Notch: A multi-functional integrating system of microenvironmental signals

Abstract

The Notch signaling cascade is an evolutionarily ancient system that allows cells to interact with their microenvironmental neighbors through direct cell-cell interactions, thereby directing a variety of developmental processes. Recent research is discovering that Notch signaling is also responsive to a broad variety of stimuli beyond cell-cell interactions, including: ECM composition, crosstalk with other signaling systems, shear stress, hypoxia, and hyperglycemia. Given this emerging understanding of Notch responsiveness to microenvironmental conditions, it appears that the classical view of Notch as a mechanism enabling cell-cell interactions, is only a part of a broader function to integrate microenvironmental cues. In this review, we summarize and discuss published data supporting the idea that the full function of Notch signaling is to serve as an integrator of microenvironmental signals thus allowing cells to sense and respond to a multitude of conditions around them.

The Cellular Microenvironment

Conditions of the local environment in which a cell resides can vary widely depending on the species and its anatomical location within the organism. In recent years, cellular microenvironments have gained wide acceptance as major determinants influencing cellular physiology, especially as it pertains to the cancer microenvironment [1], the stem cell niche [2], the vascular system [3], and wound healing/granulation tissue [4]. A multitude of components contribute to a cell's microenvironment. Extracellular matrices which surround and support cells contribute chemical and physical properties to the microenvironment. Both the chemical composition and the physical stiffness of the matrix provide signaling cues that are actively monitored by cells. Neighboring epithelial cells, endothelial cells, leukocytes, and fibroblasts are all known to influence nearby cells chemically through cytokine and hormone secretions, and physically through cell-cell interactions. Other properties of the cellular microenvironment include concentrations of dissolved gases such as O₂ and CO₂, blood sugar concentrations, temperature, shear stress, oxidative stress, the presence/absence of foreign antigens, and osmolality. Moreover, cellular microenvironments can change rapidly and dramatically in response to situations such as wounding and subsequent healing, tumor development, hypoxia, glucose availability, and fibrosis. Due to the potentially dynamic nature of the cellular microenvironment, cellular responses to both static and changing microenvironments need to be calibrated to properly and rapidly respond to these situations. Understanding how cells respond to the incredible complexity of the microenvironment requires a systems biology approach to integrate the microenvironmental information, a task that is immensely complicated.

Notch

The Notch signaling mechanism is a highly conserved developmental pathway that is used during differentiation in numerous tissues in most, if not all, multicellular organisms. Evolutionary evidence for the emergence of the Notch receptors first appears in the choanoflagellates, unicellular flagellated free-living eukaryotic cells widely considered the closest extant protist relative to metazoans [5, 6]. The genome of the choanoflagellate Monosiga brevicollis encodes three domains that show similarity to metazoan Notch receptors (Figure 1A). However, these domains are split amongst three separate transmembrane proteins in the M. brevicollis genome including one gene that encodes 36 epidermal growth factor (EGF) like domains, a second gene that encodes two Lin-12-Notch repeats (LNR domains), and a third gene that encodes six ankyrin repeats [6]. Presumably, these three ancestral partial Notch homologs were responsible for individual functions. This suggests that modern metazoan Notch receptors, which unify these domains in a single receptor, might represent an amalgamation of three independent proteins with independent ancestral functions and may help explain why Notch is capable of integrating a multitude of cellular microenvironmental signals and conditions as described in this review. Despite the lack of a bona fide Notch receptor or Notch ligands, M. brevicollis genome does encode several other components of the Notch system [7] (Figure 1B). Therefore the origin of Notch domains (if not function) likely predates the rise of the metazoans. It has been postulated that these proto-Notch receptors might have served an adhesive function that was independent of Notch signaling activity and is conserved in modern Notch function (reviewed in [8]). It was not until the rise of sponges however, that bona fide Notch receptors and ligands appeared and exhibited the developmental roles that are representative of the metazoan Notch mechanism [9]. Thereafter, Notch receptors, ligands, and other Notch processing and/or modifying proteins are expressed throughout all metazoans examined to date [7].

Figure 1.

Canonical Notch signaling and Notch conservation between human and Monosiga brevicollis. Throughout Figure 1, conservation of Notch proteins or domains between human and M. brevicollis is indicated by green (positive), yellow (unknown), or red (negative) shading according to references [6] and [7]. (A) Conservation of mammalian Notch receptor domains in M. brevicollis. Mammalian (human) Notch receptors contain 36 EGF-like repeats and three LNR or NRR (Lin-12 Notch Repeats or Negative Regulatory Region) repeats in the extracellular domain. The intracellular portion of human Notch contains seven ankyrin domains and a PEST sequence at the C-terminal. For simplicity, the intracellular RAM (RBPj Association Module) domain, two NLS (Nuclear Localization sequence) domains, and TAD (Transactivation Domain) are not shown in this figure. Please refer to references [10, 11] for complete details. Three separate proteins (N1, N2, and N3) in M. brevicollis contain six Ankyrin domains, two LNR domains, and 36 EGF-like repeats respectively [6]. (B) Model of canonical Notch activation mechanism. Notch receptors are modified in the secretory pathway (ER/golgi) by Furin cleavage (S1 cleavage) and glycosylation of EGF-like domains by O-fucosyltransferase (O-fut), Rumi/Poglut1 (Protein O-Glycosyltransferase 1), and fringe family glycosyltransferases. The Furin cleavage products remain non-covalently associated in the membrane where a pulling force initiated by Notch ligand endocytosis in sending cells enables further cleavage by α-secretase (S2 cleavage, NEXT fragment) at the LNR domain, and γ-secretase (S3 cleavage, NICD fragment) in the membrane of receiving cells. Several regulatory proteins including Numb, Notchless, and Deltex control Notch availability at the membrane. After γ-secretase cleavage, the NICD fragment translocates to the nucleus where it displaces the transcriptional co-repressor SMRT from CSL/RBP-jk. NICD participates in a transcriptional complex with CSL, MAML, and p300 to drive transcription of Notch targets such as Hes and Hey genes. NICD steady-state levels are controlled by nuclear export, ubiquitination (Ub) by Sel10, and subsequent degradation in the proteasome.

In mammals, the core of the Notch mechanism consists of five Notch ligands (Jagged1, 2 and Delta like (Dll) 1, 3, 4) present on the “signaling cell”. The Jagged and Dll 1 and 4 ligands directly interact with and activate a transmembrane Notch receptor (four different isoforms in mammals) present on the “receiving cell” [10, 11] (Figure 1B). The Dll3 ligand is a decoy receptor that interferes with Notch activation [12]. Notch receptors undergo a maturation process involving three proteolytic cleavage events that ultimately result in Notch activation. The first cleavage is performed by a furin convertase during translocation through the Golgi complex on the way to the cell membrane [13]. The resulting two Notch fragments remain non-covalently associated at the membrane where canonical Notch activation is initiated by interaction between Notch receptors and ligands. Canonical Notch activation at the membrane is commonly thought to be dependent on a physical tugging mechanism of ~ 4-12 pN [14, 15] that is initiated by Notch ligand endocytosis in the signaling cell [16]. This pulling force sets up a second cleavage by an ADAM metalloprotease (α-secretase) producing the transient NEXT (Notch Extracellular Truncation) fragment [17], and a third cleavage by γ-secretase [18] thus releasing the intracellular NICD domain of Notch that translocates to the nucleus and functions as a co-transcription factor in association with the CSL transcription factor and other co-transcription factors including MAML and p300. In addition to this canonical mechanism, evidence for several non-canonical Notch activation mechanisms have also been gaining traction. In particular, Notch activation that is independent of canonical ligands [19], NICD cleavage and transcriptional activity [20], as well as several non-canonical ligands [21] have all been described in the literature. Finally, in addition to the core receptors and ligands, a wide variety of cellular and secreted proteins have been characterized that modify Notch signaling either through direct interaction and/or modification of extracellular Notch receptor or ligand domains [22] or via post-translational modification of intracellular NICD fragments [23]. References [10, 11] provide excellent in-depth reviews of the Notch signaling mechanism.

Notch as an integrator of cellular microenvironments

While the traditional view of Notch activation focuses on Notch receptor – ligand interactions, it is becoming increasingly clear that Notch signaling is also influenced by a wide array of molecules and events in the cellular microenvironment. In particular, extracellular matrix (ECM) mediated Notch signaling is emerging as a new paradigm for controlling Notch signaling. Regulation of Notch by ECM occurs on several levels, including direct interaction between ECM and Notch receptor/ligands, transcriptional control of Notch receptors and/or ligands, and via cross-talk with other ECM stimulated signaling networks, such as integrins. In addition, Notch is engaged in crosstalk with a number of signaling pathways that are initiated by growth factors and cytokines commonly present in cellular microenvironments including, TGF-β, WNT/β-catenin, and VEGF. Finally, Notch can also be regulated by additional conditions such as shear stress, hypoxia, and hyperglycemia. These microenvironmental conditions are summarized in Table 1. Taking into consideration the wide variety of cellular microenvironmental cues that regulate Notch signaling output, a new picture of Notch is emerging which depicts Notch as an integrating system for the cellular microenvironment, which enables cells to respond appropriately to changing ECM composition, growth factor secretions, oxygen tension, shear stress, and glucose levels. Importantly, this idea is not inconsistent with the classical model of Notch receptor activation by Notch ligands on adjacent cells, but rather builds on this model since cellular neighbors are also an important part of a cell's microenvironment. The goal of this review is to summarize what is known about the role Notch signaling plays in responding to and integrating changing microenvironmental conditions, and to explore and develop the idea of Notch as a multi-functional integrating system of microenvironmental signals.

Table 1.

Basic and proposed mechanisms by which Notch responds to various microenvironmental signals. ECM – Notch (Direct interaction mechanism)

Protein Name	Basic/Proposed Mechanism(s):	Ref.
MAGP-2	-Decrease Notch in RGD/integrin dependent manner -Interacts with EGF-like domains of Notch1 + Jagged1, activates cleavage	[32] [30, 31]
EGFL7	-RGD/integrin dependent Notch activation -Interaction with Notch EGF domains and resultant Notch inhibition	[32] [40]
CCN3	-Binds Notch1 receptor and stimulates Notch receptor activation	[43, 45]
TSP-2	-Binds Notch3, activates Notch3/Jagged signaling	[53]
Syndecans	-Syndecan-2 binds Notch3, activates Notch3/Jagged signaling -Syndecan-3 interacts with Notch1	[54] [62]
Col IV	-Interacts with Notch3, blocks Jagged1 – Notch3 signaling	[55]
Col I	-Interacts with Notch3, blocks Jagged1 – Notch3 signaling	[55]
YB-1	-Binds Notch3 EGF-like repeats, blocks Notch3 but not Notch1 signaling	[59, 60]
Galectin-3	-Binds Notch1 glycosylations, increases NICD processing	[63]

ECM – Notch (Indirect transcriptional mechanism)
Protein Name	Basic/Proposed Mechanism:	Ref.
SPARC	-Suppresses Notch by blocking Notch1 transcription	[64, 65]
Fibulin-3	-Enhances Dll-4 expression, ADAM10/17 activity, and Notch signaling	[69]
Laminin α4	-Increases Dll-4 expression, stimulates Notch	[71]
Laminin-111	-Increases Dll-4 expression, stimulates Notch	[72]
MGP	-Decreases Jagged1 expression, suppresses Notch	[75, 77]

ECM – Notch (Indirect cross-talk mechanism)
Protein Name	Basic/Proposed Mechanism:	Ref.
Reelin	-Stimulates DAB-1 and SRC signaling. Increases Notch activity	[86-88]

Cross-talk with other signaling pathways
Pathway:	Basic/Proposed Mechanism:	Ref.
Integrin	-RGD/RGE control of Notch activity. Possible SRC or ILK involvement -Notch4 / R-Ras mediate increase in β1 integrin affinity	[32, 93, 95] [90, 91]
TGF-β/BMP	-Transcriptional complex between NICD and SMAD3	[110, 111, 118, 119]
WNT	-Transcriptional complex between NICD and β-catenin -β-Catenin interaction with uncleaved Notch receptor -GSK3β mediated phosphorylation of NICD, Notch inhibition -Dishevelled interaction with CSL, Notch inhibition	[128-130] [131-133] [134, 135] [136, 137]
VEGF	-Suppression of VEGFR2 receptors by Dll-4 -VEGF signaling increases Notch1 and Dll-4 expression	[140, 141] [142-144]

Other microenvironmental conditions
Condition:	Basic Mechanism:	Ref.
Shear Stress	-Rapid NICD cleavage, transcriptional control of receptor/ligands, possible involvement of VEGFR2	[167-169]
Hypoxia	-Transcriptional complex between NICD and Hif1α -Activation of γ-secretase -Increase of Dll-4 expression by Hif1α -FIH mediated hydroxylation and destabilization of NICD (Normoxia)	[173] [175] [150, 177] [179, 180]
Hyperglycemia	-VEGF mediated increase in Notch receptor expression and NICD accumulation. -Possible involvement of NICD methylation by CARM1	[183-185] [196]

ECM-Notch interactions

Extracellular matrices are a major component of a cell's microenvironment. In some instances ECM can be stable over decades. In other situations, ECM is rapidly turned over and remodeled. Therefore, cells need to be able to adjust to these stable or changing conditions. Notch responsiveness to the composition of the ECM has only recently begun to be characterized. The interactions between ECM and Notch can be summarized as either direct interactions between ECM and Notch receptors or ligands, indirect (transcriptional) responses of Notch receptors or ligands to ECM, and indirect (crosstalk) interactions between Notch and ECM stimulated signaling cascades (Figure 2). Below, we summarize and discuss these interactions between ECM and Notch signaling.

Figure 2.

Summary of ECM control of Notch signaling. Canonical activation of Notch receptors by Notch ligands can be manipulated in three ways by cellular interactions with ECM. 1.) Direct interactions between Notch receptors or ligands and various ECM molecules can either inhibit or promote activation of Notch signaling. 2.) Indirect interactions between ECM and Notch are characterized by ECM mediated increased or decreased expression of Notch ligands on sending cells or Notch receptors on receiving cells. 3.) Indirect interactions between Notch and ECM are characterized by ECM mediated activation of signaling pathways that post-translationally intersect with Notch proteins or signaling intermediates.

Direct ECM-Notch interactions that control Notch signaling

Direct interactions between Notch receptors and several ECM proteins have been described in the literature. Below, we summarize the current data available for several ECM proteins including Microfibril Associated Glycoprotein-2 (MAGP-2), Epidermal Growth Factor–like 7 (EGFL7), Nephroblastoma Overexpressed (NOV, CCN3), Thrombospondin-2 (TSP-2), syndecans 2/3, collagens type I and IV, the Y-box binding protein (YB-1), and Galectin-3. An interesting observation is that while all these proteins have been shown to regulate Notch signaling via direct interactions with Notch receptors or ligands, there is not a common thread of increased or decreased activity connecting these proteins. Thus, the Notch regulatory activities of these molecules most likely do not rely solely on mechanisms involving a simple steric hindrance model. Moreover, several of these molecules appear to control Notch via multiple mechanisms, suggesting that ECM control of Notch may be a highly regulated activity.

MAGP-2 is a component of elastic fibrils that are thought to help recruit tropoelastin to fibrillin containing microfibrils during the development of elastin networks [24]. Since elastin is critical for Windkessel function and structural integrity of the aortic wall [25], it is not surprising that loss of function of MAGP-2 is linked to aortic dilation in mice [26] and familiar thoracic aortic aneurisms in humans [27]. MAGP-2 may also serve other functions in the cardiovascular system since MAGP-2 contains an αvβ3 integrin binding RGD domain and has been shown to control angiogenesis [28] and vascular density in ovarian cancers [29]. The link between MAGP-2 and Notch was first made when MAGP-2 was identified as a Jagged1 interacting protein by yeast-two hybrid screening [30] and was shown to induce Jagged1 shedding from the cell surface [31]. Subsequent analysis determined that MAGP-2 and the related protein MAGP-1 both increased Notch signaling in COS cells [31]. Mechanistically, MAGP-2 mediated stimulation of Notch signaling was shown to involve direct binding between the MAGP-2 C-terminal domain, and the EGF-like domains of Notch1 and Jagged1 [31]. In addition, RGD→RGE mutation of the MAGP-2 integrin binding domain converted MAGP-2 from a suppressor to an activator of Notch signaling in endothelial cells suggesting that MAGP-2 may also regulate Notch via interactions with integrins [32]. This finding may help to explain the cell type dependent effects of MAGP-2 on Notch signaling previously observed [33] and suggests that cell type specific control of Notch may be dependent on several factors including integrin and Notch ligand expression profiles.

EGFL7 is a secreted protein that is specifically expressed from endothelial cells during development [34, 35]. EGFL7 is predominantly found in the vascular microenvironment where it appears to be an important regulator of elastogenesis [36] and angiogenesis [37, 38]. In particular, EGFL7 is important for the formation and maintenance of vascular lumen structures [34, 39] and suppressing angiogenic sprouting [38]. The first observations that EGFL7 could control Notch signaling were made in neural stem cell cultures where it was found that the N-terminal half of EGFL7 specifically interacted with EGF domains in Notch1-4 and inhibited Notch signaling [40]. Subsequent work showed that EGFL7 control of Notch in endothelial and placental trophoblast cells was important for placenta development and that decreased EGFL7 expression may be linked to preeclampsia [41, 42]. In addition to controlling Notch via direct interaction with Notch receptors, recent work showed that RGD→RGE mutation of the EGFL7 integrin-binding domain enhanced Notch signaling in endothelial cells [32]. By regulating Notch via direct interactions with Notch receptors and via RGD integrin binding, EGFL7 demonstrates similarities with MAGP-2 and suggest that dual control of Notch by ECM molecules is a common theme.

In addition to MAGP-2 and EGFL7, CCN3 (NOV) has also been implicated in the regulation of Notch signaling [43]. CCN3 belongs to the ECM CCN family of proteins (CCN1-6) that share a modular structure including of conserved cysteine knot C-terminal (CT) domain and are multi-functional regulators of diverse processes including development, osteogenesis, and angiogenesis [44]. The Notch regulatory activity of CCN3 appears to be important for controlling a variety of activities including osteoblast differentiation [45, 46] and trophoblast senescence [47, 48]. Regulation of Notch signaling may be a general feature of the CCN family since CCN2 (CTGF) suppresses Notch signaling [49] and CCN1 (Cyr61) is linked to suppression of Notch1 during the epithelial to mesenchymal transition (EMT) [50]. Although, mechanistic details describing how CCN3 manipulates Notch signaling are lacking, the cysteine rich C-terminal tail of CCN3 binds to Notch1 [43] and is required for Notch regulation [45]. Similar to MAGP-2 and EGFL7, CCN3 also interacts with several integrins (in an RGD independent manner) [51] although it is unknown if control of Notch by CCN3 involves integrin ligation. Interestingly however, CCN1 is highly expressed near developing blood vessels where it enhances Notch signaling in an integrin dependent manner [52].

Although MAGP-2, EGFL7, and CCN3 are the best characterized examples of ECM proteins known to regulate Notch activity via direct Notch receptor and/or ligand interactions, several other ECM molecules have also been implicated in the Notch pathway and appear to control Notch via direct interactions with Notch receptors and/or ligands. A common thread among these molecules is Notch3, which appears to be frequently targeted by ECM interactions compared to other Notch receptors. For instance, Thrombospondin-2 (TSP-2) and Syndecan-2 specifically interact with Notch3 and promote Notch3 – Jagged1 signaling [53, 54]. Conversely, collagen type I and IV also bind to Notch3 and Jagged1 but suppress downstream Notch signaling [55]. An additional example of Notch3 regulation by microenvironment is YB-1. The multi-functional YB-1 protein has widespread DNA/RNA binding activities [56] and has historically been thought of as a cold shock protein [56]. Interestingly however, YB-1 can be secreted from mesangial and immune cells after cytokine stimulation via a non-classical mechanism that involves ubiquitin E3 ligase HACE-1 mediated K27 ubiquitination and association with the Tumor Susceptibility Gene 101 (TSG101) [57, 58]. In turn, secreted YB-1 has been found to specifically interact with Notch3 EGF repeats and to control Notch3 downstream signaling, but not Notch1 signaling [59, 60]. More recently, non-secreted YB-1 was found to control Notch4 expression in triple negative breast cancer cells suggesting that YB-1 may control Notch on multiple levels [61]. Finally, while Notch3 appears to be a common target for many ECM molecules, it is not the only target. For instance, while Syndecan-2 regulates Notch3 – Jagged1 signaling [54], Syndecan-3 interacts with Notch1 receptor, regulates processing by ADAM17/TACE, and is required for Notch signaling activity in skeletal muscle progenitor cells [62]. Finally, the sugar binding protein Galectin-3 has been reported to directly interact with Notch1 in a sugar-dependent manner and to activate downstream Notch signaling without affecting expression of Notch1 receptor [63].

Collectively, these examples demonstrate that a diverse array of ECM molecules can influence Notch utilizing a wide variety of mechanisms. Given that each of these molecules exhibits tissue and/or temporal specific expression patterns, these examples serve as a dynamic illustration of how Notch responds to changing ECM microenvironments. With this understanding, it will be interesting to see how future work refines our understanding of ECM – Notch interactions.

Indirect ECM-Notch interactions that control Notch signaling (transcriptional mechanisms)

Each of the examples described thus far involve matricellular control of Notch that appears to be mediated at least in part by direct protein interactions with Notch receptors and/or Notch ligands. However, other matricellular proteins control Notch activity in a less direct manner by influencing the expression of Notch signaling components. For instance, the SPARC protein (Secreted Protein, Acidic, and Rich in Cysteine) stimulates differentiation of medulloblastoma cells by suppressing Notch signaling [64]. However, instead of direct interaction with Notch receptors or ligands, SPARC seems to transcriptionally control Notch signaling since SPARC null osteoblasts express increased Notch1 protein [65] and SPARC protein transcriptionally suppresses Notch1 expression [64]. In comparison to SPARC, which seems to control Notch via direct manipulation of Notch1 expression, other ECM proteins such as Fibulin-3, basement membrane laminins, and MGP influence Notch signaling by controlling expression of Notch ligands. Fibulin-3 is a member of the fibulin family of extracellular matrix glycoproteins that are characterized by tandem repeats of calcium binding EGF sites and a C-terminus fibulin-type module [66]. Fibulin proteins are commonly misregulated during cancer and have emerged as important microenvironmental regulators of cancer and tumor angiogenesis [67]. In particular, Fibulin-3 has emerged as a biomarker for pleural mesothelioma and malignant glioma where Fibulin-3 appears to enhance glioma malignancy by stimulating tumor cell motility and invasion [68]. Fibulin-3 also enhances tumor angiogenesis in glioma by increasing endothelial expression of the Notch ligand Dll-4 and simultaneously stimulating ADAM10/17 activity and downstream Notch signaling [69]. An interesting observation however, is that Dll-4 has been extensively shown to limit branching angiogenesis by suppressing the endothelial tip cell phenotype [70]. As an example of this idea, basement membrane laminins including Laminin-α4 and Laminin-111 promote Notch activation by increasing Dll-4 expression via interaction with integrins [71, 72]. As opposed to Fibulin-3 however, Dll-4 induction by these laminins appears to be an important mechanism to maintain endothelial quiescence by limiting tip cell behaviors. Thus, perhaps simultaneous regulation of ADAM10/17 and Dll-4 enables Fibulin-3 to behave as an angiogenic promoter in glioma, but to inhibit angiogenesis in other tumors as previously described [67]. Finally, Matrix Gla Protein (MGP) is a well known inhibitor of vascular calcification [73] that functions by binding to and suppressing the osteogenic programs initiated by BMP-2 and other BMPs [74]. In addition to suppressing vascular calcification, MGP has additional roles in the vasculature since MGP deficiency in mice leads to increased vascular densities, enhanced tumor angiogenesis [75], and the development of arteriovenous malformations [76]. Mechanistically, MGP deletion results in increased Notch signaling via enhanced expression of the Notch ligand Jagged1 [75] and accordingly, deletion of a single Jagged1 or Jagged2 allele in MGP knockout animals suppresses arteriovenous malformations [77]. Although it is not yet clear how MGP controls Jagged1 expression, it appears that MGP expression is also controlled by Notch in shear-stressed aortic valve endothelium, [78] suggesting that Notch and MGP are coordinated by a feedback regulation.

An important observation is that many of the ECM proteins discussed above not only control Notch signaling, but have also been implicated in the matricellular control of angiogenesis. Indeed, MAGP-2, EGFL7, the CCN family of proteins (CCN1, 2, 3), Thrombospondin-2, Syndecan-2, SPARC, collagens I and IV, Fibulin-3, MGP, and laminins have all been characterized as angiogenic regulators. Given that Notch has emerged as a major regulator in the cardiovascular system (discussed below), matricellular control of Notch activity may be a common mechanism whereby the vascular microenvironment exerts control over angiogenic activity. Hopefully, future research will be able to determine the relative contributions of these matrix molecules towards Notch regulation during angiogenic processes and begin to understand how these multiple signals integrate to control Notch.

Direct ECM-Notch interactions that control Notch signaling (crosstalk mechanisms)

Reelin is a secreted glycoprotein that is an important regulator of neuronal cell migration in the developing brain, [79, 80] and provides one last mechanism to demonstrate how ECM molecules may control Notch. Deletion of Reelin in mice causes an abnormal “reeling” gait referred to as a Reeler phenotype [81]. Reelin has several cell surface receptors including the VLDLR and ApoER2 lipoprotein receptors on neuronal cells that have been described by several investigators [82-84]. In addition, Reelin has also been described to interact with integrins α3β1 and α5β1 [80, 85]. Downstream from these receptors, Reelin signaling typically propagates through Disabled-1 (DAB-1) phosphorylation and downstream PI-3K, AKT, and SRC signaling mechanisms [83-88]. Early work by Keilani et al [86] and Hashimoto-Torii et al [88] suggested that Notch may be important for activities downstream of Reelin. For example, Reelin induces a radial glial phenotype in human neural progenitor cells, and this effect is phenocopied by activation of the Notch signaling cascade [86]. Moreover, the Notch NICD domain is sufficient to rescue abnormal migration in neurons from reeler mice [88]. Mechanistically, Reelin does not appear to directly interact with, or control the expression of Notch receptors and/or ligands. Instead, Reelin appears to control Notch via manipulation of downstream signaling networks. For instance, it has been shown that the downstream Reelin signaling intermediate, DAB-1, physically associates with NICD [86], that Notch works through DAB-1 to regulate axon guidance in Drosophila [89], and that DAB-1 phosphorylation and SRC activity are essential for Notch1 activation by Reelin [87]. Taken together, these results suggest that Reelin may regulate Notch via a mechanism independent of Notch expression or Notch processing, but dependent on downstream DAB-1 and/or SRC kinase activities. Although further research is required to confirm this, the molecular interaction between Notch and SRC is further explored in the Integrin/Notch crosstalk section below.

Notch crosstalk with other signaling networks. Integrins, TGF-β, WNT, and VEGF

An important distinction between Notch regulation by Reelin compared to other molecules mentioned in the previous section is that Reelin does not depend on direct Notch receptor/ligand interactions nor on transcriptional control of individual Notch signaling components. Instead, the evidence supports a mechanism whereby Reelin interaction with its cell surface receptors triggers downstream signaling (DAB-1 and SRC) that then regulates Notch via undefined mechanisms. Thus, Reelin serves as an example of how Notch signaling can be influenced by crosstalk with other signaling pathways. Similarly, crosstalk between Notch and several other signaling mechanisms initiated by molecules including integrins, TGF-β/BMP, VEGF, and WNTs in the cellular microenvironment (Figure 3) have been described and are discussed below.

Figure 3.

Crosstalk between Notch and other signaling pathways. Crosstalk between WNT and Notch occurs on several levels including the formation of a β-catenin-NICD transcriptional complex, interaction between Notch receptors and β-catenin at the membrane, phosphorylation of NICD by GSK3β, and inhibitory interactions between Dishevelled and CSL. The mechanistic interaction between integrins and Notch is poorly characterized, but existing evidence suggests ubiquitination and/or phosphorylation of NICD by SRC and ILK kinases. Interaction between the Notch and VEGF pathways involves the reciprocal transcriptional regulation of Notch ligands by VEGF, and VEGFR2 by Notch. Notch/TGF-β, or Notch/BMP crosstalk occurs downstream of ALK (TBR1/TBR2 or BMPR1/BMPR2) receptors and is dependent on RSMAD and Co-SMAD activation and subsequent formation of a SMAD/NICD transcriptional complex similar to the β-catenin/NICD complex.

Crosstalk between Notch and Integrins

The earliest evidence that integrins and Notch coordinate comes from studies which explored the effect of Notch on integrin activation. For instance, Leong et al. demonstrated that Notch4 activation in microvascular endothelium increased β1 integrin affinity for collagen [90]. This was taken one step further by Hodkinson et al., who demonstrated that activation of the small GTPase R-Ras by Notch1 resulted in increased β1 affinity for collagen [91]. Subsequent work on this topic began to uncover the reciprocal interaction wherein integrins also exert control over Notch. Initially, integrin control of Notch was focused on transcriptional regulation of Notch receptors or ligands. For example, work by Weijers et al. [92] described an effect of integrin blocking low molecular weight fibronectin fragments on the expression of the Notch ligand Dll-4 and subsequent Notch activation in endothelial cells. Similarly, Estrach et al. [72] and Stenzel et al. [71] demonstrated that Laminin 111 and Laminin α4 increase Dll-4 expression in endothelial cells via α2β1 and α6β1 integrins, and that disruption of this signaling system had dramatic complications for normal angiogenesis. While these studies suggested a functional coordination of ECM, integrins, and Notch they did little to dissect a molecular mechanism, beyond transcriptional control, through which coordination occurs. A handful of reports however have suggested that Notch control by integrins is not restricted to transcriptional regulation, but rather, may also engage Integrin Linked Kinase (ILK) and/or SRC signaling downstream from activated integrins. For instance, Mo et al. [93] observed that ILK decreased Notch signaling by stimulating ubiquitination and rapid degradation of the NICD fragment. Similarly, Suh et al. [94] showed that collagen type I increased ILK signaling and NICD accumulation through interaction with α2β1 integrin. In addition to ILK, the non-receptor tyrosine kinase SRC which is commonly activated by integrin ligation, may also regulate Notch. As eluded to above, Reelin has been shown to control Notch in a DAB-1 and SRC dependent manner [87]. Although a molecular interaction between Notch and SRC was not explored in this study, the authors did show that SRC inhibitors did not affect expression of Notch1, suggesting a more direct Reelin-SRC-Notch interaction. In support of this, SRC was found to be an important regulator of Notch S1 processing by furin and that the kinase domain of SRC binds to and phosphorylates the ankyrin domain of active NICD [95]. Moreover, a genetic interaction between SRC and Notch has been uncovered during Drosophila development that is critical for normal eye formation [96]. Taken as a whole, these publications show that Notch can control integrin adhesion (i.e. inside out signaling), and that integrins can control Notch (i.e. outside in signaling). Therefore, these data suggest that integrins and Notch are coordinated into a cellular signaling network that involves feedback control between Notch and integrins and may involve ILK and/or SRC signaling.

The implications of integrin/Notch crosstalk are potentially quite numerous. In particular, one field of research that may be impacted by this crosstalk is the study of pathological tissue fibrosis. Fibrotic diseases are defined by excessive deposition of fibrotic ECM molecules, increased tissue stiffness, and can occur in most any tissue although fibrosis of the liver, lung, kidney, and heart represent the major impacts of fibrosis on human health. Given the increased ECM present in fibrotic tissues, it is not surprising that integrins figure predominantly in the pathology of fibrosis [97-99]. Adding to this, it has become apparent that Notch is also an important regulator of fibrosis in the lung, liver, kidney, and skin [100, 101]. For example, strong expression of Notch was observed in myofibroblasts, the pathological cells associated with the progression of fibrosis, in lung specimens from patients with idiopathic interstitial pneumonias and in bleomycin-induced pulmonary model of fibrosis [102]. Moreover, in airway subepithelial fibrosis, the Notch pathway stimulated the promoter activity of collagen type I through a Hes1-dependent mechanism [103]. In the kidney, Bielesz et al. showed that upregulation of Notch pathway components (Jag1/Notch1/HeyL) regulated the development of tubulointerstitial kidney fibrosis in mice and humans [104]. In the liver, the number of Notch1, Notch3 and Notch4 positive cells were highly upregulated in CCL4 induced fibrosis [105]. Moreover, activated hepatic stellate cells (HSC) showed an increased expression of Notch2, Notch3, Hey2 and HeyL [106]. However, after blocking with the γ-secretase inhibitor DAPT, activated HSC reversed back to quiescent HSC [106] and attenuated hepatic fibrosis [105]. Collectively, these examples clearly illustrate the importance of Notch signaling during fibrotic responsis. Given the crosstalk between integrins and Notch, it will be interesting to determine if integrins have a strong impact on Notch mediated fibrosis.

Notch and TGF-β

The TGF-β superfamily encompasses more than 30 ligands including TGF-βs, BMPs, activins/inhibins, and Mullerian Inhibiting Substance (MIS) that specifically interact with at least seven ALK receptors. Activation of ALK receptors by TGF-β or other ligands stimulates SMADs to translocate to the nucleus where they coordinate transcriptional responses (reviewed in [107]). TGF-β and Notch signaling are both involved in several physiological and patho-physiological processes including embryonic development, wound healing, cancer, and fibrosis. Several lines of evidence indicate that TGF-β and Notch can engage in crosstalk (reviewed in [108, 109]). The first molecular evidence for this interaction was revealed in a series of papers showing that Notch is synergistic with both TGF-β and BMP signaling. Specifically, Blokzijl et al. demonstrated that NICD can form a transcription factor complex with SMAD3, an intracellular transducer of TGF-β signaling, in chicken embryos and in mouse myoblast C2C12 cells [110]. In this study, it was also observed that TGF-β upregulated the expression of Hes-1, a Notch target gene, and the effect was abolished by using a dominant negative form of CSL [110]. A similar interaction was observed in mouse regulatory T cells in which NICD cooperates with activated SMAD3 and accelerates its nuclear translocation [111]. The importance of TGF-β/Notch crosstalk is illustrated by several reports showing that Notch activity is required for some TGF-β effects such as TGFβ-induced EMT [112] and the well-known pro-fibrotic activity of TGF-β [105, 113-115]. Finally, although the majority of interactions between TGF-β and Notch appear to be synergistic, this may be an oversimplified view of the TGF-β/Notch interaction. In support of this, Fu et al [116], found that while Notch did enhance expression of some TGF-β responsive genes including PAI1, CTGF, and CYR61, other TGF-β responsive genes including ID1, and ID2, were decreased by Notch activity. The authors also found that Notch enhanced expression of SMAD3 while decreasing expression of SMAD1, 2, and 6, suggesting that differential regulation of R-SMADs by Notch may be responsible for positive and negative TGF-β/Notch interactions. From this analysis, it appears that the interaction between TGF-β and Notch may be more complex than currently thought.

TGF-β however is not the only member of the TGF-β superfamily that engages in crosstalk with Notch. Early work observed that BMP and Notch signals synergistically reinforced one another during various developmental processes such as Xenopus tail bud formation [117] and tooth morphogenesis [118]. Mechanistically, the BMP/Notch crosstalk involves the formation of a SMAD/NICD transcription factor complex, much like the TGF-β/Notch crosstalk mechanism. Formation of this complex was observed and found to be important for endothelial function and neuroepithelial cell differentiation [118, 119]. Follow up work has now determined that crosstalk between Notch and BMP is important for a wide variety of cellular responses including osteoblastic differentiation [120-122] and vascular biology/angiogenesis [123, 124]. Finally, besides TGF-β and BMP, little is known regarding crosstalk between other TGF-β superfamily members and Notch. However, given that the majority of the other TGF-β superfamily members utilize ALK receptors and SMAD signaling intermediates, it seems likely that future research may uncover new crosstalk mechanisms between members of the TGF-β superfamily and Notch.

Notch and WNT

Like Notch, the WNT signaling network is evolutionarily ancient and heavily utilized during development. Consisting of ~19 ligands that can bind to ~10 frizzled receptors and their co-receptors (LRP5/6), the canonical WNT signaling pathway is mediated by ligand binding to receptor, stabilization and nuclear translocation of β-catenin, and subsequent association with LEF/TCF transcription factors to activate gene specific promoters. In the absence of WNT signaling, β-catenin is phosphorylated by GSK3β which triggers β-catenin ubiquitination and rapid protein turnover (reviewed in [125]). The first evidence suggesting a crosstalk between WNT and Notch signaling was uncovered in Drosophila where it was shown that Notch and WNT cooperate to control wing development [126, 127]. The first molecular evidence supporting crosstalk between Notch and WNT was made by Ross and Kadesch [128], when N1ICD was found to increase transcriptional activity of the LEF transcription factor independently of the canonical Notch transcription factor, CSL. Instead, it was found that the NICD/WNT crosstalk was mediated by a NICD/LEF transcriptional complex that regulated a unique subset of promoters compared to the β-catenin/LEF complex [128]. Similarly, NICD/β-catenin complexes have been identified and found to be important for suppression of neural precursor cells [129] and for inducing an arterial fate in vascular progenitors [130]. Despite these results, this NICD/β-catenin complex does not appear to be required for all instances of Notch/WNT crosstalk. Instead, Hayward et al. demonstrated that membrane-bound Notch is capable of interacting with, and deactivating β-catenin at the cell membrane in a Notch ligand and cleavage independent fashion [131]. Subsequent reports reinforced this finding by showing that β-catenin's association with uncleaved Notch at the membrane is also important for β-catenin regulation in stem cells [132], and imaginal disc development in Drosophila [133]. Thus, Notch signaling can alternatively increase or decrease β-catenin function, depending on the nature of Notch/β-catenin interaction. Finally, while these reports show that β-catenin is a shared point of overlap during crosstalk between WNT and Notch, other WNT signaling intermediates have also been shown to interact with the Notch mechanism. For instance, GSK3β, a serine/threonine kinase that is inactivated by WNT signaling [125], directly phosphorylates NICD resulting in decreased NICD stability and signaling output [134, 135]. In this way, inhibition of GSK3β by WNT signaling results in a positive interaction between the WNT and Notch signaling mechanisms. In contrast to this, WNT activation of the Dishevelled protein triggers an inhibitory interaction between WNT and Notch. It is not completely clear how this is accomplished however since Dishevelled has been shown to interact both with NICD itself, and with the NICD transcriptional factor CSL in the nucleus, where it inhibits NICD/CSL mediated transcription [136, 137].

In summary, the interplay between WNT and Notch is very complex and involves at least four independent mechanisms. This extensive co-regulation may reflect the fact that both Notch and WNT are both heavily utilized during development, where WNT and Notch must cooperate for proper development [138]. Future studies will no doubt further dissect and define the relative contributions of these pathways to crosstalk between Notch and WNT signaling.

Notch and VEGF

The vascular endothelial growth factor (VEGF) signaling pathway coordinates vascular development through VEGF ligand binding to cell surface receptor tyrosine kinases. The core of VEGF signaling consists of six broadly expressed VEGF ligands and four VEGF receptors that are highly restricted to vascular and lymphatic tissues (reviewed in [139]). A flurry of publishing activity in recent years now supports a strong crosstalk between the Notch and VEGF signaling mechanisms in the vascular system. The basis for Notch/VEGF crosstalk appears to be rooted in the reciprocal transcriptional control of Notch ligands by VEGF, and VEGF receptors by Notch. For instance, early work determined that VEGF was an important regulator of Notch receptors and ligands [140, 141]. Around the same time, it was also becoming apparent that Notch activity was an important determinant of VEGF receptor expression [142-144]. It was not until later, however, when a more complete picture of the interaction between Notch and VEGF began to come into focus. The prime example demonstrating reciprocal regulation between Notch and VEGF occurs during angiogenesis, wherein Notch/VEGF crosstalk has been implicated in the selection and differentiation of tip versus stalk cells on growing columns of endothelial cells (reviewed in [145]). During tip cell selection, VEGF binding to VEGF Receptor 2 (VEGFR2) at the quiescent endothelial membrane causes a phenotypic switch into a motile cell state known as a tip cell, while also inducing the expression of Dll-4 [146-148]. Dll-4 expression in tip cells and subsequent binding to Notch receptors on adjacent endothelial cells (stalk cells) reduces stalk cell sensitivity to VEGF through the down regulation of VEGFR2, thereby preventing stalk cells from taking on the tip cell phenotype and restricting the number of new vascular branches [148, 149]. Dll-4 signaling in tip cells also increases Jagged1 expression in stalk cells which in a twist of understanding, inhibits Dll-4-Notch signaling in tip cells resulting in increased VEGFR2 expression and VEGF sensitivity [150, 151]. In this way, VEGF first elevates Dll-4 expression, which then represses VEGF sensitivity in adjacent cells, thus demonstrating reciprocal regulation between Notch and VEGF. Beyond tip/stalk cell differentiation, crosstalk between Notch and VEGF has also been shown to be an important mechanism controlling other aspects of cardiovascular biology such as arteriovenous differentiation [152], differentiation of vascular progenitors from stem cells [153], heart valve development [154], tumor angiogenesis [155], as well as neuronal development [156].

Other microenvironment conditions that control Notch (Shear stress, hypoxia, and hyperglycemia)

As an integrator of cellular microenvironments, the crosstalk between Notch and other signaling pathways is fairly well described compared to crosstalk between Notch and other microenvironmental conditions. Nonetheless, compelling evidence has been emerging in recent years, that stimuli such as shear stress in the cardiovascular system, low oxygen levels (hypoxia), and even hyperglycemia all have significant impacts on Notch signaling (Figure 4). Additionally, Notch has been reported to respond to other microenvironmental conditions including high salt in endothelial precursor cell media [157] and temperature flux in Drosophila [158, 159], however these responses will not be further discussed here. Below we summarize the data and provide mechanistic information (where possible) for interactions between Notch and shear stress, hypoxia, and hyperglycemia.

Figure 4.

Summary of microenvironmental conditions (shear stress, hypoxia, and hyperglycemia) that control Notch. Depicted is a cross-sectional view through a blood vessel showing endothelial cells (EC) and vascular basement membrane. Shear stress (laminar versus disturbed or non-laminar) controls Notch by largely undefined mechanisms that may include regulation of Notch receptors and/or ligands. Hypoxia controls Notch signaling by several mechanisms including the formation of HIF1α-NICD transcriptional complexes, HIF1α mediated stabilization of NICD, enhanced γ-secretase activity, and FIH mediated NICD destabilization. Hyperglycemia controls Notch by largely uncharacterized mechanisms that may include increased NICD stability due to decreased CARM1 expression and/or increased VEGF release from other cells in the vascular microenvironement.

Notch and Shear Stress

Notch signaling has in recent years been shown to be an extremely important regulator in the development and function of vascular systems, and many excellent reviews have been published on the role of Notch in vascular development and function [160, 161]. In addition, Notch has also been tightly linked to several vascular malfunctions including the development of atherosclerotic lesions [162]. Recently however, a previously unrecognized role for Notch in sensing shear stress in the vascular system has also begun to emerge. Shear stress in the vascular system is a mechanical force applied to endothelial cells by fluid flow and normally ranges from 1-5 Pa (10-50 dynes/cm²) in arteries and capillaries, to 0.1-0.5 Pa (1-5 dynes/cm²) in veins [163]. Shear stress is an important component of the endothelial cellular microenvironment that strongly influences endothelial cell biology. Laminar (undisturbed) shear stress provides an athero-protective signal to endothelium, while non-laminar, disturbed, or oscillatory shear stress provokes the development of endothelial dysfunction and atherogenesis [164]. Several endothelial shear stress sensors have been identified and include a wide range of transmembrane proteins on both the apical and basolateral endothelial surfaces and the intracellular kinases and signaling networks that are stimulated by these surface proteins [165]. The first demonstration that Notch can serve as a sensor for shear stress was provided by Wang et al. who showed that Notch signaling targets were differentially regulated after exposure to shear stress for as little as 10 minutes [166]. Masumura et al. subsequently showed that shear stress activates Notch signaling and that this signal is critical for embryonic stem cell differentiation to endothelium [167]. Although protein expression of Notch1 and 4, as well as the Notch ligands Dll-4, Jagged1 and 2 increased after exposure to laminar flow, increased abundance of the active Notch NICD domain was observed prior to increased Notch receptor/ligand expression, suggesting that shear stress may regulate Notch signaling on both transcriptional and post-translational levels [167]. Mechanotransduction by Notch signaling has since been demonstrated to be an important player in both vascular development [168] and dysfunction [169]. Interestingly, inhibition of Notch under atherogenic / low shear stress conditions was shown to inhibit several pro-inflammatory molecules, suggesting that inappropriate activation of Notch by low shear stress may also be linked to the early stages of atherogenesis [170]. While these findings clearly implicate Notch in endothelial shear stress responses, it is not currently understood how Notch signaling is activated by shear stress. It seems unlikely that activation of Notch is wholly dependent on transcriptional mechanisms since Notch activation is observed after as little as 10 minutes of shear stress [166]. However, it has also been shown that inhibition of VEGFR2 signaling during shear stress blocks Notch activation [167], suggesting that if Notch receptors are flow sensors, that they do not act alone during endothelial response to shear stress. Taken together, these findings illustrate the important role Notch plays in responding to shear stress in the endothelial microenvironment. Future work will hopefully further explore the mechanism by which Notch is activated by shear stress and continue to define the importance of Notch in endothelial/vascular response to shear stress.

Notch and Hypoxia

Notch signaling responds to oxygen content within the cellular microenvironment, showing differential activity under normoxic and hypoxic conditions. The first evidence suggesting that Notch might be involved in hypoxic responses came when researchers observed that Dll4 expression was increased in hypoxic tissues [171]. Soon after, Notch signaling was observed to be increased in hypoxic neuroblastoma cells [172]. Further investigation has discovered a physical interaction between NICD and HIF-1α (Hypoxia Inducible Factor α), which was promoted by hypoxia and lead to inhibition of cellular differentiation [173]. A similar observation was also made in Drosophila when it was observed that Sima (the Drosophila HIF1-α homolog) could also activate Notch receptor in a ligand-independent manner [174]. HIF-1α binds to NICD, stabilizes it, and enhances the transcriptional activation of Notch downstream genes through an association with NICD transcriptional complexes [173]. Subsequently, it was shown that hypoxia induced HIF-1α also serves to activate γ-secretase through a direct interaction, promoting invasiveness and metastasis in murine breast cancer cells [175]. Furthermore, HIF-1α can upregulate expression of the Notch ligand Dll-4 in endothelial progenitor cells [176], and lymphatic vessels [177]. Finally, further sophistication of this mechanism is achieved through FIH (Factor Inhibiting HIF-1α), an asparagine hydroxylase [178], which works to inhibit HIF-1α activity in an O₂ dependent mechanism [179]. FIH also negatively regulates Notch target gene transcription [173], likely through its hydroxylation and destabilization of NICD under normoxic conditions [180]. Collectively, the multiple mechanisms by which hypoxia controls Notch including HIF-1α association with Notch transcriptional complexes, γ-secretase induction, promotion of ligand expression, and FIH activity provide at least four independent mechanisms by which Notch cooperates in hypoxic responses.

Notch and hyperglycemia

Recent work has begun to dissect a molecular mechanism by which Notch signaling may respond to increased or decreased blood sugar and possibly play a role in diabetes and the vascular and renal complications associated with diabetes. For example, Notch signaling in hepatocytes is increased in response to high sugar concentrations [181] and hyperglycemia induced Jagged1 expression in endothelium was proposed to be an important mediator of diabetic vasculopathy [182]. Moreover, several investigators have shown that hyperglycemia elevates Notch receptor expression/signaling in cultured podocytes [183-185], and elevated Notch signaling has been linked to a loss of glomerular filtration due to a negative impact on podocyte function [186, 187]. Thus, hyperglycemic stimulation of Notch may be extremely important for understanding the pathology of diabetic nephropathy, especially since podocyte damage in diabetic kidneys has been proposed to be an early triggering event leading to other downstream renal complications [188]. In support of this, treatment of Streptozotocin induced diabetic mice with the angiotensin inhibitor Valsartan, simultaneously decreased renal damage and Notch activation [189]. Mechanistically, the link between hyperglycemia and Notch has been elusive, however it is known that hyperglycemia (as well as hypoglycemia) induce VEGF secretion and signaling [190, 191]. Given the reciprocal transcriptional regulation between VEGF and Notch (as described above), it seems likely that VEGF and Notch may pathologically synergize in hyperglycemic conditions. In support of this, a recent report by Chiu et al. found that hyperglycemia increased endothelial secretion of heparinase leading to increased VEGF release from neighboring myocytes thus enhancing endothelial Notch activity [192]. Interestingly, anti-VEGF therapies have shown some success in reducing diabetic renal dysfunction [193, 194], although it is unknown whether these anti-VEGF approaches also reciprocally decrease Notch signaling. It has also been shown that inhibition of Notch reduced the elevated VEGF secretions in podocytes cultured under hyperglycemic conditions, as well as diabetic nephropathy in diabetic rats [195]. Finally, there may be other Notch regulatory mechanisms beyond VEGF that are operant during hyperglycemic conditions. For example, hyperglycemia induced Notch activity seems to be also be linked to decreased CARM1 methyltransferase activity [196], a recently discovered negative regulator of Notch signaling [197]. Hopefully future research will be able to more clearly define the molecular mechanism by which hyperglycemia controls Notch in order to more fully understand the downstream implications of Notch signaling in diabetes.

Conclusions

Through the examples presented in this review, it is clear that Notch signaling is capable of responding to a range of changing microenvironmental conditions that go far beyond the basic model of receptor-ligand interaction for Notch activation. Instead, evidence is building that the basic model of Notch activation is manipulated by a variety of microenvironmental cues and that the basic cell-cell interaction model of Notch activation represents only a part of the broader function of Notch to sense and respond to a wide variety of microenvironmental conditions. Despite these findings, many of the results discussed here have been gained from simplified models of cellular microenvironment. In order to build a more complete understanding of how Notch serves its role as an integrator of cellular microenvironments, future studies will need to examine how Notch responds to these conditions in more complex in vivo models, an undoubtedly complex task. In addition, fleshing out the molecular underpinnings of how Notch responds to microenvironmental conditions is an important goal that should provide opportunities for pharmacological intervention in the many diseases and processes that are characterized by altered microenvironments. Lastly, efforts to therapeutically manipulate Notch are ongoing for the treatment of a wide variety of diseases (reviewed in [198]). Therefore, an understanding of the full spectrum of activities that Notch serves in the microenvironment is an important consideration in developmental biology and various Notch targeting strategies.

Highlights.

• In addition to adjacent cells, Notch responds to a wide variety of microenvironmental signals.

• Notch activation by microenvironmental signals might reflect the evolutionary origins of Notch.