We thank Drs Hemker and Beguin for their comments on our recent paper . In our introduction, we stated that “acceleration of factor (F)Xa inhibition via conformational activation of anthrombin is generally accepted as the predominant therapeutic mechanism for low-molecular-weight forms of heparin”. While our experiments primarily address the mechanisms of action for LMWH in plasma, the writer's conclusion that the predominant activity of heparin is due to ‘increased thrombin inactivation and not to decreased prothrombin conversion’ is not supported by our data.
Drs Hemker and Beguin extrapolate a pentasaccharide concentration for the unfractionated heparin (UFH) preparation in Table 1 of our paper to support their contention that the acceleration of thrombin inhibition is the primary mechanism of action for heparin. However, caution should be employed in comparing EC50 values between different heparin forms. The concentrations are expressed as molar values only to provide a common scale for these disparate forms of heparin. With the exception of Fondaparinux, these molar concentrations are based on an average molecular weight for each heparin preparation. They do not take into account the heterogeneity and relative polydispersity of UFH preparations. Thus, these molarity assumptions do not translate easily into pentasaccharide concentrations, particularly as the longer ‘high-affinity’ chains in UFH may contain more than one pentasaccharide sequence [2, 3]. The more relevant and accurate comparisons reside within each heparin type based on the tissue factor concentration (Table 1), absence or presence of antithrombin, and their respective therapeutic ranges (Table 2).
In contrast to LMWH, Fondaparinux and UFH demonstrate similar potency under FIX-dependent or independent conditions (Table 1), and are both dramatically (42- to 62-fold) less potent in the absence of antithrombin (Table 2). These results suggest that the relevant mechanism(s) of action for these heparins in human plasma do not depend on inhibition of the intrinsic tenase complex, and as expected, are highly antithrombin-dependent. In our Western blot analysis of plasma thrombin generation, 0.012 μm or 0.04 U mL–1 UFH (specific activity approximately 165 U mg–1) is present in Fig. 2E, which is approximately eight-fold below the therapeutic range. While this sub-therapeutic UFH concentration accelerated the appearance of the TAT complex, it did not prevent depletion of the prothrombin band, and only modestly reduced thrombin generation. In comparison, 0.31 μm or 0.92 U mL–1 of UFH is present in Fig. 2F, which demonstrates essentially complete inhibition of prothrombin consumption. A similar, although less complete reduction of prothrombin consumption is observed with the low-molecular-weight forms of heparin at their EC50 values (Fig. 2B–D). Although 0.92 U mL–1 UFH represents a ∼25-fold higher concentration than the EC50 value for inhibition of thrombin generation (not 1200- to 1600-fold as stated in the letter), it is only modestly higher than the therapeutic range for UFH in venous thromboembolic disease (0.3–0.7 U mL–1), and substantially lower than levels expected during cardiopulmonary bypass (> 2.0 U mL–1) . Thus, the findings in Fig. 2F are relevant to expected plasma concentrations during the clinical usage of UFH. While these data do not specifically address the potential effects of accelerated thrombin inhibition on cofactor activation in the lag (initiation) phase, they do suggest that inhibition of prothrombin consumption is a major effect of all therapeutic heparin forms.
The complexity of heparin effects in plasma makes definitive identification of the critical antithrombotic mechanisms difficult, which was part of the motivation for our paper. The clinical efficacy of Fondaparinux (which has trivial effects on thrombin inhibition) makes it difficult to argue that accelerated thrombin inhibition is required for the therapeutic effect of heparin [5, 6]. However, acceleration of thrombin inhibition may reasonably be expected to contribute to the potency of UFH in the thrombin generation assay, which depends on detection of thrombin fluorogenic activity.
As an additional note, in the preparation of this response an error was noted in the Figure 2 legend. The Figure 2 legend should indicate 0.31 μm UFH (not 0.16 μm) for panel F. The accompanying text correctly states the UFH concentration in this experiment.