School of Biological Sciences, University of Southampton, Southampton SO16 7PX, United Kingdom
The preimplantation mammalian embryo from different species appears sensitive to the environment in which it develops, either in vitro or in vivo, for example, in response to culture conditions or maternal diet. This sensitivity may lead to long-term alterations in the characteristics of fetal and/or postnatal growth and phenotype, which have implications for clinical health and biotechnological applications. We review the breadth of environmental influences that may affect early embryos and their responses to such conditions along epigenetic, metabolic, cellular, and physiological directions. In addition, we evaluate how embryo environmental responses may influence developmental potential and phenotype during later gestation. We conclude that a complex of different mechanisms may operate to associate early embryo environment with future health.
conceptus, early development, embryo, epigenetics, female reproductive tract, fetal and postnatal development, gene expression regulation, metabolism, pregnancy
Source: BIOLOGY OF REPRODUCTION 71, 1046–1054 (2004). © 2004 by the Society for the Study of Reproduction, Inc.
Several studies using animal models have shown that preimplantation embryos are sensitive to environmental conditions that can affect future growth and developmental potential, both pre- and postnatally. These studies initiated with the observation that cultured mouse embryos, after transfer, resulted in reduced fetal growth compared with that of in vivo counterparts [1, 2]. Further studies on the mouse have confirmed this effect and identified possible culture conditions, particularly inclusion of serum, which may act deleteriously [3–6; see also reviews 7, 8]. For example, embryos cultured in medium containing serum had a reduced viability and, after transfer, gave rise to fetuses at Day 14 that were 20% smaller than those derived from embryos cultured in the absence of serum or developing in vivo before transfer .
In addition to these rodent examples, early embryos from ruminants also display environmental sensitivity, which may, after transfer, lead to the condition known as large offspring syndrome (LOS). LOS has been identified in both sheep and cows and is characterized by aberrant fetal and placental development, increased fetal myogenesis, dystocia, dysfunctional perinatal pulmonary activity, organomegaly, and increased mortality in early postnatal life [9, 10]. Those animals that survive maintain features of abnormal organ sizing during later life . LOS occurrence is also associated particularly with the use of sera and complex media [12, 13]. In addition, cloned embryos derived from somatic nuclear transfer in both ruminants and mice experience similar abnormalities in fetal growth, reduced viability, and increased perinatal death [8, 14]. The human embryo must also be considered as potentially at risk to environmental conditions that may have long-term consequences [15–19]. Several, but not all, outcome studies show an increase in preterm delivery, low birth weight, and perinatal mortality in singleton pregnancies following assisted reproduction technologies (ART) compared with natural conception [18, 20–24]. However, the breadth of clinical and technical parameters that may complicate the interpretation of human studies needs to be borne in mind . Nevertheless, an association between low birth weight and ART is a cause for concern because susceptibility to adult-onset diseases such as coronary heart disease (CHD), stroke, hypertension, type II diabetes, and osteoporosis, while having adult life-style risk factors, correlate independently and more strongly with early life factors such as low weight and small size and thinness at birth [reviewed in 25, 26].
Abnormalities in developmental potential arising from embryonic environment are not limited to in vitro culture and may arise from the specific conditions experienced in vivo. Maternal diet can impact both on preimplantation phenotype and long-term development. High-protein diets in sheep during the periconceptional period have been associated with reduced developmental viability and increased fetal and birth weights as observed in LOS [27, 28]. A low-protein diet fed to rats exclusively during the preimplantation period before return to control diet for the remainder of gestation is associated with several changes in postnatal phenotype even though offspring were fed a normal diet. These effects included low birth weight, subsequent overcompensatory adolescent growth, onset of adult hypertension, and alterations in relative organ sizing in a gender-specific manner . Periconceptional maternal undernutrition in sheep has also been shown to affect later fetal development and physiology [30, 31] and can also influence the response to undernutrition in late gestation through its influence on fetal growth factor expression . Poor nutrition around the time of conception has further been reported to influence fetal growth and birth weight in the human . In a related context, maternal micronutrient intake periconceptionally, particularly of folic acid and B12 vitamin, which act as methyl donors important in DNA and protein modification, has long-lasting effects contributing to reduced incidence of neural tube defects later in development .
Collectively, these in vitro and in vivo studies lend support to the notion that an embryo's environment is critical for its future. The concept is a general one, evident across species, although species-specific differences arising in fetal growth rate are apparent. In the human, the relationship between smallness at birth and susceptibility to adult-onset disease largely derives from maternal undernutrition during pregnancy and intrauterine growth restriction. In global terms, therefore, the contribution that embryo environment in vivo may make to later fetal growth is an important issue to evaluate and the parallel analysis of in vivo and in vitro models is required to understand mechanisms. Moreover, the phenotypic consequences of early embryonic environment are potentially broad. Apart from growth, cardiovascular and metabolic conditions, low birth weight in humans may also be a significant risk factor for development of different neurological disorders, including psychopathologies associated with stress, such as depression and suicide  and the condition of autism . An important and well-controlled study has recently demonstrated that mouse embryo culture conditions resulted in an altered behavioral activity in offspring following embryo transfer .
What mechanisms relate preimplantation environment with long-term developmental potential? Is there a single underlying cause or are multiple interactive processes responsible? The links between early environment and programming of fetal or postnatal growth and phenotype are being actively pursued by several laboratories using different animal models. This review will evaluate our current understanding of underlying mechanisms (Fig. 1).
FIG. 1. Schematic representing the potential interactions between the environment of the embryo, either in vitro or in vivo, the embryo's short-term responses, and their long-term consequences. Different stages and lineages of embryo development are shown in different colors representing undifferentiated cells (pink), trophectoderm (yellow), ICM (pale blue), and primitive endoderm (deep blue)
There is growing and convincing evidence from analysis of different species that epigenetic events in the early embryo contribute to altered developmental potential [7, 38]. Imprinted genes, normally showing allele-specific expression due to epigenetic modification at regulatory CpG islands (differentially methylated regions, DMRs) mediated by the pattern of DNA methylation , appear particularly sensitive in this respect. Modification to histone proteins, including acetylation and methylation, occur dynamically and in concert with DNA methylation . The regulatory roles of imprinted genes encoding insulin-like growth factor 2 and its receptor (Igf2, Igf2R) in fetal growth [40, 41] and of Igf2, Mash-2, Ip1, Peg-3, and other genes in placental growth and nutrient supply [42–45] make these potentially sensitive targets for embryo epigenetic modification [38, 46].
Gametogenesis and early development are critical periods for the erasure, acquisition, and maintenance of genomic imprints. Although genome-wide demethylation of nonimprinted genes occurs consecutively on paternal and then maternal alleles by active and passive processes, respectively, during mouse preimplantation development, the gametic methylation pattern on imprinted genes needs to be preserved for future allele-specific expression . Preservation of imprinted gene methylation status during preimplantation development appears critically dependent on the oocyte DNA methyl transferase variant, DNMT1o [47, 48]. Thus, exclusion of DNMT1o from the nucleus during early cleavage may be responsible for global demethylation of the nonimprinted maternally derived genome .
The pattern of methylation of imprinted genes has been shown to alter in a gene-specific manner in response to embryo culture conditions, indicating effects of the environment on the maintenance of their methylation profile. Thus, mouse embryos cultured in Whitten medium show deregulation of the allele-specific methylation and expression profile of the H19 gene (regulator of IGF2), while embryos cultured in KSOM medium containing amino acids maintained the normal paternal methylated imprint at the regulatory domain and a maternal allele-specific expression profile; corresponding culture effects on the Snrpn-imprinted gene were not observed . There is growing evidence that culture-mediated epigenetic changes are heritable during cell cycling. Thus, Igf2 and H19 expression are reduced and H19 DMR hypermethylated in mouse fetuses that show reduced growth rate following embryo culture in medium containing serum compared with medium lacking serum . Conversely, Igf2R expression, but not that of Igf2, is significantly reduced in fetal sheep exhibiting LOS, thereby enhancing the bioavailability of Igf2 and likely contributing to the syndrome . Bovine fetuses derived from in vitro production also show higher expression of Igf2 in liver than do in vivo controls . Mouse extraembryonic tissues may also show disturbed H19 methylation following embryo culture . Moreover, male rat embryos derived from mothers fed a low-protein diet during preimplantation development exhibit reduced expression of H19 in the extraembryonic blastocyst trophectoderm lineage . Cloned mouse and bovine embryos also show abnormal epigenetic organization with incomplete reprogramming of imprinted gene methylation pattern in the donated somatic cell nucleus . Indeed, mice generated completely from embryonic stem (ES) cells show aberrant fetal growth and increased perinatal mortality associated with abnormal expression and methylation of Igf2 and Igf2R gene clusters in the original cultured ES cells, which is maintained during later development . Abnormal epigenetics and expression of imprinted genes are also associated with the human aberrant fetal growth and neurological syndromes such as Beckwith-Weidemann, Prader-Willi, and Angelman [19, 38].
The causal mechanisms relating embryo environmental sensitivity with abnormal fetal and postnatal development are unlikely to reside exclusively at the epigenetic level. We need to consider how embryos interact with their normal in vivo environment; how this interaction may alter under suboptimal in vitro or in vivo conditions; and what type of responses may occur in embryos to compensate, a measure of their developmental plasticity, essential for survival. Disturbance in developmental and/or nutrient cues across the maternal-embryonic interface may have long-lasting effects given the evolutionarily favorable concept that the setting of fetal growth trajectory may derive from a perception of maternal nutrient supply gained at an early stage of development. We also need to evaluate the contribution of egg quality to this assessment of plasticity and potential, an issue covered elsewhere in recent reviews [55– 57]. In this context, diets high in energy or protein applied in ruminants prior to conception reduce oocyte fertility and developmental potential due to exposure to elevated urea levels .
The ionic and nutrient composition of the oviduct and uterus experienced by cleavage-stage embryos has been reviewed recently [8, 59]. There is some evidence that energy substrate levels may vary between human oviduct and uterine fluids  and that nutrient levels are relatively low compared with those routinely present within culture media [61, 62]. However, data are scarce with regard to metabolite composition experienced by embryos in vivo at different stages of cleavage. In contrast, we know from extensive in vitro studies with human and different animal models that metabolic activity and substrate preferences appear to change between early and late cleavage with elevated glucose and oxygen consumption evident as embryos approach cavitation [63, 64; reviewed in 62]. However, nutrient consumption data vary dependent on in vitro culture composition and conditions, and the extent of exchange between the embryo and its environment is further complicated by the extent of endogenous stores [8, 62].
In vivo, other factors regulating nutrient exchange will derive from the maternal hormonal status responsive to dietary intake. Thus, progesterone level during the first 3 days of development in sheep has been shown to enhance subsequent fetal growth [65, 66]. Recently, ghrelin, the growth hormone secretogogue receptor (GHSR) ligand, implicated in modulating feeding behavior and energy metabolism, has been identified in mouse uterine fluid and endometrium as well as in morula and blastocyst stages . Embryos also express GHSR, and ghrelin is inhibitory to blastocyst development in vitro; uterine levels of ghrelin increase during fasting, suggesting it may modulate embryo metabolic demands in line with maternal nutrient availability . Although such environmental conditions will influence nutrient exchange kinetics, from the perspective of the embryo, it has been proposed that low metabolite consumption rates (quiet metabolism) more closely equate with developmental competence and viability than do high consumption rates (active metabolism), which are characteristic of metabolic stress [62, 68]. Embryo metabolic parameters may therefore be a significant contributor to the programming of future growth and physiological activity.
Hyperglycemia and Metabolic Stress
Hyperglycaemic conditions illustrate the association between embryo metabolism, stress, and future potential. Although glucose consumption rate does increase in late cleavage, oxidative phosphorylation and not glycolysis is the primary source of energy production in blastocysts across the species due to the corresponding elevation in oxygen consumption [62, 64]. In fact, elevated glucose metabolism, particularly via glycolysis, can be viewed as a stress response in embryos [15, 68], and mouse blastocysts showing high glycolytic rates have a lower capacity for implantation . Sheep embryos collected from ewes fed a high-protein diet during the periconceptional period exhibit a higher metabolic rate and consume more glucose, associated with reduced developmental potential, than do embryos from control diets .
While elevated glycolysis is an indicator of metabolic stress within embryos, the extent of glucose uptake itself has been shown to be a positive indicator of fetal developmental potential after transfer . However, in rodent diabetic models where a hyperglycemic embryo environment with significantly elevated glucose levels are utilized, embryo potential is reduced. Here, high external glucose acts to deplete glucose transporter expression at both mRNA and protein levels , leading to glucose starvation and activation of apoptotic pathways, chromatin degradation, and nuclear fragmentation , which culminate in reduced blastocyst cell numbers . Similar responses to high glucose have been identified in bovine embryos . These detrimental effects in embryos can also derive from perturbation in oocyte maturation in response to high glucose levels . Following embryo transfer, embryos derived from a diabetic environment show increased resorptions and pregnancy loss, illustrating the link between embryo stress response and future developmental capacity. Similar reduction in longer term potential has been shown to occur in response to inhibition in glucose transporter activity during mouse cleavage . A deficient inner cell mass (ICM) cell number has also been identified as a potential causative component of fetal growth retardation and large placenta, which are characteristic of the inherited BB/ E diabetic rat model .
The effect of high glucose concentration on the embryo in diabetic models may also disturb later morphogenesis due to a breakdown in normal inductive interactions between blastocyst ICM and trophectoderm lineages by downregulation of ICM Fgf-4 expression required for maintaining trophectoderm proliferative activity . Further support for the importance of embryo glucose/insulin balance has been demonstrated following short-term exposure of preimplantation mouse embryos to insulin, shown to stimulate ICM proliferation , which caused a long-term increase in fetal growth rate after transfer . Hyperglycemia is also implicated in the rat low-protein diet models. This diet fed to dams exclusively during the preimplantation period caused transient, mild maternal hyperglycemia associated with blastocyst cell number depletion and abnormal programming of postnatal growth . When the diet is applied 2 wk before mating and during postimplantation development, a transient upregulation of glycolysis is evident in isolated, intact rat conceptuses during 9.5–10.5 days of development coincident with organogenesis .
Reactive Oxygen Species
A further stress identified in embryos in response to culture conditions is an increase in hydrogen peroxide production and attendant risk from reactive oxygen species (ROS) [81, 82]. Oxidative stress in the bovine embryo has been shown to lead to DNA damage . ROS exposure will enhance the demand for antioxidant enzymes to maintain homeostatic control, which may further compromise developmental potential . ROS damage in bovine embryos is suppressed by inclusion of vitamin E in culture medium, which stimulated development both before and after embryo transfer .
The implications of impaired metabolic activity within embryos for fetal and postnatal development are potentially serious, but direct consequences are yet to be explored mechanistically. Elevated glucose levels can lead to suppression of insulin and glucokinase expression, decreased mitochondrial function, increased ROS formation, and accelerated apoptosis, as well as activation of common stress-activated signaling pathways, which could readily influence proliferative, metabolic, and neuroendocrinal axes during later development [reviewed in 86, 87]. For example, one major intracellular target of hyperglycemia is the transcription factor nuclear factor-kB (NF-kB), which in turn can regulate the expression of diverse growth factors, cytokines, and adhesion molecules, all of which have the potential to modulate the phenotypic response to early embryo environment .
Another and perhaps better example of adverse embryo programming working through metabolic pathways concerns amino acids and their turnover by the embryo. This is not surprising given the multitude of roles ascribed to amino acids in early development; in addition to protein biosynthesis, they also stimulate activation of the embryonic genome, blastocyst formation and hatching, and contribute to energy production, osmoregulation, pH control, cell homeostasis, and, perhaps significantly, to signal transduction cascades . The embryo is equipped with an array of sodium-dependent and -independent transporter systems to regulate amino acid flux and availability . Clear differences in embryo amino acid intracellular content have been demonstrated between in vivo and in vitro developing mouse embryos, indicative of environmental influence on exchange rates . Indeed, amino acid turnover has been shown to vary between individual human embryos assayed noninvasively during early cleavage. Significantly, the pattern of exchange correlated with capacity to form a blastocyst, with those embryos showing lower turnover exhibiting greater viability , supporting the concept of a quiet metabolism being linked with future potential [62, 68]. Culture of mouse embryos with amino acids has also been reported to enhance fetal development after transfer . In addition, culture of sheep embryos with amino acids leads to significantly improved fetal development than in medium containing serum . The correct balance of amino acids is important because amino acids, particularly glutamine, may spontaneously break down in culture to produce ammonium ions, which can be deleterious to development in the short- and long-term. Addition of ammonium to mouse culture produced a concentration-dependent reduction in blastocyst and ICM cell numbers, increased apoptosis, and altered the pattern of H19 gene expression; this treatment, after transfer, also reduced the implantation rate, impaired fetal development at Day 15, and caused a higher rate of exencephaly [92, 93].
The importance of amino acid environment during early development is further implicated from in vivo studies. Maternal low-protein diet administered to rats during the preimplantation period induced a transient reduction in maternal serum essential amino acid concentrations, associated with reductions in blastocyst ICM and trophectoderm (TE) cell numbers and leading to abnormal programming of growth and hypertension postnatally . Our current studies reveal, using the mouse model, that maternal protein undernutrition alters the profile of amino acids present within the uterine fluid during the period of morula and blastocyst occupancy, prior to implantation . These studies further indicate that specific amino acids, notably the branch-chain group of leucine, isoleucine, and valine, are significantly depleted in both serum and uterine fluid in response to low-protein diet [29, 94]. Significant uptake of these amino acids has been shown to occur during blastocyst formation and expansion in vitro, suggestive of a critical developmental role . Leucine in particular has a key signaling function in early development as activator of the mammalian target of rapamycin (mTOR), a serine-threonine kinase pathway that phosphorylates regulatory targets involved in protein translation and biosynthesis [96, 97]. The mTOR signaling has been associated with several growth and patterning events and phenotypic changes during development and differentiation [98, 99]. Significantly, in the early embryo, mTOR activity has been shown to be required for inducing an invasive and migratory behavior in trophoblast cells during the implantation period  and, through this activity dependent on amino acid transport, could coordinate downstream signaling pathways within the embryo .
In another mechanistic direction, changes in amino acid environment to embryos may lead to abnormal epigenetic effects. Maternal low-protein diet fed to rodents during the periconceptional period led to elevation in maternal serum homocysteine levels , which may cause folate deficiency and interfere with methyl group donation required for DNA methylation . The importance of maternal dietary methyl supplementation on epigenetic regulation and the extent of DNA methylation has been demonstrated in the Agouti mouse model [104, 105].
A further mechanism that may associate egg and embryo environment with future developmental potential is the number or pattern of calcium oscillations that take place upon oocyte meiotic maturation and activation and that control cortical granule exocytosis and cell-cycle resumption at fertilization . Altering the calcium oscillation number experimentally has also been shown to affect the relative number of ICM and TE cells in the blastocyst , the rate of apoptosis , and the extent of fetal development of parthenogenetic embryos [107, 109]. Whether these longer term effects are mediated through epigenetic pathways influencing gene expression or cell cycling remains to be determined.
The disturbance in embryo metabolism, proliferative capacity, signaling competence, and maternal developmental cues associated with a suboptimal environment, either in vitro or in vivo, will activate a condition of stress, which in turn will promote further responses within embryos to maintain homeostatic balance [15, 16]. Environmental influence on embryo gene expression is a sensitive measure of stress; increased expression of hsp70.1 and the growth-arrest gene CHOP-10 occurs in response to diverse stresses in rodent and bovine embryos [110, 111]. Expression of growth-regulating genes is also susceptible to the environment. Mouse embryos developing in culture express lower levels of the growth factors Igf-1 and -2 than do embryos in vivo [112, 113], while culture in KSOM medium with amino acids results in gene expression of Igf-1 and -2 and their receptors at a higher level that in embryos cultured in Whitten medium and more similar to in vivo embryos .
A range of metabolic and differentiation-related genes has been shown to alter their expression profile in response to culture conditions in bovine embryos [115, 116]. For example, genes critical in trophectoderm differentiation and intercellular junction formation are expressed at different levels dependent on in vivo or in vitro derivation and, for the latter, the type of medium used . In the case of the human, in vitro-cultured embryos show significant variation in expression of genes involved in apoptotic [118, 119] and differentiation [120, 121] pathways and show wide variation in apoptotic rate and blastocyst cell numbers , indicative of variable developmental potential.
Environmental factors commonly influence embryo proliferation (or apoptosis) and blastocyst ICM and trophectoderm cell numbers. This has been demonstrated in various species, including the human, following in vitro culture in response to growth factors, amino acids, glucose, and many other factors [15, 123, 124]. Similarly, in vivo embryo development in mice in response to maternal age , periconceptional protein undernutrition , or periconceptional zinc deficiency  can coincide with reduced blastocyst cell numbers.
What is the significance of depleted blastocyst cell numbers in the propagation of embryo environmental response into future development? At one level, reduced proliferation can be viewed merely as a symptom or marker of metabolic stress as discussed above. However, the stem cell pool for fetal and extraembryonic lineages is limited and, if depleted, may influence the pattern of later development. Past experimental studies using mostly the mouse have demonstrated some resilience and regulative capacity within embryos to reductions in cell numbers mediated by different manipulative procedures such as cell ablation, embryo bisection, or by mitomycin treatment to deplete the ICM [127–132]. These studies have revealed that, during postimplantation development, from about Day 12 or later, proliferative control mechanisms may compensate for early depletion of cells and result in reasonable rates of fetal survival. Nevertheless, artificial reduction in embryo cell numbers may significantly decrease the proportion of ICM derivatives , reduce the rate of primitive endoderm formation and the size of egg cylinder stage conceptuses up to Day 8 after transfer, delay the timing of gastrulation, retard morphogenesis, and significantly increase the rate of pregnancy loss [127, 129–132]. However, during postnatal development, mice derived from half embryos grow at similar rates to controls . It is also worth remembering that the early mouse embryo is equipped with self-regulating mechanisms to maintain ICM and TE cells and their ratio within relative narrow limits, indicating their importance to future development. Thus, while considerable variation may occur between embryos in allocation of cells to the ICM lineage during the 16-cell morula period, the second and final allocation at the 32-cell transition is inversely related to the first . This may compensate for interembryo variation and appears to be mediated by a combination of cell interactions, mitotic spindle orientation, and cell shape changes [134–136].
There is good evidence that blastocyst proliferative potential and relative lineage sizes influence future growth and viability. Quantitative measurements of the human ICM are highly indicative of blastocyst implantation potential with larger ICM size increasing potential . Moreover, a significant relationship between mouse blastocyst and ICM cell numbers, stimulated by amino acid availability during in vitro culture, and fetal viability after transfer has been reported . Reduced blastocyst cell numbers have also been implicated in the poorer developmental potential of in vitro-cultured bovine embryos compared with in vivo counterparts  that may lead to LOS. Similarly, reduced trophectoderm and ICM cell numbers in rat blastocysts following periconceptional maternal low-protein diet is associated with postnatal abnormal growth and hypertension . In this latter model, if low-protein diet is extended to Day 10 of development, the size of the conceptus and visceral yolk sac (VYS) are reduced compared with controls . The VYS is derived from the ICM and performs a critical role in fetal nutritive support through endocytosis of maternal proteins, their lysosomal degradation, and release as amino acids for biosynthetic activity within the fetus . Recently, we have investigated the effect of maternal low-protein diet on mouse VYS endocytic activity and found that maternal undernutrition reduces the rate of receptor-mediated protein (BSA) uptake and amino acid release per unit weight of yolk sac protein (unpublished results). This suggests that depleted ICM cell numbers can associate with altered physiological activity of derivative lineages, propagating, and indeed amplifying the effect of nutrient deprivation.
Embryos respond to their environment in vitro or in vivo in diverse ways that can influence their future growth and health. As discussed above, developmental plasticity, evident in epigenetic, metabolic, and proliferative states can lead to changes in fetal development through changes in imprinted gene expression, nutrient, and stress-related signaling pathways or cell cycling and apoptotic rates. Last, we consider the interaction between the embryo-centric phenotype changes and the maternal environment during subsequent gestation.
The search for basic mechanisms underlying the intrauterine programming of fetal and postnatal growth, cardiovascular, and metabolic physiology and ultimately the enhanced risk of adult-onset disease has indicated a complex network of interactions with a central role for maternal-fetal neuroendocrine signaling [reviewed in 140–142]. Maternal undernutrition during gestation alters maternal steroid hormone levels, including elevation of glucocorticoids (GC; corticosterone, cortisol), the stress hormones, which can profoundly influence the physiological condition of the conceptus . Fetal protection against maternal GC is enzymatically regulated by placental 11ß-hydroxysteroid dehydrogenase type 2 (11ß-HSD2), which converts GC to an inactive form . Placental 11ß-HSD2 expression and activity are reduced by maternal undernutrition, resulting in increased GC exposure to the fetus [145–147]. This exposure can alter the setting of the fetal hypothalamus-pituitary-adrenal (HPA) axis, leading to elevated fetal GC activity which, through the nuclear GC receptors, can modify the expression of many downstream genes controlling growth and metabolism, including cardiovascular and renal physiology [140–142]. Vascular function and setting of resting blood pressure can be further modified via the renin-angiotensin system and kidney nephrogenesis in response to HPA modulation [141, 148].
How important are these neuroendocrinal pathways in initiating fetal and postnatal alteration in growth, metabolic, and cardiovascular phenotype following environmental manipulation during preimplantation development? The effect of restricted periconceptional nutrition (70% control feed allowance) before mating and for 7 days after mating, prior to control feed provision for the duration of gestation, has been investigated on the development of the fetal HPA axis in sheep [30, 31]. Periconceptional undernutrition resulted in higher fetal plasma concentrations of pituitary adrenocorticotropic hormone (ACTH) between 110 and 145 days of gestation, a significantly greater cortisol response to corticotropin-releasing hormone and an increase in fetal blood pressure in twin fetuses in late gestation [30, 31]. A change in the setting of the fetal HPA axis following periconceptional undernutrition may derive from reduced trophectoderm development and a change in the secretion of placental hormones such as prostaglandin E2, which is thought to control ACTH secretion in late gestation [30, 149].
The consequences of early embryo environment on future development may also be mediated through disturbance in the balance of placental-fetal growth and function derived from i) nonequivalent epigenetic changes to imprinted genes expressed in fetal or placental pathways and/ or ii) inappropriate allocation of cells during early lineage specification [15, 44, 46]. The influence of placental function and placental/fetal exchange on fetal programming has been reviewed recently . Many imprinted genes contribute to placental function and nutrient exchange [38, 46] and may therefore amplify early epigenetic effects in embryos caused by environmental conditions with physiological impairment to growth due to reduced nutrient supply. Thus, deletion of the mouse Igf2 isoform Igf2P0, which is expressed exclusively in the labyrinthine trophoblast of the placenta, results in reduced placental growth and transport capacity and subsequently retardation in fetal growth . Abnormal allantoic development and defective placentation resulting from in vitro culture of ruminant embryos and leading to fetal pathology has been proposed as a mechanistic basis for LOS [16, 151, 152].
The embryo represents a developmental window during which susceptibility to environmental conditions is prevalent. The breadth of interacting responses within embryos to these conditions illustrates the complexity of the search for underlying mechanisms of developmental plasticity contributing to long-term changes in phenotype. From the main effects observed to date, we can surmise a multifactorial process whereby short-term epigenetic, metabolic, and proliferative conditions, possibly coupled with an altered maternal physiology, impose homeostatic changes in gene expression and setting of the neuroendocrine axis during later gestation. While the use of different animal models has contributed to our understanding of embryo sensitivity and response, we cannot assume equivalent sensitivity in the human. Nevertheless, caution in the handling and manipulation of human embryos and minimizing the duration of in vitro culture would likely reduce the potential of adverse consequences. The subject of developmental programming is still mainly descriptive in nature, and more use of genomic and proteomic tools will help identify critical gene and signaling pathways involved. The environmental sensitivity of egg polarity and establishment of developmental spatial axes  should also be investigated. There is a need for further experimental research with animal models, particularly employing intervention designs in vitro, to understand the temporal relationships of developmental plasticity.
Supported by grants from the Medical Research Centre, UK (G9800761, G0100558), to T.P.F. and by studentships funded by the School of Medicine and the School of Biological Sciences, University of Southampton, to R.P., E.U., and A.J.W.
Correspondence. FAX 44 2380 594319; firstname.lastname@example.org
Received: 19 April 2004.
First decision: 11 May 2004.
Accepted: 21 June 2004.