INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

GALLBLADDER MUCIN, among other epithelial mucins, has been proposed to play a role in the protection of the underlying epithelium (10), but also in the development of disease, particularly in the onset of gallstone formation (23, 31). In the human gallbladder, a major mucin is biosynthesized, designated previously as human gallbladder mucin (HGBM), which is secreted into bile and localized in the mucous granules at the apical region of the gallbladder epithelial cells (21). This mucin is biosynthesized as a single precursor with an apparent molecular mass of ~470 kDa (21). It is unclear what the identity of this precursor is and whether other mucin precursors are also biosynthesized in the human gallbladder. To date, different human epithelial mucin genes have been identified and named MUC1-4, MUC5AC, MUC5B, and MUC6-8 (2, 7, 8, 11-14, 22, 25, 28, 30, 34; see Ref. 39 for review). Through immunohistochemistry, Northern blotting, and in situ hybridization, several mucin genes were reported to be expressed in the human gallbladder, namely MUC1, MUC3, MUC5AC, MUC5B, and MUC6 (3, 5, 8, 16, 17, 19, 40); all are candidate genes for HGBM. In addition, MUC1-4 and MUC5B were reported to be expressed in the human intestine (3, 8, 16, 35, 38, 40). This is of interest because bovine gallbladder mucin (BGBM) is also expressed in the bovine colon (26), suggesting that HGBM may also be expressed in the intestine. Previously, we showed that we could electrophoretically distinguish between different mucin precursors (38, 40). In this study, it is our aim to identify the major mucin biosynthesized in the human gallbladder by using antibodies raised against purified HGBM and by using antibodies directed against MUC2-4, MUC5B, and MUC6. In addition, we aimed to study the biosynthesis of HGBM in the human intestine and its cellular localization in gallbladder and intestine. This information will further clarify the functions of HGBM and aid in assessing the role of HGBM in diseases of these organs.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Tissues. Samples of the gallbladder fundus (10 patients, 2 with gallstones) and proximal jejunum (2 patients) were obtained from patients undergoing a Whipple operation procedure for carcinoma of the pancreas or papilla Vater or from patients undergoing hemihepatectomy for liver carcinoma. Biopsies of the duodenum (10 patients), ascending colon (4 patients), transverse colon (1 patient), or sigmoid (23 patients) were taken by endoscopy from healthy tissue from patients admitted for reflux esophagitis, polypectomy (nonpolyposis coli), control after succesful eradication of Helicobacter pylori, control for colon carcinoma, diverticulosis, irritable bowel syndrome, or constipation. Use of tissue was approved by the ethics committee of the Academic Medical Center.

Antibodies. Monoclonal anti-MUC5B was raised against purified deglycosylated tracheobronchial mucin and recognizes the tandem repeat of MUC5B (8, 27). Monoclonal antibody WE9 [anti-MUC2(1)] recognizes a peptide epitope in the unique terminals of MUC2 (36). Anti-HCM [anti-MUC2(2)] was raised against purified human colonic mucin and recognizes the unique non-O-glycosylated terminals of human MUC2 (35, 36). Anti-HGBM was raised against purified HGBM and recognizes unique non-O-glycosylated terminals of HGBM (21, 36). Anti-BGBM antibody (anti-BGBM) was raised against deglycosylated BGBM (26). Anti-M3P (anti-MUC3) was raised against a synthetic peptide representing the tandem repeat of MUC3 (16). Anti-MUC6.1 (anti-MUC6) was raised against a synthetic peptide representing the tandemly repeated amino acid sequence of MUC6 and purified by peptide affinity chromatography (5). Anti-MUC4 antibody was raised against the synthetic peptide with the sequence TSSASTGHATLPVTDTSSASC, representing the tandemly repeated amino acid sequence of MUC4, and affinity chromatography purified (not shown). The methods used for anti-MUC4 antibody production and purification and for specificity analysis were identical to those described for the anti-MUC6.1 antibody. For a summary on the antibodies, see Table 1.

Cell culture, metabolic labeling, and immunoprecipitation. LS 174T cells were cultured as described previously (40). Mucin biosynthesis was studied by metabolic labeling of biopsies or LS 174T cells with [³⁵S]methionine/cysteine or [³⁵S]sulfate followed by immunoprecipitations with the various antibodies as described previously (35, 40).

Gel electrophoresis. Radiolabeled mucins were analyzed on 3-4% reducing SDS-PAGE as described previously (35, 40). For molecular weight markers, unreduced rat gastric mucin precursors, labeled with ³⁵S-labeled amino acids, were used, with molecular masses of 300 kDa for the monomer and 600 kDa for the dimer (6). Prestained high molecular weight markers with molecular masses between 49.5 and 205 kDa (Bio-Rad, Richmond, CA) were used. Gels were either Western blotted onto nitrocellulose membranes (35) or fixed in 10% methanol and 10% acetic acid, incubated for 10 min with Amplify (Amersham), and dried. Alternatively, reduced immunoprecipitated mature mucins were analyzed by 0.8% agarose gel electrophoresis in the presence of 0.1% SDS and run for 14 h at 20 mA in buffer containing 0.04 M Tris-borate (Merck, Darmstadt, Germany) and 2 mM EDTA (Sigma Chemical, St. Louis, MO), pH 8.0, and dried according to Thornton et al. (33). All gels were exposed to X-ray film (Biomax MR, Kodak) from 1 to 4 wk at 70°C.

Immunohistochemistry. Tissues were fixed in 4% (wt/vol) paraformaldehyde (Merck), embedded in paraffin (Sherwood Medical, St. Louis, MO), cut in 7-µm sections, and mounted. After deparaffination, sections were boiled in 0.01 M citrate buffer, pH 6.0, for 10 min and treated with 1% (vol/vol) hydrogen peroxide (BDH, Poole, UK) for 30 min to reduce endogeneous peroxidase activity. Sections were then incubated for 30 min in 10 mM Tris (Boehringer, Mannheim, Germany), 5 mM EDTA (Merck), 150 mM NaCl (Merck), 0.25% (wt/vol) gelatin (Sigma Chemical), and 0.05% (vol/vol) Tween (BDH) to reduce nonspecific binding. Sections were incubated with anti-HGBM, anti-BGBM, or anti-MUC2(2) (all 1:2,000) or with anti-MUC5B (1:100) or anti-MUC2(1) (1:500), followed by incubation with biotinylated secondary antibodies and avidin-biotin peroxidase complex (Vectastain ABC kit, Vector Laboratories, Burlingame, UK) according to the manufacturers' suggested protocol.

Immunofluorescent double labeling and confocal laser-scanning microscopy. Tissue sections, prepared as described above, were incubated with anti-HGBM (1:500) or anti-MUC2(1) (1:100), followed by incubation with fluorescein isothiocyanate-conjugated goat anti-rabbit IgG (1:300) or Cy3-conjugated rabbit anti-mouse IgG (1:300), respectively (both from Jackson ImmunoResearch, West Grove, PA). Sections were then incubated with Vectashield (Vector) and evaluated using the dual excitation and detection mode of a Leica confocal laser-scanning microscope equipped with an argon-krypton laser (Leica Lasertechnik, Heidelberg, Germany). The images were optimized for voltage, offset, and merged using multicolor analysis software (Leica Lasertechnik).

RT-PCR. RNA was isolated from mucosal scrapings of gallbladder, jejunum, and colon, as previously described (40). We transcribed 1 µg of RNA at 42°C into cDNA using 200 U Superscript reverse transcriptase (GIBCO BRL, Breda, The Netherlands) in a total volume of 20 µl according to the manufacturer's suggested protocol. The final reaction condition was 20 mM Tris · HCl, pH 8.4, 50 mM KCl, 2.5 mM MgCl₂, 0.01% BSA, 10 mM dithiothreitol, 500 nM random hexamers, 1 µg RNA, and 500 µM of each dATP, dCTP, dGTP, and dTTP. After 1 h, digestion with RNase H (GIBCO BRL) for 10 min at 42°C was carried out. This was followed by a PCR reaction in a total volume of 20 µl using 1 µl cDNA as a template in combination with either -actin primers (5'-CAAGGCCAACCGCGAGAAG-3' and 5'-CAGGGTACATGGTGGTGCC-3') or MUC5B-specific primers (5'-TGGGCCTCGAGTGCCGTG-3' and 5'-CACACGGATTTCATAGTTGAAC-3') based on a nonrepetitive sequence within MUC5B (14). Final PCR reaction conditions were 10 mM Tris · HCl, pH 8.4, 50 mM KCl, 5 mM MgCl₂, 0.01% gelatin, 0.2 U Taq polymerase, 200 nM of each primer, cDNA template, and 200 µM each of dATP, dCTP, dGTP, and dTTP. The PCR reaction was carried out as follows: 5 min at 95°C, 30 1-min cycles at 95°C, 1 min at 57°C, and 1 min at 72°C, followed by a 10-min extension step at 72°C. The PCR products were analyzed on a 2% agarose gel following standard procedures (29). The amplified MUC5B sequence was verified by digestion with restriction enzyme Pst I (Boehringer).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Biosynthesis of HGBM precursors in gallbladder and intestine. To identify HGBM, human gallbladder biopsies were pulse labeled with ³⁵S-labeled amino acids, followed by immunoprecipitations and analysis on SDS-PAGE. Using this method, we previously showed that mucin precursors, synthesized by the colonic cell line LS 174T or by human gastrointestinal tissues, have distinct molecular masses on SDS-PAGE and can thus be discriminated electrophoretically (38, 40). HGBM precursor (apparent molecular mass 470 kDa) was immunoprecipitated from radiolabeled gallbladder homogenates by anti-HGBM and displayed a similar apparent molecular mass as the mucin precursors immunoprecipitated by anti-MUC5B and anti-BGBM (Fig. 1). Anti-MUC3 antibody immunoprecipitated MUC3 precursor from gallbladder homogenate, displaying a different apparent molecular mass (~550 kDa) than MUC5B precursor (Fig. 1). In gallbladder homogenates, no MUC4 or MUC6 precursors were detectable (Fig. 1). In addition, no MUC2 precursors were detectable in gallbladder homogenates, as demonstrated by immunoprecipitations with the antibodies anti-MUC2(1) (Fig. 1) and anti-MUC2(2) (not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Tissue expression of mucin precursors. Biopsies of human gallbladder, duodenum, and sigmoid were metabolically pulse labeled with [35S]methionine/cysteine. Each panel represents immunoprecipitations from the same individual. Mucin precursors were isolated by immunoprecipitations with the indicated antibodies and analyzed on SDS-PAGE. Arrowheads indicate border between 4% running and 3% stacking gel. Molecular mass markers are at left. HGBM, human gallbladder mucin. BGBM, bovine gallbladder mucin.

Identification of HGBM as MUC5B by Western blot analysis. Western blotting was used to establish the identity of the mucin precursors. First, mucin precursors were immunoprecipitated from radiolabeled gallbladder homogenate by anti-HGBM, anti-MUC5B, and anti-BGBM (anti-BGBM not shown), followed by Western blotting of these mucin precursors with anti-MUC5B (Fig. 2, left). The Western blot showed that anti-MUC5B recognized the precursors immunoprecipiated by anti-HGBM, anti-MUC5B, and anti-BGBM (anti-BGBM not shown) antibodies from gallbladder homogenate. Moreover, the precursors all displayed similar apparent molecular masses of 470 kDa. When the blot was exposed to X-ray film, the radiolabeled bands completely coincided with the bands detected by anti-MUC5B on Western blot (Fig. 2, right). This demonstrated that the immunoprecipitated radiolabeled mucin precursors are identical to the bands that were detected by anti-MUC5B on Western blot.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   HGBM is recognized by anti-MUC5B. Biopsies of human gallbladder were pulse labeled with [35S]methionine/cysteine, followed by immunoprecipitation with anti-HGBM or anti-MUC5B as indicated, and run on SDS-PAGE. The gel was Western blotted and incubated with anti-MUC5B (left). Autoradiograph of the Western blot is shown at right.

MUC5B mRNA expression. To determine MUC5B mRNA expression in the gallbladder and the small and large intestine, an RT-PCR experiment was carried out using RNA from these tissues as templates (Fig. 3). First, cDNA was synthesized from the RNA using random primers, followed by a PCR reaction using either MUC5B primers derived from a nonrepetitive MUC5B region (8) or human -actin primers (Fig. 3). When the latter primers were used (Fig. 3, lanes a, c, and e), equal amounts of the -actin PCR fragments were observed of the expected size (587 bp), indicating that equal amounts of template RNA derived from the gallbladder and the small and large intestine were used and that RT-PCR proceeded equally efficiently. Interestingly, MUC5B PCR fragments of the expected length (148 bp) could be amplified from gallbladder and colonic cDNA, but not from jejunal cDNA. The identity of the MUC5B PCR product was further checked by digestion with Pst I, which yielded the expected 88- and 60-bp products (not shown).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   MUC5B mRNA expression assessed by RT-PCR using cDNAs synthesized from RNA of gallbladder (lanes a and b), jejunum (lanes c and d), or colon (lanes e and f). PCR reaction was performed using either MUC5B primers (lanes b, d, and f) or human -actin primers (lanes a, c, and e).

Immunolocalization of MUC5B. To localize MUC5B, sections of human gallbladder, duodenum, and sigmoid were examined using anti-HGBM, anti-MUC5B, or anti-BGBM. In the sigmoid, a granular staining pattern was observed in goblet cells when stained by anti-HGBM, with highest intensity basally in the crypts (Fig. 4A), whereas in the gallbladder an intense staining of mucous granules in all epithelial cells was observed with anti-HGBM (Fig. 4F). In the duodenum, no staining with anti-HGBM was found (Fig. 4E). In the colon, anti-MUC5B showed a perinuclear staining in goblet cells, with highest intensity in the basal crypt region, while staining was absent from the mucous granules (Fig. 4B). The gallbladder also showed a high staining intensity with anti-MUC5B in all epithelial cells (Fig. 4G), whereas no staining was detected in the duodenum (Fig. 4D). Anti-BGBM showed a similar staining pattern as anti-MUC5B in gallbladder epithelial cells (Fig. 4H) and in colon (not shown).

View larger version (188K): [in this window] [in a new window]   Fig. 4.   Immunohistochemistry of human sigmoid (A-C), duodenum (D and E), and human gallbladder (F-H) using anti-HGBM (A, E, and F), anti-MUC5B (B, D, and G), or anti-BGBM (H): In C only secondary (no primary) antibodies were added. L, lumen. CT, colonic crypt. Magnifications: ×100, except F (×400).

View larger version (76K): [in this window] [in a new window]   Fig. 5.   Immunofluorescence double labeling of human gallbladder, duodenum, and sigmoid. Anti-HGBM is shown in green and anti-MUC2(1) in red. Colocalization is rendered in shades of white, as shown in inset in c. Bars, 20 µm. a: Gallbladder, anti-HGBM. b: Duodenum, anti-MUC2(1). c: Sigmoid, anti-HGBM and anti-MUC2(1). d: Sigmoid sectioned through the bottom of a crypt, anti-HGBM and anti-MUC2(1).

Comparison of the biosynthesis and secretion of mature MUC5B and MUC2 in the colon. To compare the biosynthesis and secretion of mature MUC5B to MUC2 in the human colon, pulse labeling with ³⁵S-labeled amino acids, followed by chase incubations, was performed. After pulse labeling, MUC2 precursors (apparent molecular mass 600 kDa) and MUC5B precursors (apparent molecular mass 470 kDa) were immunoprecipitated from colonic homogenate (Fig. 6A). After a 4-h chase incubation an additional diffuse band representing mature mucin was detected, with electrophoretic mobility corresponding to 550 kDa, in immunoprecipitations performed with both anti-MUC2(2) and anti-HGBM on tissue homogenate (Fig. 6A). Precursor bands of MUC5B and MUC2 were weak after 4-h chase incubation, as a result of their conversion into mature mucins. The diffuse 550-kDa bands were also detected in the medium after immunoprecipitations with anti-MUC2(2) and anti-HGBM (Fig. 6A), indicating that mature MUC2 and MUC5B were secreted.

View larger version (46K): [in this window] [in a new window]   Fig. 6.   Synthesis and secretion of mature HGBM (MUC5B) and MUC2 analyzed by SDS-PAGE. Biopsies of human sigmoid were pulse labeled for 30 min with [35S]methionine/cysteine (A) or for 1 h with [35S]sulfate (B), followed by chase incubation in medium without radiolabel as indicated (h). Medium was collected after 4 h chase incubation (m4). MUC2 and MUC5B were immunoprecipitated from tissue and media homogenates using anti-MUC2(2) or anti-HGBM, respectively, and analyzed as in Fig. 1.

View larger version (101K): [in this window] [in a new window]   Fig. 7.   Mature MUC2 (A) and mature HGBM (MUC5B; B) analyzed by agarose gel electrophoresis. Biopsies of human sigmoid or gallbladder or LS 174T cells were pulse labeled for 1 h with [35S]sulfate and homogenized, followed by immunoprecipitations with anti-MUC2(2) or anti-HGBM, and analysis on reducing 0.8% agarose gel electrophoresis. [35S]sulfate-labeled sigmoid homogenate was also analyzed (homog). Gel was dried and exposed to X-ray film. Arrowhead indicates position of the slots. Lanes in between immunoprecipitations with anti-HGBM and anti-MUC2(2) are empty.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The prominent HGBM is identical to MUC5B. The biosynthesis of the 470-kDa HGBM precursor and its maturation, secretion, and abundance in human bile has been demonstrated previously (21). Using the same antiserum raised against purified HGBM, we aimed to identify HGBM, because its identification is essential in studying the role of this mucin in disease. In the present study, we showed that MUC2, MUC4, and MUC6 precursors were not detectable in the gallbladder, whereas MUC3 precursors were detectable, displaying a different molecular mass on SDS-PAGE than HGBM precursors. Similarly, we previously demonstrated that, for Caco-2 and LS 174T cells, HGBM and MUC3 precursors displayed different molecular masses on SDS-PAGE (40). MUC1 was also not a likely candidate for the secretory mucin HGBM, since MUC1 is a relatively small membrane-bound mucin (11). Studies (3, 4, 16, 17, 19, 22, 40) previously indicated that MUC1, MUC3, MUC5B, and MUC6 are present in the gallbladder epithelium, whereas little or no MUC2, MUC4, or MUC5AC was detected. Moreover, MUC1, MUC3, and HGBM were also detected in bile (1, 21). Our immunoprecipitation data are thus in line with reported steady-state levels of mucin mRNA or protein, except that we could not immunoprecipitate MUC6 precursor from gallbladder homogenate. Most likely, the amount of MUC6 biosynthesized in the gallbladder is very low, because we were able to detect MUC6 precursors in LS 174T cells (40) and in human stomach (38), showing that the anti-MUC6 antibody functions well in immunoprecipitations. Finally, we showed that the mucin precursors immunoprecipitated from human gallbladder, by antibodies directed against HGBM, MUC5B, and BGBM, all recognized mucin precursors with similar apparent molecular masses (~470 kDa). Moreover, by means of Western blotting with anti-MUC5B it was confirmed that these mucin precursors represented an identical mucin, namely MUC5B. In conclusion, the mucin detected by anti-HGBM antiserum is MUC5B and this mucin is biosynthesized in high amounts in the gallbladder, showing that MUC5B is the prominent HGBM. However, this does not exclude the presence of other mucins, such as MUC3, in human gallbladder.

Cellular localization of MUC5B in human gallbladder and intestine. To study the cellular localization of MUC5B, immunohistochemistry was performed. Klomp and co-workers (21) previously demonstrated an intense staining of mucous granules in gallbladder epithelial cells with anti-HGBM (21). The staining patterns of BGBM and MUC5B antibodies were similar to that of anti-HGBM: staining of all epithelial cells in the gallbladder, staining of goblet cells in the deeper crypt region of colon, and no staining in the small intestine. However, the cellular staining pattern was different: BGBM and MUC5B antibodies detected perinuclear antigens of colonic goblet cells, suggesting that rough endoplasmic reticulum-localized precursors were detected. This was anticipated because these antibodies recognize tandemly repeated amino acid sequences of BGBM and MUC5B, respectively (26, 27). These sequences are known to become masked on O-glycosylation (32), explaining the inability of anti-BGBM and anti-MUC5B to recognize mature MUC5B, which is stored in granules. In contrast, anti-HGBM recognizes the non-O-glycosylated sequences of HGBM and thus also detects mature mucin (21, 36), which explains the staining of the mucous granules by this antibody. In addition, MUC2 was contained in large mucous granules of the goblet cells of the colon and duodenum. Previously, it was demonstrated (3, 4, 16, 35) that MUC2 is expressed in all colonic goblet cells. However, MUC5B demonstrated a different distribution in colonic goblet cells. Double immunofluorescence labeling showed that MUC5B is typically expressed at a high level in the base of the colonic crypts, colocalizing to MUC2-containing goblet cells, while MUC5B expression ceases during migration of the goblet cells to the luminal side of the crypt. In addition, MUC5B is contained within small granules of the colonic goblet crypt cells.

Mature MUC5B is secreted in the colon and can be discriminated from mature MUC2 by agarose gel electrophoresis. Previously, we showed that MUC2 is the prominent secretory mucin of the human colon (35). In the present study, we demonstrated that MUC5B is also a colonic secretory mucin. Strikingly, we found that mobilities on SDS-PAGE of mature MUC2 and MUC5B are much higher than expected. Because of extensive O-glycosylation, the molecular mass of the mature fully glycosylated mucin is expected to be much larger than 550 kDa. In fact, it was demonstrated that mobilities of mature mucins on SDS-PAGE are aberrant and depend highly on intrinsic negative charge (37). We therefore used agarose gel electrophoresis, which was previously shown to distinguish between mature mucin glycoproteins (33). For comparison, we studied the biosynthesis of mature MUC5B and MUC2 in the human LS 174T colon carcinoma cell line; in this cell line we previously showed the biosynthesis of both MUC2 and MUC5B (40). In the present study it is shown that mature MUC2 and MUC5B, synthesized by both sigmoid and LS 174T cells, can be discriminated by agarose gel electrophoresis, demonstrating unequivocally that mature MUC2 and MUC5B represent different secretory human colonic mucins.

The possible roles of MUC5B in disease. In this study we showed that MUC5B is expressed in the colon, but not in the small intestine, and this suggests a specific function for this mucin in the colon. Among other mucins, MUC5B serves a role in cytoprotection (10), for instance, in protecting the gallbladder from the detergent effects of bile acids. It is tempting to speculate that MUC5B also protects the colonic epithelial cells from bile acids. For instance, MUC5B may protect the colon from unconjugated, hydrophobic bile acids, which are formed by the action of bacteria on bile acids present in the colon (18) and are potent inducers of mucin secretion (20). Colonic bile acids are causally related to the development of colorectal cancer (15, 24). Also, increased bile acid content of the stool is found in patients with inflammatory bowel disease (9). Therefore, in view of our data, it is of interest to study colonic MUC5B expression and function in these patients.