Self-Complementary AAV Vectors Cause a Substantially Heightened TLR9-Dependent Innate Immune Response In the Liver

Abstract 252

Successful gene therapy for hemophilia requires therapeutic expression in the absence of destructive immune responses. Clinical trials have revealed that adaptive immune responses eliminated therapeutic levels of factor IX (F.IX) following high-dose hepatic in vivo gene transfer using single stranded (ss) AAV serotype 2 vector. More efficient AAV vectors have been developed using AAV2 with Y-F capsid mutations, alternate serotypes such as AAV8, and self-complimentary (sc) rAAV vectors. Innate immune responses to AAV in the liver are thought to be weak and highly transient. Our study sought to investigate whether variations in the capsid or genome will alter the immune profile. AAV vectors were delivered into the portal circulation of C3H/OuJ mice at 10¹¹ vector genomes per mouse. We tested ssAAV-2, ssAAV2-triple Y-F mutant, ssAAV-8 vectors as well as scAAV-2 and scAAV-8 vectors, all expressing the hF.IX transgene. Quantitative RT-PCR array showed a mild transient up-regulation of MyD88, TLR-9, TNF-α, MCP-1, IP-10 and IFN-α/β within 2 hrs, which subsided by 6 hrs following delivery of ssAAV vectors, regardless of the capsid sequence/efficiency of hepatocyte transduction. In contrast, scAAV-2 and scAAV-8 vectors induced 4- to 8-fold increases in TLR-2, TLR-9, MyD88, TNF-α, MCP-1, IP-10, and IFN-α/β expression when compared to the ssAAV vectors. In addition, scAAV vectors induced expression of pro-inflammatory cytokines and chemokines not detected for ssAAV: IL-6, IL-12α, MIP-1, and RANTES. Paralleling this data was a 4-fold increase in local protein levels of IL-6, TNF-α, and MCP-1, and a systemic increase in IL-6. None of the vectors activated TLR-4, indicating that these effects cannot be attributed to LPS contamination. Several other innate response genes showed no differences in expression compared to PBS injected mice, which included IL-1α, IL-1β, KC, TLR-1, and TLR-3 to -8. The heightened innate response to scAAV vectors was further characterized by a 5-fold increase in macrophage and neutrophil infiltrates in liver sections, illustrating the potential for hepatotoxicity. Innate responses were absent in TLR-9 deficient C57BL/6 mice and in wild-type mice that had received an oligonucleotide ligand inhibitory to TLR-9. Compared to wild-type mice, hFIX levels were increased 2-fold in TLR-9 deficient mice following scAAV vector delivery and antibody formation against capsid was delayed. Additionally, in vitro studies with a TLR-9 reporter cell line showed significantly increased TLR-9 signaling for scAAV2 compared to ssAAV vectors, independent of the transgene/expression cassette. Neither vector activated the inflammasome. Further in vivo studies showed that the innate response to scAAV was, with the exception of TNF-a expression, Kupffer cell dependent, likely owing to the importance of these cells to sequester viral particles in the liver. Our results have several implications. Gene therapy applications based on administration of high-dose scAAV may be limited by immunotoxicities, which, however, can be prevented by inhibition of TLR-9 signaling. Of general importance, changing the genome configuration of a virus from single- to double-stranded DNA increased sensing by the endosomal receptor TLR-9, thereby enhancing innate immunity.