Insulin-dependent diabetes mellitus (IDDM) is thought to result from chronic, cell-mediated, autoimmune islet damage. Our aim was to identify the earliest T-cell autoantigen in IDDM, reasoning that this antigen could be causally involved in the initiation of the disease. Identification of the earliest beta-cell-specific autoantigen is extremely important in allowing advances in prevention and treatment of initial events in the development of inflammatory insulitis that precedes beta-cell destruction and overt diabetes. Therefore, we analyzed the proliferative responses of peripheral T-cells from young, female nonobese diabetic (NOD) mice to extracts of pancreatic beta-cell lines. We were able to demonstrate that T-cells responsive to beta-cell antigens exist in the peripheral lymphoid tissue of these mice in the absence of deliberate priming before the manifestation of histologically detectable insulitis. T-cell lines and clones isolated from the peripheral lymphatic tissues of young, unimmunized, female NOD mice were also shown to react with extracts of beta-cells. Fractionation of the beta-cell extracts showed that these T-cell clones recognized multiple beta-cell-specific autoantigens but none of the previously reported putative autoantigens (glutamic acid decarboxylase [GAD]65, GAD67, Hsp65, insulin, ICA 69, carboxypeptidase-H, and peripherin). Thus, we can conclude that these responses are specific for novel beta-cell autoantigens. Finally, NOD T-cell proliferative responses were also seen to an extract of human islets suggesting potential shared antigenic determinants between human and mouse beta-cells. Our observation that human and murine beta-cell-specific antigens are conserved offers the possibility that identification of these murine autoantigens may lead to the discovery of the human homologue. This will pave the way toward effective diagnosis and/or immunotherapy to prevent diabetes.