Inhibitory mechanism of the protein C pathway on tissue factor-induced thrombin generation. Synergistic effect in combination with tissue factor pathway inhibitor.

The effects of the components of the protein C pathway on thrombin generation were studied in a reconstituted model in which thrombin is generated by factor VIIa and relipidated tissue factor (TF) via the activation of the purified coagulation factors X, IX, VIII, V, and prothrombin. The influence of protein C and soluble thrombomodulin on thrombin generation was correlated with factor Xa generation, factor V(a) and factor VIII(a) formation/inactivation, and protein C activation. Thrombin generation initiated by low concentrations of factor VIIa·TF (1.25 pM) occurs in an explosive fashion during a propagation phase which occurs after an initiation phase of ∼1 min in which only traces of thrombin are formed. In the absence of other inhibitors, protein C (65 nM) in combination with high concentrations of soluble thrombomodulin (10 nM) resulted in a reduced rate of thrombin generation during the propagation phase without affecting the initiation phase; the activated protein C generated failed to neutralize prothrombinase activity and did not prevent prothrombin consumption. In the presence of plasma levels of the tissue factor pathway inhibitor (2.5 nM recombinant TFPI), the protein C pathway reduced the rate of thrombin generation, initiated by 1.25 pM factor VIIa·TF, and completely eliminated prothrombinase activity at soluble thrombomodulin concentrations of ≥1 nM. The neutralization of prothrombinase activity coincided with cleavages at Arg-506 and subsequent cleavage at Arg-306 of the factor Va heavy chain by activated protein C. Thus, the protein C pathway combined with TFPI creates a minimal inhibitory potential required to shut down TF-initiated thrombin generation. The protein C pathway constituents did not influence factor Xa generation or factor VIIIa degradation over the interval in which prothrombinaseactivity was neutralized. Our data thus suggest that the protein C pathway regulates thrombin generation solely by the inactivation of factor Va. At low initiating factor VIIa·TF (1.25 pM) and high thrombomodulin concentrations (10 nM), the factor Va heavy chain is cleaved before significant amounts of light chain are generated. The ability of the protein C pathway to inhibit thrombin generation was greatly reduced when the reaction was initiated in the presence of factor Va, supporting the hypothesis that effective down-regulation of thrombin generation by the protein C pathway, in reactions initiated with the procofactor, occurs by prevention of the coexistence of the factor Va heavy and light chains.

The effects of the components of the protein C pathway on thrombin generation were studied in a reconstituted model in which thrombin is generated by factor VIIa and relipidated tissue factor (TF) via the activation of the purified coagulation factors X, IX, VIII, V, and prothrombin. The influence of protein C and soluble thrombomodulin on thrombin generation was correlated with factor Xa generation, factor V(a) and factor VIII(a) formation/inactivation, and protein C activation. Thrombin generation initiated by low concentrations of factor VIIa⅐TF (1.25 pM) occurs in an explosive fashion during a propagation phase which occurs after an initiation phase of ϳ1 min in which only traces of thrombin are formed. In the absence of other inhibitors, protein C (65 nM) in combination with high concentrations of soluble thrombomodulin (10 nM) resulted in a reduced rate of thrombin generation during the propagation phase without affecting the initiation phase; the activated protein C generated failed to neutralize prothrombinase activity and did not prevent prothrombin consumption. In the presence of plasma levels of the tissue factor pathway inhibitor (2.5 nM recombinant TFPI), the protein C pathway reduced the rate of thrombin generation, initiated by 1.25 pM factor VIIa⅐TF, and completely eliminated prothrombinase activity at soluble thrombomodulin concentrations of 1 nM. The neutralization of prothrombinase activity coincided with cleavages at Arg-506 and subsequent cleavage at Arg-306 of the factor Va heavy chain by activated protein C. Thus, the protein C pathway combined with TFPI creates a minimal inhibitory potential required to shut down TF-initiated thrombin generation. The protein C pathway constituents did not influence factor Xa generation or factor VIIIa degradation over the interval in which prothrombinase activity was neutralized. Our data thus suggest that the protein C pathway regulates thrombin generation solely by the inactivation of factor Va. At low initiating factor VIIa⅐TF (1.25 pM) and high thrombomodulin concentrations (10 nM), the factor Va heavy chain is cleaved before significant amounts of light chain are generated. The ability of the protein C pathway to inhibit thrombin generation was greatly reduced when the reaction was initiated in the presence of factor Va, supporting the hypothesis that effective down-regulation of thrombin generation by the protein C pathway, in reactions initiated with the procofactor, occurs by prevention of the coexistence of the factor Va heavy and light chains.
The extrinsic pathway of blood coagulation initiated by tissue factor involves the activation of multiple coagulation factors leading to thrombin generation. The reaction starts with the binding of plasma factor VIIa to tissue factor (TF), 1 an integral membrane protein that is exposed as a result of either vessel wall injury or cytokine activation of endothelial cells and/or peripheral blood monocytes. The membrane bound factor VIIa⅐TF enzyme complex activates the zymogens factor X and factor IX by limited proteolysis (1). Factor IXa combines with factor VIIIa on the membrane surface to form the intrinsic tenase that activates factor X. The factor Xa generated associates with factor Va on a membrane surface to form prothrombinase which activates prothrombin to thrombin. (For reviews on blood coagulation and membrane dependent reactions in blood coagulation see, respectively, Refs. 2 and 3). The thrombin initially formed accelerates further thrombin generation by feedback activation of the procofactors factor V and factor VIII. Deficiencies in factors VII, X, IX, V, VIII, or prothrombin are associated with bleeding. Thrombin stimulates platelets, which secrete their granule contents and aggregate, cleaves fibrinogen to generate the fibrin network, and activates the protransglutaminase factor XIII. The fibrin-platelet aggregate, stabilized by factor XIIIa catalyzed cross-links, forms the hemostatic plug which maintains the integrity of the circulatory system after vessel wall damage.
In normal hemostasis, the procoagulant system is in balance with the anticoagulant and fibrinolytic systems. The anticoagulant system consists of several stoichiometric protease inhibitors that include, at a minimum, the tissue factor pathway inhibitor (TFPI), antithrombin-III (AT-III), and heparin cofactor-II, and the dynamic protein C pathway that involves thrombin, protein C, protein S, and thrombomodulin.
The dynamic negative feedback mechanism provided by the protein C pathway is initiated by the binding of thrombin to thrombomodulin (for a review on thrombomodulin and protein C see Ref. 4). The binding of thrombin to thrombomodulin changes the specificity of the enzyme, resulting in the loss of procoagulant functions and enhancement of protein C activation. The formation of the thrombin-thrombomodulin complex accelerates the rate of protein C activation approximately 400fold. Thrombomodulin is constitutively expressed as a transmembrane protein on the surface of vascular endothelial cells, providing an anticoagulant function at the blood/vessel wall interface (4) and is also present on platelets (5). A recent report of an individual suffering from thrombosis associated with a thrombomodulin mutation leading to a defective protein (6) and the embryonic lethality in mice lacking the thrombomodulin gene (7) identify the significance of thrombomodulin as a natural anticoagulant.
Activated protein C (APC) is thought to exert its inhibitory effect on coagulation by proteolytically inactivating the cofactors of prothrombinase and the intrinsic tenase, factor V(a) and factor VIII(a), respectively (8 -16). Protein S appears to contribute an accessory function toward the APC inactivation of factor Va and factor VIIIa (14,17,18). It has also been reported that protein S directly inhibits prothrombinase activity (19). The importance of both protein C and protein S as anticoagulants is clearly illustrated by the numerous reports of thrombotic complications in either protein C-or protein S-deficient individuals (20,21).
TFPI is a protease inhibitor that regulates coagulation by inhibition of the factor VIIa⅐TF complex activity in a factor Xa-dependent manner (for a review on TFPI see Ref. 22). TFPI is also a reversible, active site-directed inhibitor of factor Xa. The TFPI⅐factor Xa complex binds to the factor VIIa⅐TF complex resulting in the formation of an inhibited TF⅐factor VIIa⅐TFPI⅐factor Xa quaternary complex (23). Although no human deficiencies have been reported, the in vivo relevance of TFPI is supported by experiments that showed the sensitization of rabbits to TF triggered disseminated intravascular coagulation after immunodepletion of TFPI (24).
In the absence of coagulation inhibitors, thrombin generation in a reconstituted model (25) initiated by factor VIIa⅐TF is characterized by an initiation phase, in which the cofactors factor V and factor VIII are activated by traces of factor Xa and thrombin initially produced. After the initiation phase is completed thrombin generation occurs rapidly during the subsequent propagation phase, in which prothrombin is quantitatively activated. The rate of thrombin generation during the propagation phase of reactions initiated by factor VIIa⅐TF in a reconstituted model in the absence of inhibitors is only mildly influenced when the initiator concentration is varied (25). In a previous report we found that TFPI, at physiological concentrations, inhibits thrombin generation by inhibition of factor VIIa⅐TF activity and by the direct inhibition of factor Xa (26). TFPI thus extends the initiation phase of thrombin generation and reduces the rate of thrombin generation during the propagation phase; however, in the presence of the components of the factor IX pathway, thrombin generation still becomes explosive after the TFPI extended lag phase (26). The combination of TFPI/AT-III (for reviews on AT-III see Refs. 27 and 28) at their physiological concentrations (2.5 nM and 3.4 M, respectively) is effective in neutralizing explosive thrombin generation by factor VIIa⅐TF concentrations Յ 10 pM. This inhibition is overcome by increasing the initiator concentration; thus explosive thrombin generation becomes a threshold-mediated event in the presence of the combination TFPI/AT-III (26).
The present study describes the effect of components of the protein C pathway, protein C and thrombomodulin on the tissue factor pathway to thrombin, and the combined actions of the protein C pathway and TFPI.
Proteins-Human coagulation factors X, IX, and prothrombin were isolated from fresh frozen plasma using the general methods of Bajaj et al. (29) and were depleted of trace contaminants and traces of the enzyme products as described (26). Human protein C was purified by heparin-Sepharose and immuno-chromatography (monoclonal antibody ␣HPC-2 coupled to CNBr-activated CL-4B) from the protein C pool eluted in the NaCl gradient of the DEAE-Sepharose column involved in the purification of the vitamin K-dependent clotting factors. Purified protein C was treated for 1 h at 22°C with 20 M FPR-ck; residual FPR-ck and hydrolysis products were removed by extensive dialysis in 0.15 M NaCl, 0.05 M Tris, pH 7.4 (TBS). Aliquots of the protein solution were frozen in liquid nitrogen and stored at Ϫ80°C. Human protein S was purified from the protein S pool of the DEAE-Sepharose column using Blue-Sepharose chromatography as described by Dahlbä ck (30). The purified protein S was passed over anti-protein C and anti-factor X immunoaffinity columns which were previously equilibrated in TBS. The flow-through of the immunoaffinity columns was concentrated on a 3-ml Q-Sepharose column that was equilibrated in TBS. Protein S was eluted from the Q-Sepharose with TBS containing 1 M NaCl. Possible traces of enzyme activity in the preparation were inhibited by treatment for 1 h with 20 M FPR-ck at 22°C. Unreacted FPR-ck and hydrolysis products were removed by extensive dialysis in TBS, and the protein S was stored frozen in aliquots at Ϫ80°C. Human factor V was isolated from freshly frozen human plasma using the method of Nesheim et al. (31). Recombinant factor VIII and recombinant tissue factor (consisting of amino acid residues 1-242 of the human sequence, lacking the cytosolic stack) were provided as gifts from Drs. Shu Len Liu and Roger Lundblad, Hyland Division, Baxter Healthcare Corp (Duarte, CA). Factor VIII was radiolabeled with 125 I using the IODO-GEN (Pierce) method (32) as described previously (25). Factor VIII loses significant cofactor activity following iodination (25). Control experiments showed that the use of an excess of radiolabeled factor VIII to obtain the same amount of active factor VIII in the reaction results in identical thrombin generation profiles. Thus, the inactive factor VIII does not inhibit the reaction (25). Recombinant human coagulation factor VIIa was purchased from NOVO Pharmaceuticals. Recombinant soluble thrombomodulin (Solulin) was provided as a gift from Dr. J. Morser, Berlex (Richmond, CA). Recombinant full-length TFPI produced in Escherichia coli was provided as a gift from Dr. K. Johnson, Chiron Corp. (Emeryville, CA). Purified human factor Xa and purified human-activated protein C were gifts from Dr. R. Jenny and Dr. P. Haley, Hematologic Technologies Inc. (Essex Jct., VT). TLCK anticoagulant peptide (TAP) was provided as a gift from Dr. S. Krishnaswamy, Hematology/Oncology Division, Emory University (Atlanta, GA). Hirudin was provided as a gift from Genentech (South San Francisco, CA).
Coagulation Activation Experiments-Thrombin generation initiated by factor VIIa⅐TF in a reconstituted procoagulant model using plasma protein concentrations was studied as described previously (25). Tissue factor (TF) was relipidated at 0.5 nM into 400 M 75% phosphatidylcholine, 25% phosphatidylserine vesicles, for 30 min at 37°C in 20 mM Hepes, 150 mM NaCl, 2 mM CaCl 2 , pH 7.4 (Hepes/Ca 2ϩ ). The relipidated TF was incubated with factor VIIa for 20 min at 37°C to allow factor VIIa⅐TF complex formation. Factor V and factor VIII were added in microliter amounts of concentrated stock solutions to the equilibrated factor VIIa⅐TF mixture (total volume of addition not exceeding 0.25% of the final reaction volume), and immediately thereafter the reaction was started by the addition of an equal volume of a solution containing factor X, factor IX, and prothrombin, which was prepared in Hepes/ Ca 2ϩ . The zymogen solution was preheated at 37°C for 3 min before addition to the factor VIIa⅐TF, factor V, and factor VIII mixture. When protein C, TFPI, or protein S were included, they were added to the factor X, IX, and prothrombin mixture. Thrombomodulin was added to the factor VIIa⅐TF mixture.
The final concentrations of the proteins in the reaction, chosen to represent mean plasma values, are 160 nM factor X, 90 nM factor IX, 0.7 nM factor VIII, 20 nM factor V, and 1.4 M prothrombin (25). When added, protein C (65 nM), protein S (300 nM), and TFPI (2.5 nM) were also present at their respective plasma concentration (33)(34)(35) in the reaction. The factor VIIa concentrations were varied to change the concentration of the initiator complex, whereas TF was kept constant and in substantial excess at 0.25 nM. The final PCPS concentration in the reaction was 200 M. Following initiation of the reaction, aliquots were withdrawn from the reaction mixture and quenched in either 20 mM EDTA/TBS, pH 7.4, to assay for thrombin formation or in 2% SDS, 0.062 M Tris, 10% glycerol, 0.04% bromphenol blue, pH 6.8, for SDS-PAGE and immunoblotting. Samples quenched in SDS were heated for 5 min at 95°C and stored at Ϫ20°C, prior to SDS-PAGE analyses with intact or reduced disulfide bonds (2% 2-mercaptoethanol, 5 min at 95°C). Assays for thrombin activity were performed using the substrate Spectrozyme TH. Hydrolysis of the substrate was monitored by the change in absorbance at 405 nm using a Molecular Devices V max spectrophotometer. Thrombin generation was calculated from a standard curve prepared using serial dilutions of ␣-thrombin. Prothrombinase concentrations were calculated from the rate of thrombin generation using a k cat of 5016 min Ϫ1 (36). Factor Xa generation was measured in samples quenched in EDTA and excess hirudin using the substrate Spectrozyme Xa. Assay specificity for factor Xa was guaranteed by subtraction of the signal observed in the presence of excess TAP (2.31 M), a specific factor Xa inhibitor. Factor Xa concentrations were calculated from a standard curve prepared by serial dilutions of purified factor Xa. Apparent dissociation constants (K d app ) for prothrombinase were calculated using the equation for the equilibrium relation: . APC generation was measured using substrate S2366 in samples that were withdrawn from the reaction mixture and diluted 5-fold into excess hirudin (7 M), TAP (2.31 M), and 20 mM EDTA/TBS. APC generation was calculated from a standard curve prepared using serial dilutions of purified APC. Expected rates of enzyme formation were calculated using the Michaelis-Menten equation: , in which V is the velocity of product formation, V max the maximum product formation at saturating substrate concentrations, [E] the enzyme concentration, [S] the substrate concentration, and K M the Michaelis constant for the particular enzyme-substrate conversion. The samples quenched in SDS were subjected to SDS-PAGE on 4 -12% polyacrylamide gels as generally described by Laemmli (37). Following SDS-PAGE the proteins were transferred to nitrocellulose membranes for immunoblot analysis using general techniques described by Towbin et al. (38). Membranes were blocked for nonspecific binding with 5% nonfat dry milk (Carnation Co., Los Angeles, CA) (w/v) in TBS containing 0.05% Tween. Generation and inactivation of the factor Va heavy chain was followed by incubating the blocked membranes for 1 h with monoclonal antibody ␣HFV-17. This antibody recognizes an epitope on the heavy chain of factor V between residues 307 and 506 and reacts with the same fragments of the factor V heavy chain upon cleavage by APC recognized by monoclonal antibody ␣HFV-6, which was described previously (39). Generation of the factor Va light chain was followed by immunoblotting using monoclonal antibody ␣HFV-9, which recognizes an epitope on the factor Va light chain (40). The reactive bands were visualized using peroxidase-conjugated goat anti-mouse IgG at a 1:5000 dilution, Southern Biotechnology Associates, Inc. (Birmingham, AL) using the chemiluminescent substrate Luminol (Renaissance chemiluminescent reagent, DuPont NEN). Light emitted from the hydrolysis of the added Luminol substrate was exposed to Kodak X-Omat film. Films were developed in a Kodak X-Omat and scanned using a Microscan 1000 scanning densitometer (TRI, Inc). Products of radiolabeled factor VIII formed in the reaction were analyzed by SDS-PAGE under reducing conditions as described previously (25).
In some experiments the membranes that were used for immunoblotting factor V products were washed with 0.5 M NaCl, 0.1 M glycine, pH 2.7, to remove bound antibodies and probed for prothrombin products using a polyclonal burro anti-prethrombin-1 (Grettle) antibody as described (25).

Effect of the Protein C Pathway on Factor VIIa⅐TF-initiated
Thrombin Generation-The effects of protein C and thrombomodulin were studied on factor VIIa⅐TF (1.25 pM)-initiated thrombin generation as described under ''Materials and Methods.'' At this concentration of initiating complex, thrombin generation occurs explosively after approximately 60 s in the absence of inhibitors ( Fig. 1, filled circles). Neither protein C at its physiological concentration (65 nM) (Fig. 1, open squares) nor thrombomodulin (10 nM) (open triangles) influenced thrombin generation when added separately. The absence of an effect of thrombomodulin alone on thrombin generation indicates that the soluble thrombomodulin used at concentrations up to 10 nM does not interfere with the activation of factor V or factor VIII by thrombin, critical steps during the initiation phase of the reaction (25). Analyses of thrombin generation in the presence of 65 nM protein C and varying concentrations of soluble thrombomodulin are displayed in Fig. 1. The initiation phase of the reaction does not appear to be influenced in the presence of protein C at thrombomodulin concentrations up to 10 nM ( Fig.  1, inverted triangles). No effect on the rate of thrombin generation during the propagation phase is observed in the presence of protein C (65 nM) at thrombomodulin concentrations Յ1 nM ( Fig. 1, filled squares). However, at 2.5 nM thrombomodulin the rate of thrombin generation is significantly decreased during the propagation phase as compared with the control (Fig. 1, stars). At 5 and 10 nM thrombomodulin ( Fig. 1, diamonds and inverted triangles, respectively), the rate of thrombin generation declines after approximately 100 nM thrombin is formed, resulting in an extended period of relatively low prothrombinase activity. Subsequently, thrombin generation increases again. These delayed propagation phases, however, ultimately result in the quantitative activation of prothrombin after 20 min. The delayed increase in thrombin generation is most probably the result of the accumulation of factor Xa activated by the uninhibited factor VIIa⅐TF complex and the continuously generated factor IXa⅐factor VIIIa complex. The large amount of factor Xa that accumulates probably produces activity by prothrombinase complex formation with traces of intact or partially degraded factor Va. Thrombin is generated at a rate of 100 nM/min in the delayed propagation phase of the reaction, which, based on a k cat for prothrombinase of 5016 min Ϫ1 (36), indicates the presence of approximately 20 pM prothrombinase.
Effect of the Protein C Pathway in the Presence of TFPI-The influence of the components of the protein C pathway were studied under conditions in which factor VIIa⅐TF activity is only expressed transiently, i.e. in the presence of TFPI (Fig. 2). Thrombin generation initiated by factor VIIa⅐TF (1.25 pM) in the presence of 2.5 nM TFPI (Fig. 2, filled circles) occurs after an extended initiation phase (2.5 min) followed by a propagation phase in which prothrombin is quantitatively converted to thrombin with a rate 30% that of the control reaction without TFPI present (open circles). Studies with increasing concentrations of thrombomodulin from 2.5 to 10 nM in the presence of protein C (65 nM) and TFPI (2.5 nM) reveal that the protein C pathway does not affect the initiation phase of the reaction even in the presence of TFPI ( Fig. 2 inset, inverted filled triangles). This observation is consistent with the notion that the protein C pathway functions only subsequent to the generation of thrombin and, thus, acts as a dynamic feedback inhibition mechanism.
In the presence of 2.5 nM TFPI, inhibition of thrombin generation during the propagation phase is observed at a thrombomodulin concentration of 0.1 nM (Fig. 2, filled triangles); thus the lowest concentration of thrombomodulin observed to inhibit thrombin generation by the factor VIIa⅐TF pathway is an order of magnitude lower when TFPI is present (see the reaction with 2.5 nM thrombomodulin in the absence of TFPI, Fig. 1, stars). At 1 nM thrombomodulin and 2.5 nM TFPI, thrombin generation occurs with an approximately 50% decreased rate after the initiation phase (Fig. 2, filled squares) compared with the controls (in the absence or presence of protein C) without thrombomodulin. After 50 nM thrombin is formed, the thrombin generation rate declines and elimination of prothrombinase activity occurs after 6 min, when ϳ100 nM thrombin has been generated (Fig. 2, filled squares). Thus, under these conditions, the combined actions of the protein C pathway (1 nM thrombomodulin) and TFPI (2.5 nM) quench the reaction during the propagation phase. The lack of an effect of 1 nM thrombomodulin and protein C in the absence of TFPI (Fig. 1), together with the inability of TFPI to prevent quantitative prothrombin consumption (26), and the observation that the combined actions of 1 nM thrombomodulin, protein C, and TFPI are able to eliminate prothrombinase activity lead to the conclusion that the protein C pathway and TFPI act in synergy under conditions present in the model. Increasing the thrombomodulin concentration (2.5, 5 and 10 nM) results in further decreases in the rate of thrombin generation and lower maximal levels of thrombin formed (Fig. 2, stars (2.5 nM), diamonds (5 nM), and inverted filled triangles (10 nM)). These data demonstrate that the combined actions of the protein C pathway and TFPI in the absence of other inhibitors can quench tissue factor-initiated thrombin generation.
APC Generation in the Reconstituted Model-The generation of APC (open symbols) measured in reactions initiated by 1.25 pM factor VIIa⅐TF in the presence of 2.5 nM TFPI is displayed along with the observed thrombin generation (filled symbols) in Fig. 3. In the presence of protein C and the absence of thrombomodulin, detectable APC generation (0.2 nM) (open circles) is observed at 5.5 min by which time 350 nM thrombin had been generated (filled circles). At the point of nearly quantitative prothrombin activation (ϳ10 min) approximately 1.5 nM APC has been generated. The APC generation rate in the absence of thrombomodulin is insufficient to affect the thrombin generation rate (compare thrombin generation in Fig. 2, in the control reaction (filled circles), and the reaction with protein C alone (open squares)). In the presence of 1 nM thrombomodulin, APC activity (open squares) is observed at the onset of thrombin generation (1.5 min, filled squares) in the reaction, resulting in a maximal formation rate of ϳ0.7 nM APC/min. This rate of APC generation is ϳ4-fold lower than that expected, based upon the k cat of 370 min Ϫ1 and K M of 7.6 M reported for the soluble thrombomodulin-thrombin complex (41). The thrombin concentration in the reaction increases from 33 to 113 nM in the reaction from 3 to 6 min without a further significant increase in the rate of APC generation, suggesting that the thrombomodulin is saturated by the thrombin formed early in the reaction. This saturation appears to be complete when thrombin concentrations are 30 nM or greater. In separate protein C activation experiments using purified ␣-thrombin, a derived K d app of ϳ10 nM was observed for the thrombomodulin-thrombin interaction with the soluble thrombomodulin used in the experiments. Together the results suggest a somewhat lower than expected k cat for the complex with the thrombomodulin used in our experiments.
The generation of 1 nM APC (Fig. 3, open squares) in the presence of 1 nM thrombomodulin at 3 min in the reaction results in a decreased rate of thrombin generation (filled squares). In the presence of 10 nM thrombomodulin, APC generation proceeds from 30 s in the reaction with a rate of 3.1 nM  ç) thrombomodulin. An uninhibited control reaction is shown in the absence of TFPI or protein C pathway components (E). Thrombin generation is effectively shut down in the presence of TFPI and the protein C pathway at thrombomodulin concentrations Ն1 nM after an initial amount of thrombin is formed. The inset shows the lack of a major effect of the protein C pathway on the initiation phase of the reaction in the presence of TFPI. APC/min (Fig. 3, open diamonds). This results in the accumulation of ϳ9 nM APC after 3 min. The accumulation of 9 nM APC occurs prior to the end of the initiation phase in reactions without thrombomodulin. This generation of APC in the presence of 10 nM thrombomodulin terminates thrombin generation in the initiation phase (Fig. 3, filled diamonds).
Effect of the Protein C Pathway on Factor V Products-Factor V circulates as a single chain procofactor (M r ϭ 330,000), which is activated by limited proteolysis by a variety of enzymes. Thrombin and factor Xa are the most likely activators of factor V under physiological conditions. The fully active product consists of a heavy and a light chain. Cleavage of factor V at Arg-709 results in generation of the heavy chain (residues 1-709, M r ϭ 105,000). Cleavage at Arg-1018 gives rise to a light chain precursor (residues 1019 -2196). Cleavage at Arg-1545 results in the generation of the light chain (residues 1546 -2196, M r ϭ 74,000). Factor V and its products formed during reactions initiated by 1.25 pM factor VIIa⅐TF in the presence of 2.5 nM TFPI were followed by immunoblotting. Fig. 4 displays the generation of the heavy chain, using antibody ␣HFV-17, and Fig. 5 displays the generation of the factor Va light chain by immunoblotting with monoclonal antibody ␣HFV-9.
In the control reaction in the absence of the protein C pathway components, factor V is cleaved to produce the heavy chain (M r ϭ 105,000) during the initiation phase of the reaction (Fig.  4A). In the absence of protein C the heavy chain (HC) is a stable product throughout the propagation phase (Fig. 4A). The light chain (LC), is generated at a slower rate (Fig. 5A) than the HC and is a relatively stable product in the reaction. In the presence of 65 nM protein C, but in the absence of thrombomodulin, cleavage of the factor Va HC is observed (Fig. 4B). However, this occurs after most of the prothrombin has been activated to thrombin (8 min). The inactivation process is signaled by the appearance of products that result from cleavage of the HC at Arg-506, indicated by the M r ϭ 75,000 band (Fig. 4B, 6 min), and subsequent cleavage of this product at Arg-306, giving rise to a M r ϭ 30,000 immunoreactive band (Fig. 4B, 12 min) consisting of residues 307-506 (42,43) of the factor Va heavy chain.
The lifetime of the intact factor Va heavy chain is shortened in the presence of protein C and increased concentrations of thrombomodulin. At 1 nM thrombomodulin (Fig. 4C) degradation of factor Va is observed by cleavage at Arg-506 (Fig. 4C, 2.5 min) immediately after the initiation phase when thrombin generation starts to accelerate (Fig. 2, filled squares). In an identical experiment approximately 0.6 nM APC was measured in the reaction at this point (Fig. 3, open squares). Arg-306 cleavage following cleavage at Arg-506 is observed at 3.5 min. Complete attenuation of thrombin generation coincides with complete degradation of intact heavy chain at ϳ6 min (compare Fig. 2, filled squares, and Fig. 4C). Quantitative cleavage of the 1-506 M r ϭ 75,000 HC fragment of factor Va at Arg-306 is observed at ϳ8 min (Fig. 4C) at which time ϳ4 nM APC is measured in an identical experiment (Fig. 3, open squares). Evaluation of the factor Va LC in the presence of 1 nM thrombomodulin (Fig. 5C) showed a decreased rate of LC generation, probably the result of the lower concentration of thrombin available for factor V activation. However, the factor Va LC is observed as a stable product in the reaction, consistent with the lack of a cleavage site for APC in the human factor Va light chain. In the presence of 10 nM thrombomodulin, cleavage of the factor Va heavy chain at Arg-506 is observed within the 1st min (Fig. 4D). Complete degradation of the heavy chain is observed after 4 min (Fig. 4D), at which point ϳ10 nM APC has been generated in an identical reaction (Fig. 3, open diamonds). A product is observed at 4 min in the M r ϭ 54,000 -60,000 region of the gel which indicates some initial cleavage of the heavy chain at Arg-306 (43) at the high APC concentration present in the reaction at that time. However, even at these high thrombomodulin concentrations the factor Va HC inactivation occurs predominantly by initial cleavage at Arg-506. The rate of generation of the factor Va light chain is further slowed in the presence of 10 nM thrombomodulin (Fig. 5D).
The quantitative relationships between thrombin generation and the appearance/disappearance of the heavy and light chains of factor Va are displayed in Fig. 6. Thrombin generation is indicated by the filled circles and the relative factor Va HC and LC concentrations by the open squares and open circles, respectively. In the control reaction (Fig. 6A) the concentration of fully activated factor Va and prothrombin activation is limited by the slower generation of the factor Va LC (M r ϭ 74,000). The increase in light chain concentration at the start of the propagation phase coincides with the increasing rate of thrombin generation (Fig. 6A, filled circles), suggesting that thrombin generation during the initiation phase is, in part, limited by the concentration of factor Va LC in the presence of an excess of HC. Thrombin formed in the reaction as a result of the factor Xa generated feeds back to activate factor V. Hence, the reac-tion is initially limited by the amount of factor Xa and Va. The maximum concentration of factor Va, judged from the appearance of both heavy and light chains, is reached at approximately 4 min in the reaction. The factor V activation profile in the reaction in the presence of protein C without thrombomodulin (Fig. 6B) is similar to that of the control reaction; however, after reaching a maximum factor Va concentration at 4 min, the HC concentration progressively decreases. Quantitative degradation of the HC is observed after 10 min.
In the reaction with 1 nM thrombomodulin and protein C (Fig. 6C), the maximum cofactor concentration observed is decreased compared with the control reactions with and without protein C (Fig. 6, A and B). In the early phase of the reaction the factor Va concentration is limited by the appearance of the LC and in the later phase by the HC disappearance (Fig. 6C), the latter being complete at 6 min. A maximum rate of thrombin generation is observed at 3 min in the reaction, when the HC is present at approximately 40% and the light chain at 20% of their respective maximum concentrations, when compared with the maximum densities observed in the control reaction (Fig. 6A). This maximum level of factor Va, as judged from the relative concentration of HC and LC in the reaction, coincides with the maximal observed rate of thrombin generation. With initially 20 nM factor V present, the 20% level of LC would correspond to a maximum concentration of approximately 4 nM factor Va. This conclusion assumes that the LC is generated in the same molecule that possesses an intact HC. However, if the intact heavy and light chains are randomly distributed, the resulting maximal concentration of fully active factor Va would be approximately 1.6 nM. Altogether these data suggest that the greatest cofactor activity existent in the reaction would range from 1.6 to 4 nM. Since only ϳ0.15 nM factor Xa (presented later in Fig. 9) is present in the reaction at this point, it might be concluded that factor Va is not the limiting component for the kinetic expression of prothrombinase activity at this point. However, in the dynamics of the reaction, the factor Va concentration becomes a limiting component for prothrombinase activity. The neutralization of prothrombinase activity during the reaction (Fig. 6C, filled circles) occurs at 6 min coincidental with quantitative degradation of intact heavy chain (Fig. 6C, open squares).
At 10 nM thrombomodulin (Fig. 6D, filled circles) prothrombinase formation is further limited by an earlier onset of HC degradation and a further decrease in the rate of LC generation compared with the reaction in the presence of 1 nM thrombomodulin (Fig. 6D, open circles and open squares). The increased rate of disappearance of the HC and decreased rate of light chain formation are such that the coexistence of the heavy and light chain during the reaction is nearly prevented at 10 nM thrombomodulin. At a maximum 10% of the light chain and 15% of the heavy chain are present in the reaction at 2.5 min. Since 20 nM factor V was added to the reaction, the maximum level of factor Va reached could be maximally 2 nM. However, in the more likely case of a random HC-LC distribution only 10% of the intact heavy chains would be associated with a light chain, resulting in only 0.3 nM factor Va at the maximum. These data demonstrate how the slowed appearance of the light chain enables the protein C pathway to prevent active cofactor accumulation by cleavage of the HC before the LC is generated.
Analyses of prothrombin activation products reveal that the complete inhibition of thrombin generation in the presence of protein C and 1 or 10 nM thrombomodulin (Fig. 2, squares and diamonds) is associated with a lack of prothrombin consumption (results not shown). The continued presence of intact prothrombin demonstrates that the inhibitory effect of the protein C pathway on thrombin generation is due to inhibition of prothrombinase activity and not due to substrate depletion.
Effect of the Protein C Pathway on Reactions Initiated in the Presence of Factor Va-The effect of the protein C pathway on thrombin generation initiated in the presence of factor Va was . Control reactions without protein C or thrombomodulin (circles) and in the presence of protein C and 1 nM (squares) or 10 nM (triangles) thrombomodulin. Factor Va was prepared by activation of factor V (400 nM) for 5 min at 37°C with 10 nM thrombin in HBS/Ca 2ϩ . Following activation thrombin was inhibited with 30 nM hirudin.

FIG. 6. Correlation of thrombin generation with generation/inactivation of the factor Va heavy and light chains.
Thrombin generation (q) initiated by 1.25 pM factor VIIa⅐TF in the presence of 2.5 nM TFPI, in an experiment identical to that as described in Fig. 2, is displayed in correlation with the appearance and disappearance of the factor Va heavy chain (Ⅺ, arbitrary density units) and factor Va light chain (E, arbitrary density units), quantified by densitometric scanning of immunoblots. The immunoblots were first developed for the light chain as described in Fig. 5. The membranes were then stripped and developed for the heavy chain as described in Fig. 4. Control reaction (A), protein C alone (B), protein C ϩ 1 nM thrombomodulin (C), and protein C ϩ 10 nM thrombomodulin (D). evaluated to examine the effects when the heavy and light chains of the active cofactor are both present at the start of the reaction. This experiment was performed to test whether the profound inhibition of thrombin generation, observed in the presence of TFPI and 10 nM thrombomodulin (Fig. 6D), is caused by the prevention of the coexistence of the heavy and light chains under these reaction conditions. Reactions initiated by 1.25 pM factor VIIa⅐TF in the presence of TFPI and factor Va (Fig. 7) display a shorter initiation phase (filled symbols) when compared with reactions initiated in the presence of the procofactor (open symbols); however, the rate of thrombin generation during the propagation phase of the reaction is not increased by preactivation of factor V. In the presence of protein C and 1 nM thrombomodulin, the thrombin generation curves of the reactions with factor Va and factor V are similar after the onset of thrombin generation (Fig. 7, open  and filled squares, respectively), and preactivation of factor Va only leads to a 20% increase in the maximal level of thrombin generation. In contrast, in the presence of 10 nM thrombomodulin and protein C, a 15-fold higher thrombin concentration is generated when the cofactor is activated (Fig. 7, filled and open  triangles). These data show a dramatic loss of effectiveness of the protein C pathway in inhibiting thrombin generation when the reaction is started with the fully active cofactor. These data support the hypothesis that inhibition of the thrombin generation pathway at limiting factor Xa concentration is dependent upon both the slowed formation of the light chain coupled with inactivation of the heavy chain by APC.

Effect of the Protein C Pathway on Factor VIII Products-
Factor VIII consists of a light chain (M r ϭ 80,000) which is associated via a Ca 2ϩ -dependent interaction with a heavy chain that may vary in size (M r ϭ 110,000 -200,000) (10). Activation of factor VIII occurs by cleavage of the heavy chain species at Arg-740 to initially generate the A1-A2 fragment (M r ϭ 90,000) (10, 44) containing residues 1-740 of factor VIII (13). Cleavage at Arg-372 bisects the A1 (M r ϭ 50,000) and the A2 (M r ϭ 40,000) domains (10,16,44). Generation of the active species of the light chain occurs by cleavage at Arg-1689 resulting in the generation of a M r ϭ 74,000 light chain containing the carboxyl terminus of factor VIII. Active factor VIIIa is thus composed of the A2 domain noncovalently associated with the A1/light chain heterodimer. The A2 domain is associated with the A1/light chain dimer via a binding site contained within residues 337-372 of the A1 domain (45). Proteolytic inactivation of factor VIIIa by APC or factor Xa has been reported to occur by cleavage at Arg-336 in the A1 domain (13).
The factor VIII products generated during reactions (initiated with 1.25 pM factor VIIa⅐TF in the presence of 2.5 nM TFPI) with and without the components of the protein C pathway were followed using radiolabeled factor VIII (Fig. 8). Thrombin generation curves using radiolabeled factor VIII showed identical rates of thrombin generation under the different conditions when compared with reactions with unlabeled factor VIII (results not shown). In the control reaction factor VIII is activated by cleavage of the heavy chain species at Arg-740 after 1 min to generate the M r ϭ 90,000 intermediate (A1-A2 do-FIG. 8. Effect of the protein C pathway on factor VIII products. Factor VIII products formed in the reaction were followed using radiolabeled factor VIII and SDS-PAGE. The factor VIII/factor VIIIa components were visualized by autoradiography of the dried gels. Factor VIII products are displayed for reactions initiated with 1.25 pM factor VIIa⅐TF in the presence of 2.5 nM TFPI without protein C or thrombomodulin (A), in the presence of protein C (65 nM) and 10 nM thrombomodulin (B), and in the presence of protein C, 10 nM thrombomodulin, and 300 nM protein S (C). The diagram shows the domain structure of factor VIII and the involved cleavage sites. The star indicates the reported inactivating cleavage at Arg-336. The open arrow at Arg-562 indicates a suggested APC cleavage site in the A2 domain (16). The A2 domain, however, appears as a stable product in the reactions with the protein C pathway components. mains) (Fig. 8A, lane 3). Cleavage by thrombin at Arg-372 and generation of the M r ϭ 50,000 fragment (A1 domain) and the M r ϭ 40,000 fragment (A2 domain) becomes apparent at 1.5 min (Fig. 8A, lane 4). Cleavage of the light chain of factor VIII (M r ϭ 80,000) also occurs at 1.5 min to generate the active species (M r ϭ 74,000). Near quantitative activation of factor VIII is observed at 4.5 min when virtually all light chain is converted to the M r ϭ 74,000 species. The A1 domain is the reported target of proteolytic inactivation by APC, factor Xa, or thrombin by cleavage at Arg-336 (13). Cleavage at this site results in a peptide that migrates on SDS-PAGE at M r ϭ 43,000 just above the A2 domain (46). The A1 fragment seems relatively stable in the reaction, although some product is lost between 7 and 20 min. This may be the result of cleavage of the A1 domain by thrombin at Arg-226 (47) which results in low molecular weight fragments such as observed at the bottom of the gel. The M r ϭ 74,000 light chain and the A2 domain appear as stable products in the reaction (Fig. 8A). The density of the B domain decreases after 4.5 min accompanied by the appearance of a low molecular weight triplet at the bottom of the gel. Thus, the low molecular weight fragments seem to be derived from proteolysis of the B domain and the A1 domain by the high concentrations of thrombin generated in the reaction.
The generation of factor VIII products in reactions containing 10 nM thrombomodulin and protein C (Fig. 8B) showed a slower activation rate of factor VIII, associated with the lower rate of thrombin generation as a result of factor Va inactivation under these conditions. In contrast to the control reaction, no significant cleavage of the A1 domain is observed with time. A stable density is observed in the A1 domain region in the presence of the protein C pathway at this high thrombomodulin concentration (10 nM) (Fig. 8B). No new cleavages in the other products are observed in the presence of the protein C pathway compared with the control reaction (compare Fig. 8, A and B), an observation consistent with the lack of any cleavages by APC in factor VIIIa besides the A1 domain. Thus, the protein C pathway does not result in an accelerated inactivation of factor VIIIa by proteolysis. The reduced level of thrombin formed in the presence of the protein C pathway results in the persistence of the B domain compared with the control reaction (Fig. 8A); consequently, a decreased rate of appearance of the low molecular weight triplet is observed.
Protein S has been reported to facilitate inactivation of factor VIIIa by APC (14,18,46). However, in a reaction in the presence of 300 nM protein S and the protein C pathway (65 nM protein C, 10 nM thrombomodulin), the A1 domain appeared as a stable product up to 12 min in the reaction (Fig. 8C). Some density is lost in the A1 domain region between 12 and 20 min, accompanied by increasing density at M r ϭ 43,000 (indicated by 4* in Fig. 8C), indicating that cleavage at Arg-336 (13, 44 -46) occurs in the presence of protein S. However, quantitative conversion of the A1 domain is not observed in the presence of protein S over the 20-min period evaluated. No other cleavages are observed in the presence of protein S (compare Fig. 8, B and C). It is noteworthy to mention that the lack of inactivating cleavages of factor VIII during the experiments does not mean that a stable level of factor VIIIa activity is reached. Factor VIIIa at physiological concentrations (0.7 nM) loses most of its activity as a consequence of dissociation of the A2 domain from the A1/light chain heterodimer (16,44,48). These data demonstrate that in the present model the protein C pathway, even in the presence of protein S, does not result in the proteolytic inactivation of the factor VIIIa molecule in the model within the interval in which factor Va is completely inactivated by APC (4 min, Fig. 4D), and prothrombin activation is totally suppressed.

Effect of the Protein C Pathway on Factor Xa Generation-
The lack of proteolytic inactivation of factor VIIIa in the presence of the protein C pathway components suggested that factor VIIIa cofactor activity in the activation of factor X by the factor IXa⅐factor VIIIa complex is not altered by APC proteolysis. Under the experimental conditions (low factor VIIa⅐TF in the presence of TFPI), thrombin generation during the propagation phase is virtually dependent upon factor IXa⅐factor VIIIa activity (26). The effect of the protein C pathway (10 nM thrombomodulin) on factor Xa generation in reactions initiated by 1.25 pM factor VIIa⅐TF in the absence or presence of 2.5 nM TFPI is displayed in Fig. 9. Factor Xa generation in the control without inhibitors is observed after 30 s in the reaction, increases to a rate of 0.2 nM/min after 30 s, and is linear up to 8 min in the reaction (Fig. 9, filled circles). Addition of protein C and 10 nM thrombomodulin did not influence factor Xa generation at any stage in the reaction (Fig. 9, filled squares). In the presence of 2.5 nM TFPI, factor Xa generation is observed at 1.5 min and proceeds with a maximal rate of 0.062 nM/min (Fig. 9, open circles). In the presence of TFPI the factor Xa generation rate is not further affected by protein C and 10 nM thrombomodulin (Fig. 9, open squares). Thus the lack of factor VIIIa cleavage by APC (Fig. 8B) is consistent with the lack of an effect of the protein C pathway upon factor Xa generation. Overall these data lead to the conclusion that the effects of the protein C pathway on thrombin generation in this model are solely explained by the inactivation of factor Va in the reaction and not by inactivation of factor VIIIa cofactor activity.
Correlation of the Prothrombinase Components in a Reaction Containing 1 nM Thrombomodulin-A representative correlation of the prothrombinase concentration, with the concentrations of the prothrombinase components, factor Xa and factor Va, for a reaction inhibited by 1 nM thrombomodulin and protein C in the presence of TFPI, is displayed in Fig. 10. The prothrombinase concentration was calculated from the change in thrombin generation (filled circles). The fully active factor Va concentration (open squares) was deduced from the concentration of the limiting component, the light chain or heavy chain. As can be seen in Fig. 10 stantial excess of factor Va until 6 min. At the peak of prothrombinase activity at 4 min when 65 pM prothrombinase is formed, the measured factor Xa concentration is 130 pM. The assembly of factor Xa and factor Va may occur almost instantaneously under the conditions used (excess negatively charged phospholipid (49)). When entered in the equilibrium relation for prothrombinase, the concentrations measured at 4 min in the reaction of factor Va (3.3 nM), factor Xa (130 pM), and prothrombinase (65 pM) suggest a K d app of ϳ3.2 nM for prothrombinase. This observed K d app for prothrombinase as estimated in the complex reaction is in reasonable agreement with the measured K d of ϳ1 nM for prothrombinase (49). It should be mentioned that the factor Xa generation in the reaction measured by the chromogenic assay may give an overestimation of the free concentration of factor Xa because of some dissociation of the formed factor Xa⅐TFPI complexes in the presence of the competing substrate. Furthermore, a number of different factor V species are present at 4 min in the reaction, and it is uncertain at the present time whether they bind factor Xa or not. It is thus possible that some factor Xa is bound to factor V species that do bind factor Xa but lack or have reduced cofactor activity in prothrombinase. The drop in factor Va concentration after 4 min correlates with declining prothrombinase activity, and this activity ceases when factor Va is depleted from the reaction after 6 min.
Effect of the Protein C Pathway at Varying Factor VIIa⅐TF Concentrations-An experiment in which thrombin generation was evaluated at varying factor VIIa⅐TF concentrations in the presence of TFPI (2.5 nM) and protein C (65 nM) with or without 10 nM thrombomodulin is displayed in Fig. 11. In the presence of TFPI and protein C, thrombin generation is observed over a range of 0.05-100 pM factor VIIa⅐TF, showing prolonged initiation phases and decreased rates of thrombin generation during the propagation phase with decreasing factor VIIa⅐TF concentrations (Fig. 11A). The rate of thrombin generation in the propagation phase decreases 20-fold over a range of 100 -0.05 pM factor VIIa⅐TF. At 0.05 pM factor VIIa⅐TF thrombin generation occurs after a lag phase of 10 min in the presence of TFPI and protein C (Fig. 11A, inverted triangles). Despite this long initiation phase explosive thrombin generation still occurs, indicating the lack of any significant protein C activation by traces of thrombin in the absence of thrombomodulin.
When the reaction is initiated in the presence of 10 nM thrombomodulin, protein C, and TFPI, a completely different profile is observed (Fig. 11B). In the presence of 10 nM thrombomodulin/protein C/TFPI (Fig. 11B, diamonds) factor VIIa⅐TF concentrations Յ5 pM fail to provide explosive thrombin generation. Initiation with 25 pM factor VIIa⅐TF in the presence of 10 nM thrombomodulin results in the sudden attenuation of prothrombinase activity after the explosive generation of 200 nM thrombin (Fig. 11B, squares). In contrast, at 100 pM, factor VIIa⅐TF overcomes the threshold raised by the protein C pathway components and TFPI, and the result is the quantitative activation of prothrombin (Fig. 11B, triangles). The presence of 10 nM thrombomodulin prevents explosive thrombin generation at low initiating factor VIIa⅐TF concentrations by stopping the formation of active factor Va (Fig. 6D). The increased rates of formation of catalysts in the presence of high concentrations of initiator overcome the inhibitory potential of the protein C pathway. The rapid procoagulant response at high initiator concentrations does not allow the protein C pathway time to suppress the thrombin generation reaction. Of interest is the threshold for explosive thrombin generation in the presence of 10 nM thrombomodulin (ϳ10 pM factor VIIa⅐TF) is similar to the threshold for explosive thrombin generation in the presence of the combination of physiological concentrations of TFPI and AT-III (26). It appears that the threshold concentration of initiating factor VIIa⅐TF required to overcome the inhibition of explosive thrombin generation will be dependent upon the thrombomodulin concentration present at the beginning of the reaction.

DISCUSSION
The effects of the protein C pathway on thrombin, factor Xa, factor Va, factor VIIIa, and APC generation initiated by the extrinsic (factor VIIa⅐TF) pathway of coagulation were studied in reconstituted reactions. The data demonstrate the following. 1) The isolated protein C pathway (thrombomodulin and protein C) reduces the rate of thrombin generation initiated by pM concentrations of factor VIIa⅐TF, without influencing the duration of the initiation phase of thrombin generation. 2) The protein C pathway and TFPI inhibit thrombin generation in a synergistic manner; the combination of these inhibitors has the potential to neutralize prothrombinase activity initiated by the tissue factor pathway.
3) The effects of the protein C pathway on thrombin generation are solely correlated with the proteolytic inactivation of factor Va by APC. 4) The generation of the factor Va heavy chain is completed during the initiation phase of the reaction. The onset and rate of the propagation phase of thrombin generation coincides with cleavage of factor V at Arg-1545 to generate the light chain. In the presence of protein C and up to 10 nM thrombomodulin, no intermediates of factor V inactivation of a molecular weight of M r ϭ 280,000 are observed (39), demonstrating that there is no significant inactivation of intact factor V by cleavage at Arg-306. 5) The inhibition of tissue factor-initiated prothrombinase activity by the protein C pathway is correlated with cleavage of the factor Va heavy chain. The cleavage of the heavy chain occurred by initial cleavage at Arg-506 followed by cleavage at Arg-306 (Fig. 4) over a range of 0 -10 nM thrombomodulin in the present model. Thus, the mechanism of inactivation under dynamic conditions, in the presence of the other coagulation factors and in situ generated APC, is similar to that observed in a system using the isolated reaction (42,43).
These conclusions are based on results of experiments performed at saturating phospholipid concentrations, and therefore the results do not exclude other effects of the protein C pathway components on platelets or at lower phospholipid concentrations.
Although the protein C anticoagulant pathway theoretically has the potential to shut down factor VIIa⅐TF-triggered thrombin generation in the absence of other inhibitors, this is not observed in the reconstituted model. The rate of factor Xa generation is unaffected by the presence of the protein C pathway, and factor VIIIa cleavage by APC does not occur within a period relevant to the propagation phase of thrombin generation. However, nearly quantitative cleavage of the factor Va heavy chain by the protein C pathway (results not shown) in the absence of other inhibitors results in slower thrombin generation (Fig. 1). The relatively high concentrations of factor Xa, present as a consequence of uninhibited factor VIIa⅐TF and factor IXa⅐factor VIIIa, apparently saturate traces of active or partially active factor Va still present resulting in significant prothrombinase activity (20 pM).
When TFPI is the only inhibitor, the reaction proceeds through factor Xa generation via the factor IXa⅐factor VIIIa pathway after the factor VIIa⅐TF complex is inhibited by TFPI⅐factor Xa (26). The actions of TFPI and the protein C pathway are complementary, and these inhibitors behave in a synergistic manner. TFPI inactivates the factor VIIa⅐TF complex and allows the formation of only limited amounts of factor Xa and factor IXa. In the presence of TFPI and the protein C pathway the limited amount of factor Xa that escapes TFPI inhibition is deprived of its cofactor factor Va because of the inactivation of the latter by APC. The level of thrombin generated and the rate at which the initial thrombin is formed are governed by the concentration of initiating factor VIIa⅐TF and the concentration of thrombomodulin. Quenching of the thrombin generation reaction in the presence of 10 nM thrombomodulin and TFPI is observed for reactions initiated by factor VIIa⅐TF concentrations Յ20 pM.
In the present model proteolytic processing of factor VIIIa does not contribute to the neutralization process nor is factor Xa generation influenced by the protein C pathway constituents. Hence, the inhibitory effects of the protein C pathway do not correlate with either factor VIIIa cleavage or reduced factor Xa generation. Protein S enhanced factor VIIIa cleavage in the model; however, even in the presence of protein S and the protein C pathway factor VIIIa inactivation occurred only slowly in the model. These observations are consistent with previous data of our laboratory on the inactivation of factor VIIIa by APC in the presence of protein S (46). However, the present data do not predict a significant role for the synergistic cofactor effect of protein S and factor V on the inactivation of factor VIIIa by APC (50).
Based on these results we suggest the regulatory model depicted in Fig. 12. TFPI regulates the initiation phase of factor VIIa⅐TF-initiated thrombin generation by inhibition of factor VIIa⅐TF and factor Xa. The protein C pathway does not affect the initiation phase or the factor Xa generation by factor IXa⅐factor VIIIa activity required to drive the propagation phase in the presence of TFPI. The acute inhibition of the propagation phase of thrombin generation by the protein C pathway is solely dependent on the inactivation of factor Va, indicated by the negative feedback arrow from APC to factor Va. The termination of thrombin generation is dependent on the synergistic inhibitory action of TFPI and the protein C pathway.
The physiological concentration of thrombomodulin is de- FIG. 12. Schematic presentation of the combined action of the protein C pathway and TFPI in the different phases of the thrombin generation reaction initiated by factor VIIa⅐TF. Note that, based on the lack of an inhibitory effect of the generated APC on factor Xa formation and on factor VIIIa activity, no inhibitory arrow is shown pointing from APC to the factor IXa⅐factor VIIIa complex in the initiation phase. The complete neutralization of prothrombinase activity by the protein C pathway in the presence of TFPI is indicated by the inhibition of factor Va by APC and by the inhibition of factor VIIa⅐TF and factor Xa by TFPI. This scheme depicts the acute action of the mechanisms involved in the regulation of TF-initiated thrombin generation by the protein C pathway which is solely dependent on the rapid quantitative inactivation of factor Va by APC. Although the activation of the cofactors is an important step in the reaction, the activation of the cofactors is not indicated for the purpose of clarity.
fined by the endothelial cell surface/volume ratio. The concentration of thrombomodulin may exceed 10 nM in the microcirculation where the surface to volume ratio is high (51). Hence, the effects of thrombomodulin observed in the present model are potentially physiologically relevant. Moreover, intact relipidated thrombomodulin is more efficient as a cofactor in the activation of protein C when compared with the soluble form (41). Furthermore, the soluble thrombomodulin used in the present experiments lacks chondroitin sulfate, a regulatory moiety on thrombomodulin involved in the association of thrombin, which may be present on thrombomodulin in vivo (52). Taken together it is anticipated that the observed effects of soluble thrombomodulin on APC generation in the present model may occur at lower concentrations of the natural form.
TFPI is present in different forms and states in the circulation (reviewed in Ref. 53). Several COOH-terminal truncated, less potent, forms of TFPI exist in plasma, and a major portion of TFPI in plasma is associated with lipoprotein particles, either by noncovalent interactions or by disulfide linkage to the apolipoprotein moieties in the lipoprotein particles. A large pool of full-length TFPI is presumably bound in vivo to endothelial cell exposed heparan sulfate (54 -57). Displacement of this TFPI pool from the endothelium by heparin treatment leads to increases in plasma TFPI levels up to 2-6-fold. Thrombomodulin is constitutively expressed as a transmembrane protein on the surface of vascular endothelial cells (4). Hence, our data provide a rationale for the apparent co-localization of the complementary anticoagulant functions of full-length TFPI and thrombomodulin at the blood/vessel wall interface.
Evaluation of clot formation and thrombin generation in a tissue factor-activated whole blood clotting model revealed that the fibrin clot is formed by the action of only 2.5 nM thrombin (58) and that most thrombin generated is formed after clot formation. Thus, the thrombin levels observed in the present model, even in the presence of high thrombomodulin concentrations, would result in significant fibrin formation. Fibrin clots may, however, be lysed by the fibrinolytic system. Fibrin clots can be stabilized against fibrinolytic attack by the action of factor XIII and the thrombin activable fibrinolysis inhibitor. Both factor XIII (59) and thrombin activable fibrinolysis inhibitor (60) are zymogens that are specifically activated by thrombin. Thus, the reduced amount of thrombin formed by the action of the protein C pathway will limit the concentration of thrombin available for the stabilization of fibrin clots. This may result in a clot predisposed to solubilization by the fibrinolytic system.