Gliogenic and neurogenic progenitors of the subventricular zone: Who are they, where did they come from, and where are they going?


  • Christine A.G. Marshall,

    1. Center for Neurobiology and Behavior, Division of Neuropathology, Department of Pathology, Columbia University, College of Physicians and Surgeons, New York, New York
    Search for more papers by this author
  • Satoshi O. Suzuki,

    1. Center for Neurobiology and Behavior, Division of Neuropathology, Department of Pathology, Columbia University, College of Physicians and Surgeons, New York, New York
    Search for more papers by this author
  • James E. Goldman

    Corresponding author
    1. Center for Neurobiology and Behavior, Division of Neuropathology, Department of Pathology, Columbia University, College of Physicians and Surgeons, New York, New York
    • Division of Neuropathology, Department of Pathology, Columbia University, College of Physicians and Surgeons, P&S 15-420, 630 West 168th Street, New York, NY 10032
    Search for more papers by this author


The subventricular zone (SVZ) of the perinatal forebrain gives rise to both neurons and glia. The mechanisms governing the phenotypic specification of progenitors within this heterogeneous germinal zone are unclear. However, the characterization of subpopulations of SVZ cells has given us a better understanding of the basic architecture of the SVZ and presents us with the opportunity to ask more detailed questions regarding phenotype specification and cell fate. Recent work demonstrating the embryonic origins of SVZ cells is summarized, and a model describing the formation of the perinatal SVZ, noting contributions of cells from pallial as well as subpallial germinal zones, is presented. We further address differences among classes of SVZ cells based on molecular profile, phenotype, and migration behavior and present a model summarizing the organization of perinatal SVZ cells along coronal, sagittal, and horizontal axes. A detailed description of the SVZ in the adult, outlining classes of cells based on morphology, molecular profile, and proliferative behavior, was recently reported by Doetsch et al. (Proc Natl Acad Sci USA 93:14895–14900, 1997). Potential relationships among cells within the perinatal and adult SVZ will be discussed. GLIA 43:52–61, 2003. © 2003 Wiley-Liss, Inc.


Allen (1912) reported the presence of mitotic activity within subependymal cells of the postnatal rat forebrain nearly 100 years ago. Later studies of human fetuses (Rydberg, 1932; Kershman, 1938) demonstrated a similar collection of undifferentiated cycling cells, superficial to the ependymal layer, lining the lateral ventricles. Opalski (1933) and Kershman (1938) suggested that this subependymal layer persists into adulthood as a smaller vestigial layer and retains the ability to produce new cells. For this reason, the subependymal layer was considered a potential source of neoplastic cells in mature humans (Globus and Kuhlenbeck, 1944). Mitotically active subependymal cells were characterized further in rodents (Bryans, 1959; Smart, 1961; Lewis, 1968b; Altman, 1966, 1969; Privat and Leblond, 1972; Sturrock, 1985) and primates (Lewis, 1968c).

Microscopic analysis of the subependymal layer revealed a division among cells based on morphology (Smart, 1961; Privat and Leblond, 1972). Cells in a border area, separating the subependymal layer from the corpus callosum and caudate, possessed light-staining nuclei, fine processes, and glycogen granules in their cytoplasms, while many cells within the subependymal layer proper had darker-staining nuclei with smaller, round cell bodies (Privat and Leblond, 1972). Darker-nucleated cells typically possessed a higher labeling index relative to light border cells, as evidenced by the rate of tritiated (3H) thymidine incorporation (Smart, 1961). The distribution of these cell types is modulated throughout development, such that the neonatal ratio of dark (91%) to light (9%) cells decreases as the animal matures. By adulthood, light nucleated cells comprise roughly 45% of subependymal cells. Interestingly, some cells, which were thought to be ependymal cells because of their contribution to the ventricular wall, have nuclei in a more subependymal position and contain glycogen granules, a typical feature of astrocytes (Privat, 1972). These cells may, in fact, correspond to the type B astrocytic cells currently thought to possess qualities of neural stem cells (Doetsch et al, 1999).

In an effort to adopt consistent nomenclature among developmental neurobiologists, the Boulder Committee (1970) used “subventricular zone” (SVZ) to describe the layer of cells generated superficial to the embryonic ventricular zone (VZ) after cortical plate formation. Cells within the SVZ do not undergo interkinetic migrations and were thought to be relatively less pluripotent than neighboring VZ cells (Sidman and Rakic, 1973; Takahashi et al., 1995). Perinatally, the SVZ completely surrounds the lateral ventricles. Furthermore, it forms a triangular prominence, bordered by the subcortical white matter and the developing striatum, at the dorsolateral tip of the ventricles. This structure exists at the level of the optic chiasm and can best be visualized in the coronal plane (Allen, 1912).

The SVZ is highly dynamic in nature. It undergoes an exponential expansion in thickness during the perinatal period, followed by a marked postnatal reduction in size. Hence, the precise anatomic characterization of the SVZ has been elusive. We present a model for the formation of the perinatal SVZ, noting contributions of cells from pallial as well as subpallial germinal zones. Furthermore, we address differences among classes of SVZ cells based on phenotype and migration behaviors and offer a model summarizing the organization of perinatal SVZ cells along coronal, sagittal, and horizontal axes. A detailed model of the adult SVZ, outlining classes of cells based on morphology, molecular profile, and proliferative behavior, was recently prepared by Doetsch et al. (1999). Potential relationships among cells within the perinatal and adult SVZ are discussed.


It has been inferred that SVZ cells are derived from underlying VZ cells and migrate directly into the cerebrum (Rakic, 1974, 1988). This holds true for radially migrating cells, but many progenitors migrate tangentially within the SVZ before emigrating into the overlying parenchyma (Halliday and Cepko, 1992; Kakita and Goldman, 1999). Hence, these cells colonize areas of the forebrain far removed from the parent VZ cell. Increasing evidence of tangential cell migration and mixing of progenitors from germinal zones along the dorsoventral axis raises new questions regarding the embryonic origins of SVZ cells. Studies of tangential migration pathways have been reviewed elsewhere (Marin and Rubenstein, 2001; Corbin et al., 2001; Maricich et al, 2001) and therefore are not addressed in detail. In summary, cells originating within the ventral basal ganglia migrate dorsally along tangential routes into the cerebral cortex and hippocampus from early (embryonic day 11) to late (embryonic day 14–16) stages of embryogenesis (Anderson et al., 2001). While early migrating cells take a lateral route around the SVZ into the intermediate zone of the cortex, late migrators follow a medial pathway within the SVZ from ventral to dorsal areas. Hence, cells generated within the ganglionic eminences migrate into the dorsolateral SVZ en route to the cortical SVZ and hippocampus. Some of these migrating progenitors express the ventral forebrain markers Dlx1/2 (Anderson et al., 1997a,b) and give rise to interneurons, but the fate of the entire dorsally migrating population is unknown. As medial tangential migrations from the ganglionic eminences continue into the postnatal period, coinciding temporospatially with the expansion of the dorsolateral SVZ, it is reasonable to believe that progenitors from ventral regions might contribute to this prominence.

Recent analyses using molecular markers have identified two populations of cells that comprise a large majority of the perinatal dorsolateral SVZ. One population is composed of large, polygonal shaped cells situated at the borders of the SVZ that express Zebrin II (aldolase C), a brain-specific isoform of fructose 1,6-bisphosphate aldolase. These stationary Zebrin II-expressing border cells possess fine processes and resemble the light nucleated cells previously described (Smart, 1961; Privat and Leblond, 1972). Such cells are morphologically and molecularly distinct from a population of smaller, mainly unipolar, migratory progenitors that reside in the central SVZ. For the most part, these unipolar progenitors do not express Zebrin II (Staugaitis et al., 2001).

While Zebrin II-expressing cells at the SVZ borders appear to be derived from the residual VZ, the origin of the central migratory population was unclear. It was hypothesized that these Zebrin II cells originate in more ventral locations and migrate dorsally into the SVZ (Marshall and Goldman, 2002). Many cells within the perinatal SVZ then migrate further to give rise to astrocytes and oligodendrocytes throughout the cerebral cortex, white matter, and striatum (Levison and Goldman, 1993). Glial precursors within the white matter of the perinatal telencephalon were found to be immunoreactive for a pan-Dlx antibody, suggesting that some oligodendrocytes may be derived from Dlx1/2-expressing subpallial cells (He et al., 2001). Hence, a Dlx2/tauLacZ knock-in mouse (Corbin et al, 2000) was used to perform short-term lineage analysis of subpallial-derived Dlx2-expressing cells, tracing the fates of these cells and their progeny well into postnatal development (Marshall and Goldman, 2002). Migratory Zebrin II cells were identified as the progeny of Dlx2-expressing subpallial cells. Furthermore, Dlx2 descendants emigrate postnatally from the SVZ and develop into astrocytes and oligodendrocytes within the cerebral cortex, white matter, and striatum. In summary, cells derived from the ganglionic eminences migrate dorsally and intermix with Zebrin II-expressing neuroepithelial cells at the corticostriatal sulcus to form the dorsolateral SVZ. Dlx2 descendants then give rise to glia in the dorsal telencephalon (Fig. 1A).

Figure 1.

Formation of the perinatal subventricular zone (SVZ). A: E16: Neuroepithelial cells (green) express Zebrin II. A Zebrin II+ wedge of neuroepithelium emerges during corticostriatal sulcus formation. This pallial wedge is fenestrated by migratory Dlx2+ subpallial-derived cells and their progeny; both can be defined by Dlx2/tauLacZ expression (blue). P0: Cells derived from subpallial Dlx2+ cells form the central region of the perinatal SVZ. B: SVZ progenitors diverge into separate lineages. The stage at which this lineage arborization occurs is unclear.


Forebrain expansion during embryogenesis appears to contribute to the intermixing of these cell populations (Marshall and Goldman, 2002). As the lateral ganglionic eminence (LGE) enlarges during mid-gestation (E12–14), the pallial VZ folds at an acute angle, forming the corticostriatal sulcus. The sulcus creates a wedge-shaped zone of Zebrin II-expressing VZ cells (Fig. 1A, E16). During the subsequent perinatal week, this area undergoes a transformation such that Zebrin II-expressing cells become distributed at the medial, dorsal and lateral periphery of the SVZ, with a large number of Zebrin II cells populating the central region (Staugaitis et al., 2001; Marshall and Goldman, 2002). The VZ wedge becomes fenestrated as cells from subpallial areas invade it, with the most dorsolateral Zebrin II+ cells becoming displaced from more medial VZ wedge cells (Fig. 1A, P0). Analyses of the Zebrin II expression pattern at each day throughout the perinatal and early postnatal weeks reveals an accumulation of these migrating cells, progressively displacing the Zebrin II+ residual neuroepithelium laterally. This embryologic displacement should not be confused with migration. There is no evidence suggesting that these wedge cells are migratory. Rather, the lateral wedge cells, which form the dorsolateral tip of the SVZ, appear to be relatively stationary.


How do embryonic origins of SVZ cells influence their specification? The SVZ is composed of a heterogeneous mixture of cells from different lineages. Some SVZ cells become specified as astroblasts or oligodendroblasts (Levison and Goldman, 1993; Luskin and McDermott, 1994; Parnavelas, 1999). Others remain uncommitted as glioblasts until they migrate into the overlying parenchyma, where they diverge into lineages of astrocytes or oligodendrocytes (Levison et al., 1993; Zerlin and Goldman, submitted). Many SVZ cells appear to commit to a neuronal lineage and take a rostral migratory route toward the olfactory bulb, where they give rise to interneurons (Luskin et al., 1993; Lois and Alvarez-Buylla, 1994). Still, a subset of SVZ cells may remain multipotent (Fig. 1B). Clonal analyses of progenitors isolated from the LGE (He et al., 2001) and the postnatal dorsolateral SVZ (Levison and Goldman, 1997) demonstrate the potential of many of these immature cells to give rise to both neurons and glia in vitro. Phenotypic specification may occur within the SVZ or even after emigration from the SVZ. In vivo, instructive or permissive factors enable some precursor cells to migrate into the developing cerebrum while others remain within the SVZ. The extent to which a relationship exists between migration pathways and fate specification remains unknown.


Recent work has described a ventral origin and specification of telencephalic oligodendrocytes during embryonic development (for review, see Woodruff et al., 2001). Before birth, progenitors isolated from the rat striatum have a much greater competence to generate oligodendrocytes in vitro than those harvested from cerebral cortex. However, progenitors from the postnatal cerebral cortex have the potential to give rise to significant numbers of oligodendrocytes (Birling and Price, 1998). One popular interpretation of this phenomenon is that all oligodendrocyte precursors are specified by Shh and migrate into the dorsal telencephalon during the course of embryogenesis (Woodruff et al., 2001). This interpretation is based on studies in which a gradient of DM-20, an alternatively spliced isoform of myelin proteolipid protein, and platelet-derived growth factor receptor-α (PDGFR-α) expression emanates from the anterior entopeduncular area (AEP) and extends into select regions of the telencephalon (Pringle and Richardson, 1993; Spassky et al., 1998; Tekki-Kessaris et al., 2001). Further work using chick/quail and chick/mouse chimeras demonstrates that many of these early oligodendrocyte precursors, which are specified and differentiate during embryogenesis, originate within the AEP (Olivier et al., 2001). Moreover, this early population of oligodendrocytes colonizes the pallidum, striatum, lateral forebrain bundle, and lateral and ventral pallial areas of host telencephalon. However, AEP cells do not contribute oligodendrocytes to dorsal regions of the cerebral cortex.

While this model provides insight into embryonic oligodendrogenesis, it does not account for those cells specified during the peak of gliogenesis within the first few postnatal weeks. Furthermore, it is unclear whether all oligodendrocytes are specified ventrally. Postnatally, progenitors emigrate from the SVZ into the striatum and white matter as well as the medial, dorsal, and lateral regions of the cerebral cortex, where some develop into oligodendrocytes (Levison and Goldman, 1993; Levison et al., 1993; Luskin and McDermott, 1994). Many of these cells are not irrevocably committed to an oligodendrocyte fate, as SVZ progenitors generate clones containing both astrocytes and oligodendrocytes in vivo (Levison and Goldman, 1993; Parnavelas, 1999; Zerlin and Goldman, submitted for publication) and mixed neuronal-glial clones in vitro (Levison and Goldman, 1997).

How can one reconcile the ventral specification of oligodendrocytes with the emergence of both astrocytes and oligodendrocytes from the postnatal SVZ? In one possible model, progenitors originate in the ventral telencephalon and migrate dorsally into the SVZ. Some are specified as oligodendrocyte precursors, while others represent astrocyte or neuronal precursors, or perhaps remain uncommitted. Ventral progenitors, which are competent to form oligodendrocytes but do not yet express common oligodendrocyte-specific markers, appear to migrate dorsally into the forming dorsolateral SVZ (Marshall and Goldman, 2002). Pringle et al. (1992) observed a lack of PDGFR-α expression by SVZ cells and an upregulation of the receptor only by cells that migrated into the surrounding parenchyma. Hence, it is possible that cells originating ventrally become competent to give rise to oligodendrocytes by exposure to extrinsic patterning factors such as Shh without committing to an oligodendrocyte lineage or necessarily expressing early oligodendrocyte markers such as DM-20 or PDGFR-α. These oligo-competent cells may remain as multipotent progenitors or glioblasts until they have migrated dorsally and taken residence within the SVZ. Alternatively, extrinsic factors within the postnatal SVZ, or even within the white matter or cortex, might serve to specify oligodendrocytes locally.


The perinatal SVZ also generates astrocytes, which colonize the cerebral cortex, white matter, and striatum (Smart, 1961; Lewis, 1968a; Privat and Leblond, 1972; Paterson et al., 1973; Levison and Goldman, 1993; Luskin and McDermott, 1994). Zebrin II precursors that emerge from the central perinatal SVZ initiate Zebrin II expression as they differentiate into astrocytes but not oligodendrocytes (Staugaitis et al., 2001). Indeed, Zebrin II is specifically expressed by astrocytes in the telencephalon as well as Bergmann glia, Purkinje cells, and astrocytes in the cerebellum of adult mammals, including humans (Thompson et al., 1982; Kumanishi et al., 1985; Ahn et al., 1994; Walther et al., 1998). The perinatal SVZ is a secondary source of astrocytes, temporospatially distinct from the embryonic ventricular zone (Luskin et al., 1988; Price and Thurlow, 1988), which generates astrocytes via a radial glial phenotype (Voigt, 1989). Although astrocytes derived from both sources express glial fibrillary acidic protein (GFAP) and form associations with blood vessels and synapses, it is possible that they possess qualities making them distinct from one another. The comparison of diverse classes among Zebrin II-expressing astrocytes may enable us to understand better the roles astrocytes serve in the normal and pathologic CNS.


Within the SVZ, migrating progenitors express markers for undifferentiated neural populations, such as polysialylated-NCAM (PSA-NCAM), GD3 ganglioside, and the complex gangliosides recognized by the monoclonal antibody, A2B5 (Levison and Goldman, 1997; Ben-Hur et al., 1998; Marshall and Goldman, 2002). Neuronal progenitors that give rise to olfactory interneurons express neuronal marker such as class III β-tubulin (Menezes and Luskin, 1994) and GABAA receptors (Ma and Barker, 1998; Stewart et al., 2002), even though they are still cycling and migrating in the SVZ and the rostral migratory stream (RMS). Radial glia are positive for astrocytic markers such as vimentin, nestin, GLAST, intermediate filament-associated antigen recognized by RC2 antibody, and brain lipid-binding protein (BLBP) (Misson et al., 1988; Feng et al., 1994; Shibata et al., 1997; Hartfuss et al., 2001; Chanas-Sacre et al., 2000). GFAP is expressed by radial glia in human and primate (Levitt et al., 1981; Choi, 1981), but not in rodent (Pixley and de Vellis, 1984) or ferret (Voigt, 1989). Astrocytic and oligodendrocytic markers, even those considered early markers, i.e. vimentin and GLAST for the former, O4, PDGFR-α and NG2 for the latter, are expressed largely after progenitors migrate out of the SVZ into the overlying white matter or the adjacent striatum (Pringle et al., 1992; Zerlin et al., 1995; Staugaitis et al., 2001). Thus, it is difficult to identify and locate progenitors that are developing into either astrocytes or oligodendrocyte until they migrate out of the SVZ.


Glial Progenitors Migrate Primarily Within a Coronal Plane

Glia, which migrate and differentiate mostly in the postnatal period, do not colonize the neocortex in a well-organized laminar pattern, like that of neurons. Nevertheless, their migration paths from the SVZ are not random. When replication-deficient retrovirus expressing a reporter gene was placed in the dorsolateral SVZ at the coronal level crossing the septal nuclei, labeled cells migrate out of the SVZ into the dorsal white matter and cortex, the striatum, and, through the lateral migratory stream (LMS), into the lateral white matter and cortex. Their migration follows the direction of radial fibers and is primarily confined to a coronal plane, perpendicular to the rostrocaudal axis of the SVZ (Fig. 2, P10). All the labeled cells that settled in the cortex, white matter and striatum gave rise to either astrocytes or oligodendrocytes as determined morphologically and immunohistochemically (Levison and Goldman, 1993; Zerlin et al., 1995; Kakita and Goldman, 1999). Therefore, we infer that this radial migratory pattern is closely associated with glial fates. However, we cannot completely exclude the possibility that late-born neuronal progenitors also take this pathway, but silence the retrovirally transduced reporter gene (Pannell and Ellis, 2001; Svoboda et al., 2000). On occasion, a very small number of neurons were generated in the neocortex by SVZ cells (Levison and Goldman, 1993).

Figure 2.

Migration of subventricular zone (SVZ) cells in the neonatal and adult brain. P10: Astrocytic and oligodendrocytic progenitors, at least partly derived from Dlx+ cells, do not actively migrate rostrocaudally within the SVZ. They start emigration (green lines, shown for just one coronal plane) from the SVZ (yellow zones) toward the white matter, cortex and the striatum utilizing local radial glial fibers as migratory scaffolds. Neuronal progenitors, which are polysialylated-NCAM (PSA-NCAM)+/Tuj1+ (blue line), migrate rostrally within the SVZ all over the wall of the lateral ventricle toward the olfactory bulb to give rise to periglomerular and granular interneurons. They remain within the SVZ until they reach the olfactory bulb, where they turn and migrate radially into the olfactory cortex. Both glial and neuronal progenitors may migrate bidirectionally until they settle in their final destination. Adult: The genesis of neuroblasts and their migration toward the olfactory bulb continues throughout life. In adulthood, the route of neuronal migration appears as an anastomosing migratory network (blue lines) within the SVZ all over the lateral wall of the lateral ventricle, ensheathed by astrocytes derived from Zebrin II+ neuroepithelial cells.

Time-lapse video imaging has allowed a direct observation of migrating cells and thus an examination of the details and kinetics of glial progenitor migration. Nuclear translocation always occurred in the direction of leading process extension, often leaving a thin trailing process behind the cell body. Sometimes movement of process and nuclear translocation were linked, but other times they were independent. Cells moved bidirectionally, occasionally migrating back in the direction of the SVZ. A small population turned and migrated tangentially in the cortex, parallel to the pial surface. Migratory velocities were not constant but were saltatory with periods of inactivity interspersed between migratory spurts. The mean velocity was about 90 μm/h: therefore, in 3 days after viral infection, we observed a widespread distribution of the labeled cells, with some of them even reaching close to the pia (Kakita and Goldman, 1999).

A GFP-encoding retrovirus was injected at various coronal levels of the SVZ to observe the overall migration pattern. Radial migration occurred from all over the SVZ from the anterior part immediately upstream of the RMS, referred to as the SVZa (Luskin, 1993), to the coronal level crossing the dorsal hippocampus. Cells always migrated along the direction of radial fibers, thus primarily confined in a plane around the injection site (Suzuki and Goldman, in press).

Neuronal Progenitors Migrate Throughout the Rostrocaudal Extension of the SVZ

When viewed in sagittal planes, a continuous rostrocaudal stream of labeled cells is observed within the SVZ throughout the rostrocaudal extension of the SVZ, including the RMS (Fig. 2, P10; Suzuki and Goldman, in press). These cells are known to be neuronal progenitors that give rise to olfactory interneurons (Luskin, 1993). Tuj1 immunopositivity was seen all over the SVZ, from the anterior region to the caudal tip at the hippocampal border (our unpublished observations). Time-lapse video imaging showed that these cells migrate bidirectionally, parallel to the ventricular surface. The occasional cell that strays from the tangential path and moves obliquely to encounter the borders between the SVZ and the white matter or the striatum, invariably turns back to stay within the SVZ. Thus, we never encountered cells migrating rostrocaudally that then turned to emigrate radially from the SVZ into adjacent structures. The migration of this population appears to be highly restricted within the SVZ until progenitors reach the olfactory bulb, where they turn direction and migrate radially into the olfactory cortex to differentiate into interneurons. The finding is consistent with the previous observations of an extensive network of pathways for the tangential chain migration of neuronal precursors throughout the lateral wall of the lateral ventricle in the adult mouse brain (Lois and Alvarez-Buylla, 1994; Doetsch and Alvarez-Buylla, 1996) (Fig. 2, Adult). Taken all together, we posit that the early postnatal SVZ is composed of at least two populations, glial and neuronal progenitors, each of which takes a distinct migratory pathway (Fig. 2, P10).


What would determine such distinct populations and migratory behaviors among SVZ cells? Although there are a growing number of secreted and membrane-bound molecules that play roles in inducing or repressing neuronal, astrocytic or oligodendrocytic fates, rather little is know about which of these molecules is expressed in the SVZ or whether they determine cell fates within the SVZ. Here we discuss some of the relevant factors known to be localized within and around the SVZ, realizing that a detailed understanding of how these molecules are localized and how they interact with SVZ progenitors is currently unavailable.

Sonic hedgehog (Shh) and bone morphogenic proteins (BMPs), which form opposing gradients of diffusible signaling molecules along the dorsoventral axis (Kessaris et al., 2001), respectively promote oligocytogenesis during embryogenesis at the ventral telencephalon in vivo (Tekki-Kessaris et al., 2001) and astrocytogenesis (Nakashima et al., 2001; Mabie et al., 1997; Mehler et al., 2000). These molecules, in addition to the BMP antagonist, Noggin, are expressed within the SVZ. For example, BMP4 is expressed in SVZ cells from the embryonic period through adulthood, along with type I and II BMP receptors, suggestive of an autocrine loop (Gross et al., 1996). In the adult telencephalon, BMPs are expressed by SVZ cells, while the antagonist Noggin is expressed by ependymal cells adjacent to the SVZ (Lim et al., 2000). Shh has also been localized within the SVZ in the developing brain (Murray et al., 2002).

Notch receptor family proteins have been reported to promote astrocytogenesis (Tanigaki et al., 2001), while inhibiting oligodendrocytogenesis (Wang et al., 1998) and neurogenesis (Morrison et al., 2000), through binding their ligands Jagged and Delta and activating signal pathway toward their downstream effectors, Hes family basic helix-loop-helix genes (Furukawa et al., 2000; Hojo et al., 2000). The role of Notch in promoting glial fates was reviewed elsewhere in detail (Gaiano and Fishell, 2002). Expression of Notch1 and its ligands, as well as Hes5, was shown in the SVZ including the RMS in the developing and the adult brains (Stump et al, 2002). Notch expression in VZ cells may be important in promoting astrocyte development and inhibiting neuronal development in radial glia (Gaiano et al., 2000).

A number of soluble growth factors have profound effects on gliogenesis in vitro and in vivo, the latter demonstrable in transgenic mouse models. For example, platelet-derived growth factor (PDGF) is a potent mitogen for oligodendrocyte progenitors (OLPs), and mice overexpressing PDGF demonstrated increased OLPs and ectopic distribution of mature oligodendrocytes (Calver et al., 1998). In contrast, mice lacking PDGF-A show a decrease in numbers of PDGFR-α+ OLPs in the embryonic period and a dysmyelination phenotype (tremor) due to a reduction in myelinating oligodendrocytes after birth (Fruttiger et al., 1999). The sources of PDGF appear to be both neurons and astrocytes (Silberstein et al., 1996), although there is controversy about astrocyte expression (Ellison et al., 1996). However, few SVZ cells express PDGFR-α until they have migrated out of the SVZ (Pringle et al., 1992), suggesting that PDGF signaling may not play a role in fate determination within the SVZ itself, but may be critical in expanding OLP pool size in white and gray matter.

Insulin-like growth factor-1 (IGF-1), which promotes the survival and differentiation of OLPs, is expressed at the border zone of the early postnatal SVZ (Bartlett et al., 1992), although the specific cellular localization is not clear. IGF-1-null mice show reduced OPC proliferation and development, resulting in decreased myelinating oligodendrocytes during early postnatal development. Mice that overexpress IGF-I in the brain exhibit postnatal brain overgrowth without anatomic abnormality, due to increased neuron and oligodendrocyte numbers. Whether these phenotypes are due to altered IGF signaling in SVZ progenitors is unknown, however.

Epidermal growth factor receptor (EGFR) and its related ErbBs are abundantly expressed in the SVZ (Seroogy et al., 1995; Eagleson et al., 1996; Misumi and Kawana, 1998; Kornblum et al., 2000). SVZ cells are sensitive to EGF as demonstrated by the dramatic increase in numbers of SVZ cells produced by injecting EGF into the lateral ventricles (Vaccarino et al., 1999). Thus, the proliferation and possibly the survival of SVZ cells is mediated in part by EGF. Basic fibroblast growth factor (bFGF), another mitogen for OLPs and for embryonic neural progenitors, will increase numbers of cortical glia after intraventricular injections at E20.5 (Vaccarino et al., 1999). Whether and how SVZ progenitors are influenced by bFGF remain unknown.


The SVZ continues to produce olfactory interneurons in adulthood (Altman, 1969; Lois and Alvarez-Buylla, 1994). The adult SVZ consists of an extensive network of pathways for the tangential chain migration of neuronal precursors throughout the lateral wall of the lateral ventricle (Lois and Alvarez-Buylla, 1994; Doetsch and Alvarez-Buylla, 1996). This is in line with our observations on the long-distance rostrocaudal migration of neuronal progenitors in the early postnatal SVZ. Doetsch et al. (1997) described the cellular composition and organization of the SVZ in the adult mouse brain based on electron microscopy. They classified constituent cells into migrating neuroblasts (type A), astrocytes (type B), cycling precursors (type C), tanycytes (type D), and ependymal cells (type E). The type B cells had properties of neural stem cells (Doetsch et al., 1999) that give rise to olfactory interneurons in vivo and multipotent neurospheres in vitro.

Johansson et al. (1999) provided evidence that ependymal cells give rise to olfactory interneurons and spinal astrocytes. Because some reports appear to rule out the possibility that ependymal cells are stem cells (Chiasson et al., 1999; Laywell et al., 2000; Capela and Temple, 2002), and another study suggested both astrocytes and ependymal cells could serve as stem cells (Rietze et al., 2001), the subject remains controversial.

Adult SVZ cells also give rise to glia in vitro (Lois and Alvarez-Buylla, 1994; Doetsch et al., 1999). Migration of glial progenitors from the SVZ to the neocortex ceases by P14 (Levison and Goldman, 1993). While gliogenesis in the adult white matter and cortex has been demonstrated both in vivo and in vitro (Gensert and Goldman, 1996); Levison and Goldman, 1997; Gensert and Goldman, 2001), it is unclear whether adult SVZ cells give rise to glia in vivo.


Why might the SVZ cease to produce glial cells during adulthood? BMPs, as well as their cognate receptors, are expressed within the perinatal SVZ and are sufficient to specify perinatal SVZ cells to become astrocytes at the expense of neurons and oligodendrocytes (Gross et al., 1996). Starting around P7 in the mouse, the neuroepithelium matures into ciliated ependymal cells and subventricular astrocytes (Bruni, 1998; C. Marshall, unpublished observations). Ependymal cells secrete the bone morphogenetic protein (BMP) antagonist Noggin, which inhibits the gliogenic activity of BMPs and creates a neurogenic niche within the SVZ along the walls of the lateral ventricles (Lim et al., 2000). As there are fewer mature ependymal cells present during the peak of gliogenesis (P10) than during adulthood, it is reasonable to suggest that less Noggin is present within the perinatal SVZ than in the adult (Fig. 3). Furthermore, because the perinatal SVZ is much larger in area than the adult SVZ, Noggin expression influences a smaller proportion of SVZ cells, allowing more to respond to BMPs and adopt glial phenotypes. Interestingly, expression of BMP-4, which is strongly inhibited by Noggin, is upregulated perinatally within the SVZ and maintained throughout postnatal development into adulthood (Gross et al., 1996), where it is expressed primarily by SVZ astroctyes (Lim et al., 2000). The adult SVZ is a vestigial structure due to the limited requirement for newly generated cells in the mature brain (Opalski, 1933; Kershman, 1938), hence the mature ependyma produces enough Noggin to diffuse throughout the entire SVZ (Lim et al., 2000), creating a primarily neurogenic germinal zone under normal conditions.

Figure 3.

Regionalization of the subventricular zone (SVZ). The SVZ can be delineated into two subpopulations of cells based on the expression of Zebrin II (green) or the Dlx2/tauLacZ reporter (blue). Zebrin II+ residual VZ cells form the outer borders of the SVZ. Cells sharing a Dlx2-expressing subpallial origin, as defined by Dlx2/tauLacZ (βgal) expression, populate the central SVZ. These cells give rise to interneurons in the olfactory bulb from the perinatal period throughout adulthood. SVZ progenitors also generate astrocytes and oligodendrocytes, which colonize the striatum, white matter, and cerebral cortex during the peak of gliogenesis (first several postnatal weeks). Noggin (yellow shading) expression by ependymal cells (pink) creates a neurogenic niche along the walls of the lateral ventricles. The postnatal dorsolateral SVZ contains neurogenic and gliogenic niches, while the adult SVZ appears to be entirely neurogenic under normal conditions.

Such a model is supported by the migration patterns of SVZ cells discussed earlier as well as by molecular expression patterns of SVZ cells. Glial-specific molecular markers such as Olig-2, vimentin, PDGFR-α and NG2 are expressed by cells residing within the lateral regions of the perinatal SVZ (as well as by cells that have begun to migrate from the SVZ) but not by cells near the developing ependymal lining (C. Marshall, unpublished observations). A halo of Olig-2 expression appears at the dorsal and lateral borders of the SVZ perinatally. This expression expands dorsally into the white matter and all layers of the cerebral cortex and laterally into the striatum throughout the first several postnatal weeks. Interestingly, both astrocytes as well as oligodendrocytes express Olig-2 in the white matter, cortex, and striatum; however, the central SVZ and RMS are relatively devoid of Olig-2 expression (C. Marshall, unpublished observations). This recent data, taken together with other work described above, suggests that a glial compartment may exist within the dorsolateral limits of the perinatal SVZ. Recent studies of transcription factors that regulate the specification of neural progenitors (for review, see Morrison, 2001; Bertrand et al., 2002) promise to deepen our understanding of the complex cell architecture within the SVZ. By identifying the genetic pathways that specify SVZ cells, we will be able to examine further the relationship between phenotype and migratory behavior.


The authors thank Gord Fishell and Josh Corbin for the generous gift of Dlx2/tauLacZ mutant tissue, John Rubenstein and Stewart Anderson for the Dlx2 antibody, Carol Mason for graciously sharing imaging equipment, and Akiyoshi Kakita for help with time-lapse video microscopy. This work was supported by National Institutes of Health grant NS-17125 (to J.E.G.).