Sage Journals: Discover world-class research

Abstract

Malignant glial tumors are the most aggressive and difficult to treat neoplasms arising in the brain. More than 22,000 people in the United States are diagnosed with a malignant glioma annually, and most will die within the first two years from diagnosis. Traditionally, gliomas have been categorized based solely on tumor histological features. However, expression studies have found that molecular signatures can be used to categorize these tumors into subclasses that more effectively predict patient outcome. The heterogeneity between tumors as well as within individual tumors makes understanding the molecular aspects of tumorigenesis extremely important. Several genetically engineered mouse models (GEMMs) of glioma have been developed that recapitulate the molecular alterations observed in the human disease. GEMMs of glioma have allowed researchers to more closely study the role of cancer stem cells (CSC) in gliomagenesis as well as the relevance of signaling within the CSC microenvironment. Knowledge of the underlying molecular signatures of malignant glial tumors coupled with the existence of a variety of human disease-relevant GEMMs of this tumor type provide researchers and clinicians with valuable resources for the discovery of new drug targets.

Keywords

carcinogenesis genetically engineered mice molecular pathology

Introduction/Relevance

Gliomas (a primary tumor of glial cell origin) are the most common intracranial neoplasm with astrocytomas, glioblastomas, and oligodendrogliomas accounting for more than 80% (Central Brain Tumor Registry of the United States [CBTRUS] 2010; Pfister, Hartmann, and Korshunov 2009; Merchant, Pollack, and Loeffler 2010) (Figure 1 ). More than 22,000 people are diagnosed with a malignant glioma (also referred to as high-grade gliomas) in the United States each year (CBTRUS 2010; Wen and Kesari 2008), and most will die within the first two years from diagnosis despite aggressive therapy. The World Health Organization (WHO) classifies malignant gliomas as grades III and IV based on tumor histological and immunohistochemical features (Louis et al. 2007). Anaplastic astrocytomas, anaplastic oligodendrogliomas, and anaplastic oligodendroastrocytomas are classified as grade III; and glioblastoma multiforme (GBM) and other less common variants are classified as grade IV. GBMs, the most aggressive and deadly of these tumors, are the most common cancer of the central nervous system (CNS) accounting for 53% of gliomas overall (Figure 1) (CBTRUS 2010). By contrast, low-grade gliomas (WHO grades I and II) are relatively benign and afford a much better prognosis as compared to their higher-grade counterparts; however, grade II gliomas have the propensity to progress to grade IV GBMs over time, especially in the younger population (Huse and Holland 2010).

Figure 1.

Frequency of primary CNS gliomas. Pie chart showing the frequency of primary glial tumors by histology group diagnosed in the United States from 2004 to 2006 (CBTRUS 2010).

Despite more than 40 years of research, little improvement in mortality rates has been made for patients diagnosed with malignant glioma owing to the intra- and inter-tumoral heterogeneity and the propensity of the cancerous cells to infiltrate into the normal brain parenchyma. Further complicating therapy, the blood-brain barrier prevents chemotherapeutic agents from adequately penetrating into the tumor mass and leads to suboptimal therapeutic response. The challenge in providing effective therapies for the treatment of malignant gliomas is evidenced in a 5-year survival rate (less than 5% for GBM) that has remained practically unchanged over the past several decades. Undoubtedly, the key to developing effective therapies is to obtain an in-depth understanding of the biological mechanisms important to gliomagenesis.

Histological and Molecular Pathology of Malignant Gliomas

Histological Pathology

In 1979, the WHO formalized the first classification system for tumors of the CNS. This system set criteria to group CNS tumors by grade (WHO grades I through IV) (Table 1 ), with increasing grade reflecting an increase in tumor aggressiveness and a worse prognosis. Grade I glial tumors are benign, slow-growing, and well-circumscribed tumors that are typically curable with resection. Grade II glial tumors infiltrate into normal brain tissue and exhibit moderate proliferative features. These tumors have the potential to undergo malignant transformation (~70%) within 5 to 10 years from diagnosis (Furnari et al. 2007). Grades III and IV tumors are classified as malignant gliomas, with grade IV being extremely aggressive with hallmarks of vascular and uncontrolled cellular proliferation, diffuse infiltration, and necrosis. Since the inception of this system, three updates have been made and described in subsequent editions. Most recently (2007), the WHO published its fourth edition that reflects changes in tumor grade, new entities, and variants of some CNS tumors (Louis et al. 2007; Brat et al. 2008). This classification system has been used by clinicians, pathologists, and scientists to collectively propel the advancement of epidemiologic and clinical studies of brain tumors.

Table 1.

World Health Organization (WHO) classification of common glial tumors in humans.

	WHO grade	Tumor histology	Subtype by histology	Subtype/molecular signature
Low-grade	I	Benign, well circumscribed	Pilocytic astrocytoma
	II	Diffuse infiltration, low proliferation	Diffuse astrocytomaOligodendrogliomaOligoastrocytoma
High-grade	III	Highly infiltrative, high proliferation	Anaplastic astrocytomaAnaplastic oligoastrocytomaAnaplastic oligodendroglioma	Proneural/PDGF
	IV	Highly infiltrative, high vascular and cellular proliferation, necrosis	Glioblastoma multiforme	Proneural/PDGF; Mesenchymal/Nf1; Classical/EGFR

Histological features and subclasses of WHO grades I through IV gliomas and corresponding molecular subclasses. Grade IV gliomas subdivide into three distinct subclasses (Brennan et al. 2009; Verhaak et al. 2010; Phillips et al. 2006).

Molecular Pathology

Although most gliomas are sporadic, genetics plays some role in gliomagenesis. Patients with a family history of glioma can have a 5% chance of developing a glial tumor (Wen and Kesari 2008; Yanhong, Lixin, and Huaiyin 2010). Some of the familial cases have already been linked to rare inherited genetic abnormalities (i.e., neurofibromatosis [involving mutations in NF1], Turcot syndrome [abnormalities in mismatch repair genes], Li-Fraumeni syndrome [involving mutations in TP53], and melanoma-brain tumor [P16/CDKN2A mutations]), but for most the etiology remains unknown (Yanhong, Lixin, and Huaiyin 2010).

Most of the molecular alterations that have been shown to be causal in mouse models of glioma have successfully predicted the genetics in the human disease. Here, the focus will primarily be on the molecular alterations that have been identified through the use of a variety of genetically engineered glioma mouse models (Table 2 ). One example is the functional loss of the tumor suppressors RB and P53. These proteins function as G1-S phase cell cycle regulators and are common targets of inactivating mutations in malignant gliomas. The RB pathway can be targeted through the direct inactivation of RB by gene mutation or deletion (loss of 13q) or through functional inactivation by amplification of its negative regulators CDK4 (12q13-14) or, less commonly, CDK6. Inactivation of p16^Ink4a prevents negative regulation of CDK4 and CDK6 and thus represents another mechanism of RB loss of function.

Table 2.

Molecular subclasses of glial tumors and relevant genetically engineered mouse models (GEMMs).

Molecular subclass	Observed genomic aberration	Other common genomic aberrations observed in humans	Relevant GEMMs
Classical/EGFR	EGFR amplification or mutation	INK4a/ARF loss; Notch pathway activation; RB pathway activation	Localized expression of EGFRvIII in GEMMs with tumor suppressor loss
Proneural/PDGF	PDGF amplification or mutation	IDH1/2 mutations, MET, OLIG2, and PDGFRA overexpression	Localized expression of PDGFB in GEMMs with tumor suppressor loss
Mesenchymal/NF1	NF1 deletion or mutation	YKL40 and CH13L1 overexpression	GEMMs with localized Nf1 and Trp53 deletion

Molecular subclasses based on the stage of neurogenesis and defining lesion. The subclasses were identified in studies that analyzed molecular aberrations of WHO grades II, III, and IV human primary and secondary gliomas (Brennan et al. 2009; Verhaak et al. 2010; Phillips et al. 2006). Mouse models have been developed that closely mimic the glioma subclasses in humans.

Similarly, TP53 can be inactivated by direct gene mutations (or 17p loss) or by amplification of its negative regulator MDM2. P14^ARF inhibits MDM2 and represents another gene that, when inactivated, can cause P53 functional loss. P16^INK4a and P14^ARF are both encoded at the CDKN2A locus, and hypermethylation or deletion of this locus can affect both RB and TP53 expression and lead to glioma formation. Interestingly, genome-wide association studies have found that single nucleotide polymorphisms in CDKN2A and CDKN2B are associated with the risk for developing gliomas (Shete et al. 2009; Wrensch et al. 2009).

Phosphatidylinositol 3-kinase (P13K) and mitogen-activated protein kinase (MAPK) signaling pathways are important to the promotion of cell growth and survival. PTEN (phosphotase and tensin homolog) is a tumor suppressor that functions as one of the primary negative regulators of P13K-AKT-mTOR and Ras-MAPK signaling (Knobbe and Reifenberger 2003; Ohgaki et al. 2004). Analyses of the Cancer Genome Atlas (TCGA) Research Network data found that PTEN functional loss in human GBM is commonly due to monoallelic gene mutations or deletions (chromosome 10q deletion). Functional loss of PTEN has also been correlated with overexpression of miR-26a (a microRNA negative regulator of PTEN, RB, and MAP3K2) (Huse and Holland 2009). PTEN is inactivated in 50% of malignant gliomas, and loss of function is considered as a negative prognostic factor (Chakravarti et al. 2004; Tada et al. 2001).

Mouse models of glioma have been developed by overexpressing components of signal transduction pathways that promote cell proliferation and survival (Skeen et al. 2006). AKT and KRAS are protein kinases and downstream effectors of the P13K pathway and MAPK pathways, respectively. However, mutated isoforms of AKT and KRAS resulting in constitutive activation are only observed in a small percentage of glioblastomas. Nonetheless, a convincing argument can be made for the relevance of these two tumor promoters, especially when considering the ability for these lesions in generating glioma mouse models.

Receptor tyrosine kinases (i.e., epidermal growth factor receptor [EGFR] and platelet derived growth factor [PDGFR]) and their ligands (epidermal growth factor [EGF] and platelet-derived growth factor [PDGF]) can activate the P13K-AKT-mTOR and Ras-MAPK pathways. The EGF/EGFR and PDGF/PDGFR growth factor axes have been suggested to stimulate autocrine and paracrine loops when both the ligand and the receptor are overexpressed; however, mutation or genomic amplification of the receptor tyrosine kinases alone can lead to constitutive activation of the receptor. The most common EGF mutant receptor (EGFRvIII; exon 2-7 deletion) is observed in 20% to 30% of glioblastomas (Frederick et al. 2000). In younger patients, the presence of EGFRvIII equates to a less favorable prognosis (Brennan et al. 2009). Increased expression of PDGFRα and its ligands (PDGF-A and PDGF-B) are commonly observed in both low- and high-grade tumors. However, amplifying mutations are less common as compared to EGFR (TCGA Research Network 2008; Parsons et al. 2008).

Genes important to metabolic processes have also been linked to gliomagenesis. Mutations in isocitrate dehydrogenase 1 (IDH1) were shown to be present in a large percentage (~80%) of human grade II and III astrocytic and oligodendroglial tumors as well as secondary glioblastomas (Dang et al. 2009). The IDH1 mutation results in a gain-of-function of enzymatic activity to convert α-ketoglutarate to R(-)-2-hydroxyglutartate (2HG). 2HG is an oncogenic metabolite associated with an increased risk of malignant brain tumors in patients with an inborn defect in 2HG metabolism (Kolker et al. 2002; Wajner et al. 2004; Aghili, Zahedi, and Rafiee 2009). It has also been found at increased levels in human malignant gliomas that harbor IDH1 mutations and thus has been suggested as a contributor to glioma formation and malignant progression (Dang et al. 2009). Another molecular alteration that has been observed in 60% to 90% of oligodendroglial tumors is deletion of the 1p and 19q chromosome arms (Wen and Kesari 2008). This deletion has been found to predict a less aggressive tumor behavior and a more favorable response to chemotherapy (Wen and Kesari 2008). It is worth mentioning that the identity of the presumed tumor suppressors located within the 1p and 19q loci remains unknown. This has presented a challenge in the development of mouse models that recapitulate this disease in humans.

Molecular Subclasses of Glioma

Several expression studies of malignant gliomas have been conducted and have found that patterns of somatic mutations, copy number alterations, and gene and protein expression can be used to categorize these tumors into subclasses (Verhaak et al. 2010; Brennan et al. 2009; Phillips et al. 2006). These studies have linked the molecular signatures of gliomas with stages of neuroglial development (Phillips et al. 2006). In agreement with previous studies (Phillips et al. 2006; Verhaak et al. 2010), Brennan et al. (2009) found that GBMs could be categorized into subclasses based on the levels of expression and activity of core proteins of signal transduction pathways (i.e. PDGF, EGFR, and NF1). Interestingly, the molecular signatures of these subclasses overlap with the classical, proneural, and mesenchymal transcriptomal glioma subtypes (Verhaak et al. 2010; Phillips et al. 2006).

The proneural subclass is primarily defined by PDGF expression. Other genes that are elevated in this subclass include genes important to oligodendrocyte development such as OLIG2 and PDGFRA (Brennan et al. 2009). This subclass consisted primarily of WHO grade III tumors and some GBMs occurring in younger patients (Brennan et al. 2009). The classical subclass most commonly contains molecular alterations in EGFR. INK4a/ARF loss as well as RB and Notch pathway activation is also observed more commonly in these tumors. And finally, the mesenchymal subclass is defined by NF1 alterations as well as increased expression of mesenchymal genes such as YKL40 and CH13L1 (Phillips et al. 2006; Verhaak et al. 2010). Interestingly, these subclasses were shown to be more effective at predicting clinical outcome than was histological classification (Phillips et al. 2006). The heterogeneity between tumors, as well as within individual tumors, makes understanding the molecular aspects of tumorigenesis essential to finding effective therapies.

Mouse Models of Malignant Glioma

The heterogeneous nature of glial tumors dictates the need for creating a variety of genetically engineered mouse models (GEMMs). Several models have been developed over the years, each representing a variety of molecular defects and morphological characteristics as seen in the human disease (Tables 2 and 3 ). Discussed below are mouse models of malignant gliomas that represent the classical, proneural, and mesenchymal subclasses. These models are believed to be close recapitulations of the human disease because they represent a combination of distinct genetic events, some of which are introduced as somatically acquired events, similar to the presumed evolution of malignant gliomas. To date, these models have been essential to increasing our understanding of glioma biology and play an important role in preclinical testing of novel therapies both now and in the future.

Table 3.

Mouse models of malignant gliomas (astrocytomas, glioblastomas, and oligoastrocytomas).

Tumor subtype	Molecular alteration	Model	Tumor histology	Latency (weeks)	Penetrance (%)	Ref
Classical	V¹²Ras gain; EGFRvIII gain	Conventional; adeno-EGFRvIII	High-grade gliomas	8	100	Wei, Clarke, and Scheidenhelm (2006)
	EGFR gain; Ink4a/Arf loss; PTEN loss	Adeno-Cre conditional; Conventional; Adeno-Cre conditional	Glioblastoma	6-8	100	H. Zhu et al. (2009)
Proneural	PDGF-B gain	MoMuLV	Glioblastoma	14-29	40	Uhrbom et al. (1998)
	PDGF-B gain	RCAS	Oligodendroglioma; Oligoastrocytoma	12	55-100^a	Dai et al. (2001); Shih et al. (2004)
	PDGF-B gain; Ink4a/Arf loss	RCAS; Conventional	Glioblastoma	8	100	Hambardzumyan et al. (2009)
	PDGF-B gain; Ink4a/Arf loss; PTEN loss	RCAS; Conventional; RCAS-Cre conditional	Glioblastoma; Oligoastrocytoma	4	100	Hambardzumyan et al. (2009)
Mesenchymal	p53 loss; Nf1loss	Cre- conditional; Conventional	Anaplastic astrocytoma; Glioblastoma	20-40	100	Y. Zhu et al. (2005)
	p53 loss; Nf1 loss; PTEN loss	Cre- conditional; Conventional	Glioblastoma	10-20	100	Kwon et al. (2008)
	P53 loss; Nf1 loss	Adeno-Cre conditional; Adeno-Cre conditional	Anaplastic astrocytoma; Glioblastoma	16, 46^b	100	Alcantara et al. (2009)
Other	Kras gain; PTEN loss	RCAS; RCAS-Cre conditional	Glioblastoma	12	62	Hu et al. (2005)
	Kras gain; Akt gain	RCAS	Glioblastoma; Glioblastoma variants	12	21	Uhrbom et al. (2002)
	Kras gain; Akt gain; Ink4a/Arf loss	RCAS; Conventional	Glioblastoma	12	49	Uhrbom et al. (2002)
	V¹²Ras gain; PTEN loss	Conventional; Adeno-Cre conditional	Anaplastic astrocytoma	4	100	Wei, Clarke, and Scheidenhelm (2006)

Mouse models of malignant gliomas grouped by three molecular subclasses: Classical, Proneural, and Mesenchymal.

^a Tumor incidence depended on PDGF expression. Higher PDGF expression resulted in higher tumor incidence.

^b Tumor latency depended on the mouse age and location of adenoviral Cre injections.

Models of “Classical” Glial Tumors

Several mouse models of glioma have been developed that harbor aberrations in receptor tyrosine kinases (i.e., EGFR, PDGFR) or their ligands. The classical molecular subclass is defined by EGFR overexpression and represents 25% to 30% of malignant glial tumors in humans. The most commonly observed EGFR variant is EGFRvIII (constitutively active variant of EGFR). GEMMs have been created to evaluate the role of EGFRvIII in gliomagenesis (Table 3). Although aberrant expression of EGFRvIII in glial cells alone does not generate gliomas in mice, when combined with other oncogenic genetic lesions, high-grade glial tumors form (Xiao et al. 2005; Wei, Clarke, and Scheidenhelm 2006). For example, when glioma-prone GFAP:V¹²-Ha-Ras adult mice are injected with adeno-EGFRvIII viral vector (Ad:EGFRvIII) malignant gliomas form within 8 weeks post injection (Wei, Clarke, and Scheidenhelm 2006). Another model that investigates the role of EGFR overexpression in adult gliomagenesis uses the Cre-Lox system to somatically introduce EGFRvIII expression by injecting adeno-Cre viral vector into the brain parenchyma. Ink4a/Arf null;Pten ^fl°^xed adult mice harboring conditional EGFRvIII alleles (expressed upon Cre-mediated deletion of a floxed stop cassette) develop glioblastomas with high penetrance within 6 to 8 weeks post orthotopic injections of adeno-Cre into the brain (H. Zhu et al. 2009). Hence, EGFR expression combined with the loss of tumor suppressors is required for glioma formation, and the loss of multiple tumor suppressors shortens tumor latency (H. Zhu et al. 2009; Wei, Clarke, and Scheidenhelm 2006).

Models of “Proneural” Glial Tumors

The proneural subclass is defined by a subset of malignant gliomas with evidence of PDGF pathway activation (30% of human tumors) (Phillips et al. 2006; Brennan et al. 2009). GEMMs have been developed to ascertain what role PDGF and its receptors play in the development of glial tumors (Table 3). Uhrbom et al. (1998) developed a pediatric model of malignant glioma using a murine retrovirus, MoMuLV, to deliver PDGF-B via orthotopic injection into the frontal lobes of newborn mice. This model results in glioblastoma formation with a 40% penetrance after 14 to 29 weeks of age. Resulting tumors coexpressed PDGFRα supporting the concept of an autocrine mechanism of transformation involved in gliomagenesis.

The RCAS/tva system has been used to somatically deliver PDGF-B into glial cells. The RCAS/tva model system allows for somatic gene transfer of the oncogene of interest into targeted brain cells by driving expression of TVA (the RCAS receptor) under an astrocytic promoter, GFAP (G-tva), or a astrocyte progenitor promoter, Nestin (N-tva). This system was used to create another model of pediatric PDGF-induced glioma by orthotopically injecting RCAS-PDGF-B into newborn N-tva or G-tva mice (Dai et al. 2001). This model produces oligodendroglial or mixed oligoastrocytic tumors with a higher penetrance and shorter tumor latency as compared to PDGF-B delivery by MoMuLV (Table 3). Additionally, it was determined that higher gene dose correlated with higher tumor grade (Shih et al. 2004).

More recently, the RCAS/tva system has been used to generate adult models of PDGF-B induced glioma. Transgenic N-tva;Ink4a/Arf null;Pten ^fl°^xed mice developed glioblastomas within 8 weeks post injection of the RCAS-PDGFB viral vector with 100% penetrance (Hambardzumyan et al. 2009). Not surprisingly, when the RCAS-PDGFB vector and RCAS-Cre vectors (conditional deletion of PTEN) were injected, malignant glial tumors formed with equal penetrance and shorter latency (4 weeks). Transgenic G-tva;Ink4a/Arf null, G-tva;Trp53, and G-tva;Arf null mice all develop glioblastomas with regions of oliogastrocytoma and mixed astrocytoma histology.

Models of “Mesenchymal” Glial Tumors

The mesenchymal subclass is defined by malignant glial tumors harboring Nf1 alterations, which is generally characterized by low NF1 expression (Verhaak et al. 2010; Brennan et al. 2009). A mouse model of this subclass was generated by crossing mice bearing a germ-line Trp53 mutation with Nf1 floxed alleles with GFAP-Cre mice (Table 3) (Y. Zhu et al. 2005). This model generated tumors with histological features consistent with high-grade gliomas such as microvascular proliferation and pseudopalisading necrosis with a 100% penetrance and a 10-month tumor latency. Another model was developed to assess the role of Akt activation in high-grade glioma formation using the Nf1/Trp53 model (Kwon et al. 2008). Not surprisingly, when PTEN haploinsufficiency was added to this model, tumor latency decreased and tumor aggressiveness increased (Table 3). However, these models of spontaneous tumor development may be more reminiscent of an inherited tumor syndrome (i.e., neurofibromatosis) since the molecular alterations were generated as germ-line mutations. A somatic cell version of the mesenchymal model for gliomas has also been generated using adenoviral CRE delivery to mice harboring floxed versions of Nf1 and p53 (Alcantara et al. 2009 ). These mice develop gliomas very closely mimicking human glioma histology. Of note, these tumors arise only if the virus is delivered to the subventricular zone (SVZ), implying that at least in this case the cell of origin is limited to CNS stem cells in this region (Llaguno et al. 2009).

Other GEMMs of Gliomas

There are other models of glioma tumors that do not specifically fall into one of the three aforementioned subclasses or that may be less representative of the accumulation of somatic mutations that occur during gliomagenesis. However, models that utilize downstream effectors of the EGFR and PDGFR signaling pathways such as constitutive expression of AKT and KRAS are still useful tools for understanding glioma biology. For example, in a pediatric model of glioma combining Akt and Kras expression generates malignant gliomas (Uhrbom et al. 2002). In this model, newborn N-tva mice injected with RCAS-Kras and RCAS-Akt viral vectors develop malignant glial tumors with a 21% tumor incidence within 12 weeks post injection (Uhrbom et al. 2002). However, when N-tva;Ink4a/Arf null mice are injected with RCAS-Kras and RCAS-Akt viral vectors, glioblastomas form at a higher tumor penetrance (Table 3) (Uhrbom et al. 2002). Glioblastomas also formed in G-tva;Ink4a/Arf null mice (incidence 42%) with histological hallmarks of the glioblastoma variant, giant cell glioblastoma.

A similar model was developed to investigate the role of Akt downstream effectors using Ntv-a;Pten ^fl°^xed mice. When Ntv-a;Pten ^fl°^xed mice are injected with RCAS-Kras and RCAS-Cre viral vectors, glioblastomas develop with greater penetrance (62%) at 12 weeks of age (Hu et al. 2005) as compared to the N-tva;Ink4a/Arf null mice (Table 3). Finally, an adult mouse model of malignant glioma was generated by using double transgenic mice bearing constitutively expressed V¹²-H-Ras and floxed Pten alleles. This model resulted in malignant astrocytoma formation within 4 weeks post injection of the adenoviral-Cre viral vector (Ad:GFAP-Cre) (Wei, Clarke, and Scheidenhelm 2006).

Brain Cancer Stem Cells and Nitric Oxide–Induced Notch Signaling

There has been a long-standing debate over the role of stem cells in brain cancer. It has been proposed that carcinogenesis is dependent on a small population of cells termed “cancer stem cells” (CSCs) that are pluripotent and capable of self-renewal (Ignatova et al. 2002; Shen et al. 2004 ). Normal neural stem cells are located in regions of the brain (subventricular and hippocampal subgranular zones) that have capillary networks capable of providing the necessary nourishment and signaling factors required for stem cell self-renewal and survival (Doetsch 2003; Shen et al. 2004, 2008; Ramirez-Catillejo et al. 2006). Evidence has suggested that brain CSCs require a similar specialized microenvironment to maintain their “stemness” as well (Calabrese et al. 2007; Gilbertson and Rich 2007). Brain CSCs migrate toward this specialized microenvironment, which is closely positioned near areas of microvasculature proliferation referred to as the “perivascular niche” (Calabrese et al. 2007; Gilbertson and Rich 2007; Riquelme, Drapeau, and Doetsch 2008). The pathological designation of the abnormal vasculature in gliomas that correspond to the perivascular niche is “microvascular proliferation”; however, these structures are composed of many cell types including endothelial cells, reactive astrocytes, microglia, smooth muscle-actin expressing pericytes, and tumor cells. There is evidence that these structures correlate with regions of blood-brain barrier breakdown and contrast enhancement on MRI scans.

Brain CSCs have been shown to express multidrug resistance proteins (e.g., ABCG2) and to be extremely resistant to conventional genotoxic therapy (Bleau et al. 2009; B. Wang et al. 2010). In fact, treatment with temozolomide in culture increases the CSC rich side population (Bleau et al. 2009), ultimately resulting in more aggressive and treatment resistant tumors upon reoccurrence. In radiation-treated medulloblastoma, it was shown that the surviving cells were located in the perivascular niche and express stem cell markers such as Nestin (Hambardzumyan et al. 2008). The perivascular niche contains a diverse group of cell types and signaling factors that are required for sustaining the cancer cells' stem-like phenotype. Identifying signaling factors important to the maintenance of the stem cell phenotype and deciphering how they contribute to gliomagenesis is important to understanding the biology of brain CSC and the development of targeted therapies against them.

Nitric Oxide–Induced Notch Signaling

The Notch pathway is important to cell differentiation and has been linked to both normal stem cells and glioma CSC biology (Baron 2003; Z. Wang et al. 2008). It has also been shown to activate the Nestin promoter in gliomas (Shih and Holland 2006). Notch receptor-ligand binding results in the release of the Notch intracellular domain (NICD), which is subsequently translocated to the nucleus where it activates target genes. In vivo studies of NICD expression in mice showed an induction of Nestin expression and an expansion of the SVZ via cooperation with Kras (Shih and Holland 2006). Additionally, Notch signaling can also influence chemoresistance by increasing drug efflux through ABCG2, a drug efflux transporter (Bhattacharya et al. 2007; Fan et al. 2006). To further support the role of the Notch pathway in brain CSC biology, when NICD is overexpressed the ratio of stem-like cells (side population) to bulk tumor cells (main population) increases, and the opposite effect is observed when the Notch pathway is inhibited (Fan et al. 2006; Bleau et al. 2009).

Endothelial nitric oxide synthase (eNOS), a member of the nitric oxide synthase family, is responsible for producing nitric oxide (NO) in the vascular endothelium (Fukumura, Kashiwagi, and Jain 2006). NO is responsible for regulating physiological processes through the NO/cGMP/Protein kinase G pathway (Fukumura, Kashiwagi, and Jain 2006). Elevated levels of eNOS expression have been observed in several cancers including pancreatic cancer and gliomas (Bakshi et al. 1998; Broholm et al. 2003; Fukumura, Kashiwagi, and Jain 2006; Lim et al. 2008). Recently, Charles et al. (2010) conducted a study to test the hypothesis that NO in endothelial cells functions to activate signaling pathways that promote a stem-cell-like phenotype in glioma cells located in the perivascular niche. They found that when human glioma cultured cells were exposed to the NO donor GSNO (S-nitrosoglutathione), the Notch pathway was activated, while other pathways important to the stem cell phenotype (i.e., Wnt and Shh) were not. Additionally, they observed a 2-fold increase in Nestin expression as compared to controls, suggesting NO can indeed activate the Notch pathway and promote stemness as measured by side population analysis.

Because the RCAS/tva PDGF-induced mouse model of glioma closely recapitulates the microenvironment of the perivascular niche and the histological character of human gliomas (Holland 2004), this model was used to ascertain the role of eNOS and NO in gliomagenesis. It was demonstrated that increased eNOS expression was specifically localized to the tumor vascular endothelium; and cells coexpressing Nestin (a stem cell marker) and NICD were in close proximity to the tumor endothelial cells (Charles et al. 2010). The NO receptor, soluble guanylate cyclase (sGC), was also found to be expressed in the stem-like cell population. Furthermore, eNOS deficient mice had a decreased tumor initiation and a longer survival as compared to eNOS wild-type mice.

Primary cultures generated from the PDGF-induced gliomas and exposed to GSNO exhibited a dose-dependent increase in NICD, confirming a collaboration between NO and Notch signaling in these tumors (Charles et al. 2010). Additionally, GSNO treatment resulted in a 2- to 5-fold increase in the stem cell side population and increased the expression of the chemoresistance gene, ABCG2. These data support the importance of NO signaling and notch pathway activation in the maintenance of the CSC phenotype in the perivascular niche during gliomagenesis.

Discussion

The WHO classification of tumors of the central nervous system has been used for decades as the gold standard for categorizing brain tumors. The most recent revisions to the WHO classification system included changes in tumor grades and the addition of new tumor variants and entities. However, the benefit of using molecular signatures over histological characterization is beginning to be realized. Although tumor histological features have historically been used as the primary consideration for therapy selection and clinical course prediction, more recently molecular profiling has resulted in a more effective subgrouping of these tumors and has shed more light on the molecular pathways involved in gliomagenesis.

Genomics has proven to be extremely useful to understanding glioma biology. It is now well understood that the molecular signatures of adult and pediatric malignant gliomas are distinctly different, even though histologically these tumors are indistinguishable. Molecular subclasses have recently been identified that stratify malignant gliomas based on expression levels of three core proteins of signaling pathways: PDGF, EGFR, and NF1. These core proteins define subclasses that represent stages of neurogenesis (proneural, classical, and mesenchymal, respectively). Interestingly, these molecularly defined subclasses have proven to be better prognostic factors than the traditional histological groups. This further supports the importance of utilizing molecular signatures to better stratify patients to ensure the most optimal treatment approach is employed.

In vivo models of cancer are essential for addressing complex questions that require a disease-host interaction to more accurately mimic the disease in humans. Several GEMMs of malignant glioma have been developed that harbor distinct molecular defects that recapitulate the biological environment of human brain tumors. Models of pediatric and adult malignant gliomas provide a tool for investigating the molecular basis for the distinct genomic differences observed between these two groups, even though they can be histologically indistinguishable. Importantly, GEMMs have been created that reflect the molecular signatures of the classical, proneural, and mesenchymal molecular subclasses (Table 3). Genetic manipulation of murine models has allowed researchers to identify genes important to disease predisposition and drug response, as well as the discovery of novel therapeutic targets. Additionally, these models have provided a tool for investigating the role of CSCs in gliomagenesis.

It has been proposed that CSCs are the cells of origin for tumorigenesis. Studies have found that CSCs are resistant to genotoxic therapy and that treatment increases the CSC population. Mouse models of cancer have aided researchers in understanding the biology of CSC. The CSC perivascular niche is a microvascular-rich microenvironment that promotes the CSC character. Identifying the diverse cell population and signaling factors involved in maintaining the CSC phenotype within this niche can help identify novel therapeutic targets. For example, a recent study found that eNOS is responsible for NO production within the perivascular niche, and elevated levels of NO activates the Notch pathway. Additionally, NO was found to promote stem cell character and increase the stem cell population within the perivascular niche. Indeed, the Notch pathway has already shown some promise in brain tumor preclinical studies (Fan et al. 2006). However, whether the Notch signaling pathway will be a fruitful target for drug development will depend on the CSC population’s response in preclinical studies.

References

Aghili

Zahedi

Rafiee

(2009). Hydroxyglutaric aciduria and malignant brain tumor: A case report and literature review. J Neurooncol 91, 233–36.

Alcantara

L. S.

Chen

Kwon

C. H.

Jackson

E. L.

Burns

D. K.

Alvarez-Buylla

Parada

L. F.

(2009). Malignant astrocytomas originate from neural stem/progenitor cells in a somatic tumor suppressor mouse model. Cancer Cell 15, 45–56.

Bakshi

Nag

T. C.

Wadhwa

Mahapatra

A. K.

Sarkar

(1998). The expression of nitric oxide synthases in human brain tumours and peritumoral areas. J Neurol Sci 155, 196–203.

Baron

(2003). An overview of the Notch signalling pathway. Semin Cell Dev Biol 14, 113–19.

Bhattacharya

Das

Mallya

Ahmad

(2007). Maintenance of retinal stem cells by Abcg2 is regulated by notch signaling. J Cell Sci 120, 2652–62.

Bleau

A. M.

Hambardzumyan

Ozawa

Fomchenko

E. I.

Huse

J. T.

Brennan

C. W.

Holland

E. C.

(2009). PTEN/PI3K/Akt pathway regulates the side population phenotype and ABCG2 activity in glioma tumor stem-like cells. Cell Stem Cell 4, 226–35.

Brat

D. J.

Parisi

J. E.

Kleinschmidt-DeMasters

B. K.

Yachnis

A. T.

Montine

T. J.

Boyer

P. J.

Powell

S. Z.

Prayson

R. A.

McLendon

R. E.

(2008). Surgical neuropathology update: A review of changes introduced by the WHO classification of tumours of the central nervous system, 4th edition. Arch Pathol Lab Med 132, 993–1007.

Brennan

Momota

Hambardzumyan

Ozawa

Tandon

Pedraza

Holland

(2009). Glioblastoma subclasses can be defined by activity among signal transduction pathways and associated genomic alterations. PLoS One 4, e7752.

Broholm

Rubin

Kruse

Braendstrup

Schmidt

Skriver

E. B.

Lauritzen

(2003). Nitric oxide synthase expression and enzymatic activity in human brain tumors. Clin Neuropathol 22, 273–81.

10.

Calabrese

Poppleton

Kocak

Hogg

T. L.

Fuller

Hamner

E. Y.

Gaber

M. W.

Finklestein

Allen

Frank

Bayazitov

I. T.

Zakharenko

S. S.

Gajjar

Davidoff

Gilbertson

R. J.

(2007). A perivascular niche for brain tumor stem cells. Cancer Cell 11, 69–82.

11.

Cancer Genome Atlas Research Network (2008). Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 455, 1061–68.

12.

Central Brain Tumor Registry of the United States (2010). CBTRUS 2010 Statistical Report: Primary Brain Tumors in the United States. Central Brain Tumor Registry of the United States, 2004-2006. http://www.cbtrus.org/2010-NPCR-SEER/CBTRUS-WEBREPORT-Final-3-2-10.pdf (accessed June 5, 2010).

13.

Chakravarti

Zhai

Suzuki

Sarkesh

Black

P. M.

Muzikansky

Loeffler

J. S.

(2004). The prognostic significance of phosphatidylinositol 3-kinase pathway activation in human gliomas. J Clin Oncol 22, 1926–33.

14.

Charles

Ozawa

Squatrito

Bleau

A. M.

Brennan

C. W.

Hambardzumyan

Holland

E. C.

(2010). Perivascular nitric oxide activates notch signaling and promotes stem-like character in PDGF-induced glioma cells. Cell Stem Cell 6, 141–52.

15.

Dai

Celestino

J. C.

Okada

Louis

D. N.

Fuller

G. N.

Holland

E. C.

(2001). PDGF autocrine stimulation dedifferentiates cultured astrocytes and induces oligodendrogliomas and oligoastrocytomas from neural progenitors and astrocytes in vivo. Genes Dev 15, 1913–25.

16.

Dang

White

D. W.

Gross

Bennett

B. D.

Bittinger

M. A.

Driggers

E. M.

Fantin

V. R.

Jang

H. G.

Jin

Keenan

M. C.

Marks

K. M.

Prins

R. M.

Ward

P. S.

Yen

K. E.

Liau

L. M.

Rabinowitz

J. D.

Cantley

L. C.

Thompson

C. B.

Vander

Heiden

M. G.

S. M.

(2009). Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739–44.

17.

Doetsch

(2003). A niche for adult neural stem cells. Curr Opin Genet Dev 13, 543–50.

18.

Fan

Matsui

Khaki

Stearns

Chun

Y. M.

Eberhart

C. G.

(2006). Notch pathway inhibition depletes stem-like cells and blocks engraftment in embryonal brain tumors. Cancer Res 66, 7445–52.

19.

Frederick

Wang

X. Y.

Eley

James

C. D.

(2000). Diversity and frequency of epidermal growth factor receptor mutations in human glioblastomas. Cancer Res 60, 1383–87.

20.

Fukumura

Kashiwagi

Jain

R. K.

(2006). The role of nitric oxide in tumour progression. Nat Rev Cancer 6, 521–34.

21.

Furnari

F. B.

Fenton

Bachoo

R. M.

Mukasa

Stommel

J. M.

Stegh

Hahn

W. C.

Ligon

K. L.

Louis

D. N.

Brennan

Chin

DePinho

R. A.

Cavenee

W. K.

(2007). Malignant astrocytic glioma: Genetics, biology, and paths to treatment. Genes Dev 21, 2683–2710.

22.

Gilbertson

R. J.

Rich

J. N.

(2007). Making a tumour’s bed: glioblastoma stem cells and the vascular niche. Nat Rev Cancer 7, 733–36.

23.

Hambardzumyan

Amankulor

N. M.

Helmy

K. Y.

Becher

O. J.

Holland

E. C.

(2009). Modeling adult gliomas using RCAS/t-va technology. Transl Oncol 2, 89–95.

24.

Hambardzumyan

Becher

O. J.

Rosenblum

M. K.

Pandolfi

P. P.

Manova-Todorova

Holland

E. C.

(2008). PI3K pathway regulates survival of cancer stem cells residing in the perivascular niche following radiation in medulloblastoma in vivo. Genes Dev 22, 436–48.

25.

Holland

E. C.

(2004). Mouse models of human cancer as tools in drug development. Cancer Cell 6, 197–98.

26.

Pandolfi

P. P.

Koutcher

J. A.

Rosenblum

Holland

E. C.

(2005). mTOR promotes survival and astrocytic characteristics induced by Pten/AKT signaling in glioblastoma. Neoplasia 7, 356–68.

27.

Huse

J. T.

Holland

E. C

(2009). Genetically engineered mouse models of brain cancer and the promise of preclinical testing. Brain Pathol 19, 132–43.

28.

Huse

J. T.

Holland

E. C.

(2010). Targeting brain cancer: Advances in the molecular pathology of malignant glioma and medulloblastoma. Nat Rev Cancer 10, 319–31.

29.

Ignatova

T. N.

Kukekov

V. G.

Laywell

E. D.

Suslov

O. N.

Vrionis

F. D.

Steindler

D. A.

(2002). Human cortical glial tumors contain neural stem-like cells expressing astroglial and neuronal markers in vitro. Glia 39, 193–206.

30.

Knobbe

C. B.

Reifenberger

(2003). Genetic alterations and aberrant expression of genes related to the phosphatidyl-inositol-3’-kinase/protein kinase B (Akt) signal transduction pathway in glioblastomas. Brain Pathol 13, 507–18.

31.

Kolker

Pawlak

Ahlemeyer

Okun

J. G.

Hörster

Mayatepek

Krieglstein

Hoffmann

G. F.

Köhr

(2002). NMDA receptor activation and respiratory chain complex V inhibition contribute to neurodegeneration in d-2-hydroxyglutaric aciduria. Eur J Neurosci 16, 21–28.

32.

Kwon

C. H.

Zhao

Chen

Alcantara

Burns

D. K.

Mason

R. P.

Lee

E. Y.

Parada

L. F.

(2008). Pten haploinsufficiency accelerates formation of high-grade astrocytomas. Cancer Res 68, 3286–94.

33.

Lim

K. H.

Ancrile

B. B.

Kashatus

D. F.

Counter

C. M.

(2008). Tumour maintenance is mediated by eNOS. Nature 452, 646–49.

34.

Louis

D. N.

Ohgaki

Wiestler

O. D.

Cavenee

W. K.

Burger

P. C.

Jouvet

Scheithauer

B. W.

Kleihues

(2007). The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114, 97–109.

35.

Merchant

T. E.

Pollack

I. F.

Loeffler

J. S.

(2010). Brain tumors across the age spectrum: Biology, therapy, and late effects. Semin Radiat Oncol 20, 58–66.

36.

Ohgaki

Dessen

Jourde

Horstmann

Nishikawa

Di Patre

P. L.

Burkhard

Schüler

Probst-Hensch

N. M.

Maiorka

P. C.

Baeza

Pisani

Yonekawa

Yasargil

M. G.

Lütolf

U. M.

Kleihues

(2004). Genetic pathways to glioblastoma: A population-based study. Cancer Res 64, 6892–99.

37.

Parsons

D. W.

Jones

Zhang

Lin

J. C.

Leary

R. J.

Angenendt

Mankoo

Carter

Siu

I. M.

Gallia

G. L.

Olivi

McLendon

Rasheed

B. A.

Keir

Nikolskaya

Nikolsky

Busam

D. A.

Tekleab

Diaz

L. A.

Jr. Hartigan

Smith

D. R.

Strausberg

R. L.

Marie

S. K.

Shinjo

S. M.

Yan

Riggins

G. J.

Bigner

D. D.

Karchin

Papadopoulos

Parmigiani

Vogelstein

Velculescu

V. E.

Kinzler

K. W.

(2008). An integrated genomic analysis of human glioblastoma multiforme. Science 321, 1807–12.

38.

Pfister

Hartmann

Korshunov

(2009). Histology and molecular pathology of pediatric brain tumors. J Child Neurol 24, 1375–86.

39.

Phillips

H. S.

Kharbanda

Chen

Forrest

W. F.

Soriano

R. H.

T. D.

Misra

Nigro

J. M.

Colman

Soroceanu

Williams

P. M.

Modrusan

Feuerstein

B. G.

Aldape

(2006). Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 9, 157–73.

40.

Ramirez-Castillejo

Sanchez-Sanchez

Andreu-Agullo

Ferrón

S. R.

Aroca-Aguilar

J. D.

Sánchez

Mira

Escribano

Fariñas

(2006). Pigment epithelium-derived factor is a niche signal for neural stem cell renewal. Nat Neurosci 9, 331–39.

41.

Riquelme

P. A.

Drapeau

Doetsch

(2008). Brain micro-ecologies: Neural stem cell niches in the adult mammalian brain. Philos Trans R Soc Lond B Biol Sci 363, 123–37.

42.

Shen

Goderie

S. K.

Jin

Sun

Abramova

Vincent

Pumiglia

Temple

(2004). Endothelial cells stimulate self-renewal and expand neurogenesis of neural stem cells. Science 304, 1338–40.

43.

Shen

Wang

Kokovay

Lin

Chuang

S. M.

Goderie

S. K.

Roysam

Temple

(2008). Adult SVZ stem cells lie in a vascular niche: A quantitative analysis of niche cell-cell interactions. Cell Stem Cell 3, 289–300.

44.

Shete

Hosking

F. J.

Robertson

L. B.

Dobbins

S. E.

Sanson

Malmer

Simon

Marie

Boisselier

Delattre

J. Y.

Hoang-Xuan

El Hallani

Idbaih

Zelenika

Andersson

Henriksson

Bergenheim

A. T.

Feychting

Lönn

Ahlbom

Schramm

Linnebank

Hemminki

Kumar

Hepworth

S. J.

Price

Armstrong

Liu

Lau

Schoemaker

Muir

Swerdlow

Lathrop

Bondy

Houlston

R. S.

(2009). Genome-wide association study identifies five susceptibility loci for glioma. Nat Genet 41, 899–904.

45.

Shih

A. H.

Dai

Rosenblum

M. K.

Koutcher

J. A.

Holland

E. C.

(2004). Dose-dependent effects of platelet-derived growth factor-B on glial tumorigenesis. Cancer Res 64, 4783–89.

46.

Shih

A. H.

Holland

E. C.

(2006). Notch signaling enhances nestin expression in gliomas. Neoplasia 8, 1072–82.

47.

Skeen

J. E.

Bhaskar

P. T.

Chen

C. C.

Chen

W. S.

Peng

X. D.

Nogueira

Hahn-Windgassen

Kiyokawa

Hay

(2006). Akt deficiency impairs normal cell proliferation and suppresses oncogenesis in a p53-independent and mTORC1-dependent manner. Cancer Cell 10, 269–80.

48.

Tada

Shiraishi

Kamiryo

Nakamura

Hirano

Kuratsu

Kochi

Saya

Ushio

(2001). Analysis of loss of heterozygosity on chromosome 10 in patients with malignant astrocytic tumors: Correlation with patient age and survival. J Neurosurg 95, 651–59.

49.

Uhrbom

Dai

Celestino

J. C.

Fuller

G. N.

Holland

E. C.

(2002). Ink4a-Arf loss cooperates with KRas activation in astrocytes and neural progenitors to generate glioblastomas of various morphologies depending on activated Akt. Cancer Res 62, 5551–58.

50.

Uhrbom

Hesselager

Nister

Westermark

(1998). Induction of brain tumors in mice using a recombinant platelet-derived growth factor B-chain retrovirus. Cancer Res 58, 5275–79.

51.

Verhaak

R. G.

Hoadley

K. A.

Purdom

Wang

Wilkerson

M. D.

Miller

C. R.

Ding

Golub

Mesirov

J. P.

Alexe

Lawrence

O’Kelly

Tamayo

Weir

B. A.

Gabriel

Winckler

Gupta

Jakkula

Feiler

H. S.

Hodgson

J. G.

James

C. D.

Sarkaria

J. N.

Brennan

Kahn

Spellman

P. T.

Wilson

R. K.

Speed

T. P.

Gray

J. W.

Meyerson

Getz

Perou

C. M.

Hayes

D. N.

and Cancer Genome Atlas Research Network (2010). Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17, 98–110.

52.

Wajner

Kolker

Souza

D. O.

Hoffmann

G. F.

de Mello

C. F.

(2004). Modulation of glutamatergic and GABAergic neurotransmission in glutaryl-CoA dehydrogenase deficiency. J Inherit Metab Dis 27, 825–28.

53.

Wang

Yang

Huang

Y. Z.

Yan

R. H.

Liu

F. J.

Zhang

J. N.

(2010). Biologic characteristics of the side population of human small cell lung cancer cell line H446. Chin J Cancer 29, 254–60.

54.

Wang

Banerjee

Sarkar

F. H.

(2008). Exploitation of the Notch signaling pathway as a novel target for cancer therapy. Anticancer Res 28, 3621–30.

55.

Wei

Clarke

Scheidenhelm

D. K.

(2006). High-grade glioma formation results from postnatal pten loss or mutant epidermal growth factor receptor expression in a transgenic mouse glioma model. Cancer Res, 66, 7429–37.

56.

Wen

P. Y.

Kesari

(2008). Malignant gliomas in adults. N Engl J Med 359, 492–507.

57.

Wrensch

Jenkins

R. B.

Chang

J. S.

Yeh

R. F.

Xiao

Decker

P. A.

Ballman

K. V.

Berger

Buckner

J. C.

Chang

Giannini

Halder

Kollmeyer

T. M.

Kosel

M. L.

LaChance

D. H.

McCoy

O’Neill

B. P.

Patoka

Pico

A. R.

Prados

Quesenberry

Rice

Rynearson

A. L.

Smirnov

Tihan

Wiemels

Yang

Wiencke

J. K.

(2009). Variants in the CDKN2B and RTEL1 regions are associated with high-grade glioma susceptibility. Nat Genet 41, 905–8.

58.

Xiao

Yin

Yang

Di Cristofano

Pandolfi

P. P.

Van Dyke

(2005). Somatic induction of Pten loss in a preclinical astrocytoma model reveals major roles in disease progression and avenues for target discovery and validation. Cancer Res 65, 5172–80.

59.

Yanhong

Lixin

Huaiyin

(forthcoming). Perivascular epithelioid cell tumor of the heart in a child. Pediatr Dev Pathol. doi: 10.2350/09-10-0726-CR.1.

60.

Zhu

Acquaviva

Ramachandran

Boskovitz

Woolfenden

Pfannl

Bronson

R. T.

Chen

J. W.

Weissleder

Housman

D. E.

Charest

(2009). Oncogenic EGFR signaling cooperates with loss of tumor suppressor gene functions in gliomagenesis. Proc Natl Acad Sci U S A 106, 2712–16.

61.

Zhu

Guignard

Zhao

Liu

Burns

D. K.

Mason

R. P.

Messing

Parada

L. F.

(2005). Early inactivation of p53 tumor suppressor gene cooperating with NF1 loss induces malignant astrocytoma. Cancer Cell 8, 119–30.

Molecular Pathogenesis of Malignant Glial Tumors

Abstract

Keywords

Introduction/Relevance

Histological and Molecular Pathology of Malignant Gliomas

Histological Pathology

Molecular Pathology

Molecular Subclasses of Glioma

Mouse Models of Malignant Glioma

Models of “Classical” Glial Tumors

Models of “Proneural” Glial Tumors

Models of “Mesenchymal” Glial Tumors

Other GEMMs of Gliomas

Brain Cancer Stem Cells and Nitric Oxide–Induced Notch Signaling

Nitric Oxide–Induced Notch Signaling

Discussion

References