UC Berkeley

Gene therapy, the introduction of genetic material to induce a change in genetic expressionto treat an underlying cause of disease has shown great potential in clinical settings. Advances in genomic and proteomic knowledge and access to several new gene editing techniques have generated a library of new potential nucleic acid therapeutics. While the root cause of diseases is often known, it is difficult to facilitate the delivery and subsequent expression of the therapeutic nucleic acids in a safe and effective manner. Viral vectors have emerged as a promising avenue for the delivery of such cargoes by harnessing their natural ability to infect and deliver genetic material. This allows them to be used as a biological tool as well as therapeutic delivery vectors. With the rapid expansion of the scope and number of clinical trials using viral vectors to deliver genetic material, a bottleneck has emerged in the ability to manufacture sufficient viral vectors to treat certain indications. Some indications, however, necessitate the integration of nucleic acid treatments to prevent the dilution that cellular division causes. Current therapies can either integrate sufficient therapeutic nucleic acid material randomly in the genome or integrate smaller payloads site-specifically. There is a need for a gene delivery vector that can both deliver and integrate large nucleic acid payloads and integrate them in a site-specific manner. The goals of this dissertation were to develop a new gene delivery vector and improve the infrastructure of existing gene delivery vectors. To mediate stable gene expression, DNA is often delivered through viral vectors such as adeno-associated virus (AAV) or lentivirus, though the use cases often dictate the necessity to use one over the other. AAV is a naturally occurring virus which is exceedingly efficient at circumventing natural barriers and mediating stable gene expression in nondividing cells, but its use in therapeutic applications may yield a vector that can only mildly infect the target cell types or may lack the ability altogether. As a result, AAV vectors have been engineered extensively to create variants that satisfy sufficient biomedical criteria such as avoiding existing neutralizing antibodies. A technique that has proven to be successful in creating new variants capable of infecting a broad range of cell types is directed evolution which is a process that mimics natural evolution. By creating a broad library of variants and applying a selective pressure such as antibody evasion or target cell infectivity, variants most adapted to overcome that pressure will be identified. Similarly, viral vectors are created in a cell line (Human Embryonic Kidney 293 cells or HEK293s) that may have intrinsic characteristics that make it ill-suited to manufacture AAV. By using a genome-wide screen which allows for the perturbation of single genes in a large pool of cells and iteratively selecting for cells which can most efficiently produce virus, I was able to identify genetic targets, SKA2 and ITPRIP, that improved the capacity of these cells to produce up to 2-fold more virus when expressed individually, or 4-fold when expressed in tandem. I also found that upregulation of these genes increased both the production of viral genomes and viral genome loading into AAV capsids by 3-fold. To further investigate whether these results were specific to a general HEK293 line or could be applied to a clinical grade HEK293 line, the genomewide perturbation study was again done, with a focus on the ability to increase viral titer and enrich for genetic hits that would secrete virus to the supernatant of the cell, making downstream purification easier. I found increased individual expression of HS6ST3 and ST3GAL4 was able to increase genomic titer 5-fold and increase production of supernatant secreted AAV by 7-fold over current industry standards. In dividing cells there is a natural dilution of delivered nucleic acids over time which can cause the therapeutic signal to be diminished. Therefore, gene therapies capable of integrating delivered nucleic acids into the genetic material of dividing are used. Lentiviral vectors are the predominant integrative viral vector used to integrate therapeutic payloads into a cell which stably expressed the gene of interest over long time courses, though there is the potential for deleterious integration. Other non-viral gene integration technologies include the use of either DNA transposons which suffer from the same drawbacks or integration mediated through the homologydirected repair pathway, which can integrate smaller nucleic acid segments in a site-specific manner. An ideal gene therapy of this type would be able to integrate sufficient nucleic acid in a site-specific manner. Retrotransposons are a class of protein that can integrate large nucleic acids, some of which integrate site-specifically. I was able to use a well-characterized retrotransposon R2 from the silkworm Bombyx mori to create a two-vector gene delivery system capable of integrating DNA payloads of at least 10kb in a site-specific manner. This system is the first retrotransposon gene delivery vector capable of site-specifically integrating DNA and expressing proteins in mammalian cells. Using full genome sequencing, I further confirmed that R2 was able to mediate site-specific integration with a 5-fold increase in integration efficacy over a control. Additionally, I showed that R2 was capable of integrating DNA into HEK293 cells and primary T cells. In summary, my dissertation has resulted in a novel genetic manipulation tool with potential to be used as a gene therapy as well as addressed a critical need in AAV production.