Common Mutation in Methylenetetrahydrofolate Reductase
Correlation With Homocysteine Metabolism and Late-Onset Vascular Disease
Background Increased homocysteine levels are a risk factor for atherosclerosis and its sequelae. A common genetic mutation in methylenetetrahydrofolate reductase (MTHFR), an enzyme required for efficient homocysteine metabolism, creates a thermolabile enzyme with reduced activity. We determined the prevalence of this mutation in many subjects with and without vascular disease and related it to homocysteine and folate levels.
Methods and Results DNA from 247 older subjects with vascular disease and 594 healthy subjects without vascular disease (in three different control groups) was screened for the MTHFR 677 C-to-T mutation. Within each group, 9% to 17% of the subjects were homozygous for this mutation, and the mutant allele frequency was 31% to 39%. The genotype distributions, homozygote frequencies, and allele frequencies did not differ significantly between the study groups. In the vascular disease subjects, despite significantly lower folate levels in MTHFR homozygotes, there was no significant difference in homocysteine levels among the MTHFR genotype groups. The negative slope of the regression line relating homocysteine and folate was significantly steeper for those with a homozygous MTHFR mutation compared with those without this mutation.
Conclusions Although the thermolabile MTHFR mutation is very common, it does not appear to be a significant genetic risk factor for typical late-onset vascular disease. Because MTHFR homozygotes have increased homocysteine with low folate levels, this mutation may contribute to early-onset or familial vascular disease. The genotype dependence of the folate-homocysteine correlation further suggests that homozygotes for this mutation may have both an exaggerated hyperhomocysteinemic response to folic acid depletion and a better response to folic acid therapy.
Elevated levels of the thiol-containing amino acid homocysteine are an independent graded risk factor for the development of arteriosclerotic vascular disease.1 Increased homocysteine levels are commonly the result of a genetic defect in one of the enzymes promoting homocysteine metabolism. Thus, children with classic homocystinuria from an inherited deficiency in a homocysteine metabolic pathway often have severe vascular disease.2 One such pathway for homocysteine metabolism, the remethylation of homocysteine to methionine, requires the carbon donor 5-methyl tetrahydrofolate, formed by the action of the enzyme methylenetetrahydrofolate reductase (MTHFR). Hereditary deficiencies of MTHFR are thus a common cause of hyperhomocysteinemia. A thermolabile variant of MTHFR with decreased basal enzymatic activity has recently been found in a large proportion of both normal subjects (5%) and patients with vascular disease (17%).3 Carriers of this defective enzyme have higher homocysteine levels than control subjects and are at increased risk for premature coronary artery disease.3 4
The common thermolabile MTHFR variant results from a C-to-T point mutation at nucleotide 677 (changing Ala to Val), which significantly reduces the basal activity of the enzyme.5 The prevalence of this MTHFR mutation is extraordinarily high, with homozygotes representing 5% to 12% of several normal American, Canadian, and Dutch cohorts.5 6 7 8 Homozygosity for this common MTHFR mutation has also recently been shown to be associated with both higher homocysteine levels5 6 7 and an increased risk of cardiovascular disease7 in selected populations.
Because ≈10% of the population's risk for arterial vascular disease may be attributable to elevated homocysteine levels,1 the high prevalence of the MTHFR 677 C-to-T mutation may represent an important genetic risk factor for this common disease. The identification of these genetically at-risk individuals may be particularly relevant in light of the ability to lower homocysteine-associated vascular risk with folic acid therapy.1 To address this question, we present a genetic and biochemical analysis of the homocysteine metabolic pathway in a group of older patients with arterial vascular disease compared with several healthy control groups.
The 247 vascular disease subjects were recruited from the vascular surgery clinic at Oregon Health Sciences University (OHSU) and are enrolled in a prospective clinical trial of homocysteine and progression of atherosclerosis. Inclusion criteria for this study included age between 40 and 80 years (average, 68±11 years) and a history and documented vascular laboratory evidence of symptomatic atherosclerotic cerebrovascular or peripheral vascular disease. Subjects with end-stage vascular disease were excluded. Other atherosclerotic risk factors in this group included hypertension (affecting 60%), diabetes (affecting 22%), smoking (33% current smokers, 56% ex-smokers), heart disease (affecting 48%), hypercholesterolemia (mean cholesterol level, 221±45 mg/dL), and male sex (63%).
Three different healthy control groups were also studied. The Canadian newborn samples (provided as DNA by Dr Marcus Grompe) were from 293 healthy neonates randomly collected in 1 month through a genetic screening program in Quebec.9 The blood donor group was composed of 170 healthy American Red Cross volunteer blood donors (50% men; average age, 46±14 years). The 133 adult hospital control subjects (average age, 54±13 years) were healthy volunteers recruited from hospital and laboratory personnel at OHSU.
Each study subject gave informed consent for the use of his or her blood, and the study was approved by our human studies Institutional Review Board.
MTHFR Genotyping by Polymerase Chain Reaction
DNA was extracted from fresh, frozen, or dried spots of peripheral blood by one of two different standard methods: a rapid boiling method with Chelex resin used for most of the specimens or a silica resin affinity chromatography method for the rare problematic specimen. For the Chelex method, we followed the whole-blood procedure of Walsh et al10 with some minor modifications.11 For the silica column method, we used the reagents and protocols provided within a commercially available DNA preparation kit (the QIAamp blood kit, Qiagen). The region surrounding the nucleotide 677 MTHFR mutation site was polymerase chain reaction–amplified (with the primers 5′-TGAAGGAGAAGGTGTCTGCGGGA-3′ and 5′-AGGACGGTGCGGTGAGAGTG-3′) as originally described.5 Because the nucleotide 677 mutation creates a restriction site for either Taq I or HinfI, the genotype was determined by digestion with either of these enzymes, followed by agarose gel electrophoresis and ethidium bromide staining.
Plasma levels of folate, pyridoxine (B6), and B12 were determined by use of commercially available radioimmunoassay kits. Plasma total homocysteine was determined by the Clinical Research Center at OHSU with a previously described high-performance liquid chromatography method.12
Genotype or allele frequency differences between the four study groups were assessed with χ2 tests. Differences in homocysteine or vitamin levels between the genotype groups were assessed by ANOVA (for three groups) and Student's t tests (for two groups) by use of log-transformed data (because all of these frequency distributions were highly skewed). The homocysteine-folate correlation was assessed by Spearman's rank correlation test (to test the overall correlation) and multiple regression analysis (to assess the difference between genotype groups). The correlation and multiple regression analyses used log-transformed folate values to best fit a linear regression model. Statistical tests were performed with the STATVIEW 4.0 (Abacus) software package.
Prevalence of the Thermolabile MTHFR Mutation
Blood specimens from the subjects within each of the four study groups were genotyped for the presence of the 677 C-to-T MTHFR mutation (Ala to Val). As Table 1⇓ shows, the mutation was common in each group, homozygotes representing 9% to 17% of these cohorts. This high mutant allele frequency (31% to 39%) agrees with several other recent reports.5 6 7 8 The genotype distributions did not vary significantly from those predicted by the Hardy-Weinberg equilibrium. There was no significant difference in the mutant allele frequency, the homozygote frequency, or the distribution of genotypes between the four study groups (Table 1⇓). Although subjects in each of the non–vascular disease groups were not age-matched with those in our older vascular disease group, the 36% to 39% mutant allele frequency and 15% to 17% homozygosity rate in these control groups do not even approach being lower than the rates in the vascular disease subjects (with an allele frequency of 31% and a homozygosity rate of 9%).
Homocysteine and Vitamin Levels
The mean homocysteine levels in the vascular disease subjects were 14.7±7.8 μmol/L for men and 13.0±5.3 μmol/L for women. These levels are significantly increased (P<.005) compared with either a large control group (n=3318) of healthy age-matched subjects (age, 65 to 67 years; reported means, 12.9±6.2 μmol/L for men and 11.6±4.5 μmol/L for women)13 or a small control group of young (mean age, 34 years) healthy laboratory personnel previously analyzed by one of us (Dr Taylor; reported means, 9.3±1.9 μmol/L for men and 7.9±2.3 μmol/L for women).12 In this vascular disease group, although the mean homocysteine levels were apparently higher for MTHFR homozygotes (17.2±13 μmol/L) than for wild-type (13±5.5 μmol/L) or heterozygous (13.6±6.9 μmol/L) subjects, these differences were not statistically significant (P=.16; see Table 2⇓).
Consistent with a recent report in subjects without vascular disease,6 there were significant differences in circulating folate levels between the three MTHFR genotype groups (P=.041; Table 2⇑) in our vascular disease subjects. Homozygotes for the MTHFR mutation had significantly lower folate levels (9.65±5.5 μmol/L) than those without the mutation (15.2±15 μmol/L; P=.019; Table 2⇑). In these homozygotes, the inability to efficiently generate 5-methyl-tetrahydrofolate (the enzymatic product of MTHFR) then leads to low circulating folate levels, because 5-methyltetrahydrofolate is the predominant folate species in plasma. In contrast, the MTHFR genotype groups did not significantly differ in plasma levels of vitamin B12 and pyridoxine (Table 2⇑), two other vitamins that regulate homocysteine metabolism.14 To assess whether the genotype-dependent folate levels might correlate with homocysteine levels, linear regression analysis confirmed a significant negative correlation between levels of folate and homocysteine in the vascular disease cohort (r=−.52, P<.0001). The strength of this negative correlation (as measured by the slope of the regression line) was significantly greater for MTHFR homozygotes compared with wild-type (P=.006) or heterozygous (P=.02; the Figure⇓) subjects. Homozygotes for this common MTHFR mutation may be not only more susceptible to the hyperhomocysteinemic effects of poor folate intake but also more responsive to subsequent folate therapy.
In several large cohorts of North Americans, we have shown an extremely high prevalence of an MTHFR genetic mutation associated with hyperhomocysteinemia. In our study subjects, both the frequency of the MTHFR 677 C-to-T mutant allele (31% to 39%) and the homozygote frequency (9% to 17%) are similar to those recently found in several smaller North American5 8 and northern European6 7 15 studies. Because the subjects in each of these studies are primarily white, the prevalence of this allele is likely quite high in this population. This high mutant allele frequency in white subjects may vary, however, in localized geographic areas as shown by the lower homozygote frequencies in northern European groups such as the Dutch (5%)6 7 and the Irish (6%)15 compared with Americans and Canadians (12% to 17%).5 8 Because the hyperhomocysteinemia resulting from this mutant MTHFR may be associated with both neural tube defects6 and vascular disease,7 it is not yet clear why the prevalence of this allele is so high. Historically, an enhanced thrombotic tendency may have conferred a slight survival advantage on MTHFR mutation carriers.
In contrast to the results of Kluijtmans et al7 in Dutch subjects, in our older cohort we found no evidence to suggest an association between the MTHFR mutation and late-onset vascular disease. But because the healthy study subjects for this comparison were not age-matched with those in the older vascular disease group, we cannot exclude the possibility of an MTHFR-dependent survival bias causing a reduced mutation prevalence among older versus younger subjects. For the 247 vascular disease subjects, the 95% confidence intervals for the mutant allele frequency and homozygosity rate were 27% to 36% and 6% to 13%, respectively (data from Table 1⇑). To reach statistical significance, an age-matched control group would have to have a mutant allele frequency below ≈27% and a homozygosity rate below ≈6%. None of the younger North American control groups analyzed so far come close to having these characteristics (Table 1⇑). Also, should the MTHFR mutation be associated with reduced overall survival, the prevalence of this mutation would drop with increasing age. To exclude this possibility, we have found no age-dependent changes in allele frequency or homozygosity rate in either the vascular disease subjects or the hospital or laboratory controls (data not shown). In addition, among a recently analyzed group of 178 subjects with venous thrombosis (average age, 45±16 years), we also found no age-dependent changes in MTHFR homozygosity or allele frequency rates (data not shown). In these venous thrombosis control subjects, the MTHFR homozygosity and allele frequency rates (11% and 36%, respectively) were no different for those above or below age 65 (P>.05) and also no different from those in any of the study groups previously analyzed (Table 1⇑). This lack of an age-dependent decline in MTHFR mutation prevalence further supports our conclusion that the thermolabile MTHFR mutation is not a genetic risk factor for late-onset vascular disease.
In the recent Dutch report of a positive association between this MTHFR mutation and cardiovascular disease,7 the MTHFR homozygosity rate in the vascular disease group (15%) was similar to that in our control groups (15% to 17%). The Dutch control group, in contrast, showed a much lower homozygote frequency (5%) than did any of our groups. In contrast, unlike the Dutch study, which excluded those vascular disease patients with known atherosclerotic risk factors (hyperlipidemia, hypertension, or diabetes), our elderly vascular disease cohort was composed predominantly of those with multiple vascular disease risk factors (60% with hypertension, 48% with heart disease, 22% with diabetes, and many with hypercholesterolemia [averaging 221±45 mg/dL]). Much of the resulting vascular disease in our study subjects may then be attributable to non-MTHFR factors. The vascular disease cases that could be attributable to the MTHFR mutation might then be lost in this high background. Also, because the vascular disease in those subjects with other inherited homocysteine metabolic defects (homocystinuria) generally occurs at a much younger age,2 our older vascular disease group (age, 68±11 years) might be biased toward those without MTHFR-associated disease who have survived to this age. Nevertheless, unlike previous studies,7 our data suggest that the common MTHFR 677 C-to-T mutation does not predispose to typical late-onset peripheral vascular disease. Given the association of this mutation with hyperhomocysteinemia, however, the vascular disease risk of younger mutation carriers and/or those with other common atherosclerotic sequelae may, in contrast, be quite significant.
In a group of vascular disease patients, we have found that the total plasma homocysteine level did not vary (P=.16) between MTHFR genotype groups. This nonsignificant statistical conclusion was based on an analysis of log-transformed homocysteine levels (because of the marked skewness of the frequency distributions). In comparison, in the non–vascular disease subjects reported by Kluijtmans et al7 and van der Put et al,6 the “significant” homocysteine differences between MTHFR homozygous and wild-type subjects were the result of statistical analyses on untransformed (presumably skewed) homocysteine data. When appropriate statistical techniques are used, we conclude that in those with late-onset vascular disease, the MTHFR genotype dependence of the total homocysteine level is not significant. Homocysteine levels are then likely to be dependent on a combination of several different genetic and nutritional factors in addition to MTHFR genotype status. As discussed below, the significant negative correlation between homocysteine and folate levels (particularly among those with a mutant MTHFR) suggests that folate may be one such key determinant of hyperhomocysteinemia.
Because 5-methyltetrahydrofolate, the enzymatic product of MTHFR, is the predominant form of folate in plasma,16 our finding of significantly reduced plasma folate levels in MTHFR homozygotes (with decreased enzymatic activity) can be explained. Others have reported a similar MTHFR genotype dependence for folate levels in subjects without vascular disease.6 The requirement for 5-methyltetrahydrofolate as the methyl group donor for the conversion of homocysteine to methionine might explain the marginal hyperhomocysteinemia in these homozygous individuals. The known association between homocysteine levels and cardiovascular disease1 may then make the identification of those genetically at risk for hyperhomocysteinemia of great public health relevance.
Consistent with previous reports,8 12 14 17 we have confirmed a significant inverse relationship between homocysteine and folate levels. High homocysteine levels are thus correlated with low folate levels and vice versa. In addition, we have now shown for the first time that the slope of the regression line relating homocysteine to folate is steeper in those homozygous for the MTHFR mutation compared with those with one or two wild-type alleles. The significant difference in homocysteine levels between MTHFR genotype groups in those with low folate (but not those with high folate) suggests that non–vascular disease subjects may have a similar genotype dependence for their folate-homocysteine correlation.8 Because nucleotide 677 may be within a folate binding region of MTHFR,5 the mutant enzyme (with perhaps a reduced folate affinity [Km]) may require significantly higher folate levels for efficient function. Nutritional deficits may then cause the folate level to fall below the Km of the mutant but not the wild-type enzyme. MTHFR homozygotes might then be particularly prone to the hyperhomocysteinemic and atherogenic effects of folate deprivation compared with those with a normal enzyme. The more marked folate sensitivity in those with either thermolabile MTHFR18 19 or coronary artery disease17 would support this hypothesis. Given the high population prevalence of both the MTHFR mutation and vascular disease, the identification of MTHFR homozygotes might be an effective means to predict that subset of hyperhomocysteinemic patients most likely to benefit from therapeutic folate.
This work was supported by an Oregon Medical Research Foundation grant to Dr Deloughery, an American Heart Association (Oregon affiliate) grant to Dr Press, and NIH grants to Dr McWilliams (K11-CA68297) and Dr Taylor (1R01-HL-45267-01A1). Thanks go to Gary Sexton for expert statistical help and to Chris Grant for MTHFR genotype advice.
Division of Laboratory Medicine, Department of Pathology (T.G.D., A.E., A.S., R.D.P.), Oregon Health Sciences University, Portland.
- Received June 26, 1996.
- Revision received August 14, 1996.
- Accepted August 31, 1996.
- Copyright © 1996 by American Heart Association
Mudd SH, Levy HL, Skovby F. Disorders of transsulfuration. In: Scriver CR, Beadet AL, Sly WS, Valle D, eds. The Metabolic Basis of Inherited Disease. New York, NY: McGraw-Hill Publishing Co; 1989:693-734.
Kang SS, Passen EL, Ruggie N, Wong PW, Sora H. Thermolabile defect of methylene-tetrahydrofolate reductase in coronary artery disease. Circulation. 1993;88:1463-1469.
Kluijtmans LAJ, Vandenheuvel LPWJ, Boers GHJ, Frosst P, Stevens EMB, Denheijer M, Trijbels FJM, Rozen R, Blom HJ. Molecular genetic analysis in mild hyperhomocysteinemia: a common mutation in the methylenetetrahydrofolate reductase gene is a genetic risk factor for cardiovascular disease. Am J Hum Genet. 1996;58:35-41.
Jacques PF, Bostom AG, Williams RR, Ellison RC, Eckfeldt JH, Selhub J, Rozen R. Relation between folate status, a common mutation in methylenetetrahydrofolate reductase, and plasma homocysteine concentrations. Circulation. 1996;93:7-9.
Jacobsen DW, Gatautis VJ, Green R, Robinson K, Savon SR, Secic M, Ji J, Otto JM, Taylor LM Jr. Rapid HPLC determination of total homocysteine and other thiols in serum and plasma: sex differences and correlation with cobalamin and folate concentrations in healthy subjects. Clin Chem. 1994;40:873-881.
Whitehead AS, Gallagher P, Mills JL, Kirke PN, Burke H, Molloy AM, Shields DC, Scott JM. A genetic defect in 5,10 methylenetetrahydrofolate reductase in neural tube defects. QJM. 1995;88:763-766.
Hillman RS. Vitamin B12, folic acid, and the treatment of megaloblastic anemias. In: Gilman AG, Rall TW, Nies AS, Taylor P, eds. The Pharmacological Basis of Therapeutics. New York, NY: Pergamon Press; 1990:294-310.
Hopkins PN, Wu LL, Wu J, Hunt SC, James BC, Vincent GM, Williams RR. Higher plasma homocyst(e)ine and increased susceptibility to adverse effects of low folate in early familial coronary artery disease. Arterioscler Thromb Vasc Biol. 1995;15:1314-1320.