dnamethylationstatusofkeycell-cycleregulatorssuchas cdkna2/p16 and ccna1...

Personalized Medicine and Imaging

DNAMethylationStatusofKeyCell-CycleRegulatorsSuchasCDKNA2/p16 and CCNA1 Correlates with TreatmentResponse to Doxorubicin and 5-Fluorouracil in LocallyAdvanced Breast Tumors

Jovana Klajic1,2,3, Florence Busato4, Hege Edvardsen3, Nizar Touleimat4, Thomas Fleischer2,3,Ida Bukholm5,6, Anne-Lise Børresen-Dale2,3, Per Eystein Lønning7,8, J€org Tost4, andVessela N. Kristensen1,2,3

AbstractPurpose: To explore alterations in gene promoter methylation as a potential cause of acquired drug

resistance to doxorubicin or combined treatment with 5-fluorouracil and mitomycin C in human breast

cancers.

Experimental Design: Paired tumor samples from locally advanced breast cancer patients treated with

doxorubicin and 5-fluorouracil-mitomycin C were used in the genome-wide DNA methylation analysis as

discovery cohort. An enlarged cohort from the same two prospective studies as those in the discovery cohort

was used as a validation set in pyrosequencing analysis.

Results:A total of 469 geneswere differentiallymethylated after treatmentwith doxorubicin and revealed

a significant association with canonical pathways enriched for immune cell response and cell-cycle

regulating genes including CDKN2A, CCND2, CCNA1, which were also associated to treatment response.

Treatment with FUMI resulted in 343 differentially methylated genes representing canonical pathways such

as retinoate biosynthesis, gai signaling, and LXR/RXR activation. Despite the clearly different genes and

pathways involved in the metabolism and therapeutic effect of both drugs, 46 genes were differentially

methylated before and after treatmentwith both doxorubicin and FUMI.DNAmethylation profiles in genes

such as BRCA1, FOXC1, and IGFBP3, and most notably repetitive elements like ALU and LINE1, were

associated with TP53 mutations status.

Conclusion: We identified and validated key cell-cycle regulators differentially methylated before and

after neoadjuvant chemotherapy such as CDKN2A and CCNA1 and reported that methylation patterns of

these genes may be potential predictive markers to anthracycline/mitomycine sensitivity. Clin Cancer

Res; 1–10. �2014 AACR.

IntroductionInitially implemented for locally advancedbreast cancers,

primary (presurgery) systemic therapy is nowadays increas-ingly used for operable breast carcinomas. Multiple studieshave correlated a pathologic complete response to primarychemotherapy with improved prognosis (1). However,although factors like hormone receptor status, HER2 over-expression, and gene expression signatures do correlatewithpathologic complete response (2, 3), the predictive powerof each parameter is modest, and the cause of drug resis-tance remains poorly understood (4). TP53mutation statusis another strong prognostic factor, being associated withworse prognosis after treatment with anthracycline as wellas mitomycin-containing chemotherapy (5, 6).

Beside genetic abnormalities, epigenetic alterationsmay contribute to breast carcinogenesis and tumorgrowth. A large number of genes have been shown to beinactivated in breast cancer through gene silencing by

1Division of Medicine, Department of Clinical Molecular Biology andLaboratory Science (EpiGen), Akershus University Hospital, Lørenskog,Norway. 2K.G. Jebsen Center for Breast Cancer Research, Institute forClinical Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway.3Department of Genetics, Institute for Cancer Research, OUS Radiumhos-pitalet Montebello, Oslo, Norway. 4Laboratory for Epigenetics and Envi-ronment, Centre National de G�enotypage, CEA—Institut de G�enomique,Evry, France. 5Department of Surgery, Akerhus University Hospital, Oslo,Norway. 6Institute for Clinical Medicine, Faculty of Medicine, Universityof Oslo, Norway. 7Section of Oncology, Institute of Clinical Science,University of Bergen, Bergen, Norway. 8Department of Oncology, Hauke-land University Hospital, Bergen, Norway.

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

Corresponding Author: Vessela N. Kristensen, Faculty of Medicine, Insti-tute for Clinical Medicine, University of Oslo, 0372, Oslo, Norway. Phone:47-22-78-13-75/47-920-68-432; Fax: 47-22-78-13-95; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-14-0297

�2014 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org OF1

Research. on March 19, 2020. © 2014 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst October 7, 2014; DOI: 10.1158/1078-0432.CCR-14-0297

http://clincancerres.aacrjournals.org/

DNA methylation. Such findings include genes involvedin cell-cycle regulation (CDKN2A, CCND2), DNA repair(MGMT, BRCA1, MLH1, GSTP1), regulation of cell tran-scription (HOXA5), hormone and receptor-mediated cellsignaling (ER and THRb), apoptosis, and tissue invasionand metastasis (RASSF1A, RARb, TWIST, HIN1, CDH1;ref. 7). Epigenetic silencing of these genes by DNA meth-ylation is frequent and, in contrast with genetic mutation,reversible (8, 9), making them potential candidates forpharmacologic manipulation, for example, through tar-geted inhibition of DNA methyltransferases. Experimen-tal studies have indicated drug-induced alterations ingene promoter methylation as a potential cause ofacquired drug resistance (10).

The aim of this study was to explore alterations in genepromoter methylation as a potential cause of acquired drugresistance to doxorubicin or combined treatment with 5-fluorouracil (5-FU) and mitomycin C (FUMI) in humanbreast cancers. To do so, we took advantage of our tissuebank providing unique access to paired tumor samplescollected from the same patients before and after primarytreatment with doxorubicin or FUMI. We used the IlluminaInfinium27KHumanDNAMethylation BeadChip to assessto which degree the genome-wide methylation status isassociated with treatment response.

Materials and MethodsPatient cohorts

All tumor samples were collected from two prospectivestudies described in detail elsewhere (5, 6). Each patientprovided written informed consent, and the protocols wereapproved by the regional ethical committee.

Discovery cohort (genome-wide screening)Paired tumor samples from locally advanced breast can-

cer were taken before and after neoadjuvant treatment. Thestudy comprised 19 tumors treated with doxorubicin and14 tumors treated with FUMI. These patients were admittedto the Haukeland University Hospital in Bergen, Norwaybetween 1991 and 2001 (5, 6). DNA from nine normalbreast tissues were collected at the Akershus UniversityHospital (Lørenskog, Norway) and included as a controlof DNA methylation background to which tumor-specificmethylation events are compared with.

Validation cohortAn enlarged cohort comprising 85 paired tumors treated

with doxorubicin and 39 paired tumors treated with FUMIfrom the same two prospective studies as those in thediscovery cohort was used as a validation set. Samples thatwere used for 27K arrays were also among those in valida-tion cohort. Clinical and molecular parameters were avail-able for all samples. DNA from fresh-frozen tumor tissueswas used in both cohorts. Study design is shown in Fig. 1.

Classification of Treatment ResponseThe protocols enrolled patients between 1991 and 2001.

Response to therapy was classified by the UICC systemcommonly applied at that time (11) and not the morerecent RECIST system (12). Responses were classified as CR(complete response-complete disappearance of all tumorlesions), PR (partial response reduction�50% in the sumofall tumor lesions), PD (progressive disease-increase in thediameter of any individual tumor lesion by �25%), andStbD (stable disease; the condition between PR and PD).Similar to what has been done for previous reports on thesematerials (5, 6), we compared PD tumors as nonresponderswith the combined group of tumors classified as StbD/PR asresponders.

TP53 analysisMutations in the TP53 gene were analyzed with use of

genomic DNA and the temporal temperature gradient gelelectrophoresis (TTGE) strategy (5, 6). DNA fragmentscovering exons 2–11 were amplified, all with a GC clampon one of the primers, and submitted to analysis by TTGE.Mutations of the TP53genewere correlatedwith response tochemotherapy with use of the c2 method, including Yatescorrection for a limited number of observations. In addi-tion, the differences in the distribution of TP53 mutationamong patients revealing a PD and the remaining groupswere analyzed with the use of Fisher exact test.

Methylation assaysTumor and normal DNA were isolated using standard

phenol/chloroform procedure. Of note, 500 ng ofDNAwasbisulphite converted using the EpiTect 96 Bisulfite Kit(Qiagen GmbH). Effective bisulphite conversion wasverified by absolute quantification assay using AppliedBiosystem7900HT/7900HTFast Real Time PCR System

Translational RelevanceWe investigated treatment-associated changes in

the DNA methylation profiles of locally advancedbreast cancer patients treated with doxorubicin and50fluorouracil-mitomycin C, prior and after neoadju-vant chemotherapy and validates these profiles in alarger cohort. We identify and validate methylationpatterns of key cell-cycle regulators such as CDKN2Aand CCNA1 as potential predictive markers to anthra-cycline/mitomycine sensitivity. We report on the asso-ciation of differentially methylated genes before andafter treatment with canonical pathways enriched forcell-cycle regulating genes and immune cell response.We demonstrate an association between TP53 muta-tion status and DNA methylation levels in repetitiveelements like ALU and LINE1 and suggest that meth-ylation levels of these genes together with TP53 statuscould be of value in predicting response to chemo-therapy treatment. We also show an association ofCDKN2A to molecular subtypes and suggest that theLuminal B subtype could be used in defining a group,which will benefit from chemotherapy treatment.

Klajic et al.

Clin Cancer Res; 2014 Clinical Cancer ResearchOF2




Amplification and pairs of primers specific for either con-verted or unconverted DNA. Aliquote of 4 mL of bisulphate-converted DNA was used to perform genome-wide DNAmethylation profiling of 33 paired tumor samples and ninenormal breast tissues using the Illumina HumanMethyla-tion27 BeadChip. All steps were performed according to theInfinium protocol. Methylation data are deposited in GeneExpression Omnibus (GEO) repository under accessionnumber GSE59724 and may be viewed at http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc¼GSE59724. Quan-titative DNA methylation analysis of bisulphate-treatedDNA was performed in 125 tumors by pyrosequencing(13) on 22 selected genes. Thirteen genes (ABCB1, APC,BRCA1, CCNA1, CCND2, CDH1, CDKN2A, ESR1, FOSL1,GSTP1, IGFBP3, MGMT, and RARB2) were chosen forvalidation from the genome-wide analysis and seven(FOXC1, FOXC2, HMLH1, IGF2, PPP2R2B, PTEN, andRASSF1A) were selected from previous reports from us andothers for differential DNA methylation in breast cancer orbreast cancer cell lines. GlobalDNAmethylation levelswereassessed analyzing the repetitive LINE1 and ALU sequencefamilies as proxy.

Statistical analysisDataset preprocessing. All targets that contained a "zero"

value in one sample for methylated and unmethylatedsignals were removed (n ¼ 10). Second, intrasample nor-malization, which consists of color bias and backgroundlevel correction, was performed using the lumi R package.

Subsequently, quantile normalization was performed forbetween sample normalization. After preprocessing, sevenpaired samples treated with doxorubicin were excludedbecause of low-quality DNAmethylation profiles, resultingin 26 paired tumor samples that were further analyzed. Betavalues were used for further analysis.

Genes were termed differentially methylated in the dis-covery cohort when medians between the two selectedgroups (jmedian_gpA - median_gpBj) showed a differenceof at least 15% methylation. The methylation index ofsamples (Z-score) was calculated as follows: (methylationlevel of each sample � mean value of methylation levels)/SD of methylation levels. Then the sum for the genes wascalculated giving one single value (Z-score) for each sample.In the validation cohort, differences in the distribution ofmethylation between the two selected groups were assessedby the nonparametric Mann–Whitney test (on parameterswith two categories) or the Kruskal–Wallis test analysis onparameters with more than two categories. All obtained Pvalues were corrected with the Bonferroni correction meth-od in which the P values are multiplied by the number ofcomparisons.

Ingenuity pathway analysisData were also analyzed by Ingenuity Pathway Analysis

(IPA; ref. 14). Core analyses were performed to determinewhich genes from our dataset were present in alreadydefined canonical pathways. The significance of the asso-ciation between our dataset and the canonical pathways are

Locally advanced breast cancer

Discovery cohort

Doxorubicin

19 paired tumor

samples

Genome-wide screening

5-Fluouracil and

Mitomycin C

14 paired tumor samples

Validation cohort

Doxorubicin


Pyrosequencing

5-Fluouracil and

Mitomycin C


Figure 1. Study design.

DNA Methylation and Chemotherapy Response

www.aacrjournals.org Clin Cancer Res; 2014 OF3




assessed by: (i) the ratio of the number of molecules fromour gene list that map to a canonical pathway; (ii) Fisherexact test P value, P < 0.05; and (iii) FDR, FDR < 0.05.

ResultsDifferentially methylated genes in tumors before andafter treatmentwith doxorubicin and FUMI revealed byarray screen

Genome-wide DNAmethylation analysis was performedon 12 and 14 paired tumor samples from locally advancedbreast cancer taken before and after neoadjuvant treatmentwith doxorubicin and FUMI.

A total of 486 CpGs representing 469 genes were founddifferentially methylated (jmedian_gpA - median_gpBj >15% of methylation) before and after treatment with doxo-rubicin; out of these, 230 genes had higher methylationlevel before and 239 genes after treatment. The DNAmeth-ylation status was also compared before and after treatmentwith FUMI identifying 380 CpGs corresponding to 343genes to be differentially methylated (215 genes had highermethylation level before and 128 genes higher methylationlevel after treatment). Among these, 469 and343 geneswerethose involved in regulation of cell transcription, immu-noregulatory and inflammatory processes, cell-cycle regu-lators, growth factors, and tumor suppressor genes(Supplementary Table S1a and S1b).

Next, IPA was applied to the lists of differentially meth-ylated genes before/after treatment with both drugs sepa-rately. Interestingly, the pathway analysis of the 469 genesdifferentially methylated before and after treatment withdoxorubicin revealed a significant association with thefollowing canonical pathways: LXR/RXR activation, com-munication between innate and adaptive immune cells(Supplementary Fig. S1), and the role of cytokines inmediating communication between immune cells (Supple-mentary Table S2a). IPAof the 343differentiallymethylatedgenes before and after treatment with FUMI revealed asignificant association with canonical pathways such asretinoate biosynthesis, gai signaling (Supplementary Fig.S2), LXR/RXR activation, and cardiac b-adrenergic signaling(the four significant canonical pathways are given in Sup-plementary Table S2b).

Despite the clearly different genes and pathways involvedin the metabolism and therapeutic effect of both drugs, 46genes were differentially methylated before and after treat-ment with both drugs. The full list of the 46 genes and theirfunctions are given in Supplementary Table S3a and S3b.

DNAmethylated genes from array screen and responseto treatment with doxorubicin and FUMI

A total of 1,468 CpGs representing 1,310 genes werefound differentially methylated (jmedian_gpA � med-ian_gpBj > 15% of methylation) in tumors progressing ontherapy (PD) compared with tumors retaining a stabledisease or an objective response after treatment with doxo-rubicin. Out of these, 639 genes had higher methylationlevels in responders compared with nonresponders beforetreatment. Methylation levels of responders were signifi-

cantly higher than normal controls both before and aftertreatment with doxorubicin, whereas methylation levels ofnonresponders were close to those in normal controls(Fig. 2A). Differentially methylated genes and their func-tions are listed in Supplementary Table S1c. These 1,310genes were then used as input for an IPA. The top significantcanonical pathways showed a significant association withtwo canonical pathways (Supplementary Table S2c). Again,communication between innate and adaptive immune cellscame on top, as in the case ofmost differentiallymethylatedgenes found when comparing before and after treatmentwith doxorubicin.

Differentially methylated genes were also observed inresponders versus nonresponders for 1,220 genes repre-sented by 1,424 CpGs in patients treated with 5-FU/mito-mycin and 546 genes had higher methylation levels inresponders comparedwith nonresponders. Compared withthe normal controls, methylation levels of responders weresignificantly higher before treatment, whereas methylationlevels of nonresponders were close to those in normalcontrols. Interestingly after treatment, in contrast with whatwas observed for doxorubicin, methylation levels of non-responders were significantly higher compared with both,responders and normal controls (Fig. 2B). Differentiallymethylated genes and their functions are listed in Supple-mentary Table S1d. All differentially methylated 1,220genes were then used as input for IPA. The top significantcanonical pathways are shown in Supplementary Table S2d.Canonical pathway gai signaling involved in the G-proteincascade was one also found among the most differentiallymethylated genes before and after treatment with FUMI.

When comparing the differentially methylated genesassociated to treatment response to both drugs, the meth-ylation status of 333 genes was associated to response toboth treatment types. The full list of genes is given inSupplementary Table S4, including CD33, CD40, ESR2,andWT1 as well as CCNA1, CCND2, and CDKN2A, whichnotably were also found differentiallymethylated after eachtreatment. Pathway analysis on 323 genes revealed threesignificant canonical pathways: cAMP-mediated signaling,G-protein-coupled receptor signaling, and Gai signaling(Supplementary Table S5).

Validated genes differentially methylated before andafter treatment

Validation experiment was performed using the highlyquantitative pyrosequencing method in 85 paired tumorstreated with doxorubicin and 39 paired tumors treatedwith FUMI from the same two prospective studies as thosein the discovery cohort. Of the 22 genes selected forvalidation, 13 and six genes remained significantly dif-ferentially methylated either before or after treatmentwith doxorubicin and FUMI, respectively (Table 1 and2). After correction for multiple testing (Bonferroni cor-rection), the observed differences in DNA methylationlevels before and after treatment with doxorubicin werestill significant for all discovered genes except for GSTP1(marked in Table 1) and for all initially found genes

Klajic et al.





before and after treatment with FUMI. Three genes(CCNA1, CCND2, and CDKN2A) were consistentlyreported with the same results in the doxorubicin discov-ery and validation cohorts, whereas CDKN2A was foundboth in the FUMI discovery and validation cohort.

Validated genes associated with response totreatmentPyrosequencing analysis confirmed 15 and 13 genes

differentially methylated either before or after treatmentbetween responders and nonresponders in the doxorubicinand FUMI-treated tumors, respectively (Table 2). Aftercorrection for multiple testing (Bonferroni correction),differences in DNA methylation levels between respondersand nonresponders were still significant for all initiallyfound genes except for ABCB1 and FOXC1 before treatmentand CCND2, IGF2, and LINE1 after treatment with doxo-rubicin (marked in Table 2). Five of these genes (CCNA1,CCND2, CDKN2A, FOSL1, and APC) were also discoveredin the array analysis and then validated by pyrosequencingin the larger cohort. After correction for multiple testing(Bonferroni correction), differences in DNA methylation

levels between responders and nonresponders were signif-icant for all initially observed genes except forAPC,CCND2,and PTEN before treatment and CDKN2A, CCNA1, andPTEN after treatment with FUMI (marked in the Table 2).Six genes (ABCB1, BRCA1, CCNA1, CCND2, CDKN2A, andRARB) were discovered in the genome-wide analysis andconfirmed by pyrosequencing.

The DNA methylation status of cell-cycle regulatorygenes, CCNA1, CCND2, and CDKN2A was associated totreatment response for both doxorubicin and FUMI (Fig.3). These genes were significantly differentially methyl-ated before/after treatment with doxorubicin and before/after treatment between responders. Lower levels of meth-ylation were observed after treatment in the respondergroups. Changes of methylation levels in nonresponderswere significant for CDKN2A, and levels for all three geneswere close to those in normal controls (especially aftertreatment). In general, similar patterns of methylationlevels were observed before and after treatment withFUMI. The main difference was higher levels of methyl-ation of nonresponders compared with responders of theCCNA1 gene.

Met

hyl

atio

n in

dex

Met

hyl

atio

n in

dex

-

Normal Responders-after Nonresponders-after

P = 0.001

P = 0.001

Normal Responders-before Nonresponders-before

Normal Responders-after Nonresponders-afterNormal Responders-before Nonresponders-before

A

Met

hyl

atio

n in

dex

B

Met

hyl

atio

n in

dex

P = 2.9e–4

P = 0.002

Doxo before treatment Doxo after treatment

FUMI before treatment FUMI after treatment

P = 0.033

800

700

600

500

400

300

600

500

400

300

700

600

500

400

300

550

500

450

400

350

300

250

Figure 2. Boxplots illustrating significant association between methylation status of response groups and normal breast tissue controls before and aftertreatment with doxorubicin and FUMI. A, significantly higher methylation levels in responders compared with nonresponders and normal controls areobserved before and after treatment with doxorubicin. B, significantly higher methylation levels in responders compared with nonresponders, and normalcontrols before treatmentwith FUMI. Significantly highermethylation levels in nonresponders comparedwith responders and normal controls after treatment.






Correlation with TP53 mutations and gene expressionprofiles

We compared the observed DNA methylation profileswith theTP53mutations status before and after treatment inboth patient cohorts. In the patient cohort treated withFUMI, TP53-mutated tumors had significantly lower DNAmethylation levels both before and after treatment in threegenes:ABCB1,CCND2, andPPP2R2B (Fig. 4A). In addition,six genes were significantly differentially methylatedbetween TP53 wild-type and mutated tumors before treat-ment only (Supplementary Table S6). In the doxorubicin-treated cohort, tumors with TP53 mutations had signifi-cantly lower DNA methylation levels in CCND2 andRASSF1A before treatment (P ¼ 0.023 and 0.035).

We then compared the observed methylation patternswith well-defined tumor subclasses. Interestingly, onlyCDKN2A, which was found associated both with degree ofmethylation before and after treatment as well as withtreatment response to FUMI and doxorubicin, was alsodifferentially methylated between different molecular sub-classes, both before/after doxorubicin treatment. However,although the methylation levels of CDKN2A in Luminal Aand Luminal B did not differ significantly before treatment,themethylation level of this genewas significantly decreased

after treatment in the Luminal B subtype. Significant differ-ences inmethylation levels were observed between LuminalA and ERBB2 enriched subtype both before/after treatment.In all cases, themethylation levelswere lower after treatmentbut the fold change ofmethylation levelswas strongest in theLuminal B and a fraction of the basal-like subtype. Thechanges were less among tumors belonging to the LuminalA subtype (Fig. 4B).

DiscussionThe use of neoadjuvant chemotherapy for patients with

large and locally advanced breast cancer before surgery hasbecome an established practice allowing shrinkage of theprimary tumor before surgery. Whether the mechanisms of

Table 1. Differentially methylated genes beforeand after neoadjuvant chemotherapy with bothdrugs

Before/after

GeneBefore (% ofmethylation)

After (% ofmethylation) P

DoxorubicinALU 17.9 20.3 2.3e–19BRCA1 87.2 85.8 0.001CCNA1 26.8 15.6 4.7e–4CCND2 15.4 8.4 5.3e–11CDH1 11.4 7.5 6.1e–18FOSL1 9.1 1.7 6.1e–18HMLH1 4.2 3.6 8.3e–15IGF2 45.5 61.4 2.7e–11LINE1 51 40.6 7.1e–6MGMT 8.8 2.7 1.1e–17CDKN2A 49.7 30.4 1.8e–6PTEN 10.8 5.4 1.7e–5RASSF1A 27.7 12.1 7.1e–5GSTP1 11.3 4.6 0.013a

FUMIALU 40.1 19.6 6.3e–9BRCA1 4.3 7.2 1.6e–4CDH1 5.2 8 2.7e–5HMLH1 5.9 6.7 1.2e–7CDKN2A 41 31.9 2.6e–4PTEN 2.9 4.3 1.7e–4

aNot significant after Bonferroni correction.

Table 2. Differentially methylated genes beforeand after neoadjuvant chemotherapy inresponders versus nonresponders for bothdrugs

Before treatment(responders vs.nonresponders)

After treatment(responders vs.nonresponders)

Gene P P

DoxorubicinAPC 2.3e–11 1.5e–9ABCB1 0.006a 2.5e–10BRCA1 1.3e–11 2e–10CCNA1 2.1e–4CCND2 2.3e–8 0.05a

CDH1 1.8e–11 3e–9CDKN2A 4.1e–11 7.1e–8FOSL1 1.1e–11FOXC1 0.008a 7.5e–10FOXC2 5.3e–12 2.1e–10IGF2 0.014a

LINE1 0.019a

PPP2R2B 2.1e–7 1.2e–8PTEN 1.3e–11 8.3e–8RARB 5.1e–11 2.7e–10

FUMIAPC 0.01a 1.7e–6ABCB1 2.4e–4 1.8e–7BRCA1 3.2e–9 1.1e–8CCNA1 6.6e–5 0.038a

CCND2 0.05a

CDKN2A 3e-7 0.001a

CDH1 3.8e–6FOSL1 0.033FOXC1 5.5e–5 2.2e–6FOXC2 7.2e–8 6.5e–7PPP2R2B 5.3e–8 4.8e–7PTEN 0.006a 0.017a

RARB 9.8e–8 3.2e–6

aNot significant after Bonferroni correction.

Klajic et al.





primary and acquired resistance are similar is still unknown.The few studies conducted on potential associationsbetween gene expression profiles and chemoresistance havenot provided definite answers (15, 16). Genes are expressedto influence whole "pathways"; thus, it is likely that resis-tance is not related to either individual factors or expression,but rather, should be considered in the context of "func-tional cascades" (17).Functional pathways may be inactivated not only by

genetic, but also by epigenetic events. These may either bede novo or acquired in response to therapy, mutation effects,or cellular selection.Here, we took advantage of the fact thatwe have at hand before and after treatment samples fromtwo prospective studies. We found genes differentiallymethylated between responders and nonresponders work-ing in the same pathway. Notably, these relate to immuneresponse and cell-cycle control. Emerging evidence indi-cates that many chemotherapeutic agents exert immunestimulatory effects (18, 19). Studies of immunogenic celldeath after chemotherapy revealed that doxorubicin isinvolved in processes which lead to expression of toll-likereceptor 4 (TLR4) by dendritic cells (20, 21). It was alsoshown that TLR4 expressed on tumor cells contributes totumor progression by promoting tumor cell proliferation,(22, 23), and that silencing of TLR4 could reduce the

metastatic potential of metastatic tumor cells (24). In thepresent study, TLR4 was highly methylated in responderscompared with nonresponders, implying that silencing ofthis gene by promoter methylation may predict betterprognosis. Furthermore, studies are warranted to addresswhether epigenetics changes in other genes acting up ordownstream in this pathway may be related to outcome aswell.

In the present study, we identified genes differentiallymethylated before and after neoadjuvant chemotherapyinvolved in cell-cycle control such as CDKN2A, CCNA1,andCCND2. Toour knowledge, this is the first report on theDNA methylation level of these genes and treatmentresponse in locally advanced breast cancer treated withdoxorubicin and FUMI. CDKN2A promoter methylationhas been detected in breast and many other human cancers(25). We previously showed that hypermethylation ofCDKN2A is a late event during breast carcinogenesis, lead-ing to its inactivation in late stage breast cancers (26).Expression of CCND2 is lost in the majority of breastcancers, and a mechanism underlying this loss is the pro-moter hypermethylation of cyclin D2 (27). In addition,hypermethylation of CCNA1 has been correlated to breastcancer progression (28). Again, further studies are needed tovalidate these findings. A study on the DNA methylation

Ave

rag

e D

NA

met

hyl

atio

n (

%)

Ave

rag

e D

NA

met

hyl

atio

n (

%)

Ave

rag

e D

NA

met

hyl

atio

n (

%)

Ave

rag

e D

NA

met

hyl

atio

n (

%)

Ave

rag

e D

NA

met

hyl

atio

n (

%)

Ave

rag

e D

NA

met

hyl

atio

n (

%)

CCND2

CCNA1CCND2

CCNA1CDKN2A

CDKN2A

Normal All tumors Responders Nonresponders Normal All tumors Responders Nonresponders Normal All tumors Responders Nonresponders

P = 2.3e–8P = 5.3e–11

P = 4.1e–11P = 1.8e–6

P = 4.7e–4P = 2.1e–4

P = 3e–7P = 2.6e–4

P = 0.042P = 0.001 P = 0.08

before–after before–after before–after before–after before–after before–after before–after before–after before–after

before–after before–after before–after before–after before–after before–after before–after before–after before–afterNormal All tumors Responders Nonresponders Normal All tumors Responders Nonresponders Normal All tumors Responders Nonresponders

Dox

oF

UM

I

P = 7.1e–8

P = 6.6e–5

10

100

2030

4050

100

2030

4050

2030

4050

6070

100

2030

4050

6070

100

2030

4050

6070

100

2030

4050

6070

Figure 3. Average percentage of DNAmethylation levels forCDKN2A, CCND2, andCCNA1 gene in doxorubicin and 5-FUMI-treated breast cancers. For eachgene, difference in methylation between all samples before/after, between responders before/after, and between nonresponders before/after and normalsamples are shown.






profiles in breast cancer cell lines and their doxorubicin-resistant counterpart showed that methylation levels ofsome genes had a directional tendency when acquiringdoxorubicin resistance (10). Some genes (ABCB1, APC,GSTP1) were becoming hypomethylated and some(BRCA1, CDH1, ESR1) got hypermethylated in breast can-cer cells resistant to the drug. In the present study, meth-ylation levels of ABCB1 and APC were lower in nonrespon-ders compared with the responders, whereas the methyla-tion level of CHD1was higher in nonresponders comparedwith the responders.

In the present study, we observed lower levels of meth-ylation in TP53-mutated tumors compared with wild-typefor some genes both before and after treatment with bothdrugs. It was previously shown that TP53mutations, togeth-er with DNA methylation, play a role in keeping largefamilies of interspersed (SINEs) and tandem repeats tran-scriptionally dormant (29). The authors demonstrated thatTP53-mutated cells treatedwith aDNAdemethylating agent

show transcriptional activation of normally silent SINEs.Activation of repeats resulting from combined lack of TP53function and DNA methylation was followed by inductionof the type I IFN signaling pathway (29). Loss of TP53function and hypomethylation occur naturally in tumors,and transcription of normally silent and heavilymethylatedsequences that represent repeats has been shown to occur intumors (30, 31). In present study, we showed that ALU(SINEs) and LINE1 had significantly lower methylationlevels in TP53-mutated tumors compared with wild-type.Identification of these processes in tumors could be used fordiagnosis of specific tumor types and stages of progressionof the given tumor.

Studies on different expression subtypes in breast cancerhave shown that different subtypes have a different under-lying biology strongly influenced by TP53 mutation statuswhich is reflected in the methylation profile. Thus, basal-like tumors contain in general a low degree of DNA meth-ylation but frequently harbor TP53 mutations (32).

Ave

rag

e D

NA

met

hyla

tio

n (

%)

Ave

rag

e D

NA

met

hyla

tio

n (

%)

Ave

rag

e D

NA

met

hyla

tio

n (

%)

Ave

rag

e D

NA

met

hyla

tio

n (

%)

ABCB1-TP53

PPP2R2B-TP53

CCND2-TP53

wt-before

010

2030

4050

010

2030

4050

mut-before wt-after mut-after

wt-before mut-before wt-after mut-after

wt-before mut-before wt-after mut-after

P = 0.041

P = 0.044

P = 0.036

P = 0.036

P = 0.045

P = 0.017

before–after before–after before–after before–after before–afterLuminal A Luminal B ERBB2 Basal Normal-like

A

BP = 0.009

P = 0.005

Doxo before/after CDKN2A

05

1015

2025

3035

2030

4050

60

Figure 4. A, boxplots illustrating significant association between methylation status and TP53 mutation status. For each gene, difference in methylationbetween wild-type and TP53-mutated tumors before and after treatment is given. B, boxplots illustrating the differential methylation of CDKN2A betweendifferent molecular subtypes before and after treatment with doxorubicin.

Klajic et al.





Interestingly Luminal B tumors and some of the basal-liketumors had significantly lower methylation levels aftertreatment. Patients with luminal B tumors were among thebetter responders so patients of this subtype might benefitfrom treatment with doxorubicin.

ConclusionWe identified and validated key cell-cycle regulatory

genes differentially methylated before and after neoadju-vant chemotherapy such as CDKN2A and CCNA1. Thus,methylation patterns of these genes may be potentialpredictive markers to anthracycline/mitomycine sensitiv-ity. Furthermore, we identified genes differentially meth-ylated between TP53-mutated and wild-type tumorsamong which are LINE1 and ALU elements and suggestedthat methylation levels of these genes together with TP53status could be of value in predicting response to che-motherapy treatment. However, further studies are nec-essary requiring additional patient cohorts treated withdoxorubicin and FUMI to confirm predictive value of allgenes mentioned above.

Disclosure of Potential Conflicts of InterestJ. Tost reports receiving commercial research support from DBV Tech-

nologies and has provided expert testimony for Roche Diagnostics, travelexpenses to scientific conferences. No potential conflicts of interest weredisclosed by the other authors.

Authors' ContributionsConception and design: A.-L. Børresen-Dale, V.N. KristensenDevelopment of methodology: N. Touleimat, J. TostAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): F. Busato, N. Touleimat, I.R.K. Bukholm, P.E.Lønning, J. Tost, V.N. KristensenAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): J. Klajic, F. Busato, H. Edvardsen,T. Fleischer, V.N. KristensenWriting, review, and/or revision of the manuscript: J. Klajic, F. Busato,H. Edvardsen, T. Fleischer, A.-L. Børresen-Dale, P.E. Lonning, J. Tost, V.N.KristensenAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): V.N. KristensenStudy supervision: J. Tost, V.N. KristensenOther (experimental work): J. KlajicOther (supervisor of the PhD thesis of J. Klajic): V.N. Kristensen

AcknowledgmentsThe authors thank Grethe I. Grenaker Alnæs for technical support in

performing the laboratory experiment.

Grant SupportThis work was supported by grant no. 193387/V50 from the Norwegian

Research Council (to V.N. Kristensen and A.-L. Borresen-Dale) and by theNorwegian Cancer Society (Grant no.4196163832; to V.N. Kristensen).

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 February 10, 2014; revised September 25, 2014; acceptedSeptember 25, 2014; published OnlineFirst October 7, 2014.

References1. Lonning PE. Molecular basis for therapy resistance. Mol Oncol

2010;4:284–300.2. Kaufmann M, von MG, Bear HD, Buzdar A, McGale P, Bonnefoi H,

et al. Recommendations from an international expert panel on theuse of neoadjuvant (primary) systemic treatment of operablebreast cancer: new perspectives 2006. Ann Oncol 2007;18:1927–34.

3. Prat A, Perou CM. Deconstructing the molecular portraits of breastcancer. Mol Oncol 2011;5:5–23.

4. Lonning PE, Knappskog S. Mapping genetic alterations causing che-moresistance in cancer: identifying the roads by tracking the drivers.Oncogene 2013;32:5315–30.

5. Geisler S, Lonning PE, Aas T, Johnsen H, Fluge O, Haugen DF, et al.Influence of TP53 gene alterations and c-erbB-2 expression on theresponse to treatment with doxorubicin in locally advanced breastcancer. Cancer Res 2001;61:2505–12.

6. Geisler S, Borresen-Dale AL, Johnsen H, Aas T, Geisler J, Akslen LA,et al. TP53 gene mutations predict the response to neoadjuvanttreatment with 5-fluorouracil andmitomycin in locally advanced breastcancer. Clin Cancer Res 2003;9:5582–8.

7. Jovanovic J, Ronneberg JA, Tost J, Kristensen V. The epigenetics ofbreast cancer. Mol Oncol 2010;4:242–54.

8. Jones PA, Baylin SB. The fundamental role of epigenetic events incancer. Nat Rev Genet 2002;3:415–28.

9. Sandhu R, Rivenbark AG, Coleman WB. Enhancement of chemother-apeutic efficacy in hypermethylator breast cancer cells through tar-geted and pharmacologic inhibition of DNMT3b. Breast Cancer ResTreat 2012;131:385–99.

10. BoettcherM,Kischkel F,Hoheisel JD.High-definitionDNAmethylationprofiles from breast and ovarian carcinoma cell lines with differingdoxorubicin resistance. PLoS ONE 2010;5:e11002.

11. Hayward JL, Carbone PP,Heusen JC, KumaokaS, Segaloff A, RubensRD. Assessment of response to therapy in advanced breast cancer. BrJ Cancer 1977;35:292–8.

12. Therasse P, Arbuck SG, Eisenhauer EA, Wanders J, Kaplan RS, Rubin-stein L, et al. New guidelines to evaluate the response to treatment insolid tumors. European organization for research and treatment ofcancer, national cancer institute of the United States, national cancerinstitute of Canada. J Natl Cancer Inst 2000;92:205–16.

13. Tost J, Gut IG. DNA methylation analysis by pyrosequencing. NatProtoc 2007;2:2265–75.

14. Ingenuity�Systems. Ingenuity Pathway Analysis. [Accessed July 162014]. Available from: http://www ingenuity com/products/ipa.

15. Buchholz TA, Stivers DN, Stec J, Ayers M, Clark E, Bolt A, et al. Globalgene expression changes during neoadjuvant chemotherapy forhuman breast cancer. Cancer J 2002;8:461–8.

16. Chang JC, Wooten EC, Tsimelzon A, Hilsenbeck SG, Gutierrez MC,Elledge R, et al. Gene expression profiling for the prediction of ther-apeutic response to docetaxel in patients with breast cancer. Lancet2003;362:362–9.

17. Lonning PE. Genes causing inherited cancer as beacons to identify themechanisms of chemoresistance. Trends Mol Med 2004;10:113–8.

18. Emens LA, Machiels JP, Reilly RT, Jaffee EM. Chemotherapy: friend orfoe to cancer vaccines? Curr Opin Mol Ther 2001;3:77–84.

19. Lake RA, Robinson BW. Immunotherapy and chemotherapy–a prac-tical partnership. Nat Rev Cancer 2005;5:397–405.

20. Obeid M, Tesniere A, Ghiringhelli F, Fimia GM, Apetoh L, Perfettini JL,et al. Calreticulin exposure dictates the immunogenicity of cancer celldeath. Nat Med 2007;13:54–61.

21. Apetoh L, Ghiringhelli F, Tesniere A, Obeid M, Ortiz C, Criollo A, et al.Toll-like receptor 4-dependent contribution of the immune system toanticancer chemotherapyand radiotherapy.NatMed2007;13:1050–9.

22. Huang B, Zhao J, Li H, He KL, Chen Y, Chen SH, et al. Toll-likereceptors on tumor cells facilitate evasion of immune surveillance.Cancer Res 2005;65:5009–14.

23. Szczepanski MJ, Czystowska M, Szajnik M, Harasymczuk M, Boy-iadzis M, Kruk-Zagajewska A, et al. Triggering of Toll-like receptor 4expressed on human head and neck squamous cell carcinoma






promotes tumor development and protects the tumor from immuneattack. Cancer Res 2009;69:3105–13.

24. Hua D, Liu MY, Cheng ZD, Qin XJ, Zhang HM, Chen Y, et al. Smallinterfering RNA-directed targeting of Toll-like receptor 4 inhibitshuman prostate cancer cell invasion, survival, and tumorigenicity. MolImmunol 2009;46:2876–84.

25. Herman JG, Merlo A, Mao L, Lapidus RG, Issa JP, Davidson NE, et al.Inactivation of the CDKN2/p16/MTS1 gene is frequently associatedwith aberrant DNA methylation in all common human cancers. CancerRes 1995;55:4525–30.

26. Klajic J, Fleischer T, Dejeux E, Edvardsen H, Warnberg F, Bukholm I,et al. Quantitative DNA methylation analyses reveal stage dependentDNA methylation and association to clinico-pathological factors inbreast tumors. BMC Cancer 2013;13:456.

27. Evron E, Umbricht CB, Korz D, Raman V, Loeb DM, Niranjan B, et al.Loss of cyclin D2 expression in the majority of breast cancers isassociated with promoter hypermethylation. Cancer Res 2001;61:2782–7.

28. Starlard-Davenport A, Kutanzi K, Tryndyak V, Word B, Lyn-Cook B.Restoration of the methylation status of hypermethylated gene pro-moters by microRNA-29b in human breast cancer: a novel epigenetictherapeutic approach. J Carcinog 2013;12:15.

29. Leonova KI, Brodsky L, Lipchick B, Pal M, Novototskaya L, ChenchikAA, et al. p53 cooperates with DNA methylation and a suicidal inter-feron response to maintain epigenetic silencing of repeats and non-coding RNAs. Proc Natl Acad Sci U S A 2013;110:E89-E98.

30. Ting DT, Lipson D, Paul S, Brannigan BW, Akhavanfard S, Coffman EJ,et al. Aberrant overexpression of satellite repeats in pancreatic andother epithelial cancers. Science 2011;331:593–6.

31. Li TH, Kim C, Rubin CM, Schmid CW. K562 cells implicate increasedchromatin accessibility in Alu transcriptional activation. Nucleic AcidsRes 2000;28:3031–9.

32. Holm K, Hegardt C, Staaf J, Vallon-Christersson J, Jonsson G, OlssonH, et al. Molecular subtypes of breast cancer are associated withcharacteristic DNA methylation patterns. Breast Cancer Res 2010;12:R36.


Klajic et al.




Published OnlineFirst October 7, 2014.Clin Cancer Res Jovana Klajic, Florence Busato, Hege Edvardsen, et al. Advanced Breast TumorsResponse to Doxorubicin and 5-Fluorouracil in Locally

Correlates with TreatmentCCNA1/p16 and CDKNA2as DNA Methylation Status of Key Cell-Cycle Regulators Such

Updated version

10.1158/1078-0432.CCR-14-0297doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2014/10/08/1078-0432.CCR-14-0297.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/2014/11/22/1078-0432.CCR-14-0297To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-14-0297

http://clincancerres.aacrjournals.org/content/suppl/2014/10/08/1078-0432.CCR-14-0297.DC1

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/early/2014/11/22/1078-0432.CCR-14-0297


dnamethylationstatusofkeycell-cycleregulatorssuchas cdkna2/p16 and ccna1...

Documents