Abstract
Redox balance and methylation are crucial to homeostasis and are linked by the methionine-homocysteine cycle. We examined whether differences in methylation potential, measured as plasma levels of S-adenosyl methionine (SAM) and S-adenosyl homocysteine (SAH), occur at baseline and during anti-oxidant therapy with the xanthine oxidase inhibitor allopurinol in patients with heart failure with reduced ejection fraction. We analyzed plasma samples collected at baseline and 24 weeks in the Xanthine Oxidase Inhibition for Hyperuricemic Heart Failure Patients (EXACT-HF) study, which randomized patients with heart failure with reduced ejection fraction to allopurinol or placebo. Associations between plasma levels of SAM, SAH, SAM/SAH ratio, and outcomes, including laboratory markers and clinical events, were assessed. Despite randomization, median SAM levels were significantly lower at baseline in the allopurinol group. SAH levels at 24 weeks, and change in SAM from baseline to week 24, were significantly higher in the group of patients randomized to allopurinol compared to the placebo group. A significant correlation was observed between change in SAH levels and change in plasma uric acid (baseline to 24-week changes) in the allopurinol group. There were no significant associations between levels of SAM, SAH, and SAM/SAH ratio and clinical outcomes. Our results demonstrate significant biological variability in SAM and SAH levels at baseline and during treatment with an anti-oxidant and suggest a potential mechanism for the lack of efficacy observed in trials of anti-oxidant therapy. These data also highlight the need to explore personalized therapy for heart failure.
Keywords: Methylation, redox balance, anti-oxidant therapy, heart failure
Introduction
Pre-clinical and clinical investigations have demonstrated that oxidant stress due to elevated levels of various reactive oxygen species plays a major role in the pathogenesis of cardiovascular disease including heart failure (HF).(1) Clinical trials utilizing predominantly naturally occurring vitamins and minerals have examined the benefit of anti-oxidant therapy in cardiovascular diseases, yielding disappointing results.(2,3) While several potential reasons for these negative results have been proposed, such as the type of anti-oxidant used, duration of therapy, and patient selection, the underlying biologic bases for the negative results of these trials and the role of redox imbalance in HF are not well understood.
In addition to the association of severe hyperhomocysteinemia due to inborn errors of metabolism with cardiovascular disease, more modest elevations of plasma homocysteine in the mild to moderate is also associated with cardiovascular disease including HF.(4,5) Studies from our laboratory and those of others have demonstrated that moderate hyperhomocysteinemia can lead to direct adverse effects on the myocardium mediated by pro-oxidant mechanisms.(6–8) However, trials utilizing B vitamins to lower homocysteine levels did not prevent adverse cardiovascular outcomes.(9) In fact, a post-hoc analysis of a clinical trial that examined the effect of high dose B vitamins in subjects with high plasma homocysteine levels demonstrated that B-vitamin therapy was associated with an increase in left atrial size, which is a surrogate marker of worsening left ventricular function.(10) While many factors, including duration of therapy, target population, effectiveness of primary vs. secondary prevention, and adverse effects of vitamin therapy, have been proposed as potential explanations for the neutral findings, these studies indicated that elevated plasma homocysteine may not be a target but an indicator of perturbation of the methionine-homocysteine cycle that regulates both redox balance and methylation reactions.(4)
Studies with selenium supplementation as an anti-oxidant have shown the complex biology that connects oxidant stress and methylation. Selenium is incorporated as the unique amino-acid selenocysteine into selenoproteins that include major anti-oxidant enzymes, such as glutathione peroxidases and thioredoxin reductases, and hence selenium supplementation increases anti-oxidant activity.(11) However, clinical trials of selenium supplementation, similar to other anti-oxidant interventions, have not demonstrated benefit; in fact, one study suggested an increased risk of diabetes, which has also been demonstrated in pre-clinical investigations.(12–14) A preclinical study from our laboratory showed that both selenium deficiency and selenium supplementation promoted myocardial fibrosis despite opposite effects on redox balance.(15) Our results revealed a link between redox reactions and methylation balance that could offer a potential explanation for the lack of benefit of anti-oxidant therapy, i.e., the off-setting effects of altered methylation and redox balance on matrix gene expression, such as elevated expression of the genes for collagen types I and III.(4,11,15)
The Xanthine Oxidase Inhibition for Hyperuricemic Heart Failure Patients (EXACT-HF) study examined the effect of allopurinol, which inhibits the xanthine oxidase pro-oxidant pathway, in patients with heart failure with reduced ejection fraction (HFrEF) and an elevated uric acid level (a surrogate marker of activation of the xanthine oxidase pathway).(16) The results of this trial did not show an improvement in clinical outcomes with allopurinol. In the current study we conducted a post-hoc analysis of the EXACT trial to examine the hypothesis generated by the above-mentioned pre-clinical data, i.e., whether baseline differences in methylation potential and changes due to anti-oxidant therapy are a potential explanation for the negative results of the clinical trial.
Materials and Methods
Investigators from the National Heart, Lung, and Blood Institute–sponsored Heart Failure Clinical Research Network conceived, designed, and conducted the Xanthine Oxidase Inhibition for Hyperuricemic Heart Failure Patients (EXACT-HF) trial.(16,17) Inclusion and exclusion criteria for the EXACT-HF study are described in the Supplementary Data. The Duke Clinical Research Institute served as the data-coordinating center. All patients provided written informed consent. This sub-study was approved by Institutional Review Boards of Brigham and Women’s Hospital and Duke University Medical Center.
Study population
Details of the EXACT-HF study design, population, end points, and statistical considerations, as well as the primary results, have been previously published.(16,17) Among the patients enrolled in EXACT-HF (N=253), we included the subset of patients who agreed to participate in the biorepository sub-study (N=195). Patients who had high values of the ratio of S-adenosyl-methionine (SAM) to S-adenosyl-homocysteine (SAH; ratio>5000) at baseline or week 24 were considered outliers and excluded from this analysis (n=6). The final cohort for this study included 189 patients. In addition, patients who died before week 24 were excluded from all endpoint analyses that involved change between baseline and 24 weeks or 24-week results (n=8). Baseline characteristics, including demographics, vital signs and laboratory data, and past medical history were obtained from the EXACT-HF study database. Blood was collected for biomarker assays as part of the EXACT trial at baseline, 12 weeks and 24 weeks and shipped and stored at the Heart Failure Network Biorepository at the University of Vermont, and plasma samples were made available for further analyses.
Measurement of methylation potential (SAM and SAH levels)
Plasma samples (30 µL) were mixed with 200 µL of water:methanol (25:75) containing internal standard (methionine-13C, S-adenosyl-homocysteine-d4, homocysteine-d4 10 ng/mL). The samples were vortexed and incubated on ice for 20 minutes. The samples were centrifuged at 13,000 rpm at 4oC for 20 minutes; supernatant was transferred to MS vial and 5 µL of sample was injected on a UPLC (Acquity, Waters Corp) in conjunction with triple quadrupole mass spectrometer (Xevo TQ-S, Waters Corp. USA), operating in the MRM (multiple reaction monitoring) mode for targeted quantitation of SAM, SAH, and homocysteine.(18) The samples were resolved on an UPLC BEH Amide 1.7 um, 2.1×100 mm column using water with 0.2% formic acid as buffer A and 95/5 ACN/water with 1mM EDTA as buffer B at 45oC column temperature. Signal intensities from all MRM Q1/Q3 ion pairs for three metabolites and internal standards were ranked to ensure selection of the most intense precursor and fragment ion pair for MRM-based quantitation. This approach resulted in selection of cone voltages and collision energies that maximized the generation of each fragment ion species. The mobile phase composition and selected reaction monitoring ion transitions are outlined in Supplementary Tables 1 and 2. The metabolite ratios were calculated by normalizing the peak area of endogenous metabolites within plasma samples normalized to the internal standard (SAM-d4 for SAM, HC-d4 for homocysteine, and methionine-13C for SAH). The sample queue was randomized, and solvent blanks were injected to assess sample carryover using four biological replicates for each comparative group.
Statistical analyses
Baseline characteristics including patient’s demographics, medical history, vital signs, medication intake and laboratory results were stratified by treatment arm for analysis. Continuous variables were summarized using medians (Q1, Q3) and categorical variables were summarized using counts (percent). Owing to significant missing data for some outcomes (e.g., week 24 LV end-systolic volume had 49% missingness), multiple imputation by fully conditional specification with 25 multiple imputation datasets was used. The relationships of changes in SAM, SAH, and SAM/SAH ratio with the primary study endpoint were not analyzed due to variability in the timing of outcomes. The primary endpoint was a composite clinical endpoint that aimed to classify a subject’s clinical status as improved, worsened, or unchanged at 24 weeks and was based on a classification scheme based on sequential rules applied to following outcomes: death; hospitalization, emergency room visit or emergent clinic visit for worsening HF; medication change for worsening HF; and Patient Global Assessment.(17)
Wilcoxon rank-sum tests were used to assess the association between the treatment groups and the methylation markers (SAM, SAH and SAM/SAH ratio) at baseline and 24 weeks, as well as the change between baseline and the 24-week follow-up time point.
The association between the change from baseline to 24 weeks in SAM, SAH, and SAM/SAH ratio and the treatment were assessed with general linear regression modelling, adjusting for baseline values of specific markers of methylation potential, gender, age, and baseline creatinine and estimated glomerular filtration rate (eGFR). In order to meet normality assumptions for the linear models, changes in markers for methylation potential were transformed using a Box-Cox transformation.
Correlation between change of methylation markers from baseline to week 24 and change in uric acid level from baseline to week 24 was analyzed using Spearman’s correlation coefficient for the entire population as well as by treatment arm. General linear regression modelling was used to test associations between changes in the SAM/SAH ratio from baseline to week 24 and secondary endpoints, including echocardiographic parameters at week 24, change in echocardiographic parameters from baseline to week 24, Kansas City Cardiomyopathy Questionnaire (KCCQ) overall summary score at week 24, and change in distance walked during the 6-minute walk test from baseline to week 24. Performed models were controlled for gender, age, and baseline creatinine and eGFR. Furthermore, baseline value of specific echocardiographic parameters and baseline 6-minute walk distance were used as additional adjusted variables where suitable.
P-values less than 0.05 were consider as statistically significant. All analyses were conducted using SAS software version 9.4 (SAS Institute Inc., Cary NC).
Results
Study participants
Of the 253 patients enrolled in the EXACT study, 189 patients were included in this study. As shown in Table 1, there were no major differences in demographic and clinical features between patients receiving allopurinol or placebo except for an increased use of amiodarone in the placebo group, a finding that was also observed in the entire EXACT study group.(16) Unlike in the EXACT study, the allopurinol and placebo groups in this sub-study did not demonstrate a difference in left ventricular ejection fraction.(16) Baseline laboratory parameters, including uric acid, N-terminal pro-B-type natriuretic peptide, and myeloperoxidase, were not different between the allopurinol and placebo groups.
Table 1.
Baseline Characteristics
| Parameter | Allopurinol (n=92) | Placebo (n=97) |
|---|---|---|
| Age | 63 (54–72) | 63 (52–69) |
| Male sex, n (%) | 79/92 (85.9%) | 76/97 (78.4%) |
| White race, n (%) | 60/92 (65.2%) | 59/97 (60.8%) |
| Duration of HF (years) | 5.2 (2.2–10.0) | 6.0 (1.9–10.5) |
| ≥ 1 HF hospitalizations within past year, n (%) | 59/92 (64.1%) | 70/97 (72.2%) |
| NYHA functional class, n (%) | ||
| II | 46/92 (50.0%) | 48/97 (49.5%) |
| III | 46/92 (50.0%) | 48/97 (49.5%) |
| IV | 0/92 (0.0%) | 1/97 (1.0%) |
| Left ventricular ejection fraction (%) | 26 (20–33) | 23 (20–30) |
| History of moderate or severe mitral regurgitation, n (%) | 30/92 (32.6%) | 36/96 (37.5%) |
| Ischemic cardiomyopathy, n (%) | 49/92 (53.3%) | 50/97 (51.5%) |
| Diabetes mellitus, n (%) | 51/92 (55.4%) | 53/97 (54.6%) |
| Hypertension, n (%) | 75/92 (81.5%) | 77/97 (79.4%) |
| Atrial fibrillation/flutter, n (%) | 46/92 (50.0%) | 46/97 (47.4%) |
| Gout, n (%) | 18/92 (19.6%) | 21/97 (21.6%) |
| Implantable cardioverter-defibrillator, n (%) | 63/92 (68.5%) | 67/97 (69.1%) |
| Cardiac resynchronization therapy, n (%) | 30/92 (32.6%) | 26/97 (26.8%) |
| Systolic blood pressure (mm Hg) | 108 (100–122) | 108 (100–115) |
| Heart rate (beats per minute) | 72 (66–80) | 75 (65–83) |
| Weight (lbs) | 215 (169–253) | 207 (173–249) |
| Body mass index (kg/m2) | 31.0 (25.8–38.3) | 31.4 (26.4–35.8) |
| Medications | ||
| ACE inhibitor or ARB, n (%) | 76/92 (82.6%) | 81/97 (83.5%) |
| Beta-blocker, n (%) | 87/92 (94.6%) | 91/97 (93.8%) |
| Aldosterone antagonist, n (%) | 48/92 (52.2%) | 49/97 (50.5%) |
| Digoxin, n (%) | 36/92 (39.1%) | 46/97 (47.4%) |
| Hydralazine, n (%) | 22/92 (23.9%) | 16/97 (16.5%) |
| Nitrates, n (%) | 27/92 (29.3%) | 29/97 (29.9%) |
| Amiodarone, n (%) | 12/92 (13.0%) | 25/97 (25.8%) |
| Furosemide-equivalent dose (mg) | 102 (70–160) | 120 (80–160) |
| Laboratory markers | ||
| Sodium (mEq/L) | 139 (137–140) | 139 (137–140) |
| Blood Urea Nitrogen (mg/dL) | 33 (26–42) | 31 (23–42) |
| Creatinine (mg/dL) | 1.54 (1.22–1.90) | 1.49 (1.17–1.80) |
| eGFR (mL/min/1.73m2) | 50.82 (38.4–65.6) | 52.29 (39.3–69.7) |
| Hematocrit (%) | 38.2 (35.0–41.0) | 39.0 (36.0–42.4) |
| Cystatin C (mg/L) | 1.50 (1.11–1.86) | 1.40 (1.00–1.73) |
| Uric acid (mg/dL) | 11.0 (10.2–12.1) | 11.2 (10.1–12.1) |
| NT-pro-BNP (pg/mL) | 2544 (1062–4427) | 2247 (1112–3570) |
| Myeloperoxidase (ng/mL) | 32.3 (22.1–50.3) | 30.7 (22.8–54.3) |
Continuous variables are presented as median (25th – 75th percentile) and categorical variables were summarized using counts (percent).
ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; eGFR – estimated glomerular filtration rate; HF, heart failure; NYHA, New York Heart Association; NT-pro-BNP – N-terminal pro-B-type natriuretic peptide
Methylation potential at baseline and end of treatment
Table 2 demonstrates homocysteine, SAM and SAH levels at baseline, 24 weeks, and the change between these two time points. While there was a trend towards higher homocysteine level in the allopurinol group at baseline, 24-week values and the change in values from baseline to 24 weeks were similar between the two groups. Despite randomization, baseline values of SAM and SAM/SAH ratio were lower in the allopurinol group compared to the placebo group. At 24 weeks, median value of SAH was higher in the allopurinol group compared to the placebo group. Median values of the change from baseline to 24 weeks of both SAM and SAH were higher in the allopurinol group compared to placebo group, although only the change in SAM values was statistically significant.
Table 2.
Methylation parameters
| Methylation marker | Baseline values | 24-week values | Change from baseline to 24 weeks | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Allopurinol* | Placebo* | P-value** | Allopurinol* | Placebo* | P-value** | Allopurinol* | Placebo* | P-value** | |
| Homocysteine (mmol/L) |
0.10 (0.07–0.13) | 0.09 (0.06–0.11) | 0.05 | 0.09 (0.06–0.13) | 0.09 (0.06–0.12) | 0.83 | −0.01 (−0.03–0.03) | 0.01 (−0.03–0.03) | 0.21 |
| SAM (nmol/L) | 798 (576–1033) | 977 (647–1365) | <0.001 | 815 (603–1201) | 779 (596–1066) | 0.44 | 101 (−238–436) | −158 (−468–219) | 0.004 |
| SAH (nmol/L) |
4.09 (2.34–5.39) | 3.84 (2.19–5.27) | 0.47 | 4.04 (2.55–5.90) | 3.23 (2.00–4.91) | 0.03 | 0.69 (−1.23–2.22) | 0.12 (−1.84–1.51) | 0.16 |
| SAM/SAH ratio | 198 (107–362) | 236 (179–450) | 0.02 | 214 (136–341) | 253 (175–422) | 0.07 | 0.3 (−137.0–126.8) | −38.7 (−175.1–205.3) | 0.74 |
Values are presented as median (25th – 75th percentile).
Wilcoxon Rank-Sum test was used to compute p-value.
SAM, S-adenosyl methionine; SAH, S-adenosyl homocysteine
Associations of methylation potential with treatment
Table 3 demonstrates the association of treatment group with changes in methylation marker values from baseline to 24 weeks, where placebo was used as reference group. The treatment effect was close to being significantly associated only with changes in SAH values from baseline to 24 weeks (p=.0501).
Table 3.
Association between treatment and change in methylation markers from baseline to week 24.
| Model§ | Estimate (95% CI) | P-value |
|---|---|---|
| change in SAM from baseline to week 24* | 1.186 (−1.009, 3.381) | 0.2892 |
| change in SAH from baseline to week 24* | 0.810 (0.000, 1.621) | 0.0501 |
| change in SAM/SAH from baseline to week 24* | −103.110 (−261.568, 55.348) | 0.2021 |
Box-Cox transformation was applied to models that violated normality assumption.
Placebo and male groups were used as reference groups. Models adjusted for gender, age at baseline, baseline value of specific DNA methylation marker, baseline creatinine and baseline GFR.
As uric acid is a surrogate marker of xanthine oxidase activity, we examined the relationship between the changes in SAM and SAH levels and changes in uric acid levels. As shown in Table 4, a statistically significant but weak correlation was observed only between change in SAH level and change in uric acid level in the allopurinol group. In the placebo group there were no significant correlations between change in methylation markers and change in uric acid levels. These findings suggest that methylation is impacted by treatment with a xanthine oxidase inhibitor.
Table 4.
Spearman correlation coefficients between changes in methylation markers and uric acid level from baseline to week 24
| R | P-value | |
|---|---|---|
| Entire population - correlation between change in uric acid level and: | ||
| change in SAM/SAH ratio | −0.050 | 0.519 |
| change in SAM | 0.039 | 0.619 |
| change in SAH | 0.099 | 0.196 |
| Patients randomized to Allopurinol - correlation between change in uric acid level and: | ||
| change in SAM/SAH ratio | −0.126 | 0.248 |
| change in SAH | 0.223 | 0.038 |
| change in SAM | 0.195 | 0.073 |
| Patients randomized to Placebo - correlation between change in uric acid level and: | ||
| change in SAM/SAH ratio | 0.015 | 0.895 |
| change in SAH | 0.107 | 0.337 |
| change in SAM | 0.157 | 0.158 |
SAM, S-adenosyl methionine; SAH, S-adenosyl homocysteine
Associations of methylation potential with outcomes
As shown in Table 5, there were no significant associations between changes in methylation potential from baseline to week 24 and measures of left ventricular structure and function, exercise capacity, or quality of life.
Table 5.
Associations between changes in SAM/SAH ratio from baseline to week 24 and outcomes
| Outcome | Estimate (95% CI) | P-value |
|---|---|---|
| Change in left ventricular diastolic volume index from baseline to week 24 (ml/m2)* | 0.000 (−0.015, 0.015) | 0.995 |
| Change in left ventricular systolic volume index from baseline to week 24 (ml/m2)* | 0.003 (−0.010, 0.016) | 0.639 |
| Log Left ventricular mass index at week 24 (gm/m2)* | −0.004 (−0.015, 0.006) | 0.430 |
| Change in left ventricular ejection fraction from baseline to week 24 (%)* | −0.001 (−0.003, 0.001) | 0.508 |
| KCCQ overall summary score at week 24** | −0.001 (−0.029, 0.027) | 0.928 |
| Change in 6-minute walk distance* | −0.005 (−0.022, 0.012) | 0.574 |
Model adjusted for baseline value of the parameter, treatment arm, gender, age, baseline creatinine and estimated glomerular filtration rate. Appropriate transformation was used where normality assumption was violated.
Model adjusted for treatment arm, gender, age, baseline creatinine and estimated glomerular filtration rate.
Male group and placebo group were used as reference groups in all models.
KCCG, Kansas City Cardiomyopathy Questionnaire; SAM, S-adenosyl methionine; SAH, S-adenosyl homocysteine
Discussion
Our results demonstrate that there is significant variability in the body’s potential to methylate various bioactive molecules in a cohort of patients with HFrEF. Furthermore, our results demonstrate that treatment of hyperuricemic HFrEF patients with an anti-oxidant medication is associated with changes of the levels of the methyl donor SAM and the methylation inhibitor SAH. The correlation of changes in SAH levels with changes in uric acid levels in patients treated with allopurinol also suggests that changes in flux through redox regulatory pathways may directly influence methylation reactions. The biological variability in SAM and SAH levels at baseline and during anti-oxidant therapy observed in patients with heart failure underscores the limitation of treating all-comers with HFrEF similarly, and the need to move towards more personalized approaches to treatment.
Methylation reactions are crucial to homeostasis and are catalyzed by various methyl transferases that promote the transfer of a methyl group from SAM to many macromolecules, such as DNA, RNA, catecholamines, histones, and other proteins and phospholipids.(4) DNA and histone methylation play crucial roles in epigenetic modifications, which underlie embryonic developmental programs and the integration of genetic and environmental stimuli.(19) A study by Movassagh and colleagues on normal and end-stage cardiomyopathic human hearts showed that distinct epigenetic patterns of DNA and histone methylation exist in end-stage heart failure,(20) indicating that epigenetic modifications may contribute to the development of heart failure. Signaling mediated by oxygen free radicals is crucial to normal cellular functioning; both oxidant stress and reductive stress have been proposed to play a role in cardiovascular diseases including heart failure.(11,15,21) The finding that baseline SAM levels varied between the allopurinol and placebo groups indicates the variability in the methylation status in patients before the initiation of treatment, which could also have influenced the outcomes.
The methionine-homocysteine cycle serves to integrate methylation balance and redox balance by controlling the amounts of SAM and SAH and by regulating the synthesis of the anti-oxidants cysteine and glutathione (Figure 1). Several of the enzymes in the methionine-homocysteine cycle are redox sensitive.(4) The enzyme cystathionine beta synthase (CBS) catalyzes the first reaction in the transsulfuration pathway that leads to conversion of homocysteine to the antioxidants cysteine and glutathione. Oxidation of a heme moiety in a regulatory heme binding domain of CBS increases enzyme activity, while heme reduction decreases CBS activity. Similarly, methionine synthase, which remethylates homocysteine to methionine, is inactivated by oxidation, while cobalamin, a cofactor for methionine synthase activity, is also inactivated by oxidant stress. Furthermore, oxidized glutathione can inhibit the activity of the enzyme that synthesizes SAM.(22) Hence changes in oxidant status can affect methylation reactions by affecting flux through the methionine-homocysteine cycle. For example, an in vitro study in human liver cells showed that oxidant stress increased transsulfuration of homocysteine to cysteine and glutathione.(23) As shown by our results, the plasma homocysteine level may not completely reflect the levels of disturbance of flux through the methionine-homocysteine cycle, necessitating the measurement of SAM and SAH levels also.
Figure 1.
Regulation of redox and methylation balance by the methionine-homocysteine cycle and the effects of oxidant levels. Green arrows indicate positive regulation and red arrows indicate negative regulation. See text for additional explanation.
While we cannot determine a specific mechanism from this study, a decrease in oxidant status with anti-oxidant therapy may lead to less CBS activity, as well as increased flux through methionine and SAM synthesis that could contribute to higher concentrations of homocysteine and hence SAH. A preclinical study from our laboratories showed the complex relation between redox balance and methylation. In this study, we utilized dietary selenium deficiency as a model of oxidant stress and selenium supplementation as a model of anti-oxidant supplementation, and demonstrated that both dietary modifications led to the adverse phenotype of myocardial fibrosis.(15) While selenium deficiency led to increased oxidant levels, studies of methylation revealed that selenium supplementation increased the levels of SAH (a powerful inhibitor of methyltransferases), decreased DNA methyltransferase activity, and reduced DNA methylation.
Human cross-sectional studies have also demonstrated a link between redox balance and methylation. Children with autism have been found to have a significantly lower ratio of SAM to SAH and a significantly higher ratio of oxidized to reduced glutathione in plasma compared to control children.(24) A study of adults from Bangladesh examined the association between whole blood oxidized glutathione levels and peripheral blood mononuclear DNA methylation levels.(25) While a significant association was observed between oxidized glutathione status and DNA methylation, blood SAM levels were not significantly associated with DNA methylation. A translational study on the effects of fine particulate matter on cultured human neuroblastoma cells showed concomitant changes in redox balance and DNA methylation.(26) Interestingly, in a small study of 15 patients with allopurinol-induced severe cutaneous adverse reactions, genomic DNA methylation profiling demonstrated differences in DNA methylation between those with adverse cutaneous reactions and those who were tolerant of allopurinol.(27)
While we have not measured intracellular levels of SAM and SAH, previous clinical studies have shown a close correlation between plasma SAM/SAH ratio and intracellular SAM/SAH ratio in peripheral blood lymphocytes.(28) Since the affinity of methyltransferases for SAM is quite high, the concentration of the methyltransferase inhibitor SAH is a more potent influence on methylation reactions.(29) Our study demonstrated that treatment with allopurinol was associated with maintained SAH levels between baseline and 24 weeks, as opposed to a decline in the SAH level between baseline and 24 weeks in the placebo group, suggesting that anti-oxidant treatment may be associated with significant dynamic changes in methylation. This conclusion is further supported by the correlation observed between change in uric acid level and change in SAH level in the allopurinol group but not in the placebo group. While the uric acid level is not a direct measurement of oxidant levels or oxidant stress, since xanthine oxidase inhibition reduces superoxide levels by preventing the conversion of hypoxanthine to xanthine and subsequently to uric acid, these data provide indirect evidence that changes in oxidant levels lead to changes in methylation potential.
The links between redox balance and methylation balance offer potential explanations for the association of allopurinol treatment with plasma SAH levels (Figure 2). As discussed above and illustrated in Figure 1, it is possible that exogenous anti-oxidants, by reducing oxidant levels, may decrease the metabolism of homocysteine to cysteine and glutathione, and lead to an increase in the levels of SAH. As mentioned above, SAH levels affect methylation more than SAM levels. Plasma SAH levels correlate with intracellular SAH levels and reduced DNA methylation in lymphocytes.(30) In a small case-control study, Kerins and colleagues observed a significant elevation of plasma SAH levels in patients with proven cardiovascular disease compared to controls.(31) Xiao and colleagues examined plasma SAH levels in 1,003 patients undergoing coronary angiography. During a median follow-up of 3 years, they observed a significant and independent association of adverse cardiovascular events with plasma SAH concentration.(32)
Figure 2.
Potential mechanisms to explain the results of our study. HFrEF – heart failure with reduced ejection fraction.
In conclusion, our study demonstrates that there is significant biological variability of methylation potential in a cohort of patients with HFrEF and that treatment with anti-oxidant therapy is associated with changes in methylation potential. These findings offer possible explanations for the lack of efficacy of anti-oxidant interventions in clinical trials and underscore the need to move towards personalized approaches to heart failure pathobiology and therapy.
Supplementary Material
Key Points:
Redox balance and methylation reactions are linked through the methionine-homocysteine cycle.
The body’s methylation potential varies widely among individuals with heart failure
Methylation potential is altered by anti-oxidant therapy.
These findings may explain the neutral results of clinical trials of anti-oxidant therapy.
Acknowledgements
The authors gratefully acknowledge the EXACT-HF study investigators, coordinators and patients for their time and effort on behalf of the clinical trial. We also thank the staff of the core biomarker laboratory at the University of Vermont, under the direction of Russel Tracy PhD, for their handling of clinical specimens. The authors would also like to acknowledge the Metabolomics Shared Resource at Georgetown University (Washington, DC, USA) that is partially supported by NIH/NCI/CCSG grant P30-CA051008.
Funding: This work was supported by grants from the National Institutes of Health: NHLBI coordinating center: U10HL084904; and regional clinical center: U10HL110337
Footnotes
Compliance with Ethical Standards
Disclosure of potential conflicts of interest: The authors do not report any conflicts of interest.
Research involving Human Participants and/or Animals: This is a sub-study utilizing data and blood samples collected as part of the primary study. No additional data or blood collection was performed as part of this study. This sub-study was approved by Institutional Review Boards of Brigham and Women’s Hospital and Duke University Medical Center.
Informed consent: Informed consent was obtained from participants as part of the main study EXACT-HF.
References
- 1.Tsutsui H, Kinugawa S, Matsushima S. Oxidative stress and heart failure. Am J Physiol Heart Circ Physiol 2011;301:H2181–90. [DOI] [PubMed] [Google Scholar]
- 2.Vivekananthan DP, Penn MS, Sapp SK, Hsu A, Topol EJ. Use of antioxidant vitamins for the prevention of cardiovascular disease: meta-analysis of randomised trials. Lancet 2003;361:2017–23. [DOI] [PubMed] [Google Scholar]
- 3.Bjelakovic G, Nikolova D, Gluud LL, Simonetti RG, Gluud C. Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases. Sao Paulo Med J 2015;133:164–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Joseph J, Loscalzo J. Methoxistasis: integrating the roles of homocysteine and folic acid in cardiovascular pathobiology. Nutrients 2013;5:3235–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Vasan RS, Beiser A, D’Agostino RB et al. Plasma homocysteine and risk for congestive heart failure in adults without prior myocardial infarction. JAMA 2003;289:1251–7. [DOI] [PubMed] [Google Scholar]
- 6.Devi S, Kennedy RH, Joseph L, Shekhawat NS, Melchert RB, Joseph J. Effect of long-term hyperhomocysteinemia on myocardial structure and function in hypertensive rats. Cardiovasc Pathol 2006;15:75–82. [DOI] [PubMed] [Google Scholar]
- 7.Joseph J, Joseph L, Devi S, Kennedy RH. Effect of anti-oxidant treatment on hyperhomocysteinemia-induced myocardial fibrosis and diastolic dysfunction. J Heart Lung Transplant 2008;27:1237–41. [DOI] [PubMed] [Google Scholar]
- 8.Wang X, Cui L, Joseph J et al. Homocysteine induces cardiomyocyte dysfunction and apoptosis through p38 MAPK-mediated increase in oxidant stress. J Mol Cell Cardiol 2012;52:753–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Marti-Carvajal AJ, Sola I, Lathyris D, Dayer M. Homocysteine-lowering interventions for preventing cardiovascular events. Cochrane Database Syst Rev 2017;8:CD006612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rafeq Z, Roh JD, Guarino P, Kaufman J, Joseph J. Adverse myocardial effects of B-vitamin therapy in subjects with chronic kidney disease and hyperhomocysteinaemia. Nutr Metab Cardiovasc Dis 2013;23:836–42. [DOI] [PubMed] [Google Scholar]
- 11.Joseph J, Loscalzo J. Selenistasis: epistatic effects of selenium on cardiovascular phenotype. Nutrients 2013;5:340–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Stranges S, Marshall JR, Natarajan R et al. Effects of long-term selenium supplementation on the incidence of type 2 diabetes: a randomized trial. Ann Intern Med 2007;147:217–23. [DOI] [PubMed] [Google Scholar]
- 13.Lippman SM, Klein EA, Goodman PJ et al. Effect of selenium and vitamin E on risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA 2009;301:39–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Labunskyy VM, Lee BC, Handy DE, Loscalzo J, Hatfield DL, Gladyshev VN. Both maximal expression of selenoproteins and selenoprotein deficiency can promote development of type 2 diabetes-like phenotype in mice. Antioxid Redox Signal 2011;14:2327–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Metes-Kosik N, Luptak I, Dibello PM et al. Both selenium deficiency and modest selenium supplementation lead to myocardial fibrosis in mice via effects on redox-methylation balance. Mol Nutr Food Res 2012;56:1812–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Givertz MM, Anstrom KJ, Redfield MM et al. Effects of Xanthine Oxidase Inhibition in Hyperuricemic Heart Failure Patients: The Xanthine Oxidase Inhibition for Hyperuricemic Heart Failure Patients (EXACT-HF) Study. Circulation 2015;131:1763–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Givertz MM, Mann DL, Lee KL et al. Xanthine oxidase inhibition for hyperuricemic heart failure patients: design and rationale of the EXACT-HF study. Circ Heart Fail 2013;6:862–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Klepacki J, Brunner N, Schmitz V, Klawitter J, Christians U, Klawitter J. Development and validation of an LC-MS/MS assay for the quantification of the trans-methylation pathway intermediates S-adenosylmethionine and S-adenosylhomocysteine in human plasma. Clin Chim Acta 2013;421:91–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cyr AR, Domann FE. The redox basis of epigenetic modifications: from mechanisms to functional consequences. Antioxid Redox Signal 2011;15:551–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Movassagh M, Choy MK, Knowles DA et al. Distinct epigenomic features in end-stage failing human hearts. Circulation 2011;124:2411–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rajasekaran NS, Connell P, Christians ES et al. Human alpha B-crystallin mutation causes oxido-reductive stress and protein aggregation cardiomyopathy in mice. Cell 2007;130:427–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Pajares MA, Duran C, Corrales F, Pliego MM, Mato JM. Modulation of rat liver S-adenosylmethionine synthetase activity by glutathione. J Biol Chem 1992;267:17598–605. [PubMed] [Google Scholar]
- 23.Mosharov E, Cranford MR, Banerjee R. The quantitatively important relationship between homocysteine metabolism and glutathione synthesis by the transsulfuration pathway and its regulation by redox changes. Biochemistry 2000;39:13005–11. [DOI] [PubMed] [Google Scholar]
- 24.James SJ, Cutler P, Melnyk S et al. Metabolic biomarkers of increased oxidative stress and impaired methylation capacity in children with autism. Am J Clin Nutr 2004;80:1611–7. [DOI] [PubMed] [Google Scholar]
- 25.Niedzwiecki MM, Hall MN, Liu X et al. Blood glutathione redox status and global methylation of peripheral blood mononuclear cell DNA in Bangladeshi adults. Epigenetics 2013;8:730–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wei H, Liang F, Meng G et al. Redox/methylation mediated abnormal DNA methylation as regulators of ambient fine particulate matter-induced neurodevelopment related impairment in human neuronal cells. Sci Rep 2016;6:33402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sun B, Cheng L, Xiong Y et al. PSORS1C1 Hypomethylation Is Associated with Allopurinol-Induced Severe Cutaneous Adverse Reactions during Disease Onset Period: A Multicenter Retrospective Case-Control Clinical Study in Han Chinese. Front Pharmacol 2017;8:923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Melnyk S, Pogribna M, Pogribny IP, Yi P, James SJ. Measurement of plasma and intracellular S-adenosylmethionine and S-adenosylhomocysteine utilizing coulometric electrochemical detection: alterations with plasma homocysteine and pyridoxal 5’-phosphate concentrations. Clin Chem 2000;46:265–72. [PubMed] [Google Scholar]
- 29.Hoffman DR, Cornatzer WE, Duerre JA. Relationship between tissue levels of S-adenosylmethionine, S-adenylhomocysteine, and transmethylation reactions. Can J Biochem 1979;57:56–65. [DOI] [PubMed] [Google Scholar]
- 30.Yi P, Melnyk S, Pogribna M, Pogribny IP, Hine RJ, James SJ. Increase in plasma homocysteine associated with parallel increases in plasma S-adenosylhomocysteine and lymphocyte DNA hypomethylation. J Biol Chem 2000;275:29318–23. [DOI] [PubMed] [Google Scholar]
- 31.Kerins DM, Koury MJ, Capdevila A, Rana S, Wagner C. Plasma S-adenosylhomocysteine is a more sensitive indicator of cardiovascular disease than plasma homocysteine. Am J Clin Nutr 2001;74:723–9. [DOI] [PubMed] [Google Scholar]
- 32.Xiao Y, Zhang Y, Wang M et al. Plasma S-adenosylhomocysteine is associated with the risk of cardiovascular events in patients undergoing coronary angiography: a cohort study. Am J Clin Nutr 2013;98:1162–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.