Regulated Expression of Collagenases MMP-1, -8, and -13 and Stromelysins MMP-3, -10, and -11 by Human Corneal Epithelial Cells

Abstract

purpose. This study investigated the regulated expression of collagenases (MMP-1, -8, and -13) and stromelysins (MMP-3, -10, and -11) by human corneal epithelial cells treated with IL-1β, TNF-α, and doxycycline, a medication used to treat ocular surface diseases.

methods. Primary human corneal epithelial cell cultures were treated with IL-1β or TNF-α, with or without their corresponding inhibitors. Total RNA extracted from cells treated for 4 to 24 hours was subjected to semiquantitative RT-PCR and Northern hybridization. Conditioned media from 24-hour–treated cultures were evaluated for MMP production by ELISA and activity assays.

results. Semiquantitative RT-PCR and Northern hybridization revealed that the mRNAs of MMP-1, -13, -3, -10, and -11 were dose dependently upregulated by IL-1β and TNF-α, whereas MMP-8 and -14 and tissue inhibitor of metalloproteinase (TIMP)-1 were not altered, in corneal epithelial cells. MMP ELISA and activity assays confirmed this dose-dependent increase in MMP-1, -13, -3, and -10 protein production in conditioned media by IL-1β and TNF-α. This stimulated production was inhibited by their neutralizing antibodies and by IL-1 receptor antagonist. Doxycycline suppressed stimulated MMP-1, -10, and -13 production at both the mRNA and protein levels.

conclusions. This study demonstrated that IL-1β and TNF-α upregulate collagenases (MMP-1, -13) and stromelysins (MMP-3, -10, and -11) in human corneal epithelial cells. Doxycycline suppresses stimulated MMP-1, -13, and -10 at the mRNA and protein levels, which suggests that collagenases and stromelysins may play a role in the pathogenesis of sterile corneal ulceration and other ocular surface diseases.

Matrix metalloproteinases (MMPs) are a family of zinc-dependent extracellular endoproteinases that are collectively capable of degrading essentially all components of the extracellular matrix and basement membrane. They are synthesized and secreted by multiple cell types, including corneal epithelial cells and fibroblasts. These MMPs are secreted as inactive zymogens (pro-MMPs) requiring extracellular activation, and their activity is tightly regulated by specific tissue inhibitors of MMPs (TIMPs). There is much evidence to suggest that these enzymes play a vital role in several physiologic and pathologic processes, including embryonic development, tissue morphogenesis, cancer, angiogenesis, wound healing, and inflammatory diseases. To date, at least 22 different human MMPs have been identified, and they can be classified into five groups on the basis of substrate specificity: gelatinases, collagenases, stromelysins, membrane-type MMPs, and others. ^{1 2 3 4}  

The corneal epithelium is exposed to the numerous cytokines and growth factors that are present in the tear fluid and that are produced by corneal stromal fibroblasts. Tear fluid growth factors and cytokines are secreted by the lacrimal glands and are produced by the epithelial and inflammatory cells that reside on the ocular surface. As tear clearance decreases in dry eye conditions, the concentrations of proinflammatory cytokines, such as interleukin (IL)-1 and TNF-α, increase. ^{5 6 7} Tear film disorders are accompanied by MMP-mediated corneal disease, such as sterile corneal ulceration ⁸ and recurrent epithelial erosion. ⁹ The effects of proinflammatory cytokines on the regulation of MMPs have not been thoroughly investigated, especially in the ocular surface. 

Many studies have been focused on gelatinases, gelatinase A (MMP-2) and gelatinase B (MMP-9), one of the main MMP groups. Gelatinases degrade collagen types IV, V, VII, and X; elastin; and denatured collagens. They are the primary matrix-degrading enzymes produced by the corneal epithelium and fibroblasts. ^{10 11} MMP-9 has been found to be of central importance in catalyzing the cleavage of epithelial basement membrane components. ¹² They participate in extracellular matrix remodeling after wounding of the corneal surface and have been implicated in the pathogenesis of sterile corneal ulceration, ¹³ dry eye, ⁵ and other ocular diseases. We have evaluated the regulation of gelatinases by a number of cytokines and growth factors that the cornea is exposed to and have demonstrated that the inflammatory cytokines IL-1β and TNF-α upregulate MMP-9 mRNA, protein, and enzymatic activity in human corneal epithelial cells. ¹⁴  

Collagenases and stromelysins are the other two main MMP groups. Collagenases, including collagenase-1 (MMP-1, interstitial collagenase), collagenase-2 (MMP-8, neutrophil collagenase), and collagenase-3 (MMP-13), are the principle neutral proteinases capable of degrading native fibrillar collagens, which are the most abundant structural components of the human connective tissues. They all cleave type I, II, and III collagens at a specific site generating three-fourths N-terminal and one-fourth C-terminal fragments, which rapidly denature at physiological temperatures and become susceptible to degradation by other MMPs, such as gelatinases. ³ MMP-13 also cleaves type IV, X, and XIV collagens; large tenascin C; fibronectin; and aggrecan core protein and displays more than 40 times stronger gelatinase activity than MMP-1 and -8. ¹⁵ Stromelysins, including stromelysin-1 (MMP-3) and -2 (MMP-10), degrade proteoglycan core proteins, laminin, fibronectin, elastin, gelatin, and collagen types III, IV, V, VII, and IX, ^{2 4} whereas stromelysin-3 (MMP-11) is unusual and does not degrade any of the major extracellular matrix components. Collagenases and stromelysins have been found to participate in tumor invasion, vascularization, wound healing, and inflammatory diseases. On the ocular surface, MMP-13 mRNA has been detected in epithelial cells of wounded rat corneas, but not in normal control corneas. ¹⁶ MMP-10 is overexpressed in the diabetic corneal epithelium. ¹⁷ We have reported overexpression of MMP-1 and -3 in pterygium head ¹⁸ and conjunctivochalasis fibroblasts. ¹⁹ We also have observed induction of MMP-1, -13, -3, and -10 by experimentally induced corneal neovascularization in rabbits (Huang AJW, Li DQ, Shang TY, Dursun D, ARVO Abstract 661, 2001). Although the role of collagenases and stromelysins in the pathogenesis of corneal diseases, including corneal ulceration, vascularization and dry eye, has not been established, these findings led us to hypothesize that inflammatory cytokines, such as IL-1β and TNF-α, may stimulate the production of collagenases and stromelysins, in addition to gelatinase B, ¹⁴ by the human corneal epithelium. Confirming this hypothesis would support the use of MMP inhibitors in the treatment of MMP-mediated corneal diseases. Doxycycline is one such factor. This tetracycline antibiotic is well recognized for its therapeutic efficacy in treating ocular surface disease, such as rosacea and sterile corneal ulceration. ^{20 21} Doxycycline has been found to decrease the production and activity of IL-1β ²² and MMP-9 ^{14 23} in the human corneal epithelium. In the present study, we investigated the regulated production of collagenases (MMP-1, -8, and -13) and stromelysins (MMP-3, -10, and -11) by cultured human corneal epithelial cells. Our findings indicate that the corneal epithelium produces collagenases and stromelysins, and their production is regulated by the proinflammatory cytokines IL-1β and TNF-α and by doxycycline. 

Materials and Methods

Materials

Cytokines IL-1β and TNF-α, an IL-1 receptor antagonist (IL-1RA), neutralizing monoclonal antibodies against IL-1β or TNF-α, and ELISA kits for human MMP-1, -13, -3, and -10 were purchased from R&D Systems (Minneapolis, MN). Activity assay kits (Biotrak) for MMP-1 and -13 were from Amersham Pharmacia Biotech, Inc. (Piscataway, NJ). Nitrocellulose membranes were from Schleicher & Schuell (Keene, NH). The RNA-PCR kit (GeneAmp) was from Applied Biosystems (Foster City, CA). PCR purification and gel extraction kits (QiaQuick) were from Qiagen, Inc. (Valencia, CA). [α-³²P]-dCTP was from Du Pont NEN (Boston, MA). Film (XAR-5 and BioMax MS-1) and intensifying screens were from Eastman Kodak (Rochester, NY). DMEM, fetal bovine serum (FBS), HEPES buffer, Ham’s F12, amphotericin-B, phenol, DNA size marker, and random primers DNA labeling kit were from Gibco-BRL (Grand Island, NY). Dispase II was from Roche Applied Science (Indianaplis, IN). All plastic ware was from BD Biosciences (Lincoln Park, NJ). Cholera toxin subunit A, hydrocortisone, doxycycline, and all other reagents were from Sigma-Aldrich (St. Louis, MO). 

Primary Cultures of Human Corneal Epithelial Cells

Human corneal epithelial cells were cultured from explants taken from human donor corneoscleral rims, provided by the Florida Lions Eye Bank, using a previously described method. ^{14 23} In brief, each corneoscleral rim was trimmed, the endothelial layer and iris remnants were removed, and the tissue was treated with Dispase II for 15 minutes. Each rim was dissected into 12 equal segments. One corneal segment from one donor and one segment from a second donor were applied in six-well cell culture plates and covered with a drop of fetal bovine serum overnight. The explants were cultured in hormone supplemented medium, containing equal amounts of DMEM and Ham’s F12 medium, supplemented with 5% FBS, 0.5% dimethyl sulfoxide (DMSO), 2 ng/mL epidermal growth factor (EGF), 5 μg/mL insulin, 5 μg/mL transferrin, 5 ng/mL selenium, 0.5 μg/mL hydrocortisone, 30 ng/mL cholera toxin A, 50 μg/mL gentamicin, and 1.25 μg/mL amphotericin B. at 37°C under 95% humidity and 5% CO₂. The medium was renewed every 2 days. The epithelial phenotype of these cultures was confirmed by characteristic morphology and immunofluorescent staining with cytokeratin antibodies (AE-1/AE-3). Subconfluent cultures (12–14 days) were washed four times with PBS and switched to serum-free medium (the same medium as just described, but without FBS) for 24 hours before treatment. 

Cell Treatment

For MMP gene expression analysis, the corneal epithelial cells were treated with different concentrations (0.1, 1.0, or 10.0 ng/mL) of IL-1β or TNF-α, with or without addition of 10 μg/mL doxycycline for 4 to 24 hours in serum-free medium. For MMP ELISA and activity assays, the same volume of the serum-free medium (1.2 mL) was added to each corneal epithelial cell culture well (approximately 4–5 × 10⁵ cells/well). Except for the control groups that were cultured in serum-free medium alone, the cultures were treated with IL-1β or TNF-α at different concentrations (0.1, 1.0, or 10.0 ng/mL), with or without adding their specific neutralizing antibodies (5 μg/mL), IL-1RA (1 μg/mL), or doxycycline (10 μg/mL) for 24 hours. The conditioned media was then collected and centrifuged, and the supernatants were stored at −20°C until they were used for assays. The adherent cells were lysed in phosphate-buffered saline (PBS, pH 7.3), containing 1.5 M NaCl and 0.039% Triton X-100 for bicinchoninic acid (BCA) protein assay. The cellular protein concentration in each well was used to adjust the volume of the conditioned media used for MMP ELISA and activity assays. These experiments were performed at least three times on each of three separate sets of cultures that were initiated from different donor corneas. 

RNA Isolation and Semiquantitative RT-PCR

Total RNA was isolated from corneal epithelial cultures by acid guanidium thiocyanate-phenol-chloroform extraction using a previously described method. ²⁴ The PCR primers for collagenases (MMP-1, -8, and -13), stromelysins (MMP-3, -10, and -11), MMP-14, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were designed from published human gene sequences (Table 1) . 

The semiquantitative reverse transcription–polymerase chain reaction (RT-PCR) was performed to evaluate the expression of these MMP genes by corneal epithelial cells with a housekeeping gene, GAPDH, as internal control. ¹⁴ In brief, the first-strand cDNAs were synthesized from 1 μg of total RNA at 42°C for 30 minutes. PCR amplification was performed in a DNA thermal cycler (model 480; Perkin-Elmer, Boston, MA) using the following program: denaturation for 2 minutes at 95°C followed by 20 to 40 cycles of denaturation for 1 minute at 95°C, annealing for 1 minute at 60°C and extension for 1 minute at 72°C; extension for 7 minutes at 72°C. Semiquantitative RT-PCR was established by terminating reactions at intervals of 20, 24, 28, 32, 36, and 40 cycles for each primer pair to ensure that the PCR products formed were within the linear portion of the amplification curve. The fidelity of the RT-PCR product was verified by comparing the size of the amplified products to the expected cDNA bands and by sequencing the PCR products through the DNA Core Facility in the Department of Biochemistry and Molecular Biology at the University of Miami. 

Probe Preparation and Northern Hybridization

Human cDNA probes—a 541-bp fragment of MMP-1, 544 bp of MMP-8, 432 bp of MMP-3, 718 bp of MMP-10, 482 bp of MMP-14, and 498 bp of GAPDH—were purified from the RT-PCR products by electrophoresis through a 1.5% low-melting-point agarose gel using a PCR purification kit and gel extraction kit (QIAquick; Qiagen), according to the manufacturer’s protocol. The TIMP-1 probe (551 bp) was kindly provided by Zeenat Gunza-Smith (University of Miami, FL). The fidelity of the cDNA probes was established by sequencing the purified PCR products. These cDNA probes were ³²P-labled (1–2 × 10⁹ cpm/μg DNA) with a random primer DNA labeling system with [α-³²P]-dCTP (3000 Ci/mmol). 

Northern hybridization was performed by using a previously described method. ²⁴ In brief, total RNA for each group at 20 μg/lane was electrophoresed through 1.2% agarose containing formaldehyde, transferred to nitrocellulose membranes, and hybridized with a ³²P-labeled cDNA probe at 2 to 4 × 10⁶ cpm per 3 to 8 ng/mL in the hybridization solution. After visualizing the hybridization product on the x-ray film, the ³²P-label on the membrane was stripped by washing the membranes twice at 65°C for 1 hour in 5 mM Tris-HCl (pH 8.0), 0.2 mM EDTA, 0.05% sodium pyrophosphate and 0.1× Denhardt’s solution. The membranes were then rehybridized with another ³²P-labeled MMP probe or GAPDH probe, which served as a loading control. 

MMP ELISAs

Double sandwich ELISAs for human MMP-1, -13, -3, or -10 were performed with commercial kits (R&D Systems), according to the manufacturer’s protocol. In brief, 100 μL of assay buffer and 100 μL of standard dilutions of recombinant human MMP-1, -3, -10, or -13 and experimental supernatants of the conditioned media were dispensed into a 96-well microtiter plate coated with anti-MMP-1, -3, -10, or -13 monoclonal antibody, respectively. The plate was sealed, incubated at room temperature (RT) for 2 hours. After the plates were washed four times, 200 μL of rabbit anti-MMP-1, -3, -10, or -13 conjugate with horseradish peroxidase was added into each well and incubated at RT for 2 hours. Aliquots of 200 μL of the color reagent 3,3′,5,5′-tetramethylbenzidine (TMB) were then applied for 20 to 30 minutes to develop a blue color, and the reaction was stopped by adding 50 μL of 1 M H₂SO₄. Absorbance was read at 450 nm by an automatic plate reader with a reference wavelength of 570 nm. 

MMP-1 and -13 Activity Assay

The total activity levels of MMP-1 and -13 protein in supernatants of the corneal epithelial cultures were determined with commercial kits (Biotrak, Amersham), according to manufacturer’s protocol. In brief, 100 μL each of pro-MMP-1 (0.31–50 ng/mL) or pro-MMP-13 standards (0.75–24 ng/mL), culture supernatant samples, and 100 μL assay buffer as a blank were incubated at 4°C overnight in microtiter wells precoated with anti-MMP-1 or -13 antibody. Any MMP-1 or -13 present in these solutions bound to the wells, other components of the sample were removed by washing four times with 0.01 M sodium phosphate buffer[pH 7.0] containing 0.05% Tween-20. To measure the total activity of MMP-1 or -13, bound pro-MMP-1 or pro-MMP-13 was activated with 50 μL of 0.025 or 0.5 mM p-aminophenylmercuric acetate (APMA) in assay buffer at 37°C for 1 hour. Detection reagent (50 μL) was then added to each well and incubated at 37°C for 2 hours. Active MMP-1 or -13 was detected though activation of a modified prodetection enzyme and the subsequent cleavage of its chromogenic peptide substrate. The resultant color was read at 405 nm in a microtiter plate reader. The activity of MMP-1 or -13 in a sample was determined by interpolation from a standard curve. 

Statistical Analysis

Student’s t-test was used for statistical comparison of all MMP ELISA and activity assay data. 

Results

Regulation of Collagenases-1, -2, and -3 (MMP-1, -8, and -13) by IL-1β and TNF-α

Semiquantitative RT-PCR.

As shown in Figure 1 , Semiquantitative RT-PCR analysis revealed that the mRNAs of the three collagenases, MMP-1, -8, and -13, were expressed in normal corneal epithelial cells, though MMP-13 mRNA was barely detected. The transcripts of MMP-1 and -13 were upregulated in a dose-dependent manner by a 4-hour treatment with increasing concentrations (0.1, 1.0, and 10.0 ng/mL) of IL-1β or TNF-α (Figs. 1A 1C) when compared with constantly expressed GAPDH mRNA levels (Fig. 1E) , which serve as an internal control. This induction and stimulation lasted for at least 24 hours. Doxycycline (10 μg/mL) treatment inhibited the stimulated transcripts of MMP-1 and -13 by IL-1β and TNF-α at 24 hours (Fig. 2) , but it did not change their levels at 4 hours (Figs. 1A 1C) . In contrast, the expression of MMP-8 mRNA (Fig. 1B) was not affected by exposure to IL-1β or TNF-α, with or without doxycycline. 

Northern Blot Analysis.

Northern hybridization disclosed that single 2.2-kb MMP-1 transcripts and 3.3-kb MMP-8 transcripts were expressed by cultured corneal epithelial cells (Fig. 3) . The sizes of these transcripts were consistent with previous reports. ^{18 19 25} The expression of MMP-1 mRNA was stimulated in a dose-dependent manner by treatment with different doses (0.1, 1.0, and 10.0 ng/mL) of IL-1β and TNF-α. Doxycycline had no inhibitory effect at 4 hours on this stimulated expression of MMP-1 transcripts. In contrast, the expressions of MMP-8 and GAPDH mRNAs were unaltered by exposure to IL-1β or TNF-α, with or without doxycycline. 

MMP-1 and -13 ELISAs.

Treatment of corneal epithelial cells with increasing concentrations of IL-1β or TNF-α (0.1, 1.0, and 10.0 ng/mL) for 24 hours, dose dependently increased the concentrations of MMP-1 and -13 proteins that were secreted into the culture media (Fig. 4) . IL-1β at 1 and 10 ng/mL significantly increased MMP-1 production 2.31-and 3.48-fold (both P < 0.05), respectively, and TNF-α at 1 and 10 ng/mL increased MMP-1 2.35-fold and 3.07-fold (both P < 0.05), respectively. The production of MMP-13 was more markedly stimulated by 3.95- and 6.50-fold (both P < 0.005), respectively, by 1 and 10 ng/mL IL-1β, and by 3.38- and 5.04-fold (both P < 0.05) by 1 and 10 ng/mL TNF-α, respectively. This stimulated production of MMP-1 and -13 protein by 1 ng/mL IL-1β was significantly inhibited by coincubation with 5 μg/mL IL-1β–neutralizing antibody, 1 μg/mL IL-1RA, or 10 μg/mL doxycycline. However, the same concentration of doxycycline did not inhibit the stimulated production of MMP-1 and -13 by 10 ng/mL IL-1β. Similarly, the stimulated production of MMP-1 and -13 by 1 ng/mL TNF-α was completely abolished by coincubation with 5 μg/mL TNF-α neutralizing antibody and was also markedly inhibited by 10 μg/mL doxycycline, MMP-1 and -13 production stimulated by the higher dose of TNF-α (10 ng/mL) was not inhibited by this dose of doxycycline. 

MMP-1 and -13 Activity Assays.

The activity assays for MMP-1 and -13 confirmed the ELISA results. As shown in Figure 5 , MMP-1 activity in the conditioned media was significantly increased 2.23- and 3.76-fold (both P < 0.005), respectively, by IL-1β at 1 and 10 ng/mL, and 2.13-fold (P < 0.005) and 2.78-fold (P < 0.05), respectively, by TNF-α at 1 and 10 ng/mL. MMP-13 activity was also markedly stimulated 2.8- and 4.41-fold (both P < 0.005), respectively, by 1 and 10 ng/mL IL-1β, and 3.17- and 5.45-fold (both P < 0.05) by 1 and 10 ng/mL TNF-α, respectively. The increase in MMP-1 and -13 activities by 1 ng/mL IL-1β was abolished by coincubation with 5 μg/mL IL-1β–neutralizing antibody or 1 μg/mL IL-1RA and was inhibited by 10 μg/mL doxycycline. However, the same concentration of doxycycline did not inhibit the stimulated MMP-1 and -13 activities by 10 ng/mL IL-1β. Similarly, the increased MMP-1 and -13 activities by 1 ng/mL TNF-α were significantly inhibited by coincubation with 5 μg/mL TNF-α–neutralizing antibody or 10 μg/mL doxycycline, but MMP-1 and -13 activities stimulated by the higher dose of TNF-α (10 ng/mL) were not inhibited by this concentration of doxycycline. 

Regulation of Stromelysin-1, -2, and -3 (MMP-3, -10, and -11) by IL-1β and TNF-α

Semiquantitative RT-PCR.

As shown in Figure 6 , Semiquantitative RT-PCR revealed that MMP-3, -10, and -11 were expressed in normal corneal epithelial cells. The transcripts of these three stromelysins were upregulated in a dose-dependent manner by a 4-hour treatment with increasing concentrations (0.1, 1.0, and 10.0 ng/mL) of IL-1β or TNF-α (Figs. 6A 6B 6C) when compared with the levels of the constantly expressed internal control GAPDH (Fig. 6D) . TNF-α only slightly stimulated the expression of MMP-11. This stimulation lasted for at least 24 hours. Doxycycline (10 μg/mL) treatment inhibited the 1 ng/mL IL-1β– or TNF-α–stimulated MMP-10 transcripts at 24 hours (Fig. 2) , but not at the early stage (4 hours; Fig. 6B ). Doxycycline did not inhibit expression of MMP-3 and -11 (Figs. 1 6A 6C) . 

Northern Blot Analysis.

Northern hybridization disclosed that 1.9-kb transcripts of MMP-3 and -10 were expressed by cultured corneal epithelial cells (Fig. 7) . The size of these transcripts was consistent with previous reports. ^{18 19 26} Normalized with GAPDH, the levels of MMP-3 and -10 mRNAs were dose dependently increased by different doses (0.1, 1.0, and 10.0 ng/mL) of IL-1β and TNF-α, with IL-1-β having a greater stimulatory effect than TNF-α. Doxycycline had no inhibitory effect at 4 hours on the expression of these stromelysin transcripts stimulated by either IL-1β or TNF-α at concentrations of 1.0 or 10 ng/mL. 

MMP-3 and MMP-10 ELISAs.

When the corneal epithelial cells were treated with increasing concentrations of IL-1β or TNF-α (0.1, 1.0, and 10.0 ng/mL) for 24 hours, a dose-dependent increase in the concentrations of MMP-3 and -10 proteins secreted into the culture media were observed (Fig. 8) . IL-1β at 1 and 10 ng/mL significantly increased MMP-3 production 2.04- and 3.04-fold (both P < 0.05), respectively, and TNF-α at 1 and 10 ng/mL increased MMP-3 1.88- and 2.97-fold (both P < 0.05), respectively. MMP-10 production was more markedly stimulated, 3.17- and 4.48-fold (both P < 0.005), by 1 and 10 ng/mL IL-1β, respectively, and 3.64- and 5.61-fold (both P < 0.01) by 1 and 10 ng/mL TNF-α, respectively. This stimulated production of MMP-10, but not of MMP-3, by 1 ng/mL IL-1β or TNF-α was significantly inhibited by coincubation with 10 μg/mL doxycycline. However, the same concentration of doxycycline did not inhibit the stimulated MMP-10 production by 10 ng/mL IL-1β or TNF-α (Fig. 8) . 

Expression of MMP-14 and TIMP-1 by Human Corneal Epithelial Cells

Semiquantitative RT-PCR (Fig. 1D) and Northern hybridization (Fig. 9) showed that a membrane-type MMP, MMP-14 (MT-1 MMP), and a tissue inhibitor of MMP, TIMP-1, were consistently expressed in corneal epithelial cells, and were not regulated by treatment with IL-1β and TNF-α, with or without doxycycline. 

Discussion

Collagenases in the Human Corneal Epithelium

In this study, three human collagenases, MMP-1, -8, and -13, were expressed by cultured human corneal epithelial cells, though MMP-13 mRNA was barely detectable under normal conditions. The proinflammatory cytokines, IL-1β and TNF-α, stimulated mRNA expression and protein production and activity of MMP-1 and -13 by corneal epithelial cells. Semiquantitative RT-PCR disclosed that the MMP-1 and -13 transcripts were dose dependently increased by these two inflammatory cytokines (Figs. 1A 1C) . Northern hybridization supported the RT-PCR findings for MMP-1 mRNA (Fig. 3) , but it was not sensitive enough to detect the MMP-13 transcripts. The regulated expression of MMP-1 and -13 was confirmed at protein levels by ELISA and enzyme activity assays (Figs. 4 5) . The specificity of these findings was confirmed by using specific neutralizing antibodies against IL-1β and TNF-α that abolished their stimulation of MMP-1 and -13 protein and activity and by IL-1RA, which also inhibited IL-1β–stimulated production and activity of MMP-1 and -13. 

Stimulated production of MMP-1 by IL-1β and TNF-α has been reported in human corneal stromal fibroblasts ^{11 27 28} and other tissues, ^{29 30 31} but this effect has not been evaluated in human corneal epithelium. Human MMP-13 was originally identified in breast carcinomas, and it is produced mainly by malignant tumors, such as head and neck carcinomas, chondrosarcomas, and basal cell carcinomas of the skin. In all these tumors, MMP-13 expression is associated with an invasive and metastatic phenotype. ^{32 33 34} The expression of MMP-13 in tumor cell lines is enhanced by TNF-α and TGF-β. ³⁵ In contrast, MMP-13 expression was not detected by normal keratinocytes in intact or reepithelealizing epidermis, ³⁶ nor in normal fibroblasts cultures. ³⁷ MMP-13 was expressed abundantly only by fibroblasts deep in the bed of chronic cutaneous ulcers, but was not detected in the epidermis and in acute wounds. ³⁷ There is only one previously reported study evaluating MMP-13 in the cornea. MMP-13 mRNA was detected in epithelial cells of wounded rat corneas, but not in the normal control, and in the wounded corneas, MMP-13 mRNA was localized exclusively to basal epithelial cells. ¹⁶ In this study we have shown for the first time that MMP-13 is expressed by primary human corneal epithelial cultures, and its production and activity was enhanced by IL-1β and TNF-α. Our findings indicate that these collagenases may be of important in the pathogenesis of corneal epithelial diseases. Stimulated MMP-1, which is critical for reepithelialization, and MMP-13, which has potent proteolytic activity and a wide range of substrate specificity, may play roles in remodeling the collagenous matrix in chronic wounds, as well as in the matrix destruction of corneal ulceration. 

Stromelysins in Human Corneal Epithelium

This study demonstrates that human corneal epithelial cells express all three human stromelysins, MMP-3, -10, and -11. Semiquantitative RT-PCR revealed that IL-1β and TNF-α dose dependently stimulated all three stromelysin transcripts, though TNF-α increased MMP-11 slightly (Fig. 6) . Northern hybridization confirmed the RT-PCR findings (Fig. 7) . The stimulatory effects of these proinflammatory cytokines on MMP-3 and -10 protein synthesis was confirmed by ELISA (Fig. 8) . All three stromelysins are produced by a variety of tumor cells and are stimulated in wound healing and inflammation. ^{38 39 40 41} The effects of inflammatory cytokines on the regulated expression of stromelysins have not been well elucidated. Although there have been no previous reports on the regulation of stromelysin expression in corneal epithelial cells, overexpression of MMP-10 in diabetic corneal epithelium and stroma and MMP-3 expression in the corneal stroma have been recognized. ¹⁷ Overexpression of MMP-3 has been observed in fibroblasts isolated from pterygia ¹⁸ and conjunctivochalasis ¹⁹ as well as the aqueous humor of patients with uveitis. ⁴² Upregulation of MMP-3 expression by IL-1β and/or TNF-α in human ocular stromal fibroblasts ^{11 22 27 28} and other tissues ^{29 30 43} has been reported. Our study has demonstrated for the first time that production of MMP-3, -10, and -11 by human corneal epithelial cells is dose dependently stimulated by IL-1β and TNF-α. MMP-3 and -10 share 82% sequence homology, but their expression appears to be differentially regulated. It was reported that expression of MMP-10, but not MMP-3, in human keratinocytes was induced by TNF-α, but not by IL-1β. ⁴⁴ Our results indicate that these two inflammatory cytokines stimulate both MMP-10 and -3 production. These differences suggest that the regulation of stromelysin expression may be tissue specific. 

Inhibition of Protein Production and Activity by Doxycycline

Tetracyclines and chemically modified non-antimicrobial tetracyclines (CMTs) have been reported to inhibit the activity of MMPs. ^{45 46 47} We have experienced clinical success in treating patients with rosacea-associated corneal epithelial erosions with doxycycline, a long-acting, second-generation tetracycline. ⁹ We have also reported that doxycycline, at a nontoxic dose, markedly decreases the activity of pro-MMP-9 and of IL-1β– and TNF-α-induced pro-MMP-9 produced by human corneal epithelium. ^{14 23} Doxycycline and chemically modified tetracycline 3 (CMT-3) inhibited secretion and activity of both MMP-2 and -9 produced by human prostate cancer cells. ⁴⁷ Most previous reports focused on the effects of tetracyclines on the activity of gelatinases and the neutrophil collagenase MMP-8. Herein, we provide evidence that doxycycline at a pharmacologically achievable nontoxic dose (10 μg/mL) significantly inhibits the collagenases MMP-1 and -13 and the stromelysin MMP-10, at the transcript, protein production, and activity levels stimulated by IL-1β or TNF-α (at a concentration of 1 ng/mL) in cultured corneal epithelial cells (Figs. 2 4 5 8) . Our results also show that doxycycline did not inhibit MMP-1, -13, and -10 protein and activity stimulated by a higher dose (10 ng/mL) of IL-1β or TNF-α. 

Regulated Expression of MMPs by IL-1β and TNF-α

This study and our previous report ¹⁴ have demonstrated that inflammatory cytokines upregulate three main groups of MMPs, gelatinases (MMP-9), collagenases (MMP-1 and -13), and stromelysins (MMP-3, -10 and -11), but not MMP-14 and TIMP-1, in human corneal epithelial cells. It is possible that MMPs participate in the epithelial erosions and stromal ulcerations that accompanying ocular surface diseases such as Sjögren syndrome keratoconjunctivitis sicca and ocular rosacea, in which the concentrations of proinflammatory cytokines, such as IL-1, in the tear fluid or ocular surface epithelium correlate with disease severity. ^{5 6 7 13 23 48 49} Most MMPs are synthesized as inactive zymogens (pro-MMP) that undergo subsequent activation in the extracellular milieu. Pro-MMPs can be activated by various factors such as organomercurials, serine proteinases, hypochlorous acid, and acid exposure. ⁵⁰ However, findings in recent studies suggest that intermolecular activation of pro-MMP by active MMP may be the physiological mechanism of activation in vivo. The stromelysins MMP-3 and -10 are considered to play a central role in the activation of various pro-MMP, including pro-MMP-9, ^{26 51} pro-MMP-1, ⁵² pro-MMP-8, ⁵³ and pro-MMP-7 (promatrilysin). ²⁶ Pro-MMP-9 (92 kDa) is processed to its full activity as a low molecular weight species of 81 kDa and 65 kDa by MMP-3 or -10. Pro-MMP-1 is activated by MMP-3 and -10 to generate a fully active collagenase of 41 kDa. ⁵² Membrane-type metalloproteinase 1 (MT1-MMP, MMP-14) can activate pro-MMP-2 in vitro. ⁵⁰ MMP-2 and -14 are able to cleave human procollagenase-3 (pro-MMP-13, 60 kDa) to its fully active enzyme (Tyr85 N terminus, 48 kDa). ⁵⁴ We have detected the expression of MMP-2 and -14 in corneal epithelial cells, although their expression was not found to be regulated by IL-1β and TNF-α. Thus, through a coordinated proteolytic cascade, collagenases can degrade the native fibrillar collagens, which can be further degraded by gelatinases such as MMP-9. 

In conclusion, the present study and data from our previous study ¹⁴ demonstrate for the first time that the proinflammatory cytokines IL-1β and TNF-α upregulate the three major groups of MMPs: gelatinase (MMP-9), collagenases (MMP-1 and -13), and stromelysins (MMP-3, -10, and -11), by cultured human corneal epithelial cells. Doxycycline treatment suppressed IL-1β- and TNF-α–stimulated production of MMP-1, -13, and -10 at mRNA and protein levels, but it had no effect on MMP-3. These results may explain the reported therapeutic efficacy of doxycycline and provide insight into strategies for treating MMP-mediated ocular surface diseases. 

 Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2002.

 Supported in part by National Eye Institute Grant EY11915 (SCP), Grant CA61038 from the National Cancer Institute (BLL), Department of Defense Grant DAMD 17-98-18526 (BLL), and unrestricted grants from Research to Prevent Blindness, the Oshman Foundation, and the William Stamps Farish Fund.

 Submitted for publication August 26, 2002; revised February 3, 2003; accepted February 26, 2003.

 Disclosure: D.Q. Li, None; T.Y. Shang, None; H.-S. Kim, None; A. Solomon, None; B.L. Lokeshwar, None; S.C. Pflugfelder, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Human Primer Sequences Used for RT-PCR and Probe Preparation

Gene	Accession	Sense Primer	Antisense Primer	Probe (bp)
MMP-1	M13509	GGAGGGGATGCTCATTTTGATG	TAGGGAAGCCAAAGGAGCTGT	541
MMP-8	NM_002424	TGGACCCAATGGAATCCTTGC	ATAGCCACTCAGAGCCCAGTA	544
MMP-13	NM_002427	TTGTTGCTGCGCATGAGTTCG	GGGTGCTCATATGCAGCATCA	370
MMP-3	J03209	CCTGCTTTGTCCTTTGATGC	TGAGTCAATCCCTGGAAAGTC	432
MMP-10	NM_002425	CTGCCATTGAGAAAGCTCTGA	CCTGCTTGTACCTCATTTCCTC	718
MMP-11	NM_005940	AAGACTCACCGAGAAGGGGAT	CCTCGGAAGAAGTAGATCTTG	665
MMP-14	U41078	GATGTTTGTCTTCAAGGAGCG	CCTTCAGCTTCTGGTTGTTGA	482
GAPDH	M33197	GCCAAGGTCATCCATGACAAC	GTCCACCACCCTGTTGCTGTA	498

Figure 1.

 

Representative semiquantitative RT-PCR profiles of mRNA expression of collagenases and MMP-14 regulated by IL-1β or TNF-α. Total RNA was extracted from primary human corneal epithelial cells treated with IL-1β or TNF-α for 4 hours at 0.1, 1, and 10.0 ng/mL without (−) or with (+) 10 μg/mL doxycycline (Doxy). RT-PCR was performed using an equal amount (1 μg) of total RNA and was analyzed at intervals of four cycles from 20 to 40 PCR cycles. PCR products with the expected sizes were amplified for MMP-1 (A), -8 (B), -13 (C), and -14 (D) with GAPDH (E) as an internal control. A 100-bp DNA ladder was run in parallel.

Figure 2.

 

Representative RT-PCR profiles of mRNA expression of collagenases and stromelysins regulated by IL-1β or TNF-α with or without doxycycline for 24 hours. Total RNA was extracted from corneal epithelial cells in six groups: control groups (lanes 1 and 4), groups treated for 24 hours with 1 ng/mL IL-1β (lane 2) or TNF-a (lane 5), and groups treated with 10 μg/mL doxycycline plus 1 ng/mL IL-1β (lane 3) or TNF-α (lane 6). RT-PCR was performed with an equal amount (1 μg) of total RNA and was analyzed at intervals of four cycles from 20 to 40 PCR cycles. PCR products of the expected sizes were amplified for MMP-1, -13, -3, -10, and -11 with GAPDH as an internal control. A 100-bp DNA ladder was run in parallel.

Figure 3.

 

Representative Northern hybridization profiles of mRNA expression of collagenases regulated by IL-1β or TNF-α. Total RNA was extracted from corneal epithelial cells treated with IL-1β or TNF-α for 4 hours at 0.1, 1.0, and 10.0 ng/mL, without (−) or with (+) 10 μg/mL doxycycline (Doxy). Northern hybridization was performed sequentially with ³²P-labeled cDNA probes for MMP-1 or -8 or GAPDH, the latter serving as a loading control.

Figure 4.

 

MMP-1 and -13 ELISAs for protein concentration of collagenases in conditioned media. Subconfluent primary cultures of human corneal epithelial cells were switched to serum-free medium and treated with IL-1β or TNF-α at 0.1, 1.0, and 10.0 ng/mL, with or without neutralizing antibody (Ab), IL-1RA (RA), or 10 μg/mL doxycycline (D) for 24 hours. The supernatants of the conditioned media were collected for MMP ELISAs, according to the manufacturer’s protocol.

Figure 5.

 

MMP-1 and -13 activity assays for collagenase activity in conditioned media, conducted as described in Figure 4 .

Figure 6.

 

Representative semiquantitative RT-PCR profiles for mRNA expression of stromelysins regulated by IL-1β or TNF-α. Total RNA was extracted from corneal epithelial cells treated with IL-1β or TNF-α for 4 hours at 0.1, 1.0, and 10.0 ng/mL without (−) or with (+) 10 μg/mL doxycycline (Doxy). RT-PCR was performed and analyzed between 20 and 40 PCR cycles. PCR products with the expected sizes were amplified for MMP-3 (A), -10 (B), and -11 (C) with GAPDH (D) as an internal control.

Figure 7.

 

Representative Northern hybridization profiles for mRNA expression of stromelysins MMP-3 and -10 regulated by IL-1β or TNF-α. Cells were prepared as described in Figure 3 . Northern hybridization was performed sequentially with ³²P-labeled cDNA probes for MMP-3 or -10 or GAPDH, the latter serving as a loading control.

Figure 8.

 

MMP-3 and -10 ELISAs for protein concentration of stromelysins in conditioned media. Cells were prepared as described in Figure 4 . The supernatants of the conditioned media were collected for MMP-3 and -10 ELISAs, according to manufacturer’s protocol.

Figure 9.

 

Representative Northern hybridization profiles for mRNA expression of MMP-14 and TIMP-1. Cells were prepared as described in Figure 3 . Northern hybridization was performed sequentially with ³²P-labeled cDNA probes for MMP-14, TIMP-1, or GAPDH, the latter serves as a loading control.