Abstract
Gliomas represent the most common type of brain tumor, but show considerable variability in histologic appearance and clinical outcome. The phenotypic differences between types and grades of gliomas have not been explained solely on the grounds of differing oncogenic stimuli. Several studies have demonstrated that some phenotypic differences may be attributed to regional differences in the neural stem cells from which tumors arise. We hypothesized that temporal differences may also play a role, with tumor phenotypic variability reflecting intrinsic differences in neural stem cells at distinct developmental stages. To determine how the tumorigenic potential of lineally related stem cells changes over time, we used a conditional transgenic system that integrates Cre-Lox–mediated and Tet-regulated expression to drive K-rasG12D expression in neuro-glial progenitor populations at different developmental time points. Using this model, we demonstrate that K-rasG12D–induced transformation is dependent on the developmental stage at which it is introduced. Diffuse malignant brain tumors develop during early embryogenesis but not when K-rasG12D expression is induced during late embryogenesis or early postnatal life. We show that differential expression of cell-cycle regulators during development may be responsible for this differing susceptibility to malignant transformation and that loss of p53 can overcome the transformation resistance seen at later developmental stages. These results highlight the interplay between genetic alterations and the molecular changes that accompany specific developmental stages; early progenitors may lack the regulatory mechanisms present at later, more lineage-restrictive, developmental time points, making them more susceptible to transformation.
Keywords: mouse models, temporal specification, lineage, differentiation
Gliomas represent the most common type of primary brain tumor in both children and adults. These tumors have been conventionally classified according to the normal central nervous system (CNS) cell type they most resemble—astrocyte, ependymal cell, or oligodendrocyte (1). However, in most cases, it is not known which cell type has actually given rise to the tumor. Furthermore, it is unclear to what extent the histologic type and clinical behavior of the tumor are determined by its cell of origin or by the initial genetic alteration that occurred in that cell. Currently, there is evidence implying an interaction between the cell of origin, the tumor microenvironment, and specific cancer-causing genetic changes in the evolution of CNS tumors (2, 3). Additionally, it has been established that differences between histologically similar brain tumors arising from distinct CNS locations have region-specific genetic alterations and molecular signatures (4, 5). These differences in tumor biology and behavior mimic many aspects of normal neural stem cell (NSC) behavior, as these cells also harbor intrinsic region-specific molecular differences, which dictate their behavior and differentiation potential (6, 7). A pending question, however, is how the transformation potential of NSCs varies as they progress through distinct developmental stages. A good candidate neuro-glial progenitor in which to address this question is the radial glial cell, as it progresses stepwise through distinct developmental stages and is the neonatal origin of adult subventricular zone (SVZ) neural stem cells (8).
Radial glia comprise a molecularly defined cellular population playing critical roles in CNS development as both neural and glial progenitors (8, 9) and as scaffolding for migratory neurons (10). They derive from neuroepithelial cells and are first identified in the embryonic brain at the onset of neurogenesis (8, 11). The transition from neuroepithelial cells to radial glial cells is marked by the expression of glial markers such as the astrocyte-specific glutamate transporter and the brain lipid-binding protein (BLBP). Specifically, the BLBP promoter has been used extensively to study the neurogenic and gliogenic potentials of radial glial cells as well as their temporal specification (11, 12). These studies indicate that radial glial populations progress stepwise through distinct developmental stages during which particular classes of neurons or glia are generated (11). Using this and viral lineage-tracing techniques, it has been shown that both parenchymal and SVZ astrocytes in the postnatal brain appear to directly derive from radial glia. However, only SVZ astrocytes retain expression of critical determinants for their function as NSCs (8).
To understand how intrinsic differences between lineally related NSCs can influence malignant transformation, we induced expression of activated K-ras into radial glial progenitors at distinct developmental time points. Ras-pathway activation has been demonstrated in a wide spectrum of both high- and low-grade adult and pediatric gliomas (13, 14). This often occurs through overexpressed or activated receptor tyrosine kinases such as EGFR or PDGFR or through mutations in downstream pathway members such as NF1 or BRAF (15, 16). To target radial glial cells, we used the BLBP promoter, as it is expressed specifically at the neurogenic and gliogenic stages of radial glial development (11), allowing us to test the transformation potential of this progenitor population across different stages of its development.
Our work demonstrates that the potential for active K-ras–induced transformation is highly dependent on the developmental stage at which it occurs. K-ras alone is sufficient to initiate tumorigenesis when targeted to early stages of radial glial development but not at later stages, when they have committed to a glial lineage. These results are of clinical interest, as they highlight the importance of the developmental context of oncogene expression and how it may determine the phenotype and biologic aggressiveness of the tumor.
Results
Oncogenic K-Ras Efficiently Initiates Tumorigenesis When Targeted to Embryonic Neuro-Glial Progenitors.
Although it has become evident that NSCs at different developmental stages are distinct in their response to growth factors, proliferation, differentiation, and cell-cycle regulation (17), the idea that these differences might affect tumor initiation and progression has not been studied. To evaluate this, we developed a mouse model in which we could achieve spatial and temporal control of transgene expression. For this purpose, BLBP-Cre mice were intercrossed with double transgenics carrying both the tetracycline transcriptional activator (rtTA) preceded by a loxP-flanked stop signal and K-rasG12D under the control of the tetracycline response element (TRE) (Fig. S1A). After Cre-mediated excision of the loxP-flanked sequence, expression of the rtTA is driven by the ROSA26 promoter. The presence of doxycycline (Dox) results in the formation of an active transcriptional activator and the expression of K-RasG12D in BLBP-positive radial glial cells (Fig. S1B) (18).
All crosses resulted in progeny with the expected Mendelian ratios, and expression of active Ras was restricted to transgenics expressing Cre in the presence of doxycycline (Fig. S1). Active Ras expression led to the activation of downstream targets such as ERK1/2 and Akt as well as an increase in cyclin D1 (Fig. 1A).
Fig. 1.
K-rasG12D expression alone is sufficient to initiate tumorigenesis when targeted to an early neuro-glial progenitor population. (A) Ras-GTP pull-downs on forebrain derived from presymptomatic transgenic (Tg) and control (Ctr) littermates at P10. (B) Weight reduction was observed at P20. Values represent mean and SD of the measurements; control n = 6, transgenic n = 8; *P < 0.001, two-tailed paired t test. (C) Mice with expression of K-rasG12D (Tg) have shortened survival compared with control littermates. Kaplan–Meier survival curves of transgenic mice with active K-ras expression induced at E0 show a median survival of 30 d. Control n = 16, Tg n = 15; *P < 0.001, log-rank test. (D–G) Representative hematoxylin and eosin (H&E)-stained brain sections demonstrate hydrocephalus and brain hypercellularity (D and F) compared with controls (E and G). Tumor cells expressed GFAP and Nestin (H–K) as well as Ki67 as a proliferative marker (L and M). The tumors showed activation of the RAS-MAPK pathway as evidenced by p-ERK expression (N), which was not present in controls (O). (P and Q) Expression of Cre was observed in triples that developed tumors. [Scale bars, 2,000 μm (D and E) and 100 μm (Q).]
Transgenic mice where K-ras induction started at embryonic day 0 (E0) were externally indistinguishable until postnatal day 20 (P20), at which point their weight was 30% less than control littermates (Fig. 1B). The median survival for these transgenics was 30 d and none survived past 2 mo of age (Fig. 1C). Upon histological examination, the brains of all transgenic mice were hypercellular with diffuse infiltration of cells with nuclear atypia and mitotic activity reminiscent of World Health Organization grade III gliomatosis cerebri seen in human patients (19–21) (Fig. 1 D–G). Neoplastic cells showed robust expression of GFAP and Nestin, markers of human astrocytic tumors, and were highly proliferative as assessed by Ki67 staining (Fig. 1 H–M). The tumor cells show Ras-MAPK–pathway activation as evidenced by robust p-ERK expression (Fig. 1O). Collectively, these data indicate that K-RasG12D expression efficiently initiates tumorigenesis in radial glial cells when expressed during early embryonic time points.
The Tumorigenic Effects of K-RasG12D in the Brain Are Restricted to Early Developmental Time Points.
Although abnormal glial phenotypes have been described in other models of high Ras activity (22), the development of actual brain tumors has not been reported in the presence of activated Ras alone. Thus, it was thought that additional cooperating mutations were necessary to induce transformation. Interestingly, all of these experiments leading to this conclusion were carried out at early postnatal time points (23, 24). As opposed to these reports, we have shown that when K-rasG12D is targeted to an early embryonic neuro-glial progenitor, gliomas do develop. Possible explanations for this discrepancy could be differences in the promoters used to drive oncogenic K-ras. Alternatively, these differences could be of biological relevance, as they might reflect the fact that the mutations necessary for the transformation of a specific cellular population are partially determined by its developmental stage and the intrinsic changes that accompany the temporal development of NSCs. To test this hypothesis, we took advantage of the temporal inducibility of our model to target expression of K-rasG12D to specific radial glial developmental time points. In these experiments, K-rasG12D was induced at E9.5, E14, E16, E18, and P0. As shown in Fig. 2A, we noted a striking and statistically significant difference in survival dependent on the developmental stage at which active K-ras was induced; mice where K-rasG12D expression was induced at E9.5 or E14 had much shorter survival than those where K-rasG12D expression was induced at later embryonic or postnatal time points (Fig. 2A). Upon histological examination it became evident that the extent of proliferation varied depending on the developmental stage at which active K-ras was induced; mice where K-rasG12D expression was induced at earlier time points have a more prominent and diffuse proliferation compared with the more restricted proliferation observed in late embryonic and postnatal time points (Fig. 2B). This difference was not attributed to varying levels of K-ras activity. As shown in Fig. 2C, upon doxycycline induction, transgenic mice had similar levels of active Ras.
Fig. 2.
K-rasG12D tumor initiation is age-dependent. (A) Kaplan–Meier survival plot of mice with K-rasG12D expression induced at distinct developmental time points. Induction of KrasG12D at E9.5 or E14 leads to widespread neoplastic transformation and early death (median survival 27 d) in contrast to induction at E16, E18, or P0 (*P < 0.001, log-rank test: E14 vs. E18; n = 10). (B) Representative H&E-stained sections from transgenic mice killed 21 d after Dox induction. The bluer color seen in E9.5 and E14 is due to diffuse involvement by neoplastic cells in these brains vs. those where active K-ras was expressed at later developmental time points. (Scale bar, 1,000 μm.) (C) Western blot demonstrating Ras activation (Ras-GTP) at all developmental time points despite differences in tumor initiation. The two sides of the Ras pull-down image are from the same exposure of the same blot, but the samples were not adjacent to each other.
We hypothesized that these differences may be attributed to the molecular changes that accompany the maturation of radial glial cells between E14 and E16 and that have major effects on the transforming potential of this cell population. This raised the intriguing possibility that the interplay between the molecular changes that accompany the temporal development of NSCs and the oncogenic stimulus is pivotal to the process of tumorigenesis.
Differences in Tumorigenicity Reflect the Cell Context of Oncogene Expression.
When the phenotypes observed during early embryogenesis and later stages were compared, it became apparent that the extent of brain involvement by proliferative cells was lessened at later developmental stages (Fig. 3A). Proliferative lesions were not seen in the midbrain or cortex in mice where K-rasG12D was induced at E16 vs. E14 (Fig. 3A). The proliferative lesions that developed during late embryogenesis and postnatal stages were exclusively restricted to neurogenic regions (Fig. S2 C–E). As shown in Fig. 3B, radial glial cells from E14 to E16 are labeled with the glial/progenitor markers GFAP, BLBP, and Nestin and retain their radial morphology, which extends to cortical regions (Fig. 3B and Fig. S3). However, despite these morphologic similarities, by E16 cortical radial glial cells have committed to the glial lineage and most of them will differentiate and become cortical astrocytes (12), explaining the lack of transformation observed in cortical regions past E16 (Figs. S2 C–E and S4 C–E). This is in concordance with published data showing that mature mouse astrocytes are less susceptible to transformation in vivo (23, 25). However, neurogenic regions of the brain, such as the SVZ, still develop proliferative lesions upon expression of active Ras from E16 to P0 (Fig. S2). Interestingly, active Ras has even less effect at postnatal stages (Table S1); most transgenics where active Ras is expressed at postnatal stages do not develop malignancies or regions of proliferation even though progenitor populations in the SVZ have undergone excision and are capable of proliferation (Fig. S5 A and B).
Fig. 3.
Induction of K-ras at later developmental stages induces proliferative lesions that are restricted to neurogenic regions and no longer behave malignantly. (A) (Upper) Representative (H&E) images of mice where K-rasG12D was induced at E14 or E16. Mice were killed 21 d after induction. (Lower) Immunohistochemical staining for Ki67. Colored squares represent the regions of high magnification depicted below. [Scale bars, 1,000 μm (Upper) and 100 μm (Lower).] RMS, rostral migratory stream. (B) Characteristic expression of Nestin, BLBP, and GFAP in radial glial cells at E14 and E16. Arrowheads indicate cell body–associated BLBP, Nestin, and GFAP immunoreactivity. CP, cortical plate; VZ, ventricular zone. (Scale bar, 61 μm.)
An advantage of our model is that its inducibility does not enable Cre excision but instead enables transgene expression. As a result, all of the progeny deriving from BLBP-positive cells will express active K-ras upon Dox administration. This is of importance, as postnatal neural stem cells in the SVZ are lineally related to BLBP-positive radial glial cells (8, 26) and thus have undergone excision; however, they no longer undergo transformation upon Dox induction. These results suggest that as neuro-glial progenitors progress to later developmental stages they are no longer able to initiate the tumorigenic process with active K-ras alone.
We hypothesized that a variation in expression levels of major cell-cycle regulators during the temporal development of neural progenitors might be responsible for the observed differences. Ink4a/Arf serves as a critical node linking upstream signals such as those induced by active Ras to downstream effector pathways during oncogene-induced tumor suppression. Whether specific NSC populations during distinct stages of their development acquire different inherent abilities to engage these programs following oncogenic insults remains largely unknown. Such variability could have a profound influence on tumor susceptibility.
To determine whether tumor development in our model was associated with a loss of Ink4a/Arf, we looked at its expression level in presymptomatic (P10) and end-stage (P21) transgenics following K-rasG12D induction at E0. Quantitative (q)RT-PCR analysis revealed a fivefold decrease in the cyclin-dependent kinase inhibitor 2A (Cdkn2A) expression in end-stage transgenics compared with control littermates of the same age (Fig. 4A). We then went on to isolate BLBP-positive cells using BLBP-GFP transgenic mice (Fig. S6) to determine whether there was a developmental stage-dependent change in the expression of cell-cycle regulators. As shown in Fig. 4B, there is a dramatic decrease in BLBP-GFP–positive cells at postnatal time points, in accordance with the published literature (11, 12). However, at E14 and E16, mouse brains have similar levels of BLBP expression. Despite this, transgenics induced at E16 no longer succumb to the proliferative lesions that develop. Although it could be argued that the decrease in BLBP expression could explain the differences in tumorigenicity, we have shown that NSC populations at all stages express similar levels of active K-ras.
Fig. 4.
Differential expression of cell-cycle regulatory genes during development may underlie the differential malignant progression of BLBP-expressing neural progenitors in response to active Ras. (A) Comparison of log2-transformed expression ratios of Cdkn2A in presymptomatic (P10) and symptomatic (P21) transgenics. Data represent relative quantity normalized to 18S rRNA and expressed relative to control littermates of the same age. (B) Quantification of BLBP+ and BLBP− cells after sorting of each developmental stage. Data represent the mean percentage ± SD of BLBP− GFP cells after fluorescence-activated cell sorting; n = 3. (C) Log2-transformed expression ratios of cell-cycle genes in BLBP-expressing cells from mouse brains at each developmental time point as determined by qPCR. Data represent relative quantity normalized to 18S rRNA and expressed relative to E14. Error bars represent SD; n = 3.
To determine the expression levels of cell-cycle regulators at specific developmental time points, we carried out qRT-PCR analysis in GFP-sorted BLBP-positive cells. As shown in Fig. 4C, a robust increase in Arf is observed at late prenatal and postnatal time points compared with E14, and this is accompanied by the down-regulation of cell-cycle progression regulators such as Cdk4, Cdc25A, and Cdc25C as well as a decrease in cyclin-dependent kinase inhibitors (CDKIs) such as p57Kip2 and p27Kip1. The decrease in these major cell-cycle regulators correlates with the time points at which targeted BLBP-positive cells no longer undergo transformation.
An explanation for the increase in Arf induction at late prenatal and postnatal time points underlies cell type-specific locus regulation that can in turn set different thresholds for gene expression, a mode of regulation that has been shown to be important in multiple types of stem cells compared with their differentiated progeny (27). In radial glial cells, this regulation might come from differential expression of bmi-1 and ezh2 during development, as a decrease in these Arf repressors is observed as a function of age (Fig. 4C).
To demonstrate that the differences in cell-cycle regulators inherent to specific developmental stages underlie the differences in tumor susceptibility in our model, we crossed our K-rasG12D mouse model to p53 flox mice (Fig. S7 A–C), as it has been shown that the antitumorigenic effects of Arf in response to oncogenic Ras are dependent on the activation of p53 (28). In this case, Dox induction was initiated at postnatal day 0, where we previously saw a loss in transformation potential. Interestingly, when combined with loss of p53, K-rasG12D now induced tumors at P0. These tumors also showed elevated Ki67 and p-ERK expression, but contained a subpopulation of large atypical cells that were less frequently seen in the tumors induced at E14 with WT p53 (Fig. S7D). These mice succumbed to their disease 48 d after Dox administration, in contrast to wild-type p53 mice on Dox, who remained healthy.
Collectively, these data highlight the interplay between an oncogenic stimulus and the molecular changes the cell is undergoing during its progressive restriction in fate potential. This suggests that the susceptibility of neural stem cell populations to transformation is dependent on the cell’s inherent ability to engage tumor-suppressor pathways.
Discussion
As development progresses, neural progenitors, such as radial glial cells, decrease in number and their proliferation declines to low levels; the remaining neural stem cells become tightly regulated to ensure that they do not hyperproliferate in adult tissues. These control mechanisms are likely to be imposed during the cell’s progressive restriction in fate potential. Thus, it is important to understand how the susceptibility of these progenitor populations to transformation changes as a function of their maturation and thus their ability to restrict growth.
In the present study, we identify naturally occurring, developmentally dependent variability in the tumorigenic effects of active K-ras when targeted to radial glial cells. Active K-ras was chosen as an oncogenic stimulus, as Ras-pathway activation has been demonstrated in a wide spectrum of both high- and low-grade adult and pediatric gliomas (13, 14). Thus, this may provide some insight into the differences observed between pediatric and adult brain tumors, especially because Ras-pathway mutations are rare in adult glioblastomas but are more prevalent in pediatric low-grade astrocytomas (29).
We showed that active K-ras alone was able to induce diffuse malignant gliomas when targeted to early stages of radial glial development but not at later stages, when these cells have presumably become restricted in their fate potential (11, 30, 31). Although previous studies have indicated that K-ras alone is unable to induce transformation (23, 24), we believe that this discrepancy reflects the fact that these studies were carried out on newborn pups (P0 or later); at this time point, we observed only small proliferative lesions with only 20% penetrance.
As opposed to previous models, we have taken advantage of the lineage-tracing and -inducible characteristics of our model, which allow us to track the progeny that derived from BLBP+ cells and their response to the oncogenic effects of active K-ras at distinct developmental time points. This approach was validated in that, as cortical radial glial cells became terminally differentiated at E16–E18.5 (30, 31), they became refractive to the oncogenic effects of active K-ras alone (Figs. S2 and S4), in accordance with what has previously been shown to be true for differentiated astrocytes (24). Interestingly, only neurogenic regions develop proliferative lesions at late embryonic time points, and mice no longer succumb to these lesions. Even more strikingly, at postnatal time points, neural stem cell populations deriving from radial glial cells do not undergo transformation despite expression of active K-ras (Fig. 2C and Fig. S5). The difference in the transforming capacity of active K-ras between prenatal and postnatal stages led us to hypothesize that there may be differences in the ability of NSCs to engage tumor-suppressor pathways as they progress through distinct developmental stages.
By sorting the respective SVZ (BLBP+) populations at defined developmental stages, we were able to show that indeed the level of cell-cycle regulators in these cells varies as a function of age, reflecting the changes in cell-cycle kinetics undergone by radial glia during development and mirroring the ability of active K-ras to induce transformation. First, a down-regulation of cell-cycle activators in NSCs with development reflects a decrease in their proliferative potential, a process that has been linked to the expression of Ink4a/Arf in aging tissues (32). Similarly, in our study there is an increase in Arf expression of BLBP-positive cells with development. The higher levels of other CDKIs such as p27 and p57 at E14 echo their early developmental involvement in cell-cycle regulation, exit, and differentiation as observed during corticogenesis (33). Once corticogenesis ceases at postnatal stages, the expression of P27 and P57 is reduced in comparison with early embryonic stages. Thus, the decrease does not indicate a lack of importance but rather that the relative abundance of these cell-cycle regulators is directly proportional to the cell’s intrinsic propensity to differentiate.
In our model, higher expression of the tumor suppressor Arf at later prenatal and postnatal time points inversely correlates with the ability of K-ras mutations alone to initiate radial glial cell transformation. Induction of active K-ras at early prenatal stages leads to the development of malignant brain tumors accompanied by a fivefold decrease in Cdkn2A expression in tumor-bearing mice compared with presymptomatic transgenics (Fig. 4A). In contrast, induction of active K-ras at later prenatal and postnatal time points, despite similar activation levels, fails to induce malignant tumors and has no effect on survival. Thus, resistance to oncogenic K-ras may reflect a developmental activation threshold for Ink4a/Arf which might be related to the basal proliferative rate of cells at different stages in their development. This idea is not new, as it has been shown that expression of Ink4a/Arf increases with age in many tissue-specific stem cells, including the NSCs (32). However, our analysis shows that even though postnatal neural stem cells derive from embryonic radial glial cells, their response to the same oncogenic stimulus is distinct. These distinctions reflect the inherent ability of the cell to engage tumor-suppressor pathways in response to an oncogenic stimulus. A further validation of this concept comes from our cross to p53 flox mice. The absence of p53 permits active K-ras to initiate malignant tumors at postnatal day 0 which, in the presence of p53, only occurred at E14 and earlier (Fig. S7D).
Overall, our results highlight the interplay between genetic alterations and the molecular changes that accompany the temporal development of NSCs, and further emphasize the need to view the tumorigenic process of gliomas in the context of normal brain development. The cell context of oncogene expression may determine the phenotype and biologic aggressiveness of the tumor. Thus, the results of genetic or epigenetic alterations leading to brain tumors may be quite different during the course of CNS development, suggesting unique treatment targets and strategies may be required.
Materials and Methods
Mice.
Information on mouse strains is provided in SI Materials and Methods. Control and mutant mice were given doxycycline in food pellets (Bio-Serv) at a concentration of 6 g/kg. For time-course experiments, timed pregnant females (day of plug, E0) were exposed to doxycycline once embryos reached the desired stage. Females were kept on Dox throughout pregnancy and lactation. For mice where Dox was started at late embryonic and early postnatal stages, mice were killed at 12 wk of age. All mouse experiments were approved by and performed according to the guidelines of the Institutional Animal Care and Use Committee of the University Health Network (Toronto Western Hospital).
Neurosphere Cultures.
Neurospheres were derived from transgenic and control littermates at E14, E16, E18, and P0. Neurospheres were grown in the presence of basic fibroblast growth factor (10 ng/mL), EGF (10 ng/mL), and doxycycline (100 ng/mL). Primary neurospheres were dissociated 5 d after plating and were used for active Ras pull-downs (Millipore).
Fluorescence-Activated Cell Sorting.
The ventricular zone was isolated from E14, E18, P10, and P40 BLBP-GFP mice. Tissues were dissociated in 20 μg/mL papain, 1 mM l-cysteine, and 0.5 mM EDTA in Earle's Balanced Salt Solution (EBSS) (Worthington) to single cells and filtered through a 45-μm nylon mesh before final centrifugation. GFP-positive cells were sorted on a FACSAria (BD Biosciences) using 20-psi pressure and 100-μm nozzle aperture. FlowJo software (Tree Star) was used for analysis. Gates were set manually with GFP-negative control samples. Data were analyzed with FlowJo data analysis software and displayed using logical (biexponential) scaling.
Statistical Analysis.
Kaplan–Meier curves for mouse glioma latency were made using GraphPad Prism 4.0 software and analyzed with a standard log-rank test. Each experiment was performed with samples from at least three animals from independent litters. Statistical significance (P < 0.05) was determined by unpaired t test using GraphPad Prism 4.0 software. For RT-PCR, statistical analyses were done on the cycle threshold (CT) values.
Supporting Information.
Genotyping by PCR, protein extraction and immunoblots, tissue preparation, immunostaining and mRNA isolation/quantitative PCR methods, and information on mice lines are provided in SI Materials and Methods.
Supplementary Material
Acknowledgments
We dedicate this work to the memory of Dr. Abhijit Guha, who passed away during the final preparation of the manuscript. He was the leader and primary progenitor of the work described. We thank Dr. Aaron Gajadhar for critical comments on the manuscript. We are grateful to Dr. H. Varmus for providing the TRE-K-ras mice and to Dr. J. Rutka for providing the BLBP-GFP mice. This study was partially funded by Canadian Cancer Research Society [formerly National Cancer Institute of Canada (NCIC)-Canada] and Canadian Institutes of Health Research (MOP 102513).
Footnotes
The authors declare no conflict of interest.
This article is a PNAS Direct Submission. E.C.H. is a guest editor invited by the Editorial Board.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1303504110/-/DCSupplemental.
References
- 1.Kleihues P, Cavanee WK, editors. World Health Organization Classification of Tumours: Pathology and Genetics: Tumours of the Nervous System. New York: Oxford Univ Press; 2000. [Google Scholar]
- 2.Swartling FJ, et al. Distinct neural stem cell populations give rise to disparate brain tumors in response to N-MYC. Cancer Cell. 2012;21(5):601–613. doi: 10.1016/j.ccr.2012.04.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Johnson RA, et al. Cross-species genomics matches driver mutations and cell compartments to model ependymoma. Nature. 2010;466(7306):632–636. doi: 10.1038/nature09173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Taylor MD, et al. Radial glia cells are candidate stem cells of ependymoma. Cancer Cell. 2005;8(4):323–335. doi: 10.1016/j.ccr.2005.09.001. [DOI] [PubMed] [Google Scholar]
- 5.Lee Y, Gianino SM, Gutmann DH. Innate neural stem cell heterogeneity determines the patterning of glioma formation in children. Cancer Cell. 2012;22(1):131–138. doi: 10.1016/j.ccr.2012.05.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Merkle FT, Mirzadeh Z, Alvarez-Buylla A. Mosaic organization of neural stem cells in the adult brain. Science. 2007;317(5836):381–384. doi: 10.1126/science.1144914. [DOI] [PubMed] [Google Scholar]
- 7.Hochstim C, Deneen B, Lukaszewicz A, Zhou Q, Anderson DJ. Identification of positionally distinct astrocyte subtypes whose identities are specified by a homeodomain code. Cell. 2008;133(3):510–522. doi: 10.1016/j.cell.2008.02.046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Merkle FT, Tramontin AD, García-Verdugo JM, Alvarez-Buylla A. Radial glia give rise to adult neural stem cells in the subventricular zone. Proc Natl Acad Sci USA. 2004;101(50):17528–17532. doi: 10.1073/pnas.0407893101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Alvarez-Buylla A, Nottebohm F. Migration of young neurons in adult avian brain. Nature. 1988;335(6188):353–354. doi: 10.1038/335353a0. [DOI] [PubMed] [Google Scholar]
- 10.Cappello S, et al. A radial glia-specific role of RhoA in double cortex formation. Neuron. 2012;73(5):911–924. doi: 10.1016/j.neuron.2011.12.030. [DOI] [PubMed] [Google Scholar]
- 11.Anthony TE, Heintz N. Genetic lineage tracing defines distinct neurogenic and gliogenic stages of ventral telencephalic radial glial development. Neural Dev. 2008;3:30. doi: 10.1186/1749-8104-3-30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Anthony TE, Klein C, Fishell G, Heintz N. Radial glia serve as neuronal progenitors in all regions of the central nervous system. Neuron. 2004;41(6):881–890. doi: 10.1016/s0896-6273(04)00140-0. [DOI] [PubMed] [Google Scholar]
- 13.Guha A, Feldkamp MM, Lau N, Boss G, Pawson A. Proliferation of human malignant astrocytomas is dependent on Ras activation. Oncogene. 1997;15(23):2755–2765. doi: 10.1038/sj.onc.1201455. [DOI] [PubMed] [Google Scholar]
- 14.Sharma MK, Zehnbauer BA, Watson MA, Gutmann DH. RAS pathway activation and an oncogenic RAS mutation in sporadic pilocytic astrocytoma. Neurology. 2005;65(8):1335–1336. doi: 10.1212/01.wnl.0000180409.78098.d7. [DOI] [PubMed] [Google Scholar]
- 15.Feldkamp MM, Lau N, Guha A. Signal transduction pathways and their relevance in human astrocytomas. J Neurooncol. 1997;35(3):223–248. doi: 10.1023/a:1005800114912. [DOI] [PubMed] [Google Scholar]
- 16.Cancer Genome Atlas Research Network Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008;455(7216):1061–1068. doi: 10.1038/nature07385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Temple S. The development of neural stem cells. Nature. 2001;414(6859):112–117. doi: 10.1038/35102174. [DOI] [PubMed] [Google Scholar]
- 18.Belteki G, et al. Conditional and inducible transgene expression in mice through the combinatorial use of Cre-mediated recombination and tetracycline induction. Nucleic Acids Res. 2005;33(5):e51. doi: 10.1093/nar/gni051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Malamud N, Wise BL, Jones OW., Jr Gliomatosis cerebri. J Neurosurg. 1952;9(4):409–417. doi: 10.3171/jns.1952.9.4.0409. [DOI] [PubMed] [Google Scholar]
- 20.Yamada SM, et al. Histological and genetic diagnosis of gliomatosis cerebri: Case report. J Neurooncol. 2001;52(3):237–240. doi: 10.1023/a:1010601132673. [DOI] [PubMed] [Google Scholar]
- 21.Kannuki S, Hirose T, Horiguchi H, Kageji T, Nagahiro S. Gliomatosis cerebri with secondary glioblastoma formation: Report of two cases. Brain Tumor Pathol. 1998;15(2):111–116. doi: 10.1007/BF02478893. [DOI] [PubMed] [Google Scholar]
- 22.Hegedus B, et al. Neurofibromatosis-1 regulates neuronal and glial cell differentiation from neuroglial progenitors in vivo by both cAMP- and Ras-dependent mechanisms. Cell Stem Cell. 2007;1(4):443–457. doi: 10.1016/j.stem.2007.07.008. [DOI] [PubMed] [Google Scholar]
- 23.Holland EC, et al. Combined activation of Ras and Akt in neural progenitors induces glioblastoma formation in mice. Nat Genet. 2000;25(1):55–57. doi: 10.1038/75596. [DOI] [PubMed] [Google Scholar]
- 24.Uhrbom L, et al. Ink4a-Arf loss cooperates with KRas activation in astrocytes and neural progenitors to generate glioblastomas of various morphologies depending on activated Akt. Cancer Res. 2002;62(19):5551–5558. [PubMed] [Google Scholar]
- 25.Bachoo RM, et al. Epidermal growth factor receptor and Ink4a/Arf: Convergent mechanisms governing terminal differentiation and transformation along the neural stem cell to astrocyte axis. Cancer Cell. 2002;1(3):269–277. doi: 10.1016/s1535-6108(02)00046-6. [DOI] [PubMed] [Google Scholar]
- 26.Kriegstein A, Alvarez-Buylla A. The glial nature of embryonic and adult neural stem cells. Annu Rev Neurosci. 2009;32:149–184. doi: 10.1146/annurev.neuro.051508.135600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Valk-Lingbeek ME, Bruggeman SW, van Lohuizen M. Stem cells and cancer: The Polycomb connection. Cell. 2004;118(4):409–418. doi: 10.1016/j.cell.2004.08.005. [DOI] [PubMed] [Google Scholar]
- 28.Palmero I, Pantoja C, Serrano M. p19ARF links the tumour suppressor p53 to Ras. Nature. 1998;395(6698):125–126. doi: 10.1038/25870. [DOI] [PubMed] [Google Scholar]
- 29.Zhang J, et al. St. Jude Children’s Research Hospital–Washington University Pediatric Cancer Genome Project Whole-genome sequencing identifies genetic alterations in pediatric low-grade gliomas. Nat Genet. 2013;45(6):602–612. doi: 10.1038/ng.2611. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Hartfuss E, Galli R, Heins N, Götz M. Characterization of CNS precursor subtypes and radial glia. Dev Biol. 2001;229(1):15–30. doi: 10.1006/dbio.2000.9962. [DOI] [PubMed] [Google Scholar]
- 31.Malatesta P, Hartfuss E, Götz M. Isolation of radial glial cells by fluorescent-activated cell sorting reveals a neuronal lineage. Development. 2000;127(24):5253–5263. doi: 10.1242/dev.127.24.5253. [DOI] [PubMed] [Google Scholar]
- 32.Krishnamurthy J, et al. Ink4a/Arf expression is a biomarker of aging. J Clin Invest. 2004;114(9):1299–1307. doi: 10.1172/JCI22475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Tury A, Mairet-Coello G, DiCicco-Bloom E. The cyclin-dependent kinase inhibitor p57Kip2 regulates cell cycle exit, differentiation, and migration of embryonic cerebral cortical precursors. Cereb Cortex. 2011;21(8):1840–1856. doi: 10.1093/cercor/bhq254. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.