Effect of Mitomycin C on IL-1R Expression, IL-1–Related Hepatocyte Growth Factor Secretion and Corneal Epithelial Cell Migration

Abstract

Purpose.: To investigate how mitomycin C (MMC) modulates hepatocyte growth factor (HGF) and keratinocyte growth factor (KGF) secretions in human corneal fibroblasts and regulates human corneal epithelial (HCE) cell migration.

Methods.: Primary human corneal fibroblasts were treated with MMC (0.05, 0.1, or 0.2 mg/mL for 5 minutes) and were cultivated with or without interleukin (IL)-1β. Transcript and secretion of HGF and KGF were determined by quantitative real-time RT-PCR and Western blot analysis, respectively. The effect of MMC-treated fibroblasts on HCE cell migration was evaluated using a transwell migration assay. The influence of MMC on HGF expression/secretion and HCE cell migration was further confirmed by RNA interference. The number of IL-1 receptors (IL-1R) on the fibroblast surface was analyzed by flow cytometry.

Results.: MMC alone did not affect endogenous HGF expression, whereas IL-1β alone significantly upregulated HGF transcripts and secretion. By modifying IL-1R numbers, MMC further upregulated IL-1β–related HGF expression at a concentration of 0.05 mg/mL but to a lesser extent at 0.1 and 0.2 mg/mL. KGF transcripts and intracellular expression were suppressed by MMC dose dependently in the presence or absence of IL-1β, whereas KGF secretion was not affected. Conditioned medium from MMC-treated fibroblasts exerted a similar concentration-dependent effect on HCE cell migration, enhancing migration most significantly at 0.05 mg/mL MMC in the presence of IL-1β. The MMC dose-dependent modulation of HCE cell migration was abolished in HGF-silenced fibroblasts.

Conclusions.: MMC differentially modulated IL-1R expression at various concentrations and regulated HGF and KGF differently. MMC alone did not alter HGF expression. In the presence of IL-1β, MMC-treated corneal fibroblasts modified HCE cell migration through IL-1β–induced HGF secretion.

Reepithelialization is important in corneal wound healing because the cornea is more susceptible to infection and melting in the presence of an epithelial defect. The corneal stromal-epithelial dialog plays an important role in wound healing and is mediated in part by cytokines such as interleukin (IL)-1. ¹ After wounding from photorefractive keratectomy (PRK) or laser in situ keratomileusis, the injured epithelial cells release substantial amounts of IL-1 as a master regulator, ^1,2 and they upregulate hepatocyte growth factor (HGF) and keratinocyte growth factor (KGF) in fibroblasts. ^3,4  

HGF has powerful mitogenic, motogenic, and morphogenic activities in various cell types, ^5–8 and its receptor c-Met is highly expressed on corneal epithelial cells. ^9,10 HGF increases the proliferation and migration of corneal epithelial cells ^6,11 ; moreover, it inhibits their apoptosis. ¹² It is thus implicated as a regulator of wound healing. After PRK, HGF availability is dramatically increased in cornea and tear film. ¹³ Hence, it is conceivable that fibroblast-secreted HGF plays an important role in corneal epithelial wound healing. KGF, another heparin-binding growth factor, also modulates corneal epithelial proliferation, motility, and differentiation. ⁶ Topical KGF significantly impedes epithelial migration in rabbit eyes after PRK, ¹⁴ but it accelerates epithelial wound closure after n-heptanol wounding. ¹⁵  

Mitomycin C (MMC), a conventional antimetabolite from Streptomyces caespitosus, has been widely used to prevent corneal scar formation and recurrence of subepithelial fibrosis after corneal refractive surgery. ^16–19 MMC also represses the synthesis of DNA, RNA, and protein. ²⁰ Our previous studies indicated that MMC upregulates IL-8 and monocyte chemoattractant protein-1 expression in corneal fibroblasts, ²¹ suggesting that MMC modulates corneal wound healing through pathways other than keratocyte apoptosis and myofibroblast reduction. ^22,23 The effect of MMC on corneal wound remodeling may thus be more complicated than previously expected. In addition, intraoperative MMC application slows postoperative epithelial healing to various degrees. ^18,24–26 It is thus warranted to verify whether MMC modulates epithelial wound healing through the stromal-epithelial dialog. In this study, we present data that clarify previously unknown mechanisms of MMC in corneal epithelial healing. 

Materials and Methods

Culture of Human Corneal Cells

The rims of human corneas were obtained after central corneas were used for penetrating keratoplasty. The tissues were scraped to remove residual epithelium/endothelium and were treated at 37°C with 2 mg/mL collagenase (Roche, Mannheim, Germany) in DMEM (Gibco, Grand Island, NY)/10% fetal bovine serum (FBS; Biological Industries, Kibbutz Beit Haemek, Israel) with 100 U/mL penicillin G and 100 μg/mL streptomycin until a single-cell suspension of corneal keratocytes was obtained. Isolated fibroblasts were cultured at 37°C in DMEM/10% FBS without antibiotics under humidified air with 5% CO₂ and were passaged by trypsinization. The cells were used in passages 4 to 10. 

Human corneal epithelial (HCE) cells, kindly provided by Chia-Yang Liu, were maintained in DMEM/10% FBS without antibiotics at 37°C under humidified air with 5% CO₂. The medium was replaced every other day until the cells were seeded by trypsinization for cell migration assay. 

Treatment with MMC and IL-1β

To analyze the effect of MMC on HGF expression, 60% to 70% confluent fibroblasts were treated with MMC (Kyowa, Tokyo, Japan) at 0, 0.05, 0.1, or 0.2 mg/mL for 5 minutes followed by three washes with phenol red-free DMEM (Sigma, St. Louis, MO). Cells were then incubated in the presence or absence of IL-1β (R&D Systems, Minneapolis, MN) at 1 ng/mL in DMEM/0.1% FBS for 24 to 96 hours to mimic in vivo wound healing. Conditioned medium, cell lysates, and total RNA were harvested for the subsequent experiments. 

Analysis of Transcripts by qRT-PCR

To analyze the gene transcripts, total RNA was isolated with an RNA extraction reagent (REzol C&T; ProTech Technology, Taipei, Taiwan). The cDNA was prepared from 2 μg total RNA by reverse transcription at 42°C for 2 hours using oligo-(dT)₁₅ and MMLV reverse transcriptase (Promega, Madison, WI). A fluorescein PCR detection system (Clontech, Mountain View, CA) was used to determine changes of HGF, KGF, and GAPDH with their own specific primers (HGF, 5′-AGACTACATTAGAAACTGC-3′ and 5′-CTTGTGAAACACCAGG-3′; KGF, 5′-ACAAGGAAGGAAAACTC-3′ and 5′-GTTATTGCCATAGGAAGAAA-3′; GAPDH, 5′-CACCACCAACTGCTTAG-3′ and 5′-CTTCACCACCTTCTTGATG-3′). PCR was set at 94°C for 15 minutes, followed by 40 cycles at 94°C for 5 seconds, 55°C for 5 seconds, and 72°C for 15 seconds. The relative amount of mRNA was calculated by the second-derivative comparative Ct method and relative quantification with software (LightCycler, version 3.0; Roche, Indianapolis, IN). Quantitative results of HGF and KGF were normalized to GAPDH. 

Analysis of Extracellular and Intracellular Proteins by Western Blot Analysis

To analyze protein secretions, conditioned media were collected after the removal of suspended cells by centrifugation. To analyze intercellular proteins, cell lysates were harvested in lysis buffer, as described. ²⁷ The lysates were quantified (BCA Protein Assay; Pierce, Rockford, IL), and equal loading of conditioned media was subjected according to total protein of its own lysate. The lysates and conditioned media were subjected to the same volume with equality and were mixed with one-third volume of 3× sample buffer, as described. ²⁷ These samples were analyzed by SDS-PAGE and blotted onto PVDF membranes (Millipore, Bedford, MA). After blocking the nonspecific binding with 5% bovine serum albumin (BSA) in TBST, the membranes were incubated in 5% BSA/TBST with anti-HGF (H-145, 1:1000; Santa Cruz Biotechnology, Santa Cruz, CA), anti-KGF (H-73, 1:1000; Santa Cruz Biotechnology), or anti-β-actin (AC-15, 1:5000; Sigma) antibodies. The membranes were washed with TBST and incubated with horseradish peroxidase-conjugated secondary antibody (1:10,000; Santa Cruz Biotechnology). After three washes with TBST, blots were developed by chemiluminescence (Millipore), and images were captured with a Fujifilm imaging system (LAS-4000; Fujifilm, Tokyo, Japan). 

IL-1R Expressions on Corneal Fibroblasts by Flow Cytometry

To analyze the amount of IL-1 receptors (IL-1R) on corneal fibroblasts, cells were treated with MMC at 60% to 70% confluence followed by IL-1β incubation. The cells were harvested and fixed with 4% paraformaldehyde/phosphate-buffered saline (PBS) for 10 minutes. After washing with PBS, the fixed cells were blocked for 30 minutes in PBS containing 0.2% Triton X-100 and 10 mg/mL BSA. They were hybridized with goat anti–human IL-1RI antibody (50 μg/mL; R&D Systems), washed with PBS, and labeled with Alexa Fluor 488 anti–goat second antibody (50 μg/mL; Invitrogen, Carlsbad, CA). FL1 intensity was detected by flow cytometry (FACSCalibur; BD Biosciences, San Jose, CA) and analyzed by cytometry software (WinMDI). 

HCE Cell Migration Assay

To study the effect of corneal fibroblast secretions on HCE cell migration, 60% to 70% confluence fibroblasts in 24-well plates were treated with MMC followed by incubation with or without IL-1β. Before 24 hours posttreatment (hpt), HCE cells were suspended by trypsinization and adjusted to 1.0 × 10⁵ cells/mL in DMEM/0.1% FBS. HCE cell suspension (500 μL) was seeded into 8-μm transwell culture inserts (Nunc, Rochester, NY), and HCE cell migration was tested by cocultivation for another 48 hours with MMC-treated fibroblasts starting at 24 hpt. Migrated HCE cells were stained with DAPI and photographed under a fluorescence microscope at 100-fold magnification. 

HGF Silencing in Corneal Fibroblasts by Lentiviral shRNA Infection

To silence HGF expression, fibroblasts were treated with the lentiviral short-hairpin RNA (shRNA) pseudovirions of HGF from the National RNAi Core Facility (Academia Sinica, Taiwan), a member of the RNAi Consortium (TRC). ²⁸ Briefly, fibroblasts were infected for 3 days with shHGF or vehicle pseudovirions at 5 MOI, packaged from 293T cells with shHGF or vehicle clones by transfection (Fugene 6; Roche). After checking the infection efficiency of five shHGF clones, the highest efficient clone (TRCN0000003307, 5′-CAGACCAATGTGCTAATAGAT-3′) was chosen. The infected fibroblasts were used in cell migration assay. 

Statistical Analysis

Differences in transcripts and migrated HCE cell numbers under various conditions were evaluated by one-way ANOVA. P < 0.05 was considered statistically significant and was further tested with Dunnett's post hoc test. 

Results

Regulation of IL-1β–Induced HGF Secretion in Corneal Fibroblasts

IL-1β induces HGF secretion in various human fibroblasts, including cornea, ^3,15,29 skin, ³⁰ lung, ³¹ and stomach. ³² To determine how MMC affects HGF expression in corneal fibroblasts, we treated corneal fibroblasts with MMC and incubated them in the presence or absence of IL-1β. Quantitative RT-PCR and Western blot analysis showed that MMC alone did not modulate HGF transcripts (Fig. 1A) or secretion (Fig. 1B) at 72 hpt. In contrast, IL-1β significantly upregulated HGF transcripts by 3.29-fold. In the presence of IL-1β, MMC further enhanced HGF transcripts by an additional 35.6% (i.e., 4.46-fold at 0.05 mg/mL). This additional effect was less significant at higher MMC concentrations, namely 4.00-fold at 0.1 mg/mL and 2.01-fold at 0.2 mg/mL (Fig. 1A). IL-1β also upregulated HGF secretion (Fig. 1B). HGF secretion was further enhanced by 0.05 mg/mL MMC but was less upregulated by 0.1 or 0.2 mg/mL MMC in the presence of IL-1β. However, MMC did not change the cytosolic HGF expression at all tested concentrations up to 72 hpt, even in the presence of IL-1β (Fig. 1B, middle panel). 

Figure 1.

 

MMC regulates HGF mRNA level and secretion in IL-1β–incubated human corneal fibroblasts dose dependently. Human primary corneal fibroblasts were treated with the indicated concentration of MMC for 5 minutes and then were incubated with or without 1 ng/mL IL-1β in DMEM/0.1% FBS. Treated cells and conditioned medium were harvested at 72 hpt and were analyzed by (A) qRT-PCR with an HGF-specific primer set and (B) Western blotting with anti–HGF or anti–β-actin antibodies. (C) Upregulation of HGF secretion is dependent on IL-1β incubation time. Conditioned media were harvested at the indicated time points and analyzed by Western blotting with anti–HGF antibody. (A) Mean ± SEM of triplicates from an experiment that was repeated three times with similar results. **P < 0.01 and *P < 0.05 compared with IL-1β–incubated cells.

To further investigate the temporal effect of MMC on HGF secretion, we harvested culture medium from MMC-treated fibroblasts at 24 to 96 hpt (Fig. 1C). Western blot analysis revealed no significant HGF secretion up to 96 hpt after MMC treatment alone. In contrast, HGF secretion was detected after 48 hpt in the presence of IL-1β, and it showed the same dose-dependent response to MMC after 72 hpt. These results suggested that a single exposure to MMC regulated HGF secretion in a concentration-dependent pattern in the presence of IL-1β. Furthermore, this was sustained and especially significant after 72 hpt with MMC. 

Alteration of IL-1R on the Fibroblast Surface

IL-1R may play an important role in mediating IL-1β–induced HGF section in corneal fibroblasts. ⁴ Our results indicated that MMC modulated HGF secretion in IL-1β–incubated corneal fibroblasts (Figs. 1B, 1C). Therefore, we tested whether MMC modified HGF secretion by changing the IL-1R on the fibroblast surface. Using flow cytometry, we determined signal intensity of labeled IL-1R with a fluorescent dye (Fig. 2). We found that 0.05 mg/mL MMC significantly enhanced signal intensity at 48 hpt (Fig. 2A). In contrast, signal intensity after 0.2 mg/mL MMC was slightly lower than that after 0.05 mg/mL MMC, though it was still higher than that in the non-MMC group. IL-1β also enhanced signal intensity (Fig. 2B). In the presence of IL-1β, 0.05 mg/mL MMC further increased signal intensity slightly more than 0.2 mg/mL MMC. These findings suggest that MMC may modulate HGF secretion by changing IL-1R expression on corneal fibroblasts. 

Figure 2.

 

MMC enhances IL-1R expression on the fibroblast surface. (A) In the absence or (B) presence of IL-1β incubation after MMC treatment, the corneal fibroblasts were hybridized with IL-1R–specific antibody, stained with a fluorescent Alexa Fluor 488–conjugated second antibody, and analyzed by flow cytometry. This experiment was repeated three times with similar results.

Suppression of KGF Transcripts and Cytosolic KGF Expression in Corneal Fibroblasts

IL-1β stimulates KGF expression in human fibroblasts, including breast, ³³ cornea, ^3,15,29 lung, ³⁴ and skin. ³⁵ To examine whether MMC also affects KGF expression in corneal fibroblasts, we treated fibroblasts with MMC and IL-1β, as described, and analyzed KGF transcripts by qRT-PCR and KGF secretion and cytosolic protein by Western blotting (Fig. 3). Quantitative RT-PCR revealed that IL-1β upregulated KGF transcripts by 3.10-fold, whereas MMC downregulated KGF transcripts in the presence and absence of IL-1β (Fig. 3A). Western blot analysis showed that MMC also reduced endogenous KGF expression in IL-1β–stimulated fibroblasts (Fig. 3B, middle panel). However, KGF secretion was undetectable in conditioned medium after MMC treatment, even in the presence of IL-1β (Fig. 3B, top panel). There was no detectable KGF secretion on Western blot after MMC treatment (data not shown) up to 96 hours of IL-1β incubation. These results revealed that MMC downregulated KGF transcripts and endogenous KGF protein but had no noticeable effect on KGF secretion. 

Figure 3.

 

MMC reduces intracellular expression of KGF mRNA and protein in human corneal fibroblasts. Primary corneal fibroblasts were treated with MMC at the indicated concentration for 5 minutes and then incubated in DMEM/0.1% FBS with or without IL-1β. Treated cells and conditioned medium were harvested at 72 hpt and analyzed by (A) qRT-PCR with a KGF-specific primer set and (B) Western blotting with anti–KGF or anti–β-actin antibodies. (A) Mean ± SEM of triplicates from an experiment that was repeated three times with similar results. **P < 0.01 and *P < 0.05 compared with IL-1β–incubated cells.

Dose Dependent Effect of MMC-Regulated HGF Secretion in IL-1β–Incubated Corneal Fibroblasts on Corneal Epithelial Cell Migration

HGF modulates HCE cell migration ⁵ and proliferation ¹¹ and has anti-apoptotic effect as well. ¹² Our in vitro migration assay (Fig. 4A) showed that HCE cell migration was minimal in culture medium supplemented with 0.1% FBS in the absence or presence of IL-1β (Figs. 4B, 4C). However, the addition of recombinant HGF to culture medium increased HCE cell migration by 10.9-fold. Because MMC modulated HGF secretion (Figs. 1B, 1C), we investigated whether MMC-induced HGF secretion affected HCE cell migration. The addition of corneal fibroblasts with (Fig. 5) or without (data not shown) vehicle infection significantly promoted HCE cell migration. The conditioned medium from vehicle-infected MMC-treated fibroblasts did not further affect HCE cell migration (Fig. 5B, second column); however, IL-1β–incubated corneal fibroblasts increased HCE cell migration by 25.9% (Fig. 5B, third column). When fibroblasts were pretreated with 0.05 mg/mL MMC followed by IL-1β incubation, HCE cell migration was increased by an additional 11.9% (Fig. 5B, fourth column). Nevertheless, HCE cell migration was lower at MMC 0.1 and 0.2 mg/mL than at 0.05 mg/mL MMC (Fig. 5B, fifth and sixth columns). In contrast, fibroblast-conditioned medium supplemented with 10 ng/mL recombinant HGF increased HCE cell migration by 12.4% (Fig. 5B, last column). 

Figure 4.

 

Recombinant HGF is sufficient for HCE cell migration. (A) Scheme of migration assay for HCE cell migration from top to bottom of culture inserts. HCE cells were seeded into culture inserts with DMEM/0.1% FBS and cultivated in DMEM/0.1% FBS with various additives. Migrated HCE cells were stained with DAPI for convenient counting. (B) Counts of migrated HCE cells and (C) DAPI staining after 48-hour incubation in DMEM/0.1% FBS with 1 ng/mL IL-1β or 10 ng/mL HGF. Migrated HCE cells were counted and averaged from five random fields in an experiment that was repeated three times with similar results. **P < 0.01 versus the cells in DMEM/0.1% FBS. Scale bar, 0.1 mm.

Figure 5.

 

MMC-regulated HGF secretion of the IL-1β–stimulated corneal fibroblasts affects HCE cell migration dose dependently. (A) The migration assay was performed as described in Figure 3, with the addition of corneal fibroblasts pretreated with MMC and IL-1β. (B) Counts and (C) DAPI images of migrated HCE cells in conditioned medium of MMC-treated corneal fibroblasts at 48 hours after incubation. Migrated HCE cells were counted and averaged from five random fields in an experiment that was repeated three times with similar results. **P < 0.01 versus IL-1β–incubated cells. Scale bar, 0.1 mm.

To determine whether KGF modulated HCE cell migration in our model, we conducted migration assay, as described above, substituting recombinant KGF for HGF. Our results revealed no obvious difference between untreated and treated conditions with recombinant KGF (data not shown). These data revealed that KGF might not directly modulate HCE cell migration in our model. 

Effect of HCE Cell Migration in Conditioned Medium from HGF-Silenced Fibroblasts

Finally, we studied whether reducing HGF secretion from corneal fibroblasts would mitigate MMC-dependent changes in HCE cell migration. HGF gene knockdown was achieved with the shRNA system. At 3 days of IL-1β incubation 3 days after infection with shHGF pseudovirions, shHGF silencing efficiency was analyzed by qRT-PCR (Fig. 6A) and Western blotting (Fig. 6B). We achieved a 94% decrease in HGF transcripts and HGF secretion, even in the presence of IL-1β, indicating that shHGF infection affected the entire transcription/translation and secretory pathway of HGF. MMC dose-related HGF secretions were also confirmed by Western blot analysis in the presence and absence of IL-1β incubation for 72 hours (Fig. 6C). At 72 hpt with MMC, we also confirmed that MMC-related HGF secretion in vehicle-infected fibroblasts was similar to that in uninfected fibroblasts (Fig. 1B). Therefore, we compared HCE cell migration using conditioned medium from vehicle-infected (Fig. 5B) and HGF-silenced MMC-treated fibroblasts (Fig. 6D) by transwell migration assay. We found that HCE cell migration was still increased by 5.1% in conditioned medium from HGF-silenced fibroblasts in the presence of IL-1β (Fig. 6D, third column) compared with no IL-1β incubation. This was significantly less than the 25.9% increase observed in the vehicle-infected cells (Fig. 5B). Similarly, the MMC-related change in HCE cell migration seen in the vehicle-infected cells was almost abolished using HGF-silenced fibroblasts. Taken together, these results suggest that MMC-dependent HGF secretion from corneal fibroblasts played an important role in IL-1β–related HCE cell migration. 

Figure 6.

 

HGF-silenced corneal fibroblasts do not affect HCE cell migration. (A) HGF transcripts and (B) endogenous HGF expression in HGF-silenced corneal fibroblasts with or without IL-1β incubation for 72 hours were determined separately by qRT-PCR and Western blot analysis. HGF secretion in corneal fibroblasts was also suppressed by shHGF pseudovirion infection. Both vehicle and HGF-silenced corneal fibroblasts were treated with MMC with or without IL-1β incubation. (C) Secretion of HGF by the corneal fibroblasts was determined at 72 hpt of MMC by Western blot analysis with HGF-specific antibody. (D) The HCE cell migration assay was performed using HGF-silenced cells. (E) Some critical DAPI images are shown. (A) Mean ± SEM of triplicates from an experiment that was repeated three times with similar results. **P < 0.01 compared with normal cells. (C) Migrated HCE cells were counted and averaged from five random fields in an experiment that was repeated three times with similar results.**P < 0.01 and *P < 0.05 versus IL-1β–incubated shHGF-silenced cells. Scale bar, 0.1 mm.

Discussion

MMC did not affect HGF secretion in corneal fibroblasts in the absence of IL-1β. However, it regulated IL-1β–induced HGF secretion dose dependently. The change of HGF secretion in corneal fibroblasts was able to functionally modulate corneal epithelial migration. A scheme for how MMC modulates corneal stromal-epithelial interactions during wound healing is shown in Figure 7. In brief, IL-1β secretion in injured corneal epithelium (step 1) triggers HGF secretion in corneal fibroblasts by binding to IL-1R on the fibroblast surface (step 2). The paracrine effect of fibroblast-released HGF on corneal epithelial migration (step 3) leads to corneal epithelial healing (step 4). Addition of a low dose of MMC (0.05 mg/mL) significantly enhances HGF secretion in corneal fibroblasts by further increase of IL-1R (step 5a). The increase in HGF secretion promotes corneal epithelial migration. On the other hand, higher doses of MMC (0.1 and 0.2 mg/mL) promote IL-1β–induced HGF secretion to a lesser extent though a lower increase in IL-1R compared with 0.05 mg/mL MMC (step 5b) and, consequently, slow down corneal epithelial migration. It is possible that MMC changed IL-1R synthesis and degradation when it was applied at different concentrations. When applied at a low concentration (0.05 mg/mL), it affected IL-1R degradation first, resulting in overall higher numbers of IL-1R. MMC at a higher concentration (0.2 mg/mL) probably affected both IL-1R synthesis and degradation, resulting in less IL-1R than with 0.05 mg/mL, though it was still higher than in untreated cells. This alteration in IL-1R should have contributed to the altered HGF levels in our IL-1β–induced HGF secretion model. The molecular mechanisms through which MMC modulated IL-1R synthesis and degradation await further elucidation. 

Figure 7.

 

Scheme depicting MMC dose-dependent regulation of corneal epithelial migration through HGF secretion by modifying IL-1R on corneal fibroblasts.

Although MMC altered HGF transcripts and secretion in the presence of IL-1β, it did not change intracellular HGF protein homeostasis in either condition (Fig. 1B, middle panel). We propose that MMC-treated cells secreted the newly translated HGF protein into culture medium without the compromising HGF secretion mechanism. This resulted in a change in HGF secretion but comparable protein levels in cell lysates at concentrations of 0.05 to 0.2 mg/mL. Interestingly, the alteration in HGF expression had a sustained effect with a noticeable change in HGF secretion at 72 hpt with MMC. 

Some studies have shown that topical MMC application delays corneal epithelial healing ^18,24,26 by prolonged latency and decreased migration rates ¹⁸ and by delayed restoration of normal cellular appearance, ²⁴ whereas others have shown that MMC does not interfere with epithelial closure. ²⁵ However, no molecular mechanism has been proposed to account for how MMC impedes epithelial healing. In this study, we demonstrated for the first time that MMC could modulate corneal epithelial migration by modifying IL-1β–related HGF secretion in corneal fibroblasts, which subsequently exerts its paracrine effect on corneal epithelium. This MMC concentration-dependent effect might contribute to the variations in the reported results regarding corneal epithelial closure after MMC application because MMC concentrations and application durations studied are not uniform. ^18,24–26 Our results further support the current trend of applying a lower MMC dosage (0.02 mg/mL or 0.002%) during refractive surgeries to avoid the potential MMC-related side effects of delaying epithelial wound healing. ^36–38 However, whether lower MMC dosage used during refractive surgeries would facilitate postoperative epithelial closure awaits further clinical verification. 

We demonstrated that IL-1β upregulated KGF transcripts and endogenous KGF expression while MMC reduced both in a dose-dependent manner, either in the presence or absence of IL-1β. Although KGF secretion has been shown to regulate corneal epithelial migration, ¹⁰ we found no obvious changes in the number of migrated epithelial cells after coincubation with shHGF-silenced or vehicle-infected corneal fibroblasts, even when they were treated with MMC (Fig. 6D) because MMC did not change KGF secretion (Fig. 3B). In these cells, we found a 5.1% increase of migrated HCE cells (Fig. 6D, third column) compared with no IL-1β incubation. It could be attributed to the incomplete silence of HGF, which expressed approximately 6% HGF transcripts and detectable endogenous HGF protein (Figs. 6A, 6B). Additionally, recombinant KGF-enriched medium did not change the extent of epithelial migration (data not shown). These imply that KGF is not a critical factor for corneal epithelial migration in our model and support the notion that HGF rather than KGF promotes corneal epithelial migration by PKCα translocation to plasma membrane. ^6,39 The physiological functions of MMC-related downregulation of cytosolic KGF require further exploration. Given that cultured corneal fibroblasts are able to secrete IL-6 ⁴⁰ and epithelial growth factor ^15,41,42 to modulate epithelial migration, ^42–44 whether they were involved in the alteration of HCE cell migration in our model (Figs. 4B, 5B) also deserves further elucidation. 

MMC at 0.2 mg/mL induced some, but not much, apoptosis in our corneal fibroblast system. More important, it inhibited cell proliferation. Both factors contributed to the overall lower numbers of corneal fibroblasts in the 0.2-mg/mL MMC–treated groups. Moreover, the two factors also contributed to the lower levels of secreted HGF and stimulated epithelial migration in the 0.2-mg/mL MMC–treated fibroblasts in the presence of IL-1β. Given that MMC-treated tissues remain less cellular for up to several months after surgery, including the early postoperative phase when corneal epithelial migration is ongoing, we did not adjust for the corneal fibroblast cell number in our system because this might have mimicked the in vivo scenario better. It is also possible that MMC induced an IL-1R–independent overall suppression of cell function that resulted in further decreased HGF expression disproportional to the alteration in IL-1R expression. 

This is the first study demonstrating that MMC modulates corneal epithelial migration by affecting secretion in corneal fibroblasts in the presence of IL-1β, a condition simulating surgical wounding. MMC-related effects might not manifest in quiescent corneal fibroblasts. However, MMC-treated cells might react differently from untreated cells in the presence of wound-healing cytokines such as IL-1β and might manifest a sustained effect. Because MMC-dependent keratocyte loss does not vanish up to 3 months after PRK ^18,23 and these MMC-treated cells might react differently during wound healing, ⁴⁵ the MMC-dependent wound modulation effect could last longer than previously expected. HGF is involved in the production and subsequent migration of transient amplifying cells for cell renewal, whereas KGF is involved in limbal stem cell division. ²⁹ Intraoperative MMC application during refractive surgery could possibly interfere with corneal epithelial closure if higher dosages are used. Potential limbal deficiency is also possible if MMC is applied to the limbal area through subconjunctival ^46,47 or topical routes. ⁴⁸ Confirmation of our results in vivo may be necessary because cells in tissues may respond to MMC differently from cultured cells. Further studies are warranted to verify whether MMC can trigger other regulators in corneal fibroblasts to modulate corneal epithelial migration. Additional understanding of MMC-related molecular mechanisms could facilitate procedural modifications and improve outcomes of surgeries incorporating the wound-healing modulator. 

Footnotes

 Supported in part by Far Eastern Memorial Hospital Grant FEMH-96-D-002.

Footnotes

 Disclosure: T.-C. Chen, None; S.-W. Chang, None