ABSTRACT
Adeno-associated virus (AAV) vectors are in clinical development for delivery of genes to treat multiple genetic
diseases including hemophilia. While progress has been made to optimize gene delivery, in some studies the
required AAV vector doses were high, leading to toxicity and even fatal outcomes in one study. These findings
highlight the need for novel approaches to reduce the AAV vector dose to minimize liver toxicity, anti-AAV
immune responses, and genotoxicity. Our recent studies and work from others have identified an
underappreciated limitation to efficient gene correction with AAV vectors. In a long term study of AAV gene
delivery of FVIII in hemophilia A dogs, we found that most of the AAV vector genomes were highly rearranged
in transduced liver tissues. These rearrangements typically disrupted the transgene, and so would compromise
expression of the transgene product—unexpectedly, our data indicated that most of the AAV vector genomes
present did not produce functional protein after transduction. These rearranged AAV genomes were present in
integrated forms but also in AAV concatemers that may be episomal forms. It is unclear whether these
rearrangements occurred during vector production or after transduction of the target cells, though data is
accumulating that at least some of the rearrangements originate in vector producer cells. Our hemophilia A dog
study also identified integration events in the canine genome within genes linked to cell growth and cancer that
were associated with clonal expansions. Validation of integrated AAV DNA in these expanded clones by
sequence analysis showed that in all cases integrated vectors were highly rearranged, with only one of five
encoding an intact transgene. An extensive literature documents interactions of AAV with host DNA repair
pathways in both vector producer and target cells, though the influence of host factors in AAV DNA
rearrangements is mostly unstudied. We hypothesize that modulation of host cell pathways can suppress AAV
DNA rearrangements, thereby allowing improved transgene expression per vector DNA copy. In this proposal,
we will 1) implement a deep sequencing method to quantify rearrangement frequency in a statistically rigorous
fashion, 2) identify cellular pathways that can be modulated with small molecules, siRNAs, or microRNAs that
suppress vector rearrangements, and 3) devise novel delivery strategies that support efficient pathway
modulation, suppress vector rearrangement, and boost transgene output per vector copy. These methods will
be assessed during AAV vector production (Specific Aim 1) and after AAV delivery in the transduced target cells
(Specific Aim 2). Our deliverables at the end of the project will be a greatly enhanced understanding of the
interaction of AAV with host cell DNA handling pathways, and methods for modulating these pathways to allow
safe and effective gene delivery at lower vector doses.