Original Article Over-expression of TOP2A as a prognostic biomarker in patients with glioma

Original Article

Over-expression of TOP2A as a prognostic biomarker

in patients with glioma

Tianmin Zhou

, Yan Wang

, Dongmeng Qian

, Qing Liang

, Bin Wang

1_{Key Laboratory of Medicine and Biotechnology of Qingdao, Department of Microbiology, Medical College of} Qingdao University, Qingdao, Shandong, P. R. China; 2_{Department of Pathology, The Affiliated Hospital of Qingdao} University, Qingdao, Shandong, P. R. China

Received December 7, 2017; Accepted January 19, 2018; Epub March 1, 2018; Published March 15, 2018

Abstract: Topoisomerase (DNA) II alpha (TOP2A), an enzyme that controls and alters the topologic states of DNA during transcription, is aberrantly expressed in many cancers. However, few studies have investigated expression of TOP2A and its clinical significance in glioma. We retrieved six independent investigations from the Oncomine da-tabase and found that TOP2A is highly expressed in glioma tissues compared with corresponding normal controls. Similar results were also found in clinical specimens at the protein level. Immunohistochemical analysis indicated that TOP2A over expression was highly correlated with grade stage, KI67 positive percentage, IDH1 mutation, and age, but other clinical parameters such as sex distribution and tumor size were barely associated with high TOP2A gene expression. Meanwhile we used Prognos can to assess the prognostic value of TOP2A expression in glioma patients, and found that high expression was associated with poor prognosis of patients with glioma. Furthermore, we used the Gene-Cloud of Biotechnology Information (GCBI) bioinformatics platform predict the role of TOP2A in glioma. It was not only involved in DNA replication, chromosome condensation, and responses to DNA damage stimuli, but also promoted cancer cell mitotic cell cycle and apoptosis, and phosphatidylinositol-mediated signaling by regulating gene expression. By these approaches we demonstrate that TOP2A may be a reliable prognostic factor or therapeutic target in glioma.

Keywords: TOP2A, glioma, marker, prognosis

Introduction

Glioma remains the most common adult brain

tumor which is slightly more common in men

than in women [1], with 7.3 cases pr. 100,000

person-years. High-grade gliomas are present

in 85% and low-grade glioma in 15% with 5-

year overall survival of 82, 54, 22 and 3% for

grade I, II, III and IV, respectively [2]. Within a

five year follow-up time, most patients had a

recurrent tumor and died of the disease [3].

Despite much effort focused on improving the

diagnosis and therapy of glioma, the clinical

outcome of glioma is still unsatisfactory [4,

5]. Consequently, more studies are needed to

explore the potential mechanism and detect

effective diagnostic and prognostic biomark-

ers to prolong the lives of glioma patients.

TOP2A gene encodes a DNA topoisomerase

that controls and alters the topologic state of

DNA during transcription. This nuclear enzyme

is involved in processes such as chromosome

condensation, chromatid separation, and the

relief of torsional stress that occurs during

DNA transcription and replication [6]. Its

aber-rant expression has been associated with mu-

ltiple tumor prognostics [7], including lung

cancer, breast cancer, pancreatic

adenocarci-noma, and colorectal cancer [8-11]. There

are few investigations analyzing expression of

TOP2A in glioma [12], and due to the limited

number of studies and few clinical samples, no

role has been detected.

Materials and methods

Oncomine analysis

The Oncomine database (http://www.oncomi-

ne.org) incorporates 264 independent

datas-ets including 35 cancer types and supports

various methods of analysis, including

molecu-lar concepts, interactome, and meta-analysis

[13]. Thus, we detected the level of TOP2A

in different cancers including hematological

malignancies and solid tumors and adopted

the Oncomine database online tools meta-an-

alysis to estimate the expression of TOP2A in

glioma and normal brain control that come

from 6 different datasets.

PrognoScan

PrognoScan (http://www.prognoscan.org) is a

comprehensive platform for evaluating pot-

ential tumor biomarkers and therapeutic

tar-gets. Based on a large collection of cancer

microarray datasets with clinical annotation,

PrognoScan is a useful online tool for

assess-ing the association between specific gene ex-

pression and prognosis in patients with can-

cer [14]. We used the PrognoScan platform to

validate the prognostic value of TOP2A expr-

ession in patients with glioma.

Patients and specimens

Forty-six glioma tumor tissue samples were

collected from the Affiliated Hospital of

Qing-dao University at QingQing-dao in China from July

2014 to December 2016. All the patients we-

re with tumors suitable for resection accord-

ing to the Chinese guideline of surgical

treat-ment strategy for glioma in 2012, and among

them, 16 cases were in gradeII, 9 cases were

grade III, and 21 cases were glioma. Altoge-

ther, 46 cases of formalin-fixed paraffin-em-

bedded (FFPE) specimens were used for Im-

munohistochemistry.

In this study, the inclusion criteria were as

the following: (1) Pathological diagnosis of gli-

oma tumors; (2) A diagnosis that was cons-

istent with the histological diagnostic criteria

of the World Health Organization [15]; (3) Tu-

mor is the primary tumor in the brain, and pa-

tients did not receive pre-operative anti-can-

cer treatment, and there were no extrahepa-

tic metastases; (4) For the use of archived

tis-sue specimens from these patients,

permis-sion from the Affiliated Hospital of Qingdao

University was granted. (5) A representative

sample was taken from each FFPE tissue and

hematoxylin and eosin (HE) staining was

per-formed. Results were reviewed by two

experi-enced pathologists.

Immunohistochemical staining and evaluation

A traditional immunohistochemical (IHC)

stain-ing protocol was used in this study. Briefly,

tis-sue slices was deparaffinized in xylene and

rehydrated with different concentrations of et-

hanol alcohol. Then the section was treated

with 3% hydrogen peroxide, followed by anti-

gen retrieval with 10 mM citrate buffer (pH

6.0) with microwave. After being blocked with

10% goat serum for 30 min, incubation with

the primary antibody was done overnight at

4°C with anti-TOP2A antibody (mouse

onoclo-nal antibody, 1:100 dilution, ZM-0245, Zhong-

shanjinqiao, China) overnight at 4°C, followed

by a peroxidase-labeled secondary antibody.

Immunoreaction score distribution (IOD)

mea-surement, as reported previously, was used

to determine the positive-staining density [16].

A Leica CCD camera DFC420 connected to a

Leica DM IRE2 microscope (Leica Microsyst-

ems Imaging Solutions, Ltd., Cambridge, UK)

was used as an imaging system. High-power

magnification (×40) was used to photograph

representative fields. The IODs in each image

were measured and counted using Image-Pro

Plus v6.0 software (Media Cybernetics, Inc.,

Bethesda, MD, USA).

X-tile analysis

TOP2A expression was assessed with X-tile

plots. TOP2A expression was expressed as

IOD and the optimization of cutoff points was

based on outcomes. The cutoff score derived

from 46 cases of training set by a standard

log-rank method was used to assess statisti-

cal significance.

P

values were obtained from

a lookup table.

Gene-cloud of biotechnology information

(GCBI)

bioinformat-ics. It creates a “gene knowledge

base” which encompasses biolo-

gy, medicine, informatics, comput-

er science, mathematics, graphics,

and other disciplines. GCBI includ-

es more than 120 million copies

of genomic samples, approximately

90,000 copies of tumor samples

and more than 17 million copies of

genetic information. Therefore, GC-

BI is a good bioinformatics analys-

is platform and has provided data

analysis support for many studies

on cancer research [17-20]. In this

study, we used GCBI to identify

dif-f-ferentially expressed genes (DEGs)

between glioma tissue and normal

brain control and finally predict the

role of TOP2A in gliomas.

Statistical analysis

Figure 2. TOP2A expression levels in peri-tumoral tis-sues (A) and glioma (B) (×40). (C) Box plot indicating the mean staining intensity of paired glioma and peri-tumoral tissues. (D) Paired peri-peri-tumoral tissue and glioma tissue in the same section (D) (×10).

Table 1.

Meta-analysis of TO2PA gene expression from six Oncomine databases

Datasets (sample size) Comparison groups _ChangeFold P-value Overexpression _{Gene Rank} Sun Brain (180) Glioblastoma vs. Normal 102.544 2.37E-27 22 (in top 1%)

Oligodendroglioma vs. Normal 28.449 8.4E-18 27 (in top 1%) Anaplastic Astocytoma 34.640 4.10E-12 27 (in top 1%) Liang Brain (33) Glioblastoma vs. Normal 2.202 6.27E-6 115 (in top 2%) TCGA Brain (552) Glioblastoma vs. Normal 28.903 2.59E-13 115 (in top 2%) French Brain Statistics (29) Anaplastic Oligodendroglioma VS. Brain 9.685 1.15E-8 115 (in top 2%) Shai Brain Statistics (34) Glioblastoma vs. White Matter 2.291 1.17E-6 227 (in top 3%) Murat Brain Statistics (84) Glioblastoma vs. Normal 13.755 2.17E-11 262 (in top 2%) Note: A meta-analysis of TO2PA gene expression from six Oncomine databases where colored squares indicate the median rank for TOP2A (vs. Normal tissue) across 8 analyses. French Brain (1), Liang Brain (2), Murat Brain (3), Shai Brain (4), Sun Brain (5-7) and TCGA (8). The P value is given for the medianrank analysis.

Variance. Survival curves for TOP2A expres-

sion were generated using the Kaplan-Meier

method and compared using the log-rank test.

P

values < 0.05 were considered statistically

significant. Each statistical analysis of online

database was completed by usedits online tool.

Results

TOP2A is distinctively overexpressed in glioma

than normal brain

Table 2.

Association of TOP2A expression with

pathological characteristics of glioma (46

cases)

Pathological

category numberCase

TOP2A

P-value Low High

Sex Male 29 11 18 0.1264 Female 17 11 6

Age ≥ 50 23 5 18 0.0009 < 50 23 17 16

Tumor size ≥ 3 cm 21 10 11 1.0000 < 3 cm 25 12 13

IDH1 IDH1 (+) 15 12 3 0.0113 IDH1 (-) 16 5 11

Figure 3. Correlation analysis of TOP2A with clinicopathological features of glioma. A, B: Expression of TOP2A was significantly associated with grade stage. C: KI67 positive percentage.

normal tissues in a dataset with 552 samples

that derived from TCGA database [21] (

Figure

1C

). In another dataset from Sun L study [22],

TOP2A was 102.544 fold elevated in glioma

samples as compared with

normal tissues (P =

2.37*10-27) (

Figure 1D

).

Elevation of TOP2A in glioma

compared with peri-tumoral

tissues by immunochemistry

We also analyzed the

expres-sion levels of TOP2A protein in

46 glioma tumors and paired

peri-tumoral tissues by

immu-nochemistry (IHC). Expression

of TOP2A protein levels was

frequently higher in the glioma

tumor tissues (

Figure 2A

) than

in peri-tumoral tissues (

Figu-

re 2B

). These differences we-

re statistically significant (

Fi-

gure 2C

; P < 0.0001, paired t

test). Meanwhile we found th-

at expression of TOP2A is po-

sitive in all samples of glioma.

In 43 cases, glioma tissues

and peri-tumoral tissue were

located in one area, and the

picture clearly revealed that

TOP2A expression was higher

in peri-tumoral tissue than gli-

oma tissue (

Figure 2D

).

Association of TOP2A

expres-sion with clinicopathological

features of glioma patients

As show in

Figure 3A

and

3B

, Immunoreaction

score distribution (IOD) of TOP2A significant-

ly increased along with grad stage of glioma (F

= 11.91, P < 0.0001) and KI67 positive

per-centage (Spearman’s r = 0.7788, P < 0.0001).

We also learned the association of the expr-

ession of TOP2A with other clinicopathological

features in glioma, as shown in

Table 2

. Ex-

pression of TOP2A association correlated wi-

th IDH1 mutation (P = 0.0113) and age (P =

0.0009), but other clinical parameters, such as

sex distribution (P = 0.1264) and tumor size (P

= 1.0000) were barely associated with high

TOP2A gene expression.

High mRNA expression of TOP2A is an

unfavor-able prognostic factor for glioma

2925 (P < 0.01, Flod change > 2) significant

DEGs were identified, of which 1041

over-pre-sented and 1884 showed an attenuated be-

havior (

Figure 5B

and

5C

). We found that TOP-

2A was significantly overexpressed in glioma

samples (Fold Change = 6.447, t = 17.9485, P

< 0.01). We then performed a Gene Ontology

analysis of 2925 differentially expressed ge-

nes (

Figure

5D

). Fortunately, we found that

TOP2A in the enriched set was associated wi-

th mitotic cell cycle, positive regulation of tr-

anscription from RNA polymerase II promoter,

positive regulation of apoptotic process, DNA

replication, phosphatidylinositol-mediated

sig-naling, response to DNA damage stimulus,

chromosome condensation. So we speculate

that the elevation of TOP2A not only takes

part in DNA replication, chromosome

conden-sation, and response to DNA damage stimu-

lus, but also associates with mitotic cell cycle,

apoptotic processes, and

phosphatidylinositol-mediated signaling.

Discussion

Glioma, a tumor of central nervous system,

which mainly develops from the macroglial

cells, presents the highest prevalence and mo-

rtality risk. The difficult treatment of patients

with glioma is primarily attributed to recurren-

ce and resistance to chemoradiotherapy

[29-31]. Despite the challenges of treatment, it is

rewarding to illustrate the pathogenesis of gli-

oma, as well as to develop novel prognostic

strategies and discover effective therapeutic

approaches. TOP2A is a nuclear enzyme that

mainly involved in processes such as chro-

mosome condensation, chromatid separation,

and the relief of torsional stress that occurs

during DNA transcription and replication [6]. It

has been reported that TOP2A is a sensitive

and specific marker in actively proliferating

cells (in the late S, G2, and M phases of the cell

cycle) which suggests a role in a wide range

of human cancers [32, 33].

In the present study, through the Oncomine

databases, we found that TOP2A was

signifi-cantly over expressed in glioma compared with

normal brain samples. Similar results were

found in 46 gliomas and paired peri-tumoral

tissues by immunochemistry byanalysis of the

association of TOP2A expression with

clinico-pathological features of glioma patients. TOP-

2A significantly increased along with grade

Figure 4. Kaplan-Meier survival curves generated from PrognoScan for TOP2A mRNA expression in tients with glioma. A: Overall survival curves for pa-tients with glioma in the GEO dataset GSE4271. B: Patients with glioma in the GEO dataset GSE4412. HR = hazard ratio.

cordingly, we hypothesized that TOP2A expr-

ession may be a prognostic factor for patients

with glioma. We used the PrognoScan

data-base to analyze the correlation of TOP2A ex-

pression with overall survival in GSE4271 and

GSE4412 enrolled from the Gene Expression

Omnibus (GEO) datasets with a total of 153

cases. The results show that TOP2A expres-

sion is highly associated with overall survival

of glioma patients (log-rank test, P < 0.001,

P = 0.0016,

Figure 4A

and

4B

).

Prediction of the role of TOP2A in glioma

stage of glioma, KI67 positive percentage,

IDH1 mutation and age, but other clinical

parameters, such as sex distribution and tu-

mor size were barely associated with high

TOP2A gene expression. However, there was

no evidence revealing arole of highly express-

ed TOP2A in the prognosis of glioma. In our

study, according to the PrognoScan database,

we could find its prognostic value: high expr-

ession of TOP2A caused shorter overall survi-

val time and shorter disease-free survival time.

Thus, the expression of TOP2A might become

a potential biomarker for the prognosis of

glioma.

As for the functional and pathway enrichment

analysis, TOP2A was strongly emphasized in

cell cycle pathways and DNA replication

pro-cesses, which areconnected with its physiolo-

gical functions [34, 35]. In our study, GCBI

anal-ysis found that mitotic cell cycle, positive

regu-lation of transcription from RNA polymerase II

promoter, positive regulation of apoptotic

pro-cess, DNA replication, phosphatidylinositol-me-

diated signaling, response to DNA damage

stimulus, and chromosome condensation pa-

thways were highly enriched in glioma sampl-

es with TOP2A up-regulated. Furthermore,

ear-lier clinical data indicate that TOP2A mRNA

and protein expression might participate in the

process of chromosome condensation,

chro-matid separation, and the relief of torsional

stress that occurs during DNA transcription and

replication [6]. According to our analysis, we

speculate that high expression of TOP2A at

mRNA and protein levels might be one of the

causes of cell cycle apoptotic process and

phosphatidylinositol-mediated signaling.

In conclusion, we used bioinformatics and

immunochemistry analysis to define the expr-

ession level of TOP2A in glioma. TOP2A was

identified in association with the progression

and prognosis of glioma, probably regulating

cell cycle apoptotic processes and

phosphati-dylinositol-mediated signaling.

Disclosure of conflict of interest

None.

Address correspondence to: Tianmin Zhou, Key Laboratory of Medicine and Biotechnology of Qing- dao, Department of Microbiology, Medical College of Qingdao University, 308 Ningxia Road, Qingdao, P.

R. China. E-mail: 329637042@163.com; Bin Wang, School of Basic Medical Sciences, Qingdao Uni- versity, Qingdao, P. R. China. Tel: +86-131-0772-05- 70; E-mail: wangbinqindao@163.com

References

Sumer SO, Sun Y, Tasman N, Taylor BS, Voet D, Weinhold N, Weinstein JN, Yang D, Yoshihara K, Zheng S, Zhang W, Zou L, Abel T, Sadeghi S, Cohen ML, Eschbacher J, Hattab EM, Raghu-nathan A, Schniederjan MJ, Aziz D, Barnett G, Barrett W, Bigner DD, Boice L, Brewer C, Cala-tozzolo C, Campos B, Carlotti CG Jr, Chan TA, Cuppini L, Curley E, Cuzzubbo S, Devine K, DiMeco F, Duell R, Elder JB, Fehrenbach A, Fin-occhiaro G, Friedman W, Fulop J, Gardner J, Hermes B, Herold-Mende C, Jungk C, Kendler A, Lehman NL, Lipp E, Liu O, Mandt R, McGraw M, Mclendon R, McPherson C, Neder L, Nguy-en P, Noss A, Nunziata R, Ostrom QT, Palmer C, Perin A, Pollo B, Potapov A, Potapova O, Rath-mell WK, Rotin D, Scarpace L, Schilero C, Sen-ecal K, Shimmel K, Shurkhay V, Sifri S, Singh R, Sloan AE, Smolenski K, Staugaitis SM, Steele R, Thorne L, Tirapelli DP, Unterberg A, Valluru-palli M, Wang Y, Warnick R, Williams F, Wolin-sky Y, Bell S, Rosenberg M, Stewart C, Huang F, Grimsby JL, Radenbaugh AJ, Zhang J. Compre-hensive, integrative genomic analysis of dif-fuse lower-grade gliomas. N Engl J Med 2015; 372: 2481-98.

[2] Rasmussen BK, Hansen S, Laursen RJ, Kostel-janetz M, Schultz H, Nørgård BM, Guldberg R and Gradel KO. Epidemiology of glioma: clini-cal characteristics, symptoms, and predictors of glioma patients grade I-IV in the the Danish neuro-oncology registry. J Neurooncol 2017; 135: 571-579.

[3] Wion D. Therapeutic dormancy to delay post-surgical glioma recurrence: the past, present and promise of focal hypothermia. J Neuroon-col 2017; 133: 447-454.

[4] van den Bent M, Chinot OL and Cairncross JG. Recent developments in the molecular charac-terization and treatment of oligodendroglial tumors. Neuro Oncol 2003; 5: 128-38. [5] Kalpathy-Cramer J, Gerstner ER, Emblem KE,

Andronesi O and Rosen B. Advanced magnetic resonance imaging of the physical processes in human glioblastoma. Cancer Res 2014; 74: 4622-4637.

[6] Tsavaris N, Lazaris A, Kosmas C, Gouveris P, Kavantzas N, Kopterides P, Papathomas T, Ara-pogiannis G, Zorzos H, Kyriakou V and Pat-souris E. Topoisomerase I and IIα protein ex-pression in primary colorectal cancer and recurrences following 5-fluorouracil-based ad-juvant chemotherapy. Cancer Chemother Pharmacol 2009; 64: 391-398.

[7] Heestand GM, Schwaederle M, Gatalica Z, Ar-guello D and Kurzrock R. Topoisomerase ex-pression and amplification in solid tumours: Analysis of 24,262 patients. Eur J Cancer 2017; 83: 80-7.

[8] Wang TL, Ren YW, Wang HT, Yu H and Zhao YX. Association of topoisomerase II (TOP2A) and dual-specificity phosphatase 6 (DUSP6) single nucleotide polymorphisms with radiation treat-ment response and prognosis of lung cancer in Han Chinese. Med Sci Monit 2017; 23: 984-993.

[9] Dimasgonzález J, Maldonadolagunas V, Díazchávez J, Lópezarellano ME, Muñozcama-cho J, Teránporcayo MA and Lagunasmartínez A. Overexpression of p53 protein is a marker of poor prognosis in Mexican women with breast cancer. Oncol Rep 2017; 37: 3026-3036. [10] Zhou Z, Liu S, Zhang M, Zhou R, Liu J, Chang Y

and Zhao Q. Overexpression of topoisomerase 2-alpha confers a poor prognosis in pancreatic adenocarcinoma identified by co-expression analysis. Dig Dis Sci 2017; 62: 2790-2800. [11] Yang FD and Jia ZL. [Expression of

topoisomer-ase II alpha in human colorectal carcinoma and its significance]. Nan Fang Yi Ke Da Xue Xue Bao 2010; 30: 1959-1961, 1964. [12] Arivazhagan A, Kumar DM, Sagar V, Patric IR,

Sridevi S, Thota B, Srividya MR, Prasanna K, Thennarasu K, Mondal N, Hegde AS, Chandra-mouli BA, Santosh V, Rao MR, Kondaiah P and Somasundaram K. Higher topoisomerase 2 al-pha gene transcript levels predict better prog-nosis in GBM patients receiving temozolomide chemotherapy: identification of temozolomide as a TOP2A inhibitor. J Neurooncol 2012; 107: 289-297.

[13] Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, Scheithauer BW and Klei-hues P. The 2007 WHO classification of tu-mours of the central nervous system. Acta Neuropathol 2007; 114: 97-109.

[14] Mizuno H, Kitada K, Nakai K and Sarai A. Prog-noScan: a new database for meta-analysis of the prognostic value of genes. BMC Med Ge-nomics 2009; 2: 18.

[15] Zaczek AJ, Markiewicz A, Seroczynska B, Skokowski J, Jaskiewicz J, Pienkowski T, Olsze-wski WP, Szade J, Rhone P and Welnicka-Jaskiewicz M. Prognostic significance of TOP2A gene dosage in HER-2-negative breast cancer. Oncologist 2012; 17: 1246-1255.

[16] Franceschini A, Lin J, Mering CV and Jensen LJ. SVD-phy: improved prediction of protein func-tional associations through singular value de-composition of phylogenetic profiles. Bioinfor-matics 2016; 32: 1085-7.

[17] Feng A, Tu Z and Yin B. The effect of HMGB1 on the clinicopathological and prognostic features of non-small cell lung cancer. Oncotarget 2016; 7: 20507-20519.

mRNA-lncRNA signature with predictive and prognostic value. Cancer Res 2016; 76: 2105-14.

[19] Kong F, Wei H, Kai Z, Xiao W, Kou Y, You H, Zheng K and Tang R. Hepatitis B virus X protein promotes interleukin-7 receptor expression via NF-κB and Notch1 pathway to facilitate prolif-eration and migration of hepatitis B virus-relat-ed hepatoma cells. J Exp Clin Cancer Res 2016; 35: 172.

[20] Wang G, Cao R, Wang Y, Qian G, Dan HC, Jiang W, Ju L, Wu M, Xiao Y and Wang X. Simvastatin induces cell cycle arrest and inhibits prolifera-tion of bladder cancer cells via PPARγ signal-ling pathway. Sci Rep 2016; 6: 35783. [21]

https://tcga-data.nci.nih.gov/docs/publica-tions/tcga/.

[22] Sun L, Hui AM, Su Q, Vortmeyer A, Kotliarov Y, Pastorino S, Passaniti A, Menon J, Walling J, Bailey R, Rosenblum M, Mikkelsen T and Fine HA. Neuronal and glioma-derived stem cell fac-tor induces angiogenesis within the brain. Can-cer Cell 2006; 9: 287-300.

[23] Liang Y, Diehn M, Watson N, Bollen AW, Aldape KD, Nicholas MK, Lamborn KR, Berger MS, Botstein D, Brown PO and Israel MA. Gene ex-pression profiling reveals molecularly and clini-cally distinct subtypes of glioblastoma multi-forme. Proc Natl Acad Sci U S A 2005; 102: 5814-9.

[24] French PJ, Swagemakers SM, Nagel JH, Kou-wenhoven MC, Brouwer E, van der Spek P, Lu-ider TM, Kros JM, van den Bent MJ and Sillevis Smitt PA. Gene expression profiles associated with treatment response in oligodendroglio-mas. Cancer Res 2005; 65: 11335-11344. [25] Shai R, Shi T, Kremen TJ, Horvath S, Liau LM,

Cloughesy TF, Mischel PS and Nelson SF. Gene expression profiling identifies molecular sub-types of gliomas. Oncogene 2003; 22: 4918-4923.

[26] Murat A, Migliavacca E, Gorlia T, Lambiv WL, Shay T, Hamou MF, de Tribolet N, Regli L, Wick W, Kouwenhoven MC, Kouwenhoven MC, Hain-fellner JA, Heppner FL, Dietrich PY, Zimmer Y, Cairncross JG, Janzer RC, Domany E, Delorenzi M, Stupp R and Hegi ME. Stem cell-related “self-renewal” signature and high epidermal growth factor receptor expression associated with resistance to concomitant chemoradio-therapy in glioblastoma. J Clin Oncol 2008; 26: 3015-3024.

[27] Rhodes DR, Yu J, Shanker K, Deshpande N, Varambally R, Ghosh D, Barrette T, Pandey A and Chinnaiyan AM. Large-scale meta-analysis of cancer microarray data identifies common transcriptional profiles of neoplastic transfor-mation and progression. Proc Natl Acad Sci U S A 2004; 101: 9309-9314.

[28] Kapur K, Xing Y, Ouyang Z and Wong WH. Exon arrays provide accurate assessments of gene expression. Genome Biol 2007; 8: R82. [29] Fack F, Espedal H, Keunen O, Golebiewska A,

Obad N, Harter PN, Mittelbronn M, Bähr O, Weyerbrock A, Stuhr L, Miletic H, Sakariassen PØ, Stieber D, Rygh CB, Lund-Johansen M, Zheng L, Gottlieb E, Niclou SP and Bjerkvig R. Bevacizumab treatment induces metabolic ad-aptation toward anaerobic metabolism in glio-blastomas. Acta Neuropathol 2015; 129: 115-31.

[30] Chen R, Cohen AL and Colman H. Targeted therapeutics in patients with high-grade glio-mas: past, present, and future. Curr Treat Op-tions Oncol 2016; 17: 42.

[31] Grosu AL, Weber WA, Franz M, Stärk S, Piert M, Thamm R, Gumprecht H, Schwaiger M, Molls M and Nieder C. Reirradiation of recurrent high-grade gliomas using amino acid PET (SPECT)/CT/MRI image fusion to determine gross tumor volume for stereotactic fractionat-ed radiotherapy. Int J Radiat Oncol Biol Phys 2005; 63: 511-519.

[32] Ravasz E, Somera AL, Mongru DA, Oltvai ZN and Barabási AL. Hierarchical organization of modularity in metabolic networks. Science 2002; 297: 1551-5.

[33] Lazaris AC, Kavantzas NG, Zorzos HS, Tsavaris NV and Davaris PS. Markers of drug resistance in relapsing colon cancer. J Cancer Res Clin Oncol 2002; 128: 114-118.

[34] Deweese JE and Osheroff N. The DNA cleavage reaction of topoisomerase II: wolf in sheep’s clothing. Nucleic Acids Res 2009; 37: 738-48. [35] Xie ZC, Dang YW, Wei DM, Chen P, Tang RX,