Blastocyst environment and its influence on offspring cardiovascular health: the heart of the matter

Abstract

The development of adult-onset diseases such as type II diabetes, obesity and cardiovascular disease is traditionally attributed to adult lifestyle characteristics such as a lack of physical exercise, poor diet and smoking. However, evidence from both human and animal model studies has demonstrated that environmental factors such as an imbalance or reduction in maternal nutrition during gestation can have adverse effects on offspring metabolism and cardiovascular health. The severity and nature of the phenotypic changes induced in offspring is influenced by the period of gestation manipulated. In particular, the mammalian preimplantation embryo in different animal models displays particular sensitivity to environmental factors, either in vivo (maternal diet) or in vitro (embryo culture) that is associated with the onset of cardiovascular dysfunction in adult life. The detailed mechanisms by which environmental conditions can alter postnatal cardiovascular physiology are poorly understood. However, various factors including endothelial function, vascular responsiveness, the renin-angiotensin system, kidney structure and early postnatal growth dynamics have all been recognize as potential contributors. Here, we review the relationship between preimplantation embryo environment and postnatal cardiovascular disease risk, and consider biochemical, molecular, genetic and physiological pathways implicated in this association.

Introduction

Data from human and animal studies have demonstrated that a strong association exists between the onset of specific adult diseases and poor maternal nutrition during pregnancy. Initial observations made by Professor David Barker and colleagues in the late 1980s identified the relationship between low birth weight in humans and the risk of developing type II diabetes (Hales et al. 1991; Ravelli et al. 1998) and cardiovascular disease (Barker & Osmond, 1986; Barker et al. 1989). Since then, the use of animal models has further increased our understanding of the relationships between a deficient maternal environment and the long-term health of offspring. Typically, the nature of the animal studies employed have focused upon altering specific aspects of the maternal diet (e.g. reduction in protein content, global restriction of food and/or caloric intake) during critical windows of gestational (e.g. early development, periods of organogenesis, fetal growth in late gestation), and then examining the resultant health and physiology of the offspring.

In response to these findings from both human and animal studies, the Developmental Origins of Health and Disease (DOHaD) hypothesis has emerged, proposing that the fetus can adapt its own development in direct response to the uterine environment, mediated by factors such as maternal nutrition and physiology (Gluckman & Hanson, 2004; McMillen & Robinson, 2005; Hanson & Gluckman, 2008). These responses to maternal diet allow the fetus to make adjustments to its growth and metabolism so that fetal properties and size are compatible with the predicted availability of postnatal nutrients. However, under conditions of maternal undernutrition, the predictive responses may become maladaptive if the postnatal nutritional environment then provides a relative abundance of nutrients. In such circumstances, a discordance exists between nutrient intake and homeostatic processes regulating energy consumption, leading to an increased risk of adult cardiovascular and metabolic disease (Gluckman & Hanson, 2004; McMillen & Robinson, 2005; Hanson & Gluckman, 2008).

In recent years, a new focus has emerged concerning the sensitivity to environmental factors during preimplantation development and their effect on the long-term health of the offspring both in vivo (maternal diet) and in vitro (embryo culture) (Fleming et al. 2004; Sinclair & Singh, 2007; Thompson et al. 2007; Watkins et al. 2008a). During preimplantation development (between fertilization of the egg and implantation of the embryo into the uterus), the embryo forms a blastocyst in which are established the first two cell lineages; the outer polarize epithelial trophectoderm (TE) and the pluripotent inner cell mass (ICM), which give rise predominantly to placental and fetal tissues, respectively. Thus, during zygotic cleavage (approximately 4 days in the mouse and 6 days in the human), the role of the embryo has been classically viewed as initiation of the developmental programme comprising embryonic gene activation, cell cycling, lineage diversification and blastocyst morphogenesis including TE differentiation (Eckert & Fleming, 2008). However, with recent work, this concept has been revised such that the intrinsic programme appears responsive to extrinsic cues from the maternal environment to fine-tune the course of development, using a range of cellular, physiological and epigenetic mechanisms. These responses to the environment may facilitate changes in the development and behaviour of early lineages to compensate for nutrient deficits and/or to optimize phenotype for later development and the postnatal environment. For example, poor periconceptual nutrition can stimulate the extra-embryonic visceral yolk sac lineage to develop with enhanced endocytic capacity, thereby facilitating increased nutrient retrieval in later pregnancy (Watkins et al. 2008b). In a stable environment, the responses may be beneficial, acting in a compensatory way to protect fetal growth and maintain reproductive fitness and competitiveness. However, in line with the DOHaD concept, responses may become inappropriate and lead to adult disease if a mismatch exists between periconceptual and postnatal environments.

From a range of animal models and from different laboratories it has been demonstrated that manipulation of preimplantation environment through dietary or in vitro culture treatments can adversely affect postnatal phenotype (Fleming et al. 2004; Sinclair et al. 2007; Thompson et al. 2007; Watkins et al. 2008a). In particular, poor fetal or adult cardiovascular and metabolic health, and notably hypertension, constitute a common outcome phenotype from diverse preimplantation treatments (Kwong et al. 2000, 2007; Edwards & McMillen, 2002 a,b; Fernandez-Gonzalez et al. 2004, 2007; Gardner et al. 2004, 2006; Cleal et al. 2007; Sinclair et al. 2007; Watkins et al. 2007, 2008b). This is summarized in Fig. 1 and Table 1. This review focuses on the long-term changes in offspring cardiovascular phenotype following altered periconceptual and preimplantation embryo development, and the various pathways and physiological mechanisms underlying the hypertensive state induced.

Fig. 1.

Diagram representing the importance of environmental factors, either in vivo or in vitro, during the perioconceptional period on embryo development and adult cardiovascular phenotype, mediated potentially through diverse organ systems leading to hypertension and cardiovascular ill health. See Table 1 for details and references.

Table 1.

Effect of periconceptional environment on fetal and postnatal cardiovascular physiology

Model	Observed effects	Hypertension	Reference(s)
Rodent
Embryo culture	Increased serum ACE activity.	Yes	Watkins et al. (2007)
Maternal LPD	Altered nephron number.	Yes	Langley-Evans et al. (1996)
	Attenuated mesenteric responsiveness to isoprenaline and sodium nitroprusside.		Kwong et al. (2000, 2007)
	Elevated ACE activity.		Watkins et al. (2008b, unpublished observations)
Absencel of Ped gene expression	Elevated serum ACE activity.	Yes	Watkins et al. (2006)
Sheep/cattle
Maternal periconceptual undernutrition	Elevated fetal and postnatal arterial blood pressure.	Yes	Edwards et al. (2005)
	Increased activation of fetal hypothalamic-pituitary-adrenal (HPA) axis.		Edwards & McMillen (2002a)
	Increased postnatal plasma cortisol and altered cardiovascular function.		Gardner et al. (2004, 2006)
	Increased thoracic sensitivity to phenylephrine and acetylcholine.		Cleal et al. (2007)
	Increased heart wall thickness.		Sinclair et al. (2007)
Human
Assisted reproductive techniques (ART)	Elevated systolic and diastolic blood pressure and altered glucose homeostasis.	Yes	Ceelen et al. (2008)

Manipulation of maternal diet

Analysis of historical human data from women who became pregnant during the Dutch Hunger winter famine, a period of 5 months of malnutrition in the winter of 1944–45 in Amsterdam, the Netherlands, found that grown children of mothers who experienced the famine during embryonic and/or early fetal development had increased levels of coronary heart disease, an increased body mass index, and were glucose intolerant (Ravelli et al. 1999; Painter et al. 2006; De Rooij et al. 2006). Interestingly, children of mothers who became pregnant prior to the famine and experienced malnutrition during the later stages of pregnancy had reduced birth weights and became obese as adults (Ravelli et al. 1999; McMillen & Robinson, 2005). These data suggest that different periods of gestation appear differentially sensitive to the effects of maternal undernutrition with regard to the nature of the changes in offspring health and phenotype induced.

The employment of animal models has enabled the study of fetal and postnatal development in response to precise and controlled manipulation of the maternal diet to be conducted in greater detail. Along with dietary constituents, the use of rodent and sheep models permits the targeting of maternal undernutrition to discrete periods of embryonic and/or fetal development. Langley-Evans et al. (1996) demonstrate that different periods of gestation in the rat showed differential sensitivity with regard to the severity of the hypertension induced in weanling rats exposed to maternal low-protein diet (LPD). They observed that maternal LPD given at discrete 7-day intervals, or throughout gestation, resulted in hypertensive offspring. However, the magnitude of hypertension induced was greatest when LPD was given either during the last 7 days of gestation, or for the entirety of gestation. In a range of studies, LPD targeted exclusively to the preimplantation stage of rodent development has been shown to alter blastocyst cell number, increase the expression of 11β-hydroxysteroid dehydrogenase type 1 (Hsd11b1) and phosphoenolpyruvate carboxykinase (Pepck, Pck1), genes involved in activating glucocorticoid in the fetal liver and gluconeogenesis, respectively, which alter fetal and postnatal growth, induce relative hypertension in the adult offspring, increase patterns of anxiety-related behaviour in an open field test, and alter the relative sizes of organs (Kwong et al. 2000, 2006, 2007; Watkins et al. 2008b). In the sheep, maternal undernutrition from days 0 to 30/31 of pregnancy increases circulating plasma cortisol levels in female lambs and altered cardiovascular function of lambs at 1 year of age (Gardner et al. 2004, 2006), whilst in the adult, an increased interventricular septal wall thickness and increased mean left ventricular wall thickness were observed (Cleal et al. 2007). Maternal dietary deficiency in B vitamins and methionine during the periconceptional period in sheep also induced heavier and fatter offspring with associated hypertension with evidence of epigenetic changes associated with the DNA methylation profile of many genes (Sinclair et al. 2007).

Data from the human and animal studies have also revealed the significance of early postnatal growth dynamics upon adult health and physiology. In humans, children born small for gestational age and who remained small at 1 year of age but then underwent rapid gains in weight and body mass index (BMI), but not height, had a higher prevalence of coronary heart disease as adults when compared to adults of average birth weight (Eriksson et al. 1999, 2001). In mice, offspring from mothers fed LPD exclusively during preimplantation development (Emb-LPD) differ significantly in their associations between perinatal body weight and adult systolic blood pressure (SBP) from those of controls (Watkins et al. 2008b). Here, control offspring who were small at 3 weeks of age developed with a higher SBP as adults. However, Emb-LPD offspring showed positive associations, such that increased perinatal growth was associated with a higher adult SBP, highlighting the cardiovascular risk derived from excessive peri- and postnatal growth and body weight.

Not only do the very earliest stages of development appear sensitive to the programming effects of maternal undernutrition, but also dietary challenges during the period of oocyte maturation may result in changes in offspring cardiovascular health. In the mouse, maternal LPD given exclusively during oocyte maturation (3.5 days prior to mating) has been shown to result in the development of hypertensive adult offspring (Watkins et al. 2008c). In the sheep, maternal undernutrition during the periconceptual period (45 days prior to conception until 7 days after) alters maternal and placental/fetal growth dynamics, increases fetal arterial blood pressure, and increases the activation of the fetal hypothalamo-pituitary-adrenal axis (Edwards & McMillen, 2002a,b; Edwards et al. 2005, McMillen et al. 2008).

In vitro culture

In vitro culture and manipulation of mammalian preimplantation embryos is a necessary requirement for diverse procedures ranging from human assisted reproductive techniques (ART) such as IVF, the derivation of embryonic stem cell lines, biotechnological treatments to improve production of important domestic animal strains, and ultimately increasing our understanding of essential developmental pathways and mechanisms. However, whilst the development of morphologically ‘normal’ blastocysts may be achieved in vitro with relatively high success rates, the short- and long-term developmental capacity of those tissues, and ultimately the offspring derived from them, may be compromised. Perhaps the most dramatic display of how fetal and postnatal development can be altered in response to manipulating the preimplantation embryo is the phenomenon of ‘large offspring syndrome’ (LOS). LOS has been observed in cattle and sheep, and occurs as a consequence of in vitro culture in the presence of serum, the introduction of the preimplantation embryo to an altered uterine environment, or the production of cloned embryos following nuclear transfer (Holm et al. 1996; Sinclair et al. 1999, 2000; Wells et al. 1999). LOS is associated with significantly increased fetal and postnatal growth, altered organ sizing, changes in patterns of gene expression, and increased rates of perinatal death (Holm et al. 1996; Sinclair et al. 1999; Niemann & Wrenzycki, 2000; Farin et al. 2006, Smith et al. 2008). It has also been reported that mice produced from cloned embryos also display an increased rate of fetal and postnatal growth (Eggan et al. 2001).

In vitro culture and transfer of mouse preimplantation embryos has also been associated with significant changes to offspring phenotype. We reported elevation of offspring SBP following either prolonged (two-cell to blastocyst stage) or brief (2 h prior to uterine transfer) in vitro culture and embryo transfer (Watkins et al. 2007). Changes in embryo gene expression and regulation (Stojanov & O’Neill, 2001; Rinaudo & Schultz, 2004), reduced rates of development and survival post transfer (Bowman & McLaren, 1970; Li et al. 2007), altered fetal growth (Caro & Trounson, 1984; Arny et al. 1987; Khosla et al. 2001) and disturbed postnatal memory characteristics (Ecker et al. 2004; Fernandez-Gonzalez et al. 2004) have also been documented following in vitro culture of preimplantation embryos. Recently, culture and transfer of mouse preimplantation embryos was shown to result in altered postnatal growth and organ sizing in the offspring which persisted into a second generation, indicating epigenetic modifications (Mahsoudi et al., 2007).

The well-documented sensitivity of rodent and large animal embryos to differing in vitro manipulations raises several questions about the possible long-term development and cardiovascular health of children conceived using ART. Follow-up studies have demonstrated that ART children have a higher risk of being born smaller, even after accounting for the incidences of multiple pregnancies (Schieve et al. 2002). There are also several reports indicating increased rates of congenital abnormalities for infants conceived by intracytoplasmic sperm injection (ICSI) (Hansen et al. 2002; Winston & Hardy, 2002) and increased incidence of syndromes caused by errors in gene imprinting (Cox et al. 2002; De Rycke et al. 2002; DeBaun et al. 2003; Lucifero et al. 2004). Conversely, others report no differences in the psychological status of school-age children conceived by IVF as compared with children conceived naturally (Montgomery et al. 1999). At present it is difficult to gauge the detailed long-term impact of ART treatments primarily due to the relatively young age of the children born. Whilst there have been several systematic follow-up studies examining physical and cognitive development and rates of diseases such as cancer occurring in ART children (Schieve et al. 2002; Koivurova et al. 2003; Ludwig et al. 2006), there has only been one study examining postnatal blood pressure and metabolic profile. This study reported that pre-pubertal IVF children displayed significantly elevated systolic and diastolic blood pressure, and elevated fasting glucose levels independent of early life factors or parental characteristics (Ceelen et al. 2008). It will be important to extend such cardiovascular studies on ART children as they reach adulthood, particularly as the tracking of childhood blood pressure into adulthood is known to occur (Law et al. 1993).

Mechanisms central to the hypertensive state

The mechanisms and causal pathways that underlie the hypertensive state and associated cardiovascular diseases in response to developmental programming are likely to be complex and numerous and have been described in detail elsewhere (Langley-Evans, 2001; Brawley et al. 2003a, McMillen & Robinson 2005). Therefore, we highlight here the central components demonstrated to be affected in response to periconceptual environment.

Endothelial dysfunction and vascular responsiveness

As peripheral vascular dysfunction is associated with, and often precedes, the development of hypertension (Brawley et al. 2003a), an understanding of the role of vascular responsiveness and endothelial and smooth muscle function in the homeostatic regulation of blood pressure is essential. Offspring of diet-restricted mothers have consistently demonstrated elevations in their SBP (Langley-Evans et al. 1996, Kwong et al. 2000, Torrens et al. 2003, 2006; Watkins et al. 2008c) and, as a result, numerous studies have examined the production of and responsiveness to a range of vasoactive compounds. Isolated mesenteric arteries from male mice exposed to maternal LPD either exclusively during preimplantation development or throughout gestation, display attenuated responsiveness to the β-adrenoceptor agonist, isoprenaline, and the nitric oxide (NO) donor, sodium nitroprusside (Watkins et al. 2008, unpublished observations), suggestive of a vascular defect in smooth muscle signalling via cAMP. Maternal LPD given exclusively during mouse oocyte maturation results in impaired responsiveness to the endothelial-dependent vasodilator acetylcholine, and isoprenaline, in male isolated mesenteric arteries (Watkins et al. 2008c). It is of interest to note that no difference in vasoconstriction mediated through the α₁-adrenergic vasoagonist, phenylephrine, was observed in these two studies. These findings are in line with previous studies of offspring from dams fed a globally restricted diet (50% reduction) during the second half of pregnancy (Holemans et al. 1999), or LPD throughout pregnancy (Brawley et al. 2003b), where no difference in α₁-adrenoceptor-mediated constriction was observed.

The renin-angiotensin system (RAS)

Along with vascular responsiveness, the renin-angiotensin system is also sensitive to the programming effects of an altered embryo environment. Angiotensin-converting-enzyme (ACE), a metallopeptidase, has been identified as a central component of the RAS through its predominant actions of cleaving bradykinin, and the generation of angiotensin II from angiotensin I (Corvol et al. 1995; Coates, 2003). Upon binding to its receptor (AT₁), located in vascular smooth muscle cells, kidney, heart, adrenal glands and the brain, angiotensin II induces vasoconstriction, renal tubular sodium re-absorption, and aldosterone and vasopressin secretion (Unger, 2002). Angiotensin II has also been implicated in endothelial dysfunction through its ability to increase oxidative stress (Mehta & Griendling, 2007), whilst the cleavage of the vasodilator bradykinin also contributes to the establishment of a hypertensive state (Corvol et al. 1995; Coates, 2003). In mice, ACE^−/− animals are hypotensive with an SBP ~30 mmHg lower than control animals (Esther et al. 1997; Cole et al. 2000), whilst in humans, serum ACE levels are significantly higher in hypertensive subjects (Forrester et al. 1997). The RAS and ACE function have been demonstrated to be involved in the programming of hypertension following maternal dietary manipulation in rodents and sheep. Elevated activity of serum and lung ACE has been observed in mice from mothers fed LPD either exclusively during preimplantation development or throughout gestation (Watkins et al. 2008, unpbublished observations). Elevated ACE activity also occurs in mice in response to in vitro culture and embryo transfer procedures (Watkins et al. 2007) and in vivo in response to expression of the ‘Ped’ genes that contribute to embryo survival (Watkins et al. 2006). The blockade of the AT₁ receptor with the agonist losartan (Sherman & Langley-Evans, 2000), or the use of the ACE inhibitor captopril (Langley-Evans & Jackson, 1995), significantly reduces the SBP in offspring rats from LPD-fed mothers. In the sheep, significantly elevated blood pressure responses to RAS stimulation by ferusamide is observed in offspring of mothers with early gestational nutrient restriction (Cleal et al. 2007). This elevation in SBP was blocked by prior administration of captopril, further indicating the central role of the RAS in the hypertensive state.

Kidney development and function

The notion that a reduced birth weight and associated reduction in nephron number could increase the risk of adult hypertension and chronic renal disease was first proposed by Brenner & Chertow (1994). Autopsy studies from intrauterine growth-retarded (IUGR) infants have demonstrated associations with a reduction in nephron number (Hinchcliffe et al. 1991, 1992), whilst in a range of animal species, experimental maternal undernutrition or placental insufficiency have now been reproducibly associated with reduced nephron number (Jones et al. 2001; Woods et al. 2001; Bauer et al. 2002; Gilbert et al. 2005), which is then associated with elevated SBP (Woods et al. 2001; Gilbert et al. 2005). However, several studies have demonstrated programmed elevated blood pressure effects in the absence of significant changes in glomerular number (Ortiz et al. 2003, Pladys et al. 2005; Woods et al. 2005). Whilst we have found no change in glomerular number in response to maternal LPD given exclusively during the mouse preimplantation period (Watkins et al. 2008, unpublished observations), targeting LPD to the period of oocyte maturation induced both reduced kidney size and an increase in glomerula number coincident with relative hypertension (Watkins et al. 2008c). As kidney development occurs late in gestation and this periconceptional dietary challenge actually precedes the onset of development, the response of increased glomerula numbers may be compensatory to guard against adverse programming, although the mechanism of this response is currently unknown.

Conclusions

Evidence is growing from many laboratories that the periconceptual period is a window of enhanced sensitivity to environmental conditions both in vivo and in vitro. This sensitivity reflects the emerging concept that the early embryo, before implantation, responds to maternal cues important in setting homeostatic regulation of metabolic, physiologic and growth criteria for gestation. We have found that the cardiovascular system is particularly vulnerable to adverse programming during the periconceptual window (Fig. 1). Such compensatory responses by embryos identified in animal models may also occur in the human and underlie epidemiological and related studies concerning the developmental origin of adult chronic disease. Further research is clearly needed to identify and understand the range of mechanisms contributing to the link between embryo environment and adult health.