Integrated computational and experimental immunoengineering of adeno-associated virus capsid T cell epitopes in mice.

Adeno-associated virus (AAV) vectors are widely used in gene therapy, but their immunogenicity remains a significant challenge, limiting long-term efficacy and the feasibility of repeated administration. In this study, we combine computational prediction with experimental validation to engineer AAV9 capsids with reduced immunogenicity. To facilitate this, we developed the Epitope Modification and MHC Prediction (EMMP) pipeline, which systematically automates the evaluation of amino acid substitutions for their predicted effects on major histocompatibility complex (MHC) presentability. Using this pipeline, we modify a CD4⁺ T-cell epitope in the AAV9 capsid that is identified and characterized as a proof-of-concept. Two mutant variants, R312H and R312Q, are selected and evaluated for transduction efficiency in vitro and immune response modulation in vivo. Notably, R312Q shows a significant reduction in T-cell activation and anti-AAV9 antibody production, albeit with a slight reduction in transduction at low multiplicities of infection (MOI). These results demonstrate a rational approach for optimizing AAV vector design, with potential applications for improving the safety and efficacy of gene therapy.