Prenatal and early postnatal life experiences, reflected by size at birth and postnatal catch-up growth, contribute to the risk of developing the metabolic syndrome in adulthood, but their relative importance is unclear. Therefore, we determined the effects of restricted placental and fetal growth on components of the metabolic syndrome in young adult sheep and the relationships of the latter to size at birth and early postnatal growth. Fasting plasma metabolites, glucose tolerance (by intravenous glucose tolerance test, IVGTT), insulin secretion and sensitivity, and resting blood pressure were measured in 22 control and 20 placentally restricted (PR) 1-yr-old sheep. In male sheep, PR increased the initial rise in glucose during an IVGTT and reduced diastolic blood pressure, and small size at birth independently predicted reduced adult size, glucose tolerance, and fasting plasma insulin and insulin disposition of glucose metabolism but increased insulin disposition of circulating FFAs. Also in males, high fractional growth rates in early postnatal life independently predicted impaired early glucose clearance during an IVGTT. In female animals, PR increased insulin sensitivity of glucose metabolism and reduced fasting plasma FFAs, and thinness at birth predicted increased adult size, fasting blood glucose, and pulse pressure. In conclusion, PR and small size at birth are associated with more components of the metabolic syndrome in adult male than in adult female sheep, with few independent effects of early postnatal growth. These sex differences in the onset and extent of adverse metabolic consequences after prenatal restraint in the sheep are consistent with observations in humans.
- size at birth
- insulin action
- blood pressure
- glucose metabolism
small size at birth and its associated infant catch-up growth are implicated as leading to the metabolic syndrome in adult men (2, 28). This syndrome, also known as the insulin resistance syndrome or syndrome X, comprises a cluster of risk factors (markers of impaired glucose homeostasis or insulin resistance, central obesity, hypertension, elevated fasting triglycerides, and low/high-density lipoprotein cholesterol), which contribute markedly to cardiovascular disease risk (31, 42). The incidence of the metabolic syndrome in 64-yr-old men increases with decreasing birth weight (2), and insulin resistance, the central defect of the syndrome, is consistently worse in men and women who were light at birth (20). Most human intrauterine growth-restricted (IUGR) infants also undergo accelerated, or catch-up, growth in early infancy, although how this relates to later outcomes has been less studied (1, 9, 14, 40). Early life influences, measured by indicators including size at birth and childhood catch-up growth, accounted for a greater proportion of risk of the metabolic syndrome than influences attributed to adult socioeconomic status and lifestyle in a cohort of middle-aged men (28). Low birth weight and catch-up growth in the first 2–3 yr of life can independently predict the incidence of individual components of the metabolic syndrome in studies in which various outcomes were assessed separately, including insulin resistance (22), impaired glucose tolerance (4), diabetes (11), increased blood pressure (15), and obesity (21–23). However, the extent to which restricted fetal growth, in particular that arising from a common complication of pregnancy (placental restriction), and the manner in which its associated accelerated postnatal growth separately predict the whole metabolic syndrome has not been more directly investigated.
Placental insufficiency, due to poor placental growth and/or function, is a common cause of restricted fetal growth in humans (3). Experimental placental restriction (PR) in late pregnancy in the rat induces a number of elements of the metabolic syndrome in offspring, causing insulin resistance, impaired insulin secretion, diabetes, obesity, and elevated blood pressure in adult progeny in various studies (34, 36, 37). In contrast, the role of catch-up growth in early postnatal life after PR or other types of prenatal challenge, in causing the metabolic syndrome in adults, has not been directly examined in any nonhuman species to date. Unlike the human, the PR rat does not appear to undergo catch-up growth immediately after birth (37). Neonatal catch-up growth does occur after spontaneous growth restriction in the pig, particularly in male animals, and was associated with insulin resistance in young adult pigs of both sexes; however, other aspects of the metabolic syndrome were not studied, and independent effects of birth size and neonatal growth were not assessed in those studies (29, 30).
PR can be readily induced in sheep, with metabolic, endocrine, and growth consequences for the fetus similar to those of IUGR human fetuses (25, 26, 32, 33). This chronic restriction of placental and hence fetal growth in sheep reduces size at birth and importantly is followed by increased fractional growth rates (FGR) and normalization of weight and length by 45 days of age, indicating that the PR lamb undergoes catch-up growth in the first month of life (6, 7, 12). Our group (5) has recently reported that PR and small size at birth are each associated with increased insulin sensitivity of glucose metabolism in young lambs at 1 mo of age but that glucose-stimulated insulin secretion is already impaired in young lambs. In the present study, we investigated the later postnatal effects of experimentally induced PR on components of the metabolic syndrome [insulin sensitivity, secretion and action on glucose and free fatty acids (FFA), blood pressure, and glucose and lipid homeostasis] in chronically catheterized and conscious adult sheep. In addition, we assessed the independent effects of size at birth and early postnatal growth rates on each of these metabolic and cardiovascular outcomes in the adult sheep after experimental PR.
Males appear more susceptible to perinatal programming of metabolic and cardiovascular homeostasis than females in other experimental models of fetal growth restriction (39) and after intrauterine growth restriction in humans. At 20 yr of age, insulin sensitivity was reduced in men but not in women who were light or short at birth (10). Early-life influences also accounted for a greater proportion of variance in metabolic syndrome score in men (∼12%) than in women (∼5%) at ∼50 yr old, and birth weight and catch-up growth in childhood were related to central metabolic score and to fasting insulin, respectively, in men but not in women at 50 yr of age, implying that men are more susceptible than women to early life influences before and after birth (28). Therefore, the influence of sex on adult responses to PR and associations of components of the metabolic syndrome with size at birth and early postnatal growth were also examined.
MATERIALS AND METHODS
All procedures in this study were approved by the University of Adelaide Animal Experimentation and Ethics Committee and complied with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Placental growth was restricted (PR group) in 45 Merino ewes by removal of the majority of visible endometrial caruncles (65–148) from the nonpregnant uterus (33). Six to ten caruncles were left in each horn of the bicornuate uterus. In sheep, placental cotyledons form during pregnancy only at the caruncles on the maternal endometrium, and this therefore reduces the number of cotyledons formed in the subsequent pregnancy and hence placental size and function (24). General anesthesia was induced after an overnight fast by an intravenous injection of thiopentone (1–2 g; Boehringer Ingelheim, Sydney, NSW, Australia) and maintained by inhalation of 2–3% halothane in oxygen. Ewes received a daily 3.5 ml im injection of antibiotics (250 mg/ml procaine penicillin, 250 mg/ml dihydrostreptomycin, 20 mg/ml procaine hydrochloride; Ilium Penstrep, Troy Laboratories, Smithfield, NSW, Australia) for 3 days commencing on the day of surgery. After a 6-wk recovery period, these carunclectomized ewes, along with control ewes, entered a timed mating program. Estrus was synchronized by insertion of intravaginal sponges containing 4 mg of the synthetic progestagen flugestone acetate for a 12-day period, and mating was assumed to occur 2 days after removal of the sponges. Mating dates were recorded, and pregnancies were confirmed by ultrasound. One week before parturition, ewes were housed in individual pens in animal holding rooms, with a constant photoperiod (12:12-h light-dark period). Data from 42 offspring (22 control, 20 PR), with at least one measure of blood pressure or of glucose tolerance, were included in the study. Not all measures were obtained for all animals, mostly because of loss of catheter patency, and a flow diagram of animal use is shown in Fig. 1. Both treatment groups included twin lambs [5 control twin lambs (1 male and 4 females), 8 PR twin lambs (4 males and 4 females)], with the remainder singleton births. Progeny was coded as twins to reflect the number of lambs present at birth, including where one still-born and one live-born lamb were delivered. If possible, both twins were studied. Gestational age at delivery was calculated as the number of days since mating. Pregnant ewes, lactating ewes, and growing and adult progeny were fed lucerne chaff and had access to water ad libitum. Pregnant and lactating ewes also received 150 g of oats daily.
Measurement of size at birth and early postnatal growth.
Size (weight, crown-rump length, shoulder height) was measured at birth, at 5, 10, 15, 20, 30, and 45 days of age, and in adult progeny. Ponderal index was calculated as weight/(crown-rump length)3 (kg/cm3). The absolute growth rate (AGR) for each size parameter was calculated as the growth of that parameter per day and was obtained by linear regression for the early postnatal period (0–45 days of age). As previously reported, growth was linear during this period (7). The FGR for each parameter was then calculated as the early postnatal AGR divided by the size of that parameter at birth.
Surgery and catheter maintenance.
At 360 days of age, surgery was performed under general anesthesia and aseptic conditions to insert vascular catheters into the right carotid artery and jugular vein of progeny, as described previously (12). General anesthesia was induced after an overnight fast by an intravenous injection of thiopentone (0.5–1 g; Boehringer Ingelheim) and maintained by inhalation of 2–3% halothane in oxygen. Lambs received a daily 2-ml im injection of antibiotics (250 mg/ml procaine penicillin, 250 mg/ml dihydrostreptomycin, 20 mg/ml procaine hydrochloride; Ilium Penstrep, Troy Laboratories) for 3 days commencing on the day of surgery. To ensure patency, catheters were flushed with heparinized saline (500 U/ml) daily for 3 days postsurgery and then every second day for the duration of the experimental period. A 4-day recovery period was allowed before the commencement of experiments.
Measurement of systolic and diastolic blood pressure.
Measurement of the insulin sensitivity of glucose and lipid metabolism in vivo.
At 373 days of age and after a 12-h overnight fast, whole body insulin sensitivity of glucose metabolism was measured in vivo, using the hyperinsulinemic euglycemic clamp technique (8, 13). Arterial blood was sampled (2 ml) 10, 5, and 0 min before the start of the clamp for the determination of the fasting blood glucose and plasma insulin and FFA concentrations. Human insulin (Actrapid, Novo Nordisk) was infused at 2 mU·kg−1·min−1 iv for 2 h from 0 min. Arterial blood was sampled (0.2 ml) at 5-min intervals throughout the clamp, and blood glucose was rapidly measured (HemoCue). At 15 min, an intravenous infusion of glucose (25% dextrose) was commenced at 2 mg·kg−1·min−1, and this rate was adjusted every 5 min to restore and maintain euglycemia. Plateau rates of glucose infusion during the second hour of the clamp were termed steady-state glucose infusion rates. Arterial blood was sampled (2 ml) at 60, 75, 90, 105, and 120 min, and steady-state insulin and FFA concentrations were calculated as the average plasma concentrations in the second hour of the hyperinsulinemic euglycemic clamp. Calculations and equations describing the insulin sensitivity of glucose metabolism have been more fully described previously (13). Briefly, the insulin sensitivity of net whole body glucose uptake (insulin sensitivityglucose) was calculated as the steady-state glucose infusion rates needed to maintain euglycemia corrected for the steady-state plasma insulin concentration. The insulin sensitivity of lipid metabolism (insulin sensitivityFFA) was calculated as the change in plasma FFA concentrations in response to insulin divided by the steady-state plasma insulin concentration (Eq. 1). (1)
The metabolic clearance rate of insulin was calculated as the insulin infusion rate throughout the hyperinsulinemic euglycemic clamp divided by the increase in plasma insulin from fasting to steady-state concentrations (13).
Measurement of glucose tolerance and insulin secretion in vivo.
At 375 days of age and after a 12-h overnight fast, glucose tolerance and insulin secretion were measured with the use of an intravenous glucose tolerance test (IVGTT). Arterial blood was sampled (2 ml) 5, 3, and 0 min before the commencement of the IVGTT for the determination of fasting blood glucose and plasma insulin. A bolus of 0.25 g glucose (25% dextrose)/kg live weight was then infused intravenously over a 1-min period, and the start time was taken as the start of the bolus infusion. Arterial blood was sampled (2 ml) 2, 5, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100, 120, 140, 160, 180, and 210 min after administration of glucose for measurement of blood glucose and plasma insulin concentrations. Indexes of glucose tolerance and insulin secretion were calculated as described previously (13). Briefly, glucose tolerance was defined by the area under the glucose concentration curve, insulin secretion was calculated as the area under the insulin concentration curve, and relative insulin secretion was calculated as area under the insulin concentration curve divided by area under the glucose concentration curve. Area under the curve was calculated as the area between the baseline concentration and the concentration profile during the IVGTT with the use of the Sigma Scan Pro v4 software package (Jandel Scientific Software). First phase was defined as between 0 and 20 min after glucose administration, based on the presence of a trough or plateau in plasma insulin at that time in the majority of animals. Second phase was defined as between 20 and 100 min after glucose administration, when blood glucose had returned to basal levels in the majority of animals.
Calculation of glucose disappearance, posthepatic insulin delivery rates, and disposition indexes.
Glucose rate of disappearance, the rate of glucose elimination after a glucose dose, was calculated as the slope of the regression line of natural logarithm-transformed blood glucose concentration vs. time, early (2–20 min) and late (20–100 min) in the IVGTT. Posthepatic insulin delivery rates and insulin disposition indexes (DI) for glucose metabolism (DIglucose) were calculated as described previously (13). These calculations require independent measures of insulin secretion and clearance and/or sensitivity and were therefore calculated only for those animals in which intact catheters permitted both the IVGTT and hyperinsulinemic euglycemic clamp to be completed (16 control, 13 PR sheep). DIs for lipid metabolism (DIFFA) were calculated similarly by multiplying the insulin sensitivity of lipid metabolism measured during the hyperinsulinemic euglycemic clamp (insulin sensitivityFFA) by measurements of insulin delivery in the fasting and glucose-stimulated states. Basal DIFFA was calculated as insulin sensitivityFFA multiplied by the basal posthepatic insulin delivery rate (Eq. 2). Maximal DIFFA was calculated as insulin sensitivityFFA multiplied by the maximal posthepatic insulin delivery rate (Eq. 3). Change in DIFFA was calculated as insulin sensitivityFFA multiplied by the increase in plasma insulin from fasting to maximum concentration during the IVGTT (Eq. 4). (2) (3) (4)
Collection of fed blood samples for measurement of triglycerides and cholesterol.
Blood samples (5 ml) were collected from fed sheep between 0830 and 0930 at 378 days of age, which had ad libitum access to lucerne chaff (12). Plasma was stored at −20°C until analyzed as described below.
Analysis of plasma insulin and metabolite concentrations.
Plasma insulin concentrations were measured in duplicate by a double-antibody, solid-phase RIA using a commercially available kit (Phadaseph insulin RIA; Pharmacia, Upsala, Sweden). The intra- and interassay coefficients of variation for the insulin assay were 3.7 and 5.9%, respectively. This kit measures both human and ovine insulin, as previously reported by our group (13). Blood glucose concentrations were measured with a glucometer (HemoCue). Plasma triglyceride, cholesterol, and FFA concentrations were measured in duplicate by enzymatic colorimetric analysis, using commercially available kits (Unimate TRIG, Roche Diagnostic Systems, for triglycerides; CHOL, Roche Diagnostic Systems, for cholesterol; and NEFA C, Wako Pure Chemical Industries, for FFA). Plasma triglyceride and FFA concentrations were each measured in single assays with intra-assay coefficient of variation <4%.
Effects of treatment (control or PR) and sex were evaluated by two-way ANOVA using the SPSS for Windows version 13.0 (SPSS, Chicago, IL). Insulin data were log-transformed before statistical analysis, and other data were transformed before statistical analysis to obtain normal distribution and equal variance across groups when necessary. Normal distribution and equal variance could not be obtained by transformation of AGR data; therefore, this was analyzed by a Mann-Whitney U-test. Where interaction terms for sex were significant, separate ANOVAs for treatment effects were also performed for data from males and females. Blood pressures were analyzed by two-way ANOVA with and without current body weight as a covariate. Individual glucose and insulin concentrations throughout the IVGTT were also analyzed by repeated-measures analysis using the linear mixed-model procedure of SPSS for Windows version 13.0, for one within factor (sample number) and two between factors (treatment and sex) and their interactions, with a first-order autoregressive covariance structure. Untransformed blood glucose and log-transformed plasma insulin concentrations were analyzed by repeated-measures analysis across the whole profile. Insulin concentrations were also analyzed separately during first- and second-phase insulin secretion (2–20 min and 20–100 min, respectively). Treatment differences in the shape of changes with time were assessed with polynomial contrasts. The independent effects of size at birth and neonatal FGR on adult functional outcomes were assessed by multiple linear regression. One-sided P values were used to test the a priori hypotheses that small size at birth and increased rates of growth in early postnatal life would be associated with indicators of the metabolic syndrome in young adult sheep, based on similar relationships reported in humans (28). A probability of 5% or less was accepted as statistically significant. Data are presented as means ± SE.
Size at birth and early postnatal growth.
Gestational age at delivery was not affected by either PR or sex (overall mean ± SE: 151.3 ± 0.6 days; P > 0.4 for each). In this cohort of sheep that were studied at ∼1 yr of age, PR reduced birth weight by 28%, crown-rump length by 11%, and shoulder height by 6% (Table 1); size at birth did not differ between sexes for any parameter. AGR and FGR during the early postnatal period did not differ between males and females for any parameter (P > 0.1 for each), except for AGR for weight, which was higher in males than in females (males: 0.288 ± 0.014 kg/day, females: 0.251 ± 0.008 kg/day; P = 0.019, Mann Whitney U-test for nonnormally distributed data). AGR results for weight, crown-rump length, and shoulder height were similar in control and PR lambs (Table 1). FGR for weight was 43% higher in PR lambs than in control lambs, and FGR results for crown-rump length and shoulder height were not altered by PR (Table 1).
At 365 days of age, males weighed more than females (males: 47.4 ± 2.1 kg, females: 38.3 ± 1.4 kg; P = 0.001), and body weight was not altered by PR (P > 0.6). Similarly, shoulder height was greater in males than in females (males: 68.1 ± 1.2 cm, females: 63.9 ± 1.1 cm; P = 0.039) but not altered by PR (P > 0.2), whereas crown-rump length did not vary with treatment or sex (P > 0.2 for each).
Blood glucose during the IVGTT differed between males and females (P = 0.049), and the effects of PR and time during the IVGTT on blood glucose varied between sexes (P < 0.05 for each) and were therefore analyzed separately for males and females. Blood glucose changed differently with time in control compared with PR males (Fig. 2A; P < 0.001) and was higher overall in control than in PR females (Fig. 2B; P = 0.017). PR altered the early blood glucose response to IVGTT in males, increasing blood glucose at 2 (P = 0.027) and 5 (P = 0.044) min after the glucose bolus, compared with controls (Fig. 2A).
PR did not alter fasting blood glucose overall but affected the maximum change in blood glucose during the IVGTT differently with sex (P = 0.037; Table 2), inducing a greater change in PR males than in control males (P = 0.030) but not affecting the change in glucose in females (P > 0.5). The maximum blood glucose during the IVGTT was consequently higher in PR than in control males (Fig. 2A; control: 10.0 ± 0.3 mmol/l, PR: 11.6 ± 0.7 mmol/l; P = 0.027) but did not differ between PR and control females (Fig. 2B; control: 10.8 ± 0.3 mmol/l, PR: 10.4 ± 0.4 mmol/l; P > 0.4). The time for glucose to return to baseline values (91.9 ± 3.8 min) was unaffected by PR or sex (P > 0.3 for each). Glucose disappearance (Table 2) was more rapid for PR than for control sheep during the early phase of the IVGTT (2–20 min after glucose; P = 0.044) but not later (20–100 min after glucose; P > 0.8) and was similar in males and females. Glucose tolerance, assessed as glucose area under the curve, was unaffected by PR or sex.
Plasma insulin during the IVGTT differed between males and females (P = 0.018), which were therefore analyzed separately. PR did not alter fasting plasma insulin in males or plasma insulin throughout the first-phase response to glucose in either males or females (Fig. 2, C and D; P > 0.3 for each). PR altered plasma insulin differently with time during the second-phase response in males (P = 0.016; 4th-order polynomial) but did not alter that in females (P = 0.5).
Fasting insulin (8.5 ± 0.8 mU/l; Table 2) and the ratio of fasting insulin to glucose (2.30 ± 0.23 mU/mmol) were unaffected by PR or sex (P > 0.1 for each). The maximum plasma insulin concentration (50.4 ± 3.9 mU/l), the maximum change in plasma insulin (Table 2), and time for plasma insulin to return to baseline values (78.5 ± 4.4 min) after glucose were each unaffected by either PR or sex (P > 0.1 for each). First-phase insulin secretion was unaffected by PR or sex (P > 0.1 for each), but second-phase insulin secretion was reduced by PR overall (P = 0.027) and greater (P = 0.002) in male than in female sheep (Table 2). Insulin secretion throughout the entire IVGTT was greater in males than in females (P = 0.027) and unaffected by PR (Table 2). Relative insulin secretion (i.e., corrected for the area under the glucose profile) was higher in males than in females during second phase (P = 0.015) but was not affected by PR overall or during first- or second-phase insulin secretion (P > 0.1 for each) (Table 2). PR had opposite effects on basal posthepatic insulin delivery rate (Table 2) in males and females (P = 0.041), although PR did not change basal posthepatic insulin delivery within either sex (P ≥ 0.17 for each).
Insulin sensitivity and action on glucose metabolism.
PR increased the glucose infusion rate required to maintain euglycemia during the hyperinsulinemic clamp (P = 0.038; Table 3). Steady-state plasma insulin and the derived metabolic clearance rate for insulin were unaffected by PR or sex (Table 3). PR altered insulin sensitivityglucose differently with sex, such that insulin sensitivityglucose was similar in control and PR males (P > 0.7) but was higher in PR females than in controls (P = 0.038) (Table 3). Similarly, PR altered basal DIglucose (Table 3) differently with sex (P = 0.023), although this did not differ between control and PR males (P = 0.19) or control and PR females (P = 0.052). Maximal DIglucose and change in DIglucose were not altered by PR or sex (Table 3).
Plasma total cholesterol and triglycerides in the fed state were unaffected by PR or sex (Table 4). PR altered fasting plasma FFA concentration differently with sex (P = 0.044), so that fasting plasma FFA was unchanged in males but was reduced in PR females compared with controls (P = 0.024; Table 4).
Insulin sensitivity and action on lipid metabolism.
Plasma FFA concentrations were decreased to a greater extent by insulin infusion in control than in PR females (P = 0.046) but were similar in control and PR males (Table 4). Insulin sensitivityFFA and insulin DI for lipid metabolism (DIFFA) were unaffected by PR or sex (Table 4).
Systolic BP was higher in male than in female sheep (P = 0.021) but was not altered by PR (P > 0.2) (Fig. 3). When current body weight was included as a covariate, systolic blood pressure was positively predicted by current body weight (covariate; P = 0.030) and not affected by sex or PR (P > 0.2 for each) (Fig. 3). Effects of PR on diastolic blood pressure (Fig. 3) varied with sex (P = 0.012), with PR reducing diastolic BP in males (P = 0.045) but not in females (P > 0.3). When current body weight was included as a covariate, diastolic BP was positively predicted by current weight (covariate; P = 0.038) and PR altered diastolic blood pressure, adjusted for current weight, differently in males and females (P = 0.010), being reduced in males (control: 95.3 ± 3.7 mmHg, PR: 81.0 ± 5.0 mmHg; estimated means adjusted to mean male body weight of 48.0 kg; P = 0.042) but not in females (control: 80.1 ± 2.7 mmHg, PR: 83.6 ± 2.3 mmHg; estimated means adjusted to mean female body weight of 38.3 kg; P > 0.3). Pulse pressure (Fig. 3) was not affected by sex or PR with or without current weight as a covariate (P > 0.3 for all).
Size at birth, early postnatal growth, and postnatal outcomes.
Size at birth and FGR in early postnatal life independently predicted adult size in both sexes but independently predicted most other adult functional outcomes in males only. Multiple linear regressions of the effects of weight at birth and neonatal FGR in terms of weight on adult outcomes are presented in Table 5, and relationships with long bone size and thinness at birth or neonatal FGR results for long bone lengths are included in the text below only where these relationships differ markedly from those for weight.
In males, adult body weight was independently positively correlated with neonatal FGR for weight, whereas adult shoulder height was positively correlated with both birth weight and neonatal FGR for weight (Table 5). The positive relationships between adult size and size at birth or neonatal growth were consistent across various measures of size in males. In females, adult size was weakly correlated with birth weight and not correlated with neonatal FGR for weight (Table 5), but adult weight was independently and negatively predicted by ponderal index at birth and neonatal FGR for weight or shoulder height (overall and partial correlations, each P < 0.05).
In males, fasting blood glucose was weakly positively correlated with birth weight but not related to rates of neonatal growth (Table 5). In females, fasting blood glucose (overall: R = 0.615, P = 0.023) was independently and negatively correlated with ponderal index at birth (partial correlation: R = −0.452, P = 0.046) and neonatal FGR for shoulder height (partial correlation: R = −0.515, P = 0.025). Other measures of glucose homeostasis were not independently correlated with size at birth and neonatal FGR in females (Table 5). In males, first-phase glucose disappearance was independently and negatively related to neonatal FGR for weight (Table 5) and was also independently and negatively correlated (overall: R = 0.787, P = 0.001) with ponderal index at birth (partial correlation: R = −0.664, P = 0.004) and FGR for weight (partial correlation: R = −0.697, P = 0.002). Glucose tolerance in terms of area under the curve was also independently and negatively correlated with birth weight in males (Table 5).
In males, fasting insulin concentrations and their ratio with glucose were independently and positively correlated with birth weight (Table 5). Conversely, relative first-phase insulin secretion during the IVGTT was independently and negatively correlated with birth weight (Table 5). In females, birth weight was not independently correlated with measures of insulin secretion (Table 5), but thinness at birth and neonatal FGR for weight were independently and negatively correlated with the period of elevated circulating insulin during the IVGTT (overall: R = 0.615, P = 0.018; partial correlation for birth ponderal index: R = −0.532, P = 0.017; partial correlation for neonatal FGRweight: R = −0.538, P = 0.016), and similar trends were observed for fasting insulin concentrations in females (overall and partial correlations, each P < 0.1). Basal and glucose-stimulated insulin delivery in males were consistently independently and positively correlated with size at birth but not with neonatal FGR for weight and not related to size at birth and neonatal FGR in females (Table 5).
Insulin sensitivity and action on glucose metabolism.
In males, insulin sensitivityglucose was not independently correlated with birth weight and FGR for weight (Table 5). Insulin sensitivityglucose in males (overall: R = 0.711, P = 0.021) was, however, independently and negatively correlated with FGR for shoulder height (partial correlation: R = −0.646, P = 0.016) but not with birth weight (partial correlation: R = −0.118, P = 0.365). Basal and glucose-stimulated insulin action in males were consistently independently and positively correlated with size at birth but not with neonatal FGR for weight, whereas insulin action in females was not independently correlated with size at birth or neonatal growth rates (Table 5).
Fed and fasting plasma lipid concentrations were not correlated with birth weight or FGR for weight in either sex (Table 5). Fed plasma cholesterol in both sexes was independently negatively predicted by FGR for shoulder height (partial correlations: P < 0.05 for both), although not related to thinness at birth.
Insulin sensitivity and action on lipid metabolism.
Insulin sensitivityFFA was not independently correlated with size at birth or neonatal FGR in either sex (Table 5). DIFFA results were, however, independently and negatively correlated with size at birth in males only (Table 5).
Systolic and diastolic blood pressures were not correlated with birth weight and FGR for weight in either sex (Table 5). In males, diastolic and systolic blood pressures were each independently and positively correlated with birth weight and FGR for shoulder height (partial correlations: P < 0.05 for each). In females, pulse pressure (overall: R = 0.571, P = 0.015) was independently and negatively correlated with ponderal index at birth (partial correlation: R = −0.503, P = 0.012), although not significantly with FGR for shoulder height (partial correlation R = 0.348, P = 0.067).
This study shows that PR or small size at birth induces some components of the metabolic syndrome in adult male sheep, including reduced insulin delivery and action, impaired glucose homeostasis, and elevated glucose-stimulated insulin secretion. Conversely, PR reduced diastolic blood pressure and did not change fasting lipid abundance but increased insulin action on lipid metabolism in adult male sheep, which might contribute to obesity in the long term. We have reported previously that PR increased insulin action on glucose and lipid metabolism in the fasted young lamb at 1 mo of age but impaired their insulin secretion in response to a glucose challenge, and that thinness at birth was also associated with increased insulin action in the fasting and glucose-stimulated states in the young lamb (5). The present study shows that this effect of PR and small size at birth in sheep is reversed by adulthood, particularly in males, when insulin action is now impaired after PR or small size at birth.
Small size at birth is itself a predictor of rapid or “catch-up” growth in early postnatal life, and both of these predict adverse metabolic (4, 11, 22) and hypertensive (15) outcomes in human children and adults. We therefore estimated the separate effects of size at birth and early postnatal FGR on outcomes in adult sheep in the present study. FGR is a marker of the outcome of interactions between the animal's physiological capacity for growth, its anabolic drive, and the environment, particularly nutritional, that it experiences. Increased FGR is predictive of increased visceral obesity in the young lamb (7), as in humans, suggesting a potential mechanistic link between early behavior and later metabolic impairment. PR reduced size at birth in this cohort of sheep to an extent similar to that shown in other cohorts that we studied as young lambs (5, 7). Also consistent with our previous reports, PR increased FGR for weight in this cohort during the early postnatal period, indicative of attempted catch-up growth and presumably increased central obesity (5, 7). Increased size at birth was associated with greater adult size in the present study, particularly in males, whereas females who were thin at birth and had low FGR in early postnatal life were heavier as adults but did not have increased frame size, suggestive of increased adipose deposition in later life. Small size at birth in terms of weight was associated with impaired glucose tolerance in the adult male sheep, when singletons and twins were combined, reflecting primarily impaired insulin secretion, as fasting insulin and insulin disposition were reduced. These relationships appeared to be consistent between singleton and twin lambs. Similar to small size at birth, a high FGR in terms of weight during early postnatal life was predictive of slower early glucose clearance in adult male sheep, although high early postnatal FGR did not predict most components of the metabolic syndrome in adult males. Rapid early postnatal FGR of apical bone lengths, but not in terms of weight, did predict poorer whole body insulin sensitivity of glucose metabolism in adult male sheep, but whether this relates to expanded adipose tissue is not clear. Furthermore, high FGR during the early postnatal period was not predictive of insulin secretion in adult male sheep, suggesting that prenatal restraint is the predominant influence on and responsible for the impaired insulin secretion seen postnatally. Small size at birth predicted increased insulin action on FFAs in adult male sheep, independent of early postnatal FGR, suggesting enhanced sensitivity of adipose tissue to insulin, which could lead to increased lipid storage and obesity. Similar enhanced sensitivity to insulin occurs in the young lamb after PR and small size at birth and is predictive of early-onset obesity (5). The present findings suggest that this persists and, if it continues, may lead to adverse metabolic and other consequences. The nature of these relationships with markers of prenatal and early postnatal experiences in adult male sheep contrast with outcomes of human epidemiological studies, where metabolic outcomes in children and adults are independently and adversely affected by both prenatal restraint and postnatal catch-up growth (4, 11, 22), and implicate prenatal restraint as the more important factor for long-term metabolic outcomes in the sheep. Also, in contrast to human studies (15), low birth weight predicted reduced blood pressure in the adult male sheep, independent of early postnatal growth rates, although rapid growth of apical bones during early postnatal life predicted elevated blood pressure. This suggests that early postnatal life experience is an important determinant of blood pressure in the adult placentally restricted sheep. Because markers of the environment and behavior in early postnatal life are predictive of insulin action and blood pressure in both sheep and humans, outcomes of such studies suggest that interventions soon after birth may have the potential to prevent later emergence of insulin resistance and hypertension after intrauterine growth restriction, but whether the adverse consequences for insulin secretion can be reversed at this stage is less clear. It should be noted that exendin-4, a GLP analog that stimulates expansion of β-cell mass by increasing rates of neogenesis and of proliferation of existing β-cells, administered to neonatal offspring of placentally restricted rats is able to restore insulin sufficiency (38). It is unclear whether this intervention would also be successful in a species where substantial pancreatic development occurs before birth, such as the sheep and humans, in contrast to the rat. In summary, prenatal restraint seems predictive of impaired glucose homeostasis and insulin secretion and an increased propensity to accumulate lipid, whereas rapid fractional rates of growth in early postnatal life predict impaired insulin sensitivity of glucose metabolism, and increased blood pressure, in the adult male sheep.
The effects of PR and small size at birth on metabolic and cardiovascular outcomes in the adult sheep differed between males and females, with improved insulin sensitivity and secretion and glucose homeostasis as well as decreased plasma FFA in the latter. High FGR in early postnatal life also predicted lower circulating cholesterol in adult female sheep. Together with our earlier paper (7) describing a negative relationship between size at birth and insulin sensitivity in young lambs of both sexes (i.e., greatest insulin sensitivity in the smallest lambs at birth), the present data are consistent with differing timing of onset in the two sexes. Insulin sensitivity is still higher in PR females than in control females at 1 yr of age, but insulin action does not differ at this age; however, in males, the elevated insulin sensitivity seen at 1 mo is lost by 1 yr of age, and insulin action is impaired by 1 yr of age. Thinness at birth did predict some of the early signs of impaired glucose homeostasis in females, which is one of the early antecedents of the metabolic syndrome, and suggests that females that were small at birth will also go on to develop the syndrome, but later than males. Sex differences in prenatal programming of insulin sensitivity are also evident in young adult men and women. Low weight or length at birth predicts poor insulin sensitivity in young adult men, but not women, although glucose tolerance is maintained in the former by increased insulin secretion and glucose effectiveness (10). Two studies using direct measurements by hyperinsulinemic euglycemic clamp reported that intrauterine growth restriction or low birth weight decreased the insulin sensitivity of whole body glucose uptake in young adult men and women combined (17) or of glycolysis, although not whole body glucose uptake in young adult men (18). Furthermore, in the rat, exposure to a low-protein diet in gestation and lactation induces insulin resistance in young adult male but not female progeny (39). In contrast, small size at birth predicts glucose intolerance or diabetes and impaired insulin sensitivity consistently in epidemiological studies of men and women over 40 yr old (20). Together with observations in human IUGR infants and 2-yr-old children (16, 35) and in offspring of protein-deprived pregnant rats (27), these results in a range of species suggest that restricted fetal growth is followed initially by improved insulin sensitivity, which switches to insulin resistance with aging. Furthermore, the timing of this switch appears to differ between sexes, occurring by young adulthood in men and later in women. Consistent with these sex differences in programming of adult insulin sensitivity, the effects of early life influences on risk of the metabolic syndrome in middle age are significant in both sexes but are ∼2.5-fold stronger in men than in women (28).
The observed relationships between birth size and markers of insulin secretion in older men and women are variable (20), probably reflecting methodological limitations and possibly partial compensatory insulin secretion in some populations but failure of insulin secretion and progression to diabetes in others. Further studies utilizing independent measurements of insulin sensitivity and secretion, allowing insulin disposition to be calculated, are needed to determine whether insulin secretion may nevertheless be inadequate for the reduced insulin sensitivity of older adult men and women who were of low birth weight. In sheep, insulin action on glucose metabolism (disposition) is negatively related to birth weight in young lambs of both sexes (5), and in the present study we have shown that this reverses by adulthood, at least in male sheep. Similarly, low birth weight reduced insulin DI in young adult men (18) and in children at 3 but not at 1 yr of age (19). Another study found no effect of low birth weight, however, on insulin disposition in male and female prepubertal children (41).
We conclude that restriction of fetal growth in sheep, induced by chronic restriction of placental growth, induces some aspects of the metabolic syndrome in adult male offspring, particularly impaired insulin action. Like that shown in humans, impaired metabolic homeostasis appears to develop earlier in male than in female sheep after restricted fetal growth, and we predict that female sheep would also develop impaired metabolism with aging. The separate predictive effects of size at birth and early postnatal FGRs on aspects on the metabolic syndrome in the adult sheep suggest that neonatal interventions may have the potential to prevent emergence of insulin resistance and hypertension after intrauterine growth restriction but that prenatal restraint is the more important determinant of impaired insulin secretion in the sheep. We are currently using this experimental paradigm to investigate the mechanistic basis of impaired insulin secretion and action after prenatal restraint and accelerated FGR in early life.
This work was supported by the National Health and Medical Research Council of Australia (NHMRC). K. L. Gatford held the Peter Doherty Postdoctoral Fellowship of the NHMRC and subsequently the Hilda Farmer Medical Research Associateship of the University of Adelaide.
We thank Simon Fielke, Arkadi Katsman, Melissa Walker, and Damian Adams for assistance with in vivo studies and the staff of Laboratory Animal Services, University of Adelaide, for assistance with animal care.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by American Physiological Society