317. Screening S. Aureus CRISPR-Cas9 Paired-Guide RNAs for Efficient Targeted Deletion in Duchenne Muscular Dystrophy

Abstract

Duchenne muscular dystrophy (DMD) is a recessive X-linked neuromuscular disorder that results in progressive muscle degeneration and premature death. Most patients have exonic deletions in the dystrophin gene that result in a frameshift and nonfunctional protein. In contrast, Becker muscular dystrophy (BMD) patients carry a range of exonic deletions in dystrophin that do not disrupt the reading frame, leading to a much milder disease phenotype. Thus, multiplex CRISPR/Cas9 targeted deletions that restore the reading frame could convert DMD genotypes into BMD-like genotypes and potentially treat this disease. Previously, this strategy has been demonstrated in vitro with zinc finger nucleases, TALENs, and S. pyogenes Cas9 leading to restoration of dystrophin expression in DMD patient myoblasts. However, these genome-editing enzymes are limited by the difficulty of delivering large transgenes with viral vectors in vivo. Alternatively, the smaller Cas9 ortholog from S. aureus (SaCas9) can be packaged with paired sgRNAs in an all-in-one AAV vector for in vivo gene therapy. Recently, three groups have demonstrated that AAV-SaCas9 can mediate targeted deletions in mdx mice and restore dystrophin expression. As a first step towards developing a genome editing therapeutic for DMD, we conducted an efficiency screen to identify highly active paired sgRNAs for targeted deletion of exon 51 of the dystrophin gene. First, a sensitive, digital droplet PCR assay to quantify exon 51 deletion was validated with DMD patient samples. An initial pilot study of 15 sgRNAs identified a pair (01+09) that mediated 18% Exon 51 deletion in HEK293T cells three days post-transfection. Next, from the set of 10,553 21-nt SaCas9 guides targeting human DMD introns 50 and 51, we selected 53 sgRNAs (675 pairs) that met the following filtering requirements: an endogenous 5’ G, a 3’ T in the NNGRR(T) PAM, cross-reactivity with the non-human primate (NHP) genome, and no off-by-1 or off-by-2 mismatch sites in the human genome. These filters were selected to improve U6 promoter expression, improve SaCas9 cleavage efficiency, enable pre-clinical animal model studies, and minimize off-target editing concerns, respectively. Notably, the NHP cross-reactivity requirement skewed the targeted deletion lengths towards larger sizes, all greater than 12.4kb. In order to test smaller deletions, an additional 174 pairs of human-only sgRNAs were designed for deletion lengths of 0.8-14kb. From the 850 guide pairs screened, 78 pairs with a deletion efficiency Z-score >1.5 were selected for follow-up validation. Interestingly, a few individual sgRNAs consistently appeared among the top performing pairs, regardless of whether they were paired with a generally less effective partner sgRNA. The top hit (guides 68+84), a human-only guide pair with a deletion size of 2.3kb, demonstrated a reproducible deletion efficiency of 32%. The 6 best human-only guide pairs and the 6 best NHP cross-reactive guide pairs have been cloned together with SaCas9 in all-in-one AAV vectors for testing in ex vivo and in vivo models of DMD.