Vascular Endothelial Growth Factor KDR Receptor Signaling Potentiates Tumor Necrosis Factor-induced Tissue Factor Expression in Endothelial Cells*

Vascular endothelial growth factor (VEGF) and tumor necrosis factor-α (TNF-α) have been shown to synergistically increase tissue factor (TF) expression in endothelial cells; however, the role of the VEGF receptors (KDR, Flt-1, and neuropilin) in this process is unclear. Here we report that VEGF binding to the KDR receptor is necessary and sufficient for the potentiation of TNF-induced TF expression in human umbilical vein endothelial cells. TF expression was evaluated by Western blot analysis and fluorescence-activated cell sorting. In the absence of TNF-α, wild-type VEGF- or KDR receptor-selective variants induced an approximate 7-fold increase in total TF expression. Treatment with TNF alone produced an approximate 110-fold increase in total TF expression, whereas coincubation of TNF-α with wild-type VEGF- or KDR-selective variants resulted in an approximate 250-fold increase in TF expression. VEGF lacking the heparin binding domain was also able to potentiate TF expression, indicating that heparin-sulfate proteoglycan or neuropilin binding is not required for TF up-regulation. Neither placental growth factor nor an Flt-1-selective variant was capable of inducing TF expression in the presence or absence of TNF. Inhibition of protein-tyrosine kinase or protein kinase C activity significantly blocked the TNF/VEGF potentiation of TF up-regulation, whereas phorbol 12-myristate 13-acetate, a protein kinase C activator, increased TF expression. These data demonstrate that KDR receptor signaling governs both VEGF-induced TF expression and the potentiation of TNF-induced up-regulation of TF.

Vascular endothelial growth factor (VEGF) and tumor necrosis factor-␣ (TNF-␣) have been shown to synergistically increase tissue factor (TF) expression in endothelial cells; however, the role of the VEGF receptors (KDR, Flt-1, and neuropilin) in this process is unclear. Here we report that VEGF binding to the KDR receptor is necessary and sufficient for the potentiation of TNF-induced TF expression in human umbilical vein endothelial cells. TF expression was evaluated by Western blot analysis and fluorescence-activated cell sorting. In the absence of TNF-␣, wild-type VEGF-or KDR receptor-selective variants induced an approximate 7-fold increase in total TF expression. Treatment with TNF alone produced an approximate 110-fold increase in total TF expression, whereas coincubation of TNF-␣ with wild-type VEGF-or KDR-selective variants resulted in an approximate 250fold increase in TF expression. VEGF lacking the heparin binding domain was also able to potentiate TF expression, indicating that heparin-sulfate proteoglycan or neuropilin binding is not required for TF up-regulation. Neither placental growth factor nor an Flt-1-selective variant was capable of inducing TF expression in the presence or absence of TNF. Inhibition of proteintyrosine kinase or protein kinase C activity significantly blocked the TNF/VEGF potentiation of TF up-regulation, whereas phorbol 12-myristate 13-acetate, a protein kinase C activator, increased TF expression. These data demonstrate that KDR receptor signaling governs both VEGF-induced TF expression and the potentiation of TNF-induced up-regulation of TF.
Vascular endothelial growth factor (VEGF) 1 is a potent endothelial cell (EC)-specific mitogen that promotes the proliferation and migration of EC, remodeling of the extracellular matrix, formation of capillary tubules, and vascular leakage (1)(2)(3)(4)(5)(6). As a key angiogenic growth factor, VEGF plays a critical role in the development of the fetal cardiovascular system, as well as a significant role in the physiological and pathological angiogenesis (7)(8)(9)(10). Although it is well accepted that Flt-1 and KDR are both high affinity receptors for VEGF, it is not clear which receptor activates the downstream signaling pathways responsible for the diverse biological responses of VEGF. Transgenic knockout studies in mice showed that both receptors are essential for animal survival, because mouse embryos null for KDR or Flt-1 die in utero (11)(12). Experiments using receptor-specific binding variants or receptor-specific inhibitors have associated KDR receptor activity with EC proliferation, migration, vascular permeability, cell survival, and angiogenesis (13)(14)(15)(16)(17). The role of Flt-1 receptor signaling in VEGF biology is less clear. Studies indicate that Flt-1 may mediate, at least in part, chemotaxis and procoagulatant activity in macrophages and up-regulation of matrix metalloproteinases in vascular smooth muscle cells (18 -19). Upon VEGF stimulation, KDR receptor is strongly tyrosine-phosphorylated, but little or no autophosphorylation of Flt-1 occurs (19 -21). Rahimi et al. (19) recently proposed that KDR activation plays a dominant role in angiogenesis by promoting EC proliferation, whereas Flt-1 binding plays a stationary role by antagonizing the interaction of VEGF with KDR. A novel VEGF receptor with a sequence identical to that of neuropilin was recently identified by Soker et al. (22). This receptor binds to VEGF via an interaction with C-terminal heparin binding domain. The role of neuropilin in VEGF signaling is currently under active investigation (22)(23)(24).
VEGF has previously been demonstrated to regulate tissue factor (TF) expression in monocytes and ECs (25)(26)(27). Although VEGF alone induced only a moderate increase in TF expression, it significantly enhanced TNF-␣-induced TF up-regulation through a synergistic mechanism in EC (26 -27). The role that the two TNF receptors, TNFR60 and TNFR80, play in the synergy between VEGF and TNF has been studied. Clauss et al. (27) reported that stimulation of the 60-kDa TNF receptor by a mutant of TNF specific for TNFR60 induced a TF upregulation comparable with wild-type TNF. In contrast, stimulation of TNFR80 by a TNFR80-specific TNF mutant did not enhance TF expression. Thus, TNFR60 is the principle receptor involved in the synergistic up-regulation of TF induced by TNF and VEGF. The role that the VEGF receptors, KDR and Flt-1, play in this process is unknown. Therefore, in the present study we have investigated the role of KDR and Flt-1 receptor signaling in TF up-regulation and in its synergy with TNF by using VEGF receptor-selective variants. VEGF and TNF work in concert to synergistically up-regulate TF expression in human umbilical vein endothelial (HUVEC) cells. Here we demonstrate that KDR receptor signaling is necessary and sufficient for the potentiation of TNF-induced TF up-regulation.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human VEGF 165 (rhVEGF 165 ) was produced in Escherichia coli (Genentech, Inc., South San Francisco, CA). The VEGF 109 , an heparin binding domain (HBD)-deficient variant, was prepared as described previously (28). The VEGF receptor-selective variants were prepared as described by Li et al. (14). Briefly, a Flt-1selective variant (Fltsel ) was created based on a comprehensive mutational analysis of the receptor binding site of VEGF. This variant contains four substitutions (Ile-43 3 Ala, Ile-46 3 Ala, Gln-79 3 Ala, and Ile-83 3 Ala) from the wild-type protein and is heparin binding domain-deficient. Fltsel binds to Flt-1 with wild-type affinity and has 470-fold reduced affinity for KDR binding. In addition, Fltsel had no effect on autophosphorylation of KDR but significantly induced the secretion of matrix metalloprotease-9 in human aorta smooth muscle cells that express Flt-1 only. Two KDR-selective variants (KDR-sel1 and KDR-sel2 ) were created by using a competitive phage display strategy. The KDR-sel1 has the HBD, whereas KDR-sel2 is HBD-deficient. These variants have approximately wild-type affinity for KDR but 2000-fold reduced affinity for binding to Flt-1. KDR variants showed activity comparable with the wild-type VEGF in KDR autophosphorylation and EC proliferation assays but had no effect on the secretion of matrix metalloprotease-9. The heterodimeric form of recombinant human hepatocyte growth factor (HGF) was produced in and isolated from Chinese hamster ovary cells as described previously (29). Recombinant human TNF-␣, fibroblast growth factor basic (FGF), transforming growth factor (TGF)-␤1, TGF-␤2, and epidermal growth factor (EGF) were purchased from R & D Systems (Minneapolis, MN). TNF-␤ and anti-TF antibody (3D1) were from Genentech (South San Francisco, CA). Genistein, staurosporine, phorbol 12-myristate 13-acetate (PMA), PD98059, PP1, and wortmannin were purchased from BIOMOL (Plymouth Meeting, PA). All other chemicals were purchased from Sigma. All reagents were prepared as 1000ϫ stock solutions unless otherwise specified.
Cell Culture and Drug Treatment-HUVEC were purchased from Cell Systems (Kirkland, WA) and maintained as instructed by the manufacturer. Briefly, cells were first grown in a T-75 flask. When HUVEC cells reached 70 -80% confluency, they were subcultured onto 6-well tissue culture plates for Western blot (WB) analysis and fluorescence-activated cell sorting (FACS). All flasks and culture plates were precoated with an attachment factor provided by the manufacturer. HUVEC cells were used between passages 3 and 8. Fresh medium was replaced every 24 h. For drug treatment, subconfluent cells were switched to a GF/serum-free medium overnight and then treated with VEGF or other drugs as specified.
Western Blot Detection of Tissue Factor-The methods for cell lysis and WB are described in detail elsewhere (30). Briefly, cells were lysed in a buffer containing 50 mM Tris, 150 mM NaCl, 1 mM EGTA, 1% Triton X-100, 1% sodium deoxycholate, 1 mM sodium vanadate, 50 mM sodium fluoride, 2 mM EDTA (pH 8.0), 1 mM phenylmethylsulfonyl fluoride, 10 g/ml of leupeptin/pepstatin A/aprotinin for 15 min on ice. Cell lysates were clarified by centrifugation. Protein concentration in the supernatants was determined using a BCA assay (Pierce). An equal amount of protein was denatured and separated using SDS polyacrylamide gel electrophoresis (Novex, San Diego, CA), transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA), and blocked in 5% milk overnight. A monoclonal anti-TF antibody (3D1) was used at a 1:2500 dilution to probe TF protein. A secondary antibody conjugated with horseradish peroxidase (Zymed Laboratories Inc., South San Francisco, CA) and enhanced chemiluminescent kit (Amersham Phamacia Biotech) were used to visualize the TF immunoreactive bands. Multiple exposures of films were obtained to determine the optimal exposure time. In general, the blots for samples that were treated in the presence of TNF-␣ were exposed to x-ray films for ϳ2-5 min, whereas the blots for samples that were treated in the absence of TNF-␣ were exposed to x-ray films for ϳ30 -50 min because of lower TF signal. The protein bands were scanned by a densitometer, and the relative intensities were quantified using ImageQuant software (Molecular Dynamics). The relative TF expression was always normalized to controls (no VEGF or TNF) on the same blot with the same exposure.
FACS Analysis-The treated HUVEC cells were washed with a FACS buffer (phosphate-buffered saline, 0.5% bovine serum albumin) and dissociated with versene (Life Technologies, Inc.) or Accutase (Innovative Cell Technologies, La Jolla, CA) for 10 min at 37°C, harvested, pelleted at 1050 rpm for 5 min, and then resuspended in cold FACS buffer (0.1% bovine serum albumin in phosphate-buffered saline, filtered) at 3-5 million cells/ml. Cells (0.3-0.5 million) were incubated with 167 nM of 3D1 or anti-GP120 for 1 h at 4°C on a shaker. When the incubation was complete, cells were washed twice with FACS buffer and further incubated with 100 nM fluorescein isothiocyanate goat anti-mouse IgG for 1 h at 4°C on a shaker. Cells were washed and resuspended in 500 l of FACS buffer for FACS analysis.
FACS analysis was carried out with a FACS Calibur flow cytometer (Becton Dickinson, Immunocytometry Systems, San Jose, CA). Data were collected in FL-1 with a 530-nm band pass filter. For each analysis, 10,000 events were recorded. The data were gated on the main population identified on the forward scatter versus side scatter dot plot (80 -85% of total events) and analyzed with CellQuest software (Becton Dickinson, Immunocytometry Systems, San Jose, CA) for relative fluorescence intensity.
Statistics Analysis-The results are reported as means Ϯ S.D. Statistical analysis was conducted following comparison of results from the VEGF variants in the absence of TNF to the TNF minus control (no VEGF or TNF); similarly, results obtained in the presence of TNF/ VEGF variants were compared with the TNF plus control (with TNF but no VEGF). An unpaired Student's t test was used to determine significant differences. A value of p Ͻ 0.05 was considered significant.

RESULTS
To investigate the effect of VEGF on TF expression and whether VEGF interacts with TNF in regulating TF expression, subconfluent HUVEC cells were treated with VEGF, TNF-␣, or a combination of VEGF and TNF-␣ for 0 to 9 h. The VEGF concentration was selected based on a dose-response study that showed that 1000 pM of VEGF induced a maximal increase in TF expression (data not shown). The optimal TNF-␣ concentration (5 ng/ml) was selected based on previously reported studies (26 -27). Total TF expression in whole cell lysates was evaluated by WB using a specific anti-human TF antibody. In the absence of TNF-␣, VEGF induced a modest time-dependent increase in TF expression (Fig. 1, A and B). TF up-regulation was noted at 3 h and reached a peak (7-fold over control) at 6 h post-VEGF exposure. TNF-␣ treatment resulted in a more rapid and greater increase in TF expression relative to VEGF. The maximal TF expression (110-fold) was observed at 3 h post-TNF-␣ exposure (Fig. 1, A and B). Coincubation of VEGF with TNF-␣ produced an even more striking increase in total TF expression. The maximum 270-fold increase occurred at 3 h post-exposure to both agonists (Fig. 1, A and B). The TF up-regulation persisted for at least 9 h. We also investigated the effect of TNF-␤, another member of TNF family cytokine, on TF expression in this cell culture model. As shown in Fig. 1, C and D, treatment with TNF-␤ alone also resulted in a timedependent increase in TF expression, although to a lesser degree than TNF-␣ (35-fold for TNF-␤ versus 110-fold TNF-␣). When VEGF was coincubated with TNF-␤, a maximal of 130fold increase in TF expression was observed at 9 h post-exposure (Fig. 1, C and D). Taken together, these data suggest that VEGF is capable of interacting with either TNF-␣ or TNF-␤ in a synergistic manner to increase TF expression.
To elucidate the role of VEGF receptors in the synergistic up-regulation of TF induced by VEGF and TNF concomitant treatment, we treated HUVEC cells with equimolar concentrations of wild-type VEGF-, KDR-, or Flt-1 receptor-selective variants in the presence or absence of TNF-␣ for 6 h. The VEGF HBD-deficient variants (VEGF 109 , KDR-sel2 ) were also used to address the role of the HBD in TF up-regulation. Fig. 2, A and  B shows a representative WB result and quantification by densitometry of three independent experiments. In the absence of TNF-␣, KDR-selective variants induced an increase in total TF expression similar to that induced by wild-type VEGF (both ϳ7-fold); in the presence of TNF-␣, a more pronounced increase (ϳ250-fold) in total TF expression was induced by wild-type VEGF-and KDR-selective variants (Fig. 2, A and B). In contrast, the Flt-1-selective variant had no effect on TF expression in either the presence or absence of TNF-␣ (Fig. 2, A and B). The HBD did not appear to play a significant role in TF upregulation, as VEGF variants without the HBD (VEGF 109 and KDR-sel2 ) induced an increase in TF expression similar to that induced by VEGF variants with the HBD (VEGF 165 and KDR-sel1 ) (Fig. 2).
FACS analysis was used to confirm these observations and to further characterize the subcellular localization of the newly synthesized TF. In the absence of TNF-␣, KDR receptor variants induced an increase in cell surface TF expression similar to that induced by wild-type VEGF (ϳ3-fold) (Fig. 3), and the Flt-1-selective variant had no significant effect (Fig. 3). In a separate experiment, treatment with TNF-␣ alone induced an ϳ10-fold maximal increase in TF fluorescence density (data not shown). When coincubated with TNF-␣, wild-type VEGF-and KDR-selective variants both induced an ϳ20-fold maximal increase in the cell surface TF expression (Fig. 3). In contrast, Flt-selective variant did not potentiate TNF-␣-induced TF expression on the cell surface (Fig. 3). Once again, the HBD did not seem to play a role in up-regulating cell surface TF expression as VEGF variants without HBD (VEGF 109 and KDR-sel2 ) induced a similar increase in TF expression as VEGF variants with HBD (VEGF 165 and KDR-sel1 ) (Fig. 3).
VEGF ϩ TNF-␤. D, densitometeric quantification of WB from three independent experiments. -Fold increase of TF expression in treated groups was normalized to control cells that were not exposed to VEGF or TNF. Data shown are means Ϯ S.D.

FIG. 1. VEGF potentiates TNF-␣-and TNF-␤-induced TF expression.
Subconfluent HUVEC cells were starved in a serum/GF-free medium overnight and then treated with VEGF (1000 pM), TNF-␣ (5 ng/ml), TNF-␤ (5 ng/ml), or a combination of VEGF and TNF-␣, VEGF, and TNF-␤ for 0, 3, 6, and 9 h. TF expression in whole cells lysates was evaluated by WB as described under "Experimental Procedures." A, representative WB of cell lysates following treatment with VEGF, TNF-␣ alone, or the combination of VEGF and TNF. B, densitometeric quantification of WB from three independent experiments. C, representative WB of cell lysates following treatment with VEGF, TNF-␤, or

FIG. 2. KDR receptor signaling enhanced TNF-induced TF expression.
HUVEC cells were treated with equimolar concentrations (1000 pM) of the Fltsel variant, wild-type VEGF165, KDR -sel1 , or HBDdeficient variants VEGF109 and KDR -sel2 in the presence or absence of TNF-␣ (5 ng/ml) for 6 h. Control cells were treated in the presence and absence of TNF-␣ without VEGF variants. The treated cells were subjected to WB analysis for evaluating total TF expression. A, a representative WB analysis. B, densitometeric quantification of TF expression from three independent experiments. Data represent means Ϯ S.D. * indicates the TNF minus group compared with the TNF minus control; # indicates the TNF plus group compared with the TNF plus control. (* or #, p Ͻ 0.05).
To further address the role of the downstream signaling pathway following KDR binding and activation in TF up-regulation, several inhibitors to tyrosine kinases, Src kinase, PKC, or phosphatidylinositol 3-kinase (PI3K) were coincubated with VEGF and TNF. Fig. 4 shows that inhibition of protein-tyrosine kinase (PTK) with genistein, a potent tyrosine kinase inhibitor, significantly blocked VEGF/TNF-induced TF up-regulation. Similarly, inhibition of PKC with staurosporine also blocked VEGF/TNF-induced TF up-regulation. In contrast, treatment of HUVECs with only PMA, a PKC activator, resulted in an increase in TF expression similar to that induced by VEGF/ TNF. Inhibition of phospholipase C-␥ with PP1 or inhibition of mitogen-activated protein kinase with PD-98059 only partially blocked VEGF/TNF-induced TF up-regulation. Meanwhile, inhibition of PI3K with wortmannin resulted in an increase in TF expression. These data suggest that both PTK and PKC are involved in VEGF/TNF-induced TF up-regulation and that PI3K may negatively regulate TF expression in ECs.
The effect of VEGF on TF expression was also compared with several other GFs including FGF, HGF, EGF, TGF-␤1, and TGF-␤2. HUVEC cells were treated with equimolar concentrations of VEGF, FGF, HGF, EGF, or TGF-␤ in either the presence or absence of TNF-␣ for 6 h. Fig. 5 shows that only VEGF treatment increased TF expression and potentiated TNF-␣induced TF up-regulation, whereas FGF, HGF, EGF, or TGF had little or no significant effect. DISCUSSION Tissue factor is a 47-kDa transmembrane glycoprotein that plays an important role in coagulation by serving as a cofactor for coagulation factor VII activation. It is also involved in early wound healing, angiogenesis, and tumor metastasis (31)(32)(33). Normally, TF is not expressed on the surface of cells that are in direct contact with blood such as the endothelium lining vessels and circulating monocytes; however, growth factors and cytokines may induce TF expression in these cells (25)(26)(27). HUVEC cells were selected as our model system, because they have been shown to express both KDR and Flt-1 receptors and have been previously used to assess TF regulation (26 -27). VEGF and TNF have been shown to synergistically up-regulate TF expression in endothelial cells, but the role of VEGF receptors in this synergy has not been studied. Therefore, we have 1) investigated the role of the VEGF receptor(s) in TF up-regulation by using VEGF receptor-selective binding variants; 2) de- termined the contribution of PTK, PKC, and PI3K to VEGF/ TNF-induced TF up-regulation; and 3) evaluated the specificity for the effect by comparing the ability of VEGF with various other GFs to up-regulate TF. Our data demonstrate that wildtype VEGF-or KDR-selective variants alone induce a modest increase in both total and cell surface TF expression relative to TNF. In the presence of TNF, wild-type VEGF-and KDRselective variants both significantly potentiate TNF-induced TF up-regulation. Furthermore, we demonstrate that PTK and PKC signaling pathways are required for this effect. In comparison with other angiogenic growth factors, VEGF is unique in its ability to induce TF up-regulation and synergistically interact with TNF to produce TF.
TNF-␣ and TNF-␤ are structurally similar, but TNF-␣ binds to the TNFR60 receptor with higher affinity (31). Our data are consistent with this observation in that VEGF/TNF-␣ induced a higher TF expression relative to VEGF/TNF-␤. Previously, Clauss et al. (27) and Camera et al. (26) reported that VEGF or TNF-␣ alone had little effect on TF expression in HUVEC, but cells treated with both agonists showed a significant increase in TF expression. Our observations are consistent with these previously published reports in that VEGF synergistically upregulates TNF-induced TF expression; however, we were able to demonstrate an increase in TF expression with either VEGF or TNF alone.
It is well accepted that VEGF exerts its biological effects through the two high affinity receptor tyrosine kinases, KDR and Flt-1, predominately expressed on the vascular endothelial cells (2,11). However, it is unclear which VEGF receptor(s) mediates the synergy between VEGF and TNF with respect to TF up-regulation. To address this question, we have used VEGF receptor-selective variants recently developed by Li et al. (14). These receptor-selective binding variants have been demonstrated to be highly selective and bioactive (14). Treatment of HUVEC cells with KDR-selective variants resulted in an increase in both total and cell surface TF expression similar to that induced by wild-type VEGF and significantly enhanced TNF-␣-induced TF up-regulation (Figs. 2 and 3), whereas the Flt-1-selective binding variant had no significant effect on TF expression in either the presence or absence of TNF (Figs. 2 and 3). The dose (1000 pM) of KDR-selective variants used in the study was about 5-10-fold higher than the K d for KDR but well below the K d for Flt-1 considering the 2000-fold affinity reduction for Flt-1 binding. The dose (1000 pM) of Flt-1-selective variant was about 50-fold higher than the K d for Flt-1, yet still well below the K d of the KDR receptor. The HBD did not play a significant role in mediating TF up-regulation as VEGF variants without the HBD (VEGF 109 and KDR -sel2 ) induced an increase in TF expression similar to that induced by VEGF variants with HBD ( Figs. 2 and 3). These results rule out the requirement for heparin sulfate proteoglycans or neuropilin binding in TF up-regulation. Taken together, our data strongly indicate that the KDR receptor signaling mediates TF upregulation and governs the synergy between VEGF and TNF.
Consistent with this observation, Meyer et al. (32) reported that VEGF-E, a novel VEGF encoded by Orf virus and bound with high affinity to KDR but not to Flt-1, induced the production of TF in HUVEC cells. Others have reported that the Flt-1 receptor may also play a role in TF up-regulation in monocytes and ECs (25,33). Addition of placental growth factor, which binds to Flt-1 but not to KDR, resulted in a relatively small increase in TF expression in monocytes and ECs (25). In the present study, we were unable to demonstrate any increase in TF expression with the Flt-1-selective variant or placental growth factor in HUVEC cells (data not shown). Thus, our data strongly imply that KDR is the dominant receptor signaling governing TF up-regulation and mediating the synergy between VEGF and TNF in EC. Our observation adds to the growing list of actions attributed to KDR receptor signaling, which includes cell proliferation, endothelial nitric oxide synthase, and KDR up-regulation, cell survival, and vascular permeability (13, 16, 30, 34 -36).
Many studies have shown the importance of protein-tyrosine kinases and PKC in VEGF signaling (21, 30, 34 -35). KDR itself and many downstream molecules are protein-tyrosine kinases or regulated by tyrosine kinase phosphorylation (21). To address the role of these signaling molecules in VEGF/TNF-induced TF up-regulation, cells were treated in medium containing VEGF/TNF, as well as inhibitors to PTK, PKC, PI3K, phospholipase C-␥, or mitogen-activated protein kinase. As expected, inhibition of PTK with genistein significantly blocked the synergistic up-regulation of TF. Similarly, inhibition of PKC with staurosporine also significantly attenuated the synergistic up-regulation of TF. Stimulation with PMA alone resulted in a striking increase in TF expression, indicating that PKC plays a central role in VEGF/TNF-induced up-regulation of TF. Previously, several isoforms of PKC such as PKC-␣, -␤, -␥, -, and -⑀ have been shown to be activated by VEGF stimulation (34,(37)(38). Because TNF has also been shown to activate PKC (39), the synergistic up-regulation of TF induced by VEGF and TNF may occur at the level of PKC. More studies will be required to elucidate the specific role of PKC isoform(s) in TF up-regulation. Inhibition of Src kinase and mitogenactivated protein kinase partially blocked TF up-regulation, indicating that these signaling molecules may also contribute to the signaling mechanism involved in TF up-regulation. Interestingly, we found that inhibition of PI3K with wortmannin resulted in an even higher induction of TF expression than VEGF/TNF, suggesting that PI3K may negatively regulate TF production.
Other mechanisms may also contribute to the synergy between TNF and VEGF. For example, TNF-␣ has been reported to modulate VEGF action by promoting VEGF production and regulating KDR and neuropilin-1 receptor expression (40 -42). It is unlikely that TNF-induced expression of VEGF contributed to the TF up-regulation in our model system as endothelial cells express little or no VEGF, and saturating concentrations of exogenous VEGF (1000 pM) were used in the study. It is also unlikely that TNF-induced KDR expression augmented the biological response as it took ϳ24 h for TNF to significantly increase KDR expression (40), and we observed TF expression occurring as early as 3 h post-VEGF treatment.
Up-regulation of TF induced by the interplay between VEGF and TNF may be critical in physiological angiogenesis, as well as tumor angiogenesis and metastasis (43)(44)(45)(46)(47). TNF-␣ has been found to cooperate with VEGF and other GFs to induce capillary-like tubular structure of human microvascular endothelial cell growth in a three-dimensional gel of extracellular matrix proteins (47). Increased TF expression may facilitate cell migration, an important step in tube formation (48). In addition to its role in angiogenesis, TF may also play a role in cancer progression, atherosclerosis, and thrombotic complications (43, 49 -52). Both VEGF and TNF levels have been found to be increased under these pathophysiological conditions (53)(54). The elevated TNF and VEGF levels in these conditions may contribute to the increased TF expression in vivo, resulting in increased angiogenesis, thrombosis, and tumor metastasis.
In summary, we have investigated the role of VEGF receptors in mediating the synergistic up-regulation of TF induced by VEGF and TNF in human endothelial cells. Our data indicate that KDR receptor signaling governs the synergistic up-regulation of TF induced by VEGF and TNF. These data provide new insights into the role of the KDR receptor in mediating both physiological and pathophysiological processes.