Epidermal Growth Factor-induced Proliferation Requires Down-regulation of Pax6 in Corneal Epithelial Cells*

Growth factors play important roles in regulating corneal epithelial cell proliferation/differentiation during wound healing. It is suggested that PAX6 involves corneal epithelium lineage-specific differentiation (Liu, J. J., Kao, W. W., and Wilson, S. E. (1999) Exp. Eye Res. 68, 295–301); however, the regulatory mechanism and function of Pax6 in growth factor-induced corneal epithelial responses is still unknown. In the present study, we found that the mitogenic effect of epidermal growth factor (EGF) in corneal epithelial cells required suppression of PAX6 activity through cellular mechanisms involving Erk-signaling pathway-mediated increase in CTCF expression. EGF-induced CCCTC binding factor (CTCF) activation subsequently inhibited Pax6 expression by interacting with a CTCF-specific region upstream of the pax6 P0 promoter. Suppression of EGF-induced Erk activation by specific inhibitor or by the dominant expression of a silent Erk mutant effectively abolished the effects of EGF stimulation on regulations of CTCF and pax6. Apparently, down-regulation of Pax6 expression induced by EGF is required for corneal epithelial proliferation, because overexpression of pax6 in these cells attenuated EGF-induced proliferation. In contrast, knockdown of mRNA expression with pax6- or CTCF-specific small interfering RNA in corneal epithelial cells significantly promoted or attenuated EGF-induced proliferation, respectively. Thus, our results revealed a new regulatory mechanism that involves cellular signaling events and pax6 transcription regulation in growth factor-mediated proliferation. In corneal epithelial cells, this suggests that inhibition of pax6 expression is a prerequisite for EGF to elicit controls of cell growth and fate.

Corneal epithelial cells on the surface layer of the cornea form the defense line as a barrier against noxious agents. Growth factor-mediated renewal of the corneal epithelium plays a functional role in maintaining the barrier function and corneal transparency (1,2). The basal layer cells must proliferate at a fast rate to replace terminally differentiated cells in the more superficial layers and to maintain this protective function. Epidermal growth factor (EGF) 1 is found to have stimulatory effects on corneal epithelial cell proliferation and migration (3). In wound healing models employing cultured corneal epithelial cells, the healing rate can be tremendously enhanced by an optimal dose of EGF present in the medium (4). EGF receptor-linked cell signaling in corneal epithelial cells includes pathways involving stimulation of phospholipase C, D, protein kinase A, phospholipase A 2 , phosphatidylinositol 3-kinase, and limbs of the mitogen-activated protein kinase cascade (5)(6)(7)(8)(9)(10)(11)(12). Application of EGF in the serum starvation-synchronized corneal epithelial cells induces the formation of EGF receptor clusters in the cell membrane and increases cell proliferation by promoting cells entering S and G 2 /M phases of the cell cycle (12,13). The proliferative effect of EGF on corneal epithelial cell growth is a complex process including activation of the extracellular signal-regulated kinase limb of the mitogen-activated protein kinase cascade and stimulation of voltage-gated K ϩ channels and bumetanide-sensitive Na-K-2Cl cotransporters (11,(13)(14)(15)(16). However, the regulatory mechanism for the cell fade and mitogenic response to this cytokine still remains to be further investigated.
Pax6 plays a critical and evolutionarily conserved role in determining early stage cell differentiation in eye development in both vertebrates and invertebrates (17,18). Homozygous mutation of eyeless, homologue of the pax6 gene, results in missing eye structures in Drosophila (19). Overexpression of pax6 can induce fully differentiated ectopic eyes and ectopic expressions of early eye development genes, such as Otx2, Rx, and Six3, as well as the endogenous pax6 (17,20,21). In addition, specific mutation of pax6 causes a small eye (Sey) defect in mice and ocular aniridia in humans (22)(23)(24). The Pax6 gene is expressed in essentially all ocular structures of vertebrates, including the cornea, iris, lens, and retina (25)(26)(27)(28). Deficient expression of the pax6 gene causes death in mice shortly after birth. Although pax6 is down-regulated or fades away in most tissues and cell types following differentiation, it remains detectable in several mature cell types of the eye, including corneal, epithelial, lenticular, and retinal (29). In the developed cornea, PAX6 is a positive transcription factor essential for controlling the transcription of the cornea-specific differentiation marker keratin 12, suggesting that Pax6 may play an important role in early differentiation and maintaining a differentiation pattern in these cells (30).
The regulatory mechanisms of Pax6 gene expression are still largely unknown. It is apparently regulated by at least two enhancers. One has been identified as a 341-base pair enhancer located in the 5Ј region 4.2 kb upstream of the pax6 gene (31). This enhancer is highly conserved in the Pax6 gene of mouse, humans, and puffer fish (Fugu) (31,32). This enhancer is a dominant factor in the preplacodal phase of Pax6 expression and designated as the Pax6 ectoderm enhancer, whereas the existence of the other (enhancer 2) is implied. Recently, studies in our laboratory revealed that a repressor element is located at ϳ1.2 kb upstream from the pax6 P0 promoter in mouse. The repressor element contains five repeat binding sites in the 80-bp region for CTCF, a transcription factor binding to the CCCTC DNA sequence (33). Over-expression of the CTCF gene in corneal cells induces a decrease of pax6 P0 activity, and a deletion mutant of the CTCF binding sequence attenuated the decrease, indicating that CTCF serves as a repressive protein in pax6 gene expression (33).
In this report, we present important results to demonstrate that the mitogenic effect of EGF in corneal epithelial cells requires suppression of PAX6 activity. The effect of EGF on the suppression of pax6 expression is through Erk-signaling pathway-mediated activation of CTCF. EGF-induced activation of Erk cascades resulting in increases in CTCF expression subsequently inhibited pax6 transcription. Overexpression of pax6 attenuated EGF-stimulated corneal epithelial proliferation, suggesting that inhibition of pax6 expression is a prerequisite for EGF to elicit control of cell growth and fate. In addition, our results revealed new regulatory mechanisms in corneal epithelial cells involving the EGF receptor-linked mitogen-activated protein kinase signaling pathway and transcription control of Pax6 function in growth factor-stimulated proliferation.

MATERIALS AND METHODS
Cell Culture and Gene Transfection-Human and rabbit corneal epithelial (HCE and RCE, respectively) cells were cultured in Dulbecco's modified Eagle's medium/Ham's F-12 medium containing 10% fetal bovine serum and 5 g/ml insulin (Sigma) in a 37°C incubator gassed with 5% CO 2 . Before the experiments, both cells were synchronized at G 1 phase of the cell cycle by depriving serum from the culture medium for at least 24 h. Human hematopoietic myeloblast ML-1 cells were cultured in RPMI 1640 medium (Invitrogen) containing 7.5% fetal bovine serum (Invitrogen) in a 37°C incubator gassed with 5% CO 2 . Pax6 P0 reporter constructs and full-length cDNAs encoding the CTCF gene and the dominant negative Erk1 gene were transfected into RCE cells by electroporation. The cells were washed twice and resuspended in phosphate-buffered saline (PBS) with a density of 10 7 /ml on ice for 5 min. Cells in 0.8 ml of PBS containing 5 g of constructed DNAs were transferred into a 0.4-cm gap electroporation cuvette. An ECM® 630 electroporator (Genetronic, Inc.) was used with a pulse protocol using 300-V, 100-ohm, and 500-microcoulomb settings. Electroporated cells were transferred into 10 ml of normal culture medium prewarmed at 37°C and cultured in an incubator at 37°C.
Northern Blot Experiments-Total RNAs were extracted with a guanidine thiocyanate procedure (34,35). Briefly, 1 ϫ 10 7 cells were collected and rinsed with ice-cold PBS. Cells were immediately lysed with 1 ml of guanidium solution (5 M guanidine hydrochloride, 50 mM Tris-HCl, pH 8, 0.5% N-lauroylsarcosine, 100 mM ␤-mercaptoethanol). Lysates were extracted three times with 50:50 phenol:chloroform. Finally, RNAs were precipitated by centrifugation at 12,000 revolutions/min for 15 min after preincubation with ethanol at Ϫ80°C. RNA (20 g) for each sample was loaded in 1% agarose gel denatured with 2.2 M formaldehyde. The fractionated RNA was transferred onto nylon membrane. The membrane was subsequently hybridized with the corresponding ␣-32 P-labeled DNA probe using a Random Primer labeling kit (New England Biolabs, Beverly, MA). Signals in the membrane were visualized by exposure to x-ray film at Ϫ80°C overnight or longer.
Western Analysis Experiments-Western blot experiments were performed as described previously (36). In brief, 5 ϫ 10 6 cells were rinsed twice with PBS and harvested in 0.5 ml of lysis buffer (20 mM Tris, pH 7.5, 137 mM NaCl, 1.5 mM MgCl 2 , 2 mM EDTA, 10 mM sodium pyrophosphate, 25 mM ␤-glycerophosphate, 10% glycerol, 1% Triton X-100, 1 mM Naorthovanadate, 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml leupeptin). Cell lysates were precleared by centrifugation at 13,000 ϫ g for 15 min. The cell lysates were denatured by adding equal volumes of 2ϫ Laemmli buffer and by boiling for 5 min. Each sample with 20 g of protein was electrophoresed in a 10% SDS-polyacrylamide gel and then transferred to polyvinylidene difluoride membrane. The membrane was incubated with rabbit anti-Pax6 or rabbit anti-Erk1 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) and then incubated with anti-rabbit immunoglobulin G conjugated with horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA). Signals in the membrane were visualized using a horseradish peroxidase blot detection kit (Santa Cruz Biotechnology). All membranes were stripped following a standard stripping protocol. These membranes were rehybridized with mouse anti-␤-actin and goat anti-mouse IgG antibodies and visualized with the horseradish peroxidase blot detection kit (Santa Cruz Biotechnology).
Immunocomplex Kinase Assays-To detect mitogen-activated protein kinase activity, cells were harvested as described above. Erk proteins in cell lysates were immunoprecipitated by incubation with anti-Erk antibody (1:100) at 4°C overnight. Immunocomplexes of protein and antibodies were washed twice with lysis buffer and then kinase buffer (20 mM HEPES, pH 7.6, 20 mM MgCl 2 , 25 mM ␤-glycerophosphate, 100 mM sodium orthovanadate, and 2 mM dithiothreitol). Immunocomplexes were resuspended in 100 l of kinase buffer. Glutathione S-transferase/ myelin basic protein (Santa Cruz Biotechnology) was used as a substrate for kinase activity assays. The kinase reaction was initiated by adding 2 l of ATP mixture containing 20 M ATP and 10 Ci of [␥-32 P]ATP (Amersham Biosciences). The reaction was processed at room temperature for 5 min and terminated by adding 30 l of 2ϫ Laemmli buffer. Phosphorylation of myelin basic protein was displayed by PAGE and visualized by autoradiography.
Analysis of ␤-Galactosidase (␤-Gal) and Luciferase Activities-Pax6 P0 promoter and its mutant transfected cells were harvested for the Pax6 P0 promoter activity analysis in days 2 and 3 after transfection. The cells were washed twice with ice-cold PBS and suspended in icecold lysis buffer containing 100 mM potassium phosphate buffer with 1 mM dithiothreitol, pH 7.8. Cells were ruptured by three cycles of freeze/ thaw. Cell lysates were precleared by centrifugation at 13,000 ϫ g for 5 min, and the supernatants were kept at Ϫ80°C for later analysis of ␤-galactosidase activity. Galacton-star was used as a chemiluminescent substrate in a luminescent ␤-gal system (Clontech, Palo Alto, CA) for determination of ␤-galactosidase activity. Chemiluminescent signals from products of Galacton-star catalyzed by ␤-galactosidase were determined by a luminometer. An internal control vector (pBRL-TK, 0.2 g/transfection, Invitrogen) was introduced into the cells in parallel with pax6 P0 reporters for normalization of lacZ activity. The luciferase activity was measured using an assay kit supplied by Invitrogen.
MTT Cell Proliferation Assay-The MTT cell proliferation assay is a colorimetric assay system that measures the reduction of a tetrazolium component (MTT) into an insoluble formazan product by the mitochondria of viable cells. One hour prior to cell harvest, the culture medium was replaced with 1 ml of serum-free medium, and 100 l of MTT solution (5 mg/ml in PBS) was added into each of the wells and incubated in a CO 2 incubator for 1 h. The medium was replaced by 0.4 ml of acidic isopropyl alcohol (0.04 M HCl in absolute isopropyl alcohol) to solubilize the colored crystals. MTT was changed to a blue color by mitochondrial dehydrogenase. The samples were read using an enzymelinked immunosorbent assay plate reader (Labsystems Multiskan MCC/340, Fisher Scientific) at a wavelength of 570 nm, with the background subtraction at 650 nm. The amount of color produced, normalized with the background, is directly proportional to the number of viable cells and is represented as the proliferation index.

Growth Factor-induced Increase in CTCF Expression and
Decrease in pax6 Expression-The effects of serum containing growth factors on expressions of CTCF mRNA and PAX6 protein were tested by stimulating synchronized RCE and ML-1 cells with 7.5% fetal bovine serum for 6 h (Fig. 1A). Northern analysis revealed that the level of CTCF mRNA was increased in response to serum stimulation in both RCE and ML-1 cells. There was a decrease in Pax6 protein expression detected by Western analysis in RCE cells. However, Pax6 was not expressed in ML-1 cells. To further investigate the regulatory mechanism, the effect of EGF on CTCF and Pax6 activities was studied in RCE cells, because EGF is one of the important growth factors that promote corneal epithelial proliferation. The dose-response relationship and time course of EGF-stimulated increases in CTCF mRNA expression were determined in RCE cells by Northern analysis (Fig. 1, B and C, upper panels). Normalized density analyses obtained from three independent experiments were plotted to demonstrate the statistical significant difference, p Ͻ 0.05 (Fig. 1, B and C, lower panels). Expression of CTCF apparently increased in response to 10 ng/ml EGF stimulation at 3 h and reached the peak level at 12 h. On the other hand, control experiments detected a high expression level of basal pax6 protein in RCE cells using Western analysis. The effect of EGF on decrease in Pax6 expression demonstrated a dose-dependent pattern (Fig. 1D). Upon application of 10 ng/ml EGF, Pax6 expression was decreased following a time course and reached a very low level at 12 h (Fig. 1E). Statistical analysis showed the significant difference in EGFinduced changes of Pax6 expression (n ϭ 3, p Ͻ 0.05) (Fig. 1, D and E, lower panels). These data suggest that there is a correlation between EGF-induced increase in CTCF expression and decrease in pax6 expression in RCE cells. This correlation is consistent with our previous report that CTCF can inhibit pax6 transcription by binding to a repressor element located upstream of the pax6 P0 promoter.
Effect of EGF on Pax6 Promoter Activity-In our previous studies, we have demonstrated that CTCF suppresses P0 activity through interaction with a repressor element (CTCF binding sites) located in the region Ϫ1.2 kb upstream from the pax6 P0 promoter (33). To verify whether EGF-induced decrease in Pax6 expression resulted from the increase in CTCF expression, the effect of EGF on Pax6 promoter activity was examined using the previously made P4.2 ␤-galactosidase reporter and an internal deletion mutant Pxba that lacks the binding sequences for CTCF (33). These reporter constructs were transiently introduced by electroporation into RCE cells and a stable RCE cell line that was previously transfected with tetracycline (tet)-inducible CTCF cDNA for overexpression of CTCF. To test Pax6 P0 promoter activity, P4.2 and Pxba mutant-transfected cells were induced with 2 g/ml tetracycline. Tet-induced overexpression of CTCF markedly inhibited P4.2 promoter activity but did not affect Pxba mutant promoter activity ( Fig. 2A). The effect of EGF on pax6 P0 promoter activity was also investigated in RCE cells that were transiently transfected with P4.2 and Pxba mutant reporters. EGF stimulation significantly inhibited P4.2 reporter activity but had no effect on Pxba mutant reporter activity, as the binding sequence for CTCF in the mutant reporter had been deleted (Fig. 2B). The inhibitory effect of EGF on the pax6 P0 promoter was rather specific, because there were no changes in ␤-galactosidase activity in RCE cells transfected with cytomegalovirus-␤-gal control vector. The results indicate that EGF inhibits pax6 P0 promoter activity through increase in CTCF activity in RCE cells.
Effect of EGF-induced Erk Activation on CTCF and Pax6 Expression-Previous studies from our laboratory indicate that the Erk signaling pathway is one of the major EGF-induced signaling events responsible for RCE cell proliferation (11). Upon stimulation of RCE cells with 10 ng/ml EGF, Erk activity was transiently increased within 24 h (Fig. 3A). EGF-induced effects on CTCF and PAX6 expressions were markedly affected by PD98059, which is a specific inhibitor for mitogen-activated protein kinase/ extracellular signal-regulated kinase kinase, a mitogen-activated protein kinase kinase immediately upstream from Erk in the mitogen-activated protein kinase cascades (Fig. 3B). To further verify the effect of Erk on CTCF and PAX6 expressions, a tetinducible dominant negative Erk construct was introduced into RCE cells. A RCE cell line with stable transfection of dominant negative Erk1 was established by double selections with blasticidin (2 g/ml) and zeocin (50 g/ml). Tet-induced increases in Erk1 protein expression were observed following an increased dose of tetracycline in RCE cells that were stably transfected with dominant negative Erk1 (pcDNA4-ErkDN) but not in control cells transfected with the pcDNA4 vector (Fig. 3C). EGFinduced Erk activity was measured in tet-induced and uninduced RCE cells with dominant negative Erk1 stable transfection by immunocomplex assay. Myelin basic fusion protein was used as a substrate. Induction of these cells with tetracycline (2 g/ml) effectively blocked EGF-induced Erk activities (Fig. 3D). Furthermore, EGF induced a strong increase in CTCF expression and decrease in Pax6 expression. EGF-induced increase in CTCF expression and decrease in Pax6 expression were totally blocked by tet-induced overexpression of dominant negative Erk1 (Fig. 3,  E and F). Results presented here provide further evidence in RCE cells that EGF-induced suppression of Pax6 expression is through activation of CTCF and that EGF-induced Erk signaling cascades play an important role in regulating CTCF activity.
Functional Role of Pax6 in EGF-stimulated RCE Cell Proliferation-Functional role of PAX6 in EGF-induced RCE cell proliferation was studied by overexpression of the pax6 gene and by knockdown of Pax6 mRNA. RCE cell proliferation was determined by MTT and presented as the proliferation index. A cDNA fragment encoding the full-length human pax6 gene was subcloned into pcDNA4 vector that contains a cytomegalovirus promoter. The full-length human pax6 cDNA was transfected into RCE cells by electroporation. Western analysis of human pax6 cDNA-transfected cells revealed a new and larger molecular mass band (47 kDa) representing human PAX6 protein, but this band was absent in vector-transfected control cells (Fig. 4A). The other lower molecular mass band (43 kDa) was found in both human Pax6 cDNA-and vector-transfected cells, representing the endogenous rabbit PAX6 protein. Stimulation of EGF (10 ng/ml) induced a fast proliferation in days 2 and 3 in control RCE cells. However, the proliferative effect of EGF was significantly attenuated by overexpression of pax6 in human pax6 cDNA-transfected cells (n ϭ 6, p Ͻ 0.05), suggesting that PAX6 plays a functional role in the inhibition of EGFinduced cell proliferation (Fig. 4B). In contrast, transfecting RCE cells with siRNA specific to the pax6 gene knocked down the endogenous pax6 mRNA (Fig. 4C). Interestingly, knockdown of Pax6 expression with siRNA significantly promoted RCE cell proliferation in the control cells and in EGF-stimulated RCE cells in days 2 and 3 (Fig. 4D). The effect of knocking down pax6 mRNA on cell proliferation was statistically significant (n ϭ 6, p Ͻ 0.05). In addition, it resulted in the growing of RCE cells in the absence of EGF and even much faster growing in EGF-induced cells in days 2 and 3.
Effect of Knocking Down CTCF mRNA on pax6 Expression and EGF-stimulated Proliferation-To further verify the functional role of CTCF in mediating EGF-induced alteration of pax6 transcription and RCE cell proliferation, CTCF activity in RCE cells was suppressed by knockdown of CTCF mRNA using CTCFspecific siRNA. As it has been shown in Fig. 1, EGF induced increases in CTCF expression and decreases in pax6 expression. However, EGF stimulation (20 ng/ml) failed to induce the increase in CTCF expression and decrease in pax6 expression in RCE cells that were transfected with CTCF-specific siRNA, (Fig.  5A). The effect of knocking down CTCF mRNA on CTCF and Pax6 expression were plotted as means with S.E. bars in Fig. 5, B and C, respectively. The significant difference was determined by analysis of variance (n ϭ 3, p Ͻ 0.05). The effect of knocking down CTCF mRNA on EGF-induced RCE cell proliferation was also examined by MTT and presented as the proliferation index (Fig. 5D). Apparently, suppression of CTCF mRNA expression significantly inhibited RCE cell growth in both the absence and presence of EGF stimulation (n ϭ 6, p Ͻ 0.05). Results from inhibition of CTCF mRNA expression experiments provide further evidence that EGF-induced RCE cell proliferation requires downregulation of Pax6, and this action was mediated by CTCF. DISCUSSION Corneal epithelial progenitor cells proliferate and differentiate to form the corneal epithelial layer by making spatial and temporal fate decisions. This renewal process maintains the healthy condition through a dynamic wound healing process. These processes are largely controlled by intercellular signal pathways through activation of growth factor receptors. Our data indicate that serum-containing growth factors, such as EGF, stimulate RCE cell growth by inhibiting Pax6 expression. It is consistent with previous reports that pax6 involves the lineage-specific differentiation by turning on the cornea-specific differentiation marker keratin 12 (1,30). A recent study demonstrates, in Drosophila, that hyperactivation of Egfr, a homologue of the mammalian epidermal growth factor receptor (EGFR), induces inhibition of a pax6 homologue eyeless gene expression, resulting in an ectopic eye growing on an antenna (38). PAX6 plays a critical and evolutionarily conserved role in determining early stages of In the present study, we aimed to determine the functional role of PAX6 in EGF-induced proliferation. Two parallel experiments were performed in RCE cells to manipulate Pax6 activ-ities and to verify the effect of Pax6 on EGF-induced proliferation. First, PAX6 activity was altered by overexpression of the pax6 gene and by knockdown of pax6 mRNA. Second, pax6 transcription was altered by tet-inducible overexpression of CTCF and by knockdown of CTCF mRNA. We found that, in RCE cells, up-regulation of pax6 expression by overexpression of the human pax6 gene significantly suppressed EGF-induced EGF (10 ng/ml) was applied to human PAX6-transfected RCE cells, and cell proliferation was determined on days 2 and 3 by MTT cell proliferation assay. C, knockdown of pax6 mRNA using siRNA specific to PAX6. Expression levels of endogenous pax6 mRNA were detected by Northern analysis in control and pax6 siRNA-transfected RCE cells. ␤-actin levels were detected as loading controls. D, effect of knocking down Pax6 mRNA on EGF-induced cell proliferation. EGF (10 ng/ml) was applied to Pax6 siRNAtransfected RCE cells, and cell proliferation was determined on days 2 and 3 by MTT cell proliferation assay. * indicates significant difference by comparison in the absence and presence of EGF stimulation on days 2 and 3 (n ϭ 6, p Ͻ 0.05), and ** indicates significant difference by comparison of transfected and untransfected cells (n ϭ 6, p Ͻ 0.05). proliferation on both days 2 and 3 (Fig. 4B). On the other hand, down-regulation of pax6 expression by the introduction of siRNA specific to the pax6 mRNA sequence into RCE cells significantly promoted EGF-induced cell proliferation on both days 2 and 3 (Fig. 4D). On the other hand, dominant expression of tet-inducible CTCF and knockdown of CTCF mRNA by siRNA resulted in down-regulation and up-regulation of Pax6 expression, respectively. Enhancement of pax6 expression by knockdown of CTCF mRNA effectively inhibited EGF-stimulated RCE cell proliferation. In conclusion, our results provide strong evidence that the regulation of pax6 in mature corneal tissue is closely related to growth factor-induced cell proliferation and differentiation, because: 1) pax6 expression was suppressed by EGF-induced cell proliferation, 2) alteration of CTCF activity resulted in decrease in pax6 transcription, and 3) alteration of PAX6 activity affected EGF-induced cell proliferation. These results indicate that, indeed, PAX6 is involved in EGF-induced corneal epithelial cell growth control. It is understandable that cell proliferation temporarily must suppress the machinery that links to cell differentiation. In this case, pax6 expression suppressed by EGF during growth factorinduced proliferation may play an important functional role by keeping proliferative cells in the cell cycle.
Further studies are being performed to understand how pax6 is regulated and which signaling pathway is involved in regulating pax6 transcription. Our results indicate that EGF promotes corneal epithelial proliferation and induces suppression of pax6 expression. The effect of EGF on pax6 expression is closely associated with the increase in CTCF expression (Fig.  1). EGF stimulates pax6 P0 promoter activity, resulting in increases in ␤-gal expression of the pax6 reporter (P4.2 construct) in RCE cells. The mutant (Pxba) with an internal deletion of CTCF binding domain in the pax6 reporter abolished the effect of EGF on the pax6 P0 promoter (Fig. 2B). This is consistent with our previous studies that indicate there is a CTCF regulatory element (repressor) in this region upstream from the pax6 P0 promoter (33). EGFR, a receptor tyrosine kinase, regulates cell growth and fate decisions of corneal epithelial cells (11). In Drosophila, hyperactivation of Egfr induces an eye-to-antenna transformation and also an inhibition of eyeless gene expression (ey, homologue of the pax6 gene in Drosophila) (37). The role of Egfr in eye development of Drosophila is properly mediated by the extracellular signal-regulated kinase pathway (38). We have studied here the effect of EGF on RCE cell proliferation and differentiation by investigating EGFR-coupled signaling events and the regulation of related genes, including the pax6 gene. The effect of EGF-induced corneal epithelial cell growth is investigated by the detection of EGFRlinked Erk signaling cascades. EGF-induced increases in CTCF expression and decreases in Pax6 expression are mediated by the Erk signaling pathway in RCE cells (Fig. 3). Our conclusion is based on the following results. First, EGF-induced CTCF expression in RCE cells revealed a similar time course and dose-response relationship compared with its inhibitory effect on pax6 expression. Second, inhibition of mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (an immediate mitogen-activated protein kinase kinase upstream from Erk) with PD98059 markedly blocked EGF-induced upregulation of CTCF expression and down-regulation of Pax6 expression. Finally, overexpression of the tet-inducible dominant negative Erk in transfected RCE cells effectively inhibited EGF-induced up-regulation of CTCF expression and down-regulation of pax6 expression. These results provide sufficient evidence that the regulatory effect of EGF on CTCF and subsequently on pax6 expression is through the EGFR-linked Erk signaling pathway.
In summary, the present work, for the first time, demonstrates three important findings. First, EGF induces increases in CTCF expression in corneal epithelial cells through activation of the Erk signaling pathway. It is known that Erk nuclear translocation occurs after activation and regulates gene expressions, and it is possible that CTCF is one of those genes regulated by Erk. Second, EGF induces increases in CTCF expression to suppress pax6 expression by inhibition of pax6 transcription. The inhibitory effect of CTCF is accomplished by interaction with a specific DNA binding region in the pax6 P0 promoter. Finally, our data demonstrate that knockdown of pax6 expression promotes EGF-induced cell proliferation. In contrast, overexpression of pax6 expression attenuates the EGF effect on RCE cell proliferation, suggesting that PAX6 plays an important role in controlling growth factor-induced corneal epithelial cell growth. Apparently, EGF-induced RCE cell proliferation requires down-regulation of pax6 to prevent premature differentiation.