Enhancement and Application of Prime Editing to Correct Pathogenic Mutations In Vitro and In Vivo

Abstract

Since the identification of DNA as the genetic blueprint for living organisms, researchers have sought to make precise changes in the genomes of living cells to better understand biology and to treat diseases with a genetic component. Although the ability to alter mammalian genomes both in vitro and in vivo has existed for decades, recent advances in CRISPR-Cas-based systems have enabled gene editing with unprecedented precision. The focus of my doctoral thesis has been to advance and apply one such precision gene editing technology, prime editing (PE), for the correction of pathogenic mutations in vitro and in vivo. The first chapter of this thesis reviews our group’s current practices for conducting PE experiments and describes the design and optimization of pegRNAs. We also offer guidelines for how to select the proper PE system (PE1 to PE5, and twinPE) for a given application. Finally, we provide detailed instructions on how to perform PE in mammalian cells. In the second chapter, I demonstrate the systematic multi-tiered development of PE strategies to convert the CFTR F508del mutation, the predominant cause of cystic fibrosis (CF), to wild-type CFTR by precisely inserting the missing nucleotides. Initial attempts with first-generation PE systems achieved low editing efficiencies in human cells (.5%). By combining six recent advances in prime editing— epegRNAs, the PEmax architecture, MLH1dn, strategic silent edits, novel PE6 variants, and dsgRNAs— that each address different bottlenecks in PE efficiency, we increased CFTR F508del precise correction efficiencies 120-fold. Treating immortalized human bronchial epithelial cells homozygous for CFTR F508del with the resulting highly optimized PE system yielded an average of 51% precise correction to wild-type CFTR. In primary CF patient airway epithelial cells from multiple donors, this prime editing treatment yielded 25% average CFTR F508del correction efficiency and restored CFTR ion channel function to >50% of wild type levels, achieving functional rescue comparable to or higher than the outcome of treatment with the combination of small-molecule drugs elexacaftor+tezacaftor+ivacaftor. Importantly, this prime editing strategy exhibited minimal off-target effects and a 3.4-fold higher ratio of precise correction:indel byproducts compared to previous optimized nuclease-mediated homology- directed repair (HDR) approaches. Direct, efficient F508del correction in primary CF patient airway epithelial cells, restoring CFTR function, suggests the feasibility of a durable one-time treatment for CF, and provides a blueprint for optimizing efficient PE strategies to correct other pathogenic gene variants. Finally, in the third chapter I describe the application of PE to address pathogenic mutations in cell and mouse models of alternating hemiplegia of childhood (AHC). AHC is a rare neurological disorder caused by mutations in ATP1A3, which codes for the α3 subunit of the Na+/K+ ATPase, and results in devastating paroxysmal paralytic attacks, developmental delays, and cognitive deficits in AHC patients. In five cell models of individual ATP1A3 mutations that represent 70% of AHC cases, we engineered PE and base editing (BE) strategies capable of correcting over 40% and up to 70% of pathogenic mutations. We demonstrate that two of these editing strategies are highly efficient in AHC-patient-derived iPSCs, correcting 59% and 90% of pathogenic alleles with PE and BE, respectively. Encouraged by these results, we initiated studies in a mouse model that faithfully recapitulates multiple hallmarks of AHC. We engineered a PE strategy to correct the model’s Atp1a3 D801N pathogenic mutation in cells and delivered this strategy using a dual-AAV9 PE system. We observed 47% and 37% correction of pathogenic alleles in the cortex and hippocampus, respectively, and a significant rescue of hippocampal ATP1A3 ATPase activity to 90% of wild-type levels. In vivo prime editing resulted in significantly enhanced survival and behavioral improvements in AHC mice, including rescue of motor coordination to near wild- type levels, significant reduction in the occurrence of tonic-clonic seizures during triggered paroxysmal events, and significant improvement in paroxysmal event recovery time. Compared to gene therapy in similar AHC Atp1a3 D801N mice, our gene editing approach represents the first durable multi-phenotype rescue of AHC in a mouse model, provides evidence for a dominant negative effect of the D801N AHC- causing mutation, and suggests the feasibility of a one-time gene editing treatment that can mitigate or reverse the pathology of AHC.