Introduction
In the sheep, it is well established that the prepartum activation of the fetal hypothalamo–pituitary–adrenal (HPA) axis is essential for the normal timing of parturition (term =150 ± 3 days of gestation) and for a successful transition to extrauterine life.Reference Liggins1–Reference MacLaughlin and McMillen3 Interestingly, it has been demonstrated that the timing of the prepartum activation of the fetal HPA axis is different in singletons and twins.Reference Edwards and McMillen4–Reference Rumball, Oliver and Thorstensen6 Plasma adrenocorticotropic hormone (ACTH) concentrations are lower, the prepartum cortisol surge occurs later and adrenocortical responsiveness to ACTH is blunted in twin compared with singleton fetal sheep.Reference Edwards and McMillen4–Reference Rumball, Oliver and Thorstensen6 It has been suggested that there may be a programmed delay in the prepartum activation of the HPA axis, which could act to protect the twin fetuses from preterm delivery, given that twins are exposed to the additional stress of a decreased placental substrate supply in late gestation.Reference MacLaughlin and McMillen3, Reference Bloomfield, Oliver and Hawkins7 Although there is some evidence that there may be an increased impact of the negative feedback actions of cortisol on ACTH secretion in the late-gestation twin fetal sheep,Reference Rumball, Oliver and Thorstensen6 it is not known whether there are differences in the expression of factors in the fetal pituitary that regulate the synthesis and secretion of ACTH between twin and singleton fetuses in late gestation.
It has been shown that a 30% reduction in maternal nutrition during the periconceptional period (PCUN) for at least 45 days before until 7 days after conception results in an earlier activation of the HPA axis in twin but not in singleton fetuses during late gestation in sheep.Reference Edwards and McMillen4 When maternal nutrition is restricted more severely and for a period which extends beyond the preimplantation period to include the first 30 days of gestation, there is an earlier activation of the HPA axis in singleton fetuses in late gestation and this is associated with an increased probability of preterm delivery.Reference Bloomfield, Oliver and Hawkins8 One possibility is that PCUN results in a decrease in the negative feedback actions of glucocorticoids on the synthesis and secretion of ACTH in the fetal pituitary, either through a decrease in intrapituitary glucocorticoid availability or through a decrease in glucocorticoid receptor (GR) expression and that this occurs, to a greater extent, in the twin compared with the singleton fetus. There are conflicting reports, however, of the effects of PCUN on GR expression in the fetal pituitary, as maternal undernutrition from 60 days before until 30 days after conception did not result in a change in the GR expression in the fetal pituitary in late gestation,Reference Bloomfield, Oliver and Hawkins7 whereas maternal undernutrition imposed from conception to 70 days of gestation resulted in a decrease in the pituitary GR expression.Reference Hawkins, Hanson and Matthews9 Given the variation in the length of the period of exposure to periconceptional undernutrition in prior studies, it is not clear whether the effects of maternal undernutrition on the development of the HPA axis are a consequence of the effects of maternal undernutrition acting on the oocyte or on the embryo.
A series of studies have demonstrated that early environmental events can alter the central GR expression through changes in the chromatin structure and DNA methylation and have long-term effects on the HPA axis and stress response in the offspring.Reference Weaver, Cervoni and Champagne10–Reference Stevens, Begum and Cook13 It has been demonstrated in the rat, for example, that different levels of the hippocampal GR expression resulting from different patterns of early maternal care are due to changes in the expression of GR mRNA containing the exon 17 promoter.Reference Weaver, Cervoni and Champagne10 In the late-gestation fetal sheep, Stevens et al. Reference Stevens, Begum and Cook13 reported that maternal undernutrition during the periconceptional period from 60 days before until 30 days after conception resulted in the increased GR mRNA expression and hypomethylation of the GR promoter region in the hypothalamus, but no change in the hippocampal and pituitary GR mRNA expression and methylation levels. However, there have been no studies that have investigated whether there are changes in the epigenetic state of the GR gene in the HPA axis after exposure to maternal undernutrition during the periconceptional period from before until only 1 week after conception in either singleton or twin fetuses in late gestation.
In the present study, our aim was to determine the critical period of exposure to maternal undernutrition during the periconceptional period on the expression of genes within the fetal anterior pituitary, which regulates adrenal growth and steroidogenesis in late gestation. We have therefore determined the impact of maternal undernutrition during the periconceptional period in ewes from at least 45 days before to 6 days after conception (PCUN) and separately during the first 6 days only of the 16-day preimplantation period (PIUN) on the expression of ACTH precursor – proopiomelanorcortin (POMC), prohormone convertase 1 (PC1), which cleaves POMC to ACTH, two distinct isoforms of 11β-hydroxysteroid dehydrogenases (11βHSD1 and 11βHSD2) and GR in the fetal anterior pituitary in twin and singleton pregnancies at 136–138 days of gestation, that is, just before the prepartum cortisol surge.Reference McMillen, Warnes and Adams14–Reference Tomlinson and Stewart16 We have also investigated the specific effects of PCUN and PIUN on the level of methylation in the nerve growth factor-inducible protein (NGF1-A)-binding region of the GR promoter in the fetal anterior pituitary in late gestation.
Materials and methods
Animals and nutritional protocols
A total of 35 South Australian Merino ewes were used in this study. Ewes were moved into an enclosed shed and housed in pens for 2 weeks before the start of the feeding regime. During this period, ewes were acclimatized to a diet, which consisted of lucerne chaff and pellets containing cereal hay, lucerne hay, barley, oats, almond shells, lupins, oat bran, lime and molasses (Johnsons & Sons Pty Ltd., Kapunda, South Australia, Australia). Eighty per cent of the total energy requirements were obtained from the lucerne chaff and 20% of the energy requirements from the pellet mixture. The lucerne chaff provided 8.3 MJ/kg metabolizable energy, 193 g/kg crude protein and contained 85% dry matter, and the pellets provided 8.0 MJ/kg metabolizable energy, 110 g/kg crude protein and contained 90% dry matter. All ewes received 100% of nutritional requirements to provide sufficient energy for the maintenance of a non-pregnant ewe as defined by the Agricultural and Food Research Council.17 At the end of this acclimatization period, ewes were randomly assigned to one of the three feeding regimes:
• Control group (n = 12): ewes received 100% of nutritional requirements from at least 45 days before mating until 6 days after mating.
• Periconceptional undernutrition (PCUN) group (n = 14): ewes received 70% of the control allowance from at least 45 days before mating until 6 days after mating.
• Preimplantation undernutrition (PIUN) group (n = 9): ewes received 70% of the control allowance for 6 days after mating only.
All of the dietary components were reduced by an equal amount in the restricted diet. From day 7 of pregnancy, all ewes were fed a control diet (100% of requirements) until post mortem between 136 and 138 days of gestation.
Ewes were released in a group in the evening with rams of proven fertility that were fitted with harnesses and marker crayons. Ewes were individually penned the following morning and the occurrence of mating was confirmed by the presence of a crayon mark on the ewe's rump. The day of mating was defined as day 0. Ewes were weighed approximately every week after commencing the feeding regime until post mortem at 136–138 days of pregnancy. Pregnancy and fetal number were diagnosed by ultrasound between 40 and 80 days of gestation. The number of fetuses carried by each ewe was confirmed at post mortem.
Animals and surgery
Pregnant ewes (Control, n = 12; PCUN, n = 14; PIUN, n = 9) were transferred to an animal holding facility between 90 and 100 days of gestation (term = 150 ± 3 days). Surgery was performed under general anaesthesia and aseptic conditions between 105 and 110 days of gestation as previously described.Reference Edwards, Simonetta, Owens, Robinson and McMillen18 In brief, vascular catheters were inserted in a fetal carotid artery and jugular vein, a maternal jugular vein and the amniotic cavity. Vascular catheters were only inserted into one fetus in twin pregnancies. All ewes and fetal sheep received a 2 ml intramuscular injection of antibiotics (procaine penicillin 250 mg/ml, dihydrostreptomycin 250 mg/ml, and procaine hydrochloride 20 mg/ml; Penstrep Illium, Troy Laboratories, Smith-field, New South Wales, Australia) at the time of surgery. The ewes were housed in individual pens in animal holding rooms with a 12-hour light/dark cycle and were fed once daily at 1100 h with water provided ad libitum. Animals were allowed to recover from surgery for at least 4 days before experimentation.
Blood sample collection
Fetal arterial blood samples (3.5 ml) were collected in chilled tubes on the morning of post mortem between 136 and 138 days of gestation. Blood samples for cortisol assay were collected in heparinized tubes (125 IU, Sarstedt, South Australia, Australia), and blood samples for ACTH assay were collected in tubes coated in ethylenediaminetetraacetic acid (EDTA) (Sarstedt, Australia) containing aprotinin (1000 kIU/ml, Sigma Chemicals, St Louis, MI, USA). Blood samples were centrifuged at 1500 g for 10 minutes, and plasma was separated into aliquots and stored at −20°C for subsequent hormone assays.
Collection of tissues
A total of 19 singleton-bearing ewes (Control, n = 7; PCUN, n = 9; PIUN, n = 3) and 16 twin-bearing ewes (Control, n = 5; PCUN, n = 5; PIUN, n = 6) were killed with an overdose of sodium pentobarbitone (Virbac Pty Ltd., Peakhurst, NSW, Australia) between 136 and 138 days of pregnancy, and fetal sheep were immediately weighed and killed by decapitation. Fetal organs including the pituitary and adrenals were dissected and weighed. The anterior lobe was separated from the neurointermediate lobe of the fetal pituitary before snap-freezing in liquid nitrogen and storage at −80°C. The fetal anterior lobe of the pituitary was available from 17 singleton fetuses (Control, n = 6; PCUN, n = 8; PIUN, n = 3) and 26 twin fetuses (Control, n = 9; PCUN, n = 7; PIUN, n = 10).
Cortisol assay
Cortisol was extracted from fetal plasma samples in duplicate using dichloromethane and measured with a radioimmunoassay (PerkinElmer Pty Ltd., Waltham, MA, USA) previously validated for use in fetal sheep plasma.Reference MacLaughlin, Walker and Kleemann19 The efficiency of recovery of 125I-cortisol from fetal plasma using this extraction procedure was always >90%. Samples were then reconstituted in assay buffer (Tris hydrochloride, BSA and sodium azide). Rabbit anti-cortisol (1:450 dilution, MP Biomedicals, Seven Hills, NSW, Australia) was added followed by 125I-labelled cortisol (Amersham Pharmacia Biotech, Buckinghamshire, Little Chalfont, UK). Tubes were vortexed and incubated at 37°C for 1 hour before the addition of goat anti-rabbit serum (1:30 dilution, Chemicon, Billerica, MA, USA) and 20% polyethylene glycol (BDH Laboratory Supplies, Poole, Dorset, UK). After centrifugation at 3700 × g and 4°C for 30 minutes, the supernatant was aspirated and the precipitate counted on a Gamma-counter (Packard, Downers Grove, IL, USA). The sensitivity of the assay was 0.2 nmol/l, and the intraassay and interassay coefficients of variation were <5% and 6.3%, respectively.
ACTH assay
Immunoreactive ACTH concentrations in fetal sheep plasma were measured using a DiaSorin ACTH radioimmunoassay kit (DiaSorin S.A., Vercelli, Italy) previously validated for fetal sheep plasma.Reference Warnes, McMillen, Robinson and Coulter20 Rabbit anti-ACTH was added into samples and incubated overnight at 4°C followed by 125I-labelled ACTH. After a further incubation at 4°C overnight, precipitating complex including goat anti-rabbit serum and polyethylene glycol was added and incubated at room temperature for 20 minutes. Tubes were vortexed and centrifuged at 3000 rpm at 15°C for 30 minutes. The supernatant was aspirated and the precipitate counted on a Gamma-counter (Packard, Downers Grove, IL, USA). The sensitivity of the assay was 4.5 pg/ml, and the intraassay and interassay coefficients of variation were <10% and 13.9%, respectively.
Quantitative real-time RT-PCR
Total RNA was extracted from each fetal anterior pituitary using Trizol reagent (Invitrogen, Groningen, The Netherlands) and purified using the RNeasy Mini Kit (Qiagen, Basel, Switzerland).Reference MacLaughlin, Walker and Kleemann19 cDNA was synthesized by reverse transcription using superscript III (Invitrogen). Negative controls containing no RNA or superscript III were used to test for DNA contamination.
The relative abundance of GR, 11βHSD1, 11βHSD2, POMC and PC1 mRNA transcripts was measured by quantitative real-time PCR using Applied Biosystems 7500 Fast Real-Time PCR System (Applied Biosystems, Foster, CA, USA). A PCR reaction consisted of 5 μl of Fast SYBR Green Master Mix (Applied Biosystems), 1 μl each of forward and reverse primers (GeneWorks, Adelaide, South Australia, Australia) for the appropriate genes (Table 1),Reference Myers, Bell, Hyatt, Mlynarczyk and Ducsay21–Reference Fletcher, Roberts, Hartwich, Walker and McMillen24 2 μl of molecular-grade H2O, and 1 μl of cDNA (50 ng/μl) in a total volume of 10 μl. The PCR products were confirmed by sequencing. Three replicates of cDNA were performed for each gene, and controls containing no cDNA were included on each plate. Amplification efficiencies were determined from the slope of a plot of Ct (defined as the threshold cycle with the lowest significant increase in fluorescence) against the log of a series of diluted cDNA concentrations (ranging from 1 to 100 ng/μl). The abundance of each transcript relative to the abundance of the reference gene cyclophilin was calculated using Q-Gene analysis software.Reference Muller, Janovjak, Miserez and Dobbie25
Table 1 Primer sequences used for real-time RT-PCR and bisulphite PCR
PC1, prohormone convertase 1; POMC, proopiomelanorcortin; GR, glucocorticoid receptor; 11βHSD1, 11β-hydroxysteroid dehydrogenase type 1.
Methylation status of the GR promoter
To examine the methylation status of the ovine GR promoter, we first cloned the sequence upstream of the reported ovine GR cDNA (GenBank Accession EU371026) using genome walking. The cloned 4.5 kb promoter sequence has been deposited in GenBank under Accession HM204706. Genomic DNA was extracted from fetal pituitaries using standard phenol–chloroform methods and then subjected to bisulphite modification (EpiTect, Qiagen). The region of the ovine GR promoter homologous to the rat exon 17 promoter was amplified from 100 ng bisulphite-modified DNA using primers and conditions that amplified methylated and unmethylated templates with no bias (Table 1). PCR amplicons were cloned into a plasmid vector (PGEM-T Easy, Promega, Madison, WI, USA); individual clones were sequenced and analysed using BiQ analyser software (http://biq-analyzer.bioinf.mpi-inf.mpg.de/).Reference Bock, Reither and Mikeska26
Statistical analyses
All data are presented as the mean ± s.e. of the mean (s.e.m.). All variables were analysed using the Statistical Package for Social Sciences (SPSS) for Windows version 17.0 (SPSS Inc., Chicago, IL, USA). The change of maternal weight during the periconceptional period was compared between the treatment groups (Control, PCUN or PIUN) using a one-way analysis of variance (ANOVA). Two-way ANOVA was used to determine the effects of maternal nutritional treatment and fetal number on fetal weight, absolute and relative pituitary and adrenal weights, plasma cortisol and ACTH concentrations, and gene expression in the fetal anterior pituitary. The Duncan's New Multiple Range post-hoc test was used to identify significant differences between mean values. Relationships between variables were assessed by linear regression using Sigma Plot 10.0 (SPSS Inc.). A probability of 5% (P < 0.05) was taken as the level of significance for all analyses.
Results
Effects of PCUN and PIUN on maternal weight
The weights of the non-pregnant ewes assigned to the control (59.0 ± 3.2 kg), PCUN (59.0 ± 2.4 kg) and PIUN (58.7 ± 2.9 kg) groups were not significantly different before the start of the feeding regime. Ewes in the PCUN group significantly lost more weight (−2.7 ± 0.4 kg, n = 14) when compared with the control (1.9 ± 0.7 kg, n = 12) and PIUN ewes (0.8 ± 0.8 kg, n = 9) during the periconceptional period (P < 0.001). There was no significant weight change in the PIUN group compared with the control group during the period between the starting of feeding regime and 6 days after conception.
Fetal weight and fetal adrenal and pituitary weights
There was no effect of nutritional treatment or fetal number on fetal weight during late gestation (Table 2). Total adrenal weight was significantly higher in the PCUN group compared with the control group, independent of fetal number (P = 0.011, Table 2). There was no difference, however, in total adrenal weight between the PCUN and PIUN groups in either singleton or twin pregnancies. Relative adrenal weight was significantly higher in twins compared with singletons, independent of maternal nutritional treatment (P = 0.048). There was no effect of maternal nutritional treatment on either absolute or relative pituitary weight. The absolute but not relative pituitary weight was significantly lower in twin fetuses compared with singleton, independent of maternal nutritional treatment (P = 0.027, Table 2).
Table 2 Effect of maternal undernutrition (PCUN or PIUN) during the periconceptional period and fetal number on fetal weight, adrenal weight and pituitary weight during late gestation

PCUN, periconceptional undernutrition; PIUN, preimplantation undernutrition.
Values are means ± s.e.m.
a,bDifferent superscripts denote treatment groups that are significantly different from each other (P < 0.05).
*Significant difference between singleton and twin fetuses (P < 0.05).
Fetal plasma ACTH and cortisol concentrations
There was no effect of maternal nutritional treatment, fetal number, or any interaction between the effects of nutritional treatment and fetal number on fetal plasma ACTH and cortisol concentrations or the plasma cortisol: ACTH ratio within 24 hours of post mortem between 136 and 138 days of gestation (Table 3).
Table 3 Plasma cortisol and ACTH concentrations and the plasma cortisol: ACTH ratio in singleton and twin fetuses at 136–138 days of gestation

ACTH, adrenocorticotropic hormone; PCUN, periconceptional undernutrition; PIUN, preimplantation undernutrition.
Values are means ± s.e.m.
GR mRNA expression in the fetal anterior pituitary
GR mRNA expression in the fetal anterior pituitary was significantly lower in the PCUN and PIUN groups compared with controls in both singleton and twin pregnancies (P = 0.013, Fig. 1). There was no difference in pituitary GR mRNA expression between the PCUN and PIUN groups. There was a weak, inverse relationship between the pituitary GR mRNA expression (y) with total adrenal weight (x) (y = −0.26x + 0.36, r 2 = 0.14, P = 0.018) and relative adrenal weight (x) (y = −1x + 0.34, r 2 = 0.12, P = 0.026) when data from all fetuses were combined. There was also a positive correlation between the pituitary 11βHSD1 (y) and GR mRNA expression (x) when data from all nutritional treatment groups were combined (y = 0.11x + 0.05, r 2 = 0.14, P = 0.017).
Fig. 1 The pituitary GR mRNA expression in singleton and twin fetuses in the control, PCUN and PIUN groups (Control: open bar; PCUN: black-filled bar; PIUN: grey-filled bar) during late gestation. Different superscipts denote treatment groups that are significantly different from each other (P < 0.05). GR, glucocorticoid receptor; PCUN, periconceptional undernutrition; PIUN, preimplantation undernutrition.
GR promoter methylation
Bisulphite sequencing was carried out on DNA from a subset of fetal pituitaries to determine the methylation status of the GR at the region corresponding to the exon 17 promoter. This region, containing the NGF1-A consensus sequence, was largely unmethylated in control fetuses and this was unchanged in the PCUN and PIUN treatment groups (Fig. 2).
Fig. 2 Bisulphite sequencing of the region uspstream of the ovine GR in fetal pituitaries. (a) Genomic structure of the GR showing the region cloned (grey bar; GenBank Accession Number HM204706) and the sequence amplified in bisulphite PCR; grey shading within the sequence indicates the NGF1-A consensus site. (b) Bisulphite sequencing maps: each block represents sequences derived from the pituitary of a single fetus. Each horizontal row represents one sequenced allele, and each box represents a CpG dinucleotide within the amplified region: black represents methylated and white unmethylated (Control: 1, 2; PCUN: 3, 4; PIUN: 5). NGF1-A, nerve growth factor-inducible protein.
POMC, PC1, 11βHSD1 and 11βHSD2 mRNA expression in the fetal anterior pituitary
There was no effect of either maternal nutritional treatment or of being a twin on POMC or PC1 mRNA expression in the fetal anterior pituitary at 136–138 days of gestation (Table 4). There was a significant correlation between PC1 (y) and POMC (x) mRNA expression in the fetal anterior pituitary when data from all fetuses were combined (y = 0.06 x + 0.02, r 2 = 0.12, P = 0.023).
Table 4 Pituitary POMC, PC1 and 11βHSD2 mRNA expression in singleton and twin fetuses during late gestation

POMC, proopiomelanorcortin; PC1, prohormone convertase 1; 11βHSD1, 11β-hydroxysteroid dehydrogenase type 2; PCUN, periconceptional undernutrition; PIUN, preimplantation undernutrition.
Values are means ± s.e.m.
There was no effect of either maternal nutritional treatment or fetal number on the 11βHSD2 mRNA expression (Table 4); however, there was a significant interaction between the effects of nutritional treatment and fetal number on the 11βHSD1 mRNA expression in the fetal anterior pituitary. In twin fetuses, the pituitary 11βHSD1 mRNA expression was significantly higher (P = 0.015, Fig. 3) in the PIUN group compared with the PCUN but not control group.
Fig. 3 The pituitary 11βHSD1 mRNA expression in singleton and twin fetuses in the Control, PCUN and PIUN groups (Control: open bar; PCUN: black-filled bar; PIUN: grey-filled bar) during late gestation. Different superscipts denote treatment groups that are significantly different from each other in twin fetuses (P < 0.05). 11βHSD1, 11β-hydroxysteroid dehydrogenase type 1; PCUN, periconceptional undernutrition; PIUN, preimplantation undernutrition.
Discussion
A novel finding of the current study was that maternal undernutrition imposed either before conception and extending for 1 week after conception, or imposed only during the first week after conception resulted in the lower GR mRNA expression in the anterior pituitary of the singleton and twin fetuses in late gestation. This suggests that one critical period for the effects of maternal undernutrition during the periconceptional period on the development of the pituitary–adrenal axis is during the first week of embryonic life.
PCUN and PIUN and GR mRNA expression
GR mRNA expression at 136–138 days of gestation was lower in the pituitaries of singleton and twin fetuses, which had been exposed to either PCUN or PIUN some 130 days earlier. A previous study on singleton fetuses, exposed to a more severe maternal nutrient restriction from before conception to 30 days of gestation, found that there was no difference in the pituitary GR mRNA expression as determined by in situ hybridization between PCUN and control groups at 132 days of gestation.Reference Bloomfield, Oliver and Hawkins7 In contrast, an earlier study in which the period of 15% reduction in maternal nutrition was extended to 70 days of gestation reported that the pituitary GR mRNA expression was significantly lower in the fetuses exposed to the period of maternal undernutrition.Reference Hawkins, Hanson and Matthews9 This suggests that the fetal GR expression in the anterior pituitary is dependent on the timing of dietary restriction, intensity and duration of nutritional insult before conception and/or in early pregnancy.
There are a number of mechanisms that might explain the suppression of GR expression in the fetal pituitary in response to exposure to maternal undernutrition during the first week of embryonic life, including an increased negative feedback action of cortisol on the GR expression. However, there was no difference in basal plasma ACTH and cortisol concentrations between the PCUN, PIUN and control fetuses at 136–138 days of gestation. This is consistent with the previous findings that maternal undernutrition between 60 days before until 30 days after conception did not result in an increase in either basal plasma ACTH and cortisol concentrations in singleton and twin fetuses at 128–131 days of gestation.Reference Rumball, Oliver and Thorstensen6 Thus, the suppression of the pituitary GR expression does not appear to be a consequence of increased circulating cortisol at 136–138 days of gestation. There was also no evidence for a downregulation in the pituitary 11βHSD2 mRNA expression in singletons and twins in the PCUN or PIUN groups. Although the pituitary 11βHSD1 expression appeared to be upregulated in PIUN twin fetuses, there was a positive, rather than a negative, relationship between the intrapituitary 11βHSD1 and GR expression when data from all groups were combined. This suggests that the decrease in the pituitary GR expression is unlikely to be a consequence of increased negative feedback effect of intrapituitary cortisol.
We have assessed whether the effects of PIUN and PCUN on the pituitary GR expression may be a consequence of an induced change in the epigenetic state of the GR promoter. It has been demonstrated in the rat, for example, that there are different levels of GR expression in the hippocampus of offspring exposed to different patterns of maternal care in the first few weeks of life and that these are a consequence of changes in the expression of GR mRNA derived from the exon 17 promoter.Reference Weaver, Cervoni and Champagne10 Furthermore, the low hippocampal GR expression is associated with hypermethylation of the 5′ CpG dinucleotide within an NGF1-A consensus sequence located within this promoter.Reference Weaver, Cervoni and Champagne10 In the present study, however, we have found no evidence that there is an increased level of methylation at this site located 3 kb upstream of the translational start site (TSS) in exon 2 in the sheep GR gene. It has been previously shown in the rat that there are a range of GR mRNAs that encode a common protein, but that differ in their 5′-leader sequences as a consequence of alternate splicing of around 11 different exon 1 sequences, each with its own upstream promoter region. In the rat, there are significant levels of at least six first exon variants, with tissue-specific differences in promoter activity between the liver, hippocampus and thymus.Reference McCormick, Lyons and Jacobson27 Postnatal handling of the rat pup results in a selective increase in GR mRNA containing the hippocampus-specific exon 1, which results in a permanent increase in the GR expression in the hippocampus. Analysis of the 5′ region of the human GR gene has revealed nine untranslated alternative first exons and 13 splice variants and its tissue-specific promoter usage has been extensively examined in a range of human tissues.Reference Presul, Schmidt, Kofler and Helmberg28–Reference Turner, Alt and Cao30 In the fetal sheep, Stevens et al. Reference Stevens, Begum and Cook13 reported that maternal undernutrition during the periconceptional period from 60 days before until 30 days after conception resulted in the increased GR mRNA expression and hypomethylation of the GR promoter region in the hypothalamus, but no change in the hippocampal and pituitary GR mRNA expression and methylation levels. It is therefore possible that periconceptional undernutrition may downregulate the GR expression within the pituitary through an action on a pituitary-specific GR promoter that is distinct from the hippocampal exon 17 promoter, which is yet to be characterized. Thus, the possibility remains that maternal undernutrition during the first week after conception results in epigenetic changes at the GR in the developing pituitary to ensure an enhanced response to fetal and postnatal stress in anticipation of a life of continuing adversity.
PCUN and PIUN and pituitary POMC expression and circulating ACTH
A lower level of pituitary GR mRNA expression might be expected to decrease the negative feedback actions of cortisol in the PCUN and PIUN groups, leading to an increase in the pituitary POMC expression, circulating ACTH 1–39 and cortisol concentrations, and increased adrenal growth. However, the downregulation of the pituitary GR mRNA expression in the PCUN and PIUN groups was not associated, with an increase in pituitary POMC and PC1 mRNA expression in either singleton or twin fetuses. Previous studies have also reported that maternal undernutrition imposed beyond the periconceptional period (up to either 30- or 70 days of gestation) does not alter the POMC and PC1 mRNA expression in the anterior pituitary in late gestation.Reference Bloomfield, Oliver and Hawkins7, Reference Hawkins, Hanson and Matthews9 It has been demonstrated in the adult and fetal sheep pituitary that there are subpopulations of corticotrophic cell types, which are differentially sensitive to the negative feedback actions of cortisol.Reference Neill, Smith and Luque31–Reference Butler, Schwartz and McMillen33 Farrand et al.Reference Farrand, McMillen, Tanaka and Schwartz34 previously identified three major subpopulations of corticotrophs that expressed POMC and/or ACTH. One possibility is that the decrease in the pituitary GR mRNA expression in the PCUN and PIUN groups may result in an increase in the POMC expression in defined corticotroph subpopulations, and thus measurement of the POMC expression in whole-tissue samples may mask changes occurring within these subpopulations. In a previous study, we found that, whereas there was no effect of PCUN on basal plasma ACTH and cortisol concentrations in singletons, circulating immunoreactive ACTH concentrations were higher in PCUN than control twin fetuses between 110 and 147 days of gestation.Reference Edwards and McMillen4 It has been shown that there are a range of POMC-derived peptides present in the circulation of the sheep fetus in late gestation, including higher molecular weight forms of ACTH-containing peptides.Reference Ozolins, Antolovich and Browne35, Reference Ross, Bennett, James and McMillen36 It is therefore possible that PCUN and/or PIUN results in an increase in the pituitary secretion of POMC-derived peptides including the higher molecular weight forms of ACTH, rather than ACTH 1–39 alone. It may be the actions of these trophic peptides on the fetal adrenal that explain the greater adrenal cortisol response, which occurs in the PCUN fetus after exogenous corticotropin-releasing hormone stimulation.Reference Edwards and McMillen4
PCUN and PIUN and pituitary POMC expression and adrenal growth
In the present study, the total but not relative weight of the fetal adrenal was significantly higher in the PCUN group when compared with controls at 136–138 days of gestation. Interestingly, total adrenal weight in singleton and twin fetuses in the PIUN group was intermediate between that of the control and PCUN groups at this gestational age. One possibility is that exposure to maternal undernutrition during the preconceptional period has a relatively greater effect on fetal adrenal growth than on the pituitary GR expression, and that not all of the effects of PCUN may be explained by the lower expression of GR in the fetal pituitary. There was, however, an inverse relationship between the pituitary GR expression and either total or relative fetal adrenal weight when the data from all groups were combined, and thus the lower GR expression within the pituitary may play a role along with other intra-adrenal factors in determining the adrenal growth phenotype. It has previously been reported that the total or relative weight of the adrenal was not greater in PCUN fetuses at either 55Reference MacLaughlin, Walker and Kleemann19 or 132 days of gestation,Reference Bloomfield, Oliver and Hawkins7 although we have found that the total but not relative adrenal weight was higher in the PCUN singleton and twin fetuses in the current study at 136–138 days of gestation. The effects of PCUN on fetal adrenal growth may emerge in later gestation when adrenal growth and steroidogenesis are normally upregulated during the prepartum period.
In the present study, while there was no difference in the level of pituitary GR expression between twin and singleton fetuses, the relative weight of the fetal adrenal was greater in twins compared with singleton fetuses in all nutritional treatment groups. Interestingly, we reported that the relative adrenal weight was smaller in twins compared with singletons at 55 days of gestationReference MacLaughlin, Walker and Kleemann19 and this highlights the presence of factors in the twin fetus during late gestation that promote adrenal growth. Despite the relatively greater adrenal mass present in the twin fetus, it has been reported that plasma ACTH concentrations are lower, the prepartum cortisol surge occurs later and adrenocortical responsiveness to ACTH is blunted in twin compared with singleton fetal sheep in late gestation.Reference Edwards and McMillen4–Reference Rumball, Oliver and Thorstensen6 We have suggested that there may be a programmed delay in the prepartum activation of the fetal HPA axis in the twin that acts to protect the twin fetus against the effects of a decreased placental substrate supply, fetal hypoxaemia and hypoglycaemia, characteristics of a twin pregnancy that have been shown to stimulate the fetal HPA axis in late gestation.Reference MacLaughlin and McMillen3, Reference MacLaughlin, Walker and Kleemann19, Reference McMillen, Schwartz, Coulter and Edwards37–Reference Ozolins, Young and McMillen40 Thus, the effects of PCUN on fetal adrenal growth may depend on the differential actions of maternal undernutrition during the pre- and post-conceptional period on both the fetal pituitary and adrenal, as well as on fetal number and the developmental stage at which the adrenal is studied.
Implications
In the present study, we have found that maternal undernutrition imposed during either the periconceptional period (from before until 1 week after conception) or during the first week of the preimplantation period resulted in a decrease in pituitary GR mRNA expression in the fetal pituitary during late gestation. This suggests that a critical period for the effects of maternal undernutrition during the periconceptional period on the development of the pituitary–adrenal axis is during the first week of embryonic life. On the basis of the findings of the present study, we would also suggest that exposure to maternal undernutrition during the period before conception may have a relatively greater effect on fetal adrenal growth than on the pituitary GR expression, and that the GR expression within the pituitary may play a role along with other intra-adrenal factors in determining the adrenal growth phenotype. This is consistent with our findings in a recent study that maternal undernutrition in the periconceptional period resulted in increased adrenal growth and in the cortisol stress response in lambs at 3–4 months of age and that these changes were associated with epigenetic modifications of the paternally imprinted gene, IGF2, in the lamb adrenal.Reference Zhang, Rattanatray and MacLaughlin41 Interestingly, in the present study, we have found no evidence that the decrease in the pituitary GR mRNA expression after exposure to periconceptional or preimplantation undernutrition was associated with hypermethylation of the NGF1-A consensus sequence in the GR promoter. One possibility is that the early nutritional environment may act to alter the epigenetic state of an uncharacterized pituitary-specific GR promoter. The cloned 4.5 kb promoter sequence upstream of the TSS in the sheep GR gene would allow us to further investigate the epigenetic regulation of tissue-specific GR exon 1 promoters in the sheep model. The data from the present study therefore highlight that exposure of the early embryo to maternal undernutrition alters the development of the pituitary–adrenal axis in a manner that would ensure an enhanced response to fetal and postnatal stress perhaps in anticipation of a life of continuing adversity. When this prediction fails, however, as it happens when there is a mismatch between a poor prenatal and abundant postnatal nutritional environment, then the individual is at risk of hypercortisolism, central obesity, hypertension and metabolic disease.Reference Zhang, Rattanatray and MacLaughlin41
Acknowledgments
This study was supported by funding from the Australian Research Council (C.McM., C.T.R., S.K.W.) and from the National Health and Medical Research Council (C.McM.). J.L.M. was funded by the Heart Foundation South Australian Cardiovascular Research Network. The authors gratefully acknowledge the technical assistance provided by Laura O'Carroll, Pamela Sim, Andrew Snell and Bernard Chuang during the course of this study.
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit.
Conflicts of Interest
None.
Ethical Standards
All procedures were approved by the University of Adelaide Animal Ethics Committee.