Plasma Homocysteine and Polymorphisms of Genes Involved in Folate Metabolism Correlate with DNMT1 Gene Methylation Levels

DNA methyltransferase 1 (DNMT1) is responsible for the maintenance of DNA methylation patterns during cell division. Several human diseases are characterized by impaired DNMT1 gene methylation, but less is known about the factors that regulate DNMT1 promoter methylation levels. Dietary folates and related B-vitamins are essential micronutrients for DNA methylation processes, and we performed the present study to investigate the contribution of circulating folate, vitamin B12, homocysteine, and common polymorphisms in folate pathway genes to the DNMT1 gene methylation levels. We investigated DNMT1 gene methylation levels in peripheral blood DNA samples from 215 healthy individuals. All the DNA samples were genotyped for MTHFR 677C > T (rs1801133) and 1298A > C (rs1801131), MTRR 66A > G (rs1801394), MTR 2756A > G (rs1805087), SLC19A1 (RFC1) 80G > A (rs1051266), TYMS 28-bp tandem repeats (rs34743033) and 1494 6-bp insertion/deletion (indel) (rs34489327), DNMT3A -448A > G (rs1550117), and DNMT3B -149C > T (rs2424913) polymorphisms. Circulating homocysteine, folate, and vitamin B12 levels were available from 158 of the recruited individuals. We observed an inverse correlation between plasma homocysteine and DNMT1 methylation levels. Furthermore, both MTR rs1805087 and TYMS rs34743033 polymorphisms showed a statistically significant effect on DNMT1 methylation levels. The present study revealed several correlations between the folate metabolic pathway and DNMT1 promoter methylation that could be of relevance for those disorders characterized by altered DNA methylation.


Introduction
DNA methylation is one of the most investigated epigenetic mechanisms that tightly regulate gene transcription and expression levels, and it consists of the addition of a methyl group to the DNA, mediated by a class of enzymes called DNA methyltransferases (DNMTs) [1]. DNA methylation usually occurs on cytosine, forming 5-methylcytosine (5-mC) in a cytosine-phosphate-guanine (CpG) dinucleotide context, and CpG methylation in the promoter region of a gene represses gene expression [2]. The human genome encodes five DNMTs, namely, DNMT1, DNMT2, DNMT3A, DNMT3B, and DNMTL [3]. Among them, DNMT1 is the "maintenance" DNMT that maintains DNA methylation patterns during cell division, while DNMT3A and DNMT3B are responsible for the establishment of de novo DNA methylation patterns during embryogenesis. DNMT3L is not an active Figure 1. Simplified overview of the folate (light brown color) and methionine (heavenly color) cycles in one-carbon metabolism, adapted from Reference [4]. The diagram illustrates the enzymes (green color) whose polymorphisms were investigated in this article, and their metabolites or cofactors (those analyzed in the present study are in red). Enzymes: DNMTs, DNA methyltransferases; MTHFR, methylenetetrahydrofolate reductase; MTR, methionine synthase; MTRR, methionine synthase reductase; RFC1, reduced folate carrier 1; TYMS, thymidilate synthase. Metabolites: DHF, dihydrofolate; THF, tetrahydrofolate; dTMP, deoxythymidine monophosphate; dUMP, deoxyuridine monophosphate; SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine.
As discussed above, DNMT1 is the main enzyme responsible for the maintenance of DNA methylation during cell division, and previous studies revealed the existence of a CpG island in the promoter/5' untranslated region (5'UTR) of the DNMT1 gene itself, whose methylation levels regulate DNMT1 gene expression in both healthy and pathological tissues [5][6][7][8][9][10][11][12]. In this regard, increased DNMT1 promoter methylation resulting in decreased gene expression levels was observed in blood DNA of individuals with ankylosing spondylitis, a chronic inflammatory autoimmune disease [5]. Similarly, a decreased DNMT1 expression was observed in blood DNA of individuals with Kawasaki disease, an acute vasculitis syndrome [6], whilst increased DNMT1 expression was observed in blood DNA from patients with psoriasis [7]. Changes in DNMT1 promoter methylation Figure 1. Simplified overview of the folate (light brown color) and methionine (heavenly color) cycles in one-carbon metabolism, adapted from Reference [4]. The diagram illustrates the enzymes (green color) whose polymorphisms were investigated in this article, and their metabolites or cofactors (those analyzed in the present study are in red). Enzymes: DNMTs, DNA methyltransferases; MTHFR, methylenetetrahydrofolate reductase; MTR, methionine synthase; MTRR, methionine synthase reductase; RFC1, reduced folate carrier 1; TYMS, thymidilate synthase. Metabolites: DHF, dihydrofolate; THF, tetrahydrofolate; dTMP, deoxythymidine monophosphate; dUMP, deoxyuridine monophosphate; SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine.
As discussed above, DNMT1 is the main enzyme responsible for the maintenance of DNA methylation during cell division, and previous studies revealed the existence of a CpG island in the promoter/5 untranslated region (5 UTR) of the DNMT1 gene itself, whose methylation levels regulate DNMT1 gene expression in both healthy and pathological tissues [5][6][7][8][9][10][11][12]. In this regard, increased DNMT1 promoter methylation resulting in decreased gene expression levels was observed in blood DNA of individuals with ankylosing spondylitis, a chronic inflammatory autoimmune disease [5]. Similarly, a decreased DNMT1 expression was observed in blood DNA of individuals with Kawasaki disease, an acute vasculitis syndrome [6], whilst increased DNMT1 expression was observed in blood DNA from patients with psoriasis [7]. Changes in DNMT1 promoter methylation were also observed in peripheral blood DNA samples from patients with acute lymphoblastic leukemia (ALL) and were linked to the DNMT1 expression pattern in the patients [8]. Furthermore, DNMT1 expression levels in white blood cells of healthy individuals showed a large inter-individual variability [9]. Hypermethylation of the Dnmt1 promoter was also responsible for reduced gene expression levels in various tissues of animal models of asthma [10]. The investigation of DNMT1 methylation, expression, and protein levels in solid tumors showed a large variability according to the investigated cell line or to the cancer subtype [9,[11][12][13].
Little is still known concerning genetic, dietary, or environmental factors linked to changes in DNMT1 methylation levels. In this regard, a previous investigation in peripheral blood DNA samples of Alzheimer's disease (AD) individuals suggested that the promoter methylation levels of several genes, including DNMT1, could be linked to circulating markers of folate metabolism, and particularly to plasma homocysteine (Hcy) levels [14].

Distribution of the Investigated Variables in the Study Population
The demographic characteristics of the study population and the average methylation levels of the DNMT1 gene are shown in Table 1. In total, 215 healthy Italian subjects, including 85 males and 130 females of mean age 76.4 years, composed the study cohort. Plasma Hcy, serum folate, and vitamin B12 levels were available from 158 individuals, and their average values are shown in Table 1.  The promoter methylation levels of the DNMT1 gene ranged from 0% to 15% in the study population ( Figure 2), with an average value of 2.2% (Table 1), which is in agreement with previous investigations in other cohorts [14,20,21]. The genotype distribution of the investigated polymorphisms is shown in Table 2. Genotype frequencies are in agreement with those previously observed in healthy Caucasians [19,22], and conformed to Hardy-Weinberg equilibrium (HWE) expectations (p > 0.05).   Figure 3 shows the correlation between DNMT1 promoter methylation and circulating folate, Hcy, and vitamin B12 levels. A significant inverse correlation between DNMT1 promoter methylation and plasma Hcy levels was observed (r = −0.23, p = 0.008). No correlation was observed between DNMT1 promoter methylation and serum folate levels (r = 0.01, p = 0.85), whilst a trend for correlation with serum vitamin B12 was observed, although not statistically significant (r = 0.11, p = 0.19). The genotype distribution of the investigated polymorphisms is shown in Table 2. Genotype frequencies are in agreement with those previously observed in healthy Caucasians [19,22], and conformed to Hardy-Weinberg equilibrium (HWE) expectations (p > 0.05). Table 2. Distribution of genotypes in the study population; indel-insertion/deletion.  Figure 3 shows the correlation between DNMT1 promoter methylation and circulating folate, Hcy, and vitamin B12 levels. A significant inverse correlation between DNMT1 promoter methylation and plasma Hcy levels was observed (r = −0.23, p = 0.008). No correlation was observed between DNMT1 promoter methylation and serum folate levels (r = 0.01, p = 0.85), whilst a trend for correlation with serum vitamin B12 was observed, although not statistically significant (r = 0.11, p = 0.19).  Figure 4 shows the contribution of gender to circulating folate, Hcy, and vitamin B12 levels, as well as to the methylation levels of the DNMT1 gene. We observed significantly higher Hcy levels in males than in females (p = 0.002), whilst vitamin B12 levels were higher in females with respect to males (p = 0.02), and a similar trend was observed for folates, even if the gender difference was not statistically significant (p = 0.11). No gender difference was observed for DNMT1 methylation levels (p = 0.97) (Figure 4).  Figure 4 shows the contribution of gender to circulating folate, Hcy, and vitamin B12 levels, as well as to the methylation levels of the DNMT1 gene. We observed significantly higher Hcy levels in males than in females (p = 0.002), whilst vitamin B12 levels were higher in females with respect to males (p = 0.02), and a similar trend was observed for folates, even if the gender difference was not statistically significant (p = 0.11). No gender difference was observed for DNMT1 methylation levels (p = 0.97) (Figure 4). The contribution of age to the studied variables is shown in Figure 5. A positive correlation was observed between Hcy levels and age (r = 0.37, p < 0.0001). No significant correlation was observed between age and the other variables (folate, vitamin B12, and DNMT1 methylation levels).  The contribution of age to the studied variables is shown in Figure 5. A positive correlation was observed between Hcy levels and age (r = 0.37, p < 0.0001). No significant correlation was observed between age and the other variables (folate, vitamin B12, and DNMT1 methylation levels). The contribution of age to the studied variables is shown in Figure 5. A positive correlation was observed between Hcy levels and age (r = 0.37, p < 0.0001). No significant correlation was observed between age and the other variables (folate, vitamin B12, and DNMT1 methylation levels).

DNMT1 Methylation Levels and Polymorphisms of the Folate Pathway Genes
Analysis of variance (ANOVA) in the whole cohort of subjects revealed a statistically significant contribution of both the MTR 2756A > G and the TYMS 28-bp repeat polymorphisms to DNMT1 promoter methylation levels. In particular, a significant increase in DNMT1 methylation levels was observed in carriers of the MTR mutant allele (AG + GG) with respect to wild-type (AA) subjects (p = 0.01), and the TYMS 28-bp repeat polymorphism had a significant effect on DNMT1 methylation levels in our cohort, with carriers of the 3R/3R genotype showing significantly higher methylation levels than 2R/2R carriers (p = 0.03). None of the other polymorphisms showed a significant contribution to DNMT1 methylation levels ( Figure 6). Overlapping results were observed in the subgroup of 158 subjects with available data on folate, Hcy, and vitamin B12 levels (not shown).

DNMT1 Methylation Levels and Polymorphisms of the Folate Pathway Genes
Analysis of variance (ANOVA) in the whole cohort of subjects revealed a statistically significant contribution of both the MTR 2756A > G and the TYMS 28-bp repeat polymorphisms to DNMT1 promoter methylation levels. In particular, a significant increase in DNMT1 methylation levels was observed in carriers of the MTR mutant allele (AG + GG) with respect to wild-type (AA) subjects (p = 0.01), and the TYMS 28-bp repeat polymorphism had a significant effect on DNMT1 methylation levels in our cohort, with carriers of the 3R/3R genotype showing significantly higher methylation levels than 2R/2R carriers (p = 0.03). None of the other polymorphisms showed a significant contribution to DNMT1 methylation levels ( Figure 6). Overlapping results were observed in the subgroup of 158 subjects with available data on folate, Hcy, and vitamin B12 levels (not shown).

Discussion
In the present study, we investigated the contribution of circulating folate, Hcy, and vitamin B12, as well as of major polymorphisms of the folate pathway genes, as modulators of DNMT1 promoter methylation levels in blood DNA from healthy individuals, observing an inter-individual variability in DNMT1 promoter methylation levels, which ranged from 0% to 15% in the studied cohort, a significant inverse correlation with plasma Hcy levels, and a trend for a positive correlation with

Discussion
In the present study, we investigated the contribution of circulating folate, Hcy, and vitamin B12, as well as of major polymorphisms of the folate pathway genes, as modulators of DNMT1 promoter methylation levels in blood DNA from healthy individuals, observing an inter-individual variability in DNMT1 promoter methylation levels, which ranged from 0% to 15% in the studied cohort, a significant inverse correlation with plasma Hcy levels, and a trend for a positive correlation with vitamin B12 levels. Furthermore, carriers of the MTR 2756G allele (AG + GG) showed a significantly higher DNMT1 promoter methylation than non-carriers (AA), as well as carriers of the TYMS 28-bp repeat 3R/3R genotype with respect to 2R/2R carriers. None of the other investigated polymorphisms were linked to the methylation status of the DNMT1 promoter. Age and gender had no significant effect on DNMT1 promoter methylation levels in our cohort.
Present data confirm previous investigations of the methylation status of the DNMT1 promoter in blood DNA samples, all reporting low methylation levels in average, albeit with inter-individual variability [14,20,21]. Furthermore, we confirmed previous findings obtained in a cohort of 100 Alzheimer's disease individuals and matched controls, showing that elevated plasma Hcy levels in the AD subjects were linked to reduced methylation levels of several genes in blood DNA, including DNMT1 [14]. Homocysteine is a metabolite for methionine production, and methionine is the precursor of SAM that serves as the universal one-carbon donor for DNA methylation reactions (Figure 1), such that hyperhomocysteinemia (HHcy) is often linked to reduced DNA methylation potential [23], and the epigenetic dysregulation of gene expression is a pathogenic consequence of HHcy in many human diseases [24]. In this context, previous in vitro studies revealed that Hcy treatment altered the expression levels of DNMT1, leading to global or gene-specific DNA methylation changes [25][26][27]. For example, Hcy treatment enhanced DNMT1 expression in oocytes leading to hypermethylation of the mitochondrial DNA [25], upregulated DNMT1 expression leading to DNA hypermethylation of the mitofusin 2 gene promoter in vascular smooth muscle cells [26], or induced DNA hypomethylation of the cyclin A promoter in endothelial progenitor cells through downregulated expression of DNMT1 [27]. A mechanistic link between Hcy and DNA methylation is the resulting reduction of the S-adenosylmethionine/S-adenosylhomocysteine ratio following HHcy, and the subsequent accumulation of S-adenosylhomocysteine, which in turn inhibits methyl transfer reactions; however, it was also suggested that Hcy can alter the binding of transcription factors to the DNMT1 promoter [25][26][27]. The present study suggests that Hcy levels correlate with DNMT1 promoter methylation levels, thus strengthening the contribution of this metabolite to the regulation of DNA methylation.
Original findings of the present study are the correlations between DNMT1 promoter methylation levels and both MTR 2756A > G and TYMS 28-bp repeat polymorphisms. Particularly, we observed increased DNMT1 methylation levels in carriers of the MTR 2756G allele compared to the wild-type MTR 2756AA genotype. The MTR gene encodes for methionine synthase, the enzyme that catalyzes the remethylation of Hcy to methionine, and cobalamin (vitamin B12) is a cofactor in this reaction ( Figure 1). The MTR 2756A > G polymorphism impairs MTR function and stability [28,29] and is often linked to either global or gene-specific DNA methylation changes in human cells and tissues [18,[30][31][32][33]. For example, the MTR 2756A > G polymorphism was linked to increased global DNA methylation levels in human leukocytes [30], and carriers of the variant G allele had a significant increase of LINE-1 (long interspersed nuclear element 1) methylation in histologically normal breast tissues, compared to those carrying the common AA genotype [18]. Other studies reported association of this polymorphism with global methylation levels of leukocyte DNA [31], with MTHFR methylation levels in blood DNA of valproate-treated patients with epilepsy [32], and with hypermethylation of tumor suppressor genes in cancer specimens [33] or adjacent healthy tissues [17]. The present finding of a correlation between the MTR 2756A > G polymorphism and DNMT1 promoter methylation levels in blood DNA not only strengthens the contribution of this genetic variant to the DNA methylation levels in leukocytes, but it could also be of relevance for those disorders, such as hematological malignancies, autoimmune/inflammatory disorders, and solid tumors, that are linked to impaired DNMT1 methylation and/or expression [5][6][7][8][9][10][11][12]. For example, changes in DNMT1 methylation levels were observed in peripheral blood DNA samples from ALL patients [8], the MTR 2756A > G polymorphism was linked to the risk of pediatric ALL [34], and several maternal polymorphisms of the MTR gene were recently proposed as responsible of aberrant methylation of ALL-related genes in their offspring [35]. Furthermore, the MTR 2756A > G polymorphism was also linked to the risk of autoimmune/inflammatory diseases and solid tumors, all characterized by impaired DNMT activities and global and gene-specific DNA methylation changes [36][37][38][39], suggesting that the investigation of the contribution of this polymorphism to DNMT1 methylation levels is warranted in all these disorders. TYMS and MTHFR compete for 5,10-methylenetetrahydrofolate in such a way that the folate pathway can be shifted from DNA methylation to the synthesis of DNA precursors, based on cellular demands (Figure 1). In this regard, the 28-bp tandem repeat polymorphism, located in the 5 UTR of the TYMS gene, works as enhancer for TYMS transcription, with higher efficiency conferred by the 3R allele with respect to the 2R one. This enhanced expression minimizes uracil misincorporation into the DNA and increases the availability of DNA precursors, such that this polymorphism is often associated with cellular growth and risk of human malignancies, including leukemias, lymphomas, breast cancer, and thoracic neoplasms, among others [40][41][42][43]. Less is known about the contribution of TYMS polymorphisms to the DNA methylation levels, even if the TYMS 28-bp tandem repeat polymorphism was associated with gene-specific methylation levels in the blood DNA of patients systemic lupus erythematosus [44], and other TYMS polymorphisms were linked to either global or gene-specific methylation levels in both healthy and cancerous tissues, reinforcing the evidence that DNA synthesis and methylation are interconnected pathways [17,[45][46][47]. The present findings of a correlation between the TYMS 28-bp tandem repeat polymorphism and DNMT1 promoter methylation levels in blood cells highlight the contribution of this gene to DNA methylation, and suggest that further investigation is warranted in human disorders associated to this variant.
The main limitation of the present study was that we investigated an elderly population of limited sample size that, despite healthy at collection, could include individuals with age-related subclinical conditions potentially impacting the findings. Therefore, results should be considered preliminary, and replication in a larger cohort and in younger subjects is warranted. Moreover, the observed DNMT1 methylation levels were low on average, and their inter-individual variability could result from subclones of cells with high levels of DNMT1 methylation in some individuals. Further investigations of single-cell DNA methylation levels are warranted to clarify this issue and their potential correlation with preclinical conditions in humans.
In conclusion, we investigated the contribution of both metabolites and genetic polymorphisms in the folate metabolic pathway to the DNMT1 promoter methylation levels in blood DNA from healthy individuals, observing an inverse correlation with plasma Hcy levels, and significant associations for both MTR 2756A > G and TYMS 28-bp tandem repeat polymorphisms. The present study highlights the links existing between the folate metabolic pathway and the promoter methylation levels of the maintenance DNMT, suggesting that further studies are required to address this issue in human disorders characterized by changes in DNMT1 promoter methylation, such as hematological malignancies, solid tumors, and autoimmune/inflammatory diseases.

Study Population
We collected peripheral blood samples from 215 healthy individuals, including 85 males and 130 females of mean age 76.4 ± 9.6 years ( Table 1). All individuals were volunteer subjects of Italian origin, underwent a rigorous clinical and neurological examination before inclusion in the study, and were healthy at blood drawing. In addition, individuals taking drugs, substances, or vitamin supplements known or suspected to interfere with DNA methylation, such as anti-cancer drugs, anti-epileptic drugs, anti-inflammatory drugs, epi-drugs, metformin, tobacco smoking, folic acid, or other vitamins, were not enrolled in the study. Written informed consent for inclusion in the study was collected from each subject. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the Pisa University Hospital (Protocol number 3618/2012).

Collection of DNA Samples
Genomic DNA was extracted from peripheral blood samples using the QIAamp®Blood Mini Kit (Catalog No. 51104, Qiagen, Milan, Italy) following the manufacturer's instructions. DNA samples were stored at −20 • C until assayed.

Biochemical Analyses
The measurement of plasma Hcy, serum vitamin B12, and folate levels was performed with the standard diagnostic laboratory methodologies of the Pisa University Hospital, as detailed elsewhere [48].

Analysis of DNMT1 Methylation Levels
In total, 200 ng of DNA from each sample was quantified using a Nanodrop ND 2000c spectrophotometer (NanoDrop, Thermo Scientific, Wilmington, DE, USA). The DNA was subsequently treated with sodium bisulfite, using the EpiTect Bisulfite Kit (Catalog No. 59104, Qiagen, Milan, Italy), to convert unmethylated cytosines into uracil. All the samples were treated simultaneously in order to avoid potential batch effects. The bisulfite conversion efficiency, assessed using a sample of completely unmethylated human DNA (Catalog No. 59568, Qiagen, Milan, Italy), was 99% on average. For the analysis of DNMT1 methylation levels, we applied the methylation-sensitive high-resolution melting (MS-HRM) technique, using a protocol previously developed, validated, and fully described by us elsewhere [51]. All the analyses were performed in a CFX96 Real-Time PCR detection system (Bio-Rad, Milan, Italy). We investigated a CpG island in the promoter/5 UTR region of the DNMT1 gene, whose methylation levels were previously linked to DNMT1 gene expression levels [5,8]. Table 3 lists the sequence of the primers and the details of the investigated region. Each reaction was performed in duplicate, and 10% of the samples were analyzed independently on separate occasions to verify the inter-assay variability. The mean standard error for inter-assay variability was 0.1%, indicating a very good technical reproducibility. We mixed fully methylated and unmethylated DNA (EpiTectH methylated and unmethylated human control DNA, bisulfite-converted (Catalog No. 59695, Qiagen, Milan, Italy), to obtain the following ratios of methylation: 0%, 25%, 50%, 75%, and 100%. These standard DNA samples were included in each assay to generate standard curves ( Figure 2) that were used to deduce the methylation levels of each sample, as detailed elsewhere [52]. Table 3. Primers and annealing temperature (T a ) used during methylation-sensitive high-resolution melting (MS-HRM) analysis, as well as amplicon length, region analyzed respect to the Transcription Start Site (TSS), number of cytosine-phosphate-guanine (CpG) sites, accession number, and nucleotide position of the DNMT1 region analyzed. F-forward; R-reverse.

Statistical Analyses
The chi-square (X 2 ) analysis was used to investigate deviations from Hardy-Weinberg equilibrium. Linear regression analysis was used to investigate the correlation between DNMT1 promoter methylation levels and plasma Hcy, serum folate, or vitamin B12 levels. Similarly, linear regression was applied to investigate the correlation of these variables with age. Analysis of variance (ANOVA) was used to evaluate gender differences in DNMT1 promoter methylation levels and plasma Hcy, serum folate, or vitamin B12 levels. ANOVA, including age at sampling and gender as covariates (and also plasma Hcy, serum folate, and vitamin B12 levels in the subgroup of patients with available metabolites), was used to investigate the contribution of each of the studied polymorphisms to DNMT1 methylation levels, followed by post hoc Bonferroni's correction for multiple testing. The Shapiro-Wilk test was used to check for normality, and natural logarithm transformation of data that did not follow a normal distribution (DNMT1 methylation levels, plasma Hcy, serum folate, and vitamin B12) was done before the analyses. Statistical analyses were performed with the STATGRAPHICS 5.1 plus software package for Windows and the MedCalc statistical software v. 12.5. We considered Bonferroni's corrected p-values < 0.05 as statistically significant.