306. Generating Novel AAV Capsid Mutants for Large Genome Packaging Through Protein Libraries and Directed Evolution

306. Generating Novel AAV Capsid Mutants for Large Genome Packaging Through Protein Libraries and Directed Evolution

ADENOVIRUS VECTORS AND OTHER DNA VIRUS VECTORS II oversized AAV genomes. In addition, strand specific packaging of AAV genomes is a novel observation ...

245KB Sizes 0 Downloads 24 Views

ADENOVIRUS VECTORS AND OTHER DNA VIRUS VECTORS II oversized AAV genomes. In addition, strand specific packaging of AAV genomes is a novel observation and may have implications in vector production.

306. Generating Novel AAV Capsid Mutants for Large Genome Packaging Through Protein Libraries and Directed Evolution

Heikki T. Turunen,1 Luk H. Vandenberghe.1 1 Schepens Eye Research Institute, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, MA. One of the major shortcomings of the AAV as a gene therapy vector is its small size, allowing efficient packaging of vector constructs of only up to ~5 kbp in size. Although ultimately limited by the physical dimensions of the viral capsid, we hypothesize that by improving the efficiency of the viral genome packaging process a significant increase in AAV genetic payload carrying capacity can be achieved. Here, we have generated an approach combining protein libraries and directed evolution to generate novel AAV8 vectors with hopes to identify mutations enabling more efficient packaging of larger genomes. Wild type AAV8 genome was modified with silent point mutations to introduce restriction enzyme sites to facilitate insertion of library inserts. In addition, a cassette providing resistance to puromycin antibiotic was added to make the genome oversized and to allow selection for viruses carrying said oversized genomes if required. Libraries were generated to regions in Rep, Cap and AAP predicted to be involved in the genome packaging process. For each library seven amino acids were targeted with NNK mutations, generating libraries with total variability of 327, or ~3.4E10. However, due to the inherent inefficiency of restriction enzyme mediated insertion of library fragments into AAV8 genome, libraries with variability of 1E7-3E7 were achieved. Recombinant AAV8 vectors were produced by transfecting HEK-293 cells with AAV8 vectors including library mutation fragments, with adenovirus helper plasmid or wild type adenovirus. Vectors were purified and used to transduce fresh HEK293 cells treated with adenovirus helper. The cycling was repeated similarly for up to 10 cycles, with samples collected after each cycle. Samples were titered with qPCR, and library containing regions were amplified by PCR and subcloned back into the AAV8 genome to enrich library fragments allowing generation of viable vectors. No viable vectors were obtained from the Rep library. Rep is highly conserved throughout all AAV serotypes, and likely even slight changes caused by the majority of library mutations result in a non-functional protein, thus preventing amplification during cycling. However, decent titers were obtained from the Cap library, and several viable mutants have been isolated. Preliminary analyses have revealed mutants that are produced and are able to transduce cells with efficiencies comparable to wild type AAV8. All mutants have been sequenced and are analyzed to find patterns resulting in functional or non-functional viruses. Should positive patterns be detected, vectors with such targeted mutations will be made and assessed for production and transduction. Assays for packaging larger genomes are currently underway.

307. Engineering Recombinant AAV Vectors for More Efficient and Restricted Gene Expression in the CNS after Systemic Administration

Benjamin E. Deverman,1 Bryan Simpson,1 Piers Pravdo,1 Abhik Banerjee,1 Paul H. Patterson.1 1 Biology and Biological Engineering, California Institute of Technology, Pasadena, CA.

extend these findings, we are developing delivery vehicles capable of providing long-term, controllable expression of LIF, or other transgenes, that is widespread within, yet restricted to, the CNS. Recombinant adeno-associated virus 9 (rAAV9) is an attractive candidate for such a vector as it can deliver genes to neurons and glia throughout the CNS when injected intravenously (IV). For this systemic rAA9-based delivery approach to succeed in the adult, however, significant improvements in CNS transduction efficiency and selectivity are needed. In addition, it would be optimal to be able to modulate or turn off transgene expression post-delivery. We are making these enhancements by (i) using directed evolution/in vivo screening to develop AAV9-based capsids that more efficiently transduce CNS astrocytes, (ii) developing rAAV genomes that use gene regulatory elements (promoter/enhancer elements and miRNA binding sites) to restrict expression to CNS astrocytes and (ii) optimizing dox-inducible regulatory elements to provide control over transgene expression post-delivery. These optimized vectors should be useful as vehicles for the efficient and selective delivery of cytokines, trophic factors or other therapeutic cargo to the CNS for the treatment of neurological disorders.

Adenovirus Vectors and Other DNA Virus Vectors II 308. Natural Antibodies Inhibit Liver Transduction with Adenovirus Vectors in Mice

Qi Qiu,1 Zhili Xu,1 Jie Tian,1 Rituparna Moitra,1 Andrew P. Byrnes.1 1 Division of Cellular and Gene Therapies, FDA CBER, Bethesda, MD. Natural antibodies are a relatively understudied component of the innate immune system that can inhibit the efficiency of gene therapy. It is well established that liver transduction with intravenously-injected Ad5 vectors is higher in antibody-deficient mice than in wild-type mice. We have previously shown that normal wild-type mice have pre-existing natural IgM antibodies that can bind to Ad5 vectors and inhibit liver transduction. In the current work, we examined the relationship between IgM concentration and liver transduction by Ad5 vectors. We also asked how differences in IgM concentration between C57BL/6 and BALB/c mice are related to differences in liver transduction between these two mouse strains. Results: Liver transduction was evaluated after i.v. injection of Ad5 vectors into knockout (KO) mice that have abnormally low or high levels of natural antibodies. We found greatly elevated liver transduction in Rag-1 KO mice (no IgM) and CD19-cre mice (low IgM), as compared to liver transduction in control mice of the same background (C57BL/6). In contrast, mice with elevated IgM concentration (Ptpn6f/f;CD19-cre) had greatly reduced liver transduction. These results demonstrate an inverse correlation between IgM concentration and liver transduction. When Kupffer cells were depleted prior to vector injection, liver transduction was significantly enhanced and the concentration of IgM no longer influenced liver transduction. This result indicates that the inhibitory effects of IgM on liver transduction are mediated through Kupffer cells. When comparing different mouse strains, we found that BALB/c mice had higher IgM concentration and lower liver transduction than C57BL/6 mice. Crosses between these two strains confirmed that natural IgM concentration is negatively correlated with liver transduction. Conclusions: We conclude that high natural IgM concentration has a major negative effect on the ability of Ad5 vectors to transduce liver.

We previously demonstrated that delivery of the cytokine leukemia inhibitory factor (LIF) to the brain by recombinant adenovirus enhances oligodendrocyte progenitor cell proliferation and remyelination in a mouse model of multiple sclerosis (MS). To S118

Molecular Therapy Volume 22, Supplement 1, May 2014 Copyright © The American Society of Gene & Cell Therapy