MUC1 Expression Is Regulated by DNA Methylation and Histone H3 Lysine 9 Modification in Cancer Cells

Research Article

Norishige Yamada,1 Yukari Nishida,1 Hideaki Tsutsumida,1 Tomofumi Hamada,2 Masamichi Goto,1 Michiyo Higashi,1 Mitsuharu Nomoto,1 and Suguru Yonezawa1

1Department of Human Pathology, Field of Oncology and 2Department of Oral and Maxillofacial Rehabilitation, Field of Oral and Maxillofacial Surgery, Kagoshima University Graduate School of Medical and Dental Sciences, Sakuragaoka, Kagoshima, Japan

Abstract Regulation of MUC1 expression has been studied extensively, and TATA box, ubiquitous cis-acting elements, GC-boxes, and MUC1 is a transmembrane mucin that is highly expressed in binding sites for transcriptional regulators such as Sp1 and various cancers and correlates with malignant potential. MUC1 Important cancer-related genes such as p16 and E-cadherin GATA3 have been identified in the promoter (6–8). are controlled epigenetically; however, MUC1 has been over- However, the methylation status of the large number of CpG sites in the promoter region has yet to be clarified, in contrast to looked in epigenetics. Herein, we provide the first report that MUC1 MUC1 gene expression is regulated by DNA methylation and the coding region (9). Methylation of cytosine in genomic histone H3 lysine 9 (H3-K9) modification of the MUC1 DNA playsan important role in gene regulation and especiallyin promoter. The recently developed MassARRAY assay was gene silencing (10, 11), and generally, the promoter region of a transcribed gene is hypomethylated (12, 13). To examine the performed to investigate the DNA methylation status of 184 f MUC1 CpG sites from À2,753 to +263. Near the transcriptional start methylation profilesof 184 CpG sitesin 3,000 bp in the promoter in cancer cell lines, we performed a methylation analysis site, the DNA methylation level of MUC1-negative cancer cell lines (e.g., MDA-MB-453) was high, whereas that of MUC1- (MassARRAY) that uses base-specific cleavage of nucleic acids positive cell lines (e.g., MCF-7) was low. Histone H3-K9 (14). Samples were analyzed by matrix-assisted laser desorption modification status was also closely related to MUC1 gene ionization time-of-flight mass spectrometry. This method permits expression. Furthermore, MUC1 mRNA expression in MUC1- high-throughput identification of methylation sites and semiquan- negative cells was restored by treatment with the DNA titative measurement at single or multiple CpG sites. Based on the methylation inhibitor 5-aza-2¶-deoxycytidine. Our results results of the MassARRAY analysis, we designed methylation- indicate that DNA methylation and histone H3-K9 modifica- specific PCR (MSP) primers to ensure the CpG sites related to tion in the 5¶ flanking region play a critical role in MUC1 gene gene expression. expression, and this study defines MUC1 as a new member of Modification of histone tails also plays a critical role in the class of epigenetically controlled genes. An understanding epigenetic silencing (15, 16). Acetylation of lysine residues on of the epigenetic changes of MUC1 may be of importance for histone H3 leads to formation of an open chromatin structure, diagnosis of carcinogenic risk and prediction of outcome for whereasmethylation of K9 on histone H3 isa marker of heterochromatin (17). Functional interactionsamong DNA meth- cancer patients. [Cancer Res2008;68(8):2708–16] ylation, histone modification, and gene expression is the focus of studies on epigenetic control (18, 19). Therefore, to reveal the Introduction relationship between DNA methylation and histone modification, Mucinsare high molecular weight glycoproteinswith oligosac- chromatin immunoprecipitation (ChIP) primerswere designedto charides attached to serine or threonine residues of the mucin core target regions similar to those targeted by the MSP primers. We protein backbone via O-glycosidic linkages. These proteins are also treated cells that were MUC1-negative or those with low produced by variousepithelial cells.The MUC1 transmembrane expression of MUC1 with a DNA methylation inhibitor, 5-aza-2¶- glycoprotein is a member of the mucin family that is expressed at a deoxycytidine (5-AzadC), and a histone deacetylase inhibitor, basal level by normal ductal epithelial cells of secretory organs, trichostatin A, to confirm that DNA methylation and histone H3 including pancreas, breast, lung, and gastrointestinal tract (1), and modification suppress expression of MUC1 mRNA. In thisreport, is overexpressed and aberrantly glycosylated in most cases of we describe an epigenetic mechanism through which MUC1 gene adenocarcinoma (2). An elevated level of MUC1 protein playsa role expression is regulated by a tightly related combination of DNA in tumor progression, especially in the process of metastasis (3, 4). methylation and histone histone H3 lysine 9 (H3-K9) modification Our immunohistochemical studies of mucin expression in various in the 5¶ flanking region of the MUC1 promoter. human tumors, including pancreatic tumors, have also shown that MUC1 expression is related to the invasive proliferation of tumors Materials and Methods and ispredictive of a poor outcome for patients(5). Cells and treatment. Human pancreatic carcinoma cell linesHPAFII (MUC1+), BxPC3 (MUC1+), PANC1 (MUC1+/À), human breast cancer cell linesMCF-7 (MUC1+), T-47D (MUC1+), MDA-MB-453 (MUC1 À) and human colon adenocarcinoma cell linesCaco2 (MUC1 À), and LS174T (MUC1+/À) Requests for reprints: Masamichi Goto, Department of Human Pathology, Field were obtained from American Type Culture Collection. HPAFII, MCF-7, of Oncology, Kagoshima University Graduate School of Medical and Dental Sciences, LS174T, and Caco2 cellswere cultured in Eagle’sMEM (Sigma), PANC1 cells 8-35-1 Sakuragaoka, Kagoshima 890-8544, Japan. Phone: 81-99-275-5270; Fax: 81-99- were cultured in DMEM (Sigma), BxPC3 and T-47D cellswere cultured in 265-7235; E-mail: [email protected]. I2008 American Association for Cancer Research. RPMI 1640 (Sigma), and MDA-MB-453 cellswere cultured in Leibovitz’s doi:10.1158/0008-5472.CAN-07-6844 L-15 medium (Invitrogen). All media were supplemented with 10% fetal

Cancer Res 2008; 68: (8). April 15, 2008 2708 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 2008 American Association for Cancer Research. Epigenetic Control of MUC1 bovine serum (Invitrogen) and 100 U/mL penicillin/100 Ag/mL streptomy- was amplified by PCR using a Fast Cycling PCR kit (Qiagen). The target cin (Sigma). MUC1-negative cellsor cellswith low MUC1 were split regions were amplified using the primer pairs shown in Table 1. The PCR 24 h before treatment. PANC1, MDA-MB-453, and LS174T cellswere conditionswere 95 jC for 5 min, 36 cyclesat 96 jC for 5 s, 58jC for 5 s, and incubated for 5 d with 100 Amol/L 5-AzadC (Sigma) and/or for 5 d with 68jC for 3 s, with a final extension reaction at 72jC for 1 min. The amplified 500 nmol/L trichostatin A (TSA; Sigma). Caco2 cells were incubated for 5 d products were subjected to 1.5% agarose gel electrophoresis. with 100 Amol/L 5-AzadC and/or for 5 d with 50 nmol/L TSA. Media were ChIP assay. The ChIP assay was carried out using an EpiQuik Chromatin changed every 24 h. Immunoprecipitation kit (Epigentek, Inc.), according to the manufacturer’s Quantitative reverse transcription-PCR analysis. The mRNA from instructions. The nucleoprotein complexes were sonicated to reduce the cellsthat had or had not undergone 5-AzadC, TSA, or 5-AzadC/TSA sizes of DNA fragments to 300 to 500 bp using a Bioruptor (Cosmo Bio). combination treatment waspurified with a RNeasyMini kit (Qiagen). Of Four micorgrams of normal mouse IgG was used as the negative control, a total of 100 AL mRNA, 20 AL was reverse transcribed with Random- and antidimethyl histone H3-K9 antibody (Abcam) and antiacetyl histone Hexamers(Applied Biosystems).A 3.5- AL cDNA aliquot wasamplified in H3-K9 antibody (Upstate Biotechnologies) were used for each immunopre- 25 ALof2Â Taq Man Universal Master Mix, 2.5 ALof20ÂTarget Assay cipitation. Immunoprecipitated DNA was amplified by PCR using a Fast Mix, and 2.5 ALof20Â Control Assay Mix (Applied Biosystems) under the Cycling PCR kit (Qiagen). The ChIP primerswere designedto target a region following PCR conditions: 2 min at 50jC, 10 min at 95jC, 45 cyclesof 15 s similar to the target region of the MSP primers (Table 1). The PCR at 95jC, and 1 min at 60jC. The primersand probeswere designedand conditionswere 95 jC for 5 min, 38 to 40 cyclesat 96 jCfor5s,59jCfor5s, synthesized by Applied Biosystems. The product number of the Target and 68jC for 7 s, with a final extension reaction at 72jC for 1 min. The Assay Mix used for MUC1 was Hs00410317. Human glyceraldehyde-3- amplified products were subjected to 1.5% agarose gel electrophoresis. phosphate dehydrogenase (GAPDH; product number 4310884E) was used to calibrate the original concentration of mRNA; i.e., the concentration of Results mRNA in the cell wasdefined asthe ratio of target mRNA copiesversus Restoration of the MUC1 mRNA level by 5-AzadC and GAPDH mRNA copies. In this analysis, data from three separate experimentswere averaged. 5-AzadC/TSA. Experimentswere performed usingthree human Immunohistochemical staining. MUC1 protein expression levels were pancreatic cancer cell lines, HPAFII, BxPC-3, and PANC1; three assessed by immunohistochemistry. MUC1 was detected using a monoclo- human breast cancer cell lines, MCF-7, T-47D, and MDA-MB-453; nal antibody against MUC1 core glycoprotein (mouse IgG; Novocastra and two human colon adenocarcinoma cell lines, Caco2, and LaboratoriesLtd.; dilution, 1:100 for cell culture; incubation period, 1 h at LS174T. Expression levels of MUC1 mRNA and protein in the cell 37jC). Immunohistochemical staining was performed by an immunoper- lines were examined by reverse transcription-PCR (RT-PCR) oxidase method using a Vectastain Elite ABC kit (Vector Laboratories), as analysis and immunohistochemical staining (Fig. 1A and B). described previously (20). HPAFII, BxPC-3, MCF-7, and T-47D cells expressed MUC1, but MUC1 gene promoter sequencing. Genomic DNA wasextracted from the seven cell lines using a DNeasy Tissue System (Qiagen) according to the PANC1, MDA-MB-453, Caco2, and LS174T cellsdid not do so.To manufacturer’s instructions. DNA was PCR amplified using six pairs of examine the effectsof DNA methylation and histone modification sense and antisense primers (Table 1) in the full-length MUC1 promoter. on MUC1 gene expression, quantitative RT-PCR analysis was Primer sequences were based on a previously published sequence (21). PCR performed in MUC1-negative cells or cells with low expression of fragments were sequenced using single strand sequencing method MUC1 treated with a DNA demethylating agent, 5-AzadC, a hist- (Hokkaido System Science Co., Ltd.). Sequences were analyzed with an one deacetylase inhibitor, TSA, or 5-AzadC/TSA in combination ABI Prism 310 Genetic Analyzer (PE Applied Biosystems). (Fig. 1C). A 3-fold recovery of MUC1 mRNA expression was Quantitative methylation analysis. Quantitative methylation analysis of observed after 5-AzadC or 5-AzadC/TSA treatment of PANC1 cells, the MUC1 promoter was performed using the MassARRAY Compact system but there wasno effect of TSA alone. In MDA-MB-453 cells, (Hitachi High-TechnologiesCorporation; ref. 14). DNA from cell lineswas extracted using a DNeasy Tissue System (Qiagen). A 1-mg sample of DNA was treatment with 5-AzadC or 5-AzadC/TSA resulted in a 100-fold MUC1 converted with sodium bisulfite using an EZ DNA methylation kit increase in expression of mRNA compared with controls, (Zymo Research) and the modified DNA was amplified by PCR. The target whereasTSA alone gave a 3-fold restoration. Treatment of Caco2 regions were amplified using the primer pairs shown in Table 1. Each forward cells with 5-AzadC or 5-AzadC/TSA showed an increase of 10- to primer istagged with a 10 mer (5 ¶-AGG AAG AGA G-3¶) to balance the PCR 25-fold in MUC1 mRNA, whereasthe level remained constantafter and a reverse primer (5¶-CAG TAA TAC GAC TCA CTA TAG GGA GAA GGC T- treatment with TSA alone. Therefore, in these three cell lines, ¶ in vitro 3 ) with a T7-promoter tag for transcription. PCR amplification was treatment with 5-AzadC or 5-AzadC/TSA significantly restored the j performed with the following variables: hot start at 94 C for 15 min, followed MUC1 mRNA level, compared with treatment with TSA alone. In by denaturing at 94jC for 20 s, annealing at 56jC for 30 s, extension at contrast, LS174T cells showed restoration of MUC1 expression only 72jC for 1 min for 45 cycles, and final incubation at 72jC for 3 min. Unincorporated deoxynucleotide triphosphates were dephosphorylated by when treated with 5-AzadC/TSA. These results suggest that a adding 2 AL of premix including 0.3 unitsShrimp Alkaline Phosphate(SAP; combination of DNA methylation and histone modification may Sequenom). The reaction mixture wasincubated at 37j C for 40 min, and SAP affect expression of the MUC1 gene. wasthen heat inactivated for 5 min at 85 jC. After SAP treatment, 2 AL of the Next, to rule out the contribution of single nucleotide poly- PCR productswere usedasa template for in vitro transcription, and RNase A morphisms (SNP) in control of MUC1 expression, we sequenced a cleavage was used for the reverse reaction, following the manufacturer’s 2.8-kb promoter region of the human MUC1 gene from seven cell instructions (Sequenom). The samples were conditioned and spotted on a lines (Fig. 2). Sixty bases differed from the previously published 384-pad Spectro-CHIP (Sequenom) using a MassARRAY nanodispenser MUC1 promoter sequence (21), but all the sequences in the seven (Samsung), followed by spectral acquisition on a MassARRAY Analyzer cell lines were identical. Additionally, our sequence showed a 99% Compact MALDI-TOF MS (Sequenom). The resultant methylation calls were match with the sequences with Genbank accession numbers analyzed with EpiTyper software v1.0 (Sequenom) to generate quantitative NT_004487.18 and NW_925683.1. results for each CpG site or an aggregate of multiple CpG sites. MUC1 DNA extraction and DNA MSP analysis. DNA from cell lineswas Quantification of DNA methylation in the gene MUC1 extracted using a DNeasy Tissue System (Qiagen), according to the promoter. To map methylated cytosines in the promoter manufacturer’sinstructions.Bisulfitemodification of the genomic DNA in eight cancer cell lines, we used quantitative methylation analysis wascarried out usinga Epitect Bisulfitekit (Qiagen), and the modified DNA using the MassARRAY Compact system and mapped the efficacious www.aacrjournals.org 2709 Cancer Res 2008; 68: (8). April 15, 2008

Table 1. Synthetic oligonucleotides used in the study

Name Primer sequence Position

MUC1 promoter sequencing primers MUC1-S1L TTCAGGGATGAGAAGTTTTCAAG À2,769 to À2,747 MUC1-S1R GCAGTCTGAGTACTTAAAGAGCAAG À2,071 to À2,047 MUC1-S2L GAGTTGGGTTCACCTTGGAG À2,207 to À2,188 MUC1-S2R CCTGTATTCAGAAAACAGTTCTCAG À1,687 to À1,663 MUC1-S3L CACCTAGCTACTCGTGTTTTCTC À1,951 to À1,929 MUC1-S3R GGATTTACGGACACAGAACTATC À1,460 to À1,438 MUC1-S4L GTTCTTTCCAACCTGTGTCTTTG À1,713 to À1,691 MUC1-S4R GTCTGGAGGAAATGGAGCTAAG À992 to À971 MUC1-S5L GCATCTGCCTCTTAAGTACATAACG À1,112 to À1,088 MUC1-S5R TCTAGTTTTAGCCTCAGGGCAAC À304 to À282 MUC1-S6L CATTATCCAGCCCTCTTATTTCTC À457 to À434 MUC1-S6R ACTTCTCTACAGGACATTTGCTTG +331 to +354 10mer-tagged or T7-tagged primers MUC1-1L AGGAAGAGAGTGTAATTGTTTATTTTAGGGATTTTT À2,671 to À2,646 MUC1-1R CAGTAATACGACTCACTATAGGGAGAAGGCTAAATAAACCCAACTCCCCAAC À2,214 to À2,193 MUC1-2L AGGAAGAGAGTATTTTGGAGGTTAGAGAGATTTGG À2,197 to À2,173 MUC1-2R CAGTAATACGACTCACTATAGGGAGAAGGCTATTAAAAAAAACACTTTAAAAACACTTC À1,729 to À1,702 MUC1-3L AGGAAGAGAGGGGGTTGAGAATTGTTTTTTGAATATA À1,691 to À1,665 MUC1-3R CAGTAATACGACTCACTATAGGGAGAAGGCTCCAACCTAACCAAAATAATAAAACCTC À1,221 to À1,195 MUC1-4L AGGAAGAGAGTTGAATTTTTGATTTTGTGATTTATT À1,192 to À1,167 MUC1-4R CAGTAATACGACTCACTATAGGGAGAAGGCTTTACCACTCTCCACCACTAACCAA À722 to À699 MUC1-5L AGGAAGAGAGGGTAAGGAAGGATTTTAGGGTTTAT À703 to À679 MUC1-5R CAGTAATACGACTCACTATAGGGAGAAGGCTTTTTAACCTCAAAACAACAACTAAC À311 to À287 MUC1-6L* AGGAAGAGAGAGGGGGTAGAATAGATTTAGGTAGG +32 to +46 MUC1-6R* CAGTAATACGACTCACTATAGGGAGAAGGCTACTAAAACCCAAAAACCCCAAATT À287 to À264 MUC1-7L AGGAAGAGAGGGGTTTTGTTATTTGTTATTTGTTT À95 to À71 MUC1-7R CAGTAATACGACTCACTATAGGGAGAAGGCTTTCTTACCTTAACTTCATAATACCATCTAC +203 to +232 MSP primers c MSP-UL GGGGATTGGTATAAAGTGGTAGGT À37 to À13 MSP-UR AAAACAAAACAATTCAAACAAACA +20 to +44 b MSP-ML GATCGGTATAAAGCGGTAGGC À34 to À14 MSP-MR AAAACAAAACAAATTCAAACAAACG +20 to +44 ChIP primers ChIP-1L GTTTGCGGAGTCCCAGAAG À2,121 to À2,103 ChIP-1R GGAGAAAACACGAGTAGCTAGGTG À1,951 to À1,928 ChIP-2L AAGTACATAACGGCCCACACAG À1,099 to À1,078 ChIP-2R CGGGAGAGAAGGGTCACTG À886 to À868 ChIP-3L TTTGTCACCTGTCACCTGCTC À91 to À71 ChIP-3R CCCCTCACCTGTAAGCACTG +115 to +122

*MUC1-6 primers were designed for the antisense strand because PCR amplification was not obtained in the forward direction. cU indicatesthe primer for unmethylated alleles. bM indicatesthe primer for methylated alleles.

data (Fig. 3). Thisapproach hasbeen proven to be a powerful DNA were designed to target the 5¶ flanking region of the MUC1 methylation analysis tool (14). Near the transcriptional start site promoter (Fig. 4A; Table 1). Four CpG sites (Nos. 177, 178, 179, (fÀ100 to +100), a high level of CpG methylation wasobserved in and 181) were included in the MSP primer. These sites showed MUC1-negative cell lines and those with low MUC1 expression 0% to 4%, 2% to 21%, 1% to 17%, and 2% to 26% methylation, (PANC1, MDA-MB-453, and Caco2) compared with MUC1-positive respectively, in MUC1-positive cells, and 29% to 63%, 66% to cell lines(HPAFII, BxPC-3, MCF-7, and T-47D). However, LS174T 86%, 53% to 67%, and 79% to 93% methylation, respectively, in cells, which have low MUC1 expression, showed hypomethylation all MUC1-negative/low cellsexcept for LS174T cells.The LS174T of the MUC1 promoter. The methylation status of CpG sites 1 to cellsshowed1%, 5%, 4%, and 3% methylation at the respective 173 did not influence expression of the MUC1 gene. These results sites. To ensure the reliability of the MSP primer, MSP analysis indicate that hypomethylation of the 5¶ flanking region of the wasperformed on the eight cell lines(Fig. 4 B). An unmethylated promoter islinked to MUC1 expression, except in LS174T cells. band (lanesindicated by U;Fig.4B) wasclearly obtained in MSP analysis of the MUC1 gene promoter in eight cancer MUC1-positive HPAFII, BxPC-3, MCF-7, and T-47D cells, whereas cell lines. To confirm the MassARRAY results, primers for MSP a methylation band (lanesindicated by M;Fig.4B) wasobserved

Cancer Res 2008; 68: (8). April 15, 2008 2710 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 2008 American Association for Cancer Research. Epigenetic Control of MUC1 in PANC1, MDA-MB-453, and Caco2 cells. Overall, the MSP performed for three regionsusingantidimethyl H3-K9 and results were consistent with those from MassARRAY analysis and antiacetyl H3-K9 antibodies. Methylation of lysine 9 of histone showed the methylation status of the MUC1 promoter in the H3 (H3-meK9) facilitatesformation of heterochromatin, and eight cell lines. elevated levels of H3-meK9 at promoter sequences are associated Correlation between CpG methylation and histone H3-K9 with suppression of gene expression (22, 23). In MUC1-negative modification in the MUC1 promoter. Three ChIP primerswere or cells with low MUC1 expression (MDA-MB-453, Caco2, and designed to examine the relationship between DNA methylation LS174T), dimethylation of H3-K9 wasfound in all regions and histone modification. ChIP3 primer was designed to target (ChIP1–3; Fig. 4C). Acetylation of lysine 9 of histone H3 is as- regions similar to those targeted by the MSP primer (Fig. 4A; sociated with euchromatin formation, and acetylation of Table 1). To determine the profile of histone H3-K9 modification promoter-proximal histones is associated with gene expression in MUC1-positive and MUC1-negative/low cells, ChIP assays were (24). The acetylation level of H3-K9 in MUC1-positive cells

Figure 1. Expression of the MUC1 gene examined by quantitative RT-PCR and immunohistochemistry. A, quantitative RT-PCR results in eight cancer cell lines. Bars, gene expression levels relative to those in PANC1 cells. HPAFII, BxPC-3, MCF-7, and T-47D cells showed high expression of MUC1 mRNA, whereas PANC1, MDA-MB-453, Caco2, and LS174T cells had no or low expression. B, MUC1 immunoreactivity in eight cancer cell lines. MUC1 expression was consistent with the results of RT-PCR. C, quantitative RT-PCR results before and after treatment with 5-AzadC, TSA, and 5-AzadC/TSA in combination in cells with little or no MUC1 expression. After 5-AzadC and 5-AzadC/TSA treatment, PANC1, MDA-MB-453, and Caco2 cells showed significant restoration of MUC1 expression. In LS174T cells, MUC1 mRNAwas restored only when treated with 5-AzadC/TSA.

www.aacrjournals.org 2711 Cancer Res 2008; 68: (8). April 15, 2008

Figure 2. The human MUC1 gene promoter sequence, which spans positions À2,753 to +263 with respect to the transcription start site. The numbers of CpG sites and the transcriptional start site +1 (arrow) are shown.

Cancer Res 2008; 68: (8). April 15, 2008 2712 www.aacrjournals.org

(HPAFII, BxPC-3, MCF-7, and T-47D) washigh in the MUC1 can only analyze a restricted set of CpG sites in a target region. promoter (Fig. 4C). In PANC1 cells (low MUC1 expression) high Thus, we used the MassARRAY method to overcome this acetylation of H3-K9 wasobserved. Collectively, MSP and ChIP limitation. In eight cancer cell lines, methylation of CpG sites analysis revealed that DNA demethylation, histone H3-K9 174 to 182 wasinverselycorrelated with MUC1 gene expression, demethylation, and histone H3-K9 acetylation in the 5¶ flanking whereasmethylation of sites1 to 173 wasalmostunrelated to region of the MUC1 promoter may all be necessary for MUC1 gene expression (Fig. 3). Gonzalgo et al. (30) suggested that a gene expression. relatively low number of methylation errorsmay be sufficientto initiate a reduction of gene expression. Our results are in agreement with this hypothesis because we observed a high level Discussion of CpG methylation in MUC1-negative/low-expression cell lines Regulation of human MUC1 gene expression has been studied only in the vicinity of the transcriptional start site. To validate the intensively, and SP1, tumor necrosis factor, INF,and GATA3 have detailed methylation status in the 5¶ flanking region of the MUC1 been reported to activate MUC1 transcription (7, 25–27). However, promoter, we selected four CpG sites (177, 178, 179, and 181), at the methylation status of CpG islands existing in the MUC1 which methylation may affect MUC1 gene expression, and promoter hasyet to be elucidated, and our data revealed the designed a corresponding MSP primer. Our MSP results were detailsof DNA methylation and histoneH3-K9 modification on consistent with the MassARRAY data in all cancer cell lines. the full-length MUC1 promoter in MUC1-positive and MUC1- Interestingly, despite LS174T cells showing low expression of negative cancer cell linesfor the firsttime. Before beginning the MUC1, the MUC1 promoter was mostly unmethylated in these epigenetic analysis, we performed sequencing of the MUC1 cells. promoter in seven cancer cells to examine possible regulatory To examine the possibility of another form of epigenetic SNPs(rSNP). rSNPsare usuallypromoter region mutationsthat control for the MUC1 gene, we investigated histone modification cause variation in gene expression levels (28). Our results were in the eight cancer cell lines. Similarly to CpG methylation, it perfectly congruent in the seven cell lines and largely matched the has become increasingly evident that histone modification can data for Genbank accession numbers NT_004487.18 and contribute to gene regulation (31). To examine the relationship NW_925683.1, although some bases differed from the published between DNA methylation and histone modification in MUC1 MUC1 promoter sequence (21). We found no correlation between expression, a ChIP primer was designed to target regions similar rSNPs and MUC1 expression, and therefore, we examined the DNA to those targeted by the MSP primer (Fig. 4A). Nguyen et al. (22) methylation status of the MUC1 promoter in eight cancer cell showed that aberrantly silenced genes in cancer cells exhibit a lines. heterochromatic structure that is characterized by H3-K9 Methylation of cytosine residues at CpG dinucleotides is an hypermethylation. In contrast, histone H3 acetylated at lysine 9 important epigenetic change that islinked to transcriptional in a gene promoter region is associated with low nucleosome repression and regulation of chromatin structure (29). There are density near the transcription start site in human cells (32). several techniques for evaluation of CpG methylation, but most Based on these findings, we performed a ChIP assay using

Figure 3. Quantitative methylation analysis of CpG sites located in the MUC1 promoter using a MassARRAY Compact system. The efficacious data was mapped and different colors display relative methylation changes in 10% increments (green, 0%; red, 100% methylated). *, MUC1-positive cell lines. Of the cells with little or no MUC1 expression, PANC1, MDA-MB-453, and Caco2 cells showed high CpG methylation, but LS174T cells showed CpG hypomethylation in the MUC1 promoter. The methylation level in MUC1-positive HPAFII, BxPC-3, MCF-7, and T-47D cells was low. www.aacrjournals.org 2713 Cancer Res 2008; 68: (8). April 15, 2008

Figure 4. MSP and ChIP analyses in the MUC1 promoter region. A, schematic representation of the MUC1 gene promoter region, which spans positions À2,753 to +263 with respect to the transcription start site. There are 184 CpG sites, and the MSP and ChIP primers used in the study are indicated. B, MSP analysis of the MUC1 gene promoter region in eight cell lines. The PCR-products labeled M (methylated) were generated by methylation-specific primers, and those labeled U (unmethylated) were generated by primers specific for unmethylated DNA. The methylated allele was detected in PANC1, MDA-MB-453, and Caco2 cells, whereas the unmethylated allele was detected in HPAFII, BxPC-3, MCF-7, T-47D, and LS174T cells. *, MUC1-positive cell lines. C, ChIP analysis of histone H3-K9 modification in the MUC1 promoter region (ChIP regions 1–3) in eight cancer cell lines. Reactions were performed by PCR using input DNA. A negative control reaction using an aliquot precipitated with negative control normal mouse IgG antibody (Nc) is also shown. Unbound and bound fractions in each ChIP experiment are shown. In ChIP regions 1 to 3, histone H3-K9 dimethylation was observed in cells with little or no MUC1 expression, except for PANC1 cells, which showed high acetylation of histone H3-K9. Acetylation of histone H3-K9 was detected in all MUC1-positive cells. *, MUC1-positive cell lines.

Cancer Res 2008; 68: (8). April 15, 2008 2714 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 2008 American Association for Cancer Research. Epigenetic Control of MUC1 antidimethyl H3-K9 and antiacetyl H3-K9 antibodies. The 5-AzadC or a combination of 5-AzadC and TSA, but not TSA presence of dimethyl H3-K9 in MUC1-negative/low cells was alone, reactivates tumor suppressor gene expression at silenced confirmed in ChIP regions1 to 3, including in LS174T cells loci (e.g., p16). Our results in MDA-MB-453 and Caco2 cell lines (Fig. 4C). In LS174T cells, heterochromatin structure induced by are in good agreement with this observation because we showed H3-K9 methylation may down-regulate MUC1 expression. Histone that treatment with 5-AzadC or 5-AzadC/TSA significantly H3-K9 wasmore highly acetylated in MUC1-positivecellsthan in restored the MUC1 mRNA level, compared with TSA alone. MUC1-negative/low cells. Unexpectedly, PANC1 cells (low MUC1 However, treatment with TSA did not restore the MUC1 mRNA expression) also showed a high degree of H3-K9 acetylation in all level in PANC1 cells(Fig. 1 C), raising the possibility that ChIP primer regions. As contrasted with LS174T cells, methylated euchromatin formation remainsunchanged due to the acetylated CpGsin the 5 ¶ flanking region of the MUC1 promoter may be histone H3-K9 (Fig. 4C), and 5-AzadC did not restore the MUC1 related to suppression of MUC1 gene transcription in PANC1 mRNA level in LS174T cells(Fig. 1 C), which impliesthat 5-AzadC cells. alone hasno effect on an already hypomethylated MUC1 gene To investigate the combined role of DNA methylation and promoter (Fig. 3). histone modification in the regulation of MUC1 gene expression, From these results, we propose four patterns of epigenetic we also treated MUC1-negative or low-expression cancer cell lines control of MUC1 gene expression, as shown in Fig. 5. First, DNA with 5-AzadC and/or TSA (Fig. 1C). Kondo et al. (33) showed that demethylation, histone H3-K9 demethylation and histone H3-K9

Figure 5. Amodel depicting changes in chromatin modification and DNAmethylation of the MUC1 gene promoter region in eight cancer cell lines. Four mechanisms of epigenetic control of MUC1 expression are proposed. Patterns A and B indicate euchromatin states, whereas patterns C and D show heterochromatin formation. The core region is indicated by brackets. o, unmethylated CpG sites; ., methylated CpG sites. HPAFII, BxPC-3, MCF-7, and T-47D cells, in which MUC1 is overexpressed, follow pattern A. PANC1 and LS174T cells follow patterns B and C, respectively, and both express a low level of MUC1. MUC1-negative MDA-MB-453 and Caco2 cells follow pattern D.

www.aacrjournals.org 2715 Cancer Res 2008; 68: (8). April 15, 2008

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 2008 American Association for Cancer Research. Cancer Research acetylation in the 5¶ flanking region of the MUC1 promoter are epigenetic changesfor the MUC1 gene may be of importance for all necessary in cancer cells with high MUC1 expression (HPAF diagnosis of carcinogenic risk and prediction of outcomes of II, BxPC-3, MCF-7, and T-47D; Fig. 5A). Second, in cases of patients. histone H3-K9 demethylation and histone H3-K9 acetylation, MUC1 can be expressed slightly, even if CpG is methylated (PANC1; Fig. 5B). Third, with methylated histone H3-K9, MUC1 Acknowledgments expression is reduced regardless of CpG demethylation (LS174T; Received 12/26/2007; accepted 1/28/2008. Fig. 5C). Fourth, methylation of both DNA and histone H3-K9 Grant support: Scientific Research on Priority Areas 18014024 from the Ministry of Education, Science, Sports, Culture and Technology, Japan (S. Yonezawa) and the results in suppression of MUC1 expression in cells such as MDA- Kodama Memorial Foundation (N. Yamada). MB-453 and Caco2 (Fig. 5D). Taken together, these results show The costs of publication of this article were defrayed in part by the payment of page advertisement that neither DNA methylation nor histone modification alone charges. This article must therefore be hereby marked in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. fully determine the expression of MUC1, suggesting that We thank Yoshiko Arimura and Yukari Nishimura for their excellent technical expression is regulated by a combination of DNA methylation assistance, and Hiromichi Iai (Hitachi High-Technologies Corporation, Tokyo, Japan) ¶ and Katsuhiko Hashimitsu (Hitachi High-Tech Manufacturing & Service Corporation, and histone H3-K9 modification in the 5 flanking region of the Ibaraki, Japan) for their help with quantitative DNA methylation analysis. MUC1 promoter. An understanding of these intimately correlated Conflicts of interest: The authorshave no conflictsof interest.

References 12. Wolffe AP, JonesPL, Wade PA. DNA demethylation. 23. Mutskov V, Felsenfeld G. Silencing of transgene Proc Natl Acad Sci U S A 1999;96:5894–6. transcription precedes methylation of promoter DNA 1. Lan MS, Batra SK, Qi WN, Metzgar RS, Hollingsworth 13. Stirzaker C, Song JZ, Davidson B, Clark SJ. Transcrip- and histone H3 lysine 9. EMBO J 2004;23:138–49. MA. Cloning and sequencing of a human pancreatic tional gene silencing promotes DNA hypermethylation 24. Shi H, Wei SH, Leu YW, et al. Triple analysis of tumor mucin cDNA. J Biol Chem 1990;265:15294–9. through a sequential change in chromatin modifications the cancer epigenome: an integrated microarray 2. Patton S, Gendler SJ, Spicer AP. The epithelial mucin, in cancer cells. Cancer Res 2004;64:3871–7. system for assessing gene expression, DNA methyla- MUC1, of milk, mammary gland and other tissues. 14. Ehrich M, Nelson MR, Stanssens P, et al. Quantitative tion, and histone acetylation. Cancer Res 2003;63: Biochim BiophysActa 1995;1241:407–23. high-throughput analysis of DNA methylation patterns 2164–71. 3. Spicer AP, Rowse GJ, Lidner TK, Gendler SJ. Delayed by base-specific cleavage and mass spectrometry. Proc 25. Kovarik A, Lu PJ, Peat N, MorrisJ, Taylor-Papadimitriou mammary tumor progression in Muc-1 null mice. J Biol Natl Acad Sci U S A 2005;102:15785–90. J. Two GC boxes(Sp1 sites)are involved in regulation of Chem 1995;270:30093–101. 15. Wolffe AP, Matzke MA. Epigenetics: regulation the activity of the epithelium-specific MUC1 promoter. 4. Schroeder JA, Adriance MC, Thompson MC, Camenisch through repression. Science 1999;286:481–6. J Biol Chem 1996;271:18140–7. TD, Gendler SJ. MUC1 alters h-catenin-dependent tumor 16. Yamada N, Hamada T, Goto M, et al. MUC2 26. Kuwahara I, Lillehoj EP, Hisatsune A, et al. Neutrophil formation and promotescellular invasion.Oncogene expression is regulated by histone H3 modification elastase stimulates MUC1 gene expression through 2003;22:1324–32. and DNA methylation in pancreatic cancer. Int J Cancer increased Sp1 binding to the MUC1 promoter. Am J 5. Osako M, Yonezawa S, Siddiki B, et al. Immunohisto- 2006;119:1850–7. Physiol Lung Cell Mol Physiol 2005;289:L355–62. chemical study of mucin carbohydrates and core proteins 17. PetersAH, O’Carroll D, Scherthan H, et al. Lossofthe 27. Clark S, McGuckin MA, Hurst T, Ward BG. Effect of in human pancreatic tumors. Cancer 1993;71:2191–9. Suv39h histone methyltransferases impairs mammalian interferon-g and TNF-a on MUC1 mucin expression 6. Kondo M, Tamaoki J, Takeyama K, Nakata J, Nagai A. heterochromatin and genome stability. Cell 2001;107: in ovarian carcinoma cell lines. Dis Markers 1994;12: Interleukin-13 inducesgoblet cell differentiation in pri- 323–37. 43–50. mary cell culture from Guinea pig tracheal epithelium. 18. Eden S, Hashimshony T, Keshet I, Cedar H, Thorne 28. Knight JC. Regulatory polymorphisms underlying Am J Respir Cell Mol Biol 2002;27:536–41. AW. DNA methylation modelshistone acetylation. complex disease traits. J Mol Med 2005;83:97–109. 7. Abba MC, Nunez MI, Colussi AG, Croce MV, Segal-Eiras Nature 1998;394:842. 29. Holliday R, Pugh JE. DNA modification mechanisms A, Aldaz CM. GATA3 protein asa MUC1 transcriptional 19. Nan X, Ng HH, Johnson CA, et al. Transcriptional and gene activity during development. Science 1975;187: regulator in breast cancer cells. Breast Cancer Res repression by the methyl-CpG-binding protein MeCP2 226–32. 2006;8:R64. involves a histone deacetylase complex. Nature 1998;393: 30. Gonzalgo ML, Hayashida T, Bender CM, et al. The 8. Andrianifahanana M, Moniaux N, Batra SK. Regulation 386–9. role of DNA methylation in expression of the p19/p16 of mucin expression: mechanistic aspects and implica- 20. Yonezawa S, Sueyoshi K, Nomoto M, et al. MUC2 locusin human bladder cancer cell lines.Cancer Res tions for cancer and inflammatory diseases. Biochim gene expression is found in noninvasive tumors but not 1998;58:1245–52. BiophysActa 2006;1765:189–222. in invasive tumors of the pancreas and liver: its close 31. Goll MG, Bestor TH. Histone modification and 9. Zrihan-Licht S, Weiss M, Keydar I, Wreschner DH. relationship with prognosis of the patients. Hum Pathol replacement in chromatin activation. GenesDev 2002; DNA methylation status of the MUC1 gene coding for a 1997;28:344–52. 16:1739–42. breast-cancer-associated protein. Int J Cancer 1995;62: 21. Kovarik A, Peat N, Wilson D, Gendler SJ, Taylor- 32. Nishida H, Suzuki T, Kondo S, Miura H, Fujimura Y, 245–51. Papadimitriou J. Analysis of the tissue-specific promoter Hayashizaki Y. Histone H3 acetylated at lysine 9 in 10. Bird A. The essentials of DNA methylation. Cell 1992; of the MUC1 gene. J Biol Chem 1993;268:9917–26. promoter is associated with low nucleosome density in 70:5–8. 22. Nguyen CT, Weisenberger DJ, Velicescu M, et al. the vicinity of transcription start site in human cell. 11. Hamada T, Goto M, Tsutsumida H, et al. Mapping of Histone H3-lysine 9 methylation is associated with Chromosome Res 2006;14:203–11. the methylation pattern of the MUC2 promoter in aberrant gene silencing in cancer cells and is rapidly 33. Kondo Y, Shen L, Issa JP. Critical role of histone pancreatic cancer cell lines, using bisulfite genomic reversed by 5-aza-2’-deoxycytidine. Cancer Res 2002;62: methylation in tumor suppressor gene silencing in sequencing. Cancer Lett 2005;227:175–84. 6456–61. colorectal cancer. Mol Cell Biol 2003;23:206–15.

Cancer Res 2008; 68: (8). April 15, 2008 2716 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 2008 American Association for Cancer Research. MUC1 Expression Is Regulated by DNA Methylation and Histone H3 Lysine 9 Modification in Cancer Cells

Norishige Yamada, Yukari Nishida, Hideaki Tsutsumida, et al.

Cancer Res 2008;68:2708-2716.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/68/8/2708

Cited articles This article cites 32 articles, 15 of which you can access for free at: http://cancerres.aacrjournals.org/content/68/8/2708.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/68/8/2708.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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/68/8/2708. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.