Nitric oxide is an upstream signal of vascular endothelial growth factor-induced extracellular signal-regulated kinase1/2 activation in postcapillary endothelium.

We recently demonstrated that nitric oxide (NO) significantly contributes to the mitogenic effect of vascular endothelial growth factor (VEGF), suggesting a role for the NO pathway in the signaling cascade following kinase-derivative receptor activation in vascular endothelium. The aim of this study was to investigate the intracellular pathways linked to VEGF/NO-induced endothelial cell proliferation. We assessed the activity of the mitogen-activated protein kinase (MAPK) that is specifically activated by growth factors, extracellular-regulated kinase (ERK1/2), on cultured microvascular endothelium isolated from coronary postcapillary venules. ERK1/2 was immunoprecipitated, and its activity was assessed with an immunocomplex kinase assay. In endothelial cells exposed for 5 min to the NO donor drug sodium nitroprusside at a concentration of 100 microM, ERK1/2 activity significantly increased. VEGF produced a time- and concentration-dependent activation of ERK1/2. Maximal activity was obtained after 5 min of stimulation at a concentration of 10 ng/ml. The specific MAPK kinase inhibitor PD 98059 abolished ERK1/2 activation and endothelial cell proliferation in a concentration-dependent manner in response to VEGF and sodium nitroprusside. The NO synthase inhibitor Nomega-monomethyl-L-arginine, as well as the guanylate cyclase inhibitor 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one, blocked the activation of ERK1/2 induced by VEGF, suggesting that NO and cGMP contributed to the VEGF-dependent ERK1/2 activation. These results demonstrate for the first time that kinase-derivative receptor activation triggers the NO synthase/guanylate cyclase pathway to activate the MAPK cascade and substantiates the hypothesis that the activation of ERK1/2 is necessary for VEGF-induced endothelial cell proliferation.

Vascular endothelial growth factor (VEGF) 1 is a secreted protein that is a specific growth factor for endothelial cells, and it has been shown to increase vascular permeability (1,2). It is angiogenic in vivo and in in vitro assays (3,4), and its physiological importance in vasculogenesis is well documented (5,6). The action of VEGF is regulated by two receptors belonging to the tyrosine kinase family, Flt-1 and KDR (or Flk) (7,8). Flt-1, which has higher affinity for VEGF than KDR, is required for endothelial cell morphogenesis, whereas KDR is involved primarily in mitogenesis (5,6,9,10).
The postreceptor signaling pathways underlying VEGF actions on endothelial cells are still unclear. VEGF has been shown to elevate intracellular inositol 1,4,5-trisphosphate and calcium levels and to stimulate tyrosine phosphorylation and von Willebrand factor release in cultured human umbilical vein endothelial cells (11). VEGF effects on permeability (12) and vascular tone (13) are coupled to nitric oxide (NO) production. Consistent with this observation, we have recently demonstrated that NO production and cGMP elevation contribute to the angiogenic effect of VEGF (14,15). The activation of mitogen-activated protein kinase (MAPK) cascade by VEGF has been recently demonstrated (16).
MAPKs are important intermediates in signal transduction pathways that are stimulated by a variety of agents, such as growth factors, hormones, neurotransmitters, and physical and chemical stressing agents (17). Many receptor tyrosine kinase and G protein-coupled receptors have been shown to activate the MAPKs. The 44-and 42-kDa MAPK (ERK1 ⁄2 ) isoforms are ubiquitously expressed and have been shown to be activated by dual specificity MAPK kinases (MEK 1 /MEK 2 ) in response to diverse stimuli (18,19).
This study was designed to characterize the transducing pathways underlying VEGF-activated endothelial cell proliferation. Recently, we have shown that NO is a downstream signal in VEGF effects (15). Here, we have investigated the role of NO on the intracellular pathway linked to VEGF receptor activation in postcapillary endothelium. We assessed MAPK activity specifically activated by a growth factor, i.e. ERK1 ⁄2 , on cultured endothelium isolated from coronary postcapillary venules.

MATERIALS AND METHODS
Cell Line Culture Conditions and Proliferation Assay-Coronary venular endothelial cells (CVECs) were obtained and maintained in culture as described previously and characterized for their endothelial morphology by immunofluorescent staining for factor VIII antigen and uptake of acetylated low density lipoproteins (20). Cells between passages 15 and 25 were used in these experiments.
Cell proliferation was quantified by total cell number after 48 h of stimulation with test substances (14). To evaluate the effect of the MAPKK, NO synthase (NOS), and guanylate cyclase inhibitors, the drugs were added to the cells 30 min before the test substances. Proliferation is expressed as mean Ϯ S.E. of total cells counted in each well.
Immunoprecipitation and Immunocomplex Kinase Assay of ERK1 ⁄2 -CVECs were serum starved overnight. Following treatment, cells were washed twice with ice-cold Dulbecco's phosphate-buffered saline and lysed by adding 0.3 ml of buffer containing 50 mM Tris (pH 7.4), 1% Triton X-100, 1 mM EGTA, 100 mM NaCl, 1 mM Na 3 VO 4 , 0.2 M phenylmethylsulfonyl fluoride, 25 g/ml leupeptin, 10 g/ml aprotinin, and 10 mM NaF. To assess the calcium dependence of ERK1 ⁄2 activation, CVECs were stimulated in the presence of 3 mM EGTA. Cell lysate containing 100 g of protein in a total volume of 800 l were precleared with nonimmune rabbit IgG and 30 l of goat anti-rabbit IgG agarose beads on a rotating plate for 1 h at 4°C and then centrifuged at 10,000 ϫ g for 10 min. 1 g of anti-ERK 1 polyclonal antibody, which is reactive with ERK 1 and to a lesser extent with ERK 2 , and 25 l of goat anti-rabbit IgG agarose beads were added to the supernatant, and the mixture was placed on a rotating plate overnight at 4°C. Following a centrifugation at 10,000 ϫ g for 5 min, the pellet was recovered and washed twice with the lysis buffer and once with the kinase buffer containing 20 mM Hepes (pH 7.6), 20 mM MgCl 2 , and 2 mM dithiothreitol. The kinase assay was carried out at 30°C for 10 min in 30 l of assay buffer containing 5 g of myelin basic protein (MBP) as specific substrate for ERK1 ⁄2 (21), 20 M ATP, and 3 Ci of [␥ 32 P]ATP. The reaction was stopped by the addition of Laemmli's sample buffer and boiled for 5 min. The samples were resolved by 12% SDS-polyacrilamide gel electrophoresis, stained with Coomassie Brilliant Blue, and exhaustively destained. The gel was dried, and the incorporation of [␥ 32 P]ATP was visualized by autoradiography. Gel slices of the 20-kDa MBP bands were also cut out in most of the experiments, and their radioactivity was measured by liquid scintillation counting.
Inositol Phosphate Activation-CVECs seeded onto 6-well plates (3 ϫ 10 4 cells/well) after overnight incubation were labeled with [ 3 H]myoinositol (2 Ci/ml) in DMEM without cold inositol for 48 h. Excess of tritiated myoinositol was removed by three washes with cold DMEM followed by 4 h of incubation with cold DMEM at 37°C. After one wash, cells were incubated for 10 min with 20 mM LiCl to block myoinositol-1-phosphatase and then with test compounds for the designed times. Reaction was stopped by the addition of ice-cold methanol for 30 min. Cells were scraped, and cell-associated inositols were extracted by chloroform-methanol (1:1). Water-soluble fractions were applied to anion-exchange columns (Resin AG-X8, 200 -400 mesh, formate form), and water-soluble inositols were eluted by successive washes with 60 mM CH 5  Calcium Measurements-Cytosolic calcium measurements were made using the calcium-sensitive fluorescence indicator indo-1 (Molecular Probes, Eugene, OR) (22). Cells were incubated with the acetoxymethyl ester for the dye in DMEM containing 1% Me 2 SO for 40 min. At that time, the cells were rinsed with Dulbecco's phosphate-buffered saline and reincubated in DMEM without Me 2 SO, dye, or phenol red for 20 min to allow for de-esterification. The experiments were performed on a stage-scanning photometric imaging system (ACAS 570, Meridian Instruments, Okemos, MI) at 37°C. The cells were illuminated with the 355-nm line of an argon laser that was attenuated with neutral density filters and an acuosto-optic modulator to minimize bleaching. Fluorescence emission was monitored at 405 and 485 nm using narrow-band pass filters and dual photomultipliers. The fluorescent images of a suitable field of cells were captured before and after addition of VEGF, and the 405/485 nm ratio was analyzed. This ratio is a direct index to the calcium concentration within the cell at the time of image capture.
Differential Reverse Transcription PCR Analysis-Subconfluent and serum-starved cells were stimulated for 4 -24 h with VEGF in the presence of 1% serum. At the end of incubation, total RNA was isolated by the standard guanidine thiocyanate-phenol-chloroform extraction (23). cDNA was synthetized as described (24). Differential reverse transcription PCR for NOS isoforms was carried out by using 5 l of cDNA and specific primers for bovine calcium/calmodulin-dependent NOS (ecNOS) and iNOS with sequences as follows: ecNOS sense, 5Ј-GCTT-GAGACCCTCAGTCAGG-3Ј; ecNOS antisense, 5Ј-GGTCTCCAGTCTT-GAGCTGG-3Ј (25); iNOS sense, 5Ј-TAGAGGAACATCTGGCCAGG-3Ј; iNOS antisense, 5Ј-TGGCAGGGTCCCCTCTGATG-3Ј (26). Calibration was performed by co-amplification of the same cDNA samples with primers for glyceraldehyde-3-phosphate dehydrogenase as an internal standard, with sequences as described (24). For PCR amplification a Perkin-Elmer GeneAmp PCR System 2400 was used. The PCR cycles were as follows: 30 s at 94°C, 1 min at 55°C, and 1 min at 72°C. After 30 cycles of amplification, aliquots of each sample product (20 l) were electrophoresed on a 3% agarose gel and stained with ethidium bromide. The size of the amplification products was 296 for ecNOS and 372 for iNOS. Image processing and analysis of the intensity of the bands were performed as described (24). Results were evaluated as ratio between the target genes (ecNOS and iNOS) and glyceraldehyde-3phosphate dehydrogenase amplification analysis.
Determination of NOS Activity-Subconfluent CVECs in 60 -mm culture dishes were serum starved overnight and then washed and equilibrated for 20 min at 37°C with Hepes buffer containing: 145 mM NaCl, 5 mM KCl, 1 mM MgSO 4 , 10 mM Hepes, 1 mM CaCl 2 , 1 mM glucose, pH 7.4. Cells were incubated for 20 min with 1 Ci of [ 3 H]arginine plus 10 M L-arginine. To assess the calcium dependence of NOS activation, experiments were performed in calcium-free Hepes buffer containing 1 mM EGTA. Test substances were added for 5 min at 37°C. The reaction was stopped by cold Hepes buffer containing 4 mM EDTA, and the supernatant was removed. 0.5 ml of ethanol added to each monolayer was allowed to evaporate, and 2 ml of 10 mM Hepes-Na, pH 5.5, was added for 20 min. The supernatant was collected and applied to 0.8 ml of Dowex AG50WX-8 (Na-form) and vigorously shaken for 45 min. Then, 0.5 ml was collected and added to 3 ml of liquid scintillation counting mixture. NOS activity was expressed as pmol/mg of protein.
Measurement of cGMP Levels-cGMP levels were measured on cell extracts from confluent cell monolayers by radioimmunoassay using iodinated tracer (14). Cell monolayers were treated with 1 mM 3-isobutyl-5-methyl-xanthine for 15 min before stimulation. After stimulation, cells were rinsed with phosphate-buffered saline and removed by scraping in ice-cold 10% trichloroacetic acid. Following centrifugation, cGMP levels were assayed in the supernatant, and proteins were measured in the pellet by the Bradford's procedure. Data are expressed as fmol/mg of protein.

ERK1 ⁄2 Activation Induced by VEGF Is Time-dependent, Concentration-dependent and Sensitive to Inhibition of the MAPKK-
The ability of VEGF to stimulate the MAPK cascade was assessed. After 24 h of starvation, CVECs were stimulated with 10 ng/ml VEGF over a range of times from 2 to 15 min, and the activity of the immunoprecipitated ERK1 ⁄2 was measured. The activation of ERK1 ⁄2 was statistically higher than the unstimulated control within 2 min and reached the maximum at 5 min (Fig. 1a). The activation of ERK1 ⁄2 by VEGF was concentration-dependent with maximal activity at 10 ng/ml, which doubled the basal activity of ERK1 ⁄2 (Fig. 1, b and d). In the same experiments, bFGF (10 ng/ml) did not significantly induce ERK1 ⁄2 activation. When the endothelial cells were preincubated for 30 min with increasing concentrations of the MAPKK inhibitor PD 98059 (27), VEGF-induced ERK1 ⁄2 activation was inhibited (IC 50 ϭ 30 M). At 100 M, PD 98059 completely abolished the ERK1 ⁄2 activation produced by 10 ng/ml of VEGF but did not modify the unstimulated control (Fig. 1, c and e).
NO Activates ERK1 ⁄2 -In previous reports, we demonstrated that the NO pathway is necessary for the proliferative effects of VEGF on microvascular endothelial cells (14). We therefore investigated whether NO contributed to VEGF mitogenic activity by activating the MAPK cascade. For this purpose, starved and subconfluent CVECs were treated with the NO donor SNP, and the activity of ERK1 ⁄2 was measured. After 5 min of stimulation with 100 M SNP, the MAPK activity was increased by 2-fold (Fig. 2a). NO-induced ERK1 ⁄2 activation and the proliferative effect of NO were abolished by PD 98059 (Fig.  2, a and b), indicating that ERK1 ⁄2 was specifically and directly activated by NO and that this phosphorylation cascade was involved in signaling mitogenesis in postcapillary endothelial cells.
Calcium-dependent Activation of NOS and ERK1 ⁄2 by VEGF-We then characterized the NOS isoform mediating VEGF effect in CVECs. The rapid activation of ERK1 ⁄2 in response to VEGF suggested that the acute activation of NO production in CVECs by ecNOS. Differential reverse transcription PCR of total RNA indicated that this isoform was predominantly expressed in CVECs (Fig. 3a). After 4 h from VEGF administration, ecNOS expression was not modified, indicating the absence of a transcriptional event between VEGF administration and NO production. iNOS expression was not detected at any time point between 4 and 24 h of exposure to the growth factor.
CVECs preloaded with the ratiometric fluorescent indicator indo-1 exhibited a rapid calcium transient upon exposure to VEGF (Fig. 3b). The upward stroke of the calcium transient began 3 min after addition of VEGF, the peak concentration of cytosolic calcium occurred at approximately 7 min, and recov-ery occurred over the next 20 min. After continued exposure of CVECs to VEGF for 70 min, cytosolic calcium had recovered only 65%, suggesting a continuing signal for calcium-calmodulin NO production beyond the rapid peak.
Consistent with the rapid cytosolic calcium elevation, within 5 min after VEGF exposure, NOS activity increased, and EGTA abolished its elevation (Fig. 3c). EGTA also abolished ERK1 ⁄2 activation by VEGF, suggesting that calcium was required to trigger the MAPK cascade, as well as the NOS activity (Fig. 3d). The MAPKK inhibitor specifically reduced the proliferative effect of VEGF in a concentration-dependent manner, whereas it did not inhibit the growth-promoting effect of bFGF (Fig. 4a). The IC 50 for growth inhibition (10 M) was in the same range of concentration as that for ERK1 ⁄2 inhibition. At the highest concentration, PD 98059 slightly reduced the number of cells recovered under control conditions. The effect was independent from cytotoxicity, as indicated by trypan blue exclusion assays (data not shown).

MAPKK and NOS/Guanylate Cyclase Inhibitors Block VEGF-induced Endothelial
The addition of VEGF doubled cGMP levels in CVECs, an effect specifically blocked by NOS inhibitors (Table I) as previously reported (14). L-NMMA inhibited VEGF-induced growth in a concentration-dependent manner; maximal growth inhibition was obtained at 200 M (IC 50 ϭ 10 M) (Fig. 4b). Conversely, no inhibition in cell growth was produced when bFGF was used as a mitogen (Fig. 4b).
ODQ produced concentration-dependent inhibition of the FIG. 1. Time-and dose-dependent activation of ERK1 ⁄2 exposed to VEGF. a, CVECs were exposed for different times (2-5 min) to 10 ng/ml VEGF. b, cells were stimulated for 5 min with increasing concentrations (5-20 ng/ml) of VEGF and 10 ng/ml bFGF. c, effect of the MAPKK inhibitor PD 98995 (10 -100 M) on 10 ng/ml VEGF-induced ERK1 ⁄2 activation. d and e are representative autoradiographies related to b and c, respectively. ERK1 ⁄2 was immunoprecipitated, and its activity was measured with an in vitro kinase assay by using [␥ 32 P]ATP and MBP as substrate. The samples were resolved by 12% SDS-polyacrilamide gel electrophoresis followed by autoradiography. Gel slices of the 20 K D MBP bands were cut out, and the radioactivity was measured by liquid scintillation counting. n ϭ 5; mean Ϯ S.E. *, p Ͻ 0.05; ***, p Ͻ 0.001 versus unstimulated control; #, p Ͻ 0.001 versus VEGF alone (ANOVA followed by Fisher's test).
guanylate cyclase activation and cGMP levels elevation induced by VEGF as well as by the NO donor SNP (maximal effect at 10 M; IC 50 ϭ 0.5 M) ( Table I). The minimal effective concentration of ODQ that inhibited cGMP formation was sufficient to block the proliferative effect of VEGF, and lower concentrations gave the same effect (Fig. 4c). Conversely, proliferation and cGMP elevation produced by the NO donor SNP were reduced by ODQ in a concentration-dependent manner ( Fig. 4c and Table I). Maximal inhibition was obtained at 10 M, and the IC 50 was 0.5 M for both the effects. The guanylate cyclase inhibitor did not produce significant reduction of bFGFinduced growth (Fig. 4c).

MAPKK Inhibitor Does Not Affect the Biochemical Cascade of NOS/Guanylate Cyclase Elicited by VEGF and NO-
To demonstrate the exact biochemical location of the MAPK in the NO/NOS pathway in our system, the MAPKK inhibitor was tested on guanylate cyclase activation. PD 98059 did not affect the NOS/cGMP pathway activation stimulated by either VEGF or SNP on CVECs at any of the concentrations tested (Table I). Similar results were obtained when PD 98095 was assessed on NOS activity. The VEGF-induced NOS activity (223 Ϯ 11 pmol/mg of protein versus a basal value of 169 Ϯ 15 pmol/mg of protein) could be selectively blocked by 3 M L-NMMA (131 Ϯ 10 pmol/mg of protein; IC 50 ϭ 50 M) but not by 100 M PD 98059 (210 Ϯ 22 pmol/mg of protein; n ϭ 3).
The possibility that PD 98059 could affect other transducing pathways required for proliferation in our system was ruled out in parallel experiments in which inositol phosphate metabolism was assessed. PD 98059 at the concentration producing 100% reduction of the specific biochemical target (ERK1 ⁄2 activation) failed to affect the metabolism of inositol phosphate. VEGF induced inositol phosphate 1 accumulation (448 Ϯ 37 cpm/well over basal control), which was not affected by 100 M PD 98059 pretreatment (642 Ϯ 98 cpm/well; n ϭ 3).
NOS/Guanylate Cyclase Inhibitors Prevent the VEGF-induced ERK1 ⁄2 Activation-To investigate the exact location of the NOS/cGMP pathway in the phosphorylation cascade triggered by VEGF in our system, experiments were done to assess the involvement of NO pathway in VEGF-induced ERK1 ⁄2 activation. CVECs were pretreated for 30 min with 200 M L-NMMA and then stimulated for 5 min with 10 ng/ml VEGF. Data obtained showed that the pretreatment with L-NMMA abolished the increased ERK1 ⁄2 activity elicited by VEGF (Fig.  5a). This effect was selective for the VEGF effect because no inhibition was found for cells stimulated with 10% calf serum (1680 Ϯ 70 cpm and 1725 Ϯ 150 cpm, with and without L-NMMA, respectively; n ϭ 3).
Consistent with the observation that NO is the transducing molecule between the VEGF receptor and ERK1 ⁄2 , ODQ significantly inhibited the VEGF-and SNP-induced increase of ERK1 ⁄2 activity (Fig. 5b). DISCUSSION The data presented here demonstrate that the mitogenic activity of VEGF on postcapillary endothelial cells requires the activation of the MAPK cascade and that NO/cGMP production mediates the MAPK activation following VEGF receptor interaction, ultimately leading to endothelial cell growth. These conclusions are based on the following observations: 1) VEGF stimulated the MAPK specifically linked to proliferation, i.e. ERK1 ⁄2 , as did the NO-donor drug SNP; 2) blockade of the NO pathway by L-NMMA and by ODQ prevented the ERK1 ⁄2 activation by VEGF and SNP; and 3) inhibition of MAPK kinase activation, of NO synthase activity, and of cGMP production specifically blocked VEGF/NO-induced proliferation.
In rat liver sinusoidal endothelial cells, it was reported that VEGF stimulated phosphorylation of the MAPK (16). Postcapillary venular endothelium has the ability to respond promptly to mitogenic peptides. Using cultured endothelium from coronary postcapillary venules, we demonstrated that ERK1 ⁄2 activation lies upstream of the proliferative effect of VEGF. The specificity of ERK1 ⁄2 activation is confirmed by the use of the MAPKK (or MEK) inhibitor PD 98059 (27). This compound has been demonstrated to be a selective and noncompetitive MEK inhibitor in in vitro assay (30,31) without any effect on ERK. PD 98059 at concentrations above 50 M has been shown both to inhibit MEK1 ⁄2 by binding a regulatory site on the enzyme and to prevent activation by c-Raf and MEK1 ⁄2 kinase (30). Our data show that in this concentration range, PD 98059 prevented the ERK1 ⁄2 activation and the proliferative effect induced by VEGF, demonstrating that the activation of ERK1 ⁄2 is a necessary step for endothelial cell proliferation.
In previous work, we have shown that molecules able to increase NO levels induced endothelial cells proliferation and FIG. 2. Effect of NO on ERK1 ⁄2 activation and cell proliferation. a, ERK1 ⁄2 activation. CVECs were stimulated with 100 M SNP. ERK1 ⁄2 was immunoprecipitated, and its activity was measured with an in vitro kinase assay by using [␥ 32 P]ATP and MBP as substrate. The samples were resolved by 12% SDS-polyacrilamide gel electrophoresis followed by autoradiography. Gel slices of the 20-kDa MBP bands were cut out, and the radioactivity was measured by liquid scintillation counting. b, cell proliferation: 1.5 ϫ 10 3 cells resuspended in 10% calf serum were seeded in each well of 96-well plates. After adherence (3-4 h), the medium was replaced with 1% calf serum DMEM containing test substances and incubated for 48 h. After fixation and staining of cells with hematoxylin-eosin with Diff-Quik, the number of cells was counted in seven random fields of each well at ϫ 100 magnification with the aid of a 21-mm 2 ocular grid. Data are expressed as total number of cells counted/well. PD 98059 was given at 100 M for 30 min before cell stimulation. n ϭ 3; ***, p Ͻ 0.001 versus unstimulated control; #, p Ͻ 0.05 versus SNP alone (ANOVA followed by Fisher's test). migration in vivo and in vitro (32,33) and also that the activation of the NO pathway following VEGF stimulation significantly contributed to the mitogenic effect of VEGF (14). Here, we demonstrate that under the same experimental conditions, NO directly triggers the activation of the MAPK cascade. ERK1 ⁄2 activation and endothelial cell proliferation promoted by NO are selectively blocked by the MAPKK inhibitor.
ODQ, a selective and specific inhibitor of the soluble guanylate cyclase (29), blocked in a concentration-dependent manner cGMP elevation in venular endothelial cells exposed to the NO donor and to VEGF. Consistent with cGMP being required to transduce the NO-dependent proliferation signal, neither VEGF or SNP promoted ERK1 ⁄2 phosphorylation and growth in the presence of ODQ. The IC 50 for proliferation and cGMP formation overlapped when SNP was the mitogen. Interestingly, minimal reduction of cGMP levels was sufficient to completely block the proliferation signal produced by VEGF. Because in our experimental model, production of cGMP is required for VEGF-induced cell adhesion (15), the effect of ODQ on VEGF proliferation might be related to the specific requirement of cell adhesion to fulfill the growth program encoded by VEGF in postcapillary venular endothelium.
The link between VEGF stimulation of CVECs, NO release, and the rapid activation of ERK appears to be ecNOS, the calcium/calmodulin-dependent enzyme found in endothelial cells. ecNOS is the predominant isoform expressed in CVECs, and its expression is not affected by the growth factor. We show that VEGF causes a rise in cytoplasmic calcium that peaks at 7 min and triggers NOS activity within 5-10 min. Consistently, ERK1 ⁄2 activity peaks between 5 and 10 min. Thus, the time frame for increases in cytosolic calcium, NO production, and ERK1 ⁄2 activity support a KDR/calcium/ecNOS/NO/soluble guanylate cyclase/cGMP/ERK1 ⁄2 cascade activated by VEGF. The mechanism responsible for the calcium transient is not completely clear. However, as indicated by the elevation of inositol phosphate levels recovered in CVECs, release of calcium from the endoplasmic reticulum could occur by KDR-mediated activation of phospholipase C gamma 1 (34). Alternatively, VEGF may activate processes that accelerate calcium entry via plasmalemmal ion channels (35).
We recently demonstrated that NO synthase lies downstream of the angiogenesis induced by VEGF but not of that induced by bFGF (15). The present data provide a new insight on the mechanism underlying the role of NO in mediating VEGF effect by demonstrating that the NO pathway is upstream of the MAPK cascade activated by VEGF. In fact, the ERK1 ⁄2 activation and the endothelial proliferation following VEGF/receptor activation are prevented in culture conditions in which NO production and cGMP elevation are impaired by the use of NOS/cGMP selective inhibitors. Conversely, blockade of the MAPKK does not affect the NOS/guanylate cyclase. Thus, the NO pathway activation is intermediate between the VEGF receptor activation and the MAPK phosphorylation in endothelial cells.
Other observations support a link between the NO and the MAPK cascade. Sing et al. (36) recently described that ERK activation is necessary for the induction of the inducible NOS by interleukin-1␤ in myocytes and cardiac microvascular endo-  4, 9, and 24), iNOS expression at 0, 4, 9, 24 h of stimulation, respectively. b, effect of VEGF on cytosolic calcium mobilization in individual adherent CVECs. Addition of 10 ng/ml VEGF induced a synchronized rapid increase in cytosolic calcium followed by a long-lasting decline to levels above prestimulation values. Data are the means of tracing recorded from 22 individual cells. c, effect of calcium on NOS activity in CVECs. NOS activity (pmol/mg of protein) was evaluated by [ 3 H]L-arginine conversion in cells exposed to 10 ng/ml VEGF for 5 min (open columns). EGTA was used at 1 mM in calcium-free buffer (hatched columns) (n ϭ 3). d, effect of calcium on ERK1 ⁄2 activation. Unstimulated and VEGF-treated cells were lysed, ERK1 ⁄2 was immunoprecipitated, and its activity was detected as MBP phosphorylation. Cells were stimulated with 10 ng/ml VEGF in the absence (open columns) and in presence (hatched columns) of 3 mM EGTA. ERK1 ⁄2 activity is expressed as radioactivity of gel slides of phosphorylated MBP. n ϭ 2; **, p Ͻ 0.01 versus unstimulated control; #, p Ͻ 0.05 versus VEGF (ANOVA followed by Fisher's test). thelial cells. Elevated shear rate caused increased production of NO and activated the MAPK cascade in endothelial cells (37,38). Whereas in the above mentioned reports, MAPK activation anticipates and/or parallels NO production, our data indicate an upstream role for NO.
The role of NO in promoting cell growth and differentiation is controversial. In angiogenesis, NO elevation has been shown to be positively correlated with neovascularization and tumor growth (15,39,40) in adult rodent models. Conversely, in the chorionallantoic membrane of the chick embryo and during the developmental maturation of Drosophila, NO acts as an antiproliferative agent (41,42). ERK1 ⁄2 is thought to be directly involved in transmitting signals from growth factor receptors to the nucleus to regulate gene transcription and protein synthesis, leading to proliferation or differentiation and apoptosis (17,43,44,45). Recently, a difference in the actions of the ERK   and p38/JNK pathway has been demonstrated in PC12 cells; the activation of JNK-p38 cascade leads to apoptosis of PC12 cells, whereas the activation of ERK1 ⁄2 seems to be necessary for survival and/or antiapoptosis of PC12 cells (45). The data here reported support the hypothesis of NO as a "prosurvival" or antiapoptotic effector for endothelial cells. Although it is presently difficult to speculate on whether the opposing effects of NO in controlling cell growth are due to species or to differentiation diversity, nevertheless our results using venular endothelial cells continue to emphasize the importance of NO as a balancing element in the molecular events between cell proliferation and differentiation.