Vascular Endothelial Growth Factor (VEGF)-A165-induced Prostacyclin Synthesis Requires the Activation of VEGF Receptor-1 and -2 Heterodimer*

We previously reported that vascular endothelial growth factor (VEGF)-A165 inflammatory effect is mediated by acute platelet-activating factor synthesis from endothelial cells upon the activation of VEGF receptor-2 (VEGFR-2) and its coreceptor, neuropilin-1 (NRP-1). In addition, VEGF-A165 promotes the release of other endothelial mediators including nitric oxide and prostacyclin (PGI2). However, it is unknown whether VEGF-A165 is mediating PGI2 synthesis through VEGF receptor-1 (VEGFR-1) and/or VEGF receptor-2 (VEGFR-2) activation and whether the coreceptor NRP-1 potentiates VEGF-A165 activity. In this study, PGI2 synthesis in bovine aortic endothelial cells (BAEC) was assessed by quantifying its stable metabolite (6-keto prostaglandin F1α, 6-keto PGF1α) by enzyme-linked immunosorbent assay. Treatment of BAEC with VEGF analogs, VEGF-A165 (VEGFR-1, VEGFR-2 and NRP-1 agonist) and VEGF-A121 (VEGFR-1 and VEGFR-2 agonist) (up to 10–9 m), increased PGI2 synthesis by 70- and 40-fold within 15 min. Treatment with VEGFR-1 (placental growth factor and VEGF-B) or VEGFR-2 (VEGF-C) agonist did not increase PGI2 synthesis. The combination of VEGFR-1 and VEGFR-2 agonists did not increase PGI2 release. Pretreatment with a VEGFR-2 inhibitor abrogated PGI2 release mediated by VEGF-A165 and VEGF-A121, and pretreatment of BAEC with antisense oligomers targeting VEGFR-1 or VEGFR-2 mRNA reduced PGI2 synthesis mediated by VEGF-A165 and VEGF-A121 up to 79%. In summary, our data demonstrate that the activation of VEGFR-1 and VEGFR-2 heterodimer (VEGFR-1/R-2) is essential for PGI2 synthesis mediated by VEGF-A165 and VEGF-A121, which cannot be reproduced by the parallel activation of VEGFR-1 and VEGFR-2 homodimers with corresponding agonists. In addition, the binding of VEGF-A165 to NRP-1 potentiates its capacity to promote PGI2 synthesis.

We previously reported that vascular endothelial growth factor (VEGF)-A 165 inflammatory effect is mediated by acute platelet-activating factor synthesis from endothelial cells upon the activation of VEGF receptor-2 (VEGFR-2) and its coreceptor, neuropilin-1 (NRP-1). In addition, VEGF-A 165 promotes the release of other endothelial mediators including nitric oxide and prostacyclin (PGI 2 ). However, it is unknown whether VEGF-A 165 is mediating PGI 2 synthesis through VEGF receptor-1 (VEGFR-1) and/or VEGF receptor-2 (VEGFR-2) activation and whether the coreceptor NRP-1 potentiates VEGF-A 165 activity. In this study, PGI 2
Stimulation of ECs with VEGF-A 165 can promote prostacyclin (PGI 2 ) synthesis, which is a potent vasodilator and an inhibitor of platelet aggregation (8 -10). Consequently, the imbalance in PGI 2 production can be involved in the pathophysiology of many thrombotic and cardiovascular disorders. The induction of PGI 2 can be mediated upon the activation of different phospholipase A 2 enzymes that catalyze the cleavage of arachidonic acid from membrane glycerophospholipids. Subsequently, arachidonic acid is converted in PGH 2 by the action of two cyclooxygenase (COX) isoforms, either the constitutive form, COX-1, or the inducible form, COX-2. The newly formed PGH 2 then is transformed into PGI 2 by the action of the PGI 2 synthase (11)(12)(13)(14). However, it is unknown whether the members of the VEGF superfamily are mediating PGI 2 synthesis either through VEGFR-1 and/or VEGFR-2 activation and whether NRP-1 is contributing to potentiate VEGF-A 165 -mediated PGI 2 synthesis.
During last few years, we have shown that VEGF-A 165 increases vascular permeability, endothelial P-selectin translo-cation, and neutrophil adhesion upon the synthesis of plateletactivating factor (PAF) by ECs (5,15). We subsequently investigated the contribution of VEGF receptors and assessed that all of these biological activities are mediated through the activation of VEGFR-2 and that these effects are potentiated by the presence of NRP-1 (4,5,16). Consequently, by using VEGF analogs and by regulating VEGF receptor activity, either with selective inhibitors or by antisense treatment, we investigated the contribution of VEGF members and their corresponding receptors on their capacity to promote endothelial PGI 2 synthesis.
Antisense Oligonucleotide Therapy-We also used an antisense oligonucleotide therapy approach to discriminate the contribution of VEGFR-1 and VEGFR-2 on PGI 2 synthesis mediated by VEGF-A isoforms. BAEC were treated with antisense oligonucleotide sequences complementary to bovine VEGFR-1 or VEGFR-2 mRNA (GenBank TM accession numbers X94263 and 94298). Antisense oligonucleotide phosphorothioate backbone sequences targeting bovine VEGFR-1 mRNA (AS-R1: 5Ј-CAA AGA TGG ACT CGG GAG-3Ј) and VEGFR-2 mRNA (AS-R2: 5Ј-GCT GCT CTG ATT GTT GGG-3Ј) or a scrambled phosphorothioate sequence (AS-Scr: 5Ј-TGC TGG CAT GTG CGT TGT-3Ј) (AlphaDNA, Montreal, Quebec, Canada) were used. The antisense oligomers were chosen based on their capacity to selectively abrogate the protein expression of the genes targeted as described previously (16). BAEC were seeded at 5 ϫ 10 5 cells/well in 6-well plates in Dulbecco's modified Eagle's medium, 5% fetal bovine serum, and antibiotics with or without oligomers (5 ϫ 10 Ϫ7 M/daily) up to 3 days post-confluence. Culture medium was removed, cells were rinsed and stimulated in HBSS/HEPES ϩ CaCl 2 (5 mM) with PBS or VEGF-A isoforms, and PGI 2 synthesis was quantified as described above.
Preparation of Glutathione S-Transferase (GST)-VEGF-A 165 Exon 7 Fusion Protein-To evaluate the possible potentiating effect of NRP-1 on VEGF-A 165 -induced PGI 2 synthesis, we produced a GST fusion protein encoding exon 7 of human VEGF-A 165 (GST-Ex7). The construct of GST fusion protein-exon 7 was generously provided by Dr. Shay Soker, Wake Forest University, Winston-Salem, NC. Escherichia coli (DH5␣) were transformed with pGEX-2TK or p2TK-exon 7 vectors to produce GST and GST-Ex7 proteins. The recombinant proteins were purified from bacterial lysates using glutathione and heparin affinity chromatographies as described previously (5,6).
Western Blot Analyses of VEGF Receptors Expression and Phosphorylation-BAEC were cultured up to 3 days post-confluence, and cells were rinsed, incubated on ice, and stimulated in HBSS/HEPES ϩ CaCl 2 (5 mM) plus 1 mg/ml bovine serum albumin. In some experiments, BAEC were pretreated with PBS or VEGF receptor inhibitors (SU1498 or VTK) 15 min prior to the addition of VEGF-A 165 or VEGF-A 121 and the cells were kept on ice for an additional 15 min. The cells then were stimulated for 7.5 min at 37°C and placed again on ice. In another set of experiments, BAEC were stimulated as above with VEGF analogs only. Upon stimulation, the medium was removed, cells were washed, and lysates were prepared. Western blot analyses were performed as described previously (4,5,16). The primary antibodies used were mouse monoclonal anti-human VEGFR-1 (clone Flt-11, Sigma), polyclonal rabbit anti-mouse VEGFR-2, and goat anti-human NRP-1 IgG antibodies (Santa Cruz Biotechnology Inc., Santa Cruz, CA). Membranes were stripped using Re-Blot Plus Strong stripping solution (Chemicon International, Temecula, CA) for 20 min and reprobed with a mouse monoclonal anti-phosphotyrosine IgG (clone 4G10, 1:4000 dilution, Upstate Biotechnology Inc., Lake Placid, NY) to determine VEGFR-1 and VEGFR-2 phosphorylation. Kaleidoscope molecular weight markers (Bio-Rad) were used as molecular mass standards for SDS-PAGE immunoblotting experiments. Immunoreactive bands were visualized by ECL, digitized using a two-dimensional gel scanner, and quantified using Quantity One software (Bio-Rad).
In another set of experiments, we assessed by Western blot analyses the expression and activation of selective enzymes involved in the cell signaling pathway leading to PGI 2 synthesis. Studies were performed as described above. Primary antibodies used for immunoprecipitations and Western blot analyses were as follows: rabbit polyclonal antihuman phospho-p42/44 MAPK and anti-human phospho-cPLA 2 (Cell Signaling Technology Inc., Beverly, MA) and mouse monoclonal antiovine COX-1 and mouse monoclonal anti-human COX-2 (Cayman Chemicals) IgGs. The membranes then were stripped and reprobed with rabbit anti-rat p42/44 MAPK (Cell Signaling) and mouse monoclonal anti-human cPLA 2 (Santa Cruz Biotechnology).
Western Blot Analysis of PGI 2 Synthase S-Nitrosylation, and Cyclooxygenases-1 and -2-BAEC were cultured up to 3 days post-confluence and treated prior to stimulation as described above. Cells were incubated on ice for 30 min with either VEGF-A 165 or VEGF-C (10 Ϫ9 M) and then stimulated for 5-15 min at 37°C and returned to ice. Upon stimulation, the medium was removed, cells were washed, and lysates were prepared. Immunoprecipitation of cell lysate was performed with a rabbit polyclonal anti-bovine PGI 2 synthase IgG (Cayman Chemicals). Samples were separated on a 10% SDS-PAGE, and Western blot analyses were performed as described previously (4,5,16). A mouse monoclonal anti-nitrotyrosine IgG (Cayman Chemicals) was used to assess the S-nitrosylation level of PGI 2 synthase. The membranes then were stripped and reprobed with rabbit polyclonal anti-bovine PGI 2 synthase IgG as described above.
Statistical Analysis-Data are presented as the mean ϩ S.E. Statistical comparisons were made by analysis of variance followed by a Bonferroni's t test for multiple comparisons. Differences were considered significant when p Ͻ 0.05.
To determine which VEGF receptors are involved in PGI 2 synthesis, we used selective VEGF analogs. Treatment with VEGF-A 121 isoform (10 Ϫ11 -10 Ϫ8 M), which like VEGF-A 165 binds to VEGFR-1 and VEGFR-2 but not to NRP-1 coreceptor, induced a significant but reduced PGI 2 synthesis (40-fold increase at 10 Ϫ9 M) compared with VEGF-A 165 (Fig. 1C). These latter data suggest that NRP-1 coreceptor might contribute to potentiate VEGF-A 165 capacity to promote PGI 2 synthesis. To support this hypothesis, we treated BAEC with PBS, VEGF-A 165 , and VEGF-A 121 and confirmed by Western blot analysis the capacity of VEGF-A 165 as opposed to VEGF-A 121 -or PBStreated cells to promote VEGFR-2⅐NRP-1 complex formation ( Fig. 2A). BAEC then were pretreated with GST-Ex7 of human VEGF-A 165 to block the interaction of VEGF-A 165 with NRP-1. Exon 7 encodes a domain not present in VEGF-A 121 that is responsible for the binding of VEGF-A 165 to NRP-1 (3,5). Pretreatment of BAEC with GST-Ex7 (up to 10 Ϫ7 M) 15 min prior to stimulation with VEGF-A 165 (10 Ϫ9 M) reduced PGI 2 synthesis to the level induced by VEGF-A 121 (Fig. 2B). Pretreatment with GST (up to 10 Ϫ7 M), without the exon 7 insert did not alter VEGF-A 165 -induced PGI 2 synthesis, and neither GST-Ex7 nor GST (up to 10 Ϫ7 M) altered significantly the basal level of PGI 2 synthesis or VEGF-A 121 -induced PGI 2 synthesis (Fig. 2B).
Because VEGF-A 165 and VEGF-A 121 are both capable of activating VEGFR-1 and VEGFR-2, we treated BAEC with selective analogs for VEGFR-1 (PlGF and VEGF-B) and VEGFR-2 (VEGF-C) to verify the contribution of each receptor under their homodimeric conformations on PGI 2 synthesis. Treatment with PlGF, VEGF-B, or VEGF-C (10 Ϫ11 -10 Ϫ8 M) for 15 min did not promote the release of PGI 2 (Fig. 3, A-C). To assess whether a parallel activation of VEGFR-1 and VEGFR-2 homodimers may induce PGI 2 release, BAEC were treated with the combination of VEGFR-1 analogs PlGF or VEGF-B with VEGFR-2 analog VEGF-C at 10 Ϫ9 M and for 15 min. Such a combination did not increase PGI 2 synthesis (Fig. 3D).
Phosphorylation of VEGF Receptors by Corresponding VEGF Analogs-Because VEGF-A 165 and VEGF-A 121 were the only VEGF analogs capable of mediating endothelial PGI 2 synthesis, we performed Western blot analyses to confirm the expression of VEGFR-1 and VEGFR-2 and the capacity of VEGF analogs to activate them. BAEC were treated with VEGF analogs for 7.5 min, which is the suitable time to detect the phosphorylation of VEGF receptors as described previously (16). Cell lysates were immunoprecipitated either with anti-VEGFR-1 or anti-VEGFR-2 IgGs. VEGFR-1 and VEGFR-2 protein expression in BAEC was detected by immunoblotting, which is in agreement with previous reports (Fig. 4, A and B, upper bands) (17). The membranes were stripped, and the detection of VEGFR-1 and VEGFR-2 phosphorylation was performed by reprobing the membranes with anti-phosphotyrosine IgG. Treatment with VEGF-A 165 and VEGF-A 121 (10 Ϫ9 M) increased the phosphorylation of VEGFR-1 by 10.6-and 7.3-fold as compared with PBS-treated cells, whereas equivalent treat- ment with PlGF or VEGF-B did not increase VEGFR-1 phosphorylation (Fig. 4A, lower panel). Although we could not detect the phosphorylation of VEGFR-1-mediated by PlGF or VEGF-B, they were not deprived of biological activities since they were capable of promoting endothelial P-selectin translocation (data not shown and as described previously) (5). Next, we assessed the capacity of VEGF-A 165 , VEGF-A 121 , and VEGF-C (10 Ϫ9 M) to mediate VEGFR-2 phosphorylation. Such treatment increased the phosphorylation of VEGFR-2 by 57.8-, 24.7-, and 5.8-fold compared with PBS-treated cells (Fig. 4B,  lower panel). PGI 2 Synthesis Requires VEGFR-1 and VEGFR-2 Heterodimerization-Our data demonstrate that the activation of VEGFR-1 or VEGFR-2 homodimers alone or in parallel with selective VEGFR-1 or VEGFR-2 analogs did not promote PGI 2 synthesis. Consequently, we speculated that VEGF-A isoforms induce PGI 2 synthesis through the activation of VEGFR-1/R-2 heterodimer. By Western blot analysis, we observed in PBStreated cells that VEGFR-1 and VEGFR-2 subunits can constitutively be present under heterodimeric VEGFR-1/R-2 state and that one treatment either with VEGF-A 165 or VEGF-C did not modulate VEGFR-1/R-2 dimerization (Fig. 5). Next, to demonstrate that PGI 2 synthesis mediated by VEGF-A 165 and VEGF-A 121 is driven through the activation of VEGFR-1/R-2 heterodimer, BAEC were treated with selective antisense oligomers targeting VEGFR-1 or VEGFR-2 mRNA. We showed previously that such an approach at the concentration used (5 ϫ 10 Ϫ7 M/daily) abrogated selectively the protein expression of VEGFR-1 or VEGFR-2 by over 90% and the biological activities investigated by 80 -100% (16). Treatment with selective antisense oligomers targeting VEGFR-1 (AS-R1) or VEGFR-2 (AS-R2) mRNA reduced, by 79 and 71%, the synthesis of PGI 2 mediated by VEGF-A 165 and, by 73 and 62%, the synthesis of PGI 2 mediated by VEGF-A 121 , respectively (Fig. 6A). As negative control, BAEC were treated with a scrambled oligomer sequence, which did not decrease significantly the level of PGI 2 synthesis mediated by VEGF-A 165 and VEGF-A 121 . In addition, the treatment of BAEC with antisense or scrambled oligomers did not affect the basal level of PGI 2 synthesis in PBS-treated cells (Fig. 6A).
To verify the selectivity of VEGF receptor inhibitors, we assessed their corresponding inhibitory effect on VEGFR-1 and VEGFR-2 phosphorylation mediated by VEGF-A 165 . Pretreatment of BAEC with SU1498 (5 ϫ 10 Ϫ6 M) 15 min prior to stimulation with VEGF-A 165 (10 Ϫ9 M, 7.5 min) did not affect the phosphorylation of VEGFR-1 but prevented the phosphorylation of VEGFR-2. Pretreatment with VTK prevented the phosphorylation of VEGFR-1 and VEGFR-2 mediated by VEGF-A 165 (Fig. 7) by 100 and 83%.
Denitrosylation of Prostacyclin Synthase Is Required to Promote PGI 2 -Under quiescent state, PGI 2 synthase is S-nitrosylated, thus preventing its capacity to convert its substrate, PGH 2 , into PGI 2 (27). In a previous study, we have shown that VEGF-A 165 and VEGF-C are capable of mediating PAF synthesis upon VEGFR-2 homodimer activation (4,16). In the current study, we also observed that VEGF-A 165 and VEGF-C are both capable to promote the phosphorylation p42/44 MAPK and cPLA 2 (data not shown), which are essential to promote arachidonic acid release (8,28). However, VEGF-C as opposed to VEGF-A 165 cannot induce PGI 2 synthesis. Consequently, we hypothesized that VEGF-C may not be capable of promoting the S-denitrosylation of PGI 2 synthase, which is required for PGI 2 production (27,29).
Thus, we assessed, by Western blot analysis, the capacity of VEGF-A 165 and VEGF-C to regulate the level of PGI 2 synthase nitrosylation. Treatment with VEGF-A 165 (10 Ϫ9 M) induced a rapid/immediate S-denitrosylation of PGI 2 synthase by 30%, which returned to its basal nitrosylated state after 15 min of treatment. At the opposite, treatment with VEGF-C (10 Ϫ9 M) had no such effect (Fig. 9). DISCUSSION Previous studies reported the capacity of VEGF-A 165 to promote PGI 2 synthesis. However, there were no data defining the contribution of VEGF receptor(s) (8,30,31). In the current study, we observed that only VEGF-A isoforms (VEGF-A 165 and VEGF-A 121 ) were able to promote an acute endothelial PGI 2 synthesis, whereas the stimulation of BAEC with selective VEGFR-1 (PlGF or VEGF-B) or VEGFR-2 (VEGF-C) agonists, alone or combined, had no such effect. In addition, VEGF-A 165 was approximately twice potent as VEGF-A 121 to promote PGI 2 synthesis. Pretreatment of ECs with a GST-Ex7 fusion protein, which prevents the binding of VEGF-A 165 to NRP-1, reduced the capacity of VEGF-A 165 to promote PGI 2 synthesis to the level mediated by VEGF-A 121 , which does not bind to NRP-1. Together, these data suggest that the activation of VEGFR-1 and VEGFR-2 homodimers by their corresponding analogs is not sufficient to promote PGI 2 synthesis and depends on the activation of the VEGFR-1/R-2 heterodimer. Furthermore, NRP-1 potentiates the capacity of VEGF-A 165 to promote PGI 2 synthesis.
To assess the requirement of VEGFR-1/R-2 activation for promoting PGI 2 synthesis mediated by VEGF-A isoforms, we used antisense gene therapy and pharmacological approaches to regulate the expression and the activation of VEGF receptors. In a first series of experiments, BAEC were treated with antisense oligomers targeting selectively VEGFR-1 or
VEGFR-2 mRNA that we previously had defined for their capacity to down-regulate selectively their corresponding protein expression by over 90% and related biological activities (16). Such treatment with antisense oligomers for VEGFR-1 or VEGFR-2 mRNA abrogated by over 70% the synthesis of PGI 2 mediated by VEGF-A 165 and VEGF-A 121 (Fig. 6). In another set of experiments, BAEC were pretreated with selective VEGF receptor inhibitors. Pretreatment of BAEC with a VEGFR-2 inhibitor, SU1498, prevented VEGFR-2 activation by VEGF-A isoforms and PGI 2 synthesis. Together, these data demonstrate that the activation of both receptors under heterodimeric state (VEGFR-1/R-2) is required for PGI 2 synthesis. Indeed, it cannot be due either to an independent or parallel activation of VEGFR-1 and/or VEGFR-2 homodimers by VEGF-A isoforms, because the blockade of VEGFR-1 or VEGFR-2 expression and corresponding activation were sufficient to prevent PGI 2 synthesis. Furthermore, parallel activation of VEGFR-1/R-1 and VEGFR-2/R-2 by selective agonists did not promote PGI 2 synthesis.
Although we detected the phosphorylation of VEGFR-1 mediated by VEGF-A isoforms, we could not detect its phosphorylation upon a treatment with PlGF and VEGF-B. Consequently, one might argue that this could explain why PlGF and VEGF-B were unable to promote PGI 2 synthesis. However, this hypothesis does not stand because we observed in a different study that PlGF and VEGF-B at the same concentrations and within the same time period were both capable to promote a significant increase of endothelial P-selectin translocation, despite our incapacity to detect VEGFR-1 phosphorylation (5). Our data are also in agreement with previous reports that have shown that PlGF and VEGF-B were capable of promoting specific biological activities, namely on endothelial and monocyte tissue factor production and migration of monocytes, despite the fact that VEGFR-1 phosphorylation was undetectable (32)(33)(34)(35)(36). This could also be explained by the fact that the tyrosine kinase activity of VEGFR-1 is one order of magnitude lower than that of VEGFR-2 and that, in function of the ligands used, VEGFR-1 can autophosphorylate differently, rendering difficult the detection of its autophosphorylation (37,38).
The lack of PGI 2 synthesis upon stimulation with VEGF-C also was not due to its incapacity to activate VEGFR-2, because VEGF-C was capable of promoting VEGFR-2 phosphorylation (Fig. 4) and selective biological activities including endothelial P-selectin translocation, endothelial cell migration and proliferation, and PAF synthesis (4,5,16,39). Together, these observations strengthened our hypothesis that VEGF-A isoforms require the activation of VEGFR-1/R-2 heterodimer to support PGI 2 synthesis.
Our observations are also supported by recent studies that have demonstrated that, in native unstimulated and VEGF-A 165 -treated ECs, VEGFR-1 was consistently detected in anti-VEGFR-2 immunoprecipitates, indicating that both receptors spontaneously form heterodimer (38). Porcine aortic endothelial cells (PAEC) transfected with VEGFR-1 and VEGFR-2 provided as well the heterodimerization of VEGFR-1/R-2. Stimulation of all three PAEC-transfected cell lines expressing VEGFR-1, VEGFR-2, and VEGFR-1/R-2 with VEGF-A 165 resulted in signal transduction with different efficiencies and biological activities (40). For instance, migration of PAEC coexpressing VEGFR-1/R-2 toward VEGF-A 165 was more efficient than migration of PAEC expressing VEGFR-2 alone, even though similar number of VEGFR-2 subunits were expressed in transfected PAEC (40). These data suggest that the signal transduction properties of VEGFR-2 are affected by its dimerization with VEGFR-1 and its transphosphorylation. This is in agreement with a recent study reporting that VEGF-A 165 can induce a strong phosphorylation of VEGFR-1 tyrosine residue Tyr-1213 and to lesser extent Tyr-1242 and Tyr-1333, whereas PlGF induced the phosphorylation of Tyr-1309 but not Tyr-1213 (38). Such differences in the activation of VEGF receptors by various agonists termed "agonist trafficking" might explain the distinct biologic activities of VEGF-A 165 and its analogs. Because VEGF-A 165 and VEGF-A 121 are the only VEGF analogs capable to bind VEGFR-1/R-2 heterodimer, they might provide an exclusive transphosphorylation of VEGFR-1 and VEGR-2 subunits that appears to be essential to govern cell signaling leading to PGI 2 synthesis.
In previous studies, we have shown that VEGF-A 165 induces PAF synthesis upon the activation of VEGFR-2/R-2 homodimer, which is potentiated by the presence of NRP-1 coreceptor (4,16). Furthermore, the synthesis of PAF requires the activation of both p38 and p42/44 MAPKs and subsequent activation of sPLA 2 -V (20,21). Interestingly, the stimulation of VEGFR-1 did not promote PAF synthesis and was not required to support VEGF-A 165 -mediated PAF synthesis through VEGFR-2 activation (4,16).
In order to assess the different contribution of VEGF receptors for the induction of PAF and PGI 2 synthesis mediated by VEGF-A 165 , we investigated the cell signaling pathways involved in PGI 2 synthesis. Previous studies highlighted the contribution of p42/44 and/or p38 MAPK on PGI 2 synthesis (8,(41)(42)(43)(44)(45). To discriminate the potential role played by both MAPKs, we used specific inhibitors, namely PD98059, which inhibits MEK, the enzyme upstream and responsible for p42/44 MAPK activation, and SB203580, a specific inhibitor of p38 MAPK activation (20). In our study, the blockade of p42/44 MAPK activation prevented the induction PGI 2 synthesis mediated by VEGF-A 165 , whereas the blockade of p38 MAPK activity had no such effect. We previously reported that PAF synthesis mediated by VEGF-A 165 requires the activation of sPLA 2 as opposed to cPLA 2 . However, because it is well established that p42/44 MAPK can promote the phosphorylation of cPLA 2 (8,46) and that p42/44 MAPK activation is essential for PAF synthesis (20), we assessed the contribution of both PLA 2 on PGI 2 synthesis. The blockade of cPLA 2 prevented completely the synthesis of PGI 2 , whereas the inhibition of sPLA 2 induced a non-significant decrease of PGI 2 synthesis.
Because VEGF-C is also capable of promoting VEGFR-2 activation and PAF synthesis but not PGI 2 synthesis, we investigated by Western blot analyses its capacity to induce the phosphorylation p42/44 MAPK and cPLA 2 . In both cases, VEGF-C promoted the activation of these enzymes, suggesting that the activation of these enzymes is not the rate-limiting step for the incapacity of VEGF-C to mediate PGI 2 synthesis. The activation of cPLA 2 catalyzes the cleavage of arachidonic acid from the sn-2 position of phospholipids, which is then converted into PGH 2 by COX-1 and/or COX-2, which serves as substrate for PGI 2 synthase and PGI 2 production (11)(12)(13)41). By Western blot analysis, we confirmed the presence of constitutive COX-1, whereas COX-2 was not present. This is in agreement with previous reports that have shown in quiescent and unstimulated ECs that COX-2 isoform is not present and that its up-regulation by diverse cytokines including VEGF is achieved at least two hours after stimulation (12,14,(47)(48)(49). Therefore, only COX-1 isoform was involved and its contribution was essential since its blockade with indomethacin abrogated PGI 2 synthesis. Furthermore, because COX-1 activity is not repressed by VEGF isoforms (14) and because there is no alternative pathway for PGI 2 synthesis, this brings us to suggest that COX-1 activation cannot be responsible for the difference between VEGF-C and VEGF-A 165 to promote PGI 2 synthesis. Thus, the rate-limiting step can be due to a possible difference in their capacity to promote PGI 2 synthase activity. Recently, a new post-translational modification was found for the different isoforms of nitric-oxide synthase and that the nitrosylation of nitric-oxide synthase provides a negative regulatory mechanism on nitric oxide production (50,51). The same nitrosylation mechanism is also observed for the regulation of PGI 2 synthase activity upon the S-nitrosylation of the tyrosine 430 residue (52). In our study, we observed in quiescent non-stimulated EC that the PGI 2 synthase was constitutively nitrosylated, which is concordant with the very low concentration of prostacyclin in BAEC. Treatment with VEGF-A 165 induced a parallel rapid and transient denitrosylation of PGI 2 synthase and PGI 2 synthesis, whereas a treatment with VEGF-C induced neither PGI 2 synthase denitrosylation nor PGI 2 synthesis. The specific mechanism by which PGI 2 synthase denitrosylation occurs remains to be identified. However, recent studies proposed the existence of a nitrase-denitrase tyrosine activity, which is regulating the nitrosylation-denitrosylation equilibrium of selective proteins such as PGI 2 synthase (53,54). Consequently, our data suggest that the only difference in the signaling pathways leading to PGI 2 synthesis of VEGF-A 165 as opposed to VEGF-C is the capacity of VEGF-A 165 to promote, upon the activation of VEGFR-1/R-2 heterodimer, a rapid and transient of PGI 2 synthase denitrosylation and PGI 2 release (Fig. 10).
In conclusion, we observed for the first time the necessity for VEGFR-1/R-2 heterodimer activation to promote a selective biological activity in occurring PGI 2 synthesis. We have also observed that, even if VEGF-C and VEGF-A 165 are capable of activating the same signaling pathways leading to endothelial arachidonic acid release, they showed a different capacity to promote PGI 2 synthase denitrosylation and therefore PGI 2 synthesis. Such a difference would be attributable to the exclusive capacity of VEGF-A isoforms to activate VEGFR-1/R-2 het-FIG. 9. S-Nitrosylation of PGI 2 synthase in the presence of VEGF-A 165 and VEGF-C. Post-confluent BAEC were stimulated time-dependently with PBS, VEGF-A 165 , or VEGF-C (10 Ϫ9 M). Cell lysates (500 g of total proteins) were immunoprecipitated with PGI 2 synthase IgG. Western blot analyses were performed with an anti-nitrotyrosine IgG to assess the S-nitrosylation level of PGI 2 synthase, and then the membranes were stripped and reprobed with an anti-PGI 2 synthase IgG as described above. The nitrosylation level of PGI 2 synthase upon stimulation with VEGF-A 165 or VEGF-C was expressed as a function of the protein expression level of PGI 2 synthase, and the results were normalized to PBS-treated cells.
FIG. 10. Proposed signaling pathway by which VEGF-A 165 mediates PGI 2 synthesis. PGI 2 synthesis induced by VEGF-A 165 requires the heterodimerization of VEGFR-1/R-2 subunits and is potentiated by the contribution of NRP-1 coreceptor. Such a complex that can activate downstream effectors (namely p42/44 MAPK and cPLA 2 ) provides the release of arachidonic acid, which can be converted upon the activation of COX-1 and PGI 2 synthase into PGI 2 . On the other hand, the activation of VEGFR-2/R-2 homodimer by VEGF-C can promote similar cell signaling up to the release of arachidonic acid; however, VEGF-C is unable to promote S-denitrosylation of PGI 2 synthase, which is essential for PGI 2 synthesis. erodimer. These findings provide novel informative data leading to a better comprehension of the selective roles played by VEGF family members and their corresponding receptors on the regulation and maintenance of vascular tone and integrity.