Zn²(+) -containing protein S inhibits extrinsic factor X-activating complex independently of tissue factor pathway inhibitor

Abstract

Background: Protein S (PS) has direct anticoagulant activity, independently of activated protein C (APC). The mechanisms underlying this activity remain unclear, because PS preparations differ in activity, giving rise to conflicting results. Some purification procedures result in loss of intramolecular Zn²(+) , which is essential for inhibition of prothrombinase.

Objective: To investigate the inhibition of extrinsic factor (F)Xase by Zn²(+) -containing PS.

Methods: Purified component extrinsic FXase assays were used to determine FXa generation in the presence and absence of PS and/or tissue factor pathway inhibitor (TFPI). Binding assays, immunoblots and thrombin generation assays in plasma supported the FXase data.

Results: Zn²(+) -containing PS potently inhibited extrinsic FXase in the presence of saturating phospholipids, independently of TFPI, whereas inhibition of extrinsic FXase by Zn²(+) -deficient PS required TFPI. Immunoblots for FXa and functional assays showed that Zn²(+) -containing PS inhibited primarily the quantity of FXa formed by tissue factor (TF)-FVIIa, rather than FXa amidolytic activity. Zn²(+) -containing PS, but not Zn²(+) -deficient PS, bound to TF with high affinity (K(dapp) = 41 nm) and targeted TF function. Binding of PS to FVIIa was negligible, whereas PS showed appreciable binding to FX. Increasing FX concentrations 10-fold reduced PS inhibition five-fold, suggesting that PS inhibition of FXase is FX-dependent. PS also exhibited TFPI-independent and APC-independent anticoagulant activity during TF-initiated thrombin generation in plasma.

Conclusions: PS that retains native Zn²(+) also retains anticoagulant functions independently of APC and TFPI. Inhibition of extrinsic FXase by PS at saturating levels of phospholipids depends on PS retention of intramolecular Zn²(+) , interaction with FX, and, particularly, interaction with TF.

Fig. 1

Effect of PS and TFPI on extrinsic FXase activity. TF/FVIIa/phospholipids were incubated ± PS (80–100 nM) and/or TFPI (100 pM) as described in Methods. Following addition of FX, aliquots were removed over time and quenched in EDTA. FXa generation was measured with chromogenic substrate. % Inhibition was estimated by comparing curves at 100 s. (A) Zn²⁺-containing affinity-purified PS. Seven additional affinity-purified PS independently inhibited FXase, did not enhance inhibition by TFPI, and were unaffected by neutralizing antibody against TFPI. (B) PS conventionally-purified using anion exchange. (C) Commercial Zn²⁺-deficient PS before and after reconstitution of Zn²⁺. Zn²⁺-reconstituted PS inhibited FXase independently of TFPI and did not enhance TFPI inhibition. Curves are representative of two or more experiments, each performed in duplicate.

Fig. 2

PS inhibition of the generation of FXa protein by FXase. TF/FVIIa/phospholipids were incubated ± 100 nM PS prior to addition of FX as described in Methods. Aliquots were quenched in parallel in SDS or EDTA after 50–150 s. (A) Blot for FXa generated by the extrinsic FXase complex. SDS samples were subjected to immunoblotting for FXa antigen. After transfer, membranes were cut horizontally to separate FXa (MW~49 kDa) from excess FX (MW~67 kDa) prior to development with anti-FX antibodies and chemiluminescent detection. (B) EDTA-quenched samples were assayed for FXa activity. Results were confirmed in several experiments.

Fig. 2

PS inhibition of the generation of FXa protein by FXase. TF/FVIIa/phospholipids were incubated ± 100 nM PS prior to addition of FX as described in Methods. Aliquots were quenched in parallel in SDS or EDTA after 50–150 s. (A) Blot for FXa generated by the extrinsic FXase complex. SDS samples were subjected to immunoblotting for FXa antigen. After transfer, membranes were cut horizontally to separate FXa (MW~49 kDa) from excess FX (MW~67 kDa) prior to development with anti-FX antibodies and chemiluminescent detection. (B) EDTA-quenched samples were assayed for FXa activity. Results were confirmed in several experiments.

Fig. 3

Phospholipid dependence of inhibition of extrinsic FXase by PS. TF/FVIIa with varying phospholipid concentrations were incubated ± 100 nM Zn²⁺-containing PS or ± 333 nM Zn²⁺-deficient PS as described in Methods. FX was added. Aliquots were removed between 40–120 s and quenched in 20 mM EDTA. FXa generation was measured with chromogenic substrate. % inhibition relative to the controls without PS at each phospholipid concentration was calculated by comparing activity progress curves at 120 s. The TF preparation contributed a small amount of phospholipids, such that inhibition by PS in the absence of exogenous phospholipids was similar to or greater than inhibition in the presence of 5 μM exogenous phospholipids. Results are representative of several experiments.

Fig. 4

Binding of Zn²⁺-containing PS to FX and competition between FX and PS for binding to the FXase complex. (A) Fluid phase assay for binding of FX to Zn²⁺-containing PS. Assay described in Methods. Binding constant (Kd_app) was determined using nonlinear regression analysis for one site binding. Results from two experiments are combined. (B) Inhibition of extrinsic FXase in the presence of varying [FX]. TF/FVIIa/phospholipids were incubated ± PS as described in Methods except that varying concentrations of FX were used. Aliquots were removed over time and quenched in EDTA. FXa generation was measured with chromogenic substrate. Results were confirmed in a separated experiment.

Fig. 5

Functional assay to identify the most significant target for PS inhibition of FXase. The FXase assay described in Methods was modified such that different FXase components were incubated for 10 min with 100 nM PS in the presence of saturating phospholipids prior to addition of the remaining FXase components to start FXa generation. Aliquots were removed over time, quenched in 20 mM EDTA, and FXa generation was measured using a chromogenic substrate. (A) FX/phospholipid was preincubated ± PS before TF/FVIIa were added. (B) TF/phospholipid ± PS were preincubated before FVIIa/FX were added. (C) FVIIa/phospholipid were preincubated with PS before TF/FX were added. (D) TF/FVIIa/phospholipid ± PS were preincubated before FX was added. All data were collected within 1 h using the same reagents in each set, and confirmed in a separate set of experiments performed in duplicate on a different day. Lag times were taken as the time when thrombin generation reached a threshold signal of 100 fluorescence units/min.

Fig. 6

Binding of sTF to Zn²⁺-containing PS. (A) Fluid phase assay described in Methods. Data from two experiments are combined. (B) PS binding to immobilized sTF as described in Methods. Binding constants were determined using nonlinear regression analysis for one site binding. Kd_app=41±22 nM, n=9.

Fig. 7. Lag times for thrombin generation in normal plasma and depleted plasmas

(A) Thrombin generation in pooled normal human plasma (NHP, light grey bars), PS-depleted plasma (PSdP, striped bars), or TFPI-depleted plasma (TFPIdP, dark grey bars). Coagulation was initiated with 0.3 pM TF, 15 mM CaCl₂, and 10 μM phospholipids as described in Methods. Additions prior to initiation of coagulation, where indicated, were 67 nM PS, 500 pM TFPI, or sufficient anti-TFPI (αTFPI) to neutralize all TFPI. From thrombin generation profiles, the parameter most sensitive to PS was lag time, as plotted here. Experiments were performed on each of 4 days, and each bar represents 4 to 9 assays. Where statistical significance (p<0.05) was achieved in a particular plasma, the p value is given above the relevant comparison with the parent plasma. (B) Lag times for thrombin generation in PSdP reconstituted with Zn²⁺-containing PS (vertical striped bars) or Zn²⁺-deficient PS (cross-hatched bars) at the concentrations indicated (in nM). αTFPI, where indicated, was incubated in the PSdP + PS for 8 min before procoagulants were added.

Fig. 8

Schematic of mechanisms by which PS may inhibit extrinsic FXase. (Left) Competition for phospholipids was suggested in Fig. 3. (Center) TFPI-independent interaction of Zn²⁺-containing PS with TF and FX was suggested in Figs. 4–6. (Right) TFPI-dependent interaction of Zn²⁺-deficient PS was suggested in Fig. 1B,C and in published data [10]. This interaction may lead to enhanced FXa inhibition, based on published data [10,18].