Adeno-associated virus (AAV) is widely considered a promising vector for therapeutic gene delivery. This promise is based on previous studies assessing AAVs safety and toxicity, ability to infect nondividing cells, elicit a limited immune response and provide long-term gene expression. However, we now find that earlier studies underappreciated the degree of AAV immunogenicity as well as the extent to which genetic background, through regulation of immune responsiveness, influences the duration of gene expression and thereby the effectiveness of AAV-mediated gene therapy. We evaluated antibody responses in 12 mouse strains to AAV serotype 2 (AAV2) and AAV2-expressed transgene products including green fluorescent protein (GFP), human alpha1-antitrypsin and murine interleukin-10. As expected, all immunocompetent mice administered AAV2 developed serologic evidence of immune responsiveness to the virus. However, a previously unidentified serologic prozone effect was observed suggesting that the concentrations of anti-AAV2 antibodies may have historically been subject to marked underestimation. Furthermore, strains with genetic predisposition to autoimmunity (eg, NOD, NZW, MRL-lpr) specifically imparted a functionally deleterious immune response to AAV-delivered transgene products. These findings suggest that more thorough studies of anti-AAV immunity should be performed, and that genetic predisposition to autoimmunity should be considered when assessing AAV efficacy and safety in humans.