In this article, we report the identification of a new autoantigen in type 1 diabetes originating from the exocrine pancreas. This antigen is a pancreatic enzyme termed bile salt-dependent lipase (BSDL). We show that antibodies present in the sera of newly diagnosed type 1 diabetic patients recognize BSDL and more specifically the COOH-terminal mucin-like region of the protein. Therefore, we engineered the COOH-terminal peptide of BSDL and demonstrated that autoreactivity was linked to specific glycosylation sites by at least two glycosyltransferases: the Core 2 beta(1-6)N-acetylglucosaminyltransferase and the alpha(1-3) fucosyltransferase FUT7. We next examined the prevalence of circulating anti-BSDL antibodies in type 1 diabetic patients and found 73.5% positivity (25 sera among 34 patients tested) at onset, whereas only 8.4% of normal individuals (7 of 83) were positive. Within a cohort of first-degree relatives of diabetic patients followed prospectively until development of diabetes, 6 of 19 (31.6%) were also positive. Interestingly, two prediabetic individuals were already positive for anti-BSDL antibodies (Abs), while islet cell cytoplasmic Abs and antibodies to GAD65, IA-2, and insulin were not detected. Anti-BSDL autoantibodies were weakly or not detected in patients suffering from pancreatitis or pancreatic adenocarcinoma or in patients with Graves' disease. Although autoreactivity to BSDL in prediabetic and newly diagnosed diabetic patients might reflect cross-reactivity, our results strongly suggest that in addition to pancreatic beta-cells, acinar cells may be also affected in type 1 diabetes.

No abstract

The feto-acinar pancreatic protein or FAPP, the oncofetal glycoisoform of bile salt-dependent lipase (BSDL), is characterized by the presence of the J28 glycotope recognized by mAbJ28. This fucosylated epitope is carried out by the O-linked glycans of the C-terminal mucin-like region of BSDL. This glycotope is expressed by human tumoral pancreatic tissues and by human pancreatic tumoral cell lines such as SOJ-6 and BxPC-3 cells. However, it is not expressed by the normal human pancreatic tissues and by MiaPaCa-2 and Panc-1 cells. Due to the presence of many putative sites for O-glycosylation on FAPP and BSDL, the structure of the J28 glycotope cannot be attained by classical physical methods. In the first part of the present study, we have determined which glycosyltransferases were differently expressed in pancreatic tumoral cell lines compared to normal tissues, focusing in part on fucosyltransferases (Fuc-T) and core-2 beta6-N-acetylglucosaminyltransferase (Core2GlcNAc-T). Our data suggested that alpha2-Fuc-T activity was decreased in the four cell lines tested (SOJ-6, BxPC-3, MiaPaCa-2, and Panc-1). The alpha(1-3) and alpha(1-4) fucosylations were decreased in tumor cells that do not express the J28 glycotope whereas alpha4-Fuc-T and Core2GlcNAc-T activities were significantly increased in SOJ-6 cells which best expressed the J28 glycotope. Therefore, we wished to gain information about glycosyltransferases involved in the building of this structure by transfecting the cDNA encoding the mucin-like region of BSDL in CHO-K1 also expressing Core2GlcNAc-T and/or FUT3 and/or FUT7 activities. These CHO-K1 cells have been previously transfected with the cDNA encoding Core2GlcNAc-T and/or FUT3 and/or FUT7. Data indicated that the C-terminal peptide of BSDL (Cter) produced by those cells did not carry out the J28 glycotope unless Core2GlcNAc-T activity is present. Further transfection with FUT3 cDNA, increased the antibody recognition. Nevertheless, transfection with FUT3 or FUT7 alone did not generate the formation of the J28 glycotope on the C-terminal peptide. Furthermore, the Cter peptide produced by CHO-K1 cells expressing Core2GlcNAc-T was more reactive to the mAbJ28 after in vitro fucosylation with the recombinant soluble form of FUT3. These data suggested that the J28 glycotope encompasses structures initiated by Core2GlcNAc-T and further fucosylated by alpha3/4-Fuc-T such as FUT3, likely on GlcNAc residues.

No abstract

Numerous cell-based or cell-free systems for study of selectin adhesion use radiolabeled tracers. However, in addition to handling problems associated with the use of radioisotopes, these assays have difficulty relating a number of counts to a number of adherent cells. Here, we describe an assay that uses the natural fluorescence of the green fluorescent protein (GFP) to measure binding of cells to E-selectin. We elaborated an adhesion system composed of a cell monolayer expressing E-selectin ligand to which monodispersed fluorescent Chinese hamster ovary (CHO) cells expressing E-selectin are added. Due to GFP autofluorescence, adhered cells can be easily distinguished from cell monolayers by fluorescence microscopy, and adhesion can be measured by cytofluorometry. We applied this GFP-based adhesion assay to measure the adherence of a pancreatic tumor cell line and found that the binding parameters of these cells satisfy a number of E-selectin-specific criteria.