ObjectiveGM2 gangliosidosis is usually fatal by 5 years of age in its 2 major subtypes, Tay‐Sachs and Sandhoff disease. First reported in 1881, GM2 gangliosidosis has no effective treatment today, and children succumb to the disease after a protracted neurodegenerative course and semi‐vegetative state. This study seeks to further develop adeno‐associated virus (AAV) gene therapy for human translation.MethodsCats with Sandhoff disease were treated by intracranial injection of vectors expressing feline β‐N‐acetylhexosaminidase, the enzyme deficient in GM2 gangliosidosis.ResultsHexosaminidase activity throughout the brain and spinal cord was above normal after treatment, with highest activities at the injection sites (thalamus and deep cerebellar nuclei). Ganglioside storage was reduced throughout the brain and spinal cord, with near complete clearance in many regions. While untreated cats with Sandhoff disease lived for 4.4 ± 0.6 months, AAV‐treated cats lived to 19.1 ± 8.6 months, and 3 of 9 cats lived >21 months. Correction of the central nervous system was so effective that significant increases in lifespan led to the emergence of otherwise subclinical peripheral disease, including megacolon, enlarged stomach and urinary bladder, soft tissue spinal cord compression, and patellar luxation. Throughout the gastrointestinal tract, neurons of the myenteric and submucosal plexuses developed profound pathology, demonstrating that the enteric nervous system was inadequately treated.InterpretationThe vector formulation in the current study effectively treats neuropathology in feline Sandhoff disease, but whole‐body targeting will be an important consideration in next‐generation approaches. ANN NEUROL 2023