Search

Aberrant glycosylation of mucins is a common phenomenon associated with oncogenic transformation. We investigated the association between expression of the tumor-associated antigens T, Tn, and sialyl-Tn and polymorphism in the length of the MUC1 mucin tandem repeat in a series of gastric carcinomas. We further evaluated the relevance of MUC1 tandem repeat length on the expression of these tumor-associated carbohydrate antigens (TACAs) using a gastric carcinoma cell line model expressing recombinant MUC1 constructs carrying 0, 3, 9, and 42 repeats. Gastric carcinomas showed a high prevalence of Tn and sialyl-Tn antigens, whereas T antigen was less frequently expressed. The expression of T antigen was significantly higher in gastric carcinomas from patients homozygous for MUC1 large tandem repeat alleles. No significant associations were found for Tn and sialyl-Tn antigens. This novel association was reinforced by the gastric carcinoma cell line model experiments, where de novo expression of T antigen was detected in clones transfected with larger VNTR regions. Our results indicate that polymorphism in the MUC1 tandem repeat influences the expression of TACAs in gastric cancer cells and may therefore allow the identification of subgroups of patients that develop more aggressive tumors expressing T antigen.

Mucins in ulcerative colitis and colon cancer share common properties of reduced sulfation and increased oncofetal carbohydrate antigen expression. It has previously been shown that there is no simple correlation between these changes and the activity of the relevant glycosyl-, sialyl-, and sulfo-transferases. We examined mucin sulfation and expression of oncofetal Thomsen-Friedenreich (TF) antigen (galactosyl beta1-3N-acetylgalactosamine alpha-) in the goblet cell-differentiated human colon cancer cell line LS174T following treatment with bafilomycin A(1, )which raises intra-Golgi pH, or monensin, which disrupts medial-trans Golgi transport. Cells were dual-labeled with sodium [(35)S]-sulfate and D-[6-(3)H(N)]-glucosamine hydrochloride, or labeled with L-[U-(14)C]-threonine alone. Mucin was purified using Sepharose CL-4B gel filtration. Mucin sulfo-Lewis(a) and TF antigen expression were assessed using the F2 anti-sulfo-Lewis(a) monoclonal antibody and peanut agglutinin binding respectively. Bafilomycin (0.01 microM; 48 h) reduced total mucin sulfation, expressed relative to incorporation of glucosamine, to 0.50 +/- 0.04 d.p.m. [(35)S]-sulfate per d.p.m. [(3)H]-glucosamine compared to control, 0.84 +/- 0.05 (p0.001, n = 16). This was accompanied by 50.3 +/- 8.0% increased expression of TF antigen (p0.01) and 50.1 +/- 5.5% decreased expression of sulfo-Lewis(a) (p0.01). The reduced sulfate:glucosamine ratio was largely due to increased incorporation of glucosamine into newly synthesized mucin rather than reduction in total sulfate incorporation. In contrast, monensin only reduced total mucin glycosylation at concentrations0.1 microM and had no significant effect on mucin sulfation or TF expression. Intra-Golgi alkalinization affects mucin glycosylation, resulting in decreased mucin sulfation and increased expression of TF antigen, changes that mimic those seen in cancerous and premalignant human colonic epithelium.

Increased mucosal expression of TF, the Thomsen-Friedenreich oncofetal blood group antigen (galactose beta1-3 N-acetylgalactosamine alpha-) occurs in colon cancer and colitis. This allows binding of TF-specific lectins, such as peanut agglutinin (PNA), which is mitogenic to the colorectal epithelium. To identify the cell surface TF-expressing glycoprotein(s), HT29 and Caco2 colon cancer cells were surface-labeled with Na[(125)I] and subjected to PNA-agarose affinity purification and electrophoresis. Proteins, approximately 110-180 kDa, present in HT29 but not Caco2 were identified by Western blotting as high molecular weight splice variants of CD44 (CD44v). Selective removal of TF antigen by Streptococcus pneumoniae endo-alpha-N-acetylgalactosaminidase substantially reduced PNA binding to CD44v. Immunoprecipitated CD44v from HT29 cell extracts also expressed sialyl-Tn (sialyl 2-6 N-acetylgalactosaminealpha-). Incubation of PNA 15 microg/ml with HT29 cells caused no additional proliferative effect in the presence of anti-CD44v6 mAb. In colon cancer tissue extracts (N = 3) PNA bound to CD44v but not to standard CD44. These data show that CD44v is a major PNA-binding glycoprotein in colon cancer cells. Because CD44 high molecular weight splice variants are present in colon cancer and inflammatory bowel disease tissue but are absent from normal mucosa, these results may also explain the increased PNA reactivity in colon cancer and inflammatory bowel disease. The coexpression of oncofetal carbohydrate antigens TF and sialyl-Tn on CD44 splice variants provides a link between cancer-associated changes in glycosylation and CD44 splicing, both of which correlate with increased metastatic potential.