710. Optimization of AAV Vector Design for Safe Expression of β-N-Acetylhexosaminidase in the Brain for Tay-Sachs Disease Gene Therapy

The GM2 gangliosidoses are lysosomal storage disorders that encompass both Tay-Sachs and Sandhoff diseases. These diseases are associated with deficiencies in the lysosomal enzyme β-N-acetylhexosaminidase (HexA). These deficiencies result in accumulation of GM2 ganglioside (GM2) in the central nervous system leading to neuronal dysfunction and death. HexA is a heterodimer composed of α and β subunits encoded by HEXA and HEXB genes respectively. Gene therapy approaches using direct injection of AAV vectors into the brain of both small and large disease models (mice, cats, and sheep) have all been successful in treating CNS pathology as well as extending lifespan. These studies utilized two AAVrh8 vectors encoding species-specific alpha and beta subunits of HexA under a CBA promoter with a WPRE (CBA-HexA-WPRE). However, when preclinical safety studies were performed in cynomolgus macaques (cm) using the same strategy, severe neurotoxicity was observed for doses 0.1-3.2E12 vg, which are comparable to those tested in other species on a vg/kg brain weight basis. We hypothesized that the cause of unexpected toxicity was due to high expression of HexA. In order to reduce expression of HexA while maintaining our AAV dose (1.78E12-5.34E13 vg/kg brain weight), we generated a series of new vectors with different combinations of promoters and expression elements with a gradient of HexA expression levels. We tested 7 designs of AAVrh8 vectors encoding cm HexA subunits in athymic nude mice (3.3 E13 vg/kg brain weight). In mice injected with the original vector, AAVrh8-CBA-cmHexA-WPRE, we observed increased levels of reactive astrocytes (GFAP) and activated microglia (Iba1) at the injection site. We used this as a screen to test the other vectors for lower gliosis while expressing HexA activity above normal. Three vector designs emerged and were tested in cynomolgus macaques (n=2, 90 days) infused bilaterally into the thalamus and cerebral lateral ventricle at the intermediate dose (5.34 E12 vg/kg brain weight) as in the first study. The behavior of all monkeys remained normal throughout the study. An abnormal T2 weighted MRI signal was documented at day 90 post-injection in one injection site in a monkey in the cohort with the highest HexA activity (up to 88 fold over normal). This signal was absent in day 30 and 60 brain MRI. Neurohistopathological examination revealed considerable neurodegeneration at this site. The other two cohorts had minimal to no neurotoxicity associated with increased HexA expression (up to 9 fold). Two new AAV vectors have been identified for safe overexpression of HexA in the primate brain.