Site-specific noninvasive delivery of retrograde viral vectors to the brain

Neuronal activity underlies the brain function. Different behaviors, physiological processes, and disorders depend on which neurons are active at a given moment. Treating brain disorders without side effects will require exclusive control of disease-relevant neurons. Traditionally, small molecule drugs could control a subset of neurons that express a molecularly specific receptor. Local noninvasive therapies such as delivery of neuromodulatory agents with focused ultrasound blood-brain barrier opening (FUS-BBBO) also added spatial precision, allowing one to control specific brain regions without surgery. However, the final characteristic of neurons, which other neurons they connect to, remains underexplored as a therapeutic target. If targeting neurons based on their connectivity was possible noninvasively, it would open the doors to broadly deployable precise therapies that can target selected subgroups of neurons within a brain region. Such delivery could be achieved with retrograde-tracing adeno-associated viral vectors (AAVs). For noninvasive delivery with FUS-BBBO, AAV9 has emerged as the most promising serotype. However, its retrograde-tracing version, the AAV9.retro, has not been evaluated for FUS-BBBO delivery. Here, we show that following such noninvasive delivery, AAV9.retro can safely transduce neuronal projections with comparable efficiency to a direct intracranial injection. Compared to AAV8, a naturally occurring vector with low retrograde transduction, AAV9.retro offers superior retrograde transduction and comparable transduction at the site of delivery. Overall, we show that AAV9.retro is a valuable FUS-BBBO gene delivery vector, while also highlighting the surprising possibility of improved specificity of transduction of projections compared to invasive delivery.