Role of the Lewis x Glycan Determinant in Corneal Epithelial Cell Adhesion and Differentiation

In this study we demonstrate that in corneal epithelium there is cell-cell contact-regulated expression of a 145-kD glycoprotein bearing the glycan determinant Le x (Gal b (1,4)[Fuc a (1,3)]GlcNAc). This glycoprotein (Le x -GP) was expressed in confluent/contact-inhibited cultures but not in sparse cultures of corneal epithelium. In contrast, a 135-kD glycoprotein bearing precursor, unfucosylated, lactosamine-containing glycans (Gal b 1-4GlcNAc b 1-R) was expressed in sparse cultures . Immunofluorecence staining and confocal microscopy of confluent cultures revealed that in corneal epithelium, Le x -antigen is located in high density at sites of cell-cell adhesion. In in vitro cell-cell adhesion assays, anti-Le x , but not anti-sialyl-Le x mAbs, inhibited the formation of corneal epithelial cell-cell adhesion. Also, when added to confluent cultures, antibodies to Le x disrupted the monolayer and caused tightly packed polygonal cells to round up. Analysis of the expression of Fut genes that encode the enzymes which generate the Le x determinant, revealed that confluent/contact-inhibited cultures of rabbit corneal epithelium contain markedly elevated levels of Fut 4 and Fut 3/5/6 gene transcripts compared to sparse cultures. These data suggest that the Fut 4 and Fut 3/5/6 genes are targets of cell-cell contact- regulated signals and that Fut gene products direct cell-cell contact-associated expression of Le x on the Le x -GP in corneal epithelium. x (1,3)fucosyltransferase -antigen density sites adhesion, (iii) to Le x inhibit and (iv) there is expression of the x


INTRODUCTION
Corneal epithelium is a prototype stratified squamous epithelium. Its major function is to provide a barrier against fluid loss and pathogen entrance. This function requires the cells of the epithelium to remain tightly adherent to one another, as well as to the underlying basal lamina. The structural integrity of corneal epithelium is requisite for normal vision. Abnormality or loss of corneal epithelial cell-cell adhesion may lead to the development of a number of ocular surface disorders including corneal epithelial dysplasia and dysmaturation (1)(2)(3), corneal epithelial hyperplasia and corneal epithelial ingrowth, which is a rare but disastrous complication that can occur following intraocular surgery (4)(5)(6). In many of these conditions, enlarged intercellular spaces may be seen ultrastructurally in the epithelium (1)(2)(3)(4)(5)(6).
Numerous studies on skin epithelium have shown that the loss of cell-cell adhesions in the epidermis produces life-threatening blistering skin diseases known as pemphigus foliaceus and pemphigus vulgaris (7)(8)(9). An understanding of the molecular mechanisms of epithelial cell-cell adhesion is also crucial to the understanding of complications relating to the failure of re-epithelialization of wounds. In response to injury, cells at the leading edge undergo a phenotypic conversion characterized by a dramatic reorganization of the cytoskeleton, disruption of stable intercellular adhesion, and redistribution of adhesion-related molecules. In fact, the breakage of the stable intercellular contacts is prerequisite for initiating the phase of re-epithelialization. Following re-epithelialization, reversion to the epithelial phenotype including the reformation of stable intercellular contacts must occur if the function of the epithelium is to be fully restored. The molecules which mediate cell-cell adhesion contacts also play crucial roles in important biological processes such as tissue morphogenesis and cell differentiation (reviewed in [9][10][11][12]. In normal, intact epithelium, intercellular adhesion is brought about by a variety of cell by guest on March 24, 2020 http://www.jbc.org/ Downloaded from 4 adhesion molecules, including cadherins, desmogleins and desmocollins (7)(8)(9)(10)(11)(12). Cell-cell adhesion might also be mediated by the carbohydrate determinant Lewis x (Le x ; Galβ (1,4)[Fucα (1,3)

]GlcNAc).
Several studies have suggested that Le x -side chains of plasma membrane glycoconjugates play an essential role in cell-cell interactions during embryogenesis in the mouse (13)(14)(15)(16)(17). In the developing mouse embryo, the Le x antigen appears after the third cleavage (8- (14). Eggens et al. (15) and Kojima et al. (16) demonstrated that LNFP III also inhibits the intercellular adhesion of teratocarcinoma cells. In these cells and in developing mouse embryos, Le x -Le x interactions occurring between opposing homotypic cell surfaces have been postulated to mediate cell-cell adhesion (15,16). Several other studies have also implicated Le x as a possible adhesion molecule during embryogenesis (18)(19)(20). In the present study we show that, in rabbit corneal epithelium: (i) there is cell-cell contact-regulated expression of a Le x bearing 145-kD glycoprotein (Le x -GP), and of α(1,3)fucosyltransferase genes which mediate the synthesis of the Le x -side chains, (ii) Le x -antigen is located in high density at sites of cell-cell adhesion, (iii) antibodies to Le x inhibit the formation of cell-cell adhesion contacts and (iv) there is cell differentiation stage-specific and developmentally regulated expression of the Le x antigen.

EXPERIMENTAL PROCEDURES Preparation of rabbit corneal epithelial cell cultures
Rabbit corneal epithelial cells were grown in tissue culture using eyes from Pel-Freez Biologicals (Rogers, AK) as previously described (21,22). Sparse/exponentially growing cultures, and confluent/contact-inhibited cell cultures were analyzed. Sparse cultures were collected 3 to 4 days after starting the culture at which time 30%-50% of the culture dish was populated with rapidly dividing cells that were migrating away from explants ( Fig. 1). Confluent cultures were collected 10 to 12 days after starting the culture when 90%-95% of each dish was populated with tightly packed, polygonal cells (Fig. 1).

Western blot analysis of Le x -glycoproteins of confluent and sparse cultures of corneal epithelium
MAb MMA, which is specific for Le x (Galβ1-4[Fucα1-3]GlcNAcβ1-R) (23) and mAb 1B2 which binds to terminal N-acetyllactosamine disaccharide (Galβ1-4GlcNAcβ1-R), the unfucosylated precursor of Le x (24) were used. Hybridomas secreting both mAbs were purchased from American Type Culture Collection (Rockville, MD). To isolate fractions enriched in Le x -containing glycoproteins (Le x -GPs), 1.0 ml of cold radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-6 (AAL, a fucose-specific lectin; Vector Labs, Burlingame, CA) for 1 h at 4 o C. After the incubation period, the lectin-agarose beads were washed five times with RIPA buffer and 26 µl of electrophoresis sample buffer (50 mM Tris-HCl containing 2% SDS, 10% glycerol, and 100 mM dithiothreitol pH 6.8) was added. The samples were boiled for 4 min, centrifuged, and the supernatants were electrophoresed on 10% polyacrylamide gels in the presence of SDS. Proteins from the gels were blotted onto nitrocellulose paper. To identify Le x -GPs of corneal epithelium, gel blots were treated with 2% bovine serum albumin (BSA) in PBS (2h, room temperature) to block nonspecific binding and were then incubated with mAb MMA (undiluted hybridoma fluid, overnight, 4 o C). Gel blots were developed with peroxidase-labeled antimouse IgM using a chemiluminiscent detection system (25) (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD). To determine whether there is a correlation between the contact-inhibition of cell growth and the expression of the Le x -GP, cell cultures at various densities (day-4, sparse; day-10, early confluent; day-15, confluent and day-18, post confluent) were incubated in SHEM medium containing 3 H-thymidine (2µCi/ml, 6.7 Ci/mmol) for 16 to 18 h. At the end of the incubation period, cells were extensively rinsed with PBS and harvested for estimating 3 H-thymidine incorporation and Western blot analysis.
To determine whether glycoproteins containing precursor, non-fucosylated, lactosamine glycans are expressed in corneal epithelium, Western blot analysis was performed using mAb 1B2 which reacts with terminal (Galβ1-4GlcNAcβ1-R)-side chains. Briefly, detergent extracts of cell cultures of rabbit corneal epithelium were incubated with agarose beads conjugated with a βgalactose-specific lectin, Ricinus communis agglutinin-I (RCA-I). The RCA-I-bound glycoproteins were electrophoresed, and gel blots were analyzed for reactivity with mAb 1B2 as described above.  (1:1) and were diluted as needed using dialyzed EMEM:F12 containing 5% fetal bovine serum (MFS).

Cell-cell adhesion assay
To determine the effect of anti-Le x mAbs on cell-cell adhesion, primary cultures of rabbit corneal epithelium were trypsinized and were plated in 24-well plates (3 x 10 5 cells/well in 0.6 ml) in SHEM media (22). After allowing the cells to adhere for 3 h, they were rinsed with EMEM: In one experiment, to determine whether anti-Le X mAb was able to disrupt a contactinhibited monolayer, mAb MMA was added to tightly packed monolayer cultures of corneal epithelium. This assay was carried out as described above except that cell cultures in 24-well plates were incubated overnight instead of for 3 h in serum containing SHEM media to obtain monolayers consisting of tightly packed polygonal cells. These monolayers were rinsed with EMEM:F12 (1:1) and were then incubated with MFS medium in the presence and absence of mAb MMA (0.05 to 0.8mg/ml) and IgM (0.8mg/ml) for 20-24 h. The cultures were then evaluated under a phase-contrast microscope for cell rounding and lifting of cell clumps from the culture dish.

RNA isolation
Poly(A) + RNA from confluent and sparse cultures of corneal epithelium and from CHO cell mutants LEC11 (28) and LEC30 (29) was isolated using

Genomic DNA preparation
One gram of fresh liver (mouse or rabbit) was cut into small pieces and was incubated with 12 ml of digestion buffer consisting of 100 mM NaCl, 10 mM Tris HCl pH 8.0, 25 mM EDTA, 0.5% 9 SDS, and 0.1 mg/ml proteinase K for 16h at 55 o C. The digest was extracted twice with saturated phenol and twice with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1). DNA was precipitated by the addition of 2 volumes of 7.5M sodium acetate and 2 volumes of 100% ethanol.
After rinsing the precipitate with 70% ethanol, genomic DNA was dissolved in TE buffer and stored at 4 o C.

Immunohistochemical localization of Le x -glycoconjugates in developing rabbit corneas
New Zealand white rabbits (Milbrook Farms, Amherst, MA) were used throughout this study.

All animal treatments in this study conformed to the Association for Research in Vision and
Ophthalmology

Cell-cell contact-regulated expression of Le x -GP in corneal epithelium
Confluent and sparse cultures of corneal epithelium were analyzed to determine whether there Fut3/5/6 (28) respectively (Fig. 4A). Clustal W multiple sequence alignment analysis (Fig. 4B) showed that the deduced amino acid sequence of the rabbit Fut4 PCR fragment was closely related to that of human, mouse and hamster Fuc-TIV (human: 90% similarity, 88% identity; mouse and hamster: 87% similarity, 84% identity

Cell differentiation stage-specific expression of Le x determinant in adult rabbit corneas
In the avascular corneal epithelium, stem cells are confined to a vascularized outer rim of the cornea known as the limbus, the transition zone between cornea and conjunctival epithelium ( The characteristic anti-Le x staining pattern of adult corneal, limbal, and flanking conjunctival epithelium is shown in Figure 6. Because limbal epithelium overlies a "loose" connective tissue, it was easily differentiated from the corneal epithelium that overlies compact, plywood-like stroma (36). While, in corneal epithelium, Le x immunoreactivity was predominantly detected in basal and middle cell layers (Figs. 6Bi[arrowhead], 6Biii and 6Bv), in conjunctival epithelium the Le x expression was seen largely in superficial cell layers (Fig. 6Bi, black arrow). No Le x -immunoreactivity was detected in 18 basal or suprabasal cell layers of conjunctival epithelium in any of the 17 specimens analyzed. For comparison purposes, three rabbit corneas were also stained with another anti-Le x mAb, mAb 7A (37). Similar results were obtained with both mAbs 7A and MMA except that the staining intensity with mAb 7A was often less than that observed with mAb MMA. No staining was detected when tissue sections were treated with a control mAb (mAb 2D4, not shown).
To determine whether the Le x antigen is widely expressed in the stratified epithelia of adult tissues at the site of cell-cell adhesion, we surveyed a variety of epithelial tissues of three adult rabbits for Le x expression by immunohistochemical analysis. In epithelia of six different tissues analyzed, Le x immunoreactivity in the basal and the middle cell layers was seen only in the cornea. In other stratified epithelia, Le x immunoreactivity was either not detected (epidermis) or was detected only in the superficial cell layers (conjunctiva, tongue, esophagus and bladder).

Expression of Le x -glycoconjugates in the cornea is developmentally regulated
To further determine whether there is a correlation between Le x expression and the differentiation stage of corneal epithelial cells, we analyzed the Le x expression pattern of developing corneas. Frozen sections of 21-day and 27-day-old rabbit embryos and of offspring of various age groups (1 day -12 weeks) were stained with mAb MMA. Table 1 shows a summary of the Le x expression pattern of developing corneas. No Le x immunoreactivity was detected in the corneas of rabbit embryos or of 1-to 10-day-old animals. At postnatal day12, around the time of eyelid opening, Le x immunoreactivity was transiently detected in central corneal epithelium in three of the seven corneas analyzed (Fig. 7A, Table 1). In these three corneas, no immunoreactivity was detected in peripheral ( Fig.7C) corneal epithelium. In fact, the Le x distribution pattern detected in these corneas, i.e., robust expression in central corneal epithelium with the antibody staining diminishing progressively towards the periphery of the cornea (Fig. 7), was exactly the opposite of that seen in the adult corneas (Fig. 6). No Le x immunoreactivity was detected in epithelia of 14-or 15-day-old corneas (Table 1). Weak Le x immunoreactivity of peripheral corneal epithelium was first detected in 1-monthold corneas. With increasing age of the animal, the Le x immunoreactivity in peripheral corneal epithelium progressively increased. Between the second and third month of age, the Le x distribution pattern approached that of adult corneas with peripheral epithelium staining intensely with the anti-Le x mAb and epithelium in the central region either not staining or staining only weakly (Fig. 6, Table 1).

DISCUSSION
In the present study, we demonstrate cell-cell contact-regulated expression of a 145-kD glycoprotein bearing the Le x determinant (Le x -GP) in corneal epithelium. While confluent cultures of corneal epithelium were found to robustly express Le x -GP, sparse cultures did not express detectable levels.