Large-scale molecular characterization of adeno-associated virus vector integration in mouse liver

Recombinant adeno-associated virus (rAAV) vector holds promise for gene therapy. Despite a low frequency of chromosomal integration of vector genomes, recent studies have raised concerns about the risk of rAAV integration because integration occurs preferentially in genes and accompanies chromosomal deletions, which may lead to loss-of-function insertional mutagenesis. Here, by analyzing 347 rAAV integrations in mice, we elucidate novel features of rAAV integration: the presence of hot spots for integration and a strong preference for integrating near gene regulatory sequences. The most prominent hot spot was a harmless chromosomal niche in the rRNA gene repeats, whereas nearly half of the integrations landed near transcription start sites or CpG islands, suggesting the possibility of activating flanking cellular disease genes by vector integration, similar to retroviral gain-of-function insertional mutagenesis. Possible cancer-related genes were hit by rAAV integration at a frequency of 3.5%. In addition, the information about chromosomal changes at 218 integration sites and 602 breakpoints of vector genomes have provided a clue to how vector terminal repeats and host chromosomal DNA are joined in the integration process. Thus, the present study provides new insights into the risk of rAAV-mediated insertional mutagenesis and the mechanisms of rAAV integration.