1293 Using Living Human Brain Tissue to Develop Human Cell-Type Specific AAV Gene Therapy Vectors

INTRODUCTION: Gene therapy promises to improve outcomes for our most frustrating pathologies. To date, however, we have no method to deliver such sequence-based therapies selectively into the cell types involved. Virtually all prior work on adeno-associated viruses, the most common vector for neurological gene therapies, has studied capsid tropism in model organisms, primarily mice, despite abundant data showing poor translation across species. METHODS: Patients undergoing neurosurgical procedures are consented for any discarded tissue. After sectioning the tissue thin enough so that it can live for weeks in specialized culture conditions, we add panels of AAV variants. We subsequently use both imaging and single-nuclei sequencing to characterize the cell types each AAV variant is capable of entering. RESULTS: Resected human brain tissue remains living and viable for at least two weeks in specialized culture. Using temporal lobe cortex, we find that transduction with fourteen well-established AAV variants, including AAV2, the most commonly used vector in clinical trials, show quite low transduction of human neurons. For virtually all variants, astrocyte transduction is much greater than neuronal transduction, even for variants that have been designed to target neurons in non-human primates & mice. AAV6 and AAV2-retro transduce greater than 90% of GFAP+ astrocytes. AAV-PHP.s and AAV-DJ show the highest neuronal transduction, but at best transduce 10-15% of neurons. (PHP.s was notably developed in mice to target peripheral sensory neurons & shows poor brain transduction in mice, in stark contrast to human tissue.) Single nuclei sequencing is consistent with the imaging data, & sets the stage for high-throughput screening. CONCLUSIONS: We find that AAV tropism in human tissue varies markedly from mice (& monkeys), underscoring the need to use human tissue when developing targeted AAV vectors.