Glutamic acid decarboxylase 65 kDa isoform (GAD65) is a major autoantigen in type 1 diabetes (T1D). Although aluminum-formulated GAD65 (GAD-alum) induced preservation of residual insulin secretion in a previous clinical phase II trial, recent phase II and III trials failed to reach their primary end-points. The European phase III trial was therefore closed after 15 months, and the entire study period was completed only for a minority of the patients. This study aimed to characterize GAD65 autoantibodies (GADA) and Tyrosine phosphatase IA-2 autoantibody (IA-2A) levels, GADA IgG1-4 subclass distribution, B-cell frequencies/phenotypes and cytokine secretion. We also assessed whether GAD-alum preserved β-cell function in the small subgroup of Swedish patients who completed the 30 months visit. Serum samples and peripheral blood mononuclear cells (PBMC) were collected at baseline and after 1, 3, 9, 15 and 21 months from the 148 Swedish subjects included in the trial, and also at 30 months from the 45 patients who reached the final visit. Patients were randomly assigned to; i) 4 doses of GAD-alum (4D), ii) 2 doses of GAD-alum followed by two doses of placebo (2D), or iii) 4 doses of placebo.
GADA titers were induced both in the 4D and 2D group compared to placebo, and 4D patients also displayed a higher GADA fold-change after receiving the two additional injections compared to the 2D group. The 4D group switched to a higher frequency of GADA IgG4, associated to a Th2 type response at 9 months, whereas an association between GADA fold-change and GAD65-induced in vitro cytokine secretion was observed in the 2D group. These findings suggest that the humoral response, induced by the 2D treatment, seems to be associated with a GAD65-specific cellular response, while 4D induces a distinct humoral response. Even though GADA titers were elevated, no changes in B-cell frequencies or phenotype were observed in any group. IA-2A levels declined at a similar rate in all groups during the trial.The subgroup of patients who completed the 30 month visit receiving 2 doses of GAD-alum had less decline of both fasting and stimulated C-peptide after 30 months compared to placebo. These results support the concept of GAD-alum treatment, but no specific immune markers have been identified.