Experimental autoimmune diabetes is caused by IFN-γ+/CD8+ T cell immunity against a single insulin A-chain epitope

Abstract

Insulin and its primary gene product, preproinsulin (PPI), are considered central autoantigens in human type 1 diabetes (T1D), with potential value for antigen-specific immune therapy. We have previously demonstrated (Diabetes 51:3237–44) that PPI can adversely induce hyperglycemia and experimental autoimmune diabetes (EAD). Here we identify a C-terminal insulin A-chain epitope as the cause of EAD. RIP-B7.1 transgenic mice were immunized by DNA vaccination using PPI-expressing plasmid vectors, resulting in CD4+/CD8+ insulitis and hyperglycemia in >95% animals after only 2–3 weeks. Adoptive spleen cell transfer shows that autoreactive PPI-specific CD8 T cell clones can be induced in both transgenic and wild type mice, but beta cell co-stimulation (B7.1) is required for diabetes manifestation. Using gene-targeted B7.1 mice we demonstrate that interferon-γ, but not perforin, is required for autoimmune beta cell damage in EAD. Using PPI minigene constructs and a PPI peptide library, we localize and fine-map an immunodominant CD8 T cell epitope to the C-terminus of the A chain of insulin. We identify its MHC restriction as class I Db by a stabilization experiments in TAP-deficient H2 cell lines. After i.m. injection as peptide-DNA complex, this peptide is sufficient to induce insulitis and hyperglycemia in mice. Our data show that experimental autoimmune diabetes results from CD8+ T cell immunity against a Db-restricted T cell epitope of insulin, located in the C-terminal moiety of the insulin A-chain.