Complex-dependent inhibition of factor VIIa by antithrombin III and heparin.

The regulation of the factor VIIa-tissue factor complex is essential for control of the hemostatic response. However, the role of the inhibitor antithrombin III in the regulation of factor VIIa has remained in question. The inhibition of factor VIIa activity by antithrombin III and heparin in the presence and absence of tissue factor was evaluated using the fluorescent substrate m-LGR-nds. Our data show that the activity of recombinant human factor VIIa is inhibited by antithrombin III in the presence of heparin at a rate of 1.7 x 10(2) M-1 s-1. In the presence of tissue factor, the rate constant for this reaction increases to 5.6 x 10(3) M-1 s-1. A 1:1 stoichiometric complex between factor VIIa and antithrombin III, with an apparent molecular weight of 110,000, was detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A heterogeneous mixture of factor VIIa products with molecular weights between 50,000 and 80,000, most likely representing proteolytically degraded factor VIIa-antithrombin III complexes, was also observed.

The regulation of the factor VIIa-tissue factor complex is essential for control of the hemostatic response. However, the role of the inhibitor antithrombin I11 in the regulation of factor VIIa has remained in question. The inhibition of factor VIIa activity by antithrombin I11 and heparin in the presence and absence of tissue factor was evaluated using the fluorescent substrate rn-LGR-nds. Our data show that the activity of recombinant human factor VIIa is inhibited by antithrombin I11 in the presence of heparin at a rate of 1.7 X lo2 M" s-'. In the presence of tissue factor, the rate constant for this reaction increases to 5.6 X lo3 M-' s-'. A 1:l stoichiometric complex between factor VIIa and antithrombin 111, with an apparent molecular weight of 110,000, was detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A heterogeneous mixture of factor VIIa products with molecular weights between 50,000 and 80,000, most likely representing proteolytically degraded factor VIIa-antithrombin I11 complexes, was also observed.

~~ ~ ~ ~~~~~~~~~~~~~
The factor VIIa-tissue factor complex appears to play a major role in the initiation of the blood coagulation process (1)(2)(3). The expression of this complex requires the activation of the zymogen factor VI1 to the enzyme factor VIIa, the expression of tissue factor on a cell surface, and the presentation of the appropriate membrane environment. A fully formed complex efficiently activates the plasma zymogens factor X and factor IX to their respective enzyme forms, ultimately leading to the generation of a-thrombin.
Antithrombin I11 is the major inhibitor of coagulation proteases. However, although the inhibition of factor VIIa by antithrombin 111 has been reported, the role of antithrombin I11 has remained controversial (4-6). A study by Jesty (4) reported that neither factor VI1 nor factor VIIa in the presence or absence of tissue factor was inhibited by antithrombin 111. However, Broze and Majerus (5) and Kondo and Kisiel (6) reported that factor VIIa was inhibited slowly by anti- thrombin I11 in the presence of heparin. These authors concluded that this reaction proceeded so slowly that it was probably of little physiological significance (5). These authors also observed that tissue factor exhibited no protective effect in preventing antithrombin I11 inhibition of factor VIIa (5). Of the known plasma inhibitors of blood coagulation enzymes, tissue factor pathway inhibitor (TFPI)' has been the most well characterized inhibitor of the factor VIIa-tissue factor complex (7, 8). TFPI is a protease inhibitor that consists of three tandem Kunitz-type domains and may require the binding of factor Xa to TFPI before the inhibitor can interact with the factor VIIa-tissue factor complex (9, 10). In this setting, inhibition of the factor VIIa-tissue factor complex occurs only after factor Xa has been generated in the reaction, although the work of Callander et al. (11) has challenged the notion that factor Xa binding to TFPI is required prior to the binding of TFPI to the factor VIIa-tissue factor catalytic complex. These latter investigators reported that TFPI binds to and inhibits the factor VIIa-tissue factor complex directly with a dissociation constant of 11.9 nM in the absence of factor Xa, and a dissociation constant of 4.5 nM in the presence of factor Xa. These studies concluded that factor Xa is not an obligate requirement for the inhibition of the factor VIIa-tissue factor complex by TFPI but that factor Xa only enhances the overall binding affinity of TFPI for the catalytic complex.
The evaluation of the activity of factor VIIa and its inhibition has been hampered by the lack of sensitive and direct reporters of factor VIIa activity. This lack of a sensitive and direct method for quantitating factor VIIa activity has limited quantitative studies of the antithrombin I11 inhibition of factor VIIa. We have recently characterized and reported new synthetic substrates for factor VIIa enzymatic activity (12, 13). Our data demonstrated that the active site of factor VIIa is incompletely formed in the absence of Ca2+ and tissue factor. The inhibition of factor VIIa by antithrombin I11 and heparin was directly evaluated using this fluorescent substrate technology.
Proteins-Antithrombin I11 was purified from human plasma as previously described (15). Heparin sodium injection (USP) from beef lung was purchased from Organon, Inc. Recombinant human coagulation factor VIIa was purchased from Novo Pharmaceuticals. Recombinant human tissue factor was provided as a gift from Dr. Shu-Len Liu, Hyland Division, Baxter Healthcare Corp.
Factor VZZa Assays-Factor VIIa assays were conducted in 20 mM Hepes, 150 mM NaCl, 5 mM CaC12, pH 7.4 (HBS) at 37 "C. The synthesized m-LGR-nds was dissolved in dimethyl sulfoxide to a stock concentration of 10 mM. This solution was diluted in HBS to final working concentrations prior to all assays. The final concentration of reagents used in the presence of tissue factor were 50 nM recombinant factor VIIa, 50 nM tissue factor, and 100 p M PCPs. Antithrombin I11 (0.3-3 pM) and heparin (10 units/ml) were added to a mixture of factor VIIa, tissue factor, and PCPs. Subsamples (30 pl) of the reaction mixture were added to 100 p~ m-LGR-nds in a final volume of 1.3 ml. Rates of substrate hydrolysis were measured and product fluorescence monitored as previously reported for this substrate (12). The initial rate of factor VIIa substrate hydrolysis was evaluated, and the rates of the factor VIIa inhibition process for each concentration of antithrombin 111 were monitored over time following the addition of antithrombin I11 and heparin to the reaction mixture. Experiments evaluating the inhibition of factor VIIa (800 nM) by antithrombin I11 (1.6 p M ) and heparin in the absence of tissue factor were done in a manner similar to that described above. Data were reduced and rate constants estimated by the method of Downing et al. (19).
Zmmunoblot Analysis of Factor VZIa-The formation of a factor VIIa-antithrombin I11 complex was evaluated by SDS-PAGE. Antithrombin I11 (2.5 p~) and heparin (10 units/ml) were added either to factor VIIa (50 nM) or to factor VIIa preincubated with tissue factor for 15 min at 37 "C. The final concentration of tissue factor was 60 nM. Experiments were conducted in HBS with a 200-p1 final volume. After mixing the reaction components for 5 s, a 100-p1 reaction sample was removed from the mixture and rapidly quenched with an equal volume of solution containing 50 mM EDTA, 4% SDS (w/v), and 20% glycerol (v/v). The remaining reaction mixture was kept at 37 "C for 30 min and then added to the quenching solution described above. Reaction samples were then analyzed by SDS-PAGE chromatography using 8-12% polyacrylamide gels as generally described (20). Following SDS-PAGE, total protein from each gel was transferred to nitrocellulose membranes (Trans-Blot nitrocellulose membranes, Bio-Rad) for immunoblot analysis using general techniques described by Towbin et al. (21) and Lawson et al. (12).
Rate Comparisons-The inhibition of factor VIIa-tissue factor by TFPI and antithrombin-111 were compared as shown in Equations 1 and 2.

2)
u equals the rate of product formation (inhibitor-enzyme complex), k is the second-order rate constant for the reaction, and [A corresponds to the concentration of serine protease inhibitor. The relative rates for the inhibition processes can be calculated at constant factor VIIa, as shown by Equation 3.

RESULTS
The Inhibition of Factor VIIa by Antithrombin 111 in the Presence and Absence of Tissue Factor-The inhibition of factor VIIa activity by antithrombin I11 and heparin in the presence and absence of tissue factor was evaluated using the fluorescent substrate m -L G R -d s . The initial rate of factor VIIa substrate hydrolysis was evaluated, and the rates of factor VIIa inhibition for each concentration of antithrombin I11 were monitored over time following the addition of antithrombin I11 and heparin to the reaction mixture.  factor VIIa was observed. When tissue factor was added to factor VIIa, the antithrombin I11 inhibition rate in the absence of heparin proceeded at a more significant rate than was observed for antithrombin 111-heparin in the absence of tissue factor. When both tissue factor and heparin were added, a still higher rate of inhibition of factor VIIa activity was seen. These data demonstrate that the rate of factor VIIa inhibition is influenced by the concentration of antithrombin I11 added to the reaction, the presence of tissue factor in the experimental system, and the presence of heparin in the reaction system. These data were analyzed using the methods described by Downing et al. (19). This procedure yields linear plots with slopes representing ki (Fig. 1B). Table I presents the calculated second-order rate con-

Factor VIIa
Inhibition by Antithrombin 111 769 stant(s) for factor VIIa inhibition by antithrombin I11 in the presence and absence of tissue factor and heparin. These data illustrate that factor VIIa is inhibited by antithrombin IIIheparin. Furthermore, under the experimental conditions used, the addition of tissue factor enhanced the rate of factor VIIa inhibition 33-fold over that observed in the presence of heparin alone. These data provide quantitative evidence that the formation of the factor VIIa-tissue factor complex increases the susceptibility of factor VIIa for inhibition by antithrombin 111. The enhanced rate of factor VIIa inhibition, in the factor VIIa-tissue factor complex, is unique when compared with other procoagulant enzyme complexes. In other procoagulant complexes the proteases are protected from antithrombin I11 inhibition when complexed to their respective cofactors (22). The evaluation provided here can only be considered an initial venture into establishing the appropriate kinetic parameters to adequately describe the factor VIIa-tissue factor reaction with antithrombin III-heparin. The interactions of other coagulation enzyme complexes with heparin-antithrombin I11 are complex functions with concentration-dependent interaction for all species. The rate constants reported here are appropriate only for the specific conditions described.
The nature of the factor VIIa-antithrombin I11 inhibitor complex was evaluated by subjecting various reaction mixtures to SDS-PAGE/immunoblot analysis in the presence and absence of tissue factor and heparin. Fig. 2 demonstrates that a number of detergent stable species with mass greater than factor VIIa are observed when factor VIIa is treated with antithrombin I11 and heparin in the presence of tissue factor (lanes 6 and 8). The largest species ( a ) has an apparent molecular weight of 110,000, a value consistent with the predicted molecular weight of a 1:l stoichiometric complex between factor VIIa and antithrombin 111. A heterogeneous mixture of factor VIIa-related products with molecular weights between 50,000 and 80,000 are also observed ( 6 ) . These products most likely represent proteolytically degraded factor VIIa-antithrombin I11 complexes and free factor VIIa not covalently linked to antithrombin I11 (c). From kinetic studies of substrate hydrolysis (Fig. lA), under native condi- tions, all of the factor VIIa is functionally inactive under this set of conditions represented by the mixture in lanes 6 and 8. Thus, the inhibition process probably involves the formation of an inhibited non-covalent factor VIIa-antithrombin I11 complex followed by the subsequent formation of a covalent complex between the enzyme and inhibitor.

DISCUSSION
In this report we have reevaluated the inhibition of factor VIIa by antithrombin I11 and heparin in the presence and absence of tissue factor. We have previously demonstrated that factor VIIa undergoes an active site transition upon binding of both calcium and tissue factor, which enhances the fundamental catalytic properties of the enzyme (12). Data presented in this report illustrate that the interaction of the protease factor VIIa with the cofactor also enhances the susceptibility of factor VIIa for inhibition by antithrombin 111 at least 33-fold.
The highest rate constant observed in this report for the inhibition of factor VIIa-tissue factor by antithrombin IIIheparin, 5.6 X lo3 M-', s-', is small relative to that reported for the inhibition of a-thrombin by antithrombin 111-heparin (-1.5-4 x lo7 M-' s-'1 (23-25). The value observed for the factor VIIa-tissue factor reaction is in the range of that observed for factor XIa (0.6-1.5 X lo4 M-' s-I) (26, 27) and that reported for factor XIIa (4 X 10' M-' s-') (28). The rates of inhibition of a-thrombin and factor Xa by antithrombin I11 are significantly depressed by formation of complexes with their respective cofactors, thrombomodulin and factor Va (29, 30). In contrast, the rate of inhibition of factor VIIa by antithrombin I11 is positively influenced by complexation with tissue factor and further enhanced by the addition of heparin. Thus, in contrast to other coagulation serine proteases, circulating factor VIIa is relatively resistant to antithrombin IIIheparin inhibition, whereas the complex of factor VIIa-tissue factor is 33-fold more susceptible to antithrombin I11 inhibition.
The biological significance of the antithrombin 111-heparin inhibition of factor VIIa-tissue factor is not known. In the last several years, a number of investigators have concentrated on the influence of the tissue factor pathway inhibitor (TFPI) on the regulation of factor VIIa-tissue factor activity (3,8,11). However, the data reported here, when evaluated in respect to the relative concentrations of TFPI and antithrombin I11 in plasma, suggest that the role of antithrombin I11 as a factor VIIa-tissue factor inhibitor requires further exploration. The concentration of antithrombin I11 in plasma, 3 VM (22) is more than 1000 times that reported for the concentration for TFPI in normal adults (2.4 nM) (31). To our knowledge, no investigator studying the TFPI inhibition process has yet published data identifying the rate constant for inhibition of the factor VIIa-tissue factor complex by this inhibitor. Broze and colleagues (8) reported values for factor Xa inhibition by TFPI which would indicate that the secondorder rate constant for this reaction is 9 X lo5 "' s-'. If one assumes a similar rate constant for TFPI inhibition of factor VIIa, using the blood concentrations of TFPI and antithrombin 111, one would conclude that the relative velocity of inhibition of antithrombin 111-heparin toward factor VIIatissue factor would exceed that for TFPI by a factor of 7.4. Even if one assumes that the TFPI inhibition of the factor VIIa-factor Xa-tissue factor complex may occur at a diffusionally limited rate (approximately lo' M-' s-'), it would still be concluded that the relative efficiency of antithrombin IIIheparin inhibition of factor VIIa-tissue factor would be approximately equivalent to the efficiency of TFPI pathway in regulating this important reaction of blood clotting. For the diffusionally limited reaction, the relative efficiency of antithrombin 111-heparin to TFPI for inhibition would be approximately 0.7.