5,10-Methylenetetrahydrofolate reductase (MTHFR) polymorphisms implicated in the cancer development, but the published studies had yielded inconsistent results.
Pubmed was searched for all published case–control studies about MTHFR polymorphisms and prostate cancer risk.
In all, 13 studies including 5872 cases and 6255 controls described C677T genotypes, among which 9 articles, containing 2847 cases and 3657 controls described A1298C genotypes, were involved in our meta-analysis. Odds ratios (ORs) with 95% confidence intervals (CIs) were estimated to assess the association between MTHFR polymorphisms and prostate cancer risk, heterogeneity, publication bias and sensitivity were also calculated. Overall, meta-analysis indicated that the 677T allele was more likely to exert protective effect on prostate cancer risk (random-effects pooled OR, 0.78 (0.64–0.96); P=0.016 (P=0.033 for heterogeneity studies)) in a recessive genetic model, no associations were found in other genetic models or in comparing a/a versus A/A homozygous. Neither did we find any difference in effects on high or low aggressive prostate cancer. No evidence of an association of MTHFR A1298C polymorphism with prostate cancer was found.
C677T of the MTHFR gene may provide protective effects on susceptibility to prostate cancer risk.
With the increase of incidence and death rate, prostate cancer has become a serious public threat to men throughout the world,1 but the precise mechanisms of prostate carcinogenesis have not become apparent yet. Recently, there is a growing body of evidence that aberrant DNA methylation seems to have a pivotal role in prostate carcinogenesis.2, 3 5, 10-Methylenetetrahydrofolate reductase (MTHFR) is a key enzyme that is involved in the folate metabolism and thus may promote DNA methylation, suggesting that MTHFR might be an important factor that related to prostate carcinogenesis.
MTHFR C677T (Ala222Val) and A1298C (Glu429Ala), two main common types of polymorphisms, could inhibit the enzyme activity of MTHFR.4, 5 Reduced activity of MTHFR may result in DNA hypomethylation, and hence increase cancer susceptibility, but the hypomethylation of the CpG promoter sequences of tumor suppressor gene might, such as GSTP1 (glutathione-S-transferase), which is frequently silenced by hypermethylation in prostate cancer cells, protect against prostatic tumourigenesis.6, 7 Meanwhile, a high concentration of 5, 10-MTHF could maintain sufficient thymidylate, reduce DNA misrepair and chromosome breakage by uracil misincorporation, and thus maintain genetic stability.8
Studies about the association of MTHFR polymorphisms with some cancers, such as colorectal and gastric cancers, had been manifested,9, 10, 11 but the published studies about the MTHFR C677T and A1298C polymorphisms with prostate cancer risk is conflicting. In 2009, two meta-analyses had been published,12, 13 but results from those two papers were also inconsistent. Bai et al. study12 selected seven papers published from 2000 to 2007, which included 3511 cases and 2762 controls, demonstrated that 677T allele exert protective effect on prostate cancer risk with recessive genetic model, while in Collin et al. study,13 which contained nine papers published from 2000 to 2008 and included 5170 cases and 5085 controls found no effect of MTHFR C677T and A1298C on the prostate cancer risk. All the participants included in those two meta-analyses were Caucasians. During 2008–2010, four new papers had been published, which provided 702 new cases and 1170 new controls, particularly two of which included Asian participants.
In view of the conflicting results of the previous studies and the insufficient statistical power of a single study, we aimed to explore the precise relationship of MTHFR C677T and A1298C polymorphisms with risk of prostate cancer by conducting a meta-analysis of case–control studies.
Materials and methods
We searched the PubMed database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi) for all associated studies on MTHFR and prostate risk published in any language from 1999 through September 2011. The keywords we used were ‘MTHFR’, ‘polymorphisms’ and ‘prostate cancer’, and ‘prostatic cancer’. Only published studies with full text were included, abstracts or unpublished studies were excluded. We also searched the reference lists of original articles and recent reviews for additional studies, and the PubMed option ‘Related Articles’ was also used to search for potentially relevant papers. When the same patient population was included in several publications, only the most recent or complete study was included in this meta-analysis.
Studies matching the following criteria were selected as follows: (a) they were case–control studies, (b) evaluation of the MTHFR C677T or A1298C polymorphism and prostate cancer risk had been made and (c) information about the sample size, odds ratios (ORs) and their 95% confidence intervals (CIs), or other information that can help infer the results in the papers had been included. Minimum number of patients for including a study in our meta-analysis was not defined. Studies in which controls had BPH or which included female controls were also accepted.
The following data were collected from each study: first author's surname, publication year, country, ethnicity, study design, characteristics of controls, genotyping methods, total number of cases and controls with MTHFR C677T and A1298C genotypes, respectively. Ethnicity was categorized as Caucasians, Asian, African or Mixed. Where available, genotype frequencies by cancer stage and grade, that is, high aggressive/high-risk (stage III/IV or Gleason of 7) or low aggressive/low-risk cases (stage I/II and Gleason of <7) were also extracted.
Hardy–Weinberg equilibrium (HWE) in the control group was tested by the Pearson’s χ2 test. Crude ORs with 95% CIs were calculated to assess the association of MTHFR C677T and A1298C polymorphisms with prostate cancer risk, and ORs of high and low aggressive prostate cancer were also calculated. The pooled ORs for MTHFR genotypes were calculated in four models: dominant ((a/a∣a/A) versus A/A), recessive (a/a versus (a/A∣A/A)), a/a homozygous versus A/A homozygous and additive (in which genotype was coded A/A=0 (reference), a/A=1, a/a=2), the significance was determined by the Z-test. Subgroup analysis was conducted in pooled European population across the eligible studies to evaluate ethnic differences in the relationship between MTHFR polymorphisms and prostate risk.
Heterogeneity assumption was tested by a χ2-based Q-test. A P-value of >0.10 indicated a lack of heterogeneity, so the pooled estimation of the ORs was calculated by the fixed effects model (Mantel–Haenszel method); otherwise, the random effects model (DerSimonian and Laird method) was used.14
One-way sensitivity analysis was performed to assess the stability of the results, namely, a single study in the meta-analysis was deleted each time to reflect the influence of the individual data set to the pooled ORs.15
The potential publication bias was conducted by funnel plot, and an asymmetric plot suggested a possible publication bias. The funnel plot symmetric was assessed by Egger's test.16
All the statistical tests were performed using STATA version 10.0 (Stata, College Station, TX, USA). A P-value <0.05 was considered statistically significant, except where otherwise specified.
Eligible study characteristics
Our systematic review identified 13 eligible studies based on the search criteria.17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29 The basic information, including authors and published years, country, ethnicity of the study populations, study design and the numbers of cases and controls of each study is listed in Table 1.
Except Cai28 and Wu29 reported on Asians, other 10 papers had 100% Caucasian subjects, and Singal et al.20 reported on mixed ethnicities (noted as black and white), which had 67% Caucasian subjects, and no Africans were identified in our study. Because of the inability to distinguish between racial-specific allele genotyping counts from ‘mixed’ study populations, 10 eligible papers were collected to perform ethnic subgroup analysis.
Eight studies recruited prostate cancer cases from hospitals,17, 20, 23, 24, 26, 27, 28, 29 four from prospective longitudinal cohorts or cancer registries,18, 21, 22, 25 and one from family study.19 Four of the hospital case–control studies recruited men with BPH to the control group.20, 24, 26, 28 One study included women among their control group,17 and one study selected sibling brother as a control group.19
All the papers used blood samples for genotyping except one,26 which used frozen tissue samples. PCR-restriction fragment length polymorphism was used to validate genotype in nine papers,17, 18, 19, 20, 23, 26, 27, 28, 29 and Taqman single-nucleotide polymorphism genotyping assay was used for the other four papers.21, 22, 24, 25
A total of 13 eligible papers including 5872 cases and 6255 controls described C677T genotypes, among which 9 papers totaling 2847 cases and 3657 controls described A1298C genotypes. The genotype frequencies of C667T and A1298C for all cases and controls are shown in Supplementary Table 1, and genotype frequencies for advanced and localized cases in the individual study are displayed in Supplementary Table 2.
Overall, the prevalence of 677TT and 1298CC homozygosis was 10.39 and 10.34% in control subjects, respectively. There was strong evidence (P<0.001) of a departure from HWE among the control group in the study for MTHFR A1298C,26 and marginal evidence (0.03<P<0.05) among the control groups in two studies for MTHFR C677T.21, 22
The T allele was associated with the decreased risk of prostate cancer in recessive model, not in dominant modeling, or in comparing T/T versus C/C genotypes, or in an additive (per allele) model (Table 2; Figure 1). There was substantial between-study heterogeneity in the dominant, recessive, T/T versus C/C and additive models. In subgroup analysis, we also did not deduce any significant results.
In sensitivity analyses, exclusion of the two studies23, 24 that showed weak evidence of departure from HWE among the control group (P=0.05 and P=0.03, respectively) yielded pooled (random effects) ORs (95% CI) for the dominant, recessive, T/T versus C/C and additive models of the following: 0.97 (0.91–1.04), P=0.402 (P=0.012 for heterogeneity); 0.82 (0.67–1.00), P=0.046 (P=0.062 for heterogeneity); 0.84 (0.69–1.03), P=0.092 (P=0.022 for heterogeneity) and 0.94 (0.87–1.02), P=0.145 (P=0.004 for heterogeneity), respectively.
If we excluded only one study of Marchal et al.24 that showed significant HW disequilibrium (P=0.03), the pooled random-effect OR (95% CI) for the dominant, recessive, T/T versus C/C and additive models were: 0.96 (0.89–1.03), P=0.218 (P=0.004 for heterogeneity across studies); 0.82 (0.68–0.99), P=0.037 (P=0.091 for heterogeneity across studies); 0.83 (0.68–1.01), P=0.062 (P=0.029 for heterogeneity across studies) and 0.93 (0.86–1.01), P=0.09 (P=0.004 for heterogeneity across studies), respectively. Excluding the Reljic et al. study,23 which was in significant HW disequilibrium (P=0.03), there had no significant effect on the ORs for all genetic models.
In a meta-analysis based on the seven studies that provided genotype count data for high aggressive (n=2093) and low aggressive (n=2647) prostate cancer cases (3871 controls), we found no differences in effects of the T allele on these two subgroups (Table 3). Excluding the Marchal et al. study that showed some evidence of HW disequilibrium (P=0.03) had no significant effect on the ORs for all genetic models.
We found no associations with prostate cancer in dominant, recessive or additive modeling, or in comparing homozygous genotypes for MTHFR A1298C. Neither did we find any differential association with advanced versus localized prostate cancer (Table 3). Exclusion of the studies that had evidence of departure from HWE made no difference to the above null findings. The associations of the A1298C polymorphisms with prostate cancer also did not change during the sensitivity analysis (data not shown).
Begg's funnel plot and Egger's test were performed to assess the publication bias of the literature. As shown in Figure 2, the shape of the funnel plots for the comparison of the 677T allele and the 677C allele seemed symmetrical in all comparing models, which suggested the publication bias might not affect the findings of our meta-analysis. Then, Egger's test was used to provide statistical evidence for funnel plot symmetry (P=0.364). The results still did not suggest any evidence of publication bias (P=0.349 for CT versus CC, and P=0.641 for recessive model, P=0.534 for TT versus CC, P=0.361 for dominant model, respectively). Meanwhile, no publication bias was detected for associations of the A1298C polymorphisms with prostate cancer (data not shown).
On the basis of 13 case–control studies focused on MTHFR polymorphisms and prostate cancer risk, our meta-analysis provided evidence that the mutant homozygote 677TT was significantly associated with a decreased prostate cancer risk, and the 677T allele exerted the protective effect on prostate cancer risk in recessive genetic model (OR=0.78, 95% CI: 0.64–0.96, P=0.016). No associations were found for the MTHFR C677T genotype in other genetic models, and the same protective effect of T allele was not observed in the European population (OR=0.81, 95% CI: 0.64–1.02, P=0.078). Our findings suggest that, as described in Collin et al. study,13 the aberrant DNA methylation implicated by epigenetic studies of prostate cancer cells is not induced primarily by genetically determined enzymatic defects in the folate metabolic pathway, and may differ from the methylation processes postulated for colorectal and gastric carcinogenesis.
Our results about the protective role of MTHFR 677TT genotype for the prostate cancer is consistent with Singal et al. study,20 in which 81 patients with prostate cancer and 42 controls with benign prostatic hypertrophy were observed, and a previous meta-analysis12 also supported our findings about the protective role of 677T allele (OR=0.81, 95% CI=0.68–0.98), but not for the 677CT (OR=1.13, 95% CI=0.88–1.45) and the 677TT mutant genotype (OR=0.85, 95% CI=0.71–1.03). In addition, results about the 677TT genotype protective effect on colorectal cancer,8 acute lymphocyte leukemia30 and malignant lymphoma31 could also indirectly support our findings.
Furthermore, previous research6, 7 demonstrated that tumor suppressor genes in prostate cancer cells are frequently silenced, such as GSTP1, its CPG island sequences in the promoter region is often in the status of hypermethylation, and decrease its methylation of promoter could activate its role against prostatic tumorgenesis. The reduced activity of MTHFR may decrease the methylation of homocysteine to methionine and in turn the level of s-adenosyl methionine, resulting in DNA hypomethylation, thus promoting the expression of tumor suppressor genes and inhibiting the development of cancer. Meanwhile, inhibition the function of MTHFR could also arrest the proliferation of cancer cells because of the decrease of methionine supplement in vitro.32 In light of above results, we could also conclude that mutant homozygote 677TT could decrease the risk of prostate cancer by reducing the enzymatic activity of MTHFR.
However, results from Heijmans et al.18 demonstrated that 677TT genotype could significantly increase the risk of prostate cancer, and study of Johansson et al.22 also showed that patients carrying 677TT genotype displayed an increase in prostate cancer risk among men younger than 65 years, but in those two research, a deviation from HWE in the controls might induce the above results. Collin et al. study13 also demonstrated that MTHFR 677TT genotype had no effect on the risk of prostate cancer.
We found no differential effects of MTHFR C677T and A1298C gene polymorphisms on high aggressive versus low aggressive prostate cancer, suggesting that the role, if any, of the folate metabolic pathway in prostatic carcinogenesis does not change as the cancer progresses.
A test for HWE often assumes that the genotypes are a random sample from the large, randomly mating populations. Genotyping errors can distort genotype distribution and can lead to departure from HWE. In our pooled analysis, the genotypes among controls were in HWE for 10 studies. But for the three studies of HW disequilibrium,21, 22, 24 genotyping error was minimized by strict quality control, therefore, in the final analysis, these studies were also included. Furthermore, we further performed sensitivity analysis to detect the stability of the meta-analysis, and the deviations did not alter the pattern of association, which revealed that the protective effect of 677TT genotype was stable.
In addition, publication bias was not detected for both C677T and A1298C, which indicated that our findings seemed unlikely due to biased publications.
In our meta-analysis, we did not find any relation between the risk of prostate cancer and MTHFR A1298C polymorphism, it might be the little effect of A1298C on the enzyme activity, or the relatively small sample number.
Our study had several limitations that affected interpretation of the results. In this study, there were no uniform controls. Healthy populations, as well as BPH patients20, 24, 26, 28 and siblings19 were selected as control, and some women17 were also included. Some individuals in the control group are likely to develop prostate cancer in subsequent years though they had no clinical symptoms at the time of investigation, and gene effects may have been underestimated in studies that used patients with BPH as controls, but these control groups were used only by a small number of the studies. Furthermore, in the sensitivity analysis, when we omitted above studies, the pooled ORs had no obvious change. Second, folate supplement, alcohol intake could affect the function of the MTHFR on prostate cancer risk. Heijmans et al.18 demonstrated a higher relative risk for TT heterozygote in the group of low dietary intake compared with the group with high folate intake, yet, those data were not available from our included studies, and the interactions were therefore not analyzed. Further researches are required to explore the synergetic role of folate intake and MTHFR polymorphisms in prostate risk. Third, of all the included studies most were conducted in Caucasians, only two studies included Asians, and no African was collected, so subgroup meta-analyses by race were not possible, only studies on Caucasians (except for one study21 on mixed populations) were obtained, which was not representative of all ethnicities. The test for heterogeneity between subgroups may be invalid when there is substantial heterogeneity within subgroups, as observed within advanced and localized subgroups for some of the single-nucleotide polymorphisms (Table 3).
In conclusion, our meta-analysis had collected the largest samples from the published papers to explore the relationship between MTHFR polymorphisms and prostate cancer risk by now, and the results give strong support to the opinion that the MTHFR 677T allele could decrease the risk of prostate cancer. However, large trials using standardized unbiased methods, enrolling precisely defined prostate cancer patients and well-matched controls, with more detailed individual data are needed. Moreover, it has been postulated that it is the combined effects of inherited and environmental factors that induced prostate cancer, so further studies are required to manifest the interaction between MTHFR C677T and the environmental factors.
This work was supported by grants from the National Natural Science Foundation of China (no._81001185), Universities Natural Science Foundation of Jiangsu Province (no._10KJB310011) and the Social Development Foundation of Suzhou (no_YJS0905).
About this article
Supplementary Information accompanies the paper on the Prostate Cancer and Prostatic Diseases website (http://www.nature.com/pcan)