Tissue Remodeling in Ocular Mucous Membrane Pemphigoid

Abstract

Purpose: Ocular mucous membrane pemphigoid (OcMMP) is a rare eye disease characterized by relapsing–remitting or persisting long-lasting inflammatory events associated with progressive scarring. Despite long-term immunomodulating therapy, abnormal fibrosis keeps worsening in patients with OcMMP. This study investigates the fibrotic process in patients with OcMMP, as well as the critical role of the epithelium in modulating the local fibrosis.

Methods: In this prospective, observational pilot study, patients affected by long-lasting OcMMP were compared with age- and gender-matched healthy controls. Clinical grading was assessed, and conjunctival biopsy and impression cytology were performed. Conjunctival samples were used for quantifying the expression of transcripts regulating the inflammatory and fibrogenic processes.

Results: Ocular surface clinical and functional markers worsened in patients with OcMMP with fibrotic disease progression. In more advanced disease stages, both impression cytologies and conjunctival biopsies revealed increased tissue remodeling and profibrotic markers (α-SMA and TGF-β), and decreased levels of inflammatory markers (I-CAM1, IL-10, and IL-17). Increased epithelial expression of profibrotic markers and histological changes were detected.

Conclusions: Chronic OcMMP is characterized by a progressive, aberrant self-sustaining fibrotic process that worsens clinical signs and symptoms. Conjunctival epithelial cells may transdifferentiate into myofibroblast-like phenotypes when chronically exposed to high levels of inflammation, as in the case of OcMMP. Tissue remodeling markers in OcMMP could be used as early diagnostic, prognostic, and therapeutic biomarkers, harvested in a non-invasive and painless procedure such as impression cytologies.

Fibrosis is usually a severe complication of a chronic, although frequently subclinical, inflammatory response that may involve different tissues, resulting in organ or system failures,^1–4 such as idiopathic pulmonary fibrosis, keloids, cirrhosis, several cardiac and renal disorders, and systemic fibrotic diseases.⁵ On the ocular surface, severe scarring conditions mostly occur after long-lasting, subclinical inflammatory responses, leading to local homeostatic failure.⁶ 

Ocular mucous membrane pemphigoid (OcMMP) is rare disease (1:10.000 to 1:60.000⁷) belonging to the chronic cicatrizing conjunctivitis (CCC) group,⁸ characterized by a relapsing–remitting or persistent long-term inflammatory pattern, as well as progressive scarring.^9–12 It is also a significant and interesting clinical research model for systemic uncontrolled fibrosis. 

It is still debated whether such abnormal scarring reaction is a direct result of the severity of the inflammatory response or if fibrosis is a per se abnormal response in such chronic cicatrizing diseases. Moreover, no staging systems or histological procedures allow us to assess a precise relationship among inflammatory activity, fibrotic remodeling, disease progression, and clinical findings.^10,13–15 

Several mediators and cytokines^16–21 are involved in the disease progression and are released by both adaptive^22–24 and innate^19,20,25 immune cells. Furthermore, it is well known that the epithelium takes part in both ocular surface inflammatory processes^26,27 and in usual physiological tissue remodeling mechanisms.²⁸ However, its regulatory activity in pathological fibrosis is still controversial. 

Therefore, this pilot study aims to clarify the progression of fibrosis in long-lasting cicatrizing disorders, such as OcMMP, as a self-maintaining process, as well as to investigate the epithelial expression of pro-fibrotic mediators and their significance as early diagnostic, prognostic, and therapeutic biomarkers of the clinical OcMMP. 

Materials and Methods

Study Population

This prospective observational pilot study consecutively enrolled patients attending our tertiary referral center at Campus Bio-Medico University Hospital in Rome. Patients had suggestive clinical signs or histological diagnosis of OcMMP, were at least 18 years old, were able to meet the criteria throughout the entire trial, and had the physical and mental capacity to understand and provide informed consent. Patients with a history of Stevens–Johnson/Lyell syndrome, ocular or periocular malignancy, or other local or systemic chronic inflammatory, autoinflammatory, or autoimmune diseases; patients receiving anti–vascular endothelial growth factor agents in the 45 days prior to the start of the study; and patients who were in therapy with topical or systemic steroids in the last 14 days prior to the inclusion visit and/or systemic immunosuppressive or immunomodulating medications were excluded, as well as patients in pregnancy or lactation, premenopausal women, and those with simultaneous participation in other medical studies or trials. Following these criteria, we recruited a total of 26 eyes from 13 patients with OcMMP. For the control group, we selected eight age- and sex-matched subjects. The average age of the OcMMP group was 70.8 ± 14.1 years, with a male:female sex ratio of 1:2. 

Clinical Assessment, Sampling, and Ethics

All experimental practices were performed in accordance with guidelines established by ARVO and adhered to the tenets of the Declaration of Helsinki concerning human subjects/biosamples. The study protocol was approved by the Intramural Ethical Committee (University Campus Bio-Medico). All participants provided a written informed consent to undergo clinical and laboratory assessments that were conducted in accordance with the tenets of the Declaration of Helsinki. All inclusion and exclusion criteria were confirmed during the initial visit, confirming patients’ suitability for participation in this study. Patients and controls received a complete ocular examination, including acquisition of the patient's general and ocular history, assessment of best-corrected visual acuity (BCVA), measurements of the intraocular pressure (IOP), and slit-lamp evaluation of the ocular surface and internal structures of the eye. The functional activity of the tear film was also evaluated by Schirmer type I test and the tear film break-up time (TBUT) and vital staining. Subjects were also asked to complete the Ocular Surface Disease Index (OSDI), which is a 12-parameter questionnaire designed to provide a quick assessment of eye irritation symptoms consistent with dry eye disease and their impact on vision-related functioning. Patients also underwent OcMMP staging in accordance with the Foster–Tauber staging system.²⁹ Following this classification, the 26 eyes were categorized as shown in Table 1. 

Table 1.

 

Foster–Tauber Staging and Corresponding Scores

 

Microscopic Examination: Basal Histology and Immunofluorescence

Conjunctival impression cytologies (ICs) were performed on all patients with OcMMP and age/sex matched controls at the end of the inclusion visit. In addition, enrolled participants had conjunctival biopsies (CBs) performed within 30 days of the initial visit. The biopsies were required for histological examination and direct immunofluorescence analysis for detecting tissue alterations and any manifestation of autoantibody or complement deposits within the conjunctival basal membrane, according to a standard procedure.³⁰ 

ICs were subjected to immunostaining and epifluorescent analysis. Briefly, cytofixed membranes (Bio-Fix Spray; Bio-Optica, Milan, Italy) were washed in Hank's Balanced Sodium Salt (HBSS) and equilibrated in PBS (10-mM phosphate buffer and 137-mM NaCl, pH 7.5), permeabilized (0.5% Triton X-100 in PBS [0.5% TX-PBS]) and probed with the specific antibodies recognizing the human α-smooth muscle actin (α-SMA) protein (1/60 in 0.5% TX-PBS; Novocastra, Newcastle, UK), and the specific binding was detected by using Cy2-conjugated species-specific secondary antibodies (1/500; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI, 5 µg/mL; ICN, Milan, Italy). Acquisitions were carried out using the E2000U confocal microscope equipped with C1 software (Nikon, Tokyo, Japan). Control samples were stained in parallel (control irrelevant immunoglobulin G [IgG]; Vector Laboratories, Burlingame, CA, USA) and used for the channel series acquisitions (Nikon). 

RNA Extraction and Relative Real-Time Reverse-Transcription PCR

Molecular analysis on conjunctival impression cytologies and conjunctival biopsies was performed according to the relative real-time reverse-transcription PCR procedure carried out for few selected inflammatory and fibrotic biomarkers—intercellular adhesion molecule 1 (ICAM-1), interleukin (IL)-10, IL-17, transforming growth factor-β (TGF-β), and α-SMA, in the presence of H3, GAPDH, actin, and beta-2 microglobulin (β2MG) reference genes. RNA extraction (Trizol; Euroclone, Milan, Italy), reverse transcription (ExcelRT Reverse Transcriptase, 200 U/µL; BioCell, Rome, Italy) and amplification procedure (Hydra 2X Taq Master Mix; Biolab, Biocell, Rome, Italy) were carried out according to standard procedures previously developed for ICs and CBs. The reverse transcription (RT) kit was comprised of Moloney murine leukemia virus (MMLV) reverse transcriptase, 5× RT buffer (dithiothreitol), RNasin, deoxynucleoside triphosphates, and Random Primer Mix, and reactions occurred at 40°C for 45 minutes. For amplification, HotStart SYBR Green Master Mix protocol was used in 48-well Illumina Eco Real-Time PCR System (Illumina, San Diego, CA). The protocol of amplification was as follows: pre-hold (5 minutes at 50°C) and pre-incubation for 15 minutes at 95°C, followed by 39 amplification cycles consisting of 30 seconds at 94°C, followed by a specific annealing step at 61°C to 63°C, depending on primer setup. Melting curves were registered at the end of the amplification (56°C–94.1°C with one fluorescence reading every 0.3°C). Specific amplifications were tested by verifying the single curve specific for each amplicon (base pairs [bps]). Primer details are provided in Table 2. Normalized RNAs were used for amplification, and cycle threshold (Ct) values from good melting curves were utilized for the estimation of differences between groups. Target gene expressions were provided by software (row data) according to the 2^–∆∆Ct formula (∆∆Ct = ∆Ct sample − ∆Ct calibrator). The single-target gene expressions (fold changes [FCs]) were expressed in the log₂ scale, as directly provided by Illumina software with respect to the control group (normal values). REST-384 (2006) software was also used to estimate changes in transcript expression, as calculated with respect to the reference genes. 

Table 2.

 

Primer Descriptions

 

Row Data and Statistical Analysis

Subjects’ data were reported in Excel files and then transferred to the Prism 5 platform (GraphPad, San Diego, CA, USA). The Kolmogorov–Smirnov test was preliminarily run to assess the normality of the sample distribution (descriptive analysis) and the subsequent choice of the appropriate statistical test. For clinical comparative analysis, two groups were considered: controls (CTR) and patients (OcMMP); their clinical data were analyzed by linear regression and t-test (GraphPad). Molecular data were subgrouped into a control (CTR) group, a conjunctival impression cytologies (ICs) group, and conjunctival biopsies (CBs) group. Clinical and biomolecular data were also evaluated in relation to Foster–Tauber staging. Biomolecular data were analyzed by linear regression and ANOVA (GraphPad). The significance limit was set at P < 0.05 (95% confidence interval). 

Results

In Figure 1, a series of four slit-lamp photographs illustrating the progressive stages of OcMMP can be observed, from the early to the late stages. Figure 2 shows the data related to the comparison of symptoms and objective signs between controls and patients, as well as within the different stages of severity among the OcMMP group. Patients’ OSDI questionnaire values increased significantly compared to those of the control group (OcMMP vs. CTR: 44.46 ± 4.261 vs. 15.50 ± 1.899; P < 0.001). However, there were no significant differences between patients at different staging. Patients’ BCVA was also significantly reduced compared to that of the controls (OcMMP vs. CTR: 5.481 ± 0.651 vs. 9.500 ± 0.189; P < 0.05). In this case, we found a negative correlation (r² = 0.2879; P < 0.05) between Foster–Tauber clinical staging and BCVA. When compared to controls, the OcMMP group had considerably lower TBUTs (4.77 ± 0.539 vs. 8.375 ± 0.981 s; P < 0.05). There was no significant association between TBUT and Tauber staging. However, there was a clear tendency for TBUTs to decrease as the disease progressed (r² = 0.102, P = 0.1). Patients’ type I Schirmer's tests were significantly reduced compared to those of the controls (OcMMP vs. CTR: 9.268 ± 1.036 vs. 15.50 ± 1.899 mm; P < 0.05). Furthermore, we observed a negative correlation between Schirmer type I test and the Tauber staging system (r² = 0.3059; P < 0.05). IOP was normal in all the subjects (<21 mmHg). IOP values did not show variations between the control group and the OcMMP group, nor between the different stages of the disease. 

Figure 1.

 

OcMMP stages captured using slit-lamp photography following the Foster–Tauber (F-T) staging system.²⁹ Early stages are represented by the first and second stages (F-T1 and F-T2), and the advanced stages are represented by the last two stages (F-T3 and F-T4). The F-T2 and F-T3 stages are further divided based on the percentage of conjunctival involvement (a–d), as illustrated. The arrows indicate fornix shortening in stage 2 and symblepharon in stage 3.

Figure 2.

 

Comparison of symptoms (based on OSDI) (a) and objective signs of BCVA (b), TBUT (c), Schirmer type I test (d), and IOP (e) within the different Tauber stages (corresponding score) among the patient groups.

The pathologist's histological examination of the biopsies revealed several noteworthy findings, including homogenization and edema of the chorion, inflammatory infiltrates of various grades (granulocytes, lymphocytes, plasma cells, and macrophages), ectasia of the vessels, varying degrees of reduction in goblet cells, and scarring. Figure 3 shows histological findings for both the early and late stages of the disease. Direct immunofluorescence (DIF) on conjunctival biopsy samples produced negative results for IgA, IgM, IgG, C1q, and C3 deposits in most of cases (84%), and positive results in only two samples. At molecular analysis, ICAM-1, IL-10, IL-17, and TGF-β expression was increased, but no significant changes were detected when comparing IC and CB results, except for α-SMA mRNA expression, which showed higher levels in ICs (IC vs. CB: 8.751 ± 1.877 vs. C−4.092 ± 2.053; P < 0.001) (Fig. 4a). In CBs, there was no significant correlation between Foster–Tauber staging and expression levels of ICAM-1, IL-10, IL-17, or TGF-β, but there was a significant positive correlation for α-SMA (r² = 0.5407; P < 0.05) (Fig. 4b). Similarly, ICs revealed a statistically significant increase in α-SMA expression as disease severity increased, as shown in Figure 4c (r² = 0.9185; P < 0.001), and TGF- β levels appear to follow this pattern, as well (r² = 0.4340; P < 0.05) (Fig. 4d). Again, in ICs, although ICAM-1, IL-10, and IL-17 did not provide statistically significant findings, a trend to decrease with disease severity was reported, as shown in Figure 5 (ICAM-1: r² = 0.2599, P = 0.0625; IL-10: r² = 0.2257, P = 0.0860; IL-17: r² = 0.3513, P = 0.1683). 

Figure 3.

 

Histological findings in conjunctival biopsies of OcMMP: early stages (left, a, b) versus late stages (right, c, d). EP, epithelium; FT, fibrotic tissue; EV, ectasia of vessels. The black asterisk (early stages) indicates predominantly granulocytic infiltrate; the black arrow (early stages) indicates goblet cells; the blue asterisk (late stages) indicates modest, predominantly lymphocytic infiltrate; and the blue arrow (late stages) indicates long-nucleated mesenchymal cells. Note the absence of goblet cells in the late stages of the disease.

Figure 4.

 

Transcript expression in conjunctival ICs and CBs. Bar graph shows α-SMA mRNA expression in conjunctival ICs compared with CBs (a); α-SMA mRNA expression quantified in CBs in relation to Tauber stages (corresponding score) (b); and α-SMA mRNA and TGF-β mRNA expression in conjunctival ICs in relation to Tauber stages (corresponding score) (c, d). Results from statistical analyses are shown in the upper right frame.

Figure 5.

 

Transcript expression in conjunctival ICs. ICAM-1 (a), IL-10 (b), and IL-17 (c) mRNA expression in conjunctival ICs in relation to Tauber stages (corresponding score). Results from correlation studies are shown in the upper right frame.

As shown in Figure 6, the epifluorescent analysis carried out on conjunctival ICs showed the presence of α-SMA immunoreactive cells (green) over a monolayer of conjunctival epithelial cells identified by nuclear DNA intercalant dye (blue). The immunoreactivity appeared more intense in chronic stages in comparison to early ones (Figs. 6a, 6c vs. Figs. 6b, 6d); indeed, a more defined epithelial–mesenchymal transition (EMT) shape morphology appears in late stages characterized by reduced nuclear expression (Figs. 6b, 6d). 

Figure 6.

 

α-SMA expression in conjunctival ICs from early and chronic stages disease and indirect immunofluorescence specific for α-SMA contractile protein. Representative epifluorescent acquisitions for α-SMA (cy2, green) in early (a, c) and chronic (b, d) OcMMP stages (see Materials and Methods for details). The monolayers were identified by nuclear staining with DAPI (blue). Note the decreased cellularity in (b) and (d) and the increased immunostaining indicative of the presence of contractile α-SMA (indicator of EMT). Images are the result of direct acquisition with no further modifications other that the subtraction of the isotype fluorescence intensity (magnification: 400×; bar in picture).

Discussion

OcMMP is a progressive, self-maintained fibrotic process. Such a process of excessive tissue remodeling leads to a complete disruption of the ocular surface morphofunctional homeostasis and is partially sustained by epithelial activity. Our study revealed the bimodal fibrotic evolution of chronic cicatrizing conjunctivitis, such as OcMMP. First, as a result of the prolonged inflammation, we found high levels of ICAM-1, IL-10, and IL-17 expression, as well as moderate α-SMA and TGF-β expression. Expression of ICAM-1, IL-10, and IL-17 remained independent of inflammatory stimuli, and there was a significant increase in profibrotic marker expression (α-SMA and TGF-β), in addition to the appearance of aggressive and progressive clinical symptoms of scarring. As disease severity and fibrosis increased, IL-10 levels declined. Even though IL-10 has been linked to profibrotic processes in several contexts,^31,32 its exact profibrotic role warrants further investigation. 

During the different disease stages, an initial cell-mediated inflammatory response is present, which is the trigger for a subsequent worsening fibrosis.^30,33,34 We found a mild inflammation (secondary to ocular surface system homeostatic failure) that is related to scarring and tissue disruption in the later stages. In fact, although immunomodulating treatment may slow down the disease severity, the cicatrizing evolution is not stopped at all, suggesting a relative triggering role of cell-mediated inflammation in such disease and explaining the limitations of our diagnostic tests in properly grading or detecting the disease. 

In conjunctival biopsies, stromal α-SMA expression always directly increase with disease stages, although many patients with OcMMP test negative for DIF analysis. Furthermore, an increase in profibrotic markers (α-SMA and TGF-β) is better measured in IC samples, which show that the aberrant tissue remodeling in OcMMP is also regulated by epithelial layer changes. It has been suggested that the EMT reveals a hidden mechanism of disease progression that is at least partially disengaged by inflammatory autoimmune responses. This assumption states that conjunctival and corneal epithelial cells, when chronically subjected to high degrees of inflammation, as in the case of OcMMP, may transdifferentiate into myofibroblast-like phenotypes. This hypothesis, based on our biochemical and histochemical results, is illustrated in Figure 7 and finds further support in our transcriptomic analysis on α-SMA expression and related protein immunolocalization on ICs (Fig. 6). This finding has been proved in other conditions in which it seems to be involved (e.g., pulmonary, cardiac, and hepatic fibrosis). Transdifferentiated cells begin to produce extracellular matrix, characterized by tissues with exuberant fibrosis.^35,36 Our results suggest that the fibrotic process progressively replaces the inflammation state, explaining the absence of characteristic DIF findings (linear Ig and complement deposits along the epithelial basement membrane zone)^30,37 in the conjunctival biopsies of patients belonging to the OcMMP group. 

Figure 7.

 

EMT model in CCC, such as OcMMP, consistent with our histological and biochemical findings. Early stages of the disease (left) are characterized by a cell-mediated inflammation with high expression of ICAM-1, IL-10, and IL-17. At this stage, the epithelium is almost unchanged, with a reduced number of goblet cells. As the disease progresses (right), the epithelium undergoes transdifferentiation into myofibroblast-like cells (EMT), characterized by marked expression of SMA and TGF, with a reduction in ICAM-1, IL-10, and IL-17 expression. In addition to marked fibrosis and loss of normal tissue structure, complete loss of goblet cells and changes of the inflammatory infiltrate, along with vessel ectasia, are observed.

The strict correlation between α-SMA and TGF-β epithelial expression from IC samples, together with the clinical picture, suggests that these molecules could serve as diagnostic, prognostic, and therapeutic biomarkers of OcMMP and, more widely, of CCC. Hence, the ocular surface functional indices worsen in direct proportion with the aggravating scarring condition. Particularly, Schirmer test type I assessments, TBUT, and BCVA irreversibly and progressively become impaired, starting from early stages and underlying the pivotal role of the scarring process in the partial and reversible loss of function related to a chronic, clinically manifest auto-inflammatory response. 

In OcMMP, clinical findings associated with the detection of autoantibodies by DIF and histological analyses have been critical for diagnosis.³⁷ However, in recent years, it has been observed that DIF sensitivity in patients with MMP is highly variable (ranging from 20%–80%),^38–41 and a high proportion of ocular-only MMP patients (up to 50%) have no or, in some cases, intermittent positivity at biopsy.^42,43 These patients with DIF-negative MMP meet the clinical criteria for OcMMP but have negative biopsy results. These negative or inconclusive findings prevent the access of such patients to effective treatments.³⁹ Furthermore, the invasive nature of conjunctival biopsy carries with it the risk of complications, such as an excessive inflammatory response, resulting in exacerbation of the disease and worsening of the clinical picture. On the other hand, conjunctival impression cytology is a non-invasive and relatively painless procedure that can be performed, unlike conjunctival biopsy, directly in an outpatient setting. Epithelial cell changes, as in our case, may be more readily detected by ICs (Fig. 6), which excludes any confounding factors present in the deeper layers of the conjunctiva, such as excessive fibrosis. We demonstrated the non-inferiority of conjunctival impression cytology compared to conjunctival biopsy, although with ICs it is not possible to obtain the stroma, thus preventing a direct evaluation of fibrosis. Therefore, a non-invasive, simple, and repeatable IC obtained by harvesting epithelial cells may be an accurate diagnostic and prognostic method for assessing OcMMP grading and progression. 

The bimodal (first inflammatory and then fibrotic) evolution of OcMMP is also supported by a lack of worsening symptoms (based on the OSDI) within disease progression, as well as by the early corneal involvement, central corneal ulcers, or persistent epithelial defects, which is mostly not related to the inflammatory grade nor extremely responsive to immunomodulating drugs but is directly correlated to disease progression. However, although some aspects of the pathogenesis of OcMMP are understood, the precise mechanisms underlying its characteristic massive fibrosis remain unknown. Fibrotic changes in chronic inflammatory disease may evolve at least partially as an autonomous pathway from the inflammation, explaining the significant limitations in diagnosis and treatment that we still have in managing such diseases. 

This study is restricted by a relatively small sample size, but, because OcMMP is an extremely rare disease, the opportunity to obtain a substantial patient population is quite limited.^44,45 

Acknowledgments

GE and AM thank Fondazione Roma for continuous support. The study was partially supported by the Ministry of Health (Ricerca Corrente, RC 2779945; IRCCS-Fondazione Bietti). 

Disclosure: A. Di Zazzo, None; F. Cutrupi, None; M.G. De Antoniis, None; M. Ricci, None; G. Esposito, None; M. Antonini, None; M. Coassin, None; A. Micera, None; E. Perrella, None; S. Bonini, None