Association of Homocysteine, Methionine, and MTHFR 677C>T Polymorphism With Rate of Cardiovascular Multimorbidity Development in Older Adults in Sweden

Key Points Question Are serum concentrations of homocysteine and methionine associated with the rate of cardiovascular multimorbidity development in older adults? Findings In this cohort study of 1969 older adults in Sweden, high homocysteine levels, low methionine levels, and a low methionine to homocysteine ratio were associated with faster development of cardiovascular multimorbidity. The MTHFR 677C>T polymorphism was further associated with accelerated cardiovascular multimorbidity development in participants with a low methionine concentration. Meaning The findings of this study suggest that these biomarkers may have a potentially meaningful role of in the pathogenesis of cardiovascular aging, possibly through impaired methylation activity.


Introduction
Most age-related cardiovascular (CV) diseases share common underlying biological mechanisms.
Telomere attrition, 1 epigenetic modifications, 2 defects in autophagy or mitophagy, 3 and cell senescence 4 eventually contribute to a pro-inflammatory environment that is common in different disorders, such as atherosclerosis, cardiomyopathies, heart failure, ischemic heart disease, and stroke. The rate at which multiple CV diseases, ie, CV multimorbidity, accumulate is a marker of CV aging. 5 Thus, the identification of cellular and molecular biomarkers of accelerated CV multimorbidity development may be helpful for secondary prevention and prognostic counselling.
Homocysteine, an amino acid generated via demethylation of dietary methionine, is associated with atherosclerosis and its complications, such as myocardial infarction and stroke. 6,7 In fact, high concentrations of total homocysteine (tHcy) can lead to endothelial cell damage, impaired vascular compliance, and alterations of hemostasis. 8 Besides an unhealthy lifestyle, poor diet, impaired renal function, and intake of certain drugs, 9 moderate increases in tHcy have been associated with a 677C>T variant in the gene coding for methylenetetrahydrofolate reductase (MTHFR; OMIM 607093). 10 Methionine (Met), obtained both from diet and from the remethylation pathway of tHcy metabolism, is an essential amino acid functioning as a precursor for cysteine and glutathione, 2 major antioxidants. Studies in mice suggest that low Met availability affects the development of atherosclerotic lesions and the regulation of inflammatory response and hepatic cholesterol metabolism, 11,12 but its role in CV diseases is unclear. One study found an association between low blood Met levels and incident myocardial infarction 13 ; another study corroborated this association prospectively but only in subjects with high low-density lipoprotein (LDL) cholesterol levels 14 ; a third study showed no such association. 15 A positive association between low Met concentration and venous thrombosis risk has also been reported. 16 High tHcy and low Met blood concentrations could have antagonistic pleiotropic effects on CV aging, not only via atherosclerosis, the most studied mechanism, but also through other pathways, such as lipid metabolism, DNA synthesis and repair, oxidative damage, development of inflammation, and telomere attrition. 17 Therefore, it is plausible that tHcy and Met may play a role in multiple physiologic processes underlying CV aging, including left ventricular wall thickening, increased left atrial size, and vascular intimal thickening, stiffening, and fibrosis. 18 To our knowledge, no prior study has examined the association of tHcy and Met with multiple CV diseases as a group or with the rate at which these disorders accumulate.
We aimed to investigate the association of serum concentrations of tHcy and Met with the rate of CV multimorbidity development in older adults. We further intended to explore the role of MTHFR 677C>T polymorphism in this association.

Data Collection
This research is based on data from the Swedish National Study on Aging and Care in Kungsholmen (SNAC-K). This is a community-based longitudinal study of randomly selected individuals aged 60 years or older living at home or in institutions in the Kungsholmen district of Stockholm between 2001 and 2004. Of 4590 eligible participants, 3363 (73.3%) participated in the baseline examination. Participants underwent extensive clinical examinations, interviews, and assessments by physicians, nurses, and psychologists. Data on sociodemographic characteristics, medical history, drug use, laboratory test results, and cognitive function were collected according to structured protocols. Data on medical history and vital status were also obtained by linking SNAC-K data with the National Patient Register and the Swedish Cause of Death Register. Since baseline, SNAC-K participants have been followed up regularly, every 6 years for the cohorts younger than 78 years and every 3 years for those aged 78 years or older. This study included data from baseline and the first 4 follow-ups (Figure 1). In the present study, we excluded 1051 individuals (31.3%) with at least 1 CV disease at baseline plus 343 individuals (10.2%) with missing data on the exposures, leaving 1969 participants (58.5%) in the study cohort. SNAC-K baseline and follow-up protocols have been approved by the Ethics Committee at Karolinska Institutet and the Regional Ethics Review Board in Stockholm, and written informed consent was obtained from participants or their next of kin for those with cognitive impairment.

Statistical Analysis
The main outcome variable was the rate of CV multimorbidity development during the 12-year follow-up period. Linear mixed models were used to estimate β coefficients and 95% CIs for the association between baseline concentrations of the biomarkers and the rate of CV multimorbidity development. To measure the association between the exposures of interest and the mean annual increase in the number of CV diseases, interactions between follow-up time and biomarker concentrations were included as fixed effects. 22 Dyslipidemia and CRP level are 2 biomarkers that strongly and independently predict systemic atherosclerosis. 24 Cardiovascular risk factors have also been included as potential confounding or mediating factors in previous studies. 14 In a sensitivity analysis, to evaluate whether the associations between baseline concentrations of the biomarkers and the rate of CV multimorbidity development were driven by specific chronic conditions either at baseline or at follow-ups, we reran the models after removing the 8 CV diseases, 1 at a time, from the original count of CV diseases. Analyses were performed using Stata version 15 (StataCorp). Statistical significance was set at P < .05, and all tests were 2-tailed.

Results
The study population consisted of 1969 individuals without any CV disease at baseline, of whom 1261 (64.0%) were women. The mean (SD) age of the study population was 70.  Table 1). We found a significantly higher concentration of tHcy and a lower    CT/TT: β = 0.035 per year; 95% CI, 0.019-0.052; P < .001) ( Table 3). For Met, a significant association was only seen in participants with a low concentration of Met and the CT/TT genotype  (β = 0.023 per year; 95% CI, 0.006-0.041; P = .009) ( Table 3; eFigure in the Supplement). These results were further corroborated by the positive association with the rate of CV multimorbidity development in participants with low Met:tHcy ratios and both the CC group (β = 0.039 per year; 95% CI, 0.021-0.057; P < .001) and the CT/TT group (β = 0.037 per year; 95% CI, 0.021-0.054; P < .001) ( Table 3).

Discussion
In this longitudinal population-based study of older adults, a higher concentration of tHcy, a lower concentration of Met, and a lower Met:tHcy ratio (suggesting impaired methylation activity) were all associated with an increased rate of CV multimorbidity development during a 12-year period. Genetic predisposition (ie, the MTHFR 677T allele) further contributed to accelerated CV multimorbidity development in participants with low Met concentrations. Results were independent of sociodemographic, lifestyle, and clinical factors and did not seem to be driven by the incidental diagnosis of a specific CV disease but rather with the global burden of CV multimorbidity. Even if the effect sizes were relatively small, the biological coherence, the concentration-response associations, and the statistical robustness of the findings point to a potentially meaningful role of these biomarkers in the detection of CV multimorbidity and possibly also its etiology.
The prevalence of CV multimorbidity is increasing rapidly in the older population, 25 and the combination of any of the conditions included in CV multimorbidity is associated with a multiplicative mortality risk. 26 Moreover, CV multimorbidity is highly disabling in older adults and is associated with faster cognitive decline, 27,28 thus affecting quality of life as well as the ability to live independently.
Therefore, knowledge of the rate of CV multimorbidity development could help to target individuals who would benefit most from primary and secondary prevention of CV disease. In addition, if specific biological markers can be identified to capture CV aging, measuring these could track the efficacy of interventions aimed precisely at slowing CV dysfunction and decline. This will undoubtedly be the next phase of improvement in human longevity, for which CV diseases are still the leading threat. 29 The pathophysiologic mechanisms underlying the increased risk of CV diseases in individuals with hyperhomocysteinemia have been comprehensively studied. 30,31 Increased tHcy concentrations have consistently been shown to be associated with increased risk of new CV events, 32 in line with our findings. Thus, in patients with CV diseases, tHcy concentrations could be used as a prognostic factor and for stricter surveillance of lifestyle and treatment. 9 Other nonatherogenic mechanisms may explain the nonselective association between tHcy and CV disease accumulation described in our study. Indeed, hyperhomocysteinemia, which has been described to induce both ventricular systolic and diastolic dysfunction, has been associated with heart failure even in the absence of ischemic heart disease. 33,34 The increased deposition of interstitial and perivascular collagen triggered by high tHcy concentrations may explain its effect on heart failure via nonatherosclerotic mechanisms. Moreover, telomere attrition in cardiomyocytes, potentially accelerated by higher tHcy blood concentrations, can promote faster aging of myocardium, inducing pump insufficiency and chambers dilation. 1,35,36 Increased tHcy concentrations have been also associated with impaired autophagy of both cardiomyocytes and smooth vascular cells, leading to the accumulation of senescent cells and consequently organ dysfunction. 37,38 Much less is known regarding the role and pathophysiology of Met. A possible pathogenic pathway may involve glycine N-methyltransferase (GNMT), the most abundant liver methyltransferase regulating the availability of the biologic methyl donor, S-adenosylmethionine.
Met deficiency leads to reduced GNMT flux, 39 which is linked to hepatic lipid accumulation, hyperlipidemia, and the deposition of oxidized LDL in vascular walls, all of which have been associated with increased risk of atherosclerosis. 11,12,40 Another possible pathway is through the role of Met in methylation reactions. Met deficiency can induce site-specific hypomethylation, 41 leading to dysfunctional epigenetic modifications described recently in the expression of proprotein convertase subtilisin/kexin type 9, a serine protease involved in the degradation of both hepatic and extrahepatic LDL receptors, thereby increasing circulating LDL cholesterol concentrations. 42 Indeed, higher concentrations of Met and an improved methylation activity, as indicated by the Met:tHcy ratio, were both associated with a decreased rate of CV disease accumulation in our study, which is consistent with findings from animal experimental studies 11,12 and a 2019 study examining the risk of dementia development and structural brain changes. 21 Finally, Met deficiency may lead to deficiency of glutathione, a major intracellular antioxidant, which has been linked to metabolic and CV diseases. 43 The MTHFR 677C>T polymorphism leads to a modest increase in tHcy. Individuals with the homozygous TT genotype have approximately 0.34 mg/dL higher tHcy levels than those with the wild-type CC variant. 44,45 The difference depends mainly on folate status. 46 Lack of power may therefore be the reason why most studies, including ours, have failed to show any difference in the risk of CV diseases in people with this polymorphism, 32 but meta-analyses have revealed such an effect consistently, as first shown in 2002 by Klerk et al. 45 It has recently been hypothesized that the MTHFR 677C>T polymorphism could influence CV pathology through mechanisms that are independent of tHcy, such as via blood pressure disorders. 47 Interestingly, in our study, the association of low Met with the rate of CV multimorbidity development was restricted to the group with the CT/TT alleles of MTHFR, suggesting an interactive effect of both Met and genotype. Such a finding, together with the significant protective association of the Met:tHcy ratio in our models, suggests that we need to further examine methylation-derived modifications of gene expression in relation to CV multimorbidity.

Strengths and Limitations
The main strength of our study is the use of a longitudinal population-based study in older adults with detailed clinical characterization and available data on potential confounders. The identification of CV diseases was done through a clinically driven algorithm, integrating different sources of data.
Thus, the risk of misclassification and underdetection of the outcome was limited, as previously shown. 19 Individuals with CV diseases at baseline were excluded from our analyses, which minimized the risk of reverse causality.
Several limitations should be considered. Time-varying measurements for the biomarkers (ie, the marker may not be constant over time) could lead to biased associations due to regression dilution. It has been shown that failure to correct for regression dilution may underestimate the relative risks of disease by 50% after 10 years. 48 Moreover, blood measurements were performed under nonfasting conditions, which could have direct consequences on serum Met levels following a meal. 49 However, in previous studies, adjustment for fasting status did not alter the association

JAMA Network Open | Cardiology
Homocysteine, Methionine, and MTHFR 677C>T Polymorphism and Cardiovascular Multimorbidity between plasma Met concentrations and acute myocardial infarction 14 or other outcomes. 21 The low number of participants with the TT genotype forced us to explore its association in combination with the CT genotype, even if the latter had tHcy concentrations that were closer to the CC rather than the TT genotype. Moreover, we could not control for confounding from ancestry or technical bias in genotyping. Even after adjustment for potential confounders, the possibility of residual confounding cannot be discarded. Furthermore, although this is a prospective study, reverse causality should not be ruled out given that elevated tHcy is associated with risk factors often targeted for treatment in relation to CV health.

Conclusions
This study adds further epidemiologic evidence to the hypothesis that serum tHcy and Met concentrations and Met:tHcy ratio are important independent risk factors not only for the incidence of multiple CV diseases of diverse nature, but also for the rate at which these different CV diseases accumulate in older age. The interactive associations of Met concentrations and the MTHFR 677C>T polymorphism, together with the association found for the Met:tHcy ratio, point to the relevance of impaired methylation in the pathogenesis of CV aging but also call for further studies investigating alternative pathways.