Introduction
Human and animal studies have demonstrated that individuals who experienced intrauterine growth restriction (IUGR) and low birth weight have an increased risk of metabolic disease in adulthood. Reference Liao, Deng and Zhao1,Reference Christoforou and Sferruzzi-Perri2 The underlying mechanisms for the association between IUGR and metabolic disease remain poorly understood, but one potential mechanism could involve prenatal insults affecting the development of organs important for circadian rhythms and metabolism.
Circadian rhythms are natural cycles that align behaviour and physiology over the 24-h light–dark cycle. Reference Olejniczak, Pilorz and Oster3 These rhythms are regulated by the central ‘master clock’ located in the suprachiasmatic nucleus of the hypothalamus, which coordinates peripheral clocks in almost all organs and tissues. Reference Cox and Takahashi4 Circadian rhythms modulate metabolic activity and play important roles in the functioning of metabolic organs including the liver. Reference Bolshette, Ibrahim, Reinke and Asher5
The structure and function of organs associated with circadian rhythms and metabolism, such as hypothalamus and liver, are determined during development in utero. Reference Hyatt, Budge and Symonds6,Reference Bouret7 In mice, hypothalamic neurogenesis occurs between gestational day (GD) 10 and 16, Reference Shimogori, Lee and Miranda-Angulo8 and hepatoblast differentiation occurs from GD 13.5 to 18.5. Reference Gordillo, Evans and Gouon-Evans9 Adverse prenatal conditions can therefore affect hypothalamic and hepatic development, and subsequently circadian rhythms in adulthood. Reference Joseph, Mamet, Lee, Dalmaz and Van Reeth10 Such circadian disruptions are implicated in the pathophysiology of a variety of diseases including metabolic disorders such as obesity and type II diabetes. Reference Varcoe, Wight, Voultsios, Salkeld and Kennaway11,Reference Fishbein, Knutson and Zee12 In particular, misalignment between the endogenous circadian clocks across multiple organs is associated with impaired glucose tolerance and decreased insulin sensitivity, Reference Morris, Yang and Garcia13,Reference Leproult, Holmbäck and Van Cauter14 similar to the symptoms of developmentally programmed metabolic disease in IUGR offspring. Reference Xie, Lin and Zhang15,Reference Boehmer, Limesand and Rozance16
Genes involved in circadian rhythms, known as clock genes, are rhythmically expressed across the 24 h day as part of a transcription-translation feedback loop (TTFL) in virtually every nucleated cell of the body. Reference Mark, Crew, Wharfe and Waddell17 The positive arm of the circadian TTFL consists of brain and muscle Arnt-like 1 (Bmal1) and circadian locomotor output cycles kaput (Clock Reference Mark, Crew, Wharfe and Waddell17 ). The BMAL1 and CLOCK proteins heterodimerise and induce transcription of period circadian regulators (Per1-3) and cryptochrome circadian regulators (Cry1-2 Reference Mark, Crew, Wharfe and Waddell17 ). Nuclear receptor subfamily 1 group D member 1 (Reverbα) and retinoic acid receptor-related orphan receptor alpha (Rora) repress and stimulate Bmal1 transcription respectively, thereby regulating the TTFL. Reference Mark, Crew, Wharfe and Waddell17 Additionally, the peroxisome proliferator-activated receptors (PPARs) and their heterodimeric partner Pparγ coactivator 1 α (Pgc1α) are key to the bidirectional relationship between circadian rhythms and energy metabolism, and display diurnal rhythmic expression. Reference Yang, Downes and Ruth18,Reference Chen and Yang19 Various adverse conditions in utero are associated with increased Pparα, Pparγ and Pgc1α expression that could contribute to metabolic disease. Reference Myers, Hanson, Mlynarczyk, Kaushal and Ducsay20,Reference Song and Thompson21
Another potential player in the association between adverse prenatal conditions, circadian disruption and metabolic disease is the kisspeptin (Kiss1) signalling pathway in the hypothalamus, and in other metabolic organs including the liver and pancreas. Reference Hudson and Kauffman22 Kisspeptin neurons in the hypothalamus signal via the kisspeptin receptor (Kiss1r) to play key roles in reproduction, energy metabolism and circadian rhythms of behaviour. Reference Hudson and Kauffman22–Reference Yap, Wharfe, Mark, Waddell and Smith25 Adverse prenatal conditions, such as maternal under- or over-nutrition, alter Kiss1 and Kiss1r expression in the hypothalamus and liver, and drive sex-specific changes in glucose and insulin levels. Reference Iwasa, Matsuzaki and Murakami26,Reference Matuszewska, Nowacka-Woszuk and Radziejewska27
The aim of this project was therefore to characterise the impact of mid-gestation maternal hypoxia-induced IUGR on hypothalamic and hepatic clock gene, metabolic gene and Kiss1/Kiss1r gene expression in both male and female adult mice.
Materials and methods
Maternal hypoxia-induced IUGR mouse model and tissue collection
This project was approved by the Telethon Kids Institute (Project Number #264) and was carried out in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (7th Edition).
Pregnant female BALB/c mice were obtained from Animal Resources Centre (Murdoch, WA, Australia) at GD 7 and randomly allocated into Control or IUGR groups. All pregnant mice were housed at 23 ± 1 °C with a 15:9 h light:dark cycle and access to an allergen-free diet (Specialty Feeds, Glen Forrest, WA, Australia) and water ad libitum. Dams in the Control group were housed in normoxic conditions (21% oxygen) throughout gestation. As shown in Fig. 1, dams in the IUGR group were housed in a hypoxic chamber (10.5% oxygen maintained by continuous infusion of excess nitrogen gas) from GD 11 to 17.5 and then returned to normoxic conditions (21% oxygen) until birth at GD ∼ 21. Reference Looi, Kicic, Noble and Wang28,Reference Noble, Kowlessur, Larcombe, Donovan and Wang29 Offspring from dams exposed to hypoxia are referred to as IUGR offspring and offspring from dams exposed to normoxic conditions are referred to as Control offspring.
Experimental timeline. Pregnant BALB/c mice in the Control group were maintained in normoxic conditions (21% O2) throughout gestation. Mice in the IUGR group were exposed to hypoxic conditions (10.5% O2) from GD 11 to GD 17.5. Offspring were raised until euthanised at 8 weeks of age for brain and liver tissue collection. GD, gestational day; IUGR, intrauterine growth restriction; O2, oxygen.
Offspring were weighed at birth. After weaning, offspring were separated by sex to different cages, with each cage housing between 1 and 4. At 8 weeks of age, between 8 - 10 am (2 - 4 h after lights on), one male and one female offspring from each litter (Control male, n = 11; Control female, n = 12; IUGR male, n = 12; IUGR female, n = 11) were euthanised in randomised order with ketamine (240 mg/kg) and xylazine (12 mg/kg) by intraperitoneal injection. Reference Francis, Pinniger, Noble and Wang30 Body weight, abdominal circumference, crown-rump length, liver and brain weight were recorded at post-mortem. Brain tissues were flash frozen with isopentane, liver tissues were snap frozen with liquid nitrogen, and all tissues were stored at −80°C.
RNA extraction
Whole brain samples were dissected to isolate the hypothalamus which was homogenised. Reference Yap, Wharfe, Mark, Waddell and Smith25 Liver sections (approximately 50 mg) were isolated and homogenised. Total RNA for both tissues was extracted using QIAzol Lysis Reagent (Qiagen Pty Ltd., Melbourne, VIC, Australia) according to manufacturer’s instructions and then resuspended in an appropriate quantity of nuclease-free water (approximately 30 µL) and stored at −80°C. Reference Yap, Wharfe, Mark, Waddell and Smith25 The RNA samples were assessed for yield and quality using the Nanodrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA).
Reverse transcription
The RNA was DNAse treated (Promega, Madison, WI, USA) to remove any genomic DNA contamination. Hypothalamic RNA (1 µg) was reverse transcribed to cDNA using the QuantiTect Reverse Transcription Kit (Qiagen). Liver RNA (5 µg) was reverse transcribed to cDNA using Moloney Murine Leukemia Virus reverse transcriptase (Promega), purified using the QIAquick Column PCR Purification Kit (Qiagen) and eluted into nuclease-free water. Reference Yap, Wharfe, Mark, Waddell and Smith25
Real-time polymerase chain reaction (PCR)
Analyses of relative mRNA expression levels for Bmal1, Per2, Reverbα, Pparα, Pparγ, Pgc1α, Kiss1 and Kiss1r were performed through standard quantitative polymerase chain reaction (PCR) on the Rotorgene 6000 (Qiagen) using QuantiNova SYBR Green PCR Master Mix (Qiagen). Primer sequences are listed in Table 1 with Pparα and Pparγ designed using Primer-BLAST (National Centre for Biotechnology Information, Bethesda, MD, USA). The PCR cycling conditions were as follows: an initial denaturation step of 95 °C for 2 min, followed by at least 40 cycles of denaturation at 95 °C for 5 s and annealing of primers for 15 s . Melt curve analysis and DNA sequencing were performed to confirm amplification specificity.
Forward and reverse primer sequences for real-time PCR
Bmal1, brain and muscle Arnt-like 1; F, forward; Hprt, hypoxanthine-guanine phosphoribosyltransferase; Kiss1, kisspeptin; Kiss1r, kisspeptin receptor; PCR, polymerase chain reaction; Per2, period circadian regulator 2; Pgc1α, peroxisome proliferator-activated receptor γ coactivation 1 α; Ppar, peroxisome proliferator-activated receptor; Ppia, peptidylprolyl isomerase A; Rev-erbα, nuclear receptor subfamily 1 group D member 1; R, reverse; Sdha, succinate dehydrogenase complex subunit A; Tbp, TATA box binding protein.
Relative mRNA expression levels of clock (Bmal1, Per2, Reverbα), metabolic (Pparα, Pparγ, Pgc1α) and kisspeptin (Kiss1 and Kiss1r) were measured in both the liver and hypothalamus samples. Expression of the target genes in each sample was normalised using geometric means calculated from the expression of reference genes: hypoxanthine phosphoribosyltransferase (Hprt), peptidylprolyl isomerase A (Ppia), succinate dehydrogenase complex subunit A (Sdha) and TATA box binding protein (Tbp) in the liver; and Hprt, Ppia and Tbp in the hypothalamus. Reference Vandesompele, De Preter and Pattyn35
Statistical analysis
Data were transformed where necessary to ensure that the assumptions of normality and homoscedasticity of variances for the parametric tests were satisfied. Birth weight was compared using a t-test to determine the effect of treatment (Control vs IUGR). Body characteristics and gene expression in the hypothalamus and liver were compared using two-way ANOVAs to determine the effect of treatment (Control vs IUGR) and sex (male vs female) on mRNA expression of Bmal1, Per2, Reverbα, Pparα, Pparγ, Pgc1α, Kiss1 and Kiss1r. Data which do not fit a normal distribution (body weight at 8 weeks, hepatic Bmal1 expression and hypothalamic Per2 expression) were compared using two Kruskal–Wallis one-way ANOVAs. All normally distributed data are presented as mean ± SEM, data which do not fit a normal distribution are displayed as median ± interquartile range, with p-values < 0.05 considered statistically significant for all analyses. All statistical analysis was performed in SigmaPlot (Version 14.5, Grafiti LLC, Palo Alto, CA, USA). Graphs were generated using GraphPad PRISM (Version 10.3, GraphPad Software, Boston, MA, USA).
Results
Growth outcomes
Following mid-gestation hypoxia, the IUGR offspring had lower birth weight compared with Control offspring (Control (n = 57), 1.45 ± 0.02 g; IUGR (n = 49), 1.37 ± 0.03 g, p = 0.047). There was no significant difference in litter size between IUGR and Control offspring (Control (n = 16), 3.56 ± 0.38 offspring per litter; IUGR (n = 15), 3.27 ± 0.25 offspring per litter; p = 0.522).
At 8 weeks of age, IUGR offspring remained smaller than Control offspring, with lower body weight (p < 0.001, Table 2) and shorter crown-rump length (p < 0.001, Table 2), but similar abdominal circumference between groups (p = 0.722, Table 2). At 8 weeks of age, female offspring had lower body weight (p < 0.001, Table 2) and shorter crown-rump length (p = 0.023, Table 2) than male offspring, but abdominal circumference was comparable between sexes (p = 0.098, Table 2). There was no interaction between treatment and sex in crown-rump length (p = 0.662, Table 2) or abdominal circumference (p = 0.757, Table 2).
Body characteristics at 8 weeks of age
Body weight data are median ± interquartile range, crown-rump length and abdominal circumference data are mean ± SEM.
* Denotes significantly different compared with Control (p < 0.05, two-way ANOVA or Kruskal–Wallis one-way ANOVA for body weight).
# Denotes significantly different compared with males (p < 0.05, two-way ANOVA or Kruskal–Wallis one-way ANOVA for body weight). IUGR, intrauterine growth restriction.
At 8 weeks of age, absolute brain weight was similar between IUGR and Control offspring (p = 0.141, Table 3), but IUGR offspring had greater brain weight relative to body weight than Control offspring (p = 0.015, Table 3). There was no effect of sex on absolute brain weight (p = 0.810, Table 3), but female offspring had greater brain weight relative to body weight than male offspring (p < 0.001, Table 3). At 8 weeks of age, IUGR offspring had lower absolute liver weight than Controls (p = 0.016, Table 3), but no difference in liver weight relative to body weight between the groups (p = 0.874; Table 3). Female offspring had lower absolute liver weight than male offspring (p < 0.001, Table 3) and also lower liver weight relative to body weight than male offspring (p < 0.001, Table 3). There was no interaction between treatment and sex in absolute brain weight (p = 0.115, Table 3), brain relative to body weight (p = 0.203, Table 3), absolute liver weight (p = 0.695, Table 3) or liver relative to body weight (p = 0.785, Table 3).
Brain and liver weights at 8 weeks of age
Data are mean ± SEM.
* Denotes significantly different compared with Control (p < 0.05, two-way ANOVA).
# Denotes significantly different compared with males (p < 0.05, two-way ANOVA). IUGR, intrauterine growth restriction.
Clock gene expression
Expression of the core clock gene Bmal1 was lower in in the hypothalamus of IUGR offspring compared with Control (p < 0.001, Fig. 2A), but there was no difference in hepatic Bmal1 expression between the groups (p = 0.668, Fig. 2B). There was no sex effect on Bmal1 expression in either the hypothalamus (p = 0.317, Fig. 2A) or liver (p = 0.057, Fig. 2B). There was no interaction of sex and treatment on Bmal1 expression in the hypothalamus (p = 0.176, Fig. 2A).
Relative clock gene expression in mouse hypothalamus and liver. Gene expression of hypothalamic Bmal1 (A), hepatic Bmal1 (B), hypothalamic Per2 (C), hepatic Per2 (D), hypothalamic Reverbα (E) and hepatic Reverbα (F). * denotes significantly different compared with Control (p < 0.05, two-way ANOVA or Kruskal–Wallis one-way ANOVA for hepatic Bmal1 and hypothalamic Per2 expression); # denotes significantly different compared with males (p < 0.05, two-way ANOVA or Kruskal–Wallis one-way ANOVA for hepatic Bmal1 and hypothalamic Per2 expression). All data are mean ± SEM except hepatic Bmal1 (B) and hypothalamic Per2 (C) expression are median ± interquartile range. Open circles, Control males; open squares, Control females; closed circles, IUGR males; closed squares, IUGR females. Bmal1, brain and muscle Arnt-like 1; IUGR, intrauterine growth restriction; Per2, period circadian regulator 2; Reverbα, nuclear receptor subfamily 1 group D member 1.
Hypothalamic Per2 expression was comparable between treatment groups (p = 0.565, Fig. 2C) and sexes (p = 0.991, Fig. 2C). In the liver, there was no difference in Per2 expression between IUGR and Control offspring (p = 0.214, Fig. 2D), but female offspring had higher hepatic Per2 expression than male offspring (p = 0.005, Fig. 2D). There was no interaction between sex and treatment on Per2 expression in the liver (p = 0.605, Fig. 2D).
Hypothalamic Reverbα expression was lower in IUGR compared with Control offspring (p = 0.047, Fig. 2E), but hepatic Reverbα expression was similar between groups (p = 0.532, Fig. 2F). In both the hypothalamus and liver, Reverbα expression was similar between sexes (hypothalamus, p = 0.509, Fig. 2E; liver, p = 0.084, Fig. 2F). There was no interaction between sex and treatment on Reverbα expression in either the hypothalamus (p = 0.606, Fig. 2E) or liver (p = 0.686, Fig. 2F).
Metabolic gene expression
There were no differences in Pparα expression between Control and IUGR offspring in either the hypothalamus or liver (hypothalamus, p = 0.415, Fig. 3A; liver, p = 0.066, Fig. 3B). Similarly, hypothalamic and hepatic Pparγ expression were also comparable between IUGR and Control offspring (hypothalamus, p = 0.081, Fig. 3C; liver, p = 0.623, Fig. 3D). Further, there were no changes in hypothalamic or hepatic Pgc1α expression between Control and IUGR offspring (hypothalamus, p = 0.096, Fig. 3E; liver, p = 0.213, Fig. 3F).
Relative metabolic gene expression in the hypothalamus and liver. Gene expression of hypothalamic Pparα (A), hepatic Pparα (B), hypothalamic Pparγ (C), hepatic Pparγ (D), hypothalamic Pgc1α (E) and hepatic Pgc1α (F). # denotes significantly different compared with males (p < 0.05, two-way ANOVA). Data are mean ± SEM. Open circles, Control males; open squares, Control females; closed circles, IUGR males; closed squares, IUGR females. IUGR, intrauterine growth restriction; Pgc1α, peroxisome proliferator-activated receptor γ coactivation 1α; Ppar, peroxisome proliferator-activated receptor.
Metabolic gene expression was also similar between sexes in the hypothalamus (Pparα, p = 0.754, Fig. 3A; Pparγ, p = 0.977, Fig. 3C; Pgc1α, p = 0.855, Fig. 3E) and in Pparγ expression in the liver (p = 0.236, Fig. 3D). However, female offspring had higher hepatic expression of Pparα (p < 0.001, Fig. 3B) and Pgc1α (p = 0.010, Fig. 3F) than male offspring, irrespective of treatment.
There was no interaction between treatment and sex in any of the metabolic genes in the hypothalamus (Pparα, p = 0.727, Fig. 3A; Pparγ, p = 0.057, Fig. 3C; Pgc1α, p = 0.766, Fig. 3E) or liver (Pparα, p = 0.537, Fig. 3B; Pparγ, p = 0.736, Fig. 3D; Pgc1α, p = 0.843, Fig. 3F).
Kisspeptin gene expression
There was no difference in hypothalamic Kiss1 expression between IUGR and Control offspring (p = 0.220, Fig. 4A), but female offspring had increased hypothalamic Kiss1 expression compared with male offspring (p < 0.001, Fig. 4A). There was no interaction between treatment and sex on hypothalamic Kiss1 expression (p = 0.145, Fig. 4A). The IUGR offspring had decreased hypothalamic Kiss1r expression compared with Control offspring (p = 0.018, Fig. 4B), and male offspring had increased hypothalamic Kiss1r expression compared with female offspring (p = 0.003, Fig. 4B). There was no interaction between treatment and sex on hypothalamic Kiss1r expression (p = 0.128, Fig. 4B). In the liver samples, Kiss1 and Kiss1r expression were too low to obtain quantifiable data.
Relative kisspeptin gene expression in the hypothalamus. gene expression of hypothalamic Kiss1 (A) and hypothalamic Kiss1r (B).* denotes significantly different compared with Control (p < 0.05, two-way ANOVA); # denotes significantly different compared with males (p<0.05, two-way ANOVA). Data are mean ± SEM. Open circles, Control males; open squares, Control females; closed circles, IUGR males; closed squares, IUGR females. Kiss1, kisspeptin; Kiss1r, kisspeptin receptor.
Discussion
With the interest of understanding the mechanism underlying the strong associations found between IUGR and the development of metabolic disease in later life, Reference Cianfarani, Agostoni and Bedogni36 we proposed that mid-gestation maternal hypoxia-induced IUGR may disrupt circadian rhythms and Kiss1/Kiss1r signalling leading to developmentally programmed metabolic disease. Our main finding that hypothalamic, but not hepatic, Bmal1 and Reverbα expression were decreased in IUGR offspring from hypoxic pregnancies compared with Controls suggests dysregulation between the hypothalamic and hepatic circadian clocks in IUGR offspring. Furthermore, IUGR offspring had decreased hypothalamic Kiss1r expression, possibly indicating disrupted Kiss1/Kiss1r signalling in the hypothalamus and therefore altered hypothalamic control over metabolism and circadian rhythms.
Intrauterine programming of adulthood outcomes occurs during key windows of developmental plasticity, and the timing and duration of a prenatal insult can alter its long-term effects. Reference Song and Thompson21,Reference Deodati, Inzaghi and Cianfarani37 The IUGR mouse model used in this study, mid-gestation maternal hypoxia-induced IUGR, targeted key developmental windows for the liver and hypothalamus. Reference Shimogori, Lee and Miranda-Angulo8,Reference Gordillo, Evans and Gouon-Evans9 Consistent with previous publications, IUGR offspring were smaller at birth and 8 weeks of age compared with Control offspring. Reference Looi, Kicic, Noble and Wang28–Reference Francis, Pinniger, Noble and Wang30,Reference Wang, Larcombe and Berry38,Reference Kalotas, Wang, Noble and Wang39 Of relevance to this study, we reported that the IUGR offspring had greater brain relative to body weight at 8 weeks of age, consistent with asymmetric IUGR. Reference Sharma, Shastri and Sharma40 Asymmetric IUGR refers to nonuniform growth restriction wherein brain volume is spared at the expense of the viscera, caused by preferential redirection of energy to the brain when substrate delivery to a fetus is restricted in the latter half of gestation. Reference Cahill, Hoggarth, Lerch, Seed, Macgowan and Sled41 The IUGR offspring had smaller absolute liver weights at 8 weeks of age, but there was no difference in liver relative to body weight between Control and IUGR offspring. Asymmetric IUGR is commonly associated with decreased liver weight relative to body weight Reference Cahill, Hoggarth, Lerch, Seed, Macgowan and Sled41 ; however, decreased liver relative to body size in IUGR animals can resolve by early adulthood, potentially due to catch-up growth. Reference Iqbal and Ciriello42,Reference Morrison, Duffield, Muhlhausler, Gentili and McMillen43 Therefore, any decreased liver weight in the present study was likely resolved by the 8-week time point of tissue collection. Similar to previous findings that we reported in respiratory studies, Reference Looi, Kicic, Noble and Wang28–Reference Francis, Pinniger, Noble and Wang30,Reference Wang, Larcombe and Berry38,Reference Kalotas, Wang, Noble and Wang39 we observed sex-dependent effects whereby somatic and liver growth outcomes were reduced in female offspring compared with male offspring. In sum, growth parameters in IUGR compared with Control offspring were consistent with asymmetric restricted growth in utero, suggesting that hypothalamic and hepatic development in utero were likely affected by maternal hypoxia.
Exposure to adverse prenatal conditions, including hypoxic conditions, can disrupt postnatal circadian rhythms. Reference Varcoe, Wight, Voultsios, Salkeld and Kennaway11,Reference Koritala, Lee and Bhadri44–Reference Sutton, Centanni and Butler46 In the present study, hypothalamic Bmal1 and Reverbα expression were lower in IUGR offspring compared with Control offspring, although hypothalamic Per2 expression was unaffected by prenatal hypoxia at this time of collection (i.e., around 2 - 4 h after lights on). Altered expression of only two of the three clock genes was unexpected as the components of the core circadian TTFL interact to regulate each other’s expression, and changes in Bmal1 expression are generally reflected in both Per2 and Reverbα. Reference Sutton, Centanni and Butler46,Reference Meng, McMaster and Beesley47 It is possible that Per2 expression is impacted at other times of day, or changes are obscured by concomitant changes in amplitude, mesor and/or acrophase of Per2 expression. A decrease in both Bmal1 and Reverbα expression was also unexpected, given that these clock genes have opposing patterns of rhythmic oscillation whereby Bmal1 expression peaks at the end of the dark phase, while Reverbα expression peaks in the middle of the light phase and Reverbα inhibits Bmal1 expression. Reference Mark, Crew, Wharfe and Waddell17,Reference Yang, Downes and Ruth18 The decrease in both Bmal1 and Reverbα therefore suggests a potential disarrangement of the circadian TTFL in the hypothalamus of IUGR offspring, which would likely be associated with both desynchronisation of endogenous circadian clocks and metabolic disease. Reference Mazzoccoli, Pazienza and Vinciguerra48
Although relatively subtle, significant changes in hypothalamic clock gene rhythmicity were observed in IUGR offspring at 8 weeks of age. The possibility remains that these programmed changes may be exacerbated as offspring age, or through metabolic challenges such as consumption of high energy diets. Reference Sutton, Centanni and Butler46,Reference Rueda-Clausen, Dolinsky, Morton, Proctor, Dyck and Davidge49 More comprehensive, circadian studies are necessary to investigate whether the decreased Bmal1 and Reverbα expression in IUGR offspring was due to changes in amplitude, mesor, acrophase of expression, or combinations thereof. Gene expression of other components of the core circadian TTFL including Clock and Cry should be investigated in future studies.
There were no differences in hepatic clock gene expression of IUGR and Control offspring at this single time of day (around 2 - 4 h after lights on), however changes in circadian networks could still be occurring at other times of day. Independent of treatment group, female offspring had higher hepatic Per2 expression than male offspring; this sexual dimorphism was likely caused by lower amplitude of Per2 rhythmic expression in the female compared with male liver. Reference Kuljis, Loh and Truong50 These results suggest that the hepatic clock was not affected by mid-gestation maternal hypoxia-induced IUGR, potentially due to lower susceptibility to hypoxia as a circadian disruptor. Reference Tahara and Shibata51 Hypoxia exerts tissue-specific effects and, consistent with the present study, causes more pronounced disturbances to rhythmic clock gene expression in the brain than the liver. Reference Koritala, Lee and Bhadri44,Reference Manella, Aviram and Bolshette52 Interestingly, IUGR induced by a maternal protein-deficient diet from two weeks before mating and throughout pregnancy disrupts hepatic and cortical (Bmal1, Clock, Per2 and Reverbα), but not hypothalamic, clock gene expression in 8-week-old mouse offspring, which may be due to differences in the type and timing of the prenatal insult between mouse models. Reference Sutton, Centanni and Butler46
Peripheral circadian clocks can function independently from the central clock, so disruptions to the central clock do not necessarily disrupt the liver. Reference Kolbe, Leinweber, Brandenburger and Oster53 Tissue-specific change in clock gene expression suggests that the hypothalamic and hepatic circadian clocks may be misaligned in young adult IUGR offspring. Mice with misalignment between their central and peripheral circadian clocks in different tissues develop metabolic symptoms including impaired glucose tolerance, hypoinsulinemia and hyperglycaemia. Reference Kolbe, Leinweber, Brandenburger and Oster53,Reference Mukherji, Kobiita and Damara54 Therefore interventions aimed at restoring circadian alignment may provide a potential treatment for developmentally programmed metabolic disease in adulthood. Reference Sulli, Manoogian, Taub and Panda55
The Pparα, Pparγ and Pgc1α genes are recognised as links between the circadian system and energy metabolism and are sensitive to varied prenatal insults. Reference Brunton, Sullivan, Kerrigan, Russell, Seckl and Drake56–Reference Zheng, Xiao, Zhang, Yu, Xu and Wang59 To our best knowledge, the present study is the first to investigate hypothalamic and hepatic Pparα, Pparγ and Pgc1α expression using this mouse model of mid-gestation maternal hypoxia-induced IUGR. The lack of change in metabolic gene expression in IUGR offspring may be due to the specific nature, timing or duration of prenatal insult. Moreover, like circadian dysregulation, metabolic symptoms programmed in utero may be exacerbated as offspring age or when offspring face metabolic challenges such as a high fat diet. Reference Iqbal and Ciriello42,Reference Erhuma, Salter, Sculley, Langley-Evans and Bennett57,Reference Darendeliler60 At 8 weeks old, the young adult mice in this study may have been too young to display changes in Pparα, Pparγ or Pgc1α expression. Rhythmic expression of Pparα, Pparγ and Pgc1α could also have obscured changes in gene expression as tissues were only collected at a single time point (8 - 10 am, i.e., early in the light phase) when hepatic Pparα and Pparγ expression is low. Reference Yang, Downes and Ruth18 Metabolic genes, including others such as Rorα and Pparδ, should be assessed in other metabolic tissues such as the pancreas or skeletal muscle to understand how these tissues may be differentially affected by adverse prenatal conditions and interact to contribute to metabolic disease.
The higher hepatic Pparα and Pgc1α expression in female offspring was likely caused by sexual dimorphism in overall rhythmic expression of these genes, that is, higher mesor in the female compared with male liver. Reference Yang, Zhang, Esterly, Klaassen and Wan61 This sexual dimorphism may arise because Pparα protects against oestrogen-mediated hepatotoxicity, Reference Leuenberger, Pradervand and Wahli62 and Pgc1α is required for oestrogen-dependent reactive oxygen species detoxification in the liver. Reference Besse-Patin, Léveillé, Oropeza, Nguyen, Prat and Estall63
Adverse prenatal conditions can program age- and sex-specific changes in postnatal Kiss1 and Kiss1r expression, Reference Matuszewska, Nowacka-Woszuk and Radziejewska27,Reference Minabe, Iwata, Watanabe, Ishii and Ozawa64 for example, prenatal undernutrition has been shown to decrease hypothalamic Kiss1 at juvenile (2-3 weeks in rats) and mature (29 weeks in rats) ages. Reference Iwasa, Matsuzaki and Murakami26,Reference Minabe, Iwata, Watanabe, Ishii and Ozawa64 Though hypothalamic Kiss1 expression was similar in IUGR and Control offspring at 8 weeks of age, there was a prenatal hypoxia-driven decrease in hypothalamic Kiss1r expression of IUGR compared with Control offspring. Kisspeptin signalling via Kiss1r modulates the activity of orexigenic and anorexigenic neurons in the hypothalamus, thus Kiss1r is key to regulating hunger/satiety and energy balance. Reference Navarro23,Reference Fang, She and Zhao65 Decreased hypothalamic Kiss1r expression in IUGR offspring may therefore represent a deficit in central regulation of glucose homoeostasis and may contribute to developmentally programmed metabolic disease.
We also observed sexual dimorphism in hypothalamic Kiss1 and Kiss1r expression. Higher hypothalamic Kiss1 expression in female offspring may reflect the greater number of kisspeptin neurons found in the female compared with male hypothalamus. Reference Kauffman, Gottsch and Roa66 The Kiss1 neurons in the anteroventral periventricular nucleus of the hypothalamus are involved in generating the preovulatory gonadotropin-releasing hormone/luteinising hormone surge, which only occurs in females. Reference Lee, Dilower and Marsh67,Reference Smith, Popa, Clifton, Hoffman and Steiner68 Lower hypothalamic Kiss1r expression in female compared with male offspring is consistent with previous evidence that female rats have lower hypothalamic Kiss1r expression than male rats at 45 days of age and evidence of a more pronounced effect of Kiss1r knockout on male sexual differentiation than female. Reference Navarro, Castellano and Fernandez-Fernandez69,Reference Kauffman, Park and McPhie-Lalmansingh70 Moreover, a rise in Kiss1 mRNA in female offspring (and potentially kisspeptin expression) could be the stimulus for the decline in Kiss1r expression seen in our studies. Kisspeptin-induced desensitisation of Kiss1r signalling has been seen in rats, Reference Roa, Vigo and Garcia-Galiano71 and specifically results in a decline in Kiss1r mRNA expression in sheep. Reference Li, Roa, Clarke and Smith72 Previous studies have reported that both Kiss1 and Kiss1r are expressed only at low levels in liver tissue, Reference Kotani, Detheux, Vandenbogaerde, Communi, Vanderwinden, Le Poul, Brézillon, Tyldesley, Suarez-Huerta, Vandeput, Blanpain, Schiffmann, Vassart and Parmentier73,Reference Ohtaki, Shintani, Honda, Matsumoto, Hori, Kanehashi, Terao, Kumano, Takatsu, Masuda, Ishibashi, Watanabe, Asada, Yamada, Suenaga, Kitada, Usuki, Kurokawa, Onda, Nishimura and Fujino74 which may explain why we were unable to quantify hepatic Kiss1 or Kiss1r expression.
In summary, our data align with our proposal that mid-gestation maternal hypoxia-induced IUGR can program disrupted circadian rhythms and Kiss1/Kiss1r signalling, which may subsequently result in developmentally programmed metabolic disease. The present study focused on mRNA expression as a reflection of developmentally programmed changes in gene expression following adverse prenatal conditions. Reference Lillycrop, Phillips, Jackson, Hanson and Burdge75–Reference Safi-Stibler and Gabory77 With these findings, future studies could incorporate metabolic parameters in vivo such as intraperitoneal glucose tolerance testing to confirm symptoms of metabolic disease, measure protein abundance in addition to mRNA data, or employ a transcriptomic approach to potentially identify a wider range of genes developmentally programmed by mid-gestation maternal hypoxia. Further research could also measure gene expression across multiple circadian time points, at an older age (e.g., 6 months old in mice) or after a metabolic challenge to investigate the onset and persistence of IUGR-related changes in clock genes and metabolism.
Acknowledgements
We thank Dr. Darshinee Kowlessur for her assistance in animal handling and monitoring.
Financial support
This project was funded by the National Health and Medical Research Council (NHMRC) of Australia Project Grant 1120128 (K.C.W.W.). K.C.W.W. was supported by the Western Australian Future Health Research and Innovation Fund, which is an initiative of the Western Australian State Government. A.E.O was supported by The University of Western Australia School of Human Sciences Dr Margaret Loman-Hall Scholarship.
Competing interests
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national guides on the care and use of laboratory animals (Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (7th Edition)) and has been approved by the institutional committee (Telethon Kids Institute Animal Ethics Committee (Project Number #264)).
Open access
