Impact of the MTHFR C677T polymorphism on one-carbon metabolites: Evidence from a randomised trial of riboflavin supplementation
Abstract
Homozygosity for the C677T polymorphism in MTHFR (TT genotype) is associated with a 24e87% increased risk of hypertension. Blood pressure (BP) lowering was previously reported in adults with the TT genotype, in response to supplementation with the MTHFR cofactor, riboflavin. Whether the BP phenotype associated with the polymorphism is related to perturbed one-carbon metabolism is un- known. This study investigated one-carbon metabolites and their responsiveness to riboflavin in adults with the TT genotype. Plasma samples from adults (n 115) screened for the MTHFR genotype, who previously participated in RCTs to lower BP, were analysed for methionine, S-adenosylmethionine (SAM), S-adenosylhomocysteine (SAH), betaine, choline and cystathionine by liquid chromatography tandem mass spectrometry (LC-MS/MS). The one-carbon metabolite response to riboflavin (1.6 mg/d; n 24) or placebo (n 23) for 16 weeks in adults with the TT genotype was also investigated. Plasma SAM (74.7 ± 21.0 vs 85.2 ± 22.6 nmol/L, P ¼ 0.013) and SAM:SAH ratio (1.66 ± 0.55 vs 1.85 ± 0.51, P ¼ 0.043)
were lower and plasma homocysteine was higher (P ¼ 0.043) in TT, compared to CC individuals. In response to riboflavin, SAM (P ¼ 0.008) and cystathionine (P ¼ 0.045) concentrations increased, with no responses in other one-carbon metabolites observed. These findings confirm perturbed one-carbon metabolism in individuals with the MTHFR 677TT genotype, and for the first time demonstrate that SAM, and cystathionine, increase in response to riboflavin supplementation in this genotype group. The genotype-specific, one-carbon metabolite responses to riboflavin intervention observed could offer some insight into the role of this gene-nutrient interaction in blood pressure.
1. Introduction
Hypertension is a major modifiable risk factor for stroke and cardiovascular disease (CVD), and a leading cause of premature mortality worldwide, responsible for over 10 million deaths annually [1]. The pathophysiology of hypertension is complex, involving the interaction of genetics, environmental factors and physiological mechanisms [2]. Genome wide association studies (GWAS) have linked a number of genetic loci with hypertension [3,4], including a region near the gene encoding the folate metab- olising enzyme, methylenetetrahydrofolate reductase (MTHFR). The common MTHFR C677T polymorphism produces an enzyme with reduced activity [5] owing to lowered affinity for its riboflavin cofactor, (flavin adenine dinucleotide, FAD) [6]. The homozygous MTHFR 677TT genotype affects 2e32% of populations worldwide [7] and meta-analyses have estimated that the variant TT genotype is supply of SAM to methyltransferase reactions. This is achieved through the action of SAM as an allosteric inhibitor of MTHFR and an allosteric activator of CßS, thus controlling one-carbon flux and homocysteine concentrations [16].
Higher concentrations of SAM and SAH have been reported in TT relative to CC adults in some [17,18] but not all [19,20] studies. In observational analysis of 10,601 Norwegian adults elevated ho- mocysteine and decreased betaine were reported in TT compared to CC genotype groups, with no influence of genotype on other one- carbon metabolites [21e23]. Sub-optimal status of the B vitamins, folate, riboflavin, PLP and cobalamin, which act as nutritional co- factors for the key enzymes in the one-carbon pathway (Fig. 2), have been previously shown to result in elevated homocysteine in adults generally and particularly by MTHFR genotype [24,25]. The effect of intervention with one or a combination of these B vitamins has been shown to modulate homocysteine concentrations [26e28]; however, the effect on other one-carbon metabolites has not been widely investigated, and few studies have considered the effect of the MTHFR 677TT genotype.
Therefore, the aim of this study is to investigate the impact of the MTHFR C677T polymorphism on one-carbon metabolite status and the responsiveness of one-carbon metabolites to riboflavin supplementation (1.6 mg/day) in adults with the MTHFR 677TT genotype. The findings of this study could contribute to our un- derstanding of the mechanism underpinning the BP phenotype related to this gene-nutrient interaction.
2. Materials and methods
associated with 24e87% increased risk of hypertension and increased risk for CVD by up to 40% [8]. Previous studies conducted at this Centre have demonstrated that BP is highly responsive to riboflavin supplementation, with evidence that systolic BP can be lowered by between 6 and 14 mmHg in individuals with the TT genotype [9e11]. This gene-nutrient interaction thus offers a novel, nutritional approach for BP management among adults with the C677T variant in MTHFR, although the underlying mechanism re- mains unexplained. It is possible the phenotype of elevated BP and its response to riboflavin may be owing to perturbations in one- carbon metabolism in affected individuals; however, this mecha- nism has not been previously investigated.
In one-carbon metabolism, FAD-dependent MTHFR generates 5-methyltetrahydrofolate (5-MTHF), which is involved in the remethylation of homocysteine to methionine, the pre-cursor to S-adenosylmethionine (SAM; Fig. 1). As the principal methyl donor, SAM transfers methyl groups to over 100 methyl- transferases involved in numerous biochemical pathways including DNA methylation, histone modification and neuro- transmitters [12]. This transfer, in turn, leads to the formation of S-adenosylhomocysteine (SAH), which is subsequently metab- olised to homocysteine. DNA methylation, an epigenetic process involved in gene transcription and expression, has been impli- cated in a number of disease states across the life-cycle, including CVD [13]. The ratio of SAM:SAH has been sometimes used as a marker of methylation potential, although the validity of this indicator requires confirmation [14]. Choline and betaine can also serve as alternative methyl donors in homocysteine remethyla- tion as part of the betaine-homocysteine methyltransferase (BHMT) pathway [15]. Homocysteine can be removed through irreversible condensation with serine to cystathionine via the action of cystathionine ß-synthase (CßS), in the pyridoxal-50-phosphate (PLP)-dependent transsulfuration pathway. Regula- tion of the methylation cycle is essential to ensure sufficient.
2.1. Subjects and samples
Plasma samples from participants who had previously partic- ipated in studies at the Nutrition Innovation Centre for Food and Health (NICHE), Ulster University, and had been screened for the MTHFR 677TT genotype were accessed for the current study. In all cases, participants provided informed, written consent and agreed for samples to be used in subsequent studies. Samples were accessed from the GENOVIT study (ORECNI ref 08/NIR03/40) [9], the GENOVIT follow-up study (ORECNI ref 08/NIR03/40) [10] and the RIBOGENE study (ORECNI/12/0338). Ethical approval for the analysis reported in the current study was granted by Ulster University Research Ethics Committee (FCBMS-18-040). All three studies had identical inclusion (pre-screened for MTHFR C677T polymorphism) and exclusion (history of gastrointestinal, hepatic or renal disease, consumers of B vitamin supplements, use of medication known to interfere with B vitamin metabolism) criteria. Clinic BP was measured in accordance with guidelines from the National Institute of Care and Excellence (NICE) [29]. In brief, after 10 min at rest, BP was measured in the reference arm, i.e. the arm with the highest BP, with the participant in the seated
position. Mean BP was calculated as the average of two BP read- ings within 5 mmHg, with a maximum of six readings obtained. Anthropometry, health and lifestyle information and blood sam- ples were collected according to appropriate standardised oper- ating procedures as part of each study, described in detail elsewhere [9,10].
The analysis for the current study consisted of both an obser- vational and an intervention phase. In the observational phase, participants with the TT genotype were age-matched with a similar number of individuals with the CC genotype and compared for general characteristics and one-carbon metabolite biomarker sta- tus. In the intervention phase, biomarker status of methionine, SAM, SAH, SAM:SAH ratio, betaine, choline and cystathionine in response to intervention with riboflavin (n 24) and placebo (n 23) were investigated (Fig. 2).
Fig. 1. Overview of one-carbon metabolism.
Fig. 2. Flow diagram of study population. 1CC (wild type) and TT (homozygous) genotypes for the MTHFR C677T polymorphism.
2.2. Blood sampling
Venipuncture of a vein in the antecubital fossa was conducted by a trained phlebotomist with the participant in a non-fasting state. A 25 ml blood sample was obtained into two EDTA vacu- tainers (9 ml and 4 ml) and two serum vacutainers (8 ml and 4 ml). All tubes, apart from the 4 ml EDTA tube, were placed immediately on ice and centrifuged at 3000 rpm for 15 min at 4◦ Celsius, within 30 min of the venipuncture. Plasma, serum and buffy coat were removed at this stage. The erythrocytes in the 9 ml EDTA tube were thrice washed with phosphate buffered saline and these washed red cells were used for erythrocyte glutathione reductase activation coefficient (EGRac) analysis. The 4 ml EDTA tube was rolled for 30 min, and 50 ml was added to 450 ml of 1% ascorbic acid solution (1 in 10 dilution), from which red blood cell folate was determined. All fractions were labelled and stored at 80◦ Celsius in alarm- controlled freezers with batch analysis of biomarkers conducted at the end of the study. The samples did not undergo any freeze- thaw cycles between initial storage and analysis.
2.3. B vitamin biomarker analysis
Riboflavin status was determined at Ulster University using the erythrocyte glutathione reductase activation coefficient (EGRac) assay, which measures the enzyme activity of glutathione reduc- tase before and after in vitro reactivation with its prosthetic group FAD, as described elsewhere [10]. EGRac is calculated as the ratio of FAD-stimulated to unstimulated enzyme activity, with values < 1.3 indicating optimal riboflavin status, 1.3e1.4 suboptimal status and >1.4 signifying deficiency. Red blood cell folate concentrations, a long-term biomarker of folate status was measured by microbio- logical assay using Lactobacillus casei, as described by Molloy & Scott [30]. Plasma PLP, as a marker of vitamin B6 status, was ana- lysed by high performance liquid chromatography (HPLC) [31]. Plasma homocysteine was analysed by fluorescence polarisation immunoassay for plasma homocysteine [32].
2.4. Metabolite analysis
One-carbon metabolites, apart from homocysteine, were ana- lysed at the Center of Metabolomics, Baylor Scott & White Research Institute (Dallas, Texas 75226). Determination of methionine, SAM, SAH, betaine, choline and cystathionine in plasma was performed by HPLC coupled with electrospray positive ionization tandem mass spectrometry (HPLC-ESI-MS/MS) using a method previously described with some minor modification [33]. In brief, 20 ml of plasma was added to 180 ml of isotope internal standards and loaded into a microtiter plate before being centrifuged for 60 min prior to analysis. The calibration curve for SAM and SAH was 25- 400-nmol/L, for methionine, betaine and choline; 3.1e50 nmol/L and for cystathionine: 125e2000 nmol/L. Two levels of quality control samples were used to monitor within and between day precision of the method. In all cases, the coefficient of variation (cv) was less than 15% for all metabolites.
2.5. Statistical analysis
Statistical analyses were performed using Statistical Package for Social Sciences (SPSS; version 25.0; SPSS UK Ltd, Chertsey, UK). Normality tests were carried out on the data and data not normally distributed were log transformed before analysis was conducted. Differences in general characteristics and one-carbon metabolite status between genotype groups (observational cohort) were determined using independent samples t-test. Chi square test was used for comparison between categorical variables. To determine the response to intervention, within-between repeated-measures ANCOVA was used, controlling for baseline EGRac. The between- participant factor was the intervention group (placebo compared with riboflavin), and the within-participant factor was time (before compared with after intervention). Results are presented as mean (SD), unless otherwise stated. P < 0.05 was considered significant in all analysis carried out. Network analysis was performed with vis- ualisation of the networks in a circular layout in corrplot and qgraph packages from R (version 3.3.0; R Core Team 2016, Vienna, Austria; www.R-project.org). 3. Results Available plasma samples and data from adults (n 115) screened for the MTHFR genotype, and who previously participated in trials to lower BP were accessed. In the observational cohort, there were no significant differences in general characteristics between MTHFR genotype groups (Table 1). EGRac, the functional indicator of riboflavin status, was similar across the groups. PLP, serum and red blood cell folate concentrations were significantly lower in those with the TT compared to CC genotype. As previously reported, both systolic and diastolic BP were significantly elevated in the TT rela- tive to CC genotype groups (mean difference 16.6 ± 3.4 mmHg, P < 0.001; 9.0 ± 13.5 mmHg, P < 0.001, respectively), and those with the TT genotype were more likely to be classed as hypertensive according to current NICE guidelines [29]. There was no difference in use of anti-hypertensive medications between groups (75% of CC and 83% of TT genotype, P 0.308). In relation to one-carbon metabolites, elevated homocysteine (10.4 ± 3.0 vs 9.3 ± 2.5 mmol/L P 0.043), lower SAM concentra- tions (74.7 ± 21.0 vs 85.2 ± 22.6 nmol/L P 0.013) and lower SAM:SAH ratio (1.66 ± 0.55 vs 1.85 ± 0.51, P 0.043) was observed in the TT compared to the CC genotype (Table 2). No differences were observed for methionine, SAH, betaine, choline or cys- tathionine by genotype group. Network analysis showed that the nature and strength of interrelationships of metabolites and B vi- tamins within one-carbon metabolism were influenced by MTHFR genotype (Fig. 3). As previously reported [9e11], significant decreases were observed in both systolic ( 14.0 ± 15.3 mmHg, P 0.030) and diastolic BP ( 8.2 ± 11.1 mmHg, P 0.013) in response to riboflavin supplementation which resulted in a significant decrease in EGRac ( 0.15 ± 0.16, P < 0.001), indicating improved riboflavin status, in those with the MTHFR 677TT genotype (Fig. 4). No change in red blood cell folate was observed (data not shown). Response of one-carbon metabolites to riboflavin intervention among individuals with the TT genotype in MTHFR is presented in Table 3. Plasma homocysteine decreased by 0.5 ± 1.7 mmol/L in the riboflavin group, albeit an effect that was non-significant compared to the placebo group. Mean plasma SAM concentration increased significantly in response to riboflavin supplementation by 19.5 ± 20.6 (P 0.021), where the nature of this effect was only strengthened when adjusted for baseline riboflavin status (P 0.008). Plasma cystathionine concentrations increased by 50.7 ± 92.5 nmol/L (P 0.021), in response to riboflavin supple- mentation. No other metabolites were affected by riboflavin intervention. 4. Discussion The findings of the current study report for the first time that plasma concentrations of the one-carbon metabolites, SAM and cystathionine, increase significantly in response to riboflavin sup- plementation in individuals with the MTHFR C677T polymorphism. Coincident with this finding, we also observed lower concentrations of plasma SAM in TT compared to CC genotype adults. Indeed, after intervention with riboflavin in adults with the TT genotype, SAM concentrations increased to levels similar to those observed in adults with the CC genotype at baseline. The changes in plasma SAM and cystathionine concentrations in response to riboflavin intervention are consistent with the geno- type specific BP response previously reported in response to sup- plementation with riboflavin, raising the possibility that the effect of this gene-nutrient interaction on BP may be influenced by the cofactor requirements. To our knowledge, this is the first study to investigate the effect of intervention with riboflavin on SAM concentrations in adults with the MTHFR 677TT genotype. Previous investigations have, however, considered the effect of folic acid supplementation on one-carbon metabolites. In a small sub-group of MTHFR 677TT patients from the Verona Heart Study Project, 5 mg/d folic acid resulted in significant increases in SAM by 13 nmol/L and SAM:SAH ratio by 3.3, in addition to the expected reductions in homocysteine following 8 weeks of treatment [35]. The extent of the response in SAM of almost 20 nmol/L observed in the current study in response to riboflavin is even greater than these previous observations [35]. As the principal methyl donor, SAM-dependent methylation regu- lates fundamental biological processes including nuclear tran- scription, cell signalling, mRNA translation and DNA synthesis [12] and altered DNA methylation has previously been observed in TT relative to CC adults [36]. Supplementing with B vitamins to regulate concentrations of SAM in adults with perturbed one- carbon metabolism owing to genetic variants, could thus poten- tially have important implications for CVD health outcomes. Pre- vious studies have linked methylation with hypertension; however, no one has considered the C677T polymorphism in MTHFR or its relationship with SAM or BP. This is the first study to show that riboflavin supplementation in those with the mutant genotype affects concentrations of SAM and thus, possibly methylation po- tential. A recent meta-analysis reported lower global methylation levels with higher systolic BP, diastolic BP and hypertension [37]. The same meta-analysis also reported lower methylation levels of a number of candidate genes with increased BP; however, MTHFR has not yet been considered to any great extent. While hypertension was not considered in a meta-analysis by Amenyah et al., lower global methylation was reported in those with the TT genotype in combination with low folate status [38]. Fig. 3. Network analysis to show interrelationships within one-carbon metabolism by MTHFR genotype group: CC, panel a; TT, panel b. Positive and inverse associations indicated by green and red edges, respectively. Strength of association indicated by edge thickness. Abbreviations: Smk, smoking; SBP, systolic blood pressure; BMI, body mass index; Met, methionine; HCY, homocysteine; Cys, cystathionine; SAM, S-adenosylmethionine; SAH, S- adenosylhomocysteine; SSr, SAM:SAH ratio; Cho, choline; Bet, betaine; B2, riboflavin; PLP, Pyridoxal-50 -phosphate; RCF, red blood cell folate. Fig. 4. Change in riboflavin biomarker (panel a), systolic BP (panel b), and diastolic BP (panel c) in response to supplementation with placebo or riboflavin (1.6 mg/d) for 16 weeks in adults with the MTHFR 677TT genotype. For riboflavin biomarker, a decrease in EGRac indicates an improvement in riboflavin status. Choline and 5-MTHF are considered fungible methyl group sources in one-carbon metabolism, and methyl groups from choline can also facilitate homocysteine remethylation via the BHMT pathway [15]. In a study of folate-deficient males with the TT ge- notype, intervention with 2200 mg/day choline over 12 weeks was found to significantly increase plasma SAM concentrations compared to lower choline doses of 300e500 mg which were associated with a decreased SAM concentration [18]. Whilst these studies investigated one-carbon nutrients, BP was not considered. To date, research examining the effect of supplementation with B vitamins on one-carbon metabolites in adults with the MTHFR 677TT genotype has predominantly focused on the established phenotype of elevated homocysteine. Numerous meta-analyses demonstrating the responsiveness of homocysteine to supple- mentation with a combination of B vitamins have been published [24,25,39]; however, other one-carbon metabolites apart from homocysteine have received little attention in studies of this na- ture. Studies at our Centre have previously reported that riboflavin supplementation lowers homocysteine in TT, but not CC, in- dividuals, although, the response of other one-carbon metabolies were not considered [9,26]. Plasma cystathionine concentrations significantly increased in response to riboflavin supplementation in the current analysis. It is possible that increased availability of SAM, an allosteric regulator of cystathionine ß-synthase (CßS), in response to riboflavin may potentially have activated CßS, thereby increasing homocysteine elimination from the one-carbon pathway and generating cys- tathionine [16]. In addition, riboflavin administered at the same dose as the current study (1.6 mg/day) has previously been found to improve PLP status in older adults [40] and may thus augment the activity of PLP-dependent CßS. Consistent with earlier findings reported by Midttun and colleagues [23] lower PLP concentrations were observed in the current study in participants with the TT genotype. Those with the MTHFR 677TT genotype have reduced affinity for their riboflavin cofactor, FAD [6], thus are likely to have an increased requirement for riboflavin. Considering that cells appear to have a tendency to spare FAD at the expense of flavin mononucleotide (FMN) [41] it is possible that FMN-dependent pathways (such as the pathway required to convert vitamin B6 into active PLP) may be compromised in those with the mutant genotype, leading to reduced vitamin B6 metabolism and thus lower PLP concentrations. A paucity of evidence exists with respect to investigating the impact of MTHFR genotype on SAM and SAH concentrations. In a cohort of Mexican-American males, Shin et al., reported increased concentrations of SAH and decreased SAM:SAH ratio in those with the TT compared to the CC genotype [18]. Davis et al., observed elevated SAM in young females with the TT relative to CC genotype; however this was not significant [17]. This is in contrast with the findings of the current analysis, where decreased SAM was observed in the TT compared to CC genotype group. Increased trans- methylation reaction flux (i.e. conversion of SAM to SAH) has been found in females with the TT compared to the CC genotype [42]. A number of studies have found that the TT genotype is not a deter- minant of SAM or SAH [17,43] but folate status appears to be an important modulator of this effect [20]. Pertubations in one-carbon metabolism can impair the synthesis of SAM, and potentially lead to epigenetic alerations (specifically aberrant DNA methylation); correspondingly global DNA hypomethylation has been previously reported in individuals with the TTcompared to CC genotype [44,45]. The ratio of SAM:SAH has been proposed by some as an indicator of methylation potential, although confirmation of its validity remains to be established. Methylation regulation enzymes are differentially expressed in human tissues, leading to tissue-specific SAM and SAH regulation and therefore methylation capacity. Thus, systemic SAM:SAH ratio is not necessarily a meaningful indicator of methyl- ation potential in all tissues [14]. In the current analysis, lower SAM:SAH ratio was observed in the TT compared to CC genotype group, driven by the reduced SAM concentrations. However, these results are at odds with another study that reported the MTHFR genotype did not influence the ratio of SAM:SAH [43]. The observational results of the current analysis are broadly in agreement with those of the Norwegian Colorectal Cancer Pre- vention (NORCCAP) study, where differences in one-carbon metabolite status in individuals with the TT relative to the CC ge- notype were reported in 10,601 adults aged 50e64 years [21e23]. In agreement with our baseline analysis, these studies also reported the expected phenotype of elevated homocysteine, lower folate and lower PLP concentrations in the TT compared to the CC genotype. No MTHFR genotype effect was noted in relation to methionine, choline and cystathionine. One notable difference in the observed associations reported in the Norwegian cohort compared to the current cohort is betaine, where concentrations among Norwegians were significantly lower in those with the TT genotype compared to non-TT genotypes [22]. Betaine has been suggested as a preferential methyl donor in TT males relative to CC males [46]; however, in our analysis no genotype effect was noted with respect to betaine. 4.1. Strengths and limitations This is the first study to consider the effect of the MTHFR cofactor, riboflavin, on one-carbon metabolites in adults stratified by MTHFR genotype. Samples from a number of carefully conducted randomised controlled trials utilising identical dose, duration and study protocols were accessed. The one-carbon metabolite analysis, which is known to pose analytical challenges, was conducted at a centre with considerable expertise in laboratory analysis of one-carbon metabolite biomarkers. Furthermore, EGRac is considered the gold standard method for measurement of long-term riboflavin status and this measure was available for all participants. One limitation of the current study is the relatively small sample size which may have limited the ability to detect small differences in certain metabolites either between genotypes or in response to riboflavin. Additional biomarker information, in particular 5-MTHF, which is generated by the MTHFR enzyme, might further add to our understanding of the role of this gene-nutrient interaction in BP regulation. The intervention could also be extended to those with MTHFR 677CC genotype. 5. Conclusion In conclusion, this study shows evidence of perturbed one- carbon metabolism in individuals with the MTHFR C677T poly- morphism, in particular reduced concentrations of the principal methyl donor, SAM. This study provides the first evidence that altered one-carbon flux may be alleviated through riboflavin sup- plementation in individuals with the C677T variant in MTHFR. The findings of this study may shed some light on the mechanism un- derpinning the elevated BP phenotype related to this gene-nutrient interaction, which, in turn could influence health outcomes in adult cohorts. Future studies investigating the effect of riboflavin and other B vitamins on one-carbon metabolite concentrations, are needed to further explore the potential mechanisms underlying the effect of this gene-nutrient interaction on BP among individuals with S-Adenosyl-L-homocysteine the MTHFR 677TT genotype.