Early-life environment and later disease risk

An association between the quality of the early-life environment and later disease risk was first shown in a Norwegian study which identified a strong association between undernutrition and poverty during childhood and adolescence, followed by later prosperity, with CVD in late middle age⁽Reference Forsdahl⁵⁾. Subsequent work by David Barker and co-workers related the health of middle-aged individuals in Britain to their recorded birth measurements; lower birth weight was found to correlate strongly with the later risk of CVD, type 2 diabetes, hypertension and hyperlipidaemia⁽Reference Barker and Osmond^6–Reference Barker, Hales and Fall⁹⁾. Further epidemiological studies have confirmed the associations between lower birth weight and later disease risk but importantly have also shown that babies born at the highest birth weights are also at increased risk of developing diabetes or obesity in later life⁽Reference Pettitt and Jovanovic^10, Reference Ong¹¹⁾. Birth weight in these studies is thought to be an indicator of a poor intrauterine environment which may have been compromised through a variety of maternal or environmental factors such as placental insufficiency, maternal undernutrition, overnutrition, maternal obesity or gestational diabetes⁽Reference Hanson and Gluckman¹²⁾.

While these epidemiological studies were the first to show a relationship between intrauterine environment and chronic disease risk, studies of the Dutch Hunger Winter, a famine which occurred in the Netherlands during the winter of 1944, clearly demonstrated that maternal nutrition influences the health of the child in later life and that the timing of the nutritional constraint is important. Studies found that individuals whose mothers were exposed to famine periconceptually and during the first trimester of pregnancy exhibited an increased risk of obesity and CVD, whereas individuals whose mothers were exposed in the later stages of gestation showed an increased incidence of insulin resistance and hypertension in later life⁽Reference Roseboom, de and Painter^13, Reference Painter, Roseboom and Bleker¹⁴⁾. Comparable findings have now been replicated in a variety of animal models where rats or mice have been fed either a low-protein diet or a global dietary restriction during pregnancy⁽Reference Bertram and Hanson^15–Reference Bellinger, Sculley and Langley-Evans¹⁹⁾. Interestingly, experimental studies have also shown that the offspring from dams fed a high-fat or even junk food diet during pregnancy and/or lactation also exhibit similar features and develop hypertension, dyslipidaemia, obesity and insulin resistance in later life⁽Reference Samuelsson, Matthews and Argenton^20, Reference Hoile, Irvine and Kelsall²¹⁾.

The induction of different phenotypes by perturbations in early-life nutrition has been suggested to reflect a predictive adaptive response whereby the organism, acting through the process of developmental plasticity, can adjust its developmental programme in response to environmental cues in early life to aid fitness or survival in later life⁽Reference Gluckman, Hanson and Spencer²²⁾. When an organism adapts to one environment and is subsequently exposed to a different environment after birth, a mismatch occurs leaving the organism maladapted and at risk of metabolic disease in later life⁽Reference Bateson, Barker and Clutton-Brock^23, Reference Cleal, Poore and Boullin²⁴⁾. This mismatch between the prenatal and postnatal environments has been suggested to be central to the burgeoning rates of NCD observed in countries undergoing socioeconomic transition, for example, as populations move from rural to urban areas⁽Reference Norris, Osmond and Gigante^25, Reference Bavdekar, Yajnik and Fall²⁶⁾. In the western world, exposure to continuous energy/sugar-rich diets may be beyond our adaptive capacity leading to pathological changes and an increased disease risk.

Epigenetic regulation

The mechanism by which early-life nutrition may influence future disease risk has been suggested to involve the altered epigenetic regulation of genes, sections of DNA that encode for proteins⁽Reference Godfrey, Lillycrop and Burdge²⁷⁾. Epigenetic processes, which are stably inherited through cell division, regulate the expression potential of genes without altering the underlying genetic sequence. These epigenetic processes include DNA methylation, histone modifications and non-coding RNA (ncRNA). Together, they regulate all aspects of gene expression, controlling either the access of the transcriptional machinery to the underlying DNA sequence, the stability of mRNA or its translational competence, thereby defining when and where genes are expressed and levels of expression.

DNA methylation

DNA methylation is the transfer of a methyl group (CH₃) to the C₅ position of cytosine to create 5-methylcytosine⁽Reference Kumar, Cheng and Klimasauskas²⁸⁾. In mammals, the majority of methylated cytosines are found next to a guanine and are referred to as CpG (cytosine and guanine nucleotides linked by a phosphate). In general, low levels of DNA methylation in the promoter or control regions of genes are associated with transcriptional activity, while high levels of methylation are associated with transcriptional silencing⁽Reference Bird and Macleod²⁹⁾. CpG promoter methylation leads to transcriptional repression by either blocking the ability of transcription factors to bind to DNA and activate gene expression or through the recruitment of histone-modifying complexes to the DNA, which leads to a closed chromatin configuration and transcriptional repression⁽Reference Fuks, Hurd and Wolf³⁰⁾.

Establishment of DNA methylation patterns through development

DNA methylation is markedly reprogrammed during embryonic development in mammals. Following fertilisation, the methylation marks on the maternal and paternal genomes are largely erased⁽Reference Reik, Dean and Walter³¹⁾. Demethylation occurs rapidly on the paternal DNA, catalysed by the ten to eleven translocation proteins, while demethylation on the maternal DNA occurs more slowly over several cell divisions by a replication-dependent passive process⁽Reference Smith, Chan and Mikkelsen^32–Reference Smallwood, Tomizawa and Krueger³⁴⁾. However, a number of regions escape this erasure including the imprinting control regions of imprinted genes⁽Reference Lane, Dean and Erhardt³⁵⁾. After this phase of global demethylation and around the time of blastocyst implantation, the genome then undergoes de novo methylation. The pluripotency genes Oct-3/4 and Nanog, which are essential to maintaining the undifferentiated state of early stem cells, are methylated and silenced⁽Reference Farthing, Ficz and Ng³⁶⁾. Methylation of lineage-specific genes also occurs, thus genes that are not required within certain cell types are methylated and silenced. In contrast, CpG islands within housekeeping genes, constitutive genes that are essential for basic cell function, are protected from this global de novo methylation and remain unmethylated⁽Reference Smith, Chan and Mikkelsen³²⁾.

Histone modifications

In eukaryotic organisms, DNA within the nucleus is wrapped around a histone octamer consisting of two copies of each core histone (H2A, H2B, H3 and H4) forming the basic unit of chromatin, 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 and packaging of the DNA is essential to reduce its effective size⁽Reference Bednar, Horowitz and Grigoryev³⁷⁾. However, it is now clear that histone proteins are not only important for the packaging of DNA, but also play a critical role in regulating gene expression alongside DNA methylation. Histone proteins contain two domains: a globular domain and an amino tail domain. The amino tail domains of histone proteins are rich in positively charged amino acids which interact with negatively charged DNA and are now known to be subject to a large number of post-translational modifications including acetylation and methylation⁽Reference Bannister and Kouzarides^38, Reference Olins and Olins³⁹⁾. These modifications can either directly affect chromatin structure or provide binding sites for proteins involved in gene regulation⁽Reference Tessarz and Kouzarides⁴⁰⁾. For example, acetylation of the lysine residues within histone tails, which is catalysed by histone acetyltransferases, neutralises the positive charge of the lysine, thereby reducing the interaction of the histone with DNA. This leads to an opening up of chromatin, allowing access to the transcriptional machinery. Conversely, histone deacetylases remove the acetyl groups, restoring the positive charge and leading to a closing down of the chromatin and gene repression⁽Reference Turner⁴¹⁾. Histone methylation is associated with both activation and repression; methylation at lysine (K) 4 on histone H3 is an activating mark, while methylation at K9 or 27 is associated with transcriptional repression⁽Reference Lachner, O'Carroll and Rea⁴²⁾.

Non-coding RNA

Although up to 90% of the eukaryotic genome is transcribed, only 1–2 % of the genome encodes proteins⁽Reference Mouse, Stamatoyannopoulos and Snyder⁴³⁾. These ncRNA include microRNA (miRNA) which can induce mRNA degradation or translational repression and, when binding within the promoter region of a gene, induce both DNA methylation and repressive histone modifications resulting in reduced transcriptional activity or even complete repression⁽Reference Morris, Chan and Jacobsen^44–Reference Kim, Saetrom and Snove⁴⁶⁾. DNA methylation, histone modification and ncRNA often work in concert with each other to regulate gene expression. Many studies, although not all, suggest that DNA methylation may consolidate changes in gene expression induced by histone modifications and ncRNA, and that DNA methylation represents the most stable of these epigenetic marks which is required for long-term persistent repression of gene expression.

Effect of maternal diet on the epigenome: evidence from experimental models

Although DNA methylation was thought to be a very stable modification and once methylation patterns were established in early life, they were then largely maintained, there is now growing evidence that a number of environmental factors such as nutrition⁽Reference Lillycrop, Slater-Jefferies and Hanson⁴⁷⁾, body composition⁽Reference Gemma, Sookoian and Alvarinas⁴⁸⁾, endocrine disruptors⁽Reference Skinner⁴⁹⁾ and social environment⁽Reference Szyf, McGowan and Meaney⁵⁰⁾ can modulate the methylome, often resulting in long-term changes in gene expression and physiology.

Alterations in maternal diet in both rats and mice have been shown to induce changes in DNA methylation in the offspring. In adult viable yellow agouti (A^vy) mice, supplementation of the maternal diet with dietary methyl donors and cofactors (folic acid, vitamin B₁₂, choline and betaine) shifted the coat colour of the offspring from yellow (agouti) to brown (pseudo-agouti), and this was associated with the increased methylation of the agouti gene⁽Reference Wolff, Kodell and Moore⁵¹⁾. Maternal diet can also alter the methylation of key metabolic genes within the offspring. For example, feeding rats a protein-restricted (PR) diet during pregnancy induced hypomethylation of the glucocorticoid receptor and PPAR-α promoters in the livers of juvenile and adult offspring, which was associated with an increase in glucocorticoid receptor and Pparα mRNA expression⁽Reference Lillycrop, Slater-Jefferies and Hanson^47, Reference Lillycrop, Phillips and Jackson⁵²⁾. Maternal PR induced the hypomethylation of four specific CpG dinucleotides within the promoter of Pparα, two of which predicted the level of the mRNA transcript, in the juvenile offspring⁽Reference Lillycrop, Phillips and Torrens⁵³⁾. As these CpG sites lie within transcription factor binding sites, changes in the methylation status of these CpGs induced during development may affect later transcriptional induction of Pparα by specific stimuli and the capacity of the tissue to face the metabolic demand. In contrast to the effect of the maternal PR diet, maternal dietary restriction during pregnancy induced the hypermethylation of glucocorticoid receptor and Pparα promoters as well as a decrease in glucocorticoid receptor and Pparα expression⁽Reference Gluckman, Lillycrop and Vickers⁵⁴⁾. Thus, the effects of maternal nutrition on the epigenome of the offspring may depend upon the nature of the specific maternal nutrient challenge⁽Reference Lillycrop, Phillips and Jackson^52, Reference Lillycrop, Phillips and Torrens⁵³⁾.

Given the increasing consumption of energy-rich diets worldwide, much of the research has now focused on the effects of maternal high-fat feeding on DNA methylation in the offspring. For example, Vucetic et al. showed increased expression of the μ-opioid receptor and preproenkephalin in the nucleus accumbens, prefrontal cortex and hypothalamus of mice from dams that consumed a high-fat diet during pregnancy, which was accompanied by the hypomethylation of the promoter regions of these genes⁽Reference Vucetic, Kimmel and Totoki⁵⁵⁾. Hoile et al. showed that maternal high-fat feeding during pregnancy led to the reduced expression of fatty acid desaturase 2, the rate-limiting enzyme in PUFA synthesis, and the altered methylation of key CpG nucleotides within its promoter in the offspring⁽Reference Hoile, Irvine and Kelsall^21, Reference Kelsall, Hoile and Irvine⁵⁶⁾. Furthermore, Plagemann et al. showed that neonatal overfeeding induced by raising rat pups in small litters induces the hypermethylation of two CpG dinucleotides within the pro-opiomelanocortin promoter which are essential for pro-opiomelanocortin induction by leptin and insulin⁽Reference Plagemann, Harder and Brunn⁵⁷⁾.

Although most studies have concentrated on identifying changes in DNA methylation associated with perturbations in maternal diet, maternal diet has also been shown to induce changes in both miRNA and histone modifications in the offspring. Marked changes in miRNA expression have been documented in the liver and skeletal muscle of the offspring in response to maternal undernutrition during either the periconceptual or preimplantation period⁽Reference Lie, Morrison and Williams-Wyss^58, Reference Lie, Morrison and Williams-Wyss⁵⁹⁾, while a maternal high-fat diet during pregnancy and lactation has been shown to alter hepatic expression of miRNA in the adult offspring that include lethal-7a (let-7a), let-7b and let-7c ⁽Reference Zhang, Zhang and Didelot⁶⁰⁾. Substantial changes in histone modifications at the hepatic NF 4a promoter has also been reported in offspring from dams fed a PR diet during pregnancy. Here, a decrease in hepatocyte NF 4α expression was accompanied by a reduction in histone 3 lysine 4 methylation and an increase in histone 3 lysine 9 and histone 3 lysine 27 methylation⁽Reference Sandovici, Smith and Nitert⁶¹⁾ with minimal changes in DNA methylation.

Changes in histone modifications induced by variations in early-life environment may precede a change in DNA methylation. Park et al. using a model of intrauterine ligation, which results in intrauterine growth restriction and long-term metabolic changes in the offspring, found a decrease in the expression of the transcription factor, pancreatic and duodenal homeobox 1, which plays a critical role in the differentiation and function of B cells. This was accompanied by an initial decrease in histone acetylation at the promoter of pancreatic and duodenal homeobox 1 within the intrauterine growth restriction fetus. This was followed after birth with a significant increase in the repressive histone mark histone 3 lysine 9 methylation in intrauterine growth restriction islet cells. At this stage, these epigenetic changes were reversible; however, with further accumulation of histone 3 lysine 9 methylation, and methylation of the pancreatic and duodenal homeobox 1 gene, the repression of pancreatic and duodenal homeobox 1 expression becomes irreversible⁽Reference Park, Stoffers and Nicholls⁶²⁾.

The effect of maternal diet on the epigenome: evidence from human studies

There is increasing evidence that early-life nutrition can induce epigenetic alterations in human subjects. DNA methylation differences have been reported in individuals who were periconceptually exposed to famine during the Dutch Hunger Winter⁽Reference Tobi, Lumey and Talens^63, Reference Heijmans, Tobi and Stein⁶⁴⁾. Here, a decrease in methylation of the imprinted insulin-like growth factor 2 gene and increases in the methylation of IL-10, leptin and ATP-binding cassette A1 genes in genomic DNA isolated from whole blood cells from individuals who were exposed to famine in early gestation in utero compared with unexposed same-sex siblings was observed⁽Reference Heijmans, Tobi and Stein⁶⁴⁾. Such changes in DNA methylation were only observed after exposure to famine in early gestation⁽Reference Tobi, Lumey and Talens⁶³⁾ implying that the methylome is most susceptible to alterations in maternal diet in the very early stages of development. Moreover, these measurements were made 60–70 years after famine exposure in utero, suggesting as in the experimental studies that variations in maternal diet can induce persistent epigenetic changes in the offspring. Waterland et al. have also shown altered patterns of DNA methylation in individuals conceived during the protein-limited rainy season compared with those conceived in the dry harvest season in rural Gambia⁽Reference Waterland, Kellermayer and Laritsky⁶⁵⁾. DNA methylation was associated with periconceptional maternal plasma concentrations of key micronutrients involved in C₁ metabolism⁽Reference Dominguez-Salas, Moore and Baker⁶⁶⁾, which supplies the methyl groups for all methylation reactions, suggesting that the changes in DNA methylation found in this population may be linked not to the negative energy balance observed in mothers during the rainy season, but rather to the limited dietary levels of methyl donors and cofactors required for C₁ metabolism.

Consistent with the importance of C₁ metabolites, a number of studies have reported associations between the maternal intake and/or status of C₁ donors, cofactors and DNA methylation in the offspring. However, much of these data are highly variable in terms of the effect size, direction of effect and genes affected. For instance, periconceptional folic acid has been both positively⁽Reference Steegers-Theunissen, Obermann-Borst and Kremer⁶⁷⁾ and negatively⁽Reference Pauwels, Ghosh and Duca⁶⁸⁾ associated with methylation at the insulin-like growth factor 2 locus in the offspring. However, there are also studies which have reported no effect of periconceptional folic acid exposure⁽Reference Haggarty, Hoad and Campbell⁶⁹⁾ on DNA methylation. Furthermore, there are examples of maternal macronutrient intake influencing DNA methylation in the offspring. A prenatal diet high in fat and sugar was positively associated with offspring insulin-like growth factor 2 methylation⁽Reference Rijlaarsdam, Cecil and Walton⁷⁰⁾; maternal carbohydrate intake during the second trimester negatively associated with retinoid X receptor α methylation at birth⁽Reference Godfrey, Sheppard and Gluckman⁷¹⁾ and increased protein intake in pregnancy positively associated with GR methylation in the adult offspring⁽Reference Drake, McPherson and Godfrey⁷²⁾. The difficulty however with such studies is that it is not possible to know which nutrient deficits or imbalances caused the epigenetic effects.

Mechanisms by which maternal diet may alter the epigenome

The ability of nutritional factors, particularly those involved in C₁ metabolism to influence the epigenome is perhaps not unexpected, as methyl groups for all biological methylation reactions including DNA and histone methylation are primarily supplied from dietary methyl donors and cofactors via C₁ metabolism⁽Reference Scott and Weir⁷³⁾. In this pathway, methionine is converted to S-adenosylmethionine, the universal methyl donor. After transferring the methyl group, S-adenosylmethionine is converted to S-adenosylhomocysteine, which is then converted to homocysteine. Homocysteine is either recycled to methionine by the enzyme betaine homocysteine methyltransferase which uses betaine or choline, or via a folate-dependent remethylation pathway where 5-methyltetrahydrofolate is reduced to 5,10-methylene tetrahydrofolate by 5,10-methylenetetrahydrofolate reductase. This methyl group is then used by methionine synthase to convert homocysteine to methionine using vitamin B₁₂ as a cofactor.

Modulation of the epigenome may not, however, be limited to C₁ donors and cofactors, as many transcription factors which recruit the writers of the epigenetic code are regulated by nutritional factors, and indeed many of the readers and writers of the epigenetic code themselves are regulated at least in part by the concentration of specific metabolic substrates or cofactors. For instance, the PPAR family of nuclear receptors which play a key role in lipid metabolism are activated by PUFA⁽Reference Kliewer, Forman and Blumberg⁷⁴⁾, while the lysine-specific histone demethylase 1A, which demethylates histone 3 lysine 4 uses the reduction of the cofactor FAD to FADH₂. Thus, variations in dietary intake in terms of either individual components or total energy are likely to have an effect on the epigenome potentially inducing a persistent change in gene expression.

The effect of paternal diet on the offspring

To date the majority of studies have focused on the effect of maternal nutrition on the health of the child; however, it is becoming clear that paternal diet can also induce long-term effects on the health of the offspring and that these are associated with the altered epigenetic regulation of genes. Studies many years ago in Sweden showed that food availability during the pre-pubertal period of grandfathers was associated with the risk of diabetes and CVD in the grandsons, although not in the granddaughters⁽Reference Pembrey, Bygren and Kaati^75, Reference Kaati, Bygren and Edvinsson⁷⁶⁾. Subsequent experiments in rodents have found that variations in paternal diet induced phenotypic changes in the offspring. For instance, offspring from males exposed to dietary restriction had reduced birth weight and impaired glucose tolerance⁽Reference Jimenez-Chillaron, Isganaitis and Charalambous⁷⁷⁾, while feeding male rats a PR diet prior to mating led to elevated hepatic expression of genes involved in lipid and cholesterol biosynthesis and a decrease in cholesterol esters, relative to the offspring of males fed a control diet. Further studies have also reported that chronic high-fat feeding in Sprague–Dawley fathers increased body weight, adiposity and impaired glucose tolerance and insulin sensitivity in the offspring. Here, a paternal high-fat diet altered the expression of 642 pancreatic islet genes in adult female offspring; these genes were enriched for cation and ATP binding, cytoskeleton and intracellular transport. Hypomethylation of the Il13ra2 gene, which showed the highest fold difference in expression was also demonstrated⁽Reference Ng, Lin and Laybutt⁷⁸⁾.

Mechanism by which paternal environment is transmitted to the offspring

To understand how alterations in paternal diet can induce epigenetic and phenotypic changes in the offspring, many groups have examined the effect of paternal diet on DNA methylation, histone modifications and/or the ncRNA content of the sperm. Carone and co-workers showed cytosine methylation patterns were highly correlative in sperm from control, low protein or energy-restricted fathers, suggesting the sperm epigenome may be more refractory to differences in diet. However, Radford et al. reported that sperm from mice undernourished in utero showed a reduced level of DNA methylation⁽Reference Radford, Ito and Shi⁷⁹⁾. The different effects on the sperm methylome between these two studies may reflect the different diets given, exposure time to such diets and the methods used to assess DNA methylation changes. Paternal high-fat feeding has also been associated with changes in the sperm methylome. Donkin and Barres reported differential methylation of eighteen regions in the sperm of males fed a high-fat diet and their offspring, including CpGs within the solute carrier family 3 amino acid transporter heavy chain member 2, transforming growth factor β regulator 4 and major facilitator superfamily domain⁽Reference Donkin and Barres⁸⁰⁾.

Variation in the paternal diet has also been linked to changes in the sperm transcriptome. The expression of twenty-three miRNA was detected in the testis of high-fat fed fathers compared with control fed males. In particular, miRNA let-7c has been suggested as an important mediator of paternal transmission of altered offspring phenotypes. Let-7 miRNA, which is known to control lipid and glucose metabolism, was found to be differentially expressed in spermatozoa of high-fat diet fed rats as well as in the spermatozoa of their offspring⁽Reference de Castro Barbosa, Ingerslev and Alm⁸¹⁾. Let-7 miRNA were also reported to be down-regulated in spermatozoa after exposure of males to a low-protein diet⁽Reference Sharma, Conine and Shea⁸²⁾.

Further studies have identified transfer RNA (tRNA) fragments as potential contributors of the altered offspring phenotype after paternal high-fat feeding. tRNA fragments, which range in size from ten to forty-five nucleotides, are derived from the 5′ end of either mature tRNA or pre-tRNA⁽Reference Sharma, Conine and Shea⁸²⁾. Chen et al. showed that feeding a high-fat diet to males over a 6-month period substantially altered the composition of tRNA within the sperm. Moreover, injection of tRNA fragments isolated from sperm from males fed a high-fat diet into control zygotes resulted in altered expression of metabolic genes in the early embryo and metabolic disorders in the offspring, akin to those seen after natural mating of the high-fat fed fathers to control dams⁽Reference Chen, Yan and Cao⁸³⁾. These studies suggest that variations in paternal diet may alter the sperm transcriptome. Such changes may subsequently lead to the altered expression of genes and metabolism in the embryo, which may be further consolidated through modulation of histone marks and DNA methylation.