dna hypomethylation contributes to genomic instability and ... · by global dna hypomethylation,...

Research Article

DNA Hypomethylation Contributes to GenomicInstability and Intestinal Cancer InitiationKaryn L. Sheaffer, Ellen N. Elliott, and Klaus H. Kaestner

Abstract

Intestinal cancer is a heterogeneous disease driven by geneticmutations and epigenetic changes. Approximately 80% of spo-radic colorectal cancers are initiated by mutation and inactivationof the adenomatous polyposis coli (APC) gene, which results inunrestrained intestinal epithelial growth and formation of ade-nomas. Aberrant DNA methylation promotes cancer progressionby the inactivation of tumor suppressor genes via promotermethylation. In addition, global DNA hypomethylation is oftenseen before the formation of adenomas, suggesting that it con-tributes to neoplastic transformation. Previous studies employedmice with a hypomorphic mutation in DNA methyltransferase 1(Dnmt1), which exhibited constitutive global DNA hypomethyla-tion and decreased tumorigenesis in the ApcMin/þ mouse model ofintestinal cancer. However, the consequences of intestinal epithe-

lial-specific acute hypomethylation during ApcMin/þ tumor initia-tion have not been reported. Using temporally controlled intes-tinal epithelial-specific gene ablation, we show that total loss ofDnmt1 in the ApcMin/þ mouse model of intestinal cancer causesaccelerated adenoma initiation. Deletion of Dnmt1 precipitatesan acute response characterized by hypomethylation of repet-itive elements and genomic instability, which surprisingly isfollowed by remethylation with time. Two months post-Dnmt1ablation, mice display increased macroadenoma load, con-sistent with a role for Dnmt1 and DNA methylation in main-taining genomic stability. These data suggest that DNA hypo-methylation plays a previously unappreciated role in intestinaladenoma initiation. Cancer Prev Res; 9(7); 534–46. �2016 AACR.

See related article by Lee and Laird, p. 509

IntroductionIntestinal and colorectal cancers aremultifactorial diseaseswith

various inputs including diet, environment, genetic mutations,and epigenetic abnormalities. The disease first manifests as anoverproliferation defect in the formofpolyps that, if not removed,can progress to precancerous adenomas. Further transition toinvasive and metastatic cancer is due to the accumulation ofmultiple genetic mutations and epigenetic changes that alter geneexpression patterns, driving neoplastic transformation andgrowth (1). However, the extent to which epigenetic modifica-tions, and specifically DNA methylation, contribute to the initi-ation and progression of intestinal cancer is unclear.

Neoplastic tissues frequently display global DNA hypomethy-lation, which can occur passively through loss of the mainte-nance methyltransferase, DNMT1, or actively by TET enzyme–mediated oxidation of methyl cytosine, followed by base exci-sion repair (2–4). Alternatively, some tumors exhibit increased

DNA methylation on specific tumor suppressor gene promoters,which requires the de novo DNA methyltransferases, DNMT3Aand DNMT3B. Hypermethylation of tumor suppressor genepromoters causes decreased activity of the associated genes andpromotes cancer cell growth in vitro (5). Comparison of humancolon adenomas to control tissue revealed that although prox-imal promoter CpG islands are often hypermethylated intumors, DNA hypomethylation is found genome-wide, includingneighboring CpG island shores and repetitive elements (6, 7).Hypomethylated intergenic and intronic regions appear early inthe transition from normal to neoplastic (8, 9), suggesting amoreimportant role for DNA hypomethylation in cancer initiationthan previously appreciated.

Previously, the role of Dnmt1 in intestinal tumorigenesis wasstudied in the ApcMin/þ mouse model. This paradigm mimics thehereditary human colon cancer syndrome Familial AdenomatousPolyposis (FAP), which is caused by germline mutations of theAPC gene (10). LOH at the APC locus causes b-catenin stabiliza-tion and unrestricted Wnt activity, resulting in the formation ofmacroscopic adenomasby3months of age (11–13). LOH is also acommon triggering mechanism for sporadic intestinal and colo-rectal tumor development (1), rendering the ApcMin/þ mouse auseful model for the study of intestinal cancer initiation.

Multiple hypomorphicDnmt1paradigms show complete blockofmacroscopic tumor formation (14–18). The authors concludedthat Dnmt1 and DNA methylation are required for adenomadevelopment in the ApcMin/þ model. These studies employedhypomorphic Dnmt1 mice, which express constitutively reducedDnmt1 levels from earliest development onward in all tissues(15), including nonepithelial cells, such as myofibroblasts.Mesenchymally expressed genes have been shown to be strongmodifiers of the ApcMin/þ tumor load (19) and could thus con-tribute to the observed phenotype in the Dnmt1-hypomorphic

Department of Genetics and Institute for Diabetes, Obesity andMetabolism, Perelman School of Medicine, Universityof Pennsylvania,Philadelphia, Pennsylvania.

Note: Supplementary data for this article are available at Cancer PreventionResearch Online (http://cancerprevres.aacrjournals.org/).

K.L. Sheaffer and E.N. Elliott contributed equally to this article.

CorrespondingAuthor:Klaus H. Kaestner, Department of Genetics and Institutefor Diabetes, Obesity and Metabolism, Perelman School of Medicine, Universityof Pennsylvania, 12-126 Translational Research Center, 3400 Civic Center Bou-levard, Philadelphia, PA 19104. Phone: 215-898-8759; Fax: 215-573-5892; E-mail:[email protected]

doi: 10.1158/1940-6207.CAPR-15-0349

�2016 American Association for Cancer Research.

CancerPreventionResearch

Cancer Prev Res; 9(7) July 2016534

Research. on January 26, 2021. © 2016 American Association for Cancercancerpreventionresearch.aacrjournals.org Downloaded from

Published OnlineFirst February 16, 2016; DOI: 10.1158/1940-6207.CAPR-15-0349

http://cancerpreventionresearch.aacrjournals.org/

ApcMin/þ intestine. Furthermore, continuous DNA hypomethyla-tionmaymask acute affects during the neoplastic transformation,which occurs prior to reported observations ofmacroadenomas at6 months of age. Because of these limitations, the precise mech-anism of how DNA methylation acts within the intestinal epi-thelium during adenoma initiation remains to be determined.

To address this important knowledge gap, we employed atemporally controlled, intestinal epithelial-specific gene ablationmodel to delete Dnmt1 and decrease maintenance DNA methyl-ation. Ablation ofDnmt1 causes an acute phenotype characterizedby global DNA hypomethylation, genome instability, and apo-ptosis. The severe effects of Dnmt1 ablation result in a dramaticincrease in adenoma initiation by accelerated LOH at the Apclocus. These results strongly support a role for DNA hypomethy-lation in chromosomal instability and tumor initiation.

Materials and MethodsMice

Dnmt1loxP/loxP mice were provided by Rudolf Jaenisch (20),Villin-Cre-ERT2 mice were received from Sylvia Robine (21), andApcMin/þ mice from the Jackson Laboratories (10, 12). For Dnmt1deletion experiments, Cre recombination activity was induced by3 daily intraperitoneal injections of 1.6 mg tamoxifen (Sigma) inan ethanol/sunflower oil mixture. In all experiments, littermatecontrols without the VillinCreERT2 transgene were also tamox-ifen-treated. All procedures involving mice were conducted inaccordance with approved Institutional Animal Care and UseCommittee protocols.

HistologyTissues were isolated and fixed using 4% paraformaldehyde in

PBS and then embedded in paraffin. Antigen retrieval was per-formed using the 2100 Antigen-Retriever in Buffer A (ElectronMicroscopy Sciences) and standard immunostaining procedureswere performed for Dnmt1 (Santa Cruz), E-cadherin (BD Trans-duction Laboratories), Sox9 (Millipore), Ki67 (BD Pharmingen),b-catenin (BD Transduction Laboratories), CycD1 (Biocare Med-ical), Epcam (Abcam), and gH2AX (Cell Signaling Technology).TUNEL staining was performed using TUNEL Label and Enzyme(Roche) and AlexaFluor 555-aha-dUTP (Life Technologies). Per-centages of TUNELþ nuclei were determined by counting thenumber of nuclei positive for TUNEL staining and dividing bythe total number of nuclei in the image. The same method wasperformed to quantify the percentage of gH2AXþ cells. Allmicros-copy was performed on a Nikon Eclipse 80i.

Laser capture microdissectionTumor and nontumor epithelial cell DNAwas collected using a

Leica LMD7000 Laser Microdissection microscope and the Arc-turus PicoPure DNA Isolation Kit (Applied Biosystems).

Targeted bisulfite sequencingMouse gDNA (100 ng)was bisulfite converted using the Epitect

Bisulfite Kit (Qiagen). Template DNA was amplified using KAPAHIFIUracelþ (KAPA)withprimers directed to the LINE1elements(22), the H19 imprinting control region, Olfm4 and Hes1 (23),Vdr, Hic1, Sfrp5,Mgmt, Dusp6, c-Fos,Msln, and S100a4. Primer setsand regions subsequently sequenced can be found in Supplemen-tary Table S1. Sequencing libraries were prepared and analyzedusing the BiSPCR2 strategy (LINE1,H19) or the standard protocol

(Vdr, Hic1, Sfrp5, Mgmt, Dusp6, c-Fos, Msln, S100a4), describedpreviously (23, 24).

Tumor evaluationMacroscopic tumors were counted immediately after tissue

isolation using a stereoscope. Adenoma size was calculated bymeasuring the width of neoplastic lesions on hematoxylin andeosin (H&E)-stained small intestine sections using ImageJ. His-tologic pathology was performed by the Comparative PathologyCore at the University of Pennsylvania School of VeterinaryMedicine (Pennsylvania, PA) following the guidelines asdescribed previously (25). LOH was determined as describedpreviously (26).

Phosphorylated H3 cell-cycle quantificationApcMin/þ;Dnmt1loxP/loxP;Villin-Cre-ERT2 mutants and sibling

ApcMin/þ;Dnmt1loxP/loxP controls were injected with tamoxifen at1 month of age. Small intestines were collected 1 week, 1 month,and 2 months following tamoxifen administration, in separatecohorts of mice. Following fixation, tissue sections were co-stained for phosphorylated histone H3 (PH3) to identify cellsin M phase and b-catenin to identify regions of hyperproliferativeneoplastic epithelium. We counted the number of PH3þ nucleiand divided it by the total number of nuclei in the respectiveregion. This calculation was also performed for neoplastic intes-tinal epithelium at both 1 and 2 months post-tamoxifen timepoints, for controls and mutants.

qRT-PCRIntestines isolated 1 week post-tamoxifen were gently scraped

to remove villi and treated with EDTA to isolate crypt cells.Macroscopic tumors and adjacent normal tissue, from intestinesisolated 2 months post-tamoxifen, were visualized and dissectedusing a stereoscope. RNA was extracted using the TRIzol RNAisolation protocol (Invitrogen), followed by RNA cleanup usingthe RNeasy Mini Kit (Qiagen). mRNA expression was measuredusing quantitative RT-PCR, as described previously (27). TheSYBR green qPCR master mix (Agilent) was used in all qPCRreactions, and the fold change was calculated relative to thegeometric mean of Tbp and b-actin, using the DCt method. Themethod of normalizing to the geometricmean of a set of referencegenes has been described previously (28). Primer sequences canbe found in Supplementary Table S1.

Results and DiscussionDeletion ofDnmt1 in the intestinal epitheliumof adultApcMin/þ

mice induces acute global hypomethylationTo determine the role of Dnmt1 in intestinal adenoma initia-

tion, we employed ApcMin/þ;Dnmt1loxP/loxP;Villin-Cre-ERT2 miceto inducibly delete Dnmt1 throughout the intestinal epithelia.ApcMin/þ;Dnmt1loxP/loxP;Villin-Cre-ERT2 mice and their ApcMin/þ;Dnmt1loxP/loxP siblings (referred to as "mutants" and "controls,"respectively) were injected with tamoxifen at 4 weeks of age toinduce Dnmt1 deletion (Supplementary Fig. S1). On the basis ofour previous analysis of Dnmt1 ablation in the intestinal epithe-lium, we isolated the small intestine 1 week after tamoxifentreatment (23). To address the role of Dnmt1 in adenomadevelopment, we analyzed a second cohort of mice in which weharvested the intestine and colon 2 months post-tamoxifentreatment, at 3 months of age. By 3 months of age, ApcMin/þ mice

Dnmt1 in Genomic Stability during Intestinal Tumorigenesis

www.aacrjournals.org Cancer Prev Res; 9(7) July 2016 535




display macroscopic lesions (11, 12). Thus, 2 months post-tamoxifen allowed us to contrast adenoma development inDnmt1 mutants with ApcMin/þ;Dnmt1loxP/loxP controls. At 1 weekand 2 months following tamoxifen treatment, Dnmt1 deletionwas effective throughout the mutant epithelia of small intestine(Fig. 1A–D and G). Dnmt1 deletion was also maintained inmutant adenomas (Fig. 1E, F, and H). As ApcMin/þ mice exhibit

tumors predominantly in the small intestine (12, 29), we focusedour studies on adenoma development in the small intestinalepithelium. First, we analyzed the effects of Dnmt1 deletion inthe non-neoplastic epithelium of ApcMin/þ mice.

To estimate global DNA methylation levels, we performedtargeted bisulfite sequencing of theH19 imprinting control region(ICR) and the Long Interspersed Nucleotide Element 1 (LINE1).The process of imprinting is dependent on DNAmethylation andrefers to themethylation of alleles in a parental-specificmanner torestrict gene expression to one allele (30). The H19 imprintingcontrol region is methylated on the paternally inherited allele,thereby ensuring maternal allele expression (31). LINE1 retro-transposon loci comprise approximately 20% of the mouse andhuman genome, are maintained in a highly methylated state toinhibit transposase activity, and are a representative of genome-wide methylation levels (32–34). We employed laser capturemicrodissection (LCM) to collect crypts frommutant and controlintestines 1 week following tamoxifen treatment. H19 methyla-tion was maintained despite loss of Dnmt1 (Fig. 2A), implicatinga role forDnmt3a andDnmt3b inmaintaining genomic imprints,as suggested previously (35, 36). Strikingly, within 1 week fol-lowing Dnmt1 deletion, we observed a 35% reduction in DNAmethylation at the LINE1 loci inmutant crypts comparedwith thecontrol crypt epithelium (Fig. 2B). Thus, Dnmt1 is requiredacutely to maintain DNA methylation levels on LINE1 elementsbut is not required for maintenance of imprinted loci in therapidly proliferating intestinal epithelium.

To determine whether global hypomethylation of repetitiveelements was sustained over time, we isolated crypt epitheliumfrom mutants and controls 1 and 2 months following tamoxifentreatment. Surprisingly, we found no differences in methylationlevels in all conditions tested, demonstrating that DNA methyl-ation had been fully restored after temporary loss followingDnmt1 ablation (Fig. 2C–F). These results were unexpected, asDnmt1 deletion was maintained at 2 months after tamoxifenadministration (Fig. 1C–F). Recovery of DNA methylation waslikely driven by the de novoDNAmethyltransferases Dnmt3a and3b, which are also expressed in the intestinal epithelium (23).

Dnmt1 mutants display increased tumor initiationOne and 2 months following tamoxifen injection, the small

intestine was examined for neoplastic transformation. Surpris-ingly, H&E staining revealed a dramatic increase in the number ofneoplastic lesions throughout the mutant small intestine at 2months, but not 1 month, post-tamoxifen (Fig. 3A–D). Indeed,mutants displayed more than 6-fold more macroscopic adeno-mas than controls at 2months following tamoxifen injection (Fig.3E; n ¼ 17–20 per group). In addition, mutant small intestinaltumors were on average twice as large as those of controls at 2months post-tamoxifen injection (Fig. 3F; n ¼ 5–8 per group).Furthermore, we observed 58% incidence of neoplastic transfor-mation in colons ofmutant mice (n¼ 12) compared with 14% incontrols (n ¼ 7; Supplementary Fig. S2).

Previously, it had been reported that partial loss of Dnmt1produces a block in the progression of adenomas in the ApcMin/

þmouse paradigm (14–18), which is in sharp contrast to ourfindings of an increased number of macroscopic lesions in theApcMin/þ;Dnmt1loxP/loxP;VillinCreERT2 deletion model. To con-firm that these lesions are bona fide adenomas, we consultedwith a pathologist to perform histopathologic assessment on theintestines of mutant and control mice, 1 and 2 months following

Figure 1.

Conditional ablation of Dnmt1 in the ApcMin/þ intestinal epithelium. A–F, Dnmt1deletion is maintained 1 week (B) and 2 months (D) after tamoxifen treatmentin the adult mouse small intestinal epithelium. Immunohistochemicalstaining of Dnmt1 protein is evident in the crypt epithelia located just above thesubmucosa (outlined in black) of control mice at both time points (A, C) but isabsent in mutant mice (B, D). Neoplastic epithelia (outlined in red) display highlevels of Dnmt1 protein in control animals (E). Deletion of Dnmt1 protein ismaintained in the neoplastic epithelium (outlined in red) in mutant animals2 months after tamoxifen treatment (F). G and H, qRT-PCR analysiscomparing the relative gene expression levels of Dnmt1 in controls andmutants 1week (G) and 2 months (H) after tamoxifen treatment (n¼ 3–5 per genotypes).Compared with controls, Dnmt1 mutants express significantly lower levels ofDnmt1, and deletion is maintained in macroscopic tumors isolated 2 monthspost-tamoxifen. Gene expression is calculated relative to the geometric mean ofTBP and b-actin. All scale bars are 50 mm. � , P < 0.05; ��, P < 0.01; Student t test.

Sheaffer et al.

Cancer Prev Res; 9(7) July 2016 Cancer Prevention Research536




tamoxifen treatment at 4 weeks of age. Using criteria described byBiovin and colleagues (25), we found 15-fold more lesions thathad progressed to adenomas in mutant mice compared withApcMin/þcontrols (Supplementary Fig. S3B) at 2 months post-tamoxifen injection. At 1 month post-tamoxifen, all lesionsobserved in both controls and Dnmt1 mutants displayed similarhistopathology and were characterized as gastrointestinal intrae-pithelial neoplasias (Supplementary Fig. S3A).

DNA hypomethylation is observed at specific genes in tumorsDNA methylation is important for maintaining appropriate

intestinal stem cell gene expression, and loss of Dnmt1 in theintestinal epithelium results in increased crypt cell proliferation(23). Previously, we observed that 1 week following deletion ofDnmt1, DNA methylation was decreased at enhancers associated

with genes critical for intestinal stem cell proliferation, Hes1 andOlfm4 (23). Thus, altered DNAmethylation could be responsiblefor the dramatic increase in lesions observed in our mutantanimals. We employed LCM to collect crypt epithelia and neo-plastic tumors from mutant and control intestines 2 monthsfollowing tamoxifen treatment. We employed targeted bisulfitesequencing of the H19 ICR and LINE1 loci to determine whetherglobalmethylation levels were changed in tumors compared withepithelia in mice 2 months following tamoxifen injection. Inter-estingly, DNA methylation of the H19 ICR and LINE1 loci wereunchanged in all conditions (Fig. 4A and B). Control mice dis-played increased methylation at Hes1 and Olfm4 in tumorscompared with adjacent epithelia (Fig. 4C). Mutant mice exhib-ited little change in methylation when comparing control andmutant epithelia (Fig. 4C). However, mutant tumors are

Figure 2.

DNA methylation dynamics followingDnmt1 deletion. Targeted bisulfitesequencing of the H19 imprinting controlregion and the LINE1 repetitive elementswas performed to estimate maintenancemethylation ofDnmt1-deficient intestinalcrypt cells. Crypt epitheliumwas isolatedby LCM from controls and mutants at 1week (n¼ 3–4per group), 1month (n¼ 2for controls, n ¼ 4 for mutants), or 2months (n ¼ 4–7 per group) followingtamoxifen treatment. (A, C, E) Targetedbisulfite sequencing of 6 CpGs in the H19imprinting control region (Chr7:149,766,621–149,766,690). At 1week (A),1 month (C), and 2 months (E) followingtamoxifen administration, H19methylation levels are comparablebetween Dnmt1mutants and controls. B,D, F, targeted bisulfite sequencing of 9CpGs in the LINE1 repetitive elements(NCBI Accession #D84391: 976–1,072).LINE1 methylation levels 1 week aftertamoxifen injection are significantlyreduced in crypts of mutant micecompared with controls (B). One month(D) and 2 months (F) after tamoxifen-induced ablation of Dnmt1, global DNAmethylation levels have been restored tobaseline. For all graphs, data arepresented as average� SEM. �� , P < 0.01by 2-tailed Student t test.






significantly hypomethylated at the Hes1 enhancer comparedwith control tumors (Fig. 4C). Furthermore, mutant tumors aresignificantly hypomethylated at both Hes1 and Olfm4 loci com-pared with adjacent non-neoplastic epithelia (4C, dashed lines).These data are intriguing because Hes1 expression is increased instem cell–like cells in colon cancer (37), and our results suggestthat DNA methylation may play a role in activating Hes1 expres-sion in tumors.

Aberrant DNA methylation drives intestinal cancer cell growthin multiple paradigms (14–18) and could be responsible for thedramatic increase in lesion growth observed in our mutant

animals. Hypermethylation of tumor suppressor genes allowsunrestrained growth and is a hallmark of the colorectal cancerCpG island methylator phenotype (CIMP; ref. 38). We testedpromoter methylation of four genes that have been previouslyshown to be hypermethylated in adenomas, Vdr, Hic1, Sfrp5, andMgmt (39–43). To our surprise, only Vdr and Mgmt showedincreased methylation in adenomas, relative to nontumor epi-thelia (Fig. 4D). Dnmt1 deficiency reduced methylation in mut-ant tumors compared with control tumors in an inconsistentmanner (Fig. 4D), suggesting that hypermethylation at these lociis not required for tumor growth. DNA hypomethylation has also

Figure 3.

Deletion of Dnmt1 in adult ApcMin/þ

mice accelerates adenoma initiation.A–D, H&E staining of control andmutant intestinal epithelium 1 month(A and B) and 2 months (C and D) aftertamoxifen injection.H&E staining showsdisruption of normal intestinal epithelialmorphology in the small intestine oftamoxifen-treated mutants comparedwith controlmice (B vs. A, D vs. C). Scalebars are 50 mm. C, intestinal epithelial-specific Dnmt1 deletion causesincreased incidence of macroscopictumors. Total number of macroscopictumors was counted throughout theentire small intestine of mutant andcontrol mice, 1 month (n ¼ 5–7 pergroup) and 2 months (n ¼ 17–20 pergroup) following tamoxifen treatment.E, Dnmt1 deletion results in increasedsize of neoplastic lesions in the smallintestine 2 months after tamoxifentreatment. Neoplastic lesions weremeasured in control andmutantmice at1 month (n ¼ 2 for controls, n ¼ 3 formutants) and 2 months (n ¼ 5 forcontrols, n ¼ 8 for mutants) followingtamoxifen injection. For all graphs,data are presented as average � SEM.�� , P < 0.001 by one-way ANOVA.

Sheaffer et al.





Figure 4.

Dnmt1-mutant tumors show increased hypomethylation at potential oncogenes. Targeted bisulfite sequencingwas performed to estimatemaintenancemethylationof Dnmt1-deficient tumors, relative to controls tumors and to adjacent non-neoplastic crypt epithelium. Tumors and crypt epithelium were isolated byLCM from controls and mutants 2 months following tamoxifen treatment. A and B, analysis of 9 CpGs in the LINE1 repetitive elements (NCBI Accession #D84391:976–1,072) revealed no alterations in methylation of the repetitive transposable elements (A). Analysis of 6 CpGs in the H19 imprinting control region (Chr7:149,766,621–149,766,690) showed that H19 methylation levels are comparable between Dnmt1 mutants and controls, in both tumor and nontumor epithelium.C, potential enhancer regions of genes involved in intestinal stem cell identity and proliferation, Olfm4 and Hes1 were analyzed for methylation changes in allconditions. Regional averages of 7 CpGs in theOlfm4 enhancer (chr14: 80,399,983–80,400,271) and 6 CpGs in theHes1 enhancer (chr16: 30,055,572–30,055,738) areshown. The Hes1 enhancer shows increased methylation in control tumors compared with adjacent tissue, which is abolished by Dnmt1 deletion. D, geneswith cancer-specific promoter hypermethylation, Vdr, Hic1, Sfrp5, Mgmtwere analyzed for methylation changes in all conditions. Regional averages of 8 CpGs in theVdr promoter (chr15: 97,731,208–97,731,590), 16 CpGs in Hic1 promoter (chr11: 74,983,743–74,983,918), 16 CpGs in the Sfrp5 promoter (chr19: 42,276,328–42,276,453), and 19 CpGs in Mgmt promoter (chr7: 144,086,076–144,086,345) are shown. Only Vdr and Mgmt displayed significant hypermethylation in controltumors compared with adjacent tissue. Loss of Dnmt1 caused significant loss of hypermethylation at both loci, suggesting that hypermethylation of thesegenes is not required for tumor progression. E, genes with cancer-specific promoter hypomethylation, Dusp6, c-Fos, Msln, S100a4, were analyzed for methylationchanges in all conditions. Regional averages of 8 CpGs in the Dusp6 promoter (chr10: 98,728,738–98,729,028), 14 CpGs in the c-Fos promoter (chr12:86,817,243–86,817,533), 6 CpGs in theMsln promoter (chr17: 25,891,502–25,891,643), and 3 CpGs in the S100a4 promoter (chr3: 90,407,319–90,407,489) are shown.Only Dusp6 and c-Fos displayed significant hypomethylation in control tumors compared with adjacent tissue. Dnmt1 deletion caused even more drasticloss of methylation at both promoters, suggesting that loss of Dnmt1 may be driving activation of potential oncogenes during tumor growth.






been reported at several potential oncogenes including Dusp6, c-Fos, Msln, and S100a4 (43–46). Although none of these genepromoters show significant loss of methylation in tumors com-pared with the respective normal epithelia, we observed signifi-cantly decreasedmethylation atDusp6 and c-Fos inDnmt1mutantrelative to control adenomas (Fig. 4E). c-Fos is a potent driver ofcolon cancer cell growth (47), and its expression is regulated byDNAhypomethylation in several cancer types (48, 49). These datasuggest that hypomethylation at oncogenesmay be driving tumorgrowth in Dnmt1-deficient ApcMin/þ mutant animals.

Dnmt1-null tumors are initiated by LOH and Wnt activationWe further investigated tumor initiation to determine whether

the observed changes in DNAmethylation were a causative factorof neoplastic transformation. Adenoma formation in theApcMin/þ

mouse model is driven by LOH at the Apc locus, which causesnuclear accumulationofb-catenin andactivationofWnt signaling(13). LOH is observed in virtually all ApcMin/þ intestinal tumors(26, 50). To characterize adenoma initiation by LOH in detail, weemployed LCM to isolate neoplastic and normal intestinal epi-thelial cells frommutantmice and their control siblings, 2monthsfollowing tamoxifen treatment. Since the number of macroscopiclesions found in controls at 1-month post-tamoxifen is very small(Fig. 3E, Supplementary Fig. S3), we focused on the 2-month timepoint for LOH analysis. DNA was isolated from these samples(�1,000 cells per sample), and the status of the Apc alleles wasexamined as described previously (26). Mutant tumors displayedLOH at a rate similar to that of control tumors (Fig. 5A), dem-onstrating that Dnmt1-mutant ApcMin/þ mice initiate tumors bythe same mechanism.

Several lines of evidence confirm that the numerous adenomasthat develop in theDnmt1-deficientApcMin/þmice are the result ofWnt pathway activation. Mutant adenomas at both 1 and 2months following tamoxifen treatment exhibited increasednucle-ar b-catenin protein (Fig. 5B–D) relative to 2-month controls.Multiple Wnt signaling targets, such as CyclinD1 (Fig. 5E–G) andSox9 (Fig. 5H–J), were also strongly activated. These data dem-onstrate that total loss of Dnmt1 in the mature small intestinalepithelium promotes neoplastic progression, most likely througha mechanism that enhances LOH at the ApcMin/þ locus.

Dnmt1-mutantmice display acute increases in proliferation andapoptosis

Because of the dramatic difference in lesion size and hypo-methylation at potential oncogenes observed in the inducibleDnmt1-mutant intestine, we investigated whether Dnmt1 defi-ciency had a direct impact on cell proliferation in ApcMin/þ

adenomas or the non-neoplastic crypt epithelium. We deter-mined percentages of control and mutant epithelial cells in Mphase by staining for pH3, at 1week, 1month, and2months post-tamoxifen treatment. One week post-tamoxifen (Fig. 6A–C), thepercentage of pH3þ nuclei was significantly increased in theDnmt1-deficient epithelium relative to control (n¼ 4 per group).Interestingly, at 1 month (Fig. 6D–F) and 2 months (Fig. 6G–I)following tamoxifen injection, the proliferation rate was equiv-alent in Dnmt1-positive or -negative tumors and in the adjacentcrypt epithelium.

We also examined the small intestine at all three time points forsigns of cell death using terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL), which detects DNAfragmentation in apoptotic cells. We quantified the percent of

epithelial cells that were TUNELþ, similar to the pH3 stainingquantification. Mutants 1 week after tamoxifen treatment dis-played a significantly higher proportion of TUNELþ cells com-pared with controls (Fig. 7A–C). Both control and mutant ade-nomas exhibited only rare epithelial cells positive for TUNEL, andwe found no differences in cell death betweenmutant and controlneoplastic and non-neoplastic epithelia, at 1 month (Fig. 7D–F)or 2 months (Fig. 7G–I) following tamoxifen administration.However, many subepithelial cells were TUNEL-positive (Fig. 7D,E, G, and H). These data demonstrate that loss of Dnmt1 in theadult intestinal epithelium in vivo has acute temporary effects onintestinal proliferation and apoptosis but no effect on ApcMin/þ

adenoma proliferation or cell death. Overall, our results suggestthat the increase in adenoma number and size found in theDnmt1-mutant mice is driven by accelerated tumor initiationthrough LOH.

Dnmt1-deficient ApcMin/þ intestinal epithelia exhibit increasedgenomic instability

Next, we considered the molecular mechanisms behind theaccelerated LOH we observed in the Dnmt1-deficient intestinalepithelium. We hypothesized that the established role of Dnmt1in the preservation of genomic stability might contribute to thephenotype. DNAmethylation has been linked tomismatch repair(MMR)deficiency and genomic instability inmultiple contexts, inboth cell lines and disease. In the HCT116 colorectal cancer cellline, ablation of catalytically activeDNMT1 causes cell-cycle arrestand apoptosis due to increased chromosomal instability (51, 52).In mouse embryonic stem cells, loss of Dnmt1 also causes globalhypomethylation and increased mutation rates (53).

To determine levels of DNA damage in mutant and controlmice, we performed gH2AX staining, which visualizes DNAdouble-strand breaks as a marker of chromosomal instability.One week following tamoxifen administration, we discovered adramatic increase in the fraction of gH2AXþ epithelial cells inDnmt1-mutant versus controls, which contained no gH2AXþ

epithelial cells (Fig. 8A–C). Onemonth post-tamoxifen injection,mutants and controls displayed a similar number of gH2AXþ

nuclei (Fig. 8D–F). There was a significant difference betweenneoplastic and non-neoplastic epithelium in control mice, and asimilar statistically relevant change was observed inmutants (Fig.8F). Two months following tamoxifen treatment, nontumorepithelia in both the control and mutant mice showed very littlegH2AX (Fig. 8G–I). Although we found many more epithelialcells within Dnmt1-deficient tumors to be gH2AXþ comparedwith control tumors (Fig. 8I), these data were not statisticallysignificant. These results demonstrate that increased genomicinstability occurs as a result of Dnmt1 deletion in the ApcMin/þ

model and that genomic instability occurs prior to tumor devel-opment. We propose that this temporary decrease in genomicstability, combined with elevated proliferation and apoptosislevels, contributes to accelerated LOH and enhanced tumorigen-esis in the inducible ApcMin/þ;Dnmt1loxP/loxP;Villin-Cre-ERT2mice.

DiscussionThe results presented above have major implications for the

fields ofDNAmethylation and intestinal cancer biology.We showthat loss of the DNA methyltransferase Dnmt1 in the ApcMin/þ

cancer model results in acute hypomethylation and DNA damage

Sheaffer et al.





Figure 5.

Dnmt1-deficient ApcMin/þ intestinal tumors display increased Wnt signaling. A, control and mutant mice were tamoxifen-treated at 4 weeks of age, and intestineswere harvested 2 months later. Intestinal tumors from control and mutant mice were isolated by LCM for PCR testing of LOH at the Apc locus. Results ofPCR testing demonstrate that mutant tumors display LOH at a rate similar to that of controls. B–J, control and mutant mice were tamoxifen-treated at 4 weeksof age, and intestines were harvested 1 or 2 months later for immunostaining analysis of Wnt signaling targets. B–D, nuclear b-catenin, as visualized byimmunohistochemical staining, is increased in Dnmt1-mutant intestine 2 months after tamoxifen treatment (D), compared with tamoxifen-treated, age-matchedcontrol small intestine (B). One month following Dnmt1 ablation, mutant small intestine displays nuclear b-catenin levels similar to that observed inthe 2-month controls (C compared with B, respectively). E–G, expression of the Wnt signaling target cyclin D1 (CycD1) is increased in Dnmt1-deficient tumors (G)relative to age-matched control tumors (E). At 1-month post-Dnmt1 ablation, CycD1 staining appears slightly increased compared with the 2-month controls(F vs. E, respectively). CycD1 (red), DAPI (blue). H–J, Sox9 immunofluorescent staining shows increased Sox9 levels in Dnmt1-deficient tumors (J) compared withcontrol tumors (H) 2 months following tamoxifen treatment. At 1 month after Dnmt1 ablation, increased Sox9 expression is visible by immunostaining(I compared with control, H). Sox9 (red), DAPI (blue). All scale bars are 50 mm.






but does not affect tumor cell proliferation or apoptosis. We positthat decreased genomic stability accelerates LOH at the Apc locus,resulting in increased tumor initiation and larger tumors over

time. Our work adds to the body of evidence that implicates acrucial role for Dnmt1 and DNA methylation in maintaininggenome stability.

Figure 6.

Loss of Dnmt1 in the ApcMin/þ intestine causes an acute increase in crypt cell proliferation. Percentage of cells in mitosis was determined by immunostaining andcounting of pH3þ nuclei in the small intestine 1 week, 1 month, and 2 months following tamoxifen treatment. The percentage of pH3þ cells was calculatedas the number of pH3þnuclei divided by the total number of epithelial nuclei. A–C, oneweek following tamoxifen injection,mutants displayeda significantly increasedpercentage of pH3þ cells (C, n ¼ 4 per genotype). Representative images of pH3 immunostaining are shown for controls (A) and Dnmt1 mutants (B). D–F,onemonth after tamoxifen treatment, mutants (E) and controls (D) displayed comparable percentages of pH3þ cells, in both tumors and noncancerous epithelia (F).Twenty-four tumors across 5 mice were counted for Dnmt1 mutants; 4 tumors across 4 mice were counted for controls. Approximately 1,500 noncancerouscells were counted from the same biologic replicates to calculate pH3 frequency in nontumor tissue. G–I, two months after tamoxifen treatment, mutants (H)and controls (G) displayed comparable percentages of pH3þ cells, in both tumors and noncancerous epithelia (I). Twenty-four tumors across 7 mice werecounted for Dnmt1 mutants; 8 tumors across 3 mice were counted for controls. Nontumor epithelia quantified as in (D-F). E-cadherin (green) outlines theintestinal epithelium. All scale bars are 50 mm. For all graphs, data are presented as average � SEM. �� , P < 0.01 by Student t test (C). One-way ANOVA wasperformed for (F, I), and the results were not significant.

Sheaffer et al.





Importantly, our study demonstrates increased intestinal ade-noma formation after deletion of Dnmt1. Our experimentalparadigm results in intestine-specific, temporal loss of Dnmt1and acute hypomethylation, followed by recovery of global

genome DNA methylation, in contrast to prior work, which hademployed germline hypomorphic Dnmt1 alleles (14–18). Theseprevious studies thus differed both in the timing of Dnmt1deletion and in the lack of tissue specificity and described a

Figure 7.

Loss of Dnmt1 results in a short-term increase in intestinal epithelial apoptosis but is not altered 1 or 2 months following Dnmt1 deletion. TUNEL staining (red) wasperformed to detect apoptotic cells, with E-cadherin (green) to outline the epithelium. The percentage of apoptotic cells was determined by counting ofthe TUNEL-positive (TUNELþ) nuclei in the small intestine 1 week, 1 month, and 2 months following tamoxifen treatment. The percentage of TUNELþ cells wascalculated as the number of TUNELþ nuclei divided by the total number of epithelial nuclei. A–C, one week following tamoxifen treatment, Dnmt1 mutants havesignificantly higher rates of apoptosis than controls (B vs. A; quantitation in C). N ¼ 4 per group. �� , P < 0.001 by Student t-test. D–F, one month aftertamoxifen injection, both tumor and nontumor epithelia contain TUNEL-positive cells, which is unchanged in mutant mice (D vs. E, quantitation in F). n¼ 3 biologicreplicates per group, with 8 tumors counted from controls and 14 tumors counted from mutants. G–I, two months after tamoxifen injection, both tumor andnontumor epithelia contain TUNEL-positive cells at low frequency in control mice, which is unchanged in mutant mice (H vs. G, quantitation in I). n¼ 3–5 biologicalreplicates per group; two tumors were counted from controls, and 10 tumors were counted from mutants. All scale bars are 50 mm. For all graphs, data arepresented as average � SEM. One-way ANOVA test was performed in (F, I), and results were not significant.






constitutive hypomethylation phenotype which may be a con-tributing factor to our disparate results. Our Dnmt1-deficienttumors 2 months after tamoxifen administration exhibit signif-icant hypomethylation of several potential oncogenes, yet displayminimal DNA methylation changes genome-wide. We cannot

rule out that the genome remethylation occurring from 1 week to1month post-tamoxifen contributes to accelerated tumorigenesisin our conditionalDnmt1 ablation intestinal cancer model. How-ever, the significantly enhanced genomic instability, proliferation,and apoptosis observed immediately following Dnmt1 ablation,

Figure 8.

The Dnmt1-deficient ApcMin/þ intestine displays increased chromosomal instability. Mice were tamoxifen-treated at 4 weeks of age, and intestines were harvested1 week, 1 month, or 2 months later for immunofluorescent staining analysis. gH2AX (red) was used to visualize DNA double-strand breaks as a markerof genome instability, and E-cadherin (green) was used to outline the epithelium. The percentage of gH2AX-positive (gH2AXþ) cells was calculated as the number ofgH2AXþ nuclei divided by the total number of epithelial nuclei. A–C, One week following tamoxifen treatment, the number of gH2AXþ nuclei is increasedinmutant comparedwith control small intestines (B comparedwith A, quantification in C). n¼ 4 biological replicates per group. �� , P < 0.001 by Student t test. D–F,one month post-tamoxifen injection, both mutant and control non-neoplastic epithelia display low levels of DNA damage, which is significantly increased intumors of both groups (F). n¼ 3 biological replicates per group; 2 tumors were counted for controls, 8 tumors were counted for mutants. G–I, twomonths followingtamoxifen treatment, increased levels of gH2AXþ foci continue to be seen in mutant tumors compared with mutant non-neoplastic epithelium (I). There is nosignificant difference between control and mutant gH2AXþ levels (G vs. H, quantification in I). n¼ 3–4 biological replicates per group; 3 tumors were counted fromcontrols, and 12 tumors were counted from mutants. All scale bars are 50 mm. For F and I, � , P < 0.05; �� , P < 0.01; �� , P < 0.001, one-way ANOVA.


Sheaffer et al.




combined with the increased adenoma formation at 1 and 2months post-tamoxifen, strongly suggest a role for Dnmt1 inmaintaining genomic stability and preventing intestinal cancerinitiation.

Interestingly, our model results in acute global hypomethyla-tion at LINE1 repetitive elements. DNAmethylation is associatedwith the silencing of LINE1 transcription (54), and LINE1 hypo-methylation is highly variable in colorectal carcinomas (55). It hasbeen hypothesized that LINE1 retrotransposition events caninterrupt critical genes, such as APC, and drive tumorigenesis(54). In particular contexts, such as APC heterozygosity, changesin genome stability are critical drivers of cancer initiation. Ourdata support the general theory that DNA hypomethylationcauses increased DNA damage, and LINE1 hypomethylation mayalso be contributing to our observed phenotype.

The DNA demethylating agents azacytidine and decitabinehave been tested as anticancer therapeutics in colorectal cancerto abrogate DNA hypermethylation and the silencing of tumorsuppressor genes (56, 57). Decitabine and azacytidine aremost commonly used to treat acute myeloid leukemia (AML)and myelodysplastic syndromes, with variable success rates(58–60). Our results indicate that DNA demethylation actuallycontributes to cancer formation, and warrant caution in treat-ing patients with gastrointestinal cancer with a potentialtumor-enhancing drug.

In conclusion, we show that deletion of Dnmt1 in the adultintestinal epithelium of ApcMin/þ mice causes accelerated forma-tion of adenomas. Loss of Dnmt1 results in acute hypomethyla-tion and genomic instability, accompanied by increased prolif-eration and apoptosis. Although Dnmt1-deficient adenomaseventually recover global DNA methylation, they continue todisplay hypomethylation at several potential oncogenes,highlighting an unappreciated role for DNA hypomethylation inintestinal tumor development. These results support a fundamen-

tal role for DNA methylation in preserving genomic integrityduring intestinal tumor formation.

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

Authors' ContributionsConception and design: K.L. Sheaffer, K.H. KaestnerDevelopment of methodology: K.L. SheafferAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): K.L. Sheaffer, E.N. ElliottAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): K.L. Sheaffer, E.N. ElliottWriting, review, and/or revision of the manuscript: K.L. Sheaffer, E.N. Elliott,K.H. KaestnerAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): K.L. SheafferStudy supervision: K.L. Sheaffer, K.H. Kaestner

AcknowledgmentsThe authors thankHaleigh Zillges, JoeGrubb, Jonathan Schug, and Elizabeth

Buza for technical assistance. A special thank you to services provided by theUniversity of Pennsylvania Functional Genomics Core (P30-DK19525). Theyalso thank the University of Pennsylvania Department of Dermatology, espe-cially Dr. John Seykora and Dr. Stephen Prouty, for Leica LCM training andequipment access. They acknowledge the support of Dr. AdamBedenbaugh andthe Morphology Core of the Penn Center for the Study of Digestive and LiverDiseases (P30-DK050306).

Grant SupportThis work was supported by NIH grant R37-DK053839 (to K.H. Kaestner).The costs of publication of this articlewere defrayed inpart by the payment of

page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received September 21, 2015; revised January 26, 2016; accepted February 8,2016; published OnlineFirst February 16, 2016.

References1. Fearon ER. Molecular genetics of colorectal cancer. Annu Rev Pathol

2011;6:479–507.2. Smith ZD, Meissner A. DNA methylation: roles in mammalian develop-

ment. Nat Rev Genet 2013;14:204–20.3. Tahiliani M, Koh KP, Shen Y, Pastor WA, Bandukwala H, Brudno Y, et al.

Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mamma-lian DNA by MLL partner TET1. Science 2009;324:930–5.

4. Cortellino S, Xu J, Sannai M, Moore R, Caretti E, Cigliano A, et al. ThymineDNA glycosylase is essential for active DNA demethylation by linkeddeamination-base excision repair. Cell 2011;146:67–79.

5. Rhee I, Bachman KE, Park BH, Jair KW, Yen RW, Schuebel KE, et al. DNMT1and DNMT3b cooperate to silence genes in human cancer cells. Nature2002;416:552–6.

6. Berman BP, Weisenberger DJ, Aman JF, Hinoue T, Ramjan Z, Liu Y, et al.Regions of focal DNA hypermethylation and long-range hypomethylationin colorectal cancer coincide with nuclear lamina-associated domains. NatGenet 2012;44:40–6.

7. Luo Y, Wong CJ, Kaz AM, Dzieciatkowski S, Carter KT, Morris SM, et al.Differences in DNA methylation signatures reveal multiple pathways ofprogression from adenoma to colorectal cancer. Gastroenterology2014;147:418–29.

8. Goelz SE, Vogelstein B, Hamilton SR, Feinberg AP. Hypomethylation ofDNA from benign and malignant human colon neoplasms. Science1985;228:187–90.

9. Feinberg AP, Gehrke CW, Kuo KC, Ehrlich M. Reduced genomic 5-methyl-cytosine content in human colonic neoplasia. Cancer Res 1988;48:1159–61.

10. Su LK, Kinzler KW, Vogelstein B, Preisinger AC, Moser AR, Luongo C, et al.Multiple intestinal neoplasia caused by amutation in themurine homologof the APC gene. Science 1992;256:668–70.

11. Cheung AF, Carter AM, Kostova KK, Woodruff JF, Crowley D, Bronson RT,et al. Complete deletion of Apc results in severe polyposis in mice.Oncogene 2010;29:1857–64.

12. Moser AR, Pitot HC, Dove WF. A dominant mutation that predisposesto multiple intestinal neoplasia in the mouse. Science 1990;247:322–4.

13. Sansom OJ, Reed KR, Hayes AJ, Ireland H, Brinkmann H, Newton IP, et al.Loss of Apc in vivo immediately perturbsWnt signaling, differentiation, andmigration. Genes Dev 2004;18:1385–90.

14. Yamada Y, Jackson-Grusby L, Linhart H, Meissner A, Eden A, Lin H, et al.Opposing effects of DNA hypomethylation on intestinal and liver carci-nogenesis. Proc Natl Acad Sci U S A 2005;102:13580–5.

15. Laird PW, Jackson-Grusby L, Fazeli A, Dickinson SL, Jung WE, Li E, et al.Suppression of intestinal neoplasia by DNA hypomethylation. Cell 1995;81:197–205.

16. Eads CA, Nickel AE, Laird PW. Complete genetic suppression of polypformation and reduction of CpG-island hypermethylation in Apc(Min/þ)Dnmt1-hypomorphic Mice. Cancer Res 2002;62:1296–9.

17. Cormier RT, Dove WF. Dnmt1N/þ reduces the net growth rate andmultiplicity of intestinal adenomas in C57BL/6-multiple intestinal neo-plasia (Min)/þ mice independently of p53 but demonstrates strongsynergy with the modifier of Min 1(AKR) resistance allele. Cancer Res2000;60:3965–70.






18. Trasler J,DengL,Melnyk S, Pogribny I,Hiou-TimF, Sibani S, et al. Impact ofDnmt1 deficiency, with and without low folate diets, on tumor numbersand DNA methylation in Min mice. Carcinogenesis 2003;24:39–45.

19. Perreault N, Sackett SD, Katz JP, Furth EE, Kaestner KH. Foxl1 is amesenchymal Modifier of Min in carcinogenesis of stomach and colon.Genes Dev 2005;19:311–5.

20. Jackson-Grusby L, Beard C, Possemato R, Tudor M, Fambrough D, Csan-kovszki G, et al. Loss of genomic methylation causes p53-dependentapoptosis and epigenetic deregulation. Nat Genet 2001;27:31–9.

21. el Marjou F, Janssen KP, Chang BH, Li M, Hindie V, Chan L, et al. Tissue-specific and inducible Cre-mediated recombination in the gut epithelium.Genesis 2004;39:186–93.

22. LaneN,DeanW, Erhardt S,Hajkova P, Surani A,Walter J, et al. Resistance ofIAPs to methylation reprogramming may provide a mechanism for epi-genetic inheritance in the mouse. Genesis 2003;35:88–93.

23. Sheaffer KL, Kim R, Aoki R, Elliott EN, Schug J, Burger L, et al. DNAmethylation is required for the control of stem cell differentiation in thesmall intestine. Genes Dev 2014;28:652–64.

24. Bernstein DL, Kameswaran V, Le Lay JE, Sheaffer KL, Kaestner KH. TheBisPCR(2) method for targeted bisulfite sequencing. Epigenetics Chroma-tin 2015;8:27.

25. Boivin GP, Washington K, Yang K, Ward JM, Pretlow TP, Russell R, et al.Pathology of mouse models of intestinal cancer: consensus report andrecommendations. Gastroenterology 2003;124:762–77.

26. Luongo C, Moser AR, Gledhill S, Dove WF. Loss of Apcþ in intestinaladenomas from Min mice. Cancer Res 1994;54:5947–52.

27. Gupta RK, Gao N, Gorski RK, White P, Hardy OT, Rafiq K, et al. Expansionof adult beta-cell mass in response to increased metabolic demand isdependent on HNF-4alpha. Genes Dev 2007;21:756–69.

28. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A,et al. Accurate normalization of real-time quantitative RT-PCR data bygeometric averaging of multiple internal control genes. Genome Biol2002;3:RESEARCH0034.

29. Moser AR, Dove WF, Roth KA, Gordon JI. The Min (multiple intestinalneoplasia) mutation: its effect on gut epithelial cell differentiation andinteraction with a modifier system. J Cell Biol 1992;116:1517–26.

30. Plasschaert RN, Bartolomei MS. Genomic imprinting in development,growth, behavior and stem cells. Development 2014;141:1805–13.

31. Tremblay KD, Saam JR, Ingram RS, Tilghman SM, Bartolomei MS. Apaternal-specific methylation imprint marks the alleles of the mouseH19 gene. Nat Genet 1995;9:407–13.

32. Yang AS, Est�ecio MR, Doshi K, Kondo Y, Tajara EH, Issa JP. A simplemethod for estimating global DNA methylation using bisulfite PCR ofrepetitive DNA elements. Nucleic Acids Res 2004;32:e38.

33. Kuramochi-Miyagawa S, Watanabe T, Gotoh K, Totoki Y, Toyoda A, IkawaM, et al. DNA methylation of retrotransposon genes is regulated by Piwifamily members MILI and MIWI2 in murine fetal testes. Genes Dev2008;22:908–17.

34. Gregory TR. Synergy between sequence and size in large-scale genomics.Nat Rev Genet 2005;6:699–708.

35. Riggs AD, Xiong Z. Methylation and epigenetic fidelity. Proc Natl Acad SciU S A 2004;101:4–5.

36. Chen T,Ueda Y,Dodge JE,Wang Z, Li E. Establishment andmaintenance ofgenomic methylation patterns in mouse embryonic stem cells by Dnmt3aand Dnmt3b. Mol Cell Biol 2003;23:5594–605.

37. Gao F, Zhang Y,Wang S, Liu Y, Zheng L, Yang J, et al. Hes1 is involved in theself-renewal and tumourigenicity of stem-like cancer cells in colon cancer.Sci Rep 2014;4:3963.

38. Issa JP. CpG island methylator phenotype in cancer. Nat Rev Cancer2004;4:988–93.

39. Marik R, Fackler M, Gabrielson E, Zeiger MA, Sukumar S, Stearns V, et al.DNAmethylation-related vitamin D receptor insensitivity in breast cancer.Cancer Biol Ther 2010;10:44–53.

40. Abouzeid HE, Kassem AM, Abdel Wahab AH, El-mezayen HA, Sharad H,Abdel Rahman S. Promoter hypermethylation of RASSF1A, MGMT, and

HIC-1 genes in benign and malignant colorectal tumors. Tumour Biol2011;32:845–52.

41. Suzuki H, Watkins DN, Jair KW, Schuebel KE, Markowitz SD, Chen WD,et al. Epigenetic inactivation of SFRP genes allows constitutive WNTsignaling in colorectal cancer. Nat Genet 2004;36:417–22.

42. Qi J, Zhu YQ, Luo J, TaoWH.Hypermethylation and expression regulationof secreted frizzled-related protein genes in colorectal tumor. World JGastroenterol 2006;12:7113–7.

43. Grimm C, Chavez L, Vilardell M, Farrall AL, Tierling S, B€ohm JW, et al.DNA-methylome analysis of mouse intestinal adenoma identifies atumour-specific signature that is partly conserved in human colon cancer.PLoS Genet 2013;9:e1003250.

44. Rosty C, Ueki T, Argani P, Jansen M, Yeo CJ, Cameron JL, et al. Over-expression of S100A4 in pancreatic ductal adenocarcinomas is associatedwith poor differentiation and DNA hypomethylation. Am J Pathol2002;160:45–50.

45. Choi EK, Uyeno S, Nishida N, Okumoto T, Fujimura S, Aoki Y, et al.Alterations of c-fos gene methylation in the processes of aging andtumorigenesis in human liver. Mutat Res 1996;354:123–8.

46. Sato N,Maitra A, Fukushima N, van Heek NT,Matsubayashi H, Iacobuzio-Donahue CA, et al. Frequent hypomethylation of multiple genes over-expressed in pancreatic ductal adenocarcinoma. Cancer Res 2003;63:4158–66.

47. PandeyMK, LiuG,Cooper TK,Mulder KM.Knockdownof c-Fos suppressesthe growth of human colon carcinoma cells in athymic mice. Int J Cancer2012;130:213–22.

48. Shukla V, Coumoul X, Lahusen T, Wang RH, Xu X, Vassilopoulos A, et al.BRCA1 affects global DNA methylation through regulation of DNMT1.Cell Res 2010;20:1201–15.

49. Dizik M, Christman JK, Wainfan E. Alterations in expression and meth-ylation of specific genes in livers of rats fed a cancer promoting methyl-deficient diet. Carcinogenesis 1991;12:1307–12.

50. Levy DB, Smith KJ, Beazer-Barclay Y, Hamilton SR, Vogelstein B, KinzlerKW. Inactivation of both APC alleles in human andmouse tumors. CancerRes 1994;54:5953–8.

51. Chen T, Hevi S, Gay F, Tsujimoto N, He T, Zhang B, et al. Completeinactivation of DNMT1 leads tomitotic catastrophe in human cancer cells.Nat Genet 2007;39:391–6.

52. Spada F, Haemmer A, Kuch D, Rothbauer U, Schermelleh L, Kremmer E,et al. DNMT1 but not its interaction with the replication machinery isrequired for maintenance of DNA methylation in human cells. J Cell Biol2007;176:565–71.

53. Chen RZ, Pettersson U, Beard C, Jackson-Grusby L, Jaenisch R. DNAhypomethylation leads to elevated mutation rates. Nature 1998;395:89–93.

54. Miousse IR, Koturbash I. The fine LINE: methylation drawing the cancerlandscape. Biomed Res Int 2015;2015:131547.

55. Est�ecio MR, Gharibyan V, Shen L, Ibrahim AE, Doshi K, He R, et al. LINE-1hypomethylation in cancer is highly variable and inversely correlated withmicrosatellite instability. PLoS One 2007;2:e399.

56. Braiteh F, Soriano AO, Garcia-Manero G, Hong D, JohnsonMM, Silva LeP,et al. Phase I study of epigenetic modulation with 5-azacytidine andvalproic acid in patients with advanced cancers. Clin Cancer Res 2008;14:6296–301.

57. Goel A, Boland CR. Epigenetics of colorectal cancer. Gastroenterology2012;143:1442–60.e1.

58. Kadia TM, Jabbour E, Kantarjian H. Failure of hypomethylating agent-based therapy in myelodysplastic syndromes. Semin Oncol 2011;38:682–92.

59. Nebbioso A, Carafa V, Benedetti R, Altucci L. Trials with 'epigenetic' drugs:an update. Mol Oncol 2012;6:657–82.

60. Fenaux P,Mufti GJ, Hellstrom-Lindberg E, Santini V, Finelli C, GiagounidisA, et al. Efficacy of azacitidine compared with that of conventional careregimens in the treatment of higher-risk myelodysplastic syndromes: arandomised, open-label, phase III study. Lancet Oncol 2009;10:223–32.


Sheaffer et al.




2016;9:534-546. Published OnlineFirst February 16, 2016.Cancer Prev Res Karyn L. Sheaffer, Ellen N. Elliott and Klaus H. Kaestner Intestinal Cancer InitiationDNA Hypomethylation Contributes to Genomic Instability and

Updated version

10.1158/1940-6207.CAPR-15-0349doi:

Access the most recent version of this article at:

Material

Supplementary

1

http://cancerpreventionresearch.aacrjournals.org/content/suppl/2016/02/16/1940-6207.CAPR-15-0349.DCAccess the most recent supplemental material at:

Cited articles

http://cancerpreventionresearch.aacrjournals.org/content/9/7/534.full#ref-list-1

This article cites 60 articles, 22 of which you can access for free at:

Citing articles

http://cancerpreventionresearch.aacrjournals.org/content/9/7/534.full#related-urls

This article has been cited by 1 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://cancerpreventionresearch.aacrjournals.org/content/9/7/534To request permission to re-use all or part of this article, use this link



http://cancerpreventionresearch.aacrjournals.org/lookup/doi/10.1158/1940-6207.CAPR-15-0349

http://cancerpreventionresearch.aacrjournals.org/content/suppl/2016/02/16/1940-6207.CAPR-15-0349.DC1

http://cancerpreventionresearch.aacrjournals.org/content/suppl/2016/02/16/1940-6207.CAPR-15-0349.DC1

http://cancerpreventionresearch.aacrjournals.org/content/9/7/534.full#ref-list-1

http://cancerpreventionresearch.aacrjournals.org/content/9/7/534.full#related-urls

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

mailto:[email protected]

http://cancerpreventionresearch.aacrjournals.org/content/9/7/534


dna hypomethylation contributes to genomic instability and ... · by global dna hypomethylation,...

Documents