Endothelial Jagged1 Antagonizes Dll4 Regulation of Endothelial Branching and Promotes Vascular Maturation Downstream of Dll4/Notch1Significance
Objective—Notch signaling controls cardiovascular development and has been associated with several pathological conditions. Among its ligands, Jagged1 and Dll4 were shown to have opposing effects in developmental angiogenesis, but the underlying mechanism and the role of Jagged1/Notch signaling in adult angiogenesis remain incompletely understood. The current study addresses the importance of endothelial Jagged1-mediated Notch signaling in the context of adult physiological angiogenesis and the interactions of Jagged1 and Dll4 on angiogenic response and vascular maturation processes.
Approach and Results—The role of endothelial Jagged1 in wound healing kinetics and angiogenesis was investigated with endothelial-specific Jag1 gain-of-function and loss-of-function mouse mutants (eJag1OE and eJag1cKO). To study the interactions between the 2 Notch ligands, genetic mouse models were combined with pharmacological inhibition of Dll4 or Jagged1, respectively. Jagged1 overexpression in endothelial cells increased vessel density, maturation, and perfusion, thus accelerating wound healing. The opposite effect was seen in eJag1cKO animals. Interestingly, Dll4 blockade in these animals led to an increase in vascular density but induced a greater decrease in perivascular cell coverage. However, Jagged1 inhibition in Dll4 gain-of-function (eDll4OE) mutants, with reduced angiogenesis, further diminished angiogenic growth and hampered perivascular cell coverage. Our findings suggest that as Dll4 blocks endothelial activation through Notch1 signaling, it also induces Jagged1 expression. Jagged1 then blocks Dll4 signaling through Notch1, allowing endothelial activation by vascular endothelial growth factor and endothelial layer growth. Jagged1 also initiates maturation of the newly formed vessels, possibly by binding and activating endothelial Notch4. Importantly, mice administered with a Notch4 agonistic antibody mimicked the mural cell phenotype of eJag1OE mutants without affecting angiogenic growth, which is thought to be Notch1 dependent.
Conclusions—Endothelial Jagged1 is likely to operate downstream of Dll4/Notch1 signaling to activate Notch4 and regulate vascular maturation. Thus, Jagged1 not only counteracts Dll4/Notch in the endothelium but also generates a balance between angiogenic growth and maturation processes in vivo.
Increasing evidence indicates the importance of Notch signaling in the development and homeostasis of the vascular system as well as in the pathogenesis of several diseases. The Notch signaling pathway comprises 4 different transmembrane receptors (Notch1–4) and 5 ligands (Delta like 1, 3, and 4 and Jagged1 and 2). Binding of a Notch ligand to a receptor triggers a series of conformational changes and cleavages of the receptors. These lead to the release and translocation of the Notch intracellular domain to the nucleus where it controls the expression of a variety of target genes depending on cell type and on biological context.1 The Dll4 and Jag1 (encoding Jagged1) genes were shown to be fundamental in developmental angiogenesis. Dll4 ligand is required for normal arterial patterning in the embryo2 and was also shown to have a major antiangiogenic effect in wound healing in the adult.3 Jag1-null mouse mutants die at E11.5 because of heart defects and abnormal development of the yolk sac and head vasculature.4 Moreover, mutations in the human JAG1 gene cause Alagille syndrome, which comprises complex cardiac defects and vascular anomalies.5 However, little is known about the function of Jagged1 in adult neoangiogenesis.
Dll4 is predominantly expressed in the endothelium of capillaries and arteries,6 whereas Jagged1 is expressed in both endothelial and vascular smooth muscle cells (vSMCs).7 During developmental angiogenesis in the postnatal mouse retina, Dll4 is detected in endothelial tip cells at the distal end of sprouting vessels.8 In contrast, Jagged1 is found in stalk cells at the sprout base.9 A recent study has suggested that Dll4 and Jagged1 have opposite effects in mouse retinal vascular development. In this setting, Jagged1 acts as a positive regulator of sprouting and tip cell formation because of its ability to antagonize Dll4/Notch signaling, which acts as a negative regulator of these processes.9
Previous work has also shown that endothelial Jagged1 is indispensable for the development of neighboring vascular smooth muscle,10 which has been attributed to the induction of mural cell differentiation through Notch311 and RBP-Jkappa12 activation.
Despite considerable progress in understanding vascular morphogenesis during development, few studies have targeted in vivo angiogenesis in adult physiological settings. In mammals, physiological angiogenesis occurs mainly during tissue regeneration, such as in healing wounds, and in the female reproductive cycle. In particular, the wound healing assay constitutes a fast, easy, and reliable in vivo model of physiological angiogenesis for studying the molecular mechanisms involved in the formation and remodeling of vascular structures.13 During regeneration, the vascular network at the wound edges expands through sprouting into the granulation tissue, which has increased oxygen and nutrient demand. Vascular endothelial growth factor (VEGF) is the main proangiogenic signal controlling this process.14
Through the analysis of wound healing kinetics, vessel morphology, and gene expression in endothelial Jag1 or Dll4 mutants treated with ligand-specific antagonists, we provide the first evidence for an important proangiogenic role of endothelial Jagged1 in adult regenerative angiogenesis. Moreover, we present evidence supporting a synergistic effect of both ligands on mural cell recruitment and propose a new mechanism by which Dll4/Notch1 signaling promotes the recruitment of perivascular cells through the upregulation of Jag1 and Notch4 activation in endothelial cells (ECs).
Materials and Methods
Materials and Methods are available in the online-only Data Supplement.
Endothelial Jagged1 Accelerates Wound Healing by Promoting Angiogenesis and Vessel Maturation
To better characterize the role of Jagged1/Notch signaling during adult physiological angiogenesis, wound healing kinetics were analyzed in endothelial-specific Jag1 gain-of-function and loss-of-function mutant mice.
EC-specific Jag1 overexpression (endothelial-specific Jag1 gain-of-function or conditional overexpression [eJag1OE]) led to significantly accelerated wound healing from day 2 of recovery onward relative to control animals. Accordingly, closing of eJag1OE wounds occurred 2 days faster on average (Figure 1A). In contrast, loss of endothelial Jag1 (endothelial-specific Jag1 loss-of-function or conditional knockout [eJag1cKO]) led to significantly delayed healing from day 2 of recovery with an average 2-day delay in the healing process (Figure 2A).
To address whether the altered wound closure kinetics observed in these EC-specific mutants was indeed associated with altered vessel growth, the vascular morphology of the samples collected at end point (day 7 of recovery) was examined. Blood vessel endothelium in the wounded area was visualized by immunostaining against platelet endothelial cell adhesion molecule-1 (PECAM-1), whereas α-smooth muscle actin (SMA) and platelet-derived growth factor-β (Pdgfr-β) were used to reveal perivascular cell and pericyte coverage, respectively, and thereby analyze vessel maturation (Figures 1B, 1C, 2B, and 2C and Figure IA and IB in the online-only Data Supplement). In eJag1OE mutants, we observed higher vascular density (Figure 1B–1D) as well as an increase in vSMC (Figure 1B, 1C, and 1E) and pericyte coverage (Figure IA and IB in the online-only Data Supplement). As shown in detail in Figure 1C, the newly formed vasculature in eJag1OE mutants displayed highly branched networks and exhibited a greater maturation state, as indicated by the abundance of both perivascular cells and pericytes present. The opposite was seen in eJag1cKO mutant wounds, which showed a significant decrease in vascular density (Figure 2B–2D) as well as in the number of perivascular cells (Figure 2B, 2C, and 2E) and pericyte coverage (Figure IC and ID in the online-only Data Supplement). At high magnification, the vasculature of eJag1cKO wounds was sparse and with few SMCs (Figure 2C) and pericytes attached (Figure IC and ID in the online-only Data Supplement).
We also evaluated wound vessel functionality by analyzing biotinylated lectin perfusion and Evans’ blue dye vascular leakage (Figures 1F, 1H, 2F, and 2H). As shown in Figure 1F and 1G, overexpression of Jag1 in the endothelium was associated with an increase in the amount of perfused, lectin-containing vessels. At the same time, Evans’ blue extravasation was significantly reduced in comparison with controls (Figure 1H and 1I), suggesting an overall improvement in the functionality of newly formed vessels. Conversely, endothelial Jag1 loss of function led to a significant decrease in vascular perfusion (Figure 2F and 2G) and to an increase in extravasation per vessel area (Figure 2H and 2I). Taken together, endothelial Jag1 overexpression led to the formation of a dense, mature, and more functional wound vascular plexus, whereas endothelial Jag1 loss of function led to a sparse, immature, and poorly functional vessel network. These results establish that endothelial Jag1 expression is responsible for modulating vascular growth and maturation during in vivo regenerative angiogenesis, which, in turn, affects wound closure kinetics.
Regulation of Angiogenic Gene Expression by Jagged1/Notch Signaling
To better understand the molecular mechanisms behind the vascular phenotypes observed in eJag1OE and eJag1cKO mutants, we performed quantitative reverse transcriptase polymerase chain reaction analysis of selected genes involved in angiogenesis (Figure 3). RNA was extracted from EC (Lin− (ter119−cd45−) cd31+) and mural cell (Lin− (ter119−cd45−) cd146+) fluorescence-activated cell sorting sorted from wound samples collected at the experimental end point (ie, day 7 of recovery; Figure 3A). In ECs (Figure 3B), Jag1 transcript levels were largely increased (3-fold) in eJag1OE mutant samples and significantly reduced in eJag1cKO animals. In contrast, Dll4 was upregulated in eJag1cKO and downregulated in eJag1OE mutants. Transcript levels for Pdgfb (encoding PDGF-B, the endothelial ligand for PDGFRβ) and Tek (encoding the Tie2 receptor tyrosine kinase), which control the recruitment of perivascular cells and vascular permeability, respectively,15,16 were downregulated in eJag1cKO mutants and increased in gain-of-function mutants. The same was the case for the mRNAs encoding VEGF receptor-2 (Vegfr2/Kdr/Flk1) and 3 (Vegfr3/Flt4) (Figure 3B).
About known Notch effectors, we observed that Hey2 and Hes1 transcript levels were augmented in eJag1cKO samples and downregulated in eJag1OE. However, Hey1 and Hes2 were upregulated in eJag1OE and reduced in eJag1cKO wound ECs. Nrarp, a gene that is known to be rapidly upregulated in response to Dll4/Notch signaling, was strongly upregulated in the absence of endothelial Jag1 and downregulated in the Jag1 gain-of-function samples. The Notch receptor gene Notch 4 was upregulated in eJag1OE wounds, whereas in eJag1cKO only Notch1 was upregulated, and Notch4 was robustly and significantly reduced.
Furthermore, mural cell–specific transcription analysis (Figure 3C) revealed a downregulation of PdgfrB (encoding PDGFRβ), Ang1 (perivascular ligand for Tie2 receptor), Notch3 receptor, and HeyL (perivascular cell notch effector) in eJag1cKO and an upregulation in eJag1OE mutants wounds. To obtain additional validation on the specific modulation of Notch effectors by endothelial Jagged1, immunofluorescence was performed for the main endothelial Notch effectors Hey1 and Hey2 (Figure 4). Confirming the transcription results, it is clear from the quantification of double positive signal for the effectors with PECAM that in the eJag1OE wound vasculature there is increased Hey1 (Figure 4A and 4B) and decreased Hey2 (Figure 4E and 4F) levels. In contrast, eJag1cKO mutant vasculature presented decreased levels of Hey1 (Figure 4C and 4D) and increased levels of Hey2 effector (Figure 4G and 4H).
These results indicate that Jag1 modulation in the endothelium is able to ellicit changes in the expression profiles of Notch receptors and effectors as well as in genes controlling angiogenesis and the recruitment of mural cells.
Blocking Dll4 in Endothelial Jag1 Knockout Mice Rescues Angiogenesis but Not Mural Cell Coverage
The results obtained in the wound healing assays performed in Jag1 mutants together with previous results from our laboratory describing a role for Dll4 in wound healing angiogenesis3 gave rise to the hypothesis that the loss of endothelial Jag1 enables more robust Dll4/Notch signaling. In the absence of Jagged1, more Notch receptors would be left available to be activated by Dll4, which has a strong antiangiogenic function and might thereby cause a delay in wound healing.3 Conversely, in eJag1OE animals, the proangiogenic phenotype might be associated with decreased Dll4/Notch signaling. We therefore wanted to determine whether the eJag1cKO phenotype was mainly because of upregulation of Dll4/Notch signaling. We performed wound healing assays in eJag1cKO mutant mice treated with anti-Dll4 blocking antibody (anti-Dll4). As shown in Figure 5, anti-Dll4 administration increased the angiogenic growth of the endothelium, in both control (control+anti-Dll4) and eJag1cKO (eJag1cKO+anti-Dll4) mouse groups but led to significantly delayed wound healing kinetics (Figure 5A). Morphological and quantitative analysis of the wound vasculature (Figure 5B) clearly showed an increase in the vascular density on antibody blockade of Dll4 (Figure 5C). Strikingly, anti-Dll4 administration failed to rescue the defective mural cell (Figure 5B and 5D) and pericyte (Figure IIA and IIB in the online-only Data Supplement) coverage in eJag1cKO mutant mice. In fact, smooth muscle cell coverage showed a significant decrease on Dll4 blockade (Figure 5B and 5D) indicating that the reduced vascular maturation in eJag1cKO animals was further aggravated by treatment with anti-Dll4. Moreover, overall functionality of the newly formed vasculature was diminished by Dll4 blockade, as indicated by reduced perfusion (Figure IIC and IID in the online-only Data Supplement) and increased leakiness (Figure IIE and IIF in the online-only Data Supplement). No evident anti–Dll4-induced effects were observed during the experimental period in tissues other than the skin vasculature (data not shown).
With regard to ECs and mural cell–specific gene expression (Figure III in the online-only Data Supplement), we observed downregulation of Notch-related genes both in extra control mouse group (control+anti-Dll4) as in eJag1cKO (eJag1cKO+anti-Dll4) mouse groups treated with anti-Dll4 antibody. Administration of anti-Dll4 to eJag1cKO mutants reverted the upregulation of Dll4, Notch1, Hey2, Hes1, and Nrarp observed previously in these mutant mice (Figure IIIA in the online-only Data Supplement). Also, as expected, blockade of Dll4 alone (control+anti-Dll4) led to a downregulation of Jag1 transcript levels, confirming the previously obtained results indicating that Jag1 expression was downstream of Dll4/Notch signaling.3 In addition, we observed a downregulation response in the transcription of mural cell–specific genes (Figure IIIB in the online-only Data Supplement) in all analyzed mouse groups.
In summary, these observations indicate that Jag1 and Dll4 have opposing effects on regenerative angiogenesis in the adult organism. However, the data suggest a synergistic function of the 2 ligands in vessel maturation and perivascular cell recruitment to nascent vessels.
Vascular Maturation in Endothelial-Specific Dll4 Gain-of-Function or Conditional Overexpression Mice Is Impaired by Jagged1 Blockade
The results described to date show that Dll4 inhibition in eJag1cKO mice further diminishes the mural cell coverage of the vascular plexus formed during wound healing. Previous work done by our group has established that endothelial-specific overexpression of Dll4 (endothelial-specific Dll4 gain-of-function or conditional overexpression [eDll4OE]) induces an increase in vSMC coverage of the vasculature.3 To better understand the interaction between Dll4 and Jagged1 signaling in the vessel maturation process, we treated eDll4OE mutants with an anti-Jagged1 blocking antibody and analyzed the effect of this inhibition on wound healing. In accordance with our previous study,3 Dll4 overexpression in the endothelium led to a significant delay in wound closure relative to controls (Figure 6A). This reduction in the healing ability was phenocopied by Jagged1 inhibition in control animals (control+anti-Jag1), which argued further for opposing roles of Dll4 and Jag1. In addition, treatment of eDll4OE mutants with anti-Jagged1 antibody severely impaired tissue regeneration suggesting that inhibition of Jagged1 further hampers growth and function of the dermal vasculature. No evident anti–Jag1-induced effects were observed during the experimental period in tissues other than the vasculature (data not shown).
Microscopical analysis of the vasculature at high magnification revealed that both Dll4 overexpression and Jagged1 inhibition were able to induce a significant decrease in the vascular density of the granulation tissue. This effect was most pronounced after Jagged1 inhibition in eDll4OE mutant mice (Figure 6B–6D), which suggested that Dll4 overexpression and Jagged1 blockade have additive antiangiogenic effects.
Moreover, eDll4OE mutant vasculature displayed a high ratio of coverage by SMA+ cells (Figure 6B and 6C) and pericytes (Figure IVA and IVB in the online-only Data Supplement), despite the reduced vascular density observed. Conversely, Jagged1 antibody inhibition significantly reduced the proportion of vessels covered by vSMCs (Figure 6E) and pericytes (Figure IVA and IVB in the online-only Data Supplement) in both controls and eDll4OE animals. These results argue that Jagged1-mediated signaling is indispensable for the proper mural cell recruitment and sustainment of SMC and pericyte coverage during adult physiological angiogenesis. Moreover, the observed loss of maturation may be responsible for the diminished vascular perfusion and increased extravasation in animals treated with Jagged1 blocking antibody (Figure IVC–IVF in the online-only Data Supplement).
Transcription analysis revealed that inhibition of Jagged1 in eDll4OE mutants reverted the upregulation of Jag1, Pdgfb, Tek, Notch4, and Hey1 (Figure VA in the online-only Data Supplement), and of Pdgfr-β, Ang-1, Notch3, and HeyL (Figure VB in the online-only Data Supplement), that was observed in these animals. This suggests that the transcription of these genes is directly regulated by Jagged1. Conversely, inhibition of Jagged1 in eDll4OE mutants (eDll4OE+anti-Jag1) maintained the upregulation of Dll4, Notch1, Hey2, and Nrarp genes (Figure VA in the online-only Data Supplement).
In line with previous findings suggesting that Jag1 expression is positively regulated by Dll43 and with the observed Jagged1 upregulation (Figure VA in the online-only Data Supplement), immunostaining for Jagged1 in eDll4OE mutant vasculature showed much greater intensity in comparison with controls (Figure 6F). This suggests that Jagged1 is a downstream effector of Dll4/Notch signaling in the endothelium.
Endothelial Jagged1 Activates Notch3 Receptor in Perivascular Cells and Notch4 in ECs
It has been proposed that Jagged1 signals through Notch3 in perivascular cells.11 To confirm this, we immunostained the intracellular domain of Notch3 (N3ICD) in the wound samples from mutants that presented increased recruitment of mural cells: eJag1OE and eDll4OE. Endothelial Jag1 overexpression mutants presented increased perivascular N3ICD positive staining compared with the respective controls (Figure VIA and VIB in the online-only Data Supplement), measured as double positive staining for SMA and N3ICD. Conversely, administration of anti-Jagged1 to control animals (control+anti-Jag1) led to decreased N3ICD positive staining, whereas eDll4OE mutants presented increased staining (Figure VIC and VID in the online-only Data Supplement). Predictably, blocking Jagged1 in eDll4OE (eDll4OE+anti-Jag1) reversed the increased staining observed in the mutants, bringing N3ICD to levels comparable with control animals (Figure VIC and VID in the online-only Data Supplement).
Despite the evidence for the ability of Jagged1 to activate perivascular Notch3, the endothelial Notch receptor(s) for Jagged1 remained unknown. Both Notch1 and Notch4 are prominently expressed in ECs,7 therefore to better understand the downstream signaling elicited by endothelial Jagged1, immunofluorescence staining for the intracellular domains of Notch1 (N1ICD) and Notch4 (N4ICD) was performed in wound samples from eJag1 mutants (Figure 7). Jag1 overexpression led to a significant decrease in the amount of PECAM-1+ vessels showing activation of Notch1 (double positive N1ICD/PECAM-1 vessels; Figure 7A and 7B). The opposite effect was seen in eJag1cKO mutants where the proportion of N1ICD+ vessels was increased relative to controls (Figure 7E and 7F). Surprisingly, levels of activated Notch4 (N4ICD) were increased in eJag1OE mutants (Figure 7C and 7D), whereas conditional deletion of Jag1 in the endothelium was associated with a decrease in the colocalization of N4ICD and PECAM-1 staining (Figure 7G and 7H).
We also quantified endothelial N1ICD in eJag1cKO animals treated with anti-Dll4 and N4ICD in eDll4OE mutants treated with anti-Jag1 (Figure VII in the online-only Data Supplement). Anti-Dll4 administration reversed the increase of ECs with cleaved (active) Notch1 in eJag1cKO samples (Figure VIIA and VIIB in the online-only Data Supplement). Likewise, anti-Jag1 administration significantly reduced the N4ICD-positive area in the eDll4OE endothelium (Figure VIIC and VIID in the online-only Data Supplement). Moreover, anti-Dll4 administered either to control (control+anti-Dll4) or eJag1cKO (eJag1cKO+anti-Dll4) mouse groups also led to reduced levels of active endothelial Notch4 (Figure VIIIA and VIIIB in the online-only Data Supplement). In the same manner, administration of anti-Jag1 either to control (control+anti-Jag1) or eDll4OE (eDll4OE+anti-Jag1) mouse groups also showed a sustained increase in active endothelial Notch1 (Figure VIIIC and VIIID in the online-only Data Supplement). These results argue that Jagged1 is able to trigger Notch4 receptor activation in ECs, whereas the receptor Notch1 seems to be preferentially activated by Dll4.
Administration of a Notch4 Agonist to Wild-Type Mice Accelerates Wound Healing by Promoting Vessel Maturation Without Affecting Angiogenic Growth
To provide independent evidence of the role of Notch4 in the vascular response and to better ascertain whether indeed endothelial Notch4 also contributes to the vascular maturation process, we administered a Notch4 agonistic antibody to wild-type (WT) mice and evaluated the healing response. Administration of a Notch4 agonist to WT mice (WT+N4 Agonist) led to significantly accelerated wound healing from day 2 of recovery (Figure IXA in the online-only Data Supplement). Analysis of the wound vasculature revealed that despite no significant change had been observed in vascular density (Figure IXB–IXD in the online-only Data Supplement), animals injected with the agonist presented increased coverage of SMA+ cells (Figure IXA, IXB, and IXE in the online-only Data Supplement) and pericytes (Figure X in the online-only Data Supplement). Moreover, the newly formed vasculature of Notch4 agonist injected mice presented increased perfusion (Figure IXF and IXG in the online-only Data Supplement) and decreased leakage (Figure IXH and IXI in the online-only Data Supplement), consistent with an increased maturation status.
Transcription analysis further supported our previous results, with the same set of genes that were upregulated in eJag1OE also being upregulated in WT+ N4 agonist mice (Figure XIA in the online-only Data Supplement). This group of genes included all the analyzed genes involved in recruitment of perivascular cells, such as Pdgfb, Pdgfr-β, Tek, and Ang-1. It also included Jag1, Notch4, Notch3, Hey1, and HeyL. Contrastingly, Dll4, Notch1, and Hey2 transcription levels showed no significant changes in the WT+ N4 agonist group.
Accordingly to the transcription analysis, Notch4 agonist injected animals showed increased endothelial Hey1 positive area (Figure XIB and XIC in the online-only Data Supplement), whereas no significant differences were observed on Hey2 staining (Figure XID and XIE in the online-only Data Supplement).
These results reinforce the evidence for the role of endothelial Jagged1/Notch4 signaling in the process of vascular maturation.
This article addresses the role of Jagged1/Notch signaling and its relation to the other major Notch ligand, Dll4, during angiogenic processes related to skin wound healing. In particular, our results establish that the proangiogenic role of Jag1 observed in developmental processes9 is also relevant for regenerative angiogenesis in the adult. We found that endothelial Jag1 overexpression improved the healing rate of dermal wounds and was associated with an increase in the density, maturation, and functionality of the newly formed vasculature, whereas endothelial-specific Jag1 knockout had the opposite effect. VEGF receptors were upregulated in eJag1OE mutants and downregulated in eJag1cKO samples. This finding is consistent with previously published work9 showing that Jagged1 can promote VEGF signaling by upregulating the levels of Vegfr3 and Vegfr2.
Given our previous studies with endothelial-specific Dll4 mutants,3 the eJag1cKO loss-of-function endothelial phenotype is similar to that observed in eDll4OE mutants, whereas the eJag1OE phenotype corresponds to that observed in eDll4cKO mutants. Dll4 blockade with an anti-Dll4 antibody led to delayed wound healing, in accordance with our previous study where we showed the same wound healing kinetics using either endothelial-specific Dll4 loss-of-function mutant mice or administration of a soluble Dll4-Fc fusion protein.3 Administration of anti-Dll4 to eJag1cKO mutants led to a further delay in wound healing and significantly increased vascular density, demonstrating the opposing roles of the 2 Notch ligands in the endothelium, Dll4 antiangiogenic and Jagged1 proangiogenic.
In addition, our new data show that Jagged1 and Dll4 have also overlapping functional roles in the vasculature, as both ligands contributed to increased vSMC and pericyte coverage and reduced vascular leakage during wound healing angiogenesis. We propose that Jagged1/Notch signaling might be directly responsible for maturation processes in a newly formed vascular plexus acting downstream of Dll4. This is supported by the observation that the perivascular phenotype observed in eDll4OE mutants is linked to the downstream upregulation of Jag1, which was visible both at the transcript level and at the protein level, as shown in Figure 5F. Accordingly, treatment of eDll4OE mutants with anti-Jagged1 antibody led to decreased recruitment of mural support cells. In line with our observations, High et al10 have shown that endothelial-specific Jag1 knockout mutant embryos have striking deficits in vascular smooth muscle, whereas endothelial Notch1 activation and arterial-venous differentiation seemed normal. They also showed that endothelial Jag1 mutant embryos are phenotypically distinct from embryos in which Notch signaling is inhibited in the endothelium. Therefore, the primary role of endothelial Jagged1 may be to potentiate the development and differentiation of neighboring vSMCs. Thus, Jagged1 not only counteracts Dll4/Notch in the endothelium but also generates a balance between angiogenic growth and maturation processes. Our observations also suggest that endothelial Jagged1 negatively regulates the transcription and activation of Notch1, whereas it positively controls the transcription and activation of Notch4. Our colocalization analysis suggests that Jagged1 might be unable to trigger substantial Notch1 activation. Accordingly, a recent study17 demonstrated that in vitro cultivation of ECs in high glucose conditions causes increased angiogenesis because of Jagged1 overexpression and inhibition of Notch1. Conversely, eJag1cKO mutants, where the transcript levels of Dll4 were upregulated, showed a significant increase in N1ICD-positive endothelium. This increase was reverted by blocking Dll4 with anti-Dll4 antibody. In addition, antibody blocking of Jagged1 failed to revert the increased active endothelial Notch1 observed in eDll4OE mouse mutants, suggesting that Notch1 is mainly activated by Dll4. In fact, the embryonic lethal phenotype observed in Notch1−/− mice is similar to that of Dll4−/− embryos.2,18,19 From the analysis of endothelial N4ICD in the eJag1 mutants, where endothelial N4ICD positive area is higher in eJag1OE mutants and lower in eJag1cKO, it seems that Jagged1 is able to activate Notch4 in the endothelium, as previously suggested.20,21 Moreover, the vasculature of eDll4OE mutants, where Jag1 expression is high, has increased numbers of N4ICD positive ECs, which are not observed when Jagged1 signaling is blocked. Additionally, blocking of Dll4 in either control (control+anti-Dll4) or eJag1cKO (eJag1cKO+anti-Dll4) mice also led to decreased endothelial activation of Notch4, which was expected because of the downstream downregulation of Jag1 transcript levels that is observed in these mice.
Furthermore, we have also showed that administering a Notch4 agonistic antibody to WT mice accelerated the healing response, similar to what was observed in eJag1OE mutants. Vasculature analysis of the injected mice showed no alteration in vascular density but sustained increase in vSMC and pericyte coverage, suggesting this receptor to be mainly involved in the vascular maturation response. Therefore, these results support the hypothesis that the proangiogenic phenotype observed in eJag1OE mutants can be a consequence of, first, inhibiting Notch1 activation, resulting in increased angiogenic growth, and second, activating Notch4, which in turns drives vascular maturation. In addition, the immunofluorescence and transcription analysis of both eJag1 mutants and Notch4 agonist–administered mice suggest that activation of Notch4 by endothelial Jagged1 results in the transcription of specific target genes, namely Hey1.22 Conversely, the other major Notch effector, Hey2, was upregulated in eJag1cKO and in eDll4OE mutants, which in both cases is likely to be a consequence of increased Dll4/Notch1 signaling.
Evidence for a role of Notch and, more specifically, Jagged1 in endothelial contact-dependent recruitment of smooth muscle progenitor cells in vivo was provided by the deletion of Notch signaling activity in neural crest–derived smooth muscle progenitors23 and in experiments where Jag1 was deleted specifically in ECs.10 Thus, endothelial Jagged1 is also able to bind Notch3 on neural crest–derived smooth muscle progenitors, which leads to the lateral induction of Jagged1/Notch signaling and directs the expression of HeyL and other signals laterally into the growing circumferential wall, as suggested by Liu et al.11 Here, we show the ability of endothelial Jagged1 to activate perivascular Notch3. We have detected increased levels of N3ICD in SMCs in both eJag1OE and eDll4OE mutants. Moreover, administration of anti-Jagged1 to eDll4OE mutants brought N3ICD levels back to control levels. Most importantly, these results suggest that Jagged1/Notch4 endothelial signaling can also contribute to the assembly of a SMC layer by regulating the transcription of key components of pathways associated with vascular maturation and perivascular cell recruitment. In particular, these data point at the ANG-1/TIE-2 and PDGF-B/PDGFRβ pathways as possible targets of Jagged1/Notch signaling, as has been previously suggested.24 Nonetheless, further studies are required to distinguish the roles of Jagged1/Notch4 and Jagged1/Notch3 signaling in the process of vascular maturation.
In our proposed model (Figure 8), endothelial Jagged1 acts downstream of Dll4/Notch1 to produce 2 distinct effects. First, Jagged1 is responsible for antagonizing Dll4 ability to bind to and activate Notch1 in tip cells, creating a negative feedback loop in the regulation of endothelial branching, as described in the developing retina.9 Second, by activating Notch4 in ECs and Notch3 in vSMCs, Jagged1 positively regulates vascular maturation downstream of Dll4/Notch1 signaling.
In conclusion, this study is the first to demonstrate the proangiogenic role of endothelial Jagged1 in adult physiological angiogenesis and the synergistic roles of endothelial Jagged1 and Dll4 on vascular maturation. Also, it is the first study to reveal Jagged1 as a potential therapeutic target in wound healing. We have previously presented results showing that low dosage inhibition of Dll4 function could produce a proangiogenic phenotype, while maintaining an intact network of functional blood vessels.3 This approach might enable faster wound healing and regenerative processes. The current results establish that increased endothelial expression of Jagged1 has a more profound proangiogenic effect that promotes healing processes by increasing vessel density but also improved maturation of the newly formed wound vasculature. This goes way beyond what Dll4 targeting/inhibition alone can achieve. Activation of Jagged1 expression may well prove potentially useful in situations where vascular function is a limiting factor in patient recovery, like in ischemia or wound healing. In the case of diabetic retinopathy, increased Jagged1 signaling could be used to stabilize and promote the maturation of nascent blood vessels, which could help to reduce edema and hypoxia of retinal tissues in this condition. However, blocking Jagged1 is likely to be useful in cases where it is desirable to limit vascular growth such as in antiangiogenic cancer therapy.
We thank Tomás Lacerda for the contribution in the design of the proposed model. We would also like to thank S. Adams for technical and management assistance and S. Volkery for assistance in confocal imaging.
Sources of Funding
This study was supported by the Portuguese Foundation for Science and Technology (grants PTDC|/SAU-ONC/116164/2009, PTDC/SAU-OSM/102468/2008, PTDC/CVT/115703/2009 and by the individual PhD grant SFRH/BD/44964/2008). R. Diéguez-Hurtado has received funding from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme FP7/2007 to 2013/under REA grant agreement no. 317250. CIISA has provided support (Project PEst-OE/AGR/U10276/2014, funded by FCT).
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.114.304741/-/DC1.
- Nonstandard Abbreviations and Acronyms
- endothelial cell
- endothelial-specific Jag1 loss-of-function or conditional knockout
- endothelial-specific Jag1 gain-of-function or conditional overexpression
- endothelial-specific Dll4 gain-of-function or conditional overexpression
- vascular endothelial growth factor
- vascular smooth muscle cell
- Received September 29, 2014.
- Accepted February 27, 2015.
- © 2015 American Heart Association, Inc.
- Duarte A,
- Hirashima M,
- Benedito R,
- Trindade A,
- Diniz P,
- Bekman E,
- Costa L,
- Henrique D,
- Rossant J
- Trindade A,
- Djokovic D,
- Gigante J,
- Badenes M,
- Pedrosa AR,
- Fernandes AC,
- Lopes-da-Costa L,
- Krasnoperov V,
- Liu R,
- Gill PS,
- Duarte A
- Xue Y,
- Gao X,
- Lindsell CE,
- Norton CR,
- Chang B,
- Hicks C,
- Gendron-Maguire M,
- Rand EB,
- Weinmaster G,
- Gridley T
- Li L,
- Miano JM,
- Cserjesi P,
- Olson EN
- Shutter JR,
- Scully S,
- Fan W,
- Richards WG,
- Kitajewski J,
- Deblandre GA,
- Kintner CR,
- Stark KL
- High FA,
- Lu MM,
- Pear WS,
- Loomes KM,
- Kaestner KH,
- Epstein JA
- Liu H,
- Kennard S,
- Lilly B
- Doi H,
- Iso T,
- Sato H,
- Yamazaki M,
- Matsui H,
- Tanaka T,
- Manabe I,
- Arai M,
- Nagai R,
- Kurabayashi M
- Benedito R,
- Trindade A,
- Hirashima M,
- Henrique D,
- da Costa LL,
- Rossant J,
- Gill PS,
- Duarte A
- Swiatek PJ,
- Lindsell CE,
- del Amo FF,
- Weinmaster G,
- Gridley T
- Manderfield LJ,
- High FA,
- Engleka KA,
- Liu F,
- Li L,
- Rentschler S,
- Epstein JA
- Jin S,
- Hansson EM,
- Tikka S,
- Lanner F,
- Sahlgren C,
- Farnebo F,
- Baumann M,
- Kalimo H,
- Lendahl U
Notch signaling pathway is a potential target for therapeutic modulation of angiogenesis. Dll4 and Jagged1 are important Notch ligands, but although Dll4 function has been broadly investigated, Jagged1 role remains elusive namely in the adult organism and outside of the central nervous system. Here, we show for the first time that Jagged1 antagonizes the function of Dll4 in the adult organism and in the context of a physiopathological response such as skin wound healing. These findings are therefore important for potential future therapies, which might inadvertently interfere with essential roles of Jagged1 in tissue homeostasis and regeneration. In addition, an even more important part of our findings demonstrates that crucial downstream effects of Dll4 and Jagged1 in endothelial cells are mediated by distinct Notch receptors, namely Notch1 and Notch4, respectively. These data might well allow the uncoupling of different Notch responses in the angiogenic vasculature, which could prove invaluable for the design of future therapeutic strategies.