A Comparison of Biological Coatings for the Promotion of Corneal Epithelialization of Synthetic Surface In Vivo

Abstract

purpose. To investigate the effect of a range of biological coatings on corneal epithelialization of a synthetic polymer surface in vivo.

methods. Eight diverse biological factors (collagen I, collagen III, collagen IV, laminin, fibronectin, endothelial extracellular matrix, hyaluronic acid, and chondroitin sulfate) were coated individually onto the surface of polycarbonate membranes with a pore size of 0.1 μm. The coated membranes were implanted on the anterior cornea of adult cats and were clinically assessed for rapidity and extent of and persistence of epithelial overgrowth. The membranes with persistent epithelial attachment were examined histologically by immunohistochemistry and routine light and electron microscopy.

results. Collagen I, collagen IV, and laminin consistently enhanced migration and attachment of corneal epithelial cells in vivo. Multiple-layered epithelium over the collagen I–, collagen IV–, and laminin-coated membranes was demonstrated histologically. The collagen I–coated membranes performed best, in that they showed greater stratification and differentiation of the epithelium. Formation of basement membrane and adhesion complexes over the collagen I–coated membranes was detected by immunohistochemistry and electron microscopy up to 9 weeks after implantation. Membranes coated by fibronectin, endothelial extracellular matrix, hyaluronic acid, and chondroitin sulfate did not support persistent epithelial overgrowth. Compromised biostability of these coatings was mostly likely associated with postsurgical reactions of the host corneal tissue.

conclusions. A biologically modified polymer can support migration and adhesion of corneal epithelial cells in vivo. The collagen I–modified surface exhibited the most promising performance, both clinically and histologically.

Development of a successful synthetic corneal onlay for correction of refractive error entails modification of the onlay to promote corneal epithelial growth over the onlay surface. The ideal onlay should also maintain the permeability of the cornea and should remain stable and optically clear after implantation. ¹ One approach in the development of a synthetic onlay is to coat the surface with materials that mimic the basement membrane of the corneal epithelium to promote epithelial wound healing. ^{2 3 4 5}  

Corneal epithelial wound healing consists of cell migration, proliferation, and adhesion. ^{6 7} Cell migration is a major and critical component because substantial proliferation and permanent adhesion of epithelial cells will occur only when the wound area is completely covered by migrating cells. ^{6 7 8 9} Epithelial migration over the wound area is initiated by recognition between adhesion receptors of epithelial cells and extracellular matrices on the wound surface. ^{10 11 12} These extracellular matrices include fibronectin, collagens I, III, and IV, vitronectin, and laminin. ^{13 14 15 16} In addition, the presence of large quantities of hyaluronic acid in corneal wounds provides a transient matrix that facilitates cell migration during the wound-closure process. ^{16 17}  

Several in vitro studies have shown the potential of synthetic surfaces coated with components of the basement membrane to promote epithelial cell regrowth. Surfaces coated with collagen, fibronectin, and vitronectin have all been shown to promote cell migration in vitro. ^{18 19} Laminin can promote both cell migration and adhesion of epidermal tissue and corneal epithelium (Kurpakus MA. IOVS 1998;39:ARVO Abstract 5183). ²⁰ During the early stages of corneal wounding, proteinases composed of plasmin and collagenases are released from the injured epithelium and stroma, ^{21 22} and these act to remove necrotic tissue, but may also cleave the glycoproteins deposited on the wound surface and affect adhesion of the regenerated epithelial cells. ^{15 23 24} Therefore, it is necessary to test the efficacy of biological coatings for synthetic implants in vivo to ensure that proteolytic activity of the surface coating does not compromise epithelial cell regrowth. 

We have described the porosity and surface topography of synthetic materials that is necessary to promote regrowth of a healthy epithelium after implantation in vivo. ^{1 25} Short-term in vivo studies have shown that thin collagen I coatings can promote epithelial cell regrowth over synthetic surfaces and do not interfere with nutritional permeability of a permeable polycarbonate membrane. ^{5 26} Therefore, synthetic onlays coated with a biological signal appear to be biologically acceptable in vivo if the onlay material is highly permeable. 

Much of the work on epithelial cell regrowth over synthetic surfaces has focused on collagen I–coated surfaces. Little information is available on the effectiveness of other biological components found in corneal tissue to support epithelial overgrowth in vivo. This study compared the effect of a range of surface coatings consisting of major extracellular components that are found in corneal tissue on initial epithelialization of synthetic surfaces in vivo. Factors that affect the ability of the biological coatings to support in vivo epithelialization were also investigated. 

Materials and Methods

Porous polycarbonate membranes (0.1-μm pore size, polyvinylpyrrolidone [PVP] free; Poretics, Medos, Australia) were used as model onlay materials. These membranes are ultrathin and have been shown to be permeable to high molecular weight nutrients after biological modification. ²⁶ The membranes were trephined into 8-mm diameter onlays before surface treatment. Eight biological materials—collagen I, III, and IV; endothelial extracellular matrix (ECM); fibronectin; laminin; hyaluronic acid; and chondroitin sulfate—were used as surface coatings (Table 1) . Both hyaluronic acid and chondroitin sulfate are undetectable in normal cornea, but are present immediately after corneal wounding. ^{27 28 29 30} We wanted to test whether they might facilitate the attachment and migration of corneal epithelial cells at the wound periphery in the early stages of healing that followed surgery. 

The coating procedures have been detailed in previous studies. ^{19 31} Briefly, each of the purified proteins was covalently immobilized on the anterior and posterior surfaces of the onlay after plasma modification with either acetaldehyde plasma polymer (AApp) or heptylamine plasma polymer (HApp). The ECM coating was produced by culturing bovine corneal endothelial cells on one side of the onlay for 3 weeks in a manner similar to that described previously. ³² Briefly, endothelial cells adhering to the onlay were lysed with 0.02 M ammonium hydroxide, leaving behind an intact matrix on the onlay surface. The ECM-coated onlays were stored in sterile phosphate-buffered saline (pH 7.2; PBS) containing 60 μg/mL penicillin and 100 μg/mL streptomycin. All other coated onlays were stored in sterile PBS before implantation. 

Experimental Design

The animal research in this study was approved by the University of New South Wales’ Animal Care and Ethics Committee. All procedures were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Cats were used as experimental animals. The normal life span of the cat is 12 to 15 years. ³³ Fifty-two cats ranging in age from 2 to 4 years were obtained from the Animal Breeding and Holding Unit, The University of New South Wales, and were randomly assigned to eight coating groups and three control groups. The control groups included uncoated implants or AApp- or HApp-treated implants. 

Initially, three cats were assigned to each of the biological coating groups. The final number of cats for each group was determined by the clinical performance of the initial three cats. If complete epithelialization of the onlay was not achieved in all three cats, no further cats were added to the group. If a surface showed complete epithelialization in all three cats, the number of the cats in the group was increased to five. This number was further increased to at least 10 for those surfaces on which epithelial coverage of the implants persisted for at least 2 weeks. This strategy allowed the number of animals used in the entire study to be minimized, in accordance with standard ethical guidelines. 

Surgery

One eye of each cat was chosen at random for implantation. The cat was anesthetized using 10 to 15 mg/kg body weight of ketamine and 1 mg/kg body weight of xylazine intramuscularly to a depth of stage 3-plane 2. The epithelium was removed from a 6-mm diameter area in the central cornea with a Beaver blade. A trephine 4 mm in diameter was used to make a circular keratotomy, 150 to 200 μm in depth, in the central cornea. The stromal lamella within the trephined area was removed from the base of the keratotomy, leaving a keratectomy of 4 mm in diameter. A sharp corneal dissector (model E380; Karl Storz, Tuttlingen, Germany) was used to make a 2 mm wide circular pocket at the base of the keratotomy toward the limbus. The 8-mm diameter implant was tucked into the circular pocket by using two blunt spatulas. The eye was then closed by a temporary tarsorrhaphy until the animal recovered from the anesthesia (3 to 4 hours after implantation). 

After implantation 0.5% framycetin sulfate and 0.5% dexamethasone sodium eye drops were applied three times daily with nightly 0.5% chloramphenicol ointment (Chloromycetin; Monarch Pharmaceuticals, Bristol, TN) for the first week. The eye drops and ointment were reduced to once daily for a further 2 weeks. 

Clinical Evaluation

Clinical evaluation was performed daily for 3 weeks after implantation by an assessor masked to the identity of the surface coatings of the implants. This evaluation included postsurgical irritation of the eyes, physical integrity of the implants, and epithelialization of the implant surface. Slit lamp biomicroscopy with white light and fluorescein staining with blue light were used for the clinical evaluation. Assessment of epithelialization included the rapidity, extent, thickness, and persistence of epithelial coverage, using a scoring scale on which 0 corresponds to no epithelium over the onlay surface, 1 to 25% coverage, 2 to 50% coverage, 3 to 75% coverage, 4 to 100% coverage, and 5 to a multilayered epithelium over 100% of the onlay surface. 

Histology and Ultrastructure

Cats were terminated with an intramuscular injection of pentobarbitone sodium (150 mg/kg body weight). Corneas of animals with implants that supported sustained epithelial growth (a score of at least 4) were examined histologically at different time points. 

Light Microscopy

Light microscopy was used to assess morphology of sustained epithelial growth over the entire implant surface after implantation. Both the implant-receiving and contralateral eyes of the selected cats were examined. The earliest time point for histology was 2 weeks after surgery. The latest time point was at least 6 weeks, by which time epithelial stratification was expected to have occurred. ^{4 5 14 34}  

The cornea was irrigated for approximately 30 seconds with fixative (2.5% glutaraldehyde buffered with 0.1 M sodium cacodylate [pH 7.2]) before the eyeball was removed. After the eyeball was removed, approximately 0.5 to 1 mL of the fixative was injected into the anterior chamber through a limbal incision created by a beaver blade. The whole cornea was removed from the eye by cutting around the limbus with the beaver blade. The cornea was placed into fresh 2.5% fixative and left for 60 hours at 4°C. 

A suture (4-0 thread) was placed in the limbus before trimming to orient the strips after removal. The cornea was placed epithelial side down on a wax surface. Three 1-mm wide strips were cut from the endothelial surface to the epithelial surface with a surgical blade: one from the periphery to the center along the superior–inferior axis, one from the periphery to the center along the nasal–temporal axis, and one from the periphery to the center along the inferior–superior axis. Tissue strips were rinsed in 0.1 M sodium cacodylate buffer for 24 hours at 4°C, dehydrated in a graded series of ethanol (30%, 50%, 70%, and 95%) for 15 minutes each, infiltrated in 1:1 (vol/vol) mixture of 95% ethanol-2-hydroxyethylmethacrylate overnight at 4°C and infiltrated in fresh 100% historesin for 24 hours at 4°C. 

The tissue strips were then placed in a fresh solution of resin and hardener (15 mL resin plus 1 mL hardener) in molds and were covered with a stub and placed on ice for 30 minutes to allow the resin to polymerize. Polymerized blocks were stored in a dessicating atmosphere. 

The blocks were sectioned at a thickness of 3 μm with a resin microtome. Approximately 50 sections for each block were collected and mounted on slides. The slides were alternately stained with 0.1% toluidine blue and Gill’s hematoxylin-eosin. Stained sections were dried, immersed in xylene, mounted, and coverslipped. Sections were viewed and photographed by light microscope. 

Immunohistochemistry

Two cats from a coating group that showed the most consistent epithelialization were examined by immunohistochemistry 9 weeks after implantation. Freshly excised eyeballs were held in cold PBS. Tissue strips were trimmed from the bulbar conjunctiva across the limbus to the central cornea, as described for light microscopy. To preserve the antigenic sites, the strips were unfixed and frozen in optimal cutting temperature (OCT) compound (Tissue Tek; Miles, Elkhart, IN) and sectioned at 6 to 7 μm at −20°C. Sections were probed with primary antibodies against cytokeratin 3 (AE5; ICN Biomedicals, Costa Mesa, CA), laminin (Commonwealth Scientific Industrial Research Organization [CSIRO], North Ryde, Australia), bullous pemphigoid antigen (BPAg, a gift from Westmead Hospital, New South Wales, Australia), and collagen VII and α-6 integrin (both from Chemicon, Temecula, CA). Structural sites detected by these primary antibodies are presented in Table 2 . Antibody responses were labeled with a species-appropriate fluorescein isothiocyanate–conjugated second antibody, and the sections were examined by laser confocal microscopy. 

Electron Microscopy

The cornea was irrigated with fixative before removal of the eyeball. The whole cornea was removed from the eye by cutting around the limbus and was fixed for 48 hours at 4°C, as described for light microscopy. 

Tissue for ultrastructural examination was trimmed and further fixed for 3 hours and washed overnight in buffer. Blocks were postfixed in osmium tetroxide, dehydrated through a graded series of ethanol, infiltrated, and embedded in Epon Araldite, as previously described. ⁵ Sections (80–100 nm thick) were stained using uranyl acetate and Reynolds’ lead citrate and examined by electron microscopy. 

Statistical Evaluation

General factorial analysis of variance and post hoc multiple comparisons of the mean clinical scores over daily time points for each group were used to assess the ability of different coatings to promote epithelialization. The level at which the interaction between groups was significant was set at 0.05. Differences between surface coatings and between groups of coatings were considered significant at P < 0.05. 

Results

Clinical Observations

Epithelialization initially occurred at the periphery of the exposed surface of the implants and migrated toward the center (Fig. 1) . The clinical performance and final number of cats in the different groups are detailed in Table 3 . 

Surface coatings were classified into three statistically separate groups, based on their mean clinical scores over the observation period (P < 0.05). Collagens I and IV and laminin were classified as group I, ECM and fibronectin as group II, and chondroitin sulfate, collagen III, and hyaluronic acid as group III (Fig. 2 , Table 3 ). The mean clinical scores of coatings within each group did not differ from each other (P > 0.05), and each group of coatings was significantly different from the three control groups: group I (P < 0.001), group II (P < 0.01), and group III (P < 0.05). 

Classification of surface coatings into the three groups correlated well with the overall outcome of each type of coating. A greater percentage of the lenticules coated with group I and II surface coatings showed complete epithelialization in comparison to lenticules coated with group III coatings and control lenticules (Table 3) . The time required for full epithelialization varied between individual implants and between coating groups from 2 to 18 days. Epithelial growth over the lenticule surface progressed steadily only with group I coatings (Table 3) . The time required for full epithelialization was not associated with the classification of surface coatings (Table 3) . 

Mild to moderate ocular irritation was observed in 90% (47/52) of the eyes across all groups after surgery. Irritation was evidenced by chemosis and hyperemia of the bulbar conjunctiva, stromal edema, and haze anterior to and around the onlay and mucoid discharge. The remaining 10% (5/52) of eyes showed obvious ocular irritation which manifested as reflex tearing, closure of the palpebral aperture, corneal edema, and salivation. Obvious ocular irritation was only found in eyes implanted with onlays coated with collagen I and IV and hyaluronic acid (one eye each) and collagen III (two eyes). There were no signs of infection in these eyes. Irritation subsided within 1 week after surgery in all groups, except for the obvious ocular irritation in eyes with hyaluronic acid– and collagen III–coated implants. The obvious ocular irritation persisted in these eyes, despite continued treatment with antibiotics and steroids. All implants remained within the pocket throughout the observation period. 

Histology and Ultrastructure

Light Microscopy.

Routine histology was performed on group I surface coatings. Collagen I–coated implants were examined at days 12 (one cat), 21 (four cats), 44 (two cats), and 86 (one cat) after surgery. Collagen IV–coated implants were examined at days 16 (two cats), 28 (two cats), and 44 (one cat). Laminin-coated implants were examined at days 34 (one cat) and 51 (one cat). Epithelial cells had covered the entire implant surface of each cornea at each time of analysis. 

Light Microscopy of Collagen I–Coated Implants.

Day 12.

Epithelium covering the implant surface consisted of four to five cell layers at the peripheral area, and the number of cell layers became less toward the central region where only two cell layers covered the surface (Fig. 3) . At the periphery, the basal cells were cuboidal, becoming more elongated toward the central region. The basal cells had a darkly granular nucleus with a pale cytoplasm, suggesting an active metabolic state. The cells overlying the basal cell layer appeared flat and had a slightly granular nucleus. Overall, the epithelium was thinner, and cell stratification was less regular than in the central region of the cornea in the contralateral eye. 

Days 21 to 44.

Histologic results from days 21 to 44 were similar in all the sections examined. Epithelium over the peripheral area of the implant consisted of unstratified cells and had a thickness equivalent to at least 10 cell layers, which was greater than the epithelial thickness of the control cornea. The number of cell layers became less toward the central region where the thickness of epithelium was between three and five cell layers, varying with different animals of the same group. The basal cells at the periphery were cuboidal and became elongated toward the central region. These cells were lightly stained, compared with those in the contralateral eyes (Fig. 4) . There were two to four layers of squamous cells overlying the basal cell layer. These cells were flat and normally stained, compared with the superficial epithelial cells in the contralateral eyes. 

The anterior stromal flap in the day-21 corneas appeared normal in thickness, but small numbers of degenerated keratocytes accumulated at the leading edge of the flap in most of the sections examined (Fig. 5) . The leading edge of the flap in the day-44 cornea (two cats) appeared thinner than in the day-21 corneas. The stroma-implant interface was quiet in all the eyes (Fig. 4A) . There was a restricted region of epithelial ingrowth behind the peripheral region of the implant in a superior–inferior section in one animal. 

Day 86.

The epithelium over the entire implant surface appeared more regular in thickness and stratification than that in the animals examined in the earlier stages (Fig. 4C) . The epithelium consisted of five to six cell layers on both the peripheral and central regions of the implant surface, which was still slightly thinner than that of the contralateral eye. The basal cells were rectangular and had slightly granular staining in the nuclei. There were two to three layers of squamous and elongated cells overlying the basal cell layer. These cells had a slight to moderate granular staining of their nuclei in comparison with those in the contralateral eye, indicating that they were wing cells. There were two layers of superficial cells that were flat and had darkly granular staining in the nuclei. 

The leading edge of the anterior stromal flap was much thinner than that in the day-44 cornea and was triangular in cross section, indicating a progressive stromal melting from the leading edge toward the limbus. Stratification of the epithelium over the leading edge was also low. The stroma–implant interface adjacent to the stromal flap had an increased density of keratocytes. 

Light Microscopy of Collagen IV– and Laminin-Coated Implants.

Epithelial cell morphology over the collagen IV– and laminin-coated implants was similar to that of the collagen I–coated implants. Both cell stratification and differentiation of the epithelium in the collagen IV–coated implants were poorer than the collagen I–coated implants at a similar time point. However, epithelial morphology in the laminin-coated implants was very similar to that of collagen I–coated implants at a similar time point. Progressive thinning of the anterior stromal flap was observed from days 28 to 32 in both collagen IV– and laminin-coated implants. 

All the eyes examined had an intact implant that was positioned in the stromal circular pocket. The stroma posterior to the implant and the endothelium also appeared normal. 

Immunohistochemistry and Electron Microscopy.

The collagen I–coated implants were examined with immunohistochemistry and electron microscopy, because they demonstrated the most consistent epithelialization during clinical observation. 

Two animals (day 67) were used for immunohistochemistry to check the deposition of specific components of the basement membrane and adhesion complexes. The anterior surface of the implant at day 67 was covered by four to five layers of epithelial cells in both cases. Tissue apposition at the epithelium–implant interface was tight. Components of the epithelial basement membrane and adhesion complexes had been deposited at the epithelium–implant interface as evidenced by the deposition of laminin, bullous pemphigoid antigen, and collagen VII. The presence of α6-integrin indicated that the epithelial cells had recognized the onlay surface. Deposition of bullous pemphigoid antigen was interrupted and linear, whereas the deposition of collagen VII was regular and punctate, and α6-integrin deposition was interrupted and punctate (Fig. 6) . 

Two animals (day 44, one cat, and day 86, one cat) were used for electron microscopy to assess the assembly of basement membrane and adhesion complexes at the interface of the epithelium and implant surface. At day 44, there were four to five layers of epithelial cells covering the implant surface, and this tissue showed good integrity, with desmosomes evident between adjacent epithelial cells. The basal and suprabasal cells of this epithelium were flat morphologically. Electron-dense hemidesmosomal plaques were visible along the basal cell membranes, but were irregular in distribution. A band of ECM material had accumulated between the basal aspect of the epithelial tissue and the anterior surface of the implant. By day 86, the anterior surface of the implant was covered by an epithelium composed of four to six cell layers, with many desmosomes between the constituent cells. This tissue was tightly apposed to the implant surface, and a band of ECM material was also evident between the epithelium and the implant surface (similar to that seen at day 44) that was consistent with the deposition of basement membrane proteins (such as laminin and collagen VII) witnessed using immunohistochemistry. It is likely that this would have assembled to form a recognizable basement membrane in time. Hemidesmosomal plaque could be seen in a regular pattern along the basal cell membranes of the basal layer of epithelial cells that were directly apposed to the implant surface (Fig. 6B) . 

Discussion

Clinical Performance

Epithelial migration over the implant surface for all biologically coated surfaces occurred in a centripetal manner that was similar to normal corneal wound healing. ^{7 8 9} The rapidity, extent, and persistence of epithelialization varied with the different surface coatings, and the performances of the biologically coated implants were classified into three distinct groups: group I (collagens I and IV and laminin), group II (ECM and fibronectin), and group III (chondroitin sulfate, collagen III, and hyaluronic acid). In contrast, the AApp- and HApp-coated implants did not support epithelialization of the implant surface, indicating that plasma modification does not interfere with clinical assessment of the biological coatings. 

Although complete epithelialization occurred in approximately 70% to 80% of the group I– and II–coated implants, the overall performance of group I–coated implants was significantly better than in group II. Assessment of the rapidity, extent, and thickness of epithelialization was necessary to predict the overall performance of the coatings. ECM-coated implants achieved an earlier wound closure than all the other coatings; however, in overall performance, it was classified in group II. This suggests that quick wound closure does not necessarily ensure persistent epithelial coverage. This concurs with Madigan et al., ³⁵ who showed that the strength of epithelial adhesion does not necessarily occur in parallel with the wound-healing rate in vivo. Even though the laminin-coated implants appeared to have the most persistent epithelial growth, the rate of wound closure was slower for this coating in comparison with most of the other coating groups. These results coincide with previous reports ^{20 36} on epidermal wound healing where laminin was shown to play a more important role in cell adhesion than in cell migration. 

Although epithelial detachment occurred in both the collagen I– and collagen IV–coated implants, it manifested as small epithelial defects in the central region of the implant, which quickly recovered. Epithelial detachment after initial epithelialization also occurs during corneal wound healing, ^{14 34} and the detached areas vary from focal to extensive epithelial defects. ^{12 37} This is because adhesion structures are not well established in the early stage of epithelial wound healing. ^{7 15 38} It is most likely that the success of collagen I as a coating is because the migrating epithelium recognizes this signal and then responds by synthesizing and secreting or depositing the proteins it needs to reform a basement membrane and adhesion complexes. The adhesion complexes ensure the adhesion of the newly grown epithelium to the underlying surface. 

The inability of group II–coated implants to achieve persistent epithelial overgrowth during the clinical observation period may have been related to the origin of the coating material. The technique used for ECM synthesis may have caused production of cellular enzymes, which acted to denature the surface coating before in vivo application of the coated implants. ²² Fibronectin can enhance migration and adhesion of corneal epithelial cells in vitro. ^{39 40 41 42 43} However, the clinical ability of fibronectin to enhance corneal wound healing is controversial. ^{44 45 46} The fibronectin used in this study was derived from plasma and is less effective than cell-secreted fibronectin in promotion of corneal wound healing. ^{47 48} This may be one of the factors associated with the poor overall performance observed in this study. Furthermore, a mild proteolytic treatment on the corneal surface cleaves fibronectin from the cell surface, ^{49 50} suggesting that fibronectin is very susceptible to proteinases. It is assumed that in the present study the fibronectin coating may have had a weaker resistance to proteolytic activity than the other coatings, resulting in a reduction in its biological function shortly after implantation. ^{21 22}  

The group III–coated implants showed persistent postsurgical irritation and inconsistent epithelialization of the implant surface. The persistent ocular irritation in both the hyaluronic acid and collagen III–coated implants may have compromised the biological functions of the surface coatings and therefore affected the overall performance. Human lenticules made of collagens I, II, III, and IV can activate two major proteolytic enzymes (collagenases and neutral proteinases) in corneal tissue. ^{23 24} In this study, the eyes in both the collagen I and collagen IV groups were generally normal after a short period of postoperative ocular irritation that lasted for 2 to 5 days. The collagen III used in this study may be more effective than the other types of collagens in activation of the enzymatic activity, resulting in a more profound and long-lasting ocular irritation after implantation. 

Hyaluronic acid is the major proteoglycans located in the interfibrillar space of collagens in the cornea. However, the hyaluronic acid produced in the wounded area is biochemically distinct from that in the normal cornea. ⁵¹ This abnormal hyaluronic acid may disrupt the interfibrillar space and affect the nutritional flow across the cornea. This may explain why the severe postsurgical irritation with the hyaluronic acid–coated implants could not be subsidized with both antibiotics and steroids. In vivo studies have shown that topical administration of hyaluronic acid enhances epithelial wound healing. ^{30 52 53 54} This function is enhanced when hyaluronic acid binds to fibronectin at the wound surface. ^{52 54} The chondroitin sulfate–coated implants in the present study failed to support a consistent epithelialization. Chondroitin sulfate can inhibit attachment of human bone cells to fragments of fibronectin but does not affect such an attachment to intact fibronectin in vitro. ⁵⁵ Therefore, it appeared that the absence of fibronectin on the implant surface was another factor affecting the biological functions of both the hyaluronic acid and chondroitin sulfate coatings. 

Morphology and Biological Components

Histologic evaluation of the group I–coated implants indicated that migration of epithelial cells from the area peripheral to the implant was a major component of initial epithelialization at the implant surface. Epithelium overlaying collagen I–coated implant underwent continuous cell migration and proliferation from the peripheral cornea toward the implant surface and an upward differentiation from the basal cell layers. ^{8 56}  

Epithelial tissue apposition with the expression of integrin to the collagen I–coated implant surface was evidenced by immunohistochemistry 10 weeks after the implantation. This suggests that the host epithelial cells recognized the biologically modified surface. ^{10 11 12} Furthermore, ultrastructural examination revealed the assembly of sections of basement membrane and the hemidesmosomal plaque of the adhesion complexes in this stage. Therefore, corneal epithelialization over the implant surface seemed to occur in a similar manner to re-epithelialization of a corneal wound surface. ^{7 14 34}  

A progressive thinning of the anterior stromal flap was histologically detected. This accompanied a decreased epithelial thickness from the peripheral to central regions of the implant. These two changes remodeled the corneal contour and thus decreased discontinuity of the corneal surface. ⁵⁷ In some animals, small amounts of neutrophils and activated keratocytes were observed at the stroma–implant interface 2 to 3 weeks after implantation. This was probably associated with the wound-healing process ^{15 58 59} and the chemotaxis initiated by growth factors during this process. ⁶⁰  

Conclusions

In summary, in this study, collagen I was the most appropriate coating used to enhance corneal epithelialization of a synthetic onlay, in a feline model. 

 Supported by the Australian Federal Government through the Cooperative Research Centres Program.

 Submitted for publication June 10, 2002; revised January 27, 2003; accepted February 11, 2003.

 Disclosure: D.F. Sweeney, None; R.Z. Xie, None; M.D.M. Evans, None; A. Vannas, None; S.D. Tout, None; H.J. Griesser, None; G. Johnson, None; J.G. Steele, 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.

 

Table 1.

 

Biological Materials and Surface Coating Methods

Biological Material	Source	Surface Coating Method
Collagen I	Vitrogen 100, 95% bovine collagen I (Collagen Corp. Palo Alto, CA, USA)	Immobilization after AApp modification
Collagen III	Human placenta, (no. C4407; Sigma) (cat#: C4407)	Immobilization after AApp modification
Collagen IV	human placenta (no. C7521; Sigma)	Immobilization after AApp modification
ECM	Derived from cultured bovine corneal endothelial cells	Culture of cells followed by cell lysis
Fibronectin	bovine plasma (no. F4759 Sigma)	Immobilization after AApp modification
Laminin	mouse EHS tumor (no. 40232; Collaborative Biomedical Products, Walthem, MA (cat#40232)	Immobilization after AApp modification
Chondroitin sulfate	Mixture of types A and C (70%:30%) bovine trachea (no. C8529) Sigma	Immobilization after HApp modification
Hyaluronic acid	Rooster combs, (CIBA Geigy, Basel, Switzerland)	Immobilization after HApp modification

Table 2.

 

Primary Antibodies Used to Detect Corneal Structural Sites

Antibody	Structural Sites
Cytokeratin 3 (AES)	Positive identification of corneal epithelial tissue
Laminin	Basement membrane component
Bullous pemphigoid	Adhesion complex component
Collagen VII	Anchoring fibril component
α-6 Integrin	α6β4 Integrin receptors

Figure 1.

 

Clinical appearance of collagen I–coated implants under a slit lamp biomicroscope with white light. The implant was tucked within the circular stromal pocket, with a smooth surface. Epithelialization of the implant surface progressed with time after surgery (arrows).

Table 3.

 

Epithelialization over the Exposed Surface of the Lenticules

Surface Treatment	Lenticules (n)	% of Eyes Achieving Full Epithelial Cover (%)	Time until Full Epithelial Cover (d)	Eyes Achieving Full-Thickness Epithelium (%)
Biological Coating Group I
Collagen I	11	82	2–10	45
Collagen IV	10	80	3–14	50
Laminin	3	67	12–15	33
Biological Coating Group II
ECM	5	80	2–4	0
Fibronectin	5	80	4–13	0
Biological Coating Group III
Chondroitin sulfate	3	33	4	0
Collagen III	3	0	NA	NA
Hyaluronic acid	3	33	18	0
Control Group
Uncoated	3	33	7	0
AApp	3	0	NA	NA
HApp	3	0	NA	NA

Figure 2.

 

The overall performance of surface epithelialization for each coating treatment group was assessed by plotting the mean clinical scores daily throughout the observation period and using the results for statistical comparison of the persistence of epithelial overgrowth between different coating groups.

Figure 3.

 

The epithelium over the collagen I–coated implant on day 12 was four to five cell layers thick at the peripheral region and thinned to two cell layers toward the central region.

Figure 4.

 

The epithelialization over the collagen I–coated implant on days 21 (A, B) and 86 (C, D). The epithelial differentiation over the implant surface in the day-86 cornea was more obvious than in the day-21 cornea.

Figure 5.

 

Degenerated keratocytes were located at the leading edge of the anterior stromal flap overlaying the collagen I–coated implant on day 21. The epithelium at the area between the periphery of implant surface and the stromal flap thickened and was unstratified. This suggests commencement of remodeling of the stroma.

Figure 6.

 

(A) Immunohistochemistry showing the polarized expression of bullous pemphigoid antigen along the epithelial onlay interface 67 days after implantation, indicating formation of the adhesion complex component. (B) Electron micrograph of the epithelial onlay interface at 86 days after implantation. HD, hemidesmosomal plaques; BCM, basal cell membrane; ECM, extracellular matrix.