Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Thrombin Up-regulates Tissue Factor Pathway Inhibitor-2 Synthesis through a Cyclooxygenase-2-dependent, Epidermal Growth Factor Receptor-independent Mechanism

2004; Elsevier BV; Volume: 279; Issue: 7 Linguagem: Inglês

10.1074/jbc.m306679200

ISSN

1083-351X

Autores

Véronique Neaud, Jennifer Gillibert Duplantier, Claire Mazzocco, Walter Kisiel, Jean Rosenbaum,

Tópico(s)

Heparin-Induced Thrombocytopenia and Thrombosis

Resumo

The serine proteinase inhibitor tissue factor pathway inhibitor-2 (TFPI-2) inhibits the tissue factor-factor VIIa complex and thereby impairs factor Xa and subsequently thrombin generation. Here we show that thrombin itself up-regulates TFPI-2 mRNA and protein expression in human liver myofibroblasts, a cell type shown to express high levels of TFPI-2 (Neaud, V., Hisaka, T., Monvoisin, A., Bedin, C., Balabaud, C., Foster, D. C., Desmoulière, A., Kisiel, W., and Rosenbaum, J. (2000) J. Biol. Chem. 275, 35565–35569). This effect required thrombin catalytic activity, as shown by its abolition with hirudin. Although the thrombin effect could be mimicked by agonists of both protease-activated receptor (PAR)-1 and PAR-4, it was largely blocked by a PAR-1 blocking antibody. Transactivation of the epidermal growth factor (EGF) receptor has been reported as a common event in thrombin signaling. However, thrombin did not detectably transactivate the EGF receptor in liver myofibroblasts, and blocking the EGF receptor did not affect TFPI-2 induction. On the other hand, thrombin increased the expression of cyclooxygenase-2 (COX-2) mRNA via a MAPK-dependent pathway, and a specific COX-2 inhibitor abolished the effect of thrombin on TFPI-2 expression. Thus, thrombin, through PAR-1 signaling, up-regulates the synthesis of TFPI-2 via a MAPK/COX-2-dependent pathway. The up-regulation of TFPI-2 expression by thrombin could in turn down-regulate thrombin generation and contribute to limit blood coagulation. The serine proteinase inhibitor tissue factor pathway inhibitor-2 (TFPI-2) inhibits the tissue factor-factor VIIa complex and thereby impairs factor Xa and subsequently thrombin generation. Here we show that thrombin itself up-regulates TFPI-2 mRNA and protein expression in human liver myofibroblasts, a cell type shown to express high levels of TFPI-2 (Neaud, V., Hisaka, T., Monvoisin, A., Bedin, C., Balabaud, C., Foster, D. C., Desmoulière, A., Kisiel, W., and Rosenbaum, J. (2000) J. Biol. Chem. 275, 35565–35569). This effect required thrombin catalytic activity, as shown by its abolition with hirudin. Although the thrombin effect could be mimicked by agonists of both protease-activated receptor (PAR)-1 and PAR-4, it was largely blocked by a PAR-1 blocking antibody. Transactivation of the epidermal growth factor (EGF) receptor has been reported as a common event in thrombin signaling. However, thrombin did not detectably transactivate the EGF receptor in liver myofibroblasts, and blocking the EGF receptor did not affect TFPI-2 induction. On the other hand, thrombin increased the expression of cyclooxygenase-2 (COX-2) mRNA via a MAPK-dependent pathway, and a specific COX-2 inhibitor abolished the effect of thrombin on TFPI-2 expression. Thus, thrombin, through PAR-1 signaling, up-regulates the synthesis of TFPI-2 via a MAPK/COX-2-dependent pathway. The up-regulation of TFPI-2 expression by thrombin could in turn down-regulate thrombin generation and contribute to limit blood coagulation. The serine proteinase thrombin is a key factor in hemostasis as it converts fibrinogen to fibrin thus leading to clot formation. Besides this hemostatic function, thrombin is also known to elicit many cellular signals most of which originate from the interaction of thrombin with specific cellular receptors, the so-called protease-activated receptors (PAR), 1The abbreviations used are: PARprotease-activated receptorsTFPI-2tissue factor pathway inhibitor-2COX-2cyclooxygenase-2MAPKmitogen-activated protein kinaseEGFepidermal growth factorPBSphosphate-buffered salineGAPDHglyceraldehyde-3-phosphate dehydrogenaseECMextracellular matrixRTreverse transcriptaseMOPS4-morpholinepropanesulfonic acid. that belong to the G protein-coupled receptor family. Thrombin can activate PAR-1, PAR-3, and PAR-4 (1Coughlin S.R. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 11023-11027Crossref PubMed Scopus (526) Google Scholar). Binding is followed by cleavage of the amino terminus of the receptor. This will unmask a new amino terminus that behaves as a tethered ligand inducing receptor activation and downstream signalization. PAR-1 and PAR-4 appear to be the major signaling receptors, with PAR-3 being used as a co-factor for PAR-4 activation (2Nakanishi-Matsui M. Zheng Y. Sulciner D. Weiss E. Ludeman M. Coughlin S. Nature. 2000; 404: 609-613Crossref PubMed Scopus (470) Google Scholar). Many cellular functions such as migration, mitogenesis, and cytokine secretion are controlled by thrombin. These effects are mediated through several transduction pathways such as those of Ras/MAPK (3Mallat A. Gallois C. Tao J. Habib A. Maclouf J. Mavier P. Préaux A.M. Lotersztajn S. J. Biol. Chem. 1998; 273: 27300-27305Abstract Full Text Full Text PDF PubMed Scopus (97) Google Scholar, 4Guo Y.L. Peng M. Kang B. Williamson J.R. Biochem. Biophys. Res. Commun. 1997; 240: 405-408Crossref PubMed Scopus (5) Google Scholar), phosphatidylinositol 3-kinase (5Phillips-Mason P.J. Raben D.M. Baldassare J.J. J. Biol. Chem. 2000; 275: 18046-18053Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar), protein kinase C (6Ludwicka-Bradley A. Tourkina E. Suzuki S. Tyson E. Bonner M. Fenton II, J.W. Hoffman S. Silver R.M. Am. J. Respir. Cell Mol. Biol. 2000; 22: 235-243Crossref PubMed Scopus (61) Google Scholar, 7Maulon L. Mari B. Bertolotto C. Ricci J.E. Luciano F. Belhacene N. Deckert M. Baier G. Auberger P. Oncogene. 2001; 20: 1964-1972Crossref PubMed Scopus (31) Google Scholar), or JAK-STAT (8Madamanchi N.R. Li S. Patterson C. Runge M.S. J. Biol. Chem. 2001; 276: 18915-18924Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar). Thrombin has also been shown to increase the expression of cyclooxygenase-2 (COX-2), thus leading to increased prostaglandin production (3Mallat A. Gallois C. Tao J. Habib A. Maclouf J. Mavier P. Préaux A.M. Lotersztajn S. J. Biol. Chem. 1998; 273: 27300-27305Abstract Full Text Full Text PDF PubMed Scopus (97) Google Scholar). The functional relevance of these pathways could be shown by using inhibitors. Thus, blocking the Ras/MAPK pathway inhibited mitogenesis (9Molloy C.J. Pawlowski J.E. Taylor D.S. Turner C.E. Weber H. Peluso M. J. Clin. Investig. 1996; 97: 1173-1183Crossref PubMed Scopus (86) Google Scholar) or the thrombin-induced expression of PAR-1 (10Ellis C.A. Malik A.B. Gilchrist A. Hamm H. Sandoval R. Voyno-Yasenetskaya T. Tiruppathi C. J. Biol. Chem. 1999; 274: 13718-13727Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar), whereas inhibition of protein kinase C blocked thrombin-induced secretion of interleukin-8 (6Ludwicka-Bradley A. Tourkina E. Suzuki S. Tyson E. Bonner M. Fenton II, J.W. Hoffman S. Silver R.M. Am. J. Respir. Cell Mol. Biol. 2000; 22: 235-243Crossref PubMed Scopus (61) Google Scholar). The role of COX-2 was also shown as COX-2 inhibitors increase the thrombin mitogenic effect by blocking the production of prostaglandins that inhibit proliferation through enhancement of cAMP synthesis (3Mallat A. Gallois C. Tao J. Habib A. Maclouf J. Mavier P. Préaux A.M. Lotersztajn S. J. Biol. Chem. 1998; 273: 27300-27305Abstract Full Text Full Text PDF PubMed Scopus (97) Google Scholar). Finally, like other GPCR agonists, thrombin has been shown in many instances to transactivate the EGF receptor (11Daub H. Weiss F.U. Wallasch C. Ullrich A. Nature. 1996; 379: 557-560Crossref PubMed Scopus (1329) Google Scholar, 12Prenzel N. Zwick E. Daub H. Leserer M. Abraham R. Wallasch C. Ullrich A. Nature. 1999; 402: 884-888Crossref PubMed Scopus (1501) Google Scholar). This was shown to be responsible for mitogenesis (11Daub H. Weiss F.U. Wallasch C. Ullrich A. Nature. 1996; 379: 557-560Crossref PubMed Scopus (1329) Google Scholar) as well as cell migration (13Kalmes A. Vesti B.R. Daum G. Abraham J.A. Clowes A.W. Circ. Res. 2000; 87: 92-98Crossref PubMed Scopus (167) Google Scholar). protease-activated receptors tissue factor pathway inhibitor-2 cyclooxygenase-2 mitogen-activated protein kinase epidermal growth factor phosphate-buffered saline glyceraldehyde-3-phosphate dehydrogenase extracellular matrix reverse transcriptase 4-morpholinepropanesulfonic acid. The events leading to thrombin generation are very well known. The complex formed by tissue factor and factor VIIa activates factor X. Factor Xa together with factor Va then converts prothrombin to thrombin. The catalytic activity of the tissue factor-factor VIIa complex is subject to inhibition by two related molecules, the tissue factor pathway inhibitors-1 and -2. Both TFPIs contain three tandemly arranged Kunitz-type domains. Besides the tissue factor-factor VIIa complex, TFPI-2 also inhibits trypsin, plasmin, and kallikrein but not thrombin (14Petersen L.C. Sprecher C.A. Foster D.C. Blumberg H. Hamamoto T. Kisiel W. Biochemistry. 1996; 35: 266-272Crossref PubMed Scopus (139) Google Scholar). TFPI-2 exists as three isoforms of 27, 31, and 33 kDa that are synthetic products of a single gene and arise from differential glycosylation. TFPI-2 is most abundant in placenta. It is also expressed in several adult tissues, including the liver (15Miyagi Y. Koshikawa N. Yasumitsu H. Miyagi E. Hirahara F. Aoki I. Misugi K. Umeda M. Miyazaki K. J. Biochem. (Tokyo). 1994; 116: 939-942Crossref PubMed Scopus (96) Google Scholar, 16Sprecher C.A. Kisiel W. Mathewes S. Foster D.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 3353-3357Crossref PubMed Scopus (186) Google Scholar). TFPI-2 is poorly secreted and is mostly sequestered within the extracellular matrix, whereas TFPI-1 is mainly secreted. We have shown previously (17Neaud V. Hisaka T. Monvoisin A. Bedin C. Balabaud C. Foster D.C. Desmoulière A. Kisiel W. Rosenbaum J. J. Biol. Chem. 2000; 275: 35565-35569Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar) that human liver myofibroblasts synthesize TFPI-2. The goal of this study was to examine the regulation of TFPI-2 expression by thrombin. Materials—Culture medium and additives, recombinant human epidermal growth factor, and Moloney murine leukemia virus-reverse transcriptase were from Invitrogen. The Qiagen RNeasy minikit and Hot-start Taq polymerase were from Qiagen (Courtaboeuf, France). [α-32P]dCTP, Hybond N+ membrane, the ECL reagent, and the Ready-to-Go DNA labeling kit were from Amersham Biosciences. Ultrahyb solution was from Ambion (Austin, TX). The ECL reagent was from Amersham Biosciences. Recombinant human TFPI-2 was expressed in baby hamster kidney cells and purified as described previously (16Sprecher C.A. Kisiel W. Mathewes S. Foster D.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 3353-3357Crossref PubMed Scopus (186) Google Scholar) by a series of chromatographic steps. The PAR-1 agonist peptide SFLLRNPNRKYEPF and the PAR-4 agonist peptide GYPGQV were from Neosystem (Strasbourg, France). The protein assay was from Bio-Rad. Anti-phosphotyrosine antibody P-Tyr-100 was from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-phospho-p44/p42 MAPK antibody (Thr202/Tyr204) was from Cell Signaling Technology (Beverly, MA). Antitotal MAPK was from Promega (Charbonnières, France). Anti-phospho-EGF receptor recognizing phosphorylated Tyr1173 and a blocking anti-EGF receptor antibody were from Upstate Biotechnologies Inc. (Lake Placid, NY). Secondary peroxidase-conjugated antibodies were from Dako (Glostrup, Denmark). The PAR-1 cleavage-blocking antibody H-11 was from Santa Cruz Biotechnology. Immobilon membranes were from Millipore (Saint-Quentin en Yvelynes, France). S2251 was from Chromogenix (Mölndal, Sweden). NS-398 was from Biomol Research Laboratories (Plymouth Meeting, PA). PD98059 and AG1478 were from Calbiochem. Purified human thrombin (1000 NIH units/mg, T4393) and all other chemicals were from Sigma. Cell Culture—Human hepatic myofibroblasts were obtained from explants of non-tumoral liver resected during partial hepatectomy and characterized as described previously (18Win K. Charlotte F. Mallat A. Cherqui D. Martin N. Mavier P. Préaux A.M. Dhumeaux D. Rosenbaum J. Hepatology. 1993; 18: 137-145Crossref PubMed Scopus (119) Google Scholar, 19Blazejewski S. Préaux A.M. Mallat A. Brochériou I. Mavier P. Dhumeaux D. Hartmann D. Schuppan D. Rosenbaum J. Hepatology. 1995; 22: 788-797Crossref PubMed Google Scholar). Myofibroblasts were grown in Dulbecco's modified Eagle's medium containing 5% fetal calf serum, 5% pooled human AB serum, and 5 ng/ml recombinant human epidermal growth factor (EGF). When looking at TFPI-2 regulation by thrombin, cells were seeded in 35-mm dishes at an initial density of 9000/cm2 in complete medium. When confluent, they were washed and left in serum and EGF-free Waymouth medium for 3 days. Thrombin was then added in fresh Waymouth medium. Cells from five different donors were used in these experiments. ECV 304 cells are derived from a human bladder carcinoma. They were grown in Medium 199. RNA Isolation and Northern Blot—At the designated time points, total RNA was isolated using the Qiagen RNeasy minikit. For Northern blot, 2 μg of RNAs were separated on a 1.5% agarose gel containing ethidium bromide in MOPS buffer. Running buffer and gel contained 0.2 m formaldehyde. The RNAs were transferred onto a Hybond N+ membrane by downward capillary transfer in running buffer. Examination of the stained membrane under UV light was used to confirm the quality of loading and transfer. The probes used were the full-length human TFPI-2 cDNA (16Sprecher C.A. Kisiel W. Mathewes S. Foster D.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 3353-3357Crossref PubMed Scopus (186) Google Scholar), a fragment of the human TFPI-1 cDNA (a gift from P. van der Logt, Leiden, Netherlands), and a fragment of the human COX-2 cDNA (a gift from A. Dannenberg, New York). Probes were labeled with [α-32P]dCTP by random priming using the Ready-to-Go kit. Hybridization was performed using the Ultrahyb solution. The blots were washed in stringent conditions (0.1× SSC, 0.1% SDS at 65 °C). The blots were further rehybridized with a probe to human glyceraldehyde-3-phosphate dehydrogenase (GAPDH, a gift from A. Pawlak, Créteil, France). Bands were quantified after image acquisition on a Macintosh computer (Apple Computers, Les Ulis, France) by using the Kodak 1D Image Analysis software (Eastman Kodak Co.). Results are expressed as TFPI-2/GAPDH ratios and normalized according to the control value, set at 1. RT-PCR—For RT-PCR, 1 μg of total RNA was reverse-transcribed with Moloney murine leukemia virus. An aliquot was amplified by PCR with Qiagen Hot-start Taq polymerase, using the following primers that were described previously (20Shimizu S. Gabazza E.C. Hayashi T. Ido M. Adachi Y. Suzuki K. Am. J. Physiol. 2000; 279: L503-L510PubMed Google Scholar, 21Miyata S. Koshikawa N. Yasumitsu H. Miyazaki K. J. Biol. Chem. 2000; 275: 4592-4598Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar): PAR-1 sense primer, 5′-CAG TTT GGG TCT GAA TTG TGT CG-3′, and antisense primer, 5′-TGC ACG AGC TTA TGC TGC TGA C-3′; PAR-2 sense primer, 5′-GGC CAA TCT GGC CTT GGC TGA C-3′, and antisense primer, 5′-GGG CAG GAA TGA AGA TGG TCT GC-3′; PAR-3 sense primer, 5′-TCCCCTTTTCTGCCTTGGAAG-3′, and antisense primer, 5′-AAA CTG TTG CCC ACA CCA GTC CAC-3′; PAR-4 sense primer, 5′-AAC CTC TAT GGT GCC TAC GTG C-3′, and antisense primer, 5′-CCA AGC CCA GCT AAT TTT TG-3′. PCR conditions were as follows: initial denaturation for 5 min at 94 °C, followed by 35 cycles of 94 °C 30 s/55 °C (PAR-1, -3, and -4) or 60 °C (PAR-2)/30 s, 72 °C/45 s. The sizes of the amplified products are 592, 358, 513, and 542 for PAR 1–4, respectively. Negative control was performed for each reaction and included the omission of the RT step or the omission of cDNA in the PCR mix. Western Blot for TFPI-2—TFPI-2 was measured in extracellular matrix (ECM) extracts as described previously (17Neaud V. Hisaka T. Monvoisin A. Bedin C. Balabaud C. Foster D.C. Desmoulière A. Kisiel W. Rosenbaum J. J. Biol. Chem. 2000; 275: 35565-35569Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). To prepare ECM extracts for Western blotting, cells were grown to confluence, washed twice with serum-free Dulbecco's modified Eagle's medium, and incubated for 24 h in the same medium. The ECM was prepared as described by Rao et al. (22Rao C.N. Reddy P. Liu Y. O'Toole E. Reeder D. Foster D.C. Kisiel W. Woodley D.T. Arch. Biochem. Biophys. 1996; 335: 82-92Crossref PubMed Scopus (88) Google Scholar). Briefly, cells were washed 3 times with PBS and lysed with 0.5% (v/v) Triton X-100 in PBS for 20 min at room temperature. Cellular DNA was quantified as described (23Labarca C. Paigen K. Anal. Biochem. 1980; 102: 344-352Crossref PubMed Scopus (4553) Google Scholar). The remaining ECM was washed 3 times with PBS and another 3 times with 20 mm Tris-HCl, pH 7.4, containing 100 mm NaCl and 0.1% (v/v) Tween 20. ECM was then incubated for 2 h at room temperature with reducing Laemmli buffer and collected by scraping. ECM extracts normalized for cell DNA were analyzed by SDS-PAGE on 15% gels and transferred to a polyvinylidene difluoride membrane by semi-dry transfer (Transblot-SD, Bio-Rad). The membrane was blocked with 4% skimmed milk in 10 mm Tris-HCl, pH 7.4, 150 mm NaCl, 0.1% Tween 20 (TBS-Tween) for 2 h at room temperature and incubated overnight at 4 °C with anti-TFPI-2 IgG (10 μg/ml) (22Rao C.N. Reddy P. Liu Y. O'Toole E. Reeder D. Foster D.C. Kisiel W. Woodley D.T. Arch. Biochem. Biophys. 1996; 335: 82-92Crossref PubMed Scopus (88) Google Scholar) in TBS-Tween containing 1% bovine serum albumin and then 1 h at room temperature with a peroxidase-conjugated anti-rabbit IgG antibody. Detection was achieved by enhanced chemiluminescence. The signals were quantified with the Kodak 1D Image Analysis software. Preparation of Cell Extracts and Assay of Protein Phosphorylation by Western Blot—For studies of activation of MAPK, Akt, and tyrosine phosphorylation, cells were grown to confluency and serum-starved for 3 days. They were then exposed to various agonists. At the end of the incubation, cell lysates were prepared in the presence of proteases and phosphatases inhibitors as described (24Neaud V. Faouzi S. Guirouilh J. Le Bail B. Balabaud C. Bioulac-Sage P. Rosenbaum J. Hepatology. 1997; 26: 1458-1466Crossref PubMed Scopus (150) Google Scholar). Proteins were measured with a Bio-Rad assay. Equivalent amounts of proteins were separated by SDS-PAGE, transferred to polyvinylidene difluoride membranes, and analyzed by Western blotting with antibodies to phosphotyrosine, phospho-MAPK, phospho-Akt-1, or phospho-EGF receptor. The signals were revealed with a chemiluminescent substrate. The blots were then stripped and rehybridized with an antibody to total MAPK to confirm equal loading. Assay for TFPI-2 Inactivation by Thrombin—The assay was based on the ability of TFPI-2 to inhibit plasmin. Recombinant human TFPI-2 (2 μg) was incubated with 1.5 units of thrombin for 3 h at 37 °C in a total volume of 17 μl. This was followed by the addition of hirudin to neutralize thrombin. Plasmin activity was monitored by using the colorimetric substrate S2251, as described (25Lottenberg R. Christensen U. Jackson C.M. Coleman P.L. Methods Enzymol. 1981; 80: 341-361Crossref PubMed Scopus (338) Google Scholar, 26Monvoisin A. Neaud V. De Lédinghen V. Dubuisson L. Balabaud C. Bioulac-Sage P. Desmoulière A. Rosenbaum J. J. Hepatol. 1999; 30: 511-518Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar). The reaction mixture contained S2251 with either purified plasmin alone or in the presence of native or thrombin-treated TFPI-2. Optimal concentrations of reagents were determined in preliminary experiments. All assays were carried out in duplicate. An aliquot of the digestion mixtures was also analyzed by Western blot with a polyclonal anti-TFPI-2 antibody. Northern blot with total RNA showed several transcripts for human TFPI-2. In addition to the 1.8-kb transcript, the previously described 1.2-kb transcript actually resolved into separate 1.1- and 1.3-kb bands, respectively. These two species could not be discriminated enough for separate quantitation on gels. It is likely that they arise through differential polyadenylation (15Miyagi Y. Koshikawa N. Yasumitsu H. Miyagi E. Hirahara F. Aoki I. Misugi K. Umeda M. Miyazaki K. J. Biochem. (Tokyo). 1994; 116: 939-942Crossref PubMed Scopus (96) Google Scholar). Two higher molecular weight bands were also found in minor amounts. Addition of thrombin to human liver myofibroblasts in serum-free conditions dose-dependently increased the expression of TFPI-2 transcripts (Fig. 1, A and C). The increase could be observed as soon as 3 h following thrombin addition and was maximal after 24 h (Fig. 1, B and D). All transcripts were up-regulated in parallel. Variations in TFPI-2 protein expression were studied by measuring the amount of TFPI-2 deposited into the ECM, by using Western blot (Fig. 1E). When the intensity of the 31-kDa band was measured in four independent experiments conducted in quadruplicate, there was a significant increase as compared with the control (1.85 ± 0.64-fold, p = 0.006, Mann and Whitney test). TFPI-2 is closely related to TFPI-1. We thus also studied TFPI-1 expression and its regulation by thrombin by Northern blot. Our results show first that TFPI-1 is indeed expressed in human liver myofibroblasts and that its expression is also up-regulated by thrombin (Fig. 2). In order to check whether thrombin effects were specific to liver myofibroblasts, we used the human bladder carcinoma cell line ECV 304, known to express TFPI-2 (27Ott I. Miyagi Y. Miyazaki K. Heeb M.J. Mueller B.M. Rao L.V. Ruf W. Arterioscler. Thromb. Vasc. Biol. 2000; 20: 874-882Crossref PubMed Google Scholar). As shown in Fig. 3, thrombin also up-regulated TFPI-2 in ECV 304 cells. The kinetics were different, however, since expression peaked at 3 h and was back to control values at 24 h. In addition, although all transcripts were up-regulated, the 1.1/1.3-kb level remained elevated for a longer time. Most thrombin effects are believed to be mediated through protease-activated receptors. We first checked, by using RT-PCR, whether human liver myofibroblasts expressed PARs. As shown in Fig. 4A, these cells expressed the three putative thrombin receptors, PAR-1, PAR-3, and PAR-4, as well as PAR-2, a receptor for trypsin and factor Xa. In order to know whether thrombin actually used these receptors, we incubated the cells with thrombin together with hirudin, which specifically inhibits thrombin catalytic activity. As demonstrated in Fig. 4, B and C, hirudin completely abolished the stimulating effect of thrombin on TFPI-2 mRNA expression. Hirudin by itself had no effect. The PAR-1 agonist peptide SFLLRNPNRKYEPF dose-dependently up-regulated TFPI-2 mRNA expression in myofibroblasts (Fig. 4, D and E). The PAR-4 agonist peptide GYPGQV also increased TFPI-2 mRNA expression at a concentration of 100 μm. In order to discriminate the relevance of each receptor, we took advantage of the existence of an antibody that blocks the cleavage of PAR-1 by thrombin. As shown in Fig. 4F, this antibody markedly reduced thrombin-induced TFPI-2 expression. We then analyzed the transduction pathways involved in the effect of thrombin. Thrombin quickly induced the phosphorylation of both ERK1 and ERK2 (Fig. 5A). Addition of the MEK inhibitor PD98059 to thrombin-stimulated cells completely blocked TFPI-2 induction (Fig. 5, B and C). As observed previously, thrombin rapidly up-regulated the expression of COX-2 mRNA in liver myofibroblasts. The maximal increase was seen 3 h after stimulation, and COX-2 transcripts returned to control values at 24 h (Fig. 6A). Induction by thrombin was prevented by using the MAPKK inhibitor PD98059 (data not shown). In order to know whether COX-2 was involved in thrombin-induced up-regulation of TFPI-2 expression, cells were exposed to thrombin in the absence or in the presence of the specific COX-2 inhibitor NS-398. As shown in Fig. 6, B and C, NS-398 completely abolished the increase in TFPI-2 expression evoked by thrombin. NS-398 had no effect on the induction of COX-2 mRNA expression by thrombin (not shown). The inhibitory effect of NS-398 could be completely reversed by the direct addition of prostaglandin E2 to the cells (Fig. 6B). Several cellular effects of thrombin have been shown to be relayed through the transactivation of the EGF receptor by thrombin. This possibility was first evaluated in our model by investigating the phosphorylation status of the EGF receptor in thrombin-stimulated cells. When cellular extracts were analyzed by Western blotting with an anti-phosphotyrosine antibody, no increase in the phosphorylation of bands co-migrating with the EGF receptor (as detected by EGF treatment) could be seen over a period of 60 min (Fig. 7A, top panel). This result was confirmed by blotting with an antibody specifically recognizing the phosphorylated EGF receptor (Fig. 7A, bottom panel) and by immunoprecipitating tyrosine-phosphorylated proteins followed by blotting with the anti-phospho-EGF receptor antibody (Fig. 7B). In order to exclude any role for the EGF receptor, we treated thrombin-stimulated cells with AG1478, a specific inhibitor of the EGF receptor kinase. This did not modify TFPI-2 expression, although it efficiently blocked EGF-induced receptor autophosphorylation (Fig. 7C). An EGF receptor blocking antibody did not impair the effect of thrombin on TFPI-2 expression (Fig. 7C). These results suggest that transactivation of the EGF receptor is unlikely to explain the results obtained in our model. Although thrombin clearly up-regulates TFPI-2 mRNA and protein expression, the resulting functional effect could be annihilated if TFPI-2 proved to be a substrate for the proteolytic activity of thrombin. This was tested in a functional assay measuring the ability of recombinant TFPI-2 to inhibit plasmin activity toward the colorimetric substrate S2251. As shown in Fig. 8, recombinant TFPI-2 readily inhibited plasmin, and this effect was not decreased by previous treatment with thrombin. Furthermore, when thrombin-treated TFPI-2 was analyzed by Western blot, no difference was found by comparison with native TFPI-2 (not shown). Thrombin generation leads to clot formation. This important homeostatic function is tightly regulated in order to avoid excess clotting. In this article, we demonstrate a new mechanism potentially down-regulating this pathway, i.e. up-regulation by thrombin of the expression of TFPI-2. TFPI-2 inhibits the formation of factor Xa by the tissue factor-factor VIIa complex. This will result in decreased de novo thrombin generation from prothrombin. We found that thrombin increased the expression of TFPI-2 transcripts in human liver myofibroblasts as well as in the unrelated cell line ECV 304. Thrombin also increased the deposition of TFPI-2 in the extracellular matrix, as judged by Western blot. Most but not all thrombin effects are mediated through PAR receptors (28Sower L.E. Payne D.A. Meyers R. Carney D.H. Exp. Cell Res. 1999; 247: 422-431Crossref PubMed Scopus (39) Google Scholar). In our case, the effect of thrombin appeared secondary to its action on PAR-type receptors since the thrombin effect was blocked by hirudin, which specifically inhibits its catalytic activity. Both PAR-1 and PAR-4 are directly involved in thrombin signalization, whereas PAR-3 seems to act as a co-receptor for PAR-4 (2Nakanishi-Matsui M. Zheng Y. Sulciner D. Weiss E. Ludeman M. Coughlin S. Nature. 2000; 404: 609-613Crossref PubMed Scopus (470) Google Scholar). We found that the effect of thrombin could be mimicked by both PAR-1 and PAR-4 agonistic peptides. In order to know whether PAR-1 and PAR-4 were effectively used by thrombin, we incubated the cells with a PAR-1 blocking antibody. This antibody significantly blocked the thrombin effect suggesting that PAR-1 was indeed involved. However, the inhibition was not complete, which could indicate that PAR-4 also participates in this effect. Analysis of the signaling pathway showed that the thrombin effect on TFPI-2 expression could be blocked by an antagonist of MAPK activation. However, in contradiction with several published studies, thrombin failed to transactivate the EGF receptor in human liver myofibroblasts. This was shown by the lack of detectable tyrosine phosphorylation of the EGF receptor following thrombin addition. Furthermore, thrombin-induced up-regulation of TFPI-2 expression was unaffected by the EGF receptor kinase-specific inhibitor AG1478 or by an EGF receptor blocking antibody. Transactivation of the EGF receptor by thrombin appears to be a general finding since it has been observed in many different cell types (11Daub H. Weiss F.U. Wallasch C. Ullrich A. Nature. 1996; 379: 557-560Crossref PubMed Scopus (1329) Google Scholar, 12Prenzel N. Zwick E. Daub H. Leserer M. Abraham R. Wallasch C. Ullrich A. Nature. 1999; 402: 884-888Crossref PubMed Scopus (1501) Google Scholar, 13Kalmes A. Vesti B.R. Daum G. Abraham J.A. Clowes A.W. Circ. Res. 2000; 87: 92-98Crossref PubMed Scopus (167) Google Scholar, 29Daub H. Wallasch C. Lankenau A. Herrlich A. Ullrich A. EMBO J. 1997; 16: 7032-7044Crossref PubMed Scopus (588) Google Scholar), including smooth muscle cells that are closely related to myofibroblasts. It has been shown that EGF receptor transactivation by thrombin involves the metalloproteinase-mediated shedding of heparin-binding EGF from the cell membrane (12Prenzel N. Zwick E. Daub H. Leserer M. Abraham R. Wallasch C. Ullrich A. Nature. 1999; 402: 884-888Crossref PubMed Scopus (1501) Google Scholar). By using RT-PCR, we have found that liver myofibroblasts express HB-EGF (data not shown), and expression of the required proteinase, presumably ADAM-12 (30Asakura M. Kitakaze M. Takashima S. Liao Y. Ishikura F. Yoshinaka T. Ohmoto H. Node K. Yoshino K. Ishiguro H. Asanuma H. Sanada S. Matsumura Y. Takeda H. Beppu S. Tada M. Hori M. Higashiyama S. Nat. Med. 2002; 8: 35-40Crossref PubMed Scopus (641) Google Scholar), has also been demonstrated (31Le Pabic H. Bonnier D. Wewer U.M. Coutand A. Musso O. Baffet G. Clement B. Theret N. Hepatology. 2003; 37: 1056-1066Crossref PubMed Scopus (184) Google Scholar). Thus, the reason for the lack of EGF receptor transactivation by thrombin in human liver myofibroblasts is unclear. On the other hand, induction of TFPI-2 by thrombin could be blocked by the selective COX-2 inhibitor NS-398, thus suggesting that prostaglandins were involved. We show indeed that this inhibition could be reverted by adding prostaglandin E2 to the cells. COX-2 has been shown previously (3Mallat A. Gallois C. Tao J. Habib A. Maclouf J. Mavier P. Préaux A.M. Lotersztajn S. J. Biol. Chem. 1998; 273: 27300-27305Abstract Full Text Full Text PDF PubMed Scopus (97) Google Scholar) to mediate an anti-proliferative pathway of thrombin for liver myofibroblasts and thus appears as a major mediator of thrombin signaling in these cells. Besides its inhibitory effect on the procoagulant function of thrombin, the increased expression of TFPI-2 could also contribute to the fine-tuning of other thrombin functions, such as its role in extracellular matrix remodeling. Indeed, thrombin up-regulates the expression of several matrix metalloproteinases (32Duhamel-Clerin E. Orvain C. Lanza F. Cazenave J.P. Klein-Soyer C. Arterioscler. Thromb. Vasc. Biol. 1997; 17: 1931-1938Crossref PubMed Scopus (85) Google Scholar, 33Raza S.L. Nehring L.C. Shapiro S.D. Cornelius L.A. J. Biol. Chem. 2000; 275: 41243-41250Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar) and also promotes the activation of pro-matrix metalloproteinase-2 (34Galis Z.S. Kranzhofer R. Fenton J.W. Libby P. Arterioscler. Thromb. Vasc. Biol. 1997; 17: 483-489Crossref PubMed Scopus (123) Google Scholar, 35Zucker S. Conner C. DiMassmo B.I. Ende H. Drews M. Seiki M. Bahou W.F. J. Biol. Chem. 1995; 270: 23730-23738Abstract Full Text Full Text PDF PubMed Scopus (169) Google Scholar). As TFPI-2 can inhibit the activation of pro-matrix metalloproteinases (36Rao C.N. Mohanam S. Puppala A. Rao J.S. Biochem. Biophys. Res. Commun. 1999; 255: 94-98Crossref PubMed Scopus (87) Google Scholar), it could regulate the extent of matrix degradation stimulated by thrombin. Finally, depending on the cellular setting, TFPI-2 has been shown to either promote (17Neaud V. Hisaka T. Monvoisin A. Bedin C. Balabaud C. Foster D.C. Desmoulière A. Kisiel W. Rosenbaum J. J. Biol. Chem. 2000; 275: 35565-35569Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar) or inhibit tumoral cell invasion (37Rao C.N. Cook B. Liu Y. Chilukuri C. Stack M.S. Foster D.C. Kisiel W. Woodley D.T. Int. J. Cancer. 1998; 76: 749-756Crossref PubMed Scopus (73) Google Scholar, 38Konduri S.D. Rao C.N. Chandrasekar N. Tasiou A. Mohanam S. Kin Y. Lakka S.S. Dinh D. Olivero W.C. Gujrati M. Foster D.C. Kisiel W. Rao J.S. Oncogene. 2001; 20: 6938-6945Crossref PubMed Scopus (72) Google Scholar, 39Chand H.S. Du X. Ma D. Inzunza H.D. Kamei S. Foster D. Brodie S. Kisiel W. Blood. 2003; (in press)PubMed Google Scholar). Thrombin generation is a frequent event in cancers and could thus influence tumor invasion via TFPI-2 synthesis. We thank Charles Balabaud, Paulette BioulacSage, and Jean Saric for help in supplying human liver tissue.

Referência(s)