Background: Protein S (PS), a vitamin K-dependent plasma glycoprotein, functions as a cofactor for the anticoagulants activated protein C (APC) and tissue factor (TF) pathway inhibitor alpha (TFPIa), which inhibit factors Va (FVa) and Xa (FXa), respectively. Although it is unclear which of these functions is/are important in vivo, homozygous deficiency of PS is associated with life-threatening thrombosis shortly after birth. FVa and FXa form the prothrombinase complex, which generates thrombin, suggesting that PS has a role in the direct inhibition of thrombin production. However, neither the PS/APC nor PS/TFPIα system alone is effective at inhibiting thrombin generation by prothrombinase. In addition to its role in regulating coagulation, PS also functions as an inhibitor of the complement system. Approximately 60% of plasma PS circulates bound to complement factor C4bp, which blocks its anticoagulant activity. We sought to determine the impact of PS/APC on TFPIα function, and vice versa, using purified protein and plasma-based systems.
Methods and Results: To assess the effect of the combined PS/APC and PS/TFPIα systems on thrombin generation, we supplemented plasma with thrombomodulin, which promotes APC activation. In the absence of thrombomodulin, 5nM TFPIα decreased peak thrombin by 55.1% (33.1±1.9 nM in the presence of TFPIα vs. 73.7±39.9 nM in the absence) and endogenous thrombin potential (ETP) by 35.4% (475±42 nM*min vs. 735±189nM*min). In the presence of thrombomodulin, TFPIα decreased these parameters by 65.7% (11.4 ± 2.6 nM) and 77.5% (107±22 nM*min), respectively, suggesting that APC makes TFPIα a more potent inhibitor of thrombin generation.
We next sought to study each of these PS functions in a purified protein system and in plasma. To study the effect of PS/APC on TFPIα function, we produced a recombinant protein consisting of the first two epidermal growth factor-like domains of PS (EGF1-2), which contain the putative APC binding sites. In a purified protein assay, APC inhibited the rate of thrombin activation by prothrombinase by 19.62±0.01% in the absence of PS and by 34.96±0.02% in the presence of 50nM PS. EGF1-2 dose-dependently reversed the effect of PS, with 75% reversal achieved with the addition of 200nM EGF1-2. Unexpectedly, EGF1-2 had the opposite effect in plasma thrombin generation assays and potently inhibited TF-initiated thrombin generation either in the presence or absence of thrombomodulin. We hypothesize that the EGF1-2 domains also form part of the C4bp binding site, and that addition of EGF1-2 protein resulted in release of PS from C4bp and an increase in the anticoagulant PS pool. In support of this hypothesis, EGF1-2 had no effect on thrombin generation in PS-depleted plasma, which is also depleted of C4bp. When the PS-depleted plasma was supplemented with 150nM PS, EGF1-2 had the expected procoagulant activity (increasing peak thrombin from 50.4±19.9 nM to 90.4±6.0 nM). Notably, even with a saturating concentration of EGF1-2, thrombomodulin and PS significantly decreased thrombin generation, suggesting that PS-TFPIα-mediated FXa inhibition promotes APC-mediated FVa degradation, even if PS cannot directly bind APC.
We similarly assessed the impact of the PS-TFPI function, using a protein from the saliva of black flies, "black fly protease inhibitor" (BFPI), which contains the TFPIα domain that inhibits FXa but lacks the domain that binds PS. BFPI inhibits free FXa similarly to TFPIα, but PS does not promote this inhibition. Like TFPIα, BFPI is a poor inhibitor of thrombin generation by prothrombinase containing thrombin-activated FVa (5 nM BFPI had no impact on thrombin generation in the presence or absence of PS). However, in the presence of APC and PS, 5nM BFPI decreased the maximum rate of thrombin generation by 17.3±3.3%. These data suggest that PS/APC-mediated degradation of FVa promotes TFPIα-mediated inhibition of FXa, regardless of whether PS is able to bind TFPIα.
Conclusions: Our data suggest that the PS-APC and PS-TFPIα systems cooperatively regulate thrombin generation by prothrombinase. While maximal inhibition requires that PS act as a cofactor for both APC and TFPIα, PS-APC independently promotes TFPIα function, and PS-TFPIα separately promotes APC. Based on these data, we propose a model in which PS-APC-mediated inhibition of FVa renders FXa susceptible to TFPIα and vice versa.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.