human corneal and conjunctival epithelia produce a mucin-like

Human Corneal and Conjunctival Epithelia Producea Mucin-like Glycoprotein for the Apical Surface

Hitoshi Watanabe,* MicheleFabricant,* Ann S. Tisdale* SandraJ. Spurr-Michaud*Kristina Lindbergh and Ilene K. Gipson*

Purpose. The authors have determined that the corneal and conjunctival epithelia of the ratproduce a mucin-like glycoprotein at the apical surface of the epithelium. The purpose ofthis study was to determine if human ocular surface epithelium produces similar glycoproteins.

Methods. Because our initial attempts at production of monoclonal antibodies yielded bloodtype A-specific antibodies, corneal epithelia from blood type O donor eyes were used for theproduction of monoclonal antibodies. Screening of hybridoma products was accomplishedby immunofluorescence microscopy of cryostat sections of blood type O donor eyes. Themonoclonal antibody produced was used for immunofluorescence microscopy and immuno-electron microscopy top determine tissue and cellular distribution, respectively. Immunoblotanalysis of SDS-PAGE-separated proteins from corneal epithelial tissue and from culturedhuman corneal epithelium was used to determine molecular weight range and epitope bindingafter periodate oxidation, N-glycanase, and O-glycanase treatment.

Results. A monoclonal antibody, designated H185, that binds to apical cell layers of humancorneal, conjunctival, laryngeal, and vaginal epithelium was produced. The antibody bindsto 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 producesH185 antigen. By immunoblot analysis, H185 binds a high molecular weight protein, >205kD, from corneal epithelium and cultured corneal epithelium. The protein band visualizedby immunoblot analysis cannot be stained by Coomassie or silver stain on SDS-PAGE, but itdoes stain with Alcian blue followed by silver stain, indicating that the protein is highlyglycosylated. H185 binding to blotted proteins is destroyed by sodium periodate treatmentand 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 surfaceepithelium. These data suggest that the entire ocular surface epithelium produces mucinsfor the tear film. Invest Ophthalmol Vis Sci. 1995;36:337-344.

JL he presence of a stable tear film is essential to pro- the tear film associates with membranes of the apicalvide a smooth refractive surface for the corneal epithe- cells of the epithelium. This relationship must be aHum. To be stable, the tear film with its three major dynamic one because apical cells are shed from thecomponents, lipid, aqueous, and mucus, must main- surface. Several reports have shown that the glycocalyxtain an intimate association with the apical cells of the of the apical membrane of the ocular surface is richocular surface epithelium.1 Little is known about how in polyanions2'3 and in carbohydrate moieties4"8; how-

ever, specific information about the molecular natureof the cell-tear film interface is not available. Clarify-

From the* Cornea Unit, Schepens Eye Research Institute and the Department of j n g t h e molecule(s) that play a role in the mainte-Ophlhalmologf, Harvard Medical School, Boston, and from fBioSurface ° w r /Technology, inc., Cambridge, Massachusetts. nance of a stable tear fluid may lead to a better under-Supported by National Institutes of Health grant R37-EY03306 (IKC). Standing of OCUlar Surface disease resulting from drySubmitted for publication June 6, 1994; revised September 12, 1994; accepted ° 6 /September 13,1994. eye syndrome.propnetary interest categmy: N. Recently, we reported the production of a mono-Reprint requests: Ilene K. Gipson, Schepens hye Research Institute, 20 Stamford ' l "street, Boston, MA 02114. clonal antibody, R339, that binds to apical cell mem-

Investigative Ophthalmology & Visual Science, February 1995, Vol. 36, No. 2Copyright © Association for Research in Vision and Ophthalmology 337

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933410/ on 04/08/2018

338 Investigative Ophthalmology 8c Visual Science, February 1995, Vol. 36, No. 2

branes of the ocular surface epithelium of the rat.9

Immunoelectron microscopic study has shown thatthe antigen recognized by the antibody is presentalong the apical -microplical membrane in the regionof the glycocalyx. In subapical cells, the antigen is incytoplasmic vesicles. The antigen appears to be storedalong the membrane of cytoplasmic vesicles and thenmoves to the apical membrane of the ocular surfacecells when cells differentiate to the apical-most posi-tion. The antigen is prominent at the tips of micropli-cae, where they interface with the mucin layer of thetear film. It is hypothesized that the antigen recog-nized by R339 antibody facilitates tear-film spread overthe surface of the epithelium. R339 antigen has beenisolated and shown to have mucin-like characteristics(manuscript submitted for publication, 1994). Ap-proximately 60% of the mass of the molecule is carbo-hydrate. Carbohydrate and amino acid analyses indi-cate high levels of N-acetyl galactosamine and serineand threonine, respectively. R339 antibody is a usefulresearch tool; however, because it is species specific,its usefulness in the study of the human ocular surfaceepithelium is limited.

To begin to investigate the specific molecularcomposition of the apical membrane of the humanocular surface epithelium, we developed a mono-clonal antibody against the antigen expressed in theapical membrane of human ocular surface epithe-lium.10 In this article, we demonstrate the develop-ment of a human-specific antibody, designated H185,that has characteristics similar to those of R339, whichbinds to apical cells of human ocular surface epithe-lium. We describe several characteristics of the antigenrecognized by this antibody and report that the anti-gen can be produced by cultured human corneal epi-thelium as well as by native tissue.

MATERIALS AND METHODS

All investigations conformed to the ARVO Statementfor the Use of Animals in Ophthalmic and Vision Re-search and to the tenets of the Declaration of Helsinki.

Monoclonal Antibody ProductionPreparation of Immunogen and Immunization. Previ-

ous attempts to produce monoclonal antibodies tomucin-like molecules produced by human ocular sur-face epithelium yielded blood type A-specific antibod-ies." To obtain antibodies that recognize all bloodtypes, corneal epithelial cells were obtained by debrid-ing epithelium from eye bank eyes of donors withblood group type O. Eyes were obtained less than 24hours after donor death. The epithelial cells were re-suspended in equal amounts of phosphate-bufferedsaline (PBS) and Ribi adjuvant system (RIBI Immunochem Research, Hamilton, MT). The human corneal

epithelial cell suspension was injected intraperitone-ally into 6-week-old BALB/cByJ mice (Jackson Labora-tories, Bar Harbor, ME).

Cell Fusion and Hybridoma Cloning. Cell fusion wasperformed 4 days after the second boost, as previouslydescribed.9 The supernatant of growing cells wasscreened on cryostat sections of human corneas ofblood group type O by immunofluorescence micros-copy to choose hybridomas of interest. Hybridomasshowing apical cell binding by immunofluorescencemicroscopy were cloned by limiting dilution (0.5 cell/well) two consecutive times. The clones were similarlyscreened, and clone H185.4.6, herein referred to asH185, was chosen for further characterization. H185was concentrated from tissue culture medium by am-monium sulfate precipitation.12 The isotype waschecked using the Pierce Immunopure MonoclonalAntibody Isotyping Kit I (Pierce, Rockford, IL) andfound to be I g d , with kappa light chain.

Tissue Preparation and Cell CultureNormal human tissues were supplied by the NationalDisease Research Interchange (NDRI, Philadelphia,PA). They were obtained less than 24 hours after do-nor death and were frozen in OCT compound (Miles,Elkhart, IN). Human corneal epithelial cell cultureswere derived from limbal epithelium of discarded rimsof donor corneas used for penetrating keratoplasty.Cells were cultured as described by Lindberg et al.13

Primary cultures were propagated on irradiated 3T3fibroblasts. Cells at passages 3 to 6 were used for HI85characterization.

Immunohistochemistry and ImmunoelectronMicroscopy

Immunofluorescence microscopy was performed aspreviously described.9 Briefly, unfixed 6-/im cryostatsections of human cornea and conjunctiva of A, B,and O blood groups were incubated for 1 hour atroom temperature with H185. Sections were washedwith PBS and incubated for 1 hour at room tempera-ture with fluorescein isothiocyanate-conjugated don-key anti-mouse IgG (Jackson Immuno Research, WestGrove, PA). Sections were then washed with PBS andcoverslipped using a mounting medium containingPBS, glycerol, and paraphenylenediamine.14 In addi-tion, cryostat sections of human larynx, vagina, skin,esophagus, trachea, buccal mucosa, large intestine,bladder, ileum, liver, and mammary gland werescreened by immunofluorescence for H185 binding.To check for cross-reactivity with other species, cryo-stat sections of corneas of owl monkey, calf, sheep,rat, and rabbit were similarly screened for binding ofHI85. Cultures of human corneal epithelium werefixed in ice-cold methanol and rinsed with PBS beforethe usual immunofluorescence protocol to check for


Corneal and Conjunctiva! Epithelia Produce Mucin 339

H185 binding. Negative control tissue sections (pri-mary antibody omitted) were run with each experi-ment. Sections were viewed and photographed on aZeiss Photo Microscope III (Zeiss, Oberkochen, Ger-many) equipped for epi-illumination.

Immunoelectron Microscopy. Post-embedding im-munoelectron microscopic localization of H185 onfixed (0.2% glutaraldehyde, 4% paraformaldehyde,0.1 M phosphate buffer) and LR White (Ernest F.Fullam, Latham, NY)-embedded human cornea, con-junctiva, and cultured human corneal epithelium wasperformed as previously described,9 except that thesecondary antibody used was goat anti-mouse IgG con-jugated to 10-nm gold (Ted Pella, Redding, CA). Bind-ing of HI85 to human ocular surface epithelium inthe LR White-fixed samples was confirmed by immu-nofluorescence microscopy.

Electrophoresis and Immunoblotting

Corneal epithelium debrided from human eye bankeyes and cultured human corneal epithelium were sol-ubilized in 2% sodium dodecyl sulfate (SDS). Proteinconcentration was determined using the Micro BCAprotein assay reagent kit (Pierce, Rockford, IL). Theproteins were then diluted with Laemmli samplebuffer (containing SDS, /?-mercaptoedianol) and runon 6% separating, 4% stacking SDS-polyacrylamidegel electrophoresis (PAGE) using the Laemmli buffersystem.l5 After SDS-PAGE, gels were stained with Coo-massie blue, silver stain, periodic acid-Schiff,16 or by amodification of an Alcian blue-silver stainingmethod.17 For immunoblotting, proteins were electro-phoretically transferred to nitrocellulose and probedfor reactivity with H185, as previously described.9

Periodate TreatmentPeriodate Oxidation of Inununoblots. To examine if

the H185 epitope is carbohydrate, proteins trans-ferred onto nitrocellulose membranes were pre-treated with varying concentrations (0, 0.1, 1, 5, 10,and 20 raM) of sodium periodate in 50 mM sodiumacetate buffer, pH 4.5, for 1 hour at room temperaturein the dark.18 After rinses in 50 mM sodium acetatebuffer, the membranes were incubated in 50 mM so-dium borohydride in PBS for 30 minutes at room tem-perature. After further rinsing in PBS, the membraneswere probed widi HI85 as usual.

Enzyme Treatment

N-glycanase Treatment. To determine if HI85 epi-tope is an N-linked sugar, N-glycanase (Genzyme,Cambridge, MA) was used according to manufactur-er's protocols on denatured corneal protein followedby immunoblot with H185. Cultured human cornealepithelial cells were solubilized in 10 mM Tris-HCl,pH 7.5, with 10% sucrose and 0.2 mM phenylmediane

sulfonyl fluoride (Boehringer Mannheim, Indianapo-lis, IN). Insoluble material was spun out by ultracen-trifugation at 300,000gfor 1.5 hours, and the solublesupernatant was denatured by boiling for 5 minutesin the presence of 0.5% SDS and 50 mM /?-mercapto-ethanol. NP-40 was added in a sevenfold excess (byweight to SDS) to protect the N-glycanase enzymefrom denaturation by the SDS. One unit of N-glyca-nase was added, and the sample was incubated for 24hours at 37°C. As a control, one sample was treatedwith heat-inactivated N-glycanase under identical con-ditions. The reaction was stopped by die addition ofSDS-reducing sample buffer. The samples were boiledfor 5 minutes and run on 6% SDS-PAGE for analysisby immunoblotting with H185.

O-glycanase Treatment. To determine if the carbo-hydrate epitope of HI85 is O-linked, denatured cor-neal protein was digested with neuraminidase (Gen-zyme, Cambridge, MA) followed by O-glycanase (Gen-zyme, Cambridge, MA) according to manufacturer'sprotocols. To determine if digestion removed HI85binding, immunoblot analysis was used. Cultured hu-man corneal epithelial cells were solubilized in 15 mMsodium phosphate buffer, pH 7.3, widi 0.2 mM phenyl-methane sulfonyl fluoride. After centrifugation at3000 rpm for 30 minutes to spin down any insolublecomponents, the supernatant was denatured and NP-40 was added to protect die enzyme as above for N-glycanase treatment; 10 mM sodium acetate and 1 Uneuraminidase was added, and die sample was incu-bated for 60 minutes at 37°C. After this, 4 mU of O-glycanase was added and incubated overnight at 37°C.As controls, one sample was treated with neuramini-dase alone and a second was treated with heat-inacti-vated O-glycanase after neuraminidase digestion. Thereactions were stopped by the addition of SDS-reduc-ing sample buffer, and samples were run on 6% SDS-PAGE for analysis by immunoblotting with H185.

RESULTS

Immunolocalization

By immunofluorescence microscopy, we observedthat the binding of H185 was to all the flattened celllayers, or squames, along the entire ocular surfaceepithelium of the cornea (Fig. IB). In the conjunc-tiva (Fig. 1C), binding was intense along the apicalsurface of the epithelium and extended down fourcell layers. Binding ended abruptly at the lid marginat the junction with keratinized epidermal epithe-lium (Fig. 1C, inset). In conjunctiva in a portionof the specimens, the goblet cells also bound theantibody (Fig. 1C). This finding was not consistentbetween tissue samples. No binding of H185 wasobserved in basal cells in either the cornea or the


340 Investigative Ophthalmology & Visual Science, February 1995, Vol. 36, No. 2

FIGURE 1. Micrographs demonstrating binding of monoclonal antibody H185 to humantissues. (A) Light micrograph of human cornea for comparison to immunofluorescencemicrograph of H185 localization in cornea shown in (B). Note that only apical flattened cellsbind, (arrow) Region of basement membrane. (C) Binding to the conjunctival epithelium isalso to apical cells, and, in this specimen, goblet cells bind intensely (not all specimens showintense goblet cell binding). Arrow in inset indicates abrubt ending of H185 binding at lidmargin. (D) Cultured corneal epithelial cells bind HI85 to varying degrees. Cells adjacentto culture dish do not bind to the extent that stratified cells do. (E, F) Binding to sectionsof human laryngeal and vaginal epithelium, respectively. Note that apical cells of the epithe-lium bind the antibody. (A, B, D, F) Bar = 10 //m; (C, E) Bar = 50 pun.

conjunctiva. In comparison with our previous at-tempts at making human-specific antibodies, whichwere blood type A specific," the binding of H185showed the same pattern in human corneal epithe-lium regardless of blood group (A, B, or O). Bindingof H185 was not found in other species tested (owl,monkey, rabbit, rat, cow, sheep) (data not shown).Interestingly, as in rat studies with R339 antibody,HI85 binding was observed in human vaginal andlaryngeal epithelium. It was present along the apicalsurface and cell layers below the apical cells, similar

to that in ocular surface epithelium (Figs. IE, IF).Other epithelia found to be devoid of binding in-clude buccal mucosa, esophagus, ileum, colon,mammary epithelium, and liver. H185 binding wasseen in several human glandular tissues, includingeccrine sweat glands and submucosal glands of theesophagus and large intestine (data not shown).

By immunoelectron microscopy, binding was on theouter apical membrane adjacent to the tear film (Fig.2A). Binding was prominent at the tips of the micropli-cae (Figs. 2A, 2B). In the squames below the apical-most


Cornea] and Conjunctiva] Epithelia Produce Mucin 341

<•••• • * " > ? - -•«

FIGURE 2. Iinmunoclcctron microscopic localization of H185 in human cornea] epitheliumin situ (A, B) and in cultured human corneal epithelium (C). In (A), tissue has not beencounterstained with lead or uranyl acetate to demonstrate more clearly the gold particlesthat indicate sites of H185 binding. (B) Binding on counterstained tissue. Note that insubapical cells, binding is primarily associated with cytoplasmic vesicles (small arrows),whereas in apical cells, binding is particularly prevalent along the tips of the microplicae(large arrows). Cultured cells in (C) show intense binding to microplicae and to the peripheryof cytoplasmic vesicles. Note also the two adjacent cells (1 and 2) vary in the degree of H185binding, (inset) H185 binding material in the lumen of a large vacuole suggests that H185is secreted from the cell membrane. H185 binding could also be detected in the culturemedia by immunoblot analysis. Bars = 0.2 /xm.



FIGURE 3. Immunoblot analysis of H185 binding to proteinsof debrided human corneal epithelium from blood typeO (lane 1) and blood type A (lane 2) proteins, {lane 3)Immunoblot of proteins from cultured corneal epithelium.{lane 4) Alcian blue-silver stain of the SDS-PAGE of pro-teins from cultured corneal epithelium, indicating a highlyglycosylated molecule in the same molecular weight range asthe immunoreactive band, {arrow) 205-kD molecular weight.

cells, the antigen was present and was associated withsmall vesicles in the cytoplasm (Fig. 2B).

In cultures of human corneal epithelium derivedfrom limbal tissue, islands of cells develop within cul-tures.IS These islands are stratified, and it is the apicalcells of the islands that bind H185 (Fig. ID). Immuno-electron microscopy of HI85 on these cultured cellsdemonstrates binding to cytoplasmic vesicles, as wellas to the tips of microplicae on the cell membrane(Fig. 2C). Binding in the cell cultures was more in-tense with less scatter than in tissue samples (Figs. 2A,2B). This may be due to the fact that cultured cellsare fixed immediately, whereas tissue was obtained upto 24 hours after donor death. Scatter in tissue maybe due to postmortem degradation. Cells within theculture show varying degrees of binding of H185 (Fig.2C) and within large vacuoles within cultures, H185binding material was present. This observation indi-cates that H185 is secreted or shed from cell surfaces.

Electrophoretic Mobility and Immunoblots

H185 reacted with a prominent band with a molecularweight greater than 205 kD (Fig. 3). No staining wasobserved in the area of the immunoreactive band witheither Coomassie blue or silver staining of the gel.PAS staining (data not shown), as well as AJcian bluefollowed by silver staining (Fig. 3, lane 4), stained aband in the region of molecular weight similar to thatof the immunoreactive band. These data suggest thatthe antigen is a highly glycosylated glycoprotein. Asimilarly reactive band was found in cell-free, spentculture media from corneal epithelial cultures.

Periodate Incubation

To determine whether the epitope recognized byHI85 is to the carbohydrate portion of the glycopro-tein, we used sodium periodate oxidation on blots ofproteins from both native and cultured human cor-neal epithelium (Fig. 4). There was a significant de-crease in the binding of HI85 at a concentration of0.1 mM sodium periodate and a complete loss ofibind-ing at concentrations of 1 mM or more of sodiumperiodate. These data indicate that the epitope recog-nized by H185 is carbohydrate.18

Enzyme Treatments

To determine whether the epitope of the antigen rec-ognized by H185 is an N- or O-linked carbohydrate,we assessed the binding of H185 to blots of proteinssolubilized from cultured human corneal epitheliumtreated with either N- or O-glycanase. N-glycanasetreatment had no effect on the binding of H185, indi-cating that the carbohydrate epitope recognized byH185 is not N-linked to the protein core (Fig. 5A).In contrast, H185 binding was lost after O-glycanasetreatment (Fig. 5B). This result indicates that the car-bohydrate epitope recognized by HI85 is O-linked tothe protein core, indicative of mucins.

DISCUSSION

Data obtained through the use of H185 antibody indicatethat, as in the rat,9 the stratified squamous epithelia ofthe human ocular surface produces mucin-like glycopro-teins for the apical surface-tear film interface. The H185antigen has an identical cellular and tissue distributionto the rat R339 antigen. Both H185 and R339 antibodiesbind to the flattened apical squamous cells; they bothbind to cytoplasmic vesicles in subapical squames and

205 —

O m M .lfflM l m M 5mM

FIGURE 4. Periodate oxidation of nitrocellulose-blotted SDS-PAGE-separated proteins from both debrided corneal epi-thelium (CE) and cultured human corneal epithelium(CHCE). Higher concentrations of sodium periodate com-pletely removed the epitope to which H185 binds. The 205indicates molecular weight marker (MWM), and mM con-centrations indicate the amount of sodium periodate used.


Corneal and Conjunctiva! Epithelia Produce Mucin 343

205 —

N-GLY

I i

205 —

feiO-GLYB

FIGURE s. Immunoblot analysis of N-glycanase-lreated (A)and O-glycanase-treated (B) proteins from cultured humancorneal epithelium. (A) Active N-glycariase treatment (laneA) did not destroy HI 85 binding more than heat-inactivatedenzyme (lane In). Untreated control (lane Un) is an equiva-lent sample incubated for a comparable length of time with-out the added enzyme aliquot. (B) Active O-glycanase treat-ment proceeded by neuraminidase treatment (lane A) de-stroyed HI 85 binding, whereas head-inactivated'O-glycanasedid not (laneIn). Neuraminidase alone did not destroy H185binding (laneNe). Untreated control (lane Un) is an equiva-lent sample incubated a comparable length of time withoutenzyme additions or boiling after enzyme additions. Diminu-tions in binding in lanes NE and In may reflect degradationdue to boiling. MWM also contains untreated control pro-teins, as seen above the 205-kD protein.

to the apical membranes of the apical cells. Both bindprominently to the tips of the microplicae in the zone ofthe glycocalyx. H185 and R339 antigen are of similarmolecular weight, and both stain with PAS or Alcian bluefollowed by silver stain. The R339 antigen has been iso-lated and has been shown to have an approximately 60%carbohydrate:40% protein ratio, placing the molecule inthe mucin class of glycoproteins (manuscript submittedfor publication, 1994). Although H185 has not been iso-lated to the purity that allows such assay, the similarity incellular localization, size, and H185 epitope removal byO-glycanase suggests that H185, like R339, is a mucin.

The past few years has seen significant progress inunderstanding the molecular nature of mucins.'9"21 To

date, seven different mucins have been cloned. One ofthese mucins, designated MUC-1, has been reported tobe present in the stratified squamous epithelium of therat vagina.22 This so-called "epithelial" mucin is the onlymucin cloned that has a membrane-spanning region inits protein backbone.23 Mucins 2 to 7 are "secreted" mu-cins and have only been demonstrated in simple epithelia.MUC-1 may be a candidate mucin for the H185 antigen.The identification of R339 and HI85 as one of the clonedmucins, or a unique mucin, awaits molecular cloning.

The data from studies of the rat and now humansuggest that, besides goblet cells, the entire ocularsurface epithelium, including both conjunctival andcorneal epithelia, produce mucins. The observationthat the corneal epithelium produces the mucin is, toour knowledge, a new finding. This is proven by thefact that HI85 antigen is produced by corneal epithe-lium in culture, away from a potential conjunctivalor goblet cell source. That conjunctival epitheliumprovides a second source of mucin for the tear film isnot a new concept. Greiner et al5i24 proposed—basedon PAS or Alcian blue followed by PAS staining ofvesicles that appeared to discharge to the apical sur-face of human conjunctival epithelial cells—that thestratified squamous epithelium provided mucin forthe tear film. Dilly and Mackie3'25 have also providedhistochemical and morphologic evidence that vesiclesfilled with highly glycosylated material were dis-charged onto the conjunctival epithelial cells' apicalsurfaces. Our data corroborate and extend these stud-ies in that we demonstrate that the cornea, as wellas the conjunctiva, produce mucin-like glycoproteins.Designation of mucin-like glycoproteins, in this in-stance, is based on biochemical characteristics.

Two questions arise: What proportion of the mu-cus layer of the tear film is provided by goblet cells,and what proportion is provided by the stratified epi-thelium? Answers await probes that discriminate be-tween products of the two cell types. Recent data sug-gest that goblet cells are innervated and secrete inresponse to neural stimulation.26 Perhaps epitheliumsupplies tear film mucus and goblet cells provide bolusmucus secretion in response to sensory stimulation.

Data from the rat studies using R339 antibodysuggest that R339 is a component of the glycocalyxlayer.9 The glycocalyx, so elegantly demonstrated inrapid freeze-prepared electron micrographs by Nich-ols et al,27 is the layer directly adjacent to the epithe-lium. These micrographs demonstrate the glycocalyxas an electron-dense zone, particularly prominentalong the tips of the microplicae, which meshes gradu-ally and without abrupt interface with the mucus layerof the tear film in the guinea pig cornea. Both H185and R339 bind avidly to tips of microplicae in tissuesand in cells in culture. Such localization suggests thatthe function of the mucin-like glycoprotein may be to



facilitate tear film spread, providing an interface withthe mucus layer. The availability of probes to humanocular surface epithelial miicins will allow study ofocular surface disorders in which tear film spread isdisrupted because of various dry eye diseases.

Ph summary, we have developed a monoclonal an-tibody-designated HI85 that recognizes a highly glyco-sylated O-linked glycoprotein produced by the stra-tified epithelium of the cornea and conjunctiva. Theantigen in the human appears to be an analog of arecently described mucin-like glycoprotein discoveredin rat ocular surface epithelium.

,We conclude that the entire ocular surface epithe-lium produces mucin, and we hypothesize that theepithelial mucins facilitate tear film spread and stabili-zation.

Key Words

ocular surface, monoclonal antibody, tear film spread, mucus

References

1. Holly FJ, Lemp MA. Tear physiology and dry eyes.Sum Ophthalmol. 1977;22:69-87.

2. Nichols B, Dawson CR, Togni B. Surface features ofthe conjunctiva and cornea. Invest Ophthalmol Vis Sci.1983;24:570-576.

3. Dilly PN. On the nature and the role of the subsurfacevesicles in the outer epithelial cells of the conjunctiva.BrJ Ophthalmol. 1985; 69:477-481.

4. Wells PA, Hazlett LD. Complex carbohydrates at theocular surface of the mouse: An ultrastructural andcytochemical analysis. Exp Eye Res. 1984;39:19-35.

5. Greiner JV, Weidman TA, Korb DR, Allarisrriith MR.Histochemical analysis of secretory vesicles in nongob-let conjunctival epithelial cells. Ada OphthalmoL1985;63:89-92.

6. Gipson IK, Riddle CV, Kiorpes TC, Spurr SJ. Lectinbinding to cell surfaces: Comparisons between nor-mal and migrating corneal epithelium. Dev Biol.1983;96:337-345.

7. Panjwani N, Moulton P, AlroyJ, Baum J. Localizationof lectin binding sites in human, cat and rabbit cor-neas. Invest Ophthalmol Vis Sci. 1986;27:1280-1284.

8. Hazlett LD, Mathieu P. Glycoconjugates on cornealepithelial surface. Effect of neuraminidase treatment.JHistochem Cytochem. 1989;37:1215-1224.

9. Gipson IK, Yankauckas M, Spurr-Michaud SJ, TisdaleAS, Rinehart W. Characteristics of a glycoprotein inthe ocular surface glycocalyx. Invest Ophthalmol Vis Sci.1992;33:218-227.

10. Gipson IK Evidence that the entire ocular surfaceepithelium produces mucins for the tear film. ARVOAbstracts. Invest Ophthalmol Vis Sci. 1994; 35:1811.

11. Watanabe H, Gipson IK. Detection of blood groupdifferences in human corneal epithelium using amonoclonal antibody and lectins. Arch Ophthalmol.1994; 112:667-673.

12. Goding JW. Purification, fragmentation and isotopic

labelling of monoclonal antibodies. In: Monoclonal An-tibodies: Principles and Practice. 2nd ed. Orlando: Aca-demic Press; 1986:108-109.

13. Lindberg K, Brown ME, Chaves HV, Kenyon KR,Rheinwald JG. In vitro propagation of human ocularsurface epithelial cells for transplantation. Invest Oph-thalmol Vis Sci. 1993; 34:2672-2679.

14. Huff JC, Weston WL, Wanda KD. Enhancement ofspecific immunofluorescent findings with use of apara-phenylenediamine mounting buffer. / Invest Der-matol. 1982; 78:449-450.

15. Laemmli UK. Cleavage of structural proteins duringthe assembly of the head of bacteriophage TR. Nature(Land). 1970; 227:680-685.

16. Segrest JP, Jackson RL. Molecular weight determina-tion of glycoproteins by polyacrylamide gel electro-phoresis in sodium dodecyl sulfate. Methods Enzymol.1972;28(B):54-63.

17. Jay GD, Culp DJ, Jahnke R. Silver staining of exten-sively glycosylated proteins on sodium dodecyl sulfate-polyacrylamide gels: Enhancement by carbohydrate-binding dyes. Anal Biochem. 1990; 185:324-330.

18. Woodward MP, Young WWJr, Bloodgood RA. Detec-tion of monoclonal antibodies specific for carbohy-drate epitopes using periodate oxidation. / ImmunolMethods. 1985; 78:143-153.

19. Gum JR, Hicks JW, Swallow DM, et al. Molecular clon-ing of human intestinal mucus cDNAs. J Biol Chem..1989; 265:6480-6487.

20. Strous GJ, Dekker J. Mucin-type glycoproteins. Crit RevBiochem Mol Biol. 1992; 27:57-92.

21. Bobek LA, Tsai H, Biesbrock AR, Levine MJ. Molecu-lar cloning, sequence, and specificity of expression ofthe gene encoding the low molecular weight humansalivary mucin (M\JC1). J Biol Chem. 1993; 268:20563-20569.

22. Pemberton L, Taylor-Papadimitriou J, Gendler SJ. An-tibodies to the cytoplasmic domain of the MUC1 mu-cin show conservation throughout mammals. BiochemBiophysRes Commun. 1992; 185:167-175.

23. Gendler SJ, Lancaster CA, Taylor-Papadimitriou J, etal. Molecular cloning and expression of human tu-mor-associated polymorphic epithelial mucin. J BiolChem. 1990; 265:15286-15293.

24. Greiner JV, Kenyon KR, Henriquez AS, Korb DR,Weidman TA, Allansmith MR. Mucus secretory vesi-cles in conjunctival epithelial cells of wearers of con-tact lenses. Arch Ophthalmol. 1980;98:1843-1846.

25. Dilly PN, Mackie IA. Surface changes in the anestheticconjunctiva in man, with special reference to the pro-duction of mucus from a non-goblet-cell source: BrJOphthalmol. 1981;65:833-842.

26. Kessler TL, Dartt DA. Neural stimulation of conjuncti-val goblet cell mucous secretion in rats. In: SullivanDA, ed. The Lacrimal Gland, Tear Film and Dry Eye Syn-dromes: Basic Science and Clinical Relevance: New York:Plenum; 1994:393-398.

27. Nichols BA, Chiappino ML, Dawson CR. Demonstra-tion of the mucous layer of the tear film by electronmicroscopy. Invest Ophthalmol Vis Sci. 1985; 26:464-473.


human corneal and conjunctival epithelia produce a mucin-like

Documents