Nuclear Factor-κB and cAMP Response Element Binding Protein Mediate Opposite Transcriptional Effects on the Flk-1/KDR Gene Promoter
Abstract—The vascular endothelial growth factor receptor Flk-1/KDR is highly expressed during development and almost disappears in adult tissues. Despite its biological relevance, little is known about the molecular mechanisms controlling its expression. In the present work, it is shown that cAMP response element binding protein (CREB) and nuclear factor-κB (NF-κB)–related antigens bind specific sequences in the Flk-1/KDR promoter. Functional studies demonstrate that cAMP represses whereas tumor necrosis factor-α, an activator of NF-κB, stimulates promoter activity. Histone acetyltransferases (HATs) P/CAF and CBP/p300 together with p65/RelA, the catalytic subunit of NF-κB, increase Flk-1/KDR promoter activity 10- to 20-fold. Consistently, inhibition by cAMP is reverted by increasing intracellular HATs and is completely abolished by site-specific mutagenesis of the cAMP response element. In contrast, specific mutations in the NF-κB response element abolish responsiveness to p65/RelA and HATs without affecting cAMP-dependent repression. These results suggest that opposing signaling pathways, activating NF-κB or CREB and requiring HAT molecules, control Flk-1/KDR promoter activity. The full text of this article is available at http://www.circresaha.org.
Vascular endothelial growth factor (VEGF) receptor-2 (Flk-1/KDR) characterizes the hemangioblast, which is the common endothelial cell and hematopoietic cell early progenitor.1 2 In adults, the expression level of Flk-1/KDR is low,3 and enhanced KDR expression has been reported in acute myocardial infarction,4 as well as in brain5 and retinal ischemia.6 Both mouse (Flk-1) and human (KDR) genes7 8 reveal a class II promoter structure,9 characterized by the absence of a TATA box and by the presence of several DNA binding sites for general and tissue-specific transcription factors. SP1, SP3,10 and TFII-I11 modulate basal and tissue-specific Flk-1/KDR promoter activity, whereas no trans-acting factors, mediating its transcriptional response to angiogenic stimuli, have been identified to date. Tumor necrosis factor α (TNF-α) and cAMP seem to retain opposite angiogenic effects.12 13 These molecules exert their action via nuclear factor-κB (NF-κB) and cAMP response element binding protein (CREB) and may operate on opposing signaling pathways, via a competitive interaction with histone acetyltransferases (HATs),14 15 16 17 such as p30018 CBP,19 and P/CAF.20 The present study examines the role of putative NF-κB response element (κB-RE) and cAMP response element (CRE) as positive and negative regulatory elements of Flk-1/KDR promoter transcriptional activity.
Materials and Methods
Flk-1/KDR Promoter Cloning and Luciferase Constructs
Five fragments were amplified from a murine genomic clone by polymerase chain reaction (PCR) using specific oligonucleotides designed according to a published Flk-1/KDR promoter sequence.8 The fragments were cloned in PCR 2.1 vector (Invitrogen), then cloned in pGL2 basic vector (Promega), after SacI and XhoI digestion.
Cell Culture, Transfections, and Luciferase Assays
Bovine aortic endothelial cells (BAECs) were prepared as previously described.21 BAECs and murine fibroblast STO cells were cultured in complete DMEM with 10% FBS; NIH-3T3 cells were cultured in complete DMEM with 5% FBS, and human umbilical vein endothelial cells (HUVECs) were cultured in EGM-2 complete medium (Clonetics). Transient transfections were performed with LipofeCTAMINE-Plus Reagent (Life Technologies) according to the manufacturer’s instructions; 9×105 cells (passages 4 through 6) were seeded in 60-mm culture plates 24 hours before transfection. In each experiment, the total amount of DNA was always kept constant (20 μg), filling in with empty backbone vectors or pBluescript plasmid. Luciferase assays were performed using a luciferase assay system (Promega), according to the manufacturer’s instructions. Luciferase activity, expressed as single photon counts×s−1 (cps), was determined with an A9904V Top Count (Packard), and it was normalized for protein concentration and transfection efficiency, which was determined by β-galactosidase production. All transfection data are represented as the mean±SD of 3 or 4 independent experiments performed in duplicate.
Total RNA was isolated from 107 HUVECs treated for 24 hours with 20 ng/mL TNF-α, 1 mmol/L cAMP (Sigma) or control solvent, using TriZOL reagent (Life Technologies). RNA (25 μg) was electrophoresed on a 1% formaldehyde-agarose gel, capillary-blotted on Hybond-N membrane (Amersham Pharmacia Biotech), and hybridized with a 700-bp cDNA probe, which was specific for the human KDR gene.22
Stable Transfection and FACS Analysis
A −442/+297 promoter fragment was subcloned in pEGFP reporter plasmid (Clontech), generating a Flk-1GFP reporter vector. Linearized Flk-1/GFP and pEGFP plasmids (20 μg) were transfected in BAECs (107 cells, passages 3 through 5) by electroporation with a Gene Pulser II (Bio-Rad), at 200 V (capacitance 250 μF). Transfected cells were selected by adding 0.5 mg/mL of G418 (Gibco BRL). 48 hours after electroporation. Cells were analyzed for fluorescence using a FACScan (Becton Dickinson).
Nuclear Extract and Electrophoretic Mobility Shift Assays (EMSAs)
Nuclear extracts were obtained as previously described.23 Binding reactions were performed on ice; 5 μg of nuclear extract was incubated for 10 minutes in the presence or absence of cold competitor oligonucleotides or anti-CREB, anti p65 antibodies (Santa Cruz), or normal immunoglobulins used as a negative control, and then with a 32P-labeled oligoprobe (4×104 cpm/sample) as described.23 Electrophoresis was performed for 2 hours at 180 V, at 4°C. Gels were dried in a gel dryer apparatus (Bio-Rad) and exposed for 2 to 3 hours with Biomax film (Kodak).
Site-directed mutagenesis was performed using the QUICK CHANGE Site Directed Mutagenesis Kit (Stratagene), according to the manufacturer’s instructions. Mutagenic oligonucleotides were designed as follows; wild-type CRE and NF-κB consensus sequences are underlined and mutated nucleotides are shown in lowercase letters.
Mutagenized regions were sequenced to verify the accuracy of mutagenesis before use.
Results were analyzed by one-way ANOVA. Post hoc tests according to the Student-Newman-Keuls method were used to assess statistically significant differences among different groups. A value of P<0.05 was considered statistically significant.
Flk-1/KDR Promoter Isolation, Deletion Analysis, and Identification of cAMP and κB-REs
Three positive clones were isolated from a murine λ phage genomic DNA library. Specific oligonucleotides were designed to amplify by PCR portions of the Flk-1/KDR promoter to subclone in pGL2 basic vector. The promoter reporter vectors −1500/+297Flkluc, −1000/+297Flkluc, −442/+297Flkluc, and −442/+79Flkluc were tested for their transcriptional activity in BAECs and STO cells (Figure 1A⇓). The −1500/+297Flkluc construct did not exert significant transcriptional activity in agreement with previous studies that postulated the presence of upstream silencer elements in the region encompassing nucleotides −1500/−1000.7
The constructs −1000/+297Flkluc, −442/+297Flkluc, and −442+79Flkluc showed a significant level of transcription in BAECs; construct −442/+297Flkluc was the most active (Figure 1A⇑) and was used in all subsequent experiments. In STO cells, transcriptional activity of constructs −1000/+297Flkluc, −442/+297Flkluc, and −442/+79Flkluc were above control but considerably lower than in BAECs. Sequence analysis revealed the presence of a CRE, TGAGTCCT, at nucleotides +11/+18, and recent data7 revealed the presence of a putative κB-RE at position −62/−55 in the murine promoter (Figure 1B⇑). Comparison analysis showed that the CRE in the Flk-1/KDR promoter retained a high level of similarity to the CRE of other genes such as somatostatin24 and chromogranin A25 26 (Figure 2A⇓, top) and that the κB-RE was highly homologous (>80%) with functional NF-κB elements present in the promoter region of other genes (Figure 2A⇓, bottom). CRE and κB-RE probes and mutant oligonucleotides (Figure 2B⇓) were designed to perform EMSAs. By this approach, specific complexes bound to the CRE were detected in nuclear extracts from both BAECs (Figure 2C⇓, left) and STO cells (not shown). CREB-related proteins were bound to this DNA element in BAECs, as suggested by a supershift of the complex bound to CRE (Figure 2C⇓) but not in those from fibroblasts (not shown). Competition analysis (Figure 2C⇓, right) showed that oligonucleotides M1 through M6 could efficiently compete the nuclear protein binding to the wild-type sequence. In contrast, the M7 oligoprobe did not interfere with DNA/nuclear protein interactions (Figure 2C⇓), indicating the conserved G residue at position +12 as important for nuclear protein binding to this sequence. In other promoters, mutation in the CRE region encompassing the same G residue also reduced competition capacity or cAMP responsiveness.27 28
In Figure 2D⇑, it is shown that nuclear complexes containing NF-κB immunologically related proteins were bound to an oligoprobe designed across the −62/−55 region. An anti–p65 antibody (lane 5, left), raised against the DNA binding and dimerization domain of p65,29 reduced protein binding, whereas no band depletion was observed with a control IgG (lane 6); a cold κB-RE oligonucleotide (see Figure 2B⇑ for sequence) efficiently competed nuclear complex binding (Figure 2D⇑, left, lanes 2 through 4; right, lanes 2 and 3), whereas an oligonucleotide mutated in the κB-RE did not interfere with nuclear protein binding (mκB oligonucleotide, Figure 2D⇑, right, lanes 4 through 7). The M7 and the mκB mutant regions (Figure 2B⇑) were used to generate the mutated promoter constructs mCREluc and mκBluc used in transient transfections (see below).
cAMP and TNF-α Mediate Opposite Effects on Flk-1/KDR Gene Expression and Promoter Activity
The effects of TNF-α and cAMP on Flk-1/KDR mRNA and promoter activity were evaluated in HUVECs and BAECs, respectively. Dibutyryl cAMP repressed Flk-1/KDR promoter activity ≈2-fold, whereas TNF-α enhanced it ≈3-fold. Northern blot analysis performed in TNF-α–treated HUVECs showed an increase in Flk-1/KDR mRNA (≈3-fold), as previously demonstrated,12 whereas cAMP reduced Flk-1/KDR expression in parallel with the inhibition of the promoter activity (Figure 3A⇓). Consistently, FACS analysis (Figure 3B⇓) performed on BAECs stably transfected with a Flk-1GFP reporter vector showed that TNF-α increased (panel 2) whereas cAMP decreased (panel 3) the number of green fluorescent protein (GFP)–positive cells. TNF-α (panel 4) and the histone deacetylase inhibitor trichostatin A (TSA) (panel 5) treatments counteracted cAMP inhibition whereas TSA enhanced TNF-α effects (panel 6). Expression was tissue-specific, because no fluorescence was detected in NIH-3T3 cells transfected with the same reporter vector (not shown). These results indicate that signals activated by TNF-α and cAMP mediate opposite effects on Flk-1/KDR gene expression and promoter activity and suggest a role for histone acetylation in this process. The molecular mechanism of this process was examined in the following experiments.
NF-κB Requires HATs to Stimulate Transcription From the Flk-1/KDR Promoter
NF-κB is one of the main effectors of TNF-α intracellular functions, and NF-κB cooperation with HAT molecules could be important for the optimal induction of IL-6 promoter activity.16 Therefore, it was examined whether the NF-κB active subunit p65/RelA modulated Flk-1/KDR promoter activity and, eventually, whether this effect required the contribution of HATs. The direct transfection of the NF-κB active subunit p65/RelA did not stimulate promoter activity above basal level (Figure 4A⇓). In contrast, p65/RelA overexpression, in the presence of cotransfected P/CAF, p300, or CBP, produced a 6- to 10-fold increase in promoter activity, which was increased to 20-fold when p65/RelA, P/CAF, and p300 were cotransfected (Figure 4A⇓). Transfection of backbone vectors alone had no effect (not shown), and the protein level of transfected p65/RelA was not modulated by the presence of HATs, as determined by Western blot analysis (not shown). In addition, it was found that a functional HAT domain was required, because p300 mutants Δ1472 to 1522 and Δ1603 to 165330 failed to cooperate with p65/RelA (Figure 4B⇓), and cotransfection of p300 mutants with the wild-type plasmid rescued p65 promoter induction whereas transfection of backbone vectors had no effect (not shown). Notably, HAT molecules directly stimulated transcription from the −442/+297Flkluc reporter, a phenomenon that may be ascribed to their interaction with other trans-acting factors as well as members of the basal transcription machinery.14
cAMP and CREB Are Negative Regulators of Flk-1/KDR Promoter Activity
In BAECs, treatment with cAMP produced a 2-fold repression of Flk-1/KDR transcription (Figures 3A⇑ and 4C⇑); this was further increased to 5-fold by CREB, whereas CREB overexpression in the absence of cAMP did not modulate basal transcriptional activity (Figure 4C⇑). In a similar experiment, cAMP stimulated the activity of a control reporter plasmid, bearing a CRE upstream of luciferase gene (not shown). P/CAF and p300, alone or in combination, stimulated Flk-1/KDR transcription from 2- to 7-fold; however, cAMP and CREB inhibited this effect and their action was additive (Figure 4C⇑). These findings were further investigated by using a dominant-negative CREB (ΔCREB).31 The expression of ΔCREB per se produced a 2-fold increase in promoter basal activity whereas cAMP treatment still exerted its negative effect (Figure 4C⇑). ΔCREB cotransfected with HATs enhanced Flk-1/KDR promoter activity up to 50-fold. Remarkably, a cAMP inhibitory effect was still detectable despite HATs overexpression (Figure 4C⇑), indicating that DNA binding independent mechanisms could also be operating. Therefore, it was attempted to rescue the promoter activity in the presence of cAMP by increasing the total intracellular level of pCAF and p300. It was found that increasing amounts of coactivators rescued, at least up to 50%, cAMP-mediated repression (Figure 4D⇑). These results suggest that the cAMP-mediated negative transcription signals could be transduced via CREB by its functional interaction with HATs.
Site-Directed Mutagenesis of CRE and κB-RE Alters Transcription From the Flk-1/KDR Promoter
To investigate the direct role of CRE and κB-RE, specific nucleotides were mutagenized inside their core regions (Figure 5⇓), according to the results of the oligonucleotide competition experiments (Figures 2C⇑ and 2D⇑). The mCREluc and mκBluc reporter vectors were generated and used in transfection experiments. mCREluc lost sensitivity to cAMP, gaining direct inducibility by p65/RelA even in the absence of cotransfected HATs. Remarkably, p300 and p65/RelA overexpression stimulated Flk-1/KDR promoter activity ≈45- to 50-fold (Figure 5⇓), ie, at levels similar to those achieved by ΔCREB overexpression, previously shown (Figure 4C⇑). Nevertheless, cAMP still exerted its negative effect in the presence of p300, likely because cAMP-activated CREB was still able to bind p300 but not the mutated CRE. The mκBluc construct was still sensitive to cAMP-mediated repression whereas the overexpression of p65/RelA, despite the presence of p300, failed to stimulate transcription. These data indicate that the CRE, κB-RE, and their cognate binding proteins could play distinct and opposite regulatory roles on Flk-1/KDR promoter activity.
It has been previously shown that TNF-α activates NF-κB function and enhances human KDR expression,12 whereas cAMP has been described as a negative regulator of angiogenesis.13 However, their effects on Flk-1/KDR gene transcription are poorly characterized. In the present study, putative cis-acting CRE and κB-RE have been mapped at position +11/+18 and −62/−55, respectively, in the Flk-1/KDR 5′ flanking region. These predicted binding sites for CREB and NF-κB are indeed functional, acting as negative and positive regulators of Flk-1/KDR promoter activity, respectively. In fact, impairing CREB or p65/RelA binding to the DNA by site-directed mutagenesis reduces or abolishes cAMP-mediated repression of Flk-1/KDR promoter transcription as well as p65/RelA stimulatory effect without interfering with the basal level of transcription.
Because CREB and p65/RelA are both capable of HATs recruitment,19 16 we suggest that the positive/negative regulation of the Flk-1/KDR promoter relies on a competition for limited intracellular amounts of HATs. This interpretation is in agreement with a recent study in which the competitive association of CBP/p300 with CREB and p65/RelA has been shown to mediate distinct transcriptional effects on E-selectin and vascular cell adhesion molecule-1 promoters.15 A functional competition for HATs between p53 and p65/RelA has also been described.17 Those reports and the present study suggest that HATs recruitment may represent a general mechanism to control promoter responsiveness to opposing transcription signals.
It is noteworthy that the majority of the experiments reported in the present study were performed by transient transfection of supercoiled plasmid DNA. Under this condition, plasmids are expected to retain an episomal position in the cell nucleus, and this may make them less sensitive to remodeling signals than the cellular chromatin. However, transfected plasmids, which may be organized into nucleosomal-like structures,32 still seem sensitive to histone acetylation.33 Therefore, the biological relevance of our results, despite only being suggestive, is further substantiated by the in vivo regulation of the stable integrated and chromatinized Flk-1GFP reporter construct.
The hypothesis of a direct role for HAT molecules in the process of vasculogenesis and/or angiogenesis, suggested by the present study, is also supported by the analysis of transgenic mice bearing a truncated dominant-negative form of CBP and showing abnormalities in the production of endothelial precursors.34 The development of transgenic animals bearing wild-type or mutated Flk-1/ KDR promoter constructs will be useful to better understand the role of NF-κB and CRE elements in the adaptation of Flk-1/KDR gene expression to proangiogenic or antiangiogenic stimuli in vivo.
Although the regulatory mechanism described in the present study has been reported for other genes in different cell types,14 it may be relevant in a variety of clinical applications aimed at enhancing blood flow to ischemic tissue or downregulating new blood vessel development to inhibit tumor growth.
This work has been partially supported by grants from the European Union (Biomed96, No. 1160; Biomed97, No. 2270), Telethon (No. A61), and Associazione Italiana per la Ricera sul Cancro. The authors would like to thank Mauro H. Citterich for technical assistance in oligonucleotide synthesis and automated sequencing and Gabriella Ricci and Cinzia Carloni for excellent secretarial assistance.
- Received May 8, 2000.
- Accepted June 2, 2000.
- © 2000 American Heart Association, Inc.
Ziegler BL, Valtieri M, Almeida Porada G, De Maria R, Muller R, Masella B, Gabbianelli M, Casella I, Pelosi E, Bock T, Zanjani ED, Peschle C. KDR receptor: a key marker defining hematopoietic stem cells. Science. 1999;285:1553–1558.
Li J, Brown LF, Hibberd MG, Grossman JD, Morgan JP, Simons M. VEGF, flk-1 and flt-1 expression in a rat myocardial infarction model of angiogenesis. Am J Physiol. 1996;270:1803–1811.
Ronicke V, Risau W, Breier G. Characterization of the endothelium-specific murine vascular endothelial growth factor receptor-2 (Flk-1/KDR) promoter. Circ Res. 1996;79:277–285.
Gerber HP, Condorelli F, Park J, Ferrara N. Differential transcriptional regulation of the two vascular endothelial growth factor receptor genes. J Biol Chem. 1997;272:23659–23667.
Hata Y, Duh E, Zhang K, Robinson GS, Aiello LP. Transcription factors Sp1 and Sp3 alter vascular endothelial growth factor receptor expression through a novel recognition sequence. J Biol Chem. 1998;273:19294–19303.
Wu Y, Patterson C. The human KDR/Flk-1 gene contains a functional initiator element that is bound and transactivated by TFII-I. J Biol Chem. 1999;274:3207–3214.
Giraudo E, Primo L, Audero E, Gerber HP, Koowlwijk P, Soker S, Klagsburn M, Ferrara N. Bussolino F. Tumor necrosis factor-α regulated expression of vascular endothelial growth factor receptor-2 and its co-receptor neuropilin-1 in human vascular endothelial cells. J Biol Chem. 1998;273:22128–22135.
Tsopanoglou NE, Haralabopoulos GC, Maragoudakis ME. Opposing effects on modulation of angiogenesis by protein kinase C and cAMP-mediated pathways. J Vasc Res. 1994;4:195–204.
Struhl K. Histone acetylation and transcriptional regulatory mechanisms. Genes Dev. 1998;12:599–606.
Parry GCN, Mackman N. Role of cyclic AMP response element-binding protein in cyclic AMP inhibition of NF-κB mediated transcription. J Immunol. 1997;159:5450–5456.
Vanden Berghe W, De Bosscher K, Boone E, Stephanie P, Haegeman G. The nuclear factor-κB engages CBP/p300 and histone acetyltransferase activity for transcriptional activation of the interleukin-6 gene promoter. J Biol Chem. 1999;274:32091–32098.
Wadgaonkar R, Phelps K, Haque Z, Williams AJ, Silverman ES, Collins T. CREB-binding protein is a nuclear integrator of nuclear factor-κB and p53 signaling. J Biol Chem. 1999;274:1879–1882.
Eckner R, Ewen ME, Newsome D, Gerdes M, DeCaprio JA, Lawrence JB, Livingstone DM. Molecular cloning and functional analysis of the adenovirus E1A-associated 300 kD (p300) reveals a protein with properties of transcriptional adaptor. Genes Dev. 1994;8:869–884.
Mandriota SJ, Menoud PA, Pepper MS. Transforming growth factor β1 down regulates vascular endothelial growth factor receptor-2/Flk-1 expression in vascular endothelial cells. J Biol Chem. 1996;19:11500–11505.
Schreiber E, Matthias P, Muller MM, Shaffner W. Rapid detection of octamer binding proteins with “mini-extracts,” prepared from a small number of cells. Nucleic Acids Res. 1989;17:6419.
Montminy MR, Sevarino KA, Wagner JA, Mandel G, Goodman RH. Identification of a cyclic-AMP-responsive element within the rat somatostatin gene. Proc Natl Acad Sci U S A. 1986;83:6682–6686.
Eggers A, Siemann G, Blume R, Knepel W. Gene-specific transcriptional activity of the insulin cAMP-responsive element is conferred by NF-Y in combination with cAMP response element-binding protein. J Biol Chem. 1998;273:18499–18508.
Tinti C, Yang C, Seo H, Conti B, Kim C, Joh TH, Kim KS. Structure/function relationship of the cAMP response element in tyrosine hydroxylase gene transcription. J Biol Chem. 1997;272:19158–19164.
Jeong S, Stein A. Micrococcal nuclease digestion of nuclei reveals extended nucleosome ladders having anomalous DNA lengths for chromatin assembled on non-replicating plasmids in transfected cells. Nucleic Acids Res. 1994;22:370–375.
Oike Y, Takakura N, Hata A, Kaname T, Akizuki M, Yamaguchi Y, Yasue H, Araki K, Yamamura Ki, Suda T. Mice homozygous for a truncated form of CREB-binding protein exhibit defects in hematopoiesis and vasculo-angiogenesis. Blood. 1999;93:2771–2779.