Introduction

Transforming growth factor beta (TGF-) and its signaling pathways are frequently involved in cell growth, embryogenesis, differentiation, morphogenesis, extracellular matrix formation, wound healing, immune response, and apoptosis in a wide variety of cells including epithelial cells (Roberts and Sporn, 1993; Gold, 1999; Bottner et al., 2000). TGF- belongs to a superfamily of proteins that include activins, inhibins, bone morphogenetic proteins (BMP), veg-1, the Drosophila decapentaplegic complex, and Mullerian-inhibiting substance (Massague, 1998). The receptors for TGF- (TRs) are expressed in almost every mammalian cell type, including cancer cells (Massague et al., 1992; Massague and Weis-Garcia, 1996). TRI and -II are serine–threonine protein kinases that form complexes and phosphorylate the Smad transcription factors (Derynck, 1994; Hayashi et al., 1997). The only known function for the type II receptor is to activate the type I receptor (Wrana, 2000). In response to TGF-, two TRI receptor-associated Smads, Smad2 and Smad3, become phosphorylated and activated by TRI kinase. After activation, Smads form a heterodimeric complex with the common mediator Smad, Smad4, and translocate to the nucleus to participate in the transactivation of specific target genes (Zhang et al., 1997). Several adaptor proteins belonging to the cytoskeletal proteins can modulate signaling pathways (Nelson and Veshnock, 1986; De Matteis and Morrow, 2000; Bennett and Baines, 2001). Our laboratory has demonstrated that TGF- triggers phosphorylation and association of elf (embryonic liver fodrin), a -spectrin gene, with Smad3 and Smad4 resulting in nuclear translocation (Tang et al., 2003). ELF deficiency in mice results in mislocalization of Smad3 and Smad4 truncating TGF- signaling and transcriptional response. Abnormal development of primary brain vesicles occurs along with liver, heart and gastrointenstinal defects in early embryos of elf mice.

Stem cells are self-renewing progenitor cells with the capacity to generate multiple differentiated derivatives (Morrison et al., 1997). In developing embryos, tissues and organs are derived from stem cells de novo. However, in adults the stem cells undergo continuous cellular turnover providing regenerative capacities in certain tissues. Different categories of stem cells with different self-renewal and developmental capacities have been described in different tissues and at different stages of development (Weissman et al., 2001). Some of the hallmarks of the stem cells are (1) inhibition of overt differentiation, (2) maintenance of proliferative capacity, (3) maintenance of pluripotency, and (4) asymmetrical cell division. However, inhibition of differentiation does not necessarily maintain pluripotency; for example, in the central nervous system, Notch signaling pathway inhibits neuronal differentiation but stem cells do not remain multipotent (Ohtsuka et al., 1999). In practice, one of the ways of defining stem cells is the differentiation- or lineage potential. In doing so, it is also important to reference its temporal and spatial context which could define the growth factor/hormonal inductive influences. Thus, dopaminergic and serotonergic precursor cells in later stages of raphe development, as well as muscle precursor cells bearing MyoD in paraxial mesodermal somites at an earlier stage, are dependent on TGF- for survival and differentiation into the next stage of development. In each type of precursor cells, depending upon the stage of development, TGF- exhibits variable biological effects. This is exemplified by the concept of the 'stem cell niche' that provides the appropriate inductive microenvironment. Skin-derived precursors, but not the neuronal precursor cells examined in one study, proliferate as neurospheres under the influence of TGF- in combination with epidermal growth factor (EGF) and bFGF (Kawase et al., 2004). Jun protein, which is a member of the heterodimer AP1 transcriptional factor, interacts with BMP4 in antineuralizing the effect in xenopus neurogenesis and suppresses neural maintenance after stage 13 only if it is expressed from stage 11 onwards (Peng et al., 2002). During cerebral cortical development, radially migrating ventricular zone cells give rise to glutamatergic cortical cells influenced by BMP, while tangentially migrating ventral forebrain stem cells respond to SHH-mediated activation of transcriptional factor Olig2 and Mash1, and develop into GABAergic inhibitory neurons and oligodendrocytes (Yung et al., 2002). Neural crest stem cells (NCSC) under the influence of NGNs and MASH 1 produce only autonomic cells. In contrast, peripheral nerve stem cells produce sensory neurons under the influence of NGNs and autonomic neurons under the influence of MASH1 (Lo et al., 2002), a further example of stem cells obeying local rules.

Stem cell phenotype concludes with asymmetric division and postmitotic differentiation of one or both daughter cells. In some situations, this occurs a few hours after embryogenesis as in postmitotic neuronal specification of primary motor spinal neurons 12 h after the start of embryogenesis in the zebra fish. Nucleostemin, a nucleolar protein, increases in concentration before stem cell division. Its GTP-binding domain appears necessary for cells to enter mitosis. Mutation of this domain not only prevents stem cell mitosis but it also leads to p53-dependent apoptosis of stem cells (Tsai and McKay, 2002).

Patterning and growth may be linked by BMPs, which are members of the TGF- family, to generate dorsal neural fate in the developing neural tube: WNT signals appear to be important for proliferation of neuronal progenitors following fate specification by BMP (Chesnutt et al., 2004).

TGF- and Neural Stem Cells (NSCs)

Studies during the past decade have established that members of TGF- superfamily play a critical role in morphogenesis and cell lineage specification during brain development (Schier and Talbot, 2001; Munoz-Sanjuan and Brivanlou, 2002).

BMP2 subfamily, including BMP2 and BMP4, is involved in gastrulation, and neurogenesis, and BMP5 participates with BMP4 in neurogenesis.

The initial spotlight on the role of TGF- superfamily demonstrated that Activin could induce mesoderm in animal caps. The default neural fate could be induced in animal cap tissue obtained from the blastula stage of embryos by transfecting a mutant activin type II receptor that binds activin and dimerizes but fails to phosphorylate the type I receptor, thereby interfering with Activin signaling.

In serum-free cultures of rat E14 embryonic raphe, TGF-2 increased the number and relative proportion of tryptophan hydroxylase immunoreactive cells, indicating a role in serotonergic cell development (Galter et al., 1999). Murine neural crest (Sieber-Blum and Zhang, 1997) show reduced proliferation and increased proportion of SSEA-expressing sensory neurons and pigmented DBH-expressing cells in the presence of TGF-. NCSC migrate to several sites, but only under the influence of TGF- are they able to differentiate into non-neural tissue (Wurdak et al., 2005). Calcineurin signaling may be an intermediate step induced by TGF- which leads to adoption of smooth muscle phenotype in NCSC (Mann et al., 2004). Starting with Monc 1, a smooth muscle precursor cell line, it has been possible to generate a smooth muscle phenotype including contractile phenotype using TGF-. Smads 2 and 3 have been shown to be important in the signaling by TGF- in this model (Chen et al., 2004). SOX10 preserves glial and neuronal potential from extinction by lineage commitment signals acting upon neural crest stem cells, requiring induction of MASH and PHOX2b, two neurogenic transcription factors (Kim et al., 2003). Rat olfactory cell neurogenesis appears to be critically dependent upon FGF2 and TGF- sequentially (Kawauchi et al., 2004). Blocking Smad4 might not be sufficient to block mesenchymal fate in embryonic stem (ES) cells, but does promote neuronal fate (Sonntag et al., 2005). Spinal cord neurogenesis is promoted by inhibition of BMP signaling through expression of Noggin (Setoguchi et al., 2004). Extraembryonic endodermal fate of stem cells is inhibited and neuronal precursor fate is induced in other experimental systems by similar inhibition of signaling by BMP2 using noggin expression (Pera et al., 2004). Two reports suggest that specific precursors could turn towards astrocytic lineage under the influence of BMP or TGF-: cerebellar granule cell precursors that express GFAP and S100 under the influence of BMP2 (Okano-Uchida et al., 2004), and astrocytic precursors that are induced to express cystatin C by TGF- (Kumada et al., 2004). Branchial arch ectomesenchyme-derived stem cells are maintained in an undifferentiated state with Leukemia Inhibitory Factor and induced to differentiate along neuronal lineage by TGF-, while the other lineage induction requires additional factors (Deng et al., 2004). Neural tube pattern formation is dependent upon stem cell differentiation and establishment of discrete WNT responsive domains by BMP and TGF- (Chesnutt et al., 2004).

Several other inducers of neural fate have been suspected for over 10 years. These include FGF, which suppresses cranial markers in the xenopus embryo and induces caudal markers. Others include Notch and other bHLH factors whose relationship to TGF- in neural cell fate determination remains unknown.

The role of TGF- in mammalian neuronal development and/or survival is demonstrated by the fact that a large increase in the cerebellar Purkinje cell number occurs in mouse Smad4 mutants. Furthermore, the mouse TGF- mutants show highly increased neuronal apoptosis, and a TGF- family member (BMP9) controls cholinergic differentiation of mouse spinal cord and septal neurons (Lopez-Coviella et al., 2002; Brionne et al., 2003; Zhou et al., 2003).

A common theme may be imagined from the combined data of several laboratories summarized above: TGF- mediation stimulates lineage commitment in precursor cells (LCSC). The corollary that also appears to be borne out by experiments is that lineage-uncommitted stem-cell/precursor phenotype (LUSC) persists in the absence of TGF- stimulation, provided other growth factors such as bFGF are able to sustain proliferation and overcome apoptosis. The phenotype of elf^-/- mice in whom TGF- signaling is disrupted does show that proliferation of precursor cells occurs without acquiring lineage-specific phenotypic markers such as -tubulin and nestin. This observation is consistent with the schematic of TGF- involvement in neural precursor cell biology.

However, ES cell-derived neurogenesis does not entirely depend on functional Smad4, indicating that other factors, such as cell–cell contact, the cellular environment, and other signaling molecules, are involved in the process (Sonntag et al., 2005). The existence of multipotential, self-renewing NSCs in adult mammalian brain re-emerged in the past decade (Gage, 2000; Anderson, 2001; Temple, 2001; Gage, 2002; Tsai et al., 2002), confirming an earlier report on neurogenesis in the adult brain (Gross, 2000). The concept of the cancer stem cells appeared from the observation of remarkable similarities between the self-renewal patterns of stem cells and cancer cells (Reya et al., 2001; Pardal et al., 2003). Normal somatic stem cells self-renew and differentiate with a strict balance between the two directives, and cancer originates as a result of unregulated self-renewal (Reya et al., 2001). The longevity of the NSCs may allow for accumulation of genetic mutations rendering the self-renewal mechanism itself to be a primer to creating a cancer cell. Therefore, NSCs and their proliferating progenitors warrant further investigation during brain tumorigenesis. Mutations that dysregulate the pathways that control normal stem cell self-renewal (for e.g., sonic hedgehog pathway) cause brain tumors, underscoring the mechanistic similarities between normal stem cell and cancer cell self-renewal (Dahmane et al., 2001; Ruiz i Altaba et al., 2002; Pardal et al., 2003). Also, pediatric brain tumors are reported to contain stem-like cells that have the ability for sphere formation, self-renewal, and multipotential differentiation (Hemmati et al., 2003).

Recent reports elucidate a role of stem cells in brain tumor development. Glioblastoma (GB) is the most common intrinsic malignant brain tumor with a high aggressive and a strong infiltration growth. Excessive proliferation, diffuse infiltration of surrounding brain tissue, and suppression of antitumor immune surveillance contribute to the malignant phenotype of gliomas (Weller and Fontana, 1995). Silencing of TGF- expression by small interfering RNA (siRNA) technology has been reported to abrogate glioma cell tumorigenicity in vivo (Friese et al., 2004).

TGF- as the Tumor Suppressor, and Glioblastoma

TGF- has a role in arresting cell growth through two mechanisms (Gold, 1999): (1) it inhibits the expression of cyclin-dependent kinases (cdks) (Johnson et al., 1993) and (2) it downregulates cdk activity through induction of cdk inhibitors, p15, p27, and Cip/WAF1/p21 (Moses and Serra, 1996; Heldin et al., 1997; Attisano and Wrana, 1998; Sporn, 1999; Massague et al., 2000; Hu and Zuckerman, 2001; Matsuura et al., 2004). This downregulation results from Smad-mediated transcriptional activation that is pivotal for cell cycle arrest and inhibition of cell proliferation by TGF- (Figure 1). A proteasome inhibitor, PS-341, causes cell growth arrest and apoptosis in human GB that are accompanied by increased levels of p21^WAF1, p27^KIP1, cyclin B, and decreased expression of CDK2, CDK4, and E2F4 in a p53-independent fashion (Yin et al., 2005). One common feature of human GB is a hyperactive NF- pathway (Hayashi et al., 2001; Nagai et al., 2002). PS-341 treatment results in a decrease in NF- nuclear activity and activation of JNK (Yin et al., 2005) that could lead to apoptosis (Wang et al., 1998; Hideshima et al., 2003; Yang et al., 2004).

Figure 1.

A proposed schematic representation depicting potential interplay of TGF- signaling pathway in neural stem cells during normal development

Full figure and legend (138K)

Disruption of TGF- signaling occurs in all pancreatic cancers (Goggins et al., 1998; Villanueva et al., 1998) and colon cancers (Grady et al., 1999). Increased hepatocyte proliferation, decreased apoptosis in the lung and liver (Tang et al., 1998), and accelerated mammary epithelial proliferation and ductal outgrowth in response to hormones (Barcellos-Hoff and Ewan, 2000) are characteristics of the TGF- I^-/- mice.

A 10 bp polyadenine repeat within the TRII sequence encoding a part of the extracellular domain (referred to as the BAT-TRII track) results in a frameshift and a truncated, inactive TRII product (Markowitz et al., 1995). BAT-TRII mutations are found in subsets of colon cancers, gastric cancers, and gliomas with microsatellite instability (Markowitz et al., 1995; Myeroff et al., 1995; Izumoto et al., 1997). However, Smad mutation has not been reported in GB.

TGF- as the Tumor Promoter, and Glioblastoma

Although known for its growth-inhibitory function in epithelial tissues, TGF- is both a suppressor and a promoter of tumorigenesis (Cui et al., 1996; Akhurst and Derynck, 2001; Derynck and Zhang, 2003). Its tumor-suppressive function is lost in many tumor-derived cell lines (Reiss, 1997, 1999). The malignant phenotype of transformed and tumor-derived cells in culture is activated further by TGF-. Furthermore, advanced tumors show high levels of TGF- expression (Gold, 1999). The secondary role of TGF- other than its tumor-suppressive activity may, possibly, be due to its effect in enhancing cell migration, thereby promoting cellular invasion (Cui et al., 1996; Oft et al., 1996; Yin et al., 1999). The ability of TGF- to induce epithelial to mesenchymal transition that is simultaneous with a downregulation of cell adhesion proteins may also contribute to cell migration and metastasis (Oft et al., 1996) (Figure 2).

Figure 2.

A proposed pathway representing a potential role of TGF- in glioblastoma development

Full figure and legend (177K)

Deregulation of growth factor and growth factor receptor expression is one of the characteristic features of the GBs. Overexpression of constitutively active mutated receptors of the epidermal growth factor (EGF) (Humphrey et al., 1991; Kuan et al., 2001), overexpression of functionally active receptors for the neuropeptide somatostatin (Held-Feindt et al., 2000), production of high amounts of the vascular endothelial growth factor (VEGF) (Mentlein et al., 2001), and enhanced expression of the TGF-2 isoform (Bodmer et al., 1989) and of TRI and TRII (Yamada et al., 1995) occur in GBs (Figures 2, 3). Resistance to growth inhibition can be detected in malignant glioma cells with functionally active TGF- receptors (Isoe et al., 1998). In 5/6 glioma cell lines, TGF- decreases the expression of p27 (Isoe et al., 1998). There is evidence that TGF- alters collagen synthesis, integrin expression, cell adhesion to reconstituted basement membrane, and invasiveness among cultured gliomas (Merzak et al., 1994; Paulus et al., 1995). Its effect on tumor progression by modulating the extracellular matrix may also be synergized by induction of the proto-oncogene c-sis and/or the PDGF receptor subunit in the gliomas (Plate et al., 1992). In contrast, a recent report shows that EGFR but not PDGFR correlates with cell proliferation in GB (Halatsch et al., 2003). Thus, the tumor promoting activity of TGF- seems to be a fundamental target in human glioma cell malignancy.

Figure 3.

A schematic representation of TGF, elf, and other major signaling cascades in neural stem cells during normal and neuroblastoma development

Full figure and legend (31K)

Role of TGF- as a Negative Regulatory Cytokine in Glioblastoma

An important physiologic role of TGF- within the brain appears to be termination of the glial proliferative response to injury and/or cytokine-induced stimulation (Lindholm et al., 1992; da Cunha et al., 1993). Among quiescent newborn rat cerebral glia, serum induction of DNA synthesis is delayed and attenuated by TGF-1. TGF-1 and also TRI and -II are expressed only in malignant gliomas rather than in normal brain, gliosis, or low-grade astrocytomas (Samuels et al., 1989; Horst et al., 1992; Yamada et al., 1995). The expression level of TGF- is indirectly correlated with survival among malignant glioma patients. Similar opposing effects, growth inhibitory and promotive are also displayed by the oncogenic Ras that induce differentiation in parallel with the cessation of cell division and DNA synthesis in rat pheochromocytoma and human medullary thyroid carcinoma cells (Bar-Sagi and Feramisco, 1985; Nakagawa et al., 1987).

Conversion of TGF- from a Growth Inhibitor to a Mitogen in Hyperdiploid Glioblastoma (HD-GB)

Endogenously produced and activated TGF- functions as a growth inhibitor of near-diploid gliomas, which is reversed by the addition of anti-TGF--neutralizing antibodies. However, HD-GB cultures are stimulated by exogenous TGF- in monolayer culture and display enhanced anchorage-independent clonogenicity. This mitogenic response to TGF- is associated with marked anaplasia and karyotypic divergence, although no correlation exists between TR loss and subtype expression (Jennings et al., 1991). In medulloblastoma, primitive neuroectodermal tumor and ependymoma cultures hyperdiploidy are associated with mitogenic stimulation by TGF- (Jennings et al., 1994). In the near-diploid GB controls, TGF- induces only the RNA of PDGF-A and TGF-, whereas the proto-oncogene c-sis is not expressed at all. The amount of PDGF-A secreted in response to TGF-1 is insufficient to prevent the arrest of the near-diploid cultures in G1 phase. By comparison, TGF-1 induces the RNA of PDGF-A and -B as well as TGF- autoinduction in the HD-GB. The concentration of PDGF-AA secreted following TGF- treatment is sufficient to stimulate the proliferation of the HD-GB in vitro. Combinations of antibodies against PDGF-AA, -BB, -AB, PDGFR and/or PDGFR subunits effectively neutralize TGF-'s induction of DNA synthesis among the HD-GM cell lines, indicating that PDGF serves as the principal, if not only, mediator of TGF-'s growth stimulatory effect (Jennings et al., 1997).

TGF- and ELF (a -spectrin)

Spectrin is a tetrameric actin crosslinking protein and is comprised of two and two subunits. Two genes for -spectrin (Riederer et al., 1986; Goodman et al., 1995; Ziemnicka-Kotula et al., 1998) and five for -spectrin have been identified to date in both mice and humans, each of which is alternatively spliced to produce multiple spectrin isoforms (Bennett et al., 1982; Birkenmeier et al., 1988; Gallagher and Forget, 1993; Goodman et al., 1995). A novel G-spectrin lacks the C-terminal pleckstrin homology (PH) domain that is characterized by elf 1, 2, 3, and 4 (Mishra et al., 1998, 1999; Hayes et al., 2000).

Spectrin levels are elevated after gastrulation suggesting a role in cell fate determination (Moon and McMahon, 1987). ELF has been found to be a neuronal precursor cell marker in the developing mammalian brain (Tang et al., 2002). ELF is expressed in cell bodies as well as dendrites and initial axon segments, and in this sense is unique in its expression pattern from spectrin G and brain spectrin (Goodman et al., 1995). Spectrin G is present in axons and, to a lesser extent, in neuronal cell bodies but not in dendrites (Riederer et al., 1986), while ELF is expressed in both cell types. Brain spectrin (R), an alternatively spliced form of erythrocyte spectrin, is localized to the perikaryon, dendrites (like ELF/G), and also glial precursor cells. The expression pattern of ELF is more similar to R and the brain isoform of G, which is in keeping with its cDNA structure (Mishra et al., 1999).

Disruption of the adaptor protein ELF leads to the disruption of TGF- signaling by Smad proteins in mice. Elf^-/- mice exhibit a phenotype similar to Smad2^+/-/Smad3^+/- mutant mice of midgestational death due to gastrointestinal, liver, neural, and heart defects. TGF- triggers phosphorylation and association of ELF, possibly by inducing a conformational change, with Smad3 and Smad4, followed by nuclear translocation. In contrast, ELF associates with two known spectrin-binding structural proteins, ankyrin B and tropomyosin, only in the absence of TGF-1. Elf deficiency results in mislocalization of Smad3 and Smad4 and loss of the TGF--dependent transcriptional response, which can be rescued by overexpression of the C-terminal region of ELF. These results establish a novel link between a major dynamic scaffolding protein and the TGF- signaling pathway.

Recent observations from this lab suggest that ELF3, a novel -spectrin, is important for brain development from an early stage (Tang et al., 2002). Its striking localization to the axon hillock and axonal projection is consistent with its known involvement in axonal guidance as well as cell polarization (Dubreuil et al., 2000; Hammarlund et al., 2000; Moorthy et al., 2000). ELF also appears to be a marker of Purkinje cell precursors in the cerebellum and could play a direct role in these processes of brain development through its involvement in cytoskeletal architecture, based on the appearance of the labeled cells. These observations are supported by the fact that spectrin comprises approximately 2.4% of total structural protein in mammalian brain homogenates (Davis and Bennett, 1983) and is involved in the organization of presynaptic vesicles (Sikorski and Goodman, 1991; Sikorski et al., 2000), as well as in axonal transport (Levine and Willard, 1981).

Subcellular colocalization of ELF with nestin suggests a potential role of ELF in neuronal stem cell biology, since nestin is a neurofilament protein expressed in neuronal precursor cells, as well as in radial glial cells (Dahlstrand et al., 1995), in neuritis, and growth cones of primary cerebellar granule cells signifying its role in growth cone guidance during axon elongation. Microinjections of N-terminal domain-specific anti--spectrum antibody inhibits neuritis extension in neuroblastoma cells (Sihag et al., 1996).

Summary

In a normal cell, multiple cell signaling pathways work in concert. The cell functions normally until either genetic mutations or change in the cell microenvironment results in a shift in the balance leading one or more signaling pathways to accept alternative courses other than the defined ones. For example, TGF- signaling pathways can be Smad dependent or independent (Akhurst and Derynck, 2001; Derynck and Zhang, 2003). Instead of traditional Smad pathways that terminate in a growth inhibitory effect, TGF- can deviate upstream of the Smad phosphoryaltion to activate several other signaling pathways, such as MAPK (Ras-Erk) and SAPK (Rho-JNK, TAK1-p38 kinase), activating a subset of an altogether different target genes that could lead to proliferation or transformation. Another such example is observed in EGFR function in response to EGF. Proliferation of the A431 human squamous carcinoma cell line that expresses high numbers of the EGF receptor (EGFR) is actually inhibited by EGF treatment as a consequence of activation of the EGFR's tyrosine kinase induction of the Cdk inhibitor, p21^WAF/CIP1 instead of the defined EGF-EGFR-Ras-Erk activation (Fan et al., 1995). In the NSCs, TGF- could use the conventional Smad pathway, where ELF remains a major participant (Tang et al., 2003) that would translate into growth inhibition. On the other hand, if already primed for transformation due to a changed microenvironment or mutation, it might preferentially use the TGF--MAPK or TGF-SAPK pathway to enhance proliferation (Derynck and Zhang, 2003). In the latter case, whether ELF would play any role remains to be elucidated. Defined signaling pathways remain accessible to stem cells throughout development (Tsai and McKay, 2000; Panchision et al., 2001; Rajan et al., 2003), challenging these cells to make choices regarding fate and self-renewal (Hitoshi et al., 2002; Panchision and McKay, 2002). The direction of tumor development may be determined by the age of the NSCs, among other factors. In older cell populations, possibly a genetic change would prime it toward tumorigenesis and a lateral, divergent TGF--MAPK pathway instead of TGF--Smad pathway might acquire preference. However, in younger stem cell or progenitor cell populations, the reverse could be the case. This scenario might explain the dual effect, an earlier tumor-suppressive and a later tumor-enhancing property of TGF-; the former is a case of normal homeostasis characterized by the TGF--Smad linear signaling cascade, while the latter comprises a nonlinear crosssignaling to synergize with the stronger proliferative pathways (e.g., Ras-Erk pathway) in abnormal cells like in the GBs.

In the stem cells, Smad4 is dispensable for neuronal specification (Sonntag et al., 2005) implicating that there are other mechanisms that compensate for some of TGF-'s signaling impact. As for ELF, in younger stem cell populations, its contribution in propagating TGF- signals would allow the cells to divide and differentiate at a normal rate, while its deficiency may be conducive for cell proliferation and, therefore, maintaining the stem cell-like properties. Whether the resulting NSC proliferation would lead to tumor-like GB may depend on multiple factors, especially, the cellular microenvironment and the age of the NSC. Based on the late onset of a synergistic role of TGF- in tumorigenesis, although elf-deficient mice (in case of a conditional knock out for postnatal elf deletion or surviving heterozygotes) although would be expected to harbor a higher number of NSCs during early development, a predisposition toward brain tumor occurrence later in life seems plausible (Figure 3). Whether ELF's participation in TGF- signaling alone and/or its possible involvement in other signaling processes are key to the maintenance of the NSCs demands further investigation.

References

1. Akhurst RJ & Derynck R. (2001) Trends Cell Biol. 11: S44−S51. | Article | PubMed | ISI | ChemPort |
2. Anderson DJ. (2001) Neuron 30: 19−35. | Article | PubMed | ISI | ChemPort |
3. Attisano L & Wrana JL. (1998) Curr. Opin. Cell Biol. 10: 188−194. | Article | PubMed | ISI | ChemPort |
4. Bar-Sagi D & Feramisco JR. (1985) Cell 42: 841−848. | Article | PubMed | ChemPort |
5. Barcellos-Hoff MH & Ewan KB. (2000) Breast Cancer Res. 2: 92−99. | Article | PubMed | ChemPort |
6. Bennett V & Baines AJ. (2001) Physiol. Rev. 81: 1353−1392. | PubMed | ISI | ChemPort |
7. Bennett V, Davis J & Fowler WE. (1982) Nature 299: 126−131. | Article | PubMed | ISI | ChemPort |
8. Birkenmeier CS, McFarland-Starr EC & Barker JE. (1988) Proc. Natl. Acad. Sci. USA 85: 8121−8125. | PubMed | ChemPort |
9. Bodmer S, Strommer K, Frei K, Siepl C, de Tribolet N, Heid I & Fontana A. (1989) J. Immunol. 143: 3222−3229. | PubMed | ISI | ChemPort |
10. Bottner M, Krieglstein K & Unsicker K. (2000) J. Neurochem. 75: 2227−2240. | Article | PubMed | ISI | ChemPort |
11. Brionne TC, Tesseur I, Masliah E & Wyss-Coray T. (2003) Neuron 40: 1133−1145. | Article | PubMed | ChemPort |
12. Chen D, Zhao M & Mundy GR. (2004) Growth Factors 22: 233−241. | Article | PubMed | ChemPort |
13. Chesnutt C, Burrus LW, Brown AM & Niswander L. (2004) Dev. Biol. 274: 334−347. | Article | PubMed | ChemPort |
14. Cui W, Fowlis DJ, Bryson S, Duffie E, Ireland H, Balmain A & Akhurst RJ. (1996) Cell 86: 531−542. | Article | PubMed | ISI | ChemPort |
15. da Cunha A, Jefferson JA, Jackson RW & Vitkovic L. (1993) J. Neuroimmunol. 42: 71−85. | Article | PubMed | ChemPort |
16. Dahlstrand J, Lardelli M & Lendahl U. (1995) Brain Res. Dev. Brain Res. 84: 109−129. | Article | PubMed | ChemPort |
17. Dahmane N, Sanchez P, Gitton Y, Palma V, Sun T, Beyna M, Weiner H & Ruiz i Altaba A. (2001) Development 128: 5201−5212. | PubMed | ISI | ChemPort |
18. Davis J & Bennett V. (1983) J. Biol. Chem. 258: 7757−7766. | PubMed | ISI | ChemPort |
19. De Matteis MA & Morrow JS. (2000) J. Cell Sci. 113 (Part 13): 2331−2343. | PubMed | ISI | ChemPort |
20. Deng MJ, Jin Y, Shi JN, Lu HB, Liu Y, He DW, Nie X & Smith AJ. (2004) Tissue Eng. 10: 1597−1606. | PubMed | ChemPort |
21. Derynck R. (1994) Trends Biochem. Sci. 19: 548−553. | Article | PubMed | ISI | ChemPort |
22. Derynck R & Zhang YE. (2003) Nature 425: 577−584. | Article | PubMed | ISI | ChemPort |
23. Dubreuil RR, Wang P, Dahl S, Lee J & Goldstein LS. (2000) J. Cell Biol. 149: 647−656. | Article | PubMed | ISI | ChemPort |
24. Fan Z, Lu Y, Wu X, DeBlasio A, Koff A & Mendelsohn J. (1995) J. Cell Biol. 131: 235−242. | Article | PubMed | ISI | ChemPort |
25. Friese MA, Wischhusen J, Wick W, Weiler M, Eisele G, Steinle A & Weller M. (2004) Cancer Res. 64: 7596−7603. | PubMed | ChemPort |
26. Gage FH. (2000) Science 287: 1433−1438. | Article | PubMed | ISI | ChemPort |
27. Gage FH. (2002) J. Neurosci. 22: 612−613. | PubMed | ISI | ChemPort |
28. Gallagher PG & Forget BG. (1993) Semin. Hematol. 30: 4−20. | PubMed | ChemPort |
29. Galter D, Bottner M & Unsicker K. (1999) J. Neurosci. Res. 56: 531−538. | Article | PubMed | ChemPort |
30. Goggins M, Shekher M, Turnacioglu K, Yeo CJ, Hruban RH & Kern SE. (1998) Cancer Res. 58: 5329−5332. | PubMed | ISI | ChemPort |
31. Gold LI. (1999) Crit. Rev. Oncog. 10: 303−360. | PubMed | ISI | ChemPort |
32. Goodman SR, Zimmer WE, Clark MB, Zagon IS, Barker JE & Bloom ML. (1995) Brain Res. Bull. 36: 593−606. | Article | PubMed | ChemPort |
33. Grady WM, Myeroff LL, Swinler SE, Rajput A, Thiagalingam S, Lutterbaugh JD, Neumann A, Brattain MG, Chang J, Kim SJ, Kinzler KW, Vogelstein B, Willson JK & Markowitz S. (1999) Cancer Res. 59: 320−324. | PubMed | ISI | ChemPort |
34. Gross CG. (2000) Nat. Rev. Neurosci. 1: 67−73. | Article | PubMed | ISI | ChemPort |
35. Halatsch ME, Gehrke E, Borhani FA, Efferth T, Werner C, Nomikos P, Schmidt U & Buchfelder M. (2003) Anticancer Res. 23: 2315−2320. | PubMed | ChemPort |
36. Hammarlund M, Davis WS & Jorgensen EM. (2000) J. Cell Biol. 149: 931−942. | Article | PubMed | ISI | ChemPort |
37. Hayashi H, Abdollah S, Qiu Y, Cai J, Xu YY, Grinnell BW, Richardson MA, Topper JN, Gimbrone MA, Jr, Wrana JL & Falb D. (1997) Cell 89: 1165−1173. | Article | PubMed | ISI | ChemPort |
38. Hayashi S, Yamamoto M, Ueno Y, Ikeda K, Ohshima K, Soma G & Fukushima T. (2001) Neurol. Med. Chir. (Tokyo) 41: 187−195. | Article | PubMed | ChemPort |
39. Hayes NV, Scott C, Heerkens E, Ohanian V, Maggs AM, Pinder JC, Kordeli E & Baines AJ. (2000) J. Cell Sci. 113 (Part 11): 2023−2034. | PubMed | ChemPort |
40. Held-Feindt J, Krisch B, Forstreuter F & Mentlein R. (2000) J. Physiol. Paris 94: 251−258. | PubMed | ChemPort |
41. Heldin CH, Miyazono K & ten Dijke P. (1997) Nature 390: 465−471. | Article | PubMed | ISI | ChemPort |
42. Hemmati HD, Nakano I, Lazareff JA, Masterman-Smith M, Geschwind DH, Bronner-Fraser M & Kornblum HI. (2003) Proc. Natl. Acad. Sci. USA 100: 15178−15183. | Article | PubMed | ChemPort |
43. Hideshima T, Mitsiades C, Akiyama M, Hayashi T, Chauhan D, Richardson P, Schlossman R, Podar K, Munshi NC, Mitsiades N & Anderson KC. (2003) Blood 101: 1530−1534 (Epub 2002 Sep 26). | Article | PubMed | ISI | ChemPort |
44. Hitoshi S, Tropepe V, Ekker M & van der Kooy D. (2002) Development 129: 233−244. | PubMed | ISI | ChemPort |
45. Horst HA, Scheithauer BW, Kelly PJ & Kovach JS. (1992) Hum. Pathol. 23: 1284−1288. | Article | PubMed | ChemPort |
46. Hu X & Zuckerman KS. (2001) J. Hematother. Stem Cell Res. 10: 67−74. | Article | PubMed | ChemPort |
47. Humphrey PA, Gangarosa LM, Wong AJ, Archer GE, Lund-Johansen M, Bjerkvig R, Laerum OD, Friedman HS & Bigner DD. (1991) Biochem. Biophys. Res. Commun. 178: 1413−1420. | Article | PubMed | ChemPort |
48. Isoe S, Naganuma H, Nakano S, Sasaki A, Satoh E, Nagasaka M, Maeda S & Nukui H. (1998) J. Neurosurg. 88: 529−534. | PubMed | ChemPort |
49. Izumoto S, Arita N, Ohnishi T, Hiraga S, Taki T, Tomita N, Ohue M & Hayakawa T. (1997) Cancer Lett. 112: 251−256. | Article | PubMed | ISI | ChemPort |
50. Jennings MT, Hart CE, Commers PA, Whitlock JA, Martincic D, Maciunas RJ, Moots PL & Shehab TM. (1997) J. Neurooncol. 31: 233−254. | Article | PubMed | ChemPort |
51. Jennings MT, Kaariainen IT, Gold L, Maciunas RJ & Commers PA. (1994) Hum. Pathol. 25: 464−475. | Article | PubMed | ChemPort |
52. Jennings MT, Maciunas RJ, Carver R, Bascom CC, Juneau P, Misulis K & Moses HL. (1991) Int. J. Cancer 49: 129−139. | PubMed | ISI | ChemPort |
53. Johnson MD, Jennings MT, Gold LI & Moses HL. (1993) Hum. Pathol. 24: 457−462. | Article | PubMed | ChemPort |
54. Kawase Y, Yanagi Y, Takato T, Fujimoto M & Okochi H. (2004) Exp. Cell Res. 295: 194−203. | Article | PubMed | ChemPort |
55. Kawauchi S, Beites CL, Crocker CE, Wu HH, Bonnin A, Murray R & Calof AL. (2004) Dev. Neurosci. 26: 166−180. | Article | PubMed | ChemPort |
56. Kim J, Lo L, Dormand E & Anderson DJ. (2003) Neuron 38: 17−31. | Article | PubMed | ISI | ChemPort |
57. Kuan CT, Wikstrand CJ & Bigner DD. (2001) Endocr. Relat. Cancer 8: 83−96. | Article | PubMed | ISI | ChemPort |
58. Kumada T, Hasegawa A, Iwasaki Y, Baba H & Ikenaka K. (2004) Dev. Neurosci. 26: 68−76. | Article | PubMed | ChemPort |
59. Levine J & Willard M. (1981) J. Cell Biol. 90: 631−642. | Article | PubMed | ISI | ChemPort |
60. Lindholm D, Castren E, Kiefer R, Zafra F & Thoenen H. (1992) J. Cell Biol. 117: 395−400. | Article | PubMed | ChemPort |
61. Lo L, Dormand E, Greenwood A & Anderson DJ. (2002) Development 129: 1553−1567. | PubMed | ISI | ChemPort |
62. Lopez-Coviella I, Berse B, Thies RS & Blusztajn JK. (2002) J. Physiol. Paris 96: 53−59. | Article | PubMed | ChemPort |
63. Mann KM, Ray JL, Moon ES, Sass KM & Benson MR. (2004) J. Cell Biol. 165: 483−491. | Article | PubMed | ChemPort |
64. Markowitz S, Wang J, Myeroff L, Parsons R, Sun L, Lutterbaugh J, Fan RS, Zborowska E, Kinzler KW & Vogelstein B et al.. (1995) Science 268: 1336−1338. | PubMed | ISI | ChemPort |
65. Massague J. (1998) Annu. Rev. Biochem. 67: 753−791. | Article | PubMed | ISI | ChemPort |
66. Massague J, Andres J, Attisano L, Cheifetz S, Lopez-Casillas F, Ohtsuki M & Wrana JL. (1992) Mol. Reprod. Dev. 32: 99−104. | PubMed | ChemPort |
67. Massague J, Blain SW & Lo RS. (2000) Cell 103: 295−309. | Article | PubMed | ISI | ChemPort |
68. Massague J & Weis-Garcia F. (1996) Cancer Surv. 27: 41−64. | PubMed | ISI | ChemPort |
69. Matsuura I, Denissova NG, Wang G, He D, Long J & Liu F. (2004) Nature 430: 226−231. | Article | PubMed | ChemPort |
70. Mentlein R, Eichler O, Forstreuter F & Held-Feindt J. (2001) Int. J. Cancer 92: 545−550. | Article | PubMed | ChemPort |
71. Merzak A, McCrea S, Koocheckpour S & Pilkington GJ. (1994) Br. J. Cancer 70: 199−203. | PubMed | ChemPort |
72. Mishra L, Cai T, Levine A, Weng D, Mezey E, Mishra B & Gearhart J. (1998) Int. J. Dev. Biol. 42: 221−224. | PubMed | ChemPort |
73. Mishra L, Cai T, Yu P, Monga SP & Mishra B. (1999) Oncogene 18: 353−364. | Article | PubMed | ChemPort |
74. Moon RT & McMahon AP. (1987) BioEssays 7: 159−164. | Article | PubMed | ChemPort |
75. Moorthy S, Chen L & Bennett V. (2000) J. Cell Biol. 149: 915−930. | Article | PubMed | ISI | ChemPort |
76. Morrison SJ, Shah NM & Anderson DJ. (1997) Cell 88: 287−298. | Article | PubMed | ISI | ChemPort |
77. Moses HL & Serra R. (1996) Curr. Opin. Genet. Dev. 6: 581−586. | Article | PubMed | ISI | ChemPort |
78. Munoz-Sanjuan I & Brivanlou AH. (2002) Nat. Rev. Neurosci. 3: 271−280. | Article | PubMed | ISI | ChemPort |
79. Myeroff LL, Parsons R, Kim SJ, Hedrick L, Cho KR, Orth K, Mathis M, Kinzler KW, Lutterbaugh J & Park K et al.. (1995) Cancer Res. 55: 5545−5547. | PubMed | ISI | ChemPort |
80. Nagai S, Washiyama K, Kurimoto M, Takaku A, Endo S & Kumanishi T. (2002) J. Neurosurg. 96: 909−917. | PubMed | ISI | ChemPort |
81. Nakagawa T, Mabry M, de Bustros A, Ihle JN, Nelkin BD & Baylin SB. (1987) Proc. Natl. Acad. Sci. USA 84: 5923−5927. | PubMed | ChemPort |
82. Nelson WJ & Veshnock PJ. (1986) J. Cell Biol. 103: 1751−1765. | Article | PubMed | ChemPort |
83. Oft M, Peli J, Rudaz C, Schwarz H, Beug H & Reichmann E. (1996) Genes Dev. 10: 2462−2477. | PubMed | ISI | ChemPort |
84. Ohtsuka T, Ishibashi M, Gradwohl G, Nakanishi S, Guillemot F & Kageyama R. (1999) EMBO J. 18: 2196−2207. | Article | PubMed | ChemPort |
85. Okano-Uchida T, Himi T, Komiya Y & Ishizaki Y. (2004) Proc. Natl. Acad. Sci. USA 101: 1211−1216. | Article | PubMed | ChemPort |
86. Panchision DM & McKay RD. (2002) Curr. Opin. Genet. Dev. 12: 478−487. | Article | PubMed | ISI | ChemPort |
87. Panchision DM, Pickel JM, Studer L, Lee SH, Turner PA, Hazel TG & McKay RD. (2001) Genes Dev. 15: 2094−2110. | Article | PubMed | ISI | ChemPort |
88. Pardal R, Clarke MF & Morrison SJ. (2003) Nat. Rev. Cancer 3: 895−902. | Article | PubMed | ISI | ChemPort |
89. Paulus W, Baur I, Huettner C, Schmausser B, Roggendorf W, Schlingensiepen KH & Brysch W. (1995) J. Neuropathol. Exp. Neurol. 54: 236−244. | PubMed | ChemPort |
90. Peng Y, Xu RH, Mei JM, Li XP, Yan D, Kung HF & Phang JM. (2002) Neuroscience 109: 657−664. | Article | PubMed | ChemPort |
91. Pera MF, Andrade J, Houssami S, Reubinoff B, Trounson A, Stanley EG, Ward-van Oostwaard D & Mummery C. (2004) J. Cell Sci. 117: 1269−1280. | Article | PubMed | ChemPort |
92. Plate KH, Breier G, Farrell CL & Risau W. (1992) Lab. Invest. 67: 529−534. | PubMed | ChemPort |
93. Rajan P, Panchision DM, Newell LF & McKay RD. (2003) J. Cell Biol. 161: 911−921. | Article | PubMed | ISI | ChemPort |
94. Reiss M. (1997) Oncol. Res. 9: 447−457. | PubMed | ISI | ChemPort |
95. Reiss M. (1999) Microbes Infect. 1: 1327−1347. | Article | PubMed | ISI | ChemPort |
96. Reya T, Morrison SJ, Clarke MF & Weissman IL. (2001) Nature 414: 105−111. | Article | PubMed | ISI | ChemPort |
97. Riederer BM, Zagon IS & Goodman SR. (1986) J. Cell Biol. 102: 2088−2097. | Article | PubMed | ChemPort |
98. Roberts AB & Sporn MB. (1993) Growth Factors 8: 1−9. | PubMed | ISI | ChemPort |
99. Ruiz i Altaba A, Sanchez P & Dahmane N. (2002) Nat. Rev. Cancer 2: 361−372. | Article | PubMed | ChemPort |
100. Samuels V, Barrett JM, Bockman S, Pantazis CG & Allen MB, Jr. (1989) Am. J. Pathol. 134: 894−902. | PubMed | ChemPort |
101. Schier AF & Talbot WS. (2001) Int. J. Dev. Biol. 45: 289−297. | PubMed | ISI | ChemPort |
102. Setoguchi T, Nakashima K, Takizawa T, Yanagisawa M, Ochiai W, Okabe M, Yone K, Komiya S & Taga T. (2004) Exp. Neurol. 189: 33−44. | Article | PubMed | ChemPort |
103. Sieber-Blum M & Zhang JM. (1997) J. Anat. 191 (Part 4): 493−499. | Article | PubMed | ChemPort |
104. Sihag RK, Shea TB & Wang FS. (1996) J. Neurosci. Res. 44: 430−437. | Article | PubMed | ChemPort |
105. Sikorski AF & Goodman SR. (1991) Brain Res. Bull. 27: 195−198. | Article | PubMed | ChemPort |
106. Sikorski AF, Sangerman J, Goodman SR & Critz SD. (2000) Brain Res. 852: 161−166. | Article | PubMed | ISI | ChemPort |
107. Sonntag KC, Simantov R, Bjorklund L, Cooper O, Pruszak J, Kowalke F, Gilmartin J, Ding J, Hu YP, Shen MM & Isacson O. (2005) Mol. Cell Neurosci. 28: 417−429. | Article | PubMed | ChemPort |
108. Sporn MB. (1999) Microbes Infect. 1: 1251−1253. | Article | PubMed | ChemPort |
109. Tang B, Bottinger EP, Jakowlew SB, Bagnall KM, Mariano J, Anver MR, Letterio JJ & Wakefield LM. (1998) Nat. Med. 4: 802−807. | Article | PubMed | ISI | ChemPort |
110. Tang Y, Katuri V, Dillner A, Mishra B, Deng CX & Mishra L. (2003) Science 299: 574−577. | Article | PubMed | ISI | ChemPort |
111. Tang Y, Katuri V, Iqbal S, Narayan T, Wang Z, Lu RS, Mishra L & Mishra B. (2002) Oncogene 21: 5255−5267. | Article | PubMed | ChemPort |
112. Temple S. (2001) Nat. Rev. Neurosci. 2: 513−520. | Article | PubMed | ISI | ChemPort |
113. Tsai RY, Kittappa R & McKay RD. (2002) Dev. Cell 2: 707−712. | Article | PubMed | ISI | ChemPort |
114. Tsai RY & McKay RD. (2000) J. Neurosci. 20: 3725−3735. | PubMed | ISI | ChemPort |
115. Tsai RY & McKay RD. (2002) Genes Dev. 16: 2991−3003. | Article | PubMed | ISI | ChemPort |
116. Villanueva A, Garcia C, Paules AB, Vicente M, Megias M, Reyes G, de Villalonga P, Agell N, Lluis F, Bachs O & Capella G. (1998) Oncogene 17: 1969−1978. | Article | PubMed | ChemPort |
117. Wang CY, Mayo MW, Korneluk RG, Goeddel DV & Baldwin AS, Jr. (1998) Science 281: 1680−1683. | Article | PubMed | ISI | ChemPort |
118. Weissman IL, Anderson DJ & Gage F. (2001) Annu. Rev. Cell Dev. Biol. 17: 387−403. | Article | PubMed | ISI | ChemPort |
119. Weller M & Fontana A. (1995) Brain Res. Brain Res. Rev. 21: 128−151. | Article | PubMed | ChemPort |
120. Wrana JL. (2000) Cell 100: 189−192. | Article | PubMed | ISI | ChemPort |
121. Wurdak H, Ittner LM, Lang KS, Leveen P, Suter U, Fischer JA, Karlsson S, Born W & Sommer L. (2005) Genes Dev. 19: 530−535. | Article | PubMed | ChemPort |
122. Yamada N, Kato M, Yamashita H, Nister M, Miyazono K, Heldin CH & Funa K. (1995) Int. J. Cancer 62: 386−392. | PubMed | ChemPort |
123. Yang Y, Ikezoe T, Saito T, Kobayashi M, Koeffler HP & Taguchi H. (2004) Cancer Sci. 95: 176−180. | PubMed | ChemPort |
124. Yin D, Zhou H, Kumagai T, Liu G, Ong JM, Black KL & Koeffler HP. (2005) Oncogene 24: 344−354. | Article | PubMed | ChemPort |
125. Yin JJ, Selander K, Chirgwin JM, Dallas M, Grubbs BG, Wieser R, Massague J, Mundy GR & Guise TA. (1999) J. Clin. Invest. 103: 197−206. | PubMed | ISI | ChemPort |
126. Yung SY, Gokhan S, Jurcsak J, Molero AE, Abrajano JJ & Mehler MF. (2002) Proc. Natl. Acad. Sci. USA 99: 16273−16278. | Article | PubMed | ChemPort |
127. Zhang Y, Musci T & Derynck R. (1997) Curr. Biol. 7: 270−276. | Article | PubMed | ISI | ChemPort |
128. Zhou YX, Zhao M, Li D, Shimazu K, Sakata K, Deng CX & Lu B. (2003) J. Biol. Chem. 278: 42313−42320 (Epub 2003 Aug 1). | Article | PubMed | ChemPort |
129. Ziemnicka-Kotula D, Xu J, Gu H, Potempska A, Kim KS, Jenkins EC, Trenkner E & Kotula L. (1998) J. Biol. Chem. 273: 13681−13692. | Article | PubMed | ChemPort |

Acknowledgements

We would like to acknowledge Tiffany Blake and Vidhya Murugesan for their help with this article. This work was supported by NIH Grant R01 DK58637 (BM)