Comparative analysis of epitope recognition of glutamic acid decarboxylase (GAD) by autoantibodies from different autoimmune disorders

Abstract

Autoantibodies to GAD, an important marker of the autoimmune process in type I or insulin-dependent diabetes mellitus (IDDM), are also found in non-diabetic individuals with autoimmune polyendocrine syndrome type 1 (APS1), APS2, and stiff man syndrome (SMS). Most IDDM sera contain two distinct GAD antibody specificities, one of which targets an epitope region in the middle-third of GAD65 (IDDM-E1; amino acids 221-359) and one of which targets the carboxy-third of GAD65 (IDDM-E2; amino acids 453-569). Using 11 chimeric GAD65/GAD67 proteins to maintain conformation-dependent epitopes of GAD65, we compared the humoral repertoire of IgG antibodies from an individual with APS2-like disease (b35, b78, and b96) and MoAbs from an IDDM patient (MICA-2, MICA-3, and MICA-4). Neither the APS2 IgG antibodies nor the IDDM MoAbs bind the amino-terminal third of GAD65, but instead target the carboxy-terminal two-thirds of GAD65. Amino acids 270-359 (IDDM-E1) are targeted by one APS2 IgG antibody and MICA-4, while two other APS2 IgG antibodies, MICA-2 and MICA-3, target amino acids 443-585 (IDDM-E2). Using GAD65/67 chimera that span the IDDM-E2 region, we found that MICA-2 binds amino acids 514-528 of GAD65, but two APS2 IgG antibodies require this region and amino acids 529-570. In contrast, the binding of MICA-3 requires two discontinuous amino acid segments of GAD65 (452-513 and 528-569), but not amino acids 514-528. These results indicate that there are both similarities and differences in the humoral response to GAD65 in APS2 and IDDM.

Fig. 1

Mapping of GAD65 epitope regions targeted by insulin-dependent diabetes mellitus (IDDM)-derived or autoimmune polyendocrine syndrome type 2 (APS2)-derived MoAbs. Metabolically labelled GAD65, GAD67, or chimeric GAD65/67 protein was immunoprecipitated with MoAbs as reported previously [20]. Table 1 lists the chimeric GAD proteins. A schematic of the GAD chimeric protein used is shown on the left side of each figure. Above each GAD chimera is the amino acid number for GAD65 or the amino acid number for GAD67 (GAD67 numbers are in parentheses). Each chimeric protein is GAD67 (□) except for a small region of GAD65 (shaded). The dilution of MoAbs was chosen so that a similar amount of GAD65 was immunoprecipitated. The abscissa is the ct/min of GAD protein immunoprecipitated. The background binding (binding of protein A-Sepharose (PAS) to labelled GAD protein in absence of antibody) has been subtracted from each measurement. (a) Binding of MICA-2, MICA-3, MICA-4, b35, b78 and b96 to GAD65/67; the results with b35, b78 and b96 are adapted from [24] and are shown to facilitate comparison with MICA-2, MICA-3 and MICA-4. MICA-2 binding to GAD65 (1–195)/GAD67 (205–594) was not performed. In cases where binding was not different from background binding (for GAD67 and for GAD65 (1–195)/GAD67 (205–594)), the assay (duplicate determination) was performed only once or twice for certain antibodies. The error bars represent ± s.e.m. (b) Binding of b35 (□) and MICA-4 (▪) to chimeric proteins of IDDM-E1 region; a schematic of the region of GAD65 contained in the chimeric protein is shown on the left of the graph. Each chimeric protein is full-length GAD67 except for the small region of GAD65 shown by the shaded region. The bars represent mean ± s.e.m.

Fig. 2

Carboxy terminus regions of GAD65 targeted by insulin-dependent diabetes mellitus (IDDM)-derived or autoimmune polyendocrine syndrome type 2 (APS2)-derived MoAbs. As described for Fig. 1, the chimeric GAD protein is full-length GAD67 except for the shaded areas, which are GAD65. (a) Binding of b78 and b96, MICA-2 and MICA-3 (mean of duplicate determination) to chimeric proteins of IDDM-E2 region; The background binding (binding of protein A-Sepharose (PAS) to labelled GAD protein in absence of antibody) has been subtracted from each measurement. (b) Binding of b78 (□) and b96 (▪) to chimeric proteins of IDDM-E2 region. (c) Binding of MICA-3 to GAD65 amino acids 443–585; MICA-3 was preincubated with either buffer or the unlabeled chimeric GAD protein shown on the left side of the graph. The entire chimeric protein (unlabelled) was GAD67 except for the region shown in the shaded pattern which represents the GAD65 portion of the chimeric protein. After MICA-3 incubation with buffer or the unlabelled chimeric GAD protein shown in the schematic to the left of the graph, the mixture was incubated with a labelled GAD67/65 chimeric protein containing GAD65 amino acids 443–585 (GAD67 (1–451)/GAD65 (443–585)). After the addition of PAS, the amount of labelled GAD protein immunoprecipitated was quantified.

Fig. 3

Amino acid sequence comparison between GAD65 and GAD67 in regions targeted by MoAbs. The amino acid sequence of GAD65 protein (human and rat) and GAD67 protein (human) is shown with amino acid differences underlined. (a) IDDM-E1. (b) IDDM-E2.