The formation of the mammalian nervous system takes place via a number of developmental steps. All phases of brain development involve the recurrent themes of induction, cell proliferation, cell fate determination (differentiation), cell movement (migration), cell process formation, and targeting (synapse formation) (1). Signaling pathways involved in these processes are regulated not only in space, but in time and intensity as well. A vast array of signaling events are coordinated such that cells proliferate and differentiate at the correct time, space, and orientation to generate an amazingly organized structure capable of adaptability and plasticity. The Notch signaling pathway, originally discovered in Drosophila, impinges on a wide array of cellular processes including maintenance of stem cell self-renewal, proliferation, specification of cell fate or differentiation, and apoptosis. What follows is a review of the role of the Notch signaling pathway in neurodevelopmental processes and its role in the pathogenesis of certain human neurologic diseases.


Notch signaling.

The Notch gene encodes a receptor with a single transmembrane domain. Although initially synthesized as a single protein, it is cleaved in two and exists as a heterodimeric receptor embedded in the plasma membrane. Signaling is initiated when a Notch receptor on one cell interacts with Notch ligands, such as Delta or Serrate (in Drosophila), on an adjacent cell (Fig. 1). This interaction triggers two proteolytic events culminating in the release of the Notch ICD. The free intracellular fragment then translocates to the nucleus where it binds to the transcriptional regulator CSL [for CBF-1, Su(H), and LAG-1], resulting in displacement of co-repressors previously bound to CSL and recruitment of co-activators. The co-activators then induce expression of the Hairy-Enhancer of Split (HES) and Hes-related proteins (HERP) gene families, although recent data suggest that CSL-independent pathways may also exist (24). It is also important to note that in mammals there exist multiple subtypes of each “actor” in this pathway including Notch1–4, Delta-like ligands (Dll)-1,-3, and -4, and Serrate-like ligands (Jagged-1 and -2) (3). This becomes relevant later on in this review in that various human diseases are usually associated with a particular Notch gene. Notch is involved in mediating two distinct types of cell-cell signaling interactions: lateral and inductive signaling (5,6). Studies in Drosophila revealed the importance of Notch in the control of cellular differentiation by lateral inhibition, which ensures that two distinct cell types are produced in correct numbers from a population of initially equipotent cells (7). For example, in Drosophila, Delta signaling to Notch induces the Notch responding cell to remain a progenitor cell, whereas the Delta-expressing cell differentiates into a neural cell (8). Although the equipotent cells initially express equivalent levels of both ligands and receptors, via a negative feedback mechanism, the Notch responding cell down-regulates its own Delta ligand expression effectively blocking Notch receptor pathway activity in the Delta signaling cell (Fig. 2A). Inductive signaling, on the other hand, involves Notch receptor and ligand expressed on two different cell types such that Notch is only activated in the receptor-bearing cell, resulting in a cell-fate decision (Fig. 2B). This role for Notch signaling has been demonstrated in multiple settings including T-cell lineage specification (9,10), mammalian keratinocyte differentiation (11), and mammalian gliogenesis (1214). The importance of these mechanisms is illustrated by their conservation across multiple species, from Drosophila and Caenorhabditis elegans to amphibians and mice (8,1518).

Figure 1
figure 1

The heterodimeric Notch receptor, upon contact with its ligand (e.g. Delta) undergoes proteolytic cleavage first by an ADAM family protease and then by γ-secretase/presenilin1. This liberates the intracellular domain of Notch (ICD) allowing it to translocate into the nucleus where it displaces co-repressors (CoR) from the CSL transcription factor. Subsequent binding to CSL then occurs and recruitment of co-activators (CoA) results in the transcriptional activation of downstream target genes.

Figure 2
figure 2

(A) Model of lateral inhibition: Two initially equipotent cells express equal levels of the Notch receptor and Delta ligand. If levels of a “proneural” fluctuate so as to increase in the cell on the right, Delta ligand expression increases and Notch decreases in that cell, and neural differentiation commences. Secondary to increased Delta ligand exposure, an opposite pattern in the left cell results such that Delta ligand expression is down-regulated and Notch signaling predominates resulting in maintenance of an uncommitted progenitor state. (B) Example of inductive signaling: A neural stem cell expressing the Notch receptor may interact with a variety of cells depending on its environmental context. In this figure, for example, the Notch expressing cell may come into contact with an already committed neurogenic precursor expressing the Delta ligand. This contact initiates a cascade of signaling events that, along with other cell extrinsic signals, results in glial fate commitment even in the setting of a strong pro-neurogenic signal such as BMP2 (14). In the absence of this cell contact, the neural progenitor is instructed toward neuronal development.

Stem cell maintenance.

In Drosophila, Notch prevents early neurocompetence in ectodermal cells via interplay with the Wnt signaling pathway (19,20). Although this particular role has not yet been shown to be active in vertebrates, other Notch1-mediated signaling pathways are crucial for mammalian CNS development via maintenance of a neural stem cell (progenitor) state, inhibition of neuronal commitment, and promotion of glial fates (12,21). Notch1 mouse mutant embryos die before E11.5, near to the time when the first neurons express their mature, differentiated phenotype (17). However, close examination of mutant embryos revealed hypoplastic brains and neural tubes secondary to a loss of neuroblasts and premature neuronal differentiation (22). Further support for the role of the Notch pathway in the inhibition of neuronal differentiation came from studies on PS1 (presenilin), which is a component of the γ-secretase complex that allows for the efficient proteolytic release of the Notch ICD. PS1 knock-out mice (PS1−/−) also experience a reduction in the neural progenitor population (23). Consistent with these findings, overexpression of Notch also promotes maintenance of neural precursors (13,24). The downstream signaling pathways responsible for these effects are being elucidated. The basic helix-loop-helix genes Hes1 and Hes5 are the major effectors for Notch signaling (12). In Hes1/Hes5 double mutant mice, virtually all the neural stem cells differentiate prematurely into neurons resulting in severely disorganized neural tube morphology and absence of development of normal brain structures (25). However, initial stem cell generation seems independent of Notch signaling (26). Recently, two forms of the mammalian neural stem cell have been proposed: the “primitive” neural stem cell, isolated from E5.5 to E7.5 mouse embryos that possess self-renewal capacity and neural multipotentiality, but also contain some non-neural properties (27); and the “definitive” neural stem cell, which, generally isolated initially at E8.5, are FGF2 responsive (28). Notch1 appears to be required for the transition from the primitive neural stem cell to the definitive neural stem cell and subsequent maintenance of the definitive neural stem cell state (27).


Although the Notch pathway appears to be critical for neural stem cell regulation and maintenance, it also appears to have a role in later neuroglial development (2931). Cells of glial lineage comprise more than 90% of the human brain. While astrocytes and oligodendrocytes make up the bulk of these, other cell types also comprise this “non-neural” component of the brain. Microglia clear debris and mediate inflammation and may be of hematopoietic origin and the radial glia of the embryonic brain have been shown to be neural progenitors (32). In 1994, the observation was made using the P19 mouse embryonic carcinoma cell line, that Notch could inhibit retinoic acid induced P19 cell differentiation into neurons and myocytes. However, differentiation into glial (astrocyte) lineages was maintained (29). At that time, it was assumed that Notch simply “permitted” differentiation under the control of other undefined pathways. However, studies since then have revealed a more complex role for Notch and gliogenesis. Gain-of-function studies have revealed that Notch not only inhibits the differentiation of some cell types but can also promote a glial fate, such as the Müller glia cells in the retina, the radial glial cells in the neocortex, and astrocytes from the hippocampus (1214,21,31,33). Interestingly, all of these cell types in certain contexts have the potential to act as neural stem cells, thus reiterating a potential role for Notch in neural progenitor maintenance even in these “glial” specified cell types (3335). A recent study examining Dll1 (mammalian delta ligand homologue) knock-out neurospheres used time-dependent modulation of Notch pathway activation to construct a model of neural stem cell differentiation (36). The authors suggest the existence of two progenitors, arising from the primitive neural stem cell, P1 which is committed to a neural fate, and P2 which is committed to a glial fate. Notch is initially responsible for inhibiting P2 from developing into neurons. After other differentiation signals commence, Notch guides the now committed glial precursor into GFAP+ (glial fibrillary acidic protein) astrocyte rather than oligodendrocyte development. This is supported by another recent in vivo study showing that Notch signaling in the mouse embryonic brain is activated in certain GE (ganglionic eminence) cells, a known site of neural stem cells, and yet exclusive from those neuronal precursors expressing Mash1 a pro-neural bHLH (basic helix-loop-helix) gene (37). Notch expression can be observed during the maintenance and proliferation of the radial glia (glial precursors) and reappears again transiently during immature glial precursor differentiation into astrocytes.

Learning and memory.

Now that the molecular and physiologic roles of notch in the nervous system have been discussed, the implications of these roles can be discussed in the setting of learning, memory, psychiatric disorders, and specific neurologic pathology. In addition to its embryonic role in determining cell fates, Notch has also been found to play a role in postnatal developmental processes. Interest in Notch with regards to plasticity and other higher brain functioning began with its link to the presenilins that were initially discovered during early research on the genetic basis of Alzheimer's disease, which will not be further discussed, given excellent reviews elsewhere (3840). Notch has been shown to regulate cortical neurite (axon and dendrite) growth as well as dendrite branching in murine postmitotic neurons (4143). Presenilin (also known as sel-12) mutants generated in C. elegans displayed defects in their ability to “learn” to find their optimal temperature in a gradient (44). It was postulated that this neurodeficit was mediated by defects in Notch signaling. Presente et al. (45) generated temperature sensitive Notch mutants in Drosophila, and after inactivation the flies suffered from a progressive neurodegenerative syndrome characterized by loss of flying ability and early death. More recently, the same group using the temperature-sensitive mutants as well as RNA-mediated inactivation of Notch demonstrated specific neurocognitive deficits in Drosophila related to long-term memory formation (46). Similar findings have been documented in mammalian systems as well. Mice with heterozygous null mutations in the Notch1 gene (Notch1+/−) were tested using a water maze method, which revealed long-term spatial memory deficits. Identical in nature, but more severe deficits were seen in mice lacking one copy of the CSL gene (47), where CSL is the downstream effector of all four mammalian Notch genes. Wang et al. (48) used knock-down Notch mutant mice (via an anti-sense transgene) to specifically examine the effects on hippocampal neuron plasticity. Detailed neurophysiologic testing of hippocampal neurons revealed that long-term potentiation (a phenomenon associated with long-term memory formation) was inhibited in the mutants but rescued by exogenous addition of the Notch ligand, Jag-1 (48).

The study of learning and memory in humans is difficult due to the complexity of the human brain, and elucidation of the molecular processes responsible has just begun. By studying known developmental syndromes and linking their phenotype with genotype we can begin to dissect the complex pathways that lead from the embryonic neural tube to the developed brain and then to the unique attributes that define humans, the ability to learn and retrieve memories. A recent study on Down's syndrome (DS) identified a possible linkage with a presenilin1 polymorphism (49). Individuals with DS are likely to develop neuropathologic changes characteristic of Alzheimer's disease after only 40 y of age, including build-up of amyloid protein secondary to abnormal processing of the β-amyloid precursor protein (APP). Again, PS1 is a component of the γ-secretase complex that plays a role in both APP and Notch processing (39). Additionally, from birth, individuals with DS display defects in domains of learning and memory often attributed to hippocampal and prefrontal systems (50). Examining the possible role for defective Notch pathway signaling in these patients and in the DS mouse model (51) will provide for not only a deeper understanding of DS and Alzheimer's disease neuropathology but potential understanding of other developmental disorders as well.

Neurologic disease.

As described above, Notch plays key roles in both embryonic neural development and later brain plasticity in animal models. Human diseases are often multifactorial and impossible to attribute to a single gene defect. The Notch pathway is ubiquitously present throughout mammalian development and is integral to the proper formation and maintenance of multiple tissues and organs including but not limited to the hematopoietic system, the vasculoendothelial system, as well as the nervous system (3). It is becoming apparent that dysregulation of the Notch pathway can be found in a multitude of human diseases including T-cell leukemia, spondylocostal dysostosis, schizophrenia, CADASIL syndrome (to be described shortly), and Alagille syndrome among others (5256).


Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is characterized by the adult-onset of recurrent strokes, progressive vascular dementia, migraines, psychiatric disturbances, and pseudobulbar palsy. Mutations in Notch3 have been identified as the causative mutation in this syndrome (57). Although the disease is characterized by predominately neurologic pathology, the mechanism behind the disease appears to be vascular in nature. The cerebral vessels are narrowed by expansion of the extracellular matrix and by degeneration of the vascular smooth muscle cells and subsequent deposition of granular osmiophilic material (GOM). MRI reveals a microangiopathic pattern of signal abnormalities and white matter changes suggestive of infarcts and leukoencephalopathy. Interestingly, there is also an accumulation of the ectodomain of the Notch3 receptor (58,59). In fact, this accumulation in the small vessels of the skin has been used for diagnostic purposes with a MAb to the ectodomain of Notch3 (60). However, the vascular pathology begins before the deposition of either the ectodomain or GOM. There have been suggestions of an autoimmune component to the vasculopathy, however, there are no studies documenting this as yet. The exact molecular basis behind this pathology has yet to be elucidated. Primary loss or gain of function of Notch3 signaling has not consistently been demonstrated in multiple studies using CADASIL-associated Notch3 mutants, suggesting these mechanisms are unlikely candidates for the resultant pathology (6164). Another possibility is a direct toxic effect from the accumulation of the Notch3 ectodomain, as mentioned above, however the mechanism behind this aggregation has yet to be determined.


Schizophrenia is a complex mental illness with multiple phenotypic presentations. The underlying heritability of this illness likely arises from a multitude of genetic and/or epigenetic factors, and the final phenotype depends on environmental modulation as well. Early linkage studies identified a susceptibility locus on chromosome 6p (65), and subsequently a more detailed analysis looked specifically at the Notch4 gene. Linkage disequilibrium (LD) mapping of 80 British parent-offspring trios revealed LD with an A to G substitution in the promoter region and mutations in a CTG repeat site in exon 1 of the Notch4 gene as candidate susceptibility sites (56). Multiple similar analysis have been performed on various populations since then, but not all find similar linkage patterns (6669). However, a study of 210 affected individuals comparing neuropsychological testing and brain volumes showed no significant link between the presence or absence of schizophrenia, but did show an association between smaller frontal lobe volume and worse performance on the Wisconsin Card Sort Test (WCST), a measure of frontal lobe function and integrity in those patients with schizophrenia (70). The existence of subtle neurodevelopmental anomalies and anatomic findings in these patients would be consistent with either an early developmental event in which migration or neurogenesis was affected, or a later loss of brain matter, possibly due to lack of neural stem cells. Obviously, these are merely speculations, and further epidemiologic analysis and animal model development is needed to define the role of Notch signaling in schizophrenia.


Focal cortical dysplasia (FCD) was originally described as focal developmental anomalies of cortical structure characterized histologically by cortical dyslamination, the presence of abnormal giant neurons throughout the resected cortex and adjacent white matter, and accompanied in many cases by balloon-shaped cells of uncertain lineage (71). However, FCD includes a spectrum of disordered white and gray matter entities that range from mild cortical disruption to complete derangement of cortical lamination (72). Clinically, patients present usually in childhood with refractory partial epilepsy, and subsequent magnetic resonance imaging (MRI) reveals the area of disordered cortex. Treatment is centered on relieving the seizures and often involves surgical removal of the affected cortex. The defect is thought to arise from a migrational or apoptotic defect occurring early in development. In humans, the earliest migrations occur around wk 6 of gestation. At this time, radial glial cell fibers begin directing the migration of neural precursors from the ventricular to pial surface (73). PS1-deficient mice develop global cortical dysplasia characterized by overmigration of cortical plate neurons. This is associated with alterations in the distribution of Notch1 in the Cajal-Retzius neurons, cortical plate neurons responsible for regulating radial neuronal migration (74). Furthermore, specific analysis of tissue specimens from human FCD revealed greatly altered levels of both Notch-1 and Dvl-1 (disheveled), an integral effector of the Wnt pathway, in the abnormal neurons and balloon cells (75). In light of the known role that Notch plays in embryonic neural fate specification, it is not surprising that the Notch pathway plays a role in disorders of cortical formation. Further studies using animal models of cortical dysplasia are needed to define the early embryonic role of Notch in the genesis of these disorders.


In addition to playing a central role in neurodevelopmental processes as described, Notch dysregulation has been implicated in a number of human malignancies, including T-cell leukemia, breast and colon adenocarcinomas, and cervical cancer among others. The potential role for the Notch pathway in brain tumorigenesis has been given recent attention. The existence of neural stem cell-like tumor stem cells was postulated many years ago (76). Several studies have shown the existence of neural stem-like cells from human glial tumors (77,78). A recent analysis of pediatric brain tumors revealed the existence of progenitor cells with multipotentiality and certain expression characteristics similar to those of neural stem cells (79). Ignatova et al. (78) showed abnormal Notch ligand (Delta-1 and Jagged-1 and -2) expression from human tumors of glial origin (glioblastoma multiforme and anaplastic astrocytoma). In the presence of growth factors, tumor cell lines down-regulated Delta-1 expression whereas the control cell lines continued to express it suggesting a disordered signaling pathway tending toward persistent “stemness” as opposed to differentiation. Other studies have revealed elevated expression of Musashi1, an RNA-binding protein that acts to negatively regulate m-Numb translation, in medulloblastomas and high-grade astrocytomas (77,80,81). Numb is a well-known Notch1 antagonist, thus implicating up-regulation of the Notch pathway as part of the pathogenesis of these tumors. The prognosis of malignant brain tumors remains poor even with advances in surgical approaches, radiation therapy, and chemotherapy. Therapies targeted at the molecular mechanisms behind these tumors are desperately needed to improve the outcomes of these devastating tumors, underscoring the need for further examination of these crucial signaling pathways.


Although there has been significant progress in the past decade with regard to understanding the molecular mechanisms behind neurodevelopment, much still remains to be discovered. Hundreds of interacting signaling pathways likely lie behind even the most simple of neurodevelopmental events and only further analysis of animal models, basic biochemistry, and human diseases will bring us closer to elucidating these mechanisms. The Notch signaling pathway likely plays a key role in these events and further study is needed to define the developmental context in which the different Notch receptors and ligands act to direct development and affect disease.