Characterizing Factor VIII Inhibitor Resistance of Platelet-Delivered Factor VIII in the Treatment of Hemophilia A.

Gene therapy strategies directed at expressing Factor (F) VIII in megakaryocytes may have several potential advantages in the treatment of severe hemophilia A. Among these is that platelet (p) FVIII may be protected from circulating inhibitors. We have previously described a murine transgenic line that expressed human B-domainless FVIII in a megakaryocyte-specific fashion and that this pFVIII was localized to within alpha granules. We also showed that these platelets contained FVIII equivalent to an infusion of 9% human FVIII into FVIII^null mice. We further showed that these pFVIII mice, on a FVIII^null murine background, formed stable clots in a FeCl³ carotid artery injury model. We then tested the ability of infused anti-human FVIII inhibitors in this setting. Using up to 100 μL of ESH8 monoclonal antibody (Ab) to the FVIII C2 light chain (1 μg/mL), anti-human polyclonal Ab (11 mg/mL) or a monoclonal Ab to the A2 domain (5 mg/mL), we were unable to alter thrombus formation in the carotid artery model. However, by using a 1:1:1 mixture of these inhibitors, we were able to show a dose-response curve. None of these mice developed thrombocytopenia suggesting that pFVIII is not exposed on the surface of circulating platelets. We then compared these studies to an acute infusion of the same inhibitor mixture in a FVIII^null mice receiving a 25% hFVIII correction. These studies showed that pFVIII/FVIII^null mice were ∼100-fold more resistant to inhibitors than plasma FVIII infusion into a FVIII^null mice in the carotid artery injury model. Since we had shown in the pFVIII mice that the FVIII is stored in alpha-granules, which can also store circulating Ab, we wondered whether the pFVIII/FVIII^null mice would be more sensitive to inhibitors when exposed in a chronic model where animals receive repeat doses of the inhibitor mixture. We therefore infused 3 doses of the inhibitor over 10 days, measured plasma and platelet inhibitor levels, and found that despite detectable stores of inhibitor within their platelets, these mice still demonstrated a comparable ability to form thrombosis as mice in the acute model with comparable plasma inhibitor levels. These studies suggest that pFVIII provides limited improved protection in mice with inhibitors comparable to <6 Bethesda Units (BU) in both acute and chronic models of inhibitor. Also, the presence of inhibitors within alpha-granules does not significantly inhibit the ability of these mice to form a clot. Our findings differ from a recent report of the efficacy of pFVIII in a tail vein survival model where pFVIII effectively protected mice from exsanguinations even in the presence of >100 BU/mL. We propose that the difference in outcome is due to the tail model being extremely sensitive to even low levels of pFVIII as exsanguinated, hypovolemic mice likely shunt blood away from their tail veins, and platelet activation and granular release are occurring in a low flow setting, while the FeCl₃ model used in this report requires a plasma equivalency of >3–5% human FVIII to show even partial correction.