ccr focus - clinical cancer researchby gene ampliﬁcation, providing alternative mechanisms for p53...

The Evolving Role of Molecular Markers in the Diagnosis andManagement of Diffuse Glioma

Jason T. Huse1 and Kenneth D. Aldape2

AbstractWhile the classification of diffuse gliomas has relied on the examination of morphologic features

supplemented with techniques such as immunohistochemistry, there is an increasing recognition of

substantial biologic diversity within morphologically defined entities. High-throughput technologies, in

particular studies that integrate genome-wide data from diverse molecular platforms, increasingly identify

the existence of robust and distinct glioma subtypes. While treatment advances and improvement of

outcomes for patients with diffuse glioma have beenmodest, theremay be benefit to integrate findings from

biologic studies into clinical practice to enhance the precision of treatment for these diseases. Recent

examples such as the identification of mutations in IDH1 and IDH2 as an early genetic event that is

predominantly in lower-grade gliomas (grades 2 and 3) underscore the importance of molecular discovery

leading to the ability to develop subclassifications with prognostic and potentially therapeutic implications.

In contrast, glioblastoma (grade 4), the most common and aggressive glioma, typically arises without IDH

mutation, supporting the need for different therapeutic approaches. Additional genomic and epigenomic

signatures are generally nonoverlapping between IDH-mutant and IDH wild-type diffuse glioma, and

despite comparable histopathology, IDH-mutant gliomas can be considered as biologically distinct from

IDHwild-type gliomas. In this CCR Focus article, we highlight and summarize the current understanding of

recent molecular findings and the relationships of these findings to clinical trials and clinical management.

See all articles in this CCR Focus section, "Discoveries, Challenges, and Progress in Primary Brain

Tumors."

Clin Cancer Res; 20(22); 5601–11. �2014 AACR.

IntroductionDiffuse gliomas are the most common intrinsic tumors of

the central nervous system and consist of neoplasms thatexhibit a wide range of biologic and clinical features. From along-range perspective, glial tumors are broadly defined aseither circumscribed or diffuse. Circumscribed gliomasinclude entities such as pilocytic astrocytoma, World HealthOrganization (WHO)grade1. These tumors are, inprinciple,potentially curable by surgical resection and will not beconsidered further. In contrast, diffuse gliomas are by defi-nition incurable by surgery due to their infiltrative nature, aremore common, especially in adults, and are the subject ofthis review. Diffuse gliomas are categorized by malignancygradewith a range ofWHOgrades 2, 3, and 4 (glioblastoma)and exhibit a wide range of clinical behavior, ranging fromslow clinical progression in lower grade tumors, to very short

median survival times for patientswithWHOgrade 4 tumors(glioblastoma). Glioma is unusual, if not unique, as amalignant tumorwith little propensity for distantmetastasis,and prognosis is a function almost entirely of grade ratherthan stage. Malignancy grade, and to a lesser extent mor-phologic determinations of presumed histogenesis (astrocy-toma vs. oligodendroglioma vs. mixed oligoastrocytoma)accounts for some of the variation in patient outcome, butsubstantial and extensive within-grade and within-entityclinical and biologic variability exists, prompting investiga-tion intomolecular factorswhichmaybetter account for suchvariation.While in the past themolecular basis that underlaythese contrasts between glioma variants was not well under-stood, recent efforts at molecular characterization, especiallythrough high-throughput genomic and epigenomic screenshave, in their identification of distinctive and highly recur-rent genomic and epigenomic abnormalities, clarified someof this diversity and have provided the beginnings of newconcepts in tumor classification. In addition, insights gainedfrom these studies have pointed to viable avenues for ther-apeutic development.

As stated above, diffuse gliomas of adulthood are unifiedby a shared propensity to widely infiltrate surroundingnormal brain parenchyma, a property that effectively rendersthem incurable by resection. As such, a major goal of neu-rosurgical resection in cases of diffuse glioma includes

1Department of Pathology, Memorial Sloan Kettering Cancer Center, NewYork, New York. 2MacFeeters-Hamilton Brain Tumour Centre, PrincessMargaret Cancer Centre, University of Toronto, Toronto, Ontario, Canada.

Corresponding Author: Kenneth D. Aldape, Princess Margaret CancerCentre and Ontario Cancer Institute, 610 University Avenue, Toronto, ONM5G2M9, Canada. Phone: 416-634-8728; Fax: 416-352-6031; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-14-0831

�2014 American Association for Cancer Research.

CCRFOCUS

www.aacrjournals.org 5601

on June 3, 2020. © 2014 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

accurate material for diagnosis and classification as well ascytoreduction when clinical indicated; but the goal of com-plete resection with curative intent is essentially unrealisticwith diffuse glioma, as patients almost universally undergotumor recurrence even after substantial cytoreductive sur-gery. That said, the range of diffuse glioma behavior exhibitsconsiderable clinical heterogeneity. More specifically,patients with glioblastoma (WHO grade 4) demonstrateoverall survival times of approximately 15 months whilethose affected by low-grade (WHO grade 2) astrocytomasand oligodendrogliomas frequently exhibit prolonged clin-ical courses lasting years to even decades (1, 2). Anotherfeature of diffuse gliomas (again, generally in contrast tocircumscribed variants) is the phenomenon of tumor pro-gression/malignant transformation, where lower grade glio-mas (WHOgrades 2 and 3) over time not only recur but alsoprogress to higher malignancy grade (WHO grades 3–4),with associated rapidity of clinical deterioration.

Historically, gliomas were classified largely on morpho-logic and histopathologic characteristics, forming the foun-dation for further investigations.Morphologic classificationis necessarily based on interpretation of microscopic fea-tures using routinely available tools. While clinical andradiologic features are at times taken into account, theWHO-sanctioned classification is at its core, histology-based. The value of this remains within its clinical utilityby virtue of the association of tumor grade with patientoutcome (3). However, many aspects of the histologic-based classification (which is not outlined in detail here)fall short with respect to potential personalization of care,with respect tomatching of the biologic foundations of eachtumor with appropriately targeted treatment regimens.

Large-scale molecular profiling of diffuse gliomas isoccurring bothby individual laboratories aswell as nationalconsortia such as The Cancer Genome Atlas (TCGA) net-work. Glioblastoma was an early tumor type in the TCGAarmamentarium, with the recent addition of grade 2–3gliomas (referred to by TCGA as the "lower grade glioma"effort). These efforts and others have led to an enhancedunderstanding of the molecular makeup of diffuse gliomasand have uncovered a number of recurring aberrations ingenes and pathways, including mutations/abnormalities inspecific genes, multicomponent expression signatures, andDNA methylation patterns (4–7). While a number of thesemarkers and pathways have been previously implicated inglioma biology, some represent new discoveries. Moreover,the multidimensional datasets provided by TCGA andrelated efforts allow for the integration of findings fromorthogonal platforms (e.g., gene mutation, mRNA, epige-netics) within a large set of tumor samples. Such analyseswill lead to new opportunities for the identification ofrobust biologic subtypes. In turn, patterns of aberrationsare becoming evident from these data that point to the needfor new approaches to biomarker analysis that go beyondboth routine procedures that are in the comfort zone ofcurrent clinical laboratories (e.g., immunohistochemistry)as well as beyond single marker analyses. While efforts suchas TCGA and other genome-wide studies have been helpful,

it is important to put these efforts in perspective, includinglimitations such as the retrospective nature of the findings,the fact that findings from these studies are often nottransformative and, in the future, require prospective eval-uation in uniformly treated patients (Fig. 1). Efforts toinclude clinically relevant molecular alterations into theformal, WHO classification scheme for brain tumors areunderway (8). In addition, while some prognostic markersare emerging, an unmet need in the field are true predictivemarkers for diffuse glioma, thatwhen implemented, informreal-time treatment decisions for the individual patient.

Established Oncogenes and Cancer Pathways inGlioma

Early work, beginning several decades ago, identifiedrecurring molecular alterations characterizing diffuse glio-mas. A large number of such genetic events have beencharacterized to date, and patterns have emerged pointingto key pathways likely driving these tumors, some of whichhave been causally verified using mouse models (9, 10).Among these are losses of the RB1 and p53 tumor suppres-sor genes or, alternatively, alterations in genes involved inpathways related to these tumor suppressors. Mutations inthe TP53 gene have long been considered molecular hall-marks of lower grade astrocytic tumors and the secondaryglioblastomas intowhich they evolve (11–13).Whilemuta-tions in RB1 itself have been known for some time (14)not to be common in gliomas, genes encoding its upstreamregulators are frequently altered (5, 14, 15). In particular,CDKN2A, which encodes both the INK4A and ARF genes,crucial activators ofRB1 and p53, respectively, is deleted in alarge percentage of diffuse gliomas (16), particularly glio-blastomas (15, 17–19). In addition, upstream repressorssuch as CDK4 and D-type cyclins (RB inhibitors) andMDM2 (p53 inhibitor), are frequently upregulated, oftenby gene amplification, providing alternative mechanismsfor p53 and RB1 pathway silencing. Together, the sheerfrequency of these abnormalities underscores the impor-tance of the RB1 and p53 pathways as tumor-suppressivemolecular networks in glioma biology. This concept wasshown quite clearly in the initial TCGA publication onglioblastoma, which found that there were some alterationsexpected to inactivate the p53 and Rb signaling networks inat least 87% and 78% of all cases, respectively (5).

In addition to tumor-suppressive pathways, perhaps themost well-known and common activating oncogenicchanges in glioma are those involving receptor tyrosinekinases (RTK). Mouse modeling studies have underscoredthe importance of these events as gliomagenic drivers. TheEGFR frequently undergoes high-level genomic amplifica-tion in adult glioblastoma (�40%), often in conjunctionwith constitutively activating mutations in the ectodomainof the protein that include, but are not limited to, the variantIII (vIII) deletion event (5, 15, 20–22).While the expressionof EGFRvIII is complex and heterogeneous (23), it isdetected in approximately 30% to 50% of cases in whichEGFR amplification is present. In addition, high-level

CCRFOCUS

Clin Cancer Res; 20(22) November 15, 2014 Clinical Cancer Research5602



amplification of the platelet-derived growth factor receptorgene (PDGFRA) is also present, although in a smallerproportion (�13%) of adult glioblastoma (5). Analogousto EGFR, constitutively activating deletion mutants inPDGFRA have been described in receptor-amplified tumors(24). PDGFRA amplification also appears to be a commongenomicRTKalteration impacting pediatric glioblastomaaswell as diffuse intrinsic pontine glioma (18, 25, 26). Thoughmuch less common, high-level MET amplification alsooccurs in glioblastomas (5, 15, 18). More importantly,individual glioblastoma tumors can display activatinggenomic alterations inmultiple RTKs, simultaneously, withamplified receptors often segregating to distinct cellularsubpopulations (27, 28). This finding has implications onthe likely efficacy of drugs targeting single RTKs in thesetumors. Finally, while lower-grade (grade 2–3) gliomasinfrequently harbor high-level amplification in RTK genes,enhanced PDGF signaling and PDGFRA phosphorylationhas been routinely implicated in their pathogenesis (29–31). Together, the high degree of heterogeneity and com-plexity of RTK biology in glioma may account for some ofthe lack of success of anti-EGFR trials in glioblastoma.Genomic alterations involving core components of onco-

genic signaling pathways situated downstream of RTKs arecommon findings in glioblastoma. Integrated analysis inthe initial TCGAglioblastoma report described activation ofthe extended PI3K–AKT–mTOR and RAS–MAPKmolecularpathways in nearly all glioblastoma samples evaluated (9).Dysregulating alterations include activating mutations ineither the catalytic (PIK3CA) or regulatory (PIK3R1)domains of PI3K and are found in approximately 15% of

adult glioblastomas, as well as deletions and/or silencingmutations in PTEN, the primary negative regulator of PI3K–AKT signaling, (�1/3 of cases). Beyond genetic alterationsof PTEN, additional epigenomic and microRNA (miRNA)-based mechanisms of PTEN repression have also beendescribed in diffuse gliomas with the former operative insignificant numbers of adult lower-grade gliomas (WHOgrades 2 and 3, hereafter referred to as LGGs; 50%–60%;refs. 30, 32–35). Mutations in the RAS antagonist neurofi-bromin (NF1) have long been known to cause neurofibro-matosis type I, a cancer predisposition syndrome charac-terized by frequent neurofibromas and, to a lesser extent,astrocytomas (36). However, more recent studies havedemonstrated NF1 mutation/deletion in 15% to 18% ofprimary glioblastomas (5, 15), where it appears to representa key molecular alteration of the mesenchymal glioblasto-ma subclass (see below).

Gene Expression SignaturesGene expression profiling was the first high-throughput

technology to be broadly applied to human cancer. Largelybecause of its early adoption, gene expression technologyremains a standard approach enabling the identification andcharacterization of biologic subclasses in clinicopathologictumor entities. Initial successful applications of expressionprofiling to identify tumor subclasses included adenocarci-noma of the breast and diffuse large B-cell lymphoma. Inearly work on gliomas focused primarily on high-gradetumors like glioblastoma, comprehensive transcriptionalanalysis was initially used to establish molecular correlatesfor known clinical and/or histopathologic distinctions,

© 2014 American Association for Cancer Research

Elucidate biologic subtypes

Refine glioma classification

Clinical trials (“targeted” and nontargeted agents)

Mandate tissue-integrated/integral biomarkers

Match biomarker profile to targeted therapy

Test efficacy as a function of biologic subtype

Correlate biologic subtype with overall outcomes

(prognostic) and therapy-specific patient outcomes

(predictive)

Recurrent glioma

Tissue recharacterization

Clinical trials for recurrent disease

Newly diagnosed glioma

Genome-wide screens

Integrated genomic and epigenomic analysesFigure 1. Incorporation of tissuecorrelates and genomicclassification into glioma clinicaltrials. High-throughput/genome-wide methods are used tocharacterize tumors, andintegration of data from diverseplatforms is needed to definebiologic subtypes. In turn,interrogation of the relationship ofbiologic subtype with clinicaloutcomes relative to targeted andnontargeted therapy is required toenrich patient populations intoclinically relevant subgroupsmatched, to the extent possible,with efficacious therapies.Recognition that gliomas atrecurrence can be genomicallydistinct from the newly diagnosedsetting (82) requires considerationof genomic characterization atrecurrence when feasible.

Molecular Markers in Diffuse Glioma

www.aacrjournals.org Clin Cancer Res; 20(22) November 15, 2014 5603



such as WHO grade, astrocytic versus oligodendroglial mor-phology, and primary versus secondary glioblastoma status(37–45). Subsequent studies have focused more on identi-fying natural molecular distinctions within diffuse gliomasand, despite some differences, have revealed robust sub-classes within glioblastoma. The first of these studies exam-ined prognostically relevant gene signatures inWHOgrade 3and 4 diffuse gliomas, revealing three major subclasses,termed proneural, mesenchymal, and proliferative. This wasfollowed by additional investigations, most notably that ofTCGA, which delineated four transcriptional subclasses,termed proneural, neural, classical, and mesenchymal (7).In this latter study, genomic associationswere also identified,with classical, proneural, andmesenchymal tumors stronglyenriched for abnormalities in EGFR, PDGFRA and IDH1 orIDH2, andNF1, respectively. Incorporationof thesefindings,in the context of the fundamental differences between IDH-mutant and IDH wild-type glioma leads to the concept ofIDH as the first "split" into two groups with additionalsubtypes within each of these groups (Fig. 2). These andother analyses support the notion that malignant gliomascan be loosely classified into two subgroups, correspondingto proneural and mesenchymal gene signatures (Fig. 3).Analyses from both studies found that the large majority oftumors classified as either proneural ormesenchymal by onemethodology were reclassified to the same subclass by theother, and vice versa (46). The robustness of these transcrip-tional subclasseshas been further testedby the incorporationof orthogonal high-throughput data. Two recent publica-tions, one based entirely on miRNA profiling and the other

integrating mRNA, miRNA, gene copy number, and globalDNAmethylation data, reported five and three glioblastomasubclasses, respectively, emerging from unsupervised analy-sis (47, 48). Nevertheless, in both studies, tumors defined aseither proneural ormesenchymal bymRNAprofiling tendedto colocalize within the new subclassification scheme.Whilethese two subclasses appear to be reproducibly defined andcharacterized, and the relationship of IDH-mutant and G-CIMP tumors (see below) to the proneural expression classappears stable, the ability of gene expression signatures toreliably classify gliomas in a clinically relevant fashionremains uncertain. Indeed, subclass assignment has beenshown to change following treatment (4, 49). Moreover, arecent study analyzing expression signatures of single cellswithin glioblastoma samples showed substantial intratu-moral heterogeneity of expression subclasses within eachtumor (50), suggesting that bulkmRNAprofiling of a gliomasimply represents an average of a heterogeneous mix oftranscriptional signatures, likely influenced by multiple fac-tors (genetic changes, microenvironment, treatment, etc.).These considerations suggest that while gene expression is acritical component of glioma biology, alternative cancer-specific alterations in the genomic/epigenomic space mayrepresent more stable metrics for tumor classification.

Clinically Relevant Classification MarkersIDH mutation

The discovery of IDH mutations in diffuse glioma repre-sents a major paradigm shift in our understanding and


IDHmut

G-CIMP

10 loss

7 gain

9p loss

TP53 mutation

(IDHwt)

Favorable

Intermediate

Poor

Proneural

1p/19q codeletion

Proneural

TP53 mutation

Proneural

PDGFRA

CDK4

Classical

EGFRvIII

Mesenchymal

NF1

Astro grade 2/3

GBM grade 4

Driver eventsGene expression/

additional alterationsWHO classification

Oligo grade 2/3

Astro grade 2/3GBM grade 4

Patient outcome Figure 2. Subtypes and keymolecular signatures in diffuseglioma. IDH-mutant/G-CIMP–positive gliomas represent themajority of grade 2 and 3 diffusegliomas (top) and are within theproneural expression subclass.These can be further subclassifiedon the basis of 1p/19qcodeletion (frequently showingoligodendroglioma histology) orTP53 mutation (frequentlyshowing astrocytic histology).IDH-mutant glioblastoma is rare(5%–10% of glioblastomas) but iswell recognized, especially inyounger adults. IDH wild-type gliomas (bottom) showcharacteristic changes, includingas possible driver events loss of10, gain of 7, and loss of 9p/TP53mutation (56). These can be thensubdivided into additionalexpression subclasses. Themajority of IDH wild-type diffusegliomas are glioblastomas,although grade 2–3 tumors arealso represented.

CCRFOCUS




classification of these tumors. The finding of mutations inIDH1 and IDH2 in glioma was completely unexpected andmade using a genome-wide pure discovery approach (15),as comparedwith an approach taken by groups that focusedon a candidate mutation, indicating the value of discoveryscience to generate novel and important findings. IDHnormally results in decarboxylation of isocitrate to a-keto-glutarate, leading to reduced NADP in the Krebs cycle.Mutation in codon 132 disrupts the function of IDH1 toconvert isocitrate to a-ketoglutarate. In particular, IDHmutation results in an ability of the enzyme to catalyze thereducedNADP-dependent reductionofa-ketoglutarate toR(-)-2-hydroxyglutarate (2-HG). 2-HG levels are elevated inhuman malignant gliomas with IDH1 mutations, andreplacement of arginine 132 by histidine results in shiftingof residues and acquisition of the ability to convert a-keto-glutarate to 2-HG (51).With the finding of IDH1 and IDH2mutations in diffuse

gliomas and correlation with tumor grade and clinicalfeatures, several conclusions can be drawn. First, IDHmuta-tions are an early event in the pathogenesis of a subset ofdiffuse glioma. The finding that both astrocytoma andoligodendroglioma harbor IDHmutations at high frequen-cy suggests that this is an early event in molecular patho-genesis, as well as suggesting that astrocytoma and oligo-dendroglioma have a similar cell of origin. IDH-mutantgliomas will typically exhibit additional alterations, namelymutations in TP53 (corresponding to histologically diag-nosed astrocytoma) or, conversely, codeletion of chromo-somes 1p and 19q (resulting from an unbalanced t(1;19)(q10;p10) translocation (refs. 52, 53; corresponding tooligodendroglioma histology). Second, the clinicopatho-logic distinction between the de novo/primary versus sec-ondary pathways of glioma development can largely beaccounted for by the absence and presence, respectively, ofIDH mutations. Most gliomas that are identified at a lower

grade (grade 2 or 3) are IDH-mutant, whereas most cases ofde novo glioblastoma (short clinical presentation, glioblas-tomahistology at initial diagnosis) are IDHwild-type. Thereare signature DNA copy number changes that are nearlysynonymous with the molecular definitions of primaryglioblastoma include gains of chromosome 7 and loss ofchromosome 10. These occur in 80% to 90% of glioblas-toma at initial diagnosis and are essentially mutually exclu-sive with IDH mutation. In addition, amplification at 7p,harboring EGFR, occurs in 40% to 50% of glioblastoma atinitial diagnosis, but again does not occur in the setting ofIDH-mutant diffuse glioma. Conversely, patterns of recur-rent copy number alterations in IDH-mutant diffuse glio-mas, including 1p/19q codeletion, consist largely of eventsnot frequently encountered in primary glioblastoma.More-over, the glioma CpG island methylator phenotype (G-CIMP), characterized by stereotypic and concordant meth-ylation of a set of CpG sites, is almost entirely restricted toIDH-mutant tumors. This finding was functionally clarifiedto show that mutant IDH can cause changes correspondingto G-CIMP in model systems (54). Molecular changes thatoccur between IDH-mutant and wild-type diffuse gliomasare to a large extent unshared, suggesting distinct biologydespite histologic similarity (see Fig. 4). There is mutualexclusivity of 1p/19q codeletion and TP53mutation amongIDH-mutant glioma, suggesting a hierarchy of genomicevents and divergent subpathways within IDH-mutant gli-oma. While IDH/G-CIMP status can distinguish two majorsubtypes of diffuse glioma, there are some correlates withcurrent histopathologic classification. For example, at initialdiagnosis, among lower grade gliomas (grade 2–3), themajority are IDH-mutant/G-CIMP positive, with only10%–20% in the IDH wild-type/G-CIMP–negative geno-mic class. Alternatively, among newly diagnosed glioblas-toma/grade 4 tumors, IDH mutation/G-CIMP positivity isuncommon (5%–10%), with the majority of glioblastoma

PN CL MES

Verhaak expression subtype

NL

ME

SP

NP

RO

LIF

Phi

llip

s ex

pre

ssio

n su

bty

pe

Figure 3. Comparison of twopublished transcriptomalclustering schemes that havebeen applied to subclassifyingglioblastoma: Phillips et al. (4) andVerhaak et al. (7). Expressionprofiles from each study wereclassified according to bothstudies' signature sets.Correspondence is strongestbetween mesenchymal andproneural signatures (CL,classical; MES, mesenchymal;NL, neural; PN, proneural;PROLIF, proliferative. Circle sizerepresents the strength (silhouettewidth, SW) of the original classassignments. Adapted from Huseet al. (46). Copyright 2011, JohnWiley and Sons.





as IDH wild-type. Nevertheless, both genomic subtypes areobserved in all grades and serve to distinguish diffuseglioma into distinct biologic subtypes. The presence orabsence of IDH mutation provides the molecular correlatefor the previously established concepts of "secondary" or"progressive" form of glioblastoma (IDH-mutated) com-pared with the "primary" or "de novo" form (IDHwild-type;ref. 55). Investigating IDHwild-type/G-CIMP–negative dif-fuse glioma, a recent study identified loss of 10, gainof 7 andeither loss of 9p or mutation of TP53 as key driver events inthese tumors (56).

In addition to delineating a large subclass of diffuseglioma, IDHmutation status serves as a useful classificationmarker to distinguish LGGs from histologically similarentities, whether neoplastic or not. For example, oligoden-droglioma can occasionally be difficult to distinguish from

morphologically similar entities (e.g., central neurocytoma,clear cell ependymoma), and the presence of an IDHmutation essentially rules out these alternative possibilities.Occasionally, a diffuse glioma can be difficult to distinguishfrom a circumscribed glioma (e.g., pilocytic astrocytoma,pleomorphic xanthoastrocytoma, etc.) and again the pres-ence of an IDHmutation resolves this differential diagnosis.Finally, the presence of IDH-mutant tumor cells againmitigates the distinction between diffuse glioma and reac-tive conditions (astrogliosis). In this regard, the availabilityof the R132H-specific anti-IDH1 mutant antibody is espe-cially helpful, as tumor cells are often in the minority inthese cases. However, the ease of mutant IDH1 IHC shouldnot preclude the need for a full workup and/or interpreta-tion of results, as 10% to 20% of all IDH mutations (non-R132H IDH1 mutations and all IDH2 mutations) are


Cancer-prone neuroepithelial cells (multiple cells of origin?)

IDH mutation, CIMP

1p/19q loss, TERT

Common

Uncommon Common

Uncommon

TP53, ATRX

IDH-mutated GBM

?????

?

A

Astrocytoma

Oligodendroglioma

9p loss, 10q loss

7 gain (EGFR amp), 19 gain

IDH wild-type GBM

Histopath—similar

Genomics—distinct

OligodendrogliomaAstrocytoma

Secondary orprogressive pathway

BPrimary or

de novo pathway

Figure 4. IDH wild-type versus IDH-mutated pathways to glioblastoma (GBM). Previously recognized as "primary" or "de novo" (right) and "secondary"or "progressive" (left) pathways to glioblastoma, the absence/presence of IDH mutation provides the molecular correlate of this long establishedparadigm. With the identification of IDH mutation/G-CIMP as present in a subset of diffuse glioma, a distinction can be made into two fundamentallydistinct glioma types that cannot be resolved histologically. Left, the IDH-mutant/G-CIMP–positive molecular changes in glioma are illustrated. Right, IDHwt/G-CIMP–negative gliomas are illustrated. In the IDH-mutant subset, presentation as a lower grade (WHO grade 2–3) is uncommon, whereas presentationas glioblastoma is uncommon. The opposite is true for IDH wild-type glioma (right), but these differences are relative. The lower aspect of the figureillustrates the point that routine histology, on which current classification is based, cannot reliably distinguish these important subtypes of diffuseglioma. Question marks indicate changes that have not been fully elucidated or are not well characterized.

CCRFOCUS




missed by this antibody. IDH-mutant glioma, in general,carries an improved prognosis compared with IDH wild-type gliomas of similar grade, but IDHmutation representsan important oncogenic event that leads to a lethal diseaseand efforts to target it therapeutically are certainlywarranted.

1p/19q lossThe finding of combined 1p/19q codeletion and its

correlation with oligodendroglioma has been known forsome time (17) and its characteristic presence in a subsetof gliomas has spurred efforts to identify is biologic andclinical significance. Most oligodendrogliomas with 1pand 19q codeletions also carry mutations in the CIC gene,a homolog of the Drosophila gene capicua, on chromo-somal band 19q13.2, (57, 58). In addition, althoughperhaps less frequently, mutations in FUBP1 on chromo-somal arm 1p are also present in 1p/19q codeletedtumors. 1p/19q codeletion is almost invariable associatedwith IDH mutation, and within the setting of IDH-mutant glioma, 1p/19q codeleted tumors carry a betterprognosis than non-codeleted tumors when matched formalignancy grade. While it is not clear whether thenatural history of codeleted tumors differs from thatof IDH-mutant/non-codeleted tumors in the absence oftherapy, it does appear that codeletion is a marker ofimproved response to cytotoxic chemotherapy. Long-term results of two large randomized clinical trials, Euro-pean Organization for Research and Treatment of Cancer26951 and Radiation Therapy Oncology Group 9402 (59,60), showed that adjuvant PCV chemotherapy confers asurvival advantage in the setting of 1p/19q codeletion.These findings demonstrate that, for diffuse gliomas ingeneral, codeletion can be viewed as a predictive markerfor PCV chemotherapy. However, codeletion defines whatis essentially a subset of IDH-mutant glioma and thedegree to which codeletion serves as a predictive chemo-therapy marker in the context of IDH-mutant gliomasalone remains less clear. An additional unanswered ques-tion relates to the fact that PCV chemotherapy has sincebeen subsumed with temozolomide therapy, raising theissue as to whether molecular correlates related to PCVchemotherapy will correspond to patients treated withtemozolomide. Defining predictive markers are one ofthe most important goals of biomarker characterizationand personalized medicine and with activity in antian-giogenic therapy, immunotherapy, and therapies forpediatric tumors (all addressed in this edition of CCRFocus; refs. 61–63).

ATRX mutations and TERT alterationsRecent work has shown that mutations in the SWI/SNF

chromatin regulator ATRX frequently occur in gliomas, andare almost invariably associated with mutations of TP53and IDH1 genes across glioma entities (58, 64, 65). A strongcorrelation between inactivation of ATRX and the phenom-enon known as alternative lengthening of telomeres hasbeen postulated (64). Loss of ATRX function impairs the

heterochromatic state of the telomeres and leads to telo-mere destabilization, which facilitates the development ofalternative lengthening of telomeres. These same studiesshowed that ATRXmutations were a very specificmarker forastrocytic lineage tumors, including diffuse and anaplasticastrocytomas and most oligoastrocytomas. This evidencemakes ATRXmutations an appealing counterpart for 1p and19q codeletions, which seem to be mutually exclusive withATRX mutations. As most mutations detected to date aretruncating and thus lead to a reduction in protein concen-trations, immunohistochemical assessment of loss of ATRXcould be a reasonable surrogatemarker of ATRXmutations.Moreover, combining 1p/19q codeletion and ATRX assess-ments in the clinical setting could help guide diagnosiswithin the spectrum of IDH-mutant gliomas and, in thelong-term, to stratify patients for specific treatments. TERTpromotermutations are also characteristic of diffuse gliomaand are mutually exclusive with ATRX mutation in IDH-mutated diffuse glioma (66). Promoter mutations result inincreased TERT expression, with one study showing that heexpression level of TERT in tumors carrying thosemutationswas on average 6.1 times higher than that of TERTwild-typetumors (67). Within IDH-mutant glioma, TERT promotermutations are observed in almost all tumors harboringconcurrent 1p/19q loss, while IDH-mutant non-1p/19qcodeleted tumors tend to have ATRX mutation (67). Themutations are stereotypic, with the C228T mutation withinthe TERT promoter occurring 146 bp upstream of the ATGstart codon of TERT, whereas the C250T mutation occurs126 bp upstream. Both mutations generate a de novosequence, which contains the ETS transcription factor bind-ing motif, allowing recruitment of transcription factors andincreased expression (68).

MGMT promoter methylationAnumber of clinical trials and cohort studies have shown

that promoter methylation and silencing of the MGMTgene, which codes for O6-methylguanine-DNA methyl-transferase, a DNA repair enzyme, is associated with pro-longedprogression-free andoverall survival in patientswithglioblastoma who are being treated with alkylating agentchemotherapy (69–73). A landmark practice-changing trialcompared concurrent/adjuvant temozolomide during andafter radiotherapy versus radiotherapy alone for glioblas-toma patients in the newly diagnosed setting (74). Subsetanalysis of samples from this trial showed that the benefitfrom chemotherapy was almost exclusively attributable topatients withMGMT-methylated tumors (71). Furtherworkin elderly patients, in whom aggressive treatment is aclinical issue, showed improved outcome in the setting ofchemotherapy for MGMT-methylated tumors and, interest-ingly, worse survival associated with unmethylated tumors(75, 76), indicating that in this setting, MGMTmethylationis not prognostic, but rather predictive. Additional worksupports the predictive value of MGMT methylation andthere is evidence to suggest that its predictive status isconditional on IDH mutation status (predictive only inIDH-mutant glioma; refs. 77, 78). To date, the literature on





MGMT methylation in glioblastoma would suggest thatoverall, it is at least partially a predictive (as opposed tomerely prognostic) marker for chemotherapy. This concepthas led topatient selection for clinical trials omitting TMZ inpatients with glioblastoma without MGMT methylation tofind treatments for this patient groupwith no real treatmentoptions at present. Work in this area, to date presented inabstract form, awaits further maturity.

However, several important considerations and caveatsdeserve mention. First, while MGMT methylation is notnecessarily a defining marker of G-CIMP, there is clearconcordance and overlap between these two biomarkers.Specifically, whileMGMTmethylation is clearly present in asubset of G-CIMP–negative glioblastomas, MGMT methyl-ation appears to be present in nearly all cases of G-CIMP–positive glioblastomas (79). As G-CIMP itself is a favorableprognostic marker, the effect of MGMTmethylation in IDHwild-type/G-CIMP–negative glioblastoma needs to be fullydefined. Second,methodology varies with respect toMGMTmethylation detection and results are not completely uni-form, warranting some caution in the interpretation ofstudies across laboratories (80). Third, determination ofMGMT status by immunohistochemistry has notable inter-observer variability and is not reliably associated withpromoter methylation or outcome. A recent study in whichfour observers examined MGMT expression showed sub-optimal interobserver agreement, as well as only modestcorrelation with MGMT promoter methylation status andno significant association ofMGMT expression with patientoutcome (81). This may call into question, to some extent,whether the favorable outcome experienced in MGMT-methylated cases is in fact related to reduced MGMT enzy-matic activity. Finally, the lack of good options for patientswith MGMT-unmethylated tumors limits the use of thismarker to actually personalize therapy for glioblastoma.Future work to optimize laboratory detection of MGMTmethylation, reduce interlaboratory variability, and clearlydefine the relationship of MGMT methylation to patientoutcome and IDH/G-CIMP status is warranted to fullyevaluate this promising biomarker.

Concluding RemarksSubstantial progress has been made in the molecular

classification of primary brain tumors. The recent molecularcharacterization of diffuse glioma has provided not only aclarified framework for the conceptualization of thesetumors, but also revealed pathways for the development ofmore effective targeted therapeutics. Experience with severalmarkers is at the point where clinical integration is becomingstandard of care. For MGMT promoter methylation in glio-blastomas, including those arising in elderly patients, and 1pand 19q codeletions in anaplastic oligodendroglial tumors,

molecularmarkers now play amajor role in clinical decisionmaking (59, 60, 75, 76). Dependent on the outcome ofongoing phase II and III trials, biomarkers to predict resis-tance or sensitivity to angiogenesis inhibition could alsoprove useful. Meanwhile, high-throughput analyses at thegenetic, epigenetic, and expression levels have shown theirvalue in refining the classification of brain tumors andprognostication of outcome. These techniques might soonbecome more widely available, easier to standardize, andhave less bias than single marker assessments—e.g., currentmethods for MGMT methylation assessment—and mightsoon become more cost effective. Accordingly, we predictthat the current histology-dominated diagnostic assessmentof brain tumors will be increasingly supplemented bymolec-ular diagnostic tests, which eventually might be graduallyreplaced by high-throughput profiling techniques, includingarray-based approaches and next-generation sequencing.However, for those treating patients with these tumors, thereis the feeling that translation of molecular findings intomeaningful improvements in patient outcomes has beenslow. A recent study underscores some of the challenges intranslating molecular findings into clinical trials. Usingmatched primary–recurrent matched pairs. The studyshowed that recurrences did not exhibit the full set of muta-tions found in the initial tumor, suggesting that the muta-tional profile of recurrent glioma cannot be completelydeduced from the tumor at initial diagnosis (82). Combinedwith the finding that temozolomide therapy can result inmutations in theDNAmismatch repair pathway that result ina hypermutator phenotype (83). These findings emphasizethe need for tumor resampling (Fig. 1), when possible, in thesetting of clinical trials for recurrent glioma (emphasizedin Fig. 1). Advances in minimally invasive means to sampleand detect key genomic aberrations in glioma, such as cell-free nucleic acid (84, 85) or circulating tumor cells (86). Asthe process of therapeutic refinement moves forward, moreeffective preclinical models and optimal clinical trial designwill be absolutely crucial, as will the ready availability ofsophisticated genomic technology in the clinical environ-ment, starting with the use of relevant molecular markers asan objective means for tumor classification.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: K.D. AldapeAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): K.D. AldapeWriting, review, and/or revision of the manuscript: J.T. Huse,K.D. Aldape

Received August 21, 2014; revised September 22, 2014; acceptedSeptember 24, 2014; published online November 14, 2014.

References1. Lang FF, Gilbert MR. Diffusely infiltrative low-grade gliomas in adults.

J Clin Oncol 2006;24:1236–45.2. Wen PY, Kesari S. Malignant gliomas in adults. N Engl J Med 2008;

359:492–507.

CCRFOCUS




3. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A,et al. The 2007 WHO classification of tumours of the central nervoussystem. Acta Neuropathol 2007;114:97–109.

4. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD,et al. Molecular subclasses of high-grade glioma predict prognosis,delineate a pattern of disease progression, and resemble stages inneurogenesis. Cancer Cell 2006;9:157–73.

5. Cancer Genome Atlas Research N. Comprehensive genomic charac-terization defines human glioblastoma genes and core pathways.Nature 2008;455:1061–8.

6. Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara K,Berman BP, et al. Identification of a CpG island methylator phenotypethat defines a distinct subgroup of glioma. Cancer Cell 2010;17:510–22.

7. VerhaakRG,HoadleyKA,PurdomE,WangV,QiY,WilkersonMD, et al.Integrated genomic analysis identifies clinically relevant subtypes ofglioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR,and NF1. Cancer Cell 2010;17:98–110.

8. Louis DN, Perry A, Burger P, Ellison DW, Reifenberger G, von DeimlingA, et al. International society of neuropathology-haarlem consensusguidelines for nervous system tumor classification and grading. BrainPathol 2014;24:429–35.

9. Huse JT, Holland EC. Genetically engineered mouse models of braincancer and the promise of preclinical testing. Brain Pathol 2009;19:132–43.

10. Hambardzumyan D, Parada LF, Holland EC, Charest A. Geneticmodeling of gliomas in mice: new tools to tackle old problems. Glia2011;59:1155–68.

11. Louis DN. The p53 gene and protein in human brain tumors.J Neuropathol Exp Neurol 1994;53:11–21.

12. Chung R,Whaley J, Kley N, Anderson K, Louis D, Menon A, et al. TP53gene mutations and 17p deletions in human astrocytomas. GenesChromosomes Cancer 1991;3:323–31.

13. Mashiyama S, Murakami Y, Yoshimoto T, Sekiya T, Hayashi K. Detec-tion of p53 gene mutations in human brain tumors by single-strandconformation polymorphism analysis of polymerase chain reactionproducts. Oncogene 1991;6:1313–8.

14. Henson JW, Schnitker BL, Correa KM, von Deimling A, Fassben-der F, Xu HJ, et al. The retinoblastoma gene is involved inmalignant progression of astrocytomas. Annals Neurol 1994;36:714–21.

15. Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, et al.An integrated genomic analysis of human glioblastoma multiforme.Science 2008;321:1807–12.

16. Ueki K, Rubio MP, Ramesh V, Correa KM, Rutter JL, von Deimling A,et al. MTS1/CDKN2 gene mutations are rare in primary human astro-cytomas with allelic loss of chromosome 9p. Hum Mol Genet1994;3:1841–5.

17. Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein DM, Ham-mond RR, et al. Specific genetic predictors of chemotherapeuticresponse and survival in patients with anaplastic oligodendrogliomas.J Natl Cancer Inst 1998;90:1473–9.

18. Paugh BS, Qu C, Jones C, Liu Z, Adamowicz-Brice M, Zhang J, et al.Integrated molecular genetic profiling of pediatric high-grade gliomasreveals key differences with the adult disease. J Clin Oncol 2010;28:3061–8.

19. Ueki K,OnoY,HensonJW,Efird JT, vonDeimlingA, LouisDN.CDKN2/p16 or RB alterations occur in the majority of glioblastomas and areinversely correlated. Cancer Res 1996;56:150–3.

20. Frederick L, Wang XY, Eley G, James CD. Diversity and frequency ofepidermal growth factor receptor mutations in human glioblastomas.Cancer Res 2000;60:1383–7.

21. Lee JC, Vivanco I, BeroukhimR,HuangJH, FengWL,DeBiasi RM, et al.Epidermal growth factor receptor activation in glioblastoma throughnovel missense mutations in the extracellular domain. PLoS Med2006;3:e485.

22. Wong AJ, Ruppert JM, Bigner SH, Grzeschik CH, Humphrey PA,Bigner DS, et al. Structural alterations of the epidermal growth factorreceptor gene in human gliomas. Proc Natl Acad Sci U S A 1992;89:2965–9.

23. KastenhuberER,HuseJT,BermanSH,PedrazaA, ZhangJ,SueharaY,et al. Quantitative assessment of intragenic receptor tyrosine kinasedeletions in primary glioblastomas: their prevalence and molecularcorrelates. Acta Neuropathol 2014;127:747–59.

24. Ozawa T, Brennan CW,Wang L, Squatrito M, Sasayama T, NakadaM,et al. PDGFRA gene rearrangements are frequent genetic events inPDGFRA-amplified glioblastomas. Genes Dev 2010;24:2205–18.

25. Paugh BS, Broniscer A, Qu C, Miller CP, Zhang J, Tatevossian RG,et al. Genome-wide analyses identify recurrent amplifications of recep-tor tyrosine kinases and cell-cycle regulatory genes in diffuse intrinsicpontine glioma. J Clin Oncol 2011;29:3999–4006.

26. Zarghooni M, Bartels U, Lee E, Buczkowicz P, Morrison A, Huang A,et al. Whole-genome profiling of pediatric diffuse intrinsic pontinegliomas highlights platelet-derived growth factor receptor alpha andpoly (ADP-ribose) polymerase as potential therapeutic targets. J ClinOncol 2010;28:1337–44.

27. Snuderl M, Fazlollahi L, Le LP, Nitta M, Zhelyazkova BH, Davidson CJ,et al. Mosaic amplification ofmultiple receptor tyrosine kinase genes inglioblastoma. Cancer Cell 2011;20:810–7.

28. Szerlip NJ, PedrazaA,ChakravartyD, AzimM,McGuire J, FangY, et al.Intratumoral heterogeneity of receptor tyrosine kinases EGFR andPDGFRA amplification in glioblastoma defines subpopulations withdistinct growth factor response. Proc Natl Acad Sci U S A 2012;109:3041–6.

29. Di Rocco F, Carroll RS, Zhang J, Black PM. Platelet-derived growthfactor and its receptor expression in human oligodendrogliomas.Neurosurgery 1998;42:341–6.

30. Gorovets D, Kannan K, Shen R, Kastenhuber ER, Islamdoust N,Campos C, et al. IDH mutation and neuroglial developmental featuresdefine clinically distinct subclasses of lower grade diffuse astrocyticglioma. Clin Cancer Res 2012;18:2490–501.

31. Westermark B, Heldin CH, Nister M. Platelet-derived growth factor inhuman glioma. Glia 1995;15:257–63.

32. Huse JT, Brennan C, Hambardzumyan D, Wee B, Pena J, RouhanifardSH, et al. The PTEN-regulatingmicroRNAmiR-26a is amplified in high-grade glioma and facilitates gliomagenesis in vivo. Genes Dev2009;23:1327–37.

33. Kim H, Huang W, Jiang X, Pennicooke B, Park PJ, Johnson MD.Integrative genome analysis reveals an oncomir/oncogene clusterregulating glioblastoma survivorship. Proc Natl Acad Sci U S A2010;107:2183–8.

34. McBride SM, Perez DA, Polley MY, Vandenberg SR, Smith JS, ZhengS, et al. Activation of PI3K/mTOR pathway occurs in most adult low-grade gliomas and predicts patient survival. J Neuro Oncol 2010;97:33–40.

35. Wiencke JK, Zheng S, JellumaN, Tihan T, Vandenberg S, Tamguney T,et al.Methylation of thePTENpromoter defines low-grade gliomas andsecondary glioblastoma. Neuro Oncol 2007;9:271–9.

36. Gutmann DH, Rasmussen SA, Wolkenstein P, MacCollin MM,Guha A, Inskip PD, et al. Gliomas presenting after age 10 inindividuals with neurofibromatosis type 1 (NF1). Neurology 2002;59:759–61.

37. Fuller GN, Rhee CH, Hess KR, Caskey LS, Wang R, Bruner JM, et al.Reactivation of insulin-like growth factor binding protein 2 expressionin glioblastoma multiforme: a revelation by parallel gene expressionprofiling. Cancer Res 1999;59:4228–32.

38. Karcher S, Steiner HH, Ahmadi R, Zoubaa S, Vasvari G, Bauer H, et al.Different angiogenic phenotypes in primary and secondary glioblas-tomas. Int J Cancer 2006;118:2182–9.

39. Kim S, Dougherty ER, Shmulevich I, Hess KR, Hamilton SR, Trent JM,et al. Identification of combination gene sets for glioma classification.Mol Cancer Ther 2002;1:1229–36.

40. LiangY,DiehnM,WatsonN,BollenAW,AldapeKD,NicholasMK, et al.Gene expression profiling reveals molecularly and clinically distinctsubtypes of glioblastoma multiforme. Proc Natl Acad Sci U S A2005;102:5814–9.

41. Rickman DS, Bobek MP, Misek DE, Kuick R, Blaivas M, Kurnit DM,et al. Distinctive molecular profiles of high-grade and low-gradegliomas based on oligonucleotide microarray analysis. Cancer Res2001;61:6885–91.





42. Sallinen SL, Sallinen PK, Haapasalo HK, Helin HJ, Helen PT, SchramlP, et al. Identification of differentially expressed genes in humangliomas by DNA microarray and tissue chip techniques. Cancer Res2000;60:6617–22.

43. Shai R, Shi T, Kremen TJ, Horvath S, Liau LM, Cloughesy TF, et al.Gene expression profiling identifies molecular subtypes of gliomas.Oncogene 2003;22:4918–23.

44. Tanwar MK, Gilbert MR, Holland EC. Gene expression microarrayanalysis reveals YKL-40 to be a potential serum marker for malignantcharacter in human glioma. Cancer Res 2002;62:4364–8.

45. Tso CL, Freije WA, Day A, Chen Z, Merriman B, Perlina A, et al. Distincttranscription profiles of primary and secondary glioblastoma sub-groups. Cancer Res 2006;66:159–67.

46. Huse JT, Phillips HS, Brennan CW. Molecular subclassification ofdiffuse gliomas: seeing order in the chaos. Glia 2011;59:1190–9.

47. Kim TM, Huang W, Park R, Park PJ, Johnson MD. A developmentaltaxonomy of glioblastoma defined and maintained by MicroRNAs.Cancer Res 2011;71:3387–99.

48. Shen R, Mo Q, Schultz N, Seshan VE, Olshen AB, Huse J, et al.Integrative subtype discovery in glioblastoma using iCluster. PLoSONE 2012;7:e35236.

49. Bhat KP, Balasubramaniyan V, Vaillant B, EzhilarasanR, Hummelink K,Hollingsworth F, et al. Mesenchymal differentiation mediated by NF-kappaB promotes radiation resistance in glioblastoma. Cancer Cell2013;24:331–46.

50. Patel AP, Tirosh I, Trombetta JJ, ShalekAK,GillespieSM,WakimotoH,et al. Single-cell RNA-seq highlights intratumoral heterogeneity inprimary glioblastoma. Science 2014;344:1396–401.

51. Dang L, White DW, Gross S, Bennett BD, Bittinger MA, Driggers EM,et al. Cancer-associated IDH1mutations produce 2-hydroxyglutarate.Nature 2009;462:739–44.

52. Griffin CA, Burger P, Morsberger L, Yonescu R, Swierczynski S,Weingart JD, et al. Identification of der(1;19)(q10;p10) in five oligoden-drogliomas suggests mechanism of concurrent 1p and 19q loss.J Neuropathol Exp Neurol 2006;65:988–94.

53. Jenkins RB, Blair H, BallmanKV,Giannini C, Arusell RM, LawM, et al. At(1;19)(q10;p10) mediates the combined deletions of 1p and 19q andpredicts abetter prognosis of patientswith oligodendroglioma.CancerRes 2006;66:9852–61.

54. Turcan S, Rohle D, Goenka A, Walsh LA, Fang F, Yilmaz E, et al. IDH1mutation is sufficient to establish the glioma hypermethylator pheno-type. Nature 2012;483:479–83.

55. Ohgaki H, Kleihues P. The definition of primary and secondary glio-blastoma. Clin Cancer Res 2013;19:764–72.

56. Ozawa T, Riester M, Cheng YK, Huse JT, Squatrito M, Helmy K, et al.Most human non-GCIMP glioblastoma subtypes evolve from acommonproneural-likeprecursorglioma.CancerCell2014;26:288–300.

57. Bettegowda C, Agrawal N, Jiao Y, Sausen M, Wood LD, Hruban RH,et al. Mutations in CIC and FUBP1 contribute to human oligodendro-glioma. Science 2011;333:1453–5.

58. Jiao Y, Killela PJ, Reitman ZJ, Rasheed AB, Heaphy CM, de Wilde RF,et al. Frequent ATRX, CIC, FUBP1 and IDH1 mutations refine theclassification of malignant gliomas. Oncotarget 2012;3:709–22.

59. Cairncross G, Wang M, Shaw E, Jenkins R, Brachman D, Buckner J,et al. Phase III trial of chemoradiotherapy for anaplastic oligodendro-glioma: long-term results of RTOG9402. JClinOncol 2013;31:337–43.

60. van den Bent MJ, Brandes AA, Taphoorn MJ, Kros JM, KouwenhovenMC, Delattre JY, et al. Adjuvant procarbazine, lomustine, and vincris-tine chemotherapy in newly diagnosed anaplastic oligodendroglioma:long-term follow-up of EORTC brain tumor group study 26951. J ClinOncol 2013;31:344–50.

61. Batchelor TT, Reardon DA, de Groot JF, Wick W, Weller M. Antiangio-genic therapy for glioblastoma: current status and future prospects.Clin Cancer Res 2014;20:5612–9.

62. Fecci PE, Heimberger AB, Sampson JH. Immunotherapy for primarybrain tumors: no longer a matter of privelege. Clin Cancer Res2014;20:5620–9.

63. Gajjar A, Pfister SM, Taylor MD, Gilbertson RJ. Molecular insights intopediatric brain tumors have the potential to transform therapy. ClinCancer Res 2014;20:5630–40.

64. HeaphyCM, deWilde RF, Jiao Y, Klein AP, Edil BH, Shi C, et al. Alteredtelomeres in tumors with ATRX and DAXX mutations. Science2011;333:425.

65. Kannan K, Inagaki A, Silber J, Gorovets D, Zhang J, Kastenhuber ER,et al. Whole-exome sequencing identifies ATRX mutation as a keymolecular determinant in lower-grade glioma. Oncotarget 2012;3:1194–203.

66. Killela PJ, Reitman ZJ, Jiao Y, Bettegowda C, Agrawal N, Diaz LA Jr,et al. TERT promoter mutations occur frequently in gliomas and asubset of tumors derived from cells with low rates of self-renewal. ProcNatl Acad Sci U S A 2013;110:6021–6.

67. Arita H, Narita Y, Fukushima S, Tateishi K, Matsushita Y, Yoshida A,et al. Upregulatingmutations in the TERT promoter commonly occur inadult malignant gliomas and are strongly associated with total 1p19qloss. Acta Neuropathol 2013;126:267–76.

68. Reitman ZJ, Pirozzi CJ, Yan H. Promoting a new brain tumor mutation:TERT promoter mutations in CNS tumors. Acta Neuropathol2013;126:789–92.

69. Esteller M, Garcia-Foncillas J, Andion E, Goodman SN, Hidalgo OF,VanaclochaV, et al. Inactivation of theDNA-repair geneMGMTand theclinical response of gliomas to alkylating agents. N Engl J Med2000;343:1350–4.

70. Hegi ME, Diserens AC, Godard S, Dietrich PY, Regli L, Ostermann S,et al. Clinical trial substantiates the predictive value of O-6-methyl-guanine-DNA methyltransferase promoter methylation in glioblas-toma patients treated with temozolomide. Clin Cancer Res2004;10:1871–4.

71. Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M,et al. MGMT gene silencing and benefit from temozolomide in glio-blastoma. N Engl J Med 2005;352:997–1003.

72. Herrlinger U, Rieger J, Koch D, Loeser S, Blaschke B, Kortmann RD,et al. Phase II trial of lomustine plus temozolomide chemotherapy inaddition to radiotherapy in newly diagnosed glioblastoma: UKT-03.J Clin Oncol 2006;24:4412–7.

73. Weller M, Felsberg J, Hartmann C, Berger H, Steinbach JP, SchrammJ, et al. Molecular predictors of progression-free and overall survival inpatients with newly diagnosed glioblastoma: a prospective transla-tional study of the German Glioma Network. J Clin Oncol 2009;27:5743–50.

74. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, TaphoornMJ, et al. Radiotherapy plus concomitant and adjuvant temozolomidefor glioblastoma. N Engl J Med 2005;352:987–96.

75. MalmstromA, Gronberg BH,Marosi C, Stupp R, Frappaz D, Schultz H,et al. Temozolomide versus standard 6-week radiotherapy versushypofractionated radiotherapy in patients older than 60 years withglioblastoma: the Nordic randomised, phase 3 trial. Lancet Oncol2012;13:916–26.

76. WickW, PlattenM,Meisner C, Felsberg J, Tabatabai G, SimonM, et al.Temozolomide chemotherapy alone versus radiotherapy alone formalignant astrocytoma in the elderly: the NOA-08 randomised, phase3 trial. Lancet Oncol 2012;13:707–15.

77. van den Bent MJ, Erdem-Eraslan L, Idbaih A, de Rooi J, Eilers PH,Spliet WG, et al. MGMT-STP27 methylation status as predictivemarker for response to PCV in anaplastic Oligodendrogliomas andOligoastrocytomas. A report from EORTC study 26951. Clin CancerRes 2013;19:5513–22.

78. Wick W, Meisner C, Hentschel B, Platten M, Schilling A, Wiestler B,et al. Prognostic or predictive value of MGMT promoter methylationin gliomas depends on IDH1 mutation. Neurology 2013;81:1515–22.

79. Bady P, Sciuscio D, Diserens AC, Bloch J, van den Bent MJ, Marosi C,et al. MGMT methylation analysis of glioblastoma on the Infiniummethylation BeadChip identifies two distinct CpG regions associatedwith gene silencing and outcome, yielding a prediction model forcomparisons across datasets, tumor grades, and CIMP-status. ActaNeuropathol 2012;124:547–60.

80. Weller M, Stupp R, Reifenberger G, Brandes AA, van den Bent MJ,Wick W, et al. MGMT promoter methylation in malignantgliomas: ready for personalized medicine? Nat Rev Neurol 2010;6:39–51.

CCRFOCUS




81. PreusserM, Charles Janzer R, Felsberg J, Reifenberger G, HamouMF,Diserens AC, et al. Anti-O6-methylguanine-methyltransferase (MGMT)immunohistochemistry in glioblastomamultiforme: observer variabilityand lack of association with patient survival impede its use as clinicalbiomarker. Brain Pathol 2008;18:520–32.

82. Johnson BE, Mazor T, Hong C, Barnes M, Aihara K, McLean CY, et al.Mutational analysis reveals the origin and therapy-driven evolution ofrecurrent glioma. Science 2014;343:189–93.

83. Hunter C, Smith R, Cahill DP, Stephens P, StevensC, Teague J, et al. Ahypermutation phenotype and somatic MSH6 mutations in recurrenthuman malignant gliomas after alkylator chemotherapy. Cancer Res2006;66:3987–91.

84. Balana C, Ramirez JL, Taron M, Roussos Y, Ariza A, Ballester R, et al.O6-methyl-guanine-DNAmethyltransferase methylation in serum andtumor DNA predicts response to 1,3-bis(2-chloroethyl)-1-nitrosoureabut not to temozolamide plus cisplatin in glioblastomamultiforme. ClinCancer Res 2003;9:1461–8.

85. Lavon I, Refael M, Zelikovitch B, Shalom E, Siegal T. Serum DNA candefine tumor-specific genetic and epigenetic markers in gliomas ofvarious grades. Neuro Oncol 2010;12:173–80.

86. Sullivan JP, Nahed BV, Madden MW, Oliveira SM, Springer S, BhereD, et al. Brain tumor cells in circulation are enriched for mesenchy-mal gene expression. Cancer Discov. 2014 Aug 19. [Epub ahead ofprint].





2014;20:5601-5611. Clin Cancer Res Jason T. Huse and Kenneth D. Aldape Management of Diffuse GliomaThe Evolving Role of Molecular Markers in the Diagnosis and

Updated version

http://clincancerres.aacrjournals.org/content/20/22/5601

Access the most recent version of this article at:

Cited articles

http://clincancerres.aacrjournals.org/content/20/22/5601.full#ref-list-1

This article cites 85 articles, 39 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/20/22/5601.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/20/22/5601To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/content/20/22/5601.full#ref-list-1

http://clincancerres.aacrjournals.org/content/20/22/5601.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



ccr focus - clinical cancer researchby gene ampliﬁcation, providing alternative mechanisms for p53...

Documents