The Histone Demethylase JMJD2B Plays an Essential Role in Human Carcinogenesis Through Positive Regulation of Cyclin-Dependent Kinase 6

Published OnlineFirst September 19, 2011; DOI: 10.1158/1940-6207.CAPR-11-0290

Cancer Prevention Research Article Research

The Histone Demethylase JMJD2B Plays an Essential Role in Human Carcinogenesis through Positive Regulation of Cyclin-Dependent Kinase 6

Gouji Toyokawa1,3, Hyun-Soo Cho1, Yukiko Iwai1, Masanori Yoshimatsu3, Masashi Takawa1, Shinya Hayami1, Kazuhiro Maejima1, Noriaki Shimizu2, Hirotoshi Tanaka2, Tatsuhiko Tsunoda4, Helen I. Field6, John D. Kelly5,7, David E. Neal5, Bruce A.J. Ponder5, Yoshihiko Maehara3, Yusuke Nakamura1, and Ryuji Hamamoto1,5

Abstract Histone methyltransferases and demethylases are known to regulate transcription by altering the epigenetic marks on histones, but the pathologic roles of their dysfunction in human diseases, such as cancer, still remain to be elucidated. Herein, we show that the histone demethylase JMJD2B is involved in human carcinogenesis. Quantitative real-time PCR showed notably elevated levels of JMJD2B expression in bladder cancers, compared with corresponding nonneoplastic tissues (P < 0.0001), and elevated protein expression was conﬁrmed by immunohistochemistry. In addition, cDNA microarray analysis revealed transactivation of JMJD2B in lung cancer, and immunohistochemical analysis showed protein overexpression in lung cancer. siRNA-mediated reduction of expression of JMJD2B in bladder and lung cancer cell lines signiﬁcantly suppressed the proliferation of cancer cells, and suppressing JMJD2B expression lead to a decreased population of cancer cells in

S phase, with a concomitant increase of cells in G1 phase. Furthermore, a clonogenicity assay showed that the demethylase activity of JMJD2B possesses an oncogenic activity. Microarray analysis after knockdown of JMJD2B revealed that JMJD2B could regulate multiple pathways which contribute to carcinogenesis, including the cell-cycle pathway. Of the downstream genes, chromatin immuno-

precipitation showed that CDK6 (cyclin-dependent kinase 6), essential in G1–S transition, was directly regulated by JMJD2B, via demethylation of histone H3-K9 in its promoter region. Expression levels of JMJD2B and CDK6 were signiﬁcantly correlated in various types of cell lines. Deregulation of histone demethylation resulting in perturbation of the cell cycle, represents a novel mechanism for human carcinogenesis and JMJD2B is a feasible molecular target for anticancer therapy. Cancer Prev Res; 4(12); 2051–61. 2011 AACR.

Introduction

Authors' Affiliations: 1Laboratory of Molecular Medicine, Human Genome Center; 2Division of Clinical Immunology, Advanced Clinical Research Covalent histone modifications, including acetylation, Center, Institute of Medical Science, The University of Tokyo, Shirokanedai, methylation, phosphorylation, ubiquitination, glycosyla- 3 Minato-ku, Tokyo; Department of Surgery and Science, Graduate tion, and sumoylation can modulate chromatin dynamics School of Medical Science, Kyusyu University, Maidashi, Higashi-ku, Fukuoka; 4Laboratory for Medical Informatics, RIKEN, Suehirocho, and affect multiple cellular functions (1–3). Among these Tsurumi-ku, Yokohama, Kanagawa, Japan; 5Department of Oncology, modifications, histone methylation is associated with acti- 6 Cancer Research UK Cambridge Research Institute; Department of vated or repressed transcription (3). Five lysine residues Genetics, University of Cambridge; and 7Division of Surgery & Interven- tional Science, UCL Medical School, University College London, London, (H3K4, H3K9, H3K27, H3K36, and H4K20), located in the United Kingdom N-terminal tails of histones, are reported as representative Note: Supplementary data for this article are available at Cancer Pre- lysines which can become mono-, di-, or trimethylated. vention Research Online (http://cancerprevres.aacrjournals.org/). According to recent findings, H3K9, H3K27, and H4K20

G. Toyokawa and H-S Cho contributed equally to this work. methylation mainly represses transcription, whereas methylation of H3K4 and H3K36 is associated with activated Corresponding Author: Ryuji Hamamoto, Laboratory of Molecular Med- icine, Human Genome Center, Institute of Medical Science, The University transcription (3). Although histone methylation had been of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Japan. Phone: thought to be irreversible, lysine speciﬁc demethylase 1, 81-3-5449-5233; Fax: 81-3-5449-5124; E-mail: [email protected] LSD1, was discovered to be the ﬁrst example of a demethy- doi: 10.1158/1940-6207.CAPR-11-0290 lase that can reverse histone H3 lysine 4 methylation status 2011 American Association for Cancer Research. (4, 5). Later, the Jumonji C (JmjC) domain containing

www.aacrjournals.org 2051

Downloaded from cancerpreventionresearch.aacrjournals.org on September 24, 2021. © 2011 American Association for Cancer Research. Published OnlineFirst September 19, 2011; DOI: 10.1158/1940-6207.CAPR-11-0290

Toyokawa et al.

protein family, which catalyzes the hydroxylation of a lysine polymorphic short-tandem repeat (STR) markers. RERF-LC- methyl group via a radical-based mechanism, has been AI, SBC5 and Huh-7 cells were from Japanese Collection of identified as histone demethylases which differ from LSD1 Research Bioresources (JCRB) in 2001 and tested and (6, 7). Although our knowledge of the physiologic func- authenticated by DNA profiling for polymorphic STR mar- tions of histone demethylases is increasing, it still remains kers. The 253J cells were from Korean Cell Line Bank in unclear how deregulation of the enzymes is involved in 2001 and tested and authenticated by DNA profiling for human diseases such as cancer. polymorphic STR markers. ACC-LC-319 cells were from We previously reported that the histone methyltransfer- Aichi Cancer Center in 2003 and tested and authenticated ase SMYD3 plays a critical role in human carcinogenesis (8– by DNA profiling for SNP, mutation, and deletion analysis. 10). Besides our research, other groups also clarified that All cell lines were grown in monolayers in appropriate dysfunction of histone methylation status contributes to media: Dulbecco modified Eagle medium (DMEM) for human carcinogenesis (11–13), but the detailed relation- RERF-LC-AI, HepG2, Huh-7, NIH3T3, and 293T cells; ship between abnormal histone demethylation and human Eagleminimal essential medium (EMEM) for 253J, CCD- carcinogenesis is unclear. To find demethylases that con- 18, SCaBER, HeLa, SCaBER, and SBC5 cells; McCoy 5A tribute to human carcinogenesis, we examined the expres- medium for HCT116 cells; Leibovitz L-15 for SW480 and sion profiles of a number of proteins containing JmjC SW780 cells; RPMI-1640 medium for A549, H2170, and histone demethylase domains in clinical tissues and found ACC-LC-319 cells, all supplemented with 10% FBS and 1% that expression levels of JMJD2B were significantly upregu- antibiotic/antimycotic solution (Sigma). LoVo cells were lated in cancer tissues, compared with those in correspond- cultured in Ham F-12 medium supplemented with 20% FBS ing normal tissues. JMJD2B, also known as KDM4B, was and 1% antibiotic/antimycotic solution. HFL1 cells were identified in silico (14) and shown to be one of the demethy- cultures in F-12K medium supplemented with 10% FBS, 1% lases capable of removing the trimethyl group from histone antibiotic/antimycotic solution, 2 mmol/L L-glutamine, H3 lysine 9 on pericentric heterochromatin in mammalian and 1,500 mg/L sodium bicarbonate. Cells were main- cells (15). A line of recent reports indicated that hypoxic tained at 37 C in humid air with 5% CO2 condition conditions can induce the expression of some JmjC family (RERF-LC-AI, HepG2, Huh-7, NIH3T3, 293T, HeLa, SCa- members, including JMJD2B (16, 17). In fact, JMJD2B has BER, SBC5, HCT116, A549, H2170, and ACC-LC-319) or been shown to harbor HIF binding sites in their promoter without CO2 (SW480 and SW780). Cells were transfected sequences (17). However, the significance of JMJD2B in with FuGENE6 (Roche Applied Science) according to the oncogenesis and cancer progression is not fully understood manufacturer protocol. so far. Here, we showed a critical role for JMJD2B in carcino- Expression profiling in cancer using cDNA microarrays genesis, through the regulation of cancer-related down- We established a genome-wide cDNA microarray with stream genes, and suggested the possibility that JMJD2B 36,864 cDNAs selected from the UniGene database of the might be a novel therapeutic target for several types of National Center for Biotechnology Information (NCBI). cancer, especially bladder and lung cancer. This microarray system was constructed essentially as described previously (20). Briefly, the cDNAs were ampli- þ Materials and Methods fied by reverse transcriptase PCR (RT-PCR) using poly (A) RNAs isolated from various human organs as templates; the Tissue samples and RNA preparation lengths of the amplicons ranged from 200 to 1,100 bp, Bladder tissue sampling and RNA preparation were without any repetitive or poly (A) sequences. Many types of described previously (18). Briefly, 76 surgical specimens of tumor and corresponding nonneoplastic tissues were pre- primary urothelial carcinoma were collected, either at pared in 8-mm sections, as described previously (20). A total cystectomy or transurethral resection of bladder tumor, and of 30,000 to 40,000 cancer or noncancerous cells were snap frozen in liquid nitrogen. Twenty specimens of normal collected selectively using the EZ cut system (SL Microtest bladder urothelial tissue were collected from areas of mac- GmbH) according to the manufacturer protocol. Extraction roscopically normal bladder urothelium in patients with no of total RNA, T7-based amplification, and labeling of evidence of malignancy. Vimentin is primarily expressed in probes were done as described previously (20). A measure mesenchymally derived cells and was used as a stromal of 2.5-mg aliquots of twice amplified RNA (aRNA) from each marker. Uroplakin is a marker of urothelial differentiation cancerous and noncancerous tissue was then labeled, and is preserved in up to 90% of epithelially derived tumors respectively, with Cy3-dCTP or Cy5-dCTP. (19). Use of tissues for this study was approved by Cam- bridgeshire Local Research Ethics Committee (Ref 03/018). Quantitative real-time PCR As described previously, we prepared 76 bladder cancer Cell culture and 20 normal bladder tissues in Addenbrooke’s Hospital, NIH3T3, CCD-18Co, SW780, SCaBER, A549, H2170, Cambridge UK. For quantitative RT-PCR reactions, specific SW480, HCT116, LoVo, HepG2, HeLa, HFL1, and 293T primers for all GAPDH (housekeeping gene), JMJD2B, cells were from American Type Culture Collection in 2001 HDAC1, AKT3, and MAP3K1 were designed (Primer and 2003 and tested and authenticated by DNA profiling for sequences in Supplementary Table S1; ref. 21). PCR

2052 Cancer Prev Res; 4(12) December 2011 Cancer Prevention Research

Involvement of JMJD2B in Human Carcinogenesis

reactions were done using the LightCycler 480 System tary Table S2. siRNA duplexes (100 nmol/L final concen- (Roche Applied Science) following the manufacturer pro- tration) were transfected into bladder and lung cancer cell tocol. Fifty percent SYBR GREEN universal PCR Master Mix lines with Lipofectamine 2000 (Invitrogen), and cell via- without UNG (Applied Biosystems), 50 nmol/L each of the bility was examined at indicated time points using the Cell forward and reverse primers and 2 mL of reversely tran- Counting Kit-8 (Dojindo) and by colony formation assay as scribed cDNA were applied. Amplification conditions were described previously (24). 5 minutes at 95C and then 45 cycles each consisting of 10 sec at 95 C, 1 minute at 55 C and 10 sec at 72 C. Then, Immunoblotting reactions were heated for 15 sec at 95C, 1 minute at 65Cto Whole-cell lysates were prepared from the cells with draw the melting curve, and cooled to 50 C for 10 seconds. radioimmunoprecipitation assay–like buffer, and total Reaction conditions for target gene amplification were as protein was transferred to nitrocellulose membrane. The described above and the equivalent of 5 ng of reverse membrane was probed with anti-JMJD2B antibody (H-200; transcribed RNA was used in each reaction. mRNA levels Santa Cruz Biotechnology), anti-JMJD2B antibody (A301- GAPDH were normalized to expression. 478A; Bethyl Laboratories), anti-H3K9me3 (ab8898; Abcam), and anti-FLAG antibody (F7425; Sigma). ACTB Immunohistochemistry and tissue microarray (I-19; Santa Cruz Biotechnology) or histone H3 (ab1791, Immunohistochemical analysis was done using anti- Abcam) was used to ensure equal loading and transfer of JMJD2B antibody (A301-478A; Bethyl Laboratories) as proteins. Protein bands were detected by incubating with described previously (22). For clinical bladder and lung HRP-conjugated antibodies (GE Healthcare) and visualiz- cancer tissue microarray, EnVision kit/horseradish peroxi- ing with Enhanced Chemiluminescence (GE Healthcare). dase (HRP; Dako) was applied. Briefly, slides of paraffin- embedded tumor specimens were processed under high Clonogenicity assays pressure (125 C, 30 seconds) in antigen-retrieval solution, NIH3T3 cells, cultured in DMEM 10% FBS, were trans- high pH 9 (S2367; Dako), treated with peroxidase blocking fected with a p3xFLAG-JMJD2B wild-type vector or a regent, and then treated with protein blocking regent p3xFLAG-JMJD2B DJmjC mutant vector. The transfected (X0909; Dako). Tissue sections were incubated with a rabbit NIH3T3 cells were cultured for 2 days and seeded in 10-cm anti-JMJD2B polyclonal antibody followed by secondary dish at the density of 10,000 cells per 10-cm dish in antibodies conjugated to peroxidase labeled dextran poly- triplicate. Subsequently, the cells were cultured in DMEM mers (Dako). Antigen was visualized with substrate chro- 10% FBS containing 0.9 (mg/mL) geneticin/G-418 for 2 mogen (Dako liquid DAB chromogen; Dako). Finally, weeks until colonies were visible. Colonies were stained tissue specimens were stained with Mayer’s haematoxylin with Giemsa (Merck) and counted by Colony Counter (Hematoxylin QS, Vector Laboratories) to discriminate the software. nucleus from the cytoplasm. Because the intensity of staining within each tumor tissue core was mostly homoge- neous, the intensity of JMJD2B staining was semiquantita- Coupled cell-cycle and cell proliferation assay 0 0 tively evaluated using the following criteria: negative (no A5-bromo-2 -deoxyuridine (BrdU) flow kit (BD Phar- appreciable staining in tumor cells) and positive (brown mingen) was used to determine the cell-cycle kinetics and staining appreciable in more than 30% of the nucleus of to measure the incorporation of BrdU into DNA of tumor cells). proliferating cells (25, 26). The assay was done according to the manufacturer protocol. Briefly, cells (2 105 per Immunocytochemistry well) were seeded overnight in 6-well tissue culture plates HeLa cells were transfected with pCAGGS-n3FC-JMJD2B. and treated with an optimized concentration of siRNAs in Forty-eight hours after transfection, cultured cells were fixed medium containing 10% FBS for 72 hours, followed by by 4% paraformaldehyde in 0.1 mol/L phosphate buffer addition of 10 mmol/L BrdU and incubations continued (pH 7.4) at room temperature for 30 minutes and permea- for an additional 30 minutes. Both floating and adherent bilized with 0.5% Triton X-100 in PBS (Sigma). Fixed cells cells were pooled from triplicate wells per treatment were blocked with 5% bovine serum albumin in PBS for 1 point, fixed in a solution containing paraformaldehyde hour and incubated with primary antibodies overnight at and the detergent saponin, and incubated for 1 hour with 4C. Then they were incubated with Alexa Fluor–conjugated DNase at 37 C(30mg per sample). Fluorescein isothio- second antibodies (Molecular Probes; Invitrogen) and cyanate (FITC)-conjugated anti-BrdU antibody (1:50 observed using a Leica confocal microscopy (23). dilution in wash buffer; BD Pharmingen) was added and incubation continued for 20 minutes at room tempera- siRNA transfection ture. Cells were washed in wash buffer and total DNA was siRNA oligonucleotide duplexes targeting the human stained with 7-amino-actinomycin D (7-AAD; 20 mLper JMJD2B transcripts were purchased from Sigma Genosys. sample), followed by flow cytometric analysis using siEGFP and siNegative control (siNC), which is a mixture of FACScan (Beckman Coulter) and total DNA content (7- 3 different oligonucleotide duplexes, were used as control AAD) was determined using CXP Analysis Software Ver. siRNAs. The siRNA sequences are described in Supplemen- 2.2 (Beckman Coulter).

www.aacrjournals.org Cancer Prev Res; 4(12) December 2011 2053

Toyokawa et al.

Microarray hybridization and statistical analysis for clinical bladder cancer samples and found significant dif- the clarification of downstream genes ferences in expression levels between normal and cancer Microarray analysis to identify downstream genes were cells for the JMJD2B gene (data not shown). Then, we done as described previously (21). Purified total RNA was analyzed 76 bladder cancer samples and 20 normal control labeled and hybridized onto Affymetrix GeneChip U133 samples (British) and found significantly higher expression Plus 2.0 oligonucleotide arrays (Affymetrix) according to levels of JMJD2B in bladder cancer tissues compared the manufacturer instructions. Probe signal intensities were with normal bladder (P < 0.0001, Mann–Whitney U normalized by RMA and Quantile (using R and Biocon- test; Fig. 1A and B). Subclassification of tumors according ductor). Because we could also confirm the microarray data to tumor grade, metastasis status, recurrence status, and of several randomly selected candidate genes using quan- smoking history identified no significant differences, titative real-time PCR (Supplementary Fig. S4D), we assume whereas gender was significant (P < 0.0001, Mann–Whitney our microarray data are reliable. U test; Supplementary Table S3). Importantly, JMJD2B was notably overexpressed at an early stage of bladder cancer Chromatin immunoprecipitation assay (Supplementary Fig. S1). Next, we evaluated JMJD2B pro- Chromatin immunoprecipitation (ChIP) assays were tein expression levels in bladder tissues. After confirming done using ChIP Assay kit (Millipore) according to the the specificity of the antibody we used (Supplementary Fig. manufacturer protocol. Briefly, the fragment of JMJD2B S2), we carried out immunohistochemistry using the spe- and chromatin complexes was immunoprecipitated with cific JMJD2B antibody. This experiment showed strong anti-FLAG antibody 48 hours after transfection with nuclear staining of JMJD2B, specifically in bladder cancer pCAGGS-n3FC (mock) and pCAGGS-n3FC-JMJD2B vec- tissues, whereas no significant staining was observed in tors into 293T cells. After the bound DNA fragments to normal bladder tissues (Fig. 1C). Among 29 bladder cancer JMJD2B were eluted, a standardized amount was subjected tissue sections, 20 showed positive staining of JMJD2B to quantitative real-time PCR reactions. Primer sequences (70.0%; Supplementary Table S4). In addition, our micro- are shown in Supplementary Table S1. array expression analysis of a number of clinical samples indicated that JMJD2B was overexpressed in lung cancer Results (Fig. 2A and B). According to immunohistochemical analysis using clinical lung tissues, nuclear staining of JMJD2B Overexpression of JMJD2B in clinical bladder and was observed in cancer tissues, although no significant lung cancer tissues nuclear staining was evaluated in normal lung and placental We first examined expression levels of a number of tissues (Fig. 2B). Of 63 lung cancer cases, JMJD2B stained jumonji histone demethylase genes in a small subset of positively in 24 cases (38.1%; Supplementary Table S5).

A B 180 100 Mann-Whitney 160 90 U–test: U = 174.0 140 80 Z = -5.287 P < 0.0001 Figure 1. Overexpression of 120 70 mRNA level 100 Mean value JMJD2B in clinical bladder cancers.

mRNA level 60 Normal: 7.859 80 Tumor: 41.472 A, mRNA levels of JMJD2B in 50 60 representative cases of clinical JMJD2B 40 40 JMJD2B 30 normal bladder tissues and bladder 20 20 cancer tissues (British). B,

Relative Relative 0 10

1T 5T quantitative real-time PCR Relative Relative 11T 15T 27T 28T 38T 39T 43T 46T 50T 54T 58T 69T 72T 74T 83T 88T 92T 95T 0 BN1A BN4A BN6A BN9A n n measurements were carried out BN12A BN14A BN17B BN21A BN24B BN26B Normal ( = 20) Tumor ( = 76) Normal bladder Bladder tumor Bladder tissues using 76 bladder cancer samples and 20 normal bladder samples, C Normal bladder Transitional cell carcinoma Case 1 Case 2 Case 3 Case 4 Case 5 and the result is shown by box- 65/M 27/M, TisN0M0, Grade I 50/M, T1N0M0, Grade I 47/M, T2N0M0, Grade II 61/M, T2N0M0, Grade III whisker plot. For statistical analysis, the Mann–Whitney U test was adopted (mean: 41.47 (tumor) versus 7.86 (normal), P < 0.0001). C, immunohistochemical analysis of JMJD2B in bladder tissues. Clinical

Adenocarcinoma Squamous cell carcinoma information for each section is Case 6 Case 7 represented above histologic 68/M, T2N0M0, Grade III 71/M, T1N0M0, Grade I pictures. All tissue samples were purchased from BioChain. Original magniﬁcation: 200.

Original magnification: 200x

2054 Cancer Prev Res; 4(12) December 2011 Cancer Prevention Research

Involvement of JMJD2B in Human Carcinogenesis

A Non-small cell lung cancer (Japanese case) B Non-small cell lung cancer (Japanese case) 5.0 200 Mann-Whitney Figure 2. Elevated expression of U 4.5 ] 180 –test: JMJD2B in lung cancer tissues. A, 4 U = 244.0 4.0 160 Z = -2.425 expression ratio between normal P = 0.0153 3.5 140 lung and non–small cell lung cancer Mean value 3.0 120 Normal: 46430 (NSCLC) tissues. Signal intensity Tumor: 81990 2.5 100 for each sample was analyzed by 2.0 80 cDNA microarray, and the 1.5 60 expression ratio is the signal [x10 Signal intensity 1.0 40 intensity in tumor divided by that in 0.5 20 normal (1 is normal). B, comparison Expression ratio (Tumor/Normal) 0.0 0 of JMJD2B expression between Normal (n = 28) Tumor (n = 28) normal and tumor (NSCLC) lung tissues in Japanese patients. Signal NSCLC_01 NSCLC_02 NSCLC_03 NSCLC_04 NSCLC_05 NSCLC_06 NSCLC_07 NSCLC_08 NSCLC_09 NSCLC_10 NSCLC_11 NSCLC_12 NSCLC_13 NSCLC_14 NSCLC_15 NSCLC_16 NSCLC_17 NSCLC_18 NSCLC_19 NSCLC_20 NSCLC_21 NSCLC_22 NSCLC_23 NSCLC_24 NSCLC_25 NSCLC_26 NSCLC_27 NSCLC_28 intensity for each sample was C Normal lung Normal placenta Squamous cell carcinoma analyzed by cDNA microarray, and Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 65/M 29/F 65/F, mod, T1N0M0 70/M, por , T2N0M0 69/M, por , T2N1M0 63/M, mod, T3N0M0 the result is shown by box-whisker plot (median 50% boxed). Mann– Whitney U-test was used for the statistical analysis. C immunohistochemical analysis of JMJD2B in normal and lung cancer Squamous cell tissues. Clinical information for carcinoma Adenocarcinoma Adenosquamous Large cell carcinoma Atypical carcinoma each section is represented above Case 7 Case 8 Case 9 Case 10 Case 11 Case 12 40/M, well, T3N1M0 70/F, mod, T1N0M0 56/F, por, T2N0M0 70/M, mod , T2N1M0 47/F, mod , T2N0M0 65/M, mod , T3N0M0 histologic pictures. All tissue samples were purchased from BioChain. Original magniﬁcation: 200.

Original magnification: 200x

Together, expression levels of JMJD2B in bladder and lung G1 phase significantly increased, indicating that JMJD2B cancer tissues are significantly higher than those in corre- could be a critical factor in the regulation of the G1–S sponding nonneoplastic tissues at the mRNA and protein transition in cancer cells. These results suggested that levels. JMJD2B plays an essential role in the growth regulation of cancer cells by modulating the G1–S transition. JMJD2B plays an essential role in the growth regulation To examine the oncogenic activity of JMJD2B in more of cancer cells detail, we conducted a clonogenicity assay. A JMJD2B To determine the significance of JMJD2B in human expression vector was constructed, capable of driving pro- carcinogenesis, we examined whether JMJD2B is involved duction of active enzyme within transfected cell lines (Fig. in the growth regulation of cancer cells. After confirming the 4B). A wild-type JMJD2B (JMJD2B Wt) vector and an elevated expression of JMJD2B in bladder and lung cancer enzyme-dead JMJD2B (JMJD2B DJmjC) version of our cell lines, compared with the human airway epithelial cell vector were transfected into separate cultures of NIH3T3 line SAEC and the colonic fibroblast cell line CCD-18Co at cells. A clonogenicity assay was done on each culture (Fig. the protein level (Fig. 3A), we inhibited JMJD2B expression 4C; Materials and Methods). Wild-type JMJD2B protein in SW780 and A549 cells using 2 different siRNA oligonu- showed stronger oncogenic activity than enzyme-dead cleotide duplexes (Supplementary Table S2). As shown JMJD2B protein; therefore, it is the demethylase activity of in Fig. 3B, the specific siRNAs clearly knocked down JMJD2B JMJD2B that promotes oncogenesis in cells. Because expression, and subsequently, we carried out a cell growth JMJD2B is overexpressed at an early stage in cancer progres- assay 72 hours after treatment with the same siRNAs. sion, JMJD2B seems to play a crucial role in human JMJD2B knockdown significantly suppressed growth of carcinogenesis. bladder and lung cancer cells (Fig. 3C), and this result was also confirmed by a colony formation assay (Fig. 3D). JMJD2B promotes cell-cycle progression through the Similarly, growth suppression effect was confirmed in the regulation of cyclin-dependent kinase 6 other bladder cancer cell line SCaBER (Supplementary Fig. We next analyzed downstream genes and pathways asso- S3A and B). We then conducted BrdU and 7-AAD staining to ciated with JMJD2B to clarify the mechanism by which analyze the detailed cell-cycle status of cancer cells after the JMJD2B regulates cell-cycle progression. After treatment knockdown of JMJD2B and confirmed that the proportion with JMJD2B siRNA in SW780 and A549, we carried out of cancer cells at the S phase was significantly decreased Affymetrix GeneChip analysis. To exclude the secondary (Fig. 4A). Concomitantly, the percentage of cancer cells at effects of growth suppression, we picked statistically

www.aacrjournals.org Cancer Prev Res; 4(12) December 2011 2055

Toyokawa et al.

A Bladder Lung B siJMJD2B ** cancer cancer #1

siEGFP #2 * IB: JMJD2B SW780 SAEC CCD-18Co SW780 SCaBER A549 ACC-LC- 319 SBC5 IB: Histone H3 IB: JMJD2B IB: JMJD2B IB: ACTB A549 *: human small airway epithelial cells IB: Histone H3 * * : human colon fibroblast

C siEGFP siJMJD2B#1 D siNC siJMJD2B#2 1.6 siEGFP siJMJD2B#2 * 1.4 * * * * * * * * 1.2 * * * SW780 * * * * * * * * 1.0

0.8

0.6 A549 Relative cell number Relative 0.4

0.2

0 SW780 SCaBER RT4 A549 ACC-LC-319 Bladder cancer cell lines Lung cancer cell lines

Figure 3. Knockdown of JMJD2B expression leads to signiﬁcant growth suppression of cancer cells. A, validation of JMJD2B protein expression levels in various cell lines. Lysates from the human airway epithelial cell line (SAEC), the colonic ﬁbroblast cell line (CCD-18Co), 2 bladder cancer cell lines (SW780 and SCaBER) and 3 lung cancer cell lines (A549, ACC-LC-319, and SBC5) were immunoblotted with antibodies against JMJD2B and ACTB (an internal control). B, validation of JMJD2B knockdown at the protein level. Lysates from SW780 and A549 cells 72 hours after siRNA treatment were immunoblotted with anti- JMJD2B and anti-histone H3 (an internal control) antibodies. C, effects of JMJD2B knockdown on the proliferation of bladder and lung cancer cell lines (SW780, SCaBER, RT4, A549, and ACC-LC-319) measured by Cell Counting kit 8. Relative cell numbers are normalized to the number of siEGFP-treated cells (siEGFP ¼ 1): results are the mean SD of 3 independent experiments. P values were calculated using Student t test (, P < 0.05). D, colony formation assay of SW780 and A549 cells. Giemsa staining was done 5 days (SW780) and 3 days (A549) after treatment with siRNAs. IB, immunoblotting.

significant signals which were commonly decreased at 12 checkpoint pathways, as well as other important pathways and 24 hours after siRNA treatment. The knockdown of involved in carcinogenesis (Supplementary Table S6). The JMJD2B was clearly confirmed at the protein level (Fig. 5A, data revealed that JMJD2B directly activates the expression right top). Among candidate genes, we observed a signifi- of CDK6 through demethylation of histone H3 at lysine 9. cant downregulation of cyclin-dependent kinase 6 (CDK6), Because it is well-known that p16INK4a is an impor- one of the key regulators for the G1–S transition (Fig. 5A, left tant regulator of CDK6 (28–32), we examined the rela- and right bottom and Supplementary Fig. S4A; ref. 27). tionship between p16INK4A and JMJD2B in regulating Quantitative real-time PCR also confirmed a significant CDK6. To compare expression levels of p16INK4a, downregulation of CDK6 at both 12 and 48 hours following JMJD2B,andCDK6, we conducted quantitative real-time JMJD2B siRNA treatment (Fig. 5B and Supplementary Fig. PCR analysis on 14 different cell lines (Fig. 6A). Although S4B). Likewise, downregulation of CDK6 after JMJD2B p16INK4A is considered an important inactivator of knockdown was confirmed at the protein level (Supple- CDK6, we did not find a significant inverse correlation mentary Fig. S4C). In addition, we confirmed a significant between p16INK4a and CDK6 expressions; instead, a elevation of CDK6 in HeLa cells transfected with pCAGGS- slight positive correlation was observed (Fig. 6B, left). n3FC-JMJD2B compared with mock-treated cells (Fig. 5C). This implies that p16INK4a is unlikely to regulate tran- To evaluate the possibility that JMJD2B directly regulates scription levels of CDK6 to inactivate the cell-cycle depen- CDK6 expression at the transcriptional level, we conducted dent-kinase. On the contrary, expression levels of JMJD2B a ChIP assay. JMJD2B protein was highly enriched at the and CDK6 were significantly positively correlated in 14 promoter region of CDK6 after transfection with a cell lines (Fig. 6B, right), which is consistent with our pCAGGS-n3FC-JMJD2B vector, accompanied by decreased results (Fig. 5). Intriguingly, expression levels of JMJD2B levels of H3-K9 trimethylation in the region (Fig. 5D). and CDK6 were strongly correlated in cell lines expressing Consistently, signal pathway analysis using the Gene Ontol- low levels of p16INK4a (Fig. 6C; r ¼ 0.976, P < 0.0001). ogy database indicated that JMJD2B can regulate cell-cycle According to our data, JMJD2B-dependent transcriptional

2056 Cancer Prev Res; 4(12) December 2011 Cancer Prevention Research

Involvement of JMJD2B in Human Carcinogenesis

A A549 siEGFP siJMJD2B #2 B DAPI FLAG H3K9me3 (JMJD2B) Anti-BrdU-FITC DA IP FLAG 2em9K3H DNA content (JMJD2B) 100 siEGFP 80 siJMJD2B#2 60

40 IPAD FLAG em9K3H (JMJD2B) 20 H3K9m Merge

Percentage of Percentage cells 0 e2 SubG1 G1 SG2–M Cell-cycle phase

C IPAD FLAG 3em4K3H (JMJD2B) 180 P = 0.0025* 160 JMJD2B Wt 140

120 IPAD FLAG 1PH β (JMJD2B) 100 80 60

JMJD2B Number of colonies 40 Δ JmjC Scale bar: 10 µm 20 0 JMJD2B JMJD2B Wt ΔJmjC

Figure 4. Fluorescence-activated cell sorting (FACS) and clonogenicity analyses. A, effects of JMJD2B knockdown on cell-cycle kinetics in cancer cells. A549 cells were collected 72 hours after treatment with siRNAs, and cell-cycle distribution was analyzed by ﬂow cytometry after coupled staining with FITC- conjugated anti-BrdU and 7-AAD, described in Materials and Methods. Representative histograms of the FACS results (top) and numerical analysis (bottom) are shown for each experiment. B, validation of subcellular localization and an enzymatic activity of JMJD2B. 3xFLAG-JMJD2B transfected HeLa cells were stained for speciﬁc antibodies as indicated above. Immunostaining showed nuclear localization of JMJD2B and its reversing activity against tri- and di-methylated H3K9, accompanied with the increased signal of H3K9me and the decreased signal of Hp1beta. Scale bar denotes 10mm. DAPI, 40,6- diamidino-2-phenylindole. C, clonogenicity assays of NIH3T3 cells. Cells transfected with a 3xFLAG-JMJD2B vector and a 3xFLAG-JMJD2B DJmjC were cultured in DMEM 10% FBS containing 0.9 (mg/mL) geneticin/G-418 for 2 weeks and colonies were stained with Giemsa.

regulation of CDK6 seems to be affected by expression of cancers, including bladder and lung, by quantitative real- levels of p16INK4a. time PCR, cDNA microarray, or immunohistochemistry. In addition, a series of our experiments clarified that JMJD2B Discussion serves a critical role in the growth regulation of cancer cells, especially at the G1–S transition (Fig. 3 and Supplementary Among posttranslational modifications on histones, Fig. S3). This JMJD2B-dependent cell-cycle regulation could methylation has been shown to be involved in transcrip- be mediated by the downstream gene CDK6 through the tional regulation, such as X-inactivation or genomic imprint- demethylation of histone H3-K9 at the promoter region (Fig. ing (1, 2). Although it had been unclear whether histone 5), indicating that the enzyme activity of JMJD2B may be an methylation marks could be reversed, LSD1, a flavin-depen- important regulator for the G1–S transition in cancer cells. dent amine oxidase histone demethylase, was identified as Expression analysis examining the correlation between the first histone demethylase (4). Unlike LSD1, the JmjC JMJD2B and the histone methyltransferase G9a, catalyzing family, harboring the conserved JmjC domain, require Fe histone H3-K9 methylation, showed that G9a expression (II) and a-ketoglutarate to exert demethylase activity, and was not changed after treatment with JMJD2B siRNA (Sup- this demethylation process results in the generation of plementary Fig. S5). We interpret this to mean that although formaldehyde and succinate (7). Among these enzymes, the both proteins regulate the methylation status of histone JMJD2 family demethylate H3K9 and H3K36 (6), and H3K9, the target genes may be different. In fact, we have JMJD2B selectively removes H3K9me2 and H3K9me3 recently published a paper analyzing the significance of G9a (15, 17). Functional studies showed that JMJD2B contains deregulation in human carcinogenesis (33). We reported hypoxia response elements in its promoter region and, that G9a is overexpressed in various types of cancer, similarly therefore, JMJD2B is induced by Hypoxia-inducible factor to JMJD2B. Our functional analyses showed that G9a and (HIF; refs. 16, 17), but its significance in human diseases JMJD2B may regulate different genes (our previous data in such as cancer remains to be elucidated. In this study, we ref. 33 and Supplementary Table S6). Therefore, the histone showed significant upregulation of JMJD2B in various types demethylase JMJD2B and the histone methyltransferase G9a

www.aacrjournals.org Cancer Prev Res; 4(12) December 2011 2057

Toyokawa et al.

A B siJMJD2B siEGFP #1 #2 A549, 48 h SW780, 48 h IB: JMJD2B 1.6 1.6 12 h * * IB: Histone H3 1.4 1.4 * * * * * IB: JMJD2B 1.2 * 1.2 24 h 1.0 1.0 IB: Histone H3 CDK6 0.8 0.8 1.2 * * 0.6 0.6 1.0 12 h 0.4 0.4 Relative mRNA Relative levels 0.8 24 h mRNA Relative levels 0.2 0.2

0.6 0 0 JMJD2B CDK6 JMJD2B CDK6 0.4 siEGFP siNC siJMJD2B#2 0.2 Relative signal intensity Relative 0 siEGFP siJMJD2B#2 C D Ch1 CDK6 locus

TSS 100 bp TSS: transcription start site CDK6 1.4 2.0 * 0.20 * * 1.2 0.18 1.8 0.16 1.6 FLAG-Mock FLAG-JMJD2B 1.0 0.14 1.4 IB: FLAG 0.8 0.12 1.2 0.10 1.0 0.6 IB: H3K9me3 0.08 0.8 0.4 0.06 0.6 IB: ACTB Percent of Percent input 0.4 0.2 0.04 Relative mRNA Relative levels 0.02 0.2 0 0 status H3K9 trimethylation 0 FLAG- FLAG- Mock JMJD2B Mock JMJD2B Mock JMJD2B [FLAG] [FLAG] [FLAG] [FLAG] JMJD2B bound (%) Ch: FLAG H3K9me3 status (%)

Figure 5. JMJD2B transcriptionally activates CDK6 expression through demethylation of H3K9me3 at the promoter region. A, left, 2-dimensional, unsupervised hierarchical cluster analysis of SW780 and A549 mRNA expression profiles after knockdown of JMJD2B expression. Differentially expressed genes were selected for this analysis. Right top, validation of JMJD2B knockdown at the protein level. Lysates from SW780 cells 12 and 24 hours after siRNA treatment were immunoblotted with anti-JMJD2B and anti-histone H3 (an internal control) antibodies. Right bottom, signal intensity of CDK6 in SW780 cells after treatment with siEGFP (control) and siJMJD2B#2 was quantified by GeneChip U133 plus 2.0 (Affymetrix). B, relative CDK6 mRNA levels in A549 and SW780 cells 48 hours after treatment with siEGFP, siNC, and siJMJD2B#2 were analyzed by quantitative real-time PCR. Results are the mean SD in 3 independent experiments, and the P value was calculated using Student t test (, P < 0.05). C, HeLa cells transfected with pCAGGS-n3FC-mock or pCAGGS- n3FC-JMJD2B were collected 48 hours after the transfection. Left, the samples were fractionated by SDS-PAGE and immunoblotted with anti-FLAG, anti- H3K9me3 and anti-ACTB (an internal control) antibodies. Right, relative CDK6 mRNA levels were analyzed by quantitative real-time PCR. Results are the mean SD in 3 independent experiments, and the P value was calculated using Student t test (, P < 0.05). D, ChIP assay for JMJD2B at the promoter region of CDK6 gene. Top, a schematic diagram of the CDK6 promoter region. The PCR amplified fragment is positioned by nucleotide number relatives to TSS (arrow). Left bottom, quantitative real-time PCR analysis using a primer pair as described in Materials and Methods. Cross-linked and sheared chromatin was immunoprecipitated with anti-FLAG antibody (M2, Sigma). The result is shown as a percentage of the input chromatin. Right bottom, quantification of H3K9me3 ChIP at the CDK6 promoter region using quantitative real-time PCR. Cross-linked and sheared chromatin was immunoprecipitated with an anti- H3K9me3 antibody (ab8898; Abcam). IB, immunoblotting.

both regulate the status of histone H3K9 methylation and phase through promoting RB phosphorylation and subse- play important roles in human carcinogenesis but may quent detachment of E2F1 from RB (27). CDK6 has been regulate different pathways. reported to possess various physiologic functions like T-cell CDK6 is a member of the cyclin-dependent protein development (34, 35). Importantly, our microarray data kinase family and a catalytic subunit of the protein kinase showed the aberrant expression of CDK6 in several types of complex that is important for cell-cycle regulation. This human tumors, including bladder and lung cancers (Sup- kinase promotes cell-cycle progression from G1 phase to S plementary Table S7), and previous reports also described

2058 Cancer Prev Res; 4(12) December 2011 Cancer Prevention Research

Involvement of JMJD2B in Human Carcinogenesis

A 24 p16INK4a 22 JMJD2B CDK6 20 18 16 14 12 10 8

Relative mRNA Relative levels 6 4 2 0 253J A549 HeLa LoVo HFL1 SBC5 Huh-7 H2170 HepG2 SW480 HCT116 SCaBER RERF-LC-AI ACC-LC--319 B C 16 16 20 r = 0.525 r = 0.576 r = 0.976 P P * P * 18 = 0.0530 14 = 0.0295 14 < 0.0001 16 12 12 14 10 10 12 10 8 8 JMJD2B JMJD2B p16INK4a 8 6 6 6 4 4 4 2 2 2 0 0 0 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 1 2 3 4 5 6 7 8 9 10 11 CDK6 CDK6 CDK6

Figure 6. JMJD2B and CDK6 expressions are significantly correlated in various types of cell lines. A, quantitative real-time PCR analysis of one normal cell line (HFL1), 5 lung cancer cell lines (A549, RERF-LC-AI, ACC-LC-319, H2170, and SBC5), 2 bladder cell lines (253J and SCaBER), 3 colorectal cancer cell lines (SW480, HCT116, and LoVo), 2 liver cancer cell lines (HepG2 and Huh-7), and 1 cervical cancer cell line. Expression levels of p16INK4a, JMJD2B and CDK6 were examined. mRNA expression levels were normalized to GAPDH, and values are relative to HFL1. B, correlation among p16INK4a, JMJD2B,andCDK6 expressions in 14 cell lines analyzed by quantitative real-time PCR. Pearson r correlation was used to measure the relationship between the expression of the 2 genes. C, correlation between JMJD2B and CDK6 expressions in 7 cell lines expressing low levels of p16INK4a (HFL1, A549, ACC-LC-319, H2170, 253J, LoVo, and HepG2) analyzed by quantitative real-time PCR. Pearson r correlation was used to measure the expression of relationship between the expression of the 2 genes. the involvement of this gene in human carcinogenesis the mechanism for increased expression of JMJD2B in (27, 35, 36). This study shows that CDK6 is transcription- human cancers, we looked at the genome annotation ally activated by JMJD2B through demethylation of H3K9 across the JMJD2B region, spanning Ch19 4,806,367 to and that elevated CDK6 is likely to promote cell malignan- 5,154,791 using the 1 Mb CGH array data of bladder cy. This is a novel example of how deregulated histone cancer tissues. There are no single gains or amplifications; demethylation contributes to human carcinogenesis. To therefore, we cannot confirm there is a gain relating to date, the tumor suppressor gene p16INK4a has been con- gene dosage in bladder cancer. Beyer and colleagues sidered as an important negative regulator of CDK6 through previously reported that the HIF-1a binds to specific direct interaction (30, 37). We found that an inverse cor- recognition sites in the gene encoding JMJD2B and relation was not observed between p16INK4a and CDK6 induces its expression (17). All human tumors display expressions, and that JMJD2B expression was strongly cor- genomic instability, aberrant transcriptional programs, related to CDK6 expression, especially in cell lines expres- and, very often, contain areas that are insufficiently per- sing low levels of p16INK4a. These results indicate that fused, resulting in a local shortage of nutrients and expression levels of p16INK4a are likely to be a regulator oxygen (hypoxia). This leads to an activation of the of JMJD2B-dependent CDK6 transcriptional activation. Fur- transcription factor HIF, the master regulator of oxygen ther functional analysis is required to clarify the crosstalk homeostasis (38). These results imply that low-oxidation between the JMJD2B-CDK6 and the p16INK4a-RB1 concentration in cells seems to be an important mediator pathways. in activating JMJD2B expression in human carcinogene- Detailed expression analysis showed that expression sis. In addition, the Oncomine database (39) implies that levels of JMJD2B in bladder and lung cancers are signif- JMJD2B is overexpressed in various types of cancer (Sup- icantly higher than those in corresponding normal tis- plementary Fig. S6), so deregulation of JMJD2B plainly sues, and knockdown of JMJD2B resulted in the signifi- contributes to carcinogenesis in a variety of human tumor cant suppression of cancer cell proliferation. To explain types. Importantly, expression levels of JMJD2B in various

www.aacrjournals.org Cancer Prev Res; 4(12) December 2011 2059

Toyokawa et al.

types of normal tissues are significantly low (Supplemen- Disclosure of Potential Conflicts of Interest tary Fig. S7). These results indicate JMJD2B as an ideal target for cancer therapy. Our functional analyses were No potential conflicts of interest were disclosed. mainly based on in vitro models, so additional data from animal experimental models may reinforce the impor- Acknowledgments tance of JMJD2B as a therapeutic target in human cancer. Recently, inhibitors targeting DNA methyltransferases The authors thank Professor Gillian Murphy and the members of her laboratory for substantial technical support and also thank Ms. Yuka Yamane and histone deacetylases were approved by the FDA to and Ms. Haruka Sawada for technical assistance. use for patients with hematologic malignancies (40, 41). Furthermore, a novel inhibitor targeting the histone Grant Support demethylase LSD1 has been developed and a suppressive effect on tumor growth, via reexpression of epigenetically This work was supported by a grant-in aid for young scientists (A; silenced genes in colon cancer cells, was shown (42). On 22681030) from the Japan Society for the Promotion of Science. Our biorepository is supported by funding NIHR and the Cambridge Biomedical the basis of these results, it seems feasible to develop Research Centre. specific inhibitors targeting JMJD2B as anticancer agents. The costs of publication of this article were defrayed in part by the Because the development of inhibitors targeting histone payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate methyltransferases and demethylases has begun, these this fact. reagents should be further validated against the functions of this enzyme, to assure the usefulness of this approach Received June 3, 2011; revised August 16, 2011; accepted September 2, in the near future. 2011; published OnlineFirst September 19, 2011.

References 1. Jenuwein T, Allis CD. Translating the histone code. Science 2001;293: 16. Pollard PJ, Loenarz C, Mole DR, McDonough MA, Gleadle JM, Scho- 1074–80. field CJ, et al. Regulation of Jumonji-domain-containing histone 2. Kouzarides T. Chromatin modifications and their function. Cell 2007; demethylases by hypoxia-inducible factor (HIF)-1alpha. Biochem J 128:693–705. 2008;416:387–94. 3. Peterson CL, Laniel MA. Histones and histone modifications. Curr Biol 17. Beyer S, Kristensen MM, Jensen KS, Johansen JV, Staller P. The 2004;14:R546–51. histone demethylases JMJD1A and JMJD2B are transcriptional 4. Shi Y, Lan F, Matson C, Mulligan P, Whetstine JR, Cole PA, et al. targets of hypoxia-inducible factor HIF. J Biol Chem 2008;283: Histone demethylation mediated by the nuclear amine oxidase homo- 36542–52. log LSD1. Cell 2004;119:941–53. 18. Wallard MJ, Pennington CJ, Veerakumarasivam A, Burtt G, Mills IG, 5. Bannister AJ, Kouzarides T. Reversing histone methylation. Nature Warren A, et al. Comprehensive profiling and localisation of the matrix 2005;436:1103–6. metalloproteinases in urothelial carcinoma. Br J Cancer 2006;94: 6. Whetstine JR, Nottke A, Lan F, Huarte M, Smolikov S, Chen Z, et al. 569–77. Reversal of histone lysine trimethylation by the JMJD2 family of histone 19. Olsburgh J, Harnden P, Weeks R, Smith B, Joyce A, Hall G, et al. demethylases. Cell 2006;125:467–81. Uroplakin gene expression in normal human tissues and locally 7. Tsukada Y, Fang J, Erdjument-Bromage H, Warren ME, Borchers CH, advanced bladder cancer. J Pathol 2003;199:41–9. Tempst P, et al. Histone demethylation by a family of JmjC domain- 20. Kitahara O, Furukawa Y, Tanaka T, Kihara C, Ono K, Yanagawa R, et al. containing proteins. Nature 2006;439:811–6. Alterations of gene expression during colorectal carcinogenesis 8. Hamamoto R, Furukawa Y, Morita M, Iimura Y, Silva FP, Li M, et al. revealed by cDNA microarrays after laser-capture microdissection of SMYD3 encodes a histone methyltransferase involved in the prolifer- tumor tissues and normal epithelia. Cancer Res 2001;61:3544–9. ation of cancer cells. Nat Cell Biol 2004;6:731–40. 21. Yoshimatsu M, Toyokawa G, Hayami S, Unoki M, Tsunoda T, Field HI, 9. Hamamoto R, Silva FP, Tsuge M, Nishidate T, Katagiri T, Nakamura Y, et al. Dysregulation of PRMT1 and PRMT6, Type I arginine methyl- et al. Enhanced SMYD3 expression is essential for the growth of breast transferases, is involved in various types of human cancers. Int J cancer cells. Cancer Sci 2006;97:113–8. Cancer 2011;128:562–73. 10. Silva FP, Hamamoto R, Kunizaki M, Tsuge M, Nakamura Y, Furukawa 22. Toyokawa G, Masuda K, Daigo Y, Cho HS, Yoshimatsu M, Takawa M, Y. Enhanced methyltransferase activity of SMYD3 by the cleavage et al. Minichromosome maintenance protein 7 is a potential therapeutic of its N-terminal region in human cancer cells. Oncogene 2008;27: target in human cancer and a novel prognostic marker of non-small cell 2686–92. lung cancer. Mol Cancer 2011;10:65. 11. Sparmann A, van Lohuizen M. Polycomb silencers control cell fate, 23. Cho HS, Suzuki T, Dohmae N, Hayami S, Unoki M, Yoshimatsu M, et al. development and cancer. Nat Rev Cancer 2006;6:846–56. Demethylation of RB regulator MYPT1 by histone demethylase LSD1 12. Takeshita F, Minakuchi Y, Nagahara S, Honma K, Sasaki H, Hirai K, promotes cell cycle progression in cancer cells. Cancer Res 2011; et al. Efficient delivery of small interfering RNA to bone-metastatic 71:1–6. tumors by using atelocollagen in vivo. Proc Natl Acad Sci U S A 24. Sato N, Yamabuki T, Takano A, Koinuma J, Aragaki M, Masuda K, et al. 2005;102:12177–82. Wnt inhibitor Dickkopf-1 as a target for passive cancer immunother- 13. Schneider R, Bannister AJ, Kouzarides T. Unsafe SETs: histone lysine apy. Cancer Res 2010;70:5326–36. methyltransferases and cancer. Trends Biochem Sci 2002;27: 25. Hayami S, Kelly JD, Cho HS, Yoshimatsu M, Unoki M, Tsunoda T, et al. 396–402. Overexpression of LSD1 contributes to human carcinogenesis through 14. Katoh M. Identification and characterization of JMJD2 family genes in chromatin regulation in various cancers. Int J Cancer 2011;128: silico. Int J Oncol 2004;24:1623–8. 574–86. 15. Fodor BD, Kubicek S, Yonezawa M, O'Sullivan RJ, Sengupta R, Perez- 26. Hayami S, Yoshimatsu M, Veerakumarasivam A, Unoki M, Iwai Y, Burgos L, et al. Jmjd2b antagonizes H3K9 trimethylation at pericentric Tsunoda T, et al. Overexpression of the JmjC histone demethylase heterochromatin in mammalian cells. Genes Dev 2006;20:1557–62. KDM5B in human carcinogenesis: involvement in the proliferation

2060 Cancer Prev Res; 4(12) December 2011 Cancer Prevention Research

Involvement of JMJD2B in Human Carcinogenesis

of cancer cells through the E2F/RB pathway. Mol Cancer 2010; 35. Hu MG, Deshpande A, Enos M, Mao D, Hinds EA, Hu GF, et al. A 9:59. requirement for cyclin-dependent kinase 6 in thymocyte development 27. Senderowicz AM. Small-molecule cyclin-dependent kinase modula- and tumorigenesis. Cancer Res 2009;69:810–8. tors. Oncogene 2003;22:6609–20. 36. Mendrzyk F, Radlwimmer B, Joos S, Kokocinski F, Benner A, Stange 28. Alcorta DA, Xiong Y, Phelps D, Hannon G, Beach D, Barrett JC. DE, et al. Genomic and protein expression profiling identifies CDK6 as Involvement of the cyclin-dependent kinase inhibitor p16 (INK4a) in novel independent prognostic marker in medulloblastoma. J Clin replicative senescence of normal human fibroblasts. Proc Natl Acad Oncol 2005;23:8853–62. Sci U S A 1996;93:13742–7. 37. Canepa ET, Scassa ME, Ceruti JM, Marazita MC, Carcagno AL, Sirkin 29. Fahraeus R, Paramio JM, Ball KL, Lain S, Lane DP. Inhibition of pRb PF, et al. INK4 proteins, a family of mammalian CDK inhibitors with phosphorylation and cell-cycle progression by a 20-residue peptide novel biological functions. IUBMB Life 2007;59:419–26. derived from p16CDKN2/INK4A. Curr Biol 1996;6:84–91. 38. Semenza GL. Regulation of physiological responses to continuous 30. Russo AA, Tong L, Lee JO, Jeffrey PD, Pavletich NP. Structural basis and intermittent hypoxia by hypoxia-inducible factor 1. Exp Physiol for inhibition of the cyclin-dependent kinase Cdk6 by the tumour 2006;91:803–6. suppressor p16INK4a. Nature 1998;395:237–43. 39. Rhodes DR, Kalyana-Sundaram S, Mahavisno V, Varambally R, Yu J, 31. Suzuki-Takahashi I, Higashi H, Yoshida E, Nishimura S, Kitagawa M. Briggs BB, et al. Oncomine 3.0: genes, pathways, and networks in a Effects of exogenous p16INK4a on growth of cells with various status of collection of 18,000 cancer gene expression profiles. Neoplasia cell-cycle regulators. Biochem Biophys Res Commun 1997;234:386–92. 2007;9:166–80. 32. Tomita T. Cyclin-dependent kinase (cdk6) andh p16 in pancreatic 40. Cortez CC, Jones PA. Chromatin, cancer and drug therapies. Mutat endocrine neoplasms. Pathology 2004;36:566–70. Res 2008;647:44–51. 33. Cho HS, Kelly JD, Hayami S, Toyokawa G, Takawa M, Yoshimatsu M, 41. Ma WW, Adjei AA. Novel agents on the horizon for cancer therapy. CA et al. Enhanced expression of EHMT2 is involved in the proliferation of Cancer J Clin 2009;59:111–37. cancer cells through negative regulation of SIAH1. Neoplasia 2011;13: 42. Huang Y, Stewart TM, Wu Y, Baylin SB, Marton LJ, Perkins B, et al. 676–84. Novel oligoamine analogues inhibit lysine-specific demethylase 1 and 34. Grossel MJ, Hinds PW. From cell cycle to differentiation: an expanding induce reexpression of epigenetically silenced genes. Clin Cancer Res role for cdk6. Cell Cycle 2006;5:266–70. 2009;15:7217–28.

www.aacrjournals.org Cancer Prev Res; 4(12) December 2011 2061

The Histone Demethylase JMJD2B Plays an Essential Role in Human Carcinogenesis through Positive Regulation of Cyclin-Dependent Kinase 6

Gouji Toyokawa, Hyun-Soo Cho, Yukiko Iwai, et al.

Cancer Prev Res 2011;4:2051-2061. Published OnlineFirst September 19, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/1940-6207.CAPR-11-0290

Supplementary Access the most recent supplemental material at: Material http://cancerpreventionresearch.aacrjournals.org/content/suppl/2011/09/16/1940-6207.CAPR-11-0290.DC 1

Cited articles This article cites 42 articles, 12 of which you can access for free at: http://cancerpreventionresearch.aacrjournals.org/content/4/12/2051.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://cancerpreventionresearch.aacrjournals.org/content/4/12/2051.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://cancerpreventionresearch.aacrjournals.org/content/4/12/2051. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.