• Open Access

Association of dietary and genetic factors related to one-carbon metabolism with global methylation level of leukocyte DNA

Authors


To whom correspondence should be addressed.

E-mail: moiwasak@ncc.go.jp

Abstract

Global hypomethylation of leukocyte DNA has been associated with an increased risk of cancer. As dietary and genetic factors related to one-carbon metabolism may influence both the methylation and synthesis of DNA, we investigated associations between these factors and the global methylation level of peripheral blood leukocyte DNA based on a cross-sectional study of 384 Japanese women. Dietary intake of folate and vitamins B2, B6, and B12 was assessed with a validated semiquantitative food frequency questionnaire. Five polymorphisms in methylenetetrahydrofolate reductase (MTHFR) (rs1801133 and rs1801131), methionine synthase (MTR) (rs1805087), and methionine synthase reductase (MTRR) (rs10380 and rs162049) were genotyped. Global DNA methylation of leukocyte DNA was quantified using Luminometric Methylation Assay. A linear trend of association between methylation and dietary and genetic factors was evaluated by regression coefficients in a multivariable linear regression model. Mean global methylation level (standard deviation) was 70.2% (3.4) and range was from 59.0% to 81.2%. Global methylation level significantly decreased by 0.36% (95% confidence interval, 0.03–0.69) per quartile category for folate level. Subgroup analysis suggested that alcohol drinking modified the association between folate intake and global methylation level (Pinteraction = 0.01). However, no statistically significant association was observed for intake of vitamins B2, B6, and B12, alcohol consumption, or five single nucleotide polymorphisms of MTHFR, MTR, and MTRR. We found that higher folate intake was significantly associated with a lower level of global methylation of leukocyte DNA in a group of healthy Japanese females.

DNA methylation plays an important role in the epigenetic mechanism of gene regulation[1, 2] and cellular differentiation.[3] Aberrant genomic DNA methylation, both in specific genes and in the genome overall, is widely recognized to be associated with cancer.[4] For example, hypermethylation at promoter CpG islands in tumor suppressor genes is an important means of silencing transcription in carcinogenesis.[4] Global DNA hypomethylation in normally methylated regions is thought to contribute to carcinogenesis through the induction of genomic instability.[4] In addition, some previous evidence suggests that DNA hypomethylation could lead to the activation of oncogenes, and global DNA hypomethylation has been linked to hypomethylation in multiple promoter CpG islands.[4, 5] Although many studies have investigated aberrant DNA methylation at the tissue level, there is great interest in epigenetic markers in peripheral blood and several epidemiological studies have found that hypomethylation of global peripheral blood cell DNA is associated with an increased cancer risk.[6-9] However, determinants of global methylation level among healthy individuals remain largely unexplored.

Folate and vitamin Bs in one-carbon metabolism are cofactors and cosubstrates for methylation and nucleic acid synthesis and also function as regulatory molecules of these processes.[10] Accumulating epidemiological evidence has suggested that folate intake is associated with a decreased risk of some sites of cancer such as esophagus, colorectum, and pancreas,[11] which implies that folate is associated with cancer risk through the mechanisms of DNA methylation and DNA synthesis. As folate is a universal methyl donor, which is necessary in DNA methylation, it is considered to be a potential determinant of the global methylation level of leukocyte DNA.[12] Intervention studies have suggested that folate might alter DNA methylation levels, but findings have been inconsistent.[12-16] Only a few of the previous observational studies examined associations of dietary and genetic factors related to one-carbon metabolism with global methylation level of leukocyte DNA, and their overall findings showed no association.[6-8] These inconsistent findings might be explained by differences in exposure level of nutrients related to one-carbon metabolism, differences in assay methods of global methylation level, and difference in the distribution of genetic factors related to one-carbon metabolism, either alone or in combination. In particular, no study has investigated the interaction of genetic factors such as SNPs and nutrient intake related to one-carbon metabolism with DNA methylation level.

Here, we used the well-characterized control group of a breast cancer case–control study in Nagano, Japan, to carry out a cross-sectional study to evaluate the associations of dietary and genetic factors related to one-carbon metabolism with the global methylation level of peripheral blood leukocyte DNA among Japanese women.

Materials and Methods

Study subjects

Subjects were the control group in a multicenter, hospital-based case–control study of breast cancer carried out from May 2001 to September 2005 at four hospitals in Nagano Prefecture, Japan. Details of this study have been described previously.[17, 18] The study protocol was approved by the institutional review board of the National Cancer Center, Tokyo, Japan.

Briefly, healthy female individuals were selected from medical check-up examinees in two of the hospitals and confirmed to not have any cancer. Each subject was recruited as a control and matched for each case by age (within 3 years) and residential area during the study period; the cases were a consecutive series of 405 women aged 20–74 years with newly diagnosed, histologically confirmed invasive breast cancer who were admitted to one of the four hospitals during the survey period. Among potential control subjects, one declined to participate and two refused to provide a blood sample. Consequently, written informed consent was obtained from 405 matched pairs.

Data collection

Participants were asked to complete a self-administered questionnaire that included questions on demographic characteristics, anthropometric factors, smoking habit, family history of cancer, physical activity, medical history, and menstrual and reproductive history. Dietary habits were investigated using a 136-item semiquantitative FFQ that was developed and validated in a Japanese population.[19, 20] In the FFQ, participants were questioned on how often they consumed the individual food items (frequency of consumption), as well as relative sizes compared to standard portions. Daily food intake was calculated by multiplying the frequency of each food item in the FFQ by its standard portion and relative size. Daily intakes of nutrients were calculated using the 5th revised and enlarged edition of the Standard Tables of Food Composition in Japan.[21] The validity of nutrient intakes estimated from the FFQ was evaluated in a subsample of the Japan Public Health Center-based Prospective Study, which includes Nagano as one its study areas. Estimated intake according to the FFQ was compared to that in four 7-day dietary records, one carried out in each of the four seasons. Spearman's correlation coefficients between energy-adjusted intakes estimated from the FFQ and from dietary records were 0.35–0.50 for folate, 0.34–0.45 for vitamin B2, 0.36–0.47 for vitamin B6, and 0.27–0.35 for vitamin B12.[18, 19]

Participants provided blood samples at the time they returned their self-administered questionnaire. Whole blood in a 7-mL EDTA-2Na Vacutainer (Terumo, Tokyo, Japan) and serum samples were stored at −80°C until analyzed.

Laboratory analysis

Genomic DNA was extracted from the whole blood using a Qiagen FlexiGene DNA Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol.

Global DNA methylation was quantified by LUMA,[22, 23] consequent to the findings of our in-house testing of several methods which found LUMA to be relatively reliable and unbiased in assessing small differences in global methylation levels of peripheral blood leukocytes. Three hundred nanograms of genomic DNA was cleaved with HapII + EcoRI or MspI + EcoRI in two separate 20-μL reaction tubes containing 2 μL of 10× T buffer (330 mM Tris-acetate, 100 mM Mg-acetate, 660 mM K-acetate, 5 mM DTT), 2 μL of 0.1% BSA, and 5 units of each of the restriction enzymes. The reactions were set up in a PSQ96 Plate Low (Qiagen) and incubated at 37°C for 1 h. Then 20 μL annealing buffer containing 200 mM Tris-acetate and 50 mM Mg-acetate (pH 7.6) was added to the cleavage reactions, and samples were assayed using a PSQ96 MA system (Biotage, Uppsala, Sweden). The instrument was programmed to add dNTPs in six steps, consisting of: step 1, dATPαS; step 2, mixture of dGTP + dCTP; step 3, dTTP; step 4, mixture of dGTP + dCTP; step 5, water; and step 6, dATP. Peak heights were calculated using the PSQ96 software. The HapII/EcoRI and MspI/EcoRI ratios were calculated as (dGTP + dCTP)/dATP for each reaction. The HapII/MspI ratio was then calculated as (HapII/EcoRI)/(MspI/EcoRI), which corresponds to the proportion of unmethylated CCGG. Restriction enzymes (HapII, MspI, and EcoRI) were purchased from Takara Bio (1053A, 1150A and 1040A, respectively; Shiga, Japan). PyroMark Gold Q96 Reagents for pyrosequencing were purchased from Qiagen (972804). DNA quantification was carried out using a Quan-iT PicoGreen dsDNA Reagent and kit (P7581; Invitrogen, Carlsbad, CA, USA). Intra-assay CV was 6.4% at a mean methylation level of 74% (n = 20).

In the present study, we focused on three genes, MTHFR, MTR, and MTRR, which are closely related to DNA methylation in one-carbon metabolism, and selected SNPs in consideration of the availability of functional information. Five polymorphisms in MTHFR (rs1801133 and rs1801131), MTR (rs1805087), and MTRR (rs10380 and rs162049) genes were genotyped by TaqMan SNP Genotyping Assays developed by Applied Biosystems (Foster City, CA, USA). Confirmation that the genotype frequencies were in Hardy–Weinberg equilibrium was done using a χ2-test as quality control (all P values >0.05).

Statistical analysis

Nutrient intake (folate, vitamin B2, B6, and B12 intake) was adjusted for total energy intake using the residual method[24, 25] and divided into quartile categories. Adjusted mean global methylation levels of leukocyte DNA were calculated according to nutrient intake and SNPs related to one-carbon metabolism using a multivariable linear regression model. To test linear trends for mean folate intake levels, regression coefficients (β) were calculated in the multivariable linear regression model using categories of each folate intake level as ordinal variables. The following variables were used for adjustment: age (continuous); BMI (continuous); smoking (never smokers, past smokers, current smokers); alcohol drinking (non-drinkers, occasional drinkers, regular drinkers of <150 g ethanol/week, regular drinkers of ≥150 g ethanol/week); and physical activity in the past 5 years (no, ≤2 days/week, ≥3 days/week). To investigate potential effect modification, subgroup analyses were carried out by nutrient intake and SNPs related to one-carbon metabolism, and tests for interaction were carried out. All reported P-values are two-sided, and significance level was set at < 0.05. All statistical analyses were done with sas software version 9.1 (SAS Institute, Cary, NC, USA).

Results

After exclusion of subjects who reported extremely low or high total energy intake (<500 or ≥4000 kcal, respectively) or had no DNA sample, 384 healthy Japanese women were included in the present analyses. Mean age and total calorie intake of women in the present study was 54.1 years and 1947.5 kcal, respectively. Mean global methylation level (SD) was 70.2% (3.4) and range was from 59.0% to 81.2%. Table 1 shows global methylation level according to age, BMI, smoking status, and physical activity, which were used for adjustment in Tables 2 and 3. None of these factors was associated with the level of global methylation.

Table 1. Global methylation level of leukocyte DNA in Japanese women according to factors used for adjustment
FactorLevelCrudeMultivariate-adjusteda
n Methylation level (%)Methylation level (%)95% CIEffect95% CIP-value
  1. a

    Adjusted for age (continuous), body mass index (continuous), smoking (never smoker, past smoker, current smoker), alcohol drinking (non-drinker, occasional drinker, regular drinker of <150 g ethanol/week, regular drinker of ≥150 g ethanol/week), and physical activity in the past 5 years (no, ≤2 days/week, ≥3 days/week). Model for each factor listed in the table did not include the corresponding variable as adjustment. CI, confidence interval.

Age, years<402569.970.468.772.1    
40–4910970.070.669.571.7    
50–5912970.370.969.872.1    
60–699670.471.270.072.4    
≥702569.770.568.872.1    
Trend     0.15−0.20.490.403
Body mass index (quartile category)≤20.89370.170.869.671.9    
20.9–22.59771.071.770.572.9    
22.6–24.89770.170.869.671.9    
≥24.99769.570.269.171.4    
Trend     −0.26−0.570.050.104
SmokingNever35470.170.569.971.1    
Past870.870.768.373.1    
Current2070.771.369.772.9    
Trend     0.36−0.401.130.351
Physical activityNo23169.970.469.371.4    
≤2 days per week3370.971.470.072.8    
≥3 days per week12070.570.869.771.9    
Trend     0.25−0.140.630.208
Table 2. Global methylation level according to five dietary factors and five single nucleotide polymorphisms of genes associated with folate metabolic enzymes
FactorLevelCrudeMultivariate-adjusteda
n Methylation level (%)Methylation level (%)95% CIEffect95% CIP-value
  1. a

    Adjusted for age (continuous), body mass index (continuous), smoking (never smoker, past smoker, current smoker), alcohol drinking (non-drinker, occasional drinker, regular drinker of <150 g ethanol/week, regular drinker of ≥150 g ethanol/week), and physical activity in the past 5 years (no, ≤2 days/week, ≥3 days/week). CI, confidence interval; MTHFR, methylenetetrahydrofolate reductase; MTR, methionine synthase; MTRR, methionine synthase reductase.

Folate (μg/day)≤339.99670.371.270.072.3    
339.9–419.59670.671.470.272.6    
419.5–521.79670.170.869.671.9    
>521.79669.770.269.071.4    
Trend     −0.36−0.69−0.030.030
Vitamin B2 (mg/day)≤1.49669.870.669.571.7    
1.4–1.69670.070.869.672.0    
1.6–1.89670.971.470.272.6    
>1.89670.070.769.571.9    
Trend     0.08−0.240.390.636
Vitamin B6 (mg/day)≤1.49670.271.169.972.3    
1.4–1.69670.371.169.972.2    
1.6–1.89670.170.769.671.9    
>1.89670.170.669.471.7    
Trend     −0.19−0.530.150.268
Vitamin B12 (μg/day)≤6.49670.571.370.172.5    
6.4–8.39669.970.669.471.7    
8.3–10.69670.370.969.872.1    
>10.69670.070.769.571.8    
Trend     −0.14−0.460.170.370
Alcohol drinkingNon-drinker23269.970.469.471.5    
Occasional drinker3970.971.470.072.8    
Regular drinker of <150 g ethanol/week8770.871.370.172.4    
Regular drinker of ≥150 g ethanol/week2669.970.368.871.9    
Trend     0.23−0.110.570.183
MTHFR rs1801131AA25470.370.969.972.0    
AC + CC13070.070.769.671.8    
Dominant model     −0.25−0.980.470.494
MTHFR rs1801133CC11270.270.969.772.0    
CT + TT27270.170.869.871.9    
Dominant model     −0.04−0.800.720.918
MTR rs1805087AA25770.371.070.072.1    
AG + GG12669.870.569.471.6    
Dominant model     −0.53−1.260.200.156
MTRR rs162049GG11670.270.969.772.0    
AG + GG26670.270.869.871.9    
Dominant model     −0.05−0.800.700.902
MTRR rs10380CC30270.270.969.871.9    
CT + TT8170.270.869.672.0    
Dominant model     −0.09−0.930.750.834
Table 3. Association between mean global methylation level in leucocyte DNA and folate intake by factors related to one-carbon metabolism
FactorLevelNumberMultivariate-adjusteda
Effect95% CIP-value for trendP-value for interaction
  1. a

    Adjusted for age (continuous), body mass index (continuous), smoking (never smoker, past smoker, current smoker), alcohol drinking (non-drinker, occasional drinker, regular drinker of <150 g ethanol/week, regular drinker of ≥150 g ethanol/week), and physical activity in the past 5 years (no, ≤2 days/week, ≥3 days/week). CI, confidence interval; MTHFR, methylenetetrahydrofolate reductase; MTR, methionine synthase; MTRR, methionine synthase reductase.

AlcoholNon-drinker232−0.70−1.12−0.280.001 
Drinker1520.08−0.400.550.7490.013
Vitamin B2, mg/day≤1.6192−0.32−0.820.180.208 
>1.6192−0.67−1.15−0.190.0060.304
Vitamin B6, mg/day≤1.6192−0.07−0.650.500.803 
>1.6192−0.63−1.17−0.100.0200.157
Vitamin B12, μg/day≤8.3192−0.20−0.640.250.389 
>8.3192−0.53−0.99−0.080.0220.287
MTHFR rs1801131AA254−0.46−0.86−0.050.028 
AC + CC130−0.18−0.700.330.4840.400
MTHFR rs1801133CC112−0.25−0.810.310.384 
CT + TT272−0.41−0.80−0.020.0370.627
MTR rs1805087AA257−0.21−0.610.190.298 
AG + GG126−0.60−1.13−0.080.0240.233
MTRR rs162049GG116−0.49−1.050.070.084 
AG + GG266−0.30−0.680.090.1370.555
MTRR rs10380CC302−0.37−0.73−0.010.042 
CT + TT81−0.32−1.030.390.3740.892

Global methylation levels according to five dietary factors and five SNPs of genes for folate metabolic enzymes are shown in Table 2. We found a statistically significant association between folate intake level and the global methylation level of leukocyte DNA (P = 0.030). Global methylation level decreased by 0.36% (95% CI, 0.03–0.69) per quartile category for folate intake. No associations were found for vitamin B2, B6, or B12 intake, alcohol drinking, or five SNPs of MTHFR, MTR, and MTRR.

Association between mean global methylation level of leukocyte DNA and folate intake by factors related to one-carbon metabolism are shown in Table 3. Subgroup analyses revealed that alcohol drinking modified the association between folate intake and global methylation level (Pinteraction = 0.01). The global methylation level significantly decreased by 0.70% (95% CI, 0.28–1.12) per quartile category for folate intake among non-drinkers, whereas no association was observed among drinkers (0.08% [95% CI, −0.40–0.55]). Additional analysis by the four categories of alcohol drinking used in Table 2 also found a statistically significant interaction (Pinteraction = 0.002). As stated above, we observed an inverse association among non-drinkers. In contrast, the global methylation level significantly increased by 1.32% (95% CI, 0.22–2.42) per quartile category for folate intake among regular drinkers of more than 150 g ethanol/week, but no association was seen among occasional drinkers and regular drinkers of less than this amount (data not shown). No statistically significant interactions were observed for vitamin B2, B6, or B12 intake, alcohol consumption, or five SNPs of MTHFR, MTR, and MTRR.

Discussion

In this cross-sectional study among Japanese women, we found that higher folate intake was significantly associated with a lower level of global methylation of peripheral blood leukocyte DNA. Subgroup analysis suggested that alcohol drinking modified the association between folate intake and global methylation level. Because of the cross-sectional nature of the study, we were not able to determine if higher dietary folate intake leads to global hypomethylation of leukocyte DNA. Considering the role of folate in one-carbon metabolism, however, our findings suggest that dietary folate intake might modulate the global methylation level of leukocyte DNA.

Our findings appear to contradict at least some previous studies of the association between folate level and global methylation level of peripheral blood DNA. Two intervention studies showed decreased methylation of leukocyte DNA in a folate-depleted diet group.[14, 16] One of these studies provided an average of 118 μg folate per day to 33 postmenopausal women for 7 weeks,[16] and the second provided an average of 56–111 μg folate per day to eight postmenopausal women for 9 weeks.[14] Although these studies differed from our study in their method of methylation analysis (in vitro [3H]methyl incorporation assay by SssI CpG methylase) and subjects (postmenopausal or elderly women recruited in the USA), their data indicate that moderate folate depletion induces hypomethylation of leukocyte DNA. Regarding the effect of folate supplementation on the methylation level of leukocyte DNA, a randomized controlled trial of 400 μg folic acid supplementation per day (n = 15) or placebo (n = 16) for 10 weeks in patients with colorectal adenoma showed an increase in leukocyte DNA methylation level.[15] In contrast, supplementation with 2 mg folic acid and 20 μg vitamin B12 for 12 weeks did not change this variable.[13] These intervention studies suggest that the effect of folate on the methylation level of leukocyte DNA might depend on dose, but that a dose–response pattern might not be straightforward. For instance, it has been suggested that folates act as inhibitors of dihydrofolate reductase,[26] and that high folate levels could have the same functional effect as a low folate status under certain circumstances.[10, 27] In fact, several animal studies showed that the effect of isolated folate deficiency on genomic DNA methylation in rodent liver and colon was either a decrease or increase.[28, 29]

A recent cross-sectional study reported that a dietary pattern characterized by high intake of vegetables and fruits was associated with a lower prevalence of LINE-1 DNA hypomethylation.[30] In contrast, three other studies found no association between dietary folate intake and global methylation level of leukocyte DNA in the control groups of a head and neck cancer case–control study in the US, a bladder cancer case–control study in Spain, and a gastric cancer case–control study in Poland.[6-8] These findings should be interpreted cautiously, however, because the analyses of the bladder and gastric cancer case–control studies were primarily aimed at identifying potential confounders for assessing an association between global methylation level and the risk of cancer based on univariate analyses.

The important messages from this and these previous studies may be that: (i) the mechanisms of individual variation in the global DNA methylation level of peripheral blood leukocytes are complex and multifactorial in nature; and (ii) in actual daily dietary life, in Japan, folate intake may not be the major single determinant of global methylation level and may not necessarily confound association analysis between leukocyte global methylation and the risk of cancers that are associated with folate intake. Only a few observational studies have examined associations of dietary and genetic factors related to one-carbon metabolism with global methylation level of leukocyte DNA among healthy individuals based on nutrient intake estimated from the usual diet alone.[6-8] None of the five candidate SNPs examined in this study showed a statistically significant association, although rs1801131 and rs1801133 in MTHFR, for instance, have been reported to be linked to altered enzymatic activity[31, 32] and folate level.[33, 34] Given the present result of rs1801131 in MTHFR (AA genotype group: number = 254, mean = 70.25, and SD = 3.3; AC + CC genotype group: number = 130, mean = 70.01, and SD = 3.3), for example, the expected power to detect an association was 10% with a two-sided and error level of 5%. Therefore, we cannot exclude the possibility that the null findings are explained by insufficient power, and additional larger studies are needed to clarify the association between these SNPs and global methylation level.

Subgroup analyses in the present study showed that alcohol drinking modified the association between folate intake and global methylation level (Pinteraction = 0.01). The association between folate intake and global methylation level varied by alcohol drinking status: higher folate intake was significantly associated with a lower global methylation level among non-drinkers; no association was observed among occasional and light drinkers; and higher folate intake was significantly associated with a higher global methylation level among relatively heavy drinkers. Alcohol consumption interferes with folate metabolism[35] and decreases levels of serum folate.[36] Although this interaction remained inexplicable, these findings might nevertheless provide hints about its biological mechanism. Furthermore, subgroup analysis by alcohol drinking was based on a relatively small number subjects, particularly with regard to heavy drinkers (n = 26), and thus replication of this interaction in a larger study is awaited.

Several limitations of the present study warrant mention. First, misclassifications due to inaccurate measurement would result in null associations. Although dietary intakes in the present study were assessed using a validated FFQ, misclassifications may have been unavoidable. However, as reproducibility of the assay for global methylation level was relatively high in the present study (intra-assay CV, 6.4), measurement errors during laboratory assay might have been minimal. Second, the present study made multiple comparisons, which might have led to false-positive results. In this regard, we observed a statistically significant association between higher folate intake and lower level of global methylation, which might nevertheless be explained by chance. Finally, because the sample size was limited, the study might not have had sufficient statistical power to detect small associations, as mentioned above, and this is one of the possible explanations for the observed absence of associations. In particular, the results of subgroup analysis and interaction tests should be interpreted carefully.

In this cross-sectional study in 384 healthy Japanese women with validated FFQ data, we found that a higher folate intake level was associated with a lower global methylation level of leukocyte DNA. Although the data of this study and others suggest that folate intake can modulate the global methylation level of leukocyte DNA, inconsistencies among the studies have been noted, and may reflect the complex and multifactorial nature of individual variation in the global DNA methylation level of peripheral blood leukocytes.

Acknowledgments

We thank Yoko Odaka and Misuzu Okuyama for their technical assistance. This study was supported by: a Grants-in-Aid for the Third Term Comprehensive Ten-Year Strategy for Cancer Control and for Research on Applying Health Technology from the Ministry of Health, Labor and Welfare of Japan; the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation; and Grants-in-Aid for Scientific Research on Priority Areas (17015049), for Scientific Research on Innovative Areas (221S0001), and for Young Scientists (B) (22700934) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, the Japan Society for the Promotion of Science, and the Foundation for Promotion of Cancer Research in Japan.

Disclosure Statement

The authors have no conflict of interest.

Abbreviations
BMI

body mass index

CI

confidence interval

CV

coefficient of variation

FFQ

food frequency questionnaire

LUMA

LUminometric Methylation Assay

MTHFR

methylenetetrahydrofolate reductase

MTR

methionine synthase

MTRR

methionine synthase reductase

SD

standard deviation

SNP

single nucleotide polymorphism

Ancillary