nrf2 pathway activation and adjuvant chemotherapy benefit in

Personalized Medicine and Imaging

NRF2 Pathway Activation and AdjuvantChemotherapy Benefit in Lung Squamous CellCarcinomaDavidW. Cescon1,2, Desmond She3, Shingo Sakashita3,4, Chang-Qi Zhu3, Melania Pintilie5,Frances A. Shepherd1,2, and Ming-Sound Tsao3,4

Abstract

Purpose:Genomic profiling of lung squamous cell carcinomas(SCC) has identified NRF2 pathway alterations, which activateoxidative response pathways, in one third of tumors. Preclinicaldata suggest these tumors may be resistant to platinum-basedchemotherapy. We evaluated the clinical relevance of these find-ings and assessed whether NRF2 activation predicts benefit fromadjuvant chemotherapy in SCC.

Experimental Design: Logistic regression (LR) and significanceanalysis of microarrays (SAM) were applied to all 104 TCGA (TheCancer Genome Atlas) SCC cases that had microarray gene expres-sion and mutation data to identify genes associated with somaticNRF2 pathway alterations. The resulting signature (NRF2ACT) wastested in 3 independent SCC datasets to evaluate its prognostic andpredictive effects. IHC and sequencing for NRF2 and KEAP1 wereevaluated in one cohort (n¼ 43) to assess the relationship betweengene expression, mutational status, and protein expression.

Results: Twenty-eight genes were identified by overlapbetween LR (291 genes) and SAM (30 genes), and theseconsistently separated SCC into 2 groups in all datasets,corresponding to putatively NRF pathway–activated andwild-type (WT) tumors. NRF2ACT was not prognostic. Howev-er, improved survival with adjuvant chemotherapy in theJBR.10-randomized trial appears limited to patients with theWT signature (HR 0.32, P ¼ 0.16; NRF2ACT HR 2.28, P ¼ 0.48;interaction P ¼ 0.15). NRF2ACT was highly correlated withmutations in NRF2 and KEAP1, and with high NRF2 proteinexpression.

Conclusions: A gene expression signature of NRF2 pathwayactivation is associated with benefit from adjuvant cisplatin/vinorelbine in SCC. Patients with NRF2 pathway–activatingsomatic alterations may have reduced benefit from this therapy.Clin Cancer Res; 21(11); 2499–505. �2015 AACR.

IntroductionAdjuvant platinum-based chemotherapy is a standard of care

for patientswith completely resected stage II to IIIAnon–small celllung cancer (NSCLC), with an absolute 5-year survival benefit of4% to 15% in several randomized trials andmeta-analyses (1–7).However, as with most adjuvant therapies, only a subgroup ofpatients derives benefit from this intervention. Because chemo-therapy causes both short-term and long-term toxicities, predic-tive biomarkers that could identify which patients do or do notbenefit would be of great clinical utility.

Although previous efforts have evaluated potential predictivebiomarkers of adjuvant chemotherapy benefit, most tested either

single markers (8) or considered unstratified populations con-sisting of squamous cell carcinoma (SCC) and adenocarcinomaand other histologic subtypes within the spectrum of NSCLC (9,10). Following the completion of initial large-scale efforts tocharacterize the genomic and molecular alterations in NSCLC byThe Cancer Genome Atlas (TCGA) consortium and others, theexistence of major subsets of lung cancers with shared pathwayalterations has been recognized (11). This includes approximately35% of SCC where somatic alterations resulting in activation oftheNRF2pathway viamutations or amplificationof [nuclear factor(erythroid-derived 2)-like 2] NFE2L2/NRF2 ormutation or deletionof its negative regulators Kelch like-ECH-associated protein 1(KEAP1) or Cullin 3 (CUL3) have been identified (11).

The NRF2 transcription factor is a master regulator of theantioxidant response, and dysregulation of this pathway occurscommonly in cancer (12). Several lines of preclinical and clinicalinvestigation have suggested that NRF2 pathway activation con-fers resistance to chemotherapy (13–22). Furthermore, the avail-able data indicate that mutations in this pathway define a majormolecular subset of SCC. However, the critical question ofwhether this subset of patients derives differential benefit fromadjuvant chemotherapy has not been addressed. Such an analysisrequires interrogation of data from the pivotal randomized clin-ical trials of adjuvant chemotherapy, where gene expression butnot somatic mutational data are available. Although previousstudies have used groups of NRF2 target genes as a readout ofNRF2 activity in NSCLC (19), and some existing gene expressionsignatures, including the classical expression subtype of SCC,

1Departments of Medical Oncology and Hematology, Princess Mar-garet Cancer Centre, University Health Network, Toronto, Canada.2Department of Medicine, University of Toronto, Toronto, Canada.3Department of Pathology, Princess Margaret Cancer Centre, Univer-sity Health Network, Toronto, Canada. 4Department of LaboratoryMedicine and Pathobiology, University of Toronto, Toronto, Canada.5Department of Biostatistics, Princess Margaret Cancer Centre, Uni-versity Health Network, Toronto, Canada.

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

Corresponding Author: Ming-Sound Tsao, University Health Network, 200Elizabeth Street, Toronto, Ontario M5G 2C4, Canada. Phone: 416-340-4737;Fax: 416-340-5517; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-14-2206

�2015 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org 2499

on March 28, 2018. © 2015 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst March 4, 2015; DOI: 10.1158/1078-0432.CCR-14-2206

http://clincancerres.aacrjournals.org/

partially overlap with NRF2 pathway mutational status (11), noexisting signature has been specifically developed a priori toidentity NRF2 pathway–activated SCC. We therefore sought todefine a gene expression signature for these cancers, and to test thissignature as a predictor of benefit from adjuvant chemotherapy inNCIC Clinical Trials Group JBR.10 (4).

Materials and MethodsPatients and datasets

This study used patient data from four independent sources:TCGA (11); a previously published case series of resected SCCfrom Raponi and colleagues (GSE4573; ref. 23); a case series ofresected SCC from the University Health Network (UHN181;GSE50081) (24); and a subset of SCC participants in the JBR.10clinical trial (GSE14814; ref. 9). For each contributing dataset,analysis was restricted to patients with SCC for whommicroarraygene expression data were available. For analysis of the TCGAdataset, only cases with microarray gene expression and exomesequencing information available as of August 2012 were con-sidered. Demographic and clinical information for SCC patientsin each dataset is summarized in Table 1.

Development of NRF2-activated (NRF2ACT) gene expressionsignature from TCGA

TCGA SCC cases with both microarray and exome sequencingdata (as of August 2012) were used to define an NRF2 geneexpression signature by comparing cases with or without somaticalterations in NRF2 pathway genes (mutations/amplification inNFE2L2, mutations/deletions in KEAP1 or CUL3; as defined andidentified by the TCGA analysis; ref. 11). Logistic regression (LR;using P < 0.001) and significance analysis of microarrays (SAM;ref. 25) were applied independently to identify genes associatedwith somatic alterations in the NRF2 pathway. The intersection ofthese gene lists was used to define the NRF2 signature (NRF2ACT).

The NRF2ACT gene list was used to perform clustering analysisin the other datasets: NRF2-activated (NRF2ACT-high) versusnormal/wild-type (NRF2ACT-low) subgroups were defined bytheir separation at the highest level of the hierarchical tree.

Mutational analysis of UHN samplesThemutational hotspot containing exon 1 ofNRF2 and KEAP1

coding regions were sequenced for UHN cases by Sanger sequenc-ing following touchdown PCR. Amplification and sequencingprimers are included in Supplementary Table S1. Nonsynon-ymous mutations, insertions, and deletions were identified bycomparison to the reference genome (GrCh37). Mutation callingwas performed blinded to gene expression classification andclinical outcomes.

Immunohistochemical analysis of NRF2 and KEAP1Whole sections of formalin-fixed paraffin-embedded (FFPE)

samples from UHN 181 cases were stained for NRF2 (SCBTclone H-300, 1:50 dilution) and KEAP1 (Proteintech 60027,1:300 dilution) using a Benchmark XT autostainer (VentanaMedical Systems). Nuclear staining intensity (0 to 3þ) and thepercentage of cells stained were used to calculate H-scores (inten-sity score�% stained) for each sample. For each antibody, cutoffvalues were defined to categorize samples: For NRF2, sampleswere classified as high (�80) or low (<80), and for KEAP1 as high(>130), intermediate (60–130), or low (<60), based on thedistributions of scores (Supplementary Fig. S1). Interpretationwas performed blinded to other variables.

NRF2 fluorescence in situ hybridizationFluorescence in situ hybridization (FISH) studies were per-

formed with the use of dual-color DNA FISH probes. Two bac-terial artificial chromosome (BAC) clones (RP11-28M17 andRP11-317C5) that hybridize to the 2q31.2 region of chromosome2 containing the NRF2 gene and two clones (RP11-343F11 andRP11-127K18) located close to the centromeric region of chro-mosome 2 (2q11.2) were selected from the Human UCSCGenome Browser assembly (February 2009 CRch37/hg19) andobtained from TCAG Genome Resource Facility (Toronto,Canada). The BAC clones were directly labeled with SpectrumOrange (BAC clones corresponding to the NRF2 gene) or Spec-trum Green (CEP2 BAC clones) fluorochromes using a commer-cially available nick translation kit according to the manufac-turer's protocol (Abbott Laboratories). Probes were first verifiedon normal bloodmetaphases to confirm their correct allocations.Briefly, tissuemicroarray and FFPE sectionswere deparaffinized inthree changes of xylenes, dehydrated in ethanol, pretreated incitrate buffer (pH 6.8) for 50minutes at 80�C, followed by pepsindigestion for 18minutes at 37�C. Slideswere hybridized for 2days

Table 1. Demographic and clinical characteristics of SCC patients from threecohorts included in analyses

DatasetJBR.10(n ¼ 52)

Raponi(n ¼ 129)

UHN181(n ¼ 43)

Age, y (median, range) 64.2 (45.4–75.8) 68 (42–91) 69.7 (49.5–87.9)Sex (n, %)Female 6 (11.5) 47 (36.4) 18 (41.9)Male 46 (88.5) 82 (63.6) 25 (58.1)

Stage (n, %)1A 27 (20.9) 8 (18.6)1B 25 (48.1) 46 (35.7) 19 (44.2)2A 4 (7.7) 6 (4.7) 2 (4.7)2B 23 (44.2) 27 (20.9) 14 (32.6)3A 17 (13.2)

Median follow-up, yyear (range)

6.8 (2.9–9.3) 2.5 (0.2–12.0) 5.9 (1.4–12.0)

Died (n, %) 19 (36.5) 68 (52.7) 17 (39.5)

Translational Relevance

Adjuvant platinum-based chemotherapy improves survivalin patients with non–small cell lung cancer, but benefits only aminority of those treated. Integrating emerging knowledge ofmolecular alterations and their cellular consequences mayenable the identification of patients most likely to benefitfrom such treatment. Because the NRF2 pathway has beenshown to alter chemosensitivity in vitro, we felt it would beimportant to examine the impact of recently described acti-vating alterations on the benefit from adjuvant chemotherapyin patients with squamous cell lung carcinoma (SCC). To doso, we identified a set of genes whose expression was associ-ated with NRF2 pathway activation. Using this classifier, SCCpatients with the activated signature treated in the JBR.10clinical trial did not appear to derive benefit from chemother-apy. This signaturemay enablemore personalized approachesto the treatment of lung SCC, sparing chemotherapy toxicityand identifying patients for the evaluation of alternativetherapeutic strategies.

Cescon et al.

Clin Cancer Res; 21(11) June 1, 2015 Clinical Cancer Research2500




at 37�C, washed and counterstained with DAPI. Tissues and cellswere examined and scored on an Imager M1 Zeiss microscope(Carl ZeissCanada Limited) equippedwith the appropriatefilters.The JAI CV-M4þCL progressive scan monochrome camera (JAIInc.) and the MetaSystems Isis FISH Imaging software programsv5.3 (MetaSystems) were used for capturing images. The numberof red and green signals as well as their distribution was analyzedby an experienced cancer cytogenetic technician (blinded to out-comes and gene expression classification) in 50 nonoverlappingtumor cell nuclei. Amplification was defined as an NRF2:CEP2ratio of � 2.0.

Statistical analysisOverall survival (OS) calculated from the date of surgery

(Raponi, UHN) or randomization (JBR.10) to death was theprimary outcome endpoint. In JBR.10, where cause of death wasknown, disease-specific survival was used and non–lung cancerdeaths were censored at the time they occurred (9). The survivalestimates were calculated using the Kaplan–Meier method. TheCox proportional hazards model was used to test the prognosticeffect of the NRF2ACT signature as well as its predictive effect bytesting its interactionwith the treatment in the JBR.10 dataset. Theassociations between categorical variables (NRF2, KEAP1, CUL3mutations, NRF2, KEAP1 IHC with NRF2ACT signature) weretested using the Fisher exact test.

The set of genes associated with the NRF2 pathway alterationswere selected based on LR and SAM (25). The alpha level forselection of the LRwas 0.001. SAManalysiswas performed using atwo-class unpaired methodology to determine differential geneexpression, with a set false-discovery rate of 0.05 and a call for 100resamples to permit bootstrapping.Upon generation of a gene set,a fold-change cutoff of 2.3 was applied to reduce the number ofpotential probes found to be upregulated in NRF2 or KEAP1mutants. Statistical analyses were performed using the open-source software R version 2.12.1 and the publicly available samrpackage.

ResultsNRF2ACT signature in TCGA SCC

A total of 104 unique SCC cases with completemicroarray geneexpression, mutation, and copy number data were identified inTCGA from a total of 178 samples. Forty-one cases were found tocontain somatic alterations inNRF2 (mutation or amplification),or in KEAP1 or CUL3 (mutation or deletion), and the remaining63 cases were wild-type for these genes. LR identified 291 geneswith significant differential expression (P < 0.001) betweensomatically altered and wild-type cases (Supplementary TableS2a). Thirty genes were found to be upregulated in NRF2-alteredcases by SAM analysis (Supplementary Table S2b). A comparisonof these gene lists identified 28 overlapping genes, which defineour NRF2ACT signature (Supplementary Table S2c). As predicted,reclustering of the TCGA cases using the 28-gene set confirmedthat NRF2ACT identifies two major gene expression subgroups,with most NRF2-activating mutations segregating into theNRF2ACT subgroup (Fig. 1).

NRF2ACT as a marker of NRF2 pathway activationTo test the biologic relevance of genes contained in the 28-gene

NRF2ACT set, we compared this list with a previously publishedNRF2 target gene set definedby transcriptional analysis andCHIP-

seq in NRF2 and KEAP1 knockout mouse embryonic fibroblasts(26). This analysis revealed a highly significant enrichment (one-sided P¼ 5.96� 10–11; hypergeometric test), confirming that the28-gene set contains bona fide NRF2 targets.

Concordance between NRF2ACT status, NRF2/KEAP1 IHC andsomatic alterations in an independent dataset

To validate the association between the NRF2ACT signature andsomatic alterations in the NRF2 pathway in SCC, we performedhierarchical clustering using the 28-gene set, as well as NRF2 andKEAP1mutational analysis by Sanger sequencing, in 43 SCC fromtheUHN181dataset (24). As observed in the TCGAderivation set,NRF2ACT identified twomajor subgroups in theUHN181 SCC set,with similar proportions to TCGA (Fig. 2). Furthermore, theNRF2ACT-high subgroup was highly enriched for tumors withsomatic alterations in NRF2 (P ¼ 0.0001). In contrast, in theadenocarcinoma subset of UHN181 (n ¼ 130), NRF2ACT did notseparate tumors into discrete groups, indicating that this signatureis specifically relevant to SCC (Supplementary Fig. S2).

To evaluate whether IHC analysis of NRF2 or KEAP1 mightserve as a useful surrogate of NRF2 pathway activation/mutation-al status, semiquantitative IHC was performed on whole tissuesections. Representative images are shown in Fig. 3, and IHCscores are indicated for each case in Fig. 2A. Tumors with highNRF2 expression by IHC were enriched in the NRF2ACT-highsubgroup (P ¼ 0.0003). Furthermore, high NRF2 protein expres-sion was associated with the presence of either NRF2 or KEAP1alterations (P < 0.001). No association between KEAP1 proteinexpression and NRF2ACT or mutational status was observed.

Prognostic effect of NRF2ACT in SCC patients treated withsurgery alone

No prognostic association of NRF2ACT status was present inthe UHN181 dataset of 43 patients treated with surgery alone,without adjuvant chemotherapy (HR, 0.86, P ¼ 0.79; Fig. 4A).

Figure 1.Reclustering of TCGAcases usingNRF2-associated 28-gene set identifies twomajor expression subgroups. The NRF2ACT subgroup (right) contains themajority of cases with NRF2-activating somatic alterations, identified by thecolor-coded bar (top).

NRF2 Pathway and Chemotherapy in Lung Squamous Cell Carcinoma

www.aacrjournals.org Clin Cancer Res; 21(11) June 1, 2015 2501




Similarly, in a second SCC surgery-alone dataset (Raponi;n ¼ 129), NRF2ACT identified two subgroups based on geneexpression (Fig. 2B), but was not significantly associated with OS(HR, 1.43, P ¼ 0.2; Fig. 4B). A prognostic association also wasabsent in the observation (no adjuvant chemotherapy) arm ofJBR.10 (HR, 0.66, P ¼ 0.61; Fig. 4C).

Predictive effect of NRF2ACT in SCC patients treated withadjuvant chemotherapy

The predictive effect of NRF2ACT was examined in the SCCsubset of JBR.10. Patients with NRF2ACT-high tumors did notappear to benefit from adjuvant chemotherapy (HR, 2.28; 95%CI, 0.24–22; P ¼ 0.48), whereas a trend toward chemotherapybenefit was observed in NRF2ACT-low patients (HR 0.32; 95% CI,0.065–1.6; P ¼ 0.16; interaction P ¼ 0.15; Fig. 5).

DiscussionConsiderable collaborative effort to define the recurring somat-

ic alterations in human lung cancers has yielded remarkableinsights into the molecular basis of this disease. This wealth ofknowledge now enables us to delve far beyond the histologicclassifications that have, until very recently, defined clinicalapproaches to lung cancer. The ultimate goal of these efforts is

to identify actionable alterations that can be used to identify newtherapeutic targets or refine treatment approaches for individualpatients, thereby providing "personalized" or "precision" oncol-ogy and improving patient outcomes while reducing treatment-associated toxicities and costs.

The lung squamous cell carcinoma sequencing analysis hasexpanded the number of recognized putative "driver" oncogenesin this disease (11). These include a substantial number ofrecurrent but uncommon mutations in kinases, growth factorreceptors, and related genes that might be targetable with specificsmall-molecule inhibitors or monoclonal antibodies, several ofwhich currently are in clinical development (27, 28). Although itis clear that highly specific inhibitors can have dramatic efficacywhen matched to tumor genotype, as in the case of EGFR-acti-vating mutations, ALK translocations, and ROS1 rearrangementsinNSCLC (29–33), the potential impact of somatic alterations onchemotherapy efficacy, which remains the standard of care foradjuvant treatment and the mainstay of therapy for SCC, is ofadditional clinical consequence. In an effort to apply the findingsof TCGA (and the work that preceded it; refs. 13, 34) to currentclinical management, we have focused on the NRF2 pathway,which is somatically activated in over one third of SCC (11) andhas awell-characterized role in chemotherapy sensitivity based onpreclinical studies.

Figure 2.A, hierarchical clustering of UHN cases using NRF2 28-gene set identifies two major subgroups. NRF2 and KEAP1 somatic alterations (as determinedby Sanger sequencing and FISH) and protein expression (IHC) are indicated by the color-coded bars above the expression heat map. The NRF2ACT subgroup isenriched for cases with alterations in NRF2 (amplification or mutation) and KEAP1 (mutation) (P < 0.001). High NRF2 protein expression is associatedwith the NRF2ACT subgroup (P ¼ 0.0002). For KEAP1, no association between gene expression class and protein expression by IHC was observed. B, hierarchicalclustering of 129 SCC cases (Raponi) using NRF2 28-gene set identifies two major subgroups.

Cescon et al.





Using gene expression and sequencing data from TCGA, weidentified a 28-gene set (NRF2ACT) that is able to separate SCCfrom multiple datasets into two subgroups. Application ofNRF2ACT in an adenocarcinoma dataset showed no similardiscriminatory ability, indicating that this signature is specificallyrelevant for the SCC histology, and that interrogation of theNRF2 pathway in adenocarcinoma (where KEAP1 mutationspredominate) may require a similarly dedicated approach. Thesubstantial overlap between this gene set and an independentlyderived list of genes regulated by NRF2 provided strong biologicconfirmation that NRF2ACT indeed reflects transcriptional acti-vation of this pathway. In both the TCGA derivation set and anindependent SCC series from UHN, where we performed muta-tional analysis for NRF2 pathway genes (NRF2, KEAP1), weobserved a strong association between the presence of NRF2pathway–activating somatic alterations and NRF2ACT-highexpression status. The concordance of these findings supports

the use of this gene expression surrogate of somatic NRF2pathway activation to test the prognostic and predictive associa-tions of this molecular subgroup. This provides a useful tool toanalyze numerous existing datasets where gene expression, butnot somatic mutational data, is available. Furthermore, as ameasure of downstream effects of the NRF2 transcription factor,a gene expression–based classifier may provide a more function-al measure of pathway activation than mutational status alone.Indeed, it is conceivable that by capturing the set of cancerswhere NRF2 transcriptional activity is upregulated (whether viasomatic alteration of NRF2, KEAP1, or CUL3, or by alternatemechanisms; ref. 35), gene expression could be a superiorbiomarker for this purpose.

In three independent datasets of SCC patients treated withsurgery alone, we observed no significant prognostic effect of ourNRF2ACT signature. However, in keeping with our hypothesis andthe preclinical data that supported it, we observed a trend toward

Figure 4.Prognostic effect of NRF2ACT signature in patients treatedwith surgery alone. Kaplan–Meier plots are shown for UHN cohort (A), Raponi cohort (B), and observation(surgery alone) arm (C) of JBR.10. No significant association between NRF2 signature status and survival was observed in any dataset.

Figure 3.Representative images of NRF2 andKEAP1 IHC staining, showing (A)NRF2-high (3þ) and (B) NRF2-low(0þ) and (C) KEAP1-high (3þ) and(D) KEAP1-low (0þ) examples.






differential benefit from the addition of adjuvant chemotherapyin JBR.10. These results suggest that patients withNRF2-activatingsomatic alterationsmaynot benefit fromadjuvant chemotherapy,possibly due to intrinsic chemoresistance conferred by activationof the NRF2 transcriptional program. This finding is consistentwith data from other cohort studies and case series in severaltumor types, including NSCLC and esophageal cancer, eventhough different methods to assess NRF2 activation were used(14, 17). To our knowledge, ours is the first study to evaluate apredictive marker of NRF2 activation in the setting of a random-ized trial with an untreated control arm.

To explore the possibility that conventional IHC staining ofFFPE tissues could be used to assess NRF2 mutational/activationstatus, we also stained UHN SCC samples for NRF2 and KEAP1.Comprehensive evaluation of mutational status, NRF2ACT signa-ture, and IHC revealed thatNRF2protein expressiondoes, indeed,correlate with the presence of pathway-activating mutations andwith the activated gene expression signature. These results are inkeeping with the predicted effects of the somatic alterations onNRF2 stability, and the consequent transcriptional activation oftarget genes when protein levels of NRF2 are increased. Theabsence of an association between KEAP1 staining andNRF2ACT is perhaps not surprising, given the potential for somaticalterations in NRF2, CUL3, or KEAP1 itself to disrupt the rela-tionship between KEAP1 and NRF2 protein/transcription factoractivity. Feedback regulation of KEAP1 in response to down-stream activation of NRF2 or other events could also complicateany expected associations. The concordance between NRF2 IHC,NRF2ACT, and NRF2-activatingmutations provides a rationale forfurther testing of NRF2 IHC staining as a potential predictor ofadjuvant chemotherapy benefit, which, if positive, could beincorporated readily into routine pathologic evaluation.Although we were unable to directly evaluate the predictive valueof somatic alterations in NRF2, KEAP1, and CUL3 in JBR.10, thestrong association betweenmutational status and gene expressionclass observed in the UHN validation dataset suggests that muta-tional status might also have potential as a predictive biomarker.

In conclusion, our results indicate that NRF2 pathway activa-tion, as defined by the NRF2ACT gene expression signature, mightserve as a biomarker of adjuvant cisplatin-based chemotherapybenefit in SCC. Validation of these findings is necessary, and thiswork provides a foundation for the comparative evaluation ofNRF2 activation using multiple modalities (gene expression sig-

natures, IHC, and somatic alteration profiling), all of whichshould be undertaken in randomized clinical trial patients anddatasets. If confirmed, gene expression or mutational analysis forNRF2 pathway alterations/activation could have clinical utility inidentifying patients unlikely to benefit from chemotherapy. Suchpatients could be spared the toxicity and costs of ineffectivetreatment, and instead be prioritized as candidates for trials ofmuch-needed alternate therapeutic approaches.

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

Authors' ContributionsConception and design: D.W. Cescon, M.-S. TsaoDevelopment of methodology: D.W. Cescon, D. She, S. Sakashita, C.-Q. Zhu,M.-S. TsaoAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S. Sakashita, C.-Q. Zhu, F.A. Shepherd, M.-S. TsaoAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D.W. Cescon, D. She, S. Sakashita, C.-Q. Zhu,M. Pintilie, F.A. Shepherd, M.-S. TsaoWriting, review, and/or revision of the manuscript: D.W. Cescon, D. She,S. Sakashita, M. Pintilie, F.A. Shepherd, M.-S. TsaoAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): S. Sakashita, F.A. Shepherd, M.-S. TsaoStudy supervision: F.A. Shepherd, M.-S. Tsao

AcknowledgmentsThe authors gratefully acknowledge Olga Ludkovski for performing the FISH

studies.

Grant SupportThis work was supported by the Canadian Cancer Society grant #020527, the

Ontario Premier's Award to Dr. F.A. Shepherd, the Princess Margaret CancerFoundation, and the Ontario Ministry of Long Term Health. Dr. S. Sakashita issupported by the Terry Fox Foundation Training Program in Molecular Pathol-ogy of Cancer at CIHR (STP 53912). Dr. F.A. Shepherd is the Scott Taylor Chairin Lung Cancer Research. Dr. M.-S. Tsao is the M. Qasim Choksi Chair in LungCancer Translational Research.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 25, 2014; revised February 20, 2015; accepted February 23,2015; published OnlineFirst March 4, 2015.

References1. Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon J-P, Van-

steenkiste J. Cisplatin-based adjuvant chemotherapy in patients withcompletely resected non–small-cell lung cancer. N Engl J Med 2004;350:351–60.

Figure 5.Effect of adjuvant chemotherapy onsurvival in JBR.10 SCC patients withNRF2ACT-low (left) and NRF2ACT-high(right) gene expression signature. Anonsignificant trend toward benefitfrom adjuvant chemotherapy isobserved only in the NRF2-lowsubgroup.

Cescon et al.





2. Scagliotti GV, Fossati R, Torri V, Crino L, Giaccone G, Silvano G, et al.Randomized study of adjuvant chemotherapy for completely resectedstage I, II, or IIIA non–small-cell Lung cancer. J Natl Cancer Inst 2003;95:1453–61.

3. Douillard J-Y, Rosell R, De Lena M, Carpagnano F, Ramlau R, Gonzales-Larriba JL, et al. Adjuvant vinorelbine plus cisplatin versus observation inpatients with completely resected stage IB-IIIA non–small-cell lung cancer(Adjuvant Navelbine International Trialist Association [ANITA]): a rando-mised controlled trial. Lancet Oncol 2006;7:719–27.

4. Winton T, Livingston R, Johnson D, Rigas J, Johnston M, Butts C, et al.Vinorelbine plus cisplatin vs. observation in resected non–small-cell lungcancer. N Engl J Med 2005;352:2589–97.

5. Strauss GM, Herndon JE, Maddaus MA, Johnstone DW, Johnson EA,Harpole DH, et al. Adjuvant paclitaxel plus carboplatin compared withobservation in stage IB non–small-cell lung cancer: CALGB 9633 withthe cancer and leukemia group B, radiation therapy oncology group,and north central cancer treatment group study groups. J Clin Oncol2008;26:5043–51.

6. Pignon J-P, TribodetH, Scagliotti GV,Douillard J-Y, Shepherd FA, StephensRJ, et al. Lung adjuvant cisplatin evaluation: a pooled analysis by the LACECollaborative Group. J Clin Oncol 2008;26:3552–9.

7. Arriagada R, Auperin A, Burdett S, Higgins JP, Johnson DH, Le Chevalier T,et al. Adjuvant chemotherapy, with or without postoperative radiotherapy,in operable non–small-cell lung cancer: two meta-analyses of individualpatient data. Lancet 2010;375:1267–77.

8. Shepherd FA, Domerg C,Hainaut P, Janne PA, Pignon J-P, Graziano S, et al.Pooled analysis of the prognostic and predictive effects of KRAS mutationstatus and KRAS mutation subtype in early-stage resected non–small-celllung cancer in four trials of adjuvant chemotherapy. J Clin Oncol2013;31:2173–81.

9. Zhu C-Q, Ding K, Strumpf D, Weir BA, Meyerson M, Pennell N, et al.Prognostic and predictive gene signature for adjuvant chemotherapy inresected non–small-cell lung cancer. J Clin Oncol 2010;28:4417–24.

10. Van Laar RK. Genomic signatures for predicting survival and adjuvantchemotherapy benefit in patients with non–small-cell lung cancer. BMCMed Genomics 2012;5:30.

11. Cancer Genome Atlas Research Network. Comprehensive genomic char-acterization of squamous cell lung cancers. Nature 2012;489:519–25.

12. Jaramillo MC, Zhang DD. The emerging role of the Nrf2–Keap1 signalingpathway in cancer. Genes Dev 2013;27:2179–91.

13. Shibata T, Ohta T, Tong KI, Kokubu A, Odogawa R, Tsuta K, et al. Cancerrelated mutations in NRF2 impair its recognition by Keap1-Cul3 E3ligase and promote malignancy. Proc Natl Acad Sci U S A 2008;105:13568–73.

14. Shibata T, Kokubu A, Saito S, Narisawa-Saito M, Sasaki H, Aoyagi K, et al.NRF2mutation confersmalignant potential and resistance to chemoradia-tion therapy in advanced esophageal squamous cancer. Neoplasia 2011;13:864–73.

15. MaX, Zhang J, Liu S,HuangY,ChenB,WangD.Nrf2 knockdownby shRNAinhibits tumor growth and increases efficacy of chemotherapy in cervicalcancer. Cancer Chemother Pharmacol 2012;69:485–94.

16. Yang H, Wang W, Zhang Y, Zhao J, Lin E, Gao J, et al. The role of NF-E2-related factor 2 in predicting chemoresistance and prognosis in advancednon–small-cell lung cancer. Clin Lung Cancer 2011;12:166–71.

17. Solis LM, Behrens C, Dong W, Suraokar M, Ozburn NC, Moran CA, et al.Nrf2 and Keap1 abnormalities in non–small cell lung carcinoma andassociation with clinicopathologic features. Clin Cancer Res 2010;16:3743–53.

18. Jiang T, Chen N, Zhao F, Wang X-J, Kong B, Zheng W, et al. High levels ofNrf2 determine chemoresistance in type II endometrial cancer. Cancer Res2010;70:5486–96.

19. Singh A, Boldin-Adamsky S, Thimmulappa RK, Rath SK, Ashush H,Coulter J, et al. RNAi-mediated silencing of nuclear factor erythroid-2-related factor 2 gene expression in non–small cell lung cancer inhibitstumor growth and increases efficacy of chemotherapy. Cancer Res2008;68:7975–84.

20. Homma S, Ishii Y, Morishima Y, Yamadori T, Matsuno Y, Haraguchi N,et al. Nrf2 enhances cell proliferation and resistance to anticancer drugs inhuman lung cancer. Clin Cancer Res 2009;15:3423–32.

21. Wang X-J, Sun Z, Villeneuve NF, Zhang S, Zhao F, Li Y, et al. Nrf2 enhancesresistance of cancer cells to chemotherapeutic drugs, the dark side of Nrf2.Carcinogenesis 2008;29:1235–43.

22. Ohta T, Iijima K,MiyamotoM,Nakahara I, TanakaH,OhtsujiM, et al. Lossof Keap1 function activates Nrf2 and provides advantages for lung cancercell growth. Cancer Res 2008;68:1303–9.

23. RaponiM, Zhang Y, Yu J, ChenG, Lee G, Taylor JMG, et al. Gene expressionsignatures for predicting prognosis of squamous cell and adenocarcinomasof the lung. Cancer Res 2006;66:7466–72.

24. Der SD, Sykes J, PintilieM, ZhuC-Q, StrumpfD, LiuN, et al. Validation of ahistology-independent prognostic gene signature for early-stage, non–small-cell lung cancer including stage IA patients. J Thorac Oncol 2014;9:59–64.

25. Tusher VG, Tibshirani R, Chu G. Significance analysis of microarraysapplied to the ionizing radiation response. Proc Natl Acad Sci U S A2001;98:5116–21.

26. Malhotra D, Portales-Casamar E, Singh A, Srivastava S, Arenillas D, HappelC, et al. Global mapping of binding sites for Nrf2 identifies novel targets incell survival response through ChIP-Seq profiling and network analysis.Nucleic Acids Res 2010;38:5718–34.

27. Hammerman PS, Sos ML, Ramos AH, Xu C, Dutt A, Zhou W, et al.Mutations in the DDR2 kinase gene identify a novel therapeutic targetin squamous cell lung cancer. Cancer Discov 2011;1:78–89.

28. Liao RG, Jung J, Tchaicha J, Wilkerson MD, Sivachenko A, Beauchamp EM,et al. Inhibitor-sensitive FGFR2 and FGFR3 mutations in lung squamouscell carcinoma. Cancer Res 2013;73:5195–205.

29. Mok TS, Wu Y-L, Thongprasert S, Yang C-H, Chu D-T, Saijo N, et al.Gefitinibor carboplatin-paclitaxel in pulmonary adenocarcinoma.NEngl JMed 2009;361:947–57.

30. Shaw AT, KimD-W, Nakagawa K, Seto T, Crino L, AhnM-J, et al. Crizotinibversus chemotherapy in advanced ALK-positive lung cancer. N Engl J Med2013;368:2385–94.

31. Li T, Kung H-J, Mack PC, Gandara DR. Genotyping and genomic profilingof non–small-cell lung cancer: implications for current and future thera-pies. J Clin Oncol 2013;31:1039–49.

32. Shaw AT, Ou S-HI, Bang Y-J, Camidge DR, Solomon BJ, Salgia R, et al.Crizotinib in ROS1-rearranged non–small-cell lung cancer. N Engl J Med2014;371:1963–71.

33. Rosell R, Bivona TG, Karachaliou N. Genetics and biomarkers in perso-nalisation of lung cancer treatment. Lancet 2013;382:720–31.

34. Singh A, Misra V, Thimmulappa RK, Lee H, Ames S, Hoque MO, et al.Dysfunctional KEAP1–NRF2 interaction in non–small-cell lung cancer.PLoS Med 2006;3:e420.

35. Hast BE, Goldfarb D, Mulvaney KM, Hast MA, Siesser PF, Yan F, et al.Proteomic analysis of ubiquitin ligase KEAP1 reveals associated pro-teins that inhibit NRF2 ubiquitination. Cancer Res 2013;73:2199–210.






2015;21:2499-2505. Published OnlineFirst March 4, 2015.Clin Cancer Res David W. Cescon, Desmond She, Shingo Sakashita, et al. Lung Squamous Cell CarcinomaNRF2 Pathway Activation and Adjuvant Chemotherapy Benefit in

Updated version

10.1158/1078-0432.CCR-14-2206doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2015/03/05/1078-0432.CCR-14-2206.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/21/11/2499.full#ref-list-1

This article cites 35 articles, 16 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/21/11/2499.full#related-urls

This article has been cited by 5 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/21/11/2499To 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-14-2206

http://clincancerres.aacrjournals.org/content/suppl/2015/03/05/1078-0432.CCR-14-2206.DC1

http://clincancerres.aacrjournals.org/content/21/11/2499.full#ref-list-1

http://clincancerres.aacrjournals.org/content/21/11/2499.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/21/11/2499


nrf2 pathway activation and adjuvant chemotherapy benefit in

Documents