VEGF stimulates PKD-mediated CREB-dependent orphan nuclear receptor Nurr1 expression: Role in VEGF-induced angiogenesis

Authors

  • Dezheng Zhao,

    Corresponding author
    1. Division of Gastroenterolgy, Center for Vascular Biology Research, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA
    • Beth Israel Deaconess Medical Center, Center for Vascular Biology Research, 99 Brookline Avenue, RN270J, Boston, MA 02215, USA
    Search for more papers by this author
    • Tel.: 617-667-0466, Fax: 617-667-3591

  • Sachin Desai,

    1. Division of Gastroenterolgy, Center for Vascular Biology Research, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA
    Search for more papers by this author
  • Huiyan Zeng

    1. Division of Vascular Medicine, Center for Vascular Biology Research, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA
    Search for more papers by this author

Abstract

New vessel formation is critical for solid tumor growth and it is primarily stimulated by the most potent angiogenic factor vascular endothelial growth factor (VEGF or VEGF-A165). VEGF promotes endothelial cell proliferation by initiating signaling cascades to increase gene transcription. Recent works showed that VEGF potently and rapidly induces expression of orphan nuclear receptor Nurr1 in endothelial cells. However, the signaling pathway for VEGF-induced Nurr1 expression and its role in VEGF-induced endothelial cell proliferation and angiogenic response have not been examined. In our study, we first show that VEGF significantly induces expression of Nurr1 mRNA, protein and its promoter activity in cultured endothelial cells. Furthermore, the promoter analysis shows that deletion of the putative cAMP-responsive element binding protein (CREB) site in the proximal region of the promoter markedly reduces VEGF-induced promoter activity whereas deletion of the upstream NF-κB site has moderate effect. Transfection of a dominant negative CREB mutant (K-CREB) or mutation of this putative CREB site in the Nurr1 promoter attenuates VEGF-induced Nurr1 expression. VEGF also stimulates the binding of nuclear CREB protein to its site in the Nurr1 promoter in vitro and in vivo. Moreover, using pharmacological inhibitors and molecular approaches, we show that VEGF-induced CREB activation is largely mediated by protein kinase C-dependent protein kinase D activation. Finally, our data indicate that knockdown of endogenous Nurr1 expression attenuates VEGF-induced endothelial cell proliferation, migration and in vivo matrigel angiogenesis, suggesting its potential importance in mediating VEGF-induced tumor angiogenesis.

Pathological angiogenesis is a hallmark of cancer and various ischemic and inflammatory diseases. Among many angiogenic factors, vascular endothelial growth factor (VEGF) which is the major VEGF isoform or VEGF-A165 is unique in its potency and selectivity for vascular endothelium, and it is the only angiogenic factor so far recognized that renders microvessels hyperpermeable to circulating macromolecules.1–3 Extensive studies have been carried out to investigate how VEGF transmits its signal from its cell surface receptors to its target gene expression and the angiogenic response. VEGF binds to its two high affinity receptor tyrosine kinases, Flt-1 (VEGFR-1) and KDR (VEGFR-2, FLK-1 in mice).4, 5 KDR, not Flt-1, is responsible for VEGF-stimulated cell proliferation and migration in cultured endothelial cells and for microvascular permeability.6–8 However, Flt-1 functions to downregulate KDR-mediated cultured endothelial cell proliferation.6, 9 The third VEGFR-3 (or Flt-4) that binds selectively to VEGF-C and VEGF-D is primarily expressed on lymphatic endothelial cells to mediate the lymphoangiogenesis induced by VEGF-C and VEGF-D.10 The signal transduction pathways mediated by KDR involve KDR phosphorylation, phospholipase C activation, inositol 1,4,5-trisphosphate accumulation, intracellular Ca2+ mobilization, protein kinase C and MAPK activation.6–8 In addition, VEGF also activates the novel protein kinase D (PKD) in endothelial cells.

PKD (also known as PKCμ or PKD1) belongs to a family of unique serine/threonine protein kinase (PKD1, 2, 3) with structure, enzymology and regulatory properties different from the PKC family members. Its most unique feature includes the presence of a Ca2+-independent catalytic domain, a regulatory pleckstrin homology region and a highly hydrophobic stretch of amino acids in its N-terminal region.10, 11 PKD can be activated in intact cells in response to numerous extracellular stimuli such as growth factors and ligands for G-protein–coupled receptors. In all these cases, rapid PKD activation is mediated by PKC-dependent phosphorylation of Ser-744 and Ser-748 within the activation loop of the catalytic domain of PKD. PKD activation is associated with its translocation to the plasma membrane and subsequent transient accumulation in the nucleus. PKD overexpression potentiates DNA synthesis induced by the G protein-coupled receptor (GPCR) agonists such as bombesin and vasopressin by increasing the duration of MAP kinase activation.11 Interestingly, involvement of PKD in VEGF signaling was demonstrated recently by us and others showing that inhibition of PKD activation attenuates VEGF-induced endothelial cell growth.12–16 A number of PKD substrates have been identified including histone deacetylase (HDAC), troponin I and others proteins in vesicle transport (for review, see Ref. 17). Transcriptional factor cAMP-responsive element binding protein (CREB) was also shown to be a potential PKD substrate in response to thrombin and reactive oxygen species.18–20 Although CREB was shown to be activated by VEGF,21, 22 the signaling pathways leading to CREB activation is not clear.

In the effort to identify unique early response genes induced by VEGF in endothelial cells, a novel orphan nuclear receptor Nurr1 mRNA was found to be highly and rapidly induced by VEGF stimulation. In our study, we continued to investigate the signaling mechanism mediating VEGF-induced Nurr1 expression and its potential role on VEGF-mediated angiogenic response. Our data show that VEGF significantly stimulates Nurr1 protein expression and promoter activity in addition to its mRNA induction. We analyzed the role of the two putative transcriptional sites NF-κB and CREB within the human Nurr1 promoter in the Nurr1 gene transcription and found that VEGF-induced Nurr1 promoter activity is mainly mediated through the proximal CREB site as indicated by several approaches. Our in vitro and in vivo binding assays further show that VEGF increases binding of CREB to the putative CREB element with the Nurr1 promoter. Using pharmacological and molecular approaches, we show that protein kinase C-dependent PKD is the major kinase that mediates VEGF-induced CREB phosphorylation and Nurr1 gene expression. Finally, we demonstrate that knockdown of Nurr1 induction significantly attenuates VEGF-induced endothelial cell proliferation, migration and in vivo angiogenic response.

Abbreviations

VEGF: vascular endothelial cell growth factor; CREB: cAMP-responsive element binding protein; PKD: protein kinase D; NF-κB: nuclear factor-κB; PKC: Protein kinase C; HUVEC: human umbilical endothelial cells

Material and Methods

Materials

Recombinant human VEGF-A165 (or VEGF) was obtained from R&D Systems (Minneapolis, MN). The EGM-MV Bullet kit, trypsin-ethylenediamine-tetraacetic acid (EDTA) and trypsin neutralization solution were obtained from Clonetics (San Diego, CA). Vitrogen 100 was purchased from Collagen Biomaterials (Palo Alto, CA). Rabbit polyclonal antibody against Nurr1 was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies against CREB, phospho-Ser133-CREB and PKD1 were from Cell Signaling. PKD2 antibody was from Upstate Biotechnology (Lake Placid, NY). The kinase-dead mutant of CREB (K-CREB) was purchased from BD Biosciences Clontech (Palo Alto, CA). All pharmacological inhibitors were from EMD Biosciences. Nontargeted control siRNA, human PKD1 siRNA (S1 and S2) and PKD2 siRNA (S1 and S2) from Qiagen (Valencia, CA) and human PKD2 siRNA (PKD2-S3) from Dharmacon (Chicago, IL) were transfected into human umbilical vein endothelial cells (HUVEC) as described previously.16 Nurr1 shRNA and green fluorescent protein (GFP) shRNA-expressing retroviral vectors were from Origene Technologies (Rockville, MD).

Cell culture

Primary HUVEC (obtained from Clonetics) were cultured with or without transduction with retroviruses as described.9 Only cells from passage 3 to 5 that were ∼80% confluent were used for experiments.

Quantitative real-time reverse transcription-polymerase chain reaction (RT-PCR)

Applied Biosystems software was used to design optimal primer pairs for real-time RT-PCR and data calculation. The forward and reverse primers for human Nurr1 were 5′-CTTGTGTTCAGGCGCAGTATG-3′ and 5′-GAGTGGTAACTGTAGCTCTGAGAAGC-3′, respectively. 5′-/5TET/CCTCGCCTCAAGGAGCCAGCC/36-TAMNph/-3′ served as an internal probe for Nurr1 glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as an internal control. The relative Nurr1 mRNA expression level that was normalized to equal GAPDH mRNA levels was calculated according to the methods described by Livak and Schmittgen.23 Experiments were repeated three times.

Immunoblotting

HUVEC were serum-starved for 24 hr and stimulated with 20 ng/ml VEGF for various times. Equal amount of cell protein was subjected to western blot analysis using different antibodies as indicated.

Nurr1 promoter activity

The 5′-flanking sequence from −1,025 bp upstream of the first base pair of reported cDNA to its downstream +71 bp of human Nurr1 gene (GenBank access number AB017586)24 was cloned by PCR using 5′-AG TTGGTGGGCA CAGAGGAGTATC-3′ (forward primer, −1,025 bp upstream of the first base pair of the reported cDNA) and 5′-TCACGGAGGGAGGG AGCAG-3′ (reverse primer, +71 bp downstream) and subcloned to pGL3 luciferase reporter vector (Promega) and the sequence was confirmed. Series of the Nurr1 promoter deletion were made by PCR using the indicated primer sets. The Nurr1 promoter containing the mutations in the CREB site (from 5′-TCGTGACGTCAGGT-3′ to 5′-TCGTGTGCCCAGGT-3′) was also created by the standard PCR-based technique. To determine the Nurr1 promoter activity in response to VEGF, HUVEC were seeded in 12-well plates (1 × 105 cells per well) overnight and transiently transfected using Fugene-6 transfection reagent (Roche) with Nurr1 promoter luciferase constructs together with a control luciferase construct pRL-SV (Promega) or other DNA constructs as indicated. Transfected cells were serum starved for 24 hr followed by exposure to VEGF. Firefly and Renilla luciferase activities in cell extracts were measured using Dual-Luciferase Reporter Assay System (Promega). The relative luciferase activity was then calculated by normalizing Nurr1 promoter-driven Firefly luciferase activity to control Renilla luciferase activity. Data from all experiments are presented as the relative luciferase activity (mean ± SE) from at least two independent sets of experiments, each with triplicate measurements.

Electrophoretic mobility shift assays

Nuclear extracts were prepared for DNA binding assays as described previously.25 Cells were washed in phosphate buffer saline (PBS), collected into Tris-NaCl-EDTA (TNE) buffer [40 mM Tris (pH 7.4), 1 mM EDTA and 0.15 M NaCl] and centrifuged (5,000g × 10 sec). The cell pellets were incubated with buffer A [10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM ethylene glycol tetraacetic acid (EGTA) and 0.5 mM phenylmethylsulfonyl fluoride (PMSF)] for 10 min before addition of 10% NP40 for an additional 2 min. Nuclei were centrifuged (5,000g × for 10 sec), incubated with buffer B [20 mM HEPES (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA and 0.1 mM PMSF] for 45 min and centrifuged at 13,000g for 10 min. Nuclear extracts were incubated with poly(dI-dC), band shift buffer [50 mM MgCl2, 340 mM KCl and 8 μl of delta buffer (0.1 mM EDTA, 40 mM KCl, 25 mM N-2-hydroxyethylpiperazeine-N′-2-ethananesulfonic acid (HEPES; pH 7.6), 8% Ficoll 400, 1 mM dithiothreitol] at 4°C for 15 min. 32P-labeled doubled-stranded oligonucleotide probe (100,000 cpm) was added to the reaction mixture and incubated for 30 min on ice. Binding of specific nuclear protein to the probe was determined by fractionating the nuclear proteins through a nondenaturing 6% polyacrylamide gel at 200 V for 2 hr at room temperature in TBE buffer [80 mM Tris-borate and 2 mM EDTA (pH 8.0)]. The gel was dried at 80°C for 2 hr under vacuum before exposure to X-ray autoradiography film. The CREB probe corresponding to the CREB site in the human Nurr1 promoter was end labeled by T4 DNA kinase (New England Biolabs, Beverly, MA) and [γ-32P] ATP (DuPont NEN, Boston, MA).

Chromatin immunoprecipitation assays

Chromatin immunoprecipitation (ChIP) assays were carried out according to the manufacturer's protocol of the ChIP assay kit (Millipore-Upstate, Charlottesville, VA). Briefly, HUVEC pellet were sonicated and crosslinked with formaldehyde and then subjected to immunoprecipitation (IP) with an antibody against phospho-Ser133-CREB or CREB or IgG as a control overnight at 4°C. The chromosomal DNA in the immunoprecipitated complexes was extracted with phenol–chloroform. After final ethanol precipitation, IP products were resuspended in TE buffer (10 mm Tris-HCl and 1 mm EDTA) and subjected to PCR assay with the primer set (forward: 5′-AGTGAAGTGTCGCGACGCTGCG-3′ and reverse: 5′-CTGCCGCTGCCAACATGCACC-3′). The primers were designed according to the sequences of the putative CREB binding site on the proximal portion of the human Nurr1 promoter. The supernatants before IP were served as total input chromatin control.

Knockdown of human Nurr1 protein expression

Retroviral vectors expressing Nurr1 shRNA and control GFP shRNA were from Origene. The preparation of the retroviruses expressing the Nurr1 shRNA and a control GFP shRNA and HUVEC infection were done as previously described.21 The effect of Nurr1 shRNA on the endogenous Nurr1 was determined by immunoblotting.

Proliferation assays

Assays were carried out as described.26 Briefly, HUVEC were serum starved (0.2% serum) for 24 hr and then stimulated with 20 ng/ml VEGF for 20 hr. [3H] thymidine (1 μCi/ml) was added to each well and 4 hr later, the cells were washed, fixed and lysed. The data were expressed as fold activation with the stimulated cells compared to control-treated cells. The data shown represent the means with SEM of triplicate determinations per experimental condition, and the experiments were repeated three times.

Migration assay

Assays were carried out as described.27 Briefly, serum-starved HUVECs were detached from tissue culture plates, washed twice with endothelial basal medium containing 0.1% fetal bovine serum (FBS), seeded (1 × 105 cells per well) into the transwells coated with vitrogen (30 μg/ml) and the transwells were inserted into a 24-well plate containing 1 ml of the same medium. Cells over a range from 3 × 103 to 1 × 105 cells per well were seeded in a 96-well plate for the standard curve. Cells were incubated at 37°C for 1 hr to allow the cells to attach, then VEGF was added at a final concentration of 20 ng/ml. After incubation for an additional 2 hr, cells remaining on the upper surface of the transwell filter membrane were wiped off with a cotton tip. The whole transwell membrane was cut out and placed in an individual well of the 96-well plate that contained the cells for the standard curve. Cyquant DNA stain (200 μl) was added to each well containing cells or membrane, and the plate was kept at 4°C overnight. After warming to room temperature, stained cells were counted in a spectrofluorometer (SpectraFluor; TECAN) with DeltaSoft 3 software. The data are expressed as the mean ± SEM of quadruplicate values. All experiments were repeated three times.

Angiogenesis assay

Matrigel angiogenesis assays were carried out as described.13, 26 Briefly, human melanoma cell line stably transfected with VEGF-A165 (SKMEL/VEGF cells) (1 × 107), alone or mixed with 1 × 107 RetroPack PT67 cells (clontech) (PT67) which produced retroviruses expressing Nurr1 shRNA or control GFP shRNA were suspended in 0.5 ml of growth factor reduced Matrigel (BD Biosciences, Bedford, MA) and injected s.c. into Nu/Nu mice. Tissues were harvested, photographed and freshly embedded in optimal cutting temperature (OCT) media. Frozen sections were made for CD31 immunohistochemical staining. Vessel density was manually counted and calculated as number of vessels per mm2. Each experiment was replicated on eight mice.

Statistical analysis

Results were expressed as means ± SE. Data were analyzed using the Microsoft Excel statistics software program and ANOVA were used for intergroup comparisons.

Results

VEGF stimulates Nurr1 expression in cultured endothelial cells

The previous study from Liu et al. shows that VEGF markedly upregulates Nurr1 mRNA levels in HUVEC cells (130 fold relative to control-treated cells), but the Nurr1 protein expression was not shown to be induced by VEGF most likely because the antibody used did not work.28 To resolve this discrepancy, we also treated serum-starved HUVEC for various times, total RNA was isolated and subjected to quantitative real-time PCR as described in the Material and Methods Section. Consistent with the finding of Liu et al., our data show that VEGF markedly upregulates Nurr1 mRNA in a time-dependent manner with maximal induction 1 hr after stimulation (60 fold vs. control-treated cells; Fig. 1a). To determine whether dramatic upregulation of Nurr1 mRNA leads to significant induction of Nurr1 protein, cells were treated as described above and equal amount of total cell protein was subjected to Western blotting using a polyclonal antibody against Nurr1. Our results show that compared to the serum-starved control cells that had very low levels of Nurr1 protein, VEGF also markedly increases Nurr1 protein levels (Fig. 1b). To further examine whether VEGF-stimulated upregulation of Nurr1 mRNA and protein is mediated at a transcriptional level, the upstream 1,026 bp promoter sequence of human Nurr1 gene was isolated and subcloned into a luciferase reporter vector. To determine the promoter activity of Nurr1, cells were transiently transfected with the promoter construct along with an internal control plasmid. Cells were then starved and treated with VEGF for various time points. Equal amounts of cell lysates were used to measure luciferase activity. The data show that VEGF also significantly stimulates Nurr1 promoter activity begining 2 hr after treatment and reaching the maximal level 4 hr later (Fig. 1c). The kinetics of the Nurr1 promoter activity correlates with the Nurr1 protein induction and also with its mRNA expression as there is a delay from mRNA translation into protein. Lower levels of Nurr1 promoter activation by VEGF as compared to its mRNA induction might be due to a number of factors. First, the cloned 1,025 bp upstream promoter region of the Nurr1 gene may contain only a part of the regulatory elements in the Nurr1 gene. The distal region of the upstream promoter or the sequence in first intron may also participate in VEGF-induced Nurr1 transcription. Second, VEGF may also regulate Nurr1 mRNA stability. All of these regulatory mechanisms for VEGF-induced Nurr1 expression remain to be further examined in future.

Figure 1.

VEGF induces Nurr1 expression at mRNA, protein and transcriptional levels. (a) Quiescent HUVEC were treated with VEGF (20 ng/ml) for various times, and total RNA were isolated and subjected to qPCR analysis to measure Nurr1 mRNA levels. The relative Nurr1 mRNA levels were calculated by normalizing to equal levels of control GAPDH mRNA as described in the Material and Methods Section. (b) Cells were treated with VEGF (20 ng/ml) for the indicated times and equal amount of protein was subjected to Western blotting using a Nurr1 antibody or actin antibody. (c) HUVEC were transiently transfected with Nurr1 promoter–luciferase promoter construct along with an internal control plasmid. Cells were then serum starved and treated with VEGF (20 ng/ml) for the indicated times. Luciferase activity was measured and normalized to internal control. The data are expressed as relative promoter activity (mean ± SEM). Triplicate samples were used for each time points. *Denotes p < 0.01 versus control. The results are representative of three independent experiments.

VEGF-induced Nurr1 transcription requires CREB activation

The analysis of the promoter region of human Nurr1 gene shows several putative cis-acting binding elements including the upstream NF-κB site and the downstream CREB site.24 To determine the role of these binding sites in VEGF-induced Nurr1 transcription, we also made two Nurr1 promoter deletion constructs as indicated in Figure 2a. The results show that deletion of the distal NF-κB site moderately but significantly decreased VEGF-induced Nurr1 promoter activity. In contrast, deletion of the CREB site markedly inhibits VEGF-induced Nurr1 transcription (Fig. 2a) The effect of NF-κB in VEGF-mediated Nurr1 transcription is also confirmed by cotransfecting the inhibitory protein IκBα with the Nurr1 promoter construct, showing that inhibition of NF-κB activation attenuates Nurr1 promoter activity (data not shown). Then we determined the role of CREB in the VEGF-induced Nurr1 expression. To do this, cells were transiently transfected with a control plasmid or a dominant negative CREB mutant along with the Nurr1 reporter construct and an internal control plasmid. Serum-starved cells were then treated with VEGF. The results indicate that transfection of the dominant negative CREB significantly inhibits VEGF-induced Nurr1 promoter activity (Fig. 2b), suggesting that activation of CREB by VEGF mediates Nurr1 gene transcription. To further prove the role of the CREB site of the human Nurr1 promoter in Nurr1 gene transcription, we created a specific CREB site mutation construct as shown in Figure 2c (top panel). Compared to the wild-type Nurr1 promoter construct, mutation of this CREB site significantly reduces VEGF-induced Nurr1 promoter activity in HUVEC (Fig. 2c). The results indicate that CREB activation is required for VEGF-mediated Nurr1 gene transcription.

Figure 2.

Role of CREB in VEGF-induced Nurr1 transcription. (a) Cells were transfected with the three Nurr1 promoter constructs as indicated along with an internal control and stimulated with VEGF (20 ng/ml) for 4 hr. * and ** denote p < 0.05 and p < 0.001 versus −1,026 bp construct-transfected VEGF-treated cells, respectively. (b) Cells were transiently transfected with plasmids expressing control LacZ or dominant negative CREB mutant (K-CREB) together with the Nurr1 promoter construct and stimulated with VEGF (20 ng/ml) for 4 hr. * and ** denote p < 0.05 and p < 0.001 versus −1,026 bp construct-transfected VEGF-treated cells, respectively. *Denotes p < 0.05 versus VEGF-treated LacZ control. (c) Cells were transfected with the wild-type Nurr1 promoter construct or the Nurr1 promoter construct with the indicated CREB site mutations followed by VEGF stimulation. *Denotes p < 0.01 versus VEGF-treated LacZ control. The Nurr1 promoter activity was measured as described in the Material and Methods Section. The results are representative of three independent experiments.

VEGF stimulates CREB binding to the Nurr1 promoter

As VEGF was shown to stimulate CREB phosphorylation, we sought to determine whether VEGF induces binding of CREB to the putative CREB site of the Nurr1 promoter. To do this, we first performed the DNA gel electrophoretic mobility shift assay (EMSA). The oligonucleotide containing the CREB site in the Nurr1 promoter was 32P-labeled and incubated with nuclear extracts isolated from HUVEC treated with VEGF for different times. The data show that VEGF stimulates time-dependent binding of nuclear protein to the CREB oligonucleotide probe (Fig. 3a). The specificity of the binding is confirmed by inclusion of either excess amount of nonlabeled probe or CREB antibody in the reaction mixtures. The data indicate that excess of cold probe blocked the binding of the nuclear protein to the radio-labeled probe, and also the binding band was supershifted by the CREB antibody but not the control antibody (Fig. 3a). To examine whether CREB also binds to the CREB site of the Nurr1 promoter in vivo, in vivo ChIP assay was performed. HUVEC were treated with VEGF for 30 min, crosslinked with formaldehyde and then subjected to IP with antibodies against phospho-Ser133-CREB or CREB or IgG as a control. The chromosomal DNA in the immunoprecipitated complexes was extracted and subjected to PCR assay. The two primers were designed according to the sequences flanking the putative CREB binding site on the Nurr1 promoter. As shown in Figure 3b, a PCR-amplified band was detected in the immunoprecipitated complex with an antibody against CREB but not control IgG after VEGF stimulation. Total cell lysates before IP were used as equal chromatin control (Fig. 3b, lower panel). The data together show VEGF significantly increases the binding of activated phospho-Ser133-CREB to the CREB site of Nurr1 promoter.

Figure 3.

VEGF stimulates nuclear CREB binding to the putative CREB site in the Nurr1 promoter. (a) HUVEC were stimulated with VEGF (20 ng/ml) for the indicated times. Nuclear protein was isolated and subjected to EMSA using a 32P-labeled oligonucleotide probe corresponding to the CREB site of Nurr1 promoter. Excess amount of unlabeled probe or CREB antibody as well as control IgG was also included in the reaction mixtures to verify the specificity of the shifted band indicated by arrow. F means free probe. (b) HUVEC were treated with VEGF (20 ng/ml) for 30 min. Cell lysate was subjected to ChIP assay using CREB or phospho-S133-CREB antibody or their control IgG as described in the Material and Methods Section. The results are representative of three independent experiments.

VEGF-induced CREB activation requires PKC-dependent PKD pathway

The previous studies using the pharmacological inhibitors of protein kinase C21, 29 suggested that VEGF-induced CREB phosphorylation in HUVEC was partially mediated by calcium-independent PKC as intracellular calcium chelator BAPTA/AM had little effect on this CREB phosphorylation. However, the exact mechanism responsible for VEGF-induced CREB activation was still not clear. Here, we first show by using various pharmacological inhibitors that VEGF-induced CREB phosphorylation (Fig. 4a) is reduced by 20 μM protein kinase A inhibitor H-89 but not affected by the PI-3K inhibitor LY-294002 or the MAP kinase kinase (MEK) inhibitor PD98059 or p38MAP kinase inhibitor SB203580 (data not shown). Then we examined the effect of PKC inhibitors on this VEGF response. Pretreatment with the general PKC inhibitor GF109203X (0.2 μM) but not the PKCδ inhibitor rottlerin (10 μM) largely inhibits VEGF-induced CREB phosphorylation (Fig. 4b). Using another PKC inhibitor Go6976 that inhibits conventional PKC PKCα, PKCβ as well as PKCμ/PKD, we show that Go6976 almost completely inhibits VEGF-induced CREB phosphorylation even at the very low concentration (0.2 μM). On the basis of the previous observations from us and others that VEGF-induced PKD activation in HUVEC is mediated by a protein kinase C pathway, we hypothesized that PKC-dependent PKD activity is involved in VEGF-induced CREB activation. To confirm the effect, we transduced HUVEC with control retroviruses or viruses expressing PKD siRNA as previously described13 and show that knockdown of PKD expression significantly reduces VEGF-induced CREB phosphorylation (Fig. 4c, top panel). As previously described, HUVEC express both PKD/PKD1 and PKD2 isoforms,16 we examined whether PKD2 is also involved, transfection of three different PKD2 siRNA (PKD2-S1, PKD2-S2 or PKD2-S3) significantly inhibits VEGF-induced CREB phosphorylation (Fig. 4c, bottom panel). The results indicate that PKC-dependent PKD1/2 activation is required for VEGF-induced CREB pathway in endothelial cells.

Figure 4.

PKD1/2 mediates VEGF-induced CREB activation. (a) Time course of CREB activation by VEGF (20 ng/ml). (b) Cells were pretreated with the PKC inhibitors GF109203X, the PKCδ inhibitor rottlerin or PKC/PKD inhibitor Go6976 at the indicated concentrations before VEGF stimulation. (c) Cells were infected with control or PKD1 siRNA-expressing viruses (upper panel) or transfected with small PKD1 siRNA (PKD1-S1 and S2) (middle panel) or small PKD2 siRNA (PKD2-S1, S2 or S3) (bottom panel) before VEGF treatment. Equal amount of protein was subjected to Western blotting using PKD1, PKD2, actin, phospho-CREB antibody or CREB antibody. The results are representative of three independent experiments.

VEGF-induced Nurr1 expression involves the KDR-mediated PKC-dependent PKD pathway

To examine whether VEGF-induced PKC-PKD-CREB pathway affects Nurr1 expression, HUVEC were transiently transfected with the Nurr1 promoter as above and then treated with the KDR inhibitor Su5416 or GF109203X or Go6976. The data show that VEGF-induced Nurr1 gene transcription is significantly inhibited by Su5416 or GF109203X or Go6976 (Fig. 5a). To further examine the role of PKD in Nurr1 transcription induced by VEGF, cells were transfected with the Nurr1 promoter together with control plasmid LacZ, or dominant negative PKD constructs PKD-SS/AA or PKD-KD. The results show that block of PKD activation by either PKD-SS/AA or PKD-KD significantly decreases VEGF-induced Nurr1 transcription (Fig. 5b).

Figure 5.

KDR, PKC and PKD activation mediated VEGF-induced Nurr1 transcription. (a) Cells were transiently transfected with the Nurr1 promoter as above and then pretreated with the specific KDR inhibitor Su5416 (10 μM) or general PKC inhibitor GF109203X (1 μM) or PKC/PKD inhibitor Go6976 (1 μM) for 30 min before stimulation with VEGF (20 ng/ml) for 4 hr. (b) Cells were transfected with the Nurr1 promoter along with the dominant negative PKD mutant PKD-S738A/S742A or kinase-dead PKD mutant before VEGF stimulation for 4 hr. The relative Nurr1 promoter activity was determined. *, ** and *** denote p < 0.01 versus VEGF-treated controls, respectively. The results are representative of three independent experiments.

Nurr1 regulates VEGF-mediated HUVEC proliferation, migration and angiogenesis in vivo

To investigate the function of Nurr1 induced by VEGF in HUVEC, we used RNA interference to knockdown the endogenous Nurr1 protein expression. HUVEC were infected with control GFP shRNA- or Nurr1 shRNA-expressing retroviruses overnight and cultured for additional 2 days. Nurr1 protein levels were analyzed by immunoblotting. The data show that Nurr1 shRNA significantly reduces endogenous Nurr1 protein expression (Fig. 6a). Then we studied whether knockdown of Nurr1 protein expression has effect on VEGF-induced HUVEC proliferation as indicated [3H]-thymidine incorporation into HUVEC. Cells were infected with Nurr1 shRNA-expressing retroviruses or control viruses. The infected cells were subjected to [3H]-thymidine incorporation assay as described in the Material and Methods Section. The results show that knockdown of Nurr1 protein expression significantly inhibits VEGF-stimulated HUVEC proliferation (Fig. 6b). Next we determined whether Nurr1 is also important for VEGF-induced HUVEC migration. Cells were infected and subjected to migration assay as in the Material and Methods Section. Our results indicate that Nurr1 is also involved in VEGF-induced HUVEC migration (Fig. 6c).

Figure 6.

Involvement of Nurr1 in VEGF-induced HUVEC proliferation, migration and matrigel angiogenesis. (a) Cells were infected with viruses expressing control shRNA or Nurr1 shRNA. Three days later, cell lysates were subjected to Western blotting using antibodies for Nurr1 or actin for equal loading control. Efficiency of knockdown is shown on the right panel. (b) Cells infected with viruses expressing control shRNA or Nurr1 shRNA were subjected to [3H]-thymidine incorporation assay (b) or migration assay (c) as described in the Material and Methods Section. * and ** denote p < 0.01 versus VEGF-treated controls, respectively. (d) Matrigel mixtures containing VEGF-secreting SKMEL cells and virus-packaging cells making control shRNA or Nurr1 shRNA were subcutaneously implanted. Three days later, the matrigel plugs were removed, photographed and further fixed for CD31 staining. The numbers of CD31-positive vessels were counted. *Denotes p < 0.001 versus lane 1 and ** denotes p < 0.05 versus lane 2. To confirm the efficient Nurr1 knockdown, protein extracts were also made from the matrigel plugs and subjected to Western blotting using Nurr1 antibody and control actin antibody. The data show that Nurr1 shRNA efficiently knocked down VEGF-induced Nurr1 protein expression in the matrigel plugs (bottom panel). The results are representative of three independent experiments.

Next we determined whether Nurr1 also plays a role in VEGF-mediated in vivo angiogenesis. To do this, we utilized a recently modified matrigel angiogenesis assay13, 26 as described in the Material and Methods Section. As retroviruses infect preferentially dividing cells, this assay allowed us to introduce DNA into VEGF-activated endothelial cells in vivo. Briefly, SK-MEL-2 tumor cells transfected to overexpress VEGF (SK-MEL/VEGF cells) were mixed with PT67 retrovirus packaging cells that secreted viruses expressing Nurr1 shRNA or its GFP shRNA control in matrigel solutions that were then implanted in the subcutaneous space of nude mice. As previously reported,26 VEGF secreted by SK-MEL/VEGF cells induces nearby vascular endothelial cells to divide and therefore to become susceptible to infection with retroviruses secreted by PT67 packaging cells. The angiogenic response that developed after implantation of matrigel plugs containing various cell mixtures was evaluated on day 3 by macroscopy (Fig. 6d, top panels) and by histology and immunohistochemistry for the endothelial cell marker CD31 (Fig. 6d, middle panels). Matrigel plugs containing only PT67 cells packaging control retroviruses (PT67/control cells) induces minimal angiogenesis (Fig. 6d, lane 1). However, strong angiogenesis with typical “mother” vessels is induced in plugs containing SK-MEL/VEGF cells (Fig. 6d, lane 2). Mother vessels are enlarged, thin-walled, pericyte-poor vessels that are the first new vessel type induced by VEGF in vivo.30 The angiogenic response and mother vessel formation induced by SK-MEL/VEGF cells are significantly depressed by inclusion of PT67/Nurr1 shRNA cells (Fig. 6d, lane 3). The results indicate that Nurr1 expression is important for the VEGF-mediated angiogenic response.

Discussion

The importance of VEGF in vascular development and various pathological angiogenesis has been extensively investigated. The signaling transduction pathways mediating its angiogenic responses have also been extensively investigated, however, the transcriptional mechanisms remained unclear. The profiling of early responsive genes induced by VEGF has resulted in identification of several novel transcription factors, one of which is orphan nuclear receptor family protein Nurr1. Nurr1 is one member of Nur77 subfamily of nuclear proteins (Nur77, Nurr1 and Nor-1), which have completely different phenotypic functions during mouse development. Nurr1 was shown to play a vital role in development of dopaminergic neuron as the mice with its homozygous deletion lack dopaminergic neurons and die within 2 days after birth.31 Homozygous Nor-1-knockout mice die around embryonic day 8.5 of gestation.32 In contrast, Nur77-null mice are developmentally normal.33 In addition to their distinct developmental roles, recent works also pointed out that they have nonredundant effects in several pathological conditions including VEGF-mediated angiogenesis26, 29 and insulin resistance.34 In case of the VEGF-mediated angiogenesis, Nur77 and Nor-1 appear to play nonredundant or additive/synergistic role in VEGF-induced endothelial cell proliferation.26, 29 Our present study further shows that Nurr1 is also important in VEGF-induced angiogenic response as knockdown of endogenous Nurr1 expression attenuated VEGF-stimulated endothelial cell proliferation and migration as well as in vivo angiogenic response. Moreover, we further investigated the signaling mechanism mediating VEGF-induced Nurr1 expression, and show for the first time that Nurr1 induction by VEGF is primarily mediated through PKC-dependent PKD activation of the transcriptional factor CREB.

Involvement of CREB in transcription of Nurr1 has been recently suggested in Hela cells treated with PMA, EGF and TNFs35 or in synoviocytes treated with proinflammatory cytokines IL-1β or TNF or PGE2.36 Our present results demonstrate that CREB is also required for VEGF-induced Nurr1 expression. This effect is mediated at least partially at the transcriptional level because (i) VEGF significantly increased Nurr1 promoter activity and (ii) a dominant negative CREB mutant (K-CREB) or mutation of the putative CREB binding site in the promoter largely inhibited VEGF-induced Nurr1 promoter activity. The relative lower levels of Nurr1 promoter induction compared to large induction of Nurr1 mRNA could be due to the two factors: (i) low transfection efficiency with large plasmid DNA in endothelial cells and (ii) post-transcriptional regulation of Nurr1 mRNA. Post-transcriptional regulation of VEGF-induced Nurr1 mRNA expression, in particular, control of Nurr1 mRNA stability by a potential microRNA mechanism remains to be investigated. In addition to the major CREB mediator, we also showed that deletion of the NF-κB site or inhibition of NF-κB activation by its inhibitory protein IκBα slightly attenuated VEGF-induced Nurr1 promoter activity (not shown), suggesting that NF-κB activation is involved in VEGF-induced Nurr1 expression, consistent with other reports showing the importance of NF-κB in lipopolysaccharide-induced Nurr1 transcription in macrophages.36, 37

It was suggested by using pharmacological inhibitors that the signaling mechanism mediating VEGF-induced CREB phosphorylation involves several proteins kinases including protein kinase C activation (the PKC blocker GF109203X), p38 MAP kinase (SB202190), MEK1/2 (PD98059) as well as PKA (H-89).21, 29 Our experiments show that VEGF-induced CREB phosphorylation is largely inhibited by the PKC inhibitor GF109203X even at 0.2 μM concentration whereas other inhibitors such as PD98059 (50 μM) and p38 inhibitor SB203580 (10 μM) had no effect (data not shown). Moreover, we further show that another PKC inhibitor Go6976 which also inhibits PKD family of protein kinases at the nanomolar concentration blocked VEGF-induced CREB phosphorylation. Our knockdown experiments with PKD1/2 siRNA confirm that both PKD1 and PKD2 that are coexpressed in HUVEC16 are required for VEGF-induced CREB phosphorylation. CREB is known to have a consensus PKD phosphorylation site at serine 133.18, 19 However, a definitive evidence for PKD-induced CREB activation by natural ligands was not demonstrated previously although inhibition of PKD by the PKC/PKD inhibitor Go6976 blocked GPCR ligand thrombin-mediated CREB phosphorylation and the constitutive active PKD mutant PKD-S738E/S742E induced CREB phosphorylation.19 Our results using the siRNA knockdown approach provide the first convincing evidence that PKD activation causes CREB phosphorylation at least in the case of VEGF stimulation. Partial inhibition of VEGF-induced CREB phosphorylation by individual PKD1 or 2 siRNA may be explained by less than 100% liposome-mediated transfection efficiency and additive effects of these two PKD isoforms. It is established that PKD activation is primarily mediated by PKC phosphorylation of the two critical residues serine 738 and serine 742 located at its activation loop in response to both GPCR ligands.17, 38, 39 We and others also showed recently that VEGF stimulates PKD phosphorylation through PKC-dependent mechanism in endothelial cells.12, 13 Together with our present results, we propose a new pathway in which VEGF stimulates PKC-dependent PKD-mediated CREB phosphorylation, leading to Nurr1 gene expression in endothelial cells which further promotes VEGF-mediated angiogenic response.

Ancillary