CRESPO Et Al., 2015 - A1
CRESPO Et Al., 2015 - A1
CRESPO Et Al., 2015 - A1
-, - 2015
1 63
2 64
3 65
4 66
5 67
6 ajp.amjpathol.org 68
7 69
8 70
9
10
REVIEW 71
72
11 73
12
13
Molecular and Genomic Alterations in 74
75
14
15 Glioblastoma Multiforme 76
77
16 78
17 Q20 Q1 Ines Crespo,* Ana L. Vital,* María Gonzalez-Tablas,y María del Carmen Patino,z Alvaro Otero,x{ María C. Lopes,* 79
18 Catarina de Oliveira,* Patricia Domingues,*y{ Alberto Orfao,y{ and Maria D. Taberneroy{k 80
19 81
20 82
Q2 Q3 From the Centre for Neurosciences and Cell Biology,* Faculties of Pharmacy and Medicine, University of Coimbra, Coimbra, Portugal; the Department of
21 83
Medicine,y Centre for Cancer Research (CIC-IBMCC; CSIC/USAL; IBSAL), and the Department of Statistics,z University of Salamanca, Salamanca, Spain;
22 84
the Neurosurgery Servicex and the Instituto de Estudios de Ciencias de la Salud de Castilla y León and Research Laboratory,k University Hospital of
23 Salamanca, Salamanca, Spain; and the Instituto de Investigación Biomédica de Salamanca,{ Salamanca, Spain 85
24 86
25 87
26 Accepted for publication 88
February 9, 2015. In recent years, important advances have been achieved in the understanding of the molecular biology of
27 89
glioblastoma multiforme (GBM); thus, complex genetic alterations and genomic profiles, which recur-
28 Address correspondence to 90
Maria D. Tabernero, M.D., rently involve multiple signaling pathways, have been defined, leading to the first molecular/genetic
29 91
Ph.D., Research Laboratory, classification of the disease. In this regard, different genetic alterations and genetic pathways appear to
30 92
University Hospital of Sala- distinguish primary (eg, EGFR amplification) versus secondary (eg, IDH1/2 or TP53 mutation) GBM. Such
31 93
manca, Paseo San Vicente 58- genetic alterations target distinct combinations of the growth factor receptoreras signaling pathways, as
32 Q7 94
182, Salamanca, Spain. E-mail: well as the phosphoinositide 3-kinase/phosphatase and tensin homolog/AKT, retinoblastoma/cyclin-
33 95
[email protected]. dependent kinase (CDK) N2A-p16INK4A, and TP53/mouse double minute (MDM) 2/MDM4/CDKN2A-
34 96
p14ARF pathways, in cells that present features associated with key stages of normal neurogenesis and
35 97
(normal) central nervous system cell types. This translates into well-defined genomic profiles that have
36 98
been recently classified by The Cancer Genome Atlas Consortium into four subtypes: classic, mesenchymal,
37 99
proneural, and neural GBM. Herein, we review the most relevant genetic alterations of primary versus
38 100
secondary GBM, the specific signaling pathways involved, and the overall genomic profile of this
39 101
genetically heterogeneous group of malignant tumors. (Am J Pathol 2015, -: 1e14; http://dx.doi.org/
40 102
10.1016/j.ajpath.2015.02.023)
41 103
42 104
43 105
Q8 Glioblastoma multiforme (GBM) is a World Health Orga- alterations and genomic profiles (Table 1), supporting the Q9½T1
44 106
45 nization grade IV astrocytoma, which represents the most notion that the two groups of GBM arise through different 107
46 common and aggressive primary brain tumor. Most GBMs genetic pathways.23 Herein, we review current knowledge 108
47 are primary tumors that arise de novo as aggressive, highly about the signaling pathways most commonly involved in 109
48 invasive neoplasias, usually in the absence of clinical, GBM, the molecular and genetic alterations of primary and 110
49 radiological, or histopathological evidence of prior disease secondary GBM, including the clinical impact of such al- 111
50 and precedent lower-grade lesions; thus, approximately two- terations, and the most relevant gene expression profiling 112
51 113
thirds of patients with primary GBM have a clinical history of subgroups of these tumors.
52 114
53
<3 months,1,2 with rapid development of clinical symptoms. 115
54 By contrast, secondary GBMs are much less common and Oncogenic Pathways Involved in GBM 116
55 they derive from the transformation/progression of lower- 117
56 grade astrocytomas.1 Interestingly, these two subtypes of The many different genetic and molecular alterations present 118
57 glioblastoma also affect patients at different ages: primary in GBM lead to modifications of several major signaling 119
58 GBM is more common in older patients, whereas secondary 120
59 GBM tends to occur among patients <45 years. Primary Supported in part by FCT grant SFRH/BD/64799/2009 (P.D.) and by Q4
121
60 and secondary GBMs are usually indistinguishable on his- 122
IECSCYL (Soria, Spain) (M.D.T.). Q5 Q6
61 tological grounds, but they show clearly different genetic Disclosures: None declared. 123
62 124
Copyright ª 2015 American Society for Investigative Pathology.
Published by Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.ajpath.2015.02.023
REV 5.2.0 DTD AJPA2027_proof 10 May 2015 1:52 pm EO: AJP14_0661
Crespo et al
125 Table 1 Epigenetic and Genetic Alterations as Well as Gene/Protein Expression Profiles Typically Found in Primary versus Secondary 187
126 Glioblastomas 188
127 189
Variable Primary glioblastoma (95%) Secondary glioblastoma (5%) Reference
128 190
129 Promoter methylation 191
130 MGMT 36 75 3 192
131 TIMP-3 28 71 4 193
132 RB 14 43 5 194
133 CDKN2A-p14ARF 6 31 6 195
134 CDKN2A-p16INK4a 3 19 6 196
135 Genetic alterations 197
136 IDH1 mutation 5 67e85 7,8 198
137 IDH2 mutation 0 0 8,9 199
138 EGFR amplification 36e60 8 1 200
139 TERT mutation 58 28 1,10 201
140 CDKN2A-p16INK4a deletion 31e78 19 1 202
141 TP53 mutation 28 65 1 203
142 PTEN mutation 25 4 1 204
143 LOH 10p 47 8 10 205
144 LOH 10q 47; 70 54; 63 1,10 206
145 LOH 22q 41 82 4 207
146 LOH 1p 12 15 11 208
147 LOH 13q 12 38 11 209
148 LOH 19q 6 54 11 210
149 Gene/protein expression profiles 211
150 Fas (APO-1; CD95)* 100 21 12 212
151 Survivin* 83 46 13 213
152 MMP-9* 69 14 14 214
153 EGFR* 63 10 15 215
154 EGFRy High Low 16 216
155 MDM2* 31 0 17 217
156 VEGFz High Low 18 218
157 VEGF fms-related tyrosine kinase 1x High Low 19 219
158 IGFBP2x High Low 19 220
159 Tenascin-X-precursory High Low 16 221
160 Enolase 1y High Low 16 222
161 Centrosome-associated protein 350y High Low 16 223
162 TP53* 37 97 15 224
163 ASCL1{ 33 88 20 225
164 Loss of TIMP-3* 17 64 4 226
165 PDGFRA/PDGFRBz Low High 20,21 227
166 ERCC6y Low High 16 228
167 DUOX2y Low High 16 229
168 HNRPA3y Low High 16 230
169 WNT-11 protein precursory Low High 16 231
y
170 Cadherin-related tumor-suppressor homolog precursor Low High 16 232
171 ADAMTS-19y Low High 16 233
172 22
Modified from Ohgaki and Kleihues, with permission from the American Society for Investigative Pathology.
234
173 *Immunohistochemistry. 235
174 y 236
Two-dimensional protein gel electrophoresis.
175 z 237
Enzyme-linked immunosorbent assay.
x
176 cDNA array. 238
{
177 RT-PCR. 239
178 ADAMTS, ---; APO, ---; ASCL, ---; DUOX, ---; EGFR, epidermal growth factor receptor; ERCC, ---; HNRPA, ---; IGFBP, insulin- Q19
240
179 like growth factor binding protein; LOH, loss of heterozygosity; MDM, mouse double minute; MMP, matrix metalloproteinases; PDGFR, platelet-derived growth 241
180 factor receptor; TIMP, tissue inhibitor of metalloproteinases; VEGF, vascular endothelial growth factor; WNT, ---. 242
181 243
182 pathways that result in brain tumor growth and progres- interactions with additional unknown players, which poten- 244
183 ½F1 sion24,25 (Figure 1A). Despite the fact that the involvement of tially contribute to the initiation and transformation of 245
184 several well-known pathways in gliomagenesis is indubi- GBM.26 The most relevant signaling pathways involved in 246
185 table, there are complex interactions among them, including GBM include, among others, growth factor tyrosine kinase 247
186 248
249 receptor (TKR)etriggered pathways, including the Ras sar- coexist in individual GBM. The oncogenic properties of 311
250 Q10 coma (Ras) pathway, as well as the phosphoinositide 3- EGFR are associated with a constitutive and uncontrolled 312
251 kinase (PI3K)/phosphatase and tensin homolog (PTEN)/ increase in its phosphorylation (catalytic) activity. EGFR 313
252 Q11 314
AKT, retinoblastoma (RB)/cyclin-dependent kinase (CDK) gene mutations/rearrangements and expression of their
253 315
254
N2A-p16INK4a, and the TP53/mouse double minute 2 aberrant protein products are frequently observed in GBM 316
255 (MDM2)/MDM 4/CDKN2A-p14ARF pathways. so far; seven common variants have been identified, from 317
256 which variant 3 (EGFRvIII or del2-7 EGFR, DEGFR), 318
257 Growth Factor TKR Pathways which lacks a sequence of 267 amino acids in the extra- 319
258 cellular ligand-binding domain leading to a constitutively 320
259 Both the platelet-derived growth factor (PDGF) and activated EGFR and pathway, is the most frequent one (it is 321
260 epidermal growth factor (EGF) play an important role in present in 20% to 50% of GBMs that carry EGFR ampli- 322
261 normal and tumoral gliogenesis, through activation of com- fication). The introduction of this truncated receptor into 323
262 plex intracellular cascades modulated by G-proteinecoupled glioma cells dramatically enhances their tumorigenicity 324
263 receptors and second messengers that converge at multiple in vivo through both increased cellular proliferation and 325
264 326
sites. Overexpression of PDGF and EGF receptor (EGFR) in reduced apoptosis.33
265 327
266
GBM suggests that these TKR-signaling pathways are crit- Ligand-activated receptors trigger downstream signal 328
267 ical targets in gliomagenesis.27 transduction pathways, including the Ras/rapidly acceler- 329
268 The PDGF family consists of four different ligands ated fibrosarcoma (Raf)/mitogen-activated protein kinase 330
269 (PDGF-A, PDGF-B, PDGF-C, and PDGF-D) that signal (MAPK) pathway, the PI3K/AKT pathway, the protein ki- 331
270 through the PDGF receptor (PDGFR) a and PDGFRb.28 Both nase C pathway, and the STAT pathway, together with Q12 332
271 the PDGF ligands and receptors are often co-expressed in vascular endothelial growth factor production, with an 333
272 glioma cell lines and primary GBM tissues, suggesting the impact on cell proliferation, migration, invasion, resistance 334
273 establishment of both autocrine and paracrine signaling to apoptosis, and tumor neovascularization.34 335
274 loops, which may contribute to tumor formation and pro- 336
275 gression. PDGFRA and PDGFA are expressed in tumor cells, 337
276
The Ras Pathway 338
whereas PDGFB and PDGFRB have been typically found in
277 339
278
glioma-associated endothelial cells. Studies on the two new Ras is a guanosine-nucleotideebinding protein (G-protein), 340
279 PDGFR ligands PDGFC and PDGFD indicate that they may whose activation and deactivation is controlled by cycling 341
280 also play a role in the development of brain tumors. Because between the active GTP-bound and inactive GDP-bound Q13 342
281 co-expression of PDGF and PDGFR has been observed in forms. Receptor-mediated activation of Ras and Ras acti- 343
282 astrocytomas of all grades, PDGF autocrine signaling may be vation by oncogenic mutations are common in human tu- 344
283 considered as an early event, with additional secondary mors and contribute to the development and maintenance of 345
284 alterations in cell signaling being potentially required for the malignant phenotype. For example, the active Ras-GTP 346
285 progression to GBM. A subset of gliomas characterized by protein, through its downstream effectors, including the Raf, 347
286 dysregulated PDGFR activity (due to amplification and PI3K, and Ral-guanine nucleotide exchange factors, pro- 348
287 rearrangement of the PDGFRA gene locus and/or over- motes cell cycle progression, survival, and migration.35 349
288 350
expression of the PDGF ligand) has been described,29 and An increased Ras pathway activity is observed in virtu-
289 351
290
characterized by The Cancer Genome Atlas (TCGA; see ally all GBMs.36 Interestingly, RAS mutations are rarely 352
291 below).21 Interestingly, studies performed on different in vivo found in GBM. However, high levels of Ras-GTP have been 353
292 animal models have shown that PDGFA stimulates the pro- documented in GBM cell lines and primary tumors, sug- 354
293 liferation of PDGFRA-positive neural stem cells residing in gesting that this signaling pathway is activated by upstream 355
294 the subventricular zone of the adult murine brain, which are factors, such as TKR activation (eg, EGFR or PDGFR), 356
295 able to differentiate and generate glioma-like lesions,30 sug- which might be required for tumor induction, progression, 357
296 gesting a susceptibility of adult neural stem cells to oncogenic and maintenance. Ras-GTP is downstream of growth factor 358
297 transformation by PDGFRA alteration. receptors at a major signal transduction crossroad, trans- 359
298 EGFR and its ligands are variably expressed throughout lating extrinsic messages into the Raf-kinase/MAPK/ 360
299 the brain development from embryogenesis into adulthood, extracellular signaleregulated kinase pathway, PI3K/AKT, 361
300 362
suggesting a critical role of EGFR signaling in the prolif- or PI3K/Rac/Rho pathways, influencing cell proliferation,
301 363
302
eration, migration, differentiation, and survival of all types survival, and migration. In fact, MAPK/extracellular signale 364
303 of central nervous system cells and their precursors.31 In regulated kinase signaling, one of the downstream effectors 365
304 GBM cells, EGFR signaling may be activated in a ligand of Ras, is also increased in human GBM.34 In GBM, the 366
305 dependent or independent way, through overexpression of existence of mutations affecting the expression or the activity 367
306 both the ligand(s) and the receptor, leading to an autocrine of MAPKs has not been reported. However, because MAPKs 368
307 loop, and via genomic amplification of EGFR and/or mu- may contribute to cell proliferation and could be activated 369
308 tation of the receptor, leading to constitutive activation in in response to the inhibition of PI3K/mammalian target 370
309 the absence of ligand32; both types of alterations may of rapamycin, their role in gliomagenesis is currently under 371
310 372
373 435
374 436
375 437
376 438
377 439
378 440
379 441
380 442
381 443
382 444
383 445
384 446
385 447
386 448
387 449
388 450
389 451
390 452
391 453
392 454
393 455
394 456
395 457
396 458
397 459
398 460
399 461
400 462
401 463
402 464
403 465
print & web 4C/FPO
404 466
405 467
406 468
407 469
408 470
409 471
410 472
411 473
412 474
413 475
414 476
415 477
416 478
417 479
418 480
419 481
420 482
421 483
422 484
423 485
424 486
425 487
426 488
427 489
428 490
429 491
430 492
431 493
432 494
433 495
434 496
497 investigation. In addition, the Ras signaling pathway may also transcription of genes important for mitosis and, thus, pre- 559
498 be activated through the loss of function of neurofibromatosis venting progression through the G1/S cell cycle checkpoint. 560
499 type I (NF1), a negative regulator of Ras. Germline and loss- In proliferating cells, growth factors lead to the induction of 561
500 562
of-function NF1 mutations have been identified in neurofi- cyclin D1, as well as to the activation of CDK2/cyclin E
501 563
502
bromatosis type I, which includes GBM as part of the clinical through degradation of its inhibitor, p27Kip1. These activated 564
503 spectrum. In this regard, the results of TCGA Consortium CDK-cyclin complexes trigger phosphorylation of RB in late 565
504 showed that approximately 20% of GBMs harbor loss-of- G1 phase, which is maintained later in the S, G2, and M 566
505 function mutations of the NF1 gene.24 phases. Phosphorylated (inactive) RB enables E2F release, 567
506 leading to transcriptional activation of growth-promoting 568
507 The PI3K/PTEN/AKT Pathway genes, required for DNA synthesis and cell growth.37 569
508 Negative regulators of the RB signaling pathway include 570
509 Elevated signaling through the PI3K-mediated cell signaling the Ink4 family of proteins (CDKN2A-p16INK4a, CDKN2B- 571
510 pathway has been implicated in the pathogenesis of GBM. p15INK4b, CDKN2C-p18INK4c, and CDKN2D-p19INK4d) that 572
511 Recruitment of PI3K to the cell membrane activates down- compete with the D-cyclins for CDK4/6, to prevent the for- 573
512 574
stream effector molecules, such as AKT and mammalian mation of the active kinase complex that phosphorylates
513 575
514
target of rapamycin, resulting in cell proliferation and RB.41 576
515 increased cell survival by blocking apoptosis.37 The PTEN Inactivation of the RB pathway through disrupted RB- 577
516 tumor-suppressor gene negatively regulates PI3K, but its E2F interaction may occur by mutation of the RB gene itself, 578
517 function is frequently lost in GBM because of loss of het- by loss of RB expression, or by the inactivation of RB by 579
518 erozygosity (LOH) at the 10q23.3 locus or because of gene phosphorylation through CDK/cyclin complexes. CDKs can 580
519 mutations (15% to 40%),38 causing constitutive activation of be activated by an increase in the amount of the catalytic 581
520 the PI3K pathway and higher levels of activated AKT in enzyme, by an increase in the amount of their cofactors 582
521 glioma cells. Some reports also indicate that PTEN plays a (cyclins), and/or by a decrease in the amount of endogenous 583
522 significant role in inducing G1 cell cycle arrest and apoptosis, CDK inhibitors (eg, CDKN2A-p16INK4A). In addition, the 584
523 along with regulation of cell differentiation. Tumors with an CDKN2A genetic locus at chromosome 9p21 produces both 585
524 586
activated PTEN/AKT pathway may be sensitive to mamma- CDKN2A-p14ARF and CDKN2A-p16INK4a by alternative
525 587
526
lian target of rapamycin inhibitors (eg, rapamycin).39 Several splicing.42 Because p16INK4a negatively regulates CDK4 588
527 studies using viral vectors have shown that wild-type PTEN and p14ARF inhibits MDM2, leading to a rapid blockade of 589
528 suppresses tumorigenicity of glioma cells and inhibits the the ubiquitin-mediated decay of TP53, simultaneous inac- 590
529 PI3K/AKT pathway.40 Thus, the PI3K/PTEN/AKT pathway tivation of both genes by homozygous deletion dysregulates 591
530 appears to play a role in gliomagenesis, as also supported by both the RB and the TP53 pathways.43 592
531 the findings of the pilot project of the TCGA, which identified The RB pathway is altered in 78% of 206 primary GBMs,34 593
532 alterations in the EGFR/Ras/NF1/PTEN/PI3K pathway in either directly by mutations, deletions, or promoter methyl- 594
533 88% of all GBM patients.24 ation at the RB locus or indirectly through alterations on the RB 595
534 positive and negative regulators. In turn, RB promoter 596
535 597
536
The RB/CDKN2A-p16INK4a Pathway methylation and gene silencing are more frequently found in
598
secondary (43%) than in primary (14%) GBMs.44 Alterations
537 599
538
This pathway plays a central role in the regulation of cell in the RB pathway include homozygous deletion and mutation 600
539 cycle and cell proliferation, because its components are of CDKN2A-p16INK4a and RB in 52% and 11% of the samples, 601
540 activated and/or inhibited by growth-promoting as well as respectively, and homozygous deletion of CDKN2B-p15INK4b 602
541 growth-suppressing signals.41 The RB gene (13q14) encodes and CDKN2C-p18INK4c in 47% and 2% of the tumors, 603
542 the RB phosphoprotein; in quiescent cells, RB is in a hypo- respectively. In turn, CDK4, CDK6, and CCND2 (cyclin D2) 604
543 phosphorylated state (active) bound to E2F, preventing gene amplification has been found in 18%, 1%, and 2% of 605
544 606
545 Figure 1 Schematic representation of the different genetic abnormalities and the major signaling pathways involved in the pathogenesis of human
607
546 glioblastoma. A: The interactions between the major signaling pathways altered in primary glioblastoma multiforme (GBM; dark gray) and secondary GBM 608
547 (light gray). Genes that are inactivated or hyperactivated by different mechanisms are shown in green and red, respectively. In the phosphoinositide 3-kinase Q16 609
548 (PI3K)/phosphatase and tensin homolog (PTEN)/Akt signaling pathway, the growth factor receptor becomes activated and recruits PI3K to the cell membrane, 610
549 converting phosphatidylinositol-4,5-bisphosphate to the PIP3 second messenger molecule. Downstream effector molecules, such as AKT and the mammalian Q17 611
550 target of rapamycin (mTOR), are then activated, which help to induce cell proliferation and block apoptosis. PTEN terminates the PIP3 signal; inactivation 612
551 leads to increased availability of PI3K, and PI3K activates AKT, which, in turn, leads to an increased proliferative activity and survival. In the TP53 pathway, 613
552 TP53 mutation or increased degradation of TP53 because of increased mouse double minute (MDM) 2 activity interrupts the normal cell cycle arrest and 614
553 apoptosis after DNA damage. In the retinoblastoma protein (RB) pathway, cyclin-dependent kinase (CDK) 4 and CDK6 form complexes with members of the 615
cyclin D family and phosphorylate the RB, which releases the E2F transcription factor, inducing cell proliferation by transcription of genes that promote DNA Q18
554 616
synthesis. Loss of RB leads to elevated levels of E2F and, therefore, cell proliferation. B: The most frequent and relevant molecular abnormalities of primary
555 versus secondary GBM are shown. Chr, chromosome; EGF, epidermal growth factor; EGFR, EGF receptor; ERK, extracellular signaleregulated kinase; HGF,
617
556 hepatocyte growth factor; LOH, loss of heterozygosity; NF1, neurofibromatosis type I; PDGF, platelet-derived growth factor; PDGFR, PDGF receptor; TGF, 618
557 transforming growth factor; VEGF, vascular endothelial growth factor; WHO, World Health Organization. 619
558 620
621 GBMs, respectively.24 Activated AKT deregulates cell growth MDM2/MDM4/CDKN2A-p14ARF pathway in glioblas- 683
622 by stabilization of cyclin D and promotion of nuclear entry tomas of 87%, with such alterations being associated with 684
623 of MDM2, leading to degradation of TP53. AKT might TP53 mutation or homozygous deletion in 35% of cases, 685
624 686
also inhibit p21 expression through its phosphorylation and MDM2 amplification in 14%, MDM4 amplification in 7%,
625 687
626
activation of MDM2. In addition, activated AKT exerts an and CDKN2A-p14ARF homozygous deletion or mutation in 688
627 anti-apoptotic activity by phosphorylating and inactivating 49% of all GBM cases.24 Amplification of MDM4 was 689
628
Q14 pro-apoptotic signaling proteins (eg, BAD and caspase 9), and detected in 4% of GBM with neither TP53 mutation nor 690
629 it may contribute to tumor invasion and metastasis by stimu- MDM2 amplification.49 691
630 lating secretion of matrix metalloproteinases.45 692
631 693
Epigenetic Changes in GBM
632 The TP53/MDM2/MDM4/CDKN2A-p14ARF Pathway 694
633 695
634 Gliomas present a variety of epigenetic alterations that 696
The TP53 protein, coded by the TP53 gene at chromosome induce changes in normal gene expression, without altering
635 17p13.1, plays a role in the cell cycle, cellular responses to 697
636 the DNA sequence. Aberrant epigenetic mechanisms, such 698
DNA damage, cell death, and differentiation. It is a as DNA methylation, histone modifications, chromatin
637 699
638
sequence-specific nuclear transcription factor that binds as a remodeling, or altered noncoding RNA expression (eg, 700
639 tetramer to defined consensus sites within the DNA, miRNAs), are currently recognized as relevant events in 701
640 affecting the transcription of its target genes, by either tumor formation, in addition to classic genetic alterations.50 702
641 transcriptional activation or modulation of the activation of Until now, most studies about the epigenetic changes of 703
642 other proteins through direct binding.43 The TP53 network glioblastoma have focused on DNA methylation, including 704
643 is activated in response to cellular stress conditions and hypermethylation of CpG islands (associated with tumor-
705
644 facilitates DNA repair or induces cell death in case of too 706
645
suppressor gene silencing), gene-specific hypomethylation 707
much damage, preventing cells with mutated or damaged (resulting in aberrant activation of oncogenes), and genome-
646 DNA from dividing. TP53 transcriptionally regulates the 708
647 wide hypomethylation (potentially leading to chromosomal 709
promoters of potential effector genes, such as p21, which instability, loss of imprinting, and uncontrolled cell prolif-
648 710
blocks cell cycle progression at the G1 phase by binding and eration).51 To date, multiple changes in the DNA methyl-
649 711
650
inhibiting the function of cyclin-D proteins.46 This gives ation pattern of promoters of genes involved in cell cycle 712
651 time for DNA repair before replication or mitosis and, regulation (eg, CDKN2A-p16INK4a and CDKN2B-p15INK4b), 713
652 thereby, links p21 directly to the tumor-suppressor function tumor suppression (eg, RB, VHL, EMP3, RASSF1A, and 714
653 of TP53. Moreover, the TP53 and the RB pathways interact BLU), DNA repair, and genome integrity (eg, MGMT and 715
654 with each other via p21. hMLH1), as well as genes associated with regulation of 716
655 Recent findings show that TP53 also regulates prolifera- tumor invasion and inhibition of apoptosis (eg, DAPK1,
717
656 tion, differentiation, and survival of stem cells, further 718
TIMP3, CDH1, PCDHGA11, and TMS1/ASC),5,52e55 have
657 highlighting the relevance of TP53 in suppressing GBM.47 been reported in GBM.
719
658 After stress, the activity of TP53 is blocked by its nega- 720
659 tive regulator MDM2, whose transcription is induced by 721
660 Genetic Alterations of Primary versus 722
TP53, generating a negative feedback loop that regulates the
661 723
662
activity of TP53 and the expression of MDM2.43 In addi- Secondary GBM 724
663 tion, activation of TP53 is achieved through inactivation of 725
664 MDM2 by CDKN2A-p14ARF binding. The MDM4 TP53 Overall, primary GBMs typically harbor three predominant 726
665 binding protein homolog (also called MDMX) also regu- genetic alterations, as confirmed also by high-density single- 727
666 lates TP53 activity, and p14ARF is negatively regulated by nucleotide polymorphism arrays56: amplification and/or a 728
667 TP53. high rate of EGFR mutation in chromosome 7p, homozygous 729
668 The TP53 signaling pathway is disrupted in GBM due to deletion of the CDKN2A-p16INK4a gene in chromosome 9p 730
669 TP53 mutation and/or amplification, overexpression of (with absence of the CDKN2A-p16INK4a a and/or the 731
670 MDM2, and/or loss of expression of CDKN2A-p14ARF, all CDKN2A-p14ARF b transcripts), and deletion of the PTEN 732
671 such alterations blocking TP53 activity and potentially gene typically in association with monosomy 10.56e58 In 733
672 734
leading to uncontrolled cell proliferation and tumor forma- addition, amplification of the MDM2 oncogene is also pre-
673 735
674
tion. In human gliomas, TP53 mutations are missense mu- sent in a smaller, but significant, percentage of primary 736
675 tations and target the highly conserved domains of TP53 in GBMs (<15%), particularly among those primary glioblas- 737
676 exons 5, 7, and 8, which are crucial for DNA binding. At tomas that lack TP53 mutations and telomerase reverse 738
677 least one alteration in the TP53/MDM2/CDKN2A-p14ARF transcriptase (TERT) promoter mutations59,60 (Figure 1B). 739
678 pathway has been reported to occur in approximately 50% Amplification and/or mutation of the EGFR gene 740
679 of primary glioblastomas and in >70% of secondary glio- (7p11.2) occur in 36% to 60% of primary GBMs. The most 741
680 blastomas.48 Consistently, the TCGA pilot project showed common EGFR mutant type, variant 3 (EGFRvIII), is due to 742
681 an overall frequency of genetic alterations in the TP53/ an 801-bp in-frame deletion of exons 2 to 7, and it leads to a 743
682 744
745 constitutively active protein, resulting in increased prolif- frequently associated with G:C/A:T mutations of the TP53 807
746 eration and survival of mutated cells.33 In most GBM tu- gene.3,65 In addition, isocitrate dehydrogenase 1 (IDH1) and 808
747 mors, overexpression of the EGFRvIII mutant coexists with IDH2 mutations have been identified recently, as early ge- 809
748 810
EGFR gene amplification, but overexpression of the netic alterations present in most low-grade gliomas, as well as
749 811
750
EGFRvIII mutant without EGFR amplification has also in the pathway to secondary glioblastomas. However, they 812
751 been reported in a small proportion of primary GBMs.61 All are rare among primary GBMs.7 In gliomas, IDH1/2 muta- 813
752 primary glioblastomas with EGFR amplification show tions are associated with an increased DNA hyper- 814
753 EGFR overexpression, and 70% to 90% of cases with EGFR methylation profile66; therefore, it is likely that IDH1/2 815
754 overexpression have EGFR amplification. mutations are involved in oncogenesis through inactivation 816
755 Homozygous deletion of CDKN2A-p16INK4a is also more of tumor-suppressor genes via hypermethylation of their 817
756 frequently observed among primary than secondary glio- promoters, which is facilitated by the mutated gene/protein. 818
757 blastomas, whereas other alterations of this gene predominate Gliomas with mutated IDH1 and IDH2 have a better prog- 819
758 in the latter tumors.6 Chromosome 9p deletions involving the nosis compared with gliomas with wild-type IDH.8 Inacti- 820
759 CDKN2A-p16INK4a gene typically target other closely vation of the proapoptotic Harakiri (HRK) gene, which 821
760 822
located genes, such as the methylthioadenosine phosphory- encodes for a bcl2-interacting protein at 12q24,67 because of
761 823
762
lase (MTAP) gene, whose relevance in GBM remains to be hypermethylation of its promoter, could play a critical role in 824
763 established.62 the development and progression of secondary GBM via 825
764 LOH of chromosome 10 is present in up to 70% of primary abrogation of apoptosis in tumors expressing wild-type TP53. 826
765 glioblastomas,1,63 suggesting the presence of several tumor- Partial LOH of chromosome 10q is also a frequent event 827
766 suppressor genes in this chromosome, which may be in secondary glioblastomas.10 LOHs of chromosome 13q 828
767 involved in GBM tumorigenesis. Three main chromosomal (typically including the RB locus), chromosome 19q (with a 829
768 regions are commonly deleted: 10p14-15, 10q23-24 (PTEN), common deleted region at 19q13.3),11 and chromosome 22q 830
769 and 10q25-pter. Interestingly, PTEN gene mutations have (including the TIMP-3 putative tumor-suppressor gene at 831
770 been reported in approximately 25% of all glioblastomas,22 22q12.3)4 have been detected more frequently in secondary 832
771 with such mutations being almost exclusively detected versus primary GBM. Less commonly than in primary 833
772 834
among primary glioblastoma1; by contrast, PTEN homozy- GBMs, secondary GBMs also display amplification of the
773 835
774
gous deletion occurs rarely. More than half of PTEN mutations PDGFRA and/or PDGFRB genes, potentially in association 836
775 result in the introduction of premature stop codons, leading to with overexpression of the PDGF ligand, yielding to an 837
776 translation of truncated proteins. Alterations involving the autocrine loop in GBM tumor cells. Table 1 summarizes all 838
777 PTEN gene rarely occur in low-grade gliomas, whereas the of the above described and other genetic alterations and 839
778 frequency of inactivation of the PTEN gene increases in the altered gene/protein expression profiles associated with 840
779 more advanced stages of the disease (Figure 1B). Interestingly, primary versus secondary GBM.12e20 841
780 primary GBMs frequently show loss of chromosome 10 in 842
781 association with EGFR amplification, raising the possibility 843
782 that the interaction between EGFR signaling and functional Genetic and Molecular Alterations of GBM with 844
783 abrogation of relevant chromosome 10 suppressor genes may Clinical Impact 845
784 846
contribute to the aggressive features of GBM.
785 In the past decades, several alterations reported in GBM have 847
786
On the basis of the pilot project of the TCGA Consortium 848
and other studies, additional genetic alterations have been emerged as being of potential clinical relevance for more
787 849
identified in primary GBM, which include NF1 mutation/ accurate diagnostic classification, prognostic stratification,
788 850
789 homozygous deletion (18%) and PIK3R1 (regulatory sub- and/or prediction of response to therapy in GBM. Among 851
790 unit 1 of phosphatidylinositol 3-kinase) mutations (10%).24 others, amplification of the EGFR gene, IDH1/IDH2 muta- 852
791 In turn, mutation and/or amplification of the PIK3CA gene, tions, and hypermethylation of the promoter of the O(6)- 853
792 which codes for the catalytic subunit a of the phosphati- methylguanine-DNA methyltransferase (MGMT) gene have 854
793 dylinositol-4,5-bisphosphate 3-kinase, is a rare event in both become particularly relevant from the clinical point of view, 855
794 primary and secondary GBMs (approximately 5% and in addition to detection of codeletion of chromosomes 1p and 856
795 approximately 13%, respectively).64 19q in low-grade gliomas. In turn, gene expression profiles 857
796 (GEPs) are still not yet being generally used. 858
In contrast to primary GBM, TP53 mutations at chromo-
797 859
798
some 17p are most frequently seen in secondary GBM, 860
799 mainly involving codons 248 and 273 and G:C/A:T mu- EGFR Amplification and EGFR Genetic Variants 861
800 tations at CpG (cytosine-phospho-guanine) sites. TP53 mu- 862
801 tations in secondary GBM are already detectable at the early EGFR is a transmembrane glycoprotein that acts as a TKR. 863
802 stages of the disease,1 and they seem to be an early event Once bound to its ligand, EGFR becomes autophosphorylated 864
803 associated with malignant transformation in the pathway to and induces subsequent activation of signal transduction 865
804 secondary GBM (Figure 1B). Interestingly, methylation pathways involved in the regulation of cell proliferation, 866
805 of the promoter of the MGMT DNA repair gene is also differentiation, and survival. Although present in normal 867
806 868
869 cells, the EGFR gene is the most frequently overexpressed Parsons et al80 revealed, by gene sequencing of 22 GBMs, 931
870 gene, mainly in primary GBM,1,22 and its alterations in gli- the presence of recurrent (12%) point mutations in the active 932
871 omas are highly associated with high-grade malignancy.68 site of IDH1, a gene that had never been linked to cancer 933
872 934
However, controversial results exist regarding the prog- before. Furthermore, it was found that other GBMs, which do
873 935
874
nostic significance of EGFR amplification in GBM. Although not have IDH1 mutations, may show mutation of the IDH2 936
875 some studies reported that EGFR amplification is associated gene.8 IDH1 (coded at chromosome 2q33) and IDH2 (coded 937
876 with a poor prognosis and a shorter survival of GBM pa- at chromosome 15q26) mutations typically consist of mon- 938
877 tients,69 others claim that such association with survival oallelic, somatic, and missense changes. Mutations of IDH1 939
878 would not be significant or that it could even be associated almost always affect the R132 codon,80 and IDH2 mutations 940
879 with a better outcome.70 In turn, some reports have found a exclusively affect the R172 and R140 codons.8 Both IDH1 941
880 poor prognosis for GBM patients carrying amplification of the and IDH2 mutations are more frequently detected among 942
881 EGFR gene among all age groups, whereas others have found grade II to III gliomas and secondary glioblastoma (70% to 943
882 EGFR amplification to be a predictor for prolonged survival 75%), whereas they are rare in primary glioblastoma 944
883 only among older patients. Such discrepant results could (5%).7,81 Interestingly, the occurrence of IDH1 mutations in 945
884 946
potentially be associated with the specific underlying genetic diffuse gliomas is strongly associated with TP53 mutation
885 947
886
lesion targeting the EGFR gene. In this regard, the prognostic and also del(1p)/del(19q), indicating that they may represent 948
887 impact of EGFRvIII (the most common EGFR mutant variant, an early event.82 Both younger age and mutated TP53 have 949
888 which leads to constitutive activation of EGFR in a ligand- been reported to be associated with a better outcome among 950
889 independent way) has not been investigated as extensively GBM patients. Noteworthy, IDH1 mutations are present in a 951
890 as EGFR amplification. Despite this, controversial results higher fraction of younger patients, who also carry a high 952
891 have also emerged. Accordingly, the EGFRvIII variant has frequency of TP53 mutations, all being typical features of 953
892 been found not to be related to patient outcome,71 to be secondary GBM.9,83 Other genetic associations that have 954
893 associated with an unfavorable prognosis,72 or even to be a been reported in gliomas include IDH1 mutation and MGMT 955
894 molecular predictor for prolonged overall survival among promoter methylation, in the absence of chromosome 10 956
895 (conventionally) treated GBM patients.73 Some other poly- losses and EGFR amplification.84 Altogether, these associa- 957
896 958
morphic genetic variants of the EGFR gene (eg, the -191C/A tions might contribute to explain why IDH1/IDH2 mutations
897 959
898
polymorphism involving the promoter region of the gene) have been found to be positive prognostic factors.9,84 960
899 have also been associated with the prognosis of glioma pa- Efforts are ongoing to better understand the role of IDH1 961
900 tients,74 but its clinical value still deserves confirmation. mutations in gliomagenesis. Initially, it was postulated that 962
901 In addition, it has also been speculated that the EGFR mutations could abrogate the function of the protein, leading 963
902 status could contribute to predict response to EGFR-targeted to reduced synthesis of the a-ketoglutarate metabolite,85 but 964
903 therapies. In fact, several mono and multiple EGFR-targeted striking new pieces of evidence have shown that IDH1 965
904 therapies have been developed and assayed in recent years, mutants also confer an enzymatic gain-of-function pheno- 966
905 including antibodies and small-molecule inhibitors of EGFR. type, associated with production of the alternative metabo- 967
906 Among other agents, two small-molecule EGFR inhibitors lite, 2-hydroxyglutarate; because 2-hydroxyglutarate may 968
907 (gefitinib and erlotinib) have been extensively tested, but the contribute to gliomagenesis,86 a new hypothesis has been 969
908 970
response to these drugs (eg, erlotinib as a single agent) did not raised about whether treatments that would reduce produc-
909 971
910
seem to be effective.75 Other EGFR-directed therapies, such tion of 2-hydroxyglutarate could be effective against gli- 972
911 us RNA-based treatment approaches, ligand-toxin conju- omas with IDH1 mutation. 973
912 gates, and radioimmunoconjugates, are currently at various 974
913 stages of evaluation for clinical purposes.33,76 In parallel, Hypermethylation of the MGMT Promoter 975
914 efforts are being made to increase the knowledge about the 976
915 mechanisms underlying resistance to EGFR-targeted thera- The MGMT gene (coded at chromosome 10q26) is frequently 977
916 pies and to clarify the clinical value of EGFR amplification silenced in GBM by promoter hypermethylation in associa- 978
917 and its variants (eg, EGFRvIII) for prognostic stratification of tion or not with monosomy 10/del(10q). Currently, this 979
918 GBM.61 represents one of the most relevant prognostic factors in 980
919 GBM and a potent predictor of response to treatment with 981
920 982
IDH1/IDH2 Mutations alkylating agents.87 Accordingly, an association between
921 983
922
methylation of the MGMT gene promoter and response to 984
923 IDHs catalyze the oxidative decarboxylation of isocitrate to alkylating chemotherapy using nitrosourea compounds, 985
924
Q15 a-ketoglutarate and reduce NADþ and NADPþ to NADH temozolomide (TMZ), or a combination of both88 has been 986
925 and NADPH, respectively.77 IDH1 is localized in the observed in GBM. Several alkylation sites have been 987
926 cytoplasm and the peroxisomes, whereas IDH2 is in the described in the DNA, as targets for these cytotoxic com- 988
927 mitochondria,78 where they are involved in the tricarboxylic pounds, the most frequent site being the O6-position of 989
928 acid cycle, as well as in protection against oxidative stress. guanine. The use of alkylating chemotherapeutic drugs, such 990
929 IDH1 participates in the lipid and glucose metabolism.79 as TMZ, causes binding of an alkyl group to the O6-position 991
930 992
993 of guanine, thereby inducing a DNA mismatch, DNA subtypes of gliomas, and also genetic subgroups of GBM, 1055
994 double-strand breakage, and ultimately apoptosis of prolif- show unique GEP at the mRNA transcriptional level that may 1056
995 erating cells. Thus, the MGMT protein counteracts the nor- potentially be used to distinguish among them. 1057
996 1058
mally lethal effect of TMZ by repairing DNA damage. When Despite this, discordant results between morphological
997 1059
998
a tumor has a hypermethylated MGMT promoter, transcrip- and molecular tumor subtypes, as defined by GEP, have been 1060
999 tion of the gene is blocked, leading to lack of MGMT mRNA also frequently observed among histologically ambiguous 1061
1000 and protein expression, while enhancing the cytotoxic effects gliomas. In this regard, the potential diagnostic utility of GEP 1062
1001 of the alkylating drug(s). for the molecular classification of gliomas has been high- 1063
1002 Previous studies have shown that patients with hyper- lighted by Nutt et al,97 who identified two sets of genes 1064
1003 methylation of the MGMT gene promoter may have longer significantly associated with classic GBM or anaplastic oli- 1065
1004 survival rates when treated with both TMZ and radiotherapy godendroglioma, and built a class prediction model that 1066
1005 (median survival at 2 and 5 years of 49% and 14%, respec- showed an accuracy of 86% in assigning diagnostically 1067
1006 tively) instead of radiotherapy alone (median survival at 2 and challenging samples. In another collaborative study, Parsons 1068
1007 5 years of 24% and 5%, respectively).83,89,90 However, it has et al80 discovered a variety of genes that, at that time, were 1069
1008 1070
been recently reported that methylation of the MGMT pro- unknown to be altered in GBM (eg, recurrent mutations of the
1009 1071
1010
moter could also be predictive of response to radiotherapy and IDH1 gene were detected in 12% of the cases). 1072
1011 a longer survival in GBM, in the absence of adjuvant In addition, several GEP studies have identified distinct 1073
1012 chemotherapy.91 Therefore, it remains unclear whether molecular subtypes of gliomas with prognostic significance. 1074
1013 MGMT also plays a role in repairing radiotherapy-induced Early studies by Freije et al98 identified 595 differentially 1075
1014 DNA damage, whether other DNA repair genes than MGMT expressed genes that correlated with overall survival in 74 1076
1015 are also silenced by promoter hypermethylation, or whether gliomas. Such molecular subtypes [hierarchical clustering 1077
1016 the survival advantage of such MGMT hypermethylated cases (HC) 1A, HC1B, HC2A, and HC2B] could be further 1078
1017 is better explained by other coexisting prognostically favor- segregated into two groups of patients with distinct survival 1079
1018 able genetic features frequently observed in low-grade oligo- (median overall survival of 4.8 and 0.6 years) through HC. 1080
1019 dendrogliomas and secondary GBM, such as codeletion of Nutt et al97 showed that the prognostic impact of the survival 1081
1020 1082
chromosomes 1p and 19q92 and IDH1 mutation.84 cluster on the basis of tumor GEP was independent of patient
1021 1083
1022
Despite its recognition as a favorable prognostic marker, age and histological grade. Interestingly, this study also 1084
1023 MGMT testing is still not widely used in the day-to-day showed that one subtype was enriched for genes involved in 1085
1024 laboratory routine of most centers; this is mainly due to neurogenesis, whereas the poor survival subtypes were 1086
1025 the lack of technical standardization65,93 because of sample enriched for genes associated with cell proliferation and 1087
1026 heterogeneity and lack of precise identification/validation of extracellular matrix/invasion, suggesting a more versus less 1088
1027 specific MGMT promoter regions predictive of patient differentiated phenotype for the two groups, respectively. 1089
1028 response to therapy. In addition, several studies indicate that Later, Phillips et al99 subdivided GBM into three groups 1090
1029 the pattern of expression of the MGMT protein is associated [eg, mesenchymal (MQ), proliferative (PF), and proneural 1091
1030 with neither the methylation status of the MGMT promoter (PN)], according to the similarities observed between 1092
1031 nor the patient outcome, suggesting that other molecular distinct tumoral GEP and both key stages in neurogenesis 1093
1032 1094
mechanisms, in addition to MGMT promoter methylation and known cell types. MQ, PF, and PN signatures were
1033 1095
1034
and del(10q)/monosomy 10, may be involved.94 associated with those of neural stem cells, transit-amplifying 1096
1035 cells, and immature neurons, respectively. The PN signature 1097
1036 is typically overrepresented among less aggressive forms of 1098
1037 GEP of GBM and Its Molecular/Genomic high-grade gliomas, whereas the MQ and PF subtypes were 1099
1038 Subtypes enriched in more aggressive high-grade tumors; in addition, 1100
1039 such classification also showed an association with clinical 1101
1040 In the past decade, efforts have been made to further define the outcome. A subsequent meta-analysis on 267 GBMs, on the 1102
1041 impact of genetic/molecular alterations of GBM at the genomic basis of previously published data together with new GEP 1103
1042 level. Recent availability of high-throughput microarray-based data,100 identified 377 differentially expressed genes, which 1104
1043 assays allowed for genome-wide quantitative analysis of classified GBM into four distinct subtypes by HC analyses: 1105
1044 1106
thousands of genes simultaneously, providing new insights HC1A/PN, HC2A/PF, HC2B/MQ, and a fourth subtype
1045 1107
1046
into patterns of overexpressed and underexpressed genes with HC2A/HC2B hybrid features, termed PF/MQ. Survival 1108
1047 potentially involved in malignant transformation and disease analysis confirmed the more favorable outcome of HC1A/ 1109
1048 progression. Early GEP studies identified differentially PN GBM versus the other three subtypes. Moreover, 1110
1049 expressed genes among histopathologically defined subtypes younger GBM patients more frequently had PN GBM type 1111
1050 of gliomas (eg, GBM versus oligodendrogliomas,95 low- tumors, which could contribute to explain their longer life 1112
1051 versus high-grade astrocytomas,95 primary versus secondary expectancy and their better outcome. 1113
1052 GBM,96 or even GBM with versus without EGFR amplifica- Vital et al95 investigated the association between the GEP 1114
1053 tion).95 Altogether, these results indicate that histopathological and both the cytogenetics and histopathology of 40 gliomas. 1115
1054 1116
1117 Table 2 Classifications of GBM Proposed on the Basis of Different Profiles of Molecular and Genetic Alterations Identified and the 1179
1118 Relationship Existing between the Different Subtypes Identified 1180
1119 1181
GBM classifications Molecular/genetic subgroups of GBM
1120 1182
1121 Philips et al99 Proneural Proliferative and mesenchymal 1183
1122 (n Z 76) Normal EGFR EGFR AMP/normal 1184
1123 Normal PTEN þ7 and 10 1185
1124 PTEN loss 1186
1125 21 1187
Verhaak et al Proneural Classic Mesenchymal Neural
1126 (n Z 170) PDGFRA AMP EGFR AMP Loss 17q11.2 (NF1) Not distinctive* 1188
1127 IDH1 MUT Loss 10 (PTEN) 1189
1128 TP53 loss/MUT Loss 9p21 (CDKN2A) 1190
1129 PI3KCA/PIK3R1 MUT 1191
1130 56 1192
Crespo et al Pattern III Pattern I Pattern II Pattern IV Pattern V
1131 1193
(n Z 35) þ7 EGFR AMP þ7 þ7 þ7
1132 1194
del(9p21)/þ9q 10/del(10p) and del(9p21) Normal 9p21 þ9
1133 1195
Normal 10 del (10q) 10/del(10p) and 10 10
1134 1196
del(9p21) del (10q)
1135 1197
1136 *Expression of neuron histological markers. 1198
1137 AMP, gene amplification; GBM, glioblastoma multiforme; MUT, gene mutation; NF1, neurofibromatosis type I. 1199
1138 1200
1139 1201
1140 Unsupervised and supervised analyses showed significantly pathway (78%); the genes involved in these pathways 1202
1141 different GEP in low- versus high-grade gliomas, the most included the NF1 tumor-suppressor gene and the PIK3R1 1203
1142 discriminating genes including genes involved in the regu- gene, which had been previously reported to be altered in 1204
1143 lation of cell proliferation, apoptosis, DNA repair, and 80
GBM. In a more recent multidimensional study including 1205
1144 1206
signal transduction. In turn, among GBM, three subgroups previous TCGA data, together with the patterns of somatic
1145 1207
1146
of tumors were identified according to their GEP, which mutations, DNA CNA, and the tumor-associated GEP, an 1208
1147 were closely associated with the cytogenetic profile of their integrated molecular classification of GBM was proposed.21 1209
1148 ancestral tumor cell clones: i) EGFR amplification, ii) iso- This classification included four subtypes of GBM defined 1210
1149 lated trisomy 7, and iii) more complex karyotypes. In this by the patterns of alteration of the EGFR, NF1, PDGFRA, 1211
1150 regard, we have also investigated the direct impact of cy- and IDH1 genes and the GEP: classic, MQ, PN, and neural 1212
1151 togenetic alterations defined by copy number changes on the (Table 2). These subtypes also differed in their response to ½T2 1213
1152 mRNA levels of the specific genes involved.62 In that study, treatment and overall survival. Classic GBM displayed a 1214
1153 we detected recurrent amplicons for chromosomes 7 (50%), characteristic profile with high proliferation, frequently 1215
1154 12 (22%), 1 (11%), 4 (9%), 11 (4%), and 17 (4%), whereas associated with high-level EGFR amplification, monosomy 1216
1155 homozygous deletions involved chromosomes 9p21 (52%) 10, and homozygous del(9p21.3) targeting the CDKN2A- 1217
1156 INK4a 1218
and 10q (22%). Most genes that displayed a high correlation p16 gene. The MQ GBMs were defined by over-
1157 1219
1158
between DNA copy number alteration (CNA) and mRNA expression of MQ (CHI3L1/YKL40 and MET) and astrocytic 1220
1159 levels were coded in the amplified chromosomes. For some (CD44 and MERTK) markers, plus NF1 deletion or mutation 1221
1160 amplicons, the impact of DNA CNA on mRNA expression (17q11.2). The PN GBMs presented with activation of both 1222
1161 was restricted to a single gene (eg, EGFR at 7p11.2), oligodendrocytic (PDGFRA, OLIG2, TCF3, and NKX2-2) 1223
1162 whereas for others, it involved multiple genes (eg, 11 and 5 and PN (SOX, DCX, DLL3, ASCL1, and TCF4) development- 1224
1163 genes at 12q14.1-q15 and 4q12, respectively). Despite the associated genes, and they were characterized by molecular 1225
1164 fact that homozygous del(9p21) and del(10q23.31) included alterations of TP53 and PDGFRA, as well as PIK3CA/ 1226
1165 multiple genes, association between these DNA CNAs and PIK3R1 and IDH1 mutations. Finally, neural GBM lacked a 1227
1166 RNA expression was restricted to a few genes [eg, the distinctive genetic profile and display gene expression sig- 1228
1167 MTAP gene in case of del(9p21)]. natures that are similar to those found in normal brain tissue, 1229
1168 1230
Until now, the most comprehensive and reliable analysis with expression of neuron markers such as NEFL, GABRA1,
1169 1231
1170
of genomic alterations in GBM has been conducted by the SYT1, and SLC12A5. Although in primary GBM, several 1232
1171 TCGA Consortium. Initially, the TCGA Consortium pub- molecular abnormalities (eg, EGFR, MTAP, and PDGFRA 1233
1172 lished mRNA expression data and DNA CNA data on alterations) were strongly associated with specific GEPs (eg, 1234
1173 approximately 206 GBMs; at the same time, >600 genes EGFR and MTAP genetic alterations were associated with 1235
1174 were also sequenced in a subgroup of 91 GBMs.24 Overall, PF, classic, and PN GBM, respectively),21,99,100 the associ- 1236
1175 three major signaling pathways were found to be consis- ation between the genes involved in secondary GBM and 1237
1176 tently affected in many GBM tumors: the TKR-signaling their potentially associated GEP still remains to be deter- 1238
1177 pathway (88%), the TP53 pathway (87%), and the RB mined and requires further assessment. 1239
1178 1240
1241 Conclusions and IDH1 mutations refine the classification of malignant gliomas. 1303
1242 Oncotarget 2012, 3:709e722 1304
1243 8. Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA, Yuan W, 1305
Overall, GBMs carry numerous recurrent genetic alterations Kos I, Batinic-Haberle I, Jones S, Riggins GJ, Friedman H,
1244 1306
1245 that translate into different cytogenetic and genomic profiles Friedman A, Reardon D, Herndon J, Kinzler KW, Velculescu VE, 1307
1246 that are difficult to distinguish on histopathological grounds. Vogelstein B, Bigner DD: IDH1 and IDH2 mutations in gliomas.
1308
Although further validation in larger series of GBM is still N Engl J Med 2009, 360:765e773
1247 1309
9. Nobusawa S, Watanabe T, Kleihues P, Ohgaki H: IDH1 mutations as
1248 needed, current knowledge about the complex genome of 1310
molecular signature and predictive factor of secondary glioblastomas.
1249 GBM points to the existence of four major profiles: classic, Clin Cancer Res 2009, 15:6002e6007 1311
1250 MQ, PN, and neural GBM. Such genomic profiles of GBM 10. Fujisawa H, Reis RM, Nakamura M, Colella S, Yonekawa Y, 1312
1251 reflect both different cytogenetic profiles, as well as unique Kleihues P, Ohgaki H: Loss of heterozygosity on chromosome 10 is 1313
1252 GEPs associated with distinct patterns of alteration of intra- more extensive in primary (de novo) than in secondary glioblastomas. 1314
1253 Lab Invest 2000, 80:65e72 1315
cellular signaling pathways, in addition to key stages in
1254 11. Nakamura M, Yang F, Fujisawa H, Yonekawa Y, Kleihues P, 1316
neurogenesis and features of distinct normal central nervous Ohgaki H: Loss of heterozygosity on chromosome 19 in secondary
1255 1317
system cell types. Additional pieces of evidence also confirm glioblastomas. J Neuropathol Exp Neurol 2000, 59:539e543
1256 1318
1257 that the relationship between DNA (eg, copy number) alter- 12. Tohma Y, Gratas C, Van Meir EG, Desbaillets I, Tenan M,
1319
ation and gene expression at the mRNA level is not always Tachibana O, Kleihues P, Ohgaki H: Necrogenesis and Fas/APO-1
1258 (CD95) expression in primary (de novo) and secondary glioblas- 1320
1259 straightforward; distant genetic interactions and epigenetic 1321
tomas. J Neuropathol Exp Neurol 1998, 57:239e245
1260 changes also have a significant impact on the expression of 13. Xie D, Zeng YX, Wang HJ, Wen JM, Tao Y, Sham JS, Guan XY: 1322
1261 specific genes. Selection of just the most commonly ampli- Expression of cytoplasmic and nuclear Survivin in primary and sec- 1323
1262 fied and mutated genes as a therapeutic target may not be ondary human glioblastoma. Br J Cancer 2006, 94:108e114 1324
1263 sufficient. In turn, identification of driver mutations using 14. Choe G, Park JK, Jouben-Steele L, Kremen TJ, Liau LM, 1325
1264 Vinters HV, Cloughesy TF, Mischel PS: Active matrix metal- 1326
large-scale, next-generation sequencing or single-tumor cell
1265 loproteinase 9 expression is associated with primary glioblastoma 1327
sequencing, as well as genomic analyses within individual subtype. Clin Cancer Res 2002, 8:2894e2901
1266 1328
tumors, in combination with new proteomic strategies to 15. Watanabe K, Tachibana O, Sata K, Yonekawa Y, Kleihues P,
1267 1329
identify aberrant cancer peptides may contribute to the Ohgaki H: Overexpression of the EGF receptor and p53 mutations are
1268 mutually exclusive in the evolution of primary and secondary glio- 1330
1269 identification of new drugs and antibody targets that may 1331
blastomas. Brain Pathol 1996, 6:217e223. discussion 223e224
1270 help in suppressing the disease and improving the manage- 1332
16. Furuta M, Weil RJ, Vortmeyer AO, Huang S, Lei J, Huang TN,
1271 ment of GBM patients. Lee YS, Bhowmick DA, Lubensky IA, Oldfield EH, Zhuang Z: 1333
1272 Protein patterns and proteins that identify subtypes of glioblastoma 1334
1273 multiforme. Oncogene 2004, 23:6806e6814 1335
1274 References 17. Biernat W, Kleihues P, Yonekawa Y, Ohgaki H: Amplification and 1336
1275 overexpression of MDM2 in primary (de novo) glioblastomas. J 1337
1. Ohgaki H, Dessen P, Jourde B, Horstmann S, Nishikawa T, Di Neuropathol Exp Neurol 1997, 56:180e185
1276 1338
Patre PL, Burkhard C, Schuler D, Probst-Hensch NM, Maiorka PC, 18. Karcher S, Steiner HH, Ahmadi R, Zoubaa S, Vasvari G, Bauer H,
1277 1339
Baeza N, Pisani P, Yonekawa Y, Yasargil MG, Lutolf UM, Unterberg A, Herold-Mende C: Different angiogenic phenotypes in pri-
1278 Kleihues P: Genetic pathways to glioblastoma: a population-based mary and secondary glioblastomas. Int J Cancer 2006, 118:2182e2189 1340
1279 study. Cancer Res 2004, 64:6892e6899 19. Godard S, Getz G, Delorenzi M, Farmer P, Kobayashi H, Desbaillets I, 1341
1280 2. Brennan CW, Verhaak RG, McKenna A, Campos B, Noushmehr H, Nozaki M, Diserens AC, Hamou MF, Dietrich PY, Regli L, Janzer RC, 1342
1281 Salama SR, et al; TCGA Research Network: The somatic genomic Bucher P, Stupp R, de Tribolet N, Domany E, Hegi ME: Classification 1343
1282 landscape of glioblastoma. Cell 2013, 155:462e477 of human astrocytic gliomas on the basis of gene expression: a 1344
1283 3. Nakamura M, Watanabe T, Yonekawa Y, Kleihues P, Ohgaki H: correlated group of genes with angiogenic activity emerges as a strong 1345
1284 Promoter methylation of the DNA repair gene MGMT in astrocy- predictor of subtypes. Cancer Res 2003, 63:6613e6625 1346
1285 tomas is frequently associated with G: C–> A: t mutations of the 20. Somasundaram K, Reddy SP, Vinnakota K, Britto R, Subbarayan M, 1347
TP53 tumor suppressor gene. Carcinogenesis 2001, 22:1715e1719 Nambiar S, Hebbar A, Samuel C, Shetty M, Sreepathi HK,
1286 1348
4. Nakamura M, Ishida E, Shimada K, Kishi M, Nakase H, Sakaki T, Santosh V, Hegde AS, Hegde S, Kondaiah P, Rao MR: Upregulation
1287 1349
Konishi N: Frequent LOH on 22q12.3 and TIMP-3 inactivation occur of ASCL1 and inhibition of Notch signaling pathway characterize
1288 in the progression to secondary glioblastomas. Lab Invest 2005, 85: progressive astrocytoma. Oncogene 2005, 24:7073e7083 1350
1289 165e175 21. Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, 1351
1290 5. Nakamura M, Yonekawa Y, Kleihues P, Ohgaki H: Promoter Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G, 1352
1291 hypermethylation of the RB1 gene in glioblastomas. Lab Invest 2001, Lawrence M, O’Kelly M, Tamayo P, Weir BA, Gabriel S, 1353
1292 81:77e82 Winckler W, Gupta S, Jakkula L, Feiler HS, Hodgson JG, James CD, 1354
1293 6. Nakamura M, Watanabe T, Klangby U, Asker C, Wiman K, Sarkaria JN, Brennan C, Kahn A, Spellman PT, Wilson RK, 1355
1294 Yonekawa Y, Kleihues P, Ohgaki H: p14ARF deletion and methyl- Speed TP, Gray JW, Meyerson M, Getz G, Perou CM, Hayes DN: 1356
1295 ation in genetic pathways to glioblastomas. Brain Pathol 2001, 11: Integrated genomic analysis identifies clinically relevant subtypes of 1357
159e168 glioblastoma characterized by abnormalities in PDGFRA, IDH1,
1296 1358
7. Jiao Y, Killela PJ, Reitman ZJ, Rasheed AB, Heaphy CM, de EGFR, and NF1. Cancer Cell 2010, 17:98e110
1297 1359
Wilde RF, Rodriguez FJ, Rosemberg S, Oba-Shinjo SM, Nagahashi 22. Ohgaki H, Kleihues P: Genetic pathways to primary and secondary
1298 Marie SK, Bettegowda C, Agrawal N, Lipp E, Pirozzi C, Lopez G, glioblastoma. Am J Pathol 2007, 170:1445e1453 1360
1299 He Y, Friedman H, Friedman AH, Riggins GJ, Holdhoff M, Burger P, 23. Mao H, Lebrun DG, Yang J, Zhu VF, Li M: Deregulated signaling 1361
1300 McLendon R, Bigner DD, Vogelstein B, Meeker AK, Kinzler KW, pathways in glioblastoma multiforme: molecular mechanisms and 1362
1301 Papadopoulos N, Diaz LA, Yan H: Frequent ATRX, CIC, FUBP1 therapeutic targets. Cancer Invest 2012, 30:48e56 1363
1302 1364
1365 24. The Cancer Genome Atlas (TCGA) Research Network: Compre- 45. Dasari VR, Kaur K, Velpula KK, Gujrati M, Fassett D, 1427
1366 hensive genomic characterization defines human glioblastoma genes Klopfenstein JD, Dinh DH, Rao JS: Upregulation of PTEN in glioma 1428
1367 and core pathways. Nature 2008, 455:1061e1068 cells by cord blood mesenchymal stem cells inhibits migration via 1429
1368 25. Dunn GP, Rinne ML, Wykosky J, Genovese G, Quayle SN, Dunn IF, downregulation of the PI3K/Akt pathway. PLoS One 2010, 5:e10350 1430
1369 Agarwalla PK, Chheda MG, Campos B, Wang A, Brennan C, 46. Van Meir EG, Hadjipanayis CG, Norden AD, Shu HK, Wen PY, 1431
Ligon KL, Furnari F, Cavenee WK, Depinho RA, Chin L, Hahn WC: Olson JJ: Exciting new advances in neuro-oncology: the avenue to a
1370 1432
Emerging insights into the molecular and cellular basis of glioblas- cure for malignant glioma. CA Cancer J Clin 2010, 60:166e193
1371 1433
toma. Genes Dev 2012, 26:756e784 47. Meletis K, Wirta V, Hede SM, Nister M, Lundeberg J, Frisen J: p53
1372 26. Nakada M, Kita D, Watanabe T, Hayashi Y, Teng L, Pyko IV, Suppresses the self-renewal of adult neural stem cells. Development 1434
1373 Hamada J: Aberrant signaling pathways in glioma. Cancers (Basel) 2006, 133:363e369 1435
1374 2011, 3:3242e3278 48. Zawlik I, Kita D, Vaccarella S, Mittelbronn M, Franceschi S, 1436
1375 27. Zheng Y, McFarland BC, Drygin D, Yu H, Bellis SL, Kim H, Ohgaki H: Common polymorphisms in the MDM2 and TP53 genes 1437
1376 Bredel M, Benveniste EN: Targeting protein kinase CK2 suppresses and the relationship between TP53 mutations and patient outcomes in 1438
1377 prosurvival signaling pathways and growth of glioblastoma. Clin glioblastomas. Brain Pathol 2009, 19:188e194 1439
1378 Cancer Res 2013, 19:6484e6494 49. Linares LK, Hengstermann A, Ciechanover A, Muller S, Scheffner M: 1440
1379 28. Fredriksson L, Li H, Eriksson U: The PDGF family: four gene HdmX stimulates Hdm2-mediated ubiquitination and degradation of 1441
products form five dimeric isoforms. Cytokine Growth Factor Rev p53. Proc Natl Acad Sci U S A 2003, 100:12009e12014
1380 1442
2004, 15:197e204 50. Jones PA, Baylin SB: The epigenomics of cancer. Cell 2007, 128:
1381 1443
29. Alentorn A, Marie Y, Carpentier C, Boisselier B, Giry M, Labussiére M, 683e692
1382 Mokhtari K, Hoang-Xuan K, Sanson M, Delattre JY, Idbaih A: Preva- 51. Kloosterhof NK, de Rooi JJ, Kros M, Eilers PH, Sillevis Smitt PA, 1444
1383 lence clinico-pathological value, and co-occurrence of PDGFRA ab- van den Bent MJ, French PJ: Molecular subtypes of glioma identified 1445
1384 normalities in diffuse gliomas. Neuro Oncol 2012, 14:1393e1403 by genome-wide methylation profiling. Genes Chromosomes Cancer 1446
1385 30. Rheinbay E, Suva ML, Gillespie SM, Wakimoto H, Patel AP, 2013, 52:665e674 1447
1386 Shahid M, Oksuz O, Rabkin SD, Martuza RL, Rivera MN, Louis DN, 52. Horiguchi K, Tomizawa Y, Tosaka M, Ishiuchi S, Kurihara H, 1448
1387 Kasif S, Chi AS, Bernstein BE: An aberrant transcription factor Mori M, Saito N: Epigenetic inactivation of RASSF1A candidate 1449
1388 network essential for Wnt signaling and stem cell maintenance in tumor suppressor gene at 3p21.3 in brain tumors. Oncogene 2003, 22: 1450
1389 glioblastoma. Cell Rep 2013, 3:1567e1579 7862e7865 1451
31. Nicholas MK, Lukas RV, Jafri NF, Faoro L, Salgia R: Epidermal 53. Stone AR, Bobo W, Brat DJ, Devi NS, Van Meir EG, Vertino PM:
1390 1452
growth factor receptor-mediated signal transduction in the develop- Aberrant methylation and down-regulation of TMS1/ASC in human
1391 1453
ment and therapy of gliomas. Clin Cancer Res 2006, 12:7261e7270 glioblastoma. Am J Pathol 2004, 165:1151e1161
1392 32. Huang PH, Xu AM, White FM: Oncogenic EGFR signaling networks 54. Martinez R, Schackert G, Esteller M: Hypermethylation of the pro- 1454
1393 in glioma. Sci Signal 2009, 2:re6 apoptotic gene TMS1/ASC: prognostic importance in glioblastoma 1455
1394 33. Hatanpaa KJ, Burma S, Zhao D, Habib AA: Epidermal growth factor multiforme. J Neurooncol 2007, 82:133e139 1456
1395 receptor in glioma: signal transduction, neuropathology, imaging, and 55. Mellai M, Piazzi A, Caldera V, Annovazzi L, Monzeglio O, 1457
1396 radioresistance. Neoplasia 2010, 12:675e684 Senetta R, Cassoni P, Schiffer D: Promoter hypermethylation of the 1458
1397 34. Patil CG, Nuno M, Elramsisy A, Mukherjee D, Carico C, Dantis J, EMP3 gene in a series of 229 human gliomas. Biomed Res Int 2013, 1459
1398 Hu J, Yu JS, Fan X, Black KL, Bannykh SI: High levels of phos- 2013:756302 1460
1399 phorylated MAP kinase are associated with poor survival among 56. Crespo I, Vital AL, Nieto AB, Rebelo O, Tao H, Lopes MC, 1461
patients with glioblastoma during the temozolomide era. Neuro Oncol Oliveira CR, French PJ, Orfao A, Tabernero MD: Detailed charac-
1400 1462
2013, 15:104e111 terization of alterations of chromosomes 7, 9, and 10 in glioblastomas
1401 1463
35. Downward J: Targeting RAS signalling pathways in cancer therapy. as assessed by single-nucleotide polymorphism arrays. J Mol Diagn
1402 Nat Rev Cancer 2003, 3:11e22 2011, 13:634e647 1464
1403 36. Rajasekhar VK, Viale A, Socci ND, Wiedmann M, Hu X, 57. Vital AL, Tabernero MD, Crespo I, Rebelo O, Tao H, Gomes F, 1465
1404 Holland EC: Oncogenic Ras and Akt signaling contribute to glio- Lopes MC, Orfao A: Intratumoral patterns of clonal evolution in 1466
1405 blastoma formation by differential recruitment of existing mRNAs to gliomas. Neurogenetics 2010, 11:227e239 1467
1406 polysomes. Mol Cell 2003, 12:889e901 58. Appin CL, Brat DJ: Molecular genetics of gliomas. Cancer J 2014, 1468
1407 37. Furnari FB, Fenton T, Bachoo RM, Mukasa A, Stommel JM, 20:66e72 1469
1408 Stegh A, Hahn WC, Ligon KL, Louis DN, Brennan C, Chin L, 59. Nonoguchi N, Ohta T, Oh JE, Kim YH, Kleihues P, Ohgaki H: TERT 1470
1409 DePinho RA, Cavenee WK: Malignant astrocytic glioma: genetics, promoter mutations in primary and secondary glioblastomas. Acta 1471
biology, and paths to treatment. Genes Dev 2007, 21:2683e2710 Neuropathol 2013, 126:931e937
1410 1472
38. Koul D: PTEN signaling pathways in glioblastoma. Cancer Biol Ther 60. Arita H, Narita Y, Fukushima S, Tateishi K, Matsushita Y,
1411 1473
2008, 7:1321e1325 Yoshida A, Miyakita Y, Ohno M, Collins VP, Kawahara N, Shibui S,
1412 39. Vogt PK: PI 3-kinase, mTOR, protein synthesis and cancer. Trends Ichimura K: Upregulating mutations in the TERT promoter 1474
1413 Mol Med 2001, 7:482e484 commonly occur in adult malignant gliomas and are strongly asso- 1475
1414 40. Fan X, Aalto Y, Sanko SG, Knuutila S, Klatzmann D, Castresana JS: ciated with total 1p19q loss. Acta Neuropathol 2013, 126:267e276 1476
1415 Genetic profile, PTEN mutation and therapeutic role of PTEN in 61. Gan HK, Cvrljevic AN, Johns TG: The epidermal growth factor re- 1477
1416 glioblastomas. Int J Oncol 2002, 21:1141e1150 ceptor variant III (EGFRvIII): where wild things are altered. FEBS J 1478
1417 41. Knudsen ES, Wang JY: Targeting the RB-pathway in cancer therapy. 2013, 280:5350e5370 1479
1418 Clin Cancer Res 2010, 16:1094e1099 62. Crespo I, Tao H, Nieto AB, Rebelo O, Domingues P, Vital AL, 1480
1419 42. Matheu A, Maraver A, Serrano M: The Arf/p53 pathway in cancer Patino Mdel C, Barbosa M, Lopes MC, Oliveira CR, Orfao A, 1481
and aging. Cancer Res 2008, 68:6031e6034 Tabernero MD: Amplified and homozygously deleted genes in glio-
1420 1482
43. Shangary S, Wang S: Small-molecule inhibitors of the MDM2-p53 blastoma: impact on gene expression levels. PLoS One 2012, 7:
1421 1483
protein-protein interaction to reactivate p53 function: a novel e46088
1422 approach for cancer therapy. Annu Rev Pharmacol Toxicol 2009, 49: 63. Carico C, Nuno M, Mukherjee D, Elramsisy A, Dantis J, Hu J, 1484
1423 223e241 Rudnick J, Yu JS, Black KL, Bannykh SI, Patil CG: Loss of PTEN is 1485
1424 44. Grzmil M, Hemmings BA: Deregulated signalling networks in human not associated with poor survival in newly diagnosed glioblastoma 1486
1425 brain tumours. Biochim Biophys Acta 2010, 1804:476e483 patients of the temozolomide era. PLoS One 2012, 7:e33684 1487
1426 1488
1489 64. Kita D, Yonekawa Y, Weller M, Ohgaki H: PIK3CA alterations in 78. Reitman ZJ, Yan H: Isocitrate dehydrogenase 1 and 2 mutations in 1551
1490 primary (de novo) and secondary glioblastomas. Acta Neuropathol cancer: alterations at a crossroads of cellular metabolism. J Natl 1552
1491 2007, 113:295e302 Cancer Inst 2010, 102:932e941 1553
1492 65. Kitange GJ, Carlson BL, Schroeder MA, Grogan PT, Lamont JD, 79. Joseph JW, Jensen MV, Ilkayeva O, Palmieri F, Alarcon C, 1554
1493 Decker PA, Wu W, James CD, Sarkaria JN: Induction of MGMT Rhodes CJ, Newgard CB: The mitochondrial citrate/isocitrate carrier 1555
expression is associated with temozolomide resistance in glioblas- plays a regulatory role in glucose-stimulated insulin secretion. J Biol
1494 1556
toma xenografts. Neuro Oncol 2009, 11:281e291 Chem 2006, 281:35624e35632
1495 1557
66. Christensen BC, Smith AA, Zheng S, Koestler DC, Houseman EA, 80. Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P,
1496 Marsit CJ, Wiemels JL, Nelson HH, Karagas MR, Wrensch MR, Mankoo P, Carter H, Siu IM, Gallia GL, Olivi A, McLendon R, 1558
1497 Kelsey KT, Wiencke JK: DNA methylation, isocitrate dehydrogenase Rasheed BA, Keir S, Nikolskaya T, Nikolsky Y, Busam DA, 1559
1498 mutation, and survival in glioma. J Natl Cancer Inst 2011, 103: Tekleab H, Diaz LA Jr, Hartigan J, Smith DR, Strausberg RL, 1560
1499 143e153 Marie SK, Shinjo SM, Yan H, Riggins GJ, Bigner DD, Karchin R, 1561
1500 67. Inohara N, Ding L, Chen S, Nunez G: Harakiri, a novel regulator of Papadopoulos N, Parmigiani G, Vogelstein B, Velculescu VE, 1562
1501 cell death, encodes a protein that activates apoptosis and interacts Kinzler KW: An integrated genomic analysis of human glioblastoma 1563
1502 selectively with survival-promoting proteins Bcl-2 and Bcl-X(L). multiforme. Science 2008, 321:1807e1812 1564
1503 EMBO J 1997, 16:1686e1694 81. Kloosterhof NK, Bralten LB, Dubbink HJ, French PJ, van den 1565
68. Jeuken J, Sijben A, Alenda C, Rijntjes J, Dekkers M, Boots- Bent MJ: Isocitrate dehydrogenase-1 mutations: a fundamentally new
1504 1566
Sprenger S, McLendon R, Wesseling P: Robust detection of EGFR understanding of diffuse glioma? Lancet Oncol 2011, 12:83e91
1505 1567
copy number changes and EGFR variant III: technical aspects and 82. Watanabe T, Nobusawa S, Kleihues P, Ohgaki H: IDH1 mutations
1506 relevance for glioma diagnostics. Brain Pathol 2009, 19:661e671 are early events in the development of astrocytomas and oligoden- 1568
1507 69. Murat A, Migliavacca E, Gorlia T, Lambiv WL, Shay T, drogliomas. Am J Pathol 2009, 174:1149e1153 1569
1508 Hamou MF, de Tribolet N, Regli L, Wick W, Kouwenhoven MC, 83. Weller M, Felsberg J, Hartmann C, Berger H, Steinbach JP, 1570
1509 Hainfellner JA, Heppner FL, Dietrich PY, Zimmer Y, Schramm J, Westphal M, Schackert G, Simon M, Tonn JC, Heese O, 1571
1510 Cairncross JG, Janzer RC, Domany E, Delorenzi M, Stupp R, Krex D, Nikkhah G, Pietsch T, Wiestler O, Reifenberger G, von 1572
1511 Hegi ME: Stem cell-related “self-renewal” signature and high Deimling A, Loeffler M: Molecular predictors of progression-free and 1573
1512 epidermal growth factor receptor expression associated with resis- overall survival in patients with newly diagnosed glioblastoma: a 1574
1513 tance to concomitant chemoradiotherapy in glioblastoma. J Clin prospective translational study of the German Glioma Network. J Clin 1575
Oncol 2008, 26:3015e3024 Oncol 2009, 27:5743e5750
1514 1576
70. Benito R, Gil-Benso R, Quilis V, Perez M, Gregori-Romero M, 84. Sanson M, Marie Y, Paris S, Idbaih A, Laffaire J, Ducray F, El
1515 1577
Roldan P, Gonzalez-Darder J, Cerda-Nicolas M, Lopez-Gines C: Hallani S, Boisselier B, Mokhtari K, Hoang-Xuan K, Delattre JY:
1516 Primary glioblastomas with and without EGFR amplification: rela- Isocitrate dehydrogenase 1 codon 132 mutation is an important 1578
1517 tionship to genetic alterations and clinicopathological features. prognostic biomarker in gliomas. J Clin Oncol 2009, 27:4150e4154 1579
1518 Neuropathology 2010, 30:392e400 85. Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, Yu W, Li Z, Gong L, 1580
1519 71. Heimberger AB, Hlatky R, Suki D, Yang D, Weinberg J, Gilbert M, Peng Y, Ding J, Lei Q, Guan KL, Xiong Y: Glioma-derived muta- 1581
1520 Sawaya R, Aldape K: Prognostic effect of epidermal growth factor tions in IDH1 dominantly inhibit IDH1 catalytic activity and induce 1582
1521 receptor and EGFRvIII in glioblastoma multiforme patients. Clin HIF-1alpha. Science 2009, 324:261e265 1583
1522 Cancer Res 2005, 11:1462e1466 86. Dang L, White DW, Gross S, Bennett BD, Bittinger MA, 1584
1523 72. Pelloski CE, Ballman KV, Furth AF, Zhang L, Lin E, Sulman EP, Driggers EM, Fantin VR, Jang HG, Jin S, Keenan MC, Marks KM, 1585
Bhat K, McDonald JM, Yung WK, Colman H, Woo SY, Prins RM, Ward PS, Yen KE, Liau LM, Rabinowitz JD, Cantley LC,
1524 1586
Heimberger AB, Suki D, Prados MD, Chang SM, Barker FG 2nd, Thompson CB, Vander Heiden MG, Su SM: Cancer-associated IDH1
1525 1587
Buckner JC, James CD, Aldape K: Epidermal growth factor receptor mutations produce 2-hydroxyglutarate. Nature 2009, 462:739e744
1526 variant III status defines clinically distinct subtypes of glioblastoma. J 87. Zawlik I, Vaccarella S, Kita D, Mittelbronn M, Franceschi S, 1588
1527 Clin Oncol 2007, 25:2288e2294 Ohgaki H: Promoter methylation and polymorphisms of the MGMT 1589
1528 73. Montano N, Cenci T, Martini M, D’Alessandris QG, Pelacchi F, gene in glioblastomas: a population-based study. Neuroepidemiology 1590
1529 Ricci-Vitiani L, Maira G, De Maria R, Larocca LM, Pallini R: 2009, 32:21e29 1591
1530 Expression of EGFRvIII in glioblastoma: prognostic significance 88. Herrlinger U, Rieger J, Koch D, Loeser S, Blaschke B, 1592
1531 revisited. Neoplasia 2011, 13:1113e1121 Kortmann RD, Steinbach JP, Hundsberger T, Wick W, 1593
1532 74. Costa BM, Viana-Pereira M, Fernandes R, Costa S, Linhares P, Meyermann R, Tan TC, Sommer C, Bamberg M, Reifenberger G, 1594
1533 Vaz R, Pinheiro C, Lima J, Soares P, Silva A, Pardal F, Amorim J, Weller M: Phase II trial of lomustine plus temozolomide chemo- 1595
Nabico R, Almeida R, Alegria C, Pires MM, Carvalho E, Oliveira P, therapy in addition to radiotherapy in newly diagnosed glioblastoma:
1534 1596
Lopes JM, Reis RM: Impact of EGFR genetic variants on glioma risk UKT-03. J Clin Oncol 2006, 24:4412e4417
1535 1597
and patient outcome. Cancer Epidemiol Biomarkers Prev 2011, 20: 89. Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ,
1536 2610e2617 Janzer RC, Ludwin SK, Allgeier A, Fisher B, Belanger K, Hau P, 1598
1537 75. van den Bent MJ, Brandes AA, Rampling R, Kouwenhoven MC, Brandes AA, Gijtenbeek J, Marosi C, Vecht CJ, Mokhtari K, 1599
1538 Kros JM, Carpentier AF, Clement PM, Frenay M, Campone M, Wesseling P, Villa S, Eisenhauer E, Gorlia T, Weller M, Lacombe D, 1600
1539 Baurain JF, Armand JP, Taphoorn MJ, Tosoni A, Kletzl H, Cairncross JG, Mirimanoff RO: Effects of radiotherapy with 1601
1540 Klughammer B, Lacombe D, Gorlia T: Randomized phase II trial of concomitant and adjuvant temozolomide versus radiotherapy alone on 1602
1541 erlotinib versus temozolomide or carmustine in recurrent glioblas- survival in glioblastoma in a randomised phase III study: 5-year 1603
1542 toma: EORTC brain tumor group study 26034. J Clin Oncol 2009, analysis of the EORTC-NCIC trial. Lancet Oncol 2009, 10:459e466 1604
1543 27:1268e1274 90. Linz U: Commentary on Effects of radiotherapy with concomitant 1605
76. Paul I, Bhattacharya S, Chatterjee A, Ghosh MK: Current under- and adjuvant temozolomide versus radiotherapy alone on survival in
1544 1606
standing on EGFR and Wnt/beta-catenin signaling in glioma and their glioblastoma in a randomised phase III study: 5-year analysis of the
1545 1607
possible crosstalk. Genes Cancer 2013, 4:427e446 EORTC-NCIC trial (Lancet Oncol. 2009;10:459-466). Cancer 2010,
1546 77. Xu X, Zhao J, Xu Z, Peng B, Huang Q, Arnold E, Ding J: Structures 116:1844e1846 1608
1547 of human cytosolic NADP-dependent isocitrate dehydrogenase reveal 91. Rivera AL, Pelloski CE, Gilbert MR, Colman H, De La Cruz C, 1609
1548 a novel self-regulatory mechanism of activity. J Biol Chem 2004, Sulman EP, Bekele BN, Aldape KD: MGMT promoter methylation is 1610
1549 279:33946e33957 predictive of response to radiotherapy and prognostic in the absence 1611
1550 1612
1613 of adjuvant alkylating chemotherapy for glioblastoma. Neuro Oncol Nelson SF: Distinct transcription profiles of primary and secondary 1633
1614 2010, 12:116e121 glioblastoma subgroups. Cancer Res 2006, 66:159e167 1634
1615 92. Weller M, Berger H, Hartmann C, Schramm J, Westphal M, 97. Nutt CL, Mani DR, Betensky RA, Tamayo P, Cairncross JG, Ladd C, 1635
1616 Simon M, Goldbrunner R, Krex D, Steinbach JP, Ostertag CB, Pohl U, Hartmann C, McLaughlin ME, Batchelor TT, Black PM, von 1636
1617 Loeffler M, Pietsch T, von Deimling A: Combined 1p/19q loss in Deimling A, Pomeroy SL, Golub TR, Louis DN: Gene expression- 1637
1618 oligodendroglial tumors: predictive or prognostic biomarker? Clin based classification of malignant gliomas correlates better with sur- 1638
1619 Cancer Res 2007, 13:6933e6937 vival than histological classification. Cancer Res 2003, 63:1602e1607 1639
1620 93. Weller M, Stupp R, Reifenberger G, Brandes AA, van den Bent MJ, 98. Freije WA, Castro-Vargas FE, Fang Z, Horvath S, Cloughesy T, 1640
1621 Wick W, Hegi ME: MGMT promoter methylation in malignant gli- Liau LM, Mischel PS, Nelson SF: Gene expression profiling of gli- 1641
1622 omas: ready for personalized medicine? Nat Rev Neurol 2010, 6: omas strongly predicts survival. Cancer Res 2004, 64:6503e6510 1642
1623 39e51 99. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, 1643
1624 94. Fan CH, Liu WL, Cao H, Wen C, Chen L, Jiang G: O6-methyl- Misra A, Nigro JM, Colman H, Soroceanu L, Williams PM, 1644
1625 guanine DNA methyltransferase as a promising target for the treat- Modrusan Z, Feuerstein BG, Aldape K: Molecular subclasses of high- 1645
1626 ment of temozolomide-resistant gliomas. Cell Death Dis 2013, 4:e876 grade glioma predict prognosis, delineate a pattern of disease pro- 1646
1627 95. Vital AL, Tabernero MD, Castrillo A, Rebelo O, Tao H, Gomes F, gression, and resemble stages in neurogenesis. Cancer Cell 2006, 9: 1647
1628 Nieto AB, Resende Oliveira C, Lopes MC, Orfao A: Gene expression 157e173 1648
1629 profiles of human glioblastomas are associated with both tumor cy- 100. Lee Y, Scheck AC, Cloughesy TF, Lai A, Dong J, Farooqi HK, 1649
1630 togenetics and histopathology. Neuro Oncol 2010, 12:991e1003 Liau LM, Horvath S, Mischel PS, Nelson SF: Gene expression 1650
1631 96. Tso CL, Freije WA, Day A, Chen Z, Merriman B, Perlina A, Lee Y, analysis of glioblastomas identifies the major molecular basis for the 1651
1632 Dia EQ, Yoshimoto K, Mischel PS, Liau LM, Cloughesy TF, prognostic benefit of younger age. BMC Med Genomics 2008, 1:52 1652