INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

HUMAN ADOLESCENT MOTHERS have an increased risk of preterm delivery of low birth weight infants who exhibit high mortality rates within the first year of life (20). We have developed a highly controlled experimental paradigm in adolescent pregnant sheep to examine the role of maternal nutrition and growth status on the etiology of poor pregnancy outcome (38, 40). Paradoxically, in this model system, overnourishing singleton-bearing adolescent sheep throughout pregnancy to promote rapid maternal growth results in a major reduction in placental mass and a significant decrease in lamb birth weight at term relative to slow-growing, moderate-intake, adolescent dams of equivalent gynecological age (38, 40). Thus in the well-fed pregnant adolescent ewe, growth of the maternal body is prioritized at the expense of the nutrient requirements of the gravid uterus. This occurs despite the ready availability of nutrients in the maternal circulation and is characterized by attenuated fetal plasma concentrations of insulin, insulin-like growth factor (IGF)-1, and glucose during late gestation (39, 41). Physiological processes that account for the diversion of nutrients to the mother in this model are unknown and form the basis of the present studies.

Our principal interest was the role of the placenta in providing nutrients to the fetus. The growth of the placenta precedes that of the fetus, and the strong correlation between placental mass and fetal size at birth in all species studied (1, 22, 25) suggests that impaired placental growth and nutrient transfer capacity are responsible for the slowing of fetal growth during late gestation in the overnourished adolescent sheep. Normally, during late pregnancy, the highly metabolically active placenta consumes one-half of the available oxygen and two-thirds of the available glucose taken up by the gravid uterus and is the major producer of lactate (23). The uteroplacental tissues also use one-quarter of the amino acid nitrogen taken up by the uterus (5). When placental growth is diminished, the rate of nutrient supply (largely determined by uterine and umbilical blood flows) and the competition between the placenta and the fetus for these nutrients may decrease. Thus, for example in the ovine carunclectomy model of surgically reduced placental mass, resulting fetal growth restriction is associated with lower uterine and umbilical blood flows (26) and a reduced supply of glucose to the gravid uterus and fetus (27). In contrast, in heat-stressed ewes that produce placental and fetal growth restriction, glucose supply to the fetus is limited by reduced placental glucose transfer capacity (34).

The objectives of the present study were to determine the rates of uterine and umbilical blood flow and net glucose and oxygen uptakes and consumption by the gravid uterus, fetus, and uteroplacental tissues in growth-restricted pregnancies induced by overnourishing adolescent sheep. Thus we aimed to establish whether fetal growth restriction in the rapidly growing adolescent ewe is due to a reduction in placental size per se or to more subtle alterations in uteroplacental nutrient uptake and/or metabolism.

	MATERIAL AND METHODS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Animals and experimental design. All procedures were approved by the United Kingdom Home Office under the (Animals Scientific Procedures) Act of 1986. Embryos from adult ewes (Border Leicester × Scottish Blackface), inseminated by a single sire, were recovered on day 4 after estrus and transferred synchronously in singleton into the uterus of 22 recipient ewe lambs (Dorset Horn × Mule) exactly as described previously (40). This technique removes the potentially confounding influence of partial embryo loss and variation in fetal number and maximizes the homogeneity of the resulting fetuses (38). Embryo transfer was carried out during the breeding season, and the animals were housed in individual pens under natural lighting conditions at the Rowett Research Institute (57° N, 2° W). At the time of embryo transfer, the recipient ewe lambs were peripubertal, ~7 mo old, with a mean live weight of 45 ± 0.9 kg. After embryo transfer, the recipient ewes were individually offered either a high (n = 12) or moderate (n = 10) level of the same complete diet calculated to promote rapid or low maternal growth rates. The moderate dietary level was in fact a control group in that this intake level was predicted to optimize fetal growth in this genotype. The diet supplied 10.2 MJ metabolizable energy and 137 g crude protein per kilogram and was offered in two equal feeds at 0800 and 1600 daily. The diet contained 30% (wt/vol) coarsely milled hay, 50% barley, 10% molasses, 9% fish meal, 0.3% salt, 0.5% dicalcium phosphate, and 0.2% of a vitamin-mineral supplement and had an average dry matter of 86%. Animals offered moderate intakes were offered their entire ration immediately, whereas those offered high intakes had the level of feed increased over a 2-wk period until the level of daily feed refusal was ~15% of the total offered (equivalent to ad libitum intakes). The level of feed offered was reviewed three times weekly and adjusted on an individual basis as appropriate on the basis of weight change data (recorded weekly) and the level of feed refused (recorded daily). After day 100 of gestation, the feed intake of the moderate intake group was adjusted weekly to meet the estimated increasing nutrient demands of the developing fetus during the final third of gestation by maintaining body condition score during this period. Ewes were body condition scored at 2-wk intervals throughout the study. Body condition score was assessed on a five-point scale (1 = emaciated, 5 = obese), as described previously (32).

Surgery and animal care. Infusion and sampling catheters were surgically inserted at ~122 days of gestation. Water and food were withheld overnight before surgery. Anesthesia was induced by intravenous administration of thiopentone (25 mg/kg, Intraval Sodium, Merial Animal Health, Essex, UK) and maintained by inhalation of halothane (Halothane-M&B, Rhone-Mereux, Essex, UK) in a mixture of oxygen and nitrous oxide. Just after anesthesia induction, ewes received antibiotics (iv) in the form of ampicillin (fixed dose of 500 mg, Penbritin, Beecham Research, Hertfordshire, UK) and gentamicin (5 mg/kg, Pangram 5%, Virbac, Cambridge, UK) and analgesics (iv) in the form of buprenorphine (0.006 mg/kg, Temgesic, Schering-Plough, Hertfordshire, UK) and carprofen (1.4 mg/kg, Rimadyl, Pfizer, Kent, UK). With the use of methods described previously (16), a polyvinyl catheter for infusion was placed into the fetal saphenous vein via a pedal vein. Catheters for sampling the uterine circulation were placed into a maternal femoral artery and the uterine vein draining the pregnant horn. Catheters for sampling the umbilical circulation were placed into the lower fetal aorta via a pedal artery and into the common umbilical vein. The catheters were tunneled subcutaneously, exteriorized through a flank skin incision, and kept in a pouch stitched to the ewe's flank. Catheters were flushed daily with a heparin-saline solution (150 IU/ml, 0.9% wt/vol sodium chloride solution). Ewes were housed individually in polypropylene floor-level crates and allowed to recover from surgery for 5-8 days before study. The ewes had previously been acclimatized to these crates for short periods on several days before catheter insertion. During the recovery period, maternal and fetal arterial blood glucose concentrations and blood gas concentrations were monitored daily. Ewes were transferred to a three times daily feeding regimen and gradually realimented back to either a high or moderate dietary intake.

Study design. Ewes were fed as normal at 0700 on the day of study. A solution of tritiated water (³H₂O, 16.7 µCi/ml, Amersham Life Science, Buckinghamshire, UK) in saline was infused into the fetal vein catheter. A bolus of 3 ml (50 µCi) was administered over 1 min, after which the infusion rate was changed to 3 ml (50 µCi) per hour. After infusion for at least 90 min to reach steady state, blood samples were withdrawn simultaneously into heparinized syringes from the maternal artery, uterine vein, fetal artery, and umbilical vein catheters. Four to six sets of consecutive samples were withdrawn into heparinized syringes at 15- or 20-min intervals and analyzed for plasma glucose, lactate, and ³H₂O concentrations. Fetal arterial and umbilical venous whole blood oxygen contents were measured in 10 high and 6 moderate intake pregnancies, respectively. At the end of the study, the ewe and fetus were rapidly euthanized by intravenous administration of an overdose of pentobarbital sodium (Euthesate; 200 mg pentobarbitone/ml; Willows Francis Veterinary, Crawley, UK). The fetus and placental cotyledons were weighed, and catheter location was verified. The biparietal head diameter, crown rump length, and girth at the umbilicus were measured, and the major fetal organs were dissected and weighed. Maternal and fetal perirenal fat was removed and weighed.

Chemical analyses. The plasma ³H₂O concentrations were measured in triplicate using 200 µl plasma, 500 µl distilled water, and 15 ml Ultima Gold scintillation fluid (Packard Bioscience, The Netherlands) and counted on a Packard scintillation counter (Tri-carb 1900TR with internal quench correction). Plasma ³H₂O concentrations were converted to whole blood concentrations according to Van Veen et al. (36). Plasma glucose and lactate concentrations were measured in duplicate with a Yellow Springs Instruments (YSI, Yellow Springs, OH) dual biochemistry analyzer (model 2700). The YSI instrument was calibrated with known standards after every fourth determination. Whole blood oxygen content was measured in 0.2 ml blood sampled into heparinized syringes using a Radiometer OSM-3 hemoximeter (Radiometer, Copenhagen, Denmark). Maternal and fetal arterial insulin concentrations were measured in duplicate by radioimmunoassay (19). The sensitivity of the assay was 4 µU/ml, and the intra-assay coefficient of variation was 7.8%.

Calculations and data analysis. Blood flow and net substrate uptake rates were calculated as described previously (23) and according to the following equations.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Maternal dietary intakes and maternal weight and condition score changes in relation to pregnancy outcome. Mean maternal feed intakes (kg dry matter/wk) were hugely elevated in high compared with moderate intake dams between days 4 and 49 (11.3 ± 0.2 vs. 5.1 ± 0, P < 0.001), days 50 and 98 (16.0 ± 0.5 vs. 4.6 ± 0.1, P < 0.001), and days 99 and 120 (16.0 ± 0.6 vs. 6.5 ± 0, P < 0.001) of gestation. On the day of study (~day 131 of gestation), the dry matter feed intake was 1,457 ± 68 and 1,141 ± 59 g in high and moderate groups, respectively (P < 0.01).

Blood flows. Uterine and umbilical blood flows at approximately day 130 of gestation are shown in Fig. 1. The growth-restricted pregnancies in the high intake dams were associated with lower uterine (P < 0.01) and umbilical (P < 0.001) blood flow rates compared with the moderate intake dams, reductions of 36 and 37%, respectively. Although uterine blood flow normalized for the weight of the conceptus (uterus + placenta + fetus) and umbilical blood flow per placental weight (data not shown) were not different between groups, umbilical blood flow per kilogram fetal weight remained significantly lower (P < 0.05) in high compared with moderate intake dams. The uterine blood flow to umbilical blood flow ratio was not perturbed in high vs. moderate intake groups (2.5 ± 0.21 and 2.3 ± 0.18, respectively). Placental weight correlated with fetal weight (r = 0.829, n = 22, P < 0.001). Irrespective of maternal nutritional treatment, uterine and umbilical blood flows were positively correlated with both placental (r = 0.823 and r = 0.842, respectively, n = 21, P < 0.001) and fetal weight (r = 0.781 and r = 0.784, respectively, n = 21, P < 0.001). Conversely, uterine and umbilical blood flows were negatively correlated with the fetal brain-to-body weight ratio (r = 0.729 and 0.728, respectively, P < 0.001).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Uterine and umbilical blood flows expressed absolutely and per conceptus and per fetal weight, respectively, on ~day 130 of gestation in adolescent dams offered a moderate (n = 10)- or high (n = 12)-nutrient intake throughout pregnancy. ns, Not significant. * P < 0.05; ** P < 0.01; *** P < 0.001.

Nutrient concentrations and flux rates. The maternal hematocrit was significantly elevated in high compared with moderate intake dams (0.29 ± 0.012 and 0.24 ± 0.009, respectively, P < 0.01).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

The present study confirms our previous observations that high dietary intakes resulting in rapid maternal growth rates throughout pregnancy in adolescent sheep result in significant placental and fetal growth restriction (38, 40). In addition, the marked reductions in placental and fetal weight in the overnourished dams in the present study were associated with a major decrease in both uterine and umbilical blood flows. Similar absolute reductions in uterine and umbilical blood flows during late gestation have been documented when placental growth restriction has been induced by heat stress (3, 34) or premating carunclectomy (26). However, when these flows were normalized for the weight of the conceptus and fetus, respectively, they were not different between restricted and control groups (Table 6). In contrast, in the present study umbilical blood flow in the high intake dams remained significantly lower when expressed per kilogram fetus, implying a relatively inadequate perfusion of the placenta by the fetus. This was unexpected but consistent with the observation that fetal weight was relatively less perturbed than placental weight in the high intake dams, resulting in a significantly higher fetal-to-placental weight ratio than in the moderate intake dams. Despite these observations, the uterine-to-umbilical blood flow ratio was equivalent in both groups, implying that the normal relationship between uteroplacental growth and blood flow had largely been maintained. The design of this initial study does not establish when during pregnancy the changes in blood flow arise and whether they drive or merely reflect the observed placental and fetal growth restriction. Intriguingly, chronic reduction of uterine blood flow (45% decrease) between days 113 and 138 of gestation, when proliferative growth of the placenta has ceased, produced a decrease in fetal body weight (32%; 18), which was strikingly similar to that observed in the present study (28%). Moreover, the chronic reduction in uterine blood flow in the former study was associated with a decrease in relative placental mass (34% decrease), which the authors attributed to a blood flow-mediated adaptive regression of the organ. A similar phenomenon may partially underlie the placental growth restriction observed in the overnourished adolescent dam. In noncatheterized adolescent dams at days 70, 100, and 128 of gestation, total placentome mass in high compared with moderate intake dams was reduced by 18, 20, and 51%, respectively (37), suggesting that the placentae of the high intake dams become progressively more perturbed as gestation proceeds. In normal pregnancies, blood flow to the uteroplacenta increases threefold between mid-gestation and term (24). However, as high intake adolescent dams continue to deposit adipose tissue during the final third of gestation, a possible explanation that should be explored is that blood flow to the maternal tissue beds continues to increase, whereas flow to the gravid uterus is progressively compromised. The role of maternal and/or placental hormones in coordinating these changes in blood flow are as yet unknown. We previously demonstrated that in overnourished compared with moderate intake adolescent dams, circulating insulin, IGF-1, leptin, and thyroid concentrations are high, whereas progesterone, growth hormone, and placental lactogen concentrations are low (42). These endocrine perturbations generally become significantly different between groups at around day 35 of gestation, and hence it is possible that one or more of these hormones plays a role in orchestrating placental vascularization and angiogenesis and, as such, affect uteroplacental blood flow in late pregnancy.

The modest but significant reduction in placentome number in the overnourished group is in agreement with our previous data obtained after spontaneous delivery of the placenta at term (42) and at first glance implies that overfeeding the adolescent sheep has a detrimental effect on maternal caruncle occupancy and hence placentome formation during the first third of gestation. However, in adolescent dams studied at ~day 80 of gestation there are no differences in placentome number in high compared with moderate dietary intake groups (99 ± 6.3 vs. 104 ± 4.8 placentomes, respectively, n = 14/group; J. M. Wallace, unpublished data). This suggests that a normal number of caruncles are originally occupied but fewer develop to their full size and transport potential. Thus it is possible that the reduction in blood flow observed during late gestation in the overnourished dams is sufficient to induce atrophy or regression of a proportion of the small placentomes during the second half of pregnancy. The relative role of the reduction in placentome number vs. the reduction in mean placentome size in mediating the reported blood flow and nutrient uptake effects cannot be determined in this study.

The present study indicates that this form of nutritionally mediated placental growth restriction is associated with marked fetal brain sparing and consequently an asymmetric pattern of growth similar to that observed in other models of fetal growth restriction (2, 30, 43). Furthermore, the significantly negative correlation between the brain-to-fetal weight ratio and both uterine and umbilical blood flows implies that this brain sparing is due to a redistribution of blood flow in the fetal circulation as has been demonstrated in a chronic preparation involving repetitive embolization of the uterine circulation (6). The relative placental and fetal growth restriction induced by overnourishing the adolescent dams throughout pregnancy in the present study is ~10-15% less than that achieved by the premating carunclectomy (13, 29) or experimentally induced heat stress models (1, 34). However, it is important to note that in our model, the growth-restricted lambs born to overnourished adolescent dams are largely viable at birth despite being 30% smaller than their normally grown counterparts at this time (42). In a previous late gestation study where fetal growth was reduced by 37% in high vs. moderate intake adolescent dams, maternal and fetal cortisol concentrations were not significantly influenced by maternal nutritional treatment (39), implying that perturbation of the glucocorticoid axis is not central to the degree or type of growth restriction observed.

The growth-restricted fetuses of the high intake dams were hypoxic and moderately hypoglycemic and hypoinsulinemic relative to the normally grown fetuses. This confirms our previous observation of low fetal glucose and insulin concentrations in cardiac blood samples obtained before autopsy at day 128 of gestation (39). As maternal glucose and oxygen concentrations on the day of study were equivalent in high and moderate intake dams, the reduction in fetal glucose and oxygen concentrations implies that the growth-restricted placenta of the high intake dams has a reduced ability to supply these metabolic substrates to the fetus. Indeed, the Fick principle calculations reveal that uterine glucose uptake was markedly reduced (46%) in high intake dams and remained lower when expressed per kilogram conceptus. This contrasts with the reduction in uterine blood flow of 36%, which was in proportion to the reduction in conceptus mass. Previously, in a study involving mechanically reducing uterine blood flow, it has been shown that a 50% reduction in flow is required before fetal glucose uptakes are perturbed (44). Thus, although the decrease in uterine blood flow in the high intake group clearly plays a part in reducing glucose availability in the present study, the data also suggest that there is a defect in uterine glucose uptake per se. Glucose uptake by the ovine uteroplacenta is mediated by the steady-state expression and activity of membrane-localized glucose transporters, namely GLUT-1 and GLUT-3 (9, 12). The relative concentrations of these transporters have not been quantified in the growth-restricted placentae in the present study. However, it is likely that the concentration of these transporters per unit membrane will be suppressed in response to the long-term maternal hyperglycemia, which is a consistent feature of high intake adolescent dams (40, 41). Support for this supposition comes from the observation that the induction of long-term hyperglycemia in late gestation adult ewes persistently downregulated GLUT-1 and GLUT-3 expression between 10 and 21 days of treatment (7, 8). In addition to reduced uterine glucose uptake, the hypoglycemia observed in the growth-restricted fetuses may also reflect reduced placental glucose transfer capacity. The use of glucose-clamp procedures to assess transport over a range of fetal glucose concentrations would help clarify this issue.

Uteroplacental glucose consumption is a major hindrance to glucose transport into the umbilical circulation (15). However, in the adolescent paradigm used in this study, there was no evidence that uteroplacental metabolism per unit placenta was radically altered at the expense of the fetus in that uteroplacental glucose and oxygen consumption and uteroplacental lactate production were decreased in proportion to the observed reduction in placental mass. Furthermore, uteroplacental oxygen and glucose clearance rates were similar between groups, indicating normal transplacental oxygen and glucose transfer capacities in the smaller than normal placenta of the high intake group. Similar proportionate reductions in uteroplacental glucose and oxygen consumption have been reported in growth-restricted pregnancies induced by heat stress (34) and imply a basic similarity in the physiological processes underlying uteroplacental growth restriction and metabolism in the two models. Indeed, both the heat-stressed and overnourished adolescent models have a similar placental phenotype in that placental growth restriction in both cases is characterized by a reduction in the number of placental cells rather than a change in cell size and a predominance of A-type placentomes where the maternal caruncle surrounds the fetal cotyledon (11, 21, 35, 37). In contrast, in the carunclectomy model, in the most severely growth-restricted pregnancies, uteroplacental glucose consumption per kilogram placenta was reduced, whereas lactate production per kilogram placenta was increased, resulting in a preferential redistribution of glucose (directly and in the form of lactate) to the fetus (27). In addition, carunclectomy is associated with increased fetus-to-maternal clearance rates of a nonmetabolizable glucose analog per unit placental mass (28) and increased glucose extraction across the uterine circulation (29). In combination, these metabolic aspects of placental function may reflect functional conditions unique to the marked placental hypertrophy and overgrowth of the fetal cotyledons, which remain after carunclectomy (4, 31), and are in direct contrast to the placental phenotype of the adolescent model. The trend (P < 0.06) toward lower umbilical glucose uptakes and the marked reduction (P < 0.01) in net fetal lactate uptake per kilogram body weight in the growth-restricted fetuses of the high-intake dams may partially reflect the significant reduction in umbilical blood flow per kilogram fetus. However, they may also be indicative of altered placental transport capacity and/or a compensatory mechanism by the fetus per se. It is acknowledged that growth-restricted fetuses may preferentially use amino acids for fetal metabolism when glucose supply becomes limiting (33), and studies of fetal amino acid uptake in the adolescent model would help resolve this issue.

An interesting observation in the intrauterine growth-restricted (IUGR) fetuses in the present study is that the rates of glucose and oxygen utilization are normal on a body weight-specific basis. This implies that overall fetal cellular metabolism is normal for these substrates, whereas fetal growth rate and its metabolic costs are diminished. It also implies that because fetal glucose and insulin concentrations are low and thus the ratio of glucose utilization rate to glucose and/or insulin concentrations are higher than normal, glucose utilization capacity is increased. Such a condition could be produced by increased glucose uptake and/or metabolism, increased insulin sensitivity, or both. Increased basal glucose uptake capacity could be due to upregulation of GLUT-1 transporter concentrations and/or activity. A small fraction of this increase also could be due to the relatively higher than normal brain-to-body weight ratio and thus a relatively larger obligatory consumption of glucose by the relatively larger brain. This explanation is less likely, because brain weight and its rate of glucose consumption represent relatively small fractions (15%) of total fetal glucose utilization (14, 17). Increased insulin sensitivity could result from hypoglycemia-induced increases in insulin signal transduction mechanisms, culminating in increased GLUT-4 activity at cell membranes. A redistribution of body organs would not be responsible for even part of this effect, as insulin-sensitive tissues such as skeletal muscle are diminished in proportion to body weight in the IUGR fetuses. The apparently increased glucose utilization capacity also is true of two other models of long-term (nearly full gestation) IUGR, the uterine carunclectomy and the maternal hyperthermia models (Table 6). In contrast, a hypoglycemic model of glucose insufficiency (produced by maternal insulin infusion leading to maternal hypoglycemia and decreased glucose transport to the fetus) does not show increased fetal glucose utilization capacity (10). In this model, glucose transporters decrease with time (7, 8). This is a model that is produced in late gestation, perhaps after the adaptations have occurred in the other models that allow increased glucose utilization capacity. One could conclude from this last model relative to the others, therefore, that hypoglycemia per se, which is a common finding in all of the models as well as human IUGR, is not the primary cause of the increase in glucose utilization capacity, which appears to represent a longer term adaptation to placental insufficiency that begins in early gestation.