Successful Production of Canine FVIII: Biochemical and Functional Characterization in Hemophilia A Dogs.

The canine hemophilia A (HA) dog model is of pivotal importance for the development of novel therapeutics for hemophilia. Notably, the lack of species-specific reagents hampers the assessment of both efficacy and safety of many protein- or cell/gene-based approaches. Here we report the successful production and characterization of recombinant canine FVIII (cFVIII) and its biological activity in HA dogs. A B-domain deleted cFVIII derivative (cFVIII-SQ) was stably expressed in BHK cells and protein was purified using ion exchange chromatography. cFVIII-SQ expressed very well (0.5mg/L of media) compared to human FVIII-SQ (hFVIII-SQ; 0.15 mg/L). This possibly suggests that cFVIII-SQ is more efficiently secreted and/or more stable than the hFVIII-SQ. Because of sequences differences within the intracellular B-domain cleavage site relative to hFVIII-SQ, cFVIII-SQ migrated as a single band on an SDS-PAGE gel with a molecular mass of ∼170 kDa. Thrombin (IIa) cleavage resulted in the expected processed heavy (50 kDa and 43 kDa fragments) and light chains (73 kDa). N-terminal amino acid sequence analysis of the single chain protein and IIa cleavage products gave the expected results. A one-stage aPTT assay with hFVIII-deficient plasma revealed that cFVIII-SQ has specific activity ∼2-fold (13,000 U/mg) greater compared to hFVIII-SQ (6,600 U/mg) and both proteins had a high activation quotient (62 vs 45, respectively). These data were further confirmed using a fluorescent thrombin generation assay. Surprisingly, using either a clotting assay or a FX activation assay employing purified FIXa, FX, FVIII-SQ and synthetic phospholipid vesicles, we were able to show that cFVIII-SQ was much more stable following IIa activation compared to hFVIII-SQ. These data suggest that the A2 domain of cFVIII-SQ protein may dissociate less efficiently than the human A2 domain following IIa cleavage. These findings explain, at least partially, the superior clotting activity determined in the multiple tests. Using purified cFVIII-SQ we generated a series of novel antibodies and determined that cFVIII plasma concentration in samples of normal dogs ranges from 80–130ng/ml whereas FVIII levels are undetectable in HA dogs. Next, we injected two adult HA dogs (20–22kgs) at doses of 25 IU/kg of cFVIII-SQ for a total of 5 injections. The whole blood clotting time and aPTT values shortened to within the normal range. Overall, the calculated half-life of cFVIII-SQ varied from 18–22 hrs. These data suggest that recombinant cFVIII-SQ has a half-life in dogs which is comparable to hFVIII-SQ in humans. Next, we tested whether infusion of the purified cFVIII-SQ would trigger the formation of antibodies. Two HA dogs received a total of five injections of 25 IU/kg of cFVIII-SQ per injection, two-weeks apart, and blood was drawn at several time points up to day 30. There is no evidence of thrombocytopenia, excessive coagulation activation, or organ abnormal functions for periods longer than 2 wks post-infusion. Neither inhibitory antibody nor anti-cFVIII-specific IgGs were detected in these samples over time. This work fills an important void for the study of cFVIII biological functions and immune responses in HA dogs. Furthermore, the understanding of the mechanism of superior production of recombinant cFVIII may provide insights for the production of the human cognate.