Production of a recombinant hybrid molecule of cholera toxin-B-subunit and proteolipid-protein-peptide for the treatment of experimental encephalomyelitis

Mucosal administration of experimental autoimmune encephalomyelitis (EAE)-specific autoantigens can reduce the onset of disease. To examine whether cholera toxin-B-subunit (CTB)-conjugated EAE-specific T-cell epitope can reduce development of the autoimmune disease in mice, we produced a recombinant hybrid molecule of CTB fusion protein linked with proteolipid-protein (PLP)-peptide139-151(C140S) at levels up to 0.1 gram per liter culture media in Bacillus brevis as a secretion-expression system. Amino acid sequencing and GM1-receptor binding assay showed that this expression system produced a uniformed recombinant hybrid protein. EAE was induced in SJL/J mice by systemic administration with the PLP-peptide. When nasally immunized 5 times with 70 microg rCTB PLP-peptide hybrid protein, mice showed a significantly suppressed development of ongoing EAE and an inhibition of both the PLP-peptide-specific delayed-type hypersensitivity (DTH) responses and leukocyte infiltration into the spinal cord. In contrast, all mice given the PLP-peptide alone or the PLP-peptide with the free form of CTB did not suppress the development of EAE and DTH responses. These results suggest that nasal treatment with the recombinant B. brevis-derived hybrid protein of CTB and autoantigen peptide could prove useful in the control of multiple sclerosis.     

Oral administration of a cholera toxin B subunit-insulin fusion protein produced in silkworm protects against autoimmune diabetes

The oral administration of disease-specific autoantigens can induce oral immune tolerance and prevent or delay the onset of autoimmune disease symptoms. Here, we describe the construction of an edible vaccine consisting of a fusion protein composed of cholera toxin B subunit (CTB) and insulin that is produced in silkworm larvae at levels of up to 0.3 mg/ml of hemolymph. The silkworm bioreactor produced this fusion protein vaccine as the pentameric CTB-insulin form, which retained the GM1-ganglioside binding affinity and the native antigenicity of CTB and insulin. Non-obese diabetic mice fed hemolymph containing microgram quantities of the CTB-insulin fusion protein showed a prominent reduction in pancreatic islet inflammation and a delay in the development of symptoms of clinical diabetes. These results demonstrate that the silkworm bioreactor is a feasible production and delivery system for an oral protein vaccine designed to develop immunological tolerance against T-cell-mediated autoimmune diabetes by regulatory T-cell induction.     

Altered B:9-23 insulin, when administered intranasally with cholera toxin adjuvant, suppresses the expression of insulin autoantibodies and prevents diabetes

Insulin peptide B:9-23 is a major autoantigen in type 1 diabetes that contains two distinct CD4 epitopes (B:9-16 and B:13-23). One of the two epitopes, B:13-23, overlaps with a CTL epitope (B:15-23). In this study, we report that the elimination of the CTL epitope from the B:9-23 peptide by amino acid substitution (with alanine) at positions B:16 and 19 (A16,19 altered peptide ligand) or truncation of the C-terminal amino acids from the peptide (B:9-21), neither of which stimulated the proliferation of insulin B:15-23 reactive CD8 T cells, provided significant intranasally induced suppression of diabetes when coadministered with a potent mucosal adjuvant cholera toxin (CT). Intranasal treatment with A16,19 resulted in the elimination of spontaneous insulin autoantibodies, significant inhibition of insulitis and remission from hyperglycemia, and prevented the progression to diabetes. Intranasal administration of native B:9-23/CT or B:11-23/CT resulted in a significant enhancement of insulin autoantibody expression and severity of insulitis and failed to prevent diabetes. Our present study indicates that elimination of the CTL epitope from the B:9-23 peptide was critically important for mucosally induced diabetes prevention. The A16,19 altered peptide ligand, but not other native insulin peptides, suppresses insulin autoantibodies associated with protection from and remission of diabetes.     

Anergy and not clonal ignorance determines the fate of B cells that recognize a physiological autoantigen

Autoantibodies to insulin arise spontaneously in the insulin autoimmune syndrome and in type I diabetes. In addition, administration of insulin to individuals without autoimmune disease routinely results in Abs that bind autologous hormone. These observations and findings in transgenic models of tolerance led to an inference that physiological levels of hormones and growth factors, such as insulin, are not sufficient to induce tolerance in B cells, a state termed clonal ignorance. In contrast, we have discovered that virtually all conventional B cells expressing a low affinity anti-insulin transgene interact with endogenous insulin and are effectively silenced for Ig production and for T cell-dependent immune responses. A fraction of transgenic B cells escapes silencing and functions autonomously to produce insulin Abs that may lower fasting blood sugars similar to an insulin autoimmune syndrome. These B cells have characteristics of a B1-like subset and are depleted by hypotonic peritoneal lysis. These findings question the concept of clonal ignorance and show that physiological concentrations of Ag may effectively silence conventional B cells even when the affinity for autoantigen is low. Self-reactivity may arise in the repertoire because of compartmental differences that govern the fate of B cells and not as a result of true clonal ignorance.     

Expression of cholera toxin B subunit and assembly as functional oligomers in silkworm

The nontoxic B subunit of cholera toxin (CTB) can significantly increase the ability of proteins to induce immunological tolerance after oral administration, when it was conjugated to various proteins. Recombinant CTB offers great potential for treatment of autoimmune disease. Here we firstly investigated the feasibility of silkworm baculovirus expression vector system for the cost-effective production of CTB under the control of a strong polyhedrin promoter. Higher expression was achieved via introducing the partial non-coding and coding sequences (ATAAAT and ATGCCGAAT) of polyhedrin to the 5' end of the native CTB gene, with the maximal accumulation being approximately 54.4 mg/L of hemolymph. The silkworm bioreactor produced this protein vaccine as the glycoslated pentameric form, which retained the GM1-ganglioside binding affinity and the native antigenicity of CTB. Further studies revealed that mixing with silkworm-derived CTB increases the tolerogenic potential of insulin. In the nonconjugated form, an insulin : CTB ratio of 100 : 1 was optimal for the prominent reduction in pancreatic islet inflammation. The data presented here demonstrate that the silkworm bioreactor is an ideal production and delivery system for an oral protein vaccine designed to develop immunological tolerance against autoimmune diabetes and CTB functions as an effective mucosal adjuvant for oral tolerance induction.     

Expression of cholera toxin B subunit and the B chain of human insulin as a fusion protein in transgenic tobacco plants

A DNA construct containing the cholera toxin B subunit (CTB) gene genetically fused to a nucleotide sequence encoding three copies of tandemly repeated diabetes-associated autoantigen, the B chain of human insulin, was produced and transferred into low-nicotine tobaccos by Agrobacterium. Integration of the fusion gene into the plant genome was confirmed by polymerase chain reaction (PCR). The results of immunoblot analysis verified the synthesis and assembly of the fusion protein into pentamers in transgenic tobacco. GM1–ELISA showed that the plant-derived fusion protein retained GM1–ganglioside receptor binding specificity. The fusion protein accounted for 0.11% of the total leaf protein. The production of transgenic plants expressing CTB–InsB₃ offers a new opportunity to test plant-based oral antigen therapy against autoimmune diabetes by inducing oral tolerance.     

Enhancement of DTH response by cholera toxin B subunit inoculated intranasally together with influenza HA vaccine

The effects of B subunit of cholera toxin (CTB) on delayed-type hypersensitivity (DTH) response to influenza vaccine derived from influenza virus A/PR/8/34 (PR-8, HlNl) virus were investigated in B10 mice that were immunized intranasally with both influenza vaccine and CTB. The result showed that intranasal inoculation of this combination augmented DTH response to influenza vaccine, which reached its peak 6 days after inoculation, and also induced accelerated DTH response upon a second inoculation of influenza vaccine alone 4 weeks later, that the cross-reactive DTH response to PR-8 vaccine was elicited by the injection of the different influenza A-type virus vaccine into the footpad of the vaccinated mice, but was not by influenza B-type virus vaccine, that the DTH-mediating T cells were detected selectively in the lungs of mice that received the nasal inoculation of the vaccine and CTB together, but that subcutaneous inoculation of this combination failed to induce DTH-mediating T cells in the lungs. These results, together with the previous papers (Tamura et al, Vaccine 7: 257-262; 314-320, 1989), suggest that CTB could augment both humoral and DTH responses against influenza vaccine in the respiratory mucosal tract.     

Cholera toxin B pretreatment of macrophages and monocytes diminishes their proinflammatory responsiveness to lipopolysaccharide

The cholera toxin B chain (CTB) has been reported to suppress T cell-dependent autoimmune diseases and to potentiate tolerance of the adaptive immune system. We have analyzed the effects of CTB on macrophages in vitro and have found that preincubation with CTB (10 microg/ml) suppresses the proinflammatory reaction to LPS challenge, as demonstrated by suppressed production of TNF-alpha, IL-6, IL-12(p70), and NO (p < 0.01) in cells of macrophage lines. Pre-exposure to CTB also suppresses LPS-induced TNF-alpha and IL-12(p70) formation in human PBMC. Both native and recombinant CTB exhibited suppressive activity, which was shared by intact cholera toxin. In cells of the human monocyte line Mono Mac 6, exposure to CTB failed to suppress the production of IL-10 in response to LPS. Control experiments excluded a role of possible contamination of CTB by endotoxin or intact cholera toxin. The suppression of TNF-alpha production occurred at the level of mRNA formation. Tolerance induction by CTB was dose and time dependent. The suppression of TNF-alpha and IL-6 production could be counteracted by the addition of Abs to IL-10 and TGF-beta. IFN-gamma also antagonized the actions of CTB on macrophages. In contrast to desensitization by low doses of LPS, tolerance induction by CTB occurred silently, i.e., in the absence of a measurable proinflammatory response. These findings identify immune-deviating properties of CTB at the level of innate immune cells and may be relevant to the use of CTB in modulating immune-mediated diseases.     

Structure-based selection of small molecules to alter allele-specific MHC class II antigen presentation

Class II major histocompatibility molecules are the primary susceptibility locus for many autoimmune disorders, including type 1 diabetes. Human DQ8 and I-A(g7), in the NOD mouse model of spontaneous autoimmune diabetes, confers diabetes risk by modulating presentation of specific islet peptides in the thymus and periphery. We used an in silico molecular docking program to screen a large "druglike" chemical library to define small molecules capable of occupying specific structural pockets along the I-A(g7) binding groove, with the objective of influencing presentation to T cells of the autoantigen insulin B chain peptide consisting of amino acids 9-23. In this study we show, using both murine and human cells, that small molecules can enhance or inhibit specific TCR signaling in the presence of cognate target peptides, based upon the structural pocket targeted. The influence of compounds on the TCR response was pocket dependent, with pocket 1 and 6 compounds inhibiting responses and molecules directed at pocket 9 enhancing responses to peptide. At nanomolar concentrations, the inhibitory molecules block the insulin B chain peptide consisting of amino acids 9-23, endogenous insulin, and islet-stimulated T cell responses. Glyphosine, a pocket 9 compound, enhances insulin peptide presentation to T cells at concentrations as low as 10 nM, upregulates IL-10 secretion, and prevents diabetes in NOD mice. These studies present a novel method for identifying small molecules capable of both stimulating and inhibiting T cell responses, with potentially therapeutic applications.     

Dermal enhancement: bacterial products on intact skin induce and augment organ-specific autoimmune disease

The skin is both an essential barrier for host defense and an important organ of immunity. In this study, we show that the application of cholera toxin to intact mouse skin induces and enhances autoimmune diseases affecting organs at distant anatomic sites, whereas its administration by the mucosal route has been reported to have the opposite effect. First, the CNS autoantigen myelin oligodendrocyte glycoprotein 35-55, when applied repeatedly with cholera toxin to the intact skin of healthy C57BL/6 mice, induced relapsing paralysis with demyelinating immunopathologic features similar to multiple sclerosis. Second, the application of cholera toxin in the absence of autoantigen exacerbated the severity of conventional experimental autoimmune encephalomyelitis induced by myelin oligodendrocyte glycoprotein in CFA. Third, the application of cholera toxin to the intact skin of NOD/Lt mice, with or without insulin B peptide 9-23, exacerbated insulitis and T lymphocyte-derived IFN-gamma and IL-4 production in the islets of Langerhans, resulting in an increased incidence and rate of onset of autoimmune diabetes. The data presented in this study highlight the different outcomes of adjuvant administration by different routes. Because dermal application of cholera toxin, and other bacterial products with similar adjuvant activities, is being developed as a clinical vaccination strategy, these data raise the possibility that it could precipitate autoimmune disease in genetically susceptible humans.     

Transgenic plants expressing human glutamic acid decarboxylase (GAD65), a major autoantigen in insulin-dependent diabetes mellitus

Parenteral and oral administration of autoantigens can induce immune tolerance in autoimmune diseases. Prophylactic therapy based on oral administration of human autoantigens is not, however, feasible when sufficient quantities of candidate autoantigens are not available. Transgenic plants that express high levels of recombinant proteins would allow large quantities of autoantigens to be produced at relatively low costs. In addition, transgenic food would provide a simple and direct method of delivering autoantigens. The production and the characterization of transgenic tobacco and carrot plants expressing human GAD65, a major autoantigen in human insulin-dependent diabetes mellitus (IDDM), is reported. Immunogold labeling and electron microscopy of transgenic tobacco tissue shows the selective targeting of human GAD65 to chloroplast tylacoids and mitochondria. In planta expressed GAD65 has a correct immunoreactivity with IDDM-associated autoantibodies and retains enzymatic activity, a finding that suggests a correct protein folding. In transgenic tobacco and carrot the expression levels of human GAD65 varies between 0.01% and 0.04% of total soluble proteins. Transgenic edible plant organs are now available to study the feasibility of inducing immune tolerance in IDDM animals by oral administration of GAD65.     

B lymphocytes treated in vitro with antigen coupled to cholera toxin B subunit induce antigen-specific Foxp3(+) regulatory T cells and protect against experimental autoimmune encephalomyelitis

The ability of activated B cells to protect against various experimental autoimmune or allergic diseases makes them attractive for use in cell-based therapies. We describe an efficient way to generate B cells with strong suppressive functions by incubating naive B cells with a relevant Ag conjugated to cholera toxin B subunit (CTB). This allows most B cells, irrespective of BCR, to take up and present Ag and induces their expression of latency-associated polypeptide (LAP)/TGF-β and after adoptive transfer also their production of IL-10. With OVA as model Ag, when naive T cells were cocultured in vitro with B cells pretreated with OVA conjugated to CTB (OVA/CTB) Ag-specific CD4(+) Foxp3 regulatory T (Treg) cells increased >50-fold. These cells effectively suppressed CD25(-)CD4(+) effector T (Teff) cells in secondary cultures. Adoptive transfer of OVA/CTB-treated B cells to mice subsequently immunized with OVA in CFA induced increase in Foxp3 Treg cells together with suppression and depletion of Teff cells. Likewise, adoptive transfer of B cells pulsed with myelin oligodendrocyte glycoprotein peptide(35-55) (MOGp) conjugated to CTB increased the number of Treg cells, suppressed MOGp-specific T cell proliferation and IL-17 and IFN-γ production, and prevented the development of experimental autoimmune encephalomyelitis. Similar effects were seen when B cells were given "therapeutically" to mice with early-stage experimental autoimmune encephalomyelitis. Our results suggest that B cells pulsed in vitro with relevant Ag/CTB conjugates may be used in cell therapy to induce Ag-specific suppression of autoimmune disease.     

Suppression of hyperglycemia in NOD mice after inoculation with recombinant vaccinia viruses

In autoimmune (type 1) diabetes, autoreactive lymphocytes destroy pancreatic β-cells responsible for insulin synthesis. To assess the feasibility of gene therapy for type 1 diabetes, recombinant vaccinia virus (rVV) vectors were constructed expressing pancreatic islet autoantigens proinsulin (INS) and a 55-kDa immunogenic peptide from glutamic acid decarboxylase (GAD), and the immunomodulatory cytokine interleukin (IL)-10. To augment the beneficial effects of recombinant virus therapy, the INS and GAD genes were fused to the C terminus of the cholera toxin B subunit (CTB). Five-week-old non-obese diabetic (NOD) mice were injected once with rVV. Humoral antibody immune responses and hyperglycemia in the infected mice were analyzed. Only 20% of the mice inoculated with rVV expressing the CTB::INS fusion protein developed hyperglycemia, in comparison to 70% of the mice in the uninoculated animal group. Islets from pancreatic tissues isolated from euglycemic mice from this animal group showed no sign of inflammatory lymphocyte invasion. Inoculation with rVV producing CTB::GAD or IL-10 was somewhat less effective in reducing diabetes. Humoral antibody isotypes of hyperglycemic and euglycemic mice from all treated groups possessed similar IgG1/IgG2c antibody titer ratios from 19 to 32 wk after virus inoculation. In comparison with uninoculated mice, 11-wk-old NOD mice injected with virus expressing CTB::INS were delayed in diabetes onset by more than 4 wk. The experimental results demonstrate the feasibility of using rVV expressing CTB::INS fusion protein to generate significant protection and therapy against type 1 diabetes onset and progression.     

Combined insulin B:9-23 self-peptide and polyinosinic-polycytidylic acid accelerate insulitis but inhibit development of diabetes by increasing the proportion of CD4+Foxp3+ regulatory T cells in the islets in non-obese diabetic mice

Insulin peptide B:9-23 is a major autoantigen in type 1 diabetes. Combined treatment with B:9-23 peptide and polyinosinic-polycytidylic acid (poly I:C), but neither alone, induce insulitis in normal BALB/c mice. In contrast, the combined treatment accelerated insulitis, but prevented diabetes in NOD mice. Our immunofluorescence study with anti-CD4/anti-Foxp3 revealed that the proportion of Foxp3 positive CD4⁺CD25⁺ regulatory T cells (Tregs) was elevated in the islets of NOD mice treated with B:9-23 peptide and poly I:C, as compared to non-treated mice. Depletion of Tregs by anti-CD25 antibody hastened spontaneous development of diabetes in non-treated NOD mice, and abolished the protective effect of the combined treatment and conversely accelerated the onset of diabetes in the treated mice. These results indicate that poly I:C combined with B:9-23 peptide promotes infiltration of both pathogenic T cells and predominantly Tregs into the islets, thereby inhibiting progression from insulitis to overt diabetes in NOD mice.     

One amino acid difference is critical for suppression of the development of experimental autoimmune diabetes (EAD) with intravenous injection of insulinB:9-23 peptide

InsulinB:9-23 peptide (insB:9-23) reactive T cells has been reported as crucial for type 1 diabetes. In this study, experimental autoimmune diabetes (EAD) mice, which subcutaneous immunization of ins1 or 2B:9-23 induced autoimmune diabetes in F1(B7.1B6 × BALB/c), was investigated for antigen specific therapy to delete pathogenic T cells. Intravenous injection of ins1 or 2B:9-23 significantly delayed the development of diabetes on the corresponding peptide-induced EAD (ins1EAD or ins2EAD) concomitant with reduced insulitis and insulin autoantibodies expression. Population of Foxp3⁺ CD4⁺ T cell was unchanged whereas the level of anti-insB:9-23 specific IgG_2a but not IgG₁ were specifically decreased, suggesting reduction of pathogenic insB:9-23 reactive T cells. Most interestingly, intravenous administration of ins2B:9-23, whose amino acid sequence had one amino acid difference at position 9 delayed the development of diabetes in both ins1EAD and ins2EAD whereas ins1B:9-23 administration delayed diabetes in the ins1EAD but not ins2EAD, suggesting that one amino acid difference gives critical influence on the effect of intravenous injection of antigenic peptide for type 1 diabetes.     

Isoaspartyl post-translational modification triggers autoimmune responses to self-proteins.

The normal functioning immune system is programmed to attack foreign pathogens and other foreign proteins while maintaining tolerance to self-proteins. The mechanisms by which tolerance is broken in the initiation of autoimmunity are not completely understood. In the present study, mice immunized with the murine cytochrome c peptide 90-104 showed no response by the B or T cell compartments. However, immunization with the isoaspartyl form of this peptide, where the linkage of Asp(93) to Leu(94) occurs through the beta-carboxyl group, resulted in strong B and T cell autoimmune responses. Antibodies elicited by immunization with the isoaspartyl form of self-peptide were cross-reactive in binding to both isoforms of cytochrome c peptide and to native cytochrome c self-protein. In a similar manner, immunization of mice with the isoaspartyl form of a peptide autoantigen of human systemic lupus erythematosus (SLE) resulted in strong B and T cell responses while mice maintained tolerance to the normal aspartyl form of self-antigen. Isoaspartyl linkages within proteins are enhanced in aging and stressed cells and arise under physiological conditions. These post-translationally modified peptides may serve as an early immunologic stimulus in autoimmune disease.     

Deleting islet autoimmunity

Even though there are numerous autoantigens for type 1 diabetes, current evidence suggests that a single autoantigen, namely insulin, is responsible for the key initiating event in autoimmunity. If a single autoantigen is necessary for triggering the autoimmune process, then antigen-specific therapy to block or delete the immune response against that autoantigen before epitope spreading occurs, may become a larger focus of future immunotherapeutic strategies. In this article, we review current literature regarding insulin as an autoantigen and potential approaches to deleting insulin-reactive T cells through the use of peptide vaccines and targeted T cell receptor immunizations.     

Active tolerance induction and prevention of autoimmune diabetes by immunogene therapy using recombinant adenoassociated virus expressing glutamic acid decarboxylase 65 peptide GAD(500-585)

Tolerance induction of autoreactive T cells against pancreatic beta cell-specific autoantigens such as glutamic acid decarboxylase 65 (GAD65) and insulin has been attempted as a method to prevent autoimmune diabetes. In this study, we investigate whether adenoassociated virus (AAV) gene delivery of multiple immunodominant epitopes expressing GAD(500-585) could induce potent immune tolerance and persistently suppress autoimmune diabetes in NOD mice. A single muscle injection of 7-wk-old female NOD mice with rAAV/GAD(500-585) (3 x 10(11) IU/mouse) quantitatively reduced pancreatic insulitis and efficiently prevented the development of overt type I diabetes. This prevention was marked by the inactivation of GAD(500-585)-responsive T lymphocytes, the enhanced GAD(500-585)-specific Th2 response (characterized by increased IL-4, IL-10 production, and decreased IFN-gamma production; especially elevated anti-GAD(500-585) IgG1 titer; and relatively unchanged anti-GAD(500-585) IgG2b titer), the increased secretion of TGF-beta, and the production of protective regulatory cells. Our studies also revealed that peptides 509-528, 570-585, and 554-546 in the region of GAD(500-585) played important roles in rAAV/GAD(500-585) immunization-induced immune tolerance. These data indicate that using AAV, a vector with advantage for therapeutic gene delivery, to transfer autoantigen peptide GAD(500-585), can induce immunological tolerance through active suppression of effector T cells and prevent type I diabetes in NOD mice.     

T cell autoreactivity to insulin in diabetic and related non-diabetic individuals

T cell autoreactivity to insulin in type I diabetic and related non-diabetic individuals was analyzed. Peripheral T lymphocytes from both insulin-treated diabetic and untreated non-diabetic members of four families were found to proliferate in vitro in response to human insulin. T cell autoreactivity to insulin therefore does not appear to be diagnostic of the onset of type I diabetes. Highest T cell responses to human insulin were usually detected in insulin-dependent type I diabetes patients treated with a mixture of beef and pork insulin than with self insulin, the greater the dose of animal insulin the higher the T cell response. The T cell repertoires for self insulin appear to be similar in diabetics and non-diabetics based on their patterns of T cell reactivity to beef insulin, port insulin, human insulin, and various peptide of human insulin. The autoreactive T cells analyzed recognize two conformational epitopes of human insulin formed by interactions between A chain and B chain residues. One epitope is associated with the A chain loop and is present in the A1-A14/B1-B16 peptide, and the other epitope consists mainly of B chain residues located in the A16-A21/B10-B25 peptide. These two epitopes are present in amphipathic alpha-helical regions of insulin. HLA-DR (DR3, DR4, and DR5) and HLA-DQ (DQw2/DQw3) Ag can restrict these T cell responses to human insulin epitopes. The ability to detect insulin-specific autoreactive T cells in healthy non-diabetic individuals supports the hypothesis that autoreactive lymphocytes do not necessarily elicit autoimmune disease if present in an environment in which their activity is immunoregulated.     

Identification of an MHC class I-restricted autoantigen in type 1 diabetes by screening an organ-specific cDNA library.

Type 1 diabetes is an autoimmune disease in which the insulin-producing pancreatic beta cells are destroyed at an early age by an immune process that involves both CD4 and CD8 T lymphocytes. The identification of autoantigens in diabetes is very important for the design of antigen-specific immunotherapy. By screening a pancreatic islet cDNA library, we have identified the autoantigen recognized by highly pathogenic CD8 T cells in the non-obese diabetic mouse, one of the best animal models for human diabetes. This is the first identification, to our knowledge, of a CD8 T-cell epitope in an autoimmune disease. The peptide recognized by the cells is in the same region of the insulin B chain as the epitope recognized by previously isolated pathogenic CD4 T cells. This has very important implications for the potential use of insulin in preventative therapy.