Coagulation Factor Xa Induces Endothelium-Dependent Relaxations in Rat Aorta
Abstract The relaxing effect of coagulation factor Xa on phenylephrine-contracted rat aortic rings was compared with the effect of thrombin and trypsin. All three proteases induced a dose-dependent relaxation in the presence of an intact endothelium. EC50 values were 3±1, 24±9, and 16±1 nmol/L for thrombin, trypsin, and factor Xa, respectively. Whereas thrombin induced rapid relaxations followed by partial recontraction, trypsin and factor Xa induced slower sustained effects. Factor Xa–induced relaxations were not affected by hirudin at high concentrations (1 μmol/L) but were abolished by DX9065A, a specific inhibitor of the catalytic activity of factor Xa. Furthermore, no relaxations to factor Xa could be elicited in the presence of the NO synthase inhibitor Nω-nitro-l-arginine methyl ester (100 μmol/L), whereas relaxations were not altered in the presence of the inactive enantiomer Nω-nitro-d-arginine methyl ester (100 μmol/L). Addition of factor Xa together with thrombin induced relaxations that were larger than those induced by thrombin alone, whereas factor Xa had no additional effects on trypsin-induced relaxations. Furthermore, factor Xa relaxed thrombin-desensitized aortic rings but was ineffective in trypsin-desensitized tissues. These data suggest that factor Xa acts on a cleavable endothelial receptor that induces NO release, resulting in the relaxation of precontracted rat aortic rings. Factor Xa does not act through endothelial thrombin receptors but may activate another cleavable trypsin-sensitive receptor.
The serine protease factor Xa is one of the major late-stage enzymes activated during blood coagulation. Factor Xa acts in concert with factor Va to convert prothrombin into thrombin, which induces the formation of the fibrin clot by proteolysis of fibrinogen.1 Recently, it has become apparent that in addition to its hydrolytic activity on fibrinogen in the coagulation cascade, thrombin also acts through cell membrane receptors to induce various effects, such as vasorelaxation, contraction, mitogenesis, and release of various substances.2 Cloning of this cellular thrombin receptor has identified it as a G protein–coupled receptor activated through catalytic cleavage of the extracellular N-terminal loop and subsequent exposure of an endogenous activating sequence3 (tethered ligand receptor, also called PAR-1). Recently, a homologous receptor specifically activated by another serine protease, trypsin, has been cloned (PAR-2).4 This receptor has been localized in the intestine4 but is also present in cells of the vascular wall, such as endothelial cells.5 6
These cellular effects of thrombin and trypsin have prompted us to investigate whether factor Xa, the serine protease just upstream from thrombin in the coagulation cascade, could also induce effects on cells of the vascular wall. Recently, we and others have shown that endothelial cells express one receptor for factor Xa (EPR-1) and that factor Xa induces mitogenesis and an intracellular free Ca2+ increase in these cells.7 8 This suggested that the activation of factor Xa receptors on endothelial cells might induce vasoactive effects.
Materials and Methods
Male Sprague-Dawley rats (350 to 450 g, Iffa-Credo, L’Arbesle, France) were killed by cervical dislocation after pentobarbital anesthesia (50 mg/kg IP), and the thoracic aorta was rapidly removed and placed in an oxygenated (95% O2/5% CO2) Krebs’ solution composed of (mmol/L) NaCl 112, KCl 5, NaHCO3 25, KH2PO4 1, MgSO4 1.2, CaCl2 1.25, and glucose 11.5. The isolated aorta was then cut into rings (length, 3 to 4 mm), which were mounted between two stainless steel wires in an automated isolated organ system (Emka Technologies). In some cases, the endothelium was removed by gentle rubbing of the intimal surface. Rings were set up under 20 mN of tension and equilibrated for 1 hour in oxygenated Krebs’ solution, which was changed every 15 minutes. A first contraction to phenylephrine (1 μmol/L) was then recorded, followed by relaxation to acetylcholine (1 μmol/L), which induced ≈70% of relaxation in endothelium-intact rings and served as a control of endothelial integrity. Rings were then washed several times and incubated for a further 60 minutes before the beginning of the experiment. The effects of factor Xa, thrombin, and trypsin were tested on phenylephrine (1 μmol/L)–contracted rings. In order to test the effect of inhibitors on the effect of factor Xa or thrombin, rings were contracted by phenylephrine (1 μmol/L), and a first relaxation to factor Xa or thrombin was recorded. Rings were then rapidly washed and incubated for two 15-minute periods separated by another wash before a second contraction to phenylephrine. Antagonists were injected just after the last wash and were thus preincubated for 15 minutes before the injection of phenylephrine. Control experiments were performed in parallel by injecting vehicle instead of antagonists. A second relaxation was then recorded, and results were calculated as a percentage of the second relaxation relative to the first one. Contractions were analyzed in the same manner. The enzymes thrombin, trypsin, and factor Xa were kept on ice during the experiment. When two enzymes were injected together, the mixture was prepared just before the injection from ice-cold components to minimize any cross proteolysis between the enzymes. The protocols were approved by the Animal Care and Use Committee of Sanofi Recherche.
Factor Xa was from Enzyme Research, human α-thrombin was from Sigma Chemical Co, and trypsin was from Seromed. DX9065A was a kind gift from Daiichi (Tokyo, Japan), and hirudin (rHV2-Lys 47 variant) was from Sanofi Recherche (Toulouse, France).
Protease-Induced Relaxation of Phenylephrine-Contracted Aortic Rings
Thrombin, factor Xa, and trypsin induced relaxations of phenylephrine (1 μmol/L)–contracted rat aortic rings with an intact endothelium (Fig 1⇓). However, whereas thrombin induced relaxations consisting of a rapid phase followed by partial or total recontraction, factor Xa or trypsin induced slower and more sustained relaxations (Fig 1a⇓ through 1c). The relaxing effects of all three proteases were dose dependent (Fig 2⇓), with thrombin being slightly more potent (EC50, 3±1 nmol/L) than factor Xa (EC50, 16±1 nmol/L) or trypsin ( EC50, 24±9 nmol/L). However, although rings relaxed totally in response to acetylcholine (30 μmol/L, not shown, n=5), the three enzymes were not able to completely relax phenylephrine-contracted rat aortic rings. Maximal relaxations to thrombin, factor Xa, and trypsin were 45±5%, 59±7%, and 60±4% (n=7 to 15), respectively.
Factor Xa (30 nmol/L) had no effect on the basal tone of aortic rings, in the presence or absence of a functional endothelium (not shown, n=3).
Mechanism of Factor Xa–Induced Relaxations
Under physiological conditions, factor Xa cleaves circulating prothrombin to form thrombin, which may act on endothelial thrombin receptors to indirectly relax smooth muscle cells. In order to determine whether the effect of factor Xa may be due to thrombin generation in situ, aortic rings were incubated with a high concentration (1 μmol/L) of the potent thrombin inhibitor hirudin before the addition of factor Xa. As shown in Fig 3⇓, hirudin had no effect on either the contraction to phenylephrine (panel A) or the response to factor Xa (panel B), whereas thrombin (10 nmol/L)–induced relaxations, which were equivalent to those induced by 30 nmol/L factor Xa (see Fig 2⇑), were abolished by 100 nmol/L hirudin (not shown, n=3).
It was then determined whether factor Xa induced relaxations through the release of endothelium-dependent factors or through a direct effect on aortic smooth muscle cells. In endothelium-denuded aortic rings, factor Xa had no effect on phenylephrine-induced contractions (Fig 1d⇑). Furthermore, preincubation of endothelium-intact aortic rings with the NO synthase inhibitor L-NAME (100 μmol/L), but not the inactive enantiomer D-NAME (100 μmol/L), increased phenylephrine-induced contractions and abolished factor Xa–induced relaxations (Fig 3⇑.).
DX9065A is a potent inhibitor of the catalytic activity of factor Xa (Ki, 41 nmol/L)9 but has no effect on thrombin (Ki, >1 mmol/L).9 As shown in Fig 1e⇑, preincubation of rat aortic rings with DX9065A (1 μmol/L) abolished the relaxation induced by factor Xa but did not affect a subsequent relaxation to thrombin (30 nmol/L).
Actually, a quantitative analysis shows that relatively low (100 nmol/L) concentrations of DX9065A significantly decreased and higher concentrations (1 μmol/L) abolished factor Xa–induced relaxations, resulting in an IC50 value of 70±10 nmol/L (Fig 3⇑). DX9065A (1 μmol/L) had no effect on acetylcholine (1 μmol/L)–induced relaxations (not shown, n=3).
Interaction Between Factor Xa and Thrombin or Trypsin-Induced Relaxations
Since none of the proteases used was able to completely relax phenylephrine-contracted aortic rings, it was possible to test whether the addition of factor Xa would be able to increase the relaxation induced by maximally active concentrations of thrombin or trypsin. As shown in Fig 1f⇑ and Fig 4A⇓, addition of a mixture of a maximally effective concentration of thrombin (30 nmol/L) and factor Xa (30 nmol/L) resulted in a relaxation that was significantly increased compared with the relaxation induced by either thrombin or factor Xa alone. By contrast, the same concentration of factor Xa was not able to increase the relaxation induced by a maximally effective concentration of trypsin (100 nmol/L, Fig 1g⇑ and Fig 4B⇓), whereas acetylcholine (1 μmol/L) induced 95±4% relaxation in the same aortic rings.
Furthermore, when phenylephrine-contracted rat aortic rings were relaxed with a high concentration of thrombin (100 nmol/L) for >30 minutes, a second challenge with thrombin was ineffective, showing complete desensitization of the thrombin response under these conditions (Fig 5A⇓). However, sequential addition of factor Xa (30 nmol/L) in these thrombin-desensitized rings still elicited a marked relaxation (Fig 5A⇓). The relaxant response to trypsin (100 nmol/L) could also be desensitized by repeat injection of the protease, but several sequential additions of trypsin were necessary to abolish the response to the protease (Fig 5B⇓). As shown in Fig 5B⇓, factor Xa (30 nmol/L) was ineffective in trypsin-desensitized rings, whereas acetylcholine (1 μmol/L) was still able to induce nearly complete relaxation.
We show in the present study that coagulation factor Xa is a potent endothelium-dependent relaxant of rat aortic ring preparations. Factor Xa acts at concentrations similar to those expected to be generated locally during the induction of coagulation.10 Factor Xa relaxes phenylephrine-contracted rat aortic rings through a NO synthase–dependent pathway, because relaxations were abolished in the presence of L-NAME, a selective inhibitor of NO synthase, and were not affected by the inactive enantiomer D-NAME. Furthermore, the finding that factor Xa–induced relaxations were completely inhibited by the selective inhibitor DX9065A shows that the catalytic activity of factor Xa is required for the relaxant effect of factor Xa. This evidently suggests that factor Xa is acting through a cleavable receptor present on endothelial cells. This mechanism of factor Xa–induced relaxations is very similar to the mechanism reported for thrombin and trypsin.11 This evidently raises the question as to whether factor Xa is acting through thrombin receptors or the homologous trypsin receptors (PAR-2) to induce relaxations. Both of these receptors are cleavable tethered-ligand receptors that have been shown to exist on endothelial cells5 6 12 13 and would be expected, on the basis of the amino acid sequence of their cleavable region, to be activatable by other serine proteases, such as factor Xa.
Concerning the thrombin receptor, several facts would tend to exclude it as the receptor implicated in the relaxing effect of factor Xa. Indeed, factor Xa–induced relaxations differed from those induced by thrombin in numerous ways: (1) whereas the relaxations induced by thrombin were rapid, as already reported by previous work,11 and followed by partial or total recontraction, which might be due to release of contracting factors like endothelin from endothelial cells,14 factor Xa produced relaxations that were sustained or decreased only slowly with time; (2) the maximal effect of factor Xa was slightly higher than the effect of thrombin; (3) factor Xa was not acting through the generation of thrombin from prothrombin, because the extremely potent thrombin inhibitor hirudin at high concentrations had no effect on factor Xa–induced relaxations; (4) factor Xa had no contractile effect per se on deendothelialized aortic rings, whereas thrombin has been shown to induce contractions at low concentrations15 16 ; (5) factor Xa in combination with thrombin induced relaxations that were stronger than those induced by maximally effective concentrations of thrombin; (6) factor Xa was still able to relax aortic rings rendered unresponsive to thrombin by repeat administration of the protease, showing that most of the relaxant response to factor Xa was not due to thrombin receptor activation; (7) factor Xa has very low catalytic activity against the peptide representing the cleavage site of thrombin receptors17 ; and (8) preincubation with DX9065A completely blocked factor Xa–induced relaxations but did not affect subsequent relaxing effects of thrombin. Actually, factor Xa has been shown to bind to endothelial cells even in the presence of DX9065A,8 and catalytically inactivated factor Xa has been shown to inhibit Ca2+ signaling through factor Xa receptors in Mardin-Darby canine kidney cells.18 If factor Xa was acting through thrombin receptors, the complex between factor Xa and DX9065A would therefore have been expected to inhibit at least part of the thrombin response. Taken together, this whole range of converging evidence makes it highly unlikely that the receptor through which factor Xa induces relaxations in rat aortic rings is the thrombin receptor.
Factor Xa induced relaxations with a potency, maximal effect, and time course similar to those induced by trypsin. Considering the similarity between the effects of factor Xa and trypsin, the receptor for trypsin, PAR-2, would be an attractive candidate for being the receptor through which factor Xa acts on endothelial cells to induce vasorelaxation. This would be consistent with the finding that factor Xa in combination with trypsin was not able to induce increased relaxations compared with maximally active concentrations of trypsin alone and was unable to relax trypsin-desensitized aortic rings. Furthermore, lacking the “60-insertion” region and the anion-binding exosite of thrombin, the active site configuration of factor Xa is structurally somewhat more similar to trypsin than to thrombin,19 which is also confirmed by the fact that selective factor Xa inhibitors like DX9065A have some affinity for trypsin and are inactive on thrombin.9 Also, factor Xa did not induce any contractile activity per se, in contrast to thrombin and in analogy to PAR-2–activating peptides.12 16 Interestingly, we and others have recently shown that factor Xa binding sites exist on endothelial cells and are coupled to Ca2+ release from intracellular stores.7 8 Since intracellular free Ca2+ increase is generally associated with the release of NO from endothelial cells, this would be consistent with the present results showing NO-dependent relaxation of aortic rings by factor Xa. Nevertheless, 125I-factor Xa binding was inhibited by antibodies to EPR-1 and was insensitive to DX9065A, which led us to suggest that EPR-1 might be the endothelial receptor for factor Xa.8 At first sight, this hypothesis would not be consistent with the requirement for catalytically active factor Xa, because the amino acid sequence of EPR-1 does not exhibit any consensus sequence for cleavage by serine proteases. However, this apparent discrepancy between studies on cultured cells and isolated organ experiments can be resolved, because EPR-1 might be necessary to localize factor Xa in proximity to the cellular membrane, where it may then activate another cleavable receptor, which might be PAR-2. Unfortunately, this hypothesis could not be tested directly, because the antibodies against PAR-2 are not available at this time. However, the desensitization experiments clearly indicate that a trypsin-sensitive receptor is involved in the relaxant effect of factor Xa.
In conclusion, the present study shows for the first time that factor Xa at low concentrations is able to induce endothelium-dependent relaxations of rat aortic rings. Factor Xa only produces relaxations when catalytically active, does not work through thrombin receptors, and acts through a trypsin-sensitive receptor. These results therefore suggest that a cleavable receptor, similar to PAR-2, the recently described receptor for trypsin, is involved in the vasorelaxant effects of factor Xa.
Selected Abbreviations and Acronyms
|D-NAME||=||Nω-nitro-d-arginine methyl ester|
|EPR-1||=||effector cell protease receptor-1|
|L-NAME||=||Nω-nitro-l-arginine methyl ester|
- Received April 8, 1997.
- Accepted July 15, 1997.
- © 1997 American Heart Association, Inc.
Grand RJA, Turnell AS, Grabham PW. Cellular consequences of thrombin-receptor activation. Biochem J. 1996;313:353-368.
Nystedt S, Emilsson K, Wahlestedt C, Sundelin J. Molecular cloning of a potential proteinase activated receptor. Proc Natl Acad Sci U S A. 1994;91:9208-9212.
Nystedt S, Ramakrishnan V, Sundelin J. The proteinase-activated receptor 2 is induced by inflammatory mediators in human endothelial cells. J Biol Chem. 1996;271:14910-14915.
Nicholson AC, Nachman RL, Altieri DC, Summers BD, Ruf W, Edgington TS, Hajjar DP. Effector cell protease receptor-1 is a vascular receptor for coagulation factor Xa. J Biol Chem. 1996;271:28407-28413.
Nesheim ME, Tracy RP, Mann KG. ‘Clotspeed,’ a mathematical simulation of the functional properties of prothrombinase. J Biol Chem. 1994;259:1447-1453.
Rapoport RM, Draznin MB, Murad F. Mechanisms of adenosine triphosphate-, thrombin-, and trypsin-induced relaxations of rat thoracic aorta. Circ Res. 1984;55:468-479.
Hwa JJ, Ghibaudi L, Williams P, Chintala M, Zhang R, Chatterjee M, Sybertz E. Evidence for the presence of a proteinase-activated receptor distinct from the thrombin receptor in vascular endothelial cells. Circ Res. 1996;78:581-588.
Magazine HI, Srivastava KD. Thrombin-induced vascular reactivity is modulated by ETB receptor-coupled nitric oxide release in rat aorta. Am J Physiol. 1996;271:C923-C928.
Magazine HI, King JM, Srivastava KD. Protease activated receptors modulate aortic vascular tone. Int J Cardiol. 1996;53:S75-S80.
Parry MAA, Myles T, Tschopp J, Stone SR. Cleavage of the thrombin receptor: identification of potential activators and inactivators. Biochem J. 1996;320:335-341.
Camerer E, Røttingen JA, Iversen JG, Prydz H. Coagulation factors VII and X induce Ca2+ oscillations in Madin-Darby canine kidney cells only when proteolytically active. J Biol Chem. 1996;271:29034-29042.