Impact of Pre-Existing Immunogenicity to AAV on Vector Transduction By Bmn 270, an AAV5-Based Gene Therapy Treatment for Hemophilia A

Neutralizing antibodies are hypothesized to limit the transduction efficiency of adeno-associated virus (AAV) based gene therapies. This pre-existing immunogenicity is thought to be the result of prior exposure to wild type AAV virus, and seroprevalence studies estimate that between 40% and 70% of the human population has been exposed across AAV serotypes. In addition, other poorly defined plasma factors are suspected of inhibiting transduction where antibodies are not detected. BMN 270 is an investigational AAV5-mediated gene therapy for the treatment of Hemophilia A, and encodes for a codon-optimized B-domain deleted human FVIII protein (hFVIII-SQ) under control of a liver specific promoter. In the first-in-human clinical trial, BMN 270-201, prospective patients were screened and excluded on the basis of pre-existing total antibodies (TAbs) to the AAV5 vector or inhibitors of AAV5 transduction using a cell-based in vitro transduction inhibition (TI) assay. In the BMN 270-201 Phase I/II clinical trial, 2 of 21 screened patients with severe Hemophilia A were excluded from the study on the basis of having low titer TAb+ responses (titer = 94 and 96, respectively). Nonetheless, the question remained as to whether patients with pre-existing immunity against AAV5 could still be successfully dosed with BMN 270 to achieve sufficient FVIII expression. To better understand the relationship between pre-existing AAV5 immunity and the pharmacodynamics of gene delivery and human FVIII (hFVIII) expression, we studied cynomolgus monkeys with varying anti-AAV5 TAb levels and TI titers following a single infusion of BMN 270.

Over the 8 week course of the study, maximal expression of hFVIII-SQ was observed between Weeks 3 and 5, and was notably lower in monkeys that were positive for pre-existing immunity in both the TI and TAb assays (TI+/TAb+). In TI+/TAb+ monkeys, only 3 of 5 animals had detectable hFVIII-SQ expression with a mean FVIII protein level of 6.16 ng/mL at Week 4, an approximately 72% reduction compared to TI-/TAb- control animals (week 4 mean = 22.2 ng/mL). One of five TI+/TAb+ monkeys expressed hFVIII levels comparable to that in TI-/TAb- controls. At Week 4, the FVIII protein expression level in this monkey was 21.9 ng/mL, while two others had expression at much lower levels of 3.93 ng/mL and 4.96 ng/mL respectively. Two of five monkeys in the TI+/TAb+ group had no detectable FVIII protein at any time point. Vector genome copies were also lower in the livers from TI+/TAb+ animals, as measured by qPCR, and correlated with protein expression levels. In contrast, in monkeys with measurable TI titers (TI titers < 10) but no detectable TAb (TI+/TAb-), no decrease in hFVIII levels compared to TI-/TAb- controls was found, suggesting that neutralizing anti-AAV5 antibodies, as opposed to non-antibody-based AAV5 inhibitors, are mainly responsible for the diminished transduction and expression of hFVIII in vivo. These results suggest a reasonable likelihood that similar TI+/TAb- Hemophilia A patients will respond to BMN 270 gene therapy, and that the biological threshold for clinically relevant anti-AAV5 antibody levels may be different from the statistically-derived assay cut point. These data also show that individual antibody responses may vary with respect to in vivo neutralizing capacity, and support further studies in Hemophilia A patients with low levels of pre-existing AAV5 antibodies.