Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Thrombin and trypsin are two potent vasoactive peptides of the serine-protease family that activate protease-activated receptors (PARs) on vascular smooth muscle and endothelium to regulate vascular tone (Muramatsu et al., 1992; Godin et al., 1995; Komuro et al., 1997). Of the four protease-activated receptors cloned to date, thrombin has higher affinity for PAR-1 and PAR-3 than PAR-2 or PAR-4, whereas trypsin has the highest affinity for PAR-2 (Dery et al., 1998). The specific contribution of these receptors in inducing a contraction and/or relaxation to these proteases is not clear at this time because of the lack of highly potent and selective receptor antagonists and agonists. Novel protease-activated receptors, yet to be identified, may also exist and serve to modulate serine-protease-induced changes in blood vessel contractility (Hollenberg, 1999). However, the predominant evidence suggests that PAR-2 mediates vascular relaxation and PAR-1 mediates contractile responses to trypsin and thrombin (Hollenberg et al., 1996; Hwa et al., 1996). Thrombin- and trypsin-induced relaxation is endothelium and nitric oxide dependent (Muramatsu et al., 1992) and endothelium may serve to modulate contractile responses to the proteases (Sakiyama et al., 1991; Komuro et al., 1997). PAR-2 is localized on the vascular endothelium (Hwa et al., 1996), whereas PAR-1 is thought to be located on vascular smooth muscle (Muramatsu et al., 1992). Comparative vascular studies with thrombin and trypsin will assist in our understanding of the role of these PARs in the vascular effects of thrombin and trypsin.

Another "receptor" for thrombin is thrombomodulin, an integral membrane protein that binds thrombin and alters the specificity of thrombin from activation of protease-activated receptors and other procoagulant activities to activation of protein C (Esmon, 2000), forming a potent anticoagulant serine-protease. Functional thrombomodulin in vascular endothelium and plasma may be altered under pathological conditions. Thrombomodulin, expressed on the vascular endothelium (Bombeli et al., 1997) and platelets (Dittman, 1991), is down-regulated during inflammatory shock, endotoxicity, or after exposure to cytokines (Moore et al., 1987; Conway and Rosenberg, 1988; Archipoff et al., 1991). In other pathological states, such as diabetes, disseminated intravascular coagulation, respiratory stress, and pulmonary embolism, plasma concentrations of thrombomodulin increase dramatically (Asakura et al., 1991; Yamada et al., 1995; Boffa and Karmochkine, 1998). Thus, plasma thrombomodulin levels may be altered by either its administration or by pathological conditions. The effects that these changes in thrombomodulin may exert on vascular actions of serine-proteases such as thrombin are unknown.

The physical interaction of thrombin and thrombomodulin has been well studied. Thrombomodulin binds via epidermal growth factor-like domains to the highly electropositive fibrinogen recognition exosite 1 of thrombin (Sadler, 1997). This interaction of thrombomodulin and thrombin occurs with high affinity (K_d = 0.5 nM) (Dittman, 1991). Furthermore, PAR-1 and PAR-3 contain negatively charged residues immediately downstream of the cleavage site, also thought to interact with exosite 1 of thrombin (Vu et al., 1991; Dery et al., 1998). Thus, thrombin is thought to interact with thrombomodulin, PAR-1, and PAR-3 via the exosite 1 domain on thrombin, raising the possibility that increases/decreases in plasma thrombomodulin can have profound effects on thrombin-induced responses mediated via these receptors. The possibility that thrombomodulin may alter vascular contractile effects of thrombin is further supported by studies documenting the ability of thrombomodulin to alter certain nonvascular functions of thrombin, such as cell activation and cell death (Parkinson et al., 1993; Grinnell and Berg, 1996; Sarker et al., 1999).

To understand the effects of plasma thrombomodulin on thrombin-induced vascular actions and to support further the role of PAR-1 and PAR-2 in thrombin's vascular effects, we studied the effect of soluble thrombomodulin on thrombin-induced vasoconstriction and vasorelaxation. The current study was designed to explore the degree and directionality (potentiation/attenuation) of thrombomodulin-induced modulation of thrombin's vascular effects. The endothelium-denuded rabbit aorta and the endothelium intact rat aorta were used as established models for thrombin-induced contraction (Godin et al., 1995; Komuro et al., 1997) and relaxation (Muramatsu et al., 1992), respectively. To examine the specificity of the observed effects of soluble thrombomodulin and to shed light on the PARs involved in vascular responses, we also compared thrombin with trypsin with regard to vascular relaxation and contraction and then examined the effect of soluble thrombomodulin on trypsin-induced aortic contraction and relaxation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue Preparation. Thoracic aortae were dissected from male New Zealand White rabbits (2-3 kg) (Myrtle Rabbitry, Thompson Station, TN; Harlan Sprague-Dawley, Indianapolis, IN) and male Sprague-Dawley rats (0.25-0.35 kg) (Harlan Sprague-Dawley). Rats were sacrificed by cervical dislocation and rabbits were euthanized by i.v. injection of a lethal dose of sodium pentobarbital (65-100 mg/kg) into the ear vein according to animal use protocols approved by the Lilly Animal Care and Use Committee. The thoracic aorta was dissected free of surrounding tissue in modified Krebs' buffer (4.6 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 118.2 mM NaCl, 10.0 mM glucose, 1.6 mM CaCl₂·2H₂O, 24.8 mM NaHCO₃) and cut into 3- to 5-mm rings. For endothelium denudation, rabbit aortic rings were rotated 10 times on fine-point serrated forceps. Tissues were then placed between two stainless steel hooks and mounted in 10-ml organ baths filled with buffer solution. Baths were maintained at 37°C and bubbled with a 95:5% O₂:CO₂ mixture (pH 7.4). Tissues were equilibrated for 1 h and optimum passive force was produced by successively increasing the initial force to 6 g in each tissue with intermittent tissue washes.

Experimental Procedure. For every experiment, tissue viability was confirmed by a KCl (67 mM) challenge. Presence or absence of endothelium was confirmed by adding carbamylcholine (10⁶ M) to tissues precontracted to steady state with norepinephrine (10⁷ M). Serine-proteases (thrombin or trypsin) were added to the tissues in a noncumulative manner because of the rapid development of tachyphylaxis (Sakiyama et al., 1991).

Contractile Responses. Contractile responses were examined in endothelial-denuded rabbit aorta, a vascular model previously used to explore protease contractile activity (Sakiyama et al., 1991; Godin et al., 1995; Komuro et al., 1997). For each concentration of protease, noncumulative contraction was measured at steady state and was expressed as percentage of the maximal force to KCl (67 mM) generated initially in each tissue. Contraction to only one concentration of thrombin or trypsin was generated in each tissue.

Relaxant Responses. Relaxant responses were examined in endothelial-intact rat aorta, a vascular model used to explore protease relaxant activity (Muramatsu et al., 1992). Tissues were contracted with norepinephrine (10⁷ M) to a steady state followed by a single concentration of relaxant. Maximal relaxation for each concentration of protease was expressed as the percentage decrease in norepinephrine (10⁷ M)-induced force. Relaxation to only one concentration of thrombin or trypsin was generated in each tissue.

Effect of Thrombomodulin. In most experiments thrombin or trypsin was incubated with thrombomodulin or vehicle (control responses) for 30 min at 35°C before tissue exposure. In other experiments, thrombomodulin was incubated with the tissue for 30 min at 37°C before thrombin challenge.

Data Acquisition and Analysis. For all experiments, isometric force was measured with Sensotec transducers coupled to MP100 data acquisition software (BIOPAC Systems, Inc., Santa Barbara, CA). Data were analyzed off-line and expressed as mean ± S.E. Data represent aortic responses from the number of animals indicated with the number of tissues shown in parentheses. Statistical comparisons were performed with Student's t test using SigmaStat software. Differences between mean values were considered statistically significant when P < .05.

7

Proteins and Chemicals. Norepinephrine, carbamylcholine, and porcine trypsin were obtained from Sigma (St. Louis, MO). Human -thrombin was purchased from Enzyme Research (South Bend, IN). Recombinant soluble human thrombomodulin (sTM;CS+) (referred to as thrombomodulin/soluble thrombomodulin in the text) was synthesized in the Lilly Research Laboratories (Brian W. Grinnell, Research Technologies & Proteins, Eli Lilly & Company).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Thrombin and Trypsin-Induced Aortic Contraction and Relaxation. Both thrombin and trypsin relaxed rat aorta (Fig. 1, top) and contracted rabbit aorta (Fig. 1, bottom) in a concentration-dependent manner. Trypsin (EC₃₀ = 3 × 10¹⁰ M) was 10-fold more potent than thrombin (EC₃₀ = 3 × 10⁹ M) as a relaxant peptide. In contrast, trypsin (EC₃₀ = 6 × 10⁷ M) was approximately as potent a vasoconstrictor as thrombin (EC₃₀ = 2.3 × 10⁸ M). Interestingly, trypsin and thrombin were both more potent relaxant agonists in rat aorta than contractile agonists in rabbit aorta and again quantitative differences were apparent between trypsin and thrombin. Trypsin was 2000-fold more potent as a relaxant agonist than as a vasoconstrictor, whereas thrombin was only 8-fold more potent as a relaxant than contractile agonist.

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Trypsin- and thrombin-induced vasorelaxation (top) in endothelium-intact rat aorta and vasoconstriction (bottom) in endothelium-denuded rabbit aorta. Relaxant effects were represented as percentage of norepinephrine (107 M)-induced force (1.8 ± 0.1 g), whereas contractile effects were represented as percentage of KCl (67 mM)-induced maximal contraction (4.8 ± 0.3 g). The status of the endothelium was confirmed by carbamylcholine (106 M)-induced relaxation of norepinephrine (107 M) contracted tissues (insets). Points are mean values and vertical lines represent the standard error of the mean. The number of animals and number of rings (in parentheses) are indicated for each data point. * denotes P < .05.

9

8

7

6

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Time course for the mean effect of equieffective concentrations of thrombin and trypsin to induce vasorelaxation (top) and vasoconstriction (bottom). Relaxant response was represented as percentage of norepinephrine (107 M)-induced force (1.7 ± 0.1 g) and contractile response as percentage of KCl (67 mM)-induced maximal contraction (6.2 ± 0.8 g). Points are mean values and vertical lines represent the standard error of the mean. The number of animals and number of rings are indicated for each graph. Note the different time scales for each graph.

Vascular Effects of Thrombomodulin. Thrombomodulin, an endothelial transmembrane protein capable of binding thrombin, was studied for its ability to affect vasomotility. As shown in Fig. 3, thrombomodulin (10⁷ M) was ineffective either as a relaxant or as a contractile peptide in tissues that markedly relaxed to carbamylcholine (10⁶ M) or contracted to KCl (67 mM). Thus, effects of thrombomodulin to alter protease-mediated vascular responses would not result from any direct interaction of thrombomodulin with vascular smooth muscle.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Effect of thrombomodulin to induce vascular relaxation compared with carbamylcholine (106 M)-induced relaxation and vascular contraction relative to KCl (67 mM)-induced contraction. Columns are mean values and vertical lines represent the standard error of the mean of tissues from three to four animals.

Effect of Thrombomodulin on Thrombin-Induced Aortic Contraction and Relaxation. To study the effect of thrombomodulin on serine-protease-induced vascular contraction or relaxation, thrombomodulin was incubated with the proteases for 30 min (under Materials and Methods) to facilitate the interaction of thrombomodulin with thrombin before tissue exposure. Thrombomodulin (10⁸ M) did not significantly attenuate thrombin (10⁸ M)-induced relaxation (Fig. 4). However, a higher concentration of thrombomodulin (10⁷ M) significantly (P < .05) attenuated thrombin (10⁸ M)-induced vasorelaxation by 50.7 ± 5.7% (Fig. 4). In contrast, thrombomodulin was a more potent inhibitor of thrombin (10⁸ M)-induced vasoconstriction (Fig. 4) because both 10⁸ M and 10⁷ M thrombomodulin significantly (P < .05) inhibited thrombin-induced contraction by 51.9 ± 3.9 and 79.2 ± 4.2%, respectively. Thus, thrombomodulin (10⁸ M) significantly attenuated thrombin-induced vasoconstriction without any effect on thrombin-induced relaxation.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effect of thrombomodulin (108 and 107 M) on thrombin (108 M)-induced relaxation and contraction. , control; , 108 M thrombomodulin; , 107 M thrombomodulin. Endothelial integrity was determined by carbamylcholine (106 M) challenge of norepinephrine (107 M)-contracted tissues (85.7 ± 1.6% for relaxant studies and 4.9 ± 2.1% for contractile studies). Columns are mean values and vertical lines represent the standard error of the mean. The number of animals and number of rings (in parentheses) are indicated for each data point. * denotes P < .05.

8

7

8

7

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Effect of thrombomodulin (108 and 107 M) on the contractile concentration response curve to thrombin. The Schild plot (inset) was graphed from the composite curves as log [dose ratio  1] versus log [thrombomodulin] with the slope equal to 0.98. Points represent mean values and vertical lines are the standard error of the mean for the number of animals and number of rings (in parentheses) indicated.

7

7

7

7

Effect of Thrombomodulin on Trypsin-Induced Aortic Contraction and Relaxation. The specificity of thrombomodulin's effect to inhibit thrombin-induced aortic contraction was studied by examining thrombomodulin-trypsin interactions (Fig. 6). Thrombomodulin (10⁸ and 10⁷ M) was incubated with trypsin for 30 min (under Materials and Methods) before tissue exposure. As anticipated, neither 10⁸ nor 10⁷ M thrombomodulin attenuated trypsin (10⁹ M)-induced relaxation or more importantly, trypsin (10⁷ M)-induced contraction. Thus, thrombomodulin specifically and preferentially attenuated thrombin-induced vasoconstriction.

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Effect of thrombomodulin (108 and 107 M) on trypsin (109 M)-induced relaxation and trypsin (107 M)-induced contraction. , control; , 108 M thrombomodulin; , 107 M thrombomodulin. Endothelial integrity was determined by carbamylcholine (106 M) challenge of norepinephrine (107 M)-contracted tissues (78.4 ± 2.7% for relaxant studies and 3.1 ± 0.9% for contractile studies). Columns are mean values and vertical lines represent the standard error of the mean for the number of animals and number of rings indicated in parentheses.

Effect of Thrombomodulin on the Rate of Contraction or Relaxation to Thrombin and Trypsin. Because 10⁸ M thrombomodulin significantly attenuated thrombin-induced contraction without a significant effect on relaxation, we questioned whether thrombomodulin (10⁸ M) might have altered the time to attain maximal relaxation even though it had no apparent effect on maximal responses. However, thrombomodulin (10⁸ M) had no effect (Fig. 7, top) on the rate of thrombin (10⁸ M)-induced maximal relaxation (t_1/2 = 21.9 ± 2.9 and 23.4 ± 3.4 s in the presence and absence of thrombomodulin, respectively). Likewise, thrombomodulin (10⁸ M) did not significantly alter the rate of contraction to thrombin (t_1/2 = 188.1 ± 40.1 and 227.5 ± 21.5 s in the presence and in the absence of thrombomodulin, respectively) (Fig. 7, bottom) in spite of the reduction in maximal contraction to thrombin. Interestingly, thrombomodulin (10⁸ M) did attenuate the prolongation of relaxation to thrombin, consistent with the possibility of dual components to the relaxant response, a fast and a slower phase. Nevertheless, thrombomodulin (10⁸ M) did not alter the time to reach maximal responses (relaxation/contraction) to thrombin even though the maximal contractile effect was attenuated significantly by thrombomodulin (10⁸ M).

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Effect of thrombomodulin on the rate of thrombin-induced relaxation and contraction. Relaxant effects of thrombin (108 M) in the presence of thrombomodulin (108 M) were expressed as percentage loss of norepinephrine (107 M)-induced force (top). Contractile effects of thrombin (108 M) in the presence of thrombomodulin (108 M) were expressed as percentage of KCl (67 mM)-induced force (bottom). Points are mean values and vertical lines represent the standard error of the mean for the number of animals and number of rings indicated in parentheses.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Both thrombin and trypsin are serine-proteases known to possess PAR-dependent vascular effects (see the Introduction). The vascular effects of thrombin and trypsin were similar in that both proteases were more potent relaxant than contractile agonists (Fig. 1), consistent with the observations of other investigators (Muramatsu et al., 1992; Hwa et al., 1996; Komuro et al., 1997). Furthermore, the protease activity responsible for activation of contractile receptors showed similar kinetics between thrombin and trypsin and contraction was a slower response than relaxation (Fig. 2). It is possible that the release of endothelial factors may play a role in generating a faster relaxant than contractile response or that the relaxant second messenger system is more rapidly activated. The differences in the time to achieve maximal relaxant and maximal contractile responses could also be due to the cleavage site amino acid sequence differences between the relaxant and contractile PARs such that enzymatic cleavage of the relaxant receptor occurs more rapidly.

On the other hand, several marked differences between the vascular responses of thrombin and trypsin occurred. Although trypsin was more potent than thrombin as a relaxant agent, the reverse was true for contraction (Fig. 1). For relaxation, trypsin was 10-fold more potent than thrombin, whereas thrombin was more potent than trypsin as a contractile agonist. These quantitative differences support the contention that separate receptors (PARs) mediate relaxation and contraction to serine-proteases consistent with trypsin-activated PAR-2 and thrombin-activated PAR-1 being the predominant relaxant and contractile receptors, respectively (Hollenberg et al., 1996). In this regard, it is worthwhile to note that although serine-proteases such as thrombin and trypsin modulate vasomotility, the serine-protease activated protein C did not affect vascular motility (Bhattacharya et al., 2000), and the protease factor Xa was a relaxant agonist with potency between that of thrombin and trypsin (Schaeffer et al., 1997).

Furthermore, although trypsin was more potent than thrombin as a relaxant agonist, thrombin produced a significantly more rapid relaxant response than trypsin (Fig. 2, top). The difference in kinetics between trypsin- and thrombin-induced activation of the relaxant receptor may suggest subtle differences in the ability of trypsin and thrombin to interact with the same relaxant receptor, with trypsin possessing slower hydrolytic activity. Alternatively, it is possible that these kinetic differences reflect thrombin and trypsin's ability to activate distinct G-proteins coupled to the same PAR-2 and thereby activate different signal transduction mechanisms, with thrombin activating a signaling mechanism functionally more rapid than that of trypsin. The difference in the relaxant time course could also arise from potential interactions of thrombin, but not trypsin, with endothelial thrombomodulin. Last, it is possible that thrombin- and/or trypsin-induced relaxation occurred via related but distinct PARs, possibly PAR-4 or other novel PARs yet to be identified (Hollenberg, 1999).

Having characterized the vascular effects of thrombin and trypsin, we next explored the role of thrombomodulin in modulating the vascular effects of the serine-proteases. Thrombomodulin is known to bind to thrombin's exosite 1 domain, a stretch of charged residues on thrombin instrumental in docking thrombin to PAR-1 and PAR-3, but not to PAR-2 (Dery et al., 1998). Therefore, thrombomodulin could theoretically function as a competitive inhibitor of thrombin-induced activation of PAR-1 and/or PAR-3, although our data (Fig. 5) may indicate otherwise. Alternatively, the thrombin-thrombomodulin complex may augment receptor proteolysis and thereby thrombin's vascular effects, just as the complex enhanced activated protein C-dependent proteolysis (Knobe et al., 1999). A third possibility is that thrombomodulin may have no effect on the thrombin-PAR interaction just as glycoprotein-Ib, a protein also interacting with exosite 1 of thrombin, did not compete with the thrombin receptor (PAR-1) for thrombin binding (Bouton et al., 1995). Hence, understanding the effect of thrombomodulin on thrombin-induced vascular responses is important to 1) our understanding of the possible role of thrombomodulin in vascular physiology, and 2) our knowledge of the PARs responsible for contraction and relaxation to thrombin.

Interestingly, thrombomodulin preferentially inhibited thrombin-induced contractile responses, possibly mediated via PAR-1 activation and required higher concentrations of thrombomodulin to inhibit thrombin-induced relaxation (Fig. 4). Based only on two concentrations of thrombomodulin with each concentration of thrombin examined in separate tissues coupled to an inability to obtain a saturable maximal response (limited by thrombin availability), we attempted to construct an estimate of a Schild analysis to evaluate the interaction of thrombomodulin with thrombin. Inhibition of thrombin-induced contraction by thrombomodulin appeared to be competitive because the Schild analysis (Arunlakshana and Schild, 1959) resulted in a slope estimate of 0.98 (Fig. 5, inset) that does not appear to differ from unity. Furthermore, the effect of thrombomodulin was a direct result of its interaction with thrombin rather than any interaction of thrombomodulin with tissue PARs (Table 1). In addition, thrombomodulin did not modulate vascular tone induced by trypsin, a peptide lacking exosite 1 (Fig. 6). Hirudin, a thrombin inhibitor known to bind to exosite 1 on thrombin (Rydel et al., 1990) and a competitive inhibitor of thrombin-thrombomodulin interaction (Tsiang et al., 1990), also did not alter trypsin-induced vascular tone (Muramatsu et al., 1992; Hwa et al., 1996). In line with this, hirudin was observed to inhibit the binding of thrombin to the freshly excised rabbit aorta (Hatton and Moar, 1991). These observations strengthen the notion that the exosite 1 domain of thrombin is not only important for its interactions with thrombomodulin and hirudin but also for its interaction with contractile PARs such as PAR-1.

Thrombomodulin did not alter either the early fast component of thrombin-induced relaxation, maximal thrombin-induced relaxation, or the rate of thrombin-induced contraction (Fig. 7). In this regard, it is to be noted that the delayed slower component of thrombin-induced relaxation appeared inhibited by thrombomodulin. Thus, in the presence of thrombomodulin, thrombin-induced relaxation was not sustained. This observation raises the possibility that thrombin-induced relaxation was mediated by two components. This possibility is consistent with the different rates of relaxation produced by trypsin and thrombin wherein trypsin may be activating only the slow component of relaxation.

The observation that thrombomodulin preferentially attenuated thrombin-induced maximal vasoconstriction is consistent with the ability of thrombomodulin to modulate other cell surface effects of thrombin (Grinnell and Berg, 1996) and may have important physiological and pathological significance. Thrombomodulin is established to modulate thrombin's effect on activated protein C and on the coagulation pathway (Esmon, 2000), although its effect on vascular actions of thrombin has not previously been studied. The circulating concentration of thrombomodulin approximates 0.2 to 1.0 nM (Ishii et al., 1990; Takano et al., 1990; Dittman, 1991; Boffa and Karmochkine, 1998) and is known to increase by 5- to 10-fold during vascular injury associated with sepsis, disseminated intravascular coagulation, preclampsia, coronary and atherosclerotic vascular disease, and even during cardiac surgery (Asakura et al., 1991; Yamada et al., 1995; Boffa and Karmochkine, 1998). Under such pathological conditions, vascular endothelial cell injury exists and thrombin-induced vascular contraction would predominate. Thus, the present study raises the possibility that increases in plasma thrombomodulin would antagonize thrombin-induced vasoconstriction and protect vascular tissue from the contractile effects of thrombin, especially apparent during vascular injury.

The present study may also suggest a significant physiological role of thrombomodulin in the microvasculature because the concentration of surface thrombomodulin in the microvasculature is about 500 nM, whereas that in large vasculature is less than 1 nM (Esmon, 1989). The higher concentration of thrombomodulin in the microcirculation is consistent with an important role for thrombomodulin to attenuate thrombin-induced microvascular contractility based on our observation that 10⁸ M thrombomodulin inhibited thrombin-induced aortic contraction.

Furthermore, thrombomodulin is enriched in the pulmonary microvasculature (Boffa and Karmochkine, 1998) and is reported to decrease in severe pulmonary hypertension (Cacoub et al., 1996). Our results raise the possibility that the low levels of plasma thrombomodulin may be contributing to the pathology of pulmonary hypertension. Loss of cell-surface thrombomodulin or decreases in plasma thrombomodulin might minimize any proposed vascular protective effect that thrombomodulin might exert. This in effect would leave thrombin-induced contraction unopposed, possibly leading to the precipitation of pulmonary hypertension. In fact, pulmonary hypertension has been associated with vascular smooth muscle proliferation, vasoconstriction, and thrombocytopenia, all cellular events related to increased thrombin activity (Rich and Brundage, 1989; Rostagno et al., 1991). Furthermore, although controversial (Hampl et al., 1993), hirudin, a thrombin inhibitor, was effective in reducing pulmonary hypertension in the pig (Hoffmann et al., 1990), supporting the contention that thrombin may play an important role in the vasoconstriction that occurs in pulmonary hypertension.

In summary, our results reinforce the concept that thrombin- and trypsin-induced relaxation and contraction are mediated predominantly via activation of PAR-2 and PAR-1, respectively. In addition, the fact that thrombomodulin preferentially inhibited thrombin-induced vascular contraction via interaction with the exosite 1 domain of thrombin strengthened the hypothesis that PAR-1 is the vascular contractile receptor, whereas PAR-2 is the vascular relaxant receptor. Last, thrombomodulin may be important in modulating vascular contractile effects of thrombin in addition to its established role in modulating the anticoagulant actions of thrombin. This novel action of soluble thrombomodulin to protect vascular tissue from thrombin-induced vasoconstriction may have important clinical ramifications in vascular pathology.