MYC-Induced Epigenetic Activation of GATA4 in Lung Adenocarcinoma

Published OnlineFirst December 13, 2012; DOI: 10.1158/1541-7786.MCR-12-0414-T

Molecular Cancer Chromatin, Gene, and RNA Regulation Research

Ines^ C. Castro1, Achim Breiling2, Katharina Luetkenhaus1, Fatih Ceteci5, Simone Hausmann3, Sebastian Kress1, Frank Lyko2, Thomas Rudel1, and Ulf R. Rapp4

Abstract Human lung cancer is a disease with high incidence and accounts for most cancer-related deaths in both men and women. Metastasis is a common event in non–small cell lung carcinoma (NSCLC), diminishing the survival chance of the patients with this type of tumor. It has been shown that MYC is involved in the development of metastasis from NSCLC, but the mechanism underlying this switch remained to be identified. Here, we focus on GATA4 as a MYC target in the development of metastasis with origin in lung adenocarcinoma, the most common type of NSCLC. Epigenetic alterations at the GATA4 promoter level were observed after MYC expression in lung adenocarcinoma in vivo and in vitro. Such alterations include site-specific demethylation that accompanies the displacement of the MYC-associated zinc finger protein (MAZ) from the GATA4 promoter, which leads to GATA4 expression. Histone modification analysis of the GATA4 promoter revealed a switch from repressive histone marks to active histone marks after MYC binding, which corresponds to active GATA4 expression. Our results thus identify a novel epigenetic mechanism by which MYC activates GATA4 leading to metastasis in lung adenocarcinoma, suggesting novel potential targets for the development of antimetastatic therapy. Mol Cancer Res; 11(2); 161–72. 2012 AACR.

Introduction candidate tumor suppressor gene RASSF1A (3). However, Non–small cell lung carcinoma (NSCLC) is the most comparatively little is known about the role of promoter frequent type of lung cancer with high metastatic potential hypomethylation in gene activation in cancer, especially in fi NSCLC. and a low cure rate. Cancer has been classi ed as a genetic fi disease ever since the discovery of retroviral oncogenes and We previously identi ed MYC as a metastasis inducer in the finding of frequent mutations in their cellular counter- a mouse model for NSCLC and in the human NSCLC cell parts in human tumors. Growing evidence now suggests that line A549. On the basis of our observations in this mouse model, we suggested a lineage switch mechanism to be epigenetic alterations are at least as common as mutational fi events in the development of cancer (1, 2). Tumor-specific involved in progression to metastasis and identi ed GATA4 as a MYC-target in this context (4). Our earlier promoter hypermethylation is well documented (1). Epige- fi netic silencing is also known to be a frequent event in ndings of RAF/MYC cooperation in the murine hemato- poietic system revealed that MYC affects tumor cell NSCLC, that is, of p16, H-cadherin, death-associated pro- fi tein (DAP) kinase 1 (DAPK1), 14-3-3 sigma, and the reprogramming (5). Furthermore, ampli cations and rear- rangements of MYC genes were found in a fraction of human NSCLC (6). Authors' Affiliations: 1Biocenter, Department of Microbiology, University The GATA family of transcription factors consists of 6 of Wurzburg,€ Wurzburg;€ 2Division of Epigenetics, DKFZ-ZMBH Alliance, – German Cancer Research Center, Heidelberg; 3Department of General members, GATA1 6. These proteins have 2 highly con- Surgery, University Hospital Rechts der Isar, Technical University of served zinc-finger domains and recognize a consensus DNA- 4 Munich, Munich; Division Molecular Mechanisms in Lung Cancer, Max binding motif of (A/T)/GATA/(A/G; ref. 7). They are Planck Institute for Heart and Lung Research, Bad Nauheim, Germany; and 5Beatson Institute for Cancer Research, Glasgow, United Kingdom involved in the regulation of several biologic processes including organogenesis, differentiation, proliferation, and Note: Supplementary data for this article are available at Molecular Cancer – Research Online (http://mcr.aacrjournals.org/). apoptosis (7). While the GATA1 3 proteins are important for the development of the nervous system and hematopoi- I.C. Castro, A. Breiling, and K. Luetkenhaus contributed equally to this – work. etic cell lineages, GATA4 6 are involved in organogenesis, especially of the heart, ovary, and extraembryonic tissues. In Corresponding Author: Ulf R. Rapp, Division Molecular Mechanisms in Lung Cancer, Max Planck Institute for Heart and Lung Research, the murine lung, GATA6 drives the expression of the thyroid Parkstraße 1, Bad Nauheim 61231, Germany. Phone: 49-6032-705-420; transcription factor TTF-1, which is an upstream effector Fax: 49-6032-705-471; E-mail: [email protected] necessary for surfactant protein C (Sp-C) expression (8). doi: 10.1158/1541-7786.MCR-12-0414-T Furthermore, GATA6-Wnt signaling was shown to be 2012 American Association for Cancer Research. important for epithelial stem cell development and airway

www.aacrjournals.org 161

Castro et al.

regeneration (9). In contrast to GATA6, expression of tioned, Trichostatin A (TSA) (Sigma-Aldrich) was added GATA4 was detected neither in adult lung nor in lung to the culture medium at a concentration of 50 ng/mL. adenomas induced by oncogenic C-RAF. (4). GATA4 is, however, involved in the maintenance of the intestine of the Soft agar assay adult mouse. For the soft agar assay, the bottom agar was prepared using Here, we show that MYC induces epigenetic changes in an autoclaved stock of 5% Sea Plaque Agarose, which was the GATA4 promotor leading to its expression. Further- microwaved and mixed with DMEM to a final concentra- more, MYC-dependent GATA4 upregulation was accom- tion of 0.5% agarose. Five milliliter of 0.5% agarose were panied by site-specific demethylation and the acquisition of poured in a 60-mm dish. For the top agar, the 5% stock was active histone modification marks in its promoter. Our diluted to 0.6% agarose with DMEM and stored in a water results thus suggest a further autoregulation mechanism of bath at 40C. Dilutions of the cells were prepared in 1 mL of GATA4 upon MYC expression. medium (10,000 cells per 60-mm dish) and then mixed 1:1 with the 0.6% agarose. The mixture was poured on top of the fi fi Materials and Methods solidi ed bottom agar. After solidi cation of the top agar, the dishes were incubated in a wet chamber at 37C for 21 to 28 Molecular cloning days, when the colonies number was counted. The coding sequence of GATA4 was subcloned into the Eco RI site of the retroviral vector pEGZ/GFP/Neo by using Immunohistochemistry the oligonucleotides 50-CCGGAATTCCGGATGTATCA- 0 0 Immunohistochemistry was conducted on paraffin sec- GAGCTTGGCCATGG-3 and 5 -CCGGAATTCCGG- tions. After deparaffinization and rehydration, sections were TTACGCAGTGATTATGTCCCC-30. The resulting plas- fi boiled in 10 mmol/L citrate buffer for 10 to 20 minutes for mid was con rmed by sequencing. antigen retrieval. To quench the endogenous peroxidase activity, sections were incubated with methanol or PBS fi Cell culture and cell lines containing 1% to 3% H2O2. Nonspeci c antibody binding The A549 and Phoenix amphotrophic cell lines were was prevented by incubation with 5% of serum with 0.2% obtained from the American Type Culture Collection, Triton X-100 in PBS for 1 hour at room temperature. After where they are regularly verified by genotypic and phe- blocking, sections were incubated with the primary antibody notypic tests. All cell lines were cultured in Dulbecco's [GATA4, sc-1237, Santa Cruz (1:200); TTF-1, M3575, modified Eagle's medium (DMEM) supplemented with Dako (1:200); Pro Sp-C, gift from Jeffrey A. Whitsett from 10% fetal calf serum (FCS) and 1% penicillin/streptomy- the Division of Pulmonary Biology, Cincinnati Children's cin. Cells were maintained in a humidified incubator at Hospital Medical Center, Cincinnati, Ohio (1:5,000)] over- – 37 Cwith5%CO2 and harvested using 0.15% Trypsin night at 4 C. After washing, sections were incubated with EDTA. For viral transfection, the MYC-ER virus encod- the corresponding biotinylated secondary antibodies (Dako) ing a fusion of the murine estrogen receptor (ER) and at 1:200 to 1:600 for 1 hour at room temperature. For human c-MYC, GATA4 short hairpin RNA (shRNA), staining ABC reagent was applied (Vactastain Elite ABX Kit, and MYC-associated zinc finger protein (MAZ) shRNA- Vector Laboratories) and color was developed with 3,30- producing virus were independently produced in the diaminobenzidine (DAB). For counterstaining hematoxylin Phoenix amphotrophic packaging cell line. Ten micro- was used. After dehydration, the stainings were mounted grams of pBpuro c-MYC-ER (10), pLKO.1-puro- with enthellan. For immunofluorescence staining, the fol- shGATA4 (MISSION shRNA Plasmid DNA from Sig- lowing secondary antibodies (all obtained from Jackson ma-Aldrich; Clone ID:NM_002052.2-1592s1c1) and ImmunoResearch Laboratories, at 1:200 dilutions) were pLKO.1-puro-shMAZ (MISSION shRNA Plasmid DNA used: donkey anti-goat Cy5 and donkey anti-rabbit Cy3. from Sigma-Aldrich; Clone ID:NM_002383.1-761s1c1) were individually transfected into Phoenix amphotrophic Immunocytochemistry cells using Lipofectamine (Invitrogen) according to man- Immunocytochemistry was conducted on cells grown on ufacturer's instructions. After 24 hours, the medium was coverslips. After fixation using 4% PFA-PBS for 15 minutes changed. The viruses were produced overnight, and super- at room temperature, cells were washed with PBS. To natant was then harvested and used for infection of A549 prevent unspecific antibody binding, cells were blocked with or A549 J5-1 (4) cells in the presence of 8 mg/mL of 4% serum, 0.2% Triton X-100 in PBS for 1 hour at room polybrene. Twenty-four hours after infection, infected temperature. The primary antibody was diluted in 4% serum cells were selected with 250 mg/mL of puromycin. For in PBS and applied overnight at 4C [GATA4, sc-1237, viral infection of A549 cells with pEZG-GATA4, the Santa Cruz (1:250); GATA6, sc-9055, Santa Cruz (1:100)]. procedure described earlier was used. After infection, cells After washing with PBS, the cells were incubated with the were selected with 500 mg/mL of zeocin. One week after secondary antibody for 2 hours at room temperature. The selection, pools of 10 cells were seeded in a 96-well plate antibody was diluted in 4% serum in PBS (anti-goat-Alexa and for each pool GATA4 mRNA levels were measured. 647, 1:200; anti-rabbit-Cy3, 1:200). Before mounting with The pool number 11 showed the highest level of GATA4 Mowiol, cells were counterstained using 40,6-diamidino-2- mRNA and was used for further experiments. If men- phenylindole (DAPI).

162 Mol Cancer Res; 11(2) February 2013 Molecular Cancer Research

Epigenetic Switch Induced by MYC in Lung Adenocarcinoma

Luciferase reporter assay Bisulfite sequencing A 442-bp fragment upstream of the transcription start site For next generation, bisulfite-sequencing DNA from of Gata4 was amplified from genomic DNA from HeLa cells cultured cells was prepared using the DNeasy Blood & using the following primers: 50-AAAAAACTCGAGGGA- Tissue Kit (Qiagen) according to manufacturer's specifica- ACTAGCATCCAGCC-30 and 50-AAAAAAAAGCTTG- tions. DNA from paraffin section was prepared as described CTGCAGCGGCGACGAA-30. The fragment was sub- earlier. 250 ng up to 1 mg of DNA was bisulfite-treated with cloned into XhoI and HindIII sites of the pGL3 Luciferase the EpiTect Bisulfite Kit (Qiagen) according to manufac- Reporter-Basic Vector (Promega) by enzymatic reaction. turer's instructions. Deaminated DNA was amplified by A549 and A549 J5-1 cells were transfected independently PCR using the primer mentioned earlier but with the 454- with the pGL3 vector containing the GATA4 promoter and adapter sequence added and an identifier 4-base code for with the pGL3 control vector containing luciferase under each tissue sample (here referred to as XXXX): 454- control of a SV40 promoter. Caco-2 cells were transfected COGATA4-prom_up 50-GCCTCCCTCGCGCCATCA- with the same vectors and used as a positive control, as they GXXXXTAATAAAGTTGATTTTGGGTATTATAG-30- normally express GATA4 (11). To measure luciferase activ- 454-COGATA4-1_lo 50-GCCTTGCCAGCCCGCTCA- ity, ONE-Glo Luciferase Assay System was used and con- GXXXXCCCTACCTACTAAACCTAAAAATTC-30. ducted in 96-well plates according to the recommendations PCR conditions were as follows: 95C for 3 minutes fol- in the manual. The lumimometer was preheated to 37C lowed by 40 cycles at 95C for 30 seconds, annealing and the luminescence was measured with an integration time temperature for 40 seconds and 72C for 45 seconds, of 100 milliseconds. followed by a 3-minute incubation at 72C. PCR products were subsequently purified using the QIAquick Gel Extrac- Chromatin-immunoprecipitation tion Kit (Qiagen). For sequencing, equimolar amounts of all Cross-linked chromatin from A549 control cells or A549 amplicons were combined in a single tube Roche 454 FLX. J5-1 cells was prepared and immunoprecipitated as described Standard sequencing was provided by the Genomics and previously (12). Chromatin immunoprecipitation (ChIP)- Proteomics Core Facility of the DKFZ. grade antibodies specific for H3K27me3 (07-449), H3K4me2 (07-030), and H3K4me3 (07-473) were pur- Real-time PCR chased from Upstate. ChIP-grade antibodies specificfor RNA was isolated from cultured cells using TriFast H3K9me2 (ab1220) and H3K9me3 (ab8898) were obtained reagent (Peqlab). cDNA-synthesis was conducted using 1 fromAbcam.OtherantibodiesusedwereagainstMAZ mg of RNA per reaction with first strand cDNA synthesis kit (Abcam ab85725), EZH2 (Cell Signalling AC22), DNMT1 (Fermentas). Quantitative PCR (qPCR) was conducted in a (active motive 39204), DNMT3a (Abcam ab13888), 20 mL reaction using Finnzymes SYBER Green Master Mix DNMT3b (active motive 39207), POLII (4H8, Abcam (NEB) and Rox as a reference dye. The amplification fi C ab5408), GATA4 (Santa Cruz sc-1237), and GATA6 (Santa ef ciency was raised to the power of the threshold cycle ( t Cruz sc-9055). The chicken-v-MYC antibody was a gift from value) to calculate the relative amount of the transcript. This Klaus Bister from the Center for Molecular Biosciences, gives the number of cycles necessary for the product to be University of Innsbruck, Innsbruck, Austria. Immunopreci- detectable. The resulting value was normalized to the level of pitates were finally dissolved in 30 mL of Tris-EDTA (TE) the housekeeping gene. Assays were conducted in triplicates buffer (10 mmol/L Tris–HCL pH 8, 1 mmol/L EDTA). One in a StepOnePlus detection system (Applied Biosystems). microliter was analyzed by real-time PCR using a primer pair The primer sequences used were for Chicken-v-MYC 50- specific for the GATA4 promoter region (CHIPGATA4_up CGGCCTCTACCTGCACGACC-30 and 50-GACCAG- 50-CGGAGACCCCAGAGCCTG-30; CHIPGATA4_lo 50- CGGACTGTGGTGGG-30, for hsGATA4 50-CTACAT- CTCTCTACCTCCAGACAAGC-30)in10mL PCR reac- GGCCGACGTGGGAG-30 and 50-CTCGCCTCCAGA- tions, using the Absolute QPCR SYBR Green Mix (Themo GTGGGGTG-30, for hsGATA6 50-TCCATGGGGTG- Scientific) and a Roche LightCycler 480. PCR conditions: 1 CCTCGACCA-30 and 50-CCCTGAGGTGGTCGCTT- cycle: 95C 15 minutes; 42 cycles: 95C 15 seconds/ GTG, for hsmucin2 50-CGACTAACAACTTCGCCTC- 60C 40 seconds, read; melting curve 65Cto95C, read CG-30 and 50-CGCGGGAGTAGACTTTGGTG-30, for every 1C. Cycle threshold numbers for each amplification hsMYC 50-GCCCCTGGTGCTCCATGA-30 and 50-CA- were measured with the LightCycler 480 software and enrich- ACATCGATTTCTTCCTCATCTTCT-30, and for ments were calculated as percentage of the input. b-actin 50-GTCGTACCACAGGCATTGTGATGG-30 and 50-GCAATGCCTGGGTACATGGTGG-30. DNA-isolation from paraffin sections After GATA4 staining, sections were air-dried and Ethics statement GATA4-positive and -negative tumor regions were scratched All animal studies were approved by the Bavarian State from the slide using a scalpel. Material was collected in 30 mL authorities for animal experimentation. of lysis buffer (10 mmol/L Tris–HCl, 1 mmol/L EDTA, 1% w/v Tween 20, 100 mg/mL proteinaseK) and incubated Statistical analysis overnight at 37C. For inactivation of proteinaseK samples Results are expressed as the mean SEM. One-way were heated at 95C for 20 minutes. ANOVA was used to determine the differences among

www.aacrjournals.org Mol Cancer Res; 11(2) February 2013 163

Castro et al.

several groups, followed by Student–Newman–Keul test We analyzed a 389-bp long CpG-rich region, which using GraphPad InStat software. Student t test was used contains the promoter of the GATA4 gene (Supplementary for 2 group comparisons. For all tests, statistical significance Fig. S1A) by next generation bisulfite sequencing, using was considered to be at P < 0.005. genomic DNA from GATA4-positive and -negative regions of a Sp-C-c-MYC/KRas–mutant lung tumor and a GATA4- positive liver metastasis (Supplementary Fig. S1B). We did Results not observe global methylation changes for the examined GATA4 expression is accompanied by GATA4 promoter region but significantly reduced methylation levels for the demethylation in c-MYC/KRas-mutant type II CpG-sites 13 to 22 (with the exception of CpGs 15, 19, and pneumocytes 20—see Supplementary Table S1 for statistical tests) in the Previously, we observed a GATA6 to GATA4 switch in GATA4-positive tumor and metastasis material (Fig. 1B). Sp-C-c-MYC and Sp-C-CRAF-BxB/Sp-C-c-MYC lung Thus, this region showed clear hypomethylation in GATA4- tumors and liver metastasis (4). In this work, we observed expressing tissues. low or absent expression of pro Sp-C in GATA4-positive lung tumor regions from mice with the same genotype MYC is necessary to induce the expression of GATA4 and (Fig. 1A), which suggests that GATA4-positive tumor its target gene mucin2 in human lung adenocarcinoma cells might become independent of any oncogenes under cells Sp-C promoter, including c-MYC, by an epigenetic mech- If MYC expression is a prerequisite for induction of anism. We have therefore surmised an epigenetic mech- GATA4 expression in lung adenocarcinoma, the introduc- anism for the underlying mechanism of GATA4-induction of MYC into a GATA4-negative human lung adeno- tion in lung tumor. carcinoma cell line should lead to the expression of GATA4

A #22620 (Sp-C-c-MYC/KRas*)

GATA4 Sp-C Sp-C / GATA4 Red = areas of high GATA4 staining Red = areas of low Sp-C staining Green = areas of low GATA4 staining Green = areas of high Sp-C staining

B #16740 (Sp-C-c-MYC/KRas*) #16740 (Sp-C-c-MYC/KRas*) #22620 (Sp-C-c-MYC/KRas*) GATA4–neg. lung tumor GATA4–pos. lung tumor GATA4–pos. liver metastasis

1 5 9 13 16 19 22 26 30 33 1 5 9 13 16 19 22 26 30 33 1 5 9 13 16 19 22 26 30 33 66% 44% 44%

Figure 1. Ectopic expression of GATA4 in Sp-C-c-MYC/Ras–mutant lung adenocarcinoma is accompanied by site-specific GATA4 promoter demethylation. A, staining of mouse lung tumor sections for Sp-C and GATA4 shows mutually exclusive expression patterns between them. Lung tumor serial sections were stained for Sp-C and GATA4 (left and middle). GATA4-high/Sp-C–negative and GATA4-low/Sp-C–positive tumor regions were encircled with red and green dashed lines, respectively. Right, double immunofluorescence staining of a mouse lung tumor section for Sp-C and GATA4. Note that the majority of GATA4-positive tumor cells (red) are negative for Sp-C (green). DAPI (blue) shows nuclei. Scale bar, 50 mm. B, next generation bisulfite sequencing of the CpG-rich region near the GATA4 promoter (for exact primer locations see Supplementary Fig. S1A). DNA was isolated from lung tumor regions stained negatively and positively for GATA4 and from GATA4-positive liver metastasis, respectively. Sequencing results are shown as heatmaps in which each row represents one sequence read. Individual red boxes indicate methylated and green boxes indicate unmethylated CpG dinucleotides. Sequencing gaps are shown in white. CpGs showing the highest degree of change between DNA from GATA4-positive and -negative tumor tissue are boxed. The overall methylation of the region is 76% for both GATA4-positive and -negative lung tumors and 73% for the liver metastasis. A significant decrease in methylationoftheCpGs13-22 (boxed,66% inGATA4-negativetissues)wasobservedinGATA4-positivetissue(44% inGATA4-positive tissues).KRas,KRas-mutant.ForstatisticaltestsseeSupplementaryTableS1.TheDNAusedforthenextgenerationsequencingwasisolatedfromGATA4-stained paraffin sections as illustrated in Supplementary Fig. S1B.

164 Mol Cancer Res; 11(2) February 2013 Molecular Cancer Research

Epigenetic Switch Induced by MYC in Lung Adenocarcinoma

A P < 0.001 P < 0.01 P < 0.01 P < 0.01 0.6 1.0 0.025 0.005 *** ** ** **

0.8 0.020 0.004 0.4 0.6 0.015 0.003

0.4 0.010 0.002 0.2 0.2 0.005 0.001 Relative hs mucin 2 mRNA levels Relative GG MYC mRNA levels Relative hs GATA4 mRNA levels 0.0 Relative hs GATA6 mRNA levels 0.0 0.000 0.000 A549 A549 J5-1 A549 A549 J5-1 A549 A549 J5-1 A549 A549 J5-1

B A549 MYC-ER C

0.10 A549 MYC-ER + OHT P < 0.01 P < 0.01 ** A549 A549 J5-1 ** 0.08 P < 0.05 0.06 * P < 0.001 0.04 *** GATA4

0.02

Relative hs MYC mRNA levels 0.00 Week 1 Week 2 Week 3 Week 4 0.0020 P < 0.01 **

0.0015

P < 0.01 0.0010 **

ns ns 0.0005

Relative hs GATA4 mRNA levels 0.0000 Week 1 Week 2 Week 3 Week 4 GATA6 merge

0.005 P < 0.01 ** 0.004

0.003 P < 0.05 *

0.002 merge

0.001 ns ns

Relative hs mucin2 mRNA levels 0.000 Week 1 Week 2 Week 3 Week 4

Figure 2. MYC expression in A549 cells leads to upregulation of GATA4 and its target, mucin2. A, qPCR analysis of A549 cells infected with chicken-v-MYC–expressing J5 virus shows highly significant expression of exogenous MYC. The expression of MYC in the human lung adenocarcinoma cell line A549 leads to a highly significant upregulation of GATA4 and its target gene, mucin2, on mRNA level. In parallel, GATA6 was also significantly upregulated. Experiments were carried out on the basis of 3 independent RNA isolations. B, mRNA levels of human MYC, GATA4, and mucin2 in induced and not induced A549 MYC-ER cells over time. GATA4 and its target, mucin2, are upregulated after 3 weeks of MYC induction in A549 MYC-ER cells, although MYC upregulation is detected as early as 1 week after induction. C, consistent with the RNA expression, immunocytochemistry shows an induction of GATA4 expression on the protein level in the A549 J5-1 cells. Green (Cy3) colored cells are GATA4-positive; red (Alexa 647) colored cells are GATA6-positive; blue, DAPI. Scale bars, 50 mm. All error bars represent the SD of the mean. Statistical differences between groups are indicated. and its target gene mucin2 (13). To test this possibility, we in Fig. 2A, the mRNA expression of GATA4 and its target conducted qPCR and immunofluorescence staining on the gene mucin2 was significantly upregulated in MYC-expres- chicken-v-MYC–expressing A549 J5-1 cells (4). As shown sing A549 cells.

www.aacrjournals.org Mol Cancer Res; 11(2) February 2013 165

Castro et al.

Moreover, using an inducible cell line carrying a fusion of Epigenetic signature of the GATA4 promoter MYC with the ER, which is dependent on tamoxifen To complement our analysis of epigenetic modifications in presence to be located to the nucleus, we could show GATA4 the GATA4 promoter, we mapped histone modifications by and mucin2 mRNA upregulation after 3 weeks of MYC ChIP in the same region. Chromatin was prepared from A549 induction, although MYC induction occurs as early as 1 control and A549 J5-1 cells and precipitated with antibodies week after 4-hydroxytamoxifen (OHT) addition (Fig. 2B). against histone H3 dimethylated at lysine 9 (H3K9me2), This suggests that epigenetic changes at the GATA4 pro- histone H3 trimethylated at lysine 9 (H3K9me3), histone moter take effect at this time point. H3 dimethylated at lysine 4 (H3K4me2), histone H3 In contrast to our in vivo data, GATA6 was also signif- trimethylated at lysine 4 (H3K4me3), and histone H3 icantly upregulated in the A549 cells upon MYC-expression trimethylated at lysine 27 (H3K27me3). Immunoprecipi- (Fig. 2A). This suggests that other factors might superimpose tated DNA was analyzed by qPCR using a primer pair positive regulation. Furthermore, we conducted immuno- specific for the GATA4 promoter region. This PCR frag- fluorescence staining for GATA4 and GATA6 and were able ment of 372 bp spans a region starting with CpG 8 and to detect both in the A549 J5-1 cells (Fig. 2C). extending about 200 bp into the GATA4 gene body. In agreement with the strong methylation of the promoter in A549 control cells, we found mainly repressive histone GATA4 is necessary for MYC-induced anchorage- marks in the GATA4 promoter region (H3K9me3 and independent growth in lung adenocarcinoma cells H3K27me3), most prominently H3K27me3, which would The ability of cancer cells to form colonies without indicate repression by Polycomb group proteins (Fig. 5A). In anchorage support is a key aspect of the tumor phenotype, contrast, in the MYC-expressing line J5-1, the levels of particularly with respect to metastatic potential (14). repressive marks were low, whereas both active H3K4 marks A549 J5-1 cells showed an increased ability to form were enriched, indicating ongoing transcription at the colonies in soft agar in comparison with the parental cell GATA4 promoter. These results suggest a significant change line (Fig. 3A and B). The inducible cell line A549 MYC- in the epigenetic makeup of the GATA4 promoter upon ER showed the same increased ability in the presence of MYC expression, which is characterized by local DNA tamoxifen (Fig. 3A and B). The knockdown of GATA4 in hypomethylation and a switch from repressive to active A549 J5-1 (Fig. 3C) restored the behavior of wild-type histone marks. A549 cells about anchorage-independent growth ability (Fig. 3D and E). These data suggest that GATA4 might be Protein occupancy in the GATA4 promoter region necessary for the transformation phenotype of MYC- To identify proteins interacting with the GATA4 promoter expressing A549 cells. On the basis of these data, a newly in A549 control and A549 J5-1 cells, we analyzed a fragment prepared A549 cell line overexpressing GATA4 (Fig. 3F of the GATA4 promoter region corresponding to the segment and G) was tested for the same feature, which showed that amplified by the bisulfite primers with the MatInspector GATA4 alone is able to mimic MYC-induced ability to software (15). This revealed the presence of a MYC-related E- grow anchorage independently in A549 cells (Fig. 3H and box, the potential binding site for the MAZ, which is usually I), strengthening the evidence that GATA4 is necessary for shared by the SP1 transcription factor (Supplementary Fig. MYC-induced metastasis. While growing in adherent S1C). In addition, we identified a Y-box and consensus culture, the A549 GATA4-11 cells did not show a dif- sequences for the general transcription factor TFIID between ference in proliferation when compared with A549 cells CpGs 15 and 22. To validate these potential binding sites and (data not shown). to gain insight into the protein presence at the GATA4 promoter in control and MYC expressing cells, we conducted MYC is able to induce GATA4 promoter demethylation ChIP using antibodies against chicken-v-MYC (chk-MYC), in human lung adenocarcinoma cells GATA4, GATA6, MAZ, the 3 major DNA methyltrans- To test whether ectopic MYC expression is sufficient to ferases (DNMT1, DNMT3a, and DNMT3b), the histone induce demethylation of the GATA4 promoter in a human H3 lysine 27 methyltransferase Enhancer of Zeste 2 (EZH2) lung adenocarcinoma cell line, we conducted next gener- and the large subunit of the RNA-polymerase II (POL II) in ation bisulfite sequencing of the same region analyzed chromatin prepared from A549 control cells and the J5-1 cell earlier in A549 and A549 J5-1 cells. Also, in A549 cells, line. Immunoprecipitated DNA was analyzed by qPCR using the GATA4 promoter region was strongly methylated, the same primer pair as for the analysis of histone modifica- which did not change significantly upon MYC expression tions. In the repressed case (A549 control), we found strong (72% and 71%, respectively). Nevertheless, as shown binding of MAZ and EZH2 in the GATA4 promoter, which in Fig. 4A and B, we observed a significant reduction in was paralleled by the presence of DNMT1 (Fig. 5B). In A549 methylation at CpG sites 19 to 22 (with the exception of J5-1, MAZ and EZH2 binding was greatly reduced, whereas site 21—see Supplementary Table S2 for statistical tests) chk-MYC (the protein expressed in this cell line) and in these cell lines. These results thus confirm the data GATA4 itself were found enriched on the GATA4 promoter. obtained in mouse lung tumors and strongly suggest that While DNMT3a and DNMT3b were detectable only at ectopic MYC expression is sufficient to induce local background levels, DNMT1 showed a significant interaction demethylationoftheGATA4 promoter. with the GATA4 promoter region (Fig. 5B). DNMT1 levels

166 Mol Cancer Res; 11(2) February 2013 Molecular Cancer Research

Epigenetic Switch Induced by MYC in Lung Adenocarcinoma

A549 A549 EV + OHT A549 J5-1 A549 MYCER + OHT B C D P < 0.0001 ns *** P P < 0.0001 2 P < 0.0001 < 0.05 2 60 *** * 0.08 100 *** P < 0.001 P < 0.01 *** ** 80 0.06 40 60 0.04 40 20 mRNA levels mRNA 0.02 20 Relative hsGATA4 Number of colonies/78.5 mm of colonies/78.5 Number mm of colonies/78.5 Number 0 0.00 0

A549 A549 A549 A549 +OHT +OHT A549 EV +OHT +OHT A549 J5-1 A549 J5-1 A549 J5-1 A549 J5-1 A549 EV A549 J5-1 A549 J5-1 A549 MYCER A549 MYCER shGATA4 26 shGATA4 26 E F 0.004 P < 0.0001 *** 0.003

0.002

mRNA levels 0.001 Relative hs GATA4 A549 A549 J5-1 A549 J5-1 shGATA4 26 0.000

A549 G GFP GATA4 H A549 I GATA4-11 P 2 25 < 0.0001 *** 20 A549 A549 15

Number of colonies/78.5 mm of colonies/78.5 Number 0

A549 A549

GATA4-11 A549 GATA4 - 11 A549 GATA4 - 11

Figure 3. GATA4 is necessary for anchorage-independent growth induced by MYC. A, anchorage-independent growth of A549 MYC-ER cells in the presence of OHT. Colonies formed by A549 and A549 empty vector (EV) in the presence of OHT (controls), and A549 J5-1 and A549 MYC-ER in the presence of OHT, after 3 weeks in soft agar are shown. B, quantification of colonies formed by the different cell lines (conditions as indicated). A549 MYC-ER cells show, in the presence of OHT, a comparable ability to form colonies as A549 J-51 cells, indicating the induction of anchorage-independent growth by MYC. C, mRNA levels of human GATA4 in A549, A549 J5-1, and A549 J5-1 shGATA4 26. The shRNA-mediated knockdown of GATA4 in A549 J5-1 cells restored the low values of GATA4 mRNA levels of A549 wild-type cells. D, anchorage-independent growth of A549 J5-1 cells after knockdown of GATA4. Quantification of colonies formed by the 3 different cell lines indicated. The soft agar assay with the A549 J5-1 cells shows a highly significant reduction in the number of colonies after infection with the shRNA26-expressing virus. E, colonies formed by A549, A549 J5-1, and A549 J-51 shGATA4 26 cells after 3 weeks in soft agar are shown. F, A549 cells infected with the pEGZ/GATA4 vector show GATA4 expression. mRNA levels of GATA4 and EGFP in A549 cells infected with the plasmid pEGZ/GATA4 are shown. Pools of 10 cells were seeded in a 96-well plate after infection and selected with zeocin. The best candidate was pool number 11, showing the highest level of GATA4 mRNA levels and used for further experiments. G, immunocytochemistry for human GATA4 in A549 and A549 GATA4–11 cells. GATA4 and GFP expression was detected in the nucleus of A549 GATA4–11 cells. Red (Cy5) colored cells are GATA4-positive; green cells are GFP-positive; scale bars, 50 mm. H, growth in soft agar resulted in an increase of number and size of colonies in the case of A549 GATA 4–11 cells compared with the parental cell line A549. I, quantification of the colonies formed by A549 and A549 GATA4–11 showed a significant increase of the colonies formed in GATA4-expressing cells. All error bars represent the SD of the mean. Statistical differences between groups are indicated. ns, not significant.

www.aacrjournals.org Mol Cancer Res; 11(2) February 2013 167

Castro et al.

A A549 B A549 J5-1

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 59% 49% 60% 37%

Figure 4. Site-specific demethylation of CpG dinucleotides in the promoter region of GATA4 upon MYC expression in A549 cells. Next generation bisulfite sequencing of the GATA4 promoter in A549 (A) and A549 J5-1 (B) cells (for exact primer locations see Supplementary Fig. S1A). Sequencing results are shown as heatmaps in which each row represents one sequence read. Individual red boxes indicate methylated and green boxes indicate unmethylated CpG dinucleotides. Sequencing gaps are shown in white. There is no significant change in the overall methylation status of the region (72% in A549 and 71% in A549 J5-1), but significant demethylation of the CpGs 19 to 22 (right box) was observed (49% in A549 vs. 37% in A549 J5-1). For statistical tests see Supplementary Table S2.

did not differ significantly between the 2 cell lines, which are leads to specific DNA-hypomethylation of the region con- in line with the observation that the overall methylation of taining the MAZ-binding site and the E-box due to the the region did not change in MYC-expressing cells. GATA6, absence of DNMT1 recruitment. The recruitment of as GATA4 antagonist, could not be found interacting either GATA4 to its own promoter suggests some further auto- in control or in J5-1 cells. In addition to the observed regulation of this transcription factor (Fig. 6). activation of the GATA4 promoter in J5-1 cells, we found increased interaction of POL II with the promoter (Fig. 5B), Discussion which is again consistent with increased transcriptional A mutually exclusive expression of GATA4/GATA6 trig- activity. gered by MYC was previously reported both in lung tumors and corresponding metastasis from animals expressing c- MAZ plays a role in GATA4 induction by MYC MYC in type II cells under the control of the Sp-C promoter, MAZ has been described as a recruiting factor of DNMTs which suggests that this lineage switch occurring in the and histone deacetylases (HDAC; ref. 16). As we found primary tumor is important for the development of c- MAZ to be displaced from the GATA4 promoter after MYC MYC–induced metastasis (4). expression in A549 cells, we predicted that the knockdown In this work, we stained lung tumor tissues from animals of MAZ in A549 cells could also lead to GATA4 upregula- with the same genotype and observed that GATA4-positive tion. MAZ absence might result in the blocking of DNMTs regions show low pro Sp-C expression. This can represent and/or HDACs recruitment to GATA4 promoter leading to loss of differentiation or reprogramming and suggests that GATA4 transcription activation and mimicking the trans- these regions are presumably poor for any transgene expres- formation phenotype induced by GATA4. Therefore, we sion that is under Sp-C promoter, such as c-MYC (Fig. 1A). used a shRNA-mediated knockdown of MAZ in A549 cells These assumptions led us to the investigation of the epige- and which were afterward seeded in soft agar to test if the lack netic mechanism behind MYC-induced GATA4 expression of MAZ in A549 is able to induce anchorage-independent suggested in this work. growth. Depletion of MAZ was sufficient to induce anchorage-independent growth in A549 cells in a significant MYC drives GATA4 expression in human lung manner (Fig. 5C and D), suggesting that the knockdown adenocarcinoma cells of MAZ might be sufficient for the phenotype transforma- The upregulation of GATA4 and its functional target tion of A549 cells. Moreover, as postulated earlier, MAZ- mucin2 was also detected in MYC-expressing A549 cells depleted A549 cells also showed an upregulation of GATA4 (Fig. 2A and B) and occurred 3 weeks after ectopic MYC and its target mucin2 (Fig. 5E). expression (Fig. 2B), suggesting that a multistep mechanism In conclusion, our findings suggest a scenario in which might be involved in the changes induced by this oncogene. GATA4 is usually kept repressed by MAZ-, DNMT1-, and Accompanying GATA4 upregulation, MYC-expressing Polycomb group–containing complexes. Upon MYC A549 cells showed a metastatic behavior (14), in particular expression, MYC interacts with its binding site (E-box) near increased ability to grow anchorage independently in soft the GATA4 promoter, displacing MAZ and EZH2. This agar, which was reverted after GATA4 knockdown (Fig. 3A,

168 Mol Cancer Res; 11(2) February 2013 Molecular Cancer Research

Epigenetic Switch Induced by MYC in Lung Adenocarcinoma

A B 14 0.7 A549 A549 12 A549 J5-10.6 0.6 A549 J5-1

10 0.5

8 0.4

6 0.3 % Input % Input

4 0.2

2 0.1

0 0.0

MAZ EZH2 POLII K9me2 K9me3 K4me2 K4me3 Control GATA4 GATA6 DNMT1 Control K27me3 GG MYC DNMT3aDNMT3b C D E P < 0.0001 P P < 0.05 40 < 0.0001 0.0005 0.003 *** *** * 0.0004 30

2 0.002

A549 0.0003 20 0.0002 78.5 mm 0.001 mRNA levels mRNA levels mRNA 10 Relative hs GATA4 hs Relative Relative hs mucin2 hs Relative

Number of colonies/ 0.0001

0 0.0000 0.000

A549 A549 A549 A549 shMAZ 345 shMAZ A549 A549 shMAZ 345 A549 shMAZ 345 A549 shMAZ 345

Figure 5. MYC and GATA4 bind to the GATA4 promoter upon expression of exogenous MYC in the A549 cells, whereas MAZ is displaced. A, ChIP analysis of the GATA4 promoter region in control (A549) and MYC-expressing cells (A549 J5-1). ChIP was repeated at least 3 times with chromatin from biologic replicates using antisera specific for H3K9me2, H3K9me3, H3K4me2, H3K4me3, and H3K27me3. Immunoprecipitated DNA was analyzed by real-time PCR and a primer specific for the GATA4 promoter region. Enrichments are shown as percentage of the total input. Error bars show the SD of the mean. B, ChIP assay as in A monitoring occupancy of the GATA4-promoter by candidate proteins chicken-v-MYC (chk-MYC), GATA4, GATA6, MAZ, the 3 major DNA methyltransferases (DNMT1, DNMT3a, and DNMT3b), EZH2, and the large subunit of the POL II. Binding of chicken-v-MYC and GATA4 on the GATA4 promoter was significantly higher in the MYC-expressing A549 J5-1 cells than in the parental cell line. Error bars show the SD of the mean. C, anchorage-independent growth of A549 cells after knockdown of MAZ. MAZ knockdown in A549 cells leads to GATA4 activation and increased anchorage-independent growth. Pictures from the colonies formed by A549 and A549 shMAZ 345 cells after 3 weeks in soft agar are shown. D, quantification of the colonies formed by A549 cells after infection with shRNA-producing virus 354. The soft agar assay plate with the A549 shMAZ 345 cells showed a highly significant increase in the number of colonies formed, when compared with the parental cell line A549. E, GATA4 and mucin2 mRNA levels were measured in A549 and A549 shMAZ 345 cells. Results showed a significant increase in both GATA4 and mucin2 mRNA levels in MAZ-depleted cells, compared with the parental cell line A549. Error bars show the SD of the mean. Statistical differences between groups as indicated.

D, E, H, and I). These data indicate the direct involvement MYC induces GATA4 promoter demethylation of GATA4 in MYC-induced metastasis. In fact, GATA4 is a Growing evidence now suggests that epigenetic alterations lineage selector gene, and its upregulation suggests dediffer- at the promoter level are at least as common as mutational entiation and loss of organ identity, which is in line with the events in the development of cancer (1, 2). Tumor-speciﬁc theory that the metastatic process is a recapitulation of promoter hypermethylation is well documented in NSCLC ontogeny (5). GATA4 promotes the expression of the (1), but comparatively little is known about gene activation antiapoptotic factor Bcl2 and cyclin D2 (17), increases by promoter hypomethylation. In the present work, we show proliferation and migration capacities in colorectal cancer that GATA4 upregulation in lung adenocarcinoma is (18), and its altered expression is correlated with a broad accompanied by site-speciﬁc demethylation of the CpG range of tumors emerging from gastrointestinal tract, lungs, islands 13 to 22 of its own promoter upon MYC-expression ovaries, brain, and with poor prognosis in ovarian granulose in vitro, in primary tumors and in corresponding liver cell tumors (19). metastasis (Figs. 1B, 4A and B). Interestingly, this region

www.aacrjournals.org Mol Cancer Res; 11(2) February 2013 169

Castro et al.

ABMethylation GATA4 gene

GATA4 promoter MAZ DNMT1 DNMT1

MAZ MYC CpG CpG CpG CCpGpG CCpGpG CpG CpG CpG CpG CpG CpG CpG CpGCpG CpG 13 14 15 16 17 18 19 13 14 15 16 17 18 19

Replication CD MYC CpG CpG CpG CpG CpG CpGCpG CpG CpG CpG CpG CpG CpG CpG CpG 13 14 15 16 17 18 19 13 14 15 16 17 18 19 GATA 4

GATA 4 GATA 4

Figure 6. Locked Switch Model of GATA4 expression induced by MYC. A, under physiologic conditions, MAZ binds to the GATA4 promoter. MAZ recruits DNMT1, which in turn methylates the CpG-sites. B, when MYC expression increases above a given threshold, MYC binds to the GATA4 promoter displacing MAZ. As a consequence, DNMT1 is not recruited to this DNA region any more. C, after replication CpG sites remain unmethylated because of the absence of DNMT1. D, activated transcription of GATA4 might induce its autoregulation by interacting with its own promoter.

contains an E-box, a potential binding site for MYC (20, 21), oncogene as a key-player in epigenetic modifications occur- which was previously described as a key regulatory element ring at GATA4 promoter level. of GATA4 expression in vitro (22) and in vivo (23). A Y-box sequence is located directly downstream. The Y-box–bind- MYC induces the enrichment of active histone marks at ing protein 1 regulates the expression of many important the GATA4 promoter genes (24) and is involved in the development of metastasis DNA methylation acts in cooperation with histone tail in patients with gastric and breast cancer (25). This protein modifications (28). In the present work, corresponding to may bind to the hypomethylated region of the GATA4 the strong methylation of the GATA4 promoter of the A549 promoter, and thus regulate its expression, or even act as control, mainly repressive histone marks H3K9me3 (1), and a partner in metastasis development. most prominently trimethylation of lysine 27 of histone H3 Although controversial, hypermethylation of GATA4 in (29) were found in control cells (Fig. 5A). The enrichment lung cancer has been reported (18, 26). A study from Guo of H3K27me3 indicates repression by the Polycomb-repres- and colleagues on the methylation of GATA genes in lung sive complex 2 (PRC2; ref. 30), which methylates H3K27 cancer used several lung cancer cell lines revealing GATA4 via its catalytic subunit—EZH2 (29, 30), which was indeed silencing by hypermethylation in most of the tested cells, found to be enriched at the GATA4 promoter of A549 wild- including A549 (26). Curiously, the only cells extracted type cells, in contrast to the low binding of EZH2 detected from a metastatic tissue—H157—showed hypomethylation at the GATA4 promoter of MYC-expressing A549 cells. of their GATA4 promoter (26). Montavon and colleagues This supports our finding that GATA4 expression is showed that the GATA4 promoter was kept unmethylated in repressed in A549 cells, and that this event is accompanied all tissues collected from patients with a rapidly invasive by an enrichment of the methylated target of PRC2, ovarian cancer (high-grade serous ovarian cancer), whereas H3K27. tumors from patients with less invasive ovarian cancers In contrast to the enrichment of repressive histone marks showed hypermethylation of the GATA4 promoter (27). at the GATA4 promoter of wild-type A549 cells, in the This supports our observations that hypomethylation- MYC-expressing cell line J5-1, the levels of repressive marks dependent upregulation of GATA4 is involved in metastasis are low, whereas both active H3K4 marks (1) are enriched, development. Because MYC preferentially binds to promot- indicating ongoing transcription at the GATA4 promoter er regions and CpG-rich regions (21), our data highlight this (Fig. 5A). The recruitment of POL II—the key rate-limiting

170 Mol Cancer Res; 11(2) February 2013 Molecular Cancer Research

Epigenetic Switch Induced by MYC in Lung Adenocarcinoma

step in gene activation (31)—cooccurs with the enrichment MYC interacts with its binding site (E-box) near the GATA4 of the active H3K4 marks (32) and was observed in MYC- promoter, which leads to MAZ and EZH2 displacement, expressing A549 cells. Altogether, these data show significant and therefore MAZ-recruitment of DNMT1 to the MAZ- changes in the epigenetic landscape of the GATA4 promoter binding site and the E-box is blocked. Demethylation of upon MYC expression leading to the transcriptional activa- these regions occurs as a consequence of replication pro- tion of GATA4. ceeding in the absence of maintenance methylation, provided by DNMT1 (34), leading to subsequent GATA4 MYC induces changes in protein occupancy at GATA4 activation. Activated transcription of GATA4 might induce promoter its autoregulation by interacting with its own promoter (Fig. Broad changes in protein occupancy at the GATA4 6). We propose a novel molecular mechanism for GATA4 promoter of A549 cells upon MYC expression were also activation in lung adenocarcinoma, which is dependent on observed. Among other proteins tested, GATA4 markedly epigenetic changes on its own promoter and leads to met- bound to its own promoter upon MYC expression. On the astatic conversion. Moreover, we were able to identify other hand, MAZ (16) and the already mentioned EZH2 GATA4 as a MYC target in malignant lung adenocarcinoma, protein were displaced from the GATA4 promoter in MYC- and highlighted GATA4 as a potential target for antimeta- expressing A549 cells (Fig. 5B). Judging from the specific static therapy. and very restricted hypomethylation observed in MYC- positive tumors or MYC-expressing A549, most likely also Disclosure of Potential Conflicts of Interest DNMT1 is displaced from or looses access to the region No potential conflicts of interest were disclosed. encompassing CpGs 13 to 22. Nevertheless, the resolution of the ChIP-method used (ca. 400 bp) is not suitable to resolve Authors' Contributions such restricted changes in DNMT1 levels. MAZ displace- Conception and design: I.C. Castro, A. Breiling, K. Luetkenhaus, F. Lyko, U.R. ment from this region nevertheless suggests that it has a role as Rapp Development of methodology: I.C. Castro, K. Luetkenhaus, U.R. Rapp DNMT1 recruiter, which may be a key event for GATA4 Acquisition of data (provided animals, acquired and managed patients, provided repression in wild-type A549 cells. MAZ displacement might facilities, etc.): I.C. Castro, A. Breiling, K. Luetkenhaus, S. Hausmann, S. Kress prevent maintenance methylation due to the absence of Analysis and interpretation of data (e.g., statistical analysis, biostatistics, compu- tational analysis): I.C. Castro, A. Breiling, K. Luetkenhaus, F. Ceteci, S. Kress, U.R. DNMT1-recruitment, and therefore provoke very site-spe- Rapp cific hypomethylation at the GATA4 promoter. Although the Writing, review, and/or revision of the manuscript: I.C. Castro, A. Breiling, K. knockdown of MAZ in A549 cells did not lead to a very Luetkenhaus, F. Lyko, T. Rudel, U.R. Rapp Administrative, technical, or material support (i.e., reporting or organizing data, pronounced GATA4 upregulation, it led to a dramatic constructing databases): I.C. Castro, K. Luetkenhaus, F. Ceteci increase in mucin2 mRNA levels and in the ability of these Study supervision: I.C. Castro, K. Luetkenhaus, T. Rudel, U.R. Rapp cells to form colonies anchorage independently. These data suggest that moderate changes in GATA4 levels can result in Acknowledgments strong upstream effects and highlight MAZ as a key-player in The authors thank Jeffrey A. Whitsett for pro Sp-C antibody, Roland Halter for Sp- – C-c-MYC mice, Klaus Bister for chicken-v-MYC antibody, and Daniel J. Murphy for MYC-induced metastasis via GATA4 activation (Fig. 5C E). the MYC-ER construct. The authors also thank Tanja Musch and Francesca Tuorto Indeed, a previous study from Bartoszewski and colleagues for supporting bisulfite sequencing. showed that the repression of the gene coding for the cystic fi CFTR brosis transmembrane conductance regulator ( )is Grant Support maintained during endoplasmatic reticulum stress by recruit- This study was supported by Deutsche-Krebshilfe Foundation (grant-109081); ment of methyl-binding proteins and/or HDACs. The German Center for Lung Research (DZL); grants GRK1411/1, BMBF–01GU0709, CFTR and FORPROTECT (to I.C. Castro); and by the grant GRK1411/1 (to K. repressed state of depends on the binding of MAZ to Luetkenhaus). a hypermethylated region of the promoter, showing methyl- The costs of publication of this article were defrayed in part by the payment of page ation-specific repression by MAZ in this case as well (33). charges. This article must therefore be hereby marked advertisement in accordance with Our findings suggest a scenario in which GATA4 is 18 U.S.C. Section 1734 solely to indicate this fact. usually kept repressed by MAZ, DNMT1, and Polycomb Received July 3, 2012; revised November 13, 2012; accepted November 30, 2012; group containing complexes. Upon MYC expression, chk- published OnlineFirst December 13, 2012.

References 1. Jones PA, Baylin SB. The fundamental role of epigenetic events in 5. Rapp UR. Metastasis as a faulty recapitulation of ontogeny. Nature cancer. Nat Rev Genet 2002;3:415–28. Precedings 2007;258:83–136 2. Rapp UR, Ceteci F, Schreck R. Oncogene-induced plasticity and 6. Yokota J, Wada M, Yoshida T, Noguchi M, Terasaki T, Shimosato Y, cancer stem cells. Cell Cycle 2008;7:45–51. et al. Heterogeneity of lung cancer cells with respect to the ampliﬁ- 3. Risch A, Plass C. Lung cancer epigenetics and genetics. Int J Cancer cation and rearrangement of myc family oncogenes. Oncogene 1988; 2008;123:1–7. 2:607–11. 4. Rapp UR, Korn C, Ceteci F, Karreman C, Luetkenhaus K, SeraﬁnV, 7. Lowry JA, Atchley WR. Molecular evolution of the GATA family of et al. Myc is a metastasis gene for non–small-cell lung cancer. PLoS transcription factors: conservation within the DNA-binding domain. ONE 2009;4:e6029. J Mol Evol 2000;50:103–15.

www.aacrjournals.org Mol Cancer Res; 11(2) February 2013 171

Castro et al.

8. Liu C, Glasser SW, Wan H, Whitsett JA. GATA-6 and thyroid tran- 22. Mazaud Guittot S, Tetu A, Legault E, Pilon N, Silversides DW, Viger RS. scription factor-1 directly interact and regulate surfactant protein-C The proximal Gata4 promoter directs reporter gene expression to gene expression. J Biol Chem 2002;277:4519–25. sertoli cells during mouse gonadal development. Biol Reprod 2007;76: 9. Zhang Y, Goss AM, Cohen ED, Kadzik R, Lepore JJ, Muthukumar- 85–95. aswamy K, et al. A Gata6-Wnt pathway required for epithelial stem cell 23. Boulende Sab A, Bouchard M-F, Beland M, Prud'homme B, Souch- development and airway regeneration. Nat Genet 2008;40:862–70. kova O, Viger RS, et al. An ebox element in the proximal gata4 promoter 10. Littlewood TD, Hancock DC, Danielian PS, Parker MG, Evan GI. A is required for gata4 expression in vivo. PLoS One 2011;6:e29038. modified oestrogen receptor ligand-binding domain as an improved 24. Lyabin DN, Eliseeva IA, Skabkina OV, Ovchinnikov LP. Interplay switch for the regulation of heterologous proteins. Nucleic Acids Res between Y-box-binding protein 1 (YB-1) and poly(A) binding protein 1995;23:1686–90. (PABP) in specific regulation of YB-1 mRNA translation. RNA Biol 2011; 11. Fang R, Olds LC, Santiago NA, Sibley E. GATA family transcription 8:883–92. factors activate lactase gene promoter in intestinal Caco-2 cells. Am 25. Wu Y, Yamada S, Izumi H, Li Z, Shimajiri S, Wang K-Y, et al. Strong YB- J Physiol Gastrointest Liver Physiol 2001;280:G58–67. 1 expression is associated with liver metastasis progression and 12. Musch T, Oz Y, Lyko F, Breiling A. Nucleoside drugs induce cellular predicts shorter disease-free survival in advanced gastric cancer. differentiation by caspase-dependent degradation of stem cell factors. J Surg Oncol 2012;105:724–30. PLoS ONE 2010;5:e10726. 26. Guo M, Akiyama Y, House MG, Hooker CM, Heath E, Gabrielson E, 13. van der Sluis M, Melis MHM, Jonckheere N, Ducourouble M-P, Buller€ et al. Hypermethylation of the GATA genes in lung cancer. Clin Cancer HA, Renes I, et al. The murine Muc2 mucin gene is transcriptionally Res 2004;10:7917–24. regulated by the zinc-finger GATA-4 transcription factor in intestinal 27. Montavon C, Gloss BS, Warton K, Barton CA, Statham AL, Scurry JP, cells. Biochem. Biophys Res Commun 2004;325:952–60. et al. Prognostic and diagnostic significance of DNA methylation 14. Mori S, Chang JT, Andrechek ER, Matsumura N, Baba T, Yao G, et al. patterns in high grade serous ovarian cancer. Gynecol Oncol 2012; Anchorage-independent cell growth signature identifies tumors with 124:582–588. metastatic potential. Oncogene 2009;28:2796–805. 28. Esteller M. Cancer epigenomics: DNA methylomes and histone-mod- 15. Werner T. Computer-assisted analysis of transcription control regions. ification maps. Nat Rev Genet 2007;8:286–98. Matinspector and other programs. Methods Mol Biol 2000;132: 29. Cao R, Wang L, Wang H, Xia L, Erdjument-Bromage H, Tempst P, et al. 337–49. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. 16. Song J, Ugai H, Kanazawa I, Sun K, Yokoyama KK. Independent Science 2002;298:1039–43. repression of a GC-rich housekeeping gene by Sp1 and MAZ involves 30. Kuzmichev A, Nishioka K, Erdjument-Bromage H, Tempst P, Reinberg the same cis-elements. J Biol Chem 2001;276:19897–904. D. Histone methyltransferase activity associated with a human multi- 17. Kyronlahti€ A, Ram€ o€ M, Tamminen M, Unkila-Kallio L, Butzow R, protein complex containing the Enhancer of Zeste protein. Genes Dev Leminen A, et al. GATA-4 regulates Bcl-2 expression in ovarian 2002;16:2893–905. granulosa cell tumors. Endocrinology 2008;149:5635–42. 31. Ptashne M. Regulation of transcription: from lambda to eukaryotes. 18. Zheng R, Blobel GA. GATA transcription factors and cancer. Genes Trends Biochem Sci 2005;30:275–9. Cancer 2010;1:1178–88. 32. Chen Y, Jørgensen M, Kolde R, Zhao X, Parker B, Valen E, et al. 19. Anttonen M, Unkila-Kallio L, Leminen A, Butzow R, Heikinheimo M. Prediction of RNA Polymerase II recruitment, elongation and stalling High GATA-4 expression associates with aggressive behavior, where- from histone modification data. BMC Genomics 2011;12:544. as low anti-Mullerian€ hormone expression associates with growth 33. Bartoszewski R, Rab A, Twitty G, Stevenson L, Fortenberry J, Pio- potential of ovarian granulosa cell tumors. J Clin Endocrinol Metab trowski A, et al. The mechanism of cystic fibrosis transmembrane 2005;90:6529–35. conductance regulator transcriptional repression during the unfolded 20. Walhout AJ, Gubbels JM, Bernards R, van der Vliet PC, Timmers HT. c- protein response. J Biol Chem 2008;283:12154–65. Myc/Max heterodimers bind cooperatively to the E-box sequences 34. Hajkova P, Erhardt S, Lane N, Haaf T, El-Maarri O, Reik W, et al. located in the first intron of the rat ornithine decarboxylase (ODC) gene. Epigenetic reprogramming in mouse primordial germ cells. Mech Dev Nucleic Acids Res 1997;25:1493–501. 2002;117:15–23. 21. Zeller KI, Zhao X, Lee CWH, Chiu KP, Yao F, Yustein JT, et al. Global 35. Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM, mapping of c-Myc binding sites and target gene networks in human B et al. The human genome browser at UCSC. Genome Res 2002;12: cells. Proc Natl Acad Sci U S A 2006;103:17834–9. 996–1006.

172 Mol Cancer Res; 11(2) February 2013 Molecular Cancer Research

MYC-Induced Epigenetic Activation of GATA4 in Lung Adenocarcinoma

Inês C. Castro, Achim Breiling, Katharina Luetkenhaus, et al.

Mol Cancer Res 2013;11:161-172. Published OnlineFirst December 13, 2012.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-12-0414-T

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2012/12/13/1541-7786.MCR-12-0414-T.DC1

Cited articles This article cites 35 articles, 8 of which you can access for free at: http://mcr.aacrjournals.org/content/11/2/161.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/11/2/161.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://mcr.aacrjournals.org/content/11/2/161. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.