Notch Pathways Regulate Expression of Angiostatic Factor Activin A in Endothelial–Pericyte-Like Mesenchymal Stromal Cell Interactions

Abstract

Vascularization of ischemic and fabricated tissues is essential for successful tissue repair and replacement therapies. Endothelial cells (ECs) and mesenchymal stem/stromal cells (MSCs) in close proximity spontaneously organize into vessels after coimplantation in semisolid matrices. Thus, local injection of EC mixed with MSC may facilitate tissue (re)vascularization. The organization of these cells into vessels is accompanied by induction of a key regulator of vasculogenesis, activin A, in MSC through juxtacrine pathway. Mechanisms regulating activin A expression are poorly understood; therefore, the contributions of notch signaling pathways were evaluated in EC-adipose mesenchymal stromal cells (ASC) cocultures. Disruption of notch signaling in EC + ASC cocultures with a γ-secretase inhibitor, DAPT, completely abrogated both activin A induction and production, depending on the stage of vasculogenesis. While DAPT stimulated EC proliferation concurrent with increased secretion of vasculogenic factors, it also prevented the crucial transition of ASC from progenitor to smooth muscle cell phenotype, collectively resulting in ineffective tubulogenesis. Silencing Notch2 in ASC abolished activin A production in cocultures, but resulted in normal ASC maturation. In contrast, silencing Notch3 in ASC led to autonomous upregulation of mural cell markers, and intercellular contact with EC further enhanced upregulation of these markers, concurrent with amplified activin A secretion. Strong induction of activin A expression was achieved by exposing ASC to immobilized notch ligand jagged1, whereas jagged1 IgG, added to EC + ASC incubation media, prevented activin A expression. Overall, this study revealed that EC control activin A expression in ASC through trans juxtacrine notch signaling pathways, and uninterrupted notch signaling is required for activin A production, although signaling through Notch2 and Notch3 produce opposing effects.

Introduction

The human and economic costs arising from poor recovery of ischemic tissues and wound healing are enormous. Based upon recent analyses, the incidence of heart failure among U.S. adults ages 55 or older is 1 million [1], more than 230 million people worldwide live with peripheral vascular diseases [2], and 1%–2% of the general population in developed countries have chronic nonhealing wounds [3,4].

Clinical trials with defined angiogenic factors (e.g., vascular endothelial growth factor, hepatocyte growth factor, and basic fibroblast growth factors) and progenitor cells [e.g., mesenchymal stem/stromal cells (MSCs) from bone marrow and adipose tissue, cardiac progenitor cells, blood derived CD34+ cells] have been conducted attempting to treat these conditions; however, these well-controlled clinical trials had unimpressive results [5 –7]. Failure to achieve therapeutic effects is partially due to unsuccessful stimulation of the angiogenic process. However, limited understanding of vessel homeostasis and remodeling, especially in ischemic and inflammatory situations, prevents the design of therapies that effectively target the vasculature. This study investigated the complex temporal regulation of activin A, a factor that promotes formation of mature vessels, by notch signaling using an in vitro model based on coculturing endothelial cells (ECs) and adipose mesenchymal stromal cells (ASCs).

ASCs behave phenotypically and functionally as pericyte/smooth muscle cell progenitors [8]. The processes required for ASC isolation and in vitro expansion result in activation of a vasculo/angiogenic program in these cells, which translates to secretion of cytokines and growth factors that stimulate EC survival, proliferation, migration, and tubulogenesis in vitro and in vivo [8 –10].

Direct interactions of ECs + ASCs lead to spontaneous reorganization of ECs into networks of stable, connected vascular cords in vitro and mature vessels in vivo [9,10]. Concomitantly, the ASCs differentiate from progenitor cells into smooth muscle cells. Cord formation is accompanied by downregulation of the proangiogenic secretory program of ASCs, which facilitates stabilization of vascular structures [11]. Essential extracellular matrix proteins of the basement membrane of vessel walls are deposited around the cords [10,12]. In animal models, implantation of semisolid matrices containing both ECs and ASCs (or other MSCs), but neither cell type alone, generates multilayered functional vessels which are well-integrated with the host vasculature [9,13]. Interactions of endothelial and mural cells (pericytes and smooth muscle cells) guiding these complementary cells through the process of vasculogenesis depend upon autocrine, paracrine, and juxtacrine signaling, such as through connexins, eph–ephrin, and notch signaling pathways [14].

Our prior studies showed that vasculogenesis executed by ECs and ASCs is accompanied by substantial induction of activin A in ASCs [12]. Activin A is a multifunctional factor which plays an essential role in numerous physiological and pathological processes. Activin A acts as a dimer, composed of the product of the gene inhibin BA. It signals through a complex of four proteins; two type 1 (ALK4) and two type 2 receptors (either ACVRIIA or IIB) that activate a SMAD phosphorylation cascade to affect gene transcription and eventually phenomena such as cell cycling and cellular differentiation. While activin A promotes neuronal recovery following a brain insult [15,16], it also contributes to the development of pathology, as seen in acute kidney injury [17] and inflammatory processes in lungs following exposure to LPS [18] or cigarette smoke [19].

Despite multiple observations of activin A's role in many physiological processes, its role in vasculogenesis was not recognized until recently. Activin A, in a paracrine manner, shifts the ASC secretome from predominantly proangiogenic to angiostatic [11] and promotes expression of smooth muscle cell markers in ASC [12]. In EC, the primary cells of the vessel wall, activin A upregulates membrane vascular endothelial growth factor (VEGF) receptor 1 (mFLT1) and its soluble form (sFLT1), which function as scavenger receptors for VEGF to limit undesirable overstimulation with VEGF [11]. Collectively, these findings suggest that the presence of activin A in the active vasculogenic niche helps ensure formation of stable, mature multilayered vessels.

We have reported that activin A is induced in ASCs by ECs through juxtacrine mechanisms [12]; however, the precise pathway has not been identified. Some clues point to the possible involvement of the notch signaling pathway. Studies manipulating the mouse genome revealed a critical role of notch signaling in vasculogenesis/angiogenesis during development. Specifically, the notch family regulates endothelial cell tip/stalk specification in sprouting angiogenesis, arterial/venous cell specialization [20], and maintenance of EC quiescence in established vessels [21].

Brenda Lilly's group characterized the role of Notch2 and Notch3 in regulation of the differentiation of pericytes to smooth muscle cells and the central role of EC-expressed jagged1 in this process [22]. Notch signaling pathways are multifaceted; humans have four receptors (Notch 1–4) and multiple ligands, including jagged1 and 2 and DLL1-DLL4, which through both cis and trans interactions [23] lead to different and, in some cases, opposing cellular responses [24]. While ECs are characterized by expression of Notch 1 [25] and 4 [26], mural and MSCs express Notch2 and 3 [27 –29]. Collectively, these data led us to hypothesize that notch signaling regulates activin A induction. Thus, this study tested whether ASC-expressed Notch2 and Notch3 signaling was involved in activin A induction and affected vasculogenesis.

Materials and Methods

Procedures for collecting umbilical cord blood and adipose tissue were approved by the Indiana University School of Medicine Institutional Review Board. Human ASC and cord blood derived endothelial cells (CBD-ECs) were isolated and expanded as previously described [8,30]. ASCs were isolated from subcutaneous adipose tissue of healthy adult female donors undergoing elective liposuction procedures. Samples were obtained from eight patients, ages 27–41 (34.8 ± 1.3) with body mass index 20–26 (23.5 ± 0.9). To eliminate patient-dependent effects, all experiments were performed with at least two randomly selected ASC samples. CBD-ECs were isolated from the umbilical cord vein blood of healthy newborn females (38–40 weeks gestational age). ASCs, expanded in endothelial growth media (EGM-2mv) (Lonza), and CBD-ECs, expanded in EGM/10% fetal bovine serum (FBS; Lonza), were used at passages 4–6.

Coculture of EC and ASC

A previously described two-dimensional EC + ASC coculture model was used [10]. Unless specified, 6 × 10⁴ cells/cm² of ASCs and 1 × 10⁴ cells/cm² of CBD-ECs were premixed, plated, and incubated in endothelial basal media (EBM)-2/5%FBS for up to 22 days. In parallel, ASCs were plated alone and incubated under the same conditions. In subset of tests, ECs and ASCs were cocultured in the same wells, but physically separated by 0.4-μm pore membrane Transwell (Costar). Such setup allows for paracrine, but not direct contact interactions between EC and ASC.

To abrogate notch signaling, cocultures were exposed to 10 μM DAPT (Selleckchem) [31,32] or dimethyl sulfoxide (DMSO) for 2–4 days at different points of incubation. To test role of jagged1 in activin A induction, coculture incubation media was augmented with neutralizing jagged1 IgG or mouse isotype control (both 10 μg/mL; RnD Systems).

EC + ASC interactions were also evaluated in a three-dimensional model, where 6 × 10⁵ DsRed-expressing ECs and 1.5 × 10⁵ ASCs were resuspended in 60 μL fibrinogen (final concentration 5 mg/mL) and added to the wells of 96-well plates containing 3 μL of 10 U/mL thrombin. Polymerized gels were overlaid with EBM-2/5%FBS. Next day, media was exchanged to fresh media supplemented with either 20 μM DAPT or DMSO followed by incubation for 96 h. At the end of incubation, media was collected for analysis, and images of the networks were acquired using fluorescent microscopy.

Immunohistochemical evaluation of cultures

To reveal vascular networks, cocultures were fixed and probed with either biotinylated Ulex Europaeus Agglutinin I (Vector labs), followed by Streptavidin Alexa 488 (Invitrogen) or mouse anti-human CD31 (ThermoFisher Scientific), followed by anti-mouse Alexa 488 (Invitrogen). Serial digital images (covering 30% of the total well surface) of vascular networks were acquired with Nikon microscope and processed with the “Angiogenesis Tube Assay” algorithm of MetaMorph software (Molecular Devices). ASC monocultures and EC + ASC cocultures were also stained with alpha smooth muscle actin mouse IgG (Sigma), Ki67 (Cell Signaling), Inhibin Ba (GeneTex) rabbit IgG, or isotype control IgGs for 1 h, followed by incubation with anti-mouse or anti-rabbit IgG tagged with either Alexa 594 or Alexa 488 (Invitrogen) for 30 min. The nuclei were counterstained with 4′,6-diamidino-2-phenylindole.

Conditioned media generation and analysis

Conditioned media (CM) were generated by exposing ASC mono- or EC + ASC cocultures to EBM-2/5%FBS media for 48 h at different time points of incubation and then evaluated for activin A, hepatocyte growth factor (HGF), stromal cell-derived factor 1 (SDF1), and VEGF by enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN, USA).

Analysis of ASC response to immobilized jagged1

jagged1 was immobilized on high protein binding 96-well plates following the established protocol with minor modifications [33]. Plates were coated with 10 μg/mL donkey anti-human Fc IgG (Jackson Labs) overnight, followed by sequential incubation with 1% bovine serum albumin (1 h) and 5 μg/mL human jagged1-Fc (RnD Systems) for 2 h. A total of 20,000 ASCs were plated into the wells with immobilized jagged1 or wells treated just with anti-human Fc IgG (negative control). As a positive control, EC + ASC cocultures were plated in untreated wells. After 48 h of incubation, CM were collected and evaluated for activin A.

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and immunoblotting analysis

ASC and EC + ASC cultures were harvested with 1% sodium dodecyl sulfate (SDS)/ lysis buffer containing a cocktail of protease inhibitors. Proteins were fractionated on 4%–20% SDS–polyacrylamide gels by electrophoresis and transferred onto 0.45 μm nitrocellulose membranes. Membranes were incubated with anti-Notch2, Notch3 (Cell Signaling), β-tubulin (BioLegend), αSMA (Sigma), SM22α, and calponin (ThermoFisher Scientific) IgG, followed by incubation with peroxidase–conjugated anti-mouse or anti-rabbit IgGs (Cell Signaling) and visualized with SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher Scientific) on Molecular Imager Gel Doc XR+ System (Bio-Rad). Alternatively, membranes were incubated with IRDye^® 800CW anti-mouse or anti-rabbit IgG, imaged on Odyssey CLx imaging system, and intensities of bands were quantitatively assessed using Image Studio software (all from Li-Cor).

mRNA expression analyses

Total RNA was isolated from ASC and EC + ASC cultures, followed by a reverse transcription reaction. Polymerase chain reactions (PCRs) were performed using iTaq SYBR Green PCR Supermix with ROX (Bio-Rad) and oligonucleotide primers (Table 1) or TaqMan Fast Advance Master Mix with FLT-1 TaqMan Gene Expression Assay on StepOne PCR machine (all Invitrogen). Relative quantitation was performed against a standard curve constructed for each gene of interest, normalized to β-actin, and compared to control/untreated samples.

Table 1.

Sequences of Oligonucleotide Primers

	Forward	Reverse
Inhibin B_A	ggagaacgggtatgtggaga	aatctcgaagtgcagcgtct
HGF	gagagttgggttcttactgcacg	ctcatctcctcttccgtggaca
Notch2	cacagaggctgggaaaggatgata	ggccacctgaagggaagcacata
Notch3	gagccaatgccaactgaagag	ggcagatcaggtcggagatg
β-actin	caccattggcaatgagcggttc	aggtctttgcggatgtccacgt
SDF-1	gcccgtcagcctgagctaca	ttcttcagccgggctacaatct
VEGF	ttgccttgctgctctacctcca	atggcagtagctgcgctgata

Cell transfection with siRNA constructs

ASCs were plated at 4–5 × 10⁴ cells/cm² in EBM-2/5%FBS. Three hours later, ASCs were transfected with silencing RNA constructs (siRNA) for Notch2 or Notch3 or with scrambled RNA (scRNA), as a control for transfection manipulations, using Lipofectamine RNAiMAX reagent (all from Invitrogen). Cells were used for experiments 24 h later.

Analysis of cell surface markers by flow cytometry technique

ASC monocultures were harvested with 2 mM ethylenediaminetetraacetic acid, incubated with Notch2-APC (BioLegend) or Notch3-APC (eBioscience) IgG for 20 min on ice, and then analyzed on BD Accuri analyzer (BD).

Statistical analysis

Quantitative data are expressed as mean ± standard error of the mean. Statistical analysis of the data that include only two experimental groups was performed with an unpaired nonparametric Mann–Whitney test. If one group was used for normalization and expressed as “1” or “100%,” then one sample t-test was performed. For the data that had at least three groups Shapiro–Wilk test was performed to confirm normal distribution, followed by one-way analysis of variance with Tukey's multiple comparisons. If the data did not pass normal distribution test, data were analyzed with Kruskal–Wallis test. Statistical analysis was performed using Prism 9 (GraphPad).

Results

Dynamic of activin A induction in EC + ASC cocultures

As we have previously reported [10] and as shown in Fig. 1, incubation of ECs on ASC monolayers leads to spontaneous organization of ECs into networks of vascular cords and cord density increased dose-dependently with the plating density of ECs (Fig. 1A, B). Concomitantly, activin A increased in the incubation media, also dose-dependently with EC plating density (Fig. 1C). The induction of activin A was dependent upon direct contact between ECs and ASCs, since incubation of both cell types in the same well, but separated by 0.4 μm-pore membrane that prevented heterotypic cell contact while permitting humoral exchange, eliminated production of activin A (Fig. 1D). This was further supported by analysis of immunofluorescent images of EC + ASC cocultures probed for Inhibin Ba at day 5 of incubation: a strong signal for Inhibin Ba was detected in ASCs which were either in direct contact or in close proximity to EC cords (Fig. 1E).

FIG. 1.

2D and 3D cocultures of ECs and ASCs generate vascular cords and activin A. (A) Representative fluorescent images of ASC monoculture and EC + ASC coculture (5000 EC/cm² on ASC monolayer) probed for EC marker Ulex lectin (green) at day 5 of incubation. (B, C) Correlation between EC density in EC + ASC cocultures and either vessel density (B) or level of activin A in 48-h conditioned media (C) assessed at day 5 of incubation (n = 8–9). (D) Activin A level in the media conditioned by EC + ASC cocultures for the last 72 h of 6-day incubation when ECs and ASCs were incubated either with direct contact (Dir) or physically separated by a membrane with 0.4 μm pore size (TW) (n = 11–12). (E) Representative fluorescent images of ASC monocultures and EC + ASC cocultures probed for Inhibin Ba (InhBa, red) and then stained with Ulex Lectin (green) to reveal ECs and with DAPI to reveal nuclei (blue). Control EC + ASC cocultures were incubated with nonimmunogenic rabbit IgG. (F) Representative fluorescent images of vascular cords established by DsRed-expressing EC (white) when cultured alone or cocultured with ASCs in 3D fibrin gel. Images were taken on day 6 of incubation. (G) Accumulation of activin A in media conditioned for 48 h by ECs, or ASCs, or ECs + ASCs in fibrin gel assessed on day 3 of incubation. (n = 8). For all graphs: **P ≤ 0.01, ***P ≤ 0.01. ASC, adipose stromal cell; DAPI, 4′,6-diamidino-2-phenylindole; EC, endothelial cell.

Observations made with two-dimensional vasculogenesis model were further supported by the use of a complementary model where ECs and ASCs were embedded into fibrin matrices individually or in combination. ECs, embedded alone, did not form vascular cords (Fig. 1F) nor, as expected, did ASCs (data not shown), whereas mixtures of ECs and ASCs established dense vascular networks (Fig. 1F). Furthermore, media collected from EC + ASC cocultures in fibrin matrices contained high levels of activin A, but it was barely detectable in the media from monocultures. (Fig. 1G).

Effect of inhibition of γ-secretase activity on EC + ASC interaction

We have previously shown that in EC + ASC cocultures, the expression of activin A is induced only in ASCs and not in ECs, and juxtacrine interactions between ECs and ASCs are necessary for the induction [12]. To evaluate contributions of notch signaling to EC + ASC juxtacrine communication, cocultures were treated with a γ-secretase inhibitor, DAPT, between days 1 and 5 of incubation. DAPT inhibits notch signaling by preventing formation of intracellular notch signaling domains [34,35]. While incubation of cocultures in EBM-2/5%FBS alone or with vehicle (DMSO) led to EC organization into thin vascular cords, cocultures exposed to DAPT formed much thicker cords (Fig. 2A). Furthermore, DAPT abolished induction of the smooth muscle cell markers αSMA, SM22α, and calponin in ASCs which were observed in untreated cocultures (Fig. 2A, B). This suggests that disruption of notch signaling prevented EC-mediated acquisition of a smooth muscle phenotype by ASCs.

FIG. 2.

Inhibition of notch signaling with DAPT thickens vascular cords and prevents ASC acquisition of smooth muscle cell markers and secretion of activin A. (A) Representative fluorescent images of ASCs and EC + ASC cocultures incubated in EBM-2/5%FBS alone or treated with DMSO or 10 μM DAPT for the last 4 of 5 days of incubation. Cultures were stained with Ulex lectin (green) for endothelial cells and probed for αSMA (red). Nuclei were counterstained with DAPI (blue). (B) Representative immunoblot of αSMA, SM22α, and Calponin expression in ASCs and EC + ASC cocultures after exposure to DMSO or 10 μM DAPT for the last 4 of 5-day incubation. Expression of β-tubulin was used as a control for protein loading. (C) Expression of inhibin Ba mRNA in EC + ASC cocultures treated with DMSO or 10 μM DAPT for the last 48 h of 3- or 5-day incubations (n = 5–6). (D) Accumulation of activin A in media conditioned by EC + ASC cocultures treated with DMSO or 10 μM DAPT during the last 48 h of 3 or 5 day incubations (n = 8). (E) Representative fluorescent images of vascular network formed by DsRed-expressing ECs (white) and ASCs in fibrin gel over 5 days while being treated with DMSO or 10 μM DAPT. (F) Comparative analysis of activin A in incubation media of EC + ASC cocultures embedded in fibrin gel and incubated in the presence of either DMSO or 10 μM DAPT. Media was conditioned for the last 96 h of 5 day incubation. (G) Comparative analysis of activin A in the media conditioned by cocultures composed of untreated ECs and ASCs pretreated with either DMSO or 10 μM DAPT. Media was conditioned between days 1 and 3 of incubation (n = 13–14). (H) Representative fluorescent images of vascular networks formed by intact ECs and ASCs pretreated with DMSO or 10 μM DAPT at day 6 of incubation. EC cords were revealed with Ulex Lectin (white). For all graphs: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.01. αSMA, alpha smooth muscle actin; DMSO, dimethyl sulfoxide; EBM, endothelial basal media; FBS, fetal bovine serum.

Expression of the mRNA for the monomers of activin A, inhibin Ba, in cocultures and secretion of activin A into incubation media were prevented when cells were exposed to DAPT for 3 days (measured in 48-h CM collected at day 3 of incubation; Fig. 2C, D). In reversal experiments, exposure of cocultures to DAPT at day 3 of incubation, when activin A expression is well-established, almost completely silenced inhibin Ba expression and activin A secretion (Fig. 2C, D). Similar findings were observed when DAPT was introduced to fibrin gels encapsulating a mixture of ECs and ASCs a day after plating: control cocultures developed a network of dense interconnected EC cords, whereas DAPT caused formation of short, thick disconnected cords; concurrent with low, if any, activin A accumulation in the media (Fig. 2E, F). To prove that lack of activin A secretion in cocultures was due to disruption of notch signaling specifically in ASCs, these cells were pretreated with DAPT before coculturing with ECs. While ASC pretreatment did not affect EC tubulogenesis, media from these cocultures contained 70% less activin A (Fig. 2G, H).

Since prior studies suggest that activin A possesses angiostatic activities [11,36], the hypothesis that inhibition of activin A secretion in EC + ASC cocultures leads to EC proliferation was tested. EC + ASC cocultures were exposed to DAPT or DMSO for 4 days and then probed for the marker of proliferating cells, ki67. Ki67-positive ECs (Ulex-lectin positive cells) were substantially increased in DAPT-treated cocultures compared to control or DMSO-treated cultures (Cntr: 8.6 ± 3.0, DMSO: 8.8 ± 1.8, and DAPT: 72.6 ± 13.6 (ki67+EC/HPF); Fig. 3A). Complementary experiment with exposing semiconfluent EC monolayers to DAPT has clarified that DAPT did not have direct mitogenic effect on ECs (Fig. 3B).

FIG. 3.

Inhibition of activin A secretion induces EC proliferation and VEGF and HGF secretion. (A) Representative fluorescent images and quantitative analysis of prevalence of ki67+ ECs in EC + ASC cocultures exposed to EBM-2/5%FBS alone or with either DMSO or 10 μM DAPT for the last 4 of a 5 day incubation and then stained with Ulex Lectin (green) and probed for proliferation marker ki67 (red). Nuclei were counterstained with DAPI (blue). (B) Mitogenic response of EC exposed to EBM-2/5%FBS alone or with either DMSO or 10 μM DAPT for 4 days [Cntr (n = 18), DMSO (n = 16), DAPT (n = 16), Gr_M (n = 4)]. (C) Analysis of VEGF and HGF accumulation in 48 h conditioned media from EC + ASC cocultures incubated in EBM-2/5%FBS alone or with DMSO or DAPT for the last 2 days of 5 day incubation (n = 3). (D) Relative FLT-1 mRNA expression in EC + ASC cocultures exposed to EBM-2/5%FBS alone or with either DMSO or 10 μM DAPT for the last 4 days of a 5 day incubation (n = 6). For all graphs: **P ≤ 0.01, ***P ≤ 0.01. FLT-1, fms related receptor tyrosine kinase 1; HGF, hepatocyte growth factor; VEGF, vascular endothelial growth factor.

An increase in EC proliferation in DAPT-treated cocultures was accompanied by substantially higher levels of vasculogenic factors such as VEGF and HGF in corresponding incubation media (Fig. 3C). Furthermore, our prior study revealed an increase in soluble and membrane bound FLT-1 in EC + ASC cocultures, and this was attributed to activin A activity [11]. FLT-1 is a scavenger receptor for VEGF and its activity may contribute to angiostatic effects of activin A. In this study, we observed that EC + ASC cocultures treated with DAPT showed 65% decrease in FLT-1 mRNA expression (Fig. 3D).

Dynamics of Notch2 and Notch3 expression

It has been reported that mural cells primarily express Notch2 and Notch3 [27 –29]. To extend these prior observations to ASCs, they were labeled with corresponding antibodies and evaluated by flow cytometry; the majority of ASCs (>95%) robustly expressed both receptors (Fig. 4A). In EC + ASC cocultures, a threefold increase in Notch3 mRNA expression was observed over 6 days in coculture, whereas the expression of Notch2 mRNA was unchanged (Fig. 4B). Similar dynamics for Notch2 and Notch3 protein expression was observed when evaluated by western blotting (Fig. 4C, D).

FIG. 4.

Notch3, but not Notch2, increases in ASCs in EC + ASC cocultures over time. (A) Flow cytometric analysis of Notch2 and Notch3 expression by ASCs. (B) Change in Notch2 and Notch3 mRNA expression in EC + ASC cocultures over 6 days of incubation (n = 8). (C, D) Representative immunoblot (C) and quantitative analysis (D) of Notch2 and Notch3 proteins in ECs + ASCs cocultured for 6 days. Expression of β-actin served as a control for normalization. Protein expression in each sample of ASCs on day one presented as 100% (n = 5). For all graphs: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.01.

Modulation of activin A expression in ASC by Notch2 and Notch3

To assess the roles of Notch2 (N2) and Notch3 (N3) in modulation of activin A expression, ASCs were transfected with N2 or N3 siRNA. Analysis of mRNA expression in transfected cells at day 1 revealed residual expression of N2 (12%) in N2-silenced and N3 (10%) in N3-silenced ASCs, whereas surface receptors were only 20% for N2 in N2-silenced and 30% for N3 in N3-silenced ASCs of those on unsilenced cells (Fig. 5A). Silencing N2 did not affect expression of N3 and vice versa (data not shown).

FIG. 5.

Notch 2 signaling stimulated by jagged 1 is necessary for activin A secretion, normal vascular tube formation, and tempering growth factor secretion. (A) Flow cytometric analysis of Notch2 and Notch3 expression in control ASCs and ASCs transfected with scRNA or siN2 or siN3 at day two post-transfection. (B) Change in Inhibin Ba mRNA expression (n = 6–7), and (C) activin A in incubation media [Cntr (n = 25), scRNA (n = 25), siN2 (n = 12), siN3 (n = 13)] of EC+ASC cocultures established by ECs with either control, or scRNA, or siN2, or siN3 ASCs. Analysis was performed on day 3 of incubation. (D) Change in activin A in the media conditioned by EC + ASC[siN3] cocultures in the presence of DMSO or 10 μM DAPT for the last 2 days of 3 day incubation (n = 5). (E) Representative fluorescent images of vascular cords established by DsRed-expressing ECs (white) cocultured with either scRNA, or siN2, or siN3 ASCs in fibrin gel by day 5 of incubation. (F) Change in activin A accumulation in incubation media of EC + ASC cocultures embedded in fibrin gel depending on ASC type used (scRNA, siN2, or siN3). Incubation media conditioned between days 3 and 5 was analyzed (n = 8–10). (G, H) HGF, SDF-1, and VEGF mRNA expression (G, n = 4–6) and protein accumulation (H, n = 5–16) in incubation media of cocultures established by intact ECs and various ASCs (scRNA, siN2, siN3). mRNA and media conditioned for 48 h were collected on day 5 of incubation. (I) Activin A accumulation in the media conditioned (24 h) by EC + ASC cocultures or by ASCs exposed to either anti-Fc fragment IgG (aFc) or jagged1-Fc chimera (JFc) (n = 8–10). (J) Change in activin A in media conditioned (72 h) by EC + ASC cocultures in EBM-2/5%FBS alone or supplemented with goat IgG (gIgG) or anti-jagged1 IgG (αJ1) (n = 9–10). For all graphs: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. scRNA, scrambled RNA; siN2, Notch2 silencing RNA; siN3, Notch3 silencing RNA.

Silencing of N2 in ASCs significantly diminished Inhibin Ba mRNA expression (by 60%; Fig. 5B) and prevented activin A secretion (Fig. 5C) in EC + ASC cocultures. In parallel, silencing of N3 in ASCs resulted in a 60% increase in Inhibin Ba mRNA expression and a 2.5-fold increase in activin A level in EC + ASC coculture media (Fig. 5B, C) by day 3 of incubation. No spontaneous secretion of activin A was observed in monocultures of N3-silenced ASCs (ASC^siN3; not shown). To confirm that upregulation of activin A in EC + ASC^siN3 was notch dependent, these cocultures were treated with DAPT for 3 days. Blocking notch signaling prevented activin A secretion (Fig. 5D). In the three-dimensional coculture model, EC co-embedded with N2-silenced ASCs formed multiple EC clusters and wide-diameter tubes (Fig. 5E), suggestive of pathological tubulogenesis; this was accompanied with only residual level of activin A in the media (Fig. 5F).

Analysis of the correlation between N2 and N3 silencing in ASCs and expression of provasculogenic factors in corresponding cocultures revealed that N2 silencing led to almost twofold increase in HGF mRNA and protein (Fig. 5G, H). A similar effect was observed for SDF-1 mRNA, but SDF-1 protein accumulation in the media was not as impressive. In contrast, VEGF expression was unaffected by N2 knockdown. Silencing of N3 in ASCs decreased HGF mRNA and protein, lowered SDF-1 mRNA, but had no effect on SDF-1 secretion, and substantially increased VEGF secretion in EC + ASC cocultures (Fig. 5G, H).

Role of jagged1 in induction of activin A expression

It has been reported that jagged1, which is expressed on EC, is a key “ligand” for notch receptors [37]. To test whether jagged1 modulates activin A expression, ASCs were exposed to jagged1-FC or control IgG immobilized to cell culture plastic. ASCs exposed to jagged1 strongly induced activin A expression, higher than observed in control EC + ASC cocultures (Fig. 5I). In complementary test, incubation of EC + ASC cocultures in the presence of jagged1 neutralizing IgG decreased accumulation of activin A in media by more than 80% compared to control cocultures (Fig. 5J).

Role of Notch2 and Notch3 in acquisition of smooth muscle cell phenotype by ASCs

We have previously reported [12] that ASCs exposed to ECs acquire smooth muscle phenotype; however, activin A, acting in paracrine manner, is only responsible for differentiation of ASCs that were not in direct contact with ECs, whereas EC + ASC heterotypic juxtacrine interaction is responsible for differentiation of ASC adjacent to vascular cords. To evaluate whether perturbations in N2 and N3 signaling would affect ASC differentiation, modified ASCs were incubated either alone or with ECs. Unexpectedly, silencing of N3 upregulated αSMA, SM22α, and calponin 1 expression in ASC monocultures (Fig. 6A, B), whereas silencing of N2 had no effect on ASC maturation (Fig. 6A). Furthermore, exposure of N3-silenced ASCs to ECs further increased expression of smooth muscle markers in ASCs (Fig. 6A). The increase in αSMA expression in ASC^siN3 was reduced by DAPT (Fig. 6C, D), confirming that upregulation of smooth muscle markers was notch dependent.

FIG. 6.

ASCs silenced for Notch 3 spontaneously acquired smooth muscle cell phenotype and this was enhanced by coculture with ECs. (A) Representative immunoblot of αSMA expression by control, scRNA, siN2, and siN3 ASC in monocultures or cocultured with ECs at day 5 of incubation. (B) Representative immunoblot of αSMA, SM22α, and calponin expression by control, scRNA, and siN3 ASCs at day 5 post-transfection. (C, D) Representative immunoblot (C) and quantitative analysis (D) of αSMA expression in ASCs transfected with scRNA or siN3 and treated 2 days later with DMSO or 10 μM DAPT for 48 h. *P ≤ 0.05. SM22α, smooth muscle 22α.

Discussion

Vasculogenesis, both developmental and reparative, depends upon coordinated communication between endothelial and mural cells. This cross talk is through multiple autocrine, paracrine, and juxtacrine mechanisms, involving mediators such as VEGF/VEGF-Rs, PDGF-BB/PDGF-Rβ, eph–ephrin, angiopoietin-tie2, connexins, and notch pathways. Many of these pathways are interconnected and modulate each other's expression.

Members of the TGFβ/activin A superfamily play an important role in vasculogenesis by inducing differentiation of MSCs adjacent to nascent vessels to pericytes and smooth muscle cells [12,38]. We have previously reported that mural cell-derived activin A plays a central role in EC-mural cell interactions promoting formation of stable vascular structures. Activin A by modulating the phenotype of both ASCs (a subtype of mural progenitor cells) and ECs plays a key role in organizing ECs into stable vessels and acquisition of a smooth muscle phenotype by ASCs [11,12].

Interestingly, several factors that promote smooth muscle cell differentiation, including TGFβ [39,40], thrombin [41], and angiotensin II [42], also induce activin A expression in these cells [12,43], suggesting that activin A might be the final common pathway these factors use to drive the differentiation of progenitors to SMC [43]. Our prior studies revealed that secretion of activin A by ASCs is induced by ECs through direct cell–cell interactions; media conditioned by EC + ASC cocultures in the absence of direct contact of ASCs with ECs were devoid of activin A [12]. This suggests that ECs modulate expression of activin A in ASCs through juxtacrine signaling pathways.

Our observations suggest that Notch2 plays a primary role in induction and Notch3 in inhibition of activin A expression since silencing Notch2 eliminates induction of activin A in ASCs cocultured with ECs (Fig. 5B, C), whereas silencing Notch3 enhanced activin A expression. While this study confirmed that induction of activin A in ASCs is EC dependent, and regulated through a trans-juxtacrine Notch2-mediated mechanism, whether direct EC + ASC contact is required for Notch3 activity is not clear yet. We hypothesize that Notch3 may compete with Notch2 for either common extracellular ligand(s) (e.g., jaggeds and delta-like 1, 3, 4) [24] or for intracellular downstream mediators of the signaling cascade, including cofactors/transcription factors [44,45].

In EC + ASC cocultures, expression of Notch2 did not change over time, whereas expression of Notch3 increased (Fig. 4B–D). The latter has also been observed in EC-SMC cocultures [22]. These data suggest that while Notch2 is the sole mechanism inducing activin A, Notch3 fine tunes the strength of activin A expression, such that the increase in Notch3 expression may partially be responsible for the decrease in activin A in EC + ASC cocultures over time (Fig. 4D).

It has been shown that EC-expressed jagged1 activates the smooth muscle differentiation program in adjacent mural cells by binding to their notch receptors [37]. The current study suggests that jagged1 is responsible for induction of activin A in ASCs, since exposure of these cells to immobilized jagged1 produced strong activin A expression (Fig. 5I), but interference with jagged1/Notch interaction prevented activin A expression in EC + ASC cocultures (Fig. 5J).

Expression of activin A in EC + ASC coculture model was observed as early as day three, but interference with notch signaling at this point, using DAPT, a broad inhibitor of Notch signaling, completely abrogated activin A expression, suggesting that the signaling pathway responsible for Inhibin Ba mRNA expression and/or activin A secretion has a short half-life and requires uninterrupted juxtacrine signaling [24]. This further suggests that activin A expression is under tight temporal control by dynamic changes in the extracellular environment. Indeed, in normal physiological circumstances, no expression of activin A can be detected in the vascular wall.

Several other mechanisms leading to decreases in activin A in cultures and vessel walls may be driven by formation of the basement membranes of vessel walls. As has been reported, vessel formation in EC + ASC cocultures is accompanied by deposition of extracellular matrix proteins, including collagen IV, perlecan, fibronectin, and laminin [10]. Several factors have been identified within basement membranes that either interfere with Notch-ligand interactions or directly modulate notch activation [46]. Microfibril-associated glycoprotein-2 (subunit of elastic fibers) binds to jagged1 and triggers shedding of jagged1's extracellular domain, thus preventing notch pathway activity [47]. EC-secreted EGFL7 regulates blood vessel formation [48] by blocking the ligand-binding site of notch, producing an antagonistic effect [49]. TSP-2, syndecan-3, and galectin are known modulators of notch activity [46]. Finally, notch-jagged and notch-delta interactions could be blocked by collagens I and IV, although the precise mechanism remains to be defined [50].

It is well known that acute injuries or chronic pathologies of the vessel wall cause disturbance of the organization of the basement membrane increasing the chances of interactions of cells from different layers within the wall (tunica intima and tunica media). Increased expression of activin A was observed in neointima of balloon-injured rat carotid artery [43] and in neointimal SMC in the early stages of atherosclerosis [51]. Whether notch signaling is involved in upregulation of Activin A in these conditions should be explored.

While activin A induces the smooth muscle phenotype in mural cells [12], mechanisms regulating activin A expression and smooth muscle differentiation in ASCs differ and are independent. Specifically, activin A induces αSMA in ASCs that are remote from EC cords, whereas induction of αSMA in ASCs that are in direct contact with EC cords is activin A independent [12].

Current study has revealed that Notch3 deficiency in ASCs spontaneously upregulates expression of SMC markers, which is further enhanced by direct contact with ECs (Fig. 6C, D). Reports on a role for Notch3 in smooth muscle differentiation are quite controversial. While some studies claim that signaling through Notch3 upregulates SMC markers [28,29], another claims that it is inhibitory [52], and at this point it is uncertain what factors contribute to these discrepancies. In the current study we were unable to identify the mechanism responsible for induction of smooth muscle phenotype in Notch3-deficient ASCs, and further studies are needed to clarify whether lack of Notch3 leads to secretion of soluble factor, which in autocrine manner induces SMC markers in ASCs or promotes cell differentiation through either trans or cis juxtacrine signaling pathways.

This study revealed that inhibition of notch signaling with DAPT caused poor EC organization into vascular cords, including formation of thick structures and incomplete engagement of ECs in the vasculogenic process (Fig. 2A, E). Similarly, mice treated with DAPT between P2–P5 develop immature vascular plexuses in the retina [53]. Based on the fact that DAPT prevented upregulation of activin A, the driver of the angiostatic program of ASCs, it was no surprise to find that in EC + ASC cocultures, DAPT promoted EC proliferation (Fig. 3A), accompanied by increased expression of the vasculogenic factors VEGF and HGF (Fig. 3C). Similarly, in cocultures with ASCs silenced for Notch2 (no activin A) increased HGF and SDF-1 mRNA was detected, whereas in cocultures with ASCs silenced for Notch3 (activin A overexpression) mRNAs for these factors were depressed (Fig. 5G, H). These findings further support our hypothesis that activin A plays an essential role in regulating vasculogenesis.

MSCs, pericytes, and smooth muscle cells share many signaling pathways and functional properties, suggesting that findings for one cell type may be relevant to others. The current study revealed that activin A expression in mural cells depends on juxtacrine notch signaling pathways, where Notch2 induces and Notch3 suppresses its expression (Fig. 7). These data suggest that in situations requiring vasculogenesis, such as ischemic tissues or engineered tissue implants, judicious manipulation of the timing and concentration of activin A exposure or notch-mediated signaling may improve the process. It also reveals a need to study drugs that influence activin A, particularly in vivo, to exploit the potential benefits of this signaling pathway to alleviate the health and economic burden of many pathological conditions that affect vasculature.

FIG. 7.

Schematic representation of proposed Notch signaling pathway responsible for EC-mediated induction of activin A secretion in ASC and subsequent autocrine signaling of activin A. Created with BioRender.com

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by grants from NIH R01 HL77688-01, VA Merit Review grants (K.L.M.), The Cryptic Masons' Medical Research Foundation, and UF Gatorade Foundation.

References

Virani

, Alonso

, Benjamin

, Bittencourt

, Callaway

, Carson

, Chamberlain

, Chang

, Cheng

, Delling

, Djousse

, Elkind

MSV

, Ferguson

, Fornage

, Khan

, Kissela

, Knutson

, Kwan

, Lackland

, Lewis

, Lichtman

, Longenecker

, Loop

, Lutsey

, Martin

, Matsushita

, Moran

, Mussolino

, Perak

, Rosamond

, Roth

, Sampson

UKA

, Satou

, Schroeder

, Shah

, Shay

, Spartano

, Stokes

, Tirschwell

, VanWagner

, Tsao

. (2020). Heart disease and stroke statistics-2020 update: a report from the American Heart Association. Circulation, 141:e139–e596.

Aday

, Matsushita

. (2021). Epidemiology of peripheral artery disease and polyvascular disease. Circ Res, 128:1818–1832.

Heyer

, Herberger

, Protz

, Glaeske

, Augustin

. (2016). Epidemiology of chronic wounds in Germany: analysis of statutory health insurance data. Wound Repair Regen, 24:434–442.

Guest

, Ayoub

, McIlwraith

, Uchegbu

, Gerrish

, Weidlich

, Vowden

. (2015). Health economic burden that wounds impose on the National Health Service in the UK. BMJ Open, 5:e009283.

Quyyumi

, Vasquez

, Kereiakes

, Klapholz

, Schaer

, Abdel-Latif

, Frohwein

, Henry

, Schatz

, Dib

, Toma

, Davidson

, Barsness

, Shavelle

, Cohen

, Poole

, Moss

, Hyde

, Kanakaraj

, Druker

, Chung

, Junge

, Preti

, Smith

, Mazzo

, Pecora

, Losordo

. (2017). PreSERVE-AMI: a randomized, double-blind, placebo-controlled clinical trial of intracoronary administration of autologous CD34+ cells in patients with left ventricular dysfunction post STEMI. Circ Res, 120:324–331.

Rheault-Henry

, White

, Grover

, Atoui

. (2021). Stem cell therapy for heart failure: Medical breakthrough, or dead end?. World J Stem Cells, 13:236–259.

Golchin

, Farahany

, Khojasteh

, Soleimanifar

, Ardeshirylajimi

. (2019). The clinical trials of mesenchymal stem cell therapy in skin diseases: an update and concise review. Curr Stem Cell Res Ther, 14:22–33.

Traktuev

, Merfeld-Clauss

, Li

, Kolonin

, Arap

, Pasqualini

, Johnstone

, March

. (2008). A population of multipotent CD34-positive adipose stromal cells share pericyte and mesenchymal surface markers, reside in a periendothelial location, and stabilize endothelial networks. Circ Res, 102:77–85.

Traktuev

, Prater

, Merfeld-Clauss

, Sanjeevaiah

, Murphy

, Johnstone

, Ingram

, March

. (2009). Robust functional vascular network formation in vivo by cooperation of adipose progenitor and endothelial cells. Circ Res, 104:1410–1420.

10.

Merfeld-Clauss

, Gollahalli

, March

, Traktuev

. (2010). Adipose tissue progenitor cells directly interact with endothelial cells to induce vascular network formation. Tissue Eng Part A, 16:2953–2966.

11.

Merfeld-Clauss

, Lupov

, Lu

, March

, Traktuev

. (2015). Adipose stromal cell contact with endothelial cells results in loss of complementary vasculogenic activity mediated by induction of activin A. Stem Cells, 33:3039–3051.

12.

Merfeld-Clauss

, Lupov

, Lu

, Feng

, Compton-Craig

, March

, Traktuev

. (2014). Adipose stromal cells differentiate along a smooth muscle lineage pathway upon endothelial cell contact via induction of activin A. Circ Res, 115:800–809.

13.

Melero-Martin

, De Obaldia

, Kang

, Khan

, Yuan

, Oettgen

, Bischoff

. (2008). Engineering robust and functional vascular networks in vivo with human adult and cord blood-derived progenitor cells. Circ Res, 103:194–202.

14.

Lilly

(2014) We have contact: endothelial cell-smooth muscle cell interactions. Physiology (Bethesda), 29:234–241.

15.

Ata

, Lennmyr

, Funa

, Olsson

, Terent

. (1999). Expression of transforming growth factor-beta1, 2, 3 isoforms and type I and II receptors in acute focal cerebral ischemia: an immunohistochemical study in rat after transient and permanent occlusion of middle cerebral artery. Acta Neuropathol, 97:447–455.

16.

, Lai

, Hughes

, Sirimanne

, Gluckman

, Williams

. (1999). Expression of the activin axis and neuronal rescue effects of recombinant activin A following hypoxic-ischemic brain injury in the infant rat. Brain Res, 835:369–378.

17.

Maeshima

, Zhang

, Nojima

, Naruse

, Kojima

. (2001). Involvement of the activin-follistatin system in tubular regeneration after renal ischemia in rats. J Am Soc Nephrol, 12:1685–1695.

18.

Apostolou

, Stavropoulos

, Sountoulidis

, Xirakia

, Giaglis

, Protopapadakis

, Ritis

, Mentzelopoulos

, Pasternack

, Foster

, Ritvos

, Tzelepis

, Andreakos

, Sideras

. (2012). Activin-A overexpression in the murine lung causes pathology that simulates acute respiratory distress syndrome. Am J Respir Crit Care Med, 185:382–391.

19.

Verhamme

, Bracke

, Amatngalim

, Verleden

, Van Pottelberge

, Hiemstra

, Joos

, Brusselle

. (2014). Role of activin-A in cigarette smoke-induced inflammation and COPD. Eur Respir J, 43:1028–1041.

20.

Phng

, Gerhardt

. (2009). Angiogenesis: a team effort coordinated by notch. Dev Cell, 16:196–208.

21.

Dou

, Wang

, Hu

, Hou

, Wang

, Xu

, Wang

, Liang

, Yao

, Yang

, Han

. (2008). RBP-J, the transcription factor downstream of Notch receptors, is essential for the maintenance of vascular homeostasis in adult mice. FASEB J, 22:1606–1617.

22.

Liu

, Kennard

, Lilly

. (2009). NOTCH3 expression is induced in mural cells through an autoregulatory loop that requires endothelial-expressed JAGGED1. Circ Res, 104:466–475.

23.

del Alamo

, Rouault

, Schweisguth

. (2011). Mechanism and significance of cis-inhibition in Notch signalling. Curr Biol, 21:R40–R47.

24.

Kopan

, Ilagan

. (2009). The canonical Notch signaling pathway: unfolding the activation mechanism. Cell, 137:216–233.

25.

Liu

, Shirakawa

, Li

, Soma

, Oka

, Dotto

, Fairman

, Velazquez

, Herlyn

. (2003). Regulation of Notch1 and Dll4 by vascular endothelial growth factor in arterial endothelial cells: implications for modulating arteriogenesis and angiogenesis. Mol Cell Biol, 23:14–25.

26.

Shawber

, Das

, Francisco

, Kitajewski

. (2003). Notch signaling in primary endothelial cells. Ann N Y Acad Sci, 995:162–170.

27.

Tang

, Urs

, Boucher

, Bernaiche

, Venkatesh

, Spicer

, Vary

, Liaw

. (2010). Notch and transforming growth factor-beta (TGFbeta) signaling pathways cooperatively regulate vascular smooth muscle cell differentiation. J Biol Chem, 285:17556–17563.

28.

Domenga

, Fardoux

, Lacombe

, Monet

, Maciazek

, Krebs

, Klonjkowski

, Berrou

, Mericskay

, Li

, Tournier-Lasserve

, Gridley

, Joutel

. (2004). Notch3 is required for arterial identity and maturation of vascular smooth muscle cells. Genes Dev, 18:2730–2735.

29.

Baeten

, Lilly

. (2015). Differential regulation of NOTCH2 and NOTCH3 contribute to their unique functions in vascular smooth muscle cells. J Biol Chem, 290:16226–16237.

30.

Ingram

, Mead

, Tanaka

, Meade

, Fenoglio

, Mortell

, Pollok

, Ferkowicz

, Gilley

, Yoder

. (2004). Identification of a novel hierarchy of endothelial progenitor cells using human peripheral and umbilical cord blood. Blood, 104:2752–2760.

31.

Lin

, Lilly

. (2014). Endothelial cells direct mesenchymal stem cells toward a smooth muscle cell fate. Stem Cells Dev, 23:2581–2590.

32.

Wang

, Sugano

, Isago

, Murayama

, Tamai

, Tomita

. (2012). Notch signaling pathway regulates proliferation and differentiation of immortalized Müller cells under hypoxic conditions in vitro. Neuroscience, 214:171–180.

33.

, Chen

, Xiang

, Li

, Wang

. (2016). jagged1/Notch3 signaling modulates hemangioma-derived pericyte proliferation and maturation. Cell Physiol Biochem, 40:895–907.

34.

Binesh

, Devaraj

, Halagowder

. (2019). Molecular interaction of NFκB and NICD in monocyte-macrophage differentiation is a target for intervention in atherosclerosis. J Cell Physiol, 234:7040–7050.

35.

Yang

, Arslanova

, Gu

, Augelli-Szafran

, Xia

. (2008). Quantification of gamma-secretase modulation differentiates inhibitor compound selectivity between two substrates Notch and amyloid precursor protein. Molecular Brain, 1:15.

36.

Merfeld-Clauss

, Lease

, Lu

, March

, Traktuev

. (2017). Adipose stromal cells differentiation toward smooth muscle cell phenotype diminishes their vasculogenic activity due to induction of activin A secretion. J Tissue Eng Regen Med, 11:3145–3156.

37.

High

, Lu

, Pear

, Loomes

, Kaestner

, Epstein

. (2008). Endothelial expression of the Notch ligand jagged1 is required for vascular smooth muscle development. Proc Natl Acad Sci U S A, 105:1955–1959.

38.

Hirschi

, Rohovsky

, D'Amore

. (1998). PDGF, TGF-beta, and heterotypic cell-cell interactions mediate endothelial cell-induced recruitment of 10T1/2 cells and their differentiation to a smooth muscle fate. J Cell Biol, 141:805–814.

39.

Wang

, Yin

, Cen

, Liu

, Cao

, Cui

. (2010). Differentiation of adipose-derived stem cells into contractile smooth muscle cells induced by transforming growth factor-beta1 and bone morphogenetic protein-4. Tissue Eng Part A, 16:1201–1213.

40.

Harris

, Abdollahi

, Zhang

, McIlhenny

, Tulenko

, DiMuzio

. (2011). Differentiation of adult stem cells into smooth muscle for vascular tissue engineering. J Surg Res, 168:306–314.

41.

Martin

, Weiss

, Metharom

, Schmeckpeper

, Hynes

, O'Sullivan

, Caplice

. (2009). Thrombin stimulates smooth muscle cell differentiation from peripheral blood mononuclear cells via protease-activated receptor-1, RhoA, and myocardin. Circ Res, 105:214–218.

42.

Kim

, Jeon

, Kim

, Jho

, Ryu

, Kim

. (2008). Angiotensin II-induced differentiation of adipose tissue-derived mesenchymal stem cells to smooth muscle-like cells. Int J Biochem Cell Biol, 40:2482–2491.

43.

Pawlowski

, Taylor

, Valentine

, Hail

, Ferrer

, Kowala

, Molloy

. (1997). Stimulation of activin A expression in rat aortic smooth muscle cells by thrombin and angiotensin II correlates with neointimal formation in vivo. J Clin Invest, 100:639–648.

44.

Borggrefe

, Oswald

. (2009). The Notch signaling pathway: transcriptional regulation at Notch target genes. Cell Mol Life Sci, 66:1631–1646.

45.

Bray SJ. (2016). Notch signalling in context. Nat Rev Mol Cell Biol, 17:722–735.

46.

LaFoya

, Munroe

, Mia

, Detweiler

, Crow

, Wood

, Roth

, Sharma

, Albig

. (2016). Notch: A multi-functional integrating system of microenvironmental signals. Dev Biol, 418:227–241.

47.

Nehring

, Miyamoto

, Hein

, Weinmaster

, Shipley

. (2005). The extracellular matrix protein MAGP-2 interacts with jagged1 and induces its shedding from the cell surface. J Biol Chem, 280:20349–20355.

48.

Parker

, Schmidt

, Jin

, Gray

, Beis

, Pham

, Frantz

, Palmieri

, Hillan

, Stainier

, De Sauvage

, Ye

. (2004). The endothelial-cell-derived secreted factor Egfl7 regulates vascular tube formation. Nature, 428:754–758.

49.

Schmidt

MHH

, Bicker

, Nikolic

, Meister

, Babuke

, Picuric

, Muller-Esterl

, Plate

, Dikic

. (2009). Epidermal growth factor-like domain 7 (EGFL7) modulates Notch signalling and affects neural stem cell renewal. Nat Cell Biol, 11:873–880.

50.

Zhang

, Meng

, Wang

. (2013). Collagen represses canonical Notch signaling and binds to Notch ectodomain. Int J Biochem Cell Biol, 45:1274–1280.

51.

Engelse

, Neele

, van Achterberg

, van Aken

, van Schaik

, Pannekoek

, de Vries

. (1999). Human activin-A is expressed in the atherosclerotic lesion and promotes the contractile phenotype of smooth muscle cells. Circ Res, 85:931–939.

52.

Kennard

, Liu

, Lilly

. (2008). Transforming growth factor-beta (TGF- 1) down-regulates Notch3 in fibroblasts to promote smooth muscle gene expression. J Biol Chem, 283:1324–1333.

53.

Scheppke

, Murphy

, Zarpellon

, Hofmann

, Merkulova

, Shields

, Weis

, Byzova

, Ruggeri

, Iruela-Arispe

, Cheresh

. (2012). Notch promotes vascular maturation by inducing integrin-mediated smooth muscle cell adhesion to the endothelial basement membrane. Blood, 119:2149–2158.