Background Despite progress in adeno-associated virus (AAV) mediated factor VIII (FVIII) gene transfer for hemophilia A (HA), durable and sustained FVIII expression sufficient to ameliorate bleeding in all participants is an unrealized goal. Following the success of using the gain of function FIX-Padua variant in hemophilia B gene therapy, we hypothesized that use of an enhanced hemostatic function FVIII would similarly improve HA gene therapy by: 1) permitting use of lower vector doses, thereby overcoming dose-dependent safety and efficacy limitations and 2) achieving efficacy with lower FVIII antigen. We previously published a FVIII variant resistant to activated protein C (APC) cleavage (FVIII-R336Q/R562Q or FVIII-QQ) with normal specific activity and 4-5 fold improved hemostatic function in recombinant protein experiments (Wilhelm et al. Blood 2021).

Aims We evaluated the possible clinical translation of FVIII-QQ for HA AAV-mediated gene therapy by determining the hemostatic benefit and potential prothrombotic and immunological risks in mice. Additionally, FVIII-QQ steady-state expression was 2.5 to 5-fold higher than FVIII-WT. We hypothesized this was due to disrupted cleavage by APC or other serine proteases at R336 and/or R562.

Results Survival studies in HA/CD4KO and WT/CD4KO mice treated with AAV vectors to express (h)uman FVIII-WT and hFVIII-QQ in the range of mild HA, normal, and supratherapeutic FVIII levels thus far demonstrate no significant differences in survival at 7-11 month follow up or D-dimer levels after 8 months. Additionally, FVIII-QQ mice were generated using CRISPR/Cas9 to introduce R562Q (WT mice are Q336) for endogenous (m)ouse FVIII-QQ expression. Endogenous mFVIII-QQ mice were viable, fertile, normal weight, and D-dimers did not significantly differ from WT mice, while FVIII activity was ~2-fold higher than WT mice. HA immune competent mice treated with either 6 weekly doses of hFVIII-WT or hFVIII-QQ protein (0.2 μg) or AAV vector to express hFVIII-QQ or hFVIII-WT had no significant differences in inhibitor titers.

To evaluate hemostasis, HA/CD4KO mice were treated with AAV vector only differing by 2 amino acids (R336, R362) at 3 different doses (n=6-12 mice/cohort) to express hFVIII-QQ or hFVIII-WT and underwent tail clip challenge. Estimated EC50 and EC80 of vector doses required to normalize blood loss to WT mice were 5 and 10-fold lower, respectively, for hFVIII-QQ versus hFVIII-WT supporting a hemostatic benefit to using the hFVIII-QQ transgene. Interestingly, across 3 different vector constructs differing only by F8 codon optimization, we observed hFVIII-QQ antigen values 2.5 to 5-fold higher than in mice expressing hFVIII-WT. Analysis of liver tissue from HA/CD4KO mice treated with the same vector construct and dose to express hFVIII-WT or hFVIII-QQ demonstrated approximately the same mRNA values, but plasma hFVIII-QQ antigen was 4-fold higher than hFVIII-WT and intracellular hFVIII-QQ antigen was the same or lower than hFVIII-WT antigen. Consistent with our in vivo AAV experiments, preliminary studies of intracellular versus secreted hFVIII antigen in BHK stable cell lines expressing hFVIII-WT or hFVIII-QQ demonstrated similar mRNA values with higher relative secreted hFVIII-QQ protein. To further probe the mechanism of higher hFVIII-QQ steady-state expression, we evaluated a panel of serine proteases (APC, FIXa, FXa, plasmin, PACE/furin and thrombin) for cleavage of FVIII-QQ relative to FVIII-WT protein. As expected, hFVIII-QQ was resistant to APC cleavage. Surprisingly, all other serine proteases tested demonstrated cleavage of hFVIII-WT, but not hFVIII-QQ, at R336 and/or R562 suggesting protease attack at these cleavage sites may occur more than previously appreciated, though the physiologic significance remains to be determined.

Conclusion Our studies outline the use of hFVIII-QQ transgene expression thus far does not demonstrate safety concerns and confers hemostatic benefit over hFVIII-WT via: 1) improved hemostatic function and 2) higher steady-state expression. Enhanced steady state hFVIII-QQ expression relative to hFVIII-WT was observed across multiple model systems. Preliminary data demonstrate multiple serine proteases can cleave at the described FVIII APC cleavage sites (R336, R562) suggesting hFVIII-QQ may generally be more proteolytically stable thereby permitting improved steady-state expression.

George:Pfizer: Consultancy; Bayer: Consultancy; Avrobio: Consultancy, Membership on an entity's Board of Directors or advisory committees; STRM.Bio: Membership on an entity's Board of Directors or advisory committees; Spark: Consultancy; Biomarin: Consultancy; AskBio Therapeutics: Other: Licensing Fees, Research Funding; Intellia: Consultancy.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution