Document GKQGOz46VXoXwrkE3XJmbJV9Y

[CANCER RESEARCH 64, 2000 2006, March 15, 2004] Hypermethylation of the 5 CpG Island of the FHIT Gene Is Associated with Hyperdiploid and Translocation-Negative Subtypes of Pediatric Leukemia Shichun Zheng,1 Xiaomei Ma,2 Luoping Zhang,2 Laura Gunn,2 Martyn T. Smith,2 Joseph L. Wiemels,1 Kenneth Leung,3 Patricia A. Buffler,2 and John K. Wiencke1 1Laboratory for Molecular Epidemiology, Department of Epidemiology and Biostatistics, University of California at San Francisco, San Francisco, California; 2School of Public Health, University of California at Berkeley, Berkeley, California; and 3Kaiser Permanente, Pediatric Hematology/Oncology, San Francisco, California ABSTRACT human FHIT gene is a member of the histidine triad gene family (22, The human FHIT (fragile histidine triad) gene is a putative tumor suppressor gene located at chromosome region 3p14.2. Previous studies have shown that loss of heterozygosity, homozygous deletions, and abnormal expression of the FHIT gene are involved in several types of human malignancies. A CpG island is present in the 5 promoter region of the FHIT gene, and methylation in this region correlates with loss of FHIT expression. To test whether aberrant methylation of the FHIT gene may play a role in pediatric leukemia, we assessed the FHIT methylation status of 10 leukemia cell lines and 190 incident population-based cases of childhood acute lymphocytic and myeloid leukemias using methylationspecific PCR. Conventional and fluorescence in situ hybridization cytogenetic data were also collected to examine aneuploidy, t(12, 21), and other chromosomal rearrangements. Four of 10 leukemia cell lines (40%) and 52 of 190 (27.4%) bone marrows from childhood leukemia patients demonstrated hypermethylation of the promoter region of FHIT. Gene expression analyses and 5-aza-2-deoxycytidine treatment showed that promoter hypermethylation correlated with FHIT inactivation. Among primary leukemias, hypermethylation of FHIT was strongly correlated with acute lymphoblastic leukemia (ALL) histology (P 0.008), high hyperdiploid (P < 0.0001), and translocation-negative (P < 0.0001) categories. Hyperdiploid B-cell ALLs were 23-fold more likely to be FHIT methylated compared with B-cell ALL harboring TEL-AML translocations. FHIT methylation was associated with high WBC counts at diagnosis, a known prognostic indicator. These results suggest that hypermethylation of the promoter region CpG island of the FHIT gene is a common event and may play an important role in the etiology and pathophysiology of specific cytogenetic subtypes of childhood ALL. 23), the function of which remains unknown. FHIT knockout mice have an increased susceptibility to spontaneous tumors as well as being exquisitely sensitive to carcinogens (24, 25), and transfection of FHIT into tumorigenic cell lines inhibits tumorigenicity in mice (26). All of these data are compatible with the idea that FHIT is a tumor suppressor gene. The FHIT protein may function in the metabolism of polyphosphorylated diadenosine (e.g., Ap3A) substrates that can be induced in hematopoietic cells by cytokines and which may participate in intracellular signaling pathways and mediate antiviral mechanisms (2730). Other observations indicate that the activation of caspase-8 was correlated with FHIT-mediated apoptosis, which suggests that FHIT might exert a proapoptotic function through a caspase-mediated pathway (31, 32). The promoter region around exon 1 of the FHIT gene contains a CpG island that has been shown to be hypermethylated in esophageal, lung, breast, prostate, bladder, cervical, and oral cancers (3338). Aberrant methylation of the promoter region of the FHIT gene is strongly associated with gene inactivation as indicated by Northern blot, reverse transcription (RT)-PCR, and immunostaining analyses (33, 34). Hypermethylated cells can be demethylated and induced to re-express the FHIT gene products after treatment with 5-aza-2deoxycytidine. Also of interest is the finding that FHIT alterations, including methylation, may be associated with environmental exposures (34). Links between FHIT methylation and environmental exposures could make it a useful marker for epidemiological research exploring possible causal pathways in pediatric leukemia. Leukemia INTRODUCTION cases examined in this study are participants in a population-based etiological study of childhood leukemia in Northern California. We Pediatric leukemias are a heterogeneous group of malignancies. Major cytogenetic subgroups exist that are characterized by structural chromosomal abnormalities leading to the synthesis of oncogenic fusion proteins and other subgroups with nonrandom gains or losses in measured DNA methylation in the promoter region of the FHIT gene in 190 consecutive cases of childhood leukemia and assessed the relationship between FHIT hypermethylation and clinicopathological and cytogenetic parameters. chromosome number (1, 2). Some of these cytogenetic subtypes are mutually exclusive, indicating distinct pathogenetic pathways. In addition, a significant fraction of acute leukemias in children display apparently diploid karyotypes. Among the mechanisms considered in the pathogenesis of leukemias without structural alterations are aberrant methylation and associated transcriptional silencing of putative tumor suppressor genes. Several targets of epigenetic silencing have been identified in leukemia (313). Here we have focused on the fragile histidine triad (FHIT) locus because multiple reports indicate aberrant expression of the FHIT gene in leukemic cells (14 21). The Received 8/4/03; revised 11/25/03; accepted 1/9/04. Grant support: NIH grants and National Institutes of Environmental Health Sciences Grants P42ES04705, R01 ES 06717, and R01 ES 009137. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Note: X. Ma is currently at the Department of Epidemiology and Public Health, Yale University School of Medicine, 60 College Street, P.O. Box 208034, New Haven, CT 06520-8034. Requests for reprints: John K. Wiencke, Laboratory for Molecular Epidemiology, University of California San Francisco, San Francisco, CA 94143-0560. Phone: (415) 476-3059; Fax: (415) 502-7411; E-mail:wiencke@itsa.ucsf.edu. MATERIALS AND METHODS Cell Lines and DNA Isolation. Ten human leukemia cell lines (Molt-4, KG1A, Jurkat, RCH, Reh, Blin, NALM, 697, K562, and HL-60) from the American Type Culture Collection were maintained in RPMI 1640 supplemented with 10% fetal bovine serum (Hyclone, Logan, UT) and were grown at 37C in 5% CO2. DNA was isolated from cultured cells using QIAamp DNA Mini kit (Qiagen Inc.) and quantified by fluorometery. Study Population. Included in the FHIT methylation analysis were a total of 190 incident cases of childhood leukemia, patients who were enrolled in the Northern California Childhood Leukemia Study (NCCLS) and our studies used cryopreserved pretreatment bone marrow aspirates obtained from the clinical center that first diagnosed the case. A detailed description of this populationbased study design can be found elsewhere (39 41). These patients were diagnosed between August 1995 and July 2000 in nine major clinical centers in the San Francisco Bay Area and Central Valley of California and were representative of the large case population (88% of all newly diagnosed cases were included in this study). One hundred and fifty-six of the patients were diagnosed with acute lymphoblastic leukemia (ALL), 32 with acute myeloid leukemia (AML), and 2 with chronic myeloid leukemia. One hundred and four (54.7%) of the patients were male, and 86 (45.3%) were female. The 2000 DNA METHYLATION PROFILE IN PEDIATRIC LEUKEMIA Table 1 Clinical and demographic characteristics of leukemia patients Characteristics N (number of patients) Total Gender Male Female Age (years) 2 25 614 Subtype ALLa AML CML 190 104 86 19 95 76 156 32 2 a ALL, acute lyphoblastic leukemia; AML, acute myeloid leukemia; CML, chronic myeloid leukemia. age of the patients ranged from 0.2 to 14.9 years, and the mean and median ages were 6.1 and 5.0 years, respectively. Of all of the cases, 48.9% were non-Hispanic White, 33.0% were Hispanic, 5.1% were African American, 4.0% were Asian American, and 9.1% belonged to other racial/ethnic groups. The clinical characteristics of leukemia patients are given in Table 1. Immunophenotype, cell counts at diagnosis, and diagnostic cytogenetics were abstracted from patient records and merged with results from the epidemiological, fluorescence in situ hybridization (FISH), and methylation analyses. More detailed information about distribution of the immunophenotypes and cytogenetic abnormalities as well as comparison to other population-based studies can be found in Table 2. Data on p15INK4b methylation were also available for these patients. Sample Processing of Bone Marrow and Bisulfite Modification of DNA. Because bone marrow samples are very limited and there is DNA loss in any extraction procedure, a "One Step" sample processing and DNA modification technique was developed in this study to maximize DNA yield from limited samples. Briefly, to a microcentrifuge tube containing 13 l of PBS, 5 l of bone marrow aspirate and 1 l of proteinase K were added. After mixing briefly, 17 l of buffer AL (both proteinase K and AL buffer were from QIAamp DNA Mini kit; Qiagen Inc. Valencia, CA) were added and mixed, and samples were incubated at 56C for 1 h. Bisulfite modification was as described previously (42, 43). To the above reaction tube, 4 l of 3.0 M NaOH, freshly prepared, was added, and the DNA sample was denatured at 37C for 15 min. The sample was then treated with 416 l of 3.6 M sodium bisulfite solutions (pH 5.0) and 24 l of 10 mM hydroquinone, both freshly prepared, and samples were incubated under 2 drops of mineral oil at 55C for 16 h. For each set of modification, DNA extracted from peripheral blood of healthy individuals was included as control for the unmethylated version, and DNA from healthy individuals treated with Sss I methylase (New England BioLabs, Inc., Beverly, MA) was also included as positive methylation control. Bisulfite-modified DNA samples were then purified using the Wizard DNA Clean-Up System (Promega, Madison, WI) according to the manufacturer's instructions and eluted twice in a total of 60 l of 10 mM Tris (pH 7.6) preheated to 70C. Freshly prepared NaOH, to a final concentration of 0.3 M, was added, and the sample was incubated at 37C for 15 min. The solution was neutralized by addition of ammonium acetate (pH 7.0) to 3 M, and the DNA was ethanol precipitated, dried, and resuspended in 30 l of 10 mM Tris buffer. Methylation Status by Methylation-Specific PCR (MS-PCR). Detection of methylated CpG dinucleotides within the promoter region CpG island of the FHIT gene was carried out using MS-PCR (44), and primers (Qiagen Operon, Alameda, CA) for both methylated and unmethylated CpG sites were as described previously (34). Methylation detected with this assay was demonstrated previously to be significantly associated with loss of gene expression in lung cancer cell lines and primary lung and breast tumors (34); methylated CpG site, forward 5-TTGGGGCGCGGGTTTGGGTTTTTACGC-3 and reverse 5-CGTAAACGACGCCGACCCCACTA-3, unmethylated CpG site, forward 5-TTGGGGTGTGGGTTTGGGTTTTTATG-3, and reverse 5CATAAACAACACCAACCCCACTA-3; GenBank accession no. is U76263, with an amplicon of 189 262 bp relative to transcription start site. The PCR mixture contained 10 PCR buffer (Applied Biosystems), MgCl2 (1.5 mM final), deoxynucleotide triphosphates (0.2 mM each), primers (0.4 M each), 1 unit AMPliTAQ DNA polymerase treated with TaqStart Antibody (CLONTECH Laboratories, Inc., Palo Alto, CA), and 2 l of modified bone marrow DNA templates in a total volume of 25 l. The PCR reactions were cycled in a GeneAmp 9600 thermal cycler (Applied Biosystems) under the following conditions: preheat at 94C for 3 min., 94C for 30 s, 65C for 30 s, 72C for 30 s for 38 times, and a final extension at 72C for 7 min. For each PCR set, DNA samples from peripheral blood of normal blood donors treated with CpG methylase (M.Sss I; New England BioLabs) and bisulfite were included as positive controls, and no template reaction was included as negative control. In addition, we repeated MS-PCR assays on 30% of primary tumor specimens and found no discordant results among replicates. Aliquots (12 l) of MS-PCR products were analyzed on 3% agarose gel, stained with ethidium bromide, and visualized under UV illumination. Results were recorded with a digital image system. Bisulfite Genomic Sequencing. To confirm the efficiency of the bisulfite modification and the specificity of methylation-specific PCR, direct sequencing of the PCR products was carried out as described previously (43). Briefly, PCR products were ligated into the PCR 2.1-TOPO plasmid vector using TOPO TA Cloning kit (Invitrogen, Carlsbad, CA). Purified plasmid DNA containing FHIT gene amplicon was sequenced in both directions using an ABI 377 automated sequencer with standard M13 primers. Two control samples from healthy individuals, two leukemia cell lines, and four FHIT-methylated bone marrow samples were directly sequenced. FHIT Expression by RT-PCR and Western Blot. cDNAs were synthesized from 3 g of total RNA extracted from leukemia cell lines and representative primary leukemia bone marrows using Qiagen RNeasy Mini kits. The cDNA concentration was then normalized in series of PCRs with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primers [5-TCGTGGAAGGACTCATGACC-3 (sense) and 5-GGGATGATGTTCTGGAGAGC-3 (antisense); 115 bp transcript fragment] by carefully diluting cDNA samples until PCR products of different samples were similar to each other in band Table 2 Distribution of immunophenotypic and cytogenetic subgroups of acute childhood leukemia in the NCCLSa and the UKCCS Current study NCCLS UKCCSb Immunophenotype/ cytogenetic subgroup 2 Age Gender 25 614 Female Male 2 Age Gender 25 614 Female Male B-cell ALL (141/951)c T-cell ALL (14/137) AML (32/250) t(12;21) (41/139) Hyperdiploid (56/423) 15 (10.6%) 0 (0.0%) 4 (12.5%) 3 (7.3%) 5 (8.9%) 86 (61.0%) 3 (21.4%) 5 (15.6%) 34 (82.9%) 35 (62.5%) 40 (28.4%) 11 (78.6%) 23 (71.9%) 4 (9.8%) 16 (28.6%) 69 (48.9%) 4 (28.6%) 12 (37.5%) 17 (41.5) 31 (55.4%) 72 (51.1%) 10 (71.4%) 20 (62.5%) 24 (58.5%) 25 (44.6%) 109 (11.5%) 8 (5.8%) 71 (28.4%) 6 (4.3%) 29 (6.9%) 579 (60.9%) 50 (36.5%) 58 (23.2%) 95 (68.3%) 271 (64.1%) 263 (27.6%) 79 (57.7%) 121 (48.4%) 38 (27.3%) 123 (29.1%) 417 (43.8%) 47 (34.3%) 116 (46.4%) 65 (46.8%) 182 (43.0%) 534 (56.2%) 90 (65.7%) 134 (53.6%) 74 (53.2%) 241 (57.0%) a NCCLS, Northern California Childhood Cancer Center Study; UKCCS, United Kingdom Childhood Cancer Study; ALL, acute lyphoblastic leukemia; AML, acute myeloid leukemia. b UKCCS (46), B-cell ALL cases from the UKCCS (n 951) include common ALL and pro-B ALL. c Numbers shown denote cases in the current study NCCLS/number of cases in the UKCCS within each row subtype. One ALL case listed in Table 1 could not be classified further; hence the total number of ALL here for the NCCLS is 155. 2001 DNA METHYLATION PROFILE IN PEDIATRIC LEUKEMIA intensity. GAPDH-cycling parameters were preheated at 94C for 2 min, then 94C for 15 s, 60C for 30 s, and 72C for 30 s. The reaction was repeated for 33 cycles. Using the normalized cDNA as a template, FHIT transcripts were amplified with previously described primers 5RT-F/3D2 (GCTCTTGTGAATAGGAAACC-sense, TCACTGGTTGAAGAATACAGG-antisense) and cycling conditions (95C for 30 s, 58C for 30 s, and 72C for 30 s for 38 cycles; 33). This assay amplified a 532 bp FHIT transcript spanning exon 5 to exon 10. The GAPDH and FHIT PCR products were run on 3% and 2% agarose gel, respectively, and visualized by ethidium bromide staining. Western blot analysis was performed as described previously (45); briefly, protein was extracted from leukemia cell lines and representative primary leukemia bone marrow samples using Mammalian Protein Extraction Reagent (PIERCE, Rockford, IL) with additions of 150 mM sodium chloride and Halt Protease Inhibitor Cocktail (PIERCE), and quantified with bicinchoninic acid protein assay kit (PIERCE) according to the manufacturer's instructions. One hundred g of cell extract were electrophoretically separated on a 4 20% SDS-polyacrylamide gel with 150 V for 50 min and transferred to nitrocellulose filter under 100 V for 2 h at 4C. The membrane was blocked with 5% nonfat milk in PBS containing 0.1% Tween 20 for 2 h at room temperature, and the blot was then incubated overnight at 4C with anti-FHIT rabbit antibody (Zymed Laboratories, South San Francisco, CA) diluted 1:500 in PBS containing 0.1% Tween 20 and 2.5% nonfat milk. After extensive washing with PBS containing 0.1% Tween 20, the filter was incubated for 1 h with goat antirabbit IgG horseradish peroxidase conjugate (Zymed) diluted 1:2000 and washed with PBS containing 0.1% Tween 20. The immunoreactive bands were visualized with enhanced chemiluminescence detection reagent (Amersham, Arlington Heights, IL) as described by the manufacturer. The quality of the protein was assessed by incubating the filter with antitubulin antibodies instead of anti-FHIT antibodies. 5-aza-dC Treatment and RT-PCR. HL-60, Blin, Reh, Molt-4, and Jurkat leukemia cell lines were maintained in culture medium with and without 0.51.0 M 5-aza-dC (Sigma) for 6 days. RNA extraction and RT-PCR were performed as described above. FISH Detection of t(12;21) and Hyperdiploidy. The FISH method applied in this study was designed for detecting TEL-AML1 fusion genes derived from t(12;21) and high hyperdiploidy (50 chromosomes, hereafter referred to as "hyperdiploid") simultaneously. Interphase FISH probes targeted to the TEL and AML1 genes and the centromere of chromosome X (Vysis, Downer Grove, IL) were applied in bone marrow smears of childhood ALL patients. The TEL probe begins between exons 3 and 5 and extends approximately 350 kb toward the telomere of chromosome 12 and was labeled directly with SpectrumGreen. The AML1 probe labeled directly with SpectrumOrange spans the entire gene of approximately 500 kb. The centromere probe of chromosome X was labeled with SpectrumAqua. The hybridization procedures were performed according to the manufacturer's protocols. FISH signals were viewed with a quadra-band filter (Chroma, Battleboro, VT). The t(12;21) was detected by observing the TEL-AML1 fusion signals (yellow), whereas high hyperdiploidy was defined when the additional copies of both chromosomes 21 and X were found in the same cell, because over 90% of ALL cases that are high hyperdiploid include extra copies of both 21 and X (2). These cytogenetic characteristics were observed and confirmed by two experienced cytogeneticists. Statistical Analysis. Statistical analyses were carried out using SAS analysis software. 2 analyses were used to test for associations between FHIT methylation status and a variety of demographic, cytogenetic, and molecular characteristics of the patients, including age, gender, ethnicity, histological subtype of leukemia, the presence of any chromosomal translocations, t(12; 21)/TEL-AML1, and hyperdiploidy (50 chromosomes in leukemic cells). Wilcoxon rank-sum test, Students t test, and Fisher's exact test were used to test for associations of elevated WBC counts with methylation. We also tested the association of FHIT methylation with a classification scheme that combined conventional and FISH cytogenetic data and incorporated immunophenotypic data to create five major subgroups of pediatric ALL, including B-cell ALL hyperdiploid (50 68 chromosomes) positive, B-cell ALL TEL-AML1 translocation positive, B-cell ALL translocation positive & TEL-AML1 negative, B-cell ALL hyperdiploidy & translocation negative, and T-cell ALL hyperdiploidy & translocation negative. All statistical runs were done using coded patient specimens, and all methylation analysis were carried out blind with respect to the clinical, demographic, and cytogenetic status of patients. RESULTS Study Population Demographic and Clinicopathological Characteristics. Characteristics of the leukemia cases included in the study are shown in Table 1. To assess how well our cases represent pediatric leukemia, we compared our case series with another large and well-defined population. Table 2 shows the distribution of pediatric acute leukemias by age, gender, immunophenotype, and cytogenetic characteristics compared with the large database from the United Kingdom Childhood Cancer Study (46). The age distribution of the common B-cell leukemias (25 years) were very similar to the United Kingdom study as well as indicating an older age for children presenting with non-B-cell tumors. A male predominance was noted in our case series. B cell was the most common acute leukemia (71.1% United Kingdom/75.4% NCCLS), followed by AML (18.9% United Kingdom/17.1% NCCLS) and T-cell leukemia (10.2% United Kingdom/7.5% NCCLS). Among the B-cell leukemias, similar percentages of hyperdiploid tumors were found (44.5% United Kingdom/39.7% NCCLS) and t(12;21) accounted for 29.1% of B-cell leukemia in the current study compared with 14.6% in the United Kingdom study. FHIT 5 CpG Island Methylation in Cancer Cell Lines. We analyzed 10 leukemia cell lines and found methylation of cytosine residues at CpG dinucleotides in four of them (Blin, Jurkat, Nalm, MOLT-4; Fig. 1A). These cell lines contained either the methylated or the unmethylated form, except the Jurkat cell line, which contained both forms (hemimethylated). No cell line that lacked both methylated and unmethylated forms was found, indicating no homozygous deletions in the locus tested. To assess the sensitivity of the MS-PCR method, DNA samples from methylated Molt-4 and unmethylated KG1A were mixed in different ratios; a single unambiguous FHITmethylated band was detectable when methylated template was present at 1:32 (3%) of the total DNA but not at lower dilutions (data not shown). Aberrant FHIT Methylation in Primary Pediatric Leukemias. We examined 190 bone marrow aspirates of primary pediatric leukemias using the MS-PCR method and found an overall FHIT promoter methylation frequency of 27.4% (52 of 190) and 32.1% (50 of 156) for ALL, and 6.2% (2 of 32) for AML (Table 3). MS-PCR results for representative bone marrows are shown in Fig. 1B. We found no methylated products among 10 peripheral blood samples from healthy individuals. The unmethylated form of FHIT was found in 90% of the methylated leukemia bone marrow samples, indicating some contamination with normal cells, and 100% of the unmethylated samples. No homozygous deletion in the locus tested was found. In addition to MS-PCR, which was run on all cases, we also carried out bisulfite- Fig. 1. Methylation analysis of fragile histidine triad (FHIT) promoter region by methylation-specific PCR. A, methylation status of 10 leukemia cell lines. Molt-4, Nalm, and Blin are methylated; Jurkat is hemimethylated; others are not methylated. B, FHIT methylation in primary bone marrow aspirates. Patients 0036, 0141, 0206, and 0394 are methylated; positive control DNA from peripheral blood of a healthy individual was treated with Sss I methylase and then bisulfite modified; negative control is a no template control. U, unmethylated; M, methylated. 2002 DNA METHYLATION PROFILE IN PEDIATRIC LEUKEMIA Table 3 Effects of gender, age, histologic subtype and cytogenetics on the distribution of FHITa methylation in childhood leukemia FHIT methylation status (%) Patient characteristic No (unmethylated) Yes (methylated) P Gender Male Female Age (years) 2 25 614 Subtype ALL AML CML Any Translocationb No Yes t(12;21)c No Yes Hyperdiploidyc No Yes 76 (73.1) 62 (72.1) 16 (84.2) 64 (67.4) 58 (76.3) 106 (67.9) 30 (93.8) 2 (100.0) 44 (58.7) 70 (94.6) 63 (56.8) 40 (97.6) 78 (81.3) 25 (44.6) 28 (26.9) 24 (27.9) 3 (15.8) 31 (32.6) 18 (23.8) 50 (32.1) 2 (6.2) 0 (0.0) 31 (41.3) 4 (5.4) 48 (43.2) 1 (2.4) 18 (18.8) 31 (55.4) 0.88 0.21 0.008 0.0001 0.0001 0.0001 a FHIT, fragile histidine trial; ALL, acute lyphoblastic leukemia; AML, acute myeloid leukemia; CML, chronic myeloid leukemia. b These analyses were limited to a total of 149 cases whose clinical cytogenetic studies were considered adequate, i.e. at least 20 metaphase cells were analyzed. P apply to comparisons using the 2 test. c These included 152 ALL cases for whom the diagnosing hospital or the University of California Berkeley performed flourescence in situ hybridization screening for t(12;21) and hyperdiploidy and for whom the t(12;21) and hyperdiploidy information was captured from the hospital clinical cytogenetics report. To create mutually exclusive categories, we have assumed that cases who were hyperdiploidy (50 chromosomes) per hospital cytogenetics are negative for t(12;21) and that T-cell ALL cases are negative for both hyperdiploidy and t(12;21) and that cases with other types of translocations are negative for both hyperdiploidy and t(12;21). These assumptions are justified by observations reported in large multicenter clinical series (1, 2). RNA was harvested to determine whether the locus could be reactivated with the demethylating agent. RT-PCR analysis showed that, after 5-aza-dC treatment, two methylated cell lines re-expressed mRNA of the FHIT gene, and hemimethylated cell line showed detectable increased mRNA production (Fig. 4). Cytogenetic and Patient Characteristics Correlated with FHIT Methylation. As shown in Table 3, it was rare to observe FHIT methylation in patients who presented with tumors containing any chromosomal translocations (5.4% compared with 41.3% in patients without translocations, P 0.0001). On the other hand, patients with hyperdiploid leukemia were much more likely to have FHIT methylation compared with other patients (P 0.0001). FHIT methylation status was not associated with the age or gender of patients. No association was observed between FHIT methylation status and ethnicity (data not shown). The inclusion of immunophenotype data did not reveal any further associations within the ALL group; FHIT methylation was common in both CD10/CD19-positive ALL with hyperdiploidy (B cell), but it was also prevalent among the smaller Table 4 Relationship of FHITa methylation and expression in leukemia cell lines Gene expression Cell lines Classification Methylation Transcriptsb Proteinsc Molt-4 KG1A HL-60 RCH Blin Reh Jurkat 697 Nalm K562 T-ALL AML AML B-ALL B-ALL B-ALL B-ALL B-ALL B-ALL CML d e e f g a FHIT, fragile histidine triad; ALL, acute lyphoblastic leukemia; AML, acute myeloid leukemia; CML, chronic myeloid leukemia. b As detected by reverse transcription-PCR. c As detected by Western blotting. d Hemimethylated with both methylated and nonmethylated bands. e The bands were weak. f Multiple PCR bands showed products of full length and smaller molecular weight. g Multiple PCR bands showed products of small molecular weight only. Fig. 2. Representative reverse transcription (RT)-PCR analysis. In leukemia cell lines, fragile histidine triad (FHIT) RT-PCR products are absent in Blin and Molt-4 and detectable in Jurkat, KG1A, HL-60, and Reh. In primary leukemia bone marrows, FHIT RT-PCR products are absent in patients 0206 and 0141 and detectable in 0036, 0059, 0270, and 0447. BM, bone marrow; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. sequencing on four randomly selected methylated (by MS-PCR) bone marrows to confirm complete C to T conversion. Each of the four methylated bone marrows showed complete conversion except for cytosines in CpG dinucleotides. Effect of FHIT Hypermethylation on Gene Expression. By RTPCR analysis, undetectable or very low levels of FHIT transcripts Fig. 3. Western blot analysis of fragile histidine triad (FHIT) protein expression in representative leukemia cell lines and primary bone marrow samples. After SDS-PAGE gel electrophoresis and transfer, the filter was cut between FHIT protein and tubulin protein regions based on a prestained size standard. Upper and lower portions of the filter were probed with tubulin and FHIT antibodies, respectively (in this figure, FHIT is depicted above tubulin). Tubulin probe is used to confirm equal loading of protein samples. Left, the size of the two marker bands for FHIT (Mr 17,000) and tubulin (Mr 55,000) are indicated. were found in four of six leukemia cell lines and three of six repre- sentative leukemia samples (Fig. 2; Table 4). Among the six leukemia cell lines and six primary leukemia bone marrows, we found an inverse association between FHIT methylation and FHIT transcript expression. To confirm the transcript analysis results, FHIT protein detection was performed by Western blotting. As predicted, protein product was not detectable in methylated cell lines and methylated primary leukemia bone marrows (Fig. 3; Table 4). Recovery of FHIT Transcripts after Treatment with 5-aza-dC. Two methylated cell lines (Blin and Molt-4), two unmethylated cell lines (HL-60 and Reh), and one hemimethylated cell line (Jurkat) were treated with 0.51.0 M 5-aza-dC for 6 days at which time total Fig. 4. Reverse transcription (RT)-PCR analysis of re-expression of the fragile histidine triad (FHIT) transcripts by 5-aza-dC treatment in five leukemia cell lines; () no treatment, () 5-aza-dC treated. In farthest left lane, M shows molecular weight markers. Methylated Blin and Molt-4 show recovery of transcription after treatment, unmethylated HL-60 and Reh show no or minimal change, and moderate increase in transcript is shown for the hemimethylated Jurkat cell line. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. 2003 DNA METHYLATION PROFILE IN PEDIATRIC LEUKEMIA Table 5 FHITa methylation by immunophenotype and cytogenetic subgroups of childhood leukemia FHIT methylation status B cell ALL hyperdiploid (5068 chr) positive B cell ALL TEL-AML1 translocation positive B-cell ALL translocation positiveb & TEL-AML1 negative B cell ALL hyperdiploid & translocation negative T cell ALL hyperdiploid & translocation negative Myeloid leukemiac Otherd Total Methylated Unmethylated Total 31 (55.4%) 25 (44.6%) 56 1 (2.4%) 40 (97.6%) 41 1 (7.7%) 11 (37.9%) 12 (92.3%) 18 (62.1%) 13 29 2: 47.7; P 0.0001e 5 (38.5%) 8 (61.5%) 13 2 (5.9%) 32 (94.1%) 34 1 3 4 52 138 190 a FHIT, fragile histidine triad; ALL, acute lyphoblastic leukemia; AML, acute myeloid leukemia. b Any translocation other than TEL-AML1, includes nine recurrent [t(1;19)(q23;p13), t(1;19)(q21;p13), t(9;22)(q34;q11), ins(4;11)(q21;q13q23), t(8;14)(q24;q23)], and four nonrecurrent translocations [t(8;14)(q34;q32), t(3;9)(p10;q10), t(9;20)(q12;q11.2), t(12;?) not TEL/AML1] see Mitelman F, Johansson B, and Mertens, editors Mitelman database of chromosome aberrations in cancer (the November 2003 version of the database was accessed and searched on 02/10/04). Available from: http://cgap.nci.nih.gov/Chromosomes/ Mitelman. c 32 acute myeloid leukemias, 2 chronic myeloid leukemias; both methylated leukemias were AML. d Other: 1 T-cell ALL with translocation, 3 unknowns due to missing cytogenetics and/or flow cytometry data, i.e. 152 total ALL are listed of the total of 156. e 2, tests the distribution of the entire table. group of ALLs with T-cell markers (e.g., CD2, CD4, CD7, and CD8; Table 5; Fig. 5). The most positive predictive variables for FHIT methylation appear first to be lymphocytic versus myeloid lineage and then the karyotypic features of pediatric ALL and specifically the absence of translocation or other detectable structural alterations. We examined age and WBC counts at diagnosis within subgroups of pediatric leukemia. No significant differences in age at diagnosis by FHIT methylation status were observed. Higher median and mean WBC counts were observed among the methylated T- and B-cell ALLs that were negative for translocation and hyperdiploidy com- pared with unmethylated cases in this same subtype; however, the differences were not statistically significant. For example, a median value of 31.5 106/ml versus 14.2 106/ml WBC were observed for methylated and unmethylated cases, respectively (P 0.05). Similarly, the mean WBC count was higher among methylated cases (93.9 versus 32.9 106/ml), but this difference was also only borderline significantly different (P 0.06; Students t test). Significantly higher WBC counts at diagnosis were associated with FHIT methylation among hyperdiploid cases (median WBC count 11.6 106/ml in methylated versus 4.9 106/ml in unmethylated leukemias; P 0.046). Mean WBC counts were also higher in methylated (17.5 106/ml) versus unmethylated (7.5 106/ml) hyperdiploid cases (P 0.03; Students t test). It is important to note that most hyperdiploid patients do not have elevated WBC counts at diagnosis. In our study, nine patients presented with 20 106 leukocytes/ml, and importantly, eight of these were FHIT methylation positive (P 0.03; Fisher's exact test). Only 4 of 56 hyperdiploid cases presented with 40 106 WBCs/ml, and all of these were found to be FHIT methylation positive. We also had access to data on p15INK4b methylation in this series; 44 of 190 (23.2%) cases were p15 methylation positive, but no association was found between p15INK4b and FHIT methylation status (data not shown). DISCUSSION The current study emphasizes the role of aberrant methylation in the inhibition of FHIT gene expression in leukemia. We compared FHIT methylation and loss of FHIT transcripts by RT-PCR assay and found a strong concordance between methylation and lack of expression among all leukemia cell lines. Results of Western blot analyses showed a similar match between FHIT methylation and FHIT expression, except for cell line K562, which was not methylated but also did not express FHIT protein. We looked at the mRNA patterns of K562 and found only aberrant transcripts, indicating deletion or rearrangements may be the mechanism for FHIT inactivation in this cell line. Additional evidence that supports the importance of FHIT methylation in transcriptional silencing is the reexpression of the FHIT gene after treatment with 5-aza-dC, a known demethylating agent. As Fig. 5. Fragile histidine triad (FHIT) methylation in immunophenotype and cytogenetic subgroups of childhood leukemia. Histogram originate depicts the numbers of primary leukemia cases that demonstrated FHIT methylation (black) or unmethylated CpG regions (white) as described in the "Materials and Methods." Subgroups of cases were created according to conventional cytogenetic and fluorescence in situ hybridization (FISH) analyses and immunophenotype (cluster of differentiation markers). The percentages indicated above histograms apply to the percent of all B-cell acute lymphoblastic leukemia (ALL), all T-cell ALL, and all myeloid leukemias. 2004 DNA METHYLATION PROFILE IN PEDIATRIC LEUKEMIA predicted, these findings are in agreement with the results found in suggest that structural chromosomal alterations may be inversely esophageal squamous cell carcinoma and lung and breast cancers (33, correlated with FHIT abnormalities, as we found in our series. For 34). Taken together, our results demonstrate that promoter aberrant example, in previous studies of Philadelphia chromosome (Ph)-posi- methylation of FHIT is an important mechanism for inactivation of tive chronic myelogenous leukemia, intact (i.e., normal) FHIT tran- this tumor suppressor gene in hematological malignancies. scripts were observed in all cases (14). Fewer than 4% of Ph-positive The most dramatic findings of this study are the common occur- chronic myeloid leukemias showed low levels of FHIT protein by rence of FHIT methylation among specific subtypes of primary pedi- Western blot analysis (18). atric ALL and the uncommon occurrence of FHIT methylation in Decreased or absent FHIT expression could arise through either myeloid leukemias. Methylation of FHIT occurred in 27.4% of our epigenetic or genetic mechanisms affecting the FHIT locus. In one pediatric leukemia patients but was most common among the hyper- study, about half of B-cell and all T-cell ALLs examined demon- diploid ALL subgroup (55.4% methylated). In marked contrast, only strated reduced FHIT protein expression (16). Loss of FHIT was also 2.4% of t(12;21)/TEL-AML-positive cases were methylation positive, very common in AML (17, 21). These previous studies did not and interestingly the B-cell ALLs with translocations other than specifically examine pediatric leukemia as we have in our series; to t(12;21)/TEL-AML also showed a low methylation rate. Because our our knowledge, ours is the first study of FHIT methylation in pediatric case series was extensively characterized, we were able to combine leukemia. In a recent study comparing pediatric and adult ALL, the immunophenotypic data with the conventional and molecular cytogenetic prevalence of methylation of a panel of gene loci (ER, MDRT, p15, data for each patient. Considering all available patient characteristics, we C-ABL, CD10, p16, p73) was found to be very similar in 16 pediatric found that irrespective of immunophenotype, those ALL cases that did and 61 adult ALL (11). Although not a focus of the current study, we not have detectable structural chromosomal alterations were most likely found 23% of pediatric ALL to demonstrate p15INK4b methylation, to demonstrate FHIT methylation, and the lowest rates were observed in which is very similar to the previous report in childhood ALL (i.e., ALL with translocations or other structural chromosomal changes. 25% p15INK4b methylation positive; Ref. 11). We found no correla- Within the hyperdiploid cases, the data also indicated a trend toward tion of FHIT with p15INK4b methylation in our study, and hence higher WBC counts with FHIT methylation. In contrast, only 5.9% of FHIT probably represents a class of methylation targets distinct from pediatric AML demonstrated FHIT methylation. those examined previously. Additional studies of FHIT methylation The marked association of FHIT methylation with hyperdiploid and expression in adult ALL would be useful. ALL suggests an etiological role of FHIT in this type of leukemia. A Interestingly, human tumor suppressor gene FHIT is located on the unique mechanism involving a single aberrant mitotic division that short arm of chromosome 3p, a region that harbors many other leads to a highly nonrandom aneuploidy has been proposed in the potential tumor suppressor genes (53). One of most important discov- evolution of the high hyperdiploid subtype of ALL (47). It is unknown eries involving putative tumor suppressors in this region is the im- whether this mitotic event precedes or follows the methylation of portance of tumor-acquired promoter hypermethylation as an epige- FHIT in hyperdiploid ALL. Hyperdiploidy is associated with a favor- netic mechanism for inactivating the expression of these genes. able survival outcome among ALL patients (48), and possibly relevant DUTT1 (ROBO1) at 3p12, RASSF1A at 3p21.3, BLU at 3p21.3, to this is the observation that a large proportion of hyperdiploid SEMA3B at 3p21.3, HYAL1 at 3p21.3, CACNA2D2 at 3p21.3, RAR2 leukemic blasts display a high rate of spontaneous apoptosis (49). at 3p24, and VHL at 3p25.3 have been found to undergo hypermethy- Interestingly, FHIT has been implicated in apoptosis (32, 50) and in lation and are associated with absent or reduced expression in various the metabolism of Ap3A (51). IFNs and other cytokines effectively human malignancies (34, 54 58). Three of those genes, RASSF1A, induce apoptosis in hyperdiploid and other leukemias (29) and have RAR2, and FHIT, have been shown to be involved in human leuke- been shown to also induce Ap3A levels (27, 28, 52). In this regard, our mia (55, 59). Therefore, methylation of genes residing in the 3p region finding of higher WBC counts at diagnosis among FHIT-methylated should be further examined in childhood leukemia. It will be of cases may be indicative of leukemic cells that are defective in some interest to see whether a concordant pattern of methylation in the 3p FHIT-related apoptotic mechanism or pathway. region is associated with the specific cytogenetic subtypes of ALL. A similarly high rate of FHIT methylation (i.e., 37.9%) was also DNA hypermethylation, affecting the p15INK4b, p16INK4a, RB, observed among ALL cases that exhibited the normal complement of p73, NNAT, and calcitonin loci in childhood leukemia have been chromosomes (i.e., 46 XX, 46 XY). Although conventional cytoge- reported (35, 60). Links between aberrant methylation and clinico- netic analyses may fail to detect hyperdiploid leukemic cells in bone pathological features were observed, such as reduced survival and marrow aspirates, we also screened these "normal" cases with a higher relapse rate (4, 60, 61). Within ALL high peripheral WBC sensitive and specific FISH assay. This makes it unlikely that we have count at diagnosis is associated with a relatively worse survival misclassified this latter subgroup. On the basis of these observations, outcome. FHIT methylation should be investigated further as a prog- we interpret our data to indicate that the strongest association of FHIT nostic marker because we found that FHIT methylation was more methylation is with pediatric ALL of either T- or B-cell origins that do common among the hyperdiploid cases that had the highest WBC not contain structural chromosomal alterations and instead display counts at diagnosis. Moreover, the mechanisms leading to aberrant normal or hyperdiploid karyotypes. methylation in leukemia are only poorly understood and deserve much Because so few AML cases displayed FHIT methylation, our data more intensive investigation. It has been proposed that in some suggest that transcriptional silencing of FHIT by an epigenetic mech- cytogenetic subtypes of leukemia, aberrant CpG methylation is in- anism may be relatively uncommon in pediatric AML. This conclu- duced by abnormal chromosomal fusion proteins, which recruit DNA sion must be considered tentative because we did not assess protein methyltransferases to target specific promoters (59). Because we expression; however, previous studies indicate a close correlation of found FHIT methylation to be inversely associated with chromosomal FHIT methylation with loss of FHIT gene expression (34). In our translocations, it is unlikely that this mechanism is responsible for tar- study, gene expression tracked exquisitely with methylation in leuke- geting aberrant methylation within the FHIT locus. Additional studies are mia primary samples and cell lines (Figs. 2 4). If FHIT is not also necessary to determine whether FHIT methylation could be a useful down-regulated in pediatric AML, this would contrast with previous marker of etiologically important subgroups of pediatric leukemias. Com- results in adult AML that show frequent loss of FHIT expression. paring environmental exposures of leukemia cases or their parents ac- Even in adult myeloid leukemia, it is interesting that previous studies cording to the FHIT methylation status of the child's diagnostic bone 2005 DNA METHYLATION PROFILE IN PEDIATRIC LEUKEMIA marrow may be one way to gain insights into epigenetic mechanisms operating in some subtypes of pediatric leukemia. ACKNOWLEDGMENTS We thank clinical collaborators at the participating hospitals (Drs. V. Crouse, G. Dahl, J. Ducore, J. Feusner, V. Kiley, M. Loh, K. Matthay, S. Month, and C. Russo) and the staff members of those hospitals for help with ascertaining cases and providing specimens. J. Wiemels is a Scholar of the Leukemia and Lymphoma Society of America. REFERENCES 1. Harrison CJ, Foroni L. Cytogenetics and molecular genetics of acute lymphoblastic leukemia. Rev Clin Exp Hematol 2002;6:91113 discussion 200 2. 2. Harrison CJ, Martineau M, Secker-Walker LM. The Leukaemia Research Fund/ United Kingdom Cancer Cytogenetics Group Karyotype Database in acute lymphoblastic leukaemia: a valuable resource for patient management. Br J Haematol 2001;113:310. 3. Guo SX, Taki T, Ohnishi H, et al. Hypermethylation of p16 and p15 genes and RB protein expression in acute leukemia. Leuk Res 2000;24:39 46. 4. Liu M, Taketani T, Li R, et al. Loss of p73 gene expression in lymphoid leukemia cell lines is associated with hypermethylation. Leuk Res 2001;25:4417. 5. Kuerbitz SJ, Pahys J, Wilson A, Compitello N, Gray TA. Hypermethylation of the imprinted NNAT locus occurs frequently in pediatric acute leukemia. Carcinogenesis 2002;23:559 64. 6. Nakamura M, Sugita K, Inukai T, et al. p16/MTS1/INK4A gene is frequently inactivated by hypermethylation in childhood acute lymphoblastic leukemia with 11q23 translocation. Leukemia 1999;13:884 90. 7. Asimakopoulos FA, Shteper PJ, Krichevsky S, et al. ABL1 methylation is a distinct molecular event associated with clonal evolution of chronic myeloid leukemia. Blood 1999;94:2452 60. 8. Fu WN, Bertoni F, Kelsey SM, et al. Role of DNA methylation in the suppression of Apaf-1 protein in human leukaemia. Oncogene 2003;22:4515. 9. Garcia-Manero G, Daniel J, Smith TL, et al. DNA methylation of multiple promoterassociated CpG islands in adult acute lymphocytic leukemia. Clin Cancer Res 2002;8:221724. 10. 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;8:18971903. 11. Garcia-Manero G, Jeha S, Daniel J, et al. Aberrant DNA methylation in pediatric patients with acute lymphocytic leukemia. Cancer (Phila) 2003;97:695702. 12. Issa JP, Kantarjian H, Mohan A, et al. Methylation of the ABL1 promoter in chronic myelogenous leukemia: lack of prognostic significance. Blood 1999;93:2075 80. 13. Toyota M, Kopecky KJ, Toyota MO, Jair KW, Willman CL, Issa JP. Methylation profiling in acute myeloid leukemia. Blood 2001;97:28239. 14. Carapeti M, Aguiar RC, Sill H, Goldman JM, Cross NC. Aberrant transcripts of the FHIT gene are expressed in normal and leukaemic haemopoietic cells. Br J Cancer 1998;78:6015. 15. Gayther SA, Barski P, Batley SJ, et al. Aberrant splicing of the TSG101 and FHIT genes occurs frequently in multiple malignancies and in normal tissues and mimics alterations previously described in tumours. Oncogene 1997;15:2119 26. 16. Hallas C, Albitar M, Letofsky J, Keating MJ, Huebner K, Croce CM. Loss of FHIT expression in acute lymphoblastic leukemia. Clin Cancer Res 1999;5:2409 14. 17. Iwai T, Yokota S, Nakao M, et al. Frequent aberration of FHIT gene expression in acute leukemias. Cancer Res 1998;58:51827. 18. Kantarjian HM, Talpaz M, O'Brien S, et al. Significance of FHIT expression in chronic myelogenous leukemia. Clin Cancer Res 1999;5:4059 64. 19. Lin PM, Liu TC, Chang JG, Chen TP, Lin SF. Aberrant FHIT transcripts in acute myeloid leukaemia. Br J Haematol 1997;99:6127. 20. Peters UR, Hasse U, Oppliger E, et al. Aberrant FHIT mRNA transcripts are present in malignant and normal haematopoiesis, but absence of FHIT protein is restricted to leukaemia. Oncogene 1999;18:79 85. 21. Yang HW, Piao HY, Taki T, et al. Pattern of FHIT gene expression in normal and leukaemic cells. Int J Cancer 1999;81:897901. 22. Ohta M, Inoue H, Cotticelli MG, et al. The FHIT gene, spanning the chromosome 3p14.2 fragile site and renal carcinoma-associated t(3;8) breakpoint, is abnormal in digestive tract cancers. Cell 1996;84:58797. 23. Sozzi G, Veronese ML, Negrini M, et al. The FHIT gene 3p14.2 is abnormal in lung cancer. Cell 1996;85:1726. 24. Fong LY, Fidanza V, Zanesi N, et al. Muir-Torre-like syndrome in Fhit-deficient mice. Proc Natl Acad Sci USA 2000;97:47427. 25. Zanesi N, Fidanza V, Fong LY, et al. The tumor spectrum in FHIT-deficient mice. Proc Natl Acad Sci USA 2001;98:10250 5. 26. Siprashvili Z, Sozzi G, Barnes LD, et al. Replacement of Fhit in cancer cells suppresses tumorigenicity. Proc Natl Acad Sci USA 1997;94:13771 6. 27. Vartanian A, Narovlyansky A, Amchenkova A, et al. Interferons induce accumulation of diadenosine triphosphate (Ap3A) in human cultured cells. FEBS Lett 1996;381:32 4. 28. Turpaev K, Hartmann R, Justesen J. 2-Adenylated derivatives of Ap3A activate RNase L. FEBS Lett 1999;457:9 12. 29. Jedema I, Barge RM, Willemze R, Falkenburg JH. High susceptibility of human leukemic cells to Fas-induced apoptosis is restricted to G(1) phase of the cell cycle and can be increased by interferon treatment. Leukemia 2003;17:576 84. 30. Castelli J, Wood KA, Youle RJ. The 25A system in viral infection and apoptosis. Biomed Pharmacother 1998;52:386 90. 31. Dumon KR, Ishii H, Fong LY, et al. FHIT gene therapy prevents tumor development in Fhit-deficient mice. Proc Natl Acad Sci USA 2001;98:3346 51. 32. Roz L, Gramegna M, Ishii H, Croce CM, Sozzi G. Restoration of fragile histidine triad (FHIT) expression induces apoptosis and suppresses tumorigenicity in lung and cervical cancer cell lines. Proc Natl Acad Sci USA 2002;99:361520. 33. Tanaka H, Shimada Y, Harada H, et al. Methylation of the 5 CpG island of the FHIT gene is closely associated with transcriptional inactivation in esophageal squamous cell carcinomas. Cancer Res 1998;58:3429 34. 34. Zochbauer-Muller S, Fong KM, Maitra A, et al. 5 CpG island methylation of the FHIT gene is correlated with loss of gene expression in lung and breast cancer. Cancer Res 2001;61:35815. 35. Maruyama R, Toyooka S, Toyooka KO, et al. Aberrant promoter methylation profile of bladder cancer and its relationship to clinicopathological features. Cancer Res 2001;61:8659 63. 36. Maruyama R, Toyooka S, Toyooka KO, et al. Aberrant promoter methylation profile of prostate cancers and its relationship to clinicopathological features. Clin Cancer Res 2002;8:514 9. 37. Virmani AK, Muller C, Rathi A, Zoechbauer-Mueller S, Mathis M, Gazdar AF. Aberrant methylation during cervical carcinogenesis. Clin Cancer Res 2001;7:584 9. 38. Chang KW, Kao SY, Tzeng RJ, et al. Multiple molecular alterations of FHIT in betel-associated oral carcinoma. J Pathol 2002;196:300 6. 39. Zheng S, Ma X, Buffler PA, Smith MT, Wiencke JK. Whole genome amplification increases the efficiency and validity of buccal cell genotyping in pediatric populations. Cancer Epidemiol Biomark Prev 2001;10:697700. 40. Ma X, Buffler PA, Selvin S, et al. Daycare attendance and risk of childhood acute lymphoblastic leukaemia. Br J Cancer 2002;86:1419 24. 41. Ma X, Buffler PA, Gunier RB, et al. Critical windows of exposure to household pesticides and risk of childhood leukemia. Environ Health Perspect 2002;110:955 60. 42. Wiencke JK, Zheng S, Lafuente A, et al. Aberrant methylation of p16INK4a in anatomic and gender-specific subtypes of sporadic colorectal cancer. Cancer Epidemiol Biomark Prev 1999;8:501 6. 43. Zheng S, Chen P, McMillan A, et al. Correlations of partial and extensive methylation at the p14(ARF) locus with reduced mRNA expression in colorectal cancer cell lines and clinicopathological features in primary tumors. Carcinogenesis 2000;21:2057 64. 44. Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci USA 1996;93:9821 6. 45. Albitar M, Manshouri T, Gidel C, et al. Clinical significance of fragile histidine triad gene expression in adult acute lymphoblastic leukemia. Leuk Res 2001;25:859 64. 46. United Kingdom Childhood Cancer Study Investigators. The United Kingdom Childhood Cancer Study: objectives, materials and methods. Br J Cancer 2000;82:1073102. 47. Onodera N, McCabe NR, Rubin CM. Formation of a hyperdiploid karyotype in childhood acute lymphoblastic leukemia. Blood 1992;80:203 8. 48. Kaspers GJ, Smets LA, Pieters R, Van Zantwijk CH, Van Wering ER, Veerman AJ. Favorable prognosis of hyperdiploid common acute lymphoblastic leukemia may be explained by sensitivity to antimetabolites and other drugs: results of an in vitro study. Blood 1995;85:751 6. 49. Zhang Y, Lu J, van den Berghe J, Lee SH. Increased incidence of spontaneous apoptosis in the bone marrow of hyperdiploid childhood acute lymphoblastic leukemia. Exp Hematol 2002;30:3339. 50. Mady HH, Melhem MF. FHIT protein expression and its relation to apoptosis, tumor histologic grade and prognosis in colorectal adenocarcinoma: an immunohistochemical and image analysis study. Clin Exp Metastasis 2002;19:351 8. 51. Barnes LD, Garrison PN, Siprashvili Z, et al. Fhit, a putative tumor suppressor in humans, is a dinucleoside 5,5-P1,P3-triphosphate hydrolase. Biochemistry 1996; 35:11529 35. 52. Elias L, Crissman HA. Interferon effects upon the adenocarcinoma 38 and HL-60 cell lines: antiproliferative responses and synergistic interactions with halogenated pyrimidine antimetabolites. Cancer Res 1988;48:4868 73. 53. Zabarovsky ER, Lerman MI, Minna JD. Tumor suppressor genes on chromosome 3p involved in the pathogenesis of lung and other cancers. Oncogene 2002;21:691535. 54. Dallol A, Forgacs E, Martinez A, et al. Tumour specific promoter region methylation of the human homologue of the Drosophila Roundabout gene DUTT1 (ROBO1) in human cancers. Oncogene 2002;21:3020 8. 55. 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;25:3159. 56. Tse C, Xiang RH, Bracht T, Naylor SL. Human Semaphorin 3B (SEMA3B) located at chromosome 3p21.3 suppresses tumor formation in an adenocarcinoma cell line. Cancer Res 2002;62:542 6. 57. Csoka AB, Frost GI, Stern R. The six hyaluronidase-like genes in the human and mouse genomes. Matrix Biol 2001;20:499 508. 58. Virmani AK, Rathi A, Zochbauer-Muller S, et al. Promoter methylation and silencing of the retinoic acid receptor- gene in lung carcinomas. J Natl Cancer Inst (Bethesda) 2000;92:13037. 59. Di Croce L, Raker VA, Corsaro M, et al. Methyltransferase recruitment and DNA hypermethylation of target promoters by an oncogenic transcription factor. Science (Wash D C) 2002;295:1079 82. 60. Roman J, Castillejo JA, Jimenez A, et al. Hypermethylation of the calcitonin gene in acute lymphoblastic leukaemia is associated with unfavourable clinical outcome. Br J Haematol 2001;113:329 38. 61. Liu M, Li R, Hayashi Y, Zhu G, Guo S. [Study on abnormal expression of the p73 gene in childhood acute lymphoblastic leukemia]. Zhonghua Xue Ye Xue Za Zhi 2002;23:239 42. 2006