A DNA Methylation Prognostic Signature of Glioblastoma: Identiﬁcation of NPTX2-PTEN-NF-Kb Nexus

Published OnlineFirst September 27, 2013; DOI: 10.1158/0008-5472.CAN-13-0298

Cancer Integrated Systems and Technologies Research

A DNA Methylation Prognostic Signature of Glioblastoma: Identiﬁcation of NPTX2-PTEN-NF-kB Nexus

Sudhanshu Shukla1, Irene Rosita Pia Patric1, Sivaarumugam Thinagararjan1, Sujaya Srinivasan1, Baisakhi Mondal1, Alangar S. Hegde2, Bangalore A. Chandramouli3, Vani Santosh4, Arimappamagan Arivazhagan3, and Kumaravel Somasundaram1

Abstract Glioblastoma (GBM) is the most common, malignant adult primary tumor with dismal patient survival, yet the molecular determinants of patient survival are poorly characterized. Global methylation profile of GBM samples (our cohort; n ¼ 44) using high-resolution methylation microarrays was carried out. Cox regression analysis identified a 9-gene methylation signature that predicted survival in GBM patients. A risk-score derived from methylation signature predicted survival in univariate analysis in our and The Cancer Genome Atlas (TCGA) cohort. Multivariate analysis identified methylation risk score as an independent survival predictor in TCGA cohort. Methylation risk score stratified the patients into low-risk and high-risk groups with significant survival difference. Network analysis revealed an activated NF-kB pathway association with high-risk group. NF-kB inhibition reversed glioma chemoresistance, and RNA interference studies identified interleukin-6 and intercellular adhesion molecule- 1 as key NF-kB targets in imparting chemoresistance. Promoter hypermethylation of neuronal pentraxin II (NPTX2), a risky methylated gene, was confirmed by bisulfite sequencing in GBMs. GBMs and glioma cell lines had low levels of NPTX2 transcripts, which could be reversed upon methylation inhibitor treatment. NPTX2 overexpression induced apoptosis, inhibited proliferation and anchorage-independent growth, and rendered glioma cells chemosensitive. Furthermore, NPTX2 repressed NF-kB activity by inhibiting AKT through a p53-PTEN- dependent pathway, thus explaining the hypermethylation and downregulation of NPTX2 in NF-kB-activated high- risk GBMs. Taken together, a 9-gene methylation signature was identified as an independent GBM prognosticator and could be used for GBM risk stratification. Prosurvival NF-kB pathway activation characterized high-risk patients with poor prognosis, indicating it to be a therapeutic target. Cancer Res; 73(22); 1–11. 2013 AACR.

Introduction drogenase 1 (IDH1) mutation status, and Glioma-CpG Island fi The grade IV glioma (glioblastoma; GBM) is the most Methylator Phenotype (G-CIMP) have been identi ed as GBM common, malignant primary brain tumor in adults. The prog- prognostic indicators (1, 8, 9). fi fi nosis remains poor with median survival ranging from 12 to 15 Expression pro ling studies have identi ed mRNA and fi months in spite of improvements in treatment protocol (1). microRNA signatures for classi cation and prognosis in GBM – Genetic heterogeneity in GBM has been proposed to explain (10 14). However, none of the gene signatures have been the limitations in the effectiveness of current therapies, which translated into clinics, suggesting the need for more robust fi necessitates the need for prognostic gene signature (2). Many prognostic gene signature panels. Here, we have identi ed and genetic and epigenetic alterations as well as expression of some validated a 9-gene methylation signature for GBM prognosti- genes have been correlated with poor or better prognosis (3–7). cation. Multivariate analysis with all known prognostic mar- fi In addition, molecular biomarkers like methyl guanine methyl kers and signatures for GBM identi ed our methylation sig- transferase (MGMT) promoter methylation, isocitrate dehy- nature to be an independent GBM prognostic indicator. Fur- thermore, an activated NF-kB pathway was found to be associated with poor prognosis. We also show that neuronal Authors' Affiliations: 1Department of Microbiology and Cell Biology, pentraxin II (NPTX2), a component of methylation signature Indian Institute of Science; 2Sri SatyaSai Institute of Higher Medical with risky methylation, inhibits cell growth by antagonizing 3 4 Sciences; Departments of Neurosurgery and Neuropathology, National NF-kB pathway through p53-PTEN activation, thus connecting Institute of Mental Health and Neuro Sciences, Bangalore, India the methylation signature to NF-kB pathway. Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). Materials and Methods Corresponding Author: Kumaravel Somasundaram, Department of Microbiology and Cell Biology, Indian Institute of Science, Bangalore Plasmids and reporter constructs 560012, India. Phone: 91-80-23607171; Fax: 91-80-23602697; E-mail: pCMV-Entry/NPTX2 was obtained from Origene. CAPI3K [email protected] (pCDNA3-CD2p110myc) was described earlier (15, 16). CAAKT doi: 10.1158/0008-5472.CAN-13-0298 (pCDNA MYR HA AKT1) was from Addgene (Plasmid 9008). 2013 American Association for Cancer Research. PG13-Luc was described earlier (17). CAIKK, NF-kB-Luc,

www.aacrjournals.org OF1

Shukla et al.

PTEN-luc, shATM, and shATR plasmid constructs were target genes were used to generate an interaction network obtained from Profs. I. Verma, K.N. Balaji, Dr. R.C.M. Simmen, using STRING database (19) using "experiment," "textmining," Dr. Y. Shiloh, and Dr. T. De Lange, respectively. and 0.700 confidence level as input option. To make a protein– protein interaction network, STRING curates publically avail- Cell lines and reagents able databases for the information on direct and indirect Temozolomide, adriamycin, 5-aza-20-deoxycytidine, MTT, functional protein interactions. The data file of protein–pro- and EscortIII transfection reagent were purchased from Sigma. tein interaction network generated from STRING was then Glioma cell lines U138, LN18, U343, LN229, U251, U87, T98G, used in Medusa software (20) to construct a network. This U373 (all human glioma derived), 293, and C6 (Rat glioma network was then analyzed by ViaComplex software (21). derived) were obtained from the laboratory of Dr. A. Guha, ViaComplex plots the gene expression activity over the Medusa University of Toronto, Toronto, Canada, and grown in DMEM network topology. ViaCompex overlays expression data with medium. SVG cells were obtained from Dr. P. Seth, National the interaction network and constructs a landscape graph by Brain Research Center, New Delhi, and grown in MEM medi- distributing microarray signal with interaction network coor- um. The medium was supplemented with 10% FBS, penicillin, dinates. For each gene, the average of each group (high-risk and and streptomycin. No information is available about their low-risk) was plotted over the average of the same gene in authentication, although many common known mutations/ normal brain and landscape was constructed. alterations have been verified in our laboratory. Other methods and additional information Patient samples and clinical data The details for additional clinical data, genomic DNA extrac- Tumor samples used for study were obtained from patients tion, sodium bisulfite conversion, methylation array analysis, who were operated at Sri Sathya Sai Institute of Higher Medical bisulfite sequencing, RNA isolation, real-time quantitative Sciences and National Institute of Mental Health and Neuro- RT-PCR analysis, 5-Aza-2'-deoxycytidine treatment, transfec- sciences, Bangalore, India. Normal brain tissue samples (ante- tion, colony formation assays, stable cell line generation, rior temporal lobe) obtained during surgery for intractable proliferation assay, cytotoxicity assay, CHIP assay, and PTEN epilepsy were used as control samples. Tissues were bisected activity ELISA are given in the Supplementary information. and one half was snap-frozen in liquid nitrogen and stored at 80C until DNA/RNA isolation. The other half was fixed in Statistical analysis formalin and processed for paraffin sections. These sections To identify the CpGs whose methylation correlates with were used for the histopathologic grading of tumor and survival, methylation data, available for 44 patients, were immunohistochemical staining. considered for survival analysis. The probes with missing value imputed using k nearest neighbor algorithm using Network analysis Imputation package in R. The assessment of survival correla- To find out the key signaling pathways differently regulat- tion for all 27,578 probes was performed using BRB-Array- ed between low-risk and high-risk patients, we compared gene Tools survival analysis (BRB-ArrayTools is available without expression profiles of The Cancer Genome Atlas (TCGA) data charge for non-commercial applications at http://linus.nci. set (Agilent expression data, 17,814 genes) with gene expression nih.gov/BRB-ArrayTools.html). Firstly, for each CpG probe, differences between low-risk and high-risk groups by conduct- univariate Cox proportional hazard regression was done and ing t-test analysis with FDR correction. We identified 3,127 parametric P value was calculated. The survival time and genes that were significantly differentially regulated between censoring status were randomly assigned to each patient the high-risk and low-risk groups (1,983 genes upregulated and and univariate Cox proportional hazard regression was 1,144 genes were downregulated in high-risk compared to performed to re-estimate the P value for each probe. This low-risk). To functionally interpret the role of differentially step was repeated 10,000 times and permutation P value was regulated genes with respect to patient survival, we carried out calculated. To select the probes with maximum variation in biological network analysis using MetaCore from Thomson b value, a standard deviation criteria of >0.2 was applied. Reuters, a pathway mapping tool. This tool builds biological Based on these criteria, we identified 9 probes that strongly networks from the input gene set and lists the biological proces- correlated with survival. ses (Gene Ontology terms) associated with each network. From The methylation risk score was calculated as follows. Using the list of 3,127 differentially regulated genes, we chose a subset these 9 probes, we devised a formula, resulting in a methylation of 61 genes, with a log2 fold change >1.5 as the input for Meta- risk score that would combine their independent survival core software. Using variant shortest path algorithm with main predicting capabilities (based on Cox regression coefficients parameters of relative enrichments of input genes and relative derived from Cox proportional hazards analysis). Each patient saturation of networks with canonical pathways, Metacore was assigned a risk score that is a linear combination of b value identified 8 significant networks (Supplementary Table S3). of the nine selected probes weighted by their respective Cox regression coefficients. NF-kB gene/protein association network and landscape The internal validation of the risk score was carried out analysis of gene expression using ROC analysis and subset analysis. Independent valida- To construct the NF-kB gene/protein interaction network, tion of risk score in other data sets, multivariate analysis, and NF-kB target genes were downloaded (18). A total of 150 NF-kB additional details are given in the Supplementary Information.

OF2 Cancer Res; 73(22) November 15, 2013 Cancer Research

GBM Prognostic DNA Methylation and NPTX2-NF-kB Link

GBM patients our cohort (n = 44) (training set)

Whole genome methylation assay Internal validation Cox regression analysis and P value cut-off

9-gene signature Area under curve Random (AUC) of subset sampling risk-score Risk-score calculation

Validation-TCGA cohort Indirect validation of risk-score (testing set; n = 205)

Univariate and Univariate and REMBRANDT Phillips et al Stratification into Stratification into multivariate multivariate dataset (n = 157) dataset (n = 77) low- and high-risk low- and high- Cox regression Cox regression Unsupervised groups risk groups analysis analysis clustering

Stratification into low- Stratification into low and high-risk groups and high-risk groups

Differential expression analysis Validation-Bent Colleagues cohort (testing set; n = 67) between low- and high-risk groups Landscape Network analysis Landscape analysis for low- (Metacore) and high-risk groups analysis for high- risk compared to Identification of enriched compared to Univariate and k low-risk groups Stratification into networks and NF- B normal group multivariate pathway genes low- and high- Cox regression STRING and risk groups analysis Medusa software

Landscape analysis for low- and high-risk compared to normal; Identification of NF-kB activation in high-risk groups

Figure 1. Schematic diagram of the entire workﬂow as to how the 9-gene signature was identiﬁed, tested, and validated is shown.

Results indicated by very short median survival (Supplementary Fig. Identification and validation of a 9-gene prognostic S1B and S1C and Supplementary Materials). methylation signature for GBM The risk score distribution and a comparison of risk score With an aim to find methylation signature predictive of with patient survival status among GBMs of our and TCGA – GBM survival, Cox proportional hazards regression was cohort are shown (Supplementary Fig. S2A S2D). Upon corre- carried out by comparing methylation status of 27,578 CpGs lation of methylation status to patient survival, NPTX2 meth- and patient's (our cohort; n ¼ 44) survival data. The entire ylation was found to be risky, whereas methylation of the workflow as to how the methylation signature was identi- remaining 8 genes was found to be protective (Fig. 2D, Table fied, tested, and validated is shown in a schematic diagram 1, and Supplementary Fig. S2E). GBM tumors from high-risk (Fig. 1). We identified 9 probes whose methylation status patients tend to have more methylation of NPTX2 whereas correlated significantly with patients' survival (Table 1 and tumors from low-risk group tend to have more methylation of Supplementary Materials). A methylation risk score calcu- the remaining genes. Additional analysis revealed that all 9 lated for each patient by combining the effect of each of the 9 genes, which form part of the methylation signature, are needed genes using a risk score formula (Supplementary Materials) for prognostication (Supplementary Materials). divided the patients into high-risk and low-risk groups with Multivariate regression analysis indicates 9-gene significant survival difference (median survival: 12 months methylation signature is an independent prognosticator vs. 23 months; Fig. 2A and Supplementary Table S1). The Cox multivariate analysis with age, we found methylation risk strength of methylation risk score in predicting patient score to be an independent predictor of GBM patient survival in survival was further confirmed by ROC curve analysis and our patient set (P ¼ 0.001; HR ¼ 1.267; B ¼ 0.237). Additional subset analysis (Supplementary Fig. S1A and Supplementary multivariate analysis with various prognostic factors was carried Materials). The methylation risk score predicted survival in out using TCGA dataset. Although MGMT promoter methylation GBM dataset derived from TCGA cohort as well (Fig. 2B and (CpG probe ID #Cg02941816), IDH1 mutation, and Glioma-CpG Supplementary Table S1). The methylation risk score also Island Methylator Phenotype (G-CIMP) were found to be inde- divided patients significantly into 3 groups in both datasets pendent predictors when compared with age, an analysis that with the identification of a third group with very high risk as included all markers identified methylation risk score alone to be

www.aacrjournals.org Cancer Res; 73(22) November 15, 2013 OF3

Shukla et al.

signiﬁcant with MGMT promoter methylation showing a trend toward signiﬁcance(Table2).Interestingly,wefoundthatall IDH1 mutant patients (n ¼ 9) and all G-CIMP þ patients (n ¼ 11)

0.17 were a subset of low-risk group (n ¼ 16;Fig.2E). Upon correlation with gene expression subtypes, we noticed value) that although the low-risk group is highly enriched with pro- b neural GBM tumors (9 of 12; 75%), the high-risk group was moderately enriched for mesenchymal GBM tumors (16 of 30;

TCGA dataset 53%; Fig. 2E). However, survival analysis revealed that proneural (Median low-risk patients survived signiﬁcantly longer than proneural high-risk patients or all other nonproneural GBM patients (Fig. 2C). On comparison with recently published 9-gene predictor (22) using TCGA dataset, we found that both signatures were signiﬁcant in predicting survival (Supplementary Table S2). On comparison with 10-microRNA prognostic signature (14) using

0.35 0.4 0.58 TCGA dataset, we found the methylation risk score remain signiﬁcant whereas the microRNA signature showed a trend

value) toward signiﬁcance (Supplementary Table S2). Furthermore, b multivariate analysis using another cohort of glioma grade III (N ¼ 67) also identiﬁed methylation risk score to be an independent survival predictor (Supplementary Table S3 and Fig. S3; (Median

Our patient dataset ref. 23). Thus, these results put together identify methylation risk score as an independent predictor of patient survival.

Network analysis reveals an activated NF-kB pathway in high-risk patients

HR Low risk High risk Difference Low risk High risk Difference We hypothesized that the 9 genes, which form part of the methylation signature, might modify the global gene expression 2.62 0.073 0.812.86 0.057 0.572.23 0.51 0.107 0.5 0.24 0.3 0.29 0.33 0.73 0.21 0.71 0.55 0.33 0.56 0.18 0.41 0.38 0.16 2.49 0.0832.6 0.62.19 0.074 0.1122.4 0.66 0.72 0.091 0.172.32 0.69 0.31 0.098 0.42 0.61 0.46 0.43 0.35 0.28 0.3 0.72 0.22 0.67 0.81 0.33 0.26 0.7 0.3 0.57 0.66 0.47 0.48 0.36 0.35 0.24 0.21 0.31 B consequently, altering some key signaling pathways differently between low-risk and high-risk patients, thus explaining the difference in their survival. Biological network analysis using differentially regulated genes between low-risk and high-risk value 0.0018 0.0055 P 0.0037 0.0084 0.007 0.0081 0.0043 1.91 6.783 0.15 0.5 groups derived from TCGA (Supplementary Table S4) identiﬁed 8signiﬁcant networks (Supplementary Table S5). Careful analysis revealed an enrichment of NF-kB pathway members in these networks. As NF-kB pathway is associated with oncogen- esis, tumor progression, and chemoresistance in many cancers (24), we tested whether the NF-kB pathway is differentially activated between low-risk and high-risk groups. First, NF-kB gene/protein interaction network was constructed using the STRING database with 150 NF-kB target genes as input (Sup- plementary Fig. S4A). Landscape analysis of the NF-kB interaction network using GBM patient gene expression data of low- Proein 1 factor 6 protein 1 suppressor risk versus normal and high-risk versus normal from the TCGA cohort revealed that NF-kB interaction network is very mini- mally upregulated in low-risk (as seen by very little red and orange color in z-axis plot; Fig. 3A). However, the GBM tumors from the high-risk group showed much higher level of NF-kB pathway activation as evidenced by the bright red and orange color in z-axis plots (Fig. 3B), suggesting the possibility that higher activation of NF-kB pathway in high-risk GBM may be responsible for their poor treatment response and less survival. LIMD1 (cg04037228) LIM domain-containing NPTX2 (cg12799895) Neuronal pentraxin2 MOXD1 (cg13603171) DBH-like monooxygenase FMOD (cg26987645)LAD1 (cg25947945)RBPSUHL (cg21835643) Fibromodulin Recombining binding protein Ladinin-1 GSTM5 (cg04987894) Glutathione S-transferase Mu 5 0.0094 IRF6 (cg23283495)FBP1 (cg17814481) Interferon regulatory Fructose-1,6-bisphos-phatase 1 0.0056 In good correlation, pretreatment with NF-kB inhibitor made Details about 9-gene methylation signature that predicts survival in GBM only those cell lines, where NF-kB is already activated, more chemosensitive but not in other cell lines, wherein NF-kB activation is very low (Fig. 3C and D and Supplementary Fig. – k Table 1. Serial No. Symbol (CpG ID) Name 2 1 3 4 5 6 7 8 9 S4B S4D). Furthermore, silencing of p65 subunit of NF- B rendered glioma cells more sensitive to chemotherapy (Fig. 3E).

OF4 Cancer Res; 73(22) November 15, 2013 Cancer Research

GBM Prognostic DNA Methylation and NPTX2-NF-kB Link

ABLow risk (n = 12; C Low risk (n = 37; Low risk proneural (n = 9; median survival 100 median survival median survival = 53 months) = 23 months) 100 100 = 23 months) High risk proneural (n = 5; High risk (n = 32; median survival = 10 months) 80 80 median survival High risk (n = 168; 75 Non proneural (n = 28; = 12 months) median survival median survival = 17 months) 60 60 P = 0.001 = 15 months) P = 0.028 P = 0.0007 50 40 40

25 Percent survival Percent survival 20 20 Percent survival

0 0 0 01020304050 020406080100 0255075100 Over all survival (months) Over all survival (months) Over all survival (months)

D NPTX2 LIMD1 MOXD1 IRF6 FBP1 FMOD LAD1 RBPSUHL GSTM5 Low risk High risk

0 0.4 0.8 KEY Color code for heat map 65 TCGA GBM samples 0 0.4 0.8 E G-CIMP status G-CIMP status (n = 65) Glioma subtypes Cluster1-G-CIMP + (n = 11) IDH1 status Cluster3 -G-CIMP – (n = 36) Cluster2 -G-CIMP – (n = 18) NPTX2 LIMD1 Gene expression subtypes (n = 42) Proneural (n = 14) IRF6 Classical (n = 7) MOXD1 Neural (n = 5) FBP1 Mesenchymal (n = 16) FMOD N/A (n = 23) LAD1 IDH1 mutation status (n = 65) RBPSUHL IDH1 WT GSTM5 IDH1 Mu Low risk High risk

Figure 2. Risk stratiﬁcation of GBM based on methylation risk score. A and B, Kaplan–Meier graphs of (25th percentile used as cut-off) our patient cohort and TCGA cohort, respectively. C, Kaplan–Meier survival curve among proneural low-risk (green), proneural high-risk (red), and all nonproneural (blue) GBM tumors. D, heat map of methylation status (b value) of 9 genes from methylation signature (our patient set; n ¼ 44). E, heat map of methylation status (b value) of 9 genes from methylation signature (TCGA set; n ¼ 65) for whom G-CIMP status, glioma gene expression subtypes, IDH1 mutation status, and methylation b values were available. D and E, the rows represent genes and columns represent patients. A color code with yellow and blue indicating high and low methylation, respectively, was used. The red line divides patients into low-risk and high-risk groups. E, above the heat map, each sample is additionally color coded as described in the KEY.

It is interesting to note that several NF-kB target genes with genes (n ¼ 3,127) that are differentially regulated between low- antiapoptotic, prometastasis, prosurvival, and proinflamma- risk and high-risk groups in TCGA dataset (Supplementary tory functions were upregulated only in GBM tumors from Table S4) was able to divide these datasets into 2 clusters with high-risk patients (Supplementary Table S6), but were either significant survival difference and NF-kB activation (Supple- downregulated or not changed in tumors from low-risk mentary Figs. S5A and S5B and S6A–S6E). These results overall patients, when compared to normal, thus suggesting the validate the survival prediction by our methylation signature potential roles of these set of genes in imparting chemoresis- and confirm the association of activated NF-kB pathway with tance. Silencing each of the 7 selected target genes identified poorer prognosis in high-risk GBMs. interleukin (IL)-6 and intercellular adhesion molecule (ICAM)- 1 as the key NF-kB target genes to be the mediators of NPTX2 is methylated in GBM and has a growth inhibitory chemoresistance, as their downregulation rendered glioma function cells sensitive to chemotherapy consistently (Fig. 3F). NPTX2 gene, a component of methylation signature, was Association of NF-kB with high-risk GBM was further val- particularly interesting as its methylation is risky. Bisulfite idated by means of an indirect approach using REMBRANDT sequencing validated the hypermethylation in GBMs and glioma (25) and Phillips (26) GBM datasets, which had both gene cell lines as against normal brain samples (Fig. 4A). NPTX2 expression data and patients survival information (Supple- transcripts were found to be downregulated in GBM samples mentary Materials). The expression levels of the same set of and glioma-derived cell lines compared to normal brain samples

www.aacrjournals.org Cancer Res; 73(22) November 15, 2013 OF5

Shukla et al.

Table 2. Multivariate Cox regression analysis of methylation risk score and other prognostic markers using TCGA cohort

Factor No. of patients HR B (coefﬁcient) P value I. Univariate analysis TCGA dataset Age 205 1.033 0.033 <0.0001 MGMT 205 0.226 1.325 0.004 IDH1 197 0.441 0.916 <0.0001 G-CIMP 205 0.298 1.211 <0.0001 Methylation risk score 205 1.146 0.136 <0.0001 II. Multivariate analysis with TCGA dataset Age 205 1.027 0.027 <0.0001 Methylation risk score 1.109 0.103 0.005 MGMT 205 0.320 1.138 0.013 Methylation risk score 1.14 0.131 <0.0001 IDH1 65 0.476 0.472 0.316 Methylation risk score 1.026 0.029 <0.0001 G-CIMP 205 0.400 0.916 0.215 Methylation risk score 1.030 0.030 <0.0001 III. Multivariate analysis of all the markers in TCGA dataset Age 65a NA NA 0.275 MGMT NA NA 0.064 IDH1 NA NA 0.322 G-CIMP NA NA 0.416 Methylation risk score 1.250 0.223 <0.0001 III. Multivariate analysis of all the markers in TCGA dataset (except IDH1) Age 205 1.026 0.026 0.001 MGMT 0.329 1.112 0.016 G-CIMP 0.942 0.059 0.904 Methylation risk score 1.100 0.096 0.038

aMultivariate analysis that included all markers was carried out in 2 ways as IDH1 mutation was available only for 65 patients among 205 patients who had CpG methylation values for all 9 probes that formed part of methylation signature. The analysis that included MGMT, IDH1, G-CIMP, age, and methylation risk score was carried out with 65 patients, whereas the analysis involving MGMT, G-CIMP, age, and methylation risk score (without IDH1) was carried out with 205 patients.

(Fig. 4B and Supplementary S7A) and methylation inhibitor, (Fig. 5A), sequence-specific DNA-binding (Supplementary Fig. 5-aza-20-deoxycytidine, treatment resulted in reexpression of S8A), nuclear translocation of NF-kB subunit (p65); Supplemen- NPTX2 transcripts (Fig. 4C). NPTX2 transcripts levels inversely tary Fig. S9A–S9C), and repressed 6 of the 7 key NF-kBtarget correlated with NPTX2 promoter methylation (Supplementary genes significantly (Supplementary Fig. S8B). We also show that Fig. S7B). Further investigation revealed that exogenous over- NPTX2 inhibition of NF-kB could be abrogated by coexpression expression of NPTX2 suppressed colony formation (Fig. 4D). of constitutively active forms of PI3 kinase, AKT, and IKKa (Fig. NPTX2 stable clone of U343 glioma cell line, which overex- 5B). Furthermore, NPTX2 overexpression activated PG13-Luc, a presses NPTX2 transcripts and protein (Supplementary p53-dependent reporter, efficiently in U343 and U87 cells, in a Fig. S7C and S7D), showed increased apoptosis, reduced prolif- concentration-dependent manner in 293 cells as well as in a p53- eration, decreased anchorage-independent growth, and increas- dependent manner as it failed to activate in p53 mutant U251 ed chemosensitivity compared to a vector stable (Fig. 4E–H). cells (Fig. 5C) and p53 silenced cells (Supplementary Fig. S10A Furthermore, 5-aza-20-deoxycytidine pretreatment sensitized and S10D). NPTX2 overexpression also induced p53 protein U251glioma cells to chemotherapy in an NPTX2-dependent levels (but not transcript levels) and its targets, p21 and PTEN manner, as simultaneous silencing of NPTX2 resulted in chemo- protein levels (Fig. 5E and Supplementary Fig. S8C). Aptly, resistance (Fig. 4I and Supplementary S7E). NPTX2-activated expression from PTEN promoter reporter construct in U343 and U87 cells and in a concentration-depen- NPTX2 inhibits NF-kB pathway through p53-PTEN-PI3K- dent manner in 293 cells as well as in a p53-dependent manner AKT–dependent manner as it failed to activate PTEN reporter in p53 mutant U251 cells Next upon investigation of the connection between NPTX2 (Fig. 5D) and p53 silenced cells (Supplementary Fig. S10B). and NF-kB pathway, we found NPTX2 overexpression inhibited NPTX2 also induced PTEN transcript (Supplementary Fig. expression from NF-kB luciferase reporter in glioma cell lines S8D). More importantly, NPTX2 overexpression resulted in

OF6 Cancer Res; 73(22) November 15, 2013 Cancer Research

GBM Prognostic DNA Methylation and NPTX2-NF-kB Link

A B 1.0 1.0 1.0 1.0 0.8 0.8 0.8 0.8

Z-axis 0.6 Z-axis 0.6 0.6 0.6

Y-axis 0.4 0.4 Y-axis 0.4 0.4

0.2 0.2 0.2 0.2 Decrease Increase Decrease Increase 0.0 0.0 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 X-axis X-axis

C TMZ (125 µmol/L) E BAY (4 µmol/L) + TMZ (125 µmol/L) 140 P < 0.0001 P = 0.0009 P < 0.0001 P < 0.0001 120 120 P = 0 .013 P = 0.024 P = 0.0249 P < 0.001 NS NS NS 100 100 80 siControl 80 k 60 siNF- B 60 40 40 Percent viability 20 Percent viability 20 0 0 U343 T98G C6 U138 U251 SVG LN229 Temozolomide (µmol/L) 250 - 250 - NF-kB activated NF-kB not activated Adriamycin (µg/mL) - 0.1 - 0.1 U343 U138 D Adria (0.5 µg/mL) F 140 BAY (4 µmol/L) + Adria (0.5 µg/mL) 140 120 P = 0013 P = 0.013 P = 0.01 P < .003 NS NS NS 120

100 50 100 siControl 80 80 siLGALS3 siIL6

60 60 Not done siBCL3 siICAM1 siIER3 40 Percent IC 40 siCCL7 Percent viability 20 20 siF3 0 0 U343 T98G C6 U138 U251 SVG LN229 Adria (µg/mL) TMZ (µmol/L) Adria (µg/mL) TMZ (µmol/L) NF-kB activated NF-kB not activated U138 U343

Figure 3. NF-kB pathway association with high-risk GBM. A and B, landscape analysis of NF-kB network (Supplementary Fig. S4A) using NF-kB target gene expression in low-risk group compared to normal brain (A) and high-risk group compared to normal brain (B). C and D, human glioma cell lines were treated with BAY 11-7082 (4 mmol/L) for 12 hours and then treated with indicated drugs for 48 hours followed by viability measurement. The proportion of viable cells from untreated cells/Bay-treated cells was considered as 100% (NS—P > 0.05). E, U343 and U138 cells were transfected with siRNAs against NF-kB (p65) and, after 36 hours, treated with indicated drugs. The proportion of live cells was quantiﬁed after 48 hours. The proportion of viable cells from siControl was considered as 100%. F, U343 and U138 cells were transfected with siRNAs against indicated genes and after 36 hours, treated with temozolomide (125, 250, 500, and 1,000 mmol/L) and adriamycin (0.1, 0.2, 0.4, and 0.8 mg/mL). After 48 hours of the drug treatment, proportion of live cells was quantiﬁed. IC50 values of control siRNA were considered as 100% and the ratio of control siRNA IC50 compared with that of indicated siRNA is shown as percent IC50.

4.8-fold increase in p53 occupancy in PTEN promoter (Fig. 5F), Both ATM and ATR have been shown to have overlapping significant activation of PTEN phosphatase activity (3–3.5 fold) and independent functions in phosphorylation and subsequent as seen by increased PIP2 levels(Fig.5G),andanefficient activation of p53 during cellular genotoxic stress (27). To define inhibition of AKT as seen by reduced pAKT levels (Fig. 5E). the role of ATM/ATR proteins in NPTX2 activation of p53, the NPTX2 overexpression also decreased pIkB and increased total ability of NPTX2 to activate p53-dependent reporter activity IkB levels (Supplementary Fig. S9D). In good correlation, NPTX2 was measured in either ATM or ATR silenced cells. Interestingly, failed to inhibit NF-kB promoter in p53 mutant U251 cells we found that NPTX2 activation of p53 activity was compro- (Fig.5A),p53silencedcells(SupplementaryFig.S10C),andin mised significantly in ATM silenced cells (Supplementary Fig. PTEN null U87 cells (Fig. 5A). These results together suggest that S10E and S10F) but not in ATR silenced cells (data not shown). NPTX2 is a glioma growth inhibitory gene, which can modulate chemosensitivity, and its antiproliferative functions may involve Discussion inhibition of NF-kB pathway through p53-dependent PTEN In this study, using a cohort of patients of newly diagnosed activation, leading to inhibition of PI3K-AKT-IKKa signaling. GBM, treated uniformly and followed up prospectively, we

www.aacrjournals.org Cancer Res; 73(22) November 15, 2013 OF7

Shukla et al.

A BC TSS NPTX2

P = 0.002 P < 0.01 P < 0.001 5.0 CpG island 2 4.5 BSG % 4.0 Methylation 0 MI: 0.13% 3.5 Normal 0.0 0.0 3.0 brain 0.0 samples -2 0.5 ratio

2 2.5

MI: 57.00% ratio 35 2 -4 2.0

84 Log 63 1.0 GBM 58 Log 42 -6 1.5 60 0 MI: 64.75% -8 Aza (µmol/L) 105105 105105 105105 88 Normal GBM Normal GBM Normal GBM Glioma 60 Days 64 64 64 cell lines 62 Our patient Rembrandt TCGA 49 set set U251 LN18 U87 Analysed set in Infinium 0-30% 31-60% 61-100%

U343/Vector U343/NPTX2

D F H P = 0.0002 P = 0.0013 P = 0.005 Vector U343/Vector NPTX2 120 ) 2.5 U343/NPTX2 P < 0.001 P < 0.01 P < 0.001 6 100 U343/Vector 50 2.0 80 U343/NPTX2 100 60 1.5 80 40 60 1.0

Percent IC 20 40 0.5 0 20 Adria Taxol Temo

0 No. of cells (×10 0 U343 U251 LN229 Days 1 246 No. of colonies (%)

E G I P = 0.013 250 P = 0.017 200 Adriamycin 250 120 0.4 µg/mL U343/Vector 200 100 150 Temozolomide 500 µmol/L 80 150 100 60 50 100 U343/NPTX2 40 Percent survival Percent

% Colonies 0

Mean intensity 50 20 1 2 3 4 + + - - 0 0 Control siRNA U343/ U343/ U343/ U343/ NPTX2 siRNA - - + + Vector NPTX2 vector NPTX2 Chemo + + + + Aza cytidine - + - +

Figure 4. Validation of methylation status and growth inhibitory function of NPTX2 gene. A, bisulfite sequence analysis of the NPTX2 promoter. Percentage methylation was established as total percentage of methylated cytosines from 7 to 10 randomly sequenced colonies. B, NPTX2 transcript levels obtained from RT-qPCR (this study group) or microarray data (TCGA and REMBRANDT) are plotted. C, total RNA was isolated from glioma cell lines after treatment with 5aza2dC and NPTX2 transcript level was quantified by RT-qPCR. For each sample, fold change in gene expression is calculated over its mean expression in untreated sample. D, G418-resistant colonies were selected after transfection of glioma cell lines with vector or pCMV-NPTX2. Mean colony counts are displayed as percentage of vector control. E, the active caspase-3 levels were measured using FITC-DEVD-FMK in U343/Vector and U343/NPTX2 cells and shown as mean intensity. F, cell proliferation was measured as number of cells for U343/Vector and U343/NPTX2 clones and plotted. The difference in proliferation at different time points was found to be significant (P < 0.01). G, U343/Vector and U343/NPTX2 clones were subjected to soft agar colony formation and the number of colonies were counted and shown keeping U343/Vector as 100%. The difference was found to be significant (P < 0.028). H, U343/Vector and U343/NPTX2 clones were treated with varying concentrations of indicated drugs (adriamycin: 0.1, 0.2, 0.4, and 0.8 mg/mL; Taxol: 2, 4, 8, and 16 mmol/L; temozolomide: 125, 250, 500, and 1,000 mmol/L). At 48 hours after drug addition, the proportion of live cells was quantified by MTT assay. The proportion of viable cells from untreated cells was considered as 100%. Percent IC50 is shown. I, U251 cells were transfected with either control siRNA or NPTX2siRNA. At 48 hours posttransfection, the cells were either untreated or treated with 5aza2dC (5 mmol/L). After 24 hours of 5aza2dC addition, the cells were treated with various amounts of adriamycin, 0.1, 0.2, 0.4, and 0.8 mg/mL or temozolomide, 125, 250, 500, and 1,000 mmol/L. After 48 hours of the drug treatment, proportion of live cells was quantified by MTT assay. The absorbance of control cells was considered as 100%. The proportion of viable cells for indicated concentrations are shown.

have identiﬁed a 9-gene methylation signature using Cox patients, which may explain the short survival of high-risk proportional hazards model that can predict survival. Further- patients. Inhibition of NF-kB pathway made glioma cell lines more, the methylation signature was validated in independent with activated NF-kB pathway alone sensitive to chemother- cohorts. Multivariate analysis using TCGA data set with all apy. NPTX2, a gene with risky methylation, was validated for known prognostic markers, gene signatures, and microRNA its methylation and growth inhibitory functions and was found signature identiﬁed methylation signature as an independent to inhibit NF-kB activity through its ability to induce PTEN in GBM prognostic indicator. Network analysis revealed an asso- a p53-dependent manner. Thus, the NPTX2-p53-PTEN path- ciation between activated NF-kB pathway and high-risk way connected the methylation signature to NF-kB pathway.

OF8 Cancer Res; 73(22) November 15, 2013 Cancer Research

GBM Prognostic DNA Methylation and NPTX2-NF-kB Link

A Vector NPTX2 BEVector NPTX2 293 150 NS NS 150 NS NS NS Vector NPTX2 12 h 24 h 12 h 24 h U343/Vector U343/NPTX2 100 100 Adriamycin Myc/ NPTX2 50 50 P53

0 activity (%) Luc. 0 B Luc. activity (%) NF- k B Luc. p21 LN18 U343 T98G U251 U87 Vector CAPI3K CAAKT CAIKK 293 PTEN C Vector NPTX2 NS pAKT 200 150 500 300 AKT 150 400 100 200 300 100 Actin 200 100 50 50 100 FGVector NPTX2 0 0 0 0 120 80 PG13 Luc. activity (%) PG13 Luc. U343 U87 U251 293 100 Vector 60 80 NPTX2 D Vector NPTX2 60 40

NS (pmol) 200 250 150 400 2 40

200 PIP 20 150 300 20 100 Fold enrichment 150 0 100 200 0 100 a-IgG a-p53 a-IgG a-PTEN a-IgG a-PTEN 50 50 50 100 293 U343

0 0 0 0 PTEN Luc. activity (%) PTEN Luc. U343 U87 U251 293

Figure 5. NPTX2 inhibits NF-kB pathway through a p53-PTEN–dependent pathway. A, NPTX2 vector or control vector (2 mg) was transfected along with NF-kB luciferase reporter (1 mg) and, after 36 hours, luciferase activity was measured and plotted. B, NPTX2 vector or control vector (2 mg) was transfected along with NF-kB luciferase reporter (1 mg) and constitutively active form of either PI3K, AKT, or IKK (1 mg) and, after 36 hours, luciferase activity was measured and plotted. C, NPTX2 vector or control vector (2 mg) was transfected along with pG13 luciferase reporter (1 mg) and, after 36 hours, luciferase activity was measured and plotted. For 293 cells, 0.5, 1, and 2 mg of NPTX2 vector or control vector was used. D, NPTX2 vector or control vector (2 mg) was transfected along with PTEN promoter luciferase reporter (1 mg) and, after 36 hours, luciferase activity was measured and plotted. For 293 cells, 0.5, 1, and 2 mg of NPTX2 expression vector or control vector was used. For all experiments earlier, the luciferase activity of control vector was considered as 100%. E, cell extracts from 293 cells transfected with either vector or NPTX2 overexpression construct (10 mg) or U343/vector and U343/NPTX2 stables were subjected to Western blot analysis for NPTX2 (Myc tag), p53, p21, PTEN, pAKT, AKT, and actin proteins. F, The chromatin isolated from 293 cells transfected with vector or NPTX2 construct (20 mg) was used for immunoprecipitation with control or p53 antibody and used for PCR speciﬁc p53-binding region of PTEN. G, cell extracts from 293 or U343 cells transfected with vector or NPTX2 construct (20 mg) were used for immunoprecipitation using control or PTEN antibody. The immunoprecipitates were used to conduct PTEN activity ELISA. For all experiments, test was carried out to test the signiﬁcance of the observed differences between 2 conditions. , P < 0.05; , P < 0.01; , P < 0.001.

The strength of our methylation signature is that multivar- study identified specificactivationofNF-kB pathway in high- iate analysis with all known prognostic markers and signatures risk group from TCGA dataset and was further validated in identified it as an independent predictor of GBM survival. REMBRANDT and Phillips datasets. Although the association Although low-risk group showed enrichment of patients with between NF-kB pathway activation and GBM aggressiveness IDH1 mutation, proneural gene expression subtype and G- and chemoresistance as well as NF-kB as a therapeutic target CIMP positivity (8), methylation signature alone was identified is known (24, 28), our study clearly shows that NF-kB pathway as an independent survival predictor. Furthermore, compar- is activated particularly in high-risk group. Our data also isons with the 9 gene (22) and 10-microRNA panels (14) also show that NF-kB inhibition could sensitize only those glioma identified our methylation signature as an independent sur- cells having activated NF-kB pathway to chemotherapy. vival predictor. Thus, it seems that our 9-gene methylation Furthermore, we found IL-6 and ICAM1, which are upregu- signature is novel, robust, specific, and the only independent lated only in high-risk tumors, as the key mediators of predictor of survival. chemosensitivity. In good correlation, IL-6 and ICAM1 were To provide biological insight for survival prediction by found associated with invasion, angiogenesis, tumor growth, methylation signature, with the use of network analysis, our chemoresistance, and prognosis in many cancers (29–32).

www.aacrjournals.org Cancer Res; 73(22) November 15, 2013 OF9

Shukla et al.

This raises the potential use of targeting NF-kB pathway in tion signature as an independent GBM prognostic signature. treatment protocol specifically for those patients belonging It can be used to stratify the GBM patients into low-risk and to high-risk group. high-risk groups, which would facilitate individualized ther- Additional analysis of methylation signature and survival apeutic modality. Mechanistic investigation identified acti- identified NPTX2 as a very important gene as its methylation is vated NF-kB signaling as the potential cause of poor prog- found to be risky with a higher methylation in high-risk groups. nosis in high-risk patients, suggesting NF-kBasatherapeu- NPTX2 has been shown to be methylation silenced in pancre- tic target in particular for patients who do not respond to atic cancer, shown to induce BAX, possibly involving p53 current treatment protocols. activation, and shown to downregulate cyclin D1, leading to cell-cycle arrest and apoptosis (33). Our study shows that Disclosure of Potential Conflicts of Interest NPTX2 is methylation silenced in GBMs, could be reexpressed No potential conflicts of interest were disclosed. by methylation inhibitors and overexpression inhibits cell proliferation, anchorage-independent growth, and reexpres- Authors' Contributions Conception and design: S. Shukla, A. Arivazhagan, K. Somasundaram sion sensitizes glioma cells to chemotherapy. High levels of Development of methodology: S. Shukla, B. Mondal, A. Arivazhagan, NPTX2 mRNA and protein have been reported in neuronal K. Somasundaram cells and in subpopulation of glial cells in normal brain (34). Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.S. Hegde, B.A. Chandramouli, A. Arivazhagan, This may explain the reduced methylation and the consequent K. Somasundaram expression of NPTX2 in low-risk group, which is enriched for Analysis and interpretation of data (e.g., statistical analysis, biostatis- proneural gene expression type. Our work also shows that tics, computational analysis): S. Shukla, I.R. Pia Patric, S. Thinagararajan, k S. Srinivasan, V. Santosh, K. Somasundaram NPTX2 overexpression inhibits NF- B activity, suggesting that Writing, review, and/or revision of the manuscript: S. Srinivasan, its functions may involve activating or inhibiting signaling V. Santosh, A. Arivazhagan, K. Somasundaram k Administrative, technical, or material support (i.e., reporting or orga- pathways, which are modulators of NF- B pathway. p53 is nizing data, constructing databases): I.R. Pia Patric, A.S. Hegde, B.A. Chan- known to activate PTEN, leading to inhibition of PI3K-AKT dramouli, V. Santosh, A. Arivazhagan, K. Somasundaram pathway (35, 36). In our study, we show that NPTX2 activates Study supervision: A.S. Hegde, B.A. Chandramouli, K. Somasundaram PTEN at the level of transcription in a p53-dependent manner, leading to PI3 kinase inhibition. Furthermore, we also show Acknowledgments The results published here are in whole or part based upon data generated that NPTX2, through its ability to activate PTEN, inhibits by TCGA pilot project established by the NCI and NHGRI. Information about AKT, which has been shown to be a key upstream activator TCGA and the investigators and institutions who constitute the TCGA of NF-kB pathway (37, 38). Although NPTX2 expression research network can be found at http://cancergenome.nih.gov/. The use of datasets from Institute NC, 2005, and Phillips et al. (2006) are acknowl- induces p53 protein level without a change in its transcript edged. S. Shukla thanks CSIR for the fellowship. The authors thank Professors levels and may involve ATM, the exact mechanism at present M.R.S. Rao and P. Kondaiah for valuable suggestions. The authors thank Dr. remains to be investigated. These results suggest that NPTX2 H. Noushmehr for help in acquiring TCGA data. The authors also thank B. Thota,S.Bhargava,K.Chandrashekar,K.Prem,andB.C.Shailajaforthehelp inhibition of NF-kB may involve p53-PTEN activation, leading rendered. K. Somasundaram is a J.C. Bose Fellow of the Department of to PI3K-AKT-IKKa inhibition. Furthermore, it explains why Science and Technology. NPTX2 is methylation silenced in high-risk group wherein NF-kB pathway is activated. Grant Support Thus, our study uncovers a novel pathway (through This study was supported by a grant from DBT, government of India. The costs of publication of this article were defrayed in part by the NPTX2) of a complex interaction network between differ- payment of page charges. This article must therefore be hereby marked ential methylation of 9 genes and alteration of global gene advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this regulation, leading to modulation of signaling, in particular fact. k NF- B pathway, between low- risk and high-risk groups. Received January 31, 2013; revised July 17, 2013; accepted August 27, 2013; More importantly, this work identifies the 9-gene methyla- published OnlineFirst September 27, 2013.

References 1. Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ, 5. Reddy SP, Britto R, Vinnakota K, Aparna H, Sreepathi HK, Thota B, Janzer RC, et al. Effects of radiotherapy with concomitant et al. Novel glioblastoma markers with diagnostic and prognostic value and adjuvant temozolomide versus radiotherapy alone on sur- identified through transcriptome analysis. Clin Cancer Res 2008; vival in glioblastoma in a randomised phase III study: 5-year 14:2978–87. analysis of the EORTC-NCIC trial. Lancet Oncol 2009;10: 6. Santosh V, Arivazhagan A, Sreekanthreddy P, Srinivasan H, Thota B, 459–66. Srividya MR, et al. Grade-specific expression of insulin-like growth 2. Furnari FB, Fenton T, Bachoo RM, Mukasa A, Stommel JM, Stegh A, factor-binding proteins-2, -3, and -5 in astrocytomas: IGFBP-3 et al. Malignant astrocytic glioma: genetics, biology, and paths to emerges as a strong predictor of survival in patients with newly treatment. Genes Dev 2007;21:2683–710. diagnosed glioblastoma. Cancer Epidemiol Biomarkers Prev 2010;19: 3. Mizoguchi M, Betensky RA, Batchelor TT, Bernay DC, Louis DN, Nutt 1399–408. CL. Activation of STAT3, MAPK, and AKT in malignant astrocytic 7. Sreekanthreddy P, Srinivasan H, Kumar DM, Nijaguna MB, Sridevi gliomas: correlation with EGFR status, tumor grade, and survival. S, Vrinda M, et al. Identification of potential serum bio- J Neuropathol Exp Neurol 2006;65:1181–8. markers of glioblastoma: serum osteopontin levels correlate with 4. Nagarajan RP, Costello JF. Epigenetic mechanisms in glioblastoma poor prognosis. Cancer Epidemiol Biomarkers Prev 2010;19: multiforme. Semin Cancer Biol 2009;19:188–97. 1409–22.

OF10 Cancer Res; 73(22) November 15, 2013 Cancer Research

GBM Prognostic DNA Methylation and NPTX2-NF-kB Link

8. Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara K, 23. van den Bent MJ, Gravendeel LA, Gorlia T, Kros JM, Lapre L, Wessel- Berman BP, et al. Identification of a CpG island methylator phenotype ing P, et al. A hypermethylated phenotype is a better predictor of that defines a distinct subgroup of glioma. Cancer Cell 2010;17: survival than MGMT methylation in anaplastic oligodendroglial brain 510–22. tumors: a report from EORTC study 26951. Clin Cancer Res 2011; 9. Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, et al. 17:1–8. An integrated genomic analysis of human glioblastoma multiforme. 24. Nogueira L, Ruiz-Ontanon P, Vazquez-Barquero A, Moris F, Fernan- Science 2008;321:1807–12. dez-Luna JL. The NF-kB pathway: a therapeutic target in glioblastoma. 10. Freije WA, Castro-Vargas FE, Fang Z, Horvath S, Cloughesy T, Liau Oncotarget 2011;2:646–53. LM, et al. Gene expression profiling of gliomas strongly predicts 25. Institute NC. REMBRANDT home page. Available at . Accessed 2011 February 17, 2005. 11. Godard S, Getz G, Delorenzi M, Farmer P, Kobayashi H, Desbaillets I, 26. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, et al. Classification of human astrocytic gliomas on the basis of gene et al. Molecular subclasses of high-grade glioma predict prognosis, expression: a correlated group of genes with angiogenic activity delineate a pattern of disease progression, and resemble stages in emerges as a strong predictor of subtypes. Cancer Res 2003;63: neurogenesis. Cancer Cell 2006;9:157–73. 6613–25. 27. Abraham RT. Cell cycle checkpoint signaling through the ATM and 12. Nutt CL, Mani DR, Betensky RA, Tamayo P, Cairncross JG, Ladd C, ATR kinases. Genes Dev 2001;15:2177–96. et al. Gene expression-based classification of malignant gliomas 28. Zanotto-Filho A, Braganhol E, Schroder R, de Souza LH, Dalmolin RJ, correlates better with survival than histological classification. Cancer Pasquali MA, et al. NF-kB inhibitors induce cell death in glioblastomas. Res 2003;63:1602–7. Biochem Pharmacol 2010;81:412–24. 13. Rao SA, Santosh V, Somasundaram K. Genome-wide expression 29. Becker C, Fantini MC, Wirtz S, Nikolaev A, Lehr HA, Galle PR, et al. IL-6 profiling identifies deregulated miRNAs in malignant astrocytoma. signaling promotes tumor growth in colorectal cancer. Cell Cycle Mod Pathol 2010;23:1404–17. 2005;4:217–20. 14. Srinivasan S, Patric IR, Somasundaram K. A ten-microRNA 30. Liu Q, Li G, Li R, Shen J, He Q, Deng L, et al. IL-6 promotion of expression signature predicts survival in glioblastoma. PLoS One glioblastoma cell invasion and angiogenesis in U251 and T98G cell 2011;6:e17438. lines. J Neurooncol 2010;100:165–76. 15. Reif K NC, Thomas G, Hall A, Cantrell DA. Phosphatidylinositol 3- 31. Roland CL, Harken AH, Sarr MG, Barnett CC Jr. ICAM-1 expression kinase signals activate a selective subset of Rac/Rho-dependent determines malignant potential of cancer. Surgery 2007;141:705–7. effector pathways. Curr Biol 1996;6:1445–55. 32. Schroder C, Witzel I, Muller V, Krenkel S, Wirtz RM, Janicke F, et al. 16. Suvasini R SK. Essential role of PI3-kinase pathway in p53-mediated Prognostic value of intercellular adhesion molecule (ICAM)-1 expres- transcription: implications in cancer chemotherapy. Oncogene 2010; sion in breast cancer. J Cancer Res Clin Oncol 2011;137:1193–201. 29:3605–18. 33. Zhang L, Gao J, Li L, Li Z, Du Y, Gong Y. The neuronal pentraxin II gene 17. Das S, El-Deiry WS, Somasundaram K. Regulation of the p53 homolog (NPTX2) inhibit proliferation and invasion of pancreatic cancer cells in p73 by adenoviral oncogene E1A. J Biol Chem 2003;278:18313–20. vitro. Mol Biol Rep 2010;38:4903–11. 18. Pahl H. Activators and target genes of Rel/NF-kB transcription factors. 34. Moran LB, Hickey L, Michael GJ, Derkacs M, Christian LM, Kalaitzakis Oncogene 1999;18:6853–66. ME, et al. Neuronal pentraxin II is highly upregulated in Parkinson's 19. von Mering C, Jensen LJ, Snel B, Hooper SD, Krupp M, Foglierini M, disease and a novel component of Lewy bodies. Acta Neuropathol et al. STRING: known and predicted protein-protein associations, 2008;115:471–8. integrated and transferred across organisms. Nucleic Acids Res 2005; 35. Stambolic V MD, Sas D, Lin Y, Snow B, Jang Y, Benchimol S, Mak TW. 33:D433–7. Regulation of PTEN transcription by p53. Mol Cell Biol 2001;153: 20. Hooper SD, Bork P. Medusa: a simple tool for interaction graph 1369–80. analysis. Bioinformatics 2005;21:4432–3. 36. Song MS, Salmena L, Pandolfi PP. The functions and regulation of the 21. Castro MA, Filho JL, Dalmolin RJ, Sinigaglia M, Moreira JC, Mombach PTEN tumour suppressor. Nat Rev Mol Cell Biol 2012;13:283–96. JC, et al. ViaComplex: software for landscape analysis of gene expres- 37. Ozes ON, Mayo LD, Gustin JA, Pfeffer SR, Pfeffer LM, Donner DB. NF- sion networks in genomic context. Bioinformatics 2009;25:1468–9. kB activation by tumour necrosis factor requires the Akt serine-thre- 22. Colman H, Zhang L, Sulman EP, McDonald JM, Shooshtari NL, Rivera onine kinase. Nature 1999;401:82–5. A, et al. A multigene predictor of outcome in glioblastoma. Neuro Oncol 38. Romashkova JA, Makarov SS. NF-kB is a target of AKT in anti- 2009;12:49–57. apoptotic PDGF signalling. Nature 1999;401:86–90.

www.aacrjournals.org Cancer Res; 73(22) November 15, 2013 OF11

A DNA Methylation Prognostic Signature of Glioblastoma: Identification of NPTX2-PTEN-NF- κB Nexus

Sudhanshu Shukla, Irene Rosita Pia Patric, Sivaarumugam Thinagararjan, et al.

Cancer Res Published OnlineFirst September 27, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-13-0298

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2013/10/16/0008-5472.CAN-13-0298.DC1

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2013/11/07/0008-5472.CAN-13-0298. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.