Leukemia (2008) 22, 1529–1538 & 2008 Macmillan Publishers Limited All rights reserved 0887-6924/08 $30.00 www.nature.com/leu ORIGINAL ARTICLE Genome-wide identification of aberrantly methylated promoter associated CpG islands in acute lymphocytic leukemia S-Q Kuang, W-G Tong, H Yang, W Lin, MK Lee, ZH Fang, Y Wei, J Jelinek, J-P Issa and G Garcia-Manero Department of Leukemia, University of Texas MD Anderson Cancer Center, Houston, TX, USA We performed a genome-wide analysis of promoter associated CpG island methylation using methylated CpG island amplification (MCA) coupled to representational differential analysis (RDA) or a DNA promoter microarray in acute lymphoblastic leukemia (ALL). We identified 65 potential targets of methylation with the MCA/RDA approach, and 404 with the MCA/array. Thirty-six (77%) of the genes identified by MCA/RDA were shared by the MCA/array approach. Chromosomal location of these genes was evenly distributed in all autosomes. Functionally, 303 of these genes clustered in 18 molecular pathways. Of the 36 shared genes, 31 were validated and 26 were confirmed as being hypermethylated in leukemia cell lines. Expression analysis of eight of these genes was epigenetically modulated by hypomethylating agents and/or HDAC inhibitors in leukemia cell lines. Subsequently, DNA methylation of 15 of these genes (GIPC2, RSPO1, MAGI1, CAST1, ADCY5, HSPA4L, OCLN, EFNA5, MSX2, GFPT2, GNA14, SALL1, MYO5B, ZNF382 and MN1) was validated in primary ALL samples. Patients with methylation of multiple CpG islands had a worse overall survival. This is the largest published list of potential methylation target genes in human leukemia offering the possibility of performing rational unbiased methylation studies in ALL. Leukemia (2008) 22, 1529–1538; doi:10.1038/leu.2008.130; published online 5 June 2008 Keywords: acute lymphoblastic leukemia; DNA methylation; microarray; epigenetics; MCA; RDA Introduction Acute lymphoblastic leukemia (ALL) is characterized by a block in differentiation of early lymphoid progenitors, which leads to an accumulation of immature lymphoblasts.1 Together with genetic changes, epigenetic alterations such as aberrant DNA methylation of promoter associated CpG islands (CGIs) have a role in downregulation of tumor suppressor genes and in the molecular pathogenesis of ALL.2 Prior data from several laboratories have indicated that aberrant DNA methylation of multiple CGIs is common in ALL3,4 and is associated with poor prognosis.5–7 Furthermore, we have demonstrated that methylation patterns at relapse are stable in a majority of patients with ALL,8 and that the identification of residual methylation levels is associated with a higher risk of relapse in patients in morphological remission (H Yang, submitted). We have also demonstrated that the reactivation of an epigenetically silenced gene, such as p57KIP2, results in selective induction of apoptosis only in cells in which the gene is methylated.2 This body of Correspondence: Dr G Garcia-Manero, Department of Leukemia, University of Texas M.D. Anderson Cancer Center, Box 428, 1515 Holcombe Blvd, Houston, TX 77030, USA. E-mail: [email protected] Received 21 November 2007; revised 23 April 2008; accepted 24 April 2008; published online 5 June 2008 information indicates that epigenetic silencing of tumor suppressor genes is important in ALL. Until recently, one of the limitations of studying DNA methylation has been the process to select gene(s) to study. It is possible that several hundred genes could be aberrantly methylated and silenced in ALL.9 Some of these genes may be bona fide tumor suppressor genes, whose silencing could be essential to the pathogenesis of ALL, whereas others may be silenced as a consequence of global methylation dysregulation and not directly involved in oncogenesis.10 Therefore, an unbiased whole genome approach to identify methylation silenced genes in ALL may reveal novel tumor suppressor genes and enhance our understanding of the role of aberrant gene methylation in this disease. To approach this, we have performed two large-scale genome-wide studies of aberrantly methylated CpG islands in ALL. Materials and methods Cell lines and ALL patient samples The following human leukemia cell lines were studied: of lymphoid origin; MOLT4, Jurkat, Peer, T-ALL1, CEM, J-TAG, BJAB, RS4, ALL1, Raji, REH and Ramos. Of myeloid origin: K562 and BV173, HL60, NB4, THP1, U937, ML1, OCI, HEL, MOLM13 and KBM5R. All were obtained from the American Type Culture Collection and were cultured in RPMI 1640 (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal calf serum (Gemini Bio-Products, West Sacramento, CA, USA) and penicillin–streptomycin (Invitrogen). Cell suspensions from bone marrow aspiration specimens from patients with ALL were obtained prior to any therapy and were stored at established tissue banks at MD Anderson Cancer Center (MDACC) following institutional guidelines. All samples were collected using Ficoll-Paque density centrifugation. DNA was extracted from leukemia cell lines and ALL samples using standard phenol-chloroform methods. Methylated CpG island amplification/representational difference analysis The principle underlying methylated CpG island amplification (MCA) involves amplification of closely spaced methylated SmaI sites (CCCGGG) to enrich for methylated CGIs.11 About 70% of CGIs contain two SmaI sites spaced closely enough (200–1000 bases) to be amplified efficiently by PCR. Representational difference analysis (RDA) is a subtraction hybridization step used to identify CGIs methylated specifically in tester DNA. This procedure was performed as described previously.11 MCA amplicons were generated from a pool of DNA from three patients with ALL (testers). Of importance, the three ALL tester patients were selected because they were refractory to induction DNA Methylation in ALL S-Q Kuang et al 1530 therapy with hyperCVAD chemotherapy.12 These three patients were diploid (Mixed Lineage Leukemia (MLL) and Philadelphia (Ph) chromosome negative). One was a male and two females. Their median age was 33 years (range 27–66). Therefore, this represents a group of patients with very poor prognosis not currently identifiable with current molecular techniques. As control, we used MCA amplicons generated from a pool of DNA from two healthy donors (drivers). These controls were one male and one female, and their ages were 30 and 55 years (median 42.5), respectively (P ¼ 0.48). Briefly, 5 mg of genomic DNA was digested with 100 units of SmaI (New England Biolabs, Ipswich, MA, USA) twice and then further digested with 20 units of XmaI (New England Biolabs). RMCA adaptor was prepared by annealing of oligonucleotides RMCA 24 and RMCA 12, and ligated to the digested DNA fragments using T4 DNA ligase (New England Biolabs). To amplify hypermethylated DNA fragments, PCR was performed using 3 ml of each of the ligation mixture as a template in a 100 ml volume containing 100 pmol RMCA 24 primer, 5 units of Taq DNA polymerase (Life Technologies, Inc., Carlsbad, CA, USA), 4 mM MgCl2, 16 mM of NH4 (SO4)2, 10 mg/ml of bovive serum albumin, and 5% v/v dimethylsulphoxide. The reaction mixture was incubated at 72 1C for 5 min and at 95 1C for 3 min, and then was subjected to 25 cycles of 1 min at 95 1C and 3 min at 77 1C followed by a final extension of 10 min at 77 1C. Representational difference analysis was performed on the test and driver MCA amplicons as described previously.11 Primers used for the first and second rounds of RDA were JMCA24/JMCA12, and NMCA24/NMCA12. After the second round of hybridization and selective amplification, the MCA/ RDA products were cloned into a pCR 4.0-TOPO vector (Invitrogen). To screen for inserts, a total of 800 clones were sequenced at a core facility (MDACC). Sequence homologies and chromosomal localization were identified using the Human Blat program (http://www.genome.ucsc.edu). A list of primers used in this study is shown in Supplementary Table 1. MCA/human proximal promoter DNA microarray The MCA amplicons generated above were labeled with Cy5dUTP (tester) or Cy3-dUTP (driver) according to the manufacturer’s protocol (Invitrogen). Five micrograms of Cy5- and Cy3labeled MCA amplicons were mixed and hybridized to the Agilent human proximal promoter microarray slides. These contain approximately 17 000 of the best defined human transcript associated promoters covering 1.0 kb upstream and þ 0.3 kb downstream from the transcription start site (Agilent Technologies, Santa Clara, CA, USA). The hybridization reaction was carried out in rotisserie hybridization oven at 65 1C, 10 r.p.m. for 40 h and subsequently washed following the Agilent aCGH array hybridization protocol. Hybridized slides were scanned using an Agilent G2565AA scanner (Agilent Technologies). The data from scanned microarray images was extracted using Agilent feature extraction software. Image analysis was done using the Agilent G4477AA Chip Analytics 1.2 software package. Genes corresponding to each promoter DNA spotted on the array were identified and then selected by the following criteria: a tester’s Cy5 signal intensity of more than 2000 and a Cy5 signal intensity ratio of tester versus driver of more than 2. Promoter specific CGI methylation was further determined using an in silico gene-filtering search method. Genes containing CGI in their promoter regions based on results from USCS Human Blat analysis (http://www.genome.ucsc.edu) and having two putative SmaI cutting sites in their CGIs (putative MCA targets) were selected for further validation. Leukemia DNA bisulfite treatment and bisulfite pyrosequencing Bisulfite treatment of genomic DNA was performed as described.13 Please refer to Supplementary Information for further details. Bisulfite sequencing Bisulfite sequencing was performed to confirm the pyrosequencing results in selected samples. Please refer to Supplementary Information for further details. RNA preparation and real-time PCR Methods are described in Supplementary information. 5-aza-20 -deoxycytidine and/or trichostatin A treatment To study the effect of epigenetic modulation, leukemia cell lines were cultured in media supplemented with 5 mmol/l of 5-aza-20 deoxycytidine (Sigma, St Louis, MO, USA) for 4 days, 5 mmol/l of 5-aza-20 -deoxycitidine for 4 days and then 500 nmol/l trichostatin A (TSA) (ICN Biomedicals, Costa Mesa, CA, USA) for the last 24 h or 500 nmol/l TSA for 24 h alone. Cells were split 12– 24 h before treatment. Mock-treated cells were cultured similarly. Methylation loci mapping and molecular pathway analysis The MapChart program (http://www.biometris.wur.nl/UK/Software/MapChart/download) was used to generate a chromosomal map of methylation loci. The ingenuity pathway program (http://analysis.ingenuity.com) was used to cluster gene function and signaling pathway networks. Statistical analysis Statistical analyses were performed using Prism 4 (GraphPad Software, Inc.) or Statistica 6 software (Statistica for Windows version 6.0, StatSoft, Tulsa, OK, USA). The Fisher’s exact test and t-tests were used to compare gene methylation frequencies or expression levels in leukemia cell lines or ALL patients and normal control groups. The Spearman non-parametric test was used to determine correlations. Effect on survival was analyzed as previously described.5 All reported P values were two-sided, and Po0.05 was considered statistically significant. Results Global identification of hypermethylated genes in Ph negative ALL using MCA/RDA and MCA/microarray techniques First, we used the MCA/RDA approach to identify aberrantly methylated CGIs in ALL. We recovered and sequenced 800 clones. Chromosomal locations, gene homology and CGI information of the sequenced clones were analyzed using the Human Blat program (http://www.genome.ucsc.edu). Of the 800 clones, 217 (27%) contained CGIs, 74 (9.3%) were localized within or close to (o200 bp) the promoter region of identifiable genes and 62 of these 74 clones were matched to known human gene sequences (Supplementary Table 2). Of note, five of these clones contained bidirectional promoters.11 These corresponded to either two known genes, or to one known gene, one mRNA/or spliced EST. In total, therefore, 65 genes were identified. Table 1 Methylation profile of leukemia cell lines and normal controls DNA Methylation in ALL S-Q Kuang et al Bisulfite pyrosequencing was carried out to determine the methylation status of selected genes in 23 leukemia cell lines and 10 normal peripheral lymphocytic cells (control). Each column represents a cell line number, and each row indicates a gene identified by both methylated CpG island amplification/representational difference analysis (MCA/RDA) and MCA/DNA promoter microarray. Green box, methylation density o15%; yellow box, methylation density between 15–29.99%; pink grey box, methylation density 30–59.99%; red box, methylation density 460%. Methylation density 415% was used as the cutoff for hypermethylation. Methylation frequency was counted as the percentage of positive methylation numbers versus total numbers studied for each gene. P-values were calculated using a t-test comparing cell line methylatyion versus normal controls. 1531 Leukemia DNA Methylation in ALL S-Q Kuang et al 1532 Validation of methylation target genes in leukemia cell lines Due to limited yield of this approach, we performed a second genome-wide DNA methylation search using the MCA/DNA promoter microarray method. This analysis revealed that 923 (5.4%) of the 17000 promoter specific transcripts featured on the Agilent chip were expected to be hypermethylated in Ph negative ALL samples. We further reduced the number of candidate genes by excluding unknown genes (spliced EST, mRNA and hypothetical proteins), sex chromosomal genes and genes whose putative function were not related to oncogenesis, and focused on 404 autosomal genes (Supplementary Table 3). We then compared the genes identified by the two techniques: MCA/RDA and MCA/microarray. Of the 65 known genes identified by MCA/RDA, only 47 were represented in the array, 36 of them (77%) were found to be hypermethylated by the MCA/microarray approach. Information on these genes is shown in Supplementary Tables 2 and 3. 75 50 25 75 50 25 0 0 60 40 20 0 50 25 40 20 25 GNA14 50 25 100 MYO5B 75 50 25 0 100 80 ZNF38 2 60 40 20 0 Controls Patients Cell lines N=7 N=36 N=23 50 25 Controls Patients Cell lines N=10 N=41 N=23 Patients Cell lines N=44 N=23 100 Methylation (%) N=23 SALL1 75 0 Controls N=9 Methylation (%) N=35 25 Controls Patients Cell lines N=11 N=55 N=23 0 N=9 MSX2 50 N=22 Methylation (%) N=58 75 Controls Patients Cell lines N=23 0 N=11 Methylation (%) Methylation (%) 25 N=35 75 Controls Patients Cell lines 0 Methylation (%) EFNA5 50 100 50 100 N=13 100 75 Controls Patients Cell lines N=7 N=41 N=23 75 25 Controls Patients Cell lines 0 0 GFPT 50 0 Methylation (%) Methylation (%) Methylation (%) OCLN HSPA4L 75 Controls Patients Cell lines N=8 N=56 N=23 100 60 100 100 0 80 25 Controls Patients Cell lines N=8 N=45 N=23 ADCY5 75 Controls Patients Cell lines N=57 N=23 N=11 100 50 0 Methylation (%) Methylation (%) Methylation (%) 100 CAST1 80 MAGI 75 Controls Patients Cell lines N=8 N=46 N=23 Controls Patients Cell lines N=10 N=31 N=23 100 100 RSPO1 Methylation (%) 100 GIPC2 Methylation (%) Methylation (%) 100 Several of the genes identified by MCA/RDA and MCA/array in this study have already been reported in the literature to undergo promoter hypermethylation and gene silencing in various human cancers. These include, among several, p73,3 RUNX3,14 RASSF1,15 MGMT,16 EGR3,17 SLC26A418 and TERT.19 This data validated the techniques used here. To further demonstrate that other genes identified by these techniques were also aberrantly methylated in leukemia, we selected 31 genes that had been identified both by the MCA/RDA and MCA/array. These included: WNT2B, GIPC2, RSPO1, LHX4, TRIMP58, ECRG4, MAGI1, CAST1, ADCY5, HSPA4L, ZFP42, OCLN, EFNA5, MSX2, GFPT2, PNMA2, GNA14, FOXE1, CRTAC1, DPYSL4, FANK1, DCAMKL1, AGC1, SALL1, SLC13A5, PTPRM, MYO5B, FBX027, ZNF382, MN1, PCDH11. MN1 75 50 25 0 Controls Patients Cell lines N=10 N=46 N=23 Controls Patients Cell lines N=11 N=53 N=23 Figure 1 Methylation levels of 15 selected genes identified by both MCA-RDA and MCA/DNA promoter microarray in normal controls, primary acute lymphoblastic leukemia (ALL) patients and leukemia cell lines. DNA methylation was analyzed using bisulfite pyrosequencing. The name of each gene is shown in the upper left corner of each graph. N represents the number of cases studied for each gene. Each graph represents the results of an individual gene indicated in the left upper corner of each graph. Leukemia DNA Methylation in ALL S-Q Kuang et al 1533 Table 2 Methylation profile of primary ALL samples and normal controls Bisulfite pyrosequencing was carried out to determine the methylation status. Each row indicates a gene identified by both methylated CpG island amplification/representational difference analysis (MCA/RDA) and MCA/DNA promoter microarray. White box, not done; green box, methylation density o15%; yellow box, methylation density between 15–29.99%; pink grey box, methylation density 30–59.99%; red box, methylation density 460%. Methylation density 415% was used as the cutoff for hypermethylation. Methylation frequency was counted as the percentage of positive methylation numbers versus total numbers studied for each gene. MCA tester and drivers are the results on the patient samples used for the MCA experiement. P-values were calculated using a t-test comparing cell line methylatyion versus normal controls. Methylation validation was performed using bisulfite pyrosequencing in 23 leukemia cell lines, 10 normal controls and the 3 Ph negative ALL samples originally used for MCA. Methylation results are shown in Table 1, Figure 1 and Table 2. Five genes (TRIMP58, ECRG4, ZFP42, FANK1 and FBXO27) were found methylated in both leukemia cells lines and normal controls samples (Table 1), and were considered as non-informative. The other 26 genes were methylated in at least 2 of the 23 leukemia cell lines but not or rarely in normal control samples (Table 1). Spearman coefficient analysis of the 26 genes in leukemia cells revealed that methylation of these genes was generally correlated with each other (Supplementary Table 4). Analysis of gene methylation in primary ALL We subsequently evaluated the methylation status of 15 randomly selected genes of the 31 genes analyzed in cell lines in 61 pretreatment samples from patients with ALL. Patient characteristics are shown in Supplementary Table 4. Results are shown in Figure 1 and Table 2. Using a methylation cutoff of Leukemia 1534 Leukemia DNA Methylation in ALL S-Q Kuang et al Figure 2 Gene expression in leukemia cell lines and primary leukemia specimens, and effect of epigenetic modulation on gene expression. Gene expression was analyzed with real-time PCR in normal bone marrow (BM), peripheral blood specimens (PB) from normal controls, as well as primary acute lymphoblastic leukemia (ALL) samples (ALL). The figure below each sample or cell line indicates the percent of methylation. Leukemia cells were treated with 5-aza-20 -deoxycytidine (DAC) only, trichostatin A (TSA) only or both (D þ T) as described. Expression of GIPC2 (a), MAGI1 (b), ADCY5 (c), HSPA4L (d), OCLN (e), EFNA5 (f), GNA14 (g) and MYO5B (h) genes was analyzed by real-time PCR. C, control untreated cells. M % methylation density. DNA Methylation in ALL S-Q Kuang et al 1535 15% (as described previously),3 methylation frequencies (percent of patients in which a gene was methylated) ranged from 23 to 100%. Two patients had methylation of 1 gene, 2 of 2 genes, 3 of 3 genes, 14 of 4 genes, 8 of 5 genes, 8 of 6 genes, 11 of 7 genes, 7 of 8 genes, 5 of 9 genes and 3 of 10 genes. No patient had methylation of more than 11 genes. To further confirm the results of the pyrosequencing assays, DNA methylation of RSPO1, ADCY5, HSPA4L and MYO5B was confirmed in one ALL sample, one normal control and the Raji cell line using bisulfite sequencing. Sequencing results were in concordance with the results obtained by pyrosequencing (Supplementary Figure 1). Spearman coefficient analysis of all 15 genes revealed that methylation of these genes was significantly correlated with each other (Supplementary Table 4). We also observed an excellent correlation between cell line methylation results and patients (data not shown). HSPA4L, OCLN, EFNA5, GNA14 and MYO5B was restored in methylated leukemia cell lines by either 5-aza-20 -dexocytidine with or without TSA (Figure 2), a phenomenon associated with gene demethylation (data not shown). These data indicates that the DNA methylation and histone deacetylation of the genes analyzed here are associated with suppressed gene expression in leukemia cell lines. Furthermore, a relationship was also observed between methylation and gene expression in primary ALL samples (Figure 2). Physical mapping of methylated loci and pathway analysis of microarray data The 404 methylated genes identified by MCA/microarray were analyzed according to their chromosomal location and found to be distributed evenly in the 22 autosomes. The physical distribution of these loci was further mapped using MapChart program (Figure 3). We also performed pathway and functional analysis of these genes using the Ingenuity Pathway Analysis program (http://analysis.ingenuity.com). The initial analysis divided 303 out of the 404 genes into 29 functional networks, with the majority of genes clustered into 18 networks (Table 3). Genes clustered in these networks are involved in a wide variety Gene expression and effect of epigenetic modulation To examine the role of the DNA methylation in the control of gene expression, eight genes were studied.20–27 Leukemia cell lines were treated with 5-aza-20 -deoxycytidine with or without TSA. In general, expression of GIPC2, MAGI1, ADCY5, 1 2 29(7%) 31(8%) TP73 PER3 SPEN NBL1 EPHB2 E2F2 RUNX3 SESN2 BMP8A BMP8B TRIT1 PPT1 RNF11 DHCR24 GIPC2 SOX11 RRM2 ROCK2 ALK CYP1B1 MTA3 SIX2 MSH6 VRK2 REL MXD1 TGFA BIN1 ARNT EFNA4 MAPBPIP ARHGEF11 IGSF4B IGSF8 CREG1 LHX4 HOXD13 NAB1 GLS TMEFF2 SUMO1 FEV INHA IRS1 HRB NCL GBX2 HES6 PPP1R7 BOK HLX1 WNT3A ARF1 5 6 7 8 9 10 11 25(6%) 15(4%) 25(6%) 24(6%) 28(7%) 21(5%) 19(5%) DGKQ TACC3 CRMP1 HTRA3 LGI2 RBPSUH OGG1 IRAK2 TIMP4 THRB CTDSPL TSP50 MAP4 CDC25A SEMA3F HYAL2 RASSF1 CACNA2D2 ACY1 ALAS1 BAP1 IL17RB CAST PTPRG MAGI1 IGSF11 GATA2 TRH SOX14 FAM62C GMPS IL12A MAL STARD7 MAP4K4 ECRG4 GSTM1 WNT2B SYCP1 4 26(6%) 3 26(6%) KDR IGFBP7 EPHA5 CXCL6 CXCL2 FGF5 TSPAN5 PITX2 TIFA FGF2 NUDT6 HSPA4L GAB1 EDNRA NR3C2 SAP30 HAND2 VEGFC FAT ZFP42 TERT BASP1 AMACR SKP2 OCLN MSH3 RASA1 EFNA5 TNFAIP8 HSD17B4 LOX FBN2 IRF1 SPOCK MATR3 DND1 HDAC3 PCDH1 KIBRA NKX2-5 STC2 MSX2 NSD1 GRK6 GFPT2 DUSP22 IRF4 FOXQ1 FOXC1 TXNDC5 BAT2 COL11A2 PACSIN1 PTK7 GTPBP2 HTR1B ME1 PTPRK SLC22A3 PDCD2 MAFK MAD1L1 RBAK ACTB DFNA5 HOXA5 PDE1C CAMK2B BCL7B CACNA2D1 AKAP9 CDK6 NPTX2 GNB2 CUTL1 SVH SLC26A4 NRCAM CAV1 SMO MEST PODXL NOM1 HLXB9 PTPRN2 MSRA GATA4 EGR3 TNFRSF10A LOXL2 EXTL3 NRG1 ZNF703 SFRP1 SDCBP CA8 NCOA2 E2F5 SDC2 TSPYL5 RPL30 RSPO2 EXT1 NOV ADCY8 BAI1 GML SCRIB PLEC1 PSIP1 MTAP CDKN2A FANCG TPM2 MELK GNA14 TLE4 UBQLN1 GAS1 CTSL FANCC ABCA1 KLF4 TXN DBC1 GSN LHX6 HSPA5 AK1 PKN3 BARHL1 TSC1 WDR5 RXRA LHX3 TRAF2 COBRA1 PFKP IL15RA NMT2 VIM EPC1 ITGB1 NRP1 SLC18A3 NCOA4 UBE2D1 RHOBTB1 HK1 DDIT4 PAPSS2 CRTAC1 BTRC NEURL SLC18A2 FANK1 MGMT DPYSL4 USH1C PTPRJ UBE2L6 FADS1 STIP1 RPS6KA4 MEN1 EHD1 SYVN1 RELA CFL1 PACS1 MTL5 FGF3 SERPINH1 PGR PPP2R1B ZBTB16 ST14 12 13 14 15 16 17 18 19* 20 21 22 14(3.5%) 10(2.5%) 9(2.2%) 14(3.5%) 15(4%) 13(3.2%) 8(2%) 22(5.4%) 11(2.7%) 6(15%) 6(15%) WNK1 EPS8 BCAT1 COL2A1 ACVRL1 CSAD HOXC10 ERBB3 RASSF3 DYRK2 NTN4 APAF1 CRY1 RASAL1 TNFRSF19 APRIN DCAMKL1 RCBTB2 RCBTB1 GPC6 CLDN10 DZIP1 LIG4 IRS2 KIAA0323 SNX6 SSTR1 MGAT2 BMP4 PRIMA1 BCL11B CCNK YY1 THBS1 MAPK6 ALDH1A2 USP3 MEGF11 TLE3 PKM2 COX5A IMP3 ARNT2 BNC1 NTRK3 IQGAP1 NR2F2 BAIAP3 TRAF7 PKMYT1 CREBBP EEF2K MAPK3 CORO1A VKORC1 NKD1 SALL1 HERPUD1 CDH5 NFATC3 MARVELD3 APRT PLD2 SLC13A5 EFNB3 PER1 FLCN SREBF1 IGFBP4 ADAM11 DLX4DLX4 COL1A1 SFRS1 SUMO2 NPTX1 MRCL3 PTPRM ROCK1 MIB1 RBBP8 DSC3 PSTPIP2 FVT1 PALM PRG2 MUM1 GNA11 MATK ZBTB7A ICAM1 ICAM5 KEAP1 RAB3A GDF15 CCNE1 CEBPA ZNF382 SIX5 HIF3A SEPW1 ATF5 KLK10 PPP2R1A KLP1 TRIM28 RASSF2 PAX1 TP53INP2 RBL1 SRC ADA NCOA3 CEBPB PFDN4 GNAS TNFRSF6B TPTE CXADR GABPA OLIG2 ETS2 UBE2G2 MN1 NF2 TST PGEA1 L3MBTL2 TOB2 Figure 3 Chromosomal localization of 404 genes identified by MCA/DNA promoter microarray. Each chromosome is numbered at the top. The figure below the chromosome number represents the number and percentage of methylated loci identified in each chromosome. Gene names are indicated beside each methylated locus. Leukemia DNA Methylation in ALL S-Q Kuang et al 1536 Table 3 ID Genes 1 ADA, AMACR, CXCL2, EPHB2, ETS2, FEV, FGF5, FOXC1, GDF15, ICAM1, IL12A, IL15RA, IRF1, IRF4, L3MBTL2, MAFK, MAP4K4, MEN1, MXD1, NCL, NF2, RASA1, REL, RELA, SDC2, SDCBP, TACC3, TERT, TNFAIP8, TNFRSF19, TNFRSF10A (includes EG:8797), TNFRSF6B, TRAF2, TXN, ZBTB7A ACTB, ARF1, BIN1, CCNE1, CDH5, CFL1, CRY1, EGR3, EPS8, GAB1 (includes EG:2549), GABPA, GATA2, GSN, HLXB9, IGSF8 (includes EG:93185), ITGB1, KDR, KEAP1, MATK, NRG1, NRP1, PER1, PER3, PITX2, PLD2, PTPRG, PTPRJ, RAB3A, SEMA3F, SFRS1, SKP2, SRC, SUMO2 (includes EG:6613), TOB2, VEGFC AK1, CDC25A, CDK6, CDKN2A, CDKN2D, COL11A2, COL1A1, COL2A1, CREG1, E2F2, E2F5, FGF2, FGF3, IGFBP4, IGFBP7, LOX, MELK, MGMT, MSX2, NBL1, NOV, OLIG2, PLEC1, PPP2R1A, PPP2R1B (includes EG:5519), RBL1, ROCK2, RPL30, RRM2, SFRP1, TGFA, THBS1, TP73, VIM, YY1 ALDH1A2, CEBPA, CEBPB, CREBBP, CSAD, CUTL1, GATA4, GNA11, HAND2, HOXA5, HOXD13, HSPA5, INHA, IRS2, MAPK3, ME1, MN1, MTA3, NCOA2, NCOA3, NCOA4 (includes EG:8031), NFATC3, NR3C2, PCDH1, PGR, RPS6KA4, RUNX3, RXRA, SCRIB, SREBF1, SUMO1, SYCP1, THRB, TRAF7, TRH ABCA1, ACVRL1, ADA, APAF1, CCNK, COX5A, CRMP1, CTSL, DHCR24, DPYSL4, FADS1, GDF15, LOXL2, MAPK6, PFKP, PKM2, TPM2 CAV1, CXADR, CXCL6, FLCN, GNA14, GNB2, HDAC3, MAD1L1, MSRA, MTAP, OGG1, PALM, PPP1R7, SAP30, TLE3, ZBTB16, ZFP42 ARNT, ARNT2, BASP1, BMP4, CLDN10, EEF2K, GAS1, KLF4, MRCL3, NEURL, NKX2-5, NMT2, SERPINH1, STIP1, UBE2G2, UBE2L6, WNT3A ACY1, CYP1B1, DLX4, EFNA4, EFNA5, EFNB3 (includes EG:1949), EPHA5, ERBB3, GBX2, NAB1, NCL, NTRK3, PKMYT1, SLC18A3, SMO, SYVN1, BAP1, BCL11B, DDIT4, GSTM1, MAP4, MIB1, NR2F2, PACSIN1, PODXL, PSIP1, RNF11, ST14, TRIM28, UBE2D1, UBQLN1 FANCC, FANCG, GDF15, GLS, HES6, HSD17B4, HSPA4L, MATR3, NSD1, RBBP8, RBPSUH, SALL1, SPEN, SSTR1, TSC1 ABCA1, DSC3, EHD1, EXT1, GFPT2, HIF3A, HLX1, NRP1, RCBTB2, SFRP1, SLC26A4, SPOCK1, TIMP4, TST 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 APRT, CAST (includes EG:26059), DCAMKL1, EDNRA, EPS8, GRK6, HRB, MAL, MATK, NPTX2, PTPRK, SNX6, WNK1, ZNF703 GIPC2, KDR, MAGI1, NCOA2, NPTX1, NRCAM (includes EG:4897), PACS1, PDE1C, PGEA1, PTK7, SIX2, SIX5, SOX11 ARHGEF11, BCAT1, BTRC, DGKQ, FBN2, OCLN, PDCD2, PFDN4, ROCK1, SESN2, SIAHBP1, TLE4, ZBTB7A BAI1, BAIAP3 (includes EG:8938), BOK, CACNA2D2, HERPUD1, HYAL2, ICAM5, IQGAP1 (includes EG:8826), LHX4, LHX6, PTPRM, RASSF1, TNFRSF10A (includes EG:8797) AKAP9, ALAS1, ATF5, BCL7B, DYRK2, GABPA, GNAS (includes EG:2778), HK1, MEST, MTL5, PRIMA1, PTPRN2, SLC18A2 19 20 21 COBRA1, CTDSPL, EPC1, GMPS, LHX3, LIG4, MSH3, MSH6, PAX1, RBAK, RBBP8, WDR5 ALK, BNC1, CACNA2D1, CAMK2B, CORO1A, DUSP22, GNAS (includes EG:2778), IL17RB, IRAK2, IRS1, RASSF2, WNT2B IMP3 BMP8A FAT 22 23 24 25 26 27 28 29 IGSF4B PAPSS2 RASAL1 MAPBPIP (includes EG:28956) HTR1B PKN3 VKORC1 USH1C 18 Leukemia Functional class of genes identified by MCA/DNA promoter Focus genes Top functions 35 Cellular Growth and proliferation, cell death, cancer 35 Cellular growth and proliferation, cellular movement, cell morphology 35 Cell cycle, cellular growth and proliferation, cancer 35 Gene expression, viral function, developmental disorder 17 Inflammatory disease, respiratory disease, organismal development 17 Cancer, cellular growth and proliferation, hematological disease 17 Gene expression, connective tissue development and function, skeletal and muscular system development and function Nervous system development and function, tissue development, cellular development 16 15 15 14 14 13 13 13 13 12 12 1 1 1 1 1 1 1 1 1 1 1 Protein degradation, protein synthesis, posttranslational modification Genetic disorder, hematological disease, cell cycle Cardiovascular system development and function, embryonic development, tissue development Amino-acid metabolism, Post-translational modification, small molecule biochemistry Skeletal and muscular system development and function, tissue development, cancer Gene expression, cancer, reproductive system disease Cell death, renal and Urological disease, embryonic development Nervous system development and function, skeletal and muscular system development and function, cell death DNA replication, recombination and repair, cancer, gene expression Cancer, cell signaling, post-translational modification RNA post-transcriptional modification Cell morphology, gene expression Cellular compromise, Nervous system development Cancer, cell-to-cell signaling Developmental disorder, genetic disorder Cellular development Cancer, cell death, reproductive system disease Neurological disease, organismal injury Cellular assembly and organization Genetic disorder, hematological disease, Genetic disorder, neurological disease DNA Methylation in ALL S-Q Kuang et al 1537 Figure 4 Survival analysis of patients with acute lymphoblastic leukemia (ALL) based on number of methylated genes. Patient were grouped in three different sets: methylation of 0–3 genes, 4–7 genes and 8–10 genes. These could include any the genes studied. of biological functions including cell growth and differentiation, cell death, gene regulation, cell cycle, DNA replication and repair, signal transduction, transport and metabolism. Exploratory analysis of effect of DNA methylation in survival Prior studies have indicated a relationship between hypermethylation of multiple genes and worse survival.5,6 To explore this effect, we analyzed the impact of methylation of gene(s) on overall survival. To do this analysis, patients were divided into three groups: those with methylation (X15%) of 0–3 genes, 4–7 genes and 8–10 genes. The median survival of patients in the first group of patients had not been reached, whereas it was 173 weeks for the second group and 68 weeks for the third group (P ¼ 0.012) (Figure 4). Discussion Using MCA/RDA and an MCA/DNA promoter microarray, we identified a large cohort of genes that are potential targets of aberrant DNA methylation in ALL. Using MCA/RDA we identified 65 targets of aberrant DNA methylation, and with the MCA/array 404. Thirty-one of the genes (77%) were identified by both techniques. Of these 31 genes, 26 genes were found to be informative in cell lines (not methylated in normal controls), and 15 genes were confirmed to be methylated in primary ALL samples. None of the genes that were validated as informative in cell lines was discarded by the analysis of primary ALL samples. Furthermore, a number of these genes had been already shown to be hypermethylated in cancer. This evidence further demonstrates the validity of the assays used here. In the current study, and as a proof of principle, we validated the methylation of 31 genes not previously reported to be hypermethylated in ALL. Methylation analysis excluded TRIMP58, ECRG4, ZFP42, FANK, DHX32 and FBXO27 as they were found to be methylated in normal controls. The other 26 genes were found to be specifically hypermethylated in leukemia cell lines, and 15 of them were further confirmed to be hypermethylated in a panel of 61 ALL patients. Methylation of these genes was highly correlated both in cell lines and primary samples. Thus, confirming the presence of a hypermethylator phenotype in ALL, as described previously.3 To demonstrate that the level of methylation identified here was associated with gene expression inactivation, we studied the effect of treating cell lines with a hypomethylating agent with or without a histone deacetylase inhibitor. As expected, treatment of cell lines with these epigenetic modulators resulted, in general, in gene reactivation, an effect that was mediated by induction of DNA hypomethylation. An inverse relationship between methylation and gene expression was also observed in primary ALL samples. Targets of aberrant DNA methylation were evenly distributed in all autosomes with no evidence of particular chromosomal hot spots for methylation. The function of these genes clustered in 29 different networks involved in multiple cellular functions. Finally, and as described in other previous studies,5,6 methylation of multiple genes was associated with a worse outcome. Those patients with methylation of more than four genes appeared to have a significant worse outcome when treated with hyperCVAD based chemotherapy. The implications of this work are very significant. First, we have made public the largest list of potential targets of aberrant DNA methylation in ALL. This provides a significant source to perform further studies of DNA methylation in ALL, and potentially other leukemias. We provide information that these genes cluster together in specific functional networks and chromosomal locations. This provides the opportunity to investigate specific functions of genes located in specific chromosomal locations. It should be noted that this is not the first large-scale analysis of methylation in ALL. Recently, Taylor et al.28 reported on 262 methylated genes using a different CGI microarray method. We compared the gene lists identified by our method and their CGI array and found that only nine genes (NASP, FEV, HRB, EFNA5, GTPBP2, RPIB9, CAV1, NOV and SMC2L1) were concordant. This probably relates to the different methods used, in particular the regions analyzed, and the data is probably complementary and it argues that potentially other targets of methylation exist that have not been identified by these two studies combined. It also should be noted that we have identified a number of genes methylated in normal samples. The significance of this observation is not known at the present time. In summary, we have identified a group of aberrantly methylated genes that can be used as potential epigenetic biomarkers of ALL. We have demonstrated that ALL exhibits widespread epigenetic alterations. Although probably not all methylation events identified here are relevant, clinically or functionally, in ALL directly, the study of this group of genes may provide further insights into the pathogenesis of ALL. Acknowledgements This work was supported by a Leukemia SPORE (P50CA100632) Career Development Award to S-Q K, and NCI grants CA100067 and CA105771, the Leukemia & Lymphoma Society of America and a MD Anderson Cancer Center’s Physician-Scientist Award from the Commonwealth Cancer Foundation for Research (all to G G-M). References 1 Pui CH, Relling MV, Downing JR. Acute lymphoblastic leukemia. N Engl J Med 2004; 350: 1535–1548. 2 Kuang SQ, Ling X, Sanchez-Gonzalez B, Yang H, Andreeff M, Garcia-Manero G. Differential tumor suppressor properties and transforming growth factor-beta responsiveness of p57KIP2 in Leukemia DNA Methylation in ALL S-Q Kuang et al 1538 3 4 5 6 7 8 9 10 11 12 13 14 15 16 leukemia cells with aberrant p57KIP2 promoter DNA methylation. Oncogene 2007; 26: 1439–1448. Garcia-Manero G, Daniel J, Smith TL, Kornblau SM, Lee MS, Kantarjian HM et al. DNA methylation of Multiple Promoterassociated CpG Islands in Adult Acute Lymphocytic Leukemia. Clin Cancer Res 2002; 8: 2217–2224. Gutierrez MI, Siraj AK, Bhargava M, Ozbek U, Banavali S, Chaudhary MA et al. Concurrent methylation of multiple genes in childhood ALL: correlation with phenotype and molecular subgroup. Leukemia 2003; 17: 1845–1850. Shen L, Toyota M, Kondo Y, Obata T, Daniel S, Pierce S et al. Aberrant DNA methylation of p57KIP2 identifies a cell-cycle regulatory pathway with prognostic impact in adult acute lymphocytic leukemia. Blood 2003; 101: 4131–4136. Roman-Gomez J, Jimenez-Velasco A, Castillejo JA, Agirre X, Barrios M, Navarro G et al. Promoter hypermethylation of cancerrelated genes is a strong independent prognostic factor in acute lymphoblastic leukemia. Blood 2004; 104: 2492–2498. Bueso-Ramos C, Xu Y, McDonnell TJ, Brisbay S, Pierce S, Kantarjian H et al. Protein expression of a triad of frequently methylated genes, p73, p57Kip2, and p15, has prognostic value in adult acute lymphocytic leukemia independently of its methylation status. J Clin Oncol 2005; 23: 3932–3939. Garcia-Manero G, Bueso-Ramos C, Daniel J, Williamson J, Kantarjian HM, Issa JP. DNA methylation patterns at relapse in adult acute lymphocytic leukemia. Clin Cancer Res 2002; 8: 1897–1903. Herman JG, Baylin SB. Gene silencing in cancer in association with promoter hypermethylation. N Engl J Med 2003; 349: 2042–2054. Rauch T, Li H, Wu X, Pfeifer GP. MIRA-assisted microarray analysis, a new technology for the determination of DNA methylation patterns, identifies frequent methylation of homeodomain-containing genes in lung cancer cells. Cancer Res 2006; 66: 7939–7947. Toyota M, Ho C, Ahuja N, Jair KW, Li Q, Ohe-Toyota M et al. Identification of differentially methylated sequences in colorectal cancer by methylated CpG island amplification. Cancer Res 1999; 59: 2307–2312. Garcia-Manero G, Kantarjian HM. The hyper-CVAD regimen in adult acute lymphocytic leukemia. Hematol Oncol Clin North Am 2000; 14: 1381–1396, x–xi. Garcia-Manero G, Kantarjian HM, Sanchez-Gonzalez B, Yang H, Rosner G, Verstovsek S et al. Phase 1/2 study of the combination of 5-aza-20 -deoxycytidine with valproic acid in patients with leukemia. Blood 2006; 108: 3271–3279. Li QL, Ito K, Sakakura C, Fukamachi H, Inoue K, Chi XZ et al. Causal relationship between the loss of RUNX3 expression and gastric cancer. Cell 2002; 109: 113–124. Dammann R, Li C, Yoon JH, Chin PL, Bates S, Pfeifer GP. Epigenetic inactivation of a RAS association domain family protein from the lung tumour suppressor locus 3p21.3. Nat Genet 2000; 25: 315–319. Esteller M, Hamilton SR, Burger PC, Baylin SB, Herman JG. Inactivation of the DNA repair gene O6-methylguanine-DNA 17 18 19 20 21 22 23 24 25 26 27 28 29 methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res 1999; 59: 793– 797. Matsubayashi H, Canto M, Sato N, Klein A, Abe T, Yamashita K et al. DNA methylation alterations in the pancreatic juice of patients with suspected pancreatic disease. Cancer Res 2006; 66: 1208–1217. Xing M, Tokumaru Y, Wu G, Westra WB, Ladenson PW, Sidransky D. Hypermethylation of the Pendred syndrome gene SLC26A4 is an early event in thyroid tumorigenesis. Cancer Res 2003; 63: 2312–2315. Liu L, Saldanha SN, Pate MS, Andrews LG, Tollefsbol TO. Epigenetic regulation of human telomerase reverse transcriptase promoter activity during cellular differentiation. Genes Chromosomes Cancer 2004; 41: 26–37. Kirikoshi H, Katoh M. Molecular cloning and characterization of human GIPC2, a novel gene homologous to human GIPC1 and Xenopus Kermit. Int J Oncol 2002; 20: 571–576. Kotelevets L, van Hengel J, Bruyneel E, Mareel M, van Roy F, Chastre E. Implication of the MAGI-1b/PTEN signalosome in stabilization of adherens junctions and suppression of invasiveness. FASEB J 2005; 19: 115–117. Ludwig MG, Seuwen K. Characterization of the human adenylyl cyclase gene family: cDNA, gene structure, and tissue distribution of the nine isoforms. J Recept Signal Transduct Res 2002; 22: 79–110. Takahashi H, Furukawa T, Yano T, Sato N, Takizawa J, Kurasaki T et al. Identification of an overexpressed gene, HSPA4 L, the product of which can provoke prevalent humoral immune responses in leukemia patients. Exp Hematol 2007; 35: 1091–1099. Ichimiya S, Kojima T. Cellular networks of human thymic medullary stromas coordinated by p53-related transcription factors. J Histochem Cytochem 2006; 54: 1277–1289. Shi H, Guo J, Duff DJ, Rahmatpanah F, Chitima-Matsiga R, Al-Kuhlani M et al. Discovery of novel epigenetic markers in non-Hodgkin’s lymphoma. Carcinogenesis 2007; 28: 60–70. Liu AM, Wong YH. Activation of nuclear factor {kappa}B by somatostatin type 2 receptor in pancreatic acinar AR42J cells involves G{alpha}14 and multiple signaling components: a mechanism requiring protein kinase C, calmodulin-dependent kinase II, ERK, and c-Src. J Biol Chem 2005; 280: 34617–34625. McGurk L, Tzolovsky G, Spears N, Bownes M. The temporal and spatial expression pattern of myosin Va, Vb and VI in the mouse ovary. Gene Expr Patterns 2006; 6: 900–907. Taylor KH, Kramer RS, Davis JW, Guo J, Duff DJ, Xu D et al. Ultradeep bisulfite sequencing analysis of DNA methylation patterns in multiple gene promoters by 454 sequencing. Cancer Res 2007; 67: 8511–8518. Shu J, Jelinek J, Chang H, Shen L, Qin T, Chung W et al. Silencing of bidirectional promoters by DNA methylation in tumorigenesis. Cancer Res 2006; 66: 5077–5084. Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu) Leukemia
© Copyright 2024 Paperzz