Vascular Endothelial Growth Factor–Induced Genes in Human Umbilical Vein Endothelial Cells
Relative Roles of KDR and Flt-1 Receptors
Objective— This study evaluated the relative roles of the vascular endothelial growth factor (VEGF) receptors KDR and Flt-1 in the mediation of altered gene expression elicited by VEGF.
Methods and Results— We used mutants of VEGF selective for the KDR and Flt-1 receptors to differentiate gene expression patterns mediated by wild-type VEGF (VEGFwt) in human umbilical vein endothelial cells. RNA was extracted from cells treated for 24 hours with 1 nmol/L of each ligand, and gene expression was monitored by using oligonucleotide arrays (Affymetrix U95A). We report that activation of KDR was sufficient to upregulate all the genes induced by VEGFwt. In contrast, there were no genes selectively upregulated by the Flt-selective mutant. However, high concentrations of the Flt-selective mutant could augment the expression of some genes induced by submaximal concentrations of VEGFwt but not the KDR-selective mutant.
Conclusions— The binding of VEGF to its receptor, KDR, is necessary and sufficient to induce the gene expression profile induced by this growth factor. Furthermore, in human umbilical vein endothelial cells, the Flt-1 receptor appears to act as a decoy receptor, tempering the response to lower concentrations of VEGF.
Vascular endothelial growth factor (VEGF) is a major regulator of normal and pathological angiogenesis. This potent growth factor is secreted by many cell types, including a variety of tumor cells, epithelial cells, and podocytes, and it may be the primary angiogenic factor responsible for physiological angiogenic responses (such as wound healing) and the pathological angiogenesis associated with tumor progression, diabetic retinopathy, rheumatoid arthritis, and macular degeneration.1,2⇓ An unusual feature of VEGF is the observation that even loss of a single allele results in embryonic lethality, strongly suggesting the requirement for highly regulated levels of this growth factor.3,4⇓
The angiogenic effects of VEGF are believed to be mediated by 2 receptor tyrosine kinases, Flt-1 and KDR. The Flt-1 receptor has a higher affinity for VEGF,5 yet its role in the vasculature is poorly understood. Flt-1–deficient mice die in utero at embryonic day 9 and exhibit an increased number of endothelial cells and a failure of the vasculature to organize.6 However, more recent studies using animals homozygous for a mutant of Flt-1 lacking the cytoplasmic domain found the animals to be fertile and devoid of any obvious vascular abnormalities.7 KDR-deficient mice also die in utero at embryonic day 9, apparently because of a pronounced defect in vasculogenesis and blood island formation.8 Flt-1, in contrast to KDR, is not restricted to the endothelium, and its expression has been described on monocytes and vascular smooth muscle cells.9,10⇓
Potential clues to the role of Flt-1 were suggested by the phenotype of placental growth factor (PLGF)-deficient mice. PLGF, which shares 53% identity with the PDGF domain of VEGF, binds Flt-1 with high affinity but does not interact with KDR.11 PLGF exhibits minimal effects on endothelial cell growth or migration, suggesting that Flt-1 activation is insufficient to mediate these effects. PLGF-deficient mice do develop normally but exhibit defects in postnatal retinal vasculature remodeling and in tumor angiogenesis.12 The recent availability of highly selective mutants generated by phage display technology for the VEGF receptors KDR and Flt-1 have enabled a more thorough understanding of the respective roles for these 2 receptors.13 These mutants allow selective activation of Flt-1 or KDR in primary cultures of endothelial cells and thus enable the elucidation of the relevant roles of each VEGF receptor in VEGF-modulated activities in endothelial cells. For example, recent studies14 have demonstrated that KDR appears to mediate many of the known activities of VEGF in endothelial cells, including chemotaxis, survival, mitogenesis, and vascular permeability. To date, no activity for the Flt-1–selective (Flt-1 sel) mutant has been identified in normal nontransformed endothelial cells. However, the Flt-1 sel mutant stimulated matrix metalloproteinase expression in vascular smooth muscle cells13 and was a potent chemotactic agent for porcine aortic endothelial cells stably transfected with a mutant form of Flt in which the juxtamembrane region of KDR was exchanged with the Flt region.14
In the present study, we have selectively activated the Flt-1 or KDR receptor in primary cultures of human umbilical vein endothelial cells (HUVECs) and evaluated the involvement of each individual receptor in mediating VEGF-induced gene expression. We used an oligonucleotide probe array (Affymetrix U95A) to identify genes upregulated >3-fold over basal levels at 24 hours after treatment with wild-type VEGF (VEGFwt). For comparison, we also evaluated RNA generated from cells similarly treated with equimolar KDR-selective (KDR-sel) and Flt-1 sel mutants. The probe array used (U95A) contained >8900 individual expressed sequence tags or genes. A total of 40 genes were upregulated >3-fold in ≥75% of the 2×2 comparisons that were made. The expression of 12 of the identified genes was further validated by a second independent method (real-time quantitative polymerase chain reaction [Real Time PCR], Taqman, Applied Biosystems).
Human Umbilical Vein Endothelial Cells
HUVECs were routinely grown in Clonetics EGM medium supplemented with 10% FBS and endothelial cell growth supplements that were provided by the manufacturer. Cells were incubated overnight with basal medium (M199, 1× ITS, 2 mmol/L l-glutamine, 50 μg/mL ascorbic acid, 26.5 mmol/L NaHCO3, 100 U/mL penicillin, and 100 U/mL streptomycin) supplemented with 1% FBS and then incubated for 24 hours in basal medium, 1% FBS supplemented with 1 nmol/L VEGF165 (VEGFwt), KDR-sel, or Flt-1 sel. These culture conditions were selected on the basis of preliminary studies, which demonstrated that 1 nmol/L VEGFwt elicited a near-maximal increase in bromodeoxyuridine incorporation (not shown) and in the expression of a known VEGF-induced gene, CXCR4 (see Figure 1A). We selected a time point of 24 hours, which would identify those genes that demonstrate either a delayed or stable upregulation of expression (eg, Figure 1B). Total RNA was extracted from the cells with Tri-Reagent (Molecular Research Center) and purified by using an RNeasy mRNA purification kit (Qiagen).
VEGF Mutants and Related Reagents
The generation and characterization of the KDR-sel and Flt-1 sel mutants has been recently described.13,14⇓ These VEGF165 variants have 3 or 4 amino acid substitutions compared with VEGFwt. The Flt-1 sel mutant binds with native affinity to Flt-1 but exhibits ≈128-fold weaker binding to KDR compared with VEGFwt. The KDR-sel variant has 3 changes from the wild-type protein and exhibits wild-type affinity for KDR but ≈2000-fold reduced affinity for Flt-1. More details on the generation and characterization of the mutants have recently been published.13–15⇓⇓
Preparation of cRNA and Array Analysis
Preparation of cRNA, hybridization, and scanning of probe arrays were performed according to the protocols of the manufacturer (Affymetrix). Fluorescently stained probe arrays were visualized with a GeneArray scanner (Hewlett-Packard). Data analysis was performed by using GeneChipAnalysis Software (version 3.2, Affymetrix); subsequent further analysis (clustering, Venn diagrams, and gene trees) used GeneSpring (Silicon Graphics). Pairwise comparisons were made by using time-0 probe arrays as baseline. Two replicate samples were analyzed for each experimental condition. Hence, there were 2 time-0 samples for each treatment and 2 replicates of each condition. Therefore a 2×2 comparison was performed for each time point against each time-0 time point, resulting in 4 pairwise comparisons. Using Affymetrix Data Mining Tool software to examine the qualitative parameters of increase (or marginal increase), we identified those genes that increased in expression compared with the time-0 time points. We performed a nonparametric hypothesis reference versus a control test with a cutoff at a value of P=0.12, which corresponds to an increase call in at least 3 of the 4 comparisons. To further refine our analysis, we considered significant only those increase calls that averaged ≥3-fold the baseline value. This very conservative analytical approach limited the number of false-positive gene identifications. This analysis method and its validation have been previously published.16
Real Time PCR
Real Time PCR (Taqman) was performed as described previously17 with the use of an ABI Prism 7700 Sequence Detector (ABI–Perkin-Elmer). Expression levels for each gene were normalized to cyclophilin, which was unaffected in the different treatment groups. Results are expressed as the mean of 4 independent experiments, with each assayed in duplicate. Table 1 provides the sequences of the Taqman probes and primers used in the present study.
To focus on a set of genes that were significantly increased by VEGFwt, we performed a 2×2 comparison (2 treated versus 2 control) for each group, ie, a total of 4 possible comparisons. On the basis of a Mann-Whitney analysis, we accepted as significant only those genes that were increased in at least 75% (ie, 3/4) of the comparisons made. To further narrow the spectrum of genes evaluated, we then selected only those genes whose expression was elevated ≥3-fold over the baseline value. We selected the group of genes upregulated by VEGFwt and queried the database as to the respective expression of these genes in the KDR-sel and Flt-1 sel groups (we also performed the same analysis with the KDR-sel and Flt-1 sel groups, and all of the genes identified were present in the VEGFwt upregulated group [not shown]). On the basis of above criteria, 40 different genes (of >8900) were upregulated >3-fold by VEGFwt at 24 hours. (One gene, angiopoeitin-2, was detected twice in 2 independent probe sets.) A hierarchical cluster of the 40 genes identified by this analysis is depicted in Figure 2. It can be readily seen that there is a close relationship between the genes upregulated by VEGFwt and the KDR-sel mutant. In contrast, relatively few genes appeared to be upregulated by the Flt-1 sel mutant.
On the basis of Affymetrix array analysis, the genes upregulated by KDR-sel exhibited a much greater increase than those modulated by Flt-1 sel, and the majority of genes identified as upregulated by the KDR-sel ligand were not significantly modulated by Flt-1 sel. A summary of the 40 genes, their accession numbers, and known functions is provided in Table 2. As shown in Table 2, there was no predilection for any particular class of gene to be induced: the list of genes upregulated includes growth factors, receptors, protease inhibitors, actin-binding proteins, transcription factors, and a number of genes of poorly defined function. Some of those identified (eg, angiopoeitin-218 and CXCR419) have been previously reported to be upregulated by VEGF. Interestingly, a well-known VEGF-responsive gene, endothelial NO synthase,20 was not represented on this list. This is probably due to the time point (24 hours) evaluated in this experiment. Bouloumie et al20 found that VEGF upregulated eNOS mRNA in HUVECs at 4, 8, and 12 hours but that by 24 hours, eNOS mRNA levels were close to the level observed in control cells. Interestingly, the majority of the genes identified represent downstream targets of VEGF-driven endothelial activation.
To validate the gene expression data by a second, more sensitive, quantitative, and independent method, we analyzed the expression of 12 of the identified genes in VEGF-treated, KDR-sel–treated, and Flt-1 sel–treated cells by using Real Time PCR (Taqman). These analyses were performed on 4 independent experiments, and for each sample, duplicate aliquots were analyzed by Taqman. On the basis of the expression profiles depicted in Figure 2, we selected representative genes for validation that appeared to be upregulated by KDR, Flt-1 sel, or both. As can be seen in Table 3, the relative expression of all of these genes was validated by Real Time PCR, although the effects of VEGFwt on SY14 expression did not reach statistical significance. In every case, the VEGFwt and KDR-sel ligand elicited a similar magnitude of response, whereas there was no significant increase in gene expression in response to Flt-1 sel. In other experiments, we evaluated the effects of Flt-1 sel at higher concentrations (10 nmol/L) and still were unable to demonstrate a significant increase in the expression of those genes depicted in Table 3 (not shown).
To determine a possible decoy role of Flt-1, we evaluated the effects of combining VEGFwt, KDR-sel, and Flt-1 sel. At a concentration of 1 nmol/L, the effects of VEGFwt appeared to be maximal, and coincubation with KDR-sel (2 nmol/L) or Flt-1 sel (2 nmol/L or 10 nmol/L) did not result in any further increases in the mRNA levels for the 3 genes tested (thrombomodulin, STC-1, or CXCR-4; not shown). However, when we tested a lower dose of VEGFwt (0.5 nmol/L), the addition of Flt-1 sel (10 nmol/L) with the VEGFwt (0.5 nmol/L) significantly increased the expression of STC-1 and thrombomodulin (Table 4). In contrast, combining KDR-sel (0.5 nmol/L) with Flt-1 sel (10 nmol/L) did not significantly alter the response to KDR-sel (not shown).
In the present study, we have compared the patterns of gene expression in HUVECs exposed to VEGFwt, KDR-sel, or Flt-1 sel mutants at 24 hours. HUVECs are a commonly used cellular model, human in origin, and represent low passage cultures of nontransformed cells. Our group and others have used HUVECs in numerous angiogenesis-related assays to explore the functions of VEGF and other growth factors. To increase the sensitivity of the assay, we used cells that had been serum-starved and then exposed to VEGFwt or the mutants. This approach reduces background problems with serum and growth factors present in serum and provides more standardized experimental conditions. In addition, all the cells were used at the same passage number and had been grown under identical conditions (eg, the same serum lots, plastic ware, and incubator) to reduce experimental variability. The data from the present study clearly suggest that the VEGF receptor Flt-1 does not selectively signal any major pattern of gene expression in HUVECs. In previous studies by our group and others, the Flt-1 and KDR-sel ligands were both shown to be highly selective and active on their respective receptors.13,14⇓ We and others have also shown that HUVECs express KDR and Flt-1 receptors, so the selective activity of KDR cannot be explained by a lack of Flt-1 expression.1,15⇓ Previous studies (Li, Gille, and colleagues13,14⇓ and our laboratory15) using these receptor-selective variants have strongly suggested that the KDR receptor mediates the mitogenic, chemotactic, tubulogenic, and survival activities of VEGFwt.
In the present study, we used an open-ended unbiased approach to survey potential activities activated by the 2 receptors with the use of Affymetrix oligonucleotide probe arrays. Using conservative statistical models, we identified 40 genes as significantly upregulated by VEGF and the KDR-sel mutant at 24 hours. There were no genes identified (which reached our statistical cutoffs for significance) as selectively upregulated by the Flt-1 sel mutant. We confirmed the selective expression of 11 of the genes identified in the study by a second independent method, namely, Real Time PCR. In an attempt to identify genes selectively upregulated by Flt-1, we also lowered our stringency criteria (upregulated by >2 fold in 3/4 of the comparisons) and still were unable to identify any Flt-1–selective genes (not shown).
On the basis of these observations, we conclude that the Flt-1 receptor may cooperate with KDR in modulating downstream events, but there does not appear to be a selective pathway activated by this receptor in HUVECs. Our observations do not negate the possibility that Flt-1 might signal selective pathways in endothelial cells from other organs or the more likely possibility that in smooth muscle cells or monocytes, an Flt-1–selective program of gene expression is activated. Moreover, because we performed an analysis at a single time point (24 hours) and did not survey all possible genes, there still exists a remote possibility that there are a few Flt-1–selective genes in HUVECs. We did evaluate VEGFwt, KDR-sel, and Flt-1 sel gene expression on a second probe array, U95B, which contains >8900 additional probe sets, most of which are expressed sequence tags. Similar to the data with U95A, however, no Flt-1–selective genes could be identified (not shown). Thus, in an unbiased sampling of >17 800 probe sets, likely representing 30% to 40% of the estimated human genome, we were unable to detect a statistically significant Flt-1–selective signaling pathway in HUVECs.
Despite its higher affinity for VEGF, the autophosphorylation of Flt-1 in endothelial cells on ligand binding is often difficult to detect. This is thought to be due, at least in part, to the fact that there are only 2 major phosphorylation sites on this receptor.21 Signal transduction downstream from Flt-1 has been very difficult to demonstrate in endothelial cells. Our gene expression studies would concur with the signaling studies; ie, an Flt-1–selective response is difficult to demonstrate in HUVECs. Our studies demonstrate that the binding of VEGF to KDR is sufficient to elicit the specific program of VEGF-induced gene expression and that contribution or cooperation with Flt-1 is not required to activate these responses. We were able to demonstrate that at submaximal concentrations of VEGFwt, an excess of the Flt-1 sel ligand could augment the response to VEGFwt, in agreement with the possible decoy role of Flt-1 proposed by our group and others. Therefore, it follows that modulation of the relative expression of KDR and Flt-1 receptors in endothelial cells may serve as a mechanism for titration of the response to VEGFwt and could potentially modulate the spectrum of activities mediated by this potent growth factor.
Received September 4, 2002; revision accepted September 9, 2002.
- ↵Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig M, Vandenhoeck A, Harpal K, Eberhardt C, Declercq C, Pawling J, Moons L, Collen D, Risau W, Nagy A. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature. 1996; 380: 435–439.
- ↵de Vries C, Escobedo JA, Ueno H, Houck K, Ferrara N, Williams LT. The fms-like tyrosine kinase, a receptor for vascular endothelial growth factor. Science. 1992; 255: 989–991.
- ↵Hiratsuka S, Minowa O, Kuno J, Noda T, Shibuya M. Flt-1 lacking the tyrosine kinase domain is sufficient for normal development and angiogenesis in mice. Proc Natl Acad Sci U S A. 1998; 95: 9349–9354.
- ↵Shen H, Clauss M, Ryan J, Schmidt AM, Tijburg P, Borden L, Connolly D, Stern D, Kao J. Characterization of vascular permeability factor/vascular endothelial growth factor receptors on mononuclear phagocytes. Blood. 1993; 81: 2767–2773.
- ↵Wang H, Keiser JA. Vascular endothelial growth factor upregulates the expression of matrix metalloproteinases in vascular smooth muscle cells: role of flt-1. Circ Res. 1998; 83: 832–840.
- ↵Park JE, Chen HH, Winer J, Houck KA, Ferrara N. Placenta growth factor: potentiation of vascular endothelial growth factor bioactivity, in vitro and in vivo, and high affinity binding to Flt-1 but not to Flk-1/KDR. J Biol Chem. 1994; 269: 25646–25654.
- ↵Carmeliet P, Moons L, Luttun A, Vincenti V, Compernolle V, De Mol M, Wu Y, Bono F, Devy L, Beck H, Scholz D, Acker T, DiPalma T, Dewerchin M, Noel A, Stalmans I, Barra A, Blacher S, Vandendriessche T, Ponten A, Eriksson U, Plate KH, Foidart JM, Schaper W, Charnock-Jones DS, Hicklin DJ, Herbert JM, Collen D, Persico MG. Synergism between vascular endothelial growth factor and placental growth factor contributes to angiogenesis and plasma extravasation in pathological conditions. Nat Med. 2001; 7: 575–583.
- ↵Li B, Fuh G, Meng G, Xin X, Gerritsen M, Cunninghan B, de Vos A. Receptor-selective variants of human VEGF: generation and characterization. J Biol Chem. 2000; 275: 29823–29828.
- ↵Yang S, Xin X, Zlot C, Ingle G, Fuh G, Li B, Moffat B, de Vos AM, Gerritsen ME. Vascular endothelial cell growth factor–driven endothelial tube formation is mediated by vascular endothelial cell growth factor receptor-2, a kinase insert domain-containing receptor. Arterioscler Thromb Vasc Biol. 2001; 21: 1934–1940.
- ↵Jin H, Yang R, Awad TA, Wang F, Li W, Williams SP, Ogasawara A, Shimada B, Williams PM, de Feo G, Paoni NF. Effects of early angiotensin-converting enzyme inhibition on cardiac gene expression after acute myocardial infarction. Circulation. 2001; 103: 736–742.
- ↵Oh H, Takagi H, Suzuma K, Otani A, Matsumura M, Honda Y. Hypoxia and vascular endothelial growth factor selectively up-regulate angiopoietin-2 in bovine microvascular endothelial cells. J Biol Chem. 1999; 274: 15732–15739.
- ↵Salcedo R, Wasserman K, Young HA, Grimm MC, Howard OM, Anver MR, Kleinman HK, Murphy WJ, Oppenheim JJ. Vascular endothelial growth factor and basic fibroblast growth factor induce expression of CXCR4 on human endothelial cells: in vivo neovascularization induced by stromal-derived factor-1alpha. Am J Pathol. 1999; 154: 1125–1135.
- ↵Ito N, Wernstedt C, Engstrom U, Claesson-Welsh L. Identification of vascular endothelial growth factor receptor-1 tyrosine phosphorylation sites and binding of SH2 domain-containing molecules. J Biol Chem. 1998; 273: 23410–23418.