Folate metabolic gene polymorphisms and childhood acute ...

Leukemia (2007) 21, 320–325 & 2007 Nature Publishing Group All rights reserved 0887-6924/07 $30.00 www.nature.com/leu

ORIGINAL ARTICLE Folate metabolic gene polymorphisms and childhood acute lymphoblastic leukemia: a case–control study A Gast1, JL Bermejo1, T Flohr2, M Stanulla3, B Burwinkel1, M Schrappe4, CR Bartram2, K Hemminki1,5 and R Kumar1,6 1 Division of Molecular Genetic Epidemiology, German Cancer Research Center, Heidelberg, Germany; 2Institute of Human Genetics, University of Heidelberg, Heidelberg, Germany; 3Department of Pediatric Hematology and Oncology, Hannover Medical School, Hannover, Germany; 4Department of Pediatrics, University of Kiel, Kiel, Germany; 5 Center for Family Medicine, Karolinska Institute, Huddinge, Sweden and 6Department of Biosciences and Nutrition, Karolinska Institute, Huddinge, Sweden

We genotyped six folate metabolic pathway genes for 11 polymorphisms in 460 cases of childhood acute lymphoblastic leukemia (ALL) and 552 ethnically matched controls. None of the polymorphisms except the 66A4G (I22M) in the 5-methyltetrahydrofolate-homocysteine methyltransferase reductase (MTRR) gene showed any effect on disease risk. The carriers of the G-allele were associated with a marginal decreased risk of ALL (gender-adjusted global P ¼ 0.03; multiple-testing corrected P ¼ 0.25). Analysis of four polymorphisms in the MTRR gene showed statistically significant differences in haplotype distribution between cases and controls (global Po0.0001). The haplotypes GCAC (odds ratio (OR) 0.5, 95% confidence interval (CI) 0.4–0.6) and ATAC (OR 0.5, 95% CI 0.3–0.6) were associated with a reduced risk and the haplotypes ACAC (OR 2.3, 95% CI 1.8–2.9) and GTAC (OR 1.8, 95% CI 1.4–2.3) with an increased risk. The genotype-combination analyses indicated that the best model stratifies cases and controls based on the 66A4G and the 524C4T polymorphisms in the MTRR gene (global P ¼ 0.03). Our results suggest that, besides a weak association of childhood ALL with the 66A4G polymorphism, haplotypes within the MTRR gene may, in part, account for populationbased differences in risk. Leukemia (2007) 21, 320–325. doi:10.1038/sj.leu.2404474; published online 30 November 2006 Keywords: acute lymphoblastic leukemia; folate metabolic pathway; polymorphism; genotype

Common or rare genetic polymorphisms and accompanied haplotypes in critical genes are putative candidate modulators of ALL susceptibility. Earlier studies have shown risk alterations owing to polymorphisms in genes encoding various metabolic enzymes.9–12 In several investigations, variant alleles of the methylene-tetrahydrofolate reductase (MTHFR) gene have been associated with altered risk of acute leukemia whereas in some studies such effects were not observed.13–15 The effects of polymorphisms in other folate metabolic pathway genes on susceptibility to childhood leukemia have also been investigated.14,16 Earlier, we described that MTHFR polymorphisms had no effect on susceptibility of childhood ALL in a German population.17 In order to study the role of polymorphisms in other genes in the folate metabolic pathway on the same population we genotyped 460 childhood ALL patients and 552 ethnically matched control subjects for 11 variants in six genes that encode enzymes involved in the folate metabolic pathway.

Materials and methods

Study population Introduction Acute lymphoblastic leukemias (ALL) are the most common malignancies that afflict children worldwide.1 The known risk factors for childhood leukemia include ionizing radiation, Down’s syndrome and monogenic disorders like ataxia telangiectasia.2–4 Both neo- and post-natal exposures and infections have also been suggested to be associated with the disease risk.5 An increased risk in children with ALL-afflicted twins in a population-based study and a high concordance rate of common B-cell precursor ALL among monozygotic twins suggest prenatal origin of the disease.6,7 The etiology of ALL supports the argument for combinatorial impacts of exposures, inherited susceptibility and other influences.8 The inherited susceptibility can potentially modulate ALL risk from initiation through complete manifestation of the disease. Correspondence: Dr R Kumar, Division of Molecular Genetic Epidemiology, German Cancer Research Center, Im Neuenheimer Feld 580, D-69120 Heidelberg, Germany. E-mail: [email protected] Received 21 June 2006; revised 11 October 2006; accepted 12 October 2006; published online 30 November 2006

The study population consisted of 460 childhood ALL case patients (273 male and 187 female) of German origin recruited under an ongoing therapy study (Berlin–Frankfurt–Mu¨nster; BFM).18 The case patients were born between 1983 and 2003 with a mean age of 6.9 years (74.4 years). Patients included in the study comprised different risk categories and morphological sub-divisions (Table 1). Blood samples from case subjects were taken at the time of initial diagnosis. Informed consent was obtained from the guardians of the patients and the appropriate ethical institutional review board approved the study. Control subjects included 552 ethnically matched healthy individuals of German origin recruited in 2004 at the Institute of Transfusion Medicine and Immunology, Mannheim (Germany). The control subjects were born in southwest Germany between 1959 and 1986 (Table 1). Although the German population is probably genetically homogeneous, we used a genomic control approach to assess population stratification.19 All genotyped polymorphisms were used to estimate and contrast the stratification parameter (l). The estimate of the stratification parameter (l) from the genomic control method was 0.47, indicating no evidence of population stratification at the 0.05/11(number of polymorphisms investigated) ¼ 0.0045 confidence level.

Folate metabolic polymorphisms and childhood ALL A Gast et al

321 Table 1

Characteristics of cases and controls Cases

Controls

273 187

276 276

Mean age, years (s.d.)

6.9 (74.4)

32 (78.1)

Risk categories (%) Standard Intermediate High Unevaluated

144 259 31 26

(31) (56) (7) (6)

F F F F

Morphological sub-divisions (%) c-ALL Pre-B-ALL T-ALL Pre-T-ALL Undefined

269 84 74 8 25

(58) (18) (16) (2) (5)

F F F F F

Male Female

Ethnicity

German

German

ALL, acute lymphoblastic leukemia.

Genes and polymorphisms The polymorphisms in the genes encoding enzymes involved in the folate metabolic pathway were selected on the basis of published papers and information available on dbSNP database of the National Center for Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov). In total, 21 variants in eight genes were identified from different information sources (Supplementary Table 1).

Polymorphisms by direct DNA sequencing We attempted to validate all polymorphisms selected from various sources by sequencing a set of 32 DNA samples. For sequencing DNA, fragments containing variants were amplified by PCR using appropriate primers (Supplementary Table 2). The sequencing reactions were performed using a Big Dye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA), and data were analyzed as described previously.20

Genotyping Single nucleotide polymorphisms (SNPs) validated by DNA sequencing were genotyped by 50 nuclease allelic discrimination method using primers and fluorescence-labeled probes (Supplementary Table 3) and conditions described previously.20 The 50 UTR 28 bp repeat and the 30 UTR 6 bp deletion polymorphisms in the thymidylate synthase (TYMS) gene were genotyped by fragment analysis on the ABI Prism 3100 Genetic Analyzer (Applied Biosystems). PCR fragments were amplified using fluorescent-labeled forward and unlabeled reverse primers (Supplementary Table 3). PCR products were electrophoresed, and data analysis was performed with GeneScan analysis 3.7 software (Applied Biosystems). Direct DNA sequencing of at least 4% of the samples validated the results from allelic discrimination and fragment analysis. The samples that could not be genotyped after twice repeated assays and by direct DNA sequencing were used to calculate genotyping failure rate.

Statistical analysis Odds ratios (OR) adjusted for gender and the corresponding 95% confidence intervals (95% CI) for assessment of the

association of ALL with different genotypes were calculated using SAS version 9.1 (SAS Institute, Cary, NC, USA). The haplotype procedure of SAS/Genetics Software was used to estimate haplotype frequencies in the cases and the controls separately, and to infer the possible haplotype combinations for each individual. Relationships between genotype/haplotypes and ALL risk were summarized as global P-values, corrected for multiple-testing by the Westfall and Young permutation method.21 The association between the inferred genotype combinations and susceptibility to ALL was explored using a forward stepwise approach; the analyses started by considering a single SNP, and likelihood ratio tests were used to assess whether the consideration of further genetic markers improved the fit of the model.22 Linkage disequilibrium was calculated with Haploview software (www.broad.mit.edu/mpg/haploview/ documentation.php). Power calculations were carried out using the Power and Sample Size calculation software version 2.1.31.

Results In validation experiments, out of 21 polymorphisms, identified from published reports and various databases in eight genes involved in the folate metabolic pathway, only 12 polymorphisms in seven genes were detected in a set of 32 test DNA samples (Supplementary Table 1). The validated polymorphisms, with a minor allele frequency higher than 0.1, were genotyped in the childhood ALL patients and the controls (Table 2). The percentage of the samples (both cases and controls) that failed in genotyping assays ranged between 1 and 7%. In confirmatory experiments, a complete concordance of results was observed between allelic discrimination or fragment analysis methods of genotyping and DNA sequencing. However, both TaqMan assay for allelic discrimination and DNA sequencing methods failed in genotyping the 829C4T polymorphism in the 30 -UTR of the dihydrofolate reductase (DHFR) gene and was therefore excluded from the study. Genotype distributions in the controls were in accordance with the Hardy– Weinberg equilibrium for all polymorphisms included in the study. The sample size in our study permitted to detect under dominant model the effect of a polymorphism with a minor allele frequency higher than 0.1 and an associated OR equal or lower than 0.5 with an over 80% power (Type I error 0.05/ number of polymorphisms investigated ¼ 0.0045). Except for the 66A4G (I22M) in the MTRR gene, no statistically significant differences between the cases and the controls in allele and genotype frequencies were observed for any of the polymorphisms (Table 2). The frequency of the variant G-allele for the 66A4G polymorphism in the MTRR gene was lower in the controls than in the cases (gender-adjusted P ¼ 0.01, multiple-testing corrected P ¼ 0.15). Heterozygous and homozygous carriers of the variant allele showed a reduced risk of leukemia (ORAG versus AA 0.7, 95% CI 0.5–1.0; ORGG versus AA 0.6, 95% CI 0.4–0.9; gender-adjusted global P ¼ 0.03; multiple-testing corrected P ¼ 0.25). After correction for multiple comparisons, the differences in effect of MTRR genotypes on risk were no longer statistically significant. Genotype-specific risks were homogeneous for different categories and morphological sub-divisions of the disease (data not shown). The analysis that included the four polymorphisms 66A4G, 524C4T, 1049A4G and 1783C4T in the MTRR gene showed a statistically significant difference in the distribution of the haplotype frequencies between the cases and the controls. Out of the 16 possible haplotypes, 10 were detected in the controls and six in the cases. The modulatory effect of variant allele for Leukemia


322 Table 2

Genotype distributions of folate metabolic pathway genes in childhood ALL patients and controls Casesa

Controls

OR (95% CI)

P-value

MTR 2756A4G (D919G) AA AG GG A-allele G-allele

280 153 13 713 179

(62.8) (34.3) (2.9) (79.9) (20.1)

375 151 21 901 193

(68.6) (27.6) (3.8) (82.4) (17.6)

1.0 1.4 0.8 1.0 1.2

(referent) (1.0–1.8) (0.4–1.7) (referent) (0.9–1.5)

0.07c/0.49d

MTHFD1e 401G4A (R134K) GG GA AA G-allele A-allele

304 141 13 749 167

(66.4) (30.8) (2.8) (81.8) (18.2)

383 144 24 910 192

(69.5) (26.1) (4.4) (82.6) (17.4)

1.0 1.2 0.7 1.0 1.1


MTHFD1 1958G4A (R653Q) GG GA AA G-allele A-allele

130 234 91 494 416

(28.6) (51.4) (20.0) (54.3) (45.7)

166 267 115 599 497

(30.3) (48.7) (21.0) (54.7) (45.3)

1.0 1.0 1.1 1.0 1.0


SHMT1f 1420C4T (L474F) CC CT TT C-allele T-allele

233 171 49 637 269

(51.4) (37.7) (10.8) (70.3) (29.7)

267 232 42 766 316

(49.4) (42.9) (7.8) (70.8) (29.2)

1.0 0.9 1.3 1.0 1.0


RFC1g 80G4A (R27H) GG GA AA G-allele A-allele

125 251 79 501 409

(27.5) (55.2) (17.4) (55.1) (44.9)

178 256 108 612 472

(32.8) (47.2) (19.9) (56.5) (43.5)

1.0 1.4 1.0 1.0 1.1


TYMSh 2R43Ri 2R/2R 2R/3R 3R/3R 2R-allele 3R-allele

95 234 128 424 490

(20.8) (51.2) (28.0) (46.4) (53.6)

111 289 141 511 571

(20.5) (53.4) (26.1) (47.2) (52.8)

1.0 1.0 1.1 1.0 1.0


0.77/1.00

TYMS 1494del6 6bp+6bp+ 6bp+6bp6bp-6bp6bp+ _allele 6bp _allele

200 207 49 607 305

(43.9) (45.4) (10.7) (66.6) (33.4)

237 220 61 694 342

(45.8) (42.5) (11.8) (67.0) (33.0)

1.0 1.1 1.0 1.0 1.0


0.59/1.00

MTRRj 66A4G (I22M) AA AG GG A-allele G-allele

109 236 111 454 458

(23.9) (51.8) (24.3) (49.8) (50.2)

97 294 158 488 610

(17.7) (53.6) (28.8) (44.4) (55.6)

1.0 0.7 0.6 1.0 0.8


0.03/0.25

MTRR 524C4T (S175L) CC CT TT C-allele T-allele

180 218 55 578 328

(39.7) (48.1) (12.1) (63.8) (36.2)

226 251 63 703 377

(41.9) (46.5) (11.7) (65.1) (34.9)

1.0 1.1 1.1 1.0 1.1


MTRR 1049A4G (K350R) AA AG GG A-allele G-allele

342 82 5 766 92

(79.7) (19.1) (1.2) (9.3) (10.7)

424 102 8 950 118

(79.4) (19.1) (1.5) (89.0) (11.0)

1.0 1.0 0.8 1.0 1.0


Genotype b

Leukemia

0.14/0.81

0.09/0.63

0.83/1.00

0.69/1.00

0.87/1.00

0.12/0.75

0.73/1.00

0.05/0.37

0.53/0.65

0.73/1.00

0.77/1.00

0.01/0.15

0.82/1.00

0.54/1.00

0.91/1.00

0.81/1.00


323 Table 2

(continued)

Genotype MTRR 1783C4T (H595Y) CC CT TT C-allele T-allele

Casesa

Controls

OR (95% CI)

P-value

368 77 3 813 83

440 88 5 968 98

1.0 1.1 0.7 1.0 1.0

0.86/1.00

(82.1) (17.2) (0.7) (90.7) (9.3)

(82.6) (16.5) (0.9) (90.8) (9.2)


0.98/1.00

a

The number of cases and controls included in the study were 460 and 552, respectively. However, genotyping assays failed for different polymorphisms in both cases and controls, and the results are based on actual number of cases and controls genotyped successfully. b MTR, 5-methyltetrahydrofolate-homocysteine methyltransferase (alias methionine synthase). c Global P-values calculated from w2 test for the genotype (as a single event) and allele effects were adjusted for gender. d Global P-values are after adjustment for multiple comparisons by permutation. e MTHFD1, 5,10-methylenetetrahydrofolate dehydrogenase. f SHMT1, serine hydroxymethyltransferase. g RFC1, reduced folate carrier. h TYMS, thymidylate synthase. i 2R and 3R are two and three 28-bp repeats in the 50 -UTR of the TYMS gene, respectively. j MTRR, 5-methylenetetrahydrafolate-homocysteine methyltransferase reductase (alias methionine synthase reductase).

Table 3 Haplotype distribution of the four investigated MTRR SNPs 66A4G, 524C4T, 1049A4G and 1783C4T in childhood ALL patients and controls Haplotype

Cases

Controls

OR (95% CI)

P-valuea

ACAC GCAC GTAC ATAC ACGT ACGC

256 193 227 73 75 14

166 396 177 185 104 12

2.3 0.5 1.8 0.5 1.1 0.6

o0.0001

(30.5) (23.0) (27.1) (8.7) (8.9) (1.7)

(16.0) (38.1) (17.0) (17.8) (10.0) (1.2)

(1.8–2.9) (0.4–0.6) (1.4–2.3) (0.3–0.6) (0.8–1.6) (0.3–1.2)

a

Global P-value for haplotype effect calculated from w2 test.

66A>G

524C>T

1049A>G

1783C>T

1

0.1 1

0.95 1 1

0.83 0.61 0.86 1

66A>G 524C>T 1049A>G 1783C>T

considering the two other SNPs in the gene, when compared with a model where only the 66A4G polymorphism was considered, was small (P ¼ 0.05). The results indicate a protective effect of the MTRR 66G allele, in particular when combined with the adjacent 524C allele. Overall, the GG (66A4G)/CC (524C4T) genotype combination was associated with a decreased risk (OR 0.6, 95% CI 0.3–0.9), whereas AA/CC genotype combination was associated with increased risk (OR 1.8, 95% CI 1.2–2.8). Genotype combinations among the investigated polymorphisms (including those in MTHFR from our earlier study on the same population17) showed an even distribution in cases and controls (data not shown). Haplotype analysis of the two polymorphisms in the TYMS gene and the two MTHFD1 SNPs did not show any difference between cases and controls (data not shown). The values of linkage disequilibrium for the two polymorphic loci of TYMS (D0 ¼ 0.50) and the two loci of MTHFD1 (D0 ¼ 0.01) suggested an absence of a strong linkage.

Figure 1 Linkage disequilibrium (D0 ) between polymorphisms in the MTRR gene.

Discussion

66A4G polymorphism on the risk within the haplotypes showed a duality in the context of allele at 524C4T polymorphism. The haplotypes GCAC and ATAC were associated with a reduced risk of ALL, while the ACAC and GTAC haplotypes were associated with significantly increased risk (Table 3, global P ¼ o0.0001). A strong linkage disequilibrium was observed between the 66A4G locus and 1049A4G (D0 ¼ 0.95), and between 66A4G and 1783C4T (D0 ¼ 0.83). However, a relatively weak linkage was detected between 66A4G and the adjacent 524C4T polymorphism in the MTRR gene (Figure 1). The differences in haplotype distributions between the cases and the controls motivated genotype-combination analysis (Table 4). The analysis involving the single SNPs was followed by the stratification of the 66A4G genotypes according to a second SNP in the MTRR gene. The best model was found when the cases and the controls were classified according to genotype combinations encompassing the polymorphisms 66A4G and 524C4T (P ¼ 0.03). However, the statistical advantage of

The folate or one-carbon metabolic pathway that maintains an appropriate folate level in cells involves multiple enzymes.14 Imbalance in the folate metabolic pathway has been associated with various disorders including cancers.23 Contrary to earlier small-sized studies, none of the 11 polymorphisms in any of the six genes investigated in the present study, with the exception of 66A4G (I22M) polymorphism in the MTRR gene, showed association with ALL susceptibility.16 Interestingly, we found that the haplotypes based on the four SNPs in the MTRR gene, including the previous three uninvestigated non-synonymous polymorphisms, the 524C4T, 1049A4G and 1783C4T, showed significant differential distribution in the cases and the controls. The four haplotypes that accounted for nearly 90% of the studied population were strongly associated with modulation of the risk. The ACAC and GTAC haplotypes were associated with increased risk, whereas GCAC and ATAC seemed to be related to a reduced risk of ALL. Functionally, MTRR maintains adequate levels of methylcob(III)alamin, a functional cofactorial form of methionine synthase (MS) enzyme.24,25 MS in an active form is essential for maintaining apropriate levels of methionine, which Leukemia


324 Table 4 Forward analysis of effects of genotype combinations based on combinations of MTRR 66A4G (1) and 524C4T (2) polymorphisms on risk of ALL SNPs 1+2 AA/CC AA/CT AG/CC AG/CT AA/TT AG/TC AG/TT GG/CC GG/CT GG/TT

Cases

54 38 83 78 4 30 26 27 54 18

(13.1) (9.2) (20.1) (18.9) (1.0) (7.3) (6.3) (6.6) (13.1) (4.4)

Controls

39 43 122 70 7 53 32 53 68 20

(7.7) (8.5) (24.1) (13.8) (1.4) (10.5) (6.3) (10.5) (13.4) (3.9)

OR (95% CI)

1.8 1.1 0.8 1.4 0.7 0.7 1.0 0.6 1.0 1.1

LRT haplotype effecta

LRT complex and parsimonious modelsb

w2; P-value 19.03; 0.03

w2; P-value 13.91; 0.05

(1.2–2.8) (0.7–1.7) (0.6–1.1) (1.0–2.0) (0.2–2.4) (0.5–1.1) (0.6–1.7) (0.3–0.9) (0.7–1.4) (0.6–2.1)

ALL, acute lymphoblastic leukemia. a Likelihood ratio test for effect of the genotype combinations on risk of childhood ALL. b Likelihood ratio test comparing the complex and parsimonious models.

acts as a precursor for the universal methyl group donor S-adenosylmethionine. Mutations, including deletions and insertions in the MTRR gene, are associated with a rare autosomal recessive disorder cblE type of homocystinuria leading to megaloblastic anemia and developmental delay in early childhood.26,27 The variant allele for the MTRR 66A4G polymorphism has been previously associated with increased risk of lung and esophageal cancer and decreased risk of squamous cell carcinoma of head and neck.28–30 The genotype combinations of polymorphisms in the MTRR, MTHFR and MS genes have been associated with a reduced risk of adult ALL and non-Hodgkin’s lymphoma.24 In combination, the 677C4T polymorphism in the MTHFR gene and the 66A4G MTRR polymorphism have been associated with an increased plasma homocysteine concentration.31 Childhood ALL patients with the G-allele for the polymorphism showed decreased in vitro sensitivity in both short- and long-term methotrexate assays.32 The dietary folate and other nutrient intake have been shown to influence the risk of different cancers, including leukemia, associated with polymorphisms in folate metabolic pathway genes.14,16 However, in this study in the absence of such information, the gene–nutrient interaction could not be assessed and it cannot be excluded that the modulation of risk observed here is due to gene–folate interactions. The possibility that the observed differences in risk of leukemia owing to the haplotypes within the MTRR gene reflect historical linkage to other functional loci or gene(s) also cannot be ruled out. This argument is supported by predictions of such associations and existence of large haplotype blocks within the human genome.33 Moreover, analysis of genotype combination did not show as strong an effect on risk modulation as was observed with haplotypes. Further, the reversal of modulation effect of 66A4G (I22M) polymorphism within haplotypes on the allele at the 524C4T polymorphism, probably, points to additional polymorphism(s) in the MTRR gene that cause differential cancer risk either directly or through interaction with environmental effects. In conclusion, in this study, no association was detected between the polymorphisms in folate metabolic genes and childhood ALL except for the 66A4G variant in the MTRR gene. However, as a novel finding, we report a significant difference in population risk of childhood ALL associated with the haplotypes based on four polymorphisms in the MTRR gene. Our observation of a clear demarcation between the risk and the non-risk haplotypes in the MTRR gene may in part explain the Leukemia

differential susceptibility to childhood leukemia owing to background genetic composition, which, in conjunction with other factors, could drive the expanded clones with the chimeric fusion genes, which are present in about 1% of new born, to the overt childhood ALL.34

Acknowledgements We gratefully acknowledge BFM Study Group for their support. The study was supported by an EU grant New Generis (FOOD-CT2005-016320).

References 1 Canalle R, Burim RV, Tone LG, Takahashi CS. Genetic polymorphisms and susceptibility to childhood acute lymphoblastic leukemia. Environ Mol Mutagen 2004; 43: 100–109. 2 Greaves MF. Aetiology of acute leukaemia. Lancet 1997; 349: 344–349. 3 Meier M, den Boer ML, Hall AG, Irving JA, Passier M, Minto L et al. Relation between genetic variants of the ataxia telangiectasiamutated (ATM) gene, drug resistance, clinical outcome and predisposition to childhood T-lineage acute lymphoblastic leukaemia. Leukemia 2005; 19: 1887–1895. 4 Zeller B, Gustafsson G, Forestier E, Abrahamsson J, Clausen N, Heldrup J et al. Acute leukaemia in children with Down syndrome: a population-based Nordic study. Br J Haematol 2005; 128: 797– 804. 5 McNally RJ, Eden TO. An infectious aetiology for childhood acute leukaemia: a review of the evidence. Br J Haematol 2004; 127: 243–263. 6 Couto E, Chen B, Hemminki K. Association of childhood acute lymphoblastic leukaemia with cancers in family members. Br J Cancer 2005; 93: 1307–1309. 7 Greaves MF, Maia AT, Wiemels JL, Ford AM. Leukemia in twins: lessons in natural history. Blood 2003; 102: 2321–2333. 8 Greaves M. Infection, immune responses and the aetiology of childhood leukaemia. Nat Rev Cancer 2006; 6: 193–203. 9 Cheok MH, Evans WE. Acute lymphoblastic leukaemia: a model for the pharmacogenomics of cancer therapy. Nat Rev Cancer 2006; 6: 117–129. 10 Pakakasama S, Mukda E, Sasanakul W, Kadegasem P, Udomsubpayakul U, Thithapandha A et al. Polymorphisms of drug-metabolizing enzymes and risk of childhood acute lymphoblastic leukemia. Am J Hematol 2005; 79: 202–205. 11 Rocha JC, Cheng C, Liu W, Kishi S, Das S, Cook EH et al. Pharmacogenetics of outcome in children with acute lymphoblastic leukemia. Blood 2005; 105: 4752–4758.


325 12 Krajinovic M, Sinnett H, Richer C, Labuda D, Sinnett D. Role of NQO1, MPO and CYP2E1 genetic polymorphisms in the susceptibility to childhood acute lymphoblastic leukemia. Int J Cancer 2002; 97: 230–236. 13 Skibola CF, Smith MT, Kane E, Roman E, Rollinson S, Cartwright RA et al. Polymorphisms in the methylenetetrahydrofolate reductase gene are associated with susceptibility to acute leukemia in adults. Proc Natl Acad Sci USA 1999; 96: 12810–12815. 14 Robien K, Ulrich CM. 5,10-Methylenetetrahydrofolate reductase polymorphisms and leukemia risk: a Huge minireview. Am J Epidemiol 2003; 157: 571–582. 15 Krajinovic M, Lamothe S, Labuda D, Lemieux-Blanchard E, Theoret Y, Moghrabi A et al. Role of MTHFR genetic polymorphisms in the susceptibility to childhood acute lymphoblastic leukemia. Blood 2004; 103: 252–257. 16 Skibola CF, Smith MT, Hubbard A, Shane B, Roberts AC, Law GR et al. Polymorphisms in the thymidylate synthase and serine hydroxymethyltransferase genes and risk of adult acute lymphocytic leukemia. Blood 2002; 99: 3786–3791. 17 Thirumaran RK, Gast A, Flohr T, Burwinkel B, Bartram C, Hemminki K et al. MTHFR genetic polymorphisms and susceptibility to childhood acute lymphoblastic leukemia. Blood 2005; 106: 2590–2591, author reply 2591–2592. 18 Schnakenberg E, Mehles A, Cario G, Rehe K, Seidemann K, Schlegelberger B et al. Polymorphisms of methylenetetrahydrofolate reductase (MTHFR) and susceptibility to pediatric acute lymphoblastic leukemia in a German study population. BMC Med Genet 2005; 6: 23. 19 Devlin B, Roeder K. Genomic control for association studies. Biometrics 1999; 55: 997–1004. 20 Thirumaran RK, Bermejo JL, Rudnai P, Gurzau E, Koppova K, Goessler W et al. Single nucleotide polymorphisms in DNA repair genes and basal cell carcinoma of skin. Carcinogenesis 2006; 27: 1676–1681. 21 Westfall PH, Young SS. Resampling-Based Multiple Testing. John Wiley & Sons: New York, 1993, pp. 1–360. 22 Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA. Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet 2002; 70: 425–434. 23 Sharp L, Little J. Polymorphisms in genes involved in folate metabolism and colorectal neoplasia: a HuGE review. Am J Epidemiol 2004; 159: 423–443. 24 Gemmati D, Ongaro A, Scapoli GL, Della Porta M, Tognazzo S, Serino ML et al. Common gene polymorphisms in the metabolic

25

26

27

28

29

30

31

32

33 34

folate and methylation pathway and the risk of acute lymphoblastic leukemia and non-Hodgkin’s lymphoma in adults. Cancer Epidemiol Biomarkers Prev 2004; 13: 787–794. Gaughan DJ, Kluijtmans LA, Barbaux S, McMaster D, Young IS, Yarnell JW et al. The methionine synthase reductase (MTRR) A66G polymorphism is a novel genetic determinant of plasma homocysteine concentrations. Atherosclerosis 2001; 157: 451–456. Zavadakova P, Fowler B, Suormala T, Novotna Z, Mueller P, Hennermann JB et al. cblE type of homocystinuria due to methionine synthase reductase deficiency: functional correction by minigene expression. Hum Mutat 2005; 25: 239–247. Wilson A, Leclerc D, Rosenblatt DS, Gravel RA. Molecular basis for methionine synthase reductase deficiency in patients belonging to the cblE complementation group of disorders in folate/ cobalamin metabolism. Hum Mol Genet 1999; 8: 2009–2016. Stolzenberg-Solomon RZ, Qiao YL, Abnet CC, Ratnasinghe DL, Dawsey SM, Dong ZW et al. Esophageal and gastric cardia cancer risk and folate- and vitamin B(12)-related polymorphisms in Linxian, China. Cancer Epidemiol Biomarkers Prev 2003; 12: 1222–1226. Shi Q, Zhang Z, Li G, Pillow PC, Hernandez LM, Spitz MR et al. Polymorphisms of methionine synthase and methionine synthase reductase and risk of lung cancer: a case–control analysis. Pharmacogenet Genomics 2005; 15: 547–555. Zhang Z, Shi Q, Liu Z, Sturgis EM, Spitz MR, Wei Q. Polymorphisms of methionine synthase and methionine synthase reductase and risk of squamous cell carcinoma of the head and neck: a case–control analysis. Cancer Epidemiol Biomarkers Prev 2005; 14: 1188–1193. Vaughn JD, Bailey LB, Shelnutt KP, Dunwoody KM, Maneval DR, Davis SR et al. Methionine synthase reductase 66A-G polymorphism is associated with increased plasma homocysteine concentration when combined with the homozygous methylenetetrahydrofolate reductase 677C-T variant. J Nutr 2004; 134: 2985–2990. de Jonge R, Hooijberg JH, van Zelst BD, Jansen G, van Zantwijk CH, Kaspers GJ et al. Effect of polymorphisms in folate-related genes on in vitro methotrexate sensitivity in pediatric acute lymphoblastic leukemia. Blood 2005; 106: 717–720. Crawford DC, Nickerson DA. Definition and clinical importance of haplotypes. Annu Rev Med 2005; 56: 303–320. Mori H, Colman SM, Xiao Z, Ford AM, Healy LE, Donaldson C et al. Chromosome translocations and covert leukemic clones are generated during normal fetal development. Proc Natl Acad Sci USA 2002; 99: 8242–8247.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Leukemia