Optimization of Recombinant Adeno-Associated Virus (aav) Vector Production in Saccharomyces Cerevisiae

Recombinant adeno-associated viral vectors (rAAV) are emerging drugs for gene therapy applications. Their non-pathogenic status, low inflammatory potential, availability of viral serotypes with different tissue tropisms, and prospective long-lasting gene expression are important attributes that make rAAVs safe and efficient therapeutic options. One of the main limitations for bringing rAAV gene therapy to the market is the difficulty in supplying enough rAAV vector product. The high vector doses suggested by early clinical data infer the need to scale up production at high volumes in order to satisfy patient demand. Current production platforms such as HEK293 or Sf9 cells are very efficient but up to date, scalability issues limit their use to preclinical and phase I/II production campaigns. Our team recently developed a novel rAAV-producing yeast strain, which recapitulated key molecular processes for vector particle formation (Barajas et al., 2017). The use of a microbial system for vector production would represent an affordable and highly scalable platform for industrial production. Preliminary data showed low vector yields, possibly associated to very low DNA encapsidation rate. The present thesis aims at getting clues about the molecular and bioprocessing factors that could be impacting vector yield in the novel rAAV-producing yeast system. In one approach, we performed a proteomic profiling of the yeast host response to rAAV protein expression. By using mass spectrometry and bioinformatics tools, we were able to identify trends in protein expression associated to vector formation. Gene ontology enrichment and network interaction analyses highlighted five specific cellular events: protein folding activity linked to unfolded protein response, proteasomal degradation activity, oxidation-reduction processes linked to oxidative stress, protein biosynthesis, and carbon metabolism. We speculated that some of these processes might be directly or indirectly linked to vector production constraints. A protein overexpression strategy was tested by transforming yeast with 2-micron plasmids carrying expression cassettes for 19 host cell proteins identified in the profiling. Increased vector yield was obtained in yeast strains overexpressing proteins SSA2, SSE1, SSE2, CCP1, GTT1, and GAL4. On a second approach, we used the yeast system as a means to screen the effect of host protein expression modulation on rAAV DNA replication and vector yield, by using the yTHC library strains (R1158-derived) and a set of 2 plasmids that confer all rAAV genetic elements. More than 850 strains, each one with a single host gene under a TET-repressible promoter, were screened in duplicates. From preliminary screenings, we identified 22 gene candidates that improved rAAV DNA replication (rAAV-GFP/18s rDNA ratio) and vector yield (benzonase-resistant rAAV DNA vector genome titer) as high as 6-fold and 15-fold relative to control, respectively. The candidate proteins participate in various biological processes such as DNA replication, ribosome biogenesis, and RNA and protein processing. The top five candidates (PRE4, HEM4, TOP2, GPN3, SDO1) were further screened by generating overexpression mutants in another yeast strain (YPH500). Subsequent clone evaluation was performed to confirm the rAAV-promoting activity of selected candidates under plate-based and bioreactor-controlled fermentation conditions. Our results highlighted HEM4 and TOP2 proteins as enhancers of rAAV2 vector yield in the yeast model. A final approach was focused on bioprocessing studies intended to develop a fed-batch fermentation process for rAAV2 vector production. Preliminary characterization studies performed in shake flasks provided useful data regarding rAAV DNA replication and vector formation in yeast over time, as well as optimal pH and temperature values for fermentation. Results suggested extending the original process to four days of galactose induction, and operating values of starting pH and temperature of 4.8 and 30°C, respectively. An additional media optimization study was performed to identify critical media components for optimal vector yield. A 3-fold increase was obtained after supplementing the galactose induction media with lysine, pyridoxin, myo-inositol, ferric chloride, and cysteine. We were able to translate a shake flask-based, batch process with medium replacement to a bioreactor-controlled fed-batch process. Low and moderate cell culture strategies were performed, controlling pH, DO %, and temperature. Additional studies were done to optimize growth rate, glucose and galactose feed, and induction strategy. However, final yields at moderate cell densities were comparable to the ones obtained at low cell densities, suggesting the presence of unknown factors that might be impacting per cell productivity. These three independent approaches provided important information regarding molecular and process strategies to optimize rAAV vector yield. Follow-up studies need to be done to consolidate yeast strain development and fermentation development efforts into a robust yeast platform potentially useful for industrial vector production