depletion of foxm1 via met targeting underlies establishment...

16
Cancer Therapy: Preclinical Depletion of FOXM1 via MET Targeting Underlies Establishment of a DNA DamageInduced Senescence Program in Gastric Cancer Paola Francica 1,2 , Lluís Nisa 1,2 , Daniel M. Aebersold 1,2 , Rupert Langer 3 , Friedhelm Bladt 4 , Andree Blaukat 4 , Deborah Stroka 2,5 , María Rodríguez Martínez 6 , Yitzhak Zimmer 1,2 , and Michaela Medov a 1,2 Abstract Purpose: Deregulated signaling via the MET receptor tyrosine kinase is abundant in gastric tumors, with up to 80% of cases displaying aberrant MET expression. A growing body of evidence suggests MET as a potential target for tumor radiosensitization. Experimental Design: Cellular proliferation and DNA damage-induced senescence were studied in a panel of MET- overexpressing human gastric cancer cell lines as well as in xenograft models after MET inhibition and/or ionizing radia- tion. Pathways activation and protein expression were assessed by immunoblotting and immunohistochemistry. Tumor tissue microarrays (91 gastric cancer patients) were generated and copy number alteration (178 patients) and gene expression (373 patients) data available at The Cancer Genome Atlas were analyzed to assess the coalterations of MET and FOXM1. Results: MET targeting administered before ionizing radiation instigates DNA damageinduced senescence (80%, P < 0.001) rather than cell death. MET inhibitionassociated senescence is linked to the blockade of MAPK pathway, correlates with down- regulation of FOXM1, and can be abrogated (11.8% vs. 95.3%, P < 0.001) by ectopic expression of FOXM1 in the corresponding gastric tumor cells. Cells with ectopic FOXM1 expression dem- onstrate considerable (20%, P < 0.001) growth advantage despite MET targeting, suggesting a novel clinically relevant resis- tance mechanism to MET inhibition as the copresence of both MET and FOXM1 protein (33%) and mRNA (30%) overexpres- sion as well as gene amplication (24,7%) are common in patients with gastric cancer. Conclusions: FOXM1, a negative regulator of senescence, has been identied as a key downstream effector and poten- tial clinical biomarker that mediates MET signaling follow- ing iniction of DNA damage in gastric tumors. Clin Cancer Res; 115. Ó2016 AACR. Introduction Signaling networks downstream of MET that drive normal physiologic processes are exploited in cancer cells to promote key features such as tumor uncontrolled growth, survival, angiogenesis, local invasion, and systemic dissemination (1). Dysregulation of the MET/hepatocyte growth factor (HGF) axis via establishment of ligandreceptor autocrine loops, presence of activating point mutations and most frequently via receptor overexpression have emerged as common events of many human malignancies (1, 2). In that respect, MET overexpres- sion has been described in various solid tumors including colorectal, hepatocellular, ovarian, pancreatic, prostate, and gastric carcinomas (2). Gastric cancer remains the fth leading cause of cancer- related death worldwide (3). Most patients are diagnosed with locoregionally advanced or metastatic disease, and despite improvements in surgical and multimodal approaches com- bining chemotherapy and radiation, patients with nonresect- able tumors have a median survival of only 10 months (4). In this respect, approaches targeting oncogenic drivers of gastric tumors such as HER2 (overexpressed in 10%15% of cases) have emerged as promising treatment strategies (5). MET over- expression, which is reported in 18%82% of gastric cancer cases (6), correlates with advanced-stage tumors and poor prognosis (6, 7). In a subset of patients, MET overexpression is associated with increased MET copy number due to gene amplication (8, 9). In addition, about 20% of human gastric cancer cell lines exhibit MET amplication with a correspond- ing MET addiction phenotype, rendering these tumor cells extremely sensitive to MET inhibition (10). Accordingly, MET is considered to be a prognostic marker and at the same time an important therapeutic target for locally advanced or metastatic gastric cancers (11, 12). Currently, several anti-MET antibodies 1 Department of Radiation Oncology, Inselspital, Bern University Hos- pital and University of Bern, Bern, Switzerland. 2 Department of Clinical Research, University of Bern, Bern, Switzerland. 3 Institute of Pathol- ogy, University of Bern, Bern, Switzerland. 4 Merck Serono Research & Development, Merck KGaA, Darmstadt, Germany. 5 Department of Visceral Surgery, Inselspital, Bern University Hospital and University of Bern, Bern, Switzerland. 6 IBM Research, Zurich Research Labora- tory, Zurich, Switzerland. Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). Current address for F. Bladt: Bayer Pharma AG, Global Drug Discovery Clinical Sciences, Experimental Medicine ONC, Berlin, Germany. Corresponding Author: Michaela Medov a, Department of Radiation Oncology, Inselspital, Bern University Hospital, Bern University Hospital, and University of Bern, Murtenstrasse 35, Bern 3008, Switzerland. Phone: 413-1632-3565. Fax: 413-1632-3297; E-mail: [email protected] doi: 10.1158/1078-0432.CCR-15-2987 Ó2016 American Association for Cancer Research. Clinical Cancer Research www.aacrjournals.org OF1 Cancer Research. on February 15, 2019. © 2016 American Association for clincancerres.aacrjournals.org Downloaded from Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Upload: lamkhanh

Post on 16-Feb-2019

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

Cancer Therapy: Preclinical

Depletion of FOXM1 via MET Targeting UnderliesEstablishment of a DNA Damage–InducedSenescence Program in Gastric CancerPaola Francica1,2, Lluís Nisa1,2, Daniel M. Aebersold1,2, Rupert Langer3,Friedhelm Bladt4, Andree Blaukat4, Deborah Stroka2,5, María Rodríguez Martínez6,Yitzhak Zimmer1,2, and Michaela Medov�a1,2

Abstract

Purpose: Deregulated signaling via the MET receptor tyrosinekinase is abundant in gastric tumors, with up to 80% of casesdisplaying aberrant MET expression. A growing body of evidencesuggests MET as a potential target for tumor radiosensitization.

Experimental Design: Cellular proliferation and DNAdamage-induced senescence were studied in a panel of MET-overexpressing human gastric cancer cell lines as well as inxenograft models after MET inhibition and/or ionizing radia-tion. Pathways activation and protein expression were assessedby immunoblotting and immunohistochemistry. Tumor tissuemicroarrays (91 gastric cancer patients) were generated andcopy number alteration (178 patients) and gene expression(373 patients) data available at The Cancer Genome Atlas wereanalyzed to assess the coalterations of MET and FOXM1.

Results: MET targeting administered before ionizing radiationinstigates DNA damage–induced senescence (�80%, P < 0.001)

rather than cell death. MET inhibition–associated senescence islinked to the blockade of MAPK pathway, correlates with down-regulation of FOXM1, and canbe abrogated (11.8%vs. 95.3%, P<0.001) by ectopic expression of FOXM1 in the correspondinggastric tumor cells. Cells with ectopic FOXM1 expression dem-onstrate considerable (�20%, P < 0.001) growth advantagedespite MET targeting, suggesting a novel clinically relevant resis-tance mechanism to MET inhibition as the copresence of bothMET and FOXM1 protein (33%) and mRNA (30%) overexpres-sion as well as gene amplification (24,7%) are common inpatients with gastric cancer.

Conclusions: FOXM1, a negative regulator of senescence,has been identified as a key downstream effector and poten-tial clinical biomarker that mediates MET signaling follow-ing infliction of DNA damage in gastric tumors. Clin Cancer Res;1–15. �2016 AACR.

IntroductionSignaling networks downstream of MET that drive normal

physiologic processes are exploited in cancer cells to promotekey features such as tumor uncontrolled growth, survival,angiogenesis, local invasion, and systemic dissemination (1).Dysregulation of the MET/hepatocyte growth factor (HGF) axisvia establishment of ligand–receptor autocrine loops, presence

of activating point mutations and most frequently via receptoroverexpression have emerged as common events of manyhuman malignancies (1, 2). In that respect, MET overexpres-sion has been described in various solid tumors includingcolorectal, hepatocellular, ovarian, pancreatic, prostate, andgastric carcinomas (2).

Gastric cancer remains the fifth leading cause of cancer-related death worldwide (3). Most patients are diagnosed withlocoregionally advanced or metastatic disease, and despiteimprovements in surgical and multimodal approaches com-bining chemotherapy and radiation, patients with nonresect-able tumors have a median survival of only 10 months (4). Inthis respect, approaches targeting oncogenic drivers of gastrictumors such as HER2 (overexpressed in 10%–15% of cases)have emerged as promising treatment strategies (5). MET over-expression, which is reported in 18%–82% of gastric cancercases (6), correlates with advanced-stage tumors and poorprognosis (6, 7). In a subset of patients, MET overexpressionis associated with increased MET copy number due to geneamplification (8, 9). In addition, about 20% of human gastriccancer cell lines exhibit MET amplification with a correspond-ing MET addiction phenotype, rendering these tumor cellsextremely sensitive to MET inhibition (10). Accordingly, METis considered to be a prognostic marker and at the same time animportant therapeutic target for locally advanced or metastaticgastric cancers (11, 12). Currently, several anti-MET antibodies

1Department of Radiation Oncology, Inselspital, Bern University Hos-pital andUniversityofBern,Bern, Switzerland. 2DepartmentofClinicalResearch, University of Bern, Bern, Switzerland. 3Institute of Pathol-ogy, University of Bern, Bern, Switzerland. 4Merck Serono Research &Development, Merck KGaA, Darmstadt, Germany. 5Department ofVisceral Surgery, Inselspital, Bern University Hospital and Universityof Bern, Bern, Switzerland. 6IBM Research, Zurich Research Labora-tory, Zurich, Switzerland.

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

Current address for F. Bladt: Bayer Pharma AG, Global Drug Discovery – ClinicalSciences, Experimental Medicine ONC, Berlin, Germany.

Corresponding Author: Michaela Medov�a, Department of Radiation Oncology,Inselspital, Bern University Hospital, Bern University Hospital, and University ofBern, Murtenstrasse 35, Bern 3008, Switzerland. Phone: 413-1632-3565. Fax:413-1632-3297; E-mail: [email protected]

doi: 10.1158/1078-0432.CCR-15-2987

�2016 American Association for Cancer Research.

ClinicalCancerResearch

www.aacrjournals.org OF1

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 2: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

and small-molecule MET inhibitors are undergoing phase I andphase II clinical trials in different types of solid tumors,including gastric cancer (e.g., NCT02435108, NCT01468922,NCT00725712, NCT02002416), with only a modest benefitin otherwise unselected patient populations (13).

A growing body of evidence suggests MET as a potentialtarget for tumor sensitization to DNA-damaging agents (DDA),such as ionizing radiation (IR), which are commonly used incancer therapy (14–16). In 2011, Medov�a and colleagues (17)reported that inhibition of the MET receptor tyrosine kinase(RTK) impairs the ability of irradiated MET-overexpressinggastric cancer cell lines to properly execute DNA double strandbreaks (DSB) repair. It is well established that persisting DNAdamage may trigger a broad range of different outcomes onaffected cells including apoptosis, necrosis, mitotic catastrophe,or senescence (18, 19).

Cellular senescence was first described in a seminal study byHayflick and Moorhead in 1961 (20) as a limit to the replicativelifespan of somatic cells after serial cultivation in vitro. Morerecently, it was shown that senescence can be triggered also byother stimuli such as oncogene expression, oxidative stress, andDNA damage (18, 21). This premature senescence seems to play acritical role in tumor suppression by preventing cancer initiationand progression (22) and often accompanies apoptotic responseelicited by chemotherapy or IR (23, 24).

FOXM1, a member of the Forkhead box (FOX) superfamilyof transcription factors (25), is ubiquitously expressed inactively proliferating tissues and plays an essential role in theregulation of a wide spectrum of biologic processes (26).Elevated expression of FOXM1 has been detected in a broadrange of malignancies, including gastric cancers, and has beenshown to correlate with tumor progression and poor progno-sis (27). In recent years, numerous studies have reportedFOXM1 playing a crucial role in both homologous recombi-

nation repair of DNA DSBs (28, 29) and negative regulationof the senescence program (29, 30), significantly turningresearch interests toward FOXM1 as a potential target in cancertherapy.

In this study, we investigated the impact of MET inhibitionon irradiated gastric cancer models both in vitro and in vivo. Ourdata show that MET inhibition sensitizes human gastric cancercells to IR by promoting DNA damage–induced senescence,conceivably via interfering with the MAPK pathway. We reportthat impairment of the MET signaling cascade leads to accu-mulation of unrepaired DNA DSBs in parallel to a downregula-tion of FOXM1 protein levels, providing thereby for the firsttime evidence for a role for FOXM1 downstream of METsignaling. Although the regulatory mechanisms of FOXM1 withrespect to cancer development and progression are still poorlyunderstood, our current findings provide important mechanis-tic insights into the role of FOXM1 in DNA damage–inducedsenescence following MET inhibition. Furthermore, assessmentof MET and FOXM1 protein levels in gastric cancer tissuemicroarrays (TMA) and analysis of the TCGA database showthat coalteration of MET and FOXM1 is common in patientswith gastric cancer. These observations may have importanttherapeutic translational implications for the treatment ofgastric cancer through MET targeting.

Materials and MethodsCell lines

Human gastric carcinoma cells GTL-16, MKN-45, and SNU-5were provided by the laboratory of Dr. Paolo Comoglio (Med-ical School University of Torino, Torino, Italy). GTL-16 cellswere grown in RPMI medium (Gibco, Invitrogen) supplementedwith 5% FCS (Sigma) and antibiotic–antimycotic (penicillin100 U/mL, streptomycin sulfate 100 U/mL, amphotericin B asfungizone 0.25 mg/mL; Gibco). MKN-45 cells were maintainedin DMEM (Gibco, Invitrogen) with 10% FCS and antibiotic–antimycotic, while SNU-5 were grown in Iscove's modifiedDulbecco's medium with 20% FCS and antibiotic–antimycotic.SNU-638 and KATO-II gastric carcinoma cells [(from KoreanCell Line Bank and Dr. Morag Park (McGill University, Mon-treal, QC, Canada), respectively] were maintained in RPMI with10% FCS. All cell lines were obtained within the last 5 years. Nocell line authentication was performed by the authors for thisstudy. The SNU638_FOXM1 and GTL16_FOXM1 cell lines wereestablished by transfecting SNU638 and GTL16 cells with a Flag-FoxM1-NT2 expression vector (kindly provided by Prof. Ren�eMedema, Netherlands Cancer Institute, Amsterdam, the Nether-lands) that contains a geneticin selection marker. Cells werethen selected and maintained in presence of 500 mg/mL ofgeneticin (Fisher Scientific).

shRNA lentiviral transductionStable transduction with lentiviral vectors for scrambled

control and p53 short-hairpin RNA (shRNA) were describedpreviously (31).

TreatmentsTepotinib (MSC2156119, EMD1214063; Merck), AZD6244

(Selleck Chemicals), and LY294002 (Sigma) were dissolvedin DMSO. Cells were irradiated using Gammacell 40 (MDSNordion), and animals using XStrahl 150 (Xstrahl Limited).

Translational Relevance

It is widely accepted that cellular senescence may func-tion as a fundamental barrier of tumor progression.Therefore, the significance of senescence and underlyingmechanisms for particular clinical interventions, includ-ing tyrosine kinase targeting–based modalities, need to bewell defined. Interestingly, despite numerous studies eval-uating efficacies of MET-targeting strategies in both pre-clinical and clinical settings, the relevance of MET inhibi-tion–associated senescence in MET-positive tumorsremains still largely uncharacterized. Here we demonstratethat induction of a senescence program is the majorconsequence in MET-driven models of gastric cancer afterexposure to a combined regimen of the MET inhibitortepotinib and irradiation. Mechanistically, the work iden-tifies FOXM1 as a critical negative regulator of senescencethat directly associates to MET targeting–related DNAdamage. Furthermore, the work demonstrates coaberrantexpression of MET and FOXM1 in gastric cancer patientssamples and suggests FOXM1 expression level as a novelpotential marker for patients stratification for MET-basedtherapies.

Francica et al.

Clin Cancer Res; 2016 Clinical Cancer ResearchOF2

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 3: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

Western blotting and antibodiesCells were lysed [20 mmol/L HEPES (pH 8.0), 9.0 mol/L

urea, 1 mmol/L sodium orthovanadate, 2.5 mmol/L sodiumpyrophosphate, 1 mmol/L b-glycerol-phosphate] and sonicat-ed. Protein concentration was determined (Bio-Rad Laborato-ries, Inc.), proteins were resolved by SDS-PAGE, transferredonto polyvinylidene difluoride membranes, and followed byincubation with primary [rabbit phospho-MET (Tyr1234/1235), MET, Ki-67, phospho-Akt (Ser473), phospho-Erk1/2(Thr202/Tyr204), phospho-p53 (Ser15), p53, BRIP1 andBMI-1 from Cell Signaling Technology; phospho-HistoneH2AX, p21, Histone H3 (tri methyl K9), and b-actin fromMillipore Corporation; FOXM1 and p16 from Santa CruzBiotechnology] and secondary antibodies.

Cell viability/toxicity assayCell death and viability were assessed by the Live/Dead Assay

Kit (Molecular Probes). Briefly, cells were stained with greenfluorescent calcein-AM (viable cells) and red fluorescent ethidiumhomodimer-1 (dead cells). Images were analyzed with a fluores-cence microscope (Leica) at 10� magnification.

Senescence-associated b-galactosidase assayCells were seeded in 6-well plates and treated after 24 hours.

Staining for senescence-associated b-galactosidase was done asdescribed previously (32) after 7 days of treatment. Imageswere obtained with a Leica inverted microscope at 40� mag-nification. For the in vivo study, frozen sections of xenografttumors (5 mm) were fixed in 2% glutaraldehyde and stained asthe in vitro experiments and then counterstained with hema-toxylin for 1 minute.

BrdUrd incorporation assayBromodeoxyuridine (BrdUrd) incorporation assay was per-

formed in vitro using the BD Pharmingen BrdU Flow Kit (BDBiosciences). Briefly, cells were seeded in 6-well plates and treatedafter 24 hours as indicated for 3 days. BrdUrd at a final concen-tration of 10 mmol/L was added overnight to the cells beforeperforming the staining protocol. Sampleswere acquired on LSRIIflow cytometer (Beckton Dickinson) and analyzed using theFlowJo software.

In vivo modelGTL-16 and SNU-638 xenografts were grown in 12-week-old

female Rag2�/�gc�/� mice. A total of 106 cells were injectedsubcutaneously in the right flank of each mouse and the growthof the tumors was regularly monitored by caliper measure-ments. Once tumors had reached a size of 300 mm3 (day 0),mice were randomly divided in the following treatment groups:vehicle control (Solutol HS 15, BASF Chem-Trade GmbH),tepotinib only (15 mg/kg/d orally, days 1–6), radiation only(one single 6 Gy radiation dose, on day 3), and tepotinib withradiation. Tumors were harvested 24 hours after the last tepo-tinib treatment (on day 7), frozen in Tissue-Tek optimumcutting temperature medium (OCT; Sakura Finetek Europe B.V.) and stored at �20�C.

Immunohistochemical analysisTissue samples were cut into 5-mm sections with a Leica Micro-

tome and fixed in 4% paraformaldehyde (PFA) for 10 minutes.Sections were washed with 1� PBS-Tween 0.1% (PBST) and

antigen retrieval was performed in 1� PBS-Triton-X 0.1%(FLUKA) for 10 minutes. Endogenous peroxidase was blockedwith H2O2 0.1% in H2O for 10 minutes. Sections were blockedwith 1% normal goat serum in PBS for 1 hour.

Primary rabbit antibodies were incubated at room temper-ature in blocking solution overnight. Sections were washed inPBST and incubated with secondary antibody at room tem-perature for 1 hour. Signal was detected using the VectastainABC Kit (Vector Laboratories) and 3,30-Diaminobenzidine(DAB; Sigma-Aldrich). DAB incubation varied between 3 and10 minutes and was rinsed with distilled H2O before counter-staining with hematoxylin for 1 minute. Dehydrated sectionswere mounted using Eukitt (Kindler, GmbH). Images wereobtained with a Leica DMRB microscope (20� and 40�magnifications).

XTT assayCells were plated into a 96-well plate in 50 mL of the corre-

sponding medium (six wells per treatment condition) and thenext day treated with tepotinib. Seventy-two hours later, cellularproliferation was measured using the commercially available kit(based on the cleavage of the yellow tetrazolium salt XTT to forman orange formazan dye bymetabolic active cells; Roche). Briefly,at the end of treatment, cells were incubated with the XTT reagentat 37�C, 5% CO2 for 2 hours; the resulting formazan was quan-tified by measuring the absorbance at 490 nm (reference wave-length: 655 nm) using a Tecan Platereader.

Tumor tissue microarrays and IHCTumor tissue microarray (TMA) was built using a next-gener-

ation tissue microarray, and hematoxylin and eosin (H&E) slidesof all samples were scanned for morphologic assessment by aboard-certified pathologist (R. Langer). In addition, a commer-cially available TMA of gastric cancer patients was purchasedfrom Lucerna Chem.

TMA sections were deparaffinized and rehydrated. Antigenretrieval was performed by heating the slides for 20–30 min ina microwave oven (600–700 W) in either 0.01 mol/L sodiumcitrate or Tris-EDTA (pH 9.0). Sections were incubated overnightat 4�C with primary antibodies (all from Cell Signaling Techno-logy): Ki-67 (1:400; D2H10), MET (1:150; D26), and FOXM1(1:600; D1205).

Stained TMAs were scored semiquantitatively by threeindependent observers blinded to each other's results (grading:1 ¼ negative/low; 2 ¼ intermediate; 3 ¼ high).

Extraction and analysis of data from The Cancer GenomeAtlas database

RNA sequencing (RNASeq) data [297 patients sequenced inIllumina HiSeq 2000 (HiSeq) platform] was normalized accord-ing to the trimmed mean of M-values normalization methodusing the edgeR package (33).

Copy number variation (CNV) data were available from twodifferent platforms, SNP arrays and Low Pass DNASeq, with443 and 107 patients profiled on each technology, respective-ly. To avoid platform biases, we only selected data from SNParrays for subsequent analysis. We extracted CNVs that over-lapped with MET or FOXM1. Genes start and end positionswere from the UCSC Known Genes dataset, GRCh37/hg19-Feb2009 genome reference. Transcription start and end positionswere augmented by 5 kb to account for promoter and

MET, FOXM1, and DNA Damage–Induced Senescence

www.aacrjournals.org Clin Cancer Res; 2016 OF3

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 4: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

enhancers. If two or more CNVs were present in a gene in thesame sample, we only considered the CNV with highestabsolute value.

After preprocessing, RNASeq,CNV, andprotein levels datawerez-transformed,

z ¼ x�mean xð Þð Þ=std xð Þ:

We used a threshold of �1.5 on the z-scores to consider asample overexpressed or downregulated with respect to the pop-ulation, or, respectively, amplified or deleted.

Statistical analysisStatistical analysis was performed with GraphPad Prism Soft-

ware using the one-way ANOVA or Student t test as indicated inthe figure legends.

ResultsMET inhibition along with IR inflicts DNA damage–inducedsenescence rather than cell death in MET-overexpressinghuman gastric cancer cell lines

As activation of the MET receptor was reported to protectcancer cells from IR-induced cell death by enhancing DNAdamage repair activity (14, 17), here we examined the effectof MET inhibition by the small-molecule tyrosine kinase inhib-itor tepotinib (MSC2156119; Merck) alone or along with IR inMET-overexpressing human gastric cancer cell lines GTL-16,MKN-45, SNU-638, SNU-5, and KATO-II. Importantly, weobserved that exposure of cells to low doses of tepotinib aloneor in combination with IR resulted in a significant decrease incell proliferation as indicated by the loss of incorporation ofBrdUrd (Fig. 1A and Supplementary Fig. S1A and S1B). Sur-prisingly, the treatment did not induce massive cell death(Fig. 1B and Supplementary Fig. S1C and S1D). Accumulatingevidence suggests that tumor cell response to traditional cyto-toxic chemotherapy, which aims at inducing extensive DNAdamage, is not strictly confined to cell death but may also leadto the induction of cell senescence (24, 34, 35).

As irreversible cell-cycle arrest is a feature of cellular senes-cence, the panel of human gastric cancer cell lines was exam-ined for senescence-associated b-galactosidase (SA-b-Gal) activ-ity, the most widely used senescence biomarker. In all thestudied cell lines, MET inhibition enhanced SA-b-Gal stainingparticularly in combination with IR and notably, this wasaccompanied by characteristic morphologic features of cellularsenescence, for example, enlarged and flattened cells (Fig. 1Cand Supplementary Fig. S1E).

The execution of senescence programs is also associated withformation of senescence-associated heterochromatic foci(SAHF) enriched in trimethylated histone H3 (H3K9me3) andwith increased levels of DNA damage markers such as Ser139phosphorylated histone variant H2AX (gH2AX). We observedthat tepotinib treatment led to increased expression of bothH3K9me3 and gH2AX especially in combination with IR(Fig. 1D and Supplementary Fig. S1F).

MET inhibition promotes DNA damage–induced senescencein vivo

The significance of the in vitro observations was furthervalidated in vivo by establishing two independent xenografttumor models with mice bearing subcutaneous GTL-16

or SNU-638 tumors. As shown in Fig. 1E and F and Supple-mentary Fig. S2A and S2B, MET inhibition (15 mg/kg/day)caused a substantial decrease in tumor cell proliferation, asindicated by reduction in Ki-67 staining and tumor size inboth GTL-16 and SNU-638 xenografts treated with tepotinib.Moreover, our data demonstrate the ability of MET inhibitionto massively induce DNA damage in vivo: tissues derived frommice treated with tepotinib alone and particularly along withIR displayed enhanced gH2AX staining compared withuntreated controls (Fig. 1E and F). Decrease in cell prolifer-ation and elevated gH2AX levels coincided also with anincrease in SA-b-Gal activity when tepotinib was combinedwith IR (Fig. 1E and F), further indicating that inhibition ofthe MET RTK along with IR leads to DNA damage–inducedsenescence in vivo.

Characterization of pathways involved in the execution ofthe senescence program

The senescence program is predominantly established andmaintained by the p53 and the p16INK4A-RB signal transductionpathways (36). Importantly, the relative preponderance of eachpathway in the execution of senescence is variable and dependenton cell type and species (36). While some cells need both path-ways to be intact to undergo senescence, others may rely on asingle pathway (36). Furthermore, some studies indicate that cellsmay also undergo senescence in a p53- and p16-independentmanner through alternative pathways which are currently underinvestigation (36, 37).

We determined the levels of p53 as well as its downstreamtarget p21, and of p16 after induction of senescence by tepo-tinib treatment combined with IR in our panel of cell linesharboring wild-type [GTL-16 (38), MKN-45 (39) and KATO-II(data not shown)], mutated [SNU-638 (40)], or deleted [SNU-5(40)] p53. Even though there were no detectable changes inp53 levels and its phosphorylation on Ser15 when IR wascombined with MET inhibitor as compared with IR alone,an increase in p21 protein levels was observed in GTL-16,MNK-45, and SNU-638 cells upon the combined treatmentwhen compared with untreated cells or cells treated by METinhibition or IR (Fig. 2A). In KATO-II cells, this increasecorrelated with IR only (Supplementary Fig. S3), while asexpected, in SNU-5 cells, p53 and p21 were not detected (Fig.2A). Conversely, p16 expression was downregulated after METinhibition alone or in combination with IR in all the cell linesapart of SNU-638 cells, where increased protein levels wereobserved (Fig. 2A). Such distinct expression profiles of p21 andp16 proteins indicate alternative pathways regulating the senes-cence program in the human gastric cell lines used in thecurrent study.

To determine the p53 dependence of the observed senes-cence phenotype in GTL-16 cells (wild-type p53), a p53-deplet-ed isogenic cell line (GTL-16 shp53) has been used (31). Asshown in Fig. 2B and C, p53 knockdown surprisingly did notabrogate induction of senescence and exposure of GTL-16shp53 cells to the combined treatment resulted in ratherenhanced SA-b-Gal staining and increased H3K9me3 andgH2AX protein levels as compared with control (GTL-16 shc)cells. Elevated expression of p16 in shp53 cells suggests thatMETi-associated senescence is in the absence of p53 executedby the p16 pathway.

Francica et al.

Clin Cancer Res; 2016 Clinical Cancer ResearchOF4

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 5: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

Figure 1.

MET inhibition promotes DNA damage–induced senescence both in vitro and in vivo. A, percentage of BrdUrd–positive cells. Data are mean � SEM fromthree independent experiments. P values are calculated by one-way ANOVA. ��� , P < 0.001. B, representative images (10� magnification) of viableand dead cells, calcein-AM (green) or ethidium homodimer 1 (red) positive, respectively. Right, data quantification. C, representative images of SA-b-Galactivity (40� magnification). Right, percentage of SA-b-Gal–positive cells. Data are mean � SEM from three independent experiments. P values arecalculated by one-way ANOVA. ��� , P < 0.001. (Continued on the following page.)

MET, FOXM1, and DNA Damage–Induced Senescence

www.aacrjournals.org Clin Cancer Res; 2016 OF5

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 6: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

Involvement of the MAPK pathway in senescence inductionafter MET inhibition

We next aimed at investigating which pathway downstreamof MET may enforce cellular senescence after MET inhibition.To that end, two key MET-downstream networks, the MAPKand the PI3K/AKT pathways (1), were inhibited using AZD6244and LY294002, respectively. As Fig. 3A and SupplementaryFig. S4A show, treatment with the MEK1/2 inhibitor AZD6244significantly enhanced SA-b-Gal staining and led to character-istic morphologic features of cellular senescence in irradiatedGTL-16, MKN-45, KATO-II, SNU-638, and SNU-5 cells. Onthe other hand, inhibition of the PI3K pathway alone or in

combination with IR did not reproduce the effects of tepotiniband AZD6244 treatment on SA-b-Gal staining in any of the celllines (Fig. 3A and Supplementary Fig. S4A). We further evalu-ated the effects of AZD6244 and LY294002 on additionalsenescence markers in GTL-16 and SNU-638 cells (Fig. 3B).Western blot analysis revealed that AZD6244, but notLY294002, had a comparable effect to tepotinib on gH2AX,H3K9me3, and p21 protein levels, particularly in combinationwith IR. To further confirm these findings, AZD6244 treatmentwas also investigated in MKN-45, KATO-II, and SNU-5 cells,where it was shown to recapitulate the effect of tepotinib ongH2AX, H3K9me3 and p21 levels (data not shown).

Figure 1.

(Continued. ) D, whole-cell lysates were subjected to Western blotting using specific antibodies against p-MET, MET, H3K9me3, and yH2AX followingthe indicated treatments. (Continued on the following page.)

Francica et al.

Clin Cancer Res; 2016 Clinical Cancer ResearchOF6

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 7: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

Downregulation of Forkhead box protein M1 (FOXM1) afterMET inhibition is crucial to induce MET inhibition–dependentsenescence

Within a phosphoproteomics discovery survey, we observedthat MET inhibition results in a nearly 10-fold decrease inphosphorylation of the transcription factor FOXM1 (Ser 605,Ser 620) in MET-addicted human non–small cell lung cancercell line EBC-1 (data not shown). The fundamental role ofFOXM1 in both the DNA damage response and senescence(28–30) prompted us to investigate whether the same wasrelevant in gastric cancer cells and whether the aforementioneddiminished phosphorylation of FOXM1 might in fact reflectlower levels of the total protein.

Indeed, FOXM1 expression levels were downregulated inall the studied human gastric cancer cell lines after MET inhi-bition independently of IR (Fig. 4A and Supplementary Fig.S5). This observation was also confirmed in vivo in both GTL-16and SNU-638 xenografts, where no FOXM1 staining could bedetected in tumor tissues of mice treated with tepotinib(Fig. 4B).

FOXM1 is required for an efficient repair of DNADSBs throughdirect upregulation of BRIP1 aswell as indirect targeting of BMI-1,one of the major negative regulators of the senescence program(41). As Fig. 4C shows, tepotinib treatment decreased the expres-sion of BRIP1 and BMI-1 in GTL-16 as well as SNU-638 cells invivo, suggesting that downregulation of FOXM1 protein levels

followedbyMET inhibition could affect the ability of these cells torepair DNA damage induced by IR, leading to senescence.

As FOXM1 activity depends on MAPK pathway activation[following phosphorylation by phospho-ERK1/2, FOXM1translocates to the nucleus and activates the transcription ofseveral targets involved in cell proliferation, survival, invasion,and metastasis (26)], we analyzed FOXM1 levels after treatmentwith inhibitors of both ERK/MAPK and PI3K/AKT pathways.Indeed, while AZD6244 treatment recapitulated the effect oftepotinib on FOXM1 protein level, LY294002 did not affectFOXM1 expression (Supplementary Fig. S5).

To directly assess the impact of FOXM1 on MET inhibition–dependent DNA damage–induced senescence, we stablyexpressed FOXM1 in SNU-638 (SNU638_FOXM1) and inGTL16 cells (GTL16_FOXM1). Figure 5A and SupplementaryFig. S6A and S6B show that in FOXM1-overexpressing cells, METinhibition does not lead to a complete downregulation ofFOXM1 expression when compared with the empty vectorcontrol cells. Interestingly, SNU-638_FOXM1 cells displayedless H3K9me3 and p21 accumulation and significantly lessSA-b-Gal activity (Fig. 5A and B, respectively) after MET inhi-bition alone or in combination with IR when compared with theempty vector control cells (SNU638_pcDNA3). We also inves-tigated the effect of FOXM1 overexpression on the ability ofMET inhibition to induce DNA damage by evaluating gH2AXlevels in SNU638_FOXM1 (Fig. 5A and Supplementary Fig. S6B)

Figure 1.

(Continued. ) E and F, representative images (40� magnification) of p-MET, Ki-67, yH2AX staining, and SA-b-Gal activity in GTL-16 and SNU-638 xenografts,respectively, after treatment as indicated (day 7). Right, percentage of p-MET, Ki-67, yH2AX, and SA-b-Gal activity–positive cells. P values arecalculated by one-way ANOVA. � , P < 0.05, �� , P < 0.01 and ��� , P < 0.001.

MET, FOXM1, and DNA Damage–Induced Senescence

www.aacrjournals.org Clin Cancer Res; 2016 OF7

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 8: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

and GTL16_FOXM1 cells (Supplementary Fig. S6A) with orwithout the irradiation treatment. The results showed thatFOXM1 ectopic expression abrogated MET inhibition–depen-dent increase of gH2AX independently of irradiation, highlight-ing an important role of FOXM1 in the execution of DNA DSBsrepair by tumor cells following MET inhibition. To furtherdemonstrate that the lack of accumulation of gH2AX is possiblya result of a more efficient repair of DNA DSBs in SNU-638_FOXM1 and GTL-16_FOXM1 cells, we evaluated BRIP1expression by Western blot analysis and we showed that FOXM1overexpression leads to higher levels of BRIP1 protein, whichcould enable SNU-638 and GTL-16 cells to repair DNA DSBsmore effectively (Fig. 5A and Supplementary Fig. S6A and S6B).

Furthermore, cells ectopically expressing FOXM1 seem to gain asignificant growth advantage over their parental counterpartsupon MET inhibition (Fig. 5C and D and SupplementaryFig. S6C), avoiding cell-cycle arrest and subsequent senescence.

Coalterations of MET and FOXM1 are common in patientswith gastric cancer

To gain further translational insights into the potentialrelevance of the MET-FOXM1 proposed network in patientswith gastric cancer, we assessed MET and FOXM1 protein levelsin two TMAs containing gastric cancer tissues from 91 patients.High protein levels of MET and FOXM1 were detected in 33%of the samples, and 80.3% of the tumors displayed high levels

p-p53

p21

p16

β-ac�n

– 2 4 – 2 4

––– 15 15 15

– 2 4 – 2 4

––– 15 15 15

– 2 4 – 2 4

––– 15 15 15 METi (15 nmol/L)

IR (Gy)– 2 4 – 2 4

––– 15 15 15

GTL-16 SNU-638 MKN-45 SNU-5

C

METi

2 Gy 4 Gy

GTL

-16

shp5

3

– + – + – + – +

β-Ac�n

γH2AX

p-p53

p-MET

MET

H3K9me3

p21

p16

p53

METi (15 nmol/L)

IR (4 Gy)

–– + + – – + +

GTL-16 shc GTL-16 shp53

B

A

C0 Gy

p53

Figure 2.

Characterization of pathways involved in the execution of the senescence program upon MET inhibition. A, total cell proteins were subjected toWestern blotting with a specific antibody against p-p53, p53, p21 or p16 after the indicated treatments. B, top, representative images (40� magnification)of SA-b-Gal activity in GTL-16 p53 knocked down cells (GTL-16 shp53). Bottom, percentage of SA-b-Gal activity–positive cells. P values are calculatedby one-way ANOVA. ��� , P < 0.001. C, Western blot analysis showing different expression of indicated senescence markers in GTL-16 cells harboringwild-type (GTL-16 shc) or knocked down p53 (GTL-16 shp53).

Francica et al.

Clin Cancer Res; 2016 Clinical Cancer ResearchOF8

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 9: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

of MET and/or FOXM1 (Fig. 6A). We further used Ki-67 as amarker of tumor proliferation and could observe a higherprevalence of MET/FOXM1 high tumors in samples with inter-mediate or high proliferation index (Fig. 6B). Tumor- andpatient-related data were available for 72 patients. We foundthat high FOXM1 and especially high FOXM1/MET expressionwas significantly more prevalent in patients with lymph nodemetastases (P values 0.046 and 0.008, respectively; Supple-mentary Table S1).

We next extracted gene copy number (CNVs from SNP arrays)and gene expression (RNASeq) data from the TCGA database. Weselected cases with full clinical and follow-up data (178 in total).Amplification of both oncogenes was found in 24.7% of patients,representing the largest subgroup of patients followed closely onlyby the group of patients with MET amplification/FOXM1 deletion,which represented 21.9% of the cohort (Fig. 6C). RNASeq datarevealed that all cases featuring FOXM1 mRNA overexpression(30%) also displayed MET mRNA overexpression (Fig. 6D).

PI3KiMEKiMETiC

0 Gy

4 Gy

GTL-16

SNU-638

PI3KiMEKiMETiC

0 Gy

4 Gy

A

B

γH2AX

p21

β-Ac�n

H3K9me3

– + – + – + – +

–– + + – ––– METi (15 nmol/L)

IR (4 Gy)

–––– + + – – MEKi (10 μmol/L)

–––––– + PI3Ki (10 μmol/L)+

– + – + – + – +

–– + + – ––– METi (15 nmol/L)

IR (4 Gy)

–––– + + – – MEKi (10 μmol/L)

–––––– + PI3Ki (10 μmol/L)+

GTL-16 SNU-638

γH2AX

p21

β-Ac�n

H3K9me3

Figure 3.

The MAPK pathway links MET inhibition and senescence induction. A, representative images (40� magnification) of SA-b-Gal activity in GTL-16 (top)and SNU-638 (bottom) cells,. Right, percentage of SA-b-Gal–positive cells. Data are mean � SEM from three independent experiments. P values arecalculated by one-way ANOVA. ��� , P < 0.001. B, Western blot analysis showing H3K9me3, yH2AX, and p21 levels in GTL-16 (left) and SNU-638 (right)cells after the indicated treatment.

MET, FOXM1, and DNA Damage–Induced Senescence

www.aacrjournals.org Clin Cancer Res; 2016 OF9

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 10: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

FOXM1

BRIP1

BMI-1

BRIP1

BMI-1

– – + +– + – +

METi (15 nmol/L)

IR (2 Gy)

– – + +– + – +

– – + +– + – +

– – + +– + – +

– – + +– + – +

FOXM1

β-Ac�n

p-MET

GTL-16 MKN-45 SNU-5 KATO-II SNU-638

C IR METi METi + IR

FOXM1

C IR METi METi + IR

GTL-16 Xenogra�

SNU-638 Xenogra�

C IR METi METi + IR

C IR METi METi + IR

GTL-16 Xenogra�

SNU-638 Xenogra�

B

A

C

Figure 4.

MET inhibition–dependent radiosensitization through induction of DNA damage–induced senescence is mediated by FOXM1 downregulation. A, Westernblot analysis showing p-MET and FOXM1 levels following the indicated treatments. B, representative images of FOXM1 staining in GTL-16 (top) andSNU-638 (bottom) xenografts. Right, percentage of FOXM1-positive cells. P values are calculated by one-way ANOVA. �� , P < 0.01; ��� , P < 0.001. C,representative images (40�magnification) of BMI-1 and BRIP1 staining in GTL-16 (top) and SNU-638 (bottom) xenografts. Right, percentage of BMI-1 or BRIP1-positive cells. P values are calculated by one-way ANOVA. �� , P < 0.01; ��� , P < 0.001.

Francica et al.

Clin Cancer Res; 2016 Clinical Cancer ResearchOF10

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 11: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

Neither gene CNV nor mRNA levels correlated with clini-copathologic features or survival in the TCGA cohort, consist-ing of patients treated with modalities other than anti-MET–targeted therapy (data not shown).

Given the lack of data from patients treated with anti-METtherapies, the potential relevance of these findings in terms of

outcomes or treatment selection criteria needs further eluci-dation. While the main interest of the current results isdescriptive, these data show that our proposed MET–FOXM1axis is commonly deregulated in human gastric cancer, atthe genomic, transcriptomic, and protein levels. Consequent-ly, upcoming clinical trials should allow specifying the role

Figure 5.

Ectopic expression of FOXM1 abrogates MET inhibition–dependent DNA damage–induced senescence. A, whole-cell lysates were subjected to Western blottingusing specific antibodies against FOXM1, BRIP1, yH2AX, H3K9me3, and p21 following the indicated treatments. B, representative images of SA-b-Gal activity (40�magnification) in FOXM1-overexpressing SNU-638 cells (SNU-638_FOXM1), compared with empty vector control SNU-638 cells (SNU-638_pcDNA3). Right,percentage of SA-b-Gal–positive cells. Data are mean � SEM from three independent experiments. P values are calculated by one-way ANOVA. ��� , P < 0.001. C,representative images (4� magnification, 2% crystal violet staining) of SNU-638_FOXM1 and SNU-638_pcDNA3 cells following 3 days of MET inhibition treatmentand stained 3 days after the removal of the MET inhibitor. D, percentage of proliferating SNU-638_FOXM1 and SNU-638_pcDNA3 cells following 3 days ofMET inhibition. Data are mean � SEM from three independent experiments. Statistical significance was determined using the Student t test. ��� , P < 0.001.

MET, FOXM1, and DNA Damage–Induced Senescence

www.aacrjournals.org Clin Cancer Res; 2016 OF11

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 12: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

A

C

RNA Expression levelsD

MET

FOXM1

High METHigh FOXM1

High METLow FOXM1

Low METHigh FOXM1

Low METLow FOXM1

18.7%44%3.3%33%

MET High/FOXM1 high

MET High/FOXM1 normal

MET Normal/FOXM1 normal

METLow

MET Normal

METHigh

FOXM1 High

FOXM1 Norm

FOXM1 Low

FOXM1 High

FOXM1 Norm

FOXM1 Low

FOXM1 Low

FOXM1 Norm

FOXM1 High

CNV

Low

Intermed

iate

High0

20

40

60High MET/ high FOXM1High MET/low FOXM1Low MET/high FOXM1Low MET/low FOXM1

Ki-67

Num

ber

of tu

mor

s

B

Low

Intermed

iate

High

Ki-67

Ki-67

0

20

10

30

40

60

Num

ber

of tu

mor

s

Low FOXM1High FOXM1

Low METHigh MET

Low

Intermed

iate

High0

20

40

60

Num

ber

of tu

mor

s

Figure 6.

Alterations of MET and FOXM1 commonly coexist in patients with gastric cancer. A, images of MET and FOXM1 staining in 4 representative gastric tumorsincluded in the tumor TMA. B, correlation of MET and FOXM1 expression in tumors included in the TMA with low, intermediate or high proliferationrate (Ki-67 labeling index is used as a marker of tumor cell proliferation). C, MET and FOXM1 gene copy number variation (norm, normal) in a TCGA cohortof 178 gastric cancer patients. D, MET and FOXM1 RNA expression levels in a TCGA cohort of 373 gastric cancer patients.

Francica et al.

Clin Cancer Res; 2016 Clinical Cancer ResearchOF12

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 13: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

of the MET–FOXM1 axis in responses to therapy in gastriccancer.

DiscussionGastric cancer is one of the most common malignant tumors

worldwide accounting for more than 700,000 cancer-relateddeaths annually (42). Complete resection of localized tumors isthe mainstay of treatment, but patients newly diagnosed withgastric cancer tend to present with advanced and often incurabledisease (42). TheMETRTK is frequently dysregulated and stronglyimplicated in the malignant transformation and progression ofthis disease (43).

Here we examined the impact of inhibiting MET signalingby the highly selective and potent anti-MET small-moleculetepotinib (44) in MET-addicted gastric cancer models andreport for the first time that tepotinib sensitized cancer cellsto IR by promoting DNA damage–induced senescence bothin vitro and in vivo.

It is not yet fully understood what determines whether cellsundergo senescence or cell death, but the nature and intensity ofinsult/damage as well as cell type and inactivation of apoptosispathways during tumorigenesis (e.g., Bcl2 overexpression; ref. 45)may play a role (12, 14, 15). Interestingly, we observed differentexpression profiles of p21 and p16 in our panel of cell linesfollowing the combined treatment. Surprisingly, p53-null cellsunderwent senescence without increased p16 levels, suggestingthat senescence can be activated also independently of p53/p16INK4A-RB. These data are in contrast with some other studieswhere inactivation of these two pathways prevented senescenceinduction (46). However, according to a study of Wang andcolleagues (47) where induction of p16 expression in cell lineswith endogenous mutant p53 following senescence onset hasbeen reported, we similarly observed an upregulation of p16protein levels in GTL-16 shp53 cells compared with control cells.This suggests that in cancers harboring mutations of either p53 orp16, the presence of a wild-type version of the other gene couldmediate/execute the senescence program.

Even though cell death is the favored outcome of anticancertreatments, a current parallel perception advocates that activa-tion of nonapoptotic mechanisms should be considered apotential endpoint following chemo- and radiosensitizationapproaches as well (23). Using the Em-myc model of lympho-magenesis, the pioneer work of Schmitt and colleagues in 2002provided direct evidence that cellular senescence can be inducedfollowing chemotherapy in vivo and thus be a potential desirableoutcome of cancer treatment (35). Since then, several studieshave described senescence induction by chemo- or radiotherapyin a variety of cancer types as well as in cancer cells grown in vitro(19, 22, 24, 47). Similarly, we report that proliferation of cellstreated with tepotinib together with IR was dramaticallyimpaired in vitro and this effect was also observed in vivo, wheretumor volumes were 4- to 6-fold smaller compared withuntreated tumors (Supplementary Fig. S2A and S2B). The factthat senescence can be achieved by applying lower dosages ofDDAs than those required to drive short-term proliferative arrestor cell death may effectively result in reduced toxic side-effectsinherent to many therapies (48). Consistently, in the currentreport, we have observed the induction of senescence using theMET inhibitor tepotinib in concentrations as low as 10–15nmol/L for in vitro and 15 mg/kg for in vivo studies.

In recent years, many studies have focused on the importanceof FOXM1 in both DDR and negative senescence regulation(28–30). It has become evident that FOXM1 plays a fundamen-tal role in DNA DSBs repair through upregulation of key DDRproteins such as RAD51, NBS1, and BRIP1 (28, 30). Likewise,various studies described a role for FOXM1 in the regulation ofthe senescence program, mainly through its downstream targetBMI-1 (41). As Rovillain and colleagues reported, FOXM1 over-expression bypassed senescence in human fibroblasts followingactivation of the p53 and the p16INK4A-RB pathways, while Zengand colleagues (49) showed that inhibition of FOXM1 in gastriccancer–derived cell lines led to senescence in a p53- and thep16INK4A-RB–independent manner. Interestingly, our resultssuggest that FOXM1 levels are downregulated following METinhibition, thus compromising the ability of irradiated cells torepair their DNA.

To elucidate the mechanism by which MET inhibition leadsto downregulation of FOXM1 protein levels/senescence, Maand colleagues (50) reported that the nuclear translocation andactivation of FOXM1 is regulated by phospho-ERK–dependentphosphorylation by the MAPK pathway. In accordance with thisstudy, we similarly observed a decrease in FOXM1 protein levelsfollowing inhibition of the MAPK pathway by the MEK1/2inhibitor AZD6244, which also induced senescence in irradi-ated GTL-16, MNK-45, KATO-II, SNU-638, and SNU-5 cells. Onthe basis of these findings, we propose a model to describeradiosensitization via MET inhibition in the context of gastriccancer: following MET inhibition, FOXM1 is no longer phos-phorylated and activated by the MAPK pathway and therefore itdoes not translocate to the nucleus. In this scenario, the abilityof cells to repair the DNA damage inflicted by IR and potentiallyother genotoxic insults that inflict DNA damage is impaired andthe consequent accumulation of DSBs, together with the down-regulation of one of the major negative regulators of thesenescence program, will lead to DNA damage-induced senes-cence. Importantly, our current data also suggest that METtargeting–dependent FOXM1 downregulation underlies theobserved senescence phenotype, predominantly in combina-tion with an additional required genotoxic stress such as exten-sive DNA damage elicited in this case by IR. These findingssuggest a possible synergistic mechanism, which results fromthe downregulation of a critical senescence negative regulator,which also plays a positive role in DDR activation on one handand concurrent generation of DNA damage on the other hand.

From a translational perspective, available data clearly indicatethat alterations of MET and FOXM1 cooccur in a substantialsubset of patients with gastric cancer, therefore underlining thepotential relevance of our preclinical findings upon implemen-tation of anti-MET targeted therapy. While CNV andmRNA levelsof MET and FOXM1 did not correlate with tumoral features orsurvival, it is important to note that these patients did not receiveanti-MET–targeted therapy. Moreover, available data suggest thatprotein levels could be the most relevant readout in terms ofprognosis. Relevant to this, Spigel and colleagues found that METexpression levels were predictive of responses to the anti-METantibodyonartuzumab inpatientswith advancednon—small celllung cancer. Currently ongoing clinical trials aim at assessing theimpact of MET expression and amplification in gastric carcinoma(NCT02002416).

Our findings highlight the potential benefit of using METinhibitors in MET-overexpressing tumors to impair the ability of

MET, FOXM1, and DNA Damage–Induced Senescence

www.aacrjournals.org Clin Cancer Res; 2016 OF13

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 14: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

cancer cells to successfully repair damaged DNA and to enhancethe effect of radiotherapy in gastric cancer, a devastating diseasewith an urgent need for improved treatment modalities. Molec-ular stratification seems essential to achieve the best possibleoutcomes. Importantly, basedonourpreclinicalfindings, FOXM1overexpression may be a potential mechanism of resistance toMET inhibitors. This point should be explored in upcomingclinical trials.

Disclosure of Potential Conflicts of InterestA. Blaukat is listed as a co-inventor on all patents related to Merck's c-Met

inhibitors listed in this manuscript. No potential conflicts of interest weredisclosed by the other authors.

Authors' ContributionsConception and design: P. Francica, D.M. Aebersold, F. Bladt, A. Blaukat,Y. Zimmer, M. Medov�aDevelopment of methodology: P. Francica, L. NisaAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): L. Nisa, R. Langer, D. Stroka, M. Medov�aAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): P. Francica, L. Nisa, M.R. Martínez, Y. Zimmer,M. Medov�a

Writing, review, and/or revision of the manuscript: P. Francica, L. Nisa,D.M. Aebersold, M.R. Martínez, Y. Zimmer, M. Medov�aAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): P. Francica, R. Langer, Y. Zimmer, M. Medov�aStudy supervision: Y. Zimmer, M. Medov�a

AcknowledgmentsThe authors thank Prof. R. Medema (Netherlands Cancer Institute, Amster-

dam, theNetherlands) and Prof.M.P. Tschan (Institute of Pathology, Universityof Bern, Bern, Switzerland) for kindly providing the Flag-FoxM1-NT2 expressionvector and shp53 cells, respectively. The authors also thank Dr. A. Quintin andMrs. M. Leuener-Tombolini for excellent technical assistance.

Grant SupportThis study was supported by the Hedy und Werner Berger-Janser Stiftung

grant (toM.Medov�a) and by the Swiss National Science Foundation (grant. no.31003A_156816; to Y. Zimmer).

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

Received December 14, 2015; revised March 29, 2016; accepted April 20,2016; published OnlineFirst May 16, 2016.

References1. Trusolino L, Bertotti A, Comoglio PM. MET signalling: principles and

functions in development, organ regeneration and cancer. Nat Rev MolCell Biol 2010;11:834–48.

2. Ma PC, TretiakovaMS,MacKinnon AC, Ramnath N, Johnson C, Dietrich S,et al. Expression and mutational analysis of MET in human solid cancers.Genes Chromosomes Cancer 2008;47:1025–37.

3. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al.Cancer incidence and mortality worldwide: sources, methods and majorpatterns in GLOBOCAN 2012. Int J Cancer 2015;136:E359–86.

4. Dicken BJ, Bigam DL, Cass C, Mackey JR, Joy AA, Hamilton SM. Gastricadenocarcinoma: review and considerations for future directions. Ann Surg2005;241:27–39.

5. Liu YJ, ShenD, Yin X, Gavine P, Zhang T, Su X, et al. HER2,MET and FGFR2oncogenic driver alterations definedistinctmolecular segments for targetedtherapies in gastric carcinoma. Br J Cancer 2014;110:1169–78.

6. Nakajima M, Sawada H, Yamada Y, Watanabe A, Tatsumi M, YamashitaJ, et al. The prognostic significance of amplification and overexpressionof c-met and c-erb B-2 in human gastric carcinomas. Cancer 1999;85:1894–902.

7. An X, Wang F, Shao Q, Wang FH, Wang ZQ, Chen C, et al. MET ampli-fication is not rare and predicts unfavorable clinical outcomes in patientswith recurrent/metastatic gastric cancer after chemotherapy. Cancer2014;120:675–82.

8. Lee HE, Kim MA, Lee HS, Jung EJ, Yang HK, Lee BL, et al. MET in gastriccarcinomas: comparison between protein expression and gene copy num-ber and impact on clinical outcome. Br J Cancer 2012;107:325–33.

9. Graziano F, Galluccio N, Lorenzini P, Ruzzo A, Canestrari E, D'Emidio S,et al. Genetic activation of theMET pathway and prognosis of patients withhigh-risk, radically resected gastric cancer. J Clin Oncol 2011;29:4789–95.

10. Smolen GA, Sordella R, Muir B, Mohapatra G, Barmettler A, Archibald H,et al. Amplification of MET may identify a subset of cancers with extremesensitivity to the selective tyrosine kinase inhibitor PHA-665752. Proc NatlAcad Sci U S A 2006;103:2316–21.

11. Feng Y,Ma PC. Anti-MET targeted therapy has come of age: the first durablecomplete response with MetMAb in metastatic gastric cancer. CancerDiscov 2011;1:550–4.

12. Catenacci DV, Henderson L, Xiao SY, Patel P, Yauch RL, Hegde P, et al.Durable complete response of metastatic gastric cancer with anti-Mettherapy followed by resistance at recurrence. Cancer Discov 2011;1:573–9.

13. Peters S, Adjei AA. MET: a promising anticancer therapeutic target. Nat RevClin Oncol 2012;9:314–26.

14. Medova M, Aebersold DM, Blank-Liss W, Streit B, Medo M, Aebi S, et al.MET Inhibition results in DNA breaks and synergistically sensitizestumor cells to DNA-damaging agents potentially by breaching a dam-age-induced checkpoint arrest. Genes Cancer 2010;1:1053–62.

15. MedovaM, AebersoldDM, Zimmer Y. Themolecular crosstalk between theMET receptor tyrosine kinase and the DNA damage response-biologicaland clinical aspects. Cancers 2013;6:1–27.

16. Ganapathipillai SS, Medova M, Aebersold DM, Manley PW, Berthou S,Streit B, et al. Coupling of mutated Met variants to DNA repair via Abl andRad51. Cancer Res 2008;68:5769–77.

17. Medova M, Aebersold DM, Zimmer Y. MET inhibition in tumor cells byPHA665752 impairs homologous recombination repair of DNA doublestrand breaks. Int J Cancer 2012;130:728–34.

18. d'Adda di Fagagna F. Living on a break: cellular senescence as a DNA-damage response. Nat Rev Cancer 2008;8:512–22.

19. Azad A, Jackson S, Cullinane C, Natoli A, Neilsen PM, Callen DF, et al.Inhibition of DNA-dependent protein kinase induces acceleratedsenescence in irradiated human cancer cells. Mol Cancer Res 2011;9:1696–707.

20. Hayflick L, Moorhead PS. The serial cultivation of human diploid cellstrains. Exp Cell Res 1961;25:585–621.

21. ColladoM, BlascoMA, SerranoM. Cellular senescence in cancer and aging.Cell 2007;130:223–33.

22. Collado M, Serrano M. Senescence in tumours: evidence from mice andhumans. Nat Rev Cancer 2010;10:51–7.

23. Nardella C, Clohessy JG, Alimonti A, Pandolfi PP. Pro-senescence therapyfor cancer treatment. Nat Rev Cancer 2011;11:503–11.

24. te Poele RH, Okorokov AL, Jardine L, Cummings J, Joel SP. DNA damage isable to induce senescence in tumor cells in vitro and in vivo. Cancer Res2002;62:1876–83.

25. Myatt SS, Lam EW. The emerging roles of forkhead box (Fox) proteins incancer. Nat Rev Cancer 2007;7:847–59.

26. Koo CY,Muir KW, Lam EW. FOXM1: From cancer initiation to progressionand treatment. Biochim Biophys Acta 2012;1819:28–37.

27. Teh MT. FOXM1 coming of age: time for translation into clinical benefits?Front Oncol 2012;2:146.

28. Monteiro LJ, Khongkow P, Kongsema M, Morris JR, Man C, Weekes D,et al. The Forkhead Box M1 protein regulates BRIP1 expression andDNA damage repair in epirubicin treatment. Oncogene 2013;32:4634–45.

29. Alvarez-Fernandez M, Medema RH. Novel functions of FoxM1: frommolecular mechanisms to cancer therapy. Front Oncol 2013;3:30.

Francica et al.

Clin Cancer Res; 2016 Clinical Cancer ResearchOF14

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 15: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

30. KhongkowP, KarunarathnaU, KhongkowM,GongC, Gomes AR, Yague E,et al. FOXM1 targets NBS1 to regulate DNA damage-induced senescenceand epirubicin resistance. Oncogene 2013;34:4144–55.

31. Britschgi C, Rizzi M, Grob TJ, Tschan MP, Hugli B, Reddy VA, et al.Identification of the p53 family-responsive element in the promoter regionof the tumor suppressor gene hypermethylated in cancer 1. Oncogene2006;25:2030–9.

32. DimriGP, Lee X, BasileG, AcostaM, Scott G, Roskelley C, et al. A biomarkerthat identifies senescent human cells in culture and in aging skin in vivo.Proc Natl Acad Sci U S A 1995;92:9363–7.

33. RobinsonMD,McCarthyDJ, SmythGK. edgeR: a Bioconductor package fordifferential expression analysis of digital gene expression data. Bioinfor-matics 2010;26:139–40.

34. Roninson IB. Tumor cell senescence in cancer treatment. Cancer Res2003;63:2705–15.

35. Schmitt CA, Fridman JS, Yang M, Lee S, Baranov E, Hoffman RM, et al. Asenescence program controlled by p53 and p16INK4a contributes to theoutcome of cancer therapy. Cell 2002;109:335–46.

36. Adams PD. Healing and hurting: molecular mechanisms, functions, andpathologies of cellular senescence. Mol Cell 2009;36:2–14.

37. OlsenCL,Gardie B, YaswenP, StampferMR.Raf-1-induced growth arrest inhuman mammary epithelial cells is p16-independent and is overcome inimmortal cells during conversion. Oncogene 2002;21:6328–39.

38. Furlan A, Stagni V, Hussain A, Richelme S, Conti F, Prodosmo A, et al. Ablinterconnects oncogenic Met and p53 core pathways in cancer cells.Cell Death Differ 2011;18:1608–16.

39. Nabeya Y, Loganzo F Jr, Maslak P, Lai L, de Oliveira AR, Schwartz GK,et al. The mutational status of p53 protein in gastric and esophagealadenocarcinoma cell lines predicts sensitivity to chemotherapeuticagents. Int J Cancer 1995;64:37–46.

40. Ku JL, Park JG. Biology of SNU cell lines. Cancer Res Treat 2005;37:1–19.41. Li SK, SmithDK, LeungWY, Cheung AM, Lam EW,Dimri GP, et al. FoxM1c

counteracts oxidative stress-induced senescence and stimulates Bmi-1expression. J Biol Chem 2008;283:16545–53.

42. Takahashi T, Saikawa Y, Kitagawa Y. Gastric cancer: current status ofdiagnosis and treatment. Cancers 2013;5:48–63.

43. Hack SP, Bruey JM, Koeppen H. HGF/MET-directed therapeutics in gas-troesophageal cancer: a review of clinical and biomarker development.Oncotarget 2014;5:2866–80.

44. Bladt F, Faden B, Friese-Hamim M, Knuehl C, Wilm C, Fittschen C, et al.EMD 1214063 and EMD 1204831 constitute a new class of potent andhighly selective c-Met inhibitors. Clin Cancer Res 2013;19:2941–51.

45. Lowe SW, Schmitt EM, Smith SW, Osborne BA, Jacks T. p53 is requiredfor radiation-induced apoptosis in mouse thymocytes. Nature 1993;362:847–9.

46. Campisi J. Senescent cells, tumor suppression, and organismal aging: goodcitizens, bad neighbors. Cell 2005;120:513–22.

47. WangM,Morsbach F, Sander D, Gheorghiu L, Nanda A, Benes C, et al. EGFreceptor inhibition radiosensitizes NSCLC cells by inducing senescence incells sustaining DNA double-strand breaks. Cancer Res 2011;71:6261–9.

48. HavelkaAM,BerndtssonM,OlofssonMH, ShoshanMC, Linder S.Mechan-isms of action of DNA-damaging anticancer drugs in treatment of carci-nomas: is acute apoptosis an "off-target" effect? Mini Rev Med Chem2007;7:1035–9.

49. Zeng J, Wang L, Li Q, Li W, BjorkholmM, Jia J, et al. FoxM1 is up-regulatedin gastric cancer and its inhibition leads to cellular senescence, partiallydependent on p27 kip1. J Pathol 2009;218:419–27.

50. Ma RY, Tong TH, LeungWY, Yao KM. Raf/MEK/MAPK signaling stimulatesthe nuclear translocation and transactivating activity of FOXM1. MethodsMol Biol 2010;647:113–23.

www.aacrjournals.org Clin Cancer Res; 2016 OF15

MET, FOXM1, and DNA Damage–Induced Senescence

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987

Page 16: Depletion of FOXM1 via MET Targeting Underlies Establishment …clincancerres.aacrjournals.org/content/clincanres/early/2016/09/23/... · FoxM1-NT2 expression vector (kindly provided

Published OnlineFirst May 16, 2016.Clin Cancer Res   Paola Francica, Lluís Nisa, Daniel M. Aebersold, et al.   Program in Gastric Cancer

Induced Senescence−Establishment of a DNA Damage Depletion of FOXM1 via MET Targeting Underlies

  Updated version

  10.1158/1078-0432.CCR-15-2987doi:

Access the most recent version of this article at:

  Material

Supplementary

 

http://clincancerres.aacrjournals.org/content/suppl/2016/05/14/1078-0432.CCR-15-2987.DC1Access the most recent supplemental material at:

   

   

   

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

  Subscriptions

Reprints and

  [email protected] at

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

  Permissions

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

.http://clincancerres.aacrjournals.org/content/early/2016/09/23/1078-0432.CCR-15-2987To request permission to re-use all or part of this article, use this link

Cancer Research. on February 15, 2019. © 2016 American Association forclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 16, 2016; DOI: 10.1158/1078-0432.CCR-15-2987