Modulation of Corneal Epithelial Cell Functions by the Neutrophil-Derived Inflammatory Mediator CAP37

Abstract

purpose. To investigate the effect of CAP37, an inflammatory mediator in neutrophils, on three important events in corneal wound healing: proliferation, migration, and adhesion.

methods. Immortalized human corneal epithelial cells (HCEC) were treated with CAP37, and its effects on migration and proliferation were measured using the modified Boyden chemotaxis chamber and the proliferation assays (CyQUANT; Molecular Probes, Eugene, OR), respectively. Effects on adhesion were determined by measuring upregulation of adhesion molecules belonging to the selectin, integrin, and immunoglobulin superfamily using RT-PCR and flow cytometry.

results. CAP37 promoted proliferation of HCEC in a time- and dose-dependent fashion. CAP37 was maximally chemotactic for HCEC over a range of 1.3 × 10⁻⁸ to 5.2 × 10⁻⁸ M. CAP37 upregulated intercellular adhesion molecule (ICAM)-1, platelet endothelial cell adhesion molecule (PECAM)-1, and integrin molecules α3 (CD49c) and β1 (CD29). Data on migration and ICAM-1 and PECAM-1 upregulation were corroborated using primary human corneal epithelial cells.

conclusions. CAP37 modulated corneal epithelial cell proliferation and migration and upregulated adhesion molecules involved in leukocyte–epithelial and epithelial–extracellular matrix interactions.

Extravasation of leukocytes from the circulation into tissue sites is an integral feature of the host response to injury and inflammation. ¹ By virtue of their ability to engulf and destroy bacteria, eliminate toxins, and secrete numerous soluble mediators, leukocytes are capable of restricting and limiting the spread of infection. Neutrophils (PMNs) are the predominant cell type in the early phases of inflammation, soon followed by a second wave of cells composed mainly of monocytes and lymphocytes. ^{1 2} Irreversible damage to the eye can occur in cases of fulminant inflammation. Clearly the desirable outcome is one in which the immune system can control the infection, resulting in re-epithelialization and healing with minimal damage to vision. In spite of the significance of the immune system in ocular infections, the mediators that regulate leukocyte migration, epithelial–leukocyte interaction, and healing remain elusive. 

A novel inflammatory protein known as cationic antimicrobial protein of M_r 37 kDa (CAP37) plays a critical role in host defense and inflammation. ³ CAP37 was first isolated from the granule fractions of human PMN and viewed as part of the oxygen-independent killing mechanism of the PMN because of its strong antimicrobial activity. ^{4 5} CAP37 has potent chemotactic activity for monocytes ⁶ and can affect endothelial cell activity through its ability to stimulate protein kinase C activity ⁷ and to upregulate adhesion molecules. ⁸ Recent studies undertaken in our laboratory to further define the role of CAP37 in inflammation and host defense involved the use of a rabbit model of Staphylococcus aureus keratitis. ⁹ This well-characterized in vivo model of bacterial keratitis indicated the expected expression of CAP37 in the granules of migrating PMN. However, an unexpected observation was the expression of CAP37 in corneal epithelial cells, stromal keratocytes, ciliary epithelium, related limbus and ciliary vascular endothelium, and bulbar conjunctiva. Particularly striking was the extremely strong staining for CAP37 in corneal epithelium. ⁹  

An important aspect of the resolution of an inflammatory response is the healing process. Corneal epithelial wound healing consists of three sequential events: cell migration, cell proliferation, and cell adhesion. ^{10 11 12} Our study is focused on the in vitro effect of CAP37 on these three critical elements of wound healing. Migration of leukocytes from the vasculature is dependent on the upregulation of adhesion molecules; therefore we measured the effect of CAP37 on upregulation of E-selectin, intercellular adhesion molecule (ICAM)-1, vascular cell adhesion molecule (VCAM)-1, and platelet endothelial cell adhesion molecule (PECAM)-1 on human corneal epithelial cells (HCEC). Since attachment of the newly formed epithelium to the extracellular matrix is essential for completing the healing process, we measured adhesion molecules of the integrin family such as α1, α2, α3, α4, αv, β1, β2, β3, and β4 that are capable of binding to fibronectin, laminin, and other extracellular matrix proteins. ¹³ The interaction of these ligands with their receptors contribute to the formation of attachments and adhesion, thereby aiding the healing process. 

Methods

Cell Culture

Immortalized human corneal epithelial cells (HCECs,) kindly provided by Kaoru Araki-Sasaki (Suita, Japan), ¹⁴ were grown and maintained in keratinocyte-serum free media (Keratinocyte–SFM, Gibco BRL, Grand Island, NY) supplemented with prealiquoted amounts of l-glutamine, epidermal growth factor, and bovine pituitary extract (Gibco BRL) as provided by the manufacturer. ^{9 15} Media changes were made every 2–3 days and cells were subcultured (0.25% trypsin-1mM EDTA at 37°C for 5 minutes, Gibco BRL) when they reached 70% confluence at a split ratio of 1:3. Cells used were always from passage <16. 

Primary human corneal epithelial cells were cultured from normal corneas obtained from autopsy specimens within six hours post mortem (National Disease Research Interchange [NDRI], Philadelphia, PA). All experiments using primary corneal epithelial cells were performed in accordance with the University of Oklahoma Health Sciences Institutional Review Board and HIPAA (Health Insurance Portability and Accountability Act) privacy policies and procedures. Primary corneal epithelial cells for each experiment were isolated ¹⁶ from a total of 2–4 corneas. Cells were divided into tissue culture flasks (Primaria; Falcon, Franklin Lakes, NJ) and cultured as described above in keratinocyte-SFM including supplements and gentamicin (5 μg/mL; Sigma, St. Louis, MO). After 48 hours in culture, the cells were changed to basal keratinocyte-SFM for 16 hours before treatment with CAP37 for RT-PCR studies or for use in chemotaxis assays. Homogeneity of primary corneal epithelial cells was determined ^{16 17} using a mouse anti-cytokeratin epithelial monoclonal antibody (Clone AE1/AE3; Chemicon International, Temecula, CA). 

Recombinant CAP37

Functionally active recombinant CAP37 (rCAP37) was produced using a RSV-PL4 expression system in human 293 cells. ⁸ The recombinant protein was characterized by amino acid sequence, sodium dodecyl sulfate polyacrylamide gel electrophoresis, and western blots, and shown to be identical with native PMN-derived CAP37. All preparations of rCAP37 were <0.1 EU/μg as determined by the limulus amebocyte lysate assay (QCL 1000; BioWhittaker, Walkersville, MD) performed according to the manufacturer’s instructions. The range of rCAP37 used in these studies was selected based on previously published data indicating that the amount of CAP37 present in a PMN to range from 11 fg to 1 pg per PMN. ^{6 18 19} Depending on the number of PMN present it would be expected that the physiologic/normal range of CAP37 could reach levels of 5 μg/mL of blood. ¹⁹ These levels are possibly elevated in inflammatory and disease states when the induced forms of CAP37 in epithelial cells and endothelial cells could contribute to the total levels of CAP37. 

Cell Proliferation

Human corneal epithelial cells were seeded onto 48 well tissue culture plates (7.5 × 10³ cells/well, Falcon) and cultured as described above. Cultures were changed to basal keratinocyte-SFM medium containing no additives or supplements and incubated overnight before the start of the experiment. The cells were then treated with various concentrations of rCAP37 (0–2000 ng/mL) for 48–72 hours. Recombinant human epidermal growth factor (EGF; 50 ng/mL; Becton Dickinson, Bedford, MA) and recombinant human hepatocyte growth factor/scatter factor (HGF/SF 20 ng/mL; Becton Dickinson) were used as positive controls and growth factor-free basal keratinocyte-SFM medium as negative control. The medium was aspirated and new medium with rCAP37 or growth factors was added to the cultures every 24 hours. The CyQUANT Cell Proliferation Assay Kit (Molecular Probes, Eugene, OR) was used to quantify cell proliferation exactly according to the manufacturer’s specifications. Briefly, cells were frozen, thawed, and lysed with the addition of the lysis buffer containing the green fluorescent dye, CyQUANT GR (Molecular Probes), which binds to nucleic acids. The fluorescence levels were read on a fluorescent microplate reader (fmax; Molecular Devices, Sunnyvale, CA) with filters for 485 nm excitation and 538 nm emission. 

Chemotaxis Assay

Human corneal epithelial cells were cultured in basal medium overnight, detached using trypsin-EDTA as described above, and resuspended at a final concentration of 8 × 10⁵ cells/mL. Primary human corneal epithelial cells were similarly detached and resuspended at a final concentration of 2.5 × 10⁵ cells/mL. Chemotaxis assays were performed using the modified Boyden chemotaxis chamber assay described previously ^{6 20} with the following modifications. Briefly, 200 μL of cell suspension was added to the upper chamber. Chemoattractants including rCAP37 (10–5000 ng/mL) and the positive control, recombinant human platelet-derived growth factor-BB (PDGF-BB; 10 ng/mL; Collaborative Biomedical Products, Bedford MA) in 0.1% BSA (IgG-free, endotoxin-low; Sigma) in Geys’ buffer (Gibco BRL) were added to the lower chamber. The chambers were separated by an 8.0 μm pore membrane (13 mm polyvinyl pyrrolidone-free; Whatman, Clifton, NJ). Membranes were precoated with 50 μg/mL collagen type I rat tail (Collaborative Biomedical Products) in 0.02 M acetic acid at room temperature for 1 hour and then air dried. Membranes were rehydrated in basal cell culture medium immediately before commencement of each experiment. The negative control in these experiments was 0.1% BSA in Geys’ buffer. After incubation of the chambers in a humidified atmosphere (37°C, 5% CO₂) for 4 hours, the filters were removed, stained with Diff-Quick (Dade Behring, Dűdingen, Switzerland), and mounted with Permount (Fisher Scientific, Pittsburgh, PA). The filters were viewed under oil immersion (×400 magnification, BH-2; Olympus, Lake Success, NY) and the total numbers of cells migrated through to the underside of the filter were counted in five different fields on each slide. Triplicates were set up for each experimental point. 

To assess whether CAP37 had chemokinetic properties, various concentrations of rCAP37 (0, 10, 100, and 1000 ng/mL) were added to the upper chamber as well as to the lower chamber (0, 10, 100, 500, 1000 ng/mL) and a checkerboard assay performed according to the methodology of Zigmond and Hirsch. ²¹  

To determine the specific interaction of CAP37 with HCEC we used a previously characterized polyvalent, monospecific rabbit antiserum to CAP37 ⁷ to inhibit the chemotactic activity of CAP37. rCAP37 (500 ng/mL, 1.3 × 10^-8 M) was incubated with heat inactivated (56°C for 30 minutes) rabbit anti-CAP37 antiserum at concentrations of 1:10, 1:50, and 1:100, and chemotaxis assays performed as outlined above. Controls included heat-inactivated antiserum alone, rCAP37 alone, PDGF alone, and PDGF plus anti-CAP37 antiserum. 

Flow Cytometry

Flow cytometry was used to assess the upregulation of PECAM-1 (CD31), and the integrin molecules β1 (CD29), β2 (CD18), β3 (CD61), β4 (CD104), α1 (CD49a), α2 (CD49b), α3 (CD49c), α4 (CD 49d), and αv (CD51). Human corneal epithelial cells were cultured as above and treated with rCAP37 (0–2000 ng/mL) for 0 to 72 hours. A corresponding culture was left untreated at each time point. After treatment with rCAP37, cells were detached with 0.25% trypsin in 1 mM EDTA (pH 7.4; Fisher Scientific), washed twice in PBS and fixed with 0.125% paraformaldehyde (J.T. Baker, Phillipsburg, NJ) overnight at 4°C. The cells were washed in PBS and then incubated in 0.5% normal goat serum and 0.5% BSA in PBS for 30 minutes to block nonspecific binding sites. Cells were incubated in the primary antibody (at concentrations described below) at 4°C for 1 hour, followed by the secondary antibody (FITC-goat anti-mouse IgG; Pharmingen, San Diego, CA) at 0.5 μg/10⁶ cells and incubated at 4°C for 30 minutes. The isotype control for these studies was a mouse isotype IgG₁ (Pharmingen). The cells were analyzed by flow cytometry (FACSCalibur; Becton Dickinson, San Jose, CA). Ten thousand cells were analyzed for each sample. 

Antibodies

The primary antibodies and the concentrations used in the flow cytometry experiments were as follows: mouse anti-human PECAM-1 (CD31) monoclonal antibody clone HEC7 (0.5 μg/10⁶ cells; Endogen, Woburn, MA), mouse anti-human very late antigen 1α (VLA-1α, or CD49a) monoclonal antibody clone SR84 (0.5 μg/10⁶ cells; Pharmingen), mouse anti-human VLA-α₂ (CD49b) monoclonal antibody clone AK-7 (0.125 μg/10⁶ cells; Pharmingen), mouse anti-human α3 (CD49c) monoclonal antibody clone C3II.1 (0.125 μg/10⁶ cells; Pharmingen), mouse anti-human VLA-4 (α4) monoclonal antibody clone 2B4 (1 μg/10⁶ cells; R&D Systems, Minneapolis, MN), mouse anti-human α5 (CD49e) monoclonal antibody clone VC5 (0.125 μg/10⁶ cells; Pharmingen), mouse anti-human β₁ (CD29) monoclonal antibody MAR4 (2 μg/10⁶ cells; Pharmingen), mouse anti-human β₂ integrin (CD18) monoclonal antibody clone 6.7 (0.5 μg/10⁶ cells, Pharmingen), mouse anti-human α_vβ₃ (CD51/CD61) monoclonal antibody clone 23C6 (0.5 μg/10⁶ cells; Pharmingen), and mouse anti-human integrin β₄ (CD104) monoclonal antibody clone 450–11A (1.0 μg/10⁶ cells; Pharmingen). A purified mouse IgG₁, κ monoclonal immunoglobulin isotype standard (clone MOPC-31C; Pharmingen) was used as the isotype matched control in the flow cytometry experiments. 

Reverse Transcriptase–Polymerase Chain Reaction (RT-PCR)

Cultured HCEC were treated with rCAP37 (1 μg/mL) for 0–12 hours at 37°C. Primary corneal epithelial cells were treated with rCAP37 (1 μg/mL) for 6 hours at 37°C. Total cellular RNA was isolated from untreated and treated HCEC and primary corneal epithelial cells according to vendor specifications (RNeasy Mini Kit, Qiagen Inc, Valencia, CA). After reverse-transcription of 10 μg of total RNA from HCEC and 3 μg of total RNA from primary corneal epithelial cells by SuperScript II RT (Invitrogen, Carlsbad, CA), the resulting single-stranded cDNA was amplified by PCR (Biometra TGradient, Göttingen, Germany) using specific primers for ICAM-1 (5′-CAT AGA GAC CCC GTT GCC TA-3′,5′-GAA ATT GGC TCC ATG GTG AT-3′), ¹⁷ VCAM-1 (5′-AGT GGT GGC CTC GTG AAT GG-3′,5′-CTG TGT CTC CTG TCT CCG CT-3′), ²² PECAM-1 (5′-TTG CAG CAC AAT GTC CTC TC-3′,5′-AGC ACA GTG GCA ACT ACA CG-3′), ⁸ E-selectin (5′-AGA AGA AGC TTG CCC TAT GC-3′,5′-AGG CTG GAA TAG GAG CAC TCC A-3′) ²³ and β-actin (5′-TAC CTC ATG AAG ATC CTC A- 3′,5′-TTC GTG GAT GCC ACA GGA C-3′) ²⁴ synthesized by the Molecular Biology Resource Facility, University of Oklahoma Health Sciences Center. The thermocycler conditions for ICAM-1 were 95°C for 5 minutes initially, with 40 cycles at 95°C for 1 minute, 60°C for 45 seconds, 72°C for 2 minutes, with a final extension at 72°C for 5 minutes The conditions for VCAM-1 were 95°C for 5 minutes initially, with 40 cycles at 95°C for 1 minute, 58°C for 45 seconds, 72°C for 1 minute, followed by a final extension at 72°C for 7 minutes The conditions for E-selectin were 95°C for 5 minutes initially, with 40 cycles at 94°C for 1 minute, 58°C for 1 minute, 72°C for 1 minute, followed by a final extension at 72°C for 5 minutes. The conditions for PECAM-1 were 95°C for 5 minutes initially, with 40 cycles at 95°C for 45 seconds, 60°C for 1 minute, 72°C for 1 minute, followed by a final extension at 72°C for 5 minutes Amplified DNA fragments were separated by electrophoresis on a 1% agarose gel and visualized by exposure to UV after ethidium bromide (0.5 μg/mL) staining. Expected sizes for ICAM-1, VCAM-1, PECAM-1, E-selectin, and β-actin were 376 bp, 700 bp, 677 bp, 315 bp, and 267 bp, respectively. To assess the integrity of the cDNA, primers for human β-actin were used. Other controls included a positive control consisting of human umbilical vein endothelial cells treated with TNF-α, a water control, and a RNA sample with no reverse transcriptase. 

Statistical Analysis

Data from proliferation and chemotaxis studies are presented as mean ± SE. Statistical analysis was performed employing one-way analysis of variance followed by Dunnett’s multiple comparison procedure using GraphPad Prism 4.0 (GraphPad Software, Inc., San Diego, CA). P < 0.05 was considered significant. 

Results

Proliferation of HCEC in Response to CAP37

CAP37 significantly affects the proliferation of HCEC (Fig. 1) . This response is both dose- and time-dependent. At 48 hours posttreatment with rCAP37, there was a significant increase in proliferation over basal levels observed in culture medium alone. Levels of proliferation obtained with 1000–2000 ng/mL (2.7–5.4 × 10⁻⁸ M) of rCAP37 were comparable to those obtained with the two positive controls, EGF and HGF. HCEC continued to proliferate with time in response to rCAP37 and an approximately two- to threefold increase in cell numbers was obtained at 72 hours posttreatment with 1000 ng/mL and 2000 ng/mL of rCAP37, respectively. The levels obtained with EGF and HGF were similar to those obtained with 1000 ng/mL of rCAP37. 

Migration of HCEC in Response to CAP37

We investigated whether CAP37 was chemotactic for HCEC and primary corneal epithelial cells using the modified Boyden chemotaxis chamber technique. Data described in Figure 2A indicate that rCAP37 is a strong chemoattractant for HCEC. It was maximally chemotactic in the range of 500–1000 ng/mL and was reduced but still significantly active at 2000 ng/mL. The levels of migration in response to rCAP37 were comparable to those obtained with the positive control, PDGF. The dose–response obtained with rCAP37 shows the typical bell-shaped curve indicative of a chemoattractant. ² Figure 2B demonstrates that there is a four- to ninefold increase in migration of primary corneal epithelial cells in response to rCAP37 (1 μg/mL) when compared with the negative buffer control, validating the data obtained with HCEC. This range in migration is not uncommon when using primary cells from different donors. The number of primary corneal epithelial cells used in this assay was almost threefold less than the number of HCEC and was dictated by the numbers of viable cells that could be isolated from each cornea. To allow a direct comparison between numbers of HCEC and primary corneal epithelial cells migrating in response to rCAP37, chemotaxis for this series of experiments was expressed as the numbers of cells migrating as a percent of BSA control, which was arbitrarily designated as 100%. 

An important issue that requires clarification when determining movement of cells in response to a mediator is whether the migration is due to directed movement (chemotaxis) as opposed to merely accelerated random motion (chemokinesis). The checkerboard assay ²¹ has been traditionally used to distinguish chemotaxis from chemokinesis. Our experiments demonstrated that the effect of rCAP37 on HCEC was predominantly chemotactic (Table 1) . We concluded that CAP37 like other chemoattractants displays a certain level of chemokinesis particularly at higher concentrations, ² but that it contributes little to the overall chemotactic effect on corneal epithelial cells. The values in Table 1 are from a representative experiment and are expressed as total numbers of cells migrated rather than percent of control to indicate the absolute numbers of cells migrating to the underside of the filter. 

To demonstrate the specificity of the chemotactic response of HCEC to CAP37, an antibody previously shown to be specific for CAP37 ⁷ was used to inhibit the migration of cells in response to rCAP37. Figure 3 indicates a dose–response inhibition of the chemotactic response, with significant inhibition (P < 0.01 and P < 0.05) obtained with the antibody at 1:10 and 1:50 dilution, respectively. As predicted, the antibody did not have an inhibitory effect on the chemotactic activity of PDGF for HCEC. 

Effect of CAP37 on Adhesion Molecules on HCEC

RT-PCR was performed using primers specific for ICAM-1, VCAM-1, PECAM-1, and E-selectin. Treatment of HCEC with rCAP37 indicates an initial upregulation of ICAM-1 message beginning at 1 hour and lasting through 6 hours (Fig. 4A) . Maximum expression of ICAM-1 message was seen between 3 and 6 hours. The positive control TNF-α strongly upregulated ICAM-1 message expression on HCEC even at 12 hours poststimulation. PECAM-1 was also upregulated by rCAP37. Unlike the upregulation of ICAM-1 message, upregulation of PECAM-1 message was sustained. It was detected at 1 hour after stimulation, and was expressed above basal constitutive levels for as long as 12 hours after treatment with rCAP37. The kinetics of expression of PECAM-1 with the positive control were very similar to that seen with rCAP37. Interestingly, the basal constitutive levels of PECAM-1 message appeared to be influenced by prolonged culture in supplement-free basal media. rCAP37 had no effect on the induction of VCAM-1 and E-selectin message on HCEC (not shown). The data obtained with HCEC was validated using primary corneal epithelial cells. A complete time course was not performed with the primary cells due to the large numbers of corneas required for such an extensive experiment. The 6-hour time point was selected since both PECAM-1 and ICAM-1 mRNA were observed to be upregulated on HCEC at this point. As can be seen, rCAP37 (1 μg/mL) upregulated ICAM-1 and PECAM-1 on primary corneal epithelial cells (Fig. 4B) . 

The upregulation of PECAM-1 mRNA on HCEC was corroborated at the protein level using flow cytometry (Fig. 5) . Kinetic studies using 1 μg/mL of rCAP37 showed significant PECAM-1 protein expression by 6 hours. This level of expression was maintained through 12 hours and waned by 24 hours, corroborating our findings in Figure 4 . The kinetics of this response to rCAP37 closely followed that of TNF-α up to 12 hours. Thereafter the effect of TNFα was more sustained, resulting in PECAM-1 expression until 24 hours. We have previously described ICAM-1 protein expression on HCEC ⁹ in response to rCAP37. It occurs early (between 2 and 6 hours) and reaches maximum levels at approximately 24 hours, which corroborated the RT-PCR data obtained in this study. 

Upregulation of α1, α2, α3, α4, αv and β1, β2, β3, β4 integrins in response to rCAP37 was also assessed by flow cytometry. Of the eight integrin molecules analyzed, only two, CD49c (α3) and CD29 (β1), showed significant upregulation. CD49c (α3) was initially upregulated between 4 and 12 hours, and the level of protein expression was sustained through 24 hours (Fig. 6) . There was high constitutive expression of CD49c, as indicated by strong staining even in the untreated samples. CD49c protein levels on HCEC at 48 and 72 hours (not shown) returned back to constitutive levels. There was no significant difference between fluorescence intensity of untreated cells and TNF-α-treated cells under these conditions. The other integrin molecule to be upregulated by rCAP37 was CD29 (β1, Fig. 7 ). The upregulation is clearly significant by 6 hours, increases to maximum levels between 12 and 48 hours, and although reduced at 72 hours, is still significantly elevated above background constitutive levels. The flow cytometry analysis indicates constitutive expression of CD29, which remained constant throughout all time points in this experiment. TNF-α was used as the positive control in these experiments. 

Discussion

Recently, we made the novel observation that the inflammatory molecule CAP37 could be localized to the eye in response to infection. ⁹ The intrastromal injection of S. aureus resulted, within 10 hours, in the expression of CAP37 in the corneal epithelium, stromal fibroblasts, ciliary epithelium, related limbus, ciliary vascular endothelium, and bulbar conjunctiva. ⁹ The induction of CAP37 occurred well in advance of PMN migration, which in our studies was observed at 15 hours. Corroborative in vitro evidence implicated the cytokines, TNF-α, and interleukin-1β (IL-1β) in the regulation of CAP37 expression. The hypothesis we propose is that CAP37, expressed either in PMN or in extraneutrophilic sites such as the cornea and vascular endothelium of the limbic circulation, could influence ocular host defense by modulating inflammation and wound healing. CAP37 may provide more than just antimicrobial defenses to the infected cornea. ^{4 5} It may additionally regulate corneal inflammation and healing through its ability to attract leukocytes ⁶ and affect corneal epithelial cell functions. The studies in this report focus on the latter aspect and were undertaken to investigate whether CAP37 modulated corneal epithelial cell migration, proliferation, and adhesion; three critical events involved in wound healing. ^{11 12 25}  

Migration of corneal epithelial cells is an important first step in the sequence of healing. ^{11 12 25} CAP37 is a potent chemoattractant for monocytes ⁶ and microglia. ²⁶ Here we show for the first time that CAP37 is capable of acting as a chemoattractant on cells other than those belonging to the mononuclear phagocytic system. The migration of corneal epithelial cells in response to CAP37 was attributable to chemotaxis rather than chemokinesis and occurred at concentrations comparable to those effective on monocytes and microglia. ^{6 26} Changes in expression of growth factors such as EGF, fibroblast growth factor (FGF), transforming growth factor-beta (TGF-β), and PDGF ^{27 28 29} and their receptors ³⁰ have been demonstrated to be of importance in wound healing of the ocular surface, due to their effects on epithelial migration. In vivo studies show that EGF, HGF, and keratinocyte growth factor (KGF) mRNA levels increased in lacrimal glands after corneal epithelial wounding, giving further credence to this premise. ³¹ Studies performed in mouse corneas indicate that expression of HGF receptor mRNA and KGF receptor mRNA were markedly upregulated in the corneal epithelium after wounding. EGF receptor mRNA was not upregulated to the same extent, and the authors ³² suggest that this indicates a less prominent role for EGF in corneal epithelial healing. In addition to growth factors, inflammatory cytokines such as IL-1, IL-6, and TNFα also contribute to corneal epithelial wound healing through their effect on corneal epithelial migration. ²⁷ Our in vitro chemotaxis studies convincingly demonstrate that CAP37 at the concentrations used was as effective as PDGF, a well-characterized mediator of corneal epithelial cell migration. 

In addition to its effects on corneal epithelial cell migration, this study demonstrates for the first time that CAP37 modulates corneal epithelial cell proliferation. Other regulators of epithelial cell proliferation are the well-documented mitogens, KGF, HGF, ²⁷ and EGF. ²⁹ It is noteworthy that CAP37 and HGF/scatter factor, a key regulator of HCEC proliferation, motility, and differentiation bear certain structural similarities. ^{33 34} Whether this has bearing on shared functions and possibly similar signaling mechanisms has yet to be determined. CAP37, like HGF/scatter factor, shows nearly 35% overall sequence similarity with plasminogen. ^{33 34} The remarkable feature about both these molecules is that although they have strong sequence identity with serine esterases, the histidine and serine residues in the catalytic triad have been replaced, making both molecules enzymatically inactive. ^{6 35 36 37 38 39 40 41} Peptide growth factors such as TGF-α and TGF-β may also play key roles in wound healing. ²⁹ TGFβ1 and TGFβ2 act as negative modulators against the cell proliferative effects of EGF, KGF, and HGF. ⁴² Nakamura et al. ⁴³ have shown that inhibition of the EGF receptor affects epithelial cell proliferation and stratification during corneal epithelial wound healing. The authors ⁴³ suggest that it may play a role in maintaining normal corneal epithelial thickness. Our studies show that the mitogenic effect of CAP37 in vitro is comparable to HGF and EGF, two known mediators of corneal epithelial cell proliferation. 

Another important aspect of the healing process is the attachment of migrating epithelial cells to the extracellular matrix. Adhesion is important for cells in providing anchorage, signals for migration, growth, and differentiation. ⁴⁴ The epithelial–extracellular matrix interactions are facilitated mainly through interactions of adhesion molecules of the integrin family and their ligands. ^{11 45 46} Members of the integrin family are heterodimers composed of a larger 120–180 kDa α chain and a smaller 90–110 kDa β chain. ¹³ We showed that many of the assayed integrin molecules were constitutively expressed on corneal epithelial cells but only the α3 and β1 integrin molecules were upregulated on treatment with CAP37. α3β1 is the receptor for fibronectin, ⁴⁷ laminin 5, ^{48 49} and laminin 10, ⁵⁰ all major components of the corneal basement membrane. Laminin 5 is one of the first extracellular matrix proteins deposited by keratinocytes after wounding, ^{51 52} and plays a role in corneal epithelial cell migration. ⁵³ Further studies implicating the importance of laminin 5 and the integrin receptors α6β4 and α3β1 in wound healing are described by Goldfinger et al., ⁵⁴ using an in vitro wound healing model system. Although the authors found that α3β1 was primarily distributed at sites of cell–cell contact, ⁵⁴ exactly how laminin 5 and α3β1 regulate the motility of epithelial cells remains unclear. Evidence from other groups suggests that cells expressing high levels of α2β1 and α3β1 have the proliferative properties expected of epidermal stem cells, whereas cells expressing low levels have low proliferative properties. ⁴⁶ The β1 unit has been observed to be expressed basally in the first cells of the leading edge of the migrating epithelium, whereas the α2 and α3 units are not altered in their distribution after wounding. ⁴⁴  

In addition to upregulation of adhesion molecules important in epithelial-extracellular matrix interactions, CAP37 also regulates the expression of adhesion molecules of the immunoglobulin superfamily important in leukocyte–epithelial interactions. cDNA microarrays performed on healing mouse corneas after transepithelial excimer laser ablations indicate that the ICAM-1 gene is differentially upregulated, implicating this molecule in the healing process. ⁵⁵ It has been known for some time that one of the major roles of ICAM-1 ^{56 57} and PECAM-1 ⁵⁷ within the cornea is the regulation of leukocyte trafficking and accumulation ⁵⁸ in response to infection ^{57 59} and various disease states. ^{60 61 62} In addition to its traditional role in promoting leukocyte recruitment, recent data shows that ICAM-1 is expressed in HCEC in the marginal segment of outgrowing epithelial sheets, suggesting a role for this molecule in epithelial cell motility such as spreading and migration of cells. ⁶³ We speculate that the expression of ICAM-1 and PECAM-1 and the integrin molecules α₃β₁ on the surface of corneal epithelial cells may suggest a more expansive role for CAP37 in the processes of inflammation and healing in the eye through leukocyte–epithelial and epithelial–extracellular matrix interactions. 

Corneal wound healing is an extremely complex sequence of events involving inflammatory cells, growth factors, cytokines, keratocytes, epithelial cells, lacrimal glands, and tear film. ^{12 25} This multifactorial process is driven by a number of important mediators including EGF, HGF, PDGF, TNF-α, and the cytokine, IL-1, which has been referred to as a master regulator of wound healing, ²⁵ since it regulates several critical processes involved in wound healing. The burgeoning numbers of cytokines, chemokines, growth factors, and other mediators prompt the question as to which of these will be significant and relevant in the host. Clearly, until we have studied them all and dissected their roles in inflammation and wound healing it will be difficult to establish a hierarchy of importance. It is possible that one or several factors may be critical contributors to this multifactorial cascade. We believe that there are many salient features about CAP37, such as its biochemical ⁶ and functional attributes, ^{5 6 7 8 9} its cellular localization, ^{6 9 64} and kinetics of expression in response to infection, ⁹ that suggest a strong potential for its role in inflammation and healing in the eye. The mechanism by which CAP37, as an inflammatory mediator, may contribute to the process of healing is still unknown. The signaling mechanisms whereby CAP37 stimulates epithelial cells to undergo migration and proliferation and to upregulate adhesion molecules is clearly the next step in the understanding of its role in wound healing. The existence of a cellular receptor for CAP37 is strongly predicted; nevertheless it has yet to be identified. An understanding of the regulation of this molecule and its receptor has the potential to provide valuable insight into its possible role in corneal wound healing. 

 Supported in part by the Oklahoma Center for the Advancement of Science and Technology (OCAST) and Public Health Service Grant AI28018 from the National Institute of Allergy and Infectious Diseases (HAP), and a core Grant EY12190 Lew R. Wasserman Merit Award from Research to Prevent Blindness and Public Health Service (JC).

 Submitted for publication September 23, 2003; revised April 23 and July 29, 2004; accepted August 18, 2004.

 Disclosure: H.A. Pereira, None; X. Ruan, None; M.L. Gonzalez, None; I. Tsyshevskaya-Hoover, None; J. Chodosh, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Proliferation of human corneal epithelial cells (HCEC) in response to CAP37. Human corneal epithelial cells were treated with 500, 1000, and 2000 ng/mL rCAP37 for 48 and 72 hours, and proliferation determined by the CyQuant Cell Proliferation Assay kit as described in the text. Positive controls used were 50 ng/mL epidermal growth factor (EGF) and 20 ng/mL hepatocyte growth factor (HGF). Data are ± SE of four independent experiments performed in triplicate. **P < 0.01 and ***P < 0.001 compared to untreated control.

Figure 2.

 

Migration of HCEC and primary corneal epithelial cells in response to CAP37. (A) Chemotactic activity of HCEC in response to rCAP37 (10, 100, 500, 1000, 2000, and 5000 ng/mL) as determined by the modified Boyden chemotaxis chamber assay. Chemotaxis is expressed as a percent of the number of cells migrated to the underside of the filter, where 0.1% BSA in Geys buffer (BSA) serves as the negative control and is arbitrarily assigned the value of 100% migration. Data are mean ± SE of three independent experiments performed in triplicate. * P < 0.05 and *** P < 0.001 compared to the untreated buffer (BSA) control. Platelet-derived growth factor (PDGF; 10 ng/mL) served as the positive control. (B) Chemotactic activity of primary corneal epithelial cells in response to 1000 ng/mL rCAP37. Data are mean ± SE from three individual donors (#1, #2, and #3) performed in triplicate. Each data point was performed using two corneas from a single donor.

Table 1.

 

Checkerboard Assay

Concentration of rCAP37 above the Filter (ng/mL)	Number of Cells Migrated
		Concentration of rCAP37 below the Filter (ng/mL)
		0	10	100	500	1000
0	24.06 ± 2.80	28.56 ± 7.64	39.82 ± 6.20	68.89 ± 7.70***	67.34 ± 8.42***
10	ND	27.63 ± 3.03	39.80 ± 6.93	52.40 ± 15.76***	62.73 ± 9.66***
100	ND	ND	36.63 ± 6.48	ND	48.57 ± 15.58***
500	ND	ND	ND	ND	ND
1000	ND	ND	28.08 ± 2.76	ND	38.13 ± 6.07

 Representative data from three independent assays. ND, Not done.

***   P < 0.001 compared to buffer control.

Figure 3.

 

Inhibition of HCEC migration in response to CAP37 using a monospecific antiserum to CAP37. The chemotaxis assay was performed using rCAP37 alone (500 ng/mL, diagonal stripes), 500 ng/mL rCAP37 incubated with various dilutions (1:100, 1:50, 1:10) of anti-CAP37 antiserum (solid bars). Chemotaxis in response to the buffer alone (open bar), antiserum alone at 1:10 dilution (gray bar), platelet-derived growth factor (PDGF; 10 ng/mL) treated with anti-CAP37 antiserum (horizontal stripes), and PDGF alone (vertical stripes) was determined. Data are mean ± SE of three independent experiments performed in triplicate. *P < 0.05 and **P < 0.01 compared to rCAP37 alone.

Figure 4.

 

RT-PCR analysis of HCEC and primary corneal epithelial cells for the expression of adhesion molecule mRNA in response to CAP37. (A) HCEC were incubated with 1 μg/mL rCAP37 (C) for the indicated times to determine upregulation of ICAM-1 and PECAM-1 mRNA. HCEC treated with 10 ng/mL TNF-α (T) at each time point was used as positive controls, and untreated HCEC (U) at each time point was used as the negative controls. cDNA integrity was assessed by β-actin amplification. Other controls included human umbilical vein endothelial cells (+) treated with rCAP37 and a negative water control (−). This is a representative figure from six independent experiments. (B) Primary corneal epithelial cells incubated with 1 μg/mL rCAP37 (C) for 6 hours, compared with untreated (U) cells and ICAM-1 and PECAM-1 mRNA expression detected. Controls as for Figure 4A. Data obtained from cells pooled from four corneas from two donors.

Figure 5.

 

Kinetic expression of PECAM-1 on HCEC in response to CAP37 as measured by flow cytometry. HCEC were treated for the indicated times with 1 μg/mL of rCAP37 (dark solid line), TNF-α (10 ng/mL, dashed line), or left untreated (light solid line). The gray shaded area reflects data obtained with the isotype control. Data are representative from two independent experiments. Mean fluorescence intensity is indicated for each panel for untreated (Mean U) and rCAP37-treated HCEC (Mean C).

Figure 6.

 

Kinetic expression of CD49c (α3) integrin molecule on HCEC in response to CAP37 as measured by flow cytometry. HCEC were treated for the indicated times with 1 μg/mL of rCAP37 (dark solid line), or left untreated (light solid line). The gray shaded area reflects data obtained with the isotype control. Data are representative from three independent experiments. Mean fluorescence intensity is indicated for each panel for untreated (Mean U) and rCAP37-treated HCEC (Mean C).

Figure 7.

 

Kinetic expression of CD29 (β1) integrin molecule on HCEC in response to CAP37 as measured by flow cytometry. HCEC were treated for the indicated times with 1 μg/mL of rCAP37 (dark solid line), TNF-α (10 ng/mL; dashed line), or left untreated (light solid line). The gray shaded area reflects data obtained with the isotype control. Data are representative from five independent experiments. Mean fluorescence intensity is indicated for each panel for untreated (Mean U) and rCAP37-treated HCEC (Mean C).

The authors thank Jim Henthorn, Director, University of Oklahoma Health Sciences Center Flow Cytometry and Confocal Microscopy Laboratory, for assistance in obtaining flow cytometric data; Ken Jackson, Molecular Biology Resource Facility, Warren Medical Research Institute, for primer synthesis; and Philip A. McHale for his help with statistical analysis.