The folate metabolism pathway contributes to important metabolic processes, such as RNA and DNA synthesis, DNA repair, and DNA methylation (1). Previous observations have suggested a potential relationship between altered folate levels and tumorigenesis (2). Therefore, inherited genetic variation in the gene encoding methylenetetrahydrofolate reductase (MTHFR), an enzyme that regulates the main circulating form of folate, 5-methyltetrahydrofolate, as well as the synthesis of S-adenosyl-l-methionine, the methyl donor for most methyltransferase reactions, may play an important role in the etiology of cancer. Epidemiologic studies of MTHFR and breast cancer have focused on only two common gene variants: the C665T, Ala222Val polymorphism encoding a thermolabile variant allozyme with decreased enzyme activity, and the A1286C, Glu429Ala polymorphism (3). Although significant associations with breast cancer risk have been observed, at least with the 665T variant in premenopausal women (4, 5), the two common polymorphisms studied to date represent only a portion of the sequence variation present in MTHFR (6). Therefore, we set out to assess the association of common MTHFR polymorphisms and haplotypes with breast cancer using a haplotype-tagging approach.

Study Population

Cases were women diagnosed with breast cancer within the previous year seen in Medical Oncology at the Mayo Clinic (Rochester, MN). Cases were frequency matched to controls on age (5-year categories) and region of residence. Controls were selected from general medical examination appointments in the Department of Internal Medicine. All subjects were from the states of Illinois, Iowa, Minnesota, North Dakota, South Dakota, or Wisconsin. Both cases and controls were ineligible if they had a previous diagnosis of cancer (except nonmelanoma skin cancer).

MTHFR Single-Nucleotide Polymorphism Selection and Genotyping

As described elsewhere (6), MTHFR was resequenced in Coriell DNA samples from randomly selected Caucasian-American subjects. Sixteen single-nucleotide polymorphisms (SNP) that captured most of the genetic variability in the gene were selected either through the haplotype-tagging approach of Stram et al. (7) or through the LD-Select method of Carlson et al. (8). All SNPs selected through the Stram method, except I1C (+128), were in common to both methods. These 16 SNPs, plus four nonsynonymous coding SNPs (C400T Arg134Cys, G1556A Arg519His, G1743A Met581Ile, and C1958T Thr653Met), were genotyped using the SNPstream platform (Beckman Coulter, Fullerton, CA) as described elsewhere (9). MTHFR polymorphisms within exons and in 5′ and 3′ untranslated regions were numbered by designating the ‘A’ in the translation initiation codon for the cDNA encoding the 70-kDa isoform as position (+1). cDNA nucleotides located 5′ to that position were assigned negative numbers, whereas those located 3′ were assigned positive numbers. Positions within introns were numbered relative to splice junctions, with the initial 5′ nucleotide in the intron designated (+1).

Statistical Methods

Genotypes for the controls were assessed for departures from Hardy-Weinberg equilibrium. Single SNP analyses were done using logistic regression, where case-control status was the response, and genotypes were modeled as having a log-additive relationship with breast cancer case status. Empirical adjustment for the multiple SNPs tested was achieved by permuting case-control status 10,000 times, doing all single SNP tests for each permutation, and tallying the number of times the smallest resulting P was lower than the smallest observed P. We assessed effect modification of SNPs with menopausal status using standard tests for interaction and did the same permutation procedure to determine whether there was significant evidence of a MTHFR-menopausal status interaction. Finally, a whole-gene test of association between MTHFR haplotypes and breast cancer status was done using the haplotype score test of Schaid et al. (10). All analyses were adjusted for age and region of residence. Analyses were done using Statistical Analysis System (SAS Institute, Cary NC) and S-Plus (Insightful, Seattle, WA).

Power Considerations

The minimum odds ratio (OR) that would be detectable in a study of this size was estimated for a variety of minor allele frequencies using power formulae for the Armitage test for trend (11). We set the power at 80% and used two settings for type I error: 0.05 and 0.005. The more stringent P of 0.005 was selected following empirical calculations that suggested a Bonferroni correction for 10 independent tests would preserve a whole-gene type I error level of 0.05.

A total of 750 cases and 732 controls were included in these analyses. No genetic variation was observed within the three nonsynonymous coding SNPs C400T Arg134Cys, G1556A Arg519His, and G1743A Met581Ile. All SNPs met the Hardy-Weinberg equilibrium assumption except exon 6, C1056T (P < 0.01, Hardy-Weinberg equilibrium test). Associations of individual SNPs with breast cancer risk are presented in Table 1. The common C665T polymorphism that has previously been shown to confer risk for breast cancer in premenopausal women (4, 5) also showed no association. As with the single SNP main effect tests, tests for interaction between individual SNPs and menopausal status and their associations with breast cancer risk also did not reach statistical significance after accounting for multiple testing (P > 0.05 for each). Haplotype analyses for the LD-selected SNPs revealed eight haplotypes with frequencies >1% (Table 2). None of these haplotypes were significantly associated with breast cancer risk. Similar results were found for the SNPs selected by using the haplotype-tagging approach. We did multivariate logistic regression and haplotype analyses that included the following covariates found to be associated with case status in our cohort: age at menarche, menopausal status, family history, education, activity, and alcohol. Multivariate-adjusted results were similar to those presented in the tables (data not shown).

The purpose of this work was to conduct a comprehensive analysis of the possible association between MTHFR polymorphisms or haplotypes and risk for breast cancer. MTHFR plays a central role in the regulation of intracellular folate levels and is an important target for epidemiologic studies of folate metabolism and breast cancer. Previous reports of MTHFR variants and breast cancer have focused primarily on the first common variant identified, C665T, and those studies yielded inconsistent results. The effect of this polymorphism varied depending on menopausal status and folate levels. Two groups reported an increased risk for breast cancer in premenopausal individuals with the homozygous variant (4, 5). Other groups observed no such association (12, 13). The A1286C variant was associated with increased risk in smaller studies (14, 15) but not in studies with larger population samples (12, 16, 17).

Although informative, prior studies were limited to only one or, at most, two candidate polymorphisms in MTHFR. We recently published a report of the systematic resequencing of all coding exons, exon-intron splice junctions, and the 5′-flanking region of MTHFR (6). A total of 65 polymorphisms was identified in the 240 samples examined. The current analysis extends that work. For these analyses, we conducted a systematic selection of tag SNPs from these polymorphisms. We used two methods (7, 8) of SNP selection to determine if one provided greater insight into gene/disease association. Both methods selected essentially the same tagging SNPs.

In our study, we observed no significant association of any of the MTHFR variants or haplotypes with breast cancer risk. With 750 cases and 732 controls, this study had 80% power to detect an OR of ≥1.49 with an allele frequency of 0.10. As shown in Table 1, the detectable per-allele ORs for most of the SNPs were <1.5. We also did not observe any significant effect modification of risk by menopausal status.

In conclusion, our study used a broad coverage of the genetic variation in MTHFR by the use of haplotype-tagging SNPs chosen from the 65 variants identified during a gene resequencing study. No significant association was observed with any of the tagging SNPs or common MTHFR haplotypes after adjustment for multiple comparisons. These results suggest that genetic variation in MTHFR, independent of other factors, such as folate levels (which were not available in the current study), may not play a significant role in the development of breast cancer.

Grant support: NIH grants P01 CA 82267; Breast Cancer Specialized Program of Research Excellence grants P50 CA166201, R01 GM28157, and U01 GM61388; the Pharmacogenetics Research Network.

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.

1
Ulrich CM. Nutrigenetics in cancer research—folate metabolism and colorectal cancer.
J Nutr
2005
;
135
:
2698
–702.
2
Rossi E, Hung J, Beilby JP, Knuiman MW, Divitini ML, Bartholomew H. Folate levels and cancer morbidity and mortality: prospective cohort study from Busselton, Western Australia.
Ann Epidemiol
2006
;
16
:
206
–12.
3
Frosst P, Blom HJ, Milos R, et al. A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase.
Nat Genet
1995
;
10
:
111
–3.
4
Semenza JC, Delfino RJ, Ziogas A, Anton-Culver H. Breast cancer risk and methylenetetrahydrofolate reductase polymorphism.
Breast Cancer Res Treat
2003
;
77
:
217
–23.
5
Campbell IG, Baxter SW, Eccles DM, Choong DY. Methylenetetrahydrofolate reductase polymorphism and susceptibility to breast cancer.
Breast Cancer Res
2002
;
4
:
R14
.
6
Martin YN, Salavaggione OE, Eckloff BW, Wieben ED, Schaid DJ, Weinshilboum RM. Human methylenetetrahydrofolate reductase pharmacogenomics: gene resequencing and functional genomics.
Pharmacogenet Genomics
2006
;
16
:
265
–77.
7
Stram DO, Haiman CA, Hirschhorn JN, et al. Choosing haplotype-tagging SNPS based on unphased genotype data using a preliminary sample of unrelated subjects with an example from the Multiethnic Cohort Study.
Hum Hered
2003
;
55
:
27
–36.
8
Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA. Selecting a maximally informative set of single-nucleotide polymorphisms for association analyses using linkage disequilibrium.
Am J Hum Genet
2004
;
74
:
106
–20.
9
Denomme GA, Van Oene M. High-throughput multiplex single-nucleotide polymorphism analysis for red cell and platelet antigen genotypes.
Transfusion
2005
;
45
:
6606
.
10
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
–34.
11
Slager SL, Schaid DJ. Case-control studies of genetic markers: power and sample size approximations for Armitage's test for trend.
Hum Hered
2001
;
52
:
149
–53.
12
Justenhoven C, Hamann U, Pierl CB, et al. One-carbon metabolism and breast cancer risk: no association of MTHFR, MTR, and TYMS polymorphisms in the GENICA study from Germany.
Cancer Epidemiol Biomarkers Prev
2005
;
14
:
3015
–8.
13
Langsenlehner U, Krippl P, Renner W, et al. The common 677C>T gene polymorphism of methylenetetrahydrofolate reductase gene is not associated with breast cancer risk.
Breast Cancer Res Treat
2003
;
81
:
169
–72.
14
Ergul E, Sazci A, Utkan Z, Canturk NZ. Polymorphisms in the MTHFR gene are associated with breast cancer.
Tumour Biol
2003
;
24
:
286
–90.
15
Chen J, Gammon MD, Chan W, et al. One-carbon metabolism, MTHFR polymorphisms, and risk of breast cancer.
Cancer Res
2005
;
65
:
1606
–14.
16
Le Marchand L, Haiman CA, Wilkens LR, Kolonel LN, Henderson BE. MTHFR polymorphisms, diet, HRT, and breast cancer risk: the multiethnic cohort study.
Cancer Epidemiol Biomarkers Prev
2004
;
13
:
2071
–7.
17
Shrubsole MJ, Gao YT, Cai Q, et al. MTHFR polymorphisms, dietary folate intake, and breast cancer risk: results from the Shanghai Breast Cancer Study.
Cancer Epidemiol Biomarkers Prev
2004
;
13
:
190
–6.