Secondary Logo

Journal Logo

Laboratory science

Hepatocyte growth factor and keratinocyte growth factor regulation of epithelial and stromal corneal wound healing

Carrington, Louise M. PhD; Boulton, Mike PhD

Author Information
Journal of Cataract & Refractive Surgery: February 2005 - Volume 31 - Issue 2 - p 412-423
doi: 10.1016/j.jcrs.2004.04.072
  • Free
  • Take the CME Test

Hepatocyte growth factor (HGF) and keratinocyte growth factor (KGF) are classically defined as paracrine mediators of epithelial cell function in many tissues1–3 and as such have been implicated as regulators of the healing process in epidermal and esophageal wounds.2–7

Hepatocyte growth factor, a heparin-binding dimer in its mature form, is secreted predominantly by fibroblasts as an inactive, single-chain precursor.2 The high affinity receptor for HGF, a product of the proto-oncogene c-Met, is a heterodimeric transmembrane tyrosine kinase receptor that is expressed principally on epithelial cells. This pattern of epithelial cell receptor expression and ligand secretion by fibroblasts is replicated by KGF-1 (fibroblast growth factor [FGF]-7).1 Keratinocyte growth factor-1, a heparin and collagen-binding growth factor,3 signals exclusively to cells via a splice-variant of the FGF receptor 2.8 Both growth factors initiate a PI-3K signaling cascade4 and are able to regulate the proliferation, differentiation, gap junctional communication, and migration of diverse epithelia.2,3

However, the HGF/KGF signaling pattern may be more complex in the cornea. Corneal epithelial cells are capable of secreting KGF and express low levels of HGF mRNA,9–13 while keratocytes express high-affinity receptors for KGF and HGF.9,10,12,13 The tear fluid provides another source of HGF.12,14 This suggests a more complex function for these 2 growth factors in homeostasis than occurs in other tissues. Following superficial or penetrating wounds, the availability of HGF and KGF and the expression of their receptors are dramatically increased in the corneal environment.12,13 Levels of both growth factors and their receptors start to return to normal after epithelial coverage of the wound, suggesting a regulatory role in the early healing response.13,14

In this study, we investigated the role of HGF and KGF in corneal wound healing using 2 serum-free culture systems that increase in complexity and mimic the in vivo situation more closely than serum-containing techniques.15,16 The effect of HGF and KGF on key events during epithelial and stromal repair (rate of reepithelialization, epithelial stratification, keratocyte proliferation, repopulation of the stroma, and myofibroblast differentiation) and the consequences of stimulated changes in 1 cell layer on other layers are reported.

Materials and Methods

Keratocyte Cell Culture

Nonactivated bovine keratocytes were cultured using the method of Jester and coauthors.15 Briefly, the corneal stroma was isolated by scraping off epithelial and endothelial cell layers and incubated in 2 mg/mL collagenase (weight per volume modified Eagle medium [MEM] at 37°C [Gibco BRL]) overnight at 37°C. Isolated cells were plated in the wells of 24-well plates at 1 × 104 cells/well in 1 mL of serum-free MEM containing nonessential amino acids (Gibco BRL), RPMI 1640 vitamins and glutathione (Gibco BRL), glutamine, antibiotic agents, and fungizone. Cultures were maintained in MEM at 37°C in 5% carbon dioxide with 95% air in a standard incubator for 48 hours to attach. The medium was replaced with fresh medium containing 0.1, 1, and 10 ng/mL HGF or KGF or growth factor diluent (in triplicate) and cultured for up to 5 days.

Assessment of Proliferation in Cell Culture

Crystal Violet Assay.

The crystal violet assay was carried out according to the method of De Saint Jean et al.17 Briefly, cells were fixed (room temperature, 10 minutes, 70% ethanol) and incubated with a 0.5% solution of crystal violet (BDH) (room temperature, 1 minute). After the cells were washed, acetic acid (33%, volume per volume deionized H2O) was added to the wells to elute the dye. The absorbance of the wells was read at 540 nm using an enzyme-linked immunosorbent assay reader (Sigma Diagnostics), and wells containing 33% acetic acid were used as a blank; readings were converted to cell numbers using a standard curve of cells plated 24 hours in advance (range 1 × 103 – 5 × 105; n=3 per density for crystal violet and 3 to be counted using a hemocytometer).

Bromodeoxyuridine Incorporation.

Bromodeoxyuridine (BrdU) incorporation was carried out to confirm the proliferative ability of bovine keratocytes in this culture system with the crystal violet assay, as previous investigators report that this phenomenon does not occur.15,18–20 Briefly, BrdU (Sigma Diagnostics) (3.0 μg/mL) was incubated with cell cultures plated at 1 × 104 cells/well in standard incubator conditions for 24 hours following a 48-hour attachment period. Cells were fixed in acid/ethanol (room temperature, 5 minutes), permeablized with 0.1% Triton-X-100 (room temperature, 20 minutes) and DNA-denatured using 4N hydrochloride (room temperature, 30 minutes). Cells were incubated with mouse anti-BrdU 1/500 (Sigma Diagnostics) overnight at 37°C, followed by a peroxidase-conjugated, goat anti-mouse secondary (Sigma Diagnostics) (1 hour), Extravidin tertiary complex (Sigma Diagnostics) (30 minutes), and amino ethylcarbazole (Sigma Diagnostics) (5 minutes), according to the manufacturer's instructions. Photomicrographs were taken of cells using bright field (BrdU staining) and phase contrast (morphology).

Positive controls in which cells were exposed to medium containing 10% fetal calf serum (FCS) (Gibco BRL) and negative controls in which BrdU or anti-BrdU antibody was omitted from the assay were carried out concurrently.

Migration in Response to Scrape Wounding

Keratocytes were plated at 1 × 105 cells per well in a 24-well plate, the higher density of cells allowing easier quantification of migration. After a 48-hour attachment period, cultures were scrape wounded using a sterile 100 μL micropipette tip and the medium was changed. Cells were challenged with 0.1 to 100 ng/mL of HGF or KGF. Medium containing 0.1% growth factor diluent (phosphate-buffered saline [PBS] containing 0.1% bovine serum albumin) acted as a control. Cultures were photographed 4 times per well along the wound at 0 and 8 hours (empirically determined as the optimum time), and the mean distance between the leading edges of the cultures was determined using image-analysis software (ImagePro Plus, Image Solutions). Distance traveled is expressed as a percentage of the original wound width.

Identification of Myofibroblasts

Cultured cells were fixed at 0, 1, 2, 3, 4, and 5 days posttreatment with HGF, KGF, or 10% FCS as a positive control in 1% paraformaldehyde. Cells were pretreated with 0.1% Triton-X-100 for 20 minutes, incubated with a monoclonal anti-α-smooth muscle actin (SMA) antibody (Sigma Diagnostics) for 2 hours followed by a fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulin (IgG) (Sigma Diagnostics) for 1 hour. Incubation in a bisbenzimide solution (0.5 mg/mL in PBS) for 10 minutes was done to counterstain the nuclei. Cell cultures were also stained with tetramethylrhodamine isothiocyanate-conjugated phalloidin (Sigma Diagnostics) (5 μg/mL in PBS) for 2 hours and counterstained with bisbenzimide.

Organ Culture

Bovine corneas were used in the study partly because most human corneas are required for transplantation and partly because healing in the bovine cornea has been shown to be similar to that in humans.16 Briefly, bovine corneas were centrally wounded with a 5.0 mm trephine, as described,16 and the disk of epithelial/stromal tissue within the wound was excised. Sterile, serum-free Dulbecco's MEM (DMEM) containing 1% agar and 1% gelatin (BDH) was used as a support, and serum-free Trowell's T8 medium (Gibco BRL) containing antibiotics, fungizone, and glutamine was added to the dish to a level just below the limbal region. Twice daily, 100 μL of fresh, serum-free T8 medium was pipetted onto the surface of the cornea containing 1, 10, or 100 ng/mL HGF or KGF or growth factor diluent. Cultures were maintained in serum-free T8 medium for up to 5 days (n=6 per time, per treatment).


Reepithelialization was assessed as described16 using captured macro images of the wound area, in which the original wound outline and the leading edge of the epithelium could be seen. Reepithelialization was expressed as the percentage of the original wound area recovered.

Morphological Assessment of Epithelium

Corneas (wounded and unwounded) were fixed in 10% neutral buffered formalin (NBF) overnight, processed into wax, sectioned (7 μm), and floated onto 1% 3-aminopropyltriethoxysilane-coated slides (Sigma Diagnostics). After deparaffinization and rehydration, sections were stained with Harris hematoxylin and eosin (BDH).

Stromal Cell Density Beneath the Wound

Corneas were fixed in 10% NBF overnight and processed into wax. They were cut into 7 μm sections, floated onto glass slides, dewaxed, rehydrated, and incubated with 0.5 mg/mL bisbenzimide solution for 10 minutes. Photographs of 3 sections of each wound area were taken using an Olympus OM-2 camera attached to an Olympus IX-70 microscope. Photographs were scanned, reproduced digitally, and stored electronically. Boxes with internal dimensions of 100 μm × 100 μm were created using Photoshop 5.0 (Adobe). These were pasted into a transparent layer on top of the digital images, rotated, and repositioned until they abutted the wound face. The number of nuclei within each box was counted.

Identification of Myofibroblasts

Corneal organ cultures were snap frozen using liquid nitrogen and embedded in OCT (Tissue-Tek) 0, 1, 2, 3, and 5 days posttreatment, sectioned at 5 μm intervals, and floated onto poly-L-lysine-coated slides. Sections were pretreated with 0.1% Triton-X-100 for 20 minutes and incubated with a monoclonal anti-α-SMA antibody for 2 hours followed by a FITC-conjugated goat anti-mouse IgG for 1 hour. Incubation for 10 minutes in a bisbenzimide solution (0.5 mg/mL in PBS) was done to counterstain the nuclei.

Statistical Analysis

Cell counts beneath the wound were found to approximate a Poisson distribution using a Kolmogorov-Smirnov test and were compared using the nonparametric Mann-Whitney U test. Cell numbers in the crystal violet assay, percentage migration, and percentage of wound reepithelialization approximated normal distributions (Kolmogorov-Smirnov test); groups were compared using unpaired 2-way t tests (Prism 3.0, Graphpad Software). The rate of reepithelialization was assessed between 24 hours and 48 hours using a least-squares regression function of data.


Rate of Reepithelialization

The epithelium of the unwounded bovine cornea usually comprises 8 to 12 layers of cells: a basal columnar layer, 5 to 9 layers of suprabasal or wing cells, and 2 to 5 layers of flattened superficial epithelial cells (Figure 1, a). Central excisional trephine wounds in bovine corneas led to the removal of all epithelial layers along with a portion of the underlying stroma within the demarcated zone (Figure 1, b). In control corneas receiving no growth factor treatment, epithelial migration commenced within the first 12 hours and by 72 hours, all wounds had reepithelialized (Figure 2). The rate of reepithelialization was calculated to be 1.56% ± 0.07% h−1.

Figure 1.
Figure 1.:
Epithelial morphology in unwounded bovine corneas (a) showing columnar basal cells and superficial flattened epithelia sloughing from the surface (black arrow) and (b) immediately following trephine wounding and excision of epithelial/stromal disk of tissue (bars = 100 μm).
Figure 2.
Figure 2.:
Percentage of reepithelialization of bovine corneal wounds treated with 100 μL of serum-free medium twice daily containing 1, 10, or 100 ng/mL HGF or KGF. Each point represents the mean of at least 6 corneas ± SEM. Significant differences between growth factor-treated corneas and controls were determined by the Student t test ( P<.05).

Hepatocyte growth factor and KGF had opposite effects on the rate of reepithelialization, with HGF delaying and KGF accelerating the process (Figure 2). Although the percentage of the wound covered by migrating epithelia was significantly lower than that in the untreated controls for all doses of HGF at all times up to 72 hours (P<.001), the effect was most pronounced with the 10 ng/mL dose and only 10 and 100 ng/mL slowed the reepithelialization rate significantly (0.47% ± 0.1% h−1 and 0.81% ± 0.2% h−1, respectively; P<.05). In corneas receiving the highest dose of KGF (100 ng/mL), the percentage of the wound covered by 48 hours increased significantly and the reepithelialization rate increased to 2.04% ± 0.9% h−1 (P<.05). Lower doses of KGF had no effect.

Epithelial Cell Morphology

Epithelial cell morphology, differentiation, and stratification appeared normal at sites distal to the wound at 0 hour. After 4 hours, the epithelium had rounded off and retracted 100 to 200 μm (Figure 3, a). At commencement of migration, the leading edge of the epithelium was commonly 1 cell deep, smoothly increasing in thickness toward the original wound edge. Cells at the leading edge appeared more rounded with no obvious stratification; columnar, basal, flattened superficial cells could not be morphologically distinguished for 300 μm. After 24 hours, the leading edge had increased to 3 to 4 cells in depth and slight hypercellularity at the original wound edge was evident, with some thinning of the epithelial layers over the original wound edge (Figure 3, b). By 72 hours, wound closure had occurred, confirming the results of the macroscopic image analysis. The site of closure was thinner and less differentiated than the distal epithelium, and hypercellularity at the original wound edge had increased (Figure 3, c). After 5 days in culture, the entire epithelium within the wound site exhibited stratification; columnar basal cells were evident abutting the stroma, and the superficial epithelium was populated with flattened cells. Slight thickening of the epithelium within the wound was common. Controls that received medium containing growth factor diluent were identical to those that received medium alone.

Figure 3.
Figure 3.:
Epithelial morphology during wound healing. In control corneas (ac), the epithelium is seen to round and retract from the cut stromal edge (black arrow) 4 hours post wounding (a). Some stratification of the epithelium can be seen as migration continues at 24 hours (b); by 72 hours, the wound has completely reepithelialized (c). In HGF-treated corneas (df), the epithelium exhibited separation of the basal layer from the stroma and suprabasal cells (white arrows) 24 hours post wounding and later (e), with some corneas showing epithelial debris in front of the leading edge of the migrating epithelium (black arrow) (f). In KGF-treated corneas (gi), in contrast to controls, migration began 4 hours post wounding (g). Hypercellularity was observed in the epithelium surrounding the wound by 120 hours, although stratification was still evident (i) (bars = 100 μm).

The delaying effect of HGF on reepithelialization was confirmed by histologic examination. Neither retraction nor migration of the epithelium had occurred by 4 hours post wounding in HGF-treated corneas, and epithelial cells did not appear to be rounding at the edge (Figure 3, d). By 24 hours, the migrating epithelium had reached the wound face and extreme hypercellularity was evident, increasing with higher doses. Superficial, flattened epithelial cells were abundant, but basal cells were difficult to distinguish. By 48 hours, all doses revealed an extremely hypercellular leading edge, with basal cells appearing to detach from more superficial layers (Figure 3, e). In corneas treated with 100 ng/mL HGF, evidence of cellular debris in front of the leading epithelial edge could be seen, apparently attached to the stroma (Figure 3, f). This pattern continued until wound closure (120 hours). Extreme thickening of the epithelium was observed, increasing in higher doses; at 100 ng/mL, separation of the suprabasal cells from the basal cells was noted.

In contrast to control and HGF-treated corneas, migration of the epithelium into the wound area had started by 4 hours in corneas that received KGF (Figure 3, g). Minimum thinning of the epithelium was evident at 100 ng/mL but was more pronounced at lower doses. Good stratification of the epithelium was demonstrated until 72 hours (Figure 3, h), when although basal cells were clearly distinguishable, few flattened superficial epithelial cells were apparent. Thinning of the leading edge was evident until wound closure, as was hypercellularity of the distal epithelium within the wound (increasing with dose strength). This hypercellularity encompassed the entire wound area by 120 hours in corneas treated with 100 ng/mL (Figure 3, i).

Keratocyte Proliferation in Cell Culture

The culture conditions allowed proliferation of keratocytes throughout the 5 days of experimentation, increasing in a linear manner; by 120 hours, 1.8 ± 0.2 × 105 cells per well were seen (Figure 4, a and b). The proliferative ability of keratocytes in this culture system in the absence of growth factors was confirmed by BrdU incorporation in approximately 50% of nuclei during a 24-hour incubation period (Figure 4, c and d).

Figure 4.
Figure 4.:
Effect of HGF (a) and KGF (b) on keratocyte numbers in serum-free cell culture assessed using the crystal violet assay. Significant differences between growth factor-treated cultures and controls were determined by the Student t test (a and b) ( P<.05, ∗∗ P<.01). Incorporation of BrdU into the nuclei of proliferating keratocyte cells in serum-free culure: Phase contrast (c) shows all cells and their stellate morphology, and bright field (d) shows nuclei incorporating BrdU (red stain). Black arrows indicate the position of unstained, nonproliferating nuclei and red arrows, stained, proliferating nuclei.

Keratocyte proliferation was stimulated by 10 ng/mL of HGF (Figure 4, a), which increased the number of cells at 96 hours to double that seen in controls (2.8 ± 1.3 × 105 cells per well; P<.01), and increased to 3.5 ± 1.1 × 105 cells per well by 120 hours (P<.01). Lower doses had no effect.

After 96 hours in cell culture, all concentrations of KGF had increased the proliferation of keratocytes with roughly equal efficiency to almost twice the number of cells seen in control wells (P<.01) (Figure 4, b). At 120 hours, however, KGF appeared to be most effective at lower doses, with numbers in 10 ng/mL-treated wells remaining approximately constant at 2.8 ± 0.4 × 105 cells per well (P<.01 compared to controls). A dose of 1 ng/mL increased cell numbers to more than double those in control wells (3.7 ± 0.6 × 105; P<.01) and of 0.1 ng/mL, to 2.5 times (4.1 ± 1.0 × 105; P<.01).

Migration of Keratocytes

The leading edge of keratocytes in untreated cultures had migrated a mean of 73.7 ± 24 μm after 8 hours in culture and covered 33.5% ± 10.9% of the original wound. Both HGF and KGF significantly increased the rate of migration (Figure 5). Hepatocyte growth factor increased the motogenic activity of keratocytes in a dose-dependent manner, with 100 ng/mL increasing the rate of migration by 2.5 times to 23.9 ± 9.8 μm h−1. Keratinocyte growth factor stimulated migration at low doses but had no effect at 10 and 100 ng/mL.

Figure 5.
Figure 5.:
Effect of HGF and KGF on keratocyte migration following scrape wounding in serum-free cell culture. Each point represents the mean of 3 experiments of triplicate wells ± SEM. Significant differences between growth factor-treated cultures and controls were determined by the Student t test ( P<.05, ∗∗ P<.01).

Repopulation of the Stroma Proximal to the Wound

Immediately after the corneas were scraped, keratocyte numbers in the 100 μm below the wound face dropped to 55.52% ± 2.01% of the numbers in the unwounded corneas (Figure 6). A further reduction is seen at 4 hours post wounding (32.49% ± 1.47%); however, by 24 hours, keratocytes were beginning to repopulate the area under the wound (39.65% ± 2.46%). Over the next 72 hours, the repopulation continued; by 120 hours, the number of cells beneath the wound face had increased to 60.11% ± 7.45% of that in the unwounded corneas (Figure 6).

Figure 6.
Figure 6.:
The effect of HGF (a) and KGF (b) on keratocyte numbers beneath a trephine wound in organ-cultured corneas between 0 hour and 120 hours post wounding. Points represent the mean number of cells under the wound as a percentage of the number at a comparable depth in unwounded controls. Each point is the mean of 6 corneas per treatment, and error bars represent the SEM. Data were analyzed for differences between treatment groups and the control corneas using the Kruskal-Wallis test ( P<.05, ∗∗ P<.01). Significance symbols for 4-hour time points are not shown due to lack of space.

Lower doses of HGF (1 and 10 ng/mL) had no effect on the number of keratocytes under the wound (P>.05) (Figure 6, a). The addition of 100 ng/mL of HGF resulted in an increase to 52.7% ± 3.1% in unwounded controls by 120 hours (P<.05).

The lowest concentration of KGF (1 ng/mL) had no effect on keratocyte numbers beneath the wound (P>.05) (Figure 6, b). Corneas treated with 10 ng/mL had the same number of cells in the 100 μm under the wound up to 72 hours; by 120 hours, the number dropped to below the level in wounded controls (36.25% ± 3.00%; P<.05). With KGF of 100 ng/mL, the number of cells under the wound decreased at all time points post 48 hours (72 hours, 42.61% ± 1.68%, P<.05; 120 hours, 55.44% ± 6.08%, P<.01).

Myofibroblast Differentiation

Bovine keratocytes in serum-free cell culture retained a stellate morphology similar to that seen in vivo, forming a monolayer with interconnecting processes (Figure 7, a). Myofibroblasts were not seen in cells in culture in controls or in HGF-treated or KGF-treated cells at any time up to 5 days. Keratocytes maintained their stellate morphology and did not express detectable levels of α-SMA (Figure 7, j and l). Phalloidin-staining of f-actin of cells in culture showed a stellate array of actin fibers radiating from the nucleus along the cellular processes and a diffuse cortical nuclear staining in untreated and growth factor-treated cells (Figure 7, g and i).

Figure 7.
Figure 7.:
Phenotype of keratocytes in cell culture. Cells in normal culture conditions (a, d, g, j) maintained a stellate morphology up to 5 days in culture (a) in a monolayer with interconnecting processes. Cells showed predominately cortical f-actin staining with phalloidin (g) and did not stain with antibodies against α-SMA (j). In contrast, cells in the presence of 10% serum for 5 days (b, e, h, k) became activated, multilayered, and spindle-shaped (b). Phalloidin staining showed bundles of actin forming stress fibers (h), and up to 10% of cells were positive for α-SMA, mostly grouped together (k). Both HGF-treated cells (100 ng/mL) (c, f, i, l) and KGF-treated cells (not shown) showed morphology similar to that of controls and lacked staining for α-SMA. Images with blue fluorescence were stained with bisbenzimide to show nuclei (bars = 100 μm).

Serum-treated cells took on a fibroblastic morphology (Figure 7, b) and by 5 days, subpopulations in culture were seen to express α-SMA (Figure 7, k), usually in groups of elongated cells in areas of the culture that were multilayered. Actin filaments in these cells appeared to be bundled into prominent stress fibers (Figure 7, h), which together with the presence of α-SMA confirmed the presence of myofibroblasts.

Organ cultures without growth factor treatment were almost devoid of the myofibroblast marker, with occasional, isolated cells labeled 120 hours post wounding beneath the wound area in some corneas. In the KGF-treated corneas, there was a similar lack of myofibroblasts in the stroma (data not shown).

In contrast to cells in culture, HGF added to organ-cultured corneas resulted in the differentiation of keratocytes into myofibroblasts. This effect was seen only after 5 days in culture with 100 ng/mL of HGF. Swathes of cells under the wound were positive for α-SMA; these were confined to the area beneath the wound and in 2 of the 6 corneas, all cells to the depth of the endothelium expressed the myofibroblast marker (Figure 8).

Figure 8.
Figure 8.:
Phenotype of stromal cells proximal to the wound in organ-cultured corneas. Positive staining of α-SMA was evident in cells lining the lumen of blood vessels in the corneal limbus (a), providing a positive control. Control corneas, which had been wounded and maintained in organ culture receiving medium or medium plus growth factor diluent alone (b), showed no or rare isolated cells on staining with α-SMA (white arrow). In contrast, organ-cultured corneas treated with 100 ng/mL of HGF showed intense staining throughout the depth of the cornea 120 hours post wounding (c) (bars = 100 μm).


The availability of HGF and KGF and the expression of their receptors in the corneal environment increase dramatically following a wound.9–14 The reasons for this upregulation and its role in the corneal repair process remain unclear. In this study, we present evidence that both growth factors are modulators of key events in corneal epithelial and stromal wound healing, acting as autocrine and paracrine regulators. The effect of HGF and KGF on corneal epithelial cells in serum-containing in vitro culture systems is well documented; both factors stimulate proliferation, but only HGF stimulates migration.21–23 These findings suggest that both should be capable of enhancing the rate of reepithelialization, as found using KGF following photorefractive keratectomy (PRK) in animal models24 and with HGF in immersion organ-cultured rabbit corneas.4 Our findings using a defined, serum-free medium with KGF agree with this earlier work; however, the retardation of reepithelialization by HGF was surprising. The difference is likely due to the presence of serum in the other studies, which involve a variety of growth factors known to modulate KGF/HGF expression and responsiveness in the cornea.25–27 Alternatively, this could be a species-specific response.

Both growth factors are reported to signal via a PI-3K/p70 s6 kinase pathway in corneal epithelial cells,4 although HGF initiates signaling faster and to a greater extent than KGF. This difference in the degree of signaling may lead to opposing reepithelialization responses, an effect that could be mediated by matrix metalloproteinase (MMP).28–31 In corneal and other epithelia, differential expression of MMP-9 is demonstrated to be under the influence of KGF and HGF and is believed to be related to the duration of mitogen-activated protein kinase (MEK) and extracellular signal-regulated kinase (ERK) activation in these cells.28,32,33

Histologic examination of HGF- and KGF-treated corneas showed that not only did these agents affect the extent of epithelial coverage; they also altered the quality of the architecture and morphology of the epithelium. The epithelia treated by KGF were better stratified but thicker than controls. It is not clear whether this represents an enhancement of wound healing, but overexpression of/treatment with KGF results in developmentally abnormal epithelia in the corneas, vulvae, and uteri of mice.34,35 Treatment by HGF seems to have an adverse effect on the epithelial architecture, with apparent separation of the basal layer of cells from the stromal substrate and the suprabasal epithelia. This is perhaps not surprising since HGF is also known as “scatter factor” and is associated with the disassembly of cell−cell contacts.2,36 The HGF-treated epithelia appeared to heal in a “2-steps forward, 1-step back” fashion, with the cellular debris of former, failed reepithelialization located in front of the leading edge.

A recent study by Daniels et al.28 shows that high doses of HGF (>20 ng/mL) result in a delayed migration of cultured human corneal epithelial cells. Hepatocyte growth factor has been shown to mediate the disruption of many cell−cell junctions including desmosomes, hemidesmosomes, 2 and gap junctions,37 all of which are present in the corneal epithelium.38 It is likely that HGF delayed reepithelialization in our model by interrupting junctions at the leading edge and in the basal-apical axis of the cell, which would have been loosened as part of the healing process,39 causing these layers to separate and shed. The HGF-induced shedding of epithelium as it migrates is a phenomenon not seen in normal corneal wound healing and is likely to more closely resemble an aberrant wound-healing response.

Analysis of the architecture of the epithelium in animal models of spontaneous corneal erosions40 shows similarities with the abnormal epithelium seen in our model, with separation of the basal-epithelial layer from the suprabasal. In addition, recurrent corneal erosions have been successfully treated in humans by therapies that inhibit MMP-9,41 a gelatinase whose production is demonstrated to be upregulated by corneal epithelial cells following HGF treatment.25 Levels of HGF in the tear fluid/corneal microenvironment of patients suffering from recurrent/spontaneous corneal erosions are at present unknown; however, it would be interesting to determine whether they are elevated in these individuals. Published levels of HGF in the tear fluid following wounding (PRK) are in the 0.3 to 0.5 ng/mL range,14 concentrations significantly below those at which delayed reepithelialization is seen in our model, although the bovine corneas receive a total HGF dose over 12 hours similar to published levels when the rate of HGF availability (28 pg/min)14 is taken into account.

Bovine keratocytes maintained in serum-free culture were seen to proliferate, in contrast to previous reports.15,18–20,42 The reasons for this are unclear but are probably due to the experimental design including species (bovine verses rabbit)15,46 and medium differences. Reports that indicate bovine keratocytes in serum-free media do not proliferate were carried out in DMEM/Hams F12 containing high concentrations of thymidine,18–20 which probably acts as a competitive inhibitor of BrdU and tritiated thymidine incorporation in both assays, leading to an underestimation of the proliferative ability of keratocytes in this culture system.

Studies of keratocyte responses to HGF and KGF have been carried out in cells cultured in the presence of serum, which produces a myofibroblast phenotype,15 most commonly present in the corneal stroma after epithelial wound closure.43 At this point, the availability of HGF and KGF is reportedly beginning to return to nonpathologic levels.12–14,44,45 Keratocytes in a cell culture system that better maintains the in vivo phenotype15 responded both proliferatively and motogenically to HGF and KGF in our study. This is in contrast to reports of the effect of KGF and HGF on corneal fibroblasts in serum-containing medium.9,46 Therefore, it appears that the reaction of corneal cells to HGF and KGF is phenotype dependent. If this is the case, it would indicate that HGF is a key regulator of keratocyte density in the cornea: Immediately after a wound, stromal cells would be prompted to migrate and proliferate by the increase of HGF and KGF in their environment. This, in turn, would lead to repopulation of the corneal stroma proximal to the wound following the well-documented keratocyte death in this area.47 However, an early, dramatic increase in cell numbers in the wounded stroma on treatment with these growth factors was not seen. Hepatocyte growth factor prompted an increase in stromal cells by 3 days post wounding in the organ culture model, while KGF decreased the number of cells beneath the wound.

It is possible that repopulation of the stroma is inversely proportional to the extent of reepithelialization; ie, the effect of HGF on the number of cells beneath the wound edge is related to the delayed reepithelialization of the wound seen with HGF treatment and not to a direct effect of HGF on the keratocytes themselves. Alternatively, the state of the epithelium—stratified and hypercellular in KGF-treated corneas and disintegrating in HGF-treated corneas—may cause the difference as epithelial cells in varying states of differentiation have distinct cytokine production profiles.48

Both explanations may also account for the dramatic appearance of myofibroblasts in wounded corneas in response to HGF, a phenomenon not seen in cell culture or previously. An alternative explanation would be that HGF increases production of MMP-9 by corneal epithelial cells,28 which in turn cleaves the latent extracellular matrix-bound form of TGF-β,49 activating a potent signal for myofibroblast differentiation.

The results of this study demonstrate that KGF and HGF play an important role in the regulation of epithelial and stromal cell behavior. It also highlights the difference between cultured cells and those in organ culture in which the architecture and many of the cell−cell interactions are maintained. The data from the organ culture model are likely to most closely resemble the in vivo situation and suggest that neutralizing the effects of high concentrations of HGF may be a worthwhile therapeutic intervention in corneal wound repair.


1. Finch PW, Rubin JS, Miki T, et al. Human KGF is FGF-related with properties of a paracrine effector of epithelial-cell growth. Science 1989; 245:752-755
2. Grierson I, Heathcote L, Hiscott P, et al. Hepatocyte growth factor/scatter factor in the eye. Prog Retin Eye Res 2000; 19:779-802
3. Beer H-D, Gassmann MG, Munz B, et al. Expression and function of keratinocyte growth factor and activin in skin morphogenesis and cutaneous wound repair. J Invest Dermatol Symp Proc 2000; 5:34-39
4. Chandrasekher G, Kakazu AH, Bazan HEP. HGF- and KGF-induced activation of PI-3K/p70 s6 kinase pathway in corneal epithelial cells: its relevance in wound healing. Exp Eye Res 2001; 73:191-202
5. Takahashi M, Ota S, Ogura K, et al. Hepatocyte growth factor stimulates wound repair of the rabbit esophageal epithelial cells in primary culture. Biochem Biophys Res Commun 1995; 216:298-305
6. Cowin AJ, Kallincos N, Hatzirodos N, et al. Hepatocyte growth factor and macrophage-stimulating protein are upregulated during excisional wound repair in rats. Cell Tissue Res 2001; 306:239-250
7. Baatar D, Kawanaka H, Szabo IL, et al. Esophageal ulceration activates keratinocyte growth factor and its receptor in rats: implications for ulcer healing. Gastroenterology 2002; 122:458-468
8. Miki T, Bottaro DP, Fleming TP, et al. Determination of ligand-binding specificity by alternative splicing: two distinct growth factor receptors encoded by a single gene. Proc Nat Acad Sci USA 1992; 89:246-250
9. Wilson SE, Walker JW, Chwang EL, He Y-G. Hepatocyte growth factor, keratinocyte growth factor, their receptors, fibroblast growth factor receptor-2, and the cells of the cornea. Invest Ophthalmol Vis Sci 1993; 34:2544-2561
10. Sotozono C, Kinoshita S, Kita M, Imanishi J. Paracrine role of keratinocyte growth factor in rabbit corneal epithelial cell growth. Exp Eye Res 1994; 59:385-391
11. Sotozono C, Kinoshita S, Kita M, Imanishi J. Expression of keratinocyte growth factor mRNA in ex vivo and in vitro specimens of human cornea and conjunctiva. Jpn J Ophthalmol 1994; 38:353-359
12. Li Q, Weng J, Mohan RR, et al. Hepatocyte growth factor and hepatocyte growth factor receptor in the lacrimal gland, tears, and cornea. Invest Ophthalmol Vis Sci 1996; 37:727-739
13. Wilson SE, Chen L, Mohan RR, et al. Expression of HGF, KGF, EGF and receptor messenger RNAs following corneal epithelial wounding. Exp Eye Res 1999; 68:377-397
14. Tervo T, Vesaluoma M, Bennett GL, et al. Tear hepatocyte growth factor (HGF) availability increases markedly after excimer laser surface ablation. Exp Eye Res 1997; 64:501-504
15. Jester JV, Barry-Lane PA, Cavanagh HD, Petroll WM. Induction of α-smooth muscle actin expression and myofibroblast transformation in cultured corneal keratocytes. Cornea 1996; 15:505-516
16. Foreman DM, Pancholi S, Jarvis-Evans J, et al. A simple organ culture model for assessing the effects of growth factors on corneal re-epithelialization. Exp Eye Res 1996; 62:555-563
17. De Saint Jean M, Brignole F, Feldmann G, et al. Interferon-γ induces apoptosis and expression of inflammation-related proteins in Chang conjunctival cells. Invest Ophthalmol Vis Sci 1999; 40:2199-2212
18. Musselmann K, Kane BP, Hassell JR. Isolation of a putative keratocyte activating factor from the corneal stroma. Exp Eye Res 2003; 77:273-279
19. Beales MP, Funderburgh JL, Jester JV, Hassell JR. Proteoglycan synthesis by bovine keratocytes and corneal fibroblasts: maintenance of the keratocyte phenotype in culture. Invest Ophthalmol Vis Sci 1999; 40:1658-1663
20. Berryhill BL, Kader R, Kane B, et al. Partial restoration of the keratocyte phenotype to bovine keratocytes made fibroblastic by serum. Invest Ophthalmol Vis Sci 2002; 43:3416-3421
21. Wilson SE, He Y-G, Weng J, et al. Effect of epidermal growth factor, hepatocyte growth factor, and keratinocyte growth factor on proliferation, motility and differentiation of human corneal epithelial cells. Exp Eye Res 1994; 59:665-678
22. Honma Y, Nishida K, Sotozono C, Kinoshita S. Effect of transforming growth factor-β1 and -β2 on in vitro rabbit corneal epithelial cell proliferation promoted by epidermal growth factor, keratinocyte growth factor, or hepatocyte growth factor. Exp Eye Res 1997; 65:391-396
23. Nishimura T, Toda S, Mitsumoto T, et al. Effects of hepatocyte growth factor, transforming growth factor-β1 and epidermal growth factor on bovine corneal epithelial cells under epithelial-keratocyte interaction in reconstruction culture. Exp Eye Res 1998; 66:105-116
24. Sotozono C, Inatomi T, Nakamura M, Kinoshita S. Keratinocyte growth factor accelerates corneal epithelial wound healing in vivo. Invest Ophthalmol Vis Sci 1995; 36:1524-1529
25. Li D-Q, Tseng SCG. Differential regulation of keratinocyte growth factor and hepatocyte growth factor/scatter factor by different cytokines in human corneal and limbal fibroblasts. J Cell Physiol 1997; 172:361-372
26. Weng J, Mohan RR, Li Q, Wilson SE. IL-1 upregulates keratinocyte growth factor and hepatocyte growth factor mRNA and protein production by cultured stromal fibroblast cells: interleukin-1-β expression in the cornea. Cornea 1997; 16:465-471
27. Li D-Q, Tseng SCG. Differential regulation of cytokine and receptor transcript expression in human corneal and limbal fibroblasts by epidermal growth factor, transforming growth factor-α, platelet-derived growth factor B, and interleukin-1α. Invest Ophthalmol Vis Sci 1996; 37:2068-2080
28. Daniels JT, Limb GA, Saarialho-Kere U, et al. Human corneal epithelial cells require MMP-1 for HGF-mediated migration on collagen I. Invest Ophthalmol Vis Sci 2003; 44:1048-1055
29. Mohan R, Chintala SK, Jung JC, et al. Matrix metalloproteinase gelatinase B (MMP-9) coordinates and effects epithelial regeneration. J Biol Chem 2002; 277:2065-2072
30. Ye HQ, Azar DT. Expression of gelatinases A and B, and TIMPs 1 and 2 during corneal wound healing. Invest Ophthalmol Vis Sci 1998; 39:913-921
31. Fini ME, Cook JR, Mohan R. Proteolytic mechanisms in corneal ulceration and repair. Arch Dermatol Res 1998; 290(suppl 1):S12-S23
32. Li D-Q, Lokeshwar BL, Solomon A, et al. Regulation of MMP-9 production by human corneal epithelial cells. Exp Eye Res 2001; 73:449-459
33. Zeigler ME, Chi Y, Schmidt T, Varani J. Role of ERK and JNK pathways in regulating cell motility and matrix metalloproteinase 9 production in growth factor-stimulated human epidermal keratinocytes. J Cell Physiol 1999; 180:271-284
34. Hom YK, Young P, Thomson AA, Cunha GR. Keratinocyte growth factor injected into female mouse neonates stimulates uterine and vaginal epithelial growth. Endocrinology 1998; 139:3772-3779
35. Lovicu FJ, Kao WW-Y, Overbeek PA. Ectopic gland induction by lens-specific expression of keratinocyte growth factor (FGF-7) in transgenic mice. Mech Dev 1999; 88:43-53
36. Jin M, Barron E, He S, et al. Regulation of RPE intercellular junction integrity and function by hepatocyte growth factor. Invest Ophthalmol Vis Sci 2002; 43:2782-2790
37. Moorby CD, Stoker M, Gherardi E. HGF/SF inhibits junctional communication. Exp Cell Res 1995; 219:657-663
38. Kinoshita S, Adachi W, Sotozono C, et al. Characteristics of the human ocular surface epithelium. Prog Retin Eye Res 2001; 20:639-673
39. Gipson IK, Spurr-Michaud S, Tisdale A, et al. Redistribution of the hemidesmosome components α6 β4 integrin and bullous pemphigoid antigens during epithelial wound healing. Exp Cell Res 1993; 207:86-98
40. Bentley E, Abrams GA, Covitz D, et al. Morphology and immunohistochemistry of spontaneous chronic corneal epithelial defects (SCCED) in dogs. Invest Ophthalmol Vis Sci 2001; 42:2262-2269
41. Dursun D, Kim MC, Solomon A, Pflugfelder SC. Treatment of recalcitrant recurrent corneal erosions with inhibitors of matrix metalloproteinase-9, doxycycline and corticosteroids. Am J Ophthalmol 2001; 132:8-13
42. Jester JV, Ho-Chang J. Modulation of cultured corneal keratocyte phenotype by growth factors/cytokines control in vitro contractility and extracellular matrix contraction. Exp Eye Res 2003; 77:581-592
43. Jester JV, Petroll WM, Cavanagh HD. Corneal stromal wound healing in refractive surgery: the role of myofibroblasts. Prog Retin Eye Res 1999; 18:311-356
44. Wilson SE, Liang Q, Kim WJ. Lacrimal gland HGF, KGF, and EGF mRNA levels increase after corneal epithelial wounding. Invest Ophthalmol Vis Sci 1999; 40:2185-2190
45. Wilson SE, Liu JJ, Mohan RR. Stromal-epithelial interactions in the cornea. Prog Retin Eye Res 1999; 18:293-309
46. Imanishi J, Kamiyama K, Iguchi I, et al. Growth factors: importance in wound healing and maintenance of transparency of the cornea. Prog Retin Eye Res 2000; 19:113-129
47. Wilson SE, Kim W-J. Keratocyte apoptosis: implications on corneal wound healing, tissue organization, and disease. Invest Ophthalmol Vis Sci 1998; 39:220-226
48. Daniels JT, Khaw PT. Temporal stimulation of corneal fibroblast wound healing activity by differentiating epithelium in vitro. Invest Ophthalmol Vis Sci 2000; 41:3754-3762
49. Yu Q, Stamenkovic I. Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-β and promotes tumor invasion and angiogenesis. Genes Dev 2000; 14:163-176
© 2005 by Lippincott Williams & Wilkins, Inc.