16.1 Introduction

Obesity is a chronic metabolic disease that arises from the complex interplay of numerous environmental, behavioural and genetic influences. It is characterised by an abnormally high proportion of adipose tissue constitutive of total body mass, and is strongly associated with an increased risk of cardiovascular disease, type 2 diabetes, various psychiatric disorders and cancers [1, 2]. As the global prevalence of obesity continues to rise at an alarming rate [3], there has yet to be any successful national or international effort to combat this disease, and its increasing economic burden upon health-care systems with limited resources [4].

Of particular concern is the rising prevalence of obesity among women of childbearing age, which has previously been demonstrated to increase the risk of complications during pregnancy and the likelihood that these children will, in turn, suffer from obesity and its associated comorbidities in adult life [5]. Prospective studies have repeatedly demonstrated strong links between maternal body mass index (BMI) in and around pregnancy and the incidence of obesity in adolescence and adulthood. Barisione et al. reported that 22 % of children born to mothers suffering from obesity were, themselves, obese at 12 years of age [6]. However, the prevalence of obesity among their siblings, who were born following substantial surgically induced maternal weight loss, was just 3 % at the same age. This discrepancy extended even into adult life, where the mean weight and BMI for each group were 79.5 kg and 27.5 kg/m2 and 66.7 kg and 23.4 kg/m2, respectively [6]. Furthermore, animal models of maternal obesity induced by obesogenic perinatal diet have also described increased body mass and adiposity, hepatic steatosis, adverse metabolic lipid profiles and a greater response to obesogenic diet among the offspring [7–10].

The Developmental Origins of Health and Disease (DOHaD) theory suggests that maternal physiology and metabolism during the perinatal, fetal and even preconceptional phases of development are capable of modifying the metabolic profiles of their offspring by altering how different cell types express specific genes across different tissues and time [11–14]. Consequently, epigenetics has been proposed as the main molecular mechanism implicated in this perinatal programming [15].

16.2 Epigenetic Mechanisms

Epigenetics describes the translation and adaptation of genotype to phenotype, which is regulated by a complex and interacting network of covalent modifications of chromatin structure (Fig. 16.1). These epigenetic modifications determine the cellular state and the metabolism affecting gene expression patterns in a cell-specific manner whilst preserving the nucleotide sequence [16].

Fig. 16.1
figure 1

External environment interacts with genotype altering epigenetic profile. Changes in DNA methylation at CpG sites and post-translational modifications on histone tails promote alterations in chromatin condensation profile, which regulates the join of the transcriptional machinery of the genes. Furthermore, the microRNAs interfere with post-transcriptional regulation of gene expression. The combination of these epigenetic changes defines phenotypic characteristics

Full size image

The cellular state is intrinsically related to the chromatin state, which describes the association of DNA molecules with specialised proteins, including histones, which package and configure the genetic code three-dimensionally within the confines of the nucleus. These histones, and in particular their N-terminal tails, are susceptible to a variety of post-translational modifications, including phosphorylation, ubiquitinylation, acetylation and methylation [17]. Commonly referred to as histone modification marks, each is believed to contribute to the regulation of gene expression by controlling the degree of condensation of the surrounding chromatin and hence the ease of access for the transcriptional machinery. The other main class of epigenetic modification is DNA methylation. This modification is mainly found at a cytosine with a guanine as next nucleotide (CpG site) and is commonly associated with transcriptional repression. These CpG sites are abundant within and around gene promoter regions, where specific transcription factors bind to DNA in order to recruit the transcriptional machinery and orchestrate the gene expression [18]. Thus, methylation of the promoter region at these sites represses transcription by means of steric impedance of transcription factor binding or via intermediary proteins that bind methylated DNA [19]. Finally, whilst not directly interacting with DNA and thus not strictly a class of epigenetic modification, short-chain RNA molecules, referred to as microRNAs, which are not themselves translated, appear to interact with mRNA sequences and regulate protein synthesis [20].

There exist key time frames, during which a cell’s epigenetic profile is more susceptible to change and adaptation, and hence more vulnerable to environmental insults, especially during pregnancy and breastfeeding. In this way, the maternal metabolism can affect or ‘program’ the epigenetic profile of their offspring and thereby alter that child’s risk of developing obesity in adult life [21]. At such times, the rate of cellular division is dramatically increased and DNA more exposed to the chemical modification within the nucleus, which may represent an enticing opportunity for future clinical interventions [21].

Our growing understanding of the mechanisms of maternal programming in obesity may go on to explain how the features of this disease can be modified by lifestyle and nutrition, uncovering how the genetic information and environmental exposure interact at the molecular level [22, 23]. Differences in DNA methylation patterns between those who suffer from obesity and controls have been widely reported in the literature [24] and successfully used as a biomarker of dietary response in kilocalorie-restricted diets [25].

It is now apparent that epigenetic information can be passed on to the next generation. This has led some to observe that we are not only what we eat but also what our progenitors ate [26, 27]. Recent evidence suggesting that maternal programming can endure across successive generations is startling and hints at an epigenetic landscape in obesity of previously unimagined complexity and importance.

16.3 Epigenetic Programming in Maternal Obesity

Obesity induces an aggressive and degenerative physiological environment, increasing the levels of triglycerides, cholesterol, glucose and other metabolites in the plasma, raising blood pressure and causing systemic angio-dysgenesis and hypoxaemia [1, 2]. Ultimately, this can lead to multiple organ damage, including non-alcoholic fatty liver disease (NAFLD), non-alcoholic fatty pancreas disease (NAFPD) and various cancers [1, 2]. Maternal obesity at conception, during pregnancy and while breastfeeding exposes the offspring to this adverse environment, programming their physiology with a heightened susceptibility to developing metabolic diseases in their own lifetime (Fig. 16.2). These children are more likely to be born prematurely and often with abnormally high or low birthweight. This was demonstrated in a study of 319 such mother–child pairs, where global DNA methylation (quantified by Long Interspersed Nuclear Element 1, LINE-1) in samples of cord blood was found to be greater in premature and extreme birthweight newborns compared with controls, and associated with increased adiposity in later life [28].

Fig. 16.2
figure 2

Maternal obesity and other perinatal influences program offspring development (NAFPD non-alcoholic fatty pancreas disease, NAFLD non-alcoholic fatty liver disease, Hyper TG hypertriglyceridaemia, Hyper Chol hypercholesterolaemia)

Full size image

In addition to affecting the development of obesity in their offspring, maternal obesity can also trigger the physiological dysregulation of other systems in offspring. For example, different patterns of DNA methylation were found in 57 genes related to the development of the central nervous system in samples of umbilical cord blood in infants from obese mothers versus normal-weight mothers, and may well indicate the onset of abnormal nervous system development in these children [29].

As technology has evolved, genome-related massive omic tools have arisen as the most useful initial approach in the search for biomarkers of maternal programming. Genome-wide interrogation of cord blood samples from more than a thousand mother–child pairs has also identified multiple CpG sites that are concordantly methylated in mother and child, and associated with high maternal weight and offspring adiposity [30]. Genome-wide methylation analyses have also found links between maternal BMI and patterns of cord blood DNA methylation in genes related to cardiovascular disease and several malignancies [31]. However, in order to identify the specific genetic loci affected by maternal obesogenic programming, a larger pool of data encompassing more diverse populations is still required. That said, numerous plausible candidates have already been identified. For example, in cord blood, high levels of DNA methylation at the gene encoding Retinoid X Receptor-α (RXRA), a transcriptional regulator, have been associated with increased adiposity in children at 9 years of age [32]. Gemma et al. also describe a positive correlation between maternal BMI and methylation levels of the Peroxisome Proliferator-activated Receptor-gamma Co-activator 1-α (Ppargc1a) gene in cord blood, which encodes a transcriptional activator involved in regulating various metabolic processes, including energy homeostasis, hepatic gluconeogenesis and cholesterol levels [33]. Additionally, Lesseur and colleagues observed that levels of DNA methylation of the Leptin gene promoter were comparably lower in blood samples taken from obese mothers than normoweight mothers just before pregnancy, as well as cord blood samples from their children at birth, when compared to mothers of normal weight and their offspring. These neonatal methylation levels were shown to correlate with leptin concentration in maternal plasma [14]. Furthermore, differences in ambion microRNA expression profile between obese and normoweight women were associated with a downregulation of insulin and adipocytokines signalling pathways, among others [34].

The paternal epigenetic profile, carried by the sperm, may also have a role in predisposing subsequent generations to obesity. Paternal obesity prior to conception has recently been associated with higher levels of DNA methylation of several genes (MEST, PEG3, NNAT and Igf2) known to be affected by paternal imprinting in leukocytes extracted from cord blood [35, 36].

This same study found that methylation levels of different genes (PLAG1, MEG3, H19) were associated with maternal obesity prior to conception, which implies that paternal and maternal influences may affect their child’s physiology differently [35, 36].

Besides the intrauterine environment, breastfeeding also represents a critical period of epigenetic reconfiguration for the neonate in response to maternal chemical influences. A study of 120 mother–child pairs, undertaken by Obermann-Borst et al., found a negative correlation between the duration of breastfeeding and leptin methylation in whole blood samples taken from offspring at 17 months post-partum [37].

Existing, as well as novel, interventions may be targeted during key developmental windows to ameliorate the risk of maternal obesity to the unborn [38]. For example, Guenard et al. describe a significant reduction in the cardiovascular risk profile of children born after substantial maternal weight loss induced by bariatric surgery when compared to their older siblings. Subsequent, transcriptomic and epigenetic analysis of whole blood samples identified 5698 genes that were differentially methylated between these sibling pairs, many of which were related to glycaemic control, inflammation and vascular disease [39]. The transcriptional patterns of five such genes linked to the innate immune system and inflammatory response were also shown to differ significantly between these two groups [40]. The critical importance of innate immunity and its regulation in the context of maternal obesity has received further support from animal studies of maternal obesogenic feeding that demonstrate innate immune dysfunction in offspring with developmentally programmed NAFLD [41].

However, it could also be argued that any of the comorbidities associated with maternal obesity may adversely condition the epigenetic profile and affect the development of these children. Indeed, maternal diabetes mellitus and gestational diabetes have previously been described as ‘conditioning factors’ affecting development in utero. Global methylation levels of placental DNA appear to be decreased in gestational diabetes and pre-eclampsia but increased in maternal obesity [42]. This phenomenon has been associated with specific phenotypic characteristics, such as head circumference at birth and height in infancy. Reduced methylation at specific CpG sites within the leptin gene has also been demonstrated in cells derived from cord blood in the context of maternal hyperglycaemia [43]. El Hajj et al. further noted that the offspring of mothers with gestational diabetes display reduced DNA methylation of the genes Mest (mesodermic-specific transcript), Nr3c1 (nuclear receptor subfamily 3, group C, member 1) and Alu sequences in cord blood and placenta, when compared to mothers with adequate glycaemic control [44]. These results accord with prior evidence that Mest methylation is similarly reduced in blood samples taken from adults who were morbidly obese [44].

To date, research in this field is largely based upon observational studies in humans, but rodent experimental models of maternal obesity have now become the first choice for interventional studies seeking to elucidate the epigenetic mechanisms of maternal developmental programming in obesity [45]. Usually such interventions involve perinatal obesogenic feeding, enriched in simple sugars and fats, similar to the Western diet [46]. Expectant mothers are fed in this way during pregnancy and whilst breastfeeding. The simplicity and economy of rodent maintenance, their significant genetic, physiological and metabolic similarities with humans and their relatively short lifespans make them ideal for studying these phenomena over successive generations [45, 46].

The offspring of such high-Fat fed obese mice, for example, display decreased levels of methylation at the promoter of the gene encoding the zinc finger protein 423 (Zfp423), a transcription factor committing cells to the adipose lineage, in association with downregulation of histone marks H3K27me3, and higher levels of expression in fetal adipose tissue [47]. Maternal obesity during pregnancy has also been associated in rats with similarly reduced levels of DNA methylation at the genes encoding Zfp423 and C/EBP-β (CCAAT/enhancer binding protein, beta), another proadipogenic transcription factor, as well as increased levels of their respective mRNA transcripts in offspring adipose tissue [48]. Concordantly, the extent of hypothalamic DNA methylation of the genes encoding proopiomelanocortin (POMC) and neuropeptide Y (NPY), both involved in the regulation of appetite, was positively correlated with calorific intake [49, 50].

Obesogenic maternal diet prior to and after conception has also been shown to induce altered levels of microRNAs associated with cardiovascular disease within the myocardium of baboons [51]. Also, in the livers of similarly exposed neonatal rats, the expression of Cdkn1a, a gene associated with hepatocyte growth following liver damage and several cancers, was upregulated in tandem with lower levels of promoter methylation [52]. Maternal obesity appears to program a greater propensity for NAFLD in their offspring, exacerbated further by exposing them in turn to an obesogenic diet. In such circumstances, changes in the DNA methylation profile of genes related to circadian rhythmicity in liver, Bmal1 and Per2, have since been reported [53].

The key question then becomes how we might effectively intervene and attenuate the risk of maternal obesity to the next generation. Animal studies of maternal physical exercise have, for example, demonstrated successful prevention of Pparg1a hypermethylation in the offspring of mothers fed an obesogenic diet, normalising its expression and that of its target genes in skeletal muscle [54]. Maternal weight loss in sheep prior to conception has also been shown to affect hepatic insulin signalling and microRNA expression profiles in their offspring [55]. It is perhaps not surprising, then, that physical maternal exercise appears to protect the offspring from the physiological changes mediated by maternal obesity. Maternal micronutrient supplementation in rats while breastfeeding has also been found to prevent maternal obesity-induced homocysteinaemia in offspring, in association with changes in the activity of DNA methyltransferases and global levels of hepatic DNA methylation [56].

Given that the epigenetic profile of each cell defines its identity and its role within the organism, it is conceivable that the maternal nutritional condition affects different cell types in different ways [11]. When occurring in germ cells, these alterations gain the potential to endure through successive generations, extending the implications of maternal programming in obesity.

16.4 Transgenerational Epigenetic Programming in Maternal Obesity

Whilst the majority of research has sought to elucidate the mechanisms and manifestations of maternal epigenetic programming in obesity by focusing on the first generation of offspring, recent evidence suggests that these maternal programs can endure across successive generations [57]. The implications of transgenerational programming in maternal obesity are startling, hinting at an extremely complex and multidimensional epigenetic landscape that is yet to be fully understood.

The time constraints implicit in such experimental models of transgenerational obesity inevitably render rodents preferable subjects to humans. High-fat perinatal maternal feeding in mice has recently been demonstrated to increase the birthweight, adiposity and macrophage infiltration of adipose tissue across three generations of their descendants. This immunomodulation was accompanied by a decrease in the levels of promoter methylation and increased expression of Toll-like receptor 1 and 2 (TLR1 and TLR2), both involved in the activation of T cells [58]. A similar study of maternal obesogenic diet in mice prior to conception induced traits of the metabolic syndrome in five subsequent generations of their offspring, as well as altering patterns of histone marks at the genes encoding leptin and adiponectin and their expression in white adipose tissue [59]. When offspring of these animals returned to a control diet, these alterations were completely abolished after three generations, implying that effective intervention of this kind is possible.

When epigenetic modifications are induced in germ cells by the perinatal maternal environment, they gain potential transmissibility across subsequent generations [57] (Fig. 16.3). Ge et al. described significant alteration of Peg3 (paternally expressed 3) and H19 (H19, imprinted maternally expressed transcript) promoter methylation in the spermatozoa of offspring in a mouse model of maternal obesity and diabetes mellitus [12]. Oocytes harvested from obese females displayed increased methylation of the leptin promoter and reduced peroxisome proliferator-activated receptor alpha (Ppar-α) promoter methylation. Whilst similar transcriptional and epigenetic profiles were observed for Ppara-α in the livers of their female offspring, oocytes harvested from these same offspring displayed increased levels of Ppar-α promoter methylation [13]. Hence, different cell types appear to be differentially affected by maternal programming in obesity, perhaps even at different stages of development. Interestingly, maternal weight loss prior to conception appeared to reprogram patterns of DNA methylation in the liver and normalise the expression of genes related to lipid metabolism in their offspring [60]. This emphasises how changes in maternal physiology, even prior to conception, hold the potential to influence the metabolism of their offspring by affecting the epigenetic processes that regulate gene expression.

Fig. 16.3
figure 3

Transgenerational transmission by maternal environment during pregnancy

Full size image

Whilst most of the transgenerational animal models were originally designed for the study of potentially teratogenic agents [61] and may require further adaptation, they have already shed considerable light upon the nature of transgenerational epigenetic programming in maternal obesity. However, it must also be acknowledged that our mechanistic comprehension of this phenomenon remains, itself, in its infancy.

16.5 Conclusion

Maternal obesity during pregnancy, through breastfeeding and mechanisms preceding conception, can program their offspring with a physiological predisposition towards developing obesity and its associated comorbidities in adult life. Maternal programming in obesity engenders changes in the epigenetic profiles of diverse cell types, affecting how certain genes associated with obesity are expressed at different stages of a child’s development. The epigenetic programs that mediate the phenotypic characteristics of this disease appear to be transmissible and can endure across successive generations. However, they remain amenable to appropriately targeted intervention. A deeper understanding of the molecular phenomenology underlying maternal epigenetic programming in obesity is desperately needed in order to develop more effective therapeutic approaches in the management of this burgeoning global epidemic.