory (CMT) \[[@B26-sensors-18-00116]\]. The propagation constant in the ring resonator can be derived with Equation (5):$${Re}(\beta) = \frac{2\pi{Re}(n_{eff})}{\lambda} = \frac{2\pi}{\lambda} \cdot \frac{N\lambda}{2\pi R_{eff}} = \frac{N}{R_{eff}}$$ where $R_{eff}$ is the effective radius of the ring resonator, and the propagation constant of odd (even) mode in the system of two parallel single-mode waveguides is as follows \[[@B26-sensors-18-00116]\]:$$\beta_{odd} = \overline{\beta} - \Delta\beta$$ $$\beta_{even} = \overline{\beta} + \Delta\beta$$ $$\overline{\beta} = \frac{\beta_{1} + \beta_{2}}{2},\Delta\beta = \frac{1}{2}\sqrt{4K^{2} + {(\beta_{1} - \beta_{2})}^{2}}$$
Here, $\overline{\beta}$ represents the average propagation constant, $\beta_{1}$ ($\beta_{2}$) is the propagation constant for a single waveguide without coupling, and *K* is the coupling coefficient. In the condition of the codirectional coupling, *K* is a pure imaginary number \[[@B26-sensors-18-00116]\]. Equations (7) and (8) shows that the ${Re}\left( \beta_{even} \right)$ is always larger than ${Re}\left( \beta_{odd} \right)$. Since the CDRR could be considered as two parallel bend waveguides with periodicity, this mode expression can also be applied. Now, the CDRR system is considered as a kind of single-ring resonator, the effective radius of which is determined by the energy distribution. When the energy is mainly trapped in the outer side, the effective radius is larger, leading to smaller propagation constant according to Equation (6), bringing about the odd mode with anti-phase in the CDRR. Otherwise, it will be an even mode with in-phase.
By changing the radius of inner ring (which also means the change of the gap *d*), the transmission spectra correspond to different gaps is obtained in [Figure 5](#sensors-18-00116-f005){ref-type="fig"}. Obviously, the resonant wavelength will have a blue shift when increasing the gap. The [Figure 6](#sensors-18-00116-f006){ref-type="fig"}b shows the details, and we find that for even modes, the resonant wavelength has a linear relationship with gap; for odd modes, it is nonlinear. It can be explained by the effective index. When changing the radius of inner ring linearly, the effective index of outer ring will be influenced and have a nonlinear change, leading a nonlinear relationship between odd-mode (which can be regarded as mode of inner ring approximately according to the energy distribution) resonant wavelength and gap. However, the inner ring will suffer both radius and gap change, and that will keep its effective index unchanged, bringing about the linear relationship between even-mode resonant wavelength and the radius of inner ring as Equation (5). Such conclusion can be useful for designing the CDRR.
Apparently, the even mode has much narrower FWHM; therefore, the two even modes (even $TM_{3}$ and $TM_{2}$) are investigated next, which is assumed as mode 1 and mode 2, respectively. [Figure 6](#sensors-18-00116-f006){ref-type="fig"}a shows the relationship between gap $d$ and FWHM (transmission). When increasing the gap, the FWHM of even mode will be narrower, meanwhile suffering from the increase of transmission. It is due to the special mode distribution of even mode, the energy of which is mainly trapped in the inner ring, and the outer ring suppresses the energy loss, which is beneficial for energy storage. As the inner and outer rings have the same symmetry, the supermode is stable and has the single-ring mode features. Such even modes are potential to generate ultra-high Q factors, which is defined as $Q = \lambda/FWHM$. With gap 30 nm, the Q for mode 1 can reach 277 and for mode 2 is 205. The Q will be larger with increasing gap, however there is a trade-off between Q and transmission.
3. Refractive Index Sensing {#sec3-sensors-18-00116}
===========================
When considering the balance between FWHM and transmission, the gap $d$ is chosen to be 30 nm. The plasmonic sensor filled with media is shown in [Figure 7](#sensors-18-00116-f007){ref-type="fig"}, and [Figure 8](#sensors-18-00116-f008){ref-type="fig"} presents the transmission spectra with different filling dielectric in the CDRR. The refractive index of dielectric is increased from 1 to 1.1 with the step 0.025, leading to a red shift of the spectra. The sensing capabilities is defined as follows \[[@B27-sensors-18-00116]\]:$$S = \frac{\bigtriangleup \lambda}{\bigtriangleup n},{FOM} = \frac{S}{FWHM}$$
The sensitivity $S$ is described as the resonant wavelength shift when the dielectric has a unit change. Since the optical resolution is also crucial for sensors, a high figure of merit (FOM) is needed. By fitting the line in the inset of [Figure 8](#sensors-18-00116-f008){ref-type="fig"}, the sensitivity 708 nm/RIU of mode1 and 1060 nm/RIU of mode 2 are obtained. The FWHM of each is 2.6 nm and 5.2 nm, leading to the high FOM, which is 272.3 and 203.8, respectively. It must be mentioned that such value of FOM is quite larger than other recent works, as shown in [Table 1](#sensors-18-00116-t001){ref-type="table"}. Although some works achieved a higher figure of up to hundreds of thousands \[[@B19-sensors-18-00116],[@B20-sensors-18-00116]\], the FOM definition of these works is different, which always refers to ${FOM}^{*}$ \[[@B27-sensors-18-00116]\].
Due to the higher sensitivity with excellent FOM in the near-infrared region, mode 2 is further studied then. The radius of CDRR is increased with the gap d fixed at 30 nm, and the spectra are given in [Figure 9](#sensors-18-00116-f009){ref-type="fig"}a. When the outer radius raises from 350 nm to 400 nm with the interval 10 nm, the sensitivity is 1112 nm/RIU, 1152 nm/RIU, 1196 nm/RIU, 1224 nm/RIU, and 1280 nm/RIU, respectively, from [Figure 9](#sensors-18-00116-f009){ref-type="fig"}b, and FOM is 185.3, 164, 149.5, 153, and 150.4, respectively. Obviously, increasing the size of CDRR can improve the sensitivity performance of the sensor with the cost of lower FOM, which may, respectively, arise from the longer optical path and higher dissipation of energy.
Comparing with Fano resonance, which is the hotspot for sensing recently \[[@B27-sensors-18-00116]\], this structure has superiority in stability. Although Fano resonance has a narrow FWHM leading to high FOM, few works pay attention to the instability of this state. Fano resonance occurs when a discrete state interferes with a continuum band of states, the shape of the special transmission spectrum can be described by Fano formula \[[@B28-sensors-18-00116]\]:$$Y = \frac{{(q + X)}^{2}}{1 + X^{2}}$$ where $q = cot\delta$ is the Fano parameter, $\delta$ is the phase shift of the continuum between two modes, depending on geometric and material parameters of the system \[[@B28-sensors-18-00116]\], which can be caused by corrosion, fabrication errors, change of localized refractive index, and so on. [Figure 10](#sensors-18-00116-f010){ref-type="fig"} presents the spectrum shape corresponding to different *q* values. Due to the character of cot function, little phase shift will cause significant change of *q*, leading to the breaking of the high-*q* Fano shape, and then bring about the decline of sensing performance.
The supermode in CDRR is based on traditional micro-ring resonance, which is not that fragile. To test the stability of this sensor, a $5\ {nm} \times 5\ {nm}$ square defect is inserted at two different positions of CDRR, respectively, as given in [Figure 11](#sensors-18-00116-f011){ref-type="fig"}b,c. From the corresponding transmission spectrum of mode 2 shown in [Figure 11](#sensors-18-00116-f011){ref-type="fig"}a, the defect on outer (inner) ring only has influence on the resonance of outer (inner) ring, which can make resonant wavelength shift or create new mode. However, the mode 2 still exists with narrow FWHM in both cases, only suffers from little lower absorption or resonant wavelength change. The corresponding distributions $\left| H_{z} \right|$ of mode 2 with defect are presented in [Figure 11](#sensors-18-00116-f011){ref-type="fig"}d,e.
4. Bio-Sensing {#sec4-sensors-18-00116}
==============
Here, the application for bio-sensing is discussed, the same structure parameters are chosen with outer ring radius 350 nm and gap $30~{nm}$, and still only mode 2 is focused. Since water is one of the most common solvents used in chemical and biological applications, it is necessary to test the performance of the sensor in water (*n* = 1.33), as presented in [Figure 12](#sensors-18-00116-f012){ref-type="fig"}a. When considering the concentration of solution can be reflected by the refractive index (RI), a set of different RI values around the water are used to measure the sensitivity and FOM, the transmission spectrum is given in [Figure 12](#sensors-18-00116-f012){ref-type="fig"}b. According to [Figure 12](#sensors-18-00116-f012){ref-type="fig"}c, the sensitivity in the water is 1061 nm/RIU, and the FOM of 193 can be achieved, which can prove that this sensor still has excellent performance in water.
Therefore, such structure can be a sensitive label-free and compact biosensor. A special case for the detection of the DNA hybridization is shown as follows. After the DNA hybridization, i.e., when single strand DNA (ssDNA) becomes double strands DNA (dsDNA), the RI of DNA layers will change from 1.456 to 1.53 \[[@B29-sensors-18-00116]\]. If a layer of ssDNA is implanted inside the resonator, the DNA hybridization can be detected, as shown in [Figure 13](#sensors-18-00116-f013){ref-type="fig"}. After the hybridization, the resonant wavelength of mode 2 shifts from 1553 nm to 1631 nm with high optical resolution, which can be probed easily.
Another bio-application of this sensor is to detect the biomolecules attached to the inner wall of the resonator. We assume that a combined sensor layer of thickness $t_{cap}$, which consists of an activating intermediate layer and a capture layer, has been immobilized onto the wall of inner ring, which can capture a layer of biomolecules of thickness $t_{bio}$ through a selective biochemical process, such as antigen-antibody binding process \[[@B30-sensors-18-00116],[@B31-sensors-18-00116],[@B32-sensors-18-00116]\], as shown in [Figure 14](#sensors-18-00116-f014){ref-type="fig"}a. The RI and material dispersion of the biolayers will depend on how the biomolecules are oriented. Here, we neglect the dispersion of the biolayers and assume that they have the same RI of 1.45, which is close to that of silica \[[@B31-sensors-18-00116]\], which is a realistic assumption that is proved by experimental measurements with MOF biosensors \[[@B33-sensors-18-00116]\]. Assuming $t_{cap} = 10~{nm}$, the relationship between resonant wavelength and thickness of captured layer $t_{bio}$ is shown in [Figure 14](#sensors-18-00116-f014){ref-type="fig"}b, and a thickness sensitivity ($S^{\prime} = \Delta\lambda/\Delta t)$ of 4.9 nm/nm can be gained. When comparing with traditional micro-structured optical fiber (MOF) biosensors \[[@B33-sensors-18-00116]\], this plasmonic biosensor has comparable sensitivity, higher optical resolution and much smaller footprint. It has to be mentioned that from [Figure 12](#sensors-18-00116-f012){ref-type="fig"}b and [Figure 13](#sensors-18-00116-f013){ref-type="fig"}b, the FWHM of mode 2 can remain narrow within a wide environment effective RI range. Since layers effect can be equal to the change of effect RI essentially \[[@B34-sensors-18-00116]\], the detection for the layer can still possess the character of high optical resolution.
Finally, when considering a biomolecules layer attached to the wall of inner ring with a certain thickness *t*, the sensitivity for the layer RI is measured for the case *t* = 5 nm and *t* = 10 nm, shown as [Figure 15](#sensors-18-00116-f015){ref-type="fig"}. We choose both sizes because these are the most common feature scale of single biomolecule \[[@B32-sensors-18-00116]\], and the sensitivity 200 nm/RIU (*t* = 5 nm) and 370 nm/RIU (*t* = 10 nm) can be realized with high optical resolution.
5. Case of Triple Ring {#sec5-sensors-18-00116}
======================
Out of interest, the case of concentric triple rings resonator (CTRR) is proposed with *w* = 50 nm, *d* = 15 nm, *g* = 10 nm, and *R* = 350 nm. The transmission spectrum and mode distribution are presented in [Figure 16](#sensors-18-00116-f016){ref-type="fig"}. Here, assuming that the binary *0*/*1* represent a set of opposite phases, and that the array *\[1 b c\]* describes the supermode (we set outer-ring-phase as 1), where *b*, *c* give the phase of mode in middle ring and inner ring. After checking the supermodes for a wide wavelength range, we found that only three kinds supermodes exist: *\[1 1 0\]*, *\[1 1 1\]*, and *\[1 0 1\]*.
For supermode *\[1 1 1\]*, an ultra-narrow FWHM can be reached, as shown in [Figure 16](#sensors-18-00116-f016){ref-type="fig"}, when $\lambda = 591\ {nm}$ and 864 nm. When $\mathsf{\lambda} = 591\ {nm}$, the FWHM is nearly 1 nm, and an ultra-high Q = 591 can be gained. The reason is that in this supermode, most energy is trapped in the inner ring. Therefore, the outer double rings can suppress the energy loss. However, this supermode may suffer from the low energy absorption, such as that at 864 nm. Additionally, CTRR will possess a great many resonant wavelengths, it may bring about some disturbances on sensing. This structure may have potential applications to be found.
6. Conclusions {#sec6-sensors-18-00116}
==============
In summary, a sensor based on MIM waveguide coupled with a CDRR is investigated in the near-infrared region, and a high sensitivity with ultra-high FOM is obtained in RI-sensing and bio-sensing. The special features of supermodes in CDRR is discussed, presenting that the even mode can produce ultra-narrow FWHM. The special case of triple rings is also discussed. In addition, such a structure also provides the concept to design a high Q microcavity. Utilizing the centrosymmetric characters of outer and inner microcavity, an ultra-high Q factor can be achieved without breaking the original mode characteristics. This device can also be used as multi band-stop filter in plasmonic system, which may have applications in integrated optical circuits.
This work is supported by the National Natural Science Foundation of China (60907003, 61671455); the Foundation of NUDT (JC13-02-13), the Hunan Provincial Natural Science Foundation of China (13JJ3001), and the Program for New Century Excellent Talents in University (NCET-12-0142).
Zhaojian Zhang designed the simulations and wrote the paper; Junbo Yang disscussed and revised the paper; Dingbo Chen, Jingjing Zhang, Jie Huang, Xin He and Yunxin Han give some advises.
The authors declare no conflict of interest.
{#sensors-18-00116-f001}
{#sensors-18-00116-f002}
{#sensors-18-00116-f003}
{#sensors-18-00116-f004}
{#sensors-18-00116-f005}
{#sensors-18-00116-f006}
{#sensors-18-00116-f007}
{#sensors-18-00116-f008}
{#sensors-18-00116-f009}
{#sensors-18-00116-f010}
{#sensors-18-00116-f011}
{#sensors-18-00116-f012}
{#sensors-18-00116-f013}
{#sensors-18-00116-f014}
{#sensors-18-00116-f015}
{#sensors-18-00116-f016}
sensors-18-00116-t001_Table 1
######
Comparison of sensitivity (S) and figure of merit (FOM) reported in various Plasmonic sensor.
Reference S (nm/RIU) FOM
----------------------------------------------------------- ------------ -------
Chen, L., et al. \[[@B5-sensors-18-00116]\] 1496 124.6
Yan, Shu Bin, et al. \[[@B13-sensors-18-00116]\] 868.4 43.9
Tang, Y., et al. \[[@B16-sensors-18-00116]\] 1125 74
Zhang, Zhidong, et al. \[[@B17-sensors-18-00116]\] 596 7.5
Zafar, Rukhsar, and Salim, M. \[[@B18-sensors-18-00116]\] 1060 176.7
This paper 1060 203.8
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-cells-09-01451}
===============
Cancer, a significant and serious medical issue, is driving mortality throughout the globe \[[@B1-cells-09-01451]\]. It is distinguished by uncontrolled and unscheduled cell multiplication. Tumors associated with viral or bacterial infection and genetic mutations are known to influence the cancer growth rates \[[@B2-cells-09-01451]\]. Although cancer is thought to be preventable, there are several factors that will increase the risk for developing it. Various risk factors, such as smoking and chewing tobacco, alcohol, obesity, chronic inflammation, age, ethnicity, and geographical location, are the major determinants for developing cancer \[[@B3-cells-09-01451]\]. Emerging research suggests that the majority of cancers are brought about by malfunction of numerous protein coded genes, such as antiapoptotic proteins, growth factors, receptors of growth factor, tumor suppressors, and transcription factors. These factors constitute the target for cancer prevention and treatment.
When tumors are formed, their abnormal cells obtain at least 10 common characteristics that describe the malignant tumor cell's complexity, including proliferation, resistance to growth suppressors, avoidance of programmed cell death (apoptosis), unlimited ability to replicate, growth of fresh blood vessels, tissue invasion with risk of metastatic growth, genetic instability and mutation of high frequency, tumor-driven inflammation, altered metabolism, and compromised immunological surveillance \[[@B4-cells-09-01451],[@B5-cells-09-01451]\]. During tumor advancement and malignant conversion, abnormal cells evade the mechanisms of host defense. Targeting and inhibiting the previously mentioned characteristics has been perceived as potential approaches to cancer treatment. Apart from conventional treatment, an additional area of interest among researchers is the utilization of alternative therapy, particularly using dietary sources to regulate these key mechanisms.
The idea of chemoprevention is receiving extensive awareness because it is a cost-effective option for cancer control when compared to conventional chemotherapy \[[@B6-cells-09-01451],[@B7-cells-09-01451]\]. Cancer chemoprevention by utilizing natural and dietary compounds, particularly phytoconstituents, minerals, and vitamins, in different in vivo and in vitro conditions, has produced empowering results against several cancers \[[@B8-cells-09-01451],[@B9-cells-09-01451],[@B10-cells-09-01451],[@B11-cells-09-01451],[@B12-cells-09-01451]\]. Herbal medicines are emerging as new and innovative substances with astounding pharmaceutical potential that can be utilized to abate human ailments, including cancer \[[@B13-cells-09-01451],[@B14-cells-09-01451]\]. No conclusive evidence has been found that herbal medicines can cure or treat cancer by themselves. Nevertheless, certain plant extracts, fractions, and pure compounds have been found to possess anticancer properties, and are used to develop chemotherapy medicines. For example, vincristine and vinblastine, the active constituents of *Catharanthus roseus*, and taxanes present in Pacific yew tree, are approved chemotherapy drugs.
There is a correlation between dietary habits and cancer development of certain cancers. Food contains different bioactive compounds that improve health and counteract cancer and other chronic diseases. These bioactive elements are found mostly in plants \[[@B15-cells-09-01451],[@B16-cells-09-01451],[@B17-cells-09-01451],[@B18-cells-09-01451],[@B19-cells-09-01451],[@B20-cells-09-01451]\]. Various findings have led to the identification and development of phytocompounds for cancer prevention and therapy. Phytochemicals affect proliferation, differentiation, apoptosis evasion, and angiogenesis of cancer cells. Such phytoconstituents are beneficial by reducing the harmful impacts of chemotherapy and increasing the viability of traditional chemotherapy \[[@B21-cells-09-01451]\]. Dietary phytochemicals prevent carcinogenesis, a multistage mechanism that involves tumor formation, growth, malignant development, and progression. The mechanism of action of many phytochemicals that suppress the development of tumors in animal models involve modulation of molecular signal transduction pathways that react to various external signals, and are implicated in cancer development.
A significant signaling pathway implicated in initiating changes in gene expression is the Janus kinase (JAK)-signal transducer and activator of transcription (STAT) pathway. Aberrated activation of this signaling pathway promotes tumorigenesis. Abnormal activation occurs when a ligand binds to its receptor constitutively or when tyrosine kinase is inappropriately stimulated \[[@B22-cells-09-01451]\]. The JAK/STAT signaling may act either directly or indirectly by triggering nuclear factor-κB (NF-κB) activation \[[@B23-cells-09-01451]\]. Blocking JAK/STAT signaling in cancer cells can suppress the expression of target genes that control essential cell functions and hamper cancer cells from evading growth control mechanism, such as apoptosis and invasion. Therefore, antagonizing JAK/STAT signaling could obstruct the transformation of preneoplastic lesions into a malignant tumor.
Studies have demonstrated that numerous phytochemicals can interfere with the JAK/STAT signaling mechanism in human malignant cells. There are only a few previous reviews which provide an in-depth analysis of this interesting research area. Most of the prior publications focus either on a single phytochemical \[[@B24-cells-09-01451],[@B25-cells-09-01451],[@B26-cells-09-01451]\] or natural compounds that modulate various transcription factors, including STATs \[[@B27-cells-09-01451],[@B28-cells-09-01451],[@B29-cells-09-01451],[@B30-cells-09-01451],[@B31-cells-09-01451]\]. In this context, this review article aims to explore various naturally-occurring phytoconstituents that can suppress the JAK/STAT signaling pathway, thereby inhibiting growth of cancer cells caused by aberrated JAK/STAT signaling. We have presented and critically analyzed up-to-date research conducted on this topic to identify possible chemopreventive and therapeutic agents with established molecular targets.
2. Literature Search Methodology {#sec2-cells-09-01451}
================================
We have followed the preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines to perform literature searches \[[@B32-cells-09-01451]\]. In vitro, in vivo, and human studies that explored the antiproliferative potency of phytoconstituents by inhibiting JAK/STAT pathway were screened using authentic databases, such as PubMed, ScienceDirect, Web of Science, SpringerLink, Scopus, and Google Scholar. Relevant full articles published until January 2020 in peer-reviewed journals have been included. Conference abstracts, books, book chapters, and unpublished findings have not been included. Only papers published in the English language have been considered and included in this review. The major keywords used for literature search included cancer, tumor, proliferation, JAK, STAT, phytochemicals, polyphenols, terpenoids, alkaloids, saponins, steroids, in vivo, in vitro, and clinical studies.
3. JAK/STAT Signaling Pathway {#sec3-cells-09-01451}
=============================
The JAK/STAT signaling pathway serves as an important agent involved in functional responses. Three main proteins are involved in JAK/STAT signaling, including cell-surface, JAK, and STAT receptors. Upon activation of the receptor complex, STAT is phosphorylated on its retained tyrosine residues, which prompts the release of STAT from the receptor and the dimerization of two STAT molecules. The dimer then translocates into the nucleus wherein the dimer gets attached to DNA and stimulates the expression of genes that are responsive to cytokines \[[@B33-cells-09-01451],[@B34-cells-09-01451]\]. Cytokines, including interferon (IFN) and interleukin (IL), attach to their unique receptors involved in the JAK/STAT pathway. This binding induces JAK to phosphorylate one another and afterward phosphorylate the receptor by itself at the receptor STAT protein binding sites. After binding, STAT is phosphorylated by JAK and separated from the receptor where they then bind with another STAT molecule to form a dimer. The STAT complex then travels into the nucleus, gets attached to DNA, and initiates gene expression that promotes tumor cell proliferation and survival ([Figure 1](#cells-09-01451-f001){ref-type="fig"}). Such cytokines bind specific cytokine receptors, which are related to the JAK lineage \[[@B35-cells-09-01451],[@B36-cells-09-01451]\]. Four types of mammalian JAKs exist, and they include tyrosine kinases 2 (TYK2), JAK1, JAK2, and JAK3.
Transcriptional activity and its functional roles are needed for dimerization of STATs through phosphorylation of tyrosine residues. This promotes and mediates binding of regulatory receptors with the formation of phosphorylated complex \[[@B37-cells-09-01451]\]. STAT dimerization regulates the process of phosphorylation by associating with the SH2 domain. In the 1990s, it was found that STAT proteins are IFN-directed genes \[[@B33-cells-09-01451],[@B34-cells-09-01451],[@B38-cells-09-01451]\]. Among primates, STATs consist of seven structurally distinct members: STAT1, STAT2, STAT3, STAT4, STAT5A, STAT5B, and STAT6 \[[@B39-cells-09-01451]\]. STAT1, STAT3, and STAT5 were elevated in various human cancer cells \[[@B40-cells-09-01451]\]. Such proteins involve cytoplasmic transcription factors; for example, transduction of signals from hormones, cytokines, and growth factors. Furthermore, STATs have downstream effector roles \[[@B39-cells-09-01451],[@B41-cells-09-01451]\]. STAT proteins regulate numerous cellular biological functions, including organogenesis, fetal growth, programed cell death, differentiation, growth, inflammation, and the immune system \[[@B42-cells-09-01451],[@B43-cells-09-01451],[@B44-cells-09-01451],[@B45-cells-09-01451],[@B46-cells-09-01451]\]. Inside the cytoplasm, STAT proteins remain as monomers; however, they produce dimers after phosphorylation of their tyrosine domains \[[@B37-cells-09-01451]\]. The dimer travels to the nucleus to empower transcriptional functions. In the formation of tumors, STATs are constitutively stimulated by tyrosine kinases, including Cluster-Abelson breakpoint (Bcr-Abl), JAK, Src, and epidermal growth factor receptor (EGFR).
Unregulated JAK/STAT signaling contributes to proliferation, survival, inflammation, invasion, new blood vessel formation, and metastasis, which are implicated in cancer initiation, progression, and advancement \[[@B27-cells-09-01451]\].
4. JAK/STAT Inhibitors {#sec4-cells-09-01451}
======================
4.1. JAK Inhibitors {#sec4dot1-cells-09-01451}
-------------------
In 1996, it was revealed that AG490, a pan-JAK inhibitor, had the ability to impede the occurrence of recurrent B-cell leukemia based on in vitro and in vivo studies \[[@B47-cells-09-01451]\]. Since then, natural components, including resveratrol, curcumin, piceatannol, and flavopiridol, have been preclinically examined and found to inhibit innumerable inflammatory pathways \[[@B48-cells-09-01451]\]. These pathways inhibit STAT3 phosphorylation, primarily through decreased cytokine production or as a direct JAKs inhibitor \[[@B49-cells-09-01451],[@B50-cells-09-01451]\]. Further, after the JAK2 mutation was discovered in myeloproliferative disorders, orally bioavailable and more effective JAK inhibitors have been developed. According to preclinical studies in myeloproliferative models, these compounds are incredibly efficient in their ability to abrogate illness, which has stimulated clinical testing \[[@B51-cells-09-01451],[@B52-cells-09-01451],[@B53-cells-09-01451]\]. The well-known JAK1/2 inhibitor INCB018424 (Incyte Corp., Wilmington, DE, USA) has demonstrated huge clinical benefits (reduction of splenomegaly, night sweats, weakness, and irritation) correlating with a decline in proinflammatory cytokines release. INCB018424 is in the third phase of clinical trials \[[@B52-cells-09-01451]\]. Merely marginal reductions were observed in the burden of the mutated JAK2 allele. This may be due to insufficient suppression of the JAK2 kinase mutant, or because the disease is the responsibility of another signaler. The additional JAK inhibitors (XL019, CEP-701, and TG101348) have shown similar results; however, they possess different adverse events, which include anemia, thrombocytopenia, neutropenia, transaminitis, neurotoxicity, and gastrointestinal intolerance \[[@B52-cells-09-01451]\]. The potency, sensitivity (JAK1/2 versus JAK2), and half-life of both JAK inhibitors are the factors underlying their side effect variations. Critically, JAK1/2 is required for natural hematopoiesis and thrombocytopenia, and anemia may result from these medications unless specific dosing schedules or lower doses are provided \[[@B54-cells-09-01451]\]. AZD1480, a JAK 1/2 inhibitor, was evaluated in the treatment of interleukin-6 (IL-6)-driven cancers, including prostate, breast, and ovarian cancer. The results of this study showed that AZD1480 suppressed tumor growth \[[@B55-cells-09-01451]\]. Such drugs are presently being studied for solid tumors, including lymphomas, sarcomas, and carcinomas, in phase 1 clinical trials.
4.2. STAT Inhibitors {#sec4dot2-cells-09-01451}
--------------------
An ideal STAT3-binding site (like double-stranded DNA) was made and injected intratumorally or intravenously to cultured cells to aim STAT3 as a DNA-binding protein. It prevents the STAT3 dimer from binding to its targets by isolating it into the STAT3 decoy \[[@B56-cells-09-01451],[@B57-cells-09-01451],[@B58-cells-09-01451],[@B59-cells-09-01451]\]. Elevation of tyrosine phosphorylated STAT3 was analyzed preclinically by the STAT3 decoy in neck and head squamous cell carcinomas. The significant levels of tyrosine-phosphorylated STAT3 resulted in cancer cell apoptosis, resulting in lower tumor growth. Interaction between the decoy and other therapies has also been demonstrated. Currently, this STAT3 decoy is clinically studied in patients with head and neck cancer, and is less intrusive and readily accessible for local injection.
Direct STAT3 inhibitors have been explored, and the research has concentrated on designing agents influencing the SH2 domain by restricting STAT3 phosphorylation and/or dimerization. The primary STAT3 inhibitors, such as peptidomimetics and engineered small molecules, act by inhibiting phosphorylation of intracellular STAT3, DNA-binding, and transcriptional activities \[[@B60-cells-09-01451]\]. In comparison, peptidomimetics may be associated with nonphosphorylated monomeric proteins of STAT3 via pY-SH2 domain binding (the peptide or mimetic incorporates the pY motif in the SH2 domain of the STAT3 monomer) to create a hetero complex. It decreases the rates of free nonphosphorylated STAT3 monomers usable for phosphorylation and de novo activation. Some of these agents have decreased the progression of cancer in various preclinical models \[[@B60-cells-09-01451],[@B61-cells-09-01451],[@B62-cells-09-01451]\]. Regardless of the fact that those components demonstrated a fair sensitivity to repress the activity of STAT3, they were not clinically tested because of the high concentrations required for them to be effective \[[@B63-cells-09-01451]\]. The utilization of a chemical database of previously synthesized compounds in a luciferase reporter cell-based analysis prompted the discovery of transcription-dependent STAT3 and STAT5 inhibitors \[[@B64-cells-09-01451],[@B65-cells-09-01451],[@B66-cells-09-01451]\]. For example, nifuroxazide, a drug used to treat diarrhea, could effectively inhibit JAK2 and TYK2, while decreasing the p-STAT3 levels in multiple myeloma \[[@B67-cells-09-01451]\]. The antimalarial drug, pyrimethamine, has been described as an STAT3 inhibitor and a myeloma growth inhibitor. It is currently undergoing clinical trials to treat lymphocytic leukemia \[[@B65-cells-09-01451]\].
The appearance of mutant (dominant-negative forms) STATs has allowed for delineation of the specific function of domains or residues, and it has led to the discovery of the novel or noncanonical functions for the STATs as mediators of tumorigenesis \[[@B68-cells-09-01451],[@B69-cells-09-01451]\]. It has been suggested that both noncanonical and canonical functions of individual STAT play a central role in the initiation and promotion of tumors. The canonical pathway is characterized as tyrosine phosphorylated STATs functioning as transcription factors. The noncanonical signaling embodies various functions of nontyrosine phosphorylated STATs, including mediation of DNA methylation, focal adhesions activation, and regulation of mitochondrial functions \[[@B70-cells-09-01451],[@B71-cells-09-01451],[@B72-cells-09-01451],[@B73-cells-09-01451]\]. The noncanonical mechanism represents the activation of nontyrosine phosphorylated STAT3, inducing transcription in combination with NF-κB or CD44, which is situated in the mitochondria to control the synthesis of ATP, and then interacts with the stathmin (a microtubule-linking protein) to inflect cell motility \[[@B74-cells-09-01451],[@B75-cells-09-01451]\]. STAT3--tubulin associations may be affected by microtubule-targeting agents, e.g., paclitaxel \[[@B76-cells-09-01451]\]. Moreover, STAT3 phosphorylated on its serine residue, but not on its tyrosine residue, can control transcription in chronic lymphoblastic leukemia \[[@B77-cells-09-01451]\]. STAT acetylation, ubiquitylation, and sumoylation functions are being examined in controlling tumor growth and metastasis. Subsequently, both nontyrosine and tyrosine phosphorylated STAT3 perform significant roles in cancer cells, and this knowledge can be utilized to develop potential anticancer agents.
5. Phytochemicals Inhibiting the JAK/STAT Signaling Pathway {#sec5-cells-09-01451}
===========================================================
Several bioactive phytocompounds have demonstrated that the JAK/STAT pathway is inhibited by various mechanisms. The JAK/STAT pathway comprises more than one site of action that phytochemicals can target. Phytochemicals can block this signaling pathway by reducing the levels of cytokines or growth hormones which trigger JAK/STAT protein activation. Phytochemicals can also act by stopping JAK phosphorylation before STAT activation. Controlling JAK/STAT signaling can be achieved by inhibiting STAT dimerization and by preventing the translocation of STAT dimer from the cytoplasm into the nucleus. The ultimate goal in the signaling pathway is to hinder the STAT--DNA binding, which directly inhibits JAK/STAT-regulated gene transcription. Phytochemicals may also control the JAK/STAT pathway by association with JAK/STAT pathway inhibitors, such as SHP \[[@B78-cells-09-01451],[@B79-cells-09-01451],[@B80-cells-09-01451],[@B81-cells-09-01451]\]. Few phytochemicals can specifically block one target site, while others can control the JAK/STAT system at multiple sites. In the following sections, we analyze the experimental evidence of the capacity of various phytochemicals from diverse plant origin to regulate JAK/STAT pathway in various preclinical cancer models.
5.1. Phenolics and Polyphenols {#sec5dot1-cells-09-01451}
------------------------------
These phytocompounds comprise 15-C atoms as well as one or several hydroxyl groups attached to one or multiple aromatic rings and they are water soluble. More than 8000 phenolic components exist in nature, with many found in various vegetables and fruits. These are metabolites of plants that are considered to confer numerous health benefits via different biochemical and pharmacological actions, such as anti-inflammatory, immunomodulatory, antioxidant, and antimicrobial actions, as well as regulation of numerous cell signaling pathways.
### 5.1.1. Resveratrol {#sec5dot1dot1-cells-09-01451}
Resveratrol (also known as, *trans*-3,5,4′-trihydroxystilbene, [Figure 2](#cells-09-01451-f002){ref-type="fig"}), a compound present in peanuts, berries, and grapes, has been demonstrated as a potent chemopreventive and chemotherapeutic agent \[[@B82-cells-09-01451],[@B83-cells-09-01451],[@B84-cells-09-01451],[@B85-cells-09-01451]\]. Resveratrol's antiproliferative and cytotoxic functions have been associated with JAK/STAT pathway inhibition. Resveratrol obstructed JAK phosphorylation and, as a result, blocked STAT1 phosphorylation in human epidermoid carcinoma (A431) cells \[[@B86-cells-09-01451]\] ([Table 1](#cells-09-01451-t001){ref-type="table"}). The growth of human multiple myeloma cells was inhibited by resveratrol to overcome the chemoresistance by suppressing both the inducible and constitutive activation of STAT3, resulting in downregulation of antiapoptotic gene expressions \[[@B87-cells-09-01451]\]. Besides the inhibition of the JAK/STAT signaling pathway, resveratrol was also found to suppress the function of Src tyrosine kinase, which consequently blocked STAT3's action in specific cancer cells \[[@B88-cells-09-01451]\]. Resveratrol was effective in blocking both signaling channels for JAK/STAT and Src/STAT in in vitro experiments. A resveratrol analog, 3,4,5,4′-tetramethoxystilbene, repressed STAT3 phosphorylation and exhibited greater antitumor activity than resveratrol \[[@B89-cells-09-01451]\]. An experimental study showed that resveratrol was bound to the estrogen receptor, but it did not cause the proliferation of estrogenic cells. Interaction with resveratrol has been found to result in a viable IL-6 transporter, a potent STAT3 mediator \[[@B90-cells-09-01451]\]. It is important to note that resveratrol studies have been concentrated on the short-term results with constrained clinical research.
### 5.1.2. Curcumin {#sec5dot1dot2-cells-09-01451}
Curcumin is one of the most studied polyphenols for anticancer potential \[[@B91-cells-09-01451],[@B92-cells-09-01451],[@B93-cells-09-01451],[@B94-cells-09-01451]\]. Curcumin is the key curcuminoid present in *Curcuma longa* L., (family Zingiberaceae). Curcumin is known to regulate the JAK/STAT signaling pathway. Unlike resveratrol, curcumin blocked STAT3 phosphorylation, thereby blocking STAT3 dimer translocation from cytoplasm into the nucleus of human multiple myeloma cells \[[@B95-cells-09-01451]\]. Curcumin has been found to suppress STAT1, STAT3, JAK1, and JAK2 phosphorylation in microglia stimulated with gangliosides, lipopolysaccharides (LPS), and/or microglia cells. It was also found that activation of Src homology region 2 domain-containing phosphatase-2 (SHP-2), a negative JAK behavior regulator, was the probable mechanism influencing the JAK/STAT signaling pathway for curcumin-mediated inhibition \[[@B78-cells-09-01451]\]. Additional research showed curcumin's potential to reduce STAT3 phosphorylation in small-cell lung cancer cells by desensitizing the downstream target genes, which was accountable for cancer cell proliferation \[[@B96-cells-09-01451]\]. The constitutive activation of the JAK/STAT pathway was also disrupted by curcumin in T-cell leukemia, in which it decreased JAK and STAT phosphorylation and resulted in growth arrest and subsequent apoptosis \[[@B97-cells-09-01451],[@B98-cells-09-01451]\].
### 5.1.3. Ascochlorin {#sec5dot1dot3-cells-09-01451}
Ascochlorin, an isoprenoid antibiotic obtained from the fungus *Ascochyta viciae* Lib., has exhibited significant antiproliferative response in many tumor cell lines and in vivo experimental models \[[@B99-cells-09-01451]\]. Ascochlorin has been found to suppress the migration and invasion in two human glioblastoma cell lines (A172 and U373MG cells) by decreasing matrix metalloproteinase-2 gelatinolytic action and expression. To understand the underlying mechanism by which the ascochlorin smothered cell invasion and migration, the effect of ascochlorin on JAK/STAT signaling has been explored. Ascochlorin essentially diminished JAK2/STAT3 phosphorylation and blocked translocation of STAT3 to the nucleus \[[@B100-cells-09-01451]\].
### 5.1.4. Bergamottin {#sec5dot1dot4-cells-09-01451}
Bergamottin, a furanocoumarin, is present in grape juice, lime, lemon, and bergamot oils. It is a potential antioxidant, anti-inflammatory, and anticancer agent \[[@B101-cells-09-01451]\]. In human multiple myeloma (U266) cells, bergamottin has been found to suppress a constitutive activation of STAT3. This effect was achieved through resisting phosphorylation of JAK 1/2 and c-Src. Furthermore, bergamottin induced the expression of the tyrosine phosphatase SHP-1 and the silencing of the SHP-1 gene by siRNA abrogated the capability of bergamottin to hinder STAT3 actuation, which is a crucial factor in bergamottin's action on SHP-1. It has also been associated with downregulation of the expression of STAT3-directed genes, such as VEGF, COX-2, survivin, cyclin D1, Bcl-xL, Bcl-2, and inhibitor of apoptosis protein-1 (IAP-1). This is related to substantial increase in apoptosis associated with cell cycle arrest at sub-G1 phase and poly (ADP-ribose) polymerase (PARP) cleavage induced by caspase-3 \[[@B102-cells-09-01451]\].
### 5.1.5. Capillarisin {#sec5dot1dot5-cells-09-01451}
One of the principal bioactive components of *Artemisia capillaries* Thunb., capillarisin is a naturally-occurring chromone. It is an analogue of benzopyran, and a coumarin isomer. It explicitly restrained both constitutive and inducible STAT3 activation at tyrosine 705, but not at serine 727, in human multiple myeloma cells \[[@B103-cells-09-01451]\]. Other than hindering phosphorylation of STAT3, capillarisin hindered constitutive STAT3 movement and nuclear translocation. The concealment of STAT3 was interceded through the restraint of the activation of upstream JAK1, JAK2, and c-Src kinases. Interestingly, knockdown of the SHP-1 and SHP-2 genes by siRNA depressed the capability of capillarisin to hinder the activation of JAK1 and STAT3. This suggests that SHP-1 and SHP-2 play a vital role in its potential mechanism of action. Apart from this, capillarisin also downregulated the STAT3-controlled antiapoptotic (Bcl-xL, Bcl-2, survivin, and IAP-1) and proliferative (cyclin D1) gene expression levels, resulting in decrease in cell viability levels, cell cycle arrest at sub-G1 phase, and induction of cell apoptosis. Overall, capillarisin represents a potent inhibitor of STAT3 that can have a negative regulatory effect on the chemoresistance and metastasis of neoplastic cells \[[@B103-cells-09-01451]\].
### 5.1.6. Bavachin {#sec5dot1dot6-cells-09-01451}
Bavachin, a flavonoid, is derived from the seeds of *Psoralea corylifolia* Linn., (currently known as *Cullen corylifolium* (L.) Medik.). It is a phytoestrogen, which activates both the estrogen receptors, namely estrogen receptor-α (ERα) and ERβ. Bavachin diminished the proliferation of multiple myeloma cells; however, it is not cytotoxic to normal cells. It restrained the activation of STAT3 in multiple myeloma cell lines. Moreover, bavachin expanded the gene expression of phorbol-12-myristate-13-acetate-induced protein 1 (PMAIP1, alternative known as Noxa) and p53. It also diminished the expression of Bcl-2, X-linked inhibitor of apoptosis protein (XIAP), survivin, and B-cell lymphoma-extra-large (Bcl-xL). Furthermore, bavachin was found to induce apoptosis through caspase-3 and caspase-9 activation, confirming the participation of mitochondrial pathway. Overall, bavachin targets STAT3 and could be used for the treatment of multiple myeloma \[[@B104-cells-09-01451]\].
### 5.1.7. Epigallocatechin Gallate (EGCG) {#sec5dot1dot7-cells-09-01451}
EGCG is the catechin commonly present in tea (*Camellia sinensis* (L.) Kuntze), and it has demonstrated promising anticancer properties against many types of tumors \[[@B105-cells-09-01451],[@B106-cells-09-01451],[@B107-cells-09-01451],[@B108-cells-09-01451]\]. A study conducted by Tang et al. (2012) \[[@B109-cells-09-01451]\] found that both the JAK3 and STAT3 protein phosphorylation and translation were diminished by EGCG in two human pancreatic cancer cell lines, namely AsPC-1 and PANC-1. The expressions of STAT3-directed genes in pancreatic cancer cells were also inhibited by EGCG. By subduing the bioavailability of insulin-like growth factors (IGFs), it also decreased the level of IGF-mediated phosphorylated-STAT3 proteins in hepatocellular carcinoma cells \[[@B110-cells-09-01451]\]. Several investigators have found that EGCG treatment suppressed the action of STAT3 in breast, head, and neck carcinoma cells \[[@B111-cells-09-01451],[@B112-cells-09-01451]\]. Very recently, EGCG has been found to inhibit the proliferation of chronic myeloid leukemia cells by inducing apoptosis via inhibition of the Bcr/Abl oncoprotein and controlling its downstream pathways, such as p38-MAPK/JNK and JAK2/STAT3/AKT \[[@B113-cells-09-01451]\].
### 5.1.8. Emodin {#sec5dot1dot8-cells-09-01451}
Emodin (1,3,8-trihydroxy-6-methyl anthraquinone) is an anthraquinone obtained from Japanese knotweed, buckthorn, and rhubarb. It is also explicitly extracted from *Rheum Palmatum* L., as well as *Socotrine aloe*, *Barbados aloe*, and *Zanzibar aloe*. Additionally, it is found in numerous fungi species, including members of genera Pestalotiopsis, Pyrenochaeta, and Aspergillus. Emodin is known to possess antioxidant, anti-inflammatory, immunomodulatory, and antitumor properties \[[@B114-cells-09-01451],[@B115-cells-09-01451],[@B116-cells-09-01451]\]. A study was conducted by He et al., \[[@B117-cells-09-01451]\] with 26S proteasome inhibitors by utilizing cell-based screening test in which emodin was reported as an effective human 26S proteasome inhibitor. Emodin hindered chymotrypsin-like and caspase-like actions of the human 26S proteasome and enhanced the ubiquitination of endogenous cell proteins. Molecular docking studies demonstrated that the orientation/conformation in the active pocket of the 1, 2, and 5 subunits of 26S proteasome was suitable for nucleophilic attack. Emodin stimulated the antiproliferative role of interferon α/β (IFN-α/β) by increasing the STAT1 phosphorylation, diminishing the STAT3 phosphorylation, and expanding the expression of an endogenous gene activated by IFN-α. It also repressed IFN-α-stimulated ubiquitination and type I interferon receptor 1 (IFNAR1) degradation. Moreover, it promoted the antiproliferative impact of IFN-α against HeLa cells (a cervical carcinoma cell line) and diminished the development of tumors in Huh7 hepatocellular carcinoma-bearing mice. Based on the results, it is possible that emodin enhanced the antiproliferative action of IFN-α/β by the inhibition of the JAK/STAT signaling pathway through restraining 26S proteasome-stimulated IFNAR1 degradation \[[@B117-cells-09-01451]\].
### 5.1.9. Chalcones {#sec5dot1dot9-cells-09-01451}
Chalcones are flavonoids found in various fruits (oranges, strawberries, and tomatoes), vegetables (bean sprouts, shallots, and potatoes), and spices (licorice). Chalcones have been shown to be promising chemopreventive and antitumor agents due to their antioxidant, cytotoxic, apoptosis-inducing, and numerous cell signaling-modulatory properties \[[@B118-cells-09-01451],[@B119-cells-09-01451]\]. Chalcone, a α,β-unsaturated flavonoid, restrained the phosphorylation of STAT3 in LPS- and IL-6-regulated endothelial cells \[[@B120-cells-09-01451]\].
### 5.1.10. Formononetin {#sec5dot1dot10-cells-09-01451}
Formononetin, a naturally-occurring isoflavone, may be found in low quantities in numerous dietary products, such as beans, carrot, cauliflower, iceberg lettuce, green peas, and red potatoes. It is associated with the Fabaceae family and is isolated from the roots of *Astragalus mongholicus* Bunge. This root represents an essential ingredient utilized in the traditional Chinese medicine due to its suppressive effect against various malignant tumors. Formononetin has been under intense investigation during the last decade due to its ability to promote apoptosis and suppress proliferation in multiple in vitro and in vivo experimental cancer models, such as breast, colorectal, and prostate carcinoma \[[@B121-cells-09-01451],[@B122-cells-09-01451]\]. One investigation demonstrated that formononetin effectively repressed the proliferation and invasion of HCT116 and SW1116 colon carcinoma cells. It also caused cell cycle arrest at the G0--G1 stage through the downregulation of protein expressions of cyclins D1. The anticancer impact of formononetin was found to be mediated by the impairment of phosphoinositide 3-kinase (PI3K)/protein kinase B (Akt) and STAT3 signaling pathways \[[@B123-cells-09-01451]\].
### 5.1.11. Garcinol {#sec5dot1dot11-cells-09-01451}
Garcinol, a poly-isoprenylated benzophenone analog, is isolated from the dried fruit rind of *Garcinia indica* Choisy. The antioxidant, anti-inflammatory, and anticancer properties of garcinol underscore its therapeutic benefits \[[@B124-cells-09-01451],[@B125-cells-09-01451]\]. It may stop constitutive and IL-6-inducible STAT3 activation in hepatocellular carcinoma cells. The molecular docking analysis demonstrated garcinol's capability to link with STAT3 at the SH2 domain and to depress its dimerization. It additionally repressed the acetylation of STAT3 and subsequently disabled its DNA binding capacity as a result of an acetyltransferase inhibitor. The restraint of STAT3 actuation by garcinol prompted repression of the expression levels of different genes associated with proliferation, survival, and angiogenesis. Moreover, it repressed the proliferation of hepatocellular carcinoma cells by stimulating apoptosis. Astoundingly, garcinol hindered the development of human hepatocellular carcinoma xenograft tumors in athymic nu/nu mice by impairing STAT3 activation. The proapoptotic and antiproliferative impacts of garcinol in hepatocellular carcinoma were mediated through inhibition of the STAT3 signaling pathway \[[@B126-cells-09-01451]\].
### 5.1.12. Cardamonin {#sec5dot1dot12-cells-09-01451}
Cardamonin (chemically known as 3,4,2,4-tetrahydroxychalcone) is a chalconoid that has been found in a few plants, such as *Alpinia conchigera* Griff. and *Alpinia hainanensis* K.Schum. It has received considerable attention from researchers because of its putative health benefits, including anticancer potential \[[@B127-cells-09-01451]\]. A study investigated the impact of cardamonin on glioblastoma stem cells, and examined its effect on apoptosis and self-renewal and whether its activity is related to the STAT3 pathway in this cancer cell population. CD133-positive (CD133+) glioma stem-like cells (GSCs), a glioblastoma stem cell line was generated from human glioblastoma tissues. In CD133+ GSCs, cardamonin hindered proliferation and promoted apoptosis. The proapoptotic impacts of temozolomide were additionally elevated in vitro via cardamonin in U87 cells and CD133+ GSCs. Cardamonin additionally repressed STAT3 initiation by luciferase assay and stifled the downstream STAT3 gene expressions, for example, Bcl-xL, Mcl-1, Bcl-2, survivin, and VEGF. Besides, cardamonin halted STAT3's travel to the nucleus and dimerization in CD133+ GSCs. Apart from this, molecular docking study affirmed that cardamonin, with a favorable binding energy of −10.78 kcal/mol, was bound to the active sites of STAT3 \[[@B128-cells-09-01451]\]. All these results demonstrate that cardamonin has the ability to be a new anticancer agent for glioblastoma by virtue of its ability to act as a STAT3 inhibitor.
### 5.1.13. Caffeic Acid {#sec5dot1dot13-cells-09-01451}
Caffeic acid and phenethyl ester caffeic acid (CAPE) are phenolic compounds synthesized by different plant organisms. They are present in beverages, such as wine, tea (*Camellia sinensis* (L.) Kuntze) and coffee (*Coffea arabica* L.). Caffeic acid and CAPE are known to have anti-inflammatory, antioxidant, and anticancer properties \[[@B129-cells-09-01451],[@B130-cells-09-01451]\]. Caffeic acid blocked STAT3 phosphorylation by suppressing Src tyrosine kinase, which inhibited the proliferation and caused the downregulation of VEGF in the human renal carcinoma cells \[[@B131-cells-09-01451]\].
### 5.1.14. Silibinin {#sec5dot1dot14-cells-09-01451}
Silibinin, a flavanone present in Silybum marianum (L.) Gaertn. (milk thistle), is an extraordinary hepatoprotective agent with anticancer potential \[[@B132-cells-09-01451],[@B133-cells-09-01451],[@B134-cells-09-01451]\]. Silibinin's benefits were also observed when used in conjunction with traditional chemotherapy since they decreased adverse effects of anticancer drugs (e.g., nephrotoxicity, neurotoxicity, and cardiotoxicity) while preventing or even reversing chemotherapy resistance \[[@B24-cells-09-01451]\]. Silibinin blocked STAT3--DNA attachment, inhibited constitutively active STAT3 phosphorylation in prostate cancer cells, and induced caspase activation that contributed to cell death \[[@B135-cells-09-01451]\]. This research showed that in conjunction with silibinin, piceatannol (a JAK1 inhibitor) lessened STAT3 phosphorylation at Tyr705 and thus triggered caps, which induced the programmed cell death of prostate cancer cells \[[@B135-cells-09-01451]\]. Silibinin has also been shown to suppress STAT3 phosphorylation and thus decrease its transcriptional function in urethane-induced lung tumors in A/J mice \[[@B136-cells-09-01451]\].
### 5.1.15. Butein {#sec5dot1dot15-cells-09-01451}
A flavonoid named butein is isolated from the bark of *Toxicodendron vernicifluum* (Stokes) F.A. Barkley (formerly *Rhus verniciflua* Stokes.) and the flowers of *Butea monosperma* (Lam.) Kuntze. Butein, a multitargeted agent, is considered to have antioxidant, anti-inflammatory, hypotensive, antidiabetic, anticancer, and neuroprotective effects \[[@B137-cells-09-01451],[@B138-cells-09-01451]\]. In multiple myeloma cells, it restrained the constitutive and IL-6-induced activation of STAT3 through inactivation of JAK1 and c-Src in a concentration-dependent fashion \[[@B139-cells-09-01451]\]. Butein also suppressed the growth of xenografted human hepatocellular carcinoma in male nude mice \[[@B140-cells-09-01451]\].
### 5.1.16. 5,7-Dihydroxyflavone {#sec5dot1dot16-cells-09-01451}
5,7-Dihydroxyflavone is a natural flavonoid present in propolis and honey, in addition to various herbs, such as *Passiflora incarnate* L., *Passiflora caerulea* L., and *Oroxylum indicum* (L.) Kurz. Various in vivo and in vitro investigations have reported its potential for chemoprevention and therapy of various cancers \[[@B141-cells-09-01451],[@B142-cells-09-01451]\]. By reducing Akt and STAT3 phosphorylation, 5,7-dihydroxyflavone significantly suppressed the growth of human hepatocellular carcinoma (HepG2) tumor xenografts \[[@B143-cells-09-01451]\].
### 5.1.17. Honokiol {#sec5dot1dot17-cells-09-01451}
Honokiol, a phytochemical present in plants such as *Magnolia officinalis* Rehder & E.H.Wilson, and *Magnolia grandiflora* L., exhibited antiproliferative, proapoptotic, and cell-cycle-modulatory activities in gastric, colon, colorectal, hepatocellular, pancreatic, lung, glioblastoma, melanoma, leukemia, skin, breast, ovarian, renal, prostate, head and neck, mesangial, and squamous cell carcinoma through modulation of multiple oncogenic targets \[[@B144-cells-09-01451],[@B145-cells-09-01451],[@B146-cells-09-01451],[@B147-cells-09-01451],[@B148-cells-09-01451]\]. Honokiol diminished the generation of p-STAT3 tyr705 and p-JAK2 Tyr1007 in human monocytic cell line` (`THP1 cells), and human erythroleukemia cell line (HEL cells) \[[@B149-cells-09-01451]\]. It additionally restrained both constitutive and inducible STAT3 activation, resulting in other downstream impacts, such as the increased expression level of SHP1. However, SOCS1/3 protein levels and knockdown of SHP1 reversed honokiol-induced STAT3 signal hindrance \[[@B149-cells-09-01451]\]. In addition, honokiol decreased the percentage of mRNA of STAT3 target genes (cyclin D1, Bcl-2, surviving, and Bcl-xL) in a concentration-dependent fashion in myeloid leukemia cells \[[@B149-cells-09-01451]\].
### 5.1.18. Casticin {#sec5dot1dot18-cells-09-01451}
Casticin can be found in several plants, such as Achillea millefolium L., Artemisia abrotanum L., Camellia sinensis (L.) Kuntze, Centipeda minima (L.) A.Braun & Asch., Clausena excavate Burm.f., Crataegus pinnatifida Bunge., Croton betulaster Müll Arg., Daphne genkwa Siebold & Zucc., Ficus microcarpa L.f., Nelsonia canescens (Lam.) Spreng., Pavetta crassipes K.Schum., Vitex trifolia subsp. Litoralis steenis (formerly Vitex rotundifolia L.), Vitex agnus-castus L., Vitex negundo L., and Vitex trifolia L. Recent experimental evidence underscores the ability of casticin as a candidate for antineoplastic drugs \[[@B150-cells-09-01451],[@B151-cells-09-01451]\]. It showed antiproliferative, cell-cycle-regulatory, and proapoptotic activities in breast, cervical, colon, colorectal, gastric, liver, leukemia, lung, pancreatic, ovarian, and prostate cancer cells \[[@B152-cells-09-01451],[@B153-cells-09-01451]\]. It also promoted apoptosis, inhibited constitutively active STAT3, modulated STAT3 activation by modifying the activity of upstream STAT3 regulators, and abrogated IL-6-induced STAT3 activation, thereby suppressing the JAK/STAT pathway in various cancer cells, such as tongue squamous cell carcinoma (HN-9), hypertriploid renal cell carcinoma (786-O), and oral squamous cell carcinoma (YD-8) \[[@B154-cells-09-01451]\].
### 5.1.19. Apigenin {#sec5dot1dot19-cells-09-01451}
Found plentifully in many medicinal plants and in many vegetables and fruits, apigenin, a naturally-occurring flavonoid, has several physiological functions \[[@B155-cells-09-01451],[@B156-cells-09-01451]\]. It showed anticancer activity via JAK/STAT pathway inhibition by reducing phosphorylated JAK1/2 and STAT3 in breast cancer (BT-474) cells \[[@B25-cells-09-01451]\]. Apigenin induced programmed cell death of colon cancer cells by hindering STAT3 phosphorylation and hence downregulated the antiapoptotic proteins Mcl-1 and Bcl-xL \[[@B157-cells-09-01451]\].
### 5.1.20. Wedelolactone {#sec5dot1dot20-cells-09-01451}
Wedelolactone is a natural coumarin with promising anticancer activities \[[@B158-cells-09-01451]\]. Wedelolactone, obtained from *Eclipta prostrata* (L.) L. (formerly *Eclipta alba* (L.) Hassk.), and *Sphagneticola calendulacea* (L.) Pruski (formerly *Wedelia calendulacea* (L.) Pruski)*,* exhibited antiproliferative and apoptosis-inducing effects in breast, prostate, neuroblastoma, pancreatic, mammary carcinosarcoma, myeloma, and leukemia cells \[[@B159-cells-09-01451],[@B160-cells-09-01451]\]. This prevented the dephosphorylation of STAT1 and prolonged the activation of STAT1 by mutual inhibition of T-cell protein tyrosine phosphatase (TCPTP) in p-GAS-HepG2 cells \[[@B161-cells-09-01451]\].
5.2. Terpenoids {#sec5dot2-cells-09-01451}
---------------
The terpenoids, also known as isoprenoids, are one broad family of naturally occurring, terpene-derived organic compounds. With more than 40,000 structurally diverse components, the terpenoids are the largest group of secondary metabolites from the plant kingdom. The majority of terpenoids with structural differences are biologically active and are used to treat numerous diseases around the world. Several terpenoids have been developed as anticancer drugs, including taxol. Natural terpenoids and related compounds have inhibited the proliferation of various human carcinoma cells and suppressed the tumor development in experimental animals \[[@B162-cells-09-01451],[@B163-cells-09-01451],[@B164-cells-09-01451]\]. Several phytoconstituents belonging to this category can inhibit the JAK/STAT pathway by inhibiting the phosphorylation of STAT3. Various terpenoids even inhibited the phosphorylation of JAK1, JAK2, and c-Src in various carcinoma cell lines. The role of various terpenoids in the JAK/STAT signaling pathway in various cancers is the focus of the subsequent sections.
### 5.2.1. Cucurbitacins {#sec5dot2dot1-cells-09-01451}
Cucurbitacins are associated with tetracyclic triterpenoids found in the Cucurbitaceae family. There are 17 major molecules in the cucurbitacins group, which range from cucurbitacin A to cucurbitacin T with hundreds of analogs \[[@B165-cells-09-01451]\]. Cucurbitacins and their derivatives have various biological and pharmacological actions, including anticancer effects \[[@B166-cells-09-01451],[@B167-cells-09-01451]\]. Cucurbitacin B ([Figure 3](#cells-09-01451-f003){ref-type="fig"}) demonstrated anti-STAT3 effects in human pancreatic cancer cells \[[@B168-cells-09-01451],[@B169-cells-09-01451]\] and leukemia cells \[[@B170-cells-09-01451]\] ([Table 2](#cells-09-01451-t002){ref-type="table"}). Even though cucurbitacin E decreased the amount of phosphorylated STAT3 in human bladder cancer cells, it did not alter the levels of STAT3 \[[@B171-cells-09-01451]\]. Cucurbitacin I (JSI-124) reduced the phosphorylation of STAT3 and JAK in Src-transformed fibroblast and in human lung carcinoma cells. This subsequently impaired STAT3--DNA binding \[[@B172-cells-09-01451]\]. Likewise, the phosphorylated levels of STAT3 in glioblastoma multiforme cells treated with cucurbitacin I were greatly decreased. This resulted in G2/M cell cycle suspension by downregulation of the expression of cdc2 and cyclin B1 \[[@B173-cells-09-01451]\]. Moreover, the inhibitory impact on JAK/STAT by cucurbitacin I was additionally determined in Sézary (Sz) syndrome \[[@B174-cells-09-01451]\] and anaplastic large cell lymphoma \[[@B175-cells-09-01451]\]. Cucurbitacin Q hindered STAT3 levels; however, it did not influence the activation of JAK \[[@B176-cells-09-01451]\]. Based on the aforementioned reports, various cucurbitacins exhibit distinct mechanisms to function as JAK and STAT inhibitors.
### 5.2.2. Andrographolide {#sec5dot2dot2-cells-09-01451}
Andrographolide, a diterpenoid lactone extracted from *Andrographis paniculata* (Burm.f.) Nees, has strong anticancer and cancer preventive potential \[[@B177-cells-09-01451],[@B178-cells-09-01451]\]. A study performed by Zhou et al. \[[@B179-cells-09-01451]\] analyzed the impact of andrographolide on the STAT3 pathway and assessed whether restraint of STAT3 action by andrographolide could sensitize MDA-MB-231 breast cancer cells to doxorubicin, a chemotherapeutic drug. Andrographolide can stifle both constitutively activated and IL-6-induced phosphorylation of STAT3 and, consequently, its nuclear translocation into cancer cells. Such restraint was accomplished through restraint of JAK1/2 and interaction between STAT3 and gp130. To comprehend the importance of the inhibiting effect of andrographolide on STAT3, the impact of it on doxorubicin-induced apoptosis via promoting caspase activation and apoptotic pathway has been evaluated in human cancer cells (HeLa, HepG2, HCT116, and MDA-MB-231). In that test, the constitutive activation level of STAT3 was found to correspond to the interference of malignant cell growth in doxorubicin-induced apoptosis. The long-term colony formation assay in addition to the short-term cell proliferation assay demonstrated that andrographolide significantly promoted doxorubicin-induced cell death in cancer cells, showing that andrographolide improves the sensitivity of cancer cells to doxorubicin principally by means of STAT3 suppression. These results uncover a novel anticancer function of andrographolide in cancer treatment.
### 5.2.3. Betulinic Acid {#sec5dot2dot3-cells-09-01451}
A natural pentacyclic triterpenoid, betulinic acid is extracted from the exterior bark of white birch plants. In recent years, betulinic acid has gained considerable attention due to its strong cytotoxic activity against a variety of tumor cells \[[@B180-cells-09-01451],[@B181-cells-09-01451]\]. Betulinic acid hindered constitutive activation of STAT3, Src kinase, and JAK1/2 \[[@B182-cells-09-01451]\]. Betulinic-acid-induced expression of the PTP SHP-1 and silencing of the SHP-1 gene nullified the capability of betulinic acid to hinder STAT3 initiation and salvages betulinic-acid-induced cell death. It also downregulated the expression level of STAT3-directed genes, such as survivin, Bcl-2, Bcl-xL, and cyclin D1. There was also a rise in apoptosis as evidenced by cell cycle arrest at sub-G1 phase and increase in PARP cleavage induced by caspase-3. In addition, overactivation of constitutive dynamic STAT3 fundamentally diminished the betulinic acid-induced apoptosis \[[@B182-cells-09-01451]\].
### 5.2.4. γ-Tocotrienol {#sec5dot2dot4-cells-09-01451}
γ-Tocotrienol is one of the four types of tocotrienol, a type of vitamin E. γ-Tocotrienol has established potency in preclinical anticancer research \[[@B183-cells-09-01451],[@B184-cells-09-01451]\]. There are eight forms of vitamin E, each consisting of a head segment associated with either a saturated (phytyl) or unsaturated (farnesyl) tail. γ-Tocotrienol prevented the phosphorylation of molecules involved in the JAK/STAT pathway, including JAK1, JAK2, and c-Src, in human liver cancer cells \[[@B185-cells-09-01451]\]. It also induced the expression of SHP-1, which is one type of tyrosine phosphatase engaged in the JAK/STAT signal's negative regulation \[[@B79-cells-09-01451]\], in liver cancer cells in a concentration-dependent fashion \[[@B185-cells-09-01451],[@B186-cells-09-01451]\].
### 5.2.5. Cryptotanshinone {#sec5dot2dot5-cells-09-01451}
Cryptotanshinone, a quinoid diterpene, has been isolated from *Salvia miltiorrhiza* Bunge. Experimental investigations have demonstrated evidence that cryptotanshinone could be used as an anticancer agent \[[@B187-cells-09-01451]\]. In pancreatic carcinoma cells, cryptotanshinone repressed STAT3 phosphorylation via a mechanism autonomous of JAK2 phosphorylation \[[@B188-cells-09-01451]\]. It was found to bind directly with STAT3 and prevented STAT3's dimerization. Computational modelling also demonstrated that cryptotanshinone has the capacity to bind with STAT3's SH2 domain \[[@B188-cells-09-01451]\].
### 5.2.6. Nimbolide {#sec5dot2dot6-cells-09-01451}
Nimbolide, a triterpene isolated from the flowers and leaves of neem (*Azadirachta indica* A. Juss.), is generally utilized in conventional medicinal practices. Emerging evidence showed that nimbolide possessed significant cytotoxic activities against various types of cancer cells and exhibited chemopreventive potential in several animal tumor models \[[@B189-cells-09-01451],[@B190-cells-09-01451]\]. It can cause cell cycle arrest, most conspicuously at the G1/S stage of glioblastoma multiforme cancer cells that exhibit EGFRvIII \[[@B191-cells-09-01451]\]. EGFRvIII is an oncogene involved in 20% to 25% of cases of glioblastoma multiforme. Nimbolide restrained CDK4/CDK6 kinase action, which led to a decreased phosphorylation of the retinoblastoma protein. This prompted the arrest of the cell cycle at the G1/S phase and cell death \[[@B191-cells-09-01451]\]. Based on another investigation, nimbolide hindered STAT3 signaling and suppressed tumor formation in prostate cancer in vitro \[[@B192-cells-09-01451]\].
### 5.2.7. Celastrol {#sec5dot2dot7-cells-09-01451}
A pentagonal cyclic triterpenoid named celastrol is extracted from the plant *Tripterygium wilfordii* Hook.f. It exhibits various pharmacological actions, and has antiproliferative, antiangiogenic, anti-inflammatory, proapoptotic, and antimetastatic properties \[[@B193-cells-09-01451],[@B194-cells-09-01451]\]. It hindered the growth of multiple myeloma cells that are either sensitive or resistant to various chemotherapeutic drugs \[[@B195-cells-09-01451]\]. It also improved the proapoptotic actions of both bortezomib and thalidomide. In another study, celastrol inhibited STAT3 phosphorylation and STAT3-mediated IL-17 expression, and T-helper 17 (Th17) differentiation and proliferation in multiple myeloma cells \[[@B186-cells-09-01451]\].
### 5.2.8. Ursolic Acid {#sec5dot2dot8-cells-09-01451}
A pentacyclic triterpenoid compound, ursolic acid (3β-hydroxy-urs-12-en-28-oic-acid), is present in many plants (e.g., *Mirabilis jalapa* L.) including fruits and some herbs (e.g., apples, prunes, bilberries, cranberries, basil, lavender, rosemary, peppermint, thyme, elder flower, oregano, and hawthorn). Ursolic acid has shown significant promise in the treatment and prevention of various cancers \[[@B196-cells-09-01451],[@B197-cells-09-01451]\]. By suppressing Src and JAK2 phosphorylation, ursolic acid prevented the activation of STAT3 in prostate carcinoma cells \[[@B198-cells-09-01451]\]. Additionally, ursolic acid decreased xenografted prostate tumor development in transgenic adenocarcinoma of the mouse prostate (TRAMP) model without impacting the body weight via hindering the activation of STAT3 \[[@B198-cells-09-01451]\].
### 5.2.9. Brusatol {#sec5dot2dot9-cells-09-01451}
Brusatol, a natural triterpene lactone called quassinoid, is extracted from the air-borne parts of the *Brucea javanica* (L.) Merr., plant. It has been distinguished as a strong blocker of STAT3 in preventing the progression of neck and head squamous cell carcinoma cells. Brusatol also exhibited cytotoxicity, hindered the activation of STAT3 as well as upstream kinases, including JAK1, JAK2, and Src. It diminished nuclear STAT3 and its DNA binding capability. Therefore, brusatol-mediated inhibition of STAT3 could be valuable in treating neck and head squamous cell carcinoma \[[@B199-cells-09-01451]\].
### 5.2.10. Oridonin {#sec5dot2dot10-cells-09-01451}
Oridonin is a phytochemical present in *Isodon rubescens* (Hemsl.) H.Hara. Oridonin and its analogs have shown the ability to treat inflammation, neurodegeneration, and cancer \[[@B200-cells-09-01451],[@B201-cells-09-01451]\]. Oridonin inhibited the proliferation of astrocytoma, breast, colorectal, hepatoma, leukemia, lung, multiple myeloma, ovarian, pancreatic, and prostate cancer cells \[[@B202-cells-09-01451],[@B203-cells-09-01451],[@B204-cells-09-01451],[@B205-cells-09-01451]\]. It also decreased IL-6, STAT3, and p-STAT3 expression levels, implying that oridonin can downregulate the STAT3 pathway, which is accountable for the growth, invasion, and metastasis of human pancreatic carcinoma cells \[[@B206-cells-09-01451]\].
### 5.2.11. Thymoquinone {#sec5dot2dot11-cells-09-01451}
Thymoquinone is the active component of black cumin (Nigella sativa L.) seed oil, and preclinical studies have established its cancer prevention effects \[[@B26-cells-09-01451],[@B207-cells-09-01451]\]. Thymoquinone inhibited STAT3 phosphorylation in multiple myeloma cells, U266 and RPMI 8226 \[[@B208-cells-09-01451],[@B209-cells-09-01451]\]. Thymoquinone caused downregulation of STAT3 activation in addition to a decrease in activity of JAK2 and c-Src in human gastric cancer cells. It also downregulated the expression of STAT3-controlled genes, namely survivin, cyclin D, Bcl-2, and the vascular endothelial growth factor (VEGF), and activated caspase-3, caspase-7, and caspase-9 \[[@B208-cells-09-01451]\].
### 5.2.12. Parthenolide {#sec5dot2dot12-cells-09-01451}
Parthenolide, a sesquiterpene, is obtained from natural sources, such as *Tanacetum parthenium* (L.) Sch. Bip., *Tanacetum vulgare* L., *Tanacetum larvatum* (Pant.) Hayek., *Centaurea ainetensis* Boiss., and *Helianthus annuus* L. Parthenolide, a potent anticancer and anti-inflammatory agent, is presently being assessed in clinical research for cancer treatment \[[@B210-cells-09-01451]\]. It hindered the proliferation of bladder, pancreatic, lung, breast, skin, melanoma, glioma, skin, liver, and gastric cancer cells \[[@B211-cells-09-01451],[@B212-cells-09-01451],[@B213-cells-09-01451]\]. By suppressing JAK2 kinase activity, parthenolide also suppressed STAT3 phosphorylation induced by IL-6 in breast cancer (MDA-MB-231) cells \[[@B214-cells-09-01451]\].
### 5.2.13. Dihydroartemisinin {#sec5dot2dot13-cells-09-01451}
Dihydroartemisinin, is an analogue of artemisinin extracted from *Artemisia annua* L. Artemisinin and several of its analogs, including dihydroartemisinin, have been found to induce tumor cell death via inhibiting numerous tumor-related signal transduction pathways \[[@B215-cells-09-01451],[@B216-cells-09-01451]\]. Dihydroartemisinin has been found to have potent cytotoxic and proapoptotic activity by suppressing activated JAK2/STAT3 signals. It also suppressed downstream targeted proteins in the human tongue (Cal-27), colon (HCT-116), liver (HepG2) as well as head and neck (FaDu) carcinoma cell lines \[[@B217-cells-09-01451],[@B218-cells-09-01451]\].
### 5.2.14. Alantolacton {#sec5dot2dot14-cells-09-01451}
Alantolactone, a sesquiterpene, is obtained from several herbs, including *Aucklandia costus* Falc. (formerly *Aucklandia lappa* Decne.), *Inula helenium* L., *Inula japonica* Thunb. and *Inula racemose* Hook.f. It exhibits various pharmacological properties, including anti-inflammatory, antimicrobial, and anticancer activities, without significant toxicity \[[@B219-cells-09-01451]\]. It hindered the proliferation of glioblastoma, leukemia, prostate, colon, liver, and pulmonary carcinoma cells \[[@B220-cells-09-01451],[@B221-cells-09-01451]\]. The activation of STAT3 at tyrosine 705 was greatly decreased by alantolactone in MDA-MB-231 triple-negative breast carcinoma cells. Alantolactone prevented STAT3 from traveling into the nucleus and thus affecting transcription of DNA \[[@B222-cells-09-01451]\].
### 5.2.15. β-Caryophyllene oxide {#sec5dot2dot15-cells-09-01451}
β-Caryophyllene oxide is a natural bicyclic sesquiterpene. It is present in numerous essential oils, including clove oil, the oil of *Syzygium aromaticum* (L.) Merr. & L. M. Perry (cloves) stems and flowers, as well as the essential oils of *Cannabis sativa* L. (Cannabis), *Psidium guajava* L. (guava), *Origanum vulgare* L. (oregano), *Salvia Rosmarinus* Spenn. (rosemary), and *Humulus lupulus* L. The potential therapeutic utility β-caryophyllene oxide is focused mainly on its analgesic and anticancer activities \[[@B223-cells-09-01451]\]. A study demonstrated that the anticancer effect of β-caryophyllene oxide was mediated through interfering with the activation of STAT3 in carcinoma cells, such as multiple myeloma cells (U266) and human prostate cancer cells (DU145). The impact of β-caryophyllene oxide on STAT3 activation, related protein kinases and phosphatases, STAT3-directed gene products, and apoptosis was investigated utilizing both functional proteomics and different tumor cell lines. It was found that β-caryophyllene oxide decreased the constitutive STAT3 activation in multiple myeloma, prostate, and breast cancer cells, and has a critical concentration- and time-dependent response in multiple myeloma cells. The restraint was mediated by the inhibition of upstream kinase c-Src and JAK1/2 activation. β-Caryophyllene oxide brought about the expression of SHP-1, which is connected with the downregulation of constitutive STAT3 activation. The STAT3 activation by β-caryophyllene oxide hindered proliferation, increased apoptosis, and blunted the invasive capability of DU145 prostate carcinoma cells \[[@B224-cells-09-01451]\].
5.3. Alkaloids {#sec5dot3-cells-09-01451}
--------------
Alkaloids are natural organic compounds that possess numerous pharmacological actions, including chemopreventive and chemotherapeutic effects \[[@B225-cells-09-01451],[@B226-cells-09-01451]\]. These components exist in diverse plants, and several alkaloids inhibit the JAK/STAT pathway, as described below.
### 5.3.1. Capsaicin {#sec5dot3dot1-cells-09-01451}
Capsaicin ([Figure 4](#cells-09-01451-f004){ref-type="fig"}), an active element of chili peppers, exhibited strong anticancer properties in diverse cancer types \[[@B227-cells-09-01451],[@B228-cells-09-01451],[@B229-cells-09-01451]\]. Capsaicin also hindered IL6-induced STAT3 activation. It impaired the advancement of multiple myeloma xenograft tumors in male thymic mice by suppressing STAT3 activation by inactivating JAK1 and c-Srcin in various preclinical cancer models \[[@B139-cells-09-01451]\] ([Table 3](#cells-09-01451-t003){ref-type="table"}).
### 5.3.2. Evodiamine {#sec5dot3dot2-cells-09-01451}
Evodiamine, obtained from *Tetradium ruticarpum* (A. Juss.) T.G. Hartley., is a multitargeted anticancer agent \[[@B230-cells-09-01451]\]. In hepatocellular carcinoma cells (HepG2), evodiamine effectively hindered constitutive and IL-6-induced activation of STAT3 (Tyr705) phosphorylation. Evodiamine suppressed the phosphorylation of JAK2, Src, and extracellular signal-regulated kinases (ERK1/2). Additionally, evodiamine impeded STAT3--DNA binding activity \[[@B81-cells-09-01451]\]. The proliferation of HepG2 cells was inhibited by evodiamine via cell cycle arrest in the G2/M phase. It also downregulated the protein expression levels of cyclin D1, Bcl-2, Mcl-1, survivin, VEGF, XIAP, MMP-9, and HIF1-α in SMMC-7721 and HepG2 hepatocarcinoma cells \[[@B231-cells-09-01451]\].
### 5.3.3. Indirubin {#sec5dot3dot3-cells-09-01451}
Indirubin, a naturally occurring alkaloid, is isolated from indigo dye-containing herbs (*Angelica sinensis* (Oliv.) Diels). It is an important potent component of "Danggui Longhui Wan", a traditional Chinese medicine formulation used for the therapeutic intervention of inflammation, cancer, and various other chronic diseases. Recent research shows the potential of indirubin and related compounds in the treatment of cancer, especially inflammation-associated malignancies \[[@B232-cells-09-01451],[@B233-cells-09-01451]\]. It is thought to hinder numerous kinases, and it might be utilized to treat chronic myelocytic leukemia, cancer, and neurodegenerative disorders. Indirubin has been shown to impede cyclin-dependent kinase dimers, triggering cell cycle arrest at G1/S or G2/M phase. Additionally, it repressed phosphorylated STAT3 expression \[[@B234-cells-09-01451]\].
5.4. Saponins {#sec5dot4-cells-09-01451}
-------------
These components exist in diverse plants, and several alkaloids inhibit the JAK/STAT pathway as described below. Saponins are glycosidic secondary metabolites present in plants, grains, soybeans and certain herbs. Saponins may have one to three straight/branched chains of sugar (either D-glucose, D-xylose, D-galactose, L-arabinose L-rhamnose, or D-glucuronic acid). In vitro enzymatic assays and preclinical animal studies reveal that the saponins possess anticancer properties \[[@B235-cells-09-01451]\]. Interestingly, various saponins are known to augment the efficiency of several chemotherapeutic drugs, such as cisplatin, cyclophosphamide, docetaxel, doxorubicin, mitoxantrone, and paclitaxel \[[@B236-cells-09-01451]\]. The anticancer actions of saponins could be linked to their capability to inhibit the JAK/STAT pathway.
### β-Escin
β-Escin, a triterpene--saponin mixture, is extracted from the seeds of horse chestnut (*Aesculus hippocastanum* L.). β-Escin displayed anticancer action in various cancer models and increased the impact of chemotherapeutic agents \[[@B237-cells-09-01451]\]. β-Escin inhibited JAK1/2 and c-Src phosphorylation in HCC cells \[[@B238-cells-09-01451]\]. β-Escin also downregulated the expression levels of STAT3-regulated genes, for example, Bcl-2, Bcl-xL, cyclin D1, Mcl-1, survivin, and VEGF \[[@B238-cells-09-01451]\].
5.5. Steroids {#sec5dot5-cells-09-01451}
-------------
Various plant-derived steroids, including phytosterols, possess anticarcinogenic properties \[[@B239-cells-09-01451]\], and the intake of high phytosterol was inversely related to an elevated cancer risk \[[@B240-cells-09-01451]\].
### 5.5.1. Diosgenin {#sec5dot5dot1-cells-09-01451}
Diosgenin is a steroidal sapogenin that occurs in nature. It is a bioactive component present in fenugreek (*Trigonella foenum-graecum* L.) seeds. Studies show that diosgenin might confer several health benefits, including cancer preventive and anticancer effects \[[@B241-cells-09-01451],[@B242-cells-09-01451],[@B243-cells-09-01451]\]. Diosgenin suppressed the phosphorylation of *c*-Src, JAK1, and JAK2 in HCC cell lines (C3A and HepG2) \[[@B244-cells-09-01451]\]. This phytochemical increased the expression of SH-PTP2, and consequently inhibited STAT3 activation triggered by IL-6 \[[@B244-cells-09-01451]\]. It also hindered the proliferation of primary human thyrocytes by inducing apoptosis and causing cell cycle arrest in the G0/G1 phase \[[@B245-cells-09-01451]\].
### 5.5.2. Ergosterol Peroxide {#sec5dot5dot2-cells-09-01451}
Ergosterol peroxide, derived from many plants and edible mushrooms, is known to possess immune-suppressive, antimicrobial, anti-inflammatory and antitumor effects \[[@B246-cells-09-01451]\]. It inhibited the phosphorylation of STAT3 and consequently reduced the expression level of phospho-STAT3 and CD-34 (an angiogenesis marker). Ergosterol peroxide also suppressed STAT3--DNA binding in multiple myeloma cells (U266). Moreover, it blocked phosphorylation of tyrosine kinases JAK2 and Src, but elevated the expression level of SHP-1 at the transcriptional level \[[@B80-cells-09-01451]\].
### 5.5.3. Guggulsterone {#sec5dot5dot3-cells-09-01451}
Guggulsterone, an antagonist of the farnesoid X receptor, is a phytosterol present in the gum resin of guggul plant (*Commiphora mukul* (Hook. ex Stocks) Engl.). Guggulsterone is broadly utilized in Indian traditional medicine for the management of various diseases, including cancer \[[@B247-cells-09-01451],[@B248-cells-09-01451],[@B249-cells-09-01451]\]. It suppressed U266 cell proliferation by inducing apoptosis and cell cycle arrest at the sub-G~1~ phase. It decreased the levels of Bcl-xL, Bcl-2, cyclin D1, Mcl-1, and VEGF. It activated caspase-3 and cleavage of PARP protein. In particular, guggulsterone abrogated the stimulation of *c*-Src and JAK2 and therefore inactivates STAT3 in human multiple myeloma cells \[[@B250-cells-09-01451]\]. It also suppressed the activation of STAT3 triggered by IL-6. Guggulsterone also elevated the expression of SHP-1, a nontransmembrane protein and negative regulator of the JAK/STAT pathway \[[@B251-cells-09-01451]\].
5.6. Lignan {#sec5dot6-cells-09-01451}
-----------
Lignans are present in an assortment of natural products, including pumpkin seeds, flax seeds, sesame seeds, broccoli, soybean, and various berries. Secoisolariciresinoldiglucoside, the primary lignan found in flax seeds, is metabolized in the colon to produce mammalian lignans, such as enterodiol and enterolactone. Once lignan is ingested, it suppresses the development of malignant tumors, specifically hormone-sensitive ones, such as neoplasms of the prostate, breast, and endometrium \[[@B252-cells-09-01451]\].
### Arctiin
Arctiin, anarctigenin glucoside, is found in numerous herbs of Asteraceae family, especially *Centaurea imperialis* Hausskn. ex Bornm. and *Arctiumlappa* L. (greater burdock), and in *Forsythia viridissima* Lindl., *Himalaiella heteromalla* (D.Don) Raab-Straube, and *Trachelospermum asiaticum* (Siebold & Zucc.) Nakai. Arctiin drew the attention of natural product researchers due to its significant therapeutic benefits in relation to inflammation and malignancy \[[@B253-cells-09-01451]\]. In human multiple myeloma cells, it suppressed the constitutive activation of STAT3 phosphorylation at tyrosine 705 residue. Arctiin abrogated the constitutive activation of Src phosphorylation and the activation of JAK1/2. Moreover, treatment of U266 cells with arctiin elevated the level of protein tyrosine phosphatase ε (PTPε), and the silencing of PTPε produced an inversion of the arctiin-induced PTPε expression and the blockage of STAT3 phosphorylation. Arctiin additionally inhibited the proliferation of U266 carcinoma cell line via apoptosis induction and halting cells at G2/M phase \[[@B254-cells-09-01451]\].
5.7. Phytoalexin {#sec5dot7-cells-09-01451}
----------------
Phytoalexins are produced in plants when they are attacked by invading organisms. They act as toxins and are formed when the host plants come in contact with parasites. Phytoalexins have been found in at least 75 plants, including soybeans, cruciferous vegetables, tomatoes, beans, rice, garlic, potatoes, and grapes. In preclinical studies, phytoalexins demonstrated anticancer actions by hindering proliferation, invasion and metastasis, hormonal stimulation, and modulatory effects on expression of xenobiotic-metabolizing enzymes \[[@B255-cells-09-01451]\].
### Brassinin
Brassinin, a phytoalexin present in cruciferous vegetables, displayed antiproliferative, anticancer, and chemopreventive properties \[[@B256-cells-09-01451]\]. In A549 human lung carcinoma cells, it inhibited constitutive and IL-6-inducible STAT3 activation. Additionally, brassinin induced protein inhibitors of activated STATs-3 (PIAS-3) protein and mRNA, while SOCS-3 expression was diminished. Ablation of PIAS-3 by small interfering RNA inhibited brassinin-mediated cytotoxicity and the hindrance of STAT3. Suppressor of cytokine signaling-3 (SOCS-3), when overexpressed in brassinin-treated cells, increases phosphorylation of STAT3 and the viability of the cell. Brassinin downregulated the expression of STAT3-directed genes (phospho-STAT3, Ki-67, and CD31), suppressed proliferation and invasion, and induced apoptosis in the xenograft lung cancer (A549) mouse model. When administered intraperitoneally, the combination of paclitaxel and brassinin diminished tumor growth by the downregulation of phospho-STAT3, CD31, and Ki-67 in a xenograft lung cancer model \[[@B257-cells-09-01451]\].
6. Conclusions and Future Perspectives {#sec6-cells-09-01451}
======================================
Cancer is a dangerous health risk for people worldwide. The morbidity and the mortality rates connected with cancer are alarming, despite the existence of multiple treatment modalities for patients suffering from this disease. Natural substances are important components that can be used for the discovery and the progression of new anticancer medications. Consumption of dietary and medicinal plants that contain active phytoconstituents has minimal or negligible adverse effects. Natural components, when extracted, purified, concentrated, and administered at higher therapeutic doses, very often exhibit adverse side effects apart from their beneficial anticancer activities. Hence, the development of these phytoconstituents can only be progressed further following the confirmation of their safety based on toxicity studies in different preclinical animal models. Once a safety profile is established, the utilization of these substances to develop novel anticancer agents would provide a promising choice for chemoprevention and novel cancer treatment. The current therapeutic intervention for the management of cancer poses several limitations due to the use of monotargeted synthetic agents, elevated cost, low effectiveness, and dangerous adverse actions. Consequently, it is necessary to develop novel and innovative medications for prevention and treatment of cancer. Phytochemicals, including phenolics and polyphenols, terpenoids, alkaloids, saponins, and steroids, are known to prevent the development of many cancers.
JAK/STAT is a vital signaling pathway implicated in the proliferation, differentiation, apoptosis evasion, and survival of neoplastic cells. In addition, its aberrant activation results in cancer-promoting mechanisms. This review indicates the capacity of several phytochemicals to act as multitargeted agents of JAK/STAT signaling inhibition in order to impede cancerous cell growth. Most of the phytochemicals examined in this review have been found to decrease the activity of JAK2 in addition to diminishing STAT3 activation ([Figure 5](#cells-09-01451-f005){ref-type="fig"}). In addition, these phytochemicals reduce the expression of STAT3-regulated genes, namely VEGF, Bcl-2, Mcl-1, cyclin D, and survivin, and activate caspase-3, caspase-7, and caspase-9. Phenolics and polyphenols, terpenoids, alkaloids, saponins, and steroids present in natural sources inhibit the JAK/STAT signaling pathway and thereby inhibit cancer. The phenolic compounds obtained naturally include resveratrol, curcumin, EGCG, chalcone, caffeic acid, silibinin, butein, dihydroartemisinin, 5,7-dihydroxyflavone, honokiol, casticin, apigenin, and wedelolactone. All these phytoconstituents are efficient in arresting the progression of cancer cells. In addition, terpenoids, such as cucurbitacins (B, I, and Q), γ-tocotrienol, cryptotanshinone, celastrol, ursolic acid, pseudolaric acid B, oridonin, thymoquinone, parthenolide, and alantolactone, also have anticarcinogenic action against numerous malignant cells. Several alkaloids (such as evodiamine) and saponins (e.g., β-escin) also show antitumorigenic activity. Naturally occurring steroidal moieties (e.g., diosgenin, ergosterol peroxide, and guggulsterone) also exhibit anticancer activity. The aforementioned phytoconstituents act as negative regulators of the JAK/STAT pathway, which is a vital oncogenic signaling pathway for development and progression of cancer. Members of the JAK family are more commonly targeted than STATs by the negative regulators. JAK is STAT's upstream protein, and therefore must be activated before STAT can be activated as well. Other signaling pathways that are independent of JAK can also reportedly activate the STATs \[[@B258-cells-09-01451]\]. Additionally, the anticancer effects of the various phytochemicals, as described in this review, may be achieved through a variety of complementary and multitargeted mechanisms, including modulation of yet unidentified targets, and one of these mechanisms could be the inhibition of the JAK/STAT signaling pathway.
Most of the current literature on the effects of bioactive phytoconstituents on JAK/STAT signaling in cancer is limited to in vitro studies. Further in vivo studies must be carried out to confirm the potential for bioactive constituents in suppressing and/or treating cancer by inhibiting the JAK/STAT signaling pathway. The limitation of many of these plant chemical components includes low aqueous solubility, limited absorption, and inadequate bioavailability, which may restrict their therapeutic use in clinical settings. It is indeed challenging to achieve sufficient concentration and/or bioavailability of a phytochemical via dietary consumption of its corresponding source food comparable to concentrations used in a research context. Human beings would need to eat an extraordinary amount of plant-based foods, which may not be feasible, to obtain the same results as achieved in in vitro or in vivo studies. However, it is also true that, unlike drugs, which may be administered at a specific dose and at a particular time under controlled conditions, phytochemicals are generally consumed at low and variable levels as a part of a complex dietary regimen at irregular intervals over a protracted period of time. Nevertheless, alternative or coadjuvant dietary interventions might help in the prevention and/or treatment of cancer.
Further research studies need to focus on investigation of the long-term impacts of phytochemical usage, the cause of the JAK/STAT pathway inhibition, the impact of phytochemicals to prevent cancer in high-risk populations, and their effects when used in combination with existing chemotherapy and when used in conjunction with diversified phytoconstituents. The impressive amount of research findings presented here establishes the promise of phytochemicals as anticancer agents and the necessity for implementation of phytochemicals in human clinical trials. Regardless, due to the side effects and resistance to current chemotherapeutic agents, phytochemicals must be further investigated in order to combat the cancer epidemic throughout the world.
Chemopreventive phytochemicals responsible for the inhibition of the JAK/STAT pathway have been discussed throughout this article. A coordinated effort to discover new targets to boost the capability to restore normal JAK/STAT signaling will result in future chemopreventive and anticancer drugs. The findings of the studies presented in this review may enable researchers to create novel and effective cancer prevention and therapeutic approaches.
The authors of this manuscript wish to express their sincere thanks to all researchers who made significant progress in our understanding of anticancer potential of phytochemicals with modulation of JAK/STAT signaling and sincerely regret for not being able to cite every relevant publication due to space limitation.
Conceptualization, A.M. and S.B. (Sankhadip Bose); Writing---original draft preparation, A.M., S.B. (Sankhadip Bose), U.C., and S.B. (Sabyasachi Banerjee); Writing---Review and editing, A.M., J.P., C.R.C. and A.B.; Visualization, J.P. and C.R.C.; Supervision, A.M. and A.B. All authors have read and agreed to the published version of the manuscript.
This research received no external funding.
The authors declare no conflict of interest.
{#cells-09-01451-f001}
{#cells-09-01451-f002}
{#cells-09-01451-f003}
{#cells-09-01451-f004}
{#cells-09-01451-f005}
cells-09-01451-t001_Table 1
######
Potential anticancer activities of phenolics and polyphenols correlated with inhibition of the JAK/STAT signaling pathway.
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Phytochemical Name Sources Anticancer Effect Mechanism of Action EC~50~/IC~50~ References
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------- ------------------------------------- --------------------------------------------------------------------------------------------------------------------------------------------
Resveratrol *Vitis vinifera* L., *Vaccinium macrocarpon* Aiton and *Arachis hypogaea* L. Inhibited proliferation in human epidermoid carcinoma (A431) cells ┴ Phosphorylation of JAK that prevented STAT1 phosphorylation 40 µM Madan et al., 2008 \[[@B86-cells-09-01451]\]
Inhibited proliferation in human multiple myeloma (U266 and RPMI 8226) cells ┴ Constitutive and inducible STAT3 activation 50 and 50 µM Bhardwaj et al., 2007 \[[@B87-cells-09-01451]\]
Inhibited tumor growth, induction of cytotoxicity, cell cycle arrest of v-Src-transformed mouse fibroblasts (NIH3T3/v-Src) at G0-G1 phase. Showed cytotoxicity in human breast cancer (MDA-MB-231), pancreatic carcinoma (Panc-1), and prostate carcinoma (DU145) cells ┴ Src tyrosine kinase activity;\ 40, 70. and 25 µM Kotha et al., 2006 \[[@B88-cells-09-01451]\]
┴ constitutive STAT3 activation
Curcumin *Curcuma longa* L. Induction of cytotoxicity in human multiple myeloma (U266, RPMI 8226, and MM.1S) cells ┴ Constitutive and IL-6-inducible STAT3 phosphorylation;\ 13.8, 12.7 and 10 µM Bharti et al., 2003 \[[@B95-cells-09-01451]\]
┴ IFN-inducible STAT1 phosphorylation
Inhibited proliferation in human small cell lung cancer (NCI-H446 and NCI-1688) cells ┴ STAT3 phosphorylation; ↓ STAT3 downstream gene expression 15 and 15 µM Yang et al., 2012 \[[@B96-cells-09-01451]\]
Antitumor activity in T-cell leukemia, treatment causes growth arrest with increase in apoptosis ┴ Activation of the JAK/STAT pathway;\ Not specified Rajasingh et al., 2006 \[[@B97-cells-09-01451]\]; Kunnumakkara et al., 2017 \[[@B98-cells-09-01451]\]
↓ JAK and STAT phosphorylation
Ascochlorin *Ascochyta viciae* Inhibited cell migration and invasion in U373MG and A172 cancer cells ┴ MMP-9 expression,\ 10 μM Cho et al., 2018 \[[@B100-cells-09-01451]\]
┴ MMP-2 gelatinolytic action and expression;\
┴ JAK2/STAT3 phosphorylation, cancer cell migration, and nuclear translocation of STAT3
Bergamottin Grape juice, lime, lemon, and bergamot oils Antitumor activity in U266 cells, induction of substantial apoptosis at sub-G1 stage ┴ Constitutive STAT3 activation;\ 100 μM Kim et al., 2014 \[[@B102-cells-09-01451]\]
┴ Phosphorylation of JAK 1/2 and c-Src
Capillarisin *Artemisia capillaries* Thunb. Negative regulator of growth and metastasis in human multiple myeloma cells; induces apoptosis; downregulated the expression level of various STAT3-regulated proteins ┴ Constitutive and inducible STAT3 activation at tyrosine 705;\ Not mentioned Lee et al., 2014 \[[@B103-cells-09-01451]\]
┴ STAT3 phosphorylation;\
┴ Activation of JAK1, JAK2, and c-Src kinases
Bavachin *Psoralea corylifolia* Linn.\ Induced cytotoxicity in multiple myeloma cell lines; induced apoptosis by activation of caspase-3 and caspase-9;\ ┴ Activation of STAT3;\ 10 μM Takeda et al., 2018 \[[@B104-cells-09-01451]\]
\[*Cullen corylifolium* (L.) Medik.\] inhibited activation of NF-κB, expression levels of Bcl-2, X-linked inhibitor of apoptosis protein (XIAP), survivin, and B cell lymphoma-extra large (Bcl-xL) ┴ Phosphorylation of STAT3
Epigallocatechin-3-gallate (EGCG) *Camellia sinensis* (L.) Kuntze Anticancer activity in human pancreatic (AsPC-1 and PANC-1), breast (T47D), head and neck cancer (YCU-H861) cells ┴ Phosphorylation and expression of both JAK3 and STAT3 proteins 40, 40, 14.17 µM, 1 µg/mL Tang et al., 2012 \[[@B109-cells-09-01451]\]
Emodin *Rheum palmatum* L. Stimulated the antiproliferation activity of interferon α/β in cervical carcinoma cell line (HeLa) and antitumor activity in Huh7 (hepatocellular cancer cell)-bearing mice in vivo ┴ STAT3 activation,\ Not specified, 1.22 μM He et al., 2016 \[[@B117-cells-09-01451]\]
┴ 26S proteasome activity\
↑ STAT1 phosphorylation, ┴ STAT3 phosphorylation
Chalcone *Malus domestica* (Suckow) Borkh. and *Solanum lycopersicum* L. Induced cytotoxicity in Bovine aortic endothelial cells (BAEC) ┴ IL-6-induced and LPS-induced STAT3 phosphorylation 0.0069 µM Liu et al., 2007 \[[@B120-cells-09-01451]\]
Formononetin *Astragalus mongholicus* Bunge Inhibited proliferation and invasion of colon carcinoma cell lines HCT116 and SW1116, cell cycle arrest at the G0/G1 stage ┴ STAT3 signaling pathway;\ Not specified Wang et al., 2018 \[[@B123-cells-09-01451]\]
┴ STAT3 phosphorylation
Garcinol *Garcinia indica* (Thouars.) Choisy Induced cytotoxicity in hepatocellular carcinoma cells,\ ┴ Constitutive and IL-6 inducible STAT3 activation Not specified Sethi et al., 2014 \[[@B126-cells-09-01451]\]
cell cycle arrest at G0/G1 phase
Cardamonin *Alpinia conchigera* Griff., and *Alpinia hainanensis* K. Schum. Antitumor activity in U87 cells (in vitro) and CD133+ GSCs (in vitro and in vivo); induced apoptosis ┴ STAT3 signaling pathway;\ Not specified Wu et al., 2015 \[[@B128-cells-09-01451]\]
┴ STAT3 activation;\
┴ downstream STAT3 gene expressions (i.e., Bcl-xL, Bcl-2, Mcl-1, survivin, and VEGF)
Caffeic acid or its derivative 3-(3,4-dihydroxy-phenyl)-acrylic acid 2-(3,4-dihydroxy-phenyl)-ethyl ester (CADPE) *Camellia sinensis* L. Kuntze., and *Coffea arabica* L. Induction of cytotoxicity in human renal carcinoma (Caki-1) cells ┴ STAT3 phosphorylation;\ 30 µM Jung et al., 2007 \[[@B131-cells-09-01451]\]
┴ Src tyrosine kinase.
Silibinin *Silybum marianum* (L.) Gaertn. Induction of cytotoxicity in human prostate cancer (DU145) cells ┴ Constitutively active STAT3, ↑ apoptosis and ┴ constitutive STAT3--DNA binding 50 µM Agarwal et al., 2007 \[[@B135-cells-09-01451]\]
Suppressed transcriptional function in urethane-induced lung tumors in A/J mice ┴ STAT3 phosphorylation Not specified Tyagi et al., 2009 \[[@B136-cells-09-01451]\]
Butein *Toxicodendron vernicifluum* (Stokes) F.A. Barkley and *Butea monosperma* (Lam.) Kuntze Exhibited antitumor activity in human hepatocellular carcinoma (HepG2 and SNU-387) ┴ Constitutive and IL-6- induced STAT3 activation by inactivating JAK1 and c-Src. 50 and 50 µM Bhutani et al., 2007 \[[@B139-cells-09-01451]\]; Rajendran et al., 2011 \[[@B140-cells-09-01451]\]
5,7-Dihydroxyflavone *Passiflora incarnate* L., *Passiflora caerulea* L., and *Oroxylum indicum* (L.) Kurz Inhibited proliferation HepG2 tumor xenografts in vivo ↓ Phosphorylation of STAT3 20 µM Zhang et al., 2013 \[[@B143-cells-09-01451]\]
Honokiol *Magnolia officinalis* Rehder & E. H. Wilson and *Magnolia grandiflora* L. Inhibited proliferation, induced apoptosis in human leukemic cell lines (HEL and THP1), multiple myeloma cells (U266) and murine myeloid cell (32D) ┴ Constitutive and inducible STAT3 activation;\ 40 µM Bi et al., 2015 \[[@B149-cells-09-01451]\]
┴ mRNA levels of STAT3 target genes in a concentration-dependent manner;\
↓ Nuclear translocation of STAT3
Casticin *Achillea millefolium* L., *Artemisia abrotanum* L., *Vitex trifolia* subsp. *Litoralis Steenis*, *Camellia sinensis* (L.) Kuntz, *Centipeda minima* (L.) A.Braun & Asch, *Clausena excavate* Burm.f., *Crataegus pinnatifida* Bunge, *Croton betulaster* Müll Arg., *Daphne genkwa* Siebold & Zucc., *Ficus microcarpa* L.f., *Nelsonia canescens* (Lam.) Spreng., *Pavetta crassipes* K. Schum., *Vitex trifolia* subsp. *Litoralis steenis*, *Vitex agnus-castus* L., *Vitex negundo* L. and *Vitex trifolia* L. Inhibited proliferation, induced apoptosis, cell cycle arrest at G2/M phase in colon (Panc-1), breast (MCF-7), lung (A549), gastric (SGC-7901), ovarian (SKOV3), liver (HepG2), leukemia (K562) cancer cells ┴ Constitutively active STAT3 and modulates STAT3 activation by modifying upstream STAT3 regulator activity. 10, 8.5, 14.3, 1, 2.18, 30, 5.95 µM Chen et al., 2011 \[[@B152-cells-09-01451]\]; Zeng et al., 2012 \[[@B153-cells-09-01451]\]
Induced apoptosis in 786-O, YD-8, and HN-9 cancer cells ┴ constitutively activation of STAT3;\ 5 μM Lee et al., 2019 \[[@B154-cells-09-01451]\]
Modulated STAT3 activation by altering the activity of upstream STAT3 regulators, and abrogated IL-6-induced STAT3 activation;\
┴ JAK/STAT pathway
Apigenin *Petroselinum crispum* (Mill.) Fuss., *Apium graveolens* L., and *Matricaria chamomilla* L. Anticancer and antitumor activity in colon cancer (HCT-116) cells ┴ Phosphorylation of STAT3 and consequently downregulated the antiapoptotic proteins Bcl-xL and Mcl-1 47.33 µM Ozbey et al., 2018 \[[@B25-cells-09-01451]\]; Maeda et al., 2018 \[[@B157-cells-09-01451]\]
Anticancer activity in BT-474 (breast cancer) cells ┴ JAK/STAT pathway\ Not specified Ozbey et al., 2018 \[[@B25-cells-09-01451]\]
┴ STAT3 nuclear accumulation\
┴ Phosphorylation of JAK1/2, and STAT3
Wedelolactone *Eclipta prostrate* (L.) L. and *Sphagneticola calendulacea* (L.) Pruski Inhibited proliferation, induced apoptosis, causes cell cycle arrest at S and G2/M phases in breast (MDA-MB-231) and HepG2 cancer cells ┴ STAT1 dephosphorylation and prolonging STAT1 activation, ┴ T-cell protein tyrosine phosphatase Not specified in MDA-MB-231;\ Benes et al., 2011 \[[@B159-cells-09-01451]\]; Benes et al., 2012 \[[@B160-cells-09-01451]\]; Chen et al., 2013 \[[@B161-cells-09-01451]\]
20 µM (EC~50~) in HepG2 cells
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Various symbols (↑, ↓ and ┴) indicate increase, decrease and inhibition in the obtained variables, respectively.
cells-09-01451-t002_Table 2
######
Potential anticancer activities of terpenoids related to inhibition of the JAK/STAT signaling pathway.
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Phytochemical Name Sources Anticancer Effect Mechanism of Action EC~50~ /IC~50~ References
-------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------- ----------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------
Cucurbitacin B *Hemsleya endecaphylla* C.Y. Wu Induced cytotoxicity in human pancreatic cancer (MiaPaCa-2, AsPC-1) cells ┴ JAK2, ┴ STAT3, and ┴ STAT5 activation 0.278, 0.017 μM Thoennissen et al., 2009 \[[@B168-cells-09-01451]\]; Zhou et al., 2017 \[[@B169-cells-09-01451]\]
Induced cytotoxicity in leukemia K562 cells ┴ STAT3 activation 50 µM Chan et al., 2010 \[[@B170-cells-09-01451]\]
Cucurbitacin E *Wilbrandia ebracteate* Cogn. Induced cytotoxicity in human bladder cancer (T24) cells ↓ Levels of phosphorylated STAT3 (p-STAT3) 1012 nM Huang et al., 2012\
\[[@B171-cells-09-01451]\]
Cucurbitacin I *Cucumis melo* L. Induced cytotoxicity in human lung adenocarcinoma (A549) cells ↓ Phosphotyrosine STAT3, ↓ JAK levels.\ 500 nM Blaskovich et al., 2003 \[[@B172-cells-09-01451]\]
┴ STAT3-DNA binding; ┴ STAT3-mediated gene transcription
Induced cytotoxicity in glioblastoma multiforme cells,\ ┴ JAK/STAT pathway;\ Not specified Su et al., 2014 \[[@B173-cells-09-01451]\]
G2/M cell cycle arrest by downregulating cyclin B1 and cdc2 expression ↓ Phosphorylated STAT3 levels
Inhibitory activity in Sézary (Sz) syndrome and anaplastic large cell lymphoma ┴ JAK/STAT pathway;\ 30 μ[M]{.smallcaps} van Kester et al., 2008 \[[@B174-cells-09-01451]\]; Shi et al., 2006 \[[@B175-cells-09-01451]\]
↓ Phosphorylated STAT3 levels
Cucurbitacin Q *Picrorhiza kurrooa* Royle ex Benth. Antiproliferative effect in human non-small-cell lung carcinoma (A549) cells ↓ STAT3 but not JAK2 activation 3.7 µM Sun et al., 2005 \[[@B176-cells-09-01451]\]
Andrographolide *Andrographis paniculata* (Burm.f.) Nees. Enhanced anticancer activity of doxorubicin ┴ STAT3 signaling pathway;\ Not specified Zhou et al., 2010 \[[@B179-cells-09-01451]\]
┴ Constitutively actuated and IL-6-induced phosphorylation of STAT3 and subsequent nuclear translocation
Betulinic acid White-barked birch plants Inhibited proliferation and induced apoptosis in human multiple myeloma (U266) cells;\ ┴ STAT3 signaling,\ Not specified Pandey et al., 2010\
Cell cycle arrest at sub-G1 stage;\ ┴ STAT3-directed gene expression;\ \[[@B182-cells-09-01451]\]
downregulated the expression level of STAT3-regulated gene products ┴ Constitutive activation of STAT3, Src kinase, and JAK1/2
*γ*-Tocotrienol *Elaeis guineensis* Jacq. Anticancer effect in human hepatocellular carcinoma (HepG2, C3A, Hep3B, SNU-387, and PLC/PRF5) cells ┴ Both constitutive and inducible activation of STAT3;\ Not specified Rajendran et al., 2011 \[[@B185-cells-09-01451]\]; Banerjee et al., 2019 \[[@B186-cells-09-01451]\]
┴ Phosphorylation of JAK1, JAK2, and c-Src
Cryptotanshinone *Salvia miltiorrhiza* Bunge Induced cytotoxicity in human prostate cancer (DU145) cells ┴ Phosphorylation of STAT3 through an independent mechanism of JAK2 phosphorylation 7.59 µM Shin et al., 2009 \[[@B188-cells-09-01451]\]
Nimbolide *Azadirachta indica* A. Juss. Induced cytotoxicity and cell cycle arrest at G1--S stage in glioblastoma multiforme cells. It downregulates Bcl2 ┴ STAT3 pathway,\ Not specified Karkare et al., 2014 \[[@B191-cells-09-01451]\]; Zhang et al., 2016 \[[@B192-cells-09-01451]\]
┴ STAT3 phosphorylation
Celastrol *Tripterygium wilfordii* Hook.f. Induced cytotoxicity in human multiple myeloma (U266, RPMI 8226 and RPMI-8226-LR-5) cells ┴ Phosphorylation of STAT3 as well as STAT3-mediated IL-17 expression Not specified Kannaiyan et al., 2011 \[[@B193-cells-09-01451]\]
Inhibited differentiation and cell proliferation in multiple myeloma cells ┴ STAT3 phosphorylation;\ Not specified Banerjee et al., 2019 \[[@B186-cells-09-01451]\]
┴ STAT3-mediated IL-17 expression\
┴ T-helper 17 (Th17)
Ursolic acid *Mirabilis jalapa* L. Induced cytotoxicity in human prostate cancer (DU145 and LNCaP) cells ┴ Activation of constitutive and inducible STAT3; ↓ phosphorylation of Src and JAK2 80 and 47 µM Shanmugam et al., 2011b \[[@B198-cells-09-01451]\]
Inhibited tumor growth in prostate xenograft tumor in TRAMPmice in vivo ┴ JAK/STAT signaling;\ Not specified Shanmugam et al., 2011 \[[@B198-cells-09-01451]\]
┴ activation of STAT3
Brusatol *Brucea javanica* (L.) Merr. Induced cytotoxicity in head and neck squamous cell carcinoma ┴ STAT3 signaling pathway;\ Not specified Lee et al., 2019 \[[@B199-cells-09-01451]\]
┴ Activation of STAT3 as well as JAK1/2, and Src
Oridonin *Isodon rubescens* (Hemsl.) H. Hara Inhibited proliferation and induced apoptosis and cell cycle arrest at G2/M phase in breast (MCF-7), leukemia (K562), lung (A549), prostate (PC-3), liver (Bel7402), gastric (BGC823), and uterine cervix cancer (HeLa) cells\ ┴ STAT3 signaling pathway;\ 18.4, 4.3, 18.6, 15.2, 7.6, 13.7 μg/mL\ Bu et al., 2012 \[[@B202-cells-09-01451]\]; Chen et al., 2008 \[[@B203-cells-09-01451]\]; Chen et al., 2012 \[[@B204-cells-09-01451]\];\
The viability of BxPC-3 (human pancreatic cancer) cells is reduced on treatment ↓ IL-6, ↓ STAT3, and ↓ phospho-STAT3 expression level, ↓ p21, ↓ cyclin A, ↓ cyclin B1, ↓ cyclin D1, ↓ VEGF, and ↓ MMP-2 19.32 μg/mL Chen et al., 2014 \[[@B205-cells-09-01451]\]
Thymoquinone *Nigella sativa* L. Induced cytotoxicity in multiple myeloma (U266 and RPMI 8226) cells. ┴ STAT3 phosphorylation 15 and 15 µM Zhu et al., 2016 \[[@B208-cells-09-01451]\]; Li et al., 2010 \[[@B209-cells-09-01451]\]
Parthenolide *Tanacetum parthenium* (L.) Sch. Bip., *Tanacetum vulgare* L., *Centaurea ainetensis* Boiss., *Tanacetum larvatum* (Pant.) Hayek., and *Helianthus annuus* L. ┴ Proliferation, ↑ apoptosis, cell cycle arrest at G2/M phase in breast (MCF-7), skin (MDMB-231), melanoma (LCC9), malignant glioma (ABCB5+), epidermal tumorigenesis (A375), liver (1205Lu), gastric cancer (WM793) cells ┴ IL-6-induced STAT3 phosphorylation; ┴ JAK2 kinase activity\ 9.54, 10 µM, 600 nM, 12, 2.9, 6, 12 µM Cheng et al., 2011 \[[@B211-cells-09-01451]\]; Liu &Xie, 2010 \[[@B212-cells-09-01451]\]; Shanmugam et al., 2011 \[[@B213-cells-09-01451]\]
Inhibited JAK2 kinase activity in MDA-MB-231 cells ┴ IL-6-induced STAT3 phosphorylation Not specified Liu et al., 2018 \[[@B214-cells-09-01451]\]
Dihydroartemisinin *Artemisia annua* L. Inhibited tumor cell growth in human head and neck cancer (FaDu), liver cancer (Hep-G2), colon cancer (HCT-116), and tongue cancer (Cal-27) cells ┴ JAK2/STAT3 signaling activation; ↓ targeted proteins 160, 80, 25, 80 µM Jia et al., 2016 \[[@B217-cells-09-01451]\]; Wang et al., 2017 \[[@B218-cells-09-01451]\]
Alantolactone *Aucklandia costus* Falc., *Inula helenium* L., *Inula japonica* Thunb., and *Inula racemose* Hook. f. Antiproliferative effect in glioblastoma (U87), colon (HCT-8), leukemia (HL-60), liver (HepG2), lung cancer (A549) cells ┴ Both constitutive and inducible STAT3 activation at tyrosine 705;\ 135.27 µM, 5 µg/mL, 1.1, 40, 8.39 µM Pal et al., 2010 \[[@B220-cells-09-01451]\]; Khan et al., 2012 \[[@B221-cells-09-01451]\]; Chun et al., 2015 \[[@B222-cells-09-01451]\]
┴ STAT3 translocation to the nucleus; ┴ DNA-binding, and ┴ STAT3 target gene expression
*β*-Caryophyllene oxide *Cannabis sativa* L., *Humulus lupulus* L., *Origanum vulgare* L., *Psidium guajava* L., *Salvia Rosmarinus* Spenn., and *Syzygium aromaticum* (L.) Merr. & L. M. Perry Inhibited proliferation in multiple myeloma, prostate, and breast cancer cell lines ┴ STAT3 pathway;\ Not specified Kim et al., 2014 \[[@B224-cells-09-01451]\]
┴ STAT3 activation as well as JAK 1/2 and c-Src
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Various symbols (↑, ↓ and ┴) indicate increase, decrease and inhibition in the obtained variables, respectively.
cells-09-01451-t003_Table 3
######
Anticancer activities of alkaloids, saponins, steroids, lignan, and phytoalexin related to inhibition of the JAK/STAT signaling pathway.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Phytochemical Class Phytochemical Name Sources Anticancer Effect Mechanism of Action EC~50~/IC~50~ References
--------------------- ------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------ ----------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------- -------------------------------------------------
Alkaloids Capsaicin *Capsicum frutescens* L. Induction of cytotoxicity in human multiple myeloma (U266 and MM.1S) cells ┴ Constitutive and IL-6-induced activation of STAT3; ┴ JAK1 and c-Src activation 50 and 50 µM Bhutani et al., 2007 \[[@B139-cells-09-01451]\]
Evodiamine *Tetradium ruticarpum* (A. Juss.) T. G. Hartley Inhibited proliferation, and induced apoptosis;\ ↓ Constitutive and IL-6-induced activation of STAT3 tyrosine 705 (Tyr705);\ 113, 8.516, 5, 10, 27.15 µM Yang et al., 2013 \[[@B81-cells-09-01451]\]
Cell cycle arrest at G2/M phase in murine Lewis lung (LLC), hepatocellular (HepG2), leukemia (K562), gastric (SGC-7901), colon (COLO-205) cancer cells. ↓ JAK2, Src and ERK1/2;\
┴ STAT3--DNA binding activity
Indirubin *Angelica sinensis* (Oliv.) Diels Reduced cell viability in human prostate and breast cancer cells; induced apoptosis;\ ┴ STAT3 signaling;\ \~4 µM for each cell line Chen et al., 2018 \[[@B234-cells-09-01451]\]
Cell cycle arrest at G1/S or G2/M ┴ STAT3 phosphorylation\
Saponins β-Escin *Aesculus hippocastanum* L. Induction of cytotoxicity in human hepatocellular carcinoma (HepG2, PLC/PRF5, and HUH-7) cells ┴ Activation of STAT3 and induced expression of SHP-1;\ Not specified Tan et al., 2010 \[[@B238-cells-09-01451]\]
┴ Phosphorylation of JAK1, JAK2, and c-Src
Steroid Diosgenin *Trigonella foenum-graecum* L. Induced cytotoxicity in human hepatocellular carcinoma (C3A, HUH-7, and HepG2) cells ┴ Constitutive and inducible activation of STAT3 100, 100 and 50 µM Li et al., 2010 \[[@B244-cells-09-01451]\]
Ergosterol peroxide *Agaricus bisporus* Induced cytotoxicity in human multiple myeloma (U266) cells ┴ Phosphorylation of JAK2; ┴ phosphorylation;\ Not specified Rhee et al., 2012 \[[@B80-cells-09-01451]\]
┴ DNA binding activity;\
┴ Nuclear translocation of STAT3
Guggulsterone *Commiphora mukul* (Hook. ex Stocks) Engl. Induced cytotoxicity in human multiple myeloma (U266) cells ┴ Constitutive andIL-6-induced STAT3;\ 25 µM Ahn et al., 2008 \[[@B250-cells-09-01451]\]
┴ Phosphorylation of JAK2 and Src
Lignan Arctiin *Centaurea imperialis* Hausskn. ex Bornm., *Arctium lappa* L., *Forsythia viridissima* Lindl., *Himalaiella heteromalla* (D. Don) Raab-Straube., and *Trachelospermumas iaticum* (Siebold & Zucc.) Nakai Inhibited proliferation,\ ┴ STAT3 phosphorylation in tyrosine 705;\ Not specified Lee et al., 2019b \[[@B254-cells-09-01451]\]
cell cycle arrest at G2/M stage in human multiple myeloma cells ┴ Constitutive enactment of Src phosphorylation and JAKs 1/2; ↑ mRNA, ↑ protein levels of protein tyrosine phosphatase ε; ┴ STAT3 regulated gene products
Phytoalexin Brassinin *Brassica rapa* L. Induced apoptosis in lung cancer cells (A549) in an in vivo mouse model ┴ STAT3 activation;\ Not specified Lee et al., 2015 \[[@B257-cells-09-01451]\]
┴ Both constitutive and IL-6-inducible STAT3 activation; ┴ Phospho-STAT3, ┴ Ki-67 and ┴ CD31
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Various symbols (↑, ↓ and ┴) indicate increase, decrease and inhibition in the obtained variables, respectively.
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
The regulation of immune responses by the nervous system represents a spectrum of inhibitory and excitatory neural pathways. Inflammatory reflexes are fundamental neural circuits mediated by the vagus nerve and are important for immune response resolution, as they prevent excessive cytokine production and tissue damage^[@CR1]--[@CR5]^. Gateway reflexes regulate the status of the blood-brain barrier (BBB) to establish immune cell gateways and the induction of neural inflammation^[@CR6]--[@CR9]^. Activation of a gateway reflex stimulates the endothelium of specific blood vessels in the central nervous system (CNS) to secrete chemokines. This secretion allows CNS-autoreactive CD4+ T cells to breach the BBB and invade the CNS, where they cause inflammation^[@CR6],[@CR8],[@CR9]^. For example, sensory neural activation in the soleus muscles by gravity or electric stimulation induces chemokine expressions in the dorsal vessels of the fifth lumbar (L5) spinal cord via sympathetic nerve activation. During experimental autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis, chemokine up-regulation at the L5 vessels acts as a gateway for pathogenic CD4+ T cells specific for myelin-oligodendrocyte glycoprotein to invade the CNS from the L5 site^[@CR6]^. Overall, various neural stimulations create gateways at different blood vessels in the CNS. Pain and chronic stress induce distinct immune cell gateways at the ventral vessels of the L5 cord and specific vessels beside the third ventricle, dentate gyrus, and thalamus, respectively^[@CR8],[@CR9]^. Electric stimulations to muscles induce the formation of immune cell gateways at the dorsal vessels of the spinal cord where the dorsal root ganglion of the sensory neurons in the muscle is located^[@CR6]^. In general, these specific neural inputs lead to the release of neurotransmitters such as norepinephrine (NE) and/or ATP at specific vessels in the BBB, which in turn enhances the expression of chemokines in the endothelium to establish gateways through which immune cells can reach the CNS^[@CR1]--[@CR6],[@CR8],[@CR10],[@CR11]^. In addition, we have reported that stress establishes immune cell gateways at two brain vessels sites followed by the development of microinflammation^[@CR9]^. The resulting microinflammation then activates new neural pathways in a manner dependent on ATP and risks upper gastrointestinal and heart failure with sudden death. These results showed that the gateway reflex can affect the homeostasis of organs besides the brain.
To breach the BBB via a gateway reflex, the induction of massive chemokine expression by endothelial cells is critical. We identified the inflammation amplifier as the mechanism responsible. The inflammation amplifier involves co-activation of NF-κB and STAT3 in non-immune cells including endothelial cells, followed by the hyper-activation of NF-κB to express NF-κB target genes such as chemokines and IL-6^[@CR12]--[@CR14]^. Activation of the inflammation amplifier is critical for the development of mouse models of rheumatoid arthritis, multiple sclerosis, skin inflammation and allogeneic transplantation rejections^[@CR6],[@CR8],[@CR9],[@CR12]--[@CR23]^. It is known that NE and epinephrine (EPI) enhance NF-κB activation^[@CR6],[@CR24],[@CR25]^, which is a molecular basis that links gateway reflexes and the inflammation amplifier^[@CR10],[@CR11]^. The above examples all describe ways in which the BBB is breached. On the other hand, no mechanism dependent on specific neural activation that prevents the breaching has been identified.
A prominent feature of autoimmune posterior uveitis is chronic inflammation of the retina and choroid that commonly results in blindness. It is believed that autoreactive CD4+ T cells, particularly Th1 and Th17 cells, initiate the pathogenic process, and malfunction of the blood-retinal barrier (BRB) is considered a critical early phenomenon for the disease development^[@CR26]--[@CR28]^. Because retinal vessels express adrenergic receptors^[@CR29],[@CR30]^ and because NE is released in the retina from sympathetic neurons from outside the eyes^[@CR31]^ and retinal neurons themselves such as amacrine cells and horizontal cells produce both NE and EPI^[@CR32]--[@CR35]^, we hypothesized that a gateway reflex determines the BRB status by regulating NE/EPI release in the retina.
The main sensory stimulus for the retina is light, which directly affects three photoreceptor cell types in the retina: rods, cones, and melanopsin-expressing intrinsically photosensitive retinal ganglion cells (ipRGCs)^[@CR36]^. Thus, different light intensities should activate different retinal neural pathways. Consistent with this theory, photopic but not scotopic light has been shown to increase the activity of retinal tyrosine hydroxylase (TH), the enzyme that synthesizes catecholamines including NE and EPI in the retina^[@CR33],[@CR37]^.
In the present study, we employed two light intensities, photopic and mesopic light, and investigated BRB alterations in the initial phase of experimental autoimmune uveoretinitis (EAU) when retina-specific autoreactive CD4+ cells infiltrate the eye^[@CR38]--[@CR41]^. We found that exposure to photopic light significantly suppressed the inflammatory phenotype of the BRB endothelium and the recruitment of immune cells including pathogenic CD4+ T cells in the retina of EAU mice. Further, our mechanistic analysis suggests that photopic light in EAU mice down-regulates retinal α~1A~-adrenoceptor (α~1A~AR) expression, which in turn decreases chemokine and IL-6 expressions in the retina. These results suggest that photopic light can stimulate a suppressive type of gateway reflex, the light-gateway reflex, which has an anti-inflammatory effect in the retina. Our data also imply that targeting the regional adrenoceptor signal and/or retinal α~1A~AR might represent a novel preventive and/or therapeutic strategy for autoimmune diseases including those that affect organs separated by blood barriers such as the CNS and eyes.
Results {#Sec2}
=======
CD4+ T cells are the first cells to invade the eye on day 10 after immunization {#Sec3}
-------------------------------------------------------------------------------
To investigate the effect of light intensity on ocular inflammation during EAU development, we focused on the initial phase of the disease because our previous study on a multiple sclerosis model, EAE, demonstrated that neuro-immune interactions are important for triggering the breakdown of the BBB by pathogenic CD4+ T cells^[@CR6],[@CR8],[@CR9],[@CR42]^. To determine the time point of the initial immune cell infiltration during the early phase of EAU, we analyzed the number of immune cells in the eyes of mice housed under 2 lux mesopic light, which is within limits of normal light intensity values for mouse cages in our animal facility, in the light phase of the experiments and 0 lux scotopic condition in the dark phase. The number of CD4+ T cells significantly increased in the eyes from day 10 after immunization, followed by an increment of other immune cells such as CD8+ T cells and CD11b+ myeloid cells on day 14 (Fig. [1A](#Fig1){ref-type="fig"}), which is consistent with the fact that EAU is a CD4+ T cell-dependent disease^[@CR38]--[@CR41]^. Fundoscopic analysis revealed inflammatory responses in the retinal tissue and around the retinal vessels as early as day 12 after immunization (Fig. [1B](#Fig1){ref-type="fig"}).Figure 1Time course of immune cell infiltration into the eye during EAU. (**A**) Time-course of leukocyte entry into the eyes quantified by flow cytometry on days 8, 10, 12, 14, and 21 after immunization. Data are expressed as means + s.e.m. (*N* = 8--10 per time point). One representative analysis of at least two independent experiments is shown. Kruskal-Wallis non-parametric ANOVA followed by Dunn's multiple comparison test between each EAU time point and non-immunized control (C); \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001. (**B**) Representative fundus images obtained by the TEFI system from non-immunized control and EAU mice on the indicated days after immunization. In the magnified images of the inflamed areas, the severity of retinal damage is denoted by (1) yellow arrows, retinal vasculitis with mild cuffing, (2) black arrows, retinal vasculitis with moderate to severe cuffing, (3) white arrowheads, small inflammatory infiltrates in retinal tissue, (4) black arrowheads, linear retinal lesions, and (5) blue asterisk, severe inflammation in the optic disc with papilledema.
Protective effect of photopic light on EAU development {#Sec4}
------------------------------------------------------
We next examined the role of light intensity on CD4+ T cell infiltration in the retina. We applied two light intensities during the light phase: mesopic or photopic (230 lux) from day 10 after immunization (Supplementary Fig. [1](#MOESM1){ref-type="media"}), the day when CD4+ T cells infiltrate the eyes to a significant degree (Fig. [1A](#Fig1){ref-type="fig"}). All animals were housed under 0 lux condition in the dark phase. On day 14 after immunization, mice housed under photopic light showed a significantly reduced accumulation of CD4+, CD8+ and CD11b+ cells in the eyes and neural retina (Fig. [2A,B](#Fig2){ref-type="fig"}, and Supplementary Fig. [2](#MOESM1){ref-type="media"}). Consistently, photopic light from day 10 onward decreased EAU clinical scores compared with mesopic light (Fig. [2C](#Fig2){ref-type="fig"}). However, the absolute number of splenic naïve/activated CD4+ T cells, CD11b+ cells, and CD11c+ dendritic cells were comparable in EAU mice independent of the light conditions (Fig. [2D](#Fig2){ref-type="fig"}), suggesting local retinal and not systemic immune suppression under photopic light. These results suggest that retinal inflammation is suppressed by photopic light stimulation during EAU development.Figure 2Photopic light attenuated clinical disease severity and reduced accumulation of immune cells in eye and retina. (**A**,**B**) Absolute numbers of CD4^+^, CD8^+^, and CD11b^+^ cells in eyes (A) and neural retinas (B) of EAU mice housed under mesopic or photopic light. Determined by flow cytometry on day 14 after immunization. Data are expressed as means + s.e.m. (*N* = 10 each). \*\*\**P* \< 0.001 by two-tailed Student's t-test or Mann-Whitney U test. (**C**) Representative appearance of the ocular fundus recorded by the TEFI system (left) and TEFI clinical scores (right) on day 14 after immunization. Optic disc, retinal vessels and retinal tissue inflammatory changes were scored separately. The final inflammation grade represents the average of the scores of these three components. Data are expressed as means + s.e.m. (*N* = 20 each). \*\**P* \< 0.01 and \*\*\**P* \< 0.001 by two-tailed Student's t-test. (**D**) Absolute numbers of splenocytes determined by flow cytometry in EAU mice housed under mesopic or photopic light on day 14 after immunization. Naive CD4^+^ T cells (CD4^+^CD44^low^), activated CD4^+^ T cells (CD4^+^CD44^high^), CD11b^+^ (CD11b^high^CD11c^low^ macrophages and neutrophils) and CD11c^+^ (CD11c^high^MHC-II^high^ dendritic cells). Data are expressed as means + s.e.m. (*N* = 6 each). NS, not significant.
Photopic light suppresses inflammatory phenotype of BRB endothelium and retinal chemokine expression {#Sec5}
----------------------------------------------------------------------------------------------------
Because it has been reported that inflammation first develops at the optic disc in EAU mice^[@CR43]^, we examined the central retinal vessels located in the middle of the optic disc in detail. On day 14 after immunization, EAU mice housed under photopic light showed very few CD4+ T and CD11b+ cells around the central retinal vessels of the optic disc (Fig. [3A](#Fig3){ref-type="fig"}). Similarly, photopic light exposure attenuated the infiltration of CD4+ T cells in the central and peripheral retina and minimized retinal structural damage including the formation of retinal folds (Fig. [3B](#Fig3){ref-type="fig"}).Figure 3Photopic light prevented entry of immune cells in the optic disc and retina by suppressing the inflammatory status of BRB endothelium and local chemokine expression. (**A**) Inflammatory cell infiltrates around microcirculation vessels in the optic disc and central retina in EAU mice housed under mesopic or photopic light on day 14 after immunization. Green, CD4^+^ T and CD11b+ cells; red (type IV collagen; Col IV), endothelium and basal lamina; blue (Hoechst), nuclei. GCL, ganglion cell layer. (**B**) Immunohistochemical staining for the CD4^+^ cell population and retinal histopathology (arrows show retinal folds) in EAU mice housed under mesopic or photopic light. (**C**) Immunohistochemical staining for vascular marker Col IV, control rabbit IgG (rIgG), phospho-p65, and phospho-STAT3 performed on serial eye sections in EAU mice housed under mesopic or photopic light on day 14 after immunization. Red arrowheads, endothelial cells positive for p-p65 or p-STAT3 protein expression; black arrowheads, endothelial cells negative for p-p65 or p-STAT3 protein expression. The graph displays percentages of p-p65- and p-STAT3-positive vascular endothelial cells in the optic disc area in EAU mice (day 14 after immunization) exposed to mesopic or photopic light. Data are expressed as means + s.e.m. (*N* = 6 to 5 each). \**P* \< 0.05, \*\*\**P* \< 0.001 by two-tailed Student's t-test. (**D**) Chemokine and IL-6 mRNA expressions in neural retinas from intact control and EAU mice (day 14 after immunization) housed under mesopic or photopic light condition. Data are expressed as means + s.e.m. (*N* = 6 to 8 each). \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001 by two-tailed Student's t-test.
We reasoned how the immune cells infiltrating the retina passed through the BRB from the blood circulation. Because we previously discovered that gateway reflexes entail a chemokine hyper-induction mechanism (i.e. the inflammation amplifier) in endothelial cells to establish immune cell gateways at specific vessels in the BBB^[@CR6],[@CR8],[@CR9],[@CR12]--[@CR23]^, we hypothesized that a similar mechanism is critical for breaching the BRB during EAU development. Consistently, the phosphorylation of NF-κB and STAT3, which is required for activation of the inflammation amplifier, was suppressed in the retinal endothelial cells of EAU mice housed under photopic light (Fig. [3C](#Fig3){ref-type="fig"}). Moreover, we found that the expression of chemokines and IL-6 was significantly reduced in the neural retina of EAU mice under photopic light compared to mesopic light (Fig. [3D](#Fig3){ref-type="fig"}). These results strongly suggest that photopic light diminishes chemokine and cytokine expression in the retina and protects against BRB breakdown.
Photopic light increases eye norepinephrine and epinephrine in EAU mice {#Sec6}
-----------------------------------------------------------------------
Given our previous observations on the importance of NE for the gateway reflexes^[@CR1]--[@CR6],[@CR8]--[@CR11]^ and the presence of adrenergic receptors on retinal blood vessels^[@CR29],[@CR30]^, we investigated whether light-mediated sensory inputs to the retina interact with the regional noradrenergic system of the eye. First, we examined changes in NE and EPI levels in the eyes and serum after photopic and mesopic light treatment. Even though we did not observe any changes in serum levels (Fig. [4A](#Fig4){ref-type="fig"}), EAU mice housed under the photopic condition exhibited more NE and EPI levels in the eyes than EAU mice housed under the mesopic condition (Fig. [4B](#Fig4){ref-type="fig"}). No significant changes in serum NE and EPI levels also suggested that the intensity of photopic light used did not induce a systemic stress response in mice. In addition, the expression level of DBH, an enzyme necessary for NE and EPI synthesis, was significantly higher in the inner nuclear layer (INL) of EAU mice housed under photopic light compared to EAU mice housed under mesopic light (Fig. [4C](#Fig4){ref-type="fig"}). Other enzymes required for NE and EPI synthesis including TH, aromatic L-amino acid decarboxylase and phenylethanolamine N-methyltransferase were also expressed in the INL as described previously (data not shown; see^[@CR33],[@CR44],[@CR45]^). Although eyes are known to receive sympathetic innervation^[@CR31]^, unilateral removal of the superior cervical ganglion (SCG-X) did not significantly affect immune cell accumulation in the eye compared to that on the sham-operated side (Fig. [4D](#Fig4){ref-type="fig"} and Supplementary Fig. [1](#MOESM1){ref-type="media"}). It is known that retinal neurons such as amacrine cells and horizontal cells synthesize NE and EPI, and they are localized in the INL where higher DBH expression was observed^[@CR32]--[@CR35]^. These results suggest that a photopic light-induced local neural pathway in the retina leads to higher NE and EPI levels in the eyes of EAU mice at least in part via resident retinal cells in the INL.Figure 4High levels of norepinephrine and epinephrine in the eye from EAU mice under photopic condition. (**A**,**B**) Norepinephrine and epinephrine levels in serum (A) and eyes (B) of EAU mice (day 14 after immunization) housed under mesopic or photopic light. Data are expressed as means + s.e.m. (*N* = 5 each). \*\**P* \< 0.01, \*\**P* \< 0.01 by two-tailed Student's t-test. NS, not significant. (**C**) Immunofluorescence staining of DBH (green) protein expression and quantitative analysis of mean fluorescence intensities of DBH expression in the retinal layers of EAU mice (day 14 after immunization) housed under mesopic or photopic light. Blue, Hoechst staining of nuclei. Data are means + s.e.m. (*N* = 4 each). \**P* \< 0.05 by two-tailed Student's t-test. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; PS, photoreceptor segments. (**D**) Flow cytometric analysis of CD4^+^ T cells, CD8^+^ T cells, and CD11b^+^ cells in the neural retina (day 14 after immunization) after unilateral SCG-X in EAU mice exposed to mesopic or photopic light. The graph shows percent changes in cell numbers in the neural retina on the SCG-X side relative to the contralateral sham side in each mouse housed under mesopic or photopic conditions. Data are expressed as means + s.e.m. (*N* = 8--10 each). NS, not significant.
Photopic light down-regulates α~1~AR expression in the retina {#Sec7}
-------------------------------------------------------------
Increased NE and EPI signaling enhances NF-κB activation to promote local inflammation^[@CR6],[@CR24]^. On the other hand, it is reported that NE and EPI induce ligand-dependent adrenoceptor (AR) down-regulation and desensitization^[@CR46]--[@CR49]^. Therefore, we next examined AR expression in the retina in order to explain reduced eye inflammation despite higher catecholamine levels. Previously, we found a reduced accumulation of pathogenic CD4+ T cells at the L5 cord in EAE mice after treatment with prazosin, a selective α~1~AR antagonist^[@CR6]^ Another study also showed that blockade of α~1~AR suppressed the development of EAU in rats^[@CR50]^. Therefore, we focused on the α~1~AR family. Indeed, photopic light decreased total α~1~AR protein expression in endothelial cells of the central retinal vessels in EAU mice (Fig. [5A--C](#Fig5){ref-type="fig"}), which is the site where immune cell infiltration was observed (Fig. [3A](#Fig3){ref-type="fig"})^[@CR43]^. Among the three transcripts of α~1~AR subtypes (α~1A,~ α~1B~ and α~1D~), the mRNA level of Adra1a, which encodes α~1A~AR, was decreased in EAU mice housed under photopic light (Fig. [5D](#Fig5){ref-type="fig"}). Consistently, after the stimulation of endothelial cells with NE *in vitro*, the reduction of α~1A~AR protein levels was observed (Fig. [5E](#Fig5){ref-type="fig"}). Moreover, exposure to photopic light reduced the mean ocular blood flow compared with exposure to mesopic light, which may implicate a reduction of catecholamine signaling in retinal endothelial cells (Fig. [5F](#Fig5){ref-type="fig"}). Treatment with prazosin further reduced the ocular blood flow, suggesting that α~1~AR mediates ocular blood flow control and the effect of photopic light used in this study on α~1~AR down-regulation is not saturated. These results suggest that high levels of NE and EPI under photopic light in EAU mice down-regulate α~1A~AR protein levels in the central retinal vessels, which prevented the α~1A~AR signaling required for inflammation development and immune cell accumulation.Figure 5α~1A~AR expression in the retina of EAU mice exposed to mesopic or photopic light and its regulation by norepinephrine in endothelial cells. (**A**) Fluorescence microscopy images (serial sections) showing α~1~AR protein expression in vascularized areas (type IV collagen; Col IV) of the optic disc in EAU mice housed under mesopic or photopic light on day 14 after immunization. Nuclei were stained by Hoechst. (**B**) Mean fluorescence intensities of α~1~AR expressions in the optic disc vascular beds stained with Col IV in the sections shown in (**A**). Data are expressed as means + s.e.m. (*N* = 4 each). \*\*\**P* \< 0.001 by two-tailed Student's t-test. (**C**) Immunohistochemistry for vascular marker Col IV, α~1~AR, and control rabbit IgG (rIgG) performed on serial eye sections of EAU mice housed under mesopic or photopic light on day 14 after immunization. Red arrowheads, endothelial cells positive for α~1~AR expression; black arrowheads, endothelial cells negative for α~1~AR expression; white arrows, eye pigment normally present in C57BL/6 mice. The graph shows percentage of α~1~AR-positive vascular endothelial cells. Data are expressed as means + s.e.m. (N = 5 each). \**P* \< 0.05 by two-tailed Student's t-test. (**D**) mRNA levels for Adra1a in the neural retina of EAU mice (day 14 after immunization) housed under mesopic or photopic light. Data are expressed as means + s.e.m. (*N* = 6 each). \**P* \< 0.05 by two-tailed Student's t-test. (**E**) Effect of 48 h NE stimulation on total protein expression of α~1A~AR in BC1 cells determined by flow cytometry. Data are expressed as means + s.e.m. (*N* = 3 each). \**P* \< 0.05 by two-tailed Student's t-test. (**F**) Representative real and blood flow (BF) images obtained from wild-type BALB/c mice exposed to mesopic or photopic light with or without prazosin treatment. B6 mice were not suitable for this assay due to eye pigmentation. The graph shows alterations of mean ocular blood flow determined in a region of interest (depicted by the circles in the BF images). Data are presented as means + s.e.m. (*N* = 5 each). \*\**P* \< 0.01 by two-tailed Student's t-test.
Blockade of α~1~AR recapitulates protective effects of photopic light in EAU mice {#Sec8}
---------------------------------------------------------------------------------
To investigate whether the down-regulation of α~1~AR-mediated signaling can explain the beneficial effect of photopic light, we used both pharmacological and genetic approaches. Prazosin treatment in EAU mice housed under mesopic light (Supplementary Fig. [1](#MOESM1){ref-type="media"}) significantly reduced clinical scores and retinal CD4+, CD8+ and CD11b+ cell infiltrates (Fig. [6A,B](#Fig6){ref-type="fig"}), observations reminiscent of the anti-inflammatory effect of photopic light. Consistently, prazosin did not influence the number of immune cells or their activation in the spleen (Fig. [6C](#Fig6){ref-type="fig"}). Moreover, a small interfering RNA (siRNA)-mediated knockdown approach (Supplementary Fig. [1](#MOESM1){ref-type="media"}) that targeted Adra1a successfully knocked down Adra1a transcript (Fig. [6D](#Fig6){ref-type="fig"}) and significantly suppressed chemokine and IL-6 expression in the retina of EAU mice housed under mesopic light (Fig. [6E](#Fig6){ref-type="fig"}). We also examined siRNA-mediated knockdown of Rela (NF-κB p65), which is an essential transcription factor for the inflammation amplifier. As expected, intravitreal injection of Rela-targeting siRNA also significantly reduced retinal chemokine and IL-6 mRNA levels (Fig. [6D,E](#Fig6){ref-type="fig"}). All these data suggest that α~1~AR signaling and NF-κB activation in the retina have a pivotal role on EAU development and that the anti-inflammatory effect of photopic light during EAU development is at least in part mediated by α~1~AR down-regulation.Figure 6Protective effects of α~1~AR and α~1A~AR inhibition in EAU mice housed under mesopic light. (**A**) Effect of prazosin treatment on EAU clinical scores by the TEFI system on day 14 after immunization in EAU mice housed under mesopic light. Representative fundus images (left) and TEFI clinical scores (right). Optic disc, retinal vessels, and retinal tissue inflammatory changes were scored separately. The final inflammation grade represents the average of the scores of these three components. Data are presented as means + s.e.m. (*N* = 10 each). \*\**P* \< 0.01 by two-tailed Student's t-test. (**B**,**C**) Absolute numbers of CD4^+^, CD8^+^, and CD11b^+^ cells in neural retina (B) and spleen (C) determined by flow cytometry. Naive CD4^+^ T cells (CD4^+^CD44^low^), activated CD4^+^ T cells (CD4^+^CD44^high^), CD11b^+^ (CD11b^high^CD11c^low^ macrophages and neutrophils), and CD11c^+^ (CD11c^high^MHC-II^high^ dendritic cells). Data are presented as means + s.e.m. (*N* = 10 each). \* *P* \< 0.05 and \*\* *P* \< 0.01 by two-tailed Student's t-test. NS, not significant. (**D**) Knockdown efficiency of Adra1a- and Rela-targeting siRNA in the neural retina 4 days after intravitreal injection of the respective siRNAs administered on day 10 post-immunization in EAU mice housed under mesopic light. Data are expressed as means + s.e.m. (*N* = 4 each). \**P* \< 0.05, \*\*\**P* \< 0.001 by two-tailed Student's t-test. (**E**) Alterations of chemokine and IL-6 mRNA expressions in neural retinas of EAU mice housed under mesopic light 4 days after treatment with intravitreal injection of siRNA against Adra1a or Rela administered on day 10 post-immunization. Data are expressed as means + s.e.m. (*N* = 4 each). \**P* \< 0.05, \*\**P* \< 0.01, \*\*\**P* \< 0.001 by two-tailed Student's t-test.
Discussion {#Sec9}
==========
We have previously demonstrated using a mouse model of multiple sclerosis, EAE, that gateway reflexes induce NE-mediated alterations of endothelial cells at specific blood vessels in the CNS to regulate the BBB status. To date, four types of gateway reflexes have been demonstrated, and all cause CNS inflammation in the presence of CNS-autoreactive immune cells in blood^[@CR6],[@CR8],[@CR9],[@CR11]^. In the present study, we identified a suppressive type of gateway reflex by using a retinal inflammation model and showed that photopic light-mediated neural activation suppresses breaching of the BRB even in the presence of organ-specific T cells.
We examined the CD44 and CD62L levels of T cells that infiltrated the retina on day 14 after immunization (early phase). Over 90% of these T cells had the CD44^high^ CD62L^low^ activated phenotype, as expected (Supplementary Fig. [3](#MOESM1){ref-type="media"}). It is reported that splenocytes from unimmunized mice that have been adoptively transferred to IRBP-immunized mice around the peak of the disease migrate to the retina^[@CR51]^, suggesting that bystander cells including antigen-nonspecific naïve and activated T cell populations infiltrate the organ once the blood-retinal barrier is fully breached. The current study was designed to address the effect of light intensity on the initial entry phase of immune cells to the retina. It is possible that both naïve and activated T cells could be found at a later time point close to the peak of the disease, potentially affecting the disease development.
Our results suggest that photopic light treatment inhibits the accumulation of immune cells in the retina during EAU development by down-regulating α~1A~AR signaling in endothelial cells of the BRB. This conclusion is supported by four observations. First, exposure to photopic light from the onset of the pre-clinical phase of EAU resulted in a significant attenuation of immune cell infiltrates in the eyes. Second, photopic light in EAU mice suppressed the inflammatory phenotype of the BRB endothelium including the activation of NF-κB and STAT3 and the retinal gene expression of various chemokines and cytokines that attract and/or activate immune cells. Third, NE and EPI levels in the eyes were higher after photopic light treatment than under mesopic light treatment in EAU mice, which correlates with DBH expression in INL neurons and is associated with the down-regulation of retinal α~1A~AR expression. Finally, blockade of α~1~AR or retinal α~1A~AR knockdown in EAU mice housed under mesopic light showed the protective effects on EAU development seen in photopic light treatment. These results strongly suggest that the light-mediated α~1~AR-adrenergic pathway in the retina is involved in disruption of the BRB during EAU development (Fig. [7](#Fig7){ref-type="fig"}). This is in contrast to our previous studies showing that at the induction and relapse phases of the transfer EAE model, NE enhances NF-κB activity in the dorsal and ventral vessels of the fifth lumbar spinal cord, respectively, to promote disease progression^[@CR6],[@CR8]^. However, the transfer EAE model shows a transiently increasing clinical score followed by recovery from the symptoms^[@CR8]^. Accordingly, there is some negative regulation of inflammation during EAE development *in vivo*. We hypothesize that down-regulation of α~1~AR contributes to this negative regulation during the remission of EAE.Figure 7A summary diagram of suppression of retinal inflammation by photopic light. Exposure of EAU mice to photopic light causes down-regulation of retinal α~1A~AR expression by high epinephrine (EPI) and norepinephrine (NE) release from retinal neurons, which significantly inhibits activation of NF-κB and STAT3 and decreases chemokine and IL-6 expressions in the retina, leading to a reduced recruitment of immune cells including IRBP-specific pathogenic CD4+ T cells.
There are at least two potential sources of NE and EPI in the eye: sympathetic neurons^[@CR31]^ and retinal neurons such as amacrine and horizontal cells^[@CR32]--[@CR35]^. We found that eliminating sympathetic innervation of the eye by SCG-X had a minimal effect on EAU development. Because the number of accumulated immune cells in the retina of EAU mice on the SCG-X side did not significantly differ from that on the sham-operated side in either mesopic or photopic light condition, we suggest only a minor contribution by sympathetic neurons to NE and EPI content in the eye. However, EAU mice under photopic light showed high expressions of DBH enzyme in the INL, indicating that INL cells are at least in part a local producer of NE and EPI during EAU development under photopic condition. Consistent with this observation, it is known that retinal neurons located in the INL increase their activity in response to photopic light stimulation and that exposure to photopic light increases retinal EPI levels in both normal and SCG-X rats^[@CR33],[@CR52]^. Thus, we hypothesize that resident retinal cells in the INL including amacrine and horizontal cells, rather than SCG sympathetic neurons, are responsible for the photopic light-induced NE and EPI production in the eye and that this production is linked to EAU amelioration. Moreover, amacrine and horizontal cells form neurovascular units with capillaries in retinal vascular plexuses and are essential for maintaining homeostasis of the intraretinal vasculature^[@CR53]^, suggesting their ability to regulate the phenotype and remodeling of BRB endothelial cells.
Light detected in the retina activates distinct local neural retinal circuits depending on light intensity via photoreceptors. Photopic light intensity activates directly cone- as well as ipRGC-mediated retinal pathways, while single-photon scotopic light in darkness activates rod- and ipRGC-mediated retinal pathways^[@CR36],[@CR54]^. On the other hand, all three photoreceptor cell types can respond to mesopic light^[@CR36],[@CR55],[@CR56]^. Therefore, under photopic light condition, neural pathways originating most likely from cones and/or ipRGCs contribute to the DBH expression in noradrenergic amacrine and horizontal cells. Although it is known that amacrine and horizontal cells communicate directly with ipRGCs and cones, respectively, or with both photoreceptor systems through bipolar cells^[@CR57]^, the precise pathway responsible for the activation of noradrenergic neurons in the INL in response to photopic light remains elusive. Our findings here, which show the biological importance of these photopic light-activated noradrenergic retinal neurons as immune regulators to maintain retinal homeostasis, may accelerate the discovery of specific neural pathways that convert photopic light into a vascular effect that protects BRB integrity.
An earlier pharmacologic study in a rat EAU model demonstrated the amelioration of EAU after systemic treatment with the α~1~AR blocker prazosin^[@CR50]^. We confirmed this finding in EAU mice housed under mesopic condition, observing that prazosin treatment from day 10 post-immunization suppressed the local retinal but not systemic immune response in a manner similar to that found under photopic light condition. In addition, we observed that photopic light exposure reduced ocular blood flow. It is reported that after EAU induction, shear stress in retinal veins is reduced, and this reduction is negatively correlated with the rolling and sticking effectiveness of leukocytes^[@CR51]^. On the other hand, it is also known that fluid shear stress promotes lymphocyte migration across endothelial barriers in cytokine-activated human umbilical vascular endothelial cells^[@CR58]^. Since it is also suggested that retinal veins have higher shear stress than normal fenestrated endothelium^[@CR26],[@CR51]^, we hypothesize that the photopic light decreased ocular blood flow to a level that inefficiently promotes transendothelial migration.
We also investigated the effect of other adrenergic antagonists on EAU development, including the nonselective βAR blocker propranolol (20 mg/kg; daily i.p.), the β~1~AR blocker atenolol (25 mg/kg; daily i.p.), and the nonselective αAR blocker phentolamine (5 mg/kg; daily i.p.), but we did not observe any significant effects of these drugs on EAU outcome (data not shown), suggesting that α~1~AR plays a major role in inflammation development during EAU. Further, we showed the down-regulation of α~1A~AR expression after treatment with NE *in vitro*. Consistent with this finding, it is reported that the expressions of many G protein-coupled receptors, including αARs, are down-regulated after long-term exposure to an agonist^[@CR49],[@CR59],[@CR60]^. Therefore, we hypothesize that NE down-regulates α~1~AR molecules on the cell surface due to the internalization of NE-α~1~AR complexes and subsequent transcriptional regulation of α~1~AR. Consistent with this notion, it was also reported that NE treatment reduces receptor expression on the cell surface^[@CR49],[@CR59],[@CR60]^. Thus, we suggest that the higher level of NE in the eyes of EAU mice housed under photopic condition suppressed retinal α~1A~AR expression compared to EAU mice housed under mesopic condition, which resulted in a decrease of α~1~AR signaling and NF-κB activation^[@CR6],[@CR24]^.
The IRBP~1-20~-induced EAU in B6 mice that we used here is associated with chronic inflammation involving mainly the posterior segment of the eye, thereby serving as a good model of human posterior uveitis^[@CR61]^. According to reports from clinical studies, the majority of patients with uveitis are not photophobic. In one study, the percentage of cases associated with photophobia was 3.4% in pediatric uveitis^[@CR62]^ and 5.8% among autoimmune uveitis in both children and adult patients^[@CR27]^. In another study, photophobia was the presenting symptom in 1.8%, 3.2% and 6.8% cases of panuveitis, posterior uveitis and anterior uveitis, respectively^[@CR63]^, allowing the use of light stimulation for a potential treatment. Indeed, there are several papers that report the effectiveness of light treatment on patients suffering from eye inflammatory diseases^[@CR64]--[@CR67]^. However, the sensitivity to light can be different between mice and humans^[@CR68]^, and thus optimal light intensity or timing to produce the anti-inflammatory effect in humans has yet to be determined. In our study, a five-day exposure to photopic light of 230 lux starting from the preclinical stage of EAU significantly ameliorated retinal inflammation. Since the photopic light treatment used in this study modulated endothelial cells of the retinal vasculature to suppress the expression of proinflammatory factors including chemokines, we assume that photopic light treatment is effective particularly at the early stage of the disease rather than the late phase, when the BRB is completely breached. It is known that bright light of 5,000 lux activates a nociceptive pathway in trigeminal nerves^[@CR69]^, which might affect the disease status of EAU by pain-induced responses including the pain-gateway reflex^[@CR8]^. Whether various intensities or durations of photopic light exposure differently influence the outcome of EAU is an important point to address in the future.
In summary, we demonstrated a previously unknown suppressive role of light function in the regulation of immune cell accumulation across the BRB, which also underscores the role of the retinal α~1A~-noradrenergic pathway in local inflammation development in EAU mice (Fig. [7](#Fig7){ref-type="fig"}). This effect is reminiscent of gateway reflexes, in which specific regional neural activations establish immune cell gateways by changing the status of endothelial cells of the BBB in the CNS. We previously reported gravity-, electric-, pain- and stress-gateway reflexes^[@CR6],[@CR8],[@CR9],[@CR11]^. Here we propose that photopic light induces another gateway reflex, the light-gateway reflex. In contrast to previously discovered gateway reflexes, the light-gateway reflex provides the first evidence that gateway reflexes can have a suppressive effect on local inflammation. Overall, the light-gateway reflex, which is induced by an enhanced regional adrenoceptor signal followed by the down-regulation of retinal α~1A~AR, might represent a novel preventive and/or therapeutic strategy for autoimmune diseases, particularly those in which blood-tissue barrier breakdown is a key factor for the disease development.
Materials and Methods {#Sec10}
=====================
Animals {#Sec11}
-------
Female wild type C57BL/6 (B6) and BALB/c mice 6 weeks old with specific-pathogen free (SPF) status were purchased from SLC (Tokyo, Japan) and housed in an SPF animal facility at the Institute for Genetic Medicine (Hokkaido University, Sapporo, Japan). All mice were adapted to a 13 h/11 h light/dark photoperiod in the animal facility for seven days before experiments and had at libitum access to food and water. Experiments were performed in agreement with the guidelines of the Institutional Animal Care and Use Committees of Hokkaido University. Experimental protocols were approved by the Institutional Animal Care and Use Committees of Hokkaido University (No. 14-0083).
Induction of EAU and treatment with prazosin {#Sec12}
--------------------------------------------
EAU was induced in B6 mice by active immunization with 100 μL emulsion containing 200 µg human interphotoreceptor retinoid-binding protein (IRBP~1--20~) (amino acids 1--20 (GPTHLFQPSLVLDMAKVLLD); custom product by Sigma-Aldrich, Tokyo, Japan) and 1 mg Mycobacterium tuberculosis strain H37Ra (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) in complete Freund's adjuvant (CFA). The IRBP~1--20~/CFA emulsion was injected subcutaneously into the base of the tail (day 0), and mice were injected intravenously with 200 μL (0.2 μg) of Bordetella pertussis toxin (Sigma-Aldrich) dissolved in saline on days 0 and 2. All mice were housed under illumination of 2 lux from day 0 to day 10 post-immunization. Afterwards, mice were divided into two groups and were held either in mesopic (2 lux) or photopic (230 lux) condition until day 14 post-immunization. The light intensity was measured with a digital lux meter (LX-1000; Custom Japan Co., Osaka). The light level during the dark phase of the day/night cycle was 0 lux in all experiments. Treatment with prazosin was started from day 10 post-immunization in mice subjected to mesopic (2 lux) condition from day 0 until the end of the experiment (day 14). Prazosin was dissolved in 5% glucose solution acidified with acetic acid to 0.25% and administered intraperitoneally every 12 h for 5 days at a dose of 5 mg/kg body weight. Control mice received the vehicle solution.
Imaging of the eye fundus {#Sec13}
-------------------------
Clinical manifestation of EAU was scored using topical endoscopy fundus imaging (TEFI), a technique adapted from Paques *et al*.^[@CR70]^ with the following modifications. An endoscope was replaced with a 180 mm long borescope (MK Modular Mini-Scope, MK017-009-000-62; Olympus, Tokyo, Japan) with a 1.7 mm outer diameter, a direction of view of 0°, a field of view of 62° and a depth of field of 1 mm - ∞. The borescope was connected to an Olympus PEN digital camera (E-P5 with a 4/3-inch Live MOS image sensor and 16.05 million camera effective pixels) through adapters (OM Adapter MF-2 and AK-1M/SM-R; Olympus) without placing an additional magnifying lens between the borescope and the camera. A LED light source (ILD-3; Olympus) attached to the borescope through a flexible fiber optic cable was used to provide illumination of the fundus. The images were transferred to a computer, and the contrast and brightness were adjusted in Preview (Apple, Cupertino, CA, USA). Fundus images were scored according to the TEFI clinical grading system described previously^[@CR43]^.
For the fundoscopy procedure, mice were anesthetized with 2% isoflurane in an oxygen-enriched air mixture (with 0.4 L/min oxygen flow rate) inhaled by a mask, and pupils were dilated with topical tropicamide 0.4% (Sandol MY; Nitten Nippon Tenganyaku Kenkyujo, Nagoya, Japan).
Measurement of ocular blood flow {#Sec14}
--------------------------------
Ocular blood flow with particular emphasis on the retinal circulation was measured by a laser speckle flowmetry according to a published method^[@CR71]^ with modification. In brief, laser light was delivered to the retina through an optical fiber positioned on the eye in a manner to obtain signals predominantly from the retinal vessels, and signals of the choroidal vasculature were imaged minimally.
Because of the strong signal reduction due to tissue pigment in the eyes of B6 mice, albino mice (BALB/c) were used in all ocular blood flow experiments. For retina imaging, 780 nm laser light was outputted from the OmegaFlo-Lab LDF-C1 laser unit (Omegawave, Tokyo, Japan) and irradiated on the eye though a custom-made 140 μm optical fiber (Omegawave) inserted into a 25-gauge needle, as described previously^[@CR71]^. The ocular blood flow was imaged using a two-dimensional Laser Blood Flow Imager Omegazone OZ-1, and the data were analyzed with Laser Image Analyzer software (Omegawave).
Ocular blood flow was recorded under mesopic (2 lux) or photopic (230 lux) light after 1 h adaptation of the mice to the respective light condition. The effect of prazosin on ocular blood flow was investigated 1.5 h or 2.5 h after prazosin intraperitoneal injection.
Flow cytometry {#Sec15}
--------------
Mouse eyes were enucleated following transcardial perfusion with PBS. Eyes or retinas were minced and dispersed in 1 mg/mL collagenase D (Roche, Basel, Switzerland) in RPMI medium (10% FCS) for 1 h at 37 °C. Spleens were mechanically dissociated in RPMI medium. Cell surface antigens of eye or retinal cell suspensions were stained in the presence of 2.4G2 antibody and the following anti-mouse antibodies by BioLegend (San Diego, CA, USA): anti-CD11c (clone N418), anti-CD45.2 (clone 104), anti-CD3 (clone 145-2C11), anti-CD4 (clone RM4-5) and anti-CD8 (clone 53-6.7) conjugated with FITC, APC, Pacific Blue, PerCP, and PE-Cy7, respectively; and by eBioscience (San Diego, CA, USA): anti-CD19 (clone MB19-1) and anti-CD11b (clone M1/70) conjugated with FITC and PE, respectively. FITC-labeled anti-CD44 (clone IM7), PE-labeled anti-CD62L (clone MEL-14), APC-labeled CD90.2 (clone 53-2.1), APC-Cy7-labeled anti-CD45 (clone 30-F11) and Pacific blue anti-CD8 (clone 53-6.7) from BioLegend were also used. In order to analyze splenocyte populations, cells were stained in the presence of 2.4G2 antibody using the following anti-mouse antibodies: Pacific Blue-conjugated anti-CD3 (clone 145-2C11; BioLegend), PE-Cy7-conjugated anti-CD4 (clone RM4-5; eBioscience), FITC-conjugated anti-CD44 (clone IM7; eBioscience), APC-conjugated anti-CD11c (clone N418; BioLegend), APC-conjugated anti-MHC class II (clone M5/114.15.2; BioLegend), APC-conjugated anti-CD11c (clone N418; Biolegend), FITC-conjugated anti-CD11b (clone M1/70; Biolegend), Pacific Blue-conjugated anti-MHC class II (clone AF6-120.1; Biolegend), and PE-conjugated anti-CD45R/B220 (clone RA3-6B2; BD Biosciences Pharmingen, San Diego, CA, USA).
BC-1 cell suspension was incubated with anti-α~1A~AR rabbit monoclonal antibody (EPR9691(B); dilution 1:1000; Abcam) plus 2.4G2 antibody (dilution 1:200) in FACS buffer on ice for 1 h. The secondary antibody used was Alexa Fluor 488 donkey anti-rabbit IgG (H + L) (dilution 1:2000; Life Technologies) on ice for 1 h. Primary antibody unlabeled samples and positive controls (HepG2 cells) were used under the same conditions. In order to obtain a total protein expression of α~1A~AR, simultaneous fixation and permeabilization of the cells using BD Cytofix/Cytoperm™ Fixation/Permeablization Kit (BD Biosciences, San Jose, CA) were performed prior to staining with antibodies.
Flow cytometry data were acquired using CyAn Advanced Digital Processing (ADP) High-Performance Flow Cytometer (DakoCytomation, Tokyo, Japan) and analyzed by FlowJo software (Tree Star, Ashland, OR, USA).
RNA isolation and quantitative real-time PCR {#Sec16}
--------------------------------------------
On day 14 post-immunization, mice were sacrificed under mesopic (2 lux) or photopic (230 lux) light by overdose of pentobarbital and transcardially perfused with 0.01 M PBS. Collected retinas without retinal pigment epithelium were immediately processed for total RNA extraction using AllPrep DNA/RNA Mini Kit (Qiagen, Hilden, Germany). RT-qPCR was done using KAPA SYBR FAST ABI Prism qPCR Kit or KAPA PROBE FAST ABI prism qPCR kit (Kapa Biosystems, Wilmington, MA, USA) and the specific mouse primers and probes listed in Tables [1](#Tab1){ref-type="table"} and [2](#Tab2){ref-type="table"}. The relative mRNA expression of each gene was calculated using the standard curve method. All genes were normalized to the endogenous control gene *Hprt*.Table 1Sequences of primers and probes used for real-time qPCR.Gene nameSymbolPrimers (forward) 5′-3′Primers (reverse) 5′-3′TaqMan probe 5′-3′Adrenergic receptor, alpha 1aAdra1aGCTTCTTTCTGAAAATGCTTCTGAAGCCACCGAGAGGATCACTAAAFAM-CCCGCCAGCACAGGTGAACATTTCTAAGG-TAMRAC-C motif chemokine ligand 2Ccl2TGTCCCAAAGAAGCTGTAGTTTTTGGGTTCCGATCCAGGTTTTTAATGTAFAM-CCTTCTTGGGGTCAGCACAGACCTCTCT-TAMRAC-C motif chemokine ligand 3Ccl3TCCACGCCAATTCATCGTTGACGGTTTCTCTTAGTCAGGAAAATGAFAM-CCAGCAGCCTTTGCTCCCAGCCAG-TAMRAC-C motif chemokine ligand 20Ccl20AGGAAGAAAAGAAAATCTGTGTGCTCTTCTTGACTCTTAGGCTGAGGFAM-AGCCCTTTTCACCCAGTTCTGCTTTGGA-TAMRAChemokine (C-X-C motif) ligand 1Cxcl1CCACACTCAAGAATGGTCGCCGTTACTTGGGGACACCTTTTAGCFAM-TGCCTTGACCCTGAAGCTCCCTTGGTT-TAMRAChemokine (C-X-C motif) ligand 2Cxcl2GTCATAGCCACTCTCAAGGGCGTCAGTTAGCCTTGCCTTTGTTCFAM-AAAAGTTTGCCTTGACCCTGAAGCCCC-TAMRAChemokine (C-X-C motif) ligand 10Cxcl10ACGTGTTGAGATCATTGCCACGGCTAAACGCTTTCATTAAATTCTTGFAM-TCCGGATTCAGACATCTCTGCTCATCATTC-TAMRAHypoxanthine guanine phosphoribosyl transferaseHprtAGCCCCAAAATGGTTAAGGTTGCAAGGGCATATCCAACAACAAACFAM-ATCCAACAAAGTCTGGCCTGTATCCAACAC-TAMRAInterleukin 6IL-6AGGAGACTTCACAGAGGATACCAGCAAGTGCATCATCGTTGTTCAFAM-CCTGTCTATACCACTTCACAAGTCGGAGGC-TAMRATable 2Sequences of primers used for SYBR Green real-time qPCR.Gene nameSymbolPrimers (forward) 5′-3′Primers (reverse) 5′-3′Hypoxanthine guanine phosphoribosyl transferaseHprtGATTAGCGATGATGAACCAGGTTCCTCCCATCTCCTTCATGACARELA Proto-Oncogene, NF-Kappa-B p65 subunitRela (p65)GAGTTCCAGTACTTGCCAGACACTGAAAGGACTCTTCTTCATGATACTC
Immunohistochemistry {#Sec17}
--------------------
After transcardial perfusion, eyes were collected and frozen or paraffin blocks were made. The frozen blocks were cut into 10 μm sections using an adhesive film (Cryofilm type IIIC (UF16); SECTION-LAB, Hiroshima, Japan) and a cryostat microtome (Leica CM3050; Leica Microsystems) as described previously^[@CR6],[@CR8],[@CR72]^. Tissue sections were dehydrated in 100% ethanol, fixed in 4% paraformaldehyde for 15 min and transferred into PBS. The sections were blocked with 2% BSA in PBS (sections stained for cell surface markers) or Tris-buffered saline/0.1% Tween 20 (sections stained for phospho-c-Fos) for 30 min at room temperature and then incubated overnight at 4 °C with the following anti-mouse primary antibodies (diluted in 2% BSA in PBS): rabbit anti-collagen IV (dilution 1:400; Abcam, Cambridge, UK), rabbit anti-α~1~AR (dilution 1:100; Abcam), rabbit anti-dopamine beta hydroxylase (dilution 1:100; Abcam), biotinylated rat anti-CD4 (dilution 1:200; clone RM4-5; BioLegend), and biotinylated rat anti-CD11b (dilution 1:100; clone M1/70; BioLegend). In the case of using biotinylated antibody, the sections were additionally treated with Avidin/Biotin blocking kit (Vector Laboratories, Burlingame, CA, USA) before incubation with primary antibody. Subsequently, the sections were washed twice in PBS and incubated with Hoechst 33342 fluorescent stain (at 1:10000 dilution in PBS; Life Technologies) and secondary antibodies, Alexa Fluor 647 goat anti-rabbit IgG (H + L) or Alexa Fluor 546 streptavidin conjugate (at 1:500 dilution in PBS; Life Technologies) for 1 h at room temperature.
Paraffin blocks were sectioned to 5-μm thickness. After deparaffinization of the sections, antigen retrieval was performed in 10 mM sodium citrate buffer (pH 6) or 10 mM Tris/1 mM EDTA buffer (pH 8.5). Endogenous peroxidase activity was blocked by 3% H~2~O~2~ solution for 10 min, followed by incubation of the sections in blocking solution (VECTASTAIN Elite ABC kit, Vector Laboratories) for 1 h at room temperature. Next, sections were incubated overnight at 4 °C with the following rabbit primary antibodies: anti-CD4 (dilution 1:200; clone RM4-5; BioLegend), anti-phospho-(Tyr705)-STAT3 (dilution 1:200, Cell Signaling Technology), anti-phospho-(Ser276)-p65 (dilution 1:500, Sigma-Aldrich), anti-collagen IV (dilution 1:500; Abcam), and anti-α~1~AR (dilution 1:100; Abcam). Rabbit IgG (dilution 1:200; Cell Signaling Technology) was used as isotype control. VECTASTAIN Elite ABC kit (Vector Laboratories) and the manufacturer's protocol were applied for incubation with secondary antibody and streptavidin-HRP reagent. Signals were visualized using DAB substrate (ImmPACT DAB, Vector Laboratories). Hematoxylin (Wako chemicals, Richmond, VA, USA) was used for nuclear counter-staining.
Sections were examined under a BZ-9000 BioRevo fluorescence microscope (Keyence, Osaka, Japan). The acquired microscope images were analyzed using ImageJ software (NIH, Bethesda, MD, USA), and the image contrast and brightness were adjusted using BZ II Analyzer software (Keyence).
Cell culture and stimulation experiments {#Sec18}
----------------------------------------
A type 1 collagen positive endothelial cell line (BC-1)^[@CR73]^ was kindly provided by Dr. M. Miyasaka (Osaka University). BC-1 cells were seeded into 24-well plates at 50,000 cells/well and treated with L-(−)-norepinephrine-(+)-bitartrate (10, 100 and 300 μM; EMD Millipore, Billerica, MA) in Dulbecco's Modified Eagle Medium (supplemented with 10% FBS; Gibco by Life Technologies) for 48 h after 2 h serum starvation. Cells were then dissociated with 0.05% Trypsin/0.53 mM EDTA (Nacalai Tesque, Kyoto, Japan), pelleted and suspended in RP10 medium, and immediately used for flow cytometry analysis.
Measurement of catecholamine concentrations {#Sec19}
-------------------------------------------
Mice were injected with pentobarbital under mesopic or photopic condition on day 14 post-immunization between 8--11 a.m. Blood was collected by cardiac puncture in order to obtain serum samples. Afterwards, mice were perfused with 0.01 M PBS, and the eyes were enucleated, snap-frozen on dry ice, and stored at −80 °C until homogenates were prepared. Eyes were homogenized in RIPA lysis buffer (1 mL per 100 mg of tissue; Cell Signaling Technology) containing Protease Inhibitor Cocktail, Phosphatase Inhibitor Cocktail 2 and 3 (1 mL per 100 mL of lysis buffer; Sigma-Aldrich) and sodium metabisulfite (Sigma-Aldrich) at a final concentration of 4 mM to prevent the degradation of catecholamine. The homogenization was performed using a Polytron PT1600E homogenizer (Kinematica, Luzern, Switzerland) at a speed of 30,000 rpm for 1 min. Additionally, eye homogenates were sonicated in a Biodisruptor sonication device (Diagenode, Denville, NJ) for 5 min (30 s on/off per minute cycle) at 4 °C.
The lysates were centrifuged at 15,000 × g for 10 min at 4 °C, and the supernatant was used to determine catecholamine concentrations. Commercially available enzyme-linked immunosorbent assay (ELISA), 2-CAT (A-N) Research ELISA kit (\#BA E-5400; LDN, Nordhorn, Germany) was used for the quantification of NE and EPI levels in serum and eye lysates following the manufacturer's instructions. The limit of sensitivity for serum and eye lysate NE was 20 pg/mL and 10 pg/mL, respectively, and for serum and eye EPI it was 50 pg/mL and 25 pg/mL, respectively. The eye lysate catecholamine concentrations were normalized to the milligram of tissue protein content, as determined using a Bradford protein assay (Bio-Rad, Hercules, CA).
Intravitreal injection of siRNA {#Sec20}
-------------------------------
On day 10 after immunization, mice were anesthetized with 2% isoflurane in an oxygen-enriched air mixture (with 0.4 L/min oxygen flow rate) inhaled by mask and kept under anesthesia during the procedure. Mice received an intravitreal injection of Accell SMARTpool Adra1a or Accell SMARTpool Rela/p65 into the right eye and non-targeting Accell siRNA (GE Dharmacon, Lafayette, CO) into the left eye at a dose of 1 μL (100 μM siRNA dissolved in sterile, RNase free 0.01 M PBS) per eye using a graduated pulled glass pipette. A 30-gauge needle was used to poke a hole in the sclera surface at the level of pars plana. A pipette was carefully inserted into the vitreous space through the poked hole, and the siRNA solution was delivered. To avoid fluid reflux, the pipette was kept in place for 20 s and then gently withdrawn. All mice that received an intravitreal injection of siRNA were housed under mesopic light from day 0 to day 14 after immunization.
Superior cervical ganglionectomy (SCG-X) {#Sec21}
----------------------------------------
Female B6 mice underwent unilateral removal of SCG under isoflurane anesthesia using a 2.5% oxygen mixture. The surgery was performed under a dissection microscope. The animal was secured in supine position on a heating pad, the ventral neck region was shaved and disinfected, and a middle vertical incision was made. The carotid artery was identified and displaced laterally. The SCG was exposed behind the carotid bifurcation and excised by transecting the internal and external carotid nerves and the cervical sympathetic preganglionic nerves as described previously^[@CR69]^. In the contralateral, sham-operated side, the SCG was exposed without excision. The successful removal of the SCG was confirmed by blepharoptosis on the ganglionectomized side. EAU was induced 4 days after surgery, and mice were housed under mesopic condition only or mesopic condition followed by photopic condition from day 10 post-immunization until being sacrificed on day 14 post-immunization.
Statistical analyses {#Sec22}
--------------------
Statistical analyses were performed using Statistica software (StatSoft Inc, Tulsa, OK) and GraphPad software (GraphPad.com). According to the distribution of the data and number of groups, Student's t-test, Mann-Whitney U test or ANOVA tests were used as indicated in the figure legends. P values less than 0.05 were considered statistically significant.
Supplementary information
=========================
{#Sec23}
FigS1-S3
**Publisher's note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
=================================
**Supplementary information** accompanies this paper at 10.1038/s41598-019-38895-y.
We appreciate the excellent technical assistances provided by Ms. Ezawa, and Ms. Nakayama, and thank Ms. Fukumoto for her excellent assistance, Dr. J.-J. Jiang, Dr. S. Sakoda and Dr. T. Atsumi for technical advice and discussion, and Dr. P. Karagiannis (CiRA, Kyoto University, Kyoto, Japan) for carefully reading the manuscript. This work was supported by KAKENHI (D.K., Y.A., and M.M.), the Joint Usage/Research Center Institute for Genetic Medicine, Hokkaido University (M.M.), JSPS Postdoctoral Fellowship for Foreign Researchers (A.S.), Takeda Science Foundation (M.M.), Institute for Fermentation Osaka (M.M.), Mitsubishi Foundation (M.M.), Mochida Memorial Foundation for Medical and Pharmaceutical Research (D.K.), Suzuken Memorial Foundation (Y.A. and D.K.), Japan Prize Foundation (Y.A.), Ono Medical Research Foundation (Y.A.), Kanzawa Medical Research Foundation (Y.A.), Kishimoto Foundation (Y.A.), Nagao Takeshi Research Foundation (Y.A.), Japan Multiple Sclerosis Society (Y.A.), Kanae Foundation (Y.A.), Tokyo Medical Research Foundation (M.M. and Y.A.), GACR 18-11795Y (A.S.), PRIMUS/17/MED/7 (A.S.), Progres Q35 (A.S.), Uehara Memorial Foundation (M.M. and Y.A.), Japan Brain Foundation (Y.A.), Kao Foundation (Y.A.), Suhara Memorial Foundation (D.K.), Yasuda Memorial Foundation (D.K.), and Novartis Pharma Research Grants (D.K.).
M.M. and A.S. conceived the project. M.M. supervised all experiments. M.M., A.S. and D.K. wrote the manuscript. A.S. performed most experiments. D.K., T.O., M.O. and Y.A. contributed to specific experiments and analyses.
Competing Interests {#FPar1}
===================
The authors declare no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .370}
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#ss1}
============
Precipitation of second phase particles such as Ti, Nb, and V carbonitrides can have a very positive effect on mechanical properties of microalloyed steels. Low temperature precipitation (in ferrite or in α/γ interphase) leads to a distribution of fine particles, that have a strong strengthening effect ([Porter & Easterling 1992](#b20){ref-type="ref"}). Moreover carbides and carbonitrides stable at high temperature have a pinning effect on austenite grain boundaries (grain size control) leading to optimal mechanical properties ([Gladman 2002](#b11){ref-type="ref"}). For both strengthening and grain size control, it is essential to quantify size distribution and precipitated volume fraction.
Due to the small size of second phase particles (typically a few tenths of nanometers), transmission electron microscopy (TEM), and particularly conventional TEM (CTEM), appears as one of the most widely used techniques for the characterization and size measurement of precipitates in metal alloys. Among a very large literature concerning this topic, we will concentrate here on methods applied to the case of precipitation within ferrous alloys, which remains of considerable practical and industrial interest, as attested by the following non‐exhaustive list of recent examples: precipitation of complex carbonitrides in microalloyed steels ([Craven *et al.* 2000](#b9){ref-type="ref"}; [Saikaly *et al.* 2001](#b22){ref-type="ref"}; [Mishra *et al.* 2002](#b16){ref-type="ref"}), precipitation in a model austenitic steel ([Rainforth *et al.* 2002](#b21){ref-type="ref"}), precipitation of AlN in low carbon steels ([Sennour & Esnouf 2003](#b23){ref-type="ref"}), carbide precipitation in HSLA steels ([Hong *et al.* 2003](#b13){ref-type="ref"}), coherent carbonitride precipitation in commercial microalloyed steels ([Morales *et al.* 2003](#b17){ref-type="ref"}), carbide precipitation at grain‐boundaries ([Kaneko *et al.* 2004](#b14){ref-type="ref"}), Nb‐based precipitates in ferrite ([Beres *et al.* 2004](#b5){ref-type="ref"}; [Perrard *et al.* 2006](#b19){ref-type="ref"}; [Courtois *et al.* 2006](#b8){ref-type="ref"}).
CTEM on thin foils is well adapted for the observation of carbides (such as VC or NbC) in ferritic steels because the precipitate exhibit the Baker and Nutting (BN) orientation relationship (OR) with the ferritic matrix ([Baker & Nutting 1959](#b4){ref-type="ref"}). In that case, the matrix is simply orientated in a specific crystallographic direction (e.g. 〈100〉 ferrite zone axes), and usual dark‐field imaging allows to reveal each of the three variants precipitated in the observed area (see for example [Perrard *et al.* (2006)](#b19){ref-type="ref"} in the case of NbC precipitation in low carbon steels). As it will be shown in the present study, CTEM imaging is much less efficient when precipitates do not exhibit any OR with the surrounding matrix.
Recently, new TEM approaches have emerged for the quantitative analysis of precipitation in metal alloys: energy‐filtered TEM (EFTEM) and high angle annular dark field (HAADF). These techniques, when applicable, present the advantage to image correctly the precipitates even when no OR exists with respect to the matrix.
EFTEM allows second phases to be imaged in thin foils owing to their difference in chemistry with the matrix. Electrons of the primary beam that have experienced an inelastic scattering caused by a given atomic specie only present in the precipitates can be used to produce an elementary map imaging the particles of interest ([Hofer *et al.* 1996](#b12){ref-type="ref"}; [Craven *et al.* 2000](#b9){ref-type="ref"}; [Rainforth *et al.* 2002](#b21){ref-type="ref"}; [Beres *et al.* 2004](#b5){ref-type="ref"}; [Courtois *et al.* 2006](#b8){ref-type="ref"}; [Mackenzie *et al.* 2006](#b15){ref-type="ref"}) with a very good spatial resolution (i.e. near 1 nm [Mackenzie *et al.* (2006)](#b15){ref-type="ref"}). HAADF in TEM is another imaging method allowing nanometric particles or precipitates to be easily imaged within thin foils and/or when collected on a supporting film (such as extraction replicas in the case of precipitates). In this method, the intensity is roughly proportional to (where *Z* is the atomic number), and excellent contrast is obtained in the case of nanometric particles deposited on holey carbon grids ([Treacy & Rice 1989](#b24){ref-type="ref"}), or carbon or alumina extraction replicas, respectively ([Wilson & Craven 2003](#b26){ref-type="ref"}; [Courtois *et al.* 2006](#b8){ref-type="ref"}). From a practical point of view, extraction replicas represent a very attractive method, when applicable to precipitation problems (see [Bradley (1965)](#b7){ref-type="ref"} and references therein for a general presentation of the replica techniques). After a slight pre‐etching, the surface of the material is covered by a nanometric film, and the matrix is further selectively dissolved, allowing precipitates to be retained and observed on the deposited film. In the case of steels, this preparation offers two great advantages: (i) to get rid of the undesirable magnetic influence of the iron matrix during TEM work; and (ii) to allow very large numbers of particles to be measured on a single replica.
Regarding SEM, [Varano *et al.* (2005)](#b25){ref-type="ref"} showed that it is possible to detect niobium carbides using secondary electrons (SE) on replicas in a scanning electron microscope (SEM), but the contrast obtained was not good enough for size distribution measurement. The recent availability of detectors located below the sample allows SEM to produce scanning TEM (STEM) images on thin specimens, which may be of great interest in various situations ([Bogner *et al.* 2007](#b6){ref-type="ref"}). However, to the authors knowledge, STEM in SEM has not been used for the characterization of precipitates in metallic alloys.
In this paper, HAADF in TEM and STEM in SEM imaging techniques will be applied to the particular case of the characterization of carbides size distribution evolution during austenitization of FeCV and FeCVNb model alloys. These techniques will then be compared to more 'conventional' ones, with the aim to validate the STEM approach on replicas in SEM, by comparing it to HAADF in TEM.
Materials and treatments {#ss2}
========================
Model alloys FeCV and FeCVNb {#ss2-1}
----------------------------
Two high purity model alloys, FeCV and FeCVNb, were used for experimental investigation of the size and composition of vanadium and niobium carbides after different austenitization heat treatments. Compositions of these alloys are given in [Table 1](#t1){ref-type="table"}. Other elements are present in individual quantities less than 5 ppm. Hence, only VC and (V,Nb)C carbides are expected to precipitate.
######
Compositions of the FeCV and FeCVNb model alloys (in weight %).
C V Nb S O N
-------- ------- ------- ------- ---------- ---------- ----------
FeCV 0.480 0.200 0 \<0.0005 \<0.0005 \<0.0005
FeCVNb 0.470 0.190 0.045 \<0.0005 \<0.0005 \<0.0005
Plates of these model alloys were manufactured by the PECM group at the ENSM., Saint Etienne, France. The final thickness of the plates was about 2 mm. The last stage of the elaboration process was a solutionizing heat treatment at 1000°C for 30 min followed by a water quench to room temperature ([Fig. 1](#f1){ref-type="fig"}). Samples of 70 mm × 5 mm × 1 mm were taken from the core region of the plates to avoid the decarburized zone at their periphery. All different heat treatments described later were performed on these samples.
{#f1}
Heat treatments {#ss2-2}
---------------
A two‐step heat treatment was designed to precipitate almost all the vanadium and the niobium of the alloy with a high volume fraction of precipitates, and to obtain precipitate sizes as large as possible. In the first step, they were further heated at 700°C for 10 h in vacuum (quartz encapsulation). In the second step, samples were then heated at 800°C for 10 days (to allow coarsening in austenite), and finally slowly cooled down to room temperature ([Fig. 1](#f1){ref-type="fig"}). The pertinence of this two‐step procedure is detailed elsewhere ([Epicier *et al.* 2008](#b10){ref-type="ref"}). This state (called fully precipitated state \[FPS\]), was chosen as the starting point for a series of isothermal heat treatments in austenite. In order to validate the STEM approach on replicas on different precipitation states, FPS samples were heat treated at different temperatures between 870 and 1200°C, and for different holding times from 2 min to 9 days. These reversion experiments were performed either in a molten salt bath or in a furnace under quartz encapsulation. Samples were water quenched to room temperature. The obtained states will be hereafter called as partially dissolved states (PDS) ([Fig. 1](#f1){ref-type="fig"}).
Experimental techniques and procedures {#ss3}
======================================
Microscopy work {#ss3-1}
---------------
SEM and CTEM techniques are widely known and will not be detailed in this paper. It can just be noted that the introduction of field emission gun (FEG) in SEM allows a resolution of about 2 nm in the secondary electrons (SE) imaging mode.
SEM was performed with a FEG SEM (XL30 from FEI, Hillsboro, OR, USA) operating at 30 kV. SE imaging was done using an Everhart‐Thornley detector, whereas STEM imaging in this SEM microscope was done using a STEM detector. In order to check the nature of precipitates (i.e. presence of V or Nb), chemical analysis has been conducted using an analysing system from EDAX (Mahwah, NJ, USA).
CTEM imaging was performed using a 2010F microscope from JEOL (Tokyo, Japan). Precipitates were examined using conventional imaging techniques, electron diffraction and EDX with an OXFORD‐Inca analyser (OXFORD, Abington, UK) used to confirm the chemistry of precipitates. Further HAADF imaging was performed using a JEOL annular detector. This detector is placed well under the projecting lenses below the sample, and collects only incoherent electrons scattered at angles much higher than Bragg\'s angles. According to Rutherford scattering, the intensity at each probe position is given by: where is the number of atoms with atomic numbers in the irradiated volume, *C* a constant of proportionality (taken equal to unity in the following for sake of simplicity).
Because of the difference in atomic number between niobium or vanadium and carbon (respectively 41 and 23 against 6), this technique is very efficient to image vanadium or niobium carbides on carbon extraction replicas with a good contrast. Furthermore, it is particularly adapted to obtain easily size distributions because of the large number of particles that can be observed in a reasonably short time, and easily analysed with semi‐automatic image treatments, owing to their high contrast.
As permitted by the XL30 SEM, STEM imaging was also tested by using a solid‐state detector placed below the sample in the microscope ([Fig. 2](#f2){ref-type="fig"}). Its dipolar constitution allows 'bright‐field' or 'dark‐field' images to be acquired by choosing one of the two diodes (i.e. by collecting either preferentially transmitted or deflected electrons over an angular range of several degrees). In this work, only 'dark‐field' imaging has been performed. The intensity of these images depends in a complex way on the chemical composition, because deflected electrons have experienced different interaction mechanisms (elastic or inelastic scattering). Hence, it is difficult to establish a direct link between the intensity of the particles and their chemical composition as permitted in HAADF on the basis of [Eq. 1](#m1){ref-type="disp-formula"}. According to [Fig. 2](#f2){ref-type="fig"}, the collection conditions cannot be strictly controlled as in the case of an annular detector. Nevertheless, detection angles estimated in our case are ranging from 10 to a few dozens degrees in a limited angular sector. As it will be observed in the following, resulting images exhibit satisfying contrast.
{#f2}
Sample preparation {#ss3-2}
------------------
For conventional SEM observations of bulk samples, their surface was polished to 1/4 mm diamond paste, and then chemical etched with an 0.4% nital solution. Transmission electron microscopy studies were carried out on both thin foils and carbon extraction replicas. Thin foils were prepared by mechanical polishing to a final thickness of 20--30 μm (in order to minimise the undesirable magnetic effects during observation), followed by an argon ion beam thinning process in a polishing precision system (PIPS) from GATAN (Pleasantown, CA, USA). Carbon extraction replicas were prepared following a standard procedure: samples were first polished and etched in 0.4% nital. The etched surface was then coated with a thin film of carbon (≈20 nm) and then scored into 2‐mm squares. Finally, samples were etched in diluted nital to remove the thin film from the substrate. The extracted carbon replicas were then rinsed in both ethanol and methanol, placed on copper grids and dried.
Comparative study of CTEM, SE in SEM, HAADF in TEM and STEM in SEM {#ss4}
==================================================================
Summary of results {#ss4-1}
------------------
As discussed above, different combinations of imaging techniques and sample nature are possible, and were tried as summarized in [Table 2](#t2){ref-type="table"}.
######
Summary of performed experiments for testing (T) or measurements (M). STEM imaging on bulk samples in SEM (shaded box) is not possible for obvious reasons. (\*) In this table thin foils are considered as "bulk" as precipitates are embedded in iron matrix.
------------------------------------------
------------------------------------------
In order to compare these different approaches, the FPS of the FeCV alloy has been investigated using SE in SEM on bulk samples, CTEM on thin foils, HAADF in TEM on carbon replicas and STEM in SEM on carbon replicas. As pointed out in [Table 2](#t2){ref-type="table"}, the remaining possibilities were only briefly tested for reasons developed in the next sub‐sections. [Figure 3](#f3){ref-type="fig"} shows the particles size distributions obtained in each case, the number of measured particles, and the calculated mean radius are also indicated. Representative micrographs obtained with these four approaches will be reported in the following.
{#f3}
The shape of the size distribution and the value of the mean radius obtained with the four different techniques are very similar ([Fig. 3](#f3){ref-type="fig"}), and thus, they all seem to be adapted for size measurements of second phase particles such and drawbacks in each case in terms of rapidity, efficiency and facility.
SEM in SE mode {#ss4-2}
--------------
[Figure 4](#f4){ref-type="fig"} shows three micrographs of different precipitation states obtained on bulk specimen with the SEM in SE mode. Although precipitates are visible, their contrast is not sufficient for automatic image analysis. Furthermore, particles seem to be covered by a remaining layer of iron probably, leading to difficulties in clearly identifying their contour. Although particle size distributions can readily be obtained with this technique, for example see [Acevedo‐Reyes *et al.* (2005)](#b3){ref-type="ref"}, the number of particles that can be unambiguously and accurately measured from such images in a reasonable time appears to be too low in order to obtain statistically representative results (about a hundred during a half‐day observation).
{#f4}
Conventional TEM on thin foils {#ss4-3}
------------------------------
In the particular case of the present investigation, it must be recalled that carbides have lost any orientation relationship with the matrix following the ferrite to austenite phase transformation during the precipitation heat treatment. Hence, it is not possible to use the conventional dark field (DF) mode to observe easily numerous particles in the same precipitation variant.
However, as shown in [Fig. 5](#f5){ref-type="fig"}, particles are easily visible in bright field images, but their contrast is not sufficient for an automatic image analysis. Precipitates observed in bright field can look either bright or dark, depending on diffraction conditions. Combining the fact that (i) particles do not have any OR with the matrix, (ii) martensite is magnetic, and (iii) the area analysed in a thin foil is quite small, the number of particles that can be observed in a reasonable time is very low (2 or 3 dozens during a half‐day observation). Consequently, TEM on thin foils is not adapted for determining the size distribution in the present case.
{#f5}
HAADF‐STEM on thin foils in TEM {#ss4-4}
-------------------------------
As reported in [Table 2](#t2){ref-type="table"}, HAADF imaging on thin foils in a TEM was only performed for testing purposes (see [Fig. 6](#f6){ref-type="fig"}). Contrary to the case of supported carbides on extraction replicas, vanadium or niobium carbides exhibit poor contrasts in HAADF images from iron thin foils. [Equation 1](#m1){ref-type="disp-formula"} can be used to predict the contrast of VC and NbC particles, considered as perfect spheres, embedded in a thin foil of iron with a given thickness. Neglecting beam spreading (i.e. assuming a constant probe size throughout the sample), and taking into account atomic densities in numbers of atoms per unit volume,[^1^](#fn1){ref-type="fn"} elementary HAADF intensities per nm^3^ can be calculated as
{#f6}
A simple computation of the precipitate geometry allows HAADF micrographs to be simulated. [Figure 6(a)](#f6){ref-type="fig"} is a conventional bright field image of two precipitates, with a similar size φ of about 60 nm, in a FeCVNb alloy. EDX analysis (not shown here for the sake of brevity) indicates that the particle on the left is vanadium‐rich, whereas the one on the right is Nb‐rich. In the HAADF image ([Fig. 6(b)](#f6){ref-type="fig"}), the VC precipitate surprisingly vanishes. A careful examination of expressions 1 and 2 implies that this experimental result can only be explained if the matrix thickness *t* (which enters into the expression of the number of irradiated atoms of [Eq. 1](#m1){ref-type="disp-formula"}) is smaller than the precipitate size, and if the particles are not totally embedded into the matrix. Computations performed varying (i) the matrix thickness; (ii) the precipitates embedding depth within matrix lead to the reasonable configuration as shown in [Fig. 6(c)](#f6){ref-type="fig"}, where each precipitate is only partly buried on 30 nm within a 50‐nm iron matrix. Simulated images confirm that the vanadium carbide particle cannot be discerned under those HAADF conditions (the difference in intensity between the VC particle and the surrounding Fe matrix, hardly discernable on the simulation, is about 4%). According to this, HAADF imaging of iron thin foils does not appear to be well adapted to the identification of carbides in the present case.
CTEM observations on carbon extraction replica {#ss4-5}
----------------------------------------------
In order to avoid the magnetism of the matrix and to analyse a large number of particles, observations can be carried out on carbon extraction replica. For clarity of the comparison between the different observation modes applicable to carbon extraction replicas. [Figure 7](#f7){ref-type="fig"} reports a montage of micrographs from the same area imaged with different techniques. [Figure 7(a)](#f7){ref-type="fig"} is a CTEM image of such a replica. Particles are clearly observed and their size can be measured easily. The main limitation arises from diffraction conditions, which makes that some particles exhibit a contrast very close to that of the supporting carbon film. Hence, some precipitates are difficult to identify and measure accurately. By comparison, STEM images ([Fig. 7(c--d)](#f7){ref-type="fig"} exhibit stronger contrast and will thus be preferred for a statistical analysis of particles.
{#f7}
HAADF imaging of carbon extraction replica {#ss4-6}
------------------------------------------
[Figure 7(c)](#f7){ref-type="fig"} shows an HAADF micrograph of the same area than previously imaged in CTEM. The high number of particles observed and the high contrast obtained allows an automatic image analysis for the determination of the particle size distribution. Considering relations 1 and 2, and according to the basic development proposed in section '[HAADF‐STEM on Thin Foils in TEM](#ss4-4){ref-type="sec"}', a quantitative analysis of HAADF images is possible in order to relate intensities to the chemistry (V‐rich or Nb‐rich in the case of the FeCVNb alloy) of particles (work in progress).
SEM Observations on replicas: SE mode {#ss4-7}
-------------------------------------
From the above, it is clear that the particle size distribution can be easily measured by TEM in HAADF. However, a motivation of this study is to check whether such measurements can also be efficiently performed with a SEM, because the latter instrument is more accessible and easier to use especially for the industry.
[Figure 7(b)](#f7){ref-type="fig"} shows a micrograph of the same zone as previously described, obtained in a SEM in SE mode. Particles are clearly discernable, but again the contrast obtained is not high enough for allowing an automatic image analysis.
SEM Observations on replicas: STEM mode {#ss4-8}
---------------------------------------
Finally, [Figure 7(d)](#f7){ref-type="fig"} shows the STEM micrograph of the same area observed in the SE‐SEM and TEM modes. It can be noticed that the contrast is very good, which is favourable for a quantitative analysis of the precipitates. Furthermore, the image is spectacularly similar to that obtained in HAADF: only very subtle differences can be observed in the contrast of the particles or in the contrast of the replica.
The degree of resemblance between STEM‐SEM and HAADF‐TEM images is however directly influenced by the size and the chemistry of the particles. [Figure 8(a,b)](#f8){ref-type="fig"} compares HAADF‐TEM and STEM‐SEM micrographs of extraction replica from the FeCVNb alloy austenitized at 1200°C for 5 days. The particles observed in this state are significantly larger and 'heavier' (owing to the presence of niobium) than the particles showed in [Fig. 7](#f7){ref-type="fig"}. Once again, both imaging techniques show the same size of particles, but there are now obvious differences in their contrast. According to relation 1, the HAADF intensity is a monotone function of the composition and the thickness of the particles; hence, if the particles have the same chemical composition, the bigger they are, the whiter they appear. In SEM, the contrast of the biggest particles is here different: they appear dark in the centre. This inversion of contrast is partly due to the energy of the incident beam: in TEM, incident electrons have a sufficient energy (200 kV) to pass across the largest particles without any significant losses which would decrease the scattered intensity collected by the annular detector, whereas 30 kV electrons in the SEM experience efficient back‐scattering or other inelastic processes for larger particles which decrease the collectable STEM signal compared to thinner particles. Furthermore, other parameters, such as the detector geometry and the range of collection angles, might also affect the STEM contrast in the SEM.
{#f8}
Validation of STEM in SEM {#ss5}
=========================
It has been shown that the STEM imaging mode in the SEM leads to micrographs with a contrast adapted for the particle size distribution determination. However, to the authors\' knowledge, this technique has not been used for this purpose, and must be validated. Hence, particle size distributions obtained with this technique were compared with particle size distributions obtained from HAADF images ([Fig. 5](#f5){ref-type="fig"}). It should be noticed here that the observed areas with both techniques were chosen randomly.
Particles were identified and measured by a semi‐automatic image analysis. The error due to the fact that measurements were made on projections is negligible because the particles observed are spherical in most of the cases (only after long heat treatments, e.g. 1200°C 5 days for example, particles tend to lose their sphericity).
[Figure 9](#f9){ref-type="fig"} shows the particle size distributions of four different states of the FeCV alloy. To make the comparison easier between results obtained in STEM and in TEM, normalized densities of distribution have been represented. The results obtained with both techniques are in good agreement. Even when the shape of the distribution is slightly different, the maximum of the density appears for the same radius range. Furthermore, the mean radius values calculated from the particle size distributions obtained with both techniques are very similar, and the maximal difference observed in the case of the FeCV alloy is less than 20%. Observations performed on the same zone ([Fig. 8](#f8){ref-type="fig"}; STEM in SEM and HAADF in TEM) lead to a difference of 4% in mean radius. As the measurements were done on different replicas and that the number of particles observed is not always the same, we can conclude that both techniques give the same result.
{#f9}
STEM imaging mode on SEM is thus an efficient way for the determination of carbide size distributions in the present case. However, two limitations remain: (i) the resolution is limited by the SEM probe size (in our case precipitates as small as 5 nm have been imaged); (ii) in our configuration, the chemistry of the particles cannot be ascertained by a quantitative contrast analysis (as it should be possible in the TEM HAADF mode) because of the non‐annular geometry of the STEM detector.
Note that although improved by the use of thin specimens, EDX chemistry analysis in SEM is not applicable to the smallest particles. TEM is thus required when it is necessary to determine the chemical nature of the particles, such as in the case of the FeCVNb alloy where it is interesting to separate the populations of V‐rich, Nb‐rich or rather mixed V‐Nb carbides ([Perez *et al.* 2008](#b18){ref-type="ref"}).
Nevertheless, the interest of the STEM imaging technique in SEM can be illustrated in the case of the study of particle size evolution during isothermal reversion of FeCV alloy. [Figure 10](#f10){ref-type="fig"} presents the evolution of the mean radius at three temperatures: 870, 920, and 950°C. It should be emphasized that the numbers of experimental measurements are reasonably representative from a statistical point of view, which permits for example to use these data to calibrate a model of precipitation and dissolution kinetics of VC as presented elsewhere ([Acevedo, 2007](#b1){ref-type="ref"}; [Acevedo *et al.*, 2008](#b2){ref-type="ref"}).
{#f10}
Conclusions {#ss6}
===========
The aim of this paper was to compare different imaging techniques for the determination of particles size distributions in both FeCV and FeCVNb alloy. Main conclusions can be summarized as follows:
1As a general trend, SEM and TEM imaging techniques on both bulk samples and carbon extraction replica can give comparable particle size measurements, but time considerations reduce the interest of observations on bulk samples (SE in SEM and CTEM or HAADF on thin foils), because it may be extremely tedious to collect sufficiently large number of data in order to obtain representative results.2Particles size distributions can be perfectly determined in SEM when extraction replicas are feasible.3Observations on carbon replica, both in HAADF imaging technique and STEM mode in SEM, allow large populations of particles to be measured in a semi‐automatic way owing to the high contrast of micrographs.4STEM in SEM was validated for the particle size measurement by comparing it to the HAADF imaging technique.5However, the use of STEM in SEM is restricted to the cases where the knowledge of the particles chemistry is not required, according to the fact that very small precipitates can be difficult to analyse confidently in SEM by EDX. In such cases, HAADF‐TEM will be preferred.6STEM in SEM technique was successfully used to follow the evolution of the populations of VC particles during austenitization of a FeCV alloy at different temperatures.
The authors would like to extend their thanks to ASCOMETAL CREAS for the financial support of this study and thank Dr. P. Dierickx (CREAS) for fruitful discussion. The authors are grateful to the Consortium Lyonnais de Microscopie Electronique (CLYME) for the access to the electron microscopes. Finally, we want to thank Dr. T. Sourmail (CREAS) for reading and commenting on the draft of this manuscript. T. Douillard (MATEIS) is gratefully acknowledged for a preliminary study and preparation advices.
(describing the iron matrix as pure ferrite, with (with fcc VC, with *a*= 0.417 nm) and (with fcc NbC, with *a*= 0.447 nm).
| {
"pile_set_name": "PubMed Central"
} |
Identifying a small number of variants most likely to be causal in a specific genetic region (such as a gene) is a challenging problem, but of fundamental importance to the understanding of precise biological mechanisms that influence disease (MacArthur et al., [@B12]). Rare causal variants, by definition, are only seen a small number of times in any given dataset and therefore are hard to distinguish from the abundant natural variation that occurs in any particular gene. This is a difficult setting, where the causal variants are sparse and tend to be weak as well. Experimental functional studies are the gold-standard, but they are expensive and difficult to implement for large number of variants. Statistical methods to help us distinguish between a true causal variant and a null variant are therefore of great interest.
For those causal variants that are rare but appear at sufficient frequency in a dataset, it is possible to prioritize them over other variants in a gene using genetic (i.e., frequency) data alone. However, for those causal variants that are observed only a few times (e.g., singletons, doubletons etc.), clearly these frequencies on their own are insufficient to provide meaningful risk predictions and no method will perform better than random guessing based on genetic data alone. It is for these variants that functional genomic annotations are essential. Large-scale studies such as the Encyclopedia of DNA Elements (ENCODE) project (ENCODE Project Consortium et al., [@B5]) provide rich functional genomics annotations that can complement genetic data and help identify likely causal variants. Aggregate measures (Frousios et al., [@B6]; Johansen et al., [@B9]; Kircher et al., [@B10]; Ritchie et al., [@B14]) that combine information across many annotations in a single measure of deleteriousness for a variant have been shown to outperform individual annotations. The usefulness of such functional measures depends on how well they correlate with pathogenicity. Clearly this correlation will depend on the disease or trait under study and possibly the gene or region of interest (e.g., the degree to which a gene is intolerant to functional variation Petrovski et al., [@B13]).
It seems natural to expect that combining these two pieces of evidence: (1) the genetic score measuring the statistical evidence for a variant being associated with a disease of interest, and (2) various functional scores for the variant, can enhance the ability to identify rare causal variants. In Supplementary Figure [1](#SM1){ref-type="supplementary-material"} we show three possible outcomes for correctly classifying a true causal variant as likely causal as a function of both genetic and functional scores. Evidently, causal variants with low functional scores and low support from genetic data (e.g., singletons or doubletons) will be impossible to classify correctly. At the other extreme, the causal variants with strong support from both genetic and functional scores will be easy to identify. For causal variants falling in-between these two extremes for which neither the genetic data nor the functional scores on their own provide sufficient evidence of their deleteriousness we believe that integrative methods that combine genetic data with functional scores would be most useful for correctly identifying these variants. Indeed, a recent report by a working group of researchers convened by the NHGRI discussed the challenges in reliably implicating sequence variants in human disease, and highlighted the importance of performing "a combined assessment of the genetic, experimental and informatic support for individual candidate variants" (MacArthur et al., [@B12]).
Hierarchical models (Gelman et al., [@B7]) provide a natural approach for such an integrative analysis. They offer a formal statistical framework that can be used in this setting to incorporate sequencing data and functional annotations with the goal of identifying rare causal variants. The strength of the hierarchical modeling approach comes from the fact that it draws upon the information contained in functional annotations to augment the weak information from the case/control frequencies for rare variants, thus improving the risk estimation for individual variants. In our setting, the hierarchical model has two levels: the first level contains the parameters of fundamental interest, namely the relative risks conferred by the individual genetic variants, and also individual-level confounding factors. In the second level, the individual log relative risks of the variants are modeled as linear functions of relevant functional scores. More details on the hierarchical modeling approach and its properties as well as applications to two large population-based case-control studies of melanoma and breast cancer can be found in Capanu et al. ([@B1]), Capanu and Begg ([@B2]), Capanu et al. ([@B3]). A recent application of hierarchical modeling to *VPS13B*, a gene involved in Cohen syndrome and autism can be found in Ionita-Laza et al. ([@B8]).
We conjecture that a hierarchical modeling approach incorporating the sequencing data with an aggregate functional measure more accurately classifies rare causal variants compared to a ranking based solely on the variants\' functional score. We focus here on the recently proposed deleteriousness measure, the Combined Annotation-Dependent Depletion (CADD) score or C-score (Kircher et al., [@B10]), but our method can include any such genomic annotations. We investigate the discriminatory ability of this hierarchical modeling approach and that of the ranking based on the C-scores alone using simulations based on exome sequencing data for 861 individuals (part of the ARRA Autism Sequencing Collaborative project) for one gene, *VPS13B*, containing 166 observed single nucleotide variants (SNVs). We assumed 10% of SNVs to be truly causal variants, with relative risks of 1.1, 2, or 4 for the association of the C-score and the causal status of a variant. We then generated the case-control status following the same data generating mechanism as in Ladouceur et al. ([@B11]). Specifically, for carriers of causal variants we generated a continuous phenotype from a normal distribution with mean μ = 0.5 and a standard deviation of 0.2 while for non-carriers we used a standard normal distribution. This corresponds to a moderate standardized effect size of 0.5 (Cohen, [@B4]). Note that smaller standardized effect sizes were also examined and the results were qualitatively the same (data not shown). Furthermore, individuals with phenotype values above the median were classified as cases, and the remaining individuals as controls.
The ROC curves for the two different methods are displayed in Figure [1](#F1){ref-type="fig"} under different degrees of associations between the causal status and the C-scores and assuming μ = 0.5. Solid curves correspond to ROC curves based on the *z*-values estimated from a hierarchical model in which the C-scores were the solely functional predictor, while dashed curves correspond to rankings based on the C-scores of variants that occurred in at least one of the cases. In all configurations, the hierarchical modeling approach has superior discriminatory ability to the C-score method, with biggest improvements observed when the functional score is weakly associated with the causal status (*RR* = 1.1, blue curves). As the association of the C-scores with the causal status gets stronger, the discriminatory ability of the two methods also improves. Similar trends are observed for larger values of μ such as 1, 1.5, and 2 (data not shown).
{#F1}
In conclusion, aggregate functional measures such as the C-score can identify causal variants more accurately when used in conjunction with sequencing data through a hierarchical modeling approach than when used on their own. Therefore, hierarchical models and other integrative methods are promising tools in this context, and further work in developing such methods is warranted to take full advantage of rich functional annotation data and large scale sequencing studies for many complex diseases in order to understand the nature of the causal variants involved in these diseases.
Author contributions
====================
MC and II-L wrote and reviewed the manuscript.
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.
Supplementary material {#s1}
======================
The Supplementary Material for this article can be found online at: <http://journal.frontiersin.org/article/10.3389/fgene.2015.00176/abstract>
######
Click here for additional data file.
[^1]: Edited by: Hong-Wen Deng, University of Missouri - Kansas City, USA
[^2]: Reviewed by: Hao He, Tulane University, USA; Xiaoying Fu, Tulane University, USA
[^3]: This article was submitted to Statistical Genetics and Methodology, a section of the journal Frontiers in Genetics
[^4]: †These authors have contributed equally to this work.
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
According to World Health Organization (WHO) estimates, about 303,000 maternal deaths occurred in the world in 2015. Of these deaths, 66% occurred in sub-Saharan Africa followed by Southern Asia at 22%. About 99% of the maternal deaths occurred in low and middle income countries and most could have been prevented. The primary causes of maternal deaths are haemorrhage, hypertensive diseases of pregnancy and sepsis, and indirect causes, mostly due to interaction between pre-existing medical conditions and pregnancy \[[@CR1]\].
Focused antenatal care (FANC) provides a package of services that contributes to the health and wellbeing of a woman throughout her pregnancy, childbirth and postnatal period. Implementation of FANC was adopted by the WHO in 2002, replacing the traditional antenatal care (ANC) service model. In the FANC model, women are classified into either basic group (those with low risk pregnancies), or specialized group (those with high risk pregnancies) basing on the existing health risks identified on the women's first clinic visit. Women may therefore change their groups (whether basic or specialized) in subsequent visits depending on the absence or availability of pregnancy related risks and complications or changing clinical conditions. For routine FANC, four targeted antenatal clinic visits are recommended with the first visit occurring in the first trimester of pregnancy (between 8 and 12 weeks). The number of visits for women with specific conditions are more than four depending on their case needs related to their condition. In order to realize the full benefits that FANC can offer, there is a need for all pregnant women to attend FANC clinics for at least each of the four visits \[[@CR2], [@CR3]\].
Significant variations of FANC clinic attendance by women in the first trimester have however, been observed across regions with the lowest proportion being noted in the sub-Saharan Africa region. Recent reports on FANC utilization in the first trimester gleaned from different Demographic and Health Surveys (DHS) across African countries such as Ethiopia, Congo-Brazzaville, Ghana, Zambia, Tanzania and Malawi indicate that fewer than 30% of pregnant women attend in the first trimester \[[@CR4]--[@CR9]\].
Many countries in sub-Saharan Africa are stepping up efforts to encourage pregnant women to utilize FANC services. Community based health promotion programmes that aim at mobilizing women to utilize services are being undertaken using different approaches depending on local circumstances. One such intervention is to encourage pregnant women to attend FANC clinics with their male partners, to address the context of health care decisions mostly being made by men and mothers in-law \[[@CR10]\]. The dominance of men in society puts pregnant women at risk because they usually do not have adequate knowledge of reproductive health issues and their decisions may not adequately embrace the elements of optimal sexual and reproductive health services \[[@CR11]\].
In Malawi, the national coverage of first trimester FANC services is low at 12% \[[@CR12]\]. Recently, WHO has extended recommendations for more comprehensive coverage of antenatal interventions with up to eight contacts for additional components of care, but in Malawi and many sub-Saharan Africa countries, the challenge remains to achieve the earlier standard of four FANC visits \[[@CR13]\].
The national context of low FANC clinic attendance is not different in Mangochi District and has been oscillating between 6 and 10%. About 90% of pregnant women in the District are therefore denied early initiation of essential services such as early screening tests for HIV, STI and malaria and identification of obstetric complications such as pre-existing hypertension and anaemia that are normally offered by health facilities in the District \[[@CR14], [@CR15]\]. Barriers were long distance to the health facility, fear of wizards and witches affecting the pregnancy if it became known, and a cultural norm requiring women to seek permission from their husband before attending clinics in rural Malawi \[[@CR16]\].
Through the Safe Motherhood Initiative, Malawi is using a community engagement approach to increase health facility-based delivery and FANC clinic utilization by pregnant women \[[@CR17]\]. The District Health Office (DHO) in Mangochi is involving community leaders to bringing about positive changes related to safe motherhood. Community leaders and Health Centre Management Teams have enacted by-laws with socially acceptable local fines for people who do not comply with the edicts of the by-laws. For example, in one area in Mangochi, there is a fine of a goat if a pregnant woman delivers at home.
There were no active community based health communication elements observed promoting FANC attendance to complement efforts made at health facilities. In order to address this gap, a community based health communication strategy called Community Driven Total FANC Attendance (CDTFA) was developed and implemented in some communities in the District. In the course of implementing this approach, we conducted a study aiming at identifying barriers that were inadvertently working against the objective of increasing FANC attendance in the first trimester.
Methods {#Sec2}
=======
This is a cross-sectional study that collecting qualitative and quantitative data from both communities through participants during village level meetings and health facilities through registers and interviews with pregnant women. Inadequate knowledge among community members about the importance of early FANC clinic utilization was assumed by health workers during the 2013/2014 Health Management Information System (HMIS) and District Implementation Plan (DIP) review meeting to be one of the factors contributing to low FANC clinic utilization in the first trimester by pregnant women in the District. Figure [1](#Fig1){ref-type="fig"} illustrates some contributing factors to low FANC clinic utilization in the District.Fig. 1Factors contributing to low FANC utilization in the first trimester by pregnant women in Mangochi District
The setting and implementation of the CDTFA approach {#Sec3}
----------------------------------------------------
The District has a total population of 1,017,070 of which 528, 876 are females. The majority of people speak and understand the vernacular Chichewa language while English is an official language. The majority of the population have low income levels and live with less than one US Dollar (\$) per day. The District has 42 health facilities consisting of four hospitals, 28 health centers and 10 dispensaries. Of the 42 health facilities, 25 are government owned and 17 are privately owned. Among the private health facilities, 15 are owned by the Christian Health Association of Malawi (CHAM) and two by private practitioners. These health facilities cater for a total of 1254 villages in the District which are administratively grouped into 12 Traditional Authority areas \[[@CR18]\]. This study was conducted in 403 villages under 19 health facilities in the District, and was implemented under the authority of the District Council. The study team consisted of the Principal Investigator (PI), district based health workers -- two clinicians, three nurses, two Assistant Environmental Health Officers (AEHO) and two Senior Health Surveillance Assistants (SHSA) as master trainers; health facility based health workers in total consisting of -- 19 Medical Assistants, 19 nurses, five AEHO, 19 SHSA and 146 Health Surveillance Assistants (HSA) as facilitators.
Before implementation of the approach, the PI conducted a 3 day training of 'master trainers' who were trained on CDTFA meeting facilitation and triggering, supervision, and data collection. The master trainers in turn conducted a series of 2 day facility based trainings for village level CDTFA facilitators. Village level meetings and interviews were conducted by facilitators from the health facility of the same catchment area.
All data collection tools for meetings and interviews were translated from English to Chichewa by two Secondary School language teachers in the District. Both master trainers and facilitators were fluent in the Chichewa language. Participants were men and women of reproductive age (15--49 years), community leaders and older women (who have active community roles in matters related to reproductive health) who attended the CDTFA meetings. Women were asked to bring with them to the meetings their FANC health profile records booklets known as "Health Passports". Community members were also facilitated to choose volunteers called "lead mothers" who were assigned roles of identifying and encouraging pregnant women to start FANC clinics in the first trimester.
Data collection {#Sec4}
---------------
All data collection for assessment of the intervention was done alongside meetings from December, 2015 to January, 2016. Qualitative data were collected during the meetings through flexible interactive processes from a sample involving all men and women who attended village level CDTFA triggering meetings. While the meeting was in progress, facilitators gathered health profile records and noted those who had or had not started antenatal care in the first trimester in their most recent pregnancy (Additional file [1](#MOESM1){ref-type="media"}). The facilitators later asked these women to give open responses on reasons for either early or late attendance. In addition to tape recording, facilitators recorded the explanations that were given by women on pre-designed data collection forms. These forms included information on the day of the meeting, names of facilitators, attendance of the 15--49 year population at the event, the number of health profile books collected and scrutinized by facilitators, the number of first trimester attendees identified, brief notes on reasons for starting clinic attendance either early or late, and details of "lead mothers" selected by the community (Additional file [2](#MOESM2){ref-type="media"}).
Six months later, in June, 2016 as a preparation for a second round of CDTFA implementation, further data collection was undertaken. During this time, health centre FANC registers were used to identify women who initiated FANC either early or late during the preceding 6 months period. The identified names were sorted according to the women's respective villages and were submitted to the responsible HSA who then interviewed them using a pre-designed open response questionnaire in order to identify main reasons why they had attended FANC either early or late. The interviews used pre-designed forms to record responses from the women (Additional file [3](#MOESM3){ref-type="media"}).
The second CDTFA meetings were held in all the study villages soon after these interviews. The issues that were identified during interviews with the pregnant women were communicated back to their respective communities as part of feedback during the follow up village CDTFA meetings. This feedback was aimed at creating community awareness on issues associated with FANC clinic attendance at village level so that local solutions could be suggested in order to increase FANC clinic utilization in their respective localities.
Data analysis {#Sec5}
-------------
Individual participant data collected during the CDTFA meetings and interviews were anonymized for analysis, presentation and community feedback. All completed village level CDTFA meeting data collection forms were consolidated at health centers to come up with a CDTFA activity report (Additional file [4](#MOESM4){ref-type="media"}). The PI consolidated the 19 health centre field notes reports, recorded audios and filled questionnaires for analysis. Recorded audios were transcribed and translated into English. Analysis was done by inductive processes driven by the data itself. The PI familiarized himself with the data through reading and re-reading to look for patterns and important issues. The PI then manually identified and sorted related codes into meaningful groups to come up with categories. The categories were further collapsed to form initial themes. These initial themes were further refined and defined from which interpretations of the data regarding barriers to early clinic utilization were based on \[[@CR19]\]. For triangulation purposes, these processes were also repeated by a District based qualitative researcher.
Quantitative data from structured interviews were entered into Microsoft Excel. Through reading and re-reading of responses, categories of reasons for early and late initiation of FANC clinic by pregnant women were identified. All related responses were then grouped under these categories. After the grouping of responses, frequencies of responses under each category were tabulated. The identified reasons provided village level specific issues which were used as feedback to inform CDTFA facilitators about the type of health communication messages to be tackled during subsequent CDTFA meetings.
Results {#Sec6}
=======
A total of 403 CDTFA trigger meetings were conducted in each of the villages under the 19 health centers of the District. Out of a total population of 204,532 men and women of childbearing age in the villages, 49,797 (25%) attended the first CDTFA meetings. The percentage of women of childbearing age in the villages who attended the CDTFA meetings was 39%. Among the participants who attended the CDTFA meetings, the proportion of women was however, higher at 83% than that of men at 17%. Out of the 41,644 FANC health records booklets that were collected and checked by facilitators during the meetings, only 1548 (mean 4%; range 1--8) had first FANC clinic visit in the first trimester.
During follow up of the health facility registers 6 months later, a total of 3147 pregnant women were identified. Of these, 1015 (32%) women had started their FANC clinics in the first trimester. Table [1](#Tab1){ref-type="table"} shows the reasons the women gave for starting FANC early or late ranked according to frequencies.Table 1Reasons^a^ given by women for starting FANC clinic early or lateReasons for early FANC clinic start \[Number of women who mentioned the reason (%)\] *N* = 1015Reasons for late FANC clinic start \[Number of women who mentioned the reason (%)\] *N* = 21321. Had knowledge on importance ANC =477 (47)1. Husband not available =362 (17)2. Adherence to health messages/advice from health workers =142 (14)2. No money to pay for FANC services =298 (14)3. Went there because of sickness =122 (12)3. Own choice =298 (14)4. To know if I was pregnant = 92 (9)4. Unaware of pregnancy =277 (13)5. Encouraged by close relatives/husband =51 (5)5. No means of transport =256 (12)6. Fear to pay by-law fines =41 (4)6. No cloth wraps for the baby to show at FANC =192 (9)7. To be well received by nurses during delivery =30 (3)7. Lack of knowledge on importance of FANC = 107 (5)8. Encouraged by local leaders =30 (3)8. Long distance to FANC clinics =107 (5)9. No reason given =30 (3)9. To reduce FANC visits =85 (4)10. Waiting for visible signs of pregnancy = 64 (3)11. Was too sick to go for FANC = 47 (2)12. Cultural beliefs = 21 (1)13. Was outside the country = 21 (1)^a^Each woman was asked to provide one main reason
Identified barriers to utilization of FANC in the district {#Sec7}
----------------------------------------------------------
Four main themes were identified as barriers."A pregnant woman can access FANC services only if she has a spouse to accompany her to the clinic"It was reported during CDTFA meeting discussions in most health facilities that there are by-laws formulated by health officials and community leaders requiring that every pregnant woman is to be accompanied by a spouse when going to attend FANC clinics. In situations where a spouse is not available, a pregnant woman is supposed to bring with her a stamped letter from her village head attesting the unavailability of a spouse. During the CDTFA meetings, it was reported that for most women to get the letter of exemption from a village head a monetary fee had to be paid. Various amounts of fees ranging from the equivalent of 1.5--4 US\$ were reported in different areas. When the women were asked why they started attending FANC clinics late, 17% (*n* = 362) indicated the unavailability of their spouses to escort them to the health facility as per stipulations of the by-laws. One woman explained this as follows:"*"You cannot be assisted by doctors if you go to the hospital without a spouse. I did not go to ANC clinic early because my husband was in South Africa. I could not afford to pay for the letter from the village head".* -- a female participant.""Male involvement in reproductive issues contradicts their traditional norms"Another reason reported during the CDTFA meetings for women not to bring along their spouses is that according to tradition, issues related to pregnancy and childbearing are mostly viewed as women's responsibilities. Men therefore, do not want to go with their spouses to FANC clinics for fear of being laughed at and stigmatized by their fellow men."*"Our nurses should just be accepting our women to access ANC services even if they go alone. Men in this village feel shy to accompany their spouses to clinics. Starting from parents, pregnancy related issues have always been handled by women".* -- male participant.""Men view FANC clinics as platforms for forcing them to be tested for HIV"Following the national policy guidelines, PMTCT is being implemented in FANC clinics in all health facilities of the District where couple counseling and non-coercive HIV testing are promoted. Community members fear the associated stigma and discrimination should they test positive as illustrated in the quote below:"*"Is it possible to stop coercing male partners getting an HIV test when they accompany their spouses to clinics? I see that male partners do not want to accompany their spouses because they do not want to be tested for HIV"* -- a female participant.""FANC services are for wealthy people"At community level, women are encouraged to save money and supplies to be used during delivery. During CDTFA meetings, it was reported that as part of this preparedness, health workers demand that pregnant women bring with them cloth wraps for the baby (traditional *'chitenje'* singular/ *'zitenje'* plural) at their initial FANC visit. The number of *'zitenje'* requested ranges from one to four and this demand was viewed by 9% (*n* = 192) of pregnant women who started FANC clinics late as unaffordable and contributed to their delay to start FANC clinics."*"If my wife tells me that she is pregnant, it becomes difficult for me to source four 'zitenje' within three months of her pregnancy as demanded from the nurse at the hospital. In this case, I tell my wife to wait until I find money to buy the four 'zitenje'. I feel the demand is too much for poor people like us."* - a male participant."The cost of one *'chitenje'* locally was reported to be the equivalent to 1.5 US\$. Participants in CDTFA meetings stated that they could not afford the total cost of four cloths within the first 3 months of pregnancy as required by the health facilities. Due to this requirement for birth preparedness, some pregnant women especially those who did not have spouses reported it caused them to incur additional costs for them to access FANC services. If a pregnant woman did not have a spouse, or the spouse was not available during her time to start FANC clinics, she in addition to sourcing a letter from the village head, had to meet the cost of *'zitenje'*."*"We have a by-law that was enacted by the chiefs in this area after receiving concerns from the health centre that most men were not accompanying their spouses to ANC clinics. We do not punish pregnant women who do not have spouses. We advise them to go to their village head to collect letters for them to be accepted at the health centre. The fee that she is complaining of is not meant for such pregnant women only. The fee is for any stamped letter that everyone pays when one asks for it from the chief"* -- a village head*.*"While FANC health services are free at the point of use in government facilities, communities living in areas under CHAM health facilities pay user fees. Integration of services in FANC clinics means their workload has increased in terms of consultation and medical supplies. Pregnant women in these areas have to pay for FANC clinic services. During the CDTFA meetings, most women who attended FANC clinics in CHAM facilities attributed their late initiation to the user fees as evidenced by the following quote:"*"The services are not free at our health facility. We pay 2,000 Kwacha* \[3 US\$\] *to access services which many of us do not afford. Is it possible for government to be paying for us money for ANC so that we can be accessing ANC services free of charge as is done in government facilities?"* -- a community leader from an area under a CHAM health facility."
Discussion {#Sec8}
==========
Our intervention reached approximately 39% (41,477 women) of those of childbearing age. The low level of male participation (17%) in reproductive health related activities has also been reported in many sub-Saharan Africa countries contexts including Tanzania, Kenya and Cameroon \[[@CR11], [@CR20], [@CR21]\].
Six months after implementation of CDTFA trigger meetings, community based assessment of first trimester attendance at clinics showed an increase from 4 to 32%. This increase cannot however, wholly be attributed to implementation of CDTFA meetings. During the study period other interventions were deployed at some sites including urine pregnancy-testing in line with the framework shown in Fig. [1](#Fig1){ref-type="fig"}. Despite interventions, the percentage of women with delayed first attendance was still substantial at 68%. Promotion messages for men's participation in FANC clinics, routine HIV testing for couples during FANC clinics, and a requirement for pregnant women to bring cloth wraps as proof for birth preparedness, though with good intentions, have inadvertently deterred some women from accessing FANC services in the first trimester. In addition, the payment of user fees in CHAM and private health facilities has also been shown to bar pregnant women from accessing essential FANC services.
Three explanations came out of the CDTFA meetings why male partners do not escort their spouses to FANC clinics. Firstly, some men in Mangochi District live for significant amounts of time away from their homes because they mainly rely on small business, some men leave their families for neighbouring Mozambique and South Africa for employment opportunities \[[@CR18]\]. Secondly, low male involvement in the District is being linked to culture that view issues of pregnancy and childbirth as female responsibilities. Thirdly, HIV testing services being done in FANC clinics are also discouraging participation of men because of fear of stigma and discrimination. These findings are consistent with those from Kenya \[[@CR20]\]. Globally, there has been a drive to increase male involvement in reproductive health services including FANC. Studies across the sub-Saharan Africa region have shown that increasing partner involvement can influence utilization decisions by women of reproductive services like HIV testing and PMTCT measures \[[@CR11], [@CR22], [@CR23]\]. Furthermore, there is evidence that involvement of men has an effect of lowering attrition rates in FANC and PMTCT programmes \[[@CR10], [@CR24]\].
While appreciating the need for pregnant women to receive FANC services with their partners, the health authorities in the District need to identify other strategies to increase partner involvement rather than the use of by-laws. In a Tanzanian study, systematic coaching of pregnant women on how to invite their partners followed by written invitations resulted in an increase in male partner participation in FANC services \[[@CR25]\]. In a similar study on women's attitudes towards male involvement in FANC in Lesotho, improvement in structural design and clinic services and a comprehensive community mobilization on male involvement were recommended to increase partner involvement in FANC services \[[@CR26]\].
Poverty is another barrier to utilization of FANC services by communities in the District. About 73% of people in Mangochi are classified as poor and live below the World Bank prescribed 1 US\$ per day threshold \[[@CR27]\]. Despite this economic status, utilization of FANC services by pregnant women in Mangochi District is being associated with some financial costs. Mangochi District Health Office is implementing a Service Level Agreement (SLA) with CHAM and private practitioners in maternal and neonatal health services. In these contracts, the Ministry of Health through the District Health Office pays for maternal and neonatal health bills that are incurred by women in areas where there are no government health facilities. The contracts however, mostly cover costs at delivery and the neonatal period and do not extend to FANC services. A review of these contracts may be necessary to determine whether it is feasible to include FANC services in the agreements. Findings about the costs associated with FANC clinic attendance in Mangochi District are consistent with observations from a study Ghana, Kenya, Malawi and Nigeria \[[@CR28], [@CR29]\]. Socio-economic status of women and their husbands was related to the timing of initiation of FANC clinics.
From a rights based perspective it can be argued whether these prescriptive requirements are fair. The financial and logistical challenges faced by pregnant women in our District cannot easily be met through women's own resources. Universal access to health care implicates elimination of geographical, economic, social-cultural, and organizational or gender barriers that impede all people from using integral health services \[[@CR30]\].
The findings presented here reflect experience gained working with communities and were largely un-anticipated: we had not specifically intended to collect and analyze data on service related barriers in our CDTFA implementation. It is the richness of the views and explanations that were associated with timing of starting FANC clinics from community members during the initial stages of introducing CDTFA meetings that ignited our interest to collect and analyze the data presented in this paper. Taking into consideration our local context, the findings of this study are in consistent with experience from Rwanda, Nigeria and Tanzania \[[@CR2], [@CR11], [@CR15]\].
Strengths and limitations {#Sec9}
-------------------------
The participatory facilitation of the meetings allowed participants to freely give their opinions which were further discussed by all. At the end of each meeting, facilitators gave summaries to enable consensus. The separate interviews that were conducted with women provided an additional opportunity for participants to share views. This was implementation research so we considered it appropriate to use data collectors affiliated to health centers serving the same catchment areas. It is possible that additional barriers could have been identified if external data collectors had been deployed who may have been perceived as more independent.
Conclusions {#Sec10}
===========
In our setting, implementation of an internationally mandated policy has been associated with unintended barriers constraining improvement in FANC utilization. The first visit prescriptive requirements (partner involvement and cloth wraps for maternity) were reported to bar pregnant women from accessing FANC services. In addition, the fees that were being imposed on women who do not have a partner to accompany them and user fees at private health facilities were not affordable for poor pregnant women in the District. Consequently, a significant proportion of women in the District are unable to access services in the first trimester, with consequent risk of adverse outcomes. Nationally mandated programming needs to address local implementation elements in order to realize the full gains in access to antenatal services and avoid inadvertently institutionalizing barriers to access. There is need for the District authorities to explore strategies that will promote uptake of integrated services in FANC clinics without infringing women's rights to access health services. District based stakeholder engagement can help to redefine some of the community mobilization strategies to enable community members to come forward, express their needs and challenges, and engage in genuine partnership with service providers. Policies on free access to maternal health services also need to be reinforced to remove indirect burdens of the type identified here.
Additional files
================
{#Sec11}
Additional file 1:CDTFA meeting facilitation guide. (DOCX 16 kb) Additional file 2:Village meeting data collection form. (DOCX 12 kb) Additional file 3:Interview guide for pregnant women. (DOCX 17 kb) Additional file 4:Health centre village meetings summary form. (DOCX 12 kb)
ADDRF
: African Doctoral Dissertation Research Fellowship
ANC
: Antenatal Care
CDTFA
: Community Driven Total FANC Attendance
CHAM
: Christian Health Association of Malawi
DHO
: District Health Officer
DHS
: Demographic and Health Survey
DIP
: District Implementation Plan
FANC
: Focused Antenatal Care
HIV
: Human Immuno Deficiency Virus
HMIS
: Health Management Information System
HSA
: Health Surveillance Assistant
ICEIDA
: Icelandic International Development Agency
PMTCT
: Prevention of Mother to Child Transmission
US\$
: United States Dollar
WHO
: World Health Organization
**Electronic supplementary material**
The online version of this article (10.1186/s12884-017-1631-y) contains supplementary material, which is available to authorized users.
Icelandic International Development Agency (ICEIDA), Amref Health Africa, and Africa Population Health Research Centre (APHRC) for the financial support. Mangochi District Council through the District Health Office for allowing this study to be conducted in the District. The community of Mangochi District for participating in this study.
Funding {#FPar1}
=======
This work was supported by Icelandic International Development Agency (ICEIDA), Africa Population Health Research Centre (APHRC), Nairobi, Kenya and Amref Health Africa as follows:ICEIDA provided funding for meetings for authors during the design and manuscript writing stages, supported allowances for enumerators, transportation costs and stationery during implementation and data collection stages; and also assisted towards costs for services of a statistician during analysis and interpretation stages.KCM holds an African Doctoral Dissertation Research Fellowship (ADDRF) from Africa Population Health Research Centre (APHRC), Nairobi, Kenya. The fellowship provided funding to support logistics for enumerators during the implementation stage; also for acquisition of a laptop and software for data processing and analysis.Amref Health Africa procured pregnancy test kits that were used in this study.
Availability of data and materials {#FPar2}
==================================
The datasets during and/or analyzed during the current study are available from the corresponding author and can readily be accessed on reasonable request.
Study Design: KCM, ASM, WS; Data Collection and Analysis: KCM; Manuscript preparation: KCM. ASM, WS critically reviewed the manuscript for content. All authors have read and approved the final version of this manuscript.
Ethics approval and consent to participate {#FPar3}
==========================================
This work was undertaken as part of an intervention study to assess strategies that can increase Focused Antenatal Care clinic attendance in first trimester for pregnant women in Mangochi District. The protocol of the main study was reviewed and approved by the Malawi College of Medicine Research Ethics Committee. Written permission was obtained from Mangochi District Council through Mangochi District Health Research Committee. Considering that the study involved community members, meetings through District Executive Committee and Area Development Committees were conducted to explain about the study to community leadership and get their consent before the start of data collection. All participants signed a written consent form (translated into the local language of Chichewa) after receiving information about the study objectives and aims. We obtained signed written consent forms from parents or legal guardians for all participants under the legal age of 18 years.
Consent for publication {#FPar4}
=======================
This manuscript does not contain any individual person's data in any form, hence consent for publication was not sought.
Competing interests {#FPar5}
===================
The authors declare that they have no competing interests.
Publisher's Note {#FPar6}
================
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
| {
"pile_set_name": "PubMed Central"
} |
1.. Introduction {#sec1}
==================
The history of search--match in the crystallographic context coincides with the history of Hanawalt's work and the International Center for Diffraction Data (ICDD), and a good account of those early days is given in the paper by Hanawalt (1986[@bb16]).
Following the 'Hanawalt' manual search, the appearance of the personal computer revolutionized search--match, making it faster, more powerful and easier to use (Smith & Gorter, 1991[@bb41]; Langford & Louër, 1996[@bb20]). More lines could be used, subsets and chemical information helped the identification of the correct phases, and errors were taken into account. Most of the subsequent work has been devoted to correctly determining peak positions in an automated way and using, as much as possible, the intensity information. These search--match methods are conducted in two steps: the first step is a classical search--match using the identified lines, and then, in a second step, the possible structures are compared directly as lines with the powder pattern to further refine the choice manually.
The majority of search--match improvements have resulted from the algorithms and software being able to provide increasingly reliable and automated results, limiting human intervention. In parallel, another major effort has been dedicated to increasing the number of 'cards' from Hanawalt's days, using experimental data from different laboratories and collected by ICDD into the Powder Diffraction File (PDF) databases. In the past ten years, large improvements in both the number and the reliability of cards have been obtained by including calculated 'cards' or patterns. Crystal structures from different databases, notably the Inorganic Crystal Structure Database (ICSD; Hellenbrandt, 2004[@bb18]) and the Crystallography Open Database (COD; Gražulis *et al.*, 2009[@bb12], 2012[@bb13]), have been added to search--match procedures. As these databases provide crystal structures for identified phases, one can proceed with phase quantification using the Rietveld method (Rietveld, 1967[@bb37], 1969[@bb38]), which does not require any standard or knowledge of reference intensity ratio factors.
In 2002, Le Meins *et al.* (2003[@bb23]) organized a search--match round robin, and the overall result was satisfactory. It proved that 'a good search--match software, in expert hands, combined with an up-to-date PDF can perform efficiently in solid state phase identification'. But it also showed that only one participant was able to succeed in identifying all compounds, and 'in expert hands' can still be interpreted as being far from an automatic procedure for beginners. Furthermore, the investigation of nanostructured compounds, or in general not perfectly crystallized samples, was not a concern in this round robin. For such compounds, the increased peak broadening makes reliable peak finding and indexing more difficult to perform, and unfortunately increased instrument resolutions are of no help. Another general problem is that the traditional search--match, in certain cases, identifies a large number of phases, which makes it difficult for the non-expert to discern which ones are really present and which are not. This is in part due to the ranking mechanism, which uses mainly a list of identified peaks with the main emphasis on number and position as the true overlapped intensity cannot be exploited.
As a general trend, actual search--match procedures are very efficient in identifying phases in compounds with well defined and crystallized structures, coming from synthesis or chemical laboratories. But this efficiency decreases for less well defined structures, because of their poorer crystallinity or/and their strong inter- and intra-phase overlap, *e.g.* in soil samples or raw industrial materials.
In 2009, a project named 'Nanoair' was founded by the European Community to develop a fully automatic and remotely controlled instrument to collect nano-powders from the air and perform a complete analysis for particle dimensions and composition (phase composition by diffraction). The analysis part of the project focused on performing an automatic phase identification and quantification using the Rietveld method (Werner *et al.*, 1979[@bb49]; Hill & Howard, 1987[@bb19]; Bish & Howard, 1988[@bb3]). The results of the phase identification can be used in a further step for the Rietveld-based quantification only when the crystal structures of identified phases are known. Therefore, we decided to work on a new procedure to merge the phase quantification and search--match in a single step using the Rietveld method. In that project, the software/procedure FPSM (full-profile search--match) has been developed. Since then, the method and software have been tested and improved, and the full theoretical basis, details of the implementation and some test results of the search--match are presented in the following. Structural and profile parameters are not fully refined in the Rietveld quantification in order to conduct the procedure automatically. When more accurate quantification is required, we advise the use of a full Rietveld refinement employing the identified phases or application of another of the available quantification methods, in particular if unidentified phases are present (Scarlett & Madsen, 2006[@bb40]; Toraya, 2016[@bb45], 2018[@bb46]). We will not discuss the quantification here, only the search--match.
2.. Theory {#sec2}
============
2.1.. Automatic Rietveld refinement {#sec2.1}
-------------------------------------
The entire search and profile fitting algorithm is based on the Rietveld method (Rietveld, 1969[@bb38]). In particular we use a special implementation, as in the *MAUD* software (Lutterotti, 2010[@bb24]), that differs from the classical Rietveld implementations in two main points:
\(1\) The peak profile function is composed of two different contributions: the instrument and sample broadening. The two contributions are directly convolved into the resulting peak broadening, and in the analysis we do not refine the instrument broadening contribution, only the crystallite sizes and r.m.s. microstrains.
\(2\) The refinement strategy, *i.e.* how the parameters are refined in successive order, is automatically decided by the algorithm and does not require human intervention or decision.
Only the second point is strictly necessary for the Rietveld search--match algorithm, but the first is convenient as it directly provides the quantities of interest.
The general formula used for the Rietveld computation is the following:where bkg() is the background function, *I* ~0~ the diffraction part of the incident intensity, *N* ~p~ the number of phases, *f* ~*j*~ the volume fraction of phase *j*, *V* ~*j*~ its unit-cell volume, Nr~*j*~ the number of reflections of phase *j* inside the pattern range considered, *L* ~*k*~ the Lorentz--polarization factor for reflection *k*, *F* ~*k*,*j*~ its structure factor, *S* ~*k*,*j*~() the peak shape function of reflection *k* of phase *j*, *d* ~*k*,*j*~ its position in *d*-space coordinates, *P* ~*k*,*j*~ the texture or preferred orientation factor, and finally *A* ~*k*,*j*~ the absorption factor. We prefer the variable *d* ~*i*~ as coordinate for the pattern instead of 2θ because we use it here for both angular dispersive and energy dispersive \[*e.g.* neutron time-of-flight (TOF)\] techniques.
Now we detail some of the functions and models used in the Rietveld algorithm implementation.
For the background function we use a simple polynomial function plus an additional Gaussian function at zero 2θ (but not for TOF/energy dispersive patterns) to represent the decay that we can find in some experiments close to the incident beam. In the case of electron diffraction in transmission electron microscopy (TEM), such a strong contribution is always present, resulting in a strong increase of the background at large *d* values.
For the peak profile we use an asymmetrical pseudo-Voigt function as described elsewhere (Enzo *et al.*, 1988[@bb10]; Lutterotti & Scardi, 1990[@bb29]). The instrumental broadening part can be determined in advance using a standard sample (Scardi *et al.*, 1994[@bb39]), and to simplify its input it was made compatible with that in the *MAUD* software (in *MAUD* we added a menu function to export a CIF containing the data and the instrument characteristics and broadening ready to be loaded and used in the web interface of the FPSM program). If the experimental details are not supplied alongside the data with a suitable CIF, the web interface can be used to specify some general categories for instrumental broadening to represent usual cases and any additional information needed regarding the experimental conditions. An imperfect definition of the instrumental contribution does not prevent the search--match from working correctly (unless a low-resolution instrument broadening is used for a high-resolution instrument and well crystallized sample), but it will give wrong results for the crystallite size and microstrain values and is likely to make the search more difficult.
The broadening due to the crystallite size and microstrain for each phase is computed using a simple inversion of the model described by de Keijser and co-workers (de Keijser *et al.*, 1983[@bb6]; Delhez *et al.*, 1983[@bb7]). In the Rietveld algorithm we first compute the integral breadth (both the Gaussian β~g~ and Cauchy β~c~ parts) for the pseudo-Voigt function from the value of crystallite size 〈*D* ~V~〉 and r.m.s. microstrain 〈ε〉^1/2^ according to
The integral breadth is then convoluted analytically with the pseudo-Voigt function, representing the symmetric instrument broadening part, and further numerically convoluted with the asymmetry function as described by Lutterotti & Scardi (1990[@bb29]).
The Lorentz--polarization correction is computed according to four different geometries: Bragg--Brentano, Debye--Scherrer, transmission (TEM or X-ray diffraction) and neutron TOF. No correction for monochromator or highly polarized beams like on a synchrotron is used, as in nearly all cases it will not greatly affect the results. In general for a better refinement, after the phase identification and first quantification using this method, a more accurate complete Rietveld refinement may be performed using all corrections and required models.
We decided not to apply a texture correction in the present FPSM, for three reasons:
\(1\) A preferred orientation correction such as the March--Dollase (Dollase, 1986[@bb8]) requires human intervention for determining the preferred orientation direction.
\(2\) The spherical harmonic model (Popa, 1992[@bb35]) is time consuming and it is not advisable to use it in general for phase quantification as it may not correctly represent the effective sample texture (Lutterotti, 2012[@bb25]).
\(3\) The use of a true orientation distribution function requires additional experiments (Ferrari & Lutterotti, 1994[@bb11]; Wenk *et al.*, 1994[@bb48]; Lutterotti *et al.*, 1997[@bb28]) to cover the orientation space.
The approximation of a randomly oriented sample may cause problems for the identification of certain clay minerals, and at present we are working on a specific solution for this case.
For the absorption-volume correction, the formulation depends on the diffraction geometry and four models were implemented: (1) Bragg--Brentano with θ/2θ measurements, for which the absorption factor is a constant and can be ignored for bulk samples, (2) flat sample in reflection with 2θ measurement only, where the absorption factor is calculated as a function of 2θ from the incident beam angle and phase composition, (3) Debye--Scherrer camera with the absorption-volume correction for a cylinder, and (4) no absorption correction for X-ray diffraction or TEM transmission and TOF neutrons.
Finally, we explain how an automatic Rietveld refinement is performed in the code each time we evaluate the presence of a phase. The execution strategy is very similar to a Rietveld analysis focused on phase quantification and not on crystal structure refinement. We remind the reader that the goal is to obtain a quick correct analysis and avoid as much as possible a divergent solution. Thus we reduce the number of refined parameters to the minimum possible and emphasize physically meaningful parameters/formulation even if this does not result in the best possible fit of the pattern. The refinement strategy is reduced to three steps, which correspond to the first three steps performed by a Rietveld expert when performing a pattern fitting. The first step consists of refining only the background and scale parameters (each *f* ~*j*~ *I* ~0~ as phase scale parameters), the second step additionally refines peak positions, and finally in the third step the peak profile parameters are introduced. No crystal structure is refined.
To go into more details on our specific Rietveld formulation, for each step we perform the least possible number of iterations to reduce the execution time. We actually use two different types of refinements. One is used during the search or scan through all the selected structures. It is focused on speed and it does not use more than three iterations per step. The other is used for the phase quantification and it requires a better fitting. It is conducted only at the end of the selection of a new phase and thus it does not need to be as fast as possible. During the first step, three to five background parameters are refined in the polynomial function (depending on how large the data range is): the phase scale parameters and a decay function for the diffuse scattering halo around zero. This step is also used to get a first estimate of all volume fractions. As the volume fractions should amount to a total of 100%, at the end of each iteration, the volume fractions are normalized and the normalization factor is used to change the incident intensity *I* ~0~ \[see equation (1)[](#fd1){ref-type="disp-formula"}\]. We keep this particular procedure to remain compatible with the calculations in *MAUD*, in which working with volume fractions (not simply phase scale factors) is necessary when dealing with multiple data sets (from different instruments) or layered and multilayer samples.
After the first step, if there are multiple phases and one or more of them results in a volume fraction below a minimum threshold (*f* ~*R*0~, normally between 0.5 and 0.1%), these are eliminated from the refinement. For the remaining phases, in the second step we add the unit-cell parameters to the refinement, but only for the phases having a volume fraction over a given threshold (*f* ~*R*1~, around 5--10%). In addition, a classical 2θ or constant *d*-space shift is refined to account for goniometer misalignment or measurement errors. Regarding the cell parameters, it is not advisable in our case to refine all of them, especially for triclinic or monoclinic phases, so during the search algorithm we use instead a volume expansion or contraction parameter, thereby reducing the total number of refined parameters added in this cycle to (1 + *N* ~p~). In the third step we add the crystallite size and microstrain parameters to the refinement, for each phase with a volume fraction over a specific threshold (*f* ~*R*2~). During this step we refine also an overall isotropic Debye--Waller factor common to all atoms of all phases.
2.2.. Search and match algorithm structure {#sec2.2}
--------------------------------------------
We describe in this section the main algorithm used for what we call the full-profile search--match (Fig. 1[▸](#fig1){ref-type="fig"}). We start from a crystal structure database restricted to match the elemental composition conditions. In principle the method may work with all the possible phases without any composition restriction, but we have to consider the computing time, and limiting the number of phases using some restrictions on the composition or type of sample (*e.g.* mineral, inorganic...) means that a search can be performed in a few seconds or minutes.
Let us call *S* ~1~ the restricted crystal structures subset, composed of *N* potential phases, and *S* ~2~ the subset containing the identified phases, initially empty. Each phase of *S* ~1~ is then evaluated against the experimental pattern by performing an automatic Rietveld refinement as described at the end of the previous section (three-step procedure). At the end of this run, the *N* phases are ranked on the basis of a figure of merit (FoM, defined later) and the rank-1 phase is moved to *S* ~2~. All phases of *S* ~2~ (at the moment just one) are then added in turn to each of the (*N* − 1) remaining phases of *S* ~1~ for another automatic Rietveld run over these (*N* − 1) combinations of phases. The combination (*S* ~2~ + 1 phase from *S* ~1~) providing the best FoM indicates the next phase that has to be moved to *S* ~2~, leaving in *S* ~1~ only (*N* − 2) phases. At this point a more accurate Rietveld refinement is performed to evaluate the volume fractions for all the phases found up to now. If one of the *S* ~2~ phases falls below a critical threshold (*f* ~*S*0~, between 0.5 and 0.1%), the phase is moved out of *S* ~2~ and another cycle starts to evaluate the next phase to add to *S* ~2~. To avoid similar or duplicated phases in the analysis, a special check is done by comparing the newly found phase with each one in *S* ~2~, and in the case of a positive match, the phase is not added but is removed also from *S* ~1~ and the next phase in the rank is chosen. The check is done by fitting the calculated pattern of the last phase added to *S* ~2~ with the other structures in turn. If the *R* ~wp~ factor from the fitting is below a certain threshold, the two phases are considered similar. The process is carried out until either the phase removed from *S* ~2~ as a result of being below the threshold *f* ~*S*0~ is also the last one added (this means that no phases over that threshold were found in the last cycle) or there are no more phases left in *S* ~1~ or the number of phases in *S* ~2~ is over an imposed limit. We keep the last possibility as a precaution to stop the algorithm in case of a completely failed identification, giving rise to several phase additions without a real fit improvement. In our still brief experience with the algorithm, the latter happens when (i) severe errors are present in the pattern (including wrong data formats *etc*.); (ii) the pattern does not correspond to the declared experimental conditions; or (iii) neither the phase nor a closely related structure is present in the database.
By default, after each cycle, we eliminate from *S* ~1~ not only the phases moved to *S* ~2~ but also the ones resulting in a volume fraction below a threshold *f* ~*S*1~, to speed up the next Rietveld ranking by having fewer structures in *S* ~1~.
2.3.. Figure of merit {#sec2.3}
-----------------------
The simplest figure of merit that can be used to rank the crystal structures during one cycle is based on the *R* ~wp~ factor (Prince, 1983[@bb36]), computable directly by the Rietveld fitting. As we are using it to test the suitability of a phase to reproduce a pattern and we aim to decide whether the phase is present or not, it may be reasonable to use the most general and best overall Rietveld index.
In such a case the figure of merit becomesThis FoM was also the first one used during testing and tuning of the algorithm. But we found it not completely suitable for seeking and sorting out the structures that are present in a pattern for several reasons, of which the main ones are as follows:
\(1\) The *R* ~wp~ criterion does not distinguish between two similar phases from the database mainly differing by their cell parameter values. This may happen for two structures originally determined at different temperatures or with compositional modifications, for which the structure factors and unit-cell parameters do not change too much.
\(2\) *R* ~wp~ privileges strong scattering phases. Phases with larger phase fraction should have greater weight.
\(3\) Phases with substantial broadening can significantly reduce *R* ~wp~ by simply improving the background fitting.
To overcome these problems we propose a modified FoM, where the phase-related parameters are the ones of the phase for which we are evaluating the FoM:where ρ~0~ is the starting density, ρ~r~ is the refined density and *f* ~r~ is the volume fraction of the structure for which we are evaluating the FoM from the Rietveld fit. The quantities *a*, *b*, *c* and *d* are weighting coefficients to be tuned for optimal performance. δ is a strong penalty function that reduces the FoM if the mean crystallite size of the phase is below a customizable threshold value (\<*D* ~t~, default is 20 Å) and/or the r.m.s. microstrain is too large (*e.g.* \>∊~max~, default is 0.02). The term weighted by the coefficient *a* gives more importance to structures with the closest cell parameters to the observed ones. The term weighted by *b* favours the low scattering phases, so these are not excluded in virtue of the non-negligible volume fraction they are representing.
A correct tuning of the FoM parameters is important for the success of the algorithm, as the *R* ~wp~ criterion alone is not sufficient to cover all the possibilities. But there is one additional note: the FoM as described here is useful only to compare and rank the crystal structures in one cycle, but not to follow the entire search--match progress or compare different cycles. This because each cycle, to identify the next phase, is different from the previous cycles (*S* ~1~ and *S* ~2~ are different) and the new best matching structure may have a higher penalty function. To follow the overall progress, *R* ~wp~ or the WSS (weighted sum of squares; Toby, 2006[@bb44]) are better suited, and in general we always observe them decreasing quite a lot during the initial identifications and then levelling to a minimum when arriving at the minor phases.
3.. Implementation {#sec3}
====================
The adoption of the FPSM procedure as a routine tool for quick automatic finding of more likely structures depends heavily on the speed of execution of the code. For this reason we tested in the implementation different possibilities of how to perform the calculation and how to store and parse the crystal structure database. For the implementation and testing we used both a portable computer with an i7 quad-core 2.8 GHz processor and a workstation with 2 × 2.93 GHz six-core Xeon processors, resulting in 158 gigaflops. Both systems were equipped with a solid state drive (SSD). The programming language used for the main program was C++, and we employed the database management system MySQL to store/retrieve the crystal structures. All the crystallographic computations (space-group interpretation, symmetry operations, reflection list and structure factor calculation) were carried out using the *Computational Crystallography Toolbox* (*cctbx*; Grosse-Kunstleve & Adams, 2002[@bb14]; Grosse-Kunstleve *et al.*, 2002[@bb50]). The latter proved very efficient in all the calculations and we used only the C++ implementation without the Python bindings. We added the scattering factors for electron diffraction to the *cctbx* library as tabulated by Peng *et al.* (1996[@bb34]), to enable a simple kinematical calculation of structure factors.
We tested two ways of preparing intensities of individual reflections:
\(1\) Preliminarily calculate the intensities and store them in the database.
\(2\) Calculate the intensities when the structures are loaded from the database.
The full computation took more time for the former than the latter. It was not shortened even if a quick SSD was used or the database was further optimized in the former. Thus the latter was adopted in the present computation. Moreover, the latter reduces the size of the database for parsing and gives flexibility in changing the radiation type. For the nonlinear least-squares algorithm we use a Marquardt algorithm (Marquardt, 1963[@bb32]), whose implementation was derived from the *MAUD* software (Lutterotti, 2010[@bb24]). Since numerical models were partly used in the formulation, all derivatives were computed numerically.
We discovered that finding minor phases with this method required a good reproduction of the background, as just the differences between experimental and calculated data points were used. The derivative differences method (DDM; Solovyov, 2004[@bb42]), on the other hand, does not rely on a good fit of the background as it minimizes first and second derivatives instead. So in FPSM, we implemented, in addition to the classical Rietveld method, the refinement indices and weighting schemes of DDM as reported in the aforementioned paper and manual (<https://sites.google.com/site/ddmsuite/>). In the search and match one can choose the Rietveld scheme, the DDM, or both Rietveld and DDM with a weight between the two.
In this first implementation and testing, all the structures imported into our database were extracted from the COD (Gražulis *et al.*, 2009[@bb12], 2012[@bb13]) in CIF format (Hall *et al.*, 1991[@bb15]). In principle, we can use any other source providing crystallographic structures in CIF format, but for testing and our first database we used the COD because it contains the American Mineralogist Crystal Structure Database (Downs & Hall-Wallace, 2003[@bb9]) and the zeolites database (Baerlocher *et al.*, 2007[@bb2]) and it is the largest one comprising inorganics, organics and organometallics all together for which CIFs are freely available.
A dedicated interface has been written to perform the following tasks:
\(1\) Input all the CIFs from a local repository (at present, we use a clone of the COD, and we plan to implement later a function for updating the database when structures are added to the COD).
\(2\) Save the database in MySQL format and carry out all the maintenance or editing required.
\(3\) Check each structure for consistency: cell content with respect to chemical formula, space group and symmetry operations. At the present state of implementation we do not make any check for bond distances.
\(4\) Classify the phases as inorganic, organic or, when no clear distinction can be made from the raw chemical formula, unknown (organometallic mostly).
In the latest version of the database imported from the COD, 379 097 structures were loaded, 148 were removed because of retraction (Harrison *et al.*, 2010[@bb17]), 1567 were duplicates of others, and 1166 were not loaded because they contain an unrecognized space group, unrecognized atoms or chemical composition problems. Most of the unloaded structures were modulated structures (unrecognized space group) that are not supported in the present implementation. A few crystal structures were rejected because there was a significant mismatch between the chemical composition and the cell content; some were missing the atoms completely. For these files, even a check with the original article was not sufficient to solve the problem clearly (most of the time the CIF is a partial solution or a general framework with the atom-site content not specified; *e.g.* atom *X*, where *X* is one of the transition elements), and we prefer not to use them in our method. However, we could successfully identify, validate and use more than 99.2% of the structures from the COD.
The core of FPSM is devoted to the full-profile search--match only, and the implementation was based on the layout shown in Fig. 1[▸](#fig1){ref-type="fig"}. We chose to keep it independent of any graphical user interface for ease of compilation and to allow it to run on multiple platforms. This core part can be compiled and used as a library by any graphical interface and as such was included in the previously described program interface to build the database and test the algorithm. But to permit more widespread use, we have built an interface accessible through a web page.
The web interface is shown in Fig. 2[▸](#fig2){ref-type="fig"}. It has a minimal input to keep it simple and is focused mainly on usability. It has three sections for the input. In the first section the user loads the pattern to be analysed as a simple text format with two columns: 2θ (or *d* spacing in the case of TOF measurements) and intensity. After loading the pattern the user may restrict the database to some subsets (*e.g.* mineral only or organic only) and specify a possible composition, and if a list of chemical elements is entered the program will exclude all the phases having at least one atom not in the list. This may be helpful to restrict the number of phases to search from and it is strongly advised to use it. In case of doubt many additional atoms may be included to lower the restrictions. Then three thresholds can be set by the web interface: the density threshold, the minimum amount of volume fraction (*f* ~*S*1~) and the maximum number of phases. It is possible to specify if the sample contains mainly highly crystallized phases ('Crystallization: high'), normal crystallization or nanocrystallized phases. This option will change the initial value used for the crystallites and tune the FoM differently to favour the type of phases selected. A final option is to refine only a volume expansion/contraction instead or to refine all cell parameters when the box is unchecked.
In the second section the user may specify the instrument geometry and details used in the measurement. Three different radiations are available: X-ray, neutron and electron. For X-rays, the user can select one of the conventional tubes with both *K*α~1~ and *K*α~2~; otherwise a strictly monochromatic radiation can be selected and the wavelength introduced. Three different geometries are available: Bragg--Brentano (flat sample reflection geometry), Debye--Scherrer and transmission. For Bragg--Brentano both possibilities of measurement, with θ/2θ or only 2θ, are available. Finally, we can specify the instrumental broadening, from very low as in a high-resolution synchrotron beamline to very broad as in TEM. The possible choices are just broad categories as the main purpose of the program is to perform the search--match (and rapid quantification) and not an accurate line broadening analysis. If the instrument broadening specified is sufficiently close to the real one, the program will give also a good estimation of the crystallization of the compound. For nanocrystallized samples, as the broadening is much greater than the instrument broadening (unless TEM is used), an error on the instrument broadening choice will not cause a large error in the size--strain parameters at the end of the analysis.
In the latest version of the program *MAUD* (available as a free download at <http://maud.radiographema.com>), we added the possibility to export a specific datafile in CIF format containing all the instrumental information needed by FPSM. When such a file is loaded as a pattern via the web interface, FPSM will ignore the web input from the 'Experiment details' section and the instrumental characteristics and broadening from the CIF will be used instead.
In the third section the algorithm options are available: the user can choose either only the Rietveld or a mixed DDM--Rietveld refinement at each step. This is specified with some weights for each type of refinement. The Rietveld refinement has a fixed weight of 1. Using the same weight of 1 for both the first and second derivative (DDM) ensures that the Rietveld, the first-derivative and the second-derivative methods all have the same importance in the algorithm. Entering 0 in both the first- and second-derivative fields corresponds to a Rietveld-only refinement, and using some large values instead produces mostly a DDM-like refinement. The 'Weight type' option is the same as that described in the DDM manual (Solovyov, 2004[@bb42]; <https://sites.google.com/site/ddmsuite/>). The 'smooth fit' corresponds to a zero-derivative DDM-like refinement or practically fitting the smoothed pattern with a Rietveld method. Its default value is zero, as at the moment we have not found any advantage to using it.
When the input is finished in the three sections, pressing the 'Search and quantify' button will start the computation by submitting the job to a remote server. The server will select from the database the structures to be used and launch the FPSM algorithm. Another web page will appear (as will be shown later in Fig. 3 and Table 1[](#sec4){ref-type="sec"}), presenting a table giving the phases found, their percentage, and their crystallite sizes and microstrains, together with a graph illustrating the fitting result for visual evaluation.
4.. Results and discussion {#sec4}
============================
Testing was done initially using the data set 1*h* from the round robin on quantitative analysis (Madsen *et al.*, 2001[@bb31]), which is available from the IUCr Commission on Powder Diffraction (CPD) web site (<https://www.iucr.org/resources/commissions/powder-diffraction/projects/qarr/data>). This is a simple analysis with three phases (corundum, fluorite, zincite) that was useful also for testing the accuracy of phase quantification. The tests focused on the final result and the speed of execution. In the following we will report the results using the web interface when imposing some compositional restrictions and the standalone program without restriction. This is because the web interface does not return a result when the computation becomes too long and the standalone program must be used instead. For this sample the analysis correctly identified all three phases both using no composition restriction and with restriction. The only difference between the two was the time taken to complete the analysis. On the workstation the computation time was 565 s without compositional restrictions but using the inorganic and minerals subsets of the COD, with a total of 174 064 phases. When restricting the composition to only phases containing the following atoms -- Al, Ca, F, Zn, O, Mg, Na, Si, Cl -- the computation time dropped to 19 s. If we specify only the atoms really present in the sample (Al, Ca, Zn, O, F) the total time falls below 10 s, most of which is consumed by loading the database and filtering out the structures to use for the search. We could estimate from this sample that the program is able to perform an average of 30 full automatic Rietveld refinements per second, per thread.
In Fig. 3[▸](#fig3){ref-type="fig"}, and in all the following figures, the plotting output from the web page has been visually reformatted during the editing of this paper. In Table 1[▸](#table1){ref-type="table"} we report the results of the quantitative analysis and compare them with the results of the round robin. Even though the automatic Rietveld analysis was performed using a lot of restrictions, the results are satisfactory and the absolute error was within 2%. Looking at the general results of the round robin (Le Meins *et al.*, 2003[@bb23]), our automatic analysis scored much better than the average participant result. In fact, the standard deviation on the quantitative analysis results for sample 1*h* reported by the organizers of the round robin is larger than our error. In reality we think that all the restrictions actually help the automatic analysis to avoid possible errors arising from inappropriate setting or refinement of some parameters.
The first test was a simple one, mainly used to check the speed, feasibility and accuracy of the analysis. In contrast, the second part of the test was devoted to comparing the FPSM procedure with the classical search--match routines. For this reason we used the datafiles and tests provided by the search--match round robin (Le Meins *et al.*, 2002[@bb22]). The round robin considered four samples, and we used two procedures, as in the previous test. In the first procedure, no composition restriction was applied, while in the second, we used the possible elemental composition given by the round-robin proposers to restrict the number of phases during the search. The analysis could not be completed for sample 2, because it contained one unknown crystal structure. There was no similar structure in the COD and the correct phase is only reported in the JCPDS-PDF database as a card. The structure is described as a silicon oxide quinuclidine fluoride octadecasil and we have not found it even in the Cambridge Structural Database (Groom *et al.*, 2016[@bb51]). So we will report the test results on the other three samples: two inorganics and one organic.
Sample 1. This sample analysis was performed using the standalone version of the program without compositional restrictions and the web form with restrictions. Both gave the same results, so we report here only the output of the web version. The final results, in line with the round-robin solution, are reported in Fig. 4[▸](#fig4){ref-type="fig"} and Table 2[▸](#table2){ref-type="table"}. The main discrepancy in the fit visible in the diagram arises from a small difference between the souzalite structure of the real sample and that available in the database. Souzalite, or the analogue gormanite, has a high variability in composition and structure. In the analysis reported, we limited the phases to no more than five. Otherwise, FPSM will also find and add several other minor phases, all around 1%, that are indeed possible, as the round-robin organizers acknowledge, but we preferred to omit them from the results as there is no information available on them to check for correctness.
Sample 3. This sample is an organic, and according to the notes on the round-robin results it should be composed of about 50:50% of two polymorphs, α- and β-thalidomide. However, the JCPDS-PDF only contains the α polymorph, and consequently identification with only the α form was accepted. In fact, none of the round-robin participants identified the other polymorph. In our case, we did have both the α and the β form in the COD database and FPSM found both correctly (Fig. 5[▸](#fig5){ref-type="fig"} and Table 3[▸](#table3){ref-type="table"}). But in this case, even with the compositional restrictions, we could not run the search on the web interface as we have nearly 40 000 organic phases in the database having similar composition. So we only used the standalone version for this test. FPSM took about 3300 s to find the solution on the workstation. The calculation time seems relatively high, but if we consider that we went directly from the raw pattern to the final correct phase quantification, this is reasonable in absolute terms.
Sample 4 was the easiest one, being composed of only lead-based compounds, but it showed some particularities. Again we report here only the analysis done via the web form using the restricted composition. The unrestricted test gave similar results but in a much longer time. The full analysis was completed in 40 s, giving the correct phases, but in addition several other compounds like sulfur, ice and sulfuric acid were found, showing substantial broadening. These compounds were mainly used by the refinement routine to provide a better fitting of the background modulations. Using a threshold density of 2 g cm^−3^, these spurious phases were eliminated, and the total computation time dropped down to 19 s. The results for the last case are reported in Fig. 6[▸](#fig6){ref-type="fig"} and Table 4[▸](#table4){ref-type="table"}. FPSM did not find the massicot phase that the round-robin organizers reported as present. Indeed, many participants did not find it either. So we investigated the presence of massicot further by performing a manual Rietveld refinement with the found phases and the massicot crystal structure. The massicot quantity was always refined close to zero. We notice that all the peaks assigned to the massicot are also peaks of the other phases, apart from a medium-strong peak at 63° 2θ that is in fact not present in the experimental pattern. The organizers did not use the entire range at their disposal but only up to 62° 2θ from their plot. We can safely conclude that the massicot not identified by FPSM is indeed not present in the sample. This example shows an advantage of the present procedure, because by using pattern fitting it may better evaluate the effective presence of a phase by considering the real contribution of the others.
In addition to the standard and round-robin tests we have analysed other samples, and we will report here a couple of extreme cases. We anticipated that the FPSM procedure may be ideally suited for nanomaterials, where the substantial broadening and overlap make the identification of the peaks and their positions more difficult. One example (TiO~2~ rutile nanoparticles), where we used an early version of FPSM, has already been reported in a paper dealing with TEM diffraction ring analysis (Boullay *et al.*, 2014[@bb4]). We used another TiO~2~ sample prepared in different conditions (Ceccato *et al.*, 2003[@bb5]), containing more than one polymorph. The sample has been measured in a capillary using silver radiation and an INEL (now Thermo Fisher Scientific) CPS 120° detector, to reach a large range in *Q* space, the data originally being intended for a pair distribution function analysis. For both the traditional search--match and FPSM we had to impose restrictions using the known chemical elements. It was easy to identify the anatase phase by both procedures (nearly 67 wt% of the sample), while the rutile and brookite phases were ranked very low by the traditional search--match. The results from the FPSM method are reported in Fig. 7[▸](#fig7){ref-type="fig"} and Table 5[▸](#table5){ref-type="table"}. All three polymorphs have been correctly identified, and the quantification is not far from the best fit obtained when a manual Rietveld refinement of the crystal structures was conducted with an anisotropic model for the size--strain broadening and the intensity data in the full *Q* range (up to 16 Å^−1^). FPSM on the workstation took 20 s to obtain the final result.
Since the FPSM procedure has been demonstrated to work for a nanocrystalline sample, we wondered how it would perform when we have a highly crystallized sample measured on a high-resolution instrument. In such cases, the peaks become similar to delta functions and the shape of the profile is less significant than the precise reflection positions. Small differences in cell parameters or instrument misalignments may be critical for the procedure presented here. But surprisingly, in all the tests we have done we concluded that there are no drawbacks with sharp peaks. The full-profile search--match is much quicker than in the normal or nanocrystalline cases, and the identification is much easier. As an example, we present a test made on an yttrium aluminium garnet sample for which the possible atoms were Al, Y, Ca and maybe O. The pattern was collected with Co *K*α~1~ radiation (using a Johansson Ge111 crystal monochromator), a capillary and this time an INEL 500 mm radius CPS 590 detector to obtain reasonable intensities in a relatively short time (90 min) but very sharp reflections. Fig. 8[▸](#fig8){ref-type="fig"} and Table 6[▸](#table6){ref-type="table"} report the final results obtained through the web interface. As options we selected 'high' for the crystallization of the sample and 'Synchrotron' for the instrumental broadening. The entire search and quantification was completed in 1 s using the list of atoms for the composition restriction.
We have tested the method on a few more examples to check its limits. In the first case, we wanted to verify what happens when the sample contains some phases with large grains. In such cases the intensities do not match the theoretical values, and this should affect the results. We used a sample containing a nano-grained periclase, large grains of calcite (high content) and large grains of magnetite (minor content). The sample was measured with monochromated Cu radiation with a fixed incident angle and a 120° CPS detector. No spinning or movements were applied, to emphasize the grain statistics problem. The results of the search--match using FPSM are shown in Fig. 9[▸](#fig9){ref-type="fig"} and Table 7[▸](#table7){ref-type="table"}. Calcite, albeit not perfectly matching the reflection intensities, was identified first, followed by periclase. The minor magnetite phase, having large deviations from the intensities of a perfect randomly oriented powder pattern, was not identified, no matter which options we selected in the algorithm. Instead, the qandilite phase was identified, and we found it has a similar structure to the magnetite, the only difference being possible substitution in the Fe^3+^ and Fe^2+^ sites with other atomic species like Mg^2+^, Ti^3+^ or Al^3+^. The diamond phase, which was wrongly found by the algorithm, contributes only to lowering the *R* ~wp~ by better fitting of the background and some strong reflections of calcite and magnetite with high graininess. The final quantification may be wrong owing to the carbon phase and graininess, but if we ignore the carbon phase and take qandilite as a replacement for magnetite the result can be considered a partial success. We compared this with the results obtained using the traditional search--match program *QUALX* (Altomare *et al.*, 2015[@bb1]), also using the COD database. With *QUALX*, after a first automatic peak search, we had to manually remove and add reflections to get a reliable selection. The high background, the noise, and the presence of small, sharp and broad peaks did not permit a satisfactory automatic peak identification. The elemental composition was used to restrict the search as for FPSM. *QUALX* was able to find all the phases correctly, as the intensity problems were not of concern when using software that checks only the peak positions, but except for calcite, which occupied from the top to the seventh position among all candidates in a list of identified phases, the next two structures found were qandilite and finally MgO. The correct phase, magnetite, appeared only in 17th position. We conclude that in this case FPSM gave more or less the same results as the traditional search--match, even in the presence of strong intensity variations. But the traditional search--match required more assistance and human judgement to first identify the correct peak positions and subsequently select the correct phases from the large ranking list. Its computational part was much quicker, but the entire process including human action took much longer.
Finally we tested the Rietveld search--match on a sample with 50 wt% pure silica glass, the rest of the volume being crystalline corundum. The powder pattern exhibits a large bump between 15 and 25° 2θ, and the original data are available as one of the examples provided with the *MAUD* program. In the first trial we ran FPSM with the default option (50% Rietveld and 50% DDM) to see if it would be able to identify the corundum and not be biased by the amorphous background. To our surprise there was no problem in identifying the corundum as the first phase in the ranking, but FPSM also partly refined the amorphous halo using an SiO~2~ nanocrystalline model and nothing else was found. Indeed, the fit of the amorphous bump was not good. We ran FPSM again using a weight for DDM ten times larger than that for the Rietveld refinement. Again (Fig. 10[▸](#fig10){ref-type="fig"} and Table 8[▸](#table8){ref-type="table"}) FPSM immediately and easily identified the corundum, but found a much better fit for the amorphous phase using a tridymite nanocrystalline model. In fact, Le Bail (1995[@bb21]) has shown how a nanocrystalline model can be used to refine the silica glass structure using a Rietveld-like approach. From the quantitative point of view the result is even more surprising as it gives us the correct amount in weight with an absolute error smaller than 1%, even though tridymite is not the correct structure to use in this case (Lutterotti *et al.*, 1998[@bb27]).
5.. Conclusions {#sec5}
=================
We have described a new method to perform a search--match based on full-profile fitting and the Rietveld/DDM algorithms. It works quite well for highly crystallized samples, but the method looks promising also for nanosized particles and nanomaterials. In the latter case it is generally more difficult to determine unambiguously the peak positions required by the traditional search--match routines. More testing is necessary to determine the algorithm's efficiency with nanocrystalline systems of greater complexity. The method relies heavily on brute force in computation and the availability of databases of crystal structures. In the present work we used the COD database, but potentially any other structure database can be used.
The main drawback of the method, at present, is that the computation time is much longer than usual search--match programs, unless some compositional restriction can be used. But considering that the method provides a restricted range of phases and a Rietveld quantification through pattern refinement, the running time should be compared with the total time needed to complete not only a search--match but also the further phase selection and a Rietveld pre-quantification. No human intervention is required during the entire process, and therefore the time taken should not be considered problematic. We have tested the method only on a single desktop computer, but for heavy use it would not be difficult to port the system over a cluster to improve its performance and its capabilities to detect minor phases and to solve more complex nanocrystalline problems.
The method is not always able to find minor phases, especially if the background, noise or both are high and the peaks quite broad or not well defined. Removing the background before running the method may partially solve the problem. With the aim of overcoming this problem, we have implemented, alongside the Rietveld refinement, the DDM method (Solovyov, 2004[@bb42]), which is less dependent on a good background fit. We also found it useful that the DDM method can emphasize the importance of missing peaks that normally contribute marginally to the global fitting and in consequence are generally ignored by the search--match method. This helps to improve the ability to find minor phases. The method is also quite sensitive to the quality of the pattern used, in particular the noise and the background, as they affect the Rietveld refinement. On the other hand, the instrument alignment, affecting the peak positions, and the peak overlap/broadening are less of a concern.
The last major problem we found arises from the completeness of the database used. If the JCPDS-PDF database is considered quite complete, at least for the inorganic case, including both the classical experimental reflections and the ones calculated from the crystal structure databases, the same cannot be said for other crystal structure databases, especially the free ones like the COD. Using a more complete database like the ICSD for inorganics or the CSD for organics may well improve its performance, but it has not been possible for us to test them up to now.
Further improvements could be achieved by integrating a pre-search using a traditional position-based search--match to reduce the initial number of phases. We are also planning to add the ability to deal with preferred orientations as they affect the intensities. Another area to improve is the clay minerals field, for which the incorporation of a smart modelling of the planar defects, like the single-layer model (Ufer *et al.*, 2004[@bb47]; Lutterotti *et al.*, 2010[@bb30]), may be helpful. In the future, as the speed of computers grows every day, and crystallographic databases are becoming more complete, these initial concerns should no longer be a problem.
Supplementary Material
======================
Experimental datafile in CIF format used to test FPSM, results in Figure 7. DOI: [10.1107/S160057671900342X/nb5231sup1.txt](https://doi.org/10.1107/S160057671900342X/nb5231sup1.txt)
Experimental datafile in CIF format used to test FPSM, results in Figure 8. DOI: [10.1107/S160057671900342X/nb5231sup2.txt](https://doi.org/10.1107/S160057671900342X/nb5231sup2.txt)
Experimental datafile in CIF format used to test FPSM, results in Figure 9. DOI: [10.1107/S160057671900342X/nb5231sup3.txt](https://doi.org/10.1107/S160057671900342X/nb5231sup3.txt)
The authors would like to thank the referees for their overall suggestions and revisions of the paper, which have improved its readability and usefulness.
{#fig1}
{#fig2}
{ref-type="fig"}. The 1*h* diffraction datafile is from the quantitative phase analysis round robin (Madsen *et al.*, 2001[@bb31]). The figure and the table results have been visually reformatted for this article with respect to the original ones given in the output by the web interface. The same applies for all the subsequent figures.](j-52-00587-fig3){#fig3}
![Results of the FPSM on sample 1 of the search--match round robin (Le Meins *et al.*, 2003[@bb23]). The execution time was 282 s on the Nanoair workstation (2 × 6 cores, 2.93 GHz). *R* ~wp~: 0.318.](j-52-00587-fig4){#fig4}
{#fig5}
{#fig6}
{#fig7}
{#fig8}
{#fig9}
{#fig10}
###### Results of the FPSM analysis on CPD sample 1*h* from the quantitative analysis round robin (R-R) (Le Meins *et al.*, 2003[@bb23]), compared with the R-R results
Column 4 contains the mean value from all participants in the round robin. In column 5, the mean standard deviation of the analyses performed by all participants using the CPD-supplied data is reported for comparison. In columns 6 and 7, the minimum and maximum values obtained by the participants is shown.
Phase Wt% FPSM Mean R-R Std. Dev. Min Max
---------- ------- ------- ---------- ----------- ------- -------
Corundum 35.12 37.23 35.96 2.71 30.30 43.83
Fluorite 34.69 33.94 35.21 2.40 30.33 40.30
Zincite 30.19 28.83 28.93 2.56 16.49 33.81
###### Results of the FPSM analysis on sample 1 of the search--match round robin (Le Meins *et al.*, 2003[@bb23]) (see Fig. 4[▸](#fig4){ref-type="fig"})
Phase COD ID Name Vol.% Wt% Crystallites (Å) Microstrain
-------------- ----------------- ------- ------ ------------------ -------------
9000098 Siderite 39.5 46.0 770.5 0.0009
9002213 Hydroxylapatite 34.9 32.6 1000 0.0004
1011172 Quartz low 13.1 10.2 1000 0.0004
9005624 Souzalite 12.5 11.2 1000 0.0004
###### Results of the FPSM analysis on sample 3 of the search--match round robin (see Fig. 5[▸](#fig5){ref-type="fig"})
Phase COD ID Name Vol.% Wt% Crystallites (Å) Microstrain
-------------- --------------- ------- ------ ------------------ -------------
1513335 α-Thalidomide 52.3 53.3 444 0.00275
1513336 β-Thalidomide 47.7 46.7 421 0.00104
###### Results of the FPSM analysis on sample 4 of the search--match round robin (see Fig. 6[▸](#fig6){ref-type="fig"})
Phase COD ID Name Vol.% Wt% Crystallites (Å) Microstrain
-------------- ---------------------------- ------- ------ ------------------ -------------
9012702 Litharge 7.7 8.5 2669 0.004
1513337 (Pb O)4 (Pb (S O4)) 45.4 43.9 1259 0.0015
1513337 (Pb O)3 (Pb (S O4)) (H2 O) 17.1 14.4 450 0.00085
9012697 Litharge 29.8 33.2 640 0.00087
###### Results of the FPSM analysis on the nanocrystalline titanium oxide sample (see Fig. 7[▸](#fig7){ref-type="fig"})
Phase COD ID Name Vol.% Wt% Crystallites (Å) Microstrain
-------------- ---------- ------- ------ ------------------ -------------
1010942 Anatase 65.7 64.3 64.8 0.00356
9004137 Brookite 26.4 27.6 57.9 0.00123
4102355 Rutile 7.9 8.1 152.7 0.00018
###### Results of the FPSM analysis on the garnet sample (see Fig. 8[▸](#fig8){ref-type="fig"})
Phase COD ID Name Vol.% Wt% Crystallites (Å) Microstrain
-------------- ------------ ------- ----- ------------------ -------------
4312143 Al5 O12 Y3 100 100 3930 0.00026
###### Results of the FPSM analysis on the sample with both nanocrystalline periclase and other well crystallized phases (see Fig. 9[▸](#fig9){ref-type="fig"})
Phase COD ID Name Vol.% Wt% Crystallites (Å) Microstrain
-------------- ----------- ------- ------ ------------------ -------------
9000095 Calcite 35.7 31.3 3334 0.00132
9000494 Periclase 35.6 40.7 158 0.0001
1011276 Qandilite 4.4 5.1 625 0.00075
9012235 Diamond 24.3 22.9 853 0.018
###### Results of the FPSM analysis on the amorphous sample (see Fig. 10[▸](#fig10){ref-type="fig"})
Phase COD ID Name Vol.% Wt% Crystallites (Å) Microstrain
-------------- ----------- ------- ------ ------------------ -------------
5000092 Corundum 36.9 50.8 1099.5 0.0007
9005269 Tridymite 63.1 49.2 38.1 0.1124
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#S1}
============
Rupture of vulnerable atherosclerotic plaques is the leading cause of stroke and myocardial infarction ([@B10]). These serious accidents often occur when plaques in the arteries suddenly rupture, causing thrombus and leading to the obstruction of the blood flow to the brain or the heart ([@B21]). Therefore, early and accurate prediction of individuals at high risk of plaque rupture could allow preventive, therapeutic, or surgical interventions to be taken to prevent such life-threatening events happening.
It is now well established that plaque components and morphology are the main factors in the determination of plaque vulnerability ([@B34]; [@B23], [@B24]). Plaques with a large lipid core and a thin fibrous cap are more prone to rupture, whereas plaques containing calcified tissue may tend to be more stable ([@B4]). High-resolution intravascular optical coherence tomography (OCT) imaging has shown great promise in the identification and characterization of atherosclerotic plaque components, such as fibrous cap, calcification and lipid tissue, as well as the quantification of plaque areas and volume ([@B45]; [@B31]). Histological studies have shown the ability of OCT to separate fibrous, calcified and lipid tissue from the carotid artery ([@B47]; [@B28]). Although OCT images present plaque morphological information with a relatively high resolution, it still relies on interpretation of the images by trained readers for the identification and quantitation of plaque components. Therefore, development of computational techniques is important to determine plaque components.
[@B22] published a quantitative study which demonstrated that OCT tissue characterization of atherosclerotic plaques could be conducted by measuring by the backscattering and attenuation coefficients. Their work showed that the coefficients has essential differences between fibrous, lipid and calcific plaques. The optical parameters were gradually enriched for atherosclerotic characterization in OCT quantitative studies ([@B38], [@B39]; [@B44]; [@B29]; [@B41]). However, tissue quantification using only optical parameters caused significant overlaps between different tissue types. In addition, discrepancies existed because of the different light sources and physical models. Later, [@B43] proposed a different methodology using the morphology operation for semiautomatic segmentation of calcified plaques in OCT images. Then, a series of such studies were implemented to quantify plaque components, by combining optical parameters and texture features ([@B37]), k means and texture features ([@B5]), and least square optimization strategy to estimate the depth profiles in OCT data ([@B32]). Previous studies demonstrate the feasibility of atherosclerotic plaques segmentation based on machine learning algorithm, while the segmentation results were compared with manual annotation. Recently, the prevailing convolutional neural networks (CNN) were also applied to the classification of plaque components based on OCT images and demonstrated excellent results ([@B2], [@B1]; [@B12]). However, it is still challenging to segment plaque components based on OCT imaging. Moreover, lack of *ex vivo* validation on the developed imaging processing methods is the other obstacle in this area.
This study was designed to characterize and identify the fibrous, calcified and lipid tissues based on the expert annotation using histology images. In the present study, we used optical parameters, texture features combined with relative position of pixels to analyze and delineate plaque components in OCT images. Validation algorithm was performed by comparing these results with those of corresponding histological sections of the *ex vivo* carotid plaques.
Materials and Methods {#S2}
=====================
Carotid Plaque Tissue Collection {#S2.SS1}
--------------------------------
In this study, 31 patients with a high-grade (\>70%) carotid stenosis scheduled for carotid endarterectomy (CEA) from October 2015 to December 2018 were included. The dissected specimens were used for both *ex vivo* OCT imaging and histology analysis. All participants provided a written informed consent prior to the enrollment and the study protocol was approved by the institutional ethics committee.
OCT Image Protocol and Preprocessing {#S2.SS2}
------------------------------------
OCT can acquire cross-sectional images ([Figure 1C](#F1){ref-type="fig"}) of microscopic pathology structure of arterial wall. In the study, the OCT system used for the CEA plaque tissue imaging was a commercially available C7-XR with Dragonfly^TM^ catheter (2.7 F, C7-XR, St. Jude Medical Inc., St. Paul, MN, United States). The OCT system had high image resolutions of ∼15 μm, approximately 10 times finer than conventional intravascular ultrasound. Scan parameters were set as 100 frames/s, 54,000 A-scans/s, pullback speed of 20 mm/s, pullback length of ∼54.2 mm ([Figure 1B](#F1){ref-type="fig"}). In average, the *ex vivo* specimen was imaged over 50--250 frames, depending upon the actual length of the samples.
{#F1}
Prior to OCT imaging, the specimens ([Figure 1A](#F1){ref-type="fig"}) were flushed with phosphate buffered saline (PBS) to remove blood clots, washing out excess blood to reduce possible clutter artifacts. In order to accurately co-register the OCT images with the histological and immunohistochemistry sections, at each imaging site, photograph was taken and stored to eliminate the rotary bias on slices as much as possible. In each imaging site, an automatic pullback was performed to obtain images. Cross-sectional images (cartesian coordinate images) and lossless raw images (polar coordinate images) were acquired and stored on CD-ROMs for off-line process. It is worth noting that the raw images were used as input and the cross-sectional images were used as visualization.
Lumen segmentation is the primary step for plaque image analysis. The lumen boundary is solved by the classic optimization method, dynamic programming (DP) ([@B3]). More details about the algorithm refer to literature ([@B42]).
[Figure 2](#F2){ref-type="fig"} presents a flowchart of the training, testing procedures and the final performance evaluation. For the training procedure, after the preprocessing steps including automatic guide-wire and lumen segmentation, the optical parameters (OP), the texture features and relation position (RP) features were quantified. Based on these locally extracted values, the supervised pixel-wise classification was applied to train a classifier. The histology slices were used as the gold standard for manually delineating the train and test sets, and an intermediate color-coded image map depicting the different types of tissue was given by the trained classifier. The performance of the algorithm was evaluated by classification accuracy of the trained classifier output and the manual tissue map based on histology.
{#F2}
Histopathology Analysis {#S2.SS3}
-----------------------
After imaging, each carotid plaque tissue was cut into 1 cm serial transverse segments. Each embedded segment was marked in the same order with the corresponding OCT pullback and was serially sectioned into slices of 4 μm transverse sections, with a 40 μm interval perpendicular to the longitudinal axis of the artery. Sections were mounted on the slide glass substrates and then stained with Hematoxylin and Eosin (H&E), Masson's trichrome, Oil Red O to show the structural and morphology information within the plaque sample. [Figure 3](#F3){ref-type="fig"} shows the process from the carotid tissue to the slice. The histopathology and immunostained sections were examined with an Axio Lab.A1 (Carl Zeiss, Germany), and the scanned histology images were analyzed using the Pannoramic Viewer image analysis software (Budapest, Hungary). The pathological classification of the plaque components was processed based on the modified American Heart Association (AHA)-classification ([@B36]).
{#F3}
### Histopathology and OCT Image Co-registration {#S2.SS3.SSS1}
The gross morphological features of OCT images such as lumen size and shape, wall size and shape, plaque configuration, as well as calcific plaques were used to adjust the cross-sectional orientation of the slices. Manual co-registration of the OCT images and the digitized histological slices was performed by two experts. The carotid bifurcation and the narrowest lumen (maximum stenosis) were used as the reference points of matching the histology slices and OCT images. The challenge is that the thickness of OCT imaging and histology slices was different. Every 0.1 cm plaque specimen could cover 5 OCT images in one pullback and 250 consecutive histology slices. We were able to use both ends of the plaque samples as additional registration points to co-register the OCT imaging with the histology slices.
### Ground Truth Annotation {#S2.SS3.SSS2}
The images were annotated by the experts using ImageJ software ([@B13]), according to the registered histology slices. The photomicrograph representative of each tissue type is shown in [Figure 4](#F4){ref-type="fig"}. The two expert independently assessed the plaque components by examining the slice, and outlined the corresponding regions on the OCT image. The three main plaque components were color-coded with a segmentation plugin ([@B33]), such as green for the fibrous tissue, white for the calcified tissue and pink for the lipid tissue. Considering the most important morphologic features of plaque tissue are in the superficial region, whereas it was just within the current OCT imaging capabilities. Therefore, the annotation depth was less than 1 mm from the vessel lumen into the deeper tissue when we defined the main analysis region of the three plaque components. This value is in agreement with the literatures reported in [@B16]. [Figure 5](#F5){ref-type="fig"} shows the annotation results by the histology slices.
{#F4}
{#F5}
### Reproducibility of Annotations {#S2.SS3.SSS3}
Although the expert performed the annotation, the decision to delineate various tissue map was varying degree subjective and thus prone to analyst variability. To better annotation OCT images, another analyst is necessary to test the reproducibility. To determine reproducibility of annotation results, the intraclass correlation coefficient (ICC) with 95% confidence interval (CI) was calculated to measure the level of agreement between two measurements.
Pixel-Wise Feature Extraction {#S2.SS4}
-----------------------------
### Inversion Methods for Optical Parameters {#S2.SS4.SSS1}
OCT systems could acquire multiple OCT A-lines. The intensity of A-line value of the detected OCT signal \< *I*~d~(*r*) \> can be modeled using a single scattering model according to the Lambert-Beer law ([@B20]).
\<
I
d
(
r
)
≥
I
0
T
(
r
)
s
\^
(
r
)
exp
(
\-
μ
t
(
r
)
)
I
0
=
I
′
μ
b
where *r* indicates the penetration depth, μ~t~ the total attenuation coefficient and *I*~*0*~ a scale factor, *I*′locally available intensity and μ~b~ the backscattering coefficient ([@B41]). The attenuation coefficient μ~t~(*mm*^−1^) is a result of scattering and absorption. However, only scattering can be considered because the contribution of tissue absorption is very low at the near infrared wavelengths used in OCT ([@B30]). In addition, the OCT signal is influenced by focusing effects related to the confocal properties of the catheter ([@B40]), which can be described as:
T
r
=
\[
(
r
\-
z
0
z
R
)
2
\+
1
\]
\-
1
/
2
Here *T*~r~ is the longitudinal point spread function (PSF), *z*~*0*~is the position of the beam waist and *z*~R~ is the Rayleigh length. While $\hat{s}$ in (1) represents the modulation due to the OCT depth scan response and is described as follow:
s
\^
(
r
)
=
exp
\[
\-
(
r
\-
z
C
z
W
)
2
\]
While *z*~C~ is the center of the scan, and *z*~W~ is the half width of the roll-off function ([@B46]).
To reduce the dynamic range and sensitivity to noise of the fitted signal, (1) is linearized by logarithmic transformation:
log
\[
⟨
I
d
(
r
)
⟩
\]
\-
log
\[
T
r
\]
\-
log
\[
s
\^
(
r
)
\]
=
log
(
I
0
)
\-
μ
t
(
r
)
The attenuation coefficient μ~t~ and constant parameter *log*(*I*~0~) were calculated using a linear least-square fit to the OCT A-lines for different layers and for different positions of the individual layers by an optimization process.
### Texture Features {#S2.SS4.SSS2}
Texture features were extracted from the training set and the testing set images in order to be used for the classification of the plaque tissues. Texture refers to the spatial interrelationships and arrangement of the basic elements of an image. In the study, several texture features were selected as follows:
First order statistics (FOS) textures are directly related to the gray tone distribution of the pixel intensity and ignore inter-pixel correlations. In the paper, four parameters including mean value, standard deviation, skewness and kurtosis were extracted directly from the image. More details can be found in reference ([@B9]).
Gray level co-occurrence matrix (GLCM) is a powerful statistical tool for texture analysis, which is a tabulation of how often different combinations of pixel brightness values (gray levels) occur in an image ([@B35]). In this paper, the distance was 1 and angle theta were 0°, 45°, 90°, and 135°. Seven features were computed based on the probability density functions (PDFS), including correlation, contract, dissimilarity, energy, entropy, homogeneity, maximum probability.
Neighborhood gray tone difference matrix (NGTDM) corresponds to the visual properties of the texture ([@B8]). The following five texture features were extracted form NGTDM, for a neighborhood size of 3 × 3: busyness, contrast, complexity, coarseness and texture length.
Fractal dimension (FD) is an index for characterizing the fractal patterns or sets by quantifying their complexity as a ratio of the change in detail to the change in scale ([@B35]). In this paper, the image intensities were transformed to the FD domain using the differential box-counting algorithm ([@B25]) at various different scales and then displayed for plaque tissue identification.
The optical parameters were one-dimension information extracted by nonlinear fit, the texture feature sets (the four groups) were two-dimension features based on the local neighborhood operations, the sixth set was relative position of pixels (RP) (the *x* and *y* coordinate of each pixel). This RP features were used in combination with other feature sets during the experiment because of their natural and essential characterization. [Table 1](#T1){ref-type="table"} presents the details of the feature sets.
######
Feature sets included in the study.
**Feature sets** **Fewature name**
------------------ -----------------------------------------------------------------------------------------
OP μ~t~ and *b*
FOS Mean, variance, median, skewness, kurtosis
GLCM Correlation, contract, dissimilarity, energy, entropy, homogeneity, maximum probability
NGTDM Busyness, contrast, complexity, coarseness, texture length
FD H^1^, H^2^, H^3^, H^4^
RP *x* and *y* coordinate
OP, optical parameters (
van Soest et al., 2010
), are two paramenters of fit functions. FOS, first order statistics (
Christodoulou et al., 2003b
); GLCM, gray level co-occurrence matrix (
Soh and Tsatsoulis, 1999
); NGTDM, neighborhood gray tone difference matrix (
Christodoulou et al., 2003a
); FD, fractal dimension (
Liu et al., 2003
); RP, relative position of pixels.
Random Forest Classifier {#S2.SS5}
------------------------
To handle the large training set, the random forest (RF) algorithm was selected ([@B7]). RF is an ensemble of decision trees that combine a series of weak classifiers (i.e., binary trees) to achieve an accurate classification. In addition, the randomization allows the flexibility to explore a large feature space because it only considers a subset of features in each decision tree. The tuning parameters are *Ntree* (number of trees to grow) and *Mtry* (number of variables randomly sampled at each node). Each decision tree is independently produced and each node is split by the parameter *Mtry*. By growing the forest up to another parameter *Ntree*, the algorithm creates trees that have a high variance and a low bias. RF becomes increasingly popular in similar medical image classification applications because of its computational efficiency for large training data, ability to handle multiclass classification.
Statistical Analysis {#S2.SS6}
--------------------
Annotation regions were compared on a pixel-wise basis with the results of the RF classifier. The performances of the RF classification methods were compared based on the following model accuracy measures: sensitivity, specificity, accuracy (ACC) ([@B26]). In order to avoid the correlation of the results, the training set contained 24 patients, and the testing set contained other 7 patients. The 10-folds cross-validation strategies were applied to estimate the classification performance of the method. The 50 images were randomly split into 10 subsets, each from random set of 5 images. The final statistical results for characterization of plaques component were then calculated based on each testing subset data. After the cross-validation, mean, standard deviation (SD) and median values were computed from the 10 testing set estimations of overall classification accuracy. All statistical analyses were conducted by using Matlab R2018a and related toolboxes (MathWork^®^, Natick, MA, United States): image processing^TM^ and Parallel Computing^TM^ toolboxes.
Results {#S3}
=======
Reproducibility of OCT Images Annotations {#S3.SS1}
-----------------------------------------
[Figure 6](#F6){ref-type="fig"} illustrates the reproducibility results for the two annotations of the three plaque components by the two experts. Both annotation area1 and annotation area2 were in good agreement and displayed a strong linear trend (*R*^2^ = 0.99). The fit lines had slopes of 1.02, and y-intercepts of 0.06 mm^2^ ([Figure 6B](#F6){ref-type="fig"}). A Bland-Altman statistic did not show bias together with narrow limits of agreement ([Figure 6A](#F6){ref-type="fig"}).
{#F6}
Manual vs Automatic Classification Result {#S3.SS2}
-----------------------------------------
The annotation results of the two experts show that excellent reproducibility of three plaque components based on the histology slices. Therefore, we choose the annotation result of expert 1 as the benchmark. [Figure 7](#F7){ref-type="fig"} illustrates examples of the automated classification results compared with histology, respectively, as well as manual annotation.
{#F7}
Accuracy of Carotid Plaque Classification {#S3.SS3}
-----------------------------------------
[Table 2](#T2){ref-type="table"} shows the classification accuracy, sensitivity and specificity of the three plaque components using individual feature set and the combined feature set. The accuracy for the classification of the calcified tissue was lower than those of the fibrous and the lipid tissues. In general, moderate quantitative results were obtained with each individual feature set. The classification results of the three tissue types affect each other, in other words, each type tissue may be misclassified into the other two types. Especially, the fibrous components were often misclassified as the lipid components and vice versa. The highest and lowest classification accuracy of three tissue components were the lipid tissue (83.1%) and the calcified tissue (62.0%) using combination feature, respectively.
######
Results for each feature set of OCT image compared to the histology.
**Feature sets** **Fibrous tissue** **Calcified tissue** **Lipid tissue**
------------------ ----- -------------------- ---------------------- ------------------
OP+RP % Sen 62.7 14.4 66.6
Spe 59.1 91.5 70.4
Acc 62.2 13.7 66.4
FOS+RP % Sen 78.1 41.6 76.9
Spe 77.0 89.0 87.0
Acc 77.7 36.0 78.4
GLCM+RP % Sen 77.5 66.4 81.8
Spe 92.6 88.0 87.7
Acc 76.5 67.1 83.8
NGTDM+RP % Sen 83.1 37.0 74.5
Spe 76.9 93.4 86.7
Acc 82.1 35.5 75.7
FD+RP % Sen 74.1 46.4 83.6
Spe 83.2 87.3 86.1
Acc 74.5 48.0 84.6
ALL % Sen 80.5 64.7 80.2
Spe 91.2 90.7 87.5
Acc 80.0 62.0 83.1
Results of Cross Validation {#S3.SS4}
---------------------------
The results of the 10-folds cross validation are shown in [Table 3](#T3){ref-type="table"}, which reports mean, standard deviation and median values of the estimates of each classification accuracy. Compared to the fibrous tissue, classification accuracy of calcified and lipid tissue appeared a larger range. The fuse feature set (ALL) obtained the best compromising results between the three plaque tissues.
######
Result of 10 fold cross validation analysis.
**Feature sets** **Acc**
------------------ ----------- ------------ ------------- -------------
OP+RP % Mean ± SD 62.2 ± 1.6 13.7 ± 4.4 66.4 ± 11.2
Median 61.7 14.3 66.7
FOS+RP % Mean ± SD 77.7 ± 1.4 36.0 ± 11.3 78.4 ± 8.7
Median 77.7 36.1 76.5
FD+RP % Mean ± SD 74.5 ± 1.4 48.0 ± 7.6 84.6 ± 4.6
Median 74.6 46.2 81.5
GLCM+RP % Mean ± SD 76.5 ± 2.7 67.1 ± 4.9 83.7 ± 6.8
Median 76.5 65.2 83.3
NGTDM+RP % Mean ± SD 82.2 ± 2.0 3.5 ± 10.1 75.7 ± 8.0
Median 81.4 31.0 73.9
ALL % Mean ± SD 80.0 ± 2.1 62.0 ± 8.2 83.1 ± 8.6
Median 79.6 60.6 81.8
Discussion {#S4}
==========
In this paper, we presented an algorithm for the characterization of atherosclerotic plaque components in OCT images and the validation using *ex vivo* carotid plaque tissue ([Figure 2](#F2){ref-type="fig"}).
The present study shows an important finding that the combination of the optical parameters, the texture features and relative position can improve the quantification results, and the validation procedures further facilitated image interpretation. Different methods tend to extract different features, for example, the optical parameters reflect the relationship between the light and the plaque tissue in term of the physical or chemical properties, whereas the texture features reflect the spatial interrelationship of different tissue types. All misclassified tissue types were associated with the other two tissue types (e.g., the calcified tissue was often misclassified as the fibrous and the lipid tissues). The occurrence of the misclassification is mainly due to the blurring of the boundaries of the three main components. Although the literature indicates that the calcified region has a clear boundary, it is difficult to annotate an accurate tissue map in the pixel-wise way. In addition, more feature parameters may be needed to interpret the heterogeneity of OCT tissue components and further distinguish the plaque components.
Recently, CNN have shown remarkable success in medical image processing tasks such as disease classification ([@B18]), tumor segmentation ([@B17]; [@B14]). [@B2], [@B1] used CNN as feature extraction to characterize the layers of coronary arteries and the classification rate was up to 96% of second layer media. Later, they further classified the coronary artery pathological formations (calcification, fibrosis, normal intima, macrophage, media, neovascularization) using CNN as feature extractor, random forest as classifier and majority voting as classification calculation. [@B12] architected the ResNet50-32 and DenseNet-121 network in the different concatenation points and investigated the optimal abstraction level of feature fusion of polar and Cartesian OCT images. The result showed the combined model performed with an accuracy of 91.7%, a sensitivity of 90.9%, and a specificity of 92.4% of the plaque detection in OCT pullbacks. Tissue characterization by OCT images mainly relies on segmentation which is a necessary step for treatment planning in percutaneous coronary intervention (PCI). So far, the studies on the segmentation of plaque components in OCT images using deep learning approaches are very limited. The present study used histology slices as gold standard to annotate the training and testing images, which is a valuable contribution to the interpretation of OCT images and it demonstrates the feasibility of machine learning for plaque components segmentation. This study may provide a foundation for future deep learning-based OCT images classification studies, which will provide a useful tool for the identification of vulnerable plaques and aid the risk stratification of patients with luminal stenosis in the future.
Until now, most research has shown that an atherosclerotic plaque with a lipid or necrotic core and a thin fibrous cap is associated with an increased risk of plaque rupture and thrombus formation, resulting in an acute coronary event or progression of atherosclerosis ([@B15]). Therefore, it is important to differentiate the fibrous and the necrotic tissue in order to distinguish a vulnerable and a stable plaque. The developed algorithm is able to characterize the difference between the fibrous and the lipid tissues, which can help in the identification of plaque vulnerability. In addition, fibrous cap rupture and subsequent plaque thrombosis are accompanied with a high macrophage content. Macrophages are inflammatory cells which lead the plaque destabilization by releasing proteolytic enzymes and other pro-inflammatory mediators. Moreover, macrophages tend to scatter light by large organelles ([@B11]). This leads to either a high attenuation coefficient or a high backscatter coefficient of OCT images. Therefore, a large lipid core is often accompanied by a large number of macrophages. Although some literature has shown that OCT is capable of identifying macrophages, the separation of the lipid tissue from the macrophage infiltration is still challenging. In other words, it is difficult to distinguish "poor signal" region in OCT images whether led by macrophage aggregation or infiltration or lipid tissue. In future studies, more data (histology sections and OCT images) may help to better classify the lipid tissues, inflammation regions *etc*.
Although this study is based on *ex vivo* CEA plaque samples, the OCT imaging of the various *ex vivo* plaque components is in agreement with those reported for *in vivo* studies ([@B27]; [@B6]). The effectiveness and efficiency of intravascular OCT imaging device were confirmed to be at least as good as the imaging information obtained from the *ex vivo* tissues and the OCT imaging was highly reproducible. In addition, the reproducibility of the imaging findings was obtained from arterial segments in patients and animals were identical in both *in vivo* and *ex vivo* images ([@B27]). Therefore, it is possible to extend this study to the coronary arteries.
Despite the encouraging results, some limitations still remain. First, it was difficult to co-register OCT with histology because of the inherent differences in the longitudinal resolution between OCT (200 μm) and histology (4 μm). The limitation can be overcome by undertaking continuous histological slide preparation of vessel segments. Second, the data set annotation was performed according to the plaque components and shape of histology image using manual analysis by expert image readers as the ground truth. It is well known that the manual analysis of OCT images tends to cause inter-observer variability and intra-observer variability, and thus resulting in a relatively large deviation. Despite this, it is important to note that some studies have shown the influence of intra-observer on manual image analysis was scarce ([@B19]). Finally, this study is a single-center study with a relatively small study population. Future multicenter studies or a large amount of histological data would be required to fully test the developed algorithm before it can be used in clinical applications.
Conclusion {#S5}
==========
In this study, the combination of the optical parameters and the texture features of OCT images were extracted and used for characterization of carotid atherosclerotic tissue types. The algorithm was validated against histology slices, which were the "gold standard" as the evaluation criteria. This study shows that the developed approach can provide an effective tool for OCT-based plaque vulnerability assessment. Although the statistical results still need to be further improved before the computer-aided automatic segmentation method is applied in routine clinical practice, the fundamental research filled the gap in the quantification and characterization of atherosclerotic tissue types from OCT imaging.
Data Availability Statement {#S6}
===========================
The datasets generated for this study are available on request to the corresponding author.
Ethics Statement {#S7}
================
The studies involving human participants were reviewed and approved by IEC for Clinical Research of Zhongda Hospital, Affiliated to Southeast University. The patients/participants provided their written informed consent to participate in this study.
Author Contributions {#S8}
====================
CH and ZhiL presented the concept and design of the work. CH performed lumen segmentation, the texture feature extraction and RF computations and drafted the manuscript. YH performed the optical parameters extraction. CH, JW, and YY analyzed the data. JW, ZhoL, and ZhiL provided suggestion and editing assistance. ZhiL critically revised the manuscript. All the authors approved the final version and made substantial contributions to this work.
Conflict of Interest {#conf1}
====================
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 authors gratefully acknowledge financial support from the National Natural Science Foundation of China (Grant Nos. 11772093, 11972118, and 61821002) and ARC (FT140101152, DP200103492, and DP200101970).
[^1]: Edited by: Philippe Sucosky, Wright State University, United States
[^2]: Reviewed by: Alkiviadis Tsamis, University of Leicester, United Kingdom; Dimitrios Ioannis Fotiadis, Foundation for Research and Technology - Hellas, Greece
[^3]: This article was submitted to Biomechanics, a section of the journal Frontiers in Bioengineering and Biotechnology
| {
"pile_set_name": "PubMed Central"
} |
With improvement in resolution broad use of the multi-detector row CT (MDCT), has resulted in detection of an increased number of small pulmonary nodules ([@B1]). Although a majority of small pulmonary nodules are benign ([@B2]-[@B5]), a specific diagnosis for a nodule frequently remains undetermined. Diagnostic evaluation of such small, probably benign but indeterminate pulmonary nodules frequently involves serial CT scanning, to detect growth as evidence of a possible malignancy.
Reliable estimates of nodule size are important for identifying subtle growth in small nodules, to diagnose a malignancy as early as possible. Computer-aided volumetric assessment, of small pulmonary nodules, for estimating growth may provide this reliability, and is gaining wider acceptance. However, several variables affecting the relative error, in nodule volume measurement, have been recognized; these include: nodule size, section thickness, reconstruction algorithm, radiation dose, and segmentation thresholds ([@B6]-[@B9]). The error for in vivo nodules is expected to be greater than for phantoms, this is because it includes measurement error from greater partial-volume effects, vascular geometry and motion artifacts.
Precision and reproducibility are important factors in the clinical application of an automatic three-dimensional nodule volume measurement program, and have been the focus of many published reports ([@B6], [@B7], [@B10], [@B11]). Although obtaining follow-up scans at a fixed lung volume is ideal, this is difficult to achieve in practice. As a part of our ongoing investigation of intrinsic measurement variability, and imaging variables that affect volumetric measurements of lung nodules, we evaluated how changes in lung volume affect volumetric measurements of lung nodules with a multi-row detector CT.
MATERIALS AND METHODS
=====================
Subject Selection and CT Scans
------------------------------
The data used in the present study is from CT scans obtained from subjects in an on-going Severe Asthma Research Program (SARP). We included subjects with severe persistent asthma and chronic bronchitis. The subjects included in our study with severe persistent asthma were required to have a history of physician-diagnosed asthma, and to have received asthma therapy for at least 12 months. The diagnosis of asthma was based on demonstration of reversible airway obstruction (i.e., 12% increase in FEV1), and was confirmed by airway hyperreactivity to methacholine (i.e., methacholine PC20 ≤ 8 mg/ml) ([@B12]). Subjects were classified as having severe persistent asthma using the criteria of the National Asthma Education and Prevention Program (NAEPP) and the ATS (American Thoracic Society) Workshop on Refractory Asthma ([@B13]). The diagnosis of chronic bronchitis was based on the ATS definition ([@B12]) and included having a cough and sputum on most days of the month for at least three months of the year during the previous two years, as well as a current or previous smoking history with a minimum of 20 pack years. Patients were excluded from analysis if their CT scans did not meet protocol specifications. Institutional review board approval and patient informed consent were obtained for the SARP study.
At the time of this study, CT scans from 23 subjects enrolled for the SARP were available. One experienced chest radiologist (J.M.G., 13 years of experience with chest CT) reviewed CT scans and identified the presence of lung nodules with the assistance of LungCare (Siemens, Forchheim, Germany). Of the 23 subjects, 10 (4 men and 6 women; age range, 26-55 years; mean age, 42 years) who had nodules on both inspiratory and expiratory CT scans were included in this study. This group included seven subjects with severe persistent asthma and three subjects with chronic bronchitis. Calcified nodules were included in the study.
All patients underwent full inspiratory and expiratory CT of the thorax in the supine position. Images were obtained during suspended full inspiration, and suspended relaxed (not forced) expiration.
The scans were performed using a 16-detector-row CT scanner (Somatom Sensation 16; Siemens, Germany), with 16×0.75-mm collimation, 120 kVp, 50 effective mAs, 500-msec gantry rotation time, 18 mm table feed per rotation, 1-mm slice thickness, and 0.7 mm reconstruction increment. CT images were reconstructed with a field of view large enough to cover the complete lung cross-section (262-424 mm). This corresponds to an in-plane pixel size of approximately 0.51-0.83 mm. No contrast media was administered.
Measurements
------------
All CT studies were transferred from the CT scanner to a Siemens workstation, and were analyzed using LungCare including \"Nodule enhanced viewing\" and Volume software.
Lung nodule volume was measured semiautomatically. After marking a nodule with a mouse click, a volume of interest (VOI) was defined. The volume of the nodule was computed by clicking on the nodule within the VOI. No manual postprocessing of the segmentation result was performed. The apparent diameter of each nodule was calculated from the computed nodule volume on the basis of the volume-diameter relationship of spheres. Nodules were categorized as isolated, juxtavascular, or juxtapleural nodules.
For the computation of the lung volume and whole lung attenuation, lung regions were segmented from the background. The Volume program uses a threshold-based 3D region growing technique to separate the lung regions from other tissues and structures, and selects all voxels between -500 and -980 HU ([@B14]). The threshold range was chosen to exclude nonaerated lung tissue such as vessels, major airway or hyperinflated lung tissue. Thus, the segmented lung region corresponds to the sum of the voxels whose attenuation values fall within these threshold values. The volume and mean attenuation of the lung were calculated from the total count and the mean attenuation of the voxel sums. These values were defined as lung volume and whole lung attenuation.
To compute the nodule attenuation and regional attenuation of the lung, we selected a transverse section with the largest nodule cross section area. For each nodule, a region of interest (ROI) that covered about one-half of the diameter of the nodule was placed over the nodule. The edges of the nodule were avoided to prevent partial volume averaging. A larger ROI of 2.0-2.5 cm^2^ was then placed adjacent to the nodule to measure the regional lung attenuation of the lung. Care was taken to exclude extrapulmonary tissue or large vessels. These measurements were performed by one radiologist (J.M.G.). The volume and mean attenuation of the nodule, and regional lung attenuation were measured twice after an interval of several days. The mean of the two measurements was used for statistical analysis.
Statistical Analysis
--------------------
Differences between nodule volume on inspiration and expiration were compared with a paired *t*-test.
The percent difference was defined as a percentage of the absolute difference between inspiration and expiration measurements divided by the mean of the two values. The percent difference in nodule volume, nodule attenuation, lung volume, whole lung attenuation, and regional lung attenuation were computed. The association between lung nodule volume and nodule attenuation, lung volume, whole lung attenuation, or regional lung attenuation was calculated using the Pearson correlation coefficient. Data analysis was performed repeatedly for all nodules and for nodules ≥ 3 mm. A *p* \< 0.05 indicated a statistically significant difference. The statistical analysis was performed using SPSS 10.0 (SPSS Inc. Chicago, IL).
RESULTS
=======
A total of 33 nodules were identified. Nodule volumes ranged from 4.0 to 297.5 mm^3^ (mean, 28.2 mm^3^) on inspiration CT and from 4.8 to 307.4 mm^3^ (mean, 31.5 mm^3^) on expiration CT. These measures correspond to nodule diameters ranging from 2.0 mm to 8.3 mm (mean, 3.8 mm) on inspiration CT, and from 2.1 to 8.4 mm (mean, 3.9 mm) on expiration CT. There were 10 nodules ≤ 3 mm, 21 nodules 3-5 mm, and two nodules \> 5 mm. Twenty-seven were isolated, four were juxtavascular and two were juxtapleural. The mean values for lung nodule volume, nodule attenuation, lung volume, whole lung attenuation, and regional lung attenuation are summarized in [Table 1](#T1){ref-type="table"}. Ten of 33 nodules \> 60 HU were considered to be calcified nodules.
The difference between inspiration and expiration nodule volume was significant (*p* = 0.001 for all nodules; *p* = 0.005 for nodules ≥ 3 mm). The mean percent difference in lung nodule volume was 23.1% (range, 0.5-76.9%) for all nodules and for nodules ≥ 3 mm ([Table 2](#T2){ref-type="table"}, [Fig. 1](#F1){ref-type="fig"}). The volume of the nodules studied was found to measure larger on the expiration CT than on the inspiration CT (28 out of 33 nodules; 19 out of 23 nodules ≥ 3 mm).
A statistically significant correlation between the percent difference of lung nodule volume and lung volume (R = 0.45, *p* = 0.03) or regional lung attenuation (R = 0.51, *p* = 0.01) was found for nodules ≥ 3 mm. A correlation between the percent difference of lung nodule volume and nodule attenuation or whole lung attenuation for nodules ≥ 3 mm was not statistically significant (*p* \> 0.05). A correlation between lung nodule volume and nodule attenuation, lung volume, whole lung attenuation, or regional lung attenuation for all nodules was not statistically significant (*p* \> 0.05). There were statistically significant correlations between: lung volume and whole lung attenuation (R = 0.99, *p* \< 0.0001), between lung volume and regional lung attenuation (R = 0.85, *p* \< 0.0001), and between whole lung attenuation and regional lung attenuation (R = 0.86, *p* \< 0.0001).
DISCUSSION
==========
Radiologists have the common challenge of accurate characterization of lung nodules as malignant or benign; this differentiation of nodules is critical for appropriate medical management. Recently, computer-aided diagnosis (CAD) programs have assisted in nodule detection ([@B15]-[@B19]), and have been used to assess change in size of nodules on serial CT scans ([@B6], [@B7], [@B10], [@B11], [@B20]).
The use of such programs could substantively enhance the accuracy and efficiency of detecting changes in nodules over time. Several lung nodule CAD programs have been developed and are commercially available. Of primary concern is whether a nodule, that appears to have grown on the basis of sequential volume measurements, actually has grown, and whether the difference in measurements is significant or due to error. By using estimates of the variation in percentage volume change, in stable nodules as a function of initial diameter, Kostis et al. ([@B20]) suggested that one could formulate an optimal follow-up time at which growth in a nodule can be reliably detected.
Parameters for acquiring or reconstructing images, as well as measurement methods, can affect measurement error in nodule volumetry ([@B7]-[@B9]). Even when a nodule is intrinsically unchanged between two scans, it may be measured differently due to altered technique such as scanning position. For instance, Wormanns et al. ([@B11]) reported that the volume measurement error in 151 nodules ranged from -41 to 29% (mean, 0.7±10.6%) when two consecutive CT scans were performed within 10 minutes after new positioning. Ninety-five percent of all measurements had a volume measurement error between -22.5 and 24.1%. These values were comparable to our results, allowing for the difference in definition of the percent difference between the two studies. In our study, no variation was identified due to acquisition or reconstruction parameters or scanning position.
Our results demonstrated that the nodule volume measurements were significantly affected by variation in respiratory phase. Similar findings, of variability of lung nodule volume measurements with respiration, have been previously reported ([@B21], [@B22]). Novak et al. ([@B21]) reported that change in nodule volume measurements, between different respiratory phases, was significantly larger than for nodules scanned twice at full inspiration. In another study, nodule size was measured bi-dimensionally, and a significant difference in the size of the lung nodules during inspiration and expiration was identified ([@B22]).
In our study the lung nodules measured larger on expiratory scans. The explanation for this difference in measured values is speculative. One possibility is that the difference observed reflects a real change of nodule volume during the respiratory cycle; another possibility is that this difference is caused by a change in segmentation between the nodule and the background. Because many of the nodules in our study were apparently calcified nodules, the chance to identify real change in nodule volume was low. Some collapsed or partially collapsed alveoli around the nodules on expiration, however, may have contributed to apparent increased nodule volume. Our study demonstrated that volumetric measurements of pulmonary nodules were significantly correlated with changes in lung volume. Nodule volume measurement programs employ threshold techniques used to separate background voxels ([@B11]). It is well known that lung attenuation is affected by the respiratory phase ([@B23], [@B24]). A change of background attenuation may affect the segmentation process of the lung nodule, this was observed in our study. Although we measured both the regional lung attenuation and the whole lung attenuation, only the regional lung attenuation had a significant effect on the lung nodule volume measurement. By contrast, a significant correlation between lung nodule volume and nodule attenuation, lung volume, whole lung attenuation, or regional lung attenuation was not observed for all nodules studied. Because measurement variation in smaller nodules is larger, and there are many nodules \< 3 mm in all nodule groups, the effect of background attenuation may be lost as a result of the larger variation in the nodule volume measurement itself.
There are several limitations to this study. First, the number of cases in our study was small, and the sizes of all studied nodules were relatively small. Calcified nodules were included in the study. We attempted to avoid unnecessary radiation exposure to subjects by using an existing database. Second, the differences observed in lung volumes on inspiration and expiration scans are likely far larger than those typically expected on routine inspiratory follow-up clinical CT scans. The results of our study reflect upper range variations in nodule volume due to respiratory differences, and were not related to positional variation on follow up studies.
In conclusion, volumetric measurements of pulmonary nodules were significantly affected by changes in lung volume. This variability in respiration-related measurements should be considered when a change in a pulmonary nodule consistent with growth is identified.
Supported by in part NIH NHLBI, RO1 HL 69149 and by a grant from Electronics and Telecommunications Research Institute.
{#F1}
######
The Mean Values for the Lung Nodule Volumes, Nodule Attenuations, Lung Volumes, Whole Lung Attenuations, and Regional Lung Attenuations
######
The Mean Percent Differences for the Lung Nodule Volumes, Nodule Attenuations, Lung Volumes, Whole Lung Attenuations, and Regional Lung Attenuations (n = 33)
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
In experimental settings, scores that predict risk for clinical deterioration in hospitalised patients have evolved from vital sign based Early Warning Scores (EWS) to systems that utilise the large amount of patient centred data in Electronic Patient Records (EPRs) \[[@CR1]--[@CR4]\]. These systems are not yet in widespread use, however they represent a first step towards *automatically* assimilating patient data to assist clinical decision making on high risk ward patients. Each of the current, published experimental models were derived and validated on large EPR linked databases that used Intensive Care Unit (ICU) admission as one of the outcome measures. This outcome measure is commonly used (along with death and cardiac arrest) as a surrogate for confirmed clinical deterioration.
These and other prognostic models use a variety of statistical methods but multivariate regression modelling and machine learning techniques are commonly used. These methods require patient centred 'candidate' variables (such as vital signs or laboratory results) to form the component parts of the model \[[@CR5]\]. The process of selecting model candidate variables is important, however there is no consensus on how best to do this. Numerous methods have been used for multivariate logistic regression, including expert opinion, forward and backward stepwise selection and machine learning techniques \[[@CR6]\]. A logical and often used first step is evaluating univariate associations, which enables the variables to be quantified in advance of their inclusion in the model \[[@CR7]\]. This is helpful when using EPR data where there are large number of available candidate variables \[[@CR8]\]. Regardless of the method, the goal is to include the *optimal* combination of variables that maximise predictive ability, whilst avoiding unnecessary complexity \[[@CR6]\].
In this systematic review we provide a complete summary of patient centred variables with a *univariate* association with unplanned ICU admission. By providing these data, we hope to aid the development of EPR based models for the prediction of ICU admission (and therefore clinical deterioration). We anticipate these data will enhance data-driven improvements in the care of deteriorating ward patients.
Methods {#Sec2}
=======
Search and identification of studies {#Sec3}
------------------------------------
The study protocol has been published \[[@CR9]\] and follows the Preferred Reporting of Observational Studies and Meta-Analysis (PRISMA) statement \[[@CR10]\]. An experienced medical librarian helped devise the search strategy to maximise identification of relevant studies. Studies were identified by searching Medical Literature Analysis and Retrieval System Online (MEDLINE), Excerpta Medica database (EMBASE), Cumulative Index to Nursing and Allied Health Literature (CINAHL), the Cochrane Database of Systematic Reviews and the Cochrane Central Register of controlled trials (CENTRAL). We included additional studies from the references of review articles, studies identified during screening, and from the authors' personal libraries. We restricted the search to studies published since 2000. We did not apply any language restrictions. The search design is shown in the Additional file [1](#MOESM1){ref-type="media"} (SDC-1).
Inclusion criteria {#Sec4}
------------------
Included studies evaluated hospitalised, adult patients located in either the Emergency Department (ED), the general surgical or medical wards. Patients in specialist wards (such as obstetric or psychiatric) were eligible if they were evaluated as a part of the general patient population rather than disease specific sub groups of patients. Included studies required an analysis of at least two cohorts: one cohort of patients admitted to ICU (intervention) and one not admitted to ICU (control). Variables were eligible if they were patient centred and had been evaluated across both cohorts. Studies which described a univariate, statistical relationship between a patient centred variable (e.g. heart rate) and unplanned ICU admission were included. The described variables were single entities, as opposed to composites such as risk scores.
Exclusion criteria {#Sec5}
------------------
Excluded studies did not evaluate unplanned ICU as an isolated outcome measure nor did they evaluate patients requiring ICU readmission. Additionally, studies that evaluated variables related to hospital processes or environmental risk alone (e.g. staff-to-nurse ratios), carried out multivariate analyses (without describing the univariate analyses that went into selecting variables for the model) or evaluated patient groups with a single diagnosis, were also excluded. (The studies excluded via this criterion are listed in Additional file [1](#MOESM1){ref-type="media"} (SDC-2)). Patients admitted to ICU (or not) from high acuity areas such as HDU were excluded from the review as these hospital areas are often linked to ICUs and so are not always captured as admissions. Subgroups of illness acuity, such as needing an Rapid Response System (RRS) alert or being a high triage category, were not excluded. Finally, studies not published in peer-reviewed journals and those examining patients \< 15 years old were excluded.
Study selection and data abstraction {#Sec6}
------------------------------------
Two authors (JM, TB) independently screened titles and abstracts of identified studies against the inclusion and exclusion criteria. (Fig. [1](#Fig1){ref-type="fig"}) They were not blinded to the journal titles or to the study authors or institutions. If there was disagreement or uncertainty regarding eligibility, the article was included in the next stage of screening. The full text was retrieved for all articles not excluded by the initial screening and re-assessed for eligibility as before. Disagreements about eligibility were resolved by discussion between the screening authors or a third party (a senior member of our research team, PW and DY). Two authors extracted independently data from the studies and supplementary material. Any uncertainties regarding data extraction were resolved by discussion amongst the study team. DistillerSR (Evidence Partners, Canada, Ottawa, Ontario) was used to manage the data and identify duplicate search results. All screening and data extraction forms were implemented within DistillerSR.Fig. 1Flow diagram of the included and excluded studies
Risk of Bias assessment {#Sec7}
-----------------------
Two authors (JM, TB) independently assessed the risk of bias for included studies by using an adapted version of the Newcastle-Ottawa Scale (NOS) \[[@CR11], [@CR12]\]. The NOS is a scoring system designed to assess the quality of non-randomised studies in meta-analyses and systematic reviews \[[@CR13]\]. We generated a score by assessing each study cohort for representativeness of the desired study population, the comparability of the cohorts being assessed, the size of the cohorts and correction for confounding. We adapted the NOS score to show bias in the types of studies included in this review (i.e. those showing univariate associations) whilst remaining faithful to the established NOS methodology. The details of the modified NOS scale are included in the Additional file [1](#MOESM1){ref-type="media"} (SDC-3).
Data synthesis {#Sec8}
--------------
We categorised patient-derived variables as comorbidities, demographics, laboratory tests, vital signs, diagnoses, medications and symptoms/signs, in general accordance with the categories used in the included studies. To synthesise and present the large number of variables included in the results in a logical way, we adopted the recently published method of Dettmer et al. (as adopted from Zaal et al.), who combined the quality of the studies investigating the variables in question (based on the NOS risk of bias assessment) with the number of times the variable was investigated \[[@CR14], [@CR15]\]. This semi-quantitative approach enable the assignation of a 'weight of evidence' to each variable (Table [1](#Tab1){ref-type="table"}).Table 1Grading system for strength of evidenceStrength of EvidenceCriteriaStrong≥2 high-quality studies showing positive association between the presence of a variable and the outcome\
AND\
No studies showing a negative associationModerateOne high-quality AND one lesser-quality study showing association\
AND\
No studies showing negative associationWeak\> 2 low-quality studies showing positive association\
OR\
Only one high-quality study showing positive associationNegative≥ 1 high-quality study showing negative association (inverse relationship)\
AND\
No studies showing a positive associationInconclusiveAssociations present in only one low-quality study\
OR\
No studies of any quality showing univariate association\
OR\
Presence of positive and negative associations from different articles, regardless of study quality
RESULTS {#Sec9}
-------
All included studies are shown in Table [2](#Tab2){ref-type="table"} \[[@CR16]--[@CR31]\], The database search yielded 1520 unique studies; 1462 were removed after title and abstract review; 57 underwent full text screening; sixteen studies were included in this review (Fig. [1](#Fig1){ref-type="fig"}). Summary details are shown in Table [2](#Tab2){ref-type="table"} with additional study data in the Additional file [1](#MOESM1){ref-type="media"} (SDC-4). The mean study quality score was five and the mode was seven. We graded six studies high quality \[[@CR18], [@CR20], [@CR21], [@CR28]--[@CR30]\], four moderate quality \[[@CR16], [@CR23]--[@CR25]\] and six (low quality) \[[@CR17], [@CR19], [@CR22], [@CR26], [@CR27], [@CR31]\]. The results of the bias assessment for each study are shown in Table [2](#Tab2){ref-type="table"} and the Additional file [1](#MOESM1){ref-type="media"} (SDC-5). The quality of the studies is also reflected in the weight of evidence for any particular variable.Table 2Details of included studiesRefLead AuthorPublication yearTotal number of patients in studyPatients in ICU groupCountryNo of sitesBias scores (high (HQ), medium (MQ) or low quality (LQ))\[[@CR16]\]Barfod20126279102Denmark15 - MQ\[[@CR17]\]Calzavacca20129515Australia12 - LQ\[[@CR18]\]Churpek201359,6432638USA17 - HQ\[[@CR19]\]Eick20155730366Germany14 - LQ\[[@CR20]\]Escobar2012102,4883525USA147 - HQ\[[@CR21]\]Frost2009126,8261582Australia17 - HQ\[[@CR22]\]Hong20111025201Singapore14 - LQ\[[@CR23]\]Hunziker201274,7845233USA15 - MQ\[[@CR24]\]Loekiko201370,829149Australia25 - MQ\[[@CR25]\]Schuetz20157000490Swiss, France, USA35 - MQ\[[@CR26]\]Steiner2016240793Switzerland14 - LQ\[[@CR27]\]Sudarshan201552742USA13 - LQ\[[@CR28]\]Tam200894,482672Australia17 - HQ\[[@CR29]\]Tsai20141049313Taiwan17 - HQ\[[@CR30]\]Tsai2014699214Taiwan17 - HQ\[[@CR31]\]Wunderink201221471USA13 - LQ
Quantised data were presented as independent variables. For example, arterial oxygen saturation was evaluated six times for ranges of \< 80, \< 95%, 80--89%, 90--94%, mean (%) and median (%) (in each study group) across four studies and thus included six times in the initial analysis, with each of these ranges being defined as a single variable \[[@CR16], [@CR18], [@CR20], [@CR22]\]. Likewise, 'cardiovascular disorder' was included 14 times across six studies, as either a comorbidity, diagnosis or symptom/sign \[[@CR21], [@CR26]--[@CR29]\]. We recorded the statistical relationship between variables and unplanned ICU admission as *p* values, Odds Ratios (OR), Risk Ratios (RR) \[[@CR28]\] or Incidence Rate Ratios (IRR) \[[@CR21]\] (with 95% confidence intervals) and did not assign preference.
Five studies (31%) were case control studies and 11 (69%) were cohort studies. Eight (50%) were prospective and eight (50%) were retrospective. The number of participants in each study ranged between 95 in a prospective cohort study and 126,826 in a retrospective case control study. The number of patients admitted to ICU ranged between 15 and 5233, while in the control group they ranged between 80 and 125,244 (Table [2](#Tab2){ref-type="table"}). Five studies (31%) evaluated patients in emergency departments (ED) and 11 studies (69%) evaluated patients treated on hospital wards. Of the studies examining ward patients, five evaluated patients admitted via ED, two evaluated patients who had an RRS review and four evaluated patients admitted via any source (SDC-6, Additional file [1](#MOESM1){ref-type="media"}). Escobar \[[@CR20]\] studied patients in 20 centres, Schuetz \[[@CR25]\] in three centres and Loekiko \[[@CR24]\] in two centres. The remaining 13 studies were single centre (Table [1](#Tab1){ref-type="table"}).
Across the 16 studies, 189 different patient-derived variables were assessed for univariate association with unplanned ICU admission. Of these, 53 were vital signs, 42 were comorbidities, 29 were diagnosis, 26 were demographics, 25 were laboratory results, 10 were symptoms/signs and 4 were medications. One hundred and twenty-eight variables had a statistically significant positive association, two had a negative association and 59 had no association with unplanned ICU admission. Information on effect size was described as ORs, RRs or IRRs where available and is shown in Additional file [1](#MOESM1){ref-type="media"} (SCD-7).
The semi-quantitative analysis resulted in 110 variables after repeatedly measured variables were grouped together. These are shown in Table [3](#Tab3){ref-type="table"} and Additional file [1](#MOESM1){ref-type="media"} (SDC-8). Overall there were 12 variables with a strong weight of evidence (one was negative), three with a moderate weight of evidence and 33 with a weak weight of evidence for an association with unplanned ICU admission. The remaining 62 variables showed an inconclusive weight of evidence.Table 3Patient centred variables associated with unplanned ICU admissionVariableHigh Quality +\'ve Association (ref)Moderate Quality +\'ve Association (ref)Low Quality +\'ve Association (ref)Negative association (ref)OverallCategoryHistory of congestive heart failure (cardiovascular disorder)\[[@CR21], [@CR29]\]StrongComorbiditiesHistory of diabetes (metabolic disorder)\[[@CR21], [@CR29]\]StrongComorbiditiesMale\[[@CR20], [@CR21], [@CR28]\]StrongDemographicIncreasing age\[[@CR18], [@CR20], [@CR21], [@CR28]\]\[[@CR16]\]\[[@CR27]\]StrongDemographicDiagnosis of hepatic disease (gastrointestinal disorder)\[[@CR28], [@CR30]\]StrongDiagnosisHigher heart rate (\> 111 bpm or higher mean in ICU group)\[[@CR18], [@CR20], [@CR29], [@CR30]\]\[[@CR16]\]\[[@CR27]\]StrongVital signsHigher respiratory rate (\> 20 bpm or higher in ICU group)\[[@CR18], [@CR20], [@CR30]\]\[[@CR16]\]\[[@CR27]\]StrongVital signsHigher temperature\[[@CR18], [@CR20]\]StrongVital signsLower arterial oxygen saturation (\< 94% or lower in ICU group)\[[@CR18], [@CR20]\]\[[@CR16]\]StrongVital signsLower diastolic blood pressure\[[@CR18], [@CR20]\]\[[@CR16]\]\[[@CR27]\]StrongVital signsLower systolic blood pressure\[[@CR18], [@CR20]\]StrongVital signsFemale\[[@CR18], [@CR28]\]NegativeDemographicHistory of respiratory disorder\[[@CR21]\]\[[@CR27]\]ModerateComorbiditiesUrea (higher in ICU arm)\[[@CR20]\]\[[@CR24]\]\[[@CR27]\]ModerateLaboratory testsWhite cell count (higher in ICU arm)\[[@CR20]\]\[[@CR24]\]ModerateLaboratory testsThe number in the boxes are references. Both studies from Tsai et al. \[[@CR29], [@CR30]\] come from the same patient data base. In accordance with the modified Grading System for Strength of Evidence, these two studies were only counted once (and weighted as a single high-quality study when shown together)
Variables associated with unplanned ICU admission {#Sec10}
-------------------------------------------------
Variables with a strong, moderate and negative weight of evidence for association with unplanned ICU admission are summarised in Table [3](#Tab3){ref-type="table"}.
### Comorbidities, demographics and diagnosis {#Sec11}
A history of congestive heart failure and diabetes were the only comorbidities in this group. These had a significant result in two high-quality studies \[[@CR21], [@CR29]\] of the demographics, being male \[[@CR20], [@CR21], [@CR27], [@CR28]\] and an increasing age \[[@CR16], [@CR18], [@CR20], [@CR21], [@CR27], [@CR28]\] had a strong weight of evidence for association. Four studies showed a significant difference in mean or median age between ICU and control groups (higher in the ICU group) \[[@CR18], [@CR20], [@CR21], [@CR27]\] and three studies showed a significant OR or IRR for increased age quantiles with the oldest quantile being 75+ years of age \[[@CR21], [@CR28]\]. Hepatic disease was the only diagnosis strongly associated with unplanned ICU admission \[[@CR28], [@CR30]\].
### Vital signs {#Sec12}
All six vital signs had a strong association. Heart rate was studied 12 times across seven studies, 10 times as quantiles (seven times for tachycardia and three times for bradycardia) and twice as a comparison of means \[[@CR16], [@CR18], [@CR20], [@CR22], [@CR27], [@CR29], [@CR30]\]. Seven of the tachycardia and one of the bradycardia quantiles (\< 60 beats per minute) showed a positive association**.** Two high-quality studies also found a significant difference in mean heart rate (higher in the ICU group) \[[@CR18], [@CR20]\]. Elevated respiratory rate was evaluated eight times across six studies and had a strong weight of evidence \[[@CR16], [@CR20], [@CR22], [@CR27], [@CR30]\]. Five of the six quantiles showed a significant result \[[@CR16], [@CR27], [@CR30]\] and both high-quality studies examining mean respiratory rate showed a significant difference \[[@CR18], [@CR20]\]. The only non-significant result was a respiratory rate of \> 20 breaths per minute in a low-quality study \[[@CR22]\]. Systolic blood pressure (SBP) was evaluated seven times across five studies \[[@CR16], [@CR18], [@CR20], [@CR22], [@CR27]\]. Two high-quality studies showed a significant reduction in mean blood pressure \[[@CR18], [@CR20]\] and one moderate-quality study showed a significant OR for a SBP of 80-89 mmHg versus 90 mmHg and above \[[@CR16]\]. Diastolic blood pressure (DBP) and temperature were evaluated in the same two high-quality studies, both showing significant differences in mean (decreased for SBP and increased for temperature) \[[@CR18], [@CR20]\]. Both studies had very small variations, \< 0.2^o^Cand \< 2 mmHg respectively. Arterial oxygen saturation was studied six times across four studies \[[@CR16], [@CR18], [@CR20], [@CR22]\]. Lower saturation quantiles (\< 80%, 80--89% and 90--94%) and lower mean/median saturations were shown to be significant.
Variables moderately and weakly associated with unplanned ICU admission are summarised in Table [3](#Tab3){ref-type="table"} and Additional file [1](#MOESM1){ref-type="media"} (SDC-8) respectively.
Discussion {#Sec13}
==========
Statement of findings {#Sec14}
---------------------
In this systematic review of 16 observational and cohort studies evaluating ED and ward patients, we found two comorbidities (congestive cardiac failure and diabetes), two demographics (increasing age and being male), one diagnosis (hepatic disease) and six vital signs (respiratory rate, heart rate, temperature, systolic and diastolic blood pressure and arterial oxygen saturations) with a strong univariate association with unplanned ICU admission. These findings support the consensus that abnormal vital signs have significant value when predicting unplanned ICU admission. The strength of association for a history of congestive cardiac failure and diabetes and a new diagnosis of hepatic disease may reflect the high burden of care required in this patient cohort up until the terminal phase of disease. Being older and male as a risk factor for ICU admission may reflect the general hospital population as a whole. Overall this review provides a thorough summary of the candidate variables available in EPRs (and elsewhere in the clinical record) that will assist researchers to develop and evaluate predictive models for patients at risk of unplanned ICU admission.
Clinical and research implications {#Sec15}
----------------------------------
Progressing from vital sign based, EWS systems to EPR based, risk model systems has incrementally improved performance, both in terms of correctly identified deteriorating ward patients (sensitivity) and the number of 'false alarms' generated for clinical staff (specificity and positive predictive value). These performance gains have been achieved via multivariate regression models and more recently machine learning processes \[[@CR1], [@CR32]--[@CR36]\]. Regardless of the statistical approach, candidate variables should be selected in a methodologically robust way. In the published literature, univariate filter methods, that rank the strength of the statistical association, are among the most common \[[@CR7], [@CR8]\]. It is a popular approach because the univariate analysis provides a summary of the variables most likely to enhance model performance, does not involve significant computation, is relatively simple, not time consuming and produces an easily interpretable output. It does have weaknesses however, including the potential to miss variables that have no association with the outcome when evaluated *in isolation* but have an association when evaluated together with another variable (e.g. age).
Despite their performance advantage, as yet no EPR based hospital model has achieved widespread adoption. In contrast, 75% of UK National Health Service hospitals monitor ward patients using the National Early Warning Score (NEWS) \[[@CR37], [@CR38]\]. The success of NEWS, which is a simple aggregate score that uses the univariate associations of abnormal vital signs with adverse patient outcomes, highlights the importance of interpretability and generalisability in this research and clinical domain. Advanced scoring systems that rely on complex computational processes may be difficult to interpret (and trust) for clinical staff and therefore less likely to be adopted into general use. We hope the univariate associations described will provide a convenient and intuitive reference for clinicians and researchers alike to overcome such barriers to implementation.
Strengths and limitations {#Sec16}
-------------------------
The association of the variables does not infer causality. The search strategy was thorough and in accordance with current methodological guidelines but studies may have been missed. Publication bias may have affected results. The methodology of the included studies was varied, making meta-analysis inappropriate. We excluded studies examining specific sub-populations of patients only (i.e. acute liver failure) meaning the variables summarised in this review are not applicable for risk models designed for specific disease sub-groups.
There is a lack of consensus on which outcome measures to use when assessing the performance of predictive models for clinical deterioration \[[@CR18]\]. Each of cardiac arrest, in hospital death and unplanned ICU admissions represent different populations and will, therefore, have different variable associations. We selected unplanned ICU admission as an isolated outcome measure (and excluded in hospital death and cardiac arrest) in the knowledge this would reduce the number of eligible studies and therefore potential variables for inclusion in this review. We adopted this method because we aim to advance the study of models that predict clinical deterioration, specifically in those who will most benefit from an intervention such as an ICU admission. When predicting ICU admission, some authors published the univariate relationships from within their derivation databases before including them in the multivariate analysis \[[@CR20]\]. However we are not aware of any who have based selection on associations evaluated in external databases.
We evaluated a heterogeneous study population by including ED, ward, post-Medical Emergency Team (MET) and non-post-MET patients. This was done because we wish to better understand associations with clinical deterioration, which may occur at any time-point during hospital admission. Namely, early deterioration, which may occur soon after discharge from ED to the ward, in which case patient centred ED data is important. Or late deterioration, when the patient has been on the ward for some days. Studies examining sub-populations of patients (i.e. where specifically designed predictive algorithms have the potential to be more accurate than when used in a general patient population) were excluded on the basis that as a first step, we wish to isolate variables that will contribute to a hospital wide EPR based risk score.
We have deliberately avoided describing multivariate studies because we do not wish describe the models themselves. There are multiple examples of high performing, multivariate clinical predictive models in the literature, whose variables will have quantifiable associations with unplanned ICU admission. However, it is impossible to exclude collinearity in these instances, making obsolete our objective to individually quantify these variable associations as potential "building blocks" for future models. As a consequence, multivariate analysis was excluded unless the univariate associations were described.
Conclusion {#Sec17}
==========
Having abnormal vital signs, being elderly, male, having a history of heart failure or diabetes and a diagnosis of liver failure are all strongly associated with unplanned ICU admission. This systematic review is the first to comprehensively collate the evidence on patient centred variables with univariate associations with ICU admission. These results may assist the development of predictive models for hospitalised patients at risk of needing escalations in care. There is a lack of high-quality data in this field and further work is required to isolate the patient centred variables most likely to enhance model performance when predicting unplanned ICU admission.
Additional file
===============
{#Sec18}
Additional file 1:Supplementary Digital Content. SDC-1 Systematic Review Search Design. SDC-2 Studies excluded because of a single or grouped diagnosis. SDC-3 Modified Newcastle-Ottawa Scale for assessment of study quality (adapted). SDC-4 Details (additional) of included studies. SDC-5 Bias scores. SDC-6 Patient populations of included studies. SDC-7 Patient derived variables examined for an association with unplanned ICU admission. In categories and then alphabetical order. SDC-8 Strength of evidence for individual variables (weak and inconclusive results) (DOCX 90 kb)
CI
: Chief Investigator
CTRG
: Clinical Trials & Research Governance, University of Oxford
ED
: Emergency Department
EPR
: Electronic Patient Record
GCP
: Good Clinical Practice
ICU
: Intensive Care Unit
IRR
: Incidence Rate Ratios
NEWS
: National Early Warning Score
OR
: Odds Ratio
REC
: Research Ethics Committee
RR
: Risk Ratio
RRS
: Rapid Response System
JM would like to acknowledge the University of Adelaide, Department of Acute Medicine, who are administering his PhD.
Funding {#FPar1}
=======
This publication reports independent research supported by the Health Innovation Challenge Fund (HICF-R9--524; WT-103703/Z/14/Z), a parallel funding partnership between the Department of Health and Wellcome Trust. The views expressed in this publication are those of the author(s) and not necessarily those of the Department of Health or Wellcome Trust. The Wellcome Trust and the Department of Health did not play any role in the design of the study, interpretation of data or the writing of manuscript.
Availability of data and materials {#FPar2}
==================================
The data generated and analysed during this study are publicly available on application to the corresponding author.
JM, TB, PW and DY designed the study. JM and TB undertook the methodological planning. JM led the writing. GL and DC contributed to data interpretation. All authors commented on successive drafts and approved the final manuscript.
Ethics approval and consent to participate {#FPar3}
==========================================
This study was a secondary analysis of published material and did not require ethical approval.
Consent for publication {#FPar4}
=======================
Not applicable.
Competing interests {#FPar5}
===================
The authors declare that they have no competing interests.
Publisher's Note {#FPar6}
================
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
| {
"pile_set_name": "PubMed Central"
} |
Almost all electronic (<http://www.itrs.net/>) and spintronic devices[@b1] have junctions between different types of semiconductors (for example, p--n junctions) and metals and non-metals (for example, metal/semiconductor and tunnel junctions) at the core of their functionality. The efficiency and even the actual functionality of such devices are determined by the quality of the interfaces between the constituent layers. By definition, these interfaces in junction structures are buried beneath the surface layer and hence any defects, either in the form of pinholes or impurities, are impossible to be imaged without using time-consuming and complex techniques as summarized in [Fig. 1](#f1){ref-type="fig"}. These techniques are based on microscopy, spectroscopy, scattering and reflection, and electrical methods. Microscopic techniques are the most powerful ones among them, such as cross-sectional transmission electron microscopy (TEM)[@b2], and can reveal detailed information about atomic structures at the junction interfaces. However, the additional preparation required for a cross-sectional sample involves erosion and strain-induced damage of the junction and hence features that subsequently appear in the imaging process may be due to the sample fabrication process rather than being inherent in the original device. Helium ion microscopy (HIM) has also been used to observe the topology, while the helium ion beam can mill the sample *in situ*[@b3]. HIM can therefore achieve subsurface imaging on a semiconductor structure with sub-nm resolution but it is destructive to image buried junctions. To avoid such preparation processes, electron beam-induced current has been developed to image the distributions of the conductive properties of the buried junctions[@b4]. However, this technique is limited to conductive samples with the resolution of sub-micron and is difficult to make quantitative analysis on the junction properties, such as their conductivities and the position of the defects. For a semiconductor integrated circuit, voltage alteration techniques in scanning electron microscopy (SEM) can also be used but their resolution is typically in a micron scale[@b5].
On the other hand, spectroscopic techniques can disclose the chemical composition in the vicinity of the junction interfaces and can be used in combination with other techniques, such as microscopic imaging. Secondary ion mass spectroscopy[@b6], Auger electron spectroscopy[@b7], energy dispersive X-ray spectroscopy and cathode luminescence[@b8] have nanometric resolution but they require destructive sample preparation, typically Ar-ion bombardment, to expose the desired interfaces for analysis. By extending the penetration depth by employing different probing light or X-ray, attenuated total reflection-infrared spectroscopy[@b9] or X-ray photoelectron spectroscopy) (<http://srdata.nist.gov/xps/DataDefinition.aspx>) can be used in a non-destructive way. However, these techniques superimpose all the information from the layers the beam penetrates through. These techniques have their resolution to be limited by their wavelength, that is, almost a sub-micron scale.
For junction evaluation techniques based on reflection and scattering, elipsometry has both nanometric resolution and non-destructiveness[@b10]. However, it requires an analytical model for fitting and is difficult to be correlated to transport properties. Rutherford backscattering[@b11] and X-ray reflectivity[@b12] have also been commonly used to evaluate the interfacial roughness microscopically, which again requires an analytical model for data fitting. To avoid such models, X-ray topography[@b13] and grazing-incident small-angle X-ray scattering[@b14] have recently been developed. These techniques still need more improvement, particularly in their resolution. In addition, a conventional current--voltage (*I--V*) measurement has been utilized to assess a macroscopic junction potential. For example, for a tunnel junction, Simmons\' fitting[@b15] is typically used to estimate the barrier height and thickness. Such electrical methods are important to determine the device performance but they cannot reveal microscopic transport properties.
In this study, a decelerated electron beam was used to control its penetration depth into a multilayered junction. The beam generated secondary and backscattered electrons after impacting on a junction layer under evaluation. By combining the control of the electron-beam voltage and energy filters, these electrons generated at a specified layer were collected to produce an image of conductance distributions across the buried junction. Our technique can be used as a quality assurance tool in nanoelectronic device fabrication.
Results
=======
Electron-beam deceleration technique
------------------------------------
A field-emission scanning electron microscope (JEOL, JSM-7800F Prime) was used to develop a technique for potential mapping on buried junctions spanning a wide range of resistance. Here to maintain nanometric resolution, the initial electron beam was accelerated at (*V*~spec~+*V*~acc~) keV, where the bias voltage for specimen was −*V*~spec~ and the effective acceleration voltage landed on the specimen was *V*~acc~. This ensures high-imaging resolution, ∼0.7 nm resolution for *V*~spec~=5 keV, *V*~acc~=1 keV for example. In our observation, we managed to identify defects with their diameter of a few nanometres clearly ([Fig. 2b,c](#f2){ref-type="fig"}). The beam was then decelerated in the vicinity of the specimen stage (the so-called gentle beam mode). This enabled us to control the beam penetration depth to match the vertical position of the buried junction interface under evaluation and the resulting secondary electrons (SE) and backscattered electrons (BSE) to be generated near the buried junction interfaces. In combination with an energy filter placed beneath the upper electron detector (UED) and the upper secondary electron detector (USD) (see [Fig. 2a](#f2){ref-type="fig"}), more precise selection of the SE and BSE energies was achieved as detailed in Methods.
For SEM imaging, either SE, BSE or both has been commonly used. In the specimen, incident electrons are scattered inelastically and generate SE according to the energy loss during scattering events. Typically SE can be generated in the vicinity of the surface (approximately three times of the mean free path of SE, that is, around 3--10 nm depending on the materials). Emitted SE to a vacuum can then be detected to create a SEM image. At a decelerated electron-beam energy below a few 100 eV, the incident beam is easily scattered by the materials in the specimen, allowing the interactions between the beam and materials to occur within a few nanometres of the surface. The deceleration voltage between 500 eV and 5 keV has therefore been used in this study to avoid such scattering and to ensure the detection of electrons from the buried junctions.
In a junction, sharp and clean interfaces produce reproducible interfacial resistances. Such a junction may offer interfacial resistances ranging from a few pΩ m^2^ to μΩ m^2^ depending on the combinations at the junction interfaces, such as residual resists and oxides. For example, a spin--valve interface consisting of a ferromagnetic metal/non-magnetic metal/ferromagnetic metal multilayer typically shows an interfacial resistivity of over a few nΩ m^2^. However, defects, such as pinholes and impurities, show much smaller or larger resistivities, which can become of the order of a few pΩ m^2^ or larger than a few μΩ m^2^, respectively. These differences in resistivities generate a potential difference of a few eV up to a few keV between the different interfacial conditions. To detect such a potential difference, the electron beam needs to be decelerated effectively to the keV order and below. This deceleration voltage depends on the energy levels of SE and BSE from the buried interfaces to be observed. These energy levels can be simulated by using a well-established model (Monte Carlo Silulation of electron trajectory in solids, <http://www.gel.usherbrooke.ca/casino/index.html>).
Fresh junction imaging
----------------------
For a lateral spin--valve device consisting of two 200 × 200 nm Ni~0.8~Fe~0.2~/Cu junctions, SEM images were taken at two distinct effective accelerated voltages (*V*~acc~) of 1 and 5 keV as shown in [Fig. 2b,c](#f2){ref-type="fig"}, respectively. Here the bottom 30-nm-thick Ni~0.8~Fe~0.2~ nanowires were fabricated using a combination of standard electron-beam lithography and a lift-off process-first, followed by top 70-nm-thick Cu wires (see Methods for more details). The device was then characterized by the conventional four-probe method as explained in Methods, confirming the resistance to be ∼5 nΩm^2^. As simulated in [Fig. 2d,e](#f2){ref-type="fig"}, the effective voltage controls the depth profiles of the electron beam for imaging. This means the effective voltage of 1 keV allows the electrons to only penetrate into the surface Cu layer above the interface as shown in [Fig. 2d](#f2){ref-type="fig"}, while that of 5 keV allows the electrons to penetrate through the Ni~0.8~Fe~0.2~/Cu junction ([Fig. 2e](#f2){ref-type="fig"}). This clearly indicates that the effective electron-beam acceleration voltage in the vicinity of the specimen surface is one of the key parameters to determine the depth of the measurement positions for the buried junction imaging proposed in this study. This technique is much simpler than electric-field measurements using cross-sectional TEM imaging[@b16] for example.
By comparing the two images in [Fig. 2b,c](#f2){ref-type="fig"}, one can identify defects in the Cu nanowires, which can be seen as grey regions and black dots. The brightness in a SEM image created by SE is proportional to the number of generated SE in the specimen. The SE generation depends on the effects of surface morphology, specimen edges, acceleration voltages, atomic number of the specimen materials and charging-up on the specimen surface. Here the Cu and Ni~0.8~Fe~0.2~ wires are proven to have smooth surface without showing clear contrast from [Fig. 2b](#f2){ref-type="fig"}. As seen in [Fig. 2b,c](#f2){ref-type="fig"}, the edges of the device show bright contrast as expected. The two SEM images obtained at two different acceleration voltages did not show a clear difference in their general contrast. The device we observed was conductive and therefore should not induce any charging-up effect. We can hence exclude these effects and can conclude that the difference in the contrast indicates variance in the conductance across the specimen. The regions with darker contrast may therefore represent defects in the device, which generate fewer SE than the other majority regions. In [Fig. 2b](#f2){ref-type="fig"}, some minor grey regions are seen in the Ni~0.8~Fe~0.2~ wires, indicating that the Ni~0.8~Fe~0.2~ wires have some defects near the surface. [Figure 2b](#f2){ref-type="fig"} also shows that the junction regions do not have any contrast, confirming that the Cu wire has no defects near the surface. On the other hand, [Fig. 2c](#f2){ref-type="fig"} shows some grey regions in the Cu wire, indicating that either within the Cu wire or the bottom interface of the Cu wire, that is, Cu/Ni~0.8~Fe~0.2~ and Cu/Si, has some defects. Such defects can also be seen at the junction regions. The size of the defects is measured to be between 10 and 100 nm. It should be noted that no grey regions are observed in the Ni~0.8~Fe~0.2~ wires, confirming the Ni~0.8~Fe~0.2~ wires have no defects at the Ni~0.8~Fe~0.2~/Si interface.
Another key parameter in this study is the energy filtering of SE and BSE from the specimen. For lateral spin-valve junctions, the voltage is applied at 1 and 5 keV, which controls the depth profile of imaging as discussed above. The energies of the generated SE and BSE then need to be selected to represent only the conductance difference in the buried junction. A similar study has been carried out on semiconductor p--n junctions[@b17], confirming that the conductance difference induced by chemical potentials can be detected as SEM image contrast. To characterize the buried interfacial defects, more precise control of the electron energies to be detected for imaging is required, which can be carried out by an additional energy filtering at the detector and an additional decelerator attachment to the specimen stage and the control of the total layer thickness above the junctions[@b18].
[Figure 2f,g](#f2){ref-type="fig"} shows SEM images taken at *V*~acc~=5 keV using energy filters below −500 V with the SE mode and above −500 V with the BSE mode, respectively. Since SE can be emitted typically within 3 and 10 nm deep from the surface, the SE image shows almost identical contrast as that in [Fig. 2c](#f2){ref-type="fig"}. In the BSE image in [Fig. 2g](#f2){ref-type="fig"}, the defects in the Cu wire are unambiguously observed. In particular, almost a quarter of the junction is found to be covered by defects, such as at the top-right and bottom-left corners in the left hand side junction, and the left edge in the right hand side one. These defects suggest the presence of the contaminations at the junction interfaces, since they are typically non-conductive and do not produce SE and BSE. By comparing [Fig. 2b,f,g](#f2){ref-type="fig"}, the interfacial defects are only found at the bottom Cu interface. These results prove that BSE have high spatial resolution and are highly sensitive to such defects in a buried junction.
The giant magnetoresistance (GMR) behaviour of the device is shown in [Fig. 3d](#f3){ref-type="fig"}, which confirms that spin-polarized electrons are successfully injected into the Cu wire and are efficiently detected by the Ni~0.8~Fe~0.2~ wire. This is measured under the non-local geometry at a current of 100 μA with showing the magnetoresistance ratio of 1.7% at room temperature as shown in [Fig. 3d](#f3){ref-type="fig"}. This is similar to the values reported in similar devices[@b19], assuring the quality of the lateral spin--valve junctions. The junction has then been damaged during the following measurements, showing the resistance to be increased up to a few μΩ m^2^.
Broken junction imaging
-----------------------
After the junction breakdown, we observed the buried Ni~0.8~Fe~0.2~/Cu junctions as shown in [Fig. 3a](#f3){ref-type="fig"}. Here *V*~acc~ of 2.5 keV is used, resulting the penetration of the electron beam to be down to ∼25 nm below the Cu surface ([Fig. 3b](#f3){ref-type="fig"}). The penetrated electrons generate BSE in the Cu wire within ∼10 nm from the surface as simulated in [Fig. 3c](#f3){ref-type="fig"}. [Figure 3a](#f3){ref-type="fig"} reveals that many defects are formed in the broken Cu/Ni~0.8~Fe~0.2~ junctions and a part of the Cu wire is pealed off at the edges of the junctions. This suggests that the top Cu wire may be detached from the bottom Ni~0.8~Fe~0.2~ wire after junction breakdown, which is supported by the transport measurement with showing μΩ m^2^ resistivity. The black dots observed in the junctions are defects, which may be formed by detached interfaces. Such detached interfaces contain voids, which cannot generate BSE and can be shown as dark contrast in a SEM image. Similar defects are also seen in a part of Cu wire. These regions with many defects may be the area that can be detached by further current applications as seen in the right hand side junction in [Fig. 3a](#f3){ref-type="fig"}. In summary, we can conclude that the deceleration technique has the capability to reveal defects and damage within the buried junctions in combination with standard simulations on electron scattering. Such a simple and non-destructive technique can be applied for other junctions by combining and controlling the deceleration voltage and energy filtering. This technique is expected to offer quality assurance for a wide range of electronic devices, consisting of nanoscale junctions (see GaAs/Fe and Fe/MgO/Fe junctions in [Supplementary Figs 1 and 2](#S1){ref-type="supplementary-material"}, respectively). As discussed in the [Supplementary Discussion](#S1){ref-type="supplementary-material"}, these junctions also show some contrast in their SEM images with a decelerated beam, proving the validity of the non-destructive method as shown above. This may allow further stacking and miniaturization of junctions to sustain the advancement in their density and functionality.
Methods
=======
Device fabrication
------------------
The lateral spin--valve devices were fabricated by conventional electron-beam lithography and lift-off processes on a Si substrate with a thermally oxidised layer on the surface[@b20]. Two Ni~0.8~Fe~0.2~ nanowires were designed to be 30 nm thick and 200 nm wide with different shapes at their ends (square and sharp) to induce a difference in their magnetization-reversal fields. They were patterned using electron-beam lithography (JEOL, JBX-6300FS) and were deposited using high vacuum (HV) sputtering (SP) system (Leybold, UNIVEX 350). After their lift-off, these wires were bridged by a Cu nanowire (70 nm thick and 200 nm wide) made by the same manner. Before the Cu deposition, the surface of the Ni~0.8~Fe~0.2~ wires were cleaned by Ar-ion milling for 10 s at 50 W to remove oxides and contamination. Electrical contacts to these wires were finally made by photolithography (EGV, Mask Aligner) and lift-off process after the deposition of Cr (10 nm)/Au (150 nm) layers using an e-beam evaporator (Leybold, UNIVEX 350).
Magnetotransport measurement
----------------------------
The transport properties of the lateral spin--valves were assessed by non-local magnetoresistance measurements with a dc reversal method[@b21]. Here an electrical current of 100 μA and an external magnetic field of up to ±1.2 kOe along the Ni~0.8~Fe~0.2~ nanowires were used for the measurements at room temperature.
SEM observation
---------------
In the JSM-7800F Prime SEM, two detectors, an UED and an USD, are located above the lens and two additional detectors, a retractable backscattered electron detector and a lower electron detector, are installed just above the sample space (<http://www.jeol.co.jp/en/products/detail/JSM-7800FPRIME.html>). The UED becomes sensitive to reflected electrons from the specimen with energies above 10 eV when a bias voltage (*V*) is applied to the energy filter beneath, while the USD senses those below 1 eV. These detection modes with energy deceleration in the range between +500 V and −2 keV can be achieved coincidently. In addition, the acceleration voltage for the electron beam can be controlled between 10 and 30 keV. Based on such controllability, 0.7-nm resolution is guaranteed at the effective electron-beam landing voltages of both 1 and 5 keV at the specimen by the manufacturer.
Data availability
-----------------
The data that support the findings of this study are available from the corresponding author upon request.
Additional information
======================
**How to cite this article:** Hirohata, A. *et al*. Non-destructive imaging of buried electronic interfaces using a decelerated scanning electron beam. *Nat. Commun.* 7:12701 doi: 10.1038/ncomms12701 (2016).
Supplementary Material {#S1}
======================
###### Supplementary Information
Supplementary Figures 1-2, Supplementary Discussion and Supplementary References
This work is partially funded by the Engineering and Physical Sciences Research Council (EPSRC, EP/I000933/1 and EP/M02458X/1). A.H. acknowledges financial support by the Precursory Research for Embryonic Science and Technology (PRESTO) program by the Japan Science and Technology Agency (JST). B.A. thanks financial support by JEOL UK.
**Author contributions** A.H. conceived the experiment and analysed the SEM images. Y.Y. performed the SEM observation with the controlled electron-beam deceleration. B.A.M. carried out the electrical measurements. A.J.V. fabricated the devices. All authors contributed to the interpretation of the results and writing of the manuscript.
{#f1}
{#f2}
{#f3}
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
Captive wildlife programs harbor animals in fenced areas, which can allow parasite infection and re-infection to easily occur. Pasture rotation is not feasible in these programs and therefore, re-infection is often high. Several programs utilize anthelmintics to control parasites; however, in zoological parks parasites can persist regardless of quarantine and anthelmintic treatments ([@bib11]). Small concentrations of parasites can help animals immunologically build resistance to additional infections via antibody responses, so exposures can be beneficial ([@bib21]). Heavy infections can cause health issues, such as a reduction in appetite, gut function, fecundity, decrease body condition ([@bib28]), and can even cause death. Thus, management for minimal parasite loads is important to the entire herd, as well as the crucial consideration of inter-specific interactions if *O. dammah* is housed in areas containing other species.
Understanding parasite genera present in a habitat, along with potential hosts present, is necessary in order for anthelmintic treatments to be effective. Different parasite loads can even exist in animals exposed to different husbandry techniques ([@bib11]). In addition, parasite loads can be affected by grazing area size, fecal removal from pens or enclosures, animal density, pasture management, as well as inter-species interactions. Inter-species interactions must be considered due to the fact that multiple hosts can be infected by the same parasite, allowing transmission of generalist parasites ([@bib5]). Understanding the types of parasites found within the captive managed artiodactylids is important to establish an effective anthelminthic program to target unwanted parasites, particularly those that are more pathogenic than others. The four classes of anthelmintics utilize different modes of action against nematodes and have varying effects in host species and developmental stages of nematodes ([@bib3]). Anthelmintic resistance exhibited by parasites also needs to be considered. In North America, *Haemonchus contortus* has shown resistance to three of the four classes of anthelmintics in a captive giraffe and the parasite was still present in captive impala after treatment with the benzimidazole, fenbendazole ([@bib8], [@bib24]). *Haemonchus contortus* is of economic importance in domestic ruminants due to its increased pathogenicity and thus increased host mortality in comparison to other gastrointestinal nematodes.
Surveying and identifying internal parasites of *O. dammah* is important for maintaining the overall health of the animals, and for ensuring the species' productiveness within the different captive management programs. Identification of fecal parasites have economic benefits as well. Human-animal interactions are becoming increasingly important due to emerging zoonotic diseases. Determining parasite species based on morphological characteristics can be difficult and inaccurate; for this reason we utilized polymerase chain reaction (PCR) for higher accuracy than microscopy alone. Currently, parasites documented in scimitar-horned oryx (*O. dammah*) vary regionally, and are considered similar to parasites within domestic cattle ([@bib14], [@bib30]). Several species of gastrointestinal nematodes such as *Trichostrongylus* spp., *Haemonchus* spp., *Ostertagi* spp., *Trichuris* spp., *Cooperia* spp., as well as several species of protozoan parasites such as *Isospora* spp. and *Eimeria* spp., are commonly found in cattle. The hypotheses were that concentrations of parasites would vary between *O. dammah* at two locations, and that different parasite genera and species would be found between study sites. The genera of parasites isolated from *O. dammah* was also expected to be similar to those found in domestic cattle. Interactions between age, sex, season, and pregnancy status on parasite loads from specific genera were expected to be significant. Seasonal and regional variances of parasite loads and genera present are important factors to consider in developing treatment programs as well.
2. Materials and methods {#sec2}
========================
2.1. Study sites and animal care {#sec2.1}
--------------------------------
Fossil Rim Wildlife Center (FRWC; Lat 32° 10′ 56.64″ N, Lon −97° 47′ 51.864″ W) and Kansas City Zoo (KCZ; Lat 39° 0′ 22″ N, Long 94° 31′ 41″ W) maintain their *O. dammah* herds using different management styles. FRWC, a drive-through wildlife park, maintains a free-ranging herd within an approximately 100-acre pasture, and interspecies interactions can occur with white rhinoceros (*Ceratotherium simum*), blackbuck (*Antilope cervicapra*), wildebeest (*Connochaetes taurinus*), roan antelope (*Hippotragus equinus*), 1 blesbok (*Damaliscus pygargus phillipsi*), 3 ostrich (*Struthio camelus*), and native white-tail deer (*Odocoileus virginianus*). The animals free-range year round, with shelters available. The herd is supplemented daily with hay and pelleted commercial feed, containing 14% or 17% protein, depending on the season. FRWC herd is dewormed on an as-needed basis. The climate for this area includes mild winters, and hot summers with temperatures usually above 18°C.
KCZ maintains its herd in an approximately 7-acre pasture during the day, and within indoor enclosures during the evening. The herd at KCZ can interact with addax (*Addax nasomaculatus*), springbok (*Antidorcas marsupialis*), lesser kudu (*Tragelaphus imberbis*), eland (*Taurotragus oryx*), Grant\'s zebra (*Equus quagga boehmi*), and giraffe (*Giraffa camelopardalis*). KCZ is located in Kansas City Missouri. Missouri experiences strong seasonality with cold winters that average around 0°C from December to February and hot summers. *O. dammah* are maintained predominantly indoors during cold months, due to lack of adaptation for colder weather. Intact males are housed separately, either individually or in small groups of two or three. Another intact male is occasionally kept with the group that consists of neutered males, juveniles, and females for breeding purposes. The herd is given standardized feed and alfalfa daily, and are dewormed on a six week rotational basis with Safeguard^®^ and Synanthic^®^ liquid.
2.2. Sample collection {#sec2.2}
----------------------
Before collection, appropriate animal use and care approval (UCM IACUC number 13--3232) was obtained for working with live vertebrate animals and collection of biological products from the animals. *O. dammah* feces were collected four times, once per season, from October 2013 to July 2014: October (autumn, reproductive), February (winter, lower nutrition), April (spring, rainfall), and July (summer, dry season). During the four collection times, 15--25 samples were collected per herd to ensure at least 75% of the herd was represented. Effort was made to avoid collecting duplicates from the same individual per sampling time. FRWC samples were collected after direct observation of defecation by tagged individuals, and samples at KCZ were collected from different individuals utilizing direct observations of defecation, morphological differentiation of feces, and segregation of individuals. Gloves were changed for each sample in order to avoid cross-contamination. The fecal samples were collected, labeled, placed into sterile plastic bags, and chilled between 0 and 2 h after defecation. The sample date and time were also documented on a data sheet.
2.3. Sample processing {#sec2.3}
----------------------
Sample processing began immediately after all samples were collected, and included the modified McMaster method to count for ova and/or oocysts, along with wet mounts to identify potential genera. Samples were processed within a 12--24 h time frame. A small portion of each fecal sample was preserved as homogenized feces in either 10% formalin or 70% ethanol. A flotation solution of magnesium sulfate with a specific gravity of 1.24 was made by mixing 1.5 pounds of Epsom salts into 4 cups of distilled water. Next, 56 mL of flotation solution was added to a leak-proof sterile jar with 4 g of feces and shook for 5 min. A drop of fecal solution was used to make wet mounts, which was used for identification of parasite genera based on morphological characteristics. Microphotographs were taken using a canon T1i SLR camera with an Amscope Canon SLR/D-SLR camera adaptor for microscopes.
2.4. DNA extraction {#sec2.4}
-------------------
For DNA extraction of *Eimeria* and *Isospora* genera, 300 mg of homogenized feces was placed into a round bottom cryovial containing 200 μl of lysis buffer. DNA isolation followed methods described by [@bib12], with slight modifications. Sterile steel beads were added to each suspension and vortexed using a tissue lyser for 3 min, followed by the addition of 20 μl of proteinase K, and incubated for 1 h at 56°C. An additional 200 μl of lysis buffer was added and the suspension was then incubated for an additional 10 min at 70°C.
For DNA extraction of *Trichuris* genera, 300 mg of homogenized feces was placed into a round bottom cryovial with sterile steel beads. The suspension was vortexed using a tissue lyser 5 times with 1 min pauses, which followed methods described by [@bib4].
For DNA extraction of *Strongylus* genera, 300 mg of homogenized feces was placed into a round bottom cryovial with 300 μl of lysis buffer and 25 μl of proteinase K. The suspension was incubated at 56°C for 18 h, and then the proteinase K was inactivated by incubating for 10 min at 95°C, which followed methods described by [@bib1].
All DNA was isolated using a commercially available DNA isolation kit, QIAmp DNA stool MiniKit (Qiagen, Hilden, Germany), and then quantified with a nanodrop.
2.5. Molecular analysis {#sec2.5}
-----------------------
Conventional PCR was utilized to amplify all DNA isolates using published genus-specific primers ([Table 1](#tbl1){ref-type="table"}). All amplicons were electrophoresed for 1 h at 80 V, in a 1 × Tris-Acetate-EDTA (TAE) 1% agarose gel containing 50 μg/μl of ethidium bromide, and visualized via UV transilluminator. A 1 Kb ladder in each gel served as a standard in which to compare amplicon size. The PCR mixtures for each parasite genera are discussed below and shown in [Table 2](#tbl2){ref-type="table"}.
The *Eimeria* and *Isospora* reaction mixtures followed those of [@bib19] and contained a total of 20 μl. The PCR cycles also followed [@bib19] and are shown in [Table 3](#tbl3){ref-type="table"}. Next, 10 μl of the amplicons were electrophoresed.
The *Trichuris* reaction mixtures were modified slightly from [@bib4]. A total reaction of 25 μl contained 12.5 μl of master mix, 0.125 μl of 100 μM primer solutions, 2 μl of DNA template, and 10.25 μl of sterile ddH~2~O. The PCR reactions also followed [@bib4]. Then 20 μl of the amplicons were electrophoresed.
For *Strongylus* genera, a reaction total of 24 μl was used in order to amplify the DNA, and was modified slightly from [@bib22]. The reaction cycles mostly followed [@bib22], except for an annealing temperature increase to 60 °C. Lastly, 20 μl of the amplicons were electrophoresed.
All bands were extracted from the gel using scalpels, purified, and then sent to Eurofins Genomics (Eurofins MWG Operon LLC, Alabama, USA) for sequencing. All sequences were analyzed utilizing FinchTV Version 1.4.0, 4Peaks, and identified utilizing BLASTn analysis and available sequences in the NCBI GenBank.
2.6. Statistical analysis {#sec2.6}
-------------------------
All results are presented using descriptive statistics. Samples were grouped and categorized by sampling site in three different sections. The October sampling combined eggs per gram (EPG) and oocysts per gram (OPG), and was grouped and categorized by sampling site. The samplings for February, April, and May were separated by EPG results and OPG results; then categorized by sample sites. All categories were tested for normality using the Studentized Range for Normality Test (SRNT) and all groups were not normally distributed. The October samplings were compared using a Mann-Whitney Rank Sum Test. The other samplings were separated for EPG and OPG, then grouped by location. The locations were then compared between EPG and OPG using a Mann-Whitney Rank Sum Test. The genera were then separated out, tested for normality using the Shapiro-Wilks, and pairwise comparisons using the Hold-Sidak method were used to determine differences for each genera based on month and herd. These analyses were done using SigmaPlot 12.3 (Systat Software Inc., California). Additional analysis was completed for FRWC using a generalized linear model with a Poisson distribution considering ova and oocyst were counted and discrete. This determined significance of interactions between genera of parasites, season, sex, pregnancy status, and age. Due to sampling limitations, the same information was not available from the KCZ herd; therefore, the analysis was only done for FRWC. This analysis was done in SPSS^®^ 24 (SPSS Inc., Chicago, IL).
3. Results {#sec3}
==========
3.1. EPG and OPG counts {#sec3.1}
-----------------------
Summary statistics of EPGs and OPGs for all genera combined were reported for all sampling times including herd size per season ([Table 4](#tbl4){ref-type="table"}). October samplings combined EPG and OPG counts per location. For all other seasons, the EPG and OPG are separate for each location. FECs showed a higher EPG and OPG in the FRWC herd. The overall FEC for FRWC were 42.6% higher than KCZ (P = \<0.001) in October. For the samplings from February 2014, April 2014, and July 2014, the EPG counts for FRWC were 15.9% higher than KCZ (P = \<0.001) and the OPG counts for FRWC were 15.6% higher than KCZ (P = \<0.001). For inclusive EPGs, a significant difference was seen between sites (P = 0.002), but there was no significant interaction between site and season (P = 0.103). For inclusive OPGs a significant difference between sites (P = 0.002) was seen, but no significant interaction between site and season (P = 0.051) was seen.
Animals with high levels of either EPG (\>500) or OPG (\>1000) are listed in [Table 5](#tbl5){ref-type="table"}. Calves born in the same year as sampling had high levels of OPG and one calf had high levels of both EPG and OPG. Out of 55 McMaster slides viewed during February, April, and July, 56.4% of EPG counts from FRWC were ≤200, with 27.2% between 201 and 500, and only 16.4% above 500. Out of 48 McMaster slides from KCZ during February, April, and July, 66.7% did not have ova present on McMaster slides due to close proximity of deworming, therefore, 25% were ≤200, and 8.3% were between 201 and 500. The majority of the OPG counts from FRWC were ≤500 (43.6%), with 18.2% ≤200, 14.6% ≤900, and 23.6% were \>1000. For the KCZ, 77.1% of the OPG were ≤200, and 22.9% were ≤500.
3.2. Identified genera via microscopy {#sec3.2}
-------------------------------------
Parasite genera found at each location were identified using wet mounts, and densities of eggs or oocysts present for each genera were determined for samplings occurring in February, April, and July ([Table 6](#tbl6){ref-type="table"}). Microscopy ([Fig. 1](#fig1){ref-type="fig"}) revealed the presence of several genera of nematodes including *Strongylus*, *Trichuris*, and possibly *Nematodirus*. Several protozoa were also found on wet mounts that came from the class Coccidia, in particular two from the genus *Eimeria*, and one from the genus *Isospora*. The genera of parasites found varied between locations. *Strongylus*, one species from *Eimeria*, and *Isospora*, were found at KCZ. *Strongylus*, *Trichuris*, *Nematodirus*, two species from the genus *Eimeria*, and *Isospora*, were all found at FRWC. Images of parasites seen via microscopy are shown in [Fig. 1](#fig1){ref-type="fig"}A--E. [Fig. 2](#fig2){ref-type="fig"} shows the OPG differences between the two sites.
3.3. Comparison of parasite genera and analysis of FRWC {#sec3.3}
-------------------------------------------------------
All genera of parasites were compared between sites for each season except for one species of *Eimeria*, *Trichuris*, and *Nematodirus* which all appeared only in the FRWC herd. For the species of *Eimeria* found at both locations, overall there was not a statistically significant difference between the herds. There was an overall significant difference between the herds for *Isospora* (P = 0.003, t = 3.083) and *Strongylus* (P \< 0.001, t = 4.413). There were significant differences in concentrations of parasites from specific genera between herds within certain months, as well as significant differences in the amount of parasites from specific genera between months within each herd ([Table 7](#tbl7){ref-type="table"}).
Using a general linear model with a Poisson distribution to analyze the data from FRWC, several significant interactions were found in the FRWC herd between genera of parasites, age, sex, pregnancy status, and month. Significance was indicated with a p-value \<0.05. *Eimeria* (sp. 1) had a significant interaction with season (P \< 0.0005) with the highest amount occurring in February. *Eimeria* (sp. 2) had a significant interaction with season (P \< 0.0005) as it only appeared in July, sex (P = 0.001), and age, only appearing in calves (P \< 0.0005). *Strongylus* had a significant interaction with season (P = 0.036) with the highest amount occurring in July and age with the highest amounts in young calves (P = 0.001). *Isospora* had significant interactions with all factors (P \< 0.0005) and the highest amounts occurred in July.
3.4. Genera confirmed by sequencing {#sec3.4}
-----------------------------------
October samples were excluded due to the combined OPG and EPG counts. Out of the remaining 55 samples from FRWC, DNA was isolated from 9 samples. These samples were tested for all parasites observed through microscopy and 21 successful amplifications were achieved. Out of 48 samples from KCZ, DNA was isolated from 4 samples with 10 successful amplifications. Several bright electrophoresed PCR bands were seen in the 1% TAE agarose gels. In order to determine which band identified the parasite the primers were designed for, several bands were cut out using a scalpel and sent in for DNA sequencing. The results suggested the other bands as *Camelostrongylus mentulatus* and *Haemonchus contortus*.
Sequencing suggested the presence of *Strongylus vulgaris* (FRWC, KCZ; 96% homology, E-value 0.46), *Angiostrongylus cantonensis* (FRWC; 88%, E-value 0.33), *Camelostrongylus mentulatus* (KCZ; 99%, E-value 9e-111), *Haemonchus contortus* (KCZ; 98%, E-value 4e-159), *Trichuris (T. leporis, T. ovis,* and *T. discolor*) (FRWC; 100%, E-value 0), *Isospora gryphoni* (FRWC, KCZ; 98%, E-value 2e-137), and *Eimeria zuernii* (FRWC and KCZ; 99%, E-value 4e-129), and *Eimeria bovis* (FRWC; 97%, E-value 1e-129). Attempts to isolate and amplify DNA from samples that contained *Nematodirus* spp. were unsuccessful. Representative agarose gel results are shown in [Fig. 3](#fig3){ref-type="fig"}, [Fig. 4](#fig4){ref-type="fig"}. It should be noted that a BLAST detects similarities and is therefore not a definitive match.
4. Discussion {#sec4}
=============
Parasite loads and genera did differ between the management styles and although some species appeared in both herds, three species were specific to FRWC. This could be indicative of specific inter-species interactions, type of grass or food supplied, how long animals were allowed to graze and how short the grass was ([@bib17]), removal of feces from the grazing area and enclosures, a drier versus humid climate, a longer parasite season due to milder winters in Texas, and even physiological differences between the herds, such as the number of pregnant females (11 at FRWC and 3 at KCZ). FRWC manages *O. dammah* in as natural of a setting as possible. In order to accomplish this, animals are able to forage free-range within a 100-acre pasture, which reduces the high possibilities of re-infestation of intestinal parasites typically associated with close confinement in grazing areas. Animals are monitored and administered a dewormer only if monitoring indicates high levels of gastrointestinal parasites. The FRWC herd had higher EPG and OPG levels in comparison to KCZ; however, in comparing the FEC/OPG counts to wild ruminants in North America, both EPC and OPG counts were much lower than those found by [@bib13]. The highest OPG levels occurred in July and were associated with calves; the youngest calf sampled had the highest concentration of oocysts present in the feces. [@bib18] determined that coccidiosis is common in calves born in early April through June, and this study utilized several calves born from early April through June. The highest OPG levels contained high levels of *Eimeria* spp., which is also typically the most prevalent endoparasite in domestic calves ([@bib16]). High numbers of oocysts are expected in the feces of young calves, with a decrease typically beginning around 7 weeks of age ([@bib16]). This pattern was observed at FRWC during our study. The reduction of protozoa as calves mature indicates developing immunity ([@bib3], [@bib16]).
The management style at KCZ requires periodic use of dewormers because animals are housed indoors during evenings and during cold weather, resulting in a higher likelihood an animal could be either infected or re-infected. The importance of utilizing appropriate dewormers to the situation was observed in three intact males housed together in an indoor enclosure. Dewormer was placed in food for all three animals to share, but was primarily consumed by the dominant male. Not surprising, the dominant male had very low EPG and OPG levels of 400 combined, but the most subordinate male demonstrated a combined total EPG and OPG count of 2500, made up predominantly of strongyles. This indicates the importance of monitoring and dewormers. These findings support the hypothesis that the concentration of parasitic eggs and oocysts found in *O. dammah* feces would vary between study sites due to different management styles. Different genera of parasites were found between study sites; in addition, several species of parasites within the same genera were seen between sites as well supporting our second hypothesis.
Although parasite loads were substantially different between herds, both locations managed herds within or below EPG levels seen in domestic animals, or animals managed in zoological parks. Only a few individuals showed EPG levels above published guidelines at different times throughout the sampling period, which has been previously seen by [@bib16], [@bib6], and [@bib20]. The majority of the EPG counts, including ova from *Strongylus* spp., *Trichuris* spp., and *Nematodirus* spp., from FRWC were ≤500 (77%), and only 16.5% were above 500. For KCZ, 67% of samples were devoid of ova on McMaster slides due to close proximity of deworming, and 33% of the samples were ≤500. The majority of OPG counts from FRWC samples were ≤500 (62%) and 23.5% were \>1000. Counts included both species of *Eimeria* spp. and *Isospora* spp. For the KCZ samples, all OPG levels were ≤500. Individuals demonstrating high EPG or OPG counts were a subordinate male, pregnant females, or young calves.
The most commonly found helminth was the genus *Strongylus*. This supports [@bib11] previous study of 100% of *O. dammah* samples showed strongyle type parasites. Analysis indicated a significant difference in *Strongylus* levels between herds, with FRWC having higher levels. [@bib5] found that within an individual habitat, an increase in bovid species increased the individual mean strongyle numbers significantly. This could explain one reason why increased numbers at FRWC were observed, as *O. dammah* are involved in higher numbers of inter-species interactions than those at the KCZ. That study also found strongyle abundance in captively managed African bovids exceeded 500 EPG, with some individuals exceeding 1000 EPG ([@bib5]). [@bib5] also found *Coccidia* ranges from 7 to 7846 EPG, possibly due to competition with other parasites. Our study found similar results in regards to higher numbers of both strongyles and coccidia.
*Isospora* as a genus is commonly found in *O. dammah* and domestic bovines. The match to *Isospora gyrphoni* was unexpected as it is commonly recovered from goldfinches ([@bib25]). Several possibilities exist for this match, including a species of *Isospora* that has not been previously identified or accidental ingestion of ova. Hay is often utilized as a perch or for nests of various birds. The possibility of the animals acquiring the parasite from the food source should be investigated further, as the likelihood of acquiring the parasite from foraging on grass alone seems unlikely. Significant interactions existed for season, sex, pregnancy status, age, and *Isospora* in the FRWC herd. *Isospora* was detected in all but one sample throughout the study at FRWC, and counts varied tremendously so these significant interactions were expected. Although we were not able to analyze the herd at KCZ, raw data shows only 11 samples that did not identify *Isospora* through microscopy. *Isospora* spp. as with all species of Coccidia are treated with sulfas, not anthelmintics. The increased treatments at KCZ did not eliminate Coccidia; however, counts were much lower in comparison to FRWC which was indicated in the analysis with a significant difference between the herds.
Upon initial trials of identifying DNA bands amplified via PCR, one occurrence each of *Haemonchus contortus* and *Camelostrongylus mentulatus* were identified at KCZ. We did not test any further bands for ID but the presence of these two species is interesting. For example, *Camelostrongylus mentulatus* is typically seen in sheep and goat, but has been documented in giraffes ([@bib7], [@bib29]), and roe deer ([@bib26]). For the KCZ herd, the *O. dammah* are allowed to interact with giraffe, but without a parasite profile database of all animals, it would be difficult to make a positive correlation to the said interaction. *Haemonchus contortus* is widely known for causing economic losses in sheep and goats. With widespread anthelmintic resistance, monitoring levels of this parasite could become more important to ensure *O. dammah* survival and fecundity.
5. Conclusion {#sec5}
=============
With the current progress towards reintroducing the *O. dammah* to the wild, this study is important for the captive management programs that currently house these oryx to ensure healthy maintenance of the species. Understanding the variability in parasites within the species can help increase the success of anthelminthic programs, as well as reduce the introduction of non-native parasites in other regions. Our study emphasizes the need for facilities to ensure proper anthelminthic treatment for their herd due to their unique management style. Currently *O. dammah* are managed very similar to cattle and are typically treated for parasites typical of cattle. Typical anthelmintics used in zoological parks are approved for livestock ([@bib15]). Although this is an efficient management style, it is important to know that parasites outside of the range of the typical ones found in cattle can exist in *O. dammah*. A specialized anthelminthic treatment can be established once the generalist, and specialist, parasites found within a species have been determined. This study also showed the variance in genera of parasites present in *O. dammah* in different regions, which is necessary to understand for animal transportation to different regions and facilities in order to prevent the introduction of non-native parasites. These parasites are commonly transmitted to other animal species in captivity, and thus the scimitar-horned oryx can serve as a reservoir of these parasites.
Interspecies interactions for potential parasite accrual were different at each facility. The additional parasites found incidentally confirm parasites are shared between giraffe or camel and *O. dammah*, but we do not know if the *O. dammah* had acquired the parasite before being housed at KCZ, just as the *Isospora* species discovered could have been an interaction at a different park or zoo. The findings support different parasite management systems for *O. dammah* depending on the type of facility they are housed in. Over time, a parasite database would provide valuable parasite and host relationship information. Future studies should include a more sensitive PCR method, such as qPCR, in order to detect species of low prevalence and discovering co-infections that may exists ([@bib23]). Utilizing a more sensitive method can also detect species that may not be identified in GenBank, and determine a more precise concentration of parasites.
Funding {#sec6}
=======
This work was supported by the University of Central Missouri through the Willard North grant, CDPWN1.
Conflict of interest {#sec7}
====================
The authors declare no conflict of interest.
There is no financial or personal interest that could affect the outcome of this manuscript.
This manuscript has been read and approved by all authors listed. Individuals who contributed to the work are listed as authors and there were no other individuals who satisfied the criteria for authorship. We have followed all ethical protocols for our institution to carry out this research.
It is understood that the corresponding author listed is the sole contact for the editorial process. The corresponding author will keep open and direct communication with all authors in regards to any aspect of the editorial process. All information given for the corresponding author is up to date and current.
Appendix A. Supplementary data {#appsec1}
==============================
The following is the supplementary data related to this article:Online materialOnline material
We thank Dr. Kirk Suedemeyer, Adam Eyres, and Holly Haefele for help coordinating the project, insight into the project, and use of their respective facilities. We thank Robin Childers-Dean for assistance with sampling and processing.
Supplementary data related to this article can be found at [http://dx.doi.org/10.1016/j.ijppaw.2016.11.001](10.1016/j.ijppaw.2016.11.001){#intref0010}.
{#fig1}
{#fig2}
{#fig3}
{#fig4}
######
Genus-specific primer sets designed for conventional polymerase chain reaction to detect species of parasites in scimitar-horned oryx between environments.
Table 1
Genus Forward (5′-3′) Reverse (5′-3′) Size
--------------------------------------------- ----------------------- ----------------------- --------
*Strongylus* GTTGTGACGGATAACGGAGAA TGCACCACCAACCACTAAA 897 bp
*Trichuris* GATGGCCTCGGTGCTATTT GCCAGAGTCTCGTTCGTTATC 525 bp
*Eimeria*[a](#tbl1fna){ref-type="table-fn"} ATTTCGCTGCGTCCTTCA CCTGCATACTCCTTCAAA 307bp
*Isospora* TCCTGCACCTCATGAGAAATC GACACGCAAAGTCCCTCTAA 330bp
*Eimeria* primers were obtained from [@bib9].
######
PCR recipes used to identify fecal parasites in Scimitar-horned oryx. Primer quantities listed were for each primer, the forward (F) and reverse (R), added. Quantities are in microliters (μl).
Table 2
PCR ingredient *Eimeria*[a](#tbl2fna){ref-type="table-fn"} *Trichuris*[b](#tbl2fnb){ref-type="table-fn"} *Isospora*[c](#tbl2fnc){ref-type="table-fn"} *Strongylus*[d](#tbl2fnd){ref-type="table-fn"}
-------------------- --------------------------------------------- ----------------------------------------------- ---------------------------------------------- ------------------------------------------------
Mastermix 10 12.5 10 12.5
DNA 1 2 1 2
Primer (each F &R) 0.5 0.125 0.5 1.2
MgCl2 -- -- -- 2
Sterile ddH2O 8 10.25 8 8.1
([@bib9]).
([@bib4]).
([@bib27]).
([@bib2]).
######
Thermocycler parameters utilized to amplify fecal DNA and identify parasites.
Table 3
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Thermocycler\ *Eimeria*[a](#tbl3fna){ref-type="table-fn"} *Trichuria*[b](#tbl3fnb){ref-type="table-fn"} *Isospora*[c](#tbl3fnc){ref-type="table-fn"} *Strongylus*[d](#tbl3fnd){ref-type="table-fn"}
step
------------------------ --------------------------------------------- ----------------------------------------------- ---------------------------------------------- ------------------------------------------------
1 94 °C for 30 s 98 °C for 30 s 94 °C for 30 s 95 °C for 10 m
2 95 °C for 10 s 98 °C for 10 s 95 °C for 10 s 95 °C for 45s
3 50 °C for 30 s 55 °C for 30 s 55 °C for 30 s 60 °C for 45 s
4 72 °C for 30 s 72 °C for 30 s 72 °C for 30 s 72 °C for 45 s
5 72 °C for 2 m -- 72 °C for 2 m --
\# cycles (steps 2--4) 40 40 35 40
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
([@bib19]).
([@bib4]).
([@bib27]).
([@bib22]).
######
Summary statistics for fecal egg counts of gastrointestinal parasites reported as eggs per gram and oocysts per gram in scimitar-horned oryx between two study sites, Fossil Rim Wildlife Center and Kansas City Zoo. Sample size, N, is listed next to the total number of individuals in the herd during the sampling time.
Table 4
Sample N Range Median 25% Quartile 75% Quartile
------------------------- --------- ------- -------- -------------- --------------
FRWC -- Oct (EPG + OPG) 19 (25) 2600 575 412.5 700
KCZ -- Oct (EPG + OPG) 16 (24) 2300 1600 1100 2000
FRWC -- Feb (EPG) 17 (25) 850 150 50 325
FRWC -- Feb (OPG) 17 (25) 1800 450 300 750
KCZ -- Feb (EPG) 16 (24) 100 0 0 25
KCZ -- Feb (OPG) 16 (24) 450 150 62.5 237.5
FRWC -- Apr (EPG) 17 (28) 550 150 50 250
FRWC -- Apr (OPG) 17 (28) 1300 300 200 775
KCZ -- Apr (EPG) 16 (22) 450 0 0 1375
KCZ -- Apr (OPG) 16 (22) 350 200 100 250
FRWC -- July (EPG) 21 (28) 4100 350 175 600
FRWC -- July (OPG) 21 (28) 9550 450 275 2000
KCZ -- July (EPG) 16 (22) 350 0 0 875
KCZ -- July (OPG) 16 (22) 450 100 50 150
######
Scimitar-horned oryx with high levels of eggs per gram (\>500) of gastrointestinal parasites in feces or oocysts per gram (\>1000) of feces sampled from Fossil Rim Wildlife Center and Kansas City Zoo in October 2013, February 2014, April 2014, and July 2014 with coordinating medians and ranges of the herd from the same sampling.
Table 5
Sample EPG/OPG Median Range
--------------------------------- ------------------ -------- -------
Subordinate Male (KCZ) -- Oct 2650 (EPG + OPG) 1600 2300
Pregnant Female (FRWC) -- Oct 3250 (EPG + OPG) 1600 2600
Pregnant Female (FRWC) -- Oct 2200 (EPG + OPG) 1600 2600
Pregnant Female (FRWC) -- Feb 2000 (OPG) 450 1800
Pregnant Female (FRWC) -- April 1300 (OPG) 300 1300
6 week calf (FRWC) -- July 9700 (OPG) 450 9600
8 week calf (FRWC) -- July 6050 (OPG) 450 9600
12 week calf (FRWC) -- July 1500 (OPG) 450 9600
12 week calf (FRWC) -- July 1650 (OPG) 450 9600
12 week calf\* (FRWC) -- July 7350 (OPG) 450 9600
12 week calf\* (FRWC) -- July 4100 (EPG) 350 4100
13 week calf (FRWC) -- July 2850 (OPG) 450 9600
16 week calf (FRWC) -- July 2350 (OPG) 450 9600
^b^\*Same calf.
######
Genera of parasites found in the feces of scimitar-horned oryx from Fossil Rim Wildlife Center and Kansas City Zoo in February 2014, April 2014, and July 2014, reported as range (median).
Table 6
Genus Location Feb 2014 April 2014 July 2014
------------------- ---------- ------------ ------------ ------------
*Strongylus* FRWC 850 (150) 500 (0) 650 (200)
*Strongylus* KCZ 100 (0) 450 (0) 350 (0)
*Nematodirus* FRWC 0 450 (0) 50 (0)
*Trichuris* FRWC 300 (0) 50 (0) 500 (0)
*Eimeria* (sp. 1) FRWC 650 (100) 150 (50) 250 (0)
*Eimeria* (sp. 1) KCZ 250 (100) 100 (25) 100 (0)
*Eimeria* (sp. 2) FRWC 0 0 3900 (0)
*Isospora* FRWC 1700 (300) 1250 (250) 9600 (450)
*Isospora* KCZ 150 (50) 350 (150) 450 (50)
######
Significant differences in number of parasites from specific genera between months for certain herd. Significant differences in number of parasites from specific genera between herds for certain months. The alpha level was 0.05 for all.
Table 7
Genus Location(s) Month(s) P-value t
------------------- -------------- ---------------- --------- -------
*Eimeria* (sp. 1) FRWC Feb vs. April \<0.001 4.263
*Eimeria* (sp. 1) FRWC Feb vs. July \<0.001 3.804
*Eimeria* (sp. 1) KCZ Feb vs. April 0.010 2.856
*Eimeria* (sp. 1) KCZ Feb vs. July 0.016 2.856
*Isospora* FRWC July vs. Feb 0.006 3.025
*Isospora* FRWC July vs. April 0.006 3.164
*Isospora* FRWC vs KCZ July \<0.001 3.890
*Strongyles* FRWC vs. KCZ February \<0.001 3.398
*Strongyles* FRWC vs. KCZ July 0.002 3.218
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
The usage of implantable central venous ports has increased year by year because of progress in chemotherapy, leading to increased numbers of patients who receive long-term treatment, and because repeated administration of continuous intravenous infusions make it difficult to obtain peripheral vascular access [@bib1], [@bib2], [@bib3], [@bib4], [@bib5], [@bib6], [@bib7], [@bib8], [@bib9]. Home parenteral nutrition is now readily initiated in patients who require long-term central venous nutrition because of difficulty in oral or enteral nutrition, such as those with terminal cancer or short bowel syndrome, thereby eliminating the need for long-term hospitalization and allowing early return to work. In addition, the risk of extravasation has decreased even in patients in whom peripheral venous access is difficult to obtain, thereby allowing drugs to be safely and reliably administered intravenously [@bib10], [@bib11]. PowerPort is a multifunctional central venous access port that can be used to obtain blood samples, perform imaging examinations such as contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI), and administer anticancer therapy, high calorie infusion solutions, and blood transfusions. The aim of our study was to determine whether PowerPort is understood and used safely in our hospital. We studied patients in whom a new multifunctional central venous port was used and clarified the usefulness of this port and risk factors for postoperative complications.
2. Material and methods {#sec2}
=======================
The study group comprised 132 patients in whom PowerPort^Ⓡ^ implantable ports were surgically placed in our hospital from March 2014 through December 2015 ([Table 1](#tbl1){ref-type="table"}). There were 61 men (46%) and 71 women (54%), with a mean age of 64.5 years (range, 20 to 90) and a mean body mass index of 21 kg/m^2^ (range, 11 to 35). The ports were placed to provide treatment in 90 patients (68%) and nutrition in 42 (32%). The underlying diseases were digestive tract disorders in 70 patients (53%), medical disorders in 26 (20%), gynecological disorders in 17 (13%), mammary disorders in 13 (10%), urological disorders in 4 (3%), and others in 2. The approach used was the subclavian vein in 43 patients (33%), the internal jugular vein in 87 (66%), and the femoral vein in 2 (1%). As for central venous access and port placement, the approach was basically made from the right internal jugular vein under superficial ultrasonic guidance. However, the approach was determined on the basis of patient\'s condition, the lesion site, and skills of physicians in each department, and a port was placed in the precordial region. If approaches from the subclavian vein and the internal jugular vein were difficult, an approach from the femoral vein was selected, and the port was placed in the lower abdomen. After port insertion, oral antimicrobial agents were given for 3 days to prevent infection. In our hospital, clinical residents and surgeons who have practiced for 5 years are in charge of central venous port placement. Practical training in surgical techniques is included in educational programs in our university hospital. We take time to educate them. Therefore, the duration of surgery became relatively long. The used port was a PowerPort device (Bard X-Port isp, Medicon, Inc., Osaka, Japan). The presence of port infection was evaluated on the basis of abscess culture tests, serum culture tests, the detection of pathogens on culture tests after removal of the port or catheter, and the exclusion of other diagnoses on heat-source testing.
2.1. Usage {#sec2.1}
----------
Since January 2014, all implantable central venous access ports used in our hospital were switched to PowerPort devices ([Fig. 1](#fig1){ref-type="fig"}a and b). To confirm the details of PowerPort device insertion, the date of insertion, the insertion site, and the name of the surgeon were recorded in an electronic medical record. In addition, frontal plain chest radiographs were taken after port placement to confirm the port ([Fig. 2](#fig2){ref-type="fig"}). To facilitate understanding of PowerPort devices by the staff in our hospital, we conducted awareness and usage educational sessions on PowerPort, which were sponsored by the Therapy Safety Promotion Office, and prepared an original manual for usage in our hospital. As for the procedure at the time of usage, each patient\'s port certificate and bracelet were confirmed, and the medical staff then confirmed that the port was a PowerPort device on the basis of medical records and chest radiographic images. At the time of contrast-enhanced CT scanning, a pressure-resistant PowerLoc^®^ port access needle (C.R. BARD, Salt Lake City, Utah, USA) was used with PowerPort. In addition to anticancer therapy and venous nutrition, blood collection, transfusion, contrast-enhanced CT, and contrast-enhanced MRI were performed.
2.2. Primary and secondary outcomes {#sec2.2}
-----------------------------------
The evaluated variables were defined as concurrent diseases and the treatment regimens, the results of imaging studies performed during the follow-up period, and early and late postoperative complications derived from the outpatient and inpatient medical records since March 1, 2014 or from the time of initial presentation to our hospital. The primary endpoint was postoperative complications during the follow-up period. The types and incidences of postoperative complications and risk factors for postoperative complications were analyzed by univariate analysis. Secondary endpoints were the safe use of the PowerPort system in the hospital and early and late postoperative complications.
2.3. System {#sec2.3}
-----------
To ensure that the PowerPort system was understood and used correctly in our hospital, we held meetings with the department in charge of using this system. To confirm that the PowerPort system was used safely, we obtained data on all patients who underwent port insertion and followed up the patients by sharing electronic medical records. Conventional port systems were not fully managed. Many types of ports and different surgical techniques were used previously, thereby precluding a direct comparison with our results. Patients who received systemic anticancer agents using a port and those who required nutritional management were included in the study. Patients were excluded from the study if they gave no informed consent for port usage. Overall, 132 patients had no port dysfunction before and after treatment although port-related troubles occurred. The limitation of our study was a small sample size. Further studies of a larger number of patients may be needed to confirm the usefulness of the PowerPort system and risk factors for postoperative complications.
3. Results {#sec3}
==========
The median observation period after surgery was 17 months (range, 8 to 29). Postoperative complications developed in 8 patients (6%). In 4 patients the port was removed because of infection. Catheter kinking occurred in 1 patient in whom a right internal jugular vein approach was used for port placement. The port was removed and then replaced. Port extravasation occurred in 3 patients ([Fig. 3](#fig3){ref-type="fig"}a and b). In all 3 patients, port placement was performed to provide nutrition, and the ports were used after confirming reverse blood flow. There was no catheter pinch-off ([Table 2](#tbl2){ref-type="table"}).
Early complications were defined as complications that occurred within 30 days after surgery and comprised infection in 2 patients, catheter kinking in 1 patient, and port leakage in 1 patient. Late complications were defined as complications that occurred 30 or more days after surgery and comprised infection in 2 patients and port leakage in 2 patients. In our study, no patient had port-associated thrombosis, occlusion, dislodgement, migration, breakage, or rupture. Only 1 patient had catheter kinking. Port infection occurred in 4 patients, all of whom had local infection. The port was promptly removed after infection in all of these patients.
The mean operation time was 74 min (range, 32--171). No patient had intraoperative bleeding or pneumothorax. Benign disease was a risk factor for postoperative complications (p = 0.009) ([Table 3](#tbl3){ref-type="table"}). Postoperative complications developed in 3 patients with benign disease (30%). One patient had Crohn\'s disease, in 1 had ulcerative colitis, and 1 had dermatomyositis. One month after insertion, postoperative complications occurred in 37.5% of the patients, and the rate of port patency calculated with the Kaplan-Meier method was 62.5%. After 3 months, postoperative complications occurred in 12.5% of the patients, and the rate of port patency was 87.5%. Because the follow-up period was short, the 1-year survival rate was 94%. In our hospital, residents and surgeons who worked for 5 years are in charge of inserting central venous ports. Practical training in surgical techniques is included in educational programs in our university hospital. We take time to educate them. Therefore, the duration of surgery became relatively long. As for sample size, about 200 more patients may be required for a significance test. We are planning to study more patients and perform a retrospective statistical analysis. The following sentences were added to the Results section.
The causative organisms of port infection were Staphylococcus epidermidis in 1 patient, Staphylococcus capitis in 1 patient, and Staphylococcus aureus in 1 patient. The cultures in 1 patient were negative. As for the treatment of infection, the port was removed promptly after confirming symptoms such as fever, pain, and edema. Oral antimicrobial agents were given for 3--5 days. After port removal, the effect of antimicrobial agents was assessed on the basis of the results of cultures obtained from the port and catheter, and the antimicrobial agents were modified if necessary.
3.1. Statistical analysis {#sec3.1}
-------------------------
Patient data are expressed as means ± SD. Comparisons between 2 groups were done with unpaired t-tests. Incidences were compared with the use of chi-square tests with Yates\' correction. Statistical significance was evaluated with the Mann-Whitney *U* test. Port patency rates were calculated with the Kaplan-Meier method.
4. Discussion {#sec4}
=============
In our study, complications developed in 8 patients (6%) after PowerPort placement. The port was removed and replaced because of infection in 4 patients. Catheter kinking occurred in 1 patient, and the catheter was replaced. Port extravasation was detected in 3 patients. Benign disease was a risk factor for postoperative complications. Progress in chemotherapy and an increased number of drugs that can be selected have prolonged the duration of chemotherapy and increased the number of doses administered, thereby increasing the number of ports required to reliably obtain peripheral venous access. In some patients with unresectable or recurrent cancer, the gastrointestinal tract cannot be used as a nutritional route because of intestinal disease. The dissemination of home-based parenteral nutrition has contributed to improving patients\' quality of life. However, in our study benign disease was a risk factor for postoperative complications. This was probably because the daily intravenous administration of high-calorie infusion solutions, frequent port punctures, and long-term treatment are associated with a high risk of infection, including decreased disinfectant activity. The extravascular extravasation of anticancer agents can damage the soft tissue around veins and cause redness, edema, pain, sacculation, ulceration, and necrosis. In the United States, the incidence of extravascular extravasation of intravenously administered necrotic anticancer agents is estimated to range from 0.1% to 6.5%. Therefore, the placement of venous access ports has been recommended to reliably obtain venous access ports and decrease the risk of extravascular extravasation [@bib12]. The internal jugular vein and veins of the upper arm and the forearm have been advocated for catheter placement. An approach from the internal jugular vein is free of the risk of catheter pinch-off, but may require the creation a long subcutaneous tunnel for passage of the catheter at the time of port placement. Although the risks of catheter kinking and other complications remain to some extent, a bigger concern is that the large surgical invasion leads to a poor esthetic outcome of the neck caused by subcutaneous placement of the catheter [@bib13]. As compared with conventional venous access, the advantages of central venous access ports include less pain caused by frequent punctures, freedom from the need for fluid lines, less vascular pain, the ability to infuse highly irritating drugs (e.g., anticancer agents), a low incidence of infection, and reliable venous puncture. On the other hand, central venous access ports have some disadvantages, such as esthetic problems and serious complications (e.g., catheter pinch-off, infection, thrombus formation, catheter obstruction, and port extravasation) [@bib14]. In all 3 patients with port extravasation in our study, ports were placed to perform nutritional therapy, frequent port punctures were required, and the ports were used after confirming reverse blood flow. Patients and medical staff members should therefore confirm methods for puncture at a certain interval. If necessary, the appropriate puncture site should be reconfirmed on superficial ultrasonography.
A previous study reported that catheter rupture occurred when an implantable port that could not withstand the injection pressure was used to perform contrast-enhanced CT [@bib15]. The Food and Drug Administration in the United States recommended that implantable ports able to withstand the pressure created by rapid injection of contrast media, should be developed, and PowerPort ports that permit the rapid injection of contrast media were launched [@bib16]. Recent studies have reported on the safety and usage of ports that can be used to rapidly inject contrast media. PowerPort devices were not associated with any adverse events at the time of contrast-enhanced imaging, and the incidences of adverse events such as catheter obstruction and infection were similar to those of other types of ports. PowerPort ports can thus be used safely [@bib17], [@bib18], [@bib19].
Although PowerPort devices have many advantages, the features of PowerPort ports should be differentiated from those of the many other types of ports that have been used in our hospital. In our hospital, central venous access ports are placed in about 100 patients per year. In some of the patients, ports other than PowerPort devices have been used, and performing contrast-enhanced imaging via ports that are not designed to be used for contrast-enhanced CT carries the risk of accidents. We have therefore used PowerPort devices in all patients in whom implantable ports have been placed in our hospital since January 2014. At the time of introduction, we conducted adequate educational sessions and prepared an original manual to facilitate the understanding of PowerPort devices by our hospital staff.
It is important to acknowledge that postoperative complications can occur even if puncture is performed under ultrasound guidance [@bib20]. After puncture, catheter-related infection, thrombus formation, and catheter pinch-off occur at certain rates. The early management of complications is essential before they become serious. Appropriate countermeasures against such complications must therefore be clearly defined. Catheter-related infections are often caused by thrombus formation and bacterial contamination of infusion solution via the catheter insertion sites, connection sites, or three-way stopcocks. Long-term catheter placement is associated with an increased incidence of thrombus formation. One study reported that thrombus formation occurred in 84% of patients in whom a catheter was placed for 10 months or longer [@bib21]. Scolapios et al. reported that 11 (5%) of 225 patients with port placement died of catheter-related sepsis. They strongly recommended that appropriate action and early catheter removal should be performed in a timely manner in patients suspected to have catheter-related infection [@bib22]. Catheter pinch-off syndrome is a troublesome problem that occurs when a catheter placed along the subclavian vein route gets caught between the clavicle and the first rib. The incidence of port catheter rupture has been reported to range from 1.1% to 2.1% [@bib23]. No port catheter rupture occurred in our study. Fisher et al. reported that serious complications such as arrhythmias, sepsis, and cardiac puncture can occur in about 71% of patients in whom catheter fragments are released [@bib24].
Careful screening of risk factors for port-related complications can decrease patients\' discomfort associated with port insertion and management and provide appropriate treatment, permitting the maintenance of a better quality of life and allowing medical staff members to offer safe treatment.
Ethical approval {#sec5}
================
This study has been approved by Institutional Review Board for Observation and Epidemiological Study of Kitasato University, and all the subjects gave informed consent. (B16-99).
Sources of funding {#sec6}
==================
None to declare.
Author contribution {#sec7}
===================
All the authors, Takatoshi Nakamura, Jiichiro Sasaki, Yasushi Asari, Takeo Sato, Shinzo Torii, and Masahiko Watanabe, contributed to the study design, data collections, data analysis and writing of the manuscript.
Conflicts of interest {#sec8}
=====================
None to declare.
Guarantor {#sec9}
=========
Takatoshi Nakamura, First Author.
Masahiko Watanabe, Senior Author.
Trial registry number {#sec10}
=====================
N/A.
{#fig1}
{#fig2}
{#fig3}
######
Patients characteristics in our study.
Table 1
n = 132(%)
------------------------ ------------------------- --------------------
Sex Male: Female 61 (46%): 71 (54%)
Age 63.5 (±12)
BMI(kg/m^2^) 21.0(±4.5)
Purpose Chemotherapy: Nutrition 90 (68%): 42 (32%)
Operation time(min) 74.2 (±26.4)
Period(median,methods) 17.0 (8--29)
######
Patients with postoperative complications.
Table 2
Case Sex Age Disease purpose Operation time(min) Port site complication Occurrece (days)
------ -------- ----- -------------------- -------------- --------------------- -------------------------- ------------------------------------------------------ ------------------
1 Female 44 Breast cancer Chemotherapy 171 Rt.internal jugular vein infection 90
2 Male 50 Crohn disease Nutrition 44 Rt.subclavian vein infection 21
3 Male 36 Ulcerative colitis Nutrition 89 Lt.internal jugular vein Infection 150
4 Male 61 Rectal cancer Chemotherapy 62 Lt. subclavian vein Infection 28
5 Female 69 Colon cancer Chemotherapy 150 Rt.internal jugular vein Catheter kink 25
6 Female 66 Ovarian cancer Nutrition 147 Rt.subclavian vein Port extravasation[a](#tbl2fna){ref-type="table-fn"} 9
7 Female 68 Dermatom yosistis Nutrition 148 Rt.internal jugular vein Port extravasation[a](#tbl2fna){ref-type="table-fn"} 35
8 Male 81 Esophageal cancer Nutrition 160 Rt.internal jugular vein Port extravasation[a](#tbl2fna){ref-type="table-fn"} 81
Reverse blood flow was confirmed before using the port.
######
Postoperative complications.
Table 3
Factors Complications(n = 8) No commplications(n = 124) p = value
--------------------- ---------------------------------------- ---------------------- ---------------------------- -----------
Sex Male: Female 4: 4 57: 67 0.8248
Age \<65:≧65 4: 4 62: 62 0.9999
BMIkg/m^2^ \<25: ≧25 7: 1 98: 26 0.9019
Purpose Chemotherapy: Nutrition 5: 3 85: 39 0.7254
Port site Internal jugular vein: Subclavian vein 5: 3 82: 40 0.7170
Disease Malignant neoplasm: Benign 5: 3 117: 7 0.0009
Operation time(min) \<90: ≧90 5: 3 98: 26 0.5131
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1}
===============
In recent years, ultrasound diffraction tomography (UDT) has drawn more and more attention in medical imaging field. Different from traditional B-mode ultrasound technique which displays the strength of the echoes with gray scale to show anatomic structure, UDT infers the distribution of acoustic properties such as refractivity, attenuation, and density. Since these acoustic properties of normal and diseased tissues have different value ranges \[[@B1]\], UDT has the potential for providing functional information of the object. For example, in the breast cancer exam, the malignant tumor, the benign mass, and the normal tissue can be differentiated by UDT \[[@B2]\].
Under the assumption of weak scatting, the Fourier diffraction theory (FDT) \[[@B3], [@B4]\] is adopted for the image reconstruction of UDT. Firstly, the object is illuminated by plane sound wave from one certain direction and the scattering waves are measured and sampled. Secondly, the spatial Fourier transform is performed on the stored data. Under the Born or Rytov approximation \[[@B5]\], the corresponding frequency values are considered to represent the 2D Fourier transform of the object nonuniform distributed along a semicircle arc. The above processes are implemented oriented at various angles around the object to acquire sufficient spatial frequency information. Finally, the UDT image is reconstructed through inverse spatial Fourier transform. To avoid undersampling, normally more spatial frequency samples are required \[[@B6]\]; thus more views from different directions are needed. However, the scan and process time and the associated cost will increase. Besides that, it will impose rigorous requirements on controlling precision of the imaging system. Moreover, redundant information is introduced by multiple views, which results in the waste of the system resources. Hence, accurate reconstruction from nonuniform distributed frequency samples in sparse-view situation has great practical significance.
Currently, in addition to reconstructing the underlying object through beamforming \[[@B7]\], there are other two main approaches based on FDT to recover the complete information from the sparse-view data: interpolation method and iterative method. The former calculates the frequency values on the rectangular grids by a predetermined interpolation function based on sampled frequency on the circular arc grid \[[@B8], [@B9]\]. Although the interpolation method is computationally efficient, it is liable to introduce coordinate conversion error, which can severely introduce artifacts and distort the image, especially in sparse-view situation. The latter repeatedly corrects the reconstructed image by minimizing the inconsistency between the sampled and the estimated frequency. Through designing appropriate optimization strategy, this process can maximally reduce the reconstruction error and suppress the artifacts. Various researches have been done on the UDT image reconstruction with the iterative method under sparse-view situation. Bronstein et al. \[[@B10]\] proposed an iterative reconstruction framework based on nonuniform fast Fourier transform (NUFFT) and reduced the number of view with broadband sound wave. LaRoque et al. \[[@B11]\] introduced a kind of diffraction tomography with few-view (sparse-view) and limited-angle data through total variation (TV) minimization algorithm for absorberless media. Tingting Li et al. \[[@B12]\] combined TV regularization with the iterative next-neighbor regridding (INNG) algorithm to suppress artifacts and noise.
Compressed sensing (CS) \[[@B13], [@B14]\] is built on the sparse nature of real signals in certain transform domain. CS has the ability to reconstruct signals with samples which are much less than those required by the Nyquist criterion. CS brings great innovations in image reconstruction and has been widely used in medical imaging field such as MR \[[@B15]\], CT \[[@B16], [@B17]\], PET \[[@B18]\], and photoacoustic imaging \[[@B19], [@B20]\]. In recent years, some progress has been made in applying this emerging theory in ultrasound imaging field to reduce the amount of the data and the complexity of the imaging system. Schiffner et al. \[[@B21]--[@B23]\] investigated the performance of CS in solving the inverse scattering problem in pule-echo diagnostic ultrasound imaging under the constraint that the scatterer distribution is sparse. With the same assumption, Wagner et al. \[[@B24]--[@B26]\] proposed a method based on finite rate of innovation and Xampling for the reconstruction of the beamformed image from channel RF data. Shen et al. \[[@B27], [@B28]\] presented a measurement-domain adaptive beamforming approach based on distributed CS to reconstruct an image of sparse targets. Achim et al. \[[@B29]\] introduced a framework based on CS for ultrasonic signals reconstruction under the assumption of RF echoes with *α*-stable distributions. Liebgott et al. \[[@B30]--[@B32]\] studied the feasibility of CS for the reconstruction of channel RF data, the quantity of channel RF data is reduced through introducing the wave atoms as a representation basis for prebeamformed RF signal. Quinsac et al. \[[@B33]--[@B37]\] and Dobigeon et al. \[[@B38]\] applied CS theory to recover beamformed 2D RF images using conjugate gradient descent or Bayesian approach. Zobly et al. \[[@B39], [@B40]\] and Richy et al. \[[@B41]\] applied CS for doppler imaging.
In this paper, we propose an iterative reconstruction method based on CS for sparse-view UDT. The above researchers mainly focused on applying CS to the pulse-echo imaging systems. To the best of our knowledge, CS was rarely applied in the investigation of UDT image reconstruction, and this is the main motivation of this work. According to FDT, the sample points are bounded in a circle with the radius of $\sqrt{2}k_{0}$ (*k* ~0~: wave number) \[[@B9]\] for transmission UDT; this results in a sharp sampling cutoff in spatial frequency space. Thus, the reconstructed object is a low-pass version of the original and the quality of image will be distorted by the Gibbs aliasing. This problem can be further deteriorated in the sparse-view situation due to limited sample data. In this context, besides the general *l* ~1~ norm constraint, we introduce TV penalty into the cost function, which can reduce the oscillation and preserve edges of the object \[[@B42]\]. Empirical observations \[[@B10]\] showed that the majority of nature images, particularly medical images, demonstrated piecewise continuous behavior; that is, parts of anatomy structures were supposed to show uniform characteristics, which belonged to the class of functions of bounded TV \[[@B43]\]. Furthermore, to improve computational efficiency, a fast and accurate NUFFT is adopted to calculate the forward scatter field.
Through the numerical simulation, the proposed method is verified and compared with interpolation method and iteration method with broadband signal \[[@B10]\]. The quantitative evaluation and the robustness to noise of the method are discussed. Simulation shows that the object can be faithfully reconstructed in sparse-view situation without noticeable loss of image quality and the reconstructed error is reduced.
The rest of the paper is organized as follows. The basic principle of UDT and CS theory are described in [Section 2](#sec2){ref-type="sec"}. The proposed method is presented in [Section 3](#sec3){ref-type="sec"}. [Section 4](#sec4){ref-type="sec"} describes the simulation experiments for UDT image reconstruction and results. [Section 5](#sec5){ref-type="sec"} provides discussion about noise robustness and computational complexity of the proposed method. Conclusions and future work are summarized in [Section 6](#sec6){ref-type="sec"}.
2. Background {#sec2}
=============
2.1. UDT Based on FDT {#sec2.1}
---------------------
The classical 2D UDT imaging configuration is shown in [Figure 1](#fig1){ref-type="fig"}. The inhomogeneous object with a distribution function *f*(**r**) = *n* ^2^(**r**) − 1, **r** = (*x*, *y*), is surrounded by homogeneous medium such as water; the *n*(**r**) is refractive index. The reconstruction object of UDT is to infer the unknown *f*(**r**) through transmitted signals measured by ultrasound transducer.
Assume the object is illuminated by a monochromatic plane wave with wave number *k* ~0~ and angular frequency *ω* ~0~ at an angle *θ*. The wave is scattered within and at the boundary of the inhomogeneous object. Set up a rotated cartesian coordinate system (*ξ*, *η*) such that *η*-axis is coincident with the view angle as shown in [Figure 1](#fig1){ref-type="fig"}. The relation between (*x*, *y*) and (*ξ*, *η*) can be derived through coordinate transformation: *ξ* = cos(*θ*)*x* + sin(*θ*)*y*, *η* = −sin(*θ*)*x* + cos(*θ*)*y*.
The sound pressure field *u*(**r**) satisfies the following wave equation \[[@B4]\]: $$\begin{matrix}
{\left( \nabla^{2} + k_{0}^{2} \right)u\left( \mathbf{r} \right) = - k_{0}^{2}f\left( \mathbf{r} \right)u\left( \mathbf{r} \right),} \\
\end{matrix}$$ where ∇^2^ denotes the Laplacian operator. *u*(**r**) can be modeled as the superposition of the incident wave *u* ~*i*,*θ*~(**r**) and the scatter wave *u* ~*s*,*θ*~(**r**): *u*(**r**) = *u* ~*i*,*θ*~(**r**) + *u* ~*s*,*θ*~(**r**).
Under the assumption of weak scattering, *u* ~*s*,*θ*~(**r**) ≪ *u* ~*i*,*θ*~(**r**), or first order Born, or Rytov approximation, one can derive the so-called FDT which relates the scattered field measured along the line *η* = *l* to the object by Fourier transform: $$\begin{matrix}
{U_{s,\theta}\left( \kappa \right) = \frac{k_{0}^{2}U_{0}}{2jy}e^{jyl}{\iint_{- \infty}^{+ \infty}{f\left( \mathbf{r} \right)e^{- j\lbrack{\kappa\xi + ({\gamma - k_{0}})\eta}\rbrack}d\mathbf{r},}}} \\
\end{matrix}$$ where *U* ~*s*,*θ*~(*κ*) is the Fourier transform of received scattered data *u* ~*s*,*θ*~(*ξ*) with respect to *ξ*: $$\begin{matrix}
{U_{s,\theta}\left( \kappa \right) = \int\limits_{- \infty}^{+ \infty}u_{s,\theta}\left( \xi \right)e^{- j\kappa\xi}d\xi} \\
\end{matrix}$$ and $y = \sqrt{k_{0}^{2} - \kappa^{2}}$, \|*κ*\| ≤ *k* ~0~, and *U* ~0~ is the complex amplitude of the illuminating plane wave. The quantity of *U* ~*s*,*θ*~(*κ*) in ([2](#EEq2){ref-type="disp-formula"}) is known or measurable, and the integral *Q* ~*θ*~(*κ*) = ∬~−*∞*~ ^+*∞*^ *f*(**r**)*e* ^−*j*\[*κξ*+(*γ*−*k*~0~)*η*\]^ *d* **r** contains the underlying object *f*(**r**) that we want to reconstruct. Detailed analysis reveals that *Q* ~*θ*~(*κ*) is the Fourier transform of the object along a semicircle of radius *k* ~0~ and centered at −*k* ~0~ **s** ~0~ \[[@B4]\], as the arc AOB depicted in [Figure 2](#fig2){ref-type="fig"}, and **s** ~0~ is the unit vector along direction *θ*. Mathematically, the image reconstruction by ([2](#EEq2){ref-type="disp-formula"}) is a typical inverse problem which can be expressed as the following formulation: $$\begin{matrix}
{F = \Phi\left( f \right),} \\
\end{matrix}$$ where *F* is the sample data in *k*-space, Φ is the diffraction operator, and *f* is the object *f*(**r**) which we want to reconstruct. For formulation ([4](#EEq4){ref-type="disp-formula"}), classical interpolation method based on FDT is not applicable in sparse-view situation due to violating Nyquist limitation. With fast and accurate nonuniform fast Fourier transform, we propose a CS framework for UDT reconstruction in sparse-view situation to improve the image quality of the reconstruction.
2.2. Compressed Sensing: A Short Overview {#sec2.2}
-----------------------------------------
Compressed sensing is a kind of signal processing technique for efficiently acquiring and reconstructing a signal. CS has been increasingly adopted in a variety of applications by applied mathematicians, computer scientists, and engineers since it was initiated in 2006 \[[@B44]\]. The idea of CS can be expressed by the following linear measurement model: $$\begin{matrix}
{\mathbf{y} = \Phi\mathbf{x},} \\
\end{matrix}$$ where **x** ∈ ℝ^*n*^ is the unknown signal such as an image that we want to reconstruct, **y** is the measured signal, and Φ is one *m* × *n* (*m* \< *n*) measure matrix, which is decided by the imaging system. In this work, Φ is the Fourier transform operator and **y** is the corresponding Fourier transform of the received scatter field. Assume the unknown signal **x** can be sparsely represented in terms of a known basis: $$\begin{matrix}
{\mathbf{x} = \Psi\mathbf{s},} \\
\end{matrix}$$ where Ψ is the basis, **s** is the corresponding representation coefficients. Here, the sparse means that the number of nonzero coefficients of **s** is small.
Substituting ([6](#EEq6){ref-type="disp-formula"}) into ([5](#EEq5){ref-type="disp-formula"}), we obtain ([7](#EEq7){ref-type="disp-formula"}) as follows: $$\begin{matrix}
{\mathbf{y} = \Theta\mathbf{s},} \\
\end{matrix}$$ where Θ = ΦΨ; since *m* \< *n*, the matrix Θ is not invertible. Taking advantage of the sparsity of **s**, CS theory shows that if the Θ meets the so-called restricted isometry property (RIP) condition, we can exactly recover **s** with overwhelming probability by solving the following minimization problem \[[@B45], [@B46]\]: $$\begin{matrix}
{\underset{\mathbf{s}}{\min}\left\{ \left. ||\mathbf{s} \right.||_{0}:\Theta\mathbf{s} = \mathbf{y} \right\},} \\
\end{matrix}$$ where \|\|·\|\|~0~ norm counts the number of no-zero entry of a vector. Equation ([8](#EEq8){ref-type="disp-formula"}) seeks the sparsest one among all the possible solutions of **y** = Θ**s**.
However, ([8](#EEq8){ref-type="disp-formula"}) belongs to the class of NP-hard problem, which is difficult to obtain solutions for nearly all real applications. One computationally tractable alternative for ([8](#EEq8){ref-type="disp-formula"}) is to solve the following *l* ~1~ problem \[[@B13], [@B45]\]: $$\begin{matrix}
{\underset{\mathbf{s}}{\min}\left\{ \left. ||\mathbf{s} \right.||_{1}:\Theta\mathbf{s} = \mathbf{y} \right\},} \\
\end{matrix}$$ where \|\|**s**\|\|~1~ = ∑\|*s* ~*i*~\|.
The above CS theory requires **x** to be sparse and **y** = Φ**x** exactly, whereas, in most practical situations, the object is approximately sparse or compressible. Here, approximately sparse means that **x** contains a small number of components with magnitudes significantly larger than those of the rest, which are not necessarily zero; compressible means the coefficients of **s** decay exponentially in absolute value. To address these problems, Candès et al. \[[@B47]\] extended ([9](#EEq9){ref-type="disp-formula"}) to the following form: $$\begin{matrix}
{\underset{\mathbf{s}}{\min}\left\{ \left. ||\mathbf{s} \right.||_{1}:\left. ||{\Theta\mathbf{s} - \mathbf{y}} \right.||_{2}^{2} \leqslant \sigma^{2} \right\},} \\
\end{matrix}$$ where \|\|**s**\|\|~2~ = (∑\|*s* ~*i*~\|^2^)^1/2^ and *σ* ^2^ represents energy bound of error.
Considering that ([10](#EEq10){ref-type="disp-formula"}) is a convex optimization problem, it can be further recast as the following regularization equation \[[@B48]\]: $$\begin{matrix}
{\underset{\mathbf{s}}{\min}\left\{ \alpha\left. ||\mathbf{s} \right.||_{1} + \left. ||{\Theta\mathbf{s} - \mathbf{y}} \right.||_{2}^{2} \right\},} \\
\end{matrix}$$ where *α* is a regularization parameter.
3. Compressed Sensing for UDT Imaging Reconstruction {#sec3}
====================================================
In this section, we present an image reconstruction method for UDT in 2D case, while it can be readily extended to 3D. [Figure 3](#fig3){ref-type="fig"} provides the comparison between the interpolation method and the proposed scheme, which is composed of three major steps. Firstly, the object is illuminated from random angles and the number of views can be far below that restricted by the Nyquist limitation. The measured sparse-view data are processed by Fourier transform to obtain the corresponding spatial frequency samples along semicircular arcs oriented at the view angles. Secondly, the inverse problem ([4](#EEq4){ref-type="disp-formula"}) is formulated within CS framework by building the measure matrix Φ and exploiting the sparsity of the object. Thirdly, the object is reconstructed through the NUFFT and CG algorithm.
3.1. Sparse-View Data Sampling {#sec3.1}
------------------------------
As we discussed in [Section 2.2](#sec2.2){ref-type="sec"}, when the sensing matrix Θ satisfies the RIP \[[@B45], [@B46]\], the NP-hard inverse problem ([8](#EEq8){ref-type="disp-formula"}) can be transformed to computationally tractable *l* ~1~ norm minimization problem (([9](#EEq9){ref-type="disp-formula"}) or ([11](#EEq11){ref-type="disp-formula"})). However, even for moderate dimensional operators Θ, it is computationally impractical to verify the RIP. Fortunately, a few classes of matrices are shown to hold RIP for almost certainly. It is shown in \[[@B49], [@B50]\] that, when Φ is a Gaussian or partial Fourier, that is, the entries of Φ are randomly selected using a Gaussian pdf or *m* rows of Φ are randomly selected from the rows of *n* × *n* Fourier matrix, Θ satisfies RIP.
For UDT scanner, the sample operator coincides with the matrix mentioned above since the samples in spatial frequency domain are obtained through FDT, Which makes it possible to reconstruct the object through *l* ~1~ minimization with sparse data. Here, we generate the sparse-view data by randomly choosing view angles and the number of view is much less than that required by the conventional interpolation method.
3.2. Inverse Problem Formulation under CS Framework {#sec3.2}
---------------------------------------------------
The sparse-view data sampling in [Section 3.1](#sec3.1){ref-type="sec"} generates the measured value *y* in problem ([9](#EEq9){ref-type="disp-formula"}). To formulate the UDT inverse problem under CS framework, the sensing matrix Θ = ΦΨ in ([9](#EEq9){ref-type="disp-formula"}) must be constructed. Here, Φ is dependent on the imaging principle of UDT, while Ψ is the basis for the object to be reconstructed. The explicit expression of Φ and Ψ for UDT will be derived as below.
Since the object has limited physical dimensions, we assume the object *f*(**r**) or *f*(*x*, *y*) in cartesian coordinate system has bounded support \[−*C*, *C*\]×\[−*C*, *C*\]; that is, *f*(*x*, *y*) = 0 when \|*x*\| \> *C* or \|*y*\| \> *C*. Let *f* ~*d*~ ∈ ℝ^(*n*~*d*~+1)×(*n*~*d*~+1)^, (*n* ~*d*~ = 2⌈*C*/*T*⌉) be the discrete form of the underlying object that we want to reconstruct, where *T* is the sample period for *x*- and *y*-axis. *F*(*u*, *v*) and *F* ~*d*~(*u*, *v*) are the Fourier transform of *f*(*x*, *y*) and *f* ~*d*~(*n* ~1~, *n* ~2~), respectively. Assume the frequency response of *f*(*x*, *y*) is band-limited; that is, there exists a cutoff frequency *W* such that \|*F*(*u*, *v*)\| ≈ 0, when \|*u*\| \> *W* or \|*v*\| \> *W*. In practice, the loss of resolution by this band limit is negligible; the reconstructed imaging quantity is more influenced by other factors such as the aperture sizes of transmitting and receiving transducers \[[@B9]\]. Based on the Nyquist theorem, if *T* \< 1/2*W*, *F* ~*d*~(*u*, *v*) ≈ *F*(*u*, *v*) for \|*u*\| \< *W* and \|*v*\| \< *W*.
Since the receiving array of UDT has a limited number of elements, we denote the measured discrete field ${\hat{u}}_{s,\theta} = u_{s,\theta}(n\tau)$, where *n* is the number of elements and *τ* is pitch (the distance between the centers of two adjacent elements). *τ* is the spatial sampling interval of UDT system. ${\hat{U}}_{s,\theta}(\kappa)$ and *U* ~*s*,*θ*~(*κ*) are the Fourier transform of ${\hat{u}}_{s,\theta}$ and *u* ~*s*,*θ*~(*ξ*), respectively. $\left. {\hat{U}}_{s,\theta}(\kappa)\rightarrow U_{s,\theta}(\kappa) \right.$ for all *κ* ∈ ℝ as *τ* → 0. The relation between (*u*, *v*) and (*κ*, *γ*) can be formulated as the following equations: $$\begin{matrix}
{\kappa = \mathcal{U}\left( {u,v} \right),} \\
{\gamma = \mathcal{V}\left( u,v \right),} \\
\end{matrix}$$ where *𝒰*, *𝒱* are derived from the coordinate transform and ([2](#EEq2){ref-type="disp-formula"}); let arc grid points (*κ* ~1~, *γ* ~1~) and (*κ* ~2~, *γ* ~2~) (−*k* ~0~ ≤ *κ* ~1~ ≤ 0, 0 ≤ *κ* ~2~ ≤ *k* ~0~) be on half arc*AO*,*OB*, respectively; then we can get $\kappa_{i} = {( - 1)}^{i}k_{0}\sin(2\,\arcsin{{(\sqrt{u^{2} + v^{2}}}/{2k_{0}}}))$, $\gamma_{i} = \sqrt{k_{0}^{2} - \kappa_{i}^{2}}$, *i* = 1, 2. For more details about this transformation please refer to \[[@B4]\].
According to ([2](#EEq2){ref-type="disp-formula"}) and ([12](#EEq12){ref-type="disp-formula"}), if ${\hat{U}}_{s,\theta}(\kappa)$ sufficiently approximates *U* ~*s*,*θ*~(*κ*) and \|*u*\| \< *W*, \|*v*\| \< *W*, then we have ${\hat{U}}_{s,\theta}(\kappa) \approx U_{s,\theta}(\kappa) = F(\kappa,\gamma) = F(u,v) \approx F_{d}(u,v)$. Generally, this requirement can be satisfied in the real UDT system because the pitch of normal medical ultrasound probe is among 0.5*λ* \~ 2*λ* (*λ* is the sound wave length).
Let *Ω* ~1~ = {*θ* ~1~, *θ* ~2~,..., *θ* ~*N*~1~~} be the finite set of view angles. In UDT system, *U* ~*s*,*θ*~(*κ*) is obtained over *Ω* ~1~. Then the corresponding ${\hat{U}}_{s,\theta}(\kappa)$ can be calculated offline for all *κ* ∈ *Ω* ~2~ = {*κ* ~1~, *κ* ~2~,..., *κ* ~*N*~2~~} ([3](#EEq3){ref-type="disp-formula"}). Therefore, the *f* ~*d*~ can be reconstructed through the set of observation ${\hat{U}}_{s,\theta}(\kappa) \mid (\theta,\kappa) \in (\Omega_{1},\Omega_{2})$: $$\begin{matrix}
{\frac{\left( T\kappa_{0} \right)^{2}U_{0}}{2j\gamma}e^{j\gamma l}\sum\limits_{n_{1} = {\lfloor{{- C}/T}\rfloor}}^{\lceil{C/T}\rceil}\,\sum\limits_{n_{2} = {\lfloor{{- C}/T}\rfloor}}^{\lceil{C/T}\rceil}f_{d_{n_{1},n_{2}}}e^{- 2\pi j(n_{1}u + n_{2}v)T}} \\
{\quad = {\hat{U}}_{s,\theta}\left( \kappa \right) + n_{s,\theta}\left( \kappa \right).} \\
\end{matrix}$$ This set of equations can be written in matrix form: Φ*f* ~*d*~ = *F* + *n*, where *Ω* = (*Ω* ~1~, *Ω* ~2~), Φ ∈ *ℂ* ^\|*Ω*\|×(*n*~*d*~\ +\ 1)^2^^, *f* ~*d*~ ∈ ℝ^(*n*~*d*~\ +\ 1)^2^^, −*n* ~*d*~/2 ≤ *n* ~1~, *n* ~2~ ≤ *n* ~*d*~/2, *F*, *n* ∈ *ℂ* ^\|*Ω*\|^, *n* ~*s*,*θ*~(*κ*), or *n* is approximation error. With ([13](#EEq13){ref-type="disp-formula"}), Φ in inverse problem ([9](#EEq9){ref-type="disp-formula"}) is obtained.
CS theory utilizes the sparse nature of the object and reconstructs the object by minimizing the corresponding *l* ~1~ norm in transform domain. According to FDT, the spatial frequency samples are distributed along the arc AOB ([Figure 2](#fig2){ref-type="fig"}). As the incident wave revolves around the object, the AOB describes a disk of radius $\sqrt{2}k_{0}$ centered at the origin; that is, the reconstructed object is a low-pass version of the original. Besides that, we also exploit the fact that the structural morphology of human soft tissue is expected to demonstrate piecewise continuous behavior. That means the object belongs to the class of bounded TV \[[@B43]\] and the gradient of the object is sparse. In this work, the sparsity of the underlying object is exploited not only through wavelet transform which provides sparse representations for rapidly varied regions, but also by TV which affords sparse transformation for piecewise smooth object. Furthermore, TV constraint can help to suppress Gibbs effect and preserve edges \[[@B10]\].
For a discrete object *f* ~*d*~, TV is defined as $$\begin{matrix}
{\text{TV}\left( f_{d} \right) = \sum\limits_{n_{1},n_{2}}\sqrt{\left( D_{n_{1},n_{2}}^{h}f_{d} \right)^{2} + \left( D_{n_{1},n_{2}}^{v}f_{d} \right)^{2}} = \sum\limits_{n_{1},n_{2}}\left. ||{D_{n_{1}n_{2}}f_{d}} \right.||_{2},} \\
\end{matrix}$$ where *D* ~*n*~1~,*n*~2~~ ^*h*^, *D* ~*n*~1~,*n*~2~~ ^*v*^ denote the forward finite difference operator in horizontal (*h*) and vertical (*v*) coordinates, respectively. Combining *l* ~1~ norm with TV constraints, we extend the problem ([11](#EEq11){ref-type="disp-formula"}) to the following minimization problem: $$\begin{matrix}
{\underset{s}{\min}\, G\left( s \right) = \alpha\text{TV}\left( \Psi\mathbf{s} \right) + \beta\left. ||\mathbf{s} \right.||_{1} + \left. ||{F - \Phi\Psi\mathbf{s}} \right.||_{2}^{2},} \\
\end{matrix}$$ where *α*, *β* are two positive regularization parameters, *F* is spatial frequency samples, Ψ is the selected basis for sparse representation of the object *f* ~*d*~, and **s** is the coefficients of *f* ~*d*~ in basis Ψ.
3.3. Object Reconstruction {#sec3.3}
--------------------------
An iterative method based on CG is adopted to solve the inverse problem ([15](#EEq15){ref-type="disp-formula"}). For UDT system, the receive elements are equally spaced that means the *u* ~*s*,*θ*~(*ξ*) along *η* = *l* is equally spaced sampled. However, the measurements of *Q* ~*θ*~(*κ*) are unequally spaced distributed, since the measurements along the line are projected perpendicularly onto frequency domain of the object along semicircular arc. Thus, the Fourier transform must be computed for every nonuniform frequency points (ΦΨ**s** or Φ*f* ~*d*~). Although the result of direct nonuniform discrete Fourier transform (NDFT) is exact, the computation time required by the NDFT restricts its real application. To speed up, a fast NUFFT is employed to approximate NDFT in every iteration of reconstruction.
### 3.3.1. Conjugate Gradient Method {#sec3.3.1}
To solve the inverse problem ([15](#EEq15){ref-type="disp-formula"}) iteratively by CG, the gradient of the objective function *G*(**s**) must be computed as: $$\begin{matrix}
{\nabla G\left( \mathbf{s} \right) = 2\left( {\Phi\Psi} \right)^{\prime}\left( \Phi\Psi\mathbf{s} - F \right) + \alpha\nabla\left( \text{TV}\left( \Psi\mathbf{s} \right) \right) + \beta\nabla\left( \left. ||\mathbf{s} \right.||_{1} \right),} \\
\end{matrix}$$ where (·)′ represents the conjugate transpose. Since the absolute value function in *l* ~1~ norm and TV is nonsmooth function, we use approximation techniques to compute the corresponding gradient. For *l* ~1~ norm, the absolute value function is approximated with a smooth function by using the relation ${||\mathbf{z}||}_{1} \approx \sqrt{\mathbf{z}^{\prime}\mathbf{z} + ɛ}$, where *ɛ* is a small positive smooth parameter. For TV, we use the following approximation strategy to avoid a zero denominator: ${||{D_{n_{1}n_{2}}f_{d}}||}_{2} \approx \sqrt{{(D_{n_{1},n_{2}}^{h}f_{d})}^{2} + {(D_{n_{1},n_{2}}^{v}f_{d})}^{2} + \mu}$, where *μ* is a small positive parameter. Therefore the gradient of *l* ~1~ norm and TV can be calculated: $$\begin{matrix}
{\nabla\left( \left. ||\mathbf{z} \right.||_{1} \right) = \frac{\mathbf{z}}{\sqrt{\mathbf{z}^{\prime}\mathbf{z} + ɛ}},} \\
\end{matrix}$$ $$\begin{matrix}
{\nabla\left. ||{D_{n_{1},n_{2}}f_{d}} \right.||_{2} = \frac{D_{n_{1},n_{2}}^{h}f_{d}}{\left. ||{D_{n_{1},n_{2}}f_{d}} \right.||_{2}} + \frac{D_{n_{1},n_{2}}^{v}f_{d}}{\left. ||{D_{n_{1},n_{2}}f_{d}} \right.||_{2}} - \frac{D_{n_{1},n_{2} - 1}^{h}f_{d}}{\left. ||{D_{n_{1},n_{2} - 1}f_{d}} \right.||_{2}}} \\
{\quad - \frac{D_{n_{1} - 1,n_{2}}^{v}f_{d}}{\left. ||{D_{n_{1} - 1,n_{2}}f_{d}} \right.||_{2}}.} \\
\end{matrix}$$
### 3.3.2. NUFFT {#sec3.3.2}
NUFFT developed by Fessler and Sutton \[[@B51]\] is adopted. Consider the following 1D NUFFT case: $$\begin{matrix}
{F\left( \omega_{m} \right) = \sum\limits_{n = 0}^{N - 1}f_{n}e^{- i\omega_{m}n},\quad m = 1,\ldots,M,} \\
\end{matrix}$$ where **f** = (*f* ~0~,..., *f* ~*N*−1~) is a vector of equally spaced samples of a signal and **ω** = (*ω* ~1~,..., *ω* ~*M*~) is a vector of nonuniform distributed frequencies. In matrix notation $$\begin{matrix}
{F = \Phi\mathbf{f},} \\
\end{matrix}$$ where Φ ∈ *ℂ* ^*M*×*N*^ : Φ = (*ϕ* ~1~,...,*ϕ* ~*M*~)^*T*^ is nonuniform Fourier transform matrix. The NUFFT is implemented by two steps: firstly, project **f** on an oversampled uniform Fourier basis Λ ∈ *ℂ* ^*qM*×*N*^ by standard FFT $$\begin{matrix}
{Z = \Lambda\mathbf{f};} \\
\end{matrix}$$ that is, $$\begin{matrix}
{Z_{k} = \sum\limits_{n = 0}^{N - 1}f_{n}e^{- i(2\pi/K)kn},\quad k = 0,\ldots,K - 1,} \\
\end{matrix}$$ where *K* = *qM*. Secondly, approximate each *F*(*ω* ~*m*~) by interpolating the *Z* ~*k*~ using *p* uniform samples $$\begin{matrix}
{F\left( \omega_{m} \right) \simeq \hat{F}\left( \omega_{m} \right) = \sum\limits_{k = 0}^{K - 1}v_{mk}Z_{k},\quad m = 1,\ldots,M;} \\
\end{matrix}$$ that is, *F* = Φ**f**≃*V* ~*p*~Λ**f**. *V* ~*m*~ = (*v* ~*m*1~,..., *v* ~*mK*~) is the *m*th row of interpolation matrix *V* ~*p*~ which makes use of *p* neighboring uniform samples of *Z* for approximation of each nonuniform sample of *F*. In \[[@B51]\], Fessler and Sutton designed a kind of min-max criterion to choose interpolation coefficients for every *F*(*ω* ~*m*~): $$\begin{matrix}
{\underset{V_{m}}{\min}\,\underset{{||\mathbf{f}||}_{2} \leq 1}{\max}\left| V_{m}\Lambda\mathbf{f} - \phi_{m}\mathbf{f} \right|^{2}.} \\
\end{matrix}$$
The analytical solution of ([24](#EEq22){ref-type="disp-formula"}) is $$\begin{matrix}
{V_{m} = \phi_{m}\Lambda^{H}\left( {\Lambda\Lambda^{H}} \right)^{- 1},} \\
\end{matrix}$$ where *H* denotes Hermitian transpose. Fessler and Sutton have shown that the overall complexity of such method is *O*(*qN*log*N* + *pM*).
4. Simulation and Results {#sec4}
=========================
4.1. Simulation Parameters {#sec4.1}
--------------------------
In order to evaluate the performance of the proposed method for UDT reconstruction, we have performed a series of numerical experiments for the phantom in [Figure 4(a)](#fig4){ref-type="fig"}. The phantom consists of ten ellipses which looks like the well-known Shepp-Logan "head phantom" for CT imaging. However, for UDT system, we have modified the gray levels to those used by \[[@B8], [@B9]\]. The gray levels represent the relative change in refractive index from the background value of 1.0; the maximum and minimum gray intensity are set to 1.0 and 0, respectively. The speed of sound of the background media is 1500 m/s. To evaluate our method, the scattered field was calculated based on FDT under Born approximation. Although the Born approximation imposes limitation on the dimension of the object for real application \[[@B4]\] and cannot distinguish the features of the object spaced less than *λ*/2 \[[@B52], [@B53]\], it can provide a simple and direct method to reconstruct the structure of an object from the measurement of the scattered field. According to FDT, the Fourier transform of the scattered field measured on *η* = *l* is proportional to Fourier transform of the object over an arc ([2](#EEq2){ref-type="disp-formula"}), while the Fourier transform of each ellipse has simple analytical expression; hence we can generate the scattered data through inverse Fourier transform. This procedure not only is fast but also allows the scattered date to be calculated for testing the reconstruction algorithms and experiments parameters such as pitch and number of elements \[[@B4], [@B8]--[@B10]\].
In numerical experiments, the imaging system utilizes a pair of parallel linear array probes \[[@B54], [@B55]\]. Referring to exiting commercial ultrasound linear transducer and ultrasound tomography system, the frequency of incident wave is set to 1.5 MHz, and the number of elements and pitch of the probes are 128 and *λ*, respectively, where *λ* is the wavelength of incident plane wave. The distance between the two probes is 200*λ*. The phantom [Figure 4(a)](#fig4){ref-type="fig"} is discretized on a 128 × 128 Cartesian grid. According to the diffraction limitation \[[@B52], [@B53]\], the spatial sample step *T* or the resolution of the system is set to *λ*/2. It is also necessary to point out that the numbers of iterations for iterative methods are all set to 8; we did not employ error threshold as the iteration criterion, because we want to compare the iterative algorithms after the same iteration numbers. According to the recommendation of Fessler and Sutton \[[@B51]\], the values of *p*, *q* are set to 6 and 2, respectively. The regularization parameters *α* and *β* are set to 0.01 and 0.001, respectively.
4.2. Results {#sec4.2}
------------
[Figure 4](#fig4){ref-type="fig"} shows the reconstructed images through different methods from simulated sparse-view data with no added noise. [Figure 4(a)](#fig4){ref-type="fig"} is the original phantom, [Figure 4(b)](#fig4){ref-type="fig"} is the reconstructed image using bilinear frequency interpolation, [Figure 4(c)](#fig4){ref-type="fig"} is the reconstructed image using the method proposed by Bronstein et al. \[[@B10]\] with broadband incident sound wave, and [Figure 4(d)](#fig4){ref-type="fig"} is the reconstructed image using the proposed method where the sparse transform basis is Haar wavelet. The number of view is 16, and the number of iterations is 8 for Figures [4(c)](#fig4){ref-type="fig"} and [4(d)](#fig4){ref-type="fig"}. Due to sparse-view sampling, the Nyquist-Shannon sampling limitation cannot be satisfied. The reconstructed image of [Figure 4(b)](#fig4){ref-type="fig"} is severely blurred and distorted by interpolation error. The small scale features of original phantom cannot be recognized and larger ones are distorted by ring artifacts. Compared with the interpolation method, the two iterative methods can suppress the artifacts and reduce the noises and Gibbs effect as shown in Figures [4(c)](#fig4){ref-type="fig"} and [4(d)](#fig4){ref-type="fig"}. In [Figure 4(c)](#fig4){ref-type="fig"}, we adopted ten different frequency waves used by Bronstein in \[[@B10]\] to illuminate the object in each view. The image shows oversmoothing effect that the edges and details of the features are blurred. This may be caused by resolving the inverse problem of UDT under overdetermined framework; that is, there are more known *F* than unknown *f* ~*d*~.
[Figure 4(d)](#fig4){ref-type="fig"} shows the result of the proposed method. Most of the features can be clearly represented. The artifacts and oscillation noises are efficiently reduced. For the homogenous background (black region), there is no visible artifact and grayscale aberration, which accords with the characteristic of the original homogenous medium. For the narrow ring region *A*, the outer and inner boundary can be preserved. The inner background *B* is uniformly displayed except for parts of slightly blurred region. *C* is set to be a low-contrast region (the gray intensity is 0.5, while it is 0.6 for background *B*), which can be used to evaluate the sensitivity of imaging methods. In [Figure 4(d)](#fig4){ref-type="fig"}, the region *C* can be clearly distinguished from the surrounding background with a sharp edge definition. Regions *D* and *E* are low and high refractive regions, with the gray intensity 0.3 and 0.75, respectively. The boundary of *D* is preserved and the inner of *D* is clean. The small region *E* can be distinguished from surroundings. It is worth noting that the three small ellipses at region *F* can be identified in [Figure 4(d)](#fig4){ref-type="fig"}, while they are almost invisible in Figures [4(b)](#fig4){ref-type="fig"} and [4(c)](#fig4){ref-type="fig"}. This implies the proposed method can improve the resolution efficiently in sparse-view situation.
[Figure 5](#fig5){ref-type="fig"} shows the magnitude of error in the frequency domains. Since the maximum error of the interpolation method in [Figure 5(a)](#fig5){ref-type="fig"} is 4311, which is much bigger than the corresponding ones of the broadband method in [Figure 5(c)](#fig5){ref-type="fig"} (249.6) and the proposed method in [Figure 5(d)](#fig5){ref-type="fig"} (54.2), we also show the magnitude of error within the interval of \[0,300\] for the interpolation method ([Figure 5(b)](#fig5){ref-type="fig"}). From [Figure 5](#fig5){ref-type="fig"}, we can find that the error of interpolation method around center frequency is significantly larger than the remaining two methods, which accords with [Figure 4(b)](#fig4){ref-type="fig"} that the details of the features are seriously distorted comparing to other two methods. In [Figure 5(c)](#fig5){ref-type="fig"}, the maximum error around center frequency is lower than that of [Figure 5(b)](#fig5){ref-type="fig"}; however, the error of frequencies other than center frequency area is generally higher than the corresponding ones in [Figure 5(b)](#fig5){ref-type="fig"}. This means the noises in the reconstructed image through broadband method are higher than the reconstructed image through interpolation method. [Figure 5(d)](#fig5){ref-type="fig"} shows that the proposed method can markedly reduce the frequency domain error particularly in the low frequency components. This indicates that the reconstructed image can faithfully represent the feature details while efficiently suppressing the noises for sparse-view sampling data. [Table 1](#tab1){ref-type="table"} gives the relative mean square error (RMSE) in frequency domains, which is defined as $$\begin{matrix}
{\text{RMSE} = \frac{\left. ||{F - \hat{F}} \right.||_{2}^{2}}{\left. ||F \right.||_{2}^{2}};} \\
\end{matrix}$$ *F* and $\hat{F}$ are the distribution function of the original phantom and the reconstructed object in frequency domain, respectively. [Table 1](#tab1){ref-type="table"} also lists the structural similarity (SSIM) index for different methods. Compared with RMSE, the SSIM has proven to be consistent with human eye perception \[[@B56]\]. In this paper, the original object ([Figure 4(a)](#fig4){ref-type="fig"}) is the reference image for SSIM. Compared to interpolation method, the proposed method has relatively smaller RMSE and higher SSIM values which coincides with the description above.
In ([15](#EEq15){ref-type="disp-formula"}), the regularization parameters *α*, *β* determine the trade-off between the data consistency and the sparsity of the object. Furthermore, they can be used to adjust the relative weights of the different components in the cost function. [Figure 6](#fig6){ref-type="fig"} shows the images reconstructed under different penalties with 16 views, 128 elements, and pitch of *λ*. [Figure 6(a)](#fig6){ref-type="fig"} is reconstructed with TV and *l* ~1~ norm regularization terms the same as [Figure 4(d)](#fig4){ref-type="fig"}; the regularization parameters are set to 0.01 and 0.001, respectively. [Figure 6(b)](#fig6){ref-type="fig"} is reconstructed with solely TV regularization term; that is, *α* = 0.01, *β* = 0. Compared to [Figure 6(a)](#fig6){ref-type="fig"}, although the TV penalty can preserve the edges of the object, the inner of the object is blurred. [Figure 6(c)](#fig6){ref-type="fig"} is reconstructed with solely *l* ~1~ norm regularization term; that is, *α* = 0, *β* = 0.001. The quality of the image, [Figure 6(c)](#fig6){ref-type="fig"}, for inner region can match [Figure 6(a)](#fig6){ref-type="fig"}, but the edges of the object are distorted. [Figure 6(d)](#fig6){ref-type="fig"} is directly reconstructed from the limited sample data without any penalties; that is, *α* = *β* = 0. Compared to [Figure 6(a)](#fig6){ref-type="fig"}, not only are the qualities of the inner and edge seriously declined in [Figure 6(d)](#fig6){ref-type="fig"}, but also the background is filled with noise. In practical, one can choose the regularization parameters *α*, *β* empirically based on the noise level and image contrast \[[@B57]\] or adaptively adjust the regularization parameters in the process of reconstruction \[[@B58], [@B59]\]. The primary target of this work is to accurately reconstruct the object in sparse-view, the selection of the optimal regularization parameters is not the focus of this paper, but it is worth further investigation.
To fully evaluate the proposed method, we also reconstruct the object with various numbers of views ([Figure 7](#fig7){ref-type="fig"}). In [Figure 7(b)](#fig7){ref-type="fig"}, with 32 views, the smaller scale features (the three small ellipses) can be fully distinguished. Although the quality of reconstructed image for interpolation is also improved visually with the increase of views, the ring artifacts and oscillation noises cannot be eliminated even with 96 views.
5. Discussion {#sec5}
=============
Not only can the sparse-view sampling scheme save the scan time of UDT but also it can reduce the complexity of the imaging device. In [Section 4.2](#sec4.2){ref-type="sec"}, we show the feasibility of image reconstruction for UDT in sparse-view situation based on CS framework. The two main reasons that CS is efficiently employed can be concluded as follows. Firstly, the acoustic indexes of the object are sparsely represented through orthogonal transform and finite difference transform; the former provides sparse representations for rapidly varied regions through wavelet transform, while the latter affords sparse transformation for piecewise smooth regions by TV. Secondly, the acoustic index coefficients with finite main components in the transform domain can be faithfully recovered through the iterative method based on CG and NUFFT.
To full evaluate our method, we will analyze the noise robustness of our method. Besides that, the computational complexity is also discussed in this part.
5.1. Robustness to Noise {#sec5.1}
------------------------
In real situation, in addition to systematic errors such as misaligned transducers, the detected signals contain different kinds of noises such as thermal noise of transducer, electronic noise of amplifier. We added white Gauss noise in the received scatter data to test the robustness to noise of the proposed method. The signal-to-noise ratio (SNR) of simulated noisy data is 20 dB and 10 dB, respectively, where SNR is defined as $$\begin{matrix}
{\text{SNR} = 10\log\left( \frac{\left. ||F \right.||_{2}^{2}}{\left. ||n \right.||_{2}^{2}} \right).} \\
\end{matrix}$$ The pitch and the number of elements are *λ* and 192, respectively.
[Figure 8](#fig8){ref-type="fig"} shows the reconstructed images with none, 20 dB, and 10 dB noise under 16 views and 32 views. The images reconstructed from the signal with an SNR of 20 dB have hardly any difference with the images reconstructed from noise-free signals. The homogenous medium background is not contaminated by noise. The edges of different structures can be distinguished.
However, when the noise increases to 10 dB, the quality of reconstructed image is affected by granulation noise. Although the main features of the object are still visible, the small details are seen to seriously deteriorate, especially for 16 views. How to increase the robustness to noise remains a further research topic.
5.2. Computational Complexity {#sec5.2}
-----------------------------
In the proposed method, the computation time is mainly occupied by NUFFT. For an *N* × *N* digital object, the complexity of NUFFT is *O*(2*qN* ^2^log*N* + *pN* ^2^) \[[@B51]\], where *q* is the oversampling constant and *p* is the number of neighbors for interpolation. For normal reconstruction with the iteration number *m* ~1~ and the linear search times *m* ~2~ for each iteration, the total complexity of our method is estimated as *O*(2*m* ~1~ *m* ~2~(*qN* ^2^log*N* + *pN* ^2^)). For comparison, the theoretical complexity of the frequency domain interpolation requires *O*(*N* ^2^log*N* + 4*N* ^2^). Furthermore, the practical view number is generally more than 4*N* to avoid aliasing. The sample points for every view are at least two times of *N*. The complexity of filtered backpropagation (FBP) is about *O*(*N* ^3^log*N*). For the 128 × 128 tested image used in this paper, the frequency interpolation method requires about 1.80 × 10^6^ operations; spatial domain interpolation method (FBP) would require 14.68 × 10^6^ operations and our method requires about 144.17 × 10^6^ operations.
6. Conclusion {#sec6}
=============
UDT is an important image modality and can afford functional exams in application. CS is one of the most exciting advances in signal theory which takes advantage of compressibility of the object to break Nyquist limitation and recover the major component in transform domain. The paper presents one CS framework for UDT image reconstruction. The numerical experiments show that the proposed method can improve image quality. The relative error is smaller than conventional interpolation method and broadband method. Combining *l* ~1~ norm with TV, not only is the edge of the object preserved, but also the contrast and resolution are improved.
In this paper our effort mainly focuses on integrating CS and UDT to develop practical framework for image reconstruction. Future work will be done on the evaluation of our study with in vivo data. Other important works include how to choose suitable transform basis and how to choose regularization parameters.
This work is partly supported by the National 973 Project of China (Grant no. 2011CB933103), the Project of the National 12th-Five Year Research Program of China (Grant no. 2012BAI13B02), the National Project of Developing Scientific Instrument of China (Grant no. 2013YQ160551), and the Self Innovation Research Fund of Huazhong University of Science and Technology (Grant no. 2013QN089).
Conflict of Interests
=====================
The authors declare that there is no conflict of interests regarding the publication of this paper.
{#fig1}
{#fig2}
{#fig3}
{#fig4}
![Magnitude of error in frequency domain. (a) Interpolation method; (b) error value of interpolation method within \[0, 300\]; (c) broadband method; (d) CS method.](CMMM2014-329350.005){#fig5}
{#fig6}
{#fig7}
{#fig8}
######
SSIM and RMSE for different methods.
Interpolation Broadband signal CS
-------------------------- --------------- ------------------ -------
SSIM 0.291 0.346 0.820
RMSE in frequency domain 0.765 0.352 0.255
[^1]: Academic Editor: William Crum
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#sec1}
============
Gene therapy treats disease by using genetic material to change or correct gene expression to cure disease. The concept of gene therapy emerged in the early 1970s as a consequence of both increased understanding of the role of genes in human disease and the development of genetic engineering techniques.[@bib1] First attempts to treat human patients with exogenous genes were conducted in the 1980s and 1990s; however, safety issues hampered their progress.[@bib2], [@bib3] Extensive research in recent years has yielded better and safer gene delivery strategies that reduce vector immunogenicity; consequently, the gene therapy field is flourishing again, with exciting successes in treating rare genetic diseases, including a first US Food and Drug Administration (FDA)-approved gene therapy for inherited retinal dystrophy,[@bib4] and thousands of gene therapy clinical trials worldwide.[@bib5] In pursuit of improved treatments for cardiovascular disease, including complications of myocardial infarction (MI) and heart failure (HF), extensive gene therapy preclinical research resulted in numerous clinical trials aiming to promote angiogenesis[@bib6] and improve calcium homeostasis.[@bib7] Although proving safety, all cardiac gene therapy trials failed to demonstrate significant therapeutic efficacy.[@bib7] It has been suggested that inefficient gene transfer that resulted in poor expression of the target gene is a potential reason for these neutral results.[@bib7]
Gene transfer can be achieved by using viral vectors including adenovirus, retrovirus, adeno-associated virus (AAV), lentivirus, vaccinia virus, poxvirus, and herpes simplex virus or non-viral vectors including naked plasmid DNA (pDNA) or lipofection.[@bib5] Modified RNA (modRNA) is a novel vector for gene transfer to both dividing and non-dividing mammalian cells that mediate fast, robust, transient expression of proteins in the targeted cells or tissue.[@bib8] Pre-clinical attempts to use mRNA as a gene transfer vector showed successful expression of the encoded proteins[@bib9] and improved symptoms in a rodent model of diabetes insipidus.[@bib10] Yet the high immunogenicity of exogenous mRNA[@bib11], [@bib12], [@bib13], [@bib14], [@bib15] and high susceptibility to degradation by RNase have hampered progress in developing mRNA as a vector for treating humans.[@bib16], [@bib17] Since the first description of a biologically active and translatable *in vitro* transcription (IVT) product by Krieg and Melton[@bib18] in 1984, however, immense technological advances have led to synthetic mRNA's emergence as a promising vector for gene delivery.[@bib19] Two important milestones in developing IVT technology turned synthetic mRNA from a scientific tool to a potential platform for gene therapy. First, pioneering work by Karikó et al.[@bib20] demonstrated that incorporating chemically modified nucleotides in the synthetic mRNA results in a more stable and less immunogenic mRNA that mediates rapid, high expression of the encoded protein.[@bib21], [@bib22] Second, the discovery of a stable cap analog, anti-reverse cap analog (ARCA), increases synthetic mRNA's stability and translatability.[@bib23]
We and others have recently shown that modRNA can be used to express reporter proteins in rodent and porcine myocardium,[@bib8], [@bib24], [@bib25] and that functional genes can be used to promote angiogenesis,[@bib26] cardiomyocyte proliferation,[@bib27] and cardiac cell survival.[@bib28] Previously, we described a detailed protocol for synthesizing modRNA capped with ARCA 3′-O-Me-m7G(5′)ppp(5′)G and substituting 100% cytidine (C) with 5-methylcytidine (m5C) and 100% uridine (U) with pseudouridine (Ψ),[@bib29] or substituting 100% uridine with N1-methyl-pseudouridine (N1mΨ), which has been shown to be superior over the combination of Ψ and m5C.[@bib8], [@bib30] Here we describe an optimized protocol that indicates that a change in ARCA and N1mΨ ratios leads to cost-effective modRNA with higher protein expression and lower immunogenicity *in vitro*.
Results {#sec2}
=======
modRNA Yield and Quality {#sec2.1}
------------------------
Currently, there are two strategies for mRNA capping in an IVT reaction: (1) post-transcriptional capping and (2) chemical capping. In the post-transcriptional approach, the vaccinia virus capping enzyme is used to cap the enzyme after the mRNA synthesis is completed.[@bib31], [@bib32] Although effective, this procedure is time-consuming and expensive. In chemical capping, a cap analog is added directly to the IVT reaction, after which the cap structure is incorporated at the 5′ end of the mRNA. To achieve a high percentage of capped mRNA using this method, the ratio between the concentrations of the cap analog and the guanosine triphosphate (GTP) in the reaction are kept high. For IVT reactions, the ratio of ARCA to GTP should always be 1:3.7. In order to reduce the synthetic mRNA's immunogenicity and increase its translatability, we replace 100% U with N1mΨ. Previously, for the IVT reaction, we used a composition of 5 nM ARCA, 1.35 mM GTP, 7.5 mM N1mΨTP, and 7.5 mM MTP (composition 1 in [Table 1](#tbl1){ref-type="table"}, hereafter termed ARCA 5 protocol). To achieve our aim of increasing the translation capacity and lowering the cost of modRNA production, we reduce the amount of template DNA by 85% ([Figure S1](#mmc1){ref-type="supplementary-material"}). We then explored different nucleotide compositions and desalting methods for making and cleaning the modRNA ([Table 1](#tbl1){ref-type="table"}). Out of the various nucleotide stoichiometries tested, composition 5, comprising higher concentrations of ARCA, GTP, ATP, and cytidine triphosphate (CTP) and less N1mΨTP, another costly element in modRNA production, resulted in the greatest modRNA yield compared with the previously used protocol (ARCA 5 protocol), keeping the amount of T7 RNA Polymerase consistent ([Table 1](#tbl1){ref-type="table"}). To desalt the modRNA in the first cleaning process, we compared the MEGAclear RNA purification kit (used previously) with the more cost-efficient Amicon filter. We found that the nuclear GFP (nGFP) modRNA yield using the Amicon filter was equivalent to the amount generated by the MEGAclear RNA purification kit. Thus, we established that the most cost-effective protocol for IVT reaction uses nucleotide concentrations of 10 mM ARCA, 2.7 mM GTP, 8.1 mM ATP, 8.1 mM CTP, and 2.7 mM N1mΨTP followed by desalting and nucleotide clearing using the Amicon filter (this protocol will now be referred to as the ARCA 10 protocol).Table 1Nucleotide Composition and Cleaning MethodsNucleotide Composition 1 (ARCA 5)Nucleotide Composition 2Nucleotide Composition 3Nucleotide Composition 4Nucleotide Composition 5 (ARCA 10)Nucleotide Composition 6ARCA5 mM10 mM10 mM10 mM10 mM7.5 mMGTP1.35 mM2.7 mM2.7 mM2.7 mM2.7 mM2.25 mMATP7.5 mM8.1 mM15 mM15 mM8.1 mM10 mMCTP7.5 mM2.7 mM15 mM15 mM8.1 mM10 mMN1mΨTP7.5 mM2.7 mM15 mM2.7 mM2.7 mM2.25 mMTotal28.85 mM26.2 mM57.7 mM45.4 mM31.6 mM32 mMDesalting with MEGAclearTotal mg nGFP modRNA per 1-mL reaction1.333.121.192.132.941.93% of ARCA 5 protocol10023589160221145Desalting with Amicon FilterTotal mg nGFP modRNA per 1-mL reaction2.121.041.922.941.74% of ARCA 5 protocol15978144221131[^2]
We also evaluated modRNA quality using a bioanalyzer. We found no differences in the integrity of the modRNA generated using the ARCA 10 protocol compared with that produced via the ARCA 5 protocol ([Figures 1](#fig1){ref-type="fig"}A and 1B). Interestingly, when using different nucleotide compositions in the reaction, we observed high levels of uncompleted IVT products ([Figure 1](#fig1){ref-type="fig"}F).Figure 1Effect of Nucleotide Composition in IVT Reaction on modRNA Integrity(A--F) modRNA integrity was evaluated using a bioanalyzer (Tap Station 2200) of compositions 1 (A), 2 (B), 3 (C), 4 (D), 5 (E), and 6 (F).
Translational Capacity of modRNA *In Vitro* {#sec2.2}
-------------------------------------------
Next, we analyzed the translational capacity of modRNA generated with the ARCA 10 protocol by transfecting several human cell lines and primary cardiac cells isolated from neonatal rat hearts and measuring the expression levels of two different reporter genes: nGFP and firefly luciferase (Luc) ([Figure 2](#fig2){ref-type="fig"}). We observed significantly increased nGFP expression in HeLa cells (data not shown), primary rat cardiac cells, and neonatal rat cardiomyocytes (nrCM) transfected with nGFP modRNA generated with the ARCA 10 protocol compared with the ARCA 5 protocol ([Figures 2](#fig2){ref-type="fig"}C and 2D). Similar results were observed in primary rat cardiac cells, nrCM, human umbilical vein endothelial cells (HUVECs), HeLa cells, and HEK293 cells when transfected with Luc modRNA ([Figures 2](#fig2){ref-type="fig"}E--2H). We did not observe any difference in cell death levels post transfection *in vitro* as measured by DAPI or propidium iodide (PI)-positive cells (data not shown).Figure 2Effect of Nucleotide Composition in IVT Reaction on Gene Expression in Human Cell Lines and Rat Cardiac Cells(A) GFP expression in rat cardiac cells 18 h post transfection with or without nGFP modRNA generated with the ARCA 5 protocol and ARCA 10 protocol. (B) Representative image of Luc activity level in neonatal rat cardiomyocytes 24 h post transfection with or without Luc modRNA generated using the ARCA 5 protocol and ARCA 10 protocol. (C and D) Quantification of GFP expression in neonatal rat cardiac cells (C) or cardiomyocytes (D) 18 h post transfection with nGFP modRNA generated using the ARCA 5 protocol and ARCA 10 protocol. (E--H) Quantification of Luc activity level in HUVECs (E), rat neonatal cardiomyocytes (F), HeLa (G), or Hek293 (H) cells 18 h post transfection with or without Luc modRNA generated using the ARCA 5 protocol and ARCA 10 protocol. One-way ANOVA, Tukey's multiple comparison test; \*\*\*p \< 0.001, \*\*p \< 0.01, \*p \< 0.05. n = 5 (A and B); n = 4 (C--F).
Immunogenicity of modRNA {#sec2.3}
------------------------
Evaluating immunogenicity is critical to defining the safety profile of mRNA therapeutics and their use in humans. To evaluate the innate immune response *in vitro*, we measured the expression levels of three innate immune-related genes, retinoic acid-inducible gene I (RIG1), interferon alpha (IFNα), and interferon beta (IFNβ), in nrCM 8 h post transfection with Luc modRNA (made by using either the ARCA 5 protocol or the ARCA 10 protocol) or unmodified mRNA ([Figure 3](#fig3){ref-type="fig"}). Importantly, Luc modRNA made via both the ARCA 5 and ARCA 10 protocols has significantly lower levels of RIG1, IFNα, and IFNβ when compared with unmodified mRNA ([Figures 3](#fig3){ref-type="fig"}B--3D). However, Luc made using the ARCA 5 protocol was significantly more immunogenic than Luc modRNA made using the ARCA 10 protocol ([Figures 3](#fig3){ref-type="fig"}B--3D).Figure 3Immunogenicity of modRNA Compared with Unmodified mRNA(A) Luc mRNA or modRNA transfected into neonatal rat cardiomyocytes 8 h after transfection cells were collected. qPCR was performed to evaluate the gene expression of innate immune response markers. (B--D) Expression of innate immune response markers: RIG1 (B), IFNα (C), or IFNβ (D) 8 h post transfecting unmodified mRNA, Luc modRNA (ARCA 5 protocol), or Luc modRNA (ARCA 10 protocol). One-way ANOVA, Tukey's multiple comparison test; \*\*\*\*p \< 0.0001. n = 5.
Discussion {#sec3}
==========
Synthetic modRNA is now emerging as a promising vector for gene delivery.[@bib19] However, synthesizing modRNA in the high quantities required by pre-clinical and clinical trials is expensive and time-consuming. In light of recent successes in promoting angiogenesis,[@bib26] CM proliferation,[@bib27] and cardiac cell survival[@bib28] in rodents, there in an urgent need to reduce the cost and improve the quality of synthesizing modRNA.
At present, our capping strategy is chemical capping; in this capping method, the IVT reaction includes a cap analog, resulting in a cap structure at the mRNA's 5′ end. This capping method requires a high ratio of cap analog and GTP concentrations to produce a high percentage of capped mRNA.
This study aimed to optimize the protocol for synthesizing mRNA using ARCA and N1mΨ. Previously published protocols[@bib8], [@bib26], [@bib28], [@bib29] maintained the correct ratio between ARCA and GTP by keeping the GTP concentration low (1.35 mM) and the nucleotides high (7.5 mM). In this composition, the GTP becomes the limiting factor in the reaction, leading the reaction to terminate without incorporating the other nucleotides present in excess. The unincorporated nucleotides are later washed away during the cleaning process. Although the ratio between ARCA and GTP must be kept high, we realized that the composition of the other nucleotides in the reaction can be modified, leading us to explore the effect of different nucleotide compositions on the final yield and quality of the modRNA. Interestingly, we found that when doubling the concentration of all nucleotides to increase the final concentration to 57.7 mM, the final yield is lower than the original concentration of 28.85 mM. The optimal final nucleotide concentration in our study was 31.6 mM (ARCA 10 mM, GTP 2.7 mM, CTP 8.1 mM, ATP 8.1 mM, and N1mΨ 2.7 mM). Using this concentration, we were able to increase the final modRNA yield by 290%. Combining this protocol with improved cleaning methods results in modRNA with increases translatability in both human cell lines and primary cardiac cells ([Figure 2](#fig2){ref-type="fig"}), as well as reduced immunogenicity ([Figure 3](#fig3){ref-type="fig"}). A possible explanation for this result is that optimizing nucleotide composition in the IVT reaction raises the percentage of successfully capped modRNA.
Here we developed an improved protocol for large-scale production of modRNA. By increasing the final nucleotide concentrations while reducing the N1mΨ molarity and using the same amount of T7 RNA polymerase, we established a new method that is more cost-effective than the previously published approach.[@bib29] Our novel protocol generated improved modRNA with higher translatability and lower immunogenicity *in vitro*, and will allow basic and translational research labs to overcome the high cost of modRNA preparation and attain more effective modRNA for their research.
Materials and Methods {#sec4}
=====================
Synthesizing modRNA {#sec4.1}
-------------------
In brief, clean PCR products generated with plasmid templates purchased from GenScript were used as the template for mRNA. modRNAs were generated by transcription *in vitro* with a customized ribonucleoside blend of ARCA; 30-O-Me-m7G(50) ppp(50)G (catalog no. \[cat. \#\] N-1081; Trilink Biotechnologies); GTP (cat. \#am1334-5; Life Technologies); ATP (cat. \#am1334-5; Life Technologies); cytidine triphosphate (cat. \#am1334-5; Life Technologies), and N1-methylpseudouridine-50-triphosphate (cat. \#N-1081; Trilink Biotechnologies). The mRNA was purified with the MEGAclear kit (cat. \#AM1908; Life Technologies) according to the manufacturer's instructions or using Amicon Ultra-4 Centrifugal Filter Unit 4 mL,10 kDa (cat. \#UFC801024; MilliporeSigma) and treated with Antarctic Phosphatase (cat. \#M0289L; NEB). It was then re-purified with the MEGAclear kit. The mRNA was quantified using a NanoDrop spectrometer (Thermo Scientific), precipitated with ethanol and ammonium acetate, and re-suspended in 10 mM Tris-HCl and 1 mM EDTA. The open reading frame for the modRNA used is listed in [Table S1](#mmc1){ref-type="supplementary-material"}. This protocol has been described in detail elsewhere.[@bib29]
Desalting and Nucleotide Clearing {#sec4.2}
---------------------------------
Amicon Ultra-4 Centrifugal Filter Unit 4 mL, 10 kDa (cat. \#UFC801024; MilliporeSigma) was washed with 4 mL nuclease-free water two times. A total of 400 μL IVT reaction was loaded to the filter and diluted with 3.5 mL nuclease-free water. The filter was centrifuged at 3,200 × *g* for 20 min until the volume reduces to 100 μL. A total of 3.9 mL water was added to the filter and centrifuged at 3,200 × *g* for 20 min two times. The purified modRNA was incubated at 70°C for 10 min (to inactivate T7).
Neonatal Rat Cardiac Cells Isolation {#sec4.3}
------------------------------------
Neonatal rat ventricular CMs were isolated from 3- to 4-day-old Sprague-Dawley rats (Jackson ImmunoResearch Laboratories) by multiple rounds of digestion with 0.1% collagenase II (Invitrogen) in PBS. After each digestion, the supernatant was collected in horse serum (Invitrogen). The total cell suspension was centrifuged at 300 × *g* for 5 min. The supernatants were discarded and cells were resuspended in DMEM (GIBCO) supplemented with 0.1 mM ascorbic acid (Sigma), 0.5% insulin-transferrin-selenium (100×), penicillin (100 U/mL), and streptomycin (100 mg/mL). Cells were plated in plastic culture dishes for 90 min until most of the non-myocytes were attached to the dish while the myocytes remained in suspension. Myocytes were then used to seed 24-well plates at a density of 1 × 10^5^ cells/well. nrCMs were incubated for 48 h in DMEM supplemented with 5% horse serum before transfection with modRNAs. Twenty-four hours post isolation, the non-myocytes were used to seed 24-well plates at a density of 1 × 10^5^ cells/well. The non-myocytes were transfected with modRNA 24 h post plating.
*In Vitro* Transfection with modRNA {#sec4.4}
-----------------------------------
Using a 24-well plate, we complexed 2.5 μg/well mRNA with Lipofectamine RNAiMAX Transfection Reagent (cat. \#13778030; Life Technologies) and used the resulting complex to transfect either neonatal rat cardiac cells or human cell lines according to the manufacturer's instructions.
Assessing Number of nGFP-Expressing Cells {#sec4.5}
-----------------------------------------
Eighteen hours post transfection, cells were imaged using a Zeiss Slide Scanner Axio Scan. Quantification of nGFP-positive cells was performed using ImageJ software.
Detection of Luciferase Activity {#sec4.6}
--------------------------------
Bioluminescence of the transfected cells was measured 18 h post transfection. Each unit of Luc signal represents p/s/cm^2^/sr × 10^6^. Luciferin (cat. \#L9504; Sigma) (1.5 mg/mL culture media) was added to cell culture 10 s before imaging. Cells were imaged using an IVIS100 charge-coupled device imaging system every 20 s until the Luc signal reached a plateau. Imaging data were analyzed and quantified with Living Image software.
qRT-PCR {#sec4.7}
-------
Total RNA was reverse transcribed with Superscript III reverse transcriptase (cat. \#LS18080044; Invitrogen), according to the manufacturer's instructions. qRT-PCR analyses were performed on a Mastercycler Realplex 4 Sequence Detector (Eppendorf) with the HotStart-IT SYBR Green qPCR master mix (×2) (cat. \#75762; Affymetrix). Data were normalized relative to GAPDH. Fold-changes in gene expression were determined by the delta-delta-Ct (ddCT) method. The PCR primer sequences used are listed in Table S2.
Statistical Analyses {#sec4.8}
--------------------
Statistical analyses were performed with GraphPad Prism software. Values are reported as means ± SD. Two-tailed Student's t tests (\*p \< 0.05 considered significant) or one-way ANOVA with Bonferroni correction (\*p \< 0.05 considered significant) were used for comparisons between groups.
Author Contributions {#appsec3}
====================
Y.H. designed most of the experiments, performed experiments, analyzed most of the data, and wrote the manuscript. N.S. performed most of the experiments, analyzed data, and revised the manuscript. E.Y. performed experiments. M.T.K.S. prepared modRNAs and performed experiments. K.K. revised the manuscript. E.E. performed CM isolation method. L.Z. designed experiments, analyzed data, and revised the manuscript.
Supplemental Information {#appsec2}
========================
Document S1. Figure S1 and Table S1Document S2. Article plus Supplemental Information
This work was funded by a cardiology start-up grant awarded to the Zangi laboratory and also by NIH grant R01 HL142768-01.
Supplemental Information can be found online at <https://doi.org/10.1016/j.omtm.2019.07.006>.
[^1]: These authors contributed equally to this work.
[^2]: Summary of nucleotide composition and desalting (first cleansing) methods used in this study and their effects on the final modRNA yield. For modRNA integrity, see [Figure 1](#fig1){ref-type="fig"}.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .120}
{#sp2 .121}
{#f1 .120}
{#f2 .121}
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#sec1}
============
Antisocial behavior is mostly regarded as consisting of attitudes that go against the grain in society and/or disregard the rights of others ([@ref20]). Antisocial behavior can be harmful to others or their homes and possessions. Adolescents, their family and society can be burdened with substantial costs in regard to physical harm and emotional duress, with financial hardship as well ([@ref5]; [@ref34]). Antisocial behavior has become a critical concern for public health worldwide, and it is especially prevalent in youths. Antisocial behavior tendencies are seen primarily in adolescence with a high incidence ([@ref2]), so this stage in life has become very interesting for the study of this deviant behavior. A tremendous portion of adolescents display some type of antisocial behavior ([@ref40]). Children who show antisocial behavior tendencies are prone to a number of detrimental issues in their development, including dropping out of school, criminal activity, psychological problems, and substance abuse ([@ref38]; [@ref36]). Antisocial behavior tendencies in the young are known to have negative consequences, so determining the factors associated with the development of antisocial behavior in young people is still a hot research topic. There is a need to elucidate the risk factors and related mechanisms of antisocial behavior tendencies in adolescents to develop prevention as well as intervention measures.
A positive life environment is a protective factor against developing a problem behavior and its persistence as well. On the other hand, negative environmental factors can induce an antisocial behavior, where they are considered to be strong predictors of such behavior in youths ([@ref37]; [@ref22]). Many studies indicate that adverse events in youth's formative years have long-term physical and mental health consequences ([@ref24]; [@ref10]). Exposure to adversity during development leads to, among other issues, a substantial increase in the probability of developing antisocial behavior ([@ref18]). Therefore, researchers are looking closely at how to mitigate the detrimental effects of negative life events on antisocial behavior tendencies in these adolescents. Antisocial behavior tendencies in adolescence, and its continuation in later life, are believed to involve an interaction of psychological vulnerability and environmental factors ([@ref1]). Regarding psychological resources, self-esteem is thought to be an important internal resource, which has a significant effect in promoting children's development ([@ref9]). Self-esteem is usually defined as an emotional evaluation of the individual himself/herself ([@ref27]). Individuals who have high self-esteem usually have enough resources to cope with daily stresses. They are more confident in life and will face difficulties bravely, which makes them less susceptible to emotional depletion and personality problems. Earlier studies showed that self-esteem is negatively associated with antisocial behavior ([@ref3]; [@ref15]), indicating that high self-esteem is less likely to lead to antisocial behavior tendencies. Besides, previous studies have confirmed the mediating role of self-esteem between negative life events and negative emotions ([@ref11]; [@ref33]), as well as a mediating role between perceived parenting and antisocial behavior tendencies ([@ref14]). Both the correlation of negative life events with self-esteem and the correlation of self-esteem with antisocial behavior tendencies have been studied. However, the comprehensive relationship of negative life events, self-esteem, and antisocial behavior tendencies has not yet been studied, and the mediating role of self-esteem between negative life events and antisocial behavior tendencies has not yet been confirmed. We may not be able to change the negative life events that the individual has experienced in a short time. However, if there is evidence demonstrating the mediation effect of self-esteem between negative life events and antisocial behavior tendencies, increasing self-esteem among adolescents could prevent the development of antisocial behavior, since the promoting effect of negative life events on antisocial behavior tendencies could be offset to some extent. Importantly, self-esteem is not a static phenomenon; rather, it is a dynamic process that can be altered at any time. Previous studies have confirmed that self-esteem could be improved through interventions ([@ref25]; [@ref46]). Accordingly, it will be of important theoretical and practical value to lessen the harmful effect of negative life events on antisocial behavior tendencies in adolescents, once the mediating role of self-esteem in the relationship between negative life events and antisocial behavior tendencies is established.
The effect of negative life events on adolescent antisocial behavior tendencies maybe mediated by self-esteem, and adolescents are sensitive to adverse events in different ways. According to the social control theory, an individual's connection with society is critical in determining their activities ([@ref42]). Adolescents' relationship with family is important in preventing them from getting involved with delinquent activity. Prior research indicates that family support has positive effects on children ([@ref35]), which is inversely associated with antisocial behavior developing in youths ([@ref28]; [@ref7]). Previous studies have made valuable contributions. However, there are few studies with large samples that take a look at how family support influences the effect of adversity events on antisocial behavior tendencies in adolescents. Previous research has indicated that there is evidence that family support may reduce stress among youths ([@ref43]; [@ref4]), but little is known about how family support may act as a protective factor against antisocial behavior tendencies. Based on a functional model of social support processes ([@ref43]), family support allows adolescents to cope more with the usual challenges in life, especially subjective family support of adolescents. Subjective family support may be critical in improving adolescents' ability to deal with various life stressors ([@ref4]; [@ref19]), which may lessen the adverse effect of negative life events on self-esteem and social behavior. In current research, we predicted that subjective family support can have a stress-moderating effect in relation to self-esteem and then deter antisocial behavior tendencies from developing. The specific prediction is an interaction of negative life events and family support, such that the effect of negative life events on self-esteem is mitigated or abolished for adolescents with greater subjective family support. Since self-esteem may mediate the effect of negative life events on antisocial behavior tendencies, this study tests a hypothesis that the indirect association between negative life events and antisocial behavior tendencies *via* self-esteem will be moderated by family support.
According to the stress-support model and our literature review, antisocial behavior tendencies depend on the interaction of opposing factors that promote vulnerability or protection. In our study, the main factor predisposing susceptibility was negative life events, which can pose an increased risk to developing antisocial behavior in adolescents. The protective factors were family support and self-esteem, which could have an increased favorable influence, resulting in a lower risk of antisocial behavior development. The stress-support model specifically predicts that antisocial behavior tendencies are based on the interplay of factors that promote vulnerability or protection. The aim of our study was: (a) to see if self-esteem influences how negative life events affect adolescent antisocial behavior tendencies and (b) to determine if self-esteem mediation of the indirect association between negative life events and antisocial behavior tendencies is moderated by family support. A moderated mediation model based on these two research questions was thus investigated ([Figure 1](#fig1){ref-type="fig"}). Therefore, we proposed the hypotheses below:
1. Hypothesis 1: Self-esteem will mediate the link between negative life events and antisocial behavior tendencies in adolescents.
2. Hypothesis 2: Family support will moderate the mediation effect of self-esteem between negative life events and adolescent antisocial behavior tendencies.
{#fig1}
Materials and Methods {#sec2}
=====================
Participants and Procedures {#sec3}
---------------------------
We selected 8,958 adolescents (6,002 middle school students and 2,956 high school students) from Sichuan, Henan, and Shandong provinces in China. First, we randomly selected one province from Eastern China (Shandong province), Central China (Henan province), and Western China (Sichuan province). Next, we randomly selected one city (Nanchong in Sichuan, Zhoukou in Henan, and Heze in Shandong) from each province. Finally, we choose two rural middle schools and one high school randomly in each city sampled. All students in the selected schools were participants of this cross-sectional study. Each student attended a 30-min session to complete the measures in their respective classrooms alone in the absence of teachers. This study was carried out in accordance with the Declaration of Helsinki. Approval to conduct the study was granted by the research ethics committee at Jinzhou Medical University \[No: 2016-08\]. Participation was voluntary. Voluntary informed consent was obtained from the participants, their parents, or legal guardians, and the principals of the sampled schools before the administration of any study-related questionnaires. We clearly stated that all their responses were collected anonymously and kept confidential, only to be used for research. They were allowed to drop out from the study whenever.
A total of 9,675 students were enrolled in this investigation, with 8,958 returning a complete and qualified questionnaire. Finally, 8,958 students were qualified adolescents in the present study, with 2,661 being from Sichuan, 4,353 from Henan, and 1,944 from Shandong.
Measures {#sec4}
--------
### Negative Life Events {#sec5}
Negative life events were determined with the Adolescent Self-Rating Life Events Check List (ASLEC) ([@ref44]). The scale evaluates the stressful life events that an individual has lived through in the last 6 months, in terms of frequency and intensity. This is a six-point Likert-type scale comprising 27 items concerning five dimensions, including interpersonal relationship, study pressure, punishment, bereavement, and change for adaptation. An example item of interpersonal relationship stressful events was "Dispute with a classmate or close friend." An example item of study pressure events was "Failed or did poorly in the exam." An example item of punishment events was "Beaten and abused by parents or others." An example item of bereavement events was "The death of a close relative or friend." An example item of adaptation stressful events was "The daily routine (diet, rest, etc.) changed significantly." The questions are answered according to the following scale: 1 (did not occur), 2 (no effect), 3 (mild effect), 4 (moderate effect), 5 (severe effect), and 6 (very severe effect), where the scores were 0 (no occurrence), 0 (had no effect), 1 (mild effect), 2 (moderate effect), and 3 (severe effect) to 4 (very severe effect). A higher score indicated more stressful negative life event. A confirmatory factor analysis (CFA) was performed to confirm its factor structure. Since the scale comprised too many original items, item parceling was conducted, which was a recommended method to deal with the long scale ([@ref45]). Thus, negative life events comprised five dimensions as their five indicators. The CFA model demonstrated a satisfactory fit \[*χ*^2^/*df* = 20.487, *p* \< 0.001, CFI = 0.947, TLI = 0.942, SRMR = 0.061, RMSEA = 0.059 (90% CI: 0.046, 0.068)\], showing that this scale was suitable for the sample studied. The composite reliability of the assessment tool (0.7903), along with Cronbach's alpha (0.918) and average variance extracted (0.5631), suggested that the internal quality of this model was satisfactory.
### Self-Esteem {#sec6}
Five items were selected from the Rosenberg self-esteem scale to assess self-esteem in the present sample ([@ref32]). An example item was "I think I have a lot of strengths." There are four possible answers for each item indicating the degree of agreement, where 1 indicates strongly disagree to 4 indicating strongly agree, and this scale has been validated in adolescents to assess their self-confidence and self-satisfaction ([@ref41]; [@ref30]). The final score for the five items determines the self-esteem level, where a higher level of self-esteem is indicated by a higher score. This scale showed CFA fit indices indicating that the fit was acceptable \[*χ*^2^/*df* = 27.886, *p* \< 0.0001, CFI = 0.965, TLI = 0.926, SRMR = 0.055, and RMSEA = 0.046 (90% CI: 0.037, 0.058)\]. Cronbach's alpha (0.905), composite reliability (0.7346), and average variance extracted (0.4861) were also acceptable, demonstrating that the internal quality of the model was suitable.
### Family Support {#sec7}
The family support scale was selected from the Resilience Scale for Chinese Adolescent (RSCA) ([@ref13]), which was one of the dimension of RSCA to assess one's perception of family support by examining how he or she regards the supports from their parents and other family members. It comprises six Likert-type items. Examples of item were "My parents always encourage me to do my best," "My parents respect my opinion very much," etc. A five-point Likert-type scale is used to score each item, where 1 indicates completely unmatched and 5 indicates completely matching, according to the frequency of feeling satisfied with each item. High scores of the scale represent high levels of family support. The reliability and validity of this scale were acceptable in our sample according to Cronbach's alpha (0.931), composite reliability (0.679), and average variance extracted (0.452). Additionally, the CFA model was a good fit model, where the fit indices of the scale were as follows: *χ*^2^/*df* = 8.249, *p* \< 0.0001, CFI = 0.964, TLI = 0.959, SRMR = 0.049, and RMSEA = 0.043 (90% CI: 0.036, 0.054).
### Antisocial Behavior Tendencies {#sec8}
Antisocial behavior tendencies were measured using six Likert-type items, which were components of the social adjustment scale for adolescents ([@ref47]). The scale evaluates a sort of deviate behaviors that the participant has experienced, such as fighting, violating school rules, ignoring social norms, etc. An example of the item was "I often fight with others." Scores go from 1 (strongly disagree) to 5 (strongly agree) Likert scale, according to the degree of agreement with which the respondent has experienced. A higher score indicates a higher level of antisocial behavior tendencies. The fit indices of the CFA model \[*χ*^2^*/df* = 8.262, *p* \< 0.0001, CFI = 0.968, TLI = 0.943, SRMR = 0.064, and RMSEA = 0.053 (90% CI: 0.042, 0.060)\] suggested that the scale was acceptable in the current study. On the basis of composite reliability (0.751) and the average variance extracted (0.394), together with Cronbach's alpha (0.932), we concluded that the sub-scale had an acceptable reliability and validity in our sample.
Statistical Analysis {#sec9}
--------------------
SPSS 21.0 software package and Mplus version 7.4 were used to analyze the data. The multiple imputation method was used to deal with missing data, while the percentage of missing data of 8,958 qualified questionnaires was 7.23%. In the preliminary analysis phase, the reliability and validity of the scales were determined using Cronbach's alpha, composite reliability, and average variance extracted after CFA, which was conducted in Mplus 7.4. Cronbach's alpha was greater than 0.90, indicating good reliability ([@ref17]). Composite reliability was greater than 0.60, and average variance extracted was greater than 0.50, indicating that each measurement construct had great convergence validity, where an acceptable average variance extracted can be between 0.36 and 0.50 ([@ref8]). The overall fit of CFA was assessed using *χ*^2^/*df*, the comparative fit index (CFI), Tucker-Lewis index (TLI), the standardized root mean square residual (SRMR), the root mean square error of approximation (RMSEA), and the 90% confidence interval (CI) of RMSEA. SRMR and RMSEA ranging from 0 to 1 with a 0.08 cut-off for adequacy and a 0.05 cut-off being a more stringent value for goodness-of-fit, while 0.90 or greater suggesting a good fit for all the other indices ([@ref39]; [@ref21]).
On the basis of the theoretical hypothesis model ([Figure 1](#fig1){ref-type="fig"}), a structural equation model (SEM) was used to determine the moderated mediation association between negative life events, family support, self-esteem, and antisocial behavior tendencies. Considering the categorical nature and the non-normal distribution of the descriptive sat for the questionnaire items, robust maximum likelihood estimator (MLR) was performed to determine SEM in Mplus 7.4, which could provide values for the observed variables with SEs showing robustness to non-normality ([@ref23]). The same fit indices as CFA models were used to evaluate the goodness-of-fit of SEM. The hypothesized model was tested in two steps. First, the mediation model without the hypothesized interaction was assessed. The mediation of self-esteem was indicated if the indirect effect of negative life events on antisocial behavior tendencies was significant. The effect size of mediation model was measured using the traditional effect size measures (*R~M~* and *P~M~*) together with the standardized indirect effect, where *R~M~* is the indirect effect to direct effect ratio and *P~M~* is the indirect effect to the total effect ratio ([@ref29]). Next, the moderated mediation model with latent interaction was tested; while the latent interaction of family support and negative life events was created using latent moderated structural equations (LMS) in Mplus 7.4, after standardized family support and negative life events. A significant moderated mediation effect of family support could be established if the 95% CI of the interaction did not contain 0. To properly interpret it, the conditional indirect effect of negative life events on antisocial behavior tendencies *via* self-esteem was plotted for low, medium, and high family support values (high family support: values above mean + 1 SD; medium family support: values of mean ± SD; low family support: values below mean − 1 SD) ([@ref12]). The index of moderated mediation effect was used to measure the effect size of the moderated mediation model. A two-tailed *p* \< 0.05 was considered statistically significant. In addition, age, gender, study site (Sichuan/Henan/Shandong), school type (middle school/high school), whether or not live with parents (both parents/mother only/father only/neither), and education level of parents were controlled for all SEM models, while we analyzed the mediation effect of self-esteem and the moderated mediation effect of family support.
Results {#sec10}
=======
The Basic Characteristics of Subjects {#sec11}
-------------------------------------
The present study included 8,958 adolescents, of which 49.18% were males and 50.82% were females. The participants were from Henan province (48.59%), Sichuan province (29.71%), and Shandong province (21.70%). They were 10--17 years old \[mean = 15.53 ± 2.23 (SD)\]. A third of the participants (33.00%) were selected from high schools and 67.00% were from middle schools. The education level of the parents of participants was generally very low, where 54.66% of the fathers just had a middle school education and over half (53.18%) of the mothers only a primary school education or less. Around 47.35% of the adolescent lived with both parents in the past 6 months and 36.88% lived with neither of parents, while 13.69% lived with mother only and 2.08% with father only.
Correlation Analysis Between Negative Life Events, Family Support, Self-Esteem, and Antisocial Behavior Tendencies {#sec12}
------------------------------------------------------------------------------------------------------------------
Pearson correlations and partial correlations were performed to analyze the correlation between two studied variables, while partial correlations were conducted after controlling for demographic characteristics (age, gender, study site, school type, whether or not live with parents, and education level of parents), and other studied variables. The results are tabulated in [Table 1](#tab1){ref-type="table"}. Antisocial behavior tendencies of adolescents were positively related to negative life events, while inversely related to self-esteem and family support. The relationship between self-esteem and family support was also positive, while the variable negative life events were significantly and inversely related to self-esteem and family support.
######
Correlation analysis between negative life events, self-esteem, family support, and antisocial behavior tendencies.
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Variable 1\ 2\ 3\ 4\
*Pearson r* (*Partial r*) *Pearson r* (*Partial r*) *Pearson r* (*Partial r*) *Pearson r* (*Partial r*)
------------------------------------ ---------------------------------------------------- ---------------------------------------------------- ---------------------------------------------------- ---------------------------
Negative life events (1) 1.000
Family support (2) −0.312(−0.348)[^\*\*^](#tfn1){ref-type="table-fn"} 1.000
Self-esteem (3) −0.201(−0.198)[^\*\*^](#tfn1){ref-type="table-fn"} 0.291(0.321)[^\*\*^](#tfn1){ref-type="table-fn"} 1.000
Antisocial behavior tendencies (4) 0.200(0.213)[^\*\*^](#tfn1){ref-type="table-fn"} −0.245(−0.250)[^\*\*^](#tfn1){ref-type="table-fn"} −0.252(−0.282)[^\*\*^](#tfn1){ref-type="table-fn"} 1.000
-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
*p* \< 0.001.
Mediation Effect of Self-Esteem {#sec13}
-------------------------------
To test the mediation effect of self-esteem between negative life events and antisocial behavior tendencies, SEM was used with self-esteem as a mediator of the effects of negative life events on antisocial behavior tendencies ([Figure 2](#fig2){ref-type="fig"}). The model fit was good \[*χ*^2^*/df* = 9.48, *p* \< 0.001, CFI = 0.944, TLI = 0.923, SRMR = 0.051, and RMSEA = 0.047 (90% CI: 0.034, 0.061)\]. This model suggested that negative life events had a negative effect on self-esteem (*β* = −0.341, *p* \< 0.001, 95% CI: −0.367, −0.312). Self-esteem was negatively related to antisocial behavior tendencies (*β* = −0.494, *p* \< 0.001, 95% CI: −0.521, −0.467), while experiencing negative life events was positively related to antisocial behavior tendencies (*β* = 0.082, *p* \< 0.001, 95% CI: 0.052, 0.111). The indirect effect of negative life events on antisocial behavior tendencies *via* self-esteem was 0.168 (*p* \< 0.001, 95% CI: 0.146, 0.191). There were two paths by which negative life events affected antisocial behavior tendencies. First, negative life events affected antisocial behavior tendencies directly, with the direct effect (0.082) accounting for 32.80% of the total effect (0.250). Second, negative life events affected antisocial behavior tendencies indirectly through self-esteem, while the indirect effect (0.168) accounted for 67.20% of the total effect. The indirect effect was 2.049-fold greater than the direct effect ([Table 2](#tab2){ref-type="table"}). These results demonstrated the mediating role of self-esteem between negative life events and antisocial behavior tendencies.
{#fig2}
######
The mediation effect size of self-esteem in the relationship between negative life events and antisocial behavior tendencies.
Path Standardized effect *p* *P~M~* (%) *R~M~*
--------------------------------------------------------------------- --------------------- --------- ------------ --------
Negative life events → Antisocial behavior tendencies 0.082 \<0.001 32.80 \-\--
Negative life events → Self-esteem → Antisocial behavior tendencies 0.168 \<0.001 67.20 2.049
Total 0.250 \<0.001 100.00 2.049
Moderated Mediation Effect of Family Support {#sec14}
--------------------------------------------
We tested the moderated mediation model with the latent interaction of family support and negative life events ([Figure 3](#fig3){ref-type="fig"}). The results revealed that the direct link between negative life events and antisocial behavior tendencies was no longer significant (*β* = 0.032, *p* = 0.091, 95% CI: −0.005, 0.069). However, other paths that included the interaction term showed significance, as seen in [Figure 3](#fig3){ref-type="fig"}. Negative life events still had a negative effect on self-esteem (*β* = −0.385, *p* \< 0.001), while self-esteem was negatively related to antisocial behavior tendencies (*β* = −0.557, *p* \< 0.001). Negative life events had an indirect effect on antisocial behavior tendencies through self-esteem that was still significant (indirect effect = 0.214, *p* \< 0.001, 95% CI = 0.191, 0.239), thus still supporting the notion that self-esteem mediated the effect of negative life events on antisocial behavior tendencies. Moreover, the effect of latent interaction of family support and negative life events on self-esteem was negatively significant (*β* = −0.018, *p* = 0.032, 95% CI: −0.035, −0.002), suggesting the negative moderating role of family support on the effect of negative life events on self-esteem. Therefore, family support moderated the mediation effect of self-esteem on the relationship between negative life events and antisocial behavior tendencies. The moderated mediation model explained 29.5% of the variance of antisocial behavior tendencies. We then plotted the conditional indirect effect of negative life events against antisocial behavior tendencies for each combination of low (below mean − 1 SD), medium (mean ± SD), and high (above mean + 1 SD) family support values. As seen in [Figure 4](#fig4){ref-type="fig"}, the indirect relationship between negative life events and antisocial behavior tendencies *via* self-esteem appeared more evident when family support was low than when it was high. The conditional indirect effect of negative life events on antisocial behavior tendencies through self-esteem for different family support level is tabulated in [Table 3](#tab3){ref-type="table"}. A stronger indirect effect of negative life events was seen with low family support (conditional indirect effect = 0.227, 95% CI = 0.200, 0.255) than with high family support (conditional indirect effect = 0.203, 95% CI = 0.177, 0.229). The index of moderated mediation effect was −0.012, *p* = 0.033. These results verified the moderated mediation effect of family support on the relationship between negative life events and antisocial behavior tendencies *via* self-esteem.
{#fig3}
{#fig4}
######
The conditional indirect effect of negative life events on antisocial behavior tendencies through self-esteem for different family support levels.
Family support level Indirect effect Standard error *p* 95% CI
---------------------------------------------------------- ----------------- ---------------- --------- ------------------
High family support[*^a^*](#tfn2){ref-type="table-fn"} 0.203 0.013 \<0.001 (0.177, 0.229)
Medium family support[*^b^*](#tfn3){ref-type="table-fn"} 0.215 0.013 \<0.001 (0.191, 0.240)
Low family support[*^c^*](#tfn4){ref-type="table-fn"} 0.227 0.014 \<0.001 (0.200, 0.255)
Index of moderated mediation effect −0.012 0.005 0.033 (−0.023, −0.003)
Family support values above mean + 1 SD.
Family support values of mean ± 1 SD.
Family support values below mean − 1 SD.
Discussion {#sec15}
==========
The effect of negative life events on antisocial behavior has gained empirical support. But the mechanisms underlying related mediating and moderating effects are still unclear. This study put forth a moderated mediation model in a sizable sample of adolescents to test whether self-esteem mediates the effect of negative life events on antisocial behavior tendencies, and whether family support moderates the indirect effect of negative life events on antisocial behavior tendencies *via* self-esteem. The results obtained in the current study supported our hypothesis.
Consistent with previous research ([@ref18]), this study found that negative life events positively correlated with antisocial behavior tendencies in adolescents, which again confirmed the adverse effect of negative life events on adolescent development. However, it is of little practical implication to prevent antisocial behavior by reducing negative life events, as we may not be able to change the negative life events that an individual has experienced. On the contrary, it will be of important practical value to lessen the harmful effect of negative life events on antisocial behavior tendencies in adolescents. Despite the well-recognized connection between antisocial behavior and negative life events ([@ref24]; [@ref10]; [@ref33]), many individuals who deal with highly stressful situations do not develop antisocial behavior, so the particular characteristics of this association is still uncertain. As a result, there is an urgent need to clarify the influence factor and internal mechanism of negative life events leading to antisocial behavior. However, there is little evidence indicating the underlying influencing mechanisms in the association of negative life events with antisocial behavior tendencies. This is the first study to examine the mediator and moderator between negative life events and antisocial behavior tendencies in adolescents, thus extending and enriching the research on antisocial behavior. The stress-support model sees antisocial behavior as being associated with an interplay of protective and vulnerability factors. The present study treated negative life events as the primary vulnerability factor that puts the youth at a higher risk of developing antisocial behavior. The protective factor is self-esteem, which may decrease risk for antisocial behavior ([@ref6]). In line with our first hypothesis, our results here showed that negative life events correlated inversely with adolescent self-esteem, which itself was negatively associated with antisocial behavior tendencies in adolescents; that is, self-esteem intervened in the effect of negative life events on antisocial behavior tendencies in adolescents. Therefore, adolescents showing more self-esteem are less prone to developing antisocial behavior. These results are consistent with previous research in problem behavior, demonstrating that those with more self-esteem will probably cope better with stressful events, resulting in fewer psycho-social problems ([@ref3]; [@ref15]). Consequently, enhanced self-esteem may be one of the mechanisms that explain why some adolescents who experience many negative life events are still less prone to antisocial behavior. The current study showed that most of the adverse effect of negative life events on antisocial behavior tendencies is mediated by self-esteem. This result is consistent with the idea that self-esteem can be viewed as an important psychological resource and motivating factor for preventing antisocial behavior ([@ref9]). Although it is impossible to change the negative life events that an individual has experienced, antisocial behavior tendencies can be avoided once self-esteem is improved, due to the negative effect of self-esteem on antisocial behavior tendencies. In addition, previous studies have demonstrated that self-esteem can be managed and developed ([@ref25]; [@ref46]). Therefore, considering the buffering action of self-esteem on the impact of negative life events, it will be of great practical significance to prevent the development of problem behavior of adolescents, if managers (for example, teachers, government workers, etc.) could focus on promoting self-esteem and training among adolescents. The results of the present study contribute to the potential use of psychological tools in dealing with the adverse effect of negative life events on antisocial behavior tendencies. These results highlight the importance of self-esteem in combating the consequences of negative life events, helping individuals to stay healthy.
In line with our second hypothesis, our research showed that family support moderated the indirect effect of negative life events on antisocial behavior tendencies through self-esteem. Prior research has stressed that family support may be essential in preventing potential emotional disorder and psychological problems of children and young adolescents facing adversity ([@ref26]). The current study found that subjective family support weakened the adverse effect of negative life events on self-esteem and subsequent antisocial behavior tendencies as well. This finding is consistent with the stress-buffering model, which argues that adolescents with strong family support, who experience many stressful life events, would not show psychological maladjustment ([@ref16]). For example, those who have greater subjective family support rely more on approach-coping, and less on avoidance-coping ([@ref31]). Family support could enhance their belief in resources for coping with potentially stressful situations to lower their emotional impact or to prevent emotional disruptions from becoming a problem behavior ([@ref35]). Therefore, adolescents with strong subjective family support will make appraisals of life events in a less negative manner, and the adversity effect of life events on self-esteem will be weakened. The present study demonstrated the stress-moderating effect of family support in relation to self-esteem and antisocial behavior tendencies of adolescents. Family support decreased the indirect effect of negative life events on antisocial behavior tendencies in adolescents, protecting them from the possibly detrimental effect of stressing events, which points to the need to enhance subjective family support during adolescence. Consequently, family support can be viewed as a way to protect youths from the harmful effects of negative life events on adolescent psychological and social behavior adjustment. Thus, our study expands our understanding of antisocial behavior management by elucidating the boundary condition where negative life events and low self-esteem can effectively contribute to antisocial behavior tendencies. It is probable that some parents relax their involvement and support in their children at adolescence, since they consider that they no longer play an important role ([@ref26]). Contrarily, our study points out the critical role that family support appears to have in the behavioral adjustment of youths, which protects them from some negative effects of stressing events in life. Accordingly, parents should be made aware of their influence on the healthy development of adolescents, and should be guided in improving their parenting skills by enhancing their communication, affection, and monitoring. This support can be more effective through the use of programs for communication training ([@ref43]), which could be adapted to prevention programs by personal training or other ways, such as online training. Adolescents with a supportive family could show more resistance to the negative effects of some adverse situations that may often occur in adolescence.
This study extends our knowledge of how self-esteem and family support play a complex role in the effect of negative life events on antisocial behavior tendencies in adolescents. This study points out important implications for preventing antisocial behavior in adolescents. The results show us the importance of self-esteem and subjective family support to the development of adolescent, even though some faced with a variety of stressful life events. To minimize the adverse effect of life events, there is an urgent need to develop interventions aimed at enhancing self-esteem and pay enough attention to the important role of subjective family support in the development of psychological and social function among adolescents. This study clearly provides interesting results and makes important contributions to the field, but it had several limitations. First, the causality of the variables was difficult to prove using the present cross-sectional study. Longitudinal designs may be considered in any expansion of the present work to better understand the relationship of the variables. Second, the information bias may be unavoidable, since all the information was obtained through self-reporting by the adolescents. It is probable that such self-reported information from adolescents was biased by social desirability. Further studies should work with more data sources, such as parents, teachers, and peers. Besides, each studied constructs of the present research was measured by one single scale. Multiple different scales should be used to measure the same structure in future studies, which can complement each other to give a better picture of the studied constructs. Furthermore, this study only measured certain demographic as potential confounding variables. Additional confounding variables, such as socioeconomic factors, should be included in future studies. And, the inter-individual patterns of the obtained models would be worth to explore in future studies, so as to enrich the practical implications of our research. In addition, it is possible that other social supports (e.g., close friends or teachers) would buffer the lack of subjective family support. Thus, a very interesting idea would be to study how these supports can have a protective effect on adolescent emotional and behavioral adjustment. Finally, the participants of the current study were only selected from three cities in China and did not sample for general population proportion, which may affect the representation of sample in the whole adolescents of China. Further studies should recruit participants from a wider geographical range and sampled for general population proportion. In spite of the above limitations, our study provided us with preliminary and innovative insights into the underlying mechanisms that influence the relationship between negative life events and antisocial behavior tendencies in adolescent.
Conclusion {#sec16}
==========
The adverse effect of negative life events on adolescent development has been well-documented. However, little is known about the internal influence mechanisms for the relationship between negative life events and antisocial behavior tendencies in adolescents. The present study again confirmed the promoting effect of negative life events on antisocial behavior tendencies in adolescents. Self-esteem was shown to be a mediator between negative life events and antisocial behavior tendencies. In addition, family support negatively moderated the indirect effect of negative life events on antisocial behavior tendencies *via* self-esteem. Negative life events had a greater indirect impact on antisocial behavior tendencies with poorer subjective family support. Subjective family support weakened the adverse effect of stressful life events on self-esteem, and consequently susceptibility to antisocial behavior in adolescents. These results displayed the important role of self-esteem and subjective family support in preventing antisocial behavior in adolescents. To minimize the adverse effect of negative life events on problem behavior development of adolescents, interventions are needed with the aim of raising self-esteem in adolescents, and parents need to be provided with support-skill training aimed at improving subjective family support of adolescents. This study came up with important implications for preventing antisocial behavior in adolescents. However, due to the cross-sectional design of the study, studies with a longitudinal design need to be conducted to further explore the causal relations between the variables studied.
Data Availability Statement {#sec18}
===========================
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Ethics Statement {#sec19}
================
The studies involving human participants were reviewed and approved by The research ethics committee of Jinzhou Medical University (No: 2016-08). Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.
Author Contributions {#sec20}
====================
FG made an outstanding contribution to the study design, data analysis and interpretation, and manuscript drafting and revision. YY participated in the study design, interpretation of the data, and manuscript revision as well. CY, YX, and HM assisted with organization of the research and data collection. HL helped with analyzing the data and drafting the manuscript. All authors contributed to the article and approved the submitted version.
Conflict of Interest {#sec21}
--------------------
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.
We are sincerely thankful for the generous help from the staff of the selected schools, who assisted in collecting and organizing the data in the current study. We are also grateful to the parents of the adolescents for their understanding and support. Furthermore, we are very grateful to the reviewers for their comments and suggestions on our manuscript, which are of great significance to improve this study and our future researches.
**Funding.** The Social Science Planning Foundation of Liaoning Province, China (No. L16CSH002) supported the study design and data collection in the present study. The General Research of Humanities and Social Sciences, Ministry of Education, China (No. 20YJC840011) supported the data analysis and interpretation, and the publication of the paper as well.
ASLEC
: Adolescent self-rating life events check list
CFA
: Confirmatory factor analysis
CFI
: Comparative fit index
TLI
: Tucker-Lewis index
SRMR
: The standardized root mean square residual
RMSEA
: Root mean square error of approximation
CI
: Confidence interval
SEM
: Structural equation model
RSCA
: Resilience scale for Chinese adolescent
MLR
: Robust maximum likelihood estimator
LMS
: Latent moderated structural equations
SD
: Standard deviation.
[^1]: Edited by: Ana M. Martín, University of La Laguna, Spain
[^2]: Reviewed by: Qi Han Zhang, Tianjin Normal University, China; Margarita Maria Maramis, STIKES Yayasan RS. Dr. Soetomo Surabaya, Indonesia; María José Rodrigo, University of La Laguna, Spain; Moises Betancort Montesinos, University of La Laguna, Spain
[^3]: This article was submitted to Psychopathology, a section of the journal Frontiers in Psychology
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
Sterile alpha motif and HD domain-containing protein 1 (SAMHD1) is an important regulator of cellular 2′-deoxynucleoside-5′-triphosphate (dNTP) in mammalian cells^[@CR1]^. SAMHD1 is widely expressed in most tissue types^[@CR2],[@CR3]^ and catalyses the hydrolysis of dNTPs into their constituent triphosphate and 2′-deoxynucleoside^[@CR4],[@CR5]^. HIV-1 and other retroviruses replicate poorly in terminally differentiated cells of the myeloid lineage and resting T-cells^[@CR6]--[@CR12]^. This is a result of SAMHD1-dNTP triphosphohydrolase activity that blocks early stage infection through depletion of the dNTP pool to a level that cannot support viral reverse transcription^[@CR13]--[@CR16]^. In addition, SAMHD1 also inhibits replication of the DNA viruses, herpes simplex virus type 1 and vaccinia virus in monocyte-derived macrophages^[@CR17],[@CR18]^, and hepatitis B virus in hepatic cells^[@CR19]^. In HIV-2 and related simian viruses (SIVs), this restriction is overcome by the lentiviral accessory protein Vpx that targets human SAMHD1 for degradation by the proteasomal pathway^[@CR9],[@CR11],[@CR12]^ through interactions with the cullin-4 ligase substrate adaptor DCAF1^[@CR20]--[@CR22]^.
In addition to restriction of HIV-1 replication, SAMHD1 is an interferon-induced gene^[@CR2],[@CR23]^ and regulator of the innate immune response^[@CR24],[@CR25]^. Germ line mutations in SAMHD1 are associated with early onset stroke^[@CR26]^ and the rare hereditary autoimmune disease Aicardi--Goutières syndrome, which mimics congenital viral infection^[@CR25],[@CR27]^. SAMHD1 mutations have also been identified as potential drivers of chronic lymphocytic leukaemia^[@CR28]--[@CR30]^ and are present in hypermutated cancers^[@CR31]^. Conversely, high SAMHD1 expression in acute myelogenous leukaemia patients is associated with efficacy reduction of the nucleoside-analogue anti-cancer drugs Cytarabine and Clofarabine^[@CR32]--[@CR34]^ as a result of SAMHD1 hydrolysis of their triphosphorylated forms^[@CR35]--[@CR38]^. Therefore, modulation of SAMHD1 activity by drugs or potentially by Vpx targeting of SAMHD1 for proteasomal degradation has been proposed as a strategy for improving anti-cancer and anti-HIV therapies^[@CR32],[@CR36],[@CR39]^. More recently, a role for SAMHD1 in the regulation of DNA repair at stalled replication forks and the suppression of the interferon response by cytosolic nucleic acids has been proposed^[@CR40],[@CR41]^.
In humans, SAMHD1 is a 626-residue protein. It comprises an N-terminal nuclear localisation signal^[@CR42]^ and two major structural domains: the N-terminal sterile alpha motif (SAM) domain, which is a helical scaffold for protein- or nucleic acid-binding in other SAM domain-containing proteins^[@CR43],[@CR44]^, and the SAMHD1 HD phosphohydrolase domain, which is named after the two pairs of conserved histidine and aspartate residues that coordinate a metal ion at the active site^[@CR45]^. In addition, residues in a C-terminal region are targeted by lentiviral Vpx accessory proteins to recruit SAMHD1 to the proteasome for degradation^[@CR22],[@CR46]^.
Although the HD-domain contains the active site, dNTP hydrolysis requires assembly of SAMHD1 into homo-tetramers^[@CR47]^. Here, sequences at the N and C termini of the HD-domain stabilise inter-monomer protein--protein interactions and incorporate four pairs of allosteric nucleotide-binding sites, AL1 and AL2, which regulate the enzyme through combined binding of G-based (AL1) and 2′-deoxynucleoside (AL2) triphosphates^[@CR4],[@CR48]--[@CR50]^. GTP is the physiological ligand for the first allosteric site, AL1, although dGTP can also coordinate this site^[@CR16],[@CR50]^. The second allosteric site, AL2, is specific for a dNTP with the following preference order: dATP \> dGTP \> TTP \> dCTP^[@CR51]--[@CR54]^ and a magnesium ion coordinates the triphosphates of each GTP-dNTP pair in adjacent AL1 and AL2-binding sites^[@CR48],[@CR49],[@CR55]^. In addition, SAMHD1 activity is cell cycle regulated by CyclinA2/CDK2 phosphorylation at Threonine 592^[@CR56]--[@CR58]^ through effects on tetramer stability that modulate activity^[@CR16]^. Removal of regulation through T592 dephosphorylation or mutation may enable SAMHD1 to inhibit HIV-1 in cycling cells^[@CR56]^.
Given that SAMHD1 allosteric regulation, assembly and the coupling of tetramerisation to catalysis have been studied so extensively, it is surprising that the molecular details and mechanism of catalysis remain unknown. Because of the importance of SAMHD1 both as a potential therapeutic target and a key intracellular regulator of metabolism, we have investigated the catalytic mechanism and inhibition of SAMHD1 through combined structural and biochemical studies. We first employed dNTP analogues modified at the α,β-phosphoester bond to probe the enzymology of SAMHD1-dNTP triphosphohydrolase activity that cleaves the labile bond between the α-phosphorus and 5′ oxygen (P^α^-O^5′^) of a substrate dNTP. These data demonstrate the contribution of the chemical moieties that are proximal to the P^α^-O^5′^ labile bond and the role of water in the hydrolysis reaction. We next determined X-ray co-crystal structures of SAMHD1 with non-hydrolysable competitive inhibitor α,β-imido-dNTP analogues. These data reveal a previously unreported iron-magnesium bi-metallic centre at the SAMHD1-active site that positions a hydroxide nucleophile for in-line attack on the substrate dNTP P^α^-O^5′^ bond. Further structural and biophysical experiments demonstrate key residues that are important for metal binding and substrate binding at the active site and show that conserved catalytic residue Histidine 215 protonates the O^5′^ leaving group, to stabilise the negative charge that O^5′^ accumulates after adduction of water to the α-phosphate.
Together, our results now reveal the physiological binding mode of substrate dNTPs in the SAMHD1-active site, provide a chemical mechanism for SAMHD1-dNTP hydrolysis and explain the nature of SAMHD1 inhibition by α,β-imido-dNTP analogues. These observations contribute significantly to our understanding of SAMHD1 regulation of cellular dNTP, its effects on the efficacy of nucleoside-based anti-cancer and anti-viral therapies, and provide insight into the design of future SAMHD1-active site inhibitors.
Results {#Sec2}
=======
SAMHD1 and α-β-modified dNTP analogues {#Sec3}
--------------------------------------
To identify dNTP analogues useful for structural studies of the catalytic mechanism of SAMHD1 triphosphohydrolase activity, we assembled a panel of α,β-imido (dNMPNPP) and α,β-methyleno (dNMPCPP) dNTP analogues (Supplementary Fig. [1](#MOESM1){ref-type="media"}). These analogues are modified at the bridging α,β-oxygen atom that is in close proximity to the P^α^-O^5′^ phosphoester bond that is cleaved by SAMHD1 in the triphosphohydrolysis reaction. The nucleotide analogues were tested for their properties as substrates, allosteric activators or inhibitors of SAMHD1 catalytic activity using both ^1^H NMR spectroscopy and a coupled enzyme assay to monitor dNTP triphosphohydrolase activity, and size exclusion chromatography with multi-angle laser light scattering (SEC-MALLS) to assess their effects on SAMHD1 tetramerisation.
SEC-MALLS analysis revealed that all four dNMPNPP analogues, which have an imido substitution at the α,β-bridging oxygen (Supplementary Fig. [1](#MOESM1){ref-type="media"}), when combined with GTP, support the formation of stable SAMHD1 tetramers, albeit to varying degrees (Fig. [1a](#Fig1){ref-type="fig"}). Near-complete tetramerisation was observed with dAMPNPP, while dCMPNPP promoted the least tetramerisation and dGMPNPP was able to support tetramerisation without additional GTP, presumably by occupying both AL1 and AL2 allosteric sites.Fig. 1SAMHD1 tetramerisation and catalytic inhibition by dNMPNPP nucleotides.**a** SEC-MALLS analysis of SAMHD1 monomer--dimer--tetramer equilibrium upon addition of dNMPNPP nucleotides. In each panel, the solid lines are the chromatograms from the output of the differential refractometer and the black scatter points are the weight-averaged molar masses determined at 1-s intervals throughout elution of chromatographic peaks, SAMHD1 monomer--dimers elute at 14--16 min, tetramers at 12 min. In each panel, the curves shown are: (red) apo-SAMHD1; (blue) SAMHD1 and 0.2 mM GTP; (cyan) SAMHD1 and 0.5 mM indicated dNMPNPP analogue; and (black) SAMHD1, 0.2 mM GTP and 0.5 mM indicated dNMPNPP analogue. For dGMPNPP, which can induce tetramer formation in the absence of GTP, only the chromatograms for the apo (red) and after addition of dGMPNPP (cyan) are shown. **b** Steady-state kinetic analysis of GTP-stimulated hydrolysis of dATP and dAMPNPP by SAMHD1, measured using the MDCC-PBP florescence-based coupled SAMHD1-Ppx assay. The dependence of the rate on substrate concentration is plotted for (black) dATP and (cyan) dAMPNPP. For the dATP reaction the solid line is the best fit to the data using the Michaelis--Menten expression, that gives values for the derived constants *K*~M~ and *k*~cat~ of 44 ± 3 µM and 0.4 ± 0.01 s^−1^, respectively. Error bars represent the standard error of the mean (s.e.m.) of three independent measurements. **c** Determination of dNMPNPP inhibition constants (*K*~i~) using the florescence-based coupled assay. Plots show the dependence of the SAMHD1 hydrolysis rate of 0.1, 0.3 and 1 mM TTP on the concentration of each dNMPNPP. Reported *K*~i~ values (inset and Supplementary Table [1](#MOESM1){ref-type="media"}) were derived from global fitting of each 3-concentration dataset. Error bars represent the s.e.m. of three independent measurements. Source data for **b** and **c** are provided in the Source Data file.
In enzymatic assays employing GTP as an AL1-activator, no hydrolysis of any dNMPNPP was observed, when monitored either by the coupled enzyme assay or by ^1^H NMR spectroscopy (Fig. [1b](#Fig1){ref-type="fig"} and Supplementary Fig. [2](#MOESM1){ref-type="media"}). Further investigation using inhibition assays revealed that all dNMPNPP analogues were actually competitive inhibitors of SAMHD1-dNTP triphosphohydrolase activity, with inhibition constants (*K*~i~) ranging from 0.78 µM for dAMPNPP to 0.052 µM for dGMPNPP, with the following rank order of increasing affinity: dAMPNPP \< TMPNPP \< dCMPNPP \< dGMPNPP (Fig. [1c](#Fig1){ref-type="fig"} and Supplementary Table [1](#MOESM1){ref-type="media"}).
By contrast, the dNMPCPP analogues, which have a methyleno substitution at the α,β-bridging oxygen (Supplementary Fig. [1](#MOESM1){ref-type="media"}) poorly stabilise the catalytically active SAMHD1 tetramer with only dAMPCPP and dGMPCPP, combined with GTP, supporting tetramerisation sufficiently well for SAMHD1 tetramers to be observed in SEC-MALLS experiments (Fig. [2a](#Fig2){ref-type="fig"}). However, assessment of dNMPCPP analogues as SAMHD1 substrates demonstrated that all are hydrolysed by GTP-activated SAMHD1, with apparent catalytic rate constants, *k*~cat~, ranging from 0.38 to 0.075 s^−1^, with the rank order of: TMPCPP \> dAMPCPP ≈ dGMPCPP \> dCMPCPP (Fig. [2b](#Fig2){ref-type="fig"}, Supplementary Fig. [3](#MOESM1){ref-type="media"} and Supplementary Table [2](#MOESM1){ref-type="media"}). The *k*~cat~ parameters determined for dNMPCPP analogues were 3- to 8-fold lower than those determined for cognate dNTP substrates, which range from 1.43 to 0.57 s^−1^ and have the rank order: TTP \> dATP \> dGTP ≈ dCTP (Fig. [2b](#Fig2){ref-type="fig"}, Supplementary Fig. [3](#MOESM1){ref-type="media"} and Supplementary Table [2](#MOESM1){ref-type="media"}). However, addition of low concentrations of dATP (10 µM) to dNMPCPP reactions, to induce tetramerisation through enhanced AL2 binding, increased the turnover of the dNMPCPP analogues with *k*~cat~ values now ranging from 0.44 to 0.21 s^−1^ and restored the rank order of turnover to that observed with the cognate dNTPs: TMPCPP \> dAMPCPP \> dGMPCPP ≈ dCMPCPP. The most significant effect of dATP addition was observed with dCMPCPP, with *k*~cat~ being increased 2.5-fold, presumably due to the limited ability of C-based nucleotides to induce SAMHD1 tetramerisation by AL2 binding. These data suggest that dNMPCPP analogues are actually SAMHD1 substrates with *k*~cat~ comparable to their analogous cognate dNTPs, but a weak association at AL2 and a lack of capacity to induce SAMHD1 tetramerisation results in decreased catalytic turnover of dNMPCPP at the active site.Fig. 2SAMHD1 tetramerisation and hydrolysis of dNMPCPP analogues.**a** SEC-MALLS analysis of SAMHD1 monomer--dimer--tetramer equilibrium upon addition of dNMPCPP nucleotides. In each panel, the solid lines are the chromatograms from the output of the differential refractometer and the black scatter points are the weight-averaged molar masses determined at 1-s intervals throughout elution of chromatographic peaks. In each panel, the displayed curves are: (red) apo-SAMHD1; (blue) SAMHD1 and 0.2 mM GTP; and (black) SAMHD1, 0.2 mM GTP and 0.5 mM indicated dNMPCPP analogue. **b** Bar chart of the apparent *k*~cat~ values derived from ^1^H NMR data recorded for GTP-stimulated SAMHD1 hydrolysis of each dNTP, dNMPCPP and dNMPCPP/dATP reaction, see Supplementary Fig. [3](#MOESM1){ref-type="media"}. Points overlaid in each bar are the individual data measurements, the error bars represent the standard deviation (s.d.) from at least two independent measurements. The source data are provided in the Source Data file.
Taking all these data into account, it is apparent that although the dNMPCPP analogues are SAMHD1 substrates, they are poor substitutes for cognate dNTPs because they are unable to induce tetramer assembly. By contrast, dNMPNPP analogues, while not SAMHD1 substrates, make excellent candidates for co-crystallisation, as they support SAMHD1 tetramerisation, through AL2 binding, coordinate the catalytic site with sub-µM affinities and are entirely refractory to hydrolysis. Therefore, we chose these analogues for structural studies aimed at understanding the molecular details of the SAMHD1 catalytic site.
Crystal structures of SAMHD1--dNMPNPP inhibitor complexes {#Sec4}
---------------------------------------------------------
We co-crystallised the SAMHD1 catalytic domain (residues 109--626) that lacks the N-terminal SAM domain with dNMPNPP inhibitor analogues in the active site and additional nucleotides at allosteric sites to support tetramer assembly. These co-crystal structures reveal how a substrate dNTP, metal ions and water molecules are coordinated at the active site prior to catalysis.
Co-crystal structures were determined using either wild-type (wt) wt-SAMHD1(109-626) or a D137N-SAMHD1(109-626) AL1 allosteric site mutant that generally produced higher resolution diffraction than wt co-crystals. A further effect of the D137N substitution is that D137N-SAMHD1 is activated by xanthosine-5′-triphosphate (XTP) rather than GTP through AL1-binding but nevertheless maintains comparable catalytic parameters for dNTP hydrolysis as those observed for wt-SAMHD1 and GTP (Supplementary Fig. [4](#MOESM1){ref-type="media"}).
We first determined structures of two co-complexes of D137N-SAMHD1(109-626) in the presence of Mg^2+^, which is required for tetramer assembly and catalysis, with (i) XTP (AL1) and dAMPNPP (AL2 and active site), and (ii) XTP (AL1) and dGMPNPP (AL2 and active site). In addition, we determined the structures of two co-complexes of wt-SAMHD1(109-626), also in the presence of Mg^2+^, with (i) GTP (AL1), dATP (AL2) and dCMPNPP (active site) and (ii) GTP (AL1), dATP (AL2) and TMPNPP (active site). dATP was added to the latter two complexes to stimulate tetramer assembly, as dCMPNPP and TMPNPP are poor AL2 activators (Fig. [1a](#Fig1){ref-type="fig"}). The structures were solved by molecular replacement and details of the unit cell contents, data collection statistics, structure determination and model refinements are presented in Supplementary Tables [3](#MOESM1){ref-type="media"} and 4.
The resolution of these structures varied between 2.0 and 3.2 Å, but all structures contain highly homologous "closed" tetramers of the SAMHD1 catalytic domain (Fig. [3a](#Fig3){ref-type="fig"}), with pairwise root-mean-square deviations (RMSDs) of \<0.5 Å across all atoms in each tetramer. Moreover, although the D137N-XTP-dAMPNPP structure is resolved at the highest resolution, there is clear density for metal ions and nucleotides bound at the allosteric and active sites in all the structures, and the coordination of dNMPNPP nucleotides at the active sites is highly similar (Fig. [3b--d](#Fig3){ref-type="fig"} and Supplementary Fig. [5](#MOESM1){ref-type="media"}).Fig. 3Crystal structures of SAMHD1--dNMPNPP inhibitor complexes.**a** Crystal structure of D137N-SAMHD1(109-626)-XTP-dAMPNPP inhibitor complex tetramer. Three monomers are shown in surface representation with a single monomer shown in cartoon representation. Monomers are coloured cyan, dark grey, dark blue and light blue, respectively. Nucleotides bound at the active and allosteric site in the cartoon monomer are shown in stick representation and coloured yellow and red, respectively. **b** Structural alignment of the active sites of the D137N-SAMHD1(109-626)-XTP-dAMPNPP, D137N-SAMHD1(109-626)-XTP-dGMPNPP, wt-SAMHD1(109-626)-GTP-dATP-dCMPNPP and wt-SAMHD1(109-626)-GTP-dATP-TMPNPP structures. Overlapping-bound nucleotides and active-site residues are shown in stick representation. The carbon atoms of the base and sugar of the bound nucleotides are coloured: yellow, dAMPNPP; magenta, dGMPNPP; cyan, dCMPNPP: and green, TMPNPP. **c** Simulated-annealing composite omit 2*F*~o~ − *F*~c~ electron density (blue mesh) for active site-bound nucleotides in D137N-XTP-dAMPNPP, D137N-XTP-dGMPNPP, wt-GTP-dATP-dCMPNPP and wt-GTP-dATP-TMPNPP structures. Nucleotides are shown in stick representation and colour coded as in **b**. Metal ions are shown as spheres (Mg, green; Fe, brown). **d** Detailed view of the active site of the D137N-SAMHD1(109-626)-XTP-dAMPNPP inhibitor complex. The protein backbone is shown in cartoon representation, bound nucleotides and active site residues are shown in stick representation, metal ions and ordered water molecules as spheres, coloured by atom type (oxygen, red; Mg, green; Fe, brown). Hydrogen bonding and metal ion coordinate bonds are shown as dashed lines. W0 indicates the in-line attacking nucleophilic water molecule.
Active-site configuration of SAMHD1--dNMPNPP complexes {#Sec5}
------------------------------------------------------
Our SAMHD1--dNMPNPP structures share many similarities with the allosteric and active sites observed in previous structures^[@CR16],[@CR37],[@CR48],[@CR51]^. However, with regards to the active-site nucleotide configuration and metal ion content, it is apparent there are significant differences that are crucial for understanding the catalytic mechanism of SAMHD1. This is most evident in our highest resolution structure, D137N-SAMHD1(109-626)-XTP-dAMPNPP-Mg, that clearly demonstrates how three octahedrally coordinated metal ions Fe, Mg2 and Mg3, support catalysis through interactions with the bound substrate, as detailed below.
Inspection of the D137N-XTP-dAMPNPP structure (Fig. [3d](#Fig3){ref-type="fig"}) reveals the key residues that position the α-phosphate proximal to the HD motif-coordinated iron, and are likely important for dNTP hydrolysis. In particular, the nucleotide 5′ oxygen, as well as non-bridging α-phosphate oxygens of dAMPNPP make hydrogen bonds with SAMHD1 side chains Arg164, His210 and His215, whereas the β- and γ-phosphates are coordinated by Mg2 (Fig. [3d](#Fig3){ref-type="fig"}). The two non-bridging α-phosphate oxygens form coordinate bonds with Fe and Mg3 (Fig. [3b--d](#Fig3){ref-type="fig"}), one phosphate oxygen coordinates Fe with an Fe--O bond distance of 2.2 Å and the second phosphate oxygen coordinates Mg3, with an Mg--O bond distance of 2.3 Å. Notably, the Fe and Mg3 metal ions have an interatomic spacing of 3.9 ± 0.5 Å (mean ± standard deviation across all Mg^2+^-SAMHD1--dNMPNPP structures) and are bridged by the carboxyl side chain of Asp207 from the HD motif as well as coordinated by the non-bridging α-phosphate oxygens of the dAMPNPP nucleotide and a water molecule, W0 (Fig. [3d](#Fig3){ref-type="fig"}).
These data reveal that the catalytic site of SAMHD1 actually comprises a bi-metallic Fe--Mg centre and although this has never been reported in previous SAMHD1-nucleotide structures, similar bi-metallic centres are present in other HD domains and several proteins that catalyse phosphoryl transfer reactions^[@CR59]--[@CR63]^. Therefore, given our observations, we sought to validate our findings and confirm the position of metal ions by substitution of Mg^2+^ with Mn^2+^ and identify ion positions using diffraction methods.
Although Mg^2+^ is essential for SAMHD1 catalysis, Mn^2+^ still supports dNTP hydrolysis with similar catalytic parameters (Supplementary Fig. [6](#MOESM1){ref-type="media"}[a, b](#MOESM1){ref-type="media"}), as is observed in other phosphohydrolase enzymes. Therefore, Mg^2+^ ions were substituted with Mn^2+^ ions in the SAMHD1--dNMPNPP co-crystallisation conditions and the structure of D137N-SAMHD1(109-626)-Mn-XTP-dAMPNPP complex was determined at 2.25 Å (Supplementary Tables [3](#MOESM1){ref-type="media"} and [4](#MOESM1){ref-type="media"}). 2*F*~o~ − *F*~c~ maps contoured at high *σ* were then used to locate and confirm the position of the highly X-ray scattering Mn^2+^ ions in the active and allosteric sites in the structure (Supplementary Fig. [6](#MOESM1){ref-type="media"}[c, d](#MOESM1){ref-type="media"}). In addition, anomalous difference maps calculated from datasets recorded at 0.93 Å from a D137N-SAMHD1(109-626)-Mg-XTP-dAMPNPP complex and at the Mn and Fe absorbance edges on a D137N-SAMHD1(109-626)-Mn--Mg-XTP-dATP-dCMPNPP complex (Supplementary Fig. 7) were further used to locate Mn ions and confirm the presence of Fe bound at the HD motif, as observed previously^[@CR16]^. Taken together, these data demonstrate that there are three manganese ions coordinated per SAMHD1 monomer, one occupying the magnesium-binding site at the allosteric site and a further two in the active site, at the Mg2 and Mg3 positions, with Fe and Mn3 forming a bi-metallic centre analogous to the Fe--Mg3 centre that we observe in our Mg^2+^-SAMHD1--dNMPNPP structures.
The Fe--Mg centre positions a hydroxide nucleophile {#Sec6}
---------------------------------------------------
Detailed inspection of our Mg^2+^-SAMHD1-dAMPNPP crystal structure revealed that, at the Fe--Mg centre, a water species, that we term W0, bridges the two metal ions with an approximately tetrahedral Fe--O--Mg bond angle (119 ± 11°; mean ± standard deviation) and that W0 is also in-line with the dAMPNPP P^α^-O^5′^ phosphoester bond (W0-P^α^-O^5′^ angle, 176.3 ± 0.7°; mean ± standard deviation) (Fig. [4a](#Fig4){ref-type="fig"}). Therefore, given this proximity and configuration, we hypothesised that W0 is the source of the nucleophile that attacks the P^α^ during SAMHD1-catalysed dNTP hydrolysis.Fig. 4NMR analysis of water-mediated SAMHD1 catalysis.**a** Detailed view of the bi-metallic centre in the SAMHD1-active site. Metal ions and coordinated water molecules are shown as spheres, coloured by atom type. W0 is the activated water molecule positioned between the HD-bound Fe and Mg3, and orientated for nucleophilic attack on the α-phosphate. The electron density shown in blue mesh is the *F*~o~ − *F*~c~ map calculated during model building prior to inclusion of W0, contoured at 6 σ. **b** Chemical structures of the dGTP substrate and deoxyguanosine and triphosphate products of SAMHD1 hydrolysis, the α, β and γ-phosphate of dGTP together with the α\*- and β\*-phosphate of triphosphate are labelled. **c**, **d** SAMHD1 hydrolysis of dGTP monitored by ^31^P NMR. Spectra of dGTP substrate (left) and the triphosphate product (right) after 0 and 3 h incubation with SAMHD1 are shown. Assigned ^31^P resonances of the α-, β- and γ-phosphate of dGTP and α- and β-phosphate of the triphosphate product species are labelled. Product resonances are marked with asterisks. An expanded view of the α\*-phosphate doublet resonance is shown inset. In **c** the reaction was carried out in solvent comprising 1:1 mixture of H~2~O^16^ and H~2~O^18^. The satellite peaks on α\*-phosphate doublet resonance result from the O^18^ isotope effect, demonstrating that the reaction proceeds via a nucleophilic attack from a water molecule on the α-phosphate that is subsequently transferred into the triphosphate product.
Support for this notion comes from studies showing that iron and magnesium reduce the p*K*~a~ for deprotonation of coordinated water molecules to form the more nucleophilic hydroxide species^[@CR64],[@CR65]^. In addition, quantum chemical calculations show that the bridging water at the Zn--Mg or Mn--Mg bi-metallic centres in the HD domains of human 3′,5′-cyclic nucleotide phosphodiesterase (PDE) enzymes can partially dissociate at physiological pH to form a hydroxide ion^[@CR66]--[@CR68]^. Given that the SAMHD1 Fe--Mg and PDE bi-metallic centres, used in the quantum chemical calculations, are similarly coordinated (Fig. [3d](#Fig3){ref-type="fig"} and Supplementary Fig. [8](#MOESM1){ref-type="media"}), this adds further support to the idea that, in SAMHD1, the W0 Fe--Mg-bridging water also likely dissociates to form a hydroxide ion. This hydroxide ion then acts as the attacking nucleophile at the P^α^ on a substrate dNTP, resulting in cleavage of the P^α^-O^5′^ phosphoester bond.
To test this hypothesis with regard to the involvement of water in the SAMHD1 catalytic mechanism, the hydrolysis reaction of dGTP was monitored using ^31^P NMR spectroscopy (Fig. [4b--d](#Fig4){ref-type="fig"}). These solution NMR experiments clearly show the conversion of the substrate dGTP spectrum, containing the three resonances from phosphorus at positions α, β and γ, to the two peaks, α\* and β\*, from the triphosphate product (Fig. [4c](#Fig4){ref-type="fig"}).
In addition, to facilitate tracking of solvent water molecules, the hydrolysis reaction was conducted in the presence of a 1:1 mixture of H~2~^16^O and H~2~^18^O solvent. Here, the product α\*-phosphorus resonance now has a satellite peak, resulting from the isotope shift upon covalent bonding of ^18^O, rather than ^16^O, to the α\*-phosphorus, with a parent-to-satellite peak intensity ratio of \~3:1 (Fig. [4d](#Fig4){ref-type="fig"}). Therefore, these data show that the reaction proceeds through nucleophilic attack of a water molecule on the α-phosphorus, resulting in oxygen from bulk water being incorporated into the triphosphate product. Moreover, as the triphosphate product contains two equivalent α\*-phosphorous atoms, the \~3:1 ratio of parent to satellite peak intensities of the product α\*-phosphate reveals a stoichiometry of a single oxygen incorporation through initial adduction at the substrate dGTP α-phosphorus. Taken together, these observations support the notion of a Fe--Mg bi-metallic centre in the SAMHD1 catalytic site that positions a hydroxide ion for nucleophilic attack on a substrate dNTP α-phosphorus, which results in dissolution of the P^α^-O^5′^ phosphoester bond to release the triphosphate and 2′-deoxynucleoside products.
Mechanism of SAMHD1-dNTP hydrolysis {#Sec7}
-----------------------------------
We propose that the crystal structures of our SAMHD1--dNMPNPP enzyme-inhibitor (EI) complexes represent good approximations of SAMHD1-dNTP enzyme--substrate (ES) complexes. Given these structural observations and our enzymological and NMR data, we propose the mechanism for SAMHD1-catalysed hydrolysis of dNTPs presented in Fig. [5a](#Fig5){ref-type="fig"}. In the mechanism, in the absence of a substrate dNTP, both Fe and Mg3 adopt distorted octahedral coordination with ligands provided by the side chains of (i) His167, Asp311, His206, Asp207, W0 and an additional water molecule for Fe; and (ii) Asp207, W0, His233 and three additional water molecules for Mg3. This is supported by the crystal structures of unliganded SAMHD1 that have either water or negatively charged phosphate bound at the active site Fe^[@CR4],[@CR52]^.Fig. 5Catalytic mechanism of SAMHD1-dNTP hydrolysis.**a** Schematic of the chemical mechanism of SAMHD1-dNTP hydrolysis. In the apo state (E) the W0 water molecule (orange) is coordinated between the HD motif-bound Fe ion and by Mg3, further water molecules take up the remaining coordination positions on the metal ions. On substrate binding, the enzyme--substrate complex (E.S) is formed and P^α^ oxygens replace water molecules to coordinate the Mg3 and Fe ions and also position the W0 nucleophile in line with the electron deficient α-phosphate. The reaction proceeds by adduction of the W0 nucleophile to the α-phosphate to form a trigonal-bipyramidal intermediate transition state (T.S.). The resulting accumulating negative charge is relieved by protonation of the leaving nucleoside 5′ oxygen to form the enzyme product complex (E.P). **b** Schematic of SAMHD1-active site conformation during the hydrolysis reaction. A bound substrate dATP nucleotide is shown in blue, W0 in orange, the HD Fe ion in brown and active site Mg ions in green. The side chains of His233 and His215 and Asp218 required for catalysis are highlighted and the direction of electron transfer is indicated by the red arrows.
In this configuration, W0 bridges Fe and Mg3 of the bi-metallic centre. This facilitates its partial dissociation to hydroxide at neutral pH to generate the nucleophile required for the hydrolysis reaction, an idea supported by previous quantum chemical calculations^[@CR66]--[@CR68]^. Upon binding of a substrate dNTP and formation of the ES complex, the two non-bridging P^α^ oxygens of the dNTP displace two of the water molecule ligands of Fe1 and Mg3 while maintaining the W0 position, the Fe--Mg3 spacing of 3.9 Å and the octahedral coordination geometry of the bi-metallic centre. In this configuration, the activated W0 hydroxide is now 2.8 Å away from the electron deficient P^α^, and the W0-P^α^-O^5′^ angle is 176°, strongly suggesting that a W0 hydroxide is the nucleophile of the hydrolysis reaction. The reaction then proceeds by nucleophilic attack by the W0 hydroxide at P^α^, and adduction to a form a trigonal-bipyramidal transition state (Fig. [5a](#Fig5){ref-type="fig"}, transition state (TS)). Inversion of P^α^ and breakage of the P^α^-O^5′^ bond results in incorporation of W0 into the newly formed triphosphate product and concomitant release of the 2′-deoxynucleoside.
A requirement of the mechanism is that during the ES to EP transition, a general acid is required for protonation of the O^5′^ in order for the P^α^-O^5′^ phosphoester bond to break and allow leaving of the deoxynucleoside product. In our SAMHD1--dNMPNPP crystal structures, the Nε2 of the conserved His215 side chain makes a hydrogen bond with the O^5′^ and is ideally placed to stabilise evolving negative charge during the transition from ES to EP and also donate a proton to the leaving 2′-deoxynucleoside (Fig. [5b](#Fig5){ref-type="fig"}). In addition, the side chain Nδ1 proton of His215 is shared with the acidic side chain of the conserved Asp218, and this configuration likely contributes further to charge neutralisation through proton relay. Therefore, the hydrolytic mechanism we propose is consistent with our experimental data demonstrating the requirement of Mg^2+^ for catalysis, the presence of a bi-metallic centre and hydroxide nucleophile in the SAMHD1 catalytic site and the proximity of highly conserved catalytic histidine residues, His215 and His233, to the metal centre and the labile P^α^-O^5′^ bond.
Critical catalytic residues {#Sec8}
---------------------------
In order to validate our reaction mechanism, site-directed mutagenesis and enzymological studies were employed to probe the function of active site histidine residues required for different aspects of catalysis. First, to test the importance of magnesium coordination at the Fe--Mg3 centre, His233, which makes a coordinate bond with Mg3 (Figs. [4](#Fig4){ref-type="fig"}a and [5b](#Fig5){ref-type="fig"}), was mutated to alanine to perturb magnesium binding. The introduction of the H233A substitution still supported tetramerisation upon addition of GTP and dATP or dAMPNPP at levels similar to wt-SAMHD1 as observed by SEC-MALLS (Supplementary Fig. [9a--d](#MOESM1){ref-type="media"}) indicating that formation of the active tetrameric state is not compromised. However, in agreement with previous observations^[@CR52]^, the apparent *k*~cat~ for GTP-activated dATP hydrolysis was reduced \~300-fold, as determined using ^1^H NMR spectroscopy (Fig. [6a, b](#Fig6){ref-type="fig"}). Therefore, these data are consistent with our model where magnesium coordination by His233 at the SAMHD1 Fe--Mg bi-metallic centre is important for dNTP coordination and hydrolysis.Fig. 6Catalytic activity of SAMHD1-active site mutants.**a** ^1^H NMR analysis of GTP-stimulated dATP hydrolysis by wt-SAMHD1 and active site mutants H233A, H215A and D218A. ^1^H NMR data were recorded for SAMHD1 hydrolysis reactions containing wt-SAMHD1 (1 µM) or active site mutants H233A (1 and 10 µM), H215A (10 µM) and D218A (1 and 10 µM) with 0.2 mM GTP AL1-activator and 0.5 mM dATP. In each panel, the integral of resolved substrate and product peak resonances is plotted against time. Initial rates of hydrolysis were determined from slopes (red lines) derived from the data measured in the linear part of the reaction. **b** Bar chart of the apparent *k*~cat~ values derived from the data shown in **a**. Points overlaid in each bar are the individual data measurements, the error bars represent the s.d. of at least three independent measurements, n.d. (not detectable). The source data are provided in the Source Data file. **c** Overlay of the active site in the crystal structures of D137N-SAMHD1(109-626)-XTP-dApNHpp and H215A--SAMHD1(109-626)-GTP-dAMPNPP inhibitor complexes. The protein backbone is shown in cartoon representation, wt (grey) and H215A (green), bound nucleotides and active site residues are shown in stick representation, metal ions and ordered water molecules as spheres, coloured by atom type. The bound inhibitor adopts the same configuration in both the wt and H215A structures.
In our proposed mechanism, His215 is the general acid that stabilises evolving negative charge at the O^5′^ during the ES to EP transition and likely protonates the O^5′^ on the 2′-deoxynucleoside leaving group (Fig. [5b](#Fig5){ref-type="fig"}). Asp218 is another highly conserved residue that makes a charged hydrogen bond to the side chain of His215, stabilising a positive charge on the Nδ1 of His215 and making the Nε2 proton available and proximal to the O^5′^ (Fig. [5b](#Fig5){ref-type="fig"}). Therefore, to probe the contribution of His215 and Asp218 to the hydrolysis reaction, alanine substitution mutants, H215A and D218A were also assessed for tetramerisation and triphosphohydrolase activity using SEC-MALLS and ^1^H NMR spectroscopy, respectively. Both mutants, H215A and D218A, supported tetramerisation upon addition of GTP and dATP or dAMPNPP at levels similar to wt-SAMHD1 (Supplementary Fig. [9e--h](#MOESM1){ref-type="media"}) and so can form the active tetrameric state. However, no measurable triphosphohydrolase activity could be detected for the H215A mutant (Fig. [6a, b](#Fig6){ref-type="fig"}), even when elevated levels (10 µM) of H215A--SAMHD1 were employed in the assay, confirming its critical function in catalysis. Moreover, the D218A substitution, although not entirely abolishing activity, heavily reduced the apparent *k*~cat~ by \~40-fold (Fig. [6a, b](#Fig6){ref-type="fig"}), consistent with its stabilising function in catalysis through maintenance of the His215 protonation state and orientation of the side chain close to the P^α^-O^5′^ phosphoester bond.
To further test the importance of His215 in the catalytic mechanism, we also determined the crystal structure of a mutant H215A--SAMHD1(109-626)-GTP-dAMPNPP inhibitor complex, details of the structure refinement and unit cell contents are presented in Supplementary Tables [3](#MOESM1){ref-type="media"} and [4](#MOESM1){ref-type="media"}. In the structure, the bound active site dAMPNPP nucleotide adopts the same configuration as that observed in the wt and D137N structures. A 3D structural superimposition of the H215A and wt structures based on alignment of all SAMHD1 atoms reveals an overall rmsd of only 0.531 Å. In addition, the active site side chains, metal ions and water molecule positions are also maintained in the same positions as in the wt active site (Fig. [6c](#Fig6){ref-type="fig"} and Supplementary Fig. [10](#MOESM1){ref-type="media"}) further supporting His215 function in catalysis at a stage subsequent to tetramer formation and binding of the substrate.
Discussion {#Sec9}
==========
SAMHD1 regulates the cellular dNTP pool through its dNTP triphosphohydrolase activity. However, despite the availability of several SAMHD1-nucleotide co-crystal structures, the structural basis for catalysis remained unclear. Therefore, we sought to determine: (i) how a substrate dNTP is coordinated at the catalytic site; and (ii) how SAMHD1 catalyses dNTP triphosphohydrolysis. Using combined biophysical and X-ray structural studies incorporating nucleotide analogues along with mutagenesis experiments we have now discovered that SAMHD1 uses a bi-metallic Fe--Mg centre to catalyse dNTP hydrolysis in a reaction similar to those catalysed by HD domains in both human and bacterial phosphodiesterase PDE enzymes^[@CR59]--[@CR63]^. Moreover, we have obtained a clear picture of substrate and inhibitor binding at the SAMHD1-active site enabling us to propose a detailed chemical mechanism for dNTP hydrolysis catalysed by SAMHD1.
We determined co-crystal structures of wt- or D137N-SAMHD1 HD catalytic domains with the dAMPNPP, dGMPNPP, dCMPNPP and TMPNPP non-hydrolysable dNTP analogues coordinated in the catalytic site (Fig. [3](#Fig3){ref-type="fig"}). Unlike previous high-resolution structures of the human SAMHD1 HD catalytic domain, that have largely employed metal ion disrupting H206R/D207N HD-mutants to catalytically inactivate SAMHD1^[@CR37],[@CR48],[@CR51]^, in our structures the active site is entirely native. In addition, we solved a further structure of the catalytically inactive mutant H215A, which does not disrupt metal ion binding, with dAMPNPP bound in the active site. In all our structures the active site side chain configurations, nucleotide conformation and metal ion--water network are all conserved, including in the H215A structure (Fig. [6c](#Fig6){ref-type="fig"}). These observations support the notion that, although critical for catalysis, His215 is not required for substrate binding. In addition, our structures reveal that a bi-metallic Fe--Mg centre at the SAMHD1-active site positions a water moiety, W0, in-line with the scissile P^α^-O^5′^ bond of a substrate dNTP, providing a model for the SAMHD1 enzyme--substrate complex (ES) prior to formation of the transition state.
Several features of our ES complex model have been observed in previously reported SAMHD1 structures^[@CR37],[@CR48],[@CR49]^. These include selectivity for 2′-deoxyribose through hydrogen bonding between the dNTP 3′-OH with Gln149 and Glu319, and steric hindrance from residues Leu150 and Tyr374 preventing binding of a ribose (2′R)-2′-OH NTP, as well as the stabilisation of dNTP binding at the catalytic site through hydrogen bonds and salt-bridges between dNTP phosphate oxygens and residues Arg164, His210, Lys312 and Arg366. However, while 2′-deoxyribose- and base-coordination in the structures presented here share these similarities with previous structures, it is apparent that the position of the dNTP α-phosphate together with water molecule organisation and metal ion complement and coordination is very different to that observed previously. Therefore, our observations, not only build significantly on previous studies, but given we observe the same configuration in different space groups and with different nucleotides and active site mutants, we are confident that our structures provide an accurate representation of the ES complex prior to substrate dNTP hydrolysis by SAMHD1.
Our enzymological data clearly demonstrate that, although the α,β-methyleno dNMPCPP analogues are SAMHD1 substrates, the α,β-imido dNMPNPP analogues in our structures are competitive inhibitors of SAMHD1, with nM *K*~i~ values, much lower than the apparent *K*~*M*~ values reported for canonical nucleotides that are in the range of 10^−5^ to 10^−3^ M^[@CR35],[@CR37],[@CR69]^. Analysis of the dNMPNPP co-complex crystal structures now suggests why this is the case. In the active site of the dNMPNPP structures, regardless of the nucleotide base, it is apparent that the side chain of Asp311 of the HD motif not only makes a coordinate bond with the bound active site Fe but also acts as an acceptor in a hydrogen bonding interaction with the proton (H^imido^) of the α,β-imido group of the dNMPNPP (Figs. [3](#Fig3){ref-type="fig"} and [6](#Fig6){ref-type="fig"}). Moreover, the active site Fe ion is highly polarising and likely distorts the delocalised carboxyl Asp311 side chain, to further increase the strength of this hydrogen bond. By contrast, canonical dNTPs and dNMPCPP analogues that are all substrates are unable to make an equivalent hydrogen bonding interaction with Asp311. Canonical nucleotides have a α,β-bridging oxygen at this position, which has a low p*K*~a~ and so have no proton to donate and the methyleno C--H group in the dNMPCPP analogues is too weakly polarised to donate a proton in a hydrogen bonding interaction with Asp311.
Given these observations, our preferred hypothesis for competitive inhibition of SAMHD1 by dNMPNPP nucleotides is that this hydrogen bond, between Asp311^Oδ^ and H^imido^ stabilises the EI complex to the extent that the reaction cannot proceed to the trigonal-bipyramidal transition state illustrated in Fig. [5a](#Fig5){ref-type="fig"} and thus prevents P^α^ inversion and P^α^-O^5′^ bond cleavage. In addition, a smaller contribution from electronegativity differences of the α,β-bridging O, N and C in dNTPs, dNMPNPPs and dNMPCPPs respectively, and the effect of this on the reactivity of the adjacent α-phosphate, may also contribute and explain our observation that canonical dNTPs are hydrolysed \~3-fold faster than dNMPCPPs, due to the greater electronegativity of oxygen over carbon increasing the reactivity of the adjacent P^α^ for nucleophilic attack. Nevertheless, the stabilising effect of the Asp311^Oδ^-H^imido^ hydrogen bond likely dominates and most importantly the dNMPNPP analogues are non-hydrolysable competitive inhibitors of SAMHD1 that bind tightly to the active site with low *K*~i~ constants. Therefore, our SAMHD1--dNMPNPP co-crystal structures provide an opportunity for the development of SAMHD1 inhibitors that can be employed in a cellular context.
Different metal ions have been reported to coordinate the HD motif residues His167, His206, Asp207 and Asp311 at the SAMHD1 catalytic site. In wt-SAMHD1-nucleotide complexes, iron, zinc and manganese have all been reported to coordinate the HD motif^[@CR4],[@CR16],[@CR48],[@CR49]^. By contrast, in structures of the catalytically inactive double mutant H206R/D207N, the HD motif metal ion-binding site is unoccupied, and instead magnesium is coordinated by the dNTP β- and γ-phosphates, proximally to Asp309^[@CR48]^. In our SAMHD1-nucleotide structures, both of these metal ion-binding sites in the catalytic site are occupied, with iron coordinated by the HD motif, and magnesium (Mg2) by the dNTP β- and γ-phosphates (Fig. [3](#Fig3){ref-type="fig"}). However, additionally in our structures it is apparent that a further magnesium ion, Mg3, is also coordinated at the catalytic site by Asp207 of the HD motif and by the proximally located His233 (Figs. [3](#Fig3){ref-type="fig"}d and [4a](#Fig4){ref-type="fig"}). Critically, Mg3 along with Fe form the Fe--Mg bi-metallic centre that is coordinated by the carboxyl side chain of Asp207, the non-bridging oxygens of the nucleotide α-phosphate and the water molecule, W0 (Fig. [4a](#Fig4){ref-type="fig"}). Given we postulate that the Fe--Mg3-bridging W0, is activated by its environment to form the hydroxide nucleophile for in-line attack on the dNTP P^α^-O^5′^ bond, the metal ion configuration at the SAMHD1-active site has important implications for understanding how SAMHD1 interacts with nucleotides and nucleotide analogues and how it catalyses hydrolysis.
Our SAMHD1 catalytic mechanism critically relies on a hydroxyl nucleophile that bridges an Fe--Mg bi-metallic centre. In other HD domains, bi-metallic centres have also been shown to be necessary for catalytic activity^[@CR59]--[@CR63]^. Structural studies of reaction intermediates of the HD containing human 3′,5′-cyclic nucleotide phosphodiesterase 9 (PDE9) clearly demonstrate a water-bridged bi-metallic centre in the active site, where the cGMP substrate is posed for nucleophilic attack at P^α^ by a Mn--Mg-bridging W0 (Supplementary Fig. [8](#MOESM1){ref-type="media"})^[@CR59]^. Here, the activated water molecule lies approximately in-line with the P^α^-O^3′^ bond that is cleaved in the cGMP hydrolysis reaction catalysed by PDE9 and is analogous to W0 in the mechanism that we postulate for SAMHD1 catalysis of P^α^-O^5′^ bond cleavage in dNTP substrates. Further, His215 and Asp218 in SAMHD1 and His252 and Glu423 in PDE9 appear to play equivalent roles in both mechanisms, acting as a proton relay to protonate the O^5′^ of the deoxynucleoside leaving group.
In bacteria, some HD domains have been exapted into signalling pathways where the bi-metallic centres of the 7TMR-HD and HD-GYP protein families^[@CR60],[@CR70]^ catalyse phospho-hydrolysis of the chemical messengers cyclic-di-AMP and cyclic-di-GMP, respectively^[@CR60],[@CR62],[@CR63]^. Elsewhere, in the mammalian enzyme myo-inositol oxygenase (MIOX)^[@CR71]^ and the bacterial enzyme PhnZ^[@CR72],[@CR73]^ HD domains have evolved to catalyse a different chemical reaction, oxidative carbon-halogen (C-X) bond cleavage. Therefore, it is clear that the bi-metallic centres of diverse HD domain-containing enzymes are used to catalyse a broad range of chemical reactions. However, despite variations in the reactions catalysed, they appear to share the use of a carboxyl side chain of an Asp or Glu residue to bridge the adjacent metal ions. Thus, although we find that the coordination about the Fe--Mg bi-metallic centre of SAMHD1 appears chemically unique, in fact SAMHD1 catalyses nucleotide hydrolysis in a similar manner to the hydrolysis reactions carried out by the HD domains of PDEs^[@CR59],[@CR61]^ as well as 7TMR-HD^[@CR60]^ and HD-GYP^[@CR62],[@CR63]^ enzymes.
The wealth of data we have accumulated relating to SAMHD1 catalysis and catalytic mechanism now realises our understanding of how SAMHD1 catalyses dNTP hydrolysis to regulate the cellular dNTP pool. Moreover, the description of the competent configuration of SAMHD1 substrates and the ES complex informs on how to avoid the effects of SAMHD1 hydrolysis on new nucleotide based-therapies and goes some way to the development of inhibitors that might be used to modulate SAMHD1 activity in a cellular context to understand function in HIV-1 restriction, DNA repair and innate immune sensing.
Methods {#Sec10}
=======
Protein expression and purification {#Sec11}
-----------------------------------
For expression in *E. coli*, the DNA sequences coding for human SAMHD1 residues M1-M626, SAMHD1 and Q109-M626, SAMHD1(109-626), were amplified by PCR and inserted into a pET52b expression vector (Novagen) using ligation independent cloning (SAMHD1) or the *Xma*1/*Not*1 restriction sites (SAMHD1(109-626)) to produce N-terminal StrepII-tag fusion proteins. The D137N and active site point mutants were prepared from the parent SAMHD1 and SAMHD1(109-M626) constructs using the Quikchange II kit. Primer sequences for PCR and mutagenesis are provide in Supplementary Table [5](#MOESM1){ref-type="media"} and all insert sequences were verified by DNA sequencing. StrepII-tagged SAMHD1 constructs were expressed in the *E. coli* strain Rosetta 2 (DE3) (Novagen) grown at 37 °C with shaking. Protein expression was induced by addition of 0.1 mM IPTG to log phase cultures (*A*~600~ = 0.5) and the cells incubated for a further 20 h at 18 °C. Cells were harvested by centrifugation, resuspended in 50 mL lysis buffer; 50 mM Tris-HCl pH 7.8, 500 mM NaCl, 4 mM MgCl~2~, 0.5 mM TCEP, 1× EDTA-free mini complete protease inhibitors (Roche) per 10 g of cell pellet and lysed by sonication. The lysate was cleared by centrifugation for 1 h at 50,000×*g* and 4 °C, and subsequently applied to a 10 mL StrepTactin affinity column (IBA), followed by 300 mL of wash buffer (50 mM Tris-HCl pH 7.8, 500 mM NaCl, 4 mM MgCl~2~, 0.5 mM TCEP) at 4 °C. Bound proteins were eluted from the column by circulation of 0.5 mg of PreScission Protease (GE) in 25 mL of wash buffer over the column in a closed circuit overnight. The supernatant (25 mL) and an additional elution of wash buffer (15 mL) were pooled and concentrated to 2.5 mL. PreScission Protease was removed by affinity chromatography using a 1 mL GSTrap HP column (GE) and the eluent was applied to a Superdex 200 26/60 (GE) size exclusion column equilibrated with a gel filtration buffer of 10 mM Tris-HCl pH 7.8, 150 mM NaCl, 4 mM MgCl~2~, 0.5 mM TCEP. Peak fractions were concentrated to approximately 20 mg mL^−1^ and flash-frozen in liquid nitrogen in small aliquots.
Nucleotides {#Sec12}
-----------
GTP and dNTPs were purchased from ThermoFisher Scientific. XTP, and dNMPNPP and dNMPCPP dNTP analogues were purchased from Jena Bioscience, DE.
Crystallisation and structure determination {#Sec13}
-------------------------------------------
We determined the X-ray crystal structures of the following seven complexes of SAMHD1(109-626), with canonical nucleotides and dNMPNPP analogues: (i) D137N-SAMHD1(109-626)-Mg-XTP-dAMPNPP, (ii) D137N-SAMHD1(109-626)-Mg-XTP-dGMPNPP, (iii) wt*-*SAMHD1(109-626)-Mg-GTP-dATP-dCMPNPP, (iv) wt*-*SAMHD1(109-626)-Mg-GTP-dATP-TMPNPP, (v) D137N-SAMHD1(109-626)-Mn-XTP-dAMPNPP, (vi) D137N-SAMHD1(109-626)-Mn--Mg-XTP-dATP-dCMPNPP and (vii) H215A--SAMHD1(109-626)-Mg-GTP-dAMPNPP. Prior to crystallisation, protein samples were diluted to 5 mg mL^−1^ with gel filtration buffer, and incubated with the appropriate nucleotides, as detailed above, at the following concentrations: XTP (1 mM), GTP (1 mM), dATP (0.5 mM) and dNMPNPP analogues (2 mM). For complexes v and vi, MnCl~2~ was also included at 5 mM.
Crystals were produced by sitting drop vapour diffusion at 18 °C using a mosquito® crystal robot (SPT Labtech) to prepare 0.2 µL droplets containing an equal volume of the protein/nucleotide solution and mother liquor. Crystals were obtained using a mother liquor of 100 mM Bis-tris-HCl pH 6--7.3 and 15--19% (v/v) PEG 3350, which additionally contained 150 mM Li~2~SO~4~ for complexes (ii), (iii) and (iv), and 100 mM MgCl~2~ for complex (vii). For data collection, 30% (v/v) glycerol was added to the respective crystallisation condition and crystals flash-frozen in liquid nitrogen. Datasets for determining the structures were collected on beamlines i03, i04, i04-1 and i24 at Diamond Light Source, UK, at wavelengths between 0.9282 and 0.9795 Å, with additional datasets for complex (vi) recorded at 1.9075 Å and at the manganese and iron *K* absorption edges, 1.8929 and 1.704 Å respectively. Details of the data collection along with crystallographic space groups, contents of the asymmetric unit and structure refinement statistics, are in Supplementary Tables [3](#MOESM1){ref-type="media"} and [4](#MOESM1){ref-type="media"}.
Data were processed using either the autoPROC pipeline^[@CR74]^ (Global Phasing LtD) for structures i to iv, or the DIALS pipeline^[@CR75]^ for structures v, vi and vii. Internally, indexing and integration utilised XDS^[@CR76]^ or DIALS; point-group symmetry was determined with POINTLESS^[@CR77]^, isotropic scaling was carried out using AIMLESS^[@CR78]^ and structure factors generated using CTRUNCATE^[@CR79]^. For structures i-iv, the data were also anisotropically scaled in autoPROC using STARANISO (<http://staraniso.globalphasing.org/cgi-bin/staraniso.cgi>) (Global Phasing LtD).
Structures were solved by molecular replacement using the programme PHASER^[@CR80]^ implemented in the CCP4 interface^[@CR81]^ using the structure of wt-SAMHD1(113-626) (PDB code: 4BZC)^[@CR48]^ as an initial search model to solve the D137N-SAMHD1(109-626)-Mg-XTP-dAMPNPP structure. Here, Buccaneer^[@CR82]^ and manual building within the programme Coot^[@CR83]^ were combined iteratively with refinement utilising individual B-factors and TLS groups in Refmac5^[@CR84]^ to produce a final model for residues 114--275 and 284--599 (chain A), and residues 114--276 and 284--599 (chain B). The protein model produced for D137N-SAMHD1(109-626) in structure i was then used as the search model for molecular replacement with PHASER in datasets ii, v, vi and vii and the protein model produced in structure ii was used as the search model for molecular replacement with PHASER in datasets iii--iv. For refinement in Refmac5, the same common set of Free-R flags was applied to reflections shared between datasets i, v and vi, and a different common set of Free-R flags for reflections in datasets ii--iv. Simulated-annealing composite omit and anomalous difference maps were generated using phenix.maps within the Phenix software package^[@CR85]^. The programme AceDRG^[@CR86]^ was used to derive stereochemical restraint libraries for the nucleotide analogues: XTP, dAMPNPP, dGMPNPP, dCMPNPP and TMPNPP.
The coordinates and structure factors of the D137N-SAMHD1(109-626)-Mg-XTP-dAMPNPP, D137N-SAMHD1(109-626)-Mg-XTP-dGMPNPP, wt*-*SAMHD1(109-626)-Mg-GTP-dATP-dCMPNPP, wt*-*SAMHD1(109-626)-GTP-dATP-TMPNP, D137N-SAMHD1(109-626)-Mn-XTP-dAMPNPP, D137N-SAMHD1(109-626)-Mn-XTP-dATP-dCMPNPP and H215A--SAMHD1(109-626)-Mg-GTP-dAMPNPP inhibitor complexes have been deposited in the Protein Data Bank under accession numbers 6TX0, 6TXA, 6TXC, 6TXE, 6TXF, 6YOM and 6XU1, respectively.
SEC-MALLS {#Sec14}
---------
Size exclusion chromatography coupled to Multi-Angle Laser Light Scattering (SEC-MALLS) was used to determine the molar mass composition of SAMHD1 samples upon addition of nucleotide analogues and activators. 30 µM wt-SAMHD1(1-626) was incubated at room temperature for 5 min after the addition of nucleotide analogues (0.5 mM) and activators (0.2 mM GTP). Samples (100 µL) were then applied to a Superdex 200 10/300 INCREASE GL column equilibrated in 20 mM Tris-HCl pH 7.8, 150 mM NaCl, 5 mM MgCl~2~, 0.5 mM TCEP and 3 mM NaN~3~ at a flow rate of 1.0 mL min^−1^. The scattered light intensity and protein concentration of the column eluate were recorded using a DAWN-HELEOS laser photometer and an OPTILAB-rEX differential refractometer (dRI) (d*n*/d*c* = 0.186), respectively. The weight-averaged molecular mass of material contained in chromatographic peaks was determined using the combined data from both detectors in the ASTRA software version 6.1 (Wyatt Technology Corp., Santa Barbara, CA).
NMR analysis of SAMHD1 catalysis {#Sec15}
--------------------------------
1D ^1^H NMR spectroscopy was used to measure SAMHD1 hydrolysis rates of dNTPs. Reactions were prepared in NMR buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, 5 mM MgCl~2~, 2 mM TCEP, 5% D~2~O) containing 0.5 mM of each dNTP or dNTP analogue, 0.2 mM GTP and 1--10 μM of wt- or mutant SAMHD1(1-626). In experiments designed to enhance hydrolysis of dNMPCPP analogues, 10 µM dATP was also included in the reaction. ^1^H NMR spectra (2 dummy scans, 4 scans) were recorded at 30 s intervals as a pseudo 2D array at 293 K using either a Bruker Avance III 600 MHz or Avance IIIHD 700 MHz NMR spectrometer equipped with a 5 mm TCI cryoprobe. Solvent suppression was achieved using excitation sculpting^[@CR87]^. Experiments were typically carried out for between 1 and 10 h. The integrals for clearly-resolved substrate and product peaks at each time-point were extracted using the Bruker Dynamics Centre software package, and used to construct plots of substrate or product against time. Initial rates were extracted from the linear part of the curve in order to determine apparent *k*~cat~ values.
For ^31^P experiments, NMR Spectra were recorded at 25 °C on a 600 MHz Varian Inova spectrometer using a broadband X-nucleus detection probe. ^1^H decoupling, when employed, was achieved using the WALTZ scheme (B~1~ \~1600 Hz)^[@CR88]^. SAMHD1 hydrolysis of 1 mM dGTP was monitored in a reaction buffer containing 1 µM wt-SAMHD1(1-626) in 20 mM Tris pH 8.0, 150 mM NaCl, 5 mM MgCl~2~ and 2 mM TCEP by recording a continuous series of 15-minute 1D spectra over a 6-hour period. Spectra were referenced indirectly to the ^1^H peak of the water using *γ*(^31^P)/*γ*(^1^H) = 0.404808636^[@CR89]^. To assess the contribution of activated water to nucleophilic addition in the triphosphohydrolase reaction, a hydrolysis reaction was also prepared in standard reaction buffer but with a solvent comprising an equimolar mixture of H~2~^16^O and H~2~^18^O water. Hydrolysis of dGTP was measured using the same data acquisition parameters that were employed in the H~2~^16^O reaction.
Real-time measurement of triphosphohydrolase activity {#Sec16}
-----------------------------------------------------
To obtain quantitative kinetic parameters for substrate hydrolysis (*K*~M~ and *k*~cat~) and for inhibition by dNMPNPP analogues (*K*~i~) a coupled assay was employed utilising the biosensor MDCC-PBP^[@CR90],[@CR91]^ to measure phosphate release from combined SAMHD1 triphosphohydrolase and *S. cerevisiae* Ppx1 exopolyphosphatase activity^[@CR35]^. In a typical experiment, solutions containing wt-SAMHD1(1-626), Ppx1, MDCC-PBP and GTP were incubated for 5 min in assay buffer (20 mM Tris pH 8.0, 150 mM NaCl, 5 mM MgCl~2~ and 2 mM TCEP) at 25 °C before the reaction was initiated by the addition of nucleotides and nucleotide analogues. The final concentrations were 0.2 µM SAMHD1, 0.01 µM Ppx1, 40 µM MDCC-PBP, 0.2 mM GTP and varying concentrations of dNTP substrates and analogues. The fluorescence intensity was recorded at 430 nm excitation and 465 nm emission over a period of 10--30 min in a Clariostar multiwell plate reader (BMG Labtech). Steady-state rates were obtained from time courses of *P*~i~ formation by linear regression of the data points in the linear phase of the reaction. Rates were divided by the SAMHD1 concentration and plotted versus substrate concentration. Apparent dissociation constants for substrate binding (*K*~M~) and catalytic constants (*k*~cat~) were then determined by non-linear least-squares fitting using a hyperbolic function in the software package Prism 7 (Graphpad).
For inhibition studies, experiments were conducted at three constant substrate concentrations (1, 0.3 and 0.1 mM TTP), the SAMHD1, Ppx1, MDCC-PBP and GTP concentrations were maintained as above and the inhibitor concentration was varied. Data from the three different experiments were analysed by non-linear least-squares fitting using the equation for competitive inhibition; where *V*/\[SAMHD1\] is the steady-state rate, normalised to the SAMHD1 concentration, \[*S*\] is the (fixed) substrate concentration, \[*I*\] is the (variable) inhibitor concentration, *K*~*i*~ is the inhibition constant, and *k*~cat~ and *K*~M~ are the catalytic and Michaelis--Menten constants for substrate turnover in the absence of inhibitor.$$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$\frac{V}{{[{\mathrm{SAMHD1}}]}} = \frac{{k_{{\mathrm{cat}}} \cdot [S]}}{{[S] + K_{\mathrm{M}} \cdot \left( {1 + \frac{{[I]}}{{K_{\mathrm{i}}}}} \right)}}.$$\end{document}$$The fitting was performed with a fixed value of *K*~M~ for GTP-activated TTP hydrolysis, determined previously^[@CR35]^ and only *k*~cat~ and *K*~i~ were allowed to vary. The fits for the three datasets yielded invariant values for *K*~i~ and *k*~cat~ supporting a competitive mode of inhibition. Global fitting of the entire three concentration dataset was used to extract the final *K*~i~ value. All measurements were performed in at least triplicate.
Reporting summary {#Sec17}
-----------------
Further information on experimental design is available in the [Nature Research Reporting Summary](#MOESM2){ref-type="media"} linked to this paper.
Supplementary information
=========================
{#Sec18}
Supplementary Information Reporting Summary
Source data {#Sec19}
===========
Source Data
**Peer review information** *Nature Communications* thanks Masha Rosenberg and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
**Publisher's note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
These authors contributed equally: Sarah J. Caswell, Simone Kunzelmann.
Supplementary information
=========================
**Supplementary information** is available for this paper at 10.1038/s41467-020-16983-2.
We thank Peter Rosenthal for critical reading of the manuscript. We gratefully acknowledge the Diamond Light Source, Didcot, UK (Block allocation groups MX13775 and MX18566) and beamlines i03, i04, i04-1 and i24; and the Swiss Light Source, Paul Scherrer Institut, Switzerland, beamlines PXI and PXIII for access and support from Beamline staff. NMR spectra were recorded at the MRC Biomedical NMR Facility, Francis Crick Institute UK, which is funded by Cancer Research UK (FC001029), the UK Medical Research Council (FC001029) and the Wellcome Trust (FC001029). The work was supported by the Francis Crick Institute, which receives its core funding from Cancer Research UK (FC001178) the UK Medical Research Council (FC001178) and the Wellcome Trust (FC001178) and by a Wellcome Senior Fellowship to I.A.T. (108014/Z/15/Z).
All authors contributed to the design of experiments. E.R.M., S.J.C., S.K., A.G.P. and L.H.A. performed crystallographic and enzymological studies. G.K., L.H.A., E.R.M. and S.J.C. performed NMR studies. E.R.M. and I.A.T. wrote the manuscript with contributions from G.K., S.K. and S.J.C.
The structural data that support the findings of this study have been deposited in the Protein Data Bank. The coordinates for the D137N-SAMHD1, Mg, XTP, dAMPNPP; D137N-SAMHD1, Mg, XTP, dGMPNPP; wt-SAMHD1, Mg, GTP, dATP, dCMPNPP; wt-SAMHD1, Mg, GTP, dATP, dTMPNPP; D137N-SAMHD1, Mn, XTP, dAMPNPP; D137N-SAMHD1, Mn, Mg, XTP, dATP, dCMPNPP; H215A--SAMHD1, Mg, GTP, dAMPNPP inhibitor complexes have PDB accession numbers 6TX0, 6TXA, 6TXC, 6TXE, 6TXF, 6YOM and 6XU1, respectively. Source data underlying Figs. [1](#MOESM3){ref-type="media"}b, c, [2](#MOESM3){ref-type="media"}b, [6a, b](#MOESM3){ref-type="media"} and Supplementary Figs. [4](#MOESM3){ref-type="media"}b, c and [6a, b](#MOESM3){ref-type="media"} are provided as a Source Data file. Other data that support the studies' findings are available from the corresponding author upon reasonable request.
The authors declare no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
The electronic version of this article is the complete one and can be found at: <http://f1000.com/prime/reports/m/6/101.1>
Commentary on: Recent advances in understanding/management of hypospadias \[F1000Prime Reports, 2014\] {#s01}
======================================================================================================
Snodgrass and Bush have presented a remarkable long-term series on the one stage tubular incised plate repair, i.e. TIP or Snodgrass repair, to correct hypospadias. It was originally described for distal hypospadias and with time Dr. Snodgrass has extended its indications to more proximal defects and redo repairs. We all can learn from Dr. Snodgrass\' tenacity in carefully documenting what he has done at each surgery in a registry, in an effort to improve results. Along the way, Dr. Snodgrass has modified his approach and gives credit to those who have influenced these changes. One such example is the credit he has given to Japanese surgeons; when Dr. Snodgrass watched them performing a Snodgrass repair and noticed how aggressive they were in freeing up the flaps to make them large to make them generous and to facilitate a tension-free closure. This makes for better distal urethral coverage, less glans dehiscence and a less frequent retrusive meatus. Dr. Snodgrass subsequently adapted this approach to his own repair and with improved results. I can verify this approach as I have become more and more aggressive in fashioning glans flaps over the past two decades. My approach has been to make deep parallel incisions in the glans when outlining the glanular urethral plate and undermine the flaps when necessary so that the glans flaps end up unrestricted and lying flat just like an open book with a broken binding. Sometimes a slightly angled longitudinal incision needs to be made at the inside base of the flaps in order to further break and flatten the "binding".
It was Lowell King who popularized tubularizing the urethral plate in a one-stage repair leading to a succession of subsequent reports by other surgeons as regards not removing this valuable tissue and that some examples of chordee are produced only by the skin. Dr. King, however, only brought the meatus to the base of the glans \[[@bib-001]\]. In 1995 Van Horn and Kass reported on 166 cases of hypospadias repair based on King\'s approach but they extended the urethroplasty to include the glans plate and sulcus as well with advancing the meatus to the distal tip of the glans, as illustrated in drawings within the article \[[@bib-002]\]. They called the repair the "GITUP" (glanuloplasty and in situ tubularization of the urethral plate) procedure. While the majority of repairs were for distal hypospadias, 9% were mid shaft and 5% penoscrotal. Five years later, and with more experience, Kass and Chung reported on another 142 children, but now with a 1.7% fistula rate following distal repair and 7.7% following more proximal repairs \[[@bib-003]\]. Evidently, even though Snodgrass published one year earlier than Kass, both series lend evidence to the likelihood that these two surgeons were independently developing their repairs over a number of years before they were published. With Dr. Snodgrass\' midline longitudinal incision through the entire length of the plate, the use of King\'s original repair was further extended to include those with narrower plates. It is important to keep in mind that not all urethral plates need to be incised, especially when the plate is soft, elastic, and wide. In fact Dr. Kass does not find it necessary to incise the plate in more than 90% of the distal hypospadias repairs that he performs. He has also been using large glans flaps since conceptualizing the GITUP repair and reports that, as a result, he has not had any instance of separation of the glans closure (Dr Kass, personal communication, 2014). Personally, I have incised the plate in approximately 90% of my cases.
We all learn from registries, even retroactive reviews of these registries. As surgeons, we improve our own techniques by going through continuous modifications. Registries help us to compare our results over time and as well to share our findings in the literature with our colleagues. Surgeons fortunate enough to have large series of any particular surgery, for the sake of our specialty, should feel obligated to share their experience. Dr. Snodgrass obviously is a proponent.
The small glans, as Dr. Snodgrass has discussed, is still a challenge. Dr. Snodgrass\' approach certainly has made more glans grooves usable for tubularization and has simplified the approach to hypospadias.
Disclosures
===========
The author declares that he has no disclosures.
GITUP
: glanuloplasty and in situ tubularization of the urethral plate
| {
"pile_set_name": "PubMed Central"
} |
TERMS AND DEFINITIONS {#S0001}
=====================
*Chaos theory* is a set of concepts which studies the behaviour of dynamical systems that are highly sensitive to initial conditions.
*Deduction*. A form of reasoning which works from the more general to the more specific. Deduction is a tool for rigorously testing a hypothesis.
*Dynamical system* -- a set of interacting and interrelated elements that can change in time.
*Induction*. A form of reasoning whereby general propositions are derived from specific examples. Induction is when an observation leads to a hypothesis.
*Integral Theory (IT)*. States that POP and pelvic floor symptoms mainly derive from laxity in the vagina or its supporting ligaments.
*Integral Theory System(ITS) or Integral System*. A management system based on IT which diagnoses and treats lax vagina/ligaments in 3 zones of the vagina.
*Non--linearity*. A system whose output is not directly proportional to its input.
*Posterior fornix syndrome*. A system complex comprising urgency, nocturia, abnormal emptying and low dragging pelvic pain caused by laxity in the uterosacral ligaments.
*Reductionism* the practice of simplifying a complex idea to the point of minimizing, obscuring, or distorting it.
Urinary stress incontinence (USI) urine loss on effort.
AIM OF THIS EDITORIAL {#S0002}
=====================
The aim of this editorial is to introduce the concept of non--linearity (complexity) and to present a few thoughts on its impact on innovation in surgery. Some of the steps involved in the discovery of the midurethral sling and Integral Theory System are provided as examples.
INTRODUCTION {#S0003}
============
Nature works in an interconnected, holistic and non--linear way: every part of a system affects another part and also, the system itself. Yet very little is written about non--linearity in medicine or research medicine in mainstream journals.
One famous description as to how Nature works was written by Marcus Aurelius, the Roman Emperor and Stoic philosopher \[[@CIT0001]\].
Always think of the Universe as one living organism, with a single substance, and a single soul; and observe how all things are submitted to the single perceptivity of this one whole; all are moved by its single impulse, and all play their part in the causation of every event that happens. Remark the intricacy of the skein, and the complexity of the web.
This description emphasizes one fundamental scientific fact: Medicine, like Nature, consists of complex nonlinear systems operating in every part of the body. These systems are balanced and interact with each other. A pure reductionist approach, therefore, has little place in the study of such a system, as it is literally impossible to isolate just one single factor unrelated to anything else. One has to look at the whole picture.
One example of this is the practice of using single urodynamic flow rate values in articles on "obstructed" micturition. The flow of urine through the urethra is non--linear in that it is inversely related to the 5^th^ power of the radius \[[@CIT0002]\] for non--laminar flow (doubling the radius increases the resistance to urine flow by a factor of 32); the urethral tube is opened (and closed) by an external striated muscle mechanism reliant on intact suspensory ligaments, so even the slightest difference in tissue tension or timing is exponentially magnified to the 5^th^ power, ensuring that flow rate can never be adequately reproducible \[[@CIT0003]\] he same criticism applies against drawing conclusions from single pressure values of "detrusor overactivity" ("DO") obtained during urodynamic testing; the instantaneous "DO" pressure recorded ultimately reflects the instantaneous urethral radius and therefore resistance, a resultant of the interplay of the natural urethral closure mechanisms which close the urethral tube \[[@CIT0003]\] and the micturition reflex which opens it out \[[@CIT0002]\]. The characteristic sinuous curve characteristic of "DO" is consistent with the slight time delay taken to switch from domination of the closure reflex (raised urethral pressure) to domination of the micturition reflex (fall in urethral pressure). "DO" as applied today is reductionist, in that it takes a single reading from an exponentially determined dynamic process and draws important conclusions, for instance, in a patient with mixed incontinence, whether to operate or not.
First steps in innovation -- towards a hypothesis {#S20004}
-------------------------------------------------
The first step is for an "innovative" scientist is to "discover" how it works, then to apply that knowledge in solving a difficult clinical problem. Popper \[[@CIT0004]\] describes two mechanisms, induction and deduction. Induction, simplistically, takes an idea or observation towards a hypothesis. The hypothesis then has to be tested in a valid way, by deductive experimentation: "if "a" is so, then "b" follows. In 1986, the 1^st^ author, PP, noted a dense collagenous tissue reaction to implanted Teflon tape. PP had previously noted that pressing a hemostat directly upwards immediately behind the pubic bone controlled urine loss on coughing (USI). Two hypotheses were formed,
The cause of USI is a loose pubourethral ligament (PUL).
Implantation of a tape will reinforce the damaged PUL by creating an artificial collagenous neoligament.
How non--linearity (randomness) may inadvertently affect research directions
PP obtained permission from the hospital IRB to perform two prototype midurethral sling operations in 1986. Both worked immediately, the patients passed urine immediately, went home the same day and were continent 10 years later. However, when a systematic study was commenced a year later, results in the 1st case series showed only a 50% cure rate. On this basis, the probability of achieving 2/2 success in the initial 1986 operations was only 25%. Conversely, there was an equivalent 25% probability of 2/2 failures. If this had occurred, there would be no midurethral sling today. *What this means is that the Innovative Scientist cannot assume the concept is wrong (or correct) on the basis of 2 cases*. One needs some luck and a lot of persistence. Marion Sims, the fistula pioneer, had many attempts at fistula repair before he finally succeeded with silver wire.
*Variation* Nonlinearity describes variation and Chaos Theory describes the feedback control mechanisms which drive non--linear systems. The challenge is to create a management system which makes allowance for variation and to avoid the rare "Black Swan" events inherent in Chaotically controlled systems. The Integral Theory System (ITS) is to our knowledge, the only management system to date which recognizes variation and nonlinear dynamics in its diagnostic and surgical systems.
Testing the theory on which the Integral System is based {#S20005}
--------------------------------------------------------
*"Organ prolapse, symptoms of stress, urge, abnormal emptying and some types of pelvic pain arise from laxity in the vagina or its suspensory ligaments, a result of altered collagen/elastin"* -- Integral Theory.
### Pubourethral ligament
One part of the Integral System states that pubourethral ligament (PUL) laxity is the principal cause of USI and is also, an important cause of urgency, [Figure 1](#F0001){ref-type="fig"}. For the midurethral sling, which reinforces the PUL, this was done in a deductive way \[[@CIT0005]\]. A hypothesis was formed of what each factor contributed, each variable (PUL, suburethral vaginal hammock, external ligament) was tested in turn and the effect was evaluated \[[@CIT0005]\]. Reinforcement of the pubourethral ligaments with a midurethral sling and tightening the suburethral vagina in patients with USI has not so many variables and so was reasonably amenable to a series of deductive experiments, each testing a different variable \[[@CIT0005]\]. Far more complex is how to scientifically evaluate other structures, for example, uterosacral ligament (USL) function.
{ref-type="fig"}, symptom grouping is used to make the diagnosis. The supporting ligaments/structures naturally divide the vagina into 3 zones: Anterior External meatus to bladder neck, Middle Bladder neck to cervix, Posterior Cervix to perineal body. The height of the bar indicates probability of the symptom originating from damaged structures in that zone. Note that the uterus is supported partly by CL (middle zone) and USL (posterior zone).](CEJU-66-E19-g001){#F0001}
Uterosacral ligament {#S20007}
--------------------
Another part of the Integral System states that uterosacral ligament laxity causes "posterior fornix syndrome" symptoms of *urgency, frequency, nocturia, abnormal emptying, pelvic pain*, ([Figure 1](#F0001){ref-type="fig"}).
The hypothesis for symptoms arising from USL, according to the Integral System (ITS), is that adequately tensioned uterosacral ligaments provide a firm anchoring point for the muscles which open and close the urethra \[[@CIT0003], [@CIT0006]\] and supports the nociceptic fibres which run through it \[[@CIT0007]\].
The authors have found that a tensioned TFS (Tissue Fixation System) sling which reinforces the uterosacral ligaments (USL) will variously cure/improve up to 70--85% of "posterior fornix syndrome" symptoms, posterior zone, [Figure 1](#F0001){ref-type="fig"}. These results appear to support the hypothesis, but they do not prove it. The proper scientific approach to proof would be to isolate the anatomical factors deemed to influence USL tension (all the pelvic muscles, other connective tissue structures such as cardinal ligaments, rectovaginal fascia, parasympathetic nerves which influence USL contractility etc.) and then test each for its effect on function. Lack of a suitable model aside, the nonlinear variation inherent in USL anatomy \[[@CIT0008]\], plus the greatly varying sensitivity of the nociceptic fibres (pain symptoms) in USL and the stretch receptors at bladder base (urge symptoms) which are also non--linear, render any direct scientific testing impossible.
Management-- How the ITS diagnoses and restores ligamentous tension and overactive bladder symptoms {#S20008}
---------------------------------------------------------------------------------------------------
*Diagnosis*. The Pictorial Diagnostic Algorithm, [Figure 1](#F0001){ref-type="fig"} is a summary guide to diagnosis and management of pelvic floor disorders. It is what Popper would call *"*a working hypothesis", based on the outcome of the 1^st^ author's extensive surgical experience repairing specific ligaments. It incorporates nonlinearity: the bar size gives an estimate of frequency of occurrence of a symptom in a particular zone.
Non--linear surgical restoration of ligament tension by the TFS {#S20009}
---------------------------------------------------------------
The TFS prosthesis incorporates the patient's own feedback control system in its surgical methodology. Using the one--way system at the base, the tape is tightened until a resistance is felt. This is a classical application of Gordon's Law \[[@CIT0009]\]: a muscle will contract efficiently only over a defined length. If its insertion point, in this instance, the ligament/ vaginal fascia is loose, the muscle effectively lengthens. As the tape is tightened, it shortens the ligament/vaginal fascial insertion until the muscle contractility returns. This is felt by the surgeon as an increasing resistance to tightening. In this sense, the TFS harnesses the patient's own non--linear control system to assess the tightness required.
Symptomatic improvement following restoration of uterosacral ligament tension by approximation with a No1 Vicryl can be as high as 80--85% for each of the "posterior fornix" symptoms \[[@CIT0010], [@CIT0011]\] in the 1^st^ instance. By 3--6 months, the tissues begin to loosen again, with increasing failure, hence our introduction of the posterior polypropylene sling \[[@CIT0006], [@CIT0011]\] to permanently strengthen USL. This is just one example of the complex chaotic dynamical system of bladder/urethral control described by Waliszewski \[[@CIT0012]\] and in a slightly different context, Petros \[[@CIT0013]\]: USL laxity sets in motion a cascade of events resulting in the "posterior fornix symptoms", [Fig. 1](#F0001){ref-type="fig"}. Tightening the USLs reverses this cascade, resulting in symptom cure \[[@CIT0006], [@CIT0010], [@CIT0011]\].
Waliszewski at al. \[[@CIT0013]\] described the urethra and urinary bladder as a complex dynamical system. Its function was described in a holistic manner by a family of fractal dimensions. All urodynamic curves analyzed in that pilot study \[[@CIT0013]\] possessed fractal structure. That structure has been identified either in the filling phase or in the voiding one by the size--frequency algorithm. It has a cumulative normal distribution profile, and, therefore, can be described by the appropriate power law; the prerequisite condition of the existence of the scale--invariant fractal structure. In some cases of long--lasting obstruction, the urethra--urinary bladder system revealed a tendency towards chaos since one of the holistic parameters, the Hurst coefficient H, reached values greater than 0.5 in the voiding phase. Certainly, the holistic approach is a way to go in future research on urethra and urinary bladder. In particular, research *in silico* may help to identify the appropriate models, and to test hypotheses \[[@CIT0012]\].
Next steps -- a more scientific validation of the Integral System {#S20010}
-----------------------------------------------------------------
Given the issue of non--linearity, we see scientific validation of anatomically based management systems such as ITS in the future ultimately as a statistical exercise. A very large data base with many participating surgeons will be required before an accurate percentage incidence can be assigned for a symptom in a particular zone and whether a particular treatment is effective.
It is likely that at some stage an anonymous data base posted on the web by every surgeon performing a procedure may become compulsory. Relevance of the data posted will be a major challenge. Meanwhile, case report series on safety, efficacy, complication rates \[[@CIT0005]--[@CIT0008], [@CIT0010], [@CIT0011]\], subject though they may be to variation and non--linearity, remain an important method of assessment. Case report series build an accumulated body of knowledge consistent with our 2500 year old medical tradition going back to the time of Hippocrates, as does the understanding of Chaos Theory and non--linearity, as evidenced by Marcus Aurelius's writings \[[@CIT0001]\].
CONCLUSIONS {#S0011}
===========
Non--linearity and chaotically influenced feedback mechanisms are fundamental to Nature. They are the key to understanding the complexity of pelvic floor anatomy and function, day to day symptom variation in an individual patient, anomalous and varied experimental results, even how the randomness of initial experimental results may influence a whole new direction in surgery in a positive or negative way.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#s0005}
===============
Parkinson\'s disease (PD) is a neurodegenerative condition that clinically manifests with abnormalities in movement, olfactory senses, gastrointestinal function and an overall decline in cognitive abilities as the disease progresses. These symptoms are associated with the progressive loss of dopaminergic neurons in the substantia nigra [@bib1]. Sharing with other neuropathologies, including dementia with Lewy bodies and multiple systems atrophy, the majority of PD patient postmortem brains exhibit accumulation of α-synuclein (αSyn) protein in highly phosphorylated, ubiquitinated and insoluble aggregates observed as Lewy bodies and Lewy neurites [@bib2], [@bib3], [@bib4]. Recent studies indicate that αSyn may be released from neurons and seed aggregations in adjacent neurons thus propagating disease progression [@bib5].
Recently studies have started to investigate how αSyn intracellular inclusion formation impacts recipient neuronal survival using a model in which αSyn pre-formed fibrils (PFFs) can enter the cell and recruit endogenous αSyn in the neuron to form insoluble, ubiquitinated and phosphorylated high molecular weight αSyn, characteristic of those in human Lewy bodies and Lewy neurites [@bib5], [@bib6], [@bib7]. Since one of the strategies to help maintain intracellular protein homeostasis is through macroautophagy [@bib8], determining the impact of macroautophagy regulators on accumulation of αSyn aggregates and neuronal survival in response to PFF exposure is thus essential to understanding PD.
Autophagy can be activated by the inhibition of the mammalian target of rapamycin (mTOR) by rapamycin. Many *in vitro* and *in vivo* studies have reported positive results using rapamycin to mitigate the effects of αSyn toxicity [@bib9]. However, clinical use of rapamycin carries with it disadvantages that make its prolonged use in humans undesirable [@bib10]. In addition, previous studies have demonstrated that autophagy induction by starvation or rapamycin post PFF transduction did not decrease PFF-induced intracellular αSyn aggregations and further exacerbated cell death [@bib6]. Given that targeting of mTOR has limited therapeutic applications in humans in the realm of neurodegenerative treatments, compounds that activate autophagy independently of mTOR have been tested. One compound that has been found to be effective in activating autophagy in cell lines is the disaccharide trehalose [@bib11], which appears to mediate its effects through both initiation of autophagy and activation of TFEB, turning on genes for increased lysosomal biogenesis [@bib12], [@bib13], [@bib14]. Important for our study, neurons lack the enzyme trehalase that breaks down trehalose, thus the level of trehalose should be persistent over long periods of time in culture [@bib15]. In humans, a single bolus of 50 g or less has been deemed safe, and in 2000 the US FDA gave notice that trehalose is generally regarded as safe for human consumption [@bib16]. Although its administration has not been studied in the context of humans and PD, trehalose decreases α-synuclein in PC12 cells [@bib11], decreases Tau accumulation in mice [@bib17], and reduces neurodegeneration in amyotrophic lateral sclerosis [@bib18], Huntington\'s [@bib19], and Alzheimer\'s [@bib20] disease models. In this study we specifically test the effects of trehalose on decreasing αSyn aggregation and cell death upon exposure to PFFs, thus helping to evaluate the therapeutic potential of trehalose in PD.
2. Methods {#s0010}
==========
2.1. Cell culture {#s0015}
-----------------
Primary wildtype (C57BL/6 strain bred in house from mice ordered from Charles River and WT mice bred from a cathepsin D knockout colony) [@bib21], [@bib22], [@bib23], or αSyn knockout (The Jackson Labs C57BL/6N-Snca^tm1Mjff^/J) [@bib24] cortical mouse neurons were derived from p0 pups. All mouse experiments were done in compliance with the University of Alabama at Birmingham Institutional Animal Care and Use Committee guidelines (IACUC09019 and IACUC20354). Briefly, cells were seeded on plates coated with Poly-[L]{.smallcaps}-Lysine solution containing 0.1 mg/mL Poly-[L]{.smallcaps}-Lysine, 50 mM Boric acid, and 10 mM Borax for a minimum of 1 h. Cells were grown and maintained in Neural Basal A medium supplemented with Glutamax, Penn/Strep, and B27 neuronal supplement. Treatments were performed by removing half the media and adding back the same volume of media with the treatment media. Tissue culture medium and reagents were obtained from Life Technologies. Trehalose (T0167--25G), chloroquine (C6628--25G) sucrose (S-5016) and AraC (C1768-500MG) were obtained from Sigma.
2.2. Generation of αSyn PFFs {#s0020}
----------------------------
A human wildtype αSyn gene was cloned into pRK172 and expressed in *E. coli* as previously reported [@bib25]. Bacteria grown under antibiotic selection were harvested, homogenized and dialyzed before purification through size exclusion and ionic exchange columns. Five mg/mL of protein was incubated at 37 °C for 1 week to produce fibrils. Before applying to the cells, fibrils were sonicated 60 times over 40 s [@bib6]. Using the Pierce LAL chromogenic endotoxin quantification kit, we have determined that ≤0.004 ng/mL of endotoxin was present in the PFF samples.
2.3. Cell viability {#s0025}
-------------------
Cell viability was measured in two ways. First, viability was measured utilizing the trypan blue exclusion method. Cells were trypsinized and then trypan blue was added to the cells. Cells excluding the dye were counted. Second, viability was measured using the MTT cell death/proliferation assay. Briefly, a media and tetrazolium salt mixture was added to the cells where it is reduced to insoluble formazan crystals. These crystals were dissolved in DMSO and measured at 550 nm using a plate reader. Cells were plated on 96-well plates at 80,000 cells per well for both assays.
2.4. Immunocytochemistry {#s0030}
------------------------
Cells were seeded at 240,000 cells per well on autoclaved glass coverslips that were placed in 24-well plates. After treatment, cells were fixed with a mixture of 4% paraformaldehyde and 4% sucrose. The addition of 1% Triton X-100 to the fixative was used to determine the soluble from insoluble protein in the cell. After fixing, cells were permeabilized with 0.1% Triton X-100 and then blocked with 3% BSA in PBS. Cells were probed with antibodies for p-Ser-129 using either Affinity Bioreagents (PA1-4686 1:2000) for [Fig. 1](#f0005){ref-type="fig"}G or Covance-81A (MMS-5091 1:5000) for the remaining figures. For non-αSyn staining, cells were plated and fixed as above but were blocked with 10% horse serum and 5% FBS in PBS. Cells were probed with either LC3 (L8918 Sigma 1:500), MAP2 (Sigma M4403 1:1000), or GFAP (Dako Z0334 1:500). Alexa Fluor 488 (Invitrogen A11001, A11008 1:500) or 568 (Invitrogen A11004 1:500) secondary antibodies were subsequently added to the wells. Cells were then counter stained with nuclear dye Hoechst 33342 (Sigma 861405) and mounted with Fluoromount-G (Southern Biotechnology). All images were acquired using a Leica TCS SP5 V confocal laser scanning microscope.
2.5. Western blot analysis {#s0035}
--------------------------
For immunoblot analysis, cells were plated at 240,000 cells per well on 48 well plates or 480,000 per well on 24 well plates and then washed with ice cold PBS and then lysed with RIPA buffer (50 mM Tris pH 7.8, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100%, and 0.1% SDS) in the presence of protease (Roche) and phosphatase inhibitors (Sigma) after treatments. After 20 min on ice, cells were scraped from the wells and placed in 1.5 mL Eppendorf tubes for centrifugation at 16,800×*g* for 20 min at 4 °C. Protein content in supernatant was determined by DC Protein assay (Bio-Rad). Equal amounts of protein for each sample were loaded and separated by SDS-PAGE using 12% or 15% gels. Protein was wet-box transferred to PVDF membranes and probed with the following antibodies: LC3 (Sigma L8918 1:2000), p62 (Abnova H00008878-M01 1:2000), α-synuclein (Santa Cruz sc-7011-R 1:2500), and β-actin (Sigma A5441 1:2500). Blots were visualized by film or Amersham™ Imager 600 (GE Healthcare Biosciences; Pittsburg PA).
2.6. Statistical analysis {#s0040}
-------------------------
All data are reported as the mean±SEM, p values of less than 0.05 were deemed statistically significant after being analyzed by Student\'s *t*-test or ANOVA.
3. Results {#s0045}
==========
3.1. αSyn Pre-formed fibrils (PFF) induce intracellular αSyn protein aggregation {#s0050}
--------------------------------------------------------------------------------
To measure the dose-dependent effects of extracellular exposure to αSyn pre-formed fibrils, we cultured primary cortical neurons from mouse pups, and at DIV7 exposed them to human wildtype αSyn PFFs for 14 days at concentrations 0.1--2.5 µg/mL for western blot analysis of total αSyn and immunocytochemistry for insoluble S129 p-αSyn. Western blot analysis revealed a small but significant increase in the monomeric αSyn at 2.5 µg/mL of PFFs. An approximately 2-fold increase of total αSyn (monomeric and high molecular weight αSyn) was found in response to PFF at concentrations of 1.0 and 2.5 µg/mL ([Fig. 1](#f0005){ref-type="fig"}A-C). Whether accumulation of αSyn is more severe in response to PFF, compared to un-aggregated monomeric αSyn, was determined by exposing cells to similar concentrations of monomeric αSyn. In the latter case, we observed no changes in the αSyn monomer, whereas a modest increase of total αSyn was observed with 1 µg/mL and 2-fold increase with exposure to 2.5 µg/mL monomeric αSyn ([Fig. 1](#f0005){ref-type="fig"}D-F).
To determine if these αSyn aggregates shared common features with both classical Lewy bodies and previously characterized PFF-induced aggregates [@bib5], [@bib6], [@bib7], we performed immunocytochemistry analyses after exposing fixed cells to 1% Triton X-100 [@bib6]. As shown in [Fig. 1](#f0005){ref-type="fig"}G, S129 phosphorylated αSyn that is insoluble in 1% Triton X-100 was significantly increased in response to 2.5 µg/mL PFF. Under the same condition, exposure of cells to 2.5 µg/mL of αSyn monomer resulted in no p-αSyn accumulation as demonstrated by immunocytochemistry analyses. αSyn knockout cells exposed to 2.5 µg/mL PFFs also resulted in no insoluble p-αSyn under the same condition indicating endogenous αSyn is required for accumulation of insoluble p-αSyn ([Fig. 1](#f0005){ref-type="fig"}H).
3.2. Effects of PFFs on autophagy markers {#s0055}
-----------------------------------------
Previous studies have demonstrated that mouse primary hippocampal neurons exposed to 5 µg PFFs for 14 days exhibit increased LC3-II and levels of Triton X-100 insoluble p62 [@bib26]. To determine whether this occurs at a lower concentration with primary neurons from a different brain region, we performed western blot analyses of LC3 and p62 after exposing primary cortical neurons to PFFs for 14 days ([Fig. 2](#f0010){ref-type="fig"}). We found that increasing concentrations of PFFs up to 2.5 µg did not significantly change LC3-II or total levels of soluble p62 ([Fig. 2](#f0010){ref-type="fig"}A-D). The levels of LC3-II were confirmed by immunocytochemistry, where LC3-II puncta per cell were counted and no significant difference was observed between control and cells exposed to PFFs ([Fig. 2](#f0010){ref-type="fig"}E, F). Nor did we observe changes in either p62 or LC3-II in cells exposed to non-aggregated αSyn monomer for 14 days ([Fig. 2](#f0010){ref-type="fig"}G-J).
3.3. Effects of trehalose on neuronal autophagy {#s0060}
-----------------------------------------------
Previous studies have demonstrated that autophagy induction by starvation or rapamycin post PFF transduction did not decrease PFF-induced intracellular αSyn aggregations and only exacerbated cell death [@bib26]. Here we investigated whether induction of autophagy using trehalose attenuates PFF-induced αSyn accumulation thus enhancing neuronal survival. We exposed neurons to 25 mM trehalose for 14 days and found significant increases in LC3 puncta as assessed by immunocytochemistry analyses ([Fig. 3](#f0015){ref-type="fig"}A, B).
To determine if autophagic flux through the autophagy pathway was increased, we exposed cells from 1 to 14 days with 25 mM trehalose, and then with 40 µM Chloroquine (CQ) for 5 h which prevents lysosomal degradation of autophagosomes. We found that trehalose alone significantly increased LC3-II beginning at 3 days that continued to increase until the conclusion of the experiment at 14 days. CQ induced significant increases in LC3-II over non-CQ treated neurons for all time points. However, only at 14 days did the trehalose and CQ treated cells exhibit significantly more LC3-II than the CQ only control cells, indicating increased autophagic flux at 14 days ([Fig. 3](#f0015){ref-type="fig"}C, D).
3.4. Effects of trehalose on αSyn protein aggregation and autophagy {#s0065}
-------------------------------------------------------------------
To determine if trehalose impacts PFF-induced αSyn aggregation, we added 25 mM trehalose to cells at the time of PFF exposure and performed immunocytochemistry of 1% Triton X-100 insoluble p-αSyn over a range of PFF concentrations with and without trehalose for 14 days. Without trehalose, there was a strong trend toward increased p-αSyn at 0.1 µg/mL that became significant at 1.0 µg/mL and remained at similar levels at 2.5 µg/mL. In the presence of trehalose, similar trends of p-αSyn were observed with significant increases seen at 1.0 µg/mL and continuing to 2.5 µg/mL similar to cells without trehalose. No statistical differences were observed for cells exposed to PFFs with or without trehalose over a range of PFF concentrations, indicating that trehalose failed to remove p-αSyn aggregates ([Fig. 4](#f0020){ref-type="fig"}A, B). To further confirm these findings, we measured total αSyn in the presence or absence of both trehalose and PFFs by western blot. Trehalose exposure alone for 14 days did not change endogenous αSyn levels, while at 1.0 µg/mL ([Fig. 4](#f0020){ref-type="fig"}C-E) or 2.5 µg/mL ([Fig. 4](#f0020){ref-type="fig"}F, G) PFFs significantly increased total high molecular weight αSyn, even in the presence of trehalose, again suggesting trehalose fails to remove PFF induced aggregations.
Noting that trehalose alone increases LC3-II ([Fig. 3](#f0015){ref-type="fig"}), we measured whether this is still the case in the presence of PFF by western blot analyses lysates from cells exposed for 14 days to 25 mM trehalose, 2.5 µg/mL PFFs, or both. No significant changes were noted in LC3-I. However, we again observed increases in LC3-II with trehalose alone but not PFF alone. LC3-II remained elevated in the presence of both trehalose and PFFs at similar levels as trehalose alone ([Fig. 4](#f0020){ref-type="fig"}F, H, I).
3.5. Trehalose impacts cell viability {#s0070}
-------------------------------------
Although trehalose does not appear to remove αSyn aggregations, it may help remove toxic species of αSyn that are below detection sensitivity. Alternatively, it may physically interact with the αSyn aggregates reducing their toxicity or assisting in clearing cellular organelles damaged by αSyn, and thus attenuate cell death. To test whether trehalose alone alters the basal viability of neurons in culture, we exposed primary neurons to 3 different concentrations of trehalose over a 15-d period. We saw no change in cell viability at 1 mM but beginning at 10 mM and continuing to 25 mM we saw significant increases in cell viability with trehalose exposure alone ([Fig. 5](#f0025){ref-type="fig"}A). Next we determined if the enhancement of viability was specific to trehalose alone or if another disaccharide could have similar effect. We thus compared cell viability after exposing cells to either trehalose or a similar disaccharide, sucrose, for 14 days both at 25 mM concentrations. Sucrose had no impact on cell viability while 25 mM trehalose increased the number of surviving cells ([Fig. 5](#f0025){ref-type="fig"}B).
It is currently unknown whether trehalose will increase glial cell proliferation in neuronal culture, even while using Neurobasal medium without serum. To rule out the possible effects of trehalose inducing glial proliferation, we used a glial cell proliferation inhibitor, cytosine arabinoside (AraC). As before, trehalose increased cell viability at 10 and 25 mM. At 10 mM trehalose, AraC slightly blunted cell viability but still trended upwards. At 25 mM trehalose, AraC did not blunt cell viability, indicating that glial proliferation is not responsible for the enhanced viability in response to trehalose ([Fig. 5](#f0025){ref-type="fig"}C). Furthermore, we confirmed this finding by comparing the number of GFAP positive cells to the total number of nuclei and determined that only approximately 5% of cultures stained positive for GFAP and that this number did not change with Trehalose or PFF exposure ([Supplemental Fig. 1](#s0090){ref-type="sec"}).
To determine if PFFs induce cell death and are more toxic than monomeric αSyn, we compared cells exposed to both 2.5 µg/mL αSyn monomer and PFFs for 14 days. We found that PFFs induced significant cell loss, while treatment with αSyn monomer had no effect on cell viability, indicating that αSyn protein in its monomeric state is not sufficient to decrease viability ([Fig. 5](#f0025){ref-type="fig"}D). Additionally, cells from αSyn KO mice exposed to PFFs did not show any decrease in cell viability, further demonstrating that endogenous αSyn is required to potentiate cell toxicity in response to PFFs ([Fig. 5](#f0025){ref-type="fig"}E). To determine if trehalose would enhance cell survival in the presence of PFFs, we exposed cells for a period of 14 days to 3 concentrations of PFFs with or without 25 mM trehalose. Using an MTT based cell death assay, we measured the effect of PFFs on cell viability. We observed that at 2.5 µg/mL PFFs decreased cell viability. In cells exposed to trehalose, there was an increase in basal cell viability after 14 days as observed above, and an increase in cell viability after 2.5 µg/mL PFF+trehalose compared to the same concentration of PFF -trehalose ([Fig. 5](#f0025){ref-type="fig"}F). Cell survival is comparable in +trehalose+PFF compared to --PFF-trehalose. However, in the presence of trehalose, PFFs still induced cell death compared to +trehalose+PFF indicating that PFF toxicity still persists.
4. Discussion {#s0075}
=============
In this study we have shown that exogenously applied αSyn pre-formed fibrils (PFFs) induce significant protein aggregation in primary neurons ([Fig. 1](#f0005){ref-type="fig"}A-C, G) consistent with prior reports [@bib6], [@bib7]. The addition of αSyn monomer does not induce S129 phosphorylated insoluble inclusions in primary neurons ([Fig. 1](#f0005){ref-type="fig"}D-F, H). Furthermore, endogenous αSyn is required for the propagation of PFFs as αSyn KO neurons exposed to 2.5 µg/mL PFFs do not accumulate insoluble S129 phosphorylated αSyn inclusions ([Fig. 1](#f0005){ref-type="fig"}H).
In our study, PFFs did not change levels of p62 and LC3-II as previous studies have shown ([Fig. 2](#f0010){ref-type="fig"}A-D). This is likely due to a lower amount of PFFs added, and that in our system we used cortical neurons where Lewy body pathology occurs later in Parkinson\'s disease [@bib27]. Additionally, our cultures were isolated from postnatal mice opposed to embryonic cultures, thus it is possible that a higher threshold is required for PFFs to attenuate autophagy. By these measures our model used a less severe exposure to PFFs and concurrently we observed a less severe phenotype.
Most studies in cell cultures use acute high doses of ≥100 mM trehalose, demonstrating in the short term that trehalose enhances autophagy. In these models, up-regulation of autophagy was implicated as being protective [@bib28], [@bib29]. However, in animal models, chronic trehalose exposure is necessary, and may also be for future patients. To determine the effects of chronic lower concentrations of trehalose in culture, we for the first time show a low concentration of 25 mM trehalose over a 14-d period enhances autophagic flux and increases basal cell viability. And while trehalose in the presence of PFFs was unable to enhance viability to that of cells exposed only to trehalose, it was able to restore viability to the level of untreated controls ([Fig. 5](#f0025){ref-type="fig"}F). This protective effect is intriguing especially that the survival of the minor population of astrocytes was not changed by either PFF or trehalose.
It is currently unclear how trehalose enhances cell survival. Trehalose increased LC3-II in the absence and presence of PFF ([Fig. 4](#f0020){ref-type="fig"}I-K). Nonetheless, trehalose has been shown to alter the formation of A53T fibrils, tau fibrils, amyloid-beta and polyQ proteins [@bib19], [@bib20], [@bib29], [@bib30], [@bib31], and thus is proposed to act as a molecular chaperone. In yeast, it has been shown that under the protein misfolding and aggregation stress of desiccation, trehalose is essential for maintaining viability by stabilizing cellular proteins, similar in function to HSP104 [@bib32]. It also can interact in synergy with existing chaperones, as is the case with its interaction with p26 to maintain functional citrate synthase protein during heat shock *in vitro* [@bib33]. Furthermore, unlike rapamycin, trehalose did not exacerbate PFF toxicity. Additionally, trehalose has been shown effective in ameliorating symptoms in an A53T AAV rat model of PD administered *via* drinking water, suggesting that trehalose may also protect against PFF-induced neurodegeneration in animal models [@bib34]. Nonetheless, this study corroborates previous work that PFFs seem to resist degradation even with enhancement of macroautophagy. Future studies are still urgently needed to identify novel pathways that promote αSyn degradation and enhance neuronal survival in response to exposure to αSyn fibrils.
Conflict of interest {#s0080}
====================
We have no conflict of interest.
Appendix A. Supplementary material {#s0090}
==================================
Supplementary material **Figure 1:** Primary cortical neurons exposed to control, 25 mM trehalose, 2.5 µg/mL PFFs, or both for 14 days were probed for MAP2, GFAP and counterstained with nuclear dye Hoechst (blue), 40× magnification. GFAP positive cells were quantified and normalized to total cell number determined by nuclear dye. Data represent mean±SEM (n=3)..
We would like to thank Dr. Laura Volpicelli-Daley and members of the Zhang and Darley-Usmar laboratories for discussion and technical assistance. This work was supported by NIHR01-NS064090 (to JZ).
Supplementary data associated with this article can be found in the online version at [doi:10.1016/j.redox.2016.12.032](http://doi:10.1016/j.redox.2016.12.032){#ir0005}.
{#f0005}
{#f0010}
{#f0015}
{#f0020}
{#f0025}
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Intussusception is a common surgical emergency in infants and children. The incidence of intussusception is approximately one to four per 2,000 infants and children.[@b1-tcrm-11-1837] The diagnosis of intussusception was confirmed by clinical and radiological findings. Common signs and symptoms included colicky abdominal pain, vomiting, palpable abdominal mass, and currant jelly stool. A plain abdominal X-ray might show a soft tissue mass, target sign, meniscus sign, and absence of air in ascending colon, and/or small-bowel dilatation.[@b2-tcrm-11-1837] The ultrasound to diagnose intussusception was performed from the findings of the doughnut and pseudokidney signs that indicate the bowel-in-bowel condition characteristic of the intussusception.[@b3-tcrm-11-1837]
According to the Brighton Collaboration Intussusception Working Group,[@b4-tcrm-11-1837] the case definition of intussusception is given as the invagination of one segment of intestine into a segment of distal intestine. The level 1 diagnosis certainties are the surgical, and/or radiologic, and/or autopsy criteria. The level 2 diagnostic certainty is the clinical criteria which include two major criteria or one major with three minor criteria. The diagnostic major criteria consist of the evidence of intestinal obstruction, intestinal invagination, and intestinal vascular compromise or venous congestion. The minor criteria are the predisposing factors (age \<1 year, male sex), abdominal pain, vomiting, lethargy, pallor, hypovolemic shock, and abnormal nonspecific bowel gas pattern in abdominal radiograph.[@b4-tcrm-11-1837]
All of the cases that met the radiologic criteria were reviewed. After the case of intussusception was diagnosed, the method of treatment was chosen. The modality of treatments consisted of operative and nonoperative management. Contraindications for nonoperative management were hemodynamic instability, peritonitis, and/or abdominal signs of perforation on abdominal X-ray. Without contraindications, nonoperative methods included performing hydrostatic or pneumatic reduction. A previous literature review[@b5-tcrm-11-1837] of the success rates of both methods in another care unit showed significant differences. The hydrostatic reduction could be performed by the use of saline, barium, or another solution. The overall success rate of the nonoperative reduction ranged from 46% to 94% according to a review by Bekdash et al.[@b5-tcrm-11-1837]
In this study, we were interested in comparing the success rates of the hydrostatic and pneumatic reduction techniques. Air reduction was represented pneumatic reduction and barium reduction represented hydrostatic reduction. The comparative study of the success of both methods was done.
Methods
=======
This retrospective cohort study was approved by the Ethics Committee of Chiang Mai University Hospital and Siriraj Hospital. The data were obtained by chart review and electronic databases. Patient consent was not required in this retrospective study. All intussusception patients (ICD-10 code K56.1) who had visited Chiang Mai University Hospital and Siriraj Hospital from January 2006 to December 2012 were included in the study. The data collected included demographic data (sex, age, and bodyweight), symptoms (vomiting, abdominal pain, rectal bleeding, diarrhea, distention, constipation, and duration of symptoms), signs (temperature, palpable mass, and location of the mass), and investigations (white blood cell counts, neutrophils, electrolytes, abdominal radiography, and ultrasound findings). Specific radiography findings showed small-bowel obstruction and ultrasound showed poor prognostic signs such as thick peripheral hypoechoic rim, free intraperitoneum fluid, fluid trapped within intussusception, enlarged lymph node in intussusception, pathologic leading point, absence of blood flow in the intussusception.[@b1-tcrm-11-1837] Based on the methods of reduction used for treatment, the patients were grouped as pneumatic reduction group and barium reduction group. The outcome of the study was the success of nonoperative reduction.
We included all intussusception patients aged 0--15 years and excluded the patients who had contraindications for nonoperative reductions, which included peritonitis, perforation sign on abdominal radiographic study, and nonresponsive shock that required surgery.
The nonoperative methods followed were pneumatic reduction and barium reduction. These procedures were performed in well-hydrated children. The standard techniques of reduction comprised three repeated attempts of 3 minutes each. In Chiang Mai University Hospital, all patients received pneumatic reduction performed by a radiologist under fluoroscopic guidance. In Siriraj Hospital, pneumatic reduction was performed by a pediatric surgeon under ultrasound guidance and barium reduction was performed by a radiologist under fluoroscopic guidance. A Foley catheter was inserted via the anus of the patients and the buttocks were taped to prevent air or barium leakage. For the pneumatic reduction method, all patients received air pressure from 80 to 120 mmHg. For the barium reduction method, the barium bucket was hung 3 feet above the patients. Sedation drugs were given according to it's hospital sedation guidelines.
The success of reduction was determined by the disappearance of intussusception and the visualization of barium or air from cecum to ileum through ileocecal valve, or barium-or air-distended ileum and absence of intussusception after reduction by ultrasound examination.[@b6-tcrm-11-1837]
The statistical analysis was done by using commercial statistical software (STATA 11.0; StataCorp LP, College Station, TX, USA). The categorical descriptive data were reported as counts (N) and percentage (%). The categorical univariable analysis was done by Fisher's exact test. The numerical descriptive data were reported as mean and standard deviation. The numerical univariable analysis was done by Student's *t*-test or Mann--Whitney *U*-test. Many factors influence the failure of reduction techniques. It was reported that duration of symptoms, emesis, bloody stool, location of intussusception, and poor prognosis sign on ultrasound were associated with failure reduction.[@b7-tcrm-11-1837] The propensity score was used to adjust all the factors, which included sex, age group of 36 months, weight group of 8 kg, duration of symptoms for 48 hours, vomiting, abdominal pain, rectal bleeding, diarrhea, abdominal distension, constipation, temperature of 37.8°C, palpable abdominal mass, location of the mass, white blood cell count of 10,000/mm^3^, plain abdominal radiography showing bowel obstruction, and ultrasound showing poor prognostic sign by logistic regression. The propensity score was generated to estimate the probability of the choice of the method of reduction (pneumatic vs barium). The success of reduction was measured by risk ratio. A multivariable exponential risk regression analysis was performed to determine the success rates of two methods of reduction that were adjusted by propensity score. Statistical significance level was set as two-tailed with *P*-value \<0.05.
Results
=======
A total of 190 episodes of intussusception were identified among patients who visited Chiang Mai University Hospital and Siriraj Hospital. The summary of the epidemiological characteristics of all the patients are shown in [Tables 1](#t1-tcrm-11-1837){ref-type="table"} and [2](#t2-tcrm-11-1837){ref-type="table"}. Twenty patients were excluded due to contraindications and surgery after the diagnosis. One hundred and seventy episodes were included in this study ([Figure 1](#f1-tcrm-11-1837){ref-type="fig"}). The male to female ratio was 2:1. The median age of the patient was 9 months with a mean weight of 9.5 kg. The most common symptoms were vomiting, abdominal pain, and rectal bleeding (86.5%, 79.4%, and 70%, respectively). Diarrhea was found in 16.5% and constipation was found 10.6% of the patients. A palpable abdominal mass and abdominal distension were observed in 66.5% and 45.9% of the patients, respectively. The median duration of symptoms before presentation was 24 hours. Twenty two percent of the patients had fever. Plain abdominal radiography showed small-bowel obstruction in 66.9% of the patients. The ultrasonography before reduction showed at least one of the poor prognostic signs, as mentioned in the "Methods" section, in 43.9% of the patients. The most common location of the palpable mass was right upper quadrant and was found in 58% of the patients. The overall success rate in this study was 55.3%. The patients were divided into the pneumatic reduction group (111 patients) and the barium reduction group (59 patients). Comparison between the two groups is shown in [Tables 3](#t3-tcrm-11-1837){ref-type="table"} and [4](#t4-tcrm-11-1837){ref-type="table"}. Univariable analysis showed that the success rate of pneumatic reduction (61%) was significantly higher than that of barium reduction (44%) (*P*-value =0.036). The propensity score was used to control all the variables. The success rate of pneumatic reduction was 1.48 times more than that of barium reduction (*P*-value =0.036, confidence interval \[CI\] =1.03--2.13) as shown in [Tables 5](#t5-tcrm-11-1837){ref-type="table"} and [6](#t6-tcrm-11-1837){ref-type="table"}. Perforation after reduction was found in only one pneumatic reduction case, and the patient safely received a right hemicolectomy due to colonic gangrene and perforation.
Discussion
==========
The modalities for the management of intussusception included nonoperative and operative management. The patient with no contraindication received nonoperative management as the initial treatment. The choices available for nonoperative treatment of intussusception were hydrostatic and pneumatic reduction.
In 1885, intussusception was treated with laparotomy and had a high mortality rate of 70%. Treves also had some idea of reduction, but did not establish the rule. The rule of reduction might set for the pressure used and the time used for each attempts.[@b8-tcrm-11-1837] In 1935, Hipsley used hydrostatic pressure to reduce intussusception and proposed the technique of pressure reduction.[@b9-tcrm-11-1837] By that time, the reductions were performed hydrostatically. In 1986, a large intussusception study in People's Republic of China including 6,396 cases over a 13-year period were successfully reduced by air reduction with a success rate of 95%.[@b10-tcrm-11-1837] After that, there was a worldwide increase in the use of pneumatic reduction. Both hydrostatic and pneumatic reduction techniques had been performed in cases of feasibility in some health care institutes. Also, in Thailand, in 2011, Kruatrachue et al reported a switch from barium to air reduction since 1992 with the success rate of 68%.[@b11-tcrm-11-1837] In 2013, Bekdash et al collected results from series reports regarding the success rate of intussusception reduction to establish an index of successful reduction.[@b5-tcrm-11-1837] This recruited study used both air and barium for reduction depending on the radiologist or pediatric surgeon's preference, experience, and institutional setting. There were not many comparative studies between the success rate of hydrostatic and pneumatic reduction in the literature. In 2013, Fallon et al studied the risk factors for surgery in patients with intussusception and found that hydrostatic enema was a predictor for failed nonoperative reduction in univariable analysis.[@b12-tcrm-11-1837] In another collective review in 2004, Daneman and Navarro found that the success rate of pneumatic reduction was 51%--100% and that of hydrostatic reduction was 12.5%--95.5%.[@b13-tcrm-11-1837]
In this retrospective analysis, we found that the overall success rate was \~55%. The pneumatic reduction technique showed a success rate of 61% and hydrostatic reduction technique 44%. The results of our two institutional studies did not show a high success rate, which may be due to the symptom duration before hospital admission being quite a long period of time. Some of the cases were referred from remote provincial hospitals and patients had to travel long distances to receive treatment. We compared the result of pneumatic and hydrostatic reduction by multivariable analysis controlling the factors by the use of propensity score. So, the results of both the methods of reduction were adjusted to allow for comparison. The risk factors associated with failure of the reduction methods will be analyzed in future studies. The complication usually observed in both the methods was perforation after reduction, which was reported as 0%--5.9% in a previous study.[@b13-tcrm-11-1837] Our study did not focus on the complications but focused on the success rate. We found that \<1% of our study population sustained a perforation after reduction. Hence, the complication rates and the surgical findings will be discussed in the future studies.
Conclusion
==========
The method of nonoperative reduction of intussusception was dependent on the experience of the radiologist or pediatric surgeon and the hospital setting. We found the success rate of pneumatic reduction was 1.48 times more than that of barium reduction in this study. Both methods can be performed safely before operation if there are no contraindications. The risk factors associated with the failure of reduction methods will be included in our next study.
**Disclosure**
The authors report no conflicts of interest in this work.
{#f1-tcrm-11-1837}
######
Baseline characteristics of all children with intussusceptions in Chiang Mai University Hospital and Siriraj Hospital from 2006 to 2012 (all 190 cases)
Characteristics N \%
-------------------------------------------------------------------------- ------- --------
**Patient factors**
Sex
Male 128 67.37
Female 62 32.63
Age (month)[a](#tfn1-tcrm-11-1837){ref-type="table-fn"} 9 7--16
Weight (kg)[b](#tfn2-tcrm-11-1837){ref-type="table-fn"} 9.73 4.22
**Symptoms**
Vomiting 166 87.37
Abdominal pain 147 77.37
Duration of symptoms (hours)[a](#tfn1-tcrm-11-1837){ref-type="table-fn"} 24 20--48
Rectal bleeding 135 71.05
Distension 96 50.53
Diarrhea 32 16.84
Constipation 21 11.05
**Signs**
Temperature (°C)[b](#tfn2-tcrm-11-1837){ref-type="table-fn"} 37.34 0.69
Palpable mass 123 64.74
**Notes:**
Median, interquartile range;
mean, standard deviation.
######
Investigation, treatment, and outcome of all children with intussusceptions in Chiang Mai University Hospital and Siriraj Hospital from 2006 to 2012 (all 190 cases)
Characteristics N \%
------------------------------------------------------------------- -------- ---------------
**Investigations**
WBC count (/mm^3^)[a](#tfn3-tcrm-11-1837){ref-type="table-fn"} 12,000 9,030--15,800
Neutrophils (%)[b](#tfn4-tcrm-11-1837){ref-type="table-fn"} 56.56 16.78
Na (mmol/L)[b](#tfn4-tcrm-11-1837){ref-type="table-fn"} 136.67 4.26
Ultrasound (poor prognosis sign) 76 45.24
Location
Right lower quadrant 17 9.34
Right upper quadrant 101 55.49
Left upper quadrant 33 18.13
Left lower quadrant 29 15.93
In rectum 2 1.10
**Treatment**
Surgical (presence of contraindication for nonsurgical reduction) 20 10.53
Nonsurgical (170 cases) 170 89.47
Pneumatic reduction 111 65.29
Hydrostatic reduction 59 34.71
**Outcome (170 cases)**
Successful reduction 94 55.29
**Notes:**
Median, interquartile range;
mean, standard deviation.
**Abbreviation:** WBC, white blood cell count.
######
Baseline characteristics of children with intussusceptions who received nonoperative reduction by pneumatic reduction (n=111) and hydrostatic reduction (n=59)
Characteristics Pneumatic reduction, n (%) Hydrostatic reduction, n (%) *P*-value
-------------------------------------------------------------------------- ---------------------------- ------------------------------ -----------
**Patient factors**
Sex
Male 75 (65.79) 39 (34.21) 0.865
Female 36 (64.29) 20 (35.71)
Age (month)[a](#tfn6-tcrm-11-1837){ref-type="table-fn"} 10 (7--16) 8 (6--18) 0.449
Weight (kg)[b](#tfn7-tcrm-11-1837){ref-type="table-fn"} 9.23 (3.16) 9.89 (3.46) 0.211
**Symptoms**
Vomiting 96 (65.31) 51 (34.69) 1.000
Abdominal pain 85 (62.96) 50 (37.04) 0.237
Duration of symptoms (hours)[a](#tfn6-tcrm-11-1837){ref-type="table-fn"} 24 (19--48) 36 (24--48) 0.119
Rectal bleeding 76 (63.87) 43 (36.13) 0.601
Distension 41(52.56) 37 (47.44) 0.002
Diarrhea 19 (67.86) 9 (32.14) 0.831
Constipation 14 (77.78) 4 (22.22) 0.008
**Signs**
Temperature (°C)[b](#tfn7-tcrm-11-1837){ref-type="table-fn"} 37.19 (0.70) 37.47 (0.57) 0.009
Palpable mass 76 (67.26) 37 (32.74) 0.497
**Notes:**
Median (interquartile range);
mean (standard deviation).
######
Investigation of children with intussusceptions who received nonoperative reduction by pneumatic reduction (n=111) and hydrostatic reduction (n=59)
Characteristics Pneumatic reduction, n (%) Hydrostatic reduction, n (%) *P*-value
------------------------------------------------------------------ ---------------------------- ------------------------------ -----------
**Investigations**
WBC count (/mm^3^)[a](#tfn8-tcrm-11-1837){ref-type="table-fn"} 12,675 (9,260--17,040) 10,830 (8,600--14,570) 0.079
Neutrophils (%)[b](#tfn9-tcrm-11-1837){ref-type="table-fn"} 57.31 (17.07) 54.83 (15.63) 0.365
Na (mmol/L)[b](#tfn9-tcrm-11-1837){ref-type="table-fn"} 137.51 (3.89) 136.05 (4.41) 0.031
K (mmol/L)[b](#tfn9-tcrm-11-1837){ref-type="table-fn"} 4.41 (4.02) 3.84 (0.65) 0.289
Cl (mmol/L)[b](#tfn9-tcrm-11-1837){ref-type="table-fn"} 104.52 (0.48) 100.95 (5.53) \<0.001
Total CO~2~ (mmol/L)[b](#tfn9-tcrm-11-1837){ref-type="table-fn"} 19.42 (2.95) 19.5 (4.55) 0.895
Location 0.117
Right lower quadrant 6 (40.00) 9 (60.00)
Right upper quadrant 66 (68.04) 31 (31.96)
Left upper quadrant 21 (67.74) 10 (32.26)
Left lower quadrant 17 (73.91) 6 (26.09)
In rectum 0 (0.00) 1 (100.00)
Plain abdominal X-ray (small-bowel obstruction) 65 (60.75) 42 (39.25) 0.026
Ultrasound (poor prognosis sign) 11 (16.18) 57 (83.82) 0.001
**Notes:**
Median (interquartile range);
mean (standard deviation).
**Abbreviation:** WBC, white blood cell count.
######
Outcome of children with intussusceptions who received nonoperative reduction by pneumatic reduction (n=111) and hydrostatic reduction (n=59)
Characteristics Success rate, n (%) Failure, n (%) *P*-value
------------------------ --------------------- ---------------- -----------
Method of reduction 0.036
Pneumatic reduction 68 (61.26) 43 (38.74)
Hydrostatic reduction 26 (44.07) 33 (55.93)
######
Multivariable risk ratio of successful reduction of intussusception adjusted by propensity score
Characteristics Crude relative risk (95% confidence interval) *P*-value Multivariable risk ratio (95% confidence interval) *P*-value
-------------------------------------------------- ----------------------------------------------- ----------- ---------------------------------------------------- -----------
Method of reduction (pneumatic over hydrostatic) 1.39 (0.88--2.18) 0.153 1.48 (1.03--2.13) 0.036
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION {#sec1-1}
============
Malignancy of the thyroid gland is the most common in the endocrine system and has been identified into different subtypes depending on the cells or growth pattern.\[[@ref1][@ref2]\] Papillary thyroid carcinoma (PTC), a well-differentiated tumor, is the most common histological subtype including 80%--90% of all thyroid cancer.\[[@ref3][@ref4]\]
It is generally an asymptomatic, slow-growing nodule with desirable prognosis.\[[@ref2][@ref4]\] PTC almost has a lymphatic spreading to regional lymph nodes. Distant metastasis with bloodstream dissemination are less frequent\[[@ref2][@ref3]\] and they have poor prognosis and decreased survival rate.\[[@ref4]\]
Distant metastasis more frequently involves lungs and bones such as sternum, vertebrae, pelvis, ribs and femur.\[[@ref2][@ref3][@ref5]\] Mandibular metastasis is rare, and it counts about 1% of oral malignancies.\[[@ref1][@ref6]\] This tumor almost presents in the premolar-molar and angle-ramus region of the mandible.\[[@ref3][@ref7]\]
We present a rare case with the mandibular metastasis from PTC, and a systematic review of the literature is performed.
CASE REPORT {#sec1-2}
===========
A 68-year-old man was referred to the cancer institute of Tehran University of medical sciences with a history of painless swelling in the right side of the mandible \[[Figure 1](#F1){ref-type="fig"}\] and a medical history of thyroidectomy 3 years ago because of PTC. After that, he received suppression therapy with levothyroxine. One year later, follow-up in the whole body scan with I-131 showed no evidence of residual thyroid tissue in thyroid bed and negative for local and distant metastasis. On the second year after thyroidectomy, he complained of swelling in the mandible then admitted. Laboratory tests were normal except elevated thyroid-stimulating hormone level that showed subclinical hypothyroid. Abdominal-computed tomography (CT) scan with and without contrast was normal although lung CT scan shows sparse hypodense nodules in bilateral hemithorax that suspect metastasis. Neck sonography revealed disruption of the mandibular cortex and a hypoechoic lesion in the mandibular bone that posed to bony mass with a pathological fracture. Furthermore, hyperechoic of related soft tissue and multiple lymph nodes was observed. A hypoechoic region closed to the cortex probability proposed hematoma. For the second time, the whole body scan with TC-99 shows hyperactive hotspots in the right side of the mandible and degenerative changes in the body joints \[[Figure 2](#F2){ref-type="fig"}\]. No other markable abnormality in other skeletal bones were reported. Panoramic radiography shows an invasive ill-defined radiolucent lesion in the body, angle and ascending ramus of the right side of the mandible. The destruction of the lower cortex and pathological fracture of medial border of ramus was also observed. The cortical border of the inferior alveolar canal was not detected \[[Figure 3](#F3){ref-type="fig"}\]. CT scan without contrast from maxillofacial region shows an expansile lesion that measured about 50 mm × 32 mm which laterally extends and destructs the mandibular body and causes thinning buccal plate, medially extension to the mylohyoid muscle is seen. The lesion also involved the mandibular canal and mental foramen \[Figure [4a](#F4){ref-type="fig"}--[d](#F4){ref-type="fig"}\]. Perforation of the lingual plate and inferior border of the mandible were also observed \[[Figure 5](#F5){ref-type="fig"}\]. Hematologic malignancies or minor salivary gland tumors with same features can be suggested. A thick cortex 15 mm × 13 mm level I LAPD is observed on the right side.
{#F1}
{#F2}
{#F3}
{#F4}
{#F5}
Then, the incisional biopsy was performed. Microscopic examination shows bone trabeculae infiltrated by a malignant epithelial neoplasm composed of papillary structures which are characterized by distinctive nuclear features. The papillae are formed by fibrovascular cores, which are covered by cuboidal to columnar neoplastic epithelial cells. The crowded cells show round-to-ovoid nuclei with frequent indentations and typical clefted or grooved appearance. Furthermore, some nuclei have an empty or clear appearance. The cell cytoplasm is typically smooth and eosinophilic. The fibrous stroma shows lymphoplasmacytic infiltration. Perineural and lymphovascular invasion are identified \[Figure [6a](#F6){ref-type="fig"}--[d](#F6){ref-type="fig"}\]. Due to the history of malignancy of thyroid and microscopic appearance, the diagnosis was metastatic carcinoma. The mentioned diagnosis was confirmed immunohistochemically by thyroid transcription factor 1 and thyroglobulin markers \[Figure [7a](#F7){ref-type="fig"} and [b](#F7){ref-type="fig"}\].
{#F6}
{#F7}
The patient undergoes hemimandibulectomy and the right neck dissection level I-V. The mandible was reconstructed with a costochondral rib graft \[[Figure 8](#F8){ref-type="fig"}\]. The pathology showed one reactive lymph node in level I. It was reported metastatic papillary thyroid carcinoma.
{#F8}
DISCUSSION {#sec1-3}
==========
Metastatic tumor to the oral region is a rare event that comprises 1% of all oral malignancies.\[[@ref3][@ref8]\] The primary malignancy that is giving more frequently metastasis to oral cavity differed between genders. These are breast cancer for a woman and lung cancer for men.\[[@ref9][@ref10][@ref11]\]
Other common primary sites are colon, kidney, prostate, bone, liver, adrenal gland and female\'s genital organs.\[[@ref3][@ref9][@ref12]\] Metastatic thyroid carcinoma to oral tissue is not very frequent. Metastatic PTC tumors to jaw include 4%--6.5% in all metastatic tumors to the jaw.\[[@ref3][@ref9]\]
Most metastasis bones from PTC are being sternum, vertebrae, skull, pelvis, ribs and femur.\[[@ref5][@ref10][@ref13]\]
A review literature of available published cases for metastatic thyroid carcinoma to oral tissues revealed 77 cases with their summarized features in [Table 1](#T1){ref-type="table"}.
######
Summarizes of published case reports with metastatic thyroid carcinoma to the oral cavity (77 cases)
Authors Age Gender Location Histological diagnosis Time to metastasis Treatment to metastasis Other metastatic sites Outcome (length of follow-up)
---------------------------------------------- ------------------ ------------------------------ ----------------------------------------------- ------------------------------------------- ---------------------------- ----------------------------------------------- ----------------------------------- ---------------------------------
Meyer and Shklar, 1965\[[@ref3]\] 51 Female Mandible Adenocarcinoma NA NA None NA
Soumar *et al*., 1970\[[@ref16]\] 54 Female Mandible NA NA NA NA NED (36 months)
McDaniel *et al*., 1971\[[@ref3]\] 77 Female Right mandible Follicular thyroid Ca First manifestation Hemimandibulectomy and parotidectomy None NED (4 years )
48 Female Left mandible Follicular thyroid Ca First manifestation NA NA NA
Al Ani 1973\[[@ref3]\] 60 Female Right mandible Follicular thyroid Ca First manifestation NA NA NA
Ripp *et al*., 1977\[[@ref3]\] 61 Female Right mandible Thyroid Ca First manifestation Cobalt teletherapy Widespread DOD ( 1 year )
Draper *et al*., 1979\[[@ref3]\] NA Female Mandible Follicular thyroid Ca NA Radiotherapy skull NA
Stypulkowska *et al*., 1979\[[@ref16]\] NA NA Mandible Adenocarcinoma 1 year symptomatic NA DOD (≤ 1 year )
NA NA Mandible Adenocarcinoma First manifestation symptomatic NA DOD ( ≤1 year )
Osguthorpe and Bratton, 1982\[[@ref3]\] 53 Male Right mandible Follicular thyroid Ca First manifestation Partial mandibulectomy None NED ( 3 years )
Nishimura *et al*., 1982\[[@ref1]\] 74 Female Right Mandible Follicular thyroid Ca First manifestation Chemotherapy Vertebra Alive ( 4 months )
51 Female Left mandible Papillary thyroid Ca (follicular variant) First manifestation Chemotherapy, radiotherapy Ribs , Lung Alive with disease (4.5 year)
Hemimandibulectomy
Parichatikanond *et al*.,1982\[[@ref3]\] 42 Female Left mandible Follicular thyroid Ca First manifestation Hemimandibulectomy NA NA
Tovi *et al*., 1984\[[@ref3]\] 33 Male Left mandible Follicular thyroid Ca First manifestation Radioactive Vertebra, mediastinum Died of thyroid
Iodine therapy Crisis (17 days)
Parietal bone, femur
Markitziu *et al*., 1986\[[@ref3]\] 69 Female Left mandible and parotid Papillary thyroid Ca (follicular variant) First manifestation Radiotherapy and chemotherapy NA NED (1.5 years)
Kahn and McCord 1989\[[@ref3]\] 82 Female Anterior mandible Follicular thyroid Ca 32 years Radiotherapy, radioactive Mediastinum, tibia, lungs, pelvis Died, NED ( 1.5 years )
Iodine therapy, resection
Whitaker *et al*., 1993\[[@ref3]\] 87 Male Dorsal tongue and lower lip Follicular thyroid Ca First manifestation Surgical removal, Radioactive Iodine therapy Lung Alive with disease (4 months)
Hefer, 1998\[[@ref16]\] 58 Male Maxilla Follicular thyroid Ca NA Resection NA NED (2 years)
Vural and Hanna 1998\[[@ref3]\] 64 Female Right Mandible infratemporal fossa Follicular thyroid Ca First manifestation Radical resection, radioactive iodine therapy Multiple skeletal Alive (6 weeks)
Agarwal *et al*., 1998\[[@ref3]\] 45 Female Left mandible Follicular thyroid Ca First manifestation Hemimandibulectomy NA Alive (2 weeks)
Erdag *et al*., 1999\[[@ref3]\] 53 Female Right mandible Papillary thyroid Ca (follicular variant) 4 years Radical Resection, Radioactive Iodine therapy Vertebra, Humerus Alive with disease (2.5 years )
Anil *et al*., 1999\[[@ref3]\] 61 Female Right Mandible Follicular thyroid Ca 8 years NA NA NA
Piatelli *et al*., 2000\[[@ref3]\] 54 Female Right Maxillary Gingiva Medullary thyroid Ca 4 months Surgical removal Cervical lymph nodes NED (4 years)
Thomas *et al*., 2001\[[@ref13]\] 61 Female Mandibular gingiva Follicular thyroid Ca First manifestation Radio Iodine ablation None NED (5 years)
Bhansali *et al*., 2003\[[@ref5]\] 60 Female Mandible, maxilla Papillary thyroid Ca (follicular variant) NA NA Rib NA
Colella *et al*., 2003\[[@ref3]\] 50 Female Right mandible Papillary thyroid Ca 5 years NA NA NA
Ostrosky *et al*., 2003\[[@ref3]\] 72 Male Anterior mandible Follicular thyroid Ca First manifestation Resection None NA
Bonder *et al*., 2006\[[@ref3]\] 65 Male Right mandible Hurtle cell thyroid Ca NA Supportive Widespread DOD (2 months)
Liu *et al*., 2007\[[@ref26]\] 66 Male Masticator space, mandible Papillary thyroid Ca (follicular variant) 22 years Radical Resection, radioactive iodine therapy Scapula NED (3 years)
Kaveri *et al*., 2007\[[@ref14]\] 65 Male Left mandible Follicular thyroid Ca First manifestation NA Ribs NA
Tamiolakis *et al*., 2007\[[@ref3]\] 69 Female Mandible Papillary thyroid Ca NA Inoperable NA DOD (1 month)
NA NA Mandible Thyroid Ca NA Resection NA NA
Antunes and Antunes 2008\[[@ref3]\] 13 Female Maxilla Papillary Adenocarcinoma NA NA NA NA
Araki *et al*., 2008\[[@ref16]\] 55 Female Left mandible Follicular thyroid Ca NA NA NA NA
Ismail *et al*., 2009\[[@ref3]\] 70 Female Left mandible Follicular thyroid Ca First manifestation NA Hilar lymph nodes NA
Algahtani *et al*., 2009\[[@ref3]\] 66 Female Left mandible Follicular thyroid Ca First manifestation Segmental mandibulectomy Cervical lymph nodes NA
Seoane *et al*., 2009\[[@ref16]\] 58 Female Tongue Papillary thyroid Ca NA NA NA NA
69 Female Mandible Papillary thyroid Ca NA NA NA NA
Nishikawa *et al*., 2010\[[@ref3]\] 83 Female Anterior mandible Poorly differentiated thyroid Ca 2 years None Lung DOD (1 year and 7 months)
Kumar *et al*., 2010\[[@ref3]\] 58 Female Left mandible Follicular thyroid Ca First manifestation Segmental mandibulectomy None NED (2 years)
Yokoe *et al*., 2010\[[@ref27]\] 71 Female Mandible Follicular thyroid Ca First manifestation Segmental mandibulectomy NA NED (48 months)
Chaturvedy *et al*., 2010\[[@ref28]\] 68 Male Upper lip Follicular thyroid Ca 4 years Surgical excision None NA
Radioactive Iodine therapy
Daley and Darling 2011\[[@ref3]\] 57 Male Anterior mandible gingiva NA NA NA NA NA
Kim *et al*., 2011\[[@ref18]\] 46 Female Bilateral mandible Follicular thyroid Ca First manifestation Hemimandibulectomy Lumbar spine, Femur neck NED (12 months)
Radioactive Iodine therapy
Muttagi *et al*., 2011\[[@ref15]\] 60 Male Mandible Papillary thyroid Ca (follicular variant) NA NA NA NA
63 Female Mandible Papillary thyroid Ca (follicular variant) NA NA NA NA
51 Female Mandible Papillary thyroid Ca (follicular variant) NA NA NA NA
44 Female Mandible Papillary thyroid Ca NA NA NA NA
35 Female Mandible Papillary thyroid Ca NA NA NA NA
Rohilla *et al*., 2011\[[@ref24]\] 55 Female Edentulous mandible Follicular thyroid Ca 2 years Radiotherapy Skull, ileum, Pelvis, vertebrae Alive (6 month)
Narain and Batra, 2011\[[@ref29]\] 62 Female Right maxilla Follicular thyroid Ca 15 years NA lymph nodes NA
Nikitakis *et al*., 2011\[[@ref3]\] 63 Male Right posterior Papillary thyroid Ca 2 years Radiotherapy Sternum, ribs, left tibia Alive with disease (2 years)
Maxilla, maxillary sinus Chemotherapy
Slim *et al*., 2012\[[@ref11]\] 67 Female Maxilla (malar region ) Papillary thyroid Ca (follicular variant) First manifestation Radioactive iodine therapy Facial region, lung, neck NA
Bhadage *et al*., 2012\[[@ref21]\] 40 Female Left mandible Follicular thyroid Ca First manifestation NA Submandibular lymph nodes NA
Shabestari *et al*., 2012\[[@ref17]\] 21 Female Anterior left maxilla Medullary thyroid Ca 7 years Chemotherapy NA NA
Pasupula *et al*., 2012\[[@ref30]\] 40 Female Left ramus of mandible Follicular thyroid Ca First manifestation Excision Rib NA
Kotina *et al*., 2013\[[@ref6]\] 55 Female Left mandible Follicular thyroid Ca 15 years NA NA NA
Kumar *et al*., 2013\[[@ref31]\] 31 Female Right maxilla Follicular thyroid Ca First manifestation Radioactive iodine therapy Right leg NED (7 years)
Palliative chemotherapy
Vishveshwaraiah *et al*., 2013\[[@ref26]\] 56 Female Right mandible Follicular thyroid Ca NA NA None NA
Pingel *et al*., 2013\[[@ref16]\] 76 Male Posterior mandible Papillary thyroid Ca (tall cell variant) 6 years Radiotherapy Rib Left clavicle DOD (11 months)
Vazifeh mostaan *et al*., 2013\[[@ref32]\] 58 Female Right mandible Follicular thyroid Ca 12 years Segmental mandibulectomy None NA
Das *et al*., 2014\[[@ref22]\] 55 Female Right mandible Papillary thyroid Ca (follicular variant) First manifestation Segmental mandibulectomy None NED (2 years)
Radioactive Iodine therapy
Lavanya *et al*., 2014\[[@ref17]\] 76 Male Left Mandible Follicular thyroid Ca First manifestation NA NA NA
Siddique *et al*., 2015\[[@ref8]\] 71 Male Labial gingiva of lower right lateral incisor Papillary thyroid Ca 8 weeks Radioactive iodine therapy Lymph nodes NA
Kori *et al*., 2015\[[@ref19]\] 50 Female Left mandible Follicular thyroid Ca First manifestation Segmental mandibulectomy None NED (1 years)
Radioactive Iodine therapy
40 Female Left mandible ramus, maxilla Follicular thyroid Ca First manifestation Radioactive iodine therapy None NA
Bingol *et al*., 2015\[[@ref2]\] 33 Female Right mandible Papillary thyroid Ca (follicular variant) First manifestation Hemimandibulectomy Pelvic, central nervous system DOD (5 years)
Hartinie *et al*., 2015\[[@ref33]\] 41 Female Right mandible Follicular thyroid Ca First manifestation Segmental mandibulectomy, Radiotherapy Anterior mediastinal nodes NED (6 months)
Fatahzadeh *et al*., 2015\[[@ref34]\] 43 Female Right Maxilla, Maxillary sinus Papillary thyroid Ca (follicular variant) NA Palliative radiotherapy Left leg, scalp NA
Right shoulder, Bilateral hips
Krishnamurthy *et al*., 2016\[[@ref23]\] 52 Male Left mandible Follicular thyroid Ca First manifestation Hemimandibulectomy None NED (14 months)
Radioactive iodine therapy
Cai *et al*., 2016\[[@ref25]\] 59 Female Ascending Ramos of mandible Thyroid adenocarcinoma NA Segmental mandibulectomy NA Alive (27 months)
Arepen and Mohamad *et al*., 2016\[[@ref1]\] 68 Female Bilateral mandible Papillary thyroid Ca (follicular variant) 4 years Palliative radiotherapy None NA
Dholam *et al*., 2017\[[@ref10]\] 58 Female Left mandible Papillary thyroid Ca (follicular variant) First manifestation Palliative radiotherapy Sacral, lumbar Alive (6 months)
Anajar *et al*., 2017\[[@ref4]\] 52 Female Mandible Papillary thyroid Ca First manifestation Segmental mandibulectomy None NED (6 months)
Loureiro *et al*., 2017\[[@ref12]\] 54 Female Left mandible Follicular thyroid Ca First manifestation None NA DOD
Varadarajan 2017\[[@ref9]\] 73 Female Left mandible Follicular thyroid Ca First manifestation Segmental mandibulectomy Tongue, soft tissue of neck NED (18 months)
Radioactive iodine therapy
Present case, 2017 68 Male Right mandible Papillary thyroid Ca 3 years Hemimandibulectomy None Alive (12 months)
NA: Not available, DOD: Died of disease, NED: No evidence of disease
Patients' age with thyroid cancer is usually from 25 to 65 years, and especially in younger patients, the tumor tends to grow very slowly.\[[@ref2]\] As regards, the mean age for well-differentiated thyroid cancer is 40 years for papillary type and 50 years for follicular type; so, it spends the time to metastasize to other regions.\[[@ref3]\] The metastatic tumors commonly occur in 5^th^ to 7^th^ decade of life.\[[@ref14][@ref15]\]
In our review, based on available information, among 73 cases with a wide age range (13--87), the mean age of patients was 56.6 years. The mean age was 64.1 years for males and 54 years for females. Most patients (45 cases) were in 6^th^ decade (24 cases) and 7^th^ decade (21 cases), only 6 cases were younger than 40 years.
In this review, among 74 patients, 57 cases (77%) were females and 17 cases (23%) were males. Female predilection can be attributed to the more incidence of thyroid cancer in women.\[[@ref16][@ref17]\]
Distant metastasis from PTC is very rare because it usually remains intraglandular or metastasize to regional cervical lymph nodes through the lymphatic system although follicular thyroid carcinoma with hematogenous spread causes more frequent distant metastasis.\[[@ref1][@ref18][@ref19]\] Metastasis to the jaw bone develops through blood vessels, that is the most common route.\[[@ref14]\]
The most frequent subtype of metastatic thyroid carcinoma is the follicular type (39 out of 75 with available data: 52%). Its more predilection can be attributed to the bloodstream dissemination.\[[@ref3][@ref11]\] PTC along with follicular variant (19 cases) and tall cell variant (1 case) accounted for 25 cases (33%) including present reported case. Although 11 cases were reported with medullary thyroid carcinoma, hurtle cell carcinoma, poorly differentiated thyroid carcinoma and adenocarcinoma.
Hirshberg *et al*.\[[@ref20]\] reported that the first sign of distant undiscovered malignancy in 23% of cases was oral metastasis. Oral metastasis in one-third of patients may be the first manifestation of its primary tumor.\[[@ref20]\]
Mandible, maxilla and oral soft tissue are more affected sites by metastatic cancer with predilection for mandible and gingiva.\[[@ref4][@ref7]\] Almost 41% of facial bones metastasis from thyroid cancer occurs in the mandible.\[[@ref9][@ref21]\] Maxillary metastasis is rare and less than one-fifth of all jaw metastatic tumors.\[[@ref3]\]
Oral soft-tissue involvement is less frequent than the jaw bones.\[[@ref2][@ref16]\] Ramus and angle of the mandible due to rich blood circulation in the medullary cavity are most locations to metastasis\[[@ref3][@ref4][@ref22]\] and in radiography has been shown with poorly defined osteolytic lesion with ragged border.\[[@ref1]\]
In review of cases with available data, oral metastasis was the first manifestation of thyroid cancer in 37 patients (66%). In 19 patients including the present case, there is a time to diagnosis metastasis after discover primary thyroid cancer. This time has widely ranged from 8 weeks to 32 years.
The most location of oral metastasis was the mandible (80%), including our case and only 10% was in the maxilla. Soft-tissue metastasis in gingiva, tongue, lower and upper lip were seen in 7 cases.
The common symptoms of metastatic tumor to jaws are pain, swelling, tooth mobility, premature loss of teeth, paresthesia, cervical lymphadenopathy and rarely pathologic fractures.\[[@ref16][@ref17][@ref23]\]
Because of rarity of oral metastasis, the first diagnosis can be squamous cell carcinoma, that is a most common malignant lesion in jaws with same clinical features.\[[@ref24]\] The presence of pain and chin paresthesia can be due to rapid progression of intraoral and extraoral expansion.\[[@ref9]\] A granulation-like mass in oral tissue and mucosa may appear and result in bleeding, infection, dysphagia and disturbance in mastication.\[[@ref9][@ref16]\]
In the present case, painless swelling was a noticeable clinical sign and extra examination also showed pathologic fracture.
Oral metastasis is similar to an inflammatory and reactive conditions such as periapical lesions, periodontitis, osteomyelitis and pericoronitis.\[[@ref2][@ref3]\]
The treatment modalities of oral metastasis from thyroid cancer have been varied from palliative to various combination management of surgical interventions, radioactive iodine ablation, radiotherapy, chemotherapy and hormone therapy.\[[@ref7][@ref16][@ref25]\]
For better survival, it has recommended total thyroidectomy (if not performed in the past) with surgical resection of metastatic turner, followed radioactive iodine treatment or radiotherapy.\[[@ref3][@ref4]\] In younger patients with small metastasis lesion, I 131 seems to be a more effective treatment.\[[@ref16]\]
The treatment decision depends on spreading metastatic lesion, symptoms and fracture risk.\[[@ref3][@ref4]\] In widespread metastasis disease, usually palliative treatment has been carried out; on the other hand, for the solitary and accessible lesion, surgical intervention has been recommended.\[[@ref3][@ref4]\]
Free fibular flap for mandible reconstruction is a gold standard.\[[@ref4]\] In the present case, thyroidectomy was performed 3 years earlier before, and then, hemimandibulectomy with radical neck dissection of the metastatic tumor was done.
The prognosis depends on the age at diagnosis of metastasis tumor and the number of involved bones\[[@ref4]\] although patient prognosis with distant metastasis of thyroid cancer is generally poor and the survival rate is 4 years on average 40% of patients after discovering metastatic lesion.\[[@ref3][@ref25]\] Metastatic thyroid carcinoma may be present adjacent to facial structures such as orbit, paranasal sinus and salivary gland.\[[@ref3][@ref9]\]
The available follow-up information in our review revealed in 40 out of 77 cases, length of follow-up varied from 17 days to 7 years. About 25 patients were reported as alive (14 of them being free of disease) and 15 patients died (13 of them due to disease).
The review shows in 32 out of 48 cases, oral metastasis was accompanied by metastasis in other sites such as femur, tibia, lung, ribs, lymph nodes, skull and pelvis, that they are usually skeletal.
CONCLUSION {#sec1-4}
==========
Oral metastasis from PTC is a very rare event that usually accompanies with a poor prognosis. Mandible metastasis revealed various symptoms and can be similar to the other conditions if it is the first manifestation. Therefore, thorough diagnostic workup for detection of the primary and metastatic sites is necessary. A proper evaluation can help to decide the best treatment. Surgical-based treatment with reconstruction can improve surgical outcome.
Declaration of patient consent {#sec2-1}
------------------------------
The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.[@ref34]
Financial support and sponsorship {#sec2-2}
---------------------------------
Nil.
Conflicts of interest {#sec2-3}
---------------------
There are no conflicts of interest.
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Cell cycle arrest is a key mechanism that is activated by cellular stress, and is strictly controlled by many regulatory mechanisms that either permit or restrain its progression ([@b1-or-38-01-0031]). The regulatory factors controlling cell cycle progression are cyclins, cyclin-dependent kinases (CDKs), CDK inhibitor (CKI) proteins such as p27 and p21, and tumor suppressor gene products p53 and Rb. Abnormalities of the gene structures (mutations, deletions, or amplifications) and the expression levels of their products are frequently observed in various cancers, which result in cell cycle progression and genetic instability ([@b2-or-38-01-0031]). Thus, correction of the collapse of cell cycle regulation by small-molecule CKIs in these tumors has offered insight that could potentially lead to effective therapeutic strategies. Flavopiridol, a pan-CDK inhibitor, improved response rates in chronic lymphocytic leukemia ([@b3-or-38-01-0031]). Palbociclib, a CDK4 and CDK6 inhibitor, plus fulvestrant prolonged progression-free survival in breast cancer ([@b4-or-38-01-0031]). However, no predictable biomarkers of small-molecule CKIs have been detected.
Recently, CKI proteins involved in G~1~/S arrest have also been shown to induce autophagy ([@b5-or-38-01-0031]--[@b7-or-38-01-0031]). Autophagy is a catabolic process that occurs via the lysosomal degradation pathway. It is essential for cell survival, differentiation, development, and homeostasis. It not only provides nutrients for maintaining vital cellular functions during starvation but also rids cells of superfluous or damaged organelles, misfolded proteins, and invading microorganisms ([@b8-or-38-01-0031]). Autophagy also plays an important role in both tumor progression and tumor suppression ([@b9-or-38-01-0031]). Moreover, autophagy is a survival pathway for cancer cells exposed to genotoxic stress and activated oncogenic signals ([@b10-or-38-01-0031]). Several lines of evidence have indicated that autophagy could be a novel target for cancer treatment. Inhibition of stress-induced autophagy induces apoptotic cell death; therefore, the combination of autophagy inducible stress and autophagy inhibition may provide a synthetic lethal interaction that could become a rational strategy for cancer treatment ([@b11-or-38-01-0031]).
In this study, we examined whether a specific CDK4 inhibitor and a CKI with a broad range of targets (flavopiridol) could induce autophagy in human solid cancer and fibroblast cell lines. Moreover, we examined whether inhibition of the autophagy induced by small-molecule CKIs caused apoptosis in these cell lines.
Materials and methods
=====================
### Cell lines and culture
The following cell lines were used in this study: BT474; MCF7; MDA-MB231; MDA-MB435S; SKBr3 (derived from breast cancer); A431 (derived from epidermoid cancer); NCI-N87 (derived from gastric cancer); SW480 (derived from colorectal cancer); and KMST-6 (derived from a fibroblast). BT474, MDA-MB231, MDA-MB435S, SKBr3, and NCI-N87 were purchased from American Type Culture Collection. A431 and KMST-6 were provided by the Institute of Development, Aging, and Cancer, Tohoku University. SW480 was kindly donated by Professor John M. Mariandason, Ludwig Institute for Cancer Research, Austin Hospital. All cell lines were maintained in RPMI-1640 medium (Sigma-Aldrich, R8758) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco, 26140-079) or without FBS (nutrient starvation conditions). The cells were incubated at 37°C in a 5% CO~2~ humidified atmosphere.
### Small-molecule CKIs and autophagy inhibition
CDK4 inhibitor, 2-bromo-12 and 13-dihydro-5 H-indolo-dione \[2 and 3-a\] pyrrolo \[3 and 4-c\] carbazole-5 and 7 (6H)-dione, was purchased from Merck Ltd. (219476-1MG), and flavopiridol was purchased from Santa Cruz Biotechnology, Inc. (sc-202157). Both were dissolved in dimethyl sulfoxide (DMSO) (Wako Pure Chemical Industries Ltd., 043-29355) and added to the cell cultures to make a final concentration of 100 nM. CQ was purchased from Tokyo Kasei Kogyo Co., Ltd. (C2301). To inhibit autophagy, CQ was added to the cell cultures to make a final concentration of 50 µM. ATG5, BECN1, and negative control siRNAs were purchased from Signal Silence, Cell Signaling Technology (\#6345, \#6222, and \#6568). Each siRNA was introduced to the cell cultures to make a final concentration of 100 nM using an X-tremeGENE siRNA transfection reagent (Roche Applied Science Ltd., 4476093).
### Western blot analysis
The lysate was subjected to western blot analysis as previously described ([@b12-or-38-01-0031]). Cells were harvested and resuspended in lysis buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA, and 1% protease inhibitor cocktail (Sigma-Aldrich, P2714). Anti-sqstm1 (p62) (Santa Cruz Biotechnology, Inc., sc-28359), anti-β-actin (Sigma-Aldrich, A5316), anti-atg5, anti-beclin-1, anti-LC3 (Medical & Biological Laboratories Co., Ltd., PM036), and anti-cleaved caspase-3 (Cell Signaling Technology, \#2630, \#3738, \#9661) antibodies were used.
### Immunofluorescence analysis
The strategy for LC3 immunofluorescence was described in a previous study ([@b13-or-38-01-0031]). Cells were harvested 18 h after treatment and fixed with 4% paraformaldehyde before incubation for 10 min at room temperature. After washing with Phosphate-buffered saline (PBS) they were permeabilized with 50 µg/ml digitonin (Sigma-Aldrich, D141) for 15 min at room temperature. After a second wash with PBS, the cells were incubated with anti-LC3 diluted 1:250 for 1 h. Subsequently, they were washed with PBS and then incubated with anti-rabbit IgG (H + L) FITC (Beckman Coulter, 732745) for 1 h and visualized using BZ-9000 (Keyence).
### Cell cycle analysis by fluorescence-activated cell sorting (FACS)
A total of 1.5×10^4^ cells per plate were seeded and incubated in 6-cm culture plates for 24 h. The cells were further incubated in the presence of the relevant compounds for 12--72 h. Subsequently, the cells were collected and FACS analysis was performed as previously described ([@b12-or-38-01-0031]).
### Cell proliferation analysis
A total of 5×10^3^ cells per well were seeded and incubated in a 96-well plate for 24 h. They were then treated with the relevant compounds and further cultured at 37°C for 12--72 h. The cell proliferation assays were performed with a Cell Counting Kit-8 (Dojin Laboratories, CK04), as previously described ([@b14-or-38-01-0031]).
### Statistical analysis
Cell cycle analyses were performed independently in triplicate. The data of cell cycle and cell proliferation analyses are presented as mean ± standard deviation (SD). In cell cycle analyses, differences between G~1~ and S phases were analyzed using Student\'s t-test. All statistical analyses were performed using Microsoft Excel^®^. Significant difference was set at p\<0.05.
Results
=======
### Cell cycle arrest and autophagy are induced by nutrient starvation
The aim of this study was to survey the potency of autophagy in eight human solid tumor cell lines, BT474, MDA-MB435S, SKBr3, MCF7, MDA-MB231 (derived from breast cancer), A431 (derived from epidermoid cancer), SW480 (derived from colorectal cancer), and NCI-N87 (derived from gastric cancer), and a human fibroblast, KMST-6. Initially, cell cycle analysis was performed under the condition of nutrient starvation. As shown in [Fig. 1A and B](#f1-or-38-01-0031){ref-type="fig"}, and [Table I](#tI-or-38-01-0031){ref-type="table"}, the G~1~ phase fraction was increased and the S phase fraction was decreased. However, there was no clear change in the sub G~1~ phase fraction. Thus, the nutrient starvation induced cell cycle arrest, but not apoptosis, in all of the cell lines.
Next, the induction of autophagy was monitored under the same conditions by measuring the expression level of p62 protein, which is degraded by autophagy. As shown in [Fig. 1C](#f1-or-38-01-0031){ref-type="fig"}, p62 protein was reduced under the nutrient starvation condition in all of the cell lines. Thus, we concluded that the autophagy induced under starvation was associated with cell cycle arrest.
### CDK4 inhibitor arrests the cell cycle and induces autophagy in some cell lines
To examine whether CDK4 inhibitor arrested cells at the G~1~/S phase in the cell cycle and also induced autophagy, cell cycle progression and p62 expression were analyzed. As shown in [Fig. 2A](#f2-or-38-01-0031){ref-type="fig"} and [Table II](#tII-or-38-01-0031){ref-type="table"}, CDK4 inhibitor arrested cells at the G~1~/S phase in all cell lines, simulating the nutrient starvation condition.
However, under the same condition, p62 degradation was observed only in MDA-MB435S, BT474, SKBr3, A431, and SW480 lines, but not in MCF7, MDA-MB231, NCI-N87, and KMST-6 lines ([Fig. 2B](#f2-or-38-01-0031){ref-type="fig"}). In addition to p62 expression, the formation of LC3 puncta was also observed in MDA-MB435S and BT474 lines, but not in MCF7 and MDA-MB231 ([Fig. 2C](#f2-or-38-01-0031){ref-type="fig"}). Furthermore, the conversion from LC3-I to LC3-II was also observed in MDA-MB435S, but not in MCF7 ([Fig. 2D](#f2-or-38-01-0031){ref-type="fig"}). These findings indicate that autophagy induced by CDK4 inhibitor does not always occur together with G~1~/S arrest, which contrasts with events during nutrient starvation.
To determine whether the autophagy is induced in a dose-dependent manner, p62 degradation assay was performed using MCF7 and KMST-6, in which autophagy was not induced by 100 nM CDK4 inhibitor. As a result, p62 degradation was also not observed in both cell lines, despite the presence of 10 µM CDK4 inhibitor, which is a sufficient concentration to induce G~1~ arrest ([Fig. 2E](#f2-or-38-01-0031){ref-type="fig"}).
We termed the autophagy induced by CDK4 inhibitor the AIC or autophagy induced by CDK4 inhibitor. The AIC status of each cell line was as follows: AIC-positive: BT474, MDA-MB435S, SKBr3, A431, and SW480; AIC-negative: MCF7, MDA-MB231, NCI-N87, and KMST-6.
### Induction of apoptosis by CDK4 inhibitor and CQ
p16, an endogenous CKI, has been shown to induce cytoprotective autophagy ([@b7-or-38-01-0031]). To examine whether autophagy induced by CDK4 inhibitor was also cytoprotective, we performed an analysis of cell proliferation with or without CDK4 inhibitor and/or CQ. CQ is known to be a non-specific autophagy inhibitor. As shown in [Fig. 2B](#f2-or-38-01-0031){ref-type="fig"}, the addition of CQ inhibited p62 degradation by CDK4 inhibitor.
Next, cell proliferation analysis was performed under the same conditions ([Fig. 3A](#f3-or-38-01-0031){ref-type="fig"}). In AIC (+) cell lines (A431 and SW480), the combination of CDK4 inhibitor and CQ inhibited cell proliferation and reduced the number of cells relative to the baseline. In contrast, in AIC (−) cell lines (MCF7 and KMST-6), the combination of these compounds inhibited cell proliferation, but did not reduce the number of cells relative to the baseline. Similar results were also observed in the experiments performed with the other cell lines, excluding NCI-N87 ([Table III](#tIII-or-38-01-0031){ref-type="table"}). To determine the cause of the reduction in the number of cells induced by the combination of CDK4 inhibitor and CQ, cell cycle analysis was performed ([Fig. 3B](#f3-or-38-01-0031){ref-type="fig"}). In AIC (+) cell lines (A431 and SW480), the addition of CQ to CDK4 inhibitor increased the sub-G~1~ fraction. In contrast, in AIC (−) cell lines (MCF7 and KMST-6), the addition of CQ did not change the sub-G~1~ fraction. Similar results were also observed in the experiments performed with the other cell lines ([Table IV](#tIV-or-38-01-0031){ref-type="table"}). These findings indicated that CDK4 inhibitor in combination with CQ induced apoptosis and promoted a synthetic lethal interaction in AIC (+) cell lines, suggesting that the autophagy induced by CDK4 inhibitor may be cytoprotective in these cell lines.
### Induction of apoptosis by CDK4 inhibitor and ATG5 or BECN1 knockdown
To determine whether the induction of apoptosis by the combination of CDK4 inhibitor and CQ was due to the inhibition of autophagy by CQ, we analyzed cell proliferation and cell cycle progression with *ATG5* and *BECN1* siRNA instead of CQ. Atg5 is an E3 ubiquitin ligase which forms a complex with Atg12 and Atg16L1, and this complex is necessary in autophagosome elongation. Beclin-1 interacts with either BCL-2 or the phosphatidylinositol 3 kinase, and shows to be involved in autophagy induction. Both the proteins play critical roles in the regulation of autophagy. As shown in [Fig. 4A](#f4-or-38-01-0031){ref-type="fig"}, siRNA of *ATG5* and *BECN1* suppressed the expression of Atg5 and Beclin-1 proteins, respectively, and also inhibited p62 degradation induced by CDK4 inhibitor in AIC (+) cell lines, which is the same as observed with CQ ([Fig. 4B](#f4-or-38-01-0031){ref-type="fig"}). In AIC (−) cell lines, p62 was not downregulated by CDK4 inhibitor, and *ATG5* and *BECN1* siRNA did not influence p62 protein levels. Next, we examined the expression levels of Atg5 and Beclin-1 by CDK4 inhibitor and CQ. As shown in [Fig. 4C](#f4-or-38-01-0031){ref-type="fig"}, CDK4 inhibitor increased the expression of Atg5 and Beclin-1 proteins in AIC (+) cell lines, but not in AIC (−) cell lines. These results indicated that CDK4 inhibitor induced autophagy via Atg5 and Beclin-1. As shown in [Fig. 4D](#f4-or-38-01-0031){ref-type="fig"}, CQ did not change the expression of Atg5 and Beclin-1 proteins in either AIC (+) or AIC (−) cell lines.
In terms of cell proliferation, a remarkable reduction in the number of cells was observed when CDK4 inhibitor was combined with the knockdown of *ATG5* or *BECN1* in AIC (+) cell lines, but not in AIC (−) cell lines ([Fig. 5A](#f5-or-38-01-0031){ref-type="fig"}, [Table V](#tV-or-38-01-0031){ref-type="table"}). In cell cycle analysis, the combination of CDK4 inhibitor and knockdown of *ATG5* or *BECN1* resulted in an increase in the sub-G~1~ fraction in AIC (+) cell lines, but not in AIC (−) cell lines ([Fig. 5B and C](#f5-or-38-01-0031){ref-type="fig"}, [Table VI](#tVI-or-38-01-0031){ref-type="table"}). These findings indicated that the cytotoxic effect of CQ in AIC (+) cell lines was due to the autophagy inhibition by CQ.
### Confirmation of apoptosis induced by the combination of CDK4 inhibitor and autophagy inhibition
To confirm that the reduction in the number of cells observed when CDK4 inhibitor was combined with autophagy inhibition was due to apoptosis, the production of cleaved caspase-3 was measured. Cleaved caspase-3 was produced when CDK4 inhibitor was combined with autophagy inhibition, by CQ as well as by *ATG5* or *BECN1* knockdown ([Fig. 6](#f6-or-38-01-0031){ref-type="fig"}).
### Induction of apoptosis by the combination of flavopiridol and autophagy inhibition
The aim of this experiment was to determine whether flavopiridol, a small molecule CKI with a broad range of targets, also induced apoptosis in AIC (+) cell lines when the cells were subjected to autophagy inhibition. As shown in [Fig. 7A](#f7-or-38-01-0031){ref-type="fig"}, the addition of flavopiridol to MDA-MB435S cells resulted in p62 degradation. The combination of flavopiridol and autophagy inhibition by either CQ or knockdown of *ATG5* or *BECN1* resulted in an increase in the sub-G~1~ fraction ([Fig. 7B and C](#f7-or-38-01-0031){ref-type="fig"}). Thus, flavopiridol, like CDK4 inhibitor, induced cytoprotective autophagy in the AIC (+) MDA-MB435S cell line.
Discussion
==========
The cell cycle progression of cells deprived of an essential nutrient or growth factor (starvation) or damaged by ionizing radiation or a DNA-damaging agent (genotoxic stress) is blocked at the G~1~/S transition, especially at a point in mid-G~1~ ([@b15-or-38-01-0031]). This point is called the restriction point (R point) and is the point at which the cell becomes 'committed' to the cell cycle because, after this point, extracellular proliferation stimulants are no longer required. In contrast, the cell cycle progression of G~1~ cells that have not reached the R point is immediately arrested. Both starvation and genotoxic stress induce autophagy to maintain homeostasis for survival ([@b8-or-38-01-0031]). Recent research has shown that the expression or activation of CKIs involved in G~1~/S arrest, such as p27, p21, and p16, also induces autophagy ([@b5-or-38-01-0031]--[@b7-or-38-01-0031]). Thus, we hypothesized that cell cycle arrest at the G~1~/S phase was required to induce autophagy and that this was achieved by small-molecule CKIs. In our study, starvation stress resulted in cell cycle arrest at the G~1~/S phase and autophagy in all of the cell lines that we examined. However, small-molecule CKIs only have the ability to induce autophagy in some cell lines \[AIC (+) cell lines\]. It has become clear that autophagy is not necessarily induced by small-molecule CKIs in all solid cancer cells. Furthermore, in KMST-6, a cell line derived from a human fibroblast, autophagy was not induced by CDK4 inhibitor. These results indicate that autophagy induction by CDK4 inhibitor is only a feature of some types of cancer cells. In AIC (−) cell lines, autophagy could be induced by starvation stress but not by CDK4 inhibitors at a dosage sufficiently high to induce cell cycle arrest at the G~1~/S phase. Therefore, we speculate that cell cycle arrest at the G~1~/S phase is necessary but not sufficient for the induction of autophagy.
Several studies have indicated that autophagy promotes cancer cell survival after chemotherapy, radiation therapy, or endocrine therapy ([@b16-or-38-01-0031]--[@b18-or-38-01-0031]). In our study, autophagy inhibition with small-molecule CKIs promoted cell death in AIC (+) cell lines, but not in AIC (−) cell lines, excluding NCI-N87. In NCI-N87, CQ without CDK4 inhibitor reduced the number of cells relative to the baseline. It seems that a concentration of CQ sufficient to induce autophagy is cytotoxic to NCI-N87. These results indicate that autophagy induced by small-molecule CKIs promoted cancer cell survival in the same manner as other anticancer therapies.
Since autophagy is a survival pathway for cancer cells, its inhibition has been considered to be a novel target for cancer treatment ([@b9-or-38-01-0031],[@b18-or-38-01-0031],[@b19-or-38-01-0031]--[@b21-or-38-01-0031]). Although no small-molecule autophagy inhibitors have been developed to date, a number of experiments in cells and animal models have investigated the effects of the anti-malarial agents CQ and hydroxychloroquine (HCQ), which block the degradation of autophagy products by inhibiting lysosomal function ([@b9-or-38-01-0031],[@b20-or-38-01-0031]). Currently, the inhibition of autophagy is thought to contribute to cell death by mechanisms such as stress augmentation ([@b22-or-38-01-0031]) and potentiation of DNA damage ([@b23-or-38-01-0031],[@b24-or-38-01-0031]). The use of an autophagy inhibitor with a genotoxic drug, which includes the majority of cytotoxic agents, might have potential as a novel combination anti-cancer therapy. In addition, the use of an autophagy inhibitor and an inhibitor targeting the PI3K/AKT pathway, such as mTOR, PI3K, and AKT inhibitors, has also been reported as a potentially useful combination in preclinical models ([@b10-or-38-01-0031],[@b22-or-38-01-0031]--[@b24-or-38-01-0031]), as well as clinical trials ([@b9-or-38-01-0031],[@b20-or-38-01-0031]). In this study, we demonstrated that small-molecule CKIs have the ability to induce autophagy and to induce apoptosis in five out of eight cancer cell lines, when combined with autophagy inhibition (CQ or knockdown of *ATG5* or *BECN1*). These findings suggest that the combination therapy with flavopiridol or palbociclib and an autophagy inhibitor such as CQ/HCQ could be a novel antitumor therapy because all have already been tested individually in humans.
If autophagy inhibition is to become a novel target for cancer treatment, then questions arise as to what kind of stress should be combined and which types of cancer are sensitive to autophagy inhibition. According to the Sanger COSMIC Database and our previous study ([@b25-or-38-01-0031]), we did not find an association between AIC status and *p53, Rb1*, and *BECN1* mutations ([Table VII](#tVII-or-38-01-0031){ref-type="table"}). HER2 overexpression and cyclin D1 amplification also seem to have no relationship with AIC status. However, previous studies have suggested that certain cancers are susceptible to autophagy inhibition ([@b10-or-38-01-0031],[@b26-or-38-01-0031]). For example, cancer cells with activation of the KRAS oncogenic signaling pathway exhibit high basal autophagy without any additional stress and are intolerant of other stresses.
The precise molecular mechanisms of autophagy induction by small-molecule CKIs also remain to be elucidated. Orvedahl *et al* reported on the molecules required for autophagy induced by viral infection by performing a comprehensive analysis using an shRNA library ([@b27-or-38-01-0031]). Such high-throughput screening would be useful for detecting the relevant molecules associated with small-molecule CKI-induced autophagy. Once a molecule required for the autophagy induced by small-molecule CKIs is identified, a biomarker predicting the effectiveness of the combination of a small-molecule CKI and an autophagy inhibitor can be developed. At present, the degradation of p62 can act as a surrogate marker to predict subtypes susceptible to synthetic lethal interaction.
In conclusion, the combination of a small-molecule CKI and autophagy inhibition led to a synthetic lethal interaction and induced apoptosis in human solid cancer cell lines susceptible to the autophagy induced by small-molecule CKIs.
We would like to thank Eri Yokota, Satoko Aoki, and Hiromi Nakano for their technical assistance. We would also like to thank Enago for the English language review. C.I. received lecture fees from Taiho Pharmaceutical and Tokyo Clinical Oncology Group and research funding from Taiho Pharmaceutical, Chugai Pharmaceutical, Ono Pharmaceutical, Daiichi Sankyo Pharmaceutical, Takeda Pharmaceutical, and Yakult Pharmaceutical.
CDK
: cyclin-dependent kinase
CKI
: CDK inhibitor
CQ
: chloroquine
FBS
: fetal bovine serum
DMSO
: dimethyl sulfoxide
PBS
: phosphate-buffered saline
FACS
: fluorescence-activated cell sorting
AIC
: autophagy induced by CDK4 inhibitor
HCQ
: hydroxychloroquine
{ref-type="table"}. Differences in G~1~ and S phases between normal and starvation conditions were analyzed using t-test. (B) The images of flow cytometry under conditions matching those for (A). (C) Western blot analysis of p62 under normal and starvation conditions as in (A). Lanes: 1, MDA-MB435S; 2, BT474; 3, SKBr3; 4, A431; 5, SW480; 6, MCF7; 7, MDA-MB231; 8, NCI-N87; 9, KMST-6.](OR-38-01-0031-g00){#f1-or-38-01-0031}
{ref-type="table"}. Differences in the G~1~ and S phases between normal and starvation conditions were analyzed using t-test. CDK4i indicates CDK4 inhibitor. (B) Western blot analysis of p62 during treatment with CDK4 inhibitor and/or CQ. Initially, a final concentration of 50 µM CQ was administered to the cells. After 48 h, a final concentration of 100 nM CDK4 inhibitor or DMSO was administered, and the cells were subsequently cultured for a further 24 h at 37°C before collection for western blotting. (C) Using MDA-MB435S, BT474, MCF7, and MDA-MB231, the formation of LC3 puncta was analyzed by immunofluorescence assay. LC3 immunostaining was performed for the cells treated with CDK4 inhibitor, as described for (A). (D) Using MDA-MB435S and MCF7, western blot analysis of LC3 during treatment with CDK4 inhibitor and/or CQ was performed as described for (A). (E) Using AIC (−) cell lines, MCF7 and KMST-6, western blot analysis of p62 was performed during treatment with CDK4 inhibitor. CDK4 inhibitor or DMSO was administered at a final concentration of 10 nM, 100 nM, 1 µM, or 10 µM, and the cells were subsequently cultured for a further 48 h at 37°C before collection for western blotting.](OR-38-01-0031-g01){#f2-or-38-01-0031}
{ref-type="fig"}, cell proliferation analysis was performed. Cell viability just before treatment with CQ was 100%, and the vertical axis corresponds to the absorbance ratio. The x-axis indicates the kind of treatment. The y-axis indicates the proportion of cell proliferation. Values shown are mean ± SD (n=3). (A) shows the data for the cell lines A431, SW480, MCF7, and KMST-6. Data for the other cell lines are shown in [Table III](#tIII-or-38-01-0031){ref-type="table"}. (B) After treatment with CDK4 inhibitor and/or CQ as described for [Fig. 2B](#f2-or-38-01-0031){ref-type="fig"}, cell cycle analysis was performed. The x-axis indicates the phase of the cell cycle. The y-axis indicates the proportion of the cell population. Values shown are mean ± SD (n=3). (B) shows the data for the cell lines A431, SW480, MCF7, and KMST-6. Data for the other cell lines are shown in [Table IV](#tIV-or-38-01-0031){ref-type="table"}. Differences in sub-G~1~ phase between DMSO + CQ and CDK4 inhibitor + CQ were analyzed using t-test. CDK4i, CDK4 inhibitor.](OR-38-01-0031-g02){#f3-or-38-01-0031}
{#f4-or-38-01-0031}
{ref-type="fig"}. Cell viability just before treatment with siRNA was 100%, and the vertical axis corresponds to the absorbance ratio. The x-axis indicates the type of treatment. The y-axis indicates the proportion of cell proliferation. Values shown are mean ± SD (n=3). (A) shows the data for the cell lines MDA-MB435S, BT474, MCF7, and KMST-6. Data for the other cell lines are shown in [Table V](#tV-or-38-01-0031){ref-type="table"}. 1, DMSO + siControl; 2, DMSO + si*ATG*5; 3, DMSO + si*BECN1*; 4, CDK4i + siControl; 5, CDK4i + si*ATG*5; 6, CDK4i + si*BECN1*. (B) Cell cycle analysis was performed under the conditions as described for (A). The x-axis indicates the phase of the cell cycle. The y-axis indicates the proportion of the cell population. Values shown are mean ± SD (n=3). (B) shows the data for the cell lines MDA-MB435S, BT474, MCF7, and KMST-6. Data for the other cell lines are shown in [Table VI](#tVI-or-38-01-0031){ref-type="table"}. Differences in sub-G~1~ phase between DMSO + siControl and CDK4 inhibitor + siControl, between DMSO + si*ATG5* and CDK4 inhibitor + si*ATG5*, and between DMSO + si*BECN1* and CDK4 inhibitor + si*BECN1* were analyzed using t-test. Lanes: 1, DMSO + siControl; 2, DMSO + si*ATG*5; 3, DMSO + si*BECN1*; 4, CDK4i + siControl; 5, CDK4i + si*ATG*5; 6, CDK4i + si*BECN1*. (C) Images of flow cytometry under the conditions as shown in (B). Lanes: 1, DMSO + siControl; 2, DMSO + si*ATG*5; 3, DMSO + si*BECN1*; 4, CDK4i + siControl; 5, CDK4i + si*ATG*5; 6, CDK4i + si*BECN1*. CDK4i, CDK4 inhibitor.](OR-38-01-0031-g04){#f5-or-38-01-0031}
{ref-type="fig"} and [5B](#f5-or-38-01-0031){ref-type="fig"}. CDK4i, CDK4 inhibitor.](OR-38-01-0031-g05){#f6-or-38-01-0031}
{ref-type="fig"}. (B) Cell cycle analysis under the conditions as described for (A). The x-axis indicates the phase of the cell cycle. The y-axis indicates the proportion of the cell population. Values shown are mean ± SD (n=3). D, DMSO; F, flavopiridol. (C) Images of flow cytometry under the conditions as in (B).](OR-38-01-0031-g06){#f7-or-38-01-0031}
######
Analysis of the cell cycle of the nutrient starvation.
Cell line FBS (%) sub-G~1~ (%) G~1~ (%) p-value S (%) p-value G~2~/M (%)
------------ --------- -------------- ---------- --------- ---------- --------- ------------
BT474 10 1.1±0.4 54.7±4.7 \<0.01 35.6±3.9 0.03 8.6±3.4
0 1.7±0.1 68.0±4.3 19.9±3.2 10.4±6.7
MDA-MB435S 10 0.9±0.5 57.5±4.4 0.01 31.8±1.9 0.02 9.8±5.8
0 1.3±0.5 71.5±6.1 17.0±4.6 10.2±4.9
SKBr3 10 1.9±0.3 52.0±5.9 \<0.01 31.9±2.4 0.01 14.2±4.0
0 3.0±0.8 68.8±2.7 13.5±2.4 14.7±4.7
A431 10 2.3±0.6 58.0±4.9 0.01 32.6±4.6 \<0.01 7.1±3.5
0 2.5±0.7 68.0±5.0 20.0±2.8 9.5±3.1
SW480 10 1.3±0.3 63.0±4.4 \<0.01 28.3±3.8 0.02 7.4±1.7
0 1.7±0.3 70.6±4.8 17.1±3.9 10.6±1.3
MCF7 10 2.2±0.6 54.7±1.4 0.04 36.9±2.1 0.01 6.2±2.9
0 2.0±1.5 64.8±6.6 23.3±3.9 9.9±4.2
MDA-MB231 10 2.4±0.7 66.8±2.4 0.01 24.2±5.8 0.01 6.6±4.1
0 4.0±1.5 78.8±3.7 11.0±2.7 6.2±0.5
NCI-N87 10 1.6±0.4 56.4±5.3 \<0.01 33.3±3.1 0.01 8.7±2.6
0 2.2±0.4 67.0±4.1 23.3±4.8 7.5±1.1
KMST-6 10 1.2±0.5 66.0±4.3 0.01 28.0±1.9 0.03 4.8±2.9
0 1.5±0.6 76.9±3.6 15.8±3.6 5.8±4.1
Data are expressed as mean ± SD. G~1~ and S fractions were examined paired t-test between FBS 10% and FBS 0%.
######
Analysis of the cell cycle of CDK4 inhibitor.
Cell line Condition sub-G~1~ (%) G~1~ (%) p-value S (%) p-value G~2~/M (%)
------------ ----------- -------------- ---------- --------- ---------- --------- ------------
BT474 DMSO 1.5±0.4 64.6±5.0 0.01 20.6±3.3 \<0.01 13.3±2.1
CDK4i 2.6±1.1 81.9±1.2 11.1±1.8 4.4±1.7
MDA-MB435S DMSO 1.1±0.3 49.9±7.6 0.03 40.2±5.5 0.03 8.8±2.4
CDK4i 2.1±0.9 69.9±8.7 19.9±4.6 8.1±5.0
SKBr3 DMSO 5.0±1.4 51.6±4.4 \<0.01 36.3±3.5 \<0.01 7.1±0.5
CDK4i 3.9±1.7 59.6±3.0 23.6±1.6 12.9±0.3
A431 DMSO 2.8±0.7 51.1±1.9 0.02 35.9±2.6 0.04 10.2±1.4
CDK4i 3.0±0.7 63.5±3.2 26.0±2.6 7.5±1.3
SW480 DMSO 2.6±0.6 60.4±3.6 0.01 29.1±1.9 \<0.01 7.9±2.3
CDK4i 2.9±0.9 70.7±3.4 18.0±1.7 8.5±2.6
MCF7 DMSO 0.9±0.2 61.8±2.5 0.01 31.8±2.4 \<0.01 5.5±0.3
CDK4i 0.8±0.2 74.7±2.5 16.9±2.3 7.6±0.4
MDA-MB231 DMSO 1.1±0.2 63.9±3.6 \<0.01 28.5±1.7 \<0.01 6.5±2.1
CDK4i 1.1±0.4 78.5±5.5 15.4±3.3 5.1±2.6
NCI-N87 DMSO 2.0±0.6 61.0±1.3 \<0.01 28.8±2.0 \<0.01 8.2±2.7
CDK4i 2.3±0.6 74.6±2.0 14.8±2.4 8.2±3.8
KMST-6 DMSO 0.9±0.2 60.4±3.4 0.01 29.4±2.5 0.01 9.3±1.1
CDK4i 0.9±0.3 78.4±2.8 13.9±3.3 6.8±0.8
Data are expressed as mean ± SD. G~1~ and S fractions were examined paired t-test between DMSO and CDK4 inhibitor (CDK4i).
######
Cell proliferation analysis for the combination of CDK4 inhibitor and CQ.
Cell line DMSO DMSO + CQ CDK4i CDK4i + CQ
-------------- -------- ----------- -------- ------------
AIC (+)
BT474 348±59 310±49 150±39 72±19
MDA-MB435S 286±86 187±55 137±49 71±27
SKBr3 210±50 150±77 111±39 49±13
A431 330±67 288±48 126±39 37±19
SW480 247±60 219±43 119±22 50±13
AIC (−)
MCF7 329±58 260±69 149±43 98±19
MDA-MB231 295±69 210±58 140±39 120±44
NCI-N87 198±76 80±44 113±33 51±29
KMST-6 238±40 221±54 110±31 120±54
Absorbance of 72 h/absorbance of 0 h × 100%. Data are expressed as mean ± SD. AIC, autophagy induced by CDK4 inhibitor; CQ, chloroquine; DMSO, dimethyl sulfoxide; CDK4i, CDK4 inhibitor.
######
Cell cycle analysis for the combination of CDK4 inhibitor and CQ.
Cell line Condition sub-G~1~ (%) p-value G~1~ (%) S (%) G~2~/M (%)
-------------- ------------ -------------- --------- ---------- ---------- ------------
AIC (+)
BT474 DMSO + CQ 2.6±0.8 \<0.01 68.4±3.7 20.2±4.4 8.8±1.5
CDK4i + CQ 28.9±2.9 55.8±5.1 8.9±0.7 6.4±2.9
MDA-MB435S DMSO + CQ 4.0±1.6 \<0.01 56.4±2.9 33.6±4.0 6.0±2.7
CDK4i + CQ 16.6±3.5 51.0±1.5 23.7±1.3 8.7±3.7
SKBr3 DMSO + CQ 7.6±1.7 0.02 49.3±5.1 35.3±4.6 7.8±2.2
CDK4i + CQ 23.7±3.9 40.7±5.0 24.4±1.1 11.2±2.1
A431 DMSO + CQ 4.1±3.2 0.03 48.8±5.2 36.6±2.9 10.5±0.9
CDK4i + CQ 19.9±4.4 46.6±5.3 24.9±4.2 8.6±3.3
SW480 DMSO + CQ 4.3±2.5 0.01 59.5±9.1 26.8±9.9 9.4±1.7
CDK4i + CQ 27.2±3.5 47.8±4.6 17.0±1.7 8.0±2.8
AIC (−)
MCF7 DMSO + CQ 3.8±1.4 0.15 52.2±3.7 30.9±2.8 13.1±4.2
CDK4i + CQ 3.2±1.2 66.2±4.7 18.5±2.6 12.1±3.3
MDA-MB231 DMSO + CQ 1.7±0.6 0.28 62.6±3.9 29.9±3.2 5.8±1.5
CDK4i + CQ 1.9±0.2 77.2±3.8 15.7±3.8 5.2±0.2
NCI-N87 DMSO + CQ 2.3±0.9 0.45 60.7±5.2 29.0±2.9 8.0±3.2
CDK4i + CQ 2.4±0.7 75.7±3.1 14.0±2.2 7.9±1.6
KMST-6 DMSO + CQ 1.6±0.3 0.09 62.6±5.5 27.6±3.5 8.2±2.3
CDK4i + CQ 3.0±1.2 74.7±2.4 13.6±3.5 8.7±4.7
Data are expressed as mean ± SD. sub-G~1~ fraction was examined paired t-test between DMSO + CQ and CDK4 inhibitor (CDK4i) + CQ. AIC, autophagy induced by CDK4 inhibitor.
######
Cell proliferation analysis for the combination of CDK4 inhibitor and *ATG5* or *BECN1* knockdown.
Cell line DMSO + siControl DMSO + si*ATG5* DMSO + si*BECN1* CDK4i + siControl CDK4i + si*ATG5* CDK4i + si*BECN1*
-------------- ------------------ ----------------- ------------------ ------------------- ------------------ -------------------
AIC (+)
BT474 247±52 194±45 155±74 116±15 67±2 64±9
MDA-MB435S 208±10 123±14 116±22 108±23 55±22 49±14
SKBr3 171±25 129±21 116±25 118±14 62±16 65±19
A431 248±39 225±31 205±16 128±7 67±4 76±39
SW480 264±55 222±42 219±13 123±15 51±9 60±16
AIC (−)
MCF7 224±5 196±46 191±5 133±26 141±26 125±13
KMST-6 210±28 189±15 194±29 130±13 137±15 122±12
Absorbance of 72 h/absorbance of 0 h ×100%. Data are expressed as mean ± SD. AIC, autophagy induced by CDK4 inhibitor; DMSO, dimethyl sulfoxide; CDK4i, CDK4 inhibitor.
######
Cell cycle analysis for the combination of CDK4 inhibitor and *ATG5* and *BECN1* knockdown.
Cell line Condition sub-G~1~ (%) p-value G~1~ (%) S (%) G~2~/M (%)
-------------- ------------------- -------------- --------- ---------- ---------- ------------
AIC (+)
BT474 DMSO + siControl 1.4±0.6 -- 64.9±3.5 23.3±5.6 10.4±2.7
DMSO + si*ATG5* 3.9±2.4 -- 62.9±4.6 21.9±5.8 11.3±1.2
DMSO + si*BECN1* 3.3±0.8 -- 65.0±6.0 21.8±5.4 9.9±0.2
CDK4i + siControl 2.9±1.2 0.09 82.9±2.9 10.1±2.4 4.1±0.7
CDK4i + si*ATG5* 22.2±3.5 \<0.01 59.0±3.6 10.9±1.5 7.9±1.4
CDK4i + si*BECN1* 25.1±2.4 \<0.01 62.1±3.8 11.2±1.8 1.6±0.4
MDA-MB435S DMSO + siControl 2.0±0.6 -- 55.6±2.7 35.9±2.1 6.5±4.2
DMSO + si*ATG5* 2.7±0.2 -- 56.6±5.4 34.0±2.0 6.7±3.6
DMSO + si*BECN1* 3.7±1.0 -- 56.9±1.5 34.5±2.4 4.9±2.9
CDK4i + siControl 2.7±0.8 0.13 64.9±8.4 23.0±4.0 9.4±5.2
CDK4i + si*ATG5* 17.7±2.7 \<0.01 51.5±3.6 22.5±0.6 8.3±5.7
CDK4i + si*BECN1* 14.2±5.1 0.03 56.7±7.6 22.1±3.3 7.0±0.8
SKBr3 DMSO + siControl 6.3±1.1 -- 52.8±3.5 31.3±2.8 9.6±5.2
DMSO + si*ATG5* 6.6±1.0 -- 51.8±8.3 34.3±4.5 7.3±4.8
DMSO + si*BECN1* 5.7±1.1 -- 53.6±3.9 32.0±2.6 8.7±2.4
CDK4i + siControl 3.9±2.0 0.09 58.9±3.1 23.7±2.0 13.5±0.9
CDK4i + si*ATG5* 21.1±4.4 \<0.01 40.7±1.8 24.7±4.7 13.4±1.8
CDK4i + si*BECN1* 19.0±6.6 0.04 42.2±3.4 23.9±4.8 14.9±1.6
A431 DMSO + siControl 3.5±0.9 -- 52.0±4.5 34.1±3.8 10.4±0.2
DMSO + si*ATG5* 4.3±0.5 -- 48.4±3.6 36.5±2.9 10.8±1.2
DMSO + si*BECN1* 3.1±1.9 -- 50.2±4.2 35.7±3.0 11.0±3.1
CDK4i + siControl 3.6±0.3 0.42 62.6±2.5 24.7±2.4 9.1±0.4
CDK4i + si*ATG5* 16.8±6.9 0.04 52.4±6.2 22.1±2.8 8.7±2.1
CDK4i + si*BECN1* 18.3±4.2 \<0.01 49.6±3.1 25.7±2.2 6.3±0.9
SW480 DMSO + siControl 2.7±0.4 -- 61.7±3.9 27.3±2.7 8.3±1.6
DMSO + si*ATG5* 3.7±2.5 -- 58.9±4.0 28.9±3.3 8.5±1.8
DMSO + si*BECN1* 2.8±0.9 -- 63.1±3.6 26.5±4.1 7.6±1.4
CDK4i + siControl 3.4±2.5 0.35 70.6±5.4 16.9±4.4 9.1±1.5
CDK4i + si*ATG5* 23.4±4.3 \<0.01 50.6±4.1 18.6±2.2 7.4±2.4
CDK4i + si*BECN1* 22.5±4.0 \<0.01 51.4±3.1 17.7±2.0 8.5±3.0
AIC (−)
MCF7 DMSO + siControl 2.6±0.7 -- 61.9±1.6 28.8±1.6 6.7±0.9
DMSO + si*ATG5* 4.6±0.6 -- 56.2±1.8 29.1±2.9 10.1±2.2
DMSO + si*BECN1* 4.6±1.3 -- 55.1±1.7 32.0±3.3 8.3±2.0
CDK4i + siControl 2.0±0.8 0.37 72.6±2.8 17.8±4.7 7.6±2.4
CDK4i + si*ATG5* 2.3±1.0 0.06 67.1±2.0 21.1±2.2 9.5±4.4
CDK4i + si*BECN1* 2.5±1.0 0.07 66.8±4.9 21.2±4.1 9.5±5.4
KMST-6 DMSO + siControl 1.0±0.4 -- 62.2±2.5 29.2±2.9 7.6±1.6
DMSO + si*ATG5* 1.1±0.5 -- 60.7±3.4 29.9±2.4 8.3±1.8
DMSO + si*BECN1* 1.1±0.4 -- 61.0±4.3 28.6±6.3 9.3±3.6
CDK4i + siControl 0.9±0.2 0.34 76.1±2.4 14.6±2.0 8.4±0.3
CDK4i + si*ATG5* 2.9±1.2 0.06 73.3±2.0 14.6±3.0 9.2±5.3
CDK4i + si*BECN1* 3.8±2.9 0.12 73.5±3.7 13.8±1.6 8.9±4.2
Data are expressed as mean ± SD. AIC, autophagy induced by CDK4 inhibitor; DMSO, dimethyl sulfoxide; CDK4i, CDK4 inhibitor. Sub-G~1~ fraction was examined paired t-test, between DMSO + siControl and CDK4i + siControl, between DMSO + si*ATG5* and CDK4i + si*ATG5*, and between DMSO + si*BECN1* and CDK4i + si*BECN1*.
######
The characteristics of cell lines used in this study.
Cell line Origin *p53* *Rb1* *BECN1* HER2 overexpression Cyclin D1 amplification
-------------- ------------ ------- -------- --------- --------------------- -------------------------
AIC (+)
BT474 Breast Mut LOH LOH (+) (+)
MDA-MB435S Breast Mut LOH LOH (+) (−)
SKBr3 Breast Mut Normal Normal (+) (−)
A431 Epidermis Mut LOH Normal ND (+)
SW480 Colon Mut Normal Normal ND (+)
AIC (−)
MCF7 Breast wt LOH LOH (−) (+)
MDA-MB231 Breast Mut LOH LOH (−) (+)
NCI-N87 Stomach Mut Normal Normal (+) (−)
KMST-6 Fibroblast wt Normal Normal ND (−)
AIC, autophagy induced by CDK4 inhibitor; Mut, mutant type; wt, wild-type; LOH, loss of heterozygosity; ND, not determined.
| {
"pile_set_name": "PubMed Central"
} |
To the editor {#S1}
=============
The proceedings of a recent meeting on alopecia areata (AA) ([@R4]) summarized work using the surgically induced C3H/HeJ mouse model for AA ([@R13]), in which investigators found enlarged hearts in affected mice, suggesting an association between AA and cardiac findings. However, the heart lesions described are a well known strain specific disease, not limited to C3H substrains. These lesions have been described by a number of names including epicardial mineralization with fibrosis and dystrophic cardiac calcinosis ([@R6]; [@R7]). Crosses between C3H/HeJ and C57BL/6J mice have identified 4 quantitative trait loci (QTLs), designated as Dystrophic Cardiac Calcinosis 1-4 (*Dyscalc1-4*) ([@R9]). Mapping to mouse Chr. 7 ([@R10]), *Dyscalc1* was subsequently identified as being due to non-synonymous single nucleotide polymorphisms in the ATP-binding cassette, sub-family C (CFTR/MRP), member 6 (*Abcc6*) gene ([@R1]; [@R14]). Mutations in the human *ABCC6* gene and targeted mutations in the mouse *Abcc6* gene produce pseudoxanthoma elasticum (PXE) ([@R8]; [@R11]), a systemic metabolic disease with cutaneous features distinct from AA ([@R16]).
In a massive histopathological screening of all organ systems in 31 inbred strains of mice of both genders, dystrophic cardiac calcinosis was diagnosed in 8 strains ([@R3]; [@R15]). C3H/HeJ and A/J strains were found to develop both heart lesions ([@R5]) and AA ([@R12]) in the aging study, although in both cases more mice with normal skin had heart lesions than those with AA ([Table 1a](#T1){ref-type="table"}). Three strains were found to develop histologically confirmed AA (MRL/MpJ, SJL/J, and SWR/J) but none of these mice had any type of heart lesion. No correlation was found in a retired breeder study ([Table 1b](#T1){ref-type="table"}) ([@R3]) or in a large mouse cross (C3H/HeJ x C57BL/6J, C3B6F2; [Table 1c](#T1){ref-type="table"}) generating F2 females for identifying AA eQTLs. Heart lesions varied in severity and location between the strains ([@R3]). Genome-wide association mapping determined that none of the QTLs for dystrophic cardiac calcinosis corresponded to genomic regions identified to determine AA.
While it is easy to see clinical correlations between seemingly unrelated diseases in small numbers of mice undergoing experimental manipulation, it is critically important to understand strain specific background lesions. The mineralization and fibrosis phenomena among the inbred strains associated with PXE-like diseases are very complicated. Some are related to each other while others are not. The underlying genetic predisposition can be modified by the genes involved in other diseases. Such appears to be the case for *Abcc6* and PXE ([@R2]). As the complex genetics of AA in humans and mice continues to be refined, it is possible that some of the genes involved in development of heart lesions may overlap with those that determine AA, but with technologies currently available using large populations of mice it appears that cardiac mineralization and fibrosis phenotypes are not correlated with AA.
**Grants:** Research reported in this publication was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health under Award Number AR056635. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Dr. Berndt is the recipient of a fellowship by the Parker B. Francis Foundation and recipient of a North American Hair Research Society Mentorship Grant. The Jackson Laboratory Shared Scientific Services were supported in part by a Basic Cancer Center Core Grant from the National Cancer Institute (CA34196).
**Conflict of interest:** none to declare
AA
: alopecia areata
*Abcc6* (mouse gene)
: *ABCC6* (human gene), ATP-binding cassette subfamily C, member 6, gene
*Dyscalc1-4*
: dystrophic cardiac calcinosis 1-4, quantitative trait loci
PXE
: pseudoxanthoma elasticum
QTL
: quantitative trait loci
###### Lack of correlation between dystrophic cardiac calcinosis in aging mouse strains and adult onset alopecia areata
There was no correlation between histologically confirmed alopecia areata and dystrophic cardiac calcinosis in 31 inbred strains in an aging histopathology study (a), evaluation of hearts in retired breeders (b), or F2 hybrid study for mapping quantitative trait loci for alopecia areata (c).
Table 1a. There was no correlation between alopecia areata and heart lesions in mouse strains in the 31 strain aging study.
----------------------------------------------------------------------------------------------------------------------------- --------- --------- -------- ------- ------- ------- --------- --------- -------- --------
A/J 51 46 0 1 0 0 51 45 23 8
C3H/HeJ 28 29 7 3 0 0 21 26 1 3
MRL/MpJ 41 31 2 0 0 0 39 31 0 0
SJL/J 36 10 2 0 0 0 34 10 0 0
SWR/J 24 18 6 0 0 0 18 18 0 0
**Total:** **180** **134** **17** **4** **0** **0** **163** **130** **24** **11**
Table 1b. Alopecia areata was not diagnosed in any of the strains in the retired breeder survey.
-------------------------------------------------------------------------------------------------- -------- -------- ------- ------- -------- -------- ------------ -------- -------- ------- ------- -------- --------
A/J 10 10 0 0 9 10 DBA/2J 10 10 0 0 10 10
BALB/cJ 10 10 0 0 9 10 FVB/NJ 10 10 0 0 0 0
BALB/cByJ 10 10 0 0 8 10 KK/HlJ 10 10 0 0 10 9
C3H/HeJ 10 10 0 0 10 6 LP/J 10 10 0 0 0 0
C57BL/6J 10 10 0 0 0 1 PWD/PhJ 10 10 0 0 0 0
C57BL/10J 10 10 0 0 7 10 SWR/J 10 10 0 0 0 0
**Total:** **60** **60** **0** **0** **43** **47** **Total:** **60** **60** **0** **0** **20** **19**
Table 1c. There was no correlation between alopecia areata and heart disease in an F2 hybrid cross used to investigate the genetics of these diseases (p-value = 0.651 using a Fisher exact test).
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------- -------- ----- --- ----- ---
C3B6F2 195--605 Female 191 1 145 4
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
Traditional genetic approaches have played a pivotal role in establishing the requirement of individual gene activities and cell behaviors in complex morphogenetic processes^[@CR1]--[@CR9]^. More recent advances in synthetic biology are opening the possibility to engineer gene circuits^[@CR10]^, signaling systems^[@CR11],[@CR12]^, and biomaterials^[@CR13],[@CR14]^ not only to probe morphogenesis, but also to re-construct it and direct it^[@CR15],[@CR16]^. These approaches, which are converging into the nascent field of synthetic morphogenesis^[@CR17]^, will be instrumental to define the minimum set of requirements sufficient to drive morphogenesis, and will therefore, also facilitate the building of artificial tissues for potential applications in regenerative medicine. Here, we used optogenetics to synthetically reconstitute morphogenesis in the early *Drosophila* embryo. We focused on epithelial folding, a conserved morphogenetic process driving internalization of tissues during animal development^[@CR18]^. A large body of experimental evidence indicates that apical constriction driven by phosphorylation and activation of the molecular motor myosin II is required for tissue invagination^[@CR6]^. However, the extent to which apical constriction is on its own sufficient to drive tissue internalization is unknown. During this process, cells undergo a series of complex changes in shape and intracellular organization, whose causal relationship to apical constriction and inward folding remain poorly understood^[@CR19]--[@CR22]^. Furthermore, the organismal scale cells occupy defined positions and are organized in specific geometrical patterns, which might facilitate or constrain invagination. Finally, apical constriction is not always coupled with invagination and several invagination processes are independent of apical constrictions^[@CR23]^. For example, during salivary gland invagination, apical constriction and tissue invagination are uncoupled. When apical constriction is inhibited, compressing forces exerted by a supracellular myosin cable surrounding the salivary gland pit are sufficient to push cells inward^[@CR24]^. Similar actomyosin-cable-mediated forces drive neural tube closure during chick embryogenesis^[@CR25]^. Other examples of invaginations independent of apical constriction include the folding of *Drosophila* leg epithelium, which is driven by whole-cell shrinkage coupled with apoptosis^[@CR22]^, and ascidian gastrulation, which is driven by a basolateral accumulation of myosin II and apicobasal cell shortening^[@CR26]^. In addition, basal wedging rather than apical constriction seems to be the major force driving tissue internalization during mouse neuronal tube development^[@CR27]^. Even in the case of *Drosophila* ventral furrow invagination, arguably the best characterized example of epithelial folding, the extent to which apical constriction drive invagination is unknown. Computer simulations suggest the requirement of additional pushing forces exerted by lateral ectodermal cells^[@CR28],[@CR29]^, and relaxation of the basal surface of invaginating cells^[@CR30]^. At the tissue-scale, the emergence of collective contractile behavior and its relationship to tissue geometry and invagination also remains the focus of active investigations^[@CR31]--[@CR33]^. In this study, we employ an optogenetic method to synthetically reconstruct epithelial folding during early *Drosophila* embryogenesis. In this context, "synthetic" refers to guided spatio-temporal control over the signaling pathway driving apical constriction, which is otherwise dependent upon the differentiation program of the embryo. Using this approach, we test the extent to which apical constriction on its own can drive invagination, and how different contractile behaviors arise in response to different temporal and spatial patterns of optogenetic activation. Collectively, our results indicate that apical constriction is sufficient to drive tissue invagination, but it is not sufficient to fold an invagination into a tube-like shape. Furthermore, our results provide insights into the emergence of pulsatile contractions and impact of tissue geometry on coordinated contractile behavior.
Results {#Sec2}
=======
RhoGEF2 plasma membrane recruitment and tissue responses {#Sec3}
--------------------------------------------------------
To study the impact of apical constriction on tissue folding, independently of any pre-defined conditions, associated with normal invagination processes, we used an optogenetic system to activate Rho signaling^[@CR34],[@CR35]^ at the apical surface of developing *Drosophila* embryos prior to any sign of morphological differentiation (Fig. [1a, b](#Fig1){ref-type="fig"}). At the end of cellularization, the *Drosophila* embryo is composed of a monolayer of epithelial cells without any morphological difference along the antero-posterior (AP) or dorso-ventral (DV) axes. To avoid any contribution from endogenous invagination processes, experiments were performed on the dorsal side of the embryo. In this region, the twist and snail transcription factors, that control ventral furrow invagination^[@CR36]^, are not expressed and there is no morphological movement at this stage^[@CR37]^.Fig. 1Light-mediated recruitment of RhoGEF2 at the plasma membrane causes tissue-level responses. **a** Schematic illustration of the CRY2/CIBN optogenetic system used for controlling plasma membrane recruitment of the Rho1 activator RhoGEF2 during early *Drosophila* embryogenesis. The DHPH catalytic domain of the GTP Exchange factor RhoGEF2 was fused to the light-sensitive PHR domain of CRY2 and either tagged with mCherry or left without a fluorescent tag. In the absence of light, RhoGEF2-CRY2::mCherry localizes to the cytoplasm (left). Upon two-photon illumination, CRY2 undergoes a specific conformational change and binds to the plasma membrane anchor CIBN::pmGFP (right). At the plasma membrane, RhoGEF2-CRY2::mCherry triggers activation of Rho1 and its downstream effectors. **b** Cartoon illustrating the experimental setup used in this study. Stage 5 *Drosophila melanogaster* embryos expressing the optogenetic module were mounted with their dorsal epithelium facing the coverslip, and a region of interest (ROI) of the desired geometrical pattern (left) was defined to delimit the area of two-photon illumination. The plane of activation was restricted to the junctional plane at 4 µm from the apical side of the epithelium (right top); over time photo-activated cells disappeared from the plane of acquisition (right bottom). **c**--**k** Confocal still frames from time-lapse recordings of three representative embryos (*N* = 9) co-expressing CIBN::pmGFP and RhoGEF2-CRY2::mCherry. Images are presented as integrated intensity projections of 3 µm. **c** In the absence of blue-light (\~20 s before photo-activation), RhoGEF2-CRY2::mCherry localizes to the cytoplasm. Inset shows the plasma membrane localization of CIBN::pmGFP at the time of photo-activation using two-photon illumination. **d** RhoGEF2-CRY2::mCherry recruitment to the plasma membrane was restricted to the imposed circular area after eight consecutive rounds of photo-activation for a total time of 2 s. In depth, the area of activation was restricted to three consecutive z-stacks centered at 4 µm from the apices of the cells. Inset represents a magnification of the activated area. **e** Sustained photo-activation was alternated with mCherry excitation and recorded in intervals of \~21 s. After \~5 min, a hollow matching the geometrical pattern of photo-activation formed on the dorsal epithelium as photo-activated cells disappeared from the plane of acquisition. Similar experiments were carried out with a triangle (**f**--**h**) and a square (**i**--**k**). Scale bars are 10 µm
Apical constriction is initiated by activation of the small GTPase Rho1 by exchange factors of the RhoGEF family, which results in myosin activation and contraction of cortical actin filaments^[@CR38],[@CR39]^. We employed the CRY2/CIB1 protein heterodimerization system^[@CR32],[@CR40]^ to control the plasma membrane localization of *Drosophila* RhoGEF2 using light. We engineered embryos to co-express RhoGEF2-CRY2 (the catalytic domain of RhoGEF2 fused to CRY2, a light-sensitive protein domain) and CIBN::pmGFP (consisting of the N-terminal domain of the CRY2 binding partner CIB1 tagged with a plasma membrane anchor and GFP), see cartoon in Fig. [1a](#Fig1){ref-type="fig"}. In the dark, embryos developed normally, RhoGEF2-CRY2 localized to the cytoplasm, and CIBN::pmGFP was correctly anchored at the plasma membrane (Fig. [1c,f, i](#Fig1){ref-type="fig"}). Photo-activation using a two-photon based protocol (*λ* = 950 nm)^[@CR41]^ caused recruitment of RhoGEF2-CRY2 to the plasma membrane within 30 s (Fig. [1d, g, i](#Fig1){ref-type="fig"}), and the moving away of cells (within \~ 5 min) from the imaging plane in an area that precisely matched the geometry (circular, triangular, or squared) of the illuminated area (Fig. [1e, h, k](#Fig1){ref-type="fig"} and Supplementary Movie [1](#MOESM4){ref-type="media"}).
RhoGEF2 optogenetic activation causes apical constriction {#Sec4}
---------------------------------------------------------
In order to test whether optogenetic activation caused apical constriction and at the same time visualize both activated and non-activated cells, we imaged embryos co-expressing RhoGEF2-CRY2 and CIBN::pmGFP, together with the plasma membrane marker Gap43 tagged with mCherry (Gap43::mCherry). We photo-activated a rectangular area on the dorsal surface of the embryo (blue box in Fig. [2a](#Fig2){ref-type="fig"}) and followed the cell surface area over time using quantitative imaging (Fig. [2f--i](#Fig2){ref-type="fig"} and Supplementary Movie [2](#MOESM5){ref-type="media"}). Cells contained within the box of activation constricted their apical area up to \~60% with a 20% increase in eccentricity (Fig. [2c--e, g](#Fig2){ref-type="fig"}), while they ingressed towards the interior of the embryo within \~6 min (Fig. [2b, c](#Fig2){ref-type="fig"}). Consistently, myosin II accumulated specifically at the apical surface of activated cells (Fig. [2j--r](#Fig2){ref-type="fig"} and Supplementary Movie [3](#MOESM6){ref-type="media"}). Non-activated cells, located at the boundary between the activated and non-activated regions, were pulled towards the activated region and increased their apical area by 50% (Fig. [2b--g](#Fig2){ref-type="fig"}). Non-activated cells outside of the boundary region did not change their apical surface significantly over time (Fig. [2b--f](#Fig2){ref-type="fig"}). TUNEL assay revealed that photo-activation did not induce apoptosis (Supplementary Fig. [1a, c, e](#MOESM1){ref-type="media"}), and in agreement with previous reports, apoptosis could only be detected at later stages of embryogenesis^[@CR42]^ (Supplementary Fig. [1b, d, f](#MOESM1){ref-type="media"}). Thus, recruitment of RhoGEF2 to the plasma membrane is on its own sufficient to cause apical constriction and tissue-level responses that are compatible with endogenous invagination processes.Fig. 2RhoGEF2 optogenetic activation triggers apical constriction. **a**--**d** Confocal still frames from a time-lapse recording of a representative embryo (*N* = 3) co-expressing CIBN::pmGFP, RhoGEF2-CRY2::mCherry, and the plasma membrane marker Gap43::mCherry. **a** The dorsal epithelium prior to two-photon illumination as visualized by imaging Gap43::mCherry, note that with the imaging setting used, the cytoplasmic signal of RhoGEF2-CRY2::mCherry was not visible. Blue rectangle indicates area of photo-activation. Cells within the magenta rectangle were defined as boundary cells. Cells contained within the gray and magenta rectangle corresponded to non-activated cells. **b** and **c** show final still frames after 8 min of continuous two-photon illumination. **b** shows the apical-most plane. **c** shows a focal plane 10 µm basally to the focal plane in **b**. Individual cells marked with a dot (color-coded as indicated above) were tracked over-time and their apical area was quantified, as presented in **f**. Scale bars, 10 µm. **d**, **e** Detailed analysis of individual cells from the embryo shown in **a**. **d** Cell outlines of activated (top row), non-activated (middle row), and boundary (bottom row) cells after photo-activation; dashed line corresponds to the top left corner of the area of photo-activation. Scale bars are 5 µm. **e** Quantification of apical area (left) and eccentricity (right) over time for the individual cells shown in **d**. **f** Quantification of mean apical area over time upon two-photon illumination of activated (*n* = 84), non-activated (*n* = 82), and boundary cells (*n* = 28). Shaded areas indicate s.d. **g** Mean apical area eccentricity quantified for the same groups of cells as in **f**. Shaded areas indicate s.d. **h** Quantification of mean apical constriction rate over time for the photo-activated cells in **f** (*n* = 84). Shaded areas indicate s.d. **i** Quantification of the constriction rate of individual cells within the photo-activated region (*n* = 34). **j**--**r** Confocal still frames from a time-lapse recording of a representative embryo (*N* = 3) co-expressing CIBN::pmGFP, RhoGEF2-CRY2, and the myosin II regulatory light chain reporter Sqh::mCherry. Images are presented as integrated intensity projections of 5 µm. The dorsal epithelium was subjected to continuous two-photon illumination within a rectangular area (dashed line). **j**--**l** CIBN::pmGFP, (**m**--**o**) Sqh::mCh, and (**p**--**r**) composite of both (merge)
RhoGEF2 optogenetic activation and pulsatile constrictions {#Sec5}
----------------------------------------------------------
Apical constriction is often characterized by pulsatile cycles of contraction and expansion of the surface area, as seen during dorsal closure^[@CR43]^, or of contraction and stabilization, as seen during ventral furrow formation^[@CR44]^. The apical constrictions induced by optogenetic stimulation of RhoGEF2 displayed a monotonic increase and decrease in the constriction rate over a period of \~8 min (Fig. [2h](#Fig2){ref-type="fig"}) without showing any obvious pulsatile behavior in a cell-by-cell analysis of the constriction rate (Fig. [2i](#Fig2){ref-type="fig"}), thus arguing that the average contractile behavior did not obscure any underlying individual pulsatile behavior. Our initial photo-activation protocol was based on a continuous and synchronous administration of light, which might override any potential pulsatile behavior induced by endogenous biochemical or mechanical processes downstream of Rho signaling. Therefore, we monitored apical constriction under different illumination conditions. First, we tested whether changing the frequency of illumination without modifying the laser power (10 mW) would be sufficient to elicit pulsatile constrictions. The results from these experiments demonstrated that, at intervals of \~55 s of illumination, cells displayed a pulsatile behavior (Fig. [3a](#Fig3){ref-type="fig"} and Supplementary Movie [4](#MOESM7){ref-type="media"}), whereas at intervals below and above this value, cells either did not constrict or constricted continuously and invaginated. At 55 s intervals, activated cells displayed cycles of apical constriction and expansion with a mean period of 157 ± 63 s (Fig. [3d](#Fig3){ref-type="fig"}), that lasted for about 10 min before endogenous gastrulation movements caused cells to move away from the photo-activation plane, making them untrackable for a longer period of time. This contractile behavior was not sufficient to cause tissue internalization but rather resembled the pulsatile behavior of amnioserosa cells during dorsal closure or that of ventral cells in *twist* or *fog* mutant embryos, which lack the stabilization phase and fail to invaginate^[@CR44]^. However, differently from these two conditions, constrictions induced by pulsed illumination were synchronized. Cells constricted and relaxed in phase as demonstrated by the preservation of oscillatory behavior even after averaging out time-course measurements of multiple cells (Fig. [3a--c](#Fig3){ref-type="fig"}). Considering the reversion kinetics of the CRY2/CIB1 system (*T*~1/2~\~5 min)^[@CR32],[@CR40]^, it is unlikely that an administration of light at 55 s intervals was directly inducing contraction and relaxation at each light on/off cycle. Consistently, when the time-point of illumination was plotted over the time-course analysis of the surface area, no direct correlation between contraction and illumination could be observed (Fig. [3e](#Fig3){ref-type="fig"}). This suggests that this discontinuous illumination regime triggers an endogenous mechanical and/or chemical oscillatory system whose dynamics were probably altered by the continuous illumination protocol previously employed. Therefore, we tested whether either a continuous administration of light at a lower laser power or a single pulse of light at a higher laser power would also elicit pulsatile behavior. Continuous illumination at 5 mW induced cycles of contraction and relaxation with a mean period of 148 ± 41 s (Fig. [3b, d](#Fig3){ref-type="fig"} and Supplementary Movie [4](#MOESM7){ref-type="media"}). Similarly, a single pulse of light at a higher laser power (15 mW) also induced pulsatile behavior, albeit with a slightly slower mean period (176 ± 58 s) (Fig. [3c, d](#Fig3){ref-type="fig"} and Supplementary Movie [4](#MOESM7){ref-type="media"}). Taken together, these data are compatible with the presence of an endogenous mechano-chemical oscillatory system^[@CR45]^ that can be induced by Rho signaling activation up to a certain threshold, above which cells contract continuously without pulsing.Fig. 3Induction of pulsatile apical contractions. **a**--**c** Constriction rate of photo-activated cells measured by change in apical area over time. Positive values indicate contractions while negative values indicate expansions. Embryos co-expressing CIBN::pmGFP, RhoGEF2-CRY2::mCherry, and the plasma membrane marker Gap43::mCherry (*N* = 9) were subjected to different illumination protocols. In all cases, cell outlines were recorded in mCherry (Gap43::mCherry) during and after photo-activation. Data presented as mean (solid line) and s.d. (shaded area). A rectangular region on the dorsal epithelium was photo-activated using two-photon illumination in a single plane, centered at 4 µm from the apical-most plane. **a** Discontinuous photo-activation (10 mW) with a resting interval of 55 s (*N* = 3 embryos, *n* = 44, 13, 34 cells). **b** Continuous photo-activation (5 mW, *N* = 3 embryos, *n* = 26, 42, 42 cells). **c** Single initial photo-activation (15 mW, *N* = 3 embryos, *n* = 32, 61, 74, cells). **d** Box-and-whisker plot of the quantified period between peaks of constriction for the indicated protocols. The quantified constriction pulses correspond to the data in **a**--**c**, (*n* = 455, 295, 488 pulses, \*\*\**p* \< 0.001, two-tailed paired *t*-test). **e** Constriction rate over time for four individual cells during the discontinuous activation protocol (blue markers indicate time of illumination)
Next, we established a photo-activation protocol that allowed us to both induce pulsatile contractions and, at the same time, visualize cell shape and myosin II dynamics. A single pulse of two-photon optogenetic activation (15 mW) was followed by continuous illumination at a lower laser power (2 mW), which allowed the imaging of the apical surface (CIBN::pmGFP) at high temporal resolution without causing inhibition of pulsations. Myosin II labeled with mCherry (Sqh::mCherry) was visualized by alternating two-photon illumination with single photon illumination (561 nm) at 13 s intervals (Fig. [4a--f](#Fig4){ref-type="fig"} and Supplementary Movie [5](#MOESM8){ref-type="media"}). Under this condition, cells pulsed with a mean period of 164 ± 49 s (Fig. [4c](#Fig4){ref-type="fig"}) with myosin II pulsing with similar kinetics (Fig. [4b](#Fig4){ref-type="fig"}, mean period 159 ± 52 s). Myosin II accumulated at the medio-apical plane as cells contracted and then gradually disappeared as cells relaxed (Fig. [4d--f](#Fig4){ref-type="fig"}). Quantitative analysis revealed a temporal correlation between myosin II pulses and apical constriction with \~40% of constrictions occurring within 25 s of a myosin pulse (Fig. [4h](#Fig4){ref-type="fig"}), and with the average maximum correlation coefficient occurring when the myosin II accumulation rate was not shifted compared to the apical constriction rate (Fig. [4g](#Fig4){ref-type="fig"}). These results argue that optogenetic-induced pulsatile behavior arises as a consequence of synchronous myosin medio-apical accumulation and dissolution.Fig. 4Medio-apical myosin II accumulation and pulsatile changes in cell shape. **a** Merged confocal still frames from a time-lapse recording of a representative embryo (*N* = 3) co-expressing CIBN::pmGFP (cell outline, in white), RhoGEF2::CRY2 (not visible), and the myosin II regulatory light chain reporter Sqh::mCherry in red (yellow arrowheads). Images are presented as maximum intensity projections (5 µm) from the apical-most plane. Cells were subjected to an initial two-photon illumination pulse (15 mW) followed by a continuous two-photon excitation at a lower laser power (2 mW) for image acquisition. Imaging was continued for 12 min until formation of cephalic furrow and posterior dorsal folds started to displace the tissue (left and right yellow arrows, respectively). Scale bars, 10 µm. **b** Rate of myosin II accumulation over time quantified from integrated intensity projections of Sqh::mCherry signal within 5 µm of the cells' apical-most plane (three embryos, with *n* = 21, 36, and 21, cells respectively). Results are presented as mean and s.d. **c** Corresponding changes in cell-shape over time quantified by the rate of constriction for the same three embryos as in **b**. Results presented as mean and s.d. **d**, **e** Quantification of apical area and myosin II intensity (**d**) and corresponding constriction and myosin II accumulation rates (**e**) for an individual cell shown in **f**. The arrowhead in magenta indicates an example of a contractile event. **f** Detailed view of an individual cell as it undergoes cycles of contraction and expansion as quantified in **d** and **e**. Magenta arrowhead indicates a contractile event. Scale bar is 5 µm. **g** Correlation coefficient resulting from temporally shifting myosin II accumulation rate with respect to the apical constriction rate (429 pulses in 78 cells from three embryos) presented as mean and s.d. The average maximum correlation coefficient occurs when myosin accumulation rate is not shifted. **h** Distribution of time offsets at a peak of maximum correlation for individual cross-correlation curves calculated in **g**. 40% of apical contractions occurred within a 20 s interval from a myosin peak (first bin in the graph). n.d. indicates that no peak of maximum correlation was found between two consecutive contractile events (\~10%)
Continuous RhoGEF2 activation and deep tissue invagination {#Sec6}
----------------------------------------------------------
The data so far collected suggest a direct relationship between strength of Rho signaling activation and tissue invagination. To further test the relationship between signal strength and invagination, we assessed whether continuous activation of Rho signaling would be sufficient to drive complete invagination. Cells were photo-activated at the apical surface of the embryo and the photo-activation plane was adjusted throughout the course of the experiment to follow the invaginating cells as they moved away from the focal plane of imaging. A series of z-stacks covering the entire apical--basal length of the cells was acquired at different time points and projected on a *zy*-plane. After 10 min of photo-activation, cells that moved away from the surface of the embryo formed a shallow indentation (Fig. [5a--c](#Fig5){ref-type="fig"}), which rapidly (\<1 min) acquired a more convex form (Fig. [5d--f](#Fig5){ref-type="fig"}) before folding into a U-shaped invagination (Fig. [5g--i](#Fig5){ref-type="fig"}). Imaging of RhoGEF2-CRY2 and myosin II confirmed selective optogenetic activation at the apical surface, with both markers displaying a maximal spreading in z along the apicobasal axis of the cells of \~4 microns (Fig. [6a--f](#Fig6){ref-type="fig"}). Similar optogenetic-guided invaginations could also be induced at later stages of embryogenesis, when the embryo had already acquired a more complex morphology (Supplementary Fig. [2](#MOESM1){ref-type="media"}), arguing that recruitment of RhoGEF2 to the plasma membrane is on its own sufficient to cause tissue invagination on a time-scale compatible with endogenous invagination processes.Fig. 5Sustained RhoGEF2 optogenetic activation is sufficient to drive a complete invagination. **a**--**i** Confocal images of three individual embryos co-expressing CIBN::pmGFP, RhoGEF2-CRY2, and Hist2A::mRFP as nuclear marker. Within a rectangular on the dorsal epithelium, two-photon illumination was restricted to three consecutive z-stacks (1 µm spacing) centered at 4 µm from the apical-most plane, and excited for \~152 ms during each acquisition round. Sustained photo-activation was alternated with mCherry excitation at intervals of \~13 s. As the epithelium folded inwards, continuous two-photo-activation was achieved by manually adjusting the ROI such that illumination was restricted to the initial group of photo-activated cells. In addition, the focal plane of activation was manually rectified to match the cells' apical-most plane. At the indicated time points, a z-stack comprising the dorsal epithelium was acquired for both mRFP and GFP (50 planes, 1 µm spacing). Overlays of both channels are shown. **a**--**c** Three views of the dorsal epithelium after 10:00 min of continuous activation. **a** Surface view of the dorsal epithelium at a depth of 4 µm. **b** Apices of activated cells at a depth of 8 µm. **c** Cross-section view of the activated area. **d**--**f** Three views of the dorsal epithelium after 10:48 min of continuous activation. **d** Surface view of the dorsal epithelium at a depth of 4 µm. **e** Apices of activated cells at a depth of 14 µm. **f** Cross-section view of the activated area. **g**--**i** Three views of the dorsal epithelium after 12:59 min of continuous activation. **g** Surface view of the dorsal epithelium at a depth of 4 µm. **h** Apices of activated cells at a depth of 17 µm. **i** Cross-section view of the activated area. Dashed lines indicate the depths at which surface views were acquired (left and middle panels). Scale bars, 10 µmFig. 6Spatial precision of photo-activation (z-spreading). **a**, **d** Confocal still frames from a time-lapse recording of representative embryos (*N* = 6) co-expressing **a** CIBN::pmGFP, RhoGEF2-CRY2::mCherry, or (**d**) CIBN::pmGFP, RhoGEF2-CRY2, and Sqh::mCherry (myosin II). A rectangular region (blue box) spanning 5 × 15 cells (XY plane) in three consecutive z-stacks (1 µm spacing) centered at 4 µm from the apical-most plane, was excited for \~228 ms for each acquisition round and alternated with mCherry excitation at intervals of 17 s. Images represents integrated intensity projections of 4 µm of the mCherry signal along the transversal cross-section. Scale bars are 10 µm. **b** Kymograph of RhoGEF2-CRY2::mCherry recruitment along the apico-basal axis of photo-activated cells and (**c**) spreading along the apico-basal axis measured by the Full Width at Half Maximum (FWHM) obtained from Gaussian fitting of the apical intensity signal peak (*N* = 3 embryos). **d** Confocal still frames from a time-lapse recording of a representative embryo co-expressing CIBN::pmGFP, RhoGEF2-CRY2, and Sqh::mCherry. **e** Kymograph of Sqh::mCherry accumulation along the apico-basal axis and (**f**) myosin II accumulation and spreading along the cells' apico-basal axis measured by the FWHM obtained from Gaussian fitting of the apical intensity signal peak (*N* = 3 embryos). Yellow arrowhead in **d** and **e** indicates myosin-II at the basal surface of the cells
Tissue geometry and the orientation of cell contractions {#Sec7}
--------------------------------------------------------
During endogenous invagination processes, cells are arranged in predefined geometrical patterns, which seem to impact on apical contractile behavior. For example, during ventral furrow formation, cells constrict preferentially along the DV axis and elongate along the AP axis (AP anisotropic constriction)^[@CR46]^. This contractile behavior correlates with the geometry of the ventral furrow primordium, which is composed of cells organized in a rectangular shape oriented with its major axis parallel to the AP axis of the embryo^[@CR32]^. However, ventral cells might also be subject to additional forces and constraints imposed, for example, by the specific curvature of the embryo or by its overall oblong morphology. To test whether there is a direct link between tissue geometry and contractile behavior, we monitored apical constriction in synthetic furrows of different quadrilateral shapes and orientation (Fig. [7](#Fig7){ref-type="fig"} and Supplementary Movie [6](#MOESM9){ref-type="media"}). Cells arranged in a squared geometry constricted isotropically (Fig. [7c, d](#Fig7){ref-type="fig"}), while cells arranged in a rectangular pattern constricted anisotropically along an axis perpendicular to the major edge of the rectangular pattern that was designed and thus acquired an elongated shape along that edge (Fig. [7k, l](#Fig7){ref-type="fig"}). The degree of anisotropy increased as a function of the rectangularity of the furrow: the higher the ratio between the major and the minor edge, the higher the degree of anisotropy (Fig. [7e--](#Fig7){ref-type="fig"}[m](#Fig7){ref-type="fig"}). Changing the orientation of a synthetic furrow by 90° (i.e., a furrow that was designed with its major axis perpendicular to the AP axis of the embryo) caused a re-orientation of the anisotropy, with cells constricting preferentially along the AP axis and elongating along the DV axis (Fig. [7a, b, n](#Fig7){ref-type="fig"}). Thus, these data demonstrate a direct correlation between anisotropic constriction and tissue geometry, independently of the overall shape of the embryo.Fig. 7The geometry of synthetic contractile domains determines the orientation of cell contraction. **a**--**n** Confocal images of six embryos co-expressing CIBN::pmGFP, RhoGEF2-CRY2::mCherry, and Gap43::mCherry. A single rectangular area of varied width to height ratio (3:1, 1:1, 2:1, 4:1, 6:1, 9:1) was photo-activated on the dorsal epithelium of different embryos. Sustained photo-activation was alternated with mCherry excitation at intervals of \~19 s. After \~6 min, a z-stack was acquired in the mCherry channel to visualize the plasma membrane, note that at the used setting, the RhoGEF2-CRY2::mCherry signal was not visible. Dashed line indicates the area of photo-activation. Scale bars, 10 µm. **a**, **b** Surface views at a depth of 4 and 10 µm for a photo-activation area with a ratio of 1:3. **c**, **d** Surface views at a depth of 4 and 8 µm for a photo-activation area with a ratio of 1:1. **e**, **f** Surface views at a depth of 4 and 8 µm for a photo-activation area with a ratio of 2:1. **g**, **h** Surface views at a depth of 4 and 11 µm for a photo-activation with a ratio of 4:1. **i**, **j** Surface views at a depth of 4 and 9 µm for a photo-activation with a ratio of 6:1. **k**, **l** Surface views at a depth of 4 and 7 µm for a photo-activation with a ratio of 9:1. **m** Quantification of cell elongation along the anterior--posterior embryonic axis (AP elongation, *y*-axis) in relation to the geometrical pattern of photo-activation measured by the width:height ratio, *x*-axis. (3:1, *n* = 34; 1:1, *n* = 43; 1:2, *n* = 25; 1:4, *n* = 32; 1:6, *n* = 52; 1:9, *n* = 91). Linear regression analysis reveals a direct relation between cell anisotropy and geometry of activation (red dashed line, slope of 0.032 ± 0.0079, Pearson's correlation coefficient of 0.42, \*\*\**p*-value of 7.01 × 10^−14^, right-tail *t*-test against a null hypothesis of a slope equal to zero, black dashed line). Shaded areas indicate root-mean-square deviation (RMSD). **n** Comparison of AP and DV cell elongation between a longitudinal (left) and transversal (right) rectangular pattern of photo-activation. While cell elongation occurs mainly along the AP axis in a longitudinal area of photo-activation, it reorients along the DV axis in the case of a transversal area of photo-activation (\**p*-value \< 0.005, \*\*\**p*-value \< 0.0001, two-tailed paired *t*-test)
Discussion {#Sec8}
==========
Taken together the results presented in this study show that localized activation of Rho signaling at the apical surface of cells, which are otherwise not programmed to invaginate, is sufficient to cause tissue invagination and to recapitulate major cell- and tissue-level behaviors associated with endogenous invagination processes. As discussed in detail in the Introduction, mechanisms other than apical constriction control a variety of different forms of invaginations during animal development. Our results do not challenge this view, rather, they argue that if considering a monolayer of epithelial cells, apical constriction is sufficient to fold it into a U-shape invagination. However, apical constriction is not sufficient to drive closure of an invagination into a tube-like structure, as seen for example during ventral furrow formation. Additional pushing forces exerted by lateral non-invaginating cells and/or loss of myosin II from the basal surface and basal relaxation might be required to complete this step, as suggested by a recent study based on a 2D finite element model^[@CR30]^.
At the tissue-level, the contractile behavior of individual cells depends on the geometry of photo-activation. While a square box results in isotropic apical constrictions, a rectangular shape causes cells to constrict preferentially along the minor axis of the photo-activated area and to elongate along the major axis. This anisotropic contractile behavior resembles the one of ventral furrow cells, which are also organized in a rectangular pattern and constrict preferentially along the short axis of the tissue. Anisotropy in ventral furrow cells is not genetically determined but arises as a consequence of tissue geometrical and mechanical constraints^[@CR32],[@CR33]^. Consistent with these studies, the increase in the degree of anisotropic constriction as a function of the rectangularity of the photo-activated area presented in Fig. [7](#Fig7){ref-type="fig"} can be explained if considering that it is mechanically less favorable to shrink cells along the major axis of a rectangle than along the short axis. Indeed, the former deformation requires the endpoints of the constricting tissue to move farther, and thus a larger deformation of neighboring tissues along that axis.
Our results also reveal an interesting correlation between pulsatile constrictions and tissue invagination. During endogenous morphogenetic processes, two different pulsatile behaviors have been described. One is based on cycles of myosin II accumulation at the medio-apical plane of the cell, during the contraction phase, and dissolution during the relaxation phase. This type of pulsatile behavior has been first described during dorsal closure in *Drosophila* and it is not linked to tissue invagination^[@CR43]^. Another type of pulsatile behavior is based on an incremental accumulation of myosin II at each contraction, which is followed by a stabilization period of cell shape without an intervening relaxation phase (ratcheted contractions)^[@CR44]^. Ratcheted contractions require the radial polarization of Rho signaling components. However, in certain mutant conditions that interfere with the establishment of radial polarity, contractions become non-ratcheted with myosin II miss-localizing at three-ways junctional vertices^[@CR39]^. The optogenetic-induced pulsatile contractions described in Fig. [3](#Fig3){ref-type="fig"} display a non-ratchet behavior, and similarly to dorsal closure contractions, are characterized by myosin II medio-apical accumulation and dissolution in phase with contraction and relaxation of the apical surface. However, differently from dorsal closure, optogenetic-induced pulsatile contractions display a higher degree of synchrony with photo-activated cells constricting and relaxing in concert. This difference could be explained if considering that a light pulse provides a coherent and synchronous input, while activation of signaling in a developing tissue might be more subject to cell-to-cell variability. Lack of tissue internalization upon induction of pulsatile constrictions is likely due to the absence of a stabilization phase after constriction of the apical surface, which might result in a dissipation of the forces that are normally needed to build tension and drive invagination. Consistently, continuous administration of light induced synchronous contractile behavior and invagination, mimicking the activity of signaling molecules such as *fog* whose function is to control the transition from stochastic to collective contractile behavior during ventral furrow invagination^[@CR47]^. Pulsatile behavior could be elicited either by a discontinuous administration of light, or by continuous illumination at a lower laser power, or by a single pulse at a higher laser power. We interpret these results to suggest that pulsations can be induced by the stimulation of a Rho-dependent mechano-chemical oscillatory system up to a certain threshold, above which cells constrict without pulsing. Stimulation of Rho signaling above a certain threshold could override, for example, the activity of a RhoGAP, which is required to control the normal spatio-temporal dynamics of Rho GDP/GTP cycling. In agreement with this proposal, pulsatile constrictions during ventral furrow invagination require the activity of a specific RhoGAP^[@CR48]^. However, while ventral cells pulse with a mean period of \~80 s^[@CR44]^, optogenetic-induced pulsations display a mean period of \~150 s, a limit probably imposed by the reversion kinetics of the CRY2/CIB1 system in the dark^[@CR32],[@CR40]^.
In conclusion, these data illustrate the utility of applying concepts of synthetic biology (e.g., precise orthogonal control over signaling pathways, guided cell behavior) to the field of tissue morphogenesis and in particular of how the nascent field of synthetic morphogenesis can help defining the minimum set of requirements sufficient to drive tissue remodeling. Our data argue that while normally tissue differentiation and tissue shape are intimately linked, it is possible to direct tissue shape without interfering with complex layers of gene regulatory network and tissue differentiation programs. This might have important implications also for tissue engineering, where it might be desirable to shape any given tissue of interest without changing its fate.
Methods {#Sec9}
=======
Cloning {#Sec10}
-------
To generate RhoGEF2-CRY2::mCherry, the DHPH catalytic domain of RhoGEF2 was PCR-amplified from *Drosophila melanogaster* cDNA using gene-specific primers and joined with the CRY2::mCherry sequence through overlap extension PCR (aa 1177--1554, from the RhoGEF2 reference sequence NP_995869.1). Non-fluorescently-tagged RhoGEF2-CRY2 was cloned by amplifying RhoGEF2-CRY2 from RhoGEF2-CRY2::mCherry. Both constructs were cloned into the pPW vector (Drosophila Genomics Resource Center, Bloomington, IN) using the Gateway cloning system (Life Technologies) according to standard procedures.
Because the available UAS-sqh-Gap43::mCherry plasma membrane marker harboring UAS regulatory sequences caused morphological abnormalities when expressed at high level in combination with Gal4 expression; we generated a spaghetti-squash-promoter-only driven Gap43::mCherry (sqhp\>Gap43::mCherry). To generate this construct, both the *D. melanogaster* spaghetti squash promoter gene sequence (−1175 to +153 from the Sqh reference sequence NM_001298002) and mCherry were amplified with gene-specific primers that additionally contained the Gap43 sequence. The fragments were then joined together into pCasper5, cut with NotI and AgeI, using Gibson Assembly.
Fly strains and genetics {#Sec11}
------------------------
The following transgenic fly lines were obtained by standard methods and all stocks were maintained at 22 °C. Standard genetic crosses were kept in the dark and were used to generate flies having the genotypes described below.
To show the effect RhoGEF2-CRY2 plasma membrane recruitment on the dorsal epithelium:
w\[\*\]; P\[w+, UASp\>CIBN::pmGFP\]/+; P\[w+, UASp\>RhoGEF2-CRY2::mCherry\]/P\[w+, Osk\>Gal4::VP16\].
To follow cell outlines upon RhoGEF2-CRY2 activation:
P\[w+, sqhp-Gap43::mCherry\]/w\[\*\]; P\[w+, UASp\>CIBN::pmGFP\]/+; P\[w+, UASp\>RhoGEF2-CRY2::mCherry\]/ P\[w+, Osk\>Gal4::VP16\].
To visualize myosin II dynamics upon RhoGEF2-CRY2 activation:
w\[\*\]; P\[w+, UASp\>CIBN::pmGFP\]/P\[w+, Sqh::mCherry\]; P\[w+, UASp\>RhoGEF2-CRY2\]/ P\[w+, Osk\>Gal4::VP16\].
To visualize both membranes and nuclei together:
w\[\*\]; P\[w+, UASp\>CIBN::pmGFP\]/P\[w+, his2Av::mRFP1\]; P\[w+, UASp\>RhoGEF2-CRY2\]/ P\[w+, Osk\>Gal4::VP16\].
For embryo collection, flies of the desired genotype were collected into cages with apple juice plates and yeast paste. All cages were kept in the dark at 22 °C.
Fly stocks {#Sec12}
----------
w\[\*\]; If/CyO; P\[w+, UASp\>RhoGEF2-CRY2::mCherry\]/TM3, Ser (this study). UAS-Gal4-driven expression of the catalytic DHPH domain of RhoGEF2 fused upstream of CRY2 PHR domain and C-terminally tagged with mCherry fluorescent protein.
w\[\*\]; If/CyO; P\[w+, UASp\>RhoGEF2-CRY2\]/TM3, Ser (this study). UAS-Gal4-driven expression of the catalytic DHPH domain of RhoGEF2 fused upstream of CRY2 PHR domain.
P\[w+, sqhp\>Gap43::mCherry\]/Fm7; Sb/TM6 Tb (this study). Gap43 membrane marker fused to mCherry fluorescent protein in which expression is driven only by the Spaghetti-squash promoter.
w\[\*\]; P\[w+, UASp\>CIBN::pmGFP\]/Cyo; Sb/TM3, Ser. Membrane-anchored CIBN additionally fused to EGFP.
w\[\*\]; If/CyO; P\[w+, his2Av::mRFP1\]/TM3, Ser. mRFP1 tagged Histone 2A version A. (Bloomington stock number 23651.)
w\[\*\]; P\[w+, Sqh::mCherry\]/CyO; Sb/TM3, Ser. Myosin regulatory light chain Spaghetti-squash tagged with mCherry fluorescent protein.
w\[\*\]; If/CyO; P\[w+, Oskp\>Gal4::VP16\]/TM3, Ser. Maternally deposited Gal4 protein driven by the Oskar promoter. (Bloomington stock number 44242.)
Live imaging and optogenetics {#Sec13}
-----------------------------
For collection of light-sensitive embryos, cages were kept and manipulated in the dark. Stage 5 embryos were identified and mounted using a standard stereomicroscope under transmitted illumination. To prevent unwanted photo-activation, the microscope light source was replaced with a conventional red-emitting LED lamp. Following selection under halocarbon oil, the embryos were dechorionated with 90% sodium hypochlorite for 2 min, washed with water and mounted with PBS onto a 35 mm glass-bottom dish (MatTek corporation). The embryos were positioned with their dorsal side facing the coverslip.
Live imaging and photo-activation experiments, as well as image acquisition of TUNEL and antibody staining, were carried out at 20 °C with a Zeiss LSM 780 NLO confocal microscope (Carl Zeiss) equipped with a 561 diode laser, an argon laser, and a tunable two-photon laser (690--1040 nm) (Chameleon; Coherent, Inc.). A 40×/NA 1.1 water immersion objective (Carl Zeiss) was used for image acquisition. For sample location, bright field illumination was filtered through a Deep Amber lighting filter (Cabledelight, Ltd.). The microscope was controlled through the Zen Black software (Carl Zeiss) whereas photo-activation protocols were carried out with the Pipeline Constructor Macro^[@CR49]^. For all experiments, an initial mCherry frame was acquired prior to activation (acquisition at 561 nm) and followed by time-lapse recording of alternating mCherry and local two-photon excitations. For experiments in which transversal cross-sections are shown, a z-stack comprising the dorsal epithelium was acquired (69 planes, 1 µm spacing) in mCherry during photo-activation and in GFP posterior to photo-activation. Local two-photon photo-activation was achieved at 950 nm with laser power set to 10 mW and scanning direction set to bidirectional yielding a pixel dwell of 1.27 µs. Unless otherwise stated, these settings were maintained throughout experiments. Laser power measurements are reported as measured at 1 cm from the objective.
For assessing the effect of RhoGEF2-CRY2::mCherry plasma-membrane recruitment, the dorsal epithelium was activated within the given ROI (circle, triangle, square) and limited to three consecutive z-stacks (1 µm spacing) centered at 4 µm from the apical-most plane. An initial mCherry stack was acquired followed by eight consecutive rounds of two-photon activation for a total effective illumination time of 2 s (\~257 ms per illumination round at 10 mW laser power). Subsequently, a time-lapse recording of alternated mCherry and local two-photon excitation was acquired. The time between two-photon excitations was 21 s.
In order to follow apical constriction, myosin II dynamics, and any potential induction of apoptosis upon photo-activation, a ROI of activation on the dorsal epithelium was designed. An initial mCherry frame was acquired prior to photo-activation followed by a time-lapse recording of alternating mCherry and local two-photon excitations. The region for photo-activation was limited to three consecutive z-stacks (1 µm spacing) centered at 4 µm from the apical-most plane. During each round of activation, depending on the size of the region, a z-stack was excited for a total effective illumination time between 152 and 380 ms per pulse, at a laser power of 10 mW. The time between two-photon excitations was between 13 and 24 s.
In order to test the extent to which synthetic furrow could be induced to internalize, to determine the extent of z-spreading of RhoGEF2-CRY2::mCherry recruitment and myosin II accumulation, the same protocol as described above was employed with the following modifications. The ROI was manually adjusted to fit the initial area of activated cells as they moved away from the focal plane and constrict their apical area. As the epithelium folded inwards, the central plane of photo-activation was manually adjusted to fit the first four micrometers from the cells' apical-most plane.
Induction of pulsatile contractions {#Sec14}
-----------------------------------
For discontinuous photo-activation, a ROI was activated in three consecutive z-stacks (1 µm spacing) centered at 4 µm from the apical-most plane. During each round of activation, the selected region was excited for a total effective illumination time of \~228 ms per pulse, at a laser power of 10 mW. The time between two-photon excitations was 55 s. For continuous photo-activation at a lower laser power, the selected region was excited at a laser power of 5 mW. The time between two-photon excitations was 13 s, the maximum imaging resolution that could be achieved with our system configuration. For testing the effect of photo-activation upon a single illumination pulse, the region of photo-activation was excited for 15 consecutive rounds of two-photon illumination at a laser power of 15 mW, for a total effective illumination time of \~4 s. Subsequently, in all cases, cell outlines in five consecutive z-stacks (1 µm spacing) were recorded in mCherry during or after photo-activation. For following myosin II dynamics, the dorsal epithelium was subjected to an initial single illumination pulse. Two-photon excitation was restricted to three consecutive z-stacks (1 µm spacing) centered at 4 µm. The region was excited for five consecutive rounds of two-photon illumination at a laser power of 15 mW, for a total effective illumination time of \~2 s. Recordings for both a single plane of two-photon (2 mW) and five consecutive z-stacks (1 µm spacing) of mCherry were acquired alternating consecutively in intervals of \~13 s.
Image and data analysis {#Sec15}
-----------------------
Images were processed and analyzed in MATLAB (MathWorks) using a custom-written pipeline. Images in LSM format were opened and metadata extracted with MATLAB Bioformats toolbox ([www.openmicroscopy.org](http://www.openmicroscopy.org)). For image segmentation, the software Ilastik was used (ilastik.org). The resulting pixel classifier was preserved for images of individual experiments. The generated probability maps were smoothened (size = \~10 px, *σ* = 2) and thresholded (thr. = 0.5). The binary image was subsequently skeletonized and subjected to watershed transformation. This process resulted in a mask for cell boundaries of 1 px in width. The membrane mask was inverted to obtain masks for individual cells. Cell features such as area and eccentricity, major and minor axis, and orientation were computed using the function regionprops in MATLAB. Myosin II accumulation pulses were quantified from integrated intensity projections of 5--6 µm for individual cells using the mask from individual cells. We defined cell elongation relative to either AP or DV axis. In brief, AP and DV elongation correspond to the *x* and *y* components of the unit vector along the major axis of the cell weighted by cell eccentricity.
Tracking of individual cells was performed by nearest-neighbor assignment, in which the region-overlap between adjacent frames determined the cell identity. For all analysis, we filtered out objects with extreme feature measurements (\>2.5 std), as well as object tracks with gaps greater than 2 time points. Myosin intensity for individual cell-masks was measured from integrated intensity projections within 5 µm of the apical-most plane (five consecutive z-stacks, 1 µm spacing). For all measurements of apical area and myosin II accumulation, time-course data was subjected to smoothing by an averaging window of 3 time points. Both instantaneous constriction and myosin accumulation rates were calculated by subtracting consecutive time points and dividing them by their corresponding time interval. For individual pulse identification and cross-correlation analysis, we used the MATLAB functions *findpeaks* and *xcorr2* using the time-course measurements of the constriction and the myosin accumulation rates. For calculating the spreading in the *z*-axis, we used the full width at half maximum (FWHM) of a Gaussian fitted to a line intensity-plot along the cell's apico-basal axis parameterized by the peak of maximum intensity at the cell's apex.
Statistical analysis {#Sec16}
--------------------
Statistical analyses were performed in MATLAB (MathWorks). For determining the relationship between geometry and cell elongation, we first identified the activated cells at the beginning of the protocol and followed them until the formation of a furrow. At that point, we measured their elongation as described in Image and Data Analysis section. We sought for a significant linear regression between the width to height ratio and cell AP elongation using the model:$$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$Y = b_0 + b_1X$$\end{document}$$
To test whether the slope of the regression line differs significantly from zero, we used the *t* statistic defined by:$$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$t = \frac{{b_1}}{{SE}}$$\end{document}$$where$$\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$SE(b_1) = \frac{{\sqrt {\frac{{\Sigma (y_i - \hat{y}_i)^2}}{{(n - 2)}}} }}{{\sqrt {\Sigma (x_i - x)^2} }}$$\end{document}$$
Finally, we used the tcdf MATLAB function to extract the *p*-value from the upper-tail probability of *t*, corresponding to test for a positive slope against a null hypothesis of a slope equal to zero. For testing differences in the orientation of anisotropy upon change in ROI orientation, and for testing the differences in the period durations for the different illumination protocols, we used two-tailed paired Student's *t*-tests.
TUNEL and antibody double labeling {#Sec17}
----------------------------------
For detection of apoptosis, embryos expressing CIBN::pmGFP, RhoGEF2-CRY2::mCH, and the plasma membrane Gap43::mCh were dechorionated for 2 min in sodium hypochlorite solution. After activation, individual embryos were fixed in 4% paraformaldehyde (Electron Microscopy Sciences) and heptane (Sigma) for 20 min in the dark. The embryos were then devitellinized and kept in methanol at −20 °C. As a positive control for apoptosis, embryos that had passed stage 10 of embryogenesis and had therefore started apoptosis, were dechorionated and fixed as described above. Embryos were blocked in 10% bovine serum albumin (BSA) in PBS, 0.1% Triton-X-100 (Sigma) for 1 h. Fixed embryos were incubated with an anti-GFP antibody (abcam 6556, 1:1000) in PBS containing 5% BSA and 0.1% Triton-X-100 for 2 h at room temperature. Embryos were washed four times in PBS, 0.1% Triton-X-100, and incubated overnight at 4 °C in the dark with Alexa 647 anti-rabbit secondary antibodies (1:500), diluted into the appropriate TUNEL reaction mixture (In Situ Cell Death Detection Kit, Fluorescein, Roche). After four washes in PBS, 0.1% Triton-X-100 and four washes in PBS. For image acquisition, stained embryos were mounted in MatTek dishes covered with PBS and orientated to display the previously activated area.
Data availability {#Sec18}
-----------------
The authors declare that all data supporting the findings of this study are available within the article and its supplementary information files or from the corresponding author upon reasonable request.
Electronic supplementary material
=================================
{#Sec20}
Supplementary Information Peer Review File Description of Additional Supplementary Files Supplementary Movie 1 Supplementary Movie 2 Supplementary Movie 3 Supplementary Movie 4 Supplementary Movie 5 Supplementary Movie 6
**Electronic supplementary material**
**Supplementary Information** accompanies this paper at 10.1038/s41467-018-04754-z.
**Publisher\'s note:** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank all members of the De Renzis laboratory for helpful discussion and Giorgia Guglielmi for helping during the initial stage of this project. We thank M. Coppey for sharing critical information on the RhoGEF2-CRY2 construct used in this study. We thank D. Arendt, E. Furlong, J. Rink, and A. Runge for critical reading of the manuscript. We thank the advanced light microscopy core facility (EMBL, Heidelberg) for their advice and assistance, and A. Politi for providing the Pipeline Constructor macro. We thank B. Klaus from the EMBL Centre for Statistical Data Analysis for discussions related to the data analysis. We thank the Bloomington Drosophila Stock Center for providing fly stocks and the Drosophila Genomics Resource Center for providing cDNAs.
The experiments were conceived and designed by E.I. and S.D.R. Cloning, TUNEL assays, and transgenic flies by T.Q. All remaining experiments were performed by E.I. All the data were analyzed by E.I. and S.D.R., who also wrote the manuscript together.
Competing interests {#FPar1}
===================
The authors declare no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-antioxidants-08-00004}
===============
Peroxiredoxins (Prdxs) are a widely distributed family of antioxidant enzymes that use a cysteine thiol group as a catalytic center to catalyze the reduction of hydroperoxides \[[@B1-antioxidants-08-00004]\]. These enzymes function in anti-oxidant defense and they also have an important role in cell signaling through the regulation or sensing of local peroxide concentration \[[@B2-antioxidants-08-00004]\]. The mammalian peroxiredoxin (Prdx) family consists of six members with a common mechanism for their peroxidase activity. The three-step peroxidatic cycle for the Prdx proteins involves: (i) reduction of a hydroperoxide substrate through the oxidation of a catalytic Cys to a sulfenic acid, followed by; (ii) formation of a disulfide; and, (iii) reduction of the disulfide to regenerate the active protein. The last mammalian peroxiredoxin to be described was Prdx6, a protein that is expressed in essentially all tissues but at particularly high levels in lung, brain, eye, and testes \[[@B3-antioxidants-08-00004],[@B4-antioxidants-08-00004],[@B5-antioxidants-08-00004]\]. While Prdx6 is similar to Prdxs 1--5 in its use of a catalytic Cys residue (C47 in Prdx6) to reduce H~2~O~2~, short chain hydroperoxides, and peroxynitrite (step i above) \[[@B6-antioxidants-08-00004],[@B7-antioxidants-08-00004]\], it expresses several unique features that distinguish it from other Prdx family members.
A major difference between Prdx6 and Prdxs 1--5 relates to both steps (ii) and (iii) (the resolution phases) of the peroxidatic catalytic cycle. These steps in the cycle are accomplished in Prdxs 1--5 by a Cys that is intrinsic to the protein (called the resolving Cys), followed by reduction of the disulfide with thioredoxin \[[@B3-antioxidants-08-00004]\]. On the other hand, Prdx6 uses glutathione (GSH) catalyzed by GSH S-transferase (GST) to complete the reaction cycle \[[@B8-antioxidants-08-00004],[@B9-antioxidants-08-00004],[@B10-antioxidants-08-00004]\]. Based on the characteristics of the catalytic mechanism reflecting the number of conserved cysteine residues, Prdxs 1--5 have been called 2-Cys enzymes. While Prdx6 has been called 1-Cys Prdx. Another important difference between Prdx6 and other family members is that Prdx6, unlike Prdx 1--5, has the ability to reduce phospholipid hydroperoxides with a rate constant similar to that for the reduction of H~2~O~2~ (∼10^6^ M^−1^ s^−1^) \[[@B9-antioxidants-08-00004],[@B11-antioxidants-08-00004]\].
In addition to differences in the peroxidatic reaction cycle and importantly for this manuscript, Prdx6 expresses several enzymatic functions that are not expressed by Prdxs 1--5. One of these enzymatic functions of Prdx6 is a phospholipase A~2~ (PLA~2~) activity. This latter activity is not associated with C47 but is based on an enzymatically active site consisting of a Ser-His-Asp catalytic triad \[[@B12-antioxidants-08-00004]\]. Finally, Prdx6 has a third enzymatic activity that is based on a HxxxxD active site (spanning the amino acids 26--31) that catalyzes lysophosphatidylcholine acyl transferase activity (LPCAT) \[[@B13-antioxidants-08-00004]\]. Thus, Prdx6 appears to play a unique role in phospholipid metabolism with the ability to catabolize (PLA~2~ activity) and/or remodel phospholipids (LPCAT activity) as well as to maintain these phospholipids in their reduced state (peroxidase activity). Regulation of the PLA~2~ activity of Prdx6 is the major focus of this mms.
Our initial reports describing Prdx6-PLA~2~ activity indicated that the enzyme is active at acidic pH (pH 4), but it has relatively little activity at neutral pH and above \[[@B8-antioxidants-08-00004]\]. This enzymatic activity at acidic pH is consistent with the localization of the protein to acidic organelles, such as lung lamellar bodies and lysosomes, where it plays an important role in phospholipid turnover \[[@B6-antioxidants-08-00004],[@B14-antioxidants-08-00004]\]. Thus, when first described, the enzyme was given the trivial name acidic, Ca^2+^ independent PLA~2~ (aiPLA~2~) to reflect its acidic pH requirement for catalysis and its catalytic activity in the absence of Ca^2+^ \[[@B15-antioxidants-08-00004],[@B16-antioxidants-08-00004]\]. However, subsequent studies have indicated that the PLA~2~ activity of Prdx6 also plays important roles in the repair of peroxidized cell membranes \[[@B17-antioxidants-08-00004],[@B18-antioxidants-08-00004]\], as well as in the activation of NADPH oxidase (type 2) \[[@B19-antioxidants-08-00004],[@B20-antioxidants-08-00004]\], functions that presumably require PLA~2~ activity in the cytosol at approximately neutral pH. This requirement can be accomplished by several mechanisms. First, phosphorylation of the protein results in a markedly increased activity at pH 7--8 (as well as at acidic pH) \[[@B21-antioxidants-08-00004]\]; we have shown that conformational change of Prdx6 upon its phosphorylation is the basis for the enhancement of substrate binding and increased PLA~2~ enzymatic activity \[[@B22-antioxidants-08-00004]\]. Second, the presence of an oxidized substrate, e.g., a phospholipid hydroperoxide, significantly enhanced enzymatic activity at neutral pH \[[@B23-antioxidants-08-00004]\]. As a third mechanism, oxidation of the conserved C47 of Prdx6 following the treatment of cells with H~2~O~2~ resulted in increased PLA~2~ activity at neutral pH \[[@B24-antioxidants-08-00004]\].
An additional mechanism that results in increased Prdx6-PLA~2~ activity at pH 7 was discovered by serendipity. As described above, GSH is required for Prdx6 peroxidase activity; it is commonly used as a reactant in a peroxidase activity assay that is routinely carried out at pH 7 \[[@B9-antioxidants-08-00004]\]. During our use of this assay, we identified products in the incubation medium that appeared to reflect the presence of PLA~2~ activity and postulated that this might have been due to the presence of GSH. Further study indicated that GSH did indeed stimulate Prdx6-PLA~2~ activity at pH 7, while there was no effect on activity at pH 4 \[[@B25-antioxidants-08-00004]\]. The present study was undertaken to document this result and to evaluate a possible mechanism for this effect of GSH. This publication is part of a forum on Peroxiredoxin 6 as a Unique Member of the Peroxiredoxin Family.
2. Materials and Methods {#sec2-antioxidants-08-00004}
========================
2.1. Reagents {#sec2dot1-antioxidants-08-00004}
-------------
1,2-Bis palmitoyl-*sn*-glycero-3-phosphocholine (DPPC), egg yolk phosphatidylcholine (PC), phosphatidylglycerol (PG), 1-palmitoyl-2-linoleoyl-*sn*-glycero-3-phosphocholine (PLPC), phosphatidylserine (PS), and cholesterol (chol) were purchased from Avanti-Polar Lipids (Birmingham, AL). Extracellular-signal-regulated kinase (Erk2) was purchased from Upstate (Millipore, Billerica, MA, USA). Protein concentration was measured by Coomassie blue binding using bovine γ-globulin as the standard (Bio-Rad, Hercules, CA, USA). A mouse monoclonal antibody against human Prdx6 was purchased from Chemicon EMD Millipore (Billerica, MA, USA). Bromoenol lactone (BEL) was obtained from Cayman Chemical (Ann Arbor, MI, USA). Reduced glutathione (GSH), oxidized glutathione (GSSG), MJ33 (1-hexadecyl- 3-trifluoroethylglycero -*sn*-2-phosphomethanol), tris (2-carboxyethyl) phosphine (TCEP), dithiothreitol (DTT), and all other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA).
2.2. Production of Recombinant Prdx6 {#sec2dot2-antioxidants-08-00004}
------------------------------------
The expression of human codon optimized Prdx6 plasmid pJexpress 414:75271---prdx6-optEc in Escherichia coli BL21 (DE3) (Novagen, Madison, WI, USA) has been described previously \[[@B25-antioxidants-08-00004]\]. The purification and identification of recombinant human Prdx6 protein was carried out by the modification of previously described methods for isolation of the rat and human proteins \[[@B10-antioxidants-08-00004],[@B22-antioxidants-08-00004],[@B23-antioxidants-08-00004],[@B26-antioxidants-08-00004]\]. Purification of recombinant Prdx6 by chromatography utilized an ion exchange diethylaminoethyl (DEAE) column in a fast protein liquid chromatography (FPLC) AKTA purifier system controlled by Unicon 5.1 software (GE Healthcare Biosciences, Uppsala, Sweden). The column was equilibrated with 40 mM Na-acetate buffer (pH 5) containing 1 mM ethylenediaminetetraacetic acid (EDTA), 3% glycerol, and 1 mM TCEP. Since the isoelectric point for Prdx6 is ∼6.0, it does not bind to positively charged DEAE resin at pH 5 and the Prdx6 protein was collected from the flow-through. (The column was rejuvenated by eluting bound (non-Prdx6) proteins using a high salt solution (40 mM Na-acetate, 2 M NaCl, 1 mM EDTA, 3% glycerol, 1 mM TCEP, pH 5) (see [Supplemental Figure S1](#app1-antioxidants-08-00004){ref-type="app"}). The purified Prdx6 fraction was concentrated using a 10 kDa molecular mass cut-off Amicon Ultra filter (Millipore, Billerica, MA, USA) followed by dialysis against 50 mM Tris-HCl buffer (pH 8) containing 1 mM EDTA, 3% glycerol, and 1 mM TCEP with a 10 kDa molecular mass cut-off Slide-A-Lyzer® dialysis cassette (Pierce, Rockford, IL, USA). The preparation was evaluated by polyacrylamide gel electrophoresis (PAGE) stained with Coomassie blue (see [Supplemental Figure S1](#app1-antioxidants-08-00004){ref-type="app"}). Immunoblots with monoclonal anti-Prdx6 as the primary and goat anti-mouse IgG as the secondary antibody were used to confirm the presence of Prdx6 in the purified protein preparation; the blots were analyzed and the purity in each band was calculated by the two-color Odyssey technique (LI-COR, Lincoln, NE, USA) \[[@B26-antioxidants-08-00004]\]. The protein was stored at −80 °C before use. For some studies, Prdx6 was phosphorylated by incubation with ERK2 in the presence of ATP and MgCl~2~, as described previously \[[@B22-antioxidants-08-00004]\] or oxidized by treating the purified protein with 100 µM H~2~O~2~ for 15 min.
2.3. Measurement of aiPLA~2~ Activity {#sec2dot3-antioxidants-08-00004}
-------------------------------------
The measurement of Prdx6-PLA~2~ (aiPLA~2~) activity has been described previously \[[@B4-antioxidants-08-00004],[@B12-antioxidants-08-00004],[@B15-antioxidants-08-00004],[@B16-antioxidants-08-00004]\]. A liposomal preparation of lipids reflecting the composition of lung surfactant was used as the substrate for the assay. The lipid mixture contained DPPC, PC, Chol, and PG in the molar ratio 50:25:15:10; liposomes were labeled with tracer \[^3^H-9,10-palmitate\]-DPPC in the *sn*-2 position. This substrate was used for all assays, unless a different substrate is specifically indicated. Lipids that were dissolved in chloroform were evaporated to dryness under N~2~ onto the wall of a Corex glass centrifuge tube; the evaporated film was resuspended in 50 mM Tris-HCl in phosphate-buffered saline (PBS), pH 7.4, vigorously mixed, frozen, and thawed three times by alternating liquid N~2~ and a 50 °C water bath, and then extruded under pressure at 50 °C for 10 cycles through a 0.1 μm pore size polycarbonate filter. Recovery of ^3^H in the liposome preparation was \>95% of the original disintegrations per min (DPMs) that were added to the lipid mixture. Analysis with a dynamic light scattering (DLS) 90 Plus Particle size Analyzer (Brookhaven Instruments, Holtsville, NY, USA) showed a homogeneous population of unilamellar vesicles with a diameter of 100--120 nm. Liposomes were stored overnight at 4 °C before use.
In experiments, to study the role of an oxidized phospholipid substrate, oxidizable liposomes were generated with PLPC substituting for egg PC in the standard liposomes. These liposomes were exposed to an ^•^OH-generating system (10 μM Cu^2+^ in the presence of 0.2 mM ascorbate) \[[@B27-antioxidants-08-00004]\] for 45 min at room temperature. The liposomes were dialyzed for 2 h against PBS (pH 7.4) using a Slide-A-Lyzer® dialysis cassette as above. These liposomes presumably contained PLPCOOH as an oxidized lipid component.
To measure PLA~2~ activity, enzyme (Prdx6) in buffer (Tris-EGTA, pH 7.4) was pre-incubated for 15 min and the reaction was started by addition of substrate (^3^H-labeled liposomes) for a 1 h incubation. Lipids were then extracted from the liposomal suspension, separated by thin layer chromatography (TLC), and the free fatty acid band was analyzed by scintillation counting. PLA~2~ activity was calculated from the liberation of ^3^H-palmitic acid.
To measure PLA~2~ activity under anaerobic conditions, the incubation medium first was purged of air in an N-EVAP (Organomation Assoc., Northborough, MA, USA) using a continuous flow of 100% N~2~ for 10 min; N~2~ flow then was maintained during the pre-incubation and incubation periods.
2.4. Mass Spectroscopy (ESI- MS and LC-MS/MS) {#sec2dot4-antioxidants-08-00004}
---------------------------------------------
Both hPrdx6 WT and hPrdx6 C91S were gel-filtered using two Superdex 75 (GE Healthcare, Chicago, IL, USA) columns that were connected in series and equilibrated in 10 mM Tris, 100 mM NaCl, pH 7.4 (TBS), and maintained at 4 °C. Each sample was diluted to 0.55 mg/mL in TBS, with or without addition of 5 mM GSH (final concentration), pH 7.4. Samples were incubated for 1 h at 37 °C, and then stored at 0 °C. After 16--20 h of storage, TCEP (5 mM final concentration) was added to half of each sample. The samples were incubated for 30 min at 37 °C and then stored on ice until analysis by electrospray ionization mass spectrometry (ESI MS). Prior to injection onto the liquid chromatography-mass spectroscopy (LC-MS) system, samples were diluted to 0.1 mg/mL. A Dionex C4 Trap column and a self packed Poros C8 analytical column (20 cm) were interfaced directly with an Orbitrap XL operating at 100 K resolution. Thermo Xtract software (Thermo Fisher Scientific, Waltham, MA, USA) was used for deconvolution of the monoisotopic masses. A portion of each sample was alkylated with iodoacetamide, followed by denaturation with urea, reduction with DTT, and then alkylation with acrylamide. Samples were digested with trypsin and analyzed using liquid chromatography-tandem mass spectrometry (LC-MS/MS) using an LTQ-Orbitrap XL™ mass spectrometer (Thermo Fisher Scientific Waltham, MA, USA). Data was searched using MaxQuant software (Max Planck Institute of biochemistry, München, Germay). The following cysteine mass modifications were included in the search: iodoacetamide derivative, 57.0215 Da; acrylamide adduct, 71.0371 Da; 2 oxidations, 31.9898 Da; 3 oxidations, 47.9847 Da. The intensity value from the MaxQuant output files for the variable modified cysteine containing peptides DFTPVCTTELGR (C47) and DINAYNCEEPTEK (C91) were plotted with the Excel program to show the relative quantitation for each peptide that was obtained after treatment of the protein with the different buffer conditions.
2.5. Fluorescence Spectroscopy {#sec2dot5-antioxidants-08-00004}
------------------------------
Fluorescence spectroscopy was performed with a spectrofluorometer (PTI, Photon Technology International, Lawrenceville, NJ, USA) equipped with a water bath temperature-controlled sample holder, a single-photon counting system for fluorescence intensity detection, and dual fluorescence and absorbance channels. Measurements were performed with 1 μM protein at 22 °C in microquartz fluorescence cuvettes with a path length of 0.3 cm. Fluorescence was excited at 295 nm and recorded in the range of 310--450 nm (the emission spectrum of tryptophan) with a 1 nm slit width for both excitation and emission.
2.6. Circular Dichroism (CD) {#sec2dot6-antioxidants-08-00004}
----------------------------
Prdx6 (6.5 μM in 10 mM Tris buffer, pH 7) was analyzed for CD in a fused quartz cell with a path length of 0.1 cm using a Chirascan™ CD Spectrometer (Applied Photophysics Ltd. Surrey, UK). Protein with or without GSH or GSSG (5 mM) was incubated for 1 h at room temperature in 10 mM Tris buffer, pH 7. Spectra were recorded with three repeats in the far-ultraviolet region (190--260 nm) with a bandwidth of 1.0 nm, a step size of 0.5 nm, and an integration time of 30 s.
2.7. Statistical Analysis {#sec2dot7-antioxidants-08-00004}
-------------------------
Data are expressed as means ± standard error (SE). Differences between mean values were analyzed by two-tailed Student's *t* test and they were considered statistically significant at *p* \< 0.05.
3. Results {#sec3-antioxidants-08-00004}
==========
3.1. Effect of GSH on the PLA~2~ Activity of Prdx6 {#sec3dot1-antioxidants-08-00004}
--------------------------------------------------
The PLA~2~ activity of recombinant human Prdx6 measured at pH 4 was \~100 nmol/min/mg protein (not shown), but it was \<1 nmol/min/mg protein when measured at pH 7.4 using the standard liposomal (non-oxidized) substrate described in Methods ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}). The addition of GSH (5 mM) to the incubation medium had no effect on aiPLA~2~ activity at pH 4 (not shown) but increased activity at pH 7.4 to \~50% of the pH 4 value ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}). The stimulation of aiPLA~2~ activity by GSH was seen at pH \> 5 ([Figure 1](#antioxidants-08-00004-f001){ref-type="fig"}). Stimulation of activity was dependent on GSH concentration with maximal effect at 3 mM GSH ([Figure 1](#antioxidants-08-00004-f001){ref-type="fig"}). GSH-stimulated activity was sensitive to MJ33, a known inhibitor of aiPLA~2~ activity, while BEL, an inhibitor of some intracellular, Ca^2+^ independent PLA~2~ enzymes, was ineffective ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}). The addition of oxidized GSH (GSSG) or the unrelated thiols, DTT, or TCEP, to the assay resulted in a relatively minor increase in aiPLA~2~ activity that was not statistically different from control ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}).
The basal activity and stimulation of activity by GSH at pH 7.4 was similar for control liposomes (DPPC as substrate) and liposomes containing linoleoyl in the *sn*-2 position of PC (PLPC) as the substrate for the PLA~2~ reaction ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}). As reported previously, aiPLA~2~ activity at pH 7.4 with oxidized PLPC (i.e., PLPCOOH) as the liposomal substrate was significantly increased when compared to control \[[@B17-antioxidants-08-00004],[@B23-antioxidants-08-00004]\] and it was increased to a much greater extent with the phosphorylation of Prdx6 \[[@B21-antioxidants-08-00004],[@B22-antioxidants-08-00004]\]. Oxidation of the Prdx6 protein increased its aiPLA~2~ activity with reduced substrate to the same level as with oxidized liposomes. Activity with PLPCOOH substrate, phosphorylated Prdx6, or with oxidized Prdx6 was not increased further in the presence of GSH ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}).
3.2. Molecular O~2~ is Required for Generation of the Sulfinic Prdx6 {#sec3dot2-antioxidants-08-00004}
--------------------------------------------------------------------
We postulated that the generation of the sulfinic form of Prdx6 in the presence of GSH represented auto-oxidation of the sulfenic form of the protein. To test this hypothesis, the assay for aiPLA~2~ activity was carried out under anaerobic conditions. The aiPLA~2~ activity of the protein was markedly inhibited in the absence of O~2~ ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}). The failure of aiPLA~2~ activity to increase in the presence of GSH under anaerobic conditions is compatible with a requirement of molecular oxygen for the oxidation of the sulfenic Prdx6. It is likely that the actual oxidant under aerobic conditions is H~2~O~2~, generated from molecular O~2~ in the presence of trace metals \[[@B28-antioxidants-08-00004]\].
3.3. Prdx6 Modifications in the Presence of GSH Evaluated by ESI- MS and LC-MS/MS {#sec3dot3-antioxidants-08-00004}
---------------------------------------------------------------------------------
The effect of GSH on the mass of Prdx6 as well as C91S-Prdx6 was determined using ESI MS ([Figure 2](#antioxidants-08-00004-f002){ref-type="fig"}). C91 is a non-conserved Cys residue that is present in the human-derived Prdx6 protein used in the present study but in most mammalian species, including rat, mouse and bovine Prdx6; C47 is the lone Cys residue \[[@B6-antioxidants-08-00004]\]. The conserved peroxidatic Cys (C47) also is the only Cys group that is present in recombinant human C91S-Prdx6 protein. The observed monoisotopic mass for both Prdx6 and C91S-Prdx6 in Tris buffered saline (TBS) deviated by −2 Da from the predicted monoisotopic mass of the sulfhydryl ([Figure 2](#antioxidants-08-00004-f002){ref-type="fig"}A), as we have published previously \[[@B26-antioxidants-08-00004]\]. These results are consistent with oxidation of the protein to the sulfenic form, followed by the loss of H~2~O and formation of a sulfenylamide between the reactive Cys47 and an adjacent amino acid residue, possibly Thr48. In the presence of TCEP, the sulfenylamide bond is reduced, yielding proteins with the expected monoisotopic mass for Prdx6 and for C91S-Prdx6 (sulfhydryl forms).
In the presence of GSH added to Prdx6 in TBS solution, the molecular mass of the protein increased to 24,920, a gain of 34 Da ([Figure 2](#antioxidants-08-00004-f002){ref-type="fig"}B). Our interpretation of this result is that the addition of GSH leads to reversal of the sulfenylamide to reform the sulfenic (+18 mass increase) and the subsequent oxidation of the sulfenic to sulfinic with the addition of an oxygen atom (+16 Da mass increase) accounts for the total increase in mass. A similar increase in mass of 34 Da in the presence of GSH was seen for C91S-Prdx6, indicating that the non-conserved Cys91 residue does not participate in the reactions, leading to the change in mass. Thus, the C47 residue in the Prdx6 proteins is paradoxically converted to a sulfinic acid (Cys-SOOH) in the presence of GSH; this sulfinic acid is not reduced by the subsequent addition of TCEP ([Figure 2](#antioxidants-08-00004-f002){ref-type="fig"}B).
The formation of a sulfinic acid at Cys47 of Prdx6 was confirmed by alkylation of the cysteine residues in Prdx6 with iodoacetamide, followed by denaturation, reduction, and alkylation with acrylamide, trypsin digestion, and then analysis by liquid chromatography-tandem mass spectrometry (LC-MS/MS) in order to determine the levels of reduced, accessible cysteine (see [Supplemental Figure S2](#app1-antioxidants-08-00004){ref-type="app"}). The presence of a sulfinic acid in the C47-containing digestion product of Prdx6 (i.e., the peptide DFTPVCTTELGR) was observed in the presence of GSH (with or without TCEP), but sulfinic acid was not observed in the peptide containing C91 (i.e., the peptide DINAYNCEEPTEK).
3.4. GSH Induces Changes in Tryptophan Fluorescence and Circular Dichroism (CD) in hPrdx6 {#sec3dot4-antioxidants-08-00004}
-----------------------------------------------------------------------------------------
We examined potential changes in Prdx6 structure following the addition of GSH. Prdx6 has three tryptophan residues, two of which are in close proximity to the enzymatically active sites for peroxidase (C47) and PLA~2~ (S32) activities \[[@B12-antioxidants-08-00004]\]. A change in tryptophan fluorescence is presumed to reflect a change in the local environment of tryptophan residues in the protein \[[@B29-antioxidants-08-00004]\]. The addition of GSH resulted in a decrease of fluorescence intensity ([Figure 3](#antioxidants-08-00004-f003){ref-type="fig"}), which is compatible with the exposure of a tryptophan residue to a more polar environment \[[@B30-antioxidants-08-00004]\]. There was no change in Trp fluorescence following the addition of GSSG.
The far UV circular dichroism (CD) profile of Prdx6 was also used to evaluate the effect of GSH addition on protein structure. Ignoring the noise below 215 nm, the results show that GSH, but not GSSG, induces a change in signal in the 210--240 nm region of the CD spectrum ([Figure 4](#antioxidants-08-00004-f004){ref-type="fig"}). This change reflects a decrease in the α-helical content of Prdx6.
3.5. PLA~2~ Activity with Protein Modification {#sec3dot5-antioxidants-08-00004}
----------------------------------------------
The PLA~2~ activity of Prdx6 at pH 7.4 was increased by the oxidation of the protein through its exposure to H~2~O~2~ ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}). This increase in activity at pH 7.4 with oxidized Prdx6 was about double the GSH-induced increase that was seen with non-oxidized Prdx6. Activity at pH 4 with oxidized Prdx6 was 102 ± 2 (*n* = 3), similar to the value at pH 7.4. There was no effect of GSH on aiPLA~2~ activity with either the oxidized protein or the oxidized substrate at either pH.
As we have shown previously, phosphorylation of Prdx6 also results in a marked increase in its aiPLA~2~ activity \[[@B22-antioxidants-08-00004]\]. The increase in activity at pH 4 was \~20-fold and, fold-wise, was considerably greater at pH 7.4 ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}), so that the resultant activities of phosphorylated Prdx6 were similar at acidic and basic pHs \[[@B21-antioxidants-08-00004],[@B23-antioxidants-08-00004]\]. This increased activity with Prdx6 phosphorylation greatly exceeded the increase in activity with Prdx6 oxidation.
4. Discussion {#sec4-antioxidants-08-00004}
=============
The major finding of the present study was that the addition of GSH to the medium that was used for assay of the PLA~2~ activity of Prdx6 resulted in: a) a physical change of protein structure, as shown by tryptophan fluorescence and far UV circular dichroism; b) the irreversible oxidation of the protein to the sulfinic acid (as indicated by MS analysis); and, c) an increase of its PLA~2~ activity at neutral pH. The increased aiPLA~2~ activity in the presence of GSH was inhibited by MJ33, a mimic of the enzymatic transition state for some PLA~2~ enzymes and a known inhibitor of aiPLA~2~ activity \[[@B31-antioxidants-08-00004],[@B32-antioxidants-08-00004]\]. These results provide evidence that the catalytic mechanisms for stimulated PLA~2~ activity at pH 7 and basal activity at pH 4 are similar. Stimulation aiPLA~2~ activity at pH 7 appeared to be specific for GSH, since other sulfhydryls (GSSG, DTT, TCEP) were ineffective. Based on these results that confirmed a previous report \[[@B24-antioxidants-08-00004]\], we conclude that GSH exerts its stimulatory effect on aiPLA~2~ activity through oxidation of the protein (Prdx6).
Although this effect of GSH to promote Prdx6 oxidation initially was surprising, since GSH is a mild reductant, further consideration led us to propose the following mechanism. Based on the MS analysis, recombinant Prdx6 after purification and storage was oxidized to the sulfenylamide, a form that has been shown to "protect" some proteins against further auto-oxidation \[[@B33-antioxidants-08-00004]\]; in the case of Prdx6, the auto-oxidizable site is C47, the active site for peroxidase activity. The Prdx6 sulfenylamide is formed by the oxidation of the thiol to the sulfenyl that rapidly reacts with an adjacent amine in the protein to form the stable sulfenylamide \[[@B25-antioxidants-08-00004]\]; the specific amino group that was involved in this reaction has not yet been determined. The mechanism for auto-oxidation has been shown to reflect primarily the presence of trace metals \[[@B28-antioxidants-08-00004],[@B34-antioxidants-08-00004],[@B35-antioxidants-08-00004]\]. In the presence of TCEP, the sulfenylamide state of the protein can be reversed first to the sulfenic and then reduced to the sulfhydryl; auto-oxidation of the protein under these conditions is prevented by the presence of the strong reductant (TCEP). GSH also is able to reverse sulfenylamide formation, as shown in previous studies with the enzyme protein tyrosine phosphatase B1 (PTPB1) \[[@B36-antioxidants-08-00004]\]. This reversal reaction likely proceeds through the formation of a disulfide between GSH and the sulfenylamide to form a glutathionylated protein followed by hydrolysis to regenerate the sulfenyl \[[@B28-antioxidants-08-00004]\]. However, GSH is a relatively weak reductant that cannot, by itself, reduce the sulfenic \[[@B3-antioxidants-08-00004],[@B10-antioxidants-08-00004]\], and therefore cannot protect the protein against further oxidation to the sulfinic (and possibly sulfonic) form. Under physiological conditions *in vivo*, catalysis by glutathione S-transferase (GST) in the presence of GSH regenerates the sulfhydryl, thereby preventing its irreversible oxidation \[[@B10-antioxidants-08-00004],[@B37-antioxidants-08-00004],[@B38-antioxidants-08-00004]\]. However, with severe oxidative stress or with GST deficiency, oxidation of Prdx6 could occur *in vivo,* resulting in a significant increase of intracellular aiPLA~2~ activity. It is important to note that this increase in aiPLA~2~ activity would be accompanied by loss of the GSH peroxidase activity of Prdx6 due to irreversible inactivation of the catalytic Cys 47.
The reactions of Prdx6 that occur *in vitro* in the absence of GST are shown in [Figure 5](#antioxidants-08-00004-f005){ref-type="fig"}. To summarize these reactions, exposure of the protein to air during the isolation and storage of recombinant Prdx6 results in the oxidation of the thiol to an unstable sulfenylated protein that rapidly dehydrates to a stable sulfenylamide (reactions 1 and 2). The oxidant (shown as H~2~O~2~ in the figure) is generated from atmospheric O~2~ in the presence of trace metals \[[@B27-antioxidants-08-00004],[@B34-antioxidants-08-00004],[@B35-antioxidants-08-00004]\]. Following the addition of GSH, there is regeneration of the sulfenic protein (reactions 3 and 4). In the presence of GSH, the rate of H~2~O~2~ generation through metal-catalyzed auto-oxidation is greatly increased \[[@B34-antioxidants-08-00004]\], and the sulfenic can be oxidized to the sulfinic form (reaction 5). The irreversible formation of the sulfinic in the presence of GSH pulls the reaction cycle toward oxidation. Thus, the end result is Prdx6 over-oxidation following the addition of GSH to the protein *in vitro*, in the absence of GST. Under physiologic conditions in the presence of cellular GST plus GSH, the heterodimerization of Prdx6 with GST leads to a reduction of Prdx6 to the native sulfhydryl (reactions not shown) \[[@B26-antioxidants-08-00004]\].
Similarly to the effect of GSH addition to the assay medium, oxidation of the protein by treatment with H~2~O~2~ prior to the assay also resulted in increased PLA~2~ activity ([Table 1](#antioxidants-08-00004-t001){ref-type="table"}). A similar result was reported previously that used mutagenesis to show that the increase of activity with H~2~O~2~ treatment required the presence of C47 in Prdx6 \[[@B24-antioxidants-08-00004]\]. aiPLA~2~ activity in the present study also was increased in the presence of a peroxidized lipid substrate (PLPCOOH), confirming our previous report \[[@B23-antioxidants-08-00004]\]. Thus, several different reactions demonstrate a similar mechanism to stimulate aiPLA~2~ activity *in vitro*, namely the oxidation of the conserved Cys of Prdx6. These oxidative reactions include: (1) the presence of GSH; (2) pretreatment of the enzyme with H~2~O~2~; or, (3) use of PLPCOOH as the liposomal substrate. Each of these assay conditions in the absence of GST would be expected to result in the oxidation of the C47 of Prdx6.
These findings reflect on the catalytic cycle for peroxidase activity. The reaction of H~2~O~2~ or a hydroperoxide substrate with Prdx6 will result in the formation of a sulfenic acid intermediate as the first step in the reduction of the oxidized lipid through peroxidase catalysis. In the absence of GST, the reaction cycle cannot be completed and the sulfenyl protein is susceptible to further oxidation, as shown previously using specific antibodies for the over-oxidized state of Prdx6 \[[@B23-antioxidants-08-00004]\]. The presence of GST plus GSH prevents over-oxidation and it returns Prdx6 to its native sulhydryl form, thereby completing the peroxidase reaction cycle and preserving its peroxidase function.
Nevertheless, how does Cys oxidation affect aiPLA~2~ activity since the C47 residue in Prdx6, while being crucial for its peroxidatic function, has no direct role in the PLA~2~ reaction? Although our previous studies have shown that reduced (non-oxidized) phospholipid substrate does not bind at pH 7, it does bind to Prdx6 at pH 4; on the other hand, oxidized phospholipids bind to Prdx6 equally at acidic and neutral pH \[[@B23-antioxidants-08-00004]\]. C47 is present as a thiolate at pH 7, but it is protonated to a sulfhydryl at pH 4, possibly resulting in changes in structure that facilitate the binding of substrate \[[@B2-antioxidants-08-00004]\]. These binding data directly correlate with aiPLA~2~ activity. Binding of reduced substrate at pH 4 results in PLA~2~ activity, while binding of an oxidized phospholipid at pH 7 can result in PLA~2~ as well as peroxidase activity.
Crystal structure analysis has provided some insights into substrate binding \[[@B39-antioxidants-08-00004]\]. The crystal structure of the sulfenic protein shows that the active site for Prdx6 binding (and aiPLA~2~-mediated catalysis) is on the protein surface and it is connected to the peroxidatic site by a narrow pocket. We have postulated that the oxidized phospholipid substrate is positioned for activity by binding of the head group to the protein surface and the insertion of the oxidized *sn*-2 fatty acid chain inside of the pocket \[[@B12-antioxidants-08-00004]\]. On the other hand, crystal structural analysis of oxidized (sulfinic) Prdx6 indicates almost 10 angstroms distance between the H26 and S32 sites that were proposed for phospholipid binding. Because of this relatively long distance, it was suggested that the phospholipid actually may bind to the flat surface across the PLA~2~ active site and the peroxidatic active site of the other monomer \[[@B40-antioxidants-08-00004]\]. However, several factors may modify the protein structure and its interaction with substrate. For example, over-oxidation of C47 to the sulfinic alters the Prdx6 secondary structure, as shown by CD analysis in this report. Further, the oxidation of C47 results in decreased affinity for homodimerization of the protein, resulting in an increased tendency for monomer formation \[[@B26-antioxidants-08-00004]\]. Finally, phosphorylation (at T177) may lead to structural changes that alter the distance between the crucial amino acids \[[@B22-antioxidants-08-00004]\]. We have shown, while using a zero-length chemical cross-linking and homology methodology, that several regions of reduced human PRDX6 are in a substantially different conformation from that shown for the crystal structure of the peroxidase catalytic intermediate \[[@B41-antioxidants-08-00004]\]. Thus, the protein shows considerable plasticity and it can change conformation depending on many variables that could affect Prdx6-substrate interactions. The variable states of Prdx6 require further study before making definitive conclusions concerning the site and mechanism for binding of the phospholipid substrate to the enzyme.
5. Conclusions {#sec5-antioxidants-08-00004}
==============
Isolated recombinant human Prdx6 undergoing auto-oxidation to the sulfenic can form a sulfenylamide, a protected state that prevents auto-oxidation. The addition of GSH results in the regeneration of the sulfenic form that can be irreversibly oxidized to the sulfinic by oxidants that are produced through metal catalyzed auto-oxidation; the generation of oxidants through auto-oxidation occurs at a significantly increased rate in the presence of GSH. Over-oxidation of Prdx6 to the sulfinic results in a marked increase of its aiPLA~2~ activity at neutral pH, as shown previously \[[@B24-antioxidants-08-00004]\] and in this manuscript.
We thank Yefim Manevich for the original observation suggesting lipase activity during assay of the peroxidase, Shampa Chatterjee and Haitao Li for helpful discussions, Tea Shuvaeva for assistance with the preparation of recombinant Prdx6, and Dawn Williams for assistance with manuscript preparation. The current affiliation of S.Z. is St. Jude Children's Research Hospital, Memphis, TN.
The following are available online at <http://www.mdpi.com/2076-3921/8/1/4/s1>, Figure S1: Purification of codon optimized recombinant human Prdx6 with an ion-exchange (DEAE cellulose) column, Figure S2: Relative quantitation based on the intensity of modified peptides from LC-MS/MS data analysis after treatment with GSH (+/−TCEP).
######
Click here for additional data file.
Formulated research plan (A.B.F., D.W.S.); generated recombinant Prdx6 (S.I.F., S.Z.); performed experiments (C.D., S.Z., S.H.); interpreted experimental results (A.B.F., S.Z., D.W.S., S.H., H.J.F.); wrote the manuscript (A.B.F., S.Z.).
This research was supported by grant R01-HL102016 (P.I., A.B.F.) and P30-CA10815 (core Grant to the Wistar Institute), both from the National Institutes of Health of the United States.
S.I.F. and A.B.F. have a patent application pending for a peptide inhibitor of peroxiredoxin 6 PLA~2~ activity and have part ownership of a start-up company to promote clinical use of the peptide inhibitor.
{#antioxidants-08-00004-f001}
![Electrospray ionization mass spectrometry (ESI MS) of wild type and C91S Prdx6. (**A**) upper 4 panels). Effect of tris (2-carboxyethyl) phosphine (TCEP) on protein mass. The addition of TCEP resulted in an increase of 2 Da in molecular mass of both wild type and C47S-Prdx6. (**B**) lower 4 panels). Effect of glutathione (GSH) +/- TCEP on protein mass. The addition of GSH resulted in a mass increase of 34 Da; there was no further effect with addition of TCEP in the presence of GSH. TBS. tris-buffered saline. **A** and **B** were part of the same experiment; **A**, but not **B**, was published previously \[[@B26-antioxidants-08-00004]\] and it is reprinted with permission.](antioxidants-08-00004-g002){#antioxidants-08-00004-f002}
{#antioxidants-08-00004-f003}
{#antioxidants-08-00004-f004}
![Reactions of Prdx6 with GSH *in vitro*. Reactions 1 and 2 occur during isolation of Prdx6 in the absence of GSH and GST; auto-oxidation of the protein by oxidant generation (shown here as HOOH) in the presence of trace metals leads to the formation of the Prdx6 sulfenic acid (reaction 1) and then dehydration to form the sulfenylamide (reaction 2). Reaction 2 is rapid when compared to the relatively slow rate of H~2~O~2~ formation (reaction 5 in the absence of GSH). Reactions 3 and 4 occur *in vitro* following the addition of GSH in the absence of GST. The sulfenylamide is reversed through glutathionylation of the protein (reaction 3), followed by reformation of the sulfenic acid (reaction 4). The rate of H~2~O~2~ generation is greatly enhanced in the presence of GSH \[[@B34-antioxidants-08-00004]\], resulting in the oxidation of the sulfenic to the sulfinic (reaction 5 in the presence of GSH).](antioxidants-08-00004-g005){#antioxidants-08-00004-f005}
antioxidants-08-00004-t001_Table 1
######
Effect of protein and substrate modifications on phospholipase A~2~ (PLA~2~) activity of peroxiredoxin 6 (Prdx6) at pH 7.4.
--------------------------------------------------------------------------------
Condition PLA~2~ Activity at pH 7.4\
nmol/min/mg
--------------------------------- ---------------------------- -----------------
−GSH +GSH
Control 0.3 ± 0.06 (6) 51 ± 1 (6)
+MJ33 \_\_\_ 19 ± 2\* (3)
+BEL \_\_\_ 50 ± 1 (4)
+GSSG 3.2 ± 2.1 (3) \_\_
+DTT 2.0 ± 1.4 (3) \_\_
+TCEP 0.5 ± 0.1 (3) \_\_
Anaerobic 0.2 ± 0.1 (4) 0.2 ± 0.1\* (4)
Liposomes with PLPC^†^ 0.3 ± 0.02 (6) 48 ± 1 (6)
Oxidized liposomes with PLPC^†^ 100 ± 1\* (3) 100 ± 1\* (3)
Phosphorylated Prdx6 1220 ± 6\* (3) 1150 ± 3\* (3)
Oxidized Prdx6 100 ± 3\* (3) 99 ± 1\* (3)
--------------------------------------------------------------------------------
Human Prdx6 (2 µg) was pre-incubated with reagents for 15 min and then PLA2 activity was measured at pH 7.4 in Ca^2+^ free buffer. Substrate was 3H-1,2-Bis palmitoyl-sn-glycero-3-phosphocholine (DPPC) in mixed unilamellar liposomes. When added, glutathione (GSH) and other sulfhydryls were at 5 mM, MJ33 at 3 mol% of lipid, and BEL at 0.1 mM. Prdx6 was oxidized with H2O2 or was phosphorylated with Erk2. Values are mean ± SE for the number of experiments indicated in parentheses. \* Significantly different (*p* \< 0.05) from corresponding control (plus or minus GSH). † Liposomes with 1-palmitoyl, 2-linoleoyl, sn-glycero-3-phosphocholine (PLPC) replacing egg phosphatidylcholine (PC).
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
Epithelial ovarian cancer (EOC) is the leading cause of gynecological cancer death in the U.S. Approximately 22,000 women are diagnosed annually and 15,000 die from the disease \[[@B1]\]. Most women are diagnosed only after peritoneal dissemination has occurred. The standard treatment for patients with EOC is cytoreductive surgery (CRS) followed by intravenous Pt-taxane chemotherapy \[[@B2]\]. Even though initially effective, relapse from residual disease and/or drug resistant cancer reduces the 5-year survival rate to about 20% \[[@B3]\]. Despite research efforts to improve on Pt-based chemotherapy, or to develop new drugs against EOC, most patients still die from metastatic disease. Since metastatic EOC is usually confined in the peritoneal cavity, it makes theoretical sense to deliver chemotherapy intraperitoneally rather than intravenously since higher levels of drug can be delivered to the disease site by that route \[[@B4],[@B5]\]. In response to three large randomized clinical trials showing benefit to incorporating intraperitoneal (IP) delivery in EOC, the National Cancer Institute issued a clinical announcement recommending that patients with small volume disease at the end of frontline surgery be offered the chance of receiving IP chemotherapy \[[@B6]\]. Adding hyperthermia to chemotherapy agents delivered intraperitoneally (HIPEC) theoretically could improve outcome \[[@B7]-[@B9]\].
Cisplatin is a DNA damaging chemotherapeutic used to treat solid tumors including EOC. However, resistance to cisplatin limits clinical success. Mechanisms of cisplatin resistance are multi-factorial and include reduced cellular drug accumulation, enhanced drug metabolism by glutathionylation and export by multidrug resistance proteins, enhanced DNA damage tolerance and DNA repair \[[@B10]\]. Since Pt-containing chemotherapy drugs remain the major weapon against EOC, improving their efficacy could have a great impact on mortality. The combination of hyperthermia with cisplatin has been reported for the treatment of EOC \[[@B11]\]. Hyperthermia is tumoricidal alone \[[@B12]\] and has been shown to enhance cisplatin inhibition of peritoneal tumor growth by increasing tumor Pt accumulation \[[@B13]\]. Arsenic trioxide (As~2~O~3~), an FDA approved drug for the treatment of all-trans-retinoic acid-resistant acute promyelocytic leukemia \[[@B14]\] has the potential to sensitize tumors to cisplatin \[[@B15],[@B16]\]. Combination chemotherapy studies demonstrate that arsenic sensitizes cancer cells to hyperthermia, radiation, cisplatin, adriamycin, doxorubicin, and etoposide \[[@B16]-[@B19]\]. *In vitro*studies demonstrate that trivalent arsenic (As^3+^administered as arsenic trioxide \[As~2~O~3~, Trisenox^®^\] or sodium arsenite \[NaAsO~2~\]) induces apoptosis in multiple types of cancer cells including cervical, melanoma, gastric, colon, pancreatic, lung, prostate and ovarian cancer cell lines \[[@B20]-[@B23]\]. *In vivo*studies also show that arsenic inhibits the growth of orthotopic metastatic prostate cancer and peritoneal metastatic ovarian cancer \[[@B24],[@B25]\]. The mechanism of arsenic-induced cell death *in vitro*is suggested to include formation of oxidative DNA damage \[[@B26]\], activation of the Fas pathway \[[@B27]\], inhibition of DNA repair \[[@B28],[@B29]\], and causation of mitotic arrest and induction of apoptosis in the mitotic cells \[[@B20],[@B21]\].
As^3+^has biological effects similar to those of both cisplatin and hyperthermia. Like cisplatin it is detoxified by glutathionylation and exported by multidrug resistant family transport pumps \[[@B30],[@B31]\], suggesting a potential for competition for the detoxification pathway if arsenic and cisplatin are used in combination. This competition might enhance cisplatin accumulation in cells. Like hyperthermia, As^3+^induces stress response proteins and causes mitotic catastrophe \[[@B21]\]. These actions make arsenic a potentially effective agent to augment hyperthermia enhancement of cisplatin-induced cell death.
The goal of this study is to determine how sodium arsenite and hyperthermia modulate mechanisms of cisplatin resistance *in vivo*. We developed murine models of HIPEC treatment and metastatic human EOC to investigate if NaAsO~2~and hyperthermia alter the expression of DNA repair proteins and tumor platinum levels. We show that NaAsO~2~and hyperthermia either as single agents or in combination reverse key DNA repair protein responses to cisplatin responsible for cisplatin resistance and also enhanced tumor Pt uptake suggesting decreased Pt detoxification.
Methods
=======
Chemicals
---------
Cisplatin and sodium arsenite were purchased from Sigma-Aldrich (St. Louis, MO). Stock solutions (cisplatin 1 mg/mL in 1X PBS and NaAsO~2~13 mg/mL in water) were prepared freshly on the day of treatment and filter sterilized (0.22 μm) prior to use.
Cells and cell culture
----------------------
Cisplatin-resistant (A2780/CP70) human ovarian cancer cells were the kind gift of Dr. Eddie Reed. Cells were maintained in RPMI 1640 medium containing 10% fetal bovine serum, 100 μg/mL penicillin/streptomycin, 2 mM L-glutamine and 0.2 units/mL insulin. Cells were cultured in an atmosphere of 95% humidity and 5% CO~2~at 37°C. Cells were passaged twice weekly and replated at a density of 1 × 10^6^cells/150 mm dish.
Animals
-------
Female NCr athymic nude mice (7 - 9 weeks old), were purchased from Taconic (Cambridge City, IN). Animals were kept in a temperature-controlled room on a 12 h light-dark schedule. The animals were maintained in cages with paper filter covers under controlled atmospheric conditions. Cages, covers, bedding, food, and water were changed and sterilized weekly. Animals were fed autoclaved animal chow diet and water. All procedures were performed under sterile conditions. This experiment was approved by the Institutional Animal Care and Use Committee of the University of Louisville in an AALAC approved facility in accordance with all regulatory guidelines.
Establishment of intraperitoneal metastatic ovarian tumors in mice
------------------------------------------------------------------
A2780/CP70 cell suspension (1 × 10^6^cells in 500 μL of serum-free RPMI 1640 media) was injected into the peritoneum of anesthetized mice using an 18-gauge needle. The needle was flushed with 500 μL physiological saline. The abdomen of injected animals was massaged to ensure even distribution of cells. By 3 - 4 weeks after injection, the mice had developed multiple small disseminated IP tumors (1 - 7 mm) (Figure [1](#F1){ref-type="fig"}). Tumors were monitored by microCT scanning in the Brown Cancer Center Small Animal Imaging Facility.
Intraperitoneal chemotherapy
----------------------------
Tumor-bearing mice were anesthetized with 3% isoflurane in an inhalation chamber and maintained on 1% isoflurane during surgery. Incisions (\~0.5 cm) were made on both sides of the lower abdominal wall allowing entry into the peritoneal cavity (Figure [2](#F2){ref-type="fig"}). Inflow and outflow tubes were inserted into the peritoneal cavity and secured with skin sutures. The tubes were connected to a bag containing 100 mL normal saline with added cisplatin (3 mg/kg body weight (BW)) ± sodium arsenite (26 mg/kg BW) and cefazolin (0.01 mg/mL). (The dose of cisplatin used for this study was determined from human dose of cisplatin (100 mg/m^2^) administered intravenously to a 70 kg (body surface area = 1.87 m^2^) \[[@B32]\] cancer patient and sodium arsenite dose was calculated from a single daily dose of Trisenox (0.15 mg/kg/day) administered intravenously to a 70 kg acute promyelocytic leukemia patient. The underlying assumption in the calculations is that the drugs are mixed in 2 L saline solution for HIPEC therapy). The saline bag was submerged in a water bath to maintain the perfusate temperature at either 37 or 43°C. Perfusion was performed at a rate of 3 mL/min for 60 min using a Masterflex pump (Cole-Palmer Instrument Co, Cat \# 07524-50). The inflow and outflow temperatures were monitored by thermocouple probes with temperature maintained within 1°C. The core temperature of the animals was monitored using an anal temperature probe and maintained using a heating pad and heat lamp. After 60 min perfusion, most of the perfusate in the peritoneum was sucked out using sterile cotton balls with a light abdominal massage. Wounds were sutured closed and animals were injected intraperitoneally with 1 mL physiological saline containing 0.01 mg ketoprofen for pain. Mice were kept in warm cages (single mouse/cage) and monitored for recovery and discomfort. Immediately (0 h) and 24 h after perfusion, mice were euthanized and tumors, kidneys, liver, spleen, heart and brain were dissected and snap frozen in liquid nitrogen and stored at -80°C until use.
Western blot analysis
---------------------
Tumors of \~ 3-5 mm in diameter were homogenized in protein lysis solution (1 M Tris-HCl pH 7.4, 0.5 M EDTA, 10% sodium dodecyl sulfate, 180 μg/mL phenylmethylsulphonylfluoride) using a tissue grinder. After removal of debris by centrifugation (45 min, 14,000 x g), total protein concentration in supernatant was determined by bicinchoninic acid (BCA) method according to manufacturer\'s instructions (Pierce, Rockford, IL, micro-well plate protocol) \[[@B33]\]. Fifteen μg protein samples were resolved by SDS-polyacrylamide gel electrophoresis and electro-transferred to nitrocellulose membranes. Membranes were probed with antibodies to XPA (Neomarkers, MS-650-P1, dilution 1:1000), XPC (Novus, \# ab6264, dilution 1:10,000), GAPDH (Sigma, \# A 5441, dilution 1:10,000), p53 (DO-1, Cell Signaling Technology, \# 9284, dilution 1:1000), MSH2 (Santa Cruz, \# SC-494, dilution 1:1000), and ERCC1 (Santa Cruz, \# SC-10785, dilution 1:1000). Secondary antibodies (rabbit anti-mouse IgG, \# 81-6120 or goat anti-rabbit, \# 81-6120, dilutions 1:2500) conjugated to horseradish peroxidase (Zymed Laboratories, Inc. South San Francisco, CA) were bound to primary antibodies and protein bands detected using enhanced chemiluminescence (ECL) substrate (Pierce, Rockford, IL). GAPDH was used as the loading control. Films were scanned with a Molecular Dynamics Personal Densitometer SI (Molecular Dynamics, Sunnyvale, CA) and analyzed with ImageQuaNT software (Molecular Dynamics) to determine band density.
ICP-MS analysis
---------------
Samples of tumor homogenates were lyphophilized using Heto vacuum centrifuge (ATR, Laurel, MD) and 350 μL concentrated nitric acid was added to each sample. Wet weight of brain, heart, spleen, liver and kidney was recorded and concentrated nitric acid (350 - 500 μL) was added to samples. Samples were predigested overnight, and then 100 μL of each dissolved sample was transferred into 10 mL acid washed microwavable digestion tubes (triplicate for each sample). The samples were microwave-digested at 150°C for 10 min using an automated focused beam microwave digestion system (Explorer™, CEM, Matthews, NC, USA). After digestion, 1.9 mL of 18 Mohm H~2~O containing 10 ppb internal standard (SPEX CertiPrep, Metuchen, NJ) was added into every sample to give final 5% nitric acid and ICP-MS analyses was performed using Thermo X Series II ICP-MS (Thermo Fisher Scientific, Waltham, MA) at the University of Louisville Center for Regulatory and Environmental Analytical Metabolomics facility. Concentrated nitric acid was processed similarly as blank. Platinum standard (SPEX CertiPrep, Metuchen, NJ) was used to generate a standard curve. Platinum and arsenic levels in tumors and tissues were expressed as ng metal/mg protein and ng metal/mg wet weight respectively. Results are presented as the means of three ICP-MS determinations for each data point ± SD from 3 individual mice.
Immunocytochemistry
-------------------
Cells (1 × 10^5^) were plated on poly-D-lysine coated coverslips (BD Biosciences) in a 24-well plate and allowed to acclimate for 24 h. Cells were then treated with 40 μM cisplatin for 1 h. After treatment, cells were washed twice with PBS and incubated in drug-free media for 24 h. Cells were fixed in ice-cold acetone for 10 min at room temperature and washed twice with ice cold PBS and samples incubated for 10 min with PBS containing 0.25% Triton X-100 (PBST). Cells were then washed with PBS three times for 5 min and incubated in 3% hydrogen peroxide for 30 min to quench endogenous peroxidase. Cells were washed three times with PBS and incubated in 1% BSA in PBST for 30 min to block unspecific binding of the antibodies. Cells were incubated overnight at 4°C in primary antibodies (1:200 dilution in PBST containing 1% BSA). The primary antibodies used were XPA (Neomarkers, MS-650-P1), XPC (H-300, SantaCruz Biotechnology, \# sc-30156), p53 (DO-1, Cell Signaling Technology, \# 9284), MSH2 (Santa Cruz, \# SC-494) and ERCC1 (Santa Cruz, \# SC-10785). After incubation, the primary antibody solution was decanted and cells were washed three times with PBS for 5 min each wash. Cells were incubated with secondary antibodies (rabbit anti-mouse IgG, \# 81-6120 or goat anti-rabbit, \# 81-6120, dilution 1:200 in PBST containing 1% BSA) conjugated to horseradish peroxidase (Zymed Laboratories, Inc. South San Francisco, CA) for 1 h at room temperature. Secondary antibody solution was decanted and cells were washed three times with PBS for 5 min. Cells were stained with 3,3\'-diaminobenzidine (DAB) substrate solution by incubating cells in 200 μL premixed DAB solution (mix 30 μL (one drop) of the DAB liquid chromogen solution to 2 mL of the DAB liquid buffer solution (Sigma, \# D 3939)) for 10 min. DAB solution was removed and cells rinsed briefly with PBS. Cells were counterstained with 20% Wright Giemsa solution for 1 min. Coverslips were mounted on microscope slides using a drop of permount mounting medium. Slides were viewed under a Nikon Eclipse E600 Microscope (Fryer Company Inc, Scientific Instruments, Cincinnati, OH 45240) and pictures taken using MetaMorph software (Universal Imaging Corporation). DAB-positive cells were counted per 1000 cells using MetaMorph software.
Statistical analysis
--------------------
Statistical analyses were performed using wilcoxon rank sum test with significance set as p \< 0.05, n ≥ 3.
Results
=======
Murine intraperitoneal chemotherapy model
-----------------------------------------
Multiple disseminated tumors were established in the peritoneal cavity of nude mice as described in Materials and Methods. Mice were scanned using microCT scan to determine the location and estimate the size of tumors (Figure [1A](#F1){ref-type="fig"}). This was confirmed upon necropsy (Figure [1B](#F1){ref-type="fig"}). Tumor bearing mice were treated by peritoneal lavage for 1 h with cisplatin ± sodium arsenite at 37°C (normothermia) or 43°C (hyperthermia) (Figures [2A](#F2){ref-type="fig"} and [2B](#F2){ref-type="fig"}) as described in Materials and Methods. During treatment, the required inflow temperature was reached within 2-5 min after the start of perfusion. Inflow, outflow and rectal temperatures were recorded every 15 min and remained stable within 1°C throughout the 60 min perfusion (Table [1](#T1){ref-type="table"}).
{#F1}
{#F2}
######
Inflow, outflow and body temperatures of mouse during intraperitoneal perfusion
[Inflow Temperature]{.ul} [Outflow Temperature]{.ul} [Body Temperature]{.ul}
--------------------------- ---------------------------- -------------------------
37.4 ±1.1°C 36.4 ± 0.8°C 35.5 ± 1.0°C
43.0 ± 0.7°C 39.7 ± 0.6°C 36.3 ± 2.1°C
Mice were perfused for 1 h with cisplatin (CP/37; CP/43) or cisplatin + NaAsO2 (CPA/37; CPA/43) at 37 or 43°C respectively. Inflow, outflow and body temperatures were recorded every 15 min. Data are presented as means ± SD of readings taken from five mice.
Platinum and arsenic accumulation and retention in metastatic tumors
--------------------------------------------------------------------
We determined Pt and arsenic accumulation in tumors immediately (0 h) and 24 h after perfusion using ICP-MS. Pt and arsenic accumulated in tumors during treatment (0 h) and generally decreased after treatment (24 h), compared with the untreated control (Figure [3](#F3){ref-type="fig"}). Co-treatment with NaAsO~2~and cisplatin at 37°C (CPA/37) or 43°C (CPA/43) caused significantly more Pt to accumulate in tumors. By 24 h after perfusion, tumor Pt levels for CPA/37 and CPA/43 treatment conditions decreased to levels similar to CP/37. Hyperthermia did not increase tumor Pt levels nor alter Pt retention in tumors 24 h after treatment. More arsenic initially accumulated in tumors when co-treated with cisplatin and NaAsO~2~at 37°C (CPA/37) than with hyperthermia treatment (CPA/43). Arsenic decreased to similar levels at 24 h.
{#F3}
Effect of cisplatin, arsenic and hyperthermia on DNA repair protein expression
------------------------------------------------------------------------------
Cisplatin causes bulky DNA damage that is repaired mostly by the nucleotide excision repair system (NER). Cellular response to cisplatin-DNA damage involves the induction of DNA repair proteins to initiate DNA repair \[[@B10]\]. We determined if NaAsO~2~and hyperthermia modulated the expression of XPC, a platinum-DNA damage recognition protein in global genome repair (GGR) \[[@B34]\] subpathway of NER, and of ERCC1 and XPA, downstream NER proteins that have been implicated in cisplatin resistance \[[@B35]\]. We also determined the expression of p53, which is involved in the activation of the GGR pathway by transcriptionally activating XPC \[[@B36]\]. In addition to NER, decreased mismatch repair (MMR) has been implicated in cisplatin resistance \[[@B37],[@B38]\]. Thus, we also investigated the expression of MSH2, an important MMR DNA damage recognition protein. Western blot analysis of p53, XPC, XPA, ERCC1 and MSH2 revealed mouse-to-mouse and tumor-to-tumor variabilities (Figure [4A](#F4){ref-type="fig"}). Some tumors failed to express the protein of interest while others either expressed high, moderate or very low levels of the proteins. We determined band intensities for the expressed proteins by scanning the films using a Molecular Dynamics Personal Densitometer SI (Molecular Dynamics, Sunnyvale, CA) and analyzing bands of interest using ImageQuaNT software (Molecular Dynamics). Each protein value was normalized to its respective GAPDH (loading control) value. Data were further normalized to untreated control (Figure [4B](#F4){ref-type="fig"}). Tumors that failed to express the protein of interest were not considered in the densitometry analyses. P53 (Figure [4B](#F4){ref-type="fig"}, panel a) and XPC (Figure [4B](#F4){ref-type="fig"}, panel b) were significantly induced during treatment (0 h) by cisplatin at 37°C (CP/37) or 43°C (CP/43) and cisplatin plus arsenite at 43°C (CPA/43). P53 significantly decreased at 24 h after treatment with CPA/43 (Figure [4B](#F4){ref-type="fig"}, panel a). XPC decreased at 24 h after perfusion with both CP/43 and CPA/43 treatments (Figure [4B](#F4){ref-type="fig"}, panel b). P53 (Figure [4B](#F4){ref-type="fig"}, panel a) and XPC (Figure [4B](#F4){ref-type="fig"}, panel b) did not significantly increase during (0 h) and after (24 h) peritoneal lavage with NaAsO~2~and cisplatin co-treatment at 37°C (CPA/37). XPA (Figure [4B](#F4){ref-type="fig"}, panel c) was significantly induced during (0 h) and 24 h after perfusion with CP/37, CPA/37 and CPA/43 but not with CP/43. ERCC1 remained generally low for all treatment conditions except with CPA/37 (Figure [4B](#F4){ref-type="fig"}, panel d). The suppression of MSH2 by CP/37 and CP/43 treatments was not seen in tumors co-treated with arsenite (CPA/37, CPA/43) (Figure [4B](#F4){ref-type="fig"}, panel e).
{#F4}
Expression of P53, XPA and MSH2 in ovarian cancer cells
-------------------------------------------------------
Western blot determination of P53, XPC, XPA, ERCC1 and MSH2 in metastatic tumors revealed that some tumors failed to express p53 (6%), XPC (3%), XPA (8%), ERCC1 (40%) and MSH2 (9%). Failure to express these proteins could be an inherent feature of the cells that were used to establish the tumors or due to mutations and alteration of genes during tumor development that could result in lack of protein expression. We therefore performed immunocytochemical studies using A2780/CP70 cells to determine expression of P53, XPA and MSH2 in these cells (Figure [5A](#F5){ref-type="fig"}). Immunocytochemistry data revealed that 25% of cells do not express p53 as evident by lack of 3,3\'-diaminobenzidine (DAB) brown staining and \~3% and 60% of cells did not stain positive for XPA and MSH2 respectively (Figure [5B](#F5){ref-type="fig"}). Full-length western blots for XPC and ERCC1 had several non-specific bands in addition to the band of interest (data not shown) making it impossible to perform immunocytochemistry with specificity for these proteins.
{#F5}
Platinum and arsenic biodistribution in somatic tissues
-------------------------------------------------------
The clinical use of anticancer chemotherapeutic agents is limited by adverse toxicities. For cisplatin, these include toxicity to the kidney, peripheral nerves, liver, heart, bone marrow and brain \[[@B39],[@B40]\]. Clinical use of arsenic is known to cause liver, kidney and neurological damage, cardiovascular and gastro-intestinal toxicity, anemia and leucopenia \[[@B41]-[@B43]\]. Therefore, we determined cisplatin and arsenic accumulation in mouse tissues including kidney, liver, heart, spleen and brain (Figure [6A](#F6){ref-type="fig"} and [6B](#F6){ref-type="fig"}). Samples were prepared as described in Methods. During perfusion, platinum accumulated in all tissues examined regardless of the treatment condition, in the order: kidney \> liver = spleen \> heart \> brain. At 24 h after perfusion, significant decrease of platinum was observed in the kidney for all treatment conditions. The combination treatment (CPA/43) favored the removal of platinum from the liver, spleen and heart at 24 h after perfusion. Arsenic also significantly accumulated in all the tissues examined, in the order: liver \> kidney = spleen \> heart \> brain and it significantly decreased in all tissues by 24 h after perfusion.
{#F6}
Discussion
==========
Although the platinum analogues (cisplatin and carboplatin) are at the forefront of combination treatment for EOC, acquired or inherent resistance limits clinical success. In the current study, we used metastatic EOC xenograft in nude mice to investigate how NaAsO~2~and hyperthermia modulate response to cisplatin *in vivo*. We focused on three key mechanisms of cisplatin resistance: enhanced NER, diminished MMR and decreased Pt accumulation. Our data suggest that cisplatin induces resistant phenotype in metastatic tumors by inducing XPC and XPA and suppressing MSH2. Sodium arsenite alone or combined with hyperthermia inhibits mechanisms of cisplatin resistance by suppressing XPC induction, maintaining higher levels of MSH2 and increasing tumor uptake of cisplatin.
Decreased Pt accumulation is an important mechanism of cisplatin resistance. Hyperthermia has been reported to increase both cellular and DNA Pt levels*in vitro*. However, *in vivo*data remains controversial. Los et al used rats bearing metastatic colon cancer to show that hyperthermia suppressed tumor growth by increasing platinum accumulation in tumors \[[@B13]\]. Zeamari et al used a similar colon cancer xenograft model in rats and reported that hyperthermia did not increase tumor Pt levels \[[@B44]\]. Similar to Zeamari, we observed that hyperthermia does not increase Pt accumulation in tumors. The observed discrepancies with Los et al could be due to differences in how HIPEC was performed. Los et al injected hyperthermic cisplatin intraperitoneally; whereas we and Zeamari et al performed peritoneal lavage similar to what is done clinically. Unlike hyperthermia, we observed that NaAsO~2~at 37 or 43°C increased initial tumor Pt levels. Since arsenic and cisplatin are detoxified by glutathionylation and export by the multidrug resistant family proteins, potential competition for the detoxification/export pathways might have resulted in more Pt accumulating in the tumors when cisplatin is co-administered with sodium arsenite.
Cisplatin is a DNA damaging agent and p53 is implicated in platinum-DNA damage response \[[@B36]\]. P53 is frequently mutated in ovarian cancer \[[@B45]\]. The p53 phenotype of A2780/CP70 cells remains controversial. Some studies have demonstrated that A2780/CP70 cells have non-functional p53 \[[@B46],[@B47]\], while other studies have shown that these cells have wild type p53 \[[@B48],[@B49]\]. Our data indicate that A2780/CP70 cell population is heterogeneous: \~75% of cells express wild type p53 and \~25% are p53 null (Figure [5](#F5){ref-type="fig"}). In addition, 6% of the tumors derived from A2780/CP70 are p53 null (Figure [4A](#F4){ref-type="fig"}). Our *in vitro*data also demonstrate the induction of p53 target genes p21CIP1/WAF1, XPC and DDB2 in A2780/CP70 cells (data not shown), which strongly suggests that a large fraction of these cells have wild type p53. The observed heterogeneity might have resulted from mutations and alterations that occur during serial propagation of cells in culture leading to cell line drift \[[@B50]\]. The observed heterogeneity may impact response to chemotherapy and result in treatment failures because p53 wild type and null cells will respond differently to chemotherapy especially DNA damaging agents such as cisplatin. This heterogeneity explains why targeting master regulators such as p53 or AKT in cancer cells has not been successful \[[@B51],[@B52]\]. Therefore, combination chemotherapy such as cisplatin, sodium arsenite and hyperthermia with different mechanisms of action might be more beneficial than using a single drug to target a single protein or pathway.
Cisplatin predominantly forms intrastrand DNA crosslinks that are repaired by the nucleotide excision repair (NER) system. There are two subpathways of NER; transcription coupled repair (TCR) which removes damage from actively transcribing DNA and global genome repair (GGR) which removes lesions from the entire genome \[[@B53]\]. These two pathways differ only in the proteins that are involved in damage recognition. In TCR, CSA and CSB along with RNA pol II recognize damage, whereas in GGR, XPC and DDB2 are important for lesion recognition. XPC is actively involved in the recognition and initiation of cisplatin-DNA damage repair in GGR \[[@B34],[@B54]\]. Arsenic has been shown to inhibit NER by inhibiting XPC expression \[[@B29]\]. In the current study, we observed that P53 and XPC were induced by cisplatin. However, NaAsO~2~alone or in combination with hyperthermia prevented the induction of p53 and XPC by cisplatin (Figure [4B](#F4){ref-type="fig"}, panels a and b). Since p53 is known to transcriptionally induce XPC \[[@B36]\], our data suggest that NaAsO~2~± hyperthermia might be inhibiting p53, which in turn might be suppressing XPC induction. Suppression of XPC will potentially sensitize tumors to cisplatin. Our *in vitro*data suggest that inhibition of XPC using siRNA sensitizes ovarian cancer cells to cisplatin (data not shown). Therefore, the suppression of XPC could potentially sensitize tumors to cisplatin in a similar fashion. Following DNA damage recognition, downstream DNA repair proteins (XPA, RPA, TFIIH complex, ERCC1/XPF and XPG) are recruited to the DNA damage recognition complexes in both TCR and GGR to remove the damage in a common pathway. Over-expression of XPA and ERCC1 mRNA has been associated with cisplatin resistance in ovarian cancer \[[@B35]\]. In the current study, cisplatin induced XPA (Figure [4B](#F4){ref-type="fig"}, panel c) that was suppressed by hyperthermia co-treatment (Figure [4](#F4){ref-type="fig"} panel c). Suppression of XPA might decrease repair of cisplatin-DNA damage. ERCC1 was modestly induced (\<1.5 fold) by NaAsO~2~co-treatment with cisplatin at 37°C (CPA37) (Figure [4B](#F4){ref-type="fig"}, panel d).
In addition to the NER pathway, the mismatch repair (MMR) system has been implicated in cisplatin resistance \[[@B37]\]. In an effort to repair Pt-DNA damage by the MMR system, a futile MMR occurs leading to cell death \[[@B53],[@B55]\]. Ovarian cancer cells over-expressing MMR proteins are sensitive to cisplatin \[[@B55]-[@B57]\]. We report for the first time that tumors treated with cisplatin at 37°C (CP37) significantly suppressed MSH2 consistent with resistance. The observed suppression of MSH2 by cisplatin was reversed in tumors co-treated with NaAsO~2~at 37 or 43°C (CPA/37 and CPA/43 respectively) Thus, NaAsO~2~at 37 or 43°C has the potential to sensitize tumors to cisplatin by maintaining functional MMR.
Cisplatin causes serious and dose-limiting side effects including kidney damage, peripheral sensory neuropathy, cardiovascular toxicity, myelosuppression and anemia which occur as a result of diffusion of chemotherapy from the peritoneal to systemic compartment. In addition, arsenic also causes adverse side effects including cardiovascular toxicity, kidney damage, myelosuppression and anemia, liver damage and peripheral sensory neuropathy. Understanding the biodistribution of these drugs during peritoneal perfusion of chemotherapy is important in order to predict the occurrence of these adverse side effects and determine the risk:benefit balance in performing intraperitoneal perfusion with cisplatin and arsenic. For this reason, we determined platinum and arsenic accumulation in the brain, heart, liver, kidney and spleen during (0 h) and 24 h after perfusion. We observed that platinum and arsenic accumulated to similar extent in these tissues regardless of the treatment condition. The greatest accumulation of Pt was observed in the kidney, the site of Pt elimination. Likewise, greatest level of arsenic was observed in the liver, the organ for arsenic metabolism and detoxification. Even though we did not observe any toxicity with the short-term survival study, accumulation of arsenic and Pt in assayed organs suggests that potential adverse side effects such as encephalopathy, cardiotoxicity, liver damage, renal damage and myelosuppression/anemia respectively may occur during long-term survival studies.
Conclusions
===========
NaAsO~2~alone or combined with hyperthermia is most likely to enhance cisplatin efficacy because of its abilities to impair NER by inhibiting induction of p53 and XPC and to activate MMR by maintaining high levels of MSH2 and enhancing platinum accumulation in tumors. NaAsO~2~and hyperthermia might not produce added systemic toxicity to cisplatin chemotherapy; on the contrary, the combined treatment might help in the clearance of Pt from tissues. Long-term survival studies are required to determine the efficacy of this new combination chemotherapy. The murine HIPEC model may serve as a useful tool to study in vivo mechanisms of platinum resistance and explore ways to sensitize tumors to platinum chemotherapy.
Abbreviations
=============
CP: (cisplatin); CP/37: (cisplatin at 37°C) or CP/43 (cisplatin at 43°C); CPA: (cisplatin plus sodium arsenite); CPA37: (cisplatin plus sodium arsenite at 37°C) or CPA/43 (cisplatin plus sodium arsenite at 43°C); ERCC1: (excision repair cross-complementing 1); GGR: (global genome repair); HIPEC: (hyperthermic intraperitoneal chemotherapy); ICP-MS: (inductively coupled plasma mass spectrometry); NaAsO~2~: (sodium arsenite); MSH2: (human mutS homolog 2); NER: (nucleotide excision repair); Pt: (platinum); TCR: (transcription coupled repair); XPA: (xeroderma pigmentosum group A); XPC: (xeroderma pigmentosum group C).
Competing interests
===================
Dr Helm has previously received speaking honoraria from ThermaSolutions and grant support from ThermaSolutions and Sanofi-Aventis for clinical research into Hyperthermic Intraperitoneal Chemotherapy for the treatment of ovarian carcinoma.
All other authors declare that they have no competing interests
Authors\' contributions
=======================
CSM established metastatic tumor model, performed HIPEC and tissue collection, ICP-MS analysis, western blot analysis, immunohistochemical studies and drafted the manuscript. VAS established metastatic tumor model, performed HIPEC and tissue collection, took and drew pictures for figures [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}. JHM established metastatic tumor model, developed murine HIPEC model in collaboration with CWH, performed HIPEC and tissue collection. TWF provided intellectual input with ICP-MS analysis. CWH developed murine HIPEC model, established metastatic tumor model and participated in study design, coordination, data analysis and manuscript editing. JCS developed murine HIPEC model, established metastatic tumor model and participated in study design, coordination, data analysis and manuscript editing. All authors read and approved the final manuscript.
Acknowledgements
================
This work was supported in part by the National Institutes of Health Grant P30ES014443 which supported the collection and analysis of data and the National Science Foundation\'s Experimental Program to Stimulate Competitive Research Grant EPS-0447479 which provided the ICP-MS instrumentation and personnel support for the analysis of Pt and As reported in the manuscript.
Also, the authors thank Dr. Richard Higashi for technical support with ICP-MS analyses and Dr. Huaiyu Zheng of the Brown Cancer Center Small Animal Imaging Facility for technical assistance with microCT scanning of mice.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .723}
{#sp2 .724}
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
Hepatocellular carcinoma (HCC) is the sixth most common diagnosed cancer and the third leading cause of cancer death in the world \[[@CR1], [@CR2]\]. Although advances in treatment have been made in HCC in the past decades, surgical resection and liver transplantation are the primary effective approaches to treat HCC, and the overall 5-year survival rate of HCC patients remains very poor due to its advanced tumor stages at the time of diagnosis \[[@CR3], [@CR4]\]. Increasing evidence has shown that several genetic or epigenetic alterations contribute to tumorigenesis, and unrestricted cell growth and invasion are the most two key traits of HCC \[[@CR5]--[@CR7]\]. However, limited information is available regarding how to improve the probability of survival. Therefore, a better understanding of the mechanisms of key genes that can be used to provide novel therapeutic strategies is urgently needed for patients with HCC.
Emerging evidence has suggested that non-coding RNAs (ncRNAs) play an essential role in various physiological and pathological processes, including cell proliferation, differentiation, apoptosis, invasion, metabolism, developmental timing, and immune responses \[[@CR8]\]. lncRNAs, functional RNA molecules larger than 200 nucleotides in length have been shown to exert their functions as oncogenes or tumor suppressors, and their aberrant expression contributes to carcinogenesis \[[@CR9]--[@CR11]\]. For example, lncRNA GAS5 has been shown to be significantly downregulated in HCC tissues, and the overexpression of GAS5 suppresses HCC cell migration and invasion through negatively regulating the expression of miR-21 \[[@CR12]\]. lncRNA HOTAIR has been reported to be upregulated in breast cancer, pancreatic cancer, non-small cell lung cancer, and gastric cancer, and high HOTAIR expression is associated with poor prognosis \[[@CR13]--[@CR16]\].
Previous studies have shown that AFAP1-AS1 is upregulated in esophageal adenocarcinoma, pancreatic ductal adenocarcinoma, lung cancer, and nasopharyngeal carcinoma, and high AFAP1-AS1 expression is associated with lymph node metastasis, perineural invasion, and poor survival \[[@CR17]--[@CR20]\]. However, little is known concerning the potential role of AFAP1-AS1 in the development and progression of HCC. In the current study, we detected the expression of AFAP1-AS1 in HCC and found that AFAP1-AS1 was dramatically increased in HCC tissues compared with matched normal tissues. High AFAP1-AS1 expression was correlated with poor clinicopathological features and poor outcome in HCC patients. Further functional studies of AFAP1-AS1 suggested that knockdown of AFAP1-AS1 could decrease cell proliferation in vitro and in vivo.
Materials and methods {#Sec2}
=====================
Patients and tissue specimens {#Sec3}
-----------------------------
HCC tissues and normal tissues were obtained from 156 patients who had undergone surgical resection of HCC between April 2008 and February 2011 at The First Affiliated Hospital of Zhengzhou University, China. No chemotherapy or radiotherapy treatment had been administered to these patients prior to surgery. All collected specimens were immediately frozen in liquid nitrogen and stored at −80 °C until RNA extraction. Tumors and noncancerous tissues were confirmed using histological diagnoses by two experienced pathologists. Written informed consent was obtained from the patients, and this study was approved by the research ethics committee of The First Affiliated Hospital of Zhengzhou University.
Cell lines and cell culture {#Sec4}
---------------------------
Non-malignant liver cells (LO2), the human embryonic kidney cell line 293 T, and human hepatoma cell lines (SMMC-7721, Bel-7402, MHCC-97 L, and MHCC-97H) were obtained from Shanghai Cell Bank of the Chinese Academy of Sciences (Shanghai, China). All of the cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Hyclone, Logan, UT, USA) supplemented with 10 % fetal bovine serum (Invitrogen, Grand Island, NY, USA) and 1 % penicillin G and streptomycin in 37 °C humidified air containing 5 % CO~2~.
RNA isolation and quantitative real-time polymerase chain reaction (qRT-PCR) {#Sec5}
----------------------------------------------------------------------------
Total RNA was isolated from tissues and cells using Trizol Reagent (Invitrogen) according to the manufacturer's protocol. RNA concentration and integrity were determined using the NanoDrop2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and standard RNA gel electrophoresis. Real-time PCR was performed using SYBR®Green (TaKaRa, Dalian, China) and the ABI Prism 7900 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's protocol. The primers used were as follows: AFAP1-AS1, forward 5′-TCGCTCAATGGAGTGACGGCA-3′ and reverse 5′-CGGCTGAGACCGCTGAGAACTT-3′; GAPDH, forward 5′- CACCCACTCCTCCACCTTTG-3′ and reverse 5′- CCACCACCCTGTTGCTGTAG-3′. The fold-change for AFAP1-AS1 was normalized to GAPDH.
Lentiviral infection {#Sec6}
--------------------
Lentivirus-mediated suppression of human AFAP1-AS1 was constructed as previously described \[[@CR19]\]. Lentivirus packaging and cell transduction were carried out as previously described \[[@CR21]\].
Immunohistochemical staining {#Sec7}
----------------------------
Standard immunohistochemical procedures were performed using anti-Ki67 polyclonal antibody (ab66155; Abcam) as previously described \[[@CR21]\].
Cell proliferation and clone formation assay {#Sec8}
--------------------------------------------
For the cell proliferation assay, 2 × 10^3^ cells in 100 μL of complete culture were seeded in 96-well culture plates with five replicate wells. Next, 10 μL of CCK-8 (Dojindo, Kumamoto, Japan) was added to each well at 24, 48, 72, and 96 h and incubated at 37 °C for another 2 h. The absorbance was measured at 450 nm using a Multiskan® Spectrum system (Thermo Fisher Scientific).
For the clone formation assay, 600 cells were plated into 6-well plates and were maintained in media containing 10 % FBS with the medium being replaced every 3 days. Colonies were fixed with methanol and stained with 0.1 % crystal violet 2 weeks later. The visible colonies were photographed and manually counted.
Apoptosis assay {#Sec9}
---------------
Flow cytometry was performed to determine the apoptotic rate. Cells were harvested and washed three times with phosphate-buffered saline. Staining with Annexin V-PE and Propidium iodide (PI) was performed using the PE annexin V apoptosis detection kit (BD Pharmingen, San Diego, CA, USA) according to the manufacturer's recommendations, and the cells were then analyzed using a Cytomics FC-500 flow cytometer (Beckman Coulter, Inc., Fullerton, CA, USA).
Wound healing assay {#Sec10}
-------------------
Cells were plated onto in 6-well plates and cultured to 80 % confluence. Wounds were created with a sterilized pipette tip, and the detached cells were washed off thrice with PBS. Images were taken at 0 and 36 h after the scratching using digital microscopy.
Cell migration and invasion assays {#Sec11}
----------------------------------
Cell migration and invasion were performed as we have previously described \[[@CR21]\]. Briefly, 5 × 10^4^ MHCC-97 L or MHCC-97H cells transduced with si-AFAP1-AS1 or the respective control were added to the upper chamber of an insert coated with or without Matrigel (BD Biosciences, San Jose, CA, USA). Cell migration was allowed to proceed at 37 °C for 24 h, and cell invasion was allowed to occur at 37 °C for 48 h. After removing the cells on the inner surface of the filter membrane with a sterile cotton swab, cells were stained with 0.1 % crystal violet, imaged, and counted using digital microscopy.
Western blotting {#Sec12}
----------------
Western blot assays were performed as we previously described \[[@CR21]\]. The membranes were incubated with anti-BCL-2 antibody (\#2876), anti-Bax antibody (\#2772), anti-MMP-9 antibody (\#3852), and anti-GAPDH antibody (\#2118), all from CST (Cell Signaling Technology, Danvers, MA, USA), and subsequently, with the appropriate horseradish peroxidase-conjugated secondary antibody (Beyotime, Zhejiang, China).
Tumorigenicity in nude mice {#Sec13}
---------------------------
The animal experiments were approved by the Committee on the Ethics of Animal Experiments of Zhengzhou University. This study was performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. MHCC-97H cells (1 × 10^7^ cells) transduced with the negative control (Scramble) or siAFAP1-AS1 were injected into the flanks of 4-week-old nude mice. The tumor size was measured using hand calipers every week. All of the mice were sacrificed 5 weeks after inoculation. The tumor tissues were removed from each mouse, weighed, imaged, embedded in 10 % paraffin, and subjected to IHC staining.
Statistical analysis {#Sec14}
--------------------
Statistical analyses were performed using SPSS 17.0 software (SPSS Inc., Chicago, IL, USA). All of the data were expressed as the means ± SD (standard deviation), and Student's *t* test was used to analyze the differences between two groups. The correlation between AFAP1-AS1 expression and clinicopathological features was evaluated using Pearson *χ*2 test. The Kaplan-Meier method was performed to evaluate the survival rate in patients with HCC. A *P* value less than 0.05 was regarded as statistically significant.
Results {#Sec15}
=======
AFAP1-AS1 is significantly increased in HCC samples and HCC cell lines {#Sec16}
----------------------------------------------------------------------
AFAP1-AS1 expression levels in 156 HCC tissue samples and 156 matched normal liver tissue samples were examined by quantitative real-time PCR. AFAP1-AS1 levels in cancer samples were significantly higher than those in the noncancerous samples (*P* \< 0.001; Fig. [1a](#Fig1){ref-type="fig"}). Furthermore, qRT-PCR was performed to detect the expression of AFAP1-AS1 in diverse HCC cell lines and a non-malignant liver cell line. As shown in Fig. [1b](#Fig1){ref-type="fig"}, the AFAP1-AS1 expression level was significantly increased in HCC cell lines compared with a non-malignant liver cell line (LO2). These results suggested that AFAP1-AS1 might function as an oncogene in HCC.Fig. 1AFAP1-AS1 expression is increased in human HCC tissues and cell lines. **a** The relative expression of AFAP1-AS1 was detected in 156 pairs of primary HCC tissues and their corresponding adjacent tissues. **b** qRT-PCR analysis was performed to assess the AFAP1-AS1 levels in HCC cells (SMMC-7721, BEL-7402, MHCC-97 L, and MHCC-97H) and LO2. \**P* \< 0.05 and \*\**P* \< 0.01
Association between AFAP1-AS1 expression and clinicopathological features in HCC {#Sec17}
--------------------------------------------------------------------------------
To understand the significance of AFAP1-AS1 overexpression in HCC, we explored the associations between AFAP1-AS1 expression and the patients' clinicopathological features (Table [1](#Tab1){ref-type="table"}). There was a significant association between AFAP1-AS1 expression and tumor size (*P* = 0.012), vascular invasion (*P* = 0.035), and TNM stage (*P* = 0.01). However, AFAP1-AS1 expression was not associated with other parameters such as age, gender, or liver cirrhosis in HCC patients. Furthermore, according to the median AFAP1-AS1 expression in tumor tissues, the patients were classified into the high AFAP1-AS1 expression group (*n* = 78) or relative low AFAP1-AS1 expression group (*n* = 78). The patients with lower AFAP1-AS1 expression had better overall survival (OS) and disease-free survival (DFS) (Fig. [2a, b](#Fig2){ref-type="fig"}). These results indicated that AFAP1-AS1 might be a potential prognostic biomarker for HCC patients.Table 1Correlation between AFAP1-SA1 expression and the clinicopathological features in HCC patientsVariablesAFAP1-SA1\
expression levelLowHigh*P* value*n* = 78*n* = 78Gender 0.495^a^Female63Male7275Age, y 0.423≤514136\>513742Liver cirrhosis 0.375No1014Yes6864Tumor size 0.012≤5 cm6450\>5 cm1428Tumor number 0.598Single6971Multiple97Tumor envelope 0.1997Yes4536No3342TNM stage 0.010I2713II/III5165Preoperative AFP (ng/mL) 0.144≤203728\>204150Vascular invasion 0.035No5238Yes2640ALT (units/L) 0.575≤757072\>7586*ALT* alanine aminotransferase, *TNM* tumor-node-metastasis,^a^Fisher's exact tests and *Χ* ^2^ tests for all other analyses Fig. 2Kaplan-Meier analysis of HCC survival rates in relation to AFAP1-AS1 expression. The overall survival **a** and disease-free survival **b** rates of HCC patients with high (*n* = 78) and low (n = 78) levels of AFAP1-AS1 expression. The *P* values were determined using the log-rank test
Knockdown of AFAP1-AS1 inhibits HCC cell proliferation in vitro {#Sec18}
---------------------------------------------------------------
To explore the biological significance of AFAP1-AS in HCC, we transfected lentivirus-mediated siRNA into MHCC-97 L and MHCC-97H cells, which had high endogenous AFAP1-AS levels. As shown in Fig. [3a](#Fig3){ref-type="fig"}, the expression of AFAP1-AS was significantly decreased compared with that of the negative control (SCR). Cell proliferation assays were performed to investigate the effect of AFAP1-AS on cell viability using the CCK-8 assay. Our results showed that knockdown of AFAP1-AS expression significantly suppressed cell growth in MHCC-97 L and MHCC-97H cells compared with that in the respective controls (Fig. [3b](#Fig3){ref-type="fig"}). Furthermore, colony formation assays showed that downregulation of AFAP1-AS could significantly inhibit colony formation in MHCC-97 L and MHCC-97H cells (Fig. [3c](#Fig3){ref-type="fig"}). In addition, AFAP1-AS silencing induced an increase in the apoptosis levels using the Annexin V Apoptosis Assay (Fig. [3d](#Fig3){ref-type="fig"}). These data suggest that AFAP1-AS plays a critical role in HCC cell proliferation in vitro.Fig. 3Knockdown of AFAP1-AS1 decreases HCC cell proliferation in vitro. **a** qRT-PCR was performed to detect the expression of AFAP1-AS1 in MHCC-97 L and MHCC-97H cells transduced with SCR or siAFAP1-AS1. The effects of knockdown of AFAP1-AS1 in cells on cell proliferation **b** colony formation **c** and apoptosis **d** were examined. \**P* \< 0.05 and \*\**P* \< 0.01
Knockdown of AFAP1-AS1 inhibits HCC cell migration and invasion {#Sec19}
---------------------------------------------------------------
Unrestricted cell growth and motility are the two most important traits of cancers. Thus, wound healing assay and transwell assays were performed to evaluate the cell motility capability. As determined by the wound healing assay, knockdown of AFAP1-AS1 markedly inhibited the migratory ability of the MHCC-97 L and MHCC-97H cells (Fig. [4a](#Fig4){ref-type="fig"}). The transwell assay also showed that motility was significantly decreased in cells transduced with siAFAP1-AS1 compared with that in the negative control (Fig. [4b](#Fig4){ref-type="fig"}). To explore whether AFAP1-AS1 exerts its functions through the proliferation- and apoptosis-related genes that are known to be involved in cancer, we examined the expression of Ki67, Bcl-2, Bax, and MMP9. As shown in Fig. [4c](#Fig4){ref-type="fig"}, knockdown of AFAP1-AS1 resulted in a dramatic decrease in Ki67, Bcl-2, and MMP9 levels and an increase in Bax levels. Taken together, the data strongly suggest that AFAP1-AS1 promotes HCC cell proliferation and invasion through mediating proliferation- and apoptosis-related gene expression in vitro.Fig. 4The effects of AFAP1-AS1 knockdown on cell invasion. **a** Representative images of wound-healing assay showed that the migration ability of MHCC-97 L and MHCC-97H cells transduced with SCR or siAFAP1-AS1 at 0 h and 48 h time point, respectively. **b** The Boyden chamber assay was employed to examine the invasion ability of MHCC-97 L and MHCC-97H cells transduced with SCR or siAFAP1-AS1, respectively. **c** The expression of Ki-67, BCL-2, Bax, and MMP-9 in MHCC-97 L and MHCC-97H cells transduced with SCR or siAFAP1-AS1 was determined by western blotting. \*\**P* \< 0.01
Knockdown of AFAP1-AS1 suppresses tumor growth in vivo {#Sec20}
------------------------------------------------------
To further determine whether AFAP1-AS1 silencing affects tumorigenesis, MHCC-97H cells transduced with SCR or siAFAP1-AS1 were subcutaneously inoculated into nude mice. All of the mice were sacrificed 5 weeks after injection, and tumors derived from the AFAP1-AS1-silencing group developed more slowly than those from the SCR group (Fig. [5a](#Fig5){ref-type="fig"}). Consistent with tumor size, the average tumor weight in the siAFAP1-AS1 group was significantly lower than that in the SCR group (Fig. [5b](#Fig5){ref-type="fig"}). Tumor growth in the knockdown AFAP1-AS1 group was slower than that in the SCR group (Fig. [5c](#Fig5){ref-type="fig"}). Furthermore, qRT-PCR was performed to detect the expression of AFAP1-AS1 in the xenograft tumor tissues. Our results showed that the levels of AFAP1-AS1 expression in tumor tissues formed in the siAFAP1-AS1 group were lower than those of the tumors formed in the SCR group (Fig. [5d](#Fig5){ref-type="fig"}). In addition, immunostaining revealed that the tumors derived from the siAFAP1-AS1 group showed weaker Ki-67 expression than that in tumors from the SCR group (Fig. [5e](#Fig5){ref-type="fig"}). These data further supported the role of AFAP1-AS1 in HCC cell growth in vivo.Fig. 5AFAP1-AS1 silencing inhibited tumor growth in a xenograft mouse model. **a** Representative photographs of tumors are shown. Tumor weight **b** and tumor growth curves **c** in mice are shown for MHCC-97H cells transduced with SCR or siAFAP1-AS1. **d** qRT-PCR was used to detect the average expression of AFAP1-AS1 in xenograft tumors. **e** Representative images of IHC staining showed the expression of Ki67 in xenograft tumor tissues from the siAFAP1-AS1 group or SCR group. \*\**P* \< 0.01
Discussion {#Sec21}
==========
Accumulating evidence has shown that lncRNAs are associated with many functions in various aspects of cell biology, and increasing attention has been paid to their roles in tumors \[[@CR9]--[@CR11]\]. Previous studies have demonstrated that AFAP1-AS1, localized in the antisense DNA strand of the AFAP1 gene, regulates the invasion and metastasis of lung cancer and PDAC cells \[[@CR18], [@CR19]\]. Recently, Hao et al. provided evidence that AFAP1-AS1 mediated cell malignant behavior by affecting the expression of several small GTPase family members and molecules in the actin cytokeratin signaling pathway in nasopharyngeal carcinoma \[[@CR20]\]. However, the expression and functions of AFAP1-AS1 in HCC remain unclear. In the current study, we provided the first evidence that AFAP1-AS1 was significantly upregulated in HCC tissues compared with that in adjacent normal tissues, and high AFAP1-AS1 expression in HCC patients was associated with an increased tumor size, vascular invasion, advanced TNM stage, and poor prognosis, findings that were consistent with those in previous studies. These findings suggested that AFAP1-AS1 acts as an oncogene in the progression of HCC.
Because high AFAP1-AS1 expression was associated with tumor size in HCC, we surmised that AFAP1-AS1 might play a key role in tumor cell proliferation. To further elucidate the functions of AFAP1-AS1 in HCC, we knocked down AFAP1-AS1 in MHCC-97 L and MHCC-97H with high endogenous AFAP1-AS1 expression using RNAi-mediated suppression. Our results showed that AFAP1-AS1 knockdown could significantly inhibit HCC cell proliferation and invasion in vitro. In addition, flow cytometric analysis indicated the inhibitory effect of AFAP1-AS1 silencing on the proliferation of HCC cells by inducing apoptosis. Similarly, the gain- and loss-of-function strategies for AFAP1-AS1 expression resulted in changes in cell proliferation and apoptosis in esophageal cancer \[[@CR17]\] and migration and invasion in pancreatic cancer and nasopharyngeal carcinoma \[[@CR19], [@CR20]\]. It is well known that aberrant BCL-2 and Bax expression is involved in cancer \[[@CR22]\]. As we expected, AFAP1-AS1 silencing led to a decrease in Ki-67, BCL-2, and MMP-9 and an increase in Bax. In addition, AFAP1-AS1 knockdown significantly inhibited tumor growth in vivo, confirming the vitro experiments.
In conclusion, our results showed that AFAP1-AS1 is remarkably upregulated in HCC tissues and is significantly correlated with the malignancy status and poor prognosis. Furthermore, AFAP1-AS1 silencing suppressed cell proliferation and induced cell apoptosis at least partly through regulating proliferation- and apoptosis-related genes. However, dysregulation of AFAP1-AS1 is only the tip of iceberg; more studies on the functions of AFAP1-AS1 in HCC are needed.
Xu Lu, Chuang Zhou and Renfeng Li contributed equally to this work.
This study was supported by the National Natural Science Foundation of China (grant no. 81171849) and Zhengzhou Innovation Team Project (Grant No. 131PCXTD617).
Conflicts of interest {#FPar1}
=====================
None
| {
"pile_set_name": "PubMed Central"
} |
The content published in Cureus is the result of clinical experience and/or research by independent individuals or organizations. Cureus is not responsible for the scientific accuracy or reliability of data or conclusions published herein. All content published within Cureus is intended only for educational, research and reference purposes. Additionally, articles published within Cureus should not be deemed a suitable substitute for the advice of a qualified health care professional. Do not disregard or avoid professional medical advice due to content published within Cureus.
Introduction and background
===========================
Venous thromboembolism (VTE) is a condition that involves the formation of clots in the deep veins, particularly in the veins of the lower limb. This causes obstruction to blood flow resulting in symptoms like pain, swelling and discoloration \[[@REF1]\]. The most common complication of venous thrombosis is the migration of these clots into other blood vessels, called embolism, particularly pulmonary embolism. According to research, half of the hospitalized patients are at a risk of thromboembolism \[[@REF2]\]. The rate of VTE ranges from 10% to 80% percent in patients, who are not being given any prophylaxis \[[@REF3]-[@REF4]\]. Studies have proven the rate of VTE is greater in hospitalized than community patients \[[@REF5]\].
The impact of thromboprophylaxis can be ascertained from the fact that it reduces the rate of thromboembolism in both medical and surgical patients. However, it decreases the mortality rate in surgical patients only, having little or no impact on the mortality rate among medical patients \[[@REF6]-[@REF7]\]. Thromboprophylaxis is of two varieties, primary and secondary. Primary prophylaxis is the one which is given to prevent the occurrence of deep vein thrombosis (DVT), which includes pharmacologic therapy like unfractionated heparin (UFH), low molecular weight heparin (LMWH), fondaparinux or mechanical therapy like pneumatic and graduated compression stockings \[[@REF8]\]. Secondary prophylaxis involves early detection and management of venous thrombosis. The method of primary prophylaxis is determined by factors like the risk of thrombosis and hemorrhage, nature of the illness, the policy of the institution, cost and preferences. These factors help to classify patients into low, moderate and high-risk categories, each having a different method of prophylaxis. Moreover, the duration of prophylactic treatment differs from patient to patient depending on the risk classification.
Patients admitted to intensive care units (ICUs) are rated as high-risk patients even if they are being given prophylaxis \[[@REF9]-[@REF10]\]. Trials comparing the efficacy of pharmacological and mechanical thromboprophylaxis in ICU patients are but few. There is a need for further study regarding proper thromboprophylaxis in these patients in order to reduce mortality. Our study aims to review the literature from the past five years regarding the incidence, diagnosis, and prevention of VTE in ICU patients.
Review
======
Epidemiology
DVT is prevalent in ICUs, particularly in Western countries. Asian countries have lower incidence comparatively. In a study in Thai surgical ICU patients, the incidence of DVT was found to be 3.6% which is comparable to a study in Tehran in which the incidence of DVT in ICU patients was found to be 3.5 % \[[@REF11]\]. Two independent risk factors of DVT development are longer ICU stay and older age \[[@REF12]\]. Another study was conducted in Chinese cancer patients admitted to ICU for the purpose of detecting VTE. It revealed a low incidence of VTE \[[@REF13]\]. The incidence of VTE was 37.2% in patients with sepsis and septic shock \[[@REF14]\]. Studies showed comparable results in adolescents \[[@REF15]\].
Risk factors
Risk factors for thromboembolism can be divided into two groups, genetic and acquired. Genetic risk factors include loss and gain of coagulation function disorders \[[@REF16]\]. According to studies, people having factor V Leiden or prothrombin 21210 mutation have the higher risk of thromboembolism than those without them \[[@REF17]\]. Acquired risk factors include bed rest, age, hematologic cancers, immobilization, obesity, pregnancy, smoking, stroke, long-distance travel and certain inflammatory conditions \[[@REF18]\]. The risk of VTE is higher in patients admitted to ICU than others, due to the higher number of risk factors specific to ICUs. Sepsis, vasopressor use, central catheters, mechanical ventilators, respiratory, cardiac or renal failure are common ICU related risk factors of VTE \[[@REF19]\]. Research proves that patients with inserted catheters are at a three-fold greater risk for developing thromboembolism than those without them \[[@REF20]\] Similarly, the duration of mechanical ventilation has an impact on the incidence of thromboembolism \[[@REF14]\] Also, according to a recent study, the risk of developing thrombosis is directly proportional to the number of packed red blood cell (PRBC) transfusions \[[@REF21]\].
Diagnosis
The diagnosis of VTE is highly dependent on the risk factors in a particular patient and the clinical presentation of the patient, for example, swelling, tenderness, hemoptysis, dyspnea etc. Also, different scoring systems help in diagnosis, most commonly employed for DVT is Well's criteria \[[@REF22]\]. Initially, following the criteria, patients were classified into high, intermediate and low-risk groups. Currently, Doppler venous ultrasound is the best imaging modality for diagnosis of DVT \[[@REF23]\]. Others like contrast venography and magnetic resonance venography are also being employed, but they have certain limitations \[[@REF24]\]. The diagnosis of pulmonary embolism (PE) also begins with the pre-test probability which is determined by Wells score, classified into low and high probability \[[@REF25]\]. computed tomography pulmonary angiography (CTPA) is the gold standard for diagnosing PE \[[@REF26]\].
The rationale for thromboprophylaxis
DVT and PE are very common complications in critical patients, especially those in ICU. Use of thromboprophylaxis can reduce mortality in such patients. Three randomized controlled trials (RCTs) held in ICU patients concluded that the incidence of DVT was significantly lower in the thromboprophylaxis group in comparison to the control group, irrespective of the type of thromboprophylaxis used \[[@REF27]-[@REF29]\]. According to current guidelines for DVT prevention in critically ill medical patients, either UFH or LMWH may be used. The Prophylaxis of Thromboembolism in Critical Care Trial (PROTECT) is a randomized clinical trial that is being held presently to study these different modes of thromboprophylaxis \[[@REF8]\]. The protocol according to the American College of Chest Physicians (ACCP) guidelines is as follows: 1) Routine evaluation for VTE risk and thromboprophylaxis is recommended for critically ill patients (Grade 1A). 2) LMWH or low-dose UFH thromboprophylaxis should be given to patients with moderate risk for VTE (e.g, medically ill or postoperative general surgery patients); (Grade 1A). 3) Mechanical thromboprophylaxis is considered best for critical care patients who are at greater risk for bleeding, at least, until the bleeding risk decreases (Grade 1A). When the high bleeding risk decreases, pharmacologic thromboprophylaxis should be substituted for or added to the mechanical thromboprophylaxis (Grade 1C) \[[@REF8]\].
Pharmacological thromboprophylaxis
Pharmacological prophylaxis with heparin is recommended for patients at the time of admission to the ICU. Heparin should be discontinued temporarily in patients with active bleeding or severe (\<50,000/cc) thrombocytopenia. Though different thromboprophylaxis regimens have been suggested there is still no agreement based on evidence \[[@REF30]\]. Until now, the PROTECT study comprising 3764 patients is the only RCT to have compared UFH with LMWH as VTE prophylaxis in the ICU. It did not include patients who were at higher risk for bleeding \[[@REF31]\]. The patients were randomly assigned to two different groups. One group was given 5,000 IU of subcutaneous dalteparin once daily plus placebo once daily and the other received 5,000 IU of subcutaneous UFH twice daily. However, the difference between the incidence of proximal DVT in the dalteparin vs UFH group was not significant; (5.1% vs 5.8%, p=0.57). However, the rate of PE was significantly lower in the dalteparin group (1.3 %) compared with the UFH group (2.3 %) (p = 0.01). A recent meta-analysis by Park J et al. showed a significant reduction in the risk of DVT with heparin compared with the control group (LMWH: OR, 0.38; UFH: OR, 0.45), which was comparable to the previous review. However, on comparing UFH and LMWH, the efficacy was found to be comparable, which was also in agreement with the previous analysis \[[@REF32]-[@REF33]\]. Thromboprophylaxis in critically ill patients may lead to bleeding which can be fatal. Moreover, both bleeding and the resulting discontinuation of thromboprophylaxis has a negative impact on clinical outcomes in the ICU. The meta-analysis also concluded that the difference in major bleeding risk between UFH and LMWH was not significant \[[@REF32]\].
According to the results of a retrospective observational cohort study, males having a body mass index ≥ 40 kg/m^2^ have the greater risk of developing VTE compared to females. However, in patients who were given a standard dose UFH, morbid obesity did not increase VTE risk overall. Also, morbid obesity was more likely associated with greater hospital and ICU length of stay \[[@REF34]\].
A study comparing two types of LMWHs, enoxaparin and bemiparin showed that bemiparin was better than enoxaparin as a prophylactic anticoagulant for VTE in critically ill patients. It was associated with fewer local complications at the injection site. DVT was found in only 4% of the patients in the bemiparin group while 20% of the patients taking enoxaparin developed DVT. Confirmed PE was found in 14% of patients in the enoxaparin group as compared to no recorded case in the bemiparin group \[[@REF35]\].
Anti-factor Xa levels can be used to clinically measure the effectiveness of LMWH anticoagulant; levels of 0.1 to 0.3 UI/ml are considered as adequate. Proper VTE prophylaxis is crucial in trauma patients with fractures of the lower extremity and pelvis. One such study explored whether dosing prophylactic enoxaparin using anti-Xa trough levels affected the incidence of VTE in such trauma patients. The study showed that in the majority of patients (84.5%) who had anti-Xa trough levels measured, the initial enoxaparin dose being given was sub prophylactic. Patients who were given enoxaparin according to anti-Xa trough level had a significantly lower risk of VTE compared to those in which anti-Xa levels were not measured; (1.7% v. 13.9%, p = 0.03) \[[@REF36]\]. Since LMWHs are mostly excreted by the kidneys, they can accumulate in patients with renal insufficiency and thus have increased the risk of causing bleeding. Further, the risk of acute renal failure is higher in critically ill patients; at ICU admission, nearly one-third of patients have a creatinine clearance below 30 ml/minute \[[@REF37]\]. A first meta-analysis failed to demonstrate the bio-accumulation of LMWH in critically ill patients with renal insufficiency \[[@REF38]\]. Critical patients with acute kidney injury present a challenge in the provision of thromboprophylaxis as the dose has to be balanced with the risk for bleeding. A trial held in Denmark showed that the existing recommendation of 40 mg enoxaparin is not sufficient in patients with renal dysfunction and that 1 mg/kg is both safe and effective for thromboprophylaxis \[[@REF39]\]. Another study consisting of a subgroup analysis of PROTECT was held in order to find out the safety and efficacy of LMWH VTE prophylaxis in critically ill patients with abnormal renal function. When dalteparin 5000 IU daily was compared with UFH 5000 IU twice daily in patients with renal insufficiency or end-stage renal disease (ESRD), there was no significant difference in the incidence of VTE or major bleeding. Patients with severe renal dysfunction who were on dalteparin did have a greater incidence of proximal DVTs compared to those on UFH; there was no greater risk of VTE or major bleeding \[[@REF40]\].
Factor Xa inhibitors can also act as anti-coagulants by inhibiting the coagulation factor X and thus preventing clotting. In addition to heparin, their role in VTE prophylaxis should also be investigated in medically ill patients. In a case report in Pakistan, a patient was admitted to the medical ICU with hospital-acquired pneumonia. She had a low risk of VTE and so, she was advised to simply continue her rivaroxaban therapy-which she was already taking for avalvular atrial fibrillation- in addition to mechanical measures to prevent VTE. Unfortunately, she developed pulmonary VTE while she was being mechanically ventilated. This is the first case report of an incident in which a patient developed VTE despite being adequately anticoagulated with rivaroxaban \[[@REF41]\]. Since VTE can still occur sometimes despite thromboprophylaxis with heparin, there is a need to study other agents for thromboprophylaxis. Aspirin has been shown to decrease the risk of VTE in surgical and high-risk medical patients but its effects in mechanically ventilated ICU patients are unknown. A study designed to investigate the effect of aspirin on thromboembolic events in mechanically ventilated patients showed a significant reduction in the odds of finding DVT with aspirin (OR 0.39, 95% CI 0.16--0.94; p = 0.036) \[[@REF42]\]. So, aspirin may be helpful in preventing DVT in such patients.
Mechanical thromboprophylaxis
When anticoagulation is contraindicated, mechanical thromboprophylaxis using either graduated compression stockings (GCS) or intermittent pneumatic compression (IPC) may be proposed. Thromboprophylaxis by mechanical means alone is recommended for critical care patients at high risk of bleeding, in whom anticoagulants are contra-indicated \[[@REF8]\]. According to one study, the use of IPC but not GCS was associated with a significantly lower VTE risk. No real association was found between the mechanical thromboprophylaxis and the type of prophylactic heparin used, recent trauma or recent surgery \[[@REF43]\]. Pressure injuries have been found to be a notable complication of GCS in surgical ICU patients \[[@REF44]\]. A recent meta-analysis of 12 trials found a trend of reduced DVT risk with IPC, compared to the control group, but the reduction was not statistically significant. This shows that even though IPC is being commonly used, the thromboprophylactic efficacy of IPC is still questionable \[[@REF33]\]. A study in China explored the comprehension and practice of mechanical thromboprophylaxis in ICU medical staff. It found that approximately 52.30% of all surveyed medical staff often practiced mechanical thromboprophylaxis. However, 25% of the included staff had never heard of mechanical thromboprophylaxis \[[@REF45]\]. There is a need to remove concerns regarding IPC and GCS and properly educate the staff as these measures could help decrease VTE incidence in ICUs and improve the prognosis of critically ill patients by increasing the use of mechanical thromboprophylaxis.
Thromboprophylaxis compliance
The term thromboprophylaxis compliance refers to the extent to which ACCP prophylaxis guidelines are followed while administering any type of prophylaxis (pharmacologic or mechanical). According to a study conducted on 472 patients to evaluate if prophylaxis was being given in the right manner, it was concluded that 54.9% of patients were not being given appropriate prophylaxis \[[@REF46]\]. This included patients who had absolute indications for prophylaxis but were not given prophylaxis, patients who had no indications for prophylaxis but were given prophylaxis, and patients who received the incorrect type of prophylaxis. Another study involving 364 patients showed that 16% of the patients were not receiving thromboprophylaxis and 45% of patients were not receiving pharmacologic thromboprophylaxis. The most common reasons were recent bleeding or surgery, provision of mechanical prophylaxis and thrombocytopenia \[[@REF47]\]. Strategies to improve thromboprophylaxis compliance include the education of physicians and electronic reminders \[[@REF48]\].
Conclusions
===========
ICU patients are at greater risk for VTE due to additional ICU related risk factors. DVT and PE in these patients can be diagnosed with venous Doppler ultrasound and CTPA, respectively. Thromboprophylaxis in these patients poses a challenge. Pharmacologic thromboprophylaxis with all types of heparin has been proven to significantly help reduce VTE in ICU patients. The efficacy of UFH and LMWH has turned out to be comparable with no increased risk of major bleeding. Amongst LMWH, bemiparin has been shown to be superior to enoxaparin as a prophylactic anticoagulant. Dosing LMWH with anti-factor Xa levels could reduce the risk of VTE. However, the data to support the efficacy of mechanical thromboprophylaxis is not strong enough. The choice for the best method for thromboprophylaxis still needs further study and research.
The authors have declared that no competing interests exist.
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
Programs teaching mindful awareness to promote well-being have become a recent topic of inquiry among researchers \[[@B1]\]. Rooted in Buddhist traditions and formally developed by Jon Kabat-Zinn in the 1980s, mindfulness has developed secularly to describe a process of becoming aware of present experience and, through that awareness, reducing emotional and physical distress \[[@B2]\]. Mindfulness has been described as more than merely a relaxation technique; rather, it is mental training to develop awareness and acceptance skills to cope with daily events that may otherwise lead to heightened anxiety and stress \[[@B3]\]. Mindfulness-based programs enhance awareness and acceptance of conscious states and teach people to approach daily situations \"mindfully\" \[[@B4]\].
Research demonstrating the efficacy of mindfulness-based stress reduction (MSBR) in promoting health is mounting although randomized, controlled studies and the inclusion of minority older adult populations are both largely absent from the literature. Past research on mindfulness-based stress reduction (MBSR) programs has demonstrated enhanced coping, improved self-efficacy, and better symptom control in general populations without control groups \[[@B3],[@B5]\]. The few randomized controlled trials of MBSR have been in specific clinical populations such as cancer \[[@B6]\], rheumatoid arthritis patients \[[@B7]\]or those with chronic pain \[[@B8]\]. Studies have shown MBSR to be effective in treatment of chronic pain \[[@B2],[@B8]-[@B10]\], anxiety disorders \[[@B11]\], and hypertension \[[@B12]\]. Reibel et al. demonstrated that MBSR increases health-related quality of life and social functioning, while decreasing physical pain, role limitations caused by physical health, and anxiety and depression among mixed diagnosis patients \[[@B13]\]. While MBSR programs have been shown to be related to positive outcomes, the research evidence to date hass been limited to clinical populations \[[@B3]\]. Further, little evidence exists for the efficacy of MBSR programs among older adults \[[@B5]\]or in low-income, urban minority populations \[[@B14]\].
Older adults, we hypothesized, could be a particularly responsive group to train in mindfulness due to their life experience accompanied by a potentially increased ability to examine their lives as well as the capacity, honed by aging-related changes, to accept their life. Low-income minority older adults, we further hypothesized, might have more stressors\[[@B15]\], more depression \[[@B16],[@B17]\] and more social isolation \[[@B18],[@B19]\] and thus more possible benefit from mindfulness training. However, there was no literature base or prior reports of experience with this population to indicate low-income minority older adults\' acceptance of training that has formerly been associated with East Asian religious practices. Therefore, the purpose of the current study was to examine the perceptions of a mindfulness-based program in a sample of low-income minority older adults. Due to the mounting evidence regarding its effectiveness in reducing stress, increasing self-efficacy, and managing physical ailments, mindfulness was proposed as an effective intervention for minority older adults living with multiple social and physical concerns. Our team is currently using a randomized clinical trial to test these hypotheses. The goals of the present study were to assess older minority adult participants\' openness to the intervention, understand which components of the intervention were most important to them, and examine their application of the mindful awareness training outside of the intervention. Determining acceptance or openness to this intervention was an important first step in acquiring information that could be used to develop further mindfulness-based interventions for this underserved population at high risk for chronic multi-morbidities.
In order to determine openness to this intervention, we decided to ask a group of older low-income African -American adults about their perceptions of the program. Based on their answers, we could tailor future programs. We decided on a focus group methodology to accomplish this.
Focus group research is a qualitative method of collecting data that allows researchers to understand experiences from the perspective of the participants (Patton, 2002). Focus groups allow participants to interact with each other as they answer questions posed by the researcher. In doing so, they allow for individuals to disagree or expand on other participants\' comments. Focus groups are a useful approach to elicit first-hand information from participants. The rising popularity of focus groups highlights the uniqueness of their group context. This allows for the facilitation of efficient data collection and the potential to elicit participant views related to experiences involving motivations, feelings, attitudes and opinions on health-related issues among challenging and/or vulnerable target populations \[[@B20],[@B21]\]. Advantages of using focus groups include: (1) being less intimidating than one-on-one interviewing but providing more depth than questionnaires (2) having sensitivity to participants\' culture and age by acknowledging participants as experts and obtaining insight into participants\' own language and concepts (4) allowing group interactions to take place, (3) permitting researchers to learn more about the degree of consensus on a topic, and (4) encouraging dialogue and dialectic between researchers and participants in order to mutually identify, describe, analyze and attempt to resolve key issues \[[@B22],[@B23]\]. Because oral history is a strong tradition in many minority communities, inviting participants to describe their experiences is generally believed to be an effective approach to data collection with minority older adults.
Methods
=======
In order to collect data regarding the perceptions of a mindfulness-based program implemented among a group of low-income minority older adults, a focus group design was selected. Three focus groups were one hour each conducted with members of ELDERSHINE, a mindfulness-based program designed for older adults.
*The*ELDERSHINE *Program*
-------------------------
ELDERSHINE is a psycho-educational program designed to foster awareness of internal states; promote positive mental and physical health through mindfulness; and build individual and neighborhood capacity through civic engagement and the creation of a caring community. The design of a series of eight workshops creates a sequential and cohesive program. The program teaches mindfulness-based stress reduction skills adapted from Jon Kabat-Zinn\'s MBSR program and was developed by one of the co-authors (A.B.C.) to teach older adults how to meditate and use mindfulness in their daily lives. Modifications from MBSR include: briefer meditation periods; shorter weekly sessions; emphasis on seated meditations rather than mindful movements or walking meditation (due to physical limitations of participants); and no daylong retreat. In lieu of a workbook, participants received a folder and weekly handouts were given out at each session with program concepts and poetry in large type to facilitate easier reading.
ELDERSHINE invites participants to discover and mobilize their strengths and foster resiliency within themselves as individuals and as a community of elders who can share wisdom and strength as they support and nurture one another. The ELDERSHINE core program consists of eight 90 minute sessions. Each weekly session includes three guided meditations; an opportunity to discuss participants\' home practice and integration of mindful skills; a time for positive sharing (called sharing \"Victories\"); and a group process that allows participants to learn through reflection on their own experience, activity and supportive dialogue. A nourishing snack at a \"tea party\" served after each workshop encourages socialization, extended conversation among participants, and a focus on healthy eating.
Participants
------------
Participants for the study included members of an ELDERSHINE program being conducted in a low-income senior housing facility in Baltimore, Maryland. The moderator of the ELDERSHINE group verbally invited members to participate in the focus groups. The moderator also announced she would not attend the focus groups. Those interested were called the night before to increase attendance. Thirteen individuals participated in the study. Three separate small focus groups (of 4, 4, and 5 participants) were conducted to increase participation in discussion by having small groups. This is within the optimal range of 4-8 participants \[[@B23]\]. These groups included almost every regular ELDERSHINE attendee. All participants were low income African-American women over the age of 60 years old and younger than 90. They were all Protestants (Baptist, Methodist, and Apostolic). Their education level ranged from less than 8 years of school through obtaining some college education. This study was approved by the IRB at the Johns Hopkins School of Public Health. All participants provided oral consent.
Procedures
----------
Our team tape-recorded the focus group sessions, and both the moderator and the trained research assistant took notes during and after the sessions. A focus group interview guide and probes guided each focus group session. Two of the groups were moderated by one of the investigators on the project (S.L.S), and one was moderated by another investigator (R.L.P). The same trained research assistant took notes during all sessions and transcribed the sessions from the audiotapes afterwards.
For each session, the moderator started the discussion by informing the group of the purpose of the session. Any questions that the participants had were answered prior to beginning the discussion.
After all participants\' questions were answered, the following four questions were asked to guide discussion:
1\. What has ELDERSHINE meant to you?
2\. Do you feel that you have changed in any way due to your involvement in ELDERSHINE?
3\. What parts of ELDERSHINE have meant the most to you?
4\. How would you describe ELDERSHINE to others?
Analysis
--------
Following each focus group session, the audiotapes were transcribed verbatim by the research assistant into Microsoft Word 2000. Primary themes from the focus groups were identified by inductive content analysis (Patton, 2002; Straus & Corbin, 1990). The level of analysis was across cases and across sessions to allow for themes to be developed overall for all participants.
The focus group interview guide dictated the topics of data analysis.^31^The investigators first analyzed the data with multiple readings of interview transcripts to gain a general sense of the data along with review of researcher field notes which included researcher observations on communication factors including body language, gestures, tones and voice intensity \[[@B23]\]. We meticulously recorded initial findings to maintain a clear audit trail. Preliminary coding of concepts followed, with categorization and a search for themes (see steps, below). We examined the data for themes, patterns, commonalities and variation. We continually validated categories, themes, and conclusions by referring back to the data. We also examined theme variations.
We achieved data trustworthiness through collection and review of field notes by the trained research assistant at each focus group session, as noted above. Each investigator described and interpreted their own behavior and experiences in relation to the research and each participant which enhanced the credibility of the data \[[@B24]\]. We used field notes to provide a clear decision trail concerning the study, describing and justifying what was actually done and the reasons for doing so \[[@B24]\].
Judgment regarding trustworthiness and authenticity was made using member checking and participant review techniques described by Lincoln and Guba \[[@B25]\]. Participants were assured that their comments would be reported in a confidential manner, using pseudonyms in the actual group sessions, in transcripts and in research reports.
Results
=======
Analysis of the focus group responses revealed three primary themes in the ELDERSHINE program: 1) stress management; 2) learning, practicing and using mindfulness, and 3) social support. Participants in ELDERSHINE articulated perceived benefits of the program as well as their ability to acquire and transfer skills learned in the program to their everyday lives.
1) Stress management
--------------------
The first question that was asked in the focus group sessions was, \"What has ELDERSHINE meant to you?\" The primary theme that emerged across all of the focus groups was the ability to use meditation skills in coping with stressors in their life. These stressors included, growing older with some physical pain, medical tests, financial strain, and having children/grandchildren with significant mental, physical, financial or legal hardships. As an example of how ELDERSHINE participation had helped with stress management, one participant remarked:
\"And then with the meditation part of it, it just takes you away from your everyday concerns in life and for that, for that moment, you are in such a restfulness. You don\'t even know you\'re really sitting there in the chair sometimes, you\'re somewhere, wherever, \[you know\] the meditation might take you. And, I think that\'s what I like about coming to ELDERSHINE.
Another participant stated:
\"When I meditate it\'s like a soothing thing inside, and relaxing. Very relaxing. You don\'t think about, you don\'t actually hear nothing, you don\'t hear nobody. And it\'s, just like I said before, it\'s very soothing. It\'s better than taking a massage, put it to you that way. All the things that really \[bug\] and bother you at that particular time melts away and then your whole day is a good day because it doesn\'t, to me, actually, I don\'t have these pains until I maybe bend over to do something, but other than that, it don\'t bother you. It\'s just a good relaxing day. When you go to bed at night, you really can relax, you can sleep, have a very good peaceful night. And then, the next day, when you\'re not in ELDERSHINE, you can do the same thing then, but you might not do it at the same time, but at a different hour of the day.\"
Furthermore, the meditation was described as spiritual and related to God by this group. One participant remarked:
\"ELDERSHINE is a meditation of our spiritual \[lives\]. We might not mention God\'s name all the time, but He\'s there all the time. It is our spiritual self that\'s being nurtured.\"
This distinction between spirituality and religiosity is one that is often made much of in academic circles. However, from the perspective of our Protestant participants, people with traditional religious commitments are willing to learn and practice meditation without fear of supplanting their usual religious practices.
2) Applying meditation to their daily lives and stresses
--------------------------------------------------------
A second theme that emerged was that mindfulness meditation was a new practice and one which they were learning to apply to different areas of their lives. Participants revealed that they learned how to meditate and how to use meditation in their lives to reduce stress. One participant stated:
\"I had to go to the hospital for an MRI, and, I told Amy \[the program designer and interventionist\] I took her voice with me, going into the MRI you know, I love the Lord very much, and I\'m not taking away from God, but I took her with me with God into the MRI going through the tube and all the \"clunk, clunk, clunk\" noise you hear in there. I took Amy\'s voice with me to help, you know, help me, quiet me down and I wouldn\'t have to go through all that \[without it\].\"
Several other participants offered examples of ways they were applying the skills they were learning
\"And when you get depressed or you feeling sad, you meditate, it helps. I have experienced that the past week, you understand? And it\'s really helped me by meditating.\"
\"You can always learn something and learn how to meditate. Meditat\[ion\] is giving yourself a chance to think before you speak. Knowing what you\'re doing. It slows you down, \[to think\] of better things.\"
\"You catch yourself getting all upset or kind of worked up or disturbed a bit, so \[you say\] \'let me go on my break\"\... I have a special chair that I sit in now\.... All the thoughts and bad feelings you had all been melted away because your body is relaxed into your meditation. And then you come out of your meditation, you really feel like a really good person. You feel more relaxed and you can go about doing what you have to do.\"
3) Social Support
-----------------
An additional theme that emerged was that ELDERSHINE allowed participants to make friends and to share their lives. One participant described this theme in the following manner:
ELDERSHINE, it brings the different neighbors together and that they can talk and they can have some peaceful time together. Bringing people together and they feel comfortable talking about different things in life because they get to know the people in the circle.\"
In addition to sharing with each other, participants specifically mentioned sharing \"victories\" as an important component of ELDERSHINE. This is a time in which each participant describes something that brought them joy or that they accomplished during the week, even as simple as getting out of bed in the morning. One participant described this part of the weekly program when she stated:
\"You go in every week you come in and you talk about your victories, what \[you\] might have done that week, something you, in your life that was just a great victory to you. So you get to share in front of everybody your victory.\"
The building in which ELDERSHINE takes place is a high rise for low-income older adults and has no common spaces or common programs. A final theme that emerged was that being in this group increased their social support, specifically the number of people that they believed they could turn to for assistance or for whom they could provide support. One participant commented:
\"When I first came here, I didn\'t call nobody but my sister. My sister live on the seventh floor, When I \[don\'t know who\] I\'m gonna call, I call my sister. But, now I\'ve got plenty of people, I call them.\"
Another participant stated:
\"You come together, you know, and know your neighbors and talk to them, understand, you know? \[When they have\] something going on, they\'ve got problems too, it\'s really good, it\'s really good.\"
Participants reported feeling connection to and respected by others in the group. As one participant remarked,
\"I would tell them, it\'s a place to meet friends and share thoughts with one another.\"
Sharing victories and talking with each other were among the ways that the participants reported that they learned respect for one another, focused on others, and fostered valuable connections and enhanced communication skills. As one participant reported,
\"It teach\[es\] you how to respect each other, how to listen to each other and don\'t talk while somebody\'s talking.\"
Discussion
==========
In this focus group study with low-income minority older adult participants of a mindfulness-based program, we found that those who were in the intervention reported learning meditation skills, used it to cope with stressors in their lives, and increased their sense of community in a socially-isolating building.
These findings, though limited to female participants of one mindfulness program in one low-income housing project, are important. Because there is evidence that mindfulness programs may help in management of illness and low-income minority adults often suffer from multiple chronic health problems, our findings could suggest further research in an underserved population with this low-cost mindfulness-based intervention. Further, older adults are at a time in life when a reflective, stationary intervention, delivered in-residence could be an ideal mechanism to improve health. The demonstrated ability to incorporate the training into use with stressful medical tests and other challenging episodes during the course of day-to-day life shows that the benefits may reach beyond the in-class training.
Additionally, social support is important in both health promotion and during treatment for diseases like cancer, particularly among underserved populations. For example, informal and formal social support networks are believed to decrease barriers to cancer treatment in African-American older adults. In a study performed by Guidry et al, \[[@B26]\]African American cancer patients were more likely to report a need for formal and informal sources of support during treatment than were Caucasian Americans. Older African Americans may require enhanced and multi-faceted cancer support because they are known to possess generally poorer health status and fewer financial resources than their non-minority counterparts \[[@B27]\].
Future research should examine the specific psychological, social, and physiological effects of a mindfulness-based stress reduction program for older adults. This focus group study has revealed that program participants, in addition to being receptive and willing to participate in the program, enjoyed the meditation and mindfulness practices, continued to effectively utilize the techniques outside of the program, and gained social support from the program. All themes emerged in all three focus groups which encompassed the bulk of the data. There is some homogeneity of experience because of the demographic homogeneity, the common sharing of experience in the meditation group, and the fact that they all participants live in the same apartment building. Though the themes were broad, they did not warrant sub-categorization. Our team has also conducted a randomized clinical pilot trial of mindfulness meditation training with low- income minority older adults at a separate low-income housing building and with people who have never tried meditation \[[@B28]\].
A few limitations should be mentioned. The participants were only women. It is unknown whether men would have the same reactions to the ELDERSHINE program. However, low income minority communities of older adults are predominantly women. A second limitation is that the participants had been participating in ELDERSHINE for a range of time, some as many as three years. This is both a limitation and a strength. The limitation is that we cannot be sure these findings would apply to those who participated in the more typical 8 week mindfulness program. The strength is that, for the interviewed participants, the program is clearly relevant and sustainable demonstrated by the continued participation. A third limitation is the focus group participants knew each other. This may have limited negative comments due to social desirability bias \[[@B29]\]. A fourth limitation is that we did not stop the sampling due to data saturation but rather stopped due to near universality of focus group attendance by the ELDERSHINE participants. Because we offered the focus groups in the participants\' apartment building at a convenient time and offered food, the three focus groups included virtually everyone in the mediation group.
Conclusion
==========
In conclusion, this study has shown preliminary acceptability and perceived benefits evidence of a mindfulness-based stress reduction program for low-income minority older adults. Similar to research that has shown that mindfulness-based interventions may be effective to reduce stress, increase self-efficacy, and manage psychological and physical ailments among younger populations, study findings suggest that similar benefits may be achievable in older, more vulnerable populations.
Competing interests
===================
The authors declare that they have no competing interests.
Authors\' contributions
=======================
SLS provided some of the focus groups, collaborated on the analysis and drafted the manuscript. JW participated in the analysis and the writing of the manuscript. ABC provided the meditation groups and participated in the writing of the manuscript. RLP provided some of the focus groups, collaborated on the analysis and participated in the writing of the manuscript. All authors read and approved the final manuscript.
Pre-publication history
=======================
The pre-publication history for this paper can be accessed here:
<http://www.biomedcentral.com/1472-6882/11/44/prepub>
| {
"pile_set_name": "PubMed Central"
} |
Introduction
============
Argyrophilic grain disease (AGD) is a neurodegenerative condition associated with old age and characterized by the presence of abundant small spindle 4 to 8 μm argyrophilic inclusions (argyrophilic grains) in neurons of the amygdala, entorhinal cortex, hippocampus, and other limbic brain regions.[@b1-jen-7-2013-051]--[@b3-jen-7-2013-051] The neuropathological profile of AGD has been described in detail, and a highly homogeneous regional distribution of lesions allows a classification of cases according to a staging system.[@b4-jen-7-2013-051] Together with the characteristic argyrophylic grains (AG), all cases display a high density of diffusely tau-positive neurons (pretangles) and numerous ballooned neurons. Additionally, oligodendroglial tau inclusions (coiled bodies) and astrocytic inclusions (bush-like astrocytes) are invariably found in involved areas. Since argyrophylic grains and associated neuronal and glial inclusions are immunoreactive with tau antibodies,[@b5-jen-7-2013-051]--[@b7-jen-7-2013-051] AGD has been classified as a sporadic tauopathy. Tauopathies have been characterized by the presence, mainly in protein aggregates, of tau protein with either 3 or 4 repeats (3R or 4R, respectively) in the tau microtubule binding domain.[@b8-jen-7-2013-051] Tau aggregates from Alzheimer's Disease (AD) are composed of similar amounts of tau 3R and tau 4R isoforms.[@b9-jen-7-2013-051] In contrast, corticobasal degeneration (CBD) and progressive supranuclear palsy (PSP)[@b10-jen-7-2013-051] present an increase in the proportion of tau 4R, whereas Pick's disease (PiD) contains predominantly tau 3R isoforms in tau aggregates.[@b11-jen-7-2013-051] Previously, it has been suggested that AGD could be a tau 4R tauopathy.[@b2-jen-7-2013-051] Western blot analysis of tau isoforms in pure AGD has shown a double band at 64 and 69 kDa and a minor 74 kDa band, suggesting a predominance of 4R tau isoforms. Accordingly, AGD together with CBD and PSP may be considered collectively as 4R tauopathies. In fact, cases with mixed phenotypic features (eg, PSP plus CBD, or AGD plus PSP or CBD) are not infrequently observed.
AGD seems to be a sporadic disease. However, the presence of AGD pathology has been described in a patient bearing a mutation in tau gene, *MAPT* S305I.[@b12-jen-7-2013-051] Genetic risk factors for AGD show clear differences from other neurodegenerative dementias like Alzheimer's disease. A higher frequency of apolipoprotein E isoform 2 has been reported in AGD patients,[@b13-jen-7-2013-051],[@b14-jen-7-2013-051] although this point has been questioned.[@b15-jen-7-2013-051],[@b16-jen-7-2013-051] Additionally, it has been discussed whether the presence of *MAPT* haplotype H1 could be significantly increased in AGD.[@b17-jen-7-2013-051] As for other neurodegenerative dementias, aging appears to be the main risk for AGD, which is particularly prevalent in the oldest old. It has been estimated that one-third of centenarians may present AGD.[@b18-jen-7-2013-051]
Given that the presence of hyperphosphorylated tau is a dominant feature in the neuropathological and molecular profile of all tauopathies and that most probably it plays a central role in pathogenesis, some clues to the specific phenotype of AGD may be related to its putative molecular signature. However, the interpretation of molecular studies in AGD cases has been hampered by several limitations: (1) most often AGD is found in combination with other neurodegenerative pathologies and is particularly associated with other 4R tauopathies; (2) AGD involves predominantly the medial temporal lobe, a brain region where pathology of the Alzheimer's type is highly prevalent in old subjects; and (3) pure AGD cases are rare and show atypical clinical features, such as rapid progression, and postmortem studies are frequently performed under a clinical suspicion of Creutzfeldt-Jakob disease (CJD). Availability of fresh frozen tissue for molecular studies is usually limited when high-risk neuropathological autopsies are performed.
Now, in this work we have analyzed the composition of the main tau isoform present in tau aggregates obtained from a series of well-characterized cases of AGD presenting either alone or combined with other neurodegenerative conditions.
Materials and Methods
=====================
Cases
-----
Sixteen patients with a neuropathological diagnosis of AGD (neuropathological stage II or III) were included. Age at death ranged from 51 to 98 years and male to female ratio was 1:1.3 (see [Table 1](#t1-jen-7-2013-051){ref-type="table"}). AGD was the main neuropathological diagnosis in 6/16 patients, while this pathology was combined with Parkinson's disease (PD) (2 cases), amyotrophic lateral sclerosis (ALS) (2 cases), CBD (1 case), PSP (2 cases), AD (1 case), Huntington's disease (HD) (1 case) and vascular dementia (VaD) (1 case) in the rest of patients. Some degree of Alzheimer's type neurofibrillary degeneration was present in all cases ([Table 1](#t1-jen-7-2013-051){ref-type="table"}). However, in order to avoid possible contamination of results by the presence of this type of tauopathy in isocortical areas, only cases with Braak stage lower than V were included in the study. Dementia was present in 12/16 cases, and only patients with a main final diagnosis of PD or ALS (cases 9 to 12) had not developed overt dementia. In 5/6 cases where AGD was the main diagnosis, a rapidly progressive neurological disease was the presenting form, so that the postmortem study in this cases was performed with a suspected clinical diagnosis of CJD. Postmortem studies were performed either at the Fundación CIEN Tissue Bank or the Hospital Universitario de Alcorcón, Madrid (Reference center for prion diseases in Spain). All postmortem studies were performed complying with national ethical and legal regulations.
Neuropathology
--------------
In all cases, postmortem examination was limited to the cranial cavity (with spinal cord extraction only in cases with ALS) with a postmortem interval shorter than 12 hours. According to the brain bank protocol (shared by both brain banks involved), in conventional donation cases, immediately after extraction, the right half of the brain was sliced and frozen, while the left half was fixed by immersion in phosphate-buffered 4% formaldehyde for at least 3 weeks. High-risk autopsies for suspected CJD were performed following international recommendations. In these cases, fresh tissue blocks from the right frontal and parietal lobes and the ipsilateral cerebellar hemisphere were obtained for freezing. All frozen tissue was stored at −80°C. After fixation, a full neuropathological study was performed in the left half brain. For that purpose, 25 tissue blocks were obtained from cortical and subcortical brain regions for paraffin embedding. Neuropathological classification of cases was based on a thorough examination of hematoxylin-eosin stained paraffin sections of all blocks, and immunostaining with a panel of antibodies (beta-amyloid, tau AT100, alpha-synuclein, ubiquitin, prion protein, and TDP-43) in selected sections. Consensus criteria were used for disease diagnosis and staging.
Further morphological study of AG pathology was performed in 2 tissue blocks comprising (1) amygdala and (2) anterior hippocampus. Paraffin sections of these blocks were studied with Gallyas stain and with immunostaining for a panel of antibodies (tau AT8, tau AT100, tau 7.51, RD3, RD4, neurofilaments SMI 1, p62). The presence and intensity of various tau-reactive inclusions in several regions was assessed on a semiquantitative basis on AT8-immunostained sections ([Fig. 3A](#f3-jen-7-2013-051){ref-type="fig"}). For that purpose, the frequency of AG was graded on a 0 to 3 scale. Other tau-immunoreactive inclusions associated to AG, principally pretangles and coiled bodies, show parallel variations in frequency (data not shown). For the evaluation of ballooned neurons anti-tau antibodies and antibodies recognizing hyperphosphorylated neurofilaments (SMI 1) (1:100) (Sternberger Monoclonals, Covance Inc., Princeton, New Jersey) were employed.
Antibodies
----------
The following anti-tau antibodies were used ([Table 2](#t2-jen-7-2013-051){ref-type="table"}): 7.51(1:100)[@b19-jen-7-2013-051] (a kind gift of Claude M. Wischik, MRCPsych, Newcastle University, UK). AT-8 (1:100) and AT-100 (1:100) were from Innogenetics (Gent, Belgium). AT-8 recognizes tau when Ser 199/202 are phosphorylated.[@b20-jen-7-2013-051],[@b21-jen-7-2013-051] AT-100 recognizes tau when Ser212 and Thr-214 are phosphorylated. Antibody 7.51 recognizes a region containing a part of the microtubule-binding domain of human and murine tau.
PHF-1 (1:100), an antibody that reacts with tau when Ser396 and Ser404 are phosphorylated, was a kind gift of Peter Davies, Ph.D. (Albert Einstein University, NY, USA). Exon-2 (1:100) antibody was raised against human tau peptide comprising the residues 55 to 67 (numbering of the longest human tau isoform,[@b20-jen-7-2013-051] and it was obtained through Abyntek Biopharma S.L., (Derio, Spain) for our laboratory (R. Cuadros et al, unpublished data, 2011). Finally, anti-tau clone 8E6/C11 (1:500) and anti-tau clone 1E1/A6 (1:500), specific for reacting with tau 3R and tau 4R isoforms were from Merck Millipore, Billerica, Massachusetts.
Immunohistochemistry
--------------------
Immunohistochemical stains were performed on 5 μ paraffin sections by using the ABC method (Pierce Antibodies, Thermo Fisher Scientific Inc., Waltham, Massachusetts), diaminobenzidine (DAB) as developer, and Carazzi's hematoxylin as counterstain. Optimal dilutions and times and conditions of incubation of primary antibodies were applied as recommended by the suppliers. Unmasking of antigens was performed when necessary.
Preparation of brain tissue extracts
------------------------------------
Brain cell extracts were isolated by homogenizing the brain tissue in ice-cold buffer (1:10 w/v) consisting of 10 mM Tris pH7.4, 1 mM EGTA, 0.1 M NaCl and 10% sucrose plus phosphatase inhibitors (10 mM NaF, 1 mM sodium orthovanadate), and protease inhibitors (2 mM PMSF, 10 μg/mL aprotinin, 10 μg/mL leupeptin, and 10 μg/mL pepstatin).
Isolation of detergent insoluble tau aggregates
-----------------------------------------------
Tau aggregates from brain tissue samples of the patients were isolated as previously described for PHF-tau from Alzheimer disease patients by Greenberg and Davies.[@b19-jen-7-2013-051],[@b21-jen-7-2013-051] Different sources were used to look for tau aggregates: temporal (including medial temporal lobe), frontal, and parietal cortex.
Detergent-insoluble tau aggregates were visualized by electron microscopy.[@b22-jen-7-2013-051]
Western blot analysis
---------------------
Extracts for Western blot analysis were prepared by homogenizing the isolated tau aggregates samples. The samples were homogenized at 4°C and protein content determined by the Bradford method (Bio-Rad laboratories Inc., Hercules, California). Total protein (10 μg) was electrophoresed on 8% SDS-PAGE gel and transferred to a nitrocellulose membrane (Schleicher and Schuell). The experiments were performed using the following primary antibodies: AT8, 7.51, PHF-1, AT100, Tau 3R, and Tau 4R. The filters were incubated with the antibody at 4°C overnight in 5% nonfat dried milk. A secondary HRP-conjugated goat antimouse (Gibco, Life Technologies Corp., Carlsbad, California) followed by ECL detection reagents (Amersham, GE Healthcare Bio-Sciences, Piscataway, New Jersey) were used for immunodetection.
Transmission electron microscopy
--------------------------------
Immunoelectron microscopy was performed after adsorption of the samples to electron microscopy carbon-coated grids and incubation with the first antibody (AT8 or PHF-1) for 1 hour at room temperature. After extensive washing, the grids were incubated with the secondary antibody (1/40) conjugated with 10 nm diameter gold particles. Finally, the samples were stained with 2% uranyl acetate for 1 minute. Transmission electron microscopy was performed in a JEOL model 1200EX electron microscope operated at 100 kV. Electron micrographs were obtained at a magnification of × 50,000.
Apolipoprotein E (APOE) and H/H genotyping
------------------------------------------
Total DNA was isolated from peripheral blood or cerebral tissue following standard procedures. Genotyping of *MAPT* rs1052553 polymorphism, which discriminates between *MAPT* H1 and H2 haplotypes, was determined using TaqMan probes (C---7563736_10 assay, Applied Biosystems, Life Technologies Corp., Carlsbad, California) according to manufacturer instructions. *APOE* genotyping (rs429358 and rs7412) was performed by real-time polymerase chain reaction (PCR).[@b23-jen-7-2013-051]
Results
=======
Grains are stained by an antibody that recognizes phospho-tau
-------------------------------------------------------------
Histopathological analysis ([Fig. 1A](#f1-jen-7-2013-051){ref-type="fig"}) indicates the presence of characteristic silver stained grains in the brain of the patients. When immunofluorescence analyses were carried out using the AT8 antibody, which recognizes phospho-tau, a clear reaction with the grains was observed ([Fig. 1B](#f1-jen-7-2013-051){ref-type="fig"}). Furthermore, immunoelectron microscopy analyses indicate the presence of phospho-tau in grains and filamentous tau aggregates ([Fig. 1C](#f1-jen-7-2013-051){ref-type="fig"}). These filamentous polymers were also stained when another antibody, reacting with a different tau phosphoepitope (PHF- 1), was used. [Figure 1D](#f1-jen-7-2013-051){ref-type="fig"} shows the filaments in the absence (left and middle) or presence (right) of anti-PHF-1 staining ([Fig. 1D](#f1-jen-7-2013-051){ref-type="fig"}). Nevertheless, the proportion of AGD filamentous polymers found in a brain region such as the temporal lobe is much lower than that found in the same region in other tauopathies such as AD ([Fig. 2](#f2-jen-7-2013-051){ref-type="fig"}).
Brain regions containing argyrophylic grains
--------------------------------------------
In all cases and all regions examined, AG were highly associated with other histological lesions characteristic of AG pathology: ballooned neurons, diffusely tau-reactive neurons (pretangles), neurofibrillary tangles, coiled bodies, and tau-reactive astrocytes, frequently showing bush-like morphology. A consistent pattern of regional distribution of AG pathology was identified both in cases with a principal diagnosis of AGD (patients 1 to 6) and in cases with AGD combined to different neurodegenerative conditions, including other sporadic tauopathies (patients 7 to 16). Areas most intensely involved in the medial temporal lobe were the amygdaloid complex (predominantly the basolateral nuclei), gyrus ambiens, entorhinal and perirhinal cortex, and CA1 sector of the hippocampal cortex. In contrast, CA4 sector and the dentate gyrus were proportionally less affected by AG pathology. Most cases (14/16) were classified as stage III AGD, and cases with a main diagnosis of AGD displayed more intense pathology in the hypothalamus, nucleus accumbens and brainstem tegmentum ([Fig. 3A](#f3-jen-7-2013-051){ref-type="fig"}).
Characterization of the main tau isoform present in AGD tau aggregates
----------------------------------------------------------------------
Six tau isoforms are present in the human central nervous system (CNS) ([Fig. 4A](#f4-jen-7-2013-051){ref-type="fig"}) (see Goedert [@b24-jen-7-2013-051] and Goedert et al[@b25-jen-7-2013-051]). Characterization of tau isoforms present in aggregates, such as the paired helical filaments found in Alzheimer disease, reveals, by electrophoresis and Western blot analysis, the presence of 3 main bands such that each corresponds to two different phospho-tau isoforms ([Fig. 4B](#f4-jen-7-2013-051){ref-type="fig"}).
When tau isoforms present in AGD aggregates were analyzed by electrophoresis/Western blot, mainly a band with a relative electrophoretic mobility corresponding to a protein of 64K was observed. [Figure 4C](#f4-jen-7-2013-051){ref-type="fig"} shows the electrophoretic patterns of tau aggregates from different AGD patients using 7.51 antibody. In all cases, the 64K band was the most abundant band, which could result from the addition of the 2 tau isoforms containing exon 2, one of them containing and the other one lacking exon 10 ([Fig. 4D](#f4-jen-7-2013-051){ref-type="fig"}). When Western blot was carried out to check for the presence or absence of exon 10, using specific antibodies that recognize tau 3R or tau 4R isoform, only tau 4R isoform was found in the 64K band for AGD aggregates. In addition, this band also reacts with an antibody raised against residues 55 to 67 (nomenclature of the largest tau isoform present in the CNS) located at exon 2 ([Fig. 4D](#f4-jen-7-2013-051){ref-type="fig"}). This pattern was consistently found in patients with pure AGD and in cases with AG pathology combined with other neurodegenerative conditions, including 4R tauopathies. Thus, we conclude that the major tau isoform present in AGD aggregates is that containing exon 2 and exon 10 and lacking exon 3.
Status of APOE and H/H polymorphisms
------------------------------------
Genotyping for APOE and *MAPT* H/H haplotypes was performed in 12/16 cases ([Table 1](#t1-jen-7-2013-051){ref-type="table"}). All patients in this study were APOE e3/e3 (10/12, 83,3%) or APOE e3/e4 (2/12, 16,7%). No e2 cases were observed. These findings are in contrast with previous reported series based in central European populations where a significantly higher proportion of AGD cases contained the e2 haplotype.[@b13-jen-7-2013-051],[@b14-jen-7-2013-051],[@b16-jen-7-2013-051] Noteworthy, a large series based in a North American population did not confirm this association.[@b15-jen-7-2013-051] Although the present series includes a proportionally shorter number of cases, the predominance of the e3 haplotype, with a high proportion of homozygosis, may depend on the higher prevalence of this haplotype in southern European populations.[@b26-jen-7-2013-051] In the present series, 9 out of 12 cases (75%) contained the H1/H1 haplotype at the *MAPT* gene, while 3/12 cases were H1/H2. These results agree with previous studies, where a high proportion of H1/H1 haplotype has been described in association with sporadic 4R tauopathies, including AGD.[@b17-jen-7-2013-051]
In general, it is not clear how APOE isoforms or *MAPT* haplotypes might condition the development of a tauopathy like AGD. Clinical subtypes of other tauopathies like frontotemporal lobar degeneration (FTLD-Tau), display different tau and APOE genotype frequencies.[@b27-jen-7-2013-051]
On the other hand, APOE isoforms may regulate tau kinases and phospho-tau levels in neurons.[@b28-jen-7-2013-051] Additionaly, *MAPT* H1 haplotype is more efficient at driving gene expression than the H2 haplotype, and this suggests that when H2 haplotype is present a lower amount of tau protein is expressed.[@b29-jen-7-2013-051] Thus, differences in APOE isoforms and *MAPT* haplotypes may regulate the amounts of phospho-tau and total tau, two important parameters in the onset of tauopathies. How both features could affect the development of AGD remains under discussion, as previously indicated.
Discussion
==========
AGD may manifest with dementia and behavioral abnormalities such as emotional and mood imbalance.[@b3-jen-7-2013-051],[@b30-jen-7-2013-051] These behavior disorders may correlate with the appearance of tau pathology (silver stained or argyrophylic grains) at hippocampal formation and amygdala regions. On the other hand, temporal-spatial spreading of tau aggregates has been described in AGD.[@b4-jen-7-2013-051] In this work, we have found a regular pattern of involvement of anterior medial temporal lobe regions by AG pathology with further extension to limbic cortices, certain subcortical nuclei, and brainstem tegmentum. However, in our series, AG pathology usually spares the dentate gyrus, a region that could be rich in tau 3R isoforms due to the presence of adult neurogenesis. Regional distribution of pathology was found to be comparable in pure AGD cases and in AGD pathology associated with other neurodegenerative conditions but with a higher intensity of involvement of extratemporal regions in pure cases. We also found that tau aggregates obtained from patients with AGD are composed predominantly by a tau isoform containing exon 2 and exon 10 and lacking exon 3. This finding was consistent in patients (1) with pure AGD, (2) with AGD associated with another 4R tauopathy, and (3) with AGD combined with other neurodegenerative diseases with low Braak stages for Alzheimer's type neurofibrillar degeneration. Previous studies[@b31-jen-7-2013-051]--[@b33-jen-7-2013-051] have reported a double-band of tau at 64K and 69K in AGD, thus including AGD in the 4R tauopathies group together with PSP and CBD. Our findings suggest that AG pathology may be characterized by a specific pattern of tau isoforms expressed in Western blot by a predominant band at 64K.
Recently, it has been shown that the exon 2 and exon 10 inserts increase tau aggregation propensity, whereas the exon 3 insert decreases tau aggregation.[@b34-jen-7-2013-051] Also, the presence of H2 haplotype together with the expression of exon 3 seems to play a protective effect against the development of some tauopathies.[@b35-jen-7-2013-051],[@b36-jen-7-2013-051] However, in our work we found that one-fourth of our AGD patients show the H2 haplotype. Although the number of cases studied in this series is too low to draw strong conclusions, our data indicate a lower proportion of H2 haplotype compared with normal controls in a similar population.[@b37-jen-7-2013-051]
In summary, we have shown that tau aggregates of AGD enclose as a major component the phosphorylated tau isoform containing exon 2 and 10 and lacking exon 3. Phosphorylation of tau occurs, at least, in those sites recognized by antibodies AT8, AT100, and PHF-1. Despite this posttranslational modification, other modifications have been described for tau protein, and recently, one of these modifications, acetylation, appears to be absent in AGD,[@b38-jen-7-2013-051] whereas it is present in other tauopathies like Alzheimer's disease.
**COMPETING INTERESTS:** Authors disclose no potential conflicts of interest.
**Author Contributions**
Conceived and designed the experiments: AR, FH, JA. Analyzed the data: AR, MC, JA. Wrote the first draft of the manuscript: AR, RC, JA. Contributed to the writing of the manuscript: MC, FH. Agree with manuscript results and conclusions: AR, RC, MC, FH, JA. Jointly developed the structure and arguments for the paper: AR, JA. Made critical revisions and approved final version: AR, FH, JA. All authors reviewed and approved of the final manuscript.
**DISCLOSURES AND ETHICS**
As a requirement of publication the authors have provided signed confirmation of their compliance with ethical and legal obligations including but not limited to compliance with ICMJE authorship and competing interests guidelines, that the article is neither under consideration for publication nor published elsewhere, of their compliance with legal and ethical guidelines concerning human and animal research participants (if applicable), and that permission has been obtained for reproduction of any copyrighted material. This article was subject to blind, independent, expert peer review. The reviewers reported no competing interests.
**FUNDING:** This work was supported by grants from Spanish Plan Nacional (SAF2011-24841), CIBERNED and Reina Sofia Foundation.
{#f1-jen-7-2013-051}
{#f2-jen-7-2013-051}
{#f3-jen-7-2013-051}
![**A**. Scheme of the 6 tau isoforms that are expressed in human central nervous system.[@b24-jen-7-2013-051],[@b25-jen-7-2013-051] **B**. Phosphorylated tau protein present in tau aggregates from Alzheimer's disease patients and their correspondence with different phospho-tau isoforms.[@b24-jen-7-2013-051] **C**. phospho-tau isoforms (found by Western blot using ab7.51) present in tau aggregates isolated from patients: 1 (pure AGD), 8 (AGD + PSP), 12 (AGD + ALS), and 13 (AGD + HD). The presence of a main protein with an electrophoretic mobility of 64K is shown in each case. **D**. Phospho-tau isoforms present in tau aggregates from patient 9 (PD + AGD). Mainly a 64K protein was also found with 7.51 antibody (a similar picture was found when ab AT-100 was used). It may correspond to the mixture of the indicated 2 tau isoforms (isoform 2 and 5, Fig. 4B). Since that 64K protein is not reacting with tau 3R isoforms, the proposed 64K tau isoform is that containing exons 2 and 10 and lacking exon 3.](jen-7-2013-051f4){#f4-jen-7-2013-051}
######
Main characteristics of the patients analyzed in this study.
PATIENTS NP DIAGNOSIS AGD STAGE BRAAK STAGE APOE H/H DEM AGE AT DEATH GENDER
---------- -------------- ----------- ------------- ------- ------- ----------------------------------------------------- -------------- --------
1 AGD II II ɛ3/ɛ3 H1/H1 \+ 76 M
2 AGD III IV ɛ3/ɛ3 H1/H1 \+[(\*)](#tfn1-jen-7-2013-051){ref-type="table-fn"} 88 F
3 AGD III II ɛ3/ɛ3 H1/H1 \+[(\*)](#tfn1-jen-7-2013-051){ref-type="table-fn"} 70 F
4 AGD III II ɛ3/ɛ4 H1/H2 \+[(\*)](#tfn1-jen-7-2013-051){ref-type="table-fn"} 81 M
5 AGD III II ɛ3/ɛ3 H1/H2 \+[(\*)](#tfn1-jen-7-2013-051){ref-type="table-fn"} 80 M
6 AGD III IV ND ND \+[(\*)](#tfn1-jen-7-2013-051){ref-type="table-fn"} 74 F
7 AGD + CBD III IV ɛ3/ɛ3 H1/H1 \+ 79 F
8 AGD + PSP III III ND ND \+ 77 F
9 AGD + PD II II ɛ3/ɛ4 H1/H1 − 79 M
10 AGD + PD III IV ɛ3/ɛ3 H1/H1 − 80 M
11 AGD + ALS III II ɛ3/ɛ3 H1/H2 − 73 M
12 AGD + ALS III III ɛ3/ɛ3 H1/H1 − 70 F
13 AGD + HD III III ND ND \+ 51 M
14 AGD + Vasc III III ɛ3/ɛ3 H1/H1 \+ 83 F
15 AGD + Vasc III II ND ND \+ 77 F
16 AGD + AD III IV ɛ3/ɛ3 H1/H1 \+ 98 F
Rapidly progressive dementia, studied postmortem as suspected CJD.
Tau haplotypes (H1 and H2) and APOE isoforms (E3 or E4) are indicated.
**Abbreviations:** M, male; F, female; NP diagnosis, neuropathological diagnosis; AGD, argyrophylic grain disease; PD, Parkinson disease; CBD, corticobasal degeneration; PSP, Progressive Supranuclear Palsy; ALS, amyotrophic lateral sclerosis; HD, Huntington disease; Vasc, cerebral vascular disease; AD, Alzheimer's disease; Dem, dementia.
######
Tau antibodies used in this study. Localization of the antibodies that react with (**A**) phospho-tau or (**B**) with tau protein regardless of its phosphorylation status, in tau molecule.
| {
"pile_set_name": "PubMed Central"
} |
As of April 12, 2020, the Coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome (SARS)--CoV-2 virus has affected more than 550,000 Americans, and claimed more than 21,000 lives.[@bib1] ^,^ [@bib2] In just 2 months, the number of deaths is projected to reach over 60,000 --- despite current levels of social distancing and other preventive measures.[@bib3] ^,^ [@bib4] Health care workers are placing their lives at risk, and are facing enormous physical and emotional stress. The shortage of testing, masks, other personal protective equipment, and ventilators threatens to make our predicament worse. But unfortunately, we are already in the midst of another devastating problem: the shortage of medications that are critical for the management of COVID-19.TableImmediate Actions Needed to Boost Medication Supply and Ensure Equitable AccessBoost medication supplyMinimize cost•Mobilize any state or federal government stockpiles of critical medications to hospitals experiencing surges in patients.•Use a "Netflix" option for new antiviral drugs and vaccines.•Institute a 24-h turnaround for approving the importation of generic drugs or grant immediate reciprocal approval for selected generic drugs manufactured in other countries.•Authorize Medicare to negotiate prices for COVID-19--related drugs.•Work with major brand-name and generic pharmaceutical companies to boost mass manufacturing of approved drugs in short supply.•Be willing to issue compulsory licensing for medications.•Encourage and fund nonprofit generic drug manufacturing in the United States.•Prohibit deductibles and rebates for COVID-19--specific drugs.•Increase the domestic manufacturing of active pharmaceutical ingredients.•Establish centralized systems to track need, based on patient caseload, to inform manufacturing needs and shift supply rapidly and equitably.
Medication shortages, anticipated to worsen with time, affect patients with COVID-19 directly, but also pose a threat to the health and safety of patients with other diseases who do not have COVID-19. Shortages include medications that have been touted as promising therapies against COVID-19, such as chloroquine and hydroxychloroquine.[@bib5] In fact, due to off-label prescriptions and hoarding, hydroxychloroquine is now difficult to obtain for patients without COVID-19 who need this drug to manage rheumatoid arthritis and other autoimmune disorders. Even more worrisome is the shortage of sedatives such as midazolam and propofol which are needed for patients who are being intubated and placed on mechanical ventilation.[@bib6] This impacts patients seriously ill with COVID-19, as well as patients with respiratory failure due to other causes who need critical care, and those who need emergency surgeries under general anesthesia. The active pharmaceutical ingredients for many commonly used medications come from China, and many of our generics are manufactured and imported from other countries, including India. Since the pandemic has affected manufacturing and exports worldwide, the shortages exemplified by antimicrobials and sedatives will soon impact numerous medications unrelated to the treatment of COVID-19.[@bib7]
Recommendations to Boost Drug Supply {#sec1}
====================================
We are exceeding capacity in terms of our vital medication supply. The usual supply chains, mechanisms, and administrative processes are inadequate for this crisis. We must act now with urgency. The US Food and Drug Administration (FDA) must revamp its regulatory procedures and dramatically accelerate its processes to ensure that important medications are available to the public. Many of the drug shortages are going to be generic drugs. Because of various barriers for some generic drugs, we have only one or two FDA-approved versions available for sale in the United States. In contrast, in some countries in Europe and in India, each brand-name drug may have multiple generic versions made by different manufacturers. The multiplicity of manufacturers acts as a protection against shortages in these countries, and lowers cost.[@bib8] Here are the steps we as a nation must take ([Table](#tbl1){ref-type="table"}):
First, we recommend that the FDA institute a 24-hour turnaround for approving the importation of generic drugs that are in short supply from established manufacturers in other countries. Alternatively, the FDA can also grant immediate reciprocal approval for selected generic drugs manufactured in other countries.[@bib9] Not having access to a drug must be considered far worse for the public the agency is charged to protect than any safety concerns that arise with the use of a generic drug that is made by an established manufacturer hitherto not authorized to sell in the United States.
Second, the federal government must work with major brand name and generic pharmaceutical companies to boost mass manufacturing of approved drugs that are in short supply. If General Motors can repurpose its factories to make ventilators, then big pharmaceutical companies can be requested to divert their factories to manufacture generic drugs that are facing shortages. If companies that have the infrastructure are not willing to do this voluntarily in the national interest, then the administration must invoke the Defense Production Act to compel them to meet the needs of the country.
Third, we must encourage and fund nonprofit generic drug manufacturing in the United States. Civica Rx, for instance, was founded with the goal of overcoming drug shortages and to manufacture and supply at reduced cost. This concept needs to be expanded. Although it will take time to set up, government manufacturing of essential drugs is something we must consider long-term because generic drugs with small profit margins are unattractive to for-profit companies despite critical need.
Fourth, in addition to negotiating with China to increase supply, the United States must consider increasing the domestic manufacture of active pharmaceutical ingredients.[@bib10] Since factories closed in China, and India shut down exports of medications, we have been under the threat of shortages in multiple medications.[@bib11] ^,^ [@bib12] There have since been some signs of improvement, but the situation is fluid, and may worsen. This pandemic reminds us of the importance of having backup manufacturing within the United States, supported by taxpayer funds if needed.
Finally, any state or federal government stockpiles of critical medications should be mobilized, now, to hospitals experiencing surges in patients. Delay in providing these to cities in need will cause major harm and loss of trust. For critical hospital medications, the government should establish centralized systems to track need based on patient caseload to inform manufacturing requirements, and to shift supply rapidly and equitably. States and cities should be willing to share supplies of vital medications with other states as their own needs diminish when disease control is achieved within the state.
Recommendations to Control Cost {#sec2}
===============================
Even as we increase the amounts of available drugs, we must take steps to keep costs down ([Table](#tbl1){ref-type="table"}). As we have witnessed with hand sanitizers and masks, if supplies are limited, prices go up dramatically. In the case of prescription drugs, we already have major problems controlling price, even before this pandemic.[@bib13] If we fail to take strict measures, access to life-saving medications will be limited by cost, and this will disproportionately affect uninsured people and the elderly --- the ones most vulnerable to COVID-19. Making treatment (and vaccines, as they become available) accessible and affordable is therefore critical.
We must pursue contracts with manufacturers especially for new antiviral drugs and vaccines for a "Netflix" option, in which they receive a fixed reimbursement for an unlimited supply.[@bib14] The Netflix model has been very successful in providing hepatitis C drugs to the State of Louisiana. We also recommend that laws be enacted to immediately authorize Medicare to negotiate price for COVID-19--related drugs. There has always been overwhelming public support for Medicare negotiation, and this is something we must do now more than ever because it can help keep prices lower.[@bib15] As a last resort, the federal government should be willing to issue compulsory licensing if negotiations with companies fail to reach agreement on a reasonable price.[@bib16] Compulsory licensing permitted by the World Trade Organization Doha declaration in 2001 allows governments to license the use of a patented invention to a third party or government agency without the consent of the patent-holder if negotiations with the patent owner are unsuccessful.[@bib17]
Another major factor that can increase the prices of COVID-19 related drugs in the United States are pharmacy benefit managers (PBMs), who normally act as middlemen between drug manufacturers, pharmacies, and insurers. PBMs are supposed to deliver value to insurers and the insured public by identifying the best options for formulary placement and negotiating for lower prices. However, this is not always the case. PBMs may favor expensive drugs over inexpensive alternatives. The role PBMs play in contributing to increased prescription drug costs is best illustrated by insulin prices that have increased several fold over the past 10 years even for older insulin preparations.[@bib13] The resulting higher retail prices disproportionately affect uninsured and under-insured people. To prevent these problems, the federal government must prohibit deductibles and rebates for COVID-19--specific drugs.
Summary {#sec8}
=======
Similar to many at-risk resources in this crisis, medication shortages have been invisible but are threatening not only our ability to overcome this pandemic, but also the health of patients who have other diseases. The fragility of the complex medication supply chain was not created overnight, and yet we must overcome its limitations rapidly and decisively. And we must act before it is too late.
**Grant Support:** Supported in part by grants CA 107476 and CA 168762 (SVR) from the 10.13039/100000054National Cancer Institute, Rockville, MD; and 1R01DA044284-01A1 (EKC) from the National Institute on Drug Abuse, Bethesda, MD.
**Potential Competing Interests:** The authors report no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
1. Introduction {#sec1-ijerph-17-02376}
===============
Since the reform and opening up in 1978, China's economy has developed rapidly, people consume a variety of food items, and they are using more meat and dairy products. At the same time, especially after China's accession to the World Trade Organization (WTO) in 2001, there was a sharp increase in food safety risks and many food safety incidents in the country \[[@B1-ijerph-17-02376]\]. The data obtained by a big data-mining firm show the mainstream network media reported about 408,000 food safety incidents in China from 2008 to 2017 with an average of 112 food safety incidents per day \[[@B2-ijerph-17-02376]\]. One of the important reasons for the outbreak of a series of food safety incidents in China is the shortcoming of food safety laws, regulations, and supervision systems \[[@B3-ijerph-17-02376],[@B4-ijerph-17-02376]\]. From 2003 to 2012, China implemented a multisectoral segmented food safety supervision, which seriously reduced the efficiency of food safety supervision \[[@B1-ijerph-17-02376]\]. Food Safety Law of the People's Republic of China implemented in 2009 did not fundamentally improve the flaws in China's food safety supervision system. Food safety incidents not only restrict the healthy development of China's food industry, damaging the international image of China, but also attack people's confidence in food safety \[[@B3-ijerph-17-02376]\]. To improve food safety, the Chinese government has implemented a series of policies and measures, such as reforming relevant laws, strengthening major surveillance, and monitoring systems \[[@B5-ijerph-17-02376]\]. After 2013, a unified food safety supervision system was formed after a new round of food safety supervision mechanism reform. In October 2015, the newly revised Food Safety Law of the People's Republic of China, known as the most stringent food safety law in Chinese history was officially promulgated and implemented \[[@B4-ijerph-17-02376]\]. The food safety strategy has risen to become one of the major national strategies (website: <http://www.gov.cn/zhengce/2019-05/20/content_5393212.htm>). The overall situation of food safety in China has been improving in recent years. However, compared with the United States, the European Union, and other developed countries or regions, China's food safety regulatory system still needs to be further improved. Food safety concerns in China are still severe, and China is still facing threats and challenges from many food safety risk factors \[[@B3-ijerph-17-02376]\]. Food safety is still one of the extremely serious issues of public concern in China and the public perceive lots of food safety risk and anxieties \[[@B6-ijerph-17-02376]\].
Consumers are the final link in the food supply chain, their attention, and understanding of food safety information can promote producers and retailers to pay more attention to the control of food quality and safety, which is also an indispensable part in food safety governance. Therefore, the research on consumers' food safety risk information-seeking behavior is crucial. Information asymmetry is the root cause of the food safety problem \[[@B7-ijerph-17-02376],[@B8-ijerph-17-02376],[@B9-ijerph-17-02376],[@B10-ijerph-17-02376]\]. The quality and safety of food have the property of the trust. Although consumers cannot completely identify the quality and safety of the product, they could minimize the degree of information asymmetry by searching for food quality and safety information. In the research of risk-faced decision-making and behavior, Griffin et al. proposed the Risk Information Seeking and Processing (RISP) model based on the Heuristic-Systematic Model (HSM) of information processing and Theory of Planned Behavior (TPB) \[[@B11-ijerph-17-02376]\]. The RISP model has been applied in many fields \[[@B12-ijerph-17-02376]\], such as to understand people's antinuclear behavioral intentions \[[@B13-ijerph-17-02376]\], risk of drinking tap water drawn from the lakes \[[@B14-ijerph-17-02376]\], risk of flooding, global warming or climate change \[[@B15-ijerph-17-02376],[@B16-ijerph-17-02376],[@B17-ijerph-17-02376]\], disease health-related risk \[[@B18-ijerph-17-02376],[@B19-ijerph-17-02376]\], and food-related risk \[[@B20-ijerph-17-02376]\]. The general adaptability of the RISP model has been widely verified. Food safety risks are diverse and are very closely related to people's life. There is a limited study related to food safety risk based on the RISP model. Additionally, there are limited studies to validate Chinese consumers' risk-seeking behavior related to food safety, especially the risk-seeking behavior of the consumers based on their characteristics. In addition, WeChat is one of the most important social media platforms in China, and its number of monthly active users has soared from 0.36 billion in 2013 to 1.15 billion as of June 2019, ranking third in the world after WhatsApp and Facebook. According to the WeChat Influences Report 2018, the information consumption driven by WeChat reached RMB 209.7 billion, accounting for 4.7% of China's total information consumption, and it also accounts for 34% of China's mobile traffic consumption.
Under such a circumstance, this paper analyzes the food safety risk information-seeking intention and the influencing factors of consumers based on the RISP model and Multigroup Structural Equation Modeling (SEM) analysis by using WeChat users' questionnaire survey. Our goal is to understand the impact of different demographic characteristics on consumers' behavior. The application of the RISP model to the research of consumers' risk-seeking behavior in food safety is of considerable significance not only to promote its theoretical use in the field of food safety but also to improve the ability of the consumers to overcome food safety risks.
2. Theoretical Framework and Hypotheses {#sec2-ijerph-17-02376}
=======================================
Individual characteristics, risk perception, affective response, information insufficiency, informational subject norms, relevant channel beliefs, and perceived information gathering capacity affect an individual's risk information collection process \[[@B11-ijerph-17-02376]\]. These are the components in the RISP model, which are shown in [Figure 1](#ijerph-17-02376-f001){ref-type="fig"}. In the succeeding paragraphs, we outline these components and their role in the risk information collection process.
2.1. Risk Perception {#sec2dot1-ijerph-17-02376}
--------------------
Risk refers to the possibility and consequences of adverse events \[[@B21-ijerph-17-02376],[@B22-ijerph-17-02376]\], which is related to the degree of future losses \[[@B23-ijerph-17-02376]\]. From the perspective of risk theory, the risk related to food safety is defined as a function of the probability of an adverse health effect and the severity of that effect, consequential to a hazard(s) in food \[[@B24-ijerph-17-02376]\]. Risk perception is also perceived hazard characteristics, which is used to describe people's cognitive evaluations of potential risk \[[@B15-ijerph-17-02376]\]. High risk perception may increase the uncertainty of consumers, so that they will obtain more information to access the risk, thus stimulating the consumers to search for more information \[[@B25-ijerph-17-02376],[@B26-ijerph-17-02376]\]. If consumers judge that food safety risk will threaten their personal health or their families' health, they may have a stronger willingness to search for more food safety information in order to cope with risks. In addition, affective response is placed after risk perception in the RISP model, which implies an affective response is always caused by risk perception. Affective response that accompanies the risky situation is essentially negative \[[@B15-ijerph-17-02376]\]. Consumers with higher risk perception may react more negatively to food safety incidents. Therefore, the following hypotheses are advanced:
*Consumers with higher risk perception have a higher information-seeking intention about food safety risk*.
*Consumers with higher risk perception have a more negative affective response to food safety incidents*.
2.2. Affective Response {#sec2dot2-ijerph-17-02376}
-----------------------
Affective response refers to emotional reactions to risks \[[@B27-ijerph-17-02376]\]. In the RISP model, this kind of emotional reaction could affect people's judgment of the amount of risk information they feel needed, more specifically, it may influence people's sense of information sufficiency toward risk \[[@B11-ijerph-17-02376],[@B16-ijerph-17-02376]\]. Food safety is closely related to consumers' life and health. Food safety incidents that have occurred were the external manifestation of objective food safety risks \[[@B28-ijerph-17-02376]\]. In the face of food safety incidents, the affective response, such as worry, anxiety, etc., could prompt the consumers to generate a sense of the inadequacy of quality and safety information. It in turn generates the intention of food safety information so as to take effective action to deal with food safety risks. Accordingly, we forward the following hypothesis:
*Affective response to risk increases consumers' information search for food safety*.
2.3. Information Insufficiency {#sec2dot3-ijerph-17-02376}
------------------------------
Information insufficiency is the gap between the knowledge held and the information needed of consumers, which is a motivational factor of risk information-seeking behavior in the RISP model \[[@B11-ijerph-17-02376]\]. This kind of size gap will ultimately influence individuals' information seeking and processing styles \[[@B29-ijerph-17-02376]\]. The positive relationship between information insufficiency and information-seeking behavior in different risks contexts has been confirmed by many papers \[[@B17-ijerph-17-02376],[@B18-ijerph-17-02376]\]. We propose the following hypothesis:
*Information insufficiency has a positive effect on consumers' food safety risk information-seeking intention*.
2.4. Informational Subjective Norms {#sec2dot4-ijerph-17-02376}
-----------------------------------
One of the basic theories of the RISP model is the theory of planned behavior (TPB). Therefore, the concept of informational subjective norms mainly comes from the subjective norm of the TPB. Subjective norm is the person's perception of social pressures (in other words, the salient individuals or groups) put on him to perform or not perform the particular behavior. If a consumer believes he should perform a certain behavior based on what the society wants he is likely to perform it \[[@B30-ijerph-17-02376],[@B31-ijerph-17-02376]\]. Subsequently, informational subjective norms emphasize a normative influence from the social pressures they perceived in the process of risk information collection, this influence can mobilize an individual desire for information sufficiency \[[@B32-ijerph-17-02376]\]. There are two meanings of information subjective norm: one is the expectation that society wants consumers to know particular risk information; the other is the enthusiasm that consumers comply with society's expectations. Only when consumers have enough enthusiasm to defer to the expectation of knowing a certain risk that the society has placed on them, can the informational subjective norms really play a role in our seeking and processing behavior of risk information \[[@B32-ijerph-17-02376]\]. Thus, the hypotheses are as follows:
*Informational subjective norms positively influence consumers' information insufficiency about food safety risk*.
*Informational subjective norms positively influence consumers' food safety risk information-seeking intention*.
2.5. Relevant Channel Beliefs {#sec2dot5-ijerph-17-02376}
-----------------------------
Relevant channel beliefs refer to the trusting attitude of consumers to the direct government risk control department, universities, and news media that provide risk-related information. This belief mainly includes the perception of the reliability and validity of the risk information released by the above relevant channels, which could affect consumers' information seeking and processing strategies they may employ \[[@B33-ijerph-17-02376]\]. Relevant channel beliefs could influence consumers' choices for risk information seeking and processing strategies. We proposed the following hypothesis:
*Relevant channel beliefs of consumers have a positive effect on consumers' food safety risk information-seeking intention*.
2.6. Perceived Information Gathering Capacity {#sec2dot6-ijerph-17-02376}
---------------------------------------------
Perceived information gathering capacity measures the level of an individual's ability to collect relevant risk information. Generally speaking, when consumers are faced with food safety risks and need to inquire relevant information and knowledge, the fewer obstacles they face, the easier it is for them to query the information, and the higher intention to search relevant risk information they will have. Consequently, the following hypothesis is formulated:
*Perceived information gathering capacity of consumers has a positive effect on their food safety risk information-seeking intention*.
3. Materials and Methods {#sec3-ijerph-17-02376}
========================
3.1. Methodology {#sec3dot1-ijerph-17-02376}
----------------
A typical Structural Equation Modeling (SEM) not only can express the relationship between the latent variables and their indicators but also describe the relationship between endogenous and exogenous variables and the relationship among the endogenous variables \[[@B34-ijerph-17-02376]\]. Therefore, we chose the SEM as the main method and the equations used are as follows:$${\eta_{1} = \beta_{12}\eta_{2} + \gamma_{11}\xi_{1} + \gamma_{12}\xi_{2} + \gamma_{13}\xi_{3} + \gamma_{14}\xi_{4} + \zeta_{1}}{\eta_{2} = \beta_{23}\eta_{3} + \gamma_{22}\xi_{2} + \zeta_{2}}{\eta_{3} = \gamma_{31}\xi_{1} + \zeta_{3}}$$
There are seven latent variables that cannot be observed directly. Information-Seeking Intention (ISI), Information Insufficiency (II), and Affective Response (AR) are the latent variables as outcomes, i.e., exogenous variables, which are represented by $\eta_{1}$, $\eta_{2}$, and $\eta_{3}$. Risk Perception (RP), Informational Subjective Norms (ISN), Relevant Channel Beliefs (RCB), and Perceived Information Gathering Capacity (PIGC) are the latent variables as causes, i.e., endogenous variables, which are denoted by $\xi_{1}$, $\xi_{2}$, $\xi_{3}$, and $\xi_{4}$. In Equation (1), $\beta_{12}$ indicates the degree of influence of information insufficiency ($\eta_{2}$) on the information-seeking intention ($\eta_{1}$). $\gamma_{11}$, $\gamma_{12}$, $\gamma_{13}$, and $\gamma_{14}$ respectively represent the degree of influence of the exogenous variables $\xi_{1}$, $\xi_{2}$, $\xi_{3}$, and $\xi_{4}$ on the information-seeking intention ($\eta_{1}$). $\beta_{23}$ indicates the influence of affective response ($\eta_{3}$) to information insufficiency ($\eta_{2}$). $\gamma_{22}$ denotes the influence of informational subjective norms ($\xi_{2}$) on information insufficiency ($\eta_{2}$) and $\gamma_{31}$ indicates the influence of risk perception ($\xi_{1}$) on the affective response ($\eta_{3}$). Finally, $\zeta_{1}$, $\zeta_{2}$, and $\zeta_{3}$ are the residual terms of the regression equations.
3.2. Data Collection {#sec3dot2-ijerph-17-02376}
--------------------
Questionnaires were sent to consumers on WeChat between June and August 2017. Convenience sampling was used in this research and it is in accordance with the sampling methods adopted in many similar papers conducted using social media survey \[[@B35-ijerph-17-02376],[@B36-ijerph-17-02376],[@B37-ijerph-17-02376]\]. Sojump (<https://www.wjx.cn/>) is a powerful humanized questionnaire survey platform in China, which can provide a professional online questionnaire survey. After developing the questionnaire on the web page through the Sojump, we linked it into WeChat, a well-known social media platform in China, and collected 803 responses. After eliminating 29 invalid responses, there were 774 valid responses left. The reason for choosing WeChat is that WeChat is more reliable than other social media platforms. In China, WeChat Pay and Alipay are the two main electronic payment methods with very high popularity. True personal information of every WeChat user who uses WeChat Pay needs to be certificated with a series of formal procedures from the WeChat based on China's information regulation. Without real-name authentication, not only the payment function in WeChat cannot be used, but also some other functions will be restricted. Hence, we have faith in that the respondents of the study are who they say they are. The sample size meets the requirement of structural equation model analyses (generally speaking, 200 or more observations are required to conduct SEM based analyses \[[@B38-ijerph-17-02376]\]). China has a population of about 1.43 billion as of 2019, while WeChat has 1.11 billion monthly active users. Excluding the elderly people and children, the most majority of Chinese consumers use WeChat.
4. Results {#sec4-ijerph-17-02376}
==========
4.1. Descriptive Statistics of Variables and Demographic Characteristics of Consumers {#sec4dot1-ijerph-17-02376}
-------------------------------------------------------------------------------------
On the basis of the previous papers using the RISP model, we used the Likert five-point scale as the measurement method and designed an appropriate scale as far as possible in order to guarantee the content validity of the questionnaire. All the measurement items were measured as "strongly disagree" (=1), "disagree" (=2), "feel neutral" (=3), "agree" (=4), and "strongly agree" (=5) depending on the respondents' degree of agreement with the indicators, respectively. Scale design and descriptive statistics of all measurement items are displayed in [Table 1](#ijerph-17-02376-t001){ref-type="table"}.
[Table 2](#ijerph-17-02376-t002){ref-type="table"} below shows the demographic characteristics of the sample. As for the gender ratio, 54.9% of the respondents were female and 45.1% were male. In terms of age, the proportion of respondents under 35 years was 65.5% and the remaining 34.5% were equal or older than 35 years. Respondents with a bachelor's degree or higher occupied 78.9% of responses, while the respondents with below a bachelor's degree accounted for 21.1%. There were 16% who did not report their average monthly earnings, excluding these samples, the proportion of average monthly earnings below or equal to 5000 CNY (around the US \$750) is the same as that of average monthly earnings above 5000 CNY, which is 42%.
4.2. Reliability, Validity, and Confirmatory Factor Analysis {#sec4dot2-ijerph-17-02376}
------------------------------------------------------------
Before testing the research hypotheses, we conducted the reliability, validity, and confirmatory factor analysis by using statistical analysis software SPSS 22.0 and LISREL 8.7. The relevant results are reported in [Table 3](#ijerph-17-02376-t003){ref-type="table"} and [Table 4](#ijerph-17-02376-t004){ref-type="table"}. As far as the reliability of the scale is concerned, the measurement indicators mainly include the Corrected Item-Total Correlation (CITC), Cronbach's α value, and Composite Reliability (CR) \[[@B39-ijerph-17-02376],[@B40-ijerph-17-02376]\]. In [Table 3](#ijerph-17-02376-t003){ref-type="table"}, the CITC of each measurement ranges from 0.670 to 0.893, which are higher than the cutoff value of 0.5, the Cronbach's α of each latent variable ranges from 0.85 to 0.93, which greater than 0.7, and the CR of every latent variable ranges from 0.848 to 0.928, which is greater than 0.5. That means the scale of our research has high reliability. As for the validity, the Average Variance Extracted (AVE) in [Table 3](#ijerph-17-02376-t003){ref-type="table"} is greater than the recommended value of 0.5, showing that the latent variables have good convergence validity. The square root of the AVE of each latent variable is larger than the correlation coefficients between one latent variable and other latent variables, implying that the latent variables have better discriminant validity (see [Table 4](#ijerph-17-02376-t004){ref-type="table"}). Kaiser--Meyer--Olkin (KMO) measure of sampling adequacy of all the latent variables was 0.908 and Bartlett's test of sphericity was significant at 1%. According to the results of the confirmatory factor analysis, the standard factor loadings of each measurement item was greater than 0.7.
4.3. Structural Equation Model Analysis {#sec4dot3-ijerph-17-02376}
---------------------------------------
### 4.3.1. Offending Estimates Analysis {#sec4dot3dot1-ijerph-17-02376}
Offending estimates refer to the fact that the estimated coefficients output in the structural mode or the measurement mode exceeds the acceptable range. Generally, offending estimates occur in three ways: (1) there are negative or other meaningless error variances; (2) the standardization coefficients exceed or close to 1; and (3) the standard error is too big \[[@B40-ijerph-17-02376]\]. In the model, the error variances are between 0.10 and 0.43, and there are no negative or other meaningless error values. We also found that the standardization coefficients are much lower than 1, which indicates that there is no problem of offending estimates in the model. Thus, we can make further analysis.
### 4.3.2. Model Fitting Test {#sec4dot3dot2-ijerph-17-02376}
Testing the fit effect is one of the most cardinal preceding steps in SEM analysis. [Table 5](#ijerph-17-02376-t005){ref-type="table"} shows the overall fit evaluation results of the SEM model. There are three classifications of fit indices: absolute fit measures, incremental fit measures, and parsimonious fit measures \[[@B41-ijerph-17-02376]\]. As we can see from [Table 5](#ijerph-17-02376-t005){ref-type="table"}, except for a ratio of chi-square to the Degrees of Freedom (χ^2^/df) and Goodness of Fit Index (GFI), the actual values of the rest of fit indices were in line with a range of suggested values. As for χ^2^/df, its actual value was very close to the recommended values. This is because when using the indicators related to the chi-square distribution to test the fit effect of the model, it is sensitive to sample size \[[@B39-ijerph-17-02376]\]. Specifically, it means that it is easy to reject the model with a good fit effect when the sample size used is large \[[@B42-ijerph-17-02376]\]. The actual value of GFI was also very close to the suggested value. GFI ranges from 0 to 1, 0 represents the poor fit and 1 represents the perfect fit \[[@B43-ijerph-17-02376]\], and the GFI has a downward bias if the degree of freedom is large compared with the sample size (the degrees of freedom was 216 in our study). Due to the restrictiveness of the "sample-size" problem, the χ^2^/df and GFI actual values were acceptable. In general, the hypothetical model presented in this study fits well the actual observation data.
### 4.3.3. Research Hypothesis Testing {#sec4dot3dot3-ijerph-17-02376}
After the offending estimates and model fitting analysis, the research hypotheses between latent variables of the structural model were tested by using LISREL 8.7 software. The estimation results are presented in [Table 6](#ijerph-17-02376-t006){ref-type="table"} and [Figure 2](#ijerph-17-02376-f002){ref-type="fig"}. In general, except for the H8, the other seven hypotheses passed the test at different levels of significance.
The standardized coefficient of risk perception to consumers' food safety risk information-seeking intention was 0.18 and was significant at 1% (H~1~; t = 4.86). It can be seen that the higher the consumers' risk perception of food safety is, the stronger their attention to the search for food safety information. Risk perception also had a significant positive influence on the affective response, the path coefficient was 0.62 (H~2~; t = 17.32, *p* \< 0.01), which means that the consumers with higher risk perception had a more negative affective response to food safety incidents. Therefore, we failed to reject H~1~ and H~2~. Next, the affective response had a negative influence on the information insufficiency and information insufficiency had a positive influence on information-seeking intention. The standardized path coefficients of H~3~ and H~4~ were −0.07 and 0.08, which were significant at a 10% level (H~3~; t = −1.73) and 5% (H~4~; t = 2.30), respectively. This implies consumers with a stronger negative emotional response to food safety incidents are more likely to have a sense of information insufficiency, and thus have a strong intention to seek food safety information. Hence, we failed to reject H~3~ and H~4~.
Informational subjective norms can indirectly affect food safety information-seeking intention of the consumers through information insufficiency. The path coefficient of informational subjective norms to information insufficiency was 0.52 (H~5~; t = 12.37, *p* \< 0.01). Informational subjective norms also could directly influence information-seeking intention, and its standardized coefficient was 0.50 (H~6~; t = 9.26, *p* \< 0.01). Results indicate that the social pressure from important individuals or groups were one of the most important factors affecting consumer's information-seeking intention for food safety supporting hypotheses H~5~ and H~6~.
Relevant channel beliefs had a positive significant impact on consumers' information-seeking intention related to food safety. We found the path coefficient to be 0.16 (H~7~; t = 4.18, *p* \< 0.01). However, contrary to our belief, the influence of perceived information gathering capacity to information-seeking intention was not significant (H~8~; t = 1.60). Therefore, H~7~ was supported but H~8~ was not.
4.4. Multiple-Group Analysis {#sec4dot4-ijerph-17-02376}
----------------------------
Taking the demographic characteristics of consumers as the moderator variables, the multiple-group analysis results of social media users' food safety risk information-seeking intention are shown in [Table 7](#ijerph-17-02376-t007){ref-type="table"}. The analysis results of grouped samples and all the samples were generally similar.
In the path H~1~ where risk perception had a positive influence on the information-seeking attention, we could see that male (b = 0.23, *p* \< 0.01) was more significant than female (b = 0.13, *p* \< 0.05), the young group under 35 years old (b = 0.19, *p* \< 0.01) was more significant than the older group over or equal to 35 years old (b = 0.13, *p* \< 0.10). For the group with a bachelor's degree or higher (b = 0.20, *p* \< 0.01), the positive impact of risk perception on information-seeking intention was significant, while the group with below a bachelor's degree was not significant. As for the average monthly earnings, compared with the group whose average monthly earnings was less than 5000 CNY (b = 0.10, *p* \< 0.10), the positive influence of risk perception on information-seeking intention was more significant among the group whose average monthly earnings was more than 5000 CNY (b = 0.26, *p* \< 0.01). This is because the male, young consumers, the group with a higher education degree and earnings usually have a higher risk-awareness, so their intention to seeking food safety information in order to reduce risk is generally higher.
In the path H~2~ where risk perception had a significant positive influence on affective response to food safety incidents, the multiple-group results demonstrated that although the impact of risk perception on affective response was more significant in female (b = 0.63, *p* \< 0.01), young (b = 0.65, *p* \< 0.01), and high-earning group (b = 0.66, *p* \< 0.01), but the path coefficients gap was very small. In addition, the standardized coefficients of H~2~ between the group with a bachelor's degree or higher and the group with below a bachelor's degree were both 0.63 (*p* \< 0.01). This indicates that no matter which group, when people are concerned about food safety risks, they will have normal affective response such as anger, annoyance, and worries about food safety incidents.
In the path H~3~ where affective response had a negative influence on the information insufficiency, it was manifested that the path coefficients between the young group and old group, low-earning and high-earning group did not pass the significance test. The male (b = −0.13, *p* \< 0.05) and the group with below a bachelor's degree (b = −0.20, *p* \< 0.01) were significant while the female and the group with a bachelor's degree or higher were not. Men are more rational and aware of the crisis in most cases, so their negative affective response to food safety incidents could quickly translate into a sense of lack of food safety information, which drives them to search for more information. As for the low-educated group, the relevant food safety knowledge may be insufficient. Therefore, they are likely to feel that they need more food safety information under the influence of the negative emotion to food safety incidents.
In the path H~4~ where information insufficiency had a positive influence information-seeking intention, we can see that female (b = 0.12, *p* \< 0.05), young group (b = 0.09, *p* \< 0.05), highly educated (b = 0.13, *p* \< 0.01), and high-income group (b = 0.12, *p* \< 0.05) were significant. This shows that when consumers think that their knowledge of food quality and safety is insufficient, the female, young group, highly educated, and high-income group are more likely to transfer this sense of information deficiency into the seeking intention of food safety.
In the path H~5~ where informational subjective norms had a positive influence on information insufficiency, all groups were significant at 1% statistical level, indicating that the pressure from the salient individuals or groups will give the consumers in different groups a sense of information insufficiency. Compared to the female (b = 0.40, *p* \< 0.01), older group (b = 0.51, *p* \< 0.01), highly educated (b = 0.43, *p* \< 0.01), and high-income group (b = 0.26, *p* \< 0.01), this effect was greater for male (b = 0.66, *p* \< 0.01), young group (b = 0.54, *p* \< 0.01), low-educated (b = 0.83, *p* \< 0.01), and low-income consumers (b = 0.57, *p* \< 0.01).
In the path H~6~ where informational subjective norms had a direct positive influence on information-seeking intention, the results in all the groups were significant, however, the effects were slightly different in different groups. Additionally, it was higher in male (b = 0.51, *p* \< 0.01), older consumers (b = 0.63, *p* \< 0.01), low-educated (b = 0.69, *p* \< 0.01), and low-income consumers (b = 0.59, *p* \< 0.01).
In the path H~7~ where relevant channel beliefs had a direct positive influence on information-seeking intention, the male (b = 0.21, *p* \< 0.01) had a greater impact than female (b = 0.11, *p* \< 0.05), the young consumers (b = 0.18, *p* \< 0.01) had a greater impact than the older consumers (b = 0.13, *p* \< 0.05), and the high-income consumers (b = 0.22, *p* \< 0.01) had a greater impact than the low-income consumers (b = 0.11, *p* \< 0.10). In addition, the highly educated consumers (b = 0.16, *p* \< 0.01) had a significant effect but low-educated consumers did not.
Finally, the path H~8~ did not pass the significance test in the structural equation model, nevertheless, we found that in the multigroup analysis, the perceived information gathering capacity had a significant influence on seeking information intention among female (b = 0.10, *p* \< 0.05), young (b = 0.14, *p* \< 0.01), low-educated (b = 0.18, *p* \< 0.05), and low-income consumers (b = 0.23, *p* \< 0.01).
5. Discussion {#sec5-ijerph-17-02376}
=============
We used RISP as a reference to construct a conceptual framework of consumers' food safety risk information-seeking intention, and then assessed critical factors of consumers' food safety risk information-seeking intention based on the responses from 774 WeChat users. The main conclusions from this research are as follows: (1) Risk perception is an important variable in predicting people's information-seeking intention, consistent with previous studies \[[@B11-ijerph-17-02376],[@B15-ijerph-17-02376],[@B25-ijerph-17-02376]\]. On the one hand, risk perception has a direct significant positive effect on consumers' information-seeking intention of food safety risk and this effect was more pronounced among the young, male, highly educated, and high-income groups. On the other hand, the higher risk perception regarding food safety risks will make people feel more anxious and threatened, and then expand the gap between the information they need and the relevant knowledge they actually have, which will further stimulate them to seek more information. (2) The social pressure that people perceived was also a significant determinant of their information-seeking intention, supporting those of previous studies \[[@B18-ijerph-17-02376]\]. Informational subjective norms can not only directly affect consumers' information-seeking about food safety, but also indirectly affect consumers' intention through information insufficiency. Moreover, male, low-education and low-income groups were more likely to feel this social pressure from the important others. (3) The more the consumers trust the relevant channels that publish the food safety-related information, the stronger their intention to seek food safety risk information, especially for male, young, highly educated, and high-income groups. The result was consistent with previous studies \[[@B27-ijerph-17-02376]\]. Our findings provide valuable implications for the efforts of food safety risk communication in China even in other countries. Firstly, the food safety information released to consumers can combine a certain degree of possible threats and potential consequences \[[@B44-ijerph-17-02376]\]. In this way, consumers' risk perception and attention to food safety can be improved, so they could actively search for more information related to food safety and give full play to their role in food safety supervision. This is conducive to forcing food production enterprises to produce more safe and high-quality products, so as to improve the overall level of food safety in China. Secondly, reliable food safety information should be provided to consumers in accordance with the principles of science, accuracy, openness, and transparency, so as not to mislead consumers and public opinion. As for the government, a unified food safety information platform should be established to publish food safety risk warning information, investigation and handling information of major food safety accidents, and other important food safety information that need to be published. Besides, it is necessary to ensure that the information published is accurate and timely. As for news media, the report and publicity of food safety information should be true and fair. As for inspection organizations, the accuracy of the food safety inspection and testing reports they publish should be strictly guaranteed, and the contents in the report should be easy to understand and unambiguous. If necessary, the testing methods they applied should be clearly explained. Finally, strengthen food safety education for consumers and let them fully realize their role in food safety supervision.
6. Conclusions {#sec6-ijerph-17-02376}
==============
In summary, our study further verified the applicability of the RISP theoretical model in the field of food safety in China. Furthermore, the effects of consumers' characteristics such as gender, age, education, and income among different groups were analyzed. Compared with other literature, our research enriches the application of the RSIP model and also expands the depth of research by considering the impact of different demographic characteristics on consumers' intention as previous studies did not empirically analyze the moderating effects of demographic variables. However, there are still the following deficiencies in our paper. (1) We did not expand the knowledge body of the RISP model. Therefore, future research could open up the new derivative model and research vision based on the RISP model or its derivative model, making its application more scientific and targeted. (2) The RISP model involved not only the content of information-seeking behavior but also the information processing behavior after information-seeking. This study did not analyze the risk information processing behavior, which is one of the limitations of this paper. So, the researchers could adopt more scientific indicators to measure the information processing behavior of consumers and strengthen the relevant research. (3) Other issues that may limit the interpretation of the results are that an online survey may have under-coverage and self-selection bias problems \[[@B45-ijerph-17-02376]\]. Many papers that used an online survey have pointed out those issues \[[@B46-ijerph-17-02376],[@B47-ijerph-17-02376],[@B48-ijerph-17-02376]\]. Limited by some facts, for example, WeChat is not open source, and there is no specific sampling frame, we can only implement convenient sampling, which may lead to biased estimates. Hence, it is necessary to alleviate these concerns in future research. Experimental, quasi-experimental, and field-experimental design may be considered to solve this problem.
The authors would like to appreciate all the respondents in WeChat and the constructive comments of three anonymous reviewers.
Z.Y. and X.W. wrote up the paper; S.S. and Y.W. analyzed data; K.P.P. provided suggestions for this article. All authors have read and agreed to the published version of the manuscript.
This research was supported by the Key Project of the National Natural Science Foundation of China (No.71633002), the National Key Research and Development Program of China (2017YFC1601700) and the Graduate Student Overseas Study Program of South China Agricultural University (No.2018LHPY019). Paudel's time in this paper was supported by the USDA NIFA grants LAB\#94382 and LAB\#94358.
The authors declare no conflict of interest.
{#ijerph-17-02376-f001}
{#ijerph-17-02376-f002}
ijerph-17-02376-t001_Table 1
######
Scale design and descriptive statistics.
Latent Variables Measurement Items Mean S.D.
-------------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------------------- ------- -------
Risk Perception (RP) RP1. Food safety issues have a real impact on my family and me. 4.01 0.890
RP2. Food safety incidents seriously threaten my health. 3.93 0.940
RP3. Food safety incidents seriously threaten my family's health. 3.92 0.933
RP4. Food safety issues seriously threaten the whole society. 4.13 0.863
Affective Response (AR) AR1. The food safety incidents made me feel angry. 4.19 0.797
AR2. The food safety incidents made me feel annoyed. 4.16 0.814
AR3. The food safety incidents made me feel worried. 4.32 0.762
Information Insufficiency (II) II1. I have more information about food safety risks currently. 3.04 0.930
II2. I know what kinds of food safety risks I am facing. 3.18 0.931
II3. I have enough knowledge in the face of food safety risks. 2.85 0.966
Informational Subjective Norms (ISN) ISN1. People important to me think that I should stay at the top of information about food safety risk. 3.51 0.863
ISN2. My family expects me to seek more information about food safety risk. 3.72 0.834
ISN3. I think I should get more information about food safety risk. 3.97 0.805
Relevant Channel Beliefs (RCB) RCB1. I trust the food safety information issued by the government. 3.62 0.875
RCB2. I trust the food safety information issued by news media. 3.48 0.827
RCB3. I trust the food safety information issued by researchinstitutions. 3.76 0.846
Perceived Information Gathering Capacity (PIGC) PIGC1. If I want to find the information about food safety risk, I know where to find it. 3.38 0.915
PIGC2. If I want to find the information about food safety risk, I know how to find it. 3.37 0.922
PIGC3. I have already got the information I need related to food safety risk. 3.07 0.941
PIGC4. It's easy for me to obtain information about foodsafety risk. 3.02 0.988
Information-Seeking Intention (ISI) ISI1. I intend to seek information related to food safety risk. 3.75 0.786
ISI2. I plan to look for information related to food safety risk. 3.46 0.847
ISI3. I will try my best to find out information related to food safety risk in the near future. 3.70 0.776
Notes: S.D. means standard deviation.
ijerph-17-02376-t002_Table 2
######
Demographic characteristics of consumers (*N* = 774).
Characteristics Variable Classification Number Ratio (%)
----------------------------------- --------------------------- -------- -----------
Gender Female 425 54.9
Male 349 45.1
Age \<35 507 65.5
≥35 267 34.5
Education background Below a bachelor's degree 163 21.1
Bachelor's degree or higher 611 78.9
Average monthly earnings (in CNY) ≤5000 325 42.0
\>5000 325 42.0
Missing 124 16.0
Notes: Exchange rate US\$1 = 7.03 CNY (as of 11/30/2019).
ijerph-17-02376-t003_Table 3
######
Reliability, validity, and confirmatory factor analysis results.
Latent Variables Measurement Items CITC Cronbach's α If Item Deleted CR AVE KMO Bartlett's Test of Sphericity Standard Factor Loadings
---------------------------- ------------------- ------- ------------------------------ ------- ------- ------- ------------------------------- --------------------------
RP (Cronbach's α = 0.93) RP1 0.779 0.92 0.928 0.764 0.811 2802.031 0.79
RP2 0.881 0.89 0.95
RP3 0.893 0.88 0.96
RP4 0.763 0.93 0.78
AR (Cronbach's α = 0.92) AR1 0.845 0.88 0.920 0.793 0.750 1730.367 0.89
AR2 0.865 0.86 0.93
AR3 0.801 0.91 0.85
II (Cronbach's α = 0.89) II1 0.802 0.81 0.884 0.718 0.741 1276.328 0.89
II2 0.762 0.85 0.83
II3 0.758 0.85 0.82
ISN (Cronbach's α = 0.85) ISN1 0.690 0.82 0.849 0.654 0.693 1053.206 0.78
ISN2 0.795 0.71 0.88
ISN3 0.670 0.83 0.76
RCB (Cronbach's α = 0.85) RCB1 0.759 0.75 0.852 0.658 0.720 1015.103 0.86
RCB2 0.719 0.79 0.82
RCB3 0.680 0.83 0.75
PIGC (Cronbach's α = 0.90) PIGC1 0.803 0.87 0.904 0.703 0.802 2088.629 0.88
PIGC2 0.799 0.87 0.88
PIGC3 0.784 0.88 0.81
PIGC4 0.754 0.89 0.78
ISI (Cronbach's α = 0.85) ISI1 0.699 0.80 0.848 0.651 0.730 985.136 0.80
ISI2 0.730 0.78 0.81
ISI3 0.722 0.78 0.81
Notes: (1) Latent variables are RP = Risk Perception, AR = Affective Response, II = Information Insufficiency, ISN = Informational Subjective Norms, RCB = Relevant Channel Beliefs, PIGC = Perceived Information Gathering Capacity, ISI = Information-Seeking Intention. KMO = Kaiser--Meyer--Olkin measure of sampling adequacy. AVE = Average Variance Extracted (AVE). (2) Corrected Item-Total Correlation (CITC) values of each latent variable are higher than 0.5. The Cronbach's alpha values of each latent variable are higher than 0.7, and Cronbach's α if item deleted values of each measurement item are lower than the Cronbach's alpha values of each latent variable, which means all measurement items should be kept. Composite Reliability (CR) values of the latent variable are higher than 0.5. It means the scale has high reliability. AVE values show that the latent variables have good convergence validity.
ijerph-17-02376-t004_Table 4
######
The square root of AVE and correlation coefficients of latent variables.
Latent Variables RP AR II ISN RCB PIGC ISI
------------------ ------ ------ ------ ------ ------ ------ ------
RP 0.87
AR 0.62 0.89
II 0.26 0.12 0.85
ISN 0.59 0.36 0.50 0.81
RCB 0.30 0.19 0.25 0.50 0.81
PIGC 0.19 0.12 0.21 0.42 0.45 0.84
ISI 0.59 0.34 0.43 0.75 0.51 0.39 0.81
Notes: (1) Latent variables are RP = Risk Perception, AR = Affective Response, II = Information Insufficiency, ISN = Informational Subjective Norms, RCB = Relevant Channel Beliefs, PIGC = Perceived Information Gathering Capacity, ISI = Information-Seeking Intention. (2) The diagonal value at the top of each column (the square root of the AVE of each latent variable) is higher than other entries in the column (the correlation coefficients between one latent variable and other latent variables). It means that the latent variables have better discriminant validity.
ijerph-17-02376-t005_Table 5
######
Fitting indices of the Structural Equation Model.
Classification Fit Indices Suggested Value Actual Value Fit Effect
--------------------------- ------------- ----------------- -------------- ------------
Absolute Fit Measures χ^2^/df \<5.00 6.37 Approx.
RMSEA \<0.09 0.09 Accepted
AGFI \>0.80 0.80 Accepted
GFI \>0.90 0.85 Approx.
Incremental Fit Measures CFI \>0.90 0.96 Accepted
NFI \>0.95 0.96 Accepted
NNFI \>0.95 0.96 Accepted
IFI \>0.90 0.96 Accepted
Parsimonious Fit Measures PNFI \>0.50 0.815 Accepted
PGFI \>0.50 0.662 Accepted
Notes: RMSEA: Root Mean Square Error of Approximation. AGFI: Adjusted Goodness of Fit Index. GFI: Goodness of Fit Index. CFI: Comparative Fit Index. NFI: Normed Fit Index. NNFI: Non-normed Fit Index. IFI: Incremental Fit Index. PNFI: Parsimony Normed Fit Index. PGFI: Parsimony Goodness of Fit Index.
ijerph-17-02376-t006_Table 6
######
Results from hypotheses tests between different latent variables in the Structural Equation Model.
Hypothesis Standardized Coefficients Direction T-Value Test Result
-------------- --------------------------- ----------- --------- -------------
H1: RP→ISI 0.18 \+ 4.86 Support
H2: RP→AR 0.62 \+ 17.32 Support
H3: AR→II 0.07 − −1.73 Support
H4: II→ISI 0.08 \+ 2.30 Support
H5: ISN→II 0.52 \+ 12.37 Support
H6: ISN→ISI 0.50 \+ 9.26 Support
H7: RCB→ISI 0.16 \+ 4.18 Support
H8: PIGC→ISI 0.06 \+ 1.60 Not support
Notes: Latent variables are RP = Risk Perception, AR = Affective Response, II = Information Insufficiency, ISN = Informational Subjective Norms, RCB = Relevant Channel Beliefs, PIGC = Perceived Information Gathering Capacity, ISI = Information-Seeking Intention.
ijerph-17-02376-t007_Table 7
######
Multiple-group analysis results.
--------------------- ---------------------- ----------------------------------- --------------- ---------- --------------- --------- --------------- -------
**Hypothesis** **Gender** **Age**
**Female** ***p*-Value** **Male** ***p*-Value** **\<35** ***p*-Value** **≥35** ***p*-Value**
H1 0.13 \*\* ≤0.05 0.23 \*\*\* ≤0.01 0.19 \*\*\* ≤0.01 0.13 \* ≤0.1
H2 0.63 \*\*\* ≤0.01 0.61 \*\*\* ≤0.01 0.65 \*\*\* ≤0.01 0.56 \*\*\* ≤0.01
H3 −0.02 \>0.1 −0.13 \*\* ≤0.05 −0.07 \>0.1 −0.09 \>0.1
H4 0.12 \*\* ≤0.05 0.03 \>0.1 0.09 \*\* ≤0.05 0.01 \>0.1
H5 0.40 \*\*\* ≤0.01 0.66 \*\*\* ≤0.01 0.54 \*\*\* ≤0.01 0.51 \*\*\* ≤0.01
H6 0.49 \*\*\* ≤0.01 0.51 \*\*\* ≤0.01 0.44 \*\*\* ≤0.01 0.63 \*\*\* ≤0.01
H7 0.11 \*\* ≤0.05 0.21 \*\*\* ≤0.01 0.18 \*\*\* ≤0.01 0.13 \*\* ≤0.05
H8 0.10 \*\* ≤0.05 0.01 \>0.1 0.14 \*\*\* ≤0.01 −0.03 \>0.1
Hypothesis Education Background Average Monthly Earnings (in CNY)
Bachelor DegreeBlow *p*-value Bachelor Degree or Higher *p*-value ≤5000 *p*-value \>5000 *p*-value
H1 0.11 \>0.1 0.20 \*\*\* ≤0.01 0.10 \* ≤0.1 0.26 \*\*\* ≤0.01
H2 0.63 \*\*\* ≤0.01 0.63 \*\*\* ≤0.01 0.61 \*\*\* ≤0.01 0.66 \*\*\* ≤0.01
H3 −0.20 \*\*\* ≤0.01 −0.02 \>0.1 −0.09 \>0.1 −0.07 \>0.1
H4 −0.14 \>0.1 0.13 \*\*\* ≤0.01 0.01 \>0.1 0.12 \*\* ≤0.05
H5 0.83 \*\*\* ≤0.01 0.43 \*\*\* ≤0.01 0.57 \*\*\* ≤0.01 0.26 \*\*\* ≤0.01
H6 0.69 \*\*\* ≤0.01 0.48 \*\*\* ≤0.01 0.59 \*\*\* ≤0.01 0.38 \*\*\* ≤0.01
H7 0.12 \>0.1 0.16 \*\*\* ≤0.01 0.11 \* ≤0.1 0.22 \*\*\* ≤0.01
H8 0.18 \*\* ≤0.05 0.02 \>0.1 0.23 \*\*\* ≤0.01 −0.03 \>0.1
--------------------- ---------------------- ----------------------------------- --------------- ---------- --------------- --------- --------------- -------
Notes: (1) \*\*\*, \*\*, and \* indicated that *p*-values are significant at 1%, 5%, and 10% levels respectively. (2) One hundred and twenty four respondents did not report average monthly earnings, so missing samples were excluded from multiple-group based on consumers' characteristics of average monthly earnings.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION {#s1}
============
Primary central nervous system (CNS) lymphoma is a rare type of malignancy confined to the brain, spinal cord, meninges, and orbits. Most of primary CNS lymphoma is histologically classified as diffuse large B-cell lymphoma, which has a fast growing and aggressive feature. Likewise, primary CNS lymphoma is characterized by its microscopically diffuse and multifocal feature. In the United States, its incidence has steadily increased since the late 20th century \[[@R1]\]. Yet, it accounts for less than 3% of all primary intracranial malignancies \[[@R2]\]. Because of the rarity of this disease, a standard treatment confirmed by reliable randomized clinical trials has not yet been established. Historically, radiotherapy was the treatment of choice for primary CNS lymphoma \[[@R3]--[@R5]\]. However, with the improvement of chemotherapy, the universal treatment used most recently for primary CNS lymphoma is chemotherapy based on a high-dose methotrexate (MTX) with or without radiotherapy. Although studies have demonstrated prolonged progression-free survival in regimens including whole brain radiotherapy \[[@R6]--[@R9]\], the major reason for deferring radiotherapy in the treatment of primary CNS lymphoma is the late neurotoxic sequelae induced by multi-modality therapy. Consequently, the actual use of radiotherapy is diminishing and controversy on its role is ongoing. However, there are patients who are reluctant or contraindicated for chemotherapy due to comorbidities, poor performance status, or refusal. The functional outcomes with disease progression for those patients are devastating if they do not receive any treatment. Therefore, a treatment with a palliative aim at minimum must be applied for a better quality of life, and radiotherapy can be a useful treatment option to address this goal. We conducted this study to evaluate the treatment outcomes of primary CNS lymphoma that was treated with radiotherapy in patients who refused, could not tolerate, or were refractory to chemotherapy.
RESULTS {#s2}
=======
Patients {#s2_1}
--------
There were 45 primary CNS lymphoma patients who were treated with radiotherapy. Among the total patients, 13 received more than two cycles of chemotherapy and were excluded from the study. Therefore, 32 patients were evaluated. Patient characteristics are provided in Table [1](#T1){ref-type="table"}. The median age at diagnosis was 66 years old. Multiple brain lesions were observed in twenty patients (62.5%). The proportion of patients with an ECOG performance status of 2 or 3 was higher (59.4%) than 0 or 1 and poor performance status. The major reason why the patients were non-candidates for chemotherapy was their poor performance status. Nineteen patients (59.4%) were in ECOG performance status of 2 and 3. Five patients (6.6%) were inadequate for chemotherapy due to comorbidities: one patient had poor renal function, one patient had chronic hepatitis B, and three patients had cardiac problem.
###### Patient characteristics
Characteristics Number \%
--------------------------------------- -------- ------
Age
\< 60 years old 12 37.5
≥ 60 years old 20 62.5
Gender
Male 16 50.0
Female 16 50.0
ECOG Performance status
0--1 13 40.6
2--3 19 59.4
LDH level
Normal 17 53.1
Elevated 15 46.9
Number of brain lesion(s)
Single 12 37.5
Multiple 20 62.5
Whole-brain radiotherapy dose
\< 30 Gy 14 43.8
≥ 30 Gy 18 56.3
Non-candidate for chemotherapy due to
Poor performance status 19 59.4
Comorbidity 5 15.6
Progression after chemotherapy^\*^ 4 12.5
Intolerable to chemotherapy^\*^ 4 12.5
ECOG: Eastern cooperative oncology group; LDH: lactate dehydrogenase.
^\*^these patients received only one or two cycles of chemotherapy due to progression or toxicity.
Twenty patients (62.5%) had multiple brain lesions. Except for one patient, the patients presented with various neurologic symptoms at diagnosis (Table [2](#T2){ref-type="table"}). The major clinical presentations were hemiparesis, headache, and memory impairment.
###### Neurologic symptoms at presentation
Symptoms Number \%
----------------------- -------- ------
Hemiparesis 14 43.8
Headache 8 25.0
Memory impairment 8 25.0
Dizziness 5 15.6
Disorientation 4 12.5
Gait disturbance 3 9.4
Dysarthria 3 9.4
Cognitive dysfunction 2 6.3
Visual disturbance 2 6.3
Seizure 1 3.1
Treatment {#s2_2}
---------
Radiotherapy was delivered at 1.8--2 Gy per fraction, once daily. All of the patients initiated with whole brain radiotherapy (WBRT). Six patients (18.8%) were treated with WBRT only and did not received additional boost RT. Remaining 26 patients (81.2%) received boost RT to the primary mass after WBRT. When planning on the reduced field for boost RT on primary mass, gross tumor volume (GTV) was delineated based on the T1-weighted gadolinium-enhanced image on initial brain magnetic resonance imaging (MRI). Clinical target volume (CTV) was created by expanding GTV with 1--1.5 cm margin in all directions. Another 0.5 cm margin in all directions from CTV was given to create the planning target volume. In patients with multifocal disease, radiotherapy field reduction was performed as follows: when the lesions were distinct and a large distance between the lesions existed, separate fields were used; when the lesions were near enough to the coalesced in the margin, one field was used; in numerous and diffuse lesions, only WBRT was delivered.
The median treatment time was 35.5 days (range 12--48 days). There were no treatment interruptions. Patients received WBRT at a median dose of 30 Gy (range, 14.4--50 Gy) and the total median radiotherapy dose was 50 Gy (range, 30--54 Gy).
Treatment response {#s2_3}
------------------
Symptoms caused by the disease were relieved in 27 patients (84.4%). In these patients, symptom relief was observed at median 23 days after radiotherapy started. The response evaluation in all patients was performed with brain MRIs. The first assessment of treatment response was performed at median 4 months (range, 1--9 months). None of the patients presented disease progression at the first assessment after radiotherapy completion (Table [3](#T3){ref-type="table"}). Only one patient had a stable disease. Sixteen patients (50%) exhibited a complete response and fifteen patients (46.9%) had a partial response. Most of the patients were tolerable to radiotherapy and reported toxicity is described in Table [4](#T4){ref-type="table"}. Of the patients who complained of toxicity, nine patients (28.1%) experienced white matter change in post-treatment brain MRI, but did not require additional steroid therapy. One patient suffered grade 3 general weakness and another patient suffered grade 3 neutropenia. No grade 4 and above toxicity was observed.
###### Radiotherapy response
Treatment response Number \%
--------------------- -------- ------
Complete response 16 50.0
Partial response 15 46.9
Stable disease 1 3.1
Progressive disease 0 0.0
###### Toxicity after radiotherapy according to CTCAE v4.03
Toxicity Grade
-------------------------- ---------------- ----------- ---------- ------
Non-hematologic toxicity 1 2 3 4--5
Fatigue 2 2 1
Nausea 3 1
Headache 1
Dizziness 1
Anorexia 1 1
Total 8 (25%) 4 (12.5%) 1 (3.1%) 0
Hematologic toxicity 1 2 3 4--5
Neutropenia 2 1 1
Anemia 9
Thrombocytopenia 1 3
Total 12 (37.5%) 4 (12.5%) 1 (3.1%) 0
CTCAE: Common Terminology Criteria for Adverse Events.
Survival {#s2_4}
--------
The median follow-up time for all patients was 21 months. The Kaplan-Meier survival curves for PFS and OS are displayed in Figures [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}, respectively. The median PFS and OS were 15.8 (range 8.7--22.9) and 16.3 (range, 2.1--30.5) months, respectively. Since radiotherapy is a local treatment modality, we also assessed intracranial progression. Twenty patients (62.5%) showed recurrent disease at follow-up and among them, fifteen patients (46.9%) had intracranial recurrence. The median intracranial PFS was 19.3 (range, 10.2--28.4) months (Figure [3](#F3){ref-type="fig"}). A correlation between radiotherapy dose to the whole brain and survival was analyzed and the results are depicted in Figure [4](#F4){ref-type="fig"}. There was no statistical difference between the above and below 30 Gy.
{#F1}
{#F2}
{#F3}
{#F4}
DISCUSSION {#s3}
==========
The incidence of primary CNS lymphoma continues to rise \[[@R10]--[@R12]\]. The algorithms of primary CNS lymphoma treatments have changed. In the 1970s, primary CNS lymphoma was treated with radiotherapy as a single-modality treatment \[[@R13]\]. Presently, the high-dose MTX based chemotherapy is generally applied in the treatment of primary CNS lymphoma and radiotherapy is often deferred or avoided due to late neurotoxicity. The treatment outcome of patients treated with chemotherapy alone is reported as a median PFS of 21.5 months and median OS of 39.4 months. For patients who achieved a CR after chemotherapy and received radiotherapy, the median PFS and OS was 36.3 months and 38.8 months, respectively \[[@R8], [@R14]--[@R18]\]. Although treating primary CNS lymphoma with chemotherapy alone has a risk of disease progression, the overall survival rate was acceptable and this is the main reason for treating primary CNS lymphoma with chemotherapy.
Unfortunately, there are patients who are unsuitable for chemotherapy due to old age, poor performance status, comorbidities, and/or patient's unwillingness to engage in chemotherapy. Actually, in a population-base study, the proportion of the patients reluctant to chemotherapy was as high as 39% \[[@R19]\]. For the patients who are inadequate for chemotherapy, secondary treatment options can be radiotherapy, surgery, and conservative treatment. However, most of the patients in this category also cannot tolerate surgery because the patients are in poor general condition in most of these cases. With conservative treatment, including steroid therapy, the median survival is around 3 months due to the disease's fast growing nature \[[@R20]\]. When the disease progresses, the patient's expected functional outcome is devastating since the disease involves the CNS; therefore, a secondary or palliative treatment should be applied, and the remaining option is radiotherapy. The treatment duration after WBRT is short due to tumor regrowth and the reported survival with WBRT as a single-modality is 10--12 months \[[@R5], [@R8], [@R18], [@R19], [@R21], [@R22]\]. However, the nature of primary CNS lymphoma is not only chemosensitive but also radiosensitive; the tumor response rate after radiotherapy is excellent. Accordingly, neurologic symptom relief occurred in 84.4% and more than 90% of the patients achieved CR or PR in our study. Additionally, the treatment outcome after radiotherapy was fairly acceptable with a median PFS of 15.8 months and median intracranial PFS of 19.3 months.
When a patient is noncandidate for chemotherapy, radiotherapy has more benefits compared to supportive therapy by not only preserving activities of daily living, but also prolonging life. When treated with radiotherapy, a balance between the benefits and risks should be considered and the high incidence of neurotoxicity is the major reason for radiotherapy in primary CNS lymphoma is often deferred or avoided. In previously reported studies \[[@R8], [@R23]\], WBRT was delivered to a total dose of 45 Gy, which is a relatively high dose and very likely to cause critical neurotoxicity. In 2013, Morris et al. reported on the advantages of reduced-dose WBRTs \[[@R24]\]. WBRT with 23.4 Gy and 45 Gy were compared and the patients who received the reduced-dose WBRT had better PFS and OS with a decreased neurologic toxicity. Although a thorough and routine assessment on neurotoxicity was not available in our study, the reported toxicity was acceptable and these acceptable results may have come from our radiotherapy scheme with relatively low dose to the whole brain with a median 30 Gy. In our study, the correlation analysis between the radiation dose to the whole brain and survival demonstrated no statistically significant difference between the administration of above and below 30 Gy. Since WBRT of 30 Gy and below does not produce fatal neurotoxicity, it can be delivered safely without harming the treatment outcome. To validate the radiotherapy dose and scheme, further investigations with randomized trials should be executed.
Due to this study's retrospective nature, selective bias and observer bias were inevitable and these yield a limitation to our study. A relatively small number of patients is another limitation and our results should be interpreted with caution. Assessment on patients' neurocognitive function was not thoroughly performed and the data on toxicity was based on patient-reported symptoms and physician-detected signs such as radiographic changes and the use of medication. Yet, non-hematologic and hematologic toxicity were quite tolerable in most of the patients. Also, since there was no grading system for the severity of neurologic symptoms, the assessment of symptom relief could not be conducted systematically. However, when the neurologic symptoms were relieved after radiotherapy, it was specified in the medical records. This result for which there was an alleviation of neurologic symptoms and acceptable tumor response, radiotherapy in primary CNS lymphoma was worth the treatment when chemotherapy is not indicated.
Since the brain controls most of the activities of the body, the expected outcomes of untreated primary CNS lymphoma are devastating. When chemotherapy is unfeasible, treatment with a palliative aim at minimum, should be administered and radiotherapy is a practical option for satisfying symptom relief and modest disease control. Since the adequate dose for radiotherapy has not yet been established, a study with a large number of patients and long-term follow-up is necessary to determine the optimal radiotherapy scheme.
MATERIALS AND METHODS {#s4}
=====================
Patients {#s4_1}
--------
This study included data of patients with histologically confirmed primary CNS lymphoma and treated with radiotherapy from January 2005 to December 2015. Inclusion criteria were patients who received less than three cycles of chemotherapy before radiotherapy, age of 20 years or older, no evidence of systemic non-Hodgkin lymphoma, and negative human immunodeficiency virus (HIV-1) infection. Patients with a history of other malignancies were excluded. Age, performance status, history taking, physical examination, complete blood counts, and blood chemistry were evaluated. All patients had neuroimaging with brain computed tomography (CT) and MRI with T1-weighted sequences before and after a contrast injection and T2-weighted images. The work-up included a HIV blood test, lumbar puncture, and ophthalmological assessment. This study was approved by the institutional review board.
Treatment {#s4_2}
---------
Fifteen patients (46.9%) could not receive chemotherapy due to poor general condition. Five patients (15.6%) were unsuitable for toxic chemotherapy due to underlying comorbidities. Twelve patients (37.5%) were treated with chemotherapy at first, but the disease progressed after one or two cycles of chemotherapy.
Radiotherapy simulation was done in a supine position and a u-frame mask was used for immobilization. An enhanced brain CT for simulation was obtained in 3mm slices. For more delicate delineation of the targets, brain MRIs were fused to the simulation CT images. Radiation was delivered at a median total dose of 50 Gy (range, 22--54 Gy) and all patients received whole brain radiotherapy upon initiation. The median whole brain radiotherapy dose was 30 Gy (range, 14.2--50 Gy).
Assessment {#s4_3}
----------
Patients were interviewed weekly during the treatment, monthly for three months after the treatment, and every three months thereafter. Neurologic symptoms caused by the disease were queried before radiotherapy and the changes in the symptoms during and after the treatment were closely observed. A brain MRI was used for response evaluation after the completion of radiotherapy. A complete response (CR) was defined as no remaining mass. A partial response (PR) was defined as a decrease of at least 30% or more in size. Stable disease (SD) was defined as less than 30% decrease in size. Progressive disease (PD) was defined as an increased size of the initial mass or development of any new lesions. Overall survival (OS) was defined as the period from the date of diagnosis to death from any cause. Progression-free survival (PFS) was defined as the period from the date of diagnosis to progression. Assessment on toxicity was done based on medical chart records on patient-reported symptoms, physician-detected signs, and radiological changes in follow up images.
Statistics {#s4_4}
----------
The goal of this study was to assess the role of radiotherapy in symptom relief with functional recovery and tumor response to radiation in primary CNS lymphoma when standard chemotherapy cannot be administered. The OS and PFS curves were assessed with Kaplan-Meier analyses. All statistical tests with *p*-values less than 0.05 were considered statistically significant. All statistical analyses were performed using the R software, version 3.1.2 (R Foundation for Statistical Computing, Vienna, Austria; [www.r-project.org](http://www.r-project.org)).
**Author contributions**
Conception and design: BOC; Collection and assembly of data: YKK, KHC; Data analysis and interpretation: YKK, BOC; Manuscript writing: YKK, BOC; Final approval of manuscript: BOC, YKK, KHC, SYS, JHL, YHL.
**CONFLICTS OF INTEREST**
The authors declare no conflicts of interests.
CNS
: central nervous system
WBRT
: whole brain radiotherapy
PFS
: progression free survival
OS
: overall survival
MTX
: methotrexate
CT
: computed tomography
MRI
: magnetic resonance imaging
CR
: complete response
PR
: partial response
SD
: stable disease
PD
: progressive disease
ECOG
: Eastern cooperative oncology group
| {
"pile_set_name": "PubMed Central"
} |
Background
==========
Increasing number of women with HIV are choosing to become pregnant as there is reduction in vertical transmission. However, management of HIV in pregnancy still poses a variety of challenges and adverse pregnancy outcomes are still common. We aimed to explore the factors associated with adverse outcomes of pregnancy in our HIV cohort.
Methods
=======
It is a retrospective case notes review of all the women attended to our unit and had the HIV care from 2008-2011. A total of 87 women were followed up. Three women had two pregnancies during the study period. Data collected from Genitourinary Medicine and maternity records were analysed by using SPSS program.
Results
=======
Mean age was 34 yrs ranging from 20-43 yrs. Majority (91%) were of African origin; 67% had HIV subtype C; 26% resistant to one or more class of HIV drugs; 55% had a nadir CD4 fewer than 350; 44% diagnosed at an antenatal setting and 62% were planned pregnancies. Prior to the current pregnancy, these women had 121 children: 5% of the children have HIV and 33% not tested for HIV.
None of the children born during the study period were infected with HIV; there were 3 sets of twins; one still birth and one child died soon after birth.
Around 46% were on anti retroviral therapy (ART) during conception, 6% had miscarriage and 16% had emergency caesarean section.
38% of the women experienced an obstetric complication, premature labour 9%; premature rupture of membranes and gestational diabetes both accounted to 4% whilst 3% had post partum haemorrhage.
On ART during conception and late HIV diagnosis, nadir CD4, less than 350 cells were significantly associated (P\< 0.05) with having a foetal complication such as prematurity 8%, low birth weight 7% or having a foetal abnormality 2.3%. More analysis is awaited as to drug exposure and adverse outcomes.
Conclusion
==========
Late diagnosis of HIV and ART during conception is significantly associated with adverse outcomes of pregnancy. Widespread HIV testing is essential and has to be extended to non traditional settings.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .100}
{#f1 .100}
| {
"pile_set_name": "PubMed Central"
} |
Background {#Sec1}
==========
Microbial communities are known to establish sophisticated metabolic interactions in order to achieve complex and energy-expensive tasks \[[@CR1]--[@CR5]\]. These syntrophic relationships are frequently studied in bacterial pathogens and symbiotic bacteria, where the interaction with the host often drives progressive adaptation, mutation, and subsequently, gene loss. These phenomena may render the bacteria "unculturable" or difficult to grow under standard laboratory conditions \[[@CR6]--[@CR11]\]. On the contrary, the phenomena underlying metabolic cooperation and competition within environmental communities are often more complex, and their implications for microbial ecology are still poorly understood \[[@CR5], [@CR11]\]. These communities recurrently exchange metabolites or co-factors and are often associated with xenobiotic-degraders thriving in polluted environments \[[@CR5], [@CR11]--[@CR15]\]. This syntrophy has been previously observed in terephthalate-degrading communities \[[@CR1], [@CR2]\], in anammox-associated communities \[[@CR3], [@CR4], [@CR16]\], in the dichloromethane-degrader '*Candidatus* Dichloromethanomonas elyunquensis' \[[@CR17]\], and in members of the candidate phylum '*Candidatus* Latescibacteria', that thrives in hydrocarbon-impacted environments \[[@CR18], [@CR19]\]. However, to date, no representatives of these groups could be isolated as pure cultures, and their metabolic needs are difficult to assess. Terephthalate-degraders, for instance, thrive in an intricate network formed between H~2~-producing syntrophs and methanogenic archaea, with numerous other secondary interactions essential for the stability of the consortium \[[@CR1], [@CR2]\]. Anammox bacteria were shown to form stable biofilm communities with ammonia-oxidizing bacteria (AOB), that appear to be essential to protect the sensitive anammox species from atmospheric O~2~ \[[@CR3], [@CR4], [@CR20], [@CR21]\]. The evolution of these communities is driven by selective pressure and stress and may result in complex syntrophic relationships that may lead to niche-specialization and dependency on other members of the community. In order to characterize the members of these communities, cell-sorting and metagenomics approaches are being used to circumvent the need for cultivation \[[@CR15]\]. Furthermore, these studies are frequently complemented with comparative genomics which has emerged as a valuable tool to determine the evolution and functional prediction between even distantly related bacteria \[[@CR14], [@CR22], [@CR23]\]. The cultivation of several members of the ubiquitous SAR11 aquatic bacteria, with no closely related culturable relatives, has been made possible by in silico metabolic studies and next-generation sequencing approaches \[[@CR24]\]. Furthermore, the evolution of this abundant group of *Alphaproteobacteria* and their ecological importance has been further elucidated using comparative genomic approaches \[[@CR25]\]. In a previous study, we have described a microbial consortium between *Achromobacter denitrificans* strain PR1 and strain GP that depends on strain PR1's presence for growth \[[@CR26]\]. Strain GP showed the highest pairwise similarity of its 16S rRNA gene sequence to members of the genus *Leucobacter*. Independently of the tested culture media, cofactors and culture conditions no pure cultures were obtained for strain GP \[[@CR26]\]. To characterize strain GP, we have sequenced the two-member consortium and reconstructed its draft genome. Also, we have performed comparative genomic studies in order to understand its phylogenetic relationship with other members of the *Leucobacter* genus and propose the hypothesis that may allow us to understand why this strain has eluded isolation in previous studies.
Results and discussion {#Sec2}
======================
Morphological and physiological characterization of the consortium {#Sec3}
------------------------------------------------------------------
The microbial consortium between strain *A. denitrificans* and the low-abundant strain GP was visualized by Cryo-TEM during mid-stationary phase (Fig. [1a and b](#Fig1){ref-type="fig"}), as well as by FISH (see Additional file [1](#MOESM1){ref-type="media"} Figure S1). As expected, strain PR1 showed the typical morphology of Gram-negative rods with an average cell size of 801.3 ± 40.2 nm (width), 1332 ± 98.7 nm (length) and 38.2 ± 6.5 nm (periplasmic space) (Fig. [1a](#Fig1){ref-type="fig"}). Moreover, peritrichous flagella were observed by negative stain electron microscopy (FG, see Additional file [1](#MOESM1){ref-type="media"} Figure S2). Although flagella have not been previously reported for the type strain of *A. denitrificans*, their presence has been repeatedly observed in other strains from this species \[[@CR27]\] and other species of the *Achromobacter* genus \[[@CR28], [@CR29]\]. Conversely, strain GP displayed the typical morphology of Gram-positive rods. Its cells showed an average size of 506.6 ± 30.1 nm (width) and 1341.0 ± 29.7 nm (length) (Fig. [1b](#Fig1){ref-type="fig"}), and the rigid cell wall of this organism had an average thickness of 20.6 ± 2.2 nm. No flagella were observed for this bacterium, suggesting that it is non-motile, like previously reported for other members of the *Leucobacter* genus \[[@CR30], [@CR31]\]. The two members of the consortium revealed significant differences regarding their respective tolerances toward temperature, pH and salinity (Fig. [2](#Fig2){ref-type="fig"}). While the abundance of strain PR1 was constant when incubating at 22, 30 and 37 °C, respectively, strain GP abundance was significantly reduced at 37 °C (*p* \< 0.05) when compared to the other tested temperatures. Strain GP also showed a lower abundance when incubated at pH 5.5, in comparison to cultures incubated in media at neutral (pH 7.2) and basic (pH 9.5) pH values (Fig. [2](#Fig2){ref-type="fig"}). As it is typically observed for members of the *Achromobacter* genus \[[@CR32]\], NaCl concentrations up to 4% (w/v) did not influence the abundance of strain PR1; however, its abundance was significantly reduced above this value (Fig. [2](#Fig2){ref-type="fig"}). Although the absolute amount of strain GP 16S rRNA copy numbers also decreased above 4% NaCl (w/v), the relative abundance of this strain in the consortium was significantly higher (ranging from 0.24% at 0% NaCl, to a maximum of 4.26% at 8% NaCl). Interestingly, the abundance of strain GP was significantly lower in complex media (Tryptic Soy Broth, TSA; Brain-Heart Infusion, BHI; and Reasoner's 2A medium, R2A) than in mineral media with succinate and trace amounts of yeast extract (MMSY, Fig. [2](#Fig2){ref-type="fig"}). These results suggest that strain GP is possibly oligotrophic, unlike previously described for members of the *Leucobacter* genus, which thrive in complex media, such as BHI enriched with peptone and yeast extract, as observed for *L. luti* RF6^T^ \[[@CR30]\]. Fig. 1Electron micrographs of frozen hydrated *Achromobacter denitrificans* strain PR1 (**a**) and strain GP (**b**). PM -- Plasma membrane; OM -- Outer membrane; FG -- Flagellum; CW -- Cell wall; C -- Carbon support grid Fig. 2Abundance of strain PR1 and strain GP after 15 h incubation at different pH, salinity (in DLB), temperatures (in MMSY) and media (R2A, TSA, BHI and MMSY). The values for copies of the 16S rRNA gene per ml are plotted in logarithmic scale. Values are the mean values of triplicates and the error bars represent the standard deviation. Significant differences in strain GP abundance are indicated by a, b, c and d (from higher to lower values of the mean) as determined by two-way ANOVA (pH, temperature and salinity) or one-way ANOVA (PR1/GP ratio in R2A, TSA, BHI and MMSY) and the Tukey test at *p* \< 0.05 within each tested condition \[[@CR33]\]
Analysis of the metagenome-assembled genome of strain GP {#Sec4}
--------------------------------------------------------
The analysis of the metagenomic contigs with SSU finder (rRNA small subunit) from CheckM \[[@CR34]\] revealed the presence of only two phylogenetic distinct organisms: one identified as *A. denitrificans* PR1 and the other as strain GP. The reconstruction of strain GP's genome from whole-consortium sequencing generated a metagenome-assembled genome (MAG) consisting of 11 contigs, with 3.84 Mb, 3621 coding sequences (CDS), 69.68% in G + C and a total mapped coverage of 61x (Table [1](#Tab1){ref-type="table"}). In spite of an enrichment step with 2-phenylethanol, only 18.5% of the total of reads obtained with Oxford Nanopore (ONT) and Illumina technologies were mapped to strain's GP MAG, while the remaining reads mapped to the complete genome of *A. denitrificans* PR1 (148x coverage in the consortium, see Additional file [2](#MOESM2){ref-type="media"} Table S1), previously determined \[[@CR35]\]. The MAG of strain GP encoded a complete rRNA operon and harbored two copies of the 5S and one copy of the 16S and 23S rRNA subunits, respectively. Moreover, analysis with tRNAscan-SE \[[@CR36]\] identified 44 tRNA encoding for all 20 amino acids. CheckM \[[@CR34]\] analysis showed high completeness and low contamination values for this assembly, as only 7 marker genes were not detected in the draft genome and 3 markers had 2 copies in the assembly (95.9% completeness and 0.6% contamination, respectively, see Additional file [2](#MOESM2){ref-type="media"} Table S2). Therefore, according to Bowers et al. \[[@CR37]\], these findings indicate that this methodology allowed the reconstruction of a high-quality MAG for strain GP (Table [1](#Tab1){ref-type="table"}). Table 1List of named species of the *Leucobacter* genus used in the phylogenetic and comparative studies. Assembly quality was calculated using QUAST \[[@CR49]\] with a minimum contig size set to 200 bp. Completeness and contamination were computed with CheckM \[[@CR34]\]. 16S rRNA pairwise similarity was computed with the global alignment tool in the EzBioCloud web server \[[@CR126]\]Strain16S rRNA pairwise similarity to GP (%)Genome\*/ assembly accession no.Completeness (%)Contamination (%)Number contigs/ scaffoldsGenome size (Mb)Contig N50G + C content (%)Number of CDSReference *Candidatus* Leucobacter sulfamidivorax' GP--QZLF0000000095.910.58113.84956,10469.683621This study *L. aridicollis* L-9^T^96.23QYAE0000000099.270.5883.56888,84767.33212This study *L. celer* subsp. *astrifaciens* CBX151^T^95.87GCA_001273835.198.830.002354.14349,81369.13661\[[@CR122]\] *L. chironomi* DSM 19883^T^95.43GCA_000421845.1100.000.88272.96268,43869.92662\[[@CR64]\] *L. chromiireducens* subsp. *chromiireducens* L-1^T^96.23QYAD00000000100.000.5863.22623,96067.02843This study *L. chromiireducens* subsp. *solipictus* TAN 31504^T^96.15QYAC0000000099.422.05113.54451,46168.93096This study *L. chromiiresistens* J31^T^96.23GCA_000231305.1100.000.0023.212,823,34370.32895\[[@CR31]\] *L. chromiiresistens* NS35496.37GCA_001477055.195.030.001942.7935,22070.82423\[[@CR142]\] *L. komagatae* DSM 8803^T^96.79GCA_006716085.198.541.7523.753,292,53066.63253Unpublished *L. luti* RF6^T^96.81QYAG00000000100.000.5853.621,858,86469.43088This study *L. massiliensis* 122RC15^T^96.52GCA_002982315.1100.001.75243.14232,65671.02789\[[@CR143]\] *L. musarum* subsp. *japonicus* CBX130^T^95.87GCA_001273855.199.560.581443.59248,15566.83311\[[@CR144]\] *L. musarum* subsp. *musarum* CBX152^T^95.87GCA_001273845.199.850.581253.44200,78566.83147\[[@CR144]\] *L. salsicius* M1-8^T^95.94GCA_000350525.199.420.00283.18197,63764.52741\[[@CR65]\] *L. triazinivorans* JW-1^T^96.4GCA_004208635.1100.000.8813.48--69.372978\[[@CR145]\] *L. zeae* CC-MF41^T^95.88QYAB0000000099.850.5863.472,226,77270.63042This study
Analysis of mobile and conjugative elements {#Sec5}
-------------------------------------------
The identification of potential plasmids and other mobilizable elements in the genome of strain GP was performed in silico by measuring differences in coverage and G + C content between the contigs of the draft assembly. Compared to the average values for all contigs, at least three contigs (5, 7 and 9) showed a significantly higher coverage, and lower G + C content (see Additional file [2](#MOESM2){ref-type="media"} Table S1). The differential coverage among contigs was observed consistently with both Illumina and ONT libraries, which were prepared from different biological replicates of the consortium. Therefore, these differences are unlikely to arise from library preparation and sequencing bias. The differences encountered suggest that these contigs may represent potential plasmids with an average copy number per cell of approximately 2--3 (contigs 5 and 9) and 9 (contig 7), respectively. Furthermore, conserved domain search and CONJscan revealed the presence of several elements linked to plasmid replication, stability, partitioning, conjugation, and mobility (Table [2](#Tab2){ref-type="table"}). Out of these three contigs, only contig 9 (11.8 kb) was marked as circular by Circlator \[[@CR38]\]; however, it had no relevant hits to other plasmids available in the National Center for Biotechnology Information (NCBI) database. Contrarily, contig 7 (22.5 kb) featured residual homology to a new plasmid found in *Cnuibacter physcomitrellae* XA^T^ (accession number CP020716.1, 4285 bp alignment with 99% identity to this plasmid), and the plasmid pKpn-35963cz from *Klebsiella pneumoniae* Kpn-35963cz (accession number MG252894.1, 2030 bp alignment with 99% identity to this plasmid). The respective homologous regions contained genes encoding for transposases and mercury resistance. Both contigs 7 and 9 carry a gene encoding for a putative relaxase (locus tag: D3X82_18105, D3X82_18250, respectively) with a TrwC family domain (accession no. pfam08751; E-value: 3.7e-28 and 7.6e-25, respectively), commonly observed in proteins from the MOB~F~ (mobility) family (e.g., TraA from *Arthrobacter* sp. Chr15, accession no. ABR67091.1 \[[@CR39]\]). This classification was further confirmed by CONJscan \[[@CR40]--[@CR42]\], which found that both D3X82_18105 (contig 7) and D3X82_18250 (contig 9) possess a highly conserved MOB~F~ domain (E-values of 5.3e-105 and 4.1e-106, respectively). Additional mobility elements were only found in contig 7. This contig was found to harbor a putative plasmid replication protein (locus tag: D3X82_18090; Family: RepA_C; accession no: pfam04796; E-value: 9.0e-07). In this way, according to Guglielmini et al. \[[@CR42]\] and Smillie et al. \[[@CR43]\], the presence of a MOB element in contigs 7 and 9 suggests these putative elements are mobilizable but non-conjugative. Contig 5, with 74 kb, was found to contain various integrative and conjugative elements (Table [2](#Tab2){ref-type="table"}) \[[@CR44]\]. Besides, this contig contained all antimicrobial resistance genes found in the genome of strain GP (*sul*1, *tet*(33), *aad*A1, *qac*E), as well as two copies of the *sad*A gene encoding for the previously described sulfonamide monooxygenase \[[@CR26]\]. Table [2](#Tab2){ref-type="table"} Homology searches for contig 5 against the NCBI database \[[@CR45]\] revealed residual homology to *Enterobacter cloacae* strain EclC2185's genomic island (accession number MH545561.1, 5187 bp alignment with 99% identity to the genomic island of this strain) containing a class I integron with multi-drug resistance genes (*aad*A1, *sul*1, and *qac*E). Other significant alignments included regions conferring mercury resistance (*Cnuibacter physcomitrellae* XA^T^ plasmid, accession number CP020716.1, 5928 bp alignment with 99% identity to this plasmid) and intergenic regions of the new plasmid pOAD2 from *Flavobacterium* sp. KI723TI (accession number D26094.1, 14,820 bp alignment with 94% to this plasmid). According to conserved domain search and CONJscan analyses, two putative MOB elements were found in contig 5: (i) D3X82_17470, a relaxase from the MOB~F~ family with a TwrC conserved domain (CONJscan domain search: E-value 1.3e-85); (ii) D3X82_17405, a relaxase from the MOB~P1~ family (CONJscan domain search: E-value 4.2e-40). Other essential mobilizable elements detected include a type IV coupling protein (T4CP, locus tag D3X82_17390) with a conserved VirD4 domain (CONJscan domain search: E-value 5.7e-40) and a type IV secretion protein (T4SS, locus tag D3X82_17385) with a VirB4 domain (CONJscan domain search: E-value 1.4e-25). According to Smillie et al. \[[@CR43]\], these three elements (T4SS, T4CP and relaxases), represented in four locus tags in strain GP, are at the core of plasmid conjugation, however, no other known accessory proteins were detected in our analysis, presumably due to incomplete assembly and/or low identity to previously characterized proteins from the mating-pair formation (MPF) system. In this way, no complete type IV secretion systems were detected in contig 5 suggesting this element may be mobile but possibly not conjugative. Table 2Genes and corresponding conserved domains linked to integrative, conjugative and resistance elements found in contigs 5, 7 and 9 from the draft assembly of strain GP. Families and E-values in bold indicate the best hits obtained with CONJscan \[[@CR146]\]. n.a., not applicableSystemContigLocus tagDescriptionFamilyAccessionE-value Replication, stabilization and partitioning5D3X82_17420Toxin protein from a toxin/antitoxin systemZeta_toxinpfam064141.6e-195D3X82_17550Plasmid partition protein ABcsQpartition_RepACOG1192TIGR034532.2e-435.7e-165D3X82_17555Single-stranded DNA-binding proteinSSB_OBFcd044968.5e-127D3X82_18090Plasmid replication proteinRepA_Cpfam047969.0e-07 Mobilization and conjugative elements5D3X82_17410Plasmid mobilization relaxosome protein MobCMobCpfam057135.7e-055D3X82_17470Conjugal transfer proteinTrwC**MOB**~F~pfam08751n.a.9.0e-971.3e-855D3X82_17385Type IV secretion protein (T4SS)**VirB4**n.a.1.40e-255D3X82_17390Coupling protein (T4CP)VirD4**T4CP2**COG3505n.a.4.2e-165.7e-405D3X82_17405RelaxaseRelaxase**MOB**~P1~pfam03432n.a.9.8e-104.20e-407D3X82_18105Conjugal transfer proteinTrwC**MOB**~F~pfam08751n.a.3.7e-285.3e-1059D3X82_18250Conjugal transfer proteinTrwC**MOB**~F~pfam08751n.a.7.6e-254.1e-106 Multi-drug and heavy metal resistance5D3X82_17365 D3X82_17695Sulfonamide monooxygenase SadANcnHCaiAAcyl-CoA_dh_2cd01159COG1960pfam080281.52e-841.93e-305.8e-205D3X82_17485Tet(A)/Tet(B)/Tet(C) family tetracycline efflux MFS transporterMFS_TetAcd173882.4e-1455D3X82_17505ANT(3″)-Ia family aminoglycoside nucleotidyltransferase AadA1PRK13746DUF4111PRK13746pfam134270e+ 001.31e-415D3X82_17510Quaternary ammonium compound efflux SMR transporter QacE delta 1Multi_Drug_Respfam008931.75e-285D3X82_17515Sulfonamide-resistant dihydropteroate synthase Sul1Pterin_bindpfam008091.8e-875D3X82_17685Mercury(II) reductaseMerATIGR020531.2e-171
Phylogenetic analysis {#Sec6}
---------------------
As reported previously, strain GP shares the highest 16S rRNA gene sequence similarity with members of the genus *Leucobacter,* 94.6--96.9% (see Additional file [2](#MOESM2){ref-type="media"} Table S3), below the 98.7% threshold currently used to define a new species \[[@CR26], [@CR46]\] and close to the 97% threshold used to define a new genus \[[@CR47]\]. The phylogenetic analysis inferred from the alignment of the near-complete 16S rRNA gene between all fully sequenced *Leucobacter* spp. showed that strain GP indeed clusters with *Leucobacter* spp. (see Additional file [1](#MOESM1){ref-type="media"} Figure S3). Nevertheless, the ANI values between strain GP and the type strains of the validly named species of this genus ranged between 80.0 and 82.1% (Fig. [3](#Fig3){ref-type="fig"}a, Additional file [2](#MOESM2){ref-type="media"} Table S3), well below the general species delimitation thresholds (94--96%) \[[@CR48], [@CR49]\], indicating that strain GP could not be affiliated to any of these species. Average amino acid identity (AAI) comparisons between this strain and the type strains of the validly named species of this genus ranged between 64.2 and 69.1% (Fig. [3](#Fig3){ref-type="fig"}b, Additional file [2](#MOESM2){ref-type="media"} Table S3). These values are near the lower edge of the typical genus delimitation boundaries (approximately 65%) \[[@CR49]\], and the specific interspecies boundaries found between the analyzed type strains of *Leucobacter* spp. (51.0--87.3%). This result was further supported by the percentage of conserved genes (POCP) \[[@CR50]\]. POCP values ranged between 46.7 and 56.5% (Fig. [3](#Fig3){ref-type="fig"}c, Additional file [2](#MOESM2){ref-type="media"} Table S3), which is also on the lower edge of the interspecies boundaries found for this genus (42.0--81.3%) and the value suggested by Qin et al. \[[@CR50]\] for new genus delimitation (55%). The G + C content of strain GP was of 69.7% (Table [1](#Tab1){ref-type="table"}), which, according to previous studies \[[@CR51]\] is within the expected G + C interval (10%) for organisms of the same genus. In fact, for the type strains of all validly named species of the *Leucobacter* genus, G + C content ranged between 64.5 and 71.0% (Table 1). Moreover, the phylogenetic analysis of 400 conserved proteins of *Leucobacter* spp. using the PhyloPhlAn pipeline \[[@CR52]\] revealed that although strain GP appears to share a common origin with the other isolates of *Leucobacter* spp. (Fig. [4](#Fig4){ref-type="fig"}), it also does not cluster with any of the analyzed strains. Fig. 3ANI (**a**), AAI (**b**) and POCP (**c**) heatmaps comparing values between strain GP and validly named species of the *Leucobacter* genus at the time of analysis
Core and softcore genome of *Leucobacter* spp. {#Sec7}
----------------------------------------------
Orthologs gene cluster analysis with GET_HOMOLOGUES \[[@CR54]\] revealed that *Leucobacter* spp. and strain GP core and softcore genome contain 456 and 885 orthologs gene clusters, respectively (see Additional file [1](#MOESM1){ref-type="media"} Figure S4). However, only a fraction of these (approximately 50%) could be functionally annotated with eggNOG-Mapper and BlastKOALA \[[@CR55], [@CR56]\]. This analysis revealed that most of these clusters are related to central metabolic pathways \[[@CR57]\], including nucleotide and amino acid metabolism (118 clusters), and carbohydrate and lipid metabolism (16 clusters) (see Additional file [2](#MOESM2){ref-type="media"} Table S4), respectively. Furthermore, these strains lack orthologs linked to antimicrobial resistance, quorum sensing, and biofilm formation, suggesting that they form a diverse and versatile genus with specific adaptations to different environments (see Additional file [2](#MOESM2){ref-type="media"} Table S4). Only a few of the fully sequenced *Leucobacter* spp. analyzed are free-living organisms isolated from wastewater or soil. These free-living strains did not form a clade. The majority of the strains form facultative symbiotic associations with arthropods, nematodes, and plants (see Additional file [2](#MOESM2){ref-type="media"} Table S3). While *Leucobacter* sp. AEAR \[[@CR58]\], whose genome has been directly reconstructed from whole genome sequences of the nematodes *Caenorhabditis angaria* and *Caenorhabditis remanei*, could not be isolated, all *Leucobacter* spp. symbionts were able to grow independently from their hosts. Nevertheless, the analysis of strain's AEAR genome revealed that it should be able to grow independently as all essential pathways seem to be present in its draft genome \[[@CR58]\]. This observation is further supported by the analysis of the genome of this strain (see Additional file [2](#MOESM2){ref-type="media"} Table S3). Unlike obligate symbionts, which often undergo extreme genome reduction \[[@CR59]--[@CR62]\], strain AEAR possesses a genome with similar size (3.54 Mb) and genetic density when compared to its closest relatives (Fig. [4](#Fig4){ref-type="fig"}). Moreover, strain AEAR forms a monophyletic clade with *Leucobacter* sp. Ag1 (accession no. GCA_000980875.1) and other 9 strains, which are all facultative symbionts from arthropod species able to grow independently from their hosts \[[@CR63]\]. These results suggest that the facultative living style may correlate with the phylogeny of the strains. However, further studies are necessary in order to understand the link between phylogeny and lifestyle within this phylogenetic group. Interestingly, strain GP appears to share many conserved genes with *L. chironomi* DSM 19883^T^ \[[@CR64]\], a facultative symbiotic bacterium isolated from a member of the *Chironomidae* family (56.49% POCP, Fig. [3c](#Fig3){ref-type="fig"}). Bidirectional best-hits (BDBH) analysis with GET_HOMOLOGUES of these two strains showed that they share 1372 orthologs gene clusters (data not shown), amounting to 38.6% of the total CDS of strain GP. Most of these genes are linked to central metabolic pathways. As strain GP, *L. chironomi* also carries iron-heme acquisition operons *hmu*TUV (accessions no. WP_024357741.1, WP_024357742.1 and WP_029747012.1, respectively) and *efe*UOB (accessions no. WP_02436012.1, WP_024356011.1 and WP_024356010.1, respectively), and a homolog of heme oxygenase (*hmu*O, accession no. WP_024356032.1). However, unlike in strain GP, *L. chironomi* does not bear operons *efe*UOB and *hmu*TUV adjacently in its genome. The *efe*UOB operon and *hmu*O are absent from the softcore genome of *Leucobacter* spp., but are shared between several members of this genus (data not shown). Furthermore, strain GP also carries a chromate transport protein A (locus tag D3X82_06990) which was confirmed to be linked to chromate resistance and a common feature shared among several members of the *Leucobacter* genus \[[@CR65]\]. Fig. 4Phylogenomic relationships between the *Leucobacter* genus and strain GP inferred from concatenated amino acid alignments of 400 universal proteins obtained with PhyloPhlAn \[[@CR52]\]. Representative members of genera *Microbacterium*, *Leifsonia*, *Gulosibacter*, *Agromyces* a n d *Arthrobacter* were included as outgroup. *Leucobacter* spp. strains sequenced in this study are marked with an asterisk, and sulfonamide degraders are shown in bold. Node labels indicate local support values obtained with FastTree using the Shimodaira-Hasegawa test \[[@CR53]\]. The scale bar represents the number of expected substitutions per site. The tree was rooted at the outgroup node and visualized with FigTree \[[@CR125]\]
Estimation of gene loss in strain GP {#Sec8}
------------------------------------
Prior studies suggested that strain GP was obligatorily dependent on *A. denitrificans* PR1 for growth, as no isolated colonies of this organism were recovered after incubation in several conditions \[[@CR26]\]. Surprisingly, despite its dependent phenotype, strain GP did not show significant genome reduction as it is commonly reported in symbiotic bacteria \[[@CR60]\]. In fact, the number of genes and the genome size of this strain was similar to the ones found in other members of the *Leucobacter* genus (Table [2](#Tab2){ref-type="table"}). These results suggest that, despite the PR1-dependent phenotype, strain GP differs from obligate parasites that, in the process of adapting to their hosts, undergo a process called reductive genome evolution, which results in relatively small genomes (often \< 1 Mb) \[[@CR66], [@CR67]\]. (Table [3](#Tab3){ref-type="table"}) Comparative genomic analysis of the *Leucobacter* genus revealed that the pangenome consists of 12,998 orthologous gene clusters. The clusters present in at least 90% of the *Leucobacter* spp. (28 of 31 genomes) were used as reference to determine potentially missing genes in the draft genome of strain GP. These results were carefully analyzed and manually curated, due to the high frequency of annotation errors associated with draft assemblies \[[@CR68], [@CR69]\]. In this way, of all these clusters, only 141 were present in 90% of the *Leucobacter* spp. and were apparently absent from the draft genome of strain GP. From these 141 clusters, only 9 clusters were non hypothetical genes and no alternative pathways were found in the draft genome of strain GP (Table [3](#Tab3){ref-type="table"}). Among these 9 clusters, only those linked with tetrapyrrole biosynthesis (*hem*ABCE) and thiol transporters (*cyd*DC) may be linked to the incapacity of strain GP to grow independently, as both systems are essential for the synthesis and correct assembly of cytochromes \[[@CR70]--[@CR73]\]. The possible absence of these regions from the genome of strain GP was further investigated by mapping the reads of its MAG against the genome of *L. chironomi* DSM 19883^T^. Visualization of the regions corresponding to these clusters on *L. chironomi* (Table [3](#Tab3){ref-type="table"}) further showed that no reads obtained from strain GP mapped to these regions (see Additional file [1](#MOESM1){ref-type="media"} Figure S5 and S6). The CydDC complex performs the transport of glutathione and L-cysteine and is responsible for maintaining an optimal redox balance in the periplasm \[[@CR74], [@CR75]\]. This balance is crucial for the correct assembly of cytochromes in the plasma membrane, and its loss is usually associated with increased sensitivity to high temperature and oxidative stress \[[@CR71]--[@CR73]\]. *hem*ABCE encodes the proteins involved in the synthesis of tetrapyrroles and, subsequently, heme which acts as a prosthetic group in many respiratory and non-respiratory cytochromes \[[@CR70]\]. To the best of our knowledge, only a few bacterial strains have been found to be incapable of de novo heme biosynthesis \[[@CR76]\]. These strains are mainly pathogenic and affiliated to *Haemophilus influenza*, with the exception of the recently described environmental isolate *Leucobacter* sp. strain ASN212 which requires exogenous heme for growth \[[@CR76]--[@CR78]\]. These organisms rely on complex heme-acquisition systems to thrive in iron-deficient environments and to synthesize essential heme-containing proteins. Functional analysis of the draft genome of strain GP revealed the presence of a heme ABC transport operon (*hmu*TUV) that encodes for a hemin-binding periplasmic protein HmuT (locus tag D3X82_13650), a permease protein HmuU (D3X82_13655) and an ATP-binding protein HmuV (D3X82_13660), respectively. This system has been extensively described and found to be highly conserved in the actinobacterium *Corynebacterium diphtheria* \[[@CR76]\]. However, in this organism, additional heme-binding genes (*hta*ABC) and a heme oxygenase *hmu*O were found to be essential for successful heme and iron-heme acquisition \[[@CR76]\]. A homolog to *hmu*O was found in the genome of strain GP (locus tag D3X82_07630). However, the conserved *hta*ABC operon, essential for exogenous heme-binding, appeared to be missing. Instead of this operon, strain GP possesses a different adjacent gene cluster encoding for a deferrochelatase/peroxidase EfeB (D3X82_13665), an iron uptake system component EfeO (D3X82_13670) and a ferrous iron permease EfeU (D3X82_13675). These enzymes have been previously linked to ferrous/ferric iron acquisition in *Bacillus subtilis* \[[@CR79]\] and intact heme transport in *Escherichia coli* \[[@CR80]\]. However, to the best of our knowledge, the EfeUOB system has not been directly linked to intact heme-acquisition in Gram-positive bacteria. In previous studies \[[@CR26]\], we have supplied the consortium with exogenous heme and known heme precursors such as coproporphyrin III, coproporphyrin III tetramethylester and coproporphyrin I dihydrochloride, replicating the conditions that allowed the isolation of the heme-dependent *Leucobacter* sp. ASN212 \[[@CR26], [@CR77]\]. However, adding these metabolites to agar plates did not abolish the dependent phenotype of strain GP. This result was unexpected as strain GP possesses several downstream genes of the porphyrin pathway; therefore, it should at least be able to use coproporphyrin III as a heme precursor. This finding suggests that either the heme transport system of strain GP is insufficient for intact heme transport across the thick peptidoglycan cell wall or that other essential cofactors or conditions are missing. Unfortunately, the genome of strain ASN212 is not publicly available, which hinders further efforts to characterize the transport of intact heme and heme precursors across the cell wall of this heme-dependent actinobacterium. Considering that strain GP may also lack homologs to thiol ABC transporters (accession numbers WP_024357159.1 and WP_042544611.1), it may be vulnerable to oxidative stress and unable to correctly assemble cytochromes \[[@CR71]--[@CR73]\]. In other sensitive organisms, the absence of this redox regulating system has been compensated by growing deficient strains in catalase-containing media and even in co-culture with catalase-producing bacteria \[[@CR71], [@CR81]\]. However, none of these strategies allowed the independent growth of strain GP. Genomic studies are inherently limited because they rely on gene homology for functional annotation and prediction \[[@CR8]\]. Furthermore, draft genomes are known to present extensive annotation errors \[[@CR68]\], and the presence of a given gene may not even be translated into a particular phenotype. Indeed, genes can be silenced by mutations in the coding region or their promoters, a common phenomenon in the evolution of dependent and pathogenic bacteria which suffer progressive phenotypic and genetic changes due to the interaction with their hosts and the environment \[[@CR8], [@CR82], [@CR83]\]. Therefore, these preliminary results require further functional studies in order to understand gene expression and activity in both strains from the microbial consortium. Table 3Essential genes missing from the draft genome of strain GP identified by core/pangenome analysis with GET_HOMOLOGUES \[[@CR54]\] Representative accession no. *L. chironomi* DSM 19883^T^Strain PR1KO identifiersDescriptionPathway/System WP_024356349.1ASC67664.1K01476Arginase RocFL-arginine biosynthesis; Urea cycle WP_024355584.1AbsentK08963, K08964*S*-methyl-5-thioribose-1-phosphate isomerase MtnAMethionine salvage pathway WP_024357159.1ASC65015.1K06147, K06148, K16013, K16014Thiol reductant ABC exporter subunit CydDGlutathione; L-cysteine ABC transporter WP_024357158.1ASC65016.1K06148, K16012Thiol reductant ABC exporter subunit CydC WP_024356490.1ASC65168.1K02492Glutamyl-tRNA reductase HemAPorphyrin and chlorophyll metabolism WP_024356487.1ASC64797.1K01698Porphobilinogen synthase HemB WP_084705356.1ASC63016K01749Porphobilinogen deaminase HemC WP_024356489.1ASC64317.1K01599Uroporphyrinogen decarboxylase HemE WP_024356124.1ASC67862.1K02083, K06016Allantoate deiminase AllCPurine metabolism
Unique genes shared between sulfonamide degraders {#Sec9}
-------------------------------------------------
### Genetic synteny {#Sec10}
As previously discussed, strain GP was shown to be responsible for the breakdown of sulfonamides in the two-member consortium \[[@CR26]\]. Some members of the *Microbacterium* genus able to degrade sulfonamides carried a conserved gene cluster encoding for two monooxygenases (SadA and SadB, accession no. WP_100812327.1 and WP_036299419.1, respectively) and one FMN-reductase (SadC, WP_100812326.1). SadA is known to catalyze the *ipso*-hydroxylation of SMX releasing 4-aminophenol, while SadB appears to be responsible for the further oxidation of this unstable and transient metabolite \[[@CR26], [@CR84]\]. Although the former enzymes appear to be highly specific for these substrates, the role of SadC can easily be fulfilled by other enzymes with similar activity. This has previously been demonstrated in assays with transformed *E. coli* that contained an incomplete cluster encoding for SadA and SadB alone \[[@CR84]\]. The genetic synteny in *Microbacterium* spp. (Fig. [5](#Fig5){ref-type="fig"}) appears to be highly conserved \[[@CR84]\]. Indeed, all strains of this genus harbor a cluster consisting of a *trw*C relaxase (WP_100812428.1), a polyisoprenoid-binding protein *yce*I (WP_100812427.1), a *sad*A monooxygenase (WP_100812327.1), a *sad*B monooxygenase (WP_036299419.1) and a *sad*C flavin monooxygenase (WP_100812326.1). Except for the putative sulfonamide degrader *Microbacterium* sp. CJ77, that carries an additional gene within the *sad* cluster encoding for an IS3 family transposase (WP_103663393.1), located between *trw*C and *yce*I (Fig. [5](#Fig5){ref-type="fig"}). Likewise, strain GP carries homologs for most of these genes, albeit with a different synteny. This strain harbors at least two identical copies of a homolog of *sad*A (locus tag D3X82_17695 and D3X82_17365) in contig 5 (Fig. [5](#Fig5){ref-type="fig"}). Both copies of *sad*A are flanked by an SOS response-associated peptidase (D3X82_17690 and D3X82_17360, respectively) and a polyisoprenoid-binding protein *yce*I (D3X82_17700 and D3X82_17370, respectively), which, in turn, is flanked by a single IS1380 family transposase (D3X82_17710 and D3X82_17380, respectively) (Fig. [5](#Fig5){ref-type="fig"}). Fig. 5Representation of the genetic organization of the *sad* cluster in sulfonamide degraders: *Microbacterium* sp. strains BR1 and CJ77, *Arthrobacter* sp. strains D2 and D4 and strain GP. Contig numbers and locus for each region are shown next to or on top of the DNA backbone. The diagram was designed with Simple Synteny \[[@CR85]\]
### Phylogenetic analysis {#Sec11}
The analysis of proteins associated with sulfonamide degradation showed that the SadA and YceI homologs shared a high percentage of amino acid identity between *Microbacterium* sp. and strain GP, while the two *Arthrobacter* sp. isolates were more similar among themselves (see Additional file [1](#MOESM1){ref-type="media"} Figure S7a and d). The high identity between these homologs among strains affiliated to different bacterial species suggests that these genetic determinants may share a common ancestor. This hypothesis was further supported by phylogenetic studies of these proteins. For instance, the Maximum Likelihood (ML) phylogenetic tree shows that all SadA and YceI homologs form a conserved clade with high support values (Fig. [6](#Fig6){ref-type="fig"}a and d), also obtained when using both cladistic (Bayesian inference) and distance-based (Neighbor Joining, NJ) methods. Furthermore, the co-existence of both the SadA monooxygenase and the YceI transporter in all the genomes suggests that the YceI binding-protein may play a complementary role in the sulfonamides degradation by enhancing the uptake of these molecules, as previously described for other systems \[[@CR86]\]. In contrast, SadB and SadC homologs were highly identical in all *Microbacterium* sp. isolates and *Arthrobacter* sp. D2 (see Additional file [1](#MOESM1){ref-type="media"} Figure S7b and c). And the phylogenetic analysis revealed that these proteins form a highly conserved clade (Fig. [6b and c](#Fig6){ref-type="fig"}). Despite the lower identity between the SadB homologs in *Arthrobacter* sp. D4 and strain GP, in comparison to the other sulfonamide-degraders (see Additional file [1](#MOESM1){ref-type="media"} Figure S7b), these proteins also appear to share a common ancestor with their homologs in *Microbacterium* spp. (Fig. [6b](#Fig6){ref-type="fig"}). Conversely, the SadC homologs found in these two strains do not appear to share a common ancestor between themselves nor with the other sulfonamide-degrading strains. This result is in agreement with previous studies with the recombinant SadABC complex from *Microbacterium* sp. BR1 \[[@CR84]\]. This study showed that SadA and SadB are sufficient to carry out complete SMX degradation in recombinant *E. coli*, suggesting that the role of SadC could be fulfilled by other flavin reductases present in the genome of the host strain \[[@CR84]\]. Noticeable, the IS1380 transposase flanking SadA in strain GP is identical to a homolog in *Microbacterium* sp. BR1 (WP_100810554.1, amino acid identity of 100%), and these two proteins form a highly conserved clade in the phylogenetic tree (Fig. [6e](#Fig6){ref-type="fig"}). Interestingly, the IS1380 family transposase is located far from the *sad* cluster in *Microbacterium* sp. BR1 (contig 4, instead of contig 9), suggesting that this transposase may be involved in gene mobility in different species and genera of the *Actinobacteria* phylum. Fig. 6Maximum likelihood phylogenetic trees inferred from amino acid alignments with MEGA6 \[[@CR87]\] of (**a**) SadA, (**b**) SadB, (**c**) SadC, (**d**) YceI transporter and (**e**) IS1380/IS3/IS4 transposases shared between sulfonamide degraders. Strain GP is shown in bold; sulfonamide degraders are marked with an asterisk (\*); and structural homologs to these enzymes obtained with SWISS-MODEL \[[@CR88]\] are shown in bright blue. Node labels indicate: ML bootstrap support above 50% (in bold) / NJ bootstrap support values above 50% / Bayesian posterior probabilities above 70%. The scale bar represents the number of expected substitutions per site. The tree was rooted at the midpoint and visualized with FigTree \[[@CR125]\]
### Structural analysis of SadA and SadB {#Sec12}
Structurally, all SadA and SadB homologs contain an acyl-CoA dehydrogenase domain (see Additional file [1](#MOESM1){ref-type="media"} Figure S8), classifying them as Group D flavoprotein monooxygenases \[[@CR26], [@CR89]\]. Furthermore, structural homology search with SWISS-MODEL \[[@CR88]\] resulted in highest similarity with XiaF (FADH~2~) from *Streptomyces* sp. HKI0576 (PDB: 5LVW/5LVU), 4-hydroxyphenylacetate hydroxylase (4-HPA) from *Acinetobacter baumannii* (PDB: 2JBR), HsaA monooxygenase from *Mycobacterium tuberculosis* and *Rhodococcus jostii* RHA1 (PDB: 3AFE and 2RFQ, respectively). All of these monooxygenases are known to hydroxylate aromatic compounds. For instance, XiaF is likely involved in terpenoid biosynthesis in *Streptomyces* sp.; it is tetrameric and acts in a two-component system together with a flavin reductase \[[@CR90]\]. Furthermore, this monooxygenase can use indole as a surrogate substrate and form indigo and indirubin, as previously described \[[@CR90]\], while 4-HPA (EC 1.14.14.9) and HsaA monooxygenases (EC 1.14.14.12) catalyze the insertion of oxygen in the benzene ring of 4-hydroxyphenylacetate or 3-hydroxy-9,10-seconandrost-1,3,5(10)-triene-9,17-dione, respectively. The similarities between XiaF and SadA (32% amino acid identity, 51% similarity, 4% gaps and E-value of 8e-66), SadB1 (locus tag: D3X82_00235, 33% identity, 51% similarity, 2% gaps and E-value of 8e-68), and SadB2 (locus tag: D3X82_03160, 35% identity, 52% similarity, 1% gaps and E-value of 9e-69) are sufficient to suggest a high degree of confidence on the homologous relationship between these proteins \[[@CR91], [@CR92]\]. The use of XiaF as a template for modelling resulted in a robust structural prediction of these proteins with quality scores above − 4 (QMEAN, Qualitative Model Energy ANalysis) \[[@CR93]\]: − 2.45, − 1.48 and − 1.52, for SadA, SadB1 and SadB2, respectively. These results suggest that the comparison with XiaF is suitable to perform preliminary structural analysis of these monooxygenases. Phylogenetic analysis revealed that XiaF shares a common ancestor with other xenobiotic-degrading enzymes \[[@CR90]\], suggesting that both monooxygenases of the SadABC complex likely share ancestry with similar enzymes. However, pairwise sequence alignment revealed a low degree of similarity within the substrate binding pocket of XiaF and homologs regions of SadA and SadB. For instance, the presence of isoleucine I237 in an alpha-helix of XiaF (Fig. [7a](#Fig7){ref-type="fig"}) constricts the size of the substrate binding pocket of this monooxygenase. However, alanine residues neighboring I237 are substituted in strains GP, *Microbacterium* sp. BR1 and *Arthrobacter* sp. D2 and D4 by proline residues (P261 and P264 in strain GP) that may induce its structural change to a loop. This conformation probably creates a wider pocket in SadA and may allow easier access to the active site of this monooxygenase. Additional substitutions in the active site of all SadA homologs further support this hypothesis. Specifically, the serine residue S121 (Additional file [1](#MOESM1){ref-type="media"} Figure S7a and Figure S9) of XiaF is substituted by an alanine in strain GP (A145) and threonine in *Microbacterium* sp. BR1 (T144) and both *Arthrobacter* sp. (T135). The serine residue has a hydroxymethyl side chain, while alanine and threonine have a methyl and 1-hydroxyethyl groups, respectively. Serine and threonine are both polar amino acids and likely make the active site of these enzymes amenable for polar cyclic substrates while SadA of strain GP may prefer aromatic substituted substrates. Furthermore, the alanine residue (A145) in the SadA of strain GP would make the active site of this enzyme slightly larger than the active sites of SadA of *Microbacterium* sp. BR1 and XiaF. These findings may explain the differences in SMX degradation rate found between the consortium of strain GP and *A. denitrificans* PR1 and axenic cultures of *Microbacterium* sp. BR1 \[[@CR94], [@CR95]\]. For instance, in resting cells conditions the specific degradation rates of the axenic cultures and the consortium were similar (approximately 2 μmol/g~cell\ dry\ weight~ x min) \[[@CR94], [@CR95]\], however, considering that the abundance of strain GP is significantly low (1--4% relative abundance) \[[@CR26]\], this strain appears to be more efficient than *Microbacterium* sp. BR1. Furthermore, in the consortium, 4-aminophenol never accumulated in sufficient amounts to be detected and was only observed in incubations with ^14^C-SMX saturated with an excess of the unlabeled 4-aminophenol \[[@CR26]\]. Conversely, the substitutions and subsequent changes in the structure of SadB are harder to predict. For instance, in all SadB homologs, the XiaF S121 is substituted by a valine (isopropyl side chain, Fig. [7b](#Fig7){ref-type="fig"}). Furthermore, the A236 of XiaF is substituted by tryptophan in all SadB homologs (W231 in SadB1 of strain GP) suggesting that SadB's binding pocket could be significantly smaller than XiaF's and thus accommodate smaller substrates. Despite the low amino acid identity between some of the SadB homologs (see Additional file [1](#MOESM1){ref-type="media"} Figure S7b), the analysis of the conserved domains indicates that the active site could be highly conserved among these enzymes. In this way, although none of the expected metabolites were detected during 4AP degradation in strain GP \[[@CR26]\], the presence of homologs of SadB in the genome of this strain suggests that it might catalyze 4AP hydroxylation as previously described for *Microbacterium* sp. BR1 \[[@CR96]\]. Fig. 7Pairwise alignment with BLASTp of the regions of the substrate binding pocket of XiaF (accession number 5LVW) and each homolog of SadA (**a**) and SadB (**b**) in strains GP, *Microbacterium* sp. BR1 and *Arthrobacter* sp. D2 and D4. Conserved regions between the different SadA and SadB homologs are highlighted in green, nonconserved residues are highlighted in red. Residues shared by all sequences are marked with an asterisk. The diagrams were designed with Excel 2013
Taxonomic classification of strain GP {#Sec13}
-------------------------------------
The total dependency of strain GP on strain PR1 and the lack of similar organisms hinder further efforts for accurate taxonomic classification. Nevertheless, according to the recommendations of the International Committee on Systematic Bacteriology, organisms unable to grow in pure culture can have a provisional taxonomic status (*Candidatus*) \[[@CR97]\]. When comparing our data with the standards proposed by Konstantinidis et al. \[[@CR98]\] to describe uncultivated prokaryotes and/or those forming microbial consortia, we propose to classify strain GP in a provisional new species within the genus *Leucobacter*, '*Candidatus* Leucobacter sulfamidivorax'.
Description of '*Candidatus* Leucobacter sulfamidivorax' {#Sec14}
--------------------------------------------------------
'*Candidatus* Leucobacter sulfamidivorax' \[sul.fa.mi.di.vo'rax N.L. n. *sulfamidum*, sulfonamide; L. adj. *Vorax* devouring, ravenous, voracious; N.L. masc. Adj. *sulfamidivorax*, sulfonamide-degrading\]. Forms a bacterial consortium with *A. denitrificans* strain PR1 and can only be cultured in association with this *Proteobacteria*. Cells stain Gram-positive, present a rod-shaped morphology (1.3 ± 0.03 μm long and 0.5 ± 0.03 μm wide), and are probably non-motile. It produces light yellow-colored colonies with less than 1 mm in diameter on top of the colonies of *A. denitrificans* strain PR1 after 10 d of incubation on 25% (w/v) BHI plates at 30 °C. In liquid medium, it constitutes between 1 and 4% of the total cells. In medium MMSY, the aerobic growth is significantly impaired at 37 °C when compared to that at 22 and 30 °C. Growths well at neutral (pH 7.2) and basic (pH 9.5) pH values when compared to that at pH 5.5. Grows better in oligotrophic media (e.g., MMSY), in comparison to complex and rich medium (e.g., BHI and TSA). Tolerates up to 8% (w/v) NaCl. The DNA G + C content is 69.68 mol%. The representative strain, GP, which degrades sulfonamides, was obtained from a sulfamethoxazole enrichment culture produced from activated sludge from an urban WWTP, in North Portugal, in 2011.
Conclusions {#Sec15}
===========
The genomic analysis showed that strain GP carries at least two copies of *sad*A encoding for the previously described sulfonamide monooxygenase. Both copies are flanked by a single IS1380 family transposase and were found in contig 5 that represents a potential plasmid carried by strain GP. Noticeably, a highly similar *sad*A-containing cluster was also found in the genomes of *Arthrobacter* sp. D2 and D4. All sulfonamide-degrading *Actinobacteria* harbored homologs to *sad*B and *sad*C, nevertheless, in strains GP and *Arthrobacter* sp. D2 and D4 these genes were not in the vicinity of *sad*A and were not associated neither with mobile nor integrative elements. Functional analysis of strain GP genome revealed that this strain may have lost some essential genes, mainly of genes linked to tetrapyrrole biosynthesis and thiol transporters. These results strongly suggest that strain GP may be unable of synthesizing respiratory and non-respiratory cytochromes, essential for aerobic growth, and may need a helper strain to provide exogenous heme and help maintain an optimal redox balance. However, supplying strain GP with exogenous heme and catalase did not abolish this strain's dependent phenotype. Additional studies are necessary to evaluate the gene expression in strain GP and the mechanisms of intact heme acquisition in this Gram-positive bacterium. Our data suggests that strain GP should be considered as the representative strain of a putatively new species within the *Leucobacter* genus, '*Candidatus* Leucobacter sulfamidivorax'.
Methods {#Sec16}
=======
Culture conditions and DNA extraction {#Sec17}
-------------------------------------
Five type strains of the genus *Leucobacter* were selected for comparative studies based on 16S rRNA gene pairwise similarity to strain GP (Table [1](#Tab1){ref-type="table"}) \[[@CR26]\] and purchased from DMSZ (Germany). These strains were grown in Brain-Heart Infusion (BHI, Sigma) for 15 h. All incubations were carried out in the dark at 30 °C under continuous shaking at 120 rpm. The two-member consortium \[[@CR26]\] consisting of *Achromobacter denitrificans* PR1 (LMG 30905) and strain GP was incubated for 7 d in mineral medium with 0.2 g/l yeast extract, 4 mM ammonium sulfate, 700 mM succinate, 0.6 mM SMX, and 2.5 g/L 2-phenylethanol (Sigma) as an inhibitor of Gram-negative cells (MMSY-SMX-PE). Further attempts to isolate strain GP were performed by incubating the consortium in 25% BHI agar plates (v/v) with 0.6 mM SMX, heme or heme precursors (10 μg/l, coproporphyrin III, coproporphyrin III tetramethylester, coproporphyrin I dihydrochloride) \[[@CR26]\], putrescine (9 μg/l) and catalase (500 U) from Sigma, respectively. Genomic DNA extraction of the *Leucobacter* spp. type strains and the two-member consortium was performed from 2 × 10^10^ cells with GenElute Bacterial Genomic DNA Kit (Sigma) as previously described \[[@CR26]\].
Physiological characterization of the consortium {#Sec18}
------------------------------------------------
The effect of environmental parameters on the abundance of each strain of the consortium was investigated. The effect of temperature was examined by incubating the culture in MMSY at 22 °C, 30 °C and 37 °C. The influence of pH was tested at 30 °C in diluted Lysogeny broth medium (DLB, 25% w/v) with 12 mM of MES (pH 5.5), 12 mM phosphate buffer (pH 7.2) or 12 mM of CAPS (pH 9.5) at 30 °C. The tolerance to NaCl was examined in DLB supplemented with NaCl at final concentrations of 2, 4, 6, 8 or 10% (w/v) at 30 °C. To determine the influence of different standard media in the growth of strain GP, the consortium was incubated in unbuffered R2A, TSA and different dilutions of BHI (5, 25, 50, 75 and 100%). Cultures under all these conditions were incubated at 30 °C for 15 h and carried out in triplicate and in parallel to an abiotic control. The abundance of each strain in the consortium was assessed by quantitative PCR with primers targeting the 16S rRNA gene as previously described \[[@CR26]\]. Significant differences (*p* \< 0.05) between overall abundance of strain GP were determined either by two-way ANOVA (to compare 16S rRNA copies/ml of GP and PR1 at different pH, temperature and salinity) or one-way ANOVA (to compare the ratio of the 16S rRNA copies/ml of strains GP and PR1 across different media) and Tukey's tests using RStudio v 1.1.463 running with R v3.5.2 \[[@CR33], [@CR99], [@CR100]\].
Electron microscopy {#Sec19}
-------------------
The consortium was visualized in mid-stationary phase (12 h incubation, MMSY, 0.6 mM SMX) by Cryo-Transmission Electron Microscopy (Cryo-TEM) for morphological characterization. Briefly, a 4 μl aliquot of the overnight grown liquid culture was adsorbed onto a holey carbon-coated grid (Lacey, Tedpella, USA), blotted with Whatman 1 filter paper and vitrified into liquid ethane at − 180 °C using a vitrobot (FEI, USA). Frozen grids were transferred onto a Talos Electron microscope (FEI, USA) using a Gatan 626 cryo-holder (GATAN, USA). Electron micrographs were recorded at an accelerating voltage of 200 kV using a low-dose system (20 to 40 e−/Å2) and keeping the sample at − 175 °C. Defocus values were − 3 to 6 μm. Micrographs were recorded on 4 K × 4 K Ceta CMOS camera. The cell size, and periplasmic and cell wall thickness were measured with Fiji from the ImageJ platform \[[@CR101]\]. For Transmission Electron Microscopy (TEM) analyses, 4 μl aliquot of the sample was adsorbed onto a glow-discharged carbon film-coated copper grid, and subsequently negatively stained with 2% uranyl acetate. Images were recorded using Philips CM200FEG electron microscope operating at 200 kV on TemCam-F416 CMOS camera (TVIPS, Germany).
*Leucobacter* spp. type strains whole-genome sequencing and assembly {#Sec20}
--------------------------------------------------------------------
High-quality DNA of the selected *Leucobacter* spp. type strains (Table [1](#Tab1){ref-type="table"}) was used for paired-end sequencing (2 × 150 bp) with the Hiseq 2500 platform (Illumina) by GATC Biotech (Germany). Paired-end reads were adapter and quality trimmed (≥ Q20) with the BBDuk tool from the BBMap package v35.74 \[[@CR102]\]. High-quality reads were used for de novo assembly with SPAdes v3.11.1 \[[@CR103]\] with the option --careful. Contigs longer than 500 bp were used further extension with SSPACE v3.0 \[[@CR104]\] with recommended settings. All data has been deposited in NCBI under the BioProject accession number PRJNA489769.
Whole consortium sequencing {#Sec21}
---------------------------
The metagenome of the consortium was sequenced in-house in the Miseq (Illumina) and MinION (Oxford Nanopore Technologies, ONT) platforms. The paired-end Miseq library was prepared from 1 μg of high-quality DNA with KAPA HyperPrep Kit (Kapa Biosystems) and TruSeq DNA PCR-free LT Kit library adaptors (Illumina) with a few modifications. Briefly, enzymatic fragmentation of the genomic DNA was increased to 25 min, and ligation was performed for 2 h at 20 °C. Eight cycles of enrichment PCR and size selection for fragments with approximately 500--700 bp was carried out with NucleoMag magnetic beads (Macherey Nagel). Paired-end sequencing (2 × 250 bp) was performed in an Illumina Miseq system (Illumina) with a V2 MiSeq Reagent Kit (500 cycles). Two independent libraries were prepared for MinION long-read sequencing. Both libraries were prepared from 1.5 μg high-quality DNA sheared with a g-TUBE (Covaris) to approximately 8 kb fragments. The libraries were then prepared with the 1D genomic DNA sequencing kit (SQK-LSK108), pooled, loaded with the Library Loading Bead Kit (EXP-LLB001) and sequenced using a flow cell with R9.4 chemistry (FLO-MIN 106, Oxford Nanopore).
Metagenome-assembled genome (MAG) of strain GP {#Sec22}
----------------------------------------------
ONT long reads were adapter trimmed with Porechop v0.2.3 \[[@CR105]\]. Illumina paired-end reads were adapter and quality trimmed (≥Q20) with BBDuk from the BBMap package v35.74 \[[@CR102]\]. The high-quality paired-end reads were used for hybrid error correction of ONT reads with LoRDEC v0.9 \[[@CR106]\]. Resulting long reads were subsequently used for whole-metagenome assembly with Canu v1.7 \[[@CR107]\]. Metagenomics contigs were analyzed with SSU finder from CheckM v1.0.11, to determine the amount and affiliation of taxonomic bins present in the metagenome \[[@CR34]\]. The metagenome was aligned to the complete genome of strain PR1 (Genbank accession no. CP020917) \[[@CR35]\] with BLASTn v2.7.1+ to remove contigs affiliated to the proteobacterium \[[@CR108]\] (e-value, identity and hit length threshold cutoffs set to 1e-10, 80 and 30%, respectively). Contigs without significant hits were retrieved from the metagenome and used to construct the new taxonomic bin corresponding to strain GP. Both ONT Illumina-corrected and Illumina reads were used for read binning between strain PR1 and GP with GraphMap v0.5.2 \[[@CR109]\] and BWA v0.7.12 \[[@CR110]\], respectively. Reads mapping uniquely to strain GP bin were used for hybrid re-assembly with SPAdes v3.7.1 \[[@CR103]\]. High-coverage contigs (≥ 1x *k-mer* coverage) obtained with hybrid assembly were used for further scaffolding and polishing with Circlator v1.5.5 \[[@CR38]\] and four iterations with Pilon v1.22 \[[@CR111]\]. All data has been deposited in NCBI under the Bioproject accession number PRJNA490017.
Genome annotation, completeness and mobile genetic elements {#Sec23}
-----------------------------------------------------------
Quality scores of draft assemblies were assessed with QUAST v4.6.3 \[[@CR49]\]. Genome contamination and completeness were determined with CheckM v1.0.11 \[[@CR34]\], and tRNA were identified with tRNAscan-SE v2.0 \[[@CR36]\]. Open-reading frames (ORFs) were predicted and annotated with NCBI Prokaryotic Genome Annotation Pipeline (PGAP) v4.7 \[[@CR112]\] and with RASTtk on the RAST webserver v2.0 \[[@CR113]\]. Antibiotic resistance genes were confirmed by aligning amino-acid sequences with BLASTp v2.7.1+ against the Antibiotic Resistance Database (ARDB) v1.1 from July, 2009 \[[@CR114]\] and by analyzing the draft genome with the Resistance Gene Identifier (RGI) against the CARD database v3.0.1 \[[@CR115]\]. Functional annotation and KEGG Orthology (KO) assignment was further performed with eggNOG-Mapper v4.5.1 \[[@CR55]\] and BlastKOALA v2.1 \[[@CR55]\]. The presence of plasmids in the genome of strain GP was investigated by assessing differences in coverage and G + C content between contigs, and by further searching for similarities with other plasmids using NCBI BLAST against the non-redundant (*nr*) database on November, 2018 \[[@CR116]\]. The differences in coverage were identified by mapping both Illumina and ONT reads against the metagenome of the consortium (concatenated draft assembly of strain GP and complete genome of strain PR1) with BWA v0.7.12 \[[@CR110]\] or Graphmap v0.5.2 \[[@CR109]\], respectively. The coverage of sorted BAM files was evaluated with Qualimap v2.2.1 \[[@CR117]\]. Genes typically associated with plasmids \[[@CR44]\] were identified by aligning amino-acid sequences against the CDD database (v3.17) using the NCBI conserved domain search on with default settings on November, 2018 \[[@CR118]--[@CR120]\]. Conjugative elements associated with the type VI secretion systems and possible origins of replication were analyzed with CONJscan v1.0.2 using the Galaxy platform at Pasteur \[[@CR40]--[@CR42]\].
Phylogenetic analysis of strain GP {#Sec24}
----------------------------------
The full genome and the 16S rRNA gene of all fully sequenced *Leucobacter* spp. isolates and MAGs were used for phylogenetic analysis (see Additional file [2](#MOESM2){ref-type="media"} Table S3). Sequences were retrieved from the NCBI database (last accessed on November, 2018) \[[@CR45]\], except for *Leucobacter* sp. AEAR which was available on GitHub \[[@CR121]\]. Moreover, representative genomes of *Microbacterium*, *Leifsonia*, *Agromyces* and *Arthrobacter* genera, were further included to serve as outgroup (see Additional file [2](#MOESM2){ref-type="media"} Table S3). 16S rRNA sequences were used for multiple sequence alignment with MUSCLE in MEGA6 \[[@CR87], [@CR122]\]. The phylogenetic tree was inferred from maximum likelihood analyses using MEGA6 \[[@CR87]\] with the best-fitting model: Tamura-Nei \[[@CR123]\] substitution model with gamma distribution and invariant sites (TN93 + G + I). Bootstrap support values were inferred from 1000 replicates. The PhyloPhlAn pipeline v0.99 \[[@CR52]\] was used to infer phylogenomic relationships among fully sequenced members of the *Leucobacter* genus and strain GP (data available on July 2019, see Additional file [2](#MOESM2){ref-type="media"} Table S3). 400 universal proteins were identified and extracted with USEARCH v5.2.32 \[[@CR53]\] and used for amino acid alignments with MUSCLE v3.8.31 \[[@CR122]\]. The concatenated alignments were used for approximately-maximum-likelihood analysis with FastTree v2.1.8 \[[@CR124]\] and the computation of local support values was performed using the Shimodaira-Hasegawa test \[[@CR53]\]. Both the 16S rRNA phylogenic tree and the PhyloPhlAn phylogenomic tree were visualized with FigTree v1.4.3 \[[@CR125]\] and rooted at the midpoint or the outgroup, respectively. 16S rRNA gene pairwise sequence similarity, Average Nucleotide Identity (ANI), Average Amino-acid Identity (AAI), and Percentage of Conserved Proteins (POCP) \[[@CR50]\] between strain GP and the validly named and fully sequenced strains of the *Leucobacter* genus on July 2019 (see Additional file [2](#MOESM2){ref-type="media"} Table S3) were determined using the pairwise similarity tool and 16S-based ID app available on the EzBioCloud platform \[[@CR126]\], AAI/ANI-matrix from the enveomics toolbox web server \[[@CR127]\] and the POCP.sh script developed by Harris et al. \[[@CR128]\] and publicly available on figshare \[[@CR129]\]. In addition, the 16S rRNA pairwise sequence similarity, ANI, AAI and POCP values were determined for seven *Leucobacter* strains for which the genome sequences become available during the revision of the present manuscript (see Additional file [2](#MOESM2){ref-type="media"} Table S3).
*Leucobacter* spp. core and pangenome analysis {#Sec25}
----------------------------------------------
Gene search between *Leucobacter* spp. genomes (last accessed on November 2018) and strain GP (see Additional file [2](#MOESM2){ref-type="media"} Table S2) was computed with the GET_HOMOLOGUES package v3.1.4 \[[@CR54]\] using all-against-all NCBI BLAST v2.2 with default settings and Pfam-domain scanning. Clustering was performed with COGtriangles v2.1 (−G -t 0 -D), orthoMCL v1.4 (−M -t 0 -D) and BDBH (−D -e) algorithms. For comparison between two genomes only BDBH was used. For the pangenome analysis, the clusters were generated from the intersection of COGtriangles and orthoMCL using the compare_clusters.pl script (−t 0) from the GET_HOMOLOGUES pipeline \[[@CR54]\]. To determine core metabolic pathways shared between *Leucobacter* spp. and strain GP, orthologous gene clusters present in the core and softcore genome (100 or 95% of the genomes) \[[@CR130]\] were used for functional annotation with eggNOG-Mapper \[[@CR55]\] and BlastKOALA \[[@CR55]\]. The list of KO identifiers was used to visualize and analyze core metabolic pathways in KEGG \[[@CR131]\] and compared to the metabolic reconstruction obtained with Pathway Tools v22.0 \[[@CR132]\].Conversely, to evaluate possible gene loss the analysis was carried out in three different stages: (i) detection, (ii) manual curation and (iii) mapping of the metagenomics reads against a closely related *Leucobacter* spp. genome. In the first step, we applied loose criteria and determined which clusters were present in at least 90% (28 of 31 genomes) of *Leucobacter* spp. genes but absent from the genome of strain GP. The clusters were found by analyzing the pangenome with the parse_pangenome_matrix.pl script from the GET_HOMOLOGUES pipeline \[[@CR54]\]. Loose criteria instead of tight criteria (i.e. genes from the core genome) was chosen for this first stage because the vast majority of the sequences used in this analysis originated from draft genomes, which may contain significant gaps and annotation errors. In the second stage, clusters marked as missing were manually curated to exclude annotation errors in the genome of strain GP. This was done by aligning representative sequences of each cluster to the draft genome of strain GP with NCBI tBLASTn \[[@CR116]\]. Moreover, in the third stage, metagenomics reads obtained from Miseq sequencing were binned between *A. denitrificans* PR1 and a reference genome affiliated to the *Leucobacter* genus -- *L. chironomi* DSM 19883^T^ -- using BBSplit from the BBMap package v35.74 \[[@CR102]\]. Alignments of strain's GP reads against the reference genome of *L. chironomi* were inspected in the Integrative Genomics Viewer (IGV) v2.6.3 \[[@CR133]\] to further determine the presence or absence of high quality reads mapping against essential missing regions.
Whole genome comparisons and evolution of the SadABC complex {#Sec26}
------------------------------------------------------------
Whole genome comparison between strain GP and other sulfonamide degraders (see Additional file [2](#MOESM2){ref-type="media"} Table S3), was performed with the GET_HOMOLOGUES package as described above \[[@CR54]\]. The core and softcore genes shared between these strains were obtained by computing the intersection of clusters generated by COGtriangles and orthoMCL using the compare_clusters.pl script (−t 7; −t 6, respectively). Relevant gene clusters (i.e., *sad*A, *sad*B, *sad*C, *yce*I, and IS1380/IS3/IS4 transposases) were further used for homology searches with BLASTp against the NCBI non-reductant database \[[@CR116]\], structural modeling in SWISS-MODEL \[[@CR88]\] and conserved domain searches in NCBI database \[[@CR134]\]. The phylogeny and evolution of these proteins and their corresponding homologs were inferred from amino acid alignments with MUSCLE in MEGA6 \[[@CR87], [@CR122]\]. The phylogeny was estimated from combination of three methods: Maximum Likelihood (ML), Bayesian optimization and Neighbor Joining (NJ). For the ML method, the amino acid alignments were first evaluated with ProtTest v3.4.2 \[[@CR135]\] to find the best-fitting model of protein substitution. For SadA and SadB phylogeny the LG substitution model was used \[[@CR136]\] with gamma plus invariant sites heterogeneity model (G + I); for SadC and YceI the WAG model \[[@CR137]\] was used with G or G + I, respectively; and for the transposase the JTT model \[[@CR138]\] was used with the G heterogeneity model. The ML trees with bootstrap support values from 1000 replicates were constructed with MEGA6 \[[@CR87]\]. Bayesian optimization was calculated with BEAST v1.10.4 \[[@CR139]\]. Markov chain Monte Carlo (MCMC) were run using one million iterations and trees were sampled every 100 generations. The results of triplicate runs were combined with LogCombiner from the BEAST package \[[@CR139]\], and the combined output was analyzed with Tracer v1.7.1 to assess the overall quality of the estimation \[[@CR140]\]. Posterior probability support values and consensus tree was calculated from 10% of the total number of iterations (300,000). For the NJ method, the phylogenetic trees were constructed in MEGA6 using the JTT model \[[@CR138]\] with uniform rates and bootstrap support values were inferred from 1000 replicates. The ML trees were rooted at midpoint and visualized with FigTree v1.4.3 \[[@CR125]\], Bayesian posterior probability values, and ML and NJ bootstrap support values were included in the final tree. Functional comparison between strain GP and its helper strain, *A. denitrificans* PR1, was performed by submitting both genomes for annotation with RASTtk \[[@CR113]\]. Furthermore, the metabolic reconstruction and comparison between these distantly related strains was achieved with the function based comparison tool in the SEED viewer v2.0 \[[@CR141]\].
Supplementary information
=========================
{#Sec27}
**Additional file 1: Figure S1.** Fluorescence microscopy composite images of DAPI-stained cells of the microbial consortium (blue) and (A) cells hybridized with the modified ActORD1 FISH probe (stains strain GP, 5′ fluorophore: FAM, green, sequence: 5'- CACCAGGAATTCCAATCTCC-3', original probe accession number: pB-1931, reference: \[[@CR147]\]) or with (B) cells hybridized with Alca2 FISH probe (stains strain PR1; 5' fluorophore: Cy3, orange/red; sequence: 5'- CATCTTTCTTTCCGAACCGC-3'; probe accession number: pB-2127, reference: \[[@CR148]\]) **Figure S2:** Electron micrographs of negatively stained *Achromobacter denitrificans* PR1 showing the absence heatmap representation of peritrichous flagella(FG) **Figure S3:** Cladogram of the 16S rRNA gene inferred from maximum likelihood estimation with MEGA6 with the best-fitting model: TN93+G+I \[[@CR87]\]. *Leucobacter* spp. strains sequenced in this study are marked with an asterisk, and sulfonamide degraders are shown in bold. The tree was rooted at the outgroup and visualized with FigTree \[[@CR125]\]. The scale bar represents the number of expected substitutions per site. Bootstrap values were inferred from 1000 replicates, values above 70% are shown at the corresponding nodes **Figure S4:** Presence/absence heatmap representation and dendrograms of the 12,998 orthologs gene clusters found in the pangenome of Leucobacter spp. and strain GP obtained with the GET_HOMOLOGUES package \[[@CR54]\]. Each column represents a different gene cluster which can be absent (white) or present (blue) in each strain. As paralogs were included in the analysis, some clusters have more than one homolog per genome, and these are shown in darker blue **Figure S5:** Visualization of the reads of the strain GP's MAG on the Interactive Genomics Viewer (IGV) \[[@CR133]\] mapping to the reference genome and annotations of *Leucobacter chironomi* strain DSM 19883 ^T^ (assembly accession number GCA_000421845.1). This region from strain DSM 19883 (ATXU0100005.1:1..268438) contains the genes from the purine de novo biosynthetic pathway and the porphyrin and chlorophyll metabolism pathway (left to right): phosphoribosylformylglycinamidine synthase subunit PurQ (accession no. WP_017883592.1, locus tag H629_RS0106495); porphobilinogen synthase HemB (accession no. WP_024356487.1, locus tag H629_RS0106505); porphobilinogen deaminase HemC (accession no. WP_084705356.1, locus tag H629_RS14980); uroporphyrinogen decarboxylase HemE (accession no. WP_024356489.1, locus tag H629_RS0106525); glutamyl-tRNA reductase HemA (accession no WP_024356490.1, locus tag H629_RS0106530) **Figure S6:** Visualization of the reads of the strain GP's MAG on the IGV \[[@CR133]\] mapping to the reference genome and annotations of *Leucobacter chironomi* strain DSM 19883 ^T^ (assembly accession number GCA_000421845.1). This region from strain DSM 19883 (ATXU01000008.1:1..186096) contains the genes related to amino acid metabolism and from the glutathione and L-cysteine ABC transporter pathway (left to right): leucine\--tRNA ligase (accession no. WP_017793981.1, locus tag H629_RS0110150); alpha/beta hydrolase (accession no. WP_010837840.1, locus tag H629_RS011055); thiol reductant ABC exporter subunit CydC (accession no. WP_024357158.1, locus tag H629_RS0110165); thiol reductant ABC exporter subunit CydD (accession no. WP_024357159.1, locus tag H629_RS0110170) **Figure S7:** Heatmaps representing amino acid identity (BLASTp) of the SadABC (a, b and c) complex and YceI transporter (d) among isolates from the *Microbacterium* genus (strains BR1, C488, SDZm4 and CJ77), *Arthrobacter* genus (strains D2 and D4) and strain GP **Figure S8:** Amino acid alignment with MUSCLE \[[@CR122]\] of Acyl-CoA domains: Nterminal (a), middle (b) and C-terminal (c); between SadA and SadB homologs in *Microbacterium* sp. BR1, *Arthrobacter* sp. D2 and D4 and strain GP (SadB1: D3X82_00235; SadB2: D3X82_03160). Conserved regions within SadA and SadB and highlighted in green and conserved regions shared between all proteins are marked with an asterisk **Figure S9:** Close-up of the substrate-binding pocket of XiaF (PDB: 5LVW) bound to FADH and indole obtained by Kugel et al. \[[@CR90]\]. FADH is the co-factor, indole the substrate and S121 and I237 are the residues that are modified in SadA of Microbacterium sp. BR1 and strain GP. The ribbon (a) and electrostatic surface potential (b) diagrams have been prepared with PyMol \[[@CR149]\]. In b negative potential is shown in red and positive potential in blue. **Additional file 2: Table S1.** Mean coverage and GC content per strain and contig in the metagenome assembly of the consortium consisting of *Achromobacter* denitrificans PR1 and *Candidatus* Leubacter sulfamidivorax' **Table S2:** Results from CheckM evalution of the draft assembly of the '*Candidatus* Leucobacter sulfamidivorax' **Table S3:** List of all bacterial strain used for comparative genomics (^T^) type stain; (\*) sulfanomide degraders; N.A not available; (bold) strains sequenced in this study;(^1^) available on Github \[[@CR1]\]; \* the 16S rRNA gene sequence of this strain has a gap between positions 706 and 761;\*\* no rRNA was annotated in this sequence; cells highlighted in orange indicate strain for which the genome sequence became available after November 2018, and, therefore were not included in the comparative genomics studies to assess gene loss in strain GP **Table S4:** Complete and near complete (1 block missing = 1 ortholog gene missing) modules of the softcore genome of *Leucobacter* spp. and strain GP reconstructed in silico with KEGG Mapper \[[@CR56]\]
AAI
: Average Amino Acid Identity
ANI
: Average Nucleotide Identity
BHI
: Brain-Heart Infusion
Cryo-TEM
: Cryo-Transmission Electron Microscopy
DLB
: 25% Diluted Lysogeny Broth
FISH
: Fluorescent in situ hybridization
FMN
: Flavin Mononucleotide
MAG
: Metagenome-Assembled Genome
MCMC
: Markov chain Monte Carlo
ML
: Maximum Likelihood
MMSY
: Mineral medium with ammonium sulfate, succinate and yeast extract
MMSY-SMX-PE
: MMSY with sulfamethoxazole and 2-phenylethanol
MOB
: Mobility element (relaxase)
MPF
: Mating-pair formation
MTA
: 5′-methylthioadenosine
NJ
: Neighbor Joining
ONT
: Oxford Nanopore Technologies
POCP
: Percentage of Conserved Genes
R2A
: Reasoner's 2A Medium
RAxML
: Randomized Accelerated Maximum Likelihood
T4CP
: Type IV coupling protein
T4SS
: Type IV secretion system
TEM
: Transmission Electron Microscopy
TSA
: Tryptic Soy Broth
**Publisher's Note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
=========================
**Supplementary information** accompanies this paper at 10.1186/s12864-019-6206-z.
Ana Reis acknowledges the Fundação para a Ciência e a Tecnologia (FCT) for her Ph.D. scholarship (grant reference: SFRH/BD/95814/2013). We wish to thank Dr. David Waite (Australian Centre for Ecogenomics, Brisbane, Australia) for the helpful discussions regarding the assembly of genomes from metagenomics data. The authors acknowledge Professor Bernhard Schink for help with the etymology.
Availability of data and material {#FPar1}
=================================
All data is available under BioProject PRJNA489769 (types strains of the *Leucobacter* genus) and PRJNA490017 (MAG of strain GP).
ACR, OCN, BAK, and PFXC designed and planned the study. ACR conducted most of the laboratory work, performed the analysis of the data and wrote the initial draft of the manuscript. MC performed Cryo-TEM and TEM analysis and interpretation of the data. CE aided in the bioinformatics analysis of Next Generation Sequencing data. LG assisted with protein alignment and modeling and interpretation of data. All co-authors assisted with the editing and proofreading of the manuscript. All have read and approved of the manuscript.
This work was financially supported by (i) Swiss National Science Foundation (Grant No. 160332); (ii) project UID/EQU/00511/2019 by FCT/MCTES (PIDDAC); and (iii) project "LEPABE-2-ECO-INNOVATION" -- NORTE-01-0145-FEDER-000005 by NORTE 2020 through the European Regional Development Fund (ERDF).
Not applicable.
Not applicable.
The authors declare that they have no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .139}
{#sp2 .140}
{#sp3 .141}
{#sp4 .142}
{#sp5 .143}
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .353}
{#sp2 .354}
{#sp3 .355}
{#sp4 .356}
{#sp5 .357}
{#sp6 .358}
{#sp7 .359}
{#sp8 .360}
{#sp9 .361}
{#sp10 .362}
{#sp11 .363}
{#sp12 .364}
{#sp13 .365}
{#sp14 .366}
| {
"pile_set_name": "PubMed Central"
} |
Although tremendous strides have been made in total joint arthroplasty for the knee and hip, this modality is still evolving in the ankle.[@B1]-[@B3] Some of the problems that increase the difficulty in total ankle replacement (TAR) include a high prevalence of post-traumatic arthritis rather than chronic arthritis that causes poor anatomic access, a higher incidence of neuroarthropathy and vascular deficiency, particularly in diabetes-associated pathology and poor soft-tissue coverage. This is associated with an increased incidence of peri-operative complications, such as wound dehiscence and infection.
Mild to moderate ankle arthritis can often be managed with an ankle foot orthoses and a rocker-bottom shoe, provided that no significant varus/valgus deformity is present.[@B2],[@B4] Ankle arthrodesis and replacement are two common surgical treatment options for end-stage ankle osteoarthritis. The relative value of these alternative procedures is not well defined. The major drawbacks of arthrodesis are a 10-60% rate of arthrosis in the adjacent joints in the long-term,[@B5] a nonunion rate of 10 to 20%, and a postoperative infection rate of 3-25%.[@B6]
There are multiple TAR designs. However, most previous designs performed well only in the short-term and generally performed poorly in the long-term[@B7] due mostly to rapid bone loss or implant wear. The current designs maintain better pressure distribution and wear characteristics.[@B8],[@B9] The Scandinavian Total Ankle Replacement (STAR; W. Link GmbH, Hamburg, Germany) is an uncemented, unconstrained, congruent, cylindrical replacement. Its advantages over previous designs include its bone-sparing design, which can allow an easier reconstruction after failure, as well as its mobile bearing that allows easy polyethylene component exchange.
There is limited data comparing the outcomes or perioperative complications and reoperations in ankle arthroplasty designs.[@B8],[@B10],[@B11] Pyevich et al.[@B12] reported a 93% satisfaction rate at 3 to 10 year follow-up of Agility arthroplasty. Knecht et al.[@B13] reported \> 90% survival at 5 years in the same cohort, which are encouraging results for patients with modern ankle arthroplasties.
This study compared the clinical and radiographic outcomes and the early perioperative complications between a series of ankle arthroplasty and arthrodesis patients. A group of STAR patients was matched with a similar group of arthrodesis patients from the same period. It was hypothesized that total ankle arthroplasty would have similar early outcomes to ankle fusion in terms of its ability to relieve the disabling symptoms of ankle osteoarthritis.
METHODS
=======
This is a 2-6 year follow-up retrospective study of a cohort of ankle arthroplasty patients with noninflammatory arthritis. Patients with isolated ankle osteoarthritis were included but those patients with any confounding pathology were excluded. Only one type of prosthesis, the STAR device, was examined. However, many different techniques of ankle fusion were included. The ankle joint was fixed only with 6.5 or 7.3 mm cannulated screws in 14 cases, a plate and screws in 10 cases and external fixators in 3 cases.
Between January 2, 1998 and May 31, 2002, 138 patients were treated with ankle fusion or replacement at our institution. Among them, 71 had isolated posttraumatic or primary ankle arthritis. The remaining 67 were excluded. Of these patients, 12 and 55 patients had a different ankle prosthesis (Agility) and ankle arthritis with confounding pathology, respectively. The reasons for their exclusion was inflammatory arthritis (n = 20), neuropathic arthritis (n = 4), concomitant hindfoot fusions (n = 25) and revision procedures (n = 6).
Among the 71 eligible subjects, one group of 42 patients underwent STAR total ankle arthroplasty while the other group comprised of 29 patients who underwent ankle fusion. A complete follow-up could be per formed on 88% (37/42) of the TAR group and 79% (23/29) of the ankle fusion group. [Table 1](#T1){ref-type="table"} lists the subject characteristics. The fusion group has a younger age at surgery (*p* = 0.034), higher proportion of males and a higher percentage of those with posttraumatic osteoarthritis (OA). There were no significant differences in the mean BMI (*p* = 0.258) or follow up length (*p* = 0.874) between the two groups.
Detailed physical examination data, including information on perioperative and postoperative complications, was collected. The foot function was evaluated using the Ankle Osteoarthritis Scale (AOS) ([Table 2](#T2){ref-type="table"}). This is a reliable, validated, visual-analog based, disease-specific self-administered instrument, which is based on the foot function index (FFI). It is designed specifically to measure the disability and pain from ankle osteoarthritis.[@B14] It has two subscales, pain and disability. The index study revealed an effect of gender, body mass index and arthritis on the other joints. Each item of the AOS was graded, and the subscales for pain and disability were generated independently. The foot function index is a validated reliable instrument that was developed in rheumatoid arthritis patients for measuring the level of foot pain, disability and activity restriction.[@B15],[@B16] In contrast, the ankle osteoarthritis scale was developed based on the FFI, but was validated in patients with osteoarthritis.
The patient\'s general health was evaluated using the short form health survey (SF-36), which is a generic measure of mental and physical health and a well established quality of life out come measure. The SF-36 evaluates the impact of a disease on the patient, and is well supported by evidence of reliability and validity, and responsiveness in musculoskeletal disorders.[@B17] Separate mental component summary (MCS) and physical component summary (PCS) scores were generated.
For the pre-operative films, an ankle anteroposterior (AP), mortise and lateral view were taken. On the day of the study, foot lateral views were taken in the maximum dorsi- and plantar-flexion, foot AP, hindfoot alignment, and Brodens views. The pre and post-operative arthritic changes were evaluated for 4 joints: subtalar, talonavicular, calcaneocuboid, and midfoot (Lisfranc + naviculocuneiform joint). The Kellgren/Lawrence (K/L) arthritis scale was used, which is a 5-point grading of the radiographic signs of arthritis.[@B18] All X-rays were read and scored by one of the authors, limiting the interobserver variance in the results. The same procedure was carried out 2 weeks later to reduce the intraobserver variance. A second opinion was given by another physician in the rare case of different scores.
Statistical Analysis
--------------------
All analyses were performed using SAS (SAS Ins., Cary, NC, USA). The AOS and SF-36 scores were analyzed using a student\'s t-test. The AOS and SF-36 scores were also examined using a 2-tailed t-test at the 0.05 significance level, which had 0.80 power to detect a difference in a mean score of at least 0.76 SD units between the two groups. Using a standard deviation of 10 and 26 for SF-36 and AOS scores, respectively, a difference in the mean SF-36 and AOS score between the two groups of at least 7.6 and 20, respectively, could be detected with 0.80 power.
The X-ray data was arranged in a table showing the frequency of each K/L grade for each joint. There was a separate table for the pre and postoperative score. A Wilcoxon signed-rank test was used to compare the K/L grade ranked data. Power analysis revealed that a difference in the median K/L grade of at least 1 between the groups could be detected with 0.90 power at the 0.05 significance level.
RESULTS
=======
Clinical Outcomes
-----------------
[Table 3](#T3){ref-type="table"} shows the mean AOS pain and disability scores, as well as the mean SF-36 PCS and MCS scores. A perfect AOS score is zero, showing no disability and no pain. The population mean SF-36 score was 50 with higher scores representing better function. The outcomes between the TAR and fusion groups were compared using a student\'s *t*-test. For each of the four outcomes, a better outcome was observed in the TAR group. However, using a critical *p*-value of 0.05, significant differences were noted only in the SF-36 MCS and AOS pain scale.
Radiographic Outcomes
---------------------
The K/L grade was recorded at the subtalar, talonavicular, calcaneocuboid and midfoot for both the pre and postoperative X-rays. The radiographic data were evaluated by asking three questions: Is there a significant difference in the pre-operative K/L grade, post operative K/L grade, or the change in the K/L grades between the TAR and fusion groups?
[Table 4](#T4){ref-type="table"} summarizes the radiographic data. For each joint, the frequency distribution of each K/L grade (0 = Normal, 1 = Minimal, 2 = Mild, 3 = Moderate, 4 = Severe) on the preoperative and postoperative X-rays is shown.
Firstly, the hypothesis that there was a difference in the preoperative K/L grade between the fusion and TAR groups was tested. For each joint, the frequency distribution of each K/L grade was determined ([Table 4](#T4){ref-type="table"}). The distribution between the TAR and fusion groups was then compared. A Wilcoxon rank test was used, and all *p*-values are shown in [Table 4](#T4){ref-type="table"}. There was no significant difference in the preoperative OA between the TAR and fusion groups at any joint evaluated.
Secondly, the hypothesis that there was a difference in the postoperative K/L grade between the two groups was tested. Another frequency distribution was developed for each joint. The Wilcoxon rank-sum test was used. All *p*-values are shown in [Table 4](#T4){ref-type="table"}. There was no significant difference in postoperative OA between the TAR and fusion groups.
Thirdly, the hypothesis that there was a difference in the change in K/L grade between fusion and TAR groups was tested. Using a similar method, there was no difference in the change in OA between the two groups.
Perioperative Problems
----------------------
In the 37 TAR patients, fifteen additional procedures were performed after the index operation during the follow-up period. These included the following: 7 cases of debridement plus bony resection for lateral or posterior impingement; 2 cases of bone debridement for exostosis; 3 cases of osteolysis requiring poly exchange and bone grafting including one revision of tibial component with calcaneal realignment; 2 cases of revision closure for wound dehiscence; and one case of medial malleolar osteotomy revision fixation for nonunion. There were 5 intraoperative medial or posterior malleolar fractures; all were fixed internally at the index operation. Two patients developed deep venous thromboses requiring anticoagulation and one superficial wound dehiscence that healed uneventfully.
In the 23 patients with ankle fusion, there were 5 postoperative events that required surgery: 2 cases of nonunion that required revision arthrodesis; 2 cases of hardware pain that required hardware removal; and one case of naviculocuneiform joint arthritis progression that required fusion. There was one case each of leg length discrepancy, delayed union, tibia stress fracture, hardware pain and wound dehiscence, and impingement. All were treated successfully without surgery.
DISCUSSION
==========
This study demonstrated a tendency toward better clinical outcomes among the ankle arthroplasty group at the 2-6 year follow up. There was better ultimate pain relief from ankle arthroplasty than fusion. This finding is very important because the primary indication for both procedures is pain relief. The cause of continued pain for both groups is unclear because the rates of new-onset, radiographically apparent arthritis was similar in the two groups. The AOS disability scale, though not significant, also suggests a better outcome among the TAR group.
The SF-36 MCS scores for the arthroplasty patients are higher, suggesting a possible difference in life outlook with the fusion patients having marginally poorer mental health despite their younger age. This might be related to the finding that a higher percentage of the fusion group had posttraumatic rather than degenerative arthritis. The long-term effect of ankle trauma on the patient\'s function and outlook was discussed previously by Marsh et al.[@B19] and Dirschl et al.[@B20] The PCS scores also showed a trend toward a better outcome among TAR subjects but the significance was not definite.
There was a higher incidence of surgical procedures performed postoperatively in the arthroplasty group. The significance of these findings is unclear. All the procedures were small, and none involved any revision of the hardware or a conversion to fusion. However, the risk from these procedures is an important factor when comparing arthroplasty with arthrodesis.
This is a retrospective study and must be interpreted in terms of patient selection bias. It is impossible to control all the known sources of bias in such a study. One of the main concerns is that the indications for ankle fusion are different from TAR. Generally, fusion patients tend to be younger and more active with a diagnosis of posttraumatic osteoarthritis. Since obesity or severe malalignment is a relative contraindication to arthroplasty, fusion patients may be heavier and more malaligned. However, there was no significant difference in BMI in this study. Therefore, the body habitus is not believed to have affected the results. Basically, a direct comparison of both group demographics and outcomes may not be appropriate. Despite these limitations, useful comparative data was obtained on the short term results of fusion and TAR.
Another major assumption is that the AOS is a real reflection of the patient\'s condition after either fusion or TAR. This is a reasonable assumption because Domsic and Saltzman[@B14] reported it to be \"a reliable and valid instrument that specifically measures the patient\'s symptoms and disability related to ankle arthritis.\" They validated the AOS scale in patients with ankle OA.
These results are encouraging and consistent with other data published regarding this device. Hintermann[@B21] followed 48 patients who had undergone arthroplasty for 1 to 4 years. They noted 91% satisfaction using the American Orthopaedic Foot and Ankle Society-Hindfoot-Score, no migration and a good range of motion. Seven cases required revision surgery: fibula resection for lateral impingement (3 cases), posteromedial soft tissue revision for a painful restriction of dorsiflexion (2 cases), percutaneous lengthening of the Achilles tendon (1 case), and osteotomy and distraction for angular correction after a stress fracture of the distal tibia (1 case).
Implant survival was not evaluated in this early postoperative period but there are studies that showed good wear characteristics for this device. Wood et al.[@B9],[@B22] reported that at 5 years, the STAR prosthesis had superior clinical and radiographic outcomes compared to a cemented comparable implant. Andersen et al.[@B23] reviewed 51 STAR implants for 36 to 97 months, and reported a 70% 5-year survival rate. This study focused on the early clinical and radiographic outcomes, and showed a higher incidence of postoperative events that required surgery than arthrodesis. However, there was no conversion to fusion during this time period.
The clinical results of ankle arthroplasty and arthrodesis after an average 4 year follow-up were almost equivalent. The arthroplasty group had more postoperative complications that required surgery than the fusion group. On the other hand, the arthroplasty group showed better pain relief and preservation of motion.
######
Subject Characteristics of the Total Ankle Replacement and Fusion Groups
TAR: Total ankle replacement, OA: Osteoarthritis, BMI: Body mass index.
######
Ankle Osteoarthritis Scale
######
The Average AOS Pain and Disability Scores and SF-36 PCS and MCS Function Score in the TAR and Fusion Groups
AOS: Ankle osteoarthritis scale, SF-36: Short form health survey, PCS: Physical component summary score, MCS: Mental component summary score, TAR: Total ankle replacement.
######
Distribution of the Preoperative and Postoperative K/L Grade in the Subtalar, Talonavicular, Calcaneocuboid, Midfoot Joints in the TAR and Arthrodesis Groups
Values are presented as number (%).
K/L: Kellgren/Lawrence, TAR: Total ankle replacement.
| {
"pile_set_name": "PubMed Central"
} |
INTRODUCTION {#s1}
============
The p53 tumour suppressor protein is a transcription factor that activates genes responsible for acute radiation-induced death, cell cycle checkpoint function, induction of apoptosis and senescence and tumour suppression. Its inactivation is considered a key event in human carcinogenesis. As a "guardian of the genome" \[[@R1]\], in normal tissues p53 protein is rapidly activated by different cellular stress pathways, such as DNA damage, hypoxia, spindle damage or oncogenic stimuli. p53 is stabilized and post-translationally modified to serve its transcriptional activation function. The target genes of p53 include Mdm2, which targets p53 for proteasomal degradation, enabling an important negative feedback loop that restores the low basal levels of p53 after activation \[[@R2]-[@R6]\].
The p53 protein is essential for the regulation of cell proliferation, and mutant p53 over-expression is usually seen in malignant tumours. Here the missense mutant protein accumulates to high levels and extensive immunohistochemical studies have shown that this accumulation is highly variable and that mutation alone is not sufficient to drive accumulation in all cells in the tumour leading to complex and variable staining patterns \[[@R7]\]. Mutation of the p53 gene results in loss of its tumour suppressive transcriptional activation properties and gain of novel oncogenic functions that are dependent on high level expression \[[@R8]\]. Germline mutations in p53 cause Li-Fraumeni Syndrome, which is characterised by the early onset of cancer in a wide variety of possible tissue types \[[@R9],[@R10]\]. Genetic and xenograft data from mouse models have shown that the restoration of wild type p53 activity is a potential anticancer strategy \[[@R11]-[@R16]\].
Small intestinal epithelium has a remarkable rate of self-renewal which provides a daily readout of proliferative activity \[[@R17]\]. The epithelium of the murine small intestine renews every 5 days \[[@R18],[@R19]\]. Each small intestinal crypt contains approximately six long-lived stem cells (crypt base stem cell, CBC) and these cells divide every day \[[@R20]\]. Their daughter cells consist of the transit-amplifying (TA) crypt compartment and these cells divide every 12-16 hours. They perform up to five rounds of cell division while migrating upwards \[[@R21]\]. The Paneth cells located between CBC cells reside at the crypt base for 3-6 weeks \[[@R22]\]. When TA cells reach the crypt-villus junction, they rapidly differentiate as villous epithelium. We have made use of the murine small intestine system to study mutant p53 expression in order to understand p53 expression in various cell populations of morphologically normal tissues and its association with potential function in initiation of preneoplasia/neoplasia.
We generated mice carrying a missense p53 R172H mutation which corresponds to the p53 R175H hot spot mutation in human tumours, and is associated with Li-Fraumeni Syndrome \[[@R23], [@R24]\]. This particular hot spot mutation in the p53 DNA-binding domain results in a protein that is transcriptionally inactive and has both dominant negative and gain-of-function phenotypes \[[@R8]\]. It is generally considered that p53 is mainly regulated at the protein rather than the transcriptional level, and mutant p53 protein is stable in human cancers but unstable in normal tissues \[[@R25]\]. In previous studies in mouse \[[@R25]\] and zebrafish \[[@R26]\] models where the animal\'s wild type p53 genes had been eliminated and replaced with mutant p53. It was reported that the mutant p53 proteins did not accumulate in normal tissues but reached very high levels of expression in the tumours that arose in these animals. In both the mouse and Zebrafish models the inactivation of the Mdm2 gene in a homozygous mutant p53 background results in the high level expression of the mutant p53 in many normal tissues \[[@R25], [@R26]\]. In this study, we show that mutant p53 protein is detectably expressed in a specific cell population of morphologically normal small intestine and also in other tissues in both absence and presence of p53-activating signals and even in the presence of active Mdm2. This level of expression is much higher than that seen for the wild type protein in unstressed tissues. We show using the Mdm2 inhibitor Nutlin *in vivo* that the mutant p53 levels are further increased by pharmacological inhibition of Mdm2 and are also induced by DNA damage. Importantly this allowed us to explore the possibility that down-regulation of wild type and mutant p53 protein in differentiated small intestinal epithelium occurs at the transcriptional level and to detect gene dosage effects on protein expression.
RESULTS {#s2}
=======
Heterogeneous expression of p53 R172H protein in morphologically normal adult mouse tissues {#s2_1}
-------------------------------------------------------------------------------------------
R172H mutant p53 protein (mutp53) levels may be elevated in preneoplastic cells, therefore we examined morphologically normal tissues in the *p53^R172H/R172H^* mice to study the expression of mutp53. We found that in the majority of apparently normal adult mouse tissues, there is a heterogeneous expression of mutp53 and we were able to divide mutp53 accumulation in mouse organs into four groups according to staining intensity and positive cell fraction in the population (Figure [1](#F1){ref-type="fig"}). Mutp53 4+: small intestine, colon, rectum and thymus; mutp53 3+: bone marrow of vertebrae and femur, spleen, growing skin and hair follicle, mutp53 2+: kidney, nonglandular & glandular stomach and ependyma of brain; mutp53 1+: testis, pancreas & islet of Langerhan, lung and cornea. There is no detectable immunostaining of mutp53 in liver, brain (except ependyma) and skeletal muscle.
{#F1}
In *p53^R172H/R172H^* mice, mutp53 accumulation was confined to the crypts of the small intestine. In colon and rectum, mutp53+ was accumulated in lower 2/3 of crypts. Mutp53 accumulation in thymus was detected both in cortex and medulla and was more pronounced in the medullary compartment. In spleen, mutp53+ cells distribute both in red pulp and white pulp, more mutp53+ cell populations are located in the red pulp. In bone marrow of vertebrae and femur, scattered mutp53 immuno-positive cell populations were found amongst the hematopoietic cells. Mutp53 accumulation is observed in growing skin and anagen hair follicles. In kidney, mutp53 expression is only found in the proximal convoluted tubules located in renal cortex while mutp53 was undetectable in glomeruli and medulla. Mutp53 accumulated in basal layers of nonglandular stomach and scattered expression is seen in the upper part of corpus region of stomach. Mutp53 is only expressed in the spermatogonium of testis, and the ependyma of brain. Mutp53 expression is not prominent in pancreas and lung. Mutp53 is immunonegative in liver, brain (except ependyma) and skeletal muscle. There was very weak or non-identifiable p53 staining in p53 wild type mouse tissues, and as expected there was no p53 immunopositive staining found in p53 knockout mice. Intriguingly, in *p53^R172H/KO^* mice, which bear a single mutant p53 allele, p53 R172H protein was found at low levels in all the tissues in which we detected mutp53 staining in the *p53^R172H/R172H^* mice. Therefore, p53 protein levels in tissue of p53 R172H mice are dependent on gene dosage ([Suppl Figure 1](#SD1){ref-type="supplementary-material"}).
p53 R172H protein accumulates in intestinal crypts in a manner dependent on cell type and gene dosage {#s2_2}
-----------------------------------------------------------------------------------------------------
The pattern of p53 immunopositivity in the small intestine (Figure [2A](#F2){ref-type="fig"}, [3A](#F3){ref-type="fig"}, [Suppl Figure 2](#SD1){ref-type="supplementary-material"} and Cover page) was particularly interesting. No p53 protein was detected in p53-null (*p53^−/−^*) and very weak or low levels of p53 protein in p53 wild type (*p53^KO/+^ and p53^+/+^*) mouse small intestines. In *p53^R172H/+^* and *p53^R172H/KO^* mice, which bear only a single mutant p53 allele, p53 R172H protein was found at lower levels in a majority of the crypts (still higher than p53 wild type mice) with a few occasional strong immunopositive foci in crypts of *p53^R172H/KO^ mice*. In *p53^R172H/R172H^* mice, p53 R172H protein levels were elevated in all the crypts of the small intestine, the specificity of strong p53 staining by IHC in the small intestine of *p53^R172H/R172H^* mice was also verified by our home-made p53 polyclonal rabbit antibody which reacts strongly with p53 and has no background staining in p53-null mice using immunofluorescent staining ([Suppl. Figure 2](#SD1){ref-type="supplementary-material"}). Therefore, p53 protein levels in the crypts of the small intestine are dependent on genotype and gene dosage, which was further validated by Western blot when we extracted protein from duodenum and jejunum of small intestine in *p53 wild type (p53^+/+^), p53^KO/+^, p53^R172H/KO^, and p53^R172H/R172H^* mice to compare total mutp53 protein expression levels between one allele or two alleles of the p53 R172H mutant gene (Figure [2B](#F2){ref-type="fig"}). We could detect elevated and dosage dependent mutp53 protein expression in *p53^R172H/KO^ and p53^R172H/R172H^* mouse small intestine, while only weak bands were detected in *p53 wild type* mice.
{#F2}
{#F3}
We could detect increased p53 R172H protein in the intestinal crypts of *p53^R172H^* pups as early as one week after birth. The effect of gene dosage on p53 protein levels was very clear in pups at just two weeks of age (Figure [2C](#F2){ref-type="fig"}). In *p53^R172H/R172H^* pups, p53 R172H protein levels were high in every crypt and remained low but detectable in their p53R172/KO littermates. Strong mutp53 staining is also observed in pseudostratified epithelium of gut tube at embryonic day 13.5 (E13.5) and in nascent crypts after villus emergence happens at E16 in *p53^R172H/R172H^* embryos (Figure [2D](#F2){ref-type="fig"}). Taken together, these data indicate that the accumulation of mutant p53 protein in intestinal crypts is detectable throughout embryo development, and soon after birth as well as in adult mice and confirm that it is dependent on cell type and gene dosage at all ages examined.
Expression of p53 R172H protein confined to cycling and proliferating cells of morphological normal adult small intestine {#s2_3}
-------------------------------------------------------------------------------------------------------------------------
Mupt53 accumulation in crypts of the small intestine is of particular interest given the well characterised location of stem and differentiated cell populations in this dynamic tissue. This epithelium is the most rapidly self-renewing tissue in adult mammals, and can be clearly divided into three kinds of epithelium: the cycling crypt base columnar (CBC) stem cells adjacent to differentiated Paneth cells, proliferating transit-amplifying (TA) cells, and differentiated villous epithelium \[[@R20]\]. Our mutp53 *in situ* immunostaining indicated that mutp53 accumulated in the cycling CBC and rapid proliferating TA cells (Figure [3A](#F3){ref-type="fig"}, orange and brown arrows) but not in the Paneth cells and differentiated villous cells of morphological normal small intestine (Figure [3A](#F3){ref-type="fig"}, Cyanic and blue arrows). Stem cell marker Lgr5 is exclusively expressed in cycling columnar cells at the crypt base which is accepted as a stem cell marker of small intestine \[[@R20]\]. We found no change of Lgr5 mRNA expression among *p53^R172H/R172H^*, *p53^+/+^* and *p53^KO/KO^* mouse small intestine, examined by qRT-PCR and by FISH procedures. Lgr5 mRNA Expression was confined to the CBC population in all these three genotypes ([Suppl Figure 3A, 3B](#SD1){ref-type="supplementary-material"}).
To further validate mutp53 protein accumulation in stem cells of the small intestine, we isolated intestinal crypts from mice of the various *p53* genotypes and cultured them *in vitro* as organoids. We could not detect p53 by immunostaining in organoids from mice that were p53-null (data not shown). However, the organoids from *p53^R172H/R172H^* mice stained strongly for p53 while those from *p53^R172H/KO^* mice showed weak p53 immunopositivity (Figure [3B](#F3){ref-type="fig"}). Therefore, the organoids showed the same trend of mutant p53 accumulation as that seen in the mice.
Next we further explored whether p53 R172H affects cell proliferation in morphological normal tissues. We co-stained mutp53 and PCNA by immunostaining, more PCNA positive cells were found in crypts of p53 R172H mouse than in p53 wild type mice, although the intensity of PCNA staining in both p53 wild type and p53 R172H mutant is similar. The mutp53 positive staining cell population was always found within the PCNA expressing cell population (Figure [3C](#F3){ref-type="fig"}).
p53 R172H protein accumulates rapidly after induction by ionising radiation but clears more slowly than wild type p53 {#s2_4}
---------------------------------------------------------------------------------------------------------------------
Given that p53 is often stabilised upon activation \[[@R27],[@R28]\], studying p53 accumulation *in vivo* could provide information on p53 induction upon drug treatment that would be relevant for the optimisation of chemotherapeutic regimens. To this end, studying the intestine is particularly important because 20-40% of all adverse effects arising from drug treatment are attributed to events in the small and large intestine \[[@R29]\]. For example, p53 plays a major role in the apoptosis of stem cells within the intestinal crypt \[[@R30],[@R31]\], which can contribute to gastrointestinal toxicity \[[@R32]\]. The crypts of the small intestine also contain proliferating cells that are sensitive to ionizing radiation \[[@R31],[@R33]\].
Given that we observed distinct mutp53 expression in the small intestine that was dependent on gene dosage, we sought to use the small intestine of these p53 mutant mice as an experimental system to examine p53 induction. We first subjected the mice to whole body ionising radiation and performed a time course experiment to compare the response of wild type p53 protein in *p53^KO/+^* mice with that of mutant p53 in their *p53^R172H/KO^* littermates. Just 3 hours after 2 Gy ionising radiation, we could detect elevated levels of both wild type and mutant p53 protein levels in the same crypt cells in which high p53 R172H protein levels had been observed in *p53^R172H/R172H^* mice (Figure [4](#F4){ref-type="fig"}). Irradiation induces apoptosis and the accompanying loss of cellularity can complicate analysis of immunohistochemical data. Nonetheless, it is still clear that basal protein levels of wild type p53 were restored 8 hours after irradiation (Figure [4](#F4){ref-type="fig"}, left panel). In contrast, p53 R172H protein levels in most crypts took up to 25 hours to return to basal levels (Figure [4](#F4){ref-type="fig"}, right panel). Therefore, ionising radiation rapidly induced both wild type and mutant p53 protein in intestinal crypts, but wild type p53 protein returned to basal levels more quickly than did mutant p53.
{#F4}
Nutlin induces the rapid accumulation of p53 R172H protein in intestinal crypts {#s2_5}
-------------------------------------------------------------------------------
We then tested whether we could similarly detect p53 induction after drug treatment. We used the small molecule nutlin, which blocks the interaction between p53 and its major negative regulator Mdm2, thereby preventing p53 degradation and inducing its activity \[[@R16]\]. As in the experiment with ionising radiation, we compared p53 protein levels in the small intestines of *p53^KO/+^* mice with those in their *p53^R172H/KO^* littermates at various time-points after drug administration. We could detect slight accumulation of p53 wild type protein 2 hours onward after nutlin treatment (Figure [5](#F5){ref-type="fig"}, left panel) whereas there was a clear increase in mutant p53 protein levels detectable 2 hours after nutlin administration and this was sustained at the 7.5 hour time-point (Figure [5](#F5){ref-type="fig"}, right panel). These data affirm the utility of the p53 mutant mice as a sensitive system to study p53 accumulation *in vivo*, although the p53 wild type mice are still valuable for studies of specific p53 target gene activation. They also confirm that the degradation of mutant p53 is due to the action of Mdm2 since it is inhibited by nutlin.
{ref-type="fig"} and in Figure [2](#F2){ref-type="fig"}. Scale bar: 50 μm.](oncotarget-06-17968-g005){#F5}
Loss of Mdm2 results in even higher levels of p53 expression but p53 is still restricted to proliferating cells in the small intestine {#s2_6}
--------------------------------------------------------------------------------------------------------------------------------------
In prior work on both the *p53^R172H/R172H^* mice and *p53^R172H/R172H^* zebrafish we were not able to detect mutant p53 accumulation in normal tissues though it was readily seen when these animals were crossed onto Mdm2-null backgrounds \[[@R25], [@R26]\]. We at first considered that this readily detection of mutant p53 accumulation in mice with Mdm2 could be due to the effect of other background genes of the p53 mutant mice but comparison of the staining methods used encouraged us to re-examine tissues from these mice using the sensitive antibodies and processing methods developed in the current studies. These showed that the differences in the current results are indeed due to the increased sensitivity of the current protocols. As shown in Figure [6A](#F6){ref-type="fig"} staining for p53 is indeed detectable in the crypt cells of the mice previously described as negative for p53. The striking increase in intensity of p53 staining in the Mdm2-null background (Figure [6B](#F6){ref-type="fig"}) does however completely confirm the earlier conclusions of the importance of Mdm2 in regulating mutant p53 levels *in vivo*. In support however of the additional regulation of p53 by transcriptional control we note that even in these intensely stained specimens p53 is confined to the crypt cells and is not expressed in the differentiated cells of the villi.
{#F6}
DISCUSSION {#s3}
==========
Previous reports have shown that the levels of mutant p53 protein are often increased in tumours but not in normal tissue, both in mice \[[@R25]\] and in zebrafish \[[@R34]\]. Our study is the first to describe elevated mutant p53 protein levels in non-cancerous mouse tissues, which may not have previously been detected due to differences in immunostaining protocols and in the p53 antibody used \[[@R24], [@R35]\]. Mutp53 protein is regulated similarly to wild type p53 in our data (Figure [4](#F4){ref-type="fig"}, [5](#F5){ref-type="fig"}) and others \[[@R25], [@R36]\]. It is induced by stress and genotoxic damage such as irradiation and degraded after ubiquitination by its major negative regulator, the ubiquitin ligase Mdm2 \[[@R25], [@R35]\]. Previous data indicated that the transcriptionally inactive mutant p53 cannot induce Mdm2 expression, which could partly explain the longer half-life of the protein in cells after induction \[[@R35]\]. There is no expression of p21 protein in p53 R172H mice as expected (Figure [7](#F7){ref-type="fig"}), indicating p53 R172H mutant mouse model we generated also lost its p53 transcriptional capacity which is consistent with previous reports \[[@R35]\].
{#F7}
Many different types of cancer show a high incidence of p53 mutations, leading to the expression of mutant p53 proteins. Although we are unable to validate whether the properties and induction of the mutp53 protein in these morphologically normal tissues are the same as that of the mutp53 protein in tumours, we did find that increased mutp53 protein accumulation is more relevant to cell and tissue types with high proliferative rate such as small intestine, thymus, spleen and anagen hair follicles. The accumulation of mutp53 protein is not spontaneous or uncontrolled, it occurs in a cell- and tissue-specific manner (Figure [1](#F1){ref-type="fig"}). The high accumulation of the mutp53 protein within the crypts is remarkable. We stained large numbers of sections from small intestine of different p53 R172H mice, we generally only detected elevated mutp53 protein in cycling CBC and rapidly proliferating TA cells within the small intestinal crypts (Figures [3](#F3){ref-type="fig"}). Our study is the first to describe elevated mutp53 protein levels in CBC and TA cells of intestinal crypts. Although we do not understand whether and how elevated mutant p53 proteins might perturb intestinal stem cells and affect tumorigenesis, this p53 R172H mouse model will provide a valuable tool for understanding the role of mutant p53 protein in the cancer initiation. In human specimens clusters of intensely stained p53 positive nuclei have been seen in clones present in skin \[[@R37]\] and fallopian tube (p53 signatures) \[[@R38]\]. Sequencing of these signatures has shown that they contain mutant p53. While high level expression of mutant p53 is associated with pre-neoplastic conditions and mutant p53 can act as a dominant transforming oncogene our studies suggest that this is not a very penetrant state. Mice expressing high levels of mutant p53 in the stem and proliferating cells of their intestine from birth do not show overt neoplastic effects in this tissue.
The mutp53 protein was not detected in differentiated Paneth cells and villous epithelium of small intestine as well as in other differentiated cells. Differentiated Sertoli cells and spermatocytes were also immuno-negative for mutp53, while spermatogonia which are undifferentiated male germ cells express elevated mutp53 protein. In stratified epithelia, such as nonglandular stomach epithelium and skin, mutp53 was detected in basal undifferentiated layers of these epithelia but not in differentiated suprabasal layers (Figure [1](#F1){ref-type="fig"}). This observation is consistent with previously published data in mouse embryonic stem cells that p53 protein levels and activity decrease upon differentiation \[[@R39], [@R40]\] and that it is more difficult to induce p53 activity in differentiated cells \[[@R40]\]. This phenomenon was mostly considered due to the post-translational regulation, which leads to down-regulation of p53 protein in differentiated cells. However in this study, whether treated by nutlin or ionising irradiated, mutp53 protein cannot be induced in differentiated villi and is always confined to the crypts of the small intestine. This is even true when the Mdm2 gene is inactivated suggesting strongly that the lack of p53 expression may be due to inhibition of the p53 promoter in differentiated cells
The anatomy of the small intestinal crypt and villus are uniquely suited to study the properties of various cell populations such as proliferating cells in crypts and differentiated cells in villi of small intestine. It still remains to be determined how p53 is down-regulated after cell differentiation. Abnormal development and incomplete differentiation are hallmarks of cancer, loss of differentiation may be linked to p53 disruption in tumourigenesis. Enforced differentiation in p53 mutant cancer may become a powerful therapeutic tool in cancer chemotherapy and drug discovery. This is because strong evidence suggests that the high level expression on mutant p53 drives the growth of the cancer and blocking its expression inhibits tumour growth \[[@R41]\]. Our results may also go a long way towards explaining the widely discussed variability of p53 expression in human tumours \[[@R42]\]. Clearly as shown here the dosage of mutant p53, the activity of Mdm2 and the activity of the p53 promoter, which is in turn linked to the state of proliferation, can along with the sensitivity of the staining protocol used have profound effects on the intensity of p53 staining observed.
MATERIALS AND METHODS {#s4}
=====================
Mice {#s4_1}
----
All mouse experiments were approved by the A\*STAR Institutional Animal Care and Use Committee (IACUC) and performed in compliance with IACUC regulations. The *Trp53* knockout mice and *p53^R172H^* mutant mice were generated by crossing mice expressing Cre under the control of the β*-actin* promoter \[[@R43]\] with *Trp53* conditional knockout mice \[[@R44]\] and *p53^LSL^*•*^R172H^* conditional mutant mice \[[@R35], [@R45]\] respectively. The generation of the *p53R^172H/R172H^*; *Mdm2^−/−^* mice was reported previously \[[@R25]\]. *p53^R172H^* mice maintained on a mixed 129S~4~/SvJae x C57BL/6 background were used for p53 IHC staining. Genotyping was performed by PCR analysis of DNA from ear clips obtained at the time of weaning. Two-month-old (mo) mice of the appropriate genotypes were subject to total body irradiation using a single dose of 2 Gy and 8 Gy. 2 mo mice with same genotypes were treated with nutlin-3 (Nutlin, 200 mg/kg, Oral gavage). The mice were then sacrificed at the specified times. To prevent infection, mice sacrificed more than 24 hours after irradiation received 0.5 mg/ml amoxicillin, starting 3 days prior to irradiation. Mouse tissues for immunohistochemistry (IHC) and immunofluorescent staining were fixed in 10% Neutral Buffered Formalin and embedded in paraffin. Mouse tissues for fluorescence *in situ* hybridization (FISH) were fixed in cold 4% paraformaldehyde in phosphate buffered saline (PBS) and submerged overnight at 4°C in 30% sucrose/4% paraformaldehyde, embedded in OCT and stored in −80C. All tissues analysed appeared normal at the time of necropsy.
Immunohistochemistry and microscopy imaging {#s4_2}
-------------------------------------------
Immunostaining was performed on formalin-fixed paraffin-embedded (FFPE) 5 μm sections. We used commercial rabbit anti-p53 (CM5, 1:500, Leica Biosystems, Germany) primary antibody and the p53 IHC procedure was performed by auto-staining machine (Leica Bond-max, Leica Biosystems, Germany) to ensure reproducibility of staining between experiments, the procedure was performed according to the manufacturer\'s instructions. The most critical step for p53 IHC staining is that we used EDTA based pH 9.0 antigen retrieval solution to exposure antigen epitopes. Home-made polyclonal rabbit anti-p53, commercial monoclonal mouse anti-p53 (1C12, 1:200, \#2425S, Cell Signaling Technology, Danvers, USA), rabbit anti-PCNA (1:100, sc-7907, Santa Cruz), and mouse anti-p21 (1:20, F5, Santa Cruz) primary antibodies and anti-rabbit/mouse Alexa Fluor 568/488 (1:500, Invitrogen, California, USA) secondary antibodies were used for immunofluorescent staining. IHC Images were captured with a Zeiss AxioImager upright microscope using 20x and 40x objective lens. Data presented are representative of results obtained from at least 3 mice per group. Immunofluorescent staining images were observed using an Olympus FV1000 upright confocal microscope and captured at 405nm, 488nm, and 568nm using 40x and 100x objective lens, processing with FV10-ASW 3.0 Viewer software.
Protein extraction and western blot {#s4_3}
-----------------------------------
Mechanically separated crypts and villi of small intestine \[[@R46]\], were homogenized and lysed using RIPA buffer (25mM Tris•HCl pH 7.6, 150mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS, Thermo Scientific, \#89901) supplemented with protease inhibitor cocktail (Roche). Supernatants were collected after centrifugation at 14,000 r.p.m for 1 hour. Protein concentration was determined by the BCA method (Pierce, Thermo Scientific). Proteins were subjected to SDS−PAGE and immunoblot analysis. Blots were probed sequentially with primary and secondary antibodies at the following dilutions: anti-p53 at 1:1000 (1C12, Cell Signaling), anti-Actin at 1:5000 (Sigma-Aldrich Corporation, St. Louis, USA). Secondary HRP-conjugated anti-mouse and anti-rabbit were used at 1:10 000 (GE Healthcare, Chalfont St Giles, UK). Proteins were detected by incubation with ECL substrate (Amersham Bioscience, Piscataway, USA) for 5 min and chemiluminescence was visualized by STORM imaging system (Amersham, Pleasanton, USA).
Total RNA isolation and qRT-PCR {#s4_4}
-------------------------------
Separated crypts of small intestine were homogenized and put into QIAzol (Qiagen, Hilden, Germany). Total RNA was extracted using the RNeasy Universal Plus Kit (Qiagen) according to the manufacturer\'s instruction. Each quantitative PCR was performed in duplicate for each primer set. Relative transcript amounts were calculated by the ΔCT method using GAPDH as a reference gene.
Primer sequences: p53 forward GTTATGTGCACGTACTCTCCTC, reverse CGTCATGTGCTGTGACTTCT. Lgr5 forward CCACAGCAACAACATCAGGT, reverse AACAAATTGGATGGGGTTGT. GAPDH forward GGAGAAACCTGCCAAGTATGA, reverse CAACCTGGTCCTCAGTGTAGC.
FISH hybridization and microscopy imaging {#s4_5}
-----------------------------------------
The FISH procedure was performed according to previously published methods \[[@R47]\]. 7 μm cryo-sections were cut for hybridizations. Lgr5 mRNA Stellaris probe (mouse: NM_010195) was designed by Probe Designer at [www.singlemoleculefish.com](http://www.singlemoleculefish.com) and synthesized by Biosearch Technologies. The FISH probe set consists of 96 TMR fluorophore labelled oligonucleotides. DAPI nuclear dye was included during the final wash. Images were captured at 405nm and 568nm using 100x objective len by an Olympus FV1000 upright confocal microscope.
*In vitro* culture of organoids from intestinal crypts {#s4_6}
------------------------------------------------------
Intestinal crypts were isolated and cultured *in vitro* as described previously \[[@R48]\]. Organoids were washed and fixed in 4% formaldehyde for an hour, then transferred to 70% ethanol before embedding in paraffin.
SUPPLEMENTARY MATERIAL FIGURES {#s5}
==============================
The authors thank Nancy Jenkins and Neal Copeland for advice and support. We thank Keith Rogers, Susan Rogers and the Institute for Molecular and Cell Biology Histology Lab for their technical assistance, as well as Graham Wright and James Zhao of IMB Microscopy Unit. We thank David Tan for helpful discussions and Jerrold Ward for pathology expertise. We thank Huang Le-Ann for production of home-made p53 antibodies. This work was funded by the Agency for Science, Technology and Research (A\*STAR) of Singapore.
**Author contributions**
AMG and YX designed the study, performed experiments, analysed data and wrote the manuscript. ML, PCC, SAR, LL, TT and JW performed experiments. MBM, GL, KMM and NB contributed to study design and data interpretation. DPL contributed to study design, data interpretation and manuscript writing. All authors had final approval of the submitted manuscript.
**CONFLICTS OF INTEREST**
The authors declare no conflict of interest.
| {
"pile_set_name": "PubMed Central"
} |
{#sp1 .385}
{#sp2 .386}
| {
"pile_set_name": "PubMed Central"
} |
Introduction {#Sec1}
============
Parkinson's disease (PD) is the second most common age-related neurodegenerative disorder following Alzheimer's disease. The main pathological feature of PD is the progressive loss of dopaminergic neurons in the Substantia Nigra pars compacta (SNpc), with the subsequent loss of dopamine (DA) in the striatum. Although much uncertainty still exists about the aetiology of PD, available evidence suggests the implication of numerous processes such as oxidative stress, mitochondrial dysfunction, inflammation and cell apoptosis. Specifically, the deleterious role of systemic inflammation in the onset and progression of PD is becoming evident, so that the interest in its study has increased in the last years^[@CR1]^.
In this sense, it is showed that dopaminergic neurons are more vulnerable to oxidative stress and pro-inflammatory cytokines because of their low levels of intracellular glutathione concentrations^[@CR2]^. Thus, a sustained systemic or brain inflammation involves activated microglia cells that secrete pro-inflammatory factors that damage neurons^[@CR3]^. At the same time, damaged neurons release toxic factors that recruit more glial cells, resulting in a fatal vicious cycle.
To clarify the implication of systemic inflammation in PD, different combinations of experimental animal models have been reported. The most commonly used animal model to approach this issue is based on the systemic administration of lipopolysaccharide (LPS). Extensive research has shown that this endotoxin, from gram-negative bacteria, activates microglia and produces a progressive and cumulative loss of dopaminergic neurons over time^[@CR4]^. For example, Qin et al. reported that LPS stimulates cells in the liver to produce TNF-α that is distributed in the blood to the brain to induce the synthesis of more TNF-α and, consequently, damaging dopaminergic neurons^[@CR5]^. In another study, the systemic administration of LPS was combined with the induction of ulcerative colitis by the oral ingestion of dextran sulphate sodium (DSS). The results from this work showed an exacerbation of LPS-induced damage in the nigrostriatal system^[@CR6]^. Furthermore, García-Domínguez and colleagues combined LPS with the 1-methyl-4-phenyl-1,2,3,6 tetrahydropyridine (MPTP)-based model of PD. The obtained data reinforces the activation of microglial related-events together with the exacerbation of the dopaminergic neurodegeneration^[@CR7]^. Together with these observations, in our previous study, we showed an exacerbation of the dopaminergic neuronal death and the glial activation both in the SNpc and striatum when combining MPTP and DSS intoxications^[@CR8]^.
Despite the published literature, further analyses are yet required to elucidate in depth the cellular and molecular mechanisms triggered after a systemic insult in Parkinsonian mice. In the present study, we performed microarray and bioinformatics analysis to identify differentially expressed genes (DEGs) of mice treated with MPTP and/or DSS. This work's aim was to provide a better understanding of the effect of systemic inflammation on the development of PD.
Material and methods {#Sec2}
====================
Animals {#Sec3}
-------
The study was carried out on 16 three-months-old male C57BL/6J mice acquired from Charles River (Janvier, Le Genest Saint Isle, France). Animals were housed in a special room under regulated temperature (21 ± 1 °C) and 12-h light/dark cycles. The "Three R's principle" was carefully applied in our study. All procedures related to animal maintenance, care and experimentation were conducted in accordance with the European Community Council Directive (2010/63/UE) for animals to be used in preclinical studies, and were approved by the Institutional Committee on Animal Ethics of the University of Murcia (REGA ES300305440012).
Regimen of intoxication for DSS and MPTP {#Sec4}
----------------------------------------
Ulcerative colitis intoxication was induced for 8 days by oral administration of 2--2.5% of DSS (molecular weight, 36--50 kDa, MP Biomedicals LLC, OH, USA) in tap water. On day fourth, MPTP + HCl was intraperitoneal administrated with two injections at 2-h intervals in 1 day dissolved in 0.9% saline (15 mg/Kg, Sigma Aldrich, St Quentin). Mice were distributed in four experimental groups: (a) Control (n = 4), (b) DSS (n = 4), (c) MPTP (n = 5) and (d) MPTP + DSS (n = 5) (Fig. [1](#Fig1){ref-type="fig"}a). On day 8, mice were sacrificed by decapitation and brains were immediately removed and dissected into striatum and midbrain^[@CR8]^. Samples were correctly stored depending on their future use.Figure 1(**a**) Experimental design and mice distribution in the different groups. (**b**) Scheme of the main steps for obtaining and analyzing the results.
RNA extraction and purification for microarray analysis {#Sec5}
-------------------------------------------------------
Striatums were homogenated using QIAzol from the miRNeasy Mini Kit (Qiagen, Hilden, Germany) and stored at -- 80 °C until RNA extraction. Total RNA was extracted using the miRNeasy Mini Kit according to the manufacturer's instructions. RNA samples were quantitated on a NanoDrop 2000 (Thermo Fisher Scientific, Whaltham, MA). RNA quality was examined on an Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Palo Alto, CA) using the RNA 6000 Nano Kit. RNA integrity numbers (RINs) of the samples ranged between 9.2 and 9.7. Samples were stored at -- 80 °C until microarray experiments.
RNA labelling, microarray hybridization and feature extraction {#Sec6}
--------------------------------------------------------------
RNA samples were thawed and labeled using Agilent Two Color Quick Amp Labeling and RNA Spike-In kits (Agilent), according to the manufacturer\'s protocol. Experimental samples were labeled with cyanine 5-CTP and used as tests. An Agilent Universal Mouse Reference RNA was labeled with cyanine 3-CTP and used as reference. Each of the labeled test cRNAs was mixed together with labeled reference cRNA. Then, mixes were hybridized onto SurePrint G3 Mouse Gene Expression v2 8 × 60 K microarrays targeting 27,122 Entrez Gene RNAs and 4,578 lncRNAs, using the Agilent Gene Expression Hybridization kit. After hybridization, the microarray slides were washed and scanned in an Agilent G2565CA DNA Microarray Scanner. Images were analyzed with the Agilent Feature Extraction software to automatically generate the datasets. Log~10~ ratios (test vs reference) were computed after normalization correction performed by linear and Lowess methods.
Analysis of the array expression {#Sec7}
--------------------------------
The raw data obtained was analyzed with the statistical language R^[@CR9]^ and limma package was used^[@CR10]^. Intra-array normalization was performed with the method of *Loess* and inter-array normalization with the *Aquantil* method. The expression levels of the signals belonging to the same gene were grouped, and the final expression value corresponds to the mean value, as described by Limma instructions. The comparisons between the groups was: D (DSS) vs Ctrl (Control); M (MPTP) vs Ctrl; MD (MPTP + DSS) vs Ctrl; D vs M; D vs MD and M vs MD. DEGs were obtained based on the false discovery rate (FDR, \< 0.05) for each comparison.
Enrichment analysis {#Sec8}
-------------------
The enrichment analyses for the up and down DEGs were carried out with gProfileR^[@CR11]^ that brings together enrichments from several databases, those of our interest were the annotations from Gene Ontology (GO) and Reactome. From the obtained GO annotations, we carefully summarize and remove redundant terms using Revigo^[@CR12]^. These results were illustrated in a network diagram using Cytoscape^[@CR13]^ for each ontology for comparisons: biological process (BP), molecular function (MF) and cellular component (CC) (Fig. [1](#Fig1){ref-type="fig"}b). The relationships between the nodes are based on the hierarchical association between the terms of the ontology. The brightness of the nodes refers to the level of significance.
Ethics approval and consent to participate {#Sec9}
------------------------------------------
Not applicable.
Consent for publication {#Sec10}
-----------------------
Not applicable.
Results {#Sec11}
=======
Differential gene expression profiles {#Sec12}
-------------------------------------
Firstly, normalization from the raw data (Fig. [2](#Fig2){ref-type="fig"}a.i) was made within (Fig. [2](#Fig2){ref-type="fig"}a.ii) and between groups (Fig. [2](#Fig2){ref-type="fig"}a.iii). To avoid any supervised analysis, volcano plots from all the comparisons between the experimental groups showed the distributions of the intensities (Fig. [2](#Fig2){ref-type="fig"}b). Moreover, samples from the four experimental groups were distinctly separated in the principal component analysis (PCA) plot, indicating a differential gene expression profile caused by both intoxications (Fig. [2](#Fig2){ref-type="fig"}c).Figure 2Validation of the DEGs profiles. (**a**) Normalization graphs from the raw data (i) was performed within (ii) and between (iii) groups. (**b**) Volcano plots comparing the levels of genes expressions between the experimental groups (i) DSS vs control; (ii) MPTP vs control; (iii) MPTP + DSS vs control; (iv) DSS vs MPTP; (v) DSS vs MPTP + DSS and (vi) MPTP vs MPTP + DSS. (**c**) PCA plot for the four experimental groups of interest (control, DSS, MPTP and MPTP + DSS).
Overall, a total of 2,413 genes were found to be differentially expressed, with an absolute fold change set at \> 2 and *p*-value \< 0.05 in all possible comparisons between the experimental groups \[see Supplementary file [1](#MOESM1){ref-type="media"}\]. Of the total DEGs, different number of genes were specifically identified based on the experimental group: DSS (132 genes), MPTP (345 genes) and MPTP + DSS (1,357 genes) compared with the control group, and MPTP compared with MPTP + DSS (281 genes) suggesting different effects in the striatum (Table [1](#Tab1){ref-type="table"}). Moreover, in the independent hierarchical clustering represented as heat-maps, it was confirmed the DEGs within and between groups \[see Supplementary file [1](#MOESM1){ref-type="media"}\], specifically exacerbated in the comparisons between MPTP + DSS and control groups.Table 1Total number of DEGs in the different experimental groups \[see Supplementary file [2](#MOESM1){ref-type="media"}\].ComparisonsTotal DEGsDown-regulatedUp-regulatedDSS vs ctrl1326369MPTP vs ctrl345150195MPTP + DSS vs ctrl1,357634723MPTP vs MPTP + DSS281124157
GO enrichment analysis {#Sec13}
----------------------
Based on the high amount of differential expressed genes between groups, we aimed to perform a functional enrichment analysis using g:GOSt from gProfiler which provides the most enriched GO terms together with other data sources (data shown in the Supplementary Files) associated with our gene lists.
Specifically, the data from the GO enrichment analysis is organized into three categories: (i) biological process, (ii) cellular component and (iii) molecular function. All the GO terms lists obtained in each category were summarize removing redundant terms using REVIGO. The output data was used to generate the GO-terms network graphs in Cytoscape. The setting threshold for the false discovery rate (FDR at 0.05).
Biological process {#Sec14}
------------------
Gene ontology enrichment analysis revealed different profiles depending on the experimental groups. Specifically, the results that concern biological process mainly involved GO terms for response to toxic substance/stimulus, detoxification and immune system but in different way depending on the experimental groups. Thus, up-regulated DEGs, related to response to toxic substance/stimulus, were found in groups treated with DSS (DSS and MPTP vs control; and, MPTP + DSS vs MPTP, Figs. [3](#Fig3){ref-type="fig"} and [4](#Fig4){ref-type="fig"}). Interestingly, we found up-regulated DEGs related to detoxification mechanisms in animals treated with DSS compared to control mice (Table [2](#Tab2){ref-type="table"}, Fig. [3](#Fig3){ref-type="fig"}a) and MPTP + DSS, clearly observed when comparing MPTP + DSS vs MPTP (Table [3](#Tab3){ref-type="table"}, Fig. [4](#Fig4){ref-type="fig"}).Figure 3Summary of the main GO terms found for biological process in the comparisons for (**a**) DSS vs control and (**b**) MPTP vs control. The color of the circles help to better identify the groups of terms for specific processes: green for immune system process, pink for response to toxic substance, blue for apoptotic pathway and purple for response to unfolded protein. Figure 4Summary of the main GO terms found for biological process in the comparisons for MPTP + DSS vs control (**a**) and MPTP (**b**, i--iv). The color of the circles help to better identify the groups of terms for specific processes: green for immune system process (ii, **b**), pink for response to toxic substance/stimulus (ii, **b**), aquamarine for oxidative stress (iii), orange for cell activation (iv) and red for cell death (iv). Table 2The enriched GO terms for the up-regulated DEGs of DSS and MPTP vs control, respectively.GO TermDescriptionCorrected *p*-valueDEGs count**DSS vs control**GO:0098869Cellular oxidant detoxification1.08E−046GO:1990748Cellular detoxification1.10E−046GO:0098754Detoxification1.44E−046GO:0097237Cellular response to toxic substance1.00E−037GO:0015893Drug transport2.71E−036GO:0002376Immune system process8.49E−0317GO:0042493Response to drug3.23E−0210GO:0030097Hemopoiesis3.23E−029GO:0046501Protoporphyrinogen IX metabolic process3.23E−022GO:0048534Hematopoietic or lymphoid organ development4.23E−029**MPTP vs control**GO:0035966Response to topologically incorrect protein7.61E−0911GO:0006986Response to unfolded protein1.67E−0610GO:0035967Cellular response to topologically incorrect protein6.57E−058GO:0034620Cellular response to unfolded protein2.64E−047GO:0030968Endoplasmic reticulum unfolded protein response1.55E−036GO:0034976Response to endoplasmic reticulum stress6.22E−039GO:0002376Immune system process2.88E−0231GO:2001234Negative regulation of apoptotic signaling pathway3.29E−028GO:1904761Negative regulation of myofibroblast differentiation3.31E−022GO:1903573Negative regulation of response to endoplasmic reticulum stress4.80E−024 Table 3The enriched GO terms for the up-regulated DEGs of MPTP + DSS vs control and MPTP, respectively.GO TermDescriptionCorrected *p*-valueDEGs count**MPTP + DSS vs control**^a^GO:0070887Cellular response to chemical stimulus3.70E−13155GO:0098754Detoxification3.36E−1124GO:0002376Immune system process5.01E−11135GO:0006950Response to stress6.10E−11173GO:1990748Cellular detoxification1.07E−0921GO:0002682Regulation of immune system process1.65E−0986GO:0009605Response to external stimulus3.46E−09135GO:0006954Inflammatory response5.40E−0953GO:0050896Response to stimulus2.14E−08345GO:0097237Cellular response to toxic substance3.37E−0829GO:0080134Regulation of response to stress3.37E−0878GO:0009636Response to toxic substance3.37E−0845GO:0042127Regulation of cell population proliferation4.34E−0893GO:0006952Defense response4.34E−0888GO:0031347Regulation of defense response4.34E−0848GO:0098869Cellular oxidant detoxification4.52E−0818GO:0042592Homeostatic process7.16E−0898GO:0032502Developmental process1.14E−07238GO:0032101Regulation of response to external stimulus1.14E−0760GO:0001775Cell activation1.39E−0765GO:0048583Regulation of response to stimulus6.53E−07167GO:0001816Cytokine production8.61E−0749GO:0048856Anatomical structure development9.28E−07220**MPTP + DSS vs MPTP**^b^GO:0098754Detoxification5.56E−048GO:0002376Immune system process5.56E−0434GO:1990748Cellular detoxification2.93E−037GO:0055076Transition metal ion homeostasis8.11E−037GO:0097577Sequestering of iron ion9.03E−032GO:0051707Response to other organism9.03E−0321GO:0046916Cellular transition metal ion homeostasis9.03E−036GO:0046110Xanthine metabolic process9.03E−032GO:0098869Cellular oxidant detoxification9.03E−036GO:0043207Response to external biotic stimulus9.03E−0321GO:0009607Response to biotic stimulus9.03E−0321GO:0009605Response to external stimulus9.44E−0332See complete tables in the Supplementary Files.^a^Results shown for MPTP + DSS vs control with corrected *p*-value ≤ 1E-07.^b^Results shown for MPTP + DSS vs MPTP with corrected *p*-value ≤ 1E-03.
Moreover, the data revealed that genes related to immune system processes are up-regulated in DSS (Fig. [3](#Fig3){ref-type="fig"}a), MPTP (Fig. [3](#Fig3){ref-type="fig"}b) and MPTP + DSS (Fig. [4](#Fig4){ref-type="fig"}a) groups compared to the control group; and, when comparing MPTP + DSS with MPTP group (Fig. [4](#Fig4){ref-type="fig"}b). However, these genes are especially exacerbated when comparing MPTP + DSS with control and MPTP mice (Fig. [4](#Fig4){ref-type="fig"}a.ii and Table [3](#Tab3){ref-type="table"}). Together with the exacerbation of the immune system process, we also observed in this group oxidative stress processes (Fig. [4](#Fig4){ref-type="fig"}a.iii and Table [3](#Tab3){ref-type="table"}), cell activation and cell death (Fig. [4](#Fig4){ref-type="fig"}a.iii and Table [3](#Tab3){ref-type="table"}).
Cellular component {#Sec15}
------------------
Unlike in the category of biological process where in all groups came out significant up-regulation of the DEGs; in the cellular component were significantly observed only in the comparisons: DSS vs control, MPTP + DSS vs control and MPTP + DSS vs MPTP (Fig. [5](#Fig5){ref-type="fig"}). Interestingly, in the three comparisons (animals treated with DSS) the groups of genes related to haptoblobin-hemoglobin complex and the extracellular region stand out. In addition, when animals are treated together with MPTP (Fig. [5](#Fig5){ref-type="fig"}b and c.i, Table [4](#Tab4){ref-type="table"}), changes are also observed in the intracellular part.Figure 5Summary of the main GO terms obtained for cellular component in the comparisons for DSS vs control (**a**), MPTP + DSS vs MPTP (**b**) and MPTP + DSS vs control (**c**). The color of the circles correspond to different groups of annotations: purple for haptoblobin-hemoglobin complex, orange for extracellular region and green for intracellular part. Table 4The enriched GO terms for the DEGs of DSS and MPTP vs control, respectively.GO TermDescriptionCorrected *p*-valueDEGs count**DSS vs control**GO:0031838Haptoglobin--hemoglobin complex7.02E−074GO:0005833Hemoglobin complex1.46E−064GO:0005576Extracellular region1.86E−0420GO:0044445Cytosolic part2.26E−036GO:0036477Somatodendritic compartment2.77E−0311GO:0033093Weibel-palade body2.77E−032GO:0005615Extracellular space2.77E−0313GO:0044421Extracellular region part3.42E−0314**MPTP + DSS vs MPTP**GO:0005576Extracellular region2.42E−0639GO:0031838Haptoglobin--hemoglobin complex2.42E−064GO:0044421Extracellular region part3.43E−0631GO:0005615Extracellular space3.43E−0628GO:0005833Hemoglobin complex2.15E−033**MPTP + DSS vs MPTP**GO:0044421Extracellular region part6.98E−07104GO:0005615Extracellular space6.98E−0796GO:0005576Extracellular region1.73E−06122GO:0044444Cytoplasmic part1.73E−06302GO:0005737Cytoplasm3.19E−05360GO:0005829Cytosol1.01E−04147GO:0031838Haptoglobin--hemoglobin complex3.64E−045GO:0031012Extracellular matrix6.95E−0430GO:0009986Cell surface7.29E−0453GO:0008305Integrin complex2.41E−037GO:0044445Cytosolic part3.07E−0320GO:0098636Protein complex involved in cell adhesion3.58E−037GO:0005833Hemoglobin complex5.81E−034GO:0009897External side of plasma membrane7.66E−0330GO:0098552Side of membrane7.66E−0337
Molecular function {#Sec16}
------------------
Regarding molecular function GO annotations, we found terms from up-regulated DEGs mainly implicated in antioxidant activity and haptoglobin binding in the comparisons for DSS vs control, MPTP + DSS vs control and MPTP + DSS vs MPTP (Fig. [6](#Fig6){ref-type="fig"}, Table [5](#Tab5){ref-type="table"}).Figure 6Summary of the main GO terms obtained for molecular function in the comparisons for DSS vs control (**a**), MPTP + DSS vs control (**b**) and MPTP + DSS vs MPTP (**c**). Table 5The most significant enriched GO terms for the DEGs of DSS and MPTP vs control and MPTP + DSS vs MPTP.GO termDescriptionCorrected *p*-valueDEGs count**DSS vs control**^a^GO:0031720Haptoglobin binding9.59E−064GO:0005344Oxygen carrier activity3.37E−064GO:0019825Oxygen binding2.73E−054GO:0016209Antioxidant activity9.96E−055GO:0140104Molecular carrier activity1.63E−044GO:0004601Peroxidase activity2.19E−044GO:0031722Hemoglobin beta binding2.19E−042GO:0016684Oxidoreductase activity, acting on peroxide as acceptor2.52E−044GO:0031721Hemoglobin alpha binding1.02E−032GO:0030492Hemoglobin binding2.78E−032GO:0015106Bicarbonate transmembrane transporter activity2.78E−032GO:0005515Protein binding2.78E−0335GO:0005452Inorganic anion exchanger activity7.66E−032**MPTP + DSS vs control**^a^GO:0042802Identical protein binding7.99E−0797GO:0016209Antioxidant activity8.02E−0716GO:0005515Protein binding2.33E−06348GO:0030414Peptidase inhibitor activity1.78E−0522GO:0004857Enzyme inhibitor activity1.93E−0531GO:0061134Peptidase regulator activity4.69E−0523GO:0004866Endopeptidase inhibitor activity6.02E−0520GO:0061135Endopeptidase regulator activity1.28E−0420GO:0048037Cofactor binding3.16E−0437GO:0005488Binding4.53E−04460GO:0016491Oxidoreductase activity5.62E−0446GO:0003824Catalytic activity1.12E−03209GO:0030234Enzyme regulator activity1.12E−0350GO:0042803Protein homodimerization activity1.22E−0348GO:0004869Cysteine-type endopeptidase inhibitor activity1.35e−0310GO:0016722Oxidoreductase activity, oxidizing metal ions1.50E−036GO:0016684Oxidoreductase activity, acting on peroxide as acceptor1.75E−039GO:0005506Iron ion binding1.75E−0318GO:0004322Ferroxidase activity2.04E−035GO:0016724Oxidoreductase activity, oxidizing metal ions, oxygen as acceptor2.04E−035GO:1901567Fatty acid derivative binding2.53E−037GO:0031720Haptoglobin binding3.21E−034GO:0046906Tetrapyrrole binding3.49E−0316GO:0004165Dodecenoyl-coa delta-isomerase activity3.73E−033GO:0004601Peroxidase activity4.13E−038GO:0020037Heme binding6.13E−0315GO:0043177Organic acid binding6.15E−0317GO:0046914Transition metal ion binding6.71E−0350**MPTP + DSS vs MPTP**^a^GO:0031720Haptoglobin binding4.89E−043^a^Results shown with corrected *p*-value ≤ 1E−03.
Reactome enriched analysis {#Sec17}
--------------------------
The data obtained from the Reactome annotations divided the comparisons into different events in the same scenario (striatum level). Firstly, DSS compared to control stand out mechanisms related to heme dysfunction (as scavenging of heme from plasma or heme biosynthesis) possibly due to the bleeding produced by the DSS administration. Since free heme promotes the conversion of low density lipoproteins into cytotoxic products, make it toxic, it is also observed immune system processes up-regulation (Table [6](#Tab6){ref-type="table"}).Table 6Functional enriched Reactome annotations for all the comparisons.Reactome codeDescriptionCorrected *p*-valueDEGs count**DSS vs control**1247673Erythrocytes take up oxygen and release carbon dioxide7.33E−0851237044Erythrocytes take up carbon dioxide and release oxygen1.92E−0751480926O~2~/CO~2~ exchange in erythrocytes1.92E−0752168880Scavenging of heme from plasma8.25E−0552173782Binding and uptake of ligands by scavenger receptors1.52E−045189451Heme biosynthesis1.08E−022189445Metabolism of porphyrins2.54E−022216083Integrin cell surface interactions3.02E−023168256Immune system3.02E−0213**MPTP vs control**140342Apoptosis induced DNA fragmentation1.45E−0552559584Formation of senescence-associated heterochromatin foci (SAHF)4.68E−0552559586DNA damage/telomere stress induced senescence4.58E−03675153Apoptotic execution phase5.18E−035**MPTP + DSS vs control**168256Immune system1.48E−07111168249Innate immune system1.14E−06736798695Neutrophil degranulation2.75E−06471247673Erythrocytes take up oxygen and release carbon dioxide2.91E−044198933Immunoregulatory interactions between a lymphoid and a Non-lymphoid cell2.12E−03195686938Regulation of TLR by endogenous ligand2.12E−036804914Transport of fatty acids5.34E−034425397Transport of vitamins, nucleosides, and related molecules8.93E−0388978868Fatty acid metabolism1.33E−0219109582Hemostasis3.06E−0239216083Integrin cell surface interactions4.19E−02101236975Antigen processing-cross presentation4.45E−0272142688Synthesis of 5-eicosatetraenoic acids4.68E−023**MPTP + DSS vs MPTP**174577Activation of C3 and C57.20E−0331247673Erythrocytes take up oxygen and release carbon dioxide1.21E−0231480926O~2~/CO~2~ exchange in erythrocytes1.79E−0231237044Erythrocytes take up carbon dioxide and release oxygen1.79E−023
Regarding MPTP compared to control, we observed two main pathways: (i) programmed cell death by apoptosis induced DNA fragmentation, and (ii) cellular responses to external stimuli that implicated DNA damage/telomerase stress induced senescence (Table [6](#Tab6){ref-type="table"}).
On the other hand, the last events are exacerbated when comparing MPTP + DSS with both control and MPTP mice, which are the immune system processes. Specifically, it is observed: (i) the innate immune system response, with the up-regulation of the neutrophil degranulation and the toll-like receptor cascade (regulation of TLR by endogenous ligand) and, (ii) the adaptive immune system response, with the Class I MHC mediated antigen processing and presentation (antigen processing cross presentation) and immunoregulatory interactions between a lymphoid and a non-lymphoid cell (Table [6](#Tab6){ref-type="table"}).
Discussion {#Sec18}
==========
Prior work have remarked the involvement of systemic inflammation in the development of PD. The findings from the present work extend those published reports through a bioinformatics approach. We aimed to elucidate in depth the molecular and cellular mechanisms triggered after a systemic insult in the striatum of mice untreated and treated with MPTP. The importance of specifically studying the effect of both toxins in the striatum is based on previous data obtained by immunohistochemical techniques in which we observed exacerbation of the inflammatory events when combining both toxins^[@CR8]^.
The results from this study provide novel information about the DEGs in the striatum of mice intoxicated only with DSS. It is clear to observe the up-regulation of processes related to detoxification and inflammatory mechanisms (Fig. [3](#Fig3){ref-type="fig"}, Table [3](#Tab3){ref-type="table"}) mainly triggered by the bleeding and lesion of the mucosal surface caused by the administration of DSS^[@CR14]^. This is the first study to our knowledge to describe the molecular and cellular mechanisms activated in the striatum after a local insult in the colon.
Otherwise, when comparing MPTP mice with the control group, we found biological processes related to the response to unfolded protein (Fig. [3](#Fig3){ref-type="fig"}b, Table [2](#Tab2){ref-type="table"}). Interestingly, these results are in accordance with the observations of Fornai and colleagues who described that the administration of MPTP reduces the protein degradation function of the striatal ubiquitin--proteasome system^[@CR15]^. Later on, it was added that this could lead the accumulation of unfolded proteins and consequently, activates stress-induced cell death mechanisms^[@CR16]^. Moreover, it was demonstrated that oxidative stress processes may participate in the formation of cross-linked protein aggregates^[@CR17]^. In these animals, we also noticed DEGs associated with an attempt to regulate negatively the apoptotic signalling pathway and positively the activation of the inflammatory response (Table [3](#Tab3){ref-type="table"}). Interestingly, the data from Reactome analysis showed annotations related to the execution of apoptosis mechanisms induced by DNA fragmentation and the induction of cellular senescence. These results are in line with those reported to another study that suggests that the nature and severity of DNA damage may determine the cellular response^[@CR18]^. More research is needed to clarify how is the interplay of both routes after a DNA insult^[@CR19]^.
Most notably were the observations that resulted from the comparison between MPTP + DSS and control animals. As this work indicates, we obtained an exacerbation of the up-regulated DEGs in all the analysis related to detoxification, oxidative stress and immune system processes.
Our results not only reinforce those results reported in the literature but also extend detailed information on the contribution of systemic inflammation to the progression of the neurodegeneration associated with Parkinson's disease. Although our aim was highly achieved, we are aware of some intrinsic limitations such as the number of animals used in this type of studies^[@CR20],[@CR21]^ or the focus in one brain area. The relevance of this lies in the contribution of novel and detailed information that describes different expression profiles that can be used as a guide for further and specific analysis. We advise that future work should evaluate upregulated pathways in different brain areas over time. In this way, we will be able to identify promising therapeutic targets that prevent the contribution of inflammatory processes in the progression of Parkinson's disease.
Conclusions {#Sec19}
===========
Altogether, we provide functional and comprehensive bioinformatics analyses of the deleterious effect of the systemic inflammation in the striatum of MPTP intoxicated mice. Interestingly, the data showed in this study describes a scenario that becomes more complex when combining both treatments. Thus, the processes related to inflammation and oxidative stress are exacerbated resulting in a significant up-regulation of the cell death mechanisms.
Supplementary information
=========================
{#Sec20}
Supplementary Information 1. Supplementary Information 2.
**Publisher\'s note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
=========================
is available for this paper at 10.1038/s41598-020-69695-4.
Research work of the authors was supported by the Spanish Ministry of Science and Innovation (FIS PI13 01293); Fundación Séneca (19540/PI/14) and "Prediction of cognitive properties of new drug candidates for neurodegenerative diseases in early clinical development" (European Community's Seventh Framework Programme (FP7/2007-2013) for the Innovative Medicine Initiative under Grant Agreement No. 115009) to MTH.
M.T.H. and A.L.G.M. designed the research; A.L.G.M. and C.S.R. performed the research; A.L.G.M. analyzed the data and designed the figures; L.C.B., A.G.C., A.P., S.V. and E.F.V. contributed to the discussion of the in vivo results; A.L.G.M. and M.T.H. discussed all the results and wrote the paper.
All data generated or analysed during this study are included in this published article \[and its supplementary information files\].
The authors declare no competing interests.
| {
"pile_set_name": "PubMed Central"
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.