Pro-angiogenic effects of resveratrol in brain endothelial cells: nitric oxide-mediated regulation of vascular endothelial growth factor and metalloproteinases

Abstract

Resveratrol may be a powerful way of protecting the brain against a wide variety of stress and injury. Recently, it has been proposed that resveratrol not only reduces brain injury but also promotes recovery after stroke. But the underlying mechanisms are unclear. Here, we tested the hypothesis that resveratrol promotes angiogenesis in cerebral endothelial cells and dissected the signaling pathways involved. Treatment of cerebral endothelial cells with resveratrol promoted proliferation, migration, and tube formation in Matrigel assays. Consistent with these pro-angiogenic responses, resveratrol altered endothelial morphology resulting in cytoskeletal rearrangements of β-catenin and VE-cadherin. These effects of resveratrol were accompanied by activation of phosphoinositide 3 kinase (PI3-K)/Akt and Mitogen-Activated Protein Kinase (MAPK)/ERK signaling pathways that led to endothelial nitric oxide synthase upregulation and increased nitric oxide (NO) levels. Subsequently, elevated NO signaling increased vascular endothelial growth factor and matrix metalloproteinase levels. Sequential blockade of these signaling steps prevented resveratrol-induced angiogenesis in cerebral endothelial cells. These findings provide a mechanistic basis for the potential use of resveratrol as a candidate therapy to promote angiogenesis and neurovascular recovery after stroke.

Introduction

It is increasingly recognized that neurovascular responses have a critical role in the ability of the brain to recover after stroke. Signals and substrates of neurogenesis and neuroplasticity are tightly coregulated with angiogenesis and vascular remodeling (Arai et al, 2009). This may not be surprising since the fundamental mechanisms of neurogenesis and angiogenesis are evolutionarily conserved so that similar mediators are involved in both phenomena (Carmeliet and Tessier-Lavigne, 2005). After stroke, newborn neuroblasts migrate along perivascular routes (Thored et al, 2007) and promotion of angiogenesis enhances neurogenesis and functional recovery (Ohab et al, 2006; Taguchi et al, 2004). Neurovascular responses have a central role as the damaged central nervous system transitions from initial injury into repair (Lo, 2010). Hence, therapies that enhance neurovascular remodeling should provide new opportunities for stroke recovery (Xiong et al, 2010).

Accumulating data suggest that resveratrol is a potent neuroprotectant against stroke (Dong et al, 2008; Sakata et al, 2010; Shin et al, 2010; Sun et al, 2010). The underlying mechanisms are pleiotropic, including modulation of sirtuin signaling, upregulation of pro-survival mediators such as FOXO and MnSOD, dampening cell death signals such as bax and caspases (Della-Morte et al, 2009; Pallàs et al, 2009). More recently, however, it has been proposed that besides acute neuroprotection, resveratrol may also have beneficial effects on stroke recovery (Dong et al, 2008). But the mechanisms that might mediate these phenomenon remain unclear.

In the present study, we test the hypothesis that resveratrol can promote angiogenesis in cerebral endothelial cells. Our findings suggest that resveratrol activates phosphoinositide 3 kinase (PI3-K)/Akt and Mitogen-Activated Protein Kinase (MAPK)/ERK signaling pathways that upregulate endothelial nitric oxide synthase (eNOS) and increases extracellular levels of vascular endothelial growth factor (VEGF) and matrix metalloproteinases (MMPs). These data suggest that resveratrol may be a candidate therapy for stroke recovery by promoting angiogenesis via a programmed NO/VEGF/MMP signaling axis.

Materials and methods

Reagents and Cells

Resveratrol (>99%) was purchased from Sigma-Aldrich (St Louis, MO, USA). A stock solution of resveratrol was made in dimethyl sulfoxide at a concentration of 10 mM. The following pharmacologic agents were used: MEK inhibitor—U0126 (Calbiochem, San Diego, CA, USA), a PI3-K inhibitor—LY294002 (Calbiochem), VEGF antagonist (sFlt-1) (Calbiochem), VEGF receptor blocker—SU1498 (Calbiochem). A human cerebral microvascular endothelial cell line (Callahan et al, 2004) was seeded at 60% to 70% confluence and kept at 37°C in 5% CO₂. Culture media comprised RPMI 1640 containing 10% fetal bovine serum, 10% Nu-Serum, 2 mM L-glutamine, 1 mM pyruvate, essential amino acids, and vitamins. To further confirm the endothelial characteristics of this cell, we used Western blotting to document CD31, von Willebrand and VE-cadherin (Supplementary Figure 1). Our results were also replicated in a second line, the well-studied rat brain endothelial cell RBE4 (Supplementary Figure 2) (Wu et al, 2010).

Proliferation Assay

For in vitro proliferation assays, brain endothelial cells were seeded into 96-well (1 × 10⁵ cells/well) flat bottom plates in triplicate with medium alone (control) or medium containing different concentrations of recombinant resveratrol (0.1 to 10 μM). Cell proliferation was determined with the WST proliferation assay kit (Dojindo, Japan). Briefly, serum-starved cells were treated with resveratrol (0.1 to 10 μM) for 24 hours. Following resveratrol treatment, 10 μL WST reagent was added to 100 μL fresh culture medium in each well. Absorbance was recorded at 490 nm (EL800, Bio-Tek Instruments, Winooski, VT, USA), and calculated as a ratio against untreated cells.

Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis Zymography

Conditioned media were collected and concentrated via centricon (Millipore, Bedford, MA, USA). Equal amounts were mixed with 2 × sample buffer and loaded on a 10% polyacrylamide gel incorporated with 0.1% gelatin for electrophoresis. The MMP-2 and MMP-9 zymographic standards were used as positive controls (Chemicon, Bedford, MA, USA). After electrophoresis, gels were renaturated and then incubated with developing buffer at 37°C for 72 hours, and stained for 1 hour with 0.1% Coomassie blue. Gelatinolytic activity was quantified via densitometry.

Matrigel Tube Formation Assay for Angiogenesis

The standard Matrigel assay was used to assess the spontaneous formation of capillary-like structures in vitro. Cells (8 × 10⁴ cells/well) were seeded in 24-well plates in serum-free media previously coated with Growth Factor-Reduced Matrigel Matrix (BD Bioscience, San Jose, CA, USA), and incubated at 37°C for 18 hours. The degree of tube formation was determined by counting the number of tubes in four random fields from each well. Data were analyzed as tube percentage versus untreated control wells.

Endothelial Cell Migration Assay

Cells (1 × 10⁵) in RPMI containing 0.5% FBS were plated onto the upper transwell chamber (6.5 mm, 8 μM pore Corning Costar, Cambridge, MA, USA) coated with 0.1% collagen Type I (BD Bioscience). The bottom chamber contained 700 μL of medium (serum free) alone or medium containing different concentrations of resveratrol. The cells were incubated for 18 hours at 37°C and cells that did not pass through the membrane pores were removed. Migrated cells were stained with crystal violet and counted in five random fields ( × 200).

Nitric Oxide and Vascular Endothelial Growth Factor Assays

Nitric oxide concentrations in conditioned media were determined using the QuantiChrom Nitric Oxide Assay Kit (BioAssay Systems, Hayward, CA, USA). Vascular endothelial growth factor levels in supernatants were determined using a human VEGF ELISA kit (R&D System, Minneapolis, MN, USA). Known concentrations of VEGF (0 to 500 pg/mL) were used to generate standard curves.

Immunoblotting

Cells were lysed in lysis buffer (Cell Signaling Technology, Danvers, MA, USA) and clarified, and protein concentration was determined with the Bradford assay (Bio-Rad, Hercules, CA, USA). Total cell lysates (50 to 100 μg per lane) were separated in precast 4% to 12% NuPage gels (Invitrogen, Carlsbad, CA, USA), and proteins were transferred to nitrocellulose membranes (Invitrogen). After blocking with 0.2% I-block (Tropix, Applied Biosystem, CA, USA), membranes were incubated overnight at 4°C with primary antibodies, and 1 hour at room temperature with horseradish peroxidase-conjugated secondary antibodies (Amersham, Piscataway, NJ, USA). The immune complexes were visualized by enhanced chemiluminescence (Amersham). All immunoblots were repeated for at least three independent experiments.

Statistical Analysis

All of the experiments were performed in duplicate or triplicate, and repeated three to four times independently. Quantitative data were expressed as mean±s.d. and analyzed with analysis of variance followed by Tukey's test for multiple comparisons. Differences of P<0.05 were considered significant. All experiments were performed under an institutionally approved protocol.

Results

Resveratrol Augments Proliferation, Migration, and Tube Formation

Cerebral endothelial cells were incubated with different concentrations of resveratrol (0.1 to 10 μM) over 24 hours. Proliferation, migration, and matrigel tube formation assays were examined as in vitro markers of angiogenesis. Resveratrol induced a dose-dependent increase in endothelial proliferation (Figure 1A). Flow cytometry analysis confirmed that resveratrol induced a significant increase in the proliferative S phase while decreasing the resting G₀/G₁ phase of the cell cycle (Figure 1B). Along with increased proliferation, resveratrol also enhanced endothelial cell migration as quantified with a transwell assay (Figures 1C and 1D). Matrigel assays showed that resveratrol induced tube formation in a dose-dependent manner (Figures 1E and 1F). In all proliferation, migration, and tube formation assays, the most effective concentrations of resveratrol appeared to peak around 5 μM. Similar results were obtained with a rat brain endothelial cell line RBE4 (Supplementary Figure 2).

Figure 1.

Resveratrol stimulates brain endothelial cell proliferation, migration, and tube formation. (A) Serum-starved brain endothelial cells were treated with resveratrol (0.1 to 10 μM) for 24 hours and cell growth was assessed for proliferation assay by WST. Bar graph shows that the proliferation of brain endothelial cells incubated with indicated concentration of resveratrol increased. Results were expressed as percentage of cells in relation to basal (untreated) and represent the mean of triplicates. ^*P<0.05 versus control, n=4. (B) Effect of resveratrol on cell-cycle distribution. Brain endothelial cells were treated with resveratrol for 24 hours. Flow cytometry analysis showed that resveratrol significantly increased the proportion of cells in S phase (gray bar) and decreased the proportion of cells in the G₀/G₁ phase (black bar). Data are presented as the mean±s.d. ^*P<0.05, ^**P<0.01 versus control. (C) Migration of serum-starved brain endothelial cells in response to resveratrol (0.1 to 10 μM) were examined using transwell chamber assay for 10 hours. Cells undergoing migration in response do resveratrol were captured using a light microscope at × 200 magnification. (D) Increased migration of brain endothelial cells treated with resveratrol. The values represent the average number of cells migrated (mean±s.d.) of three independent experiments in duplicate. ^*P<0.01 versus control. (E) Resveratrol promoted matrigel angiogenesis. Brain endothelial cells (8 × 10⁴) were plated on Matrigel-coated, 24-well plates and were incubated for 20 hours in medium alone or with resveratrol. Photomicrographs represent the matrigel tube formation after 20-hour incubation detected by phase-contrast microscopy (magnification, × 200). (F) Bar graph shows that resveratrol promoted endothelial tube formation. ^*P<0.05, ^**P<0.01 versus control. Each experiment was performed four times in duplicate. The results are expressed as mean±s.d. (G) Cytoskeleton rearrangement in brain endothelial cells following treatment with resveratrol. F-actin rearrangement in brain endothelial cells in response to resveratrol (5 μM) was examined using immunofluorescence analysis by phalloidin staining. Endothelial cells with media alone showed stress fibers aligned across the cell axis. Stress fibers disappeared forming membrane ruffles in brain endothelial cells treated with 5 μM of resveratrol in media for 20 minutes. Immunocytochemistry to show localization of β-catenin and VE-cadherin in endothelial cells treated with resveratrol was performed. β-Catenin was localized on the endothelial membrane of cells treated with media alone, whereas it was found more in the cytoplasm and nucleus of cells treated with 5 μM of resveratrol. VE-cadherin was localized at the intercellular junctions of cells treated with media alone, whereas it was found more in the cytoplasm of cells treated with 5 μM of resveratrol.

Immunostaining confirmed that resveratrol induced morphological responses that were consistent with a pro-angiogenic phenotype in cerebral endothelial cells. In control untreated cells, normal resting state profiles were observed with F-actin mostly found in unorganized fibers and membrane structures, β-catenin localized to the cytoplasm, and VE-cadherin localized at intercellular junctions (Figure 1G). After resveratrol treatments, cytoskeletal rearrangements were observed with edge ruffling and F-actin stress fiber formation, relocation of β-catenin to the nucleus, and movement of VE-cadherin from membrane junctions into the cytoplasm (Figure 1G).

Resveratrol Upregulates Nitric Oxide via PI3-Kinase/Akt and MAPK/ERK Signaling

Resveratrol robustly activated PI3-kinase signaling in cerebral endothelial cells, leading to a rapid increase in phospho-Akt levels (Figures 2A and 2B). Activation of this pathway was significantly blocked by the specific PI3-kinase inhibitor LY294002 (Figures 2C and 2D). Additionally, the ERK MAP kinase pathway was also triggered by resveratrol, with clear increases in phospho-ERK levels (Figures 2E and 2F). ERK signaling was pharmacologically confirmed with the MEK/ERK inhibitor U0126, which significantly lowered resveratrol-induced phospho-ERK levels (Figures 2G and 2H).

Figure 2.

Regulation of phosphoinositide 3-Kinase (PI3-K) and Mitogen-Activated Protein Kinase (MAPK) signaling pathway by resveratrol. (A) Exposure to 5 μM resveratrol triggered PI3-K/Akt signaling as indicated by elevated levels of phospho-Akt for indicated time periods. Representative images from four individual experiments are shown. (B) Bar graph represent p-Akt/total Akt levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05, ^**P<0.01 versus control. (C) Serum-starved cells were pretreated with a selective PI3-K inhibitor (LY294002, 10 μM) for one hour before incubation with 5 μM resveratrol for 10 minutes. Representative blot from four individual experiments are shown. (D) Bar graph represent p-Akt/total Akt levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05, ^**P<0.01 versus control. (E) Serum-starved brain endothelial cells were treated with 5 μM resveratrol for up 30 minutes and Erk phosphorylation was measured by Western blot analysis. Representative images from four individual experiments are shown. (F) Bar graph represent p-Erk/total Erk levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05, ^**P<0.01 versus control. (G) Serum-starved brain endothelial cells were pretreated with a selective MEK inhibitor (U0126, 10 μM) for 1 hour before incubation with 5 μM resveratrol for 10 minutes. Representative images from four individual experiments are shown. (H) Bar graph represent p-Erk/total Erk levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05, ^**P<0.01 versus control. The cell lysates were subjected to Western blotting with the indicated antibodies. The results are represented as a ratio of phosphorylated to total protein and expressed as percentage of control.

Since Akt and ERK pathways are known to be related to NO signaling, we next examined the role of eNOS in resveratrol-induced angiogenesis. Incubation of cerebral endothelial cells with resveratrol rapidly increased phosphorylation of eNOS (Figures 3A and 3B). Activation of eNOS occurs downstream of PI3-K/Akt and Mitogen-Activated Protein Kinase/ERK signaling since blocking these pathways with LY294002 or U0126 significantly decreased resveratrol-induced phospho-eNOS levels (Figures 3C and 3D). Consistent with the increased phosphorylation of eNOS, resveratrol amplified NO levels in endothelial-conditioned media (Figure 3E). Cotreatment with the NOS inhibitor L-NAME suppressed the resveratrol-induced NO response (Figure 3F). The sequential activation of this PI3-kinase/Akt-ERK-eNOS cascade was confirmed with further inhibitor experiments. Cotreatments with either the PI3-K/Akt inhibitor (LY294002) or the MEK/ERK inhibitor (U0126) significantly decreased resveratrol-induced NO levels (Figure 3F). Finally, matrigel assays demonstrated that these pathways were required for resveratrol-induced angiogenesis in cerebral endothelial cells. Blockade of any of these steps in the signaling cascade (PI3-K/Akt, ERK, or eNOS) significantly suppressed resveratrol-induced tube formation (Figures 4A and 4B). Similar results were obtained with a rat brain endothelial cell line RBE4. ERK and AKt signaling was activated by resveratrol and blockade with U0126 or LY294002 decreased resveratrol-induced tube formation (Supplementary Figure 2).

Figure 3.

Endothelial nitric oxide synthase (eNOS) is regulated by resveratrol and downstream of phosphoinositide 3-Kinase (PI3-K) and Mitogen-Activated Protein Kinase (MAPK) signaling pathway. (A) Serum-starved brain endothelial cells were treated with 5 μM resveratrol for indicated time period. Representative images from four individual experiments are shown. (B) Bar graph represent p-eNOS/total eNOS levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05 versus control. (C) Serum-starved brain endothelial cells were pretreated with a selective PI3-K (LY294002, 10 μM) and MEK (U0126, 10 μM) inhibitor for 1 hour before incubation with 5 μM resveratrol for 20 minutes. Representative images from four individual experiments are shown. (D) Bar graph represent p-eNOS/total eNOS levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05, ^**P<0.01 versus control. eNOS phosphorylation on Ser1177 was measured by in-cell Western blot analysis using a specific antibody as described in the Materials and methods section. The results are represented as a ratio of phosphorylated to total protein and expressed as percentage of control. (E) Serum-starved brain endothelial cells were treated with 5 μM resveratrol for up to 30 minutes. The stable end product of nitric oxide, nitrite, was measured in the medium using the Griess reaction. Medium total nitrite content corrected for protein content of the cell layer was estimated according with the company product. Composite data from four experiments are shown relative to control (mean±s.d.). ^*P<0.05 versus control. (F) Serum-starved cells were pretreated with a selective PI3-K (LY294002, 10 μM), MEK (U0126, 10 μM), and NOS (L-NAME, 3 mM) inhibitor for 1 hour before incubation with 5 μM resveratrol for 20 minutes. Composite data from four experiments are shown relative to control (mean±s.d.). ^*P<0.05 versus control, ^#P<0.05 versus resveratrol group.

Figure 4.

Involvement of phosphoinositide 3-Kinase (PI3-K), Mitogen-Activated Protein Kinase (MAPK), and nitric oxide synthase (NOS) signaling pathway in resveratrol-induced capillary tube formation. (A) Serum-starved brain endothelial cells were pretreated with a selective PI3-K (LY294002, 10 μM), MEK (U0126, 10 μM), and NOS (L-NAME, 3 mM) inhibitors for 1 hour before incubation with 5 μM resveratrol. Photomicrographs represent the matrigel tube formation after 20-hour incubation detected by phase-contrast microscopy (magnification, × 200). (B) Resveratrol-induced brain endothelial cell capillary tube formation was significantly inhibited when cells were preincubated with inhibitors. Results are mean values±s.d. of four experiments in duplicate. ^*P<0.05 versus control; ^#P<0.05 versus resveratrol group.

Resveratrol Increases Vascular Endothelial Growth Factor and Matrix Metalloproteinases Via Nitric Oxide-Dependent Signaling

To directly link resveratrol-induced signaling with angiogenesis, we assessed the well-established pro-angiogenic mediator VEGF. Immunostaining (Figure 5A) and Western blots (Figures 5B and 5C) showed that VEGF expression in cerebral endothelial cells was strongly upregulated by resveratrol. This effect was dependent on NO. Inhibition of NOS with L-NAME significantly decreased the ability of resveratrol to upregulate VEGF (Figures 5D and 5E).

Figure 5.

Vascular endothelial growth factor (VEGF) induced by resveratrol is nitric oxide synthase (NOS) dependent. (A) Serum-starved brain endothelial cells were treated with 5 μM resveratrol for 24 hours. Immunocytochemistry for VEGF demonstrated a strong stained for resveratrol–brain endothelial cells. (B) Serum-starved cells were treated with 5 μM resveratrol for indicated time periods and performed Western blotting. Representative blots from four individual experiments are shown. Deferoxamine (DFX) was used as a positive control for VEGF generation. (C) Bar graph represent VEGF/β-actin levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05, ^**P<0.01 versus control group. The cell lysate were collected and Western blotting was performed. (D) VEGF induced by resveratrol is NOS dependent. Serum-starved cells were pretreated with a selective NOS inhibitor (L-NAME, 3 mM) for 1 hour before incubation with 5 μM resveratrol for 24 hours. Representative blots from four individual experiments are shown. (E) Bar graph represent VEGF/β-actin levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05 versus control group; ^#P<0.05 versus resveratrol group. (F) VEGF levels in the conditioned media, as confirmed by ELISA, were also higher in resveratrol and DFX-treated cells. The VEGF concentrations were calculated from the standard curve. ^*P<0.05, ^**P<0.01 versus control. (G) VEGF levels in the conditioned media, as confirmed by ELISA, were higher in resveratrol and decreased when cells were pretreated with a selective NOS inhibitor (L-NAME, 3 mM). ^*P<0.05 versus control; ^#P<0.05 versus resveratrol group. (H) Serum-starved cells were pretreated with a selective VEGF antagonist (sFLT-1, 10 μM) for 1 hour before incubation with 5 μM resveratrol for 24 hours. Representative blots for p-VEGFR2/total VEGFR2 from four individual experiments are shown. (I) Bar graph represent p-VEGFR2/total VEGFR2 levels expressed as the average percentage increase (mean±s.d.) over basal levels (n=4). ^*P<0.05 versus control group; ^#P<0.05 versus resveratrol group.

Cellular upregulation of VEGF was accompanied by a release of protein into extracellular space. Vascular endothelial growth factor levels in endothelial-conditioned media were clearly increased by resveratrol (Figure 5F). Extracellular VEGF release was also NO dependent since L-NAME significantly decreased resveratrol-induced NO in conditioned media (Figure 5G).

Consistent with elevated protein levels, activation of VEGF signaling was detected in resveratrol-treated endothelial cells. Levels of the phosphorylated receptor phospho-VEGFR2 was increased by resveratrol, indicating that active signaling was indeed taking place (Figures 5H and 5I). These pathways involved autocrine signaling since blockade of the VEGF with sFlt-1 dampened the ability of resveratrol to activate the VEGF pathway and phosphorylate VEGFR2 (Figures 5H and 5I).

Besides VEGF, another major class of pro-agiogenic mediators comprise MMPs. Exposure of cerebral endothelial cells to resveratrol upregulated both MMP-2 and MMP-9 (Figure 6A). Induction of MMPs was also dependent on NO since the inhibitor L-NAME prevented resveratrol from upregulating both MMP-2 and MMP-9 (Figures 6B and 6C).

Figure 6.

Effect of resveratrol on matrix metalloproteinase (MMP)-2 and MMP-9 secretion. Gelatin zymograph analysis (A) shows MMP-2 and MMP-9 activities in the conditioned medium for 24 hours harvested from control and different concentrations of resveratrol (0.1 to 10 μM). (B) Serum-starved brain endothelial cells were pretreated with a selective vascular endothelial growth factor (VEGF) (sFLT-1, 10 μM), VEGFR2 (SU1498, 5 μM), and nitric oxide synthase (NOS) (L-NAME, 3 mM) inhibitors for 1 hour before incubation with 5 μM resveratrol for 24 hours. VEGF was used as a positive control. Representative zymograms from four individual experiments are shown. (C) Bar graph represent MMP-2 and MMP-9 levels expressed as the average percentage increase over basal levels (n=4). ^*P<0.05 versus control group; ^#P<0.05 versus resveratrol group. Data shown are mean values±s.d. The values represented are relative to basal.

Crosstalk between VEGF and MMP signaling may exist. Exposing cerebral endothelial cells to VEGF directly upregulated MMP levels (Figures 6B and 6C). And the ability of resveratrol to induce MMPs can be prevented by either antagonizing VEGF with sFlt-1 or a blocking the VEGFR2 receptor with SU1498 (Figures 6B and 6C). Taken together, these data demonstrated that resveratrol strongly activates both VEGF and MMP pathways.

Resveratrol-Induced Angiogenesis Is Dependent on Vascular Endothelial Growth Factor and Matrix Metalloproteinases

Next, we asked whether resveratrol-induced angiogenesis in cerebral endothelial cells was indeed dependent on the control of VEGF and MMP mechanisms. As expected, resveratrol increased tube formation in matrigel assays. Blocking either VEGF or MMP signaling potently suppressed these resveratrol-induced effects. Cotreatments with the VEGF antagonist sFlt-1, or the VEGFR2 blocker SU1498, or the broad spectrum MMP inhibitor GM6001, all significantly decreased resveratrol-induced tube formation (Figures 7A and 7B).

Figure 7.

Involvement of vascular endothelial growth factor (VEGF), VEGFR2, and matrix metalloproteinases (MMPs) signaling pathway in resveratrol-induced capillary tube formation. (A) Serum-starved brain endothelial cells were pretreated with a selective VEGF (sFLT-1, 10 μM), VEGFR2 (SU1498, 5 μM), and MMPs (GM6001, 10 μM) inhibitors for 1 hour before incubation with 5 μM resveratrol. Photomicrographs represent the matrigel tube formation after 20-hour incubation detected by phase-contrast microscopy (magnification, × 200). (B) Resveratrol-induced brain endothelial cell capillary tube formation was significantly inhibited when cells were preincubated with inhibitors. Results are mean values±s.d. of four experiments in duplicate and expressed as percentage of control. ^*P<0.05 versus control; ^#P<0.05 versus resveratrol group.

Discussion

Resveratrol is a potent neuroprotectant that has been shown to reduce neuronal death in many experimental models of stroke and brain injury (Dong et al, 2008; Sakata et al, 2010; Shin et al, 2010; Sun et al, 2010). But emerging data now suggest that resveratrol may also be used to promote recovery after stroke (Dong et al, 2008). The major finding of the present study is that resveratrol can promote angiogenesis in cerebral endothelial cells. The mechanisms of this phenomenon appear to involve upstream control of NO via PI3-kinase/Akt and ERK signaling, and downstream induction of VEGF and MMPs (Figure 8). These data provide a mechanistic basis for the potential application of resveratrol as a candidate therapy for neurovascular repair.

Figure 8.

A schematic diagram showing the proposed mechanism by which intracellular signaling pathway involved in resveratrol-induced brain endothelial cell angiogenesis.

Although several groups have previously reported generation of NO by resveratrol (Andriambeloson et al, 1997; Bruder et al, 2001; Hsieh et al, 1999; Hung et al, 2000; Wallerath et al, 2002); the precise signaling pathways are not well established. Resveratrol upregulated eNOS expression and NO generation in Human Umbilical Vein Endothelial Cells (HUVEC) (Wallerath et al, 2002). But in our study, the major effects of resveratrol appeared to take place via phosphorylation rather than absolute alterations of protein levels. Endothelial nitric oxide synthase was constitutively expressed in cerebral endothelial cells but the basal level of phosphorylation was minimal. Resveratrol increased ser1177 phosphorylation of eNOS within minutes. Ser1177 phosphorylation is critical for eNOS activity (Fulton et al, 1999). In our experiments, resveratrol-induced eNOS phosphorylation led to prompt generation of NO. Of note, this response was rapid but transient so downstream effectors may be required to extend the biologic effect.

To dissect how resveratrol regulates eNOS phosphorylation, we assessed two major kinase pathways, PI3-kinase and ERK MAP kinase. Incubation of endothelial cells with the PI3-kinase inhibitor LY294002 or the MEK/ERK inhibitor U0126, both abrogated eNOS phosphorylation induced by resveratrol. It is interesting to note that activation of PI3-kinase and ERK are required for angiogenesis, since either PI3-kinase or ERK inhibitors reduced resveratrol-induced tube formation. Functionally, the balance between these two signals may allow fine-tuning of eNOS activation, and the redundancy of the two signaling pathways may ensure robust NO generation. Collectively, our observations highlight the importance of PI3-kinase/Akt and ERK pathways in resveratrol-induced NO production, and both appear to be necessary for cerebral endothelial angiogenesis.

It is well known that many important mediators execute the angiogenic program downstream of NO. Acting as a messenger molecule, NO mediates the majority of endothelium-dependent responses in the brain (Faraci and Heistad, 1998). Treatment of cells with NO donors increases VEGF (Zhang et al, 2003), and inhibitors of NO synthase such as L-NAME can block VEGF generation (Chen et al, 2005). Endothelial nitric oxide synthase-deficient mice have impaired angiogenesis as well as impaired neurogenesis and recovery of neuronal functional following experimental stroke (Murohara et al, 1998). Limbs from eNOS knockout mice exhibit significant impairment in angiogenesis response, suggesting that NO can induce angiogenesis through VEGF (Francis et al, 2001). Hence, VEGF may be an important mechanism by which resveratrol induces angiogenesis. Sen et al (2002) demonstrated upregulation of VEGF expression at dermal wound site with a combination of resveratrol and grape seed proanthocyanidins. In addition, Wang et al (2007) reported that resveratrol increased VEGF production and further induced vasculogenesis in endothelial progenitor cells. Consistent with these previous reports, our results demonstrate the induction of NO and VEGF by resveratrol. The importance of these mechanisms is confirmed by the fact that blockade of any of these NO or VEGF signaling steps potently suppressed resveratrol-induced angiogenesis in our cerebral endothelial models. Furthermore, positive feedback loops may also be involved. The receptor VEGFR2 can also be upregulated by VEGF stimulation, leading to enhanced VEGF signaling and angiogenesis (Yang et al, 2002). Our data confirm that resveratrol upregulates both VEGF and VEGFR2, and inhibition of upstream NO signals can dampen all components of the VEGF response and angiogenesis. Of course, it should be noted that although we focus on VEGFR2, it remains possible that interactions with other VEGF receptors may also be involved.

Besides VEGF, neurovascular proteolysis via MMPs represent another major mechanism for angiogenesis (Karagiannis and Popel, 2005; Moses, 1997; Segarra et al, 2005). To form new blood vessels, activated endothelial cells must degrade the basement membrane and remodel the extracellular matrix around neovascular sites. Our findings reveal that resveratrol directly upregulated MMP-2 and MMP-9 production in cerebral endothelial cells, whereas inhibition of MMPs significantly reduced resveratrol-mediated angiogenesis. Of course, the inhibitor used here, GM6001, is a broad spectrum agent. So we cannot exclude the possibility that besides MMP-2 and MMP-9, other MMPs may also be involved. Taken together, these findings support a causal link between resveratrol-induced MMP responses and angiogenesis in cerebral endothelial cells. Furthermore, MMPs are known to be regulated by a number of cytokines and growth factors (Moses, 1997). In our cerebral endothelial system, crosstalk between MMPs and VEGF was apparent. Inhibition of VEGF signaling potently suppressed resveratrol-induced MMPs, and prevented angiogenesis.

Taken together, our findings suggest that resveratrol may be a novel way to promote angiogenesis in cerebral endothelial cells. But there are several important caveats to keep in mind. First, although our data provide cellular and pharmacologic proof-of-principle for resveratrol in cerebral angiogenesis, in vivo validation of these mechanisms remain to be obtained. The pro-angiogenic utility of resveratrol as a potential stroke recovery therapy should be explored in future experiments. An important limitation, however, may be related to the poor bioavailability of oral resveratrol dosing that can only achieve 10 to 50 ng/mL levels in plasma. Our 5 μM concentrations here would correspond to almost 1 μg/mL levels. Second, our findings are only focused on endothelium. But it is now recognized that multicellular interactions between endothelial cells and astrocytes and pericytes are required for neurovascular homeostasis (Abbott et al, 2006). Further dissection of how resveratrol affects cell–cell signaling in the entire neurovascular unit is warranted. Third, although our data define downstream effects, we do not know what receptors of resverstrol might be involved. Resveratrol can activate a wide range of receptors including androgen/estrogen receptors, EGF receptors, and integrins etc. (Han et al, 2006). Fourth, our studies focused on VEGF and MMP mechanisms. But angiogenesis is known to recruit complex networks of genes and proteins (Muñoz-Chápuli et al, 2004). More powerful proteomic approaches may be required to fully catalog the effects of resveratrol on cerebral endothelium. Finally, our studies only examined the pro-angiogenic effects of resveratrol in normal cells. In the context of stroke, brain injury or neurodegeneration, future studies should ask how resveratrol may operate in the milieu of a damaged neurovascular system.

In summary, the present study provides mechanistic evidence that resveratrol promotes angiogenesis in cerebral endothelial cells via NO, VEGF, and MMP signaling. Further in vivo and clinical exploration of these pathways is warranted to validate these experimental findings and develop resveratrol as a potential neurovascular repair therapy for stroke and brain injury.

The authors declare no conflict of interest.