Elevated S-Adenosylhomocysteine Alters Adipocyte Functionality With Corresponding Changes in Gene Expression and Associated Epigenetic Marks

The “predictive adaptive response” in the context of developmental plasticity proposes that the prenatal environment, in particular the levels of maternal nutrition at periconception and during pregnancy, influence the metabolic status and degree of metabolic syndrome of the progeny into adulthood (1,2). This germinal concept has since been reiterated in numerous epidemiological studies that not only lend support to the envisaged role of environment in setting the progeny’s “epigenetic phenotype” but also collectively underpin a contemporary embodiment of the original notion, known as the developmental origins of human health and disease (1,2). Within this growing body of epidemiological data, seminal investigations have linked maternal deficiencies in key micronutrients, such as folate (vitamin B₉) and cobalamin (the cobalt-containing compounds or B₁₂ vitamins), to fetal pathologies, including preterm birth (3), intrauterine growth retardation and birth size (4–6), and neural tube defects (7), and as determinants of postnatal insulin resistance and obesity (8,9). Of significance, these micronutrients comprise the one-carbon cycle, hence highlighting the effect of one-carbon metabolism on fetal growth parameters and the potential contribution to the development of offspring insulin resistance and obesity.

Folate and cobalamins are primarily required by the body as dietary precursors for the generation of methyl donor groups in the one-carbon cycle, as reviewed by Scotti et al. (10). The predominant methyl donor is S-adenosylmethionine (SAM), a metabolic intermediate derived from methionine, and cobalamin is a necessary cofactor in the enzymatic generation of methionine from homocysteine (Hcy). In the course of this process, methionine synthase transfers a methyl group from methyltetrahydrofolate (the native form of folic acid) to Hcy. Thus, the level of Hcy in the circulation is determined by the biological activity of cobalamins and folic acid. After donating its methyl group, SAM is converted to S-adenosylhomocysteine (SAH), which can then be hydrolyzed to Hcy once again to complete the cycle (10). The methyl donor groups are used by a range of trans-methylation–dependent biochemical reactions and in the synthesis of nucleic acids during rapid cellular growth and elevated DNA synthesis. The generation of methyl donors is also critical for DNA and histone methylation, a major epigenetic process implicated in the regulation of gene expression levels. The latter includes modification of specific genomic sites in the gene body or regulatory regions and of histone proteins responsible for packing DNA into a higher order chromatin structure. In both cases, methylation changes are thought to affect the accessibility of regulatory factors required to drive or suppress the complex transcriptional hierarchies needed to define functional phenotypes.

It is generally viewed that the adoption of food fortification and folate supplementation programs has resulted in significant public health benefit, reducing the incidence of neural tube defects by as much as 48% in some countries (11). More recent data, from the U.S. National Health and Nutrition Examination Surveys, in particular, suggest however that these benefits have not been realized universally across socioeconomic and ethnic groups in the U.S. (12) and that specific targeting of high-risk and disadvantaged groups would generate further impact. Beyond the U.S., vitamin B₉ and B₁₂ deficiencies in pregnancy remain a significant public health issue in low-income countries (13) and even in some developed nations such as Japan (14). Notably in Brazilian pregnancy cohorts, women with vitamin B₁₂ deficiencies exhibit significantly decreased serum levels of the methyl donor SAM and methionine, but SAH concentrations are significantly elevated (15).

Furthermore, this maternal vitamin deficiency appears to be perpetuated intergenerationally, leading to a similar deficiency in the neonate. These findings suggest that perturbation of the one-carbon cycle resulting from maternal vitamin deficiency may lead to changes in cellular methylation in the mother and offspring. Further, they suggest a physiological linkage between vitamin deficiency and subsequent patterns of adipogenesis, a process that is known to be significantly influenced by epigenetic mechanisms (16,17) and/or adipocyte function.

Adipogenesis is driven by a gene transcriptional program that underpins the conversion of undifferentiated fibroblastic preadipocytes into spherical lipid-rich mature functioning adipocytes. Mature adipocytes acquire insulin sensitivity, express the insulin-responsive GLUT4 to mediate glucose uptake (18,19), and express enzymes involved in triacylglycerol metabolism to confer lipogenic and lipolytic functions (18). Genes that encode enzymes involved in adipocyte glucose and lipid metabolism are regulated in a highly coordinated manner by at least two families of transcription factors, CAAT enhancer-binding proteins (CEBPs) and the peroxisome proliferator–activated receptors (PPARs) (18). In particular, CEBPα has been implicated in maintenance of the adipocyte phenotype as well as in promoting insulin-sensitive glucose transport (19).

To establish a molecular link between maternal deficiencies in folate and cobalamin to the differentiation and function of adipocytes, we used the 3T3-L1 cell line, a well characterized in vitro model system of adipogenesis, to investigate the consequences of exposure to supraphysiological levels of SAH, mimicking the elevated serum levels observed as a consequence of vitamin deficiencies. Interestingly, although we did not change the rate of adipocyte differentiation, exposure to SAH led to altered functionality, with the resulting mature adipocytes exhibiting altered glucose disposal and lipolysis. Further, we defined some of the key gene regulators of this altered phenotype and established a role for epigenetic mechanisms in its manifestation.

All chemicals were purchased from Sigma-Aldrich, unless otherwise stated. Cell culture reagents were purchased from Gibco-BRL, Invitrogen. We purchased 2-deoxy-d-[1-³H]glucose (37 MBq, 8 Ci/mmol) from PerkinElmer, Waltham, MA.

Murine 3T3-L1 preadipocytes (American Type Culture Collection; http://www.atcc.org) were cultured and induced to differentiate using hormonal cocktails, as previously described (20). For all experiments, preadipocytes were seeded at 2,500 cells/cm². SAH treatment at the indicated doses was initiated 24 h after seeding and maintained throughout differentiation. Adipocytes 7–8 days after induction of differentiation were used for assays. For quantitative RT-PCR (RT-qPCR) and chromatin immunoprecipitation (ChIP) analyses, cells were harvested 5 days after differentiation.

Assay was performed as previously described (20). Cells were stimulated with 100 nmol/L insulin for 20 min at 37°C before initiation of the glucose transport assay.

Assay was performed as described previously with minor modifications (21). Differentiated adipocytes in 96-well plates were fixed in 4% paraformaldehyde (pH 7.4) for 10 min and washed once with PBS, followed by incubation with 10 μg/mL Nile red and 1 μg/mL DAPI for 15 min. Relative fluorescence intensity for Nile red was used as an indicator of lipid accumulation (surrogate for differentiation) and was measured by a BioTek Synergy 2 plate reader using 485/528 and 360/460 filter sets. Cell numbers were normalized using DAPI fluorescence.

Differentiated adipocytes were rinsed with PBS twice and incubated in Dulbecco’s modified Eagle’s medium/2% BSA for 1 h at 37°C in the presence of 10 µmol/L isoproterenol or vehicle. Media were collected, and free glycerol release was measured using Free Glycerol Reagent according to the manufacturer’s instructions. Results were normalized to total intracellular protein content.

For whole-cell lysates, adipocytes were harvested in lysis buffer containing 150 mmol/L NaCl, 0.5% Nonidet P-40, and 20 mmol/L Tris (pH 7.4). Cellular protein concentrations were determined using bicinchoninic assay according to the manufacturer’s instructions (Pierce). Proteins (5–10 µg) were resolved on SDS-PAGE gel and transferred onto nitrocellulose membrane (Bio-Rad). Antibodies in 1% skim milk were used at the following dilutions: α-H3K4me3, 1:1,500; α-H3K27me3, 1:4,000; α-H3K9me3, 1:1,000; α-H3K9Ac, 1:5,000; and α-pan H3-CT, 1:20,000 (Millipore). Results were visualized with horseradish peroxidase–conjugated goat anti-rabbit secondary antibody and enhanced chemiluminescence (Pierce). Immunoblot band intensities were quantified using ImageJ 1.45 software (National Institutes of Health).

Previous reports indicated the importance of stabilizing SAM by acidification (22,23). Hence, we prepared the cell extracts and measured the metabolites using the combined methods of previous reports, with modifications (22–24). Briefly, cells differentiated in the presence or absence of SAH in 10-cm dishes were rinsed in PBS twice and lysed in five volumes of ice-cold acidified acetonitrile (90% acetonitrile containing 0.1% formic acid), then 5 μL of the lysate was frozen for determination of intracellular protein. The remaining lysate was centrifuged at 12,000g for 5 min to clarify the extract, and the supernatant was stored at −80°C for metabolite detection. The samples were further diluted before liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis, which was performed using a Diamond Hydride HILIC column (2.1 × 100 mm, 4 μm; MicroSolv Technology Corporation). The detector was a Q-Exactive Orbitrap operated in the targeted MS/MS mode and hence allowing high-resolution MS/MS, providing greater specificity over that of traditional quadrupole MS/MS protocols (23,24).

Total RNA was extracted using the Trizol method (Invitrogen). The RNA was incubated with DNAseI (Invitrogen), and 1 μg RNA was reverse-transcribed using Superscript III according to the manufacturer’s protocol (Invitrogen). The cDNA was subjected to SYBR Green qPCR (Roche) using gene-specific primers. RNA expression was normalized to Cyclophilin A (Ppia). Results were calculated as the target-to-Ppia ratio using LightCycler 480 software (Roche). ChIP was performed as previously described (25). Sequences for RT-qPCR and ChIP primers are listed in Table 1.

Cells were harvested in lysis buffer containing 150 mmol/L NaCl, 1% SDS, and 20 mmol/L Tris (pH 8.0). Cell lysates were incubated with 0.2 mg/mL Proteinase K at 65°C overnight. RNAse A at 0.2 µg/mL was added for the final 1 h of incubation. Genomic DNA was extracted using the phenol-chloroform method. DNA methylation was measured using the Sequenom Mass Array system (Sequenom, Inc.). Sequenom assay and primer design were performed according to the manufacturer’s protocol. The EZ-DNA Methylation kit (Zymed Laboratories) was used for bisulfite conversion of 1 µg DNA. PCR amplification of bisulfite-converted DNA was performed as follows: 94°C denaturation for 20 s, 58°C annealing for 30 s, and 72°C extension for 1 min, for 45 cycles, and final extension of 72°C for 3 min. Primer sequences are listed in Table 2. MS analysis was performed using EpiTyper 1.0 software (Sequenom, Inc.). For fragments containing a single CpG site, DNA methylation was calculated by the ratio of methylated to unmethylated fragments. For cleavage products containing multiple CpGs, EpiTyper reports the methylation values as weighted averages across the fragments.

Data were analyzed using IBM SPSS Statistics 20 software. Unless indicated in the figure legends, results were analyzed using one-way ANOVA with Bonferroni post hoc test at α = 0.05. A value of P < 0.05 was considered statistically significant.

To assess the effects of SAH on adipogenic differentiation capacity, preadipocytes were cultured and differentiated into mature adipocytes in the absence or presence of SAH. Total intracellular protein and cell numbers were assessed at indicated days of differentiation (Fig. 1A). 3T3-L1 preadipocytes are known to undergo two to three rounds of cell division (i.e., mitotic clonal expansion) during the midphase of differentiation (18). We found that total cell number increased by threefold 3 days after differentiation (D3) compared with cells at the onset of differentiation (D0) in control (untreated) and SAH-exposed cells (Fig. 1A). Furthermore, control and SAH-exposed cells exhibited similar increase in total intracellular protein (1.5- to 2-fold at D3, and 5- to 6-fold at D7; Fig. 1B). Preadipocyte factor-1 (Dlk1) is downregulated during the normal process of adipogenesis (26). We found that Dlk1 expression declined significantly (P = 0.000) after differentiation in SAH-exposed adipocytes, similar to the untreated controls (Fig. 1C). These data indicate that persistent SAH exposure in differentiating preadipocytes did not alter adipogenesis per se.

After preadipocyte culture and differentiation in the presence of SAH, glucose uptake, lipid accumulation, and lipolysis were measured. Although basal glucose uptake is mainly mediated by GLUT1, preadipocytes acquire insulin sensitivity upon adipogenesis and exhibit insulin-stimulated glucose uptake via GLUT4 (27). As expected in the absence of SAH treatment, mature adipocytes exhibited a robust increase in glucose uptake upon stimulation with 100 nmol/L insulin (Fig. 1D). However, adipocytes exposed to 10 µmol/L and 100 µmol/L SAH reduced basal glucose uptake by 45% (P < 0.01) and 30% (P = 0.03), respectively (Fig. 1D). Insulin-stimulated glucose uptake was reduced by 40% at 10 µmol/L (P = 0.02) and 100 µmol/L (P = 0.004) SAH. In addition, the reduced glucose uptake appears to be independent of reduced GLUT1/4 expression and translocation to the plasma membrane, because GLUT1/4 levels in whole-cell extracts, as well as plasma membrane fractions under basal and insulin-stimulated states, from adipocytes with and without SAH exposure, were similar (data not shown).

SAH exposure (10 µmol/L and 100 µmol/L) did not alter intracellular lipid store (Fig. 1E). As expected, adipocytes in the absence of SAH showed a robust increase in isoproterenol-stimulated lipolysis (Fig. 1F). Basal lipolysis was significantly reduced by 70% (P = 0.03) in adipocytes treated with 10 μmol/L SAH and by 20% (albeit not significant) at 100 μmol/L SAH. Isoproterenol-stimulated lipolysis was significantly reduced by 50% (P = 0.01) and 20% (P = 0.03) at 10 μmol/L and 100 μmol/L SAH, respectively, compared with adipocytes without SAH exposure, suggesting a dose-dependent effect of SAH on lipolysis.

Given that SAH perturbed adipocyte function but not differentiation, we measured RNA levels of several genes involved in adipogenic function. Owing to budget limitation, we chose the candidate approach (RT-qPCR) rather than a genome-wide or global method. We found that expression of Pparγ2, a key regulator of adipogenesis (18), was not changed by SAH (Fig. 2). This finding is consistent with our earlier results that persistent SAH exposure does not inhibit adipogenesis per se. Furthermore, the expression of retinoic acid x receptor-α (Rxrα), a nuclear receptor that heterodimerizes with Pparγ2 to regulate glucose and lipid metabolism in adipocytes (28), was significantly reduced by 50% in (100 μmol/L) SAH-treated cells compared with untreated controls (P = 0.01; Fig. 2). Cebpα and Cebpβ expression, both known to play crucial roles in the maintenance of adipogenesis (19,29), were reduced by 90% (P = 0.000) and 80% (P = 0.01), respectively, at 100 μmol/L SAH compared with adipocytes without exposure. Perhaps unsurprising, we found that Kruppel-like factor-4 (Klf4) expression, which induces Cebpβ expression (30), was also significantly inhibited by 80% at both 10 μmol/L and 100 μmol/L SAH (P = 0.000).

Given that adipogenic differentiation is regulated epigenetically (16,17), we investigated whether SAH-mediated inhibition of genes expression is associated with changes in DNA methylation. After adipogenic differentiation in the absence or presence of SAH, genomic DNA was extracted, bisulfite-converted, and Klf4, Cebpα, Cebpβ, and Rxrα genomic regions were assessed for changes in CpG methylation. We found that CpG methylation was unchanged for Klf4 (Fig. 3A) and Cebpβ (Fig. 3B) promoter regions, despite the SAH-mediated reduced expression of these genes. Owing to the nature of the sequence surrounding the CpG site of interest containing repeats, we could not measure Cebpα methylation; this amplicon failed PCR amplification despite our numerous attempts (data not shown). We found that Rxrα methylation on two of five CpG residues within intron 1, proximal to the promoter, were relatively unchanged in SAH-exposed and nonexposed adipocytes (chr2:27,521,169+ and chr2:27,521,208+; Fig. 3C). In contrast, we detected a significant 10% increase in methylation of a CpG unit containing two CpG residues (chr2:27,521,057+, chr2:27,521,049+; P = 0.003) and on a single CpG residue at chr2:27,521,080+ (P = 0.02) in adipocytes exposed to 10 µmol/L and 100 µmol/L SAH compared with adipocytes in the absence of SAH.

We next examined whether histone modifications are altered in adipocytes maintained in a persistently elevated SAH milieu. H3-K27 trimethylation (H3K27me3) was suppressed by fourfold (P = 0.02) in differentiated adipocytes compared with undifferentiated preadipocytes without SAH treatment (Fig. 4A). After SAH exposure, the H3K27me3 level was unchanged in preadipocytes but was significantly increased in differentiated adipocytes by fourfold at 10 µmol/L and 100 µmol/L SAH (P = 0.02) compared with differentiated adipocytes without SAH treatment. Similarly, the H3K4me3 level was suppressed by fivefold (P = 0.04) in the absence of SAH during normal differentiation. SAH exposure maintained elevated level of H3K4me3 after differentiation at 10 µmol/L (P = 0.05) and 100 µmol/L SAH. In contrast, we found that SAH did not alter H3K9me3 or acetylation of H3-K9 (H3K9Ac) in the absence or presence of SAH, suggestive of a selective and specific effect of SAH on the modification profile present on H3 during or after differentiation.

In light of these findings, we questioned whether the SAH-mediated decrease in gene expression is associated with changes in the repressive histone H3K27me3 mark. After cell proliferation and 5 days of differentiation in the presence or absence of 100 µmol/L SAH, adipocytes underwent ChIP for H3K27me3. The amount of H3K27me3-associated Klf4, Cebpα and Rxrα genes expression were then determined. Adipocytes exposed to prolonged 100 µmol/L SAH demonstrated a marked increase in H3K27me3 occupancy at the promoter of Klf4 and Cebpα genes by fourfold compared with adipocytes without SAH exposure (P = 0.000 and P = 0.01, respectively; Fig. 4B). Similarly, H3K27me3 occupancy was significantly increased at the promoter of Rxrα by twofold in adipocytes exposed to 100 µmol/L SAH compared with adipocytes without SAH exposure (P = 0.003).

To determine whether the metabolic alterations in these adipocytes are contributed by SAH, we measured intracellular SAH, SAM and Hcy, and SAH in the media collected at the end of the treatment regimen. In media containing exogenous 10 μmol/L and 100 μmol/L SAH, we detected 0.5 ± 0.1 μmol/L and 7.0 ± 0.9 μmol/L SAH, respectively (mean ± SD; Fig. 5A). Media from untreated controls were below the LC-MS/MS limit of detection of 0.02 μmol/L (Fig. 5A). Our model reflects a physiologically relevant state given that the effective exogenous SAH concentrations used in this system are well within an order of magnitude to the previously reported circulating plasma SAH levels in human (15). We found that the intracellular SAH in adipocytes exposed to exogenous 10 µmol/L and 100 µmol/L SAH was elevated by at least two- and fivefold, respectively, over the untreated controls (Fig. 5B). Interestingly, we detected only modest increases in intracellular SAM and Hcy, at 1.3- and 1.4-fold over untreated controls for SAM and 1.3- and 1.6-fold over untreated controls for Hcy at 10 μmol/L and 100 μmol/L exogenous SAH, respectively (Fig. 5B). These results are consistent with our hypothesis that exposure to elevated SAH may induce metabolic changes in adipocytes. In addition, our data suggest that increasing intracellular SAH alone does not necessarily drive a change in the one-carbon cycle flux (i.e., methyl donor availability) given the lack of an appreciable change in intracellular SAM and Hcy.

In summary, persistent SAH exposure in differentiating preadipocytes reduces glucose uptake and lipolysis in differentiated adipocytes. The SAH-mediated reduction in adipocyte functionality likely occurs through reduced expression of nodal regulators of transcriptional pathways such as Klf4, Cebpα/β, and Rxrα, which are involved in glucose and lipid metabolism. The reduced expression of these genes appears to be epigenetically regulated by increased promoter occupancy of H3K27me3, and in the case of Rxrα, perhaps also through changes in DNA methylation.

Increasing evidence suggests that the maternal nutritional state can influence adipose tissue biology in the offspring (31–33). In this study, we found that adipocytes that were differentiated in vitro in the presence of elevated SAH express an altered adipogenic functionality and gene expression, although preadipocyte proliferation, rates of differentiation, and adipocyte cell numbers appear to be normal. Notably, the differentiated adipocytes exhibit impaired glucose uptake and lipolysis. Consistent with these impairments, we demonstrate significant SAH-induced changes in the expression of pivotal regulators of adipocyte function. Reduced Rxrα expression would alter a number of effector genes involved in adipocyte function (28). Notably, the loss of Rxrα would affect the formation of the heterodimer complex with PPARγ (28). The other critical transcriptional factor affected by SAH is CEBPα, which has been shown to promote insulin-sensitive glucose transport (19,34). Significantly, changes in histone modifications appear to be of primary importance in mediating these nutritionally driven effects on adipocyte function.

The primary role of fat tissue is to uptake and store excess glucose in the form of lipid droplets and recover this energy in times of need through the process of lipid breakdown. Both of these key functions appear to be disrupted in adipocytes differentiated in the presence of elevated levels of the one-carbon cycle intermediate SAH. Our finding that persistent SAH exposure impairs lipolysis implies that conditions that perturb one-carbon metabolism in differentiating preadipocytes may contribute to the lowering of released free fatty acids by suppressing lipolysis in these adipocytes once matured. Consistent with this notion, Li et al. (35) reported that multivitamin supplementation in a group of Chinese obese/overweight cohort improved the subjects’ lipid profile. Of significance, the ability of SAH to impair adipocyte glucose uptake, in particular under the resting state, implicates the potential adverse impact of changes in one-carbon metabolism on adipose glucose disposal. This is of huge physiological relevance given that prediabetes is a high-risk state for diabetes, especially in individuals who remain prediabetic despite intensive lifestyle intervention and that reversion to normal glucose regulation significantly reduces the risk of future diabetes (36). One might speculate that a high SAH level predicts an improvement in lipid profile, thence the metabolic status, in the subjects. However, how the impaired glucose uptake may contribute to impairment of the overall glucose disposal in subjects with elevated SAH remains unclear. How changes in one-carbon metabolism (particularly through nutritional variation) modulate adipocyte glucose disposal warrants further investigation.

Nutrient imbalances can lead to changes in the epigenetic state of tissues through genome-wide alteration of DNA methylation and chromatin chemistry (16,17,31). For example, that the depletion of folate results in hypomethylation of genomic DNA in both human (37) and animal models (38) has been appreciated for some time. DNA methylation is, however, a dynamic and reversible process, such that reinstating a folate-enriched diet is sufficient to restore the normal DNA methylation status of tissues (39). The principal metabolic role of micronutrients such as folate is to generate the one-carbon cycle intermediate and universal methyl donor SAM. Not surprisingly, more direct disruption of nutrient-dependent SAM-derived methyl group generation also results in hypomethylation of genomic DNA (40). However, changes in “global” genomic methylation do not necessarily translate to altered promoter methylation states at specific loci (41,42), demonstrating that the mechanisms underlying DNA methylation are tightly directed and regulated rather than passive. Indeed, we found that the SAH-mediated suppression of gene expression at selective loci of key adipogenic genes was not generally associated with changes in promoter DNA methylation. Notably, only Rxrα gene repression was accompanied by significantly increased methylation at specific CpG sites. Given that we have recently shown that the promoter methylation status of RXRA is predictive of future adiposity in retrospective human cohort studies (43), the current findings are consistent with and extend the notion that Rxrα gene methylation is a significant determinant of fat cell biological function.

The other widely recognized epigenetic mechanism for gene regulation involves the highly choreographed chemical modification of specific histone residues, which results in changes to chromatin structure and thus influences transcription factor access and binding. In particular, the histone modification H3K27me3 has been widely linked to the repression of transcription elongation (44), leading to inhibition of gene expression. In adipocytes undergoing differentiation in the presence of elevated SAH, we found significantly enhanced promoter occupancy by H3K27me3 in several relevant target genes (Rxrα and Klf4), suggesting that repression of these genes was occurring through histone-dependent mechanisms rather than by direct promoter DNA methylation, a caveat being that few CpG sites at the promoter of these genes were investigated. Interestingly, we also observed increased H3K4me3 in mature adipocytes differentiated under the same conditions. Although methylation is generally linked to gene repression, methylation of the H3K4 residue is considered to be a permissive mark that enhances the initiation of transcription (45). Indeed, quantitative modeling of the genome-wide distribution of the H3K4me3 mark suggests that it may even be highly predictive of expression at promoters that have a low CpG content (46). In many human cells, H3K4me3 occurs as a part of a “bivalent” chromatin structure in which its role as a primer of transcriptional activity is held in check by the presence of a more dominant repressive mark such as H3K27me3 (45). Our observations suggest that adipocyte gene repression and the resultant shift in adipocyte biology resulting from elevated SAH exposure may occur through the enhanced formation of a bivalent chromatin complex.

Our findings are consistent with those of previous studies showing that changes in SAH level alone, rather than SAM, drives changes in SAM-to-SAH ratios in various tissues (47,48), suggesting a regulatory role for SAH in determining overall levels of cellular methylation. Methylation potential could presumably be altered by influencing the availability of methyl donor groups and/or modifying the activities of various types of methyltransferases that catalyze the transfer of the methyl group from SAM to a variety of biomolecules, including histones (49). Our study suggests that elevating intracellular SAH alone does not necessarily alter global methyl donor availability. However, because SAH is an established and potent inhibitor of DNA and histone methyltransferase activities (49), we conclude that exposure to elevated SAH drives phenotypic change through its capacity to inhibit cellular methylation.

In conclusion, the current study provides evidence that nutritionally driven signaling through one-carbon metabolism can influence metabolic adaptive plasticity. Specifically, although not appearing to alter adipocyte differentiation in vitro, it may influence the functionality of the mature adipocyte and does so via epigenetic mechanisms.

Acknowledgments. The authors thank Leticia Castro (Liggins Institute) for providing technical support, Tony Pleasants (AgResearch) for assistance with statistical analysis, and Phillip Shepherd (Liggins Institute) for proofreading this manuscript and are also grateful for the useful and informative comments of the anonymous reviewers.

Funding. This work was supported by the Maurice and Phyllis Paykel Fellowship (grant 9118/3626657) and in part by the National Research Centre for Growth and Development (grant 9128/3625204).

The funders played no role in the conception or design of this study, data collection and analysis, decision to publish, or preparation of this manuscript.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.N. conceived the study, undertook the molecular biology and data analysis, wrote the first draft of the manuscript, and reviewed and edited the manuscript. X.L., R.O., and C.B. undertook the molecular biology analysis and assisted with data analysis. P.G. reviewed and edited the manuscript. A.S. conceived the study and reviewed and edited the manuscript. S.N. is the guarantor of this work and, as such, takes full responsibility for the work as a whole, including the study design, access to data, and the decision to submit and publish the manuscript.

Prior Presentation. Parts of Figure 1 and all of Figure 2 were presented as a poster at the 71st Scientific Sessions of the American Diabetes Association, San Diego, California, 24–28 June 2011.