E. Sulpice, Institut des Vaisseaux et du Sang, Centre de Recherche de l'Association Claude Bernard, Hôpital Lariboisière, 8 rue Guy Patin, 75475, Paris CEDEX 10, France. Fax: +33 1 42 82 94 73, Tel.: +33 1 45 26 21 98, E-mail: email@example.com
The mechanism by which the CXC chemokine platelet factor 4 (PF-4) inhibits endothelial cell proliferation is unclear. The heparin-binding domains of PF-4 have been reported to prevent vascular endothelial growth factor 165 (VEGF165) and fibroblast growth factor 2 (FGF2) from interacting with their receptors. However, other studies have suggested that PF-4 acts via heparin-binding independent interactions. Here, we compared the effects of PF-4 on the signalling events involved in the proliferation induced by VEGF165, which binds heparin, and by VEGF121, which does not. Activation of the VEGF receptor, KDR, and phospholipase Cγ (PLCγ) was unaffected in conditions in which PF-4 inhibited VEGF121-induced DNA synthesis. In contrast, VEGF165-induced phosphorylation of KDR and PLCγ was partially inhibited by PF-4. These observations are consistent with PF-4 affecting the binding of VEGF165, but not that of VEGF121, to KDR. PF-4 also strongly inhibited the VEGF165- and VEGF121-induced mitogen-activated protein (MAP) kinase signalling pathways comprising Raf1, MEK1/2 and ERK1/2: for VEGF165 it interacts directly or upstream from Raf1; for VEGF121, it acts downstream from PLCγ. Finally, the mechanism by which PF-4 may inhibit the endothelial cell proliferation induced by both VEGF121 and VEGF165, involving disruption of the MAP kinase signalling pathway downstream from KDR did not seem to involve CXCR3B activation.
Angiogenesis, the formation of new capillary blood vessels, is controlled by positive and negative regulators. Tumours secrete potent angiogenic factors, including fibroblast growth factors (FGFs), platelet-derived growth factor B (PDGF-B) and vascular endothelial growth factor (VEGF) [1,2]. These factors are counterbalanced by inhibitory molecules such as angiostatin, endostatin, thrombospondin, and platelet factor-4 [3–8].
Platelet factor-4 (PF-4), a member of the CXC chemokine family , inhibits fibroblast growth factor-2 (FGF2)-induced proliferation and migration of endothelial cells [10–14]. Various mechanisms by which PF-4 may inhibit endothelial cell proliferation have been proposed. Via its heparin binding property, PF-4 may inhibit FGF2-induced FGF2-receptor activation [10,11,13,15]. However, in the absence of its heparin-binding domain, PF-4 retains antiangiogenic activity, suggesting another mechanism of inhibition . Indeed, we recently showed that PF-4 inhibits cell proliferation by selectively inhibiting FGF2-induced extracellular signal-regulated kinase (ERK) activation, without affecting the FGF2-induced phosphatidylinositol 3-kinase activation . These results strongly suggest that PF-4 inhibits FGF2-induced endothelial cell proliferation via an intracellular mechanism which, independently of FGF2-induced activation of FGF2-receptors , leads to ERK inhibition.
In addition to its effects on FGF2-induced proliferation, PF-4 also inhibits the proliferation and migration of endothelial cells induced by VEGF [14,15]. VEGF is the most important angiogenic factor, and is present in diverse tumour cells. It stimulates the proliferation, migration and differentiation of endothelial cells [2,18], and is involved in angiogenesis-dependent tumour progression and other diseases associated with angiogenesis, including diabetic retinopathy and rheumatoid arthritis [2,7,19]. VEGF acts via the kinase insert domain-containing receptor (KDR) and Flt1 receptors. Several lines of evidence suggest that the KDR is solely responsible for endothelial cell proliferation [20,21]. Various forms of VEGF have been described  (VEGF121, VEGF145, VEGF165, VEGF189, and VEGF206), all produced from a single gene by alternative splicing . VEGF165 possesses a heparin-binding domain necessary for full activation of KDR  and binding to heparan sulfates on the cell surface, whereas VEGF121 does not . Consequently, VEGF121 promotes endothelial cell proliferation less efficiently than VEGF165. The VEGF-induced signalling pathways involved in endothelial cell proliferation have been extensively documented. VEGF induces the dimerization, autophosphorylation and tyrosine kinase activity of KDRs [20,27]. Phospholipase Cγ (PLCγ), a substrate of KDR kinase, is then phosphorylated and activated, leading to the activation of protein kinase C (PKC), followed by the serine/threonine kinase, Raf1 and then the threonine/tyrosine kinase, MEK1/2 (MAP kinase kinase 1/2) [28–31]. This phosphorylation cascade ultimately leads to activation of the mitogen-activated protein kinases (MAP kinases), also known as extracellular signal-regulated kinases (ERK1/2), which are essential for VEGF-induced endothelial cell proliferation . VEGF also seems to induce the phosphatidyl inositol-3 kinase (PI3-kinase) pathway [28,33]. However, inhibitors of PI3-kinase have no effect on VEGF-induced MAP kinase activation and cell proliferation .
To distinguish between the extracellular effects of PF-4 acting on ligand/receptor activation and intracellular effects on signalling cascades, we compared the effects of this molecule on the signalling pathways involved in the endothelial cell proliferation induced by VEGF165, which binds PF-4, and by VEGF121, which does not. In addition, we investigated the involvement of the newly identified chemokine receptor, the CXCR3B , in this process. PF-4 inhibited the induction of human umbilical vein endothelial cell (HUVEC) proliferation by both VEGF165 and VEGF121. VEGF121-induced KDR autophosphorylation and PLCγ phosphorylation were not affected by the presence of PF-4, whereas VEGF165-induced KDR autophosphorylation and PLCγ phosphorylation were partially inhibited. In contrast, PF-4 strongly inhibited the Raf1, MEK1/2 and ERK1/2 activation stimulated by both VEGF165 and VEGF121. Thus, PF-4 inhibited the MAP kinase pathway independently of KDR activation, showing that PF-4 exerts inhibitory effects on VEGF121-induced proliferation downstream from the receptor. Presumably this inhibition occurs at/or upstream from Raf1 and downstream from PLCγ. We found the chemokine receptor CXCR3B, a putative PF-4 receptor , in small amounts in HUVEC. However, it does not appear to be involved in the inhibitory effects of PF-4 on proliferation and MAP kinase inhibition.
Materials and methods
Recombinant human PF-4 was supplied by Serbio (Gennevilliers, France). [Methyl-3H]thymidine (TdR) was obtained from ICN Biomedical Inc. (Costa Messa, CA, USA). Cell culture medium, fetal bovine serum, human serum and SuperScript II Reverse Transcriptase were purchased from Invitrogen (Cergy Pontoise, France). VEGF165, VEGF121 and anti-CXCR3 Igs (clone 49801.111) were purchased from R & D Systems (Minneapolis, MN, USA). Anti-ERK2, anti-KDR and nonimmune Igs were supplied by Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA), anti-active (pTEpY) ERK Ig by Promega (Madison, WI, USA), anti-active MEK1/2 (phospho-Ser217/221) by Cell Signaling Technology (Beverly, MA, USA). Anti-PLCγ1, anti-phosphotyrosine (4G10) Igs and the Raf1 immunoprecipitation kinase cascade assay kit were obtained from Upstate Biotechnology (Lake Placid, NY, USA). Anti-CD-31 Ig and the isotype control were obtained from Immunotech (Luminy, France).
HUVEC were isolated from human umbilical veins by collagenase digestion and were cultured in M199 medium/15 mm Hepes, supplemented with 15% (v/v) fetal bovine serum, 5% (v/v) human serum, 2 ng·mL−1 FGF2, 2 mm glutamine, 50 IU·mL−1 penicillin, 50 µg·mL−1 streptomycin and 125 ng·mL−1 amphotericin B, in gelatin-coated flasks at 37 °C in an atmosphere containing 5% CO2. All experiments were carried out between passages 2 and 3. Umbilical cords were obtained through local maternity units (Lariboisière Hospital and Saint Isabelle Clinic) under approval, and with appropriate understanding and consent of the subjects.
HUVEC were seeded at 20 000 cells per well in M199 supplemented with 15% (v/v) fetal bovine serum, 5% (v/v) human serum and 2 ng·mL−1 FGF2. After one day of culture, the cells were deprived of serum for 24 h, then cultured for a further 20 h in the presence of VEGF165 or VEGF121 (10 ng·mL−1) and various concentrations of PF-4 (0–10 µg·mL−1) and/or anti-CXCR3 or nonimmune Igs (40 µg·mL−1). Finally, cells were incubated for 16 h with 1 µCi of [3H]TdR per dish. The [3H]TdR incorporated into the cells was counted with a liquid scintillation β-counter (Beckman Coulter Scintillation Counter LS 6500, Fullerton, CA, USA).
Cells were treated with VEGF165 or VEGF121 in the presence or absence of PF-4 (10 µg·mL−1), then lysed in RIPA buffer . Insoluble material was removed by centrifugation at 4 °C for 10 min at 14 000 g. Supernatants were incubated overnight at 4 °C with various antibodies recognizing KDRs (4 µg·mL−1) or PLCγ1 (6 µg·mL−1). The antigen–antibody complexes purified with the µMACS starting kit (Miltenyi Biotec, Bergisch Gladbach, Germany) were separated by SDS/PAGE in 10% acrylamide gels and transferred to nitrocellulose membranes.
Western blot analysis
Protein lysates and immunoprecipitates were separated by SDS/PAGE in 10% acrylamide gels and transferred to nitrocellulose membranes. The membranes were probed with antibodies against ERK-P (1 : 15 000), total ERK (1 : 15 000), phosphotyrosine (1 : 5000), KDR (1 : 1000), PLCγ (1 : 2000), or MEK-P (1 : 1000). The membranes were washed in Tris buffered saline, 0.1% (v/v) Tween-20 and then incubated with horseradish peroxidase-coupled secondary antibodies. Antigen–antibody complexes were detected with the enhanced chemiluminescence system (ECL, Amersham Pharmacia Biotech, Buckinghamshire, UK).
Raf kinase assays
Raf1 activity was measured using the Upstate Biotechnology kit, according to the manufacturer's instructions. Briefly, the serine/threonine kinase, Raf1 was immunoprecipitated with an anti-Raf1 Ig coupled to protein G Sepharose beads. Kinase reactions were performed in vitro by adding inactive GST–MEK1, inactive GST–ERK2, [32P]ATP[γP] and myelin basic protein (MBP) to immunoprecipitated material and incubating for 30 min at 30 °C. [32P]MBP was quantified with a liquid scintillation β-counter (Beckman Coulter Scintillation Counter LS 6500, Fullerton, CA, USA).
RT-PCR experiments were performed with 0.3 µg total mRNA obtained from primary cultures of HUVEC, using the SuperScript II one-step RT-PCR kit according to the manufacturer's instructions. The following primers were used: CXCR3B (forward) 5′-TGCCAGGCCTTTACACAGC-3′; (reverse) 5′-TCGGCGTCATTTAGCACTTG-3′. GAPDH (forward) 5′-CCACCCATGGCAAATTCCATGGCA-3′; (reverse) 5′-TCTAGACGGCAGGTCAGGTCCACC-3′.
Cells were removed from culture dishes by adding 5 mm EDTA in phosphate buffered saline and collecting the resulting suspension. We incubated 300 000 cells for 30 min at room temperature with phycoerythrin-conjugated specific or isotype control antibody. Finally, cells were washed and a total of 104 events were analysed on a FACScalibur cytofluorimeter (Becton Dickinson), using cellquest software.
Effect of PF-4 on the endothelial cell proliferation induced by VEGF121 and VEGF165
We first investigated the effects of VEGF165 and VEGF121 on [3H]TdR incorporation into HUVEC. In the presence of VEGF165 (10 ng·mL−1), [3H]TdR incorporation was 380 ± 33% (153 942 ± 13 401 c.p.m.) that of the control with no growth factor (100%: 40 414 ± 2961 c.p.m.) (Fig. 1A). VEGF121 (10 ng·mL−1) increased [3H]TdR uptake to a lesser extent, to only 220 ± 7% (89 238 ± 3164 c.p.m.) of control levels (Fig. 1A). We then tested the effects of various concentrations of PF-4 (1 to 10 µg·mL−1) on [3H]TdR. At a PF-4 concentration of 10 µg·mL−1, VEGF165 and VEGF121 induced DNA synthesis by only 25% and 20%, respectively, of the maximum value obtained with VEGF165 or VEGF121 alone (100%) (Fig. 1B).
These observations confirm that (a) VEGF165 and VEGF121 promote DNA synthesis in HUVEC, with VEGF121 being less potent than VEGF165 and (b) PF-4 inhibits the DNA synthesis induced by VEGF165 and VEGF121.
PF-4 does not affect VEGF121-induced KDR phosphorylation
We analysed the effects of PF-4 on the signalling pathways induced by VEGF165 and VEGF121 by investigating the effect of PF-4 on KDR activation. VEGF165 and VEGF121 (10 ng·mL−1) induced significant phosphorylation of the tyrosine residues of the KDR (Fig. 2A); VEGF121 had a weaker effect (48%) than VEGF165 (100%) (Fig. 2A,B). In the presence of PF-4 (10 µg·mL−1), VEGF165-induced phosphorylation of the KDR was inhibited by 45%, whereas VEGF121-induced phosphorylation was unaffected (Fig. 2A,B). Interestingly, the level of KDR phosphorylation induced by VEGF121 in the absence of PF-4 was similar to that obtained with a combination of VEGF165 (10 ng·mL−1) and PF-4 (10 µg·mL−1).
PF-4 has no effect on VEGF121-induced PLCγ phosphorylation
PLCγ has been reported to be a downstream target of the tyrosine kinase activity of the KDR and to be involved in VEGF-induced DNA synthesis . PLCγ phosphorylation was induced by VEGF165 (10 ng·mL−1) and VEGF121 (10 ng·mL−1) and the level of phosphorylation of PLCγ was lower with VEGF121 (30%) than with VEGF165 (100%) (Fig. 3A,B). PF-4 inhibited VEGF165-induced PLCγ phosphorylation by 66% (Fig. 3B). In contrast, the phosphorylation of PLCγ induced by VEGF121 was unaffected by 10 µg·mL−1 PF-4 (Fig. 3A,B).
PF-4 inhibits VEGF121- and VEGF165-induced MAP kinase pathway activation
We then investigated the effect of PF-4 on the ERK activation necessary for VEGF-induced proliferation of HUVEC [30,32]. In the absence of PF-4, ERK phosphorylation was induced by VEGF165 and VEGF121(Fig. 4A). The level of ERK phosphorylation was higher following VEGF165 (100%) stimulation than following VEGF121 stimulation (45%) (Fig. 4A,B). The degree of ERK phosphorylation correlated with the mitogenic effect upon VEGF165 treatment of HUVEC. In the presence of PF-4 (10 µg·mL−1), the phosphorylation of ERK induced by VEGF165 and VEGF121 was strongly inhibited, only reaching 18% and 1% of maximum stimulation, respectively (VEGF165 alone: 100%) (Fig. 4B). Thus, PF-4 acts on the MAP kinase pathways induced by VEGF121 and VEGF165.
These results were confirmed by kinetic studies of ERK activation. The ERK phosphorylation induced by VEGF165 and VEGF121 was maximal between 10 and 15 min of stimulation and decreased thereafter (Fig. 4C,E). PF-4 strongly decreased ERK phosphorylation, to only 34% (VEGF165) and 22% (VEGF121) of maximal stimulation (Fig. 4D,F).
PF-4 inhibits the VEGF121- and VEGF165-induced activation of MEK1/2 and Raf1
As ERK1/2 are phosphorylated directly and activated by MEK1/2, we investigated the phosphorylation state of these kinases in the presence of PF-4. As previously reported with ERK1/2, VEGF165 induced stronger phosphorylation of MEK1/2 (100%) than did VEGF121 (50%) (Fig. 5A,B). MEK1/2 phosphorylation induced by VEGF165 and VEGF121 was strongly inhibited in the presence of PF-4 (10 µg·mL−1) reaching, respectively, 16% and 4% of maximum stimulation (VEGF165 alone: 100%) (Fig. 5A,B). Thus, PF-4 inhibits the phosphorylation not only of ERK1/2, but also of MEK1/2, induced by VEGF121 and VEGF165.
We investigated the effect of PF-4 on Raf1 kinase, which is responsible directly for MEK1/2 phosphorylation. We found that the Raf1 activity induced by VEGF165 and VEGF121 was strongly inhibited by PF-4 (10 µg·mL−1) (Fig. 5C). The inhibition was similar for VEGF165- and VEGF121-induced Raf1 activities.
CXCR3 blocking antibody had no effect on PF-4 activity
The results described above suggest that PF-4 affected the VEGF165 and VEGF121-induced MAP kinase pathway and proliferation by an intracellular mechanism involving the modulation of Raf1 activity. The inhibition of the MAP kinase pathway by an intracellular mechanism induced by PF-4 suggests that this chemokine may induce angiostatic activity via a specific receptor. Recent data have suggested that PF-4 can bind a newly cloned chemokine receptor isoform named CXCR3B . We therefore studied the involvement of this receptor in the inhibition, by PF-4, of VEGF-induced MAP kinase activation and proliferation of HUVEC. We tested for CXCR3B mRNA in HUVEC by RT-PCR. We detected CXCR3B mRNA in HUVEC and in skeletal muscle, used as a positive control (Fig. 6A). However, FACS analysis, using an antibody that recognizes both CXCR3A and CXCR3B, indicated that only 10% of HUVEC cells were positive (Fig. 6B); all HUVEC cells expressed CD-31 (Fig. 6B). Despite few cells expressing this receptor on their surface, we investigated whether CXCR3B mediated the antiangiogenic effects of PF-4 in our model. An antibody blocking CXCR3 , was unable to reverse the inhibitory effects of PF-4 (5 µg·mL−1) on proliferation or MAP kinase activity (Fig. 6C,D), suggesting that in our model, PF-4 does not act through this receptor (CXCR3).
We recently showed that the antiangiogenic chemokine, PF-4, inhibits FGF2-induced cell proliferation via an intracellular mechanism . In this study, we investigated the effect of PF-4 on another angiogenic factor of prime importance, VEGF, and compared the mechanisms by which PF-4 inhibits the DNA synthesis induced by VEGF165 and VEGF121.
The DNA synthesis induced by VEGF165 and VEGF121 was strongly inhibited by PF-4 (10 µg·mL−1) in HUVEC. Previous work showed that PF-4 efficiently inhibits the binding of VEGF165 to its receptor, but not that of VEGF121. Thus, PF-4 may disrupt the KDR-mediated signal transduction induced by VEGF121 by means of an unknown mechanism that does not involve the disruption of VEGF121 binding . We find that PF-4 acts downstream from receptor activation under conditions of VEGF121 stimulation. In contrast, PF-4 also acts at the receptor level for VEGF165. Indeed, the level of tyrosine phosphorylation of the KDR and of PLCγ decreased significantly (45% and 66%, respectively) following the addition of PF-4 (10 µg·mL−1). This is consistent with partial inhibition of the binding of VEGF165 to its receptor . However, the levels of tyrosine phosphorylation of the KDR and PLCγ were not affected by PF-4 in conditions of VEGF121 stimulation. Thus, PF-4 disrupts KDR-mediated signal transduction at a postreceptor level following VEGF121 stimulation.
We investigated at which step VEGF165- and VEGF121-induced intracellular signalling is a target of PF-4 inhibition. Activation of the MAP kinases, ERK1/2, is important for the proliferation of HUVEC . We therefore focused on the effect of PF-4 on the kinases involved in the signalling pathways leading to ERK1/2 stimulation. The level of phosphorylation of Raf1, MEK1/2 and ERK1/2 induced by both growth factors, VEGF165 and VEGF121, was strongly decreased by PF-4. Thus, PF-4 acts directly on or upstream from Raf1 and downstream from PLCγ in the signalling cascade induced by VEGF121. This mechanism may be also involved in the inhibition of VEGF165-induced ERK activation. Indeed, PF-4 only partially inhibited the phosphorylation of KDR and PLCγ whereas the phosphorylation of Raf1, MEK1/2 and ERK1/2 activity was almost abolished.
How PF-4 regulates the activation of the MAP kinase pathway downstream from the KDR is currently under investigation. PKC and Raf1, both stimulated by VEGF and downstream from PLCγ, may be involved [28,29]. PKC is involved in MAP kinase activation by VEGF [29,31,35] but not by FGF2 [36–38]. As PF-4 inhibits both VEGF- and FGF2-induced MAP kinase phosphorylation , PF-4 may act on a target common to the FGF2 and VEGF signalling pathways. Thus, PKC does not seem to be a good candidate.
Raf1 is a key signalling molecule for both VEGF and FGF2. It is a serine/threonine kinase, regulated by phosphorylation of serine and tyrosine residues [39–43]. Ser259 is the main inhibitory site of Raf1, but the phosphorylation of this residue is not affected by PF-4 (data not shown). Thus, it is unclear how PF-4 affects Raf1 activity in HUVEC. Increases in cAMP levels and the activation of the cAMP-dependent protein kinase A (PKA) may be involved . Indeed, PKA inhibits the MAP kinase pathway by blocking Raf1 activity in many cell systems [45–47]. Moreover, PF-4 increases cAMP levels in human microvascular endothelial cells (HMEC-1 cell line) transfected with a construct encoding a new chemokine isoform receptor – CXCR3B – the only seven-transmembrane chemokine receptor able to bind PF-4 with high affinity . Alternative splicing of the CXCR3 mRNA gives rise to two different chemokine receptors: CXCR3A and CXCR3B . However, only 10% of HUVEC expressed CXCR3 (CXCR3A plus CXCR3B) on the cell surface in serum deprivation conditions. We evaluated the involvement of CXCR3 in the inhibitory effect of PF-4, using a blocking antibody . Unlike for ACHN cells under the same conditions , we were unable to reverse the inhibitory effect of PF-4 on the MAP kinase pathway and on HUVEC proliferation. Similar results were obtained with lower concentrations of PF-4 (0.5 to 5 µg·mL−1) and various concentrations (5 to 40 µg·mL−1) of blocking antibody (data not shown). This absence of effect could be explained by the restricted expression of CXCR3 in HUVEC: FACS analysis indicates that 100% of ACHN cells express CXCR3 on their surface , whereas only 10% of HUVEC were positive. Further experiments will be required to fully determine the role of CXCR3B in HUVEC, nevertheless, our findings suggest that this chemokine receptor isoform is probably not central to PF-4 induced angiostatic activity in our model. Most chemokines bind and activate different chemokine receptor isoforms [48–50], and it would be valuable to determine which bind PF-4 and are expressed in HUVEC. Studies of cAMP modulation in HUVEC upon PF-4 stimulation, and its possible effect on Raf1 inhibition may also be informative.
In conclusion, this report is the first to show that the signal transduction pathways of two isoforms of VEGF (VEGF121 and VEGF165) may be regulated by PF-4 at a postreceptor level. These results, and those for the FGF2 signalling pathway, suggest that a specific mechanism of inhibition is triggered by PF-4, blocking MAP kinase pathway activation. The ability of PF-4 to abolish the proliferation of endothelial cells induced by the two major angiogenic growth factors secreted by tumours – VEGF and FGF2 – may be useful for the development of treatments based on the inhibition of angiogenesis. Any such therapy would however, require a better understanding of the mechanism underlying this effect.
We would like to thank the maternity units of Hôpital Lariboisière and Clinique Saint Isabelle for providing the umbilical cords. This work was supported by IVS and grants from l'Association pour la Recherche sur le Cancer and from Ligue contre le Cancer (contract numbers 5820 and 7566).