Hypoxia-Induced Arterial Differentiation Requires Adrenomedullin and Notch Signaling

Abstract

Hypoxia (low oxygen) and Notch signaling are 2 important regulators of vascular development, but how they interact in controlling the choice between arterial and venous fates for endothelial cells during vasculogenesis is less well understood. In this report, we show that hypoxia and Notch signaling intersect in promotion of arterial differentiation. Hypoxia upregulated expression of the Notch ligand Dll4 and increased Notch signaling in a process requiring the vasoactive hormone adrenomedullin. Notch signaling also upregulated Dll4 expression, leading to a positive feedback loop sustaining Dll4 expression and Notch signaling. In addition, hypoxia-mediated upregulation of the arterial marker genes Depp, connexin40 (Gja5), Cxcr4, and Hey1 required Notch signaling. In conclusion, the data reveal an intricate interaction between hypoxia and Notch signaling in the control of endothelial cell differentiation, including a hypoxia/adrenomedullin/Dll4 axis that initiates Notch signaling and a requirement for Notch signaling to effectuate hypoxia-mediated induction of the arterial differentiation program.

Introduction

D uring early embryogenesis, blood vessels are formed de novo as a primary capillary plexus in a process known as vasculogenesis [1,2]. The subsequent remodeling of the primary vessels is known as angiogenesis, which includes extensive sprouting and branching as well as recruitment of vascular smooth muscle cells [2]. To become functional, the vascular system has to develop into 2 distinct parts: an arterial and a venous system. The arterial system is the afferent loop, designed to endure a high and pulsatile blood flow. The venous system, on the other hand, is a low-pressure, high-capacitance system that transports blood back to the heart. It was originally thought that the differences between the arterial and venous systems arose from the different mechanical constraints caused by the afferent and efferent blood flow. It has, however, become increasingly clear that molecular programs influence the decision to become arterial or venous endothelial cells very early in development—in fact, before blood is pumped through the system [3]. Arterial versus venous differentiation is, at least in part, controlled by Sonic Hedgehog signaling, which promotes vascular endothelial growth factor (VEGF) expression [4]. High concentrations of VEGF, in turn, promote the arterial fate and induce MEK/ERK signaling and expression of the transcription factors FoxC1 and FoxC2, which leads to activation of the Notch signaling pathway [4 –7]. Notch signaling activates its immediate downstream genes Hey 1 and Hey2 (Gridlock in zebrafish) and induces Ephrin B2, a key marker for arterial endothelial cells [8]. At the ventral side, with a low concentration of VEGF, the transcription factor COUP-TFII (Nr2f2) is expressed at high levels, accompanied by repression of Notch signaling. This leads to expression of EphB4, which is a marker for venous endothelial cells [9].

In addition to Notch signaling, the cellular response to hypoxia is an important regulator of vascular development. At the onset of vascularization, the embryo relies solely on diffusion of oxygen, which means that the early phases of vascularization take place in a hypoxic environment. Cells respond to hypoxia through a molecular machinery, in which the HIF transcription factors (HIF-1α or HIF-2α) become stabilized under low oxygen conditions. HIF, together with its partner HIF-1β (ARNT), activates downstream genes, including VEGF [10]. The important role for hypoxia in the vasculature is underscored by the effects of ablating key factors in the cellular hypoxic response. Arnt^−/− mice die within 10.5 days of embryonic development (E10.5) and are characterized by defective vascular remodeling of the yolk sac, brachial arteries, and placenta [11]. A similar vascular phenotype is observed in Hif-1α^−/− embryos that die at E11 with regressed vascular endothelium [12,13].

Notch signaling is initiated by activation of transmembrane Notch receptors by ligands on juxtaposed cells. Ligand activation leads to proteolytic cleavages of the Notch receptor, and the final cleavage, which liberates the Notch intracellular domain (Notch ICD), is executed by the γ-secretase complex, and can thus be blocked by γ-secretase inhibitors [14]. Notch ICD translocates to the nucleus, where it interacts with the DNA-binding protein CSL to control expression of Notch downstream genes, including the genes encoding the Hes and Hey transcription factors. Mouse embryos deficient for CSL or the Jagged1 ligand die around embryonic day 9 (E9) with vascular collapse, and removal of only one copy of the Dll4 ligand is sufficient to induce severe vascular deformations and death around the same time [15 –18]. Notch signaling and the cellular hypoxic response intersect at several levels in the signaling cascades. This includes the interaction between HIF and Notch ICD, and the enzymatic modification of both HIF and Notch ICD by FIH [19 –21]. There are also several indications that the 2 signaling pathways may interact during vascular development. A role for hypoxia in arterial differentiation has been proposed and linked to an intersection with the Notch signaling pathway; that is, a hypoxia-induced upregulation of the Notch downstream genes Hey1 and Hey2, coupled with a repression of COUP-TFII, was observed [22]. In another study, a link between HIF-1α and Notch via VEGF was established in choroidal neovascularization [23].

In light of the important roles of both Notch and the cellular hypoxic response during vascular differentiation and because the 2 signaling pathways interact, we have studied their roles and possible intersection during the choice between the arterial and venous fates in endothelial cells. In this report, we identify a hypoxia/adrenomedullin/Dll4 axis, in which hypoxia induces expression of Dll4 and elevates Notch signaling in an adrenomedullin-dependent manner. We also show that active Notch signaling is required for a significant portion of all hypoxia-induced genes to be upregulated by hypoxia. The set of hypoxia-induced genes that depend on Notch signaling included a set of arterial marker genes: Depp, Dll4, Connexin40 (Gja5), Cxcr4, and Hey1. We propose that the hypoxia/adrenomedullin/Dll4 axis initiates Notch signaling, which is then required for the hypoxia-induced activation of a larger arterial program. In conclusion, these data provide novel insights into the process of arterial differentiation and identifies an intricate interaction between Notch and the cellular hypoxic response in this process.

Materials and Methods

Embryonic stem cell culture and differentiation

Wild-type mouse CCE embryonic stem (ES) cells (a gift from Dr M. J. Evans, Welcome/CRC Institute, Cambridge, UK), Vegf^−/− [24], and Csl^−/− ES cells were a kind gift from Dr. Tasuku Honjo [18]. Notch1ΔE^TetOn ES cells have been previously described [25,26]. ES cells were cultured on gelatin-coated plates in the Glascow MEM (Sigma) supplemented with 15% fetal bovine serum (FBS) (Hyclone), 1,000 IU/mL leukemia inhibitory factor, 0.1 mM 2-mercaptoethanol, and 0.1 mM nonessential amino acids (Gibco). The medium was changed daily and cells passaged every second day. For differentiation, 10⁴ ES cells/mL were aggregated as embryoid bodies (EBs) in 20 μL drops and differentiated 4 days in the IMDM, 15% FBS (Gibco), 0.15 mM monothioglycerol, and 50 μg/mL ascorbic acid (Sigma). On day 4, EBs were harvested and dissociated, stained for Flk1, and sorted to purity using flow cytometry. About 5×10⁴ Flk1⁺ cells/mL were plated on collagen type IV (BD Bioscience)-coated plates and differentiated for 4 additional days with 50 ng/mL hrVEGF (R&D Systems). The endothelial CD31⁺ cells were isolated by flow cytometry for further analysis. Adrenomedullin was used at 1 μM (H-2932; BachemAG), and to block Notch signaling, the gamma-secretase inhibitor (GSI) L-685.458 was used at 4 μM (BachemAG).

Flow cytometry

Day 4 EBs were dissociated to single cells with collagenase (Sigma-Aldrich) treatment for 20 min, followed by 2-min trypsinization and passed 3 times through a 20G needle. Day 8 monolayer cultures were dissociated by brief trypsinization followed by trypsin inactivation with ice cold PBS with 4% fetal calf serum (FCS). Staining was performed as previously described [5] on ice in PBS with 4% FCS and analyzed using a FACSAria cell sorter (Becton Dickinson). Propidium iodide was used to exclude dead cells. Antibodies used include Flk1-PE (1:200) and CD31-APC (1:200), (PharMingen).

Real-time PCR

10⁴ CD31⁺ cells from each biological sample were sorted by flow cytometry directly into a 350 μL RLT-lysis buffer and total RNA was isolated using the RNeasy Micro Kit (Qiagen). cDNA was synthesized using the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen) according to the manufacturer's instructions. The real-time PCR was performed using the SYBR-GREEN PCR Master Mix and analyzed using a 7300 Real-Time PCR System (Applied Biosystems). Values were normalized to Actb housekeeping transcript levels. Primer sequences are available upon request.

Transfections and luciferase reporter assay

Flk1⁺ cells were allowed to attach over 24 h before transfection in 6-well plates using the FuGENE HD Transfection Reagent (Roche). For the reporter assay, a beta-gal construct was cotransfected as an internal control of transfection efficiency. After an additional 3 days of differentiation, the CD31⁺ cells were sorted and lysed, and their luciferase and beta-gal activity was measured on a NOVOstar microplate reader (LabVision, Beckman Coulter). HIF-1α-PPA and HIF-2α-PPA were both stabilized against normoxic prolyl hydroxylation and subsequent ubiquitin-mediated proteolytic degradation by point mutations [27].

siRNA

Cultures were transfected with siRNA following Flk1 sorting using DharmaFECT1 (T-2001–02; Thermo Scientific) according to the manufacturer's specifications. Two μL DarmaFECT and 100 μL siRNA (2 μM) were added to each well of a 6-well plate in a total volume of 2 mL. To avoid cell toxicity, the transfection media was replaced after 24 h. We used ON-TARGETplus SMARTpool oligos against adrenomedullin (L-042880-01; Thermo Scientific) and scramble oligos as control.

Immunoblotting

EBs were plated in tissue culture dishes at day 4 in a differentiation medium with 50 ng/mL VEGF. At day 10, the cultures were stimulated for 60 min with either VEGF-A165 at 50 ng/mL or adrenomedullin at 1 μM (H-2932; BachemAG). The cells were lysed in 20 mM Tris HCl (pH 7.5), 150 mM NaCl, 10% glycerol, 1% NP40, 2 mM EDTA, 500 μM Na₃VO₄, 1% aprotinin, 10 μg/mL leupeptin, and 1 mM phenylmethyl sulfonylfluoride. For immunoprecipitation, lysates were incubated with goat anti-mouse VEGFR-2 (R&D) at 4°C followed by incubation with fast-flow protein-G-Sepharose (Amersham Biosciences), end over end, at 4°C. Proteins were released by boiling in a sample buffer [59 mM Tris-HCl (pH 6.8), 1.5% sodium dodecyl sulfate (SDS), 4.35% glycerol, 4% β-mercaptoethanol, 0.0025% bromophenol blue], and separated with SDS-PAGE. Receptor activation was detected with mouse anti-phospho-tyrosine (4G10; Cell Signaling) [28]. Immunoreactivity was visualized by enhanced chemiluminescence. Quantification of immunoreactive bands was made in ImageJ.

Statistical analysis

Each experiment consists of a minimum of 3 biological replicates. The statistical significance was determined using the Student's t-test. Values are significant at ***P<0.001, **P<0.01, and *P<0.05, as indicated in the figures. Error bars represent SEM.

Transcriptome analysis

The microarrays were performed in 3 biological replicates for each treatment. >50,000 CD31⁺ cells were sorted from each sample and total RNA extracted using the RNeasy Micro kit (Qiagen). About 300 ng of total RNA from each sample was reverse transcribed into cDNA and in vitro transcribed into biotin-labeled cRNA using the Illumina TotalPrep RNA Amplification kit (Ambion). About 750 ng of cRNA from each sample was hybridized to MouseRef-8 v2.0 Expression BeadChip microarrays (Illumina) and scanned on the BeadArray Reader (Illumina) at scan factor 1. Raw intensity values were subjected to a background subtraction performed on the BeadStudio Data Analysis software (Illumina) and normalized using the cross-correlation method [29]. Differentially expressed transcripts were identified based on the average of Log2-fold change in expression values compared to the normoxia control for each biological replicate experiment, with a minimum cutoff at 1.5-fold change. The microarray data were deposited into NCBI GEO with the accession number GSE35894 (www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=pxwxxcgkgaiikfq&acc=GSE35894). Gene Ontology (GO) analysis was generated using VLAD v1.5 (MGI).

Results

A significant portion of the hypoxia-induced transcriptome in CD31⁺ endothelial cells requires functional Notch signaling to be upregulated by hypoxia

To study how hypoxia and Notch signaling affect gene expression in endothelial cells, we took advantage of the fact that ES cells can be in vitro differentiated to vascular progenitors (angioblasts), which can be further matured toward arterial or venous fates [5,30]. Mouse ES cells were cultured as EBs for 4 days to generate Flk1⁺ vascular progenitors that were sorted to purity using flow cytometry. The Flk1⁺ cells were plated on collagen type IV-coated plates and differentiated with 50 ng/mL of human recombinant VEGF for 4 days to generate CD31⁺ endothelial cells that were sorted to purity and subjected to transcriptome analysis (Fig. 1A). To investigate the interaction between hypoxia and Notch signaling, the Flk1⁺ cells were cultured in normoxia or hypoxia (1.5%–2% O₂) for 4 days before the CD31⁺ sorting, combined with normal or inhibited Notch signaling. Inhibition of Notch signaling was accomplished by treating the cells with the γ-secretase inhibitor (L-685.485), which inhibits the proteolytic processing of the Notch receptor and thus the release of the Notch ICD following ligand activation. The microarray data have been deposited in NCBI GEO with the accession number GSE35894 (www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=pxwxxcgkgaiikfq&acc=GSE35894). Transcriptome analysis showed that 840 genes were regulated >1.5-fold by hypoxia (Fig. 1B). Functional Notch signaling regulated 419 genes in hypoxia and 177 in normoxia (Fig. 1B). About 161 genes were coregulated by both hypoxia and functional Notch signaling.

FIG. 1.

A significant portion of the hypoxia-induced transcriptome in CD31⁺ endothelial cells requires functional Notch signaling. (A) Mouse embryonic stem (ES) cells were differentiated as embryoid bodies for 4 days and Flk1⁺ vascular progenitors were sorted to purity by flow cytometry. Flk1⁺ cells were further differentiated for 4 days with 50 ng/mL vascular endothelial growth factor (VEGF) in normoxia or hypoxia, with or without Notch inhibition [+/− gamma-secretase inhibitor (GSI)]. The CD31⁺ endothelial cells were sorted to purity before transcriptional analysis. (B) Venn diagrams showing overlap between hypoxia- and Notch-regulated genes identified by microarray analysis. (C) Heat map showing hypoxic induction of several known arterial (in red), but not venous (in blue) markers. There is also a reduction in genes involved in inflammation and leukocyte adhesion (in green). FCS, fetal calf serum.

GO analysis identified an enriched number of genes associated with metabolic processes in the hypoxia-regulated only group, mitotic processes in the Notch regulated in normoxia only group, and vascular processes such as patterning of blood vessels in the dataset of 161 genes regulated by hypoxia in a Notch-dependent manner (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/scd). The hypoxia- and Notch-regulated group was also highly enriched in arterial genes, such as Connexin40 (Gja5), Depp, vWF, Angpt2, Dll4, Cxcr4, and Sema3g (Fig. 1C, Supplementary Table S2) [31 –36]. In contrast, the venous marker EphB4 was not induced by hypoxia, and COUP-TFII (Nr2f2) showed the opposite pattern of expression; that is, it was reduced by hypoxia in a Notch-dependent manner (Fig. 1C). A hallmark of functional venous endothelium is high leukocyte adhesiveness in response to inflammatory stimuli [37]. Intriguingly, several genes involved in inflammation and leukocyte adhesion, such as Gucy1b3, Jam2, and Spon2, are downregulated in response to hypoxia (Fig. 1C) [38 –40]. These data indicate that a considerable portion of the hypoxia-induced genes, including a set of arterial differentiation marker genes, depends on functional Notch signaling, and that the Notch-dependent transcriptomes during normoxia and hypoxia are quite distinct.

Hypoxia-promoted arterial differentiation does not require endogenous VEGF

Hypoxia induces VEGF expression during angiogenesis, and VEGF acts upstream of Notch signaling during the early stages of vascular differentiation [9]. We have previously shown that not only is exogenous VEGF critical for endothelial differentiation, but also titrated VEGF dosage controls the arterial–venous differentiation in this system [5]. In these studies, we established that 2–10 ng/mL exogenous VEGF drives the venous endothelial differentiation, whereas 50 ng/mL promotes the arterial differentiation. Since the in vitro differentiation in the current study was already performed with arterial-promoting levels of VEGF (50 ng/mL), it was unlikely that hypoxia-induced VEGF could further drive the observed arterial response. To functionally eliminate effects of endogenous hypoxia-induced VEGF, we examined the endothelial differentiation of Vegf⁻ ^/− ES cells [24]. These cells were differentiated with 50 ng/mL exogenous VEGF in normoxic or hypoxic conditions. Expression of the Dll4, Notch4, Connexin 40, and EphrinB2 genes was induced by hypoxia to a similar extent in Vegf⁻ ^/− and wild-type ES cells, whereas Hey1 was activated to a more moderate extent in the Vegf⁻ ^/− ES cells. Expression of EphB4 was not altered and COUP-TFII was downregulated in the CD31⁺ Vegf⁻ ^/− ES cells (Fig. 2), similar to what was observed in the VEGF wild-type cells. This indicates that the hypoxia-driven arterial response is not mediated by hypoxic induction of endogenous VEGF, but through an alternative mechanism.

FIG. 2.

Hypoxia-induced arterial differentiation does not require endogenous VEGF. Transcriptional response to hypoxic culture during endothelial differentiation of wild-type (WT) and Vegf^−/− ES cells. Graphs show real-time PCR quantification of mRNA expression. Values are expressed as transcript levels relative to that of the Actb housekeeping gene. Significant values from t-test: ***P<0.001, **P<0.01, and *P<0.05.

Hypoxia-induced expression of arterial genes, with the exception of Dll4, requires functional Notch signaling

We next addressed in more detail whether activation of the hypoxia-induced genes from the transcriptome analysis required functional Notch signaling and whether hypoxia activated the Notch signaling pathway. To first assess Notch activation, we transfected a Notch reporter construct, 12XCSL-luc, which contains multimerized binding sites for the DNA-binding protein CSL [41], into the Flk1⁺ cells 24 h after plating and analyzed the reporter activity 3 days later, in the CD31⁺ cells, after exposure to normoxic or hypoxic conditions. Hypoxic culture conditions robustly increased the reporter signal, as compared to normoxia (Fig. 3A).

FIG. 3.

Hypoxia-induced expression of arterial genes, with the exception of Dll4, requires functional Notch signaling. (A) Analysis of 12XCSL-luciferase reporter activity during normoxia and hypoxia in CD31⁺ endothelial cells after normalization to beta-gal activity (B) Analysis of mRNA expression levels, relative to Actb housekeeping, in normoxia and hypoxia or in hypoxia after addition of GSI, as indicated. (C) Analysis of mRNA expression levels in Csl^−/− cells in normoxia and hypoxia. (D) Analysis of mRNA expression levels in the Notch1ΔE^TetOn cells differentiated from the Flk1⁺ angioblast stage with doxycycline-activated Notch1ΔE expression. Significant values from t-test: ***P<0.001, **P<0.01, and *P<0.05.

The hypoxia-induced activation of the arterial genes Notch4, Hey1, Connexin40, and EphrinB2 was blocked by GSI (Fig. 3B). However, Dll4, a critical factor for angiogenesis [15,16], posed an exception, as hypoxia-induced Dll4 expression was not completely blocked by GSI. To test in a more stringent manner whether hypoxia could indeed upregulate Dll4 expression in the absence of Notch signaling, we analyzed the effect of hypoxia in Csl^−/− ES cells (Fig. 3C), which are unable to transduce canonical Notch signaling [18]. Both Flk1⁺ vascular progenitors and CD31⁺ endothelial cells were generated in similar numbers to wild-type cells (data not shown). Dll4 expression was upregulated by hypoxia in the Csl^−/− ES cells, as expected (Fig. 3C). Expression of the venous marker COUP-TFII was increased in the Csl^−/− ES cells, as compared to wild-type cells (Fig. 3C), which may suggest that CSL represses expression of COUP-TFII, in keeping with a previous report [42]. Together, these data indicate that hypoxia activates the immediate Notch signaling response and that hypoxia-induced expression of arterial genes, with the exception of Dll4, requires functional canonical Notch signaling.

Notch signaling can substitute for hypoxia in promotion of arterial gene expression

While Notch signaling was necessary for the hypoxia-induced upregulation of arterial markers with the exception of Dll4, we asked whether it was also sufficient to induce arterial gene expression under normoxic conditions. To activate Notch signaling in a temporally controlled manner, we used a recently developed ES cell line, Notch1ΔE^TetOn, in which a ligand-independent membrane-tethered constitutively active form of Notch, Notch1ΔE, is induced by doxycycline via the TetOn system [25,26]. Doxycycline stimulation of Flk1⁺ Notch1ΔE^TetOn cells, followed by CD31⁺ sorting, yielded a robust activation of Hey1 and Hey2 expression, as verification of Notch induction, and an increase in expression of the arterial markers Dll4 and Connexin40, whereas the venous markers EphB4 and COUP-TFII as well as Flk1 were downregulated (Fig. 3D). In conclusion, the data show that specific activation of Notch at the Flk1⁺ stage was sufficient to mimic the hypoxia-induced upregulation of several arterial genes.

Adrenomedullin is specifically upregulated by HIF and activates Dll4 expression in a Notch-independent manner

To gain further insights into how hypoxia regulates Dll4 expression, we first tested whether HIF transcription factors directly regulated Dll4 expression. To this end, we transfected normoxically stabilized versions of HIF-1α and 2α, HIF-1α-PPA and 2α-PPA, respectively [27], into the Flk1⁺ ES cells. In agreement with the published data, HIF activated expression not only of Flk1 and VEGF, but also of Hey1 and Hey2 [22]. Dll4, however, was not directly regulated by introduction of HIF-1α-PPA and 2α-PPA (Fig. 4A). This may indicate that some intermediate component is required for activation of Dll4, and as Notch signaling was dispensable for the hypoxia-induced Dll4 upregulation, we began the search for candidate genes involved in the Dll4 induction by examining the list of the 633 genes that were modulated by hypoxia in a Notch-independent manner (Fig. 1B, C). Within the top 20 most highly induced genes, we found Adrenomedullin, which encodes a vasoactive hormone that has recently been implicated in the arterial/venous specification [43]. Mice deficient for adrenomedullin or its putative receptors, the receptor activity-modifying protein 2 (RAMP2), and the calcitonin receptor-like receptor (Crlr), show blood and lymphatic vascular defects [43,44], and in zebrafish, loss of crlr leads to vascular abnormalities and a loss of arterial identity [45]. In the light of these reports, we decided to explore whether hypoxic adrenomedullin induction played a role in Dll4 activation. We found that HIF-1α-PPA, and more potently HIF-2α-PPA, upregulated transcription of Adrenomedullin in the CD31⁺ endothelial cells (Fig. 4A). Adrenomedullin was also induced by hypoxia and, importantly, this upregulation was not blocked by addition of GSI (Fig. 4B), indicating that the upregulation of adrenomedullin by hypoxia was Notch independent. Notch induction in Notch1ΔE^TetOn cells also did not alter Adrenomedullin expression (Supplementary Fig. S1).

FIG. 4.

Adrenomedullin is specifically upregulated by HIFs and activates Dll4 expression in a Notch-independent manner. (A) Analysis of gene expression following transfection of normoxically stabilized HIF-1α and HIF-2α (HIF-1α-PPA and 2α-PPA) into the Flk1⁺ ES cells. (B) Analysis of Adrenomedullin mRNA expression in normoxia, hypoxia, or hypoxia combined with GSI, as indicated. (C) Analysis of Adrenomedullin and Dll4 mRNA expression in normoxia and hypoxia, combined with adrenomedullin siRNA (siRNA), as indicated. (D) Analysis of Dll4 mRNA expression in differentiated wild-type (WT) or Csl^−/− ES cells, combined with addition of recombinant adrenomedullin (ADM), as indicated. (E) Analysis of 12XCSL-luciferase reporter activity in differentiated Flk1⁺ cells, in response to addition of adrenomedullin (ADM). Significant values from t-test: ***P<0.001, **P<0.01, and *P<0.05.

To address whether adrenomedullin serves as a link between hypoxia and Dll4 regulation, we knocked down Adrenomedullin expression using siRNA (Fig. 4C). Downregulation of Adrenomedullin mRNA levels resulted in reduced Dll4 expression (Fig. 4C). Conversely, addition of recombinant adrenomedullin during vascular differentiation led to a more than 4-fold increase in Dll4 expression (Fig. 4D). Other Notch ligands were not affected by adrenomedullin treatment (Supplementary Fig. S2). A similar increase in Dll4 expression by adrenomedullin was observed in CD31⁺ cells derived from Csl^−/− ES cells (Fig. 4D), providing further credence to the notion that Notch signaling is not required. Addition of adrenomedullin also increased the immediate Notch downstream signaling response (Fig. 4E), suggesting that the elevated Dll4 expression led to activated Notch signaling in the endothelial cells. Adrenomedullin has been suggested to act through transactivation of VEGFR2 in HUVECs [46], but we could not detect any activation of VEGFR2 following adrenomedullin stimulation in our system (Supplementary Fig. S3). In sum, these experiments show that the hypoxia target gene Adrenomedullin is necessary and sufficient to induce Dll4 expression as well as immediate downstream Notch signaling.

Discussion

The cellular hypoxic response and Notch signaling are 2 signaling cascades important for vascular development, and in this report we have addressed how they direct arterial versus venous differentiation in endothelial cells. We unravel a hypoxia/adrenomedullin/Dll4 axis, by which hypoxia leads to upregulation of Dll4 in an adrenomedullin-dependent manner, and show that hypoxia activates an arterial expression signature in a Notch-dependent manner. A model for how the hypoxia/adrenomedullin/Dll4 axis and the Notch-dependent hypoxic induction of arterial gene expression may be linked is presented in Fig. 5.

FIG. 5.

A model —for the interplay between hypoxia and Notch signaling in arterial differentiation. Hypoxia upregulates expression of adrenomedullin, which is necessary and sufficient for activation of Dll4 expression and Notch signaling. Notch signaling, in turn, positively feeds back by upregulating and sustaining Dll4 expression (gray arrows). Notch signaling is also required for hypoxia-mediated activation of arterial markers (black arrows).

That hypoxia can activate Dll4 expression and Notch signaling via adrenomedullin receives support from a number of observations. First, hypoxia was shown to upregulate adrenomedullin and as this also occurred after GSI treatment or in Csl⁻ ^/− cells, we conclude that this upregulation did not require Notch signaling. Inhibition of adrenomedullin by siRNA abrogated the hypoxia-induced upregulation of Dll4 expression, and conversely, addition of adrenomedullin led to upregulation of Dll4 expression and the Notch immediate downstream response during normoxia. This demonstrates that adrenomedullin is both necessary and sufficient for induction of Dll4 expression. Dll4 was, however, upregulated not only by hypoxia and adrenomedullin, but also by active Notch signaling (Fig. 5). Thus, specific expression of Notch1ΔE in the Flk1⁺ vascular progenitors during normoxia was sufficient to upregulate Dll4 expression in the CD31⁺ endothelial cells. This dual regulation of Dll4 by hypoxia and Notch signaling could serve to (i) first induce Dll4 expression by hypoxia in a Notch-independent manner and (ii) sustain Dll4 expression by the activated Notch signaling, which in turn maintains Notch signaling in neighboring cells, resulting in a Dll4-Notch signaling positive feed forward loop (Fig. 5). Once Notch signaling is established and sustained, it can serve to mediate the hypoxia-induced upregulation of the other arterial markers (Fig. 5; black arrows). This notion is based on our finding that functional Notch signaling was required for hypoxia to induce expression of Depp, Connexin40, Cxcr4, Hey1, and EphrinB2, in contrast to Dll4, which as discussed above was upregulated independently of the Notch signaling status. Specific upregulation of Notch signaling was also sufficient to induce expression of the arterial markers, indicating that Notch signaling was sufficient and necessary for arterial marker expression.

The model in Fig. 5 is also corroborated by several earlier reports. A critical role for hypoxia, adrenomedullin, and Dll4 is in line with phenotypes resulting from targeting of the respective genes [12,15,16,43 –45]. In all 3 cases, gene knockouts lead to embryonic death with a severely compromised vasculature, and interestingly, both HIF-1α ^−/−, Adrenomedullin^−/− and Dll4^+/− embryos fail to remodel the vasculature from an initially, apparently, and normally formed vascular plexus. Failure to remodel the vasculature is in keeping with the roles in arterial versus venous differentiation demonstrated here, and the fact that the loss of only one Dll4 allele is sufficient to cause the phenotype suggests that regulation of Dll4 levels is critical. The hypoxia/adrenomedullin/Dll4 axis is corroborated by reports demonstrating that hypoxia regulates adrenomedullin expression [47,48]. Furthermore, adrenomedullin has been shown to control Notch signaling [36], and hypoxic induction of Dll4 has also been observed in tumor angiogenesis [49]. Moreover, hypoxia resulted in upregulation of arterial markers and downregulation of COUP-TFII [22], further supporting that the arterial fate was promoted at the expense of venous differentiation by hypoxia.

In conclusion, the data presented in this report unravel the hypoxia/adrenomedullin/Dll4 axis as an important initiator of Notch signaling during vasculogenesis, and show that Notch signaling is important for sustaining Dll4 expression and is also required for mediating the hypoxia-mediated activation of a larger set of arterial genes. This provides new insights into the complex interplay between hypoxia and Notch signaling and on the function of these 2 signaling mechanisms in the early steps of the formation of a functional vascular system.

Footnotes

Acknowledgments

We thank Dr. Henry Yang for the use of the cross-correlation method for microarray data normalization. This study was supported by grants from the Swedish Medical Research Council, the Swedish Cancer Society, Knut och Alice Wallenbergs Stiftelse, and the EC projects EuroSystem and NotchIT (UL).

Author Disclosure Statement

The authors declare no conflict of interests.

References

Risau

, Flamme

. 1995. Vasculogenesis. Ann Rev Cell Dev Biol, 11:73–91.

Carmeliet

. 2003. Angiogenesis in health and disease. Nat Med, 9:653–660.

Atkins

, Jain

, Hamik

. 2011. Endothelial differentiation: molecular mechanisms of specification and heterogeneity. Arterioscler Thromb Vasc Biol, 31:1476–1484.

Coultas

, Nieuwenhuis

, Anderson

, Cabezas

, Nagy

, Henkelman

, Hui

, Rossant

. 2010. Hedgehog regulates distinct vascular patterning events through VEGF-dependent and -independent mechanisms. Blood, 116:653–660.

Lanner

, Sohl

, Farnebo

. 2007. Functional arterial and venous fate is determined by graded VEGF signaling and notch status during embryonic stem cell differentiation. Arterioscler Thromb Vasc Biol, 27:487–493.

Peterson

, Shaw

, Peterson

, Milan

, Zhong

, Schreiber

, MacRae

, Fishman

. 2004. Chemical suppression of a genetic mutation in a zebrafish model of aortic coarctation. Nat Biotechnol, 22:595–599.

Seo

, Fujita

, Nakano

, Kang

, Duarte

, Kume

. 2006. The forkhead transcription factors, Foxc1 and Foxc2, are required for arterial specification and lymphatic sprouting during vascular development. Dev Biol, 294:458–470.

Weinstein

, Stemple

, Driever

, Fishman

. 1995. Gridlock, a localized heritable vascular patterning defect in the zebrafish. Nat Med, 1:1143–1147.

Coultas

, Chawengsaksophak

, Rossant

. 2005. Endothelial cells and VEGF in vascular development. Nature, 438:937–945.

10.

Lendahl

, Lee

, Yang

, Poellinger

. 2009. Generating specificity and diversity in the transcriptional response to hypoxia. Nat Rev Genet, 10:821–832.

11.

Adelman

, Maltepe

, Simon

. 1999. Multilineage embryonic hematopoiesis requires hypoxic ARNT activity. Genes Dev, 13:2478–2483.

12.

Iyer

, Kotch

, Agani

, Leung

, Laughner

, Wenger

, Gassmann

, Gearhart

, Lawler

, Yu

, Semenza

. 1998. Cellular and developmental control of O2 homeostasis by hypoxia-inducible factor 1 alpha. Genes Dev, 12:149–162.

13.

Ryan

, Lo

, Johnson

. 1998. HIF-1 alpha is required for solid tumor formation and embryonic vascularization. EMBO J, 17:3005–3015.

14.

Andersson

, Sandberg

, Lendahl

. 2011. Notch signaling: simplicity in design, versatility in function. Development, 138:3593–3612.

15.

Duarte

, Hirashima

, Benedito

, Trindade

, Diniz

, Bekman

, Costa

, Henrique

, Rossant

. 2004. Dosage-sensitive requirement for mouse Dll4 in artery development. Genes Dev, 18:2474–2478.

16.

Gale

, Dominguez

, Noguera

, Pan

, Hughes

, Valenzuela

, Murphy

, Adams

, Lin

et al. 2004. Haploinsufficiency of delta-like 4 ligand results in embryonic lethality due to major defects in arterial and vascular development. Proc Natl Acad Sci U S A, 101:15949–15954.

17.

Xue

, Gao

, Lindsell

, Norton

, Chang

, Hicks

, Gendron-Maguire

, Rand

, Weinmaster

, Gridley

. 1999. Embryonic lethality and vascular defects in mice lacking the Notch ligand Jagged1. Hum Mol Genet, 8:723–730.

18.

Oka

, Nakano

, Wakeham

, de la Pompa

, Mori

, Sakai

, Okazaki

, Kawaichi

, Shiota

, Mak

, Honjo

. 1995. Disruption of the mouse RBP-J kappa gene results in early embryonic death. Development, 121:3291–3301.

19.

Gustafsson

, Zheng

, Pereira

, Gradin

, Jin

, Lundkvist

, Ruas

, Poellinger

, Lendahl

, Bondesson

. 2005. Hypoxia requires notch signaling to maintain the undifferentiated cell state. Dev Cell, 9:617–628.

20.

Sahlgren

, Gustafsson

, Jin

, Poellinger

, Lendahl

. 2008. Notch signaling mediates hypoxia-induced tumor cell migration and invasion. Proc Natl Acad Sci U S A, 105:6392–6397.

21.

Coleman

, McDonough

, Hewitson

, Coles

, Mecinovic

, Edelmann

, Cook

, Cockman

, Lancaster

et al. 2007. Asparaginyl hydroxylation of the Notch ankyrin repeat domain by factor inhibiting hypoxia-inducible factor. J Biol Chem, 282:24027–24038.

22.

Diez

, Fischer

, Winkler

, Hu

, Hatzopoulos

, Breier

, Gessler

. 2007. Hypoxia-mediated activation of Dll4-Notch-Hey2 signaling in endothelial progenitor cells and adoption of arterial cell fate. Exp Cell Res, 313:1–9.

23.

Dong

, Wang

, Dou

, Hou

, Shi

, Zhang

, Ma

, Wu

, Yao

, Cai

, Zhang

. 2011. Influence of Dll4 via HIF-1alpha-VEGF signaling on the angiogenesis of choroidal neovascularization under hypoxic conditions. PLoS One, 6:e18481.

24.

Carmeliet

, Ferreira

, Breier

, Pollefeyt

, Kieckens

, Gertsenstein

, Fahrig

, Vandenhoeck

, Harpal

et al. 1996. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature, 380:435–439.

25.

Das

, Lanner

, Main

, Andersson

, Bergmann

, Sahlgren

, Heldring

, Hermanson

, Hansson

, Lendahl

. 2010. Notch induces cyclin-D1-dependent proliferation during a specific temporal window of neural differentiation in ES cells. Dev Biol, 348:153–166.

26.

Main

, Lee

, Yang

, Haapa-Paananen

, Edgren

, Jin

, Sahlgren

, Kallioniemi

, Poellinger

, Lim

, Lendahl

. 2010. Interactions between Notch- and hypoxia-induced transcriptomes in embryonic stem cells. Exp Cell Res, 316:1610–1624.

27.

Pajusola

, Kunnapuu

, Vuorikoski

, Soronen

, Andre

, Pereira

, Korpisalo

, Yla-Herttuala

, Poellinger

, Alitalo

. 2005. Stabilized HIF-1 alpha is superior to VEGF for angiogenesis in skeletal muscle via adeno-associated virus gene transfer. FASEB J, 19:1365.

28.

Jakobsson

, Kreuger

, Holmborn

, Lundin

, Eriksson

, Kjellen

, Claesson-Welsh

. 2006. Heparan sulfate in trans potentiates VEGFR-mediated angiogenesis. Dev Cell, 10:625–634.

29.

Chua

, Vijayakumar

, Nissom

, Yam

, Wong

, Yang

. 2006. A novel normalization method for effective removal of systematic variation in microarray data. Nucleic Acids Res, 34:e38.

30.

, Edholm

, Lanner

, Breier

, Farnebo

, Dimberg

, Claesson-Welsh

. 2007. Lentiviral rescue of vascular endothelial growth factor receptor-2 expression in flk1^−/− embryonic stem cells shows early priming of endothelial precursors. Stem Cells, 25:2987–2995.

31.

Yeh

, Dupont

, Coppen

, Rothery

, Severs

. 1997. Gap junction localization and connexin expression in cytochemically identified endothelial cells of arterial tissue. J Histochem Cytochem, 45:539–550.

32.

Shin

, Anderson

. 2005. Isolation of arterial-specific genes by subtractive hybridization reveals molecular heterogeneity among arterial endothelial cells. Dev Dyn, 233:1589–1604.

33.

Krebs

, Xue

, Norton

, Shutter

, Maguire

, Sundberg

, Gallahan

, Closson

, Kitajewski

et al. 2000. Notch signaling is essential for vascular morphogenesis in mice. Genes Dev, 14:1343–1352.

34.

, Drouet

, Carrier

, Rothschild

, Berard

, Rouault

, Caen

, Meyer

. 1987. Differential distribution of von Willebrand factor in endothelial cells. Comparison between normal pigs and pigs with von Willebrand disease. Arteriosclerosis, 7:47–54.

35.

Mecha Disassa

, Styp-Rekowska

, Hinz

, Da Silva-Azevedo

, Pries

, Zakrzewicz

. 2009. Differential expression of VEGFA, TIE2, and ANG2 but not ADAMTS1 in rat mesenteric microvascular arteries and veins. Physiol Res, 58:193–202.

36.

Yurugi-Kobayashi

, Itoh

, Schroeder

, Nakano

, Narazaki

, Kita

, Yanagi

, Hiraoka-Kanie

, Inoue

et al. 2006. Adrenomedullin/cyclic AMP pathway induces Notch activation and differentiation of arterial endothelial cells from vascular progenitors. Arterioscler Thromb Vasc Biol, 26:1977–1984.

37.

Kalogeris

, Kevil

, Laroux

, Coe

, Phifer

, Alexander

. 1999. Differential monocyte adhesion and adhesion molecule expression in venous and arterial endothelial cells. Am J Physiol, 276:L9–L19.

38.

Zhang

, Xiang

, Dong

, Skoda

, Daugherty

, Smyth

, Du

, Li

. 2011. Biphasic roles for soluble guanylyl cyclase (sGC) in platelet activation. Blood, 118:3670–3679.

39.

Johnson-Leger

, Aurrand-Lions

, Beltraminelli

, Fasel

, Imhof

. 2002. Junctional adhesion molecule-2 (JAM-2) promotes lymphocyte transendothelial migration. Blood, 100:2479–2486.

40.

Jia

, Li

, He

. 2005. The extracellular matrix protein mindin serves as an integrin ligand and is critical for inflammatory cell recruitment. Blood, 106:3854–3859.

41.

Kato

, Taniguchi

, Kurooka

, Minoguchi

, Sakai

, Nomura-Okazaki

, Tamura

, Honjo

. 1997. Involvement of RBP-J in biological functions of mouse Notch1 and its derivatives. Development, 124:4133–4141.

42.

You

, Lin

, Lee

, DeMayo

, Tsai

. 2005. Suppression of Notch signalling by the COUP-TFII transcription factor regulates vein identity. Nature, 435:98–104.

43.

Caron

, Smithies

. 2001. Extreme hydrops fetalis and cardiovascular abnormalities in mice lacking a functional Adrenomedullin gene. Proc Natl Acad Sci U S A, 98:615–619.

44.

Dackor

, Fritz-Six

, Dunworth

, Gibbons

, Smithies

, Caron

. 2006. Hydrops fetalis, cardiovascular defects, and embryonic lethality in mice lacking the calcitonin receptor-like receptor gene. Mol Cell Biol, 26:2511–2518.

45.

Nicoli

, Tobia

, Gualandi

, De Sena

, Presta

. 2008. Calcitonin receptor-like receptor guides arterial differentiation in zebrafish. Blood, 111:4965–4972.

46.

Guidolin

, Albertin

, Spinazzi

, Sorato

, Mascarin

, Cavallo

, Antonello

, Ribatti

. 2008. Adrenomedullin stimulates angiogenic response in cultured human vascular endothelial cells: involvement of the vascular endothelial growth factor receptor 2. Peptides, 29:2013–2023.

47.

Nakayama

, Takahashi

, Murakami

, Shirato

, Shibahara

. 1998. Induction of adrenomedullin by hypoxia and cobalt chloride in human colorectal carcinoma cells. Biochem Biophys Res Commun, 243:514–517.

48.

Cormier-Regard

, Nguyen

, Claycomb

. 1998. Adrenomedullin gene expression is developmentally regulated and induced by hypoxia in rat ventricular cardiac myocytes. J Biol Chem, 273:17787–17792.

49.

, Sainson

, Oon

, Turley

, Leek

, Sheldon

, Bridges

, Shi

, Snell

et al. 2011. DLL4-Notch signaling mediates tumor resistance to anti-VEGF therapy in vivo. Cancer Res, 71:6073–6083.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.08 MB

0.06 MB

0.05 MB

0.27 MB

Hypoxia-Induced Arterial Differentiation Requires Adrenomedullin and Notch Signaling

Abstract

Introduction

Materials and Methods

Embryonic stem cell culture and differentiation

Flow cytometry

Real-time PCR

Transfections and luciferase reporter assay

siRNA

Immunoblotting

Statistical analysis

Transcriptome analysis

Results

A significant portion of the hypoxia-induced transcriptome in CD31+ endothelial cells requires functional Notch signaling to be upregulated by hypoxia

Hypoxia-promoted arterial differentiation does not require endogenous VEGF

Hypoxia-induced expression of arterial genes, with the exception of Dll4, requires functional Notch signaling

Notch signaling can substitute for hypoxia in promotion of arterial gene expression

Adrenomedullin is specifically upregulated by HIF and activates Dll4 expression in a Notch-independent manner

Discussion

Footnotes

Acknowledgments

Author Disclosure Statement

References

Supplementary Material

A significant portion of the hypoxia-induced transcriptome in CD31⁺ endothelial cells requires functional Notch signaling to be upregulated by hypoxia