Introduction

Breast cancer is the second most common cancer in Thai women and the incidence is still increasing [1]. A wide variety of genetic damage induced by endogenous metabolites and exogenous hazards may contribute to the etiology of breast cancer. Folate is an important nutrient required for DNA synthesis, and it is also involved in the methionine metabolic pathway, which is crucial for DNA methylation [2]. At least 30 different enzymes are involved in this complex pathway including methylenetetrahydrofolate reductase (MTHFR), methionine synthase (MTR), methionine synthase reductase (MTRR), and thymidylate synthase (TYMS). Defects or polymorphic variations in the folate metabolic pathway may influence cancer susceptibility [3].

Evidence level seems to be high that alcohol increases risk of breast cancer [4]. A pooled analysis of 4,335 breast cancer cases and >300,000 controls suggested that intake of 2–5 drinks/day increased risk by roughly 40% [5]. The underlying mechanisms are not firmly established [6] but may include an influence on circulating levels of estrogens [7], immune function, enhanced permeability of chemical carcinogens, decreased absorption of essential nutrients [8], or through metabolism of alcohol to acetaldehyde, a known carcinogen [9].

A relative folate deficiency may develop in individuals who chronically consume more than moderate amounts of alcohol because of the negative effects of alcohol on folate metabolism, including malabsorption, increased excretion, or enzymatic suppression [10]. The potential for high folate intake to counteract the elevated risk of breast cancer associated with alcohol consumption has been illustrated by data from a number of studies [11–14].

To further investigate the role of these pathways in mammary carcinogenesis, we analyzed the association between breast cancer and 27 SNPs in 13 key genes involved in one-carbon and alcohol metabolism on 570 cases and 497 controls in Thai women. We also investigated gene–gene and gene–environment interactions.

Materials and methods

Study population

Cases were all new incident breast cancer patients histologically diagnosed at the National Cancer Institute in Bangkok and at the hospital in Khon Kaen province of North Eastern of Thailand during the period of May 2002–March 2004 and August 2005–August 2006, with a participation rate of 99.6% (600/602). Controls were randomly selected from healthy women who visited patients admitted to the same hospitals for diseases other than breast or ovarian cancer. All of the 642 control individuals were recruited during the same study period as the case ascertainment. The participation rate among visitors who were asked to participate was 98.9% (642/649). Informed consent was obtained from all participants and a structured questionnaire was administered by trained interviewers to collect information on demographic and anthropometric data, reproductive and medical history, residential history, physical activity and occupation as well as diet (see Table 1). Lifestyle exposure parameters were reported as follows: tobacco smoking: less than or equal to 6 months of smoking in life, or if she smokes longer than 6 months, the sum of cigarettes smoked is less than or equal to 50 per 6 months (non-smoker) versus more than 50 cigarettes over a 6-month period (smoker); involuntary tobacco smoking: less versus more than or equal to 1 h of exposure per day; alcohol consumption: less versus more than or equal to once a week for at least 6 months. Approximately 7 ml of blood were collected from participants, but 30 cases and 145 controls refused to give blood samples. In total, blood samples of 570 cases and 497 controls were included in the genotype analysis, resulting in a participation rate of 95.0% (cases) and 77.4% (controls). The study was approved by the ethical review committee for research in human subjects, Ministry of Public Health, Thailand and by the ethics committee of National Cancer Center, Japan.

Table 1 Selected characteristics of study population
Full size table

Genotyping analysis

Genomic DNA was isolated from buffy coats using a QIAmp DNA blood kit (Qiagen, Hilden, Germany). DNA concentrations were measured by PicoGreen dsDNA qualification kits (Molecular Probes, Leiden, The Netherlands). All SNPs were analyzed by TaqMan 5′ nuclease assay using the ABI PRISM 7900 HT Sequence Detection System (Applide Biosystems LLC, Foster city, CA, USA). Oligonucleotide primers and the dual labeled allele specific probes were designed by ABI. PCR were performed in 384-well plates with each plate containing four control samples. A set of three 384-well plates were prepared to accommodate 570 cases and 497 control subjects and used for genotyping. Genomic DNA (5 ng) was amplified in a total volume of 5 μl in the presence of 100 μM of each of the dNTPs, 3 pmols of each of appropriate primers, 2 pmols of each of the corresponding dual labeled probes, and 0.025 units of Taq DNA polymerase. PCR cycling consisted of 40 cycles at 94°C for 15 s, 55–60°C for 15 s, and 72°C for 15 s. The results of 5% blindly repeated samples were at least 99% concordant each other. Genotyping success rate for individual polymorphisms averaged 95%.

Statistical analyses

Hardy–Weinberg equilibrium (HWE) testing was used as one of the measures for a quality control for genotyping, and allele and genotype frequencies were calculated. The multivariate logistic regression analyses were applied to evaluate differences in genotype distributions, and the odds ratios (OR) and their 95% confidence intervals (CI) were calculated after adjustment for the following covariates: age, body mass index (BMI), smoking, pregnancy and breast feeding, family history of breast cancer in the first-degree relatives, education, and menopausal status. Alcohol consumption was not included, because it was not associated with the breast cancer in our study (Table 2). For some SNPs, additional tests were also performed by Fisher’s exact test on allelic contingency table and by Cochran–Armitage trend test to confirm the observed genotype-specific associations.

Table 2 SNPs analyzed in this study
Full size table

A stratified analysis was performed as an exploratory, adjunct analysis. Selected strata were menopausal, pregnancy, breast feeding, oral contraceptive use, estrogen receptor, and progesterone receptor status. The adjusted ORs and 95% CIs for OR were calculated by the multivariate logistic regression analyses. In addition, gene–gene interactions and gene–environment interactions were evaluated by logistic model including an interaction term between genes and those between gene and environmental factors.

The statistical significance was defined as P ≤ 0.05, and adjustment for multiple testing, which is absolutely necessary in the validation phase of association studies, was not performed due to an exploratory hypothesis-generating nature of the study.

All statistical analyses were carried out using the Statistical Analysis System (SAS) software Version 9.1 (SAS Institute Inc, Cary, NC), and partially the R suite (http://www.r-project.org/).

Results

Characteristics of the study population were compared by case–control status as shown in Table 1. The mean age of controls (43.2 ± 12.4 years) was significantly lower (P < 0.01) than that of breast cancer patients (46.0 ± 10.6 years). Pregnancy, menopausal status, breast feeding, BMI, involuntary tobacco smoking, family history of breast cancer, and education were different between cases and controls. However, as to oral contraceptive use, smoking, and alcohol consumption, no significant differences were found between cases and controls.

Frequencies of variant alleles among the control population are shown in Table 2. All SNP frequencies were in Hardy–Weinberg equilibrium (HWE) among controls, except two SNPs in MTHFR (rs1801131) and NAT2 (rs1799930). The result of association analysis of individual SNPs is shown in Table 3. Homozygotes of minor alleles of MTR SNPs (rs1770449 and rs1050993) were associated with an increased risk of breast cancer with OR = 2.21 (95% CI 1.18–4.16) and 2.24 (95% CI 1.19–4.22), respectively. The SNP in DRD3 (rs167770) was found associated with an increased risk among heterozygote carriers (OR 1.36, 95% CI 1.03–1.80). Although the DRD3 (rs167770) SNP did not show a statistically significant association for the minor allele homozygotes, this SNP was significant when tested for allelic (OR 1.24, 95% CI 1.01–1.53; P = 0.045) and recessive (OR 1.38, 95% CI 1.07–1.77; P = 0.014) models.

Table 3 Association between folate and alcohol metabolism genes and breast cancer
Full size table

A stratified analysis by menopausal status suggested tendencies that NAT2 (rs1799930), DRD2 (rs10891556), and SLC6A4 (rs140701) polymorphisms are related to breast cancer risk in premenopausal women (OR 2.70, 95% CI 1.20–6.07; OR 1.62, 95% CI 1.03–2.56; and OR 1.56, 95% CI 1.01–2.41, respectively) (Table 4). Among postmenopausal women, an increased risk of breast cancer was suggested for MTRR (rs162049) and DRD3 (rs167770) polymorphisms (OR 1.61, 95% CI 1.07–2.44 and OR 1.59, 95% CI 1.11–2.28, respectively).

Table 4 Association between folate and alcohol metabolism genes and breast cancer stratified by menopausal status
Full size table

Analyses on gene–environment interactions were performed between polymorphisms and alcohol consumption, oral contraceptive use and body mass index. They were exploratory analyses, but no strong interactions were identified (Table 5). Two-gene interactions were also analyzed between SNPs of the DRD3 (rs167770) and other genes showing significant association individually in Table 3. None of gene–gene interactions was statistically significant (Table 6).

Table 5 Association of selected SNPs and breast cancer risk by alcohol, BMI, and physical activity
Full size table
Table 6 Association of DRD3 and MTRR polymorphisms with the risk of breast cancer
Full size table

Further, DRD3, and MTR genotype were correlated with ER and PR status. Fifty-three percent of breast cancer cases were ER-positive tumors (167/314) and 39% were PR-positive tumors (118/301). The associations between DRD3 (rs167770) and MTR (rs1770449 and rs1050993) polymorphisms and breast cancer risk were not different among the ER/PR status (data not shown).

Discussion

Genetic variation in enzymes and other proteins involved in folate and alcohol metabolisms are rational candidates for studying the impact of both genetic and environmental effects and their interactions on breast cancer risk. As a systematic candidate gene approach, we analyzed 12 SNPs in four folate metabolism genes and 15 SNPs in nine alcohol metabolism and behavior genes; these SNPs were selected by our previous study to catalog candidate genes, which are potentially subjected to gene–environment interactions with regard to cancer susceptibilities among Japanese population [15]. Only the SNPs that have minor allele frequencies higher than 5% were evaluated in the study. Of the 27 SNPs, seven suggested an association with breast cancer risk (MTR rs1770449 and rs1050993, MTRR rs162049, DRD2 rs10891556, DRD3 rs167770, SLC6A4 rs140701, and NAT2 rs1799930) in an overall or stratified analysis.

MTHFR is a critical gene in the one-carbon metabolism pathway. Two non-synonymous missense SNPs, C677T (Ala222Val, rs1801133) and A1298C (Glu429Ala, rs1801131), in the coding region were extensively studied. However, the previous association studies on the MTHFR polymorphism and breast cancer risk showed inconsistent results. Several studies have reported that the MTHFR C677T variants were associated with an increased risk of breast cancer in pre-menopausal women [16, 17] or in those with bilateral breast cancer or combined breast and ovarian cancers [18], but some others showed no association between MTHFR C677T and breast cancer [19, 20]. In addition, Sharp et al. [21] reported that MTHFR 1298CC genotype and compound heterozygosity (677CT and 1298AC) were associated with a reduced risk of developing breast cancer. In this study, our finding did not support a role of MTHFR C677T or MTHFR A1298C in modifying breast cancer risk in Thai women.

The 5-methyltetrahydrofolate-homocysteine S-methyltransferase (MTR; also called methionine synthase), which is essential for maintaining adequate intracellular folate pools, catalyzes the remethylation of homocysteine to methionine and has influence on DNA methylation as well as on nucleic acid synthesis [22]. Vitamin B12 is a cofactor in this methylation process. MTR is maintained in its active from by methionine synthesis reductase (MTRR). In this study, the homozygote carriers of MTR (rs1770449 and rs105099) SNPs were associated with breast cancer with OR = 2.21 (95% CI 1.18–4.16) and OR = 2.24 (95% CI 1.19–4.22), respectively. In addition, the A allele of MTRR (rs162049) seems to be associated with breast cancer among postmenopausal women (OR 1.61, 95% CI 1.07–2.44). However, these SNPs have never been reported with breast cancer before and their functional effect on enzymatic activities remain unknown.

Alcohol drinking is among the non-hormonal risk factors for breast cancer (although there may be an indirect relationship). Genetic susceptibility for ethanol metabolism can affect breast cancer, and several genes are known to be involved in this complex pathway. Our study did not find any affect of ADH1C, ALDH2, CYP2E1, GSTP1, and NAT1 polymorphisms on breast cancer risk. However, the homozygotes of minor allele of NAT2 (rs1799930) SNP was associated with an increased risk of breast cancer among premenopausal (OR 2.70, 95% CI 1.20–6.07), although the number of the subjects were limited. Lu et al. [23] showed an interaction effect in bladder cancer between NAT2 polymorphism and alcohol drinking. In addition, Rodrigo et al. [24] reported that NAT2 activity may be a factor that determines the risk of developing alcoholic liver disease. In this study, the genotype distribution of the NAT2 (rs1799930) polymorphism departed from Hardy–Weinberg Equilibrium, and the interpretation of the result should be warranted.

Furthermore, many studies suggested that the intensity of drinking (drink per day) has more effect on risk for breast cancer than recent alcohol or duration of drinking [25, 26]. We, therefore, investigated the SNPs of DRD2, DRD3, and SLC6A4, which are implicated in drinking behavior. In animal studies, alcohol can stimulate dopaminergic neurons in the ventral tegmental area [27, 28], and the density of dopamine D2 receptors in the limbic system is lower in alcohol-preferring rats than in non-preferring rats [29, 30]. Likewise, the number of striatal dopamine D2 receptors is less in alcohol-preferring humans than in healthy control subjects. The A1 polymorphism of DRD2 TaqI A loci has been considered as a risk factor for alcohol dependence [31, 32], but the association between alcoholism and the DRD2 gene remains equivocal in many studies [33–35]. In our study, an association was suggested between the DRD3 SNP (rs167770) in an overall or in postmenopausal breast cancer (OR 1.36, 95% CI 1.03–1.80 and OR 1.59, 95% CI 1.11–2.28, respectively). The DRD2 (rs10891556) polymorphism was associated with an increased risk among premenopausal women (OR 1.62, 95% CI 1.03–2.56). The propensity for severe drinking has been also hypothesized to be regulated by differential expression of serotonin transporter gene SLC6A4 [36]. Our result suggested that SCL6A4 (rs140701) was associated with an increased risk of breast cancer among premenopausal (OR 1.56, 95% CI 1.01–2.41). As the alcohol consumption was relatively low among Thai women in our study, it may have reduced the ability to detect a modifying effect of ethanol on the association of these genes with breast cancer. The current knowledge on genetic polymorphisms related to drinking behavior is far from sufficient. The possible associations between the genetic polymorphisms implicated in drinking behavior and breast cancer risk need to be confirmed in a population with a higher prevalence of alcohol drinking than Thai women.

Certain limitations of this study should be noted. First, data were not available on detailed dietary intake of folate, plasma or erythrocyte folate levels and its precursors or metabolites such as homocysteine, limiting further examination of the gene–nutrient interactions in breast carcinogenesis. In this study, a food frequency questionnaire was administered by trained interviewers to assess dietary and alcohol intake. Unfortunately, we could not calculate the total folate in individual intake due to lack of the standard food composition in Thailand. Second, like most other case–control studies, this study may suffer from recall bias. Third, the statistical power of our study was limited in the stratified analyses because of the small sample size of the subgroups. For instance, if the simple but conservative Bonferroni correction for multiple testing is applied to our data set, none of the SNPs remained statistically significant. Although this study reports a systematic survey of genetic polymorphisms on the folate and alcohol metabolic pathways on breast cancer in Thailand for the first time, it is primarily for a hypothesis generation, and the findings need to be validated in further studies with larger sample size or in meta-analyses, which is also aimed in our laboratory for years to come.

In conclusion, our study has provided some new evidence that either folate metabolism genes or alcohol metabolism and behavior gene polymorphisms may contributes to the etiology of breast cancer among Thai women. Studies should be extended to cover SNPs of other important one-carbon or alcohol metabolism genes, and ascertainment of high quality folate intake information is expected to further elucidate gene–gene and gene–environment interactions in susceptibility of breast cancer.