In vivo cross-species evolution of heart-derived AAVs generates novel capsids with enhanced cardiac tropism (for cardiac gene therapy)

Heart failure (HF) remains a major cause of mortality worldwide, underscoring the urgent need for new treatments. Gene therapy holds promise for transforming HF care by addressing its molecular causes, yet effective application requires selective gene delivery to the heart. However, the widely used serotype AAV9 displays high liver tropism, leading to fatal liver failure, thus limiting its clinical utility. To combine enhanced cardiac specificity and reduce off-target effects, the goal of this project is to engineer novel cardiac-specific AAV capsids through an innovative experimental-bioinformatic platform.

In addition to AAV6 and AAV9, novel heart-derived AAVs (HDV) isolated from human failing myocardium were used for DNA shuffling combined with peptide display techniques to create extensive libraries exceeding 10^8 variants with high producibility and titres. These highly diverse AAV libraries were subjected to two consecutive in-vivo evolution rounds by intravenously injection into small and large animal models (mice, pigs) and by rescuing successfully transduced AAV capsids from the heart. Combined bioinformatic analysis of short- and long-read sequencing was used to map biodistribution and transduction level across the heart and other organs. Enrichment over the two selection rounds was 10⁶-fold, yielding 50 cardiac-enriched capsids. The top 10-performing AAV variants exhibited the highest cardiac transduction relative to other organs (heart/ organ ratio >1), such as the liver and spleen, each displaying unique transduction profiles. Mapping the parental sequences of these novel capsids indicate high shuffling diversity, where we see contribution from all included AAV serotypes. In fact, the highest contribution comes from our novel HDV vectors in comparison to AAV6 and AAV9. Among these, we discovered novel capsids with enhanced specificity for distinct cardiac subregions. The left ventricle- (LV) directed candidate achieved a transduction rate of 71% in the LV, with 5% in the right ventricle (RV), 10% in the atria, and 14% distributed across other organs including the liver, spleen, and kidney. The RV-directed candidate demonstrated 30% transduction in the RV, 24% in the atria, 6% in the LV, and 40% in other organs. Similarly, an atrium-specific candidate reached a transduction rate of 24% in the atria, 7% in the RV, 4% in the LV, and 65% in other organs. Of particular interest, we identified a candidate capable of evenly transducing both the heart and brain, achieving 36% transduction in the heart, 35% in the brain, and 29% in other organs, presenting potential utility for treating diseases that impact both cardiac and neurological systems.

In conclusion, we have established an efficient and adaptive platform for engineering novel AAV capsids tailored specifically for heart-directed gene therapy. In combination with an optimized promoter and therapeutic transgene, they hold significant promise for the treatment of both rare and common cardiac diseases that currently lack efficient treatments, marking a significant stride toward transforming the landscape of HF management.