Mol Biol Rep (2013) 40:6547–6560 DOI 10.1007/s11033-013-2686-0
Folate intake and MTHFR polymorphism C677T is not associated with ovarian cancer risk: evidence from the meta-analysis Chenglin Li • Peizhan Chen • Pingting Hu • Mian Li • Xiaoguang Li • He Guo • Jingquan Li Ruiai Chu • Wei Zhang • Hui Wang
•
Received: 4 January 2013 / Accepted: 14 September 2013 / Published online: 17 October 2013 Ó Springer Science+Business Media Dordrecht 2013
Abstract Folate is essential for DNA synthesis and methylation and implicated in the process of carcinogenesis. Several studies inconclusively suggested increased folate intake may reduce ovarian cancer risk. Studies concerning the association between C677T polymorphism in methylenetetrahydrofolate reductase (MTHFR), an important enzyme in folate metabolism, and ovarian cancer risk also resulted in no agreement. The meta-analysis was conducted based on current studies to assess the association between folate intake, the MTHFR C667T polymorphism and ovarian cancer risk. 1,158 cases out of 217,309 participants from four cohort studies, 4,519 cases and 6,031 controls from four case–control studies about folate intake along with 5,617 cases and 9,808 controls from 10 publications concerning the polymorphism were pooled, respectively. We detected no significant association between total folate (RR = 1.04, 95 % confidence interval (CI) = 0.87–1.23) or dietary folate (RR = 0.88, 95 % CI = 0.75–1.05) intake and ovarian cancer risk, and also no significant relationship was found between MTHFR C677T
Chenglin Li and Peizhan Chen have contributed equally to this work. C. Li P. Chen P. Hu M. Li X. Li H. Guo J. Li R. Chu H. Wang (&) Key Laboratory of Food Safety Research, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Shanghai 200031, China e-mail: [email protected] W. Zhang Institute for Food Safety and Health, Illinois Institute of Technology, Bedford Park, IL, USA H. Wang Key Laboratory of Food Safety Risk Assessment, Ministry of Health, Beijing 100021, China
polymorphism and ovarian cancer risk (TT vs. CC: odds ratio (OR) = 1.15, 95 % CI = 0.90–1.46; CT vs. CC: OR = 1.04, 95 % CI = 0.94–1.16). Our analysis indicated neither folate intake nor MTHFR C677T polymorphism is related to altered susceptibility of ovarian cancer. Keywords Folate MTHFR Ovarian cancer Meta-analysis Abbreviations MTHFR Methylenetetrahydrofolate reductase SNP Single nucleotide polymorphism OR Odds ratio 95 % CI 95 % Confidence interval
Introduction Ovarian cancer is the most malignant form of cancer in the female reproductive system, which accounts for about 3 % of all cancers in women. It has been estimated that 21,990 new cases and 15,460 deaths would occur in the US in 2011 [1]. The American Cancer Society reported the overall 1-, 5-, and 10-year survival rate of ovarian cancer patients is approximately 75, 46 and 38 %, respectively [1– 3]. Factors that have been reported to be associated with ovarian cancer include age, obesity, oral contraceptive use, parity, use of fertility drugs and family history of cancer [4]. However, such factors only account for a small proportion of ovarian cancer patients; more comprehensive studies are warranted to fully illustrate the etiology of ovarian cancer. Recently, dietary factors have been found to be associated with the susceptibility of ovarian cancer. High meat and fat intake may increase the risk of epithelial ovarian cancer, while high vitamin A and beta-carotene
123
6548
intake may modestly protect against ovarian cancer. Due to the complexity of the dietary factors, results of previously studies concerning the dietary risk factor were typically inconclusive [4–6]. Folate is a common nutrient in daily vegetables which has been found to be involved in the prevention of cardiovascular disease and neural tube defects [7, 8]. Folate is also involved in the formation of the S-adenosylmethionine, a universal methyl donor essential for the synthesis of purine and thymidine. Considering its important roles in DNA methylation and nucleic acid synthesis, some studies were conducted to evaluate its effects on carcinogenesis [9]. For instance, Kim et al. [10] found a significant protective effect for both dietary (RR = 0.92, 95 % confidence intervals [CI] = 0.84– 1.00) and total folate intake (RR = 0.85, 95 % CI = 0.77–0.95) against colorectal cancer with a meta-analysis pooled of 13 prospective studies. In contrast, Larsson et al. [11] found no relation between folate intake and breast cancer risk. Other recent studies also evaluated the association between ovarian cancer and folate intake with inconsistent results. For example, Larsson et al. [12] reported that higher dietary folate intake may reduce ovarian cancer risk, particularly for those of high alcohol consumption; however, studies by Webb et al. [13] and Harris et al. [14] found little evidence for the protective effects of folate intake against ovarian cancer. Methylenetetrahydrofolate reductase (MTHFR) is a major enzyme involved in folate-related one carbon metabolism, which is responsible for the conversion of 5,10-methylenetetrahydrofolate (5,10-methylene-THF) to 5-methyltetrahydrofolate (5-methyl-THF), the predominant form of folate in plasma. A functional polymorphism (C677T) located on the exon 4 of the MTHFR gene results in the amino acid transition from Ala to Val, further leading to the reduction of enzyme activity [15]. Individuals with homozygote 677TT genotype only have 30 % enzyme activity compared to 677CC carriers. The reduced activity of MTHFR would lead to the accumulation of 5,10-methylene-THF and reduction of 5-methyl-THF level in the cells, which may interrupt the homeostasis of folate metabolism in vivo [16–18]. The C677T polymorphism has been reported to be significantly associated with many cancer types, such as colorectal cancer, breast cancer and lung cancer [19–21]. The association of the variant with ovarian cancer risk was investigated by several studies with no consistent conclusions [22–24]. Due to fact that the folate intake and related signaling genes may play important roles in the carcinogenesis and the existing epidemiological studies do not provide comprehensive analysis on available data, here we conducted the meta-analysis to systematically evaluate the association of folate intake and the MTHFR variant C677T with ovarian cancer susceptibility.
123
Mol Biol Rep (2013) 40:6547–6560
Methods Literature search A comprehensive literature search was conducted on MEDLINE and PubMed databases by two independent researchers to identify relevant studies published prior to 31 December 2012. The search flowchart was shown in Fig. 1. We used the term ‘‘ovarian cancer’’, along with ‘‘folate’’, ‘‘folic acid’’ or ‘‘MTHFR’’ for the search without any other restrictions. The reference lists of the retrieved publications were checked to identify any missing studies in the database search. We selected eligible studies that examined the correlation between folate intake and ovarian cancer risk with the folate intake data in quantiles together with the risk estimates (RR or odds ratio [OR]) and 95 % CIs, or with the results of the highest quantile in contrast with the lowest quantile, or with data that could be used to calculate the risk estimate and its 95 % CI. Studies that examined the association of MTHFR single nucleotide polymorphism (SNP) C677T and ovarian cancer risk should provide sufficiency data for the frequency of the genotypes. We included case–control, cohort or crosssectional studies that were reported in English. Selected publications and data extraction Based on our literature search, we identified eight published studies that evaluated the association between folate intake and ovarian cancer risk, Table 1. The following items from the selected publications were extracted: the first author’s name and year of the publication, study type, location and ethnicity of participants, study period and main characteristics of the study, folate intake categories and their corresponding estimate effects (OR/RR and 95 % CI), the p trend values and related statistical adjustments. For articles related to MTHFR polymorphism and ovarian cancer risk, we selected 10 publications with 12 subgroup studies that investigated the association of C677T variant and ovarian cancer susceptibility, Table 2. The following information of the studies was extracted: the name of first author and the year of publication, study type and location, sample size and genotype distribution in case and control groups. For the 12 subgroup studies, two studies only provided the number of individuals with homozygote TT genotype and the total sample size [23, 24], so the related information was only used in the recessive genetic model rather than other models. Statistical analysis The standard inverse variance weighting method was used to calculate the pooled estimate and its 95 % CI that
Mol Biol Rep (2013) 40:6547–6560 Fig. 1 Work flow chart of the meta-analysis
6549
Publications identified from literature
Publications identified from literature
search using term “folate” or “folic
search using term “MTHFR” in
acid” in combination with “ovarian
combination with “ovarian cancer”
cancer” from PubMed (N=373)and
from PubMed (N=24) and MEDLINE
MEDLINE (N=178) up to Dec. 2012
(N=19) up to Dec. 2012
Exclude overlapped publications Excluded (N=19)
Excluded (N=178)
Check for missing studies from reference lists in publications Added (N=0)
Added (N=0)
Further detailed screening Excluded
Excluded
Not publication in English (N=18)
Not publication in English (N=2)
Not population studies (N=308)
Not population studies (N=10)
Not ovarian cancer (N=8)
Not ovarian cancer (N=1)
Not folate intake(N=31)
Not MTHFR SNP C677T (N=1)
Final data check
8 relevant publications included in meta-analysis
concerning the effect of folate intake and ovarian cancer risk. Due to the relative negligible difference between ORs from case–control studies and RRs from prospective studies numerically, the ORs were assumed as the estimates of RRs for further analysis. Each study was assigned with appropriate weighting using the inverse square of SE for each logRR. To determine the association between ovarian cancer risk and variant MTHFR SNP C677T, the ORs and corresponding 95 % CIs were used to calculate the pooled estimate and 95 % CI. The standard inverse variance weighting method was used to calculate the pooled OR and its 95 % CI under the fixed-effects model and the
10 relevant publications included in meta-analysis
DerSimonian–Laird method were used to calculate the pooled estimate under the random-effects model. Statistical heterogeneity among studies was evaluated by Q test and I2 statistics. We considered significant heterogeneity existing when p value was less than 0.05 for the Q test or I2 [25 % in I2 statistics. Results from randomeffects model were acceptable for interpretation when significant heterogeneity among studies was detected; otherwise, the results from the fixed-effects model were used. Publication bias was assessed by the Egger’s linear regression test and a p value less than 0.05 indicated significant publication bias. For the sensitivity test, we excluded a single study from the meta-analysis each time to
123
123
Study types
General Canadian population
Swedish Mammography Cohort
Prospective cohort, National Breast
Larsson (2004)
Navarro Silvera (2006)
Screening Study
Uppsala County of central Sweden and Va¨stmanland County
Prospective cohort, Iowa Women’s Health Study
Iowa, US
Location/ ethnicity
Kelemen (2004)
(1) Prospective studies
Study (first author, year)
1980–2000
1987–2003
1986–2001
Study period
264 Incident ovarian cancer cases were diagnosed among 48766 women
266 Incident cases of invasive epithelial ovarian cancer were diagnosed in 61084 women
147 Incident cases of epithelial ovarian cancer were identified in 27205 eligible postmenopausal women
Study characteristics
Dietary folate
Dietary folate
Dietary folate
Total folate
Folate intake categorized
0.75 [0.42–1.34]
[357
RR 95 % CI 1
lg/day \248
0.63 [0.38–1.05]
0.67 [0.43–1.04]
C204
301–\357
0.89 [0.62–1.29]
178–\204
0.81 [0.53–1.25]
0.92 [0.65–1.30]
155–\178
249–\301
RR 95 % CI
1.45 [0.83–2.53]
C347
1
1.45 [0.85–2.46]
290–346
lg/day
1.60 [0.98–2.61]
239–289
\155
RR 95 % CI
1.73 [0.90–3.33]
C541
1
1.24 [0.71–2.22]
331–540
lg/day
1.59 [0.97–2.61]
258–330
\238
RR 95 % CI 1
lg/day \257
Estimate effect (OR/ RR and 95 % CI)
Table 1 The main characteristics of the studies included for the association between folate intake and ovarian cancer risk
0.26
0.08
0.33
0.2
p trend
Multivariate models included age (time to event variable), packyears of smoking (never plus three levels), parity (parous vs. nulliparous), participation in vigorous physical activity (any vs. none), family history of breast cancer, menopausal status, hormone replacement therapy use (0, 1–2, [2 years), contraceptive use (ever vs. never), energy intake (continuous), vitamin C intake (quartiles), vitamin E intake (quartiles), b-carotene intake (quartiles), study center, and randomization group
Multivariable RRs were adjusted for age (5-year categories), body mass index (quartiles), educational level (less than high school, high school, university), family history of breast cancer (yes or no), parity (nulliparous, 1 or 2, 3 children), age at first birth (nulliparous, \25, 25–29, C30 years), oral contraceptive use (ever or never), age at menarche (\12, 13, C14 years), age at menopause (\50, 50–54, C55 years), postmenopausal hormone use (ever or never), and quartiles of alcohol consumption, and fruit and vegetable, lactose, and total energy intake
Adjusted for age. In addition to folate and alcohol, the model included age at menopause (four categories), physical activity (three categories), postmenopausal hormone use (never, former, current), oral contraceptive use (never, former, current), family history of breast cancer, family history of ovarian cancer, known diabetes at baseline, smoking (never, former, current), energy-adjusted intakes of total carotene, vitamin C and vitamin E (all quartiles)
Adjustment
6550 Mol Biol Rep (2013) 40:6547–6560
Nurses’ Health Study Cohort
Study types
Claudio Pelucchi (2005)
Hospital based case–control study
(2) Case–control studies
Tworoger (2006)
Study (first author, year)
Table 1 continued
Four Italian areas
US
Location/ ethnicity
1992–1999
1976–2002
Study period
1031 Cases and 2411 controls
481 Incident epithelial ovarian cancers were confirmed in 80254 Nurses’ Health Study participants
Study characteristics
Dietary folate
Dietary folate
Total folate
Folate intake categorized
OR 95 % CI 1.14 [0.83–1.56] 1.12 [0.80–1.57] 1.23 [0.85–1.80] 1.26 [0.80–1.97]
186–\231 231–\268 268–\316 C316
0.76 [0.52–1.12]
5
1
0.94 [0.67–1.31]
4
lg/day
0.84 [0.61–1.15]
3
\186
0.84 [0.62–1.13]
2
1.13 [0.83–1.53]
5 RR 95 % CI
0.84 [0.60–1.18]
4
1
0.85 [0.61–1.17]
3
Quintile
0.90 [0.66–1.22]
2
1
RR 95 % CI 1
Quintile 1
Estimate effect (OR/ RR and 95 % CI)
0.32
0.3
0.08
p trend
Estimates from unconditional logistic regression models adjusted for age, study center, year of interview, education, parity, body mass index, alcohol consumption, oral contraceptives use, physical activity, and nonalcohol energy intake. Including all the adjustments above, plus further terms for menopausal status, family history of breast and/or ovarian cancer, age at menarche, first birth and menopause, hormone replacement therapy use, diabetes, fruit and vegetable consumption, and smoking habit
Concurrent inclusion of dietary intakes of folate, methionine, and vitamin B 6, adjusting for duration of oral contraceptive use, age at first birth/parity, postmenopausal hormone use, tubal ligation, lactose intake, simply hysterectomy, and total caloric intake
Adjusted for dietary methionine intake, dietary vitamin B 6 intake, duration of oral contraceptive use, age at first birth/ parity, postmenopausal hormone use, tubal ligation, lactose intake, simply hysterectomy, and total caloric intake
Adjustment
Mol Biol Rep (2013) 40:6547–6560 6551
123
123
Australian Ovarian Cancer Study, population based case–control study
Population based New England case–control study
Harris (2011)
Study types
Webb (2011)
Study (first author, year)
Table 1 continued
Eastern Massachusetts and New Hampshire, US
Australia
Location/ ethnicity
1992–2008
2002–2005
Study period
1910 Cases and 1989 controls
1363 Cases and 1414 controls
Study characteristics
Total folate overall
Dietary folate overall
Folic acid (fortified food and supplements)
Total folate
Folate from foods
Folate intake categorized
OR 95 % CI 0.82 [0.68–0.99] 0.85 [0.70–1.02] 0.90 [0.75–1.08]
2 3 4
0.88 [0.74–1.06]
4
1
0.84 [0.69–1.01]
3
1
0.89 [0.74–1.06]
2
Quartile
OR 95 % CI
1.13 [0.91–1.40]
[204
1
1.13 [0.91–1.41]
96.0–204
Quartile
1.22 [0.98–1.51]
36.3–95.9
1
OR 95 % CI
1.05 [0.84–1.30]
[546
1
1.07 [0.86–1.33]
423–546
lg/day
1.14 [0.92–1.41]
334–422
\36.3
OR 95 % CI
1.00 [0.80–1.24]
[366
1
0.87 [0.70–1.08]
308–366
lg/day
0.94 [0.76–1.16]
252–307
\334
OR 95 % CI 1
lg/day \252
Estimate effect (OR/ RR and 95 % CI)
0.43
0.18
0.6
0.9
0.9
p trend
Tests for trend were performed using the median of the interval for each quartile and correspond to the MV adjusted analyses
Adjusted for age, study center, total energy, oral contraceptive use, parity, tubal ligation and family history of ovarian cancer.
ORs and 95 % CIs adjusted for age (years), level of education (school, diploma/certificate and university), oral contraceptive use (never, \5 and C5 years), parity (0, 1–2, C3), alcohol intake (quartiles) and total energy intake (kJ). Folate from foods is energy adjusted and excludes folic acid from fortified foods; folic acid includes folic acid from fortified foods (energy adjusted) and supplements. Total intake from foods (energy adjusted) and supplements
Adjustment
6552 Mol Biol Rep (2013) 40:6547–6560
Zhang (2012)
Study (first author, year)
Hospital based case– control study
Study types
Table 1 continued
China
Location/ ethnicity
2004–2007
Study period
215 Cases and 217 controls
Study characteristics
Dietary folate
Total folate 1998–2008
Dietary folate 1998–2008
Total folate 1992–1997
0.87 [0.69–1.09] OR 95 % CI 1 0.66 [0.41–1.05] 0.54 [0.32–0.94]
\200 200–310 [310
0.80 [0.63–1.02]
3 lg/day
0.86 [0.67–1.09]
4
1
2
1.02 [0.81–1.28]
4 1
0.96 [0.75–1.21]
3 OR 95 % CI
1.02 [0.80–1.29]
Quartile
1
2
1.12 [0.71–1.76]
4 1
0.98 [0.69–1.39]
3 OR 95 % CI
0.67 [0.48–0.94]
Quartile
1
2
0.57 [0.37–0.87]
4 1
0.65 [0.44–0.96]
3 OR 95 % CI
0.73 [0.53–1.00]
2
Quartile
OR 95 % CI 1
1
Dietary folate 1992–1997
Quartile
Estimate effect (OR/RR and 95 % CI)
Folate intake categorized
NA
0.38
0.94
0.7
0.002
p trend
Adjusted for age, tobacco smoking, alcohol consumption, number of deliveries, menopausal status, hormone replacement, oral contraceptive use, and ovarian cancer history
Tests for trend were performed using the median of the interval for each quartile
Enrollment phase was used as a proxy for exposure to folic acid grain supplementation. Participants in phase 2 (1998–2002) and phase 3 (2003–2008) were classified as post-supplementation and participants in phase 1 (1992–1997) were classified as pre-supplementation. Adjusted for age, study center, total energy, oral contraceptive use, parity, tubal ligation and family history of ovarian cancer.
Adjustment
Mol Biol Rep (2013) 40:6547–6560 6553
123
6554
Mol Biol Rep (2013) 40:6547–6560
Table 2 The main characteristics of the studies included for the association between MTHFR C677T polymorphism and ovarian cancer risk Study (first author, year)
Study type
Location
Sample size (case/ control)
Genotype distribution (case/control) CC
CT
TT
GershoniBaruch (2000)
Hospital based case–control study
Sheba and Rambam Medical Centres, Haifa, Israel/ Jewish
32/69
22/58
Jakubowska (2007)
Clinic based case–control study
Hereditary Cancer Registry at the Pomeranian Medical University in Szczecin, Poland
144/280
73/ 128
Pepe (2007)
Population based case–control study
Five centres of the Italian Consortium for Hereditary Breast and Ovarian Cancer (Aviano, Chieti, Milano, Padova and Pisa)
355/119
295/97
Terry (2010)
Population based New England-based case–control study (NEC)
Eastern Massachusetts or New Hampshire
1120/1160
427/ 499
492/ 488
140/ 138
Nested case–control Nurses’ Health Study (NHS)
11 States in the US
158/496
71/ 210
72/ 217
10/ 55
Clinic-based Mayo Clinic Ovarian Cancer Case Control Study (MAY)
Mayo Clinic in Rochester, MN
364/412
164/ 193
167/ 168
33/ 51
Webb (2011)
Population based case–control study
Australian Ovarian Cancer Study, Australia
1638/1278
744/ 571
709/ 568
185/ 139
Prasad (2011)
Population based case–control study
Andhra Pradesh, a southern state of India
80/125
72/ 116
3/8
5/1
Pawlik (2011)
Population based case–control study
Clinic of Gynecological Surgery, Poznan University of Medical Sciences, Poland
135/160
67/63
55/79
13/ 18
Zhang (2012)
Hospital based case–control study
Tianjin Medical University Cancer Institute and Hospital, China
215/218
102/ 115
94/92
19/ 11
Gao (2012)
Hospital based case–control study
Affiliated Hospital of Sichuan University, China
224/432
97/ 232
100/ 178
27/ 22
Jakubowska (2012)
–
Multicentres participating in CIMBA
1152/5059
479/ 2190
527/ 2259
146/ 610
evaluate the influence of individual study on the overall estimate. All the statistical analysis was conducted with the R software and the Meta package for R (www.r-project. org).
Results Summary of selected studies In PubMed and MEDLINE databases, we identified eight studies [12–14, 22, 25–28] evaluating the folate intake from dietary and/or total folate intake including dietary and supplemental resources and ovarian cancer risk. Among these, four were prospective cohort studies [12, 22, 25, 27] with a total of 217,309 participants and 1,158 incidence cases; and the other four were case–control studies with a total of 4,519 cases and 6,031 controls [13, 14, 26, 28]. For the cohort studies, none found a significant association between folate intake and ovarian cancer risk except for one by Larsson et al. [12], which reported a marginal
123
10/ 11 56/ 134
15/ 18 60/ 22
inverse association between higher dietary intake and ovarian cancer risk. For the four case–control studies, two studies [14, 28] reported a significant preventive effect of dietary folate intake against ovarian cancer, whereas others suggested no significant correlation between folate intake and ovarian cancer susceptibility. For the correlation between variant MTHFR C677T and ovarian cancer risk, we identified 10 eligible publications with 12 subgroup studies, Table 2 [13, 23, 24, 28–34]. Among these, five were population-based case–control studies [13, 24, 29, 30, 32], five were hospital/clinic-based case–control studies [23, 28, 31–33] and one was nested case–control study [32]. Jakubowska et al.’s study [34] is multi-centered and provided no information about the study type, and due to its overlapped individuals of BRCA1 and BRCA2 subgroups in the study, we only considered those from BRCA1 group in our analysis. A total number of 5,617 cases and 9,808 controls were included and seven of the selected studies revealed no correlation between the polymorphism and ovarian cancer risk [13, 24, 30, 32, 34], whereas two indicated the higher frequency of T allele in
Mol Biol Rep (2013) 40:6547–6560
6555
ovarian cancer individuals than in controls but the results were not statistically significant [23, 29]. And three reported a significant increased risk of ovarian cancer for the variant [28, 31, 33].
study conducted by the Kelemen et al. (RR = 0.85, 95 % CI = 0.73–0.99; p = 0.0403) which contributed to the heterogeneity between the studies with the sensitivity analysis. No publication bias was detected in our metaanalysis of the published studies.
Folate intake and ovarian cancer risk MTHFR polymorphism and ovarian cancer risk Based on our meta-analysis, we found no significant preventive effect of high total folate intake against ovarian cancer risk, with the overall RR of 1.04 and 95 % CI of 0.87–1.23 (p = 0.681) by comparing those of the highest quantile of total folate intake to those of the lowest. A slight significant heterogeneity among studies was found (Q = 4.77, df = 3, p = 0.189; I2 = 37.2 %; Table 3). A marginal inverse association without significance between dietary folate intake and ovarian cancer risk was found from overall studies (RR = 0.88, 95 % CI = 0.75–1.05; p = 0.158) under the random-effects model (Q = 12.2, df = 7, p = 0.094; I2 = 42.6 %; Fig. 2). In the stratification studies, by comparing the highest quantile dietary folate intake to the lowest quantile, we found 16 % (RR = 0.84, 95 % CI = 0.61–1.14; p = 0.265) and 8 % (OR = 0.92, 95 % CI = 0.74–1.14; p = 0.431) reduction of ovarian cancer risk for the prospective and case–control studies, respectively. Moreover, when the study of Kelemen et al., which contributed to the sensitivity of summary estimate, was excluded from the pooled prospective studies, we found high dietary folate intake was significantly associated with a reduced ovarian cancer risk (RR = 0.73, 95 % CI = 0.56–0.94; p = 0.0153) with no significant heterogeneity between the studies (Q = 0.19, df = 2, p = 0.9077; I2 = 0 %). In the subgroup analysis of studies performed in Northern America countries (three for USA and one for Canada), our results indicated a marginal protective effect of dietary folate intake on ovarian cancer risk but not for the total folate (dietary folate: RR = 0.88, 95 % CI = 0.76–1.03, p = 0.105; total folate: RR = 1.07, 95 % CI = 0.81–1.43, p = 0.621). Similarly, statistical association between high dietary folate intake and reduced ovarian cancer risk was found after the exclusion of the
Overall, we found no significant association between MTHFR variant C677T and ovarian cancer risk both using the dominant model (OR = 1.05, 95 % CI = 0.94–1.16; p = 0.379) and the recessive model (OR = 1.13, 95 % CI = 0.91–1.39; p = 0.264). Compared to the homozygote CC, neither the TT carriers nor the CT carriers showed a significant altered risk of ovarian cancer (Fig. 3). In the stratification study, we found a marginal association of allele T carriers with an increased ovarian cancer risk (CT vs. CC: OR = 1.15, 95 % CI = 0.99–1.32; p = 0.061; Table 4) in the Caucasian populations from America and significant association in Asian studies under three models (dominant: OR = 1.39, 95 % CI = 1.10–1.77, p = 0.007; recessive: OR = 2.39, 95 % CI = 1.58–3.62, p = 0.001; TT vs. CC: OR = 2.66, 95 % CI = 1.66–4.26, p = 0.001; Fig. 4). No significant association was found for the variant and ovarian cancer risk in any other stratification studies. We detected no publication bias in any study pool.
Discussion Folate deficiency may lead to neural tube defects of newborns, and folate has been reported to be preventive for cardiovascular disease and several types of cancer, therefore, it is recommended for all women of reproductive age in different countries [7]. Since January 1998, generalized folic acid fortification of flour and cereal grain products at a level of 140 lg/100 g was mandated in the US and Canada [35]. However, whether higher folate intake will benefit the women from the reduced ovarian cancer risk remains unknown. For the current meta-analysis, we found no
Table 3 Main results of effects of folate intake on ovarian cancer risk from overall and subgroup analysis Study types
Intake means
Fixed effects model OR/RR [95 % CI]
Random effects model p
OR/RR [95 % CI]
p
Heterogeneity 2
I (%)
p
Publication bias p
Prospective studies
Dietary
0.82 [0.65, 1.04]
0.100
0.84 [0.61, 1.14]
0.265
40.8
0.167
0.472
Case–control studies
Dietary
0.92 [0.81, 1.05]
0.212
0.92 [0.74, 1.14]
0.431
53.3
0.093
0.832
Overall studies
Dietary
0.90 [0.80, 1.00]
0.059
0.88 [0.75, 1.05]
0.158
42.6
0.094
0.715
American studies
Total
1.01 [0.89, 1.14]
0.914
1.04 [0.87, 1.23]
0.681
37.2
0.189
0.058
Dietary
0.88 [0.76, 1.03]
0.105
0.89 [0.72, 1.09]
0.261
23.8
0.268
0.789
Total
0.99 [0.85, 1.15]
0.862
1.07 [0.81, 1.43]
0.621
56.2
0.102
0.088
Data in bold was chosen for analysis after the consideration of heterogeneity between studies
123
6556
Mol Biol Rep (2013) 40:6547–6560
Fig. 2 Forest plot of the overall association between dietary folate intake and ovarian cancer risk
Fig. 3 Forest plots of the association between MTHFR polymorphism C677T and ovarian cancer risk from overall studies under a TT versus CC and b CT versus CC models
significant association between folate intake or the polymorphism of C677T on MTHFR and ovarian cancer risk. The meta-analysis for folate intake and ovarian cancer risk revealed no significant overall association between both high dietary and total folate intake and ovarian cancer risk. A marginal but not significant protective effect of high
123
dietary folate intake was observed rather than for the total folate intake which includes dietary and supplemental folate intake. We conducted the sensitivity analysis to test if certain study contributes most to the heterogeneity between the studies, discovering a significant inverse relationship between high dietary folate intake and reduced
Mol Biol Rep (2013) 40:6547–6560
6557
Table 4 Association between MTHFR polymorphism C677T and ovarian cancer risk from overall and subgroup studies Study types
Overall studies
American
Asian
Population based studies
Hospital/clinic based studies
Genetic model
Fixed effects model
Random effects model
Heterogeneity 2
Publication bias
OR [95 % CI]
p
OR [95 % CI]
p
I (%)
p
p
Dominant
1.05 [0.98, 1.13]
0.200
1.05 [0.94, 1.16]
0.379
34.9
0.129
0.899
Recessive
1.07 [0.96, 1.20]
0.211
1.13 [0.91, 1.39]
0.265
56.8
0.008
0.249
TT versus CC
1.11 [0.98, 1.25]
0.098
1.15 [0.90, 1.46]
0.264
62.1
0.005
0.476
CT versus CC
1.05 [0.97, 1.13]
0.227
1.04 [0.94, 1.16]
0.423
29.9
0.170
0.387
Dominant
1.07 [0.93, 1.22]
0.345
1.07 [0.93, 1.22]
0.345
0
0.556
0.351
Recessive
0.92 [0.74, 1.13]
0.413
0.81 [0.55, 1.20]
0.290
57
0.098
0.088
TT versus CC CT versus CC
1.00 [0.80, 1.25] 1.15 [0.99, 1.32]
0.991 0.061
0.86 [0.55, 1.35] 1.15 [0.99, 1.32]
0.508 0.061
64.1 0
0.062 0.686
0.059 0.413
Dominant
1.39 [1.10, 1.77]
0.007*
1.39 [1.10, 1.77]
0.007*
0
0.720
0.938
Recessive
2.39 [1.58, 3.62]
0.001*
2.39 [1.58, 3.62]
0.001*
0
0.618
0.341
TT versus CC
2.66 [1.66, 4.26]
0.001*
2.66 [1.66, 4.26]
0.001*
0
0.427
0.589
CT versus CC
1.23 [0.95, 1.58]
0.114
1.23 [0.95, 1.58]
0.114
0
0.493
0.238
Dominant
1.01 [0.91, 1.12]
0.887
0.99 [0.83, 1.18]
0.913
41.7
0.161
0.783
Recessive
1.04 [0.88, 1.22]
0.653
1.04 [0.88, 1.22]
0.671
3.2
0.388
0.485
TT versus CC
1.08 [0.91, 1.29]
0.380
1.08 [0.81, 1.44]
0.615
43.2
0.152
0.653
CT versus CC
1.02 [0.91, 1.14]
0.786
0.97 [0.78, 1.21]
0.795
56.2
0.077
0.417
Dominant
1.16 [0.98, 1.37]
0.091
1.15 [0.90, 1.46]
0.260
49.2
0.116
0.700
Recessive
1.39 [1.04, 1.85]
0.024*
1.61 [0.91, 2.84]
0.103
72
0.006
0.190
TT versus CC
1.40 [1.03, 1.91]
0.034*
1.55 [0.81, 2.99]
0.187
75.8
0.006
0.381
CT versus CC
1.12 [0.93, 1.33]
0.225
1.10 [0.87, 1.39]
0.421
40.5
0.169
0.330
Data in bold was chosen for analysis after the consideration of heterogeneity between studies * Significance
Fig. 4 Forest plots of the association between MTHFR polymorphism C677T and ovarian cancer risk from Asian studies under a dominant and b recessive models
ovarian cancer risk from the prospective studies and from the studies conducted in the US when the study by Kelemen et al. [22] was excluded from the meta-analysis
(prospective studies: RR = 0.73, 95 % CI = 0.56–0.94; Northern Americans: RR = 0.85, 95 % CI = 0.73–0.99), and after which no significant heterogeneity between
123
6558
studies was found. However, the pooled data showed that high total folate intake has no effect on ovarian cancer risk. It appears that dietary folate intake is more likely to act as a protective role than the supplemental folate intake on ovarian cancer. This may be attributed to the higher bioavailability of dietary folate than supplements of folate. Another reasonable explanation could be the co-existence of folate and other potential anticancer ingredients such as vitamin C, vitamin E and fiber in green vegetables, which may collectively contribute to the protective effect against ovarian cancer. Although we found no significant association between high total folate intake from supplemental and dietary resources and ovarian cancer risk, it should be noted that several studies have suggested that excessive folate intake from supplements may be harmful. Ulrich and Potter [7] proposed that despite the benefits of folate in neural tube defects prevention, some warning lights also existed, leading to the need for reevaluation of the potential risks of excessive folate intake. Most of the current studies that evaluated the association between folate intake and ovarian cancer risk were conducted in North America, where the folate fortification has in place for a long period of time. It was reported that the serum folate concentration for 40 % of the US populations has exceeded 40 nM after folic acid fortification, which was considered as supraphysiologic and excessive folate intake resulted in an increased unmetabolized folic acid in the circulation of healthy individuals, which could lead to the disturbance of folate metabolism homeostasis in the body [36]. These also may lead to the negative association of high total folate intake and ovarian cancer risk found here. Several studies reported an adverse effect for high folate intake. For instance, Kim [37] found that colorectal mucosa could be promoted rather than suppressed by exceptionally high supplemental folate intake after microscopic neoplastic foci are established. Miller et al. [38] also pointed out that folate deficiency rather than folate overdose inhibit the progression of mammary tumor, which may be due to the indispensable role of folate in the synthesis for DNA and RNA. However, whether high folate intake will affect the prognosis of ovarian cancer patients is still unclear. We speculate that not only the dose but also the time of folate intake should be taken into consideration in future studies to fully assess the correlations of folate intake and ovarian cancer risk. With a total of 5,617 cases and 9,808 controls from 12 subgroup studies, we found no significant correlation between the MTHFR SNP C677T and ovarian cancer risk, whereas a significant increased risk of ovarian cancer for the allele T carriers was detected in Asian populations, which may be due to the relative small sample size of all Asian population included (case:control = 551:844). However, the statistical power (p = 0.001) also reminded
123
Mol Biol Rep (2013) 40:6547–6560
us of the potential differences between Asian and other population. It is acceptable when considering the dietary structure and living habits of individuals form different regions. The consumption of alcohol varies greatly in Asian and American population, which may alter the involvement of the polymorphism in folate metabolism. The negative results of overall studies could be explained by possible alternative pathway besides MTHFR and the disparity of the populations as well as the relative small sample size of each study. It should be noted that there are several limitations in our current meta-analysis of published data. Firstly, the sample size is still relatively small both for the folate intake and the MTHFR SNP C677T and ovarian cancer risk. Secondly, the folate intake and folate metabolism pathways may be modulated by the alcohol consumption; many of the studies identified here have not taken the alcohol intake into account. Thirdly, the majority of the studies were conducted in Caucasian population, so there might be a bias for the entire population. Lastly, the diversity of population brought in difficulties in subgroup analysis. It is also debatable to pool all populations with different living habits, dietary structure and genetic backgrounds for an overall estimate. In summary, we found no significant association between the high folate intake or MTHFR SNP C677T and reduced ovarian cancer form overall population. Our results suggested that higher dietary folate intake may reduce the risk of ovarian cancer; and association between polymorphisms and cancer incidence probably should be considered under gene background as well as living habits, however, more studies are warranted to fully address the question. Acknowledgments This study was supported by grants from the Ministry of Science and Technology of China (2012BAK01B00 and 2011BAK10B00), the National Nature Science Foundation (81125020, 31101261 and 31200569), the Key Research Program (KSZD-EW-Z-021) of the Chinese Academy of Sciences, the Science and Technology Commission of Shanghai Municipality (12XD140 7000, 12431900500 and 10391902100), Director Foundation (20090 101) and the Food Safety Research Center and Key Laboratory of Food Safety Research of INS, SIBS, CAS. Peizhan Chen was partially supported by the SA-SIBS scholarship program. Conflict of interest
The authors declare no conflict of interest.
References 1. American Cancer Society (2011) Cancer facts and figures 2011. American Cancer Society, Atlanta 2. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D (2011) Global cancer statistics. CA Cancer J Clin 61(2):69–90. doi:10. 3322/caac.20107 3. Jemal A, Siegel R, Xu J, Ward E (2010) Cancer statistics, 2010. CA Cancer J Clin 60(5):277–300. doi:10.3322/caac.20073
Mol Biol Rep (2013) 40:6547–6560 4. Edmondson RJ, Monaghan JM (2001) The epidemiology of ovarian cancer. Int J Gynecol Cancer Off J Int Gynecol Cancer Soc 11(6):423–429 5. Kolahdooz F, van der Pols JC, Bain CJ, Marks GC, Hughes MC, Whiteman DC, Webb PM (2010) Meat, fish, and ovarian cancer risk: results from 2 Australian case–control studies, a systematic review, and meta-analysis. Am J Clin Nutr 91(6):1752–1763. doi:10.3945/ajcn.2009.28415 6. Kolahdooz F, Ibiebele TI, van der Pols JC, Webb PM (2009) Dietary patterns and ovarian cancer risk. Am J Clin Nutr 89(1):297–304. doi:10.3945/ajcn.2008.26575 7. Ulrich CM, Potter JD (2006) Folate supplementation: too much of a good thing? Cancer Epidemiol Biomarkers Prev Publ Am Assoc Cancer Res Cospons Am Soc Prev Oncol 15(2):189–193. doi:10.1158/1055-9965.EPI-152CO 8. Kim YI (1999) Folate and carcinogenesis: evidence, mechanisms, and implications. J Nutr Biochem 10(2):66–88 9. Choi SW, Mason JB (2000) Folate and carcinogenesis: an integrated scheme. J Nutr 130(2):129–132 10. Kim DH, Smith-Warner SA, Spiegelman D, Yaun SS, Colditz GA, Freudenheim JL, Giovannucci E, Goldbohm RA, Graham S, Harnack L, Jacobs EJ, Leitzmann M, Mannisto S, Miller AB, Potter JD, Rohan TE, Schatzkin A, Speizer FE, Stevens VL, Stolzenberg-Solomon R, Terry P, Toniolo P, Weijenberg MP, Willett WC, Wolk A, Zeleniuch-Jacquotte A, Hunter DJ (2010) Pooled analyses of 13 prospective cohort studies on folate intake and colon cancer. Cancer Causes Control 21(11):1919–1930. doi:10.1007/s10552-010-9620-8 11. Larsson SC, Giovannucci E, Wolk A (2007) Folate and risk of breast cancer: a meta-analysis. J Natl Cancer Inst 99(1):64–76. doi:10.1093/jnci/djk006 12. Larsson SC, Giovannucci E, Wolk A (2004) Dietary folate intake and incidence of ovarian cancer: the Swedish Mammography Cohort. J Natl Cancer Inst 96(5):396–402. doi:10.1093/jnci/ djh061 13. Webb PM, Ibiebele TI, Hughes MC, Beesley J, van der Pols JC, Chen X, Nagle CM, Bain CJ, Chenevix-Trench G, Australian Cancer Study, Australian Ovarian Cancer Study Group (2011) Folate and related micronutrients, folate-metabolising genes and risk of ovarian cancer. Eur J Clin Nutr 65(10):1133–1140 14. Harris HR, Cramer DW, Vitonis AF, Depari M, Terry KL (2011) Folate, vitamin B(6), vitamin B(12), methionine and alcohol intake in relation to ovarian cancer risk. Int J Cancer. doi:10.1002/ijc. 26455 15. Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, Boers GJ, den Heijer M, Kluijtmans LA, van den Heuvel LP et al (1995) A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 10(1):111–113. doi:10.1038/ng0595-111 16. Friso S, Choi SW, Girelli D, Mason JB, Dolnikowski GG, Bagley PJ, Olivieri O, Jacques PF, Rosenberg IH, Corrocher R, Selhub J (2002) A common mutation in the 5,10-methylenetetrahydrofolate reductase gene affects genomic DNA methylation through an interaction with folate status. Proc Natl Acad Sci USA 99(8):5606–5611. doi:10.1073/pnas.062066299 17. Molloy AM (2012) Genetic aspects of folate metabolism. Subcell Biochem 56:105–130. doi:10.1007/978-94-007-2199-9_7 18. Macis D, Maisonneuve P, Johansson H, Bonanni B, Botteri E, Iodice S, Santillo B, Penco S, Gucciardo G, D’Aiuto G, Rosselli Del Turco M, Amadori M, Costa A, Decensi A (2007) Methylenetetrahydrofolate reductase (MTHFR) and breast cancer risk: a nested-case–control study and a pooled meta-analysis. Breast Cancer Res Treat 106(2):263–271. doi:10.1007/s10549-0069491-6 19. Taioli E, Garza MA, Ahn YO, Bishop DT, Bost J, Budai B, Chen K, Gemignani F, Keku T, Lima CS, Le Marchand L, Matsuo K,
6559
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
Moreno V, Plaschke J, Pufulete M, Thomas SB, Toffoli G, Wolf CR, Moore CG, Little J (2009) Meta- and pooled analyses of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism and colorectal cancer: a HuGE-GSEC review. Am J Epidemiol 170(10):1207–1221. doi:10.1093/aje/kwp275 Cui LH, Shin MH, Kim HN, Song HR, Piao JM, Kweon SS, Choi JS, Yun WJ, Kim YC, Oh IJ, Kim KS (2011) Methylenetetrahydrofolate reductase C677T polymorphism in patients with lung cancer in a Korean population. BMC Med Genet 12:28. doi:10.1186/1471-2350-12-28 Zhang J, Qiu LX, Wang ZH, Wu XH, Liu XJ, Wang BY, Hu XC (2010) MTHFR C677T polymorphism associated with breast cancer susceptibility: a meta-analysis involving 15,260 cases and 20,411 controls. Breast Cancer Res Treat 123(2):549–555. doi:10.1007/s10549-010-0783-5 Kelemen LE, Sellers TA, Vierkant RA, Harnack L, Cerhan JR (2004) Association of folate and alcohol with risk of ovarian cancer in a prospective study of postmenopausal women. Cancer Causes Control 15(10):1085–1093 Gershoni-Baruch R, Dagan E, Israeli D, Kasinetz L, Kadouri E, Friedman E (2000) Association of the C677T polymorphism in the MTHFR gene with breast and/or ovarian cancer risk in Jewish women. Eur J Cancer 36(18):2313–2316 Pepe C, Guidugli L, Sensi E, Aretini P, D’Andrea E, Montagna M, Manoukian S, Ottini L, Radice P, Viel A, Bevilacqua G, Caligo MA (2007) Methyl group metabolism gene polymorphisms as modifier of breast cancer risk in Italian BRCA1/2 carriers. Breast Cancer Res Treat 103(1):29–36. doi:10.1007/ s10549-006-9349-y Navarro Silvera SA, Jain M, Howe GR, Miller AB, Rohan TE (2006) Dietary folate consumption and risk of ovarian cancer: a prospective cohort study. Eur J Cancer Prev 15(6):511–515. doi:10.1097/01.cej.0000220627.54986.bf Pelucchi C, Mereghetti M, Talamini R, Negri E, Montella M, Ramazzotti V, Franceschi S, La Vecchia C (2005) Dietary folate, alcohol consumption, and risk of ovarian cancer in an Italian case–control study. Cancer Epidemiol Biomarkers Prev Publ Am Assoc Cancer Res Cospons Am Soc Prev Oncol 14(8):2056–2058. doi:10.1158/1055-9965.EPI-05-0192 Tworoger SS, Hecht JL, Giovannucci E, Hankinson SE (2006) Intake of folate and related nutrients in relation to risk of epithelial ovarian cancer. Am J Epidemiol 163(12):1101–1111. doi:10.1093/aje/kwj128 Zhang L, Liu W, Hao Q, Bao L, Wang K (2012) Folate intake and methylenetetrahydrofolate reductase gene polymorphisms as predictive and prognostic biomarkers for ovarian cancer risk. Int J Mol Sci 13(4):4009–4020. doi:10.3390/ijms13044009 Prasad VV, Wilkhoo H (2011) Association of the functional polymorphism C677T in the methylenetetrahydrofolate reductase gene with colorectal, thyroid, breast, ovarian, and cervical cancers. Onkologie 34(8–9):422–426. doi:10.1159/000331131 Pawlik P, Mostowska A, Lianeri M, Sajdak S, Kedzia H, Jagodzinski PP (2011) Folate and choline metabolism gene variants in relation to ovarian cancer risk in the Polish population. Mol Biol Rep. doi:10.1007/s11033-011-1359-0 Jakubowska A, Gronwald J, Menkiszak J, Gorski B, Huzarski T, Byrski T, Edler L, Lubinski J, Scott RJ, Hamann U (2007) Methylenetetrahydrofolate reductase polymorphisms modify BRCA1-associated breast and ovarian cancer risks. Breast Cancer Res Treat 104(3):299–308. doi:10.1007/s10549-006-9417-3 Terry KL, Tworoger SS, Goode EL, Gates MA, Titus-Ernstoff L, Kelemen LE, Sellers TA, Hankinson SE, Cramer DW (2010) MTHFR polymorphisms in relation to ovarian cancer risk. Gynecol Oncol 119(2):319–324. doi:10.1016/j.ygyno.2010.08.007 Gao S, Liu N, Ma Y, Ying L (2012) Methylenetetrahydrofolate reductase gene polymorphisms as predictive and prognostic
123
6560 biomarkers in ovarian cancer risk. Asian Pac J Cancer Prev 13(2):569–573 34. Jakubowska A, Rozkrut D, Antoniou A, Hamann U, Scott RJ, McGuffog L, Healy S, Sinilnikova OM, Rennert G, Lejbkowicz F, Flugelman A, Andrulis IL, Glendon G, Ozcelik H, Thomassen M, Paligo M, Aretini P, Kantala J, Aroer B, von Wachenfeldt A, Liljegren A, Loman N, Herbst K, Kristoffersson U, Rosenquist R, Karlsson P, Stenmark-Askmalm M, Melin B, Nathanson KL, Domchek SM, Byrski T, Huzarski T, Gronwald J, Menkiszak J, Cybulski C, Serrano P, Osorio A, Cajal TR, Tsitlaidou M, Benitez J, Gilbert M, Rookus M, Aalfs CM, Kluijt I, BoessenkoolPape JL, Meijers-Heijboer HE, Oosterwijk JC, van Asperen CJ, Blok MJ, Nelen MR, van den Ouweland AM, Seynaeve C, van der Luijt RB, Devilee P, Easton DF, Peock S, Frost D, Platte R, Ellis SD, Fineberg E, Evans DG, Lalloo F, Eeles R, Jacobs C, Adlard J, Davidson R, Eccles D, Cole T, Cook J, Godwin A, Bove B, Stoppa-Lyonnet D, Caux-Moncoutier V, Belotti M, Tirapo C, Mazoyer S, Barjhoux L, Boutry-Kryza N, Pujol P, Coupier I, Peyrat JP, Vennin P, Muller D, Fricker JP, VenatBouvet L, Johannsson OT, Isaacs C, Schmutzler R, Wappenschmidt B, Meindl A, Arnold N, Varon-Mateeva R,
123
Mol Biol Rep (2013) 40:6547–6560
35.
36. 37.
38.
Niederacher D, Sutter C, Deissler H, Preisler-Adams S, Simard J, Soucy P, Durocher F, Chenevix-Trench G, Beesley J, Chen X, Rebbeck T, Couch F, Wang X, Lindor N, Fredericksen Z, Pankratz VS, Peterlongo P, Bonanni B, Fortuzzi S, Peissel B, Szabo C, Mai PL, Loud JT, Lubinski J (2012) Association of PHB 1630 C[T and MTHFR 677 C[T polymorphisms with breast and ovarian cancer risk in BRCA1/2 mutation carriers: results from a multicenter study. Br J Cancer 106(12):2016–2024. doi:10.1038/bjc.2012.160 Honein MA, Paulozzi LJ, Mathews TJ, Erickson JD, Wong LY (2001) Impact of folic acid fortification of the US food supply on the occurrence of neural tube defects. JAMA 285(23):2981–2986 Ulrich CM (2007) Folate and cancer prevention: a closer look at a complex picture. Am J Clin Nutr 86(2):271–273 Kim YI (2004) Folate, colorectal carcinogenesis, and DNA methylation: lessons from animal studies. Environ Mol Mutagen 44(1):10–25. doi:10.1002/em.20025 Miller JW, Borowsky AD, Marple TC, McGoldrick ET, DillardTelm L, Young LJ, Green R (2008) Folate, DNA methylation, and mouse models of breast tumorigenesis. Nutr Rev 66(Suppl 1):S59–S64. doi:10.1111/j.1753-4887.2008.00070.x
Folate intake and MTHFR polymorphism C677T is not associated with ovarian cancer risk: evidence from the meta-analysis Chenglin Li • Peizhan Chen • Pingting Hu • Mian Li • Xiaoguang Li • He Guo • Jingquan Li Ruiai Chu • Wei Zhang • Hui Wang
•
Received: 4 January 2013 / Accepted: 14 September 2013 / Published online: 17 October 2013 Ó Springer Science+Business Media Dordrecht 2013
Abstract Folate is essential for DNA synthesis and methylation and implicated in the process of carcinogenesis. Several studies inconclusively suggested increased folate intake may reduce ovarian cancer risk. Studies concerning the association between C677T polymorphism in methylenetetrahydrofolate reductase (MTHFR), an important enzyme in folate metabolism, and ovarian cancer risk also resulted in no agreement. The meta-analysis was conducted based on current studies to assess the association between folate intake, the MTHFR C667T polymorphism and ovarian cancer risk. 1,158 cases out of 217,309 participants from four cohort studies, 4,519 cases and 6,031 controls from four case–control studies about folate intake along with 5,617 cases and 9,808 controls from 10 publications concerning the polymorphism were pooled, respectively. We detected no significant association between total folate (RR = 1.04, 95 % confidence interval (CI) = 0.87–1.23) or dietary folate (RR = 0.88, 95 % CI = 0.75–1.05) intake and ovarian cancer risk, and also no significant relationship was found between MTHFR C677T
Chenglin Li and Peizhan Chen have contributed equally to this work. C. Li P. Chen P. Hu M. Li X. Li H. Guo J. Li R. Chu H. Wang (&) Key Laboratory of Food Safety Research, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Shanghai 200031, China e-mail: [email protected] W. Zhang Institute for Food Safety and Health, Illinois Institute of Technology, Bedford Park, IL, USA H. Wang Key Laboratory of Food Safety Risk Assessment, Ministry of Health, Beijing 100021, China
polymorphism and ovarian cancer risk (TT vs. CC: odds ratio (OR) = 1.15, 95 % CI = 0.90–1.46; CT vs. CC: OR = 1.04, 95 % CI = 0.94–1.16). Our analysis indicated neither folate intake nor MTHFR C677T polymorphism is related to altered susceptibility of ovarian cancer. Keywords Folate MTHFR Ovarian cancer Meta-analysis Abbreviations MTHFR Methylenetetrahydrofolate reductase SNP Single nucleotide polymorphism OR Odds ratio 95 % CI 95 % Confidence interval
Introduction Ovarian cancer is the most malignant form of cancer in the female reproductive system, which accounts for about 3 % of all cancers in women. It has been estimated that 21,990 new cases and 15,460 deaths would occur in the US in 2011 [1]. The American Cancer Society reported the overall 1-, 5-, and 10-year survival rate of ovarian cancer patients is approximately 75, 46 and 38 %, respectively [1– 3]. Factors that have been reported to be associated with ovarian cancer include age, obesity, oral contraceptive use, parity, use of fertility drugs and family history of cancer [4]. However, such factors only account for a small proportion of ovarian cancer patients; more comprehensive studies are warranted to fully illustrate the etiology of ovarian cancer. Recently, dietary factors have been found to be associated with the susceptibility of ovarian cancer. High meat and fat intake may increase the risk of epithelial ovarian cancer, while high vitamin A and beta-carotene
123
6548
intake may modestly protect against ovarian cancer. Due to the complexity of the dietary factors, results of previously studies concerning the dietary risk factor were typically inconclusive [4–6]. Folate is a common nutrient in daily vegetables which has been found to be involved in the prevention of cardiovascular disease and neural tube defects [7, 8]. Folate is also involved in the formation of the S-adenosylmethionine, a universal methyl donor essential for the synthesis of purine and thymidine. Considering its important roles in DNA methylation and nucleic acid synthesis, some studies were conducted to evaluate its effects on carcinogenesis [9]. For instance, Kim et al. [10] found a significant protective effect for both dietary (RR = 0.92, 95 % confidence intervals [CI] = 0.84– 1.00) and total folate intake (RR = 0.85, 95 % CI = 0.77–0.95) against colorectal cancer with a meta-analysis pooled of 13 prospective studies. In contrast, Larsson et al. [11] found no relation between folate intake and breast cancer risk. Other recent studies also evaluated the association between ovarian cancer and folate intake with inconsistent results. For example, Larsson et al. [12] reported that higher dietary folate intake may reduce ovarian cancer risk, particularly for those of high alcohol consumption; however, studies by Webb et al. [13] and Harris et al. [14] found little evidence for the protective effects of folate intake against ovarian cancer. Methylenetetrahydrofolate reductase (MTHFR) is a major enzyme involved in folate-related one carbon metabolism, which is responsible for the conversion of 5,10-methylenetetrahydrofolate (5,10-methylene-THF) to 5-methyltetrahydrofolate (5-methyl-THF), the predominant form of folate in plasma. A functional polymorphism (C677T) located on the exon 4 of the MTHFR gene results in the amino acid transition from Ala to Val, further leading to the reduction of enzyme activity [15]. Individuals with homozygote 677TT genotype only have 30 % enzyme activity compared to 677CC carriers. The reduced activity of MTHFR would lead to the accumulation of 5,10-methylene-THF and reduction of 5-methyl-THF level in the cells, which may interrupt the homeostasis of folate metabolism in vivo [16–18]. The C677T polymorphism has been reported to be significantly associated with many cancer types, such as colorectal cancer, breast cancer and lung cancer [19–21]. The association of the variant with ovarian cancer risk was investigated by several studies with no consistent conclusions [22–24]. Due to fact that the folate intake and related signaling genes may play important roles in the carcinogenesis and the existing epidemiological studies do not provide comprehensive analysis on available data, here we conducted the meta-analysis to systematically evaluate the association of folate intake and the MTHFR variant C677T with ovarian cancer susceptibility.
123
Mol Biol Rep (2013) 40:6547–6560
Methods Literature search A comprehensive literature search was conducted on MEDLINE and PubMed databases by two independent researchers to identify relevant studies published prior to 31 December 2012. The search flowchart was shown in Fig. 1. We used the term ‘‘ovarian cancer’’, along with ‘‘folate’’, ‘‘folic acid’’ or ‘‘MTHFR’’ for the search without any other restrictions. The reference lists of the retrieved publications were checked to identify any missing studies in the database search. We selected eligible studies that examined the correlation between folate intake and ovarian cancer risk with the folate intake data in quantiles together with the risk estimates (RR or odds ratio [OR]) and 95 % CIs, or with the results of the highest quantile in contrast with the lowest quantile, or with data that could be used to calculate the risk estimate and its 95 % CI. Studies that examined the association of MTHFR single nucleotide polymorphism (SNP) C677T and ovarian cancer risk should provide sufficiency data for the frequency of the genotypes. We included case–control, cohort or crosssectional studies that were reported in English. Selected publications and data extraction Based on our literature search, we identified eight published studies that evaluated the association between folate intake and ovarian cancer risk, Table 1. The following items from the selected publications were extracted: the first author’s name and year of the publication, study type, location and ethnicity of participants, study period and main characteristics of the study, folate intake categories and their corresponding estimate effects (OR/RR and 95 % CI), the p trend values and related statistical adjustments. For articles related to MTHFR polymorphism and ovarian cancer risk, we selected 10 publications with 12 subgroup studies that investigated the association of C677T variant and ovarian cancer susceptibility, Table 2. The following information of the studies was extracted: the name of first author and the year of publication, study type and location, sample size and genotype distribution in case and control groups. For the 12 subgroup studies, two studies only provided the number of individuals with homozygote TT genotype and the total sample size [23, 24], so the related information was only used in the recessive genetic model rather than other models. Statistical analysis The standard inverse variance weighting method was used to calculate the pooled estimate and its 95 % CI that
Mol Biol Rep (2013) 40:6547–6560 Fig. 1 Work flow chart of the meta-analysis
6549
Publications identified from literature
Publications identified from literature
search using term “folate” or “folic
search using term “MTHFR” in
acid” in combination with “ovarian
combination with “ovarian cancer”
cancer” from PubMed (N=373)and
from PubMed (N=24) and MEDLINE
MEDLINE (N=178) up to Dec. 2012
(N=19) up to Dec. 2012
Exclude overlapped publications Excluded (N=19)
Excluded (N=178)
Check for missing studies from reference lists in publications Added (N=0)
Added (N=0)
Further detailed screening Excluded
Excluded
Not publication in English (N=18)
Not publication in English (N=2)
Not population studies (N=308)
Not population studies (N=10)
Not ovarian cancer (N=8)
Not ovarian cancer (N=1)
Not folate intake(N=31)
Not MTHFR SNP C677T (N=1)
Final data check
8 relevant publications included in meta-analysis
concerning the effect of folate intake and ovarian cancer risk. Due to the relative negligible difference between ORs from case–control studies and RRs from prospective studies numerically, the ORs were assumed as the estimates of RRs for further analysis. Each study was assigned with appropriate weighting using the inverse square of SE for each logRR. To determine the association between ovarian cancer risk and variant MTHFR SNP C677T, the ORs and corresponding 95 % CIs were used to calculate the pooled estimate and 95 % CI. The standard inverse variance weighting method was used to calculate the pooled OR and its 95 % CI under the fixed-effects model and the
10 relevant publications included in meta-analysis
DerSimonian–Laird method were used to calculate the pooled estimate under the random-effects model. Statistical heterogeneity among studies was evaluated by Q test and I2 statistics. We considered significant heterogeneity existing when p value was less than 0.05 for the Q test or I2 [25 % in I2 statistics. Results from randomeffects model were acceptable for interpretation when significant heterogeneity among studies was detected; otherwise, the results from the fixed-effects model were used. Publication bias was assessed by the Egger’s linear regression test and a p value less than 0.05 indicated significant publication bias. For the sensitivity test, we excluded a single study from the meta-analysis each time to
123
123
Study types
General Canadian population
Swedish Mammography Cohort
Prospective cohort, National Breast
Larsson (2004)
Navarro Silvera (2006)
Screening Study
Uppsala County of central Sweden and Va¨stmanland County
Prospective cohort, Iowa Women’s Health Study
Iowa, US
Location/ ethnicity
Kelemen (2004)
(1) Prospective studies
Study (first author, year)
1980–2000
1987–2003
1986–2001
Study period
264 Incident ovarian cancer cases were diagnosed among 48766 women
266 Incident cases of invasive epithelial ovarian cancer were diagnosed in 61084 women
147 Incident cases of epithelial ovarian cancer were identified in 27205 eligible postmenopausal women
Study characteristics
Dietary folate
Dietary folate
Dietary folate
Total folate
Folate intake categorized
0.75 [0.42–1.34]
[357
RR 95 % CI 1
lg/day \248
0.63 [0.38–1.05]
0.67 [0.43–1.04]
C204
301–\357
0.89 [0.62–1.29]
178–\204
0.81 [0.53–1.25]
0.92 [0.65–1.30]
155–\178
249–\301
RR 95 % CI
1.45 [0.83–2.53]
C347
1
1.45 [0.85–2.46]
290–346
lg/day
1.60 [0.98–2.61]
239–289
\155
RR 95 % CI
1.73 [0.90–3.33]
C541
1
1.24 [0.71–2.22]
331–540
lg/day
1.59 [0.97–2.61]
258–330
\238
RR 95 % CI 1
lg/day \257
Estimate effect (OR/ RR and 95 % CI)
Table 1 The main characteristics of the studies included for the association between folate intake and ovarian cancer risk
0.26
0.08
0.33
0.2
p trend
Multivariate models included age (time to event variable), packyears of smoking (never plus three levels), parity (parous vs. nulliparous), participation in vigorous physical activity (any vs. none), family history of breast cancer, menopausal status, hormone replacement therapy use (0, 1–2, [2 years), contraceptive use (ever vs. never), energy intake (continuous), vitamin C intake (quartiles), vitamin E intake (quartiles), b-carotene intake (quartiles), study center, and randomization group
Multivariable RRs were adjusted for age (5-year categories), body mass index (quartiles), educational level (less than high school, high school, university), family history of breast cancer (yes or no), parity (nulliparous, 1 or 2, 3 children), age at first birth (nulliparous, \25, 25–29, C30 years), oral contraceptive use (ever or never), age at menarche (\12, 13, C14 years), age at menopause (\50, 50–54, C55 years), postmenopausal hormone use (ever or never), and quartiles of alcohol consumption, and fruit and vegetable, lactose, and total energy intake
Adjusted for age. In addition to folate and alcohol, the model included age at menopause (four categories), physical activity (three categories), postmenopausal hormone use (never, former, current), oral contraceptive use (never, former, current), family history of breast cancer, family history of ovarian cancer, known diabetes at baseline, smoking (never, former, current), energy-adjusted intakes of total carotene, vitamin C and vitamin E (all quartiles)
Adjustment
6550 Mol Biol Rep (2013) 40:6547–6560
Nurses’ Health Study Cohort
Study types
Claudio Pelucchi (2005)
Hospital based case–control study
(2) Case–control studies
Tworoger (2006)
Study (first author, year)
Table 1 continued
Four Italian areas
US
Location/ ethnicity
1992–1999
1976–2002
Study period
1031 Cases and 2411 controls
481 Incident epithelial ovarian cancers were confirmed in 80254 Nurses’ Health Study participants
Study characteristics
Dietary folate
Dietary folate
Total folate
Folate intake categorized
OR 95 % CI 1.14 [0.83–1.56] 1.12 [0.80–1.57] 1.23 [0.85–1.80] 1.26 [0.80–1.97]
186–\231 231–\268 268–\316 C316
0.76 [0.52–1.12]
5
1
0.94 [0.67–1.31]
4
lg/day
0.84 [0.61–1.15]
3
\186
0.84 [0.62–1.13]
2
1.13 [0.83–1.53]
5 RR 95 % CI
0.84 [0.60–1.18]
4
1
0.85 [0.61–1.17]
3
Quintile
0.90 [0.66–1.22]
2
1
RR 95 % CI 1
Quintile 1
Estimate effect (OR/ RR and 95 % CI)
0.32
0.3
0.08
p trend
Estimates from unconditional logistic regression models adjusted for age, study center, year of interview, education, parity, body mass index, alcohol consumption, oral contraceptives use, physical activity, and nonalcohol energy intake. Including all the adjustments above, plus further terms for menopausal status, family history of breast and/or ovarian cancer, age at menarche, first birth and menopause, hormone replacement therapy use, diabetes, fruit and vegetable consumption, and smoking habit
Concurrent inclusion of dietary intakes of folate, methionine, and vitamin B 6, adjusting for duration of oral contraceptive use, age at first birth/parity, postmenopausal hormone use, tubal ligation, lactose intake, simply hysterectomy, and total caloric intake
Adjusted for dietary methionine intake, dietary vitamin B 6 intake, duration of oral contraceptive use, age at first birth/ parity, postmenopausal hormone use, tubal ligation, lactose intake, simply hysterectomy, and total caloric intake
Adjustment
Mol Biol Rep (2013) 40:6547–6560 6551
123
123
Australian Ovarian Cancer Study, population based case–control study
Population based New England case–control study
Harris (2011)
Study types
Webb (2011)
Study (first author, year)
Table 1 continued
Eastern Massachusetts and New Hampshire, US
Australia
Location/ ethnicity
1992–2008
2002–2005
Study period
1910 Cases and 1989 controls
1363 Cases and 1414 controls
Study characteristics
Total folate overall
Dietary folate overall
Folic acid (fortified food and supplements)
Total folate
Folate from foods
Folate intake categorized
OR 95 % CI 0.82 [0.68–0.99] 0.85 [0.70–1.02] 0.90 [0.75–1.08]
2 3 4
0.88 [0.74–1.06]
4
1
0.84 [0.69–1.01]
3
1
0.89 [0.74–1.06]
2
Quartile
OR 95 % CI
1.13 [0.91–1.40]
[204
1
1.13 [0.91–1.41]
96.0–204
Quartile
1.22 [0.98–1.51]
36.3–95.9
1
OR 95 % CI
1.05 [0.84–1.30]
[546
1
1.07 [0.86–1.33]
423–546
lg/day
1.14 [0.92–1.41]
334–422
\36.3
OR 95 % CI
1.00 [0.80–1.24]
[366
1
0.87 [0.70–1.08]
308–366
lg/day
0.94 [0.76–1.16]
252–307
\334
OR 95 % CI 1
lg/day \252
Estimate effect (OR/ RR and 95 % CI)
0.43
0.18
0.6
0.9
0.9
p trend
Tests for trend were performed using the median of the interval for each quartile and correspond to the MV adjusted analyses
Adjusted for age, study center, total energy, oral contraceptive use, parity, tubal ligation and family history of ovarian cancer.
ORs and 95 % CIs adjusted for age (years), level of education (school, diploma/certificate and university), oral contraceptive use (never, \5 and C5 years), parity (0, 1–2, C3), alcohol intake (quartiles) and total energy intake (kJ). Folate from foods is energy adjusted and excludes folic acid from fortified foods; folic acid includes folic acid from fortified foods (energy adjusted) and supplements. Total intake from foods (energy adjusted) and supplements
Adjustment
6552 Mol Biol Rep (2013) 40:6547–6560
Zhang (2012)
Study (first author, year)
Hospital based case– control study
Study types
Table 1 continued
China
Location/ ethnicity
2004–2007
Study period
215 Cases and 217 controls
Study characteristics
Dietary folate
Total folate 1998–2008
Dietary folate 1998–2008
Total folate 1992–1997
0.87 [0.69–1.09] OR 95 % CI 1 0.66 [0.41–1.05] 0.54 [0.32–0.94]
\200 200–310 [310
0.80 [0.63–1.02]
3 lg/day
0.86 [0.67–1.09]
4
1
2
1.02 [0.81–1.28]
4 1
0.96 [0.75–1.21]
3 OR 95 % CI
1.02 [0.80–1.29]
Quartile
1
2
1.12 [0.71–1.76]
4 1
0.98 [0.69–1.39]
3 OR 95 % CI
0.67 [0.48–0.94]
Quartile
1
2
0.57 [0.37–0.87]
4 1
0.65 [0.44–0.96]
3 OR 95 % CI
0.73 [0.53–1.00]
2
Quartile
OR 95 % CI 1
1
Dietary folate 1992–1997
Quartile
Estimate effect (OR/RR and 95 % CI)
Folate intake categorized
NA
0.38
0.94
0.7
0.002
p trend
Adjusted for age, tobacco smoking, alcohol consumption, number of deliveries, menopausal status, hormone replacement, oral contraceptive use, and ovarian cancer history
Tests for trend were performed using the median of the interval for each quartile
Enrollment phase was used as a proxy for exposure to folic acid grain supplementation. Participants in phase 2 (1998–2002) and phase 3 (2003–2008) were classified as post-supplementation and participants in phase 1 (1992–1997) were classified as pre-supplementation. Adjusted for age, study center, total energy, oral contraceptive use, parity, tubal ligation and family history of ovarian cancer.
Adjustment
Mol Biol Rep (2013) 40:6547–6560 6553
123
6554
Mol Biol Rep (2013) 40:6547–6560
Table 2 The main characteristics of the studies included for the association between MTHFR C677T polymorphism and ovarian cancer risk Study (first author, year)
Study type
Location
Sample size (case/ control)
Genotype distribution (case/control) CC
CT
TT
GershoniBaruch (2000)
Hospital based case–control study
Sheba and Rambam Medical Centres, Haifa, Israel/ Jewish
32/69
22/58
Jakubowska (2007)
Clinic based case–control study
Hereditary Cancer Registry at the Pomeranian Medical University in Szczecin, Poland
144/280
73/ 128
Pepe (2007)
Population based case–control study
Five centres of the Italian Consortium for Hereditary Breast and Ovarian Cancer (Aviano, Chieti, Milano, Padova and Pisa)
355/119
295/97
Terry (2010)
Population based New England-based case–control study (NEC)
Eastern Massachusetts or New Hampshire
1120/1160
427/ 499
492/ 488
140/ 138
Nested case–control Nurses’ Health Study (NHS)
11 States in the US
158/496
71/ 210
72/ 217
10/ 55
Clinic-based Mayo Clinic Ovarian Cancer Case Control Study (MAY)
Mayo Clinic in Rochester, MN
364/412
164/ 193
167/ 168
33/ 51
Webb (2011)
Population based case–control study
Australian Ovarian Cancer Study, Australia
1638/1278
744/ 571
709/ 568
185/ 139
Prasad (2011)
Population based case–control study
Andhra Pradesh, a southern state of India
80/125
72/ 116
3/8
5/1
Pawlik (2011)
Population based case–control study
Clinic of Gynecological Surgery, Poznan University of Medical Sciences, Poland
135/160
67/63
55/79
13/ 18
Zhang (2012)
Hospital based case–control study
Tianjin Medical University Cancer Institute and Hospital, China
215/218
102/ 115
94/92
19/ 11
Gao (2012)
Hospital based case–control study
Affiliated Hospital of Sichuan University, China
224/432
97/ 232
100/ 178
27/ 22
Jakubowska (2012)
–
Multicentres participating in CIMBA
1152/5059
479/ 2190
527/ 2259
146/ 610
evaluate the influence of individual study on the overall estimate. All the statistical analysis was conducted with the R software and the Meta package for R (www.r-project. org).
Results Summary of selected studies In PubMed and MEDLINE databases, we identified eight studies [12–14, 22, 25–28] evaluating the folate intake from dietary and/or total folate intake including dietary and supplemental resources and ovarian cancer risk. Among these, four were prospective cohort studies [12, 22, 25, 27] with a total of 217,309 participants and 1,158 incidence cases; and the other four were case–control studies with a total of 4,519 cases and 6,031 controls [13, 14, 26, 28]. For the cohort studies, none found a significant association between folate intake and ovarian cancer risk except for one by Larsson et al. [12], which reported a marginal
123
10/ 11 56/ 134
15/ 18 60/ 22
inverse association between higher dietary intake and ovarian cancer risk. For the four case–control studies, two studies [14, 28] reported a significant preventive effect of dietary folate intake against ovarian cancer, whereas others suggested no significant correlation between folate intake and ovarian cancer susceptibility. For the correlation between variant MTHFR C677T and ovarian cancer risk, we identified 10 eligible publications with 12 subgroup studies, Table 2 [13, 23, 24, 28–34]. Among these, five were population-based case–control studies [13, 24, 29, 30, 32], five were hospital/clinic-based case–control studies [23, 28, 31–33] and one was nested case–control study [32]. Jakubowska et al.’s study [34] is multi-centered and provided no information about the study type, and due to its overlapped individuals of BRCA1 and BRCA2 subgroups in the study, we only considered those from BRCA1 group in our analysis. A total number of 5,617 cases and 9,808 controls were included and seven of the selected studies revealed no correlation between the polymorphism and ovarian cancer risk [13, 24, 30, 32, 34], whereas two indicated the higher frequency of T allele in
Mol Biol Rep (2013) 40:6547–6560
6555
ovarian cancer individuals than in controls but the results were not statistically significant [23, 29]. And three reported a significant increased risk of ovarian cancer for the variant [28, 31, 33].
study conducted by the Kelemen et al. (RR = 0.85, 95 % CI = 0.73–0.99; p = 0.0403) which contributed to the heterogeneity between the studies with the sensitivity analysis. No publication bias was detected in our metaanalysis of the published studies.
Folate intake and ovarian cancer risk MTHFR polymorphism and ovarian cancer risk Based on our meta-analysis, we found no significant preventive effect of high total folate intake against ovarian cancer risk, with the overall RR of 1.04 and 95 % CI of 0.87–1.23 (p = 0.681) by comparing those of the highest quantile of total folate intake to those of the lowest. A slight significant heterogeneity among studies was found (Q = 4.77, df = 3, p = 0.189; I2 = 37.2 %; Table 3). A marginal inverse association without significance between dietary folate intake and ovarian cancer risk was found from overall studies (RR = 0.88, 95 % CI = 0.75–1.05; p = 0.158) under the random-effects model (Q = 12.2, df = 7, p = 0.094; I2 = 42.6 %; Fig. 2). In the stratification studies, by comparing the highest quantile dietary folate intake to the lowest quantile, we found 16 % (RR = 0.84, 95 % CI = 0.61–1.14; p = 0.265) and 8 % (OR = 0.92, 95 % CI = 0.74–1.14; p = 0.431) reduction of ovarian cancer risk for the prospective and case–control studies, respectively. Moreover, when the study of Kelemen et al., which contributed to the sensitivity of summary estimate, was excluded from the pooled prospective studies, we found high dietary folate intake was significantly associated with a reduced ovarian cancer risk (RR = 0.73, 95 % CI = 0.56–0.94; p = 0.0153) with no significant heterogeneity between the studies (Q = 0.19, df = 2, p = 0.9077; I2 = 0 %). In the subgroup analysis of studies performed in Northern America countries (three for USA and one for Canada), our results indicated a marginal protective effect of dietary folate intake on ovarian cancer risk but not for the total folate (dietary folate: RR = 0.88, 95 % CI = 0.76–1.03, p = 0.105; total folate: RR = 1.07, 95 % CI = 0.81–1.43, p = 0.621). Similarly, statistical association between high dietary folate intake and reduced ovarian cancer risk was found after the exclusion of the
Overall, we found no significant association between MTHFR variant C677T and ovarian cancer risk both using the dominant model (OR = 1.05, 95 % CI = 0.94–1.16; p = 0.379) and the recessive model (OR = 1.13, 95 % CI = 0.91–1.39; p = 0.264). Compared to the homozygote CC, neither the TT carriers nor the CT carriers showed a significant altered risk of ovarian cancer (Fig. 3). In the stratification study, we found a marginal association of allele T carriers with an increased ovarian cancer risk (CT vs. CC: OR = 1.15, 95 % CI = 0.99–1.32; p = 0.061; Table 4) in the Caucasian populations from America and significant association in Asian studies under three models (dominant: OR = 1.39, 95 % CI = 1.10–1.77, p = 0.007; recessive: OR = 2.39, 95 % CI = 1.58–3.62, p = 0.001; TT vs. CC: OR = 2.66, 95 % CI = 1.66–4.26, p = 0.001; Fig. 4). No significant association was found for the variant and ovarian cancer risk in any other stratification studies. We detected no publication bias in any study pool.
Discussion Folate deficiency may lead to neural tube defects of newborns, and folate has been reported to be preventive for cardiovascular disease and several types of cancer, therefore, it is recommended for all women of reproductive age in different countries [7]. Since January 1998, generalized folic acid fortification of flour and cereal grain products at a level of 140 lg/100 g was mandated in the US and Canada [35]. However, whether higher folate intake will benefit the women from the reduced ovarian cancer risk remains unknown. For the current meta-analysis, we found no
Table 3 Main results of effects of folate intake on ovarian cancer risk from overall and subgroup analysis Study types
Intake means
Fixed effects model OR/RR [95 % CI]
Random effects model p
OR/RR [95 % CI]
p
Heterogeneity 2
I (%)
p
Publication bias p
Prospective studies
Dietary
0.82 [0.65, 1.04]
0.100
0.84 [0.61, 1.14]
0.265
40.8
0.167
0.472
Case–control studies
Dietary
0.92 [0.81, 1.05]
0.212
0.92 [0.74, 1.14]
0.431
53.3
0.093
0.832
Overall studies
Dietary
0.90 [0.80, 1.00]
0.059
0.88 [0.75, 1.05]
0.158
42.6
0.094
0.715
American studies
Total
1.01 [0.89, 1.14]
0.914
1.04 [0.87, 1.23]
0.681
37.2
0.189
0.058
Dietary
0.88 [0.76, 1.03]
0.105
0.89 [0.72, 1.09]
0.261
23.8
0.268
0.789
Total
0.99 [0.85, 1.15]
0.862
1.07 [0.81, 1.43]
0.621
56.2
0.102
0.088
Data in bold was chosen for analysis after the consideration of heterogeneity between studies
123
6556
Mol Biol Rep (2013) 40:6547–6560
Fig. 2 Forest plot of the overall association between dietary folate intake and ovarian cancer risk
Fig. 3 Forest plots of the association between MTHFR polymorphism C677T and ovarian cancer risk from overall studies under a TT versus CC and b CT versus CC models
significant association between folate intake or the polymorphism of C677T on MTHFR and ovarian cancer risk. The meta-analysis for folate intake and ovarian cancer risk revealed no significant overall association between both high dietary and total folate intake and ovarian cancer risk. A marginal but not significant protective effect of high
123
dietary folate intake was observed rather than for the total folate intake which includes dietary and supplemental folate intake. We conducted the sensitivity analysis to test if certain study contributes most to the heterogeneity between the studies, discovering a significant inverse relationship between high dietary folate intake and reduced
Mol Biol Rep (2013) 40:6547–6560
6557
Table 4 Association between MTHFR polymorphism C677T and ovarian cancer risk from overall and subgroup studies Study types
Overall studies
American
Asian
Population based studies
Hospital/clinic based studies
Genetic model
Fixed effects model
Random effects model
Heterogeneity 2
Publication bias
OR [95 % CI]
p
OR [95 % CI]
p
I (%)
p
p
Dominant
1.05 [0.98, 1.13]
0.200
1.05 [0.94, 1.16]
0.379
34.9
0.129
0.899
Recessive
1.07 [0.96, 1.20]
0.211
1.13 [0.91, 1.39]
0.265
56.8
0.008
0.249
TT versus CC
1.11 [0.98, 1.25]
0.098
1.15 [0.90, 1.46]
0.264
62.1
0.005
0.476
CT versus CC
1.05 [0.97, 1.13]
0.227
1.04 [0.94, 1.16]
0.423
29.9
0.170
0.387
Dominant
1.07 [0.93, 1.22]
0.345
1.07 [0.93, 1.22]
0.345
0
0.556
0.351
Recessive
0.92 [0.74, 1.13]
0.413
0.81 [0.55, 1.20]
0.290
57
0.098
0.088
TT versus CC CT versus CC
1.00 [0.80, 1.25] 1.15 [0.99, 1.32]
0.991 0.061
0.86 [0.55, 1.35] 1.15 [0.99, 1.32]
0.508 0.061
64.1 0
0.062 0.686
0.059 0.413
Dominant
1.39 [1.10, 1.77]
0.007*
1.39 [1.10, 1.77]
0.007*
0
0.720
0.938
Recessive
2.39 [1.58, 3.62]
0.001*
2.39 [1.58, 3.62]
0.001*
0
0.618
0.341
TT versus CC
2.66 [1.66, 4.26]
0.001*
2.66 [1.66, 4.26]
0.001*
0
0.427
0.589
CT versus CC
1.23 [0.95, 1.58]
0.114
1.23 [0.95, 1.58]
0.114
0
0.493
0.238
Dominant
1.01 [0.91, 1.12]
0.887
0.99 [0.83, 1.18]
0.913
41.7
0.161
0.783
Recessive
1.04 [0.88, 1.22]
0.653
1.04 [0.88, 1.22]
0.671
3.2
0.388
0.485
TT versus CC
1.08 [0.91, 1.29]
0.380
1.08 [0.81, 1.44]
0.615
43.2
0.152
0.653
CT versus CC
1.02 [0.91, 1.14]
0.786
0.97 [0.78, 1.21]
0.795
56.2
0.077
0.417
Dominant
1.16 [0.98, 1.37]
0.091
1.15 [0.90, 1.46]
0.260
49.2
0.116
0.700
Recessive
1.39 [1.04, 1.85]
0.024*
1.61 [0.91, 2.84]
0.103
72
0.006
0.190
TT versus CC
1.40 [1.03, 1.91]
0.034*
1.55 [0.81, 2.99]
0.187
75.8
0.006
0.381
CT versus CC
1.12 [0.93, 1.33]
0.225
1.10 [0.87, 1.39]
0.421
40.5
0.169
0.330
Data in bold was chosen for analysis after the consideration of heterogeneity between studies * Significance
Fig. 4 Forest plots of the association between MTHFR polymorphism C677T and ovarian cancer risk from Asian studies under a dominant and b recessive models
ovarian cancer risk from the prospective studies and from the studies conducted in the US when the study by Kelemen et al. [22] was excluded from the meta-analysis
(prospective studies: RR = 0.73, 95 % CI = 0.56–0.94; Northern Americans: RR = 0.85, 95 % CI = 0.73–0.99), and after which no significant heterogeneity between
123
6558
studies was found. However, the pooled data showed that high total folate intake has no effect on ovarian cancer risk. It appears that dietary folate intake is more likely to act as a protective role than the supplemental folate intake on ovarian cancer. This may be attributed to the higher bioavailability of dietary folate than supplements of folate. Another reasonable explanation could be the co-existence of folate and other potential anticancer ingredients such as vitamin C, vitamin E and fiber in green vegetables, which may collectively contribute to the protective effect against ovarian cancer. Although we found no significant association between high total folate intake from supplemental and dietary resources and ovarian cancer risk, it should be noted that several studies have suggested that excessive folate intake from supplements may be harmful. Ulrich and Potter [7] proposed that despite the benefits of folate in neural tube defects prevention, some warning lights also existed, leading to the need for reevaluation of the potential risks of excessive folate intake. Most of the current studies that evaluated the association between folate intake and ovarian cancer risk were conducted in North America, where the folate fortification has in place for a long period of time. It was reported that the serum folate concentration for 40 % of the US populations has exceeded 40 nM after folic acid fortification, which was considered as supraphysiologic and excessive folate intake resulted in an increased unmetabolized folic acid in the circulation of healthy individuals, which could lead to the disturbance of folate metabolism homeostasis in the body [36]. These also may lead to the negative association of high total folate intake and ovarian cancer risk found here. Several studies reported an adverse effect for high folate intake. For instance, Kim [37] found that colorectal mucosa could be promoted rather than suppressed by exceptionally high supplemental folate intake after microscopic neoplastic foci are established. Miller et al. [38] also pointed out that folate deficiency rather than folate overdose inhibit the progression of mammary tumor, which may be due to the indispensable role of folate in the synthesis for DNA and RNA. However, whether high folate intake will affect the prognosis of ovarian cancer patients is still unclear. We speculate that not only the dose but also the time of folate intake should be taken into consideration in future studies to fully assess the correlations of folate intake and ovarian cancer risk. With a total of 5,617 cases and 9,808 controls from 12 subgroup studies, we found no significant correlation between the MTHFR SNP C677T and ovarian cancer risk, whereas a significant increased risk of ovarian cancer for the allele T carriers was detected in Asian populations, which may be due to the relative small sample size of all Asian population included (case:control = 551:844). However, the statistical power (p = 0.001) also reminded
123
Mol Biol Rep (2013) 40:6547–6560
us of the potential differences between Asian and other population. It is acceptable when considering the dietary structure and living habits of individuals form different regions. The consumption of alcohol varies greatly in Asian and American population, which may alter the involvement of the polymorphism in folate metabolism. The negative results of overall studies could be explained by possible alternative pathway besides MTHFR and the disparity of the populations as well as the relative small sample size of each study. It should be noted that there are several limitations in our current meta-analysis of published data. Firstly, the sample size is still relatively small both for the folate intake and the MTHFR SNP C677T and ovarian cancer risk. Secondly, the folate intake and folate metabolism pathways may be modulated by the alcohol consumption; many of the studies identified here have not taken the alcohol intake into account. Thirdly, the majority of the studies were conducted in Caucasian population, so there might be a bias for the entire population. Lastly, the diversity of population brought in difficulties in subgroup analysis. It is also debatable to pool all populations with different living habits, dietary structure and genetic backgrounds for an overall estimate. In summary, we found no significant association between the high folate intake or MTHFR SNP C677T and reduced ovarian cancer form overall population. Our results suggested that higher dietary folate intake may reduce the risk of ovarian cancer; and association between polymorphisms and cancer incidence probably should be considered under gene background as well as living habits, however, more studies are warranted to fully address the question. Acknowledgments This study was supported by grants from the Ministry of Science and Technology of China (2012BAK01B00 and 2011BAK10B00), the National Nature Science Foundation (81125020, 31101261 and 31200569), the Key Research Program (KSZD-EW-Z-021) of the Chinese Academy of Sciences, the Science and Technology Commission of Shanghai Municipality (12XD140 7000, 12431900500 and 10391902100), Director Foundation (20090 101) and the Food Safety Research Center and Key Laboratory of Food Safety Research of INS, SIBS, CAS. Peizhan Chen was partially supported by the SA-SIBS scholarship program. Conflict of interest
The authors declare no conflict of interest.
References 1. American Cancer Society (2011) Cancer facts and figures 2011. American Cancer Society, Atlanta 2. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D (2011) Global cancer statistics. CA Cancer J Clin 61(2):69–90. doi:10. 3322/caac.20107 3. Jemal A, Siegel R, Xu J, Ward E (2010) Cancer statistics, 2010. CA Cancer J Clin 60(5):277–300. doi:10.3322/caac.20073
Mol Biol Rep (2013) 40:6547–6560 4. Edmondson RJ, Monaghan JM (2001) The epidemiology of ovarian cancer. Int J Gynecol Cancer Off J Int Gynecol Cancer Soc 11(6):423–429 5. Kolahdooz F, van der Pols JC, Bain CJ, Marks GC, Hughes MC, Whiteman DC, Webb PM (2010) Meat, fish, and ovarian cancer risk: results from 2 Australian case–control studies, a systematic review, and meta-analysis. Am J Clin Nutr 91(6):1752–1763. doi:10.3945/ajcn.2009.28415 6. Kolahdooz F, Ibiebele TI, van der Pols JC, Webb PM (2009) Dietary patterns and ovarian cancer risk. Am J Clin Nutr 89(1):297–304. doi:10.3945/ajcn.2008.26575 7. Ulrich CM, Potter JD (2006) Folate supplementation: too much of a good thing? Cancer Epidemiol Biomarkers Prev Publ Am Assoc Cancer Res Cospons Am Soc Prev Oncol 15(2):189–193. doi:10.1158/1055-9965.EPI-152CO 8. Kim YI (1999) Folate and carcinogenesis: evidence, mechanisms, and implications. J Nutr Biochem 10(2):66–88 9. Choi SW, Mason JB (2000) Folate and carcinogenesis: an integrated scheme. J Nutr 130(2):129–132 10. Kim DH, Smith-Warner SA, Spiegelman D, Yaun SS, Colditz GA, Freudenheim JL, Giovannucci E, Goldbohm RA, Graham S, Harnack L, Jacobs EJ, Leitzmann M, Mannisto S, Miller AB, Potter JD, Rohan TE, Schatzkin A, Speizer FE, Stevens VL, Stolzenberg-Solomon R, Terry P, Toniolo P, Weijenberg MP, Willett WC, Wolk A, Zeleniuch-Jacquotte A, Hunter DJ (2010) Pooled analyses of 13 prospective cohort studies on folate intake and colon cancer. Cancer Causes Control 21(11):1919–1930. doi:10.1007/s10552-010-9620-8 11. Larsson SC, Giovannucci E, Wolk A (2007) Folate and risk of breast cancer: a meta-analysis. J Natl Cancer Inst 99(1):64–76. doi:10.1093/jnci/djk006 12. Larsson SC, Giovannucci E, Wolk A (2004) Dietary folate intake and incidence of ovarian cancer: the Swedish Mammography Cohort. J Natl Cancer Inst 96(5):396–402. doi:10.1093/jnci/ djh061 13. Webb PM, Ibiebele TI, Hughes MC, Beesley J, van der Pols JC, Chen X, Nagle CM, Bain CJ, Chenevix-Trench G, Australian Cancer Study, Australian Ovarian Cancer Study Group (2011) Folate and related micronutrients, folate-metabolising genes and risk of ovarian cancer. Eur J Clin Nutr 65(10):1133–1140 14. Harris HR, Cramer DW, Vitonis AF, Depari M, Terry KL (2011) Folate, vitamin B(6), vitamin B(12), methionine and alcohol intake in relation to ovarian cancer risk. Int J Cancer. doi:10.1002/ijc. 26455 15. Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, Boers GJ, den Heijer M, Kluijtmans LA, van den Heuvel LP et al (1995) A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 10(1):111–113. doi:10.1038/ng0595-111 16. Friso S, Choi SW, Girelli D, Mason JB, Dolnikowski GG, Bagley PJ, Olivieri O, Jacques PF, Rosenberg IH, Corrocher R, Selhub J (2002) A common mutation in the 5,10-methylenetetrahydrofolate reductase gene affects genomic DNA methylation through an interaction with folate status. Proc Natl Acad Sci USA 99(8):5606–5611. doi:10.1073/pnas.062066299 17. Molloy AM (2012) Genetic aspects of folate metabolism. Subcell Biochem 56:105–130. doi:10.1007/978-94-007-2199-9_7 18. Macis D, Maisonneuve P, Johansson H, Bonanni B, Botteri E, Iodice S, Santillo B, Penco S, Gucciardo G, D’Aiuto G, Rosselli Del Turco M, Amadori M, Costa A, Decensi A (2007) Methylenetetrahydrofolate reductase (MTHFR) and breast cancer risk: a nested-case–control study and a pooled meta-analysis. Breast Cancer Res Treat 106(2):263–271. doi:10.1007/s10549-0069491-6 19. Taioli E, Garza MA, Ahn YO, Bishop DT, Bost J, Budai B, Chen K, Gemignani F, Keku T, Lima CS, Le Marchand L, Matsuo K,
6559
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
Moreno V, Plaschke J, Pufulete M, Thomas SB, Toffoli G, Wolf CR, Moore CG, Little J (2009) Meta- and pooled analyses of the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism and colorectal cancer: a HuGE-GSEC review. Am J Epidemiol 170(10):1207–1221. doi:10.1093/aje/kwp275 Cui LH, Shin MH, Kim HN, Song HR, Piao JM, Kweon SS, Choi JS, Yun WJ, Kim YC, Oh IJ, Kim KS (2011) Methylenetetrahydrofolate reductase C677T polymorphism in patients with lung cancer in a Korean population. BMC Med Genet 12:28. doi:10.1186/1471-2350-12-28 Zhang J, Qiu LX, Wang ZH, Wu XH, Liu XJ, Wang BY, Hu XC (2010) MTHFR C677T polymorphism associated with breast cancer susceptibility: a meta-analysis involving 15,260 cases and 20,411 controls. Breast Cancer Res Treat 123(2):549–555. doi:10.1007/s10549-010-0783-5 Kelemen LE, Sellers TA, Vierkant RA, Harnack L, Cerhan JR (2004) Association of folate and alcohol with risk of ovarian cancer in a prospective study of postmenopausal women. Cancer Causes Control 15(10):1085–1093 Gershoni-Baruch R, Dagan E, Israeli D, Kasinetz L, Kadouri E, Friedman E (2000) Association of the C677T polymorphism in the MTHFR gene with breast and/or ovarian cancer risk in Jewish women. Eur J Cancer 36(18):2313–2316 Pepe C, Guidugli L, Sensi E, Aretini P, D’Andrea E, Montagna M, Manoukian S, Ottini L, Radice P, Viel A, Bevilacqua G, Caligo MA (2007) Methyl group metabolism gene polymorphisms as modifier of breast cancer risk in Italian BRCA1/2 carriers. Breast Cancer Res Treat 103(1):29–36. doi:10.1007/ s10549-006-9349-y Navarro Silvera SA, Jain M, Howe GR, Miller AB, Rohan TE (2006) Dietary folate consumption and risk of ovarian cancer: a prospective cohort study. Eur J Cancer Prev 15(6):511–515. doi:10.1097/01.cej.0000220627.54986.bf Pelucchi C, Mereghetti M, Talamini R, Negri E, Montella M, Ramazzotti V, Franceschi S, La Vecchia C (2005) Dietary folate, alcohol consumption, and risk of ovarian cancer in an Italian case–control study. Cancer Epidemiol Biomarkers Prev Publ Am Assoc Cancer Res Cospons Am Soc Prev Oncol 14(8):2056–2058. doi:10.1158/1055-9965.EPI-05-0192 Tworoger SS, Hecht JL, Giovannucci E, Hankinson SE (2006) Intake of folate and related nutrients in relation to risk of epithelial ovarian cancer. Am J Epidemiol 163(12):1101–1111. doi:10.1093/aje/kwj128 Zhang L, Liu W, Hao Q, Bao L, Wang K (2012) Folate intake and methylenetetrahydrofolate reductase gene polymorphisms as predictive and prognostic biomarkers for ovarian cancer risk. Int J Mol Sci 13(4):4009–4020. doi:10.3390/ijms13044009 Prasad VV, Wilkhoo H (2011) Association of the functional polymorphism C677T in the methylenetetrahydrofolate reductase gene with colorectal, thyroid, breast, ovarian, and cervical cancers. Onkologie 34(8–9):422–426. doi:10.1159/000331131 Pawlik P, Mostowska A, Lianeri M, Sajdak S, Kedzia H, Jagodzinski PP (2011) Folate and choline metabolism gene variants in relation to ovarian cancer risk in the Polish population. Mol Biol Rep. doi:10.1007/s11033-011-1359-0 Jakubowska A, Gronwald J, Menkiszak J, Gorski B, Huzarski T, Byrski T, Edler L, Lubinski J, Scott RJ, Hamann U (2007) Methylenetetrahydrofolate reductase polymorphisms modify BRCA1-associated breast and ovarian cancer risks. Breast Cancer Res Treat 104(3):299–308. doi:10.1007/s10549-006-9417-3 Terry KL, Tworoger SS, Goode EL, Gates MA, Titus-Ernstoff L, Kelemen LE, Sellers TA, Hankinson SE, Cramer DW (2010) MTHFR polymorphisms in relation to ovarian cancer risk. Gynecol Oncol 119(2):319–324. doi:10.1016/j.ygyno.2010.08.007 Gao S, Liu N, Ma Y, Ying L (2012) Methylenetetrahydrofolate reductase gene polymorphisms as predictive and prognostic
123
6560 biomarkers in ovarian cancer risk. Asian Pac J Cancer Prev 13(2):569–573 34. Jakubowska A, Rozkrut D, Antoniou A, Hamann U, Scott RJ, McGuffog L, Healy S, Sinilnikova OM, Rennert G, Lejbkowicz F, Flugelman A, Andrulis IL, Glendon G, Ozcelik H, Thomassen M, Paligo M, Aretini P, Kantala J, Aroer B, von Wachenfeldt A, Liljegren A, Loman N, Herbst K, Kristoffersson U, Rosenquist R, Karlsson P, Stenmark-Askmalm M, Melin B, Nathanson KL, Domchek SM, Byrski T, Huzarski T, Gronwald J, Menkiszak J, Cybulski C, Serrano P, Osorio A, Cajal TR, Tsitlaidou M, Benitez J, Gilbert M, Rookus M, Aalfs CM, Kluijt I, BoessenkoolPape JL, Meijers-Heijboer HE, Oosterwijk JC, van Asperen CJ, Blok MJ, Nelen MR, van den Ouweland AM, Seynaeve C, van der Luijt RB, Devilee P, Easton DF, Peock S, Frost D, Platte R, Ellis SD, Fineberg E, Evans DG, Lalloo F, Eeles R, Jacobs C, Adlard J, Davidson R, Eccles D, Cole T, Cook J, Godwin A, Bove B, Stoppa-Lyonnet D, Caux-Moncoutier V, Belotti M, Tirapo C, Mazoyer S, Barjhoux L, Boutry-Kryza N, Pujol P, Coupier I, Peyrat JP, Vennin P, Muller D, Fricker JP, VenatBouvet L, Johannsson OT, Isaacs C, Schmutzler R, Wappenschmidt B, Meindl A, Arnold N, Varon-Mateeva R,
123
Mol Biol Rep (2013) 40:6547–6560
35.
36. 37.
38.
Niederacher D, Sutter C, Deissler H, Preisler-Adams S, Simard J, Soucy P, Durocher F, Chenevix-Trench G, Beesley J, Chen X, Rebbeck T, Couch F, Wang X, Lindor N, Fredericksen Z, Pankratz VS, Peterlongo P, Bonanni B, Fortuzzi S, Peissel B, Szabo C, Mai PL, Loud JT, Lubinski J (2012) Association of PHB 1630 C[T and MTHFR 677 C[T polymorphisms with breast and ovarian cancer risk in BRCA1/2 mutation carriers: results from a multicenter study. Br J Cancer 106(12):2016–2024. doi:10.1038/bjc.2012.160 Honein MA, Paulozzi LJ, Mathews TJ, Erickson JD, Wong LY (2001) Impact of folic acid fortification of the US food supply on the occurrence of neural tube defects. JAMA 285(23):2981–2986 Ulrich CM (2007) Folate and cancer prevention: a closer look at a complex picture. Am J Clin Nutr 86(2):271–273 Kim YI (2004) Folate, colorectal carcinogenesis, and DNA methylation: lessons from animal studies. Environ Mol Mutagen 44(1):10–25. doi:10.1002/em.20025 Miller JW, Borowsky AD, Marple TC, McGoldrick ET, DillardTelm L, Young LJ, Green R (2008) Folate, DNA methylation, and mouse models of breast tumorigenesis. Nutr Rev 66(Suppl 1):S59–S64. doi:10.1111/j.1753-4887.2008.00070.x
