Fetal calf serum-primed dendritic cells induce a strong anti-fetal calf serum immune response and diabetes protection in the non-obese diabetic mouse.

In recent years, several investigators have shown that transfer of dendritic cells (DC) prevents diabetes development in non-obese diabetic (NOD) mice. Accumulating evidences showing that DC cultured in medium containing fetal calf serum (FCS) can induce a dominant unspecific immune response in tumor models after i.v. injection prompted us to investigate if the protecting effect of DC on diabetes development in NOD mice might be supported by the induction of an anti-FCS immune response in recipient mice. Five-week-old NOD mice were injected i.v. with FCS-cultured bone marrow-derived DC or PBS as control. Levels of anti-FCS and anti-bovine serum albumin (BSA) antibodies were measured in the serum of recipient mice. Anti-FCS cellular immune responses were also analysed after a single DC injection using in vitro proliferation of splenocytes either in RPMI supplemented with FCS, AIMV-BSA or RPMI containing autologous mouse serum or BSA as a read out. DC injection prevented diabetes development in NOD mice and high titers of anti-FCS and anti-BSA antibodies were detected in serum of all DC-injected mice. Besides, splenocytes isolated from DC-injected mice proliferated vigorously in the presence of bovine proteins in contrast to splenocytes isolated from control mice but removing bovine proteins abrogated the high level of proliferation of those splenocytes suggesting that lymphocytes have been primed against bovine proteins in vivo after DC injection. All together, our data show that DC transfer induced cellular and humoral anti-FCS immune responses in recipient NOD mice suggesting that the protective effect of DC relies on their unspecific immunostimulatory effects.

T-splenocytes from non-obese diabetic mice binding to xenogeneic pancreatic beta-cells in vitro. Implication of the alpha/beta T-cell receptor and of major histocompatibility complex class II molecules from target cells.

As compared to several strains of control mice, NOD mice displayed an increased number (P < 10(-4)) of splenocytes binding in vitro to xenogeneic rat RIN cells or hamster HIT cells, but not to nine non-beta cell lines. The increased binding to RIN cells was abolished by competition with RIN membrane extracts. It was prevented by depletion of Thy 1-2+ splenocytes, and by blocking the T-cell receptor (TCR) complex with anti-CD3 MoAbs, anti-alpha/beta TCR MoAbs, or their F(ab)'2 fragments (P < 10(-3)), but not with anti-gamma/delta TCR MoAbs. Neither anti-V beta 8 nor anti-V beta 6 MoAbs modified the signal. MoAbs against rat MHC class II molecules, but not MoAbs against rat class I molecules, inhibited the increased RIN-adhesion of NOD splenocytes (P < 10(-3)). After 3 h or 8 h of co-incubation, the number of RIN-binding splenocytes was not different between NOD and control mice, and class II molecules were undetectable on RIN cells. Class II+ RIN cells appeared after 20 h of coculture when the increased binding was also observed. When 10,000 rad-irradiated RIN cells were used for the co-incubations, neither class II+ RIN cells nor the increased binding of NOD splenocytes were found. As revealed by immunofluorescence, MoAbs against rat class II molecules cross-reacted with 30% of NOD (but not of control) splenocytes. Conversely, anti-NOD class II MoAbs (but not MoAbs against non-NOD class II molecules) cross-reacted with 20% of RIN cells coincubated with splenocytes. Thus, despite the species barrier, T-splenocytes from NOD mice display an increased adhesion to xenogeneic beta-cells. This binding involves T splenocytes bearing alpha/beta TCRs and RIN cells induced to express MHC class II molecules. MHC restriction may be completely absent in this phenomenon. Alternatively, the rat class II products may be directly recognized by NOD T cells in a xenograft context, and this model may therefore be useful toward the comprehension of some mechanisms leading to the rejection of islet xenotransplants. Finally, because of a cross-reaction with I-Anod, these rat Class II molecules may also either be directly recognized by I-A autoreactive NOD T cells or present RIN peptides to NOD alpha/beta TCRs, and thus would be relevant to the debated ability of beta cells to function as antigen-presenting cells.

Release of carcinoembryonic antigen from human tumor cells by phosphatidylinositol-specific phospholipase C: highly effective extraction and upregulation from LS-174T colonic adenocarcinoma cells.

Carcinoembryonic antigen (CEA), produced by gastrointestinal tumor cells, is anchored to cell membrane by a glycosyl-phosphatidylinositol moiety which can be cleaved with phosphatidylinositol-specific phospholipase C (PI-PLC). We studied the extraction of CEA from living human colon carcinoma (LS-174T, HT-29, COLO-205, and HRT-18) and pancreatic carcinoma (CAPAN) cells by PI-PLC from Bacillus cereus. The total CEA content of LS-174T cells, quantitated by Triton X-114 extraction followed by radioimmunoassay or by immunohistochemistry, was 3.5-fold higher than that of other cells (P < 0.001). The spontaneous release of CEA from LS-174T cells into culture medium was also higher than from other cells (P < 0.001), reaching 620 ng/10(7) cells (approximately 28% of cellular content) after 24 h. Overall, living LS-174T cells were highly susceptible to CEA extraction by PI-PLC, which was dependent on PI-PLC dose and on treatment time, leading in optimal conditions to the solubilization of 4100 ng/10(7) cells after 24 h (approximately 75% of total CEA). After 24 h treatment at the highest PI-PLC dose, cell lines remained viable and growing, and membrane CEA expression was not exhausted but only reduced as compared to untreated cells. At the same time, the amount of CEA solubilized by PI-PLC exceeded the CEA reduction in membranes, suggesting that enzyme treatment increased CEA turnover. This was particularly true for LS-174T cells which maintained 54% of the expression of untreated cells, whereas the amount of CEA extracted by PI-PLC reached 190% of this expression. Growing LS-174T cells thus constitute an effective material for producing high quantities of CEA by PI-PLC cleavage, especially since these cells probably "regenerate" because of enhanced turnover during PI-PLC action, thus allowing continuous CEA production. These experimental conditions also provide an interesting model for studying the modulation of CEA expression and release.

Heparin attenuates low-dose streptozotocin-induced immune diabetes in mice and inhibits the beta-cell binding of T-splenocytes in vitro.

Five low doses (40 mg.kg-1.day-1) of streptozotocin were given to CD-1 mice to induce "immune" diabetes with insulitis. T-splenocytes (L3T4+ and Lyt2+) from streptozotocin-treated mice were previously reported to display in vitro an increased binding for Beta cells, preceding the onset of hyperglycaemia and of insulitis. Since heparin inhibits lymphocyte traffic, displays anti-adhesive properties, and attenuates some cell-mediated immune diseases, we have investigated the effects of heparin and N-desulphated heparin: 1) in vivo on low-dose streptozotocin-induced diabetes and insulitis, and 2) in vitro on the increased binding of T-splenocytes from streptozotocin-treated mice to rat insulinoma (RINm5F) cells. Daily subcutaneous low doses (5 micrograms or 10 micrograms) of heparin induced a delay in onset and a reduction of the severity of hyperglycaemia and insulitis (p less than 0.01), and reduced the incidence of diabetes (p less than 0.01). Similar effects were obtained with 5 micrograms daily doses of N-desulphated heparin devoid of anticoagulant activity. In contrast, lower (1 microgram) or higher (200 micrograms) doses of heparin were ineffective. Heparin (10 micrograms) did not modify the "toxic" diabetes induced by a single high dose (200 mg/kg) of streptozotocin. On the other hand, heparin dose-dependently (0.1 microgram/ml to 500.0 micrograms/ml) inhibited the increased binding of splenocytes from streptozotocin-injected mice to RIN cells as compared to splenocytes from control mice. This in vitro anti-adhesive effect was detected when either splenocytes or RIN cells were pretreated with heparin before their co-incubation, and was also obtained with N-desulphated heparin.(ABSTRACT TRUNCATED AT 250 WORDS)

Monoclonal autoantibodies from insulin-dependent diabetic patients: autoantibodies against beta-cell surface or cytoplasmic antigens.

To investigate the target antigen(s) recognized during the autoimmune process in insulin-dependent diabetes mellitus (IDDM), we produced human monoclonal antibodies by Epstein-Barr virus transformation of peripheral blood lymphocytes from a large number (n = 50) of newly diagnosed IDDM patients. Screening by indirect immunofluorescence assay, using the RINm5F rat insulinoma cell line and eight other human or rat tumour cell lines, was performed to identify monoclonal antibodies that reacted with either membrane or cytoplasmic antigens. Eighteen IgM monoclonal antibodies reacting with cytoplasmic antigens of RIN cells were obtained; 14 of them also showed a staining of the cytoplasm of various non-beta-cell lines, while four displayed a binding restricted to beta-cells among the panel tested. However, among three monoclonal antibodies reacting with the membrane of RIN cells, one (HMD-1) produced an IgG antibody with a binding restricted to the membrane of beta-cells (RIN, HIT, and normal rat islet cells). The membrane antigens of HMD-1 were identified in Western blotting as proteins with molecular weights of 64 and 70 kD. This antibody had no apparent cytotoxic effect on RIN cells. These data suggest that, apart from 'natural autoantibodies,' it is feasible to obtain human monoclonal antibodies from IDDM patients that bind specifically to the beta-cell cytoplasm or to the beta-cell membrane.