NGF Modulates trkANGFR/p75NTR in αSMA-Expressing Conjunctival Fibroblasts from Human Ocular Cicatricial Pemphigoid (OCP)

Objective In a previous study, we reported the upregulation of Nerve Growth Factor (NGF) and trkANGFR expression in Ocular Cicatricial Pemphigoid (OCP), an inflammatory and remodeling eye disease. Herein, we hypothesize a potential NGF-driven mechanism on fibroblasts (FBs) during OCP remodeling events. To verify, human derived OCP-FBs were isolated and characterized either at baseline or after NGF exposure. Materials and Methods Conjunctival biopsies were obtained from 7 patients having OCP and 6 control subjects (cataract surgery). Both conjunctivas and primary FB cultures were characterised for αSMA, NGF and trkANGFR/p75NTR expression. Subcultures were exposed to NGF and evaluated for αSMA, NGF, trkANGFR/p75NTR expression as well as TGFβ1/IL4 release. For analysis, early and advanced subgroups were defined according to clinical parameters. Results OCP-conjunctivas showed αSMA-expressing FBs and high NGF levels. Advanced OCP-FBs showed higher αSMA expression associated with higher p75NTR and lower trkANGFR expression, as compared to early counterparts. αSMA expression was in keeping with disease severity and correlated to p75NTR. NGF exposure did not affect trkANGFR levels in early OCP-FBs while decreased both αSMA/p75NTR expression and TGFβ1/IL4 release. These effects were not observed in advanced OCP-FBs. Conclusions Taken together, these data are suggestive for a NGF/p75NTR task in the potential modulation of OCP fibrosis and encourages further studies to fully understand the underlying mechanism occurring in fibrosis. NGF/p75NTR might be viewed as a potential therapeutic target.

An involvement of Nerve Growth Factor (NGF) pathway in OCP has been previously reported by our group: an increased trkA NGFR immunoreactivity has been observed in OCP conjunctival stroma and a consistent NGF release has been quantified in OCP tears [8,9]. The effect of NGF in tissue remodelling and fibroblast activity is actually controversial: NGF might exert pro/anti-inflammatory effects or profibrogenic activity, acting as a "modulator" of the local immune/inflammatory response, in a receptor expression dependent manner [10][11][12][13][14][15][16]. In the last decade, the NGF modulatory effect on Fibroblasts (FBs) and their activated myofibroblast counterpart (myoFBs) has been prospected in view of the surface trkA NGFR /p75 NTR rate expression and the NGF ability to trigger apoptosis in FBs from different tissues as well as TGFβ1-induced myoFBs [17][18][19][20][21].
To address the question as whether NGF might modulate OCP-fibrosis, activated FBs and NGF immunoreactivity were verified both in tissues and cultures. Next, the study was extended to the in vitro characterization of OCP-FBs and the potential NGF influence on OCP-FB phenotype by monitoring αSMA, trkA NGFR /p75 NTR and TGFβ1/IL4 in NGF-exposed OCP subcultures.

Ethics Statement
The study followed the guidelines of the Declaration of Helsinki for research involving human subjects and was approved by the intramural Ethical committee (UCBM). Informed written consent was signed by each patient adhering to the study.
Two fragments were produced from each biopsy: one conjunctival fragment was included in paraffin and sectioned to provide 5μm-sections for light/confocal microscopy, while the other fragment was used to achieve primary culture of conjunctival OCP FBs.
All sterile tissue culture plastic-ware and reagents were from NUNC (Roskilde, Denmark) and Serva (Heidelberg, Germany). Culture media and supplements were from Euroclone (Milan, Italy). Ultrapure/RNAse free water was provided by Direct-Q5 Apparatus (Millipore, Vimodrone, Milan, Italy).
Conditioned media were collected for ELISA/Western Blot analysis, while monolayers were directly processed for confocal analysis or treated with trypsin-EDTA solution to obtain single cells for molecular (10 5 cells) / biochemical (10 6 cells) analysis.

Flow cytometry
Single-cells were washed in HBSS containing Ca ++ /Mg ++ and fixed in 3.7% buffered PFA. After washing, cells were either directly immunostained or further incubated in 70% methanol in PBS (20°C, 24hrs). For staining, cells were equilibrated in FACS buffer [0.1% saponin and 0.1% NaN 3 in PBS; pH 7.5] and incubated with the above reported primary antibodies diluted in 0.1% BSA-FACS buffer. The specific binding was detected with Cy2/PE-conjugated specie-specific secondary antibodies (1/600-1/700; Jackson). Cells (10 4 events) were analysed using the MACSQuant flow cytometer and cell plots were arranged using the manufacturers' provided software (Miltenyi Biotech., Gladbach, Germany). Changes in Mean Fluorescence Intensity (MFI) were calculated as follows: ΔMFI = [(MFI specific / MFI not-specific ] and ΔMFI values 1 were used for statistical analysis.

Relative Real-Time PCR analysis
Total RNA was extracted from confluent monolayers using the Puregene RNA purification kit (Gentra Systems, Minnesota, USA). Total RNA samples were spectrophotometrically analysed (ND-1000; NanoDrop, Wilmington, DE; λ 260 /λ 280 >1.8). Total RNA (3 μg) samples were reverse transcribed to a final volume of 21μlL, using 50 pM oligo dT21-primer, 1 mM dNTP mix and 200 U reverse transcriptase (IMPROM; Promega, Milan, Italy) in a programmable PTC100 thermocycler (MJ Research, Watertown, MA) and 3μL were run for amplification with the specific target/referring primers in an Opticon2 MJ thermocycler (MJ Research), according to the manufacturers' instructions (Table 1). PCR experiments were carried out in a final volume of 20μL containing 3μL cDNA for target genes (or 1μL/3μL for GAPDH/H3 reference genes) and 17μL of master mix [10μL SYBR Green PCR Mix (Applied Biosystems, Foster City, CA), 0.5μL of each primer (10 pM; MWG, Biotech, Ebersberg, Germany) and milliQ-water]. Amplicons were verified for their specificity according to the Southern blotting analysis. Negative controls (without template or with total RNA) were carried out for each run, to rule out any genome contamination. Single threshold cycle values (Cts) were run in the REST 384-2006 software [24] to get increase or decrease difference in target gene expression, with respect to reference genes and compared to controls. Data are gene expression ratio provided in log2 scale.
TGFβ1 and IL4 were measured in the conditioned media by commercially available ELISA kits, according to the manufacturers' instructions (Biosource).

Statistical Analysis
All experimental procedures were conducted in triplicate, each one repeated three times. Both molecular and biochemical data (means±SD) were analysed for significant differences (p < .05), by the ANOVA-Tukey-Kramer coupled post-hoc analysis (StatView II for PC; Abacus Concepts. Inc., Barkley, CA). REST-ANOVA analysis followed by Tukey Kramer posthoc comparisons was used to validate molecular data.

Results
Only conjunctival biopsies from patients' with confirmed ophthalmic and microscopic diagnosis of OCP were included in this study. Immunofluorescent staining revealed the presence of a clear linear fluorescence alongside the BMZ, representative of autoantibodies deposition in both early (A) and advanced (B) OCP sections (Fig 1). No significant difference in IgGAM immunoreactivity was observed between early and advanced OCP, except for an increasing fluorescent signal in the epithelium and stroma (asterisk; Fig 1B). Contiguous to those IgGAM staining, basal histology was also performed by using Giemsa (Fig 1C and 1D  The amplification program was as follows: 95°C/15min, followed by 47 cycles of denaturation at 94°C/15sec, annealing Ta* at 25sec or 30sec, extension at 72°C/15sec. Melting curves for each specific primer were monitored between 60°C-90°C, at the end of amplification, before a further extension at 75°C/5min. In bold, referring gene details; *Ta, annealing temperature; bp, base-pairs of amplicons. doi:10.1371/journal.pone.0142737.t001 1F) and PAS (Fig 1G and 1H). A depletion of goblet cells was observed in early OCP (arrowheads) while the absence of goblet cells was detected in advanced OCP, as pointed in Giemsa, HE and particularly PAS panels (Fig 1C-1H). The presence of squamous metaplasia (pluristratified non-keratinised epithelial with absence of goblet cells and homogenization of connective tissue) was particularly evident in advanced OCP (see § in Fig 1D and 1F). A marked vessel ectasia was particularly evident in early OCP (see asteriks in Fig 1E). Edema and mild infiltrates with prevalence of plasmacells (arrows) were observed in early with respect to advanced OCP (Fig 1C).

Characterization of activated FBs
Adjacent conjunctival sections and primary cultures of FBs were analysed for αSMA expression by confocal microscopy and flow cytometry (FCM) analysis. As shown in Fig 2, αSMA immunoreactivity in outgrowth of OCP-tissue (B) and outgrew FBs (D), as compared to control specimens (respectively A and C; p < .05). FCM analysis confirmed the higher αSMA protein expression in OCP-FBs (Fig 2E; p < .05). FCM results were corroborated at the molecular level (+14.00±4.80 expression ratio in OCP-FBs vs. controls; p < .05).
Expression of NGF and trkA NGFR /p75 NTR OCP conjunctiva and OCP-FBs were examined for NGF-trkA NGFR /p75 NTR expression. As observed by confocal analysis, NGF expression was decreased in the epithelium and specifically increased in OCP stroma, as compared to controls (Fig 3A and 3B; p < .05). Indeed, confocal analysis showed higher cytoplasmic and perinuclear NGF immunoreactivity in outgrew OCP-FBs, supporting those data observed in OCP stroma (see arrows; Fig 3C). In addition, the conditioned media from OCP-FBs showed higher NGF levels than those of control counterparts (420.00±90.00 vs. 220.00±64.00 pg/mL; p < .05), as detected by NGF ELISA. In line, NGFmRNA was increased in these OCP-FBs, as compared to their control FBs (+5.00±1.30 expression ratio; p < .05). No significant difference was quantified between early and advanced specimens. Confocal analysis showed a high p75 NTR expression associated with a weak trkA NGFR expression in OCP-FBs (Fig 4B). By contrary, control FBs showed a huge trkA NGFR expression with respect to p75 NTR slightly expressed (Fig 4A). To support, the related FCM analysis for trkA NGFR and p75 NTR are provided in Fig 4C and 4D respectively.
αSMA and p75 NTR co-expression in OCP-derived activated FBs The higher αSMA and p75 NTR and the lower trkA NGFR protein expressions in OCP-FBs were also detected by WB analysis, as compared to control FBs (Fig 5A; p<0.05). As shown in Fig  5B, αSMA protein co-localized with p75 NTR but not with trkA NGFR , as detected by confocal analysis. This would suggest the existence of a specific trkA NGFR /p75 NTR expression ratio during the progress of fibrosis. Particularly, basal histology showed a strong infiltration of inflammatory cells (arrows pointing to plasmacells in Fig 1C), the presence of connective tissue homogenization (asterisk in 1D and 1F) in the underlying stroma ( §) and a significant decrease (absence) of goblet cells in the epithelium of advanced OCP sections (see arrowheads pointing residual goblet cells in early OCP) (Fig 1C and 1E). Activated fibroblast were also visible in early OCP (Fig 1C). No significant difference in αSMA, p75 NTR and trkA NGFR transcripts was detected in early OCP-FBs compared to controls (respectively +1.05±05, +1.30±0.50 and +2.00 ±1.00 expression ratio; p>.05). In advanced OCP-FBs, αSMA and p75 NTR transcripts were significantly increased (+9.00±0.40 and +5.00±0.50 expression ratio; p < .001) while trkA NGFR transcript was significantly decreased (-4.00±1.00 expression ratio; p < .05), as compared to controls. By FCM analysis, αSMA protein was increased in early OCP-FBs ( NGF modulation of αSMA and trkA NGFR /p75 NTR expression in OCP-FBs NGF effect on αSMA and trkA NGFR /p75 NTR modulation was thereafter investigated by exposure to 10ng/mL NGF over 24hrs. A preliminary dose response study was carried out on OCP-FBs (0-100 ng/mL NGF over 24hrs), highlighting the 10 ng/mL NGF dosage. In early OCP-FBs, decreased αSMA and p75 NTR (respectively 0.83±0.19 vs. 2.8±0.58 MFI and 1.08   24.12% co-expressing p75 NTR and 72.76% expressing only trkA NGFR . The statistical analysis showed that a decrease of 57% in trkA NGFR /p75 NTR co-expressing cells occurred in association with a shift to trkA NGFR expressing cells. The trkA NGFR /p75 NTR immunoreactivity in NGF exposed early OCP-FBs is shown (Fig 6).

Discussion
Increasing data indicate that the chronic inflammatory process occurring in OCP conjunctiva leads to FB activation and survival, with overt collagen deposition and excessive matrix deposition [3]. As a product of different structural/immune cells and FBs/myoFBs, both cytokines and growth factors actively contribute to subepithelial fibrosis and conjunctival scarring [25]. To date, different proinflammatory and profibrogenic factors have been investigated by different groups, which have focused their attention especially on receptor signalling [26]. While in a previous study we have described the trkA NGFR and NGF expression respectively in OCP conjunctivas and tears, herein we hypothesize a possible NGF role in the modulation of cultured OCP-FBs [8,9].
First of all, αSMA expression was detected in OCP conjunctiva and confirmed in primary cultures of FBs obtained from OCP explants. αSMA (α-Smooth Muscle Actin) represents the most reliable phenotypic marker for the majority of fibrotic states. Our findings indicate the presence of activated FBs inside inflamed/fibrotic OCP conjunctiva [7,27,28]. If these activated FBs are αSMA-expressing myofibroblast (myoFBs) remains to be clarified since in previous studies the possible differentiation of OCP-FBs into myoFBs was not reported [29].
An increase in the stroma and a significant decrease in the epithelium were detected for NGF in OCP conjunctival biopsies (n = 7), as compared to control ones. To the best of our knowledge, this data has not been previously described and it is supported by our recent findings showing an increased NGF content in OCP tear fluids [8,9]. The observation of NGFexpressing OCP-FBs might suggest that the NGF increase in OCP stroma as well as the increased NGF levels in OCP tears might be partially due to local activated-FBs. On the other side, the decreased NGF immunoreactivity in OCP epithelium is actually missing of explanation and not investigated/discussed in this study.
As product of stromal inflammation, it is reasonable to hypothesize that NGF might contribute to tissue remodeling by influencing the FB phenotype, as observed in previous studies conducted on other cell types [14,17,18,20]. With respect to the activated FB phenotype, NGF effects might cover either cell survival and/or soluble mediator release [14,17]. According to literature, NGF activity is driven by trkA NGFR and p75 NTR receptors, which mediate NGF signal alone or in cooperation [30][31][32][33][34][35]. As a new finding, NGF, trkA NGFR /p75 NTR and αSMA (co) expressions were detected in primary cell cultures alongside with sub-cultured OCP-FBs. Of interest, sub-cultured OCP-FBs showed the ability to retain FB phenotype upon few passages and were therefore suitable for stimulation studies. Interestingly, trkA NGFR /p75 NGF expression was found strictly dependent to the early/advanced grouping of disease as well as αSMA phenotype correlated to p75 NTR and paralleled the severity of fibrosis. Particularly, FBs from advanced OCP showed higher αSMA and p75 NTR together with lower trkA NGFR , as compared to early and control counterparts. This expression would imply a close association of p75 NTR with OCP-FB phenotype, and highlight a possible modulation of myoFB apoptosis, as observed in other systems [20,35]. To date, the role of trkA NGFR and p75 NTR in tissue remodeling remains controversial. As documented, both trkA NGFR and p75 NTR can mediate either survival or apoptosis, depending on their surface receptor (co)expression and microenvironment [33,[36][37][38]. In early healing process, high levels of trkA NGFR might drive both migration and differentiation (as initial matrix remodelling) while in late healing process the trkA NGFR downregulation might allow p75 NTR to mediate other biological activities, alone or eventually in cooperation with trkA NGFR [33,[39][40][41]. As described, fibrotic tissues appear characterized by low trkA NGFR and high p75 NTR expression [17,18,20]. In this study, a higher trkA NGFR /p75 NTR ratio (the outcome of a trkA NGFR over-expression) was observed in early OCP-FBs while lower trkA NGFR /p75 NTR ratio (the outcome of an increased p75 NTR expression) was detected in Fig 6. trkA NGFR /p75 NTR expression in NGF exposed early OCP-FBs. Confocal images showing the trkA NGFR (FC/green, A) and p75 NTR (Cy5/blue, B) immunoreactivity in early OCP-FBs exposed to NGF over 24hrs (overlays, C). Nuclei counterstained with propidium iodide are shown in all panels. The cytoplasmic and perinuclear distribution of both receptors is clearly visible. Magnifications: A-C, x400. advanced OCP-FBs, according to the clinical and histological features (infiltrates and remodelling features) [2,22,42]. To support our findings, the lower trkA NGFR /p75 NTR ratio expression in advanced OCP-FBs (the outcome of an increased p75 NTR expression) has been also reported in other fibrotic conditions, either in vitro/ex vivo [7,17,18,28,43]. A down-regulation of both αSMA and p75 NTR expression was observed in NGF-exposed early OCP-FBs, while no effect was detected in NGF-exposed advanced counterpart. The observation that NGF modulated trkA NGFR /p75 NTR ratio expression preferentially in early OCP-FBs would suggest that a potential control of activated FBs might be possible in early OCP showing a mild-moderate clinical facet, opening to potential NGF therapeutic applications. As shown, activated FBs disappear alongside "proper repair process" while αSMA-expressing activated FBs survive in pathological remodelling [7,28]. This process might be highly regulated by growth factors and cytokines, including NGF, all known to be increased in OCP tissues and tears [5,6,8,9]. Therefore, a possible cross-talk between NGF and other profibrogenic factors cannot be excluded. In line, TGFβ1 and IL4 were extensively investigated in fibrosis and are widely reported to contribute selectively to tissue remodelling and overt fibrosis in different disorders via an extensive sustaining of myoFBs [2,6,28,44]. Therefore, we wonder whether NGF might influence TGFβ1 and IL4 release from sub-cultures of OCP-FBs. The biochemical analysis highlighted a significant decrease of TGFβ1 and IL4 in the conditioned media from NGF-exposed early OCP-FBs, while only a decrease of IL4 was monitored in advanced counterparts. This selective effect holds up the potential NGF involvement in OCP remodelling, through a modulation of inflammatory/fibrogenic soluble factors, at least in early stage of disease.
Overall, OCP is a chronic inflammatory disease that slowly evolves in severe conjunctival scarring and visual impairments [1,2,45]. Most of the current OCP therapies target the suppression of inflammation, as counteracting the recurrent inflammation represents the main way to reduce progressive remodelling [46][47][48]. The findings of this in vitro study suggest a possible NGF effect on early OCP-FBs having a low trkA NGFR /p75 NTR ratio, highlighting the possible NGF effect in the modulation of FB activity during the early stages of disease. Since the topical NGF application has been suggested as a therapeutic tool in some ocular surface disorders [11,49], these findings encourage further studies to understand the underlying NGF mechanism in OCP conjunctiva in order to develop alternative strategies to counteract fibrosis.