UC San Diego

While human genetic diseases can be caused by point mutations, insertions/deletions, chromosomal translocations or copy number variations, point mutations account for 58% of the pathogenic genetic variants causing disease. Programmable nucleases such as CRISPR-Cas are powerful tools but their use for the correction of point mutations in vivo poses some major challenges, namely, their reliance on the inefficient process of homologous recombination, threat of introducing permanent off-target mutations in the genome and immunogenicity due to their prokaryotic origin. In this dissertation, we develop and characterize an RNA editing toolset of human origin for correction of guanosine-to-adenosine mutations and premature stop codons. We engineer guide RNA to recruit exogenously expressed human adenosine deaminase acting on RNA (ADAR) enzymes to target transcripts and catalyze adenosine-to-inosine (guanosine) modifications. In a proof-of-concept study, we repair disease-causing premature stop codons and splice-site mutations in mouse models of Duchenne muscular dystrophy (DMD) and ornithine transcarbamylase (OTC) deficiency respectively, via exogenously delivered ADARs and associated guide RNA. However, exogenous delivery of ADARs leads to transcriptome-wide off-targeting, and additionally, enzymatic activity on certain RNA motifs, such as adenosines flanked by a 5’ guanosine is very low, thus limiting their utility as a transcriptome engineering toolset. To solve these issues, we develop a split-ADAR system with highly improved specificity profiles and also carry out a high throughput mutagenesis screen, identifying ADAR variants with enhanced activity at adenosines flanked by a 5’ guanosine. From a gene therapy perspective, recruitment of endogenous ADAR enzymes for editing a desired transcript creates minimal perturbation for the target cells as compared to exogenously delivered ADARs. Thus, we go on to engineer novel circular guide RNAs to recruit endogenous ADAR enzymes. We demonstrate its therapeutic potential by correcting a premature stop codon in a mouse model of Hurler syndrome via delivery of only circular guide RNA. Since immunogenicity against the delivery vehicle also limits efficacy of gene therapies, we develop a programmable adeno-associated virus (AAV) for gene delivery while also modifying it to evade neutralization by pre-existing antibodies in the serum.