Ph.D., Faculty of Health Sciences, University of the Witwatersrand, 2006. / Objectives. Autoantibodies to the 65kDa isoform of glutamic acid decarboxylase (GAD65Abs) are accepted markers for type 1 diabetes and, together with autoantibodies to insulin (IAA) and a protein tyrosine phosphatase-like islet cell antigen (IA-2), predict the disease. IAA are often the first autoantibodies detected in type 1 diabetics and can be present before the onset of clinical diabetes. These autoantibodies and their epitopes are however not well characterized. We explored the use of monoclonal antibodies and their recombinant Fab (rFab) as reagents for epitope analysis. Methods and Results. Four rFab specific for insulin were cloned from murine monoclonal antibodies (mAbs) 1E2, HB-126, HB-123, HB-127, and one rFab specific for GAD65 was cloned from human mAb IgG antibody DP-D (derived from autoimmune disease patients), to characterise insulin and GAD65 autoantibodies present in the sera of patients with type 1 diabetes. Only rFab 126 and DP-D showed insulin and GAD65 specific binding, respectively in radiobinding assays. In competition experiments with sera positive for autoantibodies to insulin the rFab 126 significantly reduced the binding to 125I-insulin by sera of type 1 (n=35) and type 1.5 diabetes (or LADA) (n=14) patients (p<0.0001). There was no difference in the competition pattern in IAA positive type 1 diabetes patients (n=35) and IAA positive type 1.5 diabetes patients (n=14). The insulin epitope that the rFab binds to was mapped using competitive radiobinding assays with two monoclonal antibodies (mAb 1 and mAb 125) whose epitopes are on the B chain and A chain loop of insulin, respectively. We found the epitope of this recombinant antibody to be located on the A chain loop of the insulin molecule. The 3-dimensional structure of rFab 123, 126 and DP-D were determined using an automated homology modelling programme. Using the computer programme ‘PatchDock’ we attempted to further map the epitope that rFab 126 binds to on insulin. Of the three models generated, only one supported our findings that rFab 126 binds to the A chain loop of insulin. The binding of GAD65Ab in 61 type 1 diabetes patients to GAD65 was analyzed by competitive radioimmunoassays with rFab DP-D to ascertain disease-specific GAD65Ab binding specificities. The median binding was reduced significantly by rFab DP-D (80%) (p<0.0001). The competition pattern in type 1 diabetes patients was different from that in GAD65Ab-positive type 1.5 diabetes patients (n=44), first degree relatives (n=38), and healthy individuals (n=14) (Padoa et al., 2003). Conclusions. We have shown that rFab specific for insulin and GAD65 can be generated using PCR technology and that such agents can be used to determine the insulin/GAD65 epitopes recognized by autoantibodies from type 1 and 1.5 diabetics. These novel findings with GAD65- and insulin-specific rFab support the view that type 1 diabetes is associated with disease- and epitope-specific GAD65- and insulin-autoantibodies and supports the notion that the middle epitope of GAD65 is disease-specific. These GAD65-specific rFab should prove useful in predicting type 1 diabetes. Furthermore, rFabs may be a novel method for blocking autoimmune responses against β cell autoantigens in type 1 diabetics.