Notch regulates the angiogenic response via induction of VEGFR-1
© Funahashi et al; licensee BioMed Central Ltd. 2010
Received: 4 September 2009
Accepted: 26 January 2010
Published: 26 January 2010
Skip to main content
© Funahashi et al; licensee BioMed Central Ltd. 2010
Received: 4 September 2009
Accepted: 26 January 2010
Published: 26 January 2010
Notch is a critical regulator of angiogenesis and arterial specification. We show that ectopic expression of activated Notch1 induces endothelial morphogenesis in human umbilical vein endothelial cells (HUVEC) in a VEGFR-1-dependent manner. Notch1-mediated upregulation of VEGFR-1 in HUVEC increased their responsiveness to the VEGFR-1 specific ligand, Placental Growth Factor (PlGF). In mice and human endothelial cells, inhibition of Notch signaling resulted in decreased VEGFR-1 expression during VEGF-A-induced neovascularization. In summary, we show that Notch1 plays a role in endothelial cells by regulating VEGFR-1, a function that may be important for physiological and pathological angiogenesis.
Vascular endothelial growth factor-A (VEGF-A) is essential to the multistep process of vascular development, and proper vessel formation in a variety of settings is exquisitely sensitive to levels of VEGF-A [1–4]. VEGF-A signals through two receptor tyrosine kinases: VEGFR-1 (flt1) and VEGFR-2 (flk1), while placenta growth factor (PlGF) signals exclusively through VEGFR-1. Both VEGF-A and PlGF induce endothelial cell proliferation, survival, and migration [3, 5, 6]. The role of VEGFR-1 in angiogenesis has largely been defined in terms of its opposition to VEGFR-2. VEGFR-2 is considered the primary VEGF-A receptor that drives angiogenesis, while VEGFR-1 has high binding affinity for VEGF-A but weak kinase activity. Thus, VEGFR-1 is thought to function mainly as a decoy receptor that sequesters VEGF-A [7–11]. This concept is supported by analysis of mouse models where deletion of flt1 led to vessel overgrowth and disruption of vascular patterning . In addition, mice expressing a mutant allele of flt1 that lacks the tyrosine kinase domain (flt1TK-/-) did not exhibit the vascular patterning defects seen in flt1-/-mice, suggesting that in embryonic development, the kinase activity of VEGFR-1 was dispensable and that its predominant function is via its high affinity binding to VEGF-A . Despite this, a positive function for VEGFR-1 in angiogenesis has been demonstrated in a variety of settings. flt1TK-/- mice displayed defects in tumor vessel formation and metastasis [13, 14], and inhibition of VEGFR-1 led to defects in neovascularization of the eye . The signaling pathways that regulate VEGFR-1 expression in endothelial cells remain unclear.
Notch, a receptor that functions in cell fate decisions, has been shown to be downstream of VEGF-A in endothelial sprouting [16, 17] and arterial specification [18, 19]. The Notch proteins are highly conserved trans-membrane receptors that are required for normal embryonic development. In mammals, there are four Notch proteins (Notch1-4) that, upon binding with one of five ligands, termed Delta-like (Dll) and Jagged, are subject to a series of proteolytic cleavages by ADAM metalloproteases and gamma-secretase. Cleavage releases the intracellular domain of the Notch receptor, which translocates to the nucleus and functions as a transcriptional activator in complex with the transcription factors CSL (CBF1, Su(H), Lag-2), Mastermind, and histone acetyltransferases. To date, the importance of the Notch pathway in regulating endothelial cell response to VEGF-A has been studied with respect to its effect on VEGFR-2, as it has been shown that Delta-like 4 (Dll4) signaling represses VEGFR-2 expression [16, 20, 21]. Current models assert a role for Dll4 in restricting sprouting angiogenesis [20, 22–24], but have not identified the Notch receptors that are important for this effect, or whether Notch signaling can function positively in endothelial cell morphogenesis. In addition, whether Notch signaling through a particular receptor can regulate VEGFR-1 expression in endothelial cells has not been defined.
Using ectopic expression as well as protein-based, and pharmacological loss of Notch function, we show that VEGFR-1 expression is downstream of Notch signaling in endothelial cells. Furthermore, we define a positive role for Notch signaling in VEGF-driven morphogenesis of endothelial cells via promotion of cell extension which we demonstrate requires upregulation of VEGFR-1. Coincident with the Notch-mediated upregulation of VEGFR-1, we report Notch signaling enhances endothelial cell responsiveness to PlGF. Finally, in an assay of VEGF-A induced dermal angiogenesis, we show that a protein based Notch inhibitor, the Notch1 decoy, can reduce VEGFR-1 levels in neovessels. Collectively, our data define a role for Notch in mediating the response of endothelial to angiogenic stimuli by regulation of VEGFR-1.
ZD1893, PD166866, and SU5416 are from Eisai Co., Ltd. Compound E was obtained from the Korean Research Institute of Chemical Technology. PlGF was obtained from Research Diagnostics Institute. N1IC , LacZ, and VEGF-A constructs were engineered into pAdlox vector and adenovirus stocks were produced . Notch1 decoy has been described . Briefly, the extracellular domain of rat Notch1 (bp 241-4229, accession no. X57405) was fused to human IgG Fc and engineered into pAdlox vector (Ad-Notch1 decoy) and adenovirus stocks generated.
HUVEC were isolated from human umbilical vein as described  and cultured in complete medium (EGM-2 Bullet kit, LONZA) on porcine type I collagen (Nitta Gelatine). KP1/VEGF121 cells were provided by Eisai Co., Ltd,  and maintained in RPMI 1640 containing 10% FBS. HUVEC were infected with Ad-LacZ, Ad-N1IC, Ad-VEGF-A, Ad-GFP, or Ad-Notch1 decoy at a MOI of 40. HUVEC were co-infected with Ad-LacZ and Ad-Notch1 decoy at a MOI of 40 for each virus. HUVEC infected with Ad-LacZ at a MOI of 80 served as a control. Retroviral control and N1IC-expressing HUVEC lines were generated as previously described . Control, VEGFR-1, and VEGFR-2 siRNA (Santa Cruz) were introduced into HUVEC using Effectene Reagent (Qiagen). Total RNA or cell lysate was harvested 48 hours after siRNA transfection.
HUVEC were seeded on type I collagen gels two days after adenoviral infection or retroviral infection and 5 days later total RNA was isolated with RNeasy mini kit (Qiagen). First-strand cDNA was synthesized using SuperScript First-Strand Synthesis System (Invitrogen). For RT-PCR, primers were designed to recognize human and mouse transcripts of VEGFR-1, VEGFR-2, VEGF-A, PlGF, GAPDH and beta-actin, (primer sequence available upon request). PCR used Platinum Taq DNA polymerase (Invitrogen) and reactions performed for 25 or 30 cycles. Reactions were performed in triplicate.
HUVEC were cultured on type I collagen gels for 5 days in complete medium, then starved in serum free medium for 48 hours and cell lysates were collected with TENT lysis buffer. Western blots were performed using antibodies against Flt1 (C-17, Santa Cruz), Flk1 (C-1158, Santa Cruz), and alpha-tubulin (Sigma). To validate Notch1 decoy secretion, serum-free medium from adenovirally transduced HUVEC was used for western blot analysis using an antibody against the Fc tag (Pierce).
Adenovirus infections were performed two days before seeding on porcine type I collagen, and HUVEC morphogenesis was assessed by microscopy after 5 days, as described . Extensions were scored as number of cells with single or multiple processes per 10× microscopy field. Processes were defined as extensions at acute angles to the cell body that alter normal HUVEC morphology. For each experiment, at least five 10× fields of cultures from each condition were scored. Kinase inhibitors were added to the medium one hour after HUVEC seeding, and PlGF was added at the time of HUVEC seeding. For knockdown experiments, siRNA was transfected two days after adenvoviral infection and the cells were cultured for three days before assessment of HUVEC with cellular extensions. Cell number was measured using Cell Counting Kit-8 (Dojindo).
The Dorsal Air Sac (DAS) assay was performed as described . Millipore chambers were packed with 5.0 × 106 KP1/VEGF121 cells that were transduced (60 MOI) with either Ad-GFP or Ad-Notch1 decoy and transplanted into a DAS of C57BL/6 mice. Mice were sacrificed four days after implantation and implants harvested and embedded in OCT. Each group consisted of at 3-5 mice, and experiments done in triplicate.
5-μm serial sections of KP1/VEGF121 implants were immunostained as described . The following antibodies were used: PECAM (553370, BD Pharmingen), Flt1 (AF417, R&D Systems), Flk1 (AF644, R&D Systems). Quantitative analysis of CD31, Flk1, and Flt1 immunostaining of skin was performed on serial sections using an Eclipse E800 microscope and Nikon DXM 1200 camera, with ImagePro Plus software (Silver Spring, MD). Measurements were made in five different areas in each sample at 20× magnification and average density ratio was determined by dividing the area of specific staining by the total area of the smooth muscle layer.
2 × 105 HUVEC were seeded per well in a collagen-coated 6-well plate. 24 hrs after seeding, cells were stimulated with 50 ng/ml recombinant VEGF-A (R&D Systems) in complete medium, with or without 200 nM Compound E (Korean Research Institute). DMSO was used to treat control cells. 24 hours post-stimulation, cells were harvested with cold PBS, washed, and incubated with rabbit-anti VEGFR-1 (Santa Cruz) for 45 minutes at 4°C. After washing, cells were labeled with anti-rabbit-APC (Jackson Immunoresearch) for 25 minutes at 4°C. Flow cytometry was performed and 10,000 cells per experimental group were counted using FACSCalibur and CellQuestPro acquisition software (BD Biosciences).
Data were expressed as mean plus or minus SEM. Statistical analysis was performed by 2-tailed student t test. P value of less than 0.05 is indicated with ⋆, P value of less than 0.02 is indicated with *. All data shown is representative of at least 3 independent experiments.
This regulation was also found in cultured HUVEC, where VEGF-A-induced expression of VEGFR-1 was reduced by co-expression of the Notch1 decoy, as shown by RT-PCR (Figure 4F). In contrast, induction of VEGFR-2 by VEGF-A in HUVEC was unaffected by the Notch1 decoy (Figure 4F). Similarly, VEGFR-1 expression on the surface of VEGF-A-treated HUVEC was suppressed by treatment with a gamma secretase inhibitor (GSI), Compound E, as analyzed by flow cytometry (Figure 4G). Thus, two means of Notch inhibition were used to establish that VEGF-A induces Notch signaling which in turn regulates VEGFR-1 and that this regulatory pathway is active in both cultured endothelial cells and neovessels in mice.
Our results show that VEGFR-1 is downstream of Notch1 signaling in endothelial cells. We identify a positive role for Notch signaling in endothelial morphogenesis via the induction of cellular extensions mediated by VEGFR-1. Supporting this conclusion is the observation that Notch increases VEGFR-1 levels and this increase correlated with increased endothelial responsiveness to the VEGFR-1-specific ligand, PlGF. Using a protein-based Notch inhibitor, Notch1 decoy, or a gamma secretase inhibitor, we demonstrate that perturbation of endogenous Notch signaling resulted in reduced VEGFR-1 expression. Thus, loss- and gain- of function studies show that Notch signaling regulates VEGFR-1 expression in HUVEC and dermal neovessels.
Previous studies have demonstrated a role for the Notch ligand, Dll4, in inhibiting a tip cell phenotype in the developing vasculature of the retina [16, 17]. In addition, Harrington et al  have shown that VEGFR-1 is upregulated by Dll4, and demonstrated that Dll4 signaling inhibited sprout length in a HUVEC tubulogenesis assay. The authors suggest that Dll4 signaling inhibits angiogenesis by inducing VEGFR-1 . In summary, previous studies have found a negative role for Notch signaling in endothelial cell sprouting, and have focused on this signaling pathway at the level of the ligand, Dll4. However, in these studies, the Notch receptor responsible for these effects is not defined and the possibility of divergent effects of different Notch receptors is not addressed. By focusing on the effects of Notch signaling at the level of the receptor, our results add new insights to the role of Notch and VEGFR-1 in sprouting angiogenesis. In contrast to previous studies, our data suggest that in some settings, Notch signaling may play a positive role in endothelial cell extension of filopodia-like structures via its regulation of VEGFR-1 and supports a novel role Notch1-mediated regulation of VEGFR-1 in endothelial cell morphogenesis.
It has recently been found that VEGFR-1 promotes vascular sprout formation and branching morphogenesis [35, 36]. Kearney et al  propose that this results from VEGFR-1 binding to VEGF-A, thereby regulating the amount of VEGF-A that is available to interact with VEGFR-2. They also show that soluble VEGFR-1 (sVEGFR-1) can promote sprout formation and migration. The positive effect of Notch signaling on HUVEC sprouting that we report may therefore be due to its effect on VEGFR-1, and subsequently, on local levels and availability of VEGF-A. This may particularly be the case if the predominant effect of Notch signaling is due to regulation of sVEGFR-1. In general, the relative proportion of the membrane bound and secreted isoform of VEGFR-1 does not change significantly (data not shown, and Kappas et al ), therefore, we cannot entirely exclude the possibility that Notch-induced sprouting in HUVEC is due to sequestration of VEGF-A. However, we show that Notch-induced sprouting in HUVEC is enhanced in the presence of PlGF, a VEGFR-1 specific ligand, suggesting that signaling through the VEGFR-1 receptor itself, and not simply its function as a 'VEGF-A sink,' may be responsible for Notch-mediated sprouting. This is further supported by the fact that VEGFR-1 siRNA inhibited Notch-induced sprouting in HUVEC while VEGFR-2 siRNA had only a modest effect. Thus, our data support the conclusion that activation of Notch signaling in HUVEC can induce extensions via VEGFR-1, and highlight the possibility that Notch signaling may act through VEGFR-1 to have a positive effect on endothelial cell morphogenesis.
It has been reported that inhibition of VEGFR-1 in the developing retina does not effect sprouting and filopodia extensions in endothelial cells [3, 16]. In the retina, endothelial tip cell filopodia are guided by a gradient of VEGF-A provided by a template of astrocytes [3, 37]. However, in our model of in vitro sprouting in HUVEC, as well as in many in vivo settings of physiological and pathological angiogenesis, the source of VEGF-A is likely to be more diffuse. Notch-mediated sprouting via regulation of VEGFR-1 may constitute a mechanism for endothelial cell morphogenesis that is important in settings where Notch1 is highly expressed in the vasculature and where expression of VEGF-A is more global, and endothelial cell sprouting less controlled, than in formation of the retinal plexus. In addition, our finding that Notch-induced sprouting in endothelial cells is enhanced by PlGF may be relevant in angiogenic settings where PlGF is a major angiogenic factor. Since PlGF is upregulated in pathological conditions by various stimuli [38–40], and contributes to the angiogenic switch in various pathologies [6, 41, 42], Notch-mediated upregulation of VEGFR-1 may prove an important step in disease progression in these contexts. Furthermore, our finding that blockade of Notch signaling using a protein-based inhibitor of Notch1 (Notch1 decoy) resulted in decreased expression of VEGFR-1 in an in vivo model of angiogenesis may have important implications for the efficacy of inhibition of Notch signaling in settings where VEGFR-1 expression is prominent, such as in certain tumor types and in the initiation of premetastatic niches [43–45].
This work was supported by the following grants from the National Institutes of Health: (R01HL62454, R01CA136673) (JK), (F31HL090032-01) (HHO), (5K01DK744629, R01CA136673) (CJS), (5T32 DK07328)(MV).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.