transformation-dependent silencing of tumor...

Therapeutic Discovery

Transformation-Dependent Silencing of Tumor-SelectiveApoptosis-Inducing TRAIL by DNA HypermethylationIs Antagonized by Decitabine

Per Lund1, Irina Kotova1, Val�erie Kedinger1, Harshal Khanwalkar1, Emilie Voltz1,William C. Hahn2, and Hinrich Gronemeyer1

AbstractTNF-related apoptosis-inducing ligand (TRAIL) kills tumor cells selectively. We asked how emerging

tumor cells escape elimination by TRAIL and how tumor-specific killing by TRAIL could then be restored.We

found that TRAIL expression is consistently downregulated in HRASG12V-transformed cells in stepwise

tumorigenesis models derived from four different tissues due to DNA hypermethylation of CpG clusters

within the TRAIL promoter. Decitabine de-silenced TRAIL, which remained inducible by interferon, while

induction of TRAIL by blocking the HRASG12V-activated mitogen-activated protein kinase pathway was

subordinated to epigenetic silencing. Decitabine induced apoptosis through upregulation of endogenous

TRAIL in cooperation with favorable regulation of key players acting in TRAIL-mediated apoptosis.

Apoptosis induction by exogenously added TRAIL was largely increased by decitabine. In vivo treatment

of xenografted human HRASG12V-transformed human epithelial kidney or syngenic mice tumors by deci-

tabine blocked tumor growth induced TRAIL expression and apoptosis. Our results emphasize the potential

of decitabine to enhance TRAIL-induced apoptosis in tumors and thus provide a rationale for combination

therapies with decitabine to increase tumor-selective apoptosis.Mol Cancer Ther; 10(9); 1611–23.�2011 AACR.

Introduction

TNF-related apoptosis-inducing ligand (TRAIL) is auniquemember of the TNF family, as it is the only knownfactor that induces apoptosis selectively in a large varietyof tumor cells without affecting normal cells. Indeed,recombinant TRAIL and TRAIL-agonistic antibodiesare currently evaluated in clinical trials and early resultsare promising (1, 2). Unlike other members of this family,human TRAIL interacts with a complex system of recep-tors, comprising the 2 death-inducing receptors DR4 andDR5, 2 potentially antiapoptotic decoy receptors DcR1and DcR2, and a soluble receptor osteoprotegerin. TRAILbinding to DR4 or DR5 induces oligomerization andrecruitment, through the receptor death domains, of

specific cytoplasmic proteins that form the death-induc-ing signaling complex(es) resulting in the activation of thecaspase cascade that orchestrates the death program.

TRAIL plays a pivotal role in the innate immunesystem. Consistent with its role as cytokine, TRAIL isexpressed in several types of immune cells and is crucialfor tumor surveillance (3, 4). TRAIL�/� mice exhibitedincreased sensitivity to chemical carcinogenesis,decreased suppression of tumor growth in syngenicmodels, and developed lymphoma at later age (5–7).TRAIL suppressed spontaneous tumor formation in micethat possess only one functional p53 allele (7). TRAILaction is not limited to natural killer or natural killer Tcells, as it acts ubiquitously in a cell autonomous andparacrine mode (8).

The notion that the TRAIL signaling pathway inducesapoptosis selectively in tumor cells is supported by aplethora of in vitro and in vivo experiments with xeno-grafted human tumor cells (1, 2). The molecular basisaccounting for the cancer selectivity of TRAIL was inves-tigated on a wide spectrum of tumor cells. The divergentexpression of certain factors in normal and tumor cellshas previously been correlated with the tumor-selectiveaction of TRAIL; examples are the TRAIL decoy receptorsDcR1 and DcR2, the death receptors DR4 and DR5, MYC,FLIP or O-glycosyltransferases. However, no generalmechanism accounting for this selectivity has emergedfrom these studies (1, 2).

Authors' Affiliations: 1Department of Cancer Biology, Institut deG�en�etique et de Biologie Mol�eculaire et Cellulaire (IGBMC), CNRS/INSERM/Universit�e de Strasbourg, Illkirch, France; and 2Department ofMedical Oncology, Dana-Farber Cancer Institute, Boston, Massachusetts

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

Corresponding Author: Hinrich Gronemeyer, Department of CancerBiology, Institut de G�en�etique et de Biologie Mol�eculaire et Cellulaire(IGBMC), CNRS/INSERM/Universit�e de Strasbourg, 1 Rue Laurent Fries,BP 10142, IGBMC, 67404 Illkirch CEDEX, France. Phone: 33-0-38653473;Fax: 33-0-38653437; E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-11-0140

�2011 American Association for Cancer Research.

MolecularCancer

Therapeutics

www.aacrjournals.org 1611

on June 12, 2018. © 2011 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst June 22, 2011; DOI: 10.1158/1535-7163.MCT-11-0140

http://mct.aacrjournals.org/

The tumor-selective action of TRAIL may contribute toa number of therapeutic paradigms that exhibit somedegree of tumor selectivity. For example, it has beenreported that TRAIL can be induced by retinoids orhistone deacetylase inhibitors (9–11). A great numberof cancer therapeutics synergize with TRAIL and havebeen tested in combinatorial settings (2). The TRAILsignaling pathway seems to be a central weapon thatcan be activated by a variety of drugs and exerts itsfunction in the tumor cell through neighboring cellsand through the innate immune system.

To date, the information on the relationship betweentumorigenesis and TRAIL levels in solid cancers is lim-ited (12, 13). Considering the tumor-killing activity ofTRAIL and its implication in cancer surveillance, it isreasonable to expect that tumor cells are under selectivepressure to silence the TRAIL pathway. In this study, weinvestigated whether tumorigenesis affects the levels ofendogenous TRAIL in a cell-autonomous manner byusing stepwise tumorigenesis models in vitro and in vivo.

Materials and Methods

Cell lines and reagentsThe cell lines used in this study were provided by

William C. Hahn (Dana Faber Cancer Institute, Boston,Massachusetts). The isogenic background for each pair ofimmortalized and transformed cell lines was confirmedby analysis of autosomal short-tandem repeats and on a250K-SNP array (Affymetrix). Cell lines were immortal-izedwith SV40TAg and hTERT (HA1E, BJEL, HMLE, andSALE) and subsequently transformed by introduction ofHRASG12V (HA1ER, BJELR, HMLE/PR, and SALE-HR;refs. 14–16). They are derived from normal human epithe-lial kidney (HEK), foreskin fibroblast (BJ), mammaryepithelial (HMEC), and small airway lung epithelial(SALE) cells, respectively. Expression of SV40-ER inthe immortalized and expression of SV40-ER andHRASG12V in the transformed cells were regularly ver-ified by Western blot. The cell lines were cultivated on100 mm dishes. Immediately before use, stock solutionsof decitabine [5-Aza-20deoxycytidine; 10 mmol/L in 1 �PBS (pH 6.5) Sigma], IFNg (0.1 mg/mL in EtOH; Sigma),and U0126 (10 mmol/L in dimethyl sulfoxide; Promega)were diluted in medium. Recombinant human TRAIL(375-TEC, R&D Systems) was used for apoptosis induc-tion and DR5-Fc-chimeras (R&D Systems) for blockingTRAIL-induced apoptosis. One day before treatmentwith IFNg or the MEK1/2 inhibitor U0126 or TRAIL,3 � 105 cells were seeded. One day before treatment withdecitabine, 1 � 105 cells were seeded. Decitabine andmedium were replenished every 24 hours. For cotreat-ment with decitabine and U0126 or IFNg or TRAIL, cellswere treated with decitabine for 72 hours and, afterrenewing medium and decitabine, further incubatedfor 24 hours with decitabine and the respective agent.Decitabine and DR5-Fc–chimeras were coincubated for96 hours.

RT-qPCRReal-time quantitative PCR (RT-qPCR) was used to

quantify mRNA levels. Total RNA was extracted usingthe GenElute Kit (Sigma). Two mg RNA were reversetranscribed with AMV Reverse Transcriptase (Roche).RT-qPCR was conducted with a 1:10 dilution of thecDNA using SYBR Green (Qiagen) on a Light Cycler480 (Roche) with 15 minutes of denaturation (95�C)and 50 cycles of amplification (20 seconds, 95�C; 20seconds annealing, 62�C; and 25 seconds 72�C). Forcalculation standard, curves were run in parallel. mRNAlevels were normalized relative to human glyceralde-hyde-3-phosphate dehydrogenase (GAPDH) or mouse36B4 mRNA. Primer sequences are listed in Supplemen-tary Table S1.

ELISATRAIL was measured in cell lysates using the TRAIL

human ELISA Kit (Biomol) according to the man-ufacturer’s instruction. For normalization of TRAIL pro-tein relative to cell number, cells were trypsinized from adish seeded in parallel for identical treatment and the cellnumber calculated using a hemacytometer.

Western blotCells were scraped and lysed in RIPA-buffer supple-

mented with protease inhibitor cocktail (Roche), 25mmol/L sodium fluoride and 1 mmol/L sodium ortho-vanadate (Sigma). Forty mg of total protein from celllysates were separated on 8% to 12% SDS-PAGE andtransferred onto nitrocellulose membranes. Membraneswere blocked for 1 hour with 5% nonfat dry milk or 5%bovine serum albumin (as indicated by themanufacturer)in TBS containing 0.05% Tween 20, incubated with pri-mary antibody overnight, washed, incubated with horse-radish peroxidase-conjugated secondary antibody (CellSignaling) and visualized by chemiluminescence (Wes-tern blotting reagent; GE Healthcare). Rabbit polyclonalantibodies directed against caspase 9, cleaved caspase 3,cleaved PARP, STAT1 (all from Cell Signaling), DR5(Sigma), DR4 (Chemicon), BCL2, IRF1 (Santa Cruz), goatpolyclonal antibodies against TRAIL (R&D Systems) andb-actin (Santa Cruz), and mouse monoclonal antibodiesagainst FLIP (Alexis) were used. b-Actin was used as aloading control.

Chromatin immunoprecipitationChromatin immunoprecipitation (ChIP) analysis was

conducted according to the protocol of Upstate Biotech-nology withminormodifications. The 2.5� 106 cells wereseeded per 140 mm dish. The cross-linked chromatin wassheared by sonication using a Vibra cell sonicator (Bio-block Scientific Technology) set to 30% of maximumpower for 20 bursts of 15 seconds with 45 seconds onice. Extracts from 2 � 106 cells were incubated withantibody according to the manufacturer’s suggestions.Antibodies usedwere nonspecific rabbit IgG (Santa Cruz)as negative control, RNA-Polymerase II (Covance), IRF1,

Lund et al.

Mol Cancer Ther; 10(9) September 2011 Molecular Cancer Therapeutics1612




STAT1 (both from Santa Cruz), SP1, SP3, H3K27me1,H3K27me3 (all from Upstate) and H3K4me1,H3K4me2, H3K4me3, H3K9me1, and H3 (all fromAbcam). ChIP- and input-DNA were analyzed by RT-qPCR. A standard PCR protocol was conducted with anannealing temperature of 62�C for 50 cycles.

Bisulphite sequencingDNA was isolated using the QIAamp DNA Mini Kit

(Qiagen) and bisulphite modification conducted withthe EpiTect Bisulphite Kit (Qiagen). Eighty ng of thebisulphite-modified DNAwas used as a template in eachPCR reaction conducted with Taq-Polymerase (Roche). Astandard PCR protocol was used with primer-specificannealing for 40 cycles. Primers for amplification ofbisulphite-modified DNA were designed with the assis-tance of the Oligo3 Software and MethPrimer (17). PCRproducts were resolved by PAGE (12%), the specific bandexcised, and the PCR product purified. PCR productswere cloned using the pGem-T Vector System I (Pro-mega). Clones were sequenced on an ABIPrism3100-sequencer. Primer sequences for Bisulphite sequencingare listed in the Supplementary Table S1.

Retroviral infectionRetroviruses were packaged using FLYRD-18 cells for

infection of HeLa cells (18). Plasmid DNA (pBabepuroempty, HRASG12V, and HRASG12V/S35) was transfectedinto the packaging cell line using a standard calciumphosphate protocol. Transfection was repeated the nextday. Twenty-four hours later the supernatant was filteredthrough a 0.45 mm filter to infect HA1E cells. Polybrene(8 mg/mL) was included to enhance infection efficiency.Drug selection started 24 hours postinfection and con-tinued for 14 days with 0.5 mg/mL puromycin. Samplesfor determination of TRAIL mRNA and protein weretaken 1 passage after selection. The Ras-effector loopmutant HRASV12/S35 binds exclusively to Raf-kinasesbut not to PI3-kinase or RalGEFs, thus specifically acti-vating the mitogen-activated protein kinase (MAPK)pathway (19).

Fluorescence-activated cell sortingApoptosis wasmeasured by flow cytometry (FACScan,

Becton Dickinson) using anti-APO2.7PE-Cy5 mAb(Immunotech) that reacts with a 38 kDa mitochondrialmembrane protein (7A6 antigen).

Animal studiesThe 5 � 106 HRASG12V-transformed HA1ER cells were

subcutaneously injected into both flanks of Crl:Nu(Ico)Foxn1Nu mice (Charles River). For syngenic tumors 1 �106 C26 colon cancer cells were subcutaneously graftedinto both flanks of BALB/c mice. Treatments startedwhen tumors reached a size of 4 mm in 1 dimension.Before treatment the mice were randomized into groupswith similar tumor sizes. The tumor volume during thecourse of treatment was normalized relative to the tumor

volume on the day the treatment started (set to 100%).Decitabine (1.5 mg/kg body weight) was administeredintraperitoneally at alternate days. Initial experimentsshowed toxic side effects of decitabine at 3 or morempk. Tumor sizes were determined with a Vernier Cal-liper using the formula: tumor volume ¼ (a2 � b)/2 (a ¼tumor width, b ¼ tumor length). All animal studies wereconducted following protocols approved by the DDSV(Direction D�epartementale des Services V�et�erinaires).Samples for Western blot and RT-qPCR analysis weretaken immediately after sacrifice.

StatisticsStatistical comparisons between the means of measure-

ments in the experiments were done by using theunpaired Student’s t test.

Results

Transformation by the oncogene HRASG12V inducesdownregulation of TRAIL

The presently available information suggested thatTRAIL expression varies in normal tissue, while it islow or undetectable in tumors. This could be due to (i)selection resulting in escape from immune surveillance or(ii) a phenomenon that is linked to molecular eventsduring the transformation process. The establishmentof stepwise tumorigenesis models which recapitulatecritical steps of human tumorigenesis (20) and the obser-vation that TRAIL sensitivity was acquired in 2 of thesemodels (21) provided an ideal experimental system toaddress the question if tumorigenic transformation itselfmay, directly or indirectly, alter TRAIL expression.

To gain evidence for such a phenomenon, stepwisetumorigenesis models derived fromBJ, HEK,HMEC, andSALE (14–16)were analyzed. TRAILmRNA (Fig. 1A) andprotein (Fig. 1B and C) were significantly downregulatedwhen immortalized and HRASG12V-transformed cellswere compared, irrespective of cell type and tissue origin.As no difference in the rate of TRAIL mRNA decreasewas observed between HA1E and HA1ER (Fig. 1D), weexcluded that alteredmRNA stability in HA1ER accountsfor downregulation of TRAIL. We concluded thatHRASG12V-mediated transformation is associated withdownregulation of TRAIL expression in a cell-autono-mous manner.

Epigenetic modifications, not Interferon-signaling,are implicated in downregulation of TRAIL intransformed cells

The silencing of TRAIL could be due to interferencewith an upstream TRAIL activating pathway(s) or due toepigenetic deregulation. As the interferon pathway isknown to activate TRAIL expression (22), we testedwhether TRAIL is still inducible in HRASG12V-trans-formed cells. Indeed, IFNg induced TRAIL mRNA simi-larly in immortalized and transformed cells. Notably, theabsolute levels of TRAIL mRNA induced by IFNg in the

Epigenetic Downregulation of TRAIL in Tumor Cells

www.aacrjournals.org Mol Cancer Ther; 10(9) September 2011 1613




transformed cells remained below those of the IFNg-treated immortalized cells (Fig. 2A). It is well establishedthat STAT1 and IRF1 mediate the action of interferon atthe TRAIL promoter and Western blot analysis revealedindeed that IFNg induced IRF1, STAT1, and TRAILsimilarly in both HA1E and HA1ER (Fig. 2B).

The putative TRAIL promoter (�1 kb to þ 1kb relativeto the TSS) harbors 2 interferon response elements (IRF-E,ISRE), 2 sites for SP1, and 1 AP1 site (Fig. 2C). We testedby ChIP whether the downregulation of TRAILmRNA inthe transformed cells was due to differential recruitmentof transcription factors involved in interferon signaling.STAT1 binding was not detectable at the TRAIL promo-ter, neither without nor with IFNg stimulation, while itwas detectable on the IRF1 promoter (SupplementaryFig. S1A and B). Consistently, downregulation of STAT1by short interfering RNA did not affect TRAIL expressionin either cell line (data not shown). IRF1 was recruited tothe TRAIL promoter in HA1E after treatment with IFNg ,

but the same induction of IRF1 binding was observed forHA1ER (Fig. 2D). Differential RNA polymerase IIrecruitment to the TRAIL promoter paralleled TRAILexpression and corresponded to differences seen for theuntreated cells (Fig. 2D). These results indicate thatneither STAT1 nor IRF1 are causally implicated indownregulation of TRAIL mRNA in the tumorigenesismodels. SP1/3 recruitment to the TRAIL promoter wasreported to be crucial for histone deacetylase-mediatedTRAIL induction in breast cancer cell lines (23) butnot detected in HA1E or HA1ER (SupplementaryFig. S1C and D). Thus differential recruitment of SP1and/or SP3 did not account for silencing of TRAIL upontransformation.

Silencing of TRAIL could occur at the chromatin levelby histone modification. We therefore investigatedwhether activating or repressing histone marks wereintroduced during transformation. The levels of the acti-vating methylations of histone H3 lysine 4 in the TRAIL

BHEK HMEC

HA1E HA1ER

BJ

– +++

++

HMLE HMLE/PR

– +++

++

0

400

800

*

BJEL BJELR

– +++

++

0

10

20**

0

10

20*

0

5

10 ***

0

10

20***

– +++

++

AHEK BJ SALE

HA1E HA1ER

HMEC

– +++

++

HMLE HMLE/PR BJEL BJELR SALE SALE-HR

– +++

++

– +++

++

SV40ERHRASG12V

hTERT

SV40ERHRASG12V

hTERT

0

0.4

0.8

1.2 HA1E

HA1ER

Time (h)

C

β-Actin

TRAIL

HA

1E

HA

1E

R

CO

S7

CO

S7

D Actinomycin D

pRK5-TRAIL+–

0 1 2 4 6

– +++

++

0

1,000

2,000

*

0

50

100

*

SALE

SALE SALE-HR0

400

800***

TR

AIL

/GA

PD

H m

RN

AT

RA

IL (

pg

/10

6 c

ells

)

TR

AIL

/GA

PD

H m

RN

A

(5 μg/mL)

Figure 1. Oncogenictransformation by HRASG12V

causes downregulation of TRAIL.A, RT-qPCR for TRAIL mRNA. B,ELISA measurement of TRAILprotein in immortalized (gray bars)and transformed (black bars) HEK,HMEC, BJ, and SALE cells, n ¼ 3.C, Western blot for TRAIL inimmortalized HA1E, HRASG12V-transformed HA1ER, and COS7transfected with a pRK5-TRAILexpression vector. D, RT-qPCRfor the kinetics of TRAILexpression in immortalized (graybars) and transformed (black bars)HEK cells treated withactinomycin D, n ¼ 2. In A, levelsare normalized relative to thecorresponding transformed cells(set to 1), and in D, relative to theuntreated cells (set to 1). Data aremeans � SEM. *, P < 0.05;**, P < 0.01; ***, P < 0.001.

Lund et al.





promoter (H3K4me3: regions 3, 4, and 5; H3K4me2:region 5) were reduced in the HRASG12V-transformedcells (Fig. 2D). IFNg increased H3K4me3 (regions 4 and 5)similarly in both cell lines. No changes were observedfor H3K4me1, H3K9me1, H3K9me3, H3K27me1, and

H3K27me3 (Supplementary Fig. S1E–J). We concludedthat TRAIL silencing upon oncogenic transformation (i) isnot a consequence of altered abundance/activity of IRF1or STAT1, (ii) can be antagonized by exposure to IFNg ,and (iii) occurs at the TRAIL promoter.

0

50

100

150

200HA1E HA1ER HMLE HMLE/PR BJEL BJELR SALE SALE-HR

TR

AIL

/GA

PD

H m

RN

A

- 10 - 10 - 10 - 10 - 10 - 10 - 10 - 10

***

*

*

****

***

***

HA1E HA1ER

0

25

50

75

100

125

150H3K4m3

1 2 3 4 5 6

HA1E HA1ER

1 2 3 4 5 60

5

10

15IRF1

Fold

enrich

ment

1 2 3 4 5 6 1 2 3 4 5 6

D

0

10

20

30

40

50RNA polymerase II

1 2 3 4 5 6 1 2 3 4 5 6

Fold

enrich

ment

0

20

40

60

80H3K4m2

1 2 3 4 5 6

HA1E HA1ER

1 2 3 4 5 6

A

ISRE AP1 Sp1 IRF-E

-140 -100 -70-50

-2.0 -1.5 -1.0 -0.5 +0.5 +5.5+1

1 2 3 4 5 6qPCR ampliconExon

TRAIL-promoter

C

HA1E HA1ER

STAT1

IRF1

β-Actin

– –– ++ –COS7

pSG5-IRF1+–

B

IFNγ (10 ng/mL)

TRAIL

HA1E HA1ER

Fold

enrich

ment

Fold

enrich

ment+ IFNγ (10 ng/mL)

untreated control+ IFNγ (10 ng/mL)untreated control

+ IFNγ (10 ng/mL)untreated control

+ IFNγ (10 ng/mL)untreated control

L

Figure 2. Interferon signaling is not implicated in the downregulation of TRAIL in HRASG12V-transformed cells. A, RT-qPCR for TRAIL mRNAinduction after treatment with IFNg for 18 hours in immortalized (gray bars) and transformed (black bars) HEK, HMEC, BJ, and SALE cells, n ¼ 3.TRAIL mRNA levels after treatment with IFNg are normalized relative to the corresponding untreated transformed cell line (set to 1). Data aremeans � SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. B, Western blot for STAT1, IRF1, and TRAIL in HA1E and HA1ER treated with IFNg . COS7transfected with a pSG5-IRF1 expression vector served as positive control. C, illustration of the TRAIL proximal promoter region with the positions oftranscription factor–binding sites relative to the TSS (þ1) indicating the locations of the regions analyzed by qPCR after ChIP. D, ChIP assays on theTRAIL promoter conducted with antibodies against IRF1, RNA polymerase II, H3K4me3, and H3K4me2 for untreated HA1E and HA1ER cells (gray bars) andIFNg-treated (10 ng/mL; 18 hours) HA1E and HA1ER cells (black bars). The fold enrichment is the relative abundance of DNA fragments at the indicatedregions over a transcriptional inactive control region (TSS of Myoglobin), n ¼ 2.






DNA hypermethylation represses TRAIL expressionin RAS-transformed cells

Silencing of tumor-suppressor genes by DNA methy-lation is frequently observed in tumors and implicated intumor development (24, 25). The potential impact ofDNA methylation on HRASG12V-dependent downregu-lation of TRAIL was analyzed by treating HEK, HMEC,BJ, and SALE cells with decitabine, an established DNAmethylation inhibitor that is approved for treatment ofmyelodysplastic syndromes (26). TRAIL mRNA and pro-tein induction was significantly stronger in tumor than inimmortalized cells, resulting in TRAIL protein levelsabove those in the immortalized cells (Fig. 3A). TRAILprotein induction in the immortalized cells was low.Consistent with the kinetics expected for decitabine-mediated DNA demethylation TRAIL induction wasnot detected before 72 hours (Supplementary Fig. S2A;refs. 27, 28). Bisulfite sequencing of the TRAIL promoterand upstream region revealed hypermethylation within(i) 2 adjacent CpGs in intron 1 (region A) and (ii) the 50

part of a CpG-island (region C) upstream of the TRAILTSS in HA1ER (Fig. 3B). Decitabine reduced DNAmethy-lation in regions A andC (Supplementary Fig. S2B andC).No differences were observed in region B.

A common feature of RAS-transformed cells is theconstitutive activation of the MAPK signaling pathway,which could contribute to TRAIL silencing. We observedrepression of TRAIL in immortalized HA1E when over-expressing the RAS effector loop mutant HRASV12/S35that specifically activates the MAPK pathway (Ref. 19;Fig. 3C). Although this observation suggests a linkbetween MAPK activation and TRAIL repression, wedid not observe induction of TRAIL after treatment withtheMAPKK inhibitor U0126 in HA1ER. The repression ofTRAIL by DNA methylation inhibits derepressionthrough blocking MAPK activity. Supporting this notionthe combination of decitabine and U0126 in HA1ERsynergistically increased TRAIL mRNA (Fig. 3D).

We concluded that (i) DNA-hypermethylation silencesTRAIL in RAS-transformed cells and (ii) that TRAILinduction by MAPKK inhibition requires prior reversalof the super-posed DNA methylation lock in RAS-trans-formed cells.

Synergistic apoptosis induction in tumor cells byTRAIL and decitabine

As decitabine and IFNg induced high levels of TRAILin HA1ER, we reasoned that the combination of decita-bine or IFNg with sublethal doses of TRAIL may enhanceapoptosis. Pretreatment with decitabine followed bycoexposure to decitabine and TRAIL led to synergisticcell death (80%), whereas a 24-hour exposure to TRAILalone had no effect and decitabine alone induced 40% celldeath after 96 hours (Fig. 4A and Supplementary Fig. S3).The combination of TRAIL and decitabine induced theapoptosis cascade, including both initiator (CASP9) andexecutor (CASP3) caspase activation and resulted inenhanced PARP cleavage (Fig. 4B). Coincubation with

TRAIL-neutralizing DR5-Fc–chimeras significantlyreduced decitabine and decitabine plus TRAIL-inducedapoptosis, revealing the major contribution of TRAIL(Fig. 4C). DR5-Fc–chimeras specifically abolishedTRAIL-induced apoptosis but not other apoptosis path-ways. Apoptosis induced by the DNA-damaging agentetoposide was not inhibited by blocking the TRAIL sig-naling pathway (Fig. 4D). We concluded that TRAILinduced by decitabine is involved in decitabine-inducedapoptosis and that cotreatment with TRAIL and decita-bine significantly increases apoptosis.

Decitabine induces a gene expression patternfavoring TRAIL-induced apoptosis

Treatment with IFNg induced similar levels of TRAILas treatmentwith decitabine, but in contrast to decitabine,IFNg did not synergize with TRAIL for apoptosis induc-tion. Thus decitabine is likely to exert further apoptosis-inducing activities which enhance the effect of its TRAILinduction. To identify activities induced by decitabinebut not IFNg , we compared their regulatory effects on keymediators of the interferon and apoptosis pathways inHA1ER (Fig. 5A). Although IFNg more significantlyinduced IRF1 and STAT1, decitabine specifically inducedexpression of several proapoptotic regulators, includingDR4, DR5, BID, APAF1, and CASP8. Both compoundsinduced similarly strong expression of XAF1, which islikely to antagonize the action of the apoptosis inhibitorscIAP1, cIAP2, and XIAP that were induced by decitabine(29). DcR1 was induced by decitabine but did not inter-fere with TRAIL-induced apoptosis. Protein levels ofBCL2 and FLIP, which antagonize TRAIL-induced apop-tosis (30, 31), were dramatically reduced by decitabine,whereas those of TRAIL, DR4, and DR5 stronglyincreased (Fig. 5B). The transcriptional response patterntoward decitabine was conserved in the HRASG12V-transformed fibroblast (BJELR) and mammary epithe-lial (HMLE/PR) cells (Fig. 5C). The synergistic apopto-genic effect between decitabine and exogenous TRAILwas also observed in these cells (data not shown).Notably, silencing of BID abolished TRAIL-inducedapoptosis via the intrinsic death pathway (Fig. 5D, rightpanel). However, tumor cell killing by decitabine aloneis limited. Addition of exogenous TRAIL to decitabinelargely increased apoptosis, thus indicating that TRAILis a crucial limiting factor and that decitabine inducessubthreshold TRAIL levels. Note that also lower con-centrations of decitabine increase TRAIL and DR5expression (Fig. 5D, left panel). We concluded thatmultiple actions of decitabine together with the induc-tion of TRAIL and its cognate DR5 receptor enhance thesensitivity of tumor cells toward TRAIL and that exo-genous TRAIL enhances apoptosis induction.

Anticancer action and TRAIL induction bydecitabine in vivo

To assess whether decitabine induces TRAIL andexerts an anticancer activity in vivo, we established

Lund et al.





3q261kb

TRAIL

240bp144bp

278bpregion A

region Bregion C 1 2 3 4 5 6

1 234 5 6 7 8 9 10

1 2 345

region A

1 5432 6HA1E (n = 16)HA1ER (n = 15)

region B

HA1E (n = 14)HA1ER (n = 13)

1 5432 6 10987

Met

hyla

tion

(%)

50

50

100

1 432CpGHA1E (n = 18)HA1ER (n = 16)

region C

Exon1/Intron1CpG island 3′CpG island 5′

A

B

AN

Rm

HD

PA

G/LI

AR

T

decitabine (μmol/L)

- 10 - 10 - 10 - 10 - 10 - 10 - 10 - 100

50

100

150

200

******

* ***

*

*****

HA1E HA1ER HMLE HMLE/PR BJEL BJELR SALE SALE-HR

*

0

2

4

6

8

decitabine (μmol/L)

- 10 - 10 - 10 - 10 - 10 - 10 - 10 - 10

**

*

****

*

HA1E HA1ER HMLE HMLE/PR BJEL BJELR SALE SALE-HR

TR

AIL

(n

g/1

06

ce

lls)

0

6

12

TR

AIL

/GA

PD

H m

RN

A

empty V12 S35

HA1E

***

phospho p42/p44 MAPK

p42/p44 MAPK

TRAIL

C

β-Actin

β-Actin

0

40

80

U0126 (10 μmol/L)+ ++ +

+ ++ +

HA1E HA1ER

decitabine (10 μmol/L)

TR

AIL

/GA

PD

H m

RN

A

5x5x

13x

28x

**

**

** U0126 (10 μmol/L)decitabine (10 μmol/L)

HA1ER

+ ++ +

TRAIL

β-Actin

D

decitabine

Figure 3. DNA hypermethylation silences TRAIL in transformed cells. A, left, RT-qPCR for induction of TRAIL mRNA by decitabine in immortalized(gray bars) and transformed (black bars) HEK, HMEC, BJ, and SALE cells, n ¼ 3. TRAIL mRNA levels are normalized relative to the correspondinguntreated transformed cell line (set to 1). Right, ELISA measurement for TRAIL protein after treatment with decitabine in immortalized (gray bars) andtransformed (black bars) HEK cells, n ¼ 2. Bottom left, structure of decitabine. B, regions of the TRAIL promoter analyzed by bisulfite sequencing.Exon1 (white/gray box), the TSS (arrow), a CpG-island 2 kb upstream (dark gray box), and the localization of amplified sequences with individual CpGs (circles)numbered in 50 to 30 direction are shown. Bisulfite sequencing of the exon1/intron1 TRAIL–promoter region and 2 regions covering a CpG-island 2 kbupstream of the TSS with the methylation rate for each individual CpG in HA1E (gray bars) and HA1ER (black bars); n ¼ numbers of clones analyzed.C, top, RT-qPCR for TRAIL mRNA in HA1E after retroviral infection with a HRASG12V- and a HRASG12V/S35-expressing pBabe-plasmid (black bars)and an empty control vector (gray bar), n ¼ 2. Bottom, Western blot for TRAIL, phospho-p42/p44 MAPK, and p42/p44 MAPK. D, left: RT-qPCR for TRAILmRNA after treatment of HA1E and HA1ER with the MAPKK inhibitor U0126 or decitabine or combination of both drugs, n ¼ 3. Right, Western blot forTRAIL in HA1ER. Data are means � SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.






xenografts in immunoincompetent mice with HRASG12V-transformed human HA1ER cells. Treatment of HA1ERxenografted mice with decitabine significantly reducedtumor growth. Normalized weight profiles showed theabsence of overt toxicity at the given concentration ofdecitabine (Fig. 6A). The expression pattern for apoptosisregulators in decitabine-treated tumors, including signif-icant induction of TRAIL, DR5, XAF1, and STAT1(Fig. 6B), was virtually identical to the one observedin vitro (Fig. 5A). Concomitant with the induction ofTRAIL intratumoral PARP cleavage revealed apoptosisinduction by decitabine (Fig. 6C). Notably, decitabineinduced TRAIL and DR5 in C26 mouse colon carcinomacells and in tumors of transplanted syngenic BALB/c

mice, leading to significant reduction of tumor growthwithout toxicity (Supplementary Fig. S4A and B).

This shows that TRAIL induction by decitabine isoperative in the context of an intact murine immunesystem without side effects. These data suggested thattreatment of tumors with decitabine for reactivation ofthe tumor-selective TRAIL apoptosis pathway is a valu-able option.

Discussion

It is generally accepted that the development of humancancer is a multi-step process that involves 6 types ofessential functional alterations in cell physiology, which

B

F.L.37 kDa35 kDa

19 kDa17 kDa

decitabine (10 μmol/L)decitabine (1 μmol/L)

----

+

-

-

+

-+-+

+---

--+-

+-+-

-+--

---+ +

--+

β-Actin

CA

SP

9C

AS

P3

A

seed harvest

0 time (h)24 48 72 96 120

decitabine+ decitabine+

treatment

+ IFNγ

- + -- - -- - +-

+-

-

-+- - decitabine (10 μmol/L)


-

-

+-

+

+-

+-

+- - -

---+-

+---+

+-+-- IFNγ (10 ng/mL)

+

---

++

----

Apo2.7

posi

tive (

%)

0

50

100

0

50

100

C D

- + -- - -- -

++-

+

-


DR5-Fc-chimera (500 ng/mL)

+ + +++

*

**

TRAIL (10 ng/mL)TRAIL (100 ng/mL)

TRAIL

TRAIL (10 ng/mL)TRAIL (100 ng/mL)

TRAIL (100 ng/mL)

cleaved PARP

β-Actin

β-Actin

Apo2.7

posi

tive (

%)

0

50

100

Ap

o2

.7 p

osi

tive

(%

)

-+

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

- - +-

- +- +

72 h

TRAIL (10 ng/mL)

TRAIL (100 ng/mL)

DR5-chimera (500 ng/mL)

etoposide (5 μg/mL)

HA1ER

HA1ER

HA1ER

Figure 4. Synergistic apoptosis induction in tumor cells by TRAIL and decitabine. A, top, scheme for the treatment of HA1ER with decitabine or TRAIL or IFNgor a combination of these drugs. Bottom, apoptosis rates measured in HA1ER. IFNg-treated cells are shown in gray, n ¼ 3. B, Western blot for CASP9,cleaved CASP3, and cleaved PARP after treatment with decitabine or TRAIL or a combination of these drugs. C, DR5-Fc–chimeras antagonize apoptosisinduction by decitabine and by the combination of TRAIL and decitabine. D, DR5-Fc–chimeras specifically abolish TRAIL- but not etoposide-inducedapoptosis, n ¼ 3. Apoptosis was measured by APO2.7 staining of HA1ER by fluorescence-activated cell sorting after treatment with TRAIL, decitabine,the DR5-Fc–chimera, and combinations of these drugs, n ¼ 3. Data are means � SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Lund et al.





BATRAIL

BCL2

FLIP

IRF1

STAT1

HA1ER

10 μmol/L 1 μmol/L +

–––

–+

p89

p24

decitabine

DR4

DR5

β-Actin

0.1

1

10

100

1,000

** ** **

****

**

*** ***

** ** *

**

***

γHA1ER + IFN (10 ng/mL)HA1ER + decitabine (10 μmol/L)

Targ

et g

en

e/G

AP

DH

mR

NA

IRF

1

STA

T1

TR

AIL

DR

5

DR

4

AP

AF

1

BID

XA

F1

cIA

P1

cIA

P2

XIA

P

BC

L2

CA

SP

8

Dc-

R1

Dc-

R2

0.1

1

10

100

1,000

**

**

*** ***

* **

*

** **

**

**

**

*

** *

BJELR + decitabine (10 μmol/L) HMLE/PR + decitabine (10 μmol/L)

Interferonsignaling

proapoptotic antiapoptotic

Targ

et g

en

e/G

AP

DH

mR

NA

*

IRF

1

STA

T1

TR

AIL

DR

5

DR

4

AP

AF

1

BID

XA

F1

cIA

P1

cIA

P2

XIA

P

BC

L2

CA

SP

8

Dc-

R1

Dc-

R2

C

p43p41

p18

F.L.

Cle

ava

ge

pro

du

cts

CO

S7

+

pR

K5

-TR

AIL

si RNA BIDscr RNA BID+

–––

–+

0

50

100

Ap

o2

.7 p

osi

tive

(%

)

++– –– –– + –

+– –+––

TRAIL (10 ng/mL)

si RNA BID scr RNA BID

–

+

+

BJELR

BJELR

BID

β-Actin

β-Actin

cleavedPARP

β-Actin

β-Actin

β-Actin

β-Actin

β-Actin

CASP8

Interferonsignaling


D

0

5

10AN

Rm

HD

PA

G/5

RD

0 0.01 0.1 1 10decitabine (μmol/L)

0

20

40AN

Rm

HD

PA

G/LI

AR

T

0 0.01 0.1 1 10decitabine (μmol/L)

HA1ERHA1ER

Figure 5. Decitabine modulates the expression of apoptosis regulators to trigger execution of TRAIL-induced apoptosis. A, RT-qPCR for genes involvedin (i) putative regulation of TRAIL by interferon signaling or (ii) regulation of apoptosis after treatment with decitabine or IFNg in HA1ER, n ¼ 3. B, Westernblot for TRAIL, DR5, DR4, BCL2, FLIP, IRF1, STAT1, CASP8, and cleaved PARP after treatment with decitabine in HA1ER. C, RT-qPCR for the genestreated with decitabine in BJELR and HMLE/PR, n ¼ 3. In A and C, the fold induction is relative to the corresponding untreated cell line (set to 1). Notelogarithmic scale of y-axis. D, left, dose-dependent induction of TRAIL and DR5 mRNA after treatment with decitabine for 96 hours, n ¼ 3. Right, top,measurement of APO2.7-positive BJELR by fluorescence-activated cell sorting after treatment with TRAIL for 72 hours 1 day after transfection with BIDsiRNA and the corresponding scrambled RNA. Right, bottom, Western blot for BID after transfection with BID siRNA. Data are means � SEM. *, P < 0.05;**, P < 0.01; ***, P < 0.001.






together drive malignant growth (32). Some of thesefeatures can be mimicked in a cell-autonomous mannerby stepwise tumorigenesis models (20). We observed thatTRAIL expression was very low in tumor cells whilenormal primary cells displayed higher TRAIL levelsand reasoned that TRAIL is attenuated during tumori-genesis. Indeed, there is accumulating evidence forTRAIL silencing, as (i) intense immunohistochemicalstaining for TRAIL was reported in 35% of low-gradecervical lesions or cervical intraepithelial neoplasia (CIN)but only in 3% of high-grade CINs (33), (ii) expression ofTRAIL was decreased upon colon adenoma to carcinomaprogression (34), and (iii) loss of TRAIL expressionaccompanied oral cancer progression (35).

However, in the above cases, it is impossible to discernthe cell-autonomous action of TRAIL and its contributionto immune surveillance. We found that at the transitionfrom immortalization to HRASG12V-mediated transfor-mation TRAIL is silenced. That the TRAIL locus can beepigenetically silenced puts it in line with other tumorsuppressor genes such as CDKN2A, CDKN2B, ARHGAP,or APC (24). TRAIL is not the only gene of the TRAILregulon that is epigenetically silenced. DNA methylationof both apoptogenic and decoy receptors of TRAIL and

CASP8 has been reported for glioblastoma and small celllung carcinomas (36–38). Moreover, gene-selective epi-genetic silencing upon transformation by oncogenic Rashas been previously observed (39–41); and among thesegenes is PAR-4, which is required for TRAIL-inducedapoptosis (42). Although the mechanistic link betweenoncogenic transformation and gene-selective epigeneticsilencing has remained elusive, it is important to pointout the pleiotropic action of decitabine.

Our data support a model in which decitabineincreases the apoptogenic potential of the TRAIL signal-ing pathway concomitantly at several key positions toattain threshold levels, which can then be boosted withexogenous TRAIL or TRAIL sensitizers. TRAIL-inducedapoptosis displays cell-to-cell variability based on itscontrol by multivariate factors that are constituted bythe protein levels of apoptosis regulators (43). Thoselevels differ between individual cells in a population.There is evidence that TRAIL-mediated apoptosis is cru-cially dependent on activation of the mitochondrial deathpathway as the increase of BID levels correlated withTRAIL-mediated apoptosis induction (43). Our data con-firm that BID is indispensable for TRAIL-mediated apop-tosis. BID is increased and BCL2 decreased in all

A

Decitabine (1.5 mpk)

vehicle

0,1

1

10

100HA1ER + decitabine (1.5 mpk)

B

Targ

et gene/G

AP

DH

mR

NA

Interferonsignaling


**

**

** *

TA

IRF

1

ST

1

TR

AIL

DR

5

DR

4

AP

AF

1

BID

XA

F1

cIA

P1

cIA

P2

XIA

P

BC

L2

CA

SP

8

Dc-

R1

Dc-

R2

C

TRAIL

β-Actin

F.L. PARP

cleaved PARP

T1 T2 T3 T4 T5 T6

vehicle decitabine

HA1ER-xenografts

β-Actin 0

3

6

Vehicle Decitabine

Cle

ave

d P

AR

P/

Fu

ll-le

ngth

PA

RP

HA1ER (n = 3)

0

500

1,000

1,500

No

rma

lized

tum

or

volu

me

(m

m3)

***

HA1ER (n = 5)

Days after randomization1 7 13 19

90

100

110

No

rma

lize

d a

nim

al w

eig

hts

(m

g)

1 7 13 19

HA1ER (n = 3)

Days after randomization

+ decitabine (1.5 mpk)vehicle

+ decitabine (1.5 mpk)

vehicle

Figure 6. The anticancer action of decitabine is associated with TRAIL induction and apoptosis. A, left, photography; middle, normalized tumor volumeprofile; and right, normalized animal weight after treatment with vehicle or decitabine (1.5 mpk) for 19 days in HA1ER-xenografted nude mice. B, RT-qPCRfor regulation of genes in decitabine-treated HA1ER tumor xenografts. C, Western blot and densitometry for TRAIL, full-length and cleaved PARP inHA1ER-xenografted tumors treated with vehicle or decitabine. mRNA levels in B are normalized relative to the untreated cells (set to 1). Data aremeans � SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Lund et al.





transformed cell lines after treatment with decitabine.Thus co-drugging reduces the impact of cell-to-cell varia-bility and increases the efficiency of TRAIL-mediatedkilling of cancer cells thereby lowering the critical thresh-old for efficient killing (44). Our data show that decitabineactivates the TRAIL pathway to attain this threshold.We show a link between HRASG12V-mediated activa-

tion of the MAPK pathway and TRAIL repression. Intransformed cells, this repression is subordinated to epi-genetic silencing of TRAIL. This subordination can beunlocked by combinatorial DNA demethylation andMAPKK inhibition, resulting in upregulation of TRAIL.In summary, these data suggest the existence of anintricate network comprising a TRAIL regulon thatinvolves the MAPK pathway and the epigenetic statusof the PAR-4, TRAIL, and TRAIL-receptor gene loci andreveals that oncogenic RAS positions multiple breaks inthe death pathway, several of which can be lifted bycombinatorial pharmacologic intervention (Fig. 6D).Our data suggest that decitabine-mediated derepression

of critical components of the TRAIL regulon contributesto the antitumor activity of this drug.

Current clinical trials (45) explore the combination ofdecitabine with other epigenetic modifiers and/or cyto-toxic agents. A possible concern with the use of globaldemethylating agents is that they might not only speci-fically activate tumor-suppressor genes, but also proto-oncogenes. We tested the effect of decitabine on 7protooncogenes, 3 of them previously described to beactivated by decitabine (46–48). The effects variedbetween the 3 transformation models and did not exceeda 4-fold induction. Notably, we observed repression inseveral cases (Supplementary Fig. S5). Note that at pre-sent, there is no evidence for induction of tumorigenesisin patients treated with decitabine for myelodysplasticsyndromes (49).

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

TSS

+256 +261–2,000

CpG island

TRAIL

DNMTs

decitabine

III

III

DNMTs

decitabine

U0126

HRAS

MAPK

MAPKK

+1 +57–165

ISRE AP1 GC IRF-Ebox

PAR-4

decitabine

DNMTs

Cancer-selectiveapoptosis

GRP78

?

meCpG

block of process

activation

I-IV transcription regulatory modules of TRAIL locus

regulatory events (direct or indirect) unknown?

establishment

TRAILRECEPTORS

CASP8

Intrinsic death pathway BID

BCL2

FADD

FLIP

DISC

D

Figure 6. (Continued ) D, model of the TRAIL regulon affected by Ras transformation. Transformation by HRASG12V increases DNA methylation in regulatorymodules (I, III) of the TRAIL promoter. The PAR-4 genomic loci whose gene product and its chaperone GRP78 are both required for TRAIL-inducedapoptosis was reported to be silenced by hypermethylation in RAS-transformed cells. Decitabine treatment relieves the epigenetic repression of TRAIL andPar-4. RAS activates the MAPK pathway leading to suppression of TRAIL expression. Relief of TRAIL expression by MAPKK inhibition is subordinated to theepigenetic silencing and can be seen in tumor cells only after combinatorial treatment with decitabine and U0126. Decitabine regulates several genesfacilitating TRAIL-induced apoptosis. Blue, genes upregulated by decitabine; red, genes downregulated. DISC, death-inducing signaling complex.






Acknowledgments

We thank Cathie Erb and Mich�ele Lieb for technical assistance.

Grant Support

This work was financially supported by the Ligue National Contre le Cancer(laboratoire labelis�e), and the European Community contracts LSHC-CT-2005-

518417 "Epitron" andHEALTH-F4-2007-200767 "Apo-Sys". P. Lundwas supportedby the Institut National du Cancer, I. Kotova by the Association de RechercheContre le Cancer, and V. Kedinger by the Ligue National Contre le Cancer.

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

Received February 20, 2011; revised June 10, 2011; accepted June 10,2011; published OnlineFirst June 22, 2011.

References1. Ashkenazi A, Holland P, Eckhardt SG. Ligand-based targeting of

apoptosis in cancer: the potential of recombinant human apoptosisligand 2/tumor necrosis factor-related apoptosis-inducing ligand(rhApo2L/TRAIL). J Clin Oncol 2008;26:3621–30.

2. Johnstone RW, Frew AJ, Smyth MJ. The TRAIL apoptotic pathway incancer onset, progression and therapy. Nat Rev Cancer 2008;8:782–98.

3. Takeda K, Hayakawa Y, SmythMJ, Kayagaki N, Yamaguchi N, KakutaS, et al. Involvement of tumor necrosis factor-related apoptosis-inducing ligand in surveillance of tumor metastasis by liver naturalkiller cells. Nat Med 2001;7:94–100.

4. Takeda K, Smyth MJ, Cretney E, Hayakawa Y, Kayagaki N, Yagita H,et al. Critical role for tumor necrosis factor-related apoptosis-inducingligand in immune surveillance against tumor development. J Exp Med2002;195:161–9.

5. Cretney E, Takeda K, Yagita H, Glaccum M, Peschon JJ, Smyth MJ.Increased susceptibility to tumor initiation and metastasis in TNF-related apoptosis-inducing ligand-deficient mice. J Immunol2002;168:1356–61.

6. Sedger LM, Glaccum MB, Schuh JC, Kanaly ST, Williamson E,Kayagaki N, et al. Characterization of the in vivo function of TNF-alpha-related apoptosis-inducing ligand, TRAIL/Apo2L, using TRAIL/Apo2L gene-deficient mice. Eur J Immunol 2002;32:2246–54.

7. Zerafa N,Westwood JA, Cretney E,Mitchell S,Waring P, Jezzi M, et al.Cutting edge: TRAIL deficiency accelerates hematological malignan-cies. J Immunol 2005;175:5586–90.

8. Altucci L, Rossin A, Raffelsberger W, Reitmair A, Chomienne C,Gronemeyer H. Retinoic acid-induced apoptosis in leukemia cellsismediated by paracrine action of tumor-selective death ligand TRAIL.Nat Med 2001;7:680–6.

9. Altucci L, Rossin A, Hirsch O, Nebbioso A, Vitoux D, Wilhelm E, et al.Rexinoid-triggered differentiation and tumor-selective apoptosis ofacute myeloid leukemia by protein kinase A-mediated desubordina-tion of retinoid X receptor. Cancer Res 2005;65:8754–65.

10. Nebbioso A, Clarke N, Voltz E, Germain E, Ambrosino C, Bon-tempo P, et al. Tumor-selective action of HDAC inhibitors involvesTRAIL induction in acute myeloid leukemia cells. Nat Med 2005;11:77–84.

11. Insinga A, Monestiroli S, Ronzoni S, Gelmetti V, Marchesi F, Viale A,et al. Inhibitors of histone deacetylases induce tumor-selective apop-tosis through activation of the death receptor pathway. Nat Med2005;11:71–6.

12. Daniels RA, Turley H, Kimberley FC, Liu XS, Mongkolsapaya J, Ch'en,P, et al. Expression of TRAIL and TRAIL receptors in normal andmalignant tissues. Cell Res 2005;15:430–8.

13. Spierings DC, de Vries EG, Vellenga E, van den Heuvel FA, KoornstraJJ,Wesseling J, et al. Tissue distribution of the death ligand TRAIL andits receptors. J Histochem Cytochem 2004;52:821–31.

14. ElenbaasB, Spirio L, Koerner F, FlemingMD, ZimonjicDB,Donaher JL,et al. Human breast cancer cells generated by oncogenic transforma-tion of primary mammary epithelial cells. Genes Dev 2001;15:50–65.

15. 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.

16. 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.

17. Li LC, Dahiya R. MethPrimer: designing primers for methylation PCRs.Bioinformatics 2002;18:1427–31.

18. Cosset FL, Takeuchi Y, Battini JL, Weiss RA, Collins MK. High-titerpackaging cells producing recombinant retroviruses resistant tohuman serum. J Virol 1995;69:7430–6.

19. Boettner B, Van Aelst L. The RASputin effect. Genes Dev 2002;16:2033–8.

20. Hahn WC, Weinberg RA. Modelling the molecular circuitry of cancer.Nat Rev Cancer 2002;2:331–41.

21. Nesterov A, Nikrad M, Johnson T, Kraft AS. Oncogenic Ras sensi-tizes normal human cells to tumor necrosis factor-alpha-relatedapoptosis-inducing ligand-induced apoptosis. Cancer Res 2004;64:3922–7.

22. Clarke N, Jimenez-Lara AM, Voltz E, Gronemeyer H. Tumor suppres-sor IRF-1 mediates retinoid and interferon anticancer signaling todeath ligand TRAIL. EMBO J 2004;23:3051–60.

23. Xu J, Zhou JY, Wei WZ, Philipsen S, Wu GS. Sp1-mediated TRAILinduction in chemosensitization. Cancer Res 2008;68:6718–26.

24. Esteller M. Epigenetics in cancer. N Engl J Med 2008;358:1148–59.25. Jones PA, Baylin SB. The epigenomics of cancer. Cell 2007;

128:683–92.26. Rodriguez-Paredes M, Esteller M. Cancer epigenetics reaches main-

stream oncology. Nat Med 2011;17:330–9.27. Bender CM, Gonzalgo ML, Gonzales FA, Nguyen CT, Robertson KD,

Jones PA. Roles of cell division and gene transcription in the methyla-tion of CpG islands. Mol Cell Biol 1999;19:6690–8.

28. Stresemann C, Lyko F. Modes of action of the DNA methyltransfer-ase inhibitors azacytidine and decitabine. Int J Cancer 2008;123:8–13.

29. Leaman DW, Chawla-Sarkar M, Vyas K, Reheman M, Tamai K, Toji S,et al. Identification of X-linked inhibitor of apoptosis-associated fac-tor-1 as an interferon-stimulated gene that augments TRAIL Apo2L-induced apoptosis. J Biol Chem 2002;277:28504–11.

30. Fulda S, Meyer E, Debatin KM. Inhibition of TRAIL-induced apoptosisby Bcl-2 overexpression. Oncogene 2002;21:2283–94.

31. Sharp DA, Lawrence DA, Ashkenazi A. Selective knockdown of thelong variant of cellular FLICE inhibitory protein augments deathreceptor-mediated caspase-8 activation and apoptosis. J Biol Chem2005;280:19401–9.

32. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57–70.

33. Reesink-Peters N, Hougardy BM, van den Heuvel FA, Ten Hoor KA,Hollema H, Boezen HM, et al. Death receptors and ligands in cervicalcarcinogenesis: an immunohistochemical study. Gynecol Oncol2005;96:705–13.

34. Koornstra JJ, Kleibeuker JH, van Geelen CM, Rijcken FE, Hollema H,de Vries EG, et al. Expression of TRAIL (TNF-related apoptosis-inducing ligand) and its receptors in normal colonic mucosa, adeno-mas, and carcinomas. J Pathol 2003;200:327–35.

35. Vigneswaran N, Baucum DC, Wu J, Lou Y, Bouguot J, Muller S, et al.Repression of tumor necrosis factor-related apoptosis-inducingligand (TRAIL) but not its receptors during oral cancer progression.BMC Cancer 2007;7:108.

36. Shivapurkar N, Toyooka S, Toyooka KO, Reddy J, Miyajima K, SuzukiM, et al. Aberrant methylation of trail decoy receptor genes is frequentin multiple tumor types. Int J Cancer 2004;109:786–92.

37. Eramo A, Pallini R, Lotti F, Sette G, Patti M, Bartucci M, et al. Inhibitionof DNA methylation sensitizes glioblastoma for tumor necrosis

Lund et al.





factor-related apoptosis-inducing ligand-mediated destruction. Can-cer Res 2005;65:11469–77.

38. Hopkins-Donaldson S, Ziegler A, Kurtz S, Bigosch C, Kandioler D,Ludwig C, et al. Silencing of death receptor and caspase-8 expressionin small cell lung carcinoma cell lines and tumors by DNAmethylation.Cell Death Differ 2003;10:356–64.

39. Peli J, Schr€oter M, Rudaz C, Hahne M, Meyer C, Reichmann E, et al.Oncogenic Ras inhibits Fas ligand-mediated apoptosis by down-regulating the expression of Fas. EMBO J 1999;18:1824–31.

40. Lund P, Weisshaupt K, Mikeska T, Jammas D, Chen X, Kuban RJ,et al. Oncogenic HRAS suppresses clusterin expression throughpromoter hypermethylation. Oncogene 2006;25:4890–903.

41. Pruitt K, Ulku AS, Frantz K, Rojas RJ, Muniz-Medina VM, RangnekarVM, et al. Ras-mediated loss of the pro-apoptotic response proteinPar-4 is mediated by DNA hypermethylation through Raf-independentand Raf-dependent signaling cascades in epithelial cells. J Biol Chem2005;280:23363–70.

42. Burikhanov R, Zhao Y, Goswami A, Qiu S, Schwarze SR, RangnekarVM. The tumor suppressor Par-4 activates an extrinsic pathway forapoptosis. Cell 2009;138:377–88.

43. Spencer SL, Gaudet S, Albeck JG, Burke JM, Sorger PK. Non-geneticorigins of cell-to-cell variability in TRAIL-induced apoptosis. Nature2009;459:428–32.

44. Call JA, Eckhardt SG, Camidge DR. Targeted manipulation of apop-tosis in cancer treatment. Lancet Oncol 2008;9:1002–11.

45. Available from: http://www.clinicaltrials.gov46. Wang LH, Choi YL, Hua XY, Shin YK, Song YJ, Youn SJ, et al.

Increased expression of sonic hedgehog and altered methylation ofits promoter region in gastric cancer and its related lesions. ModPathol 2006;19:675–83.

47. Wu H, Chen Y, Liang J, Shi B, Wu G, Zhang Y, et al. Hypomethylation-linked activation of PAX2 mediates tamoxifen-stimulated endometrialcarcinogenesis. Nature 2005;438:981–7.

48. Wolff EM, Byun HM, Han HF, Sharma S, Nichols PW, Siegmund KD,et al. Hypomethylation of a LINE-1 promoter activates an alternatetranscript of the MET oncogene in bladders with cancer. PLoS Genet2010;6:e1000917.

49. Garcia JS, Jain N, Godley LA. An update on the safety and efficacy ofdecitabine in the treatment of myelodysplastic syndromes. OncoTargets Ther 2010;3:1–13.






2011;10:1611-1623. Published OnlineFirst June 22, 2011.Mol Cancer Ther Per Lund, Irina Kotova, Valérie Kedinger, et al. Antagonized by DecitabineApoptosis-Inducing TRAIL by DNA Hypermethylation Is Transformation-Dependent Silencing of Tumor-Selective

Updated version

10.1158/1535-7163.MCT-11-0140doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2011/06/16/1535-7163.MCT-11-0140.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/10/9/1611.full#ref-list-1

This article cites 48 articles, 16 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/10/9/1611.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

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

Permissions

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

.http://mct.aacrjournals.org/content/10/9/1611To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-11-0140

http://mct.aacrjournals.org/content/suppl/2011/06/16/1535-7163.MCT-11-0140.DC1

http://mct.aacrjournals.org/content/10/9/1611.full#ref-list-1

http://mct.aacrjournals.org/content/10/9/1611.full#related-urls

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

mailto:[email protected]

http://mct.aacrjournals.org/content/10/9/1611


transformation-dependent silencing of tumor...

Documents