Investigating recombinant AAV (rAAV) gene therapy viral vector production

Gene therapy is a therapeutic approach used to treat diseases by correcting an underlying genetic abnormality. Viral vector systems account for more than 70% of gene therapy delivery vehicles, with the Adeno-Associated Virus (AAV) system the most widely utilised. AAV is a non-enveloped dependovirus with a number of serotypes that can be used to target different tissues; it does not integrate into the genome and has a good safety profile. AAV consists of an icosahedral capsid that contains a linear single-stranded DNA genome of about 4.8 kb. The AAV genome encodes several protein products: three structural capsid proteins (VP1-3) and four non-structural, replication Rep proteins. For therapeutic purposes, the production of recombinant AAV (rAAV) is often achieved via triple transient co-transfection of three plasmids into a mammalian host cell line, such as HEK293 cells. The three plasmids co-transfected most often contain AAV Rep and Cap genes, adenoviral genes, and the target therapeutic gene (gene of interest, GOI) flanked by AAV inverted terminal repeats (ITRs). Different serotypes of rAAV are manufactured for therapeutic purposes (usually by utilising different serotype Cap genes but AAV2 Rep) depending upon the clinical indication and target tissue. However, production of high yields of fully packaged rAAV (AAV capsids containing the genome of interest) remains challenging and a major contributor to the cost of AAV therapies (costs can be >£100,000's per treatment) where often less than 20% of capsids are successfully packaged with the required full-length recombinant genome. The aim of the work described in this thesis was to investigate different model genome systems for rAAV production, investigate pathways and transcriptomic targets implicated in limiting rAAV production and manipulation of these and the ITRs flanking the genome, and assess any impact on production of rAAV. There was no difference in the efficiency of production or yields of rAAV when making various model transgene systems in HEK293 cells as analysed using ELISA and qPCR assay for total particle and viral genome titres. The work also showed that the recombinant AAV2 vectors carrying either a fluorescent marker GFP or therapeutic target (GAA) transgene were functional in transduced HEK293 cells. Manipulation of literature identified targets that reportedly might limit rAAV production, Sp1 and PC4, in HEK293 cells showed varying impacts on the production of rAAV2 vectors. Two further approaches were investigated to determine how these might impact rAAV production. The first was to manipulate the ITR sequence utilised to flank the target gene of interest followed by subsequent monitoring of any impact on rAAV production and genome packaging. The second was to investigate anti-viral responses that might be initiated upon production of rAAV in HEK293 cells around mRNA translation initiation and its control, and whether knockdown of kinases involved in anti-viral responses impacted subsequent rAAV titre and genome packaging. The work showed that different serotype ITR sequences flanking a model genome gave similar genome titres and packaging efficiencies into either serotype 2 or 9 rAAV capsids and thus genome packaging efficiency does not appear to be sensitive to the ITR utilised. The knockdown of three eIF2α specific kinases, that phosphorylate the eIF2α translation initiation factor in response to viral infection and general stress perception to attenuate or stop protein synthesis, was also undertaken and any impact on rAAV production investigated. The knockdown of each target achieved were only in the region of 20-50% and this was insufficient to show a significant impact on cell growth or the production of rAAV. However, there was an observation of cross talk or sensing between the different eIF2α kinase targets whereby the knockdown of one target impacted the endogenous expression of other eIF2α kinases at the mRNA and protein level. Associated with this approach, the overexpression of p58, a protein that modulates eIF2α activity and general cell stress responses, was investigated alongside the long term effects of culturing HEK293 cells in the presence of sodium butyrate or valproic acid. The data showed packaging efficiency was marginally improved in transiently p58 exogenous over-expressing cultures although there was variation between production runs. Neither sodium butyrate or valproic acid directed evolution of HEK293 cells positively influenced rAAV titres. Collectively these data confirm the challenges in developing HEK293 cell systems and bioprocesses that produce rAAV with high genome packaging efficiencies although several of the approaches show promise that could be further explored in future work as outlined in the thesis.