Doxycycline Inhibits TGF-β1–Induced MMP-9 via Smad and MAPK Pathways in Human Corneal Epithelial Cells

Abstract

purpose. To evaluate the effects of TGF-β1 and doxycycline on production of gelatinase MMP-9 and activation of Smad, c-Jun N-terminal kinase (JNK), extracellular-regulated kinase (ERK), and p38 mitogen-activated protein kinase (MAPK) signaling pathways in human corneal epithelial cells.

methods. Primary human corneal epithelial cells were cultured to confluence. The cells were treated with different concentrations of TGF-β1 (0.1, 1, or 10 ng/mL), with or without TGF-β1–neutralizing mAb (5 μg/mL), SP600125 (30 μM), PD98059 (40 μM), SB202190 (20 μM), or doxycycline (5–40 μg/mL) for different lengths of time. Conditioned media were collected from cultures treated for 24 to 48 hours to evaluate the MMP-9 production by zymography and activity assay. Total RNA was isolated from cells treated for 6 to 24 hours to evaluate MMP-9 expression by semiquantitative RT-PCR and Northern hybridization. Cells treated for 5 to 60 minutes were lysed in RIPA buffer for Western blot with phospho-specific antibodies against Smad2, JNK1/2, ERK1/2, or p38.

results. TGF-β1 increased expression, production, and activity of MMP-9 by human corneal epithelial cells in a concentration-dependent fashion. TGF-β1 also induced activation of Smad2, JNK1/2, ERK1/2, and p38 within 5 to 15 minutes, with peak activation at 15 to 60 minutes. Doxycycline markedly inhibited the TGF-β1–induced production of MMP-9 and activation of the Smad, JNK1/2, ERK1/2, and p38 signaling pathways. Its inhibitory effects were of a magnitude similar to SP600125, PD98059, and SB202190, specific inhibitors of the JNK1/2, ERK1/2, and p38 pathways, respectively.

conclusions. These findings demonstrated that doxycycline inhibits TGF-β1–induced MMP-9 production and activity, perhaps through the Smad and MAPK signaling pathways. These inhibitory effects may explain the reported efficacy of doxycycline in treating MMP-9–mediated ocular surface diseases.

The important role of matrix metalloproteinases (MMPs) has been recognized in ocular surface diseases including wound healing, dry eye, sterile corneal ulceration, recurrent epithelial erosion, corneal neovascularization, pterygium, and conjunctivochalasis. ^{1 2 3 4 5 6} Among these, gelatinase B (MMP-9) has been found to be of central importance in extracellular matrix remodeling after wounding of the corneal surface and has been implicated in the pathogenesis of sterile corneal ulceration, dry eye–associated epitheliopathy, and other ocular diseases. ^{7 8 9 10} The expression of MMP-9 is known to be modulated by cytokines and growth factors that induce cellular responses by activating a variety of intracellular signaling cascades. The pro-inflammatory cytokines IL-1β and TNF-α, as well as transforming growth factor (TGF)-β have been reported to play a role in the regulation of MMP-9 production and activity in human corneal epithelial cells. ¹¹  

TGF-β is known to play a vital role in tissue remodeling during inflammatory responses. TGF-β has been found to stimulate the expression of MMP-9 in endothelial cells, fibroblasts, and epithelial cells, including the corneal epithelium. ^{11 12 13 14} The diverse cellular responses elicited by TGF-β are triggered by activation of serine/threonine kinase TGF-β receptors and are transduced through multiple intracellular signal pathways. After ligand binding, the type II TGF-β receptor recruits the type I receptor, which is then phosphorylated by the type II receptor. The activation of the type I receptor triggers the phosphorylation of cytoplasmic proteins, including the novel Smad pathway, that are directly implicated in the transmission of the intracellular signal. ¹⁵ Many of the signaling responses induced by TGF-β are mediated by Smad proteins, but certain evidence has suggested that TGF-β can also signal independently of Smads. ¹⁶ TGF-β can activate mitogen-activated protein kinase (MAPK) signaling pathways including c-Jun N-terminal kinase (JNK), extracellular-regulated kinase (ERK), and p38 MAPK, and the specific pathway used appears to vary with and depend on the cell type and tissue function. ^{17 18} The signaling pathways that mediate the regulatory effect of TGF-β on MMP-9 are not understood. 

Doxycycline, a long-acting semisynthetic tetracycline, is well recognized for its therapeutic efficacy in treating MMP-mediated ocular surface diseases, such as rosacea, and sterile corneal ulceration. ^{19 20} Doxycycline has been found to inhibit MMP-9 activity in human endothelial cells, ²¹ skin keratinocytes, ²² and cancer cells. ^{23 24} Doxycycline was reported to significantly decrease serum MMP-9 levels (50%) after 6 months of oral administration after myocardial infarction. ²⁵ Doxycycline has been reported to decrease the activity of MMP-9 in conditioned media of cultured human corneal epithelial cells, ²⁶ but the effects and mechanisms of doxycycline on TGF-β–induced MMP-9 production are not known. 

This study investigated the effects of doxycycline on TGF-β1–induced production of MMP-9 and activation of the Smad2, JNK, ERK, and p38 MAPK signaling pathways in human corneal epithelial cells. 

Materials and Methods

DMEM/Ham’s F12, amphotericin B, phenol, DNA or RNA size markers, and a random primer DNA labeling kit were purchased from Invitrogen-Gibco (Grand Island, NY); fetal bovine serum (FBS) from Hyclone (Logan, UT); recombinant human TGF-β1 and TGF-β1 neutralizing monoclonal antibody (mAb) from R&D Systems (Minneapolis, MN); activity assay kits (Biotrak) for MMP-9 and chemiluminescence reagents (ECL Plus) from Amersham Pharmacia Biotech (Piscataway, NJ); EDTA-free protease inhibitor cocktail tablet from Roche Applied Sciences (Indianapolis, IN); the bicinchoninic (BCA) protein assay kit from Pierce Chemical (Rockford, IL); SP600125, PD98059, and SB202190 from Calbiochem (Darmstadt, Germany); rabbit polyclonal antibodies against ERK or p38 from Cell Signaling (Beverly, MA); phospho-specific mAbs for JNK, ERK, and p38 and rabbit antibodies against JNK and phospho-Smad2 from Santa Cruz Biotechnology (Santa Cruz, CA); polyvinylidene difluoride (PVDF) membranes from Millipore (Bedford, MA); nitrocellulose membranes from Schleicher and Schuell (Keene, NH); the PCR kit (GeneAmp) from Applied Biosystems (Foster City, CA); PCR purification and gel extraction kits (QIAquick) from Qiagen (Valencia, CA); [α-³²P]-dCTP from Du Pont NEN (Boston, MA); and film and intensifying screens (XAR-5 and BioMax MS-1) from Eastman Kodak (Rochester, NY). All plastic ware was from BD Biosciences (Franklin Lakes, NJ). Doxycycline, cholera toxin subunit A, hydrocortisone, 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 Lions Eye Bank of Texas, using a previously described method. ^{26 27} In brief, corneoscleral rims were trimmed, the endothelial layer and iris remnants were removed, and each limbal rim was dissected into 12 equal segments. Two segments were placed in each well of six-well culture plates, and each explant was covered with a drop of FBS overnight. The explants were then cultured in hormone supplemented medium, containing equal amounts of DMEM and Ham’s F12, supplemented with 5% FBS, 0.5% dimethyl sulfoxide, 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 in 95% humidity and 5% CO₂. The medium was renewed every 2 to 3 days. 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 just medium described, but without FBS) for 24 hours before treatment. 

Cell Treatment

Except for the control groups that were continually cultured in serum-free medium alone, the cultures were treated with TGF-β1 at different concentrations (0.1, 1.0, or 10 ng/mL), with or without the TGF-β1-neutralizing mAb (5 μg/mL), SP600125 (30 μM, a JNK/SAPK pathway inhibitor), PD98059 (40 μM, a ERK MAPK pathway inhibitor), SB202190 (20 μM, a p38 MAPK pathway inhibitor), or doxycycline (5–40 μg/mL) for different lengths of time. For gelatin zymography and MMP-9 activity assays, the same volume of the serum-free medium (1.2 mL) was added to each well of corneal epithelial cultures (∼5 × 10⁵ cells/well). The conditioned medium after 24 to 48 hours of treatment was collected and centrifuged, and the supernatants were stored at −80°C before use. 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 the BCA protein assay. The cellular protein concentration in each well was used to adjust the volume of the conditioned medium used for gelatin zymography and MMP-9 activity assays. For MMP-9 gene expression analysis, the corneal epithelial cells after treatment for 6 hours were lysed in 4 M guanidium solution and subjected to total RNA extraction. The cells treated for shorter times (5, 15, 30, and 60 minutes) were lysed in RIPA lysis buffer, containing 50 mM Tris-HCl, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 2 mM sodium fluoride, 2 mM EDTA, 0.1% SDS, and an EDTA-free protease inhibitor cocktail tablet, for Western blot analysis. These experiments were performed at least three times using separate sets of cultures that were initiated from different donor corneas. 

Gelatin Zymography

To determine the relative concentration and activity of gelatinases in the conditioned media from corneal epithelial cultures receiving various treatments, SDS-PAGE gelatin zymography was performed with a previously reported method. ²⁸ Briefly, 8 to 10 μL of each conditioned medium was used, and the volume was adjusted by cellular protein concentrations. The media supernatants were treated with SDS-PAGE sample buffer without boiling or reduction. Samples were fractionated in an 8% polyacrylamide gel containing gelatin (0.5 mg/mL) by electrophoresis at 100 V for 90 minutes at 4°C. The gels were soaked in 0.25% Triton X-100 for 30 minutes at room temperature to remove the SDS, and incubated in a digestion buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 10 mM CaCl₂, 2 μM ZnSO₄ and 0.01% Brij-35) containing 5 mM phenylmethylsulfonyl fluoride (PMSF), a serine protease inhibitor, at 37°C overnight, to allow proteinase digestion of its substrate. The gels were then stained with 0.25% Coomassie brilliant blue R-250 in 40% isopropanol for 2 hours, and destained with 7% acetic acid. Gelatinolytic activities appeared as clear bands of digested gelatin against a dark blue background of stained gelatin. 

MMP-9 Activity Assay

The total activity levels of MMP-9 protein in supernatants of the corneal epithelial cultures were determined using MMP-9 activity assay systems (Biotrak) according to manufacturer’s protocol. This assay primarily measured the activity of latent MMP-9, because the active form of MMP-9 in the media was almost undetectable. In brief, 100 μL each pro-MMP-9 standard (0.5–32 ng/mL), culture supernatant sample, or assay buffer (for blanks) was incubated at 4°C overnight in microtiter wells precoated with anti-MMP-9 antibody. Any MMP-9 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-9, bound pro-MMP-9 was activated with 50 μL of 1 mM p-aminophenylmercuric acetate (APMA) in assay buffer at 37°C for 2 hours. Detection reagent (50 μL) was then added to each well and incubated at 37°C for 6 hours. Active MMP-9 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-9 in a sample was determined by interpolation from a standard curve. 

RNA Isolation and Semiquantitative RT-PCR

Total RNA was isolated from corneal epithelial cell cultures by acid guanidium thiocyanate-phenol-chloroform extraction according to a previously described method. ²⁹ The PCR primers for MMP-9 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-9 genes by corneal epithelial cells with a housekeeping gene, GAPDH, as an internal control. ¹¹ 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 9700; Applied Biosystems, Inc.) 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, with a final 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. 

Probe Preparation and Northern Hybridization

Human cDNA probes (552-bp fragment of MMP-9 and 498 bp of GAPDH) were purified from the RT-PCR products by electrophoresis through a 1.5% low-melting-point agarose gel, with a PCR purification kit (QIAquick; Qiagen) and gel extraction kit, per the manufacturer’s protocol. Northern hybridization was performed with a previously described method. ²⁹ 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/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 GAPDH probe, which served as a loading control. 

Western Blot Analysis

The cells treated for 5 to 60 minutes were lysed in RIPA buffer, and appropriate volumes (25–30 μL) of cell extracts, adjusted to represent the same amount of total cellular protein (50 μg), were electrophoresed under reducing condition at 4°C in a 4% to 15% gradient polyacrylamide gel. After electrophoretic transfer to a PVDF membrane at 100 V for 1 hour at 4°C, the membranes were blocked with 5% nonfat milk in TTBS (50 mM Tris [pH 7.5], 0.9% NaCl, and 0.1% Tween-20) for 1 hour at RT. The primary antibody (i.e., phospho-specific antibodies against Smad2, JNKs, ERKs, or p38) in TTBS containing 5% nonfat milk was placed on each membrane and incubated at RT for 60 minutes with agitation. After three washings with TTBS over 15 minutes, the second antibody, conjugated with horseradish peroxidase (HRP), was applied and the signals on the membranes were detected by chemiluminescence reagents (ECL Plus; Amersham Pharmacia Biotech) and imaged with a digital image station (2000R; Eastman Kodak). 

Results

Effects of TGF-β1 and Doxycycline on MMP-9 Production and Activity

The protein levels of MMP-9 were determined by gelatin zymography and MMP-9 activity assay in the serum-free conditioned medium from corneal epithelial cultures. As shown in Figures 1A and 1B , zymography revealed that TGF-β1 stimulated 92-kDa MMP-9 protein production in a concentration-dependent manner by corneal epithelial cells treated with increasing concentrations of TGF-β1 (0.1, 1.0, and 10 ng/mL) for 24 hours. TGF-β1 at 1 ng/mL markedly stimulated MMP-9 production and the strongest stimulation was observed with 10 ng/mL. This stimulation at 1 ng/mL TGF-β1 was abolished by 5 μg/mL TGF-β1 neutralizing mAb or 10 μg/mL doxycycline (Fig. 1A) . In contrast, the bands of 72-kDa gelatinase A (MMP-2) were essentially unchanged. The bands in zymograms from three experiments were scanned, and the quantified data showing the relative change (x-fold) in MMP-9 protein are summarized in Figure 1B . Doxycycline at a concentration of 10 μg/mL completely abolished the stimulation of MMP-9 by 1 ng/mL TGF-β1. 

The MMP-9 activity assays confirmed the zymography results. As shown in Figure 1B , MMP-9 activities in the conditioned media from TGF-β1 (0.1, 1.0, or 10 ng/mL) treated epithelial cultures were 1.97 ± 0.20, 4.09 ± 0.10, and 4.65 ± 0.25 ng/mL (mean ± SD, n = 3) respectively, significantly higher than the control group (0.82 ± 0.09 ng/mL). MMP-9 activity in the conditioned media was significantly increased to 4.99- and 5.67-fold (both P < 0.05) by TGF-β1 at concentrations of 1 and 10 ng/mL, respectively. The upregulation of MMP-9 by 1 ng/mL TGF-β1 was inhibited 54.7% by 10 μg/mL doxycycline (P < 0.05), but MMP-9 stimulated by the higher dose of TGF-β1 (10 ng/mL) was not inhibited by this pharmacologic dose of doxycycline. 

Effects of TGF-β1 and Doxycycline on MMP-9 mRNA Expression

Semiquantitative RT-PCR analysis revealed that MMP-9 mRNA was expressed by normal corneal epithelial cells. The MMP-9 transcripts were upregulated in a concentration-dependent manner after 6 (Fig. 2A)and 24 hours (Fig. 2B)of treatment with increasing concentrations (0.1, 1.0, 10 ng/mL) of TGF-β1 when compared with constantly expressed levels of GAPDH mRNA (Fig. 2C) , which served as an internal control. The addition of doxycycline did not affect the levels of MMP-9 mRNA at 6 hours, but did inhibit MMP-9 expression at 24 hours (Fig. 2B) . The scanned and quantified data of RT-PCR profiles of cultures treated for 24 hours from three experiments (Figs. 2B 2C)are summarized in Figure 2D , which shows the relative x-fold change of MMP-9 mRNA after normalization by GAPDH mRNA. The results show that 1 ng/mL TGF-β1 increased MMP-9 mRNA by 3.3-fold, whereas doxycycline decreased this stimulation by 78.3% to 1.5-fold. Northern hybridization disclosed that a single 2.9 kb band of MMP-9 transcript was expressed by cultured corneal epithelial cells (Fig. 2E) . The size of this transcript is consistent with previous reports. ³⁰ The expression of MMP-9 mRNA was stimulated in a concentration-dependent manner by treatment with different doses (0.1, 1.0, and 10 ng/mL) of TGF-β1. In contrast, the 1.4-kb band of GAPDH transcripts was unaltered by exposure to TGF-β1. The addition of doxycycline did not affect the levels of MMP-9 expression at 6 hours (Fig. 2E) , which agreed with RT-PCR results. 

Effects of TGF-β1 and Doxycycline on Activation of Smad Pathway

Phosphorylated Smad2 protein (60 kDa) was detected in corneal epithelial cells by Western blot using a phospho-specific polyclonal antibody. As shown in Figure 3A , 10 ng/mL TGF-β1 activated phosphorylation of Smad2 by 5 minutes after exposure and activation peaked at 30 minutes. TGF-β1 concentration dependently induced Smad2 phosphorylation, which was detected after treatment with a low concentration of TGF-β1 (0.1 ng/mL) and peaked at 10 ng/mL. Doxycycline at 10 μg/mL markedly inhibited the activation of Smad2 induced by 1 ng/mL TGF-β1 (Fig. 3B) . 

Effects of TGF-β1 and Doxycycline on Activation of MAPK Pathways

The activation of MAPK signaling pathways in corneal epithelial cells was detected by Western blot using phospho-specific antibodies. As shown in Figure 4A , 10 ng/mL TGF-β stimulated phosphorylation of JNK1 (46 kDa) and JNK2 (54 kDa) as early as 5 minutes, peaking at 15 minutes. TGF-β1 at 0.1, 1, and 10 ng/mL concentration dependently induced p-JNK1/2 when compared with constant levels of total JNK1/2. SP600125, a specific inhibitor of the JNK pathway, markedly inhibited the activation of JNK1/2 induced by 1 ng/mL TGF-β1, but did not decrease levels of total JNK1/2 (Fig. 4B) . Doxycycline showed an inhibitory effect on the TGF-β1–induced activation of JNK1/2, especially at a 10-μg/mL concentration, which was similar to the inhibitory effect produced by SP600125 (Figs. 4B 4C) . Activation of 46-kDa p-JNK1 appeared more pronounced than 54-kDa p-JNK2, suggesting that p-JNK1 may be the major isoform of JNK that is activated by TGF-β1. 

As shown in Figure 5A , Western blot revealed that 10 ng/mL TGF-β1 activated phosphorylation of ERK1/2 as early as 15 minutes, peaking at 60 minutes in the time course tested. TGF-β1, at 0.1, 1, and 10 ng/mL, concentration dependently induced p-ERK1/2 when compared with relatively constant levels of total ERK1/2. PD98059, a specific inhibitor of the ERK pathway, markedly inhibited the activation of ERK1/2 induced by 1 ng/mL TGF-β1, but did not inhibit total ERK1/2 (Fig. 5B) . Doxycycline at a 5- to 40-μg/mL concentration dependently inhibited the TGFβ1-induced activation of ERK1/2 (Fig. 5C) . The inhibitory effect of doxycycline on p-ERK1/2 was similar to PD98059 (Fig. 5B)All figures show that the 42-kDa band of p-ERK2 was stronger than 44-kDa p-ERK1, which may suggest that p-ERK2 was the major isoform of ERK is activated by TGF-β1. 

TGF-β1 also induced activation of p38 MAPK in corneal epithelial cells. Western blot showed that 10 ng/mL TGF-β1 activated phosphorylation of p38 as early as 5 minutes, and the reaction peaked at 60 minutes in the time course tested (Fig. 6A) . TGF-β1 concentration dependently induced p-p38 at a low concentration of 0.1 ng/mL, with peak activation at 10 ng/mL. In contrast, levels of total p38 remained constant. SB202190, a specific inhibitor of p38 MAPK pathway, markedly inhibited p38 phosphorylation induced by 1 ng/mL TGF-β1, but did not affect levels of total p38. Doxycycline at, 5 to 40 μg/mL, concentration dependently inhibited the TGFβ1–induced activation of p-38 (Fig. 6C) . The inhibitory effect of doxycycline on p-p38 was similar to SB202190 (Fig. 6B) . 

These MAPK pathway inhibitors not only blocked the phosphorylation of the JNKs, ERKs, or p38, then also markedly inhibited MMP-9 production by corneal epithelial cells. SP600125, PD98059, and SB202190 almost abolished the production and activity of MMP-9 stimulated by TGF-β1 (1 ng/mL) in the cultures treated for 24 (Fig. 7A)and 48 (Fig. 7B)hours. SB202190 showed stronger inhibition of TGF-β1–induced MMP-9 than did SP600125 and PD98059, perhaps because SB202190 at 20 μM inhibits both the p38 and JNK pathways. ^{31 32} Doxycycline, at 5 to 40 μg/mL, concentration dependently inhibited the TGF-β1–ed production of MMP-9 (Fig. 7C) . The inhibitory effect of doxycycline at 10 μg/mL on TGF-β1–stimulated MMP-9 production was similar to that of SP600125 and PD98059, whereas it was slightly weaker than that of SB202190 at 24 and 48 hours. 

Discussion

Stimulation of MMP-9 Expression, Production and Activity by TGF-β1

In mammals there are three isoforms of TGF-β (TGF-β1, -β2, and -β3), which are functionally similar and interchangeable under many conditions, although they may possess unique functions. ^{33 34 35} All three TGF-β isoforms are expressed by ocular surface fibroblasts and epithelial cells. ^{29 35 36 37 38} Among them, TGF-β1 and -β2 are abundantly present in human tears, which are secreted by lacrimal glands and are produced by the epithelial and inflammatory cells that reside on the ocular surface. ^{29 38 39} TGF-β1 is the predominant isoform, particularly in its latent inactive form, that may be activated by proteases during inflammation and wound healing. ^{40 41 42 43} In this study, we used TGF-β1 to investigate its effects on regulation of MMP-9 production by primary cultured human corneal epithelial cells. 

The role of TGF-β1 in regulating MMP-9 has not been well understood, and previous studies have reported conflicting results. TGF-β1 has been reported to reduce the production of gelatinases (MMP-2 and -9) and enhance the expression of the tissue inhibitors of MMP (TIMP-1 and -3) in human lung fibroblasts ³³ and myometrial smooth muscle cells. ⁴⁴ TGF-β1 has also been found to stimulate production of MMP-9 in keratinocytes, ¹⁴ odontoblasts, ⁴⁵ human fibroblasts, ¹³ and oral tumor cells. ⁴⁶ These observations suggest that the effect of TGF-β1 on MMP-9 may be species specific, tissue or cell-type specific, and temporally and spatially specific. We have observed by zymography that TGF-β1 stimulates MMP-9 production. ¹¹ In the current study, we demonstrated that TGF-β1 concentration dependently increased protein production and mRNA expression of gelatinase MMP-9 in cultured human corneal epithelial cells (Figs. 1 2) . This TGF-β1–stimulated MMP-9 expression may mediate epithelial cell migration during wound healing and involve the pathogenesis of ocular surface diseases such as corneal ulceration and dry eye. ^{47 48 49}  

Activation of Smad and MAPK Pathways by TGF-β1

Ligands of the TGF-β superfamily are known to be unique, in that they signal through transmembrane receptor serine-threonine kinases, rather than through tyrosine kinases. The receptor complex couples to a signal transduction pathway involving a novel family of proteins, the Smads. ⁵⁰ Receptor-activated Smads (Smad2 and -3) are directly phosphorylated by activated TGF-β receptors and form heteromeric complexes with Smad4 that translocate into the nucleus and modulate the transcription of target genes. ^{51 52 53 54} In many TGF-β-regulated genes, Smad-binding sequences are located adjacent to AP-1-recognition sites. ^{55 56} Although transcriptional responses can result from direct Smad binding to DNA, more commonly, functional interaction of Smads with transcriptional cofactors and coactivators or corepressors is necessary. 

MAPK cascades are also a group of protein serine/threonine kinases that are activated in response to a variety of extracellular stimuli and mediate signal transduction from the cell surface to the nucleus. ⁵⁷ The kinetics of this response convincingly argue that it is a primary effect of receptor kinase activation, and the use of dominant negative type II TGF-β receptors demonstrates that the same TGF-β receptor complex is coupling to the MAPK pathways as it is to the Smad pathway. ¹⁷ It is likely that there are transcriptional responses that are uniquely Smad dependent, uniquely MAPK dependent, or interdependent on the interaction between the two. The Smad pathway is a nexus for cross talk with other signal transduction pathways and for modulation by many different interacting proteins. TGF-β receptors activate Smad-independent pathways that not only regulate Smad signaling, but also allow Smad-independent TGF-β responses. ⁵⁸ Different biological responses to TGF-β1 depend to various degrees on activation of either or both of these two pathways. 

MMP-9 has been observed to be regulated by TGF-β, ^{11 13 14} and protein kinases belonging to the MAPK family, ERK, JNK, and p38 pathways are found to involve the regulation of MMP-9 production. ^{59 60 61} The regulatory mechanisms responsible for the expression of MMP-9 by TGF-β1, however, are still poorly understood. Only a few reports have shown that TGF-β1 stimulates MMP-9 production through the Ras-ERK1/2 MAPK pathway in transformed keratinocytes. ⁶² In this study, TGF-β1 concentration dependently activated the phosphorylation of 60-kDa Smad2 as early as 5 minutes, with peak activity at 30 minutes. This finding is supported by a recent study showing that TGF-β–induced MMP-13 production is triggered by Smad2 protein in human osteoarthritic chondrocytes. ⁶³ In this study, we also showed that TGF-β1 concentration dependently activates JNK1/2, ERK1/2, and p38 MAPK pathways (Fig. 4 5 6) . The requirement for activation of these MAPK pathways in the TGF-β1–induced MMP-9 production was supported by experiments using MAPK specific inhibitors, SP600125, PD98059, and SB202190, which blocked activation of JNK/SAPK, ERK, and p38, respectively, and also inhibited TGF-β1–induced MMP-9 production (Fig. 7) . These findings demonstrate that the Smad2, JNK, ERK, and p38 pathways are involved in MMP-9 production in human corneal epithelial cells. 

TGF-β1–induced activation of the ERK and JNK pathways has been found to regulate Smad phosphorylation, which facilitates both its activation by the TGF-β receptor complex and its nuclear accumulation. ^{64 65} Activation of MAPK pathways by TGF-β may also affect transcription responses through direct effects on Smad-interacting transcription factors allowing convergence of TGF-β–induced Smad and MAPK pathways. The dual ability of TGF-β to activate Smads and MAPK signaling has a potential role in TGF-β–induced epithelial-to-mesenchymal transdifferentiation, which depends in part on the ERK and/or p38 MAPK pathways. ^{16 66 67} Although this convergence often results in synergy, these pathways may also counteract each other. In this study, the relationship between the TGF-β–induced Smad and MAPK pathways was not established, and it needs further studies in the future. 

Inhibition of TGF-β1–Induced MMP-9 by Doxycycline through Smad and MAPK Pathways

Doxycycline, a member of the tetracycline family, is well recognized for its therapeutic efficacy in treating ocular surface diseases, such as rosacea and sterile corneal ulceration. ^{19 20} Doxycycline has also been reported to inhibit the activity of MMPs, including MMP-9. ^{22 24 26} We have experienced clinical success in treating patients with rosacea-associated corneal epithelial erosions with doxycycline. ⁶⁸ We also reported that doxycycline at a nontoxic dose, markedly decreases the production and activity of MMP-9 stimulated by IL-1β and TNF-α in cultured human corneal epithelial cells, ^{11 26} but the effects and mechanism of doxycycline on TGF-β–stimulated MMP-9 have not been not elucidated. 

In this study, we provide evidence that doxycycline, at pharmacologically achievable nontoxic doses, concentration dependently inhibits MMP-9 protein production and mRNA expression stimulated by TGF-β1 at a concentration of 1 ng/mL, which is in the range of in vivo physiological and pathologic levels, in cultured corneal epithelial cells, as shown by zymography, activity assay, semiquantitative RT-PCR, and Northern hybridization (Fig. 1 2) . These results are supported by previous reports that doxycycline inhibits the activity of 92-kDa gelatinases in bone-metastasizing cells, ⁶⁹ human skin keratinocytes, ²² and corneal epithelial cells. ^{11 26} We further demonstrated that doxycycline inhibits TGF-β1–induced activation of Smad2, JNK1/2, ERK1/2, and p38 MAPKs and that these inhibitory effects are similar or equal to those of SP600125, PD98059, or SB202190, the inhibitors of the JNK1/2, ERK1/2, and p38 MAPK pathways, respectively (Figs. 4 5 6) . Therefore, doxycycline inhibits TGF-β1–induced MMP-9 production and activity, perhaps through the Smad and the three MAPK pathways. 

In conclusion, TGF-β1 concentration dependently increases the expression, production, and activity of MMP-9 through activation of both Smad and MAPK signaling pathways in human corneal epithelial cells. Doxycycline dose dependently inhibits this TGF-β1–stimulated MMP-9 by blocking the activation of Smad, as well as the JNK, ERK, and p38 MAPK pathways. These findings may explain the reported efficacy of doxycycline in treating sterile corneal ulceration and other ocular surface diseases in which TGF-β1 may play a role in pathogenesis. 

 Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2003.

 Supported part by National Eye Institute Grants EY11915 (SCP) and EY014553 (D-QL), an unrestricted Grant from Research to Prevent Blindness, the Oshman Foundation, and the William Stamps Farish Fund.

 Submitted for publication August 2, 2004; revised October 11, 2004; accepted November 23, 2004.

 Disclosure: H.-S. Kim, None; L. Luo, None; S.C. Pflugfelder, None; D.-Q. Li, 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 No.	Sense Primer	Antisense Primer	Probe (bp)
MMP-9	NM_004994	ATCCAGTTTGGTGTCGCGGAGC	GAAGGGGAAGACGCACAGCT	552
GAPDH	M33197	GCCAAGGTCATCCATGACAAC	GTCCACCACCCTGTTGCTGTA	498

Figure 1.

 

Zymography (A, B) and MMP-9 activity assay (C) showing MMP-9 production and activity in the conditioned media of human corneal epithelial cells treated by TGF-β1 (0.1–10 ng/mL), alone or with TGF-β1 (1 ng/mL) plus doxycycline (10 μg/mL) or a neutralizing mAb (5 μg/mL) against TGF-β1. The bands in zymograms from three experiments were scanned and quantified. Data showing relative change of MMP-9 protein (x-fold) are summarized in (B).

Figure 2.

 

Semiquantitative RT-PCR (A–D) and Northern blot (E) showing the regulation of MMP-9 mRNA in human corneal epithelial cells by TGF-β1 and doxycycline at 6 (A, E) and 24 hours (B–D). The scanned and quantified data of RT-PCR profiles treated for 24 hours from three experiments (B, C) are summarized in (D), showing the relative change (x-fold) of MMP-9 mRNA after normalization by GAPDH mRNA.

Figure 3.

 

Western blot showing the time course (A) and dose–response (B) of phosphorylated Smad2 (p-Smad2) activated by TGF-β1 and inhibited by doxycycline (B) in human corneal epithelial cells.

Figure 4.

 

Western blot showing the time course (A) and dose–response (B) of p-JNK1 and p-JNK2 activated by TGF-β1 and inhibited by (B) 10 and (C) 5 to 40 μg/mL doxycycline (Doxy) and 30 μM SP600125 (SP) in human corneal epithelial cells.

Figure 5.

 

Western blot showing the time course (A) and dose–response (B) of p-ERK-1 and p-ERK-2 activated by TGF-β1 and inhibited by (B) 10 and (C) 5 to 40 μg/mL doxycycline (Doxy) and 40 μM PD98059 (PD) in human corneal epithelial cells.

Figure 6.

 

Western blot showing the time course (A) and dose–response (B) of p-p38 activated by TGF-β1 and inhibited by (B) 10 and (C) 5 to 40 μg/mL doxycycline (Doxy) and 20 μM SB202190 (SB) in human corneal epithelial cells.

Figure 7.

 

Representative zymograms showing MMP-9 production in the conditioned media of human corneal epithelial cells treated with TGF-β1 (1 ng/mL), with or without 30 μM SP600125 (SP), 40 μM PD98059 (PD), 20 μM SB202190 (SB), or 10 μg/mL doxycycline (Doxy) for (A) 24 or (B) 48 hours or 5 to 40 μg/mL doxycycline for (C) 24 hours.

The authors thank the Lions Eye Bank of Texas for kindly providing human corneoscleral tissues.