To clarify the role of caveolae in VEGF/VEGF receptor-2 (VEGFR-2)-mediated signaling cascades, primary cultured human umbilical vein endothelial cells (HUVECs) were fractionated to isolate caveolae-enriched cell membranes. Interestingly, VEGFR-2, phospholipase D2 (PLD2), and Ras were enriched in caveolae-enriched fractions. Moreover, VEGF increased PLD activity in a time- and dose-dependent manner in HUVECs, whereas a ligand specific for VEGFR-1 placental growth factor did not change PLD activity. A PLD inhibitor, 1-butanol, almost completely suppressed VEGF-induced ERK phosphorylation and cellular proliferation, whereas the negative control for 1-butanol, 3-butanol, did not produce significant changes. Addition of phosphatidic acid negated the 1-butanol-induced suppression. Pharmacological analyses using several inhibitors indicated that PKC-δ regulates the VEGF-induced activation of PLD/ERK. Thus PLD2 could be involved in MEK/ERK signaling cascades that are induced by the VEGF/VEGFR-2/PKC-δ pathway in endothelial cells. Pretreatment with the cholesterol depletion agent methyl-β-cyclodextrin (MβCD) almost completely disassembled caveolar structures, whereas the addition of cholesterol to MβCD-treated cells restored caveolar structures. Pretreatment with MβCD largely abolished phosphorylation of MEK/ERK by VEGF, whereas the addition of cholesterol restored VEGF-induced MEK/ERK phosphorylations. These results indicate that intact caveolae are required for the VEGF/VEGFR-2-mediated MEK/ERK signaling cascade.
- protein kinase C-δ
- vascular endothelial growth factor
- phospholipase D
there are six known members of the VEGF family: VEGF, placental growth factor (PlGF), VEGF-B, VEGF-C, VEGF-D, and orf virus VEGF (also called VEGF-E) (8, 23, 40). The VEGF family proteins are secreted as dimeric glycoproteins. They display differential interactions with three related receptor tyrosine kinases: VEGFR-1/Flt-1, VEGFR-2/KDR/Flk-1, and VEGFR-3/Flt-4 (8, 23, 40). VEGF exerts its action by binding to VEGFR-1 and VEGFR-2 (8, 23, 40). VEGFR-1 and VEGFR-2 are expressed predominantly in vascular endothelial cells, but a few additional types of cells express one or both of these receptors (23). Of these, VEGFR-2 is expressed in hematopoietic stem cells, megakaryocytes, retinal progenitor cells, and transformed tumorigenic cell types (23). On activation of VEGFR-2 in endothelial cells, three major second messenger pathways elicit cell proliferation, migration, and survival, which are essential steps for angiogenesis (10, 14, 18, 21, 22, 28, 38, 39, 41, 42). These pathways are the following: mitogen-activated protein/ERK kinase (MEK)/ERK cascade, phosphatidylinositol (PI) 3-kinase/serine-threonine protein kinase/Akt cascade, phospholipase C-γ/intracellular Ca2+/PKC cascades (10, 14, 18, 21, 22, 28, 38, 39, 41, 42).
Caveolae, which are flask-shaped membrane invaginations, were first detected on the surface of endothelial cells (24). There is growing evidence that caveolae may act as structurally and biochemically distinct plasma membrane compartments that localize and regulate transmembrane signaling events (1, 27, 34). This fact is based on the finding that caveolae are enriched for receptors and components of signaling cascades (9, 32, 35). Caveolae are formed from lipid rafts by polymerization of caveolins, which are hairpin-like palmitoylated integral membrane proteins that tightly bind cholesterol (1, 9, 26, 27, 35, 36). Interestingly, VEGFR-2 is colocalized with caveolin-1 in the caveolae of endothelial cells (6, 7, 19). This finding suggests that the VEGFR-2 signaling machinery may also be localized in the endothelial caveolae. However, it is not known how the VEGFR-2 signaling machinery is activated in the endothelial caveolae on VEGF binding.
In this study, we examined the distributions of caveolin-1, VEGFR-2, and the key signaling molecules of VEGFR-2 in primary cultured human umbilical vein endothelial cells (HUVECs). Interestingly, VEGFR-2 is localized in the endothelial caveolae, together with phospholipase D2 (PLD2) and Ras. Furthermore, this study provides evidence that PKC and PLD, in response to activation by VEGF/VEGFR-2, participate in the activation of MEK and ERK, which control several cellular processes in vascular endothelial cells.
Materials and preparation of endothelial cells. Recombinant human VEGF165 was purchased from R&D Systems (Minneapolis, MN). Antibodies for caveolin-1 and immunoglobulin heavy chain binding protein (BiP) were purchased from BD Transduction Laboratory (San Diego, CA). Antibody for Ras was purchased from Upstate Biotechnology (Waltham, MA). Antibodies for phospho-MEK (Ser217/221), MEK, phospho-ERK1/2 (Thr202/Tyr204) and ERK1/2 were purchased from Cell Signaling Technology (Beverly, MA). Antibodies for VEGFR-2 and Raf-1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibody for PLD1/2 (15) was a kind gift of Dr. Pann-Gill Suh (POSTECH, Pohang, Korea). Staurosporine, GF109203X, BAPTA-AM, and wortmannin were purchased from Sigma-Aldrich (St. Louis, MO). SU5614, brefeldin-A, Go6976, rottlerin, RO31–8220, phosphatidic acid (PA), and methyl-β-cyclodextrin (MβCD) were purchased from Calbiochem (San Diego, CA). Media, FBS, and most other biochemical reagents were purchased from Sigma-Aldrich unless otherwise specified. HUVECs were prepared from human umbilical cords by collagenase digestion, were grown in 20% FBS-containing medium 199 (13), and were between passages 2 and 3 when used for the experiments in this study. The HUVECs were incubated in 1% FBS-containing medium for 16 h before performing each experiment.
Cell fractionation. Separation of light and heavy membrane fractions from HUVECs was performed according to Song et al. (37). Briefly, the HUVECs were suspended to 2 ml of 0.5 M sodium carbonate buffer (pH 11.0) containing phosphatase inhibitors and protease inhibitors. The cell suspension was homogenized, adjusted to 45% sucrose by the addition of 80% sucrose prepared in MBS buffer, and placed into ultracentrifugation tubes. A 5–45% discontinuous sucrose gradient was then formed on top. This sample was centrifuged at 39,000 rpm for 6 h in a SW-41 rotor (Beckman Instruments; Palo Alto, CA). The fractions were collected from the top in 1-ml amounts except for fractions 1–4, which were collected as 0.9-ml fractions. After the addition of a sample loading buffer, the fractions were boiled, separated by SDS-PAGE, and transferred to nitrocellulose membranes. The membranes were immunoblotted with the desired antibodies. All signals were detected by chemiluminescent detection according to the manufacturer's protocol (Amersham; Buckinghamshire, UK).
Assay of PLD activity. PLD activity was assayed by measuring the formation of phosphatidyl-butanol, the product of PLD-mediated transphosphatidylation, in the presence of 1-butanol, as previously described, with a slight modification (25). The HUVECs were preincubated with 3 μCi/ml of [3H]myristic acid (DuPont-NEN; Boston, MA) for 4 h. Transphosphatidylation was catalyzed in the presence of 0.4% 1-butanol, and cells were exposed to the indicated concentration of VEGF for the indicated times in the absence or presence of inhibitors. The reaction was terminated with ice-cold methanol and cells were scraped from the plates. The lipid phase was extracted and developed by thin layer chromatography on silica gel-60 plates (Merck; Darmstadt, Germany) with the use of chloroform-methanolacetic acid (9:1:1) as the solvent. The formation of [3H]phosphatidylbutanol was expressed as a percentage of the total 3H-labeled lipid to account for differences in cell labeling efficiency.
Western blot analysis. The treated HUVECs were harvested with lysis buffer (50 mM Tris·HCl, 100 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 50 mM NaF, 1 mM Na3VO4, 1 μg/ml aprotinin, 1 μg/ml pepstatin, and 1 μg/ml leupeptin). After the addition of the sample loading buffer, lysates were boiled, separated by SDS-PAGE, and transferred to a nitrocellulose membranes. The membranes were immunoblotted with the appropriate primary antibody, washed, and incubated with horseradish peroxidase-conjugated secondary antibody. Signals were visualized by chemiluminescent detection according to the manufacturer's protocol (Amersham).
Proliferation assay. For the proliferation assay, DNA synthetic activity was measured as previously described (14). In brief, HUVECs were plated in 24-well plates at a density of 2 × 104 cells/cm2 in medium 199 containing 2% FBS. After 12 h, plates were changed to fresh medium with the indicated agents. After 24 h, fresh medium and agents were added, and the cells were incubated for another 24 h. The cells were then washed with PBS, and the DNA amount was measured with PicoGreen fluorescent reagent (Molecular Probes; Eugene, OR) using a fluorescence spectrophotometer equipped with a microplate reader (Molecular Devices; Sunnyvale, CA).
Transmission electron microscopy. The HUVECs were treated with MβCD (5 mM) for 30 min or 1 h plus cholesterol (0.2 mM) for 30 min. Control and treated HUVECs were rinsed two times with PBS and fixed with 2% glutaraldehyde in sodium cacodylate buffer (0.1 M cacodylate and 0.1 M sucrose) at pH 7.4 for 12 h. Cells were subsequently postfixed with 1% osmium tetroxide in 0.1 M sodium cacodylate at pH 7.4 for 1 h. Samples were further dehydrated in graded alcohol solutions and embedded in epon. After the samples hardened, the cells were lifted with embedding medium. Transverse sections of 50 nm were cut with a diamond knife, stained first with uranyl acetate and subsequently with lead citrate, and examined by transmission electron microscopy at 80 kV (ZEOL1020 EX2, Japan).
Data analyses. Data are expressed as means ± SD. Statistical significance was tested using one-way ANOVA, followed by the Student-Newman-Keuls test. Statistical significance was set at P < 0.05.
Distribution of VEGFR-2 and its signaling molecules. To isolate caveolin-enriched endothelial cell membranes, HUVEC homogenate was fractionated using sucrose density centrifugation. Twelve fractions were recovered and subjected to Western blot analysis. Caveolin-1, the structural protein of caveolae, was dominantly enriched in fraction 5, which represented the light membrane fraction (Fig. 1). BiP, a marker for endoplasmic reticulum, was enriched in noncaveolar fractions 9–12. These data indicate that the fractionation method was able to separate the caveolae from the heavier membrane fractions. Consistent with previous reports (6, 7, 19), VEGFR-2 was dominantly enriched in fraction 5 (Fig. 1). To screen for components involved in VEGFR-2 signaling, other signaling molecules were tested. PLD2 and Ras were enriched in fraction 5, whereas Raf-1, MEK, ERK1, and ERK2 were enriched in fractions 8–12 (Fig. 1). These results suggest that VEGFR-2 may interact with PLD2 and Ras in endothelial caveolae.
VEGF increases PLD activity. Because our data suggested that PLD might be involved in VEGFR-2 signaling in endothelial cells, we next examined whether VEGF-induced VEGFR-2 activation could regulate PLD activity in HUVECs. The addition of VEGF (50 ng/ml) increased PLD activity in a time-dependent manner (Fig. 2A). VEGF increased PLD activity as early as 1 min, produced maximal activity during 1–5 min, and produced maximal accumulation of product by 5 min (Fig. 2A). The maximum accumulation of PLD product was 3.1-fold. Moreover, VEGF (10, 50, and 100 ng/ml) increased PLD activity in a dose-dependent manner, whereas the VEGFR-1-specific ligand PlGF (10, 50, and 100 ng/ml) did not change PLD activity (Fig. 2B). Furthermore, pretreatment with a specific VEGFR-2 antagonist (SU5614, 10 μM) completely inhibited the VEGF-induced PLD activity (Fig. 2B). Thus VEGF-induced PLD activity is mainly dependent on VEGFR-2 activation in endothelial cells.
VEGF-induced PLD activity is involved in ERK activation and proliferation. Because PLD is known to activate MEK/ERK signaling through mobilization of Raf-1 (29), we next examined whether VEGF-induced MEK/ERK activation may be partly attributed to VEGF-induced PLD activity. A PLD inhibitor, 1-butanol (0.4%), almost suppressed VEGF-induced ERK1/2 phosphorylation (Fig. 3, A and B). The addition of PA (200 μM) restored the suppressed effect of 1-butanol (0.4%) on VEGF-induced ERK1/2 phosphorylation. The negative control for 1-butanol, 3-butanol (0.4%), did not produce a dramatic change in VEGF-induced ERK1/2 phosphorylation. To address the specificity of 1-butanol-mediated inhibition of PLD on VEGF-induced ERK1/2 phosphorylation, we measured the effect of 1-butanol on another known VEGFR-2 signaling event, the VEGF-induced phosphorylation of Akt. Neither 1-butanol nor 3-butanol produced any effect on VEGF-induced Akt phosphorylation (Fig. 3, C and D). Thus VEGF-induced PLD activity is preferentially involved in ERK1/2 activation, but not in Akt activation, in endothelial cells. Because VEGF-induced ERK activation is closely involved in the proliferation of endothelial cells (21, 28, 38), we next examined the effect of PLD inhibition on VEGF-induced proliferation of endothelial cells by measuring DNA synthetic activity. 1-Butanol (0.04%) significantly suppressed VEGF-induced DNA synthetic activity, whereas the addition of PA (20 μM) restored the suppressed effect of 1-butanol (0.04%) on VEGF-induced DNA synthetic activity (Fig. 4). The negative control for 1-butanol, 3-butanol (0.04%), did not produce a significant change in VEGF-induced DNA synthetic activity. Thus VEGF-induced PLD activity is involved in the proliferation of endothelial cells.
PKC-δ regulates VEGF-induced PLD/ERK activation. Because PLD is regulated by ADP ribosylation factor (ARF) and PKC (4, 5, 11, 12, 16, 17, 20, 31), we determined how these transducers are involved in VEGF-induced activation of PLD and ERK. As expected, the universal PKC inhibitors staurosporine (100 nM) and GF109203X (1 μM) abolished VEGF-induced PLD activity (Fig. 5A). In contrast, intracellular Ca2+ inhibitor BAPTA-AM (1 μM), ARF inhibitor brefeldin-A (50 μg/ml), and PI3-kinase inhibitor wortmannin (30 nM) did not significantly change VEGF-induced PLD activity (Fig. 5A). Notably, PKC-δ inhibitor rottlerin (100 nM) significantly suppressed VEGF-induced PLD activity, whereas PKC-α and -β inhibitor Go69761 (100 nM) and PKC-α, -β, and -γ inhibitor RO31–8220 did not significantly change VEGF-induced PLD activity. Accordingly, pretreatment of universal PKC inhibitors staurosporine and GF109203X and PKC-δ inhibitor rottlerin dramatically abolished VEGF-induced ERK1/2 phosphorylation, whereas other inhibitors did not produce a dramatic change (Fig. 5, B and C). Thus our results suggest that PKC-δ mainly regulates VEGF-induced activation of PLD and ERK.
VEGF-induced activation of MEK/ERK requires intact caveolae. We next investigated whether intact endothelial caveolae are essential for localization of VEGFR-2, PLD, and Ras. Cholesterol is an essential component for normal assembly of caveolae (1, 3). Depletion of cholesterol using MβCD has been shown to disassemble caveolae and thus represents a useful tool to study caveolae-dependent processes (30). Transmission electron microscopy analyses revealed several vesicular invaginations (caveolae) in primary cultured endothelial cell membranes (Fig. 6A). Pretreatment with MβCD (5 mM for 30 min) almost completely disassembled these structures (Fig. 6A). Under these conditions, caveolin-1, VEGFR-2, and Ras were redistributed to fractions 8–12, and the characteristic localization of PLD2 disappeared (data not shown). These results suggest that the integrity of caveolar structures is essential for the localization of VEGFR-2 and its close signaling molecules. The addition of cholesterol (0.2 mM for 1 h) to the MβCD-treated cells restored the caveolar structures (Fig. 6A). This result confirms that cholesterol is an essential substance for maintenance and generation of caveolae. We next examined the effect of the caveolar structure on VEGF-induced activation of MEK/ERK. Pretreatment with MβCD (5 mM for 30 min) largely abolished VEGF-induced MEK and ERK1/2 phosphorylations (Fig. 6, B and C). Notably, repletion of cholesterol (0.2 mM for 30 min) restored VEGF-induced phosphorylation of MEK and ERK1/2. These results indicate that an intact caveolar structure is preferentially required for the VEGF/VEGFR-2-mediated MEK/ERK signaling cascade.
Caveolae, which are flask-shaped membrane invaginations, were first detected on the surface of endothelial cells around 50 years ago (24). In fact, vascular endothelial cells are one of the most abundant sources of caveolae. In a series of electron microscopy studies, Palade and co-workers (2, 33) provided compelling evidence for caveolae as transendothelial carriers by showing the sequential movements of probes from one side of the endothelial layer to the other. Since then, many molecular and cellular studies have brought new insights into caveolar structure and function, and these findings shed new light on many areas of cellular physiology, including signal transduction. Caveolae are enriched for components of signaling cascades, such as heterotrimeric G proteins, Src family kinases, platelet-derived growth factor receptors, epidermal growth factor receptors, and PKC (9, 26, 32, 35). Thus caveolae could be the main platform for signal transduction in endothelial cells.
The fractionation method was able to effectively separate the heavier membrane fractions containing caveolae (Fig. 1). Notably, a relatively larger subfraction of PLD2 is also enriched in caveolin-1-containing caveolae, although only a trace amount of PLD2 is present in noncaveolar fractions. In comparison, our data indicate that PLD1 is evenly distributed in noncaveolar fractions. Interestingly, Seymour et al. (31) suggested that VEGF-induced PLD activity in endothelial cells, although they did not define which type of VEGFR or PLD is involved. Therefore, we hypothesized that VEGFR-2 might regulate PLD activity through activation of caveolae-enriched PLD2. Indeed, VEGF-induced VEGFR-2 activation rapidly increased PLD activity in a dose-dependent manner, whereas PlGF-induced VEGFR-1 activation did not produce any change in PLD activity. Moreover, inhibition of VEGFR-2 activation with SU5614 completely abolished VEGF-induced PLD activity. Unfortunately, there are currently no reliable methods for discrimination between PLD1 and PLD2 activity. However, considering that PLD2 is localized in the endothelial caveolae, it is likely involved in VEGFR-2-induced activation of total PLD activity (Fig. 7).
PLD catalyzes the hydrolysis of phosphatidylcholine to produce PA and choline. PA itself can then be converted to diacylglycerol or lysophosphatidic acid, second messengers that activate various downstream signaling events (5, 11, 20). PLD is known to be involved in one of the MAPK pathways, specifically the pathway by which Raf-1 acts on MEK and ERK (29). To test this possible pathway in endothelial cells, the effect of a PLD inhibitor, 1-butanol, on VEGF-induced ERK phosphorylation was examined. The addition of 1-butanol almost totally suppressed VEGF-induced ERK phosphorylation, whereas it did not produce any significant change in VEGF-induced Akt. Moreover, the negative control for 1-butanol, 3-butanol, did not produce a significant change in VEGF-induced MEK and ERK phosphorylations. Although we extensively tried to ablate PLD1 or PLD2 expression specifically using anti-sense and interference RNA technologies, these experiments did not work well. Therefore, we cannot exclude the possibility that 1-butanol has a nonspecific effect on VEGF-induced MEK and ERK phosphorylations. However, given that addition of PA restored the 1-butanol-induced suppressive effect, VEGF-induced PA formation through activation of PLD activity may be involved in VEGF-induced ERK phosphorylation in endothelial cells (Fig. 7). Furthermore, our results indicated that PLD inhibition with 1-butanol suppressed both VEGF-induced ERK activation and VEGF-induced proliferation in endothelial cells. These two effects were temporally correlated. Given that VEGF-induced ERK activation is closely involved in the proliferation of endothelial cells (21, 28, 38), VEGF-induced PLD/ERK activation could be important for proliferation of endothelial cells. Considering that VEGF-induced proliferation of endothelial cells is one of essential steps in VEGF-induced angiogenesis, PLD in endothelial cells could play an important role in angiogenesis.
Two phosphatidylcholine-specific mammalian isoforms of PLD (PLD1 and PLD2) have been isolated and characterized (5, 20). Currently, specific PLD1 and PLD2 inhibitors are not available. PLD1 and PLD2 are activated differently by ARF, Rho, and PKC in different cell types (4, 5, 11, 12, 16, 17, 20, 31). Our data demonstrate that pretreatment with PKC inhibitor totally suppressed VEGF-induced PLD activity and ERK phosphorylation, whereas pretreatment with ARF inhibitor did not change VEGF-induced PLD activity or ERK phosphorylation. Thus, similar to previous finding (31), VEGF-induced PLD activity and ERK phosphorylation are totally dependent on PKC activation. Therefore, we further examined which PKC isoenzyme is mainly responsible for VEGF-induced PLD/ERK2 activations. With the use of several inhibitors, our analyses indicate that the PKC-δ isoenzyme mainly regulates these activations. We previously demonstrated that PKC-δ mediates the direct phosphorylation-dependent activation of PLD2 in PC12 cells (12). Likewise, PLD2 may be involved in VEGF/VEGFR2/PKC-δ-induced MEK/ERK signaling cascades in endothelial cells (Fig. 7).
We next investigated whether intact endothelial caveolae are essential for localization and signaling of VEGFR-2. When the caveolae were disassembled, the localization of VEGFR-2 was shifted to noncaveolar fractions. Furthermore, under conditions that disassemble caveolae, VEGF-induced MEK/ERK activation was largely abolished. This observation is similar to that in a recent study (19). In addition, under conditions that reassemble caveolae, VEGF-induced MEK/ERK activation almost completely recovered. This result indicates that intact caveolae are essential and critical as a platform for VEGFR-2-induced MEK/ERK signaling cascades.
Taken together, two important findings regarding VEGFR-2 signaling in vascular endothelial cells were found in this study. First, PKC-δ/PLD2 could be partly involved in VEGF/VEGFR-2-induced MEK/ERK activations and cellular proliferation. Second, normal caveolar structure is essential and critical for VEGF-induced MEK/ERK activations.
We thank Jennifer Macke for help in preparing the manuscript.
This work was supported by the Bio-Challenge Program of the Korean Ministry of Science and Technology.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by the American Physiological Society