Binding of factor VIIa to the endothelial cell protein C receptor reduces its coagulant activity

Authors

  • J. LÓPEZ-SAGASETA,

    1. Haematology Department and the Division of Cardiovascular Sciences, Laboratory of Thrombosis and Haemostasis, Clínica Universitaria/School of Medicine, Centre for Applied Medical Research, University of Navarra, Pamplona
    Search for more papers by this author
    • 1

      These authors contributed equally to this work.

  • R. MONTES,

    1. Haematology Department and the Division of Cardiovascular Sciences, Laboratory of Thrombosis and Haemostasis, Clínica Universitaria/School of Medicine, Centre for Applied Medical Research, University of Navarra, Pamplona
    Search for more papers by this author
    • 1

      These authors contributed equally to this work.

  • C. PUY,

    1. Haematology Department and the Division of Cardiovascular Sciences, Laboratory of Thrombosis and Haemostasis, Clínica Universitaria/School of Medicine, Centre for Applied Medical Research, University of Navarra, Pamplona
    Search for more papers by this author
  • N. DÍEZ,

    1. Haematology Department and the Division of Cardiovascular Sciences, Laboratory of Thrombosis and Haemostasis, Clínica Universitaria/School of Medicine, Centre for Applied Medical Research, University of Navarra, Pamplona
    2. Physiology Department, School of Medicine, University of Navarra, Pamplona, Spain
    Search for more papers by this author
  • K. FUKUDOME,

    1. Department of Immunology, Saga Medical School, Saga, Japan
    Search for more papers by this author
  • J. HERMIDA

    1. Haematology Department and the Division of Cardiovascular Sciences, Laboratory of Thrombosis and Haemostasis, Clínica Universitaria/School of Medicine, Centre for Applied Medical Research, University of Navarra, Pamplona
    Search for more papers by this author

José Hermida. Laboratory of Thrombosis and Haemostasis, Division of Cardiovascular Sciences, Centre for Applied Medical Research (CIMA), University of Navarra. Avenida Pío XII 55, 31008 Pamplona, Spain. Tel.: +34948194700; fax: +34948194716; e-mail: jhermida@unav.es

Abstract

Summary. Background: Endothelial cell protein C receptor (EPCR) binds protein C through its γ-carboxyglutamic acid (Gla) domain and enhances its thrombin–thrombomodulin complex-dependent activation. So far, only protein C/activated protein C has been shown to interact with EPCR. Factor VII (FVII), the coagulation trigger upon tissue factor (TF) interaction, is a serine protease whose Gla domain is highly homologous to the Gla domain of protein C. Objectives:  To characterize the binding of FVII/FVIIa to EPCR and its functional consequences. Methods and results:  We demonstrated by surface plasmon resonance (SPR) that FVII/FVIIa binds to EPCR through its Gla domain. At therapeutic concentrations, FVIIa reduced the activation of protein C by 40%. Soluble EPCR (sEPCR) was also able to prolong dose-dependently the clotting time induced by the FVIIa–TF complex. SPR and amidolytic experiments showed that FVIIa is able to interact simultaneously with TF and EPCR, thus ruling out the possibility that the effect of EPCR on clotting time was due to the inhibition of the binding between FVIIa and TF. sEPCR inhibited dose-dependently the activation of FX by the FVIIa–TF complex. Notably, blocking the binding site of EPCR on the endothelial surface increased the generation of FXa 2-fold. Conclusions: EPCR binds to FVII/FVIIa and inhibits the procoagulant activity of the FVIIa–TF complex.

Introduction

Endothelial cell protein C receptor (EPCR) is an endothelial cell membrane glycoprotein that shares homology with the major histocompatibility complex class 1/CD1 family [1]. EPCR binds protein C and activated protein C (APC) with high affinity [2]. Binding enables zymogen to be more efficiently activated by the thrombin–thrombomodulin complex [3]. APC inhibits coagulation by cleaving factors Va (FVa) and FVIIIa. The interaction of protein C with EPCR is Ca2+ and Mg2+ dependent, and involves the γ-carboxyglutamic acid (Gla) domain of protein C [4] and a small clustered patch of residues on both α-helices of EPCR [5]. FVIIa is a serine protease that, upon binding to tissue factor (TF), triggers the coagulation cascade. The Gla domain of FVIIa exhibits an important degree of homology with the Gla domain of protein C, and all the residues directly involved in the binding of protein C to EPCR are conserved in FVII. Moreover, the crystal structures of the Gla domains of both protein C and FVII show that the spatial disposition of the key residues necessary for protein C to bind to EPCR are conserved in FVII [1,6]. However, the ability of FVII to bind to EPCR has not been addressed so far. While our studies to test this interaction were in progress, two articles reported that binding was possible [7,8], although a detailed description of the nature of the interaction or its functional implications was not provided. We now provide evidence that EPCR is able to bind FVII and FVIIa with similar affinity to protein C/APC. Upon complex formation, FVIIa undergoes a change in catalytic activity that reduces its ability to activate FX even when complexed with TF.

Materials and methods

Expression of human soluble EPCR in Pichia pastoris

Soluble EPCR (sEPCR) was expressed in P. pastoris and purified as described elsewhere, with minor changes. sEPCR site-directed mutagenesis was performed as previously described [9,10].

Biomolecular interaction analysis by surface plasmon resonance

All interaction experiments were carried out by surface plasmon resonance (SPR) technology in a BIAcore X Biosensor (BIAcore AB, Uppsala, Sweden). Wild-type and mutant sEPCR forms were captured through the anti-EPCR monoclonal antibody (mAb) RCR-2 immobilized on a CM5 chip (BIAcore) [11]. FVII (Enzyme Research Laboratories, South Bend, IN, USA) or FVIIa (Novo Nordisk, Bagsvaerd, Denmark) was injected in HEPES buffer with 3 mm CaCl2 and 0.6 mm MgCl2 (HBS-T).

Protein C and APC (Enzyme Research Laboratories) binding to sEPCR was also analyzed under the conditions used to study the binding of FVII/FVIIa. Gla-domainless FVIIa (Enzyme Research Laboratories) and other Gla domain-containing proteins, i.e. FX, FIX, protein S and protein Z (Enzyme Research Laboratories), were also tested for binding to sEPCR.

FVIIa and APC (Lilly, Indianapolis, IN, USA) were inactivated with biotinylated Phe-Pro-Arg (FPR) chloromethylketone (PPACK-b) (Calbiochem, La Jolla, CA, USA) [12] and captured on an SA chip (BIAcore). Purified sEPCR was injected and the interaction was monitored. sEPCR alanine variants were also immobilized on RCR-2 in flow cell 2 (FC2) prior to the injection of FVIIa. Kinetic and affinity analysis was performed using biaevaluation software 3.2 RC1 (BIAcore).

Interaction of sEPCR with the FVIIa–TF complex

FVIIa-PPACK-b was captured on FC2 of an SA chip until 1500 RU were reached. Soluble TF (sTF, provided by J.J. Hathcock, Mount Sinai Medical Center, New York, NW, USA) was subsequently injected on both FC1 and FC2. Finally, sEPCR was injected over both cells and the interaction was monitored.

Flow cytometric analysis of the binding of FVIIa to EPCR on endothelial cells

FVIIa was labeled with fluorescein-PPACK (FVIIa-PP*) and incubated with endothelium-derived EA.hy926 cells (provided by C. J. Edgell, University of North Carolina, Chapel Hill, NC, USA), and fluorescence was then determined on a FACSCalibur (BD Biosciences, San Jose, CA, USA) [13]. Experiments in which the cells were preincubated with RCR-252 mAb, sEPCR, unlabeled FVIIa, APC or protein Z prior to the addition of FVIIa-PP* were also performed, as were experiments in the presence of EDTA.

A set of beads labeled with known fluorescein intensities (SpheroTM, BD Biosciences) was run with every set of samples. A calibration curve was constructed with the median channel fluorescence intensities and the corresponding numbers of molecules of equivalent fluorescein of the beads.

To calculate the apparent dissociation constant (KDapp), FVIIa-PP* was incubated with the cells at concentrations between 12.5 and 400 nm. The data were analyzed using enzfitter (Biosoft, Cambridge, UK).

Effect of FVIIa on APC generation

The effect of FVII, FVIIa or Gla-domainless FVIIa on protein C activation on the surface of EA.hy926 cells was studied. The experiment was similar to previously described procedures, except that 0.2 mm ZnCl2 was used to arrest FVIIa activity [14].

Effect of sEPCR on prothrombin time

Purified sEPCR was added to pooled human plasma, and clotting was initiated by adding a thromboplastin reagent (Innovin; Dade Behring, Schwalbach, Germany). Clotting times were determined with an ACL 200 coagulation analyzer (Instrumentation Laboratory, Warrington, UK).

Influence of sEPCR on the amidolytic activity of FVIIa in the absence or presence of TF

We initially studied the effect of increasing concentrations of sEPCR on the enzymatic activity of 100 nm FVIIa towards 8 mm S-2288 (Chromogenix, Milan, Italy). Absorbance at 405 nm was monitored in a microplate reader (iEMS Reader MF, Labsystems, Helsinki, Finland). We also performed the experiment in the presence of RCR-252 mAb, and analyzed the effect of sEPCR on Gla-domainless FVIIa amidolytic activity. The amidolysis of S-2288 and S-2366 (Chromogenix) by 100 nm FVIIa was also studied in the presence and absence of 2 μm sEPCR. Initial rates (Vo) were calculated and fitted to Michaelis–Menten curves using enzfitter.

On the other hand, FVIIa was incubated with a saturating concentration of sTF in the presence or absence of sEPCR, and its activity towards S-2288 was monitored.

Effect of sEPCR on the activation of FX by the FVIIa–TF complex

FVIIa, sTF or TF-expressing H727 cells and FX at varying concentrations, with or without sEPCR, were incubated for 5 min. EDTA, 12.5 mm, was added to arrest FVIIa activity. FXa generation was estimated by proteolysis of S-2765 (Chromogenix).

Additional experiments were performed in the presence of RCR-252. The Michaelis–Menten constants (Km) and maximum rates (Vmax) of FX activation with or without sEPCR were calculated using enzfitter. For catalytic constant (kcat) calculation purposes, a standard curve of proteolyzed S-2765 was constructed. Finally, the calculation of the inhibition constant (kI) of sEPCR for the activation of FX by FVIIa was performed according to the non-competitive inhibition model.

Effect of EPCR on FX activation on the surface of endothelial cells

EA.hy926 cells, 5 × 106 mL−1, were incubated for 15 min with FVIIa and FX in the presence or absence of RCR-252. After addition of 12.5 mm EDTA, cells were removed, and 0.6 mm S-2765 was added to monitor FXa activity in the microplate reader. Additional experiments using a mAb able to prevent binding of FVIIa to TF (American Diagnostica, Greenwich, CT, USA) were performed.

Results

Analysis of the interaction between FVII/FVIIa and sEPCR by SPR

Binding of FVIIa to sEPCR was undoubtedly detected (Fig. 1A). The experimental data do not fit a simple 1:1 Langmuir model, but do fit a two-state conformational change model (chi-squared of global fittings <0.7). According to the rate constants (Table 1), a KDapp of 37.0 ± 12.5 nm (mean ± SD, = 3) was calculated. The behavior of zymogen FVII was identical to that of FVIIa (Fig. 1A and Table 1).

Figure 1.

 Surface plasmon resonance analysis of the binding of factor VII (FVII)/FVIIa to soluble endothelial cell protein C receptor (sEPCR). sEPCR was captured on a CM5 chip through RCR-2 mAb. (A) Binding of FVIIa and FVII (10, 20, 40, 60, 80 and 100 nm) to sEPCR was registered. The binding of activated protein C (APC) and protein C at 20, 40, 60, 80 and 100 nm was also studied. (B) FVIIa-PPACK-b or APC-PPACK-b were captured on an SA chip, and sEPCR was injected at 20, 40, 60, 80 and 200 nm. A representative experiment with at least two independent repeats is shown. Black lines represent experimental data, and gray lines represent fittings.

Table 1.   Surface plasmon resonance (SPR) kinetic rate constants of the interaction of factor VII (FVII)/FVIIa and protein C/APC with sEPCR
AnalyteLigandKa1 (m−1 s−1) × 105Kd1 (s−1)Ka2 (s−1) × 10−3Kd2 (s−1) × 10−3KDapp (nm)
  1. NA, not applicable; sEPCR, soluble endothelial cell protein C receptor; APC, activated protein C.

  2. Interactions were fitted to a two-state conformational change kinetic model* or to a 1:1 model.

FVII*sEPCR7.1 ± 1.30.208 ± 0.03025.2 ± 20.91.9 ± 0.528 ± 11
FVIIa*sEPCR4.8 ± 1.60.212 ± 0.02726.8 ± 4.02.1 ± 0.437 ± 13
sEPCRFVIIa-PPACK-b13.4 ± 0.90.212 ± 0.005NANA158 ± 7
Protein C*sEPCR6.1 ± 1.60.121 ± 0.0044.4 ± 0.51.5 ± 0.368 ± 8
APCsEPCR11.0 ± 1.60.063 ± 0.003NANA58 ± 11
sEPCRAPC-PPACK-b12.9 ± 0.10.198 ± 0.013NANA156 ± 25

Ca2+ and Mg2+ were needed for optimum binding, as this was not detected in the absence of the former and was weaker (KDapp = 315.6 ± 103.0 nm, = 3) in the absence of the latter. Gla-domainless FVIIa was not able to bind sEPCR under our conditions (data not shown). Finally, we were unable to detect any interaction between sEPCR and other Gla domain-containing proteins, i.e. FX, FIX, protein S, and protein Z (data not shown).

With our settings, the interaction of sEPCR with protein C (Fig. 1A and Table 1) showed a similar affinity (KDapp = 68.2 ± 7.9 nm, = 3), which concurs with the findings of previous reports [2]. The binding of APC to sEPCR showed a similar affinity as seen for the binding of protein C and of FVII/FVIIa to sEPCR (Fig. 1A and Table 1).

When FVIIa-PPACK-b or APC-PPACK-b were immobilized and sEPCR was used as the analyte, the resulting affinities were similar (Fig. 1B and Table 1), reinforcing the notion that FVII/FVIIa and protein C/APC bind to EPCR with similar affinity.

Interaction of FVIIa with alanine mutants of sEPCR

All alanine EPCR variants apparently bound to RCR-2 mAb like wild-type sEPCR. Mutants with alanine substitution of residues L82, E86, R87, F146A, Y154A, T157A and R158A displayed an extremely reduced ability to bind FVIIa (Fig. 2A). Mutants R81A, V83A, Q85A, Q150, R156 and E160A bound FVIIa, but their sensorgrams showed major differences from those obtained for the interaction between the wild-type sEPCR and FVIIa (Fig. 2B). The P22A and Q149A mutations did not affect the binding to FVIIa (Fig. 2C). Finally, mutant Y159A displayed an altered pattern of binding, possibly related to a modification in the orientation of close residues that are directly involved in binding, i.e. T157 and R158 (Fig. 2B).

Figure 2.

 Identification of the site of binding of soluble endothelial cell protein C receptor (sEPCR) to factor VIIa (FVIIa) by surface plasmon resonance. The different sEPCR variants were captured on the RCR-2 coated CM5 chip to subsequently inject 100 nm FVIIa (gray line). As a reference, binding of FVIIa to the wild-type sEPCR (black line) was also monitored. (A) EPCR mutants with non-detectable or very reduced binding. (B) EPCR mutants with reduced binding. (C) EPCR mutants with normal binding.

Binding of protein C to different sEPCR alanine variants (R81A, L82A, V83A, E86A, R87A, F146A, Y 154A, R156A, T157A, R158A and E160A) was analyzed, and the interaction was found to be as previously reported [1,5] (data not shown).

FVIIa binding to endothelial cells

EA.hy926 cells bound FVIIa-PP* (Fig. 3A). FVIIa-PP* binding was absent in the presence of sEPCR or RCR-252, and was also abolished in the absence of Ca2+ and Mg2+ (Fig. 3B). Binding of FVIIa-PP* was unaffected by the preincubation of cells with 100 μg mL−1 anti-TF mAb (not shown).

Figure 3.

 Binding of factor VIIa (FVIIa) to endothelial cell protein C receptor (EPCR) on the endothelial surface. (A) EA.hy926 cells (1 × 106 mL−1) were incubated with (solid line) or without (dotted line) 400 nm FVIIa-PP*. A representative experiment is shown. (B) Cells were incubated with 100 nm FVIIa-PP* with or without 0.5 μm soluble EPCR (sEPCR) or 100 μg mL−1 RCR-252. FVIIa-PP* binding in the presence of 10 mm EDTA was also analyzed. The mean ± SD of three independent experiments is shown. (C) Effect of unlabeled FVIIa, activated protein C or protein Z on the binding of 100 nm FVIIa-PP* to cells. The mean ± SD of at least two independent experiments is shown. (D) Cells were incubated with increasing concentrations of FVIIa-PP*, to determine the KDapp. A representative experiment with three independent repeats is shown.

APC and unlabeled FVIIa inhibited the binding of FVIIa-PP* to cells to a similar extent, whereas protein Z had no effect (Fig. 3C). Finally, the equilibrium constant KDapp of the FVIIa-PP* binding to cells was found to be 42.1 ± 4.2 nm (= 3).

Effect of FVIIa on APC generation on endothelial cells

Increasing concentrations of FVIIa dose-dependently reduced the activation of protein C on the surface of EA.hy926 cells (Fig. 4). When protein C and FVIIa were incubated at roughly equimolar concentrations, the amount of APC produced by EPCR-dependent generation decreased to around 50%. When FVII was used, a similar inhibitory effect on APC generation was observed. APC generation was unaffected by Gla-domainless FVIIa. The experiment was also performed using human aortic endothelial cells instead of EA.hy926 cells, and the results were similar (not shown).

Figure 4.

 Effect of factor (VIIa) FVIIa on the activation of protein C on the endothelial surface. Protein C, 60 nm, was incubated with 0.4 nm thrombin on the surface of EA.hy926 cells with or without FVIIa or RCR-252. Once thrombin and FVIIa activities had been arrested, the amidolytic activity of activated protein C on 0.4 mm S-2366 was monitored to calculate the initial rate (Vo) of the activation of protein C. The Vo in the absence of FVIIa and RCR-252 was 100%. The mean ± SD of three independent experiments is shown.

Effect of sEPCR on prothrombin time

As shown in Table 2, the clotting time induced by the FVIIa–TF complex was dose-dependently prolonged by sEPCR. The effect was reversed by mAb RCR-252 (not shown).

Table 2.   Effect of sEPCR on the prothrombin time in human plasma
sEPCR (nm)Prothrombin time (s)
  1. sEPCR, soluble endothelial cell protein C receptor.

  2. Values are the mean ± SD of at least two determinations.

015.9 ± 0.7
22519.8 ± 2.5
45022.5 ± 2.8
90027.0 ± 2.5
180039.8 ± 6.4
360053.7 ± 3.5
7200101.2 ± 10.8

FVIIa complexed with TF is able to bind EPCR

sEPCR was injected onto an FVIIa-PPACK-b-captured SA chip, and its interaction was compared with that observed when sTF was injected onto the chip prior to sEPCR. Once the RU corresponding to sTF binding had been subtracted, the binding profile of sEPCR was similar in both conditions (Fig. 5A), which indicates that FVIIa is able to interact simultaneously with TF and EPCR. We analyzed the effect of sEPCR on the amidolytic activity of FVIIa complexed with saturating concentrations of sTF, and were able to demonstrate an additional increase in the amidolytic activity of the FVIIa–TF complex in the presence of sEPCR, which confirmed that FVIIa can bind TF and EPCR simultaneously (Fig. 5B).

Figure 5.

 Interaction of soluble endothelial cell protein C receptor (sEPCR) with factor VIIa (FVIIa) complexed with tissue factor (TF) and inhibitory effect on FX activation on the endothelial surface. (A) FVIIa-PPACK-b, 1500 RU, was captured onto an SA chip, and 500 nm soluble tissue factor (sTF) was injected. Subsequently, 10, 30, 60, 100, 150 and 250 nm sEPCR was injected. Once the signal of sTF was subtracted (discontinuous line), the interaction was compared with that observed in the absence of sTF (solid line). A representative experiment with two independent repeats is shown. (B) The activity of 5 nm FVIIa with or without 100 nm sTF (saturating concentration), 1 μm sEPCR or both, towards 1 mm S-2288 was monitored. The mean ± SD of three independent experiments is represented. mAU indicates milli-absorbance units at 405 nm. (C) The activation of 100 nm FX by 5 pm FVIIa with or without 50 μg mL−1 RCR-252 mAb on EA.hy926 cells was monitored.

FVIIa undergoes a change in amidolytic activity upon binding to sEPCR

Incubation of increasing concentrations of sEPCR with FVIIa demonstrated a dose-dependent increase in its amidolytic activity towards S-2288, which was prevented by the blocking mAb RCR-252 (not shown). The effect of sEPCR on FVIIa was abolished when the Gla domain was removed. sEPCR enhanced by 2.4-fold and 3.1-fold the catalytic efficiency of FVIIa towards S-2288 and S-2366 respectively (Table 3).

Table 3.   Effect of soluble endothelial cell protein C receptor (sEPCR) on the rates of factor VIIa (FVIIa)-mediated hydrolysis of chromogenic substrates
 S-2288S-2366
Km (mm)kcat (s−1)kcat/Km (m−1 s−1) Km (mm)kcat (s−1)kcat/Km (m−1 s−1)
  1. The chromogenic activity of 100 nm FVIIa towards 0–8 mm S-2288 or S-2366 was determined in the absence or presence of 2 μm sEPCR. Values are the mean ± SD of at least two determinations.

FVIIa10.7 ± 1.22.4 ± 0.6221 ± 379.6 ± 0.61.0 ± 0.1107 ± 2
FVIIa + sEPCR7.9 ± 0.04.3 ± 0.3540 ± 357.3 ± 0.82.3 ± 0.2333 ± 6

Effect of sEPCR on the coagulant activity of FVIIa

sEPCR notably reduced the ability of the FVIIa–sTF complex to activate FX by decreasing its catalytic efficiency. When H727 cells were used, sEPCR also reduced the FXa generation rate, although to a lesser extent (Table 4).

Table 4.   Effect of soluble endothelial cell protein C receptor (sEPCR) on the rates of factor VIIa (FVIIa)-mediated activation of FX in the presence of soluble tissue factor (sTF) or neoplastic cells
 Km (nm)kcat (s−1)kcat/Km (m−1 s−1 × 10−5)kI (nm)
  1. NA, not applicable.

  2. FX, 15–1000 nm, was activated with FVIIa, at 300 pm, in the presence of 25 nm sTF, with or without 30 nm sEPCR, or at 10 pm in the presence of 1 × 106 mL−1 H727 cells, with or without 125 nm sEPCR. The chromogenic activity of FXa towards 0.6 mm S-2765 was subsequently determined. The values of the kinetic constants correspond to the mean ± SD of at least two determinations.

FVIIa + sTF518.9 ± 176.00.14 ± 0.043.23 ± 2.16NA
FVIIa + sTF + sEPCR556.1 ± 170.30.08 ± 0.031.64 ± 1.1636.3 ± 5.4
FVIIa + H72797.0 ± 4.81.80 ± 0.061.86 ± 0.15NA
FVIIa + H727 + sEPCR88.3 ± 1.00.88 ± 0.180.99 ± 0.20139.1 ± 45.1

We analyzed the role of EPCR in the activation of FX in a more physiologic milieu, i.e. endothelium. FVIIa was able to activate FX on the surface of EA.hy926 cells in a TF-dependent fashion, as activation was completely arrested by an anti-TF inhibitor mAb (not shown). FXa generation underwent a 2-fold increase when RCR-252 was present, and FVIIa was therefore not allowed to bind membrane EPCR (Fig. 5C).

Discussion

The FVIIa–TF complex is the principal initiator of blood coagulation. In physiologic conditions, TF is poorly expressed on the surface of vascular cells that come into contact with flowing blood. In contrast, TF is widely expressed in atherosclerotic plaques, and plaque thrombogenicity is correlated with their TF content. Moreover, soluble forms of TF exhibiting procoagulant activity, alone or on the surface of microparticles, have been related to acute coronary events [15]. EPCR, present on the surface of endothelial cells, is known to bind to the anticoagulant protein C/APC with high affinity [2].

SPR experiments showed that binding of FVII/FVIIa is of the same order of magnitude as that observed with protein C/APC, and that it is also dependent on the presence of Ca2+ and improved by Mg2+. SPR experiments also suggested that the interaction is driven by the Gla domain of FVII/FVIIa, as it cannot be seen in the case of Gla-domainless FVIIa or in the absence of Ca2+. The flow cytometry analysis further confirmed these findings. Accordingly, the alanine scanning approach showed that the residues of EPCR previously identified as being directly involved in the interaction with protein C/APC [1,5] are also involved in the binding to FVII/FVIIa.

A first consequence of the ability of FVII/FVIIa to bind to EPCR through the binding site used by protein C/APC is that high levels of FVII/FVIIa can dose-dependently impair protein C activation by preventing its interaction with EPCR. As impairment of APC generation is associated with a protective effect against bleeding in certain in vivo situations [16–18], this could be an additional beneficial effect of FVIIa in helping to achieve the cessation of hemorrhage, and could establish the rationale for assaying an increasing dosage of FVIIa to enhance its TF-independent hemostatic effect [19].

sEPCR was found to induce a delay in the plasma prothrombin time. The ability of FVIIa to bind to sEPCR prompted us to investigate the consequences of this interaction for coagulation. Upon binding to EPCR, the ability of the FVIIa–TF complex to activate FX was reduced. It is noteworthy that the finding that the EPCR expressed on the endothelial surface was able to reduce FXa generation could be physiologically relevant, in that it could constitute an additional control mechanism to avoid excessive generation of FXa induced by traces of TF in situations when the vascular integrity is not compromised. It could also be important in situations in which sTF is present in plasma. Circulating sTF can arise not only by membrane shedding but also by alternative splicing [20]. On the other hand, taking into account the high inhibitory activity of sEPCR towards the FVIIa–TF complex, this could also play an anticoagulant role when its levels are raised.

The mechanism through which EPCR exerts its anticoagulant effect on the FVIIa–TF complex seems to be a dual one. The noticeable influence of sEPCR on the amidolytic activity of FVIIa suggests that the latter undergoes a structural rearrangement involving its active site that still influences its amidolytic activity when it is complexed to TF. The rearrangement results in the reduction of the catalytic efficiency of FX activation, leaving the Km unaltered. The fact that FVIIa interacts with EPCR and phospholipids (PL) through its Gla domain makes it conceivable that binding of FVIIa to EPCR and PL might be mutually exclusive. As the FVIIa–TF complex is a more efficient activator when FVIIa is also bound to PL [21], at least a part of the effect of EPCR on FX activation could be explained by its ability to prevent the binding of FVIIa to PL. Overall, both arguments fit well with the observation that the effect of EPCR is best seen in the absence of PL, i.e. with sTF. In this sense, it could be argued that the poor PL density on the surface of intact endothelial cells makes that environment suitable for EPCR to exert this anticoagulant effect.

In sum, FVII/FVIIa binds to EPCR with high affinity. Therapeutic levels of FVIIa can partly prevent protein C activation. An active site rearrangement affecting the coagulant activity of FVIIa occurs upon interaction with EPCR, which is still seen in the presence of TF. Endothelium-anchored EPCR may play a role in modulating the coagulant activity of FVIIa. The exogenous administration of sEPCR may be a promising strategy to reduce the coagulant activity in situations in which TF is pathologically overexpressed, such as atherosclerosis, neoplasm, or sepsis.

Acknowledgements

We thank E. Molina for her excellent technical assistance and A. Jaso for his technical help.

Disclosure of Conflict of Interests

This work was supported through the Unión Temporal de Empresas project CIMA and by grants from Instituto de Salud Carlos III (PI020125, PI051178, RECAVA RD/0014/0008), Gobierno de Navarra (12457).

Ancillary