Early life environment and future disease risk

The association between the quality of the early life environment and subsequent risk for chronic disease in later life has been well described. Such studies have shown that low birth weight even within the normal range is associated with an increased risk for CVD and the metabolic syndrome (hypertension, insulin resistance, type 2 diabetes, dyslipidaemia and obesity) in later life.Reference Godfrey and Barker ⁴ Although in many of these studies, a J- or U-shaped relationship between birth weight and disease risk was seen, with babies born with the highest birth weight also being at increased risk for disease.Reference Curhan, Willett and Rimm ^{5 ,} Reference McCance, Pettitt and Hanson ⁶ However, birth weight in these studies is thought to be only a very crude indicator of the intrauterine environment, which may have been compromised through a variety of maternal, environmental or placental factors.Reference Hanson and Gluckman ⁷

One maternal factor that has consistently been shown to influence later phenotype is maternal nutrition. This has been most clearly shown in studies of the Dutch Hunger Winter, a famine, which occurred in the Netherlands during the winter of 1944. These studies have shown that individuals whose mothers were exposed to famine periconceptually and in the first trimester of pregnancy exhibit an increased risk of obesity and CVD in adulthood, whereas individuals whose mothers were exposed in the later stages of gestation showed increased incidence of insulin resistance and hypertension in later life.Reference Roseboom, de and Painter ⁸

These findings from human epidemiological studies, which show an association between the quality of early life environment and future disease risk, have been replicated in a variety of animal models where nutrition in particular both before and after pregnancy can be precisely controlled. Early animal studies focussed on the effects of maternal undernutrition or protein restriction attempting to replicate the effects of famine on long-term offspring health, although more recently given the growing epidemic of obesity in both industrialized and developing countries, animal models have been established to investigate the effect of energy-rich diets during pregnancy on the phenotype of the offspring. Interestingly, offspring born to dams fed these different diets exhibited persistent metabolic changes often leading to features similar to human cardio-metabolic disease such as obesity, insulin resistance, hypertension and raised serum cholesterol levels.Reference Bertram and Hanson ^{9 ,} Reference Samuelsson, Matthews and Argenton ¹⁰ For instance, feeding rats a protein-restricted diet (PR) during pregnancy has been reported to result in impaired glucose homoeostasis,Reference Burns, Desai and Cohen ¹¹ vascular dysfunction,Reference Torrens, Poston and Hanson ¹² impaired immunity,Reference Calder and Yaqoob ¹³ increased susceptibility to oxidative stress,Reference Langley, Seakins, Grimble and Jackson ¹⁴ increased fat deposition and altered feeding behaviour,Reference Bellinger, Lilley and Langley-Evans ^{15 ,} Reference Bellinger, Sculley and Langley-Evans ¹⁶ whereas offspring from dams fed an obesogenic diet (16% fat, 33% sugar) also produced offspring with impaired glucose homoeostasis, increased resistance artery endothelial dysfunction, hypertension, increased adiposity and reduced locomoter activity.Reference Samuelsson, Matthews and Argenton ¹⁰ Interestingly, the metabolic alterations induced in the offspring by perturbations in the maternal diet are dependent on the timing of the nutritional challenge, which may reflect the critical phase of development of different organ systems, with respect to the timing of the nutritional constraint. Bertram et al. Reference Bertram and Hanson ¹⁷ have shown in the guinea pig model that female offspring born to dams fed a PR diet in the first half of pregnancy (1–35 days) have raised mean arterial blood pressure, which was associated with an increased intraventricular septum and anterior left ventricle wall (LVW) thickness. In contrast, the offspring from dams fed a PR diet in late gestation (36–70 days) were growth restricted but did not display alterations in blood pressure or LV structure. Animal studies have also shown an interaction between prenatal and postnatal environments.Reference Ozanne and Hales ^{18 ,} Reference Zambrano, Bautista and Deas ¹⁹ For example, dyslipidaemia and impaired glucose homoeostasis induced by feeding dams a PR diet during pregnancy were exacerbated in adult male and female rats fed a diet containing 10% (w/w) fat after weaning compared with a 4% (w/w) fat post-weaning diet.Reference Zambrano, Bautista and Deas ¹⁹

Interestingly, a number of recent studies have shown that paternal diet can also influence future disease risk in the offspring. Carone et al. Reference Carone, Fauquier and Habib ²⁰ showed that offspring of males fed a low-protein diet from weaning until sexual maturity exhibited an increase in the expression of many genes involved in lipid and cholesterol biosynthesis in the liver as well as differences in the levels of several lipid metabolites, whereas Ng et al.Reference Ng, Lin and Laybutt ²¹ have shown that a paternal high-fat diet (HFD) exposure induces increased body weight, adiposity, impaired glucose tolerance and insulin sensitivity in female offspring. The mechanism by which alterations in early life nutrition may induce such long-term changes in metabolism and phenotype has been suggested to involve the altered epigenetic regulation of genes.

Introduction to epigenetics

Conrad Waddington (1905–1975) first introduced the term epigenetics in the 1940s to describe the process by which the same genotype can give rise to multiple phenotypes during development.Reference Waddington ²² However, in recent years, the definition of epigenetics has subsequently been narrowed to ‘the study of mitotically and/or meiotically heritable changes in gene function that cannot be explained by changes in DNA sequence’.Reference Bird ²³ Epigenetic processes include DNA methylation, histone modifications and non-coding RNAs (ncRNAs). Together these processes determine when and where a gene is switched on and the level of gene activity. Epigenetic processes not only play a key role in regulating gene expression but are also central to the mechanism, which allow an organism to adapt to the environment.

DNA methylation

DNA methylation is a common modification in eukaryotic organisms. DNA methylation occurs primarily on the fifth position of cytosine (5mC) within a CpG (cytosine and guanine nucleotides linked by phosphate) dinucleotide, although non-CpG methylation is prevalent in embryonic stem cells.Reference Ramsahoye, Biniszkiewicz and Lyko ²⁴ Low levels of methylation at the promoter regions are associated generally with transcriptional activity, whereas high levels of methylation are associated with transcriptional silencing.Reference Bird ²³ DNA methylation induces transcriptional silencing by preventing the binding of transcription factors to the DNA or by recruiting the methyl CpG binding protein 2 (MeCP2) to the DNA, which in turn recruits histone-modifying complexes.Reference Fuks, Hurd and Wolf ²⁵

Methylation of CpGs is largely established during embryogenesis or during early postnatal life. Following fertilization, DNA methylation markers on the maternal and paternal genomes are largely erased with the exception of the imprinted genes. This is followed by global de novo methylation just before blastocyst implantation,Reference Santos, Hendrich, Reik and Dean ²⁶ during which 70% of CpGs are methylated, mainly in the repressive heterochromatin regions and in repetitive sequences such as retrotransposable elements. Lineage-specific methylation of tissue-specific genes also occurs throughout prenatal development and early postnatal life, which is essential for cell specification. The de novo methylation of DNA is catalysed by DNA methyltransferases (DNMT) 3a and 3b,Reference Okano, Bell, Haber and Li ²⁷ and is maintained through mitosis by methylation of the hemi-methylated DNA by DNMT1.Reference Bacolla, Pradhan, Roberts and Wells ²⁸ Once DNA methylation patterns are established during development, these markers were thought to be relatively stably and generally maintained throughout life. This premise was supported by the initial failure to identify a DNA demethylase or a mechanism for DNA demethylation.Reference Ooi and Bestor ²⁹ However, this concept has now been challenged as active demethylation and has been demonstrated in a number of studies. For example, active demethylation in mammals has been observed on the paternal genomic DNA in the zygote upon fertilization,Reference Reik, Santos, Mitsuya, Morgan and Dean ³⁰ on the synaptic plasticity gene reelin in the hippocampus upon contextual fear conditioningReference Miller and Sweatt ³¹ and on the interferon gamma (IFNγ) gene upon antigen exposure of memory CD8 T cells.Reference Kersh, Fitzpatrick and Murali-Krishna ³² Moreover, a number of potential DNA demethylases have been identified such as ten–eleven translocation (Tet) proteins,Reference Tahiliani, Koh and Shen ^{33 ,} Reference Ito, D'Alessio and Taranova ³⁴ methyl binding domain protein (MBD)2b,Reference Bhattacharya, Ramchandani, Cervoni and Szyf ³⁵ MBD4,Reference Zhu, Zheng and Angliker ³⁶ the DNA repair endonucleases XPG (Gadd45a)Reference Barreto, Schafer and Marhold ³⁷ and a G/T mismatch repair DNA glycosylase,Reference Jost ³⁸ although these function by not removing the methyl group directly from cytosine, but through a multi-step processes linked either to DNA repair mechanisms or through further modification of 5mC.

Histone modifications

DNA methylation works in concert with histone modifications to regulate gene expression. In eukaryotic cells, DNA is wrapped around a core of eight histone proteins, two molecules of histone H2A, H2B, H3 and H4. This forms the basic unit of chromatin, called a nucleosome. Each nucleosome is then folded upon itself to form a solenoid or 30-nm fibre, which is then further coiled and compacted to form a 200-nm fibre. This folding is necessary to reduce the effective size of DNA; however, it is has now become very clear that the histones also play a critical role in regulating gene expression. The amino terminal tail domain of the histones are subjected to a large number of modifications including acetylation, methylation, ubiquitination, sumoylation and phosphorylation.Reference Turner ³⁹ The establishment of these markers on the histone tails is often referred to as the histone code and leads to the binding of effector proteins that, in turn, bring about specific cellular processes. Histone acetylation, induced through histone acetyl transferases, is associated with transcriptional activity and an open chromatin state. Methylation of lysine (K) residues within the histone tail can either be an active or repressive mark depending on the specific lysine involved. For example, histone H3K4 methylation is associated with gene activation,Reference Strahl, Ohba, Cook and Allis ⁴⁰ whereas histone H3K9 methylation is linked to gene silencing.Reference Nakayama, Rice, Strahl, Allis and Grewal ⁴¹

Crosstalk between DNA methylation and histone modification is well established. Methylated CpGs are bound by MeCP2, which then recruits both histone deacetylases (HDACs) – which remove acetyl groups from the histones – and histone methyl transferases, this results in a closed chromatin structure and transcriptional silencing.Reference Fuks, Hurd and Wolf ²⁵ Recent studies have also shown that a number of histone modifying enzymes such as HDAC1 and HDAC2,Reference Fuks, Hurd, Deplus and Kouzarides ^{42 –} Reference Rountree, Bachman and Baylin ⁴⁴ as well as components of the polycomb repressive histone complex 2,Reference Vire, Brenner and Deplus ⁴⁵ can recruit DNMT1 to the DNA and induce DNA methylation.

ncRNA

Data from the ENCODE project have shown that, although over 74% of the eukaryotic genome is transcribed, only 1–2% of the genome encodes for proteins,Reference Djebali, Davis and Merkel ⁴⁶ suggesting that a large proportion of the transcripts in the cell do not encode for proteins. These RNAs have been termed ncRNAs and are grouped into two classes. Long ncRNAs are longer than 200 nucleotides, whereas short ncRNAs are <200 nucleotides. Short ncRNAs include microRNAs (miRNAs), small interfering RNAs and piwi-interacting RNAs.Reference Bartel ⁴⁷ Small ncRNAs can induce mRNA degradation and/or translational repression,Reference Bartel ⁴⁷ and when targeted to the promoter region of a gene they can induce both DNA methylation and repressive histone modifications.Reference Tan, Zhang and Wu ⁴⁸ The mechanism by which long ncRNAss work is less well understood but may involve the recruitment of histone modifying complexes to the DNA to create repressive domains covering many kilobases. In mammals, long ncRNAs have been shown to be essential for X-chromosome inactivation and genomic imprinting.Reference Mercer, Dinger and Mattick ⁴⁹

Early life nutrition and the epigenome

Although DNA methylation was originally thought to be a very stable modification and once established methylation patterns were largely maintained throughout the life-course, there is now growing evidence that a number of environmental factors such as nutrition, stress, placental insufficiency, endocrine disruptors and pollution, especially in early life, can alter the epigenome leading to long-term phenotypic changes in the offspring.Reference Feil and Fraga ⁵¹ In this study, we will focus on the effect of early life nutrition on the epigenome and its long-term effects.

One of the best examples of how nutrition can alter phenotype through the altered epigenetic regulation of genes is seen in studies on the honeybee. Female larvae incubated in the presence of Royal Jelly predominantly develop in to Queen bees, whereas those incubated in the absence of Royal Jelly develop into sterile worker bees, although they are genetically identical.Reference Maleszka ⁵² However, knockdown of DNMT3, the major DNMT in bees, increased the proportion of larvae developing into queen bees as opposed to sterile workers,Reference Kucharski, Maleszka, Foret and Maleszka ⁵³ suggesting that nutrition can profoundly affect developmental fate and it does so through the altered methylation of DNA.

Nutrition has also been shown to influence DNA methylation in rodents. In agouti viable yellow (A^vy) mice, coat colour is determined by the methylation status of an intracisternal-A particle (IAP) in the 5′ upstream region of the agouti gene that acts as a cryptic promoter directing expression of the agouti gene, which encodes for a paracrine signalling protein that induces follicular melanocytes to switch from producing black eumelanin to yellow phaeomelanin. Supplementation of the diet of the pregnant A^vy mice with folic acid, cobalamin, choline and betaine induced a graded shift in coat colour of the litter from predominately yellow (agouti) to brown (pseudo-agouti).Reference Tobi, Lumey and Talens ⁵⁴ This shift was accompanied by the hypermethylation of seven CpG dinucleotides, 600 bp downstream of the A^vy IAP insertion site.

There is also evidence in models of nutritional programming where perturbations in maternal diet are associated with persistent metabolic changes in the offspring that these metabolic changes are accompanied by epigenetic changes in key metabolic genes or genes involved in appetite control. For example, feeding pregnant rats a PR diet induced hypomethylation of the glucocorticoid receptor (GR) and peroxisome proliferator activated receptor (PPAR)α promoters in the livers of juvenile and adult offspring; this was accompanied by an increase in GR and PPARα expression and in the metabolic processes that they control.Reference Lillycrop, Phillips, Jackson, Hanson and Burdge ^{55 –} Reference Burdge, Phillips, Dunn, Jackson and Lillycrop ⁵⁷ In contrast, global dietary restriction during pregnancy increased the level of DNA methylation of PPARα and the GR in the liver of the offspring,Reference Gluckman, Lillycrop and Vickers ⁵⁸ suggesting that the effects of maternal nutrition on the epigenome of the offspring depend upon the nature of the maternal nutrient challenge. Such nutrition-related responses are consistent with the concept that induced epigenetic changes that underpin physiological change may provide a means of adapting to an adverse environment.Reference Gluckman, Hanson and Spencer ⁵⁹

Feeding a HFD during pregnancy and lactation has also been shown to induce epigenetic and phenotypic changes in the offspring. Vucetic et al. Reference Vucetic, Kimmel, Totoki, Hollenbeck and Reyes ⁶⁰ showed increased expression of the μ-opioid receptor and preproenkephalin in the nucleus accumbens, prefrontal cortex and hypothalamus of mice from dams that consumed HF diet during pregnancy, and this was accompanied by the hypomethylation of the promoter regions of these genes. Maternal obesity and diabetes has also been reported to induce subtle but widespread changes in DNA methyaltion in the liver of the offpring.Reference Li, Young and Maloney ⁶¹

Although most studies to date have concentrated on identifying changes in DNA methylation associated with perturbations in maternal diet or body compoisiton, there is growing evidence that early life nutrition can also induce persistant changes in both histone modifications and miRNAs. In islet cells, a decrease in hepatic nuclear factor (HNF) 3a expression in the liver of PR offspring was accompanied by a minimal change in DNA methylation but substantial changes in histone modifications at the HNF4a promoter, such as a reduction in H3K4 methylation and an increase in H3K9me2 and H3K27me3.Reference Sandovici, Smith and Nitert ⁶² Dietary restriction from E10 to E21, which induced an increase in IGF1A and IGF1B (insulin growth factors) expression, was also accompanied by changes in the histone markers at the IGF1 gene with a decrease in H3K4me2 and an increase in H3K4me3 being observed.Reference Tosh, Fu and Callaway ⁶³ Marked changes in miRNA expression have also been documented in both the liver and skeletal muscles of the offspring in response to matenral undernutrition during either the peri-conceptual or the pre-implanatation period. Reference Lie, Morrison and Williams-Wyss ^{64 ,} Reference Lie, Morrison and Williams-Wyss ⁶⁵

Alterations in paternal diet in rodents have also been associated with epigenetic changes in the offspring. Feeding male rats a PR diet before mating induced widespread changes in DNA methylation (10–20%) in the liver of the offspring compared with controls, including an increase in methylation at an intergenic CpG island 50 kb upstream of the PPARα gene.Reference Carone, Fauquier and Habib ²⁰ Interestingly, however, cytosine methylation patterns were highly correlative in the sperm from control, low-protein or calorie-restricted fathers, suggesting that the sperm epigenome may be more refractory to differences in diet. A decrease in H3K27 me3 levels at the monoamine oxidase (MAOA) and elongation factor Tu GTP binding domain containing 1 (EFTUD1) promoters were seen, however, in the sperm from the low-protein-fed fathers compared with controls, suggesting that differences in diet may be transmitted through modifications in the histone code.Reference Carone, Fauquier and Habib ²⁰ Paternal high-fat feeding has been reported to induce the hypomethylation of the interleukin 13 receptor alpha 2 (IL13RA2) promoter in the islet cells of female offspring.Reference Ng, Lin and Laybutt ²¹ Altered expressions of 23 miRNAs were also detected in the testis of the high-fat-fed fathers, suggesting that miRNAs may play a role in the transmission of obesity and impaired metabolic health to the offspring.

Epigenome is also susceptible to nutritional factors in later life

Although many studies have shown that the epigenome is highly sensitive to environmental factors during the perinatal and prenatal period, the period of epigenetic plasticity may extend into postnatal life. Plagemann et al. Reference Plagemann, Harder and Brunn ⁶⁶ showed that neonatal overfeeding induced by raising rat pups in small litters induces the hypermethylation of two CpG dinucleotides within the pro-opiomelanocortin promoter. Folic acid supplementation in the juvenile-pubertal period has also been shown to induce the hypermethylation of the PPARα gene; this was accompanied by a decrease in PPARα expression and levels of fatty acid β-oxidation,Reference Burdge, Lillycrop and Phillips ⁶⁷ whereas Ly et al. Reference Ly, Lee and Chen ⁶⁸ showed that folic acid supplementation during the peri-pubertal period led to an increased risk for mammary adenocarcinomas and a decrease in DNMT activity.

There is also evidence that even in adulthood there is some plasticity of the epigenome. For instance, feeding a diet deficient in methyl donors to post-weaning mice led to the permanent loss of IGF2 imprinting and dysregulation of its mRNA expression,Reference Waterland, Lin, Smith and Jirtle ⁶⁹ whereas feeding a diet deficient in choline, folate, methionine and vitamin B12 for 4 weeks to adult rats induced the hypomethylation of the proto-oncogenes c-Myc, c-Fos and c-Ras. Moreover, this effect persisted for 3 weeks after re-feeding.Reference Christman, Sheikhnejad, Dizik, Abileah and Wainfan ⁷⁰

Macronutrient intake during adulthood can also influence the epigenome. Hoile et al. Reference Hoile, Irvine and Kelsall ⁷¹ showed that feeding adult rats fish oil-enriched diet for 9 weeks led to a transient decrease in the expression of the fatty acid desaturase 2 (FADS2) gene and increase in FADS2 promoter methylation. These effects were reversed by feeding a standard diet for a further 4 weeks. Such findings suggest that, although the epigenome may be most susceptible to environmental factors during early life when methylation patterns are being established, there is also some plasticity in later life. Although these effects may be more transient, such continued plasticity in later life does offer the opportunity for intervention to reverse or reset developmentally induced epigenetic markers.

Evidence that early nutrition can alter the epigenome in humans

Studies in animal models, where the diet pre- and post-pregnancy, as well as genetic background, can be carefully controlled, have been instrumental in demonstrating long-term effects that alterations in early life nutrition can have on the phenotype and in providing insights into the mechanism underlying this phenomena. Evidence that maternal diet in humans can induce long-term epigenetic and phenotypic changes in the offspring is more limited, although alterations in the methylation of a number of genes in DNA isolated from the whole blood of individuals who were exposed peri-conceptionally to famine during the Dutch Hunger Winter have been reported.Reference Heijmans, Tobi and Stein ^{72 ,} Reference Tobi, Lumey and Talens ⁷³ Specifically, a decrease in methylation of the imprinted IGF2 gene and an increase in methylation of interleukin-10, leptin, ATP binding cassette A1 and guanine nucleotide binding protein were observed in those individuals whose mothers were exposed to famine during pregnancy compared with their non-exposed siblings. Interestingly, these methylation changes were predominantly seen in those individuals who had been exposed in utero during early gestation rather than late gestation, suggesting that the peri-conceptual period is a period of extreme sensitivity to nutritional status. Furthermore, these measurements were made 60 years after the famine, suggesting that maternal nutritional constraint induces long-term epigenetic alterations in the offspring, consistent with the observations from animal models. Altered methylation of 5 CpG sites in the IGF2 gene was also detected in the peripheral blood cells of children whose mothers took 400 µg of folic acid per day during the peri-conceptional period compared with children whose mothers did not take folic acid supplementation.Reference Steegers-Theunissen, Obermann-Borst and Kremer ⁷⁴ More recent studies have reported widespread sex-specific changes in the epigenome associated with peri-conceptual micronutrient intakeReference Khulan, Cooper and Skinner ⁷⁵ and an association between the DNA methylation status of human metastable epialleles with maternal nutritional intake at conception.Reference Dominguez-Salas, Moore and Baker ⁷⁶ Plasticity in the human epigenome may also persist into adulthood, as studies have shown that HFD fed during adulthood induced changes in the methylation of 6508 genes in skeletal muscles, and such changes were interestingly only partially reversed after 6–8 weeks of a normocaloric diet.Reference Jacobsen, Brons and Bork-Jensen ⁷⁷ The changes in DNA methylation in all these human studies are considerably smaller than those found in animal studies, which may reflect the fact that in animal models the perturbation in maternal diet is more defined, the postnatal environment is precisely controlled and the genetic background is identical. Interestingly, in a recent study, Teh et al.,Reference Teh, Pan and Chen ⁷⁸ where they screened the genotypes and methylomes of 237 neonates, found 1423 variable methylated regions, and in 75% of the cases these were best explained by an interaction between different in utero environments and genotype, suggesting that the effect of early life environment is moderated by genetic background.

Epigenetic biomarkers predictive of later disease risk

If a similar mechanism does operate in humans as in animals, with early life environment inducing the altered epigenetic regulation of genes, then it should be possible to detect these altered epigenetic markers and use them as predictors of future metabolic capacity and disease risk. The potential issue with detecting such markers in humans is that there is limited tissue availability; readily accessible tissues include cord, cord blood, placenta, buccal and blood, but as DNA methylation patterns are often tissue specific, it is not clear whether the methylation status in such tissues would reflect methylation patterns in more metabolically relevant cell types. However, a number of studies have shown inter-tissue methylation correlations.Reference Murphy, Huang and Hoyo ^{79 ,} Reference Waterland, Kellermayer and Laritsky ⁸⁰ For instance, Talens et al. Reference Talens, Boomsma and Tobi ⁸¹ have shown that, for a number of non-imprinted genes, DNA methylation levels measured in blood were also equivalent in buccal cells, despite the fact that these cell types stem from different germ layers (mesoderm and ectoderm, respectively), whereas Byun et al.Reference Byun, Siegmund and Pan ⁸² reported that, using the IlluminaGoldenGate Bead Array, which integrated 1505 CpG sites in 807 genes, the intra-individual correlations for over 11 tissues ranged from 0.738 to 0.941. It may also depend on when the environmental constraint occurred. An environmental challenge during very early development is likely to affect all germ layers and an imprint of this altered epigenetic mark maybe detectable in all tissues, whereas exposures occurring later on in gestation may induce only tissue-specific effects.

Interestingly, Godfrey et al. Reference Godfrey, Sheppard and Gluckman ⁸³ have reported that the methylation of a single CpG site in the promoter region of the nuclear receptor RXRA was strongly related to childhood adiposity in both boys and girls in two independent cohorts; with RXRA promoter methylation explaining over 25% of the variance in age- and sex-adjusted fat mass in children at 6 and 9 years of age. This suggests that a far greater proportion of individual vulnerability to NCD may arise during development than has been considered previously and that developmentally induced epigenetic changes may mediate this effect. Such findings also suggest that the detection of such methylation changes even in peripheral tissues may be useful markers of later disease risk. However, contrary to many of the initial studies, which suggest that once DNA methylation patterns are established in early life they are then stably maintained, there is growing evidence that DNA methylation can be dynamically regulated in response to a number of environmental stimuli. For instance, an acute exercise bout over a 20-min period induced hypomethylation of the peroxisome proliferator-activated receptor gamma, coactivator 1 alpha (PGC1α) and mitochondrial transcription factor A promoters in muscle tissues,Reference Barres, Yan and Egan ⁸⁴ whereas TarantiniReference Tarantini, Bonzini and Apostoli ⁸⁵ found an association between concentrations of PM₁₀ particles over a 3-day period and levels of inducible nitric oxide synthase methylation. Such sensitivity to the environment could confound the contributions that developmentally induced epigenetic markers have on later phenotype, limiting the usefulness of any epigenetic predictive markers for disease risk. The factors that induce such dynamic changes in DNA methylation are not fully understood nor why the methylation of some CpGs appears to be stable and others very plastic, primarily as data on methylation stability from longitudinal studies, especially in children when the epigenome may well be more sensitive to the environment, are limited. One study that has examined DNA methylation stability over time in children found that methylation of the genes MAOA, dopamine receptor D4 (DRD4) and serotonin transporter (SLC6A4) is highly dynamic between the age of 5 and 10 years. Although the stability of CpG loci may be very loci-dependent, Clarke-Harris et al. Reference Clarke-Harris, Wilkin and Hosking ⁸⁷ have reported year on year stability of 7 CpG sites within the PGC1α promoter in peripheral blood cells in children from 5 to 14 years of age, suggesting that at least for these CpG sites methylation is set up in early life and stably maintained thereafter, despite changes in exercise, pollution or the onset of puberty. Moreover, four of the seven CpGs analysed at 5–7 years predicted adiposity in the children aged 9–14, consistent with the paradigm that developmentally induced epigenetic markers make a significant contribution to later phenotype. Interestingly, the CpGs analysed in this latter study were located upstream of the region shown to be hypomethylated in muscles in response to acute exercise, whether the difference in stability relates to the differential location of these CpG sites within PGC1α promoter, tissue-specific differences between blood and muscle or whether the change in methylation in response to exercise occur on top of a developmentally induced imprint is not clear. A better understanding of the factors influencing variation in DNA methylation and their role in mediating DNA methylation dynamics will be important to elucidate the role of these factors in producing DNA methylation differences that may underlie the development of complex diseases as well as for the development of epigenetic biomarkers predictive of later disease risk.