Factor VIIa Induces Tissue Factor-dependent Up-regulation of Interleukin-8 in a Human Keratinocyte Line*

Tissue factor (TF), a transmembrane receptor for the serine protease coagulation factor VII(a) (FVIIa), is the main initiator of the coagulation cascade. Through incompletely elucidated mechanisms, TF serves additional functions in tumor-associated angiogenesis and metastasis. We have studied interleukin-8 (IL-8) as a possible link between TF-FVIIa complex formation and subsequent processes. Recombinant human FVIIa induced the up-regulation of both IL-8 mRNA and protein in a FVIIa dose- and time-dependent fashion. A neutralizing antibody to TF reduced this induction by 93 ± 5%. Active site-inhibited FVIIa had no stimulatory effect and completely blocked that of FVIIa. This confirms that the increased IL-8 production was dependent on the formation of TF-FVIIa complexes and the proteolytic activity of FVIIa. The IL-8 promoter contains DNA binding sites for nuclear factor-κB (NF-κB) and activator protein-1 (AP-1). In response to FVIIa, the DNA binding activity of both NF-κB and AP-1 was enhanced in an electrophoretic mobility shift assay. In addition, theIL-8 promoter was transcriptionally activated both in a luciferase reporter system and a nuclear run-off assay. Moreover, IL-8 mRNA stability was significantly enhanced by FVIIa-induced activation of the mitogen-activated protein kinases ERK1/2 and p38. Taken together, TF-FVIIa signaling induced increased transcription as well as mRNA stabilization leading to the significant up-regulation of IL-8 protein synthesis.

The binding of factor VIIa (FVIIa) 1 to its membrane receptor tissue factor (TF) initiates the blood coagulation cascade, leading to the activation of factor IX and X and generation of thrombin (1,2). In addition, TF plays an important role in tumor-associated angiogenesis (3,4) and metastasis (5). These processes are reduced by the use of TF antibodies (6) and active site-inhibited FVIIa (7). Moreover, a human melanoma cell line expressing a TF mutant, which prevents the binding of FVIIa, had significantly reduced metastatic ability when introduced into severe combined immunodeficiency (SCID) mice (8), indicating that formation of the TF-FVIIa complex is required for the full metastatic effect of TF. Several studies have emphasized the importance of catalytically active FVIIa binding in TF-dependent metastasis (5,7). The involvement of TF in regulation of blood vessel development was demonstrated by targeting the TF gene in mice (9 -11). However, the mechanisms by which TF-FVIIa enhances angiogenesis and metastasis remain poorly understood. One mechanism by which TF-FVIIa complex formation may function in these processes is through intracellular signaling events.
Intracellular signal transduction pathways induced by the TF-FVIIa complex have been extensively studied. This group first reported FVIIa-induced intracellular Ca 2ϩ oscillations (12). Subsequently, it was demonstrated that TF-FVIIa signaling activated MAPKs (p38, ERK1/2, JNK) (13,14) and altered gene expression (14,15). Recently, activations of Src-like kinase and phosphatidylinositide 3-kinase (PI 3-kinase) were suggested to be important in TF-FVIIa signaling (16). At present, it has not been fully elucidated how these cellular signals relate to TF effects under conditions such as metastasis and angiogenesis. One possibility is that TF-FVIIa signaling can induce the expression of growth regulators affecting downstream cellular processes. Therefore we employed cDNA arrays to study the alterations in gene expression in response to FVIIa in HaCaT cells. Up-regulation (Ͼ2ϫ) of more than 25 genes was seen (15), one of which was interleukin-8.
IL-8, a member of the CXC chemokine family, was originally identified as a neutrophil chemoattractant (17) and has later been shown to be a potent proangiogenic factor inducing corneal vascularization (18) and angiogenesis both in vitro and in vivo (19,20). IL-8 production correlated with the metastatic potential of tumor cells (21). Basal production of IL-8 is normally low or undetectable but can be rapidly induced by different stimuli such as cytokines, viruses, and reactive oxidants and is then secreted from a variety of cells, including keratinocytes. IL-1␤ and tumor necrosis factor ␣ (TNF␣) are the strongest known stimuli for IL-8 expression in most cell types (22). This rapid induction is primarily regulated at the transcriptional level by binding of inducible transcription factors to the IL-8 promoter. The IL-8 promoter contains a single transcriptional initiation site with a consensus TATA box. The 5Ј-regulatory region of the IL-8 gene (Ϫ1 to Ϫ133), containing binding sites for NF-B, NF-IL-6, and AP-1 (23,24), is essential and sufficient to regulate the IL-8 expression. Several AU-rich elements (AREs) located in the 3Ј-untranslated region have been reported to be involved in the post-transcriptional regulation of IL-8 expression (25).
In light of the involvement of TF and IL-8 in angiogenesis and metastasis and the up-regulation of IL-8 mRNA by formation of the TF-FVIIa complex (15), it is of interest to study how the mRNA is up-regulated. We report here that the major part of the increase is caused by post-transcriptional mRNA stabilization and is followed by significantly increased expression of IL-8 protein.
Cell Line and Cell Culture-HaCaT, an immortalized human keratinocyte cell line constitutively expressing TF (15), was kindly provided by Dr. U. Birk Jensen (University of Aarhus, Denmark) and cultured in serum-free keratinocyte medium (SFM) supplemented with bovine pituitary extract (25 g/ml) and recombinant epidermal growth factor (0.5 ng/ml). Before the addition of agonists, the medium was changed to SFM with 1.5 mM CaCl 2 for 2 h to facilitate Ca 2ϩ -dependent binding of FVIIa to TF. Pretreatment with antibodies and inhibitors was done during this time.
Quantitative Human IL-8 Immunoassay-The amount of IL-8 secreted into the culture medium was analyzed using a quantitative and specific sandwich enzyme immunoassay (QuantiGlo human IL-8 chemiluminescence kit, R&D Systems). Samples and standards were assayed in parallel.
RNA Extraction and Northern Analysis-HaCaT cells were washed with cold PBS and scraped into 400 l of lysis buffer (100 mM Tris-HCl, pH 8, 500 mM LiCl, 10 mM EDTA, 1% lithium dodecyl sulfate, 5 mM dithiothreitol). Oligo(dT)-conjugated magnetic beads (Dynal, Norway) were used for mRNA isolation according to the manufacturer's instructions. Samples were run on agarose/formaldehyde gels in MOPS buffer and blotted. Prehybridization and hybridization were performed in ExpressHyb TM solution (CLONTECH). The blots were hybridized with 32 P-labeled probes to IL-8 and GAPDH. The IL-8 probe was derived from the IMAGE consortium 328846 (HGMP Resource Centre, UK). Complete cDNA was used to generate a probe for GAPDH. The filters were scanned using a PhosphorImager STORM 860. Quantitation was done with ImageQuant 5.1 (Amersham Biosciences). Intensity values of IL-8 mRNA were normalized to those of GAPDH.
IL-8 mRNA Stability Assay-HaCaT cells were grown to 70 -80% confluency and treated with 10 nM FVIIa for 1 h to induce IL-8 mRNA production. Then actinomycin D (ActD) (10 g/ml) was added to the culture to inhibit further transcription. In order to assess the effects of activation of MAPKs on the IL-8 mRNA stability, MAPK inhibitors PD98059 (50 nM, for MEK1 and indirectly for ERK1/2) or SB203580 (20 nM, for p38) were added together with ActD. Cells were harvested at different time points (0, 1, 2, 3 h), and mRNA was extracted for Northern analysis. The ratio of IL-8 to GAPDH was calculated. The level of IL-8 mRNA at the time of addition of ActD at 0 h was set to 100%. Curves fitted by least-squares regression were used for calculation of mRNA half-life.
Transient Transfection and Luciferase Reporter Assay-At 60 -70% confluency, HaCaT cells from 6-well plates were transfected with 1 g of firefly luciferase expression vector in a 1:3 ratio of DNA/FuGENE 6 (Roche, Indianapolis) according to the manufacturer's instructions. The transfection efficiency was normalized by cotransfection with 50 ng of pRL (Renilla luciferase) reporter DNA containing the full-length Renilla luciferase gene under the control of the human EF promoter. Three firefly expression vectors under the control of the IL-8 promoter (Fig. 4A) were used: wild type-133-luc, harboring the 5Ј-flanking region Ϫ133 to ϩ44 bp of the IL-8 gene and carrying the AP-1-(Ϫ126 to Ϫ120) and NF-B-like (Ϫ80 to Ϫ71) binding sites, and the mutant AP-1-luc (TatCTCA) and mutant NF-B-luc (taAcTTTCCTC), where the AP-1 and NF-B sites were mutated, respectively (25). 24 h post-transfection, cells were treated with 10 nM FVIIa or IL-1␤ (200 units/ml) for 6 h and then lysed in 500 l of passive lysis buffer (Promega). The activities of firefly and Renilla luciferase were measured in 15 l of cell lysate by using the dual-luciferase reporter assay system (Promega) in a luminometer (Berthold). Firefly luciferase activity was normalized on the basis of the Renilla luciferase activity. All transfections were performed three times and in triplicate.
Nuclear Run-off in Vitro Transcription-HaCaT cells were treated with or without 10 nM FVIIa for 1 h before harvest. After incubation in cold PBS with 1% bovine serum albumin on ice for 5 min, the cells were scraped off and centrifuged at 1800 ϫ g for 10 min. Pellets were resuspended in 100 l of PBS and 900 l of lysis buffer (10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 , 0.5% Nonidet P-40). The cell suspension was incubated on ice for 5 min with occasional vortexing and then centrifuged at 2000 ϫ g at 4°C for 5 min to collect the cell nuclei. The nuclei were suspended in the storage buffer (50 mM Tris-HCl, pH 8.3, 40% glycerol, 5 mM MgCl 2 , 0.1 mM EDTA) and stored at Ϫ80°C.  (27). Next, unlabeled UTP (3 l of 10 mM) was added, and the reaction was continued for 15 min. The mixture was incubated with 20 l of RNase-free DNaseI (1 mg/ml) for 10 min at 27°C, and proteins were digested with proteinase K for 2 h at 42°C (10 mM Tris-HCl, pH 7.4, 15 mM EDTA, 3% SDS, 1 mg/ml proteinase K). The reaction mixture was then extracted three times with phenol/chloroform, and the RNA was precipitated with ethanol on dry ice and dissolved in 50 l of diethyl pyrocarbonate (DEPC)-treated H 2 O. Labeled nuclear RNA was hybridized to 5 g of pBR322 DNA (as a negative control) or cDNA (IL-8 and GAPDH) immobilized on nitrocellulose membranes by using a slot-blot apparatus. GAPDH was used as an internal control. Hybridization and quantification were performed as described for Northern analysis. The IL-8 signal was normalized to that of GAPDH.
Preparation of Nuclear Extracts and Electrophoretic Mobility Shift Assay (EMSA)-HaCaT cells were treated with 10 nM FVIIa for 50 min. Nuclear extracts were prepared according to the method of Lee et al. (28). Briefly, cells were washed with ice-cold PBS and pelleted. Pellets were resuspended for swelling in 500 l of buffer A (10 mM Hepes, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 1 mM dithiothreitol) on ice for 15 min. After shearing eight times through a 25-gauge needle, the nuclei were spun down and washed in buffer A to remove trapped cytoplasmic material. Nuclei were resuspended in 50 l of buffer C (20 mM Hepes, pH 7.9, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.42 M NaCl, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 25% glycerol) and incubated on ice for 30 min with occasional shaking. After centrifugation, the supernatant was harvested and stored at Ϫ80°C. Protein concentrations of nuclear extracts were measured using the Bio-Rad DC protein assay. The following double-stranded oligonucleotides (sequence derived from the IL-8 promoter) were used (binding sites, underlined): NF-B-like binding sequence, 5Ј-GCGATCGTGGAATTTCCTCTGACGC-3Ј; AP-1 binding sequence, 5Ј-GAAGTGTGATGACTCAGGTTTGCCTGA-3Ј. For the EMSA reaction, 2.5 g of nuclear protein were incubated at room temperature for 20 min in a 10-l binding reaction mixture containing 5% glycerol, 5 mM MgCl 2 , 0.5 mM EDTA, 50 mM NaCl, 10 mM Tris-HCl, pH 7.5, 50 g/ml poly(dI-dC), 0.5 mM dithiothreitol, and 30,000 -40,000 cpm of 32 P-labeled probe. For the supershift assay, antibodies (300 ng) against AP-1 (JunD, c-Fos) or NF-B (p65 and p50 subunits) were included in the binding reaction. The DNA-protein complexes were subjected to gel electrophoresis on a 4% non-denaturing polyacrylamide gel in 0.5ϫ Tris borate/EDTA buffer. Gels were dried at 80°C and developed using a PhosphorImager STORM 860.
Immunocytochemistry-HaCaT cells were grown on coverslips to 70 -80% confluency. The cells were stimulated with 10 nM FVIIa for 50 min, washed twice in PBS, fixed in 4% paraformaldehyde for 20 min, and then in 4% paraformaldehyde containing 0.2% Triton X-100 for an additional 20 min. The coverslips were washed in PBS and blocked with 1% bovine serum albumin in PBS for 10 min. The cells were then incubated for 1 h with 4 g (in 100 l of blocking buffer) of a mouse monoclonal IgG against human NF-B p50 subunit, then washed three times in PBS, and incubated with a fluorescein isothiocyanate-conjugated goat anti-mouse IgG (DAKO) for 1 h. The cells were washed three times in PBS and counterstained with DAPI before mounting. The samples were viewed with a Nikon Eclipse E400 microscope. Pictures were taken with a SPOT digital camera and combined with SPOT 32 software (Diagnostic Instrument).
Statistical Analysis-Data are presented as the mean Ϯ S.E. of three independent experiments, each performed in triplicate and analyzed with the Student's two-tailed t test. p Ͻ 0.05 was regarded as statistically significant.

Effect of FVIIa Binding to TF on IL-8 Production-
HaCaT cells were treated with 10 nM rhFVIIa (the physiological plasma level), and the amount of IL-8 released into the medium after 3.5 h was quantified by using an ELISA specific for human IL-8 (Fig. 1A). The level of IL-8 was 2.6-fold higher (145 Ϯ 5 pg/mg protein) in rhFVIIa-treated cells compared with non-treated cells (56.5 Ϯ 8 pg/mg protein, p Ͻ 0.01). No difference was seen in the level of IL-8 protein in the cell lysate of controls and rhFVIIa-treated cells (data not shown), suggesting that FVIIa increased both IL-8 synthesis and secretion. Incubation with different concentrations of rhFVIIa (10 and 25 nM) induced IL-8 production in an apparently dose-dependent fashion (Fig. 1B). To study whether IL-8 induction was TF-dependent, FVIIa binding was blocked for 2 h prior to the addition of FVIIa with HTF1-7B8 (25 g/ml), a neutralizing monoclonal antibody against human TF. HTF1-7B8 inhibits the TF procoagulant activity by competition with FVIIa for association with TF (29). The secreted IL-8 was measured 3.5 h after the addition of rhFVIIa. The HTF1-7B8 antibody abolished 93 Ϯ 5% of the IL-8 induction (p Ͻ 0.01), demonstrating that TF is necessary for FVIIa-mediated IL-8 induction (Fig. 1B). To see whether IL-8 up-regulation was dependent on the proteolytic activity of FVIIa, cells were incubated with 100 nM DEGR-FVIIa for 3.5 h. This active site-inhibited FVIIa has higher affinity for TF than FVIIa but no proteolytic activity (30). The level of secreted IL-8 was the same as that of the control (Fig.  1B), showing that in contrast to FVIIa, DEGR-inactivated FVIIa failed to induce IL-8 up-regulation. Furthermore, FXa had no effect on FVIIa-induced IL-8 induction (Fig. 1B) because preincubation with rTAP, a specific inhibitor of FXa (31), did not abolish FVIIa-induced IL-8 up-regulation. These data show that IL-8 induction in response to FVIIa is TF-dependent and that the proteolytic activity of TF-FVIIa is necessary. Furthermore, this effect is not a consequence of FXa generation.
Effects of MAPK ERK1/2 and p38 on FVIIa-induced IL-8 Up-regulation-We have previously reported that in response to 10 nM FVIIa, MAPKs were activated in HaCaT cells and IL-8 mRNA synthesis was induced 18.8 Ϯ 2.2-fold within 1 h (15). We have done Northern blots to determine whether the activated MAPKs contributed to the enhanced IL-8 production in FVIIa-stimulated cells. PD98059 (inhibiting MEK1/ERK1/2) or SB203580 (inhibiting p38) was added to cell cultures 30 min before the addition of rhFVIIa (10 nM). PD98059 (50 nM) and SB203580 (20 nM) inhibited FVIIa induction of IL-8 mRNA by 69 Ϯ 3.5% (p Ͻ 0.01) and 31.4 Ϯ 6% (p Ͻ 0.04), respectively. The FVIIa-induced up-regulation of IL-8 protein was reduced by 40% (PD98059) and 20% (SB203580). One possibility is that these MAPKs (ERK1/2 and p38) contribute to the stabilization of IL-8 mRNA, thus allowing strong and rapid gene expression. To investigate this possibility, a mRNA stability assay was performed (Fig. 2). HaCaT cells were treated with 10 nM FVIIa for 1 h to induce IL-8 mRNA production. Then transcription was stopped by addition of ActD, and the half-lives of IL-8 transcripts were studied. In unstimulated HaCaT cells, the half-life of IL-8 transcripts was around 40 min, whereas it had increased to 140 min in FVIIa-stimulated cells. When FVIIastimulated cells were treated with either SB203580 or PD98059 together with ActD, the half-life of IL-8 mRNA was significantly reduced to 80 and 43 min, respectively (Fig. 2). This explains why PD98059 inhibited IL-8 induction more effectively compared with SB203580. These experiments show that FVIIa-induced activation of MAPKs ERK1/2 and p38 contributes to IL-8 up-regulation at least partly through the significantly enhanced stability of IL-8 mRNA.
FVIIa Induces Transcriptional Activation of the IL-8 Promoter-The expression of the IL-8 gene is regulated at both the transcriptional and post-transcriptional levels. To further assess the effect of FVIIa on the IL-8 promoter, HaCaT cells were transiently transfected with a firefly luciferase reporter gene, driven by the promoter of the IL-8 gene (Ϫ133 to ϩ44) or a point-mutated IL-8 promoter (Fig. 3A). The transcriptional activity of both mutated promoters was significantly reduced compared with the wild type. Consistent with other studies, both AP-1 and NF-B were necessary for IL-8 transcription; neither AP-1 nor NF-B alone was sufficient for complete IL-8 regulation. However, NF-B appears to play a dominant role, because the NF-B binding site mutation reduced luciferase activity by 90% compared with the wild type promoter.
In view of this and the marked translocation of NF-B to the nucleus caused by addition of FVIIa to HaCaT cells (see below), it was of interest to see whether a mutated NF-B binding site was accompanied by a disproportionate effect on the inducibility of the IL-8 promoter. When we compared the increased luciferase activity induced by FVIIa (10 nM) in transfected cells (Fig. 3B) units/ml, a known stimulus for IL-8 mRNA synthesis) (32,33) in transfected cells and saw a similar minor induction. This suggested that under the conditions of our experiments the promoter of the IL-8 gene was activated by FVIIa to the same degree as with IL-1␤.
To further confirm the effect of FVIIa on the transcription of the IL-8 gene, a nuclear run-off in vitro transcription assay was performed. HaCaT cells were treated with FVIIa for 1 h. Nascent mRNA transcripts elongated in the isolated nuclei were labeled with [␣-32 P]UTP and hybridized to specific cDNA. The results showed a 15-80% increase of IL-8 transcripts in the FVIIa-treated cells (Fig. 4), confirming that increased transcriptional activity of the IL-8 promoter was only a minor cause for FVIIa-induced IL-8 up-regulation.
FVIIa Enhances DNA Binding Activity of AP-1 and NF-B-Because IL-8 transcription was found to be increased by FVIIa and because the transcription factors AP-1 and NF-B are critical for the regulation of IL-8 transcription, it was of interest to investigate whether FVIIa could increase the binding of these transcription factors to their respective DNA binding sites on the IL-8 promoter by using EMSA. The DNA-protein binding complex was increased in FVIIa-treated cells (Fig. 5). To identify specific protein complexes, we performed supershift experiments. Antibodies specific to NF-B subunits p50 and p65 retarded the mobility of the DNA-protein complexes and formed supershifts. This effect was clearly more marked in nuclear extracts from FVIIa-treated cells (Fig. 5A, lanes c and  e) than in those from non-treated cells (Fig 5A, lanes b and d). Thus FVIIa (10 nM) increased the DNA binding of NF-B p50 and p65 subunits.
Because various AP-1 proteins can bind the same DNA element with different affinities, we analyzed the composition of the AP-1-DNA complex with antibodies against the AP-1 family (JunB, JunD, c-Jun, Fra-1, Fra-2, FosB, c-Fos). The c-Fos and JunD antibodies significantly retarded the mobility of the AP-1-DNA complex, whereas the other AP-1-specific antibodies did not result in apparent mobility shifts even at a higher concentration of antibody (600 ng) (data not shown). This suggested that c-Fos and JunD are the main components of the FVIIa-induced AP-1-DNA complex. Both specific c-Fos and JunD antibodies gave stronger supershifts in FVIIa-treated cells than in non-treated cells (Fig. 5B, Upper part, lanes b-e), indicating that FVIIa enhanced DNA binding activity of c-Fos and JunD.
FVIIa Stimulates NF-B Nuclear Translocation-In resting cells, NF-B was mostly retained in the cytoplasm. Cellular activation results in the translocation of NF-B to the nucleus where it binds to its genomic targets, thus enhancing gene expression (34 -36). Because the DNA binding activity of NF-B was enhanced in response to FVIIa, it was of interest to test the effect of FVIIa on NF-B translocation. HaCaT cells were stimulated with 10 nM FVIIa for 50 min before fixation and stained with a monoclonal anti-p50 antibody. In nontreated HaCaT cells, NF-B was mainly observed in the cytoplasm (Fig. 6A). In contrast, NF-B p50 was clearly translocated to the nucleus in cells treated with FVIIa (Fig. 6, B and  C). NF-B p65 showed the same nuclear translocation in response to FVIIa (data not shown). DISCUSSION Besides its primary function in coagulation, TF has other functions in cellular processes, such as angiogenesis and me-tastasis. Although it has been shown that TF-FVIIa triggers intracellular signaling, it is still unclear how these specific events influence downstream biological functions. We report here that IL-8 production in HaCaT cells was significantly induced in a dose-and time-dependent manner by rhFVIIa stimulation. No other clotting factors were included in our system. The inhibitors hirudin (for thrombin) (15) and rTAP (for FXa) failed to suppress the FVIIa-induced accumulation of IL-8. DEGR-inactivated FVIIa had no stimulatory effect on IL-8 production. We conclude that IL-8 induction is dependent on the proteolytic activity of the TF-FVIIa complex but independent of components of the coagulation pathway below TF-FVIIa. This is in contrast to VEGF, the only other protein hitherto reported to be induced by the TF-FVIIa complex at the protein level, where FXa was recently shown to be necessary (37,38). Neither inactivated FVIIa nor FVIIa affected VEGF production (39). Regulation of VEGF synthesis must therefore be mediated at least in part through another pathway than that utilized for IL-8.
It has previously been shown that IL-8 expression is primarily regulated at the transcriptional level (22,24,40). The IL-8 promoter contains a distal AP-1 binding site and a proximal element containing binding sites for NF-B and NF-IL-6. NF-IL-6 and AP-1 have been shown to physically interact with NF-B, suggesting that their functional cooperation is critical for effective regulation of IL-8 gene expression (23). By using a luciferase reporter gene assay and a nuclear run-off assay we found that the IL-8 promoter was transcriptionally activated to a moderate degree in response to FVIIa. The NF-B binding site was the predominant cis-acting element in IL-8 gene expression, because point mutation of the NF-B binding site in the IL-8 promoter abolished 90% of the luciferase activity, whereas mutation in the AP-1 binding site was not as influential. The effects of FVIIa were essentially equal (ϳ30 -40% increase of reporter gene synthesis) in cells transfected with mutant and wild type IL-8 promoter constructs. FVIIa also enhanced the binding activity of nuclear transcription factors NF-B and AP-1 to their DNA binding sites in EMSA. Thus, both transcription factors were activated and apparently functioned together in IL-8 up-regulation driven by the FVIIa signaling pathway.
In resting cells, NF-B resides in the cytoplasm by binding to its inhibitor IB. In response to various cellular stimuli, IB is phosphorylated and degraded, leading to the translocation of NF-B to the nucleus where it binds DNA and activates transcription (41). We have found that both subunits p65 and p50 of NF-B were translocated to the nucleus in response to FVIIa, confirming the significance of NF-B in FVIIa-induced IL-8 expression. We have also found that TF-FVIIa signaling results in the phosphorylation and degradation of IB. 2 AP-1 activity is controlled at both transcriptional and posttranscriptional levels (42). Most of the genes that encode AP-1 components behave as "immediate-early" genes in which increased transcription is rapidly induced upon cell stimulation and which do not require prior protein synthesis. Because we have previously reported that both c-Fos and IL-8 mRNA are up-regulated as an immediate-early response by FVIIa binding to TF in HaCaT cells (15), the increased c-Fos binding in EMSA observed here cannot be due to de novo protein synthesis and must be due to c-Fos activation. Regulation of AP-1 activity at the post-transcriptional level has been shown to be mediated by phosphorylation (42). FVIIa-activated MAPKs may also contribute to the enhanced AP-1 activity through phosphorylation of the transactivation domain of JunD, thereby potentiating its transactivation function on the FVIIa-stimulated IL-8 promoter.
In conclusion, IL-8 gene up-regulation is in part caused by recruitment of the inducible transcription factors AP-1 and NF-B in response to FVIIa stimulation. The expression of other genes may also be affected in a similar manner, because the DNA binding activity of both NF-B p65/p50 and the AP-1 c-Fos/JunD to their consensus binding sequences was also enhanced in EMSA (data not shown). This is in accordance with our previous results, where the expression of 25 immediateearly response genes was up-regulated in response to FVIIa (15). Twelve of these were NF-B and AP-1 target genes, such as IL-8, c-fos, c-myc, fra-1, collagenase 1 (MMP1), collagenase 3 (MMP13), uPAR, and gadd45.
The transcriptional activation demonstrated to occur cannot account for more than a minor fraction of the total FVIIainduced IL-8 mRNA increase. Investigating IL-8 mRNA halflife under various conditions, we have observed an increase from 40 min in untreated HaCaT cells to 140 min after FVIIa addition. This effect was mediated via MAPK signaling, because blocking ERK1/2 and p38 separately or together reduced the half-life. This regulation has been ascribed to several AREs in the 3Ј-untranslated region of IL-8-and ARE-binding proteins. A similar effect of PD98059 and SK&F 86002 (p38 inhibitor) (43) resulted in destabilization of the transcripts of IL-1␤ and GRO. These effects strongly correlated with the increased binding of ARE-binding protein AUF1 and transcripts. AUF1 has been shown to destabilize transcripts (44). Holtmann et al. (45) documented that MAPK p38 strongly stabilized IL-1-induced IL-8 transcripts through phosphorylating and activating p38 MAP kinase-activated protein kinase MK2. It remains to be determined how these phosphorylation pathways influence ARE recognition and association with different proteins to regulate IL-8 mRNA stability. Our data show that post-transcriptional regulation plays a major role in FVIIa-induced IL-8 up-regulation.
Local tumor growth is strictly dependent on vascularization (46). Angiogenesis is also required for tumor metastasis. Angiogenic factors responsible for tumor neovascularization may be derived from tumor cells themselves or from the responding inflammatory cells (47). IL-8 has been shown to be a potent activator of corneal vascularization (18) and angiogenesis in vitro and in vivo (19). IL-8 production correlates strongly with metastatic potential in many tumors (21, 48 -50). It can induce tumor growth either through its angiogenic properties or as an autocrine growth factor (51,52). High levels of TF have been reported in a variety of tumor cells; metastatic human mela-noma cells have, for example, been shown to express a 1000fold higher level of TF than their non-metastatic counterparts (6). The expression of TF correlates with the potential for metastasis (53)(54)(55) and angiogenesis (26), and the proteolytic activity of FVIIa is crucial in TF-dependent metastasis (5,7). Here we have shown that up-regulation of the synthesis of IL-8 protein depends on the proteolytic activity of the TF-FVIIa complex. When TF is highly expressed in tumor cells or in host endothelial cells, high and specific local TF-FVIIa proteolytic activity may be achieved, thus leading to efficient induction of IL-8 production in the microenvironment and promoting the process of TF-dependent metastasis and angiogenesis.
In summary, FVIIa binding to TF, the initiation of blood coagulation, generates signals that increase the transcriptional activity of NF-B and AP-1 and stimulate the expression of IL-8 protein by enhancing IL-8 mRNA synthesis and its stability. These cellular events may affect processes like angiogenesis and metastasis in vivo.