Relieving Inhibition of Prothrombinase By TFPIα: a Procoagulant Activity of Unfractionated, Low Molecular Weight, and Nonanticoagulant Heparins

Introduction: Prothrombinase, the complex of factor Xa (FXa) and factor Va (FVa), is inhibited by tissue factor pathway inhibitor (TFPI)α during the initiation of coagulation (Wood JP et al, PNAS 2013). Efficient inhibition of thrombin generation by prothrombinase requires an interaction between the TFPIα basic C-terminus and an acidic region of the FVa B-domain. This acidic region is present in FXa-activated FVa and FVa released from activated platelets, but is rapidly removed by thrombin. Thus, prothrombinase inhibition only occurs during the initiation phase of thrombin generation. As the exosite interaction is charge-dependent, large negatively charged molecules, including unfractionated heparin (UFH), block it, prevent prothrombinase inhibition, and promote thrombin generation. Studies using the negatively charged molecule polyphosphate have suggested a size requirement for blocking this TFPIα activity (Smith SA et al, Blood2010). A similar size-dependence may exist with heparins and could have clinical implications, as currently-used heparins range from long (unfractionated heparin; UFH) to medium (low molecular weight heparins; LMWHs) to short (the antithrombin-binding pentasaccharide fondaparinux). Studies were performed to assess the ability of the LMWHs enoxaparin and dalteparin, fondaparinux, and the nonanticoagulant heparin 2-O, 3-O desulfated heparin (ODSH) to block TFPIα and promote thrombin generation through this mechanism.

Methods: TFPIα inhibition of thrombin generation by prothrombinase, assembled with a form of FVa containing the acidic region of the B domain, was measured in the absence or presence of UFH, enoxaparin, dalteparin, fondaparinux, and ODSH. The effect of these compounds on the direct inhibition of FXa by TFPIα was measured using a FXa chromogenic substrate. The effect of these compounds on thrombin generation in plasma was measured by calibrated automated thrombography using human plasma immunodepleted of antithrombin III and heparin cofactor II (AT3/HCII-depleted plasma).

Results: TFPIα inhibited prothrombinase activity (IC50 = 6.8 nM), and UFH blocked this inhibition (IC50 = 12.5 nM or 14.9 nM at 0.5 or 1 U/mL, respectively). Enoxaparin (0.8 U/mL; IC50 = 30.3 nM) and dalteparin (1 U/mL; IC50 = 29.7 nM) appeared to be more effective at reversing TFPIα inhibition. The reason for this apparent enhanced effect of LMWHs compared to UFH is not clear, as UFH and the LMWHs similarly enhanced the direct inhibition of FXa by TFPIα, and the differential activity was also observed when heparins were normalized to saccharide concentration. The same pattern was observed when measuring thrombin generation in AT3/HCII-depleted plasma, with LMWHs being more procoagulant than UFH. Consistent with TFPIα inhibition being charge-dependent, ODSH promoted thrombin generation similarly to LMWHs in both purified systems and AT3/HCII-depleted plasma. In contrast, clinical doses of fondaparinux had no effect in any assay. In a purified system, ~1000 times the clinical dose of fondaparinux was required to promote thrombin generation.

Conclusion: There is a size-dependence for blocking TFPIα inhibition of prothrombinase using heparins, as the pentasaccharide has no effect. However, both LMWHs and UFH are sufficiently long to express this procoagulant activity at therapeutic doses. In addition, the nonanticoagulant heparin ODSH blocks prothrombinase inhibition by TFPIα. This procoagulant activity is likely most clinically relevant under conditions of antithrombin deficiency, which may result from sepsis, liver failure, or administration of L-asparaginase. Under any of these conditions, UFH, LMWHs, and ODSH may have unanticipated procoagulant activity mediated by blocking TFPIα.