Beyond the Cut: Long-read sequencing reveals complex genomic and transcriptomic changes in AAV-CRISPR therapy for Duchenne Muscular Dystrophy

Adeno associated virus (AAV)-mediated delivery of CRISPR associated nucleases (AAV-CRISPR) is a promising solution to treat genetic diseases such as Duchenne Muscular Dystrophy (DMD) and is now in early clinical trials. However, genotoxicity and immunogenicity concerns have hindered clinical translation. Due to the complex etiology associated with DMD, the post-transduction consequences of double-stranded breaks induced by AAV-CRISPR in disease models are unclear. This barrier is partially conferred by conventional sequencing methods where common outcomes of AAV-CRISPR editing often escape detection. However, recent reports of novel long-read sequencing approaches permit comprehensive variant detection using a broader sequence context. Here, we comprehensively investigated genomic and transcriptomic post-AAV-CRISPR transduction consequences in myoblast cells and a DMD mouse model following intramuscular and intravenous AAV-CRISPR therapy using both long- and short-read sequencing techniques. Structural variant characterization indicates that unintended on-target large insertions and inversions are common editing outcomes. We demonstrate that combining adaptive sampling with nanopore Cas9-targeted sequencing (AS-nCATS) for long-read quantification of AAV integration is synergistic for detecting difficult-to-amplify editing events. This unbiased data suggests that full-length AAV integration is equally as probable as the on-target deletion. Further, we develop a Nanopore Rapid Amplification of cDNA Ends (nRACE-seq) pipeline for long-read detection of unknown 5’ or 3’ ends of edited transcripts. The nRACE-seq approach effectively detects the presence of AAV- Dmd chimeric transcripts, erroneous splicing events, and off-target AAV integration sites. In summary, our findings offer insights into the adaptation of AAV-CRISPR DSB-mediated therapeutics for monogenic diseases and promote the standardization of CRISPR evaluation. We highlight the importance of coupling polymerase-based and polymerase-free methods in long-read sequencing to assess editing outcomes as the field progresses toward clinical applications.