Vitamin A deficiency alters the expression of mucin genes by the rat ocular surface epithelium.

PURPOSE: To study effects of depletion of retinoic acid on expression of the mucins ASGP (rMuc4), rMuc5AC, and rMuc1, by the corneal and conjunctival epithelia of the rat. METHODS: Nineteen-day-old Sprague-Dawley male rats were fed a casein-based vitamin A- deficient diet or casein-based diet with vitamin A as control. Rats from both groups were killed at 1, 3, 5, 13, 15, 18, and 20 weeks after initiation of feeding. Expression of the three mucin genes by the ocular surface epithelium was assayed by reverse transcription-polymerase chain reaction (RT-PCR) and in situ hybridization. RESULTS: In vitamin A-deficient rats, ASGP mRNA was not detected by RT-PCR after 15 weeks of feeding. rMuc5AC mRNA was detected by RT-PCR at 15 weeks, but by 18 and 20 weeks was no longer detectable. By in situ hybridization, ASGP mRNA was localized in the entire ocular surface epithelium after 1 week of feeding, was diminished but detectable above background by 13 weeks, and was not detectable at 20 weeks. rMuc5AC mRNA was detected in the goblet cells of vitamin A- deficient rats by in situ hybridization at 13 weeks, but was lost by 20 weeks, as were identifiable goblet cells. rMuc1 mRNA were detected by RT-PCR through all time points of 1 to 20 weeks in both vitamin A-deficient and control rats, indicating no significant change in rMuc1 mRNA expression with vitamin A deficiency. CONCLUSIONS: Both the membrane-spanning mucin ASGP (rMuc4) and the secretory mucin rMuc5AC are directly or indirectly regulated by vitamin A in the ocular surface epithelium, whereas the membrane-spanning mucin rMuc1 is not.

Mucin genes expressed by human female reproductive tract epithelia.

Recent characterizations of mucins at the molecular level indicate that at least eight mucin genes are expressed by epithelia of mucosal surfaces. The purpose of this study was to determine whether these cloned mucins, designated MUC1, MUC2, MUC3, MUC4, MUC5AC, MUC5B, MUC6, and MUC7, are expressed by epithelia of the female reproductive tract. Northern blot analysis, in situ hybridization, and immunohistochemistry were performed using RNA and tissue from surgically removed human reproductive tract specimens including endocervix, ectocervix, vagina, endometrium, and fallopian tube. Complementary DNA to the tandem repeat regions of MUCs 1, 2, 3, 5AC, 5B, and 6; oligonucleotides to the tandem repeat regions of MUCs 4, 6, and 7; and antibodies that recognize unique mucin tandem repeats were used. The data demonstrate that the endocervical epithelium expresses five mucin genes: MUCs 1, 4, 5AC, 5B, and 6. The ectocervical and vaginal epithelia express MUCs 1 and 4, although vaginal expression of MUC4 appears patchy. Endometrial epithelium expresses MUC1 and low amounts of MUC6. MUC6 immunoreactivity was detected only is scattered endometrial glands located in the basalis region in specimens from pre- and postmenopausal women. The only mucin detected in the fallopian tube was MUC1. These data indicate that the endocervical epithelium expresses multiple mucin genes and that the stratified epithelia of the ectocervix and vagina also produce mucins that may function in reproductive processes and protection of the reproductive tract tissues.

Stratified squamous epithelia produce mucin-like glycoproteins.

The stratified squamous epithelia of the ocular surface, larynx, and vagina are mucus-coated epithelia, apices of which are subject to abrasive pressure from epithelia-epithelia interactions from eyelid, vocal cords, or vaginal folds, respectively. Mucus coats on these epithelia have generally been considered to be derived from the specialized mucin-producing cells embedded either in the epithelia or in adjacent tissues. Here we report the isolation, partial characterization, and cellular localization of a mucin-like glycoprotein produced by these stratified epithelia. In all three epithelia, the mucin-like molecule is present on cytoplasmic vesicles in subapical cells. As cells differentiate to their apical-most position adjacent to their mucus coat, the mucin-like molecule moves to the cell membrane where it is particularly prominent on microplicae folds. Lectin affinity chromatography was used to isolate the molecule from rat vaginal and corneal epithelium. Isolated material was approximately 60% carbohydrate and 40% protein. The major monosaccharide was N-acetylgalactosamine with lesser amounts of N-acetylglucosamine, galactose, mannose, xylose and fucose. Amino acid analysis demonstrated the predominant amino acids to be glycine, serine, threonine and proline. These data plus PAS and Alcian blue binding to the isolate indicate a mucin-like glycoprotein.

Human corneal and conjunctival epithelia produce a mucin-like glycoprotein for the apical surface.

PURPOSE: The authors have determined that the corneal and conjunctival epithelia of the rat produce a mucin-like glycoprotein at the apical surface of the epithelium. The purpose of this study was to determine if human ocular surface epithelium produces similar glycoproteins. METHODS: Because our initial attempts at production of monoclonal antibodies yielded blood type A-specific antibodies, corneal epithelia from blood type O donor eyes were used for the production of monoclonal antibodies. Screening of hybridoma products was accomplished by immunofluorescence microscopy of cryostat sections of blood type O donor eyes. The monoclonal antibody produced was used for immunofluorescence microscopy and immunoelectron microscopy to determine tissue and cellular distribution, respectively. Immunoblot analysis of SDS-PAGE-separated proteins from corneal epithelial tissue and from cultured human corneal epithelium was used to determine molecular weight range and epitope binding after periodate oxidation, N-glycanase, and O-glycanase treatment. RESULTS: A monoclonal antibody, designated H185, that binds to apical cell layers of human corneal, conjunctival, laryngeal, and vaginal epithelium was produced. The antibody binds to apical membranes of apical cells, particularly at the tips of microplicae. In subapical cells, the antibody binds to small cytoplasmic vesicles. Cultured human corneal epithelium produces H185 antigen. By immunoblot analysis, H185 binds a high molecular weight protein, > 205 kD, from corneal epithelium and cultured corneal epithelium. The protein band visualized by immunoblot analysis cannot be stained by Coomassie or silver stain on SDS-PAGE, but it does stain with Alcian blue followed by silver stain, indicating that the protein is highly glycosylated. H185 binding to blotted proteins is destroyed by sodium periodate treatment and O-glycanase incubation, suggesting that the epitope of H185 is an O-linked carbohydrate. CONCLUSION: Human corneal and conjunctival epithelia produce a molecule, similar in size, cellular localization, and distribution to the mucin-like glycoprotein of rat ocular surface epithelium. These data suggest that the entire ocular surface epithelium produces mucins for the tear film.

Altered epithelial-basement membrane interactions in diabetic corneas.

Morphologic alterations of the extracellular matrix in diabetes include thickening of the basement membrane and, as evidenced in the corneal epithelium, a decrease in the area of the basal cell membrane occupied by hemidesmosomes. To determine if the abnormal basement membrane in diabetes has altered epithelial recognition of sites for hemidesmosome formation and if diabetic epithelium can recognize sites for hemidesmosome formation on normal basement membrane, we recombined epithelial sheets and corneal stromas with denuded basement membranes in vitro using diabetic and control corneas. In recombinations of diabetic epithelium on normal stromas, the area of the basal cell membrane occupied by hemidesmosomes was 10.7%, significantly less than control recombinations (19.5%). Similarly, in recombinations of normal epithelium on diabetic stromas, the value (8.6%) was significantly decreased from control. These data may indicate altered cell-extracellular matrix interactions in diabetes.

Characteristics of a glycoprotein in the ocular surface glycocalyx.

A monoclonal antibody has been produced that binds to the apical squames (flattened cells) of the rat ocular surface epithelium and to the goblet cells of the conjunctiva. Immunoelectron microscopic localization of the antigen indicates that in apical cells it is present along the apical-microplical membrane in the region of the glycocalyx. In subapical squames, the antigen is in cytoplasmic vesicles. In some goblet cells, the antigen is in the Golgi network, and in others, it is located primarily in the membrane of the mucous granules. SDS-PAGE and immunoblot analysis demonstrate that the molecular weight of the antigen is greater than 205 kD, and the electrophoretic band stains with Alcian blue followed by silver stain. Periodate oxidation of immunoblots and cryostat sections removes antibody binding. Neuraminidase treatment of cryostat sections does not remove antibody binding, whereas N-glycanase does. Taken together, these data indicate that the antigen recognized by the monoclonal antibody is a carbohydrate epitope on a high-molecular-weight, highly glycosylated glycoprotein in the glycocalyx of the ocular surface epithelium and goblet cell mucin granule membrane. The antigen appears to be stored within cytoplasmic vesicles and reaches the glycocalyx when cells differentiate to the apical-most position where the glycocalyx interfaces with the mucin layer of the tear film.

Reassembly of the corneal epithelial adhesion structures following human epikeratoplasty.

Ten epikeratoplasty lenticules removed after surgery were obtained for immunohistochemical and electron microscopic analysis to determine the pattern of re-formation of corneal epithelial adhesion structures. Antibodies to laminin and type VII collagen were used to determine the presence of basement membrane and anchoring fibrils, respectively. Electron micrographs were used to determine the percentage of basal cell membrane occupied by hemidesmosomes, the area of basal lamina per 100 microns of basal cell membrane, and the average maximum depth of penetration of anchoring fibrils into the stoma. Nine normal corneas served as controls. Compared with normal corneas (24.5% of basal cell membrane occupied by hemidesmosomes; 32.0 microns 2 basal lamina per 100 microns of basal cell membrane), lenticules removed for optical reasons had near-normal hemidemosomes as early as 10 weeks following surgery (mean, 20.3%). The area of basement membrane was reduced (16 microns 2 basal lamina per 100 microns of basement cell membrane). During the course of 2 to 3 years, irregularities and duplications of the basement membrane were noted. Compared with normal corneas, the two lenticules removed for persistent defects had a marked reduction of hemidesmosomes and basement membrane present under epithelium at 3 and 4 weeks (9.6% of basal cell membrane occupied by hemidesmosomes and 13.6 microns 2 basal lamina per 100 microns of basal cell membrane, and 5.4% of basal cell membrane occupied by hemidesmosomes and 7.2 microns 2 basal lamina per 100 microns of basal cell membrane, respectively.

Localization of smooth muscle myosin-containing cells in the aqueous outflow pathway.

Cells containing smooth muscle myosin were localized in the human aqueous outflow pathway by immunohistochemical techniques. In the majority of eyes, immunoreactive cells were observed adjacent to the collector channels and slightly distal to the outer wall of Schlemm's canal. In a few eyes, smooth muscle myosin was localized to cells in the juxtacanalicular tissue and the trabecular meshwork. The immunoreactive cells from these regions may be true smooth muscle cells or pericytes, which can contain smooth muscle myosin. No obvious differences were observed in the pattern of distribution of smooth muscle myosin-containing cells in a comparison of age groups. In the majority of eyes, we observed an apparent direct insertion of the longitudinal portion of the ciliary muscle in the corneoscleral meshwork far internal to the scleral spur.

Decreased penetration of anchoring fibrils into the diabetic stroma. A morphometric analysis.

Morphometric measurements of epithelial adhesion structures were performed on electron micrographs of corneas of diabetic patients (ages 35 to 77 years) and controls of similar age (ages 21 to 79 years). Penetration of the anchoring fibrils from the deepest layer of the basal lamina into the stroma was significantly decreased in diabetics (0.41 +/- 0.02 microns) compared with controls of similar ages (0.65 +/- 0.03 microns). Significant thickening of the basal lamina in diabetics (50.1 +/- 7.6 microns 2/100 microns basal cell membrane) was noted compared with controls (32.0 +/- 6.3 microns 2/100 microns), but no significant differences were noted in the percentage of the basal cell membrane occupied by hemidesmosomes. Decreased penetration of the anchoring fibrils into the stroma may cause the loose adhesion between the stroma and basement membrane observed in diabetes.

Development of the anchoring structures of the epithelium in rabbit and human fetal corneas.

The sequence of development of the components of the corneal adhesion complex (hemidesmosomes, basal lamina and anchoring fibrils) was studied in rabbit and human fetal corneas using electron microscopy and histochemical localization of type VII (anchoring fibril) collagen. In the rabbit, basal lamina was present at 15 days gestation, followed by hemidesmosomes (HDs) and anchoring fibrils (AFs) at 20 days gestation. Type VII collagen was first localized at 20 days. At 25 days, HDs remained low compared to the adult value. During human corneal development, basal lamina was present at 8 weeks gestation. Through 12 weeks of gestation, no HDs or AFs were discernible nor was there any type VII localization. At 13-19 weeks, HDs and cross-banded AFs were seen, and localization of type VII collagen was first noted. A palisade of filaments extending perpendicularly from the basal lamina into the underlying stroma was discernible from 13 to 27 weeks. A distinct Bowman's layer was present at 19 weeks. By 27 weeks, HDs/micron membrane were greater than or equal to the adult value, and AF penetration into the underlying stroma was also greater than or equal to the adult value. Bowman's layer had not reached adult values by term. These data indicate that after basal lamina deposition, HDs and AFs develop synchronously in both species. In humans the palisade of filaments may be the precursor of Bowman's layer, and the AF network develops within Bowman's layer.

Hemidesmosomes and anchoring fibril collagen appear synchronously during development and wound healing.

Bullous pemphigoid antisera and monoclonal antibodies to type VII collagen were used to localize hemidesmosomes and anchoring fibrils, respectively, in tissues of developing eyes and healing corneal wounds of New Zealand white rabbits. In the 17-day fetal rabbit eye, both antibodies colocalize to the epithelial-stromal junction of the lid and conjunctival region, but neither binds to the cornea, and electron microscopy demonstrates hemidesmosomes only where the antibodies bind. By 20 days of fetal development, the antibodies colocalize in cornea, and, by electron microscopy, hemidesmosomes are shown to be present as well. In healing 7-mm corneal wounds, both antibodies colocalize at the wound periphery within 66 h. By electron microscopy, hemidesmosomes along small segments of basal lamina are also shown to be present at the wound periphery at this time. These demonstrations of the synchronous assembly of hemidesmosomes and anchoring fibrils support the hypothesis of linkage of hemidesmosomes through the basement membrane to anchoring fibrils.

An organ culture system for study of adherence of Pseudomonas aeruginosa to normal and wounded corneas.

An organ culture system has been developed to study the adherence of Pseudomonas aeruginosa to unwounded corneas and to corneas healing after a 3 mm central epithelial debridement. The Pseudomonas strain was isolated from a human corneal ulcer; suspensions containing 1 X 10(8) colony-forming units/ml (CFU/ml) of bacteria were incubated with the corneas for the last 30 min of the 18 hr culture period. The distribution pattern and number of adherent bacteria on the ocular surface were determined by morphometric analysis of scanning electron micrographs. Few bacteria (25 +/- 15/mm2) adhered to the apical cells of unwounded corneas. There was a definite region-specific distribution of adherent bacteria on healing corneas. There was a definite region-specific distribution of adherent bacteria on healing corneas. Most bacteria were found on the denuded basal lamina in front of the leading edge of the migrating epithelium (360,700 +/- 49,000/mm2). Appreciable but lower numbers adhered to the apical membrane of leading-edge cells (37,700 +/- 6,100/mm2) and to the central portion of the denuded basal lamina (28,800 +/- 10,700/mm2). No bacteria were found adherent to the apical cells of the stratified epithelium behind the leading edge of the epithelium migrating to cover the wound. A similar region-specific distribution of adherent bacteria was found when corneas were inverted in the bacterial suspension and when corneas were incubated in the bacterial suspension for 15 rather than 30 min. Corneas preincubated with the lectin, succinyl-concanavalin A, showed significantly decreased bacterial adherence, indicating a possible role for mannose moieties of wound surface glycoconjugates in bacterial adherence.

Biosynthetic responses of the rabbit cornea to a keratectomy wound.

Following a corneal wound involving removal of the epithelium and basement membrane, the epithelium must migrate across bare stroma. To examine the effect of the removal of the basement membrane on epithelial migration and on protein and glycoprotein synthesis in both the epithelium and stroma, we performed superficial keratectomies on rabbits and allowed the corneas to heal in organ culture. We then analyzed the following parameters: (1) rate of epithelial wound closure; (2) proteins synthesized during epithelial wound closure in both the epithelium and stroma using SDS-PAGE; and (3) presence of fibronectin in the epithelium and stroma using immunodot blots and immunofluorescence. We found that: (1) a 7 mm keratectomy wound heals in 66 hr with a maximal rate of epithelial migration of 0.83 mm2/hr; (2) four proteins, 400+K, 220K, 70K, and 58K, are present in the epithelium migrating to close the wound that are not seen in the control epithelium; (3) a 220K band is seen in the wounded stroma but not in control stroma; and (4) fibronectin represents 2% of the total protein in the stroma 66 hr post-keratectomy but less than 0.02% in wounded epithelium, unwounded epithelium, and unwounded stroma.

Anchoring fibrils form a complex network in human and rabbit cornea.

Anchoring fibril distribution, depth of penetration into the stroma, and pattern of histochemical localization of type VII collagen (the anchoring fibril collagen) were studied in normal human and rabbit corneas. Electron micrographs of cross sections and sections taken parallel to the basement membrane demonstrate that anchoring fibrils insert into the basal lamina and then splay out laterally. They are more readily seen in sections taken parallel to the basal lamina, where they are observed to form a complex branching and anastomosing network below the basal lamina. Distal to the basal lamina, anchoring fibrils appear to insert into patches of dense extracellular matrix termed "anchoring plaques." Average depth of penetration of anchoring fibrils into stroma is 0.60 and 0.54 microns for human and rabbit, respectively. Monoclonal antibodies to type VII collagen localize only to the basement membrane-anchoring fibril zone of both human and rabbit corneas. No obvious differences in anchoring fibril structure or distribution were observed between human corneas, which have a Bowman's layer, and rabbit corneas, which do not.

Transplant of oral mucosal epithelium to rabbit ocular surface wounds in vivo.

Sheets of epithelium freed of underlying connective tissue and basal lamina with dispase II were obtained from explants of rabbit oral mucosa. Epithelial sheets were then sutured onto abraded (basement membrane intact) or keratectomized corneal-limbal zones or abraded central corneas of anesthetized rabbits. To develop the procedure, allografts to abraded (N = 13) and keratectomized wounds (N = 3) on the corneal-limbal zone were done. All allografts were retained on the wound bed until the experiment was terminated or rejection occurred. Prior to rejection, eyes were uninflamed and quiet. Allografts on central avascular corneal wounds (N = 8) were not maintained. Autografts (N = 4) sutured to corneal-limbal abrasion wounds were maintained until animals were killed at the end of the experiment, 22, 26, 60, and 120 days after transplant. Eyes were uninflamed and quiet. A fine vascular bed developed under both allografts and autografts in the peripheral cornea. Grafts maintained some histologic characteristics of oral mucosal epithelium, even after four months. These data indicate that it is feasible to transplant oral mucosal epithelium to corneal-limbal ocular surface wounds and that the grafts will be maintained. Such transplants are, however, not maintained over central, avascular corneal regions.