Recombinant adeno-associated viral (AAV) vector is a promising gene-based strategy for the treatment of several inherited diseases. Using AAV serotype 2 (AAV-2), the most common tested vector in humans, we have determined that the risk of germline transmission and the immune responses to both transgene product and/or vector-capsid proteins are critical obstacles to the safety of this strategy (

Nat Med
). Recently, novel and more potent serotypes have emerged such as AAV-8 that allows efficient liver transduction following peripheral intravenous injection (IV). The major determinant of vector safety relies on its tissue tropism. Here, we sought to compare the efficacy and safety profile of AAV-2 and AAV-8 in a large animal model. Male rabbits (∼3 kg) received IV injection of AAV-2 (n=11) or AAV-8 (n=11) encoding human F.IX (hF.IX) at doses ranging from 1×1012 to 1×1013vg/kg. Injections with AAV-2-hF.IX resulted in 6-fold lower expression of hF.IX than AAV-8-hF.IX for both low and high dose cohorts. Notably, no neutralizing antibody to hF.IX was detected with either serotypes. Eighteen weeks following the initial injections, animals were cross-administered with either AAV-2 or AAV-8. Whereas injection of AAV-8 led a 20% increase in transgene expression in animals initially injected with AAV-2-h.FIX, AAV-2 failed to boost hF.IX expression. Regarding germline safety, the presence of vector genomes in semen samples from the high-dose cohort (6 to 10 weeks after injection) was 3-5 fold higher for AAV-8 compared with AAV-2. There were no differences in the vector clearance in the semen among rabbits from the low-dose cohorts of AAV-2 and AAV-8. After 12 weeks, all semen samples from all cohorts tested negative. In another rabbit model, vasectomized prior to vector injection, we determined that semen samples lacking germ cells were also positive for vector-DNA sequences. The kinetics of vector clearance in these samples was dose- and time-dependent and serotype-independent. Because the presence of capsid in early-time points is critical for predicting possible immune responses against the viral vector, we determine the vector biodistribution one week after injection of 1×1013 vg/kg of AAV-2 or AAV-8. Rabbits were euthanized and their organs were harvested and analyzed for vector DNA presence through real-time quantitative PCR. Comparing to AAV-2, AAV-8 genomes were 2-5 fold times higher in all organs (spleen, lung, heart, and kidney), with the exception of liver where vector-DNA content was comparable (range from 25-69 copy number/cell). In addition, testes, accessory glands, and prostates were positive for the vector DNA, albeit in very low levels (the highest level of vector DNA found in those organs was 3copy number/cell in the testis of one rabbit injected with AAV-8). These differences may reflect the distributions of cellular receptors for AAV-2 and AAV-8, which may also explain the higher content of vector genome in the semen of high-dose AAV-8 cohort. Together our findings suggest that AAV-8 ensures higher transgene expression than AAV-2 and preexisting immunity to AAV-2, a naturally acquired virus in humans, may not limit AAV-8-mediated gene delivery. The overall kinetics of AAV vector clearance in the semen seems to be independent of the presence of germ cells and vector serotype. However, early biodistribution data of AAV-8 suggests a distinct safety profile from AAV-2.

Author notes

Disclosure: No relevant conflicts of interest to declare.