Successful and Safe Treatment of Canine Hemophilia by Continuous Expression of Canine FVIIa: a Model for FVIII/FIX Gene-Based Bypassing Agents

Recombinant human activated Factor VII (rhFVIIa) is extensively used in the management of hemophilic inhibitor patients. However, the short plasma half-life and need for repeated infusions are amongst the limiting factors for its use in prophylaxis. In an attempt to overcome these, we have designed a gene transfer approach using continuous expression of a modified FVII transgene that is secreted in the active form, FVIIa. In proof-of-concept experiments, we previously demonstrated that Adeno-associated virus (AAV)–mediated gene transfer of the modified Factor VII gene can result in long-term phenotypic correction in hemophilia B mice. However, a prerequisite to a human application is the demonstration of efficacy as well as safety in a large animal model of hemophilia. Hence, we utilized the canine hemophilia model that closely resembles the human disease and has been shown to be an excellent predictor of efficacy of treatments for hemophilia in humans. As we previously reported (Blood 110 (11), p65a. 2007), we generated a canine FVII transgene that is expressed and secreted in a two-chain, activated form (cFVIIa). An AAV serotype 8 vector directing expression of this transgene was delivered via the portal vein into 4 hemophilia dogs, at vector doses of 2.06 E13 (one hemophilia B dog) − 1.25 E14 (three hemophilia A dogs) vector genomes (vg)/kg. The current abstract presents a rigorous characterization of the clotting parameters in the treated dogs, clinical safety and efficacy results of a substantially longer follow-up, and a laboratory safety profile of continuous expression of canine FVIIa. Gene expression in the treated dogs was followed by antigen and clotting assays (prothrombin time [PT], whole blood clotting time [WBCT]) as well as rotational thromboelastography (TEG). The initial dose of 2.06 E13 vg/kg resulted in a modest and transient reduction in the PT, WBCT and TEG parameters over a period of >31 months, accompanied by marginal changes in cFVIIa antigen levels. However, a considerable and sustained reduction in PT (supra-physiologic), WBCT (partial correction) and TEG parameters (near-normalization) was observed following administration of 6.25 E13 – 1.25 E14 vg/kg in three hemophilia A dogs in a cumulative total of 42 months of observation. Antigen levels in the treated hemophilia A dogs were in the range of 1–2.5 μg/ml. As untreated hemophilic dogs exhibit 5–6 spontaneous bleeds per year, we utilized this as an important efficacy endpoint. None of the treated dogs (including the lowest dose hemophilia B dog) exhibited any spontaneous bleeding episodes (34 expected within the observation period) in contrast to concurrently observed controls, which exhibited clinically detectable bleeding episodes at the predicted rate. Given the procoagulant nature of cFVIIa, the safety profile of this approach was thoroughly investigated. We did not observe any abnormal changes in serum chemistries, platelet counts, levels of D-dimer, fibrinogen or thrombin-antithrombin, arguing against a procoagulant state following cFVIIa-gene transfer. In summary, our results demonstrate for the first time that

1. AAV8 vector doses of up to 1.25 E14 vg/kg (2.6 E15 vg total) are well-tolerated in hemophilic dogs;

2. continuous expression of FVIIa at levels of 1–2.5 μg/ml is safe in hemophilic dogs, and

3. can result in a sustained, measurable improvement in the clinical bleeding course of these animals.

These data suggest that this model approach of a gene-based Factor VIII/Factor IX bypassing agent has therapeutic potential for hemophilia as has been shown for bolus infusion of rhFVIIa