Mucus and mucins in diseases of the intestinal and respiratory tracts.

This review describes the organization and importance of mucus in the intestine and lungs in relation to the diseases cystic fibrosis (CF), ulcerative colitis and chronic obstructive pulmonary disease (COPD). The inner surfaces of the body are protected by mucus built around polymeric glycoproteins called mucins. In the disease CF, the small intestinal mucus is in contrast the normal attached to the epithelium, explaining the intestinal problems at this disease. The inner of the two mucus layers of colon is normally impenetrable to bacteria, keeping the commensals away from and protecting the epithelium. This impenetrable property is dependent on the bacterial composition and the host diet, observations that can explain the increased incidence of inflammatory bowel diseases in the western world as bacteria reach the epithelial cells in active ulcerative colitis. The respiratory tract is normally cleared by thick mucus bundles that moved by the cilia sweep the epithelial surface. In CF, the bundles are nonmoving already at birth. Cholinergic stimulations stop the bundle movement explaining some of the beneficial effect of anticholinergic treatment in COPD. In this disease as well as in more developed CF, an attached mucus layer is formed. This mucus has features similar to the protective inner colon mucus and is by this able to separate bacteria from the epithelial surface. When formed in healthy individuals this mucus can be coughed up, but in chronically diseased lungs, bacteria colonizing the mucus will remain in the lungs and the resulting inflammation contribute to the destruction of the lungs.

Functional mucous layer and healing of proximal colonic anastomoses in an experimental model.

BACKGROUND: Anastomotic leakage (AL) is the most dreaded complication after colorectal surgery, causing high morbidity and mortality. Mucus is a first line of defence against external factors in the gastrointestinal tract. In this study, the structural mucus protein Muc2 was depleted in genetically engineered mice and the effect on healing of colonic anastomoses studied in an experimental model. METHODS: Mice of different Muc2 genotypes were used in a proximal colonic AL model. Tissues were scored histologically for inflammation, bacterial translocation was determined by quantitative PCR of bacterial 16S ribosomal DNA, and epithelial cell damage was determined by assessing serum levels of intestinal fatty acid-binding protein. RESULTS: Of 22 Muc2-deficient (Muc2-/- ) mice, 20 developed AL, compared with seven of 22 control animals (P < 0·001). Control mice showed normal healing, whereas Muc2-/- mice had more inflammation with less collagen deposition and neoangiogenesis. A tendency towards higher bacterial translocation was seen in mesenteric lymph nodes and spleen in Muc2-/- mice. Intestinal fatty acid-binding protein levels were significantly higher in Muc2-/- mice compared with controls (P = 0·011). CONCLUSION: A functional mucous layer facilitates the healing of colonic anastomoses. Clinical relevance Colorectal anastomotic leakage remains the most dreaded complication after colorectal surgery. It is known that the aetiology of anastomotic leakage is multifactorial, and a role is suggested for the interaction between intraluminal content and mucosa. In this murine model of proximal colonic anastomotic leakage, the authors investigated the mucous layer at the intestinal mucosa, as the first line of defence, and found that a normal, functioning mucous layer is essential in the healing process of colonic anastomoses. Further research on anastomotic healing should focus on positively influencing the mucous layer to promote better postoperative recovery.

Core 1- and 3-derived O-glycans collectively maintain the colonic mucus barrier and protect against spontaneous colitis in mice.

Core 1- and 3-derived mucin-type O-glycans are primary components of the mucus layer in the colon. Reduced mucus thickness and impaired O-glycosylation are observed in human ulcerative colitis. However, how both types of O-glycans maintain mucus barrier function in the colon is unclear. We found that C1galt1 expression, which synthesizes core 1 O-glycans, was detected throughout the colon, whereas C3GnT, which controls core 3 O-glycan formation, was most highly expressed in the proximal colon. Consistent with this, mice lacking intestinal core 1-derived O-glycans (IEC C1galt1-/-) developed spontaneous colitis primarily in the distal colon, whereas mice lacking both intestinal core 1- and 3-derived O-glycans (DKO) developed spontaneous colitis in both the distal and proximal colon. DKO mice showed an early onset and more severe colitis than IEC C1galt1-/- mice. Antibiotic treatment restored the mucus layer and attenuated colitis in DKO mice. Mucins from DKO mice were more susceptible to proteolysis than wild-type mucins. This study indicates that core 1- and 3-derived O-glycans collectively contribute to the mucus barrier by protecting it from bacterial protease degradation and suggests new therapeutic targets to promote mucus barrier function in colitis patients.

New developments in goblet cell mucus secretion and function.

Goblet cells and their main secretory product, mucus, have long been poorly appreciated; however, recent discoveries have changed this and placed these cells at the center stage of our understanding of mucosal biology and the immunology of the intestinal tract. The mucus system differs substantially between the small and large intestine, although it is built around MUC2 mucin polymers in both cases. Furthermore, that goblet cells and the regulation of their secretion also differ between these two parts of the intestine is of fundamental importance for a better understanding of mucosal immunology. There are several types of goblet cell that can be delineated based on their location and function. The surface colonic goblet cells secrete continuously to maintain the inner mucus layer, whereas goblet cells of the colonic and small intestinal crypts secrete upon stimulation, for example, after endocytosis or in response to acetyl choline. However, despite much progress in recent years, our understanding of goblet cell function and regulation is still in its infancy.

Slc26a3 deficiency is associated with loss of colonic HCO3 (-) secretion, absence of a firm mucus layer and barrier impairment in mice.

AIM: Downregulated in adenoma (DRA, Slc26a3) is a member of the solute carrier family 26 (SLC26), family of anion transporters, which is mutated in familial chloride-losing diarrhoea (CLD). Besides Cl(-) -rich diarrhoea, CLD patients also have a higher-than-average incidence of intestinal inflammation. In a search for potential explanations for this clinical finding, we investigated colonic electrolyte transport, the mucus layer and susceptibility against dextran sodium sulphate (DSS)-induced colitis in Slc26a3(-/-) mice. METHODS: HCO3 (-) secretory (JHCO3 (-) ) and fluid absorptive rates were measured by single-pass perfusion in vivo and in isolated mid-distal colonic mucosa in Ussing chambers in vitro. Colonocyte intracellular pH (pHi ) was assessed fluorometrically, the mucus layer by immunohistochemistry and colitis susceptibility by the addition of DSS to the drinking water. RESULTS: HCO3 (-) secretory (JHCO3- ) and fluid absorptive rates were strongly reduced in Slc26a3(-/-) mice compared to wild-type (WT) littermates. Despite an increase in sodium/hydrogen exchanger 3 (NHE3) mRNA and protein expression, and intact acid-activation of NHE3, the high colonocyte pH in Slc26a3(-/-) mice prevented Na(+) /H(+) exchange-mediated fluid absorption in vivo. Mucin 2 (MUC2) immunohistochemistry revealed the absence of a firm mucus layer, implying that alkaline secretion and/or an absorptive flux may be necessary for optimal mucus gel formation. Slc26a3(-/-) mice were highly susceptible to DSS damage. CONCLUSIONS: Deletion of DRA results in severely reduced colonic HCO3 (-) secretory rate, a loss of colonic fluid absorption, a lack of a firmly adherent mucus layer and a severely reduced colonic mucosal resistance to DSS damage. These data provide potential pathophysiological explanations for the increased susceptibility of CLD patients to intestinal inflammation.

Ulcerative colitis patients in remission have an altered secretory capacity in the proximal colon despite macroscopically normal mucosa.

BACKGROUND: One of the hallmarks of acute colitis is loss of epithelial transport. For unknown reasons, many patients still suffer from GI symptoms during remission, indicating a sustained imbalance between absorption and secretion. We hypothesize that the colonic epithelium becomes more reactive to secretagogues to compensate for a failing barrier. METHODS: Biopsies from ascending colon and sigmoid colon of UC patients in remission and controls were mounted in Ussing chambers. Membrane current (Im) and epithelial capacitance (Cp) were used as markers for anion secretion and mucus exocytosis. Carbachol (1 mmol L(-1) ) and forskolin (10 µmol L(-1) ) were used to study Ca(2+) and cAMP-mediated secretion. KEY RESULTS: Baseline values showed segmental patterns with higher Im in ascending colon and higher Cp in sigmoid colon of both UC patients and controls, but the patterns did not differ between the groups. The Im response to forskolin was increased (+35%) in the ascending colon of UC patients and the Im response to carbachol was decreased (-40%) in the same segment. No group differences were seen in the distal colon for either the forskolin or carbachol-induced Im responses. The Cp response to carbachol was instead up-regulated in the distal colon of UC patients, but remained unaffected in the proximal colon. CONCLUSIONS & INFERENCES: The proximal colonic mucosa of UC patients in remission seems to shift its reactivity to secretagogues, becoming more sensitive to cAMP-dependent secretion and less sensitive to Ca(2+) -dependent secretion. This phenomenon may contribute to residual diarrhea in this patient group, despite resolution of inflammation.

Importance and regulation of the colonic mucus barrier in a mouse model of colitis.

The colonic mucus layer serves as an important barrier and prevents colonic bacteria from invading the mucosa and cause inflammation. The regulation of colonic mucus secretion is poorly understood. The aim of this study was to investigate the role of the mucus barrier in induction of colitis. Furthermore, regulation of mucus secretion by luminal bacterial products was studied. The colon of anesthetized Muc2(-/-), Muc1(-/-), wild-type (wt), and germ-free mice was exteriorized, the mucosal surface was visualized, and mucus thickness was measured with micropipettes. Colitis was induced by DSS (dextran sodium sulfate, 3%, in drinking water), and disease activity index (DAI) was assessed daily. The colonic mucosa of germ-free and conventionally housed mice was exposed to the bacterial products LPS (lipopolysaccharide) and PGN (peptidoglycan). After DSS induction of colitis, the thickness of the firmly adherent mucus layer was significantly thinner after 5 days and onward, which paralleled the increment of DAI. Muc2(-/-) mice, which lacked firmly adherent mucus, were predisposed to colitis, whereas Muc1(-/-) mice were protected with significantly lower DAI by DSS compared with wt mice. The mucus barrier increased in Muc1(-/-) mice in response to DSS, whereas significantly fewer T cells were recruited to the inflamed colon. Mice housed under germ-free conditions had an extremely thin adherent colonic mucus layer, but when exposed to bacterial products (PGN or LPS) the thickness of the adherent mucus layer was quickly restored to levels observed in conventionally housed mice. This study demonstrates a correlation between decreasing mucus barrier and increasing clinical symptoms during onset of colitis. Mice lacking colonic mucus (Muc2(-/-)) were hypersensitive to DSS-induced colitis, whereas Muc1(-/-) were protected, probably through the ability to increase the mucus barrier but also by decreased T cell recruitment to the afflicted site. Furthermore, the ability of bacteria to regulate the thickness of the colonic mucus was demonstrated.

Lactobacillus and Bifidobacterium species do not secrete protease that cleaves the MUC2 mucin which organises the colon mucus.

The colon epithelium is covered by two layers of mucus built around the MUC2 mucin. An inner dense and attached mucus layer does not allow bacteria to penetrate, thus keeping the epithelial cell surface free from bacteria. An outer loose mucus layer is the habitat for the commensal bacterial microbiota. The inner mucus layer is renewed from the epithelial side and gets converted into the outer layer due to proteolytic cleavages by host proteases. We have now analysed if potential probiotic bacteria, namely Lactobacillus brevis, Lactobacillus plantarum, Lactobacillus bulgaricus and Bifidobacterium lactis, can secrete protease that cleaves the MUC2 mucin. We found that none of the potential probiotic bacteria Lactobacillus and Bifidobacterium could cleave the MUC2 core protein in the form of recombinant MUC2 N and C-termini although they secreted active proteases. This was in contrast to crude mixtures of oral and faecal bacteria that cleaved the MUC2 mucin. This observation further supports the view that these potential probiotic bacteria are of no harm to the host, as these bacteria cannot disrupt the mucin organised mucus as long as they are covered by glycans.

CD43 promotes cell growth and helps to evade FAS-mediated apoptosis in non-hematopoietic cancer cells lacking the tumor suppressors p53 or ARF.

CD43 is a highly glycosylated transmembrane protein expressed on the surface of most hematopoietic cells. Expression of CD43 has also been demonstrated in many human tumor tissues, including colon adenomas and carcinomas, but not in normal colon epithelium. The potential contribution of CD43 to tumor development is still not understood. Here, we show that overexpression of CD43 increases cell growth and colony formation in mouse and human cells lacking expression of either p53 or ARF (alternative reading frame) tumor-suppressor proteins. In addition, CD43 overexpression also lowers the detection of the FAS death receptor on the cell surface of human cancer cells, and thereby helps to evade FAS-mediated apoptosis. However, when both p53 and ARF proteins are present, CD43 overexpression activates p53 and suppresses colony formation due to induction of apoptosis. These observations suggest CD43 as a potential contributor to tumor development and the functional ARF-p53 pathway is required for the elimination of cells with aberrant CD43 expression.

Bioprocess development for the production of a recombinant MUC1 fusion protein expressed by CHO-K1 cells in protein-free medium.

The mucin MUC1 is a candidate for use in specific immunotherapy against breast cancer, but this requires the large-scale production of a MUC1 antigen. In this study, a bioprocess for the expression of a recombinant MUC1 fusion protein with a cancer associated glycosylation in CHO-K1 cells has been developed. Cells permanently expressing parts of the extracellular portion of MUC1 fused to IgG Fc were directly transferred from adherent growth in serum-containing medium to suspension culture in the protein-free ProCHO4-CDM culture medium. Using the Cellferm-pro system, optimal culture parameter as pH and pO(2) were determined in parallel spinner flask batch cultures. A pH of 6.8-7.0 and a pO(2) of 40% of air saturation was found to give best cell growth and productivity of secreted recombinant protein. Specific productivity strongly depended the pO(2) and correlated with the online monitored oxygen uptake rate (OUR) of the cells, which indicates a positive influence of the rate of oxidative phosphorylation on productivity. The optimised conditions were applied to continuous perfusion culture which gave very high cell densities and space time yields of the recombinant MUC1 fusion protein, allowing production at gram scale. The product degradation was much lower in supernatants from continuous perfusion culture compared to batch mode. Antibodies reacting with cancer associated MUC1 glycoforms strongly bound to the fusion protein, indicating that the desired glycoforms were obtained and suggesting that the recombinant MUC1 protein could be tested for use in immunotherapy.

The leukocyte antigen CD43 is expressed in different cell lines of nonhematopoietic origin.

CD43 is an abundant transmembrane sialoglycoprotein in leukocyte-type cell lines, but it has also been suggested to be present in colon adenomas and colon carcinomas. We have now shown that CD43 is expressed in a variety of cell lines of different origins (CaSKI, A549, 293, MTSV1-7, MCF7, HT-1080, Jurkat, K562, COLO 205, HT-29, Caco-2, DLD-1 and SW480). The level of expression of CD43 mRNA was analyzed by reverse transcriptase-polymerase chain reaction and that of the protein by immunoprecipitation and Western blot, flow cytometry and confocal microscopy using two monoclonal anti-CD43 antibodies (L10 and 4D2). As all cell lines expressed CD43, it is suggested that CD43 has a more fundamental function than previously believed and thus cannot be considered only as a specific leukocyte marker.

Neutralization of pH in the Golgi apparatus causes redistribution of glycosyltransferases and changes in the O-glycosylation of mucins.

Addition of the weak base ammonium chloride (NH4Cl) or the proton pump inhibitor bafilomycin A1 to cultured HeLa and LS 174T cells effectively neutralized the pH gradient of the secretory pathway. This resulted in relocalization of the three studied glycosyltransferases, N-acetylgalactosaminyltransferase 2, beta1,2 N-acetylglucosaminyltransferase I, and beta1,4 galactosyltransferase 1, normally localized to the Golgi stack, the medial/trans-Golgi and the trans-Golgi/TGN, respectively. Indirect immunofluorescence microscopy, immunoelectron microscopy, and subcellular fractionation of the tagged or native glycosyltransferases showed that NH4Cl caused a relocalization of the enzymes mainly to vesicles of endosomal type, whereas bafilomycin A1 gave mainly cell surface staining. The general morphology of the endoplasmic reticulum and Golgi apparatus was retained as judged from immunofluorescence and electron microscopy studies. When the O-glycans on the guanidinium chloride insoluble gel-forming mucins from the LS 174T cells were analyzed by gas chromatography-mass spectrometry after neutralization of the secretory pathway pH by NH4Cl over 10 days shorter O-glycans were observed. However, no decrease in the number of oligosaccharide chains was indicated. Together, the results suggest that pH is a contributing factor for proper steady-state distribution of glycosyltransferases over the Golgi apparatus and that altered pH may cause alterations in glycosylation possibly due to a relocalization of glycosyltransferases.

Tumor cell MUC1 and CD43 are glycosylated differently with sialyl-Lewis a and x epitopes and show variable interactions with E-selectin under physiological flow conditions.

The mucins secreted from the colon carcinoma cell line COLO 205 have the MUC1 and CD43 (leukosialin) as core proteins, where both carry sialyl-Lewis a and MUC1 sialyl-Lewis x epitopes. The adhesion of E-selectin expressing CHO cells to the coated mucins was analyzed in a flow system revealing that the MUC1 mucin adhered better than the CD43 mucin. One reason could be their different glycosylation, a difference that was explored by analyzing the biosynthesis of MUC1 and CD43 in COLO 205 cells. Both the MUC1 and CD43 mucins became sialyl-Lewis a reactive, but after different times as revealed by pulse-chase studies. However, only MUC1 became sialyl-Lewis x reactive. These differences suggest that MUC1 and CD43 are synthesized in different compartments of the cell. It was also observed that the mucins from colon carcinoma patients had MUC1-type mucins that carried both sialyl-Lewis a and x epitopes and CD43-type sialyl-Lewis a mucins with only low levels of sialyl-Lewis x epitopes. One could hypothesize that colon carcinoma derived MUC1 is decorated with potent E-selectin epitopes, and that this could be one of several reasons for the involvement of MUC1 in cancer development.

Sequencing of sulfated oligosaccharides from mucins by liquid chromatography and electrospray ionization tandem mass spectrometry.

As part of a strategy for profiling diverse mixtures of sulfated mucin-derived oligosaccharides, liquid chromatography coupled to electrospray ionization tandem mass spectrometry (ESI-MS/MS) in the negative ion mode has been explored. Two mixtures of sulfated oligosaccharide alditols from porcine stomach and large intestine were analyzed by straight phase chromatography using an amino-bonded column connected to a Q-TOF instrument. Nine sulfated mucin-derived oligosaccharide alditols from porcine stomach underwent extensive fragmentation allowing determination of their sequence. The fragmentation generated primary, secondary, and tertiary fragment ions informative for the elucidation of the saccharide sequence and localization of the sulfate group. From a single chromatographic analysis, the sequences of 28 different sulfated mucin oligosaccharide alditols purified from porcine large intestine were elucidated, revealing information concerning prominent core sequences and terminal blood group-type epitopes. Analysis of these two sulfated oligosaccharide mixtures demonstrated the usefulness of HPLC-ESI-MS/MS: the on-line separation of multiple isomeric suffated oligosaccharides as present in biological samples, informative fragmentation allowing the identification of the sequence of nonderivatized oligosaccharides, and a sensitivity sufficient for the analysis of quantities as obtained from natural sources.

Identification of transient glycosylation alterations of sialylated mucin oligosaccharides during infection by the rat intestinal parasite Nippostrongylus brasiliensis.

The sialylation of the oligosaccharides from small-intestinal mucins during a 13-day infectious cycle was studied in Sprague-Dawley rats with the parasite Nippostrongylus brasiliensis. Sialic acid analysis and release, permethylation and analysis by GC-MS of the sialylated oligosaccharides isolated from the 'insoluble' mucin complex revealed a relative decrease (4-7-fold) of N-glycolylneuraminic acid compared with N-acetylneuraminic acid just before parasite expulsion. Northern blots showed that this effect was due to the decreased expression of a hydroxylase converting CMP-N-acetylneuraminic acid into CMP-N-glycolylneuraminic acid. Analysis of other rat strains showed that this parasite infection also caused the same effect in these animals. Detailed analysis of infected Sprague-Dawley rats revealed four sialylated oligosaccharides not found in the uninfected animals. These new oligosaccharides were characterized in detail and all shown to contain the trisaccharide epitope NeuAc/NeuGcalpha2-3(GalNAcbeta1-4)Galbeta1 (where NeuGc is N-glycolyl neuraminic acid). This epitope is similar to the Sd(a)- and Cad-type blood-group antigens and suggests that the infection causes the induction of a GalNAcbeta1-4 glycosyltransferase. This model for an intestinal infection suggests that the glycosylation of intestinal mucins is a dynamic process being modulated by the expression of specific enzymes during an infection process.

Liquid chromatography-electrospray mass spectrometry as a tool for the analysis of sulfated oligosaccharides from mucin glycoproteins.

An approach for analyzing sulfated oligosaccharide alditol mixtures by liquid chromatography-electrospray mass spectrometry (LC-ESI-MS) is described. Two columns, an amino-bonded column and a porous graphitized carbon column (PGC) were used. Oligosaccharides were eluted with linear gradients of acetonitrile and water, with 5 mM ammonium hydrogencarbonate or formate buffers at a basic pH. The methods were evaluated on a mixture of sulfated oligosaccharide alditols prepared from mucin glycoproteins from pig stomach. Results from LC-ESI-MS of the mixture were compared with the structural information obtained by high energy collision fragmentation using fast atom bombardment tandem mass spectrometry (FAB-MS-MS). The separation ability of the two columns was also tested using a more complex mixture of sulfated oligosaccharides from pig colon, where several isomers were detected. The potential use of in-source collision-induced dissociation (CID) to gain sequence information of sulfated oligosaccharides was also evaluated. The major fragment ions obtained by in-source CID of the trisaccharide Hex-3HexNAcol6-HexNAc6-SO3 were sufficient for assigning the oligosaccharide sequence and the position of the sulfate group within the monosaccharide moiety. The LC-ESI-MS approach should be a valuable tool for characterization of mucin glycosylation and alterations during pathological conditions.

Lithium evokes a more pronounced natriuresis when administered orally than when given intravenously to salt-depleted rats.

The effects on renal sodium excretion of giving lithium chloride (LiCl; 0.75 mmol per kg body mass) by gavage or intravenously were investigated. The experiments were carried out on Wistar-Kyoto (WKY) or spontaneously hypertensive (SHR) rats in metabolic cages. The rats had been on a low-salt diet for 4 days. Urine excretion of water, sodium and potassium was followed before and for 24 h after giving LiCl. An oral dose of LiCl evoked a more pronounced renal sodium excretion in either strain of rat as compared to that following intravenous administration, in agreement with previous observations of the effects of giving sodium chloride. Choline chloride (1.5 mmol per kg body mass) given by gavage to WKY rats or SHR evoked no change in the renal excretion of sodium. Based on the results of the present study and on observations reported in the literature, we propose that the intestinal tract contains a sodium "sensor", which upon activation releases a natriuretic factor to cause renal sodium excretion. The present results indicate that the proposed "sensor" is sensitive to lithium but not chloride ions.

Detection of CD43 (leukosialin) in colon adenoma and adenocarcinoma by novel monoclonal antibodies against its intracellular domain.

CD43 is a leukocyte-associated sialoglycoprotein which is also expressed in human colon adenoma and carcinoma. To obtain monoclonal antibodies (MAbs) that would react with CD43 in a glycosylation-independent way, antibodies were raised against a peptide corresponding to a portion of the CD43 cytoplasmic domain. Hybridomas were screened on paraffin sections from CD43-positive colon tumours. The reactivity of the antibodies with CD43 was verified by Western blot analysis of lysate of CHO cells transfected with human CD43 cDNA and by immunoprecipitation of lysates from CD43+ cell lines. Epitope mapping of antibodies was done using overlapping heptameric peptides. A detailed characterisation of one of the novel antibodies (CD43-3A1) is presented. This antibody reacts with the CD43 protein regardless of its glycosylation in Western blot analysis, immunoprecipitation and immuno-histochemistry of paraffin sections. Immuno-histochemical analysis of paraffin sections from colon adenoma and carcinoma tissues as well as colon cancer cell lines revealed that CD43 was predominantly localised intracellularly, in contrast to leukocyte-type cells. The MAb reacted more efficiently with paraffin-embedded colon adenoma and carcinoma cells than previously characterised CD43-specific antibodies. This should facilitate the evaluation of a potential role of CD43 during cancer development.

Lack of evidence for an immunosuppressive role for MUC1.

The in vitro anti-proliferative properties of various supernatants from MUC1-expressing cell lines and of purified preparations of MUC1 were evaluated. We have observed that supernatants from the MUC1-and MUC3-positive cell line T47D, but not from the MUC1- and MUC4-positive cell line MCF7, were able to inhibit proliferation of cells from various haematopoietic cell lines. Although the activity of T47D supernatants could be abrogated by immunodepletion of MUC1, immunopurified MUC1 from T47D was unable to inhibit cell proliferation. Significantly, supernatants from mouse 3T3 cells transfected with a secreted form of MUC1 or from BHK-21 cells infected with a recombinant vaccinia virus coding for the secreted form of MUC1, as well as preparations of purified MUC1 from bile or urine, were likewise unable to inhibit T cell proliferation. Surprisingly, a crude mixture of bile mucins had a suppressive effect on T cell growth. Our results suggest that other molecules, such as amino sugars or other mucins, which can associate with MUC1, are likely to be responsible for the observed anti-proliferative effects of T47D cells.