Main

Epidemiologic studies show that the early life environment may have long-term effects on the risk of adult-onset diseases such as hypertension and coronary heart disease (CHD) (1). In studies of the Dutch Hunger Winter, maternal undernutrition during early gestation, when the nutrient demands of the conceptus are minimal, is associated with increased prevalence of cardiovascular disease (CV) disease in adulthood (2). In utero undernutrition in humans may result in CV and metabolic and growth adaptations (3–5) aimed at increasing energy supply or reducing energy expenditure but with postnatal consequences. The nature, timing, and intensity of an in utero nutritional challenge, as well as maternal age, parity, and offspring number and sex, will determine the degree of constraint (6) imposed on the fetus and are likely to dictate the strategy adopted by it to enhance immediate survival.

One way of reducing fetal energy expenditure is to reduce growth, as shown in several animal species including rat (7) and sheep (8). However, maternal nutrient restriction during pregnancy, in the absence of effects on fetal weight or birth weight, does result in altered CV control in the offspring. In sheep, a 15% reduction in maternal nutrient intake for the first half of gestation produced low fetal arterial blood pressure (9) and impaired small artery function (10) in late-gestation singleton fetuses, but no change in weight. This was followed by increased blood pressure in early postnatal life (11). Altered postnatal CV function has been confirmed in both rats (12) and sheep (13) to be associated with perturbations of the peri-implantation environment. Moreover, the first 30 days of gestational age (dGA) have been identified as a critical window for in utero effects on renal development and later CV function (14,15). These observations suggest that net fetal growth can be sustained during maternal nutrient restriction and that CV adaptations may be involved.

An additional adaptive strategy to enhance postnatal health/survival after undernutrition could be adjustment of gestation length. Altered gestation length will depend on the size of the nutritional challenge. In sheep, a 15% reduction in maternal body weight by undernutrition in the peri-implantation period (−60 to 30 dGA) resulted in reduced gestation length mediated by accelerated maturation of the fetal hypothalamopituitary-adrenal (HPA) axis (16). This suggests that the strategy adopted to cope with this level of maternal nutritional constraint is early delivery. In contrast, a 15% reduction in maternal nutrient intake during early gestation reduced pituitary and adrenal responses in late-gestation ovine fetuses (17). This may be expected to prolong gestation, but the effect of such a relatively mild challenge on gestation length is unknown. The aim of this study was therefore to investigate the effect of such a nutritional challenge on gestation length in relation to other constraints on the fetus, which are known to influence gestation length, such as number and sex of fetuses.

METHODS

Animals.

All procedures were carried out in accordance with the regulations of the UK Home Office Animals (Scientific Procedures) Act of 1986.

Welsh Mountain ewes in their second or third parity (following on from a pregnancy in the previous season) and of uniform good body condition (n = 88) were brought in from pasture and housed on straw in open barns at the Royal Veterinary College, Hertfordshire, UK, where they were acclimatized to a complete pelleted diet (GJW Titmus, Hertfordshire, UK). The diet consisted of barley, wheat, micronized full fat soya, grass meal, molasses, chopped straw, calcium carbonate, dicalcium phosphate salt, and sheep vitamin/mineral supplement. As fed, it provided 9.6 MJ/kg (metabolizable energy) and 14.75 g crude protein. Rations were set according to initial body weight and adjusted according to guidelines for pregnant sheep (AFRC Technical Committee, 1993). Water was provided ad libitum.

In the normal breeding season, the estrous cycle of groups of three to four ewes was synchronized by withdrawing vaginal medroxyprogesterone acetate impregnated sponges (Veramix, Upjohn Ltd., Crawley, UK) 12 d after insertion. One of three raddled Welsh Mountain rams was introduced for 3 d, and 0 dGA was taken as the first day on which an obvious raddle mark was observed. Pregnancy was confirmed by measuring plasma progesterone concentrations at 16 dGA using an enzyme immunoassay kit (Ridgeway Science Ltd., Avington, UK) and fetus number was determined by ultrasound scanning at 60 dGA. Ewes were individually housed on straw from 7 d before conception to 37 dGA and group housed thereafter with animals at a similar gestational age.

Before conception, ewes were randomly assigned to a control group or a dietary restricted group. From 1–31 dGA (term = 147 d), ewes received either 100% [control group (C), n = 41; 20 singleton (s) and 21 twin (t)] or 50% of global nutritional requirements (undernutrition group (U), n = 47; 29s and 18t)] and 100% of requirements for the remainder of gestation. Maternal body weight and body condition score (BCS on scale of 0–5, where 0 is emaciated and 5 is obese) were measured weekly.

Cortisol and adrenocorticotropin (ACTH) were measured by immunoassay (Immulite analyzer, DPC, UK; single measurement) in maternal plasma (ethylenediaminetetraacetic acid treated) collected on 0, 30, and 65 dGA. The sensitivity of the cortisol assay was 0.2 μg/dL and the intra- and interassay coefficients of variation were 4.6% and 7.9%, respectively. The sensitivity of the ACTH assay was 9 pg/mL and the intra- and interassay coefficients of variation were 1.1% and 4.6%, respectively.

At 70 dGA, a subset of pregnant ewes (C, n = 8; 5s and 3t and U, n = 11; 8s and 3t) was killed by an overdose of barbiturate (pentobarbitone sodium Ph. Eur. 160 mg /kg i.v.; Animalcare Ltd., York, UK) and fetuses were removed (C, n = 10; 5s and 5t (loss of one fetus) and U, n = 14; 8s and 6t). Fetal and placental weight and fetal crown-rump length (CRL), abdominal circumference (AC), femur length (FL), and biparietal diameter (BPD) were recorded. Fetal organs were removed and weighed.

The remaining ewes delivered and suckled their lambs naturally until weaning at 12 wk of age (C, n = 51; 15s and 36t and U, n = 51; 21s and 30t lambs). There were approximately equal numbers of males and females in each group. At birth, all lambs were weighed and CRL, AC, FL, and BPD measurements were taken. Lambs were weighed again at 12 wk of age.

Statistics.

Data are expressed as mean ± SEM. Biometric data were analyzed using an unpaired t test or univariate analysis of variance (ANOVA) full factorial model in which factors were early gestation diet, sex, and twinning (SPSS Inc., Chicago, IL). When analyzing the 70-dGA fetal biometric data, the factor of sex was excluded from the ANOVA, as there were insufficient animal numbers to allow this. Gestation length was also analyzed using Kaplan-Meier survival analysis (SPSS Inc.). Significance was accepted at p < 0.05.

RESULTS

Maternal body weight.

At conception, there was no difference in maternal body weight (C, 42.8 ± 0.7 kg; U, 41.9 ± 0.7 kg) or BCS (C, 3.0 ± 0.1; U, 2.8 ± 0.1) between the two groups. There was no difference in weight at conception between twin-bearing and singleton-bearing ewes (twin, 43.08 kg versus singleton, 41.61 kg). By 28 dGA, U ewes had gained less weight (p < 0.001, Fig. 1) and had a lower BCS (−0.04 ± 0.08 versus 0.53 ± 0.09, p < 0.001) than C ewes. In addition, twin-bearing ewes gained less weight than singleton-bearing ewes by 28 dGA (p < 0.05, Fig. 1).

Figure 1
figure 1

Maternal weight gain between 0 and 28 dGA. C (open columns: single n = 18 and twin n = 20) and U (solid columns: single n = 28 and twin n = 18) groups. *p < 0.001, U significantly different from C, two-way ANOVA. §p < 0.05, twin-bearing ewes significantly different from singleton-bearing ewes, two-way ANOVA. Data are mean ± SEM.

Full size image

Maternal cortisol and ACTH concentrations.

At 0 and 65 dGA, there were no differences in maternal baseline plasma cortisol concentrations between the dietary groups. At 30 dGA, U ewes had lower baseline plasma cortisol than C ewes (p < 0.05, Fig. 2). At 0, 30, and 65 dGA, there were no differences in maternal baseline plasma ACTH concentrations between the dietary groups (Fig. 2). No differences were found in baseline cortisol and ACTH concentrations between singleton and twin pregnancies.

Figure 2
figure 2

Maternal plasma cortisol (A) and ACTH (B) on 0, 30, and 65 dGA. C (open columns: n = 41) and U (filled columns: n = 44) groups. *p < 0.05, U significantly different from C, two-way ANOVA. Data are mean ± SEM.

Full size image

Fetal and placental biometry.

Early gestation undernutrition had no effect on maternal weight, placental weight, placentome number, total conceptus weight, fetus/placenta weight ratio, fetus weight, biometry, or organ weights (except thymus and pancreas, which were increased in U, Table 2) at 70 dGA (Table 1). In twin pregnancies, the total conceptus weight was greater, with fewer placentomes per fetus (41.4 ± 1.0 versus 79.5 ± 4.3) and a greater fetus/placenta ratio compared with singleton pregnancies (p < 0.01, Table 1). In singleton, but not twin, fetuses at 70 dGA, CRL was less (p < 0.05) in U compared with C (Table 1).

Table 2 Fetal organ weights as a percentage of body weight (and as absolute weight) in C and U sheep fetuses at 70 DGA
Full size table
Table 1 Fetal weight and biometric measurements in C and U sheep fetuses at 70 DGA
Full size table

Gestation length.

There was no difference in gestation length between the two dietary groups (C, 145.3 ± 0.4 d versus U, 146.1 ± 0.4 d). Regardless of dietary group, ewes carrying twins delivered earlier than those carrying singleton fetuses (p < 0.05, ANOVA and Kaplan-Meier analysis, Fig. 3) and there was a tendency for a shorter gestation length in male same-sex than female same-sex twins (p = 0.06, data not shown). In ewes carrying a male plus female twin combination, but not same-sex twins, there was a significant increase in gestation length in U (146.5 ± 0.6 d, n = 8) compared with group C (144.6 ± 0.6 d, n = 9) (p < 0.05, t test and Kaplan-Meier analysis, Fig. 4). In male twin, but not singleton, lambs, gestation length was inversely correlated with maternal weight gain between 0 and 28 dGA (p < 0.05, R2 = 0.14).

Figure 3
figure 3

Kaplan-Meier survival analysis of the cumulative incidence of nondelivery in ewes carrying singleton or twin fetuses. §p < 0.05, twin-carrying ewes (broken line) significantly different from singleton-carrying ewes (solid line).

Full size image
Figure 4
figure 4

Kaplan-Meier survival analysis of the cumulative incidence of nondelivery in ewes carrying different-sex twin fetuses. *p < 0.05, U (broken line) significantly different from C (solid line).

Full size image

Birth and postnatal biometry.

At birth, regardless of dietary group, twin lambs had smaller (p < 0.05) biometric measurements than singletons (Table 3). There was no difference in weight and biometry between male and female lambs (Table 3). In both singleton and twin fetuses (male and female combined), there were no differences between C and U in weight (singleton: C, 3.9 ± 0.2 kg versus U, 4.0 ± 0.2 kg; twin: C, 3.7 ± 0.1 kg versus U, 3.6 ± 0.1 kg) or biometry (data not shown) at birth. In male twin, but not singleton, lambs, birth weight was inversely correlated with maternal weight gain between 0 and 28 dGA, an effect that was stronger in same-sex male twins (Fig. 5). There was no relationship between birth weight and maternal weight change 0–28 dGA in female fetuses. In all animals together, weight gain between birth and 12 wk old was positively correlated with birth weight in U (p < 0.05, R2 = 0.10), but not C (p = 0.68, R2 = 0.004). At 12 wk of age, (weaning) twins weighed less (p < 0.001) than singletons, and females weighed less (p < 0.001) than males (Table 3). In twin, but not singleton, male offspring, weight at 12 wk and weight gain between birth and 12 wk were greater (p < 0.05) in U compared with C (Table 3).

Table 3 Birth weight and biometric measurements
Full size table
Figure 5
figure 5

The relationship between birth weight and maternal weight change during 0–28 dGA. (A) Male twin lambs R2 = 0.15, p < 0.05. (B) Same-sex male twins R2 = 0.40, p < 0.01.

Full size image

DISCUSSION

This study has shown that moderate maternal nutrient restriction from 1 to 31 dGA in sheep affects twin pregnancies to a greater extent than singleton pregnancies. Moreover, this nutrient restriction produces a small but significant increase in gestation length in different-sex twins and increased weight gain by weaning in male twins.

Maternal body composition and nutrition as well as age, parity, and offspring number and sex determine the degree of constraint imposed on the fetus and are likely to dictate the strategy it adopts to enhance immediate survival. In this study, the undernutrition given during early gestation may have acted to constrain fetal growth (although no measurements of growth were taken at this time in this study). Male embryos may be affected by undernutrition to a greater extent than females as they grow at a greater rate (18). However there were insufficient numbers in the present study to analyze according to sex. Maternal undernutrition had no effect on placental weight, fetal body weight, or biometry at 70 dGA or weight at birth. Although we cannot rule out a type II error, the undernutrition did increase the thymus and pancreas (as a percentage of body weight) at 70 dGA. Such increased pancreatic weight may result from altered maturation and lead to altered postnatal function. It may be too soon or too difficult to detect small changes in body weight or biometry at 70 dGA, but the effects of the undernutrition may become apparent in later life in terms of catch-up growth.

Constraint on fetal growth and development is likely to be greater in twin than singleton pregnancies, each of which has a distinct fetal growth pattern (19). In this study, twin fetuses were smaller than singletons as evident from biometry at birth but not mid-gestation weight or birth weight. Maternal weight gain during 0–28 dGA was less in twin-carrying compared with singleton-carrying ewes. This suggests that even around the time of implantation, a twin pregnancy has a greater demand on maternal energy expenditure or increases partitioning of nutrients toward the conceptus, each of which could deplete maternal body resources. In this study, the ewes could not increase food intake to cope with such demands as we have rationed their food intake according to 100% or 50% of recommended food intake for singleton pregnancies, as it is not technically feasible to assess fetal number by ultrasonography at the time in gestation that the challenge was carried out.

Twin fetuses may have a greater exposure to the physiologic, neuroendocrine, or metabolic adaptations made by the mother during times of suboptimal nutrition. Alternatively, twins may receive a greater nutritional insult than the singleton fetuses by virtue of a greater fetus-to-placenta weight ratio. The effects of adaptations to undernutrition when combined with twinning may become more apparent later in gestation as growth rate and energy demands increase and in utero space becomes more limiting. Such speculation is supported by the small (R2 = 0.15) but significant association between increased birth weight and reduced ewe weight gain from 0 to 28 dGA in male twin, but not singleton, lambs. Interestingly, there was a similar but stronger association (R2 = 0.4) in male-male twins, and in female offspring, no such association was seen. These data suggest that the effects of undernutrition on fetal growth are also influenced by its sex. At 12 wk of postnatal age, twin male lambs, which received early gestation undernutrition, were heavier and had an increased weight gain from birth until 12 wk old compared with controls. Hence, the increased constraint of being a twin and a male embryo (18) in a nutrient-restricted early intrauterine environment induces a phenotype more likely to gain weight once removed from the constrained in utero environment and adequate nutrition is restored. Such an increase in weight gain could be detrimental to CV function in later life, as studies in humans have shown that accelerated weight gain during early postnatal life is associated with increased risk of CV disease in adulthood (20,21). The male sheep in this study have indeed shown an association between increased weight at weaning and reduced glucose tolerance at 2.5 y of age (22), a factor associated with the metabolic syndrome.

In different-sex twin pairs only, maternal undernutrition increased gestation length. We speculate that the female fetus influences this increased gestation length because different-sex and female same-sex human twins are reported to have longer gestation than male same-sex twins (23). In keeping with this concept, there was a tendency for longer gestation length in female same-sex compared with male same-sex twins in the present study; however, the mechanisms underlying this sex-twin effect are unclear. Interestingly, in male twin, but not singleton, lambs, gestation length was inversely correlated with maternal weight gain between 0 and 28 dGA. These data suggest that, in male twins, undernutrition (reduced maternal weight gain) leads to increased birth weight and increased gestation length. These observations could be a direct consequence of alterations to the preimplantation embryo environment because in vitro culture of sheep embryos increases gestation length compared with in vivo embryos (24). The increase in the growth rate of ewes in the period after cessation of the nutritional insult may also have contributed to the increased gestation length because in humans (25) and in a high-growth mouse model [resulting from the lack of expression of Socs-2 (see Corvo et al. (26)], a greater weight gain during pregnancy is associated with a longer gestation length.

Mechanisms linking altered intrauterine environment to altered gestation length could involve the HPA axis, as the preparturient fetal cortisol surge is essential for the initiation of normal labor (27). Indeed, in the present study, nutrient restriction resulted in reduced maternal plasma cortisol levels at 30 dGA, and previously we demonstrated that mild undernutrition (a 15% reduction in maternal nutrient intake) during early gestation reduced pituitary and adrenal responsiveness in late gestation ovine fetuses (28). This is consistent with our current observation of a small increase in gestation length after a moderate maternal nutritional challenge. In contrast, accelerated maturation of the HPA axis seen with a more severe undernutrition challenge (16) appears to trigger early maturation and delivery of the fetus. Taken together, these findings suggest that the developmental strategy adopted by the fetomaternal unit depends greatly on the timing, duration, and severity of the insult. The increased gestation length, although modest, could thus be part of a strategy for allowing more time in utero for maturation and growth or could be via changes in HPA axis function, which convey some later advantage.

In summary, this study has shown that moderate early gestation maternal nutrient restriction can increase gestation length without altering mid-gestation fetal size or birth weight and can have additional influences on postnatal growth. However, these findings are influenced by fetal sex and number, indicating that aspects of the maternal diet may be transmitted in a graded manner to the embryo during the very early stages of development. The results of this study will be important for informing public health strategies on diet in pregnancy aimed at reducing the risk of later CV and metabolic disease in humans.