Vascular Endothelial Growth Factor Production by Fibroblasts in Response to Factor VIIa Binding to Tissue Factor Involves Thrombin and Factor Xa
Abstract—Tissue factor (TF) assembled with activated factor VII (FVIIa) initiates the coagulation cascade. We recently showed that TF was essential for FVIIa-induced vascular endothelial growth factor (VEGF) production by human fibroblasts. We investigated whether this production resulted from TF activation by its binding to FVIIa or from the production of clotting factors activated downstream. Incubation of fibroblasts with a plasma-derived FVIIa concentrate induced the generation of activated factor X (FXa) and thrombin and the secretion of VEGF, which was inhibited by hirudin and FXa inhibitors. By contrast, the addition of recombinant FVIIa to fibroblasts did not induce VEGF secretion unless factor X was present. Moreover, thrombin and FXa induced VEGF secretion and VEGF mRNA accumulation, which were blocked by hirudin and FXa inhibitors, respectively. The effect of thrombin was mediated by its specific receptor, protease-activated receptor-1; in contrast, the effect of FXa did not appear to involve effector cell protease receptor-1, because it was not affected by an anti–effector cell protease receptor-1 antibody. An increase in intracellular calcium with the calcium ionophore A23187 or intracellular calcium chelation by BAPTA-AM had no effect on either basal or FXa-induced VEGF secretion, suggesting that the calcium signaling pathway was not sufficient to induce VEGF secretion. Finally, FVIIa, by itself, had no effect on mitogen-activated protein (MAP) kinase activation, contrary to thrombin and FXa, which activate the p44/p42 MAP kinase pathway, as shown by the blocking effect of PD 98059 and by Western blotting of activated MAP kinases. These findings indicate that FVIIa protease induction of VEGF expression is mediated by thrombin and FXa generated in response to FVIIa binding to TF-expressing fibroblasts; they also exclude a direct signaling involving MAP kinase activation via the intracellular domain of TF when expressed by these cells.
- Received September 17, 1999.
- Accepted November 11, 1999.
Tissue factor (TF), the cell-surface receptor for activated factor VII (FVIIa), is the primary regulator of blood coagulation. The TF-FVIIa complex cleaves and activates factors IX and X into factors IXa and factor Xa (FXa), respectively, which lead to thrombin generation.1 Beyond its role as a procoagulant activator, TF participates in other cellular processes, including metastasis,2 tumor-associated angiogenesis,3 and embryogenesis.4 Accordingly, several cell functions are modified in response to FVIIa binding to TF, its receptor. These include intracellular signaling,5 activation of the p44/42 mitogen-activated protein (MAP) kinase pathway,6 induction of tyrosine phosphorylation in monocytes,7 upregulation of poly(A) polymerase,8 cell spreading and phosphorylation of focal adhesion kinase,9 enhanced expression of urokinase receptor by pancreatic cancer cell lines,10 and vascular endothelial growth factor (VEGF) production by human fibroblasts.11 The respective roles of the procoagulant activity of the TF-FVIIa complex and of intracellular signals transmitted by the intracellular domain of TF are not fully understood. In most of the above-mentioned studies, the reported effect required catalytically active FVIIa to bind to TF. The TF-FVIIa complex may act by activating clotting factors downstream in the coagulation cascade; indeed, Fisher et al2 showed that TF-initiated thrombin generation activates signaling via the thrombin receptor on malignant melanoma cells. In other studies, an FVIIa-specific reaction independent of activated clotting factors appeared to be involved.6 10
VEGF is a key regulator of angiogenesis, stimulating endothelial cell proliferation and migration and increasing their permeability. We recently showed that TF was essential for the signaling events leading to VEGF synthesis by human lung fibroblasts11 in response to a plasma-derived human FVIIa concentrate (ACSET, LFB) and that this effect was mostly dependent on the proteolytic activity of the TF-FVIIa complex. It was unclear whether this was a direct effect of FVIIa protease or an indirect effect involving production of activated clotting factors, such as thrombin or FXa. We now report that FXa and thrombin generation are involved in TF-FVIIa–dependent VEGF production by human fibroblasts.
Recombinant human FVIIa12 and Phe-Phe-Arg chloromethylketone–inactivated FVIIa (FVIIai)13 were kindly provided by Novo Nordisk (Gentofte, Denmark). Recombinant tick anticoagulant peptide (TAP) and the nematode anticoagulant peptides (NAPs) 5 and c2 were generously provided by Dr G.P. Vlasuk (Corvas Inc, San Diego, Calif). TAP and NAP5 are potent and specific inhibitors of FXa and prothrombinase activity; NAPc2 is a potent and specific inhibitor of the TF-FVIIa complex, which requires prior interaction with FXa.14 Recombinant hirudin was a gift from M. Lenoble (Hoechst, Paris, France). A pool of neutralizing anti-TF antibodies (TF8-5G9, TF8-6B4, and TF9-9C3) was generously provided by T.S. Edgington and N. Mackman (The Scripps Research Institute, La Jolla, Calif).15 Human FXa was purchased from Enzyme Research Laboratories. An antibody to the effector protease receptor-1 (EPR-1), the first-generation monoclonal antibody B6,16 was generously given by Dr D.C. Altieri (Yale University School of Medicine, New Haven, Conn). The interepidermal growth factor sequence L83GTRKL88(G) and its control scrambled variant K83FTGRLL88 were synthesized by Neosystem. Human thrombin was purchased from Sigma, and the thrombin receptor agonist peptide (TRAP, H-Ser-Phe-Leu-Leu-Arg-Asn-Pro-Asn-Asp-Lys-Tyr-Glu-Pro-Phe-OH) was from Calbiochem-Novabiochem. Chromogenic substrates S-2765 and S-2238 were purchased from Chromogenix. DX9065a was kindly given by Daichi Pharmaceuticals Co Ltd (Tokyo, Japan).
ACSET was prepared as previously described.11 The ACSET preparation contains 30% FVIIa and 70% factor VII (FVII); contaminating proteins include protein S, prothrombin, and factor X (FX).
Human FX Isolation
The FX-containing fraction was obtained during the isolation of factor IX (FIX) and corresponded to the 0.28 mmol/L NaCl eluate,17 which was applied onto a ceramic hydroxyapatite column (Bio-Rad) equilibrated with 10 mmol/L K2HPO4 and 75 mmol/L NaCl (pH 8). After a wash with 30 mmol/L K2HPO4 and 0.25 mmol/L NaCl (pH 8), FX was eluted with 0.5 mmol/L K2HPO4 (pH 8) and then dialyzed overnight against 20 mmol/L Tris and 0.15 mmol/L NaCl (pH 7) at 4°C. To remove prothrombin, the FX-containing fraction was adsorbed to a chelating Sepharose Fast Flow column saturated with Cu2+ and equilibrated with 10 mmol/L K2HPO4, 10 mmol/L sodium citrate, and 0.5 mmol/L NaCl (pH 6.5). After a wash with the equilibration buffer and then with a solution containing 10 mmol/L sodium citrate, 10 mmol/L potassium phosphate, and 0.1 mmol/L sodium chloride (pH 7), FX was eluted with the same buffer containing 5 mmol/L glycine. FX exhibited a specific activity of 130 IU/mg protein and appeared as 1 protein band of 57 kDa after 12% SDS-PAGE. FVII:Ag and FIX:Ag, evaluated by ELISA (Diagnostica Stago), were <1 ng/mL and <0.3 μg/mL, respectively. Furthermore, this fraction had no amidolytic activity on S-2765 and S-2238 chromogenic substrates, demonstrating the absence of contaminating FXa and thrombin, respectively.
Cells and Culture
Human lung fibroblasts (CCD-11Lu) were obtained from the American Type Culture Collection and grown as described.11 Subconfluent fibroblasts were starved for 24 hours in DMEM without FCS and subsequently incubated with human FVIIa in DMEM supplemented with 6 mmol/L CaCl2. Cell viability, monitored by lactate dehydrogenase release, was not altered by any of the experimental conditions (not shown).
Estimation of Cellularity
To avoid changes in VEGF production due to differences in cell proliferation according to the experimental conditions, a colorimetric method was used to count cells present in the well at the time of the assay, as previously described.11
FXa and Thrombin Generation Assays
Thrombin and FXa generated in the culture medium of ACSET-treated fibroblasts were quantified by hydrolysis of the thrombin- and FXa-sensitive chromogenic substrates S-2238 and S-2765, respectively. Changes in absorbance at 405 nm (ΔA405/min) were measured for 30 minutes at 37°C on an ELISA plate reader (Molecular Devices). A calibration curve was constructed by using serial dilutions of thrombin (Sigma) or FXa (ERL).
Human VEGF Immunoassay
Human VEGF concentrations in fibroblast culture supernatants were determined by using the Quantikine human VEGF kit (R&D Systems Europe).11
Total RNA was extracted from fibroblast cultures by using TRIzol Reagent (Life Technologies). A semiquantitative competitive reverse transcription (RT)–polymerase chain reaction (PCR) method was used as described elsewhere.11 A PCR MIMIC construction kit (Clontech Laboratories) was used to correct for variations in amplification efficiency in each reaction and to calculate relative changes in mRNA levels. The density of each band was normalized to the density of the mimic band and plotted in arbitrary units.
Measurement of Intracellular Calcium Increase
Fibroblasts cultured as described above were detached with nonenzymatic cell dissociation solution (Sigma), scraped from the flasks, and centrifuged (400g for 10 minutes). Cells were then suspended in culture medium containing 5 μmol/L fura 2-AM and incubated for 10 minutes at 25°C. Thereafter, cells were first washed in culture medium and then in buffer (10 mmol/L HEPES/NaOH [pH 7.4], 137 mmol/L NaCl, 5.4 mmol/L KCl, 0.34 mmol/L Na2HPO4, 0.44 mmol/L KH2PO4, 0.8 mmol/L MgSO4, 5.5 mmol/L glucose, and 4.2 mmol/L NaHCO3) containing 0.1 mmol/L EGTA (to avoid cell adhesion) and kept in the dark at room temperature. Experiments were carried out with constant stirring in a Perkin-Elmer LS50 B spectrofluorometer with the use of ≈300 000 cells in 3-mL fluorescence cuvettes at 37°C. Increasing concentrations of FXa were added to fura 2–loaded fibroblasts and incubated for 1 minute. [Ca2+]i was measured as described by Grynkiewicz et al.18
Western Blot Detection of Active MAP Kinases
After 24-hour FCS starvation, confluent fibroblasts were stimulated with various ligands for the indicated periods of time (1 well of a 6-well plate per experimental point). Cells were then lysed in lysis buffer already described elsewhere,19 and lysates were loaded onto an 8% SDS-polyacrylamide gel and transferred onto a nitrocellulose membrane. The total amount of p44/42 MAP kinases was detected by using a cocktail of extracellular signal–regulated kinase (ERK)-1 and ERK-2 antibodies (Santa Cruz Biotechnology, Inc). Phosphorylated p44/42 MAP kinases were detected by using a phospho-p44/42 MAP kinase monoclonal antibody (Biolabs).
Results are given as mean±SD. Data were compared by ANOVA. Where significant differences were inferred, the sample means were compared by the Fisher protected least significant difference test.
Effect of FVIIa on VEGF Secretion
Incubation of human lung fibroblasts constitutively expressing TF on their surface for 24 hours with 100 nmol/L FVIIa (a plasma-derived FVIIa concentrate provided by ACSET) resulted in a 4-fold induction of VEGF secretion (from 98±29 with control cells to 391±145 pg/mL, P<0.0001), confirming our previous results11 (Figure 1A⇓).
FXa and Thrombin Generation in Culture Medium of Fibroblasts Incubated With ACSET
Because the ACSET preparation contained significant amounts of prothrombin and FX, we first assessed whether thrombin and FXa were generated during its incubation with fibroblasts. As shown in Figure 1B⇑, thrombin and FXa activities were both detected as early as 1 minute after incubation of the cells with 100 nmol/L ACSET. Thrombin and FXa levels increased gradually, reaching 4.4±1.6 U/mL and 49±22 nmol/L (n=3), respectively, after 24 hours of incubation. In the presence of hirudin (10 U/mL), the ACSET-induced thrombin activity was almost completely eliminated (residual level 0.15±0.08 U/mL at 24 hours). The FXa inhibitor TAP was also able to prevent 95% of the thrombin generation. In the same way, ACSET-induced FXa activity was largely inhibited in the presence of TAP at 50 μg/mL (residual level 1±0.5 nmol/L at 24 hours) but not by hirudin. Inhibition of thrombin and FXa activities by hirudin and TAP, respectively, occurred after as little as 1 minute of incubation (not shown). Unstimulated cells generated insignificant levels of thrombin and FXa (0.04±0.02 U/mL and 0.8±0.5 nmol/L, respectively, after 24 hours).
Effect of Hirudin and FXa Inhibitors on ACSET-Induced VEGF Secretion
To determine the role of thrombin and FXa in ACSET-induced VEGF production, we tested the effect of hirudin (the strongest known naturally occurring inhibitor of thrombin) and that of TAP and DX9065a (specific and direct FXa inhibitors). ACSET-induced VEGF secretion was inhibited by 55±25% by 10 U/mL hirudin, pointing to an important role of thrombin in this effect (Figure 1A⇑). TAP and DX9065 caused 57±23% and 48±13% inhibition, respectively. Interestingly, the simultaneous addition of TAP and hirudin or of DX9065a and hirudin blocked 79±15% and 83±9% of VEGF secretion, respectively, in response to ACSET, suggesting that the bulk of the effect of ACSET was mediated by thrombin and FXa. DX9065a and TAP or hirudin had no effect on unstimulated cells; the baseline VEGF secretion is not modified in the presence of these FXa and thrombin inhibitors.
Thrombin Induces VEGF Secretion by Fibroblasts
As shown in Figure 2⇓, incubation of cells with purified thrombin (0.5 to 10 U/mL) induced a dose-dependent increase in VEGF secretion (r=0.87, P<0.01), which was maximal after 12 hours. Stimulation of the cells with 10 U/mL thrombin resulted in a 3.3-fold increase in VEGF secretion relative to baseline (P<0.001). The thrombin receptor agonist TRAP also induced a dose-dependent increase in VEGF secretion, with a 2-fold rise at 100 μmol/L (P<0.001). Hirudin (100 U/mL) almost completely prevented VEGF secretion in response to 10 U/mL thrombin (−93±13%, data not shown). These results show that thrombin is an agonist of VEGF production and suggest that the effect occurs through protease-activated receptor-1, a specific thrombin receptor.
FXa-Induced VEGF Secretion by Fibroblasts
We then investigated whether FXa alone was able to induce VEGF production. FXa (22.8 to 228 nmol/L) incubated with fibroblasts for 24 hours induced dose-dependent VEGF secretion, which was significant relative to baseline from 57 nmol/L FXa (P<0.05) and reached 2.5-fold at 228 nmol/L FXa (Figure 3A⇓). The FXa inhibitors TAP and DX9065a prevented VEGF secretion in response to 114 nmol/L FXa. TAP (50 μg/mL) and DX9065a (10 μmol/L) inhibited 100% (P<0.001) and 63±24% (P<0.01) of this effect, respectively. VEGF secretion was not inhibited by hirudin or by anti-TF antibodies (Figure 3B⇓). VEGF secretion induced by FXa was additive with VEGF secretion induced by thrombin (insert, Figure 3A⇓).
To characterize the interaction between FXa and fibroblasts, we first investigated whether FXa was able to induce a cellular signal. As shown in Figure 4⇓, treatment of fura 2–loaded fibroblasts with increasing concentrations of FXa (1 to 1000 nmol/L) was associated with a concentration-dependent [Ca2+]i response. FXa (100 nmol/L) induced a 50% [Ca2+]i increase over baseline, an effect that was inhibited in a concentration-dependent manner by DX9065a, with an IC50 of 120 nmol/L. Interestingly, preincubation of fibroblasts for 30 minutes with an antibody (EPR-1 monoclonal antibody B6) directed against EPR-1, a reported receptor for FXa, at concentrations of 50, 100, and 300 μg/mL inhibited the [Ca2+]i response induced by 100 nmol/L FXa by 18%, 47%, and 76%, respectively (2 experiments performed in triplicate, data not shown). These results strongly suggest that fibroblasts express EPR-1. Notably, preincubation of the cells for 1 hour with B6 at a concentration of 300 μg/mL did not influence the effect of thrombin (100 nmol/L) on [Ca2+]i, excluding a nonspecific effect of the antibody on [Ca2+]i. However, preincubation of the cells for 1 hour with B6 at concentrations up to 300 μg/mL (not shown) did not affect FXa-induced VEGF production even when the incubation time with FXa was reduced to 6 hours (instead of 24 hours). The use of B6 alone had no effect on basal VEGF secretion. Moreover, the FX peptide Leu83-Leu88, representing the interepidermal growth factor sequence in FXa that mediates ligand binding to EPR-1, neither induced VEGF secretion nor inhibited the FXa-induced VEGF secretion.
We also asked whether an increase in [Ca2+]i, as seen with the FXa binding to fibroblasts, was sufficient to induce VEGF secretion. Therefore, we used the calcium ionophore A 23187 (Sigma) to mimic an intracellular calcium increase. We were unable to induce VEGF secretion with this ionophore in concentrations up to 10 μmol/L. Moreover, an intracellular calcium chelation agent, BAPTA-AM (up to 60 μmol/L), did not prevent the FXa- or the thrombin-induced VEGF secretion (data not shown). From these data, we conclude that an increase in [Ca2+]i is not sufficient to induce VEGF secretion. Taken together, these results strongly suggest that EPR-1, an FXa receptor, does not appear to be involved in VEGF production in response to FXa.
Effect of FX Plus Recombinant FVIIa on VEGF Secretion
We then determined the effect of FVIIa and FX in combination (Figure 5⇓). Incubation of fibroblasts with 100 nmol/L FVIIa and 90 nmol/L FX increased VEGF secretion to 2.3-fold baseline (P<0.0001). By contrast, incubation of fibroblasts with 100 nmol/L recombinant FVIIa for 24 hours did not induce VEGF secretion (Figure 5⇓). The FVIIa+FX-induced VEGF secretion was associated with FXa generation (shown by hydrolysis of the specific FXa chromogenic substrate S-2765) but not with thrombin generation (no hydrolysis of the specific thrombin chromogenic substrate S-2238 after 24-hour incubation of the cells with FVIIa-FX, data not shown). Furthermore, FVIIa+FX-induced VEGF production was abolished by inhibitors of FXa (TAP, NAP5, and NAPc2) used at a concentration of 50 μg/mL, by DX9065a at 10 μmol/L, and by anti-TF-antibodies (81±13% inhibition, Figure 5⇓). Hirudin had no significant effect. The use of FVIIai in combination with FX did not induce VEGF secretion, indicating that the proteolytic activity of FVIIa was necessary, through the generation of FXa, to induce VEGF expression. The combination of FVIIa with purified FIX had no effect (data not shown).
FXa and Thrombin Induce VEGF mRNA Accumulation
To identify the level of action of thrombin, FXa, and FVIIa in combination with FX, VEGF mRNA was studied after RT and amplification by PCR. Three VEGF transcripts of 180, 312, and 384 bp coding for VEGF121, VEGF165, and VEGF189, respectively, were detected. After 24 hours of treatment with 100 nmol/L FVIIa in combination with 100 nmol/L FX, the 180-bp transcript showed a 2±0.3-fold increase (P<0.001) over baseline (unstimulated cells, Figure 6⇓). The levels of the 312-bp and 384-bp transcripts increased similarly. A 1.9±0.3-fold increase (P<0.01) in the 180-bp transcript level was observed after FXa (114 nmol/L) stimulation, and a 1.6±0.27-fold increase (P=0.02) was seen after thrombin (1 U/mL) treatment. Preincubation of the cells with TAP (50 μg/mL) completely inhibited the VEGF mRNA accumulation induced by FVIIa-FX and by FXa (P<0.0001 and P<0.05, respectively). Hirudin (10 U/mL) prevented thrombin-induced VEGF mRNA accumulation (P<0.05).
Activation of MAP Kinases ERK-1 and ERK-2 by Thrombin and FXa but Not by FVIIa
Because binding of FVIIa to cell surface TF has been shown to activate signal transduction via p44/42 MAP kinases,6 we analyzed the contribution of this pathway by using a specific antibody against the phosphorylated Thr202/Tyr204 residues of these kinases. As shown in Figure 7A⇓, exposure of the cells to 100 nmol/L of FVIIa or 100 nmol/L of FVIIai did not alter the phosphorylated p44/42 band. By contrast, thrombin and FXa transiently increased the phosphorylation of p44/42 MAP kinases, peaking at ≈2′ and 10′, respectively, whereas the amount of total MAP kinases remained essentially constant (Figure 7B⇓).
Effect of MAP Kinase Inhibitors on FXa- and Thrombin-Induced VEGF Production
To confirm the involvement of thrombin- and FXa-activated MAP kinases in VEGF expression, we used PD 98059 to selectively block the activation of p44/42 MAP kinases. PD 98059 (50 μmol/L) was preincubated with fibroblasts for 30 minutes before 24 hours of cell stimulation with either 100 nmol/L FXa or 1 U/mL thrombin (Figure 7C⇑). PD 98059 inhibited 67±52% (P<0.05) of FXa-induced VEGF production and 85±19% (P<0.0001) of thrombin-induced VEGF production, pointing to a role of the p44/42 MAP kinase signaling pathway in this production. SB 203580 (10 μmol/L),a highly specific inhibitor of MAP kinase p38, had no significant effect on VEGF production induced either by FXa or by thrombin.
We first confirmed our previous report11 that ACSET, a plasma-derived FVIIa concentrate, induces VEGF production by TF-expressing human fibroblasts. We further showed that its use was associated with significant generation of FXa and thrombin in the culture medium and that the increased VEGF secretion was largely dependent on FXa and thrombin inasmuch as it was inhibited by 80% by the simultaneous addition of hirudin and TAP or DX9065a. This small residual VEGF secretion could result from an incomplete inhibition of thrombin and FXa generated in the presence of ACSET. Indeed, the addition of recombinant FVIIa to TF-expressing fibroblasts did not induce VEGF production unless FX was added simultaneously to allow FXa generation. The lack of effect of recombinant FVIIa alone excluded a direct FVIIa protease signaling mechanism. We confirmed that inactivated FVIIa did not enable VEGF production to occur, even in combination with factor X. These results show for the first time that the TF-FVIIa complex contributes to VEGF production by human fibroblasts through FXa and thrombin generation.
Confirmation of the direct effect of thrombin and FXa on VEGF production was obtained by using purified enzymes. The effect of thrombin was significant from 0.5 U/mL, a concentration that may be reached locally in vivo, and was reproduced by the human selective peptide agonist of protease-activated receptor-1, a thrombin receptor expressed by fibroblasts.20 Möhle et al21 recently reported thrombin-induced release of VEGF by human megakaryocytes and platelets. We also found that FXa induced VEGF secretion. The FXa-induced increase in VEGF was blocked by TAP and DX9065a, clearly showing that its proteolytic activity was required. This effect was not mediated by contaminating or generated thrombin, because hirudin failed to block FXa-induced VEGF production while completely preventing the thrombin-induced increase in VEGF. In a previous study, we showed that ACSET induced VEGF production at the mRNA level.11 In the present study, we found that FXa and thrombin induced VEGF mRNA accumulation, an effect that was prevented by TAP and hirudin, respectively. Similar results were obtained by using FVIIa and factor X simultaneously.
Recently, FXa has been shown to mediate a variety of biological effects (in addition to its role in the activation of coagulation), including mitogenesis, lymphocyte activation, cytokine secretion, and adherence molecule expression.22 23 Our present results suggest that FXa not only plays a role in inflammation but may also be involved in angiogenesis. We show that EPR-1, a receptor that binds FXa and mediates several of its cellular effects,24 25 is expressed on fibroblasts, because FXa induced a calcium signal that was prevented by B6, a monoclonal antibody blocking the binding of FXa to EPR-1. B6 antibody, however, was unable to block VEGF secretion even at very high concentrations or after a short incubation time (6 hours instead of 24 hours). We were also unsuccessful either in activating directly VEGF secretion or in inhibiting FXa-induced VEGF secretion by the interepidermal growth factor peptide (Leu83-Leu88) that mediates EPR-1 recognition of FXa. Furthermore, the increase in [Ca2+]i observed when FXa binds to fibroblasts was not sufficient to induce VEGF secretion because the calcium ionophore A23187 failed to do it and because intracellular calcium chelation by BAPTA-AM was also without any effect on FXa-induced VEGF secretion. This suggests that EPR-1 is not involved directly in FXa-induced VEGF production and points to the involvement of another receptor, alone or in association with EPR-1, that could be activated by FXa.
Because binding of FVIIa to cell surface TF has been shown to activate signal transduction via p44/42 MAP kinases,6 we looked at direct FVIIa p44/42 MAP kinase activation with specific antibodies. In our cellular model, the p44/42 MAP kinase pathway was not significantly activated by FVIIa. Thrombin and FXa, by contrast, induced p44/42 MAP kinase activation, which was maximal at 2 and 10 minutes, respectively. PD 98059, a specific inhibitor of MEK1/2,26 which blocks the p44/42 MAP kinase pathway, prevented 85% of the effect of thrombin and 50% of the effect of FXa on VEGF production, whereas SB 203580, an inhibitor of p38 MAP kinase, had no effect. These results show for the first time that FXa induces the p44/42 MAP kinase activation that leads to VEGF production. Investigations are in progress to clarify the complete pathway from the cellular system receptor to the VEGF gene in human fibroblasts. They are in keeping with previous work showing that thrombin induces strong MAP kinase activation in mouse lung fibroblasts20 and that the p44/42 MAP kinase module plays a key role in transcriptional regulation of the VEGF gene.27 The present results suggest that in human lung fibroblasts, direct activation of the p44/42 MAP kinase pathway triggered by the binding of FVIIa to TF does not occur, unlike the results obtained by Poulsen et al6 in transfected BHK cells. In the present study, although the proteolytic activity of FVIIa was absolutely required, FVIIa-TF signaling was not quenched by TAP or hirudin and was independent of the cytoplasmic domain of TF because cells transfected with a cytoplasmic domain–deleted version of TF also supported FVIIa-induced MAP kinase activation.28
VEGF is a direct angiogenic agent in normal and abnormal physiological conditions. Angiogenesis is a crucial component of tumor growth and metastasis, and VEGF mRNA is markedly upregulated in the majority of human tumors.29 Interestingly, in a recent study, Fukumura et al30 suggested an important contribution of stromal cells of the tumor microenvironment to tumor angiogenesis, demonstrating the activation of the VEGF promoter in fibroblasts of the fibrotic tumor matrix. Our results contribute to clarifying the role of TF-expressing cells in tumor angiogenesis and provide a molecular explanation for reports in which blocking of the coagulation pathways at the level of TF, FXa, or thrombin inhibited hematogenous metastasis in SCID mice.2 These results do not exclude the contribution of functions of the TF cytoplasmic domain, as recently suggested by others.31 32 Recently, Abe et al33 demonstrated a significant correlation between TF and VEGF production in human melanoma cells lines with tumor angiogenesis in vivo. However, in this model, the TF procoagulant function and the proteolytic activity of bound FVIIa were not required for VEGF production, suggesting that TF-mediated VEGF synthesis in cancer cells was caused by endogenous stimuli that transduce a signal via phosphorylation of the cytoplasmic serine residues of TF.
In summary, the present study clearly shows that in human lung fibroblasts, direct FVIIa protease signaling does not contribute to VEGF production, as evidenced by the fact that FVIIa alone is inactive; in these cells, the induction of VEGF by FVIIa is indirectly mediated by FXa and thrombin. As recently suggested,34 the involvement of these proteases allows cell-cell communication via the activation of neighboring cells on local diffusion. Interestingly, in this cell type, FVIIa did not induce MAP kinase activation either.
These results thus support the role of activated clotting factors in angiogenesis. They suggest that VEGF is regulated by FXa and thrombin in normal angiogenesis but that other mechanisms may be used during tumor angiogenesis or in transformed cell lines. They also raise questions as to the other possible consequences of VEGF production by fibroblasts. A recent study clearly demonstrated VEGF mRNA expression in both macrophages and poorly differentiated smooth muscle cells of human coronary atherosclerotic plaques but not in normal coronary arteries.35 Moreover, in a porcine animal model, the response of the coronary artery to balloon angioplasty has been associated with activation of adventitial fibroblasts, which undergo differentiation to myofibroblasts and migrate into the intima.36 This observation raises the possibility that adventitial myofibroblasts contribute to local VEGF production in response to coagulation proteases. VEGF may contribute to atheromatous lesion progression by enhancing neovascularization within the lesion. Alternatively, and as suggested by initial clinical results of gene therapy,37 VEGF may have a beneficial role in reendothelialization and in improving collateral blood flow.38 39 The overall effect of VEGF production mediated by TF, FXa, and thrombin in atherosclerotic plaque remains to be established.
We thank Dr Martine Jandrot-Perrus (Faculté Xavier Bichat, Paris, France) and Dr Nigel Mackman (The Scripps Research Institute, La Jolla, Calif) for fruitful scientific discussions and critical reading of the manuscript. We also thank Dr Jamel El Benna and Cedric Dewas for technical advice concerning the MAP kinase Western blotting.
Fischer EG, Ruf W, Mueller BM. Tissue factor-initiated thrombin generation activates the signaling thrombin receptor on malignant melanoma cells. Cancer Res. 1995;55:1629–1632.
Zhang Y, Deng Y, Luther T, Muller M, Ziegler R, Waldherr R, Stern DM, Nawroth PP. Tissue factor controls the balance of angiogenic and antiangiogenic properties of tumor cells in mice. J Clin Invest. 1994;94:1320–1327.
Rottingen JA, Enden T, Camerer E, Iversen JG, Prydz H. Binding of human factor VIIa to tissue factor induces cytosolic Ca2+ signals in J82 cells, transfected COS-1 cells, Madin-Darby canine kidney cells and in human endothelial cells induced to synthesize tissue factor. J Biol Chem. 1995;270:4650–4660.
Poulsen LK, Jacobsen N, Sorensen BB, Bergenhem NCH, Kelly JD, Foster DC, Thastrup O, Ezban M, Petersen LC. Signal transduction via the mitogen-activated protein kinase pathway induced by binding of coagulation factor VIIa to tissue factor. J Biol Chem. 1998;273:6228–6232.
Pendurthi UR, Alok D, Rao LVM. Binding of factor VIIa to tissue factor induces alterations in gene expression in human fibroblast cells: up-regulation of poly(A) polymerase. Proc Natl Acad Sci U S A. 1997;94:12598–12603.
Ott I, Fischer EG, Miyagi Y, Mueller BM, Ruf W. A role for tissue factor in cell adhesion and migration mediated by interaction with actin-binding protein 280. J Cell Biol. 1998;140:1241–1253.
Taniguchi T, Kakkar AK, Tuddenham EGD, Williamson RCN, Lemoine NR. Enhanced expression of urokinase receptor induced through the tissue factor-factor VIIa pathway in human pancreatic cancer. Cancer Res. 1998;58:4461–4467.
Ollivier V, Bentolila S, Chabbat J, Hakim J, de Prost D. Tissue factor-dependent vascular endothelial growth factor production by human fibroblasts in response to activated factor VII. Blood. 1998;91:2698–2703.
Sørensen BB, Persson E, Freskgård PO, Kjalke M, Ezban M, Williams T, Rao LVM. Incorporation of an active site inhibitor in factor VIIa alters the affinity for tissue factor. J Biol Chem. 1997;272:11863–11868.
Stanssens P, Bergum PW, Gansemans Y, Jespers L, Laroche Y, Huang S, Maki S, Messens J, Lauwereys M, Cappello M, Hotez PJ, Lasters I, Vlasuk GP. Anticoagulant repertoire of the hookworm Ancylostoma caninum. Proc Natl Acad Sci U S A. 1996;93:2149–2154.
Altieri DC, Edgington TS. Identification of effector cell protease receptor-1: a leukocyte-distributed receptor for the serine protease factor Xa. J Immunol. 1990;145:246–253.
Grynkiewicz G, Poenie M, Tsien RJ. A new generation of calcium indicators with greatly improved fluorescence properties. J Biol Chem. 1985;260:3440–3450.
El Benna JE, Dang PM, Gaudry M, Fay M, Morel F, Hakim J, Gougerot-Pocidalo MA. Phosphorylation of the respiratory burst oxidase subunit p67(phox) during human neutrophil activation: regulation by protein kinase C-dependent and independent pathways. J Biol Chem. 1997;272:17204–17208.
Trejo J, Connolly AJ, Coughlin SR. The cloned thrombin receptor is necessary and sufficient for activation of mitogen-activated protein kinase and mitogenesis in mouse lung fibroblasts: loss of responses in fibroblasts from receptor knockout mice. J Biol Chem. 1996;271:21536–21541.
Möhle R, Green D, Moore MAS, Nachman RL, Rafii S. Constitutive production and thrombin-induced release of vascular endothelial growth factor by human megakaryocytes and platelets. Proc Natl Acad Sci U S A. 1997;94:663–668.
Senden NHM, Jeunhomme TMAA, Heemskerk JWM, Wagenvoord R, Vantveer C, Hemker HC, Buurman WA. Factor Xa induces cytokine production and expression of adhesion molecules by human umbilical vein endothelial cells. J Immunol. 1998;161:4318–4324.
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.
Dudley DT, Pang L, Decker SJ, Bridges AJ, Saltiel AR. A synthetic inhibitor of the mitogen-activated protein kinase cascade. Proc Natl Acad Sci U S A. 1995;92:7686–7689.
Milanini J, Vinals F, Pouyssegur J, Pages G. p42/p44 MAP kinase module plays a key role in the transcriptional regulation of the vascular endothelial growth factor gene in fibroblasts. J Biol Chem. 1998;273:18165–18172.
Sorensen BB, Freskgard PO, Nielsen LS, Rao LVM, Ezban M, Petersen LC. Factor VIIa-induced signal transduction via the p44/42 mitogen-activated protein kinase activation requires the proteolytic activity of factor VIIa, and is independent of the tissue factor cytoplasmic domain. J Biol Chem. 1999;274:21349–21354.
Ferrara N. Vascular endothelial growth factor. Eur J Cancer. 1996;32A:2413–2422.
Bromberg ME, Konigsberg WH, Madison JF, Pawashe A, Garen A. Tissue factor promotes melanoma metastasis by a pathway independent of blood coagulation. Proc Natl Acad Sci U S A. 1995;92:8205–8209.
Abe K, Shoji M, Chen J, Bierhaus A, Danave I, Micko C, Casper K, Dillehay DL, Nawroth PP, Rickles FR. Regulation of vascular endothelial growth factor production and angiogenesis by the cytoplasmic tail of tissue factor. Proc Natl Acad Sci U S A. 1999;96:8663–8668.
Inoue M, Hiroshi I, Ueda M, Naruko T, Kojima A, Komatsu R, Doi K, Ogawa Y, Tamura N, Takaya K, Igaki T, Yamashita J, Chun T-H, Masatsugu K, Becker AE, Nakao K. Vascular endothelial growth factor (VEGF) expression in human coronary atherosclerotic lesions. Circulation. 1998;98:2108–2116.
Zalewski SY. Vascular myofibroblasts: lessons from coronary repair and remodeling. Arterioscler Thromb Vasc Biol. 1997;17:417–422.
Losordo DW, Vale PR, Symes JF, Dunnington CH, Esakof DD, Maysky M, Ashare AB, Lathi K, Isner JM. Gene therapy for myocardial angiogenesis: initial clinical results with direct myocardial injection of phVEGF165 as sole therapy for myocardial ischemia. Circulation. 1998;98:2800–2804.
Tsurumi Y, Kearney M, Chen DF, Silver M, Takeshita S, Yang JH, Symes JF, Isner JM. Treatment of acute limb ischemia by intramuscular injection of vascular endothelial growth factor gene. Circulation. 1997;96:382–388.