Enhancement of vinculin synthesis by migrating stratified squamous epithelium.

A 110-115-kD protein is present at levels 27-fold higher in migratory epithelium in the rat cornea than in stationary epithelium. This protein represents 2.7% of the total protein in migratory epithelium 6-h postabrasion wound and 0.1% of the total protein in stationary epithelium. Our findings demonstrate that this 110-115-kD protein is vinculin. In Western blots comparing proteins from migratory and control epithelium, antibody against vinculin cross-reacted with the 110-115-kD protein. Using immunoslot blots, vinculin was determined to be present at maximal levels 6 h postabrasion wound, at levels 22- and 8-fold higher than control at 18 and 48 h, respectively, returning to control levels 72 h postwounding. Vinculin was also localized by indirect immunohistochemistry in migrating corneal epithelium. 3-mm scrape wounds were allowed to heal in vivo for 20 h. In flat mounts of these whole wounded corneas, vinculin was localized as punctate spots in the leading edge of migrating epithelium. In cryostat sections, vinculin was localized as punctate spots along the basal cell membranes of the migrating sheet adjacent to the basement membrane and in patches between cells as well as diffusely throughout the cell. Only very diffuse localization with occasional punctate spots between adjacent superficial cells was present in stationary epithelium. The increased synthesis of vinculin during migration and the localization of vinculin at the leading edge of migratory epithelium suggest that vinculin may be involved in cell-cell and cell-substrate adhesion as the sheet of epithelium migrates to cover a wound.

Role of calcium and calmodulin in hemidesmosome formation in vitro.

Intact epithelial sheets were removed from rabbit corneas using Dispase II, a bacterial neutral protease. The freed sheets were placed on denuded corneal basal laminae and incubated at 35 degrees C for 3, 6, 18, or 24 h. Epithelial-basal lamina preparations were incubated in culture medium that either contained (a) varying concentrations of Ca2+ ions, (b) calmodulin antagonists, (c) exogenous calmodulin following an initial 6-h incubation in the presence of antagonists, or that lacked (d) Mg2+ ions. Tissues were processed for electron microscopy, and micrographs were taken of basal cell membranes. At least four experiments were conducted for each treatment, and for each experiment the total number of hemidesmosomes were counted along the basal membrane-basal lamina surface of eight cells. The number of hemidesmosomes formed was directly proportional to the increasing concentration of Ca2+. The presence of absence of Mg2+ ions did not change the numbers of hemidesmosomes formed. Calmodulin antagonists inhibited hemidesmosome formation, and this inhibition was reversed by the addition of calmodulin. Thus, hemidesmosome formation is Ca2+ dependent and appears to be mediated by a calmodulin-regulated mechanism.

Hemidesmosome formation in vitro.

Intact, viable sheets of adult rabbit corneal epithelium, 9 mm in diameter, were prepared by the Dispase II method (Gipson, I. K., and S. M. Grill, 1982, Invest. Ophthalmol. Vis. Sci. 23:269-273). The sheets, freed of the basal lamina, retained their desmosomes and stratified epithelial characteristics, but lacked hemidesmosomes (HD). Epithelial sheets were placed on fresh segments of corneal stroma with denuded basal laminae and incubated in serum-free media for 1, 3, 6, 18, or 24 h. Tissue was processed for electron microscopy, and the number of HD/micron membrane, the number of HDs with anchoring fibrils directly across the lamina densa from them, and the number of anchoring fibrils not associated with HDs were counted. After 6 h in culture, the number of newly formed HD was 82% of controls (normal rabbit corneas), and by 24 h the number had reached 95% of controls. At all time periods studied, 80-86% of HDs had anchoring fibrils directly across the lamina densa from them. Anchoring fibrils not associated with HDs decreased with culture time. These data indicate that the sites where anchoring fibrils insert into the lamina densa may be nucleation sites for new HD formation. Corneal epithelial sheets placed on two other ocular basal laminae, Descemet's membrane and lens capsule, had not formed HDs after 24 h in culture. These two laminae do not have anchoring fibrils associated with them. Rabbit epithelial sheets placed on the denuded epithelial basal lamina of rat and human corneas formed new HDs. Thus, at least in these mammalian species, HD formation may involve some of the same molecular components.

Basement membrane components in healing rabbit corneal epithelial wounds: immunofluorescence and ultrastructural studies.

The nature of the substrate that supports epithelial migration in vivo is of interest, particularly with respect to mechanisms of wound healing. Immunofluorescence and electron microscopy were used to search for common substrate components in prototype rabbit corneal wounds: epithelial scrape wounds, in which the corneal or conjunctival epithelium migrated over the denuded lamina densa of the corneal basement membrane (CBM), and superficial keratectomy, in which the corneal epithelium migrated over a bare stroma without CBM. The corneal epithelium moved rapidly over the CBM or stroma to cover the defect within 2-3 d, whereas the conjunctival epithelium required 1-2 wk. In all wounds, fibronectin and fibrin/fibrinogen were deposited onto the bare surface within 8 h after wounding and persisted under the migrating epithelium until migration was complete. Bullous pemphigoid antigen (BPA), a normal component of the CBM, was removed with the epithelium upon scrape wounding and reappeared in the CBM after migration was completed. In contrast, the conjunctival epithelium had a continuous subepithelial band of BPA out to the migrating tip. Laminin, also a normal component of the CBM, was not removed in the scrape wounds, indicating that the region of least resistance to shear stress was between the BPA and laminin layers. Laminin was removed by superficial keratectomy and was not detectable under the leading edge of the migrating cells. Laminin and BPA were restored in the CBM by 2-4 wk. Type IV collagen could not be detected in normal CBM, but was conspicuously present in conjunctival basement membrane and in blood vessels. Focal bands of type IV collagen did appear in the newly synthesized CBM 2-4 wk after keratectomy. These results argue that BPA, laminin, and type IV collagen are not essential for the migration of corneal epithelium during wound healing and support the hypothesis that fibronectin and fibrin/fibrinogen are the common, perhaps the essential, components of the provisional matrix that serves as a substrate until the permanent attachment components are regenerated.

Traction forces mediated by alpha6beta4 integrin: implications for basement membrane organization and tumor invasion.

The integrin alpha6beta4, a laminin receptor that stabilizes epithelial cell adhesion to the basement membrane (BM) through its association with cytokeratins, can stimulate the formation and stabilization of actin-rich protrusions in carcinoma cells. An important, unresolved issue, however, is whether this integrin can transmit forces to the substrate generated by the acto-myosin system. Using a traction-force detection assay, we detected forces exerted through alpha6beta4 on either laminin-1 or on an anti-alpha6 antibody, demonstrating that this integrin can transmit forces without the need to engage other integrins. These alpha6beta4-dependent traction forces were organized into a compression machine localized to the base of lamellae. We hypothesized that the compression forces generated by alpha6beta4 result in the remodeling of BMs because this integrin plays a major role in the interaction of epithelial and carcinoma cells with such structures. Indeed, we observed that carcinoma cells are able to remodel a reconstituted BM through alpha6beta4-mediated compression forces by a process that involves the packing of BM material under the cells and the mechanical removal of BM from adjacent areas. The distinct signaling functions of alpha6beta4, which activate phosphoinositide 3-OH kinase and RhoA, also contribute to remodeling. Importantly, we demonstrate remodeling of a native BM by epithelial cells and the involvement of alpha6beta4 in this remodeling. Our findings have important implications for the mechanism of both BM organization and tumor invasion.

Suppression of Toll-like receptor-mediated innate immune responses at the ocular surface by the membrane-associated mucins MUC1 and MUC16.

Membrane-associated mucins (MAMs) expressed on the ocular surface epithelium form a dense glycocalyx that is hypothesized to protect the cornea and conjunctiva from external insult. In this study, the hypothesis that the MAMs MUC1 and MUC16, expressed on the apical surface of the corneal epithelium, suppress Toll-like receptor (TLR)-mediated innate immune responses was tested. Using an in vitro model of corneal epithelial cells that are cultured to express MAMs, we show that reduced expression of either MUC1 or MUC16 correlates with increased message and secreted protein levels of the proinflammatory cytokines interleukin (IL)-6, IL-8, and tumor necrosis factor-α (TNF-α) following exposure of cells to the TLR2 and TLR5 agonists, heat-killed Listeria monocytogenes and flagellin, respectively. As mice express Muc1 (but not Muc16) in the corneal epithelium, a Muc1(-/-) mouse model was used to extend in vitro findings. Indeed, IL-6 and TNF-α message levels were increased in the corneal epithelium of Muc1(-/-) mice, in comparison with wild-type mice, following exposure of enucleated eyes to the TLR2 and TLR5 agonists. Our results suggest that the MAMs MUC1 and MUC16 contribute to the maintenance of immune homeostasis at the ocular surface by limiting TLR-mediated innate immune responses.

The Amount of MUC5B mucin in cervical mucus peaks at midcycle.

The physical character and amount of mucus secreted by the endocervix changes dramatically during the menstrual cycle to facilitate sperm migration at the time of midcycle ovulation. Mucins are highly glycosylated, high-molecular-weight proteins, which are the major structural components of the protective mucus gel covering all wet-surfaced epithelia, including that of the endocervix. We have previously demonstrated that the endocervical epithelium expresses messenger RNA (mRNA) of three of the large gel-forming mucins, designated MUC5AC, MUC5B, and MUC6, with mRNA of MUC5B predominating. Because mucin protein levels may be regulated posttranscriptionally, measurement of MUC5B protein levels with cycle are needed for correlation to mRNA levels. Measurement of specific mucin gene products within mucus secretions has been limited by availability of specific, well-characterized antibodies and by volume requirements of the isolation protocols for mucins, which include CsCl density centrifugation and fraction isolation. To measure MUC5B protein within the cervical mucus through the hormone cycle, we developed a polyclonal antibody specific to the mucin. The antibody, designated no. 799, is to a synthetic peptide mimicking a 19-amino-acid segment of an intercysteine-rich region within the D4 domain in the 3' region of the MUC5B protein. It recognizes native as well as denatured MUC5B on immunoblot, is preadsorbable with its peptide, and binds to apical secretory vesicles of epithelia expressing MUC5B. We used the MUC5B antibody along with a cervical mucin standard cervical mucin isolate in enzyme-linked immunosorbent assay to determine the relative amount of MUC5B mucin in samples of human cervical mucus taken through the menstrual cycle. We demonstrate a peak of MUC5B mucin in human cervical mucus collected at midcycle, compared with mucus from early or late in the cycle. This peak in MUC5B content coincides with the change in mucus character that occurs at midcycle, suggesting that this large mucin species may be important to sperm transit to the uterus.

Mucins of the human endocervix.

The physical character and amount of mucus secreted by the endocervix changes dramatically at midcycle to facilitate the reproductive process. Mucins expressed by the endocervical epithelium contribute to this all-important physiologic event. This review summarizes work from our laboratory demonstrating the mucin gene expression profile of cervical epithelium and mucin levels in cervical mucus through the menstrual cycle. mRNA levels of the gel-forming mucin MUC5B, the major gel-forming mucin expressed by the endocervical epithelium, peak before midcycle and the amount of MUC5B protein per unit total protein in cervical mucus peaks at midcycle. Message levels for MUC4, a major membrane-spanning mucin of the endocervix, peak at midcycle, but protein levels of MUC4 in human cervical mucus have not been measured. Message for each mucin diminishes dramatically as progesterone levels increase in the blood. These data suggest hormonal regulation of the two mucin genes in the endocervix, but there is no information on their regulation at the biosynthetic level via genomic hormone response elements. Perhaps, through its hydrophilicity, the MUC5B mucin holds water in place at the endocervical canal surface at midcycle, keeping the canal patent for sperm motility. A second potential role of the increased mucins at midcycle is to protect the cervix and uterus at the time when increased water is secreted into the cervical canal to facilitate sperm penetrance. Pathogens and other seminal fluid components may be excluded from entering the uterus by mucin trapping. Studies to determine the mechanism of hormonal regulation of mucins as well as the function of individual mucins are needed.

Response of the lysosomal system of the corneal epithelium to tyrosine-induced cell injury.

Rats fed excess tyrosine develop corneal epithelial disease which parallels that found in humans with tyrosine aminotransferase deficiency (tyrosinosis). In the rat, focal lesions develop within the central epithelium and contain crystals (presumably tyrosine) that disrupt cells. We have studied these lesions and localized acid phosphatase and aryl sulfatase at the electron microscope level. Within 60 hr after initiation of diet, cells within the lesions showed an increase in lysosomal enzyme activity. This activity was localized inside some of the crystal ghosts and in numerous lysosomes, including autophagic vacuoles and multivesicular bodies. After 84 hr on diet the entire central corneal epithelium was disrupted and polymorphonuclear leukocytes had infiltrated the area. Crystals were phagocytosed by or developed within polymorphonuclear leukocytes. We hypothesize that crystals form within epithelial cells, disrupt first their lysosomes, and then cells, leading to externalization of lysosomal enzymes. This extracellular lysosomal enzyme release may be responsible for the acute inflammatory response that ensues.

Protein synthesis during corneal epithelial wound healing.

Previous investigations have shown that corneal epithelium, migrating to cover a wound, synthesizes protein and glycoprotein at a faster rate than does normal stratified epithelium. The authors have found that the maximal rate of synthesis, as indicated by the incorporation of leucine and glucosamine, occurs 16 hr after wounding, 6 hr before wound closure. A comparison of total protein and protein synthesized during migration indicates that the increased synthesis is the result of the enhanced synthesis of many of the proteins present in unwounded epithelia. However, one protein band with a molecular weight of 110 K daltons was present to a much greater extent in migrating tissue than in normal epithelium. A time course analysis indicates that this band is apparent during migration and is not present either before wounding or 24 hr after wound closure.

Actin filaments in cells of human trabecular meshwork and Schlemm's canal.

With subfragment-one of myosin used as the histochemical marker, actin filament distribution was mapped in cells of the aqueous outflow pathway. In uveal, corneal-scleral and juxtacanalicular meshwork cells, bundles of actin filaments were present along the basal cytoplasm. Some of these bundles terminated at adhesion plaques. Juxtacanalicular meshwork cells contained, in addition to bundles, randomly oriented actin filaments at the end of cell extensions. Giant vacuoles in the inner wall of Schlemm's canal did not have actin filaments associated with their membranes; there were, however, intermediate filaments present along the membrane. Bundles of actin filaments were also found within the cells of both inner and external walls of Schlemm's canal. Cells of both the meshwork and inner wall of Schlemm's canal had actin filaments extending out into the cytoplasm from cell junctions. In all areas of the meshwork and especially in the area external to Schlemm's canal, cells were observed which had a cytoplasm, very rich in actin filaments. These cells had characteristics common to smooth muscle cells. The actin filaments of cells of the aqueous outflow pathway could have any or all of several functions. Bundles of filaments in the meshwork could contract to lift and separate trabecular sheets and/or, through their association with adhesion plaques, they could stabilize cells on their collagen substrate, thus acting as cytoskeletal struts. Randomly oriented actin filaments, noted particularly in cells of the juxtacanalicular meshwork, may play a role in phagocytosis. Actin filaments found in association with junctions may be important structures for maintaining cell-to-cell contacts. Lack of actin filaments around giant vacuoles in Schlemm's canal indicates that they do not play a role in shuttling aqueous across the endothelium of the canal. The significance of "actin-rich" cells in the meshwork and those external to Schlemm's canal is unknown.

Effects of cytochalasins and colchicine on the ultrastructure of migrating corneal epithelium.

To determine the effects of cytochalasins and colchicine on the ultrastructure of migrating corneal epithelium of the rat, abraded corneas healing in organ culture were cultured for 30 min in the presence of these cytoskeletal-perturbing drugs. Transmission and scanning electron microscopy of these corneas indicate the following. (1) Cytochalasins B (1 microgram/ml of medium) and D (0.1 microgram/ml of medium) caused drastic alteration in structure of cells of the leading edge only. Stratified layers of cells behind the leading edge were unaffected. This observation may indicate that it is the cells of the leading edge which have a rapid turnover of actin filament formation during epithelial sheet movement. (2) Both cytochalasins caused surface blebs or zeiotic processes to form on cells of the leading edge. The shape of the processes caused by the cytochalasins differed, however. Actin filaments accumulated in the cytoplasm under the zeiotic processes. (3) Colchicine had no effect on the ultrastructure of migrating epithelium.

A technique for obtaining sheets of intact rabbit corneal epithelium.

We have developed a technique for obtaining sheets of intact epithelium from rabbit corneas. Nine-millimeter corneal buttons are removed and placed in culture medium containing 1.2 U/ml Dispase II, a bacterial neutral protease. The posterior half of the stroma is removed with forceps. The anterior half is incubated in the Dispase medium for 1 hr at 35 degrees C. The epithelial sheet is then removed by gentle probing with forceps between the epithelium and the stroma. Sheets so obtained have intact basal cells and desmosomes, and the free cell surface of basal cells send out cytoplasmic blebs. The action of the enzyme appears to be at the level of hemidesmosome basement membrane attachment. Polarity of the sheets is easily determined because the cut edges of the sheet curl inward toward the apical surface. These sheets provide excellent viable epithelium for studies of epithelial adhesion and synthesis and pure epithelium for culture.

Effect of lectins on migration of the corneal epithelium.

Rat corneas with 3 mm central corneal abrasions were organ-cultured in the presence of four plant lectins, concanavalin A (ConA), wheat germ agglutinin (WGA), lotus lectin, and soybean agglutinin (SBA), ConA at 20 and 50 micrograms/ml of culture medium slowed epithelial migration. WGA at 20 and 50 micrograms/ml of medium completely inhibited migration. The effects of the ConA at both 20 and 50 micrograms concentrations and of WGA at 20 micrograms were reversible. Lotus lectin and SBA did not effect motility at concentrations up to and including 50 micrograms/ml of medium. The specific sugar to which SBA binds, N-acetyl-galactosamine, did however, slow epithelial migrations. This study demonstrates that blocking glucose, mannose, and glucosamine sugar moities of glycoproteins on the cell surface and/or basement membrane inhibits epithelial cell migration during healing of corneal abrasions.

A technique for obtaining basal corneal epithelial cells.

A technique has been developed for obtaining a cell suspension enriched (89%) in basal corneal epithelial cells. Eleven millimeter corneal buttons were removed and placed in culture medium containing low (10 microM) calcium. The posterior half of the stroma was removed with forceps. Three superficial cuts were made with a Bard-Parker blade on the anterior half of the cornea, which was then incubated for 18 hr at 35 degrees C. Nonadherent cells were brushed off after the incubation and basal cells were harvested after a 1-hr incubation in Dispase II. Cell viability estimated by Erythrocin beta exclusion was 90%. Further evidence of viability was that the cells adhered to their native substrate, the denuded basal lamina. The authors protocol provides a method for analyzing the biochemistry of a known population of epithelial cells and makes available a defined source of cells for culture.

Removal of viable sheets of conjunctival epithelium with dispase II.

A technique (based on Gipson and Grill, Invest Ophthalmol Vis Sci 23:269, 1982) has been developed to obtain pure, viable, intact sheets of rabbit conjunctival epithelium free of the underlying basement membrane. After preparing a full thickness eye wall resection, Dispase, grade II (neutral protease-Bacillus polymyxa), 1.2 Units/ml MEM is injected intrasclerally. The conjunctiva and sclera are pinned in agar and incubated in the dispase with MEM for 1 hour. A 2 X 3 mm sheet of conjunctival epithelium can be dissected bluntly. Light microscopy shows a two- to three-layered epithelium with many goblet cells. Transmission electron microscopy reveals blebbing at the freed basal epithelial cell membrane, intact desmosomes, and intact goblet cells. The conjunctival sheets were cultured on epithelial-scraped corneal stromal carriers in vitro. Numerous goblet cells were present up to 12 hours on 4-mm carriers and 24 hours on 1-mm carriers. With this technique, pure populations of conjunctival epithelium can be isolated for further characterization and tissue culture.

Isolation of corneal epithelium with Dispase II or EDTA. Effects on the basement membrane zone.

An ultrastructural and immunohistochemical comparison was made between the effects of Dispase II and EDTA on the basement membrane zone of corneal epithelium. The comparison was made on intact corneas as well as on freed epithelial sheets and remnant stromas that had been separated using either the enzyme or chelator. At the ultrastructural level, incubation with Dispase II disrupted the lamina densa allowing in situ blebbing of the basal cells. After separation of the epithelium and stroma, the epithelial sheets showed extensive blebs with extracellular matrix trapped between the blebs. The remnant stromas completely lacked lamina densa, but anchoring fibrils remained. Immunofluorescent studies with antibodies against laminin and BM-1 antigen (an antibody to the protein core of heparan sulfate proteoglycan) revealed that laminin antibody binding was present on the freed epithelial sheets and absent from the remnant stromas, whereas BM-1 antibody binding was absent from both the freed sheets and the remnant stromas after incubation with Dispase II. Incubation with EDTA did not disrupt the basal lamina. There was no in situ blebbing of the basal cells, but immediately after epithelial sheet removal, extensive blebbing occurred. Freed sheets lacked attached segments of extracellular matrix as seen with Dispase II treatment. Remnant stromas showed intact basal laminae and anchoring fibrils. Immunofluorescent studies revealed binding of both the laminin and BM-1 antibodies to the remnant stromas, but not to the freed epithelial sheets. Although EDTA removal of epithelial sheets gave a cleaner separation between basal cell membranes and the basal laminae, more basal cells were disrupted than after separation with Dispase II.

C57BL/6 mice lacking Muc1 show no ocular surface phenotype.

PURPOSE: To test the hypothesis that a membrane-spanning mucin, Muc1, facilitates the spread of tear film and protects against bacterial adherence. METHODS: Age-matched, Muc1 null mice and wild-type mice of C57BL/6 genetic background were used for comparison. Eyes were examined by slit lamp biomicroscopy with fluorescein solution to assess epithelial damage and tear film stability. Structure of the ocular surface epithelia was examined by light microscopy, scanning and transmission electron microscopy, and wholemount confocal microscopy. Bacterial adherence assay was performed on in vivo corneas with Pseudomonas aeruginosa containing a plasmid encoding green fluorescent protein, followed by wholemount confocal microscopy. Real-time reverse transcription-polymerase chain reaction was performed using Muc4-specific primers to quantitate Muc4 mRNA expression in ocular surface tissues. RESULTS: No differences were found between Muc1 null and control mice in any parameter tested. Ocular surface epithelia of Muc1 null mice of the C57BL/6 strain had a normal appearance of surface microplicae, a well-developed glycocalyx on the apical cell membrane, and a normal appearance of goblet cell mucin packets. There was no convincing evidence that bacterial adherence on the cornea was increased in Muc1 null mice. Muc4 mRNA expression was not upregulated in Muc1 null mice compared with control. No ocular surface infections were observed in Muc1 null mice of the C57BL/6 strain (n = 204), which were housed in the animal facility over a period of 26 months. CONCLUSIONS: Muc1 null mice of C57BL/6 background appeared normal in all respects tested. These data differ from the reported phenotype in the mice of the C57BL/6 x SVJ129 background, which show development of blepharitis and conjunctivitis.

Eyelid opening induces expression of a glycocalyx glycoprotein of rat ocular surface epithelium.

PURPOSE: We previously characterized a monoclonal antibody against a glycoprotein of rat ocular surface glycocalyx (ROSG). This monoclonal antibody recognizes a carbohydrate epitope on a glycoprotein expressed in apical cells of the differentiated ocular surface epithelium. Our goal was to determine the developmental appearance of the ROSG glycoprotein. METHODS: We localized the antigen immunohistochemically and immunocytochemically in newborn (day 1) rats and in rats 2-5, 7, and 11-15 days after birth. RESULTS: Before eyelid opening (days 12-14), the ROSG antigen was localized in the palpebral conjunctiva near the lid margin. The binding extended to layers of the subapical flattened cells. However, the antigen was not found in the corneal epithelium while the eyelid remained closed. In contrast, at the first day of eyelid opening (days 12-15), the antigen was present contiguously in several cell layers from the eyelid margin along the entire ocular surface, including the corneal epithelium. Thus, the binding pattern seen upon eyelid opening was similar to that of adult rats. The phenomenon of appearance of the glycoprotein upon eyelid opening was further demonstrated in rats with asynchronous eyelid opening. Artificial, premature eyelid opening at days 8-11 also induced the expression of the antigen along the entire ocular surface epithelium, similar to the binding in naturally opened eyes. CONCLUSION: Eyelid opening may induce glycosylation or expression of a glycocalyx component that we hypothesize to be involved in mucin spread.

Differential gene expression in healing rat corneal epithelium.

PURPOSE: The authors used and validated a recently developed method, mRNA differential display, to detect and clone genes that are differentially expressed in healing compared to stationary corneal epithelium. METHODS: RNAs from unwounded and 18-hour postwound corneal epithelia were isolated and subjected to mRNA differential display analysis. The generated cDNAs were used as probes in Northern blot analysis and in situ hybridization to confirm their differential expression and to clone longer or full-length cDNAs from a healing corneal epithelial cDNA library. RESULTS: Changes in the pattern of gene expression in healing epithelium, compared with that in stationary cells, were noted. To date, 15 combinations of 5'- and 3'- primers were used with approximately 1500 mRNA species screened. Differential expression of nine mRNA species were observed. These included four known proteins. They are nonmuscle tropomyosin TM-1, cytokeratin K14, small GTP binding protein rab 11, and amyloid beta-A4 precursor-like protein-2. One is a sequence with homology to type II cytokeratin, and four represent genes with sequences that are unreported. The differential expression of five of these genes was confirmed by Northern blot analysis, in situ hybridization, or both. CONCLUSION: mRNA differential display provides a unique and powerful experimental system to study differential gene expression in wound healing and cell migration. Using this system, differential expression of nine genes was observed. Detection of genes differentially expressed in healing epithelium may prompt studies that will define the specific role of each of the proteins in wound healing.