Role of the Small GTP-Binding Protein Rho in Epithelial Cell Migration in the Rabbit Cornea

Abstract

purpose. To determine the role of the small guanosine triphosphate (GTP)-binding protein Rho in the migration of corneal epithelial cells.

methods. The presence of the Rho target proteins Rho-associated coiled coil-containing protein kinase (ROCK)-1 and ROCK-2 in rabbit cornea was examined by immunohistochemical analysis, and that of the corresponding mRNAs in rabbit corneal epithelial cells was determined by reverse transcription–polymerase chain reaction analysis. The effects of various agents on epithelial cell migration were investigated by measuring the length of the migration path in rabbit corneal blocks in culture.

results. Both ROCK-1 and ROCK-2 were detected in the rabbit corneal epithelium at both protein and mRNA levels. The Rho activator lysophosphatidic acid (LPA) stimulated corneal epithelial migration in a dose-dependent manner, whereas exoenzyme C3, a Rho inhibitor, inhibited epithelial migration also in a dose-dependent manner. The stimulatory effect of LPA on corneal epithelial migration was prevented by exoenzyme C3. Both cytochalasin B, an inhibitor of actin filament assembly, and ML-7, an inhibitor of myosin light chain kinase, also prevented LPA stimulation of epithelial migration.

conclusions. These results suggest that Rho mediates corneal epithelial migration in response to external stimuli by regulating the organization of the actin cytoskeleton.

Many disorders of the corneal epithelium are caused by delayed epithelial migration. It is therefore important to understand the mechanism that underlies successful migration of the corneal epithelium. Proteins that likely contribute to such migration include: extracellular signaling molecules, such as cytokines and growth factors; components of the intracellular apparatus that mediates cellular motility; and extracellular matrix proteins that act as a substrate. ^{1 2 3}  

Intracellular cytoskeletal elements, including actin filaments, microtubules, and intermediate filaments, play important roles in the maintenance of cell shape, in cell motility, and in cell adhesion. ^{4 5} However, the dynamic reorganization of actin filaments is predominantly responsible for such processes, and the regulation of these filaments is therefore likely a critical factor in the migration of corneal epithelial cells. 

The Rho family of small guanosine triphosphate (GTP)-binding proteins has recently been implicated in cell migration. ^{6 7 8 9} Rho is activated in response to the initial attachment of cell surface integrins to extracellular matrix proteins, which is an important early step in corneal epithelial migration. Stimulation of cells by specific extracellular signals results in the conversion of Rho from its inactive, guanosine diphosphate (GDP)-bound form to its active, GTP-bound form. Activated Rho then regulates the formation of both actin filaments and cell adhesion complexes through the activation of target proteins such as Rho-associated coiled coil-containing protein kinase (ROCK) and the myosin-binding subunit of myosin phosphatase. Rho therefore likely participates in integrin-mediated cell-substrate adhesion during epithelial migration. SundarRaj et al. ¹⁰ showed that ROCK-1 is localized in the epithelium of the cornea, and suggested that this protein may play an important role in corneal differentiation and maintenance. However, the role of Rho and its target proteins in corneal epithelial cell migration remain unclear. 

We have now investigated the possible contribution of Rho to corneal epithelial migration. In particular, we have examined the effects of an activator and an inhibitor of Rho on the length of the path of epithelial migration in rabbit corneal blocks maintained in culture. 

Methods

Materials

Albino rabbits with body masses of 2 to 3 kg were obtained from Kitayama Labs (Kyoto, Japan). Care and treatment of animals conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Normal rabbit corneal epithelial cells and rabbit corneal growth medium (RCGM) were from Kurabo (Osaka, Japan). TC-199 culture medium was obtained from the Research Foundation for Microbial Diseases of Osaka University (Suita, Osaka, Japan), lysophosphatidic acid (LPA) and cytochalasin B from Sigma (St. Louis, MO), exoenzyme C3 transferase of Clostridium botulinum (exoenzyme C3) from Upstate Biotechnology (Lake Placid, NY); and ML-7 from Calbiochem (La Jolla, CA). 

LPA was dissolved in 20% ethanol at a concentration of 20 mM, and this stock solution was stored at −80°C and diluted with unsupplemented TC-199 before use. Exoenzyme C3 was dissolved in unsupplemented TC-199 before use. Cytochalasin B and ML-7 were dissolved in dimethyl sulfoxide at concentrations of 10 mg/ml and 1 mM, respectively, and these stock solutions were stored at −80°C and diluted with unsupplemented TC-199 before use. Vehicle preparations had no effect on corneal epithelial migration. 

Immunohistochemistry for ROCK-1 and ROCK-2

Rabbits were killed by intraperitoneal injection of sodium pentobarbital. The eyes were enucleated, embedded immediately in optimal cutting temperature compound (OCT; Miles, Elkhart, IN), and frozen in a mixture of acetone and dry ice. Sections (7 μm thick) were cut from each eye with a microtome cryostat (CM3000; Leica, Wetzlar, Germany) and mounted on silane-treated glass slides. They were then fixed with ice-cold acetone for 10 minutes, rinsed with phosphate-buffered saline (PBS), and incubated for 30 minutes at room temperature with 0.3% H₂O₂ in methanol to quench endogenous peroxidase activity. After incubation for 1 hour at room temperature with PBS containing 1% bovine serum albumin (fraction V; Sigma) and 1.5% horse serum to block nonspecific binding sites, the sections were washed with PBS and incubated overnight at 4°C in a moist chamber with goat antibodies to ROCK-1 (K-18) or to ROCK-2 (C-20; Santa Cruz Biotechnology, Santa Cruz, CA) diluted 1:1000 (0.2 μg/ml) with blocking solution. For staining controls, normal goat immunoglobulin G (Zymed, San Francisco, CA) at a concentration of 0.2 μg/ml in blocking solution was used in place of the specific primary antibodies. All specimens were then rinsed three times with PBS containing 0.05% Tween-20 before incubation for 30 minutes at room temperature with secondary antibodies (Vectastain Elite ABC kit; Vector, Burlingame, CA). They were again rinsed three times with PBS containing 0.05% Tween-20, incubated for 30 minutes at room temperature with the ABC reagent and for an additional 10 minutes in the presence of the diaminobenzidine (DAB) substrate solution, rinsed with deionized water, and counterstained with hematoxylin. The sections were finally mounted and examined with an epifluorescence microscope (BX50; Olympus, Tokyo, Japan). Photographs were taken with reversal film (Fujichrome Provia 100; ISO 100; Fuji Film, Tokyo, Japan). 

RT-PCR Analysis of ROCK-1 and ROCK-2 mRNAs

Normal rabbit corneal epithelial cells were cultured in RCGM until they achieved 70% to 80% confluence. Total RNA was then extracted from the cells with the use of Isogen (Nippongene, Toyama, Japan) and quantitated spectrophotometrically by measurement of absorbance at 260 and 280 nm. The abundance of ROCK-1 and ROCK-2 mRNAs was estimated by reverse transcription and polymerase chain reaction (RT-PCR) analysis with a Takara RNA PCR kit (AMV; Takara Shuzo, Otsu, Shiga, Japan). Total RNA (1 μg) was subjected to RT by incubation with 5 U of avian myeloblastosis virus reverse transcriptase XL in a reaction mixture (20 μl) containing 2 μl of 10 × RNA PCR buffer, 5 mM MgCl₂, 1 mM of each deoxynucleoside triphosphate, random nine-nucleotide oligomers (2.5 picomoles/μl), and 20 U of RNase inhibitor. The reaction mixture was incubated at 30°C for 10 minutes and at 42°C for 30 minutes, and the reaction was then terminated by incubation at 99°C for 5 minutes and then cooling to 4°C. PCR was performed with (GeneAmp PCR System 9600; Perkin-Elmer, Foster City, CA) by adding 49 μl of a reaction mixture containing 5 μl of 10 × LA PCR buffer, 2.5 U of LA Taq polymerase, 2.5 mM MgCl₂, 2.5 mM of each deoxynucleoside triphosphate, and 50 picomoles each of sense and antisense primers to 1 μl of the RT-generated cDNA. The primers used for PCR reactions were as follows: ROCK-1 sense, 5′-TGCGGGAGTTACAAGATCAGCT-3′; ROCK-1 antisense, 5′-TTTCCGTCAGTCTCATCAGCAC-3′; ROCK-2 sense, 5′-TCTGAAAGGAGGGACCGAACC-3′; ROCK-2 antisense, 5′-GTTCCTGTTTGTGTCGAGCCATCA-3′; glyceraldehyde-3-phosphate dehydrogenase (G3PDH, internal control) sense, 5′-ACCACAGTCCATGCCATCAC-3′; and G3PDH antisense, 5′-TCCACCACCCTGTTGCTGTA-3′. These primers generated the expected PCR products of 828 bp for ROCK-1 cDNA, 996 bp for ROCK-2 cDNA, and 450 bp for G3PDH cDNA. The PCR protocol comprised an initial incubation for 5 minutes at 94°C; 30 cycles (for ROCK-1 and ROCK-2) or 25 cycles (for G3PDH) of 45 seconds at 94°C, 45 seconds at 5 5°C, and 2 minutes at 72°C and a final incubation for 7 minutes at 72°C. The PCR products were subjected to electrophoresis on a 1% agarose gel (Sigma) containing ethidium bromide (1 μg/ml; Nippongene) and visualized with an ultraviolet transilluminator. The identity of the PCR product for ROCK-1 was confirmed by DNA sequencing. Although the sequence of rabbit ROCK-2 cDNA has not been described, the DNA sequence of the PCR product for ROCK-2 was more than 90% identical with those of human, rat, and mouse ROCK-2 cDNAs, suggesting that we had indeed amplified rabbit ROCK-2 cDNA in our analysis. 

Detection of Actin Stress Fiber Formation

Rabbit corneal epithelial cells were plated on four-well Laboratory-tech chamber slides (Nunc, Napierville, IL) and maintained in unsupplemented TC-199 medium for 72 hours before treatment with LPA (2 μM) for 1 hour. The cells were then washed with PBS and fixed with ice-cold acetone for 5 minutes After washing with PBS, the slides were incubated at room temperature first for 20 minutes with PBS containing 1% bovine serum albumin, and then for 30 minutes in a moist chamber with the same solution containing rhodamine-conjugated phalloidin (1:100 dilution; Molecular Probes, Eugene, OR). Finally, the cells were rinsed three times with PBS containing 0.05% Tween-20 and examined with an epifluorescence microscope (BX50). Photographs of random fields were taken with reversal film (Provia 100; Fujichrome). 

Epithelial Migration Assay

The length of the path of epithelial migration over the corneal stroma exposed on the sides of a block of cultured rabbit cornea was measured as described previously. ^{11 12} In brief, rabbits were killed with an overdose of pentobarbital sodium injected intravenously and both eyes were enucleated. The sclerocorneal rim was cut, and the cornea was excised and washed several times with sterile PBS. Six blocks (each ∼2 × 4 mm) were cut from each cornea with a razor blade. The size of the corneal blocks was previously shown not to affect the rate of epithelial migration. ¹¹ Each corneal block was placed epithelial side up in a well of a 24-well tissue culture plate with unsupplemented TC-199 culture medium (control) or TC-199 containing the various test agents (LPA at concentrations of 0.02, 0.2, or 2 μM; exoenzyme C3 at 1, 2, or 4 μg/ml; cytochalasin B at 0.03, 0.1, or 0.3 μg/ml; or ML-7 at 0.01, 0.1, or 1 μM). Each treatment group consisted of three blocks, each from a different cornea. After incubation for 24 hours at 37°C under a humidified atmosphere of 5% CO₂ in air, blocks were fixed overnight at 4°C with a mixture of glacial acetic acid and absolute ethanol (5:95, vol/vol). They were then dehydrated by exposure to a graded series of ethanol solutions, immersed in xylene, and embedded in paraffin. Three thin sections (4 μm) were cut at 250-μm intervals from each block and, after the removal of paraffin, stained with hematoxylin-eosin. The sections were examined with a light microscope and photographed, and the length of the path of corneal epithelial migration down both sides of each section (toward the endothelial side) was measured from the photographs. Preliminary experiments had shown that the lengths of the paths of epithelial migration down each side of a corneal block are independent of each other. We therefore averaged the results obtained from the three sections for each side of each block separately. Data are expressed as means ± SEM of the six determinations (one averaged value for each side of each of the three blocks in a treatment group). Experiments were performed in a double-blind manner to prevent any bias. 

Statistical Analysis

Statistical analysis was performed with the unpaired Student’s t-test for comparison of two groups and by the Dunnett multiple comparison test for comparison of three or more groups. 

Results

Expression of ROCK-1 and ROCK-2 in Rabbit Corneal Epithelium

The expression of the Rho target proteins ROCK-1 and ROCK-2 in rabbit cornea was examined by immunohistochemical analysis. Immunoreactivity corresponding to each protein was detected in all cell layers of the corneal epithelium (Fig. 1) , being most abundant in the basal cell layers. Positive staining for ROCK-1 and ROCK-2 was also apparent in keratocytes and corneal endothelial cells. Our detection of ROCK-1 in the corneal epithelium is thus consistent with the similar observation of SundarRaj et al. ¹⁰  

The presence of mRNAs encoding ROCK-1 and ROCK-2 in cultured rabbit corneal epithelial cells was examined by RT-PCR. The cells were shown to contain both ROCK-1 and ROCK-2 transcripts (Fig. 2) . 

Effect of LPA on Actin Stress Fiber Assembly

We examined the effect of the Rho activator LPA on the formation of actin stress fibers in cultured rabbit corneal epithelial cells. Actin stress fibers were labeled with rhodamine-conjugated phalloidin and observed by epifluorescence microscopy (Fig. 3) . Exposure of cells to 2 μM LPA for 1 hour induced a marked increase in the number of actin stress fibers. 

Effects of Rho Inhibition and Activation on Corneal Epithelial Migration

We next investigated the effect of LPA on epithelial migration in cultured corneal blocks. Histologic photographs revealed that incubation of the tissue with LPA for 24 hours induced a dose-dependent increase in the extent of epithelial migration (Fig. 4) . Conversely, incubation of the corneal tissue for 24 hours with the Rho inhibitor exoenzyme C3 inhibited corneal epithelial migration in a dose-dependent manner (Fig. 5) . 

Quantitative analysis revealed that the dose-dependent inhibition of epithelial migration by exoenzyme C3 was approximately linear for concentrations up to 4 μg/ml (Fig. 6A ), with the effect being statistically significant at concentrations of 2 and 4 μg/ml. Exoenzyme C3 also inhibited LPA-induced epithelial migration in a dose-dependent manner (Fig. 6A) , with complete inhibition of the effect of 2 μM LPA apparent at an exoenzyme C3 concentration of 2 μg/ml. 

The stimulatory effect of LPA on corneal epithelial migration was significant at concentrations of 0.2 and 2 μM (Fig. 6B) . However, at concentrations up to 2 μM, LPA failed to stimulate epithelial migration in the presence of exoenzyme C3 at a dose of 2 μg/ml. 

Roles of Actin Filament Assembly and Myosin Light Chain Kinase in Epithelial Migration

The role of the actin cytoskeleton in the effect of Rho on corneal epithelial migration was examined with the use of cytochalasin B, an inhibitor of actin filament assembly. Cytochalasin B inhibited epithelial migration in a dose-dependent manner, with the effect being statistically significant at concentrations of 0.1 and 0.3 μg/ml (Fig. 7) . Cytochalasin B also inhibited the stimulatory effect of 2 μM LPA on epithelial migration, with complete inhibition of the effect of LPA apparent at a cytochalasin B concentration of 0.1 μg/ml. 

The role of cell-matrix interaction in the stimulatory effect of Rho on corneal epithelial migration was investigated with the use of ML-7, an inhibitor of myosin light chain kinase (MLCK). ML-7 inhibited epithelial migration in a dose-dependent manner, with the effect being significant at a concentration of 1 μM (Fig. 8) . This agent also inhibited the stimulatory effect of 2 μM LPA on epithelial migration, with complete inhibition apparent at an ML-7 concentration of 0.1 μM. 

Discussion

We have shown that epithelial migration in the rabbit cornea was inhibited by the Rho blocker exoenzyme C3 and promoted by the Rho activator LPA. Our data also suggest that the stimulatory effect of Rho on corneal epithelial migration is mediated by the formation of actin filaments and by MLCK. Various cytokines, growth factors, and neural factors stimulate corneal epithelial migration. ^{1 2 3} We therefore propose that Rho participates in the regulation of corneal epithelial migration by such extracellular stimuli. 

Corneal epithelial wound healing occurs in three phases. ^{13 14} In the first phase, the remaining epithelial cells attach to and migrate over the defective area, in which fibronectin appears as a provisional matrix. ^{15 16} After the epithelial defect is resurfaced with a monolayer of epithelial cells, in what constitutes the second phase of wound healing, the cells begin to proliferate. In the final phase, cellular differentiation results in the establishment of a well-layered epithelial structure. The attachment of epithelial cells to the underlying provisional extracellular matrix is required for cell spreading and migration during the first phase of epithelial wound healing. Both fibronectin and fibronectin receptors (integrins) play an important role in this process. ^{17 18} However, the regulatory mechanisms of corneal epithelial cell migration have remained unclear. 

The interconversion of members of the Rho family of small GTP-binding proteins, including Rho, Rac, and Cdc42, between their active GTP-bound and inactive GDP-bound forms allows them to function as switches in intracellular signaling. ^{7 9} These proteins play important roles in signaling pathways that link extracellular signals to changes in the organization of actin filaments. In fibroblasts, Rho triggers the formation of both stress fibers and focal contacts, Rac regulates the formation of lamellipodia, and Cdc42 promotes the formation of filopodia. ^{19 20 21 22} Rho regulates various cellular functions, including integrin-mediated cell-to-substrate adhesion and cell migration, in many cell types. ²³ Many of these cellular roles of Rho have been revealed by inactivating the endogenous protein with exoenzyme C3. ²⁴ Thus, our observation that exoenzyme C3 inhibited corneal epithelial migration suggests that Rho plays a regulatory role in this process. The inhibitory effect of exoenzyme C3 on the migration of corneal epithelial cells is consistent with its similar effects in fibroblasts and leukocytes. ^{25 26}  

The biological activities of Rho are mediated by its interaction with specific target proteins. The serine-threonine protein kinase isozymes ROCK-1 and ROCK-2 have been identified as Rho effectors, and are thought to mediate the stimulatory effects of Rho on the formation of stress fibers and focal adhesions in fibroblasts and epithelial cells. ^{27 28 29} SundarRaj et al. ¹⁰ previously showed that ROCK-1 is expressed in the corneal epithelium. These researchers identified the 160-kDa protein in corneal epithelial extracts and demonstrated a marked increase in the abundance of this protein associated with the transition from the limbal to the corneal epithelium. They suggested that ROCK-1 signaling pathways play important roles in corneal epithelial differentiation and maintenance. We detected both ROCK-1 and ROCK-2 in the rabbit corneal epithelium at both the protein and mRNA levels, and therefore propose that these ROCK isozymes also function in corneal epithelial migration. 

LPA, an intermediate in de novo lipid biosynthesis, is also a platelet-derived serum factor that exhibits a wide range of biologic activities, including stimulation of cell proliferation and of cell motility. ^{30 31 32 33} LPA acts at a cell surface receptor and activates intracellular signaling pathways. It thus induces the tyrosine phosphorylation of focal adhesion kinase and paxillin in cultured Swiss 3T3 fibroblasts. The observation that these effects are inhibited by exoenzyme C3 suggests that they are mediated by Rho. ^{34 35} We have now shown that LPA stimulates corneal epithelial migration in an exoenzyme C3-sensitive manner, again suggesting that this effect is mediated by Rho. Tyrosine kinases also function in corneal epithelial migration. Thus, we recently showed that genistein and herbimycin A, inhibitors of tyrosine kinases, inhibited epithelial migration in the cultured rabbit cornea. ³⁶ Tyrosine phosphorylation was also recently shown to contribute to the LPA-induced reformation of the actin cortical mat and actin bundle reorganization in isolated embryonic corneal epithelia. ³⁷  

Reorganization of intracellular actin filaments plays an important role in epithelial cell attachment to the extracellular matrix and in cellular migration. ³⁸ We previously showed that cytochalasin B inhibits the attachment and spreading of cultured corneal epithelial cells on a fibronectin matrix, as well as the migration of these cells. ^{39 40} Other investigators have described similar results. ^{41 42 43} Furthermore, in the present study, LPA-induced corneal epithelial migration was inhibited by cytochalasin B. LPA also triggered the formation of actin stress fibers in cultured rabbit corneal epithelial cells. These observations suggest that the assembly of actin filaments is essential for corneal epithelial cell migration, and that such filament assembly is a downstream effect of the LPA signaling pathway in these cells. 

Myosin II is colocalized with actin filaments at the leading edge of migrating corneal epithelial cells, suggesting that actin-myosin II interactions are fundamental to the contractile “purse-string” machinery. ⁴³ Rho-associated kinase has been shown to determine the extent of phosphorylation of the light chain of myosin II by direct phosphorylation of this protein as well as by the inactivation of myosin phosphatase through the phosphorylation of its myosin-binding subunit. ^{44 45 46} In the present study, the stimulatory effect of LPA on corneal epithelial migration was inhibited by ML-7. Thus, MLCK also appears to be a downstream target of the LPA signaling pathway responsible for the stimulation of corneal epithelial migration. 

 Supported in part by Grant-in-Aid for Scientific Research 09470381 from the Ministry of Education, Science, Sports, and Culture of Japan.

 Submitted for publication January 26, 2000; revised July 5 and December 11, 2000; accepted December 20, 2000.

 Commercial relationships policy: E (MN, TNa), N (TC), P (TNi).

 

Figure 1.

 

Immunohistochemical analysis of ROCK-1 and ROCK-2 expression in rabbit cornea. Corneal sections were stained with antibodies to ROCK-1 (A) or to ROCK-2 (B), or with normal goat immunoglobulin G (C). Bar, 50 μm.

Figure 2.

 

RT-PCR analysis of ROCK-1 and ROCK-2 mRNAs in cultured rabbit corneal epithelial cells. RT-PCR products were fractionated by agarose gel electrophoresis and stained with ethidium bromide. The sizes of the specific products for ROCK-1 (lane 1), ROCK-2 (lane 2), and G3PDH (lane 3) are 828, 996, and 450 bp, respectively.

Figure 3.

 

Effect of LPA on the formation of actin stress fibers in rabbit corneal epithelial cells. Cells were cultured for 1 hour in the absence (A) or presence (B) of 2 μM LPA. They were then stained with rhodamine-conjugated phalloidin and examined by epifluorescence microscopy. Bar, 20 μm.

Figure 4.

 

Histologic photographs of the extent of epithelial migration in rabbit corneal blocks after incubation for 24 hours in the absence (A) or presence of LPA at concentrations of 0.02 μM (B), 0.2 μM (C), or 2 μM (D). Bar, 200 μm.

Figure 5.

 

Histologic photographs of the extent of epithelial migration in rabbit corneal blocks after incubation for 24 hours in the absence (A) or presence of exoenzyme C3 at concentrations of 1μ g/ml (B), 2 μg/ml (C), or 4 μg/ml (D). Bar, 200 μm.

Figure 6.

 

(A) Effects of exoenzyme C3 on corneal epithelial migration in the absence or presence of LPA. Rabbit corneal blocks were cultured for 24 hours in the absence or presence of 2 μM LPA and in the presence of the indicated concentrations of exoenzyme C3. (B) Effects of LPA on corneal epithelial migration in the absence or presence of exoenzyme C3. Corneal blocks were cultured for 24 hours in the absence or presence of exoenzyme C3 (2 μg/ml) and in the presence of the indicated concentrations of LPA. Data are means ± SEM of four to six determinations. *P < 0.01 versus corneal blocks cultured with medium alone; #P < 0.01 versus corresponding corneal blocks cultured in the absence of LPA (A) or exoenzyme C3 (B).

Figure 7.

 

Effects of cytochalasin B on corneal epithelial migration. Corneal blocks were cultured for 24 hours in the absence or presence of 2 μM LPA and in the presence of the indicated concentrations of cytochalasin B. Data are means ± SEM of four to six determinations.* P < 0.01 versus corneal blocks cultured with medium alone; #P < 0.01 versus corresponding corneal blocks cultured without LPA.

Figure 8.

 

Effects of ML-7 on corneal epithelial migration. Corneal blocks were cultured for 24 hours in the absence or presence of 2 μM LPA and in the presence of the indicated concentrations of ML-7. Data are means ± SEM of six determinations. *P < 0.01 versus corneal blocks cultured with medium alone; #P < 0.01 versus corneal blocks cultured without LPA.

The authors thank Megumi Kawahara for technical assistance and Michiyo Suetomi for help in preparation of the manuscript.