Biodistribution of Adeno-Associated Virus Gene Therapy Following Cerebrospinal Fluid-Directed Administration

Adeno-associated virus (AAV)-based gene therapies, exemplified by the approved therapy for spinal muscular atrophy, have the potential to deliver disease-course-altering treatments for central nervous system (CNS) indications. However, several clinical trials have reported severe adverse events, including patient deaths following high-dose systemic administration for muscle-directed gene transfer, highlighting the need to explore approaches utilizing lower doses when targeting the CNS. Animal models of disease provide insight into the response to new AAV therapies. However, translation from small to larger animals and eventually to humans is hampered by anatomical and biological differences across the species and their impact on AAV delivery. We performed a literature review of preclinical studies of AAV gene therapy biodistribution following cerebrospinal fluid (CSF) delivery (intracerebroventricular, intra-cisterna magna, and intrathecal lumbar). The reviewed literature varies greatly in the reported biodistribution of AAV following administration into the CSF. Differences between studies, including animal model, vector serotype used, method used to assess biodistribution, and route of administration, among other variables, contribute to differing outcomes and difficulties in translating these preclinical results. For example, only half of the published AAV-based gene therapy studies report vector copy number, the most direct readout following administration of a vector; none of these studies reported details such as the empty:full capsid ratio and quality of encapsidated genome. Analysis of the last decade's literature focusing on AAV-based gene therapies targeting the CNS underscores limitations of the body of knowledge and room for continued research. In particular, there is a need to understand the biodistribution achieved by different CSF-directed routes of administration and determining if specific cell types/structures of interest will be transduced. Our findings point to a clear need for a more systematic approach across the field to align the assessments and elements reported in preclinical research to enable more reliable translation across animal models and into human studies.