Molecular basis for the susceptibility of fibrin-bound thrombin to inactivation by heparin cofactor ii in the presence of dermatan sulfate but not heparin.

Although fibrin-bound thrombin is resistant to inactivation by heparin.antithrombin and heparin.heparin cofactor II complexes, indirect studies in plasma systems suggest that the dermatan sulfate.heparin cofactor II complex can inhibit fibrin-bound thrombin. Herein we demonstrate that fibrin monomer produces a 240-fold decrease in the heparin-catalyzed rate of thrombin inhibition by heparin cofactor II but reduces the dermatan sulfate-catalyzed rate only 3-fold. The protection of fibrin-bound thrombin from inhibition by heparin.heparin cofactor II reflects heparin-mediated bridging of thrombin to fibrin that results in the formation of a ternary heparin.thrombin.fibrin complex. This complex, formed as a result of three binary interactions (thrombin.fibrin, thrombin.heparin, and heparin.fibrin), limits accessibility of heparin-catalyzed inhibitors to thrombin and induces conformational changes at the active site of the enzyme. In contrast, dermatan sulfate binds to thrombin but does not bind to fibrin. Although a ternary dermatan sulfate. thrombin.fibrin complex forms, without dermatan sulfate-mediated bridging of thrombin to fibrin, only two binary interactions exist (thrombin.fibrin and thrombin. dermatan sulfate). Consequently, thrombin remains susceptible to inactivation by heparin cofactor II. This study explains why fibrin-bound thrombin is susceptible to inactivation by heparin cofactor II in the presence of dermatan sulfate but not heparin.

Heparin, a sulfated polysaccharide, acts as an anticoagulant by accelerating the inhibition of thrombin and factor Xa by antithrombin (1). Although heparin is widely used for the treatment of acute coronary ischemic syndromes, it has limitations in patients undergoing percutaneous coronary interventions (2) or when used as an adjunct to thrombolytic therapy (3)(4)(5). These limitations have been attributed to the inability of the antithrombin⅐heparin complex to inactivate clotting enzymes bound to components of the thrombus, particularly thrombin bound to fibrin (6,7).
Resistance of fibrin-bound thrombin to inactivation by the antithrombin⅐heparin complex reflects the incorporation of thrombin into a ternary heparin⅐thrombin⅐fibrin complex (8)(9)(10). To form this complex, heparin interacts with both exosite II on thrombin (11)(12)(13) and the D domain of fibrin (14), thereby bridging thrombin to fibrin via exosite II (8). This heightens exosite I-mediated binding of thrombin to fibrin and likely increases the overall affinity of thrombin for fibrin. The formation of the ternary complex, therefore, is a consequence of the presence of two exosites on thrombin, which independently bind fibrin and heparin. Recently, we demonstrated that protection requires ligation of both of thrombin's exosites within the ternary heparin⅐thrombin⅐fibrin complex, a process that impairs access of inhibitor-bound heparin to exosite II on thrombin (15). Thrombin within the ternary complex is protected from inactivation by the heparin⅐heparin cofactor II (HCII) 1 complex to a greater extent than the heparin⅐antithrombin complex (15), because access of the amino terminus of HCII to exosite I on thrombin, an obligatory part of the HCII inhibitory mechanism (16,17), is reduced. Allosteric changes in the active site of thrombin induced upon formation of the ternary heparin⅐thrombin⅐fibrin complex may also contribute to the protection of fibrin-bound thrombin from inactivation by heparin⅐antithrombin and heparin⅐HCII complexes by limiting inhibitor reactivity with fibrinbound thrombin (18).
Dermatan sulfate (DS), a sulfated glycosaminoglycan that has antithrombotic activity in laboratory animals (19,20) and in humans (21)(22)(23), acts as an anticoagulant by catalyzing only HCII. Because thrombin is the exclusive plasma target of HCII, DS is considered a selective inhibitor of thrombin (16). Although fibrin-bound thrombin is protected from inactivation by the heparin⅐HCII complex (15), indirect studies done in plasma systems suggest that fibrin-bound thrombin is susceptible to inactivation by the DS⅐HCII complex (24,25). The purpose of this study was to confirm these findings using purified reagents and to elucidate the mechanism by which the DS⅐HCII complex, but not the heparin⅐HCII complex, inactivates fibrinbound thrombin.
Preparation of Soluble Fibrin-Human fibrinogen, treated with gelatin-agarose to remove fibronectin, was used to prepare soluble fibrin (SF), as described previously (15). Briefly, fibrin clots were centrifuged and dialyzed versus water. The fibrin was then dissolved by dialysis versus 20 mM acetic acid, aliquoted, and stored at a concentration of about 100 M at Ϫ70°C. Polymerization of SF was blocked by addition of 5 mM GPRP-NH 2 (26), and the material was neutralized with 1 M Tris-HCl, pH 7.5, in a volume corresponding to 40% of the volume of SF, just prior to use. A molecular weight of 340,000 and ⑀ 280 1% value of 14.0 were used to calculate the soluble fibrin concentration.

Methods
Rates of Thrombin Inhibition by HCII in the Absence or Presence of SF, Glycosaminoglycans, or Both-The influence of varying concentrations of heparin or DS on the second-order rate constants (k 2 ) for inhibition of thrombin by HCII were determined under pseudo-firstorder conditions in the absence or presence of SF. Thrombin (10 nM) was incubated for 5 min at room temperature in TBS containing 0.6% PEG 8000 in the presence of various concentrations of heparin or DS (0 -11 M), SF (0 -4 M), and 10 mM GPRP-NH 2 . Reaction mixtures (10 l) were aliquoted to 96-well round bottom microtiter plates and an equal volume of HCII (in a concentration at least 10-fold higher than that of thrombin) was added to each well at time intervals ranging from 2 s to 5 min. All reactions were terminated concomitantly by the addition of 200 M chromogenic substrate (tGPR-pNA) in 200 l of TBS containing 10 mg/ml Polybrene. Residual thrombin activity was calculated by measuring absorbance at 405 nm for 5 min using a Spectra Max 340 Microplate Reader (Molecular Devices, Menlo Park, CA). The pseudofirst-order rate constants (k 1 ) for thrombin inhibition were determined by fitting the data to the equation k 1 ⅐t ϭ ln([P] o /[P] t ), where [P] o is initial thrombin activity and [P] t is thrombin activity at time t. The secondorder rate constant, k 2 , was then determined by dividing k 1 by the HCII concentration (27).
Effect of Heparin and DS on the Binding of 125 I-FPR-thrombin to FM-Sepharose-Thrombin blocked at its active site with FPRCK (FPRthrombin) was radiolabeled with Na 125 I as described previously (28). The binding of 125 I-FPR-thrombin to FM-Sepharose in the absence or presence of either heparin or DS (from 0 to 10 M) was studied in TBS containing 0.6% PEG 8000 and 0.01% Tween 20. FM-Sepharose (9 M) in 0.5 ml of buffer was incubated with 50 nM 125 I-FPR-thrombin by inverted mixing for 2 min. The tube was centrifuged in a microcentrifuge for 2 min at 15,000 ϫ g, and a 100-l aliquot of supernatant was removed for gamma counting. The aliquot was returned to the tube. This procedure was repeated following each addition of heparin or DS to the tube. The fraction of 125 I-FPR-thrombin bound to FM-Sepharose as a function of the total amount of 125 I-FPR-thrombin present was calculated.
Effect of Thrombin or ␥-Thrombin on the Binding of 125 I-Labeled DS to FM-Sepharose-DS (3 mg) was labeled with 1.0 mCi of Na 125 I (PerkinElmer Life Sciences, Markham, Ontario, Canada) in a volume of 300 l using the IODO-BEAD iodination reagent as described by the manufacturer (Pierce, Rockford, IL). Labeled DS catalyzed the inhibi-tion of thrombin by HCII to the same extent as its unlabeled counterpart, indicating that the labeling procedure did not influence DS activity. A mixture of 500 nM 125 I-labeled DS and 5 M FM-Sepharose was titrated with increasing concentrations of thrombin or ␥-thrombin (from 0 to 3 M). After centrifugation, residual 125 I-labeled DS in the supernatant was quantified by gamma counting and used to determine the fraction of 125 I-labeled DS bound to FM-Sepharose.

Effect of Ternary Thrombin⅐Fibrin⅐Heparin or Thrombin⅐Fibrin⅐DS Complex Formation on the Active Site Environment of Thrombin-
Thrombin labeled at the active site with anilinonaphthalene-6-sulfonic acid (ANS-FPR-thrombin) was prepared using ATA-FPR (N-(acetylthio)acetyl-D-Phe-Pro-Arg-CH 2 Cl) as described by the supplier (Molecular Innovations Inc., Royal Oak, MI). Briefly, 10.6 M thrombin was incubated for 30 min at 23°C with a 2.5-fold molar excess of ATA-FPR in 20 mM HEPES-NaOH, pH 7.0, 100 mM NaCl, 1 mM EDTA (HBSE buffer). The reaction mixture was dialyzed against HBSE buffer, incubated with a 10-fold molar excess of 2-(4-(iodoacetamide)anilino)naphthalene-6-sulfonic acid (Molecular Probes Inc.) in the presence of hydroxylamine for 60 min at 23°C in the dark and then redialyzed. 800 l of 140 nM ANS-FPR-thrombin in 50 mM HEPES, pH 7.5, 0.125 M NaCl, 1 mg/ml PEG 8000, 5 mM GPRP-NH 2 was added to a semi-micro quartz cuvette. Using a PerkinElmer Life Sciences LS50B luminescence spectrometer with excitation wavelength set to 334 nm and excitation and emission slit widths set to 5 nm, the fluorescence emission spectrum from 380 -580 nm of ANS-FPR-thrombin was monitored before and after the addition of 310 nM SF and/or 100 nM heparin or 3 M DS. Soluble fibrin was neutralized by the addition of 40% v/v 1 M Tris-HCl, pH 7.5, just prior to use. Addition of 1 M Tris-HCl, pH 7.5, had no effect on the fluorescence spectrum of ANS-FPR-thrombin (data not shown).

Determination of the Affinities of f-heparin, f-DS, and f-FPR-thrombin for Fibrin
Clots-Fluorescent derivatives of heparin and DS were prepared as follows (29). 10 mg of heparin or DS were incubated with 15 mg of FITC in 2.5 ml of 1.0 M Na 2 CO 3 , pH 9.0 for 5 h at 23°C. After centrifugation at 13,000 ϫ g for 5 min, 1 ml of the supernatant was applied to duplicate PD-10 gel filtration columns (Millipore Corp., Bedford, MA), equilibrated with H 2 O, and eluted with H 2 O under gravity. 0.5-ml fractions were collected, frozen, and lyophilized. Recovered material was pooled, weighed, and dissolved in TBS to a stock concentration of 10 mg/ml. FITC-labeled active site-blocked thrombin (f-FPRthrombin) was prepared as previously described (28). Briefly, thrombin was incubated with fluorescein-D-Phe-Pro-Arg chloromethyl ketone (f-FPRCK) until no residual thrombin chromogenic activity was detected. After dialysis to remove unincorporated f-FPRCK, the concentration of fluorescently labeled thrombin was determined by measuring absorbance using ⑀ 280 1% of 1.8 after correction for light scatter at 320 nm using the relationship A 280 corr ϭA 280 Ϫ 1.7 ϫ A 320 . The affinity of FITC-labeled ligands for fibrin was determined by measuring unbound ligand in supernatants of clots prepared by clotting varying concentrations of fibrinogen with 1 nM thrombin. Briefly, 200 l of 2 mM CaCl 2 , 100 nM FITC-labeled ligand, various concentrations of fibrinogen (30 -3000 nM), and 1 nM thrombin were mixed in a series of microcentrifuge tubes. After 1-h incubation at 23°C, fibrin was pelleted by centrifugation for 5 min at 15,000 ϫ g and 100 l of supernatant was removed and added to 300 l of TBS. The fluorescence intensity of the samples was monitored with excitation and emission wavelengths set to 492 and 522 nm, respectively, and excitation and emission slit widths both set to 15 nm. K d values were calculated by plotting I/I o versus fibrinogen concentration, where I o and I represent the fluorescence intensities in the absence and presence of the varying concentrations of fibrinogen, respectively. The parameters K d and ⌬I were calculated by nonlinear regression ("Tablecurve," Jandel Scientific, San Rafael, CA) using the equation, where ⌬I is the maximum fluorescence change, P is the initial concentration of f-heparin, f-DS, or f-FPR-thrombin, L is the total fibrinogen concentration, and a stoichiometry of 1 is assumed (28).
Determination of the Affinities of Thrombin for Heparin and DS-The affinities of heparin or DS for thrombin were determined by monitoring changes in intrinsic thrombin fluorescence when the enzyme was titrated with heparin or DS. The initial intensity reading of 100 nM thrombin in a 2-ml quartz cuvette (I o ) was determined with excitation and emission wavelengths set to 280 and 340 nm, respectively, and excitation and emission slit widths set to 6 nm. Aliquots of either heparin or DS were then added to the cuvette, and, after mixing, changes in fluorescence were monitored (I). K d values were calculated by plotting I/I o versus glycosaminoglycan concentration and the data were fit by nonlinear regression to the equation given above.
Determination of the Affinities of HCII for Heparin, DS, and SF-The association between HCII and either heparin, DS, or fibrin was monitored by the ligand-dependent fluorescence intensity change of anilinonaphthalene-6-sulfonic acid labeled HCII (ANS-HCII). ANS-HCII, prepared as previously described (30), was added to a concentration of 100 nM to a quartz cuvette. The initial fluorescence intensity (I o ) of ANS-HCII was determined at excitation and emission wavelengths set to 280 and 437 nm, respectively, and excitation and emission slit widths set to 10 nm, and an emission filter of 290 nm. Known quantities of either heparin, DS, or SF were then added to the cuvette and, after mixing, changes in fluorescence was monitored (I). K d values were calculated by plotting I/I o versus ligand concentration, and the data were fit by nonlinear regression to the equation given above.
Displacement of Thrombin from FM-Sepharose by HCII-500-l suspensions of FM-Sepharose (9 M FM) containing 50 nM 125 I-FPR-thrombin without glycosaminoglycan or with 2.5 M DS in TBS/0.6% PEG/ 0.01% Tween 20 were mixed for 2 min. The amount of unbound 125 I-FPR-thrombin in the suspension after each addition of an aliquot of 80 M HCII was determined as described above.

Comparison of the Effect of SF on Heparin-and DS-catalyzed
Rates of Thrombin Inhibition by HCII-To verify the protective effect of SF on thrombin inhibition by HCII, we determined the rates of thrombin inhibition in the absence or presence of SF and heparin. As shown in Fig. 1A, SF caused a dose-dependent decrease in the heparin-catalyzed rates of thrombin inhibition by HCII, which by two-way analysis of variance performed using Minitab (software version 11, State College, PA), was highly significant (p Ͻ 0.001). At 1 M heparin and 4 M SF, a maximal 240-fold decrease in the rate was observed, a value consistent with that reported previously (15). In contrast, at a concentration of 4 M SF, a significant (p Ͻ 0.001) but modest 3-fold decrease in the DS-catalyzed rates of thrombin inhibition by HCII was observed (Fig. 1B).
Quantification of Binary Interactions That Comprise Ternary Thrombin⅐Fibrin⅐Glycosaminoglycan Complexes-The assembly of the ternary heparin⅐thrombin⅐fibrin complex is postulated to occur through a series of binary interactions between thrombin⅐fibrin, thrombin⅐heparin, and fibrin⅐heparin (8,10,18). In this study, we determined the dissociation constants for these interactions, as well as those involving DS. The affinities of DS and heparin for thrombin were determined by monitoring changes in intrinsic protein fluorescence of thrombin when titrated with DS or heparin (not shown). DS and heparin bind saturably to thrombin with K d values of 2600 and 116 nM, respectively (Table I). The K d value for the thrombin⅐heparin interaction is in agreement with that determined by titration of ANS-labeled thrombin with increasing concentrations of heparin (K d ϭ 59 nM) (10).
Current evidence suggests that both heparin and DS bind to exosite II on thrombin (11)(12)(13)31). This was confirmed in two ways. First, competitive binding studies were performed where thrombin-bound f-heparin was displaced by DS (Fig. 2). In this experiment, addition of thrombin to f-heparin resulted in an approximate 3.5% decrease in fluorescence intensity. Subsequent titration of the sample with DS resulted in a dose-dependent increase of fluorescence intensity to the value observed for f-heparin prior to binding to thrombin (I/I o of 1). Likewise, in the reciprocal experiment, the fluorescence decrease that occurred upon addition of thrombin to f-DS was negated in a dose-dependent fashion by titration with heparin (not shown). In a second approach, we used thrombin variants with impaired exosites to identify the DS binding site (not shown). DS binds to ␥-thrombin, a proteolytic derivative of ␣-thrombin lacking exosite I (11), with a 2.6-fold lower affinity than ␣-thrombin (K d values of 6.8 M and 2.6 M, respectively).
In contrast, DS binds RA-thrombin, a thrombin variant with three point mutations in exosite II that lower its affinity for heparin 20-fold (32), with a 7-fold lower affinity (K d ϭ 17 M). These studies confirm that, like heparin, DS also binds to exosite II on thrombin, albeit with lower affinity (Table I).
The affinities of f-FPR-thrombin, f-heparin, and f-DS for fibrin were monitored by clotting varying concentrations of fibrinogen with a catalytic amount of thrombin in the presence of a fluorescently labeled ligand, and quantifying unbound ligand in the clot supernatant (not shown). As listed in Table I (8,10). In contrast, f-DS does not bind to fibrin. Both heparin and DS bind to ANS-HCII, but the affinity of heparin for HCII is 5-fold higher (13 and 71 M, respectively). Thus, DS binds to both HCII and thrombin with lower affinity than heparin. No binding of ANS-HCII to fibrin was detected ( Table I).
Effect of Heparin and DS on the Binding of 125 I-FPR-thrombin to Fibrin-It has been shown previously that heparin enhances the binding of thrombin to fibrin, an effect that occurs regardless of whether heparin has high or low affinity for antithrombin, but occurring only with heparin chains of 11,200 Da or more (8). In this study, we compared the ability of DS to promote thrombin binding to FM-Sepharose with that of heparin and low-molecular-weight heparin. As shown in Fig. 3, DS has little effect on 125 I-FPR-thrombin binding to FM-Sepharose, even at concentrations up to 10 M. In contrast, at concentrations up to 250 nM, heparin enhances 125 I-FPR-thrombin binding to FM-Sepharose in a dose-dependent manner. At heparin concentrations above 250 nM, 125 I-FPR-thrombin binding to fibrin monomer decreases, likely reflecting the accumulation of distinct heparin⅐fibrin and heparin⅐thrombin populations (8). When thrombin and FM-Sepharose were titrated with lowmolecular-weight heparin (enoxaparin), there was only a small increase in the amount of thrombin bound, comparable to that observed with DS. Similar results were obtained with fibrin clots in place of FM-Sepharose (not shown).
Incorporation of Thrombin into Ternary Glycosamino-glycan⅐Thrombin⅐Fibrin Complexes-Our binding studies indicate that thrombin can interact with both fibrin and DS, via exosites I and II, respectively. These findings suggest that, like heparin, DS can form a three-component complex with thrombin and fibrin, even though DS does not bind fibrin. To explore this possibility, a mixture of 500 nM 125 I-labeled DS and 5 M FM-Sepharose was titrated with increasing concentrations of thrombin or ␥-thrombin (Fig. 4). After centrifugation, residual 125 I-labeled DS in the supernatant was quantified and used to determine the fraction of 125 I-labeled DS bound to FM-Sepharose. In the absence of thrombin, minimal amounts of 125 Ilabeled DS bind to FM-Sepharose. With thrombin addition, 125 I-labeled DS binds to FM-Sepharose in a concentration-dependent fashion. In contrast, when ␥-thrombin, a proteolytic derivative of thrombin that lacks exosite 1, is substituted for thrombin, there is no increase in the amount of 125 I-labeled DS that binds to FM-Sepharose. These data indicate that a DS⅐thrombin⅐fibrin complex can form despite the lack of direct DS⅐fibrin interactions.

Influence of Ternary Thrombin⅐Fibrin⅐Glycosaminoglycan Complex Formation on the Conformation of the Active Site of
Thrombin-Incorporation of thrombin into a ternary heparin⅐thrombin⅐fibrin complex alters the rate of thrombinmediated hydrolysis of chromogenic substrates (9,10,15), suggesting that ternary complex formation induces conformational changes in the active site of thrombin that may limit its reactivity with macromolecular inhibitors (18). To determine whether thrombin in a DS⅐thrombin⅐fibrin complex also experiences conformational changes in its active site, the fluorescence of ANS-FPR-thrombin was monitored upon addition of DS and/or SF, as was described previously for heparin or SF (10). As illustrated in Fig. 5A, addition of 310 nM SF or 100 nM heparin to 140 nM ANS-FPR-thrombin produces a 19 and 48% increase in fluorescence, respectively, compared with the fluorescence of ANS-thrombin alone. When 100 nM heparin is added to the ANS-FPR-thrombin/SF mixture, however, there is an 86% increase in fluorescence. The observation that the addition of both SF and heparin produces a greater change in the emission spectrum of ANS-FPR-thrombin than the two components alone is in agreement with a previous report by Hogg et al. (10) and suggests that formation of the ternary heparin⅐ thrombin⅐fibrin complex induces conformational changes in the active site of thrombin beyond those produced by formation of binary thrombin⅐fibrin or thrombin⅐heparin complexes. In contrast to the results obtained with heparin, the addition of 3 M DS, a concentration above the K d value for the DS⅐thrombin interaction (Table I), to either ANS-FPR-thrombin alone or to the ANS-FPR-thrombin/SF mixture has minimal effect on the emission spectra (Fig. 5B). These findings suggest that, unlike heparin, DS does not induce conformational changes in the active site environment of thrombin in the absence or presence of fibrin.
Effect of Heparin and Low-molecular-weight Heparin on the Rate of Inhibition of Fibrin-bound Thrombin by the DS⅐HCII Complex-Although fibrin-bound thrombin is susceptible to inactivation by the DS⅐HCII complex (Fig. 1), we predicted that heparin, by virtue of its higher affinity for both fibrin and thrombin than DS, would impair DS activity by bridging thrombin to fibrin, thereby enhancing the fibrin⅐thrombin interaction. As a control, we used low-molecular-weight heparin, because only heparin chains of Ͼ11.2 kDa are of sufficient length to bridge thrombin to fibrin (8). In the presence of 4 M SF, heparin concentrations of 0.1 M or higher cause a decrease in the 3.3 M DS-catalyzed rate of thrombin inhibition by HCII (Fig. 6). In contrast, a low-molecular-weight heparin, enoxaparin, has no effect on the rate of inhibition of fibrin-bound thrombin by the DS⅐HCII complex. Neither heparin nor lowmolecular-weight heparin influenced the DS-catalyzed rate of thrombin inhibition in the absence of SF (data not shown).
Displacement of IIa from FM-Sepharose by HCII-Because catalysis of thrombin inhibition by HCII is not impaired by fibrin in the presence of DS, exosite I on thrombin must be accessible to the DS⅐HCII complex even though thrombin binds to fibrin via this site. This observation predicts that the DS⅐HCII complex is capable of displacing thrombin from fibrin by binding the amino-terminal tail of HCII to exosite I of thrombin. This was tested by monitoring the amount of 125 I-FPR-thrombin displaced from FM-Sepharose by increasing concentrations of HCII in the absence or presence of DS (Fig.  7). HCII alone, at concentrations up to three times physiological concentration, had limited capacity to displace thrombin from FM-Sepharose. However, in the presence of 2.5 M DS, dose-dependent displacement of thrombin by HCII was observed. Maximal displacement was achieved with physiological concentrations of HCII, with half-maximal effect at about 250 nM HCII. Thus, the amino terminus of HCII is able to compete effectively with fibrin for binding to thrombin exosite I.

DISCUSSION
In this study, we demonstrate that fibrin-bound thrombin is readily inhibited by the DS⅐HCII complex, but not by the heparin⅐HCII complex, even though heparin and DS demonstrate comparable catalysis of HCII in the absence of fibrin (Fig. 1). The resistance of fibrin-bound thrombin to inhibition by the heparin⅐HCII complex is a consequence of formation of a ternary heparin⅐thrombin⅐fibrin complex (15). This ternary complex forms as a result of mutual binary interactions between each of the three components, resulting in increased affinity of each of the individual interactions (10,18). Formation of this ternary complex protects thrombin from inhibition by heparin⅐HCII by reducing access of the amino terminus of HCII to exosite I on thrombin, impairing the ability of HCIIbound heparin to bridge HCII to exosite II on thrombin, and inducing allosteric changes in the active site of thrombin that may limit its reactivity with HCII (10,15).
Like heparin, DS binds to exosite II on thrombin, even when thrombin is fibrin-bound (Fig. 2). However, DS and heparin are distinguished by the fact that only heparin binds fibrin (Table  I). With high affinity for thrombin and fibrin, heparin augments exosite I-mediated binding of thrombin to fibrin, impairing the ability of HCII to bind and inhibit thrombin (Fig. 3). Because DS does not heighten the thrombin⅐fibrin interaction, the amino terminus of DS-activated HCII retains access to exosite I on thrombin. The failure of DS to induce a conformational change in the active site of thrombin (Fig. 5B) may also permit reactivity of HCII with fibrin-bound thrombin. These results confirm the hypothesis that all three binary interactions (thrombin⅐fibrin, thrombin⅐glycosaminoglycan, glyco-saminoglycan⅐fibrin) within the ternary complex are integral to convey protection of fibrin-bound thrombin from inhibition by heparin-catalyzed inhibitors (15).
Heparin compromises the DS-catalyzed rate of thrombin inhibition by HCII in the presence of fibrin (Fig. 6). This reduction in rate is consistent with the formation of a ternary thrombin⅐fibrin⅐heparin complex. Heparin can bridge thrombin to fibrin in the presence of DS, because the affinity of heparin for thrombin is higher than that of DS (Table I). In contrast, low-molecular-weight heparin (enoxaparin) does not reduce the DS-catalyzed rate of thrombin inhibition by HCII, because only heparin chains greater than 11.2 kDa are of sufficient length to bind simultaneously to thrombin and fibrin (8).
The DS⅐HCII complex displaces thrombin from FM-Sepharose with an IC 50 value of 0.17 M (Fig. 7). This value reflects the ability of the amino terminus of DS-activated HCII to bind to thrombin exosite I. A previous study determined that a 22-residue peptide, corresponding to amino acids 54 -75 of HCII, inhibited the clotting activity of thrombin with an IC 50 of 28 M (33). The lower IC 50 value with intact HCII suggests that there are additional interactions with thrombin provided by the remainder of the amino terminus or the body of the HCII molecule. A homologous peptide derived from hirudin residues 54 -66 inhibited thrombin clotting activity with an IC 50 of 0.8 M (33). This compares with the K i value of 0.6 M derived for the initial encounter of intact hirudin with thrombin (34). Thus, in addition to its ability to inhibit fibrin-bound thrombin, the DS⅐HCII complex displaces thrombin from fibrin allowing both HCII and antithrombin to inhibit thrombin. A similar displacement role for the hirudin 53-64 peptide has been proposed (35).
Our observation that fibrin-bound thrombin is susceptible to inactivation by the DS⅐HCII complex suggests that DS should be effective for the prevention and treatment of thrombosis. In a randomized trial comparing subcutaneous DS with low dose subcutaneous heparin for thromboprophylaxis in patients un-dergoing general surgical procedures, both agents had comparable efficacy (22). More recently, DS was shown to be more effective than heparin at preventing postoperative venous thrombosis in patients undergoing general surgical procedures (36). Studies in rabbits suggest that DS suppresses thrombus accretion to a greater extent than heparin when the two agents are given in doses that have the same inhibitory activity against thrombin in vitro (25).
DS produces only minimal inhibition of platelet and fibrin deposition in a high shear, arterial-type thrombosis model in baboons (37). The limited efficacy of DS in this setting may reflect the fact that DS only inhibits thrombin and has no effect on factor Xa-mediated thrombin generation, an important contributor to thrombus growth (38). Consequently, if selective thrombin inhibitors are used as monotherapy for treatment of arterial thrombosis, high concentrations may be necessary to prevent thrombus propagation. This concept is supported by our recent studies demonstrating that both DS and hirudin, agents that inhibit fibrin-bound thrombin, as well as free thrombin, are limited in their ability to block clotting in a thrombogenic extracorporeal circuit (30). In contrast, Vasoflux, a drug that not only inhibits fibrin-bound thrombin but also blocks factor Xa generation, prevents clotting in the circuit (30). These findings raise the possibility that adding heparin to DS may improve its efficacy, because heparin will block factor Xa-mediated thrombin generation. Our observation that unfractionated heparin compromises the rate of thrombin inhibition by the DS⅐HCII complex, whereas low-molecular-weight heparin does not (Fig. 6), suggests that low-molecular-weight heparin may be a better adjunct to DS than unfractionated heparin. The observation that DS and lowmolecular-weight heparin display additive anticoagulant activity in vitro provides further support for this approach (39).
In summary, our results indicate that fibrin-bound thrombin is readily inhibited by the DS⅐HCII complex because, unlike heparin, DS does not bridge thrombin to fibrin. Although DS and heparin can form thrombin⅐fibrin⅐glycosaminoglycan complexes, the additional heparin⅐fibrin interaction strengthens the ternary complex, thereby limiting accessibility of thrombin exosites and inducing conformational changes in the active site of thrombin. These findings reveal that resistance of thrombin to inhibition by serpins results from a glycosaminoglycan⅐fibrin interaction, as well as interactions between thrombin⅐fibrin and thrombin⅐glycosaminoglycan. These mechanistic insights also demonstrate the advantages of glycosaminoglycans whose activities are not compromised by fibrin.