Epigenetic Basis for Fetal Origins of Age-Related Disease

Introduction

Perturbations of the intrauterine environment can have major effects in determining long-term disease susceptibility, particularly in regard to type 2 diabetes and cardiovascular disease (CVD). Although the mechanism for this remains imprecise, permanent alterations in gene expression implicate epigenetic regulation, which may serve as a biological memory of an aberrant intrauterine milieu. The involvement of multiple organ systems and susceptibility to chronic disease are analogous to the decline in resistance to disease that is typical of normal aging. Here, we describe the current evidence for the role of epigenetic dysregulation in fetal origins of adult disease. In addition, we present a new perspective on the induction of epigenetic alterations in utero, which create an adult phenotype that is more susceptible to age-related diseases.

Barker's Hypothesis and Epidemiological Support of Adult Disease Risks

The thrifty phenotype hypothesis, first proposed at a keynote lecture in Dublin in 1991, ¹ stated that “poor fetal and early post-natal nutrition imposes mechanisms of nutritional thrift on the growing individual.” ² This idea, particularly novel at the time, unified a variety of otherwise disconnected research and explained that inadequate early nutrition can have long-lasting consequences, most notably an increased susceptibility to the development of type 2 diabetes (T2DM) and CVD. Barker's hypothesis originally evolved from epidemiological observations that the same regions in England that had the highest rates of mortality from coronary heart disease (CHD) in the 1980s had the highest infant mortality rates 70 years earlier. ³ With low birth weight as the most common cause of infant mortality in the early 20th century, Barker reasoned that low birth weight babies who survived infancy harbored an increased risk of CHD later in life. In particular, poor nutrition during fetal development was postulated to cause permanent changes in the structure and function of specific organs and tissues, with the pattern of metabolic and functional abnormalities seen in later life dependent on the precise timing and pattern of developmental deficiency.

The hypothesis was also applied in the context of T2DM, where poor fetal nutrition was postulated to cause a quantitative deficiency of pancreatic β cells or, alternatively, change the cells themselves or the overall pancreatic islet architecture. In order to explain the diabetic consequences of nutritional thrift during fetal development, Barker invoked Neel's thrifty genotype hypothesis, which asserts that diabetogenic gene(s) persist at high levels in populations exposed to famine because they may confer survival advantage during times of nutritional deprivation. ⁴ The resultant thrifty phenotype hypothesis explained that diabetes may be a consequence of intrauterine undernutrition when developmental adaptations that occur to compensate for the paucity of nutrients early in life become maladaptive later. Since then, the concept of predictive adaptive response was developed, which suggests that the developing fetus may acquire lifelong changes that are adaptive for a nutritionally poor postnatal environment and, conversely, maladaptive for a subsequent environment of abundance or excess. ⁵ This concept was illustrated by an additional study that found that almost 10% of lean, famine-exposed Ethiopian Jews developed diabetes shortly after immigrating to Israel. ⁶

A large number of clinical studies using completely independent cohorts have since confirmed the relationships between deficient intrauterine growth and subsequent disease risk, correcting for many confounding variables. Similar studies in Finland, ⁷ Norway, ⁸ Sweden, ⁹ Scotland, ¹⁰ and the entire U.K. ¹¹ confirm this association in independent populations, with subjects varying in age from 45 to 76 years. In addition to these studies of predominantly Caucasian Western European populations, investigators in Japan, ¹² China, ¹³ and India ¹⁴ concluded that intrauterine growth restriction (IUGR) increases diabetes risk in adult life. A significant association between IUGR and other diseases, including hypertension, ¹⁵ CVD, ^{16 –19} stroke, ²⁰ and osteoporosis, ²¹ has also been found. Taken together, these studies demonstrate that low birth weight, and IUGR in particular, predisposes to a constellation of age-related diseases. This pathological association is independent of many secondary factors, including country of origin, socioeconomic status (SES), physical activity, and others. Moreover, the correlation between impaired fetal environment and later onset of disease is both reproducible and robust.

Rodent Models of Intrauterine Growth Restriction

A number of animal models have been developed in order to better understand the molecular basis of the relationship between early growth restriction and the subsequent development of disease. These model systems, while recapitulating some of the major hallmarks of IUGR, provide a means of controlling the environment (e.g., nutrition) and genetic background, either or both of which may otherwise affect the outcome of a study. Severe maternal calorie restriction models induce significant growth restriction with slight hypertension in adulthood, ²² elevated fasting plasma insulin concentrations, and profound hyperphagia increasing with age. ²³ Less severe maternal calorie restriction causes 70% reduction of β cell mass by 3 weeks and loss of glucose tolerance by 12 months. ²⁴ Development of the endocrine pancreas is particularly susceptible to amino acid levels, and maternal protein restriction, another model of IUGR, causes a reduction in β cell proliferation, islet size, islet vascularization, islet function, and pancreatic insulin content. ²⁵ Maternal protein restriction induces structural and functional changes in the liver, ²⁶ and the offspring go on to develop a constellation of physiological changes, including altered hepatic glucose production, ²⁷ adipose depot-specific insulin resistance, ²⁸ hypertension, ²⁹ and progressive loss of glucose tolerance. ³⁰ Similar findings have been shown in other models of vascular injury-induced IUGR (bilateral uterine artery ligation) in rodents. ^31,32

Developmental Plasticity

The principle that the intrauterine environment is somehow able to instigate a stable thrifty phenotype in the offspring is actually part of a larger conceptual framework of developmental plasticity. Developmental plasticity refers to the phenomenon by which one genotype can give rise to a variety of different physiological or morphological states in response to different environmental conditions during development. Just as is seen with IUGR, additional instances of developmental plasticity suggest that there exists a critical window when a system is uniquely sensitive to the environment, after which point the developmental program and future functional capacity become fixed.

Although research pertaining to the concept of fetal origin of adult disease has largely been focused on fetal growth restriction, paradoxically, infants with excessive growth or large for gestational age (LGA) infants are also at increased risk for T2DM, obesity, and CVD. ^{33 –35} Animal models of maternal high-fat feeding have shown that the offspring have characteristics resembling the metabolic syndrome, ³⁶ increased adiposity, ^37,38 abnormal glucose tolerance, ^37,39 elevated blood pressure, ^40,41 and abnormal lipid profiles. ^37,40 The most overwhelming population studies come from those of the Pima Indians, which show that children born to mothers with diabetes during pregnancy are more likely than their unexposed siblings to have insulin resistance. ⁴² In the increasingly obesigenic environment that exists today, ⁴³ recent studies of populations with high rates of obesity have demonstrated associations between high birth weight and T2DM. ^44,45 Thus, the relationship between birth weight and risk for all causes of mortality in contemporary populations is U-shaped, with increased risk of premature death in adults born with both the lowest and highest birth weights. ³⁵

The impact of overgrown infants has important implications for future generations because although the percentage of infants weighing <2500 g at term (considered small for gestational age [SGA]) remains stable, the number of LGA infants is increasing substantially. In the United States, full-term infants weighing >4 kg increased from 9.3% to 11.7% between 2000 and 2006. ^46,47 In some countries, up to 20% of infants have birth weights > the 90th percentile based on traditional nomograms. ⁴⁸ The increase in prevalence of LGA infants likely reflects the increasing prevalence of overweight and obesity in reproductive age women and related comorbidities, such as gestational diabetes. The distressing rise in obesity earlier among children and young adults foreshadows escalating concerns as these individuals mature. ⁴⁹

Biological Memory: Permanent Changes in Gene Expression

Although the exact mechanism for the memory of early life events remains unclear, persistent and potentially progressive changes in gene expression implicate a fundamentally biological basis. These changes have been shown to affect a number of metabolically active tissues. For instance, gene expression programs in the endocrine pancreas appear to be particularly sensitive to the in utero environment in all major animal models. Offspring of maternal protein-restricted rodents have impaired pancreatic development secondary to decreased β cell proliferation, ²⁵ with stably reduced expression of the pancreatic duodenal homeobox-1 (Pdx-1) transcription factor, measured both at birth and at 1 month of age. ⁵⁰ Pdx-1, which is required for pancreatic development and regulates expression in the mature pancreas, ⁵¹ is similarly dysregulated in pancreatic islets from other rodent models with persistent and progressive decreases in mRNA levels in IUGR animals. ^52,53 Although β cell proliferation is reduced in parallel with stable changes in Pdx-1 expression, increased levels of apoptosis also contribute to decreased islet size and function. ⁵⁴ No corresponding studies have questioned the relationship of IUGR to altered gene expression programs in the human pancreas; however, the critical role of Pdx-1 in pancreatic development and functional maintenance is well accepted. ⁵⁵

IUGR also has long-term effects on expression in and function of the liver, ^{32,56 –59} adipose tissue, ^{60 –64} skeletal muscle, ^{65 –68} kidney, ^{69 –71} and the central nervous system, in particular the hypothalamus. ^{72 –74} Similarly, changes in gene expression have been seen in many rodent models of maternal nutritional excess. ^{36,38,39,41,75}

In summary, IUGR and maternal nutrient excess cause a number of significant changes in gene expression, demonstrated in a variety of animal models, at a variety of time points, and in a variety of metabolically active tissues. These changes in gene expression are often apparent at birth and are maintained through juvenile and adult life, sometimes associated with progressive dysregulation. Although the gene expression programs in many tissues each appears to be affected differently following in utero stress, these changes in expression have been robustly described and are characterized as persistent alterations originating during a period of transient sensitivity, most often in fetal life.

Epigenetics and the Fetal Origins of Age-Related Disease

The persistence of developmentally induced effects on genetic programs throughout life suggests an epigenetic phenomenon. Epigenetics is a composite term using the Latin prefix “epi” and literally translates to “on” or “upon” genetics. Strictly speaking, a gene under epigenetic control is defined as one regulated by a mechanism that is independent of the DNA sequence itself. Instead, epigenetic control relies on additional information associated with the DNA, which may govern the relative accessibility of a gene to transcription. Every cell in the body contains a near-identical blueprint (i.e., genomic DNA sequence), but this DNA blueprint is regulated epigenetically and used very differently depending on the structural and functional requirements of the cell.

Studies of methyl donor dietary supplementation provide additional support for an epigenetic basis of IUGR-mediated effects, with particular emphasis on DNA methylation. Maternal methyl donor supplementation, which increases offspring DNA methylation at both the Axin-fused and agouti loci in mice, ^76,77 was shown to modulate IUGR-induced epigenetic changes in both the F₁ and F₂ generations. Methyl donor (i.e., folic acid) supplementation of maternal protein-restricted rats protected the immediate offspring from IUGR-associated changes in both methylation and gene expression at a number of loci, including PPARα. ^78,79 However, these changes were not reversed by postnatal folate supplementation. ⁸⁰ Maternal methyl donor supplementation also prevented the transgenerational amplification of obesity inherited with DNA methylation state at the agouti locus. ⁸¹

Conversely, maternal exposure to folic acid antagonists (e.g., sulfamethoxazole-trimethoprim or phenobarbital) significantly increased the risk of adverse pregnancy outcomes and fetal growth restriction in humans. ⁸² Intrauterine treatment with 5-AZA-2′-deoxycytidine, a potent DNA demethylating agent, similarly induced growth restriction in F₁ mice, with consequent lower serum insulin-like growth factor-1 (IGF-1) levels in adult males (measured at 5 months). ⁸³ Taken together, these studies support the role of epigenetics in mediating the effects of IUGR, and direct studies of the epigenome in IUGR animals reveal a number of global as well as specific changes in DNA methylation and histone tail marks. ^{84 –87}

Preliminary studies in humans and nonhuman primates also suggest that IUGR affects both DNA methylation and histone tail modification. A follow-up study of the Dutch Famine cohort revealed that individuals who had been prenatally exposed to famine had ∼5% lower levels of DNA methylation at the IGF2 DMR than their unexposed, same-sex siblings. ⁸⁸ In a separate study, loss of imprinting at the H19 differentially methylated regions (DMR) with IUGR in placenta was associated with hypomethylation and biallelic expression of the H19 gene. ⁸⁹ Dysregulation of imprinted gene expression was also observed in a microarray-based study of IUGR and control placentas, with IUGR tissue showing decreased mRNA expression of IGF2, MEST, MEG3, GATM, GNAS, and PLAGL1. ⁹⁰ Finally, in utero exposure to maternal high-fat diet in a nonhuman primate model of IUGR induced site-specific alterations in fetal hepatic H3 acetylation, with corresponding changes in gene expression. ⁹¹

Taken together, these studies show that fetal environment can affect epigenetic states, with long-term consequences for gene regulation and age-related disease. These changes are persistent and potentially progressive throughout life. In some studies, DNA methylation is clearly implicated, whereas in others, changes in histone tail modifications predominate. The observed changes are widespread but site specific and often progressive with age.

Fetal Origins of an Aging Phenoytpe

Normal aging is a heterogeneous, individual-specific process, which is associated with a constellation of progressive and degenerative conditions. The age-related degeneration associated with each specific condition is typically paralleled by accumulated cellular deficits, which underlie the physiological dysregulation that occurs over time. All cells accumulate unrepaired cellular and molecular damage with the repetitive replication of normal aging. In turn, the accumulation of degenerative cellular changes impairs cell function, replicative capacity, and potentially even the cell's ability to sustain itself. Cellular senescence, which is intimately associated with the aging process, can affect tissue function by diminishing the ability of component cells to participate in regular tissue turnover and renewal ⁹² and impairing their maintenance of and interaction with the tissue microenvironment, potentially affecting neighboring cell function as well. ⁹³

Age-related diseases result from an intricate interplay between both genetic and environmental factors. However, studies of monozygotic (MZ) twins, who share nearly identical copies of their DNA, reveal that genetics likely accounts for only a fraction of the phenotypic variability observed (e.g., different onset and extent of cataracts, gray hair). ^94,95 Instead, extensive epigenetic differences seen in MZ twins are likely to play an important role in establishing these phenotypic distinctions, ⁹⁶ and with increasing age, MZ twins exhibit increasingly variable epigenotypes. ⁹⁷ Thus, epigenetic changes accumulate over time and can account for observed interindividual differences in phenotype. ^{97 –99} Applied more broadly, epigenomic dysregulation is likely an important determinant of aging and disease susceptibility.

Similarly, a spectrum of intrauterine conditions may predispose individuals to a constellation of age-related diseases. The induced adult phenotypic traits include altered activity of metabolic pathways, changes in tissue structure and function, and dysregulation of the epigenome. Although the phenotypes induced with seemingly polar intrauterine stresses (i.e., growth restriction and overgrowth) share many features (increased risk for diabetes, adiposity/obesity, CVD) leading to increased risk of premature death, ³⁵ some differences do exist (e.g., increased risk of osteoporosis in IUGR ²¹ and increased risk for certain cancers in LGA). ³⁵ The concurrence of increased susceptibility to chronic disease, involvement of multiple organ systems, and epigenetic dysregulation is analogous to the normal decline of resistance to disease that occurs with aging. The induction of epigenetic alterations in utero may presage the age of an individual, and therefore, the susceptibility to age-related diseases (Fig. 1).

OPEN IN VIEWER

FIG. 1.

Epigenetic dysregulation paradigm for fetal origins of age-related disease. The overall progression of epigenetic dysregulation with natural aging is shown. The consequences of intrauterine stress (arrowhead with asterisk) on early epigenetic maintenance, which expedites progression toward age-related disease are compared with the natural aging process in individuals unexposed to similar intrauterine stress.

Future research will undoubtedly expand our understanding of the complex epigenetic underpinnings of fetal origins of adult disease. Ultimately, fetal origins of adult disease may come to be viewed in a similar framework as other progressive disorders defined by epigenetic dysregulation, such as aging and cancer. The fundamental mechanisms in which early life events create an adult phenotype that is more susceptible to age-related disease remain to be elucidated. Further, insight into the contributions of early life events to chronic diseases, such as T2DM, may further enhance our understanding of each specific disorder.