96. Genomic Integration Profile of AAVHSC Vectors in Human CD34+ Long-Term Engrafted Hematopoietic Stem Cell Xenografts in Immune-Deficient Mice

Adeno-associated virus vectors (AAV) are proving to be promising and powerful for gene delivery due to their ability to safely transduce both non-dividing and dividing cells. While AAV mainly persists in episomal copies in non-dividing cells, genomic integration has been reported including in CD34+ hematopoietic stem cells (HSCs). Characterizing the vector integration profile is important to ensure safety since genomic integration has the potential for inducing insertional mutagenesis and genotoxicity. The absence of genotoxicity of AAV has been documented in multiple studies. Our lab has recently described the isolation and gene transfer properties of human HSC derived AAVs (AAVHSCs) (Smith LJ, et al. Mol Therapy 2014). To test the ability of the AAVHSCs to transduce long-term engrafted HSCs, we transduced human CD34+ cells with AAV vectors encoding a luciferase transgene prior to transplantation into irradiated immunodeficient NOD/SCID mice. Through serial bioluminescent imaging, we found that the AAVHSCs supported higher levels of expression than AAV2, 7 and 8. Of the AAVHSC serotypes tested, AAVHSC17 was found to support the highest level of transgene expression in vivo, for at least six months post-transplantation. Since vector genomes were found to persist in both long-term engrafted CD34+ cells as well as their differentiated hematopoietic progeny, we hypothesized that the vector genomes likely persisted as chromosomal integrants since episomal copies would have been lost during mitosis. Here, we evaluated the vector genome integration profile in the long-term engrafted CD34+ HSCs transduced with AAVHSC17. High molecular weight DNA was extracted from flow sorted human CD34+ cells isolated from engrafted mouse marrow and sonicated to fragment the DNA before ligation of double stranded DNA adaptors. Using LM-PCR we amplified AAV-chromosomal sequences using AAV-specific and adaptor-specific primers, and further enriched vector-chromosome junctions with nested PCR. The PCR generated library was then used for paired-end Illumina sequencing. The sequence reads were paired using CLC Genomics Workbench. High-throughput analysis resources including PLAN, BLAST (NCBI), and ENSEMBL were used to determine the locations of the AAV-chromosomal junctions. AAV integration was found to occur randomly throughout the genome, including the X and Y chromosomes. Similar to previous studies, AAV integration junctions were largely found in non-coding and intronic regions. Integration was also found to occur within microsatellite repeat regions and GC rich regions. Thus far, no integration events have been found in coding regions or near oncogenes. Due to the lack of detectable genotoxicity resulting from AAV integration in CD34+ HSCs and the ability to support long term in vivo transgene expression, we conclude that these vectors are well-suited for stem cell gene delivery.