It Takes a Village: Constructing the Neurogenic Niche

Abstract

While many features of neurogenesis during development and in the adult are intrinsic to the neurogenic cells themselves, the role of the microenvironment is irrefutable. The neurogenic niche is a melting pot of cells and factors that influence CNS development. How do the diverse elements assemble, and when? How does the niche change structurally and functionally during embryogenesis and into adulthood? In this review, we focus on the impact of non-neural cells that participate in the neurogenic niche, highlighting how cells of different embryonic origins influence this critical germinal space.

Introduction - Formation and Maturation of the Neurogenic Niche

Neurogenesis encompasses the entire set of events that lead to the formation of new neurons from their stem and more committed progenitor cells. This includes cell division, production of migratory precursors and progeny, differentiation and integration into circuits. Neurogenesis occurs en masse during development to build the nervous system and persists in select regions of the adult brain. The term ‘neurogenic niche’ typically refers to the complex microenvironment that supports the neural progenitor cells (NPCs, which include neural stem cells (NSCs) and their progeny). The niche informs their decision to either remain dormant or divide, and provides signals that guide early stages of differentiation. Furthermore, the germinal niche must be dynamic in order to accommodate vast changes in neurogenesis that occur during embryogenesis and into adulthood, and vary in different locales to produce area-appropriate progeny. While numerous excellent reviews have covered the structure of central nervous system (CNS) germinal zones from a neural-centric viewpoint, the impact of diverse non-neural derived elements is increasingly being appreciated. This review will describe the forebrain embryonic ventricular/subventricular zone (VZ/SVZ), focusing on the mouse cerebral cortex, and how each of its elements influences NPC proliferation, migration and differentiation. Finally, we will briefly discuss the continuing function of these elements in the adult, mainly in the mouse subventricular zone (SVZ). Although we limit ourselves to the study of these specific neurogenic niches, much of this knowledge is applicable to other CNS neurogenic areas and other stem cell (SC) niches.

Embryonic Ventricular and Subventricular Zone (VZ/SVZ)

At the onset of developmental neurogenesis (around embryonic day (E) 9.5), the primary precursors of the CNS are neuroepithelial cells (NECs) that form a tube with a central canal. Like other epithelial cells, NECs make lateral connections to each other though adherens and tight junctions, and display apico-basal polarity. They form a pseudostratified epithelium, as their nuclei migrate with cell cycle stage, being situated near the apical side during mitosis and more basally during S phase. At the anterior end of the central canal, NECs are particularly proliferative, dividing symmetrically to vastly expand the anlage of the future brain. This foundational process forms several layers surrounding the lumen of the developing nervous system, and the innermost apical layer where the principal progenitor cells reside is called the VZ.

After their initial rounds of division, NECs transform into radial glial cells (RGCs). The RGC somata and apical processes lie in the VZ, while their elongated basal processes extend to the outer (pial) surface of the brain (Fig. 1). RGCs appear to divide both symmetrically and asymmetrically, and their basal processes are used by newborn neurons as guiding scaffolds during their migration away from the germinal niche towards the pial surface to differentiate and form connections.

Fig. 1.

Illustration depicting the appearance and maturation of neurogenic niche constituents derived from (a) the neuroepithelium and (b) of non-neural origin in the mouse cerebral cortex. (c) Illustration depicting the wave of angiogenesis in a coronal view of the mouse telencephalon (adapted from Shen et al., 2004). Also note the position of the ChPs during early embryogenesis (purple). VZ: ventricular zone; SVZ: subventricular zone; CP: cortical plate; MZ: marginal zone; V-R: Virchow-Robin space.

Evidence from in vitro slice imaging and clonal analyses conducted in vitro and in vivo suggest that asymmetric divisions of RGCs can produce another RGC and a neuron or an intermediate progenitor cell (IPC) that in turn gives rise to a small number of neurons. In the forebrain, IPCs accumulate above the VZ, forming a second germinal zone, the SVZ. Unlike VZ cells, the nuclei of IPCs do not display intracellular nuclear migration. In some species, including humans, this secondary germinal zone in the cerebral cortex is greatly expanded and is separated into outer and inner SVZ layers (Florio and Huttner, 2014; Lui et al., 2011). The inner cortical SVZ is similar to the mouse SVZ, but the outer SVZ is different in including numerous basal RGCs which retain their basal process but lose their apical contact, and more prolific IPCs that have also been termed transit amplifying progenitor cells. This huge expansion of the SVZ niche is thought to underlie much of the evolutionary expansion of the cerebral cortex.

Postnatally, SVZ progenitor cells produce largely glial cells, and the remaining RGCs differentiate into either ependymal cells, which line the ventricles, or into glial cells, including astrocytes and oligodendrocytes. The NSCs in most CNS regions largely extinguish after development, but in many species including mice they remain throughout life and are actively neurogenic in the striatal SVZ and the hippocampal dentate gyrus.

These neural cells interact via cell-cell contacts and factor production to regulate neurogenesis in the niche, as reviewed previously (Taverna et al., 2014), and they are impacted by several non-neural niche components, the development of which we will now describe.

Vascularized Choroid Plexus & Cerebrospinal Fluid (CSF)

As the neural tube closes, it envelops amniotic fluid that fills the lumen and is later actively modified. Early in development, NECs and RGCs contribute to the composition of this liquid milieu, but this task soon falls to the choroid plexuses (ChPs). The ChPs are folded structures residing in the brain ventricles that consist of a single layer of highly active epithelium sandwiching an elaborate vascular network; this vascular-neural composite controls passage of molecules into the CSF, which impacts neurogenic zones throughout life (Zappaterra and Lehtinen, 2012; Redzic et al., 2005; Liddelow, 2011).

There are four ChPs in the mammalian brain. During development, the 4^th ventricle ChPs appears first, followed by the two lateral ventricle ChPs and finally the 3^rd ventricle ChP (Dziegielewska et al., 2001; Saunders et al., 2012). Bmp7 and Otx2 expression reveal the NECs that will give rise to the ChP already by E4 in chick embryos, (reviewed in Zappaterra and Lehtinen, 2012). Otx2 is a critical regulator of ChP development and function, as deletion at early stages of embryogenesis impairs development of all ChPs (Johansson et al., 2013). Lateral ventricle ChPs are derived from NECs in the roof plate soon after neural tube formation. The correct balance of Sonic hedgehog (Shh) signaling in the dorsal telencephalic midline is critical, as loss or over-activation reduces ChP development (Himmelstein et al., 2010). By E12, meninges and capillaries already invest the interior of the ChP, and the entire structure proliferates rapidly. The developing ChPs provide an entry site for macrophages to access the ventricles, where many reside throughout life as supraependymal and epiplexus (Kolmer) cells (Sturrock, 1979). Structural studies in sheep, rats and humans suggest that the embryonic ChP epithelium develops tight junctions early in their differentiation, further supported by permeability studies, although it is more permeable to small molecular weight molecules than adult ChP (Dziegielewska et al., 2001). CSF turnover is slow in the embryonic brain, and the developing CSF is relatively protein-rich (Saunders et al., 2012). CSF production helps shape the developing CNS through hydrostatic forces and can stimulate progenitor cell proliferation and promote neuronal survival during development (Zappaterra and Lehtinen, 2012). The vascular endothelial cells within the ChP are unique in the brain in having fenestrations, so the tight junction barrier between ChP cells is a key aspect of the blood brain barrier.

Growth of vascular cells into the ChP depends on midline signaling. For example, Shh produced by the hindbrain ChP epithelium induces the extensive vascular growth required for ChP functions. Interestingly, Shh acts via ChP pericytes that in turn act on the vascular endothelial cells to induce this growth (Nielsen and Dymecki, 2010). Continued, high production of VEGF by the ChP epithelium acts on the ChP vascular endothelium to maintain their characteristic fenestrations, which are important for high vascular permeability in this locale (Esser et al., 1998). In addition to regulating passage of specific molecules from the vasculature into the CSF, including key hormones such as T4, the ChPs themselves generate and secrete important growth factors (Fig. 3). ChPs are also a rich source of BMP7 and IGF2 throughout life. BMP7 is made by the ChPs from early stages and is secreted at detectable levels in the CSF. Reduction of BMP7 in the embryo causes impaired and disorganized RGCs that lack normal attachment to the pial surface, reduced cortical progenitor cells and lowered production of Tbr2⁺ IPCs, resulting in a microcephaly phenotype (Segklia et al., 2012). IGF2 released from the ChP epithelium into the CSF, promotes NEC proliferation in an age-dependent manner, and acts in concert with other ChP-derived factors yet to be identified (Lehtinen et al., 2011). The ChPs secrete several FGF, Wnt and Shh pathway components (Table 1), all of which have been shown to increase NPC proliferation. As well as having a role in progenitor cell proliferation, the ChP impacts cell exit from the germinal zones. ChP cells secrete Slit2 that acts as a chemorepulsive factor for neuronal migration (Hu, 1999). Given the central position of the ChP in contributing factors to the neurogenic niche via the CSF, further investigations of the ChP secretome are undoubtedly worthwhile.

Fig. 3. Cellular components of the neurogenic niche.

Cellular components of the embryonic VZ/SVZ (top) and the adult SVZ (bottom), along with selected factors with which they affect neurogenesis.(For references, see the body of the review.)

Table 1.

Signaling molecules associated with the V-SVZ neurogenic niche of the healthy CNS.

Source	Known Timing (rodent)	Signals	Effect	Specific Factors	Reference
Choroid plexus/CSF	Embryonic	CSF	NPC proliferation and survival		Zappaterra and Lehtinen, 2012
			NPC proliferation	BMP7	Segklia et al., 2012
			NEC proliferation	IGF2	Lehtinen et al., 2011
			neuron migration	Slit2	Hu, 1999
			mitogen	Wnt	Johansson et al., 2013
			mitogen	Shh	Huang et al., 2010
	Adult		neuron migration	CSF flow propelled by ependymal cilia	Sawamoto et al., 2006
			neuron migration	Slit1/2	Nguyen-Ba-Charvet et al., 2001
			mitogen	FGF2	Hayamizu et al., 2001
			mitogen	amphiregulin	Falk and Frisén, 2002
			mitogen	TGFα	Seroogy et al., 1993; Tropepe et al., 1997
			NSC self-renewal, mitogen	IGF2	Ziegler et al., 2012
			negatively affects neurogenesis	TGFβ	Knuckey et al., 1996; Wachs et al., 2006
Vascular endothelial cells, SVZ neurospheres	Embryonic	secreted factors	NSC self-renewal, neurogenesis		Palmer, 2000; Shen et al., 2004; Daneman et al., 2010a
			neurogenesis	VEGF	Li et al., 2013
			neural differentiation	Jagged1	High et al., 2008
	Adult		neural migration, self-renewal, neurogenesis		Nam et al., 2007; Engelhardt and Liebner, 2014; Palmer et al, 2000; Shen et al., 2004
			neural migration	SDF1	Kokovay et al., 2010
			self-renewal	PEDF	Ramírez-Castillejo et al., 2006
			proliferation	betacellulin	Codega et al., 2014; Gómez-Gaviro et al., 2012
			NSC maintenance	NT3	Delgado et al., 2014
Meninges	Embryonic	secreted factors	neurogenesis		Decimo et al., 2012
			RGC attachment, neural migration	perlecan, laminin	Halfter et al., 2002
			differentiation	RA	Siegenthaler et al., 2009
			neural migration	Cxcl12	Borrell and Marín, 2006; Paredes et al., 2006
Microglia	Embryonic	secreted factors	differentiation, proliferation		Aarum et al., 2003
			NPC phagocytosis		Cunningham et al., 2013
			astrogliogenesis, NSC maintenance	LIF, bFGF	Zhu et al., 2008; Antony et al., 2011
	Adult	neurotrophic factors	Increased neurogenesis	NGF, BDNF, NT3, GDNF, bFGF, HGF, plasminogen, and cytokines	Harry, 2013; Kim & de Vellis, 2005
			decreased NPC proliferation and differentiation	TNFα, β, IL-1α, β and IL-6	Ye & Johnson, 2001 reviewed in Harry, 2013
			NPC phagocytosis		Sierra et al., 2013
Neurons	Adult		reduced or increased proliferation	dopamine	Höglinger et al., 2004; Kippin et al., 2005
			dose-dependent effect on proliferation	NO	Romero-Grimaldi, 2008
			NSC proliferation	serotonin	Tong et al., 2014
			proliferation	ChAT+	Paez-Gonzalez et al., 2014
Ependymal Cells	Adult		promotes or inhibits neurogenesis	Noggin	Lim et al., 2000; Colak et al., 2008
			NSC self-renewal	PEDF	Ramírez-Castillejo et al., 2006
Fractones	Adult		Inhibit SVZ neurogenesis	Bmp-4, -7	Mercier et al., 2014; Douet et al., 2012
			SVZ NPC proliferation	FGF2	Douet et al., 2013

Neurovasculature, Endothelial Cells, Pericytes

Neural germinal zones are invested with a rich vascular supply and associated pericytes that must develop and grow as the VZ/SVZ emerges and expands. In the mouse, angioblasts approach the outside of the neural tube. They ensheath it to form a meshwork sleeve and undergo vasculogenesis to generate the perineural vascular plexus (PNVP), which occurs at approximately E8.5–9.5 in the caudal regions, then sweeps rostrally (reviewed by Engelhardt and Liebner, 2014).

Angiogenesis begins around E9.5–10, progressing dorsally and rostrally (Marin-Padilla, 1985). Intriguingly, the timing of angiogenesis approximates the gradient of neurogenesis in the forebrain. Vascular sprouts from the PNVP led by endothelial tip cells first contact and breach the pial basement membrane (BM) and the glia limitans, then infiltrate the nervous tissue centripetally; stalk cells, which lie just behind the tip cells, proliferate to elongate the growing vessels. These pioneering vessels plunge directly towards the VZ, and extend branches tangentially above it that anastomose with neighboring branches to produce the subventricular plexus (SVP, Engelhardt and Liebner, 2014). This entire process is very rapid, for example, in the mouse hindbrain it takes little more than a single day (Ziegler et al., 2014). Growth of additional capillary meshes bolsters the early vascular framework as the CNS expands. The endothelial cells develop tight junctions as an essential aspect of the blood brain barrier (BBB), and even invading vascular sprouts in the embryo appear not to be leaky (Saunders et al., 2012).

The vasculature close to the neurogenic niche contributes a variety of factors that impact neural progenitor behavior, and vice versa. As the brain grows and simple diffusion becomes insufficient to deliver oxygen, HIF-1a is induced in NPCs by low oxygen levels to increase vascular development, in part through expression of VEGF and erythropoietin (Lee et al., 2009). RGCs help stabilize the nascent blood vessels by modulating canonical Wnt signaling, and disruption of RGCs causes vessel regression (Ma et al., 2013). RGCs also produce retinoic acid (RA) that acts on the brain vessels and helps develop tight junctions, ensuring BBB function (Mizee et al., 2013). TGFβ2 deletion from forebrain NPCs causes diminished VEGFA, FGF2 and IGF production, which contribute to vessel malformation and hemorrhaging (Hellbach et al., 2014).

Vascular endothelial cells promote embryonic NSC self-renewal and neurogenesis (Shen et al., 2004). Transcriptome analysis of isolated cells show that brain endothelial cells secrete numerous factors (Daneman et al., 2010a), several of which have been implicated in regulating the germinal niche (Table 1). The SVZ plexus endothelial cells have a unique transcriptome that differs from those at the pial surface (Won et al., 2013) and further mining of such information will be valuable to understand the distinctive properties of niche blood vessels. For example, although VEGF is not normally produced by endothelial cells, during forebrain embryogenesis it is, and selective deletion of VEGF from endothelial cells perturbs cerebral cortical neurogenesis, cytoarchitecture and axon tract formation (Li et al., 2013). Brain endothelial cells also express high levels of the Notch ligands Jagged1, Jagged 2 and Dll4 compared to surrounding brain tissue (Daneman et al., 2010a) which may impact Notch receptor on contacting NPCs (Thomas et al., 2013).

Invading vasculature is accompanied by numerous support cells. Pericytes are more common along the neurovasculature than anywhere else in the body, even early in development. CNS pericytes likely originate from the neural crest, although pericytes may be derived from a number of sources (Armulik et al., 2011). Pericytes and vascular smooth muscle cells (vSMCs) accompany invading blood vessels in response to PDGFβ secreted by the endothelial cells and especially the tip cell. They undergo rapid proliferation from E11.5 to E14.5, which recedes by E18.5 (Hellström et al., 1999). During embryogenesis, pericytes are primarily involved in promoting BBB maturation (Daneman et al., 2010b) and thereby indirectly affect neurogenesis by limiting the blood-borne factors available to NPCs. As part of their role in angiogenesis, they secrete TGF-β (Dohgu et al., 2005) and IGF-2 (Lehtinen et al., 2011), although any direct effect on neurogenesis remains to be identified. A brain pericyte specific transcriptome should help identify candidate factors (Zhang et al., 2014). Low pericyte number in the VZ/SVZ is postulated to contribute to the vulnerability of these vessels to hemorrhage, which is a problem encountered in premature infants (Braun et al., 2007). The pericyte complement of the developing neurogenic zones still remains to be fully defined and characterized.

Meninges

The meninges encase and protect the CNS and are a source of critical neurogenic factors during embryogenesis (reviewed in Decimo et al., 2012). They contribute to the basement membrane (BM) of the glia limitans that delimits the outermost boundary of the CNS, providing an attachment point for the basal processes of NECs and RGCs. In the mouse, a reticulum of meningeal cells is evident over the telencephalon between E9–10 (McLone and Bondareff, 1975). Meningeal cells covering the forebrain originate from neural crest cells, while meninges covering more posterior CNS regions derive from cephalic and somatic mesoderm (Siegenthaler and Pleasure, 2011). Initially, the meninges form a thin sheet encasing the CNS, closely apposed to the underlying nervous tissue. The assembling angioblasts that will form the PNVP are also integrated in this sheet (Decimo et al., 2012). As the vasculature begins to infiltrate the brain, meningeal cells become more adherent, with increasing ECM anchorage to the underlying glia limitans. The original simple meningeal layer becomes stratified as it continues to proliferate, giving rise to the arachnoid layer around E13.5, at which point fibrous ECM begins to accumulate. By E14.5 the three layers of the meninges (see Fig. 1) can be distinguished: the pia mater immediately lining the CNS surface, the dura mater closest to the developing skull, and the arachnoid between them. By E16 the subarachnoid space expands, forming a fluid-filled cavity crisscrossed by fine arachnoid trabeculae composed of meningeal cells clinging to thin struts of ECM, and becomes invested with macrophages postnatally. By P21, meningeal cell proliferation is effectively complete. The pia mater follows the contours of the outer brain surface, and reaches into the CNS tissue (albeit still separated by the glia limitans) by lining recesses like the Virchow-Robin space that forms next to penetrating arteries (see Fig. 1) and by filling the interior of the ChPs (Decimo et al., 2012).

The meninges are absolutely essential for proper cerebral cortical neurogenesis through secreted signaling and ECM factors (Table 1). During early development, disruption of several BM components associated with the meninges, including perlecan and the nidogen binding domain of laminin, leads to RGC detachment and process retraction, along with incomplete or ectopic migration of both Cajal-Retzius and cortical plate neurons (Halfter et al., 2002). Similar results have been observed upon deleting RGC receptors for BM proteins, such as integrin β1 (Radakovits et al., 2009). The meninges are a rich source of RA which promotes cortical neuronal differentiation by stimulating RGC cell cycle exit (Siegenthaler et al., 2009); as seen with angiogenesis, the synthesis of RA begins laterally and moves medially between E12–13, matching the neurogenic gradient. Pial cells also secrete Cxcl12, a chemoattractant involved in guiding migrating Cajal-Retzius cells to the marginal layer of the cortex beneath the pial surface beginning at approximately E10 in mice (Borrell and Marín, 2006; Paredes et al., 2006), and subsequently (around E12–13 through E16) promoting proper positioning of GABAergic interneurons in the cortical plate (López-Bendito et al., 2008). Given its cellular complexity and the changing structure of the meninges during niche development, much remains to be understood regarding the role of meningeal-derived factors in embryonic neurogenesis.

Microglia

Microglia are the resident immune cells of the CNS, participating in both innate and adaptive immune responses. Groundbreaking research over recent years has begun to shed light on the origins, migratory routes and persistence of embryonic microglia (reviewed by Ginhoux et al., 2013). Microglia originate from the yolk sac around E7.5–8.5 in mice, migrating to the neural tube around E8.5–9.5 reliant on newly forming vasculature (Ginhoux et al., 2010). Here they congregate along both the outer surface and the inner ventricular wall (Fig. 1), proliferating and poised for invasion (Swinnen et al., 2013). A few microglia infiltrate the neural tissue parenchyma via ventricular or parameningeal routes prior to the onset of vascularization, but most begin to migrate concomitant with angiogenesis (Harry, 2013). Microglia colonize the brain in dorsal-to-ventral and rostral-to-caudal waves (Arnold and Betsholtz, 2013). The number of CNS microglia remains low between E10.5 and E14.5, with a rapid expansion from E14.5–15.5, followed by a reduced and relatively stable proliferation rate (Swinnen et al., 2013). From early colonization of the cerebral cortex until the end of corticogenesis, very few microglia inhabit the cortical plate; rather, most of them accumulate in the germinal zones (Cunningham et al., 2013; Swinnen et al., 2013).

Embryonic microglia display an amoeboid morphology as they migrate and proliferate in the cortex, with punctuated bouts of locomotion and interrogation, extending and retracting processes in a highly dynamic manner (Swinnen et al., 2013). By E16.5, roughly three quarters of the microglial population have extended a single projection or more, and eventually all adopt the classic highly branched or ‘ramified’ appearance of the mature, surveillant cells (Harry, 2013). Microglial invasion is region-specific; for example, in the spinal cord, it begins slightly later, around E11.5, when neuronal circuitry is being established and spontaneous activity is already observed (Rigato et al., 2011).

During embryogenesis, microglia play roles reflecting their phagocytic and secretory capacities. They secrete several growth factors and cytokines that may influence neurogenesis, such as TGFβ (Battista et al., 2006) (Table 1). In vitro, microglia cultured from E16–17 mice produce soluble factors that promote differentiation of cortical NPCs into neurons, however the identity of the factor(s) involved remains unknown (Aarum et al., 2003). Newborn rat ramified microglia produce secreted factors in vitro that promote NSC maintenance and astrocyte differentiation of E17 striatal NPCs by activating STAT3 via the Jak/Stat pathway (Zhu et al., 2008). Consistent with this, absence of microglia in embryonic mouse cortical cultures leads to decreased NPC proliferation and reduced astrocyte production, while increased microglia in culture leads to increased proliferation and astrogenesis (Antony et al., 2011). Fascinatingly, microglia phagocytose NPCs in the embryonic cortex of macaque and rat, even cells that do not appear to be undergoing apoptosis, suggesting that they control the number of NPCs contributing to brain growth (Cunningham et al., 2013).

After neurogenesis microglia are no longer concentrated in proliferative zones but are distributed throughout the cortex (Cunningham et al., 2013). They have been reported to accumulate near the forming deep layers in the cortex, and to secrete factors such as IGF1 that maintain survival of neurons (Ueno et al., 2013). Microglia also actively participate in later stages of neuronal circuitry formation by pruning synapses in an activity and complement-dependent manner (Schafer et al., 2012). Increased microglia activation and density have been reported in the autistic dorsolateral prefrontal cortex (Morgan et al., 2010), raising the possibility that they could lead to important changes in brain development, either through genetic variation or events such as infection that alter microglial activity.

Adult Subventricular Zone: Structure & Composition

Adult NSCs persist in a shallow niche along the striatal walls of the lateral ventricles, bounded on one side by the ependymal surface lining the CSF-filled ventricles, and on the other by a planar SVZ vascular plexus (Mirzadeh et al., 2008; Shen et al., 2008) (Fig. 2). Quiescent NSCs (type B cells) become activated to divide, giving rise to type C transit amplifying progenitor cells. Type C cells undergo rapid divisions and give rise to neuroblasts (type A cells) or glia (oligodendrocytes or astrocytes). Guided by ensheathing astrocytes that form channels, as well as by the vasculature in some regions of the SVZ, type A cells assemble into long chains that migrate toward the olfactory bulb, where they integrate into the existing neuronal circuitry as periglomerular or granule interneurons (Kriegstein and Alvarez-Buylla, 2009).

Fig. 2. Cellular components of the adult SVZ niche.

The niche is situated at the striatal side of the lateral ventricles (box, top left). The niche is innervated by axons from the substantia nigra and the raphe nucleus innervate (bottom left), as well as SVZ Chat⁺ neurons. B: type B cells (NSCs); C: type C cells; A: type A cells; m: microglia; BV: blood vessel; p: pericyte; n: neuron; ep: epiplexus cell; a: astrocyte; f: fibroblast. Top right brain section modified from Allen Institute for Brain Science.

Type B cells exhibit hybrid characteristics of astrocytes and immature progenitors (Kriegstein and Alvarez-Buylla, 2009). Type B cell somata typically rest just beneath the ependymal cell layer, and many exhibit polarity reminiscent of their embryonic precursors, with a basal process on the SVZ vasculature, and a thin apical process projecting through the ependymal cell layer to contact the CSF directly. As a result, these cells are poised to receive cues from both vascular and CSF compartments. The apical process, which often includes a single cilium, is decorated with molecules that could sense environmental cues, such as VCAM1 (Kokovay et al., 2012). Type B cell apical processes can form bundles at the center of a ‘pinwheel’ of ependymal cells, a structure seen tiled repeatedly on the ventricular surface (Mirzadeh et al., 2008).

Thus, in many respects, the adult SVZ niche can be considered a continuation of the embryonic VZ/SVZ but with differences due to state of maturation and location. For example, ependymal cells mature postnatally and are more important in the adult SVZ, and the adult SVZ is invested with mature innervation. On the other hand, the meningeal layers are distant from the adult SC niches, and play a less prominent direct role compared to the embryonic niche. Next, we summarize some of the key niche components that have a specialized role in regulating adult neurogenesis.

Innervation

The SVZ is directly innervated from dopaminergic projections from the substantia nigra and the ventral tegmental area (Baker et al., 2004). Additionally, adult SVZ type C cells express D2-like dopaminergic receptors. Conflicting data have been presented on the effects of dopamine on adult NPCs, with dopamine depletion reducing proliferation, which was restored by activation of D2 receptors (Höglinger et al., 2004), while haloperidol, a D2 receptor antagonist, was shown to increase NPC proliferation (Kippin et al., 2005). An attempt to reconcile these seemingly opposite results is the theory that dopamine affects type B and C cells in different ways (Berg et al., 2013), for example, via a different mix of receptors. D2 stimulation of SVZ proliferation appears dependent on CNTF, an NPC mitogen expressed by astrocytes that is upregulated by D2 agonists (Yang et al., 2008).

Differentiated neurons that express nitric oxide synthase are found near the adult SVZ, and low doses of nitric oxide increased NPC proliferation while higher doses inhibited it. This effect is seen in adult but not earlier postnatal SVZ, indicating it is a regulatory mechanism acquired with maturation (Romero-grimaldi and Morenolo, 2008).

Recently it has been discovered that the adult mouse SVZ is also innervated by serotonergic axons from a small group of neurons in the raphe and by choline acetyl transferase positive (ChAT⁺) axons originating in the SVZ itself. Serotonergic axons form an extensive network along most of the ventricular surface and physically contact NSCs and ependymal cells. Type B cells express serotonin receptors, activation of which was shown to increase NPC proliferation in the SVZ, while their inactivation had the opposite effect (Tong et al., 2014). SVZ ChAT⁺ neurons were shown to release acetylcholine locally in an activity-dependent fashion. Activation and inhibition experiments on these neurons showed that they directly control SVZ proliferation, an effect attenuated by FGF receptor activation (Paez-Gonzalez et al., 2014) (Fig. 3).

Ependymal cells

There are no fully differentiated ependymal cells during embryogenesis, but these cells are critical components of the adult SVZ niche. These cells have microvilli and tufts of motile cilia that contribute to CSF hydrodynamic flow (Spassky et al., 2005). Mature ependymal cells form a simple cuboidal to low columnar epithelium, connected via gap junctions, and are permissive to intercellular fluid transport. Elaborate adherens junctions but lack of tight junctions between adjoining lateral membranes, as well as expression of several channel proteins, suggest they participate in moderating exchange between CSF and the interstitial fluid (ISF).

Whether ependymal cells can act as NSCs has long been contested (Chojnacki et al., 2009), and even the proliferative capacity of ependymal cells in adult mice remains controversial. Thorough analyses by Spassky et al., 2005 indicate that differentiated ependymal cells are postmitotic. As damage to the ependymal layer accrues with aging, type B cells shore up gaps, adopting an ependymal cell phenotype (Luo et al., 2008). Canonical Notch signaling among adult mouse ependymal cells may enforce quiescence, although severe insults such as stroke could interrupt this state, stimulating re-entry into the cell cycle and promoting neurogenesis (Carlén et al., 2009).

Ependymal cells affect adult neurogenesis in a number of ways. They express Noggin, a BMP antagonist (Lim et al., 2000). The effects of BMP signaling on adult neurogenesis remain controversial, as studies have claimed either that it inhibits neurogenesis (Lim et al., 2000) or promotes it. Such conflicting results have led to hypotheses that BMP signaling either affects the various adult NPCs differently, or that it has a dose-dependent effect (Colak et al., 2008). Ependymal cells also secrete factors that affect neurogenesis. For example, they secrete PEDF, which was shown to promote NSC self-renewal (Ramírez-Castillejo et al., 2006) (Fig. 3).

Choroid plexus and cerebrospinal fluid

In the adult, the lateral ventricle ChPs run most of the length of the ventricle alongside the SVZ. Fluid flow is thought to travel from the neurovasculature to the ISF and then to drain into the CSF, however changes in hydrostatic or osmotic pressure can reverse this flow (Redzic et al., 2005). Thus, the SVZ is uniquely situated to experience the ebb and flow of tides and signals from vascular flow, ISF drainage, and CSF production. Intriguingly, migrating neuroblasts in the SVZ travel in the direction of CSF flow, and neuroblast chains become disoriented when motile cilia that propel CSF are disrupted (Sawamoto et al., 2006).

The adult ChP is responsible for continuously producing and modifying the bulk of the CSF, which is rich in nutrients and growth factors (Redzic et al., 2005). Recent transcriptome studies on adult ChP (Marques et al., 2011) and comparing embryonic to adult ChP (Liddelow et al., 2012), have revealed expression profiles rich in secreted growth factors and signaling molecules, several with known effects on neurogenesis, as well as chemorepulsive signals including Slit1/2 that help press adult neuroblasts to migrate rostrally (Nguyen-Ba-Charvet et al., 2004). It secretes mitogens such as FGF-2, IGF2, amphiregulin and TGFα, as well as TGFβ superfamily members which can act as negative regulators of neurogenesis (Table 1).

With aging, the ChP shows a marked type 1 interferon dependent gene expression profile, seen in mouse and humans. Blocking this pathway within the aged brain helped restore cognitive function and hippocampal neurogenesis, demonstrating the critical role of the ChP in aging-related changes in adult SC behaviors (Baruch et al., 2014).

Neurovasculature, endothelial cells, pericytes

In the adult SVZ, NPCs are ‘sandwiched’ between the lateral ventricle wall and a roughly parallel SVZ vascular plexus, suggesting that SVZ progenitor cells reside in a vascular niche (Shen et al., 2008; Tavazoie et al., 2008). Architecturally, blood vessels in the dorsal SVZ can serve as a migratory scaffold to guide neuroblast chains as they migrate dorsally and anteriorly along the rostral migratory stream and later within the olfactory bulb itself (Nam et al., 2007; Engelhardt and Liebner, 2014). The blood vessels are oriented to create a vascular scaffold following a path laid out by VEGF-secreting astrocytes, and disruption of VEGF impairs scaffold formation and neuron migration into the olfactory bulb (Bozoyan et al., 2012).

Actively dividing NPCs associate closely with the vascular endothelium. Endothelial cells have been shown to promote self-renewal and subsequently to expand neurogenesis of adult NSCs (Palmer et al., 2000; Shen et al., 2004). Transplanted SVZ NPCs home to the vascular niche, attracted by endothelial secreted SDF1 (Kokovay et al., 2010). The vasculature is a rich source of secreted factors that promote NPC proliferation, migration, chemotaxis and adhesion to endothelial cells (Table 1). For example, endothelial cells express PEDF, which promotes SVZ type B cell renewal in vitro, and intraventricular injections of PEDF increase the number of activated type B cells within the SVZ (Ramírez-Castillejo et al., 2006). Betacellulin secreted by vascular endothelia regulates proliferation of EGFR+ NPCs and type A neuroblasts (Gómez-Gaviro et al., 2012). Vascular endothelial cells in the SVZ plexus and in the ChP produce NT3 which is required for the maintenance of adult NSCs, acting by promoting production of nitric oxide which is a cytostatic factor (Delgado et al., 2014). Nitric oxide contributes to production of reactive oxygen species that are higher in NSCs and important for NSC self-renewal (Le Belle et al., 2011). In addition, circulating blood carries factors that can influence neurogenesis, including hormones, cytokines, metabolites and gases.

Pericytes are present along the adult SVZ vascular plexus, where they may regulate capillary blood flow and thus access to blood-borne metabolites and neurogenic signals (Armulik et al., 2011). In addition, they continue to secrete growth factors including TGF-β to maintain BBB homeostasis (Dohgu et al., 2005). Pericytes remain relatively undifferentiated, and during adulthood it has been suggested that they retain the capacity to differentiate into fibroblasts, vSMCs, macrophages, and even neural-related progeny (Dore-duffy and Cleary, 2011); although prior studies have refuted the idea that mesenchymal cells can generate bona fide neural lineages, this might be possible if such pericytes have a neural crest origin. The identification of the various pericyte populations is hampered by a paucity of cell-specific markers and by the presence of other cell types including fibroblasts and macrophages in the periendothelial space (see Armulik et al., 2011 for an excellent review).

ECM fractones

In the lateral and third ventricles of the adult rat, ECM extensions known as fractones have been identified that project from the blood vessels of the subventricular plexus as thin highly-branching ECM stalks that expand into bulbs where they contact the basal surface of the ependymal layer (Mercier et al., 2002). Fractones are enriched in laminin, heparan sulfate, perlecan, nidogen and collagens (Kerever et al., 2014; Mercier et al., 2002). In addition to a possible structural role in guiding cell migration, they bind several growth factors (Table 1), suggesting fractones may play a role in concentrating, activating, and presenting trophic factors to cells within the niche.

Microglia

Microglia undergo a second expansion shortly after birth and by postnatal day 14, approximately 95% of the total population have been generated (reviewed in Ginhoux et al., 2013). In the adult, including in neurogenic zones, microglia are ubiquitous and ‘tiled’ - establishing non-overlapping territories which they continuously survey. Adult microglia serve both neuroprotective and neurotoxic roles depending on the circumstances. They secrete a variety of neurotrophic factors including NGF, BDNF, NT3, GDNF, bFGF, HGF, plasminogen, and cytokines, typically in response to environmental insults and in the case of NT3 and NGF, in a brain region-specific manner (reviewed in (Harry, 2013; Kim and de Vellis, 2005).

A prominent role for microglia in neurogenesis in healthy adults has been established in the hippocampus. Most newborn hippocampal granule neurons undergo apoptosis, while very few survive beyond two weeks and integrate into the existing neural circuitry. Surveillant (not activated) microglia phagocytose apoptotic newborn cells (Sierra et al., 2013), and may secrete pro- or anti-neurogenic factors depending on the environmental context (Table 1). Beyond the role of surveillant cells in regulating neurogenesis, microglial activation and inflammation can result in decreased proliferation and differentiation of NPCs (Sierra et al., 2014). In the adult SVZ, microglia have been shown to play a role in promoting neural differentiation in vitro (Walton et al., 2006), although for the most part microglia appear to play more of a role in stimulating neurogenesis and recruiting newly formed neurons to sites of CNS injury. Microglia in the aging CNS become less active and motile as they senesce, and appear less capable of removing pathogens and debris. They assume a more reactive profile, with elevated expression of pro-inflammatory cytokines including TNFα, TNFβ, IL-1α, IL-1β, and IL-6, along with reduced expression of anti-inflammatory cytokines including IL-10 (reviewed in Harry, 2013). There are clear indications that these factors contribute to the decline of neurogenesis seen with aging (Gemma et al., 2010). Revitalization of neurogenesis in aged animals can occur by replacing some of the non-neural elements, as seen by refreshing blood borne factors with anastomosis of young and old mice (Katsimpardi et al., 2014), and restoration of microglial function would be an interesting test of their contribution to the aged niche.

Concluding Remarks

Neurogenesis is a marvelously intricate process. In the embryo, neurogenic zones develop in concert with, and are instructed by, signals that arise from early mesendodermal structures. With elaborate choreography, further derivatives of these different germ layers arise, invade and interact, building the complexity of the niche with heterogeneous components until its fully mature state, which then suffers the wear and tear of aging. The result is an ever-changing diorama of shifting cells and cell types, miraculously accomplishing a task that to date all human expertise and technology cannot replicate artificially: building an integrated, multi-cellular neural system. Research over the past two decades has illuminated the diverse types of stem and progenitor cells present in the developing and adult niches. More recently, the impact of surrounding tissues, innervation and non-neural elements has been revealed. Now with the application of techniques that measure changes in gene and protein expression and molecular modifications down to the single cell level, we are poised to reveal far more about the sources and actions of molecules regulating the niche, for both those that are present continuously and others that define unique stages of niche development, maturation, homeostasis and aging. Teasing apart the complex functional interactions between these different niche elements will be no small challenge, but highly valuable. Motivating a deeper understanding of neurogenesis is its therapeutic potential. Many studies described here have demonstrated increased neurogenesis in response to the administration of defined exogenous niche factors. Most of these studies have been conducted in vitro or in vivo in the short-term, and given the complexity of the system and the amount of negative feedback loops in place to deter rampant neurogenesis, a key question is whether those manipulations would have the same effect in vivo and over the longer term. Nevertheless, we should be planning research around these opportunities, as niche rejuvenation is a distinct possibility, and worth exploring given the rising prevalence of neurodegenerative illness.

List of abbreviations

Acronyms

BBB

blood-brain barrier

BM

basement membrane

CNS

central nervous system; a term that, in vertebrates, is used to describe the brain and spinal cord

CSF

cerebrospinal fluid; the fluid that fills the ventricles of the brain

E

embryonic day; days post fertilization

IPC

intermediate progenitor cell; a term used to describe a neural progenitor cell arising from radial glial cells and residing in the embryonic subventricular zone

ISF

interstitial fluid

NEC

neuroepithelial cells; the stem cells of the nervous system residing in the ventricular zone

NPC

neural progenitor cell; a term used in this text to describe neural stem cells and their undifferentiated progeny

NSC

neural stem cell; multipotent precursors with the capacity for self-renewal and differentiation into neural cell types

PNVP

perineural vascular plexus

RA

retinoic acid

RGC

radial glial cells; bipolar cells whose somata reside in the ventricular zone but whose projections span the width of the cortex and act as neural stem cells; the fate of their progeny is more restricted than that of neuroepithelial cells

SC

stem cell; undifferentiated cells that are capable of self-renewal and production of certain more differentiated cell types

Shh

Sonic hedgehog

SVP

SVZ plexus

SVZ

subventricular zone; an adult brain structure situated at the lateral walls of the lateral ventricles

vSMC

vascular smooth muscle cells

VZ/SVZ

ventricular and subventricular zone; a neurogenic region in the developing brain that is in contact with the ventricular system in the brain