Impact of Methyl-Donor Micronutrient Supplementation on DNA Methylation Patterns: A Systematic Review and Meta-Analysis of in vitro, Animal, and Human Studies

Epigenetics involves the study of heritable changes in gene expression without altering the DNA sequence [1] and is an essential component of an organism’s development [2]. In spite of that, due to its reversible characteristic, the epigenetic machinery can suffer from dysregulation, which is linked to the origin and progression of human disease [3]. Thus, beyond the individual’s genotype, epigenetic mechanisms may additionally reflect environmental exposures, considering lifestyle factors such as diet, smoking habits, and alcohol intake [4]. Recent studies have shown that irregular patterns in DNA methylation are directly associated with metabolic disorders and diseases such as cancer [5], systemic lupus erythematosus [6], obesity, and cardiovascular diseases [7].

DNA methylation is a heritable epigenetic mechanism characterized by covalent transfer of methyl groups onto the C5 position of the cytosine in CpG sites. When it occurs in a gene promoter region, it may result in gene silencing [8]. This biological pathway is regulated by DNA methyltransferase enzymes that catalyze the transfer of a methyl group from S-adenosyl methionine (SAM) to form 5-methylcytosine [9]. These methyl groups are derived from 1-carbon metabolism, a process involved in amino acid and nucleotide metabolism that comprises a group of biochemical reactions responsible for transferring a single carbon (1C) unit for biosynthetic processes such as purine synthesis and homocysteine remethylation [10]. During this process, some nutrients, such as folate, vitamin B12, vitamin B6, betaine, choline, and methionine, play a significant role as cofactors or methyl-donors [10].

The status of methyl-donor micronutrients may be associated with DNA methylation profiles. Disturbances in 1C metabolism due to (i) low intake of methyl-donor nutrients; (ii) malabsorption of these nutrients due to disease or cellular conditions; and (iii) common polymorphisms in genes that encode important enzymes may have implications on DNA methylation status [11, 12]. Multiple studies have shown that altered ingestion of dietary folate, choline, betaine, B vitamins, and methionine may cause both global changes and changes in the promoter regions of specific genes in animals and humans [13‒15]. For example, there is evidence showing that folate-deficient diets may induce DNA hypomethylation [16‒18]. Additionally, other studies have demonstrated possible associations between methyl-donor status – characterized by serum levels of folic acid and B vitamins – and DNA methylation patterns [19, 20].

Despite the above evidence, it is currently unclear whether DNA methylation profiles are susceptible to changes in response to methyl-donor micronutrient supplementation, even though methyl donor supplementation could be a promising agent for precision nutrition. Given this, we conducted a systematic review and meta-analysis of intervention studies investigating methyl-donor supplements on DNA methylation, with a particular focus on folic acid.

This review was carried out following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline [21]. The protocol was registered in the PROSPERO database (CRD42022346561).

Eligibility criteria were defined according to the PICO framework: Population (humans, animals, or cell culture), Intervention (methyl-donor micronutrient supplementation or treatment), Comparison (placebo, different dose), and Outcome (DNA methylation levels). Studies were eligible if they (1) examined DNA methylation level in any sample or tissue and (2) performed any kind of methyl-donor micronutrient supplementation or treatment. As such, we included (1) in vitro studies of cell lines exposed to methyl-donor treatment; (2) animal model studies that examined the effects of methyl-donor nutrient supplementation; and (3) human clinical trials including healthy or unhealthy individuals who received any methyl-donor micronutrient supplementation (folic acid, vitamin B12, choline, methionine, betaine, methyltetrahydrofolate). In particular, we included animal model studies examining murine models, animal, or other models closest to human homeostasis. We defined dietary supplementation as a modification of the methyl-donor content in the diet (in the case of animal studies), or as the provision of an oral supplement (in the case of human clinical trials). Single case reports, human clinical trials modifying methyl-donor intake exclusively through dietary changes (i.e., without supplementation) or involving co-supplementation with other nutrients in addition to methyl-donors were excluded. Review papers, editor letters, and conference abstracts were also excluded.

The search for relevant studies was performed in four databases (Medline, Embase, Scopus, and Web of Science) from inception to May 2022, without language restrictions. The search strategy used a combination of terms related to DNA methylation (“DNA methylation” OR methylation OR “Genomic DNA”[MeSH terms]) AND supplementation (Supplement OR supplements OR “dietary supplements” OR “food supplement”[MeSH terms] OR “food supplements”[MeSH terms]) AND methyl-donor nutrients (“Methyl donor” OR “Methyl-group donor” OR “folic acid” OR “vitamin B9”[MeSH terms] OR folate [MeSH terms] OR “vitamin B12” OR cobalamin [MeSH terms] OR cyanocobalamin [MeSH terms] OR “vitamin B2” OR “vitamin B6” OR pyridoxine [MeSH terms] OR choline OR betaine OR micronutrient OR methionine OR l-methionine [MeSH terms]). A detailed description of all search strategies is described in Supplementary File 1 (for all online suppl. material, see https://doi.org/10.1159/000533193). In addition, we also carried out a manual search to retrieve any other relevant studies.

As our selected outcome of interest, we evaluated DNA methylation levels using any valid technique (direct or indirect measurement) with a comparison between a control and treated group, or between groups following a treatment period (pre-post design). This included event levels of global DNA methylation and also methylation of specific genes or CpG sites. In the case that a given study provided more than one outcome (e.g., one effect size comparing folic acid supplementation to control for 10 weeks, and another effect size supplementing for 20 weeks), both outcomes were extracted and coded appropriately, and were later analyzed in the meta-analytical model (discussed in more detail below).

All identified studies were imported into Rayyan, a specific electronic application for systematic review and meta-analysis (https://rayyan.qcri.org/welcome). After removal of duplicates, two independent reviewers (J.C.N.L.M. and L.M.C.) applied the inclusion criteria and screened all titles and abstracts. Full texts were read, evaluated, and assessed for inclusion independently by both reviewers. Disagreements were resolved by discussion or referral to a third reviewer (C.F.N.).

Three independent reviewers performed the data extraction of included studies (J.C.N.L.M., L.M.C., and A.A.R.). For each study, we extracted information about research features (name of the first author, year of publication, research design, country), DNA methylation assessment method, DNA methylation outcome (global or locus specific, including gene of interest when appropriate), and global DNA methylation levels pre- and post-intervention or between groups. For human clinical trials, additional information was extracted regarding participants’ characteristics (i.e., age, gender, health, or disease status), sample size, doses, and type of methyl-donor supplements, their administration frequency, and follow-up duration. For the in vitro and animal model studies, information was extracted on the cell or animal model (species), study design (in vitro: assay, exposure time, animal: tissue), and intervention protocol (in vitro: type of compounds and concentration/dose tested; animal: type and doses of methyl-donor nutrients and experiment duration).

Risk of bias of human clinical trials was performed using the Revised Cochrane risk-of-bias tool for randomized trials (RoB 2) [22] or risk of bias in Non-randomized Studies of Interventions (ROBINS-I) [22]. Risk of bias of animal model intervention studies was assessed using the Systematic Review Centre for Laboratory Animal Experimentation (SYRCLE) tool [23].

We judged each domain source of bias as high, low, or unclear. We summarized the risk of bias judgments across different studies for each of the domains listed. All analyses were performed independently by two authors (C.F.N. and S.M.S.). Any discrepancies were addressed by the re-evaluation of the original article, discussed, and solved by consensus. Studies were not excluded based on their quality.

Considering that individual studies evaluated different target genes, we chose to carry out the meta-analysis only for the global DNA methylation outcome. Furthermore, we only conducted a meta-analysis for the most common intervention (i.e., folic acid supplementation). Separate random-effects meta-analyses were performed for animal model and human clinical trials. A random-effects model, rather than a fixed-effects model, was selected as it more appropriately gives the potential heterogeneity contained in the selected studies. In the case of animal model studies, means and standard deviations extracted from experimental groups (either folate restriction or supplementation through diet modification) were compared to a control group (normal folate content), and standardized mean differences (Hedges’ g) were calculated by subtracting the means from each group and dividing by using a pooled standard deviation value. Effect sizes were classified according to dose category (low dose/restriction if folate content was lower than 2 mg/kg of diet; high dose when folate content was higher than 2 mg/kg of diet; and very high dose when content was higher than 20 mg/kg of diet). Thereafter, we conducted meta-analyses using a random-effects model with meta-regression, in which the dose category (low dose/folate restriction, high dose, and very high dose) was added to the model as a fixed factor. The moderator’s test statistic and regression coefficients for each category, including p values, are provided in a regression table.

In the case of human clinical trials, these employed a placebo controlled, longitudinal design. As such, standardized mean differences were calculated using the d_ppc2 effect size and its accompanying variance according to Morris [24]. To calculate the variance, a pre-post correlation value was necessary, but since no study reported such a value, we assumed 0.7 as a pre-post correlation value, which is a default value and is likely to be a conservative estimate. Additionally, meta-regressions for dose and duration were performed when at least 3 outcomes were present for each category level.

In many cases, individual studies provided multiple outcomes. There are many options when dealing with dependent effect sizes in meta-analysis, but a commonly used one is three-level meta-analytical models, in which outcomes are considered the second level, and studies the third level of the model. Simulation studies have shown that such multi-level models are robust and capable of providing accurate effect size and error estimates in these and other likely scenarios [24, 25]. This was our selected approach, and models were carried using the rma.mv function in R package metafor [23], which employs equations 3, 4, and 5 described by Noortgate et al. [25]. Herein, we report overall random-effects estimates obtained from each model, alongside 95% confidence intervals (95% CI), and the estimated variance within levels 2 and 3 (i.e., variance between outcomes and variance between studies). In forest plots, outcomes from the same studies are represented by having the same study label, followed by numbering. Additionally, heterogeneity was assessed using Cochrane’s Q-test for heterogeneity, alongside tau² values and its confidence intervals. Small-study effects were assessed by visual inspection of funnel plots and by the Egger’s regression intercept test [26]. Statistical significance was set at p < 0.05 prior to analysis. All data were analyzed using R and Rstudio software (R version 4.2.0, R Foundation for Statistical Computing, Vienna, Austria; Rstudio version 1.4.1103, PBC, USA) and the rma.mv function from the metafor package [27], and the tidyverse [28].

A total of 2,422 studies (2,036 after removing duplicates) were identified using the search strategy. A flow diagram showing literature searches, selection process, and study number is presented in Figure 1. Briefly, 1,957 articles were excluded based on titles and abstract. Seventy-nine articles were identified as potentially eligible studies. Twenty-seven were excluded for various reasons after detailed reading. Finally, five studies (four human clinical trials and one animal study) were manually included after checking the references of included studies and consulting experts in the field. Therefore, 57 studies were included, among which 18, 35, and 4 involved humans, animals, and in vitro models, respectively. Collectively, in vitro, animal model studies and human clinical trials suggest that methyl-donor micronutrient supplementation may impact DNA methylation levels, although significant heterogeneity and inconclusive results exist and are further discussed in detail below.

All four in vitro studies used folic acid as a treatment. Folic acid doses were very different and the duration of treatment ranged from one [29] to six [30] days. One study [31] evaluated only global DNA methylation levels; another assessed both global and specific gene methylation levels [30], and two studies [29, 32] assessed DNA methylation of specific target genes (Table 1).

Effects of folic acid treatment on global DNA methylation were contradictory, with studies demonstrating either no effect [31] or changes in different gene regions [30]. Also, Price et al. [32] did not find any effect of different concentrations of folic acid treatment breast cancer-associated genes 1 and 2 (BRCA1; BRCA2) methylation levels in different cell lines. On the other hand, Cui et al. [29] observed that folic acid treatment increased DNA methylation levels in promoter regions of MCP1 and VEGF genes.

Of the 35 animal model studies, 14 were conducted with mice (C57BL/6J, CBA/J, Swiss Albino, C57BL/6, AxinFu/þ, and/or INS-GAS) [33‒46] and 21 with rats (Sprague-Dawley, Long-Evans and Wistar) [47‒67]. The route of supplement administration included adding to water [33, 34, 43] or diet (chow) [35‒38, 40, 42‒45, 47, 49‒56, 58, 59, 61‒67], oral gavage [41] and subcutaneous injection [57]. Nineteen studies supplemented only folic acid [35, 39, 41–45, 48, 50, 54, 56, 59, 60, 62, 64–67], four only choline [33, 57, 61, 63], three only betaine [47, 49, 55], three only methionine [37, 38, 58], and six studies used a mix of methyl-donor supplements [34, 36, 40, 51‒53]. Samples included maternal placenta and diverse fetal or own animal tissues such as liver, brain, kidney, adipose tissue. Most of them (n = 22) evaluated the effect of maternal supplementation on DNA methylation levels in offspring [33, 35, 36, 39–42, 45, 47–51, 54–56, 59, 61–63, 65, 67]. Considering methylation measurements, 11 studies reported on only global DNA methylation [38, 39, 41, 42, 48, 56, 57, 59, 60, 64, 66], three evaluated global DNA methylation in addition to gene-specific targets [33, 45, 52], three articles performed genome-wide assessment [43, 46, 54], and 19 evaluated DNA methylation in specific genes or imprinted loci [33–37, 40, 42, 47, 49–51, 53, 55, 58, 61–63, 65, 67] (Table 2).

Changes in global DNA methylation level were observed after choline supplementation [33, 57] and after a treatment of a mix of methyl-donor nutrients [52]; however, no effects on global DNA methylation levels were observed after methionine supplementation [38]. The effect of folic acid was diverse, with studies reporting either no effect [60, 66], increase [39, 48, 64], or decrease [41, 59] of global DNA methylation levels following supplementation.

In human clinical trials, the supplementation protocol, the method used to assay DNA methylation, and the tissue/sample used varied widely. Ten studies were parallel-group randomized controlled trials (RCTs) [68‒77] and one was a crossover RCT [78]. The remaining studies were case-control [79, 80] or non-randomized clinical trials [81‒85]. In total, 15 studies used folic acid as the supplementation agent [68–71, 73–75, 77–79, 81–85], whereas two studies involved a combination of folic acid and vitamin B12 [73, 77], and one study used a combination of folic acid and methionine [80]. The supplementation protocol had important variation, with the dosages used ranging from 400 μg to 5 mg, and the duration of supplementation ranging from 8 weeks to 3 years (Table 3). Samples analyzed included blood [68, 69, 71, 74‒77, 79‒85], cord blood [68, 70], saliva [73], sperm [84], rectal [72, 78], and colonic tissues [74, 77]. The number of participants ranged from 7 to 216, and the age ranged from 27.2 to 71.1 years. Four studies analyzed DNA methylation in a genome-wide spectrum [68, 73, 76, 85], five assessed DNA methylation in specific genes [69, 70, 72, 80, 84], and nine evaluated global DNA methylation [71, 74, 75, 77–79, 81–83].

Results after folic acid supplementation were divergent. Despite some studies reporting an increase in global DNA methylation in blood and/or rectal/colonic mucosa [77, 78, 83], other studies showed no significant differences in blood [71, 75, 82], or even a decrease in DNA methylation [74, 79, 81]. Moreover, van der Kooi et al. [80] found no significant changes in the methylation levels of D4Z4 allele in both patients and controls after folic acid (5 mg) and methionine (3 g) supplementation. Likewise, Van Den Donk et al. [72] did not observe any effect of folic acid (4.6 mg) and B12 vitamin (1.1 mg) supplementation in promoter regions of specific genes. Aarabi et al. [84], after assessing DNA methylation levels of several imprinted loci, observed a significant decrease in different regions of the sperm genome in response to supplementation. Interestingly, some authors [68‒70] supplemented healthy pregnant women with folic acid (400 μg) and reported different results. Irwin et al. [68] observed an increase in methylation level of ZFP57 zinc finger protein gene (ZFP57) in cord blood after maternal supplementation, whereas Caffrey et al. [70] observed lower DNA methylation levels of insulin-like growth factor 2 (IGF2), brain-derived neurotrophic factor (BDNF), and long interspersed nuclear elements (LINE-1) in cord blood of newborns from mothers who received supplementation. In addition, Harrison et al. [69] did not find a difference in the level of DNA methylation of candidate genes in maternal blood after supplementation.

Risk of bias assessment is shown in Figure 2. According to “missing outcome data,” “randomization process,” and “measurement of the outcome” domains, most human clinical trials presented a “high risk” of bias. Considering animal model studies, in the domains “selection bias,” “performance bias,” “detection bias,” and “attrition bias,” more than 90% of the studies were judged as having “some concerns.”

Despite 9 animal studies having evaluated folic acid supplementation and global DNA methylation [39, 41, 45, 48, 56, 59, 60, 64, 66], two studies [39, 41] were not included in the meta-analysis because the supplementation protocol differed in some characteristic (secondary intervention following folic acid supplementation and folic acid administration by oral gavage instead diet). Thus, a total of seven studies were included, totaling twenty-two outcomes [45, 48, 56, 59, 60, 64, 66]. Meta-analysis with meta-regression for the effect of dose category showed that folic acid dose significantly affected DNA methylation (p value for moderator test = 0.008), with “high” and “very high” doses leading to a significant increase in DNA methylation as compared to “low” doses (high dose coefficient: 0.97 [95% CI: 0.30; 1.63], p = 0.001; very high dose coefficient: 1.15 [95% CI: 0.11; 2.20], p = 0.03) (Table 4). Figure 3a shows a forest plot with each individual outcome, and an adjusted estimate of each dose category (note that these adjusted estimates display the difference of each category against the null, and not each other). There was significant heterogeneity (p value for Cochrane’s Q test <0.0001), with more heterogeneity between outcomes (tau² = 0.19 [95% CI: 0.01; 0.76]) than between studies (tau² = 0.11 [95% CI: 0.00; 0.98]). Visual inspection of a funnel plot, including all outcomes or separately by dose category showed a fairly symmetrical distribution of effects sizes and did not suggest small-study effects (online suppl. File 2). An Egger’s intercept test was not significant (estimate = 0.465 [95% CI: −1.27; 2.20], p value = 0.58).

A total of four parallel group RCTs [71, 74, 75, 77] and one crossover RCT [78] evaluated global DNA methylation in humans and were included in a meta-analysis, totaling six outcomes. Meta-analysis showed that folic acid supplementation did not lead to statistically significant changes in DNA methylation when compared to placebo (0.20 [95% CI: −0.42; 0.82], p value = 0.52, Fig. 3b). There was significant heterogeneity (p value for Cochrane’s Q test <0.0001), with more heterogeneity between studies (tau² = 0.38 [95% CI: 0.00; 2.97]) than between outcomes (tau² = 0.04 [95% CI: 0.00; 1.96]). Potential sources that can explain this heterogeneity are discussed in the following section. Meta-regressions for supplementation dose and length were not carried out due to insufficient number of outcomes. Visual inspection of a funnel plot did not suggest small-study effects (online suppl. File 3). An Egger’s intercept test was not used due to the low number of analyzed outcomes.

This systematic review aimed to compile evidence of potential changes in DNA methylation induced by methyl-donor micronutrient supplementation. Overall, our findings provide evidence that these nutrients could change DNA methylation levels in vitro, in animal models and in human clinical trials; however, the studies investigated variable exposures and outcomes and showed heterogeneous results. The meta-analysis of animal studies evidenced that “high” and “very high” doses of folic acid supplementation positively affects global DNA methylation levels when compared to “low” doses.

Multiple investigations have shown that epigenetic mechanisms can be modulated, directly or indirectly, by external agents, such as diet [86]. Considering that, methionine is regenerated by methylation of homocysteine via folic acid and vitamin B12-dependent reactions, a diet deficient in either nutrient could modulate DNA methylation process [87], negatively affecting gene expression [88]. For example, folic acid depletion has been shown to cause DNA hypomethylation in lymphocytes from healthy women [89, 90]. Also, an animal model study demonstrated a positive correlation between placental DNA methylation, plasma folate levels, and SAM:SAH (S-adenosyl-homocysteine) ratio [91].

Taken together, 44 studies (77.2%) (one in vitro, 30 in animal model, and 13 human clinical trials) observed some change in DNA methylation levels after methyl-donor micronutrient supplementation. As previously mentioned, such changes were heterogeneous, with reductions and increases in DNA methylation levels being observed. Meta-analysis revealed that “high” and “very high” doses of folic acid supplementation increased global DNA methylation levels significantly in animal samples as compared to “low” doses. Coefficients of “high” and “very high” doses were very similar, suggesting that there is not an additive effect of “very high” over “high” dosages. In human clinical trials, meta-analysis showed that folic acid supplementation did not change global DNA methylation levels after supplementation when compared to placebo. These results do not support a dose-response of folic acid supplementation on global DNA methylation levels, at least concerning very high levels.

Animal model studies included in the meta-analysis used folic acid doses ranging from 3.5 (high) to 40 (very high) mg/kg of diet. According to the authors, these dosages were based on previous studies. On the other hand, folic acid doses used in human clinical trials ranged from 0.1 to 50 mg/day. Considering that the recommended dietary allowance for folic acid is 400 μg/day for men and women aged 14 or more years, we can observe that from the 19 human clinical trials included, only three studies [71, 74, 85] supplemented with a dose above recommended dietary allowance. Moreover, 10 studies [72, 73, 75, 78–84] supplemented with doses above the tolerable upper intake level (UL). Given this heterogeneity, we could not establish a direct relationship between the dose supplemented and the changes in DNA methylation profiles.

This heterogeneity extends to other methodological aspects. First, a vast array of methodologies has been used to assess DNA methylation and specific genomic loci, including simple and traditional methods focusing on candidate genes to modern technologies allowing for untargeted epigenome-wide investigation. Second, there was a broad diversity of analyzed tissue samples, encompassing less (e.g., blood and saliva) and more invasive (e.g., biopsy) methods. It is important to point out that there are locus-specific DNA methylation differences across tissues [92]; therefore, the effects of supplementation on DNA methylation may be dependent on the target tissue. In addition, human samples were collected from both healthy and individuals with specific diseases, whereas animal models differed according to lineage type, and in vitro studies diverged regarding cell type. Finally, it is important to highlight possible interactions between supplemented micronutrients, knowing that possible perturbations on the metabolism of one pathway may result in compensatory changes in others. Thus, the effect of a specific nutrient may differ when supplemented alongside another. As such, the results from this review should be interpreted with caution as variability in results is likely to be influenced by the heterogeneity in study methods.

Other literature weaknesses include the lack of adjusting for important confounding variables, such as dietary factors (e.g., folic acid and vitamin B12 dietary intake) known to alter epigenetic markers, and also the serum concentration of these nutrients. Also, it should be noted that a genetic variation in methylenetetrahydrofolate reductase (MTHFR) gene is relevant to the methyl-donor cycle. Considering that MTHFR enzyme catalyzes the reduction of 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate [10], individuals homozygous for the C667T variant present low blood folic acid concentrations. Thus, response of DNA methylation upon supplementation with folic acid may differ according to MTHFR genotype [10], which has not been evaluated in the literature.

Several diseases are associated with DNA hypomethylation; however, the loss of DNA methylation level can also have an impact on disease risk and development, potentially changing the patient’s prognosis [93]. Considering that epigenetic marks are potentially reversible, the management of DNA methylation levels emerges as a promising strategy to prevent the development of chronic human diseases or to change a disease’s prospective course. In this scenario, clinical epigenetics is a promising field of research and has been linked to personalized medicine and nutrition [94]. Personalized epigenetics is an emerging approach to assess an individual’s response to specific nutritional interventions considering individual genetic and epigenetic variability, promoting personalized strategies useful in disease prevention and also to maximize the results.

Nutrition plays a crucial role in modulating the dynamic DNA methylation profile, and an “epigenetic diet” may modify epigenetic homeostasis [95]. The application of nutrition strategies such as those involving micronutrient supplementation may be of value for epigenetic changes and may provide the possibility of designing individualized epigenetic interventions. However, despite this promising insight about how nutri-epigenomics could provide a health target from a nutritional standpoint, the knowledge about the appropriate dosages of methyl-donor micronutrients is still limited. In our systematic review, appropriate dose or timing for supplementation could not be established, and the heterogeneity of available studies makes it difficult to make any clinically oriented recommendations at this stage. Timing of supplementation and also the window of exposure may modulate the effects of supplementation on DNA methylation. In addition, it is important to consider that the epigenetic process is tissue-sensitive and is influenced by several individual-related features (genetic background, nutritional status, physical activity practice, inflammatory condition, metabolic diseases, stress), and external environmentals or exposures to factors such as drugs, cigarettes, and environmental pollutants. Considering that individualized medicine is a promising approach to determine precision nutrition recommendations, future human research should consider both healthy individuals and specific diseases, taking into account the supplementation of single and/or combined intake of methyl donors micronutrients, time and dose-response relationship, and also genetic variations, such as specific polymorphism in genes related to one-carbon cycling.

To our knowledge, this is the first systematic review and meta-analysis to examine the effect of methyl-donor micronutrient supplementation on DNA methylation. Collectively, results from this systematic review considering all included articles indicate that methyl-donor micronutrient supplementation may change DNA methylation levels in vitro, animal models, and in human clinical trials in different ways. However, meta-analyses indicated that folic acid supplementation only affects global DNA methylation levels in animals supplemented with high and very high doses. Considering the range of methodological factors and supplementation protocol/treatment divergences, it is difficult to establish a unifying conclusion. From a practical perspective, these effects, even if small, might be of clinical importance in the management of patients with diseases related to DNA hypomethylation. Thus, in the future, well-designed RCT studies are necessary to increase knowledge regarding the effects of methyl-donor micronutrient supplementation on DNA methylation in humans and its potential therapeutic effects.

None to declare.

An ethics statement is not applicable because this study is based exclusively on published literature.

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this manuscript.

This study was supported by São Paulo Research Foundation (FAPESP) (grants number #2020/01893-2, #2020/15126-3, #2021/09745-5, #2021/09753-8, and #2021/09777-4).

C.F.N. conceived the idea for the manuscript. A.A.R., L.M.C., J.C.N.L.M., and C.F.N. performed the literature review and wrote the manuscript. S.M.S. and G.P.E. performed meta-analysis. K.F.G. and B.G. edited and critically reviewed the manuscript. All authors approved the final version of the manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.