Expression of MMPs and TIMPs in Human Pterygia and Cultured Pterygium Epithelial Cells

Abstract

purpose. Pterygia are a common, benign, fibrovascular, and infiltrative process of the corneal–conjunctival junction of unknown pathogenesis. Matrix metalloproteinases (MMPs) are a family of proteolytic enzymes active against all components of the extracellular matrix, whose activity is specifically neutralized by tissue inhibitors of MMPs (TIMPs). In the current study the hypothesis was that MMPs and TIMPs may actively participate in the formation and progression of pterygia.

methods. In this study, 25 pterygium specimens and 15 normal conjunctival biopsies obtained from subjects undergoing surgery for glaucoma and cataract, were processed for immunohistochemistry or in situ hybridization. Pterygium epithelial cells (PECs) were cultured under serum-free conditions and exposed to proinflammatory cytokines to determine both the mRNA and protein expression profiles of MMPs and TIMPs.

results. Collagenase-1 and gelatinase A were expressed in all pterygia examined, specifically localized to the epithelium (directly adjacent to collagen type III), with gelatinase B expression exclusively associated with neutrophils. No collagenase-1 or gelatinase A was detected in normal conjunctiva. TIMP-1 and -3 were localized to epithelial cells with additional TIMP-3 immunoreactivity detected in the extracellular matrix, endothelial cells and leukocytes of all diseased tissue. TIMP-3 protein was evident in 4 of 15 normal conjunctiva. Induction of collagenase-1, gelatinase A, and TIMP-1 mRNA and protein was demonstrated in epithelial cells treated with tumor necrosis factor-α and interleukin-1α, whereas TIMP-3 expression was unaltered.

conclusions. This is the first study to document the cellular expression of MMPs and TIMPs in pterygia and cultured human PECs. MMPs and TIMPs may contribute to the inflammation, tissue remodeling, and angiogenesis that characterize pterygia. Understanding the role these proteins play may lead to novel therapies intended to reduce the progressive nature of pterygia.

Pterygia belong to the family of sunlight-related eye diseases known as ophthalmohelioses. ¹ They are a common, degenerative and fibrovascular process thought to originate at the corneal–conjunctival junction, where it is proposed that altered limbal stem cells migrate centripetally to encroach on the normal cornea. ² The origin of pterygia remains controversial, and the pathogenesis is unknown. Recently, we proposed an explanation for the shape and location of pterygia and suggested that UV-irradiation might be the key initial event in the pathogenesis of this disease. ^{3 4} Other investigators have postulated a role for angiogenic factors, ⁵ because the normally avascular cornea is invaded by new blood vessels. The presence of infiltrating T lymphocytes in pterygia is suggestive of a delayed-type hypersensitivity reaction, ⁶ whereas the increased presence of mast cells in pterygia ⁷ may indicate environmental irritants as causative factors. 

Recently, several cytokines such as tumor necrosis factor (TNF)-α, basic fibroblast growth factor, and transforming growth factor-β, have been localized to both resident and inflammatory cells in pterygia. ⁸ Such cytokines and growth factors have been shown to play a key role in inflammatory, fibrogenic, and angiogenic ⁹ processes, all of which are commonly observed in pterygia. 

Matrix metalloproteinases (MMPs) are a family of neutral proteolytic enzymes capable of denaturing most components of the extracellular matrix. ¹⁰ At least 17 members have been cloned and grouped according to their substrate specificity. These include the collagenases, capable of cleaving intact fibrillar collagen, and the gelatinases, which can further degrade these collagens, and basement membrane collagen type IV. The third group of MMPs comprises the stromelysins, which possess broad substrate specificity and can cleave fibronectin, laminin, and proteoglycans, and the membrane-associated MMPs with poorly defined substrate specificities, although some group members display collagenolytic activity ¹¹ and can activate other MMPs. ¹²  

MMPs are regulated at several levels, including the transcriptional level, where they are modulated by various cytokines and growth factors ¹³ ; at the level of posttranscriptional processing, where they require activation in the extracellular space by other proteases ^{12 14} ; and finally, at the level of inhibition, where they are regulated by their specific tissue inhibitors (TIMPs). To date, four members of this family have been cloned and characterized. Their principal function is inhibition of MMP activity; however, some studies have demonstrated potent growth factor–like activities toward several cell types, ¹⁵ including corneal epithelium. ¹⁶  

MMPs and TIMPs have been implicated in many physiological and pathologic processes, including embryologic development, ¹⁷ angiogenesis, ¹⁸ wound healing, ¹⁹ rheumatoid arthritis, ²⁰ and cancer, ²¹ and in several ocular disorders, including scleritis, ²² uveitis, ²³ and corneal disease. ²⁴ We hypothesize that MMP and TIMP molecules may be active participants in the extensive matrix turnover and infiltration that characterize pterygia. 

The purposes of this study were to determine the expression of MMPs and TIMPs in pterygia and characterize their cellular sources and to establish an in vitro model to determine the profile of secreted MMPs and TIMPs from cytokine-induced pterygium epithelial cells (PECs). 

Materials and Methods

Patients

Fresh surgical specimens of pterygia (n = 15) and normal conjunctiva from patients undergoing surgery for glaucoma or cataract (n = 15) were collected from Greenoaks Hospital, Sydney, Australia, and fixed in Carnoy’s solution. Informed consent was obtained from each subject. Formalin-fixed paraffin-embedded archival pterygium specimens (n = 10) were obtained from the Department of Anatomic Pathology, Prince of Wales Hospital, Sydney, Australia. All research protocols were conducted and patients were treated in accordance with the tenets of the Declaration of Helsinki. 

Preparation of RNA Probes and In Situ Hybridization

Plasmid cDNAs for human interstitial collagenase-1 (2.5 kbp) and TIMP-1 (0.9 kbp) were manipulated to generate digoxigenin-labeled sense and antisense collagenase-1 and TIMP-1 riboprobes, as previously described. ^{22 25 26} Briefly, a 530-bp fragment of human interstitial collagenase-1, and a 633-bp fragment of human TIMP-1 were subcloned into a transcription vector (Bluescript-II [SK]; Stratagene, La Jolla, CA). The new plasmid constructs were linearized within the multiple cloning site using appropriate restriction enzymes, phenol-chloroform extracted, ethanol precipitated and transcribed in vitro. 

Nonisotopic in situ hybridization was performed (Digoxigenin Nucleic Acid Detection Kit; Boehringer–Mannheim, Sydney, Australia), as previously described. ^{22 25 26} Briefly, 4-μm serial sections were cut, deparaffinized, rehydrated, and equilibrated in phosphate-buffered saline. Sections were subsequently deproteinized, then treated with 5 μg/ml proteinase K (Boehringer–Mannheim) for 30 minutes at 37°C. Tissue sections were postfixed in fresh 4% paraformaldehyde, acetylated, and equilibrated in standard saline citrate buffer. ²² Sections were prehybridized for 2 hours at 42°C, then hybridized overnight at the same temperature with 50 ng digoxigenin-labeled sense or antisense probe. After hybridization, sections were stringently washed and treated with 100 μg/ml RNase A(Boehringer–Mannheim). 

Immunohistochemical Analysis

Pterygia and normal conjunctiva were cut (2–4 μm thick) and processed for immunohistochemistry as previously described. ²⁶ Colon cancer and inflamed tonsil served as control tissue. Mouse and rabbit primary monoclonal or polyclonal antibodies (Abs; see Table 1 for source and dilution) were incubated on tissue sections overnight at 4°C, then extensively washed in 0.05 M Tris-buffered saline (pH 7.6), before the addition of biotinylated secondary Abs (goat anti-mouse for monoclonal Abs or goat anti-rabbit for polyclonal Abs). Sections were again washed and incubated with horseradish peroxidase–conjugated streptavidin (Dako, Carpinteria, CA), and the immunoreactivity was revealed by adding 3-amino-9-ethylcarbazole (Sigma, Sydney, Australia). Control reactions included incubating sections with an isotype Ab (see Table 1 ) or omitting the primary Ab. Sections were counterstained with hematoxylin. 

Cell Culture

Epithelial cell growth was established from explants of fresh pterygium tissue (1–2 mm²), by using a technique previously reported from our laboratory. ^{27 28} PECs were subsequently expanded in 75-cm² tissue culture flasks (Nunc, Roskilde, Denmark) in Eagle’s minimal essential medium (Trace Biosciences, Sydney, Australia) supplemented with 10% fetal bovine serum (Trace Biosciences) and 100 U/ml penicillin and 100μ g/ml streptomycin (Trace Biosciences). Flow cytometric analysis revealed cytokeratin immunoreactivity in 98% of these cells, suggesting these cells comprised a pure population of epithelial cells. ²⁸ All cell culture media and solutions were filtered to minimize endotoxin, as previously described. ^{22 26} For cytokine stimulation assays, PECs were treated as in our previous investigations. ^{22 26} Briefly, cells were counted and seeded at 0.5 × 10⁶ cells per flask. On reaching semiconfluence they were placed in serum-free medium (0.2% bovine serum albumin in Eagle’s minimal essential medium) for 24 hours. This medium was removed, and cells were washed again with phosphate-buffered saline and placed in fresh serum free media with or without of recombinant human TNF-α (50 ng/ml) and interleukin (IL)-1α (20 ng/ml; R&D Systems, Minneapolis, MN) or phorbol myristate acetate (10 ng/ml; Sigma). Supernatants and RNA were harvested at specified time points and stored in aliquots at −70°C until used. Some PECs were cultured in chamber slides (Nunc) in the presence of 3 μM monensin (Sigma) for 8 hours, extensively washed in phosphate-buffered saline, fixed in 100% methanol for 5 minutes, and processed immunohistochemically. 

In other cell culture studies, calf skin collagen type I (Sigma) was coated onto 60-cm² culture dishes for several hours at room temperature, under sterile conditions. Excess collagen solution was removed and the dishes dried overnight in a laminar flow hood. PECs were seeded onto coated or uncoated dishes and the supernatants removed after 2 to 3 days to determine the amount of collagenase-1 secreted by these cells. 

Extraction of RNA and Reverse Transcription–Polymerase Chain Reaction Analysis

Total RNA was extracted from PECs, as previously described, ²⁹ and electrophoresed through 0.8% agarose, 2.2 M formaldehyde gels, ³⁰ to visualize RNA integrity. RNA samples were stored frozen at −70°C in aliquots until used in reverse transcription–polymerase chain reaction (RT-PCR). RT was performed according to the manufacturer’s instructions, using a cDNA synthesis kit (Preamplification System for First-Strand cDNA; Gibco, Gaithersburg, MD). Aliquots (1 μl) of cDNA were amplified by PCR using 100 nM each of the forward and reverse gene-specific primers (see Table 2 ), using conditions similar to those previously described. ²⁶ Semiquantitative RT-PCR was established by terminating reactions at regular intervals of 10, 15, 20, 25, 30, and 35 cycles for each primer pair to ensure that the products formed were within the linear portion of the amplification curve. Briefly, a 2-minute hot start at 95°C was performed, followed by 20 cycles of PCR (each cycle: 95°C, 30 seconds; 55°C, 30 seconds; and 72°C, 30 seconds), and terminated with a 2-minute extension at 72°C. Enzyme digests were performed on the products, as previously described, ²⁶ to confirm the identity of the amplicons. Products were visualized on 1.2% agarose gels stained with ethidium bromide. 

Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis Gelatin Substrate Zymography and Reverse Zymography

Gelatin substrate zymography was performed as previously described. ²² For reverse zymography, 12% acrylamide resolving gels impregnated with both gelatin and rh-progelatinase A (Calbiochem, Sydney, Australia) were used as previously described. ³¹ Semiquantitative data were generated by scanning negative film exposures of zymograms using a densitometer (model GS-300; Hoefer, San Francisco, CA). 

Western Blot Analysis

Western blot analysis was performed as previously described, ^{22 26} by using a mouse anti-human collagenase-1 monoclonal Ab (ICN, Sydney, Australia). Membranes were placed in a chemiluminescent reagent (duPont, Sydney, Australia), then exposed to x-ray film. Prestained broad or low weight markers (Bio-Rad, Sydney, Australia) were run in adjacent lanes. 

Collagenase-1 Enzyme-Linked Immunosorbent Assay

Supernatants derived from PEC cultured in the presence or absence of interstitial collagen were analyzed by commercial enzyme-linked immunosorbent assay (ELISA; Biotrak; Amersham Pharmacia Biotech, Sydney, Australia) as described by the manufacturer. This assay does not cross-react with other MMPs and detects free and TIMP complexed collagenase-1. 

Statistical Analysis

Triplicate values obtained from the collagenase-1 ELISA were expressed as mean ± SD. and the t-test was used to determine the level of significance. 

Results

Localization of MMPs in Human Pterygia

Typical histologic features of pterygia are demonstrated in a low-power micrograph (Fig. 1C ) and include a layer of epithelial cells (e), regions of denatured matrix (dm) proteins (elastosis) represented by coiled collagen, with increased vasculature (arrowheads) and many intravascular leukocytes (arrows). Regions of collagen accumulation devoid of inflammatory cells with abundant resident fibroblast-like cells were also obvious (micrograph not shown). Initially, in situ hybridization was performed on formalin-fixed, paraffin-embedded tissue, and collagenase-1 mRNA transcripts were specifically localized to PECs (Fig. 1A) . No collagenase expression was found in any normal conjunctival tissue examined (data not shown). Similarly, sections hybridized with the sense riboprobe showed no signal (Fig. 1B) . Collagenase-1 (Fig. 1D) and gelatinase A (Fig. 1G) proteins were localized to PECs in all (15/15) specimens, whereas little or no collagenase-1 (Fig. 1E) and gelatinase A (micrograph not shown) were observed in normal conjunctiva. Although collagenase-1 was generally found throughout the entire epithelium, gelatinase A was predominantly localized to basal PECs (Fig. 1G) . No immunoreactive signal was noted in sections of normal (micrographs not shown) or diseased tissue incubated with an isotype control Ab (Fig. 1C) . After using a panel of monoclonal Abs to several human MMPs (see Table 1 ), the only other proteinase detected in pterygia was gelatinase B. This enzyme was present in most intra- and some extravascular neutrophils (Fig. 1F) , which were characterized both morphologically and histochemically using a neutrophil elastase Ab (micrographs not shown). In addition, it was noted that when intravascular neutrophils were in contact with the vessel lumen, a diffuse extracellular immunoreactive staining pattern was observed, both surrounding the neutrophil (Fig. 1F , arrowheads) and apparently associated with the luminal endothelial cells (Fig. 1F , arrow). Gelatinase B staining by marginating neutrophils was also occasionally observed (inset 1F, arrow). Extracellular matrix substrates for the MMPs were examined using a panel of monoclonal Abs to collagens (see Table 1 ). Collagen type III was by far the most abundant collagen in pterygia (Fig. 1H) , specifically localized adjacent to collagenase-1– and gelatinase A–producing PECs. Collagen types I, II, and IV were found in fibrous regions among resident fibroblasts and collagen type IV stained basement membranes (data not shown). 

TIMP-1 and -3 Are Expressed in Pterygia

TIMP-1 mRNA (data not shown) and protein expression closely resembled that of collagenase-1 in pterygia; it localized specifically to PECs (Fig. 2B ) in all pterygia. In contrast, no TIMP-1 protein was detected in normal conjunctiva (Fig. 2A) . TIMP-3 expression, although similar to TIMP-1, was predominantly expressed by basal epithelial cells, which exhibited more intense staining than suprabasal cells (Fig. 2C) . TIMP-3 was also expressed by perivascular inflammatory cells, (Fig. 2C , arrowheads), some of which were also present in the connective tissue matrix. Using a panel of cell-type specific Abs (see Table 1 ) on serial tissue sections, the majority of these cells displayed CD3 positivity (Fig. 2D) . Occasional intraepithelial T-cells were also noted (micrograph not shown). Specific immunoreactivity for this inhibitor was also found in connective tissue matrix (Fig. 2F ) and associated with endothelial cells, apparently in the basement membranes (micrograph not shown). TIMP-3 protein was demonstrated in 4 of 15 normal conjunctival specimens, whereas TIMP-1, -2, and -4 were not detected in any of these tissue samples. TIMP-2 was detected in tonsillar tissue, however (data not shown). 

Cultured Human PECs Produce MMPs and TIMPs

Previously, we have cultured and characterized epithelial cells derived from pterygia. ²⁸ These cells were therefore studied for their capacity to express MMPs and TIMPs. To our surprise, no collagenase-1 immunoreactivity was detected in these cells (Fig. 3A) . However, because these proteinases are synthesized and rapidly secreted, very little intracellular storage can be expected. For this reason, PECs were exposed to monensin (an Na⁺ ionophore that inhibits protein secretion). This treatment resulted in an accumulation and subsequent detection of cytoplasmic collagenase-1 (Fig. 3B) , gelatinase A (Fig. 3C) , stromelysin-1 (Fig. 3D) , and TIMP-1 (Fig. 3E) that would otherwise not have been detected. Similar immunoreactivity for TIMP-3 was observed in both control (Fig. 3F) and monensin-treated cultures (micrograph not shown). 

In some experiments, PEC were cultured on a dry film of collagen and the supernatants analyzed by ELISA to determine whether cell-to-matrix interactions modulated the level of collagenase-1. Supernatants from cells cultured on tissue culture plastic alone or on interstitial collagen contained 25.42 ± 1.12 and 36.03 ± 1.53 ng/ml collagenase-1, respectively. This was shown to be statistically significant (P < 0.001). 

Proinflammatory Cytokines Modify the Expression of MMPs and TIMPs

PECs were cultured in the presence or absence of the proinflammatory cytokines, which previously have been localized in pterygia. ⁸ The optimal concentration and kinetics of exposure to these cytokines for MMP induction in several cell lines have been established in our laboratory. ^{22 26} Semiquantitative RT-PCR analysis on total RNA extracted from PECs after cytokine treatment demonstrated not only the constitutive expression of MMPs and TIMPs (Fig. 4A 4B 4C 4D , lane 2), but also the induction of collagenase-1 (Fig. 4A , lane 3), gelatinase A (Fig. 4B , lane 3) and the expression of TIMP-1 (Fig. 4C , lane 3) mRNA. In contrast, there was no significant modulation of TIMP-3 mRNA expression between control- (Fig. 4D , lane 2) and cytokine-stimulated PECs (Fig. 4D , lane 3). In addition, exposure to phorbol myristate acetate resulted in a potent induction of collagenase-1 (Fig. 4A , lane 4) and TIMP-1 (Fig. 4C , lane 4). However, gelatinase A (Fig. 4B , lane 4) was marginally induced, and no apparent modulation of TIMP-3 mRNA (Fig. 4D , lane 4) was demonstrated by this treatment. 

Culture media derived from PECs stimulated with proinflammatory cytokines over a 72-hour period were harvested and biochemically analyzed. The results showed the constitutive expression of collagenase-1 over the time course (Fig. 5A , lanes 1–4) by unstimulated epithelial cells and a 1.7-fold induction of this proteinase by cytokine treatment during the same period (Figs. 5A , lanes 5–8, 5D). Similarly, gelatinase A was increased approximately sixfold (Figs. 5B 5D) and TIMP-1 fourfold (Figs. 5C 5D) . In contrast, TIMP-3 expression was not modulated between control- (Fig. 5C , lanes 1–4) and cytokine-treated PECs (Fig. 5C , lanes 5–8; 5D). The absence of an inhibitor band at approximately 22-kDa indicated the absence of TIMP-2 in the supernatants. 

Discussion

To the best of our knowledge this is the first study to localize and identify the cellular source of several MMPs and TIMPs in human pterygia. In addition, corroborating in vitro data were obtained, whereby MMP and TIMP expression was identified in cultured pterygium-derived epithelial cells in both a control and inflammatory setting. In a model proposed by Dushku and Reid, ³² it was suggested that fibroblast-derived collagenase could be actively involved in the formation of pterygia; however, no attempt was made to localize this enzyme in diseased tissue. Only one other study has examined the possible role of MMPs in pterygia. In that study, total protein was extracted from pterygia and increased expression of gelatinase B was demonstrated in contrast to that in normal conjunctiva. ³³ Unlike the present study, the cellular source and role of other MMPs and TIMPs was not evaluated. 

An important aspect of the present study was the identification of potential MMP substrates in the pterygium connective tissue. Although collagenase-1 has been reported to possess the highest specific activity against type III collagen, neutrophil collagenase (collagenase-2) and collagenase-3 have been shown to preferentially digest collagen types I and III, respectively. ¹¹ However, MT1-MMP ¹¹ and gelatinase A ³⁴ are also active against native fibrillar collagen. Therefore, it is tempting to speculate that the in situ production of collagenase-1 and gelatinase A by epithelial cells may be partially responsible for the extensive collagen degradation in pterygia. Although collagens make up a large component of the pterygium extracellular matrix, the importance of other matrix molecules such as proteoglycans and glycoproteins (which are also denatured by MMPs) should not be dismissed or underestimated. 

The directed growth of pterygia (predominantly from nasal to corneal) is an interesting, yet poorly understood process. Recently, Pilcher et al. ³⁵ demonstrated that collagenase-1 activity was required by keratinocytes to migrate on a matrix consisting of collagen type I. Other studies have established that gelatinase A and B are expressed by migrating and regenerating epithelial cells during corneal wound healing in a rat model of keratectomy. ³⁶ Directed cell migration has been previously described in vitro, wherein collagenase-1, gelatinases, and the collagen fragments generated by these proteinases have been reported to show potent chemotactic and chemokinetic activities on tumor cells. ³⁷ The results of the present study suggest that cell-to-matrix interactions may be an alternative pathway by which collagenase-1 is induced and may facilitate the directed cell migration over the collagen matrix, promoting pterygium invasion into the cornea. Similarly, cell-to-cell contact induced gelatinase B on vascular endothelial cells (Fig. 1F , arrow). In this study, gelatinase B was specifically localized to intravascular neutrophils. We and others have proposed that gelatinase B may be required for T-cell ^{22 26 38} and neutrophil ³⁹ migration and extravasation (Fig. 1F , inset), because it possesses proteolytic activity against basement membrane collagens. 

Previous investigations have shown that the expression of TIMP-3 by uterine cells protects the uterine wall from invasive cytotrophoblasts, which themselves produce high levels of gelatinases. ⁴⁰ Similarly, the localization of TIMP-1 and -3 in pterygium tissue suggests a potential protective role against MMP activity. Furthermore, the absence of TIMPs in normal conjunctival tissue suggests that these gene products may be associated with the pathologic process in pterygia. The only exception was TIMP-3, which was detected in a small proportion of normal conjunctiva. This expression could result from the effects of UV exposure in control subjects. Alternatively, a similar pattern of staining for TIMP-3 has been demonstrated in pathologic and normal human corneas ²⁴ and in wounded and normal mouse corneas. ⁴¹ Similarly, no modulation of TIMP-3 expression has been demonstrated in cultured human epidermal keratinocytes treated with TNF, IL-1, and phorbol myristate acetate, ⁴² results that corroborate the findings of the present study (see Figs. 4 and 5 ). These observations are also consistent with recent reports on the regulation of human TIMP-3, which indicate novel regulation mechanisms for TIMP-3 in comparison with that of TIMP-1 and -2. ⁴³ The immunodetection of TIMP-3 in epithelial and endothelial cells and connective tissue in pterygia is in accordance with other studies. ^{24 41 43 44 45} However, its expression by inflammatory cells is less well characterized. To our knowledge this is the first demonstration of the expression of TIMP-3 by T cells and suggests an additional role for this protease inhibitor in inflammation. In addition to their function as specific MMP inhibitors, TIMP-1 and -2 have been shown to increase the proliferative rate of cultured rabbit corneal epithelial cells. ¹⁶ In this respect, the abundant expression of both TIMP-1 and -3 in pterygia could provide growth factor–like activity in pterygia. 

Factors that induce collagenase-1 include cytokines and growth factors. ¹³ In addition, UV-A has been shown to increase collagenase-1 expression as much as 10 times and TIMP-1 expression 2 times in cultured keratinocytes. ⁴⁶ Similarly, Scharffetter et al. ⁴⁷ reported a dose-dependent induction of collagenase-1 after short-term exposure to UV-A irradiated dermal fibroblasts, whereas the expression of collagen type I was unaffected. They suggested this as a possible mechanism of actinic damage. The same group presented evidence that UV-A induction of collagenase-1 was mediated by IL-1 and IL-6. ⁴⁸ More recently, UV irradiation has been shown to induce the production of IL-1, -6, and -8 and TNF-α in cultured human corneal stromal cells and whole human corneas, ⁴⁹ a mechanism that may be relevant to the pathogenesis of pterygia. 

MMPs are synthesized and rapidly secreted, making intracellular detection often difficult. In the present study, we chose to treat PECs with monensin. This agent is an Na⁺ ionophore often used in cell biology to retard or block the passage of secretory proteins (such as proteases) through the Golgi apparatus. ⁵⁰ Previously, we ²⁶ and others ⁵¹ have used this agent to detect cytoplasmic MMPs in various cells. There is, however, one study that suggests that monensin can directly enhance the expression of MMPs. ⁵² The investigators observed that monensin induces MT-1 MMP protein and mRNA, but does not modulate the levels of gelatinase A or TIMP-2. 

The establishment of pure cultures of PECs provides an in vitro model to help elucidate the pathogenesis of pterygia. Our in vivo and in vitro data provide strong evidence to implicate MMPs, TIMPs, and their regulatory molecules (proinflammatory cytokines) in the degenerative and progressive nature of this disease. Understanding the mechanisms involved in the formation of pterygia is essential to providing a better means for prevention and treatment. Treating these patients with potent inhibitors of MMPs and angiogenesis ⁵³ may halt the progressive nature of pterygia and provide an alternative and more efficient form of therapy. 

 Supported by the Ophthalmic Research Institute of Australia (ORIA).

 Submitted for publication May 14, 1999; revised August 16, 1999; accepted September 19, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Antibodies Used for Immunohistochemical Analyses

Antibody	Source	DF
Collagenase-1	ICN	1:100
Stromelysin-1	ICN	1:100
Gelatinase A	ICN	1:100
Gelatinase B	ICN	1:200
TIMP-1	ICN	1:100
TIMP-2	ICN	1:80
TIMP-3	ICN	1:300
TIMP-4	Chemicon	1:50
Cytokeratin	Dako	1:400
vWF	Dako	1:50
Neutrophil elastase	Dako	1:100
CD3	Dako	1:50
CD20	Dako	1:30
CD68	Dako	1:50
IgG	Dako	1:50
Tryptase	Dako	1:100
Collagen type I	AMRAD	1:100
Collagen type II	AMRAD	1:100
Collagen type III	AMRAD	1:100
Collagen type IV	AMRAD	1:100
Mouse IgG1	Dako	1:100
Rabbit IgG	Dako	1:100

 All antibodies used are specific to human antigens as specified by the manufacturers and are mouse monoclonal antibodies with the exception of CD3 and TIMP-4, which are rabbit polyclonal antibodies. DF, dilution factor; vWF, von Willibrand factor. AMRAD, Dako, ICN Biomedicals (Sydney, Australia), R&D Systems (Minneapolis, MN), and Chemicon, (Temecula, CA).

Table 2.

 

Primer Pairs Used for PCR Analysis

Collagenase-1 21	F	5′-GGT GAT GAA GCA GCC CAG-3′
	R	5′-CAG TAG AAT GGG AGA GTC-3′
Gelatinase A 21	F	5′-CCA CGT GAC AAG CCC ATG GGG CCC C-3′
	R	5′-GCA GCC TAG CCA GTC GGA TTT GAT G-3′
TIMP-1 26	F	5′-TGC ACC TGT GTC CCA CCC CAC CCA CAG ACG-3′
	R	5′-GGC TAT CTG GGA CCG CAG GGA CTG CCA GGT-3′
TIMP-3 24	F	5′-CCA TCA AGC AGA TGA AGA TGT ACC-3′
	R	5′-GGT AGT AGC AGG ACT TGA TCT TGC-3′
GAPDH 17	F	5′-TGA TGA CAT CAA GAA GGT GGT GAA G-3′
	R	5′-TCC TTG GAG GCC ATG TGG GCC AT-3′

 Previous investigators (studies referenced) have successfully used these primer sets to amplify the respective gene products. GAPDH, Glyceraldehyde 3 phosphate dehydrogenase; F, forward primer; R, reverse primer.

Figure 1.

 

Expression of MMPs and collagens in pterygia. Pterygium tissue (A through D, F, G, and H) and normal conjunctiva (E) were analyzed by in situ hybridization (A, B) or by immunohistochemistry (C through H). For in situ hybridization, tissue sections were counterstained with neutral red. Probe hybridization is denoted by blue-purple cytoplasmic staining. Sections incubated with a digoxigenin-labeled antisense riboprobe to human collagenase-1 resulted in specific cytoplasmic signal in epithelial cells (A). Serial sections hybridized with the corresponding sense probe demonstrated no hybridization (B). Similar in situ hybridization results were obtained with tissue incubated with TIMP-1 sense and antisense riboprobes (micrographs not shown). For immunohistochemistry, all tissue sections were counterstained with hematoxylin. Immunoreactivity is denoted by the red cytoplasmic signal. Collagenase-1 (D) and gelatinase A (G) proteins were detected in epithelial cells of pterygia, but no collagenase-1 reactivity was found in normal conjunctiva (E). Immunoreactivity for gelatinase B was detected in most intravascular neutrophils (F). Collagen type III was the predominant interstitial collagen expressed in pterygia (H). Sections where the primary Ab was omitted (micrographs not shown) or that were incubated with an isotype control monoclonal Ab (C) had no immunoreactivity. Similar results were obtained with all other pterygia and normal conjunctival tissue examined. Arrowheads: intravascular leukocytes; arrows: blood vessels. Original magnification, (A, B, and D through H)× 640; (C) ×160; (inset, F)× 1000. Epithelium (e), denatured matrix (dm).

Figure 2.

 

Expression of TIMP-1, and -3 in pterygia. Pterygium tissue (B through F) and normal conjunctiva (A) were analyzed by immunohistochemistry to determine the expression of TIMP-1 (A, B) and TIMP-3 (C, F). To characterize TIMP-3–expressing cells, a panel of Abs (see Table 1 ) including; an anti-CD3 Ab (D) and an anti-cytokeratin Ab (E) were used on serial tissue sections. TIMP-3 immunoreactivity was also displayed on connective tissue (F). When the primary Ab was omitted (data not shown) or when an isotype control Ab was used (inset, D), no immunoreactive signal was observed. All tissue sections were counterstained with hematoxylin. These data are representative of all pterygia examined. Original magnification,× 640.

Figure 3.

 

Immunolocalization of MMPs and TIMPs to cultured PECs. Human PECs were cultured in the presence (B through E) or absence (A, F) of monensin. In this set of experiments PECs were cultured in chamber slides, fixed in methanol and stained for collagenase-1 (A, B), gelatinase A (C), stromelysin-1 (D), TIMP-1 (E), and TIMP-3 (F). When fixed cells were incubated with an isotype control monoclonal Ab, no signal was observed (inset, F). All cells were counterstained with hematoxylin. These results are representative of four separate experiments. Original magnification, ×400.

Figure 4.

 

RT-PCR analysis for MMP and TIMP mRNAs in cytokine-stimulated PECs. Equal amounts of total RNA were reverse transcribed from unstimulated (lane 2), TNF-α + IL-1α–stimulated (lanes 3, 6, 7), phorbol myristate acetate–treated (lane 4) PECs and TNF-α + IL-1α–stimulated human scleral fibroblasts (lane 5) after 48 hours of cytokine exposure. When no reverse transcriptase enzyme and no gene specific primers were included, no PCR product formed (lanes 6, 7, respectively). Otherwise, PCR products at the expected size were amplified for collagenase-1 (A), gelatinase A (B), TIMP-1 (C), TIMP-3 (D), and GAPDH (E). A 100-bp ladder (Gibco) was run in parallel (lane 1). The same results were obtained in at least three separate experiments.

Figure 5.

 

Modulation of MMPs and TIMP-1 production by proinflammatory cytokines. Supernatants from unstimulated (for 0, 24, 48, and 72 hours, lanes 1, 2, 3, 4, respectively) and TNF-α + IL-1α–stimulated (for 0, 24, 48, and 72 hours, lanes 5, 6, 7, 8, respectively) PECs were analyzed by western blot analysis for collagenase-1 (A), by gelatin substrate zymography for gelatinolytic activity (B), and by reverse zymography for the detection of TIMPs (C). Note that only TIMP-3 was unresponsive to cytokine stimulation, denoted by the similar intensity of the inhibitor band migrating at approximately 24 kDa. These biochemical analyses are representative of three independent experiments. Semiquantitative data representing MMP and TIMP induction were generated by scanning densitometry (D).

The authors thank Kakesk K. Kumar, School of Pathology, University of New South Wales, for assistance in establishing epithelial cell cultures.