an integrative analysis of the tumorigenic role of taz in ... · biology of human tumors an...

Biology of Human Tumors

An Integrative Analysis of the Tumorigenic Role of TAZ inHuman Non–Small Cell Lung Cancer

Satoshi Noguchi1, Akira Saito1,2, Masafumi Horie1, Yu Mikami1, Hiroshi I. Suzuki3, Yasuyuki Morishita3,Mitsuhiro Ohshima4, Yoshimitsu Abiko5, Johanna Sofia Margareta Mattsson6, Helena K€onig7, Miriam Lohr7,Karolina Edlund8, Johan Botling6, Patrick Micke6, and Takahide Nagase1

AbstractPurpose: TAZ, also knownasWWTR1, has recently been suggested as anoncogene in non–small cell lung

cancer (NSCLC). We investigated the clinical relevance of TAZ expression and its functional role in NSCLC

tumorigenesis.

Experimental Design:We characterized TAZ at the DNA (n¼ 192), mRNA (n¼ 196), and protein levels

(n ¼ 345) in an NSCLC patient cohort. Gene expression analysis was complemented by a meta-analysis of

public datasets (n ¼ 1,382). The effects of TAZ on cell proliferation and cell cycle were analyzed in cell

cultures and on tumor growth inmice. TAZ-dependentmicroarray-based expression profiles inNSCLC cells

were combined with molecular profiles in human NSCLC tissues for in silico analysis.

Results:Higher TAZmRNAandprotein levelswere associatedwith shorter patient survival. Transduction

of TAZ enhanced cell proliferation and tumorigenesis in bronchial epithelial cells, whereas TAZ silencing

suppressed cell proliferation and induced cell cycle arrest in NSCLC cells. Microarray and cell culture

experiments showed that ErbB ligands (amphiregulin, epiregulin, and neuregulin 1) are downstream targets

of TAZ. Our in silico analysis revealed a TAZ signature that substantiated the clinical impact of TAZ and

confirmed its relationship to the epidermal growth factor receptor signaling pathway.

Conclusion: TAZ expression defines a clinically distinct subgroup of patients with NSCLC. ErbB ligands

are suggested to mediate the effects of TAZ on lung cancer progression. Our findings emphasize the

tumorigenic role of TAZ and may serve as the basis for new treatment strategies. Clin Cancer Res; 20(17);

4660–72. �2014 AACR.

IntroductionLung cancer is the leading cause of cancer-related death

worldwide. Despite the high histologic heterogeneity, theoverall prognosis for lung cancer is extremely poor, withnon–small cell lung cancer (NSCLC) representing 85% ofall diagnosed cases (1). Through the thorough characteri-

zation of human cancer tissues, distinct molecular aberra-tions have been identified in small subgroups of patientsand are now being successfully exploited for therapeuticintervention, including epidermal growth factor receptor(EGFR) mutations, anaplastic lymphoma kinase, and c-rosoncogene 1 fusion proteins (2, 3). However, for the vastmajority of patients, treatment options are scant and theoverall prognosis remains poor, emphasizing the need touncover additional molecular mechanisms important inlung cancer tumorigenesis.

The transcriptional coactivator with PDZ-binding motif(TAZ), also knownasWW-domain containing transcriptionregulator-1 (WWTR1), was originally identified as a 14-3-3interacting protein (4). The regulation of TAZ and its para-log, Yes-associated protein (YAP), occurs primarily via theHippo pathway, which is conserved from flies to humans(5). TAZ acts mainly through the TEAD family of transcrip-tion factors to stimulate expression of genes that promotecell proliferation and control organ size (6). TAZ has alsobeen suggested to be involved in other biologic processes,including mesenchymal stem cell differentiation via mod-ulation of Runx2- and PPARg-dependent gene expression(7), self-renewal of human embryonic stem cells via regu-lation of the localization of Smad (8), and mechanotrans-duction (9). There is emerging evidence of crosstalk with

1Department of Respiratory Medicine, Graduate School of Medicine, TheUniversity of Tokyo, Bunkyo-ku, Tokyo, Japan. 2Division for Health ServicePromotion, The University of Tokyo, Bunkyo-ku, Tokyo, Japan. 3Depart-ment ofMolecular Pathology, Graduate School of Medicine, The Universityof Tokyo, Bunkyo-ku, Tokyo, Japan. 4Department of Biochemistry, OhuUniversity School of Pharmaceutical Sciences, Tomitamachi, Koriyama,Fukushima, Japan. 5Department of Biochemistry and Molecular Biology,Nihon University School of Dentistry at Matsudo, Matsudo, Chiba, Japan.6Department of Immunology, Genetics and Pathology, Uppsala University,Uppsala, Sweden. 7Department of Statistics, TU Dortmund University,Dortmund, Germany. 8Leibniz Research Centre for Working EnvironmentandHuman Factors (IfADo), TUDortmundUniversity, Dortmund, Germany.

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

Corresponding Author: Akira Saito, Department of Respiratory Medicine,Graduate School of Medicine, The University of Tokyo, 7-3-1 Hongo,Bunkyo-ku, Tokyo 113-0033, Japan. Phone: 81-3-3815-5411; Fax: 81-3-3815-5954; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-13-3328

�2014 American Association for Cancer Research.

ClinicalCancer

Research

Clin Cancer Res; 20(17) September 1, 20144660

on March 30, 2020. © 2014 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 20, 2014; DOI: 10.1158/1078-0432.CCR-13-3328

http://clincancerres.aacrjournals.org/

other signaling pathways, such as those of the protease-activated receptors (PAR), G-protein–coupled receptors(GPCR), and Wnt/b-catenin (6, 10).To investigate the physiologic role of TAZ, we previously

analyzed Taz-knockout mice (11). TAZ deficiency causedabnormal alveolarization during lung development, lead-ing to enlarged airspaces with low elastance. Taz-heterozy-gous mice were more resistant to bleomycin-induced lungfibrosis. We identified connective tissue growth factor(CTGF) as a downstream target of TAZ. TAZ has also beenshown to interact with thyroid transcription factor-1(TTF-1; ref. 12).Despite its obvious role in lung morphogenesis and

homeostasis under physiologic conditions, recent studiesshowed that TAZ is overexpressed in various human can-cers. TAZ contributes to the tumorigenesis of breast cancercells by promoting cell migration, invasion, anchorage-independent growth, and epithelial–mesenchymal transi-tion (EMT; refs. 13 and 14). In addition, TAZ confers cancerstem cell–related traits to breast cancer cells (15), furtherhighlighting its importance in tumor initiation and pro-gression. TAZ also plays a role in malignant glioma, colo-rectal cancer, and papillary thyroid cancer cells (16–18).Few studies have analyzed the impact of TAZ in lung

cancer, and these were restricted mainly to cell lines orincluded a limited number of NSCLC specimens (19–21).The aim of our study was to analyze the role of TAZ inNSCLC using an integrative strategy, combining resultsfrom in vitro and in vivo analyses withmolecular and clinicalinformation of human lung cancer tissues.

Materials and MethodsGlobal gene expression andgene copynumber analysesin human NSCLC tissuesThe study population consisted of patients with NSCLC

surgically treated between 1995 and 2005 with available

fresh frozen tissue in the Uppsala Biobank in the Depart-ment of Pathology (Uppsala lung cancer cohort; ref. 22).The study was approved by the Uppsala Ethical ReviewBoard (2006/325). RNA and DNA were isolated from196 frozen tissue samples and used for microarray analysisas described previously (GEO accessionGSE37745; ref. 22).

Themeta-analysis was performed as described previously(22) and included 9 publicly available Affymetrix expres-sion datasets [the Uppsala cohort, GSE37745; GSE14814(23); GSE3141 (24); GSE19188 (25); GSE4573 (26);GSE31210 (27); Shedden and colleagues 2008 (28);GSE29013 (29); GSE31547], together comprising 1,382patients with NSCLC and 22,277 probe sets overlappingthe arrays.

Whole-genome SNP array experiments were performedaccording to standard protocols for Affymetrix Gene ChipMapping 250K Nsp I arrays (Affymetrix) on 192 of thesecases, as described previously (30). Copy number estimatesfor each SNP were calculated as the ratio of the sample chipeffect in theUppsala cohort and the robust average across 90HapMap samples as the reference (31). These non-log scalevalues were multiplied by 2, and then CBS segmentationusing default parameters was applied to the SNP-wise copynumbers (32). TAZ copy number values werematched withthe gene expression probe sets using the mean chromo-somal position of the probe sets. Subsequently, the Spear-man rank correlation coefficient (r) between TAZ geneexpression and the copy number of the 3 probe sets wascalculated.

Tissue microarray construction andimmunohistochemistry

A tissue microarray (TMA) encompassing 355 NSCLCcases was constructed as described previously (33). Ofthese cases, 189 were also analyzed for global geneexpression and gene copy number changes. Immunohis-tochemistry was performed according to standard proce-dures (33). In brief, after antigen retrieval and blocking ofendogenous peroxidase, slides were incubated with theanti-TAZ antibody (1:40 dilution, HPA007415; SigmaAldrich) and the anti-amphiregulin antibody (1:10 dilu-tion, AF262; R&D Systems) for 30 minutes at roomtemperature (RT). After incubation with the secondaryantibody (Dako REAL EnVision; Dako) for 30 minutes atRT, the staining was developed using diaminobenzidinesolution for 10 minutes.

The immunostaining was evaluated under the supervi-sion of a lung pathologist (P. Micke). The staining intensitywas scored manually using a 4-grade scale: negative (0),weak (1), moderate (2), and strong (3). The fraction ofstained tumor cells was scored as follows: no positive cells(0), 1% to 9% (1), 10% to 49% (2), and 50% to 100% (3).For all stainings, one score was used per duplicate, repre-senting the same tumor sample, on the arrays. The ordinalscores for the staining intensity and the fraction of stainedtumor cellsweremultiplied, obtaining values in the range of0 to 9. For survival analysis, the TAZ staining score wasdichotomized at the median value of 6 (low: 0–6, high: 9).

Translational RelevanceTAZ has emerged as a key player in organ growth and

tumorigenesis. We combined in vitro, in vivo, and in situanalyses to define the role of TAZ in human lung cancer.The oncogenic properties of TAZ were confirmed in cellculture experiments and in a mouse model. The analysisof a large cohort of patients with non–small cell lungcancer (NSCLC) revealed interindividual diversity inTAZ protein and mRNA levels, and that patients withNSCLCwith higher tumor TAZ expression levels have anunfavorable prognosis. Moreover, we demonstrated thatTAZ regulates ErbB ligands, including amphiregulin,epiregulin, and neuregulin 1, and that the EGFR signal-ing pathway is activated in NSCLC tissues with high TAZexpression. The results not only confirm the clinicalrelevance of TAZ in patients with NSCLC but also offera conceptual basis for novel therapeutic strategies insubgroups of patients with NSCLC.

Tumorigenic Role of TAZ in NSCLC

www.aacrjournals.org Clin Cancer Res; 20(17) September 1, 2014 4661




Cell linesThe human lung cancer cell lines A549 and NCI-H441

(H441) were purchased from the American Type CultureCollection. Thehumanbronchial epitheliumcell lineBEAS-2B was obtained from the European Collection of CellCultures. The 293FT cells were from Invitrogen. All celllines were cultured according to the manufacturer’sprotocols.

Small interfering RNA experimentSmall interfering RNAs (siRNA) against human TAZ

(Stealth RNAi, si TAZ #1: HSS119545; si TAZ #2:HSS119546) and the negative control (Stealth RNAi, siNTC: #12935-300) were purchased from Invitrogen.A549 and H441 cells were transfected with 20 nmol/LsiRNA using Lipofectamine RNAiMAX (Invitrogen) accord-ing to the manufacturer’s instructions.

Construction of the microRNA expression vectorsThe microRNA (miRNA) expression vectors were con-

structed using the BLOCK-iT Pol II miR RNAi ExpressionVector Kit with EmGFP (Invitrogen). Three pairs of senseand antisense oligonucleotides were designed for target-ing TAZ (mi TAZ #1, #2, and #3; Supplementary TableS1) using BLOCK-iT RNAi Designer (http://rnaidesigner.Invitrogen.com/rnaiexpress/). The designed oligonucleo-tides were annealed, followed by ligation into thepcDNA6.2-GW/EmGFP-miR vector (Invitrogen). As thecontrol, pcDNA6.2-GW/EmGFP-miR negative controlplasmid (mi NTC; Invitrogen) was used. The sequencecontaining the miRNA coding region was transferred tothe lentivirus vector CSII-EF-RfA via the Gateway cloningsystem (Invitrogen) according to the manufacturer’sinstructions.

Cloning of human TAZ cDNATheopen reading frameof TAZ, derived from the cDNAof

A549 cells, was amplified by PCR using high-fidelity DNApolymerase (PrimeSTAR Max; Takara Bio). The purifiedPCR fragment was subcloned into the lentiviral expressionvector CSII-EF-RfA. The final insert was confirmed by DNAsequencing using the ABI PRISM 310 Genetic Analyzer. Agreen fluorescent protein (GFP) expression vector was usedas the control.

Lentiviral infectionRecombinant lentiviruses were produced by transient

transfection of 293FT cells with the lentiviral expressionvectors, pCMV-VSV-G-RSV-Rev and pCAG-HIVgp, usingLipofectamine 2000 reagent (Invitrogen). After 72 hoursof transfection, the medium was collected, and A549 orBEAS-2B cells were incubated with lentivirus-containingmedium.

Immunoblot analysisThe detailed procedures for immunoblot analysis were

described previously (34). In brief, cells were lysed, fol-lowed by SDS gel electrophoresis and semidry transfer of

the proteins to polyvinylidene difluoride membranes.They were probed with anti-TAZ (#2149; Cell Signaling),anti-phospho-EGFR Y1068 (#3777; Cell Signaling), anti-total EGFR (#E12020; Transduction Laboratories), oranti-a-Tubulin (Sigma-Aldrich) antibodies and theappropriate horseradish peroxidase–conjugated second-ary antibodies.

Cell proliferation analysisThe siRNA-transfected A549 and H441 cells and lentivi-

rally transduced BEAS-2B cells were seeded at a density of7.5 � 103, 5 � 104, and 2 � 104 per well, respectively, inquadruplicate wells of a 12-well plate. Cell numbers werecounted on days 2, 4, and 6 after seeding.

Colony formation assayThe miRNA-transduced cells were suspended in culture

medium containing 0.4% (w/v) agar (SeaPlaque GTG aga-rose; Lonza) and layered on top of culturemedium contain-ing 0.53% (w/v) agar at a density of 1.2 � 104 per well intriplicate wells of a 6-well plate. Colonies were allowed toform for 14 days and were then stained with MTT dye.Colonies �200 mm in diameter were counted using ImageJsoftware (http://imagej.nih.gov/ij/).

Cell cycle analysisThe siRNA-transfected A549 and H441 cells and lentivi-

rally transduced BEAS-2B cells were seeded at a density of4 � 105 per well into triplicate wells of a 6-well plate. After48 hours, cells were harvested, washed with phosphate-buffered saline (PBS), and fixed with 70% ethanol. Aftertreatment with RNaseA (Sigma-Aldrich), the cells werestained with propidium iodide (Sigma-Aldrich), and flowcytometry analysis was performed.

Apoptosis assayThe siRNA-transfected A549 and H441 cells were seeded

at a density of 4 � 105 per well into a 6-well plate. Afterovernight incubation, the cells were serum starved for 24hours, and Annexin V–stained apoptotic cells were detectedusing the Annexin V-FITC Apoptosis Detection Kit (Bio-Vision) according to the manufacturer’s instructions.

Tumorigenicity assayMale BALB/c nu/nu mice, 6 to 8 weeks old, were pur-

chased from the Charles River Laboratories Japan, Inc. Atotal of 1 � 107 A549 cells transduced with miRNAs (6–8mice per group) or BEAS-2B cells transduced with GFP orTAZ (10 mice per group) in 100-mL PBS were injected s.c.into mice. After 1 or 6 months respectively, the mice weresacrificed and the tumors excised and snap frozen. Thetumor volume was approximated as a � b � c/2 (wherea, b, and c represent the 3 different axes), and the tumorsamples were examined for cell morphology using hema-toxylin and eosin staining. The animal experiment wasconducted in accordance with national guidelines andapproved by the Ethics Committee for Animal Experimentsat the University of Tokyo.

Noguchi et al.

Clin Cancer Res; 20(17) September 1, 2014 Clinical Cancer Research4662




Gene expression profiling in A549 and H441 cellsTotal RNA was extracted from A549 and H441 cells 48

hours afterNTCor TAZ siRNA transfectionusing theRNeasyMini Kit (Qiagen) and subjected to the GeneChip HumanGene 1.0 ST Array (Affymetrix). The expression levels of19,734 genes were analyzed using the GeneSpring GXsoftware (Agilent Technologies). The dataset was depositedin the Gene Expression Omnibus database (GSE51270). Toidentify the TAZ-regulated genes, we integrated 2 publiclyavailable TAZ-dependent gene expression datasets derivedfrom MCF10A (13) and SW480 cells (GSE39904; ref. 35).Gene ontology (GO) analysis and gene set enrichmentanalysis (GSEA) were performed as described previously(36, 37).

Quantitative reverse transcription PCRTotal RNA was extracted from the cells, and first-strand

cDNA was synthesized using SuperScript III Reverse Tran-scriptase (Invitrogen). Quantitative reverse transcription(RT) PCR analysis was performed using the Mx-3000PqPCR System (Stratagene) andQuantiTect SYBRGreen PCR(Qiagen) according to the manufacturer’s protocol. Theexpression levels were normalized to that of glyceralde-hyde-3-phosphate dehydrogenase (GAPDH). The PCR pri-mers used are listed in Supplementary Table S2.

ELISAThe siRNA-transfected A549 cells were seeded at a density

of 1.5 � 105 per well into triplicate wells of a 6-well plate.After 48 hours, the culture medium was changed to serum-free medium. The cells were cultured for another 48 hours,and the culture supernatant was collected. Lentivirallytransduced BEAS-2B cells were seeded at a density of 1.5 �105 per well in epidermal growth factor (EGF)–deprivedmedium, and the culture medium was collected after 48hours. The concentration of amphiregulin was determinedusing the Human Amphiregulin DuoSet (R&D Systems)following the manufacturer’s protocol.

Statistical analysisThe correlations between TAZ expression levels and

clinicopathologic parameters or between TAZ andamphiregulin protein expression levels were analyzed bythe x2 test. Kaplan–Meier plots were used to visualize thedifference in survival between the 2 groups. The statisticalsignificance of the differences was determined by the log-rank test. Furthermore, univariate and multivariate Coxmodels were fitted with inclusion of established prog-nostic parameters: age, patient performance status, andstage at diagnosis. Spearman r was used to quantify thecorrelation between gene–gene and gene–protein expres-sion levels. The differences between the experimentalgroups were examined for statistical significance usingthe Student t test, the Wilcoxon rank-sum test, or ANOVAwith the Tukey post hoc test. The analyses were conductedusing JMP, version 9.0.3 (SAS Institute Inc.). P < 0.05 wasconsidered significant. All data are expressed as means �SEM.

ResultsTAZ gene expression and its clinical correlations inhuman NSCLC

Our study population was based on a clinically well-characterized lung cancer cohort comprising 196 patients(Uppsala cohort). The tumor tissues were annotated previ-ously for KRAS and EGFR mutations, whole genome genecopy number changes, and gene expression profiles(22, 38). Three probe sets (202132_at, 202133_at, and202134_s_at) with strong correlation (r ¼ 0.52–0.83)represent TAZ on the Affymetrix U133 Plus 2.0 array(GSE37745). Higher transcript levels of TAZ correlatedwithsquamous cell carcinoma histology (P ¼ 0.002) and youn-ger age (P ¼ 0.011, Supplementary Table S3). TAZ geneexpression levels were lower in tumor samples harboringEGFR mutations than in EGFR/KRAS mutation-negativesamples (P ¼ 0.046, Supplementary Fig. S1A). No correla-tions between TAZ mRNA expression levels and the otherevaluated parameters (sex, smoking status, stage, and per-formance status) were found (Supplementary Table S3).The correlation between gene expression profiles and genecopy number levels, determined from the Affymetrix 250KSNParrays, indicated that themRNA levels can be explainedin part by differences in the gene copy number of thecorresponding TAZ gene locus (P < 0.001; Fig. 1A). Uni-variate survival analysis of the Uppsala cohort revealed thatthe 202134_s_at probe set was significantly associated withsurvival [HR, 1.25; confidence interval (CI), 1.04–1.50; P¼0.021]. The expression of the other 2 TAZ probe sets did notshowprognostic impact (202132_at,P¼0.489; 202133_at,P ¼ 0.339). The 202134_s_at association remained signif-icant when TAZ expression levels were adjusted for impor-tant prognostic parameters, including stage, performancestatus, and age, in the multivariate analysis (HR, 1.24; CI,1.02–1.50; P ¼ 0.035).

To substantiate the potential prognostic impact of theprobe set 202134_s_at, a meta-analysis was performed.The analysis, which included 9 independent, publiclyavailable lung cancer datasets, confirmed the associationof higher TAZ mRNA levels with shorter survival (Fig. 1B;HR, 1.21; CI, 1.10–1.33; P < 0.001). This prognosticimpact was pronounced in the adenocarcinoma subgroup(Fig. 1C; HR, 1.33; CI, 1.16–1.52; P < 0.001) but couldnot be demonstrated when the squamous cell carcinomasubgroup was analyzed alone (HR, 1.05; CI, 0.90–1.22; P,0.568).

TAZ protein levels in human NSCLCTo transfer our findings from mRNA to in situ protein

levels, immunohistochemistry was performed on a TMAcomprising tissue cores from 355 patients. Staining wasevaluated for 345 cases and 18 additional nonmalignantlung tissue samples.

In the nonmalignant lung tissues, alveolar cells andbronchial epithelial cells showed positive nuclear staining.In addition, stromal cells, such as inflammatory cells, endo-thelial cells, and fibroblasts, demonstrated frequent nuclearpositivity (Supplementary Fig. S2A). Lung cancer tissues






revealed variable staining patterns, ranging from stronglypositive (Fig. 2A) to completely negative (Fig. 2B). Immu-nohistochemical staining scores correlated well with geneexpression levels in the same cancer sample, indicating thereliability of the antibody (Supplementary Fig. S2B).

For statistical analysis, the scores were dichotomized intolow versus high TAZ protein levels in the NSCLC cases.Correlation with clinical parameters revealed that highlevels were more common in squamous cell carcinoma(Supplementary Table S4). No correlations were observedbetween TAZ protein levels and other clinical parameters,including EGFR and KRASmutation status (SupplementaryFig. S1B). In correspondence with the gene expressionresults, high TAZ protein levels were associated with shortersurvival in theunivariate analysis (Table 1, left;HR, 1.40;CI,1.06–1.82; P ¼ 0.019), as illustrated by the Kaplan–Meieranalysis (Fig. 2C). When we evaluated the adenocarcinomaand squamous cell carcinoma subgroups separately, a trendwas observed in the univariate analysis (P ¼ 0.160 and0.087, respectively). The prognostic relevance in the wholecohort was retained in the multivariate analysis (Table 1,right; HR, 1.34; CI, 1.01–1.76; P ¼ 0.040).

In accordance with mRNA expression, gene copy num-bers also correlated with the TAZ protein expression level(r ¼ 0.17, P ¼ 0.025). Thus, TAZ expression could beexplained, at least in part, by gene copy number changes.These results indicated differential expression of TAZ inNSCLC and suggested that high TAZ expression at thegenomic, mRNA, and protein levels might define a clinicalsubset of patients with NSCLC.

In vitro analysis of TAZ in lung cancer cell lines andbronchial epithelial cells

To generate TAZ loss-of-function models, A549 andH441 human lung adenocarcinoma cells, which expressendogenous TAZ, were used for the silencing experiments.Cells were transiently transfected with NTC or TAZ siRNA.A549 cells were also lentivirally transduced with artificialmiRNAs against TAZ for colony formation assay and tumor-igenicity assay. To establish TAZ gain-of-function models,immortalized human bronchial epithelial cells (BEAS-2B)were lentivirally transduced with control GFP or TAZ.Knockdown or overexpression of TAZ was confirmed byimmunoblotting (Fig. 3A and Supplementary Fig. S3A). The

202132_at 202133_at 202134_s_at

TAZ gene copy number

TAZ

gen

e ex

pre

ssio

n

2 3 2 3 2 35

6

7

8

9

10

11

5

6

7

8

9

10

11

7

8

9

10

11

12

TAZ

gen

e ex

pre

ssio

n

TAZ

gen

e ex

pre

ssio

n

TAZ gene copy number TAZ gene copy number

P < 0.001r = 0.37

P < 0.001r = 0.44

P < 0.001r = 0.35

GSE14814

GSE19188

GSE29013

GSE31210

GSE3141

GSE31547

GSE4573

Shedden et al.

Uppsala

Study n

0.2

82

55

204

110

30

130

442

196

Fixed-effect model

Random-effects model

133

0.5 1 2 5

HR

1.24

1.07

1.65

1.29

1.17

0.94

1.38

1.25

1.00

(0.85–1.81)

(0.73–1.56)

(1.07–2.54)

(1.03–1.61)

(0.20–6.97)

(0.75–1.18)

(1.08–1.76)

(1.04–1.50)

(0.53–1.89)

HR

(95% CI)

(1.10–1.33)

(1.09–1.35)

1.21

1.21

All NSCLCs

GSE14814

GSE19188

GSE29013

GSE31210

GSE3141

GSE31547

Shedden et al.

Uppsala

Study n

0.2

40

30

204

58

30

442

106

Fixed-effect model

Random-effects model

71

0.5 1 2 5

HR

1.20

0.77

1.65

1.45

1.17

1.38

1.21

1.16

(0.64–2.25)

(0.39–1.51)

(1.07–2.54)

(1.10–1.92)

(0.20–6.97)

(1.08–1.76)

(0.92–1.60)

(0.53–2.56)

HR

(95% CI)

P < 0.001

(1.16–1.52)

(1.16–1.52)

1.33

1.33

AdenocarcinomaB C

A

P < 0.001

Figure 1. Correlation between TAZ gene copy number and gene expression level, and a meta-analysis of the prognostic impact of TAZ gene expression. A,scatter plots for TAZ gene copy numbers and gene expression levels. Spearman correlation coefficients (r) andP valueswere calculated for each probe set. Band C, forest plots visualizing the results from the meta-analysis of the TAZ probe set 202134_s_at, including 1,382 NSCLC (B) or 981 adenocarcinoma (C)patients from 9 studies. The overall effect was quantified by the P value of the random effects model.

Noguchi et al.





cells did not show morphologic changes after siRNA trans-fection or lentiviral transduction.We investigated the effect of TAZ on cell proliferation.

TAZ knockdown in A549 and H441 cells inhibited cellproliferation (Fig. 3B), whereas TAZ overexpression inBEAS-2B cells enhanced cell proliferation slightly (Fig.3C). Soft agar colony formation assays showed that A549cells with miRNA-mediated TAZ knockdown exhibited areduction in anchorage-independent growth on soft agar(Supplementary Fig. S3B).To explore themechanisms by which TAZ promotes lung

cancer cell proliferation, we carried out cell cycle analysis.In line with the findings of the cell proliferation assay, theG0/G1 fraction was increased by TAZ knockdown in A549cells (Fig. 3D), whereas G1/S transition was promoted byTAZ overexpression in BEAS-2B cells (Fig. 3E).To investigate the involvement of TAZ in apoptosis, A549

and H441 TAZ knockdown cells were serum starved, andapoptotic cells were detected by Annexin V staining. We

found a significant increase in apoptosis induced by serumdeprivation in A549 and H441 cells after TAZ knockdown(Supplementary Fig. S3C). These results supported that theTAZ oncogene promotes lung cancer proliferation andsurvival.

Tumorigenic impact of TAZ in a mouse modelTo assess the potential tumorigenicity of TAZ in vivo, a

xenograft mouse model was generated by injecting nudemice with miRNA-transduced A549 cells or TAZ-trans-duced BEAS-2B cells. All the mice injected with miRNA-transduced A549 cells developed tumors, and TAZ knock-down reduced tumor growth (Fig. 3F). On the otherhand, during the observation period, 5 of 10 miceinjected with TAZ-transduced BEAS-2B cells developedtumors, whereas only 1 mouse demonstrated minimaltumor growth after 6 months in the control group (Sup-plementary Fig. S3D and S3E). Tumor samples showedsimilar cell morphology (Supplementary Fig. S3F). These

A

B

100 μm

100 μm

CTAZ protein high

TAZ protein low

0 1 2 3 4 5 60.0

0.2

0.4

0.6

0.8

1.0

Pro

po

rtio

n a

live

7 8 9 10 11

Years

100 μm

100 μm

P = 0.016

Figure 2. TAZ protein expression inNSCLC. A and B, the TMA,consisting of 345 evaluableNSCLC cases, was stained with apolyclonal antibody againstTAZ, and representativeimmunostaining images arepresented for high (A) and low TAZexpression (B). C, the stainingscores (0–9 as a product of thestaining intensity and theproportion of stained tumor cells)were dichotomized (0–6 vs. 9) andused for stratification in theKaplan–Meier analysis. P valuewas obtained using a log-rank test.






0

30

10

20

40

024 48 72 96 120 144 0 24 48 72 96 120 144

20

10

30

0

A549 H441

30

10

20

40

0

50

0 24 48 72 96 120 144

Time (h)

Time (h)

Time (h)

B

CGFP

TAZ*

*si NTC

si TAZ #1

si TAZ #2

si NTC

si TAZ #1

si TAZ #2

*

*

*

AC

ell n

um

ber

Cel

l nu

mb

er

Cel

l nu

mb

er

Cel

l nu

mb

er

DNA content0 1,024 DNA content0 1,024 DNA content0 1,024 DNA content0 1,024

si NTCG0/G1 S G2/M

G0/G1 S G2/M G0/G1 S G2/M

G0/G1 S G2/M G0/G1 S G2/M G0/G1 S G2/M

Cel

l pro

po

rtio

n (

%)

512

0

512

0

512

0

512

0

si TAZ #1 GFP TAZ

0

10

20

30

40

50

60

A549

Cel

l pro

po

rtio

n (

%)

0

10

20

30

40

50

D E

*

*

*si NTC

si TAZ #1

GFP

TAZ

BEAS-2B

BEAS-2B

Cel

l nu

mb

er (1

04)

Cel

l nu

mb

er (1

04)

Cel

l nu

mb

er (1

04)

α-Tubulin

TAZ

si NTC si TAZ #1

A549

GFP TAZ

BEAS-2B

si TAZ #2 si NTC si TAZ #1

H441

si TAZ #2

F

mi NTC mi TAZ #1 mi TAZ #2 mi TAZ #3

0

40

80

120

160

**

*

mi NTC mi TAZ #1 mi TAZ #2 mi TAZ #3

0

40

80

120

160

Tu

mo

r vo

lum

e (m

m3 )

Tu

mo

r w

eig

ht (

mg

)

200

Noguchi et al.





results indicated the potential role of TAZ in tumorigen-esis of bronchial epithelial cells.

Gene expression profiling in TAZ knockdown lungcancer cellsTo identify the downstream targets of TAZ in lung cancer

cells, we performed microarray-based expression profilinganalysis in A549 cells after TAZ knockdown. Following TAZsiRNA transfection, TAZ expression was less than 20%. TAZsuppression led to increased expression (>1.5-fold) of 227transcripts and decreased expression (<0.67-fold) of 131transcripts. Key molecules involved in lung epithelial celldifferentiation, survival, and proliferation are shown in Fig.4A and Supplementary Fig. S4A. Similar TAZ knockdownexperiments were performed in H441 cells, which demon-strated increased expression (>1.5-fold) of 69 transcriptsand decreased expression (<0.67-fold) of 138 transcripts.The differentially expressed genes after TAZ knockdown inA549 and H441 cells showed remarkable overlap (30transcripts).To validate our results, we integrated 2 additional gene

lists from cell lines with altered TAZ expression, yielding atotal of 4 gene lists: the genes upregulated (>2.0-fold) inMCF10A cells (breast epithelial cell) by retroviral trans-duction of TAZ (13) and the genes downregulated (<0.67-fold) in A549, H441, and SW480 (colon cancer) cells bysiRNA-mediated TAZ knockdown (35). There were 259overlapping genes in at least 2 cell lines, defined as TAZ-regulated genes (Supplementary Table S5). To character-ize these 259 genes, we performed GO analysis and foundthat a large portion of the TAZ-regulated genes is associ-ated with cell cycle regulation (Fig. 4B). The in vitrofindings were recapitulated in human NSCLC tissues ofthe Uppsala cohort using GSEA. TAZ-regulated genes wereclearly enriched in NSCLC with high TAZ expression (Fig.4C). These genes were also enriched in NSCLC tissues of

patients with poor prognosis (Fig. 4D), suggesting thatTAZ target genes might be responsible for the aggressivephenotype of NSCLC.

The downregulated genes in the A549 cells after TAZknockdown included the cell cycle regulators, cyclin D1and D3 (CCND1 and CCND3), as well as ErbB ligands,amphiregulin (AREG), epiregulin (EREG), andneuregulin 1(NRG1), which were also listed as TAZ-regulated genes inother cell types (Supplementary Table S5). Because theEGFR family and its ligands play an important role in lungcancer, these findings prompted us to validate the micro-array data and to explore the TAZ-mediated regulation ofthe 3 ErbB ligands in detail.

TAZ regulates expression of ErbB ligands in lungepithelial cells

The gene expression differences determined by microar-ray analysis of the ErbB ligands inA549 cellswere confirmedby quantitative RT-PCR (Fig. 5A). We also found that theywere upregulated in TAZ-overexpressing BEAS-2B cells.ELISA assays showed that TAZ knockdown led to decreasedamphiregulin protein levels in the culture medium of A549cells (Fig. 5B). In line with this, TAZ overexpressionincreased amphiregulin protein levels in the supernatantof cultured BEAS-2B cells. We found that the promoterregions of the 3 ErbB ligands contain putative recognitionmotifs for TEAD, amajor transcription factor that binds TAZ(Supplementary Fig. S4B and S4C).

Correlation between TAZ and ErbB ligand expressionin NSCLC

We next investigated whether there is a correlationbetween TAZ and ErbB ligand expression using publiclyavailable microarray datasets. There were positive correla-tions in the NSCLC cell lines according to 2 datasets,GSE4127 (39) and GSE4824 (40), except for neuregulin

Table 1. Univariate and multivariate Cox regression model of 345 NSCLC cases including TAZ expression(high expression vs. low expression) and themost important prognostic parameters (dichotomized stage II–IV vs. I, age >70 vs. �70 years, performance status I–III vs. 0)

Univariate Multivariate

HR (95% CI) P value HR (95% CI) P value

TAZ 1.40 (1.06–1.82) 0.019 1.34 (1.01–1.76) 0.040Stage 1.59 (1.21–2.06) <0.001 1.67 (1.28–2.17) <0.001Performance status 1.79 (1.38–2.31) <0.001 1.68 (1.30–2.18) <0.001Age 1.63 (1.25–2.12) <0.001 1.52 (1.16–1.99) 0.003

Figure 3. Effects of TAZ on lung cell proliferation, cell cycle progression, and tumor formation. A, immunoblot analysis showed the effects of TAZ knockdownand overexpression. A549 and H441 cells were transfected with si NTC (control siRNA) or si TAZ (siRNA for TAZ). BEAS-2B cells were transducedwithGFPor TAZ.a-Tubulinwasusedas the loading control. BandC,numbersof siRNA-transfectedA549andH441cells (B) and lentivirally transducedBEAS-2B cells (C) were counted on days 2, 4, and 6 after seeding. D and E, A549 cells transfected with si NTC or si TAZ (D) and BEAS-2B cells transducedwith GFP or TAZ (E) were stained with propidium iodide, and flow cytometry analysis was performed. Data are representative of three independentexperiments. Error bars: SEM. F, a total of 1�107A549 cells transducedwithmiRNAswere injected subcutaneously into nudemice. Tumor volume and tumorweight weremeasured 1month later. �,P < 0.05; Student t test; si NTC versus si TAZ #1 or #2 (B, D); GFP versus TAZ (C, E); andmi NTC versusmi TAZ #1, #2,or #3 (F).






1 in GSE4127, in which only a trend toward a positivecorrelationwas observed (Fig. 5C). The correlation betweenTAZ and amphiregulin expression was also analyzed inhuman cancer tissues using immunohistochemistry in344 evaluable cases of the Uppsala cohort (SupplementaryFig. S5). Amphiregulin staining correlated well with geneexpression levels within the same cancer samples (r¼ 0.22,P ¼ 0.003), confirming the validity of the antibody. Fur-thermore, a positive correlation between TAZ and amphir-egulin protein levels was found (P < 0.001; Fig. 5D).

To evaluate whether the EGFR pathway was activated inNSCLC with high TAZ expression, we again used GSEA. AnEGFR signaling–related gene set from a previous study (41)was enriched in NSCLC with high TAZ expression in theUppsala cohort (P ¼ 0.004; Fig. 5E). Three other EGFRsignaling–related gene sets also showed positive correla-tions with high TAZ expression, but they did not reachstatistical significance (Supplementary Fig. S6A; refs. 42and43). Immunoblot analysis showed that TAZoverexpres-sion increased EGFR phosphorylation in BEAS-2B cells,

A

BRank Name Gene Ontology ID P value

1 Mitotic cell cycle GO:0000278 3.462E–182 Cell cycle process GO:0022402 2.457E–153 Cell cycle process GO:0007049 2.563E–154 Cell cycle phase GO:0022403 1.932E–125 Mitotic prometaphase GO:0000236 3.464E–106 Regulation of cell cycle GO:0051726 6.886E–107 Regulation of cell cycle process GO:0010564 9.157E–108 M phase of mitotic cell cycle GO:0000087 1.111E–99 M phase GO:0000279 1.460E–910 Cell cycle checkpoint GO:0000075 9.838E–9

C DGene set: TAZ-regulated genesTAZ-high vs. TAZ-low, P = 0.002

Gene set: TAZ-regulated genespoor prognosis vs. good prognosis, P = 0.014

TAZ-low

Poor prognosis

Good prognosis

TAZ -high

Fold changeQuantitative RT-PCR

Gene symbol Gene name Affymetrix chip si TAZ #1 si TAZ #2 Pathway involvedAREG Amphiregulin 0.47 0.51 0.40 Cell proliferationEREG Epiregulin 0.60 0.52 0.73 Cell proliferationNRG1 Neuregulin 1 0.66 0.53 0.56 Cell proliferationDKK1 Dickkopf homolog 1 0.66 0.39 0.45 Wnt signalingCYR61 Cysteine-rich, angiogenic inducer, 61 0.72 0.46 0.49 Cell proliferationCTGF Connective tissue growth factor 0.67 0.28 0.37 Cell proliferationCCND1 Cyclin D1 0.60 0.60 0.71 Cell cycleCCND3 Cyclin D3 0.56 0.64 0.77 Cell cycleANKRD1 Ankyrin repeat domain 1 0.41 0.16 0.12 MechanotransductionSPDEF SAM pointed domain containing ets transcription factor 1.61 4.8 3.1 Cell differentiationROR1 Receptor tyrosine kinase-like orphan receptor 1 1.68 2.5 1.6 Cell survivalTGFBR1 Transforming growth factor, beta receptor 1 1.74 1.9 6.6 TGFβ signalingSMAD5 SMAD family member 5 2.09 2.7 2.3 TGFβ signalingRASSF2 Ras association (RalGDS/AF-6) domain family member 2 1.54 1.9 1.4 Apoptosis

Figure 4. Analyses of TAZ-regulated genes. A, selected TAZ-regulated genes in A549 cells. B, GO analysis of TAZ-regulated genes obtained from mergingof the data from the A549, H441, MCF10A, and SW480 cells. C and D, GSEA was performed to examine the enrichment of TAZ-regulated genes inNSCLC samples of the Uppsala cohort. Patients were divided into 2 groups according to the expression level of TAZ or overall survival. The enrichment ofTAZ-regulated genes is shown schematically with the genes that correlated best with high TAZ expression (C) or poor prognosis (D) on the left and thegenes that correlated best with low TAZ expression (C) or good prognosis (D) on the right.

Noguchi et al.





A549

BEAS-2B

Rel

ativ

e ex

pre

ssio

n

0.5

1.0

0

1.5

2.0

0.5

1.0

0

1.5

2.0

0.5

1.0

0

1.5

2.0

AREG EREG

GFP TAZ

NRG1R

elat

ive

exp

ress

ion

si NTC si TAZ #1 si TAZ #2

0.6

0.2

0.4

0.8

0

1.0

1.2

0.6

0.2

0.4

0.8

0

1.0

1.2

0.6

0.2

0.4

0.8

0

1.0

1.2

Rel

ativ

e ex

pre

ssio

n

Rel

ativ

e ex

pre

ssio

n

si NTC si TAZ #1 si TAZ #2 si NTC si TAZ #1 si TAZ #2

GFP TAZ

Rel

ativ

e ex

pre

ssio

n

Rel

ativ

e ex

pre

ssio

n

GFP TAZ

A

A549

si NTC si TAZ #1

GFP TAZ

AR

EG

(p

g/m

L)

400

800

0

1,200

1,600

2,000

2,400

AR

EG

(p

g/m

L)

40

80

0

120

160

200

240

280 *

si TAZ #2

**B

TAZ gene expression

AR

EG

gen

e ex

pre

ssio

n

6 7 8 9 105

6

8

4

10C

P = 0.002r = 0.43

2

12

4 11

12

P = 0.005r = 0.39

P = 0.647r = 0.11

P = 0.001r = 0.69

P = 0.017r = 0.33

GSE4127(n = 19)

GSE4824(n = 52)

2

6

9P = 0.003r = 0.64

5 6 7 8 9 10

6 7 8 9 105 6 7 8 9 105

6

8

4

10

7

8

ER

EG

gen

e ex

pre

ssio

n

NR

G1

gen

e ex

pre

ssio

nTAZ gene expression TAZ gene expression

5

8

6

7

2

6

8

4

10

0

-2

6

8

4

10

4 115 6 7 8 9 104 115 6 7 8 9 10

TAZ gene expression

AR

EG

gen

e ex

pre

ssio

n

ER

EG

gen

e ex

pre

ssio

n

NR

G1

gen

e ex

pre

ssio

n

TAZ gene expression TAZ gene expression

D E Gene set: EGFR signaling-related genesTAZ-high vs. TAZ-low, P = 0.004

n (%) AREG staining scoreTotal

0 1–2 ≥3TAZ staining score 0–3 89 (66.4%) 20 (14.9%) 25 (18.7%) 134 (100%)

4–6 55 (48.3%) 29 (25.4%) 30 (26.3%) 114 (100%)9 28 (29.2%) 30 (31.2%) 38 (39.6%) 96 (100%)

Total 172 79 93 344

TAZ-low

TAZ -high

BEAS-2B

* *

**

**

**

Figure 5. TAZ regulates the expression of ErbB ligands. A, the relative expression levels of the indicated transcripts in A549 cells transfected with si NTC or siTAZ and in BEAS-2B cells transduced with GFP or TAZ. B, the culture media were collected from A549 cells transfected with si NTC or si TAZ and BEAS-2Bcells transduced with GFP or TAZ. The concentration of amphiregulin was measured by ELISA. The results represent the concentration per 1 � 106

cells. Error bars: SEM. �, P < 0.05; Student t test. C, the correlations between TAZ and ErbB ligand expression from 2 microarray datasets from NSCLCcell lines, GSE4127 and GSE4824. Probe sets: TAZ, 202132_at; amphiregulin, 205239_at; epiregulin, 205767_at; neuregulin 1, 208241_at. Spearmancorrelation coefficients (r) and the P values were calculated. D, the correlation between TAZ protein expression and amphiregulin in the TMA. E, GSEA wasperformed to examine the enrichment of EGFR signaling related genes in NSCLC samples of the Uppsala cohort. Patients were divided into 2 groupsaccording to the TAZ expression level. The EGFR signaling signature was based on the genes whose expression peaked 8 hours after stimulation ofMCF10Acells with EGF. The enrichment of EGFR signaling-related genes is shown schematically with the genes that correlated best with high TAZ expression on theleft and those that correlated best with low TAZ expression on the right.






although we failed to observe decreased EGFR phosphor-ylation after siRNA-mediated TAZ knockdown (Supple-mentary Fig. S6B). Altogether, these findings suggested thatTAZ activates the EGFR signaling pathway through theinduction of EGFR ligands.

Correlation between TAZ expression and sensitivity toEGFR inhibitors

To investigate whether EGFR inhibition is a reasonabletreatment strategy for NSCLC with high TAZ expression,drug sensitivity data and gene expression profiles in NSCLCcells were obtained from the Cancer Cell Line Encyclopedia(CCLE) database (44). We found that TAZ expressiontended to be higher inNSCLC cell lines with high sensitivityto 2 EGFR tyrosine kinase inhibitors (EGFR-TKIs), whereasthere were no apparent associations between TAZ expres-sion and sensitivity to other anticancer drugs (Supplemen-tary Fig. S7A). In addition, 139 genes that define gefitinibsensitivity (45) were enriched in NSCLC with high TAZexpression (Supplementary Fig. S7B). Thus, the in silicoanalysis indicated a rational strategy for EGFR inhibitionin TAZ-positive NSCLC.

DiscussionUsing an integrated approach, we characterized TAZ

expression at the genomic, mRNA, and protein levels in alarge, well-annotated NSCLC patient cohort to combinecomprehensivemolecular data with clinical parameters andto identify TAZ as an independent prognostic marker. Con-sistent with the in situ findings in human lung cancer tissue,TAZ promoted lung cancer/epithelial cell proliferation invitro and tumor development in an in vivo mouse model.Finally, based on multiple experimental lines, tumorigeniceffects might be regulated through EGFR signaling.

The association of higher TAZ expression with shortersurvival in patients with lung cancer was described previ-ously in one study (20). However, there were certain differ-ences of interest. First, the staining pattern of the antibodyused in our study clearly indicated TAZ expression innormal alveolar and bronchial epithelial cells as well as instromal cells. TAZ protein expression in nonneoplastic lungtissue was not described in the previous report. Second, asopposed to theprevious study, TAZexpression levelswere ingeneral higher in squamous cell carcinoma in our study,which was also supported by analyses of all other micro-array datasets (data not shown). This discrepancy might becausedbydifferences in the clinicopathologic features of thestudy cohorts and/or the methods used for immunohis-tochemistry, including the use of different antibodies.

We found a clinical subset of patients with NSCLC withhigh TAZ expression at the genomic, mRNA, and proteinlevels in the Uppsala cohort. The previous studies showedthat high level of TAZ gene amplification analyzed byGISTIC algorithm was observed in 35 of 178 squamouscell carcinomas (19.7%; ref. 46) and 1 of 182 adenocarci-nomas (0.5%; ref. 47). This is consistent with the findingthat TAZ expression levels were higher in squamous cellcarcinoma in our study. They also showed that TAZ muta-

tion was found in only 1 case (0.6%) of squamous cellcarcinomas, and 2 cases (1.1%) of adenocarcinomas, sug-gesting TAZ activation may not be generally caused bygenetic mutations.

We showed that miRNA-mediated TAZ knockdown inA549 cells exhibited tumor growth reduction in a xenograftmodel, which was consistent with the previous study (19).Furthermore, we found that TAZ-transduced BEAS-2B cellsshowed a higher frequency of tumor formation in nudemice. Because lentiviral transduction of BEAS-2B cells mayitself induce tumorigenesis (48), the results in this modelshould be interpreted with caution and will require furtherinvestigation.

To further explore which signaling pathways TAZ regu-lates, we performed a comparative gene expression analysisand identified amphiregulin, epiregulin, and neuregulin 1as candidate genes. We demonstrated that TAZ inducesamphiregulin secretion from lung epithelial cells. Ourresults were consistent with a previous report that TAZincreased amphiregulin production in mammary epithelialcells (49). Epiregulin and neuregulin 1 were also among thepotential targets found in the microarray screening of A549cells, as well as MCF10A and SW480 cells, with TAZ knock-down (13, 35). Importantly, their promoter regions werelikely to harbor TEAD-binding sites, suggesting direct reg-ulatory interactions.

EGFR-TKIs are now widely accepted as effective thera-peutic strategies for patients with EGFR mutation-positiveNSCLC. However, some EGFR mutation–negative patientsalso benefit from EGFR-TKI treatment (50). Indeed, a pre-vious study linked increased expression of the EGFR ligandamphiregulin to improved clinical outcomes inEGFRmuta-tion-negative patients treated with EGFR-TKI (51). TAZexpression may help to identify those patients who wouldbenefit from treatment with EGFR-TKIs. We report heresupportive findings using data from the CCLE database.However, it should be stressed that other strategies havebeen suggested to inhibit oncogenic TAZ activation, includ-ing direct targeting of TAZ, Rho, and Rock inhibitors, andupstream WNT inhibition (6).

Indeed, our integrative microarray analysis indicatedseveral other mechanisms of TAZ-induced tumorigenesis.TAZ may facilitate cancer progression through activation oftumor-promoting genes, such as CTGF, Cyr61, and cyclins(CCND1 and CCND3). Further studies are clearly war-ranted to identify the TAZ-dependent regulatory networkin cancer.

In summary, our study confirms the tumorigenic impactand clinical significance of TAZ expression in NSCLC andsheds light on the regulatory mechanisms and the potentialfor TAZ as a therapeutic target in the era of personalizedmedicine.

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

Authors' ContributionsConception and design: S. Noguchi, A. SaitoDevelopment of methodology: S. Noguchi, Y. Abiko

Noguchi et al.





Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S. Noguchi, A. Saito, K. Edlund, J. Botling, P.MickeAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): S. Noguchi, A. Saito, H.I. Suzuki,Y. Morishita, J.S.M. Mattsson, H. K€onig, M. Lohr, P. MickeWriting, review, and/or revision of themanuscript: S. Noguchi, A. Saito,Y. Mikami, H.I. Suzuki, M. Ohshima, J.S.M. Mattsson, H. K€onig, J. Botling,P. Micke, T. NagaseAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): A. Saito, M. Horie, Y. Morishita,M. Ohshima, K. Edlund, T. NagaseStudy supervision: A. Saito, T. Nagase

AcknowledgmentsThe authors thank Asayo Imaoka and Makiko Sakamoto for their tech-

nical assistance. The authors also thank Akihisa Mitani for the usefuldiscussions.

Grant SupportThis work was supported by Scandinavia-Japan Sasakawa Foundation

(to S. Noguchi), KAKENHI (Grants-in-Aid for Scientific Research) fromthe Ministry of Education, Culture, Sports, Science, and Technology ofJapan (Grant #23790899 to A. Saito, Grant #23249045 to T. Nagase), anda grant to the Respiratory Failure Research Group and the Research onAllergic Disease and Immunology, Health and Labour Sciences ResearchGrants from the Ministry of Health, Labour and Welfare, Japan (toT. Nagase).

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 December 9, 2013; revised May 14, 2014; accepted June 5, 2014;published OnlineFirst June 20, 2014.

References1. Herbst RS, Heymach JV, Lippman SM. Lung cancer. N Engl J Med

2008;359:1367–80.2. Davies KD, Doebele RC.Molecular pathways: ROS1 fusion proteins in

cancer. Clin Cancer Res 2013;19:4040–5.3. Gainor JF, Varghese AM, Ou SH, Kabraji S, Awad MM, Katayama R,

et al. ALK rearrangements are mutually exclusive with mutations inEGFR or KRAS: an analysis of 1,683 patients with non-small cell lungcancer. Clin Cancer Res 2013;19:4273–81.

4. Kanai F, Marignani PA, Sarbassova D, Yagi R, Hall RA, Donowitz M,et al. TAZ: a novel transcriptional co-activator regulated by interactionswith 14-3-3 and PDZ domain proteins. EMBO J 2000;19:6778–91.

5. Zhao B, Tumaneng K, Guan KL. The Hippo pathway in organ sizecontrol, tissue regeneration and stem cell self-renewal. Nat Cell Biol2011;13:877–83.

6. Piccolo S, Cordenonsi M, Dupont S. Molecular pathways: YAP andTAZ take center stage in organ growth and tumorigenesis. Clin CancerRes 2013;19:4925–30.

7. Hong JH, Hwang ES, McManusMT, AmsterdamA, Tian Y, KalmukovaR, et al. TAZ, a transcriptional modulator of mesenchymal stem celldifferentiation. Science 2005;309:1074–8.

8. Varelas X, Sakuma R, Samavarchi-Tehrani P, Peerani R, Rao BM,DembowyJ, et al. TAZcontrolsSmadnucleocytoplasmic shuttling andregulates human embryonic stem-cell self-renewal. Nat Cell Biol2008;10:837–48.

9. Dupont S,Morsut L, AragonaM, Enzo E, Giulitti S, Cordenonsi M, et al.Role of YAP/TAZ in mechanotransduction. Nature 2011;474:179–83.

10. Mo JS, Yu FX, Gong R, Brown JH, Guan KL. Regulation of the Hippo-YAP pathway by protease-activated receptors (PARs). Genes Dev2012;26:2138–43.

11. Mitani A, Nagase T, Fukuchi K, Aburatani H, Makita R, Kurihara H.Transcriptional coactivator with PDZ-binding motif is essential fornormal alveolarization in mice. Am J Respir Crit Care Med 2009;180:326–38.

12. Park KS, Whitsett JA, Di Palma T, Hong JH, Yaffe MB, Zannini M. TAZinteracts with TTF-1 and regulates expression of surfactant protein-C.J Biol Chem 2004;279:17384–90.

13. Zhang H, Liu CY, Zha ZY, Zhao B, Yao J, Zhao S, et al. TEADtranscription factors mediate the function of TAZ in cell growthand epithelial-mesenchymal transition. J Biol Chem 2009;284:13355–62.

14. ChanSW, LimCJ,GuoK,NgCP, Lee I, HunzikerW, et al. A role for TAZin migration, invasion, and tumorigenesis of breast cancer cells.Cancer Res 2008;68:2592–8.

15. Cordenonsi M, Zanconato F, Azzolin L, Forcato M, Rosato A, FrassonC, et al. The Hippo transducer TAZ confers cancer stem cell-relatedtraits on breast cancer cells. Cell 2011;147:759–72.

16. Wang L, Shi S, GuoZ, ZhangX, HanS, YangA, et al. Overexpression ofYAP and TAZ is an independent predictor of prognosis in colorectalcancer and related to the proliferation and metastasis of colon cancercells. PLoS ONE 2013;8:e65539.

17. de Cristofaro T, Di Palma T, Ferraro A, Corrado A, Lucci V, Franco R,et al. TAZ/WWTR1 is overexpressed in papillary thyroid carcinoma. EurJ Cancer 2011;47:926–33.

18. Bhat KP, Salazar KL, Balasubramaniyan V, Wani K, Heathcock L,Hollingsworth F, et al. The transcriptional coactivator TAZ regulatesmesenchymal differentiation in malignant glioma. Genes Dev 2011;25:2594–609.

19. Zhou Z, Hao Y, Liu N, Raptis L, Tsao MS, Yang X. TAZ is a noveloncogene in non-small cell lung cancer. Oncogene 2011;30:2181–6.

20. Xie M, Zhang L, He CS, Hou JH, Lin SX, Hu ZH, et al. Prognosticsignificance of TAZ expression in resected non-small cell lung cancer.J Thorac Oncol 2012;7:799–807.

21. Lin CW, Chang YL, Chang YC, Lin JC, Chen CC, Pan SH, et al.MicroRNA-135b promotes lung cancer metastasis by regulating mul-tiple targets in the Hippo pathway and LZTS1. Nat Commun 2013;4:1877.

22. Botling J, Edlund K, Lohr M, Hellwig B, Holmberg L, Lambe M, et al.Biomarker discovery in non-small cell lung cancer: integrating geneexpression profiling, meta-analysis, and tissue microarray validation.Clin Cancer Res 2013;19:194–204.

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

24. Bild AH, YaoG,Chang JT,WangQ, Potti A, ChasseD, et al. Oncogenicpathway signatures in human cancers as a guide to targeted therapies.Nature 2006;439:353–7.

25. Hou J, Aerts J, den Hamer B, van Ijcken W, den Bakker M, Riegman P,et al. Gene expression-based classification of non-small cell lungcarcinomas and survival prediction. PLoS ONE 2010;5:e10312.

26. Raponi M, Zhang Y, Yu J, Chen G, Lee G, Taylor JM, et al. Geneexpression signatures for predicting prognosis of squamous cell andadenocarcinomas of the lung. Cancer Res 2006;66:7466–72.

27. Okayama H, Kohno T, Ishii Y, Shimada Y, Shiraishi K, Iwakawa R,et al. Identification of genes upregulated in ALK-positive and EGFR/KRAS/ALK-negative lung adenocarcinomas. Cancer Res 2012;72:100–11.

28. Shedden K, Taylor JM, Enkemann SA, Tsao MS, Yeatman TJ, GeraldWL, et al. Gene expression-based survival prediction in lung adeno-carcinoma: a multi-site, blinded validation study. Nat Med 2008;14:822–7.

29. XieY, XiaoG,CoombesKR,BehrensC, Solis LM,RasoG, et al. Robustgene expression signature from formalin-fixed paraffin-embeddedsamples predicts prognosis of non-small-cell lung cancer patients.Clin Cancer Res 2011;17:5705–14.

30. Micke P, Edlund K, Holmberg L, Kultima HG, Mansouri L, Ekman S,et al. Gene copy number aberrations are associated with survival inhistologic subgroups of non-small cell lung cancer. J Thorac Oncol2011;6:1833–40.

31. International HapMap Consortium. The International HapMap Project.Nature 2003;426:789–96.






32. Olshen AB, Venkatraman ES, Lucito R, Wigler M. Circular binarysegmentation for the analysis of array-based DNA copy number data.Biostatistics 2004;5:557–72.

33. Edlund K, Lindskog C, Saito A, Berglund A, Ponten F, Goransson-Kultima H, et al. CD99 is a novel prognostic stromal marker in non-small cell lung cancer. Int J Cancer 2012;131:2264–73.

34. Horie M, Saito A, Mikami Y, OhshimaM, Morishita Y, Nakajima J, et al.Characterization of human lung cancer-associated fibroblasts inthree-dimensional in vitro co-culture model. Biochem Biophys ResCommun 2012;423:158–63.

35. Azzolin L, Zanconato F, Bresolin S, Forcato M, Basso G, Bicciato S,et al. Role of TAZ as mediator of Wnt signaling. Cell 2012;151:1443–56.

36. Berriz GF, Beaver JE, Cenik C, Tasan M, Roth FP. Next generationsoftware for functional trend analysis. Bioinformatics 2009;25:3043–4.

37. 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.

38. Tan P, Planck M, Edlund K, Botling J, Micke P, Isaksson S, et al.Genomic and transcriptional alterations in lung adenocarcinoma inrelation to EGFR and KRAS mutation status. PLoS ONE 2013;8:e78614.

39. Gemma A, Li C, Sugiyama Y, Matsuda K, Seike Y, Kosaihira S, et al.Anticancer drug clustering in lung cancer based on gene expressionprofiles and sensitivity database. BMC Cancer 2006;6:174.

40. Zhou BB, Peyton M, He B, Liu C, Girard L, Caudler E, et al. TargetingADAM-mediated ligand cleavage to inhibit HER3 and EGFR pathwaysin non-small cell lung cancer. Cancer Cell 2006;10:39–50.

41. Amit I, Citri A, Shay T, LuY, KatzM, ZhangF, et al. Amodule of negativefeedback regulators defines growth factor signaling. Nat Genet 2007;39:503–12.

42. Zwang Y, Sas-Chen A, Drier Y, Shay T, Avraham R, Lauriola M, et al.Two phases of mitogenic signaling unveil roles for p53 and EGR1 inelimination of inconsistent growth signals. Mol Cell 2011;42:524–35.

43. Nagashima T, Shimodaira H, Ide K, Nakakuki T, Tani Y, Takahashi K,et al. Quantitative transcriptional control of ErbB receptor signalingundergoes graded to biphasic response for cell differentiation. J BiolChem 2007;282:4045–56.

44. Barretina J, Caponigro G, Stransky N, Venkatesan K,Margolin AA, KimS, et al. The Cancer Cell Line Encyclopedia enables predictive model-ling of anticancer drug sensitivity. Nature 2012;483:603–7.

45. Yamauchi M, Yamaguchi R, Nakata A, Kohno T, Nagasaki M, Shima-mura T, et al. Epidermal growth factor receptor tyrosine kinase definescritical prognostic genes of stage I lung adenocarcinoma. PLoS ONE2012;7:e43923.

46. Cancer Genome Atlas Research Network. Comprehensive genomiccharacterization of squamous cell lung cancers. Nature 2012;489:519–25.

47. Imielinski M, Berger AH, Hammerman PS, Hernandez B, Pugh TJ,Hodis E, et al. Mapping the hallmarks of lung adenocarcinoma withmassively parallel sequencing. Cell 2012;150:1107–20.

48. Fehse B, Roeder I. Insertional mutagenesis and clonal dominance:biological and statistical considerations. Gene Ther 2008;15:143–53.

49. Yang N, Morrison CD, Liu P, Miecznikowski J, Bshara W, Han S, et al.TAZ induces growth factor-independent proliferation through activa-tion of EGFR ligand amphiregulin. Cell Cycle 2012;11:2922–30.

50. Bonomi PD, Buckingham L, Coon J. Selecting patients for treatmentwith epidermal growth factor tyrosine kinase inhibitors. Clin CancerRes 2007;13:s4606–12.

51. Yonesaka K, Zejnullahu K, Lindeman N, Homes AJ, Jackman DM,Zhao F, et al. Autocrine production of amphiregulin predicts sensitivitytobothgefitinib andcetuximab inEGFRwild-typecancers.ClinCancerRes 2008;14:6963–73.


Noguchi et al.




2014;20:4660-4672. Published OnlineFirst June 20, 2014.Clin Cancer Res Satoshi Noguchi, Akira Saito, Masafumi Horie, et al.

Small Cell Lung Cancer−Non An Integrative Analysis of the Tumorigenic Role of TAZ in Human

Updated version

10.1158/1078-0432.CCR-13-3328doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2014/06/23/1078-0432.CCR-13-3328.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/20/17/4660.full#ref-list-1

This article cites 51 articles, 19 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/20/17/4660.full#related-urls

This article has been cited by 4 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/20/17/4660To 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-13-3328

http://clincancerres.aacrjournals.org/content/suppl/2014/06/23/1078-0432.CCR-13-3328.DC1

http://clincancerres.aacrjournals.org/content/20/17/4660.full#ref-list-1

http://clincancerres.aacrjournals.org/content/20/17/4660.full#related-urls

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/20/17/4660


an integrative analysis of the tumorigenic role of taz in ... · biology of human tumors an...

Documents