AUTHORS
Johnathan Morris, PhD, Connor Watson, PhD2, Robert Davidson, MS3, Lindsey George, MD4
AFFILIATIONS
1 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States, University of Pennsylvania, Philadelphia, Pennsylvania, United States
2 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States, University of Pennsylvania, Philadelphia, Pennsylvania, United States
3 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
4 Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, United States, University of Pennsylvania, Philadelphia, Pennsylvania, United States
ABSTRACT
Background: Elevated Factor VIII (FVIII) activity is an established risk factor for venous thrombosis (VT), with VT risk exceeding all other procoagulant factors, except von Willebrand Factor (vWF). However, because FVIII circulates in a high-affinity complex with vWF, and both proteins are elevated during physiologic stress, it remains unclear whether elevated FVIII activity independently promotes VT or primarily reflects an underlying inflammatory state. Consistent with epidemiologic data, naturally occurring human FVIII variants, including FVIII Padua and the recently described FVIII Aurora, demonstrate that elevated FVIII activity due to increased expression or enhanced potency, respectively, is associated with thrombosis despite physiologic vWF levels. Despite these observations, the biochemical mechanisms linking elevated FVIII activity to VT remain unclear. We therefore investigated the procoagulant effect of elevated FVIII in plasma compared to equal-fold increases in FIX and FV and the prothrombotic potential of elevated FVIII in hemophilia A (HA) mice with and without vWF deficiency.
Methods: Pooled normal human plasma (NHP) was supplemented with (r)ecombinant FVIII, FIX, or FV to achieve 1–10-fold normal levels and coagulation was initiated with tissue factor. Thrombin, FIXa, and FXa generation with increased FVIII was compared to equal-fold increases in FIX, an established independent VT risk factor, and FV, which is not associated with increased VT risk. FIXa and FXa generation were measured by ELISA-based quantification of FIXa-antithrombin (AT) and FXa-AT complexes. The in vivo prothrombotic potential of elevated FVIII activity was assessed in HA and HA/vWF deficient (HA/vWF-/-) mice using a ferric chloride-induced femoral vein thrombosis model. Mice were infused with rFVIII to achieve 0%, 30%, 100%, or 500% of normal FVIII activity. Platelet and fibrin accumulation were quantified by intravital confocal microscopy. Thrombus formation was compared with homozygous FV-Leiden (FVL+/+) mice, which have a robust prothrombotic phenotype. To additionally isolate the contribution of FVIII from vWF, HA mice were infused with FVIII-Y1680F, which has ≥10-fold reduced vWF binding efficiency.
Results: Increasing concentrations of FVIII and FIX in NHP enhanced thrombin, FIXa, and FXa generation in a dose-dependent manner. Equal-fold increases in FVIII and FIX concentrations produced similar peak thrombin and endogenous thrombin potential, while elevated FVIII produced shorter lag times at higher factor concentrations, suggesting more rapid propagation of coagulation. Ten-fold normal FVIII concentrations resulted in FIXa and FXa generation of 200% and 170% of NHP, respectively, while 10-fold normal FIX concentrations resulted in FIXa and FXa generation that was 160% and 120% of NHP, respectively. In contrast, 1–10-fold normal FV concentrations produced no appreciable change in thrombin, FIXa, or FXa generation. In HA mice, increasing FVIII activity produced dose-dependent increases in platelet and fibrin accumulation following femoral vein injury. HA mice with 100% FVIII activity restored thrombus formation to that of WT mice. HA mice with 500% FVIII activity exhibited platelet and fibrin accumulation that was 6-fold and 3-fold greater than that of WT mice, respectively, producing a prothrombotic phenotype indistinguishable from FVL+/+ mice (P < 0.1246 and P < 0.5557, respectively). Importantly, the prothrombotic effect of elevated FVIII was preserved in HA/vWF-/- mice. HA/vWF-/- mice with 500% FVIII activity exhibited 4-fold greater platelet and 2-fold greater fibrin accumulation over HA/vWF-/- mice with 100% FVIII activity. Similarly, HA mice with 500% FVIII-Y1680F activity exhibited platelet and fibrin accumulation comparable to HA/vWF-/- mice with 500% FVIII-WT activity, further supporting that increased thrombus formation due to elevated FVIII activity is independent of vWF.
Conclusions: By directly comparing equal-fold increases in FVIII, FIX, and FV, these studies demonstrate that FVIII disproportionately amplifies coagulation relative to equal-fold increases in FIX and FV. Our in vivo studies demonstrate that elevated FVIII activity is sufficient to drive venous thrombosis independent of vWF, supporting FVIII as an independent mechanistic driver of VT rather than a biomarker of inflammation. These findings provide a mechanistic explanation for longstanding epidemiologic observations linking elevated FVIII to VT.
DISCLOSURES
Johnathan Morris, PhD: No relevant conflicts of interest to declare.