(560h) Directed Evolution of Adeno-Associated Virus for Enhanced Evasion of Human Neutralizing Antibodies

Gene delivery vectors based on adeno-associated viruses (AAV) have yielded positive results in numerous preclinical disease models and recently in human clinical trials for several disease targets, leading to the first market approval of a clinical gene therapy. However, the high prevalence of anti-capsid neutralizing antibodies, due to widespread exposure to numerous AAV variants and serotypes within the human population, decrease the efficacy of AAV gene therapy. This pre-existing immunity, as well as the subsequent development of immunity due to vector administration, can impede the broader implementation of AAV gene therapy and should be addressed to build upon successful AAV results in immune privileged sites. Directed evolution has proven to be a powerful approach to generate AAV vectors with novel capabilities, and our results demonstrate that AAV can be engineered to significantly overcome neutralization by anti-AAV antibodies, both in vitro and in vivo.

Starting with an AAV2 cap gene containing point mutations isolated from previous selections using individual human serum samples, the cap gene was then subjected to saturation mutagenesis at additional immunogenic amino acid positions determined by ourselves and others. In addition to AAV2 mutagenesis, a shuffled AAV library was selected for resistance to IVIG neutralization. Of the twelve variants chosen for individual analysis from the saturation mutagenesis and shuffled libraries after nine rounds of screening against human IVIG, all twelve could resist higher antibody concentrations than wild-type AAV2. Variant Shuffle 100-3, which withstood a 35-fold higher in vitro IVIG concentration for neutralization than wild-type AAV2, was still capable of transducing approximately 10% of cells in the presence of 1 mg/mL IVIG. In addition, variant SM 10-2 required a 6-fold higher in vitro IVIG concentration for neutralization than wild-type AAV2. Furthermore, variants Shuffle 100-3 and SM 10-2 showed enhanced transduction in the presence of sera samples from individual patients excluded from a hemophilia B clinical trial. For in vivo analysis, recombinant Shuffle 100-3 and SM 10-2 encoding luciferase were administered via a tail vein injection in naïve mice. SM 10-2 displayed similar in vivo tropism to AAV2, except for 7-fold higher transduction of the heart, 5-fold higher transduction of the lungs, and 4.5-fold lower transduction of the liver. By comparison, Shuffle 100-3 exhibited 4-fold higher transduction of the brain, 3-fold higher transduction of the lungs, and 27-fold higher transduction of muscle than AAV2. The ability of SM 10-2 and Shuffle 100-3 to evade antibody neutralization in vivo was examined by passively immunizing mice with human IVIG prior to AAV injection. SM 10-2 had significantly higher heart, liver, and muscle transduction than AAV2 in the presence of IVIG, as measured by luciferase enzyme activity. Additionally, Shuffle 100-3 had significantly higher heart and muscle transduction compared to AAV2. Analysis of the post-AAV serum from these mice showed that the variants required equal or higher in vitro serum concentrations for neutralization than wild-type AAV. These findings represent the most broadly evasive variants described to date. The isolation of such novel clones resistant to anti-AAV antibodies may enable the future treatment of patients with high antibody titers that are currently ineligible for AAV gene therapy.