Efficacy and safety of a SOD1-targeting artificial miRNA delivered by AAV9 in mice are impacted by miRNA scaffold selection

Toxic gain-of-function mutations in superoxide dismutase 1 (SOD1) contribute to ∼2-3% of all amyotrophic lateral sclerosis (ALS) cases. Artificial microRNAs (amiRs) delivered by adeno-associated virus (AAV) have been proposed as a potential treatment option to silence SOD1 expression and mitigate disease progression. Primary microRNA (pri-miRNA) scaffolds are used in amiRs to shuttle a hairpin RNA into the endogenous miRNA pathway, but it is unclear whether different primary miRNA scaffolds impact the potency and safety profile of the expressed amiR in vivo. In our process to develop an AAV amiR targeting SOD1, we performed a preclinical characterization of two primary miRNA scaffolds, miR155 and miR30a, sharing the same guide strand sequence. We report that while the miR155-based vector, compared to the miR30a-based vector, leads to a higher level of the amiR and more robust suppression of SOD1 in vitro and in vivo, it also presents significantly greater risks for CNS-related toxicities in vivo. Despite miR30a-based vector showing relatively lower potency, it can significantly delay the development of ALS-like phenotypes in SOD1-G93A mice and increase survival in a dose-dependent manner. This data highlights the importance of scaffold selection in the pursuit of highly efficacious and safe amiRs for RNAi gene therapy.