A Polymorphism of the Methionine Synthase Gene: Association with Plasma Folate, Vitamin B₁₂, Homocyst(e)ine, and Colorectal Cancer Risk¹

Methionine synthase, a vitamin B₁₂-dependent enzyme, plays an important role in folate metabolism (1). It catalyzes the transfer of a methyl group from 5-methyltetrahydrofolate to homocysteine, producing methionine and tetrahydrofolate. Methionine synthase is critical for maintaining adequate intracellular methionine, an essential amino acid and the precursor of SAM.⁴ SAM is a crucial methyl group donor involved in over 100 methylation reactions including DNA methylation. Hypomethylation of the promoter regions of proto-oncogenes (2) or hypermethylation of these regions of tumor suppressor genes (3) may cause selective growth and transformation of cells. De novo methylation of CpG islands within the promoter regions of growth-regulatory genes, which may inactivate their transcription, is frequently observed in colonic tumors (4). Methionine synthase is also essential for maintaining adequate intracellular folate pools and ensuring that homocysteine concentrations do not reach toxic levels. Severe deficiency of vitamin B₁₂ or of methionine synthase causes hypomethioninemia, hyperhomocysteinemia, and homocystinuria (5). It may also result in an accumulation of 5-methyltetrahydrofolate and depletion of intracellular folate derivatives including 5,10-methylenetetrahydrofolate required for thymidylate biosynthesis, the basis of the well-documented “methyl folate trap” leading to deoxynucleotide pool imbalances and megaloblastic anemia (1, 6, 7). This leads to the accumulation of deoxyuridylate in DNA, and removal of this abnormal base may damage DNA, perhaps leading to strand breaks commonly seen in colorectal cancers (7, 8). Pernicious anemia, caused by vitamin B₁₂ malabsorption, has been associated with an elevated risk of cancer of the esophagus, stomach, and colon (9, 10). Recently, a polymorphism in the methionine synthase gene (MTR; 2756A→G), resulting in the substitution of aspartic acid (D919) by glycine (G), was identified in patients with methionine synthase deficiency and was found to be polymorphic among healthy controls (11).

We previously reported, from the same group of participants (12, 13), that a polymorphism (677C→T) that encodes a thermolabile MTHFR and causes a decrease in plasma 5-methyltetrahydrofolate was associated with a lower risk of colorectal cancer. This protective effect could result from increased availability of intracellular 5,10-methylenetetrahydrofolate and consequent reductions of uracil incorporation into DNA or from decreased SAM levels leading to DNA hypomethylation. Low folate intake or high alcohol consumption (which interferes with folate metabolism) seemed to negate some of the protective effect (12, 13). In the present study nested in two large cohorts, we examined the association of the MTR polymorphism with the risk of colorectal cancer and whether the association differs by plasma levels of folate, vitamin B₁₂, tHcy, or alcohol intake. We hypothesize that, if the variant genotype (gly/gly) of this MTR polymorphism is associated with a decreased activity of methionine synthase, men with the gly/gly genotype would have lower cellular methionine and folate derivatives, elevated tHcy levels, and an increased risk of colorectal cancer. Alternatively, lower methionine and SAM may lead to DNA hypomethylation, which would modify the cancer risk. For comparison, we also present some results for the MTHFR polymorphism.

The PHS is a randomized, double-blind, placebo-controlled 2 × 2 factorial trial of low-dose aspirin and β-carotene among 22,071 predominantly Caucasian-American male physicians, ages 40–84 years. Blood samples were collected at baseline, in 1982, from 14,916 (68%) of the randomized physicians. Alcohol consumption (drinks of beer, wine, or liquor) was ascertained from the baseline questionnaire. The men were subsequently followed for incident cancer through biannual mailed questionnaires. After 13 years of follow-up, 212 cases of colorectal cancer were identified and confirmed by medical records. Three hundred forty-six men who were free from diagnosed cancer at the time of case ascertainment were selected as controls and were matched on age (1 year) and smoking status (never, past, current).

The HPFS is a prospective study of 51,529 predominantly Caucasian-American male health professionals, ages 40–75, enrolled in 1986. Alcohol consumption (drinks of beer, wine, or liquor) was ascertained from a semiquantitative food frequency questionnaire at baseline. Blood samples were collected between 1993 and 1994 from 18,025 participants, among whom 144 had been diagnosed with colorectal cancer between 1986 and 1994; 130 cancer-free men were selected as controls. These cancer cases were also confirmed by medical records. The participants were predominantly Caucasians (over 95% in both the PHS and the HPFS).

DNA from cases and controls from both studies was extracted and MTR genotype was analyzed in Dr. Rozen’s laboratory; investigators and laboratory personnel were blinded to case-control status. The presence of the mutation was determined by PCR of genomic DNA, followed by HaeIII restriction digestion, as previously described (11). Because plasma levels of folate, vitamin B₁₂, and tHcy may be altered by change of diet or treatment after the diagnosis of cancer, these biomarkers were measured only for the PHS participants. Samples from cases and their matched controls were assayed in the same batch to minimize interassay variability, and aliquots from a pool of quality control plasma were inserted randomly. Laboratory personel were unable to distinguish among case, control, and quality control samples. They were also unaware of the genotype status. Plasma levels of folate and vitamin B₁₂ were determined using a radioassay kit (Ciba-Corning, Walpole, MA) in the laboratory of the Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA. Total tHcy levels were determined in the same laboratory as described previously(14). The average intra-assay coefficient of variation for folate, vitamin B₁₂, and tHcy were 6.8%, 3.0%, and 2.9%, respectively.

We examined the age-adjusted OR and 95% CI for the association of the MTR genotype with the risk of developing colorectal cancer in the PHS and HPFS separately and combined. We conducted unconditional logistic regression analysis because of the combined study population. Using conditional logistic analysis yielded similar results when the matched case-control subgroup was examined. Prospective data on blood nutrients levels were availible only for the PHS colorectal cancer cases and controls. The age-adjusted geometric mean of plasma folate (because of the skewed distribution), mean vitamin B₁₂, and tHcy concentration within strata of the MTR and MTHFR genotypes by case-control status were calculated by analysis of covariance. In the PHS, we assessed the age-adjusted ORs for the joint effect of the MTR and MTHFR genotypes and status of plasma folate, vitamin B₁₂, and tHcy (categorized into two groups based on control tertile distribution, lower one-third versus upper two-thirds for folate and vitamin B₁₂, lower two-thirds versus upper one-third for tHcy) using an indicator variable for each category in logistic regression models. Because alcohol consumption was measured prospectively in both the PHS and the HPFS, we assessed the joint effect of the MTR genotype and alcohol consumption (<1 drink/day, and ≥1 drink/day) within the PHS as well as the PHS and the HPFS combined. We compared the log likelihood statistics of the main effect model with the joint effect model to assess interaction. Because of small numbers in some of the stratified analyses, we used an exact statistic method (LogXact; Ref. 15) and obtained virtually identical results; we, therefore, present all of the results from unconditional logistic regression analyses. All of the Ps are two-sided, and all of the analyses were done using SAS (16).

The overall frequency of the homozygous variant gly/gly genotype was 5% among controls and 3% among cases (Table 1). The gly allele frequencies were 0.19 among controls and 0.17 among cases, not materially different from the 0.15 reported by Leclerc et al. (11) and the 0.16 reported by van der Put et al. (17). The genotype and allele frequencies were similar in the two cohorts. Overall, there was a nonsignificant inverse association of the gly/gly genotype with risk of colorectal cancer (OR, 0.59; 95% CI, 0.27–1.27) compared with the asp/asp genotype (Table 1).

Plasma folate, tHcy, and B₁₂ were measured only among the PHS participants. There were no apparent associations between the MTR homozygous variant gly/gly genotype and plasma levels of folate and tHcy (Table 2). Similar to our previous observation using the microbiological method for folate measurement (13), the MTHFR val/val genotype was significantly associated with lower plasma folate levels determined by the radioassay method, among both cases and controls (Table 2). The observation that the MTHFR homozygous variant genotype val/val was significantly associated with lower plasma folate levels is also consistent with findings from a separate case-control study of myocardial infarction nested in the PHS (18). In the present study, we also measured plasma vitamin B₁₂ and tHcy levels and observed significantly higher tHcy levels among cases with the MTHFR val/val genotype (Table 2). Although men with the MTR homozygous variant gly/gly genotype had higher levels of vitamin B₁₂ and men with the MTHFR val/val genotype had lower levels, none of these differences were statistically significant.

We examined the association of the MTR genotype and the risk of colorectal cancer according to plasma levels of folate, vitamin B₁₂, and tHcy in comparison with the MTHFR genotype. Because of the low prevalence of the gly/gly genotype, we categorized plasma folate, vitamin B₁₂, and tHcy into two groups based on the distribution among controls. For the MTHFR genotype, as we previously reported (13), there was a significant 70% decrease (OR, 0.29; 95% CI, 0.12–0.73) in risk among men with the val/val genotype and with plasma folate levels in the upper two tertiles of the control distribution compared with those with the ala/ala or val/ala genotype in the same folate category (Table 3). In the present analysis, we found that men with the homozygous val/val genotype and plasma tHcy levels in the lower two tertiles also had a significant 75% decrease (OR, 0.25; 95% CI, 0.08–0.74) in risk (Table 3). For the MTR genotype, we observed a nonsignificant 50% decrease (OR, 0.51; 95% CI, 0.14–1.90) in risk among men with the gly/gly genotype and with plasma folate levels in the upper two tertiles compared with those with the gly/asp or asp/asp genotype in the same folate category (Table 3). However, there was no apparent interaction between the MTR genotype and plasma tHcy (Table 3) on the risk of colorectal cancer. No apparent interactions were observed between vitamin B₁₂ and both genotypes (data not shown). Adjustment for body mass index, alcohol intake, multivitamin intake, and aspirin assignment did not change the results.

Alcohol consumption was assessed at baseline in both the PHS (1982) and the HPFS (1986). We examined the alcohol-genotype interaction within the PHS cohort as well as combined with the HPFS participants. Similar to our previous observation of the interaction between the MTHFR genotype and alcohol intake on risk (12, 13), we found a significant interaction between the MTR genotype and alcohol intake (P_interaction < 0.01) in the PHS. Among men with alcohol intake less than 1 drink/day, those with the gly/gly genotype had an OR of 0.10 (95% CI, 0.01–0.81), those with the gly/asp genotype had an OR of 0.70 (95% CI, 0.44–1.12) compared with those with the asp/asp genotype. Compared with the same reference group, among men with alcohol intake one drink/day or more, those with the gly/gly genotype had an OR of 3.79 (95% CI, 0.71–20.22), those with the gly/asp genotype had an OR of 1.14 (95% CI, 0.61–2.13). The combined PHS and HPFS data showed a similar association (P_interaction = 0.04; Table 4). Excluding men with MTHFR val/val genotype (n = 50) or further adjusting for body mass index, multivitamin intake, and aspirin assignment yielded similar results.

Although the 2756A→G polymorphism of MTR was first identified among patients with a deficiency of methionine synthase and among normal controls, the biological impact of this polymorphism is unknown (11). This A→G substitution at bp 2756 causes a substitution of glycine for aspartic acid (D919G). D919 corresponds to Q893 in the cobalamin-dependent Escherichia coli methionine synthase. In this highly homologous bacterial enzyme, this residue is at the penultimate position in a long helix that leads out of the cobalamin domain to the SAM-binding domain (19). It has been postulated that the glycine residue, a strong helix breaker compared with aspartic acid, could affect the secondary structure of the protein and, therefore, have functional consequences, perhaps leading to altered levels of vitamin B₁₂, folate, or tHcy (17). However, our observation that the MTR polymorphism was not associated with plasma levels of folate, vitamin B₁₂, or tHcy suggests that this aspartic acid-to-glycine change may not significantly deteriorate methionine synthase activity. In a recent Dutch study of patients and mothers of children with neural tube defects, patients with arterial disease, and population-based controls, the MTR polymorphism was also not associated with plasma tHcy levels (17). One alternative explanation is that homocysteine may be remethylated to methionine through an alternative pathway by betaine-homocysteine methyltransferase (Fig. 1;20). Although the main physiological role of that enzyme is to catabolize excess betaine (21), it also participates in regulating tissue levels of methionine and in removing excess homocysteine during stress (22). Under normal conditions, the activity of this enzyme increases substantially as a result of (a) inadequate methionine intake (23); (b) inactivation of methionine synthase by nitrous oxide (24); (c) impaired methionine synthase due to ethanol administration (25); or possibly (d) being in the presence of the variant MTR genotype.

Overall, we found that the MTR gly/gly genotype was associated with a nonsignificant 40% decrease in risk of colorectal cancer, in the same direction as we previous reported from the same participants for the MTHFR polymorphism (12, 13). The decreased risk was consistent among the PHS and the HPFS participants analyzed separately, but the CIs were wide, reflecting the small number of participants carrying the gly/gly genotype. Decreased activity of MTR or MTHFR would be expected to lower SAM levels leading to decreased DNA methylation. DNA hypomethylation has been suggested to suppress tumor growth (26, 27). Alternatively, because the activities of MTHFR and methionine synthase were significantly increased in tumor-bearing animals because of increased methylation reactions for tumor growth (28, 29, 30, 31, 32), these polymorphisms may inhibit the excessive use of methionine by tumor cells. This inhibition probably acts on a late stage of tumorigenesis because neither the MTHFR nor the MTR polymorphisms are associated with risk of colorectal adenoma, the immediate precursor of colorectal cancer (33).

Because plasma folate levels reflect both genetic and dietary variation, and ethanol can interfere with folate and methyl group metabolism as well as methionine synthase activity, moderate enzymatic changes due to genetic polymorphisms may behave differently according to the intake of alcohol (12, 13). The interactions between the MTR genotype and alcohol intake are consistent with what we observed for the MTHFR genotype. The increased risk for colorectal cancer conferred by high alcohol intake may overcome the protective effect of these polymorphisms (12, 13). Also, the influence of alcohol on cancer risk may differ among individuals with different genetic susceptibility. Among 10 cases and 21 controls who had the MTR variant genotype gly/gly in our analysis, men reporting one or more drinks/day had about a 10-fold higher risk than those who drank less. Although we observed similar results within the two cohorts as well as in the combined analysis, our study has limited statistical power for both main effect and stratified analyses because of the low prevalence of the variant MTR genotype, which is reflected in the wide CIs. The possible association of this MTR genotype with the risk of colorectal cancer, especially among individuals with low alcohol intake, merits further study in larger populations.

We thank the participants of the Physicians’ Health Study for their cooperation and participation. The authors are also indebted to Kathryn Starzyk, Rachel Adams, Xiaoyang Liu and Stefanie Parker for their expert and unfailing assistance. We are grateful to Dr. Klaus Lindpaintner who prepared the DNA, and to Nelly Sabbaghian and Marie Nadeau for technical assistance.