Applying precision genome editing to disease-causing variant correction and variant-to-function studies

Abstract

Though the first renditions of DNA-modifying tools such as zinc finger nucleases and transcription activator-like effector nucleases were discovered only approximately a quarter century ago, they have exploded in popularity, especially since the detection of CRISPR Cas proteins in the early 2010s. Broad interest in their utilization for disease-modifying therapies has led to their rapid characterization and evolution. Now, several tools capable of editing the mitochondrial and nuclear genomes exist and are being used clinically for the reduction or amelioration of disease symptoms as well as aberrant pathway correction. In this thesis, I describe the optimization of several such genome editing tools for the study and correction of rare and common disease-causing mutations. In Chapter 2, I describe potential therapeutic correction strategies for Metachromatic Leukodystrophy (MLD). MLD is a rare neurodegenerative disease primarily caused by mutations in the arylsulfatase A (ARSA) gene, but hundreds of such mutations are implicated in the disease’s hallmark progressive demyelination of the central and peripheral nervous system. Treatment options are largely symptom-focused, and the disease is fatal because remyelination strategies do not currently exist. Though a gene therapy exists for MLD, it is not approved for all disease subtypes and can only be used in asymptomatic or early symptomatic individuals, which are a very small minority of patients. Thus, it has very limited utility. We developed base editing (BE) and prime editing (PE) strategies to correct the two most common ARSA mutations known to be causal for MLD. Efficient correction of these mutations was seen in patient-derived fibroblasts with base editing and prime editing. Humanized mouse models containing these mutations, the phenotypes of which are currently being characterized, were also developed to enable future in vivo studies on MLD. I also describe the development and evaluation of a novel, precise, and CRISPR-free tool capable of editing organelle and nuclear DNA without inducing double-stranded breaks. This completely protein-based tool, comprised of a split dsDNA-specific cytidine deaminase (DddA), an uracil glycosylase inhibitor (UGI), and zinc finger pairs, mediates precise C-G to T-A editing of the mitochondrial and nuclear genome, and is aptly named a ZF-DdCBE. We show methods of optimizing editing efficiencies as well as its capability of installing and correcting disease-associated mutations in vitro and in vivo. In Chapter 3, I describe the optimization of prime editing in endothelial cells (ECs). Cardiovascular disease (CVD) is complex and heterogeneous, which has made it difficult to identify pathways, genes, and variants involved in disease susceptibility and progression, and is further complicated by the varying impact of these factors by cell type. Endothelial cell dysfunction is one of the first signs of coronary artery disease (CAD), a type of CVD, thereby making the study of these cells crucial to a more complete understanding of the disease. However, some gene expression levels and variants associated with increased risk for CAD are inversely linked to risk of cerebrovascular diseases like cerebral cavernous malformation (CCM), thus highlighting a need for precise variant-to-function studies. We selected control variants to use to optimize and maximize prime editing (PE) in ECs to serve as proof of concept for editing variants of interest in a relevant cell type. Following extensive component modulation and evaluation of multiple delivery modalities, PE enabled efficient and precise installation of these variants in hTERT-immortalized human aortic endothelial cells (teloHAECs). In Chapter 4, I discuss further characterization of previously prioritized CAD and CCM variants. After using PE to install the CCM variants in teloHAECs, we evaluated their effect on gene expression using RT-qPCR and showed that they robustly altered the expression of target genes. We also used PE to install the CAD variants in teloHAECs and assessed their effect on target gene expression using the novel Variant-EFFECTS (formerly Variant-FlowFISH) method. The high degree of sensitivity of the Variant-EFFECTS pipeline combined with the precision of prime editing allows researchers to quantify small changes in gene expression without the need for clonal cell lines. We determined that the CAD variants can affect the expression of genes of interest, further solidifying their hypothesized role in disease pathology.