Influences during early life have substantial impact upon adult health and disease.
This concept is most commonly known as the developmental origins of health and disease.
Environmental, genetic and epigenetic factors, as well as interactions among these factors, underlie this association.
Modification of these factors in early development has the potential to influence an individual’s health across the life course.
An association between the early life environment and health in later life has been known at least since the time of Hippocrates. Contemporary research in this field grew following on from the seminal work of David Barker and colleagues in England in the 1980s. Barker found that the geographical pattern of adult ischaemic heart disease mortality matched that of infant mortality several decades earlier. The most common cause of infant mortality at the time was low birth weight, leading to the ‘Barker hypothesis’ that exposure to an intrauterine environment leading to poor fetal growth caused metabolic changes that persisted into adulthood and increased the risk for ischaemic heart disease.
This initial work was then validated across several populations and continuing refinement led to the coining of the term ‘fetal origins of adult disease’. Further work extended this theory, and it became evident that early life impacts upon later life disease were not limited to pregnancy but also the maternal preconception environment and early postnatal life. Gluckman and Hanson first described the DOHaD theory to better incorporate these influences outside of the fetal period upon adult health.
Much scientific research now focuses on DOHaD across a wide range of fields from both clinical and basic science arenas, and complex analysis of associations has led to new frontiers in statistics. The implications of the research extend now to the development of public health policy and global recommendations about pregnancy and childhood care. A large current thrust of the research focuses upon identifying the underlying biology of DOHaD in the hope that this might identify novel approaches to the prevention of the associated adult diseases. It is now clear that complex gene–environment interactions lie at the heart of this process.
History of the Developmental Origins of Health and Disease Theory
Historical works from the time of Hippocrates link the health of a woman during pregnancy, the condition of the child at birth and the growth of the infant to well-being in adulthood. Attributed to Hippocrates, Airs, Waters and Places describes the association between an adverse environment during pregnancy and the ongoing health of the offspring, noting that:
In the first place women who happen to be with child, and whose accouchement should take place in springs . . . have feeble and sickly children, so that they either die presently or are tender, feeble, and sickly, if they live.
Kermack and Forsdahl
During the 20th century, a link between early life and adult health was proposed by several researchers. As early as 1934, reductions in all-cause mortality in Europe were thought to be related to improved childhood conditions. Forsdahl later recognised an association between infant mortality rates and subsequent ischaemic heart disease decades later, insightfully hypothesising that poor early life nutrition may result in later life susceptibility to a mismatched environment of nutritional plenty.
The Barker Hypothesis
It was not for another decade that, with the recognition in 1986 by Barker and colleagues of an association between low birth weight and adult cardiovascular disease, this area began to receive substantial research interest. In his analysis of trends in mortality across geographical regions of England and Wales, Barker found that the areas with the highest mortality rates due to ischaemic heart disease were the same as those that had the highest rates of infant mortality decades earlier. The most common cause of infant death at the time was ‘low birth weight’, and the ‘Barker hypothesis’ was formulated, suggesting that events contributing to low birth weight also contributed to the development of cardiovascular disease in adulthood ( Fig 47.1 ). Reflection upon previous research and the continued work of Barker and others developed this theory by confirming these associations across various cohorts.
The Fetal Origins of Adult Disease
Continuing work demonstrated associations with an increasing number of adult diseases. This led to the proposal in 1992 of the ‘fetal origins of adult disease’ theory, which expanded upon the Barker hypothesis to acknowledge that impacts upon the fetus had the potential to affect a variety of adult diseases.
The Developmental Origins of Health and Disease
The field continued to grow as evidence mounted across various populations, species and diseases. By early in the new millennium, it was clear that adult outcomes were related to both specific diseases and overall well-being and were associated not only with the fetal environment but also with those of the preconception period and early childhood. To better acknowledge these factors, Gluckman and Hanson coined the phrase ‘developmental origins of health and disease’ as the field is now most commonly known.
Mechanisms of the Developmental Origins of Health and Disease
One principle underlying the DOHaD phenomenon is that of ‘developmental plasticity’, the notion that a range of phenotypes can result from the same genotype as a result of altered environmental exposures during an organism’s development. In the human and in the DOHaD realm, this plasticity manifests as altered disease risk in later life. The association with low birth weight does not necessarily reflect a causal role of abnormal fetal growth in the development of disease but, rather, abnormal fetal growth is but one measure of an abnormal intrauterine environment, exposure to which induces phenotypic variation leading to increased disease risk. Developmental plasticity is a well-recognised phenomenon in nonhuman species, with substantial animal evidence of the environmental impact upon gene expression and phenotype.
Plasticity is time dependent and can only take place during critical periods of organogenesis. For example, brain development takes place over a longer period than cardiac development, and environmental impacts only result in changes to cardiac structure when present during earlier development compared with those which may produce neurologic changes. Periods of plasticity may vary between the structural and functional development of an organ, however. For example, although structural development of the fetal heart is complete relatively early in gestation, late insults may alter terminal cardiac myocyte differentiation, leading to more subtle functional changes. The physiological regulation of metabolism and inflammation is relatively late to develop and is therefore subject to environmental influences for much of the early life period and more susceptible to perturbations resulting in later disease.
Gluckman and Hanson describe an important distinction in this observation, that between developmental adaptation and developmental disruption. Not all plasticity results in advantageous changes for the developing organism, and in some cases, changes are the result of environmental damage to normal developmental processes, as in teratogenesis. It is important to consider that deleterious effects of environment on phenotype may represent subtle changes resulting from developmental disruption rather than an adaptation to better suit an environment. Experimental evidence must take this into account; however, developmental plasticity which occurs at a certain point may confer an advantage at one point but a change in environment may render that adaptation harmful, especially if it occurs after the window of plasticity, preventing a reversal of the phenotypic expression. This trade-off is the basis of the thrifty phenotype hypothesis.
Waddington described ‘homeorhesis’ in 1957 (cited by Gluckman and Hanson ) as long-term change to a phenotypic developmental trajectory as a response to prolonged exposure to an adverse environment, as opposed to a short-term change in response to a brief environmental exposure that has no extended consequences. A trade-off exists when a change has immediate advantage but later disadvantage. This is best illustrated in humans by two observations. First, children born after experiencing fetal growth restriction (FGR) who have slower growth trajectories postnatally enter puberty earlier than their normally grown counterparts. A trade-off exists that allows survival under poor nutrition but results in poor postnatal growth. Earlier reproductive development confers an evolutionary advantage by allowing the individual to reproduce even under poor developmental conditions. Second, growth-restricted fetuses may undergo ‘rescue by preterm birth’ in which the initiation of preterm delivery is a trade-off which allows the fetus to escape the hostile intrauterine environment and stillbirth but exposes it to the challenges of prematurity.
Predictive Adaptive Responses
Gluckman and Hanson described a different form of developmental plasticity in which the developing organism makes phenotypic adaptations better suited to the predicted adult environment, anticipated from signals within the intrauterine environment. These ‘predictive adaptive responses’ (PARs) do not confer benefit to the developing organism but, rather, better equip the organism for adult life. The association with adult survival advantage requires the prediction of the adult environment to be accurate. If an inappropriate PAR is made, this may manifest as altered disease risk, as in the metabolic disease associated with FGR. Accurate predictions of adult environment may be impaired by disordered maternal–placental–fetal signalling in the case of maternal–placental disease. Here, the fetus undernourished by a dysfunctional placenta prepares (by becoming insulin and leptin resistant, among other adaptations) for an extrauterine life of nutritional paucity but, when born into a mismatched environment of caloric plenty, develops metabolic disease.
Postnatal ‘Catch-up’ Growth
Prader and colleagues coined the term ‘catch-up growth’ as a description of increased growth velocity in small children after the treatment of a disease limiting their growth, suggesting that this was a healthy phenomenon. However, in the context of a growth-restricted fetus, growth trajectories during infancy were found epidemiologically to modulate the effect of the abnormal intrauterine growth on adult disease, with the highest risk in children who gained the most weight in childhood ( Fig 47.2 ). These associations were replicated experimentally, with rapid postnatal growth associated with an increase in adult disease. Furthermore, adult disease can be induced by accelerating infant growth even in those of normal birth weight, suggesting that the period in early life during which developmental plasticity acts extends beyond birth into infancy. Given this postnatal plasticity, there may be potential to introduce postnatal interventions to modulate the risk for adult disease.
The ‘Thrifty Genotype’ Hypothesis
The thrifty genotype hypothesis, initially proposed as a purely genetic explanation for regional variation in insulin resistance, suggested that certain populations had evolutionarily selected genetic insulin resistance, which allowed them to cope better with periods of nutritional deficit ( Fig 47.3 ). Neel suggested that this insulin resistance predisposed these populations to type 2 diabetes in the presence of modern-day nutritional abundance. A modification to this hypothesis was proposed in the early days of the DOHaD field whereby the genes thought responsible for the insulin resistance could also explain the observed changes in birth weight. This was certainly biologically valid, given the known interplay between insulin and insulin-like growth factor-1 (IGF-1) and the role of IGF-1 in the regulation of fetal growth. Genetic polymorphisms must play a role in DOHaD, but they are not enough alone to describe all of the epidemiologic evidence, in particular the consistency of associations across ethnic populations and the speed with which DOHaD manifests in socioeconomically transitioning populations. Moreover, the array of evidence for ‘developmental programming’ induced by environmental change highlights the importance of a place for environment in any complete theory.
The ‘Thrifty Phenotype’ Hypothesis
Hales and Barker, in response to the inadequacies of the thrifty genotype hypothesis, proposed the thrifty phenotype hypothesis as one of the first theories to explain the underlying physiology of their observed associations between the fetal environment and adult disease ( Fig 47.4 ). They proposed a mechanism by which a fetus exposed to intrauterine undernutrition grew more slowly, resulting in a trajectory to small growth, and developed insulin resistance and other changes appropriate to a nutritionally poor environment. A developmental mismatch then occurred in later life in the presence of nutritional abundance, where the fetal adaptations increased risk for adult disease.
This theory, however, also does not fully explain the observations from epidemiology. In particular, although it explains the association between severe growth restriction and adult disease, it remains uncertain if it can account for observed changes in disease risk across the range of normal growth. For example, why should an individual with a birth weight on the 30th centile have an increased risk compared with one on the 70th centile, as the early epidemiologic data suggested, when both are clearly ‘normally’ grown at birth? Regardless, it is increasingly clear that the relationships between birth weight and subsequent disease risk extends across the birth weight spectrum rather than being confined to the extremes. Furthermore, this hypothesis does not explain observed changes in adult systems that provide no apparent fetal survival advantage.
The Fetal Insulin Hypothesis
Hattersley and Tooke proposed the fetal insulin hypothesis along similar lines to the thrifty genotype hypothesis, suggesting that genetically determined insulin resistance could explain both FGR and later glucose intolerance. Evidence for this hypothesis arose through the study of monogenic conditions where a single defect in a gene results in disease. Maturity-onset diabetes of the young (MODY) is a rare form monogenic diabetes characterised by single mutations in a range of individual genes. MODY-causing gene variants have been directly implicated in lower birth weight for affected individuals. Neonates with HNF1β MODY mutations have been shown to be up to 900 g lighter than negative control participants.
Further support for a genetic role in fetal growth has come from genome-wide association studies of birth weight, which have demonstrated an association between a specific allelic variant in the ADCY5 gene and reduced fetal growth. This variant had previously been demonstrated to be associated with an increased risk for adult type 2 diabetes mellitus (T2DM).
Neither of the ‘thrifty’ hypotheses fully explains that which we observe epidemiologically and experimentally. Clearly, genotype is important because not all individuals of low birth weight go on to develop adult disease. Similarly, environment also plays a role given the rapid changes observed during short periods of famine and with rapid population socioeconomic transformation. Wells also suggests that the two hypotheses are not mutually exclusive, with genetics possibly explaining the long-term changes in entire populations and environment explaining acute changes in individuals.
In defence of the importance of environment underlying the DOHaD phenomenon, Hales and Barker acknowledged the role of genetics in every human disease. Indeed, they conceded that malaria and tuberculosis, among the most environmental of diseases, still affected individuals differently depending on their genetic susceptibility. However, they maintained, environmental exposure to the pathogen was the key factor in disease and compared this to gene–environment interactions in DOHaD. By contrast, Eriksson et al. elegantly demonstrated the importance of gene–environment interactions underlying DOHaD in demonstrating an association with the combination of PPARG gene polymorphisms and low birth weight with adult T2DM. In this study, neither the polymorphisms nor low birth weight alone were associated with T2DM, but the combination of gene and environment resulted in the phenotype of disease. This confirmed, at least for this example, the combined roles for genetics and environment underlying DOHaD phenomena.
Genomic Variation and Epigenetics
The two main areas of genetic interest in the DOHaD field are genomic variation and epigenetics. The genes that make up our genetic code vary in size from a few hundred DNA bases to more than 2 million. Each autosomal gene is represented by two copies, or alleles, one inherited from each parent. Although most alleles are the same, about 1% vary among people (genetic polymorphism). Alleles vary through small difference in their DNA sequence. The major allele refers to the one most commonly displayed in the population. These small genetic differences can contribute to each person’s unique features. In their rare forms, as described for monogenic diabetes, they can be disease causing. Genomic polymorphism may affect gene function in two main ways. First, DNA sequence variation in coding regions of genes can result in amino acid substitutions, which can change the structure and function of the protein products of the gene. Second, and probably more commonly, polymorphisms within noncoding regions of DNA, such as transcription factor binding sites (and other regulating regions), may affect the expression of proteins, with more quantitative than qualitative effects on the protein product of the gene.
Genomic variation is clearly important in DOHaD. Several studies have demonstrated associations between genetic polymorphisms, abnormal fetal growth and adult disease. The genetic regulation of both fetal growth and adult metabolic disease is complex, involving multiple genes, and it is evident that the genetic mechanisms underlying their association will also be complex, with multiple variations across multiple genes each making a small contribution to the phenotypic outcome.
Epigenetic phenomena, on the other hand, do not alter the genomic DNA sequence but involve changes to the structure and function of the DNA complex. Epigenetic mechanisms enable developing organisms to produce disparate cellular phenotypes from the same genotype. These modifications take the form of DNA methylation, chromatin remodelling, covalent modifications to histones, parental imprinting and X-chromosome inactivation.
Epigenetic events may explain the relationship between an individual’s genetic background, the environment, aging and disease. Although a DNA sequence always remains the same, cells in a specific tissue have the ability to vary their epigenetic state and hence gene expression through their life course. It is therefore likely that these mechanisms are important in DOHaD and that they interact with each other.
Traditionally, genomic variation and epigenetics have been seen as competing theories of DOHaD mechanisms, but in reality, the two are probably interdependent. For example, DNA methylation is controlled by the genes encoding the DNA methyltransferase enzymes, and evidence links polymorphisms in these genes with altered DNA methylation patterns in endometrial, pancreatic, gastric and colorectal cancers.
Technological advances in epigenomic sequencing have provided a better understanding of the complex interplay between genotype and epigenotype and have more firmly established that genetic polymorphisms can affect methylation. For example, the gene SERPINA3 (Serpin peptidase inhibitor clade A member 3) encodes a protease inhibitor involved in a wide range of biological processes. It has been found to be upregulated in human placental diseases (specifically growth restriction) in association with hypomethylation of a regulating part of the gene in the presence of a specific genetic variant.
This interaction between methylation and genotype which could lead to alteration of gene expression and therefore disease provides a novel insight into how communicating genetic and epigenetic mechanisms could be involved in placental disorders such as FGR.
Genomic imprinting is a process through which the expression of a gene is dependent on the sex of the parent from which it was inherited. Unlike the majority of genes in which both alleles are expressed in imprinted genes, only one allele is expressed. If the allele inherited from the father is imprinted, it is thereby silenced (paternally imprinting), and only the allele from the mother is expressed. If the allele from the mother is imprinted, then only the allele from the father is expressed. Imprinted genes are highly conserved and expressed in the placenta. These genes are effectively driven by epigenetic regulation theorised to be controlled by differential methylation of the relevant genes. It has been hypothesised that sporadic loss of imprinting induced through methylation change could occur in human placentas and contribute to abnormal placental development and consequentially decreased fetal growth. Within the placenta, regulation of imprinted gene expression appears to be less stable than in the fetus itself. This allows the placenta to better adapt to changing physiological environments but with potential adverse consequences such as poor growth.
The paternally expressed insulin-like growth factor 2 ( IGF2 ) gene is imprinted and is a major modulator of both placental and fetal growth. IGF2 is regulated by a differentially methylated binding region known as imprinting control region 1 (ICR1). In growth-restricted placentas, the ICR1 region has been found to be significantly less methylated with resultant reduced expression of IGF2 compared with normal grown placentas . Decreased placental methylation at the IGF2 imprinting control region is associated with intrauterine growth restriction but not preeclampsia and complete loss of placental IGF2 expression is associated with FGR in mice.
Epigenetics and Other Interactions.
Maternal diet is one of the ways in which the supply of nutrients vital to DNA methylation may be affected. Famine provides an extreme example of disruption to the supply of nutrients essential in the methylation pathway. Individuals exposed to famine in utero have been shown to have lower birth weights, and furthermore, exposure to famine during any stage of gestation was associated with glucose intolerance later in life. Individuals born during the Dutch famine born 60 years ago were later found to have less DNA methylation of the imprinted IGF2 gene compared with their famine-unexposed same-sex siblings. This observation supports the hypothesis that exposure to adverse environmental factors in utero could permanently alter epigenetic marks, which might have a bearing in adult disease risk.
DNA Methylation in Placental Diseases.
DNA hypomethylation at gene enhancer regions has been identified between placentas from pregnancies affected by pre-eclampsia and control participants. In a small cohort of term pregnancies affected by FGR, methylation differences were found in the pathway involving hepatocyte nuclear factor 4α ( HNF4A ) gene. Because HNF4A is a candidate gene for T2DM, this observation opens up the possibility that differential methylation of genes important in diabetes might also play a role in fetal growth.
Animal Models of DNA Methylation in Fetal Growth Restriction.
In an animal model of FGR, cytosine methylation in pancreatic islet cells was differentially methylated at approximately 1400 loci in male rats at 7 weeks of age, before they went on to develop diabetes. Of the top 53 genes, almost half of those tested were associated together in a single functional network centred on a collection of important metabolic and cellular regulators. The majority of changes occurred in evolutionarily conserved DNA sequences with some loci in proximity to genes manifesting changes in gene expression which were enriched near genes regulating vascularisation, proliferation and insulin secretion.
Furthermore, in growth-restricted rodent offspring, histone modifications in association with DNA methylation differences have been identified in the promoter region of the gene PDX1 , eventuating in permanent gene silencing and a diabetic phenotype. In health, PDX1 regulates pancreatic β-cell differentiation, but when silenced, T2DM eventuates. This study offered a new insight in epigenetic mechanisms linking growth restriction to diabetes development.
Given the ability of sequence variation in genes influencing epigenetic changes to affect phenotypic outcome, it is likely that sequence variation will underlie some of the association seen between epigenetics and DOHaD, with an individual’s susceptibility to epigenetic changes being controlled to some extent by their genotype.
Breastfeeding appears to be protective against the development of adult obesity. This protection is dose responsive, with greater benefit in those breastfed for longer durations and those exclusively breastfed without formula supplementation. The precise underlying mechanism by which infant nutrition modifies obesity risk is difficult to elucidate given the multiple biological, environmental, genetic, epigenetic and social contexts. There are likely to be significant contributions from all of these factors ( Fig 47.5 ). So far there is evidence to demonstrate an impact of early nutrition upon metabolic programming, upon epigenetic modification within the IGF axis and upon hormonal pathways including insulin and adiponectin.