a comparative and integrative approach identi atpase family, … · encoded cyclins (ccna2, ccnb1,...

Human Cancer Biology

A Comparative and Integrative Approach Identifies ATPaseFamily, AAA Domain Containing 2 as a Likely Driver of CellProliferation in Lung Adenocarcinoma

Robert Fouret1, Julien Laffaire2, Paul Hofman6, Mich�ele Beau-Faller7, Julien Mazieres8, Pierre Validire3,Philippe Girard3, Sophie Camilleri-Br€oet4, Fabien Vaylet9, Francois Leroy-Ladurie10, Jean-Charles Soria11,13,and Pierre Fouret5,12

AbstractPurpose: To identify genetic changes that could drive cancer pathogenesis in never and ever smokerswith

lung adenocarcinoma.

Experimental Design: We analyzed the copy number and gene expression profiles of lung adenocarci-

nomas in 165 patients and related the alterations to smoking status. Having found differences in the tumor

profiles, we integrated copy number and gene expression data from 80 paired samples.

Results: Amplifications at 8q24.12 overlapping MYC and ATAD2 were more frequent in ever smokers.

Unsupervised analysis of gene expression revealed two groups: in the group withmainly never smokers, the

tumors expressed genes common to normal lung; in the group with more ever smokers, the tumors

expressed "proliferative" and "invasive" gene clusters. Integration of copy number and gene expression data

identified one module enriched in mitotic genes and MYC targets. Its main associated modulator was

ATAD2, a cofactor ofMYC. A strong dose–response relationship between ATAD2 and proliferation-related

gene expression was noted in both never and ever smokers, which was verified in two independent cohorts.

Both ATAD2 andMYC expression correlated with 8q24.12 amplification and were higher in ever smokers.

However, only ATAD2, and notMYC, overexpression explained the behavior of proliferation-related genes

and predicted a worse prognosis independently of disease stage in a large validation cohort.

Conclusions: The likely driving force behindMYC contribution to uncontrolled cell proliferation in lung

adenocarcinoma is ATAD2. Deregulation of ATAD2 is mainly related to gene amplification and is more

frequent in ever smokers. Clin Cancer Res; 18(20); 5606–16. �2012 AACR.

IntroductionThe majority of lung cancers are caused by tobacco

smoking.However, even in peoplewhohave never smoked,lung cancer would rank as the seventhmost common causeof cancer deathworldwide (1). In 2000, lung cancer innever

smokers accounted in France for 17% cancer deaths inwomen and 4% in men (2).

For 4 major genes involved in the pathogenesis of lungcancers, ALK, EGFR, KRAS, and TP53, striking differences inthe molecular alterations of these genes have been found inlung cancers in never and ever smokers (3, 4). Molecularalterations include translocations for ALK or point muta-tions for EGFR, KRAS, and TP53 (5, 6). In addition, copynumber changes contribute through associated gene dereg-ulation to the malignant phenotype. For instance, MYC isfrequently amplified and overexpressed in lung cancer (7).No study has reported definitive associations betweenamplifications or deletions and smoking status (8).

We analyzed the copy number and gene expressionprofiles of lung adenocarcinomas and related the alterationsto smoking status. Having found differences in the tumorprofiles, we integrated copy number and gene expressiondata to identify genetic changes that could drive cancerpathogenesis. The present study differed from previousstudies on 2 aspects. First, the number of tumors fromneversmokerswas greater in our study than in previous studies (8,9). Second, to control for potential bias, the ever smokergroup was constructed by matching ever smokers to never

Authors' Affiliations: 1DCom, T�el�ecom ParisTech; 2Programme Carted'Identit�e des Tumeurs, Ligue Nationale Contre le Cancer; 3Institut Mutua-liste Montsouris; 4Hopital Europ�een George Pompidou; 5Universit�e Pierreet Marie Curie, Paris; 6CHU Nice, Nice; 7CHU Strasbourg, Strasbourg;8CHU Toulouse, Toulouse; 9Hopital d'instruction des arm�ees Percy, Cla-mart; 10Centre Chirurgical Marie-Lannelongue, Le Plessis-Robinson;11Institut Gustave-Roussy; 12INSERM G�en�etique des tumeurs, Villejuif,and 13Université Paris XI, Le Kremlin-Bicêtre, France

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

R. Fouret and J. Laffaire are first co-authors.

Corresponding Author: Pierre Fouret, INSERM G�en�etique des tumeursU985, Institut Gustave-Roussy, 114 rue E. Vaillant, 94805 Villejuif Cedex,France. Phone: 33-0-1-42-177782; Fax: 33-0-1-42-177777; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-12-0505

�2012 American Association for Cancer Research.

ClinicalCancer

Research

Clin Cancer Res; 18(20) October 15, 20125606

on September 6, 2021. © 2012 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst August 22, 2012; DOI: 10.1158/1078-0432.CCR-12-0505

http://clincancerres.aacrjournals.org/

smokers, such that the whole cohort was enriched in neversmokers and the group of ever smokers had clinical char-acteristics (sex, disease stage) identical to never smokers.

Materials and MethodsDetailed information on patients, samples, andmethods

used in copy number, gene expression, and survival anal-yses is available as Supplementary Materials and Methods.

Patients and samplesAll 165 study patients were treated by surgery for lung

adenocarcinoma without prior chemotherapy. Fifty-eightpatients received cisplatin-based adjuvant chemotherapy.Never smoking status was defined by a lifetime exposure ofless than 100 cigarettes. The tumors were classified accord-ing to the tumor–node–metastasis (TNM) system in use atthe time of diagnosis (10). The pathologic diagnoses werereviewed according to current histologic classification forlung carcinoma (11, 12). Cases for which a doubt about theprimary site in the lung remained were excluded. All ade-nocarcinomas were invasive. A bronchiolo-alveolar compo-nent was recorded when a noninvasive lepidic growth wasseen adjacent to a component of invasive adenocarcinoma.This studywas part of the LungGenes (LG) project, which

was approved by the Institut National du Cancer reviewboard (Programme National d’Excellence Sp�ecialis�e Pou-mon). Informed consent was obtained frompatients for theuse of their lung surgical samples.Only cases with an average of tumor cells equal to or

above 50% were included. Genomic DNA and RNA wereextracted and assessed for integrity and quantity followingstringent quality control criteria (cit.ligue-cancer.net).

Genomic DNA analysisGenomic DNAs were hybridized on Illumina SNP

HumanCNV370 chips (Illumina). The GISTIC version2.0 algorithm (www.broadinstitute.org/cancer/pub/GIS-TIC2) was used to identify significant regions of amplifica-tion or deletion. The frequencies of aberrations contribut-ing to significant peak regionswere compared using c2 tests.

Gene expression analysisTotal RNAs were hybridized to Affymetrix Human

Genome U133 Plus 2.0 GeneChip (Affymetrix). Unsuper-vised hierarchical clustering analysis of tumor samples fromthe LG cohort and normal lung samples from 11 femaleAsian never smokers (accession number: GSE 19804) wasconducted on the most variant probe sets. Differencesbetween sample clusters were tested using the c2 test. Hyper-geometric enrichment for Gene Ontology sets (GeneOntol-ogy.org) and MYC target genes (www.myccancergene.org)were calculated with false discovery rate (FDR) correction ofP values. Literature Vector Analysis (LitVAn) was used toinfer gene cluster functionality with an evaluation of thesignificance of their scores (litvan.bio.columbia.edu).

Both genomic and gene expression datawere deposited inArrayExpress database (accession number: E-MTAB-923).

ATAD2 relative expression was measured by real-timereverse transcriptase PCR (RT-PCR) using the Hs00204205TaqMan probe (Applied Biosystems).

Integration of copy number and gene expression dataWe used Copy Number and Expression In Cancer (CON-

EXIC) to integratematched copy number (amplifications ordeletions) and gene expression data from 80 paired sample(13).

As described by Akavia and colleagues, CONEXIC isbased on the following assumptions: (i) a driver mutationin a "modulator" gene should be associated (correlated)with a group of genes that form a "module" and (ii) copynumber aberrations often influence the expression of genesin the module via changes in expression of the modulator.

The CONEXIC learning algorithm consists of 3 key steps:

1. Selection of candidate genes that are recurrentlyamplified or deleted in tumors.

2. Single Modulator step that creates an initialassociation between expression of candidate driversand expression of genes modules.

3. An iterative Network Learning step to improve theinitial model.

During the Single Modulator and the Network learningsteps, the search is driven by the optimization of a Bayesianscoring function similar to Module Network (14). For eachnode, thedriver–split combination that achieves thehighestscore is selected as long as it is verified to be statisticallysignificant. Significance is tested using Lee and colleaguespermutation test (15); up to 3 top-scoring modulator genesare tried, and if none of them pass the permutationssignificance test, no more splits are added to the driver tree.In addition to significance testing, nonparametric bootstrapserves to eliminate spurious correlations.

The output is a driver network that divides the expressedgenes into modules and associates each module with adriver tree. Each node in the tree is associated with a drivergene (modulator gene) and a threshold expression level(split value) and divides the expression values of the mod-ule’s members into samples in which the modulator’s

Translational RelevanceOur results suggest that the aberrant expression of

MYC targets that participate in the program responsiblefor uncontrolled proliferation may be attributed toATAD2-deregulated expression. This further suggeststhat ATAD2 levels may predict the MYC dependency oflung adenocarcinoma, which should be exploited fortherapeutic purposes. While MYC has been consideredas a frequent and very relevant therapeutic target in lungcancer, specific inhibition ofMYChas not been achievedand no MYC inhibitor is currently in the clinic. ATAD2is worthwhile to investigate as a therapeutic target,which appears feasible given its ATPase activity and itsbromodomain.

ATAD2 Drives Cell Proliferation in Lung Adenocarcinoma

www.aacrjournals.org Clin Cancer Res; 18(20) October 15, 2012 5607




expression is below the threshold and those in which themodulator’s expression is above the threshold. Each side ofthe split at the first root of the tree (herein designated as thefirst-order split) can contain further splits (secondary splits)using other modulator/expression threshold pairs.

A detailed description of the selection of candidate genesand of the specified parameters for the Single Modulatorand theNetwork Learning steps aswell as a discussion of thesignificance levels under which modulators were identifiedare available as Supplementary Information (Supplemen-tary Materials and Methods).

The modules and their modulators were visualized usingGenatomy (www.c2b2.columbia.edu/danapeerlab/html/genatomy.html).

To nominate themodule associated with smoking status,we used gene set enrichment analysis (GSEA; ref. 16).

For validation,weused thepublicly available gene expres-sion data from 68 lung adenocarcinomas (accession num-ber: GSE 12667; ref. 17) and from 391 lung adenocarcino-mas (caarraydb.nci.nih.gov, pId ¼ 1015945236141280;ref. 18). The linear relationship between a modulator andits associated genes was measured using the Pearson corre-lation coefficient.

Survival analysisThe univariate overall survival analyses were conducted

using the Kaplan–Meier method and log-rank tests. In the

multivariate proportional hazard Cox overall survival anal-ysis, ATAD2 expression was studied together with age, sex,and disease stage.

ResultsDifferent frequency of aberrations in significant peakregions between never and ever smokers

A total of 121 high-quality genomic profiles wereobtained. Frequent aberrations (frequency >25%) includedgains on 1p, 1q, 5p, 5q, 6p, 7p, 7q, 8q, 14q, 16p, 17q, 20p,and 20q and losses on 1p, 3p, 4q, 5q, 6q, 8p, 9p, 10q, 12p,13q, 15q, 17p, and 18q. The GISTIC 2.0 algorithm wasapplied to identify regions that were significantly amplifiedor deleted (Fig. 1). A total of 59 significant peak regionswitha frequency of 13% to 84% were identified, including 22regions thatwere amplifiedand37 regions thatweredeleted.

The frequency of amplifications or deletions in the 59significant regions was compared between never and eversmokers. After adjustment for multiple comparisons usingBonferroni method, only 2 regions were differentiallyamplified or deleted according to smoking status: amplifi-cations were more frequent in ever smokers (83%) com-pared with never smokers (52%) at 8q24.12 (q-value ¼0.02), whereas deletions were more frequent in ever smo-kers (50%) compared with never smokers (13%) at 4q35.2(q-value ¼ 0.0006).

Figure 1. GISTIC 2.0 analysis ofcopy number changes in 121 lungadenocarcinomas. Plots of the G-scores (top) and q-values (bottom)with respect to amplifications (A) ordeletions (B) over the entire regionanalyzed. The significance level fortheq-value is indicatedbya verticaldotted line. Chromosomepositions are indicated along the y-axis with centromere positionsindicated by horizontal dottedlines. The locations of the peakregions are indicated on the right ofeach panel.

Fouret et al.

Clin Cancer Res; 18(20) October 15, 2012 Clinical Cancer Research5608




Two groups of tumors with distinctive geneexpression clusters and different clinicopathologicannotationsUnsupervised hierarchical clustering of gene expression

in 103 high-quality tumor profiles and 11 normal lungsamples (GSE19804) revealed 2 groups of tumors (Fig. 2).The partition was stable as assessed by resampling. The firstgroup of tumors was characterized by the expression of agene cluster (cluster c) that was common to normal lungsamples and mainly absent from tumors in the secondgroup. Two genes clusters (cluster f and cluster i) wereoverexpressed in the second group of tumors, whereas theywere both expressed at low levels in most tumors of the firstgroup and in normal lung samples.The cluster f was enriched for GeneOntology (GO) terms

"cell cycle process" (q-value ¼ 1.7E-17) and "mitotic cellcycle" (q-value ¼ 1.8E-11) and designated as the "prolifer-

ative" cluster. Typical genes in the proliferative clusterencoded cyclins (CCNA2, CCNB1, and CCNE2), thecyclin-dependent kinase CDK6, E2F transcription factors(E2F7, E2F8), and 12 proteins involved in mitosis. Thecluster i was enriched for the GO term "extra-cellularmatrix" (7.0E-14). Typical genes in cluster i encoded mem-bers of the disintegrin and metalloproteinase (ADAM12,ADAMDEC1, ADAMTS5) or matrix metalloproteinase(MMP1, MMP3, MMP11, MMP12, MMP13) families.

Using LitVAn, significant terms associated with the pro-liferative cluster were "cyclin" and "mitotic" as well as"spindle," reflecting the enrichment for genes participatingto the mitotic spindle (BUB1, CENPF, KIF14, KIF15,NDC80, NEK2, NUF2, SKA1, SPC25, TPX2, TTK; genome.ucsc.edu). LitVAn significant terms for cluster i included"invasion," favoring its designation as the "invasive" genecluster (Supplementary Table S1).

Figure 2. Unsupervised analysis of gene expression in 103 lung adenocarcinomas from the LGcohort and11normal lungs fromAsian female never smokers. Inthe heatmap, each cell represents the expression value for a probe in a sample. The largest expression values are in red, and the lowest expression values arein green. The sample clusters shown at the top are colored in red or blue for tumor samples and in green for normal lung samples, wherein theblue-colored tumor samples and the green-colored normal lung samples are clustered together. Each box below the sample clusters represents the valuefor a discrete clinicopathologic annotation in a sample. A black box denotes presence of a bronchiolo-alveolar component, ever-smoker status, male sex,EGFR mutation, or KRAS mutation. The P values associated with annotations are obtained by comparing the 2 groups of tumor samples using c2 tests.The gene clusters shown on the left of the heat map are labeled A to J.






The first group of tumors whose gene expression resem-bled normal lung comprised 35 never smokers (74%) and12 ever smokers, whereas the second group comprised 28never smokers (50%) and28 ever smokers (P¼0.01). In thefirst group, the tumors more frequently presented with abronchiolo-alveolar component (P¼ 5E-6) or harbored anEGF receptor (EGFR) mutation (P¼ 0.0002), whereas in thesecond group, they more often harbored a KRAS mutation(P ¼ 0.02).

ATAD2 as a likely driver of cell proliferationEight-hundred and eighteen genes overlapped significant

aberrations less than a third of chromosome length, includ-ing 350 genes overlapping 19 amplifications and 468 genesoverlapping 34 deletions. Among these genes, 109 genesoverlapped the 2 regions that were differentially alteredbetween never and ever smokers. The expression of 175genes, including 35 that overlapped the 8q24.12 and4q35.2 smoking status–related aberrations (Table 1), wassignificantly altered (P < 0.05) by either their amplificationstatus or deletion status.

Using CONEXIC, we found a model comprising 67modules that were associated with 31 main modulators(i.e., likely drivers at the first-order split of the regulatoryprograms) explaining the behavior of 10,001 genes.

We wished to uncover the likely drivers of differentiallyexpressed clusters, whichwere revealed by the unsupervisedanalysis of gene expression in the whole cohort. We had 2collections of gene sets, one provided by the unsupervisedanalysis in 103patients (the a to j gene clusters), the other byCONEXIC in 78 patients with paired genomic and geneexpression data (the 67 modules). Both collections weredetermined without the help of clinical or biologic annota-tions.We thus conducted hypergeometric enrichment anal-ysis to identify which of the gene sets overlapped in the 2datasets and thus indicate likely drivers of the overlappingclusters in the whole cohort. The proliferative cluster (clus-ter f in the unsupervised analysis) intersected very signifi-cantly (q-value¼ 2.0E–37) with CONEXICmodule 62 (Fig.3A). Twenty-eight of 46 genes of cluster f were identified asmodule 62 genes. The proliferative cluster did not intersectwith any other CONEXIC modules. Module 62 genes wereenriched in the GO term "cell cycle process" (q-value ¼1.2E-87). The main modulator associated with module 62wasATPase family, AAA domain containing 2 (ATAD2), a genelocated at 8q24.12.

A linear relationship between the expression of ATAD2and the expression of proliferation-related genes inboth never and ever smokers

To verify the association between ATAD2 expression andgenes in its module, the module 62 regulatory programsidentified in the LG cohort were applied to gene expressiondata from 68 lung adenocarcinomas of the Ding cohort(17). The profiles of module 62 genes were compared usingidentical split expression values for the modulators. Therelationship between high ATAD2 levels and overexpres-sion of module 62 genes was verified in the Ding cohort

(Supplementary Fig. S1A).WhenATAD2was low, however,the second-order regulatory programs depending onTUBB3 did not classify samples in the Ding cohort as wellas in the LG cohort, suggesting that this secondary regulatorwas not optimally chosen by CONEXIC. We replacedTUBB3 expression by ATAD2 expression, which improvedthe classification of samples in the Ding cohort (Supple-mentary Fig. S1B)without altering substantially the originalmodule 62 profiles in the LG cohort (Supplementary Fig.S1C). These results suggested that ATAD2 expression alonecould explain the behavior of module 62 genes acrossdatasets.

Table 1. Association of gene expression withsmoking status–related copy numberalterations in the LG cohort

Aberration Cytoband Gene name Welch t test (P)a

Amplification 8q24.12 ATAD2 1.1E-04Amplification 8q24.12 DERL1 1.5E-04Amplification 8q24.12 DSCC1 2.2E-05Amplification 8q24.12 FAM83A 5.4E-08Amplification 8q24.12 FAM91A1 1.5E-09Amplification 8q24.12 KIAA0196 1.3E-05Amplification 8q24.12 MRPL13 1.6E-05Amplification 8q24.12 MYC 8.0E-05Amplification 8q24.12 NDUFB9 2.3E-04Amplification 8q24.12 NSMCE2 0.002Amplification 8q24.12 RNF139 0.01Amplification 8q24.12 SQLE 1.9E-04Amplification 8q24.12 TATDN1 9.8E-05Amplification 8q24.12 TMEM65 8.1E-06Deletion 4q35.2 ACSL1 5.5E-06Deletion 4q35.2 ANKRD37 0.002Deletion 4q35.2 CCDC111 5.2E-04Deletion 4q35.2 CDKN2AIP 2.2E-05Deletion 4q35.2 CYP4V2 4.6E-04Deletion 4q35.2 DCTD 8.0E-09Deletion 4q35.2 F11 0.02Deletion 4q35.2 FRG1 4.0E-04Deletion 4q35.2 GALNT7 0.009Deletion 4q35.2 GPM6A 0.03Deletion 4q35.2 HMGB2 0.047Deletion 4q35.2 HPGD 3.6E-07Deletion 4q35.2 ING2 0.04Deletion 4q35.2 IRF2 1.1E-04Deletion 4q35.2 NEIL3 0.009Deletion 4q35.2 RWDD4A 7.5E-08Deletion 4q35.2 SNX25 1.7E-05Deletion 4q35.2 SORBS2 0.01Deletion 4q35.2 STOX2 0.04Deletion 4q35.2 TLR3 1.6E-06Deletion 4q35.2 UFSP2 3.5E-04

aComparing amplification versus normal or deletion versusnormal.

Fouret et al.





To enable the confirmation of a linear relationshipbetween ATAD2 and the proliferative cluster (cluster f),Pearson correlation coefficients were calculated in the LGcohort, the Ding cohort, and a third independent cohort of

391 patient (18). There was a strong dose–response rela-tionship between ATAD2, and the proliferative cluster inevery cohort as the expression of genes belonging to theproliferative cluster increased with higher ATAD2 levels in

Figure 3. Integrated analysis of paired copy number and gene expression data in 80 lung adenocarcinomas belonging to the LG cohort. A, CONEXICanalysis. Genatomy module network view of CONEXIC module 62. Each row of the heat map corresponds to the expression of a gene across the 80samples. Gene names indicated at the right of the heat map are sorted by alphabetical order. Sample names are indicated above the heat map. Samplesare ordered according to the regulatory programs found by CONEXIC and shown above sample names. The modulators include ATAD2, TUBB3,SLC25A21, and KCNMB4, wherein ATAD2 increased expression at the first-order split and, at the right, second-order split is associated with increasedexpression of genes in the module. Yellow dotted lines partition the samples according to split values of the modulators. B to D, analysis of geneexpression of the proliferative cluster (cluster f) in the LG cohort according to ATAD2 expression and according to smoking status, 8q24.12amplification, or MYC expression. Proliferative cluster genes are those included in the final set of genes after processing of LG gene expressiondata during the CONEXIC procedure. Gene names indicated at the right of the heat map are sorted by alphabetical order. B, samples are sorted in ever(blue) or never smokers (white) and then sorted within each smoking status category into 4 groups of increasing ATAD2 expression levels (fromwhite to red). The 4 ATAD2 groups are sorted using the split values found by CONEXIC in the analysis of the LG cohort. C, samples are sortedaccording to 8q24.12 amplification, comparing amplification (blue) versus no amplification (white), then within each category into 4 groups of increasingATAD2 expression levels as above. D, samples are sorted according to increasingMYC expression levels using quartile values as split values (from lightto dark blue), then within each category into 4 groups of increasing ATAD2 expression levels as above.






the LG (correlation coefficient 0.85), Ding (correlationcoefficient 0.77), and Shedden (correlation coefficient0.75) cohorts.

In the LG cohort, the expression of the proliferativecluster increased with higher ATAD2 levels in never (cor-relation coefficient 0.83) and in ever smokers (correlationcoefficient 0.87; Fig. 3B). A similarly strong linear relation-ship was noted in both the Ding and the Shedden cohortsfor never and ever smokers and patients with unknownsmoking status (Table 2).

In the LG cohort, the expression of ATAD2 stronglycorrelated with the expression of the proliferative cluster

in every subgroup defined by sex, disease stage, bronchiolo-alveolar component, EGFR, or KRAS status (Table 2 andSupplementary Fig. S2). Although high ATAD2 was lessfrequent among tumors without 8q24.12 amplification,ATAD2 was differentially expressed and strongly correlatedwith expression of the proliferative cluster in tumorswith orwithout the amplification (Fig. 3C and Table 2).

ATAD2 and module 62 relationships with 8q14.12amplification and smoking status

On the basis of the array data, ATAD2 expression wasassociated with 8q24.12 amplification status (P ¼ 1.1E-

Table 2. Correlation ofATAD2orMYCexpressionwith the expression of genesof the proliferative cluster inthe LG, Ding, and Shedden cohorts

ATAD2 MYC

Cohort variable

Pearsoncorrelationcoefficient P

Pearsoncorrelationcoefficient P

LGAll 0.85 <0.0001 0.39 0.0003Smoking statusNever smoker 0.83 <0.0001 0.28 0.06Ever smoker 0.87 <0.0001 0.36 0.04

SexFemale 0.86 <0.0001 0.37 0.002Male 0.80 0.003 0.45 0.16

Disease stageEarly (I or II) 0.87 <0.0001 0.33 0.01Late (III) 0.79 <0.0001 0.58 0.004

Bronchiolo-alveolar componentYes 0.77 <0.0001 0.28 0.09No 0.86 <0.0001 0.40 0.008

EGFRMutated 0.77 <0.0001 0.49 0.001Wild-type 0.88 <0.0001 0.26 0.11

KRASMutated 0.89 <0.0001 0.67 0.009Wild-type 0.85 <0.0001 0.26 0.04

8q24.12 amplificationYes 0.82 <0.0001 0.37 0.006No 0.86 <0.0001 0.02 0.92

DingAll 0.77 <0.0001 0.28 0.02Smoking statusNever smoker 0.7 0.05 0.1 0.81Ever smoker 0.78 <0.0001 0.25 0.12Unknown 0.67 0.001 0.29 0.21

SheddenAll 0.75 <0.0001 0.28 <0.0001Smoking statusNever smoker 0.7 <0.0001 0.1 0.54Ever smoker 0.76 <0.0001 0.26 <0.0001Unknown 0.72 <0.0001 0.2 0.06

Fouret et al.





4; Table 1) and was higher in ever smokers (P ¼ 0.0004). Itwas neither associated with CDKN2A nor RB1 overlappingdeletions.In a subset of 76 patients with available RNA for real-time

RT-PCR analysis, ATAD2 expression was increased in eversmokers comparedwithnever smokers (fold change¼1.75,P ¼ 0.02) and in patients with 8q24.12 amplificationcompared with no amplification (fold change ¼ 2.27, P¼ 0.001).As the number of ever smokers was higher in the group of

tumors expressing the proliferative cluster f, we tested theassociation of smoking status with each CONEXIC moduleusing GSEA. The profiles in ever smokers as compared withnever smokers were significantly enriched in themodule 62gene set to which the highest enrichment score was given(0.86 for the enrichment score, 1.67 for the normalizedenrichment score, 0.02 for the nominal P, and 0.22 for theFDR).

RelationshipsbetweenATAD2,MYC, andproliferation-related genesModule 62 genes were enriched (q-value ¼ 3.2E-4) for

genes of the MYC target database (www.myccancergene.org). Three CONEXICmodules other than module 62 wereassociated with ATAD2 as their main modulator, one ofwhich was also enriched for MYC targets (q-value¼ 0.005).Enrichments for the proliferative cluster genes, GO termscontaining "cell cycle process" and MYC targets werealigned only for module 62 (Supplementary Table S2).Amplifications of 8q24.12 included both ATAD2 and

MYC in every sample save one. Like ATAD2, MYC expres-sion was associated with 8q24.12 amplification (P ¼ 8.0E-5; Table 1) and was higher in ever smokers than in neversmokers (P ¼ 0.002).The correlation of the proliferative cluster with MYC

(correlation coefficient 0.39) was less strong, however, thanwith ATAD2 (correlation coefficient 0.85). Remarkably, thecorrelation of MYC targets in the proliferative cluster wasless strong with MYC (correlation coefficient 0.33) thanwithATAD2 (correlation coefficient 0.83). Likewise,mitotic

spindle genes correlated weakly with MYC, whereas theycorrelated strongly with ATAD2 (Supplementary Table S3).Overexpression of the proliferative cluster occurred intumors with low MYC and it correlated with ATAD2(Fig. 3D). Themodest correlation of the proliferative clusterwith MYC was verified in both Ding and Shedden cohorts(Table 2).

Survival of patientsIn the 78 patients from the LG cohort with paired copy

number and gene expression array data, the survival rateswere 0.75 [95% confidence interval (CI), 0.65-0.88] at 3years and 0.61 (95% CI, 0.48–0.77) at 4 years for the lowATAD2 group and 0.59 at 3 years (95% CI, 0.43–0.83) and0.44 (95% CI, 0.28–0.71) at 4 years for the high ATAD2group (Fig. 4A). Survival was not significantly differentaccording to ATAD2 expression (P ¼ 0.14). Late diseasestagewas associatedwith a shorter survival (P¼0.01).Noneof the other clinical or biologic variables, including 8q24.12amplification (P ¼ 0.58) and MYC expression (P ¼ 0.77),was associated with survival.

In the 75patientswhowere studiedusing PCR tomeasureATAD2 expression, neither the PCR data (P¼ 0.41) nor theATAD2 array data (P ¼ 0.43) were associated with survival.

In the 349 patients from the Shedden cohort, the survivalrates were 0.74 (95% CI, 0.68–0.81) at 3 years and 0.65(95%CI, 0.58–0.73) at 4 years for the lowATAD2 group and0.58 (95%CI, 0.51–0.66) at 3 years and 0.5 (95%CI, 0.43–0.58) at 4 years for the highATAD2 group (Fig. 4B). Survivaltime was longer in the low ATAD2 group than in the highATAD2 group (P ¼ 0.002). Late disease stage was stronglyassociated with shorter survival (P ¼ 1E-16).

Multivariate proportional hazardCoxmodels were testedto investigate the association of ATAD2 array data withsurvival and to adjust for age, sex, and disease stage in the78 patients from the LG cohort and in the 349 patients fromthe Shedden cohort. In the LG cohort, the best model(likelihood P ¼ 0.008) included ATAD2, age, and diseasestage. An older age (HR, 2.18; 95%CI, 1.04–4.55; P¼ 0.04)and late disease stage (HR, 2.32; 95%CI, 1.17-4.6;P¼ 0.02)

Figure 4. Kaplan–Meier curves ofoverall survival rates according toATAD2 levels in the LG cohort,comparing high versus low ATAD2levels. Shown is the log-rankP value.A, LG cohort. B, Shedden cohort.

A B

ATAD2 low

ATAD2 low

ATAD2 high

ATAD2 high

P value 0.14 P value 0.002

Time (d) Time (mo)

Surv

ival ra

te

1.0

0.8

0.6

0.4

0.2

0.0

0 1,000 2,000 3,000 4,000 0 10 20 30 40 50 60

1.0

0.8

0.6

0.4

0.2

0.0

Surv

ival ra

te






were associatedwith a shorter survival.HighATAD2wasnotsignificantly associated with survival (HR, 2.04; 95% CI,0.98-4.27; P¼ 0.06). Removing ATAD2 reduced slightly themodel likelihood (likelihood P ¼ 0.02).

In the Shedden cohort, the bestmodel (likelihoodP¼1E-13) included ATAD2 and stage. High ATAD2 (HR, 1.68;95% CI, 1.22-2.32; P ¼ 0.002) and late disease stage (HR,3.86; 95% CI, 2.74-5.42; P ¼ 8E-15) were significantlyassociated with survival. Removing ATAD2 reduced slightlythe model likelihood (likelihood P ¼ 2E-12).

DiscussionOur study reveals that ATAD2 is a likely driver of cell

proliferation in lung adenocarcinoma. ATAD2 overexpres-sion both explains the behavior of cell cycle genes and,mostlikely, results primarily from amplification, thereby con-necting proliferation of lung cancer cells to a unique geneticaberration. Furthermore, our results suggest that the onco-geneMYC, which is located 4.3Mb distal to ATAD2 (http://genome.ucsc.edu/), is involved in that pathway as theATAD2 associated proliferative signature includes MYCtargets involved in cell cycle. Before our study, it was knownthat ATAD2 is upstreamofMYCand that it can exert a role inthe proliferation of normal and cancer cells, strongly sup-porting our conclusion (19–21). The present study is thefirst to provide evidence suggesting that amplified ATAD2 isthe main driving force behindMYC contribution to uncon-trolled cell proliferation in lung adenocarcinoma. The cru-cial driving function shown here for ATAD2 may havetherapeutic implications. While MYC has been consideredas a frequent and very relevant therapeutic target in lungcancer, specific inhibition of MYC has not been achievedand no MYC inhibitor is currently in the clinic. ATAD2 isworthwhile to investigate as a therapeutic target, whichappears feasible given its ATPase activity and its bromodo-main (22). Moreover, ATAD2 expression predicts theexpression of mitotic spindle genes, whose products par-ticipate to a network vulnerable to inhibition of SUMOyla-tion (23–25).

A key factor in uncovering the contrasted phenotypes thatare summarized by gene clusters in the unsupervised anal-ysis of gene expression is the comparison of normal lungand tumors, many of which were from never smokers. Lungadenocarcinomas in never smokers present typically with abronchiolo-alveolar component with well-differentiatedtumor cells, whereas in ever smokers growth is usually"fully invasive," that is, consists exclusively of invasivecomponent (11, 26). Consistent with histology, geneexpression in the group where never smokers were numer-ous resembled that of normal lung. In contrast, the groupwith more ever smokers expressed proliferative or invasivegene clusters.

Most of the frequent aberrations shown in this cohorthave been previously reported (8). However, the frequen-cies of aberrations in 2 significant regions of amplificationor deletion differ according to smoking status. These resultsraised the question whether the deregulation of genes over-

lapping differentially altered regions explain the differencesin expression profiles between tumors.

To identify driving mutations and the processes theyinfluence, we integrated copy number and gene expressiondata using the recently developed algorithm CONEXIC(13). With this approach, the starting list of candidatedrivers includes only the genes within or near-significantregions of copy number changes. As a result, CONEXICwould not detect drivers that are typically associated withpoint mutations.

We identify ATAD2 as a likely driver whose expressionexplains the behavior of differentially expressed prolifer-ation-related genes. Indeed, an ATAD2-associated moduleoutputted by CONEXIC contained a majority of genes ofthe proliferative cluster identified in the unsupervisedanalysis of gene expression, an enrichment very unlikelycaused by chance. A strong dose–response relationshipbetween ATAD2 levels and those of genes belonging tothe proliferative cluster is shown in the LG cohort and isverified in 2 independent validation cohort (17, 18). Therelationship between ATAD2 and proliferation-relatedgenes is neither affected by smoking status nor smokingstatus–associated characteristics including KRAS or EGFRmutation.

ATAD2 is correctly associated by CONEXIC with genesthat it is known to regulate. ATAD2 has been identified as acofactor for MYC-dependent transcription by Cir�o andcolleague (19). Here, ATAD2-associated genes were signif-icantly enriched in MYC targets. Kalashnikova and collea-gues using chromatin immunoprecipitation assays showedthat ATAD2 occupies the proximal promoter regions ofseveral key cell-cycle regulators (BUB1, CCNA2, KIF15,MCM10, and TOP2A), which we show linearly related toATAD2 expression in lung adenocarcinoma (27).

ATAD2 overlaps the 8q24.12 region which was morefrequently amplified in ever smokers. ATAD2 expressionwas associated with 8q24.12 amplification, suggesting thatATAD2 deregulation occurs primarily through copy num-ber changes. Nevertheless, it wasATAD2 expression and not8q24.12 amplification that correlatedwith the expressionofproliferation-related genes. Discrepancies between 8q24.12amplification andATAD2overexpression point to addition-al genetic or epigenetic events contributing to ATAD2expression. It has been suggested ATAD2 deregulation maybe the consequence of the loss of retinoblastoma (RB)-mediated control in a subset of highly aggressive breastcancer (27). We checked that ATAD2 expression was notassociated with deletions targeting the RB pathway.

MYC and ATAD2 are frequently co-amplified in cancers(www.broadinstitute.org/tumorscape), a consistent findingin this cohort. Co-amplification may be selected in tumorsas a way to concomitantly overexpress not too far apartcooperating genes. Like ATAD2,MYC was overexpressed inever smokers, andMYC overexpression was associated with8q24.12 amplification. Increased expression ofMYC targetsappears necessary to the association of ATAD2 with cellproliferation as other ATAD2-associated modules thatwere not enriched in MYC targets were not enriched in

Fouret et al.





proliferation-related genes. As compared with ATAD2,expression of MYC only weakly correlated, however, withexpression of the proliferative cluster, includingMYC targetgenes and mitotic spindle genes. These results suggest thatATAD2 levels through MYC activity are more importantthan MYC levels to drive cell proliferation in lung adeno-carcinoma. Assuming that MYC protein levels are roughlyequivalent tomRNA levels, itmay seem surprising thatMYCactivity is not directly related toMYC expression. However,using a novel in vivo model of Myc-induced tumorigenesis,Murphy and colleagues reported that low levels of deregu-lated Myc are competent to drive ectopic proliferation ofsomatic cells and lung oncogenesis (28).High ATAD2 is associated with poor survival of patients

with breast cancer (19, 27). Caron and colleagues reportedthat high ATAD2 (E. and C. Brambilla, unpublished data)predicts a shorter survival of patients with lung cancer (29).In the LG cohort, ATAD2 is not significantly associated withsurvival, although there is a trend when the array data areadjusted for disease stage andage. There ismuchuncertaintyin the results in small groups. In the larger Shedden cohort,high ATAD2 predicts a shorter survival, which is consistentwith the reported prognostic value of a cluster of cellproliferation–related gene (18). The prognostic value ofATAD2 is independent of disease stage in that cohort.ATAD2 is a co-activator, which can control MYC-depen-

dent transcription (19, 27). Through MYC and E2F tran-scription factors, ATAD2 increases the expression of prolif-eration-related and anti-apoptotic genes in many differenttypes of cancer, including hormone-dependent prostate orbreast carcinomas, estrogen receptor–negative breast carci-noma, cervical carcinoma, glioblastoma, osteosarcoma,and non–small cell lung carcinoma (19, 20, 21, 27, 29,30). Although these data strongly support that ATAD2maydrive cell proliferation in various cancers,more experimentsare needed to investigate the mechanisms by which ATAD2likely influences the biologic consequences of MYC dereg-ulation in the context of lung cancer cells.In summary, ATAD2 is identified by a comparative and

integrative approach as a likely driver of cell proliferation inlung adenocarcinoma. MYC is co-amplified with ATAD2and, like ATAD2, is overexpressed in ever smokers. How-ever, it is ATAD2 and not MYC expression that is stronglyrelated to the expression of proliferation-related genes,especially mitotic spindle genes. These results suggest thatthe aberrant expression of MYC targets that participate in

the program responsible for uncontrolled proliferationmaybe attributed toATAD2-deregulated expression. This furthersuggests that ATAD2 levels may predict aMYC dependencyof lung adenocarcinoma, which should be exploited as apriority target for therapeutic purposes.

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

Authors' ContributionsConception and design: J.-C. Soria, P. FouretDevelopment of methodology: R. Fouret, J. Laffaire, P. FouretAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): P. Hofman, M. Beau-Faller, J. Mazieres, P. Vali-dire, P. Girard, S. Camilleri-Br€oet, F. Vaylet, F. Leroy-Ladurie, P. FouretAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis):R. Fouret, J. Laffaire, P. Hofman,M. Beau-Faller, F. Vaylet, P. FouretWriting, review, and/or revision of the manuscript: J. Laffaire, P. Hof-man, J. Mazieres, P. Girard, F. Vaylet, J.-C. Soria, P. FouretAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): M. Beau-Faller, P. ValidireStudy supervision: P. Fouret

AcknowledgmentsThe following investigators participated in the Lung Genes (LG) project:

CentreChirurgicalMarie-Lannelongue, Le Plessis-Robinson: P.Dartevelle, E.Dulmet, F. Leroy-Ladurie, and V. de Montpreville; Centre Hospitalier Inter-communal Cr�eteil: I. Monnet; Centre Hospitalo-Universitaire Dijon, Paris:A. Bernard and F. Piard; CentreHospitalo-UniversitaireHotel-Dieu, Paris:M.Alifano, S. Camilleri-Bro€et, D. Damotte, and J.F. R�egnard; Centre Hospitalo-Universitaire Nice: P. Hofman, V. Hofman, and J. Mouroux; Centre Hospi-talo-Universitaire Saint-Louis, Paris: J. Tr�edaniel; Centre Hospitalo-Univer-sitaire Strasbourg: M. Beau-Faller, G. Massard, and A. Neuville; CentreHospitalo-Universitaire Tenon, Paris: M. Antoine and J. Cadranel; CentreHospitalo-Universitaire Toulouse: L. Brouchet, J. Mazi�eres, and I. Rouquette;DCom, T�el�ecomParisTech, Paris: R. Fouret; Hopital d’instruction des arm�eesPercy, Clamart: P. Saint-Blancard and F. Vaylet; Institut Gustave-Roussy,Villejuif: A. Berhneim, P. Dessen, F. Dufour, N. Dorvault, P. Fouret, B. Job, L.Lacroix, V. Lazar, C. Richon, V. Roux, P. Saulnier, J.C. Soria, E. Taranchon, S.Toujani, and A. Valent; Institut Mutualiste Montsouris, Paris: P. Girard, D.Gossot, and P. Validire; and Ligue Nationale Contre le Cancer: J. Laffaire andA. de Reyn�es. The authors thank D. Simon (Laboratoire Probabilit�es etMod�eles Al�eatoires, Universit�e Pierre et Marie Curie, Paris, France) for helpin conducting the bootstrap.

Grant SupportThe study was supported by Institut National du Cancer (Programme

National d’Excellence Sp�ecialis�e Poumon); LigueNationale Contre le Cancer(Programme Carte d’Identit�e des Tumeurs); and Association pour laRecherche sur leCancer (grant number SFI20101201740) grants to P. Fouret.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received February 15, 2012; revised August 1, 2012; accepted August 5,2012; published OnlineFirst August 22, 2012.

References1. Subramanian J,GovindanR. Lung cancer in never smokers: a review. J

Clin Oncol 2007;25:561–70.2. International Agency for Research on Cancer. Water, air, soil

and food pollutants. In: Attributable causes of cancer in Francein the year 2000. Geneva, Switzerland: WHO Press; 2007. p. 97–102.

3. Sun S, Schiller JH, Gazdar AF. Lung cancer in never smokers–adifferent disease. Nat Rev Cancer 2007;7:778–90.

4. Shaw AT, Yeap BY, Mino-Kenudson M, Digumarthy SR, Costa DB,Heist RS, et al. Clinical features and outcome of patients with non-

small-cell lung cancer who harbor EML4-ALK. J Clin Oncol 2009;27:4247–53.

5. Soda M, Choi YL, Enomoto M, Takada S, Yamashita Y, Ishikawa S,et al. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Nature 2007;448:561–6.

6. Herbst RS, Heymach JV, Lippman SM. Lung cancer. N Engl J Med2008;359:1367–80.

7. Little CD, Nau MM, Carney DN, Gazdar AF, Minna JD. Amplificationand expression of the c-myc oncogene in human lung cancer cell lines.Nature 1983;306:194–6.






8. Weir BA, Woo MS, Getz G, Perner S, Ding L, Beroukhim R, et al.Characterizing the cancer genome in lung adenocarcinoma. Nature2007;450:893–8.

9. Wong MP, Fung L-F, Wang E, Chow W-S, Chiu S-W, Lam W-K, et al.Chromosomal aberrations of primary lung adenocarcinomas in non-smokers. Cancer 2003;97:1263–70.

10. Hermanek P, Hutter R, Sobin L, Wagner G, Wittekind C, editors. TNMAtlas. Guide illustr�e de la classification TNM/pTNM des tumeursmalignes. 4th ed. Paris, France: Springer-Verlag; 1998.

11. Travis WD, Brambilla E, M€uller-Hermelink HK, Harris CC. InternationalAgency for Research onCancer. Pathology and genetics of tumours ofthe lung, pleura, thymus and heart. Lyon, France: IARC; 2004.

12. Travis WD, Brambilla E, Noguchi M, Nicholson AG, Geisinger KR,Yatabe Y, et al. International association for the study of lung cancer/American thoracic society/European respiratory society internationalmultidisciplinary classification of lung adenocarcinoma. J ThoracOncol 2011;6:244–85.

13. Akavia UD, Litvin O, Kim J, Sanchez-Garcia F, Kotliar D, Causton HC,et al. An integrated approach to uncover drivers of cancer. Cell2010;143:1005–17.

14. Segal E, Shapira M, Regev A, Pe'er D, Botstein D, Koller D, et al.Module networks: identifying regulatory modules and their condition-specific regulators from gene expression data. Nat Genet 2003;34:166–76.

15. Lee S-I, Pe'er D, Dudley AM, Church GM, Koller D. Identifyingregulatory mechanisms using individual variation reveals key rolefor chromatin modification. Proc Natl Acad Sci U S A 2006;103:14062–7.

16. SubramanianA, TamayoP,Mootha VK,Mukherjee S, Ebert BL,GilletteMA, et al. Gene set enrichment analysis: a knowledge-based approachfor interpreting genome-wide expression profiles. Proc Natl Acad SciU S A 2005;102:15545–50.

17. Ding L, Getz G, Wheeler DA, Mardis ER, McLellan MD, Cibulskis K,et al. Somatic mutations affect key pathways in lung adenocarcinoma.Nature 2008;455:1069–75.

18. Shedden K, Taylor JMG, Enkemann SA, Tsao M-S, Yeatman TJ,Gerald WL, et al. Gene expression-based survival prediction in lungadenocarcinoma: a multi-site, blinded validation study. Nat Med2008;14:822–7.

19. Cir�o M, Prosperini E, Quarto M, Grazini U, Walfridsson J, McBlane F,et al. ATAD2 is a novel cofactor for MYC, overexpressed and amplifiedin aggressive tumors. Cancer Res 2009;69:8491–8.

20. Zou JX, Revenko AS, Li LB, Gemo AT, Chen H-W. ANCCA, anestrogen-regulated AAAþATPase coactivator for ERalpha, is requiredfor coregulator occupancyandchromatinmodification. ProcNatl AcadSci U S A 2007;104:18067–72.

21. Zou JX, Guo L, Revenko AS, Tepper CG, Gemo AT, Kung H-J, et al.Androgen-induced coactivator ANCCA mediates specific androgenreceptor signaling in prostate cancer. Cancer Res 2009;69:3339–46.

22. Hsia EY, Goodson ML, Zou JX, Privalsky ML, Chen H-W. Nuclearreceptor coregulators as anewparadigm for therapeutic targeting. AdvDrug Deliv Rev 2010;62:1227–37.

23. Joukov V, Groen AC, Prokhorova T, Gerson R, White E, RodriguezA, Walter JC, Livingston DM. The BRCA1/BARD1 heterodimermodulates ran-dependent mitotic spindle assembly. Cell 2006;127:539–52.

24. KiyomitsuT,ObuseC,YanagidaM.HumanBlinkin/AF15q14 is requiredfor chromosome alignment and the mitotic checkpoint through directinteraction with Bub1 and BubR1. Dev Cell 2007;13:663–76.

25. Kessler JD, Kahle KT, Sun T, Meerbrey KL, Schlabach MR, SchmittEM, et al. A SUMOylation-dependent transcriptional subprogram isrequired for Myc-driven tumorigenesis [Internet]. Science 2012;335:348–53.

26. Miller VA, Kris MG, Shah N, Patel J, Azzoli C, Gomez J, et al. Bronch-ioloalveolar pathologic subtype and smoking history predict sensitivityto gefitinib in advanced non-small-cell lung cancer. J Clin Oncol2004;22:1103–9.

27. Kalashnikova EV, Revenko AS, Gemo AT, Andrews NP, Tepper CG,Zou JX, et al. ANCCA/ATAD2 overexpression identifies breast cancerpatients with poor prognosis, acting to drive proliferation and survivalof triple-negative cells through control of B-Myb and EZH2. CancerRes 2010;70:9402–2.

28. MurphyDJ, JunttilaMR, Pouyet L, Karnezis A, ShchorsK, Bui DA, et al.Distinct thresholds govern Myc's biological output in vivo. Cancer Cell2008;14:447–57.

29. Caron C, Lestrat C, Marsal S, Escoffier E, Curtet S, Virolle V, et al.Functional characterization of ATAD2 as a new cancer/testis factorand a predictor of poor prognosis in breast and lung cancers. Onco-gene 2010;29:5171–81.

30. Revenko AS, Kalashnikova EV, Gemo AT, Zou JX, Chen H-W. Chro-matin loading of E2F-MLL complex by cancer-associated coregulatorANCCA via reading a specific histone mark. Mol Cell Biol 2010;30:5260–72.

Fouret et al.





2012;18:5606-5616. Published OnlineFirst August 22, 2012.Clin Cancer Res Robert Fouret, Julien Laffaire, Paul Hofman, et al. Lung Adenocarcinoma

as a Likely Driver of Cell Proliferation inAAA Domain Containing 2ATPase Family,A Comparative and Integrative Approach Identifies

Updated version

10.1158/1078-0432.CCR-12-0505doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2012/08/23/1078-0432.CCR-12-0505.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/18/20/5606.full#ref-list-1

This article cites 27 articles, 11 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/18/20/5606.full#related-urls

This article has been cited by 7 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

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

Permissions

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

.http://clincancerres.aacrjournals.org/content/18/20/5606To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-12-0505

http://clincancerres.aacrjournals.org/content/suppl/2012/08/23/1078-0432.CCR-12-0505.DC1

http://clincancerres.aacrjournals.org/content/18/20/5606.full#ref-list-1

http://clincancerres.aacrjournals.org/content/18/20/5606.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/18/20/5606


a comparative and integrative approach identi atpase family, … · encoded cyclins (ccna2, ccnb1,...

Documents