Quantification of MUC2 and MUC5AC Transcripts in Human Conjunctiva

Abstract

purpose. Transcripts of mucins 1, 4, and 5AC have been identified in human conjunctival tissue. Of these, only MUC5AC has been localized to goblet cells. MUC2 is a goblet cell mucin originally identified in the intestinal mucosa. The presence of MUC2 transcripts and levels of MUC2 and MUC5AC transcripts in normal human conjunctiva, as determined by quantitative polymerase chain reaction (PCR), is reported.

methods. RNA from conjunctivae of six donors (3 men, 3 women, 44 to 69 years, all white) was isolated and subjected to competitive reverse transcription–PCR. Internal standards, which are dsDNA molecules with ends complementary to a given primer pair but containing nonhomologous central sequences, were prepared for each gene assayed. Titration of a constant amount of cDNA against serial dilutions of the internal standard allowed quantification of the template cDNA. MUC2 and MUC5AC levels were compared to levels of the “housekeeping” gene,β ₂-microglobulin (β₂M). The identity of PCR products was confirmed by sequencing.

results. In the six individual samples tested, β₂M mRNA is expressed, on average, at approximately 10⁻²⁰ moles per sample (1 μg RNA) or approximately 63.5 × 10⁴ molecules. The mRNA encoding MUC5AC, a relatively abundant ocular mucin, exists at levels 10-fold lower than β₂M. In contrast to previous reports of MUC2 mRNA being absent at the ocular surface, these results show that MUC2 transcripts are present and expressed at levels 5900-fold lower than for MUC5AC. Apparently, MUC2 transcripts exist on the order of only approximately 100 to 1000 molecules/μg of RNA in the analyzed samples.

conclusions. MUC2 transcripts are present in human conjunctival tissue and their abundance is much lower than that of MUC5AC. This is the first application of competitive PCR to the quantitative analysis of mucin expression in human ocular tissue. The sensitivity of this method allows the detection of MUC2 transcripts that were not detected by Northern blot analysis or in situ hybridization in previous studies. It also makes possible the comparison of relative levels of expression for ocular mucins.

The preocular tear film is a trilayered structure consisting of an outermost layer of Meibomian lipids, a middle aqueous layer, and an inner, hydrated mucus layer. The mucus layer, a fluid-filled meshwork of glycoproteins, functions to lubricate and protect the ocular surface. It anchors the aqueous tear film to the underlying conjunctival and corneal surfaces and protects this epithelium from sheer force damage, drying, and bacterial invasion. The tear film is the first protective layer encountered by particulates, pathogens, and noxious agents. In other tissues, ^{1 2 3 4} mucins function as antioxidants and inhibit bacterial adherence; similar roles are likely for tear film mucins. Recently, the thickness of the mucus layer has been reexamined. The depth of this layer was revised to between 30 and 40 μm, ^{5 6} four- to sixfold thicker than the 7 μm previously reported. However, a recent report suggests that the depth is closer to original estimates. ⁷ Resolution of these differences will require further study. Dilly ⁶ proposes that the mucus layer is stable and anchored to the ocular surface epithelium and that the aqueous layer, containing dissolved mucins, provides a cleavage plane for lid movement. The more viscous mucus layer dissipates these movements to protect the underlying epithelium. Although the structure of the tear film is becoming more clearly understood, the mucin composition and extent to which these molecules contribute to the functions of the tear film is less clear. 

Mucins comprise a heterogeneous family of glycoproteins expressed by most specialized epithelial tissues of mucosal surfaces. To date, nine distinct mucin genes have been reported (MUC1 to 4, MUC5AC, MUC5B, MUC6 to 8). Historically, expression of these genes was thought to be tissue-specific. However, it appears that many of the mucin genes are expressed in a wide variety of tissues with overlapping patterns of expression. ^{8 9 10} Of these, three have been detected on the ocular surface: MUC1, MUC4 (ASGP-2), and MUC5AC. ^{11 12 13 14} MUC1 and ASGP-2 are membrane-associated mucins produced by nongoblet conjunctival epithelial cells. ^{12 15 16 17} Conjunctival goblet cells are thought to be the major source of the soluble ocular mucin MUC5AC. ^{14 18}  

Of the known mucins, only MUC2, MUC5AC, and MUC5B are consistently associated with goblet cells. In situ hybridization studies by Inatomi et al. ¹⁸ show that mRNA for MUC5AC is localized in human conjunctival goblet cells. These authors, using Northern blotting techniques were unable to detect any MUC2 transcripts in normal human conjunctiva. Previously reported studies from our laboratory, using reverse transcription–polymerase chain reaction (RT-PCR), a more sensitive technique, clearly demonstrate the presence of MUC2 transcripts in normal human conjunctival tissues. ¹³ The present studies therefore were undertaken to confirm the presence of MUC2 transcripts in human conjunctiva, to quantify the abundance of MUC2 mRNA relative to that of MUC5AC, and to verify the presence of the corresponding secretory product in the normal human tear film. 

Methods

Tissue Samples and Cell Lines

Human donor conjunctivae were obtained from the Kentucky Lions Eye Bank in accordance with protocols approved by the University of Louisville Human Studies Committee. Conjunctival tissue was collected from Eye Bank donors (identified by age, gender, race, cause and time of death). Tissues from 3 women, ages 44 years [postmortem excision time (PET), 3.0 hours], 66 years (PET, 3.0 hours), and 66 years (PET, 1.5 hours) and 3 men, ages 65 years (PET, 5.0 hours), 67 years (PET, 4.0 hours), and 69 years (PET, 4.0 hours) were studied. Tissue (approximately 500 mg) was collected by excision and included conjunctiva from the superior and inferior bulbar regions and conjunctiva from the nasal fornix (plica luminaris). Excised tissue was placed in ice-cold tissue culture medium (DMEM/F12 containing 5% fetal bovine serum). Tissue destined for RNA analysis was processed immediately on receipt (generally within 4 hours postmortem). Tissue to be used for Western blot analysis was homogenized in lysis buffer (10 mM Tris, pH 8.0, and 1 mM EDTA) and stored at −70°C. 

Human corneal endothelial cells (HCN; Steve Wilson, University of Washington, Seattle, WA) and LS180 cells (ATCC, Manassas, VA) were cultured in HEM (DMEM/F12 containing 10 ng/ml epidermal growth factor, 1 μg/ml insulin, 5% fetal bovine serum, and 1% antibiotic–antimycotic mix) before extraction of RNA. 

Collection of Human Tear Samples

Tear samples were collected as previously described. ¹⁴  

RNA Isolation

Total RNA was isolated using a guanidinium isothiocyanate protocol (RNeasy; Qiagen, Santa Clarita, CA). The resultant RNA was treated with RNase-free DNase, extracted twice with phenol:chloroform:isoamyl alcohol (24:24:1), precipitated with ethanol, dissolved in RNase-free water, and quantified spectrophotometrically. Average yields observed were approximately 1 μg RNA per milligram of tissue (wet weight). Total RNA from human salivary gland and trachea was used for controls (Clontech, Palo Alto, CA). Each RNA preparation was characterized qualitatively by electrophoresis on agarose gels stained with ethidium bromide. 

Competitive RT-PCR

Quantitative analysis of mucin transcripts was carried out by MIMIC competitive PCR (Clontech). The method is a co-PCR reaction with target cDNA and a competing exogenous template serving as an internal standard. ^{19 20 21} The competitive MIMIC template is a dsDNA construct with ends engineered to be complimentary to the mucin primers, but containing a nonhomologous central sequence (v-erbB). Coamplification of a constant amount of conjunctival cDNA in the presence of serial dilutions of this internal standard allows for quantification of mucin mRNA.β ₂-Microglobulin (β₂M), a ubiquitously expressed protein, was used as a “housekeeping gene” for comparative purposes. After reverse transcription, 1 μl (∼1 ng) of conjunctival cDNA was amplified and analyzed. Primers for both MUC2 and MUC5AC were as previously published. ²¹ Primers forβ ₂M were purchased from Clontech. PCR parameters consisted of an initial denaturation at 94°C for 10 minutes to heat activate the DNA polymerase (Amplitaq Gold; Perkin Elmer, Foster City, CA). This was followed by 30 cycles of denaturation at 94°C for 1 minute (40 cycles for MUC2), annealing at 59°C for 1 minute for MUC2 (60°C for MUC5AC, 54°C forβ ₂M), and extension at 72°C for 1 minute with a final extension at 72°C for 7 minutes. MgCl₂ was included at 2.5 mM for β₂M and MUC5AC (1.5 mM for MUC2). Identities of PCR products were confirmed by sequencing. After electrophoresis in 1% agarose, bands corresponding to target cDNAs and MIMIC competitors were stained with ethidium bromide and photographed with Polaroid Type 667 film. The photographs were digitized with an Epson Action Scanner II digital image processor and analyzed using SigmaGel (Jandel, San Rafael, CA) software. Levels of mucin and β₂M PCR products were determined by densitometric analysis and normalized to corresponding MIMIC standards. Results of MUC2 and MUC5AC amplification were compared to those obtained from β₂M amplification to determine the relative abundance of mucin transcripts. Reported numbers of transcripts are based on the assumption of 100% efficiency of cDNA synthesis although actual efficiency is lower. Therefore, the estimated transcript numbers represent the minimum number of mucin-specific mRNAs in the tissue samples. 

SDS-PAGE and Western Blot Analysis

Human conjunctival tissue and tears were prepared in SDS-PAGE sample buffer. All samples and standards were boiled for 5 minutes before loading onto 4% to 15% polyacrylamide gradient gels (Bio-Rad, Hercules, CA). Gels were run for 1 hour at 200 mV, washed in transfer buffer, and electroblotted onto nitrocellulose membranes for 1 hour at 100 mV. Blots were blocked overnight at 4°C with TTBS containing 2% ovalbumin and then incubated with the anti-MUC antisera at 1:1000 or with the corresponding normal serum at the same dilution. After three rinses, blots were incubated with horseradish peroxidase–conjugated rabbit anti-mouse IgG (1:5000) for 1 hour and developed using a chemiluminescent substrate (Dupont NEN, Boston, MA). The MUC1 antibody (DF3) was purchased from Signet (Dedham, MA). The MUC2 antibody, 4F1, is a kind gift of Shirley Bolis, Peter Devine, and Carol Morris, University of Queensland. The MUC5AC antibody was prepared and characterized as described previously. ¹⁴  

Results

Identification of MUC2 Transcripts in Human Tissues and Cell Lines

To determine whether the MUC2 gene contributes to mucin biosynthesis at the ocular surface, expression of the gene was analyzed in human conjunctiva by RT-PCR. RNA preparations from human conjunctiva, trachea, salivary gland, a colorectal cancer cell line (LS180), and from primary cultures of human corneal endothelial cells were reverse transcribed, and the resultant cDNAs were amplified. In agreement with previous findings in human airway tissues, ^{8 21 22 23} we detected the predicted 438 bp MUC2 PCR product in trachea (Fig. 1) . LS180 cells also express MUC2, the predominant gastrointestinal tract mucin, as has been previously reported. ²⁴ As expected, no MUC2 signal was detectable in salivary gland or corneal endothelial cells, which served as negative controls. However, contrary to a previous report of an absence of MUC2 on the human ocular surface, ¹⁸ a positive PCR signal was detected in conjunctival RNA. The identity of this PCR product was confirmed by sequencing. 

MUC5AC expression was analyzed in the same set of tissues and cells. As reported previously, ¹⁴ primers for MUC5AC amplify the predicted sized product (680 bp) in human conjunctival cDNA as well as in the positive control tracheal cDNA. ^{25 26} MUC5AC is expressed in the gastrointestinal tract ^{27 28} and not surprisingly, in LS180 cells. The MUC5AC PCR product was not detected in salivary gland extracts, thus confirming previous reports. ²¹ As expected, we did not detect PCR product in human corneal endothelial cells. 

Quantification of Mucin Transcripts by Competitive RT-PCR

A competitive RT-PCR method (MIMIC; Clontech) was used to determine levels of MUC2, MUC5AC, and β₂M transcripts in conjunctival samples from six donors. For results of the competitive amplification to be valid, both the competing MIMIC template and the target cDNA must amplify with comparable efficiencies. Target cDNA and MIMIC template were amplified independently, and levels of PCR products were visualized on ethidium bromide–stained agarose gels. The number of cycles and starting amounts of MIMIC template were determined empirically such that, after amplification, levels of target cDNA products and MIMIC products were comparable. Both cDNA and MIMIC templates for the two mucins, MUC2 and MUC5AC, and forβ ₂M, amplify with comparable efficiencies. This can be seen in Figure 2 , as cDNA- and MIMIC-amplified products increase in parallel as the cycle number increases. 

To assess levels of mucin expression in human conjunctiva, cDNA from each of six donors was amplified in the presence of competing MIMIC templates. Tenfold serial dilutions of MIMIC templates were titrated against constant levels of cDNA (data not shown). These results were then used to determine the MIMIC concentrations at which the cDNA:MIMIC ratios were close to 1. Twofold serial dilutions of MIMIC in this range were prepared and coamplified with cDNA. Typical results for each mucin are shown in Figure 3 . The double log plot allows determination of the number of MIMIC molecules at which the cDNA/MIMIC ratio is 1 (log 1 = 0). The results of this analysis indicate that for the individuals shown MUC2, MUC5AC, and β₂M transcript levels in 1 μg of RNA are approximately 5 × 10¹, 5.6 × 10⁴, and 4.5 × 10⁵ molecules, respectively. 

The results for all six individuals tested are summarized in Table 1 . Transcript levels (per microgram of RNA) are presented as well as relative expression levels (mucin levels normalized to levels of the housekeeping gene, β₂M). MUC2 transcripts are clearly detectable in each sample tested. The absolute values vary over two orders of magnitude from less than 20 for subject 2 to between 1000 and 2000 for subject 6. MUC5AC transcript levels, however, show less variation among the six individuals. In the sample group tested, men show higher levels of MUC2 than do age-matched women, whereas two of three women show higher levels of MUC5AC than do men. 

To determine whether differences in general transcriptional activity contribute to the observed variation, we normalized transcript numbers of MUC2 and MUC5AC to those of β₂M. β₂M transcription has previously been shown to remain relatively constant over time and independent of treatment compared to other typical housekeeping genes, such as β-actin and glyceraldehyde-3-phosphate dehydrogenase. ²¹ Mucin levels expressed relative toβ ₂M, however, continue to show variation over two orders of magnitude for MUC2. The range of MUC5AC non-normalized values extends from 2.3 to 61.4 × 10⁴, slightly more than a 26-fold difference. The MUC5AC values, when expressed relative to those of β₂M, continue to show about the same level of variation, from 4.6 to 180.5, or slightly less than a 40-fold range. The observed variation is likely due to differences in expression of these genes between individuals rather than due to differences in sample collection or general transcriptional activity. 

Presence of MUC2 Glycoprotein in Human Conjunctiva and Tears

To determine whether the MUC2 protein is present in normal human conjunctival tissue and in tears, samples were analyzed by Western blotting (Fig. 4) . Two other mucins were analyzed as positive controls: MUC1, a transmembrane mucin expressed in conjunctiva, ²⁹ and MUC5AC, shown by previous studies in our laboratory to be present in conjunctiva and tears. ¹⁴ Western blot analysis showed that all three mucin antibodies bind high-molecular-mass components in conjunctival samples. The MUC1 antibody (Ab) did not bind to any soluble high-molecular-mass component in tears, consistent with MUC1’s role as a transmembrane molecule. Antibodies for MUC2 and MUC5AC bound specifically to high-molecular-weight components of tears. 

Discussion

The long-term goal in our laboratory is the understanding of the roles of ocular mucins in human health and disease. Dry eye diseases affect an estimated six million patients in the United States alone, and these patients frequently receive only palliative treatment. ³⁰ The understanding of the functional roles of the individual tear film components and the regulatory steps governing their synthesis and secretion is necessary to developing effective treatment strategies. 

Transcripts of MUC1, MUC4, and MUC5AC have been identified in normal human conjunctiva. We now add MUC2 as a component of the conjunctival mucin profile. MUC2 is transcribed in human conjunctival tissue, and its corresponding protein is present in the tear film. The application of competitive PCR to the quantitative analysis of MUC2 and MUC5AC expression in ocular tissue demonstrates the relatively low abundance of MUC2. The sensitivity of this method allows the detection of transcripts for MUC2, which have not been detected by either Northern blot analysis or in situ hybridization. It also has made possible the comparison of relative levels of ocular mucin transcripts. 

Although both MUC2 and MUC5AC transcripts are expressed in the normal human conjunctiva, levels of MUC2 mRNA are considerably lower than those of MUC5AC. Of the individuals tested, MUC5AC levels exceed those of MUC2 by 24-fold (subject 6, highest MUC2 level observed) to 3 × 10⁴-fold (subject 2, lowest MUC2 level and highest MUC5AC level observed). On average, the MUC5AC values exceed MUC2 levels by 5.9 × 10³-fold. Interestingly, men in the study express higher levels of MUC2 than do women. The mean level of MUC2 for men (807 transcripts per microgram of RNA) exceeds that of women (mean value of 38 transcripts per microgram of RNA) by 21-fold. Normalized MUC2 values show the same trend—a 17-fold difference between men and women. For MUC5AC, women show slightly higher levels than men (3.2-fold difference for nonnormalized values and 7.9-fold difference for normalized values). Admittedly, the sample size is small, and further testing is needed to determine whether the observed gender differences extend to protein expression and are representative of the normal population. 

It is possible that transcripts from different mucin genes differ in stability. This could lead to artifactual differences in mRNA expression levels in postmortem tissue. However, given the magnitude of the difference (5.9 × 10³-fold) between MUC2 and MUC5AC transcript levels, it is unlikely that it is a result of differential stability. Additionally, as partially degraded transcripts can be amplified by RT-PCR, it is even less likely that differential stability is involved. 

Previous studies in our laboratory have shown that MUC5AC glycoprotein is readily detectable in the normal human tear film. This study demonstrates, by Western blot analysis, the presence of MUC2 protein in conjunctival tissue and in tears. An immunoreactive band of high molecular mass (>210 kDa) was detected as well as smaller bands that likely represent MUC2 variants, degradation products, or artifacts of sample preparation. The increased levels of these smaller bands in tears suggest that they may be products of protein degradation by proteinases contained in the tear film. 

We are currently investigating the hypothesis that MUC2 is expressed in a subset of conjunctival goblet cells. In a developmental study of respiratory tract mucins, Buisine et al. ³¹ noted that MUC2 expression precedes that of MUC5AC and is present in differentiating cells, but only in a subset of mature goblet cells. As goblet cells of the conjunctiva apparently arise from precursors localized in the fornix, we predict that this region may be the source of conjunctival MUC2. It is of interest to note that the tissues used in the present study for mRNA isolation included both fornical and bulbar conjunctiva. Little is known about possible mechanisms governing MUC2 secretion. MUC5AC secretion is stimulated by adenine nucleotides via P2Y₂ receptors. ³² Because MUC2, like MUC5AC is considered a goblet cell mucin, we hypothesize that its secretion, too, may be triggered by activation of P2Y₂ receptors. 

As RT-PCR becomes more widely applied to detection of mucin transcripts in ocular tissues, it is likely that mucins previously thought not to exist at the ocular surface will be identified. Previous reports by Inatomi and associates ¹⁸ indicated that of mucins 2 through 7, only MUC4 and MUC5AC are present in human conjunctiva. Northern blot analysis of these mucins showed positive signals for MUC4 and MUC5AC. MUC7 showed a positive signal only after extended exposure. Further analysis by in situ hybridization indicated the presence of MUC4 in stratified conjunctival epithelium and MUC5AC in conjunctival goblet cells. MUC7 was not detected. However, the more sensitive RT-PCR method is capable of detecting fewer than 10 transcripts in a sample. Using this method, we have detected both MUC2 (present results) and more recently, MUC7 (unpublished data) in human conjunctiva. We are currently testing for presence of additional mucin transcripts and secreted mucin proteins in human conjunctiva and in tears. 

As additional mucins are identified on the ocular surface, it is becoming clear that the mucin profile in the tear film is complex. Analysis of the cellular sources and functional contributions of individual tear film mucins are the ongoing goals of our studies. 

 Supported by National Institutes of Health, National Eye Institute Grant EY10736, the Kentucky Lions Eye Foundation, and an unrestricted grant from Research to Prevent Blindness.

 Submitted for publication June 16, 1999; revised September 1 and October 12, 1999; accepted October 27, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Identification of MUC2 and MUC5AC transcripts in human tissues and cell lines by RT-PCR. (A) MUC2 transcripts were detected in cDNA from human conjunctiva (lane 1), human trachea (lane 2), and in LS180 cells (lane 3). The amplified PCR product is 438 bp. No MUC2 PCR product was detected in human salivary gland cDNA (lane 4) or human corneal endothelial cell cDNA (lane 5). (B) MUC5AC transcripts were detected in the same tissues/cell lines (680-bp amplified PCR product). Lanes M in (A) and (B) contain 100-bp molecular weight ladders (500- and 1000-bp bands are indicated).

Figure 2.

 

Amplification efficiencies of MUC2 (A), MUC5AC (B), and β₂M (C) cDNAs and corresponding MIMIC standards. Levels of PCR-amplified products from mucin cDNAs and MIMIC templates were determined after a selected number of cycles to confirm that amplification efficiencies were comparable. Negative images of ethidium bromide–stained gels are shown as insets (M, MIMIC PCR products; C, cDNA PCR products). Bands were quantified by densitometric scanning (□, MIMIC; ▪, cDNA). Note: in (B) (MUC5AC), the 4th cDNA PCR product failed to develop and was not included in the analysis.

Figure 3.

 

Representative results for competitive RT-PCR of MUC2 (A), MUC5AC (B), and β₂M (C). Twofold serial dilutions of MIMIC were titrated against a constant amount of cDNA and amplified as described (Methods). Ethidium bromide–stained gels are shown as insets (M, MIMIC PCR products; C, cDNA PCR products). The equivalence point, at which cDNA-amplified product levels are equal to MIMIC-amplified product levels (cDNA/MIMIC = 1, log 1 = 0), is represented by an asterisk.

Table 1.

 

MUC2, MUC5AC, and β₂M Transcript Levels in Conjunctiva from Six Individuals

Subject No.	Sex	Age	MUC2 (×101)*	MUC5AC (×104)*	β2M (×104)*	Normalized Values (to β2M)
												MUC2 (10−5)	MUC5AC (10−2)
1	F	44	50 ± 13.9	8.8 ± 0.7	41.6 ± 2.7	12.0	21.1
2	F	66	19 ± 2.4	61.4 ± 1.5	34.0 ± 7.5	5.6	180.5
3	F	66	44 ± 5.6	2.3 ± 0.3	50.0 ± 7.8	8.8	4.6
4	M	65	398 ± 118.4	4.9 ± 0.4	73.5 ± 11.3	54.1	6.6
5	M	67	248 ± 13.7	13.5 ± 0.1	136.0 ± 33.3	18.2	9.9
6	M	69	1776 ± 548.0	4.4 ± 0.8	45.6 ± 3.4	389.4	9.6

 All individuals were tested in triplicate for each of the three gene products.

*  Number of transcripts from 1 μg of total RNA expressed as mean ± SEM.

Figure 4.

 

Antibody labeling of human conjunctival extracts and human tear samples. Extracts and samples were separated by SDS-polyacrylamide gel electrophoresis and subjected to Western blot analysis using anti-MUC1 (lane 1), anti-MUC2 (lane 2), or anti-MUC5AC (lane 3). The antibody to the membrane mucin, MUC1, labeled high molecular weight components of conjunctival homogenate (A) but not of tears (B). MUC2 and MUC5AC antisera labeled high-molecular-weight components in both conjunctival extracts and tears. Positions of marker proteins are indicated.

The authors thank Celia G. Emberts for excellent technical assistance.