A Refined Plasma Thrombin Generation Assay Identifies Functional Importance of the First Two EGF-like Domains of the Anticoagulant Protein S

Protein S (PS) is an anticoagulant cofactor for the serine protease Activated Protein C (APC) and the serine protease inhibitor tissue factor pathway inhibitor (TFPI), which inhibit coagulation factors Va (FVa) and Xa (FXa), respectively. FVa and FXa function together to activate thrombin, the enzyme that produces a clot. In addition to these roles, PS also controls the complement system, with ~60% of plasma PS circulating bound to the complement inhibitor C4‐binding protein (C4BP). The remaining ~40% of PS is the anticoagulant pool. Despite its many functions, the physiologic role of PS in coagulation is unclear, because PS concentration does not correlate with plasma thrombin production ex vivo . We sought to develop a thrombin generation assay that is sensitive to PS concentration and to use it to assess the importance of the PS/APC interaction in human plasma. We hypothesized that standard thrombin generation assays are insensitive to PS because APC activation requires the endothelial cell protein thrombomodulin (TM). Thus, we supplemented human plasma with TM. This resulted in a dose‐dependent decrease in thrombin generation when PS was added to PS‐depleted plasma, with a physiologic concentration of 150nM PS causing a 75% reduction. In the absence of TM, only a 12% decrease was observed. We next validated the assay by measuring PS concentration and thrombin generation in plasma samples from 10 healthy controls and 27 patients with Human Immunodeficiency Virus, a condition known to cause PS deficiency. 63% of the patient samples were deficient in PS compared to the control samples (p=0.039). In the absence of TM, plasma PS did not correlate with thrombin generation, consistent with previous reports. In the presence of TM, there was a negative correlation, with decreased PS being associated with increased thrombin peak (p=0.024) and total thrombin produced (p=0.011). Finally, 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 inhibition of thrombin activation was greater in the presence of PS. EGF1‐2 dose‐dependently reversed the effect of PS, with 75% reversal achieved with 200nM EGF1‐2. Unexpectedly, EGF1‐2 had the opposite effect in plasma thrombin generation assays and displayed potent anticoagulant activity. To explain these results, we hypothesize that the EGF1‐2 domains also form part of the C4BP binding site and its addition resulted in release of PS from C4BP, increasing the anticoagulant PS pool. Consistent with this, EGF1‐2 had no effect on thrombin generation in PS‐depleted plasma, which is also depleted of C4BP, but upon PS supplementation, had the expected procoagulant activity. We have developed a PS‐sensitive plasma thrombin generation assay, which we validated using a cohort of PS‐deficient HIV patients. We have also produced a recombinant EGF1‐2 protein which competed with PS for APC binding and inhibited APC function, but had anticoagulant effect in plasma. These data suggest that attempts to pharmacologically inhibit PS must adjust for the C4BP‐bound pool. Support or Funding Information Supported by NHLBI grants HL056652, HL138179, HL129193.