We previously found that adeno-associated viral vector serotype 2 (AAV-2) muscle gene delivery of GAD 500-585 autoantigen efficiently prevented autoimmune diabetes in NOD mice. Recent reports suggest that AAV vectors based on serotype 1 (AAV-1) transduce murine skeletal muscle much more efficiently than AAV-2, with reported increases in expression ranging from 2 to 1000-fold. To determine whether this increased efficacy of AAV-1 could result in increased therapeutic effects in mice, we constructed rAAV1/GAD 500-585 vectors and compared their effects in preventing autoimmune diabetes in NOD mice with those of rAAV2/GAD 500-585 after muscle injection. rAAV(1)/GAD(500-585) gene therapy prevented diabetes in NOD mice. However, although much higher level of GAD 500-585 expression was found in mice using AAV-1 as gene delivery vector than those using AAV-2, no increased efficiency of AAV-1 vectors were found in their capability to prevent autoimmune diabetes, as higher titers of rAAV1/GAD 500-585 virus (3x10(11)v.g./mouse) were needed to obtain therapeutic effects in NOD mice, a titer not different from that of AAV-2. Protection resulted from rAAV1/GAD 500-585 gene therapy were marked by enhanced Th2 immune response and up-regulated CD4+ Foxp3+T regulatory cells, which might actively suppress effector T cells in NOD mice. As here we found that the therapeutic effects of AAV1 were not positively correlated to it's transduction efficiency, our data suggested that the safety and other factors besides efficiency should be considered when use different AAV serotype to treat autoimmune disease.