.gif?ad=16248&adview=true)
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
REVIEW
Vascular Development Laboratory, Cancer Research UK London Research Institute, London WC2A 3PX, United Kingdom
 |
Abstract |
|---|
[Keywords: Dll4; Notch; angiogenesis; endothelial cell; sprouting]
; Jain 2003; Adams and Alitalo 2007). During fetal and postnatal growth, the size and transport capacity of the vasculature need to expand significantly. Although various mechanisms for the formation of new blood vessels are known, angiogenesis—that is, the expansion of a pre-existing vessel network through a combination of sprouting, proliferation, and remodeling processes—predominates in most settings. In the adult, blood vessels acquire a quiescent, nonangiogenic state but retain considerable growth potential that is activated during wound healing as well as in certain physiological processes. Unfortunately, angiogenesis is also part of the pathogenesis of several human diseases and plays a key role in tumor growth and metastasis. While it is desirable to block the growth of new blood vessels under these circumstances, the controlled stimulation of angiogenesis might be beneficial in ischemic conditions when the local blood supply is impaired.
The Notch pathway is an evolutionary highly conserved signaling machinery with roles in invertebrates as well as in almost every vertebrate organ and tissue. Notch receptors are transmembrane proteins with large extracellular domains containing numerous epidermal growth factor (EGF)-like repeats (Fig. 1). In mammals, four Notch molecules (Notch1–Notch4) interact with five ligands, namely Delta-like 1, Delta-like 3, Delta-like 4, Jagged1, and Jagged2. Like their receptors, Notch ligands are also EGF repeat-containing transmembrane proteins, and they are orthologs of invertebrate Delta and Serrate/Lag-2 proteins, respectively (Fig. 1; Weinmaster 2000; Lai 2004; Le Borgne et al. 2005; Bray 2006; Hurlbut et al. 2007). Several Notch receptors and ligands as well as components of the downstream signaling machinery and various accessory proteins are indispensable for vascular morphogenesis. In this review, we summarize some of the principles controlling blood vessel growth, highlight the function of Notch signaling in the regulation of physiological and pathological angiogenesis, and compare these roles with other processes of tubulogenesis.
|
|
The Notch signaling pathway |
|---|
; Le Borgne et al. 2005; Bray 2006; Hurlbut et al. 2007). The NECD domain contains the large array of (depending on the particular Notch receptor) 29–36 EGF-like repeats, some of which are essential for ligand binding. Three cysteine-rich Notch/LIN-12 repeats help to prevent ligand-independent signaling (Weinmaster 2000; Le Borgne et al. 2005; Bray 2006; Hurlbut et al. 2007). The NICD includes six tandem ankyrin repeats, a glutamine-rich domain, and a C-terminal PEST sequence that facilitates rapid proteolytic degradation of the protein (Fig. 1). The ligands, collectively referred to as DSL (for Delta/Serrate/Lag-2), contain an N-terminal, extracellular DSL motif mediating receptor binding followed by a variable number of EGF-like repeats (eight in mammalian Delta-like and 15–16 in Jagged ligands) (Bray 2006). Jagged/Serrate proteins have an extra cysteine-rich domain with homology with the von Willebrand factor in proximity to the transmembrane region (Fig. 1). Little is known about the DSL cytoplasmic domain except for a conserved C-terminal PDZ-binding motif. Binding and ubiquitination of DSL proteins by the E3 ubiquitin ligases Neuralized (Neur) and Mind bomb (Mib) are essential for ligand activation (Le Borgne and Schweisguth 2003; Le Borgne et al. 2005; Pitsouli and Delidakis 2005; Chitnis 2006). This ubiquitination and the activity of the endocytic ubiquitin-binding protein Epsin are requirements for DSL internalization from the cell surface and Notch signaling (Overstreet et al. 2004; Tian et al. 2004; Wang and Struhl 2004, 2005; Le Borgne et al. 2005; Chitnis 2006).
Upon ligand binding, Notch undergoes two further proteolytic cleavages. Extracellular proteases of the ADAM/TACE/kuzbanian family sever the Notch extracellular domain. Due to this cleavage and/or conformational changes triggered by ligand-dependent trans-endocytosis of the NECD, Notch becomes susceptible to processing by 
-secretase, which releases the NICD (Fig. 1; Weinmaster 2000; Selkoe and Kopan 2003; Schweisguth 2004; Nichols et al. 2007a). Following translocation of the NICD into the nucleus, its interaction with the DNA-binding protein RBP-J [also named CSL after mammalian CBF1, Drosophila Su(H), and Caenorhabditis elegans LAG-1] and its coactivator Mastermind (Mam) leads to the displacement of corepressor proteins and associated chromatin-modifying factors from RBP-J/CSL. This process triggers the transcription of Notch target genes such as the basic helix–loop–helix (bHLH) proteins Hairy/Enhancer of Split (Hes) and Hes-related proteins (Hey), which, in turn, act as transcriptional regulators of further downstream genes (Lai 2002a; Fischer and Gessler 2003; Iso et al. 2003; Bray 2006; Ehebauer et al. 2006). Some evidence suggests that Notch can also signal in a RBP-J/CSL-independent mode (Martinez Arias et al. 2002).
The ankyrin repeats within the NICD are necessary for its signal transduction activity and are also targets for modification (Fig. 1; Struhl et al. 1993). Ubiquitination by the HECT domain family E3 ligases Su(dx) and NEDD4 promotes Notch degradation and thereby negatively regulates signaling (Lai 2002b; Sakata et al. 2004; Wilkin et al. 2004; Le Borgne et al. 2005; Nichols et al. 2007b). In contrast, ubiquitination within the NICD ankyrin repeats by the Ring finger E3 ligase Deltex can antagonize Su(dx) and, depending on cell type and context, promote or decrease Notch activity (Lai 2002b; Le Borgne et al. 2005; Wilkin and Baron 2005; Bray 2006; Nichols et al. 2007b).
Notch receptors are also modulated by glycosylation within several of the extracellular EGF repeats by the glycosyltransferase Fringe in the Golgi (Fig. 1). These modifications inhibit the ability of Notch to be activated by Jagged/Serrate ligands, whereas they potentiate activation by Delta-like proteins (Haines and Irvine 2003; Bray 2006).
Tissue-specific expression of Notch and Delta-like/Jagged molecules, the modulation of their signaling by post-translational modification, and differences in the regulation of downstream target genes can help to explain the many different functional roles of the pathway. For example, initially subtle differences in Notch expression and signaling can single out individual cells from a group of equivalent neighbors and thereby impose differential behavior and differentiation fates (Lai 2004; Schweisguth 2004). Asymmetrical inheritance of Notch regulators, such as of the cytoplasmic adapter protein and Notch inhibitor Numb, is capable of directing the progeny of cell divisions into distinct lineages. Notch signaling can also separate subpopulations of cells and thereby help to generate tissue boundaries (Irvine 1999; Lai 2004; Schweisguth 2004). Dynamic and sometimes even oscillating expression of Notch pathway molecules leads to complex spatiotemporal patterns of Notch function during tissue morphogenesis (Holley and Takeda 2002; Pourquie 2003; Claxton and Fruttiger 2004; Giudicelli and Lewis 2004). Such elaborate, dynamic, and partially overlapping expression of multiple receptors (Notch1, Notch3, and Notch4) and ligands (Dll1, Dll4, Jagged1, and Jagged2) in vascular cells suggests important roles for Notch in the growth and differentiation of blood vessels.
|
Angiogenic sprouting and tubulogenesis |
|---|
; Gariano 2003; Fruttiger 2007). This unusual—compared with most other tissues—growth pattern facilitates the analysis of the initially planar, two-dimensional vasculature in the retina at great detail. Previous work has shown that this rapid and highly directional growth of blood vessels is controlled by a spatial concentration gradient of matrix-anchored vascular endothelial growth factor A (VEGF-A) released by astrocytes in response to local hypoxia (Fig. 2; Stone et al. 1995; Dorrell et al. 2002; Gerhardt et al. 2003). VEGF-A is a very potent chemoattractive signal for endothelial cells (ECs) and promotes the polarized sprouting of specialized ECs, so-called tip cells, at the leading edge of the vascular plexus (Fig. 2). Tip cells lack a lumen, are not perfused by blood, and extend numerous long filopodia to explore growth factor signals and guidance cues in the surrounding tissue (Fig. 2; Dorrell et al. 2002; Ruhrberg et al. 2002; Gerhardt et al. 2003).
|
; Gerhardt et al. 2003), suggesting that the interaction of filopodial processes and, subsequently, cell bodies leads to bridge-like structures in which ECs lack long filopodia and no longer display tip cell features. It may be at this stage that sprouts are converted into new tubules, form anastomoses, and become part of a simple, plexus-like network at the vascular perimeter (Fig. 2). ECs in this plexus are sometimes referred to as "stalk cells" in the literature because, in contrast to tip cells, they lack long filopodia, enclose a luminal space, and thereby form a patent tubule at the stalk of vascular sprouts (Gerhardt et al. 2003). Stalk ECs in the retina also show a high level of proliferation, which supports the growth of the network (Gerhardt et al. 2003; Dorrell and Friedlander 2006). The emergence of new endothelial tip cells from the stalk plexus will promote further extension of the vascular network so that growth of the vasculature is presumably achieved through repeated cycles of EC sprouting and tubulogenesis. In this model of angiogenic growth, sprouting and tubulogenesis are equally essential, but play rather opposite roles and need to be carefully balanced. |
Angiogenic remodeling of blood vessels |
|---|
). The plexus zone in the retinal vasculature has a similar fate. While growth into the periphery continues, the region of the plexus that is closer to the center of the retina is remodeled by pruning and secondary sprouting processes (Fig. 2). Pruning is most pronounced in areas in close proximity to arteries so that the periarterial vasculature is relatively sparse and contains only few branchpoints. In contrast, the perivenous vasculature remodels at a much slower rate, is a site of extensive EC proliferation, retains a higher density of vessels and branchpoints, and contains numerous secondary sprouts emerging from capillaries (Fig. 2). Although it is unclear whether these regional features are caused by hemodynamic factors (i.e., differences in the pressure and speed between arterial and venous blood flow), the concentration of oxygen in the blood stream, differential gene expression in arteries and veins, or a combination of all three, angiogenic remodeling around the arterial branch appears to strongly favor tube formation over new sprouting whereas the perivenous vasculature is permissive for both processes (Fig. 2).
A key step of endothelial tubulogenesis is the conversion of small capillary-like microvessels that contain only one or two ECs in each cross-section into larger tubes in which many ECs make tightly sealed lateral contacts with their neighbors. This circumferential growth occurs both within the arterial and venous branch of the vasculature and crucially increases blood flow toward and from capillary beds, respectively. Defects in angiogenic sprouting frequently lead to a strong increase in vessel diameter, presumably because additional ECs are incorporated into the wall of existing tubes and are no longer directed into new sprouts and branches (Ruhrberg et al. 2002; Nakatsu et al. 2003; Erber et al. 2006). Such phenotypes suggest that the balance between endothelial proliferation and new sprouting might also be a determinant of the normal circumferential growth seen in larger blood vessels and, particularly, in arteries and veins.
|
Regulation of arteriovenous (AV) identity by Notch |
|---|
). Suggesting a role of genetic pathways in AV differentiation, arteries and veins not only differ in their function and morphology but also, even before the onset of blood circulation, in their gene expression profiles. Such artery-specific expression has been reported for several Notch receptors and ligands such as Dll4 in mouse and zebrafish (Shutter et al. 2000; Mailhos et al. 2001; Leslie et al. 2007; Siekmann and Lawson 2007). Lawson et al. (2001) have shown that disruption of Notch signaling due to mutations in the zebrafish mindbomb gene leads to loss of artery-specific markers such as ephrin-B2, a small transmembrane protein and ligand for Eph family receptor tyrosine kinases. In contrast, EphB4, a receptor for ephrin-B2 and normally most prominently expressed on veins, is misexpressed in mindbomb mutant arteries. Conversely, ectopic activation of the Notch pathway represses expression of EphB4 and the VEGF receptor flt4 (VEGFR-3), another venous marker in zebrafish, and thereby imposes a more artery-like molecular profile on veins (Lawson et al. 2001). Changes in the AV gene expression profile in mindbomb mutants are accompanied by vascular malformations. For example, AV shunts between the dorsal aorta and the posterior cardinal vein prevent normal blood circulation in the trunk (Fig. 3; Lawson et al. 2001).
|
, 2001). Gridlock is expressed in the arterial endothelium but also in the lateral plate mesoderm and therefore presumably in the angioblasts that will form the dorsal aorta and cardinal vein (Zhong et al. 2001). The AV fate of these EC precursors appears to be predetermined in a Notch-dependent fashion. Knockdown of gridlock suppresses the expression of arterial markers (such as ephrin-B2) and leads to an expansion of the contiguous regions of the vein. In contrast, gridlock overexpression suppresses veins without increasing arteries, suggesting that the role of gridlock is the repression of venous differentiation rather than the induction of arterial differentiation (Zhong et al. 2001).
Zebrafish genetics also allowed some insight into the genetic regulation upstream of Notch. The morphogen Sonic Hedgehog (Shh) is responsible for the induction of VEGF-A in the somitic tissue close to the aorta. In turn, this expression of VEGA-A up-regulates Notch pathway components in the aorta and thereby controls AV differentiation (Lawson et al. 2002).
Genetic experiments in mice have shown that the orphan nuclear receptor COUP-TFII (encoded by the gene Nr2f2) promotes venous EC differentiation by suppressing Notch signaling (You et al. 2005). Consequently, veins acquire arterial features in Nr2f2 knockout mice, whereas arterial markers are suppressed by EC-specific overexpression of the orphan receptor (You et al. 2005).
Studies with various Notch reporters have shown that arteries are also major sites of Notch signaling in mice (Duncan et al. 2005; Vooijs et al. 2007), and arterial specification is defective in various Notch pathway mutants. For example, inactivation of the murine Rbpsuh gene (which encodes RBP-J) or the two transcription factor genes Hey1 and Hey2 leads to defects in vascular remodeling, hemorrhaging, disruption of the dorsal aorta, and strong reduction of arterial markers (Fischer et al. 2004; Krebs et al. 2004). Similar results have been reported in knockout mice lacking Mindbomb1 (which is essential for DSL activation), protein O-fucosyltransferase 1, Jagged1, Notch1, or both Notch1 and Notch4 (Xue et al. 1999; Krebs et al. 2000; Shi and Stanley 2003; Koo et al. 2005). The Dll4 gene, which is expressed in arterial ECs, is haploinsufficient, so that even heterozygosity causes malformation of arteries, including stenosis and atresia of the aorta, defective arterial branching, and loss of arterial markers such as ephrin-B2 (Gale et al. 2004; Krebs et al. 2004).
|
Notch controls arteriogenesis in the adult |
|---|
). This enlargement of pre-existing vessels is initiated by physical forces such as increased shear stress, but it also involves inflammatory cells, cell proliferation, and the remodeling of the extracellular matrix (Fig. 4; Heil et al. 2006). Arteriogenesis is probably not only relevant for tissue repair processes but may be also part of the vascular expansion program triggered by tissue growth or physical exercise.
|
). In the vasculature, Dll1 is largely confined to the arterial endothelium, and expression is strongly up-regulated during ischemia-induced arteriogenesis (Beckers et al. 1999; Limbourg et al. 2007). Postnatal functional studies in null mutants are precluded by the embryonic lethality of Dll1 knockout mice (Hrabe de Angelis et al. 1997). However, Dll1+/– heterozygotes, which survive to adulthood, show reduced formation of collateral arteries and fail to restore blood flow into the periphery in a hindlimb ischemia model (Limbourg et al. 2007). Loss of a single Dll1 allele is also sufficient to prevent ischemia-induced activation of Notch signaling and up-regulation of ephrin-B2 expression in arteries. In cultured cells, Dll1 expression is increased in response to VEGF-A and fibroblast growth factor-2 (FGF-2) and controls EC tube formation through ephrin-B2 (Limbourg et al. 2007). Recovery of blood flow in the hindlimb ischemia model is also compromised in Notch1 heterozygous knockout mice but not in Notch4-null mutants (Takeshita et al. 2007).
Vascular smooth muscle cells (vSMCs), contractile and matrix-producing support cells that are associated with arteries and, less prominently, with veins, also play critical roles in vascular maturation and arteriogenesis (Heil et al. 2006). In adult Notch3 knockout mice, some arteries are dilated, covered by a thinner than normal smooth muscle cell coat, and vSMCs fail to properly integrate into the vessel wall (Domenga et al. 2004). The main role of Notch3 in this process is apparently the regulation of vSMC differentiation. Notch3 mutant vSMCs fail to acquire mature markers such as the contractile protein smoothelin, whereas their proliferation and survival appear unaffected (Domenga et al. 2004).
|
Selection of sprouting ECs |
|---|
; Hellström et al. 2007; Hofmann and Iruela-Arispe 2007; Lobov et al. 2007). Spatial analysis of Notch activation by NICD antibodies or in TNR1 transgenic Notch reporter mice reveals a similar mosaic pattern of positive stalk and tip ECs at the vascular front (Hellström et al. 2007). These expression data indicate that ECs within the angiogenic vasculature of the retina are heterogeneous with respect to Notch signaling but offer few clues as to how this might be linked to specific functions within the angiogenic program.
More insight into the role of Dll4–Notch signaling was provided by genetic studies in mouse and fish. Although the functional characterization of Dll4 in postnatal mice is complicated by the embryonic lethality and haploinsufficiency of the knockout (Duarte et al. 2004; Gale et al. 2004; Krebs et al. 2004), the number of tip cells and branchpoints, endothelial proliferation, and the density of the retinal vasculature are all significantly increased in the fraction of surviving heterozygotes (Hellström et al. 2007; Lobov et al. 2007; Suchting et al. 2007). Molecular markers that are strongly, albeit not exclusively, expressed by tip cells, such as transcripts for the growth factor Pdgfb or the netrin receptor Unc5b, are up-regulated in the Dll4 mutant retinal endothelium (Hellström et al. 2007; Suchting et al. 2007). A similar phenotype of enhanced tip cell formation and endothelial sprouting results from the administration of recombinant, soluble Dll4/Fc protein, which acts as an antagonist of endogenous Dll4–Notch interactions, anti-Dll4 antibodies, or pharmacological blockage of Notch cleavage with the -secretase inhibitor DAPT (Ridgway et al. 2006; Hellström et al. 2007; Lobov et al. 2007; Suchting et al. 2007).
Studies in zebrafish have provided further insight into the role of Notch. In zebrafish embryos, intersegmental vessels sprout from the dorsal aorta and grow through the gaps between the somites into the dorsal trunk. Subsequently, a series of vessel fusion and pruning processes will give rise to a network of alternating arteries and veins that mediate the dorsoventral transport of blood (Fig. 3; Isogai et al. 2003). Blocking Dll4 expression or Rbpsuh-mediated Notch signaling during this process leads to enhanced sprouting, an increase in the number of tip cells, and aberrant branching, similar to what has been observed in the mouse retina (Leslie et al. 2007; Siekmann and Lawson 2007). Finally, time-lapse videomicroscopy of growing intersegmental vessels has directly demonstrated in vivo that the knockdown of Rbpsuh or Dll4 expression enhances the explorative behavior of endothelial sprouts (Leslie et al. 2007; Siekmann and Lawson 2007).
To address the question of whether the suppression of the tip cell phenotype requires cell-autonomous Notch activity, Siekmann and Lawson (2007) analyzed the distribution of Rbpsuh-deficient cells in mosaic fish embryos. Consistent with a model in which Notch promotes a quiescent, nonsprouting phenotype, ECs lacking Rbpsuh are preferentially found at the terminal tip of growing sprouts (Siekmann and Lawson 2007). Hellström et al. (2007) performed a similar analysis in EC-specific, inducible Notch1 mutant mice. In this chimeric setting resulting from the incomplete and patchy inactivation of the gene, Notch1-deficient cells are also preferentially found in tip cells within the retinal endothelium. In addition, studies utilizing three-dimensional in vitro EC culture systems, which permit the analysis of EC function in a simpler (albeit nonphysiological) context, had shown previously that ectopic Notch activation blocks VEGF-A-induced endothelial sprouting in a cell-autonomous fashion (Leong et al. 2002; Taylor et al. 2002; MacKenzie et al. 2004; Sainson et al. 2005). All these data from zebrafish embryos and the mouse retina together indicate that the activation of Notch signaling imposes a quiescent, nonsprouting phenotype on some ECs whereas adjacent, presumably transiently Dll4-expressing ECs are selected as tips (Fig. 5). This role offers an attractive explanation for the highly mosaic patterns of Dll4 expression and Notch activation in the retinal endothelium (Claxton and Fruttiger 2004; Hellström et al. 2007; Hofmann and Iruela-Arispe 2007; Lobov et al. 2007) and indicates that Dll4 and Notch might help to maintain the integrity of angiogenic blood vessels by confining the capacity to induce and lead new vascular sprouts to a small fraction of cells.
|
|
Cross-talk between VEGF and Notch signaling |
|---|
; Patel et al. 2005; Hainaud et al. 2006; Diez et al. 2007; Lobov et al. 2007), which might also explain why the ligand is concentrated at the periphery of the growing retinal vasculature. The link between VEGF signaling and Dll4 appears to be reciprocal because expression of the VEGF receptor Vegfr2, which is normally concentrated within the sprouting and plexus zone, is up-regulated and transcripts are distributed more widely within the retinal endothelium of Dll4 heterozygotes (Suchting et al. 2007). At the same time, the level of Vegfr1, a receptor that acts as an antagonistic inhibitor of VEGF signaling during development due to its relatively weak signaling activity, is reduced (Suchting et al. 2007). Dll4–Notch signaling also down-regulates VEGFR-2 expression in cultured human ECs (Williams et al. 2006). As further evidence for a direct functional link between Dll4 activity and VEGFR-2, the administration of VEGF-A antagonists such as soluble VEGFR1 (sFlt1) protein, a soluble chimeric receptor called VEGF trap, or antibodies blocking VEGFR2 inhibit tip cell formation and filopodia extension in the retina of Dll4 heterozygotes (Lobov et al. 2007; Suchting et al. 2007).
Despite these striking findings, a model in which Dll4–Notch signaling controls tip cells selection solely through VEGFR-2 expression is too simplistic (Fig. 5). Live video microscopy in zebrafish has shown that tip cells can quickly revert from their invasive phenotype and integrate into a blood vessel (Torres-Vazquez et al. 2004; Leslie et al. 2007). Dll4 is also regulated not only by VEGF-A but also through other pathways including Notch signaling (Shawber et al. 2003; Carlson et al. 2005), which might translate into a very transient pattern of Dll4 expression and activity (Fig. 5). Furthermore, the many connections between Notch and other pathways, such as signaling by TGF
, Hedgehog, or Wnt (Bray 2006; Hurlbut et al. 2007), may contribute to proliferation/apoptosis, cell migration, and branching morphogenesis in the vasculature. Notch has been also linked to the regulation of the cell cycle and, in particular, p21WAF1/CIP1, a regulator of cyclins and cyclin-dependent kinases (Noseda et al. 2004; Devgan et al. 2005; Sarmento et al. 2005), which may in part explain the elevated EC proliferation seen in Dll4 heterozygous retinae.
Tip cell selection within the angiogenic front is unlikely to be the only function of Dll4 in the vasculature. Expression of Dll4 is high in the arterial endothelium of the retina and other tissues (Claxton and Fruttiger 2004; Hellström et al. 2007; Hofmann and Iruela-Arispe 2007; Lobov et al. 2007). Moreover, defective AV marker expression and aorta defects in Dll4 knockout mice (Gale et al. 2004; Krebs et al. 2004) suggest that the ligand is involved in the morphogenesis of these vessels. It is appealing to speculate that high levels of Notch activation in arteries, triggered by Dll4 and perhaps Dll1 and Jagged1, might be at least in part responsible for the low levels of sprouting and endothelial proliferation in these vessels. In response to overexpression of Dll4 in experimental tumors, intratumor blood vessels are straight with relatively few side branches, appear well perfused, and contain a reduced number of proliferating ECs, and endothelial sprouts and filopodia are rare (Noguera-Troise et al. 2006; J.-L. Li and A.L. Harris, pers. comm.). Consistent with a putative role of Notch in tubulogenesis, conspicuously poorly branched and large (described as "dilated") vessels have been reported in mice overexpressing the Notch4 ICD in ECs (Uyttendaele et al. 2001; Carlson et al. 2005). Although this question still requires direct experimental confirmation, we speculate that Notch-mediated suppression of endothelial sprouting might help to promote tubulogenesis in the growing vasculature and thereby integrate these two processes in the angiogenic growth program.
|
Notch controls epithelial sprouting and branching |
|---|
; Ghabrial et al. 2003; Ribeiro et al. 2004). During the branching morphogenesis of the tracheal system in fly larvae, epithelial cells migrate and sprout in a highly stereotypic pattern and can acquire one of three different fates within the network. General tube cells assemble into main tubes and secondary branches. Specialized terminal cells at the ends of these tubes form long cellular, lumen-containing extensions into the target tissues. Due to the segmented organization of insects, a third type of cell, termed a fusion cell, is required to interconnect the tracheal tubes of adjacent body segments (Zelzer and Shilo 2000; Ghabrial et al. 2003). Tracheal cells express the receptor tyrosine kinase Breathless (Btl), and it has been shown that migration, sprouting, and fate determination processes of these cells are critically controlled by the expression of the corresponding ligand, an FGF-like molecule termed Branchless, in the surrounding tissues. Surprisingly, Breathless activity is essential only in the single leading cell within each tracheal sprout whereas the other, trailing epithelial cells can follow this tip cell and form the stalk even when they lack Btl (Ghabrial and Krasnow 2006). Similar to the connection between VEGF-A signaling and Dll4 in the vertebrate vasculature, Branchless induces the expression of the Notch ligand Delta in the tip of growing tracheal sprouts. In Notch mutant Drosophila embryos, many cells try to lead the tracheal branches and compete for the lead position (Ghabrial and Krasnow 2006). Conversely, expression of constitutively active Notch prevents outgrowth, presumably due to the lack of leading tip cells (Ghabrial and Krasnow 2006). The Dll–Notch signal also prevents the stalk epithelium from becoming terminal or fusion cells (Fig. 6; Zelzer and Shilo 2000). The exact fate of tip cells is determined by other signals such as expression of the zinc finger transcription factor Escargot in presumptive fusion cells. In contrast, the long hollow terminal cells can be identified based on the expression of the FGF antagonist sprouty and the transcriptional regulators pointed (pnt) and serum response factor (Blistered/Pruned) (Zelzer and Shilo 2000).
|
; Wan et al. 2000). |
Dll4–Notch signaling in tumor angiogenesis |
|---|
). Suggesting that the important role of Dll4 and Notch signaling is not restricted to only developmental angiogenesis, Dll4 is prominently expressed in tumor blood vessels in a VEGF-dependent fashion (Patel et al. 2005, 2006; Hainaud et al. 2006; Noguera-Troise et al. 2006). Indeed, blockade of Dll4–Notch signaling with the Dll4/Fc antagonist, anti-Dll4 antibodies, or -secretase inhibitor promotes endothelial sprouting and thereby increases vascular density in xenograft tumor models in mice (Noguera-Troise et al. 2006; Ridgway et al. 2006; Scehnet et al. 2007). Surprisingly, inhibition of Dll4 with these strategies yields significantly smaller tumors compared with control animals. The analysis of blood vessel functionality provides an explanation for this seemingly paradoxical effect. Although Dll4 targeting strongly enhances angiogenic growth, the resulting blood vessel network is poorly perfused, and hypoxia in the surrounding tumor tissue is increased (Noguera-Troise et al. 2006; J.-L. Li and A.L. Harris, pers. comm.). These findings demonstrate that the stimulation of nonproductive angiogenesis is a feasible strategy in the fight against cancer, which might in the future complement anti-angiogenic therapeutic approaches.
Two other findings highlight the potential benefit of therapies aiming at the Notch pathway. Firstly, despite the expression of Dll4 in normal tissues, the viability of the treated animals is not compromised by the antagonistic effect of Dll4/Fc fusion protein or anti-Dll4 antibodies (Noguera-Troise et al. 2006; Ridgway et al. 2006). Whereas disruption of Notch signaling with the -secretase inhibitor dibenzazepine (DBZ) disturbs homeostasis in the mouse small intestine, administration of anti-Dll4 antibody twice weekly for 6 wk has no such adverse effects (Ridgway et al. 2006). Even more importantly, Dll4 inhibition reduces the growth of solid tumors that are resistant to VEGF blockade and therefore not susceptible to treatment with avastin or other VEGF antagonists (Noguera-Troise et al. 2006; J.-L. Li and A.L. Harris, pers. comm.). Future work will have to address whether anti-Dll4 therapy is equally successful for other mouse tumor models and in human patients. Given the functional roles of Dll4 in hematopoiesis and T-cell differentiation (Amsen et al. 2004; Radtke et al. 2004), the safety of long-term treatment needs to be investigated in greater detail. It is also unclear whether tumors can acquire resistance against Dll4 disruption; for example, by up-regulating other Notch ligands. In this context, it is noteworthy that Notch responses may be ligand dependent. The overexpression of Jagged1 by carcinoma cells promotes angiogenesis and tumor growth (Zeng et al. 2005).
|
Role of Notch in vascular disease |
|---|
AGS is a congenital disorder caused by mutations in the human gene for Jagged1 (JAG1). Symptoms include abnormalities affecting the liver, heart, eye, and skeleton (Oda et al. 1997). The most frequent cardiovascular anomalies in AGS patients are peripheral pulmonic stenosis, coarctation of the aorta, atrial and ventricular septal defects, and tetralogy of Fallot (Krantz et al. 1999; McElhinney et al. 2002). AGS is caused by JAG1 haploinsufficiency, but the exact role of Jagged1 in the affected tissues is unclear (Spinner et al. 2001). While heterozygous Jagged1 knockout mice do not recapitulate AGS, the additional introduction of a single Notch2 hypomorphic allele leads to Alagille-like developmental abnormalities (Xue et al. 1999; McCright et al. 2002).
CADASIL is a hereditary vascular degenerative disorder caused by mutations in the human NOTCH3 gene. This syndrome is characterized by arteriopathy that affects mainly the small cerebral arteries and leads to stroke and dementia in humans (Joutel et al. 1997, 2004; Kalaria et al. 2004). Consistent with the expression pattern of Notch3 within blood vessels, degeneration and loss of vSMCs are main features of CADASIL and are accompanied by the progressive accumulation of extracellular matrix and the appearance of perivascular GOM deposits (i.e., granular osmiophilic material visible by electron microscopy). These processes cause the thickening of the vessel wall and narrow the lumen of the affected arteries (Chabriat et al. 1995). The genetic defects in CADASIL are typically missense mutations, most of which translate into amino acid exchanges within the first five EGF-like repeats of the NOTCH3 ectodomain (Joutel et al. 1997). It is thought that CADASIL mutations alter the intracellular trafficking and maturation but not the signaling by NOTCH3 (Haritunians et al. 2002; Karlstrom et al. 2002). Smooth muscle-specific overexpression of NOTCH3 containing a CADASIL mutation in transgenic mice reproduces many of the characteristic disease symptoms (Ruchoux et al. 2003).
The etiology of AGS and CADASIL remain insufficiently understood, so that the recent advancements in understanding the role of Notch in the vasculature may allow new insight into the pathogenesis of these important diseases.
|
Perspectives |
|---|
 TopAbstractThe Notch signaling pathway |
|---|
Abstract
The Notch signaling pathway
