Dr Janna L Morrison, Heart Foundation Research Fellow, Early Origins of Adult Health Research Group, Sansom Institute, City East Campus, University of South Australia, GPO Box 2471, Adelaide, SA 5001, Australia. Email: Janna.Morrison@unisa.edu.au
1Intrauterine growth restriction (IUGR) has been associated with poor perinatal health outcomes. Animal models have been used to investigate why IUGR is associated with a poor prognosis. The sheep has been used extensively as an experimental model for IUGR with poor placental substrate supply to the fetus induced using a range of methods, including the surgical ablation of the majority of endometrial caruncles prior to conception, experimental induction of maternal hyperthermia, ligation of an umbilical artery or embolization of the placenta in late gestation and maternal overnutrition in the pregnant adolescent ewe.
2Fetal adaptations to fetal hypoxia and hypoglycaemia include activation of the fetal hypothalamic–pituitary–adrenal (HPA) axis and sympathetic nervous system and an associated increase in circulating cortisol and noradrenaline concentrations. Fetal cardiovascular responses vary according to the method used to induce placental dysfunction.
3Although an array of experimental models has been used to induce placental dysfunction at different stages of fetal development, each leads to remarkably similar fetal growth, metabolic, neuroendocrine and cardiovascular adaptations and consequences. The extent and range of the fetal physiological adaptations to chronic placental insufficiency are determined by the duration of exposure and the degree of the severity of substrate supply restriction.
4The present review summarizes how sheep models of IUGR have provided an increased understanding of the nature of the fetal adaptations to IUGR, their longer-term physiological consequences and how to improve clinical management of IUGR in human pregnancies.
Intrauterine growth restriction (IUGR) can be defined as a birth weight that falls in the 10th centile for any given gestational age (Table 1). In human fetuses, IUGR can be caused by factors that are of maternal, placental or fetal origin (Fig. 1), including living at high altitude, maternal undernutrition, hyperthermia, drug and alcohol abuse, placentation abnormalities, intrauterine infection and fetal chromosomal abnormalities.1–3 The most common cause of IUGR is placental insufficiency.4 A reduction in placental size or transport capacity leads to impaired substrate transfer from the mother to the fetus. Intrauterine growth restriction has been associated with a high incidence of perinatal morbidity and mortality because IUGR babies have a greater risk of hypoxic–ischaemic encephalopathy, intraventricular haemorrhage and necrotizing enterocolitis.5 In addition, these babies have long hospital stays5 and higher health care costs compared with infants who have appropriate growth for their gestational age.6,7 Research into fetal adaptations to poor substrate supply is aimed at improving health outcomes.
Table 1. Growth parameters in the growth-restricted fetus at 130–140 days gestation
Sheep models of IUGR
Data are presented as control versus intrauterine growth-restricted (IUGR) fetuses.
Epidemiological studies have shown that growth-restricted babies are at an increased risk of diseases in adult life, including hypertension,8 coronary artery disease,9 Type 2 diabetes,10 central obesity11 and the metabolic syndrome.3,11 The association between low birth weight and poor adult health outcomes has attracted significant attention from epidemiologists, clinicians and physiologists.12,13 Large-scale epidemiological studies investigating the association between low birth weight and adult health outcomes have been performed in the UK,14,15 Finland,16,17 Sweden,18,19 India,20 the US8,10 and Australia.21,22 However, the physiological mechanisms underlying these associations are poorly understood. Therefore, there is significant interest in understanding how the fetus adapts to a poor substrate supply in utero and the role that these adaptations play in adverse health consequences in adult life.
Analysis of cord blood samples at birth shows that IUGR fetuses are hypoxic, hypoglycaemic and have higher circulating plasma cortisol and noradrenaline concentrations compared with appropriately grown fetuses (Fig. 2; Tables 2, 3).23–25 In order to survive chronic periods of poor placental substrate supply, the fetus adapts by redistributing fetal cardiac output (Fig. 2; Table 4)5,26–28 and slowing body growth (Fig. 3). Changes in blood flow in the middle cerebral, adrenal, coronary, hepatic and mesenteric arteries26,27,29,30 result in sparing of brain, adrenal and heart growth and a decrease in fetal gut growth. Doppler ultrasound studies have shown that abnormal umbilical artery blood flow can predict poor perinatal outcomes5,31,32 and that there is a relationship between increasing placental vascular impedance and brain sparing in growth-restricted fetuses.31,33
Table 2. Fetal blood gas status in the growth-restricted fetus at 130–140 days gestation
Sheep models of IUGR
Data are presented as control versus intrauterine growth-restricted (IUGR) fetuses.
Birth weight is used as a measure of fetal growth in pregnancy, but it is also understood to be an imprecise measure because it represents fetal growth at one point in time at the end of gestation and does not provide an insight into the pattern of fetal growth throughout gestation.3,34,35 For example, IUGR may result from poor placental development early in pregnancy leading to a reduced substrate supply for optimal fetal growth in late gestation (chronic substrate restriction, Fig. 4). Alternatively, placental growth may have occurred normally in early pregnancy, but an insult in late gestation results in a severe but short-lived reduction in placental substrate delivery to the fetus leading to IUGR (acute substrate restriction). Similarly, there may have been several short but severe reductions in substrate supply to the fetus over late gestation (repeated acute substrate restriction). This raises the question of the definition of acute versus chronic substrate restriction. An acute event is usually defined as one that lasts minutes,36–38 but a chronic event may last hours,39 days40–42 or even weeks.43–48
There are many different experimental animal models of human IUGR involving studies on rats, mice, guinea-pigs and sheep. Most studies are performed in rodents because they have a short gestation period and are excellent for follow-up studies in which it is important to determine adult health outcomes after IUGR. Furthermore, molecular mechanisms can be interrogated relatively easily given the range of genetic information available about the mouse and rat. The disadvantage of these models is that because of the small size and relative immaturity of the fetus, they are not suitable for studies that focus on the functional response of the fetus to poor placental substrate supply. In the rodent, renal49 and cardiovascular50,51 development occurs throughout gestation and during the first weeks of life, whereas in the human these systems mature in prenatal life.52 In contrast with the rodent, in large animal models, such as the sheep, the functional responses of the fetus to short or long periods of placental insufficiency can be studied in utero. In addition, like the human, organogenesis occurs during early gestation and the renal and cardiovascular systems are functionally relatively mature in late gestation.53–55 Moreover, it has been demonstrated that the sheep fetus can also mount neuroendocrine and cardiovascular responses to intrauterine insults.56–58 Not only are the biological responses to IUGR similar in the sheep and human, but, importantly, the large size of the sheep fetus and its accessibility for the surgical implantation of vascular catheters and other instrumentation mean that it is possible to measure changes in blood gases, hormones and cardiovascular parameters in response to periods of induced placental dysfunction in late gestation. For these reasons, the present review focuses on experimental models of placental dysfunction leading to IUGR in the pregnant ewe and her fetus.
SHEEP MODELS OF IUGR
Placental dysfunction, resulting in the restriction of fetal substrate supply, is a major cause of altered or reduced fetal growth. It is well established in the sheep that variations in placental weight explain up to 80% of the variation in fetal weight from early in gestation.59–61 Several experimental approaches have been used in sheep to mimic human IUGR caused by a range of factors (see Fig. 1). These include removal of most of the endometrial caruncles from the non-pregnant uterus prior to conception, experimental induction of maternal hyperthermia, ligation of an umbilical artery or embolization of the placenta in late gestation and maternal overnutrition in the pregnant adolescent ewe. Each model results in altered placental transfer of oxygen and nutrients to the fetus and has specific strengths and weaknesses in relation to mimicking human IUGR. All these experimental models of IUGR result in chronic changes (i.e. those lasting more than 1 week) in placental development and function and are associated with a reduction in fetal growth. Each of these models contributes specific information to our understanding of the mechanisms that the fetus uses to survive chronic periods of decreased placental substrate supply. The fetal neuroendocrine and cardiovascular responses to the reduced substrate supply induced in each of these experimental situations are outlined below.
Carunclectomy involves the surgical removal of most of the endometrial caruncles from the uterus of the non-pregnant ewe prior to mating.2,46,47,62–64 This restricts the number of placentomes formed from the beginning of pregnancy, thereby limiting placental growth and function throughout pregnancy and leading to restriction of the placental delivery of oxygen and glucose to the fetus.2 Ewes are typically allowed to recover from carunclectomy surgery for a minimum of 10 weeks before they enter a mating programme and the pregnancy rate is reduced after this procedure.
Impact of placental insufficiency on fetal substrate supply and growth
Placentally restricted (PR) fetuses have reduced placental and fetal weights at term.65 In addition, PR fetuses are hypoxic but not markedly hypercapnic or acidotic.48 The fetal blood gas status and metabolic state of the PR sheep fetus is similar to that observed in percutaneous fetal blood sampling studies of small-for-gestational age infants.23 Hypoxia in the PR fetus is chronic and stable (see Fig. 5a) from the earliest time points in gestation that measurements have been made (> 102 days gestation).47,48,66 The PR fetus also shows evidence of a redistribution of cardiac output leading to asymmetrical growth restriction, evidenced by brain sparing (Fig. 6).2 There is a strong inverse relationship between fetal bodyweight and relative brain weight in both PR and control fetal sheep during late gestation, which indicates that brain weight is maintained within an optimal range across a wide range of fetal weights. Placentally restricted fetuses have reduced birth and placental weight at term.65
Absolute uterine and umbilical blood flow are reduced in the PR fetus compared with controls; however, when normalized to fetal weight, there is no difference at 122 days gestation.67 On a per kg fetal weight basis, there is also no difference in glucose utilization between PR and control fetuses.67
Fetal adaptations to placental restriction: Neuroendocrine
The carunclectomy model of IUGR leads to a range of fetal neuroendocrine adaptations, including decreased plasma insulin-like growth factor (IGF)-1 and IGF-2 concentrations,68 key regulators of fetal growth. Plasma IGF-1 and -2 concentrations are positively correlated with fetal hypoxia and hypoglycaemia.68 There is a significant relationship between fetal plasma IGF-1 and IGF-1 mRNA in the liver, kidney and skeletal muscle,69 organs and tissues in which growth is sacrificed if substrate supply is markedly reduced.2 Fetal plasma prolactin concentrations are reduced by approximately 50% in the PR fetus throughout late gestation70 and there is a 10% decrease in the expression of hepatic suppressor of cytokine signalling-3 (SOCS-3) mRNA levels in the fetal liver,71 suggesting that liver growth in response to reduced substrate supply may be regulated by prolactin or other cytokines and growth factors.
Although the growth of some organs is decreased, there is a relative sparing of adrenal growth in PR fetuses, with an increase in relative adrenal weight.72 Fetal plasma ACTH and cortisol concentrations are elevated in the PR fetus compared with control fetuses in late gestation,64 indicating activation of the hypothalamic–pituitary–adrenal (HPA) axis. This occurs, however, in the face of a decrease in adrenal IGF-2, adrenocorticotrophin (ACTH) receptor and CYP11A1 mRNA expression at 140 days gestation.73
Fetal adaptations to placental restriction: Cardiovascular
There is no difference in mean arterial blood pressure between normally grown and PR fetal sheep.47,48 There is, however, a direct relationship between basal blood pressure and mean arterial Po2 in control fetuses, which is not present in PR fetuses.48 Infusion of the angiotensin-converting enzyme inhibitor captopril after the onset of the prepartum increase in fetal cortisol concentrations from approximately 135 days gestation results in a greater hypotensive response in PR fetuses compared with normoxic fetuses.48 Although there is no difference in the expression of renal renin mRNA between PR and control fetal sheep, there is an increase in the capacity of the kidney to synthesize prostaglandin in the PR sheep fetus and this is directly related to renin mRNA expression.74 These changes may underlie the increased dependence on the renin–angiotensin system for maintenance of blood pressure in late gestation in the PR fetus compared with the control fetus.
Plasma noradrenaline concentrations are significantly higher in PR than control fetuses between 110 and 140 days gestation and, for every 1 mmHg decrease in arterial Po2, noradrenaline concentrations increase by 0.4 pmol/mL under basal conditions in both PR and control fetuses.66 Throughout late gestation, the mean arterial Po2 is approximately 8 mmHg lower in PR than control fetuses and this accounts for most of the difference in noradrenaline concentrations measured between PR and normally growing fetal sheep.66 Therefore, it appears that chronic hypoxia is a major factor that stimulates an increase in circulating noradrenaline concentrations in the PR fetuses in late gestation (Fig. 7).
Intrafetal infusion of tyramine, which acts to displace noradrenaline from catecholamine-containing vesicles within post-ganglionic sympathetic neurons, results in the same proportional increase in circulating noradrenaline concentrations in PR as in control normoxic fetuses, which suggests that one source of the circulating noradrenaline in the PR fetus may be post-ganglionic sympathetic neurons.66 A study in the chick embryo has shown that exposure of the embryo to chronic moderate hypoxia during late embryonic life leads to sympathetic hyperinnervation of the peripheral arteries.75 Therefore, one possibility is that the increase in circulating noradrenaline concentrations in the chronically hypoxic sheep fetus is a consequence of an increased density of nerve terminals in those circulatory regions that are important in ensuring that a redistribution of fetal cardiac output occurs away from the gut and periphery (Fig. 7).
Importantly, in a recent study in which we infused phentolamine,47 an α-adrenoceptor antagonist, in PR and control fetuses, we demonstrated that the maintenance of mean arterial pressure in the PR fetal sheep was dependent on α-adrenoceptor activation to a significantly greater extent than in control fetuses, that the hypotensive response to α-adrenoceptor blockade was present before the onset of the prepartum cortisol increase and that there was a direct relationship between the magnitude of the fetal hypotensive response and fetal arterial Po2 (Fig. 8). There was no difference in the mean arterial blood pressure response to an intrafetal bolus of the α-adrenoceptor agonist phenylephrine in PR and control fetal sheep, suggesting that there is no significantly enhanced sensitivity of the peripheral vasculature to noradrenaline in the PR fetus.47
These data suggest that despite the absence of differences in mean arterial blood pressure or heart rate in PR compared with control fetuses, arterial blood pressure is maintained to a greater extent by the renin–angiotensin system and sympathetic nervous system in the PR fetus compared with the normoxic, normally grown fetus. One possibility is that the physiological adaptations that act to maintain arterial blood pressure in the IUGR fetus persist after birth, or result in longer-term changes within the cardiovascular system, which explains the association between being born small and an increased risk of hypertension in adult life.
The PR fetus has a reduced placental mass resulting in chronic fetal hypoxia and hypoglycaemia across late gestation and IUGR. The PR fetus responds to the reduction in substrate supply by activation of the HPA axis and sympathetic nervous system. There is an interaction between the sympathetic nervous system and the renin–angiotensin system that results in maintenance of arterial blood pressure in the PR fetus.
Ambient temperature during pregnancy influences fetal growth. In tropical regions, IUGR babies are more likely to be born in the dry than wet season.76,77 High ambient temperature in the first trimester of pregnancy is associated with lower birth weight.78 The physiological mechanisms that may explain this relationship have been investigated in the sheep by exposing pregnant ewes to a high ambient temperature (hyperthermia). In this model, the ambient temperature is increased to 40°C for 12 h and then to 35°C for 12 h of each 24 h period, with relative humidity maintained at 30–40%. When ewes are exposed to hyperthermia in very early pregnancy, there is a high rate of fetal loss and, so in most studies, ewes are exposed to an elevated ambient temperature and moderate humidity after 45 days gestation for variable durations, but typically until 120 days gestation.79–81
Maternal adaptations to hyperthermia
Under the hyperthermia regimen described above, maternal core body temperature rises by 0.6–1.0°C within 7 days and this is associated with severe IUGR.82 In some, but not all, studies, hyperthermic ewes decrease their feed intake,83 so it is important that control ewes in these studies are pair fed to match the intake of their hyperthermic counterparts. By 120 days gestation, maternal core body temperature reaches 40°C, 1°C higher than controls, and maternal respiratory rate is markedly higher (161 breaths/min) compared with controls (37 breaths/min), although respiration returns to normal by 134 days gestation.79,81,84 The normal rise that occurs across gestation in maternal progesterone, prolactin and placental lactogen concentrations in control ewes is abolished in the presence of maternal hyperthermia.85 Despite this, there is no change in cotyledonary prolactin mRNA or protein expression.85
There is a reduction in both absolute uterine and umbilical blood flow in the hyperthermic fetuses.83 However, when normalized to fetal weight, there is no change in blood flow in IUGR fetuses compared with controls.80 The glucose transfer capacity of the placenta is reduced,83 but the fetal hypoglycaemia promotes glucose flux from mother to fetus.80 Interestingly, there is also a decrease in the expression of placental the glucose transporter GLUT-8, which is localized to the chorionic epithelial cells, in hyperthermic pregnancies at 135 days gestation that may account for the decline in glucose transport.86
Impact of placental insufficiency on fetal substrate supply and growth
Placental and fetal weight are reduced by approximately 50% in IUGR fetuses of hyperthermic ewes at 135 days gestation.80 The reduction in placental weight is not a consequence of a decrease in the number of placentomes, but rather the result of a marked reduction in placentome size.80 There is an increase in the fetal to placental weight ratio, with a relative sparing of fetal heart growth, and an increase in the ratio of brain to liver weight, which occurs from as early as 55 days gestation. By 90 days gestation, fetal weight is decreased.85,87–89 Fetal arterial Po2 is significantly lower in the IUGR fetus (11 mmHg) of the hyperthermic ewe compared with control fetuses (19 mmHg),79,81,83 with no significant difference in haematocrit, pH or Pco2. As expected, fetal plasma glucose and insulin concentrations are also significantly lower in the IUGR fetus of the hyperthermic ewe compared with controls.79,81,83 Together, these data show that maternal hyperthermia interferes with placental development, resulting in smaller placentomes with a reduced capacity for the transfer of oxygen and nutrients such as glucose to the fetus, resulting in IUGR.
At 90 days gestation, there is no difference in fetal umbilical vein IGF-1 concentrations between hyperthermic and control animals, but there is an increase in cotyledon IGF-2 and caruncle IGF binding protein (IGFBP)-4 expression in hyperthermic ewes as early as 55 days gestation.89 Together, these data suggest that placental development is altered from early in pregnancy when ewes are exposed to hyperthermia and that this results in reduced placental transfer of nutrients and the subsequent decrease in fetal growth.
In hyperthermic ewes, there is also a decline in umbilical oxygen uptake,81 but there is a concomitant increase in the transplacental oxygen diffusion gradient due to the fetal hypoxia.84 The expression of vascular endothelial growth factor (VEGF) in the placenta exhibits a biphasic profile, with increased expression of cotyledonary VEGF mRNA occurring at 55 days gestation in the hyperthermic group, followed by a decrease in expression by 135 days gestation.84 In addition, although VEGF mRNA expression does not differ between experimental and control animals at 90 days gestation, mRNA expression of VEGF receptor 1 and 2 is lower in the placental cotyledons of hyperthermic ewes at this age.88 These changes in the expression of a key placental vascular growth factor and its receptors in mid- and late gestation may reflect a compensatory response that ultimately contributes to the decrease in placentome size.
Fetal adaptations to maternal hyperthermia: Neuroendocrine
At 90 days gestation, there is no difference between fetal plasma IGF-1 concentrations in hyperthermic and control ewes, despite the presence of fetal growth restriction in the hyperthermic group.89 Interestingly IGF-1 protein levels are higher in the uterine caruncle of the hyperthermic ewe at 90 days gestation.89
Fetal plasma insulin concentrations are lower in IUGR fetuses than controls at 90 days gestation,89 as well as at 135 days gestation,90 and pancreatic β-cell mass and area are reduced in the IUGR fetus of the hyperthermic ewe.90 Short exposures to maternal hyperthermia do not result in changes in either maternal or fetal cortisol concentrations;91 however, when the period of maternal hyperthermia is sufficient to induce IUGR, there is a sex-dependent effect on fetal plasma cortisol concentrations where only male IUGR fetuses have elevated cortisol concentrations.82 Fetal plasma noradrenaline concentrations are also inversely associated with fetal arterial oxygen content in this IUGR model, as in the model of experimental restriction of placental growth discussed above.90
Fetal adaptations to maternal hyperthermia: Cardiovascular
Sparing of brain growth is observed as early as 55 days gestation in hyperthermic IUGR fetuses.88 Exposure to maternal hyperthermia results in a rise in umbilical vascular resistance relative to either placental or fetal weight.81 In contrast with the carunclectomy model, fetal arterial blood pressure is higher in IUGR fetuses compared with controls during late gestation (134 days gestation) and this may be a consequence of the increase in umbilical resistance.81 There is no change in heart rate in the IUGR fetus in this experimental model.92
Maternal hyperthermia in early pregnancy leads to reduced placental size and, thus, a reduced capacity for substrate transfer to the fetus. The fetal adaptations cause a slowing of fetal growth to match placental substrate supply. In male fetuses, there is an activation of the HPA axis. There is activation of the sympathetic nervous system, with increased arterial blood pressure and no change in heart rate.
SINGLE UMBILICAL ARTERY LIGATION
Single umbilical artery ligation (SUAL) is another technique for inducing IUGR in the sheep fetus. At 108–119 days gestation,93 one umbilical artery is isolated and ligated close to the fetal abdomen. This results in a partial infarction of the placenta, which may reduce the transfer capacity of the placenta. This model was developed in the late 1960s94 and has been used recently to induce IUGR.95,96 Interestingly, this is also a model of preterm labour, with a mean gestational age at onset of labour of 129 ± 3 days (range 119–141 days).95
Umbilical blood flow is reduced in SUAL pregnancies, resulting in reduced transfer of substrates from ewe to fetus.93 Relative to fetal weight, there is a decrease in umbilical blood flow with increasing gestational age.93 These results suggest that the fetal IUGR is a result of reduced umbilical blood flow and, thus, reduced fetal substrate supply.
Impact of placental insufficiency on fetal substrate supply and growth
On average, SUAL fetuses are 22% smaller than control fetuses95 and this growth restriction is observed as early as 117 days gestation.97 Interestingly, the SUAL fetuses are asymmetrically growth restricted, with relatively greater brain and adrenal weights than in control fetuses.95,97 The SUAL fetuses have a Po2 of approximately 16 mmHg from surgery at 105–110 days to 135 days, but are not acidotic.95
Fetal adaptations to ligation of an umbilical artery: Neuroendocrine
Circulating plasma cortisol concentrations are higher in SUAL fetuses at labour compared with control fetuses prior to postmortem.95 This increase may be related to the earlier onset of labour in these fetuses because the SUAL fetuses were delivered, on average, 6 days before the age when control fetuses underwent postmortem.
Fetal adaptations to ligation of an umbilical artery: Cardiovascular
There is no difference in basal fetal heart rate or mean arterial blood pressure between SUAL and control fetuses.95,97 This is similar to the carunclectomy model, where there is a decrease in placental size with no change in relative umbilical blood flow and no change in fetal arterial blood pressure. Interestingly, SUAL fetuses have significantly higher carotid artery blood flow than control fetuses97 and this may be the basis for the observed increase in relative brain weight.
Capacity of the IUGR fetus to adapt to additional insults after SUAL
In cases of threatened preterm delivery, mothers are treated with antenatal synthetic glucocorticoids to promote fetal respiratory development and improve neonatal outcomes.98 Studies in sheep suggest that synthetic glucocorticoids may reduce brain growth99,100 and, therefore, the question has been raised about the risks of this therapy in the IUGR fetus in which there are a series of physiological adaptations designed to maintain cerebral blood flow and brain growth.101 The SUAL model has been used to investigate the effects of administration of glucocorticoids on the IUGR fetus and the results suggest that the IUGR fetus may be at greater risk of brain injury after treatment with synthetic glucocorticoids than the normally grown fetus.97
Single umbilical artery ligation causes reduced placental blood flow and, thus, a reduction in substrate transfer from the ewe to fetus. This results in chronic hypoxia and IUGR. The fetal adaptations to this insult include early and pronounced activation of the HPA axis.
In order to mimic the late gestation onset of IUGR, embolization of the placental vasculature has been performed. This involves the repeated injection of microspheres (15 µm) into the placenta via the umbilical artery, with the source of administration being either through a catheter inserted in the descending aorta102,103 or the fetal umbilical vein.45,104–109 The microspheres block the capillaries in the placenta, reducing the surface area for the transfer of oxygen and nutrients from the ewe to the fetus, resulting in IUGR.
Impact of placental insufficiency on fetal substrate supply and growth
Placental embolization in late gestation results in low fluctuating fetal arterial Po2 levels,110,111 as shown in Fig. 5b. Repeated placental embolization causes a decrease in fetal oxygenation for periods of less than 24 h for up to 1 month. Placental embolization in late gestation results in fetal hypoxia,111–113 hypoglycaemia111 and IUGR.111 This sheep model of IUGR mimics several specific features observed in human IUGR, such as abnormal umbilical artery Doppler flow velocity waveforms.45 There is sparing of kidney growth, without any changes in the number of nephrons, glomerular capillary length or surface area in embolized compared with control fetuses.114
A recent study has shown that 4 days embolization results in altered placental morphology.115 Short-term embolization results in microspheres embedded in the fetal cytotrophoblast layer of the placenta. The cytotrophoblast layer appears normal, but there is a loss in the distinct separation between fetal and maternal cell layers. Embolization had no effect on fetal kidney morphology, but reduced the number of haematopoietic cell clusters in the liver.115
Fetal adaptations to uteroplacental embolization: Neuroendocrine
Placental embolization between 120 and 129 days gestation results in no change in fetal plasma ACTH or cortisol concentrations in late gestation, although cortisol is higher on the last 2 days of embolization.110 Despite this, expression of pro-opiomelanocortin (POMC) mRNA is decreased in the pars intermedia of the embolized fetus.110 Placental embolization between 120 and 135–8 days gestation results in increased plasma cortisol concentrations.111 As in the carunclectomy, maternal hyperthermia and SUAL models of IUGR, fetal plasma noradrenaline concentrations are higher in embolized fetuses compared with controls, but there is no change in fetal plasma adrenaline concentrations.113 Maternal tryptophan loading in these hypoxic, hypoglycaemic and IUGR fetuses results in increased kynurenine, a metabolite of tryptophan, in the fetal brain,111 reflecting compromised hepatic function. Kynurenine is further metabolised into kynurenic acid (KA), which may have a neuroprotective effect in the immature brain, and quinolinic acid (QA), which may play a role in excitotoxic brain damage. Kynurenic acid is decreased in the hippocampus and hypothalamus, whereas QA is increased in the medulla, pons, midbrain, hypothalamus and hippocampus of the embolized fetus.116 This may reflect an increased neural vulnerability to asphyxial injury in IUGR fetuses.
Fetal adaptations to uteroplacental embolization: Cardiovascular
In control fetuses, there is usually an increase in fetal blood pressure with increasing gestational age. In embolized fetuses, the increase in blood pressure with gestational age is augmented across the 21 day embolization protocol in late gestation.113 This is similar to changes in fetal arterial blood pressure after exposure of the ewe to hyperthermia and may be related to increased umbilical resistance, which occurs in both models. In contrast with other sheep models of IUGR, there is a 20–30 b.p.m. decrease in heart rate in embolized fetuses compared with control fetuses on Days 5, 7, 9, 15 and 17 of embolization.113 As a result, the rate–pressure product, an index of myocardial oxygen demand, is increased in the first half of the 21 day protocol,113 suggesting an increase in myocardial oxygen demand that returns to control levels. There is no change in rate–pressure product in control fetuses across the 21 day period of the experimental protocol.113 Importantly, placental embolization results in abnormal umbilical artery flow velocity indices, which indicate an increase in placental vascular resistance,106 as observed in human IUGR.5 In addition, these fetuses have an attenuated redistribution of cardiac output to the adrenal gland compared with controls, suggesting that they are less able to mount an appropriate cardiovascular response to an additional acute hypoxic response.108 Embolized fetuses that are exposed to an additional acute hypoxic insult have an equivalent rise in plasma cortisol, but an attenuated rise in plasma ACTH concentrations, compared with control animals.108 Changes in nitric oxide synthase (NOS) expression may be involved in hypoxic brain damage incurred in this model through excessive stimulation of neurons by increased neuronal NOS activity and compromised neuroprotective effects by reduced endothelial NOS activity.117
Repeated placental embolization results in repeated acute decreases in placental substrate supply leading to repeated fetal hypoxia and IUGR, with changes in umbilical artery blood flow occurring that are similar to those observed in the human IUGR fetus. The fetal adaptations include activation of the HPA axis and sympathetic nervous system. In fetal life, there is an increase in blood pressure.
OVERFEEDING OF THE ADOLESCENT EWE
Increased nutrient intake during pregnancy in adolescents results in increased maternal weight gain but decreased placental growth and, thus, IUGR.118 This clinical situation has been mimicked in a model of adolescent overfeeding in the sheep.119 This is accomplished by the use of a single sire and embryo transfer on Day 4 to an induced oestrous cycle in a pubertal adolescent ewe.120 The recipient ewes are then maintained on either a moderate or high amount of a complete diet until late gestation, when the moderate feed intake group is adjusted to continue the maintenance of healthy maternal weight and placental and fetal growth.
Both uterine and umbilical blood flow are reduced in the high maternal feed intake group; however, when normalized to fetal weight, there is no difference compared with that in the moderate maternal feed intake group.121
Impact of placental insufficiency on fetal substrate supply and growth
Overfeeding of adolescent ewes during pregnancy is associated with increased spontaneous abortion at 125 days gestation.120 Placental and fetal weight are decreased from as early as 95 days gestation and birth weight is decreased significantly.119 The decrease in placental weight is due to a decrease in both the number and weight of placentomes.119 In addition, the fetus is hypoxic121 and exhibits brain sparing.122 There is a decrease in umbilical uptake of both oxygen and glucose in the IUGR fetuses of high compared with moderate nutrient intake adolescent ewe fetuses.119 However, there is no difference in the glucose transfer capacity on a placental weight basis between the two groups.121 In addition, there is no difference in placentome GLUT-1 or GLUT-3 mRNA expression in IUGR fetuses at 81 or 133 days gestation.119 This finding is important because it suggests that the IUGR fetus is the result of a small placenta rather than altered placental function.
Fetal adaptations to overfeeding the adolescent ewe: Neuroendocrine
Fetal plasma IGF-1 concentrations are lower, albeit not significantly, in the high maternal feed intake compared with the moderate maternal feed intake fetuses at 52 days gestation.123
Adolescent overfeeding leads to decreased placental size, not placental function, which results in reduced fetal substrate supply and IUGR. In response to the reduced substrate supply, the fetus does not activate the HPA axis, contrary to observations in the other four sheep models of IUGR, possibly due to a more moderate degree of chronic fetal hypoxia.
SIMILARITIES AND DIFFERENCES IN FETAL ADAPTATIONS TO SUBSTRATE RESTRICTION IN SHEEP MODELS OF IUGR
Each of the five sheep models of IUGR discussed in the present review has specific advantages and disadvantages in relation to reproducing clinical scenarios that result in IUGR in the human. First, there are a series of technical or experimental issues associated with each model, separate from their biological relevance. The carunclectomy model requires the ability to return sheep to the paddock after the initial surgery for recovery. The maternal hyperthermia model requires a temperature- and humidity controlled chamber of sufficient size to house several sheep for an extended period. One limitation of the SUAL model is that all fetuses deliver early, usually between 119 and 141 days gestation,95 so if one is interested in investigating the effects of IUGR in late gestation or in the early postnatal period, this model may not be appropriate. Alternatively, this may be an optimal model for study of preterm delivery. The embolization model is similar to human IUGR in terms of umbilical vein blood flow; however, the embolization results in repeated placental damage. The adolescent maternal overfeeding model is important in the context of the adolescent pregnancy in humans, but will not reflect changes observed in women who are past their pubertal growth spurt.
A strength of these sheep models of IUGR is the ability to fully investigate the mechanistic basis of the fetal adaptations that occur when poor placentation leads to fetal hypoxia, hypoglycaemia and IUGR. In addition, these models may be useful in developing clinical tools for detecting and treating IUGR. For example, recently the SUAL model has been used to induce IUGR in order to determine whether maternal activin A, a glycoprotein in the transforming growth factor-β family that is elevated in the fetus and amniotic fluid in response to acute hypoxia, can be used as a marker of IUGR.95
INTO THE FUTURE
The data presented herein have been derived from five different sheep models of IUGR and, together, provide compelling evidence that there are common physiological factors and responses that play a key role in the fetal response to placental substrate restriction. Placental insufficiency has an impact on the growth and development of a range of organ systems, including the pancreas, HPA axis, renin–angiotensin system, sympathetic nervous system, lung, brain, heart and peripheral vasculature (see Tables 1–4). As more work defines the full, integrated physiological response of the fetus to a suboptimal intrauterine environment, it will become possible to determine whether these responses in fetal life initiate events that lead to the development of a cardiovascular disease, obesity and diabetes in later life. There are, however, a range of issues that require further investigation or investigation in more than one model. These include, but are not limited to: (i) early regulators of placental development and their role in determining poor placental growth and function; (ii) placental- and organ-specific expression and function of specific nutrient transporters (i.e. glucose and amino acids); (iii) a complete understanding of adrenal steroidogenesis across models; (iv) the nature of the oxygen-sensing mechanisms present in the placenta and IUGR fetus during chronic or repeated acute periods of hypoxia; and (v) the role of other neurotransmitters and hormones, such as nitric oxide and endothelin, in the fetal cardiovascular responses to chronic placental substrate restriction.
The aim of the present review was to summarize the available data collected with different sheep models of human IUGR that may contribute to the elucidation of the complex physiological responses of the IUGR fetus to an adverse environment. This is important because IUGR in the human fetus occurs as a consequence of a range of environmental, maternal, placental and fetal factors and is not an outcome of one common condition. An enhanced understanding of the common fetal adaptations to a range of causes of IUGR will enable the development of targeted interventions to support the IUGR fetus, both before and immediately after birth.
JLM was supported by a fellowship from the Heart Foundation (PF 03A 1283) and acknowledges research funding from the National Health and Medical Research Council of Australia. The author thanks Professors Caroline McMillen and Doug Brooks (Sansom Institute, University of South Australia) for helpful discussions.