Generation of Long-Acting Coagulation Factor VIII for the Treatment of Haemophilia A

Background: Haemophilia A is an X-linked bleeding disorder resulting from dysregulation of clotting factor VIII (FVIII) during haemostasis. Predominant mode of treatment is by replacement therapy with recombinant FVIII (rFVIII) on demand or as a prophylaxis (Berntorp et al., 2003, Gouider et al., 2015). While there are a number of rFVIII products either in development or recently licensed with delayed clearance, none have a plasma half-life of >1.6 times that of native FVIII due to the dominant influence of von Willebrand factor (VWF) as all current commercial rFVIII products depend on binding to VWF for intravascular stability. However, the interaction facilitates plasma clearance via the VWF-mediated pathway, thus limiting the plasma half-life (Fischer et al., 2009). Here, a new technology for generating a long-acting rFVIII that functions independent of VWF was investigated. Objective: To develop a novel longer acting FVIII with a limited interaction with VWF and an extended plasma half-life via a fusion of rFVIII with an inactivated rFIX. Hypothesis: The fusion of FVIII to inactive FIX will create steric hindrance for binding to VWF and FVIII clearance receptors and thereby delay the in vivo clearance of FVIII. Methods: Two novel rFVIII molecules were generated by fusion of FVIII to an inactivated factor IX designated 13A1FVIII and 13B1FVIII. Both FVIII fusions were cloned and expressed in mammalian cells (CHO Flp-In and HEK293c18), protein expression was confirmed by Western blotting and biological activity assessed in a chromogenic assay. In vitro haemostatic potential of the lead FVIII-fusion (13A1FVIII) was evaluated in an activated partial thromboplastin time (aPTT) and thrombin generation (TGA) assays while the in vivo efficacy was assessed in a tail vein transection (TVT) bleeding model. The plasma clearance was assessed in both a preliminary and repeat PK/PD studies in VWF/FVIII double knock out mice and FVIII single knockout mice relative to Advate FVIII. Results: 13A1FVIII demonstrated more biological activity than 13B1FVIII in a chromogenic assay (5.58 IU/mL vs 1.22 IU/mL respectively) and so 13A1FVIII was taken forward for in vivo studies. Protein purification of 13A1FVIII was achieved by affinity chromatography with >95% purity as judged by silver staining. In vitro, 13A1FVIII was stable on storage and biological activity was confirmed in TGA and aPTT assays. In a pilot PK study using VWF/FVIII double knock out (DKO) mice, 13A1FVIII exhibited a reduced plasma clearance over Advate FVIII (terminal half-life of 32 hours vs 12 hours, respectively). However, the data from a repeat PK study in F8KO/DKO mice showed that the plasma clearance of 13A1FVIII is remarkably similar to that of Advate FVIII; both demonstrating half-lives of 7.2 vs 8.8 hours for F8KO and 51 vs 59 hours in DKO mice, respectively. In the TVT bleeding model there was no significant difference in bleed times, although Advate FVIII demonstrated a better haemostatic efficacy in TVT at 15 minutes whereas 13A1FVIII FVIII-showed reduced blood loss and bleeding time at 24 hours. Conclusions: The 13A1FVIII fusion of FVIII to inactivated FIX was stable and biologically active in vitro. However, in vivo, the 13A1FVIII fusion had similar clearance and activity to a third generation FVIII. The results suggest the FVIII and inactivated FIX fusion is being inactivated through the same pathway as native FVIII in vivo and that the current fusion molecule is not stable in vivo