Secondary Logo

Journal Logo


Genetic hypothesis for the developmental origins of health and disease theory

Zhao, Xinzhia,b,*

Author Information
doi: 10.1097/JBR.0000000000000056
  • Open



Non-communicable diseases (NCDs), also known as chronic diseases, are characterized by slow progression and long duration. Prominent NCDs include cancer, cardiovascular diseases, diabetes mellitus, and major psychiatric diseases. Over the past two centuries, dramatic improvements in healthcare have increased the global life expectancy and changed the patterns of disease. NCDs account for a substantial proportion of the global disease burden and are currently responsible for over 63% of the world's deaths.[1]

The etiologies of NCDs are generally complex and consist of multiple genetic and environmental factors. The discovery that disorders of intrauterine development were associated with risk of NCDs in adult life was an important finding of the late 20th century. David Baker and colleagues reported an increased rate of death from ischemic heart disease in individuals with low birth weight in an English cohort,[2] and later this association was replicated in several different populations.[3,4] Low birth weight was also found to be associated with disorders characterized by insulin resistance, such as type 2 diabetes (T2D), hypertension, and dyslipidaemia.[5] Based on these observations, Barker formulated the hypothesis of the developmental origins of health and disease (DOHaD), which proposes that exposure to environmental factors during specific sensitive periods of development might predispose an organism to diseases in adult life.[6] An increasing number of studies have connected exposure to stresses during early life after conception, or even during preconceptional gametogenesis, with increased risks for chronic conditions.

Although the association has been well established by many epidemiology studies, the underlying mechanisms of the DOHaD theory are not fully understood. Some important hypotheses are based on developmental plasticity.[7] Many organisms have been found to express adaptive responses to their environments and to form specific characteristics. However, if the environment changes, these characteristics could result in maladaptation, and could increase the risk of contracting disease. Hales and Baker[8] put forward the thrifty phenotype hypothesis to explain the associations between poor fetal and infant growth and an increased risk of developing impaired glucose tolerance or metabolic syndrome in adult life. The hypothesis states that poor nutrition in fetal and early infant life induces specific programming of development and function of pancreatic β-cells, as well as kidney, muscle, liver, vascular, hypothalamic-pituitary-adrenal axis, and sympathetic systems. The programming leads to a thrifty phenotype with reduced fetal growth and poor organic function. These individuals would be well adapted to living in a continuously poorly nourished state, but would be maladapted to overnutrition in adult life, putting them at risk of developing a metabolic syndrome.

An alternative explanation to the DOHaD theory is genetic pleiotropy, the phenomenon in genetics whereby a DNA variant influences multiple traits. Given the limited number of genes in the human genome, and the enormous dimensionality of the phenome, it is reasonable to expect that many functional variants have pleiotropic effects.[9] Those genetic variants that affect development and influence adaptive responses to environmental stresses at an early age may also contribute to the genetic risk of NCD later on in life. The association between changes early in development and the onset of NCDs later in life may be a coupled phenomenon caused by a common mechanism.

In this review, I focus on the following potential genetic factors in the DOHaD theory, 1) the genetic variants linking early life development, mainly during intrauterine growth, and the risk of NCDs, 2) possible antagonistic pleiotropy and positive/negative selection for these genetic variants, and 3) the interaction between genetic and environmental factors. I also discuss how the growing population of individuals born as a result of assisted reproduction provides a unique opportunity to investigate the genetic mechanisms of DOHaD theory.

Database search strategy

In this review, the author performed a search for original and review articles published in English between January 1989 and June 2019 focusing on genetic hypothesis of DOHaD theory in PubMed database. The following search terms were used alone or in combination: “genetics in developmental origins of health and disease”, “genetic pleiotropy”, “genetics of intrauterine growth”, “genetics of birth weight”, “positive selection”, “assisted reproduction”. The bibliographies of pertinent articles were also examined to identify further relevant papers.

The “thrifty genotype” hypothesis

As early as 1962, the American geneticist James Neel[10] put forth the “thrifty genotype” hypothesis. He proposed that genetic variants that reduce glucose uptake and limit body growth facilitated survival during periods of famine, and therefore these thrifty alleles would be positively selected in preindustrialized societies, in which famine is frequent. When individuals who have the thrifty genotype encounter a modern industrialized environment with plentiful food/low energy expenditure, they are at risk of developing metabolic syndromes. This hypothesis was initially used to explain the increasing prevalence of T2D among many indigenous populations, and was modified in the following years in conjunction with new information regarding the complexity of metabolic disease.[11]

The thrifty genotype hypothesis was applied to the mechanism of DOHaD by Hattersley and Tooke.[12] They proposed the fetal insulin hypothesis, which states that genetic factors are important both as determinants of birth weight and in evaluating the risk of adult metabolic syndromes. In other words, they proposed that the “thrifty phenotype” is the result of a “thrifty genotype”. Insulin pathways are considered to be excellent candidates for a common link because insulin plays a key role in both metabolism syndrome and fetal growth.

The thrifty genotype hypothesis employed an evolutionary framework to explain the growing burden of NCDs in contemporary populations. Nevertheless, many critiques of the hypothesis[13] arose due to the general lack of genetic evidence until recently. However, in the past 15 years, high throughput genomic platforms have generated huge amounts of genetic data enabling numerous genome-wide association studies (GWAS). A number of genetic variants that increase NCDs risk in old age have been found to be associated with increases in survival, fertility or lifetime reproductive success (LRS).

The genetic variants linking early development and adult diseases

Low birth weight has been identified as an important risk factor for many NCDs, and the association is primarily the result of growth retardation, rather than preterm birth.[14] Birth weight is a complex multifactorial trait, and genetic factors play an important role that is independent of the intra-uterine environment.[15] This trait is extremely polygenic, and is controlled by large numbers of loci of small effect. As a result, it is quite difficult to identify the genetic loci that influence birth weight. Fortunately, birth weight is an easy-accessible demographic trait. Meta-analyses of GWAS data from multiple large-scale studies have yielded sufficient power to detect the genetic variants associated with birth weight, and many loci have been identified. A complex correlation between birth weight and adult disease risk has been demonstrated.

Using the GWAS data from a cohort of 5465 Caucasian children with recorded birth weights, Zhao et al[16] investigated the association between birth weight and previously reported T2D-associated variations at 20 loci. They found that the minor allele of a single nucleotide polymorphisms (SNPs) rs7756992 at the CDKAL1 locus was associated with lower birth weight, although the association did not reach genome-wide significance.[16] Later, this finding was replicated in a Danish cohort consisting of 4744 individuals.[17] The researchers also reported that birth weight was inversely associated with the T2D risk alleles of rs11708067 in ADCY5. Freathy et al[18] searched for common genetic variants associated with birth weight in 38,214 individuals of European descent. They identified that rs900400, located near CCNL1, and rs9883204, located in ADCY5, were robustly associated with birth weight. The SNP rs9883204 showed a relatively strong linkage disequilibrium with rs11708067, which implied that it is involved in the regulation of glucose levels and thus T2D susceptibility.[19] The inverse genetic correlations between birth weight and T2D at the ADCY5 and CDKAL1 loci were confirmed in an expanded GWAS (with 69,308 individuals of European descent).[20] These results provided direct support for the fetal insulin hypothesis.[12]

The birth weight loci are not all associated with NCDs. For example, rs900400, which is located near the CCNL1 locus, is not known to be associated with any other traits.[20] Moreover, some alleles associated with higher birth weight may confer risk of NCDs. rs1801253 in ADRB1 has been strongly associated with birth weight, as well as systolic and diastolic blood pressure.[19,21] However, the birth weight-lowering allele at rs1801253 is associated with lower blood pressure in later life.[20] Importantly, birth weight may be influenced by both fetal and maternal genotype, and through the cellular mechanisms of gametogenesis and fertilization, fetal genotype is correlated with maternal genotype (r ≈ 0.5). In some cases, functional alleles may exert different effects on birth weight depending on whether they are carried by the mother or fetus. For instance, rare heterozygous mutations in the glucokinase gene of a fetus may result in a reduction of approximately 530 g in birth weight, while mothers who carry a glucokinase mutation may have offspring that weigh 600 g more than average due to maternal hyperglycemia.[22]

In a recent multi-ancestry GWAS meta-analysis of birth weight in 153,781 individuals,[23] the authors identified 7 previously reported and 53 novel loci that were associated with birth weight at a genome-wide significance level for populations with either European ancestry or trans-ancestry. Further, three of the loci harbored multiple distinct association signals attaining genome-wide significance. The lead SNPs in these loci were almost common variants and they individually had modest effects on birth weight (10–26 g per allele). Using a linkage disequilibrium score regression method for the data from samples with European ancestry, the researchers found that birth weight was inversely correlated with genetic indicators of adverse metabolic and cardiovascular health, reflecting the impact of shared genetic variants that influence both sets of phenotypes. However, they also observed locus-specific heterogeneity in the genetic relationships between birth weight and health-related traits, including the replication of rs1801253 in ADRB1 that the birth weight-lowing allele is associated with lower blood pressure.[20] Meanwhile, the researchers found that both maternal and fetal genetic effects connect birth weight to later T2D risk, albeit acting in opposing directions, by analyzing the loci associated with both traits in mother-child pairs.

These studies provide compelling evidence that fetal genotype plays an important role in early growth, as measured by birth weight. Further, the loci that impact birth weight may contribute to the adult risk of metabolic diseases, providing some support for the idea of the “thrifty genotype”. However, these findings can only explain a small part of the variance in birth weight, and the effect of any individual loci is very small. For example, 62 distinct and significant genome-wide signals that were identified in more than 150,000 individuals account for approximately 2% of the variance.[23] Moreover, the genetic correlation between birth weight and other adult health-related traits is complex, and is likely to be indirectly influenced by the maternal genome.

Positive selection for the risk alleles of susceptibility to NCDs

While NCDs have very clear negative impacts, they also have a large genetic component. The cost to fitness for the individuals carrying risk alleles of susceptibility to NCDs may be relatively small because they would have transmitted the alleles to the next generation before the onset of age-related NCDs. A key point of the “thrifty genotype” hypothesis is that the risk alleles of susceptibility to NCDs may confer advantages during early life in a severe environment, and therefore contribute to a net fitness benefit. These risk alleles may even be positively selected. Natural selection is the principal underlying force molding life via traits associated with survival and reproduction.[24] Therefore, it is possible that the thrifty genes may contribute to sustaining fecundity besides survival during famine, as mentioned by Prentice.[25]

Considering that the quantitative traits associated with early development and NCSs are extremely polygenic, positive selection of the loci may occur via polygenic adaptation, and the selection signals may go largely undetected by conventional methods.[26] An allele that is positively selected may rise in prevalence rapidly such that recombination does not substantially break down the association with alleles at nearby loci on the ancestral chromosome, and so a long haplotype is formed. Based on this theory, an integrated Haplotype Score method has been developed to estimate positive selection for genetic polymophisms.[27] This method is typically used to detect candidate adaptive SNPs where the selected alleles may not have reached fixation, as it is well suited for detecting recent selection signals. Another haplotype-based method for computing nSL scores (the number of segregation sites by length) is more robust than the integrated Haplotype Score and does not depend on a genetic map.[28] Most selected regions have been found to be limited to a specific population,[27] suggesting adaption to local environments. Moreover, LRS, the number of children per parent per lifetime, is a prerequisite for responses to selection, and provides a measure of fitness that combines survival and reproduction.

Although several early studies failed to detect the selection signatures in metabolic genes,[29,30] compelling evidence was found in a Samoan population. Samoan people represent a unique founder population with a high prevalence of obesity. A recent GWAS identified a functional variant, rs373863828, with a large effect size (p.Arg457Gln, meta P = 1.4 × 10−20, 1.36–1.45 kg/m2 per copy of the risk allele) in CREBRF that strongly influences body mass index in Samoan population.[31] Interestingly, while the variant is common in Samoan people, it is extremely rare in other populations. The body mass index-raising allele of rs373863828 showed decreased energy use and increased fat storage in an adipocyte cell model, and therefore was considered to be a thrifty variant. Positive selection of this thrifty variant was observed according to the integrated Haplotype Score and nSL score.

The positive selection of risk alleles associated with NCDs is not limited to obesity or specific populations. Evidence of positive selection was found in 40 of the 76 genes known to be associated with the risk of coronary artery disease in a GWAS of 12 human populations.[32] The coronary artery disease genes under positive selection were enriched in a number of traits associated with reproduction and pregnancy outcomes, including twinning, reproductive timing, lactation capacity, pregnancy loss, intrauterine growth restriction (IUGR), and preeclampsia. Meanwhile, the authors found antagonistic relationships between coronary artery disease and reproductive success in women in the Framingham Heart Study.

Human reproductive behavior including age at first birth and the number of children ever born (same as LRS) is strongly related to fitness. A GWAS that combined genome data from more than 300,000 individuals identified 12 independent loci that were significantly associated with age at first birth and/or number of children ever born.[33] Genetic correlation estimates found that the alleles linked to lower age at first birth in both men and women were associated with a higher genetic risk of smoking and T2D, while the number of children ever born was negatively genetically correlated with years of education and age at first sexual intercourse. No significantly enriched genes, tissue sets, or biological functions were found to be enriched for the genes associated with reproductive behavior. These results indicate that human reproductive behavior is influenced by a mixture of biological, psychological, and socio-environmental factors in contemporary populations.

Although limited in number, some reliable studies have identified specific genes associated with the risk of NCDs that are under positive selection, and may contribute to LRS. The positive signals of selection or association with LRS are generally very weak for a single locus. Large databases that contain information on genomics, fertility, mortality, and health status from hundreds of thousands of individuals are required to identify these genes.

Antagonistic pleiotropy in the maternal and fetal genome

Recent studies have suggested a substantial role of genetics in the association between early development and adult NCDs. However, the effect sizes of the candidate loci are so small that they can only be identified in very large cohort studies, and some locus-specific heterogeneity has also been observed. As mentioned above, the quantitative traits related to early development and NCDs are extremely polygenic and have complex gene-environment interactions. Moreover, evidence of substantial crosstalk between the genomes of parents and offspring has been observed during pre-implantation and development. Here, as an example of how crosstalk between the mother and fetus shapes the intrauterine growth environment and impacts early development, I describe the antagonistic responses between the mother and fetus during adaptation to defective deep placentation.

Human pregnancy is characterized by deep placentation, in which placental trophoblasts invade up to a depth of one-third of the thickness of the myometrium. This process is accompanied by transformation of the uterine spiral arteries from a low-velocity to a high-flow chamber. Transformation of the spiral arteries facilitates increases in uterine blood flow to enable perfusion of the intervillous space of the placenta and support fetal growth. Defective deep placentation, which is defined as a significantly increased number of spiral arteries with absent or partial transformation, results in placental ischemia, and is associated with a spectrum of obstetric disorders including preeclampsia, IUGR, preterm labor, preterm premature rupture of membranes, late spontaneous abortion, and abruptio placentae.[34]

Clinical outcomes of defective deep placentation are determined by two major factors. The first is the degree to which physiological transformation of the spiral arteries is restricted. For example, 90% of the myometrial spiral arteries are fully transformed in normal pregnancy. However, in cases of severe early-onset preeclampsia, which is a pregnancy-induced hypertension syndrome associated with IUGR, only a few spiral arteries may be fully transformed. In late-onset preeclampsia (LOPE) without IUGR, IUGR without hypertension, and preterm labor, defective deep placentation may only partially affect the spiral arteries.[34] The trophoblast invasion plays a key role in the transformation of the spiral arteries. Importantly, the placenta is genetically identical to the fetus, and is semi-allogeneic to the mother. Thus, the trophoblast invasion challenges the maternal immune system.[35] Complete failure of immunoregulatory mechanisms could lead to spontaneous abortion, whereas partial failure could lead to a continued pregnancy with a small, insufficient placenta. Thus, there exists a continuum between abortion, pre-eclampsia, and other disorders associated with deep placentation. HLA-C expressed by trophoblasts can be recognized by decidual killer immunoglobulin-like receptors, in a process involved in maternal-fetal tolerance. Combinations in which the mother carries two killer immunoglobulin-like receptor AA haplotypes and the fetus inherits paternal HLA-C2 haplotypes are the most susceptible to preeclampsia and recurrent miscarriages.[36]

A second factor in determining the clinical outcome in cases of defective deep placentation is the maternal and fetal adaptive responses to the compromised placental blood supply. Unlike diseases in the non-pregnant state, obstetric disorders develop in a unique biological situation in which a mother and fetus with different genomes coexist.[37] Compromised placental blood supply restricts the transportation of oxygen and nutrients, which can lead to ischemia–reperfusion injury and retarded fetal growth. Meanwhile, a series of adaptive responses may occur in placental trophoblast cells: hypoxia-inducible factors may increase as a response to cellular oxygen deprivation; antiangiogenic factors including soluble vascular endothelial growth factor receptor1 (sVEGFR-1) and soluble endoglin may be up-regulated; endoplasmic reticulum stress induced by ischemia–reperfusion injury may suspend protein folding, leading to trophoblast apoptosis; and the production of reactive oxygen species may increase, inducing the release of proinflammatory cytokines and chemokines, as well as trophoblast debris. These placental responses release soluble components into the maternal circulatory system and may trigger a non-specific, systemic (vascular), inflammatory response leading to clinical symptoms in the mother.[38]

Among clinical measures of maternal health status, hypertension may be the most important because it can reflect an important fetal adaptive response – increased placental perfusion. In a meta-analysis examining the relationship between fetoplacental growth and the use of oral antihypertensive medication to treat mild-to-moderate hypertension during pregnancy in 45 randomized controlled trials, treatment-induced decreases in maternal blood pressure appeared to adversely affect fetal growth. Specifically, greater decreases in mean arterial pressure from study enrolment to delivery were associated with a higher risk of IUGR (P = 0.006, 14 trials) and lower mean birthweight (P = 0.049, 27 trials).[39] This suggests that fetal growth benefits from maternal hypertension. In a GWAS on preeclampsia, genetic variants in the fetal genome near the FLT1 gene were significantly associated with the syndrome in 4380 cases and 310,238 controls. Interestingly, the association was strongest in offspring of LOPE patients without IUGR. FLT1 encodes antiangiogenic sVEGFR-1 in trophoblasts. The risk C allele of the lead SNP rs4769613 (P = 5.4 × 10−11) was associated with high concentrations of sVEGFR-1 in maternal blood in control pregnancies (P = 0.04).[40] Maternal plasma levels of sVEGFR-1 are higher in cases of severe preeclampsia compared with cases without severe features, and are also higher in cases of early-onset preeclampsia compared with cases of LOPE.[41] There are several possible explanations for the stronger genetic association in cases of LOPE without IUGR. First, while placental sVEGFR-1 is shed in normal pregnancy, levels are significantly increased in cases of preeclampsia. Thus, sVEGFR-1 may be a physiological regulator of maternal blood pressure that is released by the fetus (placenta) during normal pregnancy. In cases of partially defective deep placentation, placental ischemia tends to be mild. Fetuses that carry preeclampsia-risk alleles at the FLT locus tend to secrete more sVEGFR-1 to elevate maternal blood pressure, which leads to increased placental perfusion and fetal growth. Therefore, these alleles may simultaneously increase the risk of preeclampsia and decrease the risk of IUGR.

Although maternal hypertension may benefit fetal growth, it is likely to be detrimental to the mother. In severe cases, hypertensive crisis could lead to cerebrovascular accident and maternal death. Thus, there might be a maternal response to fetal adaption to placental ischemia. It has been suggested that a maternal inflammatory response may counteract the fetus. A recent study using transcriptome profiling of unsupervised cluster placentas identified 3 clinically probable etiologies of preeclampsia. A subgroup of preeclampsia patients showed severe fetal growth restriction and evidence of maternal antifetal rejection, while the maternal parameters of disease severity, such as blood pressure and proteinuria levels, were less severe in these cases compared with canonical patients.[42] One possible explanation is that the maternal condition is relieved by restricting fetus growth. This view is supported by the observation that resolution of maternal hypertension follows fetal death in some patients with preeclampsia.[43] Our recent study identified a number of consistently hypomethylated probes that were associated with early-onset preeclampsia in different populations. The methylation levels of the validated probes were associated with clinical severity, and the samples with intermediate changes in methylation showed antagonistic fetal/maternal outcomes.[44]

Successful reproduction is the common interest of the mother and fetus. However, conflict between the mother and her fetus takes place when their interests diverge, resulting in antagonistic responses. Variations in these adaptive responses impact pregnancy outcomes. Meanwhile, common phenotypes of pregnancy outcome may represent the manifestation of different adaptive responses. For example, IUGR may be caused by severely defective deep placentation, the insensitivity of the fetus to ischemia, or the tolerance of the mother to placental releasing factors. Thus, individuals born with IUGR may have different genotypes that influence fetal growth and future predisposition to NCDs. This complexity represents a major challenge in identifying the genetic correlation between early development and NCD risk.

Epigenetic modifications link both genetic and environmental factors to long-term health

Epigenetics refers to “the study of molecules and mechanisms that can perpetuate alternative gene activity states in the context of the same DNA sequence.”[45] Epigenetic modifications, including DNA methylation, histone modifications, and noncoding RNAs, modify DNA bases and chromatin. This enables the establishment and maintenance of chromatin states that regulate gene expression transmitted across cell divisions.

For most differentiated somatic cells, epigenetic modifications buffer environmental variations and act as barriers to prevent changes in gene expression and cell identity. However, adaption responses to extreme environmental conditions during specific sensitive periods in early development can induce epigenetic programming, which can lead to long-lasting changes in the epigenome and predispose an organism to disease in adult life.[46]

Epigenetic modifications are thought to be reversible and are governed by a series of writers (that deposit them), readers (that interpret them), and erasers (that remove them). Both genetic and environmental factors are involved in epigenetic variations. Variations in the DNA sequence can influence epigenetic modifications in two ways: 1) changing the target sequences of modifications or 2) altering the genes encoding the epigenetic writers, readers, or erasers. Rare genetic mutations of implicated regions can also cause epigenetic disorders. For example, DNA methylation is one of the best-characterized epigenetic modifications. DNA methylation in the human genome predominantly occurs at the C5 position of cytosine in CpG dinucleotides.[47] Fragile X syndrome is a mental condition caused by the excessive expansion of a CGG repeat in the 5’ untranslated region of the FMR1 gene.[48] More than 200 repeats of the trinucleotide induce mRNA-mediated DNA methylation in the CGG repeat, resulting in gene silencing.[49] DNA methylation is catalyzed by the DNA methyltransferase family, which includes DNMT1, DNMT3A, and DNMT3B (the writers). Rare de novo heterozygous mutations in the DNMT3A gene cause Tatton-Brown-Rahman syndrome, a overgrowth disorder characterized by a distinctive facial phenotype and intellectual disability.[50] Further, somatic mutations of the same gene are frequent in cases of acute myeloid leukemia.[51] Interestingly, two other genes that encode epigenetic writers, EZH2 and NSD1, are both associated with developmental growth disorders and hematological malignancies.[52] MECP2, which binds methylated CpGs, is a reader of DNA methylation that can both activate and repress transcription.[53] Mutations in the MECP2 gene are the main genetic causes of Rett syndrome, a progressive neurologic developmental disorder with X-linked dominant inheritance.[54] The ten-eleven translocation (TET) enzymes, which are encoded by TET1, TET2, and TET3, oxidize the C5 position of cytosines and are important erasers of DNA methylation. Somatic mutations in TET genes are frequent in several kinds of hematological cancer.[55] In contrast, transposable elements are frequent targets of epigenetic silencing. Abnormal epigenetic modifications may activate transposable elements, and therefore influence genome integrity and induce de novo mutations. For instance, human neuronal progenitor cells carrying MeCP2 mutations have been found to have heightened susceptibility to LINE-1 retrotransposition.[56]

Environmental factors can have direct influences on the cellular epigenome. Most epigenetic writers and erasers are enzymes for which activity is mediated by the availability of substrates, cofactors, and allosteric regulators. A number of metabolites derived from diverse metabolic pathways, including the one-carbon metabolism, the tricarboxylic acid cycle, β-oxidation, glycolysis, and hexosamine biosynthesis, are used as substrates and cofactors in epigenetic modifications.[57] Therefore, environmental factors, such as diet, microbiome, temperature, malnutrition, and chemical exposure can affect cellular metabolism and change the activity of chromatin-modifying enzymes, leading to changes in the epigenome. For example, epigenetic changes have been found to persist for periods longer than 60 years in individuals prenatally exposed to the Dutch Hunger Winter.[58]

Among epigenetic modifications, DNA methylation is more stable and more suitable for high-resolution assay platforms than histone modifications or non-coding RNA expression. High throughput platforms based on bisulfite conversion are highly quantitative and reproducible, offering high sensitivity to detect small changes from limited amounts of DNA.[59] Epigenome-wide association studies have found many loci with small (<10%) changes in intermediate methylation levels that are associated with complex phenotypes of NCDs.[60] This paradigm is in stark contrast to the observation of imprinted genes or malignant cells that have large differences in methylation level in genomic regulatory regions and clear gene expression changes. Although some changes in DNA methylation may reflect fetal genetic predisposition to disease,[61] others have been found to be reproducibly associated with specific environmental factors.[62] The presence of differential methylated cytosines in NCDs reflects the comprehensive effect of genetic and environmental factors. Therefore, the DNA methylome in peripheral, easy-to-access tissues in early stages of life, such as the placenta, umbilical cords, and fetal membranes, may be ideal for identifying the biomarkers for both genetic risk and early developmental stress associated with NCDs.

Genetic studies of children born after assisted reproduction

The use of assisted reproductive technologies (ART) has grown dramatically, contributing to millions of successful birth worldwide and accounting for more than 4% of all newborns in some European countries.[63] Assisted reproduction is redefining human society and biology, as many infertile couples may transmit their genes to subsequent generations. The long-term health of children born via ART is still unclear because they are too young to evaluate their incidence of age-related NCDs. However, ART children have a specific genetic background, and they tend to be exposed to some environmental stressors. ART children may be a promising cohort for studying the genetic basis of DOHaD theory.

ART procedures are generally used by couples who are infertile, which is defined as “failure to achieve clinical pregnancy after 12 months or more of regular unprotected sexual intercourse”’.[64] The infertile population is known to have an increased prevalence of NCDs. For example, polycystic ovary syndrome (PCOS), which affects 5% to 20% of women of reproductive age worldwide, is the most common cause of anovulatory infertility.[65] Patients with PCOS are at risk of many other NCDs such as coronary heart disease, stroke, and obesity.[66] Genetic studies have indicated causal roles in PCOS etiology for elevations in body mass index and insulin resistance. However, PCOS susceptibility is associated with alleles that raise the menopausal age and genes involved in DNA repair, suggesting a mechanism resulting in the retardation of ovarian ageing.[67] ART children are likely to inherit genes associated with susceptibility to infertility and other NCDs, or quantitative traits associated with infertility. Meanwhile, the population of parents using ART tends to be older, which is associated with an increased number of de novo mutations in offspring.[68] Moreover, social-economical factors affect the accessibility of ART treatments. ART clinic remains absent, inaccessible or unaffordable for many infertile couples in the world.[69] Those healthy subfertile couples with higher socioeconomic status have more opportunities to transmit their genes. Although the genetic background of ART children is a confounding factor in many epidemiological studies, but it can be overcomed in genetic studies using family-based genetic data. For example, live birth is a key phenotype for ART children. The classic transmission/disequilibrium test can be used to identify the genetic loci associated with live birth in nuclear families and investigate changes of gene pool between parents and offspring.

Compared with naturally-conceived children, ART children are exposed to many environmental stressors associated with the assisted reproductive procedure. For example, a pivotal step in most in vitro fertilization protocols is controlled ovarian stimulation (COS), in which exogenous gonadotrophins are used to retrieve multiple oocytes. COS may be implicated in adverse perinatal outcomes because it involves altering the embryonic genome and epigenome.[70] The oocyte yield after COS is highly variable, and in some cases, more than twenty mature oocytes are acquired in one cycle. However, in a natural menstrual cycle, only one or two oocytes from dominant follicles would undergo ovulation simultaneously. The number of oocytes retrieved after COS, which is influenced by genetic factors,[71] is positively associated with the live birth rate up to 20 oocytes.[72] However, the incidence of severe ovarian hyperstimulation syndrome, which is associated with adverse developmental outcomes in offspring, increases significantly with the number of oocytes, particularly if more than 18 oocytes are retrieved.[72] Children born to ovarian-hyperstimulated women display cardiovascular dysfunctions and reduced intellectual ability.[73,74] Therefore, the genetic factors associated with oocyte number in COS have pleiotropic effects: although women who carry the genes associated with a higher number of oocytes retrieved have a greater chance of giving birth to a child after assisted reproduction, their children may have a greater risk of NCDs later in life. Besides COS, other steps in ART, including in vitro maturation of oocytes, intracytoplasmic sperm injection, in vitro culture of the embryo, cryopreservation and thawing of the gamete and embryo, and preimplantation genetic testing, may also negatively impact the early development of ART children.[75] Population-level changes in the genetic responses to these environmental factors would be very informative with respect to the mechanisms of development.

Although it is quite difficult to conduct detailed investigations regarding implantation failure events before initiation of pregnancy in naturally-conceiving women, such medical records often exist for individuals in the ART cohort. Thus, it is possible to investigate the genetic factors influencing success of embryo implantation leading to pregnancy in ART-treated women. A genetic association study found that a common functional p53 Pro allele was associated with increased rates of blastocyst implantation failure in in vitro fertilization patients. Selected alleles in SNPs in the LIF, Mdm2, Mdm4, and Hausp genes, each of which regulates p53 levels in cells, were also enriched in in vitro fertilization patients.[76] Despite these costs, the p53 Pro allele is associated with an increased lifespan, and previous positive selection for alleles in the p53 pathway has been identified in contemporary white and Asian populations. ART itself may also influence human evolution, as has been previously discussed.[77] There are several steps of natural and artificial selection during ART treatment and favored traits may differ between natural reproduction and ART. As a result, only approximately 5% of fresh oocytes produce a baby in ART cycles.[78] Identifying the genetic factors that confer benefits to reproductive success during ART and their potential pleiotropic effects on NCDs will be helpful in predicting the long-term health of ART children. Additionally, such information may also provide insight regarding the genetic mechanisms of DOHaD theory.


The DOHaD theory was formulated based on the well-established association between adverse early environmental events and the increased risk of adult NCDs. In this article, I have reviewed the genetic hypothesis of the DOHaD theory, namely, that genetic variants associated with early development have pleiotropic effects that contribute to the risk of NCDs. This hypothesis is not incompatible with the emphasis on environmental factors in the DOHaD theory, as evidence suggests that the environment of early human development is shaped by crosstalk between the fetal and the maternal genome. Moreover, the interplay between DNA sequence variations and environmental stresses programs the epigenome, resulting in long-term effects on phenotypes. Interestingly, some risk alleles of NCDs are under positive selection, suggesting that they play beneficial roles in reproduction and early development.

Until now, few genetic loci linking early development and adult NCDs had been identified. Both kinds of phenotypes are highly polygenic and any associated genetic loci have very small effect sizes. Importantly, the genetic correlation between early development and adult NCDs will be caused directly by the genome of the child, or indirectly by the genomes of the parents. New methods are needed to improve the power of screening for genetic correlations between pleiotropic effects. Meanwhile, certain intermediate phenotypes during reproduction may help to identify subclasses of cohorts.



Author contributions

XZ drafted and revised this manuscript and approved the final submission.

Financial support

This work was supported by the National Key Research and Development Program of China (No. 2018YFC1005001), the National Natural Science Foundation of China (No. 81871180) and the Innovative Research Team of High-Level Local Universities in Shanghai, China; all to XZ.

Conflicts of interest

The authors declare that they have no conflicts of interest.


[1]. GBD 2016 Causes of Death Collaborators. Global, regional, and national age-sex specific mortality for 264 causes of death, 1980–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017;390:1151–1210.
[2]. Barker DJ, Winter PD, Osmond C, et al Weight in infancy and death from ischaemic heart disease. Lancet 1989;2:577–580.
[3]. Leon DA, Lithell HO, Vâgerö D, et al Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of 15 000 Swedish men and women born 1915–29. BMJ 1998;317:241–245.
[4]. Fan Z, Zhang ZX, Li Y, et al Relationship between birth size and coronary heart disease in China. Ann Med 2010;42:596–602.
[5]. Hales CN, Barker DJ, Clark PM, et al Fetal and infant growth and impaired glucose tolerance at age 64. BMJ 1991;303:1019–1022.
[6]. Carpinello OJ, DeCherney AH, Hill MJ. Developmental origins of health and disease: the history of the barker hypothesis and assisted reproductive technology. Semin Reprod Med 2018;36:177–182.
[7]. Hanson MA, Gluckman PD. Early developmental conditioning of later health and disease: physiology or pathophysiology? Physiol Rev 2014;94:1027–1076.
[8]. Hales CN, Barker DJ. The thrifty phenotype hypothesis. Br Med Bull 2001;60:5–20.
[9]. Gratten J, Visscher PM. Genetic pleiotropy in complex traits and diseases: implications for genomic medicine. Genome Med 2016;8:78.
[10]. Neel JV. Diabetes mellitus: a “thrifty” genotype rendered detrimental by “progress”? Amer J Hum Genet 1962;14:353–362.
[11]. Reddon H, Patel Y, Turcotte M, et al Revisiting the evolutionary origins of obesity: lazy versus peppy-thrifty genotype hypothesis. Obes Rev 2018;19:1525–1543.
[12]. Hattersley AT, Tooke JE. The fetal insulin hypothesis: an alternative explanation of the association of low birthweight with diabetes and vascular disease. Lancet 1999;353:1789–1792.
[13]. Gosling AL, Buckley HR, Matisoo-Smith E, et al Pacific populations, metabolic disease and ’Just-So Stories’: a critique of the ’Thrifty Genotype’ hypothesis in Oceania. Ann Hum Genet 2015;79:470–480.
[14]. Barker DJ, Osmond C, Simmonds SJ, et al The relation of small head circumference and thinness at birth to death from cardiovascular disease in adult life. BMJ 1993;306:422–426.
[15]. Klebanoff MA, Mednick BR, Schulsinger C, et al Father's effect on infant birth weight. Am J Obstet Gynecol 1998;178:1022–1026.
[16]. Zhao J, Li M, Bradfield JP, et al Examination of type 2 diabetes loci implicates CDKAL1 as a birth weight gene. Diabetes 2009;58:2414–2418.
[17]. Andersson EA, Pilgaard K, Pisinger C, et al Type 2 diabetes risk alleles near ADCY5, CDKAL1 and HHEX-IDE are associated with reduced birthweight. Diabetologia 2010;53:1908–1916.
[18]. Freathy RM, Mook-Kanamori DO, Sovio U, et al Variants in ADCY5 and near CCNL1 are associated with fetal growth and birth weight. Nat Genet 2010;42:430–435.
[19]. Dupuis J, Langenberg C, Prokopenko I, et al New genetic loci implicated in fasting glucose homeostasis and their impact on type 2 diabetes risk. Nat Genet 2010;42:105–116.
[20]. Horikoshi M, Yaghootkar H, Mook-Kanamori DO, et al New loci associated with birth weight identify genetic links between intrauterine growth and adult height and metabolism. Nat Genet 2013;45:76–82.
[21]. Johnson AD, Newton-Cheh C, Chasman DI, et al Association of hypertension drug target genes with blood pressure and hypertension in 86,588 individuals. Hypertension 2011;57:903–910.
[22]. Hattersley AT, Beards F, Ballantyne E, et al Mutations in the glucokinase gene of the fetus result in reduced birth weight. Nat Genet 1998;19:268–270.
[23]. Horikoshi M, Beaumont RN, Day FR, et al Genome-wide associations for birth weight and correlations with adult disease. Nature 2016;538:248–252.
[24]. Roff DA. Contributions of genomics to life-history theory. Nat Rev Genet 2007;8:116–125.
[25]. Prentice AM. Early influences on human energy regulation: thrifty genotypes and thrifty phenotypes. Physiol Behav 2005;86:640–645.
[26]. Pritchard JK, Di Rienzo A. Adaptation - not by sweeps alone. Nat Rev Genet 2010;11:665–667.
[27]. Voight BF, Kudaravalli S, Wen X, et al A map of recent positive selection in the human genome. PLoS Biol 2006;4:e72.
[28]. Ferrer-Admetlla A, Liang M, Korneliussen T, et al On detecting incomplete soft or hard selective sweeps using haplotype structure. Mol Biol Evol 2014;31:1275–1291.
[29]. Ayub Q, Moutsianas L, Chen Y, et al Revisiting the thrifty gene hypothesis via 65 loci associated with susceptibility to type 2 diabetes. Amer J Hum Genet 2014;94:176–185.
[30]. Steinthorsdottir V, Thorleifsson G, Sulem P, et al Identification of low-frequency and rare sequence variants associated with elevated or reduced risk of type 2 diabetes. Nat Genet 2014;46:294–298.
[31]. Minster RL, Hawley NL, Su CT, et al A thrifty variant in CREBRF strongly influences body mass index in Samoans. Nat Genet 2016;48:1049–1054.
[32]. Byars SG, Huang QQ, Gray LA, et al Genetic loci associated with coronary artery disease harbor evidence of selection and antagonistic pleiotropy. PLoS Genet 2017;13:e1006328.
[33]. Barban N, Jansen R, de Vlaming R, et al Genome-wide analysis identifies 12 loci influencing human reproductive behavior. Nat Genet 2016;48:1462–1472.
[34]. Brosens I, Pijnenborg R, Vercruysse L, et al The “Great Obstetrical Syndromes” are associated with disorders of deep placentation. Am J Obstet Gynecol 2011;204:193–201.
[35]. Hsu P, Nanan RKH. Innate and adaptive immune interactions at the fetal-maternal interface in healthy human pregnancy and pre-eclampsia. Front Immunol 2014;5:125.
[36]. Hiby SE, Apps R, Sharkey AM, et al Maternal activating KIRs protect against human reproductive failure mediated by fetal HLA-C2. J Clin Invest 2010;120:4102–4110.
[37]. Romero R. Prenatal medicine: the child is the father of the man. 1996. J Matern Fetal Neona 2009;22:636–639.
[38]. Chaiworapongsa T, Chaemsaithong P, Yeo L, et al Pre-eclampsia part 1: current understanding of its pathophysiology. Nat Rev Nephrol 2014;10:466–480.
[39]. von Dadelszen P, Ornstein MP, Bull SB, et al Fall in mean arterial pressure and fetal growth restriction in pregnancy hypertension: a meta-analysis. Lancet 2000;355:87–92.
[40]. McGinnis R, Steinthorsdottir V, Williams NO, et al Variants in the fetal genome near FLT1 are associated with risk of preeclampsia. Nat Genet 2017;49:1255–1260.
[41]. Leaños-Miranda A, Campos-Galicia I, Berumen-Lechuga MG, et al Circulating angiogenic factors and the risk of preeclampsia in systemic lupus erythematosus pregnancies. J Rheumatol 2015;42:1141–1149.
[42]. Leavey K, Benton SJ, Grynspan D, et al Unsupervised placental gene expression profiling identifies clinically relevant subclasses of human preeclampsia. Hypertension 2016;68:137–147.
[43]. Sarhanis P, Pugh DH. Resolution of pre-eclampsia following intrauterine death of one twin. Br J Obstet Gynaecol 1992;99:159–160.
[44]. Wang T, Xiang Y, Zhou X, et al Epigenome-wide association data implicate fetal/maternal adaptations contributing to clinical outcomes in preeclampsia. Epigenomics 2019;11:1003–1019.
[45]. Cavalli G, Heard E. Advances in epigenetics link genetics to the environment and disease. Nature 2019;571:489–499.
[46]. Heindel JJ, Vandenberg LN. Developmental origins of health and disease: a paradigm for understanding disease cause and prevention. Curr Opin Pediatr 2015;27:248–253.
[47]. Schübeler D. Function and information content of DNA methylation. Nature 2015;517:321–326.
[48]. Hagerman RJ, Berry-Kravis E, Hazlett HC, et al Fragile X syndrome. Nat Rev Dis Primers 2017;3:17065.
[49]. Colak D, Zaninovic N, Cohen MS, et al Promoter-bound trinucleotide repeat mRNA drives epigenetic silencing in fragile X syndrome. Science 2014;343:1002–1005.
[50]. Tatton-Brown K, Seal S, Ruark E, et al Mutations in the DNA methyltransferase gene DNMT3A cause an overgrowth syndrome with intellectual disability. Nat Genet 2014;46:385–388.
[51]. Ley TJ, Ding L, Walter MJ, et al DNMT3A mutations in acute myeloid leukemia. N Engl J Med 2010;363:2424–2433.
[52]. Tatton-Brown K, Rahman N. The NSD1 and EZH2 overgrowth genes, similarities and differences. Am J Med Genet C Semin Med Genet 2013;163C:86–91.
[53]. Chahrour M, Jung SY, Shaw C, et al MeCP2, a key contributor to neurological disease, activates and represses transcription. Science 2008;320:1224–1229.
[54]. Shah RR, Bird AP. MeCP2 mutations: progress towards understanding and treating Rett syndrome. Genome Med 2017;9:17.
[55]. Rasmussen KD, Helin K. Role of TET enzymes in DNA methylation, development, and cancer. Genes Dev 2016;30:733–750.
[56]. Muotri AR, Marchetto MCN, Coufal NG, et al L1 retrotransposition in neurons is modulated by MeCP2. Nature 2010;468:443–446.
[57]. Reid MA, Dai Z, Locasale JW. The impact of cellular metabolism on chromatin dynamics and epigenetics. Nat Cell Biol 2017;19:1298–1306.
[58]. Heijmans BT, Tobi EW, Stein AD, et al Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci U S A 2008;105:17046–17049.
[59]. Teschendorff AE, Relton CL. Statistical and integrative system-level analysis of DNA methylation data. Nat Rev Genet 2018;19:129–147.
[60]. Leenen FAD, Muller CP, Turner JD. DNA methylation: conducting the orchestra from exposure to phenotype? Clin Epigenetics 2016;8:92.
[61]. Liu Y, Aryee MJ, Padyukov L, et al Epigenome-wide association data implicate DNA methylation as an intermediary of genetic risk in rheumatoid arthritis. Nat Biotechnol 2013;31:142–147.
[62]. Wahl S, Drong A, Lehne B, et al Epigenome-wide association study of body mass index, and the adverse outcomes of adiposity. Nature 2017;541:81–86.
[63]. European IVFMC, European Society of Human R, Embryology, et al. Assisted reproductive technology in Europe, 2011: results generated from European registers by ESHRE. Hum Reprod. 2016;31:233–248.
[64]. Zegers-Hochschild F, Adamson GD, de Mouzon J, et al International Committee for Monitoring Assisted Reproductive Technology (ICMART) and the World Health Organization (WHO) revised glossary of ART terminology, 2009. Fertil Steril 2009;92:1520–1524.
[65]. McCartney CR, Marshall JC. Clinical practice. Polycystic ovary syndrome. N Engl J Med 2016;375:54–64.
[66]. de Groot PCM, Dekkers OM, Romijn JA, et al PCOS, coronary heart disease, stroke and the influence of obesity: a systematic review and meta-analysis. Hum Reprod Update 2011;17:495–500.
[67]. Azziz R. PCOS in 2015: New insights into the genetics of polycystic ovary syndrome. Nat Rev Endocrinol 2016;12:74–75.
[68]. Kong A, Frigge ML, Masson G, et al Rate of de novo mutations and the importance of father's age to disease risk. Nature 2012;488:471–475.
[69]. Inhorn MC, Patrizio P. Infertility around the globe: new thinking on gender, reproductive technologies and global movements in the 21st century. Hum Reprod Update 2015;21:411–426.
[70]. Xiong Y, Wang J, Liu L, et al Effects of high progesterone level on the day of human chorionic gonadotrophin administration in in vitro fertilization cycles on epigenetic modification of endometrium in the peri-implantation period. Fertil Steril 2017;108:269–276. e1.
[71]. Derks-Smeets IAP, van Tilborg TC, van Montfoort A, et al BRCA1 mutation carriers have a lower number of mature oocytes after ovarian stimulation for IVF/PGD. J Assist Reprod Genet 2017;34:1475–1482.
[72]. Magnusson Å, Källen K, Thurin-Kjellberg A, et al The number of oocytes retrieved during IVF: a balance between efficacy and safety. Hum Reprod 2018;33:58–64.
[73]. Xu GF, Zhang JY, Pan HT, et al Cardiovascular dysfunction in offspring of ovarian-hyperstimulated women and effects of estradiol and progesterone: a retrospective cohort study and proteomics analysis. J Clin Endocrinol Metab 2014;99:E2494–E2503.
[74]. Xu GF, Zhou CL, Xiong YM, et al Reduced intellectual ability in offspring of ovarian hyperstimulation syndrome: a cohort study. EBioMedicine 2017;20:263–267.
[75]. Feuer SK, Camarano L, Rinaudo PF. ART and health: clinical outcomes and insights on molecular mechanisms from rodent studies. Mol Human Reprod 2013;19:189–204.
[76]. Kang HJ, Feng Z, Sun Y, et al Single-nucleotide polymorphisms in the p53 pathway regulate fertility in humans. Proc Natl Acad Sci U S A 2009;106:9761–9766.
[77]. Hanevik HI, Hessen DO, Sunde A, et al Can IVF influence human evolution? Hum Reprod 2016;31:1397–1402.
[78]. Patrizio P, Sakkas D. From oocyte to baby: a clinical evaluation of the biological efficiency of in vitro fertilization. Fertil Steril 2009;91:1061–1066.

assisted reproduction; DOHaD; genetic pleiotropy; non-communicable diseases; positive selection

Copyright © 2020 The Chinese Medical Association. Published by Wolters Kluwer Health, Inc.