NGS-based Engineering of Potent Muscle-Tropic Gene Therapy Vectors and Synthetic Controls for Viral Surveillance

Abstract

Advanced genomic approaches have impacted nearly every field of biology in the past decade, and present an especially powerful tool to detect, study, and engineer viruses. Through the application of novel technologies leveraging high-throughput, synthetic genomic techniques applied to viruses, I have developed new tools to treat genetic myopathies and track infectious diseases. Although normally associated with illness, viruses can also be useful tools for treating disease and deepening our molecular understanding of biology. Viral vectors, such as recombinant adeno-associated virus (rAAV), utilize a viral capsid as a vessel to deliver therapeutic genetic cargo directly into the cells of patients with disease-causing mutations. rAAVs hold great promise due to their low immunogenicity, stable transgene expression, and broad tissue tropism. Effective treatments for genetic myopathies such as Duchenne Muscular Dystrophy has been a long sought after goal for rAAV-based therapies. However, due to their broad tissue tropism, these naturally occurring capsids transduce off-target organs, such as the liver, leading to the need for very high vector doses to efficiently transduce large anatomically distributed skeletal muscle. These high vector doses pose serious health risks for the patient and are difficult to manufacture. By engineering more targeted AAV capsids this risk can be offset. In Chapter 2, I describe how we use an mRNA-based selection strategy termed DELIVER (directed evolution of AAV capsids leveraging in vivo expression of transgene RNA) to select for capsid variants that potently and selectively transduce muscle fibers and cardiomyocytes. Through the insertion of a random 7-mer library into the AAV9 capsid we generated a diverse library of more than 5,000,000 capsid variants with which to begin our selections. After systemic administration into mice and cynomolgus macaques, we used next generation sequencing (NGS) to survey the capsid variants which transduced our tissues of interest while being detargeted from the liver. The sequencing results informed the design of a smaller second round library, which resulted in the discovery of a class of capsids containing an RGD-motif which transduced our tissues of interest with high specificity. We characterize some of the top performing RGD-motif containing capsids by probing the mechanism, biodistribution, and therapeutic benefit in mouse models of disease. Ultimately, we describe four top performing RGD-motif containing capsid variants arising from our non-human primate (NHP) selections that are likely to have the greatest therapeutic benefit for human applications. The global spread and continued evolution of SARS-CoV-2 made clear that standard approaches for viral genomic surveillance were insufficient, yet they play a critical role in informing public health decisions during a pandemic. With the onset of the COVID-19 pandemic, my research shifted to respond to an increased need for SARS-CoV-2 surveillance. Having accurate methods for NGS-based viral surveillance is crucial as contamination can alter the integrity of the viral genomes leading to misinterpretation of the data. In order to respond quickly to the rapid rise in case numbers, high throughput amplicon-based sequencing methods became more commonplace. These methods are relatively low cost and have greater sensitivity than unbiased metagenomic sequencing but are more prone to contamination due to the numerous cycles of virus specific PCR. In Chapter 3, I describe our work developing 96 synthetic DNA spike-ins (SDSIs) to track samples and detect inter-sample contamination throughout a SARS-CoV-2 amplicon sequencing workflow. The SDSIs did not have a negative impact on genome recovery and were easily integrated into the workflow. We validated these SDSIs across 6,676 diagnostic samples at multiple laboratories and demonstrated their value in identifying multiple modes of sample contamination. Our method, termed SDSI + AmpSeq, provides increased confidence in genomic data and can be easily implemented in a variety of other amplicon sequencing workflows. These chapters demonstrate two different viewpoints surrounding viruses, one in which they can be used as vectors to treat disease and another in which they are causing disease and must be precisely tracked to control a pandemic. In both chapters, NGS is an invaluable tool to quickly and precisely characterize both the AAV capsid and SARS-CoV-2 viral variants. As illustrated in this dissertation, continued advances in the field of virology can serve multiple purposes all of which having the ultimate goal of treating human disease.