Antigen-specific immune tolerance promises to provide safe and effective therapies to prevent type 1 diabetes (T1D). Antigen-specific therapy requires two components: well-defined, clinically relevant autoantigens; and safe approaches to inducing tolerance in T cells specific for these antigens. Proinsulin is a critical autoantigen in both NOD mice, based on knockout mouse studies and induction of immune tolerance to proinsulin preventing disease whereas most antigens cannot, and also in human T1D based on proinsulin-specific T cells being found in the islets of affected individuals and the early appearance of insulin autoantibodies. Effective antigen-specific therapies that prevent T1D in humans have not yet been developed although doubt remains about the best molecular form of the antigen, the dose and the route of administration. Preclinical studies suggest that antigen specific therapy is most useful when administered before onset of autoimmunity but this time-window has not been tested in humans until the recent "pre-point" study. There may be a 'window of opportunity' during the neonatal period when 'vaccine' like administration of proinsulin for a short period may be sufficient to prevent diabetes. After the onset of autoimmunity, naive antigen-specific T cells have differentiated into antigen-experienced memory cells and the immune responses have spread to multiple antigens. Induction of tolerance at this stage becomes more difficult although recent studies have suggested generation of antigen-specific TR1 cells can inhibit memory T cells. Preclinical studies are required to identify additional 'help' that is required to induce tolerance to memory T cells and develop protocols for effective therapy in individuals with established autoimmunity.