Oncogenes and Tumor Suppressors

Human Lung Epithelial Cells Progressed to Malignancythrough Specific Oncogenic Manipulations

Mitsuo Sato1,2, Jill E. Larsen1,2, Woochang Lee1,2, Han Sun3, David S. Shames1,2, Maithili P. Dalvi1,2,Ruben D. Ramirez1,2,5, Hao Tang3, John Michael DiMaio6, Boning Gao1,2,4, Yang Xie3, Ignacio I. Wistuba9,Adi F. Gazdar1,2,7, Jerry W. Shay8,10, and John D. Minna1,2,4,5

AbstractWeusedCDK4/hTERT–immortalized normal human bronchial epithelial cells (HBEC) from several individuals

to study lung cancer pathogenesis by introducing combinations of common lung cancer oncogenic changes (p53,KRAS, andMYC) and followed the stepwise transformation ofHBECs to full malignancy. This model showed that:(i) the combination of five genetic alterations (CDK4, hTERT, sh-p53, KRASV12, and c-MYC) is sufficient for fulltumorigenic conversion of HBECs; (ii) genetically identical clones of transformed HBECs exhibit pronounceddifferences in tumor growth, histology, and differentiation; (iii) HBECs from different individuals vary in theirsensitivity to transformation by these oncogenic manipulations; (iv) high levels of KRASV12 are required for fullmalignant transformation of HBECs, however, prior loss of p53 function is required to prevent oncogene-inducedsenescence; (v) overexpression of c-MYC greatly enhances malignancy but only in the context of sh-p53þKRASV12;(vi) growth of parental HBECs in serum-containing medium induces differentiation, whereas growth of onco-genically manipulated HBECs in serum increases in vivo tumorigenicity, decreases tumor latency, produces moreundifferentiated tumors, and induces epithelial-to-mesenchymal transition (EMT); (vii) oncogenic transformationofHBECs leads to increased sensitivity to standard chemotherapy doublets; (viii) an mRNA signature derived bycomparing tumorigenic versus nontumorigenic clones was predictive of outcome in patients with lung cancer.Collectively, our findings show that thisHBECmodel system can be used to study the effect of oncogenicmutations,their expression levels, and serum-derived environmental effects in malignant transformation, while also providingclinically translatable applications such as development of prognostic signatures and drug response phenotypes.

Visual Overview: http://mcr.aacrjournals.org/content/11/6/638/F1.large.jpg.Mol Cancer Res; 11(6); 638–50. �2013 AACR.

IntroductionHuman lung cancer develops as a multistep process, usually

after prolonged smoke-related tobacco exposure resulting inspecific proto-oncogene and tumor suppressor gene alterationsin lung epithelial cells (1). Genome-wide analyses have iden-tified multiple genetic and epigenetic alterations in lungtumors (2–5). To translate these findings to the clinic, how-ever, it is essential to identify the best targets for early detectionand therapeutic intervention by determining which alterationsrepresent "driver" and which represent "passenger" changes.Functional tests are a critical step in determining "driver"

status and are most commonly conducted by genetically orpharmacologically "correcting" the defect in a lung cancerline and monitoring its effect. Another approach is tointroduce putative oncogenic changes into normal lungepithelial cells and ascertain their contribution to malignan-cy. To undertake the latter, we previously developed an invitromodel of immortalized human bronchial epithelial cells(HBEC) establishing cell lines from more than 30 differentindividuals (6). HBECs were immortalized by overexpres-sing Cdk4 and human telomerase reverse transcriptase(hTERT) to emulate 2 of the earliest events and almost

Authors'Affiliations: 1HamonCenter for TherapeuticOncologyResearch;2Simmons Comprehensive Cancer Center; 3Center for Biostatistics andClinical Science; Departments of 4Pharmacology, 5InternalMedicine, 6Car-dio Thoracic Surgery, 7Pathology, and 8Cell Biology, The University ofTexas Southwestern Medical Center, Dallas; 9Department of Pathology,The University of TexasMDAndersonCancer Center, Houston, Texas; and10Center for Excellence in Genomic Medicine Research, King AbdulazizUniversity, Jeddah, Saudi Arabia

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

M. Sato and J.E. Larsen contributed equally to this work.

Current address for M. Sato: Department of Respiratory Medicine, NagoyaUniversity Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku,Nagoya 466-8550 Japan; current address for W. Lee, Department ofLaboratory Medicine, University of Ulsan College of Medicine and AsanMedical Center, Seoul, 138-736 Korea; current address for D.S. Shames,Oncology Diagnostic, Genentech Inc., South San Francisco, CA 94080;and current address for R.D. Ramirez, Department of Molecular and CellBiology, The University of Texas at Dallas, 800 West Campbell Road,Richardson, TX 75080.

Corresponding Author: John D. Minna, Hamon Center for TherapeuticOncology Research, The University of Texas Southwestern Medical Cen-ter, 6000 Harry Hines Blvd., NB8.206, Dallas, TX 75390. Phone: 214-648-4900; Fax: 214-648-4940; E-mail: John.Minna@utsouthwestern.edu

doi: 10.1158/1541-7786.MCR-12-0634-T

�2013 American Association for Cancer Research.

MolecularCancer

Research

Mol Cancer Res; 11(6) June 2013638

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

universal events in lung cancer pathogenesis: abrogation ofthe p16/Rb cell-cycle checkpoint pathway and bypass ofreplicative senescence. Importantly, this was the first reportof immortalizing lung epithelial cells in the absence of viraloncoproteins, as used previously by other groups (7, 8). Viraloncoproteins such as SV40 early region (containing large Tand small t antigens) are known to cause malignant trans-formation through inactivation of retinoblastoma (Rb) and/or p53, as well as inhibit the tumor suppressor gene PP2Aphosphatase (9, 10). They also have other less characterizedinteractions with cellular processes, thereby contributing tocancer pathogenesis in unknown ways. Thus, the criticalgenetic manipulations that lead tomalignant transformationof epithelial cells immortalized with viral oncoproteins arenot completely "defined." Moreover, they are likely moreoncogenically progressed than nonviral models as shown inmammary epithelial cell systems where full transformationwith SV40 early region required 3 manipulations (hTERT,SV40 early region, and RAS; ref. 8) as compared with 6manipulations (hTERT, dominant-negative p53, CDK4, c-MYC, and cyclin D1) in the absence of SV40 early region(11).While HBECs used in our study are immortalized, a

phenotype they share with cancer cells, they do not displayother cancer cell phenotypes such as disruption of the p53pathway, extensive copy number changes, lack of contact-inhibition, anchorage-independent growth, and the abilityto form tumors in immunodeficient mice (6, 12). Further-more, we have shown that immortalized HBECs retain theability to differentiate into structures found in the normalbronchial epithelium (13, 14). These features make HBECsa physiologically appropriate in vitromodel for studying thetransformation process of bronchial epithelial cells to lungcancer.Transformation of lung epithelial cells to full malignancy

using defined genetic manipulations has only beendescribed in 2 studies (8, 15). One study using viraloncoproteins-transformed tracheobronchial and small air-way epithelial cells with hTERT, SV40 early region, andoncogenic HRAS or KRAS (8). The second study showedfully transformed human small airway epithelial cells(HSAEC) occurred with CDK4, hTERT, mutant p53,mutant KRAS, and either c-MYC, PIK3CA, cyclin D1, orLKB1 manipulations (15). In our bronchial epithelial cellmodel, we have previously shown the combination ofCDK4, hTERT, and p53 knockdown with mutant EGFR,or moderate levels of mutant KRASV12 progressed cellspartially but not completely to malignancy, as evidenced bythe failure to form tumors in immunodeficient mice (12).As lung cancer develops in both the central bronchus andperipheral small airways, development of an in vitro modelof malignant transformation in bronchial epithelial cells isessential.In the present study, we aimed to confirm the genetic

tractability of HBECs and fully transform cells to malig-nancy using oncogenic manipulations commonly found inlung tumors. Loss of p53 function and oncogenic KRAS are2 well-known genetic alterations in lung cancer occurring in

approximately 50% and 30% of non–small cell lungcarcinoma (NSCLC), respectively (16, 17). Aberrantexpression of c-MYC, through amplification or overex-pression, is found in approximately 20% of NSCLCs (1).While it is known that protein levels of oncogenic RAScan influence its oncogenic ability (18), expression ofoncogenic KRASV12 can also lead to premature senescenceof normal human epithelial cells (19). The prevalence ofKRAS alterations in NSCLC indicates however that malig-nant transformation requires the cell to adapt to thisoncogenic stress, perhaps assisted through preceding onco-genic transformations (20). Here, we present one of thefirst reports of full malignant transformation of lungepithelial cells with defined genetic manipulations. Fur-thermore, we characterize the effect of oncogenic stressand environmental effects such as growth factors upontumorigenic transformation in HBECs, illustrate diver-gent clonal heterogeneity, and determine the capability ofthis in vitro model for developing and testing lung cancertherapeutics.

Materials and MethodsCells and culture conditionsHBEC3 (HBEC3KT), HBEC4 (HBEC4KT), and

HBEC17 (HBEC17KT) immortalized normal HBEC lineswere established by introducing mouse Cdk4 and hTERTinto normal HBECs (6). HBECs were cultured with ker-atinocyte serum-free medium (KSFM; Life TechnologiesInc.) media containing 50 mg/mL of bovine pituitary extract(BPE; Life Technologies Inc.) and 5 ng/mL of EGF (LifeTechnologies Inc.). Partially transformed HBECs (soft agarclones) were also cultured with RPMI-1640 (Life Technol-ogies Inc.) media supplemented with 10% FBS (R10). Lungcancer cell lines were established by our laboratory andmaintained in RPMI-1640 (Life Technologies Inc.) with5% FBS (21, 22). All cell lines were DNA-fingerprinted(PowerPlex 1.2 Kit; Promega) andMycoplasma-free (e-MycoKit; Boca Scientific).

Viral vector construction and viral transductionStable p53 knockdown and moderate expression of

KRASV12 was achieved as described previously (12). Expres-sion of high KRASV12 levels used a lentiviral vector, pLenti6-KRASV12, as described previously (23). Lentiviral vectorsexpressing KRASWT, KRASC12, and KRASD12 were con-structed from pLenti6-KRASV12 as described previously(24). A c-MYC-expressing retroviral vector (designatedpMSCV-MYC) was constructed by ligating a BamHI/SalI–digested c-MYC insert from pCTV3K (ref. 25; a giftfrom Dr. J. Michael Ruppert, University of Alabama,Tuscaloosa, AL) into BglII/XhoI–digested pMSCV-hyg(Clontech). Lentivirus and retrovirus-containing mediumwere produced as described previously (12). Transducedcells were selected with zeocin (12.5 mg/mL), blasticidin (2mg/mL), or hygromycin (20 mg/mL) for 7 to 14 days. Thepresence of mutant KRASV12 in stable cell lines was con-firmed using a reverse transcription PCR (RT-PCR)/RFLPassay, as described previously (12).

Oncogenic Transformation of Human Lung Epithelial Cells

www.aacrjournals.org Mol Cancer Res; 11(6) June 2013 639

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

Immunoblotting and southern blottingPreparation of total cell lysates andWestern blotting were

carried out as described previously (26). Primary antibodiesare listed in Supplementary Table S1 and were detected withhorseradish peroxidase (HRP)–conjugated anti-rabbit oranti-mouse secondary antibodies (1:2,000 dilution; ThermoFisher Scientific). Actin orHSP90 protein levels were used asloading controls. Southern blotting for HBEC clonesderived from large soft agar colonies was carried out asdescribed previously (27) using Phototope-Star DetectionKit (New England Biolabs). DNA was probed with theblasticidin-resistance gene amplified from pLenti6-KRASV12 vector using 50-ATGGCCAAGCCTTTGTCT-CAAG-30 and 50-TTAGCCCTCCCACACATAACC-30primers.

Biochemical and in vitro transformation assaysSenescent cells were stained with Senescence b-Galacto-

sidase Staining Kit (Cell Signaling) and blue-stained senes-cent cells were counted under a microscope (�20 totalmagnification). Percentage of positively stained cells wasaveraged across 6 fields. Cell-cycle analysis was conducted onsubconfluent populations of cells harvested 48 hours afterseeding, unless otherwise stated, and cell-cycle analysis wasconducted as described previously (28). Cell proliferationassays were conducted by seeding 5,000 cells in 6-well platesand counting cells every 4 days. Cells were expanded whensubconfluent, as necessary. Anchorage-dependent colonyformation assays were conducted as previously described(12); 200 to 600 viable cells were seeded in triplicate in 100-mm plates and cultured for 2 weeks before staining colonieswith methylene blue. Acute KRASV12 toxicity assays wereconducted by transducing cells with KRASV12 or LacZlentivirus and selecting for 3 days with blasticidin beforeseeding anchorage-dependent colony formation assays.Anchorage-independent (soft agar) growth assays were con-ducted as previously described (12) seeding 1,000 viable cellsin 12-well plates. MTS assays to measure drug response tostandard platinum-based doublets [paclitaxel–carboplatin(2/3 wt/wt), gemcitabine–cisplatin (25/2 wt/wt), and peme-trexed–cisplatin (20/3 wt/wt)] were conducted as previouslydescribed (29). Cells were treated for 96 hours with 4-folddilutions from a maximum dose of 1,000/3,501 nmol/L(paclitaxel/carboplatin), 1,000/298 mmol/L (pemetrexed–cisplatin), or 2,000/140 nmol/L (gemcitabine/cisplatin).Each experiment was carried out in quadruplicate with 8replicates per experiment. Fitting of data to dose–responseinhibition curves, calculation of ED50 values, and compar-isons based on one-way ANOVA with Dunnett posttestwere conducted using GraphPad Prism version 5.00 forWindows (GraphPad Software).

In vivo tumorigenicity assays and histologic analysisIn vivo tumorigenicity was evaluated by injection of cells

in female 5- to 6-week-old nonobese diabetic/severe com-bined immunodeficient (NOD/SCID) mice. Each mousewas given a subcutaneous injection on its flank of 3 to 5 �106 viable cells in 0.2mLof PBS.Miceweremonitored every

2 to 3 days for tumor formation for up to 6 months. Allanimal care was in accord with institutional guidelines andapproved Institutional Animal Care and Use Committee(IACUC) protocols. Formalin-fixed, paraffin-embedded(FFPE) xenograft tumor tissue was sectioned and stainedwith hematoxylin and eosin (H&E), Alcian Blue–PAS, PAS,and mucicarmine for histologic analysis. Immunohisto-chemical staining for p63 (BioCare; clone BC4A4), NapsinA, CK5, and CK7 was carried out commercially by ProPath.Terminal deoxynucleotidyl transferase–mediated dUTPnick end labeling (TUNEL) staining was conducted usingDeadEnd Fluorometric TUNEL System (Promega).

Statistical testing and microarray analysisFor comparison of colony formation and senescent cells

between different geneticallymanipulated cell strains,we useda two-tailed Student t test and P < 0.05 was consideredsignificant. mRNA microarray profiling was conducted withIllumina HumanHT-12 v4 Expression BeadChips (IlluminaInc.) following the manufacturer's guidelines and analyzedwith in-house Visual Basic softwareMATRIXV1.483. Func-tional analysis of differentially expressed genes was conductedusing Ingenuity Pathway Analysis (IPA; Ingenuity Systems,Inc.). The data discussed in this publication have been madeavailable in the National Center for Biotechnology Informa-tion's Gene Expression Omnibus (GEO) public repository(http://www.ncbi.nlm.nih.gov/geo/) and are accessiblethrough GEO Series accession number GSE40828. Thepredictive ability of the soft agar gene signature was testedusing 2 independent mRNAmicroarray lung tumor datasets;209 primary lung adenocarcinomas and squamous cell car-cinomas (SCC;SPOREdataset)GSE41271, and442primarylung adenocarcinomas [National Cancer Institute (NCI)Director's Challenge Consortium dataset; ref. 30). A detaileddescription of these 2 datasets and the prediction analysis isgiven in the Supplementary Methods.

ResultsOncogenic transformation of HBECs correlates withlevel of exogenous KRASV12 expressionNSCLC cell lines harboring a mutation in KRAS express

wide-ranging endogenous levels of the protein (Fig. 1A).Thus, to create HBECs that express comparable levels ofoncogenic KRAS protein, we used 2 different expressionvectors: a retroviral vector, which resulted inmodest levels ofexpression, and a lentiviral vector that resulted in high levelsof expression (Fig. 1A). Compared with HBEC3 expressingmoderate levels (retroviral-mediated) of KRASV12 protein,HBEC3 expressing high levels of KRASV12 (lentiviral-mediated) exhibited a significant increase in soft agar colo-nies in the background of both wild-type p53 (HBEC3)and p53 knockdown (HBEC3p53; Fig. 1B). To confirmthe contribution of the level of KRASV12 expressiontoward HBEC transformation, we examined 7 clonal popu-lations of HBEC3p53,KRAS (sh-p53 and lentiviral-deliveredKRASV12; Fig. 1C). Clones with high levels of KRASexpression had increased anchorage-dependent (liquid) col-ony-forming ability (Fig. 1D and Supplementary Fig. S1A).

Sato et al.

Mol Cancer Res; 11(6) June 2013 Molecular Cancer Research640

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

High levels of mutant KRASV12 induce senescence inHBECs, in a p53-dependent mannerLentiviral-mediated introduction of high levels of

KRASV12 caused a large subset of HBEC3 cells to displaysignificant morphologic changes, including flattened,enlarged shape, and a vacuole-rich cytoplasm, suggestive ofoncogene-induced senescence. This effect was not observedfollowing exogenous expression of moderate levels ofKRASV12 (Supplementary Fig. S1B). Senescence-associatedb-galactosidase (SA-b-gal) staining confirmed the morpho-logic changes correspondedwith senescence (SupplementaryFig. S1C). Significantly less RAS-induced senescence wasobserved in HBEC3 with stable p53 knockdown(HBEC3p53) compared with HBEC3 with wild-type p53(Fig. 1E and Supplementary Fig. S1C). An acute, senes-cence-associated p53-mediated KRASV12 toxicity wasshown in anchorage-dependent colony formation assays

seeded 4 days after transduction with KRASV12 lentivirus.Significant toxicity was observed in HBECs with wild-typep53 (HBEC3 and another HBEC line, HBEC4) followingtransduction with KRASV12 lentivirus compared with con-trol LacZ lentivirus but not inHBECs with p53 knockdown(Fig. 1F and Supplementary Fig. S1D and S1E), suggestinginduced senescence was p53-dependent. p53-mediated mit-igation of oncogenic KRAS stress was also observed withother codon 12 KRAS mutants (KRASC12 and D12; Sup-plementary Fig. S1F and S1G), whereas overexpression ofwild-type KRAS in HBEC3 did not induce a senescencephenotype (Supplementary Fig. S1G). Immunoblotting forcleaved caspase-3 confirmed KRASV12-induced toxicity wasnot apoptosis-associated (Fig. 1G), whereas cell-cycle anal-ysis found no increased accumulation of cells in G1-phase inHBEC3p53 cells infected with KRASV12 as compared withthe control vector (Supplementary Fig. S1H).

Figure 1. High exogenous levels ofKRASV12, comparable withendogenous levels found in mutantKRAS NSCLC cell lines, increasestransformation of HBECs andinduces senescence, which is largelybypassed with p53 knockdown. A,immunoblot for KRAS proteinexpression in HBEC3 cells infectedwith KRASV12 using either amoderately expressing retroviral(pBabe-hyg-KRASV12) or high-expressing lentiviral (pL6-KRASV12)vector. Actin was used as loadingcontrol. B, anchorage-independent(soft agar) colony formation inHBEC3 with high (lentiviral) ormoderate (retroviral) levels ofKRASV12 in the background of bothwild-type p53 and p53 knockdown(sh-p53; t test). C, immunoblot ofHBEC3p53,KRAS soft agar clonesconfirming p53 and KRASV12

manipulations. The presence (þ) orabsence (�) of each manipulation isshown. D, anchorage-dependent(liquid) colony formation ability ofHBEC3p53,KRAS soft agar clones. E,quantification of SA-b-gal stainingfound KRASV12-inducedsenescence in HBEC3 cells wassignificantly lower in cells with p53knockdown comparedwith p53wild-type (t test). F, anchorage-dependentcolony formation assay to compareacute KRASV12-induced toxicity inHBEC3 and HBEC4 with wild-typep53 or p53 knockdown. G,immunoblot of HBEC3 cell lysatesharvested 7 days after infection withKRASV12 or LacZ lentivirus.�,P < 0.05; ��,P < 0.01; ���,P < 0.001.Full-length blots are presented inSupplementary Fig. S8.

A B

Clo

ne 1

Clo

ne 5

Clo

ne 1

1

Clo

ne 6

Clo

ne 7

Clo

ne 8

Clo

ne 9

KRAS

Actin

p53- ++++++ ++KRASV12

sh-p53 - ++++++ ++

C D

0.0

0.2

0.4

0.6

Col

ony

form

ing

effic

ienc

y

Clo

ne 1

Clo

ne 5

Clo

ne 1

1C

lone

6C

lone

7C

lone

8C

lone

9

0

100

200

300

No.

col

onie

s (>

50 c

ells

)

-+

+--Mod. KRASV12

sh-p53

High KRASV12

*

**

-+

+-+-

-+-

Vec

tor

pBab

e-K

RA

SV

12

pLen

ti-K

RA

SV

12

H20

09

H17

92

H13

73

H46

0

H16

93

H28

82

KRAS

Actin

HBEC3 NSCLC cell lines

MutantKRAS

Wild-typeKRAS

0

50

100

150

No.

sen

esce

nt c

ells

/fiel

d

+-sh-p53KRASV12 ++

0.0

0.2

0.4

0.6

0.8

1.0

Col

ony

form

ing

effic

ienc

y LacZ KRASV12- ++-sh-p53

KRASV12 - +-+

p53

KRAS

p16

Caspase-3

Cleavedcaspase-3

***

- +-+sh-p53HBEC3 HBEC4

GFE

+-

Oncogenic Transformation of Human Lung Epithelial Cells

www.aacrjournals.org Mol Cancer Res; 11(6) June 2013 641

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

These results show that oncogenic KRAS mediates apotent cellular stress response in CDK4/hTERT–immor-talized HBECs, whereby the cells resist oncogenic transfor-mation by engaging cellular senescence. Loss of p53 functionhowever impedes this cellular senescence response, indicat-ing that in these cells, p53 signaling is a primary mediator ofRAS-induced senescence. Furthermore, as high levels ofoncogenic KRAS expression are required for malignanttransformation, it shows that loss of p53 function is a criticalco-oncogenic step in the malignant transformation of thelarge majority of HBECs.

The combination of p53 knockdown and KRASV12 inHBECs significantly increases in vitro transformation,which is further augmented with c-MYC overexpressionThe single introduction of p53 knockdown, mutant

KRASV12 or c-MYC overexpression resulted in quantitativelymodest but significant increases in soft agar colony number(Fig. 2A–C). The combination of p53 knockdown andKRASV12 (HBEC3p53,KRAS) resulted in a significant increasein transformation not observed in other dual combinations[p53 knockdownþc-MYC (HBEC3p53,MYC) or KRASV12þc-MYC (HBEC3KRAS,MYC)], whereas introduction of all 3manipulations (HBEC3p53,KRAS,MYC) resulted in the mosttransformed phenotype (Fig. 2B). Higher expression ofc-MYC was achieved if overexpressed in HBEC3p53 (p53knockdown), HBEC3KRAS (KRASV12), or HBEC3p53,KRAS

(p53 knockdown and KRASV12) cells (Fig. 2A), suggestingpartial transformation of immortalized HBEC3 with eitherp53 knockdown or addition of mutant KRASV12 is requiredfor the cells to tolerate high levels of c-MYC.

Combination of sh-p53þKRASV12 orsh-p53þKRASV12þc-MYC oncogenically transformsHBEC3 in vivoTo test the tumor-forming ability of HBEC3p53,KRAS

and HBEC3p53,KRAS,MYC, cells grown in defined KSFMmedia were injected subcutaneously into NOD/SCIDmice. We had previously found HBEC3p53 with moderatelevels of oncogenic KRASV12 (using a retroviral expressionvector) failed to form tumors when injected into immu-nodeficient mice (12). In contrast, transforming withhigher levels of KRASV12 resulted in tumor growth in5 of 24 (21%) injections (Table 1). Despite the significantincrease in in vitro anchorage-independent growthobserved in HBEC3p53,KRAS,MYC compared withHBEC3p53,KRAS (Fig. 2B), HBEC3p53,KRAS,MYC was onlyslightly more tumorigenic in vivo with tumors in 3 of 10(30%) injections (Table 1). Histopathologic analysisshowed that the oncogenically manipulated HBEC popu-lations formed SCCs, adenocarcinomas, adenosquamouscarcinomas, and poorly differentiated carcinomas withtypical morphologic features of each histology (Table 1and Fig. 3). Squamous and adenocarcinoma differentia-tion was confirmed by p63 and mucicarmine/Alcian Blue–PAS positivity, respectively, and pathology was verified byan independent lung cancer pathology expert. Adenocar-cinomas were found to be strongly cytokeratin 5–positive,whereas negative for Napsin A and the squamous cellmarker, cytokeratin 7 (Supplementary Fig. S2A). In allcases, adenocarcinoma differentiation was TITF1(NKX2-1)–negative (data not shown), most likely a reflection ofthe HBECs being derived from central airway cells. Thefact that glandular cells are negative for both TITF-1 andNapsin A is not unexpected as the vast majority ofadenocarcinomas (and glandular component in adenos-quamous carcinomas) in whole tissue sections are eitherpositive for both or negative for both. The development ofdifferent tumor histologies from the same HBEC-manip-ulated population suggests either clonal selection or thecells respond to differentiation signals in vivo.

Exploration of interindividual differences in HBECtransformationTo compare interindividual differences in malignant

transformation using the same combination of oncogenicchanges, we tested HBEC17, derived from a differentindividual than HBEC3.We observed a similar transformedphenotype in vitro (Supplementary Fig. S2B) however, incontrast to HBEC3, HBEC17 was significantly more resis-tant to full in vivo transformation (Table 1). The differencein tumor formation rate between HBECs derived fromdifferent donors suggests the existence of interindividualdifferences in susceptibility to specific oncogene-inducedmalignant transformation.

KRAS

Actin

KRASV12sh-p53

- ++-+-- +

cMYC

p53cMYC

- +-+-+- +- -++--+ +

0

200

400

600

800 **

*

****

No.

col

onie

s >

50 c

ells

)

KRASV12

sh-p53

- ++-+-- +cMYC- +-+-+- +- -++--+ +

Vector sh-p53

sh-p53KRASV12

sh-p53cMYC

sh-p53KRASV12

cMYC

KRASV12 cMYC

KRASV12

cMYC

A

B

C

Figure 2. Stepwise in vitro transformation of HBEC3 with sh-p53,KRASV12, and c-MYC. A, immunoblot of isogenic derivatives of HBEC3with sh-p53 or overexpression of KRASV12 or c-MYC, alone or incombination. The presence (þ) or absence (�) of each manipulation isshown. B, transformation as defined by anchorage-independent growthin soft agar assays for HBEC3 with each manipulation alone or incombination. C, representative photographs of soft agar assays showingthe formation of large, macroscopic (>1 mm) colonies in HBEC3p53,KRAS

and HBEC3p53,KRAS,MYC (�4 magnification). Full-length blots arepresented in Supplementary Fig. S8.

Sato et al.

Mol Cancer Res; 11(6) June 2013 Molecular Cancer Research642

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

Growth in serum-containing medium enhances fulltumorigenic transformation of HBECsIn addition to viral oncoproteins, many previous studies

that succeeded in transforming normal cells to malignancyalso used serum-containing medium instead of definedserum-free medium (31–33). We have previously shownthat CDK4/hTERT–immortalized HBECs with no addi-

tional oncogenic manipulations require EGF, a supplementpresent in KSFM medium, for cell growth, but oncogenictransformation with p53 knockdown and KRASV12 allowsthe cells to become EGF-independent (12). In the presentstudy, we show while HBEC3 cannot tolerate growth inRPMI-1640 supplemented with serum (FBS), media com-monly used for growth of cancer cell lines, HBEC3p53,KRAS,

Table 1. In vivo tumorigenicity of manipulated HBECs

Cell line Mediuma Tumor formation rateb Latency, dc Histology

HBEC3 p53,KRASPopulation KSFM 5/24 (21%) 152 SCC (2)

Adenosquamous (2)Poorly differentiated carcinoma (1)

Population R10 7/17 (41%) 128 Large cell/giant cell carcinoma (3)Adenocarcinoma (2)SCC (1)Adenosquamous (1)

Clone 1 R10 8/8 (100%) 38 Large cell/giant cell carcinoma (7)SCC (1)

Clone 5 R10 7/8 (88%) 57 Large cell/giant cell carcinoma (5)Papillary adenocarcinoma (1)Poorly differentiatedcarcinoma (1)

Clone 11 R10 6/8 (75%) 117 Large cell/giant cell carcinoma (3)Poorly differentiated carcinoma (2)Adenocarcinoma (1)

Clone 6 R10 0/6 (0%) — —

Clone 7 R10 0/7 (0%) — —

Clone 8 R10 0/8 (0%) — —

Clone 9 R10 0/9 (0%) — —

HBEC3 p53,KRAS,MYCPopulation KSFM 3/10 (30%) 115 Adenosquamous (2)

SCC (1)Population R10 12/12 (100%) 27 Large cell/giant cell carcinoma (5)

Poorly differentiated carcinoma (4)SCC (2)

Clone 1 R10 9/9 (100%) 119 Large cell/giant cell carcinoma (7)Adenocarcinoma (2)

Clone 2 R10 3/8 (38%) 90 SCC (2)Large cell/giant cell carcinoma (1)

Clone 7 R10 10/10 (100%) 37 Large cell/giant cell carcinoma (7)SCC (3)

Clone 8 R10 1/6 (17%) 78 Large cell/giant cell carcinoma (1)HBEC17 p53,KRASPopulation KSFM 0/9 (0%) — —

Population R10 0/8 (0%) — —

HBEC17 p53,KRAS,MYCPopulation KSFM 0/10 (0%) — —

Population R10 3/8 (38%) n.d. SCC (2)Sarcomatoid carcinoma features (1)

aR10, RPMI-1640 supplemented with 10% FBS.bNumber of subcutaneous tumors/number of injections (%).cMedian time (days) for subcutaneous xenografts to reach 250 mm3; n.d., not determined.

Oncogenic Transformation of Human Lung Epithelial Cells

www.aacrjournals.org Mol Cancer Res; 11(6) June 2013 643

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

is adaptable and will proliferate (Supplementary Fig. S3A).Serum-supplemented media has been shown to inducedifferentiation of epithelial cells in culture (34). Therefore,to further delineate the differences between HBEC3 com-pared with HBEC3p53,KRAS and HBEC3p53,KRAS,MYC fol-lowing growth in serum-supplementedmedia, cells grown ineither serum-free KSFM or serum-supplemented (5% FBS)KSFM for 96 hours were then analyzed for expression of apanel of lung differentiation and cancer stem cell (CSC)markers by quantitative RT-PCR (qRT-PCR). All 3 cell

lines expressed high levels of basal markers (KRT5 andKRT14), low levels of central airway epithelial cell markers(MUC1 and TUBB4), with undetectable expression ofperipheral airway markers (CC10 and SPC) in line withcells derived from bronchial epithelial cells (SupplementaryFig. S3B). Aldehyde dehydrogenase (ALDH) activity hasbeen shown to be a marker of CSCs in the lung (28). Wetherefore measured the expression of 2 ALDH isozymes,ALDH1A1 and ALDH1A3, to find all the 3 cell lines thatexpressed ALDH1A3 with much lower expression of

Adenosquamous cell carcinoma

Adenocarcinoma

Squamous cell carcinoma

Large cell carcinoma (LCC)

LCC with giant cell component

AB-PAS

H&E Muc P63

H&E H&E

H&EH&E P63

Figure 3. Representative FFPE sections of subcutaneous xenografts derived from HBEC3p53,KRAS and HBEC3p53,KRAS,MYC. HBEC3p53,KRAS andHBEC3p53,KRAS,MYC formed subcutaneous tumor reflective of naturally arising lung carcinomas with adenosquamous differentiation (top), adenocarcinoma(middle left), and squamous differentiation (bottom left), as well as undifferentiated large cell carcinomas, some of which also exhibited a giant cellcomponent (middle and bottom on right). Squamous and adenocarcinoma differentiation was confirmed with p63 and mucicarmine and/or Alcian Blue PASstaining, respectively. The example of adenosquamous cell carcinoma (top) clearly shows dual differentiation of peripheral squamous/basal-like cells(p63þ/mucin�) and central glandular cells (p63�/mucinþ). Muc, mucicarmine; AB-PAS, Alcian Blue PAS. Original magnification of images at �10 exceptadenosquamous H&E and P63 (�20); large cell carcinoma with giant cell component H&E (�20); and large cell carcinoma H&E (�40).

Sato et al.

Mol Cancer Res; 11(6) June 2013 Molecular Cancer Research644

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

ALDH1A1. Growth in serum-containing medium resultedin all 3 cell lines exhibiting significant decrease in theexpression of basal markers (KRT5 and KRT14), mostnotably in HBEC3 (Supplementary Fig. S3C). Expressionof central airway markers, particularly MUC1, increasedmore than 10-fold in HBEC3 when cells were grown inserum-supplemented medium, whereas HBEC3p53,KRAS

and HBEC3p53,KRAS,MYC showed little if any increase inthese markers. Expression of CSC markers ALDH1A1 andALDH1A3 increased in HBEC3 and HBEC3p53,KRAS butnot in HBEC3p53,KRAS,MYC. Morphologically, growth inserum-supplemented medium resulted in HBEC3 cellsdeveloping a flattened morphology representative of adifferentiated state (Supplementary Fig. S3D),whereas this was largely absent in HBEC3p53,KRAS cellsand completely absent in HBEC3p53,KRAS,MYC. Further-more, HBEC3p53,KRAS,MYC and to a lesser extentHBEC3p53,KRAS developed an elongated mesenchymalmorphology after short-term growth in serum (Supple-mentary Fig. S3D). Together, these data show thatserum-supplemented medium induces differentiation inparental HBEC3 cells; however, oncogenic transforma-tion enables the cells to resist serum-induced differenti-ation and instead undergo epithelial-to-mesenchymaltransition (EMT).We reasoned that partially transformed HBECs that have

adapted to growth in serum-supplemented RPMI-1640maydiffer in tumorigenicity compared with cells grown inKSFM. Therefore, to compare the effect of genetic andenvironmental manipulations, HBEC3p53,KRAS andHBEC3p53,KRAS,MYC were grown in either defined KSFM(serum-free) medium or RPMI medium supplemented with10% FBS (R10) for at least 3 weeks, then injected subcu-taneously in NOD/SCIDmice (Table 1). Growth in serum-containing medium markedly increased in vivo tumorige-nicity, decreased tumor latency, and tumors in general weremore undifferentiated (e.g., large cell and giant cell carci-nomas), particularly in HBEC3p53,KRAS,MYC (Table 1and Fig. 3). This effect is similar to what is observed inpatients where poorly differentiated lung tumors aregenerally associated with aggressive tumor growth. Todetermine if the rate of apoptosis differed between poorand well-differentiated xenograft tumors, FFPE sectionswere analyzed with a TUNEL assay. There was no signif-icant difference in the rate of apoptosis in relation todifferentiation although a moderate well-differentiated SCCand giant cell carcinoma showed the greatest amount ofTUNEL staining (Supplementary Fig. S3E). Overall, theincrease in tumorigenicity of oncogenically progressedHBECs after growth in serum shows the influential role ofexogenous serum-derived factors in the malignant progres-sion of lung cancer.

HBEC3p53,KRAS + 10% FBS+ c-MYC

N-cadherin

Vimentin

E-cadherin

HSP90

- + -+ - -+++10% FBS

KRASV12sh-p53

- +--++- +cMYC

+ ++++++ +HBEC3 HBEC17

+ ++++++ +

A

B

C

CDH2 NM_001792 3.89 2.27 COL5A2 NM_000393 6.65 GNG11 NM_004126 2.57 IGFBP4 NM_001552 7.93 4.11 ITGB1 NM_133376 3.17 MAP1B NM_005909 3.62 MMP2 NM_004530 3.53 MMP3 NM_002422 3.42 2.99 SPARC NM_003118 3.36 SPP1 NM_001040058 5.60 5.41 TCF4 NM_003199 2.72 TGFB2 NM_003238 4.47 3.20 VCAN NM_004385 7.13 WNT5A NM_003392 4.05 WNT5B NM_032642 CDH1 NM_004360 –6.69 –6.76 COL3A1 NM_000090 –2.84 DSP NM_004415 –2.59 F11R NM_016946 –2.19 FGFBP1 NM_005130 –6.14 FN1 NM_054034 –2.11 IL1RN NM_173843 –2.06 KRT14 NM_000526 –5.32 –4.90 MMP9 NM_004994 –2.32 –3.74 MST1R NM_002447 –3.32 OCLN CR606900 –2.07 PLEK2 NM_016445 SNAI2 NM_003068 –3.02 –2.60 TWIST1 NM_000474 –2.36

HBEC3 genotype

p53,KRAS p53,KRAS,MYCSymbol Accession

Log2 expression level (RPMI+5%FBS/KSFM)

Figure 4. c-MYC overexpression or growth in serum-containing mediainduces EMT in HBEC3p53,KRAS. A, phase contrast photomicrographsshowing the morphologic effect observed in HBEC3p53,KRAS (left)following overexpression of c-MYC (middle) or switching from definedserum-free media to serum-containing media (right; �20 magnification).B, immunoblot for EMT markers in oncogenically manipulated HBEC3and HBEC17 grown in either KSFM (serum-free) or serum-containing(R10) media. The presence (þ) or absence (�) of serum is shown.C, EMT-related genes altered 4-fold or greater in pairwise analysis of

HBEC3p53,KRAS and HBEC3p53,KRAS,cMYC comparing cells grown inserum or defined medium (KSFM). Values log2 transformed. �, P < 0.05;��, P < 0.01 (t test). Full-length blots are presented in SupplementaryFig. S8.

Oncogenic Transformation of Human Lung Epithelial Cells

www.aacrjournals.org Mol Cancer Res; 11(6) June 2013 645

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

Malignant transformation of HBECs is enhanced byepithelial-to-mesenchymal transition, induced by eitherc-MYC or growth in serumIn HBEC3p53,KRAS cells, overexpression of c-MYC (in

defined KSFM medium) or growth in serum-containingRPMI-1640 medium led to increased oncogenic transfor-mation, as shown by soft agar colony formation and in vivotumor growth. Both of these manipulations also led toHBEC3p53,KRAS cells exhibiting a more mesenchymal-likemorphology (Fig. 4A). Loss of E-cadherin (epithelial mark-er) and gain of vimentin and N-cadherin (mesenchymalmarkers) confirmed an EMT (Fig. 4B). c-MYC- or serum-induced EMT was also seen in HBEC17 cells, derived fromanother individual, with p53 and KRASV12 manipulations(HBEC17p53,KRAS; Fig. 4B). Whole-genome mRNA profil-ing of HBEC3p53,KRAS and HBEC3p53,KRAS,MYC cells grownin KSFMor R10 confirmed a significant over-representationof EMT-related genes (P ¼ 1.04 � 10�16 and P ¼ 1.38 �10�09 for HBEC3p53,KRAS and HBEC3p53,KRAS,MYC,respectively; Fig. 4C), with upregulation of EMT-promot-ing genes following growth in serum-containing media.

Isogenic soft agar clones of oncogenically manipulatedHBEC3 represent independent genetic events withdistinct in vivo growth, tumor histology, anddifferentiationThe genetic combinations of shp53þKRASV12 or

shp53þKRASV12þc-MYC in HBEC3 led to the forma-tion of a small subset (0.5%–1.5% of all soft agar colonies)of very large, macroscopically visible (>1 mm) colonieswhen grown in soft agar, which was not observed withany single manipulation (shp53, KRASV12, or c-MYC)nor dual combination of sh-p53 or KRASV12 with c-MYC (Fig. 2C). Seven HBEC3p53,KRAS clones and 4HBEC3p53,KRAS,MYC clones were isolated from these largecolonies and repeat soft agar assays confirmed that thelarge colony phenotype was maintained (Fig. 5A). South-ern blotting showed the clones represent independenttransformation events as indicated by discrete digestionpatterns (Supplementary Fig. S4A) and immunoblottingconfirmed HBEC3p53,KRAS and HBEC3p53,KRAS,MYC

clones maintained their exogenously introduced oncogen-ic manipulations (Figs. 1C and 5B, respectively; summa-rized in Supplementary Table S2). While immortalizedbut nontransformed HBECs preferentially grow in serum-free conditions, they require serum for anchorage-inde-pendent growth (Supplementary Fig. S4B). Remarkably,following isolation from soft agar (in KSFMþ20% FBS) 9of 11 clones were serum growth factor–dependent as they

A B

Clone 1 Clone 8

HBEC3p53,KRAS

KRASV12sh-p53

- ++++++cMYC

- ++++++- ++++++

Clo

ne 1

Clo

ne 2

Clo

ne 7

Clo

ne 8

KRAS

HSP90

p53

c-MYC

Subclone of large colonies

0.0

0.1

0.2

0.3

0.4

0.5

Col

ony

form

ing

effic

ienc

y

Clo

ne 1

Clo

ne 5

Clo

ne 1

1C

lone

6C

lone

7C

lone

8C

lone

9

Tumor-igenic

Nontumorigenic

HBEC3p53,KRAS

Clo

ne 1

(1)

Clo

ne 1

(2)

Clo

ne 5

(1)

Clo

ne 5

(2)

Clo

ne 1

1(1)

Clo

ne 1

1(2)

Clo

ne 6

(1)

Clo

ne 6

(2)

Clo

ne 7

(1)

Clo

ne 7

(2)

Clo

ne 8

(1)

Clo

ne 8

(2)

Clo

ne 9

(1)

Clo

ne 9

(2)

Tumorigenic Nontumorigenic

C

D

E

F

0 20 40 60 80 1001200.0

0.2

0.4

0.6

0.8

1.0

Months

P = 0.0026

Consortium to SPORE

Sur

viva

l

Months

P = 0.00011

SPORE to Consortium0 50 100 150 200

0.0

0.2

0.4

0.6

0.8

1.0

Sur

viva

l

Low High

Figure 5. Isolation of large soft agar clones from HBEC3p53,KRAS andHBEC3p53,KRAS,MYC identifies tumorigenic and nontumorigenic clonesand genome-wide mRNA expression profiling of HBEC3p53,KRAS

soft agar clones identifies a clinically applicable signature ofprognosis. A, uncloned, parental populations of HBEC3p53,KRAS andHBEC3p53,KRAS,MYC form very large (>1-mm diameter) soft agarcolonies (arrowhead). These large colonies were isolated, expanded,and retested for soft agar colony formation where they maintained theability to form large soft agar colonies (representative soft agarpictures of 2 clones; �4 magnification). B, immunoblot ofHBEC3p53,KRAS,MYC soft agar clones confirming p53, KRASV12, andc-MYC manipulations. The presence (þ) or absence (�) of eachmanipulation is shown. C, anchorage-independent soft agar colonyformation ability of HBEC3p53,KRAS soft agar clones. D, unsupervisedhierarchical clustering of whole-genome mRNA expression profiles ofHBEC3p53,KRAS soft agar clones harvested at 2 time points [denoted"(1)" and "(2)"] spanning a 3-week interval. E, a supervised analysiscomparing HBEC3p53,KRAS tumorigenic (clone 1, 5, and 11) withHBEC3p53,KRAS nontumorigenic (clone 6, 7, 8, and 9) clones identified203 probes, representing 171 unique genes, significantly differentiallyexpressed (SAM, FDR ¼ 5%). Samples (represented horizontally: red,tumorigenic clones; green, nontumorigenic clones) and probes(represented vertically) were clustered using centered Pearson

clustering. F, Kaplan–Meier log-rank analysis of overall survival inpatients with lung cancer predicted to have good (black) or poor (red)outcome using the 171 probe signature HBEC3p53,KRAS soft agarsignature. A supervised principal component analysis was used totrain the model in one dataset (Consortium) and test in a seconddataset (SPORE; top) then the datasets were reversed to test formodel robustness (bottom). Full-length blots are presented inSupplementary Fig. S8.

Sato et al.

Mol Cancer Res; 11(6) June 2013 Molecular Cancer Research646

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

could no longer tolerate growth in serum-free KSFMmedium (Supplementary Table S2). Henceforth, allclones were grown in RPMI-1640 medium containing10% FBS (R10).Injection of the large, soft agar clones into NOD/SCID

mice found all 4 HBEC3p53,KRAS,MYC clones were tumor-igenic, whereas only 3 of 7 HBEC3p53,KRAS clones formedtumors (Table 1). Thus, even with 4 oncogenic changes andbiologic selection there is dramatic clonal heterogeneity.This also showed that the tumorigenicity of the clonesreflected the tumorigenicity of the parental population(when grown in R10) where approximately 40% of miceinjected with HBEC3p53,KRAS grew tumors compared with100% of mice injected with HBEC3p53,KRAS,MYC (Table 1).Growth rate and tumor histology differed between tumor-igenic clones. While subcutaneous tumors were largelyundifferentiated (e.g., large cell and giant cell carcinomas),some xenograft tumors exhibited squamous and adenocar-cinoma differentiation (Table 1). This differentiation wasclone specific, as no clone differentiated into multiplehistologies.

Oncogenic transformation of HBECs increasessensitivity to standard lung cancer chemotherapiesTo determine if oncogenic transformation of HBECs

altered their sensitivity to standard lung cancer chemothera-pies, we tested 3 platinum-based doublets (paclitaxel–car-boplatin, gemcitabine–cisplatin, and pemetrexed–cisplatin)currently in use for NSCLC treatment. Oncogenic manip-ulation of HBEC3 with sh-p53 and KRASV12 resulted in asignificant increase in sensitivity to gemcitabine–cisplatinand pemetrexed–cisplatin doublet therapy in vitro (approx-imately 6- and 10-fold, respectively) but not paclitaxel–carboplatin (Supplementary Table S3 and SupplementaryFig. S5; one-way ANOVA; P < 0.001). Overall, the soft agarclones of HBEC3p53,KRAS showed sensitivities comparablewith the parentalHBEC3p53,KRAS cell line with clone 5 and 7showing intermediate sensitivity. Thus, these tumorigeni-cally progressed HBECs could provide a cell contextappropriate, isogenic model system for identifying geneticdifferences regulating sensitivity and resistance to platinumdoublet chemotherapy used in the treatment of NSCLCs.

mRNA profiles of tumorigenic versus nontumorigenicHBEC3p53,KRAS soft agar clones predict prognosis andhistology in clinical lung tumor specimensThe identification of tumorigenic and nontumorigenic

clones of HBEC3p53,KRAS that share the same geneticmanipulations (sh-p53 and KRASV12) and biologic selection(large, soft agar colonies) presents a unique cell model tocharacterize spontaneous transformation events that prog-ress HBECs to full malignancy. Biochemical assays suggestthe expression level of KRASV12 is a major contributortoward full transformation (Figs. 1C and D and 5C), where-as other tumorigenic events such as dysregulation of cell cycledid not differ between clones (Supplementary Fig. S6;summarized in Supplementary Table S2). To analyzemolec-ular differences between tumorigenic versus nontumorigenic

clones, the clones were profiled with whole-genome mRNAexpression microarrays. The mRNA expression profile ofeach clone remained stable in culture as shown by comparingmRNAs collected at a 3-week interval in unsupervisedclustering (Fig. 5D). Tumorigenic and nontumorigenicclones separated into 2 distinct clusters suggesting the clonesexhibit a strong expression profile associated with theirtumorigenic phenotype, and supervised analysis comparingtumorigenic with nontumorigenic clones found 171 uniquegenes (203 probes) that were differentially expressed [Sig-nificance Analysis of Microarrays (SAM), False DiscoveryRate (FDR) ¼ 5%; Supplementary Table S4; Fig. 5E].We tested the ability of the expression patterns of the 171

genes to predict overall survival and disease-free survival in 2independent lung tumor cohorts; the SPORE dataset ofresected early-stage NSCLCs (adeno- and squamous carci-nomas; n ¼ 209) and a second dataset of primary lungadenocarcinoma samples (n ¼ 442; NCI Director's Chal-lenge Consortium dataset). Prediction models were builtusing supervised principal component analysis and datasetswere interchanged as training and testing datasets to test forrobustness. Irrespective of dataset, the soft agar clone tumor-igenic versus nontumorigenic signature was able to identifypatients with significantly worse overall (Fig. 5F) and recur-rence-free (Supplementary Fig. S7) survival. The successfulapplication of a gene signature derived from an isogenic invitromodel of HBEC transformation in predicting outcomein clinical lung tumor samples indicates the power of such asystem and provides a preclinical model for testing thefunctional importance of the genes in the signature.

DiscussionIn this study, we sought to characterize the stepwise

progression of lung cancer pathogenesis by introducingdefined genetic manipulations commonly found in lungcancer into an in vitro HBEC model system (Fig. 6). Wefound the expression level of mutant KRAS is a criticaltransformative factor in HBECs, however, inactivation ofp53 signaling is required to evade the tumor-suppressivebarrier of RAS-induced senescence.We also show that EMTis an important oncogenic step, where c-MYC or growth inserum-containing medium both spontaneously induced anEMT and led to increased tumorigenicity. In HBECsderived from multiple individuals, we show 5 geneticchanges (hTERT andCdk4 to immortalize the cells, followedby p53 knockdown, mutant KRASV12, and c-MYC over-expression) together with serum-induced EMT are able totransform cells to a fully malignant state. Tumor xenograftsof transformedHBECswere typical of lung cancer but variedin histology, suggesting histologically distinct lung cancersfrom the central bronchus may originate from a multipotentstem-like cell. Interestingly, clonal analysis of sh-p53þKRASV12–transformed HBECs found the isogeniccells exhibited distinct phenotypes in terms of in vivotumorigenicity, xenograft histology, and drug response. ThemRNA profile that distinguished tumorigenic from non-tumorigenic clones was also able to identify a subset ofprimary lung tumor patients with significantly poorer

Oncogenic Transformation of Human Lung Epithelial Cells

www.aacrjournals.org Mol Cancer Res; 11(6) June 2013 647

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

survival. Together, this shows the use of theHBEC system asa preclinical model of lung cancer to understand transfor-mation events.Mutant KRAS lung cancers exhibit marked differences in

the expression level of KRAS. We found the expression levelof oncogenic KRASV12 was an important factor in tumor-igenic transformation of HBECs, suggesting it exerts itsoncogenic effect in a dose-dependent manner, similar toHRASV12 in a human mammary epithelial cell (HMEC)transformationmodel (18).While higher levels of oncogenicKRAS enhanced HBEC transformation, it also triggeredoncogene-induced senescence. Oncogene-induced senes-cence serves as an important tumor-suppressive barrier inresponse to persistent oncogenic insult by engaging prolif-erative arrest through p53 or p16/Rb (35). We found thatimmortalized HBECs, where the p16/Rb pathway isbypassed as a result of overexpression of CDK4, could largelyevade KRASV12-induced senescence with knockdown ofp53. Loss of p53 function, therefore, is a key step in themalignant transformation of HBECs by allowing the cells to

tolerate high levels of oncogenic KRAS, a driver ofmalignanttransformation. Transformation studies in HMECs, whichundergo spontaneous methylation-mediated p16 silencing,also report that high levels, but not low levels, of oncogenicRAS engage in senescence machinery (36, 37). However, inHMECs, HRAS-induced senescence is mediated throughTGF-b signaling in a p53-independent manner. Thus, themechanism of RAS-induced senescence differs between invitro epithelial cell models, most likely a consequence ofimmortalization methods or the cell of origin.Our data show microenvironmental signals such as those

provided by growth media can strongly influence the trans-formation ofHBECs. Switching immortalized, but nontrans-formed HBECs from defined, serum-free culture medium toserum-containing medium induced inhibition of cell growthwith induction of central airway differentiation markers.When immortalized HBEC with additional oncogenicmanipulations are switched to serum-containing medium,however, the cells are able to bypass serum-induced differen-tiation and instead become mesenchymal and more tumor-igenic, with a greater frequency of undifferentiated tumors.Overexpression of c-MYC inHBEC3p53,KRAS also induced anEMT and increased tumorigenicity. c-MYC has been shownto induce EMT inTERT-immortalizedHMECs (38). In ourstudy, c-MYC induced EMT only in the presence of p53 andKRASV12 alterations and not with c-MYC overexpressionalone (data not shown). While c-MYC overexpression orgrowth in serum-containing media both causedHBEC3p53,KRAS to undergo EMT and increase tumorigenic-ity, the presence of bothmanipulations had a synergistic effectresulting in full malignant transformation in HBECs from 2individuals. This suggests their tumor-promoting effects sig-nal through mutually exclusive pathways.In the present study, we found interindividual differences

in HBEC transformation suggesting HBECs derived fromdifferent donors can vary in their susceptibility to specificoncogene-induced malignant transformation. In terms of invitro anchorage-independent transformation, neither cellline showed anchorage-independent growth followingimmortalization (CDK4 and hTERT), yet the combinationof 5 genes (Cdk4, hTERT, sh-p53, KRASV12, and cMYC)resulted in a colony-forming efficiency of approximately60% in HBEC3 compared with less than 15% inHBEC17.These differences could potentially stem from multiplefactors including germline polymorphisms [such as single-nucleotide polymorphisms (SNP)], somatically acquiredmutations derived from either the patient (such as age orenvironmental exposures) or laboratory practices (such astime in culture), or epigenetic mechanisms all of which maypredispose the cells to oncogenic transformation. Thepatients from whom HBEC3 and HBEC17 were estab-lished, differ in respect to known risk factors for lung cancersuch as age and smoking history, and it is likely they alsodiffer in respect to unknown germline and/or somaticalterations. Thus, a comprehensive survey of genomic altera-tions present in HBECs before genetic manipulation (suchas by whole-genome sequencing) would provide betterindication of the level of existing premalignant

+cdk4+hTERT

+sh-p53

+cMYC

+KRASV12 +KRASV12

Moderate High HighModerate

Serum-induced EMT

Clonal selection

Oncogenic transformation phenotypeOncogenic transformation phenotypeImmortalized, nontransformedIn vitro transformation

Full in vivo transformationPartial in vivo transformation

EpithelialMesenchymal

Senescent

HBEC

Figure 6. Model of malignant transformation of in vitro HBECs followingstepwise introduction of common lung cancer mutations. Theexperimental data presented in this article identify the following steps:step 1, CDK4 and hTERT immortalized, HBECs are nontransformed andlack of anchorage-independent growth in soft agar; step 2, in vitrotransformation as defined by anchorage-independent growth in soft agaris achieved with the single manipulation of loss of p53, moderateKRASV12 expression or both, whreas expression of high levels ofKRASV12 expression leads to in vitro transformation with significantcellular senescence; step 3, partial in vivo transformation withsubcutaneous tumor growth in immunocompromised mice in 30% to80% of injections is observed with the combination of p53 loss and highKRASV12; step 4, an EMT occurs following overexpression of cMYC orgrowth in serum-containing media; step 5, combination of cMYCoverexpression and growth in serum-containing media results incomplete oncogenic transformation of HBECs with tumor growth in vivoobserved in more than 80% of injections in immunocompromised mice.Clonal selection of partially transformed HBECs identifies tumorigenicand nontumorigenic clones.

Sato et al.

Mol Cancer Res; 11(6) June 2013 Molecular Cancer Research648

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

susceptibilities, but these experiments are beyond the scopeof the current study.The lung can be divided into central and peripheral

compartments (39). SCCs usually arise from the centralcompartment, whereas adenocarcinomas may arise fromeither compartment, illustrating the importance of studyingoncogenic transformation in both central and peripherallung epithelial cell models. Many HBEC-derived tumors inour study were p63-positive with squamous differentiation,which likely reflects a stem/basal cell origin. A fewer numberof tumors showed distinct dual squamous and adenocarci-noma differentiation or adenocarcinoma differentiationalone. A study using SV40-immortalized tracheobronchialand small airway epithelial cells found both cell types couldbe fully transformed with oncogenic RAS (8). Another studyusing nonviral oncoproteins immortalized HSAECs withCDK4, hTERT, and a dominant-negative form of p53(p53CT) and transformed the cells using low levels ofKRASV12 plus c-MYC, PIK3CAH1047R, cyclin D1, orLKB1D194A (15). In HBECs, we have shown p53 mutationis not required for immortalization (6), and moreover, itincreases oncogenic transformation (6, 12). This disparitymay signify intrinsic differences between centrally andperipherally derived immortalized epithelial cell models.Taken together, however, our study of defined oncogenictransformation of HBECs [both in the present study andpreviously (12)] and the study in HSAECs by Sasai andcolleagues (15) largely correspond with respect oncogenictransformation of bronchial epithelial cells. We previouslyshowed transformation of HBECs with low levels ofKRASV12 (with CDK4, hTERT, and p53 knockdown)resulted in a modest increase in anchorage-independentgrowth but no tumor formation in vivo (12). Sasai andcolleagues reported that transformation of HSAECs withlow levels of KRASV12 (with CDK4, hTERT, and p53CT)failed to yield anchorage-independent or in vivo growth. Theauthors were able to fully transform HSAECs using lowlevels of KRASV12 only with additional genetic alterations.In the present study, we used a different approach byincreasing the level of KRASV12 expression to simulate thehigh amplification-associated expression often observed inlung cancer (40). We found that higher levels of theoncogene resulted in increased in vitro and in vivo transfor-mation compared with lower levels of KRASV12, whichcould be further increased with a fifth genetic alteration,such as cMYC overexpression. Thus, studies in HBECs andHSAECs show that increasing the number of oncogenicalterations will increase cellular transformation, whereasin the present study, we also show that increasing the levelof certain oncogenic alterations can also increase transfor-mation.

In conclusion, using the HBEC system as a progressionmodel of lung cancer, we were able to study early transfor-mative events in bronchial epithelial cells and the mechan-isms used to overcome tumor-suppressive barriers. We showHBECs can be transformed to full malignancy by introduc-ing defined genetic manipulations to produce histologicallysimilar lung tumors in xenografts, indicating our in vitroHBEC model retains characteristics of the tissue of origin.Furthermore, we show HBECs can model preneoplasticchanges and spontaneous transformation events such asoncogene-induced senescence and EMT and have clinicallytranslatable applications as shown in isogenic clones exhibit-ing distinct drug response and tumorigenic phenotypes.Thus, genetically defined in vitro models such as HBECswill be an invaluable tool to assess the contribution of specificgenes toward lung cancer pathogenesis, pertinent to recentwhole-genome sequencing efforts, and to screen for noveltherapeutic compounds directed at oncogenically acquired,tumor-specific vulnerabilities.

Disclosure of Potential Conflicts of InterestJ.D. Minna has ownership interest (including patents) in NCI and University of

Texas Southwestern Medical Center. No potential conflicts of interest were disclosedby the other authors.

Authors' ContributionsConception and design: M. Sato, J.E. Larsen, A.F. Gazdar, J.W. Shay, J.D. MinnaDevelopment of methodology:M. Sato, J.E. Larsen, R.D. Ramirez, A.F. Gazdar, J.W. ShayAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.):M. Sato, J.E. Larsen, W. Lee, M.P. Dalvi, J.M. DiMaio, I.I. Wistuba,A.F. GazdarAnalysis and interpretation of data (e.g., statistical analysis, biostatistics, compu-tational analysis):M. Sato, J.E. Larsen, H. Sun, D.S. Shames, M.P. Dalvi, H. Tang,B. Gao, Y. Xie, I.I. Wistuba, A.F. Gazdar, J.W. Shay, J.D. MinnaWriting, review, and/or revision of the manuscript: M. Sato, J.E. Larsen, D.S.Shames, R.D. Ramirez, B. Gao, Y. Xie, I.I. Wistuba, A.F. Gazdar, J.W. Shay, J.D.MinnaAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): J.E. Larsen, A.F. GazdarStudy supervision: J.E. Larsen, J.W. Shay, J.D. Minna

AcknowledgmentsThe authors thank Natasha Rekhtman (Memorial Sloan-Kettering Cancer Center,

New York, NY) for her kind assistance in analyzing immunohistochemical sections.The authors also thank themany current and pastmembers of theMinna laboratory fortheir technical assistance and article comments, particularly Luc Girard, Suzie Hight,and Michael Peyton.

Grant SupportThis workwas supported by theNCI LungCancer Specialized Program ofResearch

Excellence (SPORE; P50CA70907), NASA NSCOR (NNJ05HD36G). J.E. Larsenwas supported by NHMRC Biomedical Fellowship (494511) and TSANZ/Allen &Hanburys Respiratory Research Fellowship.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be herebymarked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Received November 6, 2012; revised January 24, 2013; accepted January 25, 2013;published OnlineFirst February 28, 2013.

References1. Larsen JE, Minna JD. Molecular biology of lung cancer: clinical impli-

cations. Clin Chest Med 2011;32:703–40.2. 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.

3. 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.

4. Rudin CM, Durinck S, Stawiski EW, Poirier JT, Modrusan Z, ShamesDS, et al. Comprehensive genomic analysis identifies SOX2 as a

Oncogenic Transformation of Human Lung Epithelial Cells

www.aacrjournals.org Mol Cancer Res; 11(6) June 2013 649

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

frequently amplified gene in small-cell lung cancer. Nat Genet2012;44:1111–6.

5. Govindan R, Ding L, Griffith M, Subramanian J, Dees ND, Kanchi KL,et al. Genomic landscape of non–small cell lung cancer in smokers andnever-smokers. Cell 2012;150:1121–34.

6. Ramirez RD, Sheridan S, Girard L, Sato M, Kim Y, Pollack J, et al.Immortalization of human bronchial epithelial cells in the absence ofviral oncoproteins. Cancer Res 2004;64:9027–34.

7. Reddel RR, KeY,Gerwin BI,McMenaminMG, Lechner JF, SuRT, et al.Transformation of human bronchial epithelial cells by infection withSV40 or adenovirus-12 SV40 hybrid virus, or transfection via strontiumphosphate coprecipitationwith aplasmid containingSV40 early regiongenes. Cancer Res 1988;48:1904–9.

8. 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.

9. Cheng J, DeCaprio JA, Fluck MM, Schaffhausen BS. Cellular trans-formation by Simian virus 40 and murine polyoma virus T antigens.Semin Cancer Biol 2009;19:218–28.

10. Arroyo JD, Hahn WC. Involvement of PP2A in viral and cellular trans-formation. Oncogene 2005;24:7746–55.

11. Kendall SD, Linardic CM, Adam SJ, Counter CM. A network of geneticevents sufficient to convert normal human cells to a tumorigenic state.Cancer Res 2005;65:9824–8.

12. Sato M, Vaughan MB, Girard L, Peyton M, Lee W, Shames DS, et al.Multiple oncogenic changes (K-RAS(V12), p53 knockdown, mutantEGFRs, p16 bypass, telomerase) are not sufficient to confer a fullmalignant phenotype on human bronchial epithelial cells. Cancer Res2006;66:2116–28.

13. Vaughan MB, Ramirez RD, Wright WE, Minna JD, Shay JW. A three-dimensional model of differentiation of immortalized human bronchialepithelial cells. Differentiation 2006;74:141–8.

14. Delgado O, Kaisani AA, Spinola M, Xie XJ, Batten KG, Minna JD, et al.Multipotent capacity of immortalized human bronchial epithelial cells.PLoS ONE 2011;6:e22023.

15. Sasai K, Sukezane T, Yanagita E, Nakagawa H, Hotta A, Itoh T, et al.Oncogene-mediated human lung epithelial cell transformation pro-duces adenocarcinoma phenotypes in vivo. Cancer Res 2011;71:2541–9.

16. Toyooka S, Tsuda T, Gazdar AF. The TP53 gene, tobacco exposure,and lung cancer. Hum Mutat 2003;21:229–39.

17. Sato M, Shames DS, Gazdar AF, Minna JD. A translational view ofthe molecular pathogenesis of lung cancer. J Thorac Oncol 2007;2:327–43.

18. Elenbaas B, Spirio L, Koerner F, Fleming MD, Zimonjic DB, DonaherJL, et al. Human breast cancer cells generated by oncogenic trans-formation of primary mammary epithelial cells. Genes Dev 2001;15:50–65.

19. SerranoM,LinAW,McCurrachME,BeachD, LoweSW.Oncogenic rasprovokes premature cell senescence associated with accumulation ofp53 and p16INK4a. Cell 1997;88:593–602.

20. Heighway J, HasletonPS. c-Ki-ras amplification in human lung cancer.Br J Cancer 1986;53:285–7.

21. Gazdar AF, Girard L, LockwoodWW, LamWL, Minna JD. Lung cancercell lines as tools for biomedical discovery and research. J Natl CancerInst 2010;102:1310–21.

22. Phelps RM, Johnson BE, Ihde DC, Gazdar AF, Carbone DP, McClin-tock PR, et al. NCI-Navy Medical Oncology Branch cell line data base.J Cell Biochem Suppl 1996;24:32–91.

23. Vikis H, Sato M, James M, Wang D, Wang Y, Wang M, et al. EGFR-T790M is a rare lung cancer susceptibility allele with enhanced kinaseactivity. Cancer Res 2007;67:4665–70.

24. Ihle NT, Byers LA, KimES, Saintigny P, Lee JJ, BlumenscheinGR, et al.Effect of KRAS oncogene substitutions on protein behavior: implica-tions for signaling and clinical outcome. J Natl Cancer Inst 2012;104:228–39.

25. Louro ID, Bailey EC, Li X, South LS, McKie-Bell PR, Yoder BK, et al.Comparative gene expression profile analysis of GLI and c-MYC in anepithelial model of malignant transformation. Cancer Res 2002;62:5867–73.

26. Sato M, Girard L, Sekine I, Sunaga N, Ramirez RD, Kamibayashi C,et al. Increased expression and no mutation of the Flap endonu-clease (FEN1) gene in human lung cancer. Oncogene 2003;22:7243–6.

27. Sato M, Sekido Y, Horio Y, Takahashi M, Saito H, Minna JD, et al.Infrequent mutation of the hBUB1 and hBUBR1 genes in human lungcancer. Jpn J Cancer Res 2000;91:504–9.

28. Sullivan JP, Spinola M, Dodge M, Raso MG, Behrens C, Gao B, et al.Aldehyde dehydrogenase activity selects for lung adenocarcinomastem cells dependent on notch signaling. Cancer Res 2010;70:9937–48.

29. GreerRM,PeytonM, Larsen JE,Girard L, XieY,GazdarAF, et al. SMACmimetic (JP1201) sensitizes non–small cell lung cancers to multiplechemotherapy agents in an IAP-dependent but TNF-alpha-indepen-dent manner. Cancer Res 2011;71:7640–8.

30. 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.

31. Hahn WC, Counter CM, Lundberg AS, Beijersbergen RL, Brooks MW,Weinberg RA. Creation of human tumour cells with defined geneticelements. Nature 1999;400:464–8.

32. Boehm JS, Hession MT, Bulmer SE, Hahn WC. Transformation ofhuman andmurine fibroblasts without viral oncoproteins. Mol Cell Biol2005;25:6464–74.

33. Campbell PM, Groehler AL, Lee KM, Ouellette MM, Khazak V, Der CJ.K-Ras promotes growth transformation and invasion of immortalizedhuman pancreatic cells by Raf and phosphatidylinositol 3-kinasesignaling. Cancer Res 2007;67:2098–106.

34. Masui T,Wakefield LM, Lechner JF, LaVeckMA, SpornMB, Harris CC.Type beta transforming growth factor is the primary differentiation-inducing serum factor for normal human bronchial epithelial cells. ProcNatl Acad Sci U S A 1986;83:2438–42.

35. Courtois-Cox S, Jones SL, Cichowski K. Many roads lead to onco-gene-induced senescence. Oncogene 2008;27:2801–9.

36. Cipriano R, Kan CE, Graham J, Danielpour D, Stampfer M, JacksonMW. TGF-beta signaling engages an ATM-CHK2-p53-independentRAS-induced senescence and prevents malignant transformation inhuman mammary epithelial cells. Proc Natl Acad Sci U S A2011;108:8668–73.

37. Lin S, Yang J, Elkahloun AG, Bandyopadhyay A, Wang L, Cornell JE,et al. Attenuation of TGF-beta signaling suppresses premature senes-cence in a p21-dependent manner and promotes oncogenic Ras-mediated metastatic transformation in human mammary epithelialcells. Mol Biol Cell 2012;23:1569–81.

38. Cowling VH, Cole MD. E-cadherin repression contributes toc-Myc-induced epithelial cell transformation. Oncogene 2007;26:3582–6.

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

40. Sunaga N, Kaira K, Imai H, Shimizu K, Nakano T, Shames DS, et al.Oncogenic KRAS-induced epiregulin overexpression contributes toaggressive phenotype and is a promising therapeutic target in non–small-cell lung cancer. Oncogene. 2012 Sep 10. [Epub ahead ofprint].

Sato et al.

Mol Cancer Res; 11(6) June 2013 Molecular Cancer Research650

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T

2013;11:638-650. Published OnlineFirst February 28, 2013.Mol Cancer Res   Mitsuo Sato, Jill E. Larsen, Woochang Lee, et al.   Specific Oncogenic ManipulationsHuman Lung Epithelial Cells Progressed to Malignancy through

  Updated version

  10.1158/1541-7786.MCR-12-0634-Tdoi:

Access the most recent version of this article at:

  Material

Supplementary

  http://mcr.aacrjournals.org/content/suppl/2013/02/28/1541-7786.MCR-12-0634-T.DC1

Access the most recent supplemental material at:

   

  Overview

Visual 

http://mcr.aacrjournals.org/content/11/6/638/F1.large.jpgA diagrammatic summary of the major findings and biological implications:

   

  Cited articles

  http://mcr.aacrjournals.org/content/11/6/638.full#ref-list-1

This article cites 39 articles, 15 of which you can access for free at:

  Citing articles

  http://mcr.aacrjournals.org/content/11/6/638.full#related-urls

This article has been cited by 27 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

  .pubs@aacr.org

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://mcr.aacrjournals.org/content/11/6/638To request permission to re-use all or part of this article, use this link

on October 9, 2020. © 2013 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2013; DOI: 10.1158/1541-7786.MCR-12-0634-T