Genomic Abnormalities in Gliomas

In general, studies reveal that cancer arises through genetic and epigenetic alterations that affect specific genes within a given cell type. These changes involve a gain of function when the alterations involve oncogenes, a loss of function when the target genes are tumor suppressor genes. Thus, both genetic and epigenetic changes promote the instability of cellular homeostasis (Bello & Rey, 2006; Richardson, 2003; Sugimura & Ushijima, 2000). This lack of stability reflects the complexity of cancer because the loss of controlled cell growth occurs due to changes in one or more genes. These genetic and epigenetic events are followed by the growing accumulation of changes in hundreds, if not thousands, of genes. Over time, this accumulation causes the tumor to reach its highest degree of malignancy, which usually culminates in metastasis (Bartek & Lukas, 2001). However, in recent years, a subpopulation of tumor cells has been found to display a slow rate of cell division, high tumorigenic potential and characteristics similar to those of normal stem cells. This discovery has changed the concept of metastasis to one associated with strictly terminal states (Stiles & Rowitch, 2008). Hanahan & Weinberg (2000) proposed that all tumor cells must acquire six essential alterations in cell physiology that collectively dictate malignant growth: (1) loss of normal signaling for cell proliferation arrest, (2) loss of signaling for cell differentiation, (3) autocrine signaling for cell division, (4) reduction in apoptosis, (5) ability to enter the basement membrane and other tissues and organs, and (6) induction of angiogenesis. All of these processes involve biochemical pathways that form a complex network of cell signaling. These processes are usually altered in tumors because the genes that compose them are changed, resulting in cell cycle dysregulation. Therefore, the identification and determination of oncogenes and tumor suppressor genes is essential for understanding both cancer biology and clinical applications, such as the identification of therapeutic targets, early detection, and prediction of the disease course. Studies concerning cell cycle control genes have served as a starting point for identifying genes related to tumorigenesis and their biochemical role in various pathways (Paige, 2003).


Introduction
In general, studies reveal that cancer arises through genetic and epigenetic alterations that affect specific genes within a given cell type. These changes involve a gain of function when the alterations involve oncogenes, a loss of function when the target genes are tumor suppressor genes. Thus, both genetic and epigenetic changes promote the instability of cellular homeostasis (Bello & Rey, 2006;Richardson, 2003;Sugimura & Ushijima, 2000). This lack of stability reflects the complexity of cancer because the loss of controlled cell growth occurs due to changes in one or more genes. These genetic and epigenetic events are followed by the growing accumulation of changes in hundreds, if not thousands, of genes. Over time, this accumulation causes the tumor to reach its highest degree of malignancy, which usually culminates in metastasis (Bartek & Lukas, 2001). However, in recent years, a subpopulation of tumor cells has been found to display a slow rate of cell division, high tumorigenic potential and characteristics similar to those of normal stem cells. This discovery has changed the concept of metastasis to one associated with strictly terminal states (Stiles & Rowitch, 2008). Hanahan & Weinberg (2000) proposed that all tumor cells must acquire six essential alterations in cell physiology that collectively dictate malignant growth: (1) loss of normal signaling for cell proliferation arrest, (2) loss of signaling for cell differentiation, (3) autocrine signaling for cell division, (4) reduction in apoptosis, (5) ability to enter the basement membrane and other tissues and organs, and (6) induction of angiogenesis. All of these processes involve biochemical pathways that form a complex network of cell signaling. These processes are usually altered in tumors because the genes that compose them are changed, resulting in cell cycle dysregulation. Therefore, the identification and determination of oncogenes and tumor suppressor genes is essential for understanding both cancer biology and clinical applications, such as the identification of therapeutic targets, early detection, and prediction of the disease course. Studies concerning cell cycle control genes have served as a starting point for identifying genes related to tumorigenesis and their biochemical role in various pathways (Paige, 2003).
Changes in oncogenes can lead to constitutive activation, which involves activation under conditions in which an oncogene would normally be inactive. For this process to occur, a cell needs only one allele of an oncogene to be altered, resulting in a selective growth advantage. In contrast, changes in tumor suppressor genes often reduce the gene product and consequently its activity. For this reason, cells that develop a selective advantage with changes in tumor suppressor genes usually require inactivation of both alleles of the target gene. Conceptually, tumor suppressor genes can be subdivided into two categories: "gatekeepers" that directly inhibit tumor growth and thereby suppress tumor formation and "caretakers" that ensure DNA integrity by repairing damage or preventing genomic instability (Vogelstein & Kinzler, 2004). Studies of hereditary and sporadic forms of tumors, particularly retinoblastoma, culminated in the formulation of Knudson's "two events" model in 1971. In hereditary tumors, the first mutation occurs in one allele in the germline and results in a predisposition to develop tumors. Throughout development, a second change (mutation or loss of heterozygosity) inactivates the other allele and silences the altered gene. In contrast, sporadic tumors acquire the two allelic alterations that lead to gene silencing throughout the organism's development (Knudson, 1971). Over time, the neoplastic transformation and metabolicphenotype of the cell can evolve. Ultimately, these changes result in a cancer in which clonal expansion of modified somatic cells destroys the adjacent normal tissue (Bartek & Lukas, 2001). For many years, research in cancer genetics has prioritized understanding the role of genetic alterations in carcinogenesis. Studies revealed that base deletions, insertions, recombination and amplification in oncogenes and tumor suppressor genes were related to metastasis and invasion. These changes were also closely related to tumorigenesis and tumor progression. For this reason, the scientific community accepted that genetic changes almost exclusively explained the process of carcinogenesis (Sugimura & Ushijima, 2000). However, studies also indicated that embryogenesis and differentiation, which are characterized by specific patterns of gene expression in tissues and organs, can occur without changes in the DNA sequence. This notion has interested the scientific community in potential epigenetic mechanisms of carcinogenesis (Jones & Buckley, 1990;Rush & Plass, 2002). An epigenetic phenomenon is defined as a change in gene function that is heritable through mitosis or meiosis but cannot be explained by changes in the DNA sequence. Aberrant epigenetic mechanisms, such as promoter hypermethylation, histone modifications, or noncoding RNA expression, are known to be important for tumor formation and comprise the "third pathway" in Knudson's model. These mechanisms result in transcriptional repression equivalent to that observed with the mutations and deletions proposed in Knudson's model (Jones & Baylin, 2007). DNA methylation, the main epigenetic modification studied, occurs at cytosine residues in the cytosine-guanine sequences (CpG) of DNA through the action of an enzyme family called DNA methyltransferases (DNMT). In humans, approximately 70% of CpG sites, which are generally located in repetitive DNA sequences, are methylated. Clusters of unmethylated CpG sites are present in the genome as well, and these clusters are referred to as CpG islands. Approximately 60% of genes have CpG islands in the promoter regions and in the first exon. CpG islands are often dimethylated when associated with housekeeping genes. Moreover, CpG islands are tissue specific and are generally methylated except in those tissues where the associated gene is expressed (Cross & Bird, 1995;Gonzalez-Gomez et al., 2003). A recent genome-wide analysis revealed that CpG islands are also found in non-promoter regions. In addition, epigenetic abnormalities causing loss of gene function are more frequent than genetic abnormalities in cancer cells (Schuebel et al., 2007). Thus, cellular epigenetic inheritance mediated by aberrant DNA methylation resulting in gene silencing, gene imprinting, and⁄or activation of cancer-associated genes is now accepted as an important factor defining the transformed phenotype (Natsume et al., 2010)

Tumors of the central nervous system
Tumors of the central nervous system (CNS) are relatively rare and represent approximately 5-9% of all cancers, with an estimated incidence of 4.2 to 5.4 per 100,000 people/year. Moreover, tumors of the CNS carry a very poor prognosis and are associated with considerable morbidity and mortality. They are a leading cause of childhood cancer deaths, the second leading cause of cancer-related death in men aged 20-39, and the fifth leading cause of cancer-related death in women aged 20-39 (Ohgaki & Kleihues, 2005). Although the incidence of CNS tumors is small compared with the incidence of other cancers, CNS tumors are among the most serious human malignancies because they affect the organ responsible for the coordination and integration of all biological activities. Moreover, as each region of the brain has a vital function, therapies used to treat other cancers (e.g., total surgical removal of an organ or tumor with a generous margin of normal tissue) cannot be applied to brain tumors. The inability to use these therapies hinders quality of life and patient survival . In CNS tumors, the histopathological classifications are extensive and based primarily on descriptive morphology. Because the histogenesis of these tumors is unique and heterogeneous, it is difficult to characterize several of the tumor subtypes, which is reflected in the difficulties encountered in tumor diagnosis (Gilbertson, 2002). In contrast with the first World Health Organization (WHO) classifications for CNS tumors (Kleihues et al., 1993;Zülch, 1979), the third edition by Kleihues & Cavenee (2000) incorporated genetic profiles as additional aids in defining brain tumors. The fourth edition of the WHO classifications for CNS tumors, which was published in 2007, lists several new characteristics. The fourth edition is based on consensus from an international working group of 25 pathologists and geneticists, as well as contributions from more than 70 international experts. Currently, this edition is the standard for defining brain tumors for clinical oncology and cancer research communities world-wide (Louis et al., 2007).

Gliomas
Gliomas are the most common tumors of the CNS. However, in spite of marked progress in characterizing the molecular pathogenesis of gliomas, these tumors remain incurable. In most cases, gliomas are also refractory to treatment because of their molecular heterogeneity. Gliomas rarely metastasize outside of the brain but instead infiltrate extensively into the surrounding normal brain. Therefore, surgery is not curative but can establish the diagnosis and relieve symptoms by decompressing the brain, which is located in the rigid intracranial cavity. Radiation therapy and chemotherapy increase survival; however, disease recurrence is frequently inevitable (Park & Rich, 2009). The following four degrees of malignancy are recognized by the WHO: grades I and II (lowgrade), which are biologically less aggressive and grades III and IV (high-grade), which are the most aggressive. The histological criteria for grading malignancies are not uniform for www.intechopen.com all subtypes of gliomas. Thus, all tumors should be classified before the degree of malignancy is determined. This classification is made according to the cell type thought to be responsible for the tumor and based on the characteristics exhibited by astrocytes, oligodendrocytes, ependymal cells, or their neuronal progenitors (Louis et al., 2007). Table 1 shows the heterogeneous WHO classification for gliomas according to the degree of malignancy. Gliomas of astrocytic, oligodendroglial, and ependymal origin account for 80% of CNS tumors. For this reason, some morphological and genetic characteristics of these tumors are discussed below.

Astrocytomas
Astrocytomas represent the vast majority of gliomas and account for 70% of the total gliomas seen in patients. Astrocytomas can be further characterized as pilocytic astrocytomas (WHO grade I) or diffuse astrocytomas, including low-grade astrocytomas (WHO grade II), anaplastic astrocytomas (WHO grade III) and glioblastomas (WHO grade IV) (Kleihues et al., 2002). Pilocytic astrocytomas are more commonly seen in children and carry a good prognosis because of their biology. Patients with neurofibromatosis type 1, a familial syndrome caused by germline mutations in the gene NF1 (neurofibromin 1), have an increased incidence of pilocytic astrocytomas. These tumors are usually not aggressive and stand out among astrocytomas because they have maintained their WHO grade I status for years and even decades, in contrast to diffuse astrocytic tumors (WHO grades II-IV). However, some cases can progress to a higher degree of malignancy, though such a progression is rare (Listernick et al., 1999). More than 100 cases of pilocytic astrocytomas were analyzed by cytogenetics and many others were used for comparative genomic hybridization (CGH); however, the vast majority of the results indicated normal patterns (Bigner et al., 1997;Sanoudou et al., 2000;Zattara-Cannoni et al., 1998). In adults, genetic changes were more frequent but were still rare. The few molecular genetics studies on these tumors indicated allelic loss of both gene loci TP53 (tumor protein p53) and NF1 in regions 17p and 17q, respectively. In sporadic tumors, few mutations were reported in the TP53 locus and none in NF1 (Gutmann et al., 2000;. The relevance of a malignancy-grading scheme based on histopathology is indicated by the correlation with patient survival. Patients with low-grade astrocytomas (WHO grade II) have a median survival of approximately seven years, whereas patients with anaplastic astrocytomas (WHO grade III) have a mean survival of half that time (McCormack et al., 1992). Patients with glioblastomas have a median survival time of 9 to 11 months (Simpson et al., 1993). Unlike pilocytic astrocytomas, diffuse astrocytic tumors are often seen in adults. Low-grade astrocytomas have a peak incidence between 25 and 50 years of age, whereas glioblastomas have a peak incidence between 45 and 50 years (Colins, 2004). Ng & Lam (1998) suggested dividing glioblastomas into two distinct molecular and clinical entities: primary or de novo glioblastomas, which occur in elderly patients and are clinically very aggressive and secondary glioblastomas, which develop from low-grade astrocytomas and have a more prolonged clinical course. Many mechanisms are involved in the initiation and progression of secondary glioblastomas, including the loss of NF1 and TP53 genes and the activation of signal transduction pathways, such as PDGF (platelet-derived growth factor) and its receptor PDGFR (PDGF receptor). These pathways are involved in the induction of low-grade tumors (e.g., pilocytic astrocytomas), which can progress to high-grade tumors (e.g., anaplastic astrocytomas and secondary glioblastoma). This progression is associated with the lack of a functional RB1 (retinoblastoma 1) because of the loss of RB1 or gene amplification/overexpression of CDK4 (cyclin-dependent kinase 4) (Fig. 1a). In primary glioblastomas, the same genetic pathways are disrupted but by different mechanisms. For example, reduction of the TP53 pathway generally occurs through the loss of the gene ARF4 (ADP-ribosylation factor 4) or less frequently through amplification of the gene MDM2 (transformed 3T3 cell double minute 2). The lack of RB1 also occurs via a loss of the gene CDKN2A (cyclin-dependent kinase inhibitor 2A). In primary glioblastomas, amplification and/or mutation of EGFR (epidermal growth factor receptor) and loss of PTEN (phosphatase and tensin homolog) are the most frequently observed genetic defects (  (Zhu & Parada, 2000).
Sequencing of the genome recently identified mutations in the IDH1/IDH2 genes (isocitrate dehydrogenase 1 and 2 genes) that occur in the majority of WHO grade II-III gliomas and secondary glioblastomas Yan et al., 2009), all of which harbor a better prognosis compared with the wild-type cases ). However, pilocytic astrocytomas (WHO grade I) that are potentially curable by complete resection rarely harbor IDH mutations. IDH appears to function as a tumor suppressor when inactivated through mutation, rendering the IDH enzyme unable to catalyze conversion of isocitrate to alphaketoglutarate. This process also induces HIF1-alpha (hypoxia-inducible factor), which triggers the angiogenic process. However, the precise mechanism of its effect on tumor biology remains unclear (Dang et al., 2009). Aberrant activation of the BRAF proto-oncogene (v-raf murine sarcoma viral oncogene homolog B1) at 7q34, which is most commonly caused by gene duplication and fusion or less frequently by point mutation, has only recently been identified as the characteristic genetic aberration in pilocytic astrocytomas. BRAF abnormalities occur in 60-80% of pilocytic astrocytomas but almost never in diffuse, infiltrating astrocytomas . Thus, testing for BRAF gene alterations might be helpful for differentiating during diagnosis between pilocytic astrocytomas and low-grade, diffuse astrocytomas (Korshunov et al., 2009). The importance of silencing DNA repair pathways, especially the DNA-repair enzyme AGAT (O6-alkylguanine DNA alkyltransferase), which is encoded by the gene MGMT (O6methylguanine-DNA-methyltransferase), has been the subject of substantial debate in recent years (Hofer & Lassman, 2010). The MGMT gene is frequently silenced by promoter www.intechopen.com hypermethylation in diffuse gliomas, and this hypermethylation has been pinpointed as an epigenetic mechanism that reduces MGMT expression levels. There are 97 CpG islands in the MGMT promoter, and these CpG islands are further divided into two hypermethylated regions (Nakagawachi et al., 2003). Because of its critical role in DNA repair, the epigenetic silencing of MGMT is associated with an increased number of mutations and with a poorer outcome in glioblastomas. Thus, MGMT silencing is considered to be a biomarker for poor prognosis (Komine et al., 2003). However, an association between MGMT promoter methylation and the response of malignant gliomas to alkylating chemotherapy using nitrosourea compounds, temozolomide, or a combination of both has been observed (Esteller et al., 2000;Herrlinger et al., 2006). Furthermore, Hegi et al. (2005) reported that patients treated with radiotherapy and temozolomide, and whose tumors had a methylated MGMT promoter (which is seen in approximately 40% of primary glioblastomas), survived significantly longer than did patients whose tumors lacked MGMT promoter methylation. Rivera et al. (2010) recently reported that MGMT promoter methylation in anaplastic gliomas (WHO grade III) is also predictive of the response to radiotherapy and linked to longer survival in the absence of adjuvant chemotherapy. The use of temozolomide based on MGMT methylation status highlights the importance of understanding epigenetic changes in glioblastomas for the discovery of novel therapies and prognostic factors for the treatment of this deadly cancer (Komine et al., 2003;Nakagawachi et al., 2003)

Oligodendrogliomas
Oligodendrogliomas represent approximately 10-15% of gliomas, are more common in adults, and can be divided into two histological subtypes: low-grade (WHO grade II) and anaplastic (WHO grade III) (Kleihues et al., 2002). Low-grade oligodendrogliomas are less biologically aggressive than are astrocytic tumors. Therefore, the prognosis is quite favorable and survival beyond 15 years is achieved in up to 90% of cases that receive a complete surgical resection. There is potential for malignancy, but even the aggressive tumors respond well to additional treatments (e.g., radiation and chemotherapy). Anaplastic oligodendrogliomas have a more aggressive course; however, survival is still five to eight years longer than that observed with anaplastic astrocytomas (Reifenberger & Louis, 2003). In 1990, the PCV chemotherapy regimen (procarbazine, carmustine, and vincristine) was shown to result in a dramatic tumor response in oligodendrogliomas. Since that time, the identification of all forms of gliomas with oligodendroglial components became crucial (Macdonald et al., 1990). Importantly, these studies indicated that the prognostic power of oligodendroglial components was independent of whether radiotherapy, chemotherapy or combined radio-chemotherapy was used ). This phenomenon is likely due to oligodendrogliomas exhibiting specific genetic abnormalities that distinguish them from other gliomas. Reifenberger et al. (1994), after a thorough analysis of the genome, reported a loss of genetic information in the 1p and 19q loci in oligodendrogliomas, the so-called chromosome 1p/19q co-deletion. This loss was later linked with a good response to PCV and provided the first molecular indicator of treatment response in brain tumors (Cairncross et al., 1998;Reifenberger et al., 2003). Further studies corroborated these findings, and it is now known that the chromosomal loss results from an unbalanced translocation (Franco-Hernandez et al., 2009;Jenkins et al., 2006). Approximately 85% of low-grade oligodendrogliomas and 65% of anaplastic oligodendrogliomas present with 1p/19q codeletions .
Low-grade oligodendrogliomas and astrocytomas present a loss of ARF4 expression and overexpression of EGFR and PDGF signaling. Malignant progression is associated with additional genetic abnormalities that are similar to those described above for astrocytomas, including a lack of the RB1 pathway, loss of RB1, or gene amplification/overexpression of the CDK4 gene (Franco-Hernandez et al., 2007;Reifenberger & Louis, 2003).

Ependymomas
Ependymomas arise in or near the ependymal surface, and these tumors can occur anywhere in the ventricular system, spinal cord and even occasionally at extraneural sites. The most common location is in the fourth ventricle, followed by the spinal cord, the lateral ventricles and the third ventricle. These tumors are more common in children but can also occur in adults (Ebert et al., 1999). WHO classification identifies four major subtypes of ependymomas: subependymomas (WHO grade I), myxopapillary ependymomas (WHO grade I), low-grade ependymomas (WHO grade II) and anaplastic ependymomas (WHO grade III). Subependymomas are intraventricular in location, while myxopapillary ependymomas are commonly found in the cauda equina. The low-grade ependymomas can be differentiated from their anaplastic counterparts based on the low rate of mitosis and the low level of nuclear polymorphism; however, the distinction between the two tumors remains poorly defined (Kleihues et al., 2002). In ependymomas, chromosomal abnormalities detected by classic cytogenetics and CGH involve chromosomes 1, 6, 7, 9, 10, 13, 17, 19 and 22. Deletions are the most commonly observed changes, and chromosome 22 losses are common in adults (50%) but rare in pediatric ependymomas (Kraus et al., 2001;Lamszus et al., 2001;von Haken et al., 1996). The target genes, located in regions of chromosomal gain or loss, are unknown, with the exception of cases in which both copies of the wild-type NF2 gene (neurofibromin 2) are lost in intramedullary ependymomas (Alonso et al., 2002). Isolated cases of MEN1 gene (multiple endocrine neoplasia I) loss have also been reported (Urioste et al., 2002). Germline mutations in TP53 are uncommon, in contrast with those seen in diffuse astrocytomas (Nozaki et al., 1998). When reviewed together, the data on genetic and epigenetic abnormalities presented above allow us to define four molecular biomarkers: MGMT hypermethylation in glioblastomas and anaplastic gliomas, IDH1 and IDH2 mutations in diffuse gliomas, BRAF aberrations in pilocytic astrocytomas, and combined deletions of chromosome arms 1p and 19q in oligodendroglial tumors. These biomarkers and their clinical significance are summarized in Table 2.

Single nucleotide polymorphisms and gliomas
A single nucleotide polymorphism (SNP) is generally defined as a stable replacement of only one DNA base, with a frequency greater than 1% in at least one population (Taylor et al., 2001). In human genetics studies, SNPs are simply referred to as bi-allelic markers because tri-and tetra-allelic markers are rare (Brookes, 1999). Initially, only a few thousand SNPs were thought to exist in the entire genome. However, since 2000, that number has increased about one thousand-fold. In 2001, an international consortium on mapping SNPs described 1.42 million polymorphic loci. More important than this large number is the precision of their placement in the genome; there is approximately one polymorphism every 1.91 Kb. Therefore, 90% of sequences greater than 20 Kb in length have at least one SNP, and this density can be higher in genic regions. Of the known genes, 93% contain SNPs and 98% are at least 5 Kb away from a SNP. Soon, almost all genes or gene regions will be marked by one of these variable sequences (Sachidanandam et al., 2001).

Molecular marker Clinical significance
IDH1/IDH2 mutation  Diagnostic marker for diffuse WHO grade II and III gliomas, as well as secondary glioblastomas, and associated with a better prognosis in these tumors  Rare in primary glioblastomas, but when present, it is associated with a more favorable outcome  Not predictive for response to a particular type of therapy BRAF duplication/fusion  Diagnostic marker for pilocytic astrocytomas and helpful in distinguishing these tumors from diffuse astrocytomas  Prognostic significance within the group of pilocytic astrocytoma patients is unknown

MGMT promoter methylation
 Predictive for response of glioblastomas to alkylating chemotherapy  Associated with longer survival in glioblastoma patients treated with radiotherapy combined with concurrent and adjuvant temozolomide  Prognostic in anaplastic glioma patients treated with radiotherapy and/or alkylating chemotherapy 1p/19q co-deletion  Associated with improved prognosis in oligodendroglial tumor patients receiving adjuvant radiotherapy and/or chemotherapy  Not predictive for response to a particular type of therapy Table 2. The four most relevant markers for the molecular diagnosis of gliomas (Hofer & Lassman, 2010;Riemenschneider et al., 2010).
Because they are found throughout the genome, some alleles containing SNPs produce functional or physiologically relevant gene products. For example, SNPs in a coding region can affect the coded protein. When located in an intron, SNPs can influence the splicing mechanism, and when located in the promoter, SNPs can alter gene transcription (Krawczak et al., 1992). For this reason, SNPs are recognized as important tools in human genetics and medicine and have been widely used in genetic association studies of various complex diseases, including cancer. In humans, several reviews of SNPs have been carried out in an attempt to determine the patterns of SNP haplotypes in different populations (Conrad et al., 2006;Gonzalez-Neira et al., 2006;Jakobsson et al., 2008;Nothnagel & Rohde, 2005;Salisbury et al., 2003). Data from these tests are extremely useful for studying the genetic basis of cancer. For this reason, several research groups have focused on elucidating the role of SNPs in different genes related to the initiation and progression of gliomas in different populations. We performed association studies between SNPs, the risk of developing gliomas, and the prognosis for gliomas in a Brazilian population. Brazilians form one of the most heterogeneous populations in the world, which is the result of five centuries of interethnic crosses of peoples from three continents: the European colonizers who are mainly represented by the Portuguese, the African slaves, and the autochthonous Amerindians (Parra et al., 2003). Until recently, we were the only laboratory investigating the association between WRN Cys1367Arg, the risk for brain tumor development, and the prognosis of brain tumors, especially with regard to gliomas (Pinto et al., 2008a). Werner syndrome (WS) is a premature aging disorder characterized by early onset of symptoms related to normal aging and is caused by inherited, recessive mutations in the WRN gene. The WRN gene encodes a member of the RecQ family of helicases involved in DNA replication and in maintaining the integrity of the genome (Harrigan et al., 2006). The cells of WS patients exhibit a high level of chromosomal translocations and deletions, and these patients present an increased predisposition to various types of cancer, including CNS tumors (Kobayashi et al., 1980). However, despite its putative tumor suppressor function, little is known about the contribution of the WRN protein to sporadic human malignancies. Taking into account that almost all cancers occur in the elderly and that mutations in the WRN g e n e l e a d t o accelerated aging, it has been suggested that polymorphisms of the WRN gene, similar to Cys1367Arg, might be associated with age-related pathologies and cancer predisposition. However, our data indicate that neither glioma risk (OR = 1.38; 95% CI, 0.78-2.43; P = 0.334) nor patient survival (overall and disease-free survival, P = 0.396 and P = 0.843, respectively) was associated with variant alleles. Similar results were found when we evaluated the genotype distribution of TP53 Pro47Ser and Arg72Pro SNPs for their involvement in susceptibility to gliomas and in determining the oncologic prognosis of patients (Pinto et al., 2008b). A critical site in the TP53 protein for apoptosis signaling is a proline-rich region located between codons 64 and 92. Dumont et al. (2003) reported that the homozygous Arg72 allele induces apoptosis at a rate that is 15-fold higher than the Pro72 allele. According to Leu et al. (2004), the apoptosis-inducing ability of the Arg72 allele is in part due to its mitochondrial location, which makes it possible for TP53 to directly interact with the pro-apoptotic protein, BAK. However, the TP53 Pro47Ser SNP resulted in a significantly decreased ability of the TP53 protein to induce apoptosis. A critical event in TP53-induced apoptosis is phosphorylation of the serine residue at codon 46. This region is where allele Pro47 acts as a substrate for proline-directed kinases such as the MAPK1 protein. Li et al. (2005) reported that the Ser47 allele, which is a poor substrate for MAPK1, has an apoptosis-inducing ability that is 5-fold lower than that of the wild-type Pro47 allele. However, our data again indicated that neither glioma susceptibility nor patient survival was associated with the TP53 Arg72Pro or Pro47Ser alleles in the Brazilian population. In 2009, we investigated the role of EGF +61 A>G as a potential risk factor and/or prognostic marker for gliomas in the Brazilian population. The EGF gene encodes a ligand for EGFR that activates a cascade of events responsible for promoting cell proliferation, inhibition of apoptosis, and differentiation. Alterations in the EGF/EGFR signaling pathway are associated with tumor progression in a variety of human cancers. Therefore, high expression of EGF may play a key role in glioma development and progression (Salomon et al., 1995). Shahbazi et al. (2002) first reported that the +61 A>G SNP in the 5'-UTR region of EGF is associated with increased EGF production and risk of malignant melanoma. Since that discovery, other research groups have obtained conflicting findings regarding the relationship of this functional SNP with different human cancers, including gliomas Bhowmick et al., 2004;Costa et al., 2007;Liu et al., 2009;Vauleon et al., 2007;Wang www.intechopen.com et al., 2010). In our results, the genotype and allele frequencies between cases and controls were similar, indicating no significant association with glioma risk (P = 0.94 and P = 0.887, respectively) and suggesting that EGF +61 A>G may not significantly contribute to the susceptibility to gliomas in the Brazilian population. This result is consistent with that of Vauleon et al. (2007) and Liu et al. (2009) in French and Chinese populations, respectively. However, we found that the major +61G allele (frequency among controls, 0.51) was associated with a shorter overall survival in patients (P = 0.023). Thus, with regard to patient survival, our results corroborate those of Bhowmick et al. (2004) in a population of North American patients. We have also studied the GSTP1 gene, which encodes a protein accounting for approximately 90% of the enzymatic activity of the glutathione S-transferase (GST) family (Custodio et al., 2010). GSTs constitute a superfamily of ubiquitous, multifunctional enzymes that are involved in cellular detoxification of a large number of endogenous and exogenous chemical agents that possess electrophilic functional groups (Ryberg et al., 1997). The GSTP1 protein is a pi-class enzyme and GSTP1 structure has been extensively examined in association with the risk of cancer (White et al., 2008). The influence of the GSTP1 Ile105Val SNP on cancer has been reported with inconsistent results from different parts of the world . Our results demonstrate that the Val105 allele was more frequent in a population of cancer patients than in a healthy population (0.29 and 0.06, respectively; P < 0.001) and that the presence of this genotype may increase the risk of developing astrocytomas and glioblastomas (OR = 8.60; 95% CI, 4.14-17.87; P < 0.001). However, we did not find an association between the GSTP1 Ile105Val SNP and patient survival. Recently, we began studying SNPs in DNA repair genes and we performed association analysis of SNPs in genes for the XRCC (X-ray cross-complementing) family, XRCC1 and XRCC3, in a series of gliomas (unpublished data). Human tumors may develop through alterations to the DNA repair system, which is crucially important for cellular life (Kawabata et al., 2005). To ensure the integrity of the genome, a complex system of DNA repair was developed. Base excision repair is the first defense mechanism of cells against DNA damage and a major means for preventing mutagenesis (Hu et al., 2005). Repair genes may play an important role in maintaining genomic stability through different pathways mediating base excision repair . For this reason, much attention has been given to the study of SNPs in XRCCs and their involvement in different types of cancer, including gliomas. We performed analysis of the Arg194Trp and Arg399Gln SNPs in XRCC1 and the Thr241Met SNP in XRCC3 to assess their roles in the risk and prognosis for gliomas in Brazilians. Our results provide evidence that the XRCC1 Arg194Trp SNP may contribute to the etiology of human gliomas because the Trp194 allele was strongly associated with risk and the Gln399 allele revealed a small, increased risk for tumor development. In regard to the XRCC3 Thr241Met SNP, we also found evidence that XRRC3 Thr241Met may contribute to the etiology of human gliomas. However, when the Arg194Trp and Arg399Gln SNPs in XRCC1 and the Thr241Met SNP in XRCC3 were considered together, we did not find statistical difference between genotypes and patient survival. Around the globe, other groups have analyzed the association between SNPs found in XRCCs genes and the risk of developing gliomas. Kiuru et al. (2003) evaluated the association between the XRCC1 Arg194Trp, Arg280His, and Arg399Gln SNPS, the XRCC3 www.intechopen.com Thr241Met SNP, and glioma risk in a prospective, population-based, case-control study conducted in Denmark, Finland, Sweden, and the UK. They found no significant association with gliomas for any of the SNPs when examined individually. However, the results indicated possible associations between combinations of XRCC1 and XRCC3 SNPs and the risk of glioma development, as carriers of both homozygous variant genotypes, i.e., XRCC1 Gln399Gln and XRCC3 Met241Met were associated with a three-fold increased risk of glioma (OR = 3.18; 95% CI, 1.26-8.04). In a haplotype-based approach in a Chinese population, Liu et al. (2007) investigated the role of 22 tagging SNPs (tSNPs) of XRCC5, XRCC6 and XRCC7. They found that glioma risk was significantly associated with three of the XRCC5 tSNPs (rs828704, rs3770502 and rs9288516, P = 0.005, 0.042 and 0.003, respectively), one XRCC6 tSNP (rs6519265, P = 0.044), and none of the XPCC7 tSNPs in a single-locus analysis. Haplotype-based association analysis revealed that glioma risk was significantly associated with one protective XRCC5 haplotype "CAGTT," which accounted for a 40% reduction (OR = 0.60, 95% CI, 0.43-0.85) in glioma risk. In a study of North Americans, Wang et al. (2004) found that the variant XRCC7T allele of the XRCC7 G6721T SNP was significantly more common in the glioma cases than in the controls (P = 0.045). The XRCC7 genotype frequency was also significant when comparing the cases and controls (P = 0.040). Likewise, the difference in distribution of the combined T-variant genotype (GT + TT) between the cases and controls was also statistically significant (P = 0.012), suggesting that the T allele may be a risk factor for glioma. In addition to the studies mentioned above, several others were published indicating the results of associations, positive or negative, of genomic variations with the risk of developing gliomas. However, the ethnic variations, methodological variations, and the presence of responsible, functionally unknown SNPs in linkage disequilibrium with those SNPs analyzed have contributed to the dissemination of conflicting results in different parts of the world. Gu et al. (2009) presented a review including a list of eight literature-defined, putative, functional, SNPs associated with gliomas in at least two populations from casecontrol studies. A summary of this list is presented in Table 3. Gene selection for association studies has previously been based on studies reporting the role of genes and their SNPs in the regulation of cellular functions. However, after the completion of the human genome project and the development of analytical platforms capable of parallel genotype processing, which resulted in new selection strategies for identifying susceptibility genes in many complex genetic disorders, gene selection is currently based on genome-wide association (GWA) studies. The two glioma GWA studies performed so far were published in 2010 in the same issue of Nature Genetics. In the first pages, Shete et al. (2009) presented the results of a meta-analysis of two GWA studies that involved the genotyping of 454,576 tSNPs in a total of 1,878 glioma cases and 3,670 controls, with posterior validation in three additional independent series totaling 2,545 cases and 2,953 controls. The authors identified five risk loci for glioma at 5p15.33 (TERT rs2736100), 8q24.21 (CCDC26 rs4295627), 9p21.3 (CDKN2A-CDKN2B rs4977756), 11q23.3 (PHLDB1 rs498872), and 20q13.33 (RTEL1 rs6010620). In the second study, Wrensch et al. (2009)

Cancer stem cells
Both the invasive nature of the tumor and its heterogeneity probably contribute to the poor response to the treatment regimens available today. Tumor heterogeneity is traditionally attributed to the accumulation of regional variations in the tumor microenvironment and the diversity of subpopulations of cancer cells, which result from random genetic changes (Reya et al., 2001). The majority tumors consist of a heterogeneous population of cells with different proliferative potential, as well as the ability to re-form the tumor upon transplantation into immunodeficient mice (Visvader & Lindeman, 2008). Recently, evidence has accumulated that tumors contain a population with characteristics similar to normal stem cells called www.intechopen.com cancer stem cells, which are also multipotent cells. This subpopulation has the ability to repair itself and is believed to control tumor initiation, a process that is responsible for tumor recurrence and the resistance to therapy observed in different tumor types, including gliomas (Bao et al., 2006). The observation that normal stem cells and cancer stem cells share common features (e.g., undifferentiated state and unlimited capacity for self-regeneration) led to the hypothesis of cancer stem cells (Park & Rich, 2009). A practical component of the cancer stem cell hypothesis is directed to the matter of intrinsic resistance to radiation and chemotherapy. Cancer stem cells are predicted to be difficult targets for tumor therapy because they exhibit a slowed cell cycle and high levels of drug export. Furthermore, these cells may not express or may not be dependent on the oncoproteins that are targeted by the most recent generation of cancer drugs (Cheng et al., 2010). Cancer stem cells, similar to normal stem cells, are dependent on the microenvironment in which they are located. This microenvironment, formed by cells and extracellular matrix, controls the maintenance of organ functions. Therefore, disturbance of the local microenvironment in neoplastic processes can trigger tumor development (Barcellos-Hoff et al., 2009). In glioma, for example, studies have shown that the microenvironment is surrounded by blood vessels, which provide access to signaling molecules, nutrition, and possibly to the use of the nascent vasculature for migration, which provides direct cell contact and secreted factors that are responsible for maintaining the state of quiescence of cancer stem cells, regulating their self-renewal and multipotency (Gilbertson & Rich, 2007;Jandial et al., 2008). Thus, one can say that cancer stem cells and the microenvironment are parts of the tumor. Therefore, knowledge of the associated characteristics will lead to a new understanding of tumor biology and the development of new therapeutic strategies against these cells. For this reason, it is extremely important to characterize the different subpopulations of cancer stem cells that contribute to tumor formation (Denysenko et al., 2010).

Conclusion
The development and progression of gliomas may likely be due to a multistep process that involves the functional inactivation of tumor suppressor genes and DNA repair genes, as well as the activation of oncogenes. Given the limitations of current therapies, understanding the pathways that lead to tumor progression should remain a high priority in cancer research. If the mechanisms that culminate in metastasis are fully understood, the development of new diagnostic and therapeutic methods may allow for a substantial improvement in the quality of life of affected patients and a better means of predicting patient prognosis. Genetic and epigenetic studies involving large cohorts of glioma patients in different populations have provided important information for understanding the role of key genes in the development and risk of gliomas. Because it is considered a work in progress, the WHO classification for brain tumors may soon incorporate molecular data to refine the classification of these diseases, which are complex from a therapeutic standpoint. SNPs have become increasingly popular in the genetic study of gliomas because of the quick, inexpensive and accurate analysis of SNPs. The identification of SNPs as risk factors for different glioma subtypes can be important for prevention, diagnosis and prognosis. Variations in the genomic sequence contribute to phenotypic diversity and susceptibility to or protection against many complex diseases. Thus, it is estimated that the risk of gliomas www.intechopen.com may be strongly influenced by the patterns of SNPs in certain key susceptibility genes; these SNPs are still being identified. The same reasoning can be applied to inter-individual variations in genetic responses to medications, which is a field of great interest to the pharmaceutical industry. The benefit of having a SNP map of different populations is that it allows for coverage of the entire genome so that researchers can compare the patterns and frequencies of SNPs in their patients and associate these patterns with the disease concerned. The GWA studies with significant numbers and carefully matched controls have become a powerful tool in identifying genes involved in common genetic diseases, including gliomas. The identification of susceptibility alleles provides a greater understanding of gliomagenesis and provides target genes for potential therapeutic intervention. Unlike environmental exposure, SNPs do not change during the process of tumorigenesis. Therefore, SNPs may be useful as indicators of risk.
The main features of normal stem cells are the capacity for self-regeneration and differentiation to different cell types. These characteristics are heavily regulated by the local microenvironment. Because studies have shown that cancer stem cells behave like normal stem cells, understanding the regulatory mechanisms of cancer stem cells and their microenvironment has changed our understanding of the biology of gliomas and has precipitated a reassessment of current therapies. The cure of glioma will require the elimination of all tumor cells, including cancer stem cells. Therefore, further studies to provide a better understanding of the origin of cancer stem cells and their interactions with the microenvironment are needed. These findings hold great promise for the development of new therapies that can help us improve the results achieved with current therapies and thereby prolong patient survival.