Differential Regulation of Membrane-Associated Mucins in the Human Ocular Surface Epithelium

Abstract

purpose. Membrane-associated mucins present in the apical cells of the ocular surface epithelium (MUC1, -4, and -16) are believed to contribute to the maintenance of a hydrated and wet-surfaced epithelial phenotype. Serum and retinoic acid (RA) have been used to treat drying ocular surface diseases. The goal of this study was to determine whether serum or RA regulates the production of membrane-associated mucins in human conjunctival epithelial cells.

methods. A telomerase-immortalized human conjunctival epithelial cell line (HCjE) was used. Cells were cultured in serum-free medium to confluence and then cultured with either 10% calf serum or with 100 nM RA for 0 to 72 hours. Conventional RT-PCR was used to determine the expression of retinoic acid receptors (RARs) and quantitative real-time PCR was used to investigate the mRNA expression of MUC1, -4, and -16. Protein levels were assayed by immunoblot analysis, using the antibodies HMFG-2, 1G8, or OC125, which are specific to MUC1, -4 and -16, respectively. To determine whether RA-associated MUC4 mRNA induction is a direct or indirect effect, HCjE cells were treated with RA and the protein synthesis inhibitor cycloheximide (1.0 μg/mL) for 12 hours.

results. MUC1 and -16, but not -4, mRNAs were detectable in HCjE cells grown in serum-free medium. Real-time PCR revealed that MUC4 mRNA was significantly induced by serum 3 hours after its addition, and that MUC1 and MUC16 mRNA levels were significantly upregulated at 72 hours. Western blot analysis demonstrated that the MUC1, -4, and -16 proteins increased over time after addition of serum. Conventional RT-PCR analysis demonstrated that RAR-α and -γ mRNA were expressed in native human conjunctival tissue as well as in the HCjE cells. Treatment with RA upregulated the expression of both MUC4 and -16 mRNA and protein, but MUC1 was unaffected. Because the protein synthesis inhibitor cycloheximide did not prevent the RA-associated induction of MUC4 mRNA, the action of RA on the MUC4 promoter may be direct.

conclusions. The membrane-associated mucins of the ocular surface epithelia, MUC1, -4, and -16, are differentially regulated by serum and RA in the telomerase-immortalized human conjunctival epithelial cell line. Serum derived from vessels in the conjunctiva may play an important role in mucin regulation in the ocular surface epithelia. These data also support the clinical efficacy of autologous serum and RA application in patients with ocular surface diseases. Furthermore, the data suggest that MUC4 and -16 are particularly important hydrophilic molecules involved in maintenance of a healthy ocular surface.

Mucins are large and highly glycosylated glycoproteins present at the interface between wet-surfaced epithelia and their extracellular environments. At the ocular surface, mucins are envisioned as concentrated in the layer of the tear film adjacent to and in the glycocalyx of the apical cells of the corneal and conjunctival epithelia. ^{1 2 3} They are believed to attract and hold water because of their hydrophilic character, thus preventing desiccation of the epithelial surface. ^{1 4} Mucins also lubricate the ocular surface during the eyelid blink, help to maintain a smooth refractive surface, and provide a barrier to pathogen penetration. ^{1 5 6}  

Based on the presence of structural motifs within their amino acid sequence, mucins have been classified as membrane-associated (MUC1, -3A, -3B, -4, -11, -13, -15, -16, and -17) ^{7 8 9 10 11 12} or secreted. The latter include the large gel-forming mucins produced by goblet cells throughout the body (MUC2, -5AC, -5B, and -6) ⁷ and the small, soluble mucins (MUC7 and -9). ^{7 13} Membrane-associated mucins have a short cytoplasmic tail, a transmembrane domain, and a long extracellular domain that is present in the glycocalyx. ⁷ Many of the membrane-associated mucins have a potential cleavage site in their extracellular domain and are thought to be shed from the apical portion of epithelial cells as the soluble form of the mucin. ^{14 15 16} The stratified epithelia of cornea and conjunctiva express the membrane-associated mucins MUC1, -4, and -16. ^{1 17 18} MUC1 and -16 are present along the apical membrane of the apical and subapical cells in human ocular surface epithelia and in suprabasal cells in conjunctival epithelium, whereas MUC4 is present throughout the entire epithelium except for the central cornea in humans, where mRNA levels for the mucin are lower. ^{1 5 17 18}  

Serum contains a number of growth factors, vitamin A, and anti-inflammatory factors that have the potential to maintain a healthy ocular surface. Several studies have examined the effects of the application of serum and retinoic acid (RA), the biologically active form of vitamin A, in maintaining a healthy and hydrated ocular surface epithelium, ^{19 20 21 22 23 24} and autologous serum has been used as a treatment of severe dry eye. ^{23 24} Although their efficacy has been reported, little is known regarding how these agents specifically work to enhance epithelial health, and it is not known whether serum or RA regulate membrane-associated mucins in these epithelia. 

Lack of sufficient vitamin A causes abnormal differentiation of the ocular surface, resulting in keratinization of both conjunctival and corneal epithelial cells, termed xerophthalmia. ²⁵ The effects of RA in cells are mediated by members of a superfamily of nuclear receptors, the retinoic acid receptors (RAR)-α, -β, and -γ. Bossenbroek et al. ²⁶ demonstrated RAR-α, -β, and -γ in rabbit corneal epithelium and fibroblasts and in conjunctival fibroblasts, and Mori et al. ²⁷ reported RAR-α and -γ in mouse corneal epithelium and stroma and in conjunctival tissue. It is not known, however, which RAR subtypes exist in human conjunctival epithelia. 

The purpose of this study was to determine whether membrane-associated mucins are regulated by serum or RA in ocular surface epithelia. Study of membrane-associated mucin regulation in human ocular surface epithelia has been difficult due to the lack of availability of appropriate cell lines. We have recently characterized mucin expression in a telomerase-immortalized conjunctival epithelial cell line (HCjE) to facilitate studies of mucin gene expression. ²⁸ We investigated RAR expression and serum and RA-mediated regulation of the expression of the membrane-associated mucins, MUC1, -4, and -16, in this human conjunctival epithelial cell line, using quantitative real-time PCR and Western blot analysis. 

Methods

Tissue Collection

All tissue was obtained in accordance with good clinical practice, Institutional Review Board and informed consent regulations of the Schepens Eye Research Institute and the Massachusetts Eye and Ear Infirmary, and the tenets of the Declaration of Helsinki. Small pieces of bulbar conjunctiva were removed from the superior temporal conjunctiva of patients undergoing routine cataract surgery to obtain RNA for comparisons of RARs in native tissue with those in cultured HCjE cells. 

Cell Culture

The conjunctival epithelial cell line used was the HCjE cell line, the derivation and mucin gene expression profile of which was previously reported. ^{28 29}  

HCjE cells were cultured in keratinocyte serum-free medium (K-sfm; Gibco-Invitrogen Corp., Rockville, MD) in six-well plates (5 × 10⁴ cells/cm²) at 37°C in a 5% carbon dioxide atmosphere, followed by culture in a 1:1 mixture of K-sfm and low calcium DMEM/F12 (Gibco-Invitrogen Corp.) to confluence. At confluence, the cells were switched to stratification medium, DMEM/F12 with 1 mM CaCl₂ and 10 ng/mL EGF (Hyclone, Logan, UT) and 10% calf serum (Invitrogen, Rockville, MD), and cultured for 0, 3, 6, 12, 24, 48, and 72 hours. In other experiments, the cells were cultured in DMEM/F12 without serum and EGF but with 100 nM of all-trans RA (Sigma-Aldrich, St. Louis, MO) or vehicle dimethylsulfoxide (DMSO; Sigma-Aldrich) for 0, 3, 6, 12, 24, 48, and 72 hours. Because RA was dissolved in DMSO and diluted 1:20,000 in culture medium to achieve the final concentration of 100 nM, DMSO alone was also diluted 1:20,000 in culture media as a control for the effect of vehicle. Serum and RA experiments were done twice and three times respectively, each experiment being done in duplicate. 

Cycloheximide Treatment

To investigate whether RA regulation of MUC4 occurs directly on the MUC4 promoter or indirectly through newly synthesized transcription factors, the effect of the protein synthesis inhibitor cycloheximide (CHX; Sigma-Aldrich) was determined by incubating HCjE cells with 1.0 μg/mL of CHX and 100 nM of RA for 12 hours, using methods previously reported. ³⁰ Briefly, HCjE cells were treated with CHX for 12 hours in the presence of 0.5 μCi/mL [³H]leucine. The cells were washed twice with PBS, proteins precipitated with 10% trichloroacetic acid, and centrifuged at 14,000 rpm for 15 minutes at 4°C. The precipitates were resolubilized with 50 mM Tris buffer (pH 7.2), and the radioactivity was quantitated with a liquid scintillation counter (Beckman Coulter, Fullerton, CA). Cycloheximide experiments were done twice, in duplicate. 

Isolation of RNA and Reverse Transcription–Polymerase Chain Reaction

After culture, total RNA from the cells was isolated using TRIzol reagent (Invitrogen), according to the manufacturer’s recommended protocol. RNA was digested with DNase I (Amplification Grade; Invitrogen), before reverse transcription (RT) to remove any residual genomic contamination. Total RNA (2.0 μg) from the HCjE cells was reverse transcribed using the first-strand synthesis system for RT-PCR (SuperScript; Invitrogen) and random hexamer primers according to the manufacturer’s instructions, as previously described. ³¹ Expression of RARs was determined by conventional RT-PCR using published primers. ³² Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) primers (Applied Biosystems, Foster City, CA) were used to confirm the integrity of the cDNA. Amplifications were performed in a programable thermal cycler (TouchDown Thermal Cycler System; Hybaid, Ashford, UK). PCR conditions for RARs were modified as follows from published methods. Briefly, PCR amplification reactions were conducted in 50 μL reaction volumes containing 5 μL of 10x Taq buffer, 5 μL of 10 mM deoxynucleoside triphosphates, 2 μL of first-strand HCjE cell line cDNA, 5 μL of 50 mM MgCl₂, 10 pmol of each primer, and 2 units of Taq DNA polymerase (AmpliTaq Gold; Applied Biosystems). The mixture was denatured at 96°C for 10 minutes, followed by 35 cycles at 96°C for 30 seconds, 60°C for 30 seconds, and 72°C for 1 minute. Ten microliters of PCR mixture were electrophoresed on 1.0% agarose gel and stained with ethidium bromide. 

Real-Time PCR

The relative amounts of MUC1, -4, and -16 mRNA in the HCjE cells were determined by real-time PCR using a sequence detection system (TaqMan Chemistry and GeneAmp 7900HT; Applied Biosystems). After reverse transcription of total RNA (2.0 μg) from the samples, PCR amplification was performed, as described previously, ^{31 33} in the presence of double-labeled fluorogenic probes (TaqMan Probes; Applied Biosystems) that allow the relative quantitation of gene expression in real-time. The primers and TaqMan probes used (MUC1, -4, and -16 and GAPDH) have been published. ^{18 31 34} Validation experiments were performed to confirm equivalent PCR efficiencies for GAPDH and the target genes. For relative quantitation, we used the ΔC_T method (Applied Biosystems) reported previously. ²⁸ The C_T value is the fractional cycle number at which the amount of amplified target reaches a fixed threshold of detectable fluorescence. The threshold is set in the midlinear phase of the amplification plot. To standardize the amount of sample cDNA added to each reaction, the amount of target gene in each sample was normalized to the endogenous control (GAPDH) by subtracting the C_T of GAPDH from that of the target gene (equals ΔC_T). For quantitation, the amount of mRNA for each target gene was expressed relative to the amount present in a calibrator sample (ΔΔC_T method). For the serum and RA studies, MUC1 expression in HCjE cells in the 0 hour control was used as the calibrator. Thus, the level of mRNA for the MUC1 0 hour control was set at 1, and all other conditions (not only MUC1, but also -4 and -16) were expressed relative to it. Using the MUC1 0 hour control as calibrator for all conditions allows determination of relative expression levels between the different membrane-associated mucin genes as well as between the different time points for each individual mucin gene. For the cycloheximide study, which examined only MUC4 expression, the MUC4 0 hour control was used as the calibrator. Samples were assayed in duplicate in a total volume of 50 μL, using thermal cycling conditions comprised of 2 minutes at 50°C, 10 minutes at 95°C followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. No template controls were run in each assay to confirm lack of DNA contamination in reagents used for amplification. 

The effect of serum or RA on the expression of the individual mucin genes was assessed by statistical comparisons of the mRNA levels for each of the time points to that of their own 0 hour baseline level. Statistical comparisons of results obtained by real-time PCR were performed with the Fisher protected least-significant difference (Fisher’s protected least-significant difference [PLSD]) test (Statview 5.0 for Macintosh; SAS Institute Inc., Cary, NC). P < 0.05 was considered significant. 

SDS-PAGE and Western Blot Analysis

Protein was extracted from the treated cells with RIPA buffer (50 nM Tris, 0.1% SDS, 0.5% deoxycholate, 1% NP-40, 150 nM NaCl) plus complete protein inhibitor cocktail (Roche Biochemical, Indianapolis, IN). Protein concentration was determined with BCA Protein Assay Reagent Kit (Pierce, Rockford, IL). Fifty micrograms of total protein were separated under reducing conditions on 4% stacking and 7.5% separating SDS-polyacrylamide gels, according to the Laemmli system ³⁵ and transferred to nitrocellulose membranes by conventional methods. ³⁶ Primary antibodies used for MUC1, -4, and -16 and GAPDH are listed in Table 1 . For assay of MUC1 and -16 and GAPDH protein, membranes were blocked with 5% (w/v) nonfat milk in Tris-buffered saline–0.1% Tween 20 (5% BLOTTO; Santa Cruz Biotechnology, Santa Cruz, CA). After 1 hour’s incubation with primary antibody diluted in 5% Blotto (1:100 dilution for MUC1, 1:1000 for MUC16, and 1:2000 for GAPDH), the membranes were incubated with horseradish peroxidase–conjugated goat anti-mouse IgG₁ (Santa Cruz Biotechnology) for MUC1 and -16, and with horseradish peroxidase–conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology) for GAPDH, diluted in nonfat dried milk (5% Blotto; 1:5000 dilution). For assay of MUC4 protein, we modified the previously published methods. ³⁷ Briefly, membranes were blocked with 5% (wt/vol) nonfat dry milk in TBS-0.5% Tween 20. After 1 hours’ incubation with anti-MUC4 monoclonal antibody 1G8, diluted in 3% BSA/Tris-buffered saline/0.5% Tween 20 (1:1000 dilution), the membranes were incubated with horseradish peroxidase-conjugated goat anti-mouse IgG₁ (Santa Cruz Biotechnology) diluted 1:5000 in 3% BSA, Tris-buffered saline, and 0.5% Tween 20. Protein bands were detected using chemiluminescence techniques ³⁸ with chemiluminescent substrate (SuperSignal West Pico; Pierce) and then exposed on film (Hyperfilm; Amersham Biosciences, Buckinghamshire, UK). Band intensities were quantified with NIH Image software (v1.62; available in the public domain at http://rsb.info.nih.gov/nih-image/ National Institutes of Health, Bethesda, MD). Serum experiments were done twice, and RA experiments were repeated three times. Western blot analyses were performed for each experiment. 

Results

Effect of Serum on Membrane-Associated Mucin Gene Expression

The expression of membrane-associated MUC1, -4, and -16 mRNA was determined by real-time PCR in confluent HCjE cells cultured in medium with or without 10% calf serum. Expressing all mRNA levels relative to the 0-hour, no-serum control MUC1 level allowed convenient comparisons of the relative amounts of MUC1, -4, and -16, expressed by these cells, while still allowing comparison of the effect of treatment on the individual mucin genes over the experiment’s time course. Figure 1 shows the relative expression of mucin mRNA in HCjE cells cultured with (Fig. 1A) and without (DMEM/F12 alone; Fig. 1B ) serum over the time course. No MUC4 mRNA was present without serum (0.002, relative amount), whereas MUC1 and -16 mRNA were present in cells grown in the absence of serum (1.00 and 3.97, respectively). MUC4 mRNA was significantly induced by serum after 3 hours, and the amount continued to increase in a time-dependent manner (P < 0.05). The expression of MUC1 and -16 mRNA started to increase 24 hours after addition of serum and was significantly increased from baseline (0 hour) at 72 hours (P = 0.0014 and 0.0036, respectively). No significant change was detected in mRNA of any of the mucins in HCjE cells cultured without serum (Fig. 1B) . This study also revealed that MUC16 transcripts were the most prevalent of the membrane-associated mucin mRNA in HCjE cells until 72 hours after addition of serum, at which time MUC4 mRNA became the most prevalent (Fig. 1A) . 

Effect of Serum on Membrane-Associated Mucin Protein Expression

Western blot analysis of the MUC1, -4, and -16 proteins was performed to determine the effect of serum on mucin protein production. Figure 2 shows the immunoblots and densitometric analyses of the effect of serum on mucin production over time, from 0 to 72 hours. As a control, HCjE cells were cultured without serum for 72 hours (Fig. 2 , 72c). The ratio of the densitometric value of the mucin protein to that of the GAPDH band was determined at each time point to compare the relative amount of mucin protein present. MUC1 protein was upregulated by serum after 12 hours of treatment, whereas no upregulation was found in the HCjE cells cultured without serum for 72 hours. Using the 1G8 antibody to MUC4, which recognizes a 120- to 140-kDa portion of one of the subunits of rat Muc4, ascites sialoglycoprotein (ASGP-2), two bands of different molecular mass, 250 (Fig. 2A) and 114 (Fig. 2B) kDa, were observed. The lower molecular mass band (114 kDa), which is the expected mass of ASGP-2, more closely correlated with mRNA expression patterns, whereas the 250-kDa band did not. The 114-kDa MUC4 protein was absent at the 0 hour (Fig. 2B) , which correlates to the data showing that MUC4 mRNA was also absent at the 0 hour, although both protein bands were upregulated by serum in a time-dependent manner. The 114-kDa band gradually increased over time, reaching its highest levels at 72 hours—again, correlating with mRNA levels (see Fig. 1A ). In the control, the 114-kDa MUC4 protein was undetectable (Fig. 2B) , as was mRNA in the cells cultured without serum for 72 hours. MUC16 protein was undetectable in the 0 hour control and at 72 hours cultured without serum, even though mRNA was present. MUC16 protein was not detectable until 12 hours after treatment and increased dramatically with time after 24 hours of serum treatment. 

Expression of RARs in HCjE Cells

Although RARs have been identified in rabbit conjunctival fibroblasts using Northern blot analysis ²⁶ and in mouse conjunctival tissue using immunohistochemistry, ²⁷ RARs in human conjunctival epithelial cells have not been demonstrated. Before studying the role of RA and mucin gene expression in these cells, we looked for the presence of RARs in the HCjE cells as well as in native human conjunctiva. In native human conjunctival biopsy tissue, RAR-α and -γ but not RAR-β are expressed as shown by conventional RT-PCR (Fig. 3A) . Conventional RT-PCR also clearly demonstrated the presence of RAR-α and -γ mRNA in HCjE cells, whereas mRNA for all the RAR subtypes (RAR-α, -β, and -γ) were expressed in a telomerase-immortalized tracheal bronchial cell line used as a positive control (Fig. 3B) . We used conventional RT-PCR to investigate the presence of RAR mRNA in untreated HCjE cells (0 hour) and those treated with 100 nM RA for 96 hours. The results showed similar levels of RAR-α and -γ in RA-treated and untreated cells, with no induction of RAR-β by RA treatment (Fig. 3C) . 

Effect of RA on Membrane-Associated Mucin Gene Expression

Having demonstrated the presence of RARs in the cell line, we examined the effect of RA on membrane-associated mucin gene expression. Cultures were treated with 100 nM RA or DMSO vehicle (control) in the absence of serum. Figure 4 shows the relative expression of mucin mRNA cultured with 100 nM RA (Fig. 4A) or vehicle (DMSO; Fig. 4B ) over the 72-hour time course, as determined by real-time PCR. Mucin mRNA values are expressed relative to the MUC1 0-hour, no-RA control. This revealed that transcripts for MUC16 were the most prevalent of the three membrane-associated mucins expressed by these cells. No change was detected in the expression of MUC1 mRNA after treatment with either RA or DMSO vehicle. On the contrary, the expression of MUC4 mRNA was induced by RA and significantly increased from 3 to 72 hours (P < 0.05, Fig. 4A ). MUC16 mRNA was expressed in the absence of RA or vehicle, and message levels were increased by RA 48 and 72 hours after its addition (6.50 and 11.79, respectively, Fig. 4A ). Only the upregulation after 72 hours of RA treatment was significantly different (P = 0.0074) from baseline (Fig. 4A) . 

Effect of RA on Membrane-Associated Mucin Protein

Western blot analysis of MUC1, -4, and -16 was performed to determine the effect of RA on the protein production of these mucins. Figure 5 shows the immunoblots and densitometric analyses of the mucin protein produced after treatment with 100 nM RA over time and after treatment with vehicle (DMSO) for 72 hours. Densitometric comparisons of mucins normalized to GAPDH were obtained at each time point. MUC1 protein levels were not affected by 100 nM RA treatment over time, which correlates with PCR data. As with serum experiments, multiple bands were seen on MUC4 immunoblots of HCjE proteins from cells treated with RA. The 250-kDa (Fig. 5A) and the 114-kDa (Fig. 5B) bands seen with serum treatment were also present after RA treatment. In addition, after RA treatment, a 93-kDa (Fig. 5C) band was observed. Each of the three molecular weight bands of MUC4 protein was absent at the 0 hour, which correlates to the data showing that MUC4 mRNA was also absent at the 0 hour, and although each of the proteins was upregulated by RA in a time-dependent manner, the increases in the 114- and the 93-kDa proteins over time correlate more closely with the mRNA data for RA treatment. MUC16 protein was not readily observed in blots until 12 hours after RA treatment and increased dramatically from 48 to 72 hours. MUC16 protein was detectable in the cells cultured with vehicle for 72 hours, but the amount was much less than with RA treatment for 72 hours (Fig. 5) . 

Effect of Cycloheximide on MUC4 mRNA Levels

Because MUC4 was the only mucin gene that was upregulated by RA within a comparatively short time period, we sought to determine whether RA acts directly on the MUC4 promoter or through an intermediate transcription factor. MUC4 mRNA expression was examined by real-time PCR, after treatment of HCjE cells with RA in the presence of 1.0 μg/mL of the protein synthesis inhibitor CHX for 12 hours. This concentration of CHX inhibited incorporation of [³H]leucine into the protein by 91.0% during the 12-hour incubation. No significant difference in the amount of MUC4 mRNA was found between the RA and RA+CHX treatments (P = 0.68, Fig. 6 ). Therefore, inhibition of protein synthesis did not prevent the RA-associated increase in the MUC4 mRNA level. These results indicated that de novo protein synthesis is not necessary for the increased MUC4 gene expression, thus suggesting that RA acts directly on the MUC4 promoter. 

Although human conjunctival goblet cells also produce the secreted mucin MUC5AC, this study focused on the membrane-associated mucins. We have previously shown that HCjE cells do not express MUC5AC mRNA when grown on plastic. ²⁸ In this study, we found that neither serum nor RA induced MUC5AC expression in HCjE cells grown on plastic (data not shown). 

Discussion

The major findings of this research show that serum and the derivative of vitamin A, all-trans RA, upregulated the expression of membrane-associated mucins expressed by the human ocular surface epithelium at both the mRNA and protein levels. We also found that the pattern of regulation of each mucin was different from the others, suggesting that the three membrane-associated mucins are independently regulated. 

Several studies have examined the regulation of MUC1, -4, and -16 expression in various cell lines and tissues, and the data suggest that regulation is epithelium specific. ^{32 40 41 42 43 44} For example, estrogen and progesterone induced Muc4 in mouse reproductive tract epithelia, whereas it has no effect on Muc4 mRNA levels of ocular surface epithelia. ⁴⁰ Thus, it is necessary to determine regulators of ocular surface epithelial mucin genes, especially because levels of mucins are affected by ocular surface disease. ^{31 45} The ocular surface epithelial cell line CCL 20.2 (Chang conjunctival cell line; ATCC, Manassas, VA) was reported to increase expression of MUC1 after application of human serum, as determined by flow cytometry. ²³ However, there is a question as to whether this cell line is a true conjunctival epithelial cell line, because it has an HeLa cell contaminant (see description of CCL 20.2 at ATCC, available at http://www.atcc.org), and many researchers consider the cells to be fibroblastic. Thus, the data presented herein can be used to initiate studies of mucin gene regulation by the human ocular surface epithelium. 

Although the data from this study show that serum upregulated each of the membrane-associated mucins, the patterns of their regulation were different. MUC1 and -16 mRNAs were detectable in HCjE cells grown in serum-free conditions, whereas MUC4 mRNA was not. MUC1 protein was easily detected without serum, but MUC4 protein was not, and although MUC16 mRNA was present, protein was not detectable, even with the highly sensitive chemiluminescence method. After addition of serum, the expression of MUC1, -4, and -16 mRNA and protein were increased in a time-dependent manner, but the time point at which they were significantly upregulated differed, with MUC4 reaching significance much earlier (3 hours), compared with 72 hours for MUC1 and -16. Early compared with late upregulation of mRNA suggests direct and indirect promoter activation, respectively. 

When HCjE cells were cultured in the absence of serum, MUC4 mRNA was not detected; however, it was detected 3 hours after addition of serum. These data are comparable to the finding in human pancreatic tumor cells (CD18/HPAF-SF) that MUC4 mRNA was not expressed without serum, but was expressed within 24 hours of culture, in the presence of serum. ³² This requirement of serum for MUC4 expression may explain the distribution of MUC4 mRNA on the human ocular surface. That is, human MUC4 is highly expressed in conjunctiva, which is rich in vessels, and in limbal cornea, which is easily exposed to serum from conjunctival vessels, whereas MUC4 expression is minimal in central cornea. ^{5 17}  

Human MUC4 is highly homologous to rat Muc4, a well-characterized membrane-associated mucin—also known as sialomucin complex (SMC). SMC is composed of two subunits: ascites sialoglycoprotein (ASGP)-1 and ASGP-2. ^{46 47} ASGP-1, which is a rat homologue of human MUC4α, contains the heavily glycosylated mucin domain, ⁴⁸ and ASGP-2, which is a rat homologue of human MUC4β, contains the transmembrane and two EGF-like domains. ⁴⁹ We used a monoclonal antibody (1G8) that recognized ASGP-2 to detect MUC4. ³⁷ We detected not only a protein of the expected molecular mass (120–140 kDa for ASGP-2), but also larger (250 kDa) and smaller (93 kDa) molecular weight proteins. It is possible that these bands represent nonspecific binding—especially the 250-kDa band. The expression patterns of the 250-kDa protein did not correlate as closely to the mRNA expression pattern as did the 114- and 93-kDa bands. A previous study using the same MUC4 antibody (1G8) also showed several protein bands in neutrophil elastase–treated human bronchial epithelial cells. ³⁷ Fischer et al. ³⁷ suggest that these bands may be differently glycosylated forms. Alternatively, the different forms may represent splice variants or proteins that bind the antibody nonspecifically. 

Although MUC16 mRNA was detected in the 0-hour control condition by RT-PCR, MUC16 protein was undetectable. After addition of serum (12 hours), translation of MUC16 mRNA was induced, and MUC16 protein increased in a time-dependent manner. Thus, the expression pattern of MUC16 transcripts was different from that of the protein. These results suggest that the regulation of MUC16 in human conjunctival epithelial cells may occur posttranscriptionally. Several reports have shown that rat Muc4 mucin expression is regulated by posttranscriptional and posttranslational mechanisms, resulting in discordance between protein and mucin mRNA levels, ^{50 51 52} but to our knowledge, posttranscriptional regulation of MUC16 has not been reported. To determine what influences MUC16 translation in the human ocular surface, further investigation is needed. 

We found that MUC4 and -16 mRNA and protein, but not MUC1 mRNA or protein, were upregulated by 100 nM RA. Because the protein synthesis inhibitor CHX did not prevent the RA-associated MUC4 mRNA induction, RA may act on the MUC4 promoter directly, not through other transcriptional proteins. MUC16 mRNA was significantly upregulated by RA only at 72 hours, but, because HCjE could not survive treatment with 1.0 μg/mL CHX for that period, we could not determine whether RA regulation of MUC16 occurred directly or indirectly. The RA-associated expression pattern for MUC16 transcripts was different from that of the protein—similar to the situation found with serum. Thus, RA may also regulate MUC16 posttranscriptionally. 

It is well known that the ocular surface has an absolute requirement for vitamin A. Previous studies have examined the effect of vitamin A deficiency on the ocular surface using animal models. For example, Tei et al. ⁵³ demonstrated that vitamin A deficiency in rats results in a decrease in expression of rat Muc4 and -5AC mRNA in the conjunctival epithelium, whereas MUC1 mRNA was unaffected. In the present study, we examined the relationship between vitamin A and regulation of membrane-associated mucins in the human ocular surface epithelium, because they may be involved in tear film retention on the surface of the eye. ¹ As in the rat, MUC1 was unaffected by RA, leading us to suggest that MUC4 and -16 are more important in vitamin-A–associated ocular surface wettability. 

RA works through RARs -α, -β, and -γ, which are members of a superfamily of nuclear receptors that includes steroid-thyroid hormone and vitamin D receptors. ^{54 55} Although it has been reported that RAR-α, -β, and -γ are expressed in rabbit corneal epithelium and conjunctival fibroblasts ²⁶ and that RAR-α and -γ are expressed in mouse conjunctiva ²⁷ it was not known which RAR subtypes exist in human conjunctiva. Our study found that RAR-α and -γ, but not RAR -β, were expressed in HCjE cells and human conjunctival tissue, and that, unlike in rabbit corneal and conjunctival cells ²⁶ or F9 embryonic carcinoma cells, ⁵⁶ RAR -β was not induced by RA in HCjE cells. There appears to be species variation in the presence of RAR-β and its inducibility by RA. 

To date, the efficacy of applying autologous serum for the treatment of ocular surface disease has been reported by several investigators in clinical trials. ^{22 23 24} Furthermore, Tsubota et al. ²³ reported that treatment with autologous serum is also effective in patients with Sjögren syndrome. Our data suggest that the mechanism of this efficacy may be related to the upregulation of expression of the membrane-associated mucins MUC1, -4, and -16 in the human ocular surface epithelia. 

In contrast, topical RA has been reported to be effective as a treatment for severe squamous metaplasia, ^{20 21} but not for keratoconjunctivitis sicca. ^{21 57} As our data showed, RA-associated MUC4 and-16 mRNA induction from baseline (0 hour control) was less than the serum-associated induction (Figs. 1 4) . This may be one reason why topical RA therapy for patients with dry eye is not as effective as autologous serum therapy. We hypothesize that RA-associated MUC4 and -16 induction may contribute to the efficacy of topical RA therapy for severe squamous metaplasia through their hydrophilicity, which helps maintain tear fluid on the surface of the eye. 

In summary, we report that the telomerase-immortalized human conjunctival epithelial cell line HCjE was useful in the study of regulation of mucin expression. The expression of all the membrane-associated mucins in the human ocular surface, MUC1, -4, and -16, are upregulated by serum at the mRNA and protein levels, but at various times after culture. Thus, serum from the conjunctival vessels may play an important role in the regulation of membrane-associated mucins in the human ocular surface. Two types of RARs (-α and -γ) are expressed by HCjE cells, and 100 nM all-trans RA induces the expression of MUC4 and -16, but not MUC1. Moreover, RA-associated MUC4 mRNA upregulation is through a direct effect on the MUC4 promoter. We infer that MUC4 and -16 are particularly important hydrating molecules of the ocular surface, due to the drying seen on the ocular surface with vitamin A deficiency. 

 Supported by National Eye Institute Grant EY03306 and a Research to Prevent Blindness Senior Scientific Investigator Award (IKG); and an ARVO/Japan National Society for The Prevention of Blindness Research Fellowship (YH).

 Submitted for publication August 19, 2003; revised September 23, 2003; accepted September 29, 2003.

 Disclosure: Y. Hori, None; S. Spurr-Michaud, None; C.L. Russo, None; P. Argüeso, 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.

 

Table 1.

 

Primary Antibodies Used for MUC1, -4, and -16: and GAPDH for Western Blot Analysis

Mucin	Ab	Company	Host	Epitope	Reference
MUC1	HMFG-2	Biodesign	Mouse IgG	Tandem repeat	39
MUC4	1G8	Zymed	Mouse IgG	ASGP-2	37
MUC16	OC125	Dako	Mouse IgG	Tandem repeat	16
GAPDH	—	Abcam	Rabbit IgG	—	Manufacturer’s protocol (Abcam)

Figure 1.

 

Real-time PCR analysis of MUC1, -4, and -16 expression in HCjE cells grown with 10% calf serum (A) or without serum (B) over a 0- to 72-hour time course. Levels of each mucin mRNA are expressed relative to MUC1 mRNA at 0 hour (no serum). Statistical analysis to determine the effect of serum on the individual mucin genes were determined by comparing the cultures at each time point to that of their own 0 hour level baseline. (A) MUC4 mRNA was undetectable in cells cultured without serum, whereas MUC1 and -16 mRNA were present. MUC4 mRNA was significantly induced by 3 hours and increased in a time-dependent manner (*P < 0.05), thereafter. The expression of MUC1 and -16 mRNA were significantly upregulated 72 hours after addition of serum (†P = 0.014 and **P = 0.0036, respectively). (B) No change was detected for any of the mucins (MUC1, -4, or -16) in HCjE cells cultured without serum. Error bars, SEM

Figure 2.

 

Western blot and densitometric analysis of MUC1, -4, and -16 protein from HCjE cells grown with 10% calf serum over a 0- to 72-hour time course or without serum for 72 hours (72c). MUC1 protein was upregulated by serum after 12 hours of treatment. Two bands of molecular masses of MUC4 protein, an (A) 250- and (B) 114-kDa band, were observed. The 114-kDa band closely correlated with MUC4 mRNA expression. MUC16 protein was not detected until 12 hours after treatment and increased thereafter. Error bars, SEM

Figure 3.

 

Expression of retinoic acid receptors (RARs) by human conjunctival tissue, HCjE cells, and telomerase-immortalized human tracheal bronchial cells, as determined by conventional RT-PCR. (A) RAR-α and -γ mRNA were expressed in human conjunctival tissue. (B) Only RAR-α and -γ mRNA were expressed in the HCjE cells (C), whereas mRNA for all the RAR subtypes (RAR-α, -β, and -γ) were expressed in a tracheobronchial cell line (T). (C) Comparison of the expression of RARs in HCjE cells untreated (0h) and treated for 96 hours (96h) with 100 nM RA. No changes in the presence of any of the RARs were seen after RA treatment.

Figure 4.

 

Real-time PCR analysis of MUC1, -4, and -16 expression in HCjE cells grown with (A) 100 nM retinoic acid (RA) or (B) vehicle (DMSO), both in the absence of serum. (A) No change was detected in the expression of MUC1 mRNA with RA. The expression of MUC4 mRNA was upregulated by RA after 3 hours and increased in a time-dependent manner (*P < 0.05). The expression of MUC16 was significantly upregulated by 100 nM RA at 72 hours after addition of RA (**P = 0.0074). (B) No change was detected for any of the mucins (MUC1, -4, and -16) in HCjE cells cultured with vehicle (DMSO) alone. Error bars, SEM

Figure 5.

 

Western blot and densitometric analysis of MUC1, -4, and -16 protein from HCjE cells grown with 100 nM retinoic acid (RA) over time (0 to 72 hours) or with vehicle alone (DMSO) for 72 hours (72v). The amount of MUC1 protein did not change with RA treatment over time. MUC4 antibody bound to three bands of different molecular masses—(A) 250, (B) 114 and (C) 93 kDa. MUC4 protein was upregulated by RA in a time-dependent manner, although increases in the 114- and 93-kDa bands more closely correlated to mRNA changes (see Fig. 4 ). MUC16 protein was not readily observed in blots until 12 hours after RA treatment. MUC16 protein was also upregulated by vehicle (DMSO), but the amount was less than that by RA. Error bars, SEM

Figure 6.

 

Real-time PCR analysis of MUC4 mRNA expression in HCjE cells cultured with retinoic acid (RA; 100 nM) and CHX (1.0 μg/mL) for 12 hours. No significant changes were detected between RA group and RA+CHX (P = 0.64). Error bars, SEM