Detection of pancreatic islet 64,000 M(r) autoantigens in insulin-dependent diabetes distinct from glutamate decarboxylase

Patients with insulin-dependent diabetes (IDDM) possess antibodies to islet proteins of M(r)-64,000. Potential autoantigens of this M(r) include glutamate decarboxylase (GAD) and 65 kD heat shock protein. We have detected two distinct antibody specificities in IDDM that bind 50,000 M(r) or 37,000/40,000 M(r) proteolytic fragments of 64,000 M(r) proteins. In this study, we investigated relationships of these proteolytic fragments to GAD and heat shock proteins. Polyclonal antibodies to GAD bound 50,000 M(r) fragments of islet antigen. Recombinant GAD65, but not GAD67, blocked binding to this antigen, suggesting that 50,000 M(r) fragments are derived from islet GAD65. In contrast, GAD antibodies did not recognize 37,000/40,000 M(r) fragments, and neither GAD isoforms blocked autoantibody binding to precursors of these fragments. The 37,000/40,000 M(r) fragments, but not the 50,000 M(r) fragments, were detected after trypsin treatment of immunoprecipitates from insulinoma cells that lacked expression of major GAD isoforms. Antibodies in IDDM did not bind native or trypsinized islet heat shock proteins. Thus, IDDM patients possess antibodies to GAD, but also distinct antibodies to a 64,000 M(r) protein that is not related to known GAD isoforms or heat shock proteins.

Interleukin 4 reverses T cell proliferative unresponsiveness and prevents the onset of diabetes in nonobese diabetic mice

Beginning at the time of insulitis (7 wk of age), CD4+ and CD8+ mature thymocytes from nonobese diabetic (NOD) mice exhibit a proliferative unresponsiveness in vitro after T cell receptor (TCR) crosslinking. This unresponsiveness does not result from either insulitis or thymic involution and is long lasting, i.e., persists until diabetes onset (24 wk of age). We previously proposed that it represents a form of thymic T cell anergy that predisposes to diabetes onset. This hypothesis was tested in the present study by further investigating the mechanism responsible for NOD thymic T cell proliferative unresponsiveness and determining whether reversal of this unresponsiveness protects NOD mice from diabetes. Interleukin 4 (IL-4) secretion by thymocytes from > 7-wk-old NOD mice was virtually undetectable after treatment with either anti-TCR alpha/beta, anti-CD3, or Concanavalin A (Con A) compared with those by thymocytes from age- and sex-matched control BALB/c mice stimulated under identical conditions. NOD thymocytes stimulated by anti-TCR alpha/beta or anti-CD3 secreted less IL-2 than did similarly activated BALB/c thymocytes. However, since equivalent levels of IL-3 were secreted by Con A-activated NOD and BALB/c thymocytes, the unresponsiveness of NOD thymic T cells does not appear to be dependent on reduced IL-2 secretion. The surface density and dissociation constant of the high affinity IL-2 receptor of Con A-activated thymocytes from both strains are also similar. The patterns of unresponsiveness and lymphokine secretion seen in anti-TCR/CD3-activated NOD thymic T cells were also observed in activated NOD peripheral spleen T cells. Exogenous recombinant (r)IL-2 only partially reverses NOD thymocyte proliferative unresponsiveness to anti-CD3, and this is mediated by the inability of IL-2 to stimulate a complete IL-4 secretion response. In contrast, exogenous IL-4 reverses the unresponsiveness of both NOD thymic and peripheral T cells completely, and this is associated with the complete restoration of an IL-2 secretion response. Furthermore, the in vivo administration of rIL-4 to prediabetic NOD mice protects them from diabetes. Thus, the ability of rIL-4 to reverse completely the NOD thymic and peripheral T cell proliferative defect in vitro and protect against diabetes in vivo provides further support for a causal relationship between this T cell proliferative unresponsiveness and susceptibility to diabetes in NOD mice.

Thymic T cell anergy in autoimmune nonobese diabetic mice is mediated by deficient T cell receptor regulation of the pathway of p21ras activation

Thymic T cell anergy, as manifested by thymocyte proliferative unresponsiveness to antigens expressed in the thymic environment, is commonly believed to mediate the acquisition of immunological self-tolerance. However, we previously found that thymic T cell anergy may lead to the breakdown of tolerance and predispose to autoimmunity in nonobese diabetic (NOD) mice. Here, we show that NOD thymic T cell anergy, as revealed by proliferative unresponsiveness in vitro after stimulation through the T cell receptor (TCR), is associated with defective TCR-mediated signal transduction along the PKC/p21ras/p42mapk pathway of T cell activation. PKC activity is reduced in NOD thymocytes. Activation of p21ras is deficient in quiescent and stimulated NOD T cells, and this is correlated with a significant reduction in the tyrosine phosphorylation of p42mapk, a serine/threonine kinase active downstream of p21ras. Treatment of NOD T cells with a phorbol ester not only enhances their p21ras activity and p42mapk tyrosine phosphorylation but also restores their proliferative responsiveness. Since p42mapk activity is required for progression through to S phase of the cell cycle, our data suggest that reduced tyrosine phosphorylation of p42mapk in stimulated NOD T cells may abrogate its activity and elicit the proliferative unresponsiveness of these cells.

p21ras and protein kinase C function in distinct and interdependent signaling pathways in C3H 10T1/2 fibroblasts

Both p21ras and protein kinase C (PKC) are believed to function downstream of plasma membrane-associated tyrosine kinases in cellular signal transduction pathways. However, it has remained controversial whether they function in the same pathway and, if so, what their relative position and functional relationship in such a pathway are. We investigated the possibilities that p21ras and PKC function either upstream or downstream of each other in a common linear pathway or that they function independently in colinear signal pathways. Either decreased expression of endogenous normal ras in fibroblasts transfected with an inducible antisense ras construct or overexpression of a mutant ras gene reduced the capacity of the phorbol ester tetradecanoyl phorbol acetate to trigger expression of the tetradecanoyl phorbol acetate-responsive and ras-dependent reporter gene osteopontin (OPN). PKC depletion decreased basal OPN mRNA levels, and the overexpression of ras restored OPN expression to the level of non-PKC-depleted cells. We propose a model in which ras and PKC function in distinct and interdependent signaling pathways.

Processing and presentation of insulin. III. Insulin degrading enzyme: a neutral metalloendoproteinase that is non-homologous to classical endoproteinases mediates the processing of insulin epitopes for helper T cells

Presentation of a protein antigen to T cells generally requires that the antigen be enzymatically processed into an immunogenic peptide(s). The identification of a protease(s) and its mechanism of action in the proteolysis of such an antigen is therefore a primary goal in the study of antigen processing. We show here that insulin degrading enzyme (IDE), a neutral thiol metalloendoproteinase that is structurally non-homologous to the classical metallo, thiol, acid, or serine proteinases, is relatively specific in its proteolytic activity for insulin and digests human insulin (H(I)) into peptides that are presented by murine TA3 B cell antigen presenting cells (APCs) to HI/I-Ad-reactive T cells. These peptides are, however, not presented by fixed TA3 APCs. Anti-IDE mAbs, after their internalization by TA3 cells, significantly inhibit the presentation of H(I) by these APCs. Immunoblotting experiments demonstrate that this inhibition is mediated by the reactivity of these mAbs with a 110 kDa protein, the known M(r) of IDE. These data show that IDE is an endoproteinase that is involved in the processing of insulin and that this IDE-mediated proteolysis is necessary but not sufficient for the recognition of insulin by T cells. Furthermore, we demonstrate that reduction of the disulfide bonds of a pre-processed A-loop containing heterodimeric insulin peptide is required to further process insulin into a T cell epitope.

Antigen-induced Fc receptor-dependent and -independent B cell desensitization. An elevation in [Ca2+]i is not sufficient and protein kinase C activation is not required for these pathways of surface IgM-mediated desensitization

The interaction of an Ag ligand with its B cell surface Ig (sIg) receptor can occur via an FcR-dependent or -independent pathway. We previously found that transfected TNP-specific B cells undergo both Ca2+ signaling and desensitization upon interaction with the thymus-dependent Ag TNP-OVA. Similarly, we showed that these B cells can also be desensitized by cross-linking sIg to the Fc gamma R via the formation of an Ag-antibody bridge. Thus, Ag-specific B cells can be desensitized by two different Ag-dependent events, one mediated by Ag-sIg interaction and the other by sIg-Fc gamma R cross-linking. Inasmuch as Ag-sIg and sIg-Fc gamma R interactions lead to positive and negative signaling, it was of interest to determine whether B cell desensitization mediated by these interactions occurs by one of the well known signaling pathways in B cells. We found that Ag-induced changes in [Ca2+]i could be readily dissociated from Ag-induced desensitization, indicating that a Ca(2+)-independent pathway is likely responsible for this pathway of desensitization. To determine if PKC plays a role in B cell desensitization mediated by either Ag or sIg-Fc gamma R interaction, PKC was downregulated by long term exposure to 12-O-tetradecanoylphorbol 13-acetate or inhibited by exposure of cells to staurosporine. The PKC down-regulated and inhibited cells underwent similar Ag- and Fc gamma R-dependent desensitization compared to cells containing active PKC. Taken together, these data indicate that Ag-induced desensitization of B cell signaling likely involves an event(s) that occurs either upstream or independent of Ag-induced elevations in [Ca2+]i and PKC activation.

Antigen-induced Fc receptor-dependent and -independent B cell desensitization. An elevation in [Ca2+]i is not sufficient and protein kinase C activation is not required for these pathways of surface IgM-mediated desensitization

The interaction of an Ag ligand with its B cell surface Ig (sIg) receptor can occur via an FcR-dependent or -independent pathway. We previously found that transfected TNP-specific B cells undergo both Ca2+ signaling and desensitization upon interaction with the thymus-dependent Ag TNP-OVA. Similarly, we showed that these B cells can also be desensitized by cross-linking sIg to the Fc gamma R via the formation of an Ag-antibody bridge. Thus, Ag-specific B cells can be desensitized by two different Ag-dependent events, one mediated by Ag-sIg interaction and the other by sIg-Fc gamma R cross-linking. Inasmuch as Ag-sIg and sIg-Fc gamma R interactions lead to positive and negative signaling, it was of interest to determine whether B cell desensitization mediated by these interactions occurs by one of the well known signaling pathways in B cells. We found that Ag-induced changes in [Ca2+]i could be readily dissociated from Ag-induced desensitization, indicating that a Ca(2+)-independent pathway is likely responsible for this pathway of desensitization. To determine if PKC plays a role in B cell desensitization mediated by either Ag or sIg-Fc gamma R interaction, PKC was downregulated by long term exposure to 12-O-tetradecanoylphorbol 13-acetate or inhibited by exposure of cells to staurosporine. The PKC down-regulated and inhibited cells underwent similar Ag- and Fc gamma R-dependent desensitization compared to cells containing active PKC. Taken together, these data indicate that Ag-induced desensitization of B cell signaling likely involves an event(s) that occurs either upstream or independent of Ag-induced elevations in [Ca2+]i and PKC activation.

Defective thymic T cell activation by concanavalin A and anti-CD3 in autoimmune nonobese diabetic mice. Evidence for thymic T cell anergy that correlates with the onset of insulitis

In nonobese diabetic (NOD) mice, T cells play a major role in mediating autoimmunity against pancreatic islet beta-cells. We and others previously reported that age-related alterations in the thymic and peripheral T cell repertoire and function occur in prediabetic NOD mice. To study the mechanism responsible for these T cell alterations, we examined whether a defect exists in the thymus of NOD mice at the level of TCR-mediated signaling after activation by Con A and anti-CD3. We found that thymocytes from NOD mice respond weakly to Con A- and anti-CD3-induced proliferation, compared with thymocytes from control BALB/c, BALB.B, (BALB.B x BALB.K)F1, C57BL/6, and nonobese non-diabetic mice. This defect correlates with the onset of insulitis, because it can be detected at 7 to 8 weeks of age, whereas younger mice displayed a normal T cell responsiveness. Thymic T cells from (NOD x BALB/c)F1 mice, which are insulitis- and diabetes-free, exhibit an intermediate stage of unresponsiveness. This T cell defect is not due to a difference in the level of CD3 and IL-2R expression by NOD and BALB/c thymocytes, and both NOD CD4+ CD8- and CD4- CD8+ mature thymic T cells respond poorly to Con A. BALB/c but not NOD thymic T cells respond to Con A in the presence of either BALB/c or NOD thymic APC, suggesting that the thymic T cell defect in NOD mice is intrinsic to NOD thymic T cells and is not due to an inability of NOD APC to provide a costimulatory signal. The defect can be partially reversed by the addition of rIL-2 to NOD thymocytes. To determine whether a defect in signal transduction mediates this NOD thymic T cell unresponsiveness, we tested whether these cells elevate their intracellular free Ca2+ ion concentration in response to Con A. An equivalent Con A-induced increase in Ca2+ ion concentration in both NOD and BALB/c thymocytes was observed, suggesting a normal coupling between the CD3 complex and phospholipase C in NOD thymocytes. In contrast to their low proliferative response to Con A or anti-CD3, NOD thymocytes respond normally (i.e., as do BALB/c thymocytes) to the combinations of PMA plus the Ca2+ ionophore ionomycin and PMA plus Con A but weakly to Con A plus ionomycin. Our data suggest that the age-related NOD thymocyte unresponsiveness to Con A and anti-CD3 results from a defect in the signaling pathway of T cell activation that occurs upstream of protein kinase C activation.

Influence of antigen processing on thymic T-cell selection

The design of a specific blocking peptide for the immunosuppressive therapy of an autoimmune disease requires the identification of peptides of an autoantigen that are physiologically processed in vivo and bind to MHC-encoded membrane glycoproteins. However, knowledge of how an antigen is physiologically processed by antigen-presenting cells (APC) in vivo, particularly in the thymus, is lacking. It is also unknown whether the processing of an antigen by different APC in the thymus can influence thymic T-cell selection. This is an important consideration for attempts to delete or inactivate autoreactive T cells that elicit autoimmune disease. To address these issues, we investigated the processing of biosynthetically labelled recombinant human insulin (rHI), a model autoantigen, injected into mice and characterized the insulin peptides associated with MHC class II molecules on thymic epithelial cells and dendritic cells. These APC were found to differ in the way they process insulin. The detection of MHC-class-II-bound insulin peptides on the surface of the epithelial cells but not the dendritic cells correlated with their capacity to either present or not present insulin to T cells, respectively. Thus, antigen processing may control the appearance of different peptide-MHC class II complexes on thymic APC that mediate positive and negative selection, and thereby influence the development of the T-cell repertoire. Our findings could have important bearing on the future design of synthetic blocking peptides that reduce or eliminate the onset of autoimmune disease.

Defective thymic T cell activation by concanavalin A and anti-CD3 in autoimmune nonobese diabetic mice. Evidence for thymic T cell anergy that correlates with the onset of insulitis

In nonobese diabetic (NOD) mice, T cells play a major role in mediating autoimmunity against pancreatic islet beta-cells. We and others previously reported that age-related alterations in the thymic and peripheral T cell repertoire and function occur in prediabetic NOD mice. To study the mechanism responsible for these T cell alterations, we examined whether a defect exists in the thymus of NOD mice at the level of TCR-mediated signaling after activation by Con A and anti-CD3. We found that thymocytes from NOD mice respond weakly to Con A- and anti-CD3-induced proliferation, compared with thymocytes from control BALB/c, BALB.B, (BALB.B x BALB.K)F1, C57BL/6, and nonobese non-diabetic mice. This defect correlates with the onset of insulitis, because it can be detected at 7 to 8 weeks of age, whereas younger mice displayed a normal T cell responsiveness. Thymic T cells from (NOD x BALB/c)F1 mice, which are insulitis- and diabetes-free, exhibit an intermediate stage of unresponsiveness. This T cell defect is not due to a difference in the level of CD3 and IL-2R expression by NOD and BALB/c thymocytes, and both NOD CD4+ CD8- and CD4- CD8+ mature thymic T cells respond poorly to Con A. BALB/c but not NOD thymic T cells respond to Con A in the presence of either BALB/c or NOD thymic APC, suggesting that the thymic T cell defect in NOD mice is intrinsic to NOD thymic T cells and is not due to an inability of NOD APC to provide a costimulatory signal. The defect can be partially reversed by the addition of rIL-2 to NOD thymocytes. To determine whether a defect in signal transduction mediates this NOD thymic T cell unresponsiveness, we tested whether these cells elevate their intracellular free Ca2+ ion concentration in response to Con A. An equivalent Con A-induced increase in Ca2+ ion concentration in both NOD and BALB/c thymocytes was observed, suggesting a normal coupling between the CD3 complex and phospholipase C in NOD thymocytes. In contrast to their low proliferative response to Con A or anti-CD3, NOD thymocytes respond normally (i.e., as do BALB/c thymocytes) to the combinations of PMA plus the Ca2+ ionophore ionomycin and PMA plus Con A but weakly to Con A plus ionomycin. Our data suggest that the age-related NOD thymocyte unresponsiveness to Con A and anti-CD3 results from a defect in the signaling pathway of T cell activation that occurs upstream of protein kinase C activation.

Altered processing of human insulin by B lymphocytes from an immunologically insulin-resistant type I diabetic patient

Immunologically insulin resistant (IIR) type I diabetic patients possess significantly elevated levels of anti-insulin serum autoantibodies. We investigated whether altered insulin processing by B lymphocytes contributes to this form of insulin resistance. A comparison was made of the 125I-labelled human insulin (HI) peptides processed by Epstein-Barr virus (EBV)-transformed B lymphocytes derived from HLA-identical and nonidentical IIR with non-IIR type I diabetic patients on insulin therapy and healthy non-diabetic individuals. Several peptides detected in the extracellular, membrane-associated and intracellular compartments of B cells from an IIR type I diabetic patient differed from those found in the corresponding compartments of B cells from two non-IIR type I diabetic patients and two normal individuals. These data suggest that HI is processed differently by B cells from an IIR type I diabetic patient compared with B cells from non-IIR type I diabetic patients and normal individuals. Further, we found that two of the five plasma-membrane associated processed HI peptides on the IIR patients' B cells were absent from the membrane compartments of the other B cell lines examined. Thus, it is possible that one or both of these peptides, unique to the IIR patient's B cells, consist of an immundominant epitope(s) that stimulates the production of insulin autoantibodies which mediate the onset of IIR in type I diabetes.

Altered thymic and peripheral T-lymphocyte repertoire preceding onset of diabetes in NOD mice

Insulitis occurs by 5 wk of age in all NOD mice. However, diabetes is detectable only after 3-5 mo of age and only in approximately 50% of females and 10% of males in our colony. Therefore, it is predictable that changes in the T-lymphocyte repertoire of diabetes-prone mice occur and predispose them to disease. We demonstrate here that an altered (with respect to control BALB/cJ mice) thymic T-lymphocyte maturation reflected by a depletion (approximately 12%) of CD4+CD8+ T lymphocytes and a reciprocal increase in CD4-CD8- T lymphocytes precedes the onset of diabetes. This depletion was detected only approximately 3 mo after insulitis and is manifested by a specific loss (approximately 3%) of immature T lymphocytes bearing V beta 8lo (lo is a relative level of expression) T-lymphocyte receptor. By onset of diabetes, an even greater decrease (approximately 35%) of CD4+CD8+ and a reciprocal increase of CD4-CD8- T lymphocytes were apparent and accompanied by the same depletion (3%) of V beta 8 lo T lymphocytes. Administration of cyclophosphamide (CY), which accelerates the appearance of diabetes in NOD mice, caused similar depletions of CD4+CD8+ and V beta 8lo thymic T lymphocytes. The same alterations in the distribution of these thymic T-lymphocyte subsets were evident even earlier in insulitis- and diabetes-free NON mice, indicating that these changes in thymic T-lymphocyte development may be necessary but not sufficient to give rise to diabetes. Despite the common genetic origin of NOD and NON mice, differences at their MHC-linked and -unlinked loci may account for their differential susceptibility to diabetes.(ABSTRACT TRUNCATED AT 250 WORDS)

Distinct antibody specificities to a 64-kD islet cell antigen in type 1 diabetes as revealed by trypsin treatment

Type 1 diabetes is associated with antibodies that immunoprecipitate a 64-kD islet cell membrane protein from detergent extracts of pancreatic islets. In this study we have determined whether mild trypsin treatment of islet membranes can release fragments of the antigen that bind antibodies in the serum of Type 1 diabetic patients. Partial tryptic proteolysis of [35S]methionine-labeled 64-kD antigen immunoprecipitated from detergent extracts of rat islets resulted in the formation of 50-, 40-, and 37-kD fragments. Similar sized fragments were recovered when sera from diabetic patients were employed to immunoprecipitate polypeptides solubilized by mild trypsin treatment of a particulate fraction of radiolabeled rat islets. Of 27 diabetic patients, 22 possessed antibodies to the 50-kD polypeptide and 21 to the 40- and 37-kD polypeptides. A positive association was found between 64k antibodies and antibodies to the 50-kD fragment but not between 64k antibodies and antibodies to the 40- or 37-kD fragments. Some 64k antibody negative patients possessed antibodies that efficiently immunoprecipitated the latter fragments. Serum from 25 of 27 (93%) diabetic patients immunoprecipitated at least one of the three tryptic polypeptides. One of 20 nondiabetic controls immunoprecipitated a 50-kD polypeptide and all controls were negative for antibodies to 40- and 37-kD fragments. Thus, Type 1 diabetes is associated with the presence of at least two antibody reactivities to distinct determinants of the 64-kD antigen, and some patients may possess antibodies to a cryptic epitope on the detergent-solubilized molecule. These data suggest that the detection of antibodies (present in 93% of patients) to epitopes on tryptic polypeptides of the 64-kD antigen may be of even greater diagnostic value for the onset of Type 1 diabetes than analyses of antibodies reactive with the intact 64-kD antigen.

Antigen-induced Ca2+ signaling and desensitization in B cells

Cross-linking of B cell surface Ig (sIg) by anti-Ig results in transmembrane signaling. However, the capacity of a thymus-dependent (TD) Ag to mediate B cell signal transduction has been less well documented. Therefore, we examined Ag-induced intracellular free calcium concentration [( Ca2+]) in B cells by using TD Ag that would be expected to either cross-link or not cross-link sIgM and/or induce the coupling of sIgM to FcR. Stimulation of mouse TA3 hybridoma B cell transfectants that express the SP6 anti-TNP specific sIgM with either TNP-OVA or anti-IgM antibodies resulted in a maximal fourfold increase in [Ca2+]i. The net increase in [Ca2+]i in response to TNP-OVA was dependent upon both the Ag dose and the TNP:OVA molar ratio. Because occupancy of several cell-surface receptor types leads to a loss of response to subsequent stimulation by ligand (homologous desensitization), we examined the ability of Ag to induce homologous desensitization of sIgM in these B cells. TNP1-OVA at all concentrations tested (up to 500 micrograms/ml) did not lead to any change in [Ca2+]i or desensitization. Cross-linking of TNP1-OVA (10 micrograms/ml) with F(ab')2 of anti-OVA antibody induced both a rise in [Ca2+]i and homologous desensitization of sIg, suggesting that cross-linking of sIgM by Ag is sufficient to induce both these processes. TNP6-OVA at a concentration of 10 micrograms/ml induced changes in [Ca2+]i and partially desensitized TNP-specific B cells to stimulation by anti-IgM. Interestingly, a high dose (180 micrograms/ml) of TNP6-OVA stimulated minimal changes in [Ca2+]i yet did not lead to desensitization. However, cross-linking of TNP6-OVA at this high dose with F(ab')2 of rabbit anti-OVA elevated [Ca2+]i and elicited partial desensitization. Complete desensitization of sIgM by Ag was achieved when intact (Fc-containing) anti-OVA antibody was used, suggesting that the FcR can play a role in desensitization. Ag- and antibody-mediated desensitization was not caused by steric hindrance of sIg. Thus, we have observed two forms of Ag-induced desensitization of sIgM, both of which involve sIg cross-linking and one of which is mediated by the physiologic coupling of sIg to FcR.

Antigen-induced Ca2+ signaling and desensitization in B cells

Cross-linking of B cell surface Ig (sIg) by anti-Ig results in transmembrane signaling. However, the capacity of a thymus-dependent (TD) Ag to mediate B cell signal transduction has been less well documented. Therefore, we examined Ag-induced intracellular free calcium concentration [( Ca2+]) in B cells by using TD Ag that would be expected to either cross-link or not cross-link sIgM and/or induce the coupling of sIgM to FcR. Stimulation of mouse TA3 hybridoma B cell transfectants that express the SP6 anti-TNP specific sIgM with either TNP-OVA or anti-IgM antibodies resulted in a maximal fourfold increase in [Ca2+]i. The net increase in [Ca2+]i in response to TNP-OVA was dependent upon both the Ag dose and the TNP:OVA molar ratio. Because occupancy of several cell-surface receptor types leads to a loss of response to subsequent stimulation by ligand (homologous desensitization), we examined the ability of Ag to induce homologous desensitization of sIgM in these B cells. TNP1-OVA at all concentrations tested (up to 500 micrograms/ml) did not lead to any change in [Ca2+]i or desensitization. Cross-linking of TNP1-OVA (10 micrograms/ml) with F(ab')2 of anti-OVA antibody induced both a rise in [Ca2+]i and homologous desensitization of sIg, suggesting that cross-linking of sIgM by Ag is sufficient to induce both these processes. TNP6-OVA at a concentration of 10 micrograms/ml induced changes in [Ca2+]i and partially desensitized TNP-specific B cells to stimulation by anti-IgM. Interestingly, a high dose (180 micrograms/ml) of TNP6-OVA stimulated minimal changes in [Ca2+]i yet did not lead to desensitization. However, cross-linking of TNP6-OVA at this high dose with F(ab')2 of rabbit anti-OVA elevated [Ca2+]i and elicited partial desensitization. Complete desensitization of sIgM by Ag was achieved when intact (Fc-containing) anti-OVA antibody was used, suggesting that the FcR can play a role in desensitization. Ag- and antibody-mediated desensitization was not caused by steric hindrance of sIg. Thus, we have observed two forms of Ag-induced desensitization of sIgM, both of which involve sIg cross-linking and one of which is mediated by the physiologic coupling of sIg to FcR.

Persistence of serum antibodies to 64,000-Mr islet cell protein after onset of type I diabetes

Antibodies to an islet protein of 64,000 Mr (64K antibodies) were measured in 15 diabetic children who were followed prospectively for up to 3 yr after onset of type I (insulin-dependent) diabetes. Of the 15 children, 12 were positive for 64K antibodies at diagnosis. Those patients who were negative for these antibodies at onset remained negative throughout the study. Modest increases in 64K antibodies were observed in 7 patients within 1 mo of diabetes onset, concomitant with an increase in C-peptide concentrations. All antibody-positive patients were still positive at the end of the study, with no significant decrease in antibody levels relative to those at diagnosis, whereas C-peptide concentrations decreased between 3 and 24 mo after onset. Islet cell antibodies, measured by immunochemical staining on sections of rat pancreas, were detected in 9 of 15 patients at onset, whereas only 3 of 11 patients were still positive after 3 yr. In an additional group of 11 patients with diabetes for 6-7 yr, when basal and stimulated C-peptide concentrations were undetectable, 4 patients were still positive for 64K antibodies. These results demonstrate that levels of 64K antibodies persist during the first 3 yr of diabetes, despite declining beta-cell function and decreased immune responses to other islet antigens, but decrease during the next 3-4 yr as the remaining functional beta-cells disappear.

Sulfated beef insulin treatment elicits CD8+ T cells that may abrogate immunologic insulin resistance in type I diabetes

The in vitro responses of T cells from 13 insulin-nonresistant and 1 immunologically insulin-resistant (IIR) type I diabetes patients to sulfated beef insulin (SBI) were analyzed. Insulin A-loop specific CD4+ T cells from these patients did not respond to SBI. After 1 yr of treatment with SBI the IIR patient's T cell and antibody responses to beef, pork, and human insulin progressed from very high to nondetectable levels. This occurred in parallel to the appearance of her insulin-specific CD8+ T cells, which inhibited the response of her A-loop-specific CD4+ T cells to insulin. A transient increase in her CD8+ anti-insulin antibody activity coincided with a relative lack of her CD8+ T cell activity. CD8+ T cells that regulate T cell responsiveness to insulin are probably present but difficult to detect in most type I diabetes patients. These T cells were identified in only 2 of 13 insulin-nonresistant patients who presented with lipoatrophy and insulin allergy, respectively, and who possessed high-titered, anti-insulin antibodies. Our data demonstrate that CD8+ T cells play an important role in controlling peripheral tolerance to insulin and may abrogate IIR in a diabetic patient treated with SBI.

Sequence and gene transfer analyses of HLA-CwBL18 (HLA-C blank) and HLA-Cw5 genes. Implications for the control of expression and immunogenicity of HLA-C antigens

Our previous studies suggested that a serologically undetectable HLA-C blank allele (HLA-CwBL18) is either a variant Cw5 allele or a novel HLA-C Ag. To examine these possibilities, the CwBL18 and Cw5 genes from the TCC (HLA-A1, -A2, -B52, -B18, -Cw-, -Cw-) and QBL (HLA-A26, -B18, -Cw5) EBV-transformed B lymphoblastoid cell lines (LCL) were cloned, sequenced, and transferred into HLA-A, -B, -C null LCL mutant .221 cells. The CwBL18 Ag was detected on the cell surface of CwBL18 transferents by flow cytometry with the anti-class I mAb W6/32 but not by complement-mediated cytotoxicity with currently available HLA-C specific antisera. Sequence analysis of the Cw-BL18 gene indicated that the CwBL18 Ag is "C"-like because it contains all C-locus-specific residues and amino acid replacements commonly found in HLA-C alleles. However, the amino acid sequence of the CwBL18 Ag is unusual; CwBL18 lacks unique allele-specific residues when compared with the sequences of other HLA-C alleles. Moreover, apart from the C-locus-specific differences, the sequence of CwBL18 is identical to the HLA class I consensus sequence. This striking homology of CwBL18 to other HLA class I alleles suggests that CwBL18 may be a weak Ag. Taken together, these data demonstrate that CwBL18 is not a variant Cw5 Ag but is a newly described HLA-C Ag. In contrast to CwBL18, the Cw5 Ag is serologically detectable on the cell surface of Cw5 transferents with HLA-specific allo-antisera. Rather unexpectedly, Cw5 was usually expressed at a lower level than CwBL18 on the surface of .221 transferents as evaluated by W6/32 mAb binding analyses. The sequence of Cw5 revealed several unique amino acid replacements. Two of these substitutions, at residue 35 of the alpha 1 domain and residue 275 of the transmembrane domain, may be responsible for the reduced cell surface expression of Cw5. Additional unique replacements at residues 138 and 177 of the alpha 2 domain suggest that these amino acids may be important in the formation of an epitope recognized by a Cw5-specific antibody.

Sequence and gene transfer analyses of HLA-CwBL18 (HLA-C blank) and HLA-Cw5 genes. Implications for the control of expression and immunogenicity of HLA-C antigens

Our previous studies suggested that a serologically undetectable HLA-C blank allele (HLA-CwBL18) is either a variant Cw5 allele or a novel HLA-C Ag. To examine these possibilities, the CwBL18 and Cw5 genes from the TCC (HLA-A1, -A2, -B52, -B18, -Cw-, -Cw-) and QBL (HLA-A26, -B18, -Cw5) EBV-transformed B lymphoblastoid cell lines (LCL) were cloned, sequenced, and transferred into HLA-A, -B, -C null LCL mutant .221 cells. The CwBL18 Ag was detected on the cell surface of CwBL18 transferents by flow cytometry with the anti-class I mAb W6/32 but not by complement-mediated cytotoxicity with currently available HLA-C specific antisera. Sequence analysis of the Cw-BL18 gene indicated that the CwBL18 Ag is "C"-like because it contains all C-locus-specific residues and amino acid replacements commonly found in HLA-C alleles. However, the amino acid sequence of the CwBL18 Ag is unusual; CwBL18 lacks unique allele-specific residues when compared with the sequences of other HLA-C alleles. Moreover, apart from the C-locus-specific differences, the sequence of CwBL18 is identical to the HLA class I consensus sequence. This striking homology of CwBL18 to other HLA class I alleles suggests that CwBL18 may be a weak Ag. Taken together, these data demonstrate that CwBL18 is not a variant Cw5 Ag but is a newly described HLA-C Ag. In contrast to CwBL18, the Cw5 Ag is serologically detectable on the cell surface of Cw5 transferents with HLA-specific allo-antisera. Rather unexpectedly, Cw5 was usually expressed at a lower level than CwBL18 on the surface of .221 transferents as evaluated by W6/32 mAb binding analyses. The sequence of Cw5 revealed several unique amino acid replacements. Two of these substitutions, at residue 35 of the alpha 1 domain and residue 275 of the transmembrane domain, may be responsible for the reduced cell surface expression of Cw5. Additional unique replacements at residues 138 and 177 of the alpha 2 domain suggest that these amino acids may be important in the formation of an epitope recognized by a Cw5-specific antibody.

Processing and presentation of insulin. II. Evidence for intracellular, plasma membrane-associated and extracellular degradation of human insulin by antigen-presenting B cells

To study the biochemistry of processing of a soluble protein Ag by an APC, we investigated how 125I-labeled human insulin (HI) is processed in situ by TA3 mouse hybridoma B cells. Fractionation of TA3 cells into their extracellular, plasma membrane-associated and intracellular compartments coupled with the use of HPLC enabled us to analyze several peptides derived from each compartment. One HI peptide found in all three compartments is composed of residues A1-A14 disulfide-linked to B7-B26 (A1-A14/B7-B26). The presence of this peptide in the extracellular compartment likely resulted from digestion of HI by an enzyme(s) released from the APC. Extracellular processing of radiolabeled HI was inhibited completely by unlabeled HI and N-ethylmaleimide, an inhibitor of a previously described insulin-specific protease, partially by lysozyme but not by BSA or OVA. This suggests that the enzyme involved in the extracellular processing of insulin is relatively insulin-specific and gives rise to the A1-A14/B7-B26 peptide. The processing of HI both at the plasma membrane and intracellularly was inhibited by chloroquine, monensin, and NH4Cl, suggesting that both intracellular pH changes and endocytic and exocytic events may be required for these compartments to process insulin. Kinetic analyses revealed that the processing of insulin into the A1-A14/B7-B26 peptide is first detected at the plasma membrane then intracellularly and finally in the extracellular compartment. This unlabeled A1-A14/B7-B26 peptide was purified from the extracellular compartment of TA3 APC by HPLC; when presented by TA3 APC this peptide effectively stimulated pork insulin (PI/I-Ad) specific Th cells to secrete IL-2. These data, taken together with the identification of another processed insulin peptide, A7-A11/B7-B26, have enabled us to elucidate the first steps in the biochemical pathway(s) of processing of insulin as an Ag in a B cell APC.

Processing and presentation of insulin. II. Evidence for intracellular, plasma membrane-associated and extracellular degradation of human insulin by antigen-presenting B cells

To study the biochemistry of processing of a soluble protein Ag by an APC, we investigated how 125I-labeled human insulin (HI) is processed in situ by TA3 mouse hybridoma B cells. Fractionation of TA3 cells into their extracellular, plasma membrane-associated and intracellular compartments coupled with the use of HPLC enabled us to analyze several peptides derived from each compartment. One HI peptide found in all three compartments is composed of residues A1-A14 disulfide-linked to B7-B26 (A1-A14/B7-B26). The presence of this peptide in the extracellular compartment likely resulted from digestion of HI by an enzyme(s) released from the APC. Extracellular processing of radiolabeled HI was inhibited completely by unlabeled HI and N-ethylmaleimide, an inhibitor of a previously described insulin-specific protease, partially by lysozyme but not by BSA or OVA. This suggests that the enzyme involved in the extracellular processing of insulin is relatively insulin-specific and gives rise to the A1-A14/B7-B26 peptide. The processing of HI both at the plasma membrane and intracellularly was inhibited by chloroquine, monensin, and NH4Cl, suggesting that both intracellular pH changes and endocytic and exocytic events may be required for these compartments to process insulin. Kinetic analyses revealed that the processing of insulin into the A1-A14/B7-B26 peptide is first detected at the plasma membrane then intracellularly and finally in the extracellular compartment. This unlabeled A1-A14/B7-B26 peptide was purified from the extracellular compartment of TA3 APC by HPLC; when presented by TA3 APC this peptide effectively stimulated pork insulin (PI/I-Ad) specific Th cells to secrete IL-2. These data, taken together with the identification of another processed insulin peptide, A7-A11/B7-B26, have enabled us to elucidate the first steps in the biochemical pathway(s) of processing of insulin as an Ag in a B cell APC.

Influence of the valency and hydrophobicity of an antigen on its efficiency of processing and presentation by antigen-specific B cells

We examined the influence of antigen structure on the efficiency of antigen-specific B-cell processing and presentation. The ability of untransfected and anti-TNP surface IgM (sIgM) bearing transfected TA3 B-hybridoma cells to process and present various conjugates of 2,4,6-trinitrophenyl-pork insulin (TNP-PI) to PI/I-Ad-specific T cells was investigated. Similar antigen concentrations were required for presentation of underivatized antigen by untransfected and transfected cells (antigen-non-specific presentation). Transfected cells present TNP-PI at about a 50-fold lower concentration and TNP-ovalbumin (TNP-OVA), at a 600-fold lower concentration than do untransfected cells (antigen-specific presentation). PI was derivatized at 3 possible residues, A1, B1 and B29, and the different TNP-PI conjugates obtained were separated by hydrophobic interaction high-pressure liquid chromatography. All TNP1-PI conjugates were presented equally by transfected and untransfected cells. Transfected TA3 cells presented a TNP2-PI conjugate derivatized at A1 and B29, TNP2(A1,B29)-PI, at about a 50-fold lower concentration than was required by untrasfected cells. Another TNP2-PI conjugate, derivatized at residues A1 and B1, TNP2(A1,B1)-PI, and a TNP3(A1,B1,B29)-PI conjugate were presented at about the same concentration by transfected vs untransfected B cells. Of the various TNP conjugates tested, only TNP2(A1,B29)-PI, which is more hydrophobic than any of the other TNP-PI conjugates, was processed more efficiently by transfected anti-TNP-specific TA3 cells vs untransfected cells. The efficiency of presentation of these TNP-PI conjugates was directly proportional to their rate of processing. Our data demonstrate that the valency and relative hydrophobicity of an antigen control its binding to sIgM and the efficiency of processing and presentation by B cells.

IL-2 secretion is pertussis toxin sensitive in a T lymphocyte hybridoma

Interaction of specific ligands with TCR initiates a cascade of biochemical events which leads to expression of high affinity IL-2R and subsequent IL-2 secretion. Activation of phospholipase C (PL-C) is considered to be a key event in the initiation of this cascade. However, in addition to this PL-C-dependent pathway, PL-C-independent pathways have been hypothesized. Identification of the steps constituting these PL-C-independent pathways has been difficult because activation of PL-C and the subsequent cascade of events mask the effects of such pathways. Specific inhibitors for PL-C, or mutants defective in, the PL-C pathway would facilitate delineation of alternative activation pathways. We have identified a murine pork insulin/IAd-specific T cell hybridoma, B8P3.11, in which perturbation of the B8P3.11 TCR by either Ag in association with Ia, anti-CD3 antibodies, or a mitogenic lectin does not induce increases in myo-inositol 1,4,5-triphosphate production or cytosolic free calcium, yet it does lead to IL-2 secretion. Treatment of B8P3.11 with pertussis toxin, at concentrations which ADP-ribosylate GTP-binding proteins, inhibits IL-2 secretion. Thus, signal transduction resulting in IL-2 secretion by B8P3.11 likely involves a G protein. In contrast, TCR/ligand interaction activates the PL-C-dependent pathway in LBRM 331A5, a T cell lymphoma. Furthermore, pertussis toxin treatment, which blocks IL-2 secretion by B8P3.11, does not alter IL-2 secretion by LBRM 331A5. However, similar pertussis toxin substrates are present in both cells. Therefore, B8P3.11 T cells should help to elucidate PL-C-independent activation pathways.

IL-2 secretion is pertussis toxin sensitive in a T lymphocyte hybridoma

Interaction of specific ligands with TCR initiates a cascade of biochemical events which leads to expression of high affinity IL-2R and subsequent IL-2 secretion. Activation of phospholipase C (PL-C) is considered to be a key event in the initiation of this cascade. However, in addition to this PL-C-dependent pathway, PL-C-independent pathways have been hypothesized. Identification of the steps constituting these PL-C-independent pathways has been difficult because activation of PL-C and the subsequent cascade of events mask the effects of such pathways. Specific inhibitors for PL-C, or mutants defective in, the PL-C pathway would facilitate delineation of alternative activation pathways. We have identified a murine pork insulin/IAd-specific T cell hybridoma, B8P3.11, in which perturbation of the B8P3.11 TCR by either Ag in association with Ia, anti-CD3 antibodies, or a mitogenic lectin does not induce increases in myo-inositol 1,4,5-triphosphate production or cytosolic free calcium, yet it does lead to IL-2 secretion. Treatment of B8P3.11 with pertussis toxin, at concentrations which ADP-ribosylate GTP-binding proteins, inhibits IL-2 secretion. Thus, signal transduction resulting in IL-2 secretion by B8P3.11 likely involves a G protein. In contrast, TCR/ligand interaction activates the PL-C-dependent pathway in LBRM 331A5, a T cell lymphoma. Furthermore, pertussis toxin treatment, which blocks IL-2 secretion by B8P3.11, does not alter IL-2 secretion by LBRM 331A5. However, similar pertussis toxin substrates are present in both cells. Therefore, B8P3.11 T cells should help to elucidate PL-C-independent activation pathways.