403. Defined Vector Substrates Used To Probe rAAV DNA Integration

Abstract

We have previously used specifically oriented self-complementary rAAV (scAAV) vectors to quantify the dynamics of vector DNA circle and concatemer formation in cultured cells. We have now extended these studies to the dynamics of vector DNA integration. Consistent with earlier studies, rAAV vectors integrated at a rate of approximately one per 1000 infectious units (IU). The infectous titers of single-stranded rAAV (ssAAV) and scAAV vectors were both measured using the replication center assay, which minimizes the differences between the two through the effects of the helper adenovirus on second-strand DNA synthesis. In a comparison of ssAAV to scAAV, we found that both yielded the same number of integrations per IU. This was surprising, given that the scAAV yields 5-50 fold higher transductions per IU, and implies either that ssAAV genomes that have undergone conversion to double-strand DNA are primed for integration, or that integration can be initiated using a single-strand vector genome as substrate. The ssAAV and scAAV vectors were also evaluated for the ratio of concatemeric to monomeric integrations using a split-GFP reporter that required concatemer formation to express the marker gene from the linear proviral state. Approximately 20% of integrated scAAV genomes, and 6% of integrated ssAAV genomes, were concatemeric, although concatemers comprised only 2% of the scAAV episomal forms. This suggested that either the concatemers were 10-fold more likely to integrate than monomoers, or that the monomeric forms were amplified by replication during or after the integration process. The latter hypothesis was tested by co-infecting cells with pairs of scAAV vectors, each containing half of the marker gene, such that genome amplification could not lead to gene expression. We found that the previously observed number of integrated concatemers could be accounted for by vector genomes joined end-to-end prior to integration, suggesting that amplification was not a significant factor. The implications of these studies for gene therapy are that the scAAV vector will result in fewer integrations per transducing unit than the single-strand vector, and that high multiplicities of infection that promote concatemer formation may also lead to disproportionately higher rates of integration.