AAVs Targeting Human Carbonic Anhydrase IV Enhance Gene Delivery to the Brain

Current gene therapies using natural AAV serotypes have limited use in the brain because they target poorly, require high doses, and pose safety risks. Directed evolution in mice or non-human primates has produced capsids with much better performance in those species, but inter-species differences make it hard to translate these vectors to humans. To overcome this, we engineered AAV capsids to engage human carbonic anhydrase IV (CA-IV), a blood-brain barrier receptor that is largely restricted to brain endothelial cells. Because the AAV binding site on CA-IV is not conserved across species, capsids optimized in mice show reduced activity on the human receptor. We therefore adopted a two-phase strategy: first, we used in vitro selection to remove variants that fail to bind human CA-IV; second, we conducted in vivo selection in transgenic mice expressing human CA-IV in their brain endothelium. Surprisingly, variants that were only weakly enriched in pull-down assays outperformed strong binders in vivo. Our lead candidate, AAV-hCA4-IV77, achieved a 100-fold increase in brain transduction compared with AAV9, providing broad neuronal and astrocytic expression across multiple regions. These results refine our understanding of receptor-targeted capsid design and underscore the therapeutic potential of AAVs that engage human CA-IV.