A Meta-analysis of Lung Cancer Gene Expression Identiﬁes ... · Tumor and Stem Cell Biology A...

Tumor and Stem Cell Biology

A Meta-analysis of Lung Cancer Gene Expression IdentifiesPTK7 as a Survival Gene in Lung Adenocarcinoma

Ron Chen1, Purvesh Khatri2,3, Pawel K. Mazur1, Melanie Polin1, Yanyan Zheng1, Dedeepya Vaka1,ChuongD. Hoang4, Joseph Shrager4, Yue Xu4, Silvestre Vicent1, Atul J. Butte5, and E. Alejandro Sweet-Cordero1

AbstractLung cancer remains the most common cause of cancer-related death worldwide and it continues to lack

effective treatment. The increasingly large and diverse public databases of lung cancer gene expression con-stitute a rich source of candidate oncogenic drivers and therapeutic targets. To define novel targets for lungadenocarcinoma, we conducted a large-scale meta-analysis of genes specifically overexpressed in adeno-carcinoma. We identified an 11-gene signature that was overexpressed consistently in adenocarcinomaspecimens relative to normal lung tissue. Six genes in this signature were specifically overexpressed inadenocarcinoma relative to other subtypes of non–small cell lung cancer (NSCLC). Among these genes wasthe little studied protein tyrosine kinase PTK7. Immunohistochemical analysis confirmed that PTK7 ishighly expressed in primary adenocarcinoma patient samples. RNA interference–mediated attenuation ofPTK7 decreased cell viability and increased apoptosis in a subset of adenocarcinoma cell lines. Further, lossof PTK7 activated the MKK7–JNK stress response pathway and impaired tumor growth in xenotransplan-tation assays. Our work defines PTK7 as a highly and specifically expressed gene in adenocarcinoma and apotential therapeutic target in this subset of NSCLC. Cancer Res; 74(10); 2892–902. �2014 AACR.

IntroductionLung cancer is the leading cause of cancer death in the United

States and worldwide (1). Despite intensive basic and clinicalresearch, the overall 5-year survival rate of the major histologicsubtype, non–small cell lung cancer (NSCLC) has only improvedfrom 14% to 18% since 1975 (1). Recently, targeted treatmentbased on patient-specific molecular aberrations has led tosignificant response rates in subsets of patients with NSCLC(2, 3). However, about half of all patients do not harbor known"driver" mutations and cannot be treated with targeted agents(4). Thus, new approaches for identification of novel regulatorsand potential targets for treatment of lung cancer are needed.

Gene expression analysis has been used to classify cancers,predict clinical outcomes, and discover disease-associated bio-markers. However, gene expression experiments are usuallyanalyzed in isolation and are limited to a small number ofsamples.Meta-analysis approachesmake it possible to combinemultiple gene expression datasets and increase the statisticalpower for gene discovery. Such meta-analysis approaches havebeen successfully used for cancers of the breast (5, 6), prostate(7), liver (8), and lung (9), as well as broadly across cancers (10–12). Statistically, individual studies of gene expression in cancerare limited bybothbiologic (e.g., samplingof a particular patientpopulation) and technical (e.g., only using one expression ana-lysis platform) biases that hinder the broader application oftheir findings and ultimate translation into clinical practice.Meta-analysis can control for such confounding factors byincreasing the statistical power to detect consistent changesacross multiple datasets. Several groups have made availablelarge NSCLC gene expression datasets that consist of tumor-to-normal comparisons (9, 13–21). These datasets represent a largeand as yet not fully tapped resource for discovering novel genesrelevant to the pathogenesis of lung cancer. We reasoned that acareful meta-analysis that combined multiple patient popula-tions across many institutions, platforms, and data procure-ment methods would uncover genes with functional relevanceto lung cancer that may have been otherwise overlooked by theisolated analysis of individual gene expression studies.

We applied a recently proposed meta-analysis approach (22)to 13 gene expression datasets consisting of 2,026 lung samples,which enabled the discovery and validation of commonly over-expressed genes in lung adenocarcinoma, the predominantsubtype of NSCLC. Among themost consistently overexpressed

Authors' Affiliations: 1Cancer Biology Program, Division of Hematology/Oncology, Department of Pediatrics; 2Center for Biomedical InformaticsResearch, Department of Medicine; 3Institute for Immunity, Transplant andInfection; 4Department of Cardiothoracic Surgery; and 5Division of Sys-tems Medicine, Department of Pediatrics, Stanford University School ofMedicine, Stanford, California

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

R. Chen and P. Khatri contributed equally to this work.

Current address for S. Vicent: Division of Oncology, Center for AppliedMedical Research (CIMA), Pamplona, Spain.

Corresponding Authors: E. Alejandro Sweet-Cordero, Stanford Univer-sity, 265 Campus Drive, LLSCRB 2078B, Stanford, CA 94305. Phone: 650-725-5901; Fax: 650-736-0195; E-mail: [email protected]; and Atul J.Butte, Stanford University School of Medicine, 1265Welch Road, Room X-163MS-5415, Stanford CA 94305-5415. Phone: 650-723-3465; Fax: 650-723-7070; Email: [email protected]

doi: 10.1158/0008-5472.CAN-13-2775

�2014 American Association for Cancer Research.

CancerResearch

Cancer Res; 74(10) May 15, 20142892

on November 27, 2020. © 2014 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst March 20, 2014; DOI: 10.1158/0008-5472.CAN-13-2775

http://cancerres.aacrjournals.org/

genes was PTK7, a member of the receptor tyrosine kinasefamily conserved across Hydra, Drosophila, Japanese pufferfish, chicken, and human (23). PTK7 contains seven immuno-globulin (Ig) domains, a transmembrane domain, and a cata-lytically inactive kinase domain within the cytoplasmic tail(24). It was first discovered in melanocytes (24) and subse-quently found to be overexpressed in colon carcinoma (25). InXenopus, PTK7 acts at the level of Frizzled and Dishevelled toregulate the Wnt/planar cell polarity (PCP) pathway (26). Ptk7knockout mice die perinatally and display developmentaldefects of the inner ear and of neural tube closure, consistentwith a role in regulating PCP (27). Overexpression of PTK7 hasbeen described in several cancers, including colon (25), gastric(28), esophageal (29), and acute myelogenous leukemia (AML;ref. 30). However, the precise role of PTK7 in regulatingoncogenesis remains unclear. In colon cancer, PTK7 may playa role in the Wnt/b-catenin pathway (31), and knockdownleads to caspase-10–mediated apoptosis (32).Consistent with the gene expression analysis, we found

consistently elevated expression of PTK7 protein in primaryadenocarcinoma patient samples. Knockdown of PTK7 dem-onstrated that it is essential for the viability of a subset ofNSCLC cell lines, and PTK7 disruption increased Mapk kinasekinase-7 (MKK7)–JNK (c-jun-NH2-kinase) pathway activity.Xenotransplantation studies revealed the requirement of PTK7in tumor growth. These results demonstrate the power of usingpublicly available patient data to uncover oncogenic driversand suggest that PTK7 may represent a novel therapeutictarget in adenocarcinoma.

Materials and MethodsData collection, preprocessing, and normalizationGene expression data for 13 human lung cancer studies

were downloaded from the National Center for BiotechnologyInformation Gene Expression Omnibus (GEO; accession num-bers GSE10072, GSE2514, GSE7670, GSE19188, GSE11969,GSE21933, GSE42127, GSE41271, GSE37745, GSE28571, andGSE20853) and websites http://www.broadinstitute.org/mpr/lung/ and https://array.nci.nih.gov/caarray/project/details.action?project.id¼182. The histologic phenotypes weredefined as in the corresponding original publications. Datasetswere curated to include only normal, adenocarcinoma, squa-mous cell carcinoma (SCC), and large cell carcinoma (LCC)samples. All datasets were normalized individually using Gua-nineCytosine RobustMulti-ArrayAverage (gcRMA; ref. 33). ForPTK7 gene expression, RNA was extracted and prepared usingestablished protocols and hybridized to Affymetrix HumanGene 1.0 ST Array. Raw data are available in GEO (GSE50138).

Meta-analysis of gene expression dataTwo meta-analysis approaches were applied to the normal-

ized data (22). The first approach combines effect sizes fromeach dataset into a meta-effect size to estimate the amount ofchange in expression across all datasets. For each gene in eachdataset, an effect sizewas computed usingHedges adjusted g. Ifmultiple probes mapped to a gene, the effect size for each genewas summarized using the fixed effect inverse-variance model.Next, study-specific effect sizes were combined to obtain the

pooled effect size and its standard error using the randomeffects inverse-variance technique. The z-statistic was com-puted as a ratio of the pooled effect size to its standard error foreach gene, and the result was compared with a standardnormal distribution to obtain a nominal P value. P values werecorrected formultiple hypotheses testing using the Benjamini–Hochberg correction (34).

A second nonparametric meta-analysis that combinesP values from individual experiments to identify genes witha large effect size in all datasets was also used. A t-statistic wascalculated for each gene in each study. After computing one-tail P values for each gene, these were corrected for multiplehypotheses using the Benjamini–Hochberg correction. Next,Fisher sum of logs method (35) was applied. Briefly, thismethod sums the logarithm of corrected P values across alldatasets for each gene, and compares the sum against a c2

distribution with 2k degrees of freedom, in which k is thenumber of datasets used in the analysis.

Leave-one-out validation and classificationTo control for the influence of single large experiments on

the meta-analysis results, leave-one-out meta-analysis wasperformed. One dataset at a time was excluded and bothmeta-analysismethodswere applied to the remaining datasets.We hypothesized that the minimal set of genes that aresignificantly overexpressed, irrespective of the set of datasetsanalyzed, would constitute a robust gene expression signatureof adenocarcinoma across multiple independent cohorts.A very stringent threshold [false discovery rate (FDR) � 1 �10�5] for selecting differentially overexpressed genes in ade-nocarcinoma was used. Furthermore, we analyzed heteroge-neity of the effect sizes across all studies. Geneswith significantheterogeneity (P� 0.05) were removed from the overexpressedgenes identified using stringent FDR criteria. The geometricmean of the remaining significant genes was computed andused to create a univariate binomial linearmodel for classifyinga lung sample as a normal or adenocarcinoma sample, or asadenocarcinoma or SCC sample.

ImmunohistochemistryImmunohistochemistry was performed as previously

described (36) with the following antibodies: rabbit antibodyto phospho-histone H3 (pHH3; 1:500; Upstate), rabbit antibodyto cleaved caspase-3 (CC3; 1:400; Cell Signaling Technology;9664), rabbit antibody to PTK7 (1:1,000; Sigma; SAB3500340).PTK7 staining was performed using pepsin antigen retrieval.

Human primary adenocarcinoma samplesThis study compliedwith federal, state, and local regulations

of the Human Research Protection Program and was approvedby the Stanford Institutional Research Board. Informed con-sent was obtained from all patients included in the study.

Establishment of patient-derived xenograft tumors fromprimary human adenocarcinoma

Surgically removed human NSCLC tumor tissues were keptin ice-cold Hank's Balanced Salt Solution (Life Technologies)until use. Tumors were cut into 1-mm pieces and implanted in

PTK7 Lung Adenocarcinoma Meta-analysis

www.aacrjournals.org Cancer Res; 74(10) May 15, 2014 2893




the subrenal capsules in NOD-SCID-IL2Rg (NSG) mice (TheJackson Laboratory).

Tissue microarrayLung adenocarcinoma tissue microarray with normal lung

tissue, containing 20 cases of lung adenocarcinoma and 10normal lung tissue (BC04119b; US Biomax) was immunohis-tochemically stained for PTK7. Staining was scored as negative(0), weak (1), or strong (2).

Cell cultureAll NSCLC cell linesweremaintained in RPMI supplemented

with 10% FBS and 1% penicillin–streptomycin. Cell linesincluded NCI-H1299, NCI-H2009, NCI-H23, A549, NCI-H441,NCI-H460, NCI-H1792, NCI-H1975, NCI-H2126, NCI-H358, NCI-H727, NCI-H1568, NCI-H1650, and NCI-H2087. Immortalizednormal lung epithelial cell line SALE (37) was cultured inserum-free media SAGM (CC-3118; Lonza).

Short hairpin RNA and virus productionHuman short hairpin RNA (shRNA) constructs against PTK7

were purchased from OpenBiosystems. The human PTK7shRNA set was cat # RHS4533-NM_002821. Of this set,TRCN0000006434 and TRCN0000006435 were labeled asshPTK7–1 and 2, respectively. Control hairpins against GFPand luciferase in the pLKO.1 backbone were used (see Sup-plementary Methods for target sequences). Transfection-qual-ity DNA was extracted using Qiagen DNA kits. Lentivirus wasproduced by transfection into 293FT cells as previouslydescribed (38), filtered, and applied directly to cells. Puromycinselection was started 2 days after lentiviral infection for aduration of 2 days at 2 mg/mL.

Quantitative real-time PCR analysisRNA was isolated 5 days after lentiviral infection and puro-

mycin selection with TRIzol reagent (Invitrogen) following themanufacturer's protocol. cDNA was synthesized with a DyNA-mo cDNA Synthesis Kit (F470; New England Biolabs), andquantitative real-time PCR (qRT-PCR) was carried out intriplicate by SYBR Green (Quanta Biosciences) using aC1000 Thermal Cycler (BioRad). See Supplementary Methodsfor primer sequences.

Cell proliferation assayCells were trypsinized and plated in triplicate into 96-well

plates (day 0). Cell viability wasmeasured by treating cells withMTT. Absorbance measurements were recorded using a Spec-traMax 340 (Molecular Devices) at 570 nm on the indicateddays (Cell Proliferation Kit I; Roche). All data were normalizedto experimental day 0.

Fluorescence-activated cell sorting analysisCells were scraped from the plate and single cell suspensions

were made by passing cells through 28G1/2 insulin syringes(BD) 10 times. Cells were then washed in PBS and resuspendedin PF10 (10% FBS in PBS), stained at 4�C for 30 minutes in thedark with rabbit anti-PTK7 (GTX104510; GeneTex) or IgG2acontrol (Clone 188B isotype:mouse, BD). For Annexin-V stain-

ing, cell lines were plated at 5 � 105 cells per 6-cm plates andstained with Annexin-V (FITC-Annexin-V; 556419; BD Bio-sciences) and propidium iodide (50 pl/mL final concentration)following themanufacturer's protocol at indicated time points.A C6 Flow Cytometer (Accuri) was used for fluorescence-activated cell sorting (FACS) analysis and the manufacturer'ssoftware and FlowJo v5 for analysis.

ImmunoblottingCells were scraped and lysed in radioimmunoprecipita-

tion assay buffer with protease inhibitor cocktail (Roche),25 mmol/L sodium fluoride, 1 mmol/L phenylmethylsulfo-nylfluoride, and 1 mmol/L sodium orthovanadate (Sigma-Aldrich). Protein samples were resolved by SDS-PAGE, trans-ferred to Amersham Hybond-P membranes (GE Healthcare),and blocking buffer (5% bovine serum albumin in TBS-T) for1 hour before addition of primary antibody, which wasapplied overnight at 4�C. The following antibodies from CellSignaling Technology were used (1:1,000 dilution unlessotherwise noted): rabbit anti-Cleaved PARP (#5625), mouseanti-pERK (# 9106), rabbit anti-ERK (#4695), rabbit anti-pAKT (T308; #9275; 1:500), rabbit anti-AKT (#9272), rabbitanti-pMKK7 (S271/T275; #4171), rabbit anti-MKK7 (#4172),rabbit anti-pJNK (T183/Y185; #9251), rabbit anti-JNK (#9252), rabbit anti-pJUN (S73; # 9164), rabbit anti-p-P38(T180/Y182; #9211), and rabbit anti-P38 (#9212). Mouseanti-b-actin (Sigma-Aldrich, clone 1A4; 1:10,000) was usedas a loading control.

XenograftCell lines infected with specific shRNA were resuspended in

serum-free RPMI and injected subcutaneously at 1million cellsper tumor into the two lower flanks of nudemice (The JacksonLaboratory). Between 4 and 8 tumors were injected for eachcondition. One week after injection, tumor dimensions weremeasured approximately every 3 days and tumor volume wascalculated using the formula p/6 � [(lengthþwidth)/2] (39).All animal experiments were approved by the Stanford Uni-versity School of Medicine Committee on Animal Care.

Statistical analysesUnpaired two-tailed t tests were used for comparisons

between different groups. Error bars correspond to SEM.Significant P values in the text correspond to <0.05 (�),<0.01 (��), or <0.001 (��).

ResultsMeta-analysis of five adenocarcinoma gene expressiondatasets identifies 11 significantly overexpressed genesin adenocarcinoma

We identified and manually curated seven published andpublicly available adenocarcinoma datasets (9, 13–15, 17–19).Only datasets with accessible unprocessed raw data and con-taining both normal and adenocarcinoma samples were usedfor further analysis. We identified five datasets that met thesecriteria (743 adenocarcinoma, 127 normal; SupplementaryTable S1; refs. 9, 14, 17–19). We applied a recently proposedmeta-analysis approach (22) to these five datasets as outlined

Chen et al.

Cancer Res; 74(10) May 15, 2014 Cancer Research2894




in (Fig. 1A). Because an experiment with a large number ofsamples can have significant influence on meta-analysisresults, we performed a leave-one-out analysis, in which weexcluded one dataset at a time and applied our meta-analysisapproach on the remaining datasets (see Materials and Meth-ods). We hypothesized that the minimal set of genes identifiedas significantly overexpressed by bothmethods, irrespective of

the datasets used for analysis, would constitute a robust geneexpression signature of adenocarcinoma across multiple inde-pendent cohorts.

Because of the heterogeneity of gene expression within andamong cancer samples and datasets, we used a very stringentthreshold (FDR� 1� 10�5) for bothmeta-analysis methods toselect overexpressed genes in adenocarcinoma. This approach

Download 870 ADC

and normal samples in

5 data sets from GEO

Manual data curation

Data normalization(gcRMA, LOESS)

Probe-to-gene mapping

Meta-analysis bycombining effect size

Meta-analysis bycombining P values

meta-analysisLeave-one-out

Discovery

Curation of public

genomic data

Preliminary candidate genes

(11 genes)

In vitro validation

in ADC cell linesIn vivo validation

Validation in 3 data setsconsisting of 237 samples

PTK7

Validation

Specificity to ADC vs. other NSCLC subtypes

in 6 data sets consisting of 932 samples

(11 genes)

(7 genes)

Validated candidate genes

Validation by IHC in

patient derived xenografts

and tissue microarray

consisting of 32 samples

FDR < 0.00001%

B

ADCNormal

Data

preprocessing

In s

ilico

bio

info

rma

tics d

ata

an

aly

sis

Bio

log

ic v

alid

atio

n

A

FDR < 0.00001

FDR < 0.00001%

SULT1C2

SLC2A5

RUNX1

PTK7

MPZL1

KCNK1

EPHB2

COL5A2

CACNB3

BLNK

ARHGEF16

Su et al.

–4 –2 0 2 4

Stearman et al.

Landi et al.

SULT1C2

SLC2A5

RUNX1

PTK7

MPZL1

KCNK1

EPHB2

COL5A2

CACNB3

BLNK

ARHGEF16

Shedden et al.

Bhattacharjee et al.

SULT1C2

SLC2A5

RUNX1

PTK7

MPZL1

KCNK1

EPHB2

COL5A2

CACNB3

BLNK

ARHGEF16

–4 0 2 4–2

–5 0 5 –5 0 5

–4 –2 0 2 4

Figure 1. Meta-analysis of NSCLC gene expression datasets identified 11 significantly overexpressed genes in NSCLC. A, meta-analysis and functionalvalidation pipeline. In silico bioinformatics data analysis workflow consisted of the curation of five publicly available datasets, data preprocessing,discovery meta-analysis, independent validation, and subtype analysis followed by validation in primary human patient samples and in vitro andin vivo functional validation. B, the 11 genes cluster adenocarcinoma (ADC) versus normal across each dataset. Each column is a sample and each rowis the expression level of a gene. The color scale represents the raw Z-score ranging from blue (low expression) to yellow (high expression).Dendrograms above each heatmap correspond to the Pearson correlation-based hierarchical clustering of samples by expression of the 11 genes.






identified 14 genes significantly and highly overexpressed inadenocarcinoma across all five datasets. In addition, we esti-mated between-studies heterogeneity in effect sizes. Threegenes (ARHGAP8, CEACAM1, and DDR1) had significantbetween-studies heterogeneity (P � 0.05) and were removedfrom the further analysis because we did not have sufficientinformation available to further explore heterogeneity in thesegenes (Supplementary Table S1 and Supplementary Fig. S1).As expected, hierarchical clustering of the five discovery data-sets using the remaining 11 genes distinguished normallung from adenocarcinoma samples (Fig. 1B). We furthervalidated the 11 genes in three additional independent datasets(16, 20, 21; 146 adenocarcinoma, 91 normal; SupplementaryTable S2). All genes were significantly overexpressed in thevalidation datasets at FDR� 5%with at least onemeta-analysismethod, and all except three (CACNB3, SLC2A5, and SULT1C2)were significantly overexpressed as assessed by both methods(Supplementary Fig. S2 and Supplementary Table S3). Geo-metric mean of the 11 overexpressed genes was significantlyhigher for adenocarcinoma compared with normal samples ineach of the three validation datasets (Fig. 2A). A univariate

linear regressionmodel using the geometricmean had the areaunder receiver operating characteristic (ROC) curve rangingfrom 0.809 to 0.993 (Fig. 2B). However, because of a very smallnumber of normal samples in the Takeuchi and colleaguesdataset, high area under the curve (AUC) value should beinterpreted with caution. Thus, publicly available datasets canconsistently identify a small set of genes as overexpressed inhuman adenocarcinoma. However, as only adenocarcinomasamples were used in this analysis, it is possible that these 11genes are upregulated in other cancers compared with normaltissue. Therefore, we sought to determine whether the 11-genesignature was specific for adenocarcinoma or whether it wasalso upregulated in other NSCLC subtypes.

Six of the 11 genes are overexpressed in adenocarcinomacompared with other NSCLC subtypes

We analyzed an additional eight gene expression datasetsconsisting of 1,040 human NSCLC samples (16, 20, 40–43;687 adenocarcinoma, 353 SCC; Supplementary Table S4). Sixof the 11 genes were significantly overexpressed in adenocar-cinoma compared with SCC, whereas two genes were

4.0

4.5

5.0

5.5

6.0

Normal ADC

Geom

etr

ic m

ean

Normal

ADC

Hou et al.A

B

0.00

0.25

0.50

0.75

1.00

0.00 0.25 0.50 0.75 1.00

False positive rate (1−specificity)

Tru

e p

ositiv

e r

ate

(S

ensitiv

ity)

AUC = 0.94 (95% CI, 0.96 − 0.91)

Hou et al.

0.2

0.4

0.6

Normal ADC

P = 4.8E–04

Geom

etr

ic m

ean

Takeuchi et al.

0.00

0.25

0.50

0.75

1.00

0.00 0.25 0.50 0.75 1.00


Tru

e p

ositiv

e r

ate

(S

ensitiv

ity)

AUC = 0.97 (95% CI, 0.99 − 0.95)

Takeuchi et al.

3

4

5

6

7

8

Normal ADC

P = 3.64E–03

Geom

etr

ic m

ean

Lo et al.

0.00

0.25

0.50

0.75

1.00

0.00 0.25 0.50 0.75 1.00


Tru

e p

ositiv

e r

ate

(S

ensitiv

ity)

AUC = 0.81 (95% CI, 0.9 − 0.73)

Lo et al.

P = 1.05E–14Normal

ADC

Normal

ADC

Figure 2. Geometric mean of the 11 overexpressed genes were significantly higher in adenocarcinoma (ADC) samples than normal samples in threeindependent data sets. A, violin plots summarizing the geometric mean of the 11-gene signature in normal and adenocarcinoma. Each dot is asample, and black horizontal line is the mean of geometric mean in a given sample group. B, performance of a univariate classification model forpredicting adenocarcinoma versus normal based on gene expression of the 11-gene signature across the independent datasets measured byROC curves.

Chen et al.





significantly underexpressed in adenocarcinoma comparedwith SCC (Supplementary Table S5). One of the six overex-pressed gene, SULT1C2, had significant between-studies het-erogeneity (P � 0.05) in the eight datasets. Therefore, it wasremoved from further analysis. Geometricmean of the remain-ing five overexpressed genes was significantly higher for ade-nocarcinoma comparedwith SCC in seven of the eight datasets(Supplementary Fig. S3A). The geometric mean of the 5-genesignature in adenocarcinomawas also significantly higher thanthat in LCC in the two additional datasets (46 LCC; Supple-mentary Fig. S3A). A univariate linear regression model usingthe geometric mean had the area under ROC curve rangingfrom 0.603 to 0.766 for discriminating between adenocarcino-ma and other NSCLC subtypes (Supplementary Fig. S3B).In summary, a meta-analysis of 13 independent NSCLC

datasets consisting of 2,026 human lung samples (1,430adenocarcinoma, 353 SCC, 46 LCC, and 197 normal) iden-tified five genes (ARHGEF16, CACNB3, MPZL1, PTK7, andRUNX1) that are significantly and consistently overexpressedin adenocarcinoma across all datasets, and are able todistinguish adenocarcinoma samples from normal lungsamples or other NSCLC subtypes. Consistent overexpres-sion of these genes despite the potential presence of biologicand technological confounding factors, including differentcohorts, treatments, and microarray technologies stronglysuggests that at least some of these genes may be function-ally important in adenocarcinoma and may play a role in thepathogenesis of this disease.

The receptor tyrosine kinase, PTK7, is overexpressed inadenocarcinomaPTK7 belongs to the receptor tyrosine kinase family. It is

located on chromosomal locus 6p21, which iswithin the top 4%of all regions of most frequent copy number gain in anindependent lung cancer dataset (ref. 44, analyzed usingOncomine; ref. 45, data not shown). In addition, 6p21 has beenassociated with poor survival in NSCLC (46). PTK7 was sig-

nificantly overexpressed in NSCLC samples compared withnormal samples irrespective of subtypes in all discovery (Sup-plementary Fig. S1) and validation datasets (SupplementaryFig. S2). Furthermore, PTK7 is specifically overexpressed inadenocarcinoma compared with SCC (Supplementary Fig.S3C). Therefore, we focused on further investigations on PTK7.

Overexpression of PTK7 in primary human and mouselung adenocarcinoma specimens

Increased expression of PTK7 (Fig. 3A) was assessed byimmunohistochemical staining. First, two patient-derivedxenografts of human lung adenocarcinoma were analyzed andnoted to stain positively for PTK7 to varying intensities(Fig. 3B). Staining patterns were heterogeneous, with bothnuclear and cytoplasmic staining present within each speci-men. Next, a tissue microarray consisting of human samplesacross various stages and grades of adenocarcinoma was usedto determine PTK7 expression. Fourteen of the 20 adenocar-cinoma samples stained positively for PTK7, of which fivestained strongly and nine stained weakly to moderately(Fig. 3C; Supplementary Fig. S4A and Supplementary TableS6). In contrast, all normal lung specimens were negative forepithelial staining of PTK7 (Supplementary Fig. S4A).We foundno correlation between PTK7 staining and stage or grade,although this may be due to limited sample size. Notably,samples with weak staining tended to have more cytoplasmicrather than nuclear PTK7 (Supplementary Table S6). Thus,PTK7 seems to be overexpressed in most human patientprimary adenocarcinoma samples and a significant subsetdemonstrates high levels of staining.

Consistent with these results, a well-established mousemodel of adenocarcinoma (47) also exhibited increased mRNAtranscript levels of Ptk7 (Supplementary Fig. S4C) and positivePTK7 protein staining by immunohistochemistry (Supplemen-tary Fig. S4D) relative to normal lung. These results suggestthat PTK7 is overexpressed in both human and mouse ade-nocarcinoma specimens.

APTK7

C Normal Grade 2 Grade 2–3Grade 1

Standardized mean difference (log2 scale)

−1.0 −0.5 0.0 0.5 1.0 1.5 2.0

Bhattacharjee et al.

Hou et al.

Landi et al.Shedden et al.

Stearman et al.Su et al.Takeuchi et al.

Summary

Lo et al.

BPDX 1 PDX 2Figure 3. PTK7 is overexpressed in

human adenocarcinoma. A, forestplot of PTK7 expression across alldiscovery and validation meta-analysis datasets. The x-axis is thestandardized mean differencebetween adenocarcinoma (ADC)and normal on a log2 scale. Thus, avalue of 1 signifies a 2-folddifference in gene expressionbetween cancer and normal. B,representative images of PTK7expression in two patient-derivedxenograft (PDX) specimens. C,representative images of PTK7staining of normal lung andadenocarcinoma spanning grades1 to 3. Scale bars, 50 mm.






Loss of PTK7 in lung cancer cell lines decreases cellviability and induces apoptosis

To determine whether PTK7 could play a functional role inthe maintenance of adenocarcinoma, two shRNAs directedagainst PTK7 were used to decrease expression in a panel oflung cancer cell lines (Supplementary Fig. S5A). Importantly,PTK7 knockdown had no effect on normal lung epithelial cells(Fig. 4A). However, PTK7 depletion using two independentshRNAs led to a significant decrease in viability in a subset ofNSCLC cell lines (Fig. 4A, Supplementary Table S7). The

dependency on PTK7 was not determined by its basal expres-sion (Supplementary Fig. S5B; P ¼ 0.51) or mutations ofany known cancer genes (Supplementary Fig. S5C). However,cell lines with chromosomal gains of the locus harboringPTK7 tended to be more sensitive to PTK7 knockdown (P ¼0.047; Fig. 4B).

To further delineate how PTK7 affects cell viability,we chose the three cell lines most sensitive to PTK7 knock-down (H1299, H2009, and H23) for further analysis. Sub-stantial PTK7 knockdown was validated at the RNA level by

510

310

0.0

0.5

1.0

1.5

2.0

Rela

tive

via

bili

ty

A B

0 2 4 6

0.0

0.5

1.0

1.5

H1299H2009

R = 0.67

H23

PTK7 copy number

shGFP shPTK7-1 shPTK7-2


Co

unt

(% o

f m

ax)

H1299

H2009

C D

PE

Co

unt

(% o

f m

ax)

PE


Count

(% o

f m

ax)

H23

PE

H1299

H2009

H23

sh

PT

K7

-1

sh

PT

K7

-2

sh

GF

P

ΔPARP

β-Actin

sh

GF

P

sh

PT

K7

-1

sh

PT

K7

-2

sh

GF

P

sh

PT

K7

-1

sh

PT

K7

-2

β-Actin

β-Actin

ΔPARP

ΔPARP

E



H1299

H2009

PI

Annexin-V

0.5

1.3

1.4

4.8

11.6

4.3

2.5

3.6

6.0

3.7

8.1

4.0


H23

4.9

0.6

9.4

2.1

18.2

3.2

H12

99

H20

09H23

H35

8

H72

7

H16

50

H20

87

H46

0

H17

92

H15

68

H44

1

H19

75

A549

H21

26

Nor

mal lu

ng

epith

elial c

ells

PT

K7 s

ensitiv

ity

(via

bili

ty a

fter

PT

K7 K

D)shPTK7-1

shPTK7-2

PTK7

IgG2a

510

310

510

310

510

310

PTK7IgG2a

510

310

510

310

IgG2aPTK7

510

310 10

310

510

310

PI

Annexin-V

PI

Annexin-V

5

Figure 4. Disruption of PTK7 in three NSCLC cell lines leads to a decrease cell viability and increase in apoptosis. A, knockdown of PTK7with two independenthairpins across a panel of NSCLC cell lines and assessment of cell viability byMTT. B, scatter plot of PTK7 sensitivity versus copy number in the panel of celllines. PTK7 sensitivitywas calculated by averaging the relative viability across both hairpins after PTK7 knockdown. Copy number at the 6p21PTK7 locuswasqueried in the Sanger Catalogue of Somatic Mutations in Cancer (COSMIC) database. R, the linear correlation coefficient. C, PTK7 knockdown validation byFACS. D, FACS analysis of Annexin-V–positive cells. E, cleaved DPARP immunoblotting and b-actin loading control.

Chen et al.





qRT-PCR (Supplementary Fig. S5D), coinciding with adecrease in protein expression (Fig. 4C). Knockdown ofPTK7 in all three cell lines led to an increase in apoptosisas detected by Annexin-V staining (Fig. 4D), as well as bycleaved PARP immunoblotting (Fig. 4E). Thus, knockdownof PTK7 in several adenocarcinoma cell lines in vitro resultsin decreased cell viability and increased apoptosis where-as normal lung epithelial cells are unaffected. These resultssuggest that PTK7 may be involved in signaling cascadesthat regulate survival of a subset of adenocarcinoma.

Disrupted PTK7 expression activates MKK7–JNKsignalingDuring normal development PTK7 is thought to play a role

in mediating planar cell polarity (26, 27). However, it remainsunclear why disruption of this signaling pathway is importantin cancer cells as it is not evident why planar cell polaritymechanisms would be involved in oncogenesis. To understandthe signaling pathways that are perturbed upon PTK7 disrup-tion, we surveyed the extracellular signal–regulated kinase(ERK), AKT, JNK, and p38MAPK pathways by immunoblotting(Fig. 5). We found no significant change in ERK and AKTsignaling, and no consistent change in p38 phosphorylationstatus across the three cell lines. However, phosphorylation ofMKK7 at Ser171/Thr275, two residues crucial for MKK7 activ-ity, was clearly increased after PTK7 knockdown. Consistentwith the activation of MKK7, phosphorylation of JNK, a down-stream target ofMKK7, was also increased in response to PTK7knockdown in the three cell lines tested with two differentshRNAs. Activation of MKK7 was also observed when pooledsiRNAs were used to deplete PTK7 expression across all three

cell lines (Supplementary Fig. S6A–S6F), arguing against an off-target effect.

Upon phosphorylation by MKK7, JNK translocates into thenucleus, leading to the heterodimerization of c-jun with c-fosto form the AP-1 transcriptional complex. Gene expressionprofiling following PTK7 knockdown in the H1299 and H2009cell lines (Supplementary Fig. S6G) revealed an overrepresen-tation of AP-1 binding sites associated with the differentiallyexpressed genes (Supplementary Fig. S6H). Thus, in someadenocarcinoma cell lines, loss of PTK7 leads to dysregulationof MKK7–JNK and possibly other signaling pathways thatultimately lead to decreased proliferation and increasedapoptosis.

Xenotransplantation of lung cancer cells depleted ofPTK7 reduces tumor burden

The above studies demonstrate that PTK7 is required for cellviability in a subset of adenocarcinoma cell lines in vitro. To testwhether PTK7 is important for tumor maintenance in vivo, weexamined the effect of PTK7 depletion on xenograft adeno-carcinoma tumor growth. Xenotransplantation of H1299(Fig. 6A and B), H2009 (Fig. 6C and D), and H23 (Fig. 6E andF) cell lines infected with shRNA hairpins against PTK7 intoimmunocompromised mice resulted in a significant decreasein tumor growth relative to control shRNA-infected cells. PTK7loss was confirmed by immunohistochemistry for the twoPTK7-specific hairpins relative to control (Fig. 6G). The controlexperiment of infecting normal lung epithelial cell lines wasnot performed, as these cell lines do not readily grow in vivo.PTK7 depletion led to a significant decrease in pHH3, amarker of mitosis (Fig. 6G and H), and a concomitantincrease in CC3, a marker of apoptosis (Fig. 6G and I).Taken together, these studies suggest that PTK7 expressionmay be required for tumor maintenance in vivo for at least asubset of adenocarcinoma.

DiscussionThe goal of this study was to identify candidate driver genes

in lung adenocarcinoma through a meta-analysis of NSCLCgene expression data and to functionally validate their biologicsignificance. The meta-analysis method described here wasdesigned to account for biases inherent in single studies andnominate commonly overexpressed genes. The explosion oflarge amounts of gene expression data from many differentplatforms inmultiple independent cohorts has enabled the useof in silico tests across numerous patient populations, manymore so than a clinical trial could accomplish. With theincreasingly widespread availability of gene expression data,a paradigm shift toward inclusion of a "pre-validation" meta-analysis step in biologic studies may accelerate translationalresearch.

PTK7 seems to have a context-specific value as a prognosticmarker. In gastric cancers, it is a favorable marker of differ-entiated cancers (28), but in AML and esophageal cancer, PTK7is a poor prognostic marker associated with reduced disease-free survival (29, 30). We were unable to find an associationbetween high PTK7 expression and patient survival in NSCLC

pAKT (T308)

pERK

AKT

ERK

pJNK

p-p38

p38

pMKK7

pJUN

β-Actin

H1299

shG

FP

shP

TK

7-1

shP

TK

7-2

shG

FP

shP

TK

7-1

shP

TK

7-2

H2009 H23

shG

FP

shP

TK

7-1

shP

TK

7-2

JNK

MKK7

Figure 5. PTK7 regulates MKK7–JNK signaling. Assessment of the ERK,AKT, JNK, and p38 signaling pathways by immunoblotting in threeNSCLC cell lines.






gene expression datasets with clinical annotations (data notshown), although one report suggested a prognostic signifi-cance in NSCLC (48).

Functional studies in vitro identified several NSCLC cell linesthat were highly sensitive to depletion of PTK7. Reducedexpression of PTK7 led to increased cell death both in vitroand in vivo. This increase in cell death was accompanied byincreased signaling through the JNK pathway. However, JNKinhibition alone did not rescue cell viability after PTK7 loss(data not shown), suggesting that PTK7 loss alters othersignaling pathways and that JNK activation, while present, isnot sufficient to induce apoptosis. Further studies will beneeded to identify other signaling pathways downstream ofPTK7. Importantly, in NSCLC, PTK7 loss did not consistentlyalter signaling through the ERK, AKT, or p38MAPK pathways.

Approximately 10% of the human kinome consists of pro-teins classified as pseudokinases on the basis of alterations inresidues that are thought to be critical to kinase catalyticactivity. The pseudokinase PTK7 is most closely related toactive tyrosine kinases such as Ros, ALK, and LTK (49). Despitethe absence of a catalytic domain, increasing evidence suggestsa key role for pseudokinases in many biologic processes (50).The best-described pseudokinase with an oncogenic role isHER3, which is a member of the EGF receptor family. Afterligand binding, HER3 heterodimerizes with HER2, resulting inthe interaction of the pseudokinase domain of HER3 and theactive kinase domain of HER2, leading to the autophosphor-ylation of HER2 and activation of downstream phosphoinosi-tide 3-kinase and ERK signaling pathways (51, 52). Pseudoki-nases can also function as scaffold proteins. For example, KSRproteins interact with protein kinases of themitogen-activatedprotein kinase pathway to localize the signaling components to

the membrane (53). In contrast with HER3, no known ligandsare known to bind PTK7 in the human setting, and it is unclearwhether PTK7 plays a role as a signal modulator, scaffoldprotein, or another role.

In vertebrates, the PCP pathway regulates convergent exten-sion movements and neural tube closure. PCP also is involvedin regulating orientation of stereociliary bundles of sensoryhair cells in the inner ear. Studies of Ptk7 C-terminal knockoutmice firmly establish this pseudokinase as a regulator of PCP inmammals (27). The precise signaling role of PTK7 in develop-ment is not clear, although the drosophila homologue off-track(otk) may act as a coreceptor for plexins tomediate semaphor-ing signaling (54). Most recently, PTK7/OTK has been sug-gested to act by inhibiting canonical Wnt signaling (55).

Although loss of PTK7 has been associatedwith activation ofWNT in Xenopus, we did not detect consistent activation ofWNT signaling in the absence of PTK7 (data not shown). Thus,the effect of PTK7 loss on lung cancer pathogenesis is likely tobe independent of the regulation of WNT signaling. In sum-mary, the data presented here describe PTK7 as a functionallyrelevant protein in NSCLC.We demonstrate that its disruptionin vitro and in vivo decreases tumor cell growth. These obser-vations provide a framework for further studies to characterizePTK7 as a potentially therapeutic target in adenocarcinoma.

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

Authors' ContributionsConception and design: R. Chen, P. Khatri, A.J. Butte, E.A. Sweet-CorderoDevelopment of methodology: P. Khatri, M. Polin, J. Shrager, A.J. Butte,E.A. Sweet-Cordero

A B

0 5 10 15 20 25 300

500

1,000

1,500

2,000

H1299

H2009C D

shLUCshPTK7-1shPTK7-2

*****

shLUC

shPTK7-1

shPTK7-2

shLUC

shPTK7-1

shPTK7-2

H1299

H2009

G

shP

TK

7-2

shLU

CshP

TK

7-1

CC3pHH3

Days

Tum

or

volu

me (

mm

3)

Tum

or

volu

me (

mm

3)

Tum

or

volu

me (

mm

3)

0

20

40

60*

*

pH

H3-p

ositiv

e

cells

/F.O

.V.

shPTK7-

2

shLU

C

shPTK7-

1

H IH23

E F

shLUC

shPTK7-1

shPTK7-2

H23shLUCshPTK7-1shPTK7-2

0 5 10 15 20 25 30 35 400

100

200

300

400shLUCshPTK7-1shPTK7-2

Days

*****

***

1 cm

1 cm

1 cm

shLUCshPTK7-1shPTK7-2

010 20 30 40 50 60 70 800

50

100

150

PTK7

0

50

100

150*

* *

CC

3-p

ositiv

e

cells

/F.O

.V.

shPTK7-

2

shLU

C

shPTK7-

1

Days

(no growth)

Figure 6. PTK7 depletion in threeNSCLC cell lines decreases tumorburden in a xenotransplantationassay. A, tumor volume over timeof xenografted H1299 with controlshRNA against luciferase (shLUC),or two independent hairpinsagainst PTK7. Error bars representSEM and significant P valuescorrespond to <0.01 (��) or <0.001(��) when compared with theshLUC control. Values are mean�SEM (n¼4–8). B, representative exvivo images of H1299 tumors. Cand D, same figures for H2009 cellline. E and F, same figures for H23cell line. G, staining of PTK7,pHH3, and CC3 byimmunohistochemistry in theH1299 xenografts. H,quantification of pHH3-positivecells per field of view (n ¼ 6). I,quantification of CC3-positivecells per field of view (n ¼ 6).

Chen et al.





Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): R. Chen, P. Khatri, P.K.Mazur,M. Polin, Y. Zheng, C.D.Hoang, J. Shrager, Y. Xu, E.A. Sweet-CorderoAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): R. Chen, P. Khatri, M. Polin, D. Vaka, S. Vicent,E.A. Sweet-CorderoWriting, review, and/or revision of the manuscript: R. Chen, P. Khatri,Y. Zheng, J. Shrager, S. Vicent, A.J. Butte, E.A. Sweet-CorderoAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): P. Khatri, J. Shrager, S. VicentStudy supervision: R. Chen, P. Khatri, A.J. Butte, E.A. Sweet-Cordero

AcknowledgmentsThe authors thank Julien Sage, Laura Attardi, Steve Artandi, Monte Winslow,

and all members of the Sweet-Cordero laboratory for helpful discussions.

Grant SupportR. Chen was funded by the Stanford Graduate Fellowship and Cancer

Biology Training Grant. P. Khatri, A. Butte, and E.A. Sweet-Cordero werefunded by the US National Cancer Institute (R01 CA138256). P.K. Mazur andS. Vicent were funded by the Tobacco-Related Disease Research Program ofCalifornia (TRDRP). P.K. Mazur was also funded by the Stanford UniversityDean's Fellowship and the Child Health Research Institute and LucilePackard Foundation for Children's Health Postdoctoral Fellowship. Y. Zhengwas funded by the American Lung Association.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received September 25, 2013; revised January 21, 2014; accepted February 27,2014; published OnlineFirst March 20, 2014.

References1. Howlader N, Noone AM, Krapcho M, Neyman N, Aminou R, Altekruse

SF, et al. (eds). SEER cancer statistics review, 1975–2009 (Vintage2009 Populations). Bethesda, MD: National Cancer Institute [cited2012 Aug 13]. Available from: http://seer.cancer.gov/csr/1975_2009_pops09/index.html.

2. Keedy VL, Temin S, Somerfield MR, Beasley MB, Johnson DH,McShane LM, et al. American Society of Clinical Oncology provisionalclinical opinion: epidermal growth factor receptor (EGFR) mutationtesting for patients with advanced non-small-cell lung cancer consid-ering first-line EGFR tyrosine kinase inhibitor therapy. J Clin Oncol2011;29:2121–7.

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

4. PaoW,HutchinsonKE. Chipping away at the lung cancer genome. NatMed 2012;18:349–51.

5. Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, et al.Repeated observation of breast tumor subtypes in independentgene expression data sets. Proc Natl Acad Sci U S A 2003;100:8418–23.

6. West M, Blanchette C, Dressman H, Huang E, Ishida S, Spang R,et al. Predicting the clinical status of human breast cancer by usinggene expression profiles. Proc Natl Acad Sci U S A 2001;98:11462–7.

7. Rhodes DR, Barrette TR, Rubin MA, Ghosh D, Chinnaiyan AM. Meta-analysis of microarrays: interstudy validation of gene expressionprofiles reveals pathway dysregulation in prostate cancer. Cancer Res2002;62:4427–33.

8. Hoshida Y, Nijman SM, Kobayashi M, Chan JA, Brunet JP, Chiang DY,et al. Integrative transcriptome analysis reveals common molecularsubclasses of human hepatocellular carcinoma. Cancer Res 2009;69:7385–92.

9. 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.

10. Ramaswamy S, Ross KN, Lander ES, Golub TR. Amolecular signatureof metastasis in primary solid tumors. Nat Genet 2003;33:49–54.

11. Rhodes DR, Yu J, Shanker K, Deshpande N, Varambally R, Ghosh D,et al. Large-scale meta-analysis of cancer microarray data identifiescommon transcriptional profiles of neoplastic transformation andprogression. Proc Natl Acad Sci U S A 2004;101:9309–14.

12. Segal E, Friedman N, Koller D, Regev A. A module map showingconditional activity of expression modules in cancer. Nat Genet2004;36:1090–8.

13. Beer DG, Kardia SL, Huang CC, Giordano TJ, Levin AM, Misek DE,et al. Gene-expression profiles predict survival of patients with lungadenocarcinoma. Nat Med 2002;8:816–24.

14. Bhattacharjee A, RichardsWG, Staunton J, Li C, Monti S, Vasa P, et al.Classification of human lung carcinomas by mRNA expression profil-ing reveals distinct adenocarcinoma subclasses. Proc Natl Acad Sci US A 2001;98:13790–5.

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

16. 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.

17. Landi MT, Dracheva T, Rotunno M, Figueroa JD, Liu H, Dasgupta A,et al. Gene expression signature of cigarette smoking and its role inlung adenocarcinoma development and survival. PloS ONE 2008;3:e1651.

18. Stearman RS, Dwyer-Nield L, Zerbe L, Blaine SA, Chan Z, Bunn PA Jr,et al. Analysis of orthologous gene expression between human pul-monary adenocarcinoma and a carcinogen-induced murine model.Am J Pathol 2005;167:1763–75.

19. Su LJ, Chang CW, Wu YC, Chen KC, Lin CJ, Liang SC, et al. Selectionof DDX5 as a novel internal control for Q-RT-PCR frommicroarray datausing a block bootstrap re-sampling scheme. BMC Genomics 2007;8:140.

20. Takeuchi T, Tomida S, Yatabe Y, Kosaka T, Osada H, Yanagisawa K,et al. Expression profile-defined classification of lung adenocarcinomashows close relationship with underlying major genetic changes andclinicopathologic behaviors. J Clin Oncol 2006;24:1679–88.

21. Lo FY, Chang JW, Chang IS, Chen YJ, Hsu HS, Huang SF, et al. Thedatabaseof chromosome imbalance regions andgenes resided in lungcancer from Asian and Caucasian identified by array-comparativegenomic hybridization. BMC Cancer 2012;12:235.

22. Khatri P, Roedder S, Kimura N, De Vusser K,Morgan AA, Gong Y, et al.A common rejection module (CRM) for acute rejection across multipleorgans identifies novel therapeutics for organ transplantation. J ExpMed 2013;210:2205–21.

23. Grassot J, Gouy M, Perriere G, Mouchiroud G. Origin and molecularevolution of receptor tyrosine kinases with immunoglobulin-likedomains. Mol Biol Evol 2006;23:1232–41.

24. Park SK, Lee HS, Lee ST. Characterization of the human full-lengthPTK7 cDNA encoding a receptor protein tyrosine kinase-like moleculeclosely related to chick KLG. J Biochem 1996;119:235–9.

25. Mossie K, Jallal B, Alves F, Sures I, Plowman GD, Ullrich A. Coloncarcinoma kinase-4 defines a new subclass of the receptor tyrosinekinase family. Oncogene 1995;11:2179–84.

26. Shnitsar I, Borchers A. PTK7 recruits dsh to regulate neural crestmigration. Development 2008;135:4015–24.

27. Lu X, Borchers AG, Jolicoeur C, Rayburn H, Baker JC, Tessier-LavigneM. PTK7/CCK-4 is a novel regulator of planar cell polarity in verte-brates. Nature 2004;430:93–8.

28. LinY, Zhang LH,WangXH, XingXF, ChengXJ, DongB, et al. PTK7 as anovelmarker for favorable gastric cancer patient survival. J SurgOncol2012;106:880–6.

29. ShinWS, Kwon J, LeeHW, KangMC, NaHW, Lee ST, et al. Oncogenicrole of protein tyrosine kinase 7 in esophageal squamous cell carci-noma. Cancer Sci 2013;104:1120–6.

30. Prebet T, LhoumeauAC,ArnouletC,AulasA,MarchettoS,Audebert S,et al. The cell polarity PTK7 receptor acts as a modulator of the






chemotherapeutic response in acute myeloid leukemia and impairsclinical outcome. Blood 2010;116:2315–23.

31. Puppo F, ThomeV, Lhoumeau AC, CiboisM, Gangar A, Lembo F, et al.Protein tyrosine kinase 7 has a conserved role in Wnt/beta-catenincanonical signalling. EMBO Rep 2011;12:43–9.

32. Meng L, Sefah K, O'Donoghue MB, Zhu G, Shangguan D, Noorali A,et al. Silencing of PTK7 in colon cancer cells: caspase-10-depen-dent apoptosis via mitochondrial pathway. PLoS ONE 2010;5:e14018.

33. Zhijin W, Irizarry RA, Gentleman R, Martinez-Murillo F, Spencer F. Amodel-based background adjustment for oligonucleotide expressionarrays. J Am Stat Assoc 2004;99:909–17.

34. Storey J. A direct approach to false discovery rates. J R Stat Soc2002;64:479–98.

35. Fisher RA. Statistical methods for research workers. Edinburgh, Lon-don: Oliver and Boyd; 1925.

36. Mazur PK, Einwachter H, Lee M, Sipos B, Nakhai H, Rad R, et al.Notch2 is required for progression of pancreatic intraepithelial neo-plasia and development of pancreatic ductal adenocarcinoma. ProcNatl Acad Sci U S A 2010;107:13438–43.

37. Lundberg AS, Randell SH, Stewart SA, Elenbaas B, Hartwell KA,Brooks MW, et al. Immortalization and transformation of primaryhuman airway epithelial cells by gene transfer. Oncogene 2002;21:4577–86.

38. RubinsonDA,DillonCP, Kwiatkowski AV, SieversC, Yang L, Kopinja J,et al. A lentivirus-based system to functionally silence genes in primarymammalian cells, stem cells and transgenic mice by RNA interference.Nat Genet 2003;33:401–6.

39. Tomayko MM, Reynolds CP. Determination of subcutaneous tumorsize in athymic (nude) mice. Cancer Chemother Pharmacol 1989;24:148–54.

40. Sato M, Larsen JE, LeeW, Sun H, Shames DS, Dalvi MP, et al. Humanlung epithelial cells progressed to malignancy through specific onco-genic manipulations. Mol Cancer Res 2013;11:638–50.

41. Tang H, Xiao G, Behrens C, Schiller J, Allen J, Chow CW, et al. A 12-gene set predicts survival benefits from adjuvant chemotherapy innon-small cell lung cancer patients. Clin Cancer Res 2013;19:1577–86.

42. 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.

43. 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.

44. Weir BA, Woo MS, Getz G, Perner S, Ding L, Beroukhim R, et al.Characterizing the cancer genome in lung adenocarcinoma. Nature2007;450:893–98.

45. Rhodes DR, Kalyana-Sundaram S, Mahavisno V, Varambally R, Yu J,Briggs BB, et al. Oncomine 3.0: genes, pathways, and networks in acollection of 18,000 cancer gene expression profiles. Neoplasia2007;9:166–80.

46. Kim TM, YimSH, Lee JS, KwonMS, Ryu JW, KangHM, et al. Genome-wide screening of genomic alterations and their clinicopathologicimplications in non-small cell lung cancers. Clin Cancer Res 2005;11:8235–42.

47. Jackson EL, Willis N, Mercer K, Bronson RT, Crowley D, Montoya R,et al. Analysis of lung tumor initiation andprogressionusing conditionalexpression of oncogenic K-ras. Genes Dev 2001;15:3243–48.

48. Endoh H, Tomida S, Yatabe Y, Konishi H, Osada H, Tajima K, et al.Prognostic model of pulmonary adenocarcinoma by expression pro-filing of eight genes as determined by quantitative real-time reversetranscriptase polymerase chain reaction. J Clin Oncol 2004;22:811–19.

49. Boudeau J, Miranda-Saavedra D, Barton GJ, Alessi DR. Emergingroles of pseudokinases. Trends Cell Biol 2006;16:443–52.

50. Zeqiraj E, van Aalten DM. Pseudokinases-remnants of evolution or keyallosteric regulators? Curr Opin Struct Biol 2010;20:772–81.

51. Holbro T, Beerli RR, Maurer F, Koziczak M, Barbas CF 3rd, Hynes NE.The ErbB2/ErbB3 heterodimer functions as an oncogenic unit: ErbB2requires ErbB3 to drive breast tumor cell proliferation. Proc Natl AcadSci U S A 2003;100:8933–8.

52. Dar AC. A pickup in pseudokinase activity. Biochem Soc Trans 2013;41:987–94.

53. Kolch W. Coordinating ERK/MAPK signalling through scaffolds andinhibitors. Nat Rev Mol Cell Biol 2005;6:827–37.

54. WinbergML, Tamagnone L, Bai J, Comoglio PM,Montell D, GoodmanCS. The transmembrane protein off-track associates with Plexins andfunctions downstreamof Semaphorin signaling during axon guidance.Neuron 2001;32:53–62.

55. Peradziryi H, Kaplan NA, Podleschny M, Liu X, Wehner P, Borchers A,et al. PTK7/Otk interacts with Wnts and inhibits canonical Wnt sig-nalling. EMBO J 2011;30:3729–40.


Chen et al.




2014;74:2892-2902. Published OnlineFirst March 20, 2014.Cancer Res Ron Chen, Purvesh Khatri, Pawel K. Mazur, et al. as a Survival Gene in Lung Adenocarcinoma

PTK7A Meta-analysis of Lung Cancer Gene Expression Identifies

Updated version

10.1158/0008-5472.CAN-13-2775doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2014/03/20/0008-5472.CAN-13-2775.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/74/10/2892.full#ref-list-1

This article cites 53 articles, 22 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/74/10/2892.full#related-urls

This article has been cited by 16 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://cancerres.aacrjournals.org/content/74/10/2892To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-13-2775

http://cancerres.aacrjournals.org/content/suppl/2014/03/20/0008-5472.CAN-13-2775.DC1

http://cancerres.aacrjournals.org/content/74/10/2892.full#ref-list-1

http://cancerres.aacrjournals.org/content/74/10/2892.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/74/10/2892


A Meta-analysis of Lung Cancer Gene Expression Identiﬁes ... · Tumor and Stem Cell Biology A...

Documents

Transcript of A Meta-analysis of Lung Cancer Gene Expression Identiﬁes ... · Tumor and Stem Cell Biology A...