MUC16 Mucin Is Expressed by the Human Ocular Surface Epithelia and Carries the H185 Carbohydrate Epitope

Abstract

purpose. H185 antibody has been shown to recognize a carbohydrate epitope on a membrane-associated mucin in the apical surfaces of the corneal and conjunctival epithelia. The distribution of this antibody is altered on the surfaces of conjunctival epithelial cells of dry eye patients. The purpose of this work was to determine whether the H185 antibody recognizes the recently cloned membrane-associated mucin MUC16 (formerly CA125 antigen).

methods. To determine whether ocular surface epithelia express MUC16, the relative expression of the MUC16 mucin gene was determined by real-time PCR on reverse transcription products from RNA isolated from human corneal and conjunctival tissues, as well as from immortalized human corneal-limbal epithelial cell (HCLE) cultures. To determine the distribution of MUC16 mRNA and protein in the ocular surface epithelia, in situ hybridization and immunohistochemistry were performed on sections of corneal and conjunctival epithelia using, respectively, a MUC16 antisense oligoprobe and the antibodies OC125, VK-8, and R16 raised against the MUC16 mucin. Determination of whether MUC1 and MUC16 mucins carry the H185 carbohydrate epitope was achieved with the respective mucins isolated from HCLE protein extracts, using one- or two-step immunoprecipitation assays and immunodepletion experiments followed by Western blot analysis.

results. MUC16 mucin transcripts were detected in the human ocular surface epithelia and in corneal cell cultures. MUC16 mRNA and protein localized to the apical cell layers of the cornea and to the suprabasal region of the conjunctival epithelium. In HCLE cultures, MUC16 protein was detected in apical cells of islands of stratified cells. Immunofluorescence microscopy demonstrated exact colocalization of the MUC16 mucin and the H185 carbohydrate epitope in sections of human corneal tissue. Immunoprecipitated MUC16 mucin was recognized by the H185 antibody and vice versa, indicating that MUC16 mucin carries the H185 epitope. Immunodepletion with H185 antibody resulted in no OC125 antibody reactivity. No cross-reactivity between immunoprecipitated MUC1 and the H185 antibody was observed.

conclusions. This study demonstrates that the membrane-associated mucin MUC16 is expressed by the human ocular surface epithelia and that MUC16 carries the H185 carbohydrate epitope. Future studies on the expression of MUC16 and the characterization of the molecular structure of the H185 carbohydrate epitope will determine their biological significance on the healthy ocular surface and in dry eye syndrome.

The wet-surfaced epithelia, including the corneal and conjunctival epithelia, produce a group of highly glycosylated protective glycoproteins termed mucins. ^{1 2} On the ocular surface, mucins lubricate the apical surfaces of the epithelium during the eyelid blink, provide a barrier against pathogen invasion, and, due to their hydrophilic nature, prevent the desiccation of the preocular tear film. ^{3 4} Structurally, mucins are high molecular weight glycoproteins that contain a variable number of tandem-repeat domains rich in proline, serine and threonine. N-acetylgalactosamine is O-linked to the serine and threonine residues and serves to anchor a variety of carbohydrates, which ultimately constitute up to 80% of the mass of the mature mucin molecule, as reviewed by Gendler and Spicer. ¹  

Based on their protein structure, two types of mucins, secreted and membrane-associated, have been identified. Secreted mucins are synthesized and secreted onto surfaces of epithelia by goblet cells. Five of these mucins have been cloned. Four are large gel-forming mucins, MUC2, -5AC, -5B, and -6, and one is a small monomeric mucin MUC7, as reviewed by Moniaux et al. ⁵ Membrane-associated mucins are synthesized by the simple epithelia of the respiratory, gastrointestinal, and reproductive tracts, as well as by the wet-surfaced stratified epithelia of these systems and also those of the cornea and conjunctiva. To date, eight membrane-associated mucins (MUC1, -3A, -3B, -4, -12, -13, -16, and -17) have been identified. ^{5 6 7 8} They have short cytoplasmic domains with the major heavily glycosylated region extending away from the cell surface into the glycocalyx. Although the function of membrane-associated mucins is not completely clear, evidence suggests that because of their large extended heavily glycosylated conformation, they have antiadhesive properties, provide a protective barrier for the cell membrane, and prevent cell–cell and cell–protein interactions. ^{9 10 11}  

Several years ago, our laboratory produced a monoclonal antibody, designated H185, that recognized a carbohydrate epitope present on a mucinlike glycoprotein expressed on apical cells of corneal and conjunctival epithelia. ¹² The H185 antigen was subsequently purified and found to have characteristics consistent with a membrane-associated mucin. ² Numerous attempts to clone and characterize the molecule have not been successful, in part because of insufficient starting material and its heavy glycosylation. Efforts to characterize the H185 antigen have continued, however, especially because it has an altered distribution on apical cells of conjunctival epithelia of patients with dry eye. ¹³ Use of impression cytology in conjunction with immunohistochemistry and immunoelectron microscopy demonstrated that the H185 antibody did not bind to squamous cells on the conjunctival surface of eyes in patients with non-Sjögren’s dry eye. This indicates an alteration in the glycosylation of the H185 mucin or the absence of expression of the mucin protein itself. 

CA125 antigen is a tumor cell marker, antibodies to which have been widely used to detect ovarian cancer in sera of patients. ^{14 15} A 20-year effort to characterize the antigen was recently successful, with its designation as a newly identified membrane-associated mucin, MUC16. ^{7 16 17} The initial reported sequence of MUC16 includes a 5797-bp sequence from the 3′ end of the gene. ⁷ A second report describes the entire coding sequence and partial genomic structure of the mucin. ¹⁷ This mucin is unusual among membrane-associated mucins, in that it contains long, partially conserved tandem repeat units (156 amino acids), has a high content of leucine, and does not have the epidermal growth factor (EGF)–like domain that is commonly found in the C-terminal non–tandem-repeat region of other membrane-associated mucins. The MUC16 gene has been localized to the region 13.3 of the short arm of chromosome 19. ⁷  

We have suggested that the H185 antigen at the ocular surface is not the membrane-associated mucin MUC1 and that it could be a different membrane-associated mucin. ² In studies to determine the presence of newly described mucins in ocular surface epithelium, we found that MUC16 message is present and that the antibody to CA125 antigen binds the ocular surface epithelia in a pattern identical with that of the H185 antibody. We thus sought to determine whether H185 antigen is carried on the mucin MUC16. Herein, we report the results of studies that indicate that the H185 carbohydrate epitope is present on MUC16. 

Methods

Tissue and Cell Lines

The study was conducted in compliance with Good Clinical Practices, Institutional Review Board regulations, informed consent regulations, and the provisions of the Declaration of Helsinki. Corneal tissue from cadavers was obtained from the Lions Eye Bank. Conjunctival biopsy samples were taken from the bulbar conjunctiva of patients who were undergoing cataract surgery. Conjunctival impression cytology samples were collected from healthy individuals as previously described. ¹⁸ Human corneal and conjunctival tissues were flash frozen for isolation of RNA and frozen in optimal cutting temperature (OCT) compound (Miles; Elkhart, IN) for immunohistochemistry. 

Primary cultures of human corneal epithelial cells derived from limbal epithelium of discarded rims of donor corneas used for transplant were cultured as described by Lindberg et al. ¹⁹ Because of the limited replicative life span of the primary cultures of human corneal epithelia, immortalized corneal epithelial cells (designated HCLE) were also used. ^{20 21} Briefly, HCLE cells derived from primary cultures of human corneal-limbal epithelial cells were immortalized by transductions that abrogate p16 and p53 function in the cell cycle, followed by transduction with the catalytic subunit of the human telomerase (hTERT). ²⁰ HCLE cultures were grown in a medium optimized for proliferation of keratinocytes (keratinocyte serum-free medium [K-sfm]; Gibco-Invitrogen, Carlsbad, CA) and switched at approximately half-confluence to a 1:1 mixture of K-sfm and low calcium DMEM/F12 (Gibco-Invitrogen) for 24 hours to achieve confluence. After reaching confluence, cells were switched to DMEM supplemented with 10% calf serum and 10 ng/mL EGF for 7 days, which promotes stratification of corneal cells and differentiation. ²¹  

RNA Isolation and cDNA Synthesis

Total RNA was extracted from tissues and confluent HCLE cells using extraction reagent (TRIzol; Gibco) according to the manufacturer’s protocol. Residual genomic DNA in the RNA preparation was eliminated by digestion with amplification-grade DNase I (Invitrogen). Reverse transcription of 1 microgram of total RNA was performed with random hexamer primers and reverse transcriptase (SuperScript II; Gibco), according to the manufacturer’s protocol, as previously described. ¹⁸  

Conventional PCR and Real-Time PCR

The MUC16 primers and probe used for mucin gene amplification and detection, respectively, were designed with the assistance of a computer program (Primer Express software; Applied Biosystems; Foster City, CA). Conventional RT-PCR experiments were performed to confirm that only a single band would be obtained when amplifying corneal and conjunctival cDNA with the MUC16 primers used in this study. A 114-bp fragment containing a MUC16 region flanking the tandem-repeat domain was generated by using the forward 5′-GCCTCTACCTTAACGGTTACAATGAA-3′ and reverse 5′-GGTACCCCATGGCTGTTGTG-3′ primers. BLASTN searches against nucleotide databases were performed to confirm the sequence specificity of the chosen nucleotide sequences (http://www.ncbi.nlm.nih.gov/blast/; provided in the public domain by the National Center for Biotechnology Information; Bethesda, MD). PCR reactions were performed in a thermal cycler (TouchDown; Hybaid, Middlesex, UK) using Taq DNA polymerase (AmpliTaq Gold; Applied Biosystems) and the following parameters: 10 minutes at 95°C, followed by 35 cycles of 15 seconds at 95°C and 1 minute at 60°C. The MUC16 PCR product was purified and sequenced at the DNA Sequencing Core Facility of Massachusetts General Hospital (Boston, MA) to verify its identity. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an endogenous reference to determine the integrity of the mRNA in each sample. 

Detection of MUC16 gene expression in multiple samples was performed by real-time PCR in the presence of the MUC16 primers described earlier and a double-labeled fluorogenic MUC16 probe (5′-AGATGAGCCTCCTACAACTCCCAAGCCAG-3′). Amplification was performed in triplicate with 0.8 μL of cDNA in a total volume of 50 μL (TaqMan chemistry; Applied Biosystems). Assays were performed using ABI Prism 5700 Sequence Detection System (Applied Biosystems). The average threshold cycle (C_T) values for GAPDH were used as an internal calibrator to correct for differences in the integrity and amount of total RNA added to each reaction. ¹⁸ To validate the relative quantitation, the efficiency of the target gene amplification was compared with the efficiency of the GAPDH amplification, as described in the manufacturer’s protocol (Applied Biosystems). The comparative C_T method was used for relative quantitation of the number of MUC16 mucin transcripts in corneal and conjunctival tissue, as well as in corneal cell cultures—selecting the relative mucin mRNA level in corneal tissue as the calibrator. 

Fluorescence In Situ Hybridization

Human cornea and conjunctiva were fixed in RNase-free 4% paraformaldehyde and embedded in paraffin for fluorescence in situ hybridization (FISH) of mucin mRNA, essentially as described. ²² Briefly, FISH was performed using digoxigenin (DIG)-labeled antisense and sense 42-mer oligoprobes to the tandem-repeat region of human MUC16 designed from GenBank Accession number AF361486 (http://www.ncbi.nlm.nih.gov/genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). ⁷ The MUC16 antisense probe used was 5′-GTTGGTGATGGTAAAGTTGAGGGTGAATGGTATCAAGAGAGG-3′ and the sense probe was 5′-CCTCTCTTGATACCATTCACCCTCAACTTTACCATCACCAAC-3′. The oligoprobes were labeled by tailing with the DIG Oligonucleotide Tailing Kit following the protocol (Roche Applied Science, Indianapolis, IN). Eight-micrometer sections of tissue were deparaffinized, rehydrated through an alcohol series, and successively washed in PBS, PBS with 100 mM glycine, PBS with 0.3% Triton X-100, and PBS. Sections were then treated with 10 μg/μL Proteinase K, postfixed in 4% paraformaldehyde, washed with PBS and acetylated before hybridization. Hybridization was performed at 37°C in a solution containing 2× SSC, 1× Denhardt’s solution, 10% dextran sulfate, 50 mM phosphate buffer (pH 7.0), 50 mM dithiothreitol, 250 μg/mL tRNA, 100 μg/mL polyadenylic acid, 500 μg/mL denatured salmon sperm DNA, and 47% deionized formamide. After hybridization, sections were sequentially washed at 37°C in 2×, 1×, and 0.25× SSC, successively, to remove nonspecific binding of probes. Sections were then rinsed with 100 mM Tris-buffered saline (pH 7.5) and blocked with 1% normal sheep serum and 0.1% Triton X-100 before incubation with 20 μg/mL rhodamine-conjugated anti-digoxigenin antibody (Roche Applied Science) to disclose hybridized oligoprobes. Sections were coverslipped with antifade mounting medium containing DAPI (Vectashield; Vector Laboratories, Burlingame, CA) to visualize cell nuclei. 

Antibodies

The monoclonal mouse anti-human H185 antibody was obtained from culture supernatants of hybridoma cells, as described previously. ¹² The monoclonal mouse anti-human MUC16 antibody, OC125, was purchased from Dako Corp. (Carpinteria, CA). Two additional antibodies against the MUC16 mucin (monoclonal mouse VK-8, and polyclonal rabbit R16) ^{7 16 23} were kindly provided by Kenneth O. Lloyd (Sloan-Kettering Institute, New York, NY). Anti-CA125 antibodies are mainly against the repeated peptide sequences, although carbohydrate may influence their reactivity. ^{7 16 17} Evidence indicating that a peptidic sequence is required for recognition of these antibodies include: (1) digestion of recombinant MUC16 with endoprotease Lys-C or the protease Asp-N destroyed epitope recognition by CA125 (OC125) antibodies ¹⁷ —a cysteine-enclosed loop within the tandem-repeat domain of MUC16 (5′ to the heavily O-glycosylated region within the repeat) is the proposed binding site for OC125 ¹⁷ ; (2) the rabbit antiserum R16 was generated from VK-8 affinity-purified CA125 and was used to identify proteins in an Escherichia coli cDNA expression library from OVCAR-3 cells which led to the cloning of the MUC16 gene ⁷ ; and (3) transfection of CA125-negative cell lines with a construct containing three tandem-repeat regions and the non–tandem-repeat region of MUC16 resulted in positive binding of the OC125, R16, and VK-8 antibodies compared with cells transfected with the empty vector, which did not bind the antibodies. ¹⁶ Monoclonal antibodies against the membrane-associated MUC1, designated HMFG-1 and -2, were purchased from Biodesign International (Saco, ME). HMFG-1 and -2 are antibodies widely used to detect MUC1. They are directed to two different epitopes found in the tandem-repeat domain of MUC1 (formerly polymorphic epithelial mucin). ^{24 25 26} Proof of specific binding of HMFG-1 and -2 to MUC1 is as follows: (1) The antibodies react with a synthetic peptide with an amino acid sequence corresponding to that predicted by the tandem repeat of MUC1 ²⁶ ; (2) COS cells transfected with a plasmid containing the full-length cDNA of MUC1 bind the monoclonal antibodies HMFG-1 and -2 ²⁷ ; and, (3) the expression of MUC1 mRNA correlates with HMFG-1 and -2 binding in seven cell lines assayed. ²⁸ Anti-mouse IgG and anti-rabbit IgG secondary antibodies conjugated to fluorescein isothiocyanate (FITC), tetramethylrhodamine isothiocyanate (TRITC), or peroxidase were purchased from Jackson ImmunoResearch (West Grove, PA) or Sigma (St. Louis, MO). 

Immunofluorescence

Immunohistochemistry was performed as previously described ¹² with H185 (1:10 dilution), OC125 (1:50), R16 (1:100), or VK-8 (1:10) used as primary antibodies. To determine whether H185 antigen colocalizes with MUC16, double-labeling studies were conducted. A mixture of H185 antibody and R16 antibody was applied to the sections, followed by a mixture of TRITC-conjugated donkey anti-mouse IgG and FITC-conjugated donkey anti-rabbit IgG. Coverslips were applied to the slides using antifade mounting medium with DAPI (Vectashield; Vector Laboratories) to visualize the nuclei of the cells. Incubation with the primary antibody was omitted in control experiments. 

Immunoprecipitation

The H185, MUC1, and MUC16 mucins were purified from protein extracts of either primary or immortalized corneal cell cultures by immunoprecipitation techniques, as detailed later, using the H185, HMFG-1, HMFG-2, and OC125 monoclonal antibodies, respectively. Protein extracts were obtained from cell cultures in T75 flasks by scraping the cells from the flask after addition of 1 mL of 2% SDS in the presence of a protease inhibitor cocktail (Complete Mini; Roche Applied Science) or phenylmethylsulfonyl fluoride (PMSF). The scraped cells were then homogenized with a ground glass homogenizer, and centrifuged at 35,016g (relative centrifugal force; RCF) for 45 minutes. The resulting supernatant was recovered, and the pellet was homogenized a second time and centrifuged as just described to increase the yield of extracted mucin. After centrifugation, supernatants were pooled and protein concentration determined using a bicinchoninic acid (BCA) protein assay (Pierce, Rockford, IL), according to the manufacturer’s recommendations, with bovine serum albumin used as the standard. 

The MUC16 mucin was immunoprecipitated by using the OC125 antibody, as previously described. ²⁹ For immunoprecipitation of H185 mucin and MUC1, their respective antibodies, H185, and HMFG-1 or -2, were incubated with anti-mouse IgG agarose particles (Sigma) for 1 hour at 4°C. One hundred micrograms of protein from the extract was diluted in 500 μL of mucin isolation buffer (0.1 M NH₄HCO₃, 2.0 mM PMSF, 0.5 M NaCl, 5 mM EDTA, 2 mM N-ethylmaleimide, and 0.02% NaN₃ and protease inhibitor cocktail [Complete Mini; Roche Applied Science]). Dilution of the protein extract in mucin isolation buffer without surfactants is required to reduce the concentration of SDS, which interferes with binding of H185 antibody to its carbohydrate epitope. The protein extract in mucin isolation buffer was added to the conjugated beads and incubated for 2 hours on a rocker at room temperature. After three washes with a buffer containing 10 mM Tris-HCl, 2 mM EDTA, 0.1% Triton X-100, 0.1% SDS (pH 7.4), the beads were boiled with 30 μL of 2× Laemmli buffer ³⁰ and centrifuged, and the supernatant was loaded on a gel for SDS-PAGE and Western blot as described later. 

For serial immunoprecipitation experiments, the H185 mucin was immunoprecipitated as just described and eluted from the agarose beads, with a solution containing 4 M urea in Tris-buffered saline (TBS; pH 7.5), and protease inhibitor cocktail. After centrifugation, salts in the supernatant were removed using separation tubes with 10-kDa molecular weight cutoff membranes (Nanosep 10 K, Omega; Gelman Laboratory, Ann Arbor, MI). The H185 mucin isolate was dissolved and immunoprecipitated once again, using the OC125 antibody as previously described. ²⁹ As a control, protein extracts were treated with anti-mouse IgG agarose particles in the absence of the primary immunoglobulins. 

In immunodepletion experiments, the H185 antigen and the MUC16 mucins were removed from protein extracts by immunoprecipitation techniques involving the H185 and OC125 antibodies, respectively, as described earlier. After immunoprecipitation, the immunodepleted protein extract was concentrated using separation tubes (10-kDa cutoff; Nanosep; Gelman Laboratory) and boiled with 10 μL of 2× Laemmli buffer, and the supernatant was loaded on a gel for SDS-PAGE and Western blot. 

Electrophoresis and Western Blot

Mucin in protein extracts and immunoprecipitated material was separated under reducing conditions on 6% separating, 4% stacking SDS-polyacrylamide gels and blotted onto nitrocellulose membranes. ¹⁸ After blotting, membranes were blocked with 10% normal horse serum in TBS (pH 7.5), for 30 minutes and incubated with the primary antibody (undiluted for H185; 1:10 dilution for OC125, R16, VK-8, and HMFG-1 and -2) for 1 hour at room temperature. After a wash in TBS, the membranes were incubated with the appropriate peroxidase-conjugated secondary antibody. Positive binding was detected colorimetrically with diaminobenzidine peroxidase substrate (Bio-Rad Laboratories, Hercules, CA). Prestained molecular weight markers (Precision Protein Standards) were purchased from Bio-Rad. 

Results

MUC16 Mucin mRNA Expression in Ocular Surface Epithelia

The expression of MUC16 mucin gene transcripts in corneal and conjunctival epithelia was initially determined by conventional RT-PCR (Fig. 1A) . Amplification of cDNA products from (lane 1) a normal corneal button, (lane 2) immortalized HCLE cells, (lane 3) conjunctival tissue, and (lane 4) conjunctival epithelium collected by impression cytology in healthy donors produced a single 114-bp band corresponding to the predicted product size (Fig. 1A) . The amplification of the housekeeping GAPDH gene, which was performed with all samples, demonstrated the quality and the amount of starting mRNA. 

The use of real-time PCR in a larger number of samples demonstrated that MUC16 mucin transcripts were consistently present (Fig. 1B) . A higher level of MUC16 mucin mRNA was observed in conjunctival samples obtained from impression cytology specimens compared with conjunctival biopsy specimens, presumably because the biopsy sample is from a mixture of cell types, not just epithelial cells. In addition, the presence of mostly apical epithelial cells in the impression cytology samples, which are known to produce higher levels of other membrane-associated mucins compared with the basal epithelial cells, may lead to the higher number of mucin transcripts. The nontemplate control, included in all the experiments performed with real-time PCR, confirmed the absence of DNA contamination of the reagents used for amplification. Furthermore, DNase treatment of the RNA before reverse transcription assured that there was no DNA contamination in the RNA used for the assays. 

Distribution of MUC16 Mucin mRNA and Protein in Human Ocular Surface Epithelia

The tissue distribution of MUC16 mRNA in human ocular surface epithelia was determined using FISH, with a digoxigenin-labeled oligonucleotide probe corresponding to the tandem-repeat region of the mucin. In cornea, MUC16 mRNA was primarily detected within the flattened cells of the apical cell layer of the epithelium and occasionally in the supranuclear region of the basal cells (Fig. 2A) . In conjunctiva, the MUC16 mRNA localized throughout all cell layers of the epithelium, but apical cell binding of the probe was the most intense (Fig. 2B) . The sense sequence was run as the negative control in cornea and conjunctiva and showed no binding (Figs. 2C 2D , respectively). 

Immunolocalization of the MUC16 mucin in corneal and conjunctival epithelia was performed with three different antibodies specific to the MUC16 molecule: OC125, VK-8, and R16. ^{7 16 23} These antibodies detect protein epitopes on the mucin (see the Methods section). The MUC16 mucin was found along the entire human ocular surface epithelium, including that of the cornea and conjunctiva, as well as on HCLE cells (Fig. 3) . As shown in Figures 3A 3B 3C , the OC125, VK-8, and R16 antibodies had the same tissue distribution and recognized specifically the apical and subapical flattened cell layers of the human corneal epithelium. In conjunctival tissue, OC125 bound along the apical surface of the epithelium and along the cell membranes of the suprabasal cells (Fig. 3F) . Results of the immunohistochemical studies in the ocular surface epithelium paralleled and corroborated the in situ hybridization data, except that the OC125, VK-8, or R16 antibodies did not bind to basal cells of either cornea or conjunctival epithelia. It is possible that some basal cells transcribe but do not translate the MUC16 mucin. Whereas binding of the anti-MUC16 antibodies in cornea and conjunctiva was particularly prominent in the cytoplasmic membranes of the squamous epithelial cells, some but not all the goblet cells in the conjunctival sections also bound the anti-MUC16 (OC125 and R16) antibodies. The binding in the goblet cells was to the mucin packets (Fig. 3F , inset). In 7-day confluent cultures of HCLE cells, the OC125 antibody bound to a distinct population of cells. Cells adjacent to the substrate did not bind the OC125, whereas binding was observed in apical cells of scattered islands of stratified cells (Fig. 3G) . 

Colocalization of MUC16 Mucin with the H185 Epitope

The pattern of binding of the anti-MUC16 antibodies to the stratified human corneal and conjunctival epithelia and cell cultures was similar to that previously reported for the H185 monoclonal antibody. ¹² Because the R16 antibody is a rabbit polyclonal antibody to MUC16 mucin and because the H185 antibody is a mouse monoclonal antibody, it was feasible to determine whether the two antibodies colocalize in the same sections of ocular surface epithelia. In these experiments, bleed-through between the different fluorophores was evaluated and found to be negative. As demonstrated in Figure 4 , antibody to the membrane-associated MUC16 exactly colocalized with the H185 antibody in sections of corneal epithelium. Both antibodies bound to the apical squamous cells of the corneal epithelium when mixtures of the two primary antibodies and then the secondary antibodies were applied to the sections. Similar results were obtained in the stratified conjunctival epithelium, where both the H185 and R16 antibodies colocalized within the suprabasal cells of the epithelium (data not shown). In goblet cells, however, some bound the R16 antibody but not the H185 antibody and vice versa, and we found some in which the two antibodies colocalized (data not shown). Previous reports show heterogeneity in mucin O-glycosylation ^{12 31} and expression of membrane-associated mucins in conjunctival goblet cells (i.e., MUC4 ³² ). It is possible to interpret the goblet cell data in several ways, perhaps MUC5AC carries the H185 epitope in native tissue or the R16 antibody recognizes differently glycosylated MUC16. 

Demonstration that MUC16 Protein in Corneal Epithelial Cells Carries the H185 Epitope

The presence of the MUC16 mucin in HCLE cells was demonstrated by immunoprecipitation experiments in which the CA125 antigen was immunoprecipitated with OC125 (capture antibody) and the material analyzed by immunoblot using the OC125 as detecting antibody (Fig. 5) . A high-molecular-weight band (>250 kDa) corresponding to the MUC16 mucin was detected in the upper region of the separating gel (Fig. 5A , lane 1). A band of similar electrophoretic mobility was obtained when analyzing the CA125 antigen in ovarian cancer cell lines. ^{16 23}  

Experiments using immunoprecipitation methods demonstrated that the H185 carbohydrate epitope is carried by the MUC16 mucin in HCLE cell lysates. The H185 and OC125 antibodies each recognized the immunoprecipitated CA125 and H185 antigens, respectively (Fig. 5A , lanes 2 and 4). Cross-reaction was confirmed by two-step immunoprecipitation experiments, in which the immunoprecipitate obtained with the first capture antibody (H185) was extracted from the agarose beads, then immunoprecipitated with a second capture antibody (OC125) and analyzed by immunoblot with the OC125 antibody (Fig. 5A , lane 5). When the H185 antigen was removed from corneal cell protein extracts H185 (control) and OC125 did not bind in immunodepletion experiments (Fig. 5B , lane 2), confirming that MUC16 mucin carries the H185 carbohydrate epitope. When the MUC16 mucin was removed using the OC125 antibody, the results were no OC125 reactivity (control) to corneal cell protein extracts and weak binding of the H185 antibody (Fig. 5B , lane 4), suggesting that either other mucin(s) in HCLE cells may carry the H185 epitope or the presence of a highly glycosylated MUC16 glycoform that cannot be immunoprecipitated with the MUC16 antibody. Additional results identifying the H185 antigen on the mucin MUC16 were obtained using the VK-8 and R16 antibodies as detection antibodies after H185 mucin immunoprecipitation (Fig. 6) . 

As demonstrated by real-time PCR, immunohistochemistry, and immunoblot analysis, the HCLE cells also produce the membrane-associated mucin MUC1. ²¹ Therefore, we also used immunoprecipitation techniques to test the possibility that MUC1 carries an H185 epitope. Immunoprecipitated MUC1 mucin (by the HMFG-1 antibody) was not recognized by the H185 antibody and vice versa (Fig. 7 , lanes 2 and 3), demonstrating that the H185 carbohydrate epitope is not carried by the MUC1 mucin in HCLE cells. In a similar experiment with HMFG-2 used as the capture antibody, immunoprecipitated MUC1 was not recognized by the H185 antibody (Fig. 7 , lane 6) supporting the previous experiment with HMFG-1 and indicating that there is no identity between H185 and MUC1 antigens. 

Nonspecific binding of mucin from the cell culture lysates to the control agarose beads used for immunoprecipitation was tested and found to be negative. When present, MUC16, MUC1, and H185 antigens were detected in the upper region of the separating gel although sporadically, some material that did not enter the stacking gel was found at the top of the wells (data not shown). 

Discussion

This study demonstrates the presence of the recently cloned membrane-associated mucin MUC16 (formerly named CA125 antigen) in the stratified squamous epithelia of the human ocular surface and identifies MUC16 as a mucin carrying the previously described H185 carbohydrate antigen. Two major types of data indicate that H185 antigen is present on the MUC16 mucin. First, in immunoprecipitation experiments, the isolated H185 antigen from cultures of corneal epithelium reacted with the OC125 monoclonal antibody as well as two additional antibodies to the MUC16 mucin, and MUC16 antigen isolated from the same cultures was recognized by H185 antibody. Second, CA125 and H185 antigens exactly colocalized in apical cells of the corneal epithelium, indicating identical cellular distribution. These coimmunoprecipitation and colocalization studies conclusively demonstrate that the H185 antigen, an epitope with expression that is altered in the ocular surface of patients with dry eye symptoms, is carried by the membrane-associated mucin MUC16. 

A major significance of the identification of the H185 antigen on MUC16 centers around the possibility of now answering the question regarding whether the alteration in H185 antibody distribution in non-Sjögren’s dry eye ¹³ is a result of altered mucin gene expression or altered mucin glycosylation. Quantitation of MUC16 mRNA from conjunctival epithelium derived by filter paper stripping (impression cytology) from patients can be achieved by the same real-time PCR methods used in quantitation of MUC5AC message in patients with Sjögren’s dry eye. ¹⁸ In addition, methods for quantitation of MUC16 protein can be developed. If there is no alteration in MUC16 expression or protein levels in non-Sjögren’s dry eye, we can assume that the glycosylation of the mucin is altered, because H185 antibody recognizes an O-linked carbohydrate epitope ¹² on MUC16. If the altered distribution of H185 antibody binding in patients with non-Sjögren’s syndrome dry eye is due to altered glycosylation of MUC16, expression of the enzymes that add sugars to the mucins may be altered in the disease. In fact, alterations of patterns of expression of GalNAc-transferases has been demonstrated in ocular cicatricial pemphigoid, which results in a dry-keratinized conjunctival epithelium. ³¹  

A second reason that identification of H185 epitope on MUC16 is significant is that with MUC16, a third membrane-associated mucin is demonstrated to be present in apical cells of the ocular surface epithelia. Do the three membrane-associated mucins—MUC1, MUC4, and MUC16—have unique functions, and if so, what are they? MUC16 does appear to be structurally unique from MUC1 and MUC4, and may therefore serve different functions on the ocular surface. The complete coding sequence of MUC16 (CA125 antigen) is now available. The 3′ cytoplasmic domain has a potential tyrosine phosphorylation motif, which may be involved in signaling, and the amino terminus, 5′ to the tandem-repeat domains, is very serine- and threonine-rich, indicating another region of heavy O-glycosylation. ^{16 17} Unlike MUC4, MUC16 does not have EGF-like domains but instead has SEA modules, which are characteristic of membrane-associated glycoproteins with high levels of O-linked carbohydrates. ³³ These modules are a four-amino-acid sequence believed to be susceptible to proteolytic cleavage. CA125 is obviously released from the cell surface, because it is used as a serum marker of ovarian cancer. H185 antigen has been detected in the tear film, which also suggests that MUC16 is cleaved from the apical cell membranes of the ocular surface epithelia (Argueso P, Spurr-Michaud SJ, Gipson IK, ARVO Abstract 351, 2000). The questions regarding the structure and function of membrane-spanning mucins on the ocular surface remain unanswered, but corneal and conjunctival cell lines that express all three mucins have been developed and will provide systems for study of their function(s). ²¹  

Because our data indicate that anti-CA125 and H185 antibodies bind the same molecule, comparison of their tissue distribution would be informative. The CA125 antigen, now known as MUC16, is a high-molecular-weight glycoprotein, the antibodies to which have been widely used in a serum assay to detect ovarian cancer. ^{14 15} The CA125 antigen was initially detected by the murine monoclonal antibody OC125 in ovarian carcinoma cell lines of epithelial origin and in the luminal surface of tumor tissues taken from patients with ovarian cancer. ³⁴ Although no reactivity was originally observed in a number of normal tissues, ³⁴ further studies revealed that the CA125 antigen is present in the tall columnar cells of the endocervical epithelium, endometrium, pleura, pericardium, and peritoneum, ³⁵ as well as in seminal plasma, ³⁶ milk secretions, ³⁷ and cervical mucus. ^{35 36 37} Of these tissues and secretions, H185 antigen is found in cervical mucus and endo- and ectocervical epithelia (Gipson IK, Spurr-Michaud SJ, Tisdale A, Argüeso P, unpublished results, 1996–2000), whereas the rest have not been assayed for H185 antibody binding. Binding of the CA125 antibody to the conjunctival epithelium has also been suggested, ³⁸ and indeed data from this study document the presence of MUC16 in conjunctival epithelium. Besides ocular surface epithelia, human vaginal and laryngeal epithelia ¹² show H185 antibody binding. Thus, the only normal tissues that have been assayed for both H185 antigen and MUC16 antigen to date include ocular surface and reproductive tract epithelia, where both antigens are present. Comparison of MUC16 mucin and H185 antibody localization will help to determine whether the H185 carbohydrate constitutes an epitope associated to MUC16 in other epithelia. 

Despite our data demonstrating that the carbohydrate epitope recognized by the H185 antibody is present on MUC16, the possibility remains that the two other membrane-associated mucins present on the ocular surface, MUC1 ³⁹ and MUC4, ^{40 41} as well as the gel-forming mucin MUC5AC in goblet cells, ⁴¹ carry the H185 antigen. Because immunoprecipitated MUC1 was not recognized by H185 antibody and vice versa, we ruled out MUC1 as carrying the H185 antigen. Immunolocalization studies with MUC4 antibodies ⁴⁰ indicate that the mucin has a very different localization within the corneal epithelium (binding throughout the entire stratified epithelium) compared with the H185 epitope (apical cell binding). This suggests that the H185 epitope is either not carried by MUC4 or that it is carried only on MUC4 that is present in apical cells of the cornea and in the suprabasal cells of the conjunctiva, not the MUC4 that is found throughout all the cells of cornea and conjunctiva. ⁴⁰ Because some conjunctival goblet cells bind the H185 antibody and because some goblet cells bind anti-MUC16 antibodies, there is the possibility that the H185 epitope is present on the goblet cell specific mucin MUC5AC and/or on membrane-associated MUC16 mucin produced by the goblet cells. Also, we cannot exclude the possibility that there may be additional, as yet unidentified, membrane-associated mucins expressed by the ocular surface epithelia that carry the H185 epitope. Cloning and sequencing of the H185 antigen is the most definitive proof of identity. Attempts by our group to clone the H185 antigen(s) have been limited by the complex structure and high degree of glycosylation associated with mucins and the limited amount of starting material. In the case of the cloning of CA125 antigen (MUC16), 31 L of supernatant medium from the NIH:OVCAR-3 tumor cell line was needed to purify sufficient mucin antigen. ¹⁶ Nevertheless, the coimmunoprecipitation and colocalization data from this study provide strong evidence indicating that the H185 antigen is present on the membrane-associated mucin MUC16. 

In conclusion, we have demonstrated that the recently cloned membrane-associated mucin MUC16 is expressed by human ocular surface epithelia and we have identified MUC16 as a carrier of the H185 carbohydrate epitope. Studies of the regulation of the expression of MUC16 and characterization of the repertoire of carbohydrates on MUC16 will contribute to the understanding of its function in normal and pathologic states. 

 Supported by CIBA Vision and National Eye Institute Grant No. R01-EY03306 (IKG).

 Submitted for publication August 22, 2002; revised November 22, 2002, and January 23, 2003; accepted January 31, 2003.

 Disclosure: P. Argüeso, None; S. Spurr-Michaud, None; C.L. Russo, None; A. Tisdale, None; I.K. Gipson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

MUC16 mRNA expression in human corneal and conjunctival epithelia and confluent HCLE cultures. (A) Conventional RT-PCR demonstrating that the MUC16 primers used for relative quantitation in real-time PCR produced the expected amplicon size in samples analyzed in (B). (B) Real-time PCR analysis of (lane 1) cornea (n = 1), (lane 2) HCLE cultures (n = 4), (lane 3) conjunctival biopsy specimens (n = 3), and (lane 4) conjunctival impression cytology samples (n = 11). Nontemplate control experiments (lane 5) showed absence of DNA contamination. Results are expressed on a logarithmic scale. Error bars, SEM.

Figure 2.

 

Distribution of MUC16 mRNA in human corneal and conjunctival epithelium as determined by fluorescence in situ hybridization. (A) In cornea, MUC16 mRNA was demonstrated in the most apical cells of the epithelium as well as in a few of the subapical cells (insert). Some apical cells in cornea sections were lost after proteinase K treatment, resulting in a discontinuous presence of apical cells in the section. (B) In conjunctiva, MUC16 mRNA was present in all layers of the epithelium, but the most intense binding was in the apical cells. No binding was observed when using a MUC16 sense oligonucleotide probe as negative control in cornea (C) or conjunctiva (D). Dashed lines: basement membrane region of the epithelia. Scale bar, 10 μm.

Figure 3.

 

Immunofluorescence micrographs demonstrating the presence of the membrane-associated mucin MUC16 in corneal and conjunctival epithelial cells. Binding of the monoclonal antibody OC125 (A) or VK-8 (B) or the polyclonal antibody R16 (C) raised against the MUC16 mucin was observed in apical cells of the corneal epithelium. There was no background binding associated with the secondary antibodies, anti-mouse IgG (D), and anti-rabbit IgG (E). Binding of the OC125 antibody to the suprabasal cells of the conjunctival epithelium is demonstrated in (F). Bottom inset: Intracellular binding of the OC125 antibody was observed in some goblet cells. Top inset: Phase-contrast image of the section in the bottom inset demonstrating that the cell shown in the bottom inset is a goblet cell. Binding of the OC125 antibody to scattered islands of cells in HCLE cultures is shown in (G). Propidium iodide was included in the mounting medium to localize the position of the nuclei of cells in culture (G, arrows). Dashed lines: basement membrane region of the epithelia. Scale bar, 25 μm.

Figure 4.

 

Immunohistochemical colocalization of MUC16 and H185 antibodies in the apical cells of the corneal epithelium. A mixture of a rabbit polyclonal antibody (R16 used as a marker for the membrane-associated mucin MUC16) and the mouse monoclonal H185 antibody was incubated on sections of cornea. The R16 antibody binding was detected with a FITC-conjugated secondary antibody (green), and the H185 antibody binding was detected using a TRITC-conjugated secondary antibody (red). The areas of colocalization of the R16 and H185 antibodies are shown in the merged image in yellow. DAPI was included in the mounting medium to localize the position of the nuclei in the section (blue). Scale bar, 25 μm.

Figure 5.

 

Western blot analysis of immunoprecipitated mucin demonstrating that the H185 epitope is on MUC16. (A) One-step immunoprecipitation of HCLE protein extracts using the MUC16-specific antibody OC125, followed by detection with OC125 (OC125/OC125 [capture Ab/detection Ab]) (lane 1), OC125/H185 (lane 2), H185/H185 (lane 3) and H185/OC125 (lane 4). Two-step immunoprecipitation using H185-OC125/OC125 (first capture Ab-second capture Ab/detection Ab) (lane 5). Demonstration of nonspecific binding of mucins in the sample to agarose beads used in two-step immunoprecipitation (lane 6). (B) Immunodepletion experiments demonstrating that removal of H185 epitope from corneal cell protein extracts resulted in absence of H185 (control) and OC125 antibody binding (lanes 1 and 2). Removal of MUC16 mucin results in absence of OC125 antibody binding (control) and weak binding of the H185 antibody (lanes 3 and 4). The low molecular weight bands (∼75 kDa) observed in (A) correspond to the primary antibodies used in the mucin immunoprecipitation assay and to secondary antibodies removed from agarose beads by the Laemmli buffer. Arrowheads: molecular weight of the mucins analyzed.

Figure 6.

 

Western blot analysis of H185-immunoprecipitated mucin with the R16 and VK-8 antibodies. Immunoprecipitation of HCLE protein extracts with the H185 antibody followed by detection with R16 (H185/R16, lane 1) and VK-8 (H185/VK-8, lane 2) and demonstration of nonspecific binding to beads (lane 3). Arrowhead: molecular weight of the mucins analyzed.

Figure 7.

 

Western blot analysis demonstrating no identity between H185 immunoprecipitate and MUC1. The HMFG-1 immunoprecipitate was recognized by HMFG-1 antibody (HMFG-1/HMFG-1, lane 1) but not by the H185 antibody (HMFG-1/H185, lane 2). The H185 immunoprecipitate was not recognized by the HMFG-1 antibody (lane 3). Lane 4: absence of nonspecific binding of mucin to beads. Immunoprecipitation of MUC1 with HMFG-2 resulted in positive binding of HMFG-1 (lane 5) and no reaction with the H185 antibody (lane 6). The low-molecular-weight bands (∼75–130 kDa) correspond to the primary antibodies used in the mucin immunoprecipitation assay and to secondary antibodies removed from agarose beads by the Laemmli buffer. Arrowheads: molecular weight of the mucins analyzed.

The authors thank C. Stephen Foster (Massachusetts Eye & Ear Infirmary, Boston, MA) for providing the conjunctival biopsy specimens.