Numerous studies have focused on the link between diet and DNA methylation in mammalians to elucidate the dietary exposures that may have lifelong consequences on epigenetic marks [
12,
49,
50]. Different types of researchers (in vitro and in vivo) have presented the relationship between nutrition and DNA methylation, including prenatal and postnatal periods, showing that diets deficient in methyl donors and proteins may cause global DNA hypomethylation, or that high-fat diet consumption may result in changes in DNA methylation [
1,
51,
52,
53,
54]. One of the most popular models that has studied the link between diet and DNA methylation is the ‘yellow agouti (A
vy) mice’ model. The agouti gene is responsible for the regulation of brown/black (eumelanin) and yellow (pheomelanin) pigmentation in the mammalian coat. It has been shown that dietary methyl donors’ supplementation of dams can change the coat color by correlating with the A
vy methylation status [
55]. However, the agouti mice model and most of these studies were incapable of showing the underlying epigenetic mechanisms regarding whether the DNA methylation occurred due to the expression or inhibition of special binding sites of methylation enzymes, substrates, cofactors, or something else. Besides, there are other questions that need to be clarified, especially regarding the nutrient doses and the duration of dietary exposure to DNA methylation [
56].
3.1. How Diet Influences Methylation Cycle and Methyl Donors
The key methyl donor for DNA and protein methyltransferases, SAM, is synthesized in the methionine cycle while accompanied by various nutrients present in the diet, including methionine, folate, choline, betaine, vitamins B2, B6, and B12 [
57]. These nutrients act as precursors and contribute to the production of SAM, although they enter the cycle at different sites [
1]. Therefore, any deficiencies in these nutrients may result in changes in the SAM pool, which can influence DNMTS’ reaction kinetics and DNA methylation, as well. Taking this opinion into account, many studies have started to investigate the link between SAM availability and its dietary sources, together with endogenous genetic factors [
4].
Furthermore, after the removal of the methyl group, SAM is transformed into SAH, which is a strong competitive inhibitor of almost all methylation reactions, and also competes with SAM for the active site on the methyltransferase enzyme [
58,
59]. Since the supply of SAM and removal of SAH is necessary for DNMT activity, the SAM/SAH ratio has been suggested as a ‘methylation index’ to show the probability of DNA hypermethylation or hypomethylation [
50]. Remarkably, some earlier studies have shown that SAH is an inhibitor of the DNMT-mediated DNA methylation [
60,
61]. Additionally, moderate elevations in plasma homocysteine concentrations have been shown to be associated with increased levels of SAH, but not SAM, and increased SAH levels have been associated with global DNA hypomethylation [
59].
So far, the methyl and folate-deficient diets have been found to be largely associated with reduced levels of SAM, increased levels of SAH, and decreased SAM/SAH ratios in the livers of male rats and mice [
62,
63,
64]. The changes in SAM and SAH levels also showed irreversible alteration in hepatic DNA methylation [
63]. Moreover, a study showed that zinc deficiency has reduced the use of methyl groups from SAM in rat liver and resulted in global DNA hypomethylation [
65]. A low-protein diet or undernutrition during gestation in mice and in utero in human studies resulted in both hypomethylation and hypermethylation at specific loci in offspring [
66,
67,
68,
69]. Although one study hypothesized that the hypomethylation of certain promoters upon protein restriction may be a consequence of decreased methyl group availability [
66], most of the studies remained unclear regarding how diet changes the activity of DNA methylation, and they did not observe the upregulation of SAM, SAH, or DNMTs.
On the other side, high dietary methionine intake is believed to increase DNA methylation, and the methyl groups that are transferred in mammalian DNA methylation reactions are believed to eventually derive from methionine [
55]. High doses of folate supplementation showed an increase in methylation and normalized gene expression at specific loci, which is believed to induce a substantial increase of the intracellular pool of the SAM and SAM/SAH ratio [
70,
71]. However, little is known regarding the effect of methionine or methyl donors’ supplementation, and the mechanisms of action on DNA methylation are not clear [
50]. Very few studies have examined the epigenetic mechanisms of the effects of high methionine intake on DNA methylation. In an epigenetic mouse model, Tremolizzo et al. [
72] studied the effect of methionine on SAM, SAH, methylation status, and the expression of the reelin gene in the frontal cortex. The study showed interesting results. For example, after 15 days of methionine treatment, brain SAH was found to be double, whereas SAM was not affected. The reduction in the SAM/SAH ratio would be expected to hypomethylate DNA, but it has been found that specific CpG sites in the reelin promoter were actually hypermethylated in the cortex of methionine-treated mice. Hence, the significant increase in CpG methylation appeared to downregulate reelin expression. A follow-up study from Dong et al. [
73] showed that a 15-day methionine (MET) treatment increased the binding of methyl CpG binding protein 2 (MeCP2) to the reelin promoter, which is thought to be the factor behind hypermethylation. However, the same effect was not found in other control genes (
Gad65 and
β-
globin). Another model examining MET-induced alterations in DNA methylation found no significant dietary effects on genome-wide DNA methylation, although methionine supplementation significantly decreased the SAM/SAH ratio in the liver and brain [
74]. The problem with determining the SAM, SAH, and their ratio in order to examine nutritional influences on DNA methylation is complicated for several reasons. To start with, each mammalian cell is responsible for synthesizing its own SAM, and SAM cannot cross the plasma membrane. However, SAH does leak from the cell with excessive accumulation. Thus, interpreting the SAM/SAH ratio on a tissue-specific basis and the ratio in plasma may not provide a meaningful indication of systemic methylation [
75].
Betaine is an important methyl donor, which can be produced by choline or taken through diet. Betaine converts into dimethylglycine (DMG) after it provides a methyl group to homocysteine [
11]. Studies have suggested that plasma DMG is a good indicator of betaine utilization as a methyl donor [
76,
77]. Moreover, SAM can inhibit BHMT and reduce the usage of betaine as a methyl donor [
10], and it is important for SAM to stimulate the BHMT pathway in order to sustain its concentrations [
78]. Choline methyl-deficient diets showed reduced hepatic concentrations of SAM and increased levels of SAH in the livers of mice [
62]. A rat study evaluating the choline-deficient diet for seven days also showed that the effects of choline deficiency on reducing liver methionine formation by 20–25%, SAM by 60%, and increasing liver SAH by 50% were significant [
79]. Plasma SAM levels were found to be significantly correlated with plasma levels of choline and DMG, but not with betaine [
80]. To date, evidence has also shown that folate deficiency may lower choline and betaine levels in liver, or that choline deficiency may decrease hepatic folate stores, and thus can affect the methyl transfer of one carbon cycle in the liver [
81,
82]. On the other hand, a study showed that folic acid-supplemented, BHMT-deleted mice have produced more hepatic SAM compared to BHMT-deleted mice fed a folate-deficient diet or a control diet [
83]. It has been a long time since a diet very low in choline and methionine resulted in the decreased methylation of cytosine in the liver [
84,
85,
86,
87]. However, studies have failed to show the direct interaction between choline, biotin, and DNA methylation through SAM and SAH activities or different mechanisms, if available.
Ultimately, most of the studies did not show the biochemical mechanisms of how methyl donors lead to aberrant DNA methylation. They relied heavily on assumptions. It is not clear how reduced levels of SAM or increased levels of SAH were causing global hypomethylation. Is it because there were not enough methyl donors to bind DNMTs? Alternatively, perhaps SAH was inhibiting the entry of the DNA nucleotide cytosine into enzymes’ active sites. It is believed that there is not a simple correlation between methyl donors and DNA methylation. Hence, more studies are warranted to explain the underlying mechanisms in order to contribute to set patterns of DNA methylation in cells.
3.2. What Are the Diet-Related Cofactor and Enzyme Activities in One-Carbon Cycles?
Enzymes taking a role in the folate cycle (MTHFR, MTR, MS, SHMT, etc.) are regulated by micronutrients such as vitamins B2, B6, and B12. It is assumed that supplementing diets with these micronutrients may contribute to the maintenance of DNA methyl marks and therefore regulate DNA methylation [
71]. Additionally, it is believed that variations in the bioavailability of these micromolecules may affect DNA methylation by altering the activity of the one-carbon cycle and the production of SAM [
1].
MTHFR is an essential enzyme for the maintenance of the folate cycle and methylation of CpG islands [
88]. SAM is a strong inhibitor of MTHFR, which also makes it the major regulator of folate-dependent homocysteine remethylation [
89]. MTHFR activity may deteriorate due to an excess concentration of methionine and SAM or polymorphisms, or a low concentration of its cofactor vitamin B2, which decreases the synthesis of 5-methyltetrahydrofolate and thus the remethylation of homocysteine [
90]. Conversely, when SAM concentrations are low and cofactor levels are high, the remethylation of homocysteine may be favored [
89].
Moreover, a reduction of MTHFR activity increases the 5,10-methylenetetrahydrofolate levels while it drops the 5-methyltetrahydrofolate levels, which in return may favor the synthesis of deoxythymidine triphosphate (dTTP) over the methylation of CpG, and therefore alter DNA methylation [
88,
90]. Additional research has suggested that subjects who are homozygous for the polymorphism (
C677T) in the
MTHFR gene exhibited a significantly lower level of methylated DNA, but only under conditions of low folate status [
91]. In tissue culture, a study has shown that folic acid, vitamin B2, and
MTHFR C677T polymorphism affect genome instability, and that high B2 concentration may increase the activity of MTHFR, which may lead folate to provide methyl groups for the methionine synthesis enzyme instead of for thymidylate synthase [
88]. Furthermore, it has been suggested that low vitamin B2 concentration in the presence of low folate may maximize the risk of genome hypomethylation [
88]. However, this study did not measure DNA methylation directly. Instead, it measured several markers related to genome stability and linked it with methylation. Unfortunately, most of the evidence from in vivo studies has not clarified the direct link between folate cycle enzymes or cofactors and DNA methylation. More studies are warranted in order to evaluate the interaction between diet-enzyme activities in the one-carbon cycle and DNA methylation.
3.3. How Diet Affects the DNA Methyltransferase Activity
Li et al. were the first scientists showing the
DNMT1 gene leading to the genome-wide loss of DNA methylation and embryonic lethality in mice [
92]. Numerous other studies later underlined the link between DNMTs and DNA methylation [
93]. Besides, it is believed that those genetic modifications and the DNMT’s activity can be modified by nutritional factors. Animal studies reported that feeding methyl-deficient diets for nine weeks or longer caused DNA hypomethylation, which was associated with the suppressed expression of DNMT1 [
94,
95]. Lillycrop et al. showed a significant decrease of DNA methylation following a protein-restricted diet in pregnant rats, and indicated that altered DNMT1 expression may provide a mechanism for the induction of the hypomethylation of specific genes and individual CpG, although they did not show how such targeting may occur [
96]. In this section, potential nutrient-based epigenetic mechanisms mostly involving the inhibition of DNMTs and altered DNA methylation have been evaluated.
The studies outlined in
Table 1 suggest that several diet compounds may directly affect the expression of DNMT, or that methyl donors from the diet may indirectly modify DNMT activity by changing the intracellular concentration of SAM [
97]. These assumptions have been demonstrated for several bioactive food components such as epigallocatechin-3-gallate (EGCG), genistein, caffeic acid, ascorbate, etc. [
1]. A study found that each of the tea polyphenols (catechin, epicatechin, and EGCG) and bioflavonoids (quercetin, fisetin, and myricetin) inhibited SssI DNMT and DNMT1-mediated DNA methylation in a concentration-dependent manner. EGCG was found to be a more potent inhibitor that had direct inhibitory interaction with the DNMTs and the catalytic site of the human DNMT1. Additionally, when epicatechin was used as a model inhibitor, kinetic analyses indicated that this catechol-containing dietary polyphenol inhibited enzymatic DNA methylation (indirect) in vitro, largely by increasing the formation of SAH. [
98]. Moreover, the treatment of the human esophageal KYSE 510 cell line with EGCG showed a dose and time-dependent reversal of hypermethylation and the re-expression of mRNA of
p16INK4a,
RARβ,
MGMT, and
hMLH1 genes. Reactivation of some methylation-silenced genes by EGCG was also demonstrated in human colon cancer HT-29 cells, prostate cancer PC3 cells, and KYSE cells [
99]. Both studies tried to explain the underlying mechanisms between EGCG and DNMT by using the structural model, molecular docking, and binding energy analysis. They revealed that EGCG shows competitive inhibition of DNMT1 by forming hydrogen bonds within the DNMT1 catalytic-binding region, thus blocking the entry of the DNA nucleotide cytosine into its active site, and inhibiting the methylation process [
98,
99]. Several other studies also revealed that EGCG decreased global DNA methylation levels, and also showed a protective effect by inhibiting the promoter hypermethylation of specific genes. These effects were attributed to the decreased mRNA and protein expression activity of DNMT1 and EGCG inducing the binding domain of DNMT1 to the promoter of the specific genes [
100,
101,
102,
103].
Genistein also showed a dose-dependent inhibitory effect on recombinant DNMT1 activity, and also decreased DNMT activity in nuclear extracts from KYSE cells, but this activity was found to be weaker than that of EGCG. However, six days of genistein treatment did not affect the mRNA expression levels of DNMTs and the methyl-CpG binding domain 2. Although genistein was found to have a synergistic or additive effect on DNMT inhibitors because it is a weak inhibitor of DNMTs, genomic global hypomethylation was not expected to occur after the dietary intake of soy isoflavones [
104]. Another study showed that a genistein diet (300 mg of genistein/kg) was positively correlated with alterations in prostate DNA methylation at CpG islands of specific mouse genes. However, the mechanistic role of genistein was not examined [
105].
Lee et al. revealed the effect of several other catechol polyphenols on DNMT activity. It has been shown that quercetin, fisetin, and myricetin may inhibit DNMT activity by transferring SAM to SAH [
98]. The same group also showed that two common coffee polyphenols, caffeic acid and chlorogenic acid, have inhibited DNA methylation, which was catalyzed by prokaryotic CpG methylase (M.Sssl) DNMT and human DNMT1. The inhibition of DNA methylation by caffeic acid or chlorogenic acid was found to be concentration-dependent, and the inhibition was predominantly through a non-competitive mechanism, which suggested that it was due to the increased formation of SAH [
106]. Eventually, caffeic acid/chlorogenic acid treatment in cultured human breast cancer cells showed no significant change in the global methylation status. However, the concentration-dependent inhibition of DNA methylation in the promoter region of the
RARβ gene was detected, which showed a potential inhibition effect in the promoter region [
106].
Curcumin, an antioxidant component of a spice called turmeric, has been investigated by some study groups for its effect on DNA methylation [
107]. Liu et al. suggested that curcumin covalently blocks the catalytic thiolate of DNMT1 to exert its inhibitory effect on DNA methylation by using molecular docking [
108]. Moreover, a combination of curcumin with the hypomethylating agents increased the response to the drug in breast cancer patients [
109]. However, Medina-Franco et al. suggested that curcumin has no significant effect on DNMT inhibition and global hypomethylation after following a multistep docking approach [
110]. Thus, more studies are required to detect an interaction between curcumin and DNA methylation.
Parthenolide, a component of a plant called feverfew, has been used for the treatment of several diseases. It has been suggested that parthenolide may have a potential role in inhibiting the activity of DNMT1 by blocking the enzyme’s catalytic site, and a study indicated that dose and cell type-dependent parthenolide treatment decreased DNMT1 protein levels and induced a decrease in global DNA methylation. The same study showed that parthenolide inhibited the DNMT1 analog M.SssI by blocking the functional thiolate of the enzyme. Although parthenolide’s binding energy is not as strong as EGCG, it has been suggested that it may be an effective DNA methylation inhibitor [
111].
Mahanine is found in several Asian herbs and species, and it is an alkaloid from the leaves of the curry leaf tree (
Murraya koenigii) and lime berry (
Micromelum minutum). It is mostly studied for its anti-inflammatory and anti-mutagenic activity. [
112,
113]. Mahanine is thought to have an anti-proliferative activity, which was associated with the inhibition of DNMT activity, and hence, may prevent the hypermethylation of a specific gene in the prostate cancer cell line [
114]. However, the mechanisms of action were not clarified.
Eventually, studies evaluating the consumption of polyphenols showed that in general, EGCG and several other polyphenols are promising candidates, especially for future cancer therapies, based on their influence on the epigenetic pathway. Most of these studies showed kinetics and possible mechanisms that alter DNA methylation. These include increasing SAH, inhibiting DNMT’s catalytic base, blocking the promoter sites of specific genes, or covalently binding to thiol groups of enzymes/transcriptional factors. However, future studies evaluating the underlying mechanisms are still needed in order to clarify the pathways of epigenetics.
3.4. Is There a Link between Diet and DNA Demethylation?
The reversal of DNA methylation is crucial, and abnormalities are often observed in anomalies and diseases. Genetic and modifiable factors such as diet may affect the regulation of DNA demethylation, and thus genetic regulations. However, DNA demethylation’s interaction with modifiable factors in mammals is still not well understood [
37]. Recent epigenetic studies have tried to investigate the link between nutrition and active DNA demethylation, which is believed to lead to several modifications in DNA methylation. One study tried to clarify the DNA methylation status of the liver of mice fed the methionine–choline-deficient (MCD) diet (for a week) by measuring the amount of 5mC and investigating the involvement of the active DNA demethylation. The results showed that the expression of DNMT1 and DNMT3a was significantly increased on the MCD diet. In addition, mRNA expression of Tet2 and Tet3 was significantly upregulated on the MCD diet. However, no statistical differences for 5mC content and other demethylation enzymes were found [
115]. It is believed that for better epigenetic investigations, long-term studies are necessary. The deletion of Tet2 was found to cause an extensive loss of 5hmC, which was accompanied by enhancer hypermethylation and delayed gene induction in the early steps of differentiation [
116]. It is assumed that methyl-deficient diets that alter the expression Tet2 may contribute to hypermethylation in specific areas [
115].
Some studies have shown that the presence of ascorbate (vitamin C) may modify the status of DNA methylation [
117,
118]. In embryonic stem cells, ascorbate caused the widespread DNA demethylation of nearly 2,000 genes [
118]. However, it remains unknown whether the effect of ascorbate on DNA demethylation is due to an enhanced hydroxylation of 5mC. A study showed that ascorbate enhances 5hmC generation, most likely by acting as a cofactor for Tet methylcytosine dioxygenase to hydroxylate 5mC in mouse embryonic fibroblasts [
117].
Pogribny et al. evaluated epigenetic changes during hepatocarcinogenesis, which was induced by diets deficient in methyl donors, in his review, and he commented that methyl donors’ deficiency sustains the demethylation of genomic DNA that occurs in methyl-deficient animal’s cytosine in their liver [
87]. Further, the results of past studies have suggested that demethylation may be associated with decreased levels of SAM, increased levels of SAH, a decreased SAM/SAH ratio [
86], and the changed activity of DNMTs [
119]. However, the latest studies have demonstrated that DNA hypomethylation or demethylation induced by methyl-deficient diets might be attributed to the induction of uracil, 5hmC, and 8-oxodeoxyguanosine [
95,
120]. The presence of these products may significantly coordinate with DNMT1 and lead to the demethylation of DNA [
121].
Less is known about the role of nutrition in the base excision repair system. In one of the few studies that has examined the five genes (
SMUG1,
TDG,
UNG,
MBD4, and
DUT) that are involved in the repair system to identify polymorphisms and establish whether one-carbon nutrient status can further alter their effects, single nucleotide polymorphisms in
SMUG1,
DUT, and
UNG genes showed an association with DNA uracil concentration. However, one-carbon nutrient status was not associated with DNA uracil concentration, and did not modify the effect of the single nucleotide polymorphisms [
122]. An older study showed that folate deficiency impairs the DNA excision repair system in rat colonic mucosa [
123], and folate status was found to be associated with uracil misincorporation and genomic instability in humans. However, both studies were not linked to DNA demethylation. Together, the evidence suggests that more studies are required in order to understand the demethylation pathways and the part that dietary factors play in demethylation.