Endothelial-Tropic AAVs for Genetic Access to Whole-Brain Vasculature in Wild-Type Mouse Strains Following Non-Invasive Systemic Delivery

Abstract

The neurovascular unit (NVU) is a vital yet understudied component of the nervous system. Malfunction of non-neuronal cell types within the NVU, including endothelial cells, can facilitate the progression of neurological disorders (Yu et al, Frontiers in Neuroscience, 2020), but limited options for cell-type specific transgene delivery hamper its study. Adeno-associated virus (AAV) vectors for gene delivery to the brain are commonly administered via intra-cranial injections, resulting in tissue damage and limited, uneven spatial coverage. Systemic AAV delivery provides a non-invasive, brainwide alternative for genetic access. Having engineered vectors that efficiently cross the blood-brain-barrier (BBB) with broad tropism in rodents (e.g. AAV-PHP. eB), we turned our focus to engineering cell-type-specific vectors that could access vasculature without targeting other components of the NVU. Using M-CREATE directed evolution (Kumar et al, Nature Methods, 2020), we identified a family of endothelial-enriched capsid variants, including one named AAV-CAP.X1. Following intravenous (I.V.) injection, AAV-CAP.X1 targets vasculature with high cell-type specificity and efficiency throughout the body, including the brain. After injecting 3E11 vg total of AAV-CAP.X1 packaging CAG-GFP into adult C57BL/6J mice, 97% (+/- 0.8%) of the GFP+ area in the hippocampus are CD31+ (demonstrating specificity), and 73% (+/- 9.1%) of the CD31+ area in the hippocampus is GFP+ (proving efficiency; note that an increased dosage of 1E12 vg per mouse resulted in even greater CD31+ labeling without losing specificity). As AAV-CAP.X1 vascular infectivity in the periphery may complicate applications that focus on brain-specific endothelial transduction, we introduced point mutations on the AAV-CAP.X1 capsid and incorporated microRNA target sites into the cargo genome that successfully de-target AAV-CAP.X1 from the liver without impairing brain transduction. AAV-CAP.X1 can be used across multiple genetically diverse mouse strains, with efficient labeling of both capillaries and arteries in the brains of C57BL/6J, FVB/NJ, CBA/J, and BALB/cJ mice following I.V. administration. We also observed a significant increase in transduction compared to its parent capsid AAV9 on multiple human-derived cell lines in vitro. In its brain-targeted form, AAV-CAP.X1 could be paired with pre-clinical therapeutic cargo both to probe vascular contributions to neurological disease and to inform intervention strategies. More broadly, gene delivery via endothelial-tropic AAV capsids could, in principle, be applied to study diverse pathologies that may benefit from vascular remodeling. Our evolving knowledge regarding vascular pathology in COVID-19 that could underlie generalized organ dysfunction demonstrates the timeliness and potential importance of such vectors.