DNMT3B Overexpression by Deregulation of FOXO3a-Mediated Transcription Repression and MDM2 Overexpression in Lung Cancer

Original Article

DNMT3B Overexpression by Deregulation of FOXO3aMediated Transcription Repression and MDM2 Overexpression in Lung Cancer Yi-Chieh Yang, MS,* Yen-An Tang, PhD,† Jiunn-Min Shieh, MD,‡ Ruo-Kai Lin, PhD,§ Han-Shui Hsu, MD, PhD,‖ and Yi-Ching Wang, PhD*†

Introduction: DNA methyltransferase 3B (DNMT3B) contributes to de novo DNA methylation and its overexpression promotes tumorigenesis. However, whether DNMT3B is upregulated by transcriptional deregulation remains unclear. Methods: We studied the transcriptional repression of DNMT3B by forkhead O transcription factor 3a (FOXO3a) in lung cancer cell, animal, and clinical models. Results: The results of luciferase reporter assay showed that FOXO3a negatively regulated DNMT3B promoter activity by preferentially interacting with the binding element FOXO3a-E (+166 to +173) of DNMT3B promoter. Ectopically overexpressed FOXO3a or combined treatment with doxorubicin to induce FOXO3a nuclear accumulation further bound at the distal site, FOXO3a-P (−249 to −242) by chromatin-immunoprecipitation assay. Knockdown of FOXO3a resulted in an open chromatin structure and high DNMT3B mRNA and protein expression. Abundant FOXO3a repressed DNMT3B promoter by establishing a repressed chromatin structure. Note that FOXO3a is a degradation substrate of MDM2 E3-ligase. Cotreatment with doxorubicin and MDM2 inhibitor, Nutlin-3, further enforced abundant nuclear accumulation of FOXO3a resulting in decrease expression of DNMT3B leading to synergistic inhibition of tumor growth and decrease of methylation status on tumor suppressor genes in xenograft specimens. Clinically, lung cancer patients with DNMT3B high, FOXO3a low, and MDM2 high expression profile *Department of Pharmacology, College of Medicine, National Cheng Kung University, Tainan, Taiwan; †Institute of Basic Medical Science, College of Medicine, National Cheng Kung University, Tainan, Taiwan; ‡Division of Chest Medicine, Department of Internal Medicine, Chi Mei Medical Center, Tainan; The Center of General Education, Chia Nan University of Pharmacy & Science, Tainan, Taiwan; §Graduate Institute of Pharmacognosy, Taipei Medical University, Taipei, Taiwan; and ‖Division of Thoracic Surgery, Taipei Veterans General Hospital; Institute of Emergency and Critical Care Medicine, National Yang-Ming University School of Medicine, Taipei, Taiwan. Yi-Chieh Yang, Yen-An Tang, and Jiunn-Min Shieh contributed equally to this work. This work was supported in part by grants CMNCKU10308 from the Chi Mei Medical Center, Taiwan (J.M.S.) and NSC102-2627-B-006-010 from the National Science Council, Taiwan (Y.C.W.) and DOH101-TD-I111-TM001 from the Department of Health, Taiwan (Y.C.W.). Disclosure: The authors declare no conflicts of interest. Address for correspondence: Yi-Ching Wang, PhD, Department of Pharmacology, National Cheng Kung University, No. 1, University Road, Tainan 70101, Taiwan. E-mail: [email protected] Copyright © 2014 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/14/0909-1305

correlated with poor prognosis examined by immunohistochemistry and Kaplan-Meier survival analysis. Conclusions: We reveal a new mechanism that FOXO3a transcriptionally represses DNMT3B expression and this regulation can be attenuated by MDM2 overexpression in human lung cancer model. Cotreatment with doxorubicin and Nutlin-3 is a novel therapeutic strategy through epigenetic modulation. Key Words: DNMT3B, FOXO3a, MDM2, Doxorubicin, Nutlin-3, Prognosis, Lung cancer (J Thorac Oncol. 2014;9: 1305–1315)

I

n the past 10 years, aberrant DNA methylation in cancer cells has been widely studied. The cancer genome shows genomewide hypomethylation and accompanies by hypermethylation of CpG islands in the tumor suppressor genes (TSGs).1,2 The promoter hypermethylation of CpG islands in TSGs is associated with transcriptional silencing and appears to promote tumor formation and progression.2 DNA methylatransferases (DNMTs) are the enzymes that methylate C5 position of cytosine bases in CpG dinucleitides, including five members (DNMT1, 2, 3A, 3B, and 3L) distinct by their different functions.3,4 Among them, DNMT3 family generates the initial methylation pattern de novo.5,6 DNMT3B knockout mice show global demethylation of DNA and die during embryogenesis or during the first week after birth.5 Loss of DNMT3B activity in human leads to immunodeficiency, centromere instability, and facial anomalies syndrome, a genetic disorder accompanied by low methylation in pericentromeric satellite regions.7 DNMT3B overexpression either at mRNA or protein level has been shown in several human cancer cells, including lung, hepatoma, colorectal, retinoblastoma, and breast cancers.8–12 Some oncogenic pathways such as PI3K/AKT and estrogen receptor signaling have been reported to be involved in stimulating the DNMT3B expression.13,14 However, the underlying mechanisms involved in DNMT3B transcriptional regulation remains to be elucidated. Recent study shows that DNMTs transcript levels negatively correlated with forkhead O transcription factor 3a (FOXO3a) in oocyte development, implicating that FOXO3a may transcriptionally repress the DNMT genes.15 FOXO3a belongs to a large protein family of transcriptional regulators characterized by a conserved DNA-binding domain termed

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

1305

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

Yang et al.

the “forkhead-box”.16,17 To date, many reports indicate a tumor suppressor role of FOXO3a. For example, ectopic overexpression of FOXO3a significantly impairs tumor growth in cell and xenograft models in breast cancer and promotes apoptosis in leukemia and prostate cancer cells.18,19 In addition, repression of FOXO3a transcriptional activity in cancer cells results in promotion of angiogenesis and tumor progression.20 FOXO3a has been shown to transcriptionally upregulate apoptotic-related genes such as p27, Bim, and Fas ligand.18,21,22 In contrast, FOXO3a could transcriptionally repress miR21, which suppresses the expression of Fas ligand.23 FOXO3a is also a transcriptional repressor of Skp2 gene expression to inhibit Skp2 SCF E3 ligase toward p27 ubiquitination and degradation.24 Of note, gene deletion of FOXO3a is found in early-stage lung adenocarcinoma of smokers and tobacco carcinogen-initiated lung tumors in mice.25,26 Restored FOXO3a in FOXO3a-deficient lung cancer cells increases the cell apoptotic response to nicotine-derived nitrosamino ketone-mediated DNA damage.27 The last-mentioned studies implicate that loss of FOXO3a may contribute to lung cancer pathogenesis. Many activated oncogenic kinase pathways in human cancers, including AKT, ERK, and IKKβ had been reported to target FOXO3a regulation.28–30 These oncogenic kinases could phosphorylate FOXO3a at different sites retaining it in the cytoplasm localization and subsequently leading to its degradation. AKT-dependent phosphorylation of human FOXO3a at Thr31, Ser253, and Ser315 promotes FOXO3a interaction with 14-3-3, the scaffold protein within the cytoplasm, and elicits the export of FOXO3a from the nucleus into the cytoplasm. On the other hand, ERK phosphorylates FOXO3a at Ser294, Ser344, and Ser425 thereby enhancing the interaction with MDM2 through conformational change and results in promoting degradation of FOXO3a through the ubiquitination-proteasome-mediated pathway. It is worth mentioning that doxorubicin, a topoisomerase II inhibitor, had been reported to induce FOXO3a translocation into the nucleus and activate FOXO3a to regulate the expression of downstream genes.23,31 Based on the observation of an inverse expression of DNMT and FOXO3a and frequent loss of FOXO3a in cancer cells, the present study aims to elucidate the mechanism of transcriptional repression of DNMT3B by FOXO3a in lung cancer cell, xenograft, and patient models. Given that FOXO3a the potent antitumor growth activity, we proposed to also investigate whether promotion of FOXO3a nuclear accumulation using the drugs doxorubicin and Nutlin-3, an MDM2 inhibitor32 could reactivate the DNMT3B transcriptional regulation by FOXO3a thereby reversing the DNA methylation status of TSGs.

MATERIALS AND METHODS Cell Culture Lung cancer cell lines A549 and CL1-5 were cultured in Dulbecco’s Modified Eagle’s Medium (Invitrogen, Carlsbad, CA) and incubated in a humidified incubator containing 5% CO2. All cells were supplemented with 10% fetal bovine serum (Invitrogen) and 1% penicillin/streptomycin (100 units/ ml penicillin and 100 μg/ml streptomycin, Invitrogen). A549 cell line was purchased from the American Type Culture Collection (Manassas, VA). CL1-5 cell line was obtained

1306

from Dr. Pan-Chyr Yang (Department of Internal Medicine, Medical College, National Taiwan University, Taiwan).

Treatment with Doxorubicin and Nutlin-3 CL1-5 and A549 cells were cultured overnight before exposure to 2 μM doxorubicin (Sigma-Aldrich, MO) for 4 hours to induce nuclear localization of FOXO3a. A549 cells were treated with 5 μM Nutlin-3 (Sigma-Aldrich) for 16 hours to inhibit MDM2-mediated FOXO3a degradation. In combined treatment, Nutlin-3 was added first and doxorubicin was added at the final 4 hours.

Plasmid, RNAi, Transfection and Promoter Activity Assay The expression vectors and interference RNA (RNAi) used in the present study are shown in SDC 1 (http://links. lww.com/JTO/A616) and SDC 2 (http://links.lww.com/ JTO/A617). Transfection of RNAi was carried out with Lipofectamine 2000 (Invitrogen) according to the manufacture’s protocol. Transfection of pGL4-derivatives promoter construct vectors, pGFP, and pGFP-FOXO3a was carried out using Turbofect (MBI Fermentas, Flamborough, Ontario, Canada) according to the manufacture’s protocol. CL1-5 cells and A549 cells were transfected with siFOXO3a for 24 hours or FOXO3a expression vector for 18 hours. Then 2 μg/well of pGL4 vector, pGL4-full length (−483 to +315), pGL4-P region (−483 to +50), or pGL4-E region (+31 to +315) of DNMT3B promoter was transfected along with 1 μg/well of pGL4-renilla construct (as internal control) and incubated for another 24 hours. Dual luciferase reporter assay system (Promega Corp., Madison, WI) was used to perform the luciferase assay according to the manufacturer’s instruction. The promoter constructs were generated by polymerase chain reaction (PCR) cloning and the sequences of the primers used are listed in SDC 3 (http://links.lww.com/ JTO/A618).

RNA Extraction and Quantitative ReverseTranscriptase PCR (RT-qPCR) Assay Total RNA were extracted using Trizol reagent (Invitrogen). Total RNA (4 μg) was reverse transcribed into cDNA using SuperScript reverse tanscriptase (Invitrogen) and amplified using SYBR Green PCR Master Mix (Roche, Palo Alto, CA) and specific primers for each candidate genes. The primer sequences and the annealing temperature are listed in SDC 3 (http://links.lww.com/JTO/A618). The relative amount of amplified gene products was calculated using the relative standard curve method and normalized to β-actin as detected in the same run.

Quantitative MethylationSpecific PCR (qMSP) Assay Sodium bisulfite modification of genomic DNA was performed using EZ DNA Methylation-Gold kit (Zymo Research, Irvine, CA) according to the manufacturer’s protocols. The qMSP assay was conducted using FastStart Universal SYBR Green Master Mix (Roche) and StepOnePlus Real-Time PCR

Copyright © 2014 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

System (Applied Biosystems, Foster City, CA). The total 20 μl reaction mix included 1 μl bisulfite-treated DNA (50 ng), 0.5 μl forward methylation-specific primer (10 μM), 0.5 μl reverse methylation-specific primer (10 μM), 10 μl SYBR Green Master Mix, and 8 μl ddH2O. The qPCR conditions were initiated with a denaturing step at 95°C for 10 minutes, followed by 45 cycles at 95°C for 20 seconds, 65°C for 20 seconds, and 72°C for 20 second. The fluorescence intensity was obtained and analyzed by StepOnePlus System. DNA of normal lung cell IMR90 and fully methylated IMR90 DNA treated with SssI enzyme served as control for unmethylated and methylated alleles, respectively. We generated serial dilution (100%, 75%, 50%, 25%, and 0% DNA methylation) of standards using methylated DNA (SssI-treated IMR90 DNA) and unmethylated DNA (IMR90 DNA), and performed qMSP analyses using BLU, RASSF1A, SLIT2, and β-actin genes using specific primers as listed in Supplemental Digital Content 3 (http://links.lww.com/JTO/A6180). For the internal reference gene β-actin, the primers were designed to amplify and detect a region devoid of CpG nucleotides. Thus, amplification of β-actin by qMSP occurs independently of its methylation status. The qPCR analysis for methylation status of BLU, RASSF1A, and SLIT2 genes were normalized to β-actin, and the relative methylation ratio was identified by 2−△△CT determination method.

Chromatin-Immunoprecipitation (ChIP) and Western Blot Assays Chromatin DNA were immunoprecipitated at 4°C for 16 to 18 hours using anti-FOXO3a, anti-HDAC3, anti-trimethylation of histone 3 lysine 27 (H3K27me3), anti-acetylation of histone 3 lysine 9 and 14 (H3Ace), and normal IgG (as negative control). Immunoblotting was performed using antibodies against FOXO3a, DNMT3B, MDM2, or HDAC3. β-actin was an loading control. Antibodies and conditions are listed in SDC 4 (http://links.lww.com/JTO/A619). The primer sequences and the annealing temperature to amplify the ChIPDNA are listed in SDC 3 (http://links.lww.com/JTO/A618).

Immunofluorescence Staining and Confocal Microscopic Analysis Cells were cultured overnight and treated with or without 2 μM doxorubicin at indicated times. Cells were then incubated with primary antibody and followed by incubation with DAPI and fluorescent secondary antibody for 1 hour and images were obtained by an OLYMPUS FV1000 confocal microscope (Olympus America Inc., Melville, NY) to detect localization of FOXO3a. Antibodies and conditions are listed in SDC 4 (http://links.lww.com/JTO/A6190).

Immunohistochemistry Assay

Paraffin blocks of tumors were sectioned into 5 μm slices and processed using standard deparaffinization and rehydration techniques. Antibodies against FOXO3a, DNMT3B, and MDM2 were used and conditions for each are listed in SDC 4 (http://links.lww.com/JTO/A6190). The sections were then sequentially incubated with biotinylated

FOXO3a Transcriptionally Represses DNMT3B

secondary antibody for 40 minutes followed by streptavidinhorseradish peroxidase for 20 minutes. 3, 3′-diaminobenzidine tetrahydrochloride (DAKO LSAB kit K0675, Dako, Glostrup, Denmark) was used as the chromagen. The evaluation of the immunohistochemistry staining was conducted blindly, without prior knowledge of the clinical and pathologic characteristics of the cases or treatments given to the mice. The surrounding non-neoplastic stroma served as an internal control for each slide. The staining “intensity” measurements were translated into the four-tier system as strong (3+), moderate (2+), weak (1+), or negative (–) staining. DNMT3B stains were graded as “overexpression” if nucleus staining “intensity” were 2+ and 3+. FOXO3a stains were graded as “overexpression” if nucleus staining “intensity” were 1+. For MDM2 protein expression level, the stains were graded as “overexpression” if more than 50% tumor cells were immunostaining-positive.

Xenograft Mice Experiments Female BALB/cAnN.Cg-Foxnlnu/CrlNar mice, 4–5 weeks of ages, were raised in pathogen-free conditions within the animal resources center at National Cheng Kung University after obtaining appropriate institutional review board permission. The animals were implanted subcutaneously with 5 × 106 A549 cells in 0.1 ml Hanks’ balanced salt solution in one flank per mouse. The size of the tumor mass was measured and the tumor volume was calculated as 1/2 × length × width2 in mm3. When tumors attained a mass ~50 mm3, the animals were treated intraperitoneally with doxorubicin (1 mg/kg; twice per week for 3 weeks; from Selleck, Houston, TX) or/and Nutlin-3 (20 mg/kg; triplicate per week for 3 weeks; from Selleck) or DMSO as control. The tumor size and body weight of mice were measured on days 1 and 5 of each week till the mice were killed by cervical dislocation. Resected tissues were fixed with 10% formaldehyde for at least over night before embedded in paraffin for histopathologic examination. Tumors were harvested and subjected to immunohistochemistry analyses or qMSP for TSGs as described above.

Patient Information Surgically resected tumor samples from 74 patients diagnosed with primary lung cancer admitted to Veterans General Hospital, Taipei, were collected between 1993 and 2007 after obtaining appropriate institutional review board permission and informed consent from patients. The followup of 72 patients was performed at 2-month intervals in the first year after surgery and at 3-month intervals thereafter at outpatient clinics or by routine phone calls. The mean followup period for all patients was 50 months (range 1–96 months). For the 37 patients who survived the follow-up period (censored patients), the mean follow-up period was 78 months. For the 35 patients who died during the follow-up period, the mean follow-up period was approximately 30 months.

Statistical Analysis The SPSS program (SPSS Inc. Chicago, IL) was used for all statistical analysis. The two-tailed student t test was used to analyze the changes in expression and chromatin

Copyright © 2014 by the International Association for the Study of Lung Cancer

1307

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

Yang et al.

marks levels in cell models. The combination index (CI) was calculated according to CalcuSyn software. CI values less than 1.0 suggested synergy, and values = 1.0 indicated additivity, whereas values more than 1.0 suggested antagonism. The Pearson χ2 test was used to analyze the correlation between protein expression levels in lung cancer patients. Survival curves were calculated according to the KaplanMeier method, and comparison was performed using the log-rank test. Differences were considered significant when p less than 0.05 (*), p less than 0.01 (**) or p less than 0.001 (***). Three independent experiments for cell studies and six mice per group for animal studies were analyzed unless indicated otherwise.

RESULTS FOXO3a Represses the mRNA and Protein Expression of DNMT3B As model cells, A549 and CL1-5 lung cancer cells, which showed low and high FOXO3a basal expression level, respectively, in cytoplasm and nucleus (Fig. 1A, DMSO panels) have been selected, whereas doxorubicin enforced FOXO3a into nuclei in both cells as seen in immunofluorescence staining assay (Fig. 1A, doxorubicin panels). These two cells were used for further FOXO3a overexpression and knockdown experiments.

The RT-qPCR and Western blot assays showed that overexpression of FOXO3a by doxorubicin treatment or GFPFOXO3a transfection reduced DNMT3B mRNA and protein expression. DNMT3B expression was reduced further by ­co-treatment of A549 cells with doxorubicin and FOXO3a transfection (Fig. 1B). In contrast, FOXO3a knockdown induced higher mRNA and protein expression of DNMT3B than control in CL1-5 cells (Fig. 1C).

FOXO3a Negatively Regulates the Promoter Activity of DNMT3B by Binding to Its Putative Binding Sites in DNMT3B Promoter We further analyzed for the presence of FOXO3a binding sites (consensus DNA sequence [G/C/A] [T/C/A] AAA [T/C] A)33 in DNMT3B promoter. Two putative FOXO3a binding sites (FOXO3a-P, −249 to −242 and FOXO3a-E, +166 to +173) in DNMT3B promoter were predicted using the PROMO software (Fig. 2A). To confirm the transcriptional effects of FOXO3a on DNMT3B promoter, we constructed three DNMT3B promoter plasmids. The full length promoter (−483 to +315) contains both putative binding sites, P region promoter (−483 to +50) contains the FOXO3a-P site, and E region promoter (+31 to +315) contains the FOXO3a-E site (Fig. 2A). Dual luciferase reporter assays were performed in A549 (Fig. 2B) and CL1-5 (Fig. 2C) cells to quantify the

FIGURE 1.  FOXO3a represses the mRNA and protein levels of DNMT3B. A, Immunofluorescence analysis of A549 and CL1-5 cells treated with DMSO or doxorubicin. The staining of FOXO3a is shown in green and the nuclei (DAPI) in red. The intensities of endogenous FOXO3a in the A549 cells are lower than that of CL1-5 cells (upper panel). Upon doxorubicin treatment, FOXO3a translocated into the nuclei in both cell lines (lower panel). Scale bar: 10 μm. (B) DNMT3B mRNA and protein level in A549 cells transfected with GFP-FOXO3a or treated with doxorubicin alone or in combination. (C) DNMT3B mRNA and protein level in CL1-5 cells transfected with si-FOXO3a or treated with doxorubicin. Relative DNMT3B mRNA expression was detected by RT-qPCR and the quantified data are shown in the Y-axis. The p values are as indicated and were calculated by two-tailed student t test based on three independent experiments (*p < 0.05; **p < 0.01; *** p < 0.001; upper panel). Quantification of relative protein levels on three different western blots is shown below the blots (lower panel).

1308

Copyright © 2014 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

FOXO3a Transcriptionally Represses DNMT3B

FIGURE 2.  FOXO3a binds to the DNMT3B promoter containing its putative binding sites. A, indicates FOXO3a putative binding sites (FOXO3a-P and FOXO3a-E) and their locations are shown as nucleotide numbers relative to transcriptional start site (+1). Three different DNMT3B promoter constructs were generated: Full length promoter containing both the FOXO3a-P and FOXO3a-E sites, P region containing only FOXO3a-P site, and E region containing only FOXO3a-E site. (B) Dual luciferase assay was performed after cells were transfected with vector control or GFP-FOXO3a followed by promoter firefly luciferase promoter constructs and pGL4-renilla vector in A549 cells. (C) CL1-5 cells were transfected with the si-control or si-FOXO3a followed by promoter constructs and pGL4-renilla vector. (D, E) ChIP to detect FOXO3a binding to P and E regions of DNMT3B promoter in A549 and CL1-5 cells. The binding of FOXO3a to DNMT3B promoters was stronger at E region than P region when expressed at the basal level (lanes 1 versus 3, A549 cell; lanes 1 versus 4, CL1-5 cell). pGFP-FOXO3a transfection and doxorubicin treatment led to high level of FOXO3a, which enhanced its binding to both P region (lanes 1 versus 2) and E region (lanes 3 versus 4) in A549 cells. Doxorubicin treatment alone also enhanced FOXO3a binding to P region (lanes 1 versus 2) and E region (lanes 4 versus 5) in CL1-5 cells, whereas the binding to both P and E regions was abolished by si-FOXO3a with a significant decrease at E region (lanes 4 versus 6). The p values are as indicated and were calculated by two-tailed student t test based on three independent experiments (*p < 0.05; **p < 0.01; ***p < 0.001). n.s., non-significant.

transcriptional regulation of FOXO3a on DNMT3B promoter activity. The cells were co-transfected with pGL4-renilla and either full length, P region or E region promoter plasmids. Vector control or GFP-FOXO3a expression vector was transfected into A549, and si-control or si-FOXO3a was transfected into CL1-5 cells. The data indicated that overexpression of FOXO3a impaired promoter activities of all three promoter constructs. However, the activity of P region promoter, which contains only the putative binding site FOXO3a-P, was less responsive to FOXO3a regulation than other promoter constructs (Fig. 2B). In contrast, knockdown of FOXO3a significantly induced activity of full length and E region promoters containing the putative binding site FOXO3a-E and this induction was less dramatic in P region promoter (Fig. 2C). The data implicated that FOXO3a could

transcriptionally repress the DNMT3B mainly through the putative binding site, FOXO3a-E. We further performed ChIP assay to determine whether FOXO3a could directly bind to the predicted FOXO3a sites in DNMT3B promoter in A549 (Fig. 2D) and CL1-5 (Fig. 2E). Primers specific for P region (−252 to −13) containing FOXO3a-P site and for E region (+71 to +222) containing FOXO3a-E site were used. The data indicated that the basal binding level of FOXO3a to the P region was lower than to the E region (p = 0.031 for A549, lanes 1 versus 3; p = 0.047 for CL1-5, lanes 1 versus 4). However, binding to the P region could be elicited by enforcing abundant FOXO3a into nuclei by treatment with doxorubicin in both cell lines (p = 0.040 for A549, lanes 1 versus 2; p = 0.017 for CL1-5, lanes 1 versus 2).

Copyright © 2014 by the International Association for the Study of Lung Cancer

1309

Yang et al.

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

Endogenous FOXO3a bound to the E region in control cells (lane 3) and the binding of this region was increased by promoting exogenous FOXO3a into nuclei in A549 (p = 0.011, lanes 3 versus 4; Fig. 2D). The binding of FOXO3a to the E region (lane 4) was abolished by si-FOXO3a in CL1-5 cells (p = 0.013, lanes 4 versus 6; Fig. 2E). Together, these data suggested that FOXO3a preferentially binds to the E region when expressed at a low level and binds to both the P and E regions when the nuclear expression level is high.

Binding of FOXO3a Renders DNMT3B Promoter to the Condense Chromatin Status It has been reported that FOXO3a could regulate the transcriptional activities for downstream genes through interaction with other chromatin modifiers such as p300 (a histone acetyltransferase).34,35 Thus, we performed ChIP using anti-histone deacetylase 3 (HDAC3; a corepressor), anti-­ trimethylation of histone 3 lysine 27 (H3K27me3; a chromatin-condensed marker), and anti-acetylation of histone 3 lysine 9 and 14 (H3Ace; a loose chromatin marker) antibodies to assess whether FOXO3a could regulate DNMT3B mRNA expression through chromatin remodeling. ChIP assays were performed for the P and E regions in A549 (Fig. 3A) and CL1-5 (Fig. 3B) cells. The data indicated that the basal binding level of HDAC3 to the P region was lower than the E region (p = 0.039 for A549, lanes 1 versus 3; p = 0.002 for CL1-5, lanes 1 versus 4), and enforcing FOXO3a into nuclei by doxorubicin treatment promoted the binding of HDAC3 to both the P and E regions (p = 0.030 and p = 0.023 for A549, lanes 1 versus 2 and lanes 3 versus 4, Fig. 3A; p = 0.027 and p = 0.049 for CL1-5, lanes 1 versus 2 and lanes 4 versus 5, Fig. 3B). In addition, the binding of HDAC3 and the chromatin-condensed modifications (increased binding of H3K27me3 and decreased binding of H3Ace) on both the P and E regions of DNMT3B promoter was evident by enforcing FOXO3a into nuclei in A549 cells (For HDAC3: p = 0.030, lanes 1 versus 2; p = 0.023, lanes 3 versus 4. For H3K27me3: p = 0.017, lanes 1 versus 2; p = 0.045, lanes 3 versus 4. For H3Ace: p = 0.011, lanes 1 versus 2; p = 0.035, lanes 3 versus 4 in Fig. 3A) and CL1-5 cells (For HDAC3: p = 0.027, lanes 1 versus 2; p = 0.049, lanes 4 vs. 5. For H3K27me3: p = 0.003, lanes 1 versus 2; p = 0.023, lanes 4 versus 5. For H3Ace: p = 0.020, lanes 1 versus 2; p = 0.014, lanes 4 versus 5 in Fig. 3B). Moreover, HDAC3 could not bind to the E region of DNMT3B promoter to form a repressive chromatin structure after FOXO3a knockdown (For HDAC3: p = 0.003, lanes 4 versus 6. For H3K27me3: p = 0.026, lanes 4 versus 6. For H3Ace: p = 0.035, lanes 4 versus 6 in Fig. 3B). These data suggested that FOXO3a may cooperate with HDAC3 to bind to DNMT3B promoter and form a condensed chromatin structure thereby to repressing DNMT3B expression.

Inhibition of MDM2 Induces FOXO3a Expression Resulting in Reduction of DNMT3B Expression FOXO3a is one of the substrates of E3-ubiquitin ligase MDM2.29 To examine whether MDM2 could regulate DNMT3B expression through repression of FOXO3a, we

1310

FIGURE 3.  FOXO3a modulates the chromatin modifications of DNMT3B promoter regions. ChIP for FOXO3a binding to DNMT3B promoter in A549 (A) and CL1-5 (B) cells. Antibody to HDAC3, trimethylation of histone 3 lysine 27 (H3K27me3), and acetylation of histone 3 lysine 9 and 14 antibodies (H3Ace) were used to immunoprecipitate the DNA which was then amplified to detect the P region and E region of DNMT3B promoter. IgG serves as a negative control. The binding of HDAC3 to DNMT3B promoter was stronger at E region than P region when FOXO3a was expressed at the basal level (lanes 1 versus 3, A549 cell; lanes 1 versus 4, CL1-5 cell). pGFP-FOXO3a transfection and doxorubicin treatment led to a repressed chromatin structure manifested by increased HDAC3 binding and H3K27me3 mark along with decreased H3Ace mark at P region (lanes 1 versus 2) and E region (lanes 3 versus 4) in A549 cells. Similar repressed chromatin status was observed in CL1-5 cells (lanes 1 versus 2 and lanes 4 versus 5). This repressed chromatin structure was abolished in si-FOXO3a treated CL1-5 cells especially at the E region (lanes 4 versus 6). The p values were calculated by two-tailed student t test based on at least three independent experiments (*p < 0.05; **p < 0.01; ***p < 0.001).

knocked down MDM2 and found an increase in FOXO3a protein expression along with a decrease in DNMT3B protein level in A549 cells (Fig. 4A). Furthermore, we treated cells with the MDM2 inhibitor, Nutlin-3, which reduced DNMT3B mRNA levels (Fig. 4B).

Copyright © 2014 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

FOXO3a Transcriptionally Represses DNMT3B

FIGURE 4.  Co-treatment with doxorubicin and Nutlin-3 diminishes tumor growth, decreases DNMT3B expression and leads to demethylation of tumor suppressor genes in vitro and in vivo. A, Western blot showing knockdown of MDM2 induced FOXO3a and reduced the protein levels of DNMT3B in A549 cells. (B) Quantitative analysis of relative DNMT3B mRNA expression in A549 cells treated with either doxorubicin, Nutlin-3 or both drugs. All treatments reduced the DNMT3B mRNA expression, and co-treatment with doxorubicin and Nutlin-3 rendered the strongest reduction. The p values are as indicated and were calculated by two-tailed student t test based on three independent experiments. (C) Balb/c nude mice bearing the established A549 tumors were treated intraperitoneally with doxorubicin (1 mg/kg; twice per week) and/or Nutlin-3 (20 mg/kg; three times per week) for 3 weeks. The tumor volumes of mice were measured at the indicated times. Points, mean; bars: SD (n = 6). The p values are as indicated. (D) Immunohistochemical analysis of FOXO3a and DNMT3B protein expression in xenograft from DMSO and co-treated mice groups. Low expression of FOXO3a and overexpression of DNMT3B were found in DMSO control mice (xenograft 1 and 2). After treatment with doxorubicin and Nutlin-3 (xenograft 3 and 4), FOXO3a expression increased and DNMT3B expression decreased. (Original magnification × 200). (E) qMSP assay to detect methylation level of the BLU, RASSF1A, and SLIT2 genes. β-actin is used for internal control. The methylation status of target genes versus β-actin in DMSO control is normalized to “1”. Doxorubicin and Nutlin-3 decreased the methylation status of the three tumor suppressor genes in treated xenograft. The p values are as indicated and were calculated by two-tailed student t test based on three independent experiments. n.s., non-significant.

Copyright © 2014 by the International Association for the Study of Lung Cancer

1311

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

Yang et al.

Co-treatment with Doxorubicin and Nutlin-3 Diminishes Tumor Growth, Decreases DNMT3B Expression and Leads to Demethylation of TSGs In Vitro and In Vivo The above findings prompted us to hypothesize that combination treatment with Nutlin-3 and doxorubicin could be a potential therapeutic strategy based on its ability to decrease DNMT3B expression and DNA methylation level of TSGs through induction of FOXO3a transcriptional activity. Therefore, we treated cells with either Nutlin-3 or doxorubicin alone or in combination. The results showed that comparing to cells treated with Nutlin-3 or doxorubicin alone, the combination treatment further augmented the reduction of DNMT3B mRNA (p < 0.001, p = 0.001, and p < 0.001 for comparisons between combined treatment to control, doxorubicin, and Nutlin-3, respectively; Fig. 4B). The combination index of 0.615 indicated synergism between doxorubicin and Nutlin-3. To confirm the synergistic antitumor activity of combined treatment of doxorubicin and Nutlin-3 in vivo, A549bearing Bulb/c nude mice were injected intraperitoneally with doxorubicin (1 mg/kg; twice per week) and/or Nutlin-3 (20 mg/kg; three times per week) for a total of three weeks. Co-treatment with doxorubicin and Nutlin-3 significantly inhibited tumor growth by almost 60% of tumor volume compared with DMSO control (p < 0.001; Fig. 4C). Moreover, tumor weight of mice treated with the combination was markedly reduced compared with other treatments. To explore whether the in vivo antitumor growth activities of co-treatment with doxorubicin and Nutlin-3 resulted from inducing FOXO3a and reducing DNMT3B expression, tumor sections in xenograft from DMSO control and cotreatment groups were collected for immunohistochemistry analysis of FOXO3a and DNMT3B. Combined treatment with doxorubicin and Nutlin-3 promoted the nucleus expression of FOXO3a leading to the reduction of DNMT3B expression (Fig. 4D, xenograft 3 and 4), whereas low FOXO3a expression accompanied with high DNMT3B expression was seen in DMSO control (Fig. 4D, xenograft 1 and 2). To investigate whether the methylation status of the TSGs could be reversibly reduced by treatment with these drugs, we

detected the methylation level in the promoters of three TSGs, including BLU, RASSF1A, and SLIT2, which had been reported to be hypermethylation in lung cancer cells.36–38 The validation and primary data of qMSP assay are shown in SDC 5 (http:// links.lww.com/JTO/A620). Combined treatment with doxorubicin and Nutlin-3 significantly decreased the methylation status of BLU, RASSF1A and SLIT2 (p < 0.001 for comparisons between combined treatment and DMSO control, Fig. 4E). Together, our data suggested that combined treatment with doxorubicin and Nutlin-3 synergistically reduces the tumor growth and reverses the hypermethylation status of TSGs in vivo.

Lung Cancer Patients with High DNMT3B, Low FOXO3a and High MDM2 Expression Profile Correlate with Poor Prognosis To analyze the correlation between DNMT3B, FOXO3a, and MDM2 expressions in lung cancer patients, immunohistochemistry was performed on tumor from 74 lung cancer patients. An inverse correlation between FOXO3a and DNMT3B expression, (p = 0.026) and between FOXO3a and MDM2 (p = 0.032) was found, whereas positive correlation between the DNMT3B and MDM2 expression was revealed (p = 0.004) (Table 1 and Fig. 5A). Importantly, the Kaplan-Meier survival analysis showed that lung cancer patients with high DNMT3B, low FOXO3a, and high MDM2 expression profile correlated with worse prognosis (p = 0.027; Fig. 5B). Together, these data suggested that the signature of a DNMT3B high, FOXO3a low, and MDM2 high expression profile correlated with poor survival and may serve as a prognostic marker in patients with lung cancer.

DISCUSSION It is a challenging question as to how the expression of DNMTs is regulated in normal cells and dysregulated in tumor cells. Our previous studies show the transcriptional deregulation of DNMT1 by p53 and DNMT3A by RB.39,40 Here, we reveal a new mechanism that FOXO3a binds to the FOXO3a-E region of the DNMT3B promoter to repress DNMT3B expression when FOXO3a is expressed at low level (Fig. 5C). Enforcing abundant FOXO3a to nuclei by doxorubicin or

TABLE 1.  Correlation Between DNMT3B, FOXO3a, and MDM2 Protein Expression in Lung Cancer Patientsa DNMT3B Protein Normal Expression Characteristics Overall FOXO3a protein   Low expression   Normal expression

MDM2 Protein

Over expression

Total

N

(%)b

N

(%)b

74

33

(44.6)

41

(55.4)

42 32

14 19

(33.3) (59.4)

28 13

(66.7)0.026 (40.6)

Over expression

Normal Expression

FOXO3a protein   Low expression   Normal expression DNMT3B protein   Overexpression   Normal expression

Total

N

(%)b

N

(%)b

74

29

(39.2)

45

(60.8)

42 32

12 17

(41.4) (58.6)

30 15

(66.7)0.032 (33.3)

41 33

10 19

(34.5) (65.6)

31 14

(68.9)0.004 (31.3)

These results were analyzed by the χ2 test. The p values with significance are shown as superscripts. Number (N) and percentage (%) of patients analyzed.

a

b

1312

Copyright © 2014 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

FOXO3a Transcriptionally Represses DNMT3B

FIGURE 5.  Patients with DNMT3B high, FOXO3a low, and MDM2 high expression profile correlate with poor survival. (A) The representative immunohistochemistry figures for DNMT3B, FOXO3a and MDM2 are shown for two lung cancer patients. FOXO3a nuclear positive immunoreactivity (+) and DNMT3B and MDM2 negative nuclear immunoreactivity (−) were found in patient 1, whereas patient 2 showed reverse expression pattern. (Original magnification × 200) Arrows indicate the magnified area (magnification × 450) of the original figures. (B) Kaplan-Meier survival curves shows that the patients with a profile of DNMT3B high, FOXO3a low, and MDM2 high expression correlate with worse overall survival compared with other patients. (C–E) Model for the transcription regulation mediated by FOXO3a in relation to doxorubicin and MDM2 controls on DNMT3B promoter. (C) FOXO3a at low basal level preferentially binds on the E site (+166 to +173) of DNMT3B promoter and leads to repression of DNMT3B mRNA and protein expression levels. (D) This negative regulation can be enhanced by abundant FOXO3a nuclear expression when cells are treated with doxorubicin or MDM2 inhibitor. Abundant FOXO3a will be recruited to the distal putative binding site, P site (−249 to −242). (E) Transcriptional repression of DNMT3B expression by FOXO3a can be diminished by MDM2 overexpression through ubiquitination-mediated degradation of FOXO3a protein.

Nutlin-3 treatment further binds to the FOXO3a-P region leading to the formation of condensed chromatin structure and DNMT3B silencing (Fig. 5D). This regulation can be attenuated by MDM2 overexpression in lung cancer cells (Fig. 5E). The present study provides a new dimension for dissecting the deregulation of DNMT3B expression by FOXO3a/MDM2 regulation control. Co-treatment with Nutlin-3 and doxorubicin may inhibit tumor growth through epigenetic modulation, a novel molecular pharmacology mechanism. It has been reported that FOXO3a functions as a transcription factor that could directly bind to the promoter of the downstream regulated genes. In general, FOXO3a acts as a transcriptional activator by interacting with some co-activators, such as p300.34,35 However, there are evidences showing that FOXO3a could suppress downstream target genes.23,24 Our ChIP results suggested that FOXO3a could coordinate with HDAC3 to form a repressive complex that transcriptionally suppresses DNMT3B expression. HDAC3 belongs to class I HDACs, which have been reported to localize in the

nucleus and play crucial roles in transcriptional regulation.41 To verify whether HDAC3 could coordinate with FOXO3a to transcriptionally repress DNMT3B expression, we determined the DNMT3B mRNA and protein expression in CL1-5 cells after knock down of either FOXO3a or HDAC3 or both proteins. The data indicated that both FOXO3a and HDAC3 were involved in transcriptionally repressing DNMT3B expression (SDC 6, http://links.lww.com/JTO/A621). Further studies are needed to verify the interaction between FOXO3a and HDAC3, and perhaps other chromatin modifiers. For example, a study demonstrated that FOXO3a can interact with SIRT1 protein, a class III HDAC, leading to deacetylation of FOXO3a thereby attenuating the transcriptional regulatory activities of FOXO3a.42 In addition, FOXO3a may be a target of miR-155 for gene silencing in breast cancer and miR-155 is considered an “oncomir.” Upregulation of miR-155 has been reported to correlate with poor prognosis in lung cancer.43,44 Whether miR-155 could regulate DNMT3B expression is worthy of further investigation.

Copyright © 2014 by the International Association for the Study of Lung Cancer

1313

Yang et al.

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

Clinically, there are contradictory reports on the prognostic effect of DNMT3B overexpression. For example, DNMT3B status in tumors did not correspond with survival distribution in colorectal cancer and disease recurrence in prostate cancer.10,45 However, overexpressed DNMT3B significantly correlated with shorter overall survival in B-cell lymphomas.46 Notably, our clinical results showed a subgroup of patient with gene expression signature of DNMT3B high, FOXO3a low, and MDM2 high expression profile correlating with poor survival. This defined signature may serve as a prognostic marker in patients with lung cancer. Loss of transcriptional repression of DNMT3B by low FOXO3a mediated by high MDM2 is a clinically relevant regulator of lung cancer that eventually leads to poor prognosis. Our animal study provides a novel pharmacology mechanism for the conventional anticancer drug, doxorubicin, and target therapy drug, Nutlin-3, that could inhibit tumor growth through epigenetic modulation in vivo. To our knowledge, the dosage of individual compound in the combined treatment we used in this study was all below the previous studies, and yet we observed the synergistic decrease of tumor growth and the reduction of methylation status in TSGs. In fact, mice intraperitoneally treated with doxorubicin at a dose below 1.2 mg/kg every 3 days, would not cause significant side effects and toxicity.47 The oral dose reported for Nutlin-3 is 200 mg/kg, twice a day or 40 mg/kg, intrapeitoneally every 2 days in clinical trial studies.31 Some well known antitumor drugs have been reported to activate FOXO3a in cells through different signal transduction pathways, such as cisplatin (DNA damage response), paclitaxel (decrease AKT and increase JNK pathway), and tamoxifen (inhibition of PI3K/AKT pathway).48,49 These signal transduction pathways could phosphorylate FOXO3a thus eliciting the export of FOXO3a from nuclei into the cytoplasm. Note that MDM2, an E3-ubiquitin ligase, physically interacts with FOXO3a resulting in ubiquitin-proteasome-­mediated degradation of FOXO3a.29 Overexpression of MDM2 has been demonstrated in many human cancers.50 Inhibition of MDM2 by Nutlin-3, an MDM2 inhibitor, also resulted in the downregulation of DNMT3B expression and demethylation of TSG promoters. These data further confirmed that FOXO3a transcriptionally represses DNMT3B expression and this regulation can be attenuated by MDM2 expression. We previously showed a link between overexpression of DNMT3A and deregulation of MDM2/RB control in lung cancer.40 In addition, DNMT1 is transcriptional suppressed by p53, another degradation substrate of MDM2, in lung cancer.39 Further investigation as to whether combination of Nutlin-3 with other conventional drugs or small molecular inhibitor could be potent therapeutic approaches to restore DNMT dysregulation in cancers is warranted. In conclusion, we reveal the mechanism by which FOXO3a transcriptionally represses DNMT3B expression and this regulation can be attenuated by MDM2 overexpression in human lung cancer model. Combined treatment with doxorubicin and Nutlin-3 is a novel therapeutic strategy through epigenetic modulation warranting further investigation in lung cancer treatment.

1314

ACKNOWLEDGMENTS

The authors thank the patients and clinicians at the Department of Thoracic Surgery, Taipei Veterans General Hospital for their contributions to this study. REFERENCES 1. Laird PW. Cancer epigenetics. Hum Mol Genet 2005;14 Spec No 1:R65–R76. 2. De Carvalho DD, Sharma S, You JS, et al. DNA methylation screening identifies driver epigenetic events of cancer cell survival. Cancer Cell 2012;21:655–667. 3. Cheng X, Blumenthal RM. Mammalian DNA methyltransferases: a structural perspective. Structure 2008;16:341–350. 4. Siedlecki P, Zielenkiewicz P. Mammalian DNA methyltransferases. Acta Biochim Pol 2006;53:245–256. 5. Okano M, Bell DW, Haber DA, Li E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 1999;99:247–257. 6. Oka M, Meacham AM, Hamazaki T, Rodić N, Chang LJ, Terada N. De novo DNA methyltransferases Dnmt3a and Dnmt3b primarily mediate the cytotoxic effect of 5-aza-2’-deoxycytidine. Oncogene 2005;24:3091–3099. 7. Ehrlich M. The ICF syndrome, a DNA methyltransferase 3B deficiency and immunodeficiency disease. Clin Immunol 2003;109:17–28. 8. Lin RK, Hsu HS, Chang JW, Chen CY, Chen JT, Wang YC. Alteration of DNA methyltransferases contributes to 5’CpG methylation and poor prognosis in lung cancer. Lung Cancer 2007;55:205–213. 9. Saito Y, Kanai Y, Sakamoto M, Saito H, Ishii H, Hirohashi S. Expression of mRNA for DNA methyltransferases and methyl-CpG-binding proteins and DNA methylation status on CpG islands and pericentromeric satellite regions during human hepatocarcinogenesis. Hepatology 2001;33:561–568. 10. Nosho K, Shima K, Irahara N, et al. DNMT3B expression might contribute to CpG island methylator phenotype in colorectal cancer. Clin Cancer Res 2009;15:3663–3671. 11. Qu Y, Mu G, Wu Y, et al. Overexpression of DNA methyltransferases 1, 3a, and 3b significantly correlates with retinoblastoma tumorigenesis. Am J Clin Pathol 2010;134:826–834. 12. Girault I, Tozlu S, Lidereau R, Bièche I. Expression analysis of DNA methyltransferases 1, 3A, and 3B in sporadic breast carcinomas. Clin Cancer Res 2003;9:4415–4422. 13. Cui M, Wen Z, Yang Z, Chen J, Wang F. Estrogen regulates DNA methyltransferase 3B expression in Ishikawa endometrial adenocarcinoma cells. Mol Biol Rep 2009;36:2201–2207. 14. Mei C, Sun L, Liu Y, et al. Transcriptional and post-transcriptional control of DNA methyltransferase 3B is regulated by phosphatidylinositol 3 kinase/Akt pathway in human hepatocellular carcinoma cell lines. J Cell Biochem 2010;111:158–167. 15. Lees-Murdock DJ, Lau HT, Castrillon DH, De Felici M, Walsh CP. DNA methyltransferase loading, but not de novo methylation, is an oocyte-autonomous process stimulated by SCF signalling. Dev Biol 2008;321:238–250. 16. Kaestner KH, Knochel W, Martinez DE. Unified nomenclature for the winged helix/forkhead transcription factors. Genes Dev 2000;14:142–146. 17. Tsai KL, Sun YJ, Huang CY, Yang JY, Hung MC, Hsiao CD. Crystal structure of the human FOXO3a-DBD/DNA complex suggests the effects of post-translational modification. Nucleic Acids Res 2007;35:6984–6994. 18. Medema RH, Kops GJ, Bos JL, Burgering BM. AFX-like Forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature 2000;404:782–787. 19. Modur V, Nagarajan R, Evers BM, Milbrandt J. FOXO proteins regulate tumor necrosis factor-related apoptosis inducing ligand expression. Implications for PTEN mutation in prostate cancer. J Biol Chem 2002;277:47928–47937. 20. Potente M, Urbich C, Sasaki K, et al. Involvement of Foxo transcription factors in angiogenesis and postnatal neovascularization. J Clin Invest 2005;115:2382–2392. 21. Dijkers PF, Medema RH, Pals C, et al. Forkhead transcription factor FKHR-L1 modulates cytokine-dependent transcriptional regulation of p27(KIP1). Mol Cell Biol 2000;20:9138–9148.

Copyright © 2014 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology  ®  •  Volume 9, Number 9, September 2014

22. Stahl M, Dijkers PF, Kops GJ, et al. The forkhead transcription factor FoxO regulates transcription of p27Kip1 and Bim in response to IL-2. J Immunol 2002;168:5024–5031. 23. Wang K, Li PF. Foxo3a regulates apoptosis by negatively targeting miR21. J Biol Chem 2010;285:16958–16966. 24. Wu J, Lee SW, Zhang X, et al. Foxo3a transcription factor is a negative regulator of Skp2 and Skp2 SCF complex. Oncogene 2013;32:78–85. 25. Herzog CR, Blake DC Jr, Mikse OR, Grigoryeva LS, Gundermann EL. FoxO3a gene is a target of deletion in mouse lung adenocarcinoma. Oncol Rep 2009;22:837–843. 26. Mikse OR, Blake DC Jr, Jones NR, et al. FOXO3 encodes a carcinogen-activated transcription factor frequently deleted in early-stage lung adenocarcinoma. Cancer Res 2010;70:6205–6215. 27. Blake DCJr, Mikse OR, Freeman WM, et al. FOXO3a elicits a pro-apoptotic transcription program and cellular response to human lung carcinogen nicotine-derived nitrosaminoketone (NNK). Lung Cancer 2010;67:37–47. 28. Brunet A, Bonni A, Zigmond MJ, et al. Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor. Cell 1999;96:857–868. 29. Yang JY, Zong CS, Xia W, et al. ERK promotes tumorigenesis by inhibiting FOXO3a via MDM2-mediated degradation. Nat Cell Biol 2008;10:138–148. 30. Hu MC, Lee DF, Xia W, et al. IkappaB kinase promotes tumorigenesis through inhibition of forkhead FOXO3a. Cell 2004;117:225–237. 31. Hui RC, Francis RE, Guest SK, et al. Doxorubicin activates FOXO3a to induce the expression of multidrug resistance gene ABCB1 (MDR1) in K562 leukemic cells. Mol Cancer Ther 2008;7:670–678. 32. Vassilev LT, Vu BT, Graves B, et al. In vivo activation of the p53 pathway by small-molecule antagonists of MDM2. Science 2004;303:844–848. 33. Pierrou S, Hellqvist M, Samuelsson L, Enerbäck S, Carlsson P. Cloning and characterization of seven human forkhead proteins: binding site specificity and DNA bending. EMBO J 1994;13:5002–5012. 34. Emerling BM, Weinberg F, Liu JL, Mak TW, Chandel NS. PTEN regulates p300-dependent hypoxia-inducible factor 1 transcriptional activity through Forkhead transcription factor 3a (FOXO3a). Proc Natl Acad Sci U S A 2008;105:2622–2627. 35. Wang F, Marshall CB, Li GY, Yamamoto K, Mak TW, Ikura M. Synergistic interplay between promoter recognition and CBP/p300 coactivator recruitment by FOXO3a. ACS Chem Biol 2009;4:1017–1027. 36. Chang JW, Hsu HS, Ni HJ, et al. Distinct epigenetic domains separated by a CTCF bound insulator between the tandem genes, BLU and RASSF1A. PLoS One 2010;5:e12847.

FOXO3a Transcriptionally Represses DNMT3B

37. Hsu HS, Chen TP, Hung CH, et al. Characterization of a multiple epigenetic marker panel for lung cancer detection and risk assessment in plasma. Cancer 2007;110:2019–2026. 38. Tseng RC, Lee SH, Hsu HS, et al. SLIT2 attenuation during lung cancer progression deregulates beta-catenin and E-cadherin and associates with poor prognosis. Cancer Res 2010;70:543–551. 39. Lin RK, Wu CY, Chang JW, et al. Dysregulation of p53/Sp1 control leads to DNA methyltransferase-1 overexpression in lung cancer. Cancer Res 2010;70:5807–5817. 40. Tang YA, Lin RK, Tsai YT, et al. MDM2 overexpression deregulates the transcriptional control of RB/E2F leading to DNA methyltransferase 3A overexpression in lung cancer. Clin Cancer Res 2012;18:4325–4333. 41. Xu WS, Parmigiani RB, Marks PA. Histone deacetylase inhibitors: molecular mechanisms of action. Oncogene 2007;26:5541–5552. 42. Motta MC, Divecha N, Lemieux M, et al. Mammalian SIRT1 represses forkhead transcription factors. Cell 2004;116:551–563. 43. Kong W, He L, Coppola M, et al. MicroRNA-155 regulates cell survival, growth, and chemosensitivity by targeting FOXO3a in breast cancer. J Biol Chem 2010;285:17869–17879. 44. Yanaihara N, Caplen N, Bowman E, et al. Unique microRNA molecular profiles in lung cancer diagnosis and prognosis. Cancer Cell 2006;9: 189–198. 45. Hoffmann MJ, Engers R, Florl AR, Otte AP, Muller M, Schulz WA. Expression changes in EZH2, but not in BMI-1, SIRT1, DNMT1 or DNMT3B are associated with DNA methylation changes in prostate cancer. Cancer Biol Ther 2007;6:1403–1412. 46. Amara K, Ziadi S, Hachana M, Soltani N, Korbi S, Trimeche M. DNA methyltransferase DNMT3b protein overexpression as a prognostic factor in patients with diffuse large B-cell lymphomas. Cancer Sci 2010;101:1722–1730. 47. Zhang L, Hannay JA, Liu J, et al. Vascular endothelial growth factor overexpression by soft tissue sarcoma cells: implications for tumor growth, metastasis, and chemoresistance. Cancer Res 2006;66:8770–8778. 48. Fernandez de Mattos S, Villalonga P, Clardy J, et al. FOXO3a mediates the cytotoxic effects of cisplatin in colon cancer cells. Mol Cancer Ther 2008;7:3237–3246. 49. Sunters A, Fernández de Mattos S, Stahl M, et al. FoxO3a transcriptional regulation of Bim controls apoptosis in paclitaxel-treated breast cancer cell lines. J Biol Chem 2003;278:49795–49805. 50. Momand J, Jung D, Wilczynski S, Niland J. The MDM2 gene amplification database. Nucleic Acids Res 1998;26:3453–3459.

Copyright © 2014 by the International Association for the Study of Lung Cancer

1315