Full Text (PDF)

Published OnlineFirst May 1, 2014; DOI: 10.1158/0008-5472.CAN-13-3147

Cancer Tumor and Stem Cell Biology Research

Loss of the Polycomb Mark from Bivalent Promoters Leads to Activation of Cancer-Promoting Genes in Colorectal Tumors

Maria A. Hahn1, Arthur X. Li2, Xiwei Wu3, Richard Yang1, David A. Drew4, Daniel W. Rosenberg4, and Gerd P. Pfeifer1

Abstract In colon tumors, the transcription of many genes becomes deregulated by poorly defined epigenetic mechanisms that have been studied mainly in established cell lines. In this study, we used frozen human colon tissues to analyze patterns of histone modification and DNA cytosine methylation in cancer and matched normal mucosa specimens. DNA methylation is strongly targeted to bivalent H3K4me3- and H3K27me3-associated promoters, which lose both histone marks and acquire DNA methylation. However, we found that loss of the Polycomb mark H3K27me3 from bivalent promoters was accompanied often by activation of genes associated with cancer progression, including numerous stem cell regulators, oncogenes, and proliferation-associated genes. Indeed, we found many of these same genes were also activated in patients with ulcerative colitis where chronic inflammation predisposes them to colon cancer. Based on our findings, we propose that a loss of Polycomb repression at bivalent genes combined with an ensuing selection for tumor-driving events plays a major role in cancer progression. Cancer Res; 74(13); 3617–29. 2014 AACR.

Introduction cells (21, 22). The acquisition of the more permanent silencing Tumorigenesis is a complex process that is driven by a mark, 5mC, at the promoters of Polycomb target genes does number of genetic and epigenetic alterations, which often not fundamentally change their expression levels although result in aberrant gene expression (1–3). Mutations are gen- plasticity of expression will be reduced. For these reasons, it erally considered to be the primary drivers of tumorigenesis (4, has remained unclear whether methylation of Polycomb target 5). However, dysregulation of epigenetic regulatory mechan- genes plays an essential role in tumor promotion and is in fact a isms also contributes to malignant transformation. Methyla- tumor-driving event (23). fi tion of cytosine at CpG sequences in DNA leading to the Recently, genome sequencing has identi ed a high frequen- formation of 5-methylcytosine (5mC) is one of the most stable cy of mutations in epigenetic regulatory factors or chromatin and most widely studied epigenetic modifications (6). In most structural elements in human malignancies (24, 25). However, fi human tumors, widespread hypermethylation of CpG-rich technical limitations have made it dif cult to directly examine sequences (CpG islands) is observed along with genome-wide chromatin changes in normal and malignant tissues, with most in vitro DNA hypomethylation (1, 7–14). DNA hypermethylation in studies being limited to cell culture models. To deter- fi cancer is not a random event; it commonly affects specific mine the role of histone modi cations in targeting DNA gene classes and is seen most frequently at targets of the methylation and in altering gene expression patterns in human fi Polycomb repression complex (15–18) including numerous primary tumors, we have conducted the rst comprehensive homeobox genes (19, 20). However, many of the genes marked analysis of frozen tissues from patients with colorectal cancer by Polycomb in normal tissues and cell types, for example focusing on two histone methylation marks, the activating embryonic stem cells, are expressed at very low levels in these histone H3 lysine 4 trimethylation (H3K4me3) and the repressive histone H3 lysine 27 trimethylation (H3K27me3), which are thought to be critical for gene regulation. We found that Authors' Affiliations: Departments of 1Cancer Biology, 2Information genes carrying both histone modifications (bivalent genes) in 3 Sciences, and Molecular Medicine, Beckman Research Institute, City of normal tissue are characterized by substantial variability and Hope, Duarte, California; and 4Center for Molecular Medicine, University of Connecticut Health Center, Farmington, Connecticut undergo reorganization of these modifications in colorectal cancer, leading to activation of cancer-promoting genes as one Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). important outcome that can confer tumor-driving properties onto an emerging malignant cell population. M.A. Hahn and A.X. Li contributed equally to this work. Corresponding Author: Gerd P. Pfeifer. Department of Cancer Biology, Materials and Methods Beckman Research Institute, 1450 East Duarte Road, City of Hope, Duarte, CA 91010. Phone: 626-301-8853; Fax: 626-358-7703; E-mail: Human tissue samples [email protected] Human Duke's stage II colon cancers and matching normal doi: 10.1158/0008-5472.CAN-13-3147 mucosa were obtained from the Cooperative Human Tissue 2014 American Association for Cancer Research. Network (CHTN).

www.aacrjournals.org 3617

Hahn et al.

Gene expression analysis 300 mmol/L NaCl at 65C overnight. Primer sequences for Total RNA from patient samples was purified by using the qPCR of individual genes are available upon request. Analysis mirVana Kit (Ambion; Life Technologies). For whole genome was done with tissues from patient no. 14. Results are for expression analysis, GeneChip Human Gene 2.0 ST arrays triplicate experiments (SD). (Affymetrix) were used. Validation of gene expression changes by real-time reverse transcription-PCR was performed as Bioinformatics analysis described previously (18). All expression data were normalized All analysis was performed using R statistical language, to GAPDH in the same sample. except for gene ontology analysis, which was performed using DAVID annotation tools and Ingenuity Pathway Analysis. The Analysis of 5mC patterns heat maps were generated with Cluster v3.0 and Java Treeview For analysis of genomic 5mC patterns, the methylated CpG v2.0. Refseq genes were downloaded from the UCSC hg18 fi þ island recovery assay (MIRA) was used as described previously annotation database. Promoter was de ned as 1.4 to 0.6 (26). After genome amplification, the methylated DNA fraction kb relative to the transcription start site (TSS). For heat maps was hybridized versus input DNA on human CpG island/ with epigenetic marks, only genes with gene body length promoter microarrays (NimbleGen). The observed MIRA pat- greater than 2 kb were included. Genes that were differentially fi terns for single CpG islands were validated by combined expressed during colorectal cancer formation were identi ed > bisulfite restriction analysis (COBRA) as described previously as those with a log2-fold change 1. Probes were considered (18). Primer sequences are available upon request. positive if their normalized log2 ratios were above 1.5-fold. Peaks in each sample were defined as four or more consecutive Chromatin immunoprecipitation positive probes with either one or no gaps. A promoter was fi For chromatin immunoprecipitation (ChIP), frozen tissues de ned as bivalent if it contained overlapping H3K4me3 and < < were crushed with a plastic pestle in ice-cold phosphate- H3K27me3 peaks at expanded promoter areas ( 2.4 kb TSS buffered saline (PBS) and fixed for 10 minutes at room 0.6 kb). For heat-map analysis, genes were assumed activated temperature in 1% formaldehyde. ChIP and genome ampli- in all tumors and bivalent in mucosa if those genes had a fication protocols were performed as described previously promoter bivalent status in at least 3 of 4 normal colon samples (26). The following antibodies were used: anti-H3K4me3 and tumor-associated gene activation of log2 0.4 in four (39159; Active Motif), and anti-H3K27me3 (07-449; Millipore). patient sample pairs. For obtaining the H3K4me3 profile, H3K4me3 antibodies Affymetrix microarray data were processed by Expression fi were preblocked with an H3K9me3 peptide (Abcam) to Console v 2.0 with default settings. Transcription pro les of remove minor cross-reactivity of this antibody. After genome unique genes were obtained by averaging the transcriptional amplification, immunoprecipitated DNA was hybridized ver- levels of all different transcript isoforms of the same gene sus input DNA on human CpG island/promoter microarrays according to Affymetrix's annotation. Genes with log2 intensity < (NimbleGen). 5 in more than 10% of biopsies were considered uninforma- For sequential ChIP (ReChIP), the first immunoprecipita- tive and were removed. Affymetrix expression array data were tion was performed as described above except for the elution used to generate heat maps as described previously (27). All step, which was performed with SDS lysis buffer (1% SDS, 10 array data were deposited in the GEO database (accession mmol/L EDTA, 50 mmol/L Tris-HCl, pH 8.1, 1 cOmplete number GSE47076). Protease Inhibitor Cocktail; Roche Applied Science) for 10 minutes at 68C on a shaker at 1,000 rpm. After removal of Results beads, the samples were diluted with 1:10 ChIP dilution buffer Chromatin mapping studies with frozen tissue (0.01% SDS, 1.1% Triton X-100, 1.2 mmol/L EDTA, 16.7 mmol/L Previous correlative studies, mostly performed with cul- Tris-HCl, pH 8.1, 167 mmol/L NaCl, 22 mg/mL BSA, 1 cOm- tured cells (e.g., embryonic stem cells), indicated that the plete Protease Inhibitor Cocktail) and incubated with protein Polycomb-associated chromatin modification, H3K27me3, A/G PLUS-agarose (sc-2003; Santa Cruz Biotechnology), pre- predisposes sequences toward DNA hypermethylation in can- blocked with BSA and herring sperm DNA, for 30 minutes at cer (15–17, 20). To study aberrations of chromatin and DNA 4C. After removal of beads, chromatin was incubated with methylation in primary tumors, we directly profiled H3K4me3, anti-H3K4me3 (39159; Active Motif), anti-H3K27me3 (17-622; H3K27me3, and DNA methylation in 4 archived frozen colo- Millipore), or IgG (17-622, Millipore) antibodies overnight at rectal tumors with diagnosed Duke's stage II and 4 matching 4C followed by further incubation with protein A/G PLUS- normal mucosa samples, using human promoter and CpG agarose. Beads were washed once with low-salt immune island microarrays containing 28,000 CpG islands and complex wash buffer (0.1% SDS, 1% Triton X-100, 2 mmol/L 20,000 promoters. EDTA, 20 mmol/L Tris-HCl, pH 8.1, 150 mmol/L NaCl, 1 We used the methylated-CpG island recovery assay (MIRA; cOmplete Protease Inhibitor Cocktail) and twice with high salt ref. 19) to comprehensively identify DNA methylation changes immune complex wash buffer (0.1% SDS, 1% Triton X-100, in the analyzed tissues. By using an established bioinformatics 2 mmol/L EDTA, 20 mmol/L Tris-HCl, pH 8.1, 500 mmol/L approach (28), we identified between 484 and 2,098 tumor- NaCl, 1 cOmplete Protease Inhibitor Cocktail). Chromatin associated hypermethylated DNA regions in the four tumors was eluted with SDS lysis buffer for 10 minutes at 68Cona (tumor #6: 2,098 regions, tumor #8: 1,552 regions, tumor #12: shaker at 1,000 rpm and crosslinks were reversed in presence of 484 regions, and tumor #14: 1,676 regions). We validated the

3618 Cancer Res; 74(13) July 1, 2014 Cancer Research

Activation of Bivalent Genes in Colorectal Cancer

A Patient#6_N, R = –0.59, P < 0.0001 Patient#6_T, R = –0.57, P < 0.0001 C

2 2 Patient#6 Patient#8 Patient#12 Patient#14 2 2 1 1

0 0

–1 –1 5mC signal, log 5mC signal, log –2 –2

–2 –1 0 1 2 –2 –1 0 1 2

H3K4me3 signal,log2 H3K4me3 signal,log2

Patient#8_N, R = –0.62, P < 0.0001 Patient#8_T, R = –0.57, P < 0.0001

2 2 2 2

1 1

0 0

–1 –1 5mC signal, log 5mC signal, log

–2 –2

–2 –1 0 1 2 –2 –1 0 1 2

H3K4me3 signal,log2 H3K4me3 signal,log2

2 Patient#6, R = –0.7, P < 0.0001 2 Patient#8, R = –0.65, P < 0.0001 3 3

2 2

1 1

0 0

–1 –1

–2 –2

–3 –3 –2–3–1 0 1 2 3 –2–3–1 0 1 2 3 Fold change H3K4me3 (T vs. N), log change H3K4me3 (T vs. Fold N), log change H3K4me3 (T vs. Fold Fold change 5mC (T vs. N), log2 Fold change 5mC (T vs. N), log2

Figure 1. Verification of chromatin mapping with frozen tissues. A, DNA methylation analysis and H3K4me3 histone mapping was conducted with normal (N) and tumor (T) samples from patients #6 (top) and #8 (bottom). An inverse correlation between 5mC and H3K4me3 signal is observed for each sample. B, changes in 5mC signal between tumor (T) and normal (N) samples are plotted versus changes in H3K4me3 signal for patients #6 and #8. An inverse correlation is observed. Analysis was performed for promoters overlapping with CpG islands. The Pearson correlations between 5mC and H3K4me3 datasets are indicated. C, a heat-map analysis of changes (T vs. N) in H3K4me3, 5mC, and gene expression levels is shown for four patient sample pairs. H3K4me3 changes positively correlate with changes in gene expression but negatively correlate with changes in 5mC. Promoters of genes longer than 2 kb were sorted by cancer-associated H3K4me3 changes. Each row represents a gene promoter. Red indicates a gain and green represents a loss. aberrant methylation patterns by COBRA for several differ- P < 0.0001; Supplementary Fig. S2A). Our data revealed the well- entially methylated regions located in the TMEFF2, LIFR, established negative correlation (R ¼0.6, P < 0.0001) between CDKN2A, WNT5A, and MLH1 promoters (Supplementary Fig. 5mC and H3K4me3 at CpG islands (Fig. 1A: refs. 29 and 30). S1). In total, we analyzed 11 regions by bisulfite-based meth- Furthermore, changes of H3K4me3 at promoters occurring ods and observed that the two approaches yielded excellent between tumors and matching normal tissues were strongly correlation for all regions examined. linked to transcriptional changes analyzed at the RNA level To judge the quality of ChIP conducted on the frozen tissues, (Fig. 1C). Genes gaining H3K4me3 at promoters were frequent- we performed several bioinformatics comparisons of our ChIP ly activated in tumors and genes, which lost H3K4me3, were data with DNA methylation and gene expression data obtained repressed. Genome-wide, DNA methylation changes at pro- with the presumably more stable DNA or RNA material moters also showed a strong negative correlation with obtained from the same samples. As shown in Supplementary H3K4me3 changes between tumors and matching normal Fig. S2, ChIP patterns from frozen tissues are very reproducible tissues (Fig. 1B). Taken together, these data confirmed the between samples and provide remarkable consistency among reliability of our chromatin mapping approach using frozen independent ChIP analyses within the same tissue (R ¼ 0.89, tissue specimens.

www.aacrjournals.org Cancer Res; 74(13) July 1, 2014 3619

Hahn et al.

Variability of H3K27me3 Bivalent chromatin in embryonic stem cells and colonic Analysis of H3K27me3-marked genomic regions showed mucosa many similarities between tumor and nonmalignant matching Bivalent (H3K4me3 and H3K27me3 containing) chromatin specimens (e.g., Supplementary Figs. S2D and S3A). However, status is a key regulatory mechanism during embryonic stem in contrast to the generally more invariant H3K4me3 (ES) cell maintenance and differentiation (31, 32). We first patterns, we observed a substantial degree of instability of the asked to what extent bivalent promoter status is retained in H3K27me3 mark when comparing normal and tumor samples human colonic mucosa versus ES cells. We found that the (Fig. 2A and B and Supplementary Fig. S3). The Pearson number of bivalent promoters varies between 3,414 and 3,967 correlation coefficients of H3K27me3 when comparing within the normal colonic mucosal samples (Supplementary between tumors and matching normal mucosa were signifi- Fig. S4A). Approximately 50% to 60% of these bivalent pro- cantly reduced (0.75) in comparison to the Pearson correla- moters in normal colon are also bivalent in ES cells, whereas tion coefficients for H3K4me3 when comparing between 40% to 50% of bivalent promoters are specific for colon. Gene matching tumor and normal samples (0.91; Supplementary ontology analysis using DAVID (http://david.abcc.ncifcrf.gov/) Fig. S3A). Inspection of H3K27me3 patterns revealed that the showed that the gene group retaining the bivalent state from H3K27me3 mark was frequently lost along larger genomic ES cells to colon is more enriched with developmental genes in areas such as the HOX and histone gene clusters in tumors comparison to the gene groups that have lost the bivalent state (Fig. 2B and Supplementary Fig. S3B), similar as has been during development, and those, which perhaps obtained it de reported recently by Bert and colleagues (10). However, these novo in the colon (Supplementary Fig. S4B). Previously, aber- long-range areas of H3K27me3 loss represented only a rela- rant DNA methylation in cancer has been linked to the bivalent tively small fraction of all sequences that lost H3K27me3. state of promoters (15). To further assess this correlation, we Quite surprisingly, a detailed examination of the chroma- compared the bivalent promoter state in normal tissue to tin profiles also revealed that loss of the H3K27me3 mark aberrant methylation in matching tumors (Supplementary Fig. over larger sequence blocks can be alternatively associated S5A) and observed that genes with bivalent promoters com- with accumulation of aberrant DNA methylation at some prise the vast majority of promoters with cancer-associated promoters or with accumulation of the H3K4me3 mark at DNA methylation. Strikingly, the presence of H3K27me3 at other (even nearby) promoters (Fig. 2B and Supplementary promoters in normal colon epithelial tissue was associated Fig. S3B). Moreover, even a switch between transcriptional with 90% of all hypermethylated sites (Supplementary Fig. S5B) isoforms can occur in tumors when alternative routes and approximately 70% of aberrantly methylated promoters between H3K4me3 and 5mC deposition are used upon had a bivalent state in normal colon tissue (Supplementary Fig. H3K27me3 loss. For example, a loss of H3K27me3 at the S5A). HOXC cluster in sample #6 is associated with aberrant DNA Because the simultaneous presence of H3K4me3 and methylation at the short HOXC9 isoform promoter, at the H3K27me3 at promoters might reflect different cell popula- HOXC10 promoter and at the HOXC4 short isoform (tran- tions, we examined chromatin state at several candidate genes script variant 2) promoter. The same genomic locus shows by sequential immunoprecipitation (Fig. 3A and Supplemen- accumulation of H3K4me3 at the HOXC5 promoter (tran- tary Fig. S4C). These experiments confirmed presence of both script variant 1), HOXC6 promoter (transcript variant 1), histone marks on the same immunoprecipitated DNA frag- HOTAIR long isoform (transcript variant 1) promoter, the ments in normal mucosa. HOXC4 long isoform (transcript variant 1) promoter, and at the long HOXC9 isoform promoter (Fig. 2B). Thus, a loss of Chromatin changes at bivalent promoters H3K27me3 at distinct promoters can be involved in We next evaluated all epigenetic changes that target bivalent H3K4me3 accumulation or aberrant DNA methylation and promoters in colorectal cancer. We generated a heat map can have a major impact on transcript isoform usage in containing information for H3K27me3, H3K4me3, and 5mC colorectal tumors. To evaluate if these observations apply to changes occurring at promoters with bivalent status in all the cancer epigenome as a whole, and to determine whether analyzed normal colonic mucosa samples (Fig. 3B and Sup- alterations of the H3K27me3 patterns reflect transcriptional plementary Table S1). We found that the majority of bivalent changes, we used the heat-map approach for promoter promoters undergo H3K27me3 changes at least in 1 of 4 regions and sorted genes by H3K27me3 changes (Fig. 2C analyzed tumors, pointing to a high instability or variability and D). We examined the 15% of genes that showed the of bivalent promoters in colorectal cancer. Detailed analysis greatest loss of H3K27me3 at promoters, and we additionally revealed that some genes such as DMRTA2 and HOXD11 lose sorted these genes by H3K4me3 changes (Fig. 2C and D). We H3K27me3 in tumors in comparison to normal tissue and may found that H3K4me3 alterations within this group of genes undergo either gene activation associated with accumulation negatively correlated with accumulation of aberrant DNA of H3K4me3 or repression associated with aberrant DNA methylation at promoters. At the same time, accumulation methylation in different individual patients (Supplementary of H3K4me3 at promoters, which lost H3K27me3 in tumors, Fig. S3C). This fact may indicate a random nature of the was frequently associated with gene activation (Fig. 2D). consequences of Polycomb loss at promoters, at least for Thus, at a genome-wide level, we confirmed the two different certain genes. Clustering analysis indicated a large group of correlates of H3K27me3 loss in colorectal cancer, gain of promoters characterized by a loss of H3K27me3, loss of 5mC, or gene activation. H3K4me3, and gain of 5mC in all tumors (Fig. 3B; promoter

3620 Cancer Res; 74(13) July 1, 2014 Cancer Research

Activation of Bivalent Genes in Colorectal Cancer

Figure 2. Frequent variability of the Polycomb mark H3K27me3 in colon tumors. A, representative snapshots showing loss of H3K27me3 at the HOXB13 locus in 3 of 4 analyzed tumors (T) compared with matching normal mucosa (N) samples. The snapshots indicate log2 ratios of signal of IP versus input. Transcriptional start sites and transcript directions are shown. B, loss of H3K27me3 in tumors is linked to aberrant DNA methylation or accumulation of H3K4me3 at promoters. Representative snapshots of the HOXC cluster from patient #6 are shown. The colored circles mark changes of H3K4me3 or 5mC levels in tumor (T) versus normal (N) samples. Transcriptional start sites and transcript directions are shown. C, sorting of the genome by cancer- associated H3K27me3 changes at promoters. Each row represents a gene promoter. Values for each epigenetic mark were calculated as a mean of signal intensity log2 ratio for probes located at promoters. Red indicates a gain, and green represents a loss. For analysis, we used genes larger than 2 kb and with positive values of H3K27me3 in at least one of the tissues (T or N). D, loss of H3K27me3 in tumors is frequently associated either with a gain of H3K4me3 and gene activation or with aberrant DNA methylation and gene silencing. The top 15% of genes with greatest loss of H3K27me3 from the sorting in C were sorted by H3K4me3 changes. Each row represents a gene promoter. Values for each epigenetic mark were calculated as a mean of signal intensity log2 ratio for probes located at promoters. Red indicates a gain, and green represents a loss of the respective mark. groups marked by blue bars; Supplementary Fig. S5C). Fur- 5mC in some tumors but accumulation of H3K27me3 in other thermore, we observed a group of genes, which are character- tumors (Fig. 3B; promoter groups marked with black bars) ized by a loss of H3K4me3 in all tumors and accumulation of providing alternative routes toward gene silencing (33). Both

www.aacrjournals.org Cancer Res; 74(13) July 1, 2014 3621

Hahn et al.

Figure 3. Patterns of instability of bivalent promoters in colorectal cancer. A, ChIP and ReChIP of the FOXQ1 promoter for matched tumor and normal samples. Standard ChIP with anti-H3K4me3 and anti-H3K27me3 antibodies is shown in the left panel and sequential ChIP with both antibody combinations is shown in the right panel. B, clustering analysis of epigenetic changes at bivalent promoters in colon tumors. Analysis was done for promoters with bivalent status in all analyzed normal tissue samples. Each vertical row represents a gene promoter. Values for each epigenetic mark were calculated as a mean of signal intensity log2 ratio for probes located at promoters. Red indicates a gain, and green represents a loss. C, transcriptional changes of bivalent promoters, which undergo similar epigenetic changes during colorectal cancer progression. The box-plots represent transcriptional levels for three groups of genes with bivalent promoter status in all analyzed normal tissues: genes with a gain H3K27me3 at promoters in tumors, genes that lose H3K27me3 at promoters and acquire cancer- associated DNA methylation, and genes that lose H3K27me3 and gain H3K4me3 at promoters. The data are merged for four patient samples. D, activation or repression of genes with bivalent promoters in normal mucosa. The graph indicates the fraction of bivalent genes, which become activated or repressed at least 2-fold in tumor versus matching mucosa.

types of alterations, accumulation of H3K27me3 or of 5mC, are S5D). This phenomenon, which has been referred to as "Poly- indeed characterized by gene repression (Fig. 3C). This obser- comb switching" (34), was true for promoters with and without vation suggests that the variability of chromatin at bivalent bivalent state (Supplementary Fig. S5C and S5D). promoters can lead to two opposite outcomes, a gain or a loss of H3K27me3, and may result in gene repression according to Gene activation at bivalent promoters in cancer both scenarios if 5mC is gained in the latter. A majority of Our study identiﬁed a novel and distinct group of promoters promoters with aberrant DNA methylation in tumors was where a loss of H3K27me3 is associated with accumulation or associated with a loss of H3K27me3 (Supplementary Fig. retention of H3K4me3 in tumors but generally little or no

3622 Cancer Res; 74(13) July 1, 2014 Cancer Research

Activation of Bivalent Genes in Colorectal Cancer

change in DNA methylation levels (Fig. 3B; promoter group Interestingly, the genes coding for both transcription factors marked by a green bar). These promoters were associated with were activated because of loss of H3K27me3 in all tumors. In strong gene activation (Fig. 3C). Analysis of the transcriptome addition, two other intestinal stem cell signature genes, OLFM4 of bivalent promoters revealed a very high transcriptional and EPHB3 were activated together with a loss of the instability of bivalent genes in tumors. Approximately 26% to H3K27me3 mark in colorectal cancer (Fig. 4D). It is remarkable 46% of bivalent promoters were affected by transcriptional that the well-known proliferation marker, MKI67 (Ki67) and alterations (at least 2-fold) in each tumor (Fig. 3D). We found cancer stem cell markers such as Prominin 1 (CD133) and that 15% to 20% of the bivalent genes undergo transcriptional ALCAM (CD166) also became activated together with a loss of activation in tumors (Fig. 3D). bivalent state in tumors (Table 1 and Supplementary Table S3). To further elucidate the potential functional impact of These data suggest that loss of the bivalent state accompanied bivalent promoter activation on colorectal cancer progression, by activation of critical cell cycle drivers and tumor-promoting we performed a detailed analysis of genes that are bivalent or genes is a general and pervasive mechanism in colorectal nonbivalent in colonic mucosa and become activated at least 2- cancer. fold in at least two out of four analyzed tumors (Supplementary To evaluate the possibility that the apparent loss of Table S2). Nonbivalent genes activated in tumors were H3K27me3 in tumors reflects an intestinal stem cell origin enriched for genes important in mitotic progression including and presence of H3K27me3 in mucosa reflects a differenti- mitotic kinases, mitotic cyclins and kinesins, and checkpoint ation process in which this mark is acquired at promoters proteins (Supplementary Fig. S6A and Table S2). The activated of stem cell genes, we determined if cancer-associated bivalent genes were strongly associated with the colonic transcriptional changes in tumors carry primarily a gene transcriptome and transcriptional activity in colorectal cancer expression signature indicative of cells at the crypt base (Fig. 4A). Importantly, this group of genes was strongly where such stem cells reside. We used previously published enriched with genes involved in transcriptional regulation and expression profiling of human crypt top versus bottom associated with early development (Supplementary Fig. S6B samples (38). From 3,157 crypt bottom-specific genes, of and Table S2). The genes that become repressed in tumors and which 306 genes carry a bivalent status in normal mucosa, carry a bivalent status in mucosa were strongly linked to brain we found only 26 crypt bottom genes including LGR5, ASCL2, development (Supplementary Fig. S6C), whereas repressed MKi67,andCLDN1 to be activated in all tumors and bivalent nonbivalent genes did not show any functional categories with in mucosa. At the same time, very few genes (10) of 2,865 major enrichment (data not shown). Significantly, the non- crypt top-specific genes, of which 251 genes carry bivalent bivalent genes did not show any enrichment for transcription character in mucosa, were repressed in all tumors and were factors (Supplementary Table S2), neither for activation nor for bivalent in mucosa (Supplementary Fig. S6F). These data repression gene groups, suggesting that the enrichment of clearly suggest that the tumors we analyzed do not express a mitotic proteins in the nonbivalent group could be down- predominant stem cell–like phenotype. Analysis of genes, stream of crucial regulators (e.g., cyclin D) encoded by the which are not associated with crypt top or bottom according activated bivalent genes. According to Ingenuity pathway to Kosinski and colleagues (38), have bivalent status in analysis (www.ingenuity.com), the group of genes, which mucosa and are activated in all tumors, revealed a strong become activated in cancer and are bivalent in mucosa, was association of these genes with cancer and epithelial neo- strongly enriched with genes associated with cancer as a plasia (Supplementary Table S4). This fact suggests that this disease (Fig. 4B) and with important aspects of the trans- group of genes may play an important role in cancer formed phenotype (Supplementary Table S3). For example, one development. of the main regulators of the epithelial to mesenchymal We then tested if activation of bivalent genes may occur at transition (EMT), SNAI2 was activated together with a loss of an early stage of colorectal cancer progression. Based on bivalent status in tumor tissues (Supplementary Table S3). This publically available gene expression profiles of 32 human group of genes contains many genes encoding transcription adenomas and matching mucosa (39), we evaluated if bivalent factors and genes associated with cellular growth and prolif- genes, which become activated in all tumors, undergo tran- eration such as the oncogene cyclin D1 (CCND1), ribonucle- scriptional changes in colorectal adenomas. Bivalent genes otide reductase (RRM2), and MKi67, genes associated with cell were activated slightly preferentially over nonbivalent genes in adhesion and invasion including claudin1 (CLDN1) and EPCAM adenomas (18% vs. 15%). However, the majority (54%) of and other genes strongly implicated in cancer such as COX2/ bivalent genes activated in stage II colorectal cancer were also PTGS2 and the MET oncogene (Fig. 4D, Table 1, Supplementary activated in adenoma. Among the activated genes were stem Table S3, and Fig. S6D). cell markers including LGR5, SOX9, and ASCL2 and prolifera- Three crypt stem cell markers, ASCL2, LGR5, and SOX9 (35– tion-associated genes such as MKI67 and RRM2 (Fig. 4E). 37) belong to the group of most frequently activated genes (in all tumors) with bivalent promoters in normal colonic epithe- Transcriptional alterations of bivalent genes in inflamed lium (Fig. 4C and D, Table 1, and Supplementary Table S3). This mucosa fact was also confirmed for LGR5 in 11 sample pairs by qRT- Chronic inflammation is a strong risk factor for cancer PCR (Fig. 4C). Analysis of transcription factor binding sites by development (40). Similar to colon cancer, chronic inflam- DAVID for bivalent tumor-activated genes showed enrichment mation of the digestive tract is associated with aberrant for SOX9 and FOXQ1 binding sites (Supplementary Fig. S6E). DNA methylation (41, 42), which frequently coincides with a

www.aacrjournals.org Cancer Res; 74(13) July 1, 2014 3623

Hahn et al.

Figure 4. Gene activation associated with bivalent promoters in colon tumors affects cancer-promoting and stem cell genes. A, genes that become activated in colorectal cancer and carry bivalent promoter status in matching mucosa are strongly linked to colon and colorectal tumor transcriptomes. The top 10 transcriptomes are indicated. Analysis was done for genes, which are at least 2-fold activated in at least 2 of 4 tumors and are bivalent in matching mucosa. We performed data analysis by using the Unigene tissue speciﬁcity database (http://www.ncbi.nlm.nih.gov/unigene) incorporated into DAVID (http://david.abcc.ncifcrf.gov/). B, top 5 diseases and disorders (www.ingenuity.com) associated with genes activated at least 2-fold in 2 of 4 tumors and having bivalent promoter status in matching mucosa. The numbers of genes in each subgroup are indicated. C, activation of LGR5 in colorectal cancer. Transcript levels of LGR5 in colon tumors and matching mucosa were validated by qRT-PCR with further normalization to GAPDH.D,lossof H3K27me3 affects stem cell marker genes and genes strongly linked to colorectal cancer. Representative snapshots for patient #8 displaying epigenetic changes associated with gene activation in colon tumors are shown. E, bivalent genes are activated in adenomas. The heat map displays transcriptional changes in colorectal adenomas for bivalent genes, which are activated in colorectal tumors and are also activated at least 2-fold in at least one third of the adenomas. Expression data for 32 human adenomas and matching mucosa were obtained from GEO dataset GDS2947. For clustering analysis, we used only genes with bivalent state in at least 3 of 4 mucosa tissues and activated in all four analyzed colorectal cancers. Gene symbols are indicated.

3624 Cancer Res; 74(13) July 1, 2014 Cancer Research

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2014 American Association for Cancer Research. Downloaded from www.aacrjournals.org cancerres.aacrjournals.org Published OnlineFirstMay1,2014;DOI:10.1158/0008-5472.CAN-13-3147

Table 1. Bivalent genes activated in colorectal cancer

Wnt-pathway Activation Gene Classiﬁcation Activation target in UC LGR5a Leucine-rich repeat containing Crypt stem cell marker Colon cancer stem cells, Yes G protein–coupled receptor 5 crypt stem cells on September 30, 2021. © 2014American Association for Cancer Research. SOX9a SRY (sex determining region Y)-box 9 Crypt stem cell marker, Crypt stem cells, colon cancer Yes transcription factor ASCL2, aAchaete-scute complex homolog 2 Crypt stem cell marker, Crypt stem cells, colon cancer Yes transcription factor CD133 (PROM1), Prominin1 Stem cell marker Colon cancer stem cells, crypt stem cells FOXQ1, aForkhead box Q1 Transcription factor Colorectal cancer Yes Yes COX2 (PTGS2), prostaglandin-endoperoxide Inﬂammation and cancer-associated Colon cancer Yes Yes synthase 2 cyclooxygenase-2 MET, Met proto-oncogene (hepatocyte Oncogene, hepatocyte growth Colon cancer, cancer stem cells, Yes growth factor receptor) factor receptor stem cells

CCND1, cyclin D1 Cyclin, oncogene Colon cancer Cancer Colorectal in Genes Bivalent of Activation CLDN1, aClaudin1 Tight junction molecule Colon cancer Yes Yes EPCAM, epithelial cell adhesion molecule Adhesion molecule Colon cancer, cancer stem cells MKI67, antigen identiﬁed by monoclonal General proliferation marker Proliferating cells Yes antibody Ki67

a acrRs 41)Jl ,2014 1, July 74(13) Res; Cancer Activated in all four tumors. 3625 Published OnlineFirst May 1, 2014; DOI: 10.1158/0008-5472.CAN-13-3147

Hahn et al.

Figure 5. Activation of genes with bivalent promoters in UC. A, transcriptional changes in UC for genes with bivalent promoter status in mucosa. Expression profiles for 13 human colon biopsies from noninflamed controls, 15 inflamed mucosa samples from patients with UC, and 7 unaffected mucosa tissues from patients with UC were obtained from GEO dataset GSE38713. Bivalent genes were divided into three groups: genes that become activated at least 2-fold inat least one third of inflamed tissues versus noninflamed controls, genes that became repressed at least 2-fold, and genes that were not activated or repressed. B, differentially expressed bivalent genes in inflamed tissues from patients with UC. Clustering analysis was performed for bivalent genes, which underwent at least 2-fold transcriptional activation or repression in at least one third of inflamed tissues from patients with UC versus noninflamed controls. For this analysis, we used genes with bivalent state in at least 3 of 4 mucosal tissues. C, top 5 diseases and disorders (www.ingenuity.com) associated with genes activated at least 2-fold in inflamed biopsies versus noninflamed controls and having bivalent promoter status in all four analyzed mucosal tissues. The numbers of genes in each subgroup are indicated. D, bivalent genes, which become activated in colorectal cancer and are activated also in inflamed tissues from patients with UC. Clustering analysis is shown for bivalent genes, which underwent at least 2-fold transcriptional changes in at least one third of inflamed UC tissues and were activated in all tumors. Clustering analysis was done for genes that carry a bivalent promoter status in 3 of 4 mucosal samples.

loss of the Polycomb mark, H3K27me3 (18). We determined cancer. We used publically available expression profiles for if activation of H3K27me3-marked genes occurs during 15 inflamed mucosal samples from patients with ulcerative inflammation and if these activated genes play a role in colitis (UC), 7 unaffected mucosal tissues from patients with

3626 Cancer Res; 74(13) July 1, 2014 Cancer Research

Activation of Bivalent Genes in Colorectal Cancer

tasis. Among the activated bivalent genes were several key components of the WNT signaling pathway including LGR5, CD133, LEF1, TCF7, WNT2, and WNT3. This pathway is commonly found to have a high level of constitutive activity in colorectal cancer. Also, many cell-cycle drivers and proliferation-associated genes were found in this category, including cyclin D, MKi67, RRM2, ETV4, and FGF19 along with many transcription factor genes such as SOX9, TP73, MYB, FOXA2, ETV4, and TEAD4, a transcription factor negatively regulated by the Hippo tumor suppressor pathway (Supplementary Table S3). It is easily understandable that selection for activation of these sets of genes would provide a growth advantage to the tumor cell population. However, the DNA methylation pathway, if operative at bivalent promoters after loss of Figure 6. Fate of bivalent promoters in colorectal cancer. The model H3K27me3, may not immediately provide such a selective depicts the state of promoters containing H3K4me3 (green) and advantage because the affected genes are already expressed H3K27me3 (red) in normal colonic mucosa and how this state has at a very low level while in the bivalent state in normal cells. resolved in cancer cells. The disappearance of the H3K27me3 mark Two nonexclusive models of colorectal cancer initiation can be associated with DNA methylation and permanent repression of have been discussed. According to mouse models, it has been genes (bottom) or, alternatively, with activation of growth-promoting þ genes (top). suggested that Lgr5 cells from the crypt bottom comprise the tumor-initiating cell population (44). Other models favor a UC and 13 human colon biopsies from noninflamed controls "top-down" model, in which dysplastic cells can originate in the obtained from GEO dataset GSE38713 (43). Analysis of upper areas of crypts (45, 46). The latter model invokes a expression profiles for genes with bivalent status in mucosa dedifferentiation process in which the cells re-acquire stem – revealed that almost 20% of bivalent genes already undergo cell like properties. Our data are consistent with a model in transcriptional changes during inflammation (Fig. 5A and which colorectal cancer can arise from differentiated epithelial Supplementary Table S1). Bivalent genes were not activated cells in which the bivalent chromatin state resolves into active preferentially over nonbivalent genes (11% vs. 10%, respec- or inactive forms. It is likely that the instability of bivalent tively). Remarkably, patterns of activation and repression of genes is a mechanism that predisposes the colonic epithelium fl bivalent genes strongly divided biopsies with and without to cancer by being operative already during in ammatory k inflammation (Fig. 5B). According to ingenuity pathway processes. Elevated NF- B signaling, as found in chronic fl analysis, these activated bivalent genes are strongly linked in ammation, has been shown to enhance Wnt activation and to cancer (Fig. 5C and Supplementary Fig. S6G) and are to induce dedifferentiation of non-stem cells in the colon that fl associated with functions associated with cancer develop- acquire tumor-initiating capacity (45). In ammation has been ment such as cellular movement, cell death and survival, and linked to methylation of Polycomb-marked genes in the intes- proliferation (Supplementary Fig. S6H). Moreover, approxi- tinal epithelium of mice (18). We show here that activation of mately 40% of genes (31/80) that are activated in all colo- cancer-relevant genes occurs at bivalent promoters in fl rectal tumors and have bivalent status in adjacent mucosa in amed differentiated mucosa from patients with UC. fl are activated in inflamed tissues from patients with UC (Fig. Besides in ammation, aging is a major risk factor for cancer. fl 5D). Among these genes, we found FOXQ1, CLDN1, LEF1,and Similar to in ammatory conditions, the aging process has been MKI67.However,incontrasttoadenoma,expressionofstem associated with DNA hypermethylation of Polycomb target cell markers including SOX9, LGR5,andASCL2 was not genes (47, 48). If the loss of the Polycomb mark underlies both activated in inflamed tissues. These data provide additional DNA methylation silencing and the activation of bivalent genes fl evidence for bivalent gene activation in differentiated muco- in cancer cells, a mechanistic connection between in amma- sa, in the absence of overt tumor formation, and demon- tion, aging, and cancer can be proposed, in which gene strates that the phenomenon of bivalent promoter activation activation because of H3K27me3 loss may easily be a dominant occurs already during chronic inflammation, a condition driving force for the malignant change. strongly correlated with cancer predisposition. Questions remain as to the possible mechanisms under- lying the high variability of the Polycomb mark. Earlier studies showed that the H3K27 methyltransferase EZH2 is Discussion downregulated in stressed and senescing populations of Our data suggest that loss of the bivalent (H3K4me3 and cells, which coincided with decreased levels of H3K27me3 H3K27me3) chromatin state at promoters is accompanied by at the INK4A-ARF locus (49). However, it has often been either gain of 5mC (when both marks are lost) or by gene observed that EZH2 is overexpressed in human tumors (50, activation (when H3K27me3 is lost; Fig. 6). The latter event, as 51). In our samples, we observed that EZH2 was expressed at we demonstrate here for the first time, is likely to be a critical higher levels in tumors than in normal mucosa but we did step in cancer pathogenesis responsible for the activation of not find substantial differences in expression of the crucial genes leading to progression to malignancy and metas- H3K27me3 demethylases KDM6A or KDM6B.Itremainsto

www.aacrjournals.org Cancer Res; 74(13) July 1, 2014 3627

Hahn et al.

be determined whetherstructuralorfunctionalaberrations Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. Wu, D.A. Drew, D.W. Rosenberg of the Polycomb complex itself underlie the cancer-associ- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, ated dysfunction of Polycomb marking at bivalent promo- computational analysis): M.A. Hahn, A.X. Li, X. Wu, R. Yang, G.P. Pfeifer ters or if bivalency of these loci itself predisposes them to Writing, review, and or revision of the manuscript: M.A. Hahn, A.X. Li, X. Wu, D.A. Drew, D.W. Rosenberg, G.P. Pfeifer inherent or transcription factor–driven variability and loss Administrative, technical, or material support (i.e., reporting or orga- of the H3K27me3 mark, resulting in activation of genes with nizing data, constructing databases): A.X. Li Study supervision: G.P. Pfeifer functional importance for tumor progression.

Disclosure of Potential Conﬂicts of Interest Grant Support G.P. Pfeifer has ownership interest (including patents) from Active Motif. G.P. This work was supported by NIH grant CA 084469 to G.P. Pfeifer. Pfeifer is a consultant/advisory board member of Zymo Research. No potential The costs of publication of this article were defrayed in part by the payment of ﬂ con icts of interest were disclosed by the other authors. page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Authors' Contributions Conception and design: M.A. Hahn, G.P. Pfeifer Received November 4, 2013; revised March 18, 2014; accepted April 3, 2014; Development of methodology: M.A. Hahn, A.X. Li, G.P. Pfeifer published OnlineFirst May 1, 2014.

References 1. Jones PA, Baylin SB. The epigenomics of cancer. Cell 2007;128: 19. Rauch T, Li H, Wu X, Pfeifer GP. Mira-assisted microarray analysis, a 683–92. new technology for the determination of DNA methylation patterns, 2. Shen H, Laird PW. Interplay between the cancer genome and epigen- identifies frequent methylation of homeodomain-containing genes in ome. Cell 2013;153:38–55. lung cancer cells. Cancer Res 2006;66:7939–47. 3. Timp W, Feinberg AP. Cancer as a dysregulated epigenome allowing 20. Rauch T, Wang Z, Zhang X, Zhong X, Wu X, Lau SK, et al. Homeobox cellular growth advantage at the expense of the host. Nat Rev Cancer gene methylation in lung cancer studied by genome-wide analysis with 2013;13:497–510. a microarray-based methylated CpG island recovery assay. Proc Natl 4. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA Jr, Acad Sci U S A 2007;104:5527–32. Kinzler KW. Cancer genome landscapes. Science 2013;339:1546–58. 21. Simon JA, Kingston RE. Occupying chromatin: Polycomb mechan- 5. Fearon ER. Molecular genetics of colorectal cancer. Annu Rev Pathol isms for getting to genomic targets, stopping transcriptional traffic, 2011;6:479–507. and staying put. Mol Cell 2013;49:808–24. 6. Suzuki MM, Bird A. DNA methylation landscapes: provocative insights 22. Aloia L, Di Stefano B, Di Croce L. Polycomb complexes in stem cells from epigenomics. Nat Rev Genet 2008;9:465–76. and embryonic development. Development 2013;140:2525–34. 7. Rauch TA, Zhong X, Wu X, Wang M, Kernstine KH, Wang Z, et al. High- 23. Kalari S, Pfeifer GP. Identification of driver and passenger DNA resolution mapping of DNA hypermethylation and hypomethylation in methylation in cancer by epigenomic analysis. Adv Genet 2010; lung cancer. Proc Natl Acad Sci U S A 2008;105:252–7. 70:277–308. 8. Berman BP, Weisenberger DJ, Aman JF, Hinoue T, Ramjan Z, Liu Y, 24. You JS, Jones PA. Cancer genetics and epigenetics: two sides of the et al. Regions of focal DNA hypermethylation and long-range hypo- same coin? Cancer Cell 2012;22:9–20. methylation in colorectal cancer coincide with nuclear lamina-asso- 25. Suva ML, Riggi N, Bernstein BE. Epigenetic reprogramming in cancer. ciated domains. Nat Genet 2012;44:40–6. Science 2013;339:1567–70. 9. Estecio MR, Issa JP. Dissecting DNA hypermethylation in cancer. 26. Hahn MA, Qiu R, Wu X, Li AX, Zhang H, Wang J, et al. Dynamics of 5- FEBS Lett 2011;585:2078–86. hydroxymethylcytosine and chromatin marks in mammalian neuro- 10. Bert SA, Robinson MD, Strbenac D, Statham AL, Song JZ, Hulf T, et al. genesis. Cell Rep 2013;3:291–300. Regional activation of the cancer genome by long-range epigenetic 27. Hahn MA, Wu X, Li AX, Hahn T, Pfeifer GP. Relationship between gene remodeling. Cancer Cell 2013;23:9–22. body DNA methylation and intragenic H3K9me3 and H3K36me3 11. Weichenhan D, Plass C. The evolving epigenome. Hum Mol Genet chromatin marks. PLoS ONE 2011;6:e18844. 2013;22:R1–6. 28. Wu X, Rauch TA, Zhong X, Bennett WP, Latif F, Krex D, et al. CpG island 12. Putiri EL, Robertson KD. Epigenetic mechanisms and genome stabil- hypermethylation in human astrocytomas. Cancer Res 2010;70: ity. Clin Epigenetics 2011;2:299–314. 2718–27. 13. Nagarajan RP, Fouse SD, Bell RJ, Costello JF. Methods for cancer 29. Ooi SK, Qiu C, Bernstein E, Li K, Jia D, Yang Z, et al. Dnmt3l connects epigenome analysis. Adv Exp Med Biol 2013;754:313–38. unmethylated lysine 4 of histone H3 to de novo methylation of DNA. 14. Hughes LA, Melotte V, de Schrijver J, de Maat M, Smit VT, Bovee JV, Nature 2007;448:714–7. et al. The CpG island methylator phenotype: what's in a name? Cancer 30. Edwards JR, O'Donnell AH, Rollins RA, Peckham HE, Lee C, Milekic Res 2013;73:5858–68. MH, et al. Chromatin and sequence features that define the fine and 15. Ohm JE, McGarvey KM, Yu X, Cheng L, Schuebel KE, Cope L, et al. A gross structure of genomic methylation patterns. Genome Res stem cell-like chromatin pattern may predispose tumor suppressor 2010;20:972–80. genes to DNA hypermethylation and heritable silencing. Nat Genet 31. Bernstein BE, Mikkelsen TS, Xie X, Kamal M, Huebert DJ, Cuff J, et al. A 2007;39:237–42. bivalent chromatin structure marks key developmental genes in 16. Schlesinger Y, Straussman R, Keshet I, Farkash S, Hecht M, Zimmer- embryonic stem cells. Cell 2006;125:315–26. man J, et al. Polycomb-mediated methylation on Lys27 of histone 32. Voigt P, Tee WW, Reinberg D. A double take on bivalent promoters. H3 pre-marks genes for de novo methylation in cancer. Nat Genet Genes Dev 2013;27:1318–38. 2007;39:232–6. 33. Kondo Y, Shen L, Cheng AS, Ahmed S, Boumber Y, Charo C, et al. 17. Widschwendter M, Fiegl H, Egle D, Mueller-Holzner E, Spizzo G, Marth Gene silencing in cancer by histone H3 lysine 27 trimethylation inde- C, et al. Epigenetic stem cell signature in cancer. Nat Genet 2007; pendent of promoter DNA methylation. Nat Genet 2008;40:741–50. 39:157–8. 34. Gal-Yam EN, Egger G, Iniguez L, Holster H, Einarsson S, Zhang X, et al. 18. Hahn MA, Hahn T, Lee DH, Esworthy RS, Kim BW, Riggs AD, et al. Frequent switching of polycomb repressive marks and DNA hyper- Methylation of polycomb target genes in intestinal cancer is mediated methylation in the PC3 prostate cancer cell line. Proc Natl Acad Sci U S by inflammation. Cancer Res 2008;68:10280–9. A 2008;105:12979–84.

3628 Cancer Res; 74(13) July 1, 2014 Cancer Research

Activation of Bivalent Genes in Colorectal Cancer

35. Barker N, van Es JH, Kuipers J, Kujala P, van den Born M, Cozijnsen M, ulcerative colitis in remission reveals lasting epithelial cell alterations. et al. Identification of stem cells in small intestine and colon by marker Gut 2013;62:967–76. gene Lgr5. Nature 2007;449:1003–7. 44. Barker N, Ridgway RA, van Es JH, van de Wetering M, Begthel H, van 36. van der Flier LG, van Gijn ME, Hatzis P, Kujala P, Haegebarth A, Stange den Born M, et al. Crypt stem cells as the cells-of-origin of intestinal DE, et al. Transcription factor achaete scute-like 2 controls intestinal cancer. Nature 2009;457:608–11. stem cell fate. Cell 2009;136:903–12. 45. Schwitalla S, Fingerle AA, Cammareri P, Nebelsiek T, Goktuna SI, 37. Stange DE, Engel F, Longerich T, Koo BK, Koch M, Delhomme N, et al. Ziegler PK, et al. Intestinal tumorigenesis initiated by dedifferentiation Expression of an ASCL2 related stem cell signature and IGF2 in and acquisition of stem-cell-like properties. Cell 2013;152:25–38. colorectal cancer liver metastases with 11p15.5 gain. Gut 2010;59: 46. Shih IM, Wang TL, Traverso G, Romans K, Hamilton SR, Ben-Sasson 1236–44. S, et al. Top-down morphogenesis of colorectal tumors. Proc Natl 38. Kosinski C, Li VS, Chan AS, Zhang J, Ho C, Tsui WY, et al. Gene Acad Sci U S A 2001;98:2640–5. expression patterns of human colon tops and basal crypts and BMP 47. Beerman I, Bock C, Garrison BS, Smith ZD, Gu H, Meissner A, et al. antagonists as intestinal stem cell niche factors. Proc Natl Acad Sci U S Proliferation-dependent alterations of the DNA methylation landscape A 2007;104:15418–23. underlie hematopoietic stem cell aging. Cell Stem Cell 2013;12: 39. Sabates-Bellver J, Van der Flier LG, de Palo M, Cattaneo E, Maake C, 413–25. Rehrauer H, et al. Transcriptome profile of human colorectal adeno- 48. Maegawa S, Hinkal G, Kim HS, Shen L, Zhang L, Zhang J, et al. mas. Mol Cancer Res 2007;5:1263–75. Widespread and tissue specific age-related DNA methylation changes 40. Terzic J, Grivennikov S, Karin E, Karin M. Inflammation and colon in mice. Genome Res 2010;20:332–40. cancer. Gastroenterology 2010;138:2101–14 e5. 49. Bracken AP, Kleine-Kohlbrecher D, Dietrich N, Pasini D, Gargiulo G, 41. Issa JP, Ahuja N, Toyota M, Bronner MP, Brentnall TA. Accelerated Beekman C, et al. The polycomb group proteins bind throughout the age-related CpG island methylation in ulcerative colitis. Cancer Res Ink4a-Arf locus and are disassociated in senescent cells. Genes Dev 2001;61:3573–7. 2007;21:525–30. 42. Chiba T, Marusawa H, Ushijima T. Inflammation-associated can- 50. Bracken AP, Pasini D, Capra M, Prosperini E, Colli E, Helin K. Ezh2 is cer development in digestive organs: mechanisms and roles for downstream of the PRB-E2F pathway, essential for proliferation and genetic and epigenetic modulation. Gastroenterology 2012;143: amplified in cancer. EMBO J 2003;22:5323–35. 550–63. 51. Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Kumar-Sinha 43. Planell N, Lozano JJ, Mora-Buch R, Masamunt MC, Jimeno M, Ordas I, C, Sanda MG, et al. The polycomb group protein EZH2 is involved in et al. Transcriptional analysis of the intestinal mucosa of patients with progression of prostate cancer. Nature 2002;419:624–9.

www.aacrjournals.org Cancer Res; 74(13) July 1, 2014 3629

Loss of the Polycomb Mark from Bivalent Promoters Leads to Activation of Cancer-Promoting Genes in Colorectal Tumors

Maria A. Hahn, Arthur X. Li, Xiwei Wu, et al.

Cancer Res 2014;74:3617-3629. Published OnlineFirst May 1, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-13-3147

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2014/05/01/0008-5472.CAN-13-3147.DC1

Cited articles This article cites 51 articles, 21 of which you can access for free at: http://cancerres.aacrjournals.org/content/74/13/3617.full#ref-list-1

Citing articles This article has been cited by 12 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/74/13/3617.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/74/13/3617. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.