Quantifying and mitigating motor phenotypes induced by antisense oligonucleotides in the central nervous system

Antisense oligonucleotides (ASOs) are emerging as a promising class of therapeutics for neurological diseases. When injected directly into cerebrospinal fluid, ASOs distribute broadly across brain regions and exert long-lasting therapeutic effects. However, many phosphorothioate (PS)-modified gapmer ASOs show transient motor phenotypes when injected into the cerebrospinal fluid, ranging from reduced motor activity to ataxia or acute seizure-like phenotypes. Using a behavioral scoring assay customized to reflect the timing and nature of these effects, we show that both sugar and phosphate modifications influence acute motor phenotypes. Among sugar analogs, DNA induces the strongest motor phenotypes while 2′-substituted RNA modifications improve the tolerability of PS ASOs. Reducing the PS content of gapmer ASOs, which contain a stretch of PS-DNA, improves their toxicity profile, but in some cases also reduces efficacy or duration of effect. We show that this acute toxicity is not mediated by major nucleic acid sensing immune pathways. Formulating ASOs with divalent ions before injection and avoiding phosphate-based buffers modestly improved tolerability through mechanisms at least partially distinct from reduced PS content. Overall, our work identifies and quantifies an understudied aspect of oligonucleotide toxicology in the CNS, explores its mechanism, and presents platform-level medicinal chemistry and formulation approaches that improve tolerability of this class of compounds. [Display omitted] Watts and colleagues present a behavioral scoring assay to measure transient motor phenotypes induced by ASOs in the central nervous system. Reducing phosphorothioate content in ASO gapmers improves their tolerability in the CNS. Buffer formulation also affects acute phenotypes, with phosphate-based diluents showing lower safety than HEPES or lactate-based formulations.