Dietary Perturbation of One - carbon Metabolism
Methylation of DNA (and histones) is intimately depend-ent on nutrition. The methyl groups that are incorporated into DNA ultimately come from the diet (for example from methionine, choline, and betaine) and many nutri-ents, including folate, vitamin B 12 , and vitamin B 6 , are
FIG. 2.2 The mammalian transmethylation pathway. Several nutrients including folate (shown here as 5 - methyltetrahydrofolate), vitamin B 12 , vitamin B 6 , choline, betaine, and methionine serve as either methyl donors or cofactors in the transmethylation pathway. DNA is one of many biological substrates (indicated here as “ X ” ) for methyltransferase reactions which transfer methyl groups from S - adenosylmethionine (SAM).
Glycine N - methyltransferase (reaction 7) is distinguished from other methyltransferases to emphasize its apparent role in regulating intracellular SAM concentrations.
Methionine Adenosyltransferase
blood DNA was inversely correlated with cord plasma homocysteine. The combined effects of dietary and genetic promotion of hyperhomocysteinemia by cystathione β synthase (Cbs) haploinsuffi ciency (see Figure 2.2 ) were studied in mouse liver, kidney, brain, and testis (Caudill et al. , 2001 ). Despite dramatic differences in plasma total homocysteine concentrations, however, in most tissues global DNA methylation did not differ between wild type and Cbs+ /− mice. Similarly, human patients with CBS defi -ciency showed no changes in either global or locus - specifi c DNA methylation (Heil et al. , 2007 ), despite exception-ally elevated plasma homocysteine concentrations. A small study of male patients with hyperhomocysteinemia and uremia, however, did fi nd reduced genomic DNA meth-ylation in patients relative to controls (Ingrosso et al. , 2003 ); patients also displayed biallelic expression of H19 in peripheral blood mononuclear cells. ( H19 is a genomi-cally imprinted gene, at which monoallelic expression is normally regulated by allele - specifi c methylation at a differentially methylated region [DMR]). Remarkably, in three patients who initially showed strong biallelic expression of H19 , 60 days of treatment with oral 5 - methyltetrahydrofolate (15 mg/day) restored normal monoallelic expression (Ingrosso et al. , 2003 ). The effects of hyperhomocysteinemia on epigenetic regulation of H19 were subsequently examined in a mouse model; Cbs hap-loinsuffi ciency combined with a hyperhomocysteinemic diet resulted in a 20 - fold elevation in plasma total homo-cysteine concentrations in adult mice (Devlin et al. , 2005 ).
As expected, DNA methylation at the H19 DMR was signifi cantly reduced in the liver of hyperhomocysteinemic mice relative to controls. Paradoxically, however, in aorta and brain, H19 DNA methylation was signifi cantly ele-vated (dramatically so in brain).
Choline, required for the synthesis of phosphatidylcho-line and acetylchophosphatidylcho-line, is also an important methyl donor (Zeisel, 2006 ) (also, see Chapter 26 ). Choline appears to be particularly important for brain development (Zeisel, 2006 ); prenatal choline supplementation in rats leads to enhanced cognitive functioning in adulthood (Meck and Williams, 2003 ). In mice, maternal choline defi ciency initi-ated in mid - pregnancy reduced DNA methylation at spe-cifi c genes in the fetal hippocampus at 17 days of fetal development (Niculescu et al. , 2006 ). It remains unclear, however, the extent to which the long - term cogni-tive benefi ts of choline supplementation are attributable to effects on epigenetic mechanisms such as DNA methy-lation. Following choline supplementation during fetal various forms of polyglutaminated pteridine - p -
aminoben-zoic acid, some of which (such as 5 - methyltetrahydrofolate) are methyl donors. “ Folic acid ” , on the other hand, is a synthetic form of folate that is not a dietary methyl donor (Smith et al. , 2008 ); ingested folic acid must be reduced and methylated (probably in the liver) before it can serve as a cofactor in the methionine synthase reaction (Figure 2.2 ) (also, see Chapter 21 ). An early study demonstrated that interindividual genetic variation interacts with folate status to determine DNA methylation. Healthy adults were assessed for plasma folate concentrations and periph-eral blood global DNA methylation, and genotyped for the C677T polymorphism in the gene encoding methyl-enetetrahydrofolate reductase (MTHFR). Plasma folate was positively correlated with global DNA methylation, but only in individuals homozygous for the T variant of MTHFR (which causes reduced enzymatic activity) (Friso et al. , 2002 ). This gene – nutrient – epigenome interaction was corroborated in a subsequent folate depletion – repletion study (Shelnutt et al. , 2004 ). Recent data indi-cate that maternal supplementation with folic acid before and during pregnancy may affect epigenetic regulation in the offspring. Methylation of Long Interspersed Elements (LINE - 1) (an indicator of global DNA methylation) in cord blood DNA at delivery was slightly elevated in off-spring whose mothers took folic acid supplements (Steegers - Theunissen et al. , 2009 ).
Since folate is a critical cofactor in one - carbon metabo-lism, DNA methylation is generally expected to correlate positively with folate status. This is not always the case. In a large epidemiologic study, dietary folate intake was inversely correlated with LINE - 1 methylation in colonic epithelium (Figueiredo et al. , 2009 ); the association was strongest in DNA from the ascending colon. Similarly, in a rat model, feeding a folate - free synthetic diet from age 21 – 56 days induced a persistent increase in hepatic global DNA methylation (Kotsopoulos et al. , 2008 ). One poten-tial explanation for this unexpected fi nding is that dietary methyl donor defi ciency induces dramatic increases in expression of the DNA methyltransferases Dnmt1 and Dnmt3a in rat liver (Ghoshal et al. , 2006 ).
Elevated plasma homocysteine is associated with both nutritional and genetic factors, and is a risk factor for vascular disease (Selhub, 2006 ). Because homocysteine is a secondary product of biological methylation reactions (Figure 2.2 ), elevated homocysteine should impair DNA methylation. Indeed, in a small study of ∼ 20 newborn infants (Fryer et al. , 2009 ), LINE - 1 methylation in cord
ured; in most cases DNA methylation was higher in famine - exposed individuals (Tobi et al. , 2009 ).
The effects of maternal undernutrition before and during pregnancy have been studied extensively in rodents.
Providing a low - protein diet (9% protein compared with 18% control) to mothers before and during pregnancy induces persistent physiological changes in the offspring.
Effects of maternal low - protein diet on epigenetic regula-tion in the offspring are now being explored. Global DNA methylation in liver (but not heart or kidney) of day 21 fetal rats was reported to be elevated in response to mater-nal low - protein diet (Rees et al. , 2000 ). Focusing on the promoter of the Ppar α gene, it was reported that specifi c CpG sites in hepatic DNA show reduced methylation in rat offspring of mothers fed a low - protein diet (Lillycrop et al. , 2008 ). Although the average methylation in this region was less than 10%, the epigenetic consequences of maternal diet were persistent; the same CpG sites showed reduced methylation in low - protein offspring at 34 and 80 days of age. Epigenetic effects of prenatal undernutri-tion have also been investigated in sheep (Stevens et al. , 2010 ). Ewes undernourished during the fi rst month of pregnancy were subsequently allowed free access to feed until their fetuses were collected at fetal day 130. Despite the 100 - day recuperation period, fetuses undernourished during early fetal development had increased expression of glucocorticoid receptor in the hypothalamus, associated with reduced DNA methylation and increased histone 3 lysine 9 (H3K9) acetylation of the glucocorticoid receptor promoter.
In a rat model, fetal growth restriction by intrauterine artery ligation cause permanent impairment of glucose tolerance. The developmental dynamics of epigenetic changes induced by intrauterine growth restriction were characterized at the Pdx1 promoter in isolated pancreatic islets (Park et al. , 2008 ). Pdx1 is a homeobox transcription factor that regulates pancreas development and β cell dif-ferentiation. In pancreatic islets of 14 - day - old rats that had been growth restricted in utero , reduced Pdx1 expression was associated with lower density of several histone modi-fi cations at the Pdx1 promoter, but no differences in DNA methylation. By age 6 months, however, extensive DNA methylation at the Pdx1 promoter was found only in islets of previously growth - restricted rats (Park et al. , 2008 ), leading the authors to propose that aberrant epigenetic silencing at Pdx1 is initiated by histone modifi cation and subsequently stabilized by DNA methylation. Epigenetic changes associated with intrauterine growth retardation development, adult rats were found to have increased
hip-pocampal neurogenesis (Glenn et al. , 2007 ), suggesting that early choline supplementation may induce persistent epi-genetic changes in the neurogenic stem cell population.
The substantial body stores of creatine and phospho-creatine, which play central roles in energy metabolism, are spontaneously lost to creatinine and must be continuously synthesized (Brosnan and Brosnan, 2007 ). Creatine syn-thesis consumes about 40% of all S - adenosylmethionine in young adults (Brosnan and Brosnan, 2007 ), a huge methyl sink. Creatine supplementation therefore has the potential to facilitate DNA methylation by sparing S - adenosylmethionine. A recent study in adult rats, however, reported just the opposite. After just 14 days of supplementation with creatine monohydrate (2% wt/wt in diet), global DNA methylation in peripheral blood was signifi cantly reduced (Taes et al. , 2007 ).
Prenatal Energy Restriction/Growth Retardation
Retrospective studies of individuals exposed to famine in early life have provided extensive insights into the poten-tial for nutrition during critical periods of development to induce persistent changes in human metabolism and disease risk. A recent study of survivors of the Dutch famine of 1944 – 1945 tested the hypothesis that maternal exposure to famine leaves a permanent epigenetic mark on the offspring (Heijmans et al. , 2008 ). In peripheral blood from individuals almost 60 years old, bisulfi te sequencing was used to measure DNA methylation at a DMR within the imprinted IGF2 gene. Recognizing the potential for genetic variation to obscure induced epigenetic variation, IGF2 DMR methylation in each periconceptionally exposed individual was compared with his or her same - sex unexposed sibling. The IGF2 DMR was slightly but sig-nifi cantly less methylated in individuals exposed to famine periconceptionally, relative to unexposed siblings, provid-ing the fi rst convincprovid-ing demonstration that an environ-mental exposure during early development can cause a persistent epigenetic alteration in humans (Heijmans et al. , 2008 ). Famine exposure during late fetal development was not associated with IGF2 methylation, indicating that the period of susceptibility for this effect is limited to early embryonic development. The same group subsequently reported a study of DNA methylation at 15 candidate genes implicated in metabolic and cardiovascular disease (Tobi et al. , 2009 ). Periconceptional exposure was signifi -cantly associated with methylation at six of 15 genes
meas-humans is from microbial digestion of dietary fi ber in the colon (Wachtershauser and Stein, 2000 ); butyrate is the dominant energy source for colonic mucosal enterocytes.
In rats fed diets of various wheat bran contents for 2 weeks, global histone H4 acetylation in colonic epithelial cells was directly correlated with butyrate concentration in the colonic lumen (Boffa et al. , 1992 ), providing an intriguing potential epigenetic link between dietary fi ber intake and risk of colon cancer. Another dietary HDAC inhibitor, sulforaphane, an isothiocyanate found in cruciferous veg-etables, is also implicated in cancer prevention (Ho et al. , 2009 ). A recent in vitro study (Meeran et al. , 2010 ) showed that sulforaphane inhibits the proliferation of breast cancer cells but not normal breast cells. Further, sulforaphane caused a dose - responsive increase in DNA methylation and epigenetic silencing of the gene encoding telomerase reverse transcriptase, offering a potential explanation for how sulforaphane specifi cally affects cancer cells.