In human type 1 diabetes and animal models of the disease, a diverse assortment of immune cells infiltrates the pancreatic islets. CD8(+) T cells are well represented within infiltrates and HLA multimer staining of pancreas sections provides clear evidence that islet epitope reactive T cells are present within autoimmune lesions. These bona fide effectors have been a key research focus because these cells represent an intellectually attractive culprit for beta cell destruction. However, T cell receptors are highly diverse in human insulitis. This suggests correspondingly broad antigen specificity, which includes a majority of T cells for which there is no evidence of islet-specific reactivity. The presence of "non-cognate" T cells in insulitis raises suspicion that their role could be beyond that of an innocent bystander. In this perspective, we consider the potential pathogenic contribution of non-islet-reactive T cells. Our intellectual framework will be that of a criminal investigation. Having arraigned islet-specific CD8(+) T cells for the murder of pancreatic beta cells, we then turn our attention to the non-target immune cells present in human insulitis and consider the possible regulatory, benign, or effector roles that they may play in disease. Considering available evidence, we overview the case that can be made that non-islet-reactive infiltrating T cells should be suspected as co-conspirators or accessories to the crime and suggest some possible routes forward for reaching a better understanding of their role in disease.

Type 1 diabetes (T1D) is an autoimmune disease characterized by the infiltration of lymphocytes into the insulin-producing β-cells in the pancreas. We have isolated live T cells sorted or grown directly from the isolated, handpicked islets of human donors with T1D. We received ~500 islet equivalent EQ of variable purity (10-90%) from 12 donors with T1D (disease duration 0.42-20 years) and from seven control donors and two donors with type 2 diabetes (T2D). A total of 321 T cell lines and clones were derived from the islets of donors with T1D (3 lines from the 9 control donors). These are 131 CD4+ lines and clones, 47 CD8+ lines and 143 lines that contain both CD4+ and CD8+ T cells. From 50 lines and clones examined to date, we have determined the autoreactivity of 19 and have seen a broad repertoire of T cell autoreactivity in the islets, including characterized targets and post-translationally modified targets. Autoreactivity of CD4+ T cell lines was to three different peptides from glutamic acid decarboxylase 65 (GAD; GAD115-127, GAD274-286, GAD555-567), proinsulin76-90, and to chromogranin A or proinsulin expressed by DR4+DQ8+ B cells transduced with lentivirus containing constructs with the open reading frames corresponding to whole autoantigens. Reactivity to modified peptides included the glucose-regulated protein 78 and islet amyloid polypeptide with arginine to citrulline modifications (GRP78292-305(Arg-Cit297) and IAPP65-84(Arg-Cit 73, 81)), deaminations (IA-2545-562(Gln-Glu 548, 551, 556), and to several insulin hybrid peptides. These autoreactive CD4+ T cell lines and clones secreted only pro-inflammatory cytokines (IFN-γ, TNFα) upon peptide stimulation. For CD8+ T cells from islets, from one donor with T1D, we saw binding of a pool of HLA-A2 pentamers loaded with insulin B10-18, IA-2797-805 and insulin specific glucose-6-phosphatase catalytic subunit related protein, IGRP265-273. These results have implications for the development of successful prevention and reversal therapeutic strategies in T1D.