547. AAV Preparations Contain Contamination from DNA Sequences in Production Plasmids Directly Outside of the ITRs

Despite having the best safety profile of any current clinical viral vector, it is known that AAV preps contain contaminating sequences that are packaged alongside the expression cassette at a low rate. These sequences can originate from production plasmid DNA, or chromosomal DNA from producer cell lines. It has been reported that sequences from the expression cassette producer plasmid are more likely to be packaged into AAV than DNA from other producer plasmids, or chromosomal DNA. We hypothesised that in the expression cassette plasmid, backbone sequences directly flanking the ITRs might be packaged into AAV at a higher rate than sequences further away from either ITR. We first confirmed the presence of these sequences via PCR amplification of non-expression cassette DNA flanking the ITR sequence up to 1.5kb in length in an AAV prep. Sequential qPCR assays showed that plasmid sequences at a range of distances up to 2kb from the ITR make up between 1% and 9% of AAV particles. Most significantly, there was an observable decreasing trend in contaminant titer as distance from the ITR increased. Contaminant sequences closer to the ITRs (within 1kb) are detected at a 100 fold greater rate than distal plasmid DNA (9kb from ITRs on the same plasmid). The disparity in the levels of ITR adjacent DNA sequences, compared to sequences 9kb from either ITR, suggest that the origin of this DNA is from within AAV particles rather than residual plasmid DNA remaining after purification procedures. ITR adjacent contamination is present at both a TRS mutated ITR (required for self-complementary vectors) and non TRS mutated ITRs. Contaminating plasmid sequences were present when the transgene was half of the packaging capacity (2.3kb FIX prep) and at the full capacity (5kb FVIII prep) at comparable levels, suggesting that increasing the transgene size with stuffer DNA to create a full genome will not solve this issue. Previous studies have concluded that increasing the size of the backbone with stuffer DNA reduces the level of plasmid backbone contamination, as the two ITRs are then not in range of each other to facilitate reverse packaging. However, the total plasmid size in our studies was u003e20kb. Therefore, the ITRs should not be in range for this to occur. We hypothesise that these ITR adjacent sequences are either a product of read-through from the expression cassette into flanking sequences, due to inefficient cleavage at the ITR breakpoint, or from reverse priming mediated by only 1 functional ITR. In the current expression cassette plasmids examined, the Kanr gene and bacterial f1origin of replication are within the range of the flanking sequences that could be packaged. With current, unsolved clinical challenges for AAV, including transaminitis post high dose infection, it is clear that clinical AAV vectors should be designed to contain as little contamination as possible. We conclude that newly designed AAV production plasmids should contain significant lengths of stuffer DNA flanking each ITR (at least 2kb) to ensure that bacterial sequences are not packaged into AAV preps. Further research into vector design is required to eliminate this source of non-functional DNA from AAV produced for the clinic.View Large Image | Download PowerPoint Slide