Corneal Epithelial Wound Healing Impaired in Keratinocyte-Specific HB-EGF–Deficient Mice In Vivo and In Vitro

Abstract

Purpose.: To study the role played by HB-EGF in corneal epithelial wound healing.

Methods.: A 2-mm corneal epithelial wound was created in keratinocyte-specific, HB-EGF–deficient mice—HB^lox/lox:K-5Cre (HB^−/−)—and the speed of wound healing was compared with that in wild-type (WT) mice. Cultured confluent mouse corneal epithelial cells (MCECs) from WT and HB^−/− mice were scraped, and the bare area was measured. The proliferation of MCECs was determined by BrdU incorporation. The degree of attachment of WT and HB^−/− MCECs was also determined. The mRNA expression of EGF family and cell adhesion molecules was determined by real-time PCR.

Results.: Corneal epithelial wound healing was significantly delayed in HB^−/− mice, and the expression of HB-EGF was detected at the leading edge of the wound in HB^lox/+:K5-Cre (HB^+/−) mice by the presence of lacZ staining. The wound closure was significantly impaired in HB^−/− MCECs and was improved by adding HB-EGF. The number of BrdU-positive MCECs of WT and HB^−/− mice was not significantly different, and both increased to different degrees when HB-EGF was added. The adhesion of isolated HB^−/− MCECs was lower than that of WT MCECs, but the degree of adhesion was restored by adding HB-EGF. The expression of epiregulin was upregulated in HB^−/− MCECs, and α6- and β1-integrin were upregulated by adding HB-EGF.

Conclusions.: HB-EGF plays an important role in corneal epithelial cell healing by enhancing cellular attachment and part of cell proliferation.

Corneal epithelial wound healing is an important process for maintaining the homeostasis of the cornea, and impaired wound healing increases the risk of corneal infection that can then lead to a decrease in visual acuity. Various growth factors participate in corneal epithelial wound healing, and extensive research has been conducted on one family of growth factors: the epidermal growth factor (EGF) family. The EGF family consists of EGF, ¹ TGF-α, ² HB-EGF, epiregulin, ³ amphiregulin, ⁴ betacellulin, ⁵ and neuregulin. ^6,7 The EGF family acts through EGF receptors (EGFRs), which are divided into four subtypes; EGFR/Erb1/HER1, Erb2/HER2/neu, Erb3/HER3, and Erb4/HER4. ^{8 –11}  

Among the members of the EGF family, many studies have demonstrated the importance of HB-EGF in the pathophysiology of malignant tumors (e.g., hepatic ¹² and pancreatic ¹³ cancers). These findings are well supported by the observations that HB-EGF–knockout mice die shortly after birth, ¹⁴ whereas knockout mice for the other EGFR binding growth factors, such as EGF, TGFα, amphiregulin, and epiregulin, do not die and have different phenotypes. ^{15 –17} This knowledge has led researchers to infer that, of all the members of the EGF family, HB-EGF plays the most role in growth and development. 

Wilson et al. ¹⁸ reported that HB-EGF enhances the proliferation of cultured human corneal epithelial cells, and the results of more recent studies showed that HB-EGF promotes wound healing in cultured human corneal epithelial cells ¹⁹ and in porcine cornea. ²⁰ Thus, HB-EGF may also be an important growth factor in the corneal epithelial cells. 

However, it is difficult to investigate the role of HB-EGF in specific organs and tissues in HB-EGF–knockout mice, because the animals die shortly after birth. Recent technology has been developed to overcome this problem by establishing lines of tissue-specific, gene-deficient animals for use in investigating the role of the genes in specific tissues. Thus, we have used keratinocyte-specific HB-EGF–deficient mice created by Cre/loxP technology in combination with the keratin 5 promoter. ²¹ The keratinocyte–specific HB-EGF–deficient mice are suitable for investigating the pathogenesis of pathologic abnormalities in the skin and cornea, because the mice grow normally and do not show apparent abnormalities. These mice do not express HB-EGF in the corneal epithelial cells and the epidermal keratinocytes, which was confirmed by RT-PCR in this study. 

The purpose of this study was to investigate the role of HB-EGF in the corneal wound-healing process in keratinocyte-specific, HB-EGF–deficient mice, in vivo and in vitro. In addition, experiments were performed on cultured mouse corneal epithelial cells (MCECs) to allow us to conduct a more detailed investigation of the role of HB-EGF. 

Materials and Methods

Animals

Because HB-EGF–null mice do not survive past the fetal period, we used keratinocyte-specific, HB-EGF–deficient mice (HB^lox/lox:K-5Cre; HB^−/−) that were created by Cre/loxP technology in combination with the keratin-5 promoter. ²¹ The targeting construct of HB^lox/lox is described in detail in Figures 1A–D and elsewhere. ¹⁴  

Figure 1.

 

Structure and genotype of keratinocyte-specific HB-EGF–deficient (HB^−/−) mice and the expression of HB-EGF in MCECs. (A–D) Structure of the transgene for the creation of HB^−/− mice. (A) Structure of the K5Cre transgene: a 14-kb fragment of human K5 promoter is used to control the expression of the Cre transgene. (B) The mouse Hb-egf gene. (C) Structure of the HB^lox/lox allele: mouse HB-EGF cDNA flanked by loxP sequences fused with the mouse Hb-egf gene. The lacZ gene was inserted downstream of the HB-EGF cDNA. (D) Structure of HB^−/− allele: Cre-mediated recombination results in the deletion of the HB-EGF cDNA and the expression of the lacZ gene. E, EcoRI; H, HindIII; K, KpnI; S, SacII; V, EcoRV; and X, XhoI. (E, F) The genotyping of keratinocyte-specific HB-EGF–deficient (HB^−/−) mice. HB^−/− mice were confirmed by PCR as K5Cre recombinase positive (E) and lox homozygous (F). (G) RT-PCR analysis of HB-EGF mRNA in MCECs. Representative images of RT-PCR in HB^−/− and WT MCECs. All PCR products were analyzed by 1.5% agarose gel electrophoresis. P1, passage 1; P2, passage 2; P3, passage 3.

Homozygous HB^lox/lox mice were bred with K5 promoter-driven Cre-recombinase transgenic mice, ²² to generate K5-Cre-HB^lox/+ mice. Subsequently, K5-Cre-HB^lox/+ mice were bred with HB^lox/lox mice to generate HB^lox/lox:K5-Cre (HB^−/−) mice. The genotype of each mouse was confirmed by PCR analysis (Figs. 1E, 1F). The primer pairs for PCR are listed in Table 1. 

Table 1.

 

Primer Sequences for PCR and RT-PCR

Target	Primer	Sequence (5′→3′)	Product Size (bp)
PCR
Wild-type HB-EGF	Forward	CATGATGCTCCAGTGAGTAGGCTCTGATTAC	350
Wild-type HB-EGF	Reverse	AGGGCAAGATCATGTGTCCTGCCTCAAGCC
lox HB-EGF	Forward	ATGGGATCGGCCATTGAACA	800
lox HB-EGF	Reverse	GAAGAACTCGTCAAGAAGGC
K5cre-recombinase	Forward	TTACCGGTCGATGCAACGAGTGATG	400
K5cre-recombinase	Reverse	TTCCATGAGTGAACGAACCTGGTCG
RT-PCR
HB-EGF	Forward	GACCCATGCCTCAGGAAATA	232
HB-EGF	Reverse	TGAGAAGTCCCACGATGACA
Epiregulin	Forward	TACCGCCTTAGTTCAGATGG	166
Epiregulin	Reverse	ACATCGCAGACCAGTGTAGC
TGF-α	Forward	GGAATTCCTAGCGCTGGGTATCCTGTTA	148
TGF-α	Reverse	CAAGCTTACCACCACCAGGGCAGTGATG
EGF	Forward	CCCAGGCAACGTATCAAAGT	203
EGF	Reverse	GGTCATACCCAGGAAAGCAA
β-actin	Forward	CCTGTATGCCTCTGGTCGTA	260
β-actin	Reverse	CCATCTCCTGCTCGAAGTCT
integrinα6	Forward	GTCACCGCTGCTGCTCAGAATA	146
integrinα6	Reverse	AGCATCAGAATCCCGGCAAG
integrinβ1	Forward	ATCCCAGCCAGTCCCAAGTG	142
integrinβ1	Reverse	TCCTGCAGTAAGCGTCCATGTC
GAPDH	Forward	TGTGTCCGTCGTGGATCTGA	150
GAPDH	Reverse	TTGCTGTTGAAGTCGCAGGAG

To confirm that the K5 promoter was functioning in the corneal epithelium, we bred K5 Cre-recombinase transgenic mice with ROSA-LacZ reporter mice (B6;129-Gt(ROSA)26Sor^tm1Sho /J; The Jackson Laboratory, Bar Harbor, ME), ²³ to generate K5-Cre-ROSA-LacZ (K5-Rosa) mice. The mouse eyes were enucleated, fixed in 0.2% glutaraldehyde for 30 minutes, washed with PBS, and stained with 5-bromo-4-chloro-3-idol β-d-galactoside (X-gal) overnight using the β-Gal staining set (Roche Diagnostics, Indianapolis, IN) to demonstrate β-galactosidase (β-gal) activity. After the stained corneas were photographed under a stereomicroscope (SteREO Lumar.V12; Carl Zeiss Meditec, Tokyo, Japan), they were postfixed in 4% paraformaldehyde and embedded in paraffin. Sections (5 μm) were cut, stained with nuclear red, and observed with a conventional light microscope. 

All mice were handled in accordance with the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Expression of HB-EGF during Corneal Epithelial Wound Healing

The targeting vector contained the lacZ gene as a reporter gene for the expression of HB-EGF, ¹⁴ and X-gal staining was performed on the corneas of HB^lox/+:K5-Cre (HB^+/−) mice during corneal wound healing. Six- to eight-week-old HB^+/− mice were anesthetized, and a 2.0-mm diameter corneal epithelial defect was manually created in the right eyes. The eyes were enucleated 0, 12, and 24 hours after the injury and stained for X-gal. 

In Vivo Corneal Epithelial Cell Wound Healing

Six- to eight-week-old HB^−/− mice (n = 9) and WT C57BL/6J mice (n = 9; CLEA Japan Inc, Tokyo, Japan) were studied. With the subject under general anesthesia, a 2.0-mm diameter corneal epithelial defect was made in the right eye. The injured area was photographed daily for 5 days after the injury, and the size of the injured area was measured on the images with technical illustration and graphic software (Canvas, ver 6.0; Deneba Softwear, Miami, FL). 

MCECs in Culture

We have recently established an MCEC culture system that enabled us to obtain enough MCECs from HB^−/− and WT mice for in vitro experiments. ²⁴ Briefly, eyes removed from HB^−/− or WT mice were incubated overnight (18 hours at 4°C) in DMEM/F12 (1:1 mixture, Invitrogen, Tokyo, Japan) containing 15 mg/mL of Dispase II (Roche Diagnostics), 100 mM sorbitol, and antibiotic-antimycotic (1×; Invitrogen). ²⁵ Then, the corneal epithelium was removed as a sheet and dissociated into single cells in 0.25% trypsin (Invitrogen). The dissociated cells were seeded into 24-well plates and cultured in CnT-50 (CELLnTEC, Bern, Switzerland), a serum-free, low bovine pituitary extract (BPE) medium, as the primary culture. After they reached subconfluence, the cells were subcultured, and third- to fifth-passage cells were used for the experiments. With this MCEC culture system, MCECs from WT and HB^−/− mice grew at approximately the same rate, and the cell densities at confluence of the fifth passage were 7.0 to 8.1 × 10⁴/cm² (n = 6) in WT and 6.2 to 8.0 × 10⁴/cm² (n = 6) in HB^−/− mice. The difference was not significant. After each subculture, RT-PCR was used to confirm that the mRNA of HB-EGF was not expressed in the isolated HB^−/− MCECs (Fig. 1G). 

RNA Extraction and Reverse Transcription–Polymerase Chain Reaction

Total RNA was extracted from the MCECs (RNeasy Kit; Qiagen, Valencia, CA), and the components were reverse-transcribed to cDNAs (SuperScript Vilo cDNA Synthesis Kit; Invitrogen) according to the instructions of the manufacturer. PCR was performed under the following conditions: 15 minutes at 95°C for the initial denaturation, followed by forty 10-second cycles at 95°C, 20 seconds at 64°C for the annealing step, and 20 seconds at 72°C for the extension. The primer pairs are listed in Table 1. 

In Vitro Corneal Epithelial Cell Wound Healing

HB^−/− and WT MCECs were seeded into 12-well plates coated with type I collagen. After reaching 100% confluence, the cells were starved overnight in CnT-20 (CELLnTEC), a culture medium containing 4 mg/mL insulin and no other supplements (CnT-20-IM). The surface of the plate was scraped with a 1-mL pipette tip (ART 1000E; Molecular Bio Products, San Diego, CA) to create an in vitro corneal epithelial bare area. After the freed cells were washed with PBS, the medium was changed to CnT-20-IM, either with or without 10 ng/mL recombinant human HB-EGF ²⁶ (R&D Systems, Minneapolis, MN). The size of the acellular area was photographed and measured at 0, 6, 12, and 24 hours after the scraping. 

Cellular Proliferation Assay

Cellular proliferation assays were performed with the 5-bromo-2′-deoxy-uridine (BrdU) incorporation assay. HB^−/− and WT MCECs were seeded into six-well plates at a density of 1 × 10⁵ cells/well, cultured until they reached confluence, and then growth factor starved in CnT-20-IM for 24 hours. The confluent cells were scraped to create a bare linear area, and then fresh Cnt-20-IM, with or without 10 ng/mL of recombinant human HB-EGF, was added and incubated for 24 or 48 hours. Unscraped confluent MCECs were used as the control. The cells were then incubated with BrdU labeling reagent for 1 hour. BrdU was detected immunohistochemically with a 5-bromo-2′-deoxy-uridine kit (Labeling and Detection Kit 2; Roche), according to the instructions of the manufacturer. BrdU-positive cells were counted under a microscope in eight random fields that included the leading edge of the wound. 

Cell Adhesion Assay

The cell adhesion assay was performed as described in detail elsewhere. ^27,28 Briefly, HB^−/− and WT MCECs were seeded into 10-cm dishes, cultured until they reached subconfluence, and then growth factor starved with CnT-20-IM for 24 hours. The cells were preincubated in CnT-20-IM, with or without 10 ng/mL HB-EGF for 24 hours and then seeded into 96-well plates at a density of 5 × 10⁴ cells/well. After incubation for 2 hours, the cells were rinsed in PBS (−). The cells adhering to the dish were detached by 0.25% trypsin-EDTA and counted in a counting chamber. Measurements were performed twice for each well. The number of cells adhering to the plate was expressed as a percentage of the number of cells initially seeded (cell adhesion rate). 

Quantitative Real-Time PCR for EGF family

MCECs from HB^−/− and WT mice were seeded into 12-well plates coated with type I collagen. After reaching 100% confluence, the cells were starved overnight in CnT-20-IM. Confluent MCECs were scraped to create a perpendicular bare linear area. After the freed cells were washed off the plate with PBS, the medium was changed to CnT-20-IM, with or without 10 ng/mL recombinant human HB-EGF. The cells were collected at 0.5, 1, 3, 6, and 24 hours. Untouched MCECs were used as the control. Total RNA was extracted from MCECs (RNeasy Kit; Qiagen) and the components were reverse-transcribed to cDNA (SuperScript VILO cDNA Synthesis Kit; Invitrogen). Real-time PCR was then performed with a qPCR kit (DyNAmo STBR Green; Finnzymes, Espoo, Finland) as follows: 95°C for 15 minutes, 40 cycles of denaturation at 95°C for 10 seconds, annealing at 60°C for 20 seconds, and extension at 72°C for 30 seconds (OPticon2 DNA Engine; Bio-Rad, Hercules, CA). The primer pairs of the EGF family used for real time PCR are listed in Table 1. The Ct values were determined by the system software (OPticon2; Bio-Rad), and the amount of each mRNA was calculated relative to the amount of β-actin mRNA in the same samples. ²⁹ Each run was completed with a melting curve analysis, to confirm the specificity of amplification and lack of primer dimmers. 

Expression of Cell Adhesion Molecules in MCECs

To investigate the upregulation of cell adhesion molecules in cultured HB-EGF, we collected MCECs from HB^−/−, and WT mice at 3 hours after creating a bare area. The expression of 88 cell adhesion-related molecules was examined by real-time PCR (Primer Array cell adhesion molecules, Mouse; Takara Bio, Inc., Otsu, Japan), according to the instructions of the manufacturer. Real-time PCR was also performed with a second master mix (SYBR Premix EX Taq; Takara Bio, Inc.) as follows; 95° for 30 seconds, 40 cycles of denaturation at 95° for 5 seconds, and annealing at 60° for 30 seconds (OPticon2 DNA Engine; Bio-Rad). The primer pairs for real-time PCR are listed in Table 1. 

Statistical Analyses

All data are presented as the mean ± SD or SEM. Statistical analyses to determine the significance of the differences between the experimental groups were performed by Student's t-test. The statistical significance level was set at P < 0.05. 

Results

K5-Cre-ROSA-LacZ (K5-Rosa) Mice

We first examined whether the K5 promoter was functional in corneal epithelial cells of K5-Rosa mice. The results showed that β-gal-positive cells were found in almost every layer of the corneal epithelium from the central to the peripheral cornea (Figs. 2A, 2B), the conjunctival epithelium (Fig. 2C), and the epidermis (Fig. 2D). β-gal-positive cells were not found in control mice (Figs. 2E, 2F). These results indicate that K5 Cre-recombinase transgenic mice were able to generate keratinocyte-specific knockout mice. 

Figure 2.

 

X-gal staining of K5-Rosa mice. Uninjured eyes were used for staining. Rosa mice (K5Cre(−)) were used as the control. Stereomicroscopic view of the cornea (A, E) and nuclear red–stained sections (B–D, F) of the central cornea (B, F), conjunctiva (C), and epidermis (D). Scale bars: (B–D, F) 50 μm.

Expression of HB-EGF during Corneal Epithelial Wound Healing

To investigate the expression of HB-EGF during corneal epithelial wound healing, we performed β-gal staining on the corneas of HB^lox/+:K5-Cre (HB^+/−) mice during corneal wound healing. We found very few β-gal-positive cells in the uninjured cornea (Figs. 3A, 3D). Twelve hours after the injury, a larger number of β-gal-positive corneal epithelial cells were seen, predominantly at the leading edge of the wound (Figs. 3B, 3E). The number of β-gal-positive cells was fewer at 24 hours after the injury (Figs. 3C, 3F). 

Figure 3.

 

X-gal staining of HB^−/− mice during corneal wound healing. An epithelial injury of 2.0 mm in diameter was created at the center of each cornea. Stereomicroscopic view (A–C) and nuclear red–stained sections (D–F). Uninjured cornea (A, D) and corneas at 12 (B, E) and 24 (C, F) hours after injury. Arrowheads: leading edge of the wound. Scale bar: (D–F) 100 μm.

In Vivo Corneal Epithelial Wound Healing in HB^−/− Mice

No apparent abnormalities were observed in the HB^−/− mice. Transmission electron microscopy also did not show any abnormalities in the corneal epithelial cells (data not shown). 

To investigate the role of HB-EGF during corneal wound healing, we created a 2-mm-diameter wound and measured the injured area daily. The degree of wound healing was significantly delayed in the HB^−/− mice compared with the WT mice at 18 hours and later. Although the wounds were completely healed by 24 hours in WT mice, 25.1% of the original wound remained in the HB^−/− mice, even 60 hours after the injury (Figs. 4A, 4B). 

Figure 4.

 

Corneal epithelial wound healing in HB^−/− (n = 9) and WT mice (n = 9). (A) Epithelial defects 2.0 mm in diameter were manually created at the center of each cornea. The area of the wound was determined by fluorescein staining. (B) Measurements of the injured area during wound healing. The injured areas were examined daily for 5 days (120 hours) after the injury. Corneal wounds healed after an average of 48 hours in the WT mice; however, wound healing took 86.7 hours in the HB^−/− mice. Data are expressed as the mean + SEM. *P < 0.05.

In Vitro Corneal Epithelial Cell Wound Healing

The role played by HB-EGF in the wound-healing process of MCECs was examined in experiments on cultured MCECs obtained from WT and HB^−/− mice. First, confluent MCECs were scraped to create a bare area, and the area was measured at 6, 12, and 24 hours after the creation (Fig. 5A). With the standard medium without HB-EGF, the size of the acellular area for the MCECs obtained from HB^−/− mice was 70.1% of the original area at 12 hours and 47.5% at 24 hours. In the WT mice, the acellular area had decreased to 49.4% of the original acellular area at 12 hours and to 17.2% at 24 hours. At both times, the acellular area in HB^−/− MCECs was significantly larger than that in the WT MCECs (Fig. 5A). 

Figure 5.

 

Corneal epithelial wound healing is also impaired in HB^−/− MCECs. (A) Representative phase microscopic images showing wound healing in HB^−/− and WT MCECs. (B) Statistical analysis of the extent of MCEC wound healing. Growth factor–starved MCECs were scraped and allowed to heal in Cnt-20-IM in the absence (−) or presence (10 ng/mL; +) of HB-EGF. The remaining acellular area of WT and HB^−/− MCECs was photographed and measured at 0, 6, 12, and 24 hours after scraping. Data are expressed as the mean ± SEM. All experiments were performed in triplicate. *P < 0.01, **P < 0.001. Scale bar, 500 μm.

When human recombinant HB-EGF was added to the WT MCEC media, the acellular area was 78.3% at 6 hours, 39.0% at 12 hours, and 9.1% at 24 hours. In the MCECs obtained from the HB^−/− mice that were exposed to human recombinant HB-EGF, the acellular area was 53.7% at 6 hours, 8.5% at 12 hours, and 0.2% at 24 hours. The acellular areas were significantly smaller in the group that was exposed to HB-EGF than that in the groups that were not exposed to HB-EGF in both WT and HB^−/− MCECs (Fig. 5B). 

In Vitro Cell Proliferation Assays

We determined whether HB-EGF influences the proliferation of MCECs during the wound-healing process by creating an acellular area in confluent MCECs and adding BrdU to the media. The number of BrdU-positive cells was counted at 24 and 48 hours after the acellular area was created, with or without the addition of 10 ng/mL of HB-EGF. When cells were not exposed to HB-EGF, BrdU-positive cells in WT and HB^−/− MCECs were 31.6/1000 and 26.6/1000 at 24 hours, and 19.2/1000 and 16.8/1000 cells at 48 hours after scraping. The differences between the WT and HB^−/− mice were not significant at both 24 and 48 hours. When WT and HB^−/− MCECs were treated with HB-EGF, BrdU-positive cells in WT and HB^−/− MCECs were 83.2/1000 and 49.7/1000 at 24 hours and 28.5/1000 and 17.9/1000 cells at 48 hours after scraping (Fig. 6). The addition of HB-EGF increased the number of BrdU-positive cells significantly in both WT and HB^−/− MCECs at 24 hours, but only in WT at 48 hours after scraping. 

Figure 6.

 

BrdU incorporation during corneal epithelial wound healing in HB^−/− and WT MCECs. Statistical analyses of the number of BrdU-positive cells/1000 cells are shown. Growth factor–starved MCECs were scraped to create a bare linear area, and then fresh Cnt-20-IM, with or without 10 ng/mL of recombinant human HB-EGF, was added and the cells were incubated for 24 or 48 hours. Unscraped confluent MCECs were used as the control (Con). Growth factor–starved unscraped MCECs were incubated with or without HB-EGF for 24 or 48 hours. BrdU-positive cells were counted under a microscope. Data are expressed as the mean ± SD of counts in eight random fields, and each contained the leading edge of a wound. NS, not significant. *P < 0.05, **P < 0.01. All experiments were performed in triplicate.

In addition, BrdU-positive cells in unscraped, confluent MCECs, with or without HB-EGF were also examined. When the cells were not exposed to HB-EGF, a few BrdU-positive cells were observed in unscraped WT and HB^−/− MCECs at 24 and 48 hours. Although the addition of HB-EGF increased the number of BrdU-positive cells significantly in both WT and HB^−/− MCECs at 24 and 48 hours, the number of BrdU-positive cells were fewer than those of MCECs after the scraping. (Fig. 6). 

In Vitro Cell Adhesion Assay

To determine whether HB-EGF influences the MCEC adhesion during the wound-healing process, we performed cell adhesion assays on subconfluent WT and HB^−/− MCECs. The MCECs were seeded onto noncoated 96-well plates, and the cell adhesion rate was calculated after 2 hours. When the cells were not treated with HB-EGF, the cell adhesion rate was significantly lower in the HB^−/− MCECs (12.7%) than in the WT cells (21.8%). The HB^−/− MCECs that were preincubated for 24 hours in HB-EGF showed a significantly higher cell adhesion rate (25.1%) than cells that were not exposed to HB-EGF (Fig. 7). 

Figure 7.

 

Adhesion rates of MCECs. Growth factor–starved MCECs were preincubated in CnT-20-IM, with or without 10 ng/mL HB-EGF for 24 hours and then seeded into 96-well plates at a density of 5 × 10⁴cells/well. After incubation for 2 hours, the cells adhering to the dish were counted, and the result was expressed as a percentage of the number of cells initially seeded (cell adhesion rate). Data are expressed as the mean ± SD of eight samples. *P < 0.01, **P < 0.001. All experiments were performed in triplicate.

Expression of Cell Adhesion Molecules during Corneal Epithelial Wound Healing In Vitro

To investigate which cell adhesion molecules were regulated by HB-EGF during corneal epithelial wound healing, we examined 88 cell adhesion–related molecules by real-time PCR (Primer Array; Takara Bio) in an in vitro wound-healing model. Many cell adhesion molecules were upregulated in the WT MCECs after the injury without HB-EGF treatment. Among these molecules, those that were upregulated by more than two-fold in the HB^−/− MCECs when treated with HB-EGF are shown in Table 2. More complete data are provided in the Supplementary Material. Among these cell adhesion molecules, α6- and β1-integrin were associated with cell-matrix adhesion. ³⁰ Therefore, an upregulation of α6- and β1-integrin was confirmed by regular real-time PCR analysis. The α6-integrin was upregulated by 4.4-fold in WT MCECs and 5.6-fold in HB^−/− MCECs at 3 hours after wounding, and β1-integrin was upregulated by 4.4-fold in WT MCECs and 3.2-fold in HB^−/− MCECs at 3 hours after wounding, when the MCECs were treated with HB-EGF (data not shown). 

Table 2.

 

Primer Array for Cell Adhesion Molecules

Cell Adhesion Molecules	WT		HB−/−
	No Treatment with HB-EGF	HB-EGF 10 ng/mL	No Treatment with HB-EGF	HB-EGF 10 ng/mL
Itga6	3.94	8.51	2.97	7.16
Cd80	ND	ND	3.25	5.06
Itgb1	2.73	6.87	1.51	3.14
Cldn7	2.27	5.39	1.78	4.26
Pvrl3	2.89	16.56	1.31	3.61
Jam2	ND	ND	1.69	5.03

 Cell adhesion molecules upregulated more than twofold by the addition of HB-EGF in HB^−/− MCECs are shown. Each number indicates the amount of each mRNA calculated relative to the amount of GAPDH mRNA in the same samples. Untouched MCECs were used as the control. Itga6, integrin alpha6; Cd80, CD80 antigen; Itgb1, integrin beta 1; Cldn7, claudin 7; Pvrl3, poliovirus receptor–related 3; Jam2, junction adhesion molecule 2; ND, not detected.

Expression of EGF Family in Mouse during Corneal Epithelial Wound Healing In Vitro

To investigate the auto- and cross-induction of the EGF family by HB-EGF during corneal epithelial wound healing, we determined the expression of the mRNA of EGF, TGFα, epiregulin, and HB-EGF by real-time PCR. In WT MCECs, HB-EGF was upregulated by 6.2-fold at 1 hour, and epiregulin was upregulated by 2.7-fold at 3 hours after the injury, whereas HB-EGF was upregulated by 53.4-fold, and epiregulin was upregulated by 20.5-fold at 1 hour after wounding, when the WT MCECs were treated with HB-EGF. However, the expression of EGF and TGF-α was not changed during wound healing, even after exposure to HB-EGF (Figs. 8A, 8B). 

Figure 8.

 

mRNA expression of the EGF family in MCECs during wound healing. The mRNA expression of HB-EGF, epiregulin, EGF, and TGF-α was analyzed by RT-PCR. Total RNA was harvested at each time point in WT (A, B) and HB^−/− (C, D) MCECs, in the absence (A, C) or presence (B, D) of HB-EGF (10 ng/mL). Untouched MCECs were used as the control (Con). Each number represents the amount of each mRNA calculated relative to the amount of GAPDH mRNA in the same samples. Data are expressed as the mean ± SEM. All experiments were performed in triplicate.

On the other hand, epiregulin was upregulated by 4.9-fold in HB^−/− MCECs without HB-EGF and by 38.2-fold at 1 hour after exposure to HB-EGF during wound healing. However, the expression of EGF and TGF-α was not changed during wound healing in HB^−/− MCECs (Figs. 8C, 8D). 

Discussion

We studied K5 promoter-driven, keratinocyte-specific, HB-EGF–deficient mice created by Cre/loxP technology. The results obtained from the K5-Rosa mice showed the efficacy of K5 promoter in the corneal epithelial cells, and the RT-PCR results demonstrated that HB^lox/lox:K5-Cre (HB^−/−) mice had a total loss of HB-EGF in the corneal epithelial cells. This indicated that these K5 promoter-driven Cre-recombinase transgenic mice can be used for corneal epithelial cell research, especially when a particular gene is crucial and a deficiency of that gene is lethal, as in HB-EGF. 

We investigated the role played by HB-EGF in each of the processes of corneal wound healing in the keratinocyte-specific, HB-EGF–deficient mice in vivo and in vitro. First, we demonstrated that corneal epithelial wound healing was significantly delayed in HB^−/− mice compared with WT mice. In addition, the expression of HB-EGF was observed in the corneal epithelial cells predominantly at the leading edge of the wound of the HB^+/− mice. These in vivo results suggested that HB-EGF may play an important role in corneal epithelial wound healing, and thus we performed in vitro studies to investigate the precise mechanism of the delayed wound healing in HB^−/− mice. 

The results of the subsequent wound-healing experiments on cultured MCECs confirmed the in vivo findings. The findings were consistent with those of an in vitro study on rabbit CECs reported by Block et al., ¹⁹ who found that HB-EGF neutralizing antibodies had a greater suppressive effect on injured cultured rabbit CECs than the neutralizing antibodies for other members of the EGF family. Xu et al. ^20,31 and Yin et al. ³² also reported that the migration of injured SV-40 immortalized human CECs was impaired by the inhibitors of HB-EGF, and the addition of HB-EGF enhanced human CECs wound healing. 

We next performed cell adhesion assays, and the in vivo cell adhesion assay demonstrated impaired cell attachments in HB^−/− MCECs, indicating that the delayed corneal wound healing observed in HB^−/− mice may be due to impaired cell adhesion. These findings supported those of Faull et al. ³³ who reported that cell adhesion was enhanced by HB-EGF in human peritoneal mesothelial cells. 

We further examined the cell adhesion molecules that are regulated by HB-EGF, and α6- and β1-integrin were found to be upregulated by HB-EGF exposure in WT MCECs and HB^−/− MCECs. α6-Integrin has been reported to be expressed at high levels in corneal epithelial cells, the α6β1 heterodimer has been found in the basal and suprabasal epithelial cells, and the α6β4 heterodimer has been found in the basal epithelial cells of the cornea. ³⁴ β1-Integrin has been reported to be expressed ubiquitously in corneal epithelial cells and to interact with many α subunits to yield the heterodimers. ³⁵ Thus, the α6β1 heterodimer may be the most likely subunit candidate, of those that are regulated by HB-EGF, to be involved in the attachment of mouse corneal epithelial cells. However, more definitive investigations are needed. 

On the other hand, the cell proliferation assay showed that the number of BrdU-positive cells was not significantly different between WT and HB^−/− MCECs without HB-EGF treatment. This result indicates that HB-EGF did not promote cell proliferation but most likely promoted MCEC migration by increasing cell adhesion. Our results are in good agreement with those of an earlier report showing that the delayed wound healing in the skin of HB^−/− mice is due to an impairment of cell migration rather than to proliferation. ²¹  

The degree of wound healing improved when the HB^−/− MCECs were exposed to HB-EGF. We also found that the addition of HB-EGF increased the number of BrdU-positive cells, although the increased number of BrdU-positive cells was more significant in WT MCECs than in HB^−/− MCECs. HB-EGF exists as a membrane-anchored protein, pro-HB-EGF, and is released into the extracellular domain by injury or pharmacological stimulation (ectodomain shedding). ^36,37 Soluble HB-EGF binds to the EGFRs in an autocrine or paracrine manner. It has been reported that the intracellular domain of the pro-HB-EGF translocates from the plasma membrane into the nucleus after shedding, where it deactivates the suppression of cellular proliferation. ^38,39 Thus, when injured HB^−/− MCECs were treated with external soluble HB-EGF, the signaling pathways through the EGFRs were activated, but the intracellular domain of pro-HB-EGF was not activated, because of the lack of pro-HB-EGF. Therefore, it is possible that the intracellular domain of pro-HB-EGF is necessary for cell proliferation, because the intracellular domain of pro-HB-EGF can translocate from the plasma membrane to the nucleus after shedding and deactivate the suppression of cell proliferation in WT MCECs. 

One other factor that should be considered is the auto- and/or cross-induction function in the EGF family shown in recent studies. ^40,41 Our results indicated that the increased expression of HB-EGF by auto-induction when treated with HB-EGF may promote MCEC proliferation by a translocation of the intracellular domain of pro-HB-EGF in WT MCECs. Despite the lack of HB-EGF upregulation and the absence of an intracellular domain for pro-HB-EGF in HB^−/− MCECs, an increase in the number of BrdU-positive cells was observed in HB^−/− MCECs exposed to HB-EGF. However, the increase was less than that in WT MCECs. These findings may be explained by a cross-induction function of the EGF family members, because the mRNA expression of epiregulin in HB^−/− MCECs was increased by 4.9-fold when injured and as much as 38.2-fold when exposed to HB-EGF. 

It has been demonstrated that epiregulin dose-dependently increases the proliferation of CECs. ⁴⁰ The upregulation of epiregulin may compensate for the impaired proliferation ability due to lack of HB-EGF in HB^−/− MCECs. Such cross-induction among the EGF family may be a compensation mechanism to maintain homeostasis of CECs. Further investigations are needed to determine whether this mechanism contributes to corneal epithelial cell proliferation. 

In summary, during corneal epithelial wound healing, HB-EGF was found to be more closely involved with promoting cell migration by cell attachment than with cell proliferation. The impaired proliferative ability of HB^−/− MCECs could be compensated for by other members of the EGF family. Future research should focus on clarifying the intracellular signaling pathways of HB-EGF in the corneal epithelial cells. 

Supplementary Materials

Supplementary Table S1 - (.xls) - The whole results of Primer Array® for cell adhesion molecules. Each number indicates the amount of each mRNA calculated relative to the amount of GAPDH mRNA in the same samples. n.d.: not detected. 

Supplementary Table S2 - (.xls) - Full spelling of cell adhesion molecules. 

Footnotes

 Supported by a grant-in-aid for Scientific Research (C) 20659720 from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Footnotes

 Disclosure: R. Yoshioka, None; A. Shiraishi, None; T. Kobayashi, None; S. Morita, None; Y. Hayashi, None; S. Higashiyama, None; Y. Ohashi, None