Epigenetic Changes in the Hypothalamus of Offspring Following Maternal Undernutrition

Epigenetic changes in the hypothalamus of offspring following maternal undernutrition

A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Life Sciences

Ghazala Begum September 2013

Contents

Page No. List of Figures 5 List of Tables 8 Abstract 9 Declaration 10 Copyright Statement 10 Abbreviations 11 Nomenclature 13 Acknowledgements 14

Chapter 1 15 Introduction 15 1.1 The implications of maternal programming 16 1.1.1 Maternal undernutrition induces programming 17 1.1.2 Maternal overnutrition induces programming 18 1.1.3 Stress-induced maternal programming 19 1.1.4 Twinning as a programming paradigm 20 1.2 Programming of the hypothalamic energy regulating pathway 21 1.2.1 Hypothalamic energy regulating pathway 21 1.2.2 Role of glucocorticoids in the hypothalamic energy regulating pathway 26 1.2.3 Effects of maternal undernutrition on the offspring’s energy regulating 27 athway pathway 1.3 Programming of the HPA axis 31 1.3.1 Key components of the HPA axis 31 1.3.2 Effects of maternal undernutrition on the offspring’s HPA axis 32 1.4 The potential involvement of epigenetics in programming 34 1.4.1 DNA methylation 35 1.4.2 Histone Modifications 37 1.4.3 The GR gene as a potential target for epigenetic modification 38 1.4.4 Epigenetic alterations of fetal GR following maternal programming 39 1.4.5 POMC as a potential target of epigenetic modification 40 1.4.6 Epigenetic alterations of fetal POMC following maternal nutritional 40 dsjdsddsdinsults 1.5 Overview 42 1.5.1 Maternal undernutrition sheep model 42 1.5.2 Aims 44 1.6 Alternative format 47

Chapter 2 49 Methods 49 2.1 Animal Management 50 2.2 DNA, RNA and protein purification from brain tissues 51 2.3 Whole blood DNA and protein purification 52 2.4 Bioinformatic analysis 53 2.5 DNA methylation analysis 53 2.6 mRNA expression analysis 55 2.6.1 mRNA expression analysis during the fetal study 55

2.6.2 mRNA expression analysis during the twin study 55 2.6.3 mRNA expression analysis during the adult study 56 2.7 Chromatin immunoprecipitation 57 2.7.1 Chromatin immunoprecipication for fetal tissues 57 2.7.2 Chromatin immunoprecipitation for adult brain tissues 58 2.8 Western blots 59 2.9 POMC ELISA 60 2.10 DNA methyltransferase activity/inhibition assay 61 2.11 Buffers 61 2.12 Statistical analysis 62

Chapter 3: Publication I 63 Epigenetic changes in the hypothalamic pro-opiomelanocrotin and glucocorticoid ssreceptor genes in the ovine fetus after periconceptional undernutrition

Chapter 4: Publication II 64 Epigenetic changes in fetal hypothalamic energy regulating pathways are dlassociated with maternal undernutrition and twinning 4.1 Supplemental data 65

Chapter 5: Publication III 67 Maternal undernutrition programs tissue-specific epigenetic changes in the bglucocorticoid receptor in adult offspring

Chapter 6 68 Maternal undernutrition induces tissue specific changes in POMC in adult 68 gioffspring 6.1 Introduction 69 6.1.1 Aims 70 6.2 Materials and methods 71 6.2.1 Animal management 71 6.2.2 Isolation of DNA and RNA and protein from tissue and blood 71 6.2.3 DNA methylation enrichment 71 6.2.4 Chromatin immunoprecipitation of H3K9AC and H3K27me3 71 6.2.5 qRT-PCR analysis 72 6.2.6 Western blot 72 6.2.7 Statistical analysis 72 6.3 Epigenetic changes in hypothalamic POMC 73 6.4 Comparable levels of hypothalamic POMC protein 75 6.5 Epigenetic changes in the POMC promoter amplicon in the pituitary of 76 fddffdpericonceptionally undernourished adult offspring 6.6 POMC mRNA and protein levels in the pituitary of adult offspring 77 6.7 POMC status in the peripheral leukocytes of maternally undernourished adult 78 vcvccoffspring 6.8 Summary 79

Chapter 7 81 Discussion 7.1 The impact of the length and timing of maternal undernutrition on the 83

3 hghghoffspring in sheep 7.2 Epigenetic changes in hypothalamic GR as a consequence of maternal 84 ghghgprogramming persisting from fetal life to adulthood 7.3 Epigenetic changes in hypothalamic POMC as a consequence of maternal 87 ghghhprogramming 7.4 Tissue specific changes in GR and POMC 90 7.5 NPY expression in the hypothalami of maternally underfed offspring 92 7.6 Sex-specific outcomes in the maternally undernourished offspring 93 7.7 Twinning as an intrauterine programming paradigm 94 7.8 Phenotypic outcome in the adult offspring subject to periconceptional 95 cvcvcundernourishment 7.9 Mechanisms of action underlying epigenetic changes in the offspring as a 97 cvcvvconsequence of maternal programming 7.10 Conclusion 98 7.11 Future work 99 7.11.1 Analysis of the role of exon 1 of the GR gene in the hypothalamus 99 7.11.2 POMC hypothalamic specific enhancer region 102 7.11.3 Candidate gene approach verses genome wide analysis 103 7.11.4 Potential models of maternal nutritional programming 104

Chapter 8 107 References

Chapter 9 128 Appendix

Final word count is 52,545.

List of Figures

Page No. Page No. in thesis in paper Chapter 1: Figure 1.01. Programming paradigms 17

Figure 1.02. Tissue specific POMC processing 22

Figure 1.03. Hypothalamic appetite regulatory pathway 25

Figure 1.04. The HPA axis 31

Figure 1.05. DNA methylation 37

Figure 1.06. Histone modifications 38

Figure 1.07. The GR exon 1 promoter region 39

Figure 1.08: Maternal programming sheep model 43

Chapter 2: Figure 2.01. RNA integrity gel 52

Figure 2.02. RNA integrity of whole and ventral hypothalamic 52 samples

Chapter 3: FIG. 1. POMC and GR gene region screening to identify highly 3656 conserved, CpG-rich regions.

FIG. 2. HPA axis activity in fetal sheep from control ewes or 3657 ewes subjected to periconceptional undernutrition (underfed from 60 d before conception to 30 d after conception).

FIG. 3. Epigenetic changes associated with the POMC gene in 3658 the fetal hypothalamus. Fetal hypothalamic tissue samples were obtained from normal and underfed maternal sheep (underfed from 60 d before to 30 d after conception).

FIG. 4. Expression of the POMC gene in the fetal hypothalamus. 3658 Fetal hypothalamic tissue samples were obtained from control and underfed (from 60 d before conception to 30 d after conception).

FIG. 5. Presence of H3K9Ac associated with the GR gene 3659 promoter in the fetal hypothalamus

Page No. Page No. in thesis in paper FIG. 6. Methylation of GR gene promoter region in the fetal 3659 hypothalamus. Hypothalamic tissue samples were obtained from control and underfed (from 60 d before conception to 30 d after conception) fetal sheep.

FIG. 7. Expression of GR gene in the fetal hypothalamus. Fetal 3660 hypothalamic tissue samples were obtained from control and underfed maternal sheep (from 60 d before conception to 30 d after conception); 2 μg of total RNA were used to quantify expression levels of GR.

FIG. 8.The effect of different periods of periconceptional 3661 undernutrition on hypothalamic GR, POMC, and NPY methylation and gene expression. The −60 to +30 group (UN −60 to +30) was underfed from 60 d before conception to 30 d after conception. The −60 to 0 group (UN −60 to 0) were fed the same diet as the −60 to +30 group but were allowed to feed ad libitum from conception. The −2 to +30 group (UN −2 to +30) were fed the same diet for 30 d after conception. Fetal ventral hypothalamic tissue samples, enriched for the ARC, were obtained from all sample groups at d 131 of gestation

Chapter 4: Figure 1.Summary of ENCODE data from different human cell 1696 lines. Data analysis from the UCSC genome browser depicts cell line-specific changes in chromatin over the POMC promoter marker region.

Figure 2.Histone modifications of the POMC promoter in 1698 response to twinning and periconceptional maternal undernutrition.

Figure 3.DNA methylation and expression levels of fetal 1699 hypothalamic neuropeptides

Figure 4.Changes in the histone patterns of the GR promoter as a 1700 result of twinning and maternal periconceptional undernutrition.

Figure 5.Effects of twinning and periconceptional maternal 1700 undernutrition on fetal hypothalamic GR promoter methylation and GR mRNA expression levels.

Figure 6.Fetal HPA axis dynamics following twinning and 1701 maternal periconceptional undernutrition.

4.1 Supplemental data: Fig. S1. Summary of ENCODE data 65

Page No. Page No. in thesis in paper Fig. S2. NPY mRNA expression levels determined using qRT- 66 PCR. Groups were compared to the control via the one-way ANOVA with Tukey HSD Post Hoc test. *p<0.05, **p<0.01, ***p<0.005.

Chapter 5: Figure 1. Periconceptional undernutrition is associated with 4 altered glucocorticoid receptor (GR) epigenetic status in the ventral hypothalamus of adult offspring.

Figure 2. Altered glucocorticoid receptor (GR) expression levels 5 in the ventral hypothalamus of adult offspring following periconceptional undernutrition.

Figure 3. Ventral hypothalamic neuropeptide mRNA expression. 6

Figure 4. Altered epigenetic and expression status of 6 hippocampal glucocorticoid receptor (GR) epigenetic and expression status in offspring from undernourished mothers.

Figure 5. No change in glucocorticoid receptor (GR) promoter 7 methylation or protein expression in leukocytes.

Figure 6. Increased fat:lean mass ratio in adult males whose 8 mothers were periconceptionally undernourished

Chapter 6: Figure 6.01. Altered epigenetic status of the POMC gene in the 74 ventral hypothalamus of maternally undernourished adult offspring.

Figure 6.02. Similar levels of hypothalamic POMC protein in 75 adult animals following periconceptional undernutrition.

Figure 6.03. Gender specific epigenetic changes in the POMC 76 promoter region in the pituitary in maternally undernourished adult offspring.

Figure 6.04. Comparable POMC mRNA and protein levels in the 77 pituitary.

Figure 6.05. POMC promoter methylation and protein 78 expression in peripheral circulating leukocytes

List of Tables

Page No. Page No. in thesis in paper Introduction:

Table 1.01: Examples of hypothalamic neurotransmitters, amines and 23 amino acids and their implications on the regulation of food intake

Table 1.02: The effects of maternal undernutrition on the offspring. 28

Table 1.03: The effects of maternal overnutrition on the offspring. 29

Table 1.04: Changes observed in the maternally undernourished 44 offspring.

Chapter 2:

Table 2.01: Primers used for methylation analysis 54

Table 2.02: qRT-PCR primers utilised in the twin study. 55

Table 2.03: Preparation mastermix for qRT-PCR reactions. 56

Table 2.04: qRT-PCR primers utilised in the adult study. 57

Table 2.05: List of primary antibodies used for western blotting. 60

4.1 Supplemental data:

Table S1. Maternal (kg) and fetal (kg) weights in singletons and 66 twins.

Chapter 5:

Table 1. Epigenetic and glucocorticoid receptor (GR) expression 7 levels in the pituitary of control and maternally programmed adult offspring.

Abstract ‘Epigenetic changes in the hypothalamus of offspring following maternal undernutrition’, Ghazala Begum- September 2013 Doctor of Philosophy at the University of Manchester

Epidemiological studies show that offspring subjected to maternal undernutrition during early pregnancy are prone to developing obesity and other diseases in adulthood. The hypothalamic energy regulating pathway may be altered in these offspring, with epigenetic changes as a core mechanism. Therefore, this thesis aimed to determine if epigenetic changes are present in this pathway in the hypothalami from offspring subjected to maternal undernutrition. The investigations are focused on the glucocorticoid receptor (GR) as an inhibitor of the anorexigenic neuropeptide pro-opiomelanocortin (POMC), with potential modifications leading to increased food intake and the development of obesity.

To achieve this, an established sheep model developed by our collaborators was used, during which maternal ewes were undernourished periconceptionally to produce a 10-15% decrease in body weight. We found that hypothalami from fetal offspring had greater epigenetic modifications when this reduction in maternal body weight was maintained from 60 days before conception until 30 days into pregnancy, with lower levels of POMC and GR promoter methylation. This was associated with increased GR mRNA expression. Other regions of the brain that also express POMC and GR, did not exhibit these epigenetic modifications. This study revealed that maternal undernutrition induces tissue specific epigenetic changes in fetal hypothalami which may contribute to disease in later life.

Twins have been shown to have similar phenotypic characteristics as maternally undernourished offspring and therefore it has been suggested that they may also be programmed, but by intrauterine growth restriction. Consequently, extensive methylation and histone analysis of GR and POMC promoter regions was carried out in twin fetal hypothalami and compared to maternally undernourished groups. Interestingly, the decreased POMC and GR methylation of our amplicons in the maternally undernourished fetal hypothalami was also observed in twin fetal hypothalamic. This was concomitant with histone modifications and alterations in overall DNA methyltransferase activity. However, it was found that there were no changes in the POMC and GR mRNA expression levels in twin fetuses, but we postulate that this may occur later in life.

To determine if changes in the fetal epigenetic status of hypothalamic GR and POMC impacted the adult progeny, tissues were obtained from adult offspring of maternally undernourished ewes. Epigenetic changes in the hypothalamic GR promoter observed in the fetal group persisted into adulthood, with concurrent increases in GR mRNA and GR protein expression. Of these groups the undernourished adult male offspring had decreased hypothalamic POMC expression and increased fat mass, changes that are consistent with an obese phenotype. The epigenetic and expression status of GR in the hippocampus and pituitary were modified, but in a tissue and sex specific manner. POMC epigenetic changes in the brain were complex, with various levels of epigenetic and expression changes.

Overall periconceptional undernutrition induces hypothalamic specific changes in the epigenetic status of the GR gene which is known to regulate energy balance. Hypothalamic changes were persistent from the fetal stage into adulthood, with modifications in other tissues occurring after birth. These adaptations have the potential to increase the offspring’s propensity to develop obesity and altered stress regulation in later life.

Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and she has given The University of Manchester certain rights to use such Copyright, including for administrative purposes.

ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made.

iii. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual- property.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on presentation of Theses

Abbreviations

3V- third ventricle

5HT1A and 5HT1B- Seretonin receptors 5mec- 5 methyl cytosine ACTH- Adrenocorticotropic hormone ADX- Adrenalectomy AgRP- Agouti related peptide α-MSH- Alpha Melanocyte stimulating hormone ARC- Arcuate nucleus AVP- Arginine vasopressin β-EP- β-endorphin β-LPH- β-lipotrophin β-MSH- β melanocyte stimulating hormone BMI- Body mass index CA1, CA2, CA3- Cornu Ammonis CART- Cocaine-amphetamine related transcript

CCKA and CCKB- Cholecystokinin receptors ChIP- Chromatin Immunoprecipitation CLIP- corticotropin-like Intermediate peptide c-myb- c- myeloblastosis CpG- Cytosine phosphate Guanine CRF1, CFF2- Corticotrophin releasing factor receptors CRH- Corticotrophin releasing hormone D1 and D2- Dopamine receptos Dex- Dexamethasone DOHAD- Developmental origins of health and disease DMH- Dorsomedial hypothalamus DNMT- DNA methyl transferases ETS- E-twenty six FFA- Free fatty acids GABA- -aminobutyric acid γ-LPH- γ-lipotrophin γ-MSH- γ melanocyte stimulating hormone

GC- Glucocorticoids GLP-1- Glucagon-like peptide-1 receptor GR- Glucocorticoid receptor GRE- Glucocorticoid responsive element H1-H4- Histamine receptors H3K9ac- Histone 3 Lysine 9 acetylation H3k27me3- Histone 3 Lysine 27 trimethylation H3k4me3- Histone 3 Lysine 4 trimethylation HAT- Histone acetyltransferases HDAC- Histone deacetylase HMT- Histone methyltransferases HPA- Hypothalamic, Pituitary, adrenal axis IRS-1- Insulin receptor substrate-1 Jak/STAT- Janus kinase/signal transducer and activator of transcription JP- Joining peptide LHA-Lateral hypothalamus MC3R and MC4R- melanocortin 3 and 4 receptors (respectively) MR- mineralocorticoid receptor MMTV- mammary tumour virus long terminal repeat NGFI-A- Nerve growth factor inducible A N-POC- N-terminal POMC fragment NPY- Neuropeptide Y NTS- Nucleus tractus solitarii Ob-Rb- Leptin receptor OCT4- Octamer 4 PBMCs- Peripheral blood mononuclear cells PC1- Prohormone convertase 1 PCR- Polymerase chain reaction PD- Anterior lobe PI- Intermediate lobe POMC- Pro-opiomelanocortin hormone PRMT- Protein arginine N-methyltransferase 1 PVN- Paraventricular nucleus qRT-PCR- Quantitavitve real time polymerase chain reaction 12

SAM- S-adenosylmethionine S.E.M- Standard error of the mean

TAT3- tyrosine amino-transferase VMH- Ventomedial hypothalamus

Nomenclature:

Standard gene nomenclature has been used throughout this thesis. Genes in the text were italicised and proteins were not italicised.

Acknowledgements

I would like to begin by thanking my supervisor Professor Anne White for giving me the opportunity to take on this incredible project. It has been an absolute pleasure to work on. I would also like to thank Anne for her dedication in ensuring that I not only get a good PhD but providing me with the very best training and career opportunities. Most notably her unwavering support and confidence in me has been amazing.

I would also like to say a special thank you to Professor Frank Bloomfield and Professor John Challis for collaborating with us. Their advice and support throughout my project has been fantastic. I especially admire their ability to reply to emails no matter what part of the world they are in or even if they are in the air or on the ground. Most importantly, I would like to thank Frank and the whole of the New Zealand team for providing us with the samples to make this project possible.

I would like to give a special thanks to my colleagues Dr Adam Stevens for being a bioinformatics tour de force and to Alison Davies for being my knight in shining armour when it comes to everything RNA. Special thanks must be extended to the rest of my amazing colleagues who it has been a privilege to work with. Especially, Dr. George Schlossmacher who always listened to my little rants and with his quiet intelligence got me through my first stages of PCR. Many thanks to Dr. Suzanne Meredith, Jennifer Bryant, Dr Erika Harno, Dr Rachel Stovold and Saba Khan for providing many many laughs and for their incredible cooking skills. Yummy. I’d also like to thank our lunch time ‘peeps’ Amanda Patist and Lee Dunham for their excellent company.

Finally I would like to say a massive thank you to my family and friends, including my parents and my sister Asia for their support and love. I would also like to say a special thank you to my Fiancé Mohammed Shiaraz. His ability to steer me out of my little strops whilst writing my thesis and his constant reassurance and belief in me that I can do this has been phenomenal. Lastly, I would like to thank my teacher Shaykh Dr. Muhammad bin Yahya al-Husayni al- Ninowy whose advice resulted in me applying for a PhD. Without their words of wisdom and kind smile this would have been a very difficult journey to get through.

Chapter 1: Introduction

1.1 The implications of maternal programming

The epidemic of obesity in the modern world is increasingly becoming a burden on health services and individual quality of life. It is estimated that at least a third of adults in America are obese and 60% of adults in the UK are overweight or obese (Ogden et al., 2012; Department of health 2012). The sharp escalation in the levels of obesity has not only arisen because of lifestyle choices and social pressures, but also because of genetic factors. In addition to this, epidemiological evidence suggests that the development of obesity is correlated with programming at the maternal and paternal level (Barker et al., 2002; Ng et al., 2010). Maternal programming is associated with changes in the maternal environment, altering maternal fetal dynamics, and thus, impeding the developmental trajectory of the fetus (Hyatt et al., 2010). This is thought to occur to allow the fetus to survive in adverse maternal conditions and to prepare the offspring for any unfavourable conditions after birth.

Programmed offspring are commonly thought to have low birth weight, and this is correlated with the increased chance of developing hypertension, obesity and impaired glucose tolerance in adulthood. This was first determined by Barker and colleagues when they observed an association between low birth weight and the increased risk of developing cardiovascular disease and diabetes in men in Hertfordshire (Hales et al., 1991). Since then further epidemiological evidence has been provided by other long term studies such as the Helsinki cohort, where subjects born in Helsinki between 1934 and 1944 with low birth weights are at a greater risk of developing type 2 diabetes and adult onset disease (Eriksson et al., 2011).

The leading maternal insults that have been described in the literature as having the most detrimental effect on the fetus, have been maternal undernutrition or overnutrition and changes in maternal stress levels (Figure 1.01). The levels of disease progression within the offspring as a consequence of these changes are further influenced by the timing of the maternal insult, the gender of the offspring and fetal number. The timing of the insult is critical in determining the pathway that is affected as the maturation of various organs occurs at different periods during gestation. As a result delineating the pathways that are affected and the implications of these changes on the offspring is extremely complex and requires further work. This thesis describes maternal programming and the potential changes in the offspring, whilst focusing on maternal undernutrition and what is currently known about its effects on the offspring.

Figure 1.01: Programming paradigms The figure depicts the different complications that may occur in adulthood following fetal exposure to an altered maternal environment.

1.1.1 Maternal undernutrition induces programming

Maternal undernutrition is known to induce fetal programming and increase the susceptibly of the offspring to develop coronary complications, diabetes and hypertension (Barker, 2005; Le Clair et al., 2009). It is also termed ‘the developmental origins of health and disease’ (DOHAD) hypothesis. Pioneering human studies determining this have been based on the Dutch famine, which occurred during 1944-1945 (Ravelli et al., 1976). Importantly it was demonstrated that maternal undernutrition at different periods of gestation had different outcomes on the offspring (Ravelli et al., 1976). For example, those mothers exposed to famine early in gestation gave birth to babies with normal birth weights but with an increased likelihood to develop obesity (Ravelli et al., 1976). This is in comparison to those mothers who experienced undernutrition in the second and third trimester who had children with lower birth weights (Ravelli et al., 1976;

Sharkey et al., 2009) Further analysis at approximately 50 years of age of the adult offspring born from this cohort identified impaired glucose tolerance and increased insulin concentrations (Mostyn and Symonds, 2009; Ravelli et al., 1999). It has been suggested that these changes may be due to impaired insulin secretion (de Rooij et al., 2006).

Despite extensive epidemiological evidence the precise mechanisms and pathways in which maternal undernutrition is affecting the fetus are yet to be fully elucidated. It is important to determine these mechanisms as areas of the world are still experiencing famine and women of a reproductive age in developed countries are encouraged to diet, which may induce programming effects within their offspring. Due to the limitations of human studies, several models have been proposed in order to investigate this mechanism, including a rat model. However, unlike humans the rat produces immature young. As a result the fetal developmental events observed in the rat during gestation, are significantly different to those observed in humans (Koutcherov et al., 2003; Muhlhausler et al., 2004; Symonds and Budge, 2009). A more reliable model would be to use an animal that is developmentally similar to humans such as sheep. This is because like humans, sheep give birth to mature young and have a long gestational period. Also ewes like humans, have been found to demonstrate a similar increase in plasma glucocorticoid levels near to the point of birth (Kapoor et al., 2006). Despite this, the rodent model is extremely useful in elucidating programming mechanisms, as it is more accessible and functionally and developmentally more comparable to the human placenta then sheep.

1.1.2 Maternal overnutrition induces programming

Due to the rise in obesity in developed parts of the world such as the USA and Europe, increasing numbers of investigations are focusing on the effects of maternal overnutrition on the fetus. This is in part due to the combination of maternal high fat diet and the lack of exercise. Epidemiological evidence has been essential in demonstrating a link between increased body weight during pregnancy and the development of obesity in the offspring’s adult life (Villamor and Cnattingius, 2006). This is demonstrated by mothers who were subject to bariatric surgery having children who had a lower susceptibility to develop obesity when compared to those offspring born before the surgery (Patti, 2013). Furthermore, mothers who have a higher body mass index (BMI) have progeny with increased adiposity in early childhood and adulthood (Drake and Reynolds, 2010; Oken et al., 2008; Reynolds et al., 2010).

Rodent studies have also been vital in showing the consequences of maternal high fat diet on the offspring not only during pregnancy, but also during lactation (Guo and Jen, 1995; Samuelsson et al., 2008; Tamashiro et al., 2009). The pathways that might be affected in the maternally overnourished offspring are similar to those in the maternally undernourished offspring and will be described in more detail in the following sections.

1.1.3 Stress-induced maternal programming

Several studies in humans have looked at the impact of prenatal maternal stress in the offspring. These studies have correlated maternal stress with a decrease in the offspring’s birth weight (Sandman et al., 1997). More importantly, the long term impact of maternal stress on the offspring causes an increased likelihood of the progeny to develop cardiovascular disease as well as delays in cognitive progression, which could be linked to psychopathology in later life (Buitelaar et al., 2003; Gutteling et al., 2005; Meaney et al., 2007). Furthermore, mothers with increased cortisol levels during pregnancy give birth earlier than mothers with lower levels of cortisol, with the offspring from mothers with higher cortisol levels, displaying challenging behaviour (de Weerth et al., 2003).

Due to the limitations of human studies, in order to further elucidate the implications of maternal stress on fetal activity, several animal models have been proposed. One of these models is the guinea pig. Like humans the guinea pig produces mature offspring. In addition, the stages of neuroendocrine development in the guinea pig have been extensively documented. The study of Kapoor and Matthews (2005) involved giving pregnant guinea pigs a moderate stressor during gestational days 50, 51 and 52 (PS50), when maximal brain development occurs, as well as gestational days 60, 61 and 62 (PS60), where maximal myelination and glial formation occurs. They found that both PS50 and PS60 offspring had decreased birth weight. PS50 fetuses demonstrated increased basal plasma cortisol levels and anxiety behaviour and therefore potentially increased activity in their stress pathway (the hypothalamic, pituitary, adrenal axis). This indicates that maternal stress induces alterations in fetal development (Kapoor and Matthews, 2005).

Once a set of PS50 and PS60 offspring had become adults the group did a follow-up study to determine the effects of prenatal stress on spatial learning (Kapoor et al., 2009). They concluded that the PS60 offspring were quicker in creating a spatial learning strategy than PS50 offspring indicating that acute exposure to stress at different periods during development can induce diverse programming effects in the offspring.

Thus, the guinea pig model provides evidence of increased activity of the offspring’s’ stress pathway in response to stress, leading to altered learning behaviour in the offspring. The above model also supports the view that the timing of the stressor is critical to the outcome, with the effects of maternal stress still being evident in the adult offspring.

1.1.4 Twinning as a programming paradigm

The rise in women giving birth to twins has increased in developed countries, which is potentially due to women using artificial reproductive technology (Blondel et al., 2002; Muhlhausler et al., 2011). This has also been attributed to women becoming pregnant later in life which is further complicated by multiple ovulations. Current research suggests that twinning may also be a programming paradigm, with similarities to models of intrauterine growth restriction such as low birth weight and preterm delivery (Ombelet et al., 2006). Despite this, there is contradictory evidence in the literature regarding the development of cardiovascular disease, diabetes and obesity in adult twins (Muhlhausler et al., 2011; Phillips et al., 2001). This may be because of the lack of singletons used in these studies and the lack of differentiation between the effects of the twin’s genotype or its environment (Muhlhausler et al., 2011). As a result more investigations need to be carried out to determine the effects of twinning as a model of intrauterine growth restriction.

1.2 Programming of the hypothalamic energy regulating pathway

1.2.1 Hypothalamic energy regulating pathway

The hypothalamic energy regulating pathway which is primarily based within the arcuate nucleus (ARC) is predicted to be a target for maternal programming. The ARC consists of an important network of neuropeptides including, pro-opiomelanocortin (POMC), cocaine- amphetamine regulated transcript (CART), neuropeptide Y (NPY) and agouti related peptide (AgRP) (Challis and Yeo, 2002). POMC has been shown to be anorexigenic with the ability to inhibit food intake and stimulate energy expenditure. To do this POMC must undergo extensive hypothalamic post-translational cleavage by a family of serine proteases (PC1/3), to produce active peptides (Bergendahl et al., 1992; Hagan et al., 1999; Pritchard and White, 2007). Furthermore, as depicted in figure 1.02, the amount of cleavage is tissue specific (Benjannet et al., 1991; Pritchard et al., 2003; Pritchard and White, 2007; Rousseau et al., 2007). Despite this, it has been shown that the amount of unprocessed POMC is greater than the amount of its cleaved peptides. Thus, POMC is likely to be synthesised, stored and then processed as and when it is required (Pritchard et al., 2003; Pritchard and White, 2007; Wardlaw et al., 1998).

In the hypothalamus, POMC processing results in the production of α-melanocyte stimulating hormone (α-MSH). It has been shown that central administration of this peptide can result in a reduction in food intake (Ludwig et al., 1998). POMC neurons extend to a large number of locations in the hypothalamus including the paraventricular nucleus (PVN), which is also essential for further regulation of energy balance. Upon stimulation, POMC neurons projecting to the PVN release α-MSH from their synaptic terminals, which then binds to the melanocortin 3 and 4 receptors (MC3R and MC4R respectively), activating further anorexigenic processes. However, in situations where food has not been consumed for a period of time the orexigenic pathway is triggered. This induces greater food consumption by competitively inhibiting αMSH by AgRP, thus, preventing αMSH from binding to the MC3R and MC4R receptors in the PVN (Figure 1.03; Cripps et al., 2005).

NPY exerts its effects by acting on NPY receptors and has extensive projections to other areas of the brain that are essential for appetite control such as the PVN, DMH (Dorsomedial hypothalamus) and LHA (lateral hypothalamus). It has also been shown that AgRP/ NPY inhibit POMC neurones by local GABA ( -aminobutyric acid) release allowing the neurones to

21 influence subsequent food intake (Cowley et al., 2001; Horvath et al., 1997). The hypothalamic energy regulating pathway is highly complex and in addition to the neuropeptides above there are a number of influencing factors involved in this system outlined in table 1 that are beyond the scope of this thesis.

Pituitary

Figure 1.02: Tissue specific POMC processing PC1/3 cleaves POMC to give the final peptides Adrenocorticotropic hormone (ACTH), N-terminal POMC fragment (N-POC), joining peptide (JP) and β-lipotrophin (β-LPH). Further cleavage in the hypothalamus and the skin occurs via PC2 and other enzymes to produce γ-lipotrophin (γ-LPH), β-endorphin (β-EP) and the melanocyte stimulating hormones γ-MSH, β-MSH and α- MSH as well as corticotropin-like Intermediate peptide (CLIP) (adapted from Pritchard and White, 2007).

Neurotransmiter, Receptor Actions Amines and Amino acids

Orexin-A OX1 and OX2 Orexigenic

Melanin-concentrating SLC-1 Orexigenic hormone Galanin GALR1, GALR2, GALR3 Orexigenic

Glutamate N-methyl-D-aspartate Anorexigenic

γ-aminobutyric acid GABAA, GABAB Anorexigenic

Dynorphin κ-opioid, μ-opioid, δ- Anorexigenic opioid , N-methyl-D- aspartic acid (NMDA)- type glutamate receptor Serotonin 14 receptors with 5- Anorexigenic

HT1A and 5-HT1B implicated in feeding and control Glucagon-like peptide-1 Glucagon like peptide-1 Anorexigenic receptor Ghrelin growth hormone Orexigenic by activating secretagogue receptor NPY/AgRP and inhibiting POMC

Cholecystokinin CCKA and CCKB receptors Anorexigenic

Corticotrophin-releasing CRH1 and CRH2 Anorexigenic factor

Dopamine D1 and D2 Initiates feeding

Noradrenaline Dependent on site of action Increase or decrease feeding depending on the site of action

Histamine H1-H4 Suppression of food intake

Nicotine α7, α4β2, β4 Reduces appetite and alters feeding patterns

Table 1.01: Examples of hypothalamic neurotransmitters, amines and amino acids and their implications on the regulation of food intake. Information was compiled and adapted from Meister, 2007; Williams et al., 2000. Dopamine receptor (D1 and D2), Histamine receptors (H1-H4), Corticotrophin releasing factor receptors (CRH1 and CRH2), Galanin receptor 1, 2,3 (GALR1, GLAR2, GALR3), Cholecystokinin receptors (CCKA and CCKB), Glucagon-like peptide-1 (GLP-1) receptor Serotonin receptors (5HT1A and 5HT1B).

The ARC has a specifically modified blood brain barrier allowing peptides through, effecting energy balance. Consequently this pathway can be influenced by a number of peripheral factors such as adipocyte derived leptin, which can pass through this barrier and bind to hypothalamic leptin receptors on the anorexigenic and orexigenic neurones. This allows leptin to influence transcriptional activity by stimulating the anorexigenic neurones and inhibiting orexigenic neurones by the Janus kinase/signal transducer and activator of transcription (Jak/STAT) signalling pathway (Plum et al., 2006). Glucose has been shown to induce central affects by glucose excitatory and inhibitory neurones. These neurones are present in many areas of the hypothalamus, with POMC/CART and NPY/AgRP being a prime example of glucose sensing neurones. Initial studies found that these neurones responded to increased concentrations of applied glucose, where AgRP/NPY neurones were inhibited and POMC neurones were excited (Ibrahim et al., 2003; Mountjoy et al., 2007; Muroya et al., 1999). Furthermore, following food consumption increases in glucose levels result in the release of insulin from pancreatic beta cells. Centrally, insulin has been shown to induce anorectic effects leading to reduced food intake and body weight (Brown et al., 2006). It does this by acting on the insulin receptors present on the POMC and AgRP/NPY neurones, which stimulates the insulin receptor substrate-1 (IRS-1). The insulin pathway then converges on the leptin pathway leading to the activation of phosphatidylinositol-3-OH kinase. However, insulin and leptin can also act by different methods such as through adenosine monophosphate-dependent kinase and the mammalian target of rapamycin pathways in the hypothalamus to regulate food intake and glucose homeostasis, as extensively reviewed in the literature (Varela and Horvath, 2012; Williams and Elmquist, 2012). Overall, the actions lead to the inhibition of AgRP and the stimulation of POMC.

The anorexigenic and orexigenic pathways have extensive projections not only from the ARC to the PVN which connects to regions in the spinal cord, but also the ARC to the LHA and then the nucleus tractus solitarii (NTS) (Marino et al., 2011; Sohn et al., 2013). It is through these connections and further central and peripheral extensions that POMC/CART and NPY/AgRP neurons are able to manipulate adiposity, thermogenesis, peripheral glucose regulation and pancreatic secretion (Buijs et al., 2001; Konner et al., 2007; Xu et al., 2010). In this way changes in the hypothalamus are able to influence the development of obesity and impaired glucose tolerance. Indeed the most striking studies demonstrating this have been in humans where the subjects have had mutations in the POMC gene, it’s processing enzymes or leptin (Challis et al., 2002; Farooqi et al., 2006; Farooqi et al., 2007). The overall outcome has been the development of obesity, due to the impaired ability to inhibit food intake. Furthermore, hypothalamic gene 24 delivery of POMC has been shown to improve glucose tolerance in rats with age-related obesity (Li et al., 2005).

Figure 1.03: Hypothalamic appetite regulatory pathway In situations of positive energy balance, leptin, insulin and glucose stimulate POMC neurones and inhibit NPY/AGRP neurones. POMC is released and is cleaved to α-MSH, which binds to MC3R and MC4R. This results in the activation of the satiety centre, reducing food intake (Adapted from Sandoval et al., 2009; Stevens et al., 2010).

1.2.2 Role of glucocorticoids in the hypothalamic energy regulating pathway

Glucocorticoids (GC) have a well-recognised role in regulating hepatic glucose output but they are also known to have effects on food intake, and the development of obesity. Animal studies have shown that the removal of glucocorticoids by adrenalectomy (ADX) causes a decrease in food intake and body weight (Dallman et al., 2004; Strack et al., 1995). It has been found that associated changes in obese rodent models, have been reversed following ADX and the obesity was correlated with GC excess (Makimura et al., 2000). Furthermore, continuous administration of GCs can trigger increased food consumption and consequently excess body weight gain (Zakrzewska et al., 1999).

The hypothalamus has a selective permeable blood brain barrier, which allows GCs to enter the brain (Karssen et al., 2002). This enables GCs to influence the energy regulating pathway via the glucocorticoid receptor (GR) present on anorexigenic and orexigenic neurons in the ARC. However, the role of the GCs acting on the POMC gene in this pathway is contentious and yet to be fully delineated. It has been shown that increasing the levels of hypothalamic GCs, leads to a decrease in POMC expression (Beaulieu et al., 1988; Gyengesi et al., 2010; Sato et al., 2005; Zakrzewska et al., 1999). Furthermore, following ADX there is greater anorexigenic tone in the POMC neurones and less inhibitory synapses indicating that GCs inhibit POMC and thereby cause an increase in food intake (Gyengesi et al., 2010; Rorato et al., 2008). In contrast to these findings another study found that ADX resulted in a decrease in hypothalamic Pomc expression in rats (Savontaus et al., 2002). The differences in results could be due to different experimental conditions and alternative models being used.

In comparison to POMC the role of GCs on NPY and AgRP neurones is more apparent. Constant intracerebroventricular infusion of synthetic GCs (dexamethasone) in the ARC of the hypothalamus leads to an increase in Npy (Zakrzewska et al., 1999). Furthermore, incubation of hypothalamic organotypic cultures with dexamethasone leads to an increase in Npy and Agrp (Goto et al., 2006; Sato et al., 2005; Shimizu et al., 2010). Overall, from the evidence presented, it can be suggested that GCs in the brain act to inhibit POMC and stimulate NPY and AgRP neurones to increase food intake, promoting body weight gain.

1.2.3 Effects of maternal undernutrition on the offspring’s energy regulating pathway

The establishment of the fetal hypothalamic energy regulating pathway during gestation implies that any perturbations in the maternal nutritional environment could alter its developmental organisation and trajectory. Evidence for this is outlined in table 1.02, showing that maternal undernutrition can alter fetal hypothalamic cell proliferation leading to changes in the number of NPY and α-MSH neurones (Garcia et al., 2010). Moreover, in postnatal offspring there was decreased cell proliferation and axon elongation leading to impaired anorexigenic and orexigenic nerve fibre density (Breton et al., 2009; Delahaye et al., 2008). Therefore, the offspring’s hypothalamic responses to feeding and fasting may be altered in a manner that could lead to increased food intake. Indeed, Garcia et al., 2010 observed an increase in cumulative food intake in their model, suggesting that the offspring have an increased propensity to develop obesity in adulthood.

The authors propose that one of the mechanisms by which hypothalamic development is impaired in response to maternal undernourishment, is changes in leptin levels. In rodents the leptin surge is essential in aiding fibre density and axonal projections from the ARC to the PVN, DMH and LHA, as well as neuronal differentiation and migration in the hypothalamus (Bouret et al., 2004; Udagawa et al., 2007). Thus, any modifications in the levels of leptin could significantly impact on the offspring’s hypothalamic development. Consequently, studies detailed in table 1, observed changes in leptin levels, which could be responsible for the impaired nerve fibre projections in those offspring, leading to altered food intake (Delahaye et al., 2008; Garcia et al., 2010).

Experimental models also show that maternal nutrient restriction can lead to decreased POMC expression and increased NPY expression in the baboon fetus (Li et al., 2013). With the exception of one study, other models indicate that the decreased anorexigenic neuropeptide expression and increased orexigenetic expression can also be observed in maternally undernourished postnatal offspring (Breton et al., 2009; Cripps et al., 2009; Delahaye et al., 2008; Lopez et al., 2005; Shin et al., 2012). The overall alterations vary from one investigation to another, due to the timing and level of the nutritional insult, but still indicate that the changes are consistent with what is expected of an obese phenotype. However, longitudinal studies are required to demonstrate the further impact of these changes on the adult offspring.

Animal Nutritional Fetal Postnatal Authors Model Programming Hypothalamic hypothalamic changes changes Sheep 50% reduction in food At one week of age the Sébert et intake during 30-80 offspring had a ↓ in al., 2009 days of pregnancy - NPY mRNA levels. ↑ melanocortin 4 and insulin receptors Baboon Mothers given 70% of ↑ NPY and Li et al., control diet GR - 2013 ↓ POMC and STAT3 Rat 50% reduction in ↓ POMC mRNA Delahaye maternal food intake ↓ POMC nerve fibre et al., from embryonic day - projections to the PVN 2008 14 until the end of ↓ Plasma leptin levels lactation. Rat 70% reduction in After fasting Breton et maternal food intake - ↑ NPY mRNA and fibre al., 2009 from day1 to day 21 intensity during gestation Rats 50% calorie restriction ↑AgRP expression only Shin et mid to late gestation - in females. al., 2012 and cross fostering ↓ Plasma leptin levels after birth in females and males Rats Offspring of control ↑NPY, AgRP and leptin Cripps et dams nursed by - receptor al., 2009 mothers on low ↓POMC and CART protein diet. Rats 20% calorie restricted ↓ NPY and α- García et dams during the first MSH neurones al., 2010 12 days of pregnancy -

Rats Postnatal ↑short isoforms of the Lopez et undernutrition from - leptin receptor. ↑ NPY al., 2005 day 2 following birth and AgRP

Table 1.02: The effects of maternal undernutrition on the offspring. The table shows the different studies that have been carried out looking at programming effects in the offspring’s hypothalamus. The table details the model and the effects on fetal and postnatal genes in the hypothalamus.

The results from maternal overnutrition models depicted in table 1.03 are more difficult to interpret with impaired anorexigenic and orexigenic projections from the ARC to the PVN and increased POMC mRNA and decreased NPY in the offspring. As with the maternally undernourished models, the offspring from maternally overnourished mothers had altered

hypothalamic projections and fibre densities that may be associated with leptin resistance (Kirk et al., 2009).

Animal Nutritional Fetal Postnatal Authors Model Programming Hypothalamic hypothalamic changes changes Non- HFD up to 4 years. ↑POMC and Grayson et human Fetuses collected at MC4R, no al., 2010 primates gestational day 130 change in POMC cleaved - peptide levels ↓AgRP fibre projections Rats Diet induced obese ↓AgRP fibres in PVN. Bouret et al., mothers throughout ↓density of α-MSH 2008 gestation and lactation - fibres innervating PVN

Rats Mothers fed a HFD 5 ↓POMC, NPY, Leptin Morris and weeks before gestation - receptor Chen, 2009 and then throughout gestation. Rat Maternal obesogenic diet ↓AgRP projections to Kirk et al., given for 6 weeks before - PVH 2009 and provided through Leptin resistance gestation and lactation Rat Fed on a high-fat ↑ POMC mRNA Chen et al., cafeteria diet for 5 weeks ↓ NPY mRNA 2008 before conception, - through gestation and until the end of lactation. Rat Postnatal overnutrition ↓ long isoform of the Lopez et al., from day 2 following - leptin receptor. 2005 birth ↑CART, NPY and AgRP Sheep Overnutrition at 160% - ↑ POMC mRNA Muhlhausler et al., 2006 Sheep Maternal ewes were ↑ POMC Muhlhausler subjected to glucose mRNA - et al., 2005 infusion from 130 to 140 days during gestation.

Table 1.03: The effects of maternal overnutrition on the offspring. Evidence for alterations in hypothalamic nerve fibre density and neuropeptide expression in maternally overnourished offspring.

The primary outcome of the studies described above is the development of an obese phenotype and impaired glucose regulation and tolerance. However, it is apparent from the rodent models that prenatal under- or over-nutrition alone is not adequate in producing an obese trait in the offspring, but the offspring must be subject to both prenatal and postnatal nutritional changes. For example, the offspring from rats on a high fat diet during gestation and 24 weeks after birth had an increased propensity to develop greater fat mass gain and increase food intake (Ikenasio- Thorpe et al., 2007). One possible explanation for this is that in rodents neurogenesis and neural migration occur during gestation and further axonal projections and synaptic connections occur postnatally (Coupe and Bouret, 2013; Markakis, 2002; Terroni et al., 2005). Thus, to significantly impair the pathway all 3 states may need to be targeted. Unlike rodents, in human fetuses, the 3 stages occur during gestation and therefore, some studies have utilised a model following a similar developmental trajectory such as the sheep (Muhlhausler et al., 2004; Symonds and Budge, 2009). Nevertheless, as shown above, the rodent model has been influential in pioneering studies analysing the effects of nutritional insults on the offspring’s hypothalamic pathways.

To further our knowledge in this field, undernutrition models require more focused investigations on the timing of the dietary changes, the levels of nutritional restriction and the potential of these changes progressing into adulthood. Furthermore, more work needs to be done to characterise possible changes in the regulation of the neuropeptides by GCs in response to maternal under- or over-nutrition. This is becausefrom the literature currently available only one study has described changes in hypothalamic GR following maternal nutritional programming and this has been at the fetal stage (Li et al., 2013).

1.3 Programming of the HPA axis

1.3.1 Key components of the HPA axis

The hypothalamic, pituitary, adrenal (HPA) axis is implicated in the stress response and is also considered to undergo a range of complex modifications in response to maternal programming (Figure 1.04; Tsigos and Chrousos, 2002).

Figure 1.04: The HPA axis In a stressful situation the hippocampus innervates the hypothalamus to release corticotrophin releasing hormone (CRH) and arginine vasopressin (AVP) from the paraventricular nucleus (PVN), which in turn acts on the pituitary corticotrophs to synthesise and release adrenocorticotrophic hormone (ACTH). ACTH then acts on the adrenal cortex to release cortisol. Cortisol then induces negative feedback inhibition of the HPA axis via the glucocorticoid and mineralocorticoid receptors (GR and MR respectively), by acting on the hippocampus, hypothalamus and the pituitary. In addition to this cortisol acts on a wide range on tissues to exert metabolic effects. (Free fatty acids (FFA)).

GCs are an integral component of the HPA axis and are synthesised from cholesterol and released by the adrenal gland de novo. They have a circadian rhythm, with peak blood levels in

31 the morning, and decreased nadir levels in the evening. As shown in figure 1.04, GCs are involved in a range of processes such as carbohydrate, protein and glucose metabolism, with effects on the cardiovascular system and inflammation. More specifically during the stress response, GCs will exert feedback inhibition on the hippocampus, the hypothalamus and the pituitary by binding to the glucocorticoid receptor (GR). As a result GCs can cause a reduction in POMC expression, so that less adrenocorticotrophic hormone (ACTH) is produced and less GCs are released. GCs also act to decrease the amount of corticotrophin releasing hormone (CRH) production in the hypothalamus (Sapolsky et al., 2000). Therefore, the actions of GCs result in the down-regulation of the HPA axis.

1.3.2 Effects of maternal undernutrition on the offspring’s HPA axis

The components of the HPA axis are necessary for a wide range of fetal processes. For example, near term the expression of hippocampal GR mRNA is decreased, resulting in a reduction in negative feedback, allowing an increase in the levels of GCs. The elevation of GCs is known as the prepartum surge. This up-regulation is necessary for the development and maturation of the brain and many other organs such as the lungs, kidney and liver (Kapoor et al., 2006). Furthermore, GCs are necessary for triggering the onset of parturition (McMillen et al., 2004). Another example is fetal POMC, which undergoes posttranslational modification to produce β- endorphin, which is necessary for the regulation of proliferating cells and neuronal differentiation (Angelogianni et al., 2000). Due to the developmental implications of these components, modifications in their expression levels due to programming could have harmful effects on the fetus.

The HPA axis of the offspring is subject to changes as a consequence of maternal undernutrition (Edwards and McMillen, 2002; Lesage et al., 2001; Rumball et al., 2008). However, there is conflicting evidence in the literature regarding how maternal undernutrition would programme this pathway. For example, in an investigation by our collaborators using sheep, maternal ewes were undernourished to achieve a 10-15% reduction in body weight which was maintained for 60 days before and 30 days after conception. When compared to the control fetuses, the undernourished fetuses had increased basal cortisol levels, increased expression of pituitary POMC mRNA and higher levels of pituitary PC1. Furthermore, in the pituitary there was no change in the levels of GR mRNA in the pars distalis and the pars intermedia had no detectable

32 expression of GR mRNA. These findings indicate a possible increase in ACTH secretion, which would result in more GC secretion. The overall effect would be an over-activation of the HPA axis, coupled with decreased negative feedback. The increase in GCs could lead to a prepartum surge, which would result in the mother giving birth earlier (Bloomfield et al., 2004).

In contrast, an earlier study, which also involved under-nourishing the ewes to achieve a 15% reduction in body weight, but between 0 to 70 days of gestation, there was no change in POMC mRNA expression in the pituitary. However, GR mRNA was reduced in the anterior pituitary of the undernourished fetuses compared to the controls (Hawkins et al., 2001).

In experiments that utilised the rat model, pregnant dams were fed half their daily food intake from the last week of gestation to weaning. The undernourished fetuses had a reduced body weight compared to the controls. This study demonstrated a different outcome to previous studies with no alteration in pituitary POMC mRNA expression (Lesage et al., 2001). Additionally, in a similar model it was found that maternal undernutrition enhanced the expression of POMC and did not affect the expression of PC1 in the anterior pituitary in malnourished fetuses (Sebaai et al., 2002).

As well as the hippocampus, hypothalamus and pituitary it has been implied that maternal nutritional changes could be affecting the offspring’s HPA axis at the level of the adrenal gland. The adrenal gland has been shown to have an elevated level of activity during early gestation. Evidence for this was provided by Long et al., 2012, where the offspring of obese ewes had higher ACTH levels in response to a CRH/AVP challenge, but the cortisol levels stayed the same, indicating altered adrenal responsiveness. Further longitudinal studies investigating the effects of brief maternal undernutrition during late gestation characterised the offspring at 30 months of age. It was found that they also had an increase in ACTH levels in response to a CRH/AVP challenge, but there was no comparable increase in cortisol levels (Bloomfield et al., 2004). Therefore, maternal under- and over-nutrition has the capacity to impair adrenal responsiveness in the offspring’s’ HPA axis. This in part has been explained by impaired adrenal growth and steroidogenesis (MacLaughlin et al., 2007).

The impact of maternal undernutrition on the fetal HPA axis has been shown to vary with fetal number and sex. In order to understand these differences, the fetal HPA activity has been investigated under normal conditions in twins and singletons and in both males and females. In 33 twins there is a reduction in HPA activity during late gestation, with the length of gestation being longer compared to singletons. Furthermore, the prepartum cortisol surge in twins occurs asynchronously and postnatally the offspring have increased central HPA activity (Rumball et al., 2008). Gender differences in singletons included plasma ACTH levels being higher in the males then the females, with no differences in the levels of fetal GCs (Edwards and McMillen, 2002).

The effects of periconceptional undernutrition in sheep on sex and fetal twins were determined by Edwards and McMillen in 2002. The HPA activity of the different genders following periconceptional undernutrition was the same. In twins there was an increase in ACTH concentrations. However, in another study this increase in ACTH concentration was correlated with a decrease in pituitary responsiveness (Rumball et al., 2008).

Gender dependent differences in cortisol in periconceptionally undernourished fetal sheep has also been shown. In 12 month old female offspring from undernourished mothers the baseline cortisol levels were significantly higher than males and controls. However, there was no change in the baseline ACTH concentrations (Gardner et al., 2006). Therefore, these studies provide some evidence that the different changes in the HPA activity following periconceptional undernutrition might be influenced by fetal sex and whether they are singletons or twins.

As mentioned previously the differences in the outcome between the experiments could be due to the differences in the length and timing of the insult impacting on different stages of the development of the fetus or potential sex specific differences. Therefore, further investigations are required to provide greater clarity on the topic and the potential of these changes to persist into adulthood. Furthermore, more work needs to be carried out to determine whether twinning effects are based on genetic or environmental changes.

1.4 The potential involvement of epigenetics in programming

Over the years, epigenetic alterations have been found to underlie diseases and their progression, most notably cancer (Egger et al., 2004; Feinberg et al., 2006). More recently epigenetic changes within the fetus have been described as occurring as a result of maternal programming. Epigenetics is considered to be the ability of processes to regulate transcriptional activity of a

34 gene by altering the chemical composition or the accessibility of the DNA. The epigenetic changes may be stable and constant or temporary and dynamic. Epigenetic processes are also involved in a number of associated mechanisms regulating chromatin assembly, chromatin segregation and the replication and repair of DNA (Delage and Dashwood, 2008). The epigenetic status of a gene is maintained during cell division and more importantly along the offspring’s germline. Consequently fetal programming has the potential to be inherited by subsequent generations.

Epigenetic regulation is based on the chromatin state, which comprises a DNA nucleosome of 147 base pairs of DNA wrapped around a composite of histone proteins, 2A, 2B, 3 and 4. The accessibility of the DNA by the opening and closing of this chromatin complex is vital in allowing transcription factors to influence gene expression, without directly affecting the DNA sequence (Delage and Dashwood, 2008). This thesis will describe two methods by which epigenetic changes can occur, DNA methylation and histone modifications (Figure 1.05 and Figure 1.06, Tang and Ho, 2007). These epigenetic processes are highly interlinked, thus any alterations in one may affect the state of the other allowing the reinforcement and stability of any changes that may occur (Kaelin and McKnight, 2013). However, there is no clear evidence as to which epigenetic event must occur first to influence any changes in the chromatin structure.

1.4.1 DNA methylation

The primary role of DNA methylation is to transcriptionally silence genes, by condensing heterochromatin. The patterns of methylation are highly reproducible and are maintained faithfully, but aberrant changes can occur in disease systems (Egger et al., 2004). DNA methylation occurs at bases where cytosine and guanine are separated by phosphate (CpG). Collectively these bases can form CpG islands and “shores” that are approximately 200-2000 bp long. These regions consist of a high content of cytosines and guanines, at levels above those that are present in normal vertebrate DNA (Newell-Price, 2003). The islands make up approximately 1% of the human genome, with 70% of the islands being based in the first and second exons, the promoter region and the first intron region. Most CpG islands are spread over transcriptional start sites; consequently they can influence gene regulation (Tang and Ho, 2007). However, the majority of these regions are unmethylated, with high levels of methylation being associated with diseases such as cancer (Feinberg et al., 2006; Jones and Baylin, 2002).

The methylation reaction is catalysed by DNA methyl transferases (DNMT) and requires accessory proteins, which allow the covalent addition of a methyl group to the fifth carbon of a cytosine ring to give 5-methyl cytosine (5mec) (Tang and Ho, 2007). The primary methyl donor for this process is S-adenosylmethionine (SAM), which is derived from the condensation of methionine (Kaelin and McKnight, 2013).

Methylated gene silencing can occur directly or indirectly. The direct process blocks the binding of a transcription factor to a recognition element, present in methylated CpG islands. Indirectly, specific binding proteins will bind to methylated DNA, which attract DNMT and histone deacetylases (HDAC) to produce inactive chromatin (Meaney et al., 2007).

DNA methylation configurations by DNMT3A and DNMT3B are set during embryonic development and the patterns are maintained and reproduced by DNMT1 (Bird, 2002; Reik et al., 1999). In this way DNA methylation undergoes a process of demethylation, followed by de novo methylation and specific demethylation to lay down the epigenetic foundation in the fetus (Newell-Price, 2003). It has been proposed that changes in the nutritional status in the maternal environment could alter the levels of methyl donors available. As a result the patterns of methylation established during development could be disrupted. Consequently there could be impaired progression of epigenetic development in the appetite pathway and/or HPA axis in the fetal offspring.

Figure 1.05: DNA methylation Epigenetic modifications can come in the form of DNA methylation, which occurs via the addition of methyl groups to CpG islands preventing transcription (Adapted from Stevens et al., 2011)

1.4.2 Histone Modifications

Epigenetic changes can also occur in the form of modifications to histone acetylation and histone methylation. In contrast to DNA methylation, histone acetylation causes transient transcriptional activation of the gene. Acetylation is associated with modifications in the histone complex by a set of enzymes known as histone acetyltransferases (HATs). The HATs acetylate specific residues like lysine (K) and arginine (R). These positively charged amino acids reside on histone tails such as histone 3 (H3) and histone 4 (H4). Acetylation of the amino acids allows the chromatin to open so that transcription factors can bind to the DNA (Figure 1.06). In contrast to HATs, HDACs reduce acetylation of the histones. As a result the chromatin is closed, preventing transcription from occurring (Meaney et al., 2007).

In comparison to histone acetylation, histone methylation is more complex as the methylation of different histones can act to either open or close chromatin. It mainly occurs at histones 3 and 4 of lysine and arginine residues. Histone methyltransferases (HMT) methylate the lysine residues and protein arginine N-methyltransferase 1 (PRMT) catalyses the methylation of arginine residues. The lysine residues can be mono-, di- or trimethylated, in comparison to the arginine residues, which can only be mono- or di- methylated (Imhof, 2006). Histone methylation can be reversible as histone demethylases have been found to remove methyl groups via oxidative mechanisms (Tsukada and Nakayama, 2010).

Figure 1.06: Histone modifications Histone modifications commonly occur at histone tails. The diagram depicts 3 of the most common forms of histone modifications and the potential impact on the opening and closing of the chromatin (Adapted from Stevens et al., 2011).

1.4.3 The GR gene as a potential target for epigenetic modification

There are currently very limited studies determining the potential of epigenetic changes in GR in offspring as a consequence of maternal undernutrition. However, there are more studies detailing epigenetic changes of GR due to maternal stress. The changes are extremely complex as the GR gene produces approximately 11 tissue specific mRNA variants, which are highly conserved across species allowing translatable studies (Turner and Muller, 2005; Turner et al., 2008). Each of these variants has its own CpG dense promoter region regulating its transcription in the exon 1 region of the GR gene but the variants differ in their transcriptional regulators that might be recruited (Figure 1.07). Furthermore, as reviewed by Oakley and Cidlowski, (2011) the level of expression of each variant is tissue specific. For example, in mice the 1C variant is more abundantly expressed in the pancreas, lung and colon, but less expressed in the liver. These levels may be altered as a consequence of time and signal dependent changes. Exon 1E and 1F were the most abundantly expressed in the human hippocampus (Oakley and Cidlowski, 2011; Turner and Muller, 2005), despite this, more work needs to be done to determine the expression status of the different variants in other regions of the brain. Importantly each of these 5’ end variants induces the production of the same protein. However, 3’ end variants have been shown

38 to produce different GR proteins, related to adverse health outcomes such as leukaemia (Longui et al., 2000).

Figure 1.07: The GR exon 1 promoter region The diagram shows the different exon 1 mRNA transcript initiating sites. (Adapted from Turner et al., 2008).

1.4.4 Epigenetic alterations of fetal GR following maternal programming

The most striking investigations demonstrating a link between epigenetic status of the GR gene and maternal programming have been based on stress models and maternal care. These studies have focused on charactering the state of the GR gene in the hippocampus and in particular the equivalent areas of exon 1F in the human and exon 17 of the GR promoter in the rat (McGowan et al., 2009; Weaver et al., 2004). This promoter region is regulated by the transcription factor, nerve growth factor inducible-A (NGF1-A). It was found that there were different levels of hippocampal CpG methylation sites in GR 17 in offspring subjected to high or low licking and grooming by their mothers. These changes altered the binding of NGF1-A, with increased binding in the offspring who had high levels of attention and relative increases in GR gene expression (Meaney et al., 2007). The group then went on to translate these findings in suicide victims who had been subject to child abuse and found an increase in hippocampal exon 17 methylation and consequently a decrease in GR expression (McGowan et al., 2009). It has been suggested that this led to a disruption in the feedback of GCs on the central regulation of the HPA axis in these victims.

Evidence of alterations in the epigenetic status of GR as a consequence of maternal nutritional restriction is only just emerging. An initial study found that rats fed a low protein diet gave birth to offspring that had decreased methylation in GR promoter regions in the liver which correlated with a 200% rise in GR expression (Lillycrop et al., 2007). Additionally, another study investigated the effects of intrauterine growth restriction on the levels of GR mRNA transcript expression in the offspring’s hippocampus and leading to HPA axis programming. They found differential levels of epigenetic changes in various histone proteins in the different exon 1

39 promoter regions, which led to overall increases in GR mRNA and protein at birth and at day 21 in growth restricted progeny (Ke et al., 2010). These studies indicate that maternal undernutrition has the potential to alter the epigenetic status of GR, however, more work needs to be done to determine if these changes are present in the hypothalamic appetite signalling pathway and how they might contribute to adult onset obesity in the offspring.

1.4.5 POMC as a potential target of epigenetic modification

The POMC gene consists of three exons, with 2 CpG islands identified in the intron 2 and exon 3 regions (Newell-Price, 2003). Furthermore, POMC gene expression is centrally regulated in the hypothalamus by its two highly conserved distal enhancer regions npe1 and npe2, with a further enhancer region downstream of these regions still being characterised (de Souza et al., 2005; Langlais et al., 2011). Consequently changes in the methylation status in hypothalamic promoter regions of this gene could impact on the development of central obesity. This has recently been shown in an elegant clinical study by Kuehnen et al. 2012, who found that hypermethylation of the 2 POMC CpG islands mentioned above, was significantly associated with obesity. This may have been due to the inability of transcription factors to bind to these elements, thus blocking transcription of POMC and preventing the inhibition of food intake. As a result potential epigenetic changes in key elements of the POMC promoter, as a consequence of maternal programming, could induce an obese phenotype in the adult offspring. Furthermore, GR is able to bind to regions of the POMC gene in a tissue specific manner influencing its transcriptional capability (Drouin et al., 1989). Thus, any epigenetic changes in either gene could influence the regulation of the appetite signalling pathway.

1.4.6 Epigenetic alterations of fetal POMC following maternal nutritional insults

Current literature regarding the general effects of maternal undernutrition on the epigenetic status is conflicting and limited. Initial pioneering studies have demonstrated that rats on a low protein diet have offspring with decreased methylation in the POMC regions over a transcription start site. It was also found that there were no alterations in the NPY methylation status (Coupe et al., 2010). In comparison to this, an investigation on neonatal overfeeding observed hypermethylation at a Sp-1 binding site on the POMC gene, critical for the regulation of POMC

40 by leptin and insulin. Regardless of this, it was found that there was no alteration in POMC gene expression, but these changes could become manifest in adulthood (Plagemann et al., 2009). In another study pre- and post-gestation calorie restriction in rats produced no changes in methylation in the POMC or NPY gene in the offspring, but there was an increase in POMC mRNA expression (Shin et al., 2012).

Again the differences in the outcomes of these studies may in part be due to the varying forms and levels of nutritional insult and difference in the timing of the insult. Consequently more research is required to define potential epigenetic changes in the appetite signalling pathway that may confer the offspring to adult onset disease.

1.5 Overview

Investigating the effects of maternal undernutrition is not only important for countries exposed to famine, but it is also applicable to developed countries where women of a reproductive age are encouraged to diet due to social pressures. This global prevalence of maternal undernutrition makes it necessary to understand the underlying mechanisms leading to increased disease susceptibility in the offspring. As a result there have been a large number of studies in the literature focused on determining the implicated pathways. However, the outcomes from these investigations have been conflicting due to different forms and levels of nutritional restriction being used. Additionally, different studies have exposed the animals to the nutritional insult at varying time points and durations during gestation, impacting on different periods of the developmental trajectory of the fetus. This has led to complex conclusions when comparing the different models. A number of diverse pathways have been hypothesised as being affected by maternal undernutrition, with research only recently becoming more focused on the hypothalamic energy regulating pathway. These studies have demonstrated an association between changes in this pathway and the development of obesity and impaired glucose tolerance in the affected offspring (Breton et al., 2009; Cripps et al., 2009; Delahaye et al., 2008; Lopez et al., 2005). Whilst there are many targets within the energy regulating pathway, none of these studies at the time of our initial investigations examined potential changes in GR as one of the regulators of energy balance, with only a few focused on altered expression of POMC. Additionally, epigenetic modifications in these genes are hypothesised as being induced by maternal undernutrition, yet these mechanisms had not been fully elucidated. Consequently, there are a number of key questions that will be addressed by this thesis to aid further understanding of maternal programming. Defining these mechanisms is important in supporting preventative measures, allowing better care to be available to mothers that may be subject to maternal undernutrition.

1.5.1 Maternal undernutrition sheep model

To determine the effects of maternal undernutrition on the offspring the work in this thesis will be based on a well-established sheep model developed by our collaborators in New Zealand (Bloomfield et al., 2003; Bloomfield et al., 2004). As mentioned previously (Section 1.2.3) there are many advantages to using sheep in our investigations as they have a long gestation period

42 and the fetal brain develops in a similar manner to the human brain (Muhlhausler et al., 2004; Symonds and Budge, 2009). Thus, due to our interests in maternal programming of fetal hypothalami, sheep are an ideal animal to study. As demonstrated in figure 1.08, the maternal ewes in our studies will undergo undernourishment to achieve a 10-15% reduction in body weight which is maintained during the periconceptional period. The fetal samples are then taken at day 135 before the cortisol surge and parturition. For further analysis separate batches of mothers are kept who give birth to offspring which are kept for a number of months/ years.

Figure 1.08: Maternal programming sheep model. Maternal ewes were maternally undernourished to achieve a 10-15% reduction in body weight which was maintained from -60 to 30 days after which the mothers were fed ad libitum. Control mothers were fed ad libitum throughout the study. The fetal offspring were then collected at day 135. During the fetal period additional analysis was carried out by our collaborators at days 85 and 125 (Bloomfield et al., 2003; Bloomfield et al., 2004). Another batch of maternal ewes gave birth and the offspring were analysed at 10 months of age and approximately 5 years of age (Todd et al. 2009; Jaquiery et al. 2012; Oliver et al., 2012).

Experiments by our collaborators were carried out during the study to determine the effects of maternal undernutrition on the HPA axis and glucose and insulin homeostasis (Bloomfield et al., 2003; Bloomfield et al., 2004; Connor et al., 2009; Todd et al. 2009; Jaquiery et al. 2012). The results from these studies are outlined in table 1:04. It was found that that there are clear alterations in the HPA axis, with maternally undernourished offspring demonstrating an overactive HPA axis during fetal life (Bloomfield et al., 2004). However, subsequent analysis after birth showed that the HPA axis becomes suppressed during adulthood, with reduced responses to isolation stress (Oliver et al., 2012). Additionally, impaired insulin responses to a glucose challenge at 10 months of age were also observed in this model in the maternally undernourished male and female offspring (Todd et al., 2009; Jaquiery et al., 2012). There was also increased fat mass at 3-4 years of age in the maternally undernourished males only (Todd et al., 2009; Jaquiery et al., 2012). Combined these changes suggest altered regulation of essential pathways as a potential consequence of modifications in key genes in regions of the brain. These

43 areas include the hippocampus and pituitary in relation to the stress axis and the hypothalamus for the control of energy balance. However, further analysis of these pathways is essential in aiding our understanding of the underlying mechanisms for these changes.

Fetal offspring Adult offspring 125 days of gestation increased fetal Increased weight gain at 10 months of cortisol (Bloomfield et al., 2003) age in undernourished males only (Todd et al., 2009) Increased POMC and PC-1 mRNA in the pars intermedia of the pituitary in UN Higher glucose levels in response to a fetuses (Bloomfield et al., 2004) glucose tolerance test and decreased insulin:glucose ratios in UN females and Decreased 11β HSD2 in the placenta of males at 10 months of age (Todd et al., maternally undernourished animals at 2009) day 85 of gestation (Connor et al., 2009) Reduced first phase insulin response to a bolus of glucose in UN females onlt at 10 months of age (Todd et al., 2009)

Increased % fat mass and perirenal fat mass at 3-4 years of age in the UN males only (Jaquiery et al., 2012)

Suppressed cortisol responses in UN males and females in response to CRH and AVP challenge at 10 months of age (Oliver et al., 2012)

Table 1.04: Changes observed in the maternally undernourished offspring. The effect of maternal periconceptional undernutrition in ewes from -60 to 30 days on the fetal offspring and offspring monitored at 10 months and 3-4 years later. (Bloomfield et al., 2003; Bloomfield et al., 2004; Connor et al., 2009; Todd et al. 2009; Jaquiery et al. 2012; Oliver et al., 2012).

1.5.2 Aims

The main aim of this thesis is to determine if moderate maternal undernutrition in sheep alters the epigenetic status of key genes in the hypothalamus including the neuropeptide POMC and its regulator GR, leading to programming of the hypothalamic energy regulating pathway.

1. Epidemiological studies have clearly demonstrated an association between the development of obesity in the adult offspring subject to maternal undernutrition during preconception and early gestation (Ravelli et al., 1976; Roseboom et al., 2006). Consequently, our initial studies focused on investigating the impact of the timing of

moderate maternal undernutrition pre and/or post conception on the hypothalamic energy regulating pathway. This was carried out in a well-established sheep model of maternal undernutrition provided by Professor Frank Bloomfield (Bloomfield et al., 2004). These studies took a candidate gene approach with the anorexigenic neuropeptide POMC and its regulator GR being the main targets. This allowed us to interpret any possible changes that may confer maternally undernourished offspring to increase food intake and decrease energy expenditure. It also determined the period of maternal undernutrition during early gestation that induces the greatest changes, which was important for our future studies. For epigenetic analysis, the promoter regions for both genes were characterised based on mammalian conservation and where a clear framework for the genomic sequence was available, as the entire sheep genome is currently not available for public access. Additionally this study determined if any changes seen in the POMC and GR genes are tissue specific by analysing the epigenetic and expression status of these genes in other regions of the brain such as the hippocampus and the pituitary.

2. Twins have been found to have an increased susceptibility of developing obesity and diabetes in adulthood (Muhlhausler et al., 2011; Phillips et al., 2001), similar to offspring subject to maternal undernutrition. This has been hypothesised to be a result of twins having to compete for space and nutrition within the womb (Rumball et al., 2008). Accordingly, we postulated that if twins are competing for nutrients, they may have a programming paradigm similar to nutritional restriction. To determine this, the next aim of my thesis was to characterise the epigenetic and expression status of POMC and GR in twin hypothalami and to compare it with maternally undernourished hypothalami. The epigenetic analysis was also extended to measure associated histone expression and overall DNA methyltransferase activity in the hypothalamus. Furthermore, we extensively investigated the potential of these changes to be tissue specific.

3. Most programming studies have failed to determine the long term progression of any changes that have been found during fetal life into adulthood. Therefore, the final aim of this thesis was to decipher if any epigenetic modifications found at the fetal stage in our candidate genes could be maintained into adulthood in maternally undernourished offspring. These animals, up to 5 years in age, were also analysed to determine if there is an associated phenotype. To accomplish this we carried out comparable epigenetic analysis in the control and maternally undernourished adult offspring similar to those 45 implemented at the fetal stage. In addition to the mRNA expression, the study also measured protein levels of GR and POMC in the hypothalamus and pituitary, and GR in the hippocampus. More recently research has focused on using non-invasive tissues to measure epigenetic changes as markers of aberrant changes in the brain associated with disease outcomes such as obesity (Kuehnen et al., 2012). Consequently, our investigations also considered the possibility of using peripheral circulating leukocytes from maternally undernourished offspring as a biomarker for changes in different regions of the brain that express GR and POMC.

1.6 Alternative format

This thesis examines the effects of maternal undernutrition on the offspring from fetal life through to adulthood, with specific emphasis on the hypothalamic energy regulating pathway. The work has identified important epigenetic changes in regulatory factors in the hypothalamus, which are implicated in obesity observed in the adult offspring.

As a result the investigations carried out during my PhD form a cohesive body of work that would suit the style of an alternative format thesis, with 2 of the studies already published and one currently accepted for publication. More specifically the first results chapter examines the effects of maternal undernutrition on fetal pathways; the second is based on the impact of maternal undernutrition and twinning. The third chapter determines whether the epigenetic changes identified in the fetus persist through to the adults. The final results chapter is unpublished work written in standard thesis format outlining changes in the POMC gene in the adult, which is work that has continued on from the fetal study presented in the first results chapter. All of these chapters are focused on epigenetic and expression changes in GR and related neuropeptides and how they progress from fetal life into adulthood.

Chapter 3: The effect of different times of maternal undernutrition on the epigenetic status of POMC and GR in the fetal hypothalamic appetite signalling pathway Endocrinology, 2010.

Adam Stevens, Ghazala Begum, Alice Cook, Kristin Connor, Christopher Rumball, Mark Oliver, John Challis, Frank Bloomfield, and Anne White

Contribution of authors: Epigenetic and expression analysis of genes in the hippocampus and pituitary was carried out by G.B, as well as all the work on the groups undernourished at 0 to 30 and -60 to 0 days and replicate hypothalamic work. A.S designed the experiments, analysed the data and wrote the paper. A.C (Student project) carried out the initial hypothalamic gene expression and epigenetic work. The in vivo model was designed and created by F.B, K.C, C.R and M.O. With funding obtained and provided by F.B, M.O and J.C for the in vivo model. J.C

47 and F.B helped to evaluate the paper. A.W determined the experimental hypothesis, acquired funding for the study and wrote the paper.

Chapter 4: Epigenetic changes in the fetal hypothalamic appetite regulating pathway associated with maternal undernutrition and twinning FASEB 2012

Ghazala Begum, Adam Stevens, Emma Bolton Smith, Kristin Connor, John R. G. Challis, Frank Bloomfield, and Anne White

G.B helped to devise the hypothesis, carried out all the molecular experiments and wrote the paper. A.S helped to analyse the data as well as the bioinformatics analysis. E.B.S provided experimental support. F.B and K.C provided funding for the in vivo study and produced the model. F.B and J.C advised with the study and the paper. A.W developed the hypothesis, gained funding for the molecular studies and wrote the paper.

Chapter 5: Maternal undernutrition programs tissue-specific epigenetic changes in the glucocorticoid receptor in adult offspring Accepted for publication in Endocrinology 2013

Ghazala Begum, Alison Davies, Adam Stevens, Mark Oliver, Anne Jaquiery John Challis, Jane Harding, Frank Bloomfield & Anne White

G.B performed the methylation, chromatin immunoprecipitation (ChIP) and western blot experiments, overall data analysis and wrote the paper. A.D carried out RNA purification and qPCR optimisation and analysis. A.S provided bioinformatics support and helped provide supplementary figure 1 for the paper. J.H., M.O., J.C. and F.H.B. designed the in vivo studies which were undertaken by M.O., A.J. and F.H.B. with funding obtained by J.H., M.O. and F.H.B. J.C and F.H.B. critiqued the paper. A.W devised the hypothesis, obtained funding for the molecular studies and wrote the paper.

Chapter 2

Methods

2.1 Animal management

Authorisation for the study was granted by the Animal Ethics Committee at The University of Auckland, New Zealand. The animals study was developed by our collaborators and maintained in New Zealand (Bloomfield et al. 2004, Jaquiery et al. 2012). The methods they used are described as follows. Multiparous Romney ewes were fed a concentrate feed consisting of Lucerne (65%), barley (30%) limestone, molasses and trace elements (CamTech, Cambridge, New Zealand) (Bloomfield et al,. 2004). The ewes were then randomly separated into four groups and treated accordingly Firstly, controls which were fed ad libitum at 3-4% of body weight per day; secondly, undernourished ewes from 60days before to 30days after mating (-60 to +30); thirdly, undernourished ewes from 60days prior to and until mating (-60 to 0) and lastly, undernourished ewes from 2days before mating up until 30days after mating (0 to +30). Undernourishment of the ewes was achieved by fasting the ewes for 2days, followed by individually determined concentrate feeds to induce and sustain a 10-15% reduction in maternal body weight. Initial food intake was at 1-2% of body weight per day, rising to 80% to that of the controls. After the relative periods of undernourishment the ewes were fed ad libitum.

An ultrasound scan at 55days post conception determined fetal number. Term in untreated ewes was approximately 147days, and so therefore, fetal tissue was collected at 132days (in twin fetuses) and 135days (in singleton fetuses) respectively by giving a lethal dose of intravenous pentobarbitone to the pregnant ewes. Subsequently, fetal hypothalamic, pituitary and hippocampal samples were harvested and frozen until assayed.

Batches of the undernourished and control cohorts were kept for out adult studies and the mothers were allowed to give birth to their offspring naturally. When these groups of offspring had reached 3-4 years of age, body composition was measured by dual-emission X-ray absorptiometry (DXA, Norland XR-800, Cooper Surgical Lts., Fort Atkinson, WI, USA). Before the scan occurred the animals were given access to water but no food overnight. They were then sedated with an intravenous injection of an equivolume solution of diazepam, 5mg/mL-1 and ketamine 100mgmL-1. The lean and fat mass was then quantified using the Norland software (Jaquiery et al., 2012).

For the adult study the offspring were reared for up to a period of 5 years. Animals were then euthanised using pentobarbitone after which tissue and blood samples were immediately collected and frozen until assayed.

The hypothalamus was dissected to provide arcuate nucleus enriched tissue from the mammillary body through to the optic chiasm. This was done by isolating the ventral region of the hypothalamus, with the 3rd ventricle clearly visible and used as a guide. For hippocampal analysis the right lateral hippocampal region was taken for all animals which contain the dentate gyrus and the Cornu Ammonis regions (CA1, CA2 and CA3).

2.2 DNA, RNA and protein purification from brain tissues

DNA and RNA were isolated from 15mg of frozen fetal and adult hypothalamic, pituitary and hippocampal tissue using the Qiagen AllPrep DNA/RNA/protein Mini kit (Qiagen, West Sussex, UK) in accordance with the manufacturer’s instructions). The RNA and DNA yield was measured using a spectrophotometer (NanoDrop ND-1000, Delaware, USA) and normalised appropriately with DEPC water according to individual experimental protocols.

For the adult study, RNA was extracted and purified from brain tissue in a two-step procedure. Firstly, a trizol extraction step was carried out whereby1.2ml of Trizol (Invitrogen, UK) was added to the tissue and homogenised using a handheld homogeniser (Qiagen TissueRuptor 230V, 50/6-Hz, UK) for 20sec. Samples were incubated at room temperature for 5minutes after which 240µl of chloroform was added. Samples were shaken vigorously for 15sec and then incubated at room temperature for 2-3 minutes. Samples were then centrifuged at 10,000 rpm for 15mins at 4oC. The samples subsequently formed 3 layers and the resultant aqueous upper layer containing the RNA was carefully harvested and placed into a fresh tube. This was further purified using an Rneasy kit (Qiagen, West Sussex, UK) which was carried out in accordance to the manufacturer’s protocol beginning with the addition of 70% ethanol. Following the procedure RNA levels were then measured and normalised using the Nanodrop (Nanodrop ND1000, Delawear, USA).

RNA integrity was determined by running 200ng of each sample on a 2% agarose gel containing ethidium bromide. Total RNA was deemed to be intact when 2 clear bands were observed representing 28S and 18S, with 28S being twice the size of 18S (Figure 2.01). RNA was 51 considered to be degraded if a low molecular weight smear was seen. If this occurred then RNA was purified from the remaining homogenised tissue and checked for integrity again.

Figure 2.01: RNA integrity gel. Gel depicting the RNA integrity of 6 control adult samples and 5 maternally undernourished samples. 2 bands are present on the gel representing 28S and 18S.

RNA from a subset of hypothalamic fetal samples used in the fetal study (Stevens et al., 2010) were tested for RNA degradation (figure 2.02). Whilst some degradation has occurred this may have happened over time. However, there is still a clear distinction of the 2 bands even after 5 years and multiple freeze thawing.

(a), (b),

Figure 2.02: RNA integrity of whole and ventral hypothalamic samples. RNA samples were measured on a 2% agarose gel from (a) whole and (b) ventral hypothalami from control and -60 to 30 fetal samples used in the Stevens et al. 2010 paper.

To prepare protein extract 10mg of homogenous hypothalamic, pituitary or hippocampal tissue was incubated on ice for 15 minutes with 500μl of extraction buffer (section 2.11). Cell debris was then removed from the samples by vortexing and centrifugation for 4 minutes at 8000rpm. The supernatant was removed and quantified for protein by using the Nanodrop (Nanodrop ND1000, Delawear, USA) at absorbance 280nm. Once the measurements had been obtained the samples were then normalised.

2.3 Whole blood DNA and protein purification

In order to isolate DNA and protein from whole blood the Gentra Puregene Blood kit (Qiagen, West Sussex, UK) was utilised. DNA was extracted from 3ml of whole blood according to the manufacturer’s protocol. Protein was also extracted using this method; however the protocol was

52 optimised with the following modifications as the kit only details the isolation of DNA. Consequently, the resultant protein precipitate that was generated was added to 2ml of buffer AG (section 2.11) and heated for 1 minute at 65oC and then vortexed 4-6 times at medium speed. This was repeated until the protein pellet had dissolved into solution. This purified protein solution extracted from whole blood was used in subsequent downstream experiments.

2.4 Bioinformatic analysis

In order to determine the sequences used for primer design the human and the bovine genome sequence was used to predict the GR, POMC and Oct4 promoter regions using the Encyclopaedia of DNA Elements (encode). These regions were located downstream of the ATG translational start site by approximately 12kb. The University of California, Santa Cruz (UCSC) website was used to determine the CpG content and mammalian conservation. Sequence homology was determined by aligning the cow, dog, human, mouse and platypus sequence and the level of conservation was measured by the phylogenetic hidden Markov model, phastcons (Siepel et al., 2005). This enabled primers to be designed specifically in the areas of conserved DNA, which allowed amplification of the sheep DNA. These samples were then ran on an agarose gel and the bands were purified using the Qiaquick PCR purification kit (Qiagen, West Sussex, UK) and sent for sequencing (in house service). The obtained sequence was then analysed and matched for homology with conserved consensus sequence using ClustalW 2.0 (European Molecular Biology Laboratory-European Bioinformatics Institute).

2.5 DNA methylation analysis

The MethylCollector kit (Active Motif, Carlsbad, CA) was used to separate methylated genomic DNA from total tissue specific genomic DNA. The experiment involved digesting 44ng/µl of total genomic DNA using the restriction endonuclease Mse1 (New England Biolabs (NEB) in order to produce CpG methylated DNA fragments. These fragments along with the positive and negative control DNA provided in the kit were incubated with a His-tagged recombinant Methyl- CpG binding domain 2b (MBD2b) protein. Nickel-coated magnetic beads were used to capture the His-tagged MBD2b-DNA complexes. The beads were then washed with a high salt buffer to remove unmethylated fragments and incubated with Proteinase K to allow the elution of the methylated DNA.

In order to determine the amount of methylated DNA extracted from the total genomic DNA, polymerase chain reaction (PCR) analysis was performed. The initial Mse1 digested DNA was compared to the amount of tissue specific methylation enriched DNA and represented as the respective input and output DNA values. The PCR reactions contained 3µl of template DNA, 2.5 U of Taq polymerase (Qiagen), 10X PCR buffer (Qiagen), 200µmol/L of each dNTP, 10 pmol/l of each primer and enough nuclease free water to make the reactions up to 50µl. The reaction cycle involved initial denaturation (94oC, 5 minutes), followed by 30 cycles of 94 oC 2 minutes, 55 oC for 30 seconds, 72 oC for 30 seconds and the final extension of 72 oC for 5 minutes.

In order to target the POMC and GR promoter regions the primers used in the reaction were designed by using the human genome sequence (build Feb.2009 (GRch37/hg19) and the bovine genome sequence (build Oct 2007 Baylor 4.0/bos Tau4) (Stevens et al., 2010). Primers were also generated targeting the promoter region of Octamer 4 (OCT4) which is generally found to be hypermethylated (Schneider et al., 2010). This provided a further control in the experiment. Primers used to generate the amplicons were as follows:

Gene Sequence Size (bp) GR 5’-TTTGGAGGGACTGTGGTCC-3’, 230 5’-AGCAGGAGGTGGCAGGCC-3’ POMC 5’-ACCCTCAGAGGTGAGAAGCT-3’, 160 5’- GGAAGGAGACCGGAGCCG-3’ Oct 4 5’- CCTGGATGAGCTTCCAAGG-3’, 223 5’-CCTCGGAGTTGCTCTCCCAC-3’ Table 2.01: Primers used for methylation analysis

The PCR reactions were then resolved on a 2% agarose gel, which was maintained at 80v for one hour. ImageJ software (developed by Wayne Rasband, National institutes of Health, Bethesda, MD) was then used to quantify the intensity of the bands.

2.6 mRNA expression analysis

2.6.1 mRNA expression analysis during the fetal study

During the fetal work tissue specific messenger RNA expression was quantified using the Quantigene II system (Panomic Inc., Freemont, CA). The assay was carried out as outlined in the manufacturer’s instructions. In summary 2µg of RNA in 10µl of nuclease free water was added to the wells of a 96 well capture plate. The samples were then incubated at 56oC overnight with label probes and with target probes specifically designed to target sheep POMC, GR, NPY and GAPDH (probes designed by Panomics with the accession numbers: POMC-NM001009266, GR-NM001114186, NPY-NM001009452, or GAPDH-AF030943). The wells were then washed several times and incubated again at 56oC for 30 mins with a chemiluminescent substrate. The signal was then measured using the Mithras luminometer at 0.2 secs (Berthold, Pforzheim, Germany). The luminescent signal thus generated directly corresponded to the amount of targeted specific mRNA in the sample.

2.6.2 mRNA expression analysis during the twin study

During the twin study a maximum of 1µg of RNA from hypothalamic and hippocampal tissue was reverse transcribed to cDNA according to the manufacturer protocol using the QuantiTect reverse transcription kit (Qiagen). The cDNA was then added to a SYBR green mastermix (QuantiTect SYBR Green PCR; Qiagen). Primers targeting POMC, GR, NPY and 18S were taken from a previous publication (ebert et al., 2009). They are as follows: Gene Primer Sequence GR forward 5′-ACTGCCCCAAGTGAAAACAGA-3′ reverse 5′-ATGAACAGAAATGGCAGACATTTTATT-3′ POMC forward 5′-GCTGCTGGTCTTGCTGCTTC-3′ reverse 5′-CCTGACACTGGCTCGTCTCC-3′ NPY forward 5′-TCATCACCAGGCAGAGATACGG-3′ reverse 5′-GAGCAAGTTTCCCATCACC-3′ 18s forward 5′-GATGCGGCGGCGTTATTCC-3′ reverse 5′-CTCCTGGTGGTGCCCTTCC-3′ Table 2.02: qRT-PCR primers utilised in the twin study.

As a preliminary experiment to ensure primer specificity, an indiscriminate selection of samples were assayed using the SYBR green mix on the StepOnePlus Real-Time PCR system thermal cycling block (Applied Biosystems, Foster City, CA, USA). After which the samples were run on a 2% agarose gel at 90V for 40 minutes. The bands were then cut out and purified using the QIAquick gel extraction kit (Qiagen) and sent for sequencing to confirm specificity and reliability. Subsequent samples were assayed using the SYBR green mastermix and run on the thermal cycling block and the mRNA expression was calculated using the 2−ΔCt method (Livak and Schmittgen, 2001).

2.6.3 mRNA expression analysis during the adult study mRNA expression analysis was further developed for the samples in the adult study. RNA samples were purified using the trizol method as described above. The samples were then reverse transcribed and taken through the PCR reaction in the same step using the power SYBR green RNA to Ct 1-step kit (Life Technologies, Paisley, UK) for GR, NPY, HPRT and 18S. During this method RNA was aliquoted and diluted to a 100ng/µl stock with DEPC water from which a 5ng/µl stock was prepared and plated into the appropriate wells in a 96 well plate on ice. To this the ‘plus RT’ reaction mix was added to the samples and the ‘minus RT’ mix to the controls of the reaction in accordance with table 2.03. The primers used for the different genes are stated in table 2.04.

plus RT x1 minus RT x 1 125x RT Enzyme mix 0.2 0 2 x RT-PCR mix 12.5 12.5 3µM f/r primer mix 2.5 2.5 Nuclease free water 5.8 6 Total 21 21 Table 2.03: Preparation mastermix for qRT-PCR reactions.

Gene Primer Sequence GR forward 5’-GGACCACCTCCCAAACTCTG-3’ reverse 5’GCTGTCCTTCCACTGCTCTT-3’ NPY forward 5’-CGGAGGACTTGGCCAGATAC-3’ reverse 5’-CTCGGGGCTAGATCGTTTCC-3’ HPRT forward 5’-TTTATTCCTCATGGACTAATTATGGA-3’ reverse 5’-GCCACCCATCTCCTTCATCAC-3’ 18s forward 5’-GATGCGGCGGCGTTATTCC-3’ reverse 5’-CTCCTGGTGGTGCCCTTCC-3’ Table 2.04: qRT-PCR primers utilised in the adult study.

Plates were sealed and then centrifuged briefly at 500rpm, 30 secs at 4oC. The plate was then run on the StepOnePlus Real-Time PCR system thermal cycling block (Applied Biosystems, Foster City, CA, USA) as described in the manufacturers protocol. Results were analysed according to the 7 point standard curve that was also included in the plates and made from a stock of pooled RNA at 80ng/µl, which was serially diluted.

RNA samples were quantified for POMC levels via a 5’ nuclease assay using a FAM reporter dye and Quantitect probe one step qRT-PCR reagents (Qiagen). POMC primers and probes for this were as follows POMC forward: 5'-GCTACGGCGGGTTCATGA-3' POMC reverse: 5'-TTCTTGATGATGGCGTTTTTGA-3' POMC probe: 5'fam-AGCCAAACGCCCCTTGTCACGC-methyl red3'; Again the levels of POMC were determined by the use of a 7 point standard curve of known concentrations.

2.7 Chromatin immunoprecipitation

2.7.1 Chromatin immunoprecipitation for fetal tissues

During the fetal and twin studies ChIP analysis was carried out using the Imprint ChIP kit (Sigm, St. Louis, MO, USA) according to the manufacturer’s instructions. To begin with chromatin from 20mg of hypothalamic, hippocampal or pituitary tissue was released and cross-linked with 1% formaldehyde. These samples were then incubated with micrococcal nuclease (2U/ml; Sigma) to digest the DNA, producing ~500bp fragments of genomic DNA. The fragments were

57 then immunoprecipitated with 2µg with either one of the target rabbit polyclonal histone antibodies, H3K27me3, H3K4me3, H3K9ac (39917, 39155, 39159 respectively; Active Motif, Carlsbad, CA, USA). Samples were also immunoprecipitated with either 1µg of the positive control anti RNA polymerase II antibody (R1350; Sigma) or 1µg of normal mouse IgG (M8695; Sigma), as the negative control. Crosslinks were then hydrolysed, releasing the DNA for collection and ran in PCR reactions for GR, POMC and OCT4 as a control. PCR reactions were set up as described for the methylation assay (Section 2.5), using the same set of primers as those used in that assay.

PCR samples were then run on a 2% agarose gel, as 80V for 1hour. The imageJ software (developed by Wayne Rasband, National institutes of Health, Bethesda, MD) was then used to measure the intensity of the bands samples were normalised to the positive control.

2.7.2 Chromatin immunoprecipitation for adult brain tissues

The ChIP assay outlined above was discontinued during the adult study and as a result we used a different protocol (The Imprint ChIP kit (Sigma) has since been reinstated). This protocol was developed by Dhal and Collas, 2008. During this protocol hypothalamic, hippocampal and pituitary samples were lysed and sonicated for 15secs at 50% pulser, 15% power. The lysate was centrifuged and the supernatant was harvested. RIPA buffer (section 2.11) was then added to the samples, which were sonicated again at 6x10sec pulses at 50% pulser, 15% power to achieve ~500bp chromatin fragments. For the antibody complexes a slurry of dynabeads-protein G (Life Technologies) was prepared with 20µl being used per sample. These magnetic beads were then washed with RIPA buffer 3 times using a magnetic rack. The RIPA buffer was added again and the beads were vortexed and placed on ice. This slurry was split into separate tubes for the samples and the positive and negative controls. Our targeted antibodies for histone analysis were added to the beads. For this the samples were incubated with either one of these antibodies: 2µg H3K9ac (39917; Active Motif, La Hulpe, Belgium) or 2µg H3K27me3 (39155; Active Motif, La Hulpe, Belgium). 1µg RNA polymerase II (R1530 Sigma, Dorset, UK) and 1µg normal mouse IgG (M8695 Sigma, Dorset, UK) were used as positive and negative controls respectively. The antibody bead complexes were incubated for 2 hours at 4oC. Beads were captured on the magnetic rack and the buffer was removed after which the beads were resuspended in 20µl per sample of RIPA buffer. 60µl of chromatin was added to these samples and incubated at 4oC overnight. Samples were washed 3 times with RIPA buffer and then finally with TE buffer.

Beads were captured and the supernatant was removed. 1.5ml of complete elution buffer (section 2.11) was then added and incubated for 2hr at 68oC with agitation. Beads were captured and transferred to new tubes. During these last stages input samples were also prepared by adding 1.5volumes of elution buffer and 7.5µl/100µl proteinase stock to 100µl of chromatin and cortex and incubated for 2hr at 68oC under agitation. Finally the eluted and DNA samples underwent DNA purification using the Qiaquick PCR purification kit (Qiagen). Samples were then run in PCR reactions and tested for GR and POMC using the same primers as those stated for use in the methylation assay. Output samples were compared to input samples and normalised to controls using image J software (developed by Wayne Rasband, National institutes of Health, Bethesda, MD).

2.8 Western blots

Western blots were carried out to measure GR, POMC and NPY in the hypothalamus, GR and POMC in the pituitary and leukocytes and GR in the hippocampus of control and maternally undernourished adult animals. To do this protein content was normalised in all the samples using the nanodrop (Nanodrop ND1000, Delawear, USA) at A280. However, we were unable to normalise protein leukocyte samples as the haemoglobin interfered with fluorescent measurements. The protein samples were boiled at 95°C for 5 minutes in laemmli buffer consisting of 12.5% β-mercaptoethanol. Whilst the samples were boiling a precast 4-12% gradient Novex Bis-Tris gels (Invitrogen) was placed into the appropriate tank with running buffer (Section 2.11). 20µg of the relevant boiled sample was then added to a well in the gels. The ColourPlus Prestained Protein ladder (New Endland Biolabs) was also added. Gels were run at 150V for 60minutes to allow for protein separation. Following the run, the gels were placed onto a nitrocellulose membrane and transferred for 60V for 90 minutes in transfer buffer (section 2.11). The membranes were then blocked in PBS-Tween (0.05% Tween20 [Sigma]) containing 1% milk for 3 hours. Membranes were incubated overnight with the relevant antibody (concentrations outlined in table 2.05), prepared in a 1% milk solution. After the incubation, the membranes were washed 3 times in PBS-Tween and then incubated with the Mouse IgG secondary antibody at 1:2000 (Invitrogen) or the Rabbit IgG antibody (1:5000, Abcam) both of which are conjugated with horse radish peroxidase, in 1%milk for 45 minutes. Membranes were washed as previously described 3 times and incubated with chemiluminescent reagents for HRP (EZ-ECL, Invitrogen). Blots were then visualised by overlaying on X-ray films (Kodak, X- Omart AR). These films were then developed with the Compact X2 automatic developer (X- 59

OGraph, Wiltshire, UK). Images were then scanned and analysed using the image J software (developed by Wayne Rasband, National institutes of Health, Bethesda, MD).

Antibody name Species Supplier Dilution GR (clone 41) Mouse BD Biosciences 1:2000 monoclonal POMC Mouse In house 1:5000 (A1A12) monoclonal NPY Rabbit Sigma 1:2000 polyclonal α-Tubulin Mouse Sigma 1:5000 monoclonal β-actin Mouse Santa Cruz 1:5000 monoclonal Table 2.05: List of primary antibodies used for western blotting.

2.9 POMC ELISA

In order to measure POMC within the circulation of our animals, a double sited ELISA assay established in our laboratory was used (Crosby et al., 1988). This assay has previously been demonstrated for effective analysis of POMC in sheep plasma (Schwartz and Rose, 1998). The assay is based on the monoclonal antibodies A1A12 which target ACTH10-18 of POMC and N1C11 which detects the γ-MSH area of POMC. 2.5mg/ml of the A1A12 antibody was used to coat a 96 well microtitre plate (Nunc Immnunomaxisorp). 100μl of sheep plasma sample was then added as well as standards for a standard curve. 100µl of sample diluent (Section 2.11) was added to all of these wells. Plates were incubated overnight at 4°C, following which they were washed 3 times with Octeia buffer (Section 2.11). 200μl of a biotin labelled N1C11 antibody (250ng/ml) was then added and the plate was incubated at room temperature for 2hrs. After this period, plates were washed again and 200μl/well of Avidin-HRP (Vector) was added and left for 30mins. The plates were then washed and 200μl of 3,3’,5,5’-Tetramethylbenzidine (TMB, Europa Bio-products) was added for 30mins. This was followed by the addition of 100μl of 0.5M HCl to stop the reaction, after which the plates were measured immediately at 450-650nm using an absorbance plate-reader (ELx50, Biotek, Vermont, USA). Samples were quantified using the standard by reading the values across from the known standard curve that was generated.

2.10 DNA methylatransferease activity/inhibition assay

DNA methyltransferase activity was measured using the EpiQuik nuclear extraction kit (Epigentek Group, Farmingdale, NY, USA) as outlined by the instruction manual. Nuclear protein was extracted from 20mg of hypothalamic tissue using the EpiQuik nuclear extraction kit. These extracts were then used to measure activity in the EpiQuik DNMT/inhibition assay kit (Epigentek). As per protocol, the protein samples were incubated in cytosine-rich DNA substrate coated with wells, with Adomet with contained methyl groups. 5-methyl cytosine was then added to the wells, with the subsequent addition of a capture antibody. The provided developing solution was then added, which enabled the activity in the wells to be measured colour- metrically.

2.11 Buffers

Protein extraction buffer: 1M Tris HCl buffer pH 8, 5M ethylenediaminetetraacetic acid (EDTA) (Fisher), 0.1% Triton X-100 and 100mM Trehalose made up to 100ml with distilled water. Protease inhibitor (at manufacturers recommended concentration) (Roche) was added before use.

AG buffer for protein purification from whole blood: 50 mM Na2HPO4,1 mM EDTA, 0.3 mM phenlymethylsulfonyl fluoride (PMSF), 0.1% Triton X-100, 0.5% Nonidet P-40, 0.9% NaCl with a pH of 7.4.

RIPA buffer for ChIP assays- A stock of RIPA buffer was made with 500 μl Tris-HCl 1 M, pH 7.5, 1.4 ml NaCl 5 M, 100 μl EDTA 0.5 M, 250 μl EGTA 0.1 M, 500 μl Triton X-100, 500 μl SDS 10% w/v, 500 μl sodium deoxycholate and made up to 50 ml deionised water. 5ml of this solution was then taken forward and 100 μl sodium butyrate 1 M, 25 μl Phenylmethylsulphonyl fluoride (PMSF, Sigma) 200 mM, 50 μl protease inhibitor cocktail was added to this just before use.

TE Buffer for ChIP assys was made up to 50 ml of deionized water by adding 500 μl Tris-HCl 1 M, pH 8.0, 100 μl EDTA 0.5 M.

Elution buffer for ChIP assays – 10 ml of elution buffer was made by adding 200 μl Tris-HCl 1 M, pH 7.5, 100 μl EDTA 0.5 M, 100 μl NaCl 5 M and made up to 10ml. Before use 5ml of this solution was taken and 100 μl sodium butyrate 1 M, 500 μl SDS 10% w/v, 125 μl proteinase K (2 mg/ml) was added.

Running Buffer for Western Blotting: 20x buffer: 104.6g 3-(N-Morpholino) propane-sulfonic acid (MOPS, Fisher), 60.6g Tris base, 10g sodium dodecyl sulphate (SDS) (Fisher), 3g EDTA made up to 500ml with deionised water.

Transfer Buffer for Western Blotting: 6.1g Tris base, 28.8g glycine, 400ml methanol, made up to 2L with deionised water.

Sample diluent: 0.5g sodium azide, 3.75g BSA (ICN), 18750USP units of sodium heparin, 5mg Mouse IgG and made up to 500ml in PBS.

Octeia wash buffer: 20x buffer: 29.6g Dipotassium phosphate (K2HPO4) (BDH), 4.7g

Potassium dihydrogen phosphate (KH2PO4) (BDH), 180g Sodium Chloride (NaCl) (BDH), 5ml Tween-20, 2.5ml Proclin-300 (Supelco) and made up to 1L in dH2O.

2.12 Statistical analysis

All data was analysed using the GraphPad Prism 5 software (GraphPad Software, Inc., La Jolla, CA) and represented as mean and standard error of the mean. All experiments were carried out 3 times. StepOne software and Microsoft Excel were used to calculate Quantitative real time PCR data and results were determined using a 7 point serial dilution curve for each gene of interest. Data was analysed using the One way ANOVA allowing comparisons between multiple groups followed by a Tukey post hoc test. In cases where 2 groups of data were analysed an unpaired students T test was used. The relevant controls were used throughout these investigations and where stated the data was normalised to the appropriate control. Results were deemed statistically significant when the p value was <0.05* <0.01** <0.005***.

Chapter 3: Publication I

Epigenetic Changes in the Hypothalamic Proopiomelanocortin and Glucocorticoid Receptor Genes in the Ovine Fetus after Periconceptional Undernutrition

Adam Stevens, Ghazala Begum, Alice Cook, Kristin Connor, Christopher Rumball, Mark Oliver, John Challis, Frank Bloomfield, and Anne White Faculty of Life Sciences (G.B., A.W.) and Faculty of Medical and Human Sciences (A.S., A.C., A.W.), University of Manchester, Manchester M13 9PT, United Kingdom; Department of Physiology (K.C., J.C.), University of Toronto, Ontario, M5S 1A1 Canada; and Liggins Institute (K.C., C.R., M.O., F.B.), University of Auckland, Auckland 1142, New Zealand

Endocrinology

ENERGY BALANCE-OBESITY

Epigenetic Changes in the Hypothalamic Proopiomelanocortin and Glucocorticoid Receptor Genes in the Ovine Fetus after Periconceptional Undernutrition

Adam Stevens, Ghazala Begum, Alice Cook, Kristin Connor, Christopher Rumball, Mark Oliver, John Challis, Frank Bloomfield, and Anne White

Faculty of Life Sciences (G.B., A.W.) and Faculty of Medical and Human Sciences (A.S., A.C., A.W.), University of Manchester, Manchester M13 9PT, United Kingdom; Department of Physiology (K.C., J.C.), University of Toronto, Ontario, M5S 1A1 Canada; and Liggins Institute (K.C., C.R., M.O., F.B.), University of Auckland, Auckland 1142, New Zealand

Maternal food restriction is associated with the development of obesity in offspring. This study examined how maternal undernutrition in sheep affects the fetal hypothalamic glucocorticoid receptor (GR) and the appetite-regulating neuropeptides, proopiomelanocortin (POMC) and neuropeptide Y, which it regulates. In fetuses from ewes undernourished from Ϫ60 to ϩ30 d around conception, there was increased histone H3K9 acetylation (1.63-fold) and marked hypomethylation (62% decrease) of the POMC gene promoter but no change in POMC expression. In the same group, acetylation of histone H3K9 associated with the hypothalamic GR gene was increased 1.60-fold and the GR promoter region was hypomethylated (53% decrease). In addition, there was a 4.7-fold increase in hypothalamic GR expression but no change in methylation of GR gene expression in the anterior pituitary or hippocampus. Interestingly, hypomethylation of both POMC and GR promoter markers in fetal hypothalami was also identified after maternal undernutrition from Ϫ60 to 0 d and Ϫ2toϩ30 d. In comparison, the Oct4 gene, was hypermethylated in both control and underfed groups. Periconceptional undernutrition is therefore associated with marked epigenetic changes in hypothalamic genes. Increase in GR expression in the undernourished group may contribute to fetal programming of a predisposition to obesity, via altered GR regulation of POMC and neuropeptide Y. These epigenetic changes in GR and POMC in the hypothalamus may also predispose the offspring to altered regulation of food intake, energy expenditure, and glucose homeostasis later in life. (Endocrinology 151: 3652–3664, 2010)

here is growing concern about the marked rise in in- tries where women of child-bearing age are encouraged to Tcidence of patients with metabolic syndrome, which diet. It is thought that programming of the fetus generates affects up to 25% of the population in the United States changes that allow it adaption to decreased nutrition, (1). At the core of the problem is the association between which therefore are an advantage to a fetus born into an obesity and insulin resistance, which leads to increased environment with limited access to food. This is described risk of developing cardiovascular disease and diabetes (2). as the “thrifty phenotype” (7). However, subsequent in- There is considerable evidence to suggest that maternal creases in nutritional availability lead to increased risk of undernutrition impacts on fetal development, leading to developing metabolic syndrome. adult obesity (3–6). This is relevant not only for areas of Mechanisms associated with the programming of meta- the world where famine exists but also for developed coun- bolic syndrome are unclear, but alterations in the fetal hy-

ISSN Print 0013-7227 ISSN Online 1945-7170 Abbreviations: ARC, Arcuate nucleus; ChIP, chromatin immunoprecipitation; GR, glu- Printed in U.S.A. cocorticoid receptor; H3K9Ac, acetylation of histone H3; HPA, hypothalamic-pituitary- Copyright © 2010 by The Endocrine Society adrenal; HSD, honestly significant difference; NPY, neuropeptide Y; POMC, proopiomel- doi: 10.1210/en.2010-0094 Received January 25, 2010. Accepted May 20, 2010. anocortin; UCSC, University of California, Santa Cruz. First Published Online June 23, 2010

3652 endo.endojournals.org Endocrinology, August 2010, 151(8):3652–3664 Endocrinology, August 2010, 151(8):3652–3664 endo.endojournals.org 3653 pothalamic-pituitary-adrenal (HPA) axis (8, 9) have been lead to inappropriate responses to high-fat diets in adult implicated (9, 10). Glucocorticoids have well-defined direct offspring. However, epigenetic programming of candidate negative feedback on the hypothalamus and pituitary (11) genes in hypothalamic neurons seems a likely mechanism. and also act on the hippocampus, which expresses high levels Patterns of genomic DNA methylation are established of glucocorticoid receptors (GRs) in both adult and fetal tis- during early embryonic development and then faithfully sues and is a target for fetal programming of glucocorti- maintained through life by DNA methyltransferases coid responses (12–14). Periconceptional undernutrition (29–31). Acetylation of histones, including histone H3 in sheep accelerates maturation of the fetal HPA axis in (H3K9Ac) opens compact chromatin structures and al- late gestation, associated with elevated fetal baseline cor- lows transcription (32). The association of epigenetic tisol concentrations (15). Approximately half of these changes with changes in mRNA expression is not neces- ewes deliver early, suggesting a potential link between sarily acute; for instance, methylation of the GR exon 1 periconceptional maternal nutrition and preterm birth promoter region has been associated with long-term (16). Therefore, programming of the HPA axis may also changes in gene expression due to changes in transcription impact on mechanisms that predispose to metabolic syn- factor binding (33). Therefore, a change in genomic DNA drome in adult offspring. methylation can only result in a change in transcription However, it is likely that effects of maternal undernu- when the transcription factor in question is present, trition manifest at the level of the fetal hypothalamus given whether this is acute or chronic depends on the transcrip- that energy homeostasis is tightly controlled in the hypo- tion factor. thalamus to cope with fluctuations in energy consumption Changes in histone acetylation have been associated and expenditure. In particular, maternal undernutrition with the transcriptional regulation of POMC by glucocor- could affect the arcuate nucleus (ARC) anorexigenic neu- ticoids (34). Recently, a rat model of neonatal overfeeding rons, which express proopiomelanocortin (POMC) and has shown hypermethylation of the hypothalamic POMC the orexigenic neurons, which express neuropeptide Y promoter compared with constant hypomethylation of (NPY) (17). POMC plays a central role in energy ho- the NPY gene, implicating plasticity in the POMC pro- meostasis (18) as evidenced by POMC knockout mice moter as a critical “set-point” for body weight regulation that develop obesity (19) and enhanced responsiveness (35). However, nothing is currently known about com- to high-fat feeding (20). POMC mutations have also parable changes after undernutrition. been associated with obesity in humans (21, 22). In The GR has been implicated in programming with addition, hypothalamic POMC has been shown to reg- methylation of the GR exon 1 promoter region being as- ulate glucose homeostasis directly, and this function is sociated with long-term changes in gene expression (30). impaired in obesity (23). Fetal exposure to maternal glucocorticoids also affects the There is evidence, from rodent studies, that maternal HPA axis (36, 37). There are well-established causal re- perinatal undernutrition disturbs the hypothalamic lationships between the epigenetic state, GR expression, POMC anorexigenic circuit in newborn pups (24) and in and the effects of maternal care on stress responses in ro- adult offspring, where it is associated with changes in the dent offspring (38). Undernutrition is known to decrease regulation of food intake, body weight, and energy me- hypothalamic GR expression in adult rats (39, 40). Al- tabolism (25). Also, rats exposed to prenatal undernutri- though glucocorticoids are key regulators of hypotha- tion and postnatal high-fat diet exhibited increased food lamic neuropeptides (41, 42), the potential role of epige- intake and fat mass, which correlated with an increase in netic programming of the GR on regulation of energy hypothalamic POMC (26). homeostasis is unclear. In sheep, fetal effects of midgestation nutrient restric- We and others have shown that hypothalamic neu- tion showed no change in fetal hypothalamic POMC ex- ropeptides are altered in obese Zucker rats (43–45), sug- pression at birth but a significant increase in expression at gesting that neurons in this region may be an important 1 yr of age when exposed to an obesogenic environment, target for epigenetic changes associated with maternal un- thus suggesting that the expression of the POMC gene was dernutrition. It has also been shown that sheep whose programmed despite no initial change and that different mothers were subjected to periconceptional undernutri- windows of sensitivity may exist in different animal mod- tion are heavier than controls at 10 months of age (28). els (27). There is, also, evidence that periconceptional un- Therefore, in this study, we examined whether epigenetic dernutrition in sheep results in glucose intolerance in post- changes in POMC and GR were present in the hypothal- pubertal offspring (28). ami of late gestation fetuses from undernourished mothers It is not clear how maternal undernutrition would af- using a sheep paradigm, chosen because of the similarity fect fetal hypothalamic feeding centers in a way that would with the human fetal HPA axis and placental function 3654 Stevens et al. Epigenetic Changes in Hypothalamic POMC and GR Endocrinology, August 2010, 151(8):3652–3664

(46). We found that both POMC and GR genes are asso- served consensus sequence was analyzed using ClustalW 2.0 ciated with changes in histone acetylation and promoter (European Molecular Biology Laboratory-European Bioin- methylation. In addition, GR expression is already in- formatics Institute). Primers used to generate the amplicons were as follows: GR, creased in fetal hypothalamic feeding centers. Perhaps 5Ј-TTTGGAGGGACTGTGGTCC-3Ј,5Ј-AGCAGGAGGTG- more importantly, the hypomethylation of these POMC GCAGGCC-3Ј, size: 230 bp; for POMC, 5Ј-ACCCTCAGAGGT- and GR promoter markers could influence POMC regu- GAGAAGCT-3Ј,5Ј- GGAAGGAGACCGGAGCCG-3Ј, size: 160 lation and glucocorticoid effects on other hypothalamic bp; for Oct4, 5Ј- CCTGGATGAGCTTCCAAGG-3Ј,5Ј-CCTCG- Ј energy regulating pathways. Early life epigenetic regula- GAGTTGCTCTCCCAC-3 , size: 223 bp. tion of these neuropeptides by maternal nutritional status Hormone and peptide assays may therefore have implications for the glucose homeosta- The presence of POMC in fetal plasma samples was measured sis, food intake, and energy balance of the adult offspring. by ELISA, based on the immunoradiometric assay described previously (49) and shown to work in sheep (50). The assay does not cross-react with the bioactive peptide derivatives of POMC, Materials and Methods ACTH (Ͻ3.6%) or ␣-MSH (Ͻ2.2%), but does recognize both POMC and pro-ACTH (100%) (51). The limit of sensitivity of Animal management, nutritional manipulation, the POMC ELISA was 10 pmol/liter. and surgery Plasma ACTH was measured using a commercial RIA (Dia- sorin, Stillwater, MN) previously validated for use in the sheep Experiments were approved by the Animal Ethics Committee (52). Intra- and interassay coefficients of variation for the ACTH at the University of Auckland. Multiparous 4–5 yr old Romney RIA were 11.8 and 5.6%. ewes were acclimatized to a concentrate feed (CamTech, Cam- Cortisol was measured using mass spectrometry as previously bridge, New Zealand) (47), then randomly divided into four described (53). groups: controls (ad libitum feeds at 3–4% of body weight per day); undernutrition from 60 d before until mating (Ϫ60 to 0), from 2 d before mating until 30 d after (Ϫ2toϩ30), or from 60 d Quantitative analysis of mRNA expression before until 30 d after mating (Ϫ60 to ϩ30). Undernutrition RNA from tissue was extracted with RNeasy kit (QIAGEN, comprised a 2-d fast, and then concentrates were individually Valencia, CA). Quantitative expression analysis of mRNA was adjusted to achieve and maintain 10–15% body weight reduc- undertaken using the Quantigene II system (Panomics, Inc., Free- tion (47). Food intake was initially 1–2% of body weight per day, mont, CA) (54). Quanti-Gene assays were performed according increasing to approximately 80% of controls (47). Ewes were fed to the manufacturer’s protocol. Briefly, 2 ␮g of total RNA in 10 ad libitum when not undernourished. After ultrasound scanning ␮l of ribonuclease-free water was loaded into each well of a at 55 d, only singleton-bearing ewes were retained. Fetal and 96-well plate and incubated with sheep target probes [probes maternal catheterization was carried out at d 127 of gestation. synthesized by Panomics, Inc. from the following accession The fetus was exposed through a midline incision in the maternal numbers: POMC, NM001009266; NPY, NM001009452; abdomen, and polyvinyl catheters were inserted into the tarsal GR, NM001114186; or glyceraldehyde-3-phosphate dehydro- artery and vein of both fetal hind limbs. Maternal femoral artery genase, AF030943], in conjunction with dendrimer DNA am- and vein catheters were also inserted. Fetal arterial blood sam- plifier and labeled probes at 56 C overnight. After washes, chemi- ples (4 ml) were collected into chilled heparinized tubes twice luminescent substrate was added to the wells and incubated at 56 daily at 0800 and 2000 h from d 131 to 135 of gestation for later C for 30 min, then read on a Mithras luminometer (Berthold, plasma analysis. To collect fetal tissue, ewes were killed on ges- Pforzheim, Germany). tational d 131 and 135 by a lethal iv dose of injectable pentobarbitone. Fetal hypothalami and pituitaries were dissected and Methylation-specific PCR analysis immediately frozen until assay (47). Tissue specific methylated genomic DNA was enriched from total genomic DNA using the MethylCollector kit (ActiveMotif, Bioinformatic analysis of sheep genome Carlsbad, CA). Briefly, CpG methylated DNA was digested with The human genome sequence (build GRCh37/hg19) and bo- MseI restriction endonuclease (New England Biolabs, Hitchin, vine genome sequence (build Baylor 4.0/bos Tau4) were used as UK), bound specifically to His-tagged recombinant methyl- a base to map the promoter regions of the POMC, GR, and Oct4 CpG-binding domain protein 2b protein, captured with nickel- genes 12 kb downstream of the ATG translational start site. CpG coated magnetic beads, and subsequent wash steps were per- content was examined by sequence analysis and by GC content formed with a stringent high-salt buffer to remove fragments in a 5-bp widow [University of California, Santa Cruz (UCSC), with little or no methylation. The methylated DNA was then please see http://genome.ucsc.edu/]. Areas of highly conserved eluted from the beads. PCR was performed on input MseI di- mammalian sequence homology were identified from the UCSC gested DNA and compared with methylation-enriched DNA to website using multiple alignments of five vertebrate species (cow, establish relative differences in methylation state. PCR was per- dog, human, mouse, and platypus) with a measure of evolution- formed in a final volume of 50 ␮l, containing 10 pmol of each ary conservation, based on a phylogenetic hidden Markov primer, 200 ␮mol/liter of each dNTP, 2.5 U of Taq polymerase model, phastCons (48). Primers designed from the consensus (QIAGEN), and 3 ␮l of DNA template. The initial denaturation sequence of these conserved homologous regions were used to (97 C, 5 min) was followed by 30 cycles of 1 min at 95 C, 1 min amplify ovine genomic DNA. Successfully amplified ovine at 58 C, 1 min at 72 C, and a final extension step at 72 C for 10 genomic DNA was sequenced and homology with the con- min. PCR band intensity for POMC, GR, and Oct4 amplicons Endocrinology, August 2010, 151(8):3652–3664 endo.endojournals.org 3655

was quantified using ImageJ software (developed by Wayne Ras- (relative to human POMC sequence), where there is an band; National Institutes of Health, Bethesda, MD). associated CpG island studied previously in human cells and associated with expression of the POMC gene (56, Chromatin immunoprecipitation (ChIP) 57). This sequence was CpG rich and located 2.5 kb down- ChIP studies were performed using the Imprint ChIP kit (Sigma, stream from regions homologous to the conserved hypo- St. Louis, MO). Homogenized hypothalamic tissue (20 mg) was cross-linked with 1% formaldehyde, the chromatin isolated, incu- thalamic enhancer regions (58, 59) and their associated bated with cell lysis buffer, and digested with micrococcal nuclease CpG islands (Fig. 1A). (2 U/ml; Sigma) to generate fragments of genomic DNA with an The GR gene promoter marker identified was located 5 average size of 500 bp. Equal amounts of digested chromatin were kb upstream of the translational start site in exon 2 (Fig. 1C) ␮ immunoprecipitated with 1 g of each of the following antibodies: within the 5Ј region of the CpG island associated with the mouse IgG (M8695; Sigma), RNA polymerase II (R1530; Sigma), transcriptional start sites of the exon 1 region as mapped by or Rabbit polyclonal histone H3 acetylK9 (H3K9Ac; Abgene, Rochester, NY; ab12178-50). The chromatin was washed, and the homology to the human genome (Fig. 1D). This amplicon cross-links were hydrolyzed. The DNA was then used for PCR anal- was CpG rich and proximal to glucocorticoid regulatory re- ysis using the same amplicons and conditions for GR and POMC as gions defined in the human genome (60). used for the methylation PCR (described above). PCR band inten- The Oct4 gene has been shown to be a hypermethylated sity was quantified using ImageJ software. pluripotency gene in humans (61) and was selected for this study to act as a control for methylation. A CpG-rich region Statistical analysis is present in the Oct4 gene over the translational start site that All data are presented as mean Ϯ SEM. Statistical analysis was shows conserved homology between human, cow, dog, and performed using the unpaired Student’s t test, repeated measures ANOVA or one-way ANOVA, and a Tukey honestly significant platypus genomes (data not shown). A sheep amplicon was difference (HSD) post hoc test as appropriate using the Graph- selected within this Oct4 CpG island 1 kb downstream of the Pad Prism software (GraphPad Software, Inc., La Jolla, CA). P Ͻ translational start identified by conserved homology. 0.05 was considered statistically significant. Baseline HPA axis activity is not changed in fetuses from mothers with periconceptional Results undernutrition Identification of POMC and GR amplicons for ChIP There was no significant difference in fetal plasma POMC and methylation analysis of ovine genomic DNA concentrations between fetuses from the maternal underfed The choice of DNA amplicons in this study was limited and control groups over 96 h between d 131 and 135 of by the lack of available sheep genomic sequence. How- pregnancy (Fig. 2A). This time period was selected because it ever, we were able to locate amplicon sequences based on was before the cortisol surge that correlates with parturition conservation of sequence homology with mammalian ge- (62). Similarly, there was no difference in total amounts of nomes. The cow genome in particular is highly homolo- fetal plasma POMC, ACTH, and cortisol present over this gous to the sheep and has been used to generate virtual period (area under the curve analysis) (Fig. 2B). models of the sheep genome to aid in the creation of a Immunohistochemistry demonstrated POMC expression genomic “scaffold” (55). in the anterior pituitary and intermediate lobe of fetal tissue Homology analysis of the POMC gene region identified from d 135 (data not shown). Methylation and mRNA ex- one candidate region (Fig. 1A) of highly conserved se- pression of the POMC promoter marker was assessed in fetal quence (Fig. 1B) that was proximal to CpG islands. The pituitary tissue for comparison with hypothalamic data. same approach was used to screen the GR and Oct4 gene Methylation of the POMC promoter region was not altered regions (the latter used as a control), and similarly, one in anterior pituitary fetal tissue from underfed mothers com- candidate region was identified in each gene (Fig. 1C) pared with controls (Fig. 2C), and POMC mRNA expression (data not shown) with highly conserved sequence that was was unchanged between the groups (Fig. 2D). POMC ex- within a CpG island (Fig. 1D) (data not shown). Primers pression in the pituitary was up to 200-fold greater than were designed using the consensus sequence and used to observed in the hypothalamus as expected. amplify ovine genomic DNA, the ovine sequence was subsequently confirmed by sequencing (these sequence data H3K9Ac associated with the hypothalamic POMC have been submitted to the GenBank database under ac- promoter in fetuses from mothers with cession nos. GR, HM118850; POMC, HM118849; and periconceptional undernutrition Oct4, HM118848). Hypothalami from d 135 fetal sheep expressed POMC, The marker of the ovine POMC promoter region used NPY, and GR protein (data not shown), markers of a in this study started 6.5 kb from the 5Ј end of exon 1 functional nutritional regulatory pathway in the hypo- 3656 Stevens et al. Epigenetic Changes in Hypothalamic POMC and GR Endocrinology, August 2010, 151(8):3652–3664

FIG. 1. POMC and GR gene region screening to identify highly conserved, CpG-rich regions. A, The POMC gene region with GC content and mammalian sequence homology mapped from the bovine genome (UCSC database). B, Direct comparison of the POMC amplicon marker region in cow, sheep, mouse, and human sequences. C, The GR gene region with GC content and mammalian sequence homology mapped from the bovine genome (UCSC database). D, Direct comparison of the GR amplicon marker region in cow, sheep, mouse, and human sequences. Sequence homology mapping was performed using the phastCons software (48). Striped box indicates promoter marker region. nPE, Neuronal POMC enhancer. thalamus. POMC expression was localized, by immuno- H3K9Ac on the POMC promoter marker was exam- histochemistry, mainly in the ARC of the fetal sheep ven- ined using ChIP (Fig. 3A). Chromatin from the ventral tral hypothalamus (data not shown). hypothalami of d 131 fetuses from groups of control ewes Endocrinology, August 2010, 151(8):3652–3664 endo.endojournals.org 3657

tion of promoter elements distant to the transcriptional start site have also been shown to effectively immunoprecipi- tate with RNA Polymerase II in a genomic screen (63).

Methylation of the hypothalamic POMC gene in fetuses from undernourished ewes Initially, whole hypothalami were taken from fetuses of a group of normally nourished and underfed animals at d 135 of pregnancy. Analysis of the POMC marker region in the whole hypothalami revealed a 64% decrease in methylation in the fetal hypothalami from the underfed group compared with the controls (P Ͻ 0.01) (Fig. 3C). POMC from tissue restricted mainly to the ARC was also assessed in a group of animals subjected to periconceptional undernutrition (from Ϫ60 to ϩ30 d relative to conception) but killed at d 131 of pregnancy. Genomic DNA from ventral hypothalamic slices containing the third ventricle revealed a 62% decrease in POMC marker methylation in the fetal hypothalami from FIG. 2. HPA axis activity in fetal sheep from control ewes or ewes subjected to the underfed group compared with the periconceptional undernutrition (underfed from 60 d before conception to 30 d after control group (P Ͻ 0.001) (Fig. 3D). conception). A, Plasma POMC concentration over 96 h starting at d 131 of pregnancy (control group, n ϭ 8; undernutrition group, n ϭ 6). B, Area under the curve (AUC) analysis over 96 h from d 131 of pregnancy for POMC (pmol/h ⅐ liter), ACTH (pmol/h ⅐ liter), and Quantitative analysis of POMC cortisol (ng/h ⅐ ml). C, Pituitary tissue samples were obtained from the fetuses for promoter gene expression in the region methylation analysis of the POMC gene promoter marker amplicon in the pituitary hypothalamus of fetuses from (control, n ϭ 6; underfed, n ϭ 6). D, Expression levels of the POMC gene in total RNA purified from pituitary samples (control, n ϭ 6; underfed, n ϭ 6). undernourished ewes There was no difference in expression of POMC in the maternally under- and groups of ewes subjected to periconceptional underfed group compared with the control group in the whole Ϫ ϩ nutrition from 60 to 30 d relative to conception was hypothalami or the ventral hypothalami (from d 135 and immunoprecipitated with a polyclonal antibody to his- 131 fetuses, respectively) (Fig. 4). NPY expression was not tone H3 acetyl-K9 (Fig. 3B). There was a 1.63-fold in- altered in either whole hypothalami (Fig. 4A) or the arc- crease of H3K9Ac in ventral hypothalami from the fetuses uate-enriched ventral region (Fig. 4B). of the underfed group compared with the control group (P Ͻ 0.001). H3K9Ac on the hypothalamic GR gene in fetuses RNA Polymerase II was used as a positive control iden- from undernourished ewes tifying binding to the POMC marker region when it is Because glucocorticoids are known to play a role in ap- associated with the transcriptional machinery (Fig. 3A). petite regulation, we assessed the effect of periconceptional This occurs via folding of the promoter to bring the marker undernutrition on the hypothalamic GR gene. In rats, a CpG region in close proximity to the transcription start site. island is found approximately 2.5-kb 5Ј of exon 2 associated This mechanism has been shown previously for POMC, with the complex of different exon 1 regions. Hypomethy- when RNA Polymerase II was immunoprecipitated with a lation of this region has been associated with increased GR promoter region fragment approximately 400 bp up- expression (64). The GR gene exon 1 region is well charac- stream of the transcriptional start site (34). The associa- terized in humans and rats and as the start site for GR tran- 3658 Stevens et al. Epigenetic Changes in Hypothalamic POMC and GR Endocrinology, August 2010, 151(8):3652–3664

FIG. 4. Expression of the POMC gene in the fetal hypothalamus. Fetal hypothalamic tissue samples were obtained from control and underfed (from 60 d before conception to 30 d after conception). Total RNA was FIG. 3. Epigenetic changes associated with the POMC gene in the used to quantify expression levels of POMC and NPY genes. POMC and fetal hypothalamus. Fetal hypothalamic tissue samples were NPY expression levels were normalized to glyceraldehyde-3-phosphate obtained from normal and underfed maternal sheep (underfed from dehydrogenase (GAPDH). A, Expression levels of POMC and NPY in total 60 d before to 30 d after conception). A, Genomic DNA (gDNA) RNA purified from whole hypothalami at gestational d 135 (control, n ϭ was purified from the tissue. DNA immunoprecipitated with a 3; underfed, n ϭ 3). B, Expression levels of POMC and NPY in total RNA histone H3K9 acetylation antibody was then used for PCR to detect purified from ventral hypothalamic sections at gestational age d 131 the presence of the POMC gene promoter marker. An antibody to enriched for ARC (control, n ϭ 9; underfed, n ϭ 11). RNA Polymerase II (RNA Pol II) was used as a positive control and mouse IgG as a negative control. B, Ratio of PCR signal from H3K9Ac immunoprecipitated DNA to total genomic DNA for DNA Hypomethylation of the hypothalamic GR gene in purified from gestational age d 131 ventral hypothalamic sections (control, n ϭ 9; underfed, n ϭ 11). C, Ratio of POMC gene fetuses from undernourished ewes promoter marker PCR signal from methylated genomic DNA to total Analysis of the sheep GR promoter marker region in the genomic DNA for whole hypothalami at gestational age d 135 whole hypothalami revealed a 40% decrease in methyl- ϭ ϭ (control, n 3; underfed, n 3). D, Ratio of POMC gene promoter Ͻ marker PCR signal from methylated genomic DNA to total genomic ation in underfed fetuses compared with controls (P DNA for ventral hypothalamic sections at gestational age d 131 0.05) (Fig. 6A). GR is ubiquitously expressed in the hy- (control, n ϭ 9; underfed, n ϭ 11). **, P Ͻ 0.01; ***, P Ͻ 0.001. pothalamus with a higher level of expression in the paraventricular nucleus, where it is involved with regulation of scription (65). The sheep GR promoter region marker iden- POMC and NPY (41). GR expression levels are also al- tified in this study was located in a similar CpG region and tered in the ventral hypothalamus in food restricted rats was used to examine the presence of H3K9Ac, as a marker (39). Therefore, we examined tissue from the ventral hypo- of transcriptional activity (Fig. 1C). thalamus that contained mainly the ARC as previously used To investigate GR gene promoter H3K9Ac status, ho- for POMC analysis. Genomic DNA was extracted from the mogenized tissue was used from the same ventral hypo- ventral hypothalamic slice and enriched for methylated thalamic samples taken for the POMC analysis (Fig. 5A). DNA. There was a 53% decrease in methylation in the fetal Genomic DNA fragments pulled down with the immuno- ventral hypothalami from the underfed group compared precipitation were used as PCR template for the 229-bp with the control group (P Ͻ 0.05) (Fig. 6B). sheep GR promoter region marker (Fig. 1C). There was a In comparison, methylation of this GR promoter region 1.60-fold increase of H3K9Ac in ventral hypothalami was shown to be unaltered in fetal pituitary and hip- from the fetuses of the underfed group compared with the pocampal tissue in the group from underfed mothers com- normal group (P Ͻ 0.001) (Fig. 5B). pared with controls (Fig. 6, C and D). Endocrinology, August 2010, 151(8):3652–3664 endo.endojournals.org 3659

FIG. 5. Presence of H3K9Ac associated with the GR gene promoter in the fetal hypothalamus. B, Ratio of PCR signal from H3K9Ac immunoprecipitated DNA to total genomic DNA purified from ventral hypothalamic sections enriched for ARC (normal, n ϭ 9; underfed, n ϭ 11). ***, P Ͻ 0.001. Fetal hypothalamic tissue samples were obtained from normal and underfed maternal sheep (underfed from 60 d before conception to 30 d after conception) at d 131 of gestation. Genomic FIG. 6. Methylation of GR gene promoter region in the fetal DNA was used for ChIP analysis with an antibody specific to H3K9Ac. hypothalamus. Hypothalamic tissue samples were obtained from control A, PCR confirmed the presence of a 229-bp marker fragment of the and underfed (from 60 d before conception to 30 d after conception) fetal GR gene promoter CpG island. RNA Pol II, RNA Polymerase II. sheep. A, Ratio of GR amplicon signal from methylated genomic DNA (Meth gDNA) to total genomic DNA for DNA purified from the whole hypothalami (control, n ϭ 3; underfed, n ϭ 3). B, Ratio of GR amplicon The Oct4 gene is hypermethylated across a range of signal from methylated genomic DNA to total genomic DNA for DNA purified from ventral hypothalamic sections enriched for ARC (control, human tissues (61). A sheep amplicon that marked the n ϭ 9; underfed, n ϭ 11). *, P Ͻ 0.05. C, Ratio of GR amplicon from CpG island surrounding the translational start site of the methylated genomic DNA to total genomic DNA for DNA purified from Oct4 gene was shown to be hypermethylated in sheep fetal pituitary samples at gestational age d 131 (control, n ϭ 12; underfed, n ϭ hypothalamic tissue in both undernourished and control 12). D, Ratio of GR amplicon from methylated genomic DNA to total genomic DNA for DNA purified from hippocampal samples at gestational groups with no difference between groups (Fig. 6E). age d 131 (control, n ϭ 12; underfed, n ϭ 12). E, Ratio of Oct4 amplicon from methylated genomic DNA to total genomic DNA for DNA purified ϭ Quantitative analysis of GR gene expression in from ventral hypothalamic samples at gestational age d 131 (control, n 4; underfed, n ϭ 6). hypothalami of fetuses from undernourished ewes In whole hypothalami, there was a 1.8-fold increase in the expression of the GR gene in the underfed compared 7B), where we had previously shown no change in NPY with the control group (P Ͻ 0.05) (Fig. 7A), where we had gene expression (Fig. 4B). previously shown no change in expression of the NPY In pituitary and hippocampal tissue, fetal GR gene ex- gene (Fig. 4A). In the ventral hypothalami, there was a pression was unchanged between the fetuses from the ma- 4.7-fold increase in GR gene expression (P Ͻ 0.05) (Fig. ternal underfed group and the control group (Fig. 7C). 3660 Stevens et al. Epigenetic Changes in Hypothalamic POMC and GR Endocrinology, August 2010, 151(8):3652–3664

periconceptional maternal undernutrition. Expression of GR mRNA is increased in fetal hypothalami collected 100 d after the periconceptional undernutrition of the mother ceased, and this is associated with promoter hypomethylation and increased H3K9 acetylation. In contrast, the POMC gene from late-gestation fetal hypothalami after maternal periconceptional undernutrition shows no change in expression of mRNA, although the POMC gene promoter region is hypomethylated and H3K9 acetylation levels are increased. This provides evidence for epigenetic changes that could act as a programming mechanism for the hypothalamic POMC and GR genes and predispose hypothalamic feeding centers to abnormal regulation later in life. Indeed, 10-month-old offspring from ewes undernourished using the same protocol were found to have abnormal glucose tolerance and increased body weight (28). The hypomethylation of both POMC and GR genes occurs after maternal undernutrition irrespective of the developmental window during which the undernutrition occurred. The observation of similar levels of hypomethylation in both fetal POMC and GR genes across different periods of periconceptional undernutrition suggests that short time frames either side of conception are sufficient to alter regulatory patterns of FIG. 7. Expression of GR gene in the fetal hypothalamus. Fetal hypothalamic tissue samples were obtained from control and underfed fetal hypothalamic genes. This is consistent with obser- maternal sheep (from 60 d before conception to 30 d after vations of the differences in fetal glucose-insulin and ␮ conception); 2 g of total RNA were used to quantify expression levels HPA axis responses between these groups (47). of GR. A, Expression levels of GR and NPY in total RNA purified from the whole hypothalami (control, n ϭ 3; underfed, n ϭ 3). B, Expression It would appear that an early nutritional insult alters levels of GR in total RNA purified from ventral hypothalamic sections the development of fetal hypothalamic appetite regulation enriched for ARC (control, n ϭ 9; underfed, n ϭ 11). C, Expression centers to increase postnatal survival, assuming a contin- levels of GR in pituitary and hippocampal tissue from fetal sheep at gestational d 131. GAPDH, Glyceraldehyde-3-phosphate dehydrogenase. uation of the undernutrition. This suggests that an adaptive response by the fetus to undernutrition allows an in- The effect of duration of undernutrition on creased chance of postnatal survival (7). However, if such hypothalamic GR and POMC genes changes persist in adulthood when food is abundant, a The investigation of epigenetic changes was extended defect in normal appetite regulation may subsequently to other periods of periconceptional undernutrition, i.e. lead to overeating and obesity. This may explain several Ϫ Ϫ from 60 d to conception ( 60 to 0 d) or from conception findings that animals undernourished prenatally are hy- ϩ to 30 d after (-2 to 30d). Methylation of both POMC and perphagic when given hypercaloric or high-fat diet post- GR promoter markers was significantly reduced in all un- natally, compared with control animals (31, 66). The cur- dernutrition groups compared with controls (Fig. 8A). rent study has not addressed fetal weight and catch up POMC expression was not increased in any of the under- growth, but in a previous study, using the same protocol, nutrition groups compared with controls. However, there fetuses from undernourished mothers grew more slowly was a marked increase in the expression of the GR gene in than well-nourished controls, but this was not reflected by Ϫ ϩ the 60 to 30 d group (Fig. 8B). The expression of the size in late gestation (47). NPY gene in the enriched hypothalami was not signifi- Significant hypomethylation of the POMC promoter cantly different to controls (Fig. 8B). marker was observed in this study, which would suggest that the promoter region is made more accessible to regulation, Discussion but this would not necessarily lead to altered expression of the POMC gene at this prenatal stage. Hypomethylation of This study has identified important epigenetic changes in this marker may indicate hypomethylation across elements POMC and GR genes in the fetal hypothalamus after of the POMC promoter CpG island. This may therefore be Endocrinology, August 2010, 151(8):3652–3664 endo.endojournals.org 3661

methylated CpG and the binding of a transcription factor, as has been shown with the interaction between the nerve growth factor inducible-A transcription factor and the GR gene (68, 69). Plage- mann et al. (35) have examined the role of methylation in the regulation of hypothalamic POMC expression in a rat neonatal overfeeding model and demonstrated hypermethylation in SP1 binding regions. However, we were not able to examine these regions directly because there is no ovine genomic sequence currently available for the homologous regions. Highly conserved hypothalamic- specific POMC regulatory regions have been identified approximately 9 kb upstream of the start of exon 1 (Fig. 1A) (58, 59), and it is possible that the POMC promoter marker region used in this study acts as a marker of methylation changes within this region or in as yet undiscovered tissue specific regulatory regions. Rats undernourished perinatally have decreased hypothalamic POMC expression with maternal undernutrition (24), and the long-term appetite-regulatory system of offspring is altered (25). In the FIG. 8. The effect of different periods of periconceptional undernutrition on hypothalamic GR, POMC, and NPY methylation and gene expression. The Ϫ60 to ϩ30 group (UN Ϫ60 to current study, there was no significant ϩ30) was underfed from 60 d before conception to 30 d after conception. The Ϫ60 to 0 difference in hypothalamic POMC ex- group (UN Ϫ60 to 0) were fed the same diet as the Ϫ60 to ϩ30 group but were allowed to pression between the maternally under- feed ad libitum from conception. The Ϫ2toϩ30 group (UN Ϫ2toϩ30) were fed the same diet for 30 d after conception. Fetal ventral hypothalamic tissue samples, enriched for the fed and the control group, in line with ARC, were obtained from all sample groups at d 131 of gestation. A, Methylation of GR and previous work (26, 37). However, when POMC promoter marker in ventral hypothalami from controls (n ϭ 7), Ϫ60 to ϩ30 (n ϭ 9), these offspring were fed a high-fat diet, Ϫ ϭ Ϫ ϩ ϭ 60 to 0 (n 8), and 2to 30 (n 7) feeding regimens. The 229-bp marker of the GR there were alterations in POMC gene ex- gene and the 160-bp marker of the POMC gene promoter region CpG islands were used to compare the ratio of methylated with unmethylated DNA. One-way ANOVA with Tukey HSD pression (26). Delahaye et al. (24) mea- post hoc test compared with control group. *, P Ͻ 0.05; ***, P Ͻ 0.001. B, Transcriptional sured POMC in postnatal life as opposed expression of POMC, GR, and NPY in the ventral hypothalami from controls (n ϭ 7), Ϫ60 to to fetal life, and therefore, it may be that ϩ30 (n ϭ 11), Ϫ2toϩ30 (n ϭ 7), and Ϫ60 to 0 (n ϭ 8) feeding regimens. One-way ANOVA with Tukey HSD post hoc test compared with all other groups. *, P Ͻ 0.05. Meth gDNA, in the current model, changes may be ob- Methylated genomic DNA; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; UN, served at an equivalent developmental Underfed. postnatal age (24, 47). Caution must be observed with the comparison of data associated with the changes in transcription factor binding generated in rats with observations made in sheep and hu- identified by Newell-Price et al. (57) and also mark the region associated with cortisol-dependent demethylation identified mans in that the hypothalamic neuroendocrine circuitry is by Mizoguchi et al. (66). Recently, hypomethylation of established after birth in rodents and before birth in both part of the proximal POMC gene promoter was asso- sheep and humans, potentially making the sheep model a ciated with food restriction, in a rat model of maternal better comparison with humans than a rat model (46). undernutrition, while equivalent regions in CART and The importance of glucocorticoids in the control of en- NPY were not altered (67). ergy homeostasis is exemplified by Zucker rats, where ad- Silencing mechanisms associated with methylated renalectomy normalizes body weight and glucocorticoid genomic DNA can involve direct interference between a replacement results in increased weight gain (70, 71). The 3662 Stevens et al. Epigenetic Changes in Hypothalamic POMC and GR Endocrinology, August 2010, 151(8):3652–3664 role of glucocorticoids is highlighted in POMC knockout creased glucocorticoid signaling in NPY neurons would mice when pituitary POMC expression is restored. This up-regulate NPY in animals given a high caloric diet (78). activates adrenal glucocorticoid production and develop- In summary, pre- and periconceptional maternal un- ment of obesity (72). dernutrition are associated with changes in late gestation The GR marker region used in this study is close to the in fetal hypothalamic genes involved in feeding regulatory networks. This suggests that minimizing risk of program- rat GR 17 promoter region, where epigenetic changes have been found concomitant with changes in hippocampal GR ming effects in humans needs to consider interventions expression associated with altered maternal care (69). before conception. The epigenetic modification of the Also, recently, glucocorticoid responsive elements have genes is suggested by the H3K9Ac associated with POMC been identified proximal to this region that are active in the and GR genes and the hypomethylation of POMC and GR auto-regulation of GR expression in human lymphoblas- gene markers in the fetal hypothalamic neurons. Given that POMC and GR are centrally involved in modulating toid cells (60). In the current study, there was a marked food intake, this potential for long-term programming increase in the H3K9Ac of the GR gene promoter in the could have implications for hypothalamic regulation of hypothalami from fetuses exposed to maternal undernu- energy balance in adult offspring. trition. This histone modification would suggest open chromatin in the GR promoter region (69) and, together with observed hypomethylation of the GR promoter marker, would predict that GR gene expression is up- Acknowledgments regulated. This change should be faithfully replicated dur- Address all correspondence and requests for reprints to: Professor ing mitosis (69). Indeed, in this study, an increase in hy- Anne White, Faculties of Life Sciences and Medical and Human pothalamic GR gene expression was associated with these Sciences, Manchester Academic Health Sciences Centre, University observed epigenetic changes. of Manchester, 3.016 AV Hill Building, Manchester M13 9PT, Because glucocorticoids can regulate both POMC and United Kingdom. E-mail: [email protected]. NPY, it is difficult to speculate on the implications of in- This work was supported by the Health Research Council of creased GR expression in the hypothalamus, because these New Zealand, the National Research Centre for Growth and neuropeptides have opposing effects on energy balance. It Development, New Zealand, the Canadian Institutes of Health Research, and the National Institute for Health Research is well known that glucocorticoids decrease POMC gene Manchester Biomedical Research Centre. expression in the pituitary, and therefore, it might be pre- Disclosure Summary: The authors have nothing to disclose. dicted that decreased POMC expression in the hypothalamus could up-regulate food intake (41, 73). However, there are tissue specific promoter regions in the POMC References gene. It has been shown that 13 kb of 5Ј POMC gene flanking sequence are required for appropriate spatial and 1. Ford ES, Giles WH, Dietz WH 2002 Prevalence of the metabolic temporal expression in the hypothalamus (74), and re- syndrome among US adults: findings from the third National Health cently, hypothalamic specific control regions have been and Nutrition Examination Survey. JAMA 287:356–359 2. Whitlock G, Lewington S, Sherliker P, Clarke R, Emberson J, Halsey well defined within this region (58, 59). J, Qizilbash N, Collins R, Peto R 2009 Body-mass index and cause- Therefore, it may not be surprising that glucocorticoids specific mortality in 900 000 adults: collaborative analyses of 57 prospective studies. Lancet 373:1083–1096 have been shown to up-regulate POMC gene expression in 3. Ravelli GP, Stein ZA, Susser MW 1976 Obesity in young men after the hypothalamus as demonstrated in adrenalectomised famine exposure in utero and early infancy. N Engl J Med 295:349– rats, where the loss of glucocorticoids results in a decrease 353 4. Barker DJ, Winter PD, Osmond C, Margetts B, Simmonds SJ 1989 in POMC expression and this effect is completely reversed Weight in infancy and death from ischaemic heart disease. Lancet by replacement of physiological doses of glucocorticoid 2:577–580 (75). POMC mRNA levels are also positively regulated by 5. Birch NP, Tracer HL, Hakes DJ, Loh YP 1991 Coordinate regulation of mRNA levels of pro-opiomelanocortin and the candidate glucocorticoids in the rat ARC, in that adrenalectomy pro- processing enzymes PC2 and PC3, but not furin, in rat pituitary duced a marked decrease in POMC mRNA, and this was intermediate lobe. Biochem Biophys Res Commun 179:1311–1319 reversed by treatment with dexamethasone (76). Simi- 6. Bloomfield FH, Oliver MH, Giannoulias CD, Gluckman PD, Harding JE, Challis JRG 2003 Brief Undernutrition in late-gesta- larly, glucocorticoids up-regulate ACTH production in tion sheep programs the hypothalamic-pituitary-adrenal axis in rodent hypothalamic arcuate neurons (77). Therefore, this adult offspring. Endocrinology 144:2933–2940 evidence would predict glucocorticoid up-regulation of 7. Hales CN, Barker DJ 2001 The thrifty phenotype hypothesis. Br Med Bull 60:5–20 POMC expression, which would subsequently lead to a 8. Bouret SG, Simerly RB 2004 Minireview: leptin and development of lean phenotype (72). In comparison, it is known that in- hypothalamic feeding circuits. Endocrinology 145:2621–2626 Endocrinology, August 2010, 151(8):3652–3664 endo.endojournals.org 3663

9. Buckley AJ, Jaquiery AL, Harding JE 2005 Nutritional program- 26. Ikenasio-Thorpe BA, Breier BH, Vickers MH, Fraser M 2007 Pre- ming of adult disease. Cell Tissue Res 322:73–79 natal influences on susceptibility to diet-induced obesity are medi- 10. Challis JR, Bloomfield FH, Bocking AD, Casciani V, Chisaka H, ated by altered neuroendocrine gene expression. J Endocrinol 193: Connor K, Dong X, Gluckman P, Harding JE, Johnstone J, Li W, Lye 31–37 S, Okamura K, Premyslova M 2005 Fetal signals and parturition. J 27. Se´bert SP, Hyatt MA, Chan LL, Patel N, Bell RC, Keisler D, Obstet Gynaecol Res 31:492–499 Stephenson T, Budge H, Symonds ME, Gardner DS 2009 Maternal 11. Stevens A, White A 2010 ACTH: cellular peptide hormone synthesis nutrient restriction between early and midgestation and its impact and secretory pathways. Results Probl Cell Differ 50:63–84 upon appetite regulation after juvenile obesity. Endocrinology 150: 12. Kapoor A, Leen J, Matthews SG 2008 Molecular regulation of the 634–641 hypothalamic-pituitary-adrenal axis in adult male guinea pigs after 28. Todd SE, Oliver MH, Jaquiery AL, Bloomfield FH, Harding JE prenatal stress at different stages of gestation. J Physiol 586:4317– 2009 Periconceptional undernutrition of ewes impairs glucose tol- 4326 erance in their adult offspring. Pediatr Res 65:409–413 13. Setiawan E, Jackson MF, MacDonald JF, Matthews SG 2007 Effects 29. Sinclair KD, Allegrucci C, Singh R, Gardner DS, Sebastian S, of repeated prenatal glucocorticoid exposure on long-term poten- Bispham J, Thurston A, Huntley JF, Rees WD, Maloney CA, Lea tiation in the juvenile guinea-pig hippocampus. J Physiol 581:1033– RG, Craigon J, McEvoy TG, Young LE 2007 DNA methylation, 1042 insulin resistance, and blood pressure in offspring determined by 14. Sloboda DM, Moss TJ, Li S, Matthews SG, Challis JR, Newnham maternal periconceptional B vitamin and methionine status. Proc JP 2008 Expression of glucocorticoid receptor, mineralocorticoid Natl Acad Sci USA 104:19351–19356 receptor, and 11␤-hydroxysteroid dehydrogenase 1 and 2 in the fetal 30. Meaney MJ, Szyf M, Seckl JR 2007 Epigenetic mechanisms of peri- and postnatal ovine hippocampus: ontogeny and effects of prenatal natal programming of hypothalamic-pituitary-adrenal function and glucocorticoid exposure. J Endocrinol 197:213–220 health. Trends Mol Med 13:269–277 15. Bloomfield FH, Oliver MH, Hawkins P, Holloway AC, Campbell 31. Szyf M, Weaver I, Meaney M 2007 Maternal care, the epigenome M, Gluckman PD, Harding JE, Challis JR 2004 Periconceptional and phenotypic differences in behavior. Reprod Toxicol 24:9–19 undernutrition in sheep accelerates maturation of the fetal hypo- 32. Sterner DE, Berger SL 2000 Acetylation of histones and transcrip- thalamic-pituitary-adrenal axis in late gestation. Endocrinology tion-related factors. Microbiol Mol Biol Rev 64:435–459 145:4278–4285 33. Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, Sharma S, 16. Bloomfield FH, Oliver MH, Hawkins P, Campbell M, Phillips DJ, Seckl JR, Dymov S, Szyf M, Meaney MJ 2004 Epigenetic program- Gluckman PD, Challis JR, Harding JE 2003 A periconceptional ming by maternal behavior. Nat Neurosci 7:847–854 nutritional origin for noninfectious preterm birth. Science 300:606 34. Bilodeau S, Vallette-Kasic S, Gauthier Y, Figarella-Branger D, Brue 17. Pritchard LE, White A 2005 Agouti-related protein: more than a T, Berthelet F, Lacroix A, Batista D, Stratakis C, Hanson J, Meij B, melanocortin-4 receptor antagonist? Peptides 26:1759–1770 Drouin J 2006 Role of Brg1 and HDAC2 in GR trans-repression of 18. Cone RD, Lu D, Koppula S, Vage DI, Klungland H, Boston B, Chen the pituitary POMC gene and misexpression in Cushing disease. W, Orth DN,Pouton C, Kesterton RA 1996 The melanocortin re- Genes Dev 20:2871–2886 ceptors: agonists, antagonists, and the hormonal control of pigmen- 35. Plagemann A, Harder T, Brunn M, Harder A, Roepke K, Wittrock- tation. Recent Prog Horm Res 51:287–317; discussion 318 Staar M, Ziska T, Schellong K, Rodekamp E, Melchior K, 19. Yaswen L, Diehl N, Brennan MB, Hochgeschwender U 1999 Obe- Dudenhausen JW 2009 Hypothalamic proopiomelanocortin pro- sity in the mouse model of pro-opiomelanocortin deficiency re- moter methylation becomes altered by early overfeeding: an epi- sponds to peripheral melanocortin. Nat Med 5:1066–1070 genetic model of obesity and the metabolic syndrome. J Physiol 20. Challis BG, Coll AP, Yeo GS, Pinnock SB, Dickson SL, Thresher RR, 587:4963–4976 Dixon J, Zahn D, Rochford JJ, White A, Oliver RL, Millington G, 36. Godfrey KM, Barker DJ 2000 Fetal nutrition and adult disease. Am J Aparicio SA, Colledge WH, Russ AP, Carlton MB, O’Rahilly S 2004 Clinical Nutrition, 71:1344–1352 Mice lacking pro-opiomelanocortin are sensitive to high-fat feeding 37. Ikenasio-Thorpe BA, Breier BH, Vickers MH, Fraser M 2007 Pre- but respond normally to the acute anorectic effects of peptide-YY(3- natal influences on susceptibility to diet-induced obesity are medi- 36). Proc Natl Acad Sci USA 101:4695–4700 ated by altered neuroendocrine gene expression. J Endocrinol 193: 21. Challis BG, Pritchard LE, Creemers JW, Delplanque J, Keogh JM, 31–37 Luan J, Wareham NJ, Yeo GS, Bhattacharyya S, Froguel P, White A, 38. Szyf M, Weaver IC, Champagne FA, Diorio J, Meaney MJ 2005 Farooqi IS, O’Rahilly S 2002 A missense mutation disrupting a di- Maternal programming of steroid receptor expression and pheno- basic prohormone processing site in pro-opiomelanocortin (POMC) type through DNA methylation in the rat. Front Neuroendocrinol increases susceptibility to early-onset obesity through a novel mo- 26:139–162 lecular mechanism. Hum Mol Genet 11:1997–2004 39. Lindblom J, Haitina T, Fredriksson R, Schio¨thHB2005 Differential 22. Farooqi IS, Drop S, Clements A, Keogh JM, Biernacka J, Lowenbein regulation of nuclear receptors, neuropeptides and peptide hor- S, Challis BG, O’Rahilly S 2006 Heterozygosity for a POMC-null mones in the hypothalamus and pituitary of food restricted rats. mutation and increased obesity risk in humans. Diabetes 55:2549– Brain Res Mol Brain Res 133:37–46 2553 40. Brady LS, Smith MA, Gold PW, Herkenham M 1990 Altered ex- 23. Parton LE, Ye CP, Coppari R, Enriori PJ, Choi B, Zhang CY, Xu C, pression of hypothalamic neuropeptide mRNAs in food-restricted Vianna CR, Balthasar N, Lee CE, Elmquist JK, Cowley MA, Lowell and food-deprived rats. Neuroendocrinology 52:441–447 BB 2007 Glucose sensing by POMC neurons regulates glucose ho- 41. Pritchard LE, Turnbull AV, White A 2002 Pro-opiomelanocortin meostasis and is impaired in obesity. Nature 449:228–232 processing in the hypothalamus: impact on melanocortin signalling 24. Delahaye F, Breton C, Risold PY, Enache M, Dutriez-Casteloot I, and obesity. J Endocrinol 172:411–421 Laborie C, Lesage J, Vieau D 2008 Maternal perinatal undernutri- 42. Beaulieu S, Gagne´ B, Barden N 1988 Glucocorticoid regulation of tion drastically reduces postnatal leptin surge and affects the devel- proopiomelanocortin messenger ribonucleic acid content of rat hy- opment of arcuate nucleus proopiomelanocortin neurons in neona- pothalamus. Mol Endocrinol 2:727–731 tal male rat pups. Endocrinology 149:470–475 43. Pritchard LE, Oliver RL, McLoughlin JD, Birtles S, Lawrence CB, 25. Breton C, Lukaszewski MA, Risold PY, Enache M, Guillemot J, Turnbull AV, White A 2003 Proopiomelanocortin-derived peptides Rivie`re G, Delahaye F, Lesage J, Dutriez-Casteloot I, Laborie C, in rat cerebrospinal fluid and hypothalamic extracts: evidence that Vieau D 2009 Maternal prenatal undernutrition alters the response secretion is regulated with respect to energy balance. Endocrinology of POMC neurons to energy status variation in adult male rat off- 144:760–766 spring. Am J Physiol Endocrinol Metab 296:E462–E472 44. Kalra SP 2008 Disruption in the leptin-NPY link underlies the pan- 3664 Stevens et al. Epigenetic Changes in Hypothalamic POMC and GR Endocrinology, August 2010, 151(8):3652–3664

demic of diabetes and metabolic syndrome: new therapeutic ap- Frauenknecht K, Riepert T, Tresch A, Muller AM, Coerdt W, Zechner proaches. Nutrition 24:820–826 U, Haaf T 1 March 2010 Spatial, temporal and interindividual epige- 45. Beck B 2000 Neuropeptides and obesity. Nutrition 16:916–923 netic variation of functionally important DNA methylation patterns. 46. Symonds ME, Budge H 2009 Nutritional models of the develop- Nucl Acids Res 10.1093/nar/gkq126 mental programming of adult health and disease. Proc Nutr Soc 62. Symonds ME, Sebert SP, Hyatt MA, Budge H 2009 Nutritional 68:173–178 programming of the metabolic syndrome. Nat Rev Endocrinol 47. Rumball CW, Bloomfield FH, Oliver MH, Harding JE 2009 Dif- 5:604–610 ferent periods of periconceptional undernutrition have different ef- 63. Tummala P, Mali RS, Guzman E, Zhang X, Mitton KP 2010 Tem- fects on growth, metabolic and endocrine status in fetal sheep. Pe- poral ChIP-on-Chip of RNA-polymerase-II to detect novel gene ac- diatr Res 66:605–613 tivation events during photoreceptor maturation. Mol Vis 16:252– 48. Siepel A, Bejerano G, Pedersen JS, Hinrichs AS, Hou M, 271 Rosenbloom K, Clawson H, Spieth J, Hillier LW, Richards S, 64. Lillycrop KA, Slater-Jefferies JL, Hanson MA, Godfrey KM, Jackson Weinstock GM, Wilson RK, Gibbs RA, Kent WJ, Miller W, Haussler AA, Burdge GC 2007 Induction of altered epigenetic regulation of the D 2005 Evolutionarily conserved elements in vertebrate, insect, worm, hepatic glucocorticoid receptor in the offspring of rats fed a protein- and yeast genomes. Genome Res 15:1034–1050 restricted diet during pregnancy suggests that reduced DNA methyl- 49. Crosby SR, Stewart MF, Ratcliffe JG, White A 1988 Direct mea- transferase-1 expression is involved in impaired DNA methylation and surement of the precursors of adrenocorticotropin in human plasma changes in histone modifications. Br J Nutr 97:1064–1073 by two-site immunoradiometric assay. J Clin Endocrinol Metab 67: 65. Turner JD, Schote AB, Macedo JA, Pelascini LP, Muller CP 2006 1272–1277 Tissue specific glucocorticoid receptor expression, a role for alter- 50. Schwartz J, Rose JC 1998 Development of the pituitary adrenal axis native first exon usage? Biochem Pharmacol 72:1529–1537 in fetal sheep twins. Am J Physiol 274:R1–R8 66. Mizoguchi Y, Kajiume T, Miyagawa S, Okada S, Nishi Y, Kobayashi 51. Gibson S, Crosby SR, Stewart MF, Jennings AM, McCall E, White M 2007 Steroid-dependent ACTH-produced thymic carcinoid: regu- A 1994 Differential release of proopiomelanocortin-derived pep- lation of POMC gene expression by cortisol via methylation of its pro- tides from the human pituitary: evidence from a panel of two-site moter region. Horm Res 67:257–262 immunoradiometric assays. J Clin Endocrinol Metab 78:835–841 67. Coupe´ B, Amarger V, Grit I, Benani A, Parnet P 2010 Nutritional 52. Jeffray TM, Matthews SG, Hammond GL, Challis JR 1998 Diver- programming affects hypothalamic organization and early response gent changes in plasma ACTH and pituitary POMC mRNA after to leptin. Endocrinology 151:702–713 cortisol administration to late-gestation ovine fetus. Am J Physiol 68. McGowan PO, Meaney MJ, Szyf M 2008 Diet and the epigenetic 274:E417–E425 (re)programming of phenotypic differences in behavior. Brain Res 53. Connor KL, Challis JR, van Zijl P, Rumball CW, Alix S, Jaquiery 1237:12–24 AL, Oliver MH, Harding JE, Bloomfield FH 2009 Do alterations in 69. Weaver IC 2009 Epigenetic effects of glucocorticoids. Semin Fetal placental 11␤-hydroxysteroid dehydrogenase (11␤HSD) activities Neonatal Med 14:143–150 explain differences in fetal hypothalamic-pituitary-adrenal (HPA) 70. Marchington D, Rothwell NJ, Stock MJ, York DA 1983 Energy function following periconceptional undernutrition or twinning in balance, diet-induced thermogenesis and brown adipose tissue in sheep? Reprod Sci 16:1201–1212 lean and obese (fa/fa) Zucker rats after adrenalectomy. J Nutr 113: 54. Arikawa E, Sun Y, Wang J, Zhou Q, Ning B, Dial SL, Guo L, Yang 1395–1402 J 2008 Cross-platform comparison of SYBR Green real-time PCR 71. York DA, Holt SJ, Marchington D 1985 Regulation of brown ad- with TaqMan PCR, microarrays and other gene expression measurement technologies evaluated in the microarray quality control ipose tissue thermogenesis by corticosterone in obese fa/fa rats. Int (MAQC) study. BMC Genomics 9:328.:328 J Obes 9(Suppl 2):89–95 55. Dalrymple BP, Kirkness EF, Nefedov M, McWilliam S, Ratnakumar 72. Smart JL, Tolle V, Low MJ 2006 Glucocorticoids exacerbate obesity A, Barris W, Zhao S, Shetty J, Maddox JF, O’Grady M, Nicholas F, and insulin resistance in neuron-specific proopiomelanocortin-de- Crawford AM, Smith T, de Jong PJ, McEwan J, Oddy VH, Cockett ficient mice. J Clin Invest 116:495–505 NE 2007 Using comparative genomics to reorder the human genome 73. Krude H, Gru¨ ters A 2000 Implications of proopiomelanocortin sequence into a virtual sheep genome. Genome Biol 8:R152 (POMC) mutations in humans: the POMC deficiency syndrome. 56. Newell-Price J 2003 Proopiomelanocortin gene expression and Trends Endocrinol Metab 11:15–22 DNA methylation: implications for Cushing’s syndrome and be- 74. Young JI, Otero V, Cerda´n MG, Falzone TL, Chan EC, Low MJ, yond. J Endocrinol 177:365–372 Rubinstein M 1998 Authentic cell-specific and developmentally reg- 57. Newell-Price J, King P, Clark AJ 2001 The CpG island promoter of ulated expression of pro-opiomelanocortin genomic fragments in the human proopiomelanocortin gene is methylated in nonexpress- hypothalamic and hindbrain neurons of transgenic mice. J Neurosci ing normal tissue and tumors and represses expression. Mol Endo- 18:6631–6640 crinol 15:338–348 75. Wardlaw SL, McCarthy KC Conwell IM 1998 Glucocorticoid reg- 58. Santangelo AM, de Souza FS, Franchini LF, Bumaschny VF, Low ulation of hypothalamic proopiomelanocortin. Neuroendocrinol- MJ, Rubinstein M 2007 Ancient exaptation of a CORE-SINE ret- ogy 67:51–57 roposon into a highly conserved mammalian neuronal enhancer of 76. Pelletier G 1993 Regulation of proopiomelanocortin gene expres- the proopiomelanocortin gene. PLoS Genet 3:1813–1826 sion in rat brain and pituitary as studied by in situ hybridization. Ann 59. de Souza FS, Santangelo AM, Bumaschny V, Avale ME, Smart JL, NY Acad Sci 680:246–259 Low MJ, Rubinstein M 2005 Identification of neuronal enhancers 77. Ozawa H, Lion J, Xie CX, Nishi M, Steinbusch H, Kawata M 1999 of the proopiomelanocortin gene by transgenic mouse analysis and Down-regulation of ACTH and glucocorticoid receptor immuno- phylogenetic footprinting. Mol Cell Biol 25:3076–3086 reactivity in hypothalamic arcuate neurons after adrenalectomy in 60. Geng CD, Schwartz JR, Vedeckis WV 2008 A conserved molecular the rat. Neuroreport 10:1571–1575 mechanism is responsible for the auto-up-regulation of glucocorti- 78. Jeanrenaud B, Rohner-Jeanrenaud F 2000 CNS-periphery relation- coid receptor gene promoters. Mol Endocrinol 22:2624–2642 ships and body weight homeostasis: influence of the glucocorticoid 61. Schneider E, Pliushch G, El HN, Galetzka D, Puhl A, Schorsch M, status. Int J Obes Relat Metab Disord 24(Suppl 2):S74–S76

Chapter 4: Publication II

Epigenetic changes in fetal hypothalamic energy regulating pathways are associated with maternal undernutrition and twinning

Ghazala Begum, Adam Stevens, Emma Bolton Smith, Kristin Connor, John R. G. Challis, Frank Bloomfield and Anne White Faculty of Life Sciences (G.B., E.B.S., A.W.) and Faculty of Medical and Human Sciences (A.S., A.W), Department of Endocrinology and Diabetes, University of Manchester, Manchester, UK; Liggins Institute (K.C., F.B) and Child and Youth Health, Department of Paediatrics, University of Auckland, Auckland, New Zealand; National Research Centre for Growth and Development, Auckland, New Zealand; and Department of Physiology (J.C), University of Toronto, Toronto, Ontario, Canada

FASEB

The FASEB Journal • Research Communication

Epigenetic changes in fetal hypothalamic energy regulating pathways are associated with maternal undernutrition and twinning

ʈ Ghazala Begum,* Adam Stevens,† Emma Bolton Smith,* Kristin Connor, ,¶ ʈ John R. G. Challis,¶ Frank Bloomﬁeld,‡,§, and Anne White*,†,1 *Faculty of Life Sciences and †Faculty of Medical and Human Sciences, Department of Endocrinology and Diabetes, University of Manchester, Manchester, UK; ‡Liggins Institute and §Child and Youth Health, Department of Paediatrics, University of Auckland, Auckland, New ʈ Zealand; National Research Centre for Growth and Development, Auckland, New Zealand; and ¶Department of Physiology, University of Toronto, Toronto, Ontario, Canada

ABSTRACT Undernutrition during pregnancy is impli- Key Words: proopiomelanocortin ⅐ glucocorticoid receptor cated in the programming of offspring for the develop- ⅐ methylation analysis ment of obesity and diabetes. We hypothesized that maternal programming causes epigenetic changes in fetal hypothalamic pathways regulating metabolism. This study There is emerging evidence that epigenetic modifi- used sheep to examine the effect of moderate maternal cations of genes are implicated in fetal programming. undernutrition (60 d before to 30 d after mating) and Maternal events that result in programming of the fetus twinning to investigate changes in the key metabolic are clearly important factors in increasing the risk of regulators proopiomelanocortin (POMC) and the gluco- the offspring developing type 2 diabetes (1) and car- corticoid receptor (GR) in fetal hypothalami. Methylation diovascular disease as adults (2). The developmental of the fetal hypothalamic POMC promoter was reduced origins of health and disease (DOHaD) hypothesis in underfed singleton, fed twin, and underfed twin groups predicts that fetal recognition of maternal signals relat- (60, 73, and 63% decrease, respectively). This was associ- ing to the intrauterine environment, such as decreased ated with reduced DNA methyltransferase activity and nutrition, results in adaptations in fetal development (3). These adaptations may be beneficial for the altered histone methylation and acetylation. Methylation intrauterine environment; however, if the postnatal of the hypothalamic GR promoter was decreased in both environment has plentiful nutrition, the adaptations twin groups and in maternally underfed singleton fetuses may cause the offspring to overcompensate (4). This (52, 65, and 55% decrease, respectively). This correlated has been demonstrated in several human and animal with changes in histone methylation and acetylation and studies with the development of obesity and impaired increased GR mRNA expression in the maternally under- glucose tolerance or type 2 diabetes in the adult fed singleton group. Alterations in GR were hypothalamic offspring (5–7). specific, with no changes in hippocampi. Unaltered levels Twinning presents an ideal model to study an alter- of OCT4 promoter methylation indicated gene-specific native effect of nutritional programming. Twinning is effects. In conclusion, twinning and periconceptional known to result in reduced birth weight, which is a undernutrition are associated with epigenetic changes in common trait associated with fetal changes due to fetal hypothalamic POMC and GR genes, potentially maternal programming (8, 9). Furthermore, twins have resulting in altered energy balance regulation in the an increased propensity to develop abdominal obesity offspring.—Begum, G., Stevens, A., Smith, E. B., Connor, and type 2 diabetes in their adult life (5, 10). This K., Challis, J. R. G., Bloomfield, F., White, A. Epigenetic suggests that twinning is an alternative impact that changes in fetal hypothalamic energy regulating pathways would cause epigenetic changes in the hypothalamus. are associated with maternal undernutrition and twinning. Twin studies have also been vital for demonstrating that FASEB J. 26, 1694–1703 (2012). www.fasebj.org the occurrence of type 2 diabetes is strongly influenced by the fetal environment. This is exemplified by studies

Abbreviations: ChIP, chromatin immunoprecipitation; 1 Correspondence: Endocrinology and Diabetes, Faculties DNMT, DNA methyltransferase; DOHaD, developmental or- of Life Sciences and Medical and Human Sciences, Man- igins of health and disease; ENCODE, Encyclopedia of DNA chester Academic Health Sciences Centre, University of Man- Elements; GR, glucocorticoid receptor; H3K9AC, histone 3 chester, AV Hill Bldg., Oxford Rd., Manchester, M13 9PT, lysine 9 acetylation; H3K4me3, histone 3 lysine 4 trimethyla- UK. E-mail: [email protected] tion; H3K27me3, histone 3 lysine 27 trimethylation; HPA, doi: 10.1096/fj.11-198762 hypothalamic-pituitary-adrenal; nPE, neuronal promoter en- This article includes supplemental data. Please visit http:// hancer region; POMC, proopiomelanocortin. www.fasebj.org to obtain this information.

1694 0892-6638/12/0026-1694 © FASEB demonstrating strong relationships between coeffi- MATERIALS AND METHODS cients of birth weight within-twin pairs and adult insulin sensitivity. There are also studies showing that the twin Animal management with the lower birth weight developed type 2 diabetes (11). If twinning is a nutritional paradigm, then any Authorization for the study was provided by the Animal Ethics changes that we observe in the fetal hypothalami from Committee at the University of Auckland. Multiparous Rom- mothers subjected to periconceptional undernutrition ney ewes were fed a concentrate feed of 65% lucerne, 30% will be mirrored in twin fetal hypothalami. Therefore, barley, and limestone, molasses, and trace elements (CamTech, Cambridge, New Zealand). The ewes were then we determined the epigenetic effects of twinning and randomly separated into 2 groups: controls [fed ad libitum at of maternal undernutrition during the periconcep- 3–4% of body weight per day (bw/d)] and undernourished tional period in sheep. from 60 d before to 30 d after mating (Ϫ60 to ϩ30). Studies investigating the effects of twinning and Undernourishment of the ewes was achieved by withholding maternal nutritional insults on the development of food for 2 d, followed by individually determined concentrate obesity and type 2 diabetes have focused on periph- feeds to induce and sustain a 10–15% reduction in maternal eral metabolic pathways. While this is important, body weight. Initial food intake was at 1–2% bw/d, rising to 80% of that of the controls. After the period of undernour- there is the potential for these insults to act at the ishment, the ewes were fed ad libitum. arcuate nucleus in the hypothalamus, which is the Ultrasound scanning was used to determine fetal number central regulator of energy balance. A likely candi- at 55 d. Fetal samples were collected by catheterization at date for epigenetic changes is the hypothalamic 128 d for plasma analysis. Term in untreated ewes is ϳ147 d, neuropeptide proopiomelanocortin (POMC), which and so fetal tissue was collected at 131 d (twin fetuses) and acts to inhibit food intake and modify glucose han- 135 d (singleton fetuses) by giving a lethal dose of intrave- dling (12). Furthermore, the glucocorticoid receptor nous pentobarbitone to maternal ewes. Subsequently, the fetuses were weighed and dissected. As previously established, (GR) is known to be a regulator of this gene (13–15). the hypothalamus was dissected and frozen to give an arcuate Both these genes have been implicated as targets for nucleus enriched ventral hypothalamic region, using the programming particularly in rodent studies (16–19). third ventricle as a guide (21, 22, 29). The hippocampus was We have chosen to use sheep because their hypothal- dissected by taking the right lateral hippocampal region ami develop prenatally (20), as in humans, and the encompassing the dentate gyrus and CA1, CA2, and CA3 model has been fully validated by our previous regions. The hypothalamic and hippocampal fetal samples studies (21). Also, POMC expression within the from the singleton control and maternally undernourished samples used in this study were investigated previously (21), sheep hypothalamus has previously been shown as but for this comparison with twin samples, separate tissue early as 110 d during gestation (22). sections were used. The POMC region selected is a highly conserved Maternal ewes carrying singletons or twins and subjected to enhancer region that binds to RNA polymerase II and control or maternal undernutrition had similar weights at alters POMC expression in the pituitary (21, 23). The postmortem (Supplemental Table S1). As expected, there was region is also closely associated with the neuronal a reduction in fetal weights in the twins compared with promoter enhancer regions 1 and 2 (nPE1 and nPE2), singletons. known to regulate the expression of hypothalamic POMC (24). The GR region is also highly conserved Bioinformatic analysis and is associated with a CpG island linked to the The POMC and GR marker regions used in this work were transcriptional start sites (25). The interacting histone compared with genomic data from the Encyclopedia of DNA modifications play a vital role in the overall state of Elements (ENCODE) Consortium using the University of Cali- gene activation or repression (26). Key chromatin fornia, Santa Cruz (UCSC; Santa Cruz, CA, USA) genome modifications of gene promoters include histone 3 browser. The ovine promoter sequence for POMC, GR, and lysine 9 acetylation (H3K9AC) and histone 3 lysine 4 OCT4 was located and determined as described previously (21). trimethylation (H3K4me3). These histone states are associated with chromatin opening and gene activation Chromatin immunoprecipitation (ChIP) (27, 28). In addition, histone 3 lysine 27 trimethylation (H3K27me3) has been extensively reported as being ChIP analysis was carried out with the Imprint ChIP kit associated with gene inactivation and compact chroma- (Sigma, St. Louis, MO, USA). To release the chromatin, 20 tin (27). mg of hypothalamic or hippocampal tissue was homogenized and cross-linked with 1% formaldehyde. This was then di- The findings from this investigation show that gested with micrococcal nuclease (2 U/ml; Sigma) to provide twinning and undernutrition in the mother around ϳ500 bp of fragmented genomic DNA. Fragmented chroma- the time of conception alter the epigenetic status of tin was immunoprecipitated with either 1 ␮g of RNA poly- POMC and GR in the hypothalamus. The changes merase 11 (R1530; Sigma) as the positive control, mouse IgG have the potential to result in altered regulation of (M8695; Sigma) as the negative control, or rabbit polyclonal food intake in offspring when they reach adulthood, histone H3K27me3, H3K4me3, and H3K9AC (39917, 39155, leading to the development of obesity. The outcome and 39159, respectively; Active Motif, Carlsbad, CA, USA). Following hydrolyzation of the crosslinks, the DNA was col- of this study presents twinning as a nutritional para- lected and utilized for PCR reactions for POMC and GR. digm, leading to specific epigenetic changes in hypo- OCT4 was used as the control gene for the assay. The PCR thalamic genes centrally involved in the regulation of reactions were set up as described previously (21), and the energy balance. primers used were as follows: GR, 5Ј-TTTGGAGGGACTGTG-

EPIGENETIC CHANGES IN FETAL HYPOTHALAMI 1695 GTCC-3Ј and 5Ј-AGCAGGAGGTGGCAGGCC-3Ј, size ϭ 230 bp; POMC, 5Ј-ACCCTCAGAGGTGAGAAGCT-3Ј and 5Ј-GGAA- GGAGACCGGAGCCG-3Ј, size ϭ 160 bp; OCT4, 5Ј-CCTGGAT- GAGCTTCCAAGG-3Ј and 5Ј-CCTCGGAGTTGCTCTCCCAC- 3Ј, size ϭ 223 bp. The PCR reactions were then resolved on a 2% agarose gel, which was maintained at 80 V for 1 h. ImageJ software (developed by Wayne Rasband, U.S. National Institutes of Health, Bethesda, MD, USA) was then used to quantify the intensity of the bands.

PCR-based methylation analysis

Genomic DNA was isolated from hypothalamic and hippocampal tissue using the AllPrep DNA/RNA kit (Qiagen, Valencia, CA, USA). The MethylCollector kit (Active Motif) was then used to separate methylated genomic DNA from total tissue-specific genomic DNA. The experiment involved digesting 50 ng/␮l of total genomic DNA via an MSE1 digest (New England Biolabs, Beverly, MA, USA) to produce CpG methylated DNA fragments. These fragments, along with the positive and negative control DNA provided in the kit, were incubated with a His-tagged recombinant methyl-CpG binding domain 2b (MBD2b) protein. Nickel-coated magnetic beads were used to capture the His-tagged MBD2b-DNA complexes. The beads were then washed with a high-salt buffer to remove unmethylated fragments and incubated with proteinase K to allow the elution of the methylated DNA. To determine the amount of methylated DNA extracted from the total genomic DNA, PCR analysis was performed for POMC, Figure 1. Summary of ENCODE data from different human GR, and OCT4, as described above. The initial MSE1-digested cell lines. Data analysis from the UCSC genome browser DNA was used as the input value and compared with the depicts cell line-specific changes in chromatin over the amount of tissue-specific methylation enriched DNA as the POMC promoter marker region. A) ChiP:Seq from multiple output value. ENCODE tracks. B) Nucleosome occupancy (28–30).

DNA methyltransferase (DNMT) activity/inhibition assay CCTTCC-3Ј. The reactions were then run on a StepOnePlus Nuclear proteins were extracted from 20 mg hypothalamic Real-Time PCR system thermal cycling block (Applied Biosys- tems, Foster City, CA, USA). mRNA expression was calculated tissue using the EpiQuik nuclear extraction kit (Epigentek Ϫ⌬Ct Group, Farmingdale, NY, USA). The extracts were then according to the 2 method (33). utilized in the EpiQuik DNMT/inhibition assay kit (Epigen- tek). Brieﬂy, the extracts were incubated in wells coated with Statistical analysis a cytosine-rich DNA substrate and Adomet containing methyl groups. 5-Methyl cytosine antibody was added, followed by a All statistical analysis was carried out using GraphPad Prism capture antibody. A developing solution was then applied, software (GraphPad, La Jolla, CA, USA). All data are shown as allowing the absorbance to be measured colorometrically. means Ϯ se. Analysis was carried out via the 2-way ANOVA. This was followed by the Bonferroni post hoc test. Values of P Ͻ POMC ELISA 0.05 were considered signiﬁcantly different.

POMC levels in fetal samples were quantiﬁed by ELISA, as described previously (30, 31). The lower level of sensitivity of RESULTS the POMC ELISA was determined as 10 pM. Selection of the POMC and GR gene promoter mRNA expression analysis marker regions for epigenetic analysis

RNA from hypothalamic and hippocampal tissue was isolated The availability of sequence data in the sheep is limited. using the AllPrep DNA/RNA kit (Qiagen). RNA samples were Nevertheless, the POMC marker region selected shows reverse transcribed with the QuantiTect reverse transcription high homology across all species analyzed (Fig. 1A). We kit (Qiagen). The cDNA was used in a SYBR green mastermix examined the available data from ENCODE, and while (QuantiTect SYBR Green PCR; Qiagen). The primers used in there was no data on hypothalamic cells, ENCODE data the mastermix have been determined previously (32) and are as Ј Ј were available on 3 cell lines (34–37). follows: POMC, forward 5 -GCTGCTGGTCTTGCTGCTTC-3 Multiple sets of ENCODE ChIP sequence (ChIP:seq) and reverse 5Ј-CCTGACACTGGCTCGTCTCC-3Ј; GR, forward 5Ј-ACTGCCCCAAGTGAAAACAGA-3Ј and reverse 5Ј-ATGAA- data demonstrated cell type-specific changes in histone CAGAAATGGCAGACATTTTATT-3Ј; neuropeptide Y (NPY), modifications (Fig. 1A; refs. 37, 38). The presence of forward 5Ј-TCATCACCAGGCAGAGATACGG-3Ј and reverse 5Ј- histone H3K4me3 and H3K3AC was identified in an GAGCAAGTTTCCCATCACC-3Ј; 18S, forward 5Ј-GATGCG- embryonic stem cell line, a breast cell line, and a liver GCGGCGTTATTCC-3Ј and reverse 5Ј-CTCCTGGTGGTGC- cell line, indicating a potentially active enhancer region

1696 Vol. 26 April 2012 The FASEB Journal ⅐ www.fasebj.org BEGUM ET AL. (26). In contrast, other cell lines demonstrated an ison with the control singleton group, H3K27me3 was increase in H3K27me3 (26) or interaction with the found to be markedly decreased in the control twin and transcriptional repressor CTCF, indicating transcrip- underfed twin fetal groups to a level similar to the tionally inactive chromatin (39). The ChIP:seq data singleton underfed group (34, 39, and 31% decrease, also correlated with ENCODE chromatin state segmen- respectively; Fig. 2F, G). The interaction of singleton- tation (CSS) data (40, 41). Open heterochromatin was twin status was also found to be significantly different shown to be present across the POMC marker region in (PϽ0.01). To provide evidence for the role of the HepG2 cells, H1-hESC cells, and HMEC cells, whereas selected POMC region in transcription, RNA polymer- repressed chromatin was present in GM12878 cells and ase II was used as a positive control in the ChIP assay. It K562 cells (Fig. 1A). CSS also highlighted a putative was observed that RNA polymerase II consistently weak enhancer site immediately 5Ј of the POMC bound to the selected POMC region (Fig. 2A). marker region. To establish whether the epigenetic changes we Nucleosome positioning within promoter and en- found were gene specific, OCT4, a transcription factor hancer regions is essential for control of transcriptional necessary for pluripotency, was used as a control gene activity due to regulation of accessibility to the DNA by for methylation analysis (42). OCT4 presented itself as transcription factors (34). Nucleosome occupancy an ideal candidate gene, as it has previously been within the POMC marker region was found in A375 and shown to be hypermethylated in humans (43). The MDA-kb2 cells but not in mammary epithelial cells. This marker chosen to target this gene is situated within the suggests that there is some cell type-dependent variation OCT4 CpG island, with a high level of mammalian in nucleosome positioning at this site (Fig. 1B). conservation (21). The H3K4me3 and H3K27me3 lev- The POMC gene marker region used in this study els of this region remained unchanged across all fetal also colocalized with an enhancer region that has groups (Fig. 2E, H). recently been shown to increase pituitary expression of POMC, with a preference for corticotrope cells (23). The GR marker region showed evidence of distinct Twinning alters DNA methylation levels of the enhancer/promoter activity in all cell lines with avail- hypothalamic POMC region in the fetuses able information within the ENCODE database (Sup- plemental Fig. S1). To further characterize the epigenetic status of the POMC gene as a result of twinning, PCR-associated Histone modifications of the fetal POMC enhancer methylation analysis was carried out. POMC enhancer region in response to twinning and undernutrition in methylation levels in the arcuate nucleus enriched the mother around conception hypothalamic region demonstrated a significant reduction in the control (60%) and twin underfed (73%) twin groups to a level similar to the singleton group To establish a mechanistic insight into the chromatin from mothers underfed around conception (63%; modifications of the hypothalamic POMC gene, mater- Fig. 3A). There was also a significant interaction nal ewes were fed ad libitum or periconceptionally between nutritional status and singleton-twin status undernourished from 60 d before until 30 d during (PϽ0.01). A decrease in POMC promoter methyl- pregnancy to maintain a 10–15% reduction in body ation levels in these groups was also associated with weight. Fetal ventral hypothalami enriched for the a reduction in the overall hypothalamic levels of arcuate nucleus from twins were compared with tissue DNMT activity (Fig. 3C). samples taken from the ventral hypothalami of control To determine whether the methylation changes associ- and maternally undernourished singleton fetuses opti- ated with the POMC promoter were gene specific, we mized previously (21). This allowed us to determine measured OCT4 amplicon methylation status. No reproducibility of analysis of multiple hypothalamic changes were found in the levels of OCT4 marker meth- biopsies. ylation across any of the groups (Fig. 3B), suggesting that Initially, ChIP analysis of the selected POMC region twinning and maternal undernutrition induce gene-spe- was assessed for H3K9AC and H3K4me3 as markers of open chromatin in the hypothalami from singleton and cific alterations. twin fetuses. H3K9AC levels for the POMC amplicon were found to be significantly increased in the control Transcriptional activity of the POMC locus is not twin (43%), underfed twin (40%), and underfed sin- activated at the fetal stage gleton (44%) fetal groups (Fig. 2A, B), compared with the control singletons. Analysis by 2-way ANOVA revealed a significant interaction of singleton-twin status We analyzed mRNA expression of the POMC gene in the (PϽ0.05), in that maternal undernutrition did not ventral hypothalamus by quantitative RT-PCR at the fetal induce a further additive effect to twinning in the levels stage of 131 d to establish whether the epigenetic changes of POMC H3K9AC levels. Another marker for tran- translated into changes in transcriptional activity of scriptional activation, H3K4me3, demonstrated no al- POMC at this early stage. No changes in POMC were terations in POMC for any of the groups (Fig. 2C, D). found in the twins or in the groups from undernourished To further analyze the open or closed status of the mothers (Fig. 3D). Similarly, no changes were found in POMC chromatin, we measured the H3K27me3 levels NPY, analyzed as an alternative neuropeptide with effects as a marker for transcriptional inactivation. In compar- on food intake (Supplemental Fig. S2).

EPIGENETIC CHANGES IN FETAL HYPOTHALAMI 1697 Figure 2. Histone modiﬁcations of the POMC promoter in response to twinning and periconceptional maternal undernutrition. A, C, F) POMC amplicon levels via ChIP analysis as input genomic DNA compared with the enriched output DNA from the singleton control, singleton underfed, twin control, and twin underfed fetal groups for H3K9AC (A), H3K4me3 (C), and H3K27me3 (F). RNA polymerase II was used as positive control; mouse IgG as negative control. Results are representative of nϭ3 experiments. B, D, G) ChIP analysis of hypothalamic POMC promoter H3K9AC (B), H3K4me3 (D), and H3K27me3 (G) in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs, i.e., 14), and twin underfed (nϭ8 pairs) groups. E, H) ChIP analysis of the OCT4 promoter H3K4me3 (E) and H3K27me3 (H) levels in the singleton and twin control and underfed groups. Groups were compared with control via 2-way ANOVA with Bonferroni post hoc test. *P Ͻ 0.05, ***P Ͻ 0.005 for nutritional effect; †P Ͻ 0.05, ††P Ͻ 0.01 for twin effect.

Twinning and nutritional status are associated with site was used (Supplemental Fig. S1). Previously pub- histone modifications of the hypothalamic GR lished data on the sheep GR promoter revealed this region to be highly conserved and CpG dense (21). Because of the well-established influence of GR on Epigenetic interactions with the GR gene in response regulation of food intake, it is a key gene that may be to twinning were assessed by ChIP analysis of H3K9AC, influenced by twinning and maternal undernutrition in H3K4me3, and H3K27me3. The selected GR region the periconceptional period. To study the epigenetic was found to have higher levels of H3K9AC in the status of GR, a marker for the GR gene promoter hypothalami from the control and maternally underfed situated 5 kb upstream of the exon 2 translational start twin groups compared with the control and maternally

1698 Vol. 26 April 2012 The FASEB Journal ⅐ www.fasebj.org BEGUM ET AL. Figure 3. DNA methylation and expression levels of fetal hypothalamic neuropeptides. A) POMC promoter methylation levels, in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups. B) OCT4 methylation levels in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups. C) Overall DNMT activity in the fetal hypothalamus following twinning and periconceptional maternal undernutrition. Levels of DNMT activity were established in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups. D) POMC mRNA expression levels determined using qRT-PCR. Groups were compared with control via 2-way ANOVA with Bon- ferroni post hoc test. **P Ͻ 0.01, ***P Ͻ 0.005 for nutritional effect; †P Ͻ 0.05, ††P Ͻ 0.01 for twin effect.

underfed singleton groups (Fig. 4A, B). Singleton fe- levels in the singleton underfed group (singleton con- tuses from maternal ewes subjected to periconceptional trol nϭ7 and singleton underfed nϭ9; PϽ0.005) but undernutrition also had higher levels of H3K9AC com- not the twin groups (Fig. 5B). pared with singleton controls (singleton controls nϭ7 and singleton underfed nϭ9; PϽ0.005). The interac- Changes in the GR locus are tissue specific and do not tion was established as being significantly different affect the hypothalamic-pituitary-adrenal (HPA) axis (PϽ0.005). Hypothalamic GR H3K4me3 levels were significantly greater in the control twin (29%) and The epigenetic status of the GR marker region in the maternally underfed twin (36%) and underfed single- hippocampus was not affected by twinning or maternal ton (35%) groups (Fig. 4C, D). The levels of variation periconceptional undernutrition (Fig. 6A, B). Follow- between being a singleton or twin was found to be Ͻ ing qRT-PCR, it was observed that the levels of hip- significantly different (P 0.05). The changes in the pocampal GR expression were similar across all groups markers for increased chromatin accessibility were cor- (Fig. 6C). To determine whether the HPA axis activity related with a decrease in GR H3K27me3 levels in the was altered in response to twinning and maternal singleton underfed and control twin groups and the undernutrition, the plasma POMC levels were mea- underfed twin groups (53, 53, and 57% decrease, sured in the control and underfed fetal sheep groups. It respectively; Fig. 4E, F). was found that the levels of POMC were also similar across all groups (Fig. 6D). All the hypothalamic and Twinning and maternal undernutrition around hippocampal data were further analyzed according to conception influence DNA methylation and gender. However, no apparent differences were found transcriptional activity of the GR locus between male and female fetuses (data not shown).

Ventral hypothalamic GR promoter methylation analysis showed that relative to the total genomic DNA, the DISCUSSION levels of methylated GR marker DNA were significantly decreased in the control and underfed twin groups (52 We provide the first evidence of epigenetic changes in and 65% decrease, respectively) and the singleton fetal hypothalamic pathways regulating energy balance underfed group (55% decrease) compared with the as a consequence of twinning and maternal undernu- control singleton group (Fig. 5A). This was associated trition around conception. The importance of these with a significant interaction in singleton-twin status findings lie in the potential long-term consequences for (PϽ0.01). Quantitative (q)RT-PCR analysis also dem- the fetus should it, in postnatal life, be exposed to a onstrated a 5-fold increase in the GR mRNA expression nutritional environment that is mismatched to that

EPIGENETIC CHANGES IN FETAL HYPOTHALAMI 1699 Figure 4. Changes in the histone patterns of the GR promoter as a result of twinning and maternal periconceptional undernutrition. A, C, E) PCR blots following ChIP enrichment for GR promoter H3K9AC (A), H3K4me3 (C), and H3K27me3 (E). RNA polymerase II was used as positive control; mouse IgG as negative control. B, D, F) Levels of fetal hypothalamic H3K9AC (B), H3K4me3 (D), and H3K27me3 (F) of the GR promoter in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups following ChIP analysis. Groups were compared with control via 2-way ANOVA with Bonferroni post hoc test. *P Ͻ 0.05, ***P Ͻ 0.005 for nutritional effect; †P Ͻ 0.05, †††P Ͻ 0.005 for twin effect. experienced in utero, when these changes in pathways of all the epigenetic changes identified suggests that regulating energy balance may lead to an increased the POMC and GR chromatin is open, increasing the propensity to become obese and/or develop type 2 accessibility of transcriptional activators and repressors. diabetes. It is difficult to speculate on the mechanism whereby The association of twinning with abdominal obesity twinning and alterations in maternal diet result in (5, 10) led us to hypothesize that twinning is a nutri- changes in the fetal epigenome. However, we found a tional paradigm that may affect the hypothalamic path- significant reduction in the levels of ventral hypotha- ways involved in energy balance. The overall net effect lamic DNMT activity in twins and in fetuses of pericon-

Figure 5. Effects of twinning and periconceptional maternal undernutrition on fetal hypothalamic GR promoter methylation and GR mRNA expression levels. A) GR promoter methylation levels were determined for the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) fetal groups. B) Fetal hypothalamic GR mRNA analysis in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups established using qRT-PCR. Groups were compared with control via 2-way ANOVA with Bonferroni post hoc test. **P Ͻ 0.01, ***P Ͻ 0.005 for nutritional effect; ††P Ͻ 0.01 for twin effect.

1700 Vol. 26 April 2012 The FASEB Journal ⅐ www.fasebj.org BEGUM ET AL. Figure 6. Fetal HPA axis dynamics following twinning and maternal periconceptional undernutrition. A) Fetal hippocampal GR H3K9AC levels were determined using ChIP analysis for the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups. B) Fetal hippocampal GR methylation levels were established using PCR-based methylation enrichment analysis for the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups. C) Hippocampal GR mRNA expression was quantitated using qRT-PCR in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups. D) Fetal circulating POMC plasma levels in the singleton control (nϭ7), singleton underfed (nϭ9), twin control (nϭ7 pairs), and twin underfed (nϭ8 pairs) groups was measured using an ELISA-based method. Groups were compared with control via 2-way ANOVA with Bonferroni post hoc test. ceptionally undernourished mothers. In addition, alter- opment are understood to be maintained into adult- ations in intermediate metabolites can influence the hood (47). As a result, the observed epigenetic changes activity of histone-associated enzymes (44). Therefore, in the current study would have implications for the it is possible that twinning and maternal undernutrition control of food intake and energy balance after birth, cause changes in fetal metabolites that alter epigenetic when the offspring feeds independently, and any ab- mechanisms impacting on hypothalamic pathways. The normalities would be manifest later in life due to greater implication of these findings is that twin con- age-dependent or pathologically induced changes in ception results in programming of fetuses that result in transcription factor expression. epigenetic changes similar to that of maternal under- In our study, hypomethylation of the hypothalamic nutrition. These data suggest that twins are more likely POMC region did not translate into changes in POMC to develop obesity and consequential diseases later in mRNA expression in the fetuses, as expected. This has life regardless of additional periconceptional maternal also been observed in a study looking at prenatal undernutrition effects. undernutrition in rats (48). Any resulting changes in In focusing on regulation of hypothalamic neuropep- POMC expression due to epigenetic effects may be- tides involved in energy balance, we chose POMC come apparent after birth when they could modulate because it is a necessary component of the hypotha- transcription factor accessibility. For example, in ro- lamic networks, acting to reduce food intake and dents where the mothers were undernourished, the increase energy expenditure (12). There is good evi- postnatal offspring had decreased POMC (16). These dence in rodents for the programming of POMC by data also reinforce the suggestion that due to decreased maternal undernutrition (16), supporting the concept methylation, the increased accessibility of the marker of epigenetic changes. Neonatal overfeeding in rodents region in the POMC promoter could lead to a down- has also identified hypermethylation of the POMC regulation of the POMC gene, causing an increase in promoter (18), and maternal low-protein food restric- food consumption. tion in rats is associated with reduced methylation in POMC neurons in the arcuate nucleus also have specific CpG sites in the POMC promoter in the direct projections to the dorsal vagal complex, and offspring (17). We have also previously shown de- POMC is known to decrease hepatic glucose produc- creased methylation in the POMC promoter in fetal tion and peripheral glucose uptake (49). Therefore, sheep hypothalami (21). Our current data further the once the offspring are feeding independently, the limited evidence of the epigenetic alterations of POMC observed changes in epigenetic status of POMC could in the hypothalamus. The changes suggest that the induce alterations in the regulation of glucose homeo- selected POMC promoter chromatin is open, increas- stasis. Indeed, twins and singletons from mothers un- ing the possibility of regulation of this region and dernourished around conception, which were on the transcription factor binding (45, 46). same regime as the sheep in this study, have impaired Epigenetic methylation patterns during fetal devel- glucose tolerance in adult life (7).

EPIGENETIC CHANGES IN FETAL HYPOTHALAMI 1701 The POMC promoter region chosen for this study is paired glucose tolerance, which was found in late gesta- thought to be associated with the hypothalamic-specific tion fetuses and adult offspring (7, 54). regulatory region (24). Hypomethylation of our chosen In summary, our work demonstrates that twinning and promoter marker region could also reflect epigenetic periconceptional maternal undernutrition induce spe- changes present in this hypothalamic-specific promoter cific epigenetic modifications in the fetal hypothalamic region, altering the control of food intake. It has been pathways regulating energy balance. This suggests that shown that GR binds proximally to our selected POMC twins may undergo a nutritional programming event promoter marker, which is an enhancer region in rats leading to altered physiology of the hypothalamic path- (23). GR is also predicted to bind within the marker ways. As a result, twins and maternally undernourished region used in this work (glucocorticoid response ele- offspring may have an increased propensity to develop ments with Ͻ15% dissimilarity; refs. 50, 51). Therefore, it obesity and/or type 2 diabetes later in life. is interesting to speculate that if a GR-regulated functional enhancer region is epigenetically modified in the arcuate The study was supported by the UK National Institute of nucleus, it may allow GR inhibition of POMC expression, Health Research Manchester Biomedical Research Center, with a concomitant increase in food intake. These the Health Research Council of New Zealand, the New changes would only become apparent after birth, when Zealand National Research Centre for Growth and Develop- ment, and the Canadian Institutes for Health Research (to energy balance is regulated by the hypothalamus. J.R.G.C.). The authors also thank Ngapouri Research Station There is good evidence that glucocorticoids regulate staff and Hui Hui Phua for technical assistance. POMC in the hypothalamus (13, 14). Therefore, changes in the epigenetic status of GR could lead to altered regulation of energy balance increasing the likelihood of the development of obesity. The epigenetic changes in GR found in this REFERENCES study indicate that the chromatin is open, thus allowing for changes in gene expression modulated by the change in transcription factor accessibility. The promoter region is 1. Hales, C. N., and Barker, D. J. (1992) Type 2 (non-insulin- dependent) diabetes mellitus: the thrifty phenotype hypothesis. closely associated with the GR 17 promoter region, which has Diabetologia 35, 595–601 been linked to altered hippocampal GR expression in re- 2. Gluckman, P. D., Hanson, M. A., Cooper, C., and Thornburg, sponse to changes in maternal care (19). Therefore, the K. L. (2008) Effect of in utero and early-life conditions on adult promoter region in this investigation could be reflective of health and disease. N. Engl. J. Med. 359, 61–73 3. Barker, D. J. (2007) The origins of the developmental origins changes in the GR 17 promoter region. theory. J. Intern. Med. 261, 412–417 In the current model, it could be hypothesized that in the 4. Hales, C. N., and Barker, D. J. (2001) The thrifty phenotype adults subjected to maternal programming, GR in the hypo- hypothesis. Br. Med. Bull. 60, 5–20 thalamus modulates neuropeptides to cause an increase in 5. Poulsen, P., Grunnet, L. G., Pilgaard, K., Storgaard, H., Alibe- food intake. However, the precise mechanism for GR regu- govic, A., Sonne, M. P., Carstensen, B., Beck-Nielsen, H., and Vaag, A. (2009) Increased risk of type 2 diabetes in elderly twins. lation of POMC neurons is conflicting with different studies Diabetes 58, 1350–1355 showing an increase or a decrease in POMC expression 6. The, N. S., Adair, L. S., and Gordon-Larsen, P. (2010) A study of following adrenalectomy (which would remove the gluco- the birth weight-obesity relation using a longitudinal cohort and corticoids; refs. 13, 14). It may be that GR regulation of NPY sibling and twin pairs. Am. J. Epidemiol. 172, 549–557 7. Todd, S. E., Oliver, M. H., Jaquiery, A. L., Bloomfield, F. H., and is more relevant, given that there is evidence that NPY Harding, J. E. (2009) Periconceptional undernutrition of ewes requires the presence of glucocorticoids to increase food impairs glucose tolerance in their adult offspring. Pediatr. Res. intake (15). Therefore, the epigenetic alterations in hypo- 65, 409–413 thalamic GR may influence neuropeptide regulation of 8. Buckler, J. M., and Green, M. (2004) A comparison of the early energy balance by decreasing POMC and increasing NPY in growth of twins and singletons. Ann. Hum. Biol. 31, 311–332 9. Van Dommelen, P., de Gunst, M., van der Vaart, A., van Buuren, offspring of mothers that were underfed. S., and Boomsma, D. (2008) Growth references for height, Previous evidence has indicated that maternal under- weight and body mass index of twins aged 0–2.5 years. Acta nutrition induces programming effects in the HPA axis Paediatr. 97, 1099–1104 (52). However, in this investigation, maternal undernutri- 10. Poulsen, P., Vaag, A. A., Kyvik, K. O., Moller Jensen, D., and Beck-Nielsen, H. (1997) Low birth weight is associated with tion before and during early pregnancy did not produce NIDDM in discordant monozygotic and dizygotic twin pairs. any epigenetic changes in the GR marker in the hip- Diabetologia 40, 439–446 pocampus. This is in line with another study that used the 11. Monrad, R. N., Grunnet, L. G., Rasmussen, E. L., Malis, C., Vaag, same levels of maternal undernutrition over the same A., and Poulsen, P. (2009) Age-dependent nongenetic influences of birth weight and adult body fat on insulin sensitivity in period and found similar levels of plasma ACTH and twins. J. Clin. Endocrinol. Metab. 94, 2394–2399 cortisol concentrations (53). This suggests that twinning 12. Brady, L. S., Smith, M. A., Gold, P. W., and Herkenham, M. (1990) and maternal undernutrition in the current model in- Altered expression of hypothalamic neuropeptide mRNAs in food- duce specific hypothalamic programming effects. restricted and food-deprived rats. Neuroendocrinology 52, 441–447 We hypothesize that the epigenetic changes found at 13. Wardlaw, S. L., McCarthy, K. C., and Conwell, I. M. (1998) Glucocorticoid regulation of hypothalamic proopiomelanocor- the fetal stage in this study may persist into adulthood, tin. Neuroendocrinology 67, 51–57 influencing the development of metabolic disease. This is 14. Beaulieu, S., Gagne, B., and Barden, N. (1988) Glucocorticoid because sheep subjected to the same level of maternal regulation of proopiomelanocortin messenger ribonucleic acid undernutrition as those in this study have produced content of rat hypothalamus. Mol. Endocrinol. 2, 727–731 15. Zakrzewska, K. E., Sainsbury, A., Cusin, I., Rouru, J., Jeanre- offspring that are overweight compared with control naud, B., and Rohner-Jeanrenaud, F. (1999) Selective depen- offspring at 10 mo of age (7). Furthermore, twins and dence of intracerebroventricular neuropeptide Y-elicited effects singletons from undernourished mothers also have im- on central glucocorticoids. Endocrinology 140, 3183–3187

1702 Vol. 26 April 2012 The FASEB Journal ⅐ www.fasebj.org BEGUM ET AL. 16. Delahaye, F., Breton, C., Risold, P. Y., Enache, M., Dutriez- methods for large-scale structural analysis of mammalian chro- Casteloot, I., Laborie, C., Lesage, J., and Vieau, D. (2008) matin. Genome Res. 17, 928–939 Maternal perinatal undernutrition drastically reduces postnatal 36. Ozsolak, F., Song, J. S., Liu, X. S., and Fisher, D. E. (2007) leptin surge and affects the development of arcuate nucleus High-throughput mapping of the chromatin structure of hu- proopiomelanocortin neurons in neonatal male rat pups. Endo- man promoters. Nat. Biotechnol. 25, 244–248 crinology 149, 470–475 37. Bernstein, B. E., Mikkelsen, T. S., Xie, X., Kamal, M., Huebert, D. J., 17. Coupe, B., Amarger, V., Grit, I., Benani, A., and Parnet, P. Cuff, J., Fry, B., Meissner, A., Wernig, M., Plath, K., Jaenisch, R., (2010) Nutritional programming affects hypothalamic organi- Wagschal, A., Feil, R., Schreiber, S. L., and Lander, E. S. (2006) A zation and early response to leptin. Endocrinology 151, 702–713 bivalent chromatin structure marks key developmental genes in 18. Plagemann, A., Harder, T., Brunn, M., Harder, A., Roepke, K., embryonic stem cells. Cell 125, 315–326 Wittrock-Staar, M., Ziska, T., Schellong, K., Rodekamp, E., 38. Bernstein, B. E., Kamal, M., Lindblad-Toh, K., Bekiranov, S., Bailey, Melchior, K., and Dudenhausen, J. W. (2009) Hypothalamic D. K., Huebert, D. J., McMahon, S., Karlsson, E. K., Kulbokas, E. J., 3rd, proopiomelanocortin promoter methylation becomes altered Gingeras, T. R., Schreiber, S. L., and Lander, E. S. (2005) Genomic by early overfeeding: an epigenetic model of obesity and the maps and comparative analysis of histone modifications in human metabolic syndrome. J. Physiol. 587, 4963–4976 and mouse. Cell 120, 169–181 19. Weaver, I. C., Cervoni, N., Champagne, F. A., D’Alessio, A. C., 39. Ohlsson, R., Bartkuhn, M., and Renkawitz, R. (2010) CTCF Sharma, S., Seckl, J. R., Dymov, S., Szyf, M., and Meaney, M. J. shapes chromatin by multiple mechanisms: the impact of 20 (2004) Epigenetic programming by maternal behavior. Nat. years of CTCF research on understanding the workings of Neurosci. 7, 847–854 chromatin. Chromosoma 119, 351–360 20. Symonds, M. E., and Budge, H. (2009) Nutritional models of 40. Ernst, J., and Kellis, M. (2010) Discovery and characterization of the developmental programming of adult health and disease. chromatin states for systematic annotation of the human ge- Proc. Nutr. Soc. 68, 173–178 nome. Nat. Biotechnol. 28, 817–825 21. Stevens, A., Begum, G., Cook, A., Connor, K., Rumball, C., 41. Ernst, J., Kheradpour, P., Mikkelsen, T. S., Shoresh, N., Ward, Oliver, M., Challis, J., Bloomfield, F., and White, A. (2010) L. D., Epstein, C. B., Zhang, X., Wang, L., Issner, R., Coyne, M., Epigenetic changes in the hypothalamic proopiomelanocortin Ku, M., Durham, T., Kellis, M., and Bernstein, B. E. (2011) and glucocorticoid receptor genes in the ovine fetus after Mapping and analysis of chromatin state dynamics in nine periconceptional undernutrition. Endocrinology 151, 3652–3664 human cell types. Nature 473, 43–49 22. Muhlhausler, B. S., McMillen, I. C., Rouzaud, G., Findlay, P. A., 42. Surani, M. A., Hayashi, K., and Hajkova, P. (2007) Genetic and Marrocco, E. M., Rhind, S. M., and Adam, C. L. (2004) Appetite epigenetic regulators of pluripotency. Cell 128, 747–762 regulatory neuropeptides are expressed in the sheep hypothal- 43. Schneider, E., Pliushch, G., El Hajj, N., Galetzka, D., Puhl, A., amus before birth. J. Neuroendocrinol. 16, 502–507 Schorsch, M., Frauenknecht, K., Riepert, T., Tresch, A., Muller, A. M., 23. Langlais, D., Couture, C., Sylvain-Drolet, G., and Drouin, J. (2011) A Coerdt, W., Zechner, U., and Haaf, T. (2010) Spatial, temporal and pituitary-specific enhancer of the POMC gene with preferential activ- interindividual epigenetic variation of functionally important DNA ity in corticotrope cells. Mol. Endocrinol. 25, 348–359 methylation patterns. Nucleic Acids Res. 38, 3880–3890 24. De Souza, F. S., Santangelo, A. M., Bumaschny, V., Avale, M. E., 44. Turner, B. M. (2009) Epigenetic responses to environmental Smart, J. L., Low, M. J., and Rubinstein, M. (2005) Identification change and their evolutionary implications. Philos. Trans. R. Soc. of neuronal enhancers of the proopiomelanocortin gene by Lond. B Biol. Sci. 364, 3403–3418 transgenic mouse analysis and phylogenetic footprinting. Mol. 45. Newell-Price, J. (2003) Proopiomelanocortin gene expression Cell. Biol. 25, 3076–3086 and DNA methylation: implications for Cushing’s syndrome and 25. Geng, C. D., Schwartz, J. R., and Vedeckis, W. V. (2008) A beyond. J. Endocrinol. 177, 365–372 conserved molecular mechanism is responsible for the auto-up- 46. Ho, S. M., and Tang, W. Y. (2007) Techniques used in studies of regulation of glucocorticoid receptor gene promoters. Mol. epigenome dysregulation due to aberrant DNA methylation: an Endocrinol. 22, 2624–2642 emphasis on fetal-based adult diseases. Reprod. Toxicol. 23, 267–282 26. Portela, A., and Esteller, M. (2010) Epigenetic modifications 47. Razin, A., and Cedar, H. (1993) DNA methylation and embryo- and human disease. Nat. Biotechnol. 28, 1057–1068 genesis. EXS 64, 343–357 27. Barski, A., Cuddapah, S., Cui, K., Roh, T. Y., Schones, D. E., 48. Ikenasio-Thorpe, B. A., Breier, B. H., Vickers, M. H., and Fraser, Wang, Z., Wei, G., Chepelev, I., and Zhao, K. (2007) High- M. (2007) Prenatal influences on susceptibility to diet-induced resolution profiling of histone methylations in the human obesity are mediated by altered neuroendocrine gene expres- genome. Cell 129, 823–837 sion. J. Endocrinol. 193, 31–37 28. Berger, S. L. (2007) The complex language of chromatin 49. Marino, J. S., Xu, Y., and Hill, J. W. (2011) Central insulin and regulation during transcription. Nature 447, 407–412 leptin-mediated autonomic control of glucose homeostasis. 29. Stevens, A., Begum, G., and White, A. (2011) Epigenetic Trends Endocrinol. Metab. 22, 275–285 changes in the hypothalamic pro-opiomelanocortin gene: a 50. Messeguer, X., Escudero, R., Farre, D., Nunez, O., Martinez, J., mechanism linking maternal undernutrition to obesity in the and Alba, M. M. (2002) PROMO: detection of known transcrip- offspring? Eur. J. Pharmacol. 660, 194–201 tion regulatory elements using species-tailored searches. Bioin- 30. Schwartz, J., and Rose, J. C. (1998) Development of the pituitary formatics 18, 333–334 adrenal axis in fetal sheep twins. Am. J. Physiol. 274, R1–R8 51. Farre, D., Roset, R., Huerta, M., Adsuara, J. E., Rosello, L., Alba, 31. Crosby, S. R., Stewart, M. F., Ratcliffe, J. G., and White, A. (1988) M. M., and Messeguer, X. (2003) Identification of patterns in Direct measurement of the precursors of adrenocorticotropin biological sequences at the ALGGEN server: PROMO and in human plasma by two-site immunoradiometric assay. J. Clin. MALGEN. Nucleic Acids Res. 31, 3651–3653 Endocrinol. Metab. 67, 1272–1277 52. Edwards, L. J., and McMillen, I. C. (2002) Impact of maternal 32. Sebert, S. P., Hyatt, M. A., Chan, L. L., Patel, N., Bell, R. C., undernutrition during the periconceptional period, fetal num- Keisler, D., Stephenson, T., Budge, H., Symonds, M. E., and ber, and fetal sex on the development of the hypothalamo- Gardner, D. S. (2009) Maternal nutrient restriction between pituitary adrenal axis in sheep during late gestation. Biol. Reprod. early and midgestation and its impact upon appetite regulation 66, 1562–1569 after juvenile obesity. Endocrinology 150, 634–641 53. Rumball, C. W., Bloomfield, F. H., Oliver, M. H., and Harding, 33. Livak, K. J., and Schmittgen, T. D. (2001) Analysis of relative J. E. (2009) Different periods of periconceptional undernutri- gene expression data using real-time quantitative PCR and the tion have different effects on growth, metabolic and endocrine 2[-delta delta C(T)] method. Methods 25, 402–408 status in fetal sheep. Pediatr. Res. 66, 605–613 34. Gupta, S., Dennis, J., Thurman, R. E., Kingston, R., Stamatoy- 54. Rumball, C. W., Harding, J. E., Oliver, M. H., and Bloomfield, F. H. annopoulos, J. A., and Noble, W. S. (2008) Predicting human (2008) Effects of twin pregnancy and periconceptional undernutri- nucleosome occupancy from primary sequence. PLoS Comput. tion on maternal metabolism, fetal growth and glucose-insulin axis Biol. 4, e1000134 function in ovine pregnancy. J. Physiol. 586, 1399–1411 35. Dennis, J. H., Fan, H. Y., Reynolds, S. M., Yuan, G., Meldrim, J. C., Richter, D. J., Peterson, D. G., Rando, O. J., Noble, W. S., Received for publication October 19, 2011. and Kingston, R. E. (2007) Independent and complementary Accepted for publication December 19, 2011.

EPIGENETIC CHANGES IN FETAL HYPOTHALAMI 1703 4.1 Supplemental data

Fig. S1. Summary of ENCODE data from different human cell lines demonstrating cell line specific changes in chromatin over the GR promoter marker region. Data analysis from UCSC genome browser; nucleosome occupancy; ChiP:Seq from multiple ENCODE tracks.

Fig. S2. NPY mRNA expression levels determined using qRT-PCR. Groups were compared to the control via the one-way ANOVA with Tukey HSD Post Hoc test. *p<0.05, **p<0.01, ***p<0.005.

Table S1. Maternal (kg) and fetal (kg) weights in singletons and twins. Data is shown as the mean ± S.E.M. Results were analysed by two way ANOVA with Bonferroni post hoc tests. ††† p<0.005 comparing singleton and twin fetal weights at postmortem in controls. *** p<0.005 comparing singleton and twin fetal weights at post-mortem in the maternally undernourished groups.

Chapter 5: Publication III

Maternal undernutrition programs tissue-specific epigenetic changes in the glucocorticoid receptor in adult offspring

Ghazala Begum, Alison Davies, Adam Stevens, Mark Oliver, Anne Jaquiery, John Challis, Jane Harding, Frank Bloomfield & Anne White Faculty of Life Sciences (G.B, A.D, A.W), Faculty of Medical and Human Sciences (A.S, A.W), University of Manchester, UK. Liggins Institute (M.O, A.J, J.H, F.B), University of Auckland, New Zealand. Gravida: National Centre for Growth and Development (M.O, A.J, F.B), New Zealand. Department of Paediatrics: Child and Youth Health (A.J, F.B), University of Auckland, New Zealand. Department of Physiology (J.C), University of Toronto, Canada

Endocrinology

Endocrinology. First published ahead of print September 24, 2013 as doi:10.1210/en.2013-1693

ENERGY BALANCE-OBESITY

Maternal undernutrition programs tissue-specific epigenetic changes in the glucocorticoid receptor in adult offspring

Ghazala Begum1, Alison Davies1, Adam Stevens2, Mark Oliver3,4, Anne Jaquiery3,4,5 John Challis6, Jane Harding3, Frank Bloomfield3,4,5 , and Anne White1,2

1Faculty of Life Sciences, 2Faculty of Medical and Human Sciences, University of Manchester, UK, M13 9PT. 3Liggins Institute, University of Auckland, New Zealand. 4Gravida: National Centre for Growth and Development, New Zealand. 5Department of Paediatrics: Child and Youth Health, University of Auckland, New Zealand. 6Department of Physiology, University of Toronto and Faculty of Health Sciences, Simon Fraser University Vancouver, Canada.

Epidemiological data indicate that an adverse maternal environment during pregnancy predisposes offspring to metabolic syndrome with increased obesity, and type 2 diabetes. The mechanisms are still unclear although epigenetic modifications are implicated and the hypothalamus a likely target. We hypothesised that maternal undernutrition around conception (UN) in sheep would lead to epigenetic changes in hypothalamic neurones regulating energy balance in the offspring, up to five years after the maternal insult. We found striking evidence of decreased glucocorticoid receptor (GR) promoter methylation, decreased H3K27 trimethylation and increased H3K9 acetylation in hypothalami from male and female adult offspring of UN mothers. These findings are entirely compatible with the increased GR mRNA and protein observed in the hypothalami. The increased GR predicted the decreased hypothalamic pro-opiomelanocortin expression and increased obesity we observed in the 5 year old adult males. The epigenetic and expression changes in GR were specific to the hypothalamus. Hippocampal GR mRNA and protein were decreased in UN offspring, whereas pituitary GR was altered in a sex-specific manner. In peripheral polymorphonuclear leukocytes there were no changes in GR methylation or protein, indicating that this epigenetic analysis did not predict changes in the brain. Overall, these results suggest that moderate changes in maternal nutrition, around the time of conception, signal life- long and tissue-specific epigenetic alterations in a key gene regulating energy balance in the hypothalamus.

he overwhelming increase in obesity in human popu- leading to a metabolic phenotype. The contribution of Tlations is undoubtedly occurring because of both com- epigenetic changes to underlying mechanisms is now plex genetic traits and environmental factors (1). Addi- emerging. Given that the hypothalamus is the main site for tionally, there are important influences from both paternal regulation of energy balance, it is a likely target for epi- and maternal environments (2, 3). Epidemiological stud- genetic alterations that predispose to the development of ies demonstrate that early life events experienced in utero obesity and type 2 diabetes (4). Indeed, most of the mu- increase the incidence of obesity and type 2 diabetes (4, 5). tations causing monogenetic obesity are in genes ex- The most influential studies have been on the Dutch fam- pressed in the hypothalamus, including pro-opiomelano- ine, clearly demonstrating that maternal undernutrition cortin (POMC) and the leptin receptor (6, 7). during pregnancy has a profound effect on the offspring, Consequently, programming studies have investigated the

doi: 10.1210/en.2013-1693 Endocrinology endo.endojournals.org 1

Copyright (C) 2013 by The Endocrine Society 2 Programming tissue specific changes in GR Endocrinology effects of maternal nutritional insults on hypothalamic modifications in GR in the hippocampus and pituitary, to neuropeptide expression, with associated changes in food determine if the changes are widespread or restricted to the intake and the development of obesity in the offspring hypothalamus. This led us to investigate whether GR in (8–12). There is considerable evidence that glucocortico- noninvasive samples such as peripheral circulating leuko- ids regulate neuropeptides and that this is key to control- cytes, could be utilized as an epigenetic biomarker of ad- ling energy balance (13–15). The glucocorticoid receptor verse maternal programming. (GR) is highly expressed in many tissues including the arcuate nucleus of the hypothalamus and is known to be subject to very specific epigenetic changes. For example, Materials and Methods both altered maternal care and child abuse have been shown to induce defined changes in the epigenetic status of Animal management. All experiments were approved by the hippocampal GR in rats and humans, with associated University of Auckland animal ethics committee. Our model of periconceptional undernutrition has been extensively described changes in GR expression levels (16, 17). Taken together and validated in detail in previous publications (24–27). Briefly, these studies suggest that hypothalamic GR could be an for this investigation Romney ewes at 5 years of age were as- ideal candidate for epigenetic targeting inducing long- signed normal feed (control) or their feed was altered to allow for term modifications in the regulation of energy balance, a 10%–15% reduction in bodyweight (UN). This was com- predisposing the adult offspring to obesity and diabetes. menced 60 days before conception and continued up until 30 days of gestation. Lambs were born at term (approximately 148 However, the long-term impact of programming on GR in days) and only singleton offspring were carried through the hypothalamic neurones regulating energy balance in the study. Offspring were euthanized with pentobarbitone at ap- offspring remains to be determined. proximately 5 years of age and brain tissues and blood samples Previously, we used a well-established sheep model to were taken immediately. Hypothalamic dissection occurred to study epigenetic mechanisms associated with program- isolate the ventral hypothalamic region containing the arcuate ming at the fetal stage. This model was chosen as hypo- nucleus at coordinates 31.56–33.06. The right lateral hippocampal area was dissected containing the CA1, CA2 and CA3 and thalamic maturation in sheep follows a similar develop- dentate gyrus regions. mental trajectory to that of the human, but differs to that in the rat and mouse (18). Sheep are also more likely to Methylation enrichment analysis. Methylation enrichment have singleton pregnancies then polytocous rodents and of the chosen GR amplicon from total genomic DNA isolated previously we have shown that having more than one off- from brain tissues and leukocytes was carried out as described spring may alter and interfere with fetal development, previously using the MethylCollector kit (Active Motif, La which could skew the data and any conclusions from it Hulpe, Belgium) (19, 21). (19). Furthermore, our model focuses on phenotypic al- Chromatin Immunoprecipitation (ChIP). ChIP experi- terations in the offspring as a consequence of maternal ments for H3K9 acetylation and H3K27me3 were performed undernutrition around the time of conception. Previous according to previous publications (16, 28). In summary, 50 mg studies have shown that this period is associated with an of dissected brain tissue was sonicated while on ice. Purified increased chance of these offspring developing obesity in chromatin was then immunoprecipitated with either an antibody adulthood (4, 20). We found that maternal undernutrition against H3K9ac (39917; Active Motif, La Hulpe, Belgium) or H3K27me3 (39155; Active Motif, La Hulpe, Belgium). RNA can induce epigenetic modifications and associated polymerase II (R1530 Sigma, Dorset, UK) and normal mouse changes in GR gene expression in hypothalamic regions IgG (M8695 Sigma, Dorset, UK) were used as positive and neg- involved in regulating energy balance (19, 21). However, ative controls respectively. Cross links were then hydrolyzed to whether these epigenetic changes persist long-term into allow DNA elution. Input digested gDNA and output methyl- adulthood to influence pathogenic effects on metabolism ated DNA was then analyzed by using the following primers: GR, Ј Ј Ј is unclear. Indeed, offspring in our model have impaired 5 -TTTGGAGGGACTGTGGTCC-3 and 5 -AGCAG- GAGGTGGCAGGCC-3Ј , size, 230 bp; POMC, 5Ј -AC- glucose tolerance that increases in severity with age (22) CCTCAGAGGTGAGAAGCT-3Ј and 5Ј - and adult males exhibit decreased physical activity (23) GGAAGGAGACCGGAGCCG-3Ј, size, 160 bp. The reactions and increased fat mass (24). We hypothesized that these were run on an agarose gel. This was followed by densitometric metabolic effects in adult offspring are associated with analysis of the bands using image J (developed by Wayne Ras- altered epigenetic status of the GR gene in the hypothal- band, U.S. National Institutesof Health, Bethesda, MD, USA). amus. Therefore, we have carried out detailed examina- All experiments were carried out in triplicate. tion of the epigenetic status of the GR gene in the hypo- mRNA expression analysis. RNA for qRT-PCR analysis was thalamus and investigated changes in gene and protein purified from brain tissues using the Allprep DNA/RNA kit (Qia- expression in adult sheep up to 5 years of age. Since GR has gen, West Sussex, UK). RNA samples then underwent reverse a well-established role in the HPA axis, we have examined transcription and PCR with the power SYBR green RNA to Ct doi: 10.1210/en.2013-1693 endo.endojournals.org 3

1-step kit (Life Technologies, Paisley, UK) for GR, NPY, HPRT 1E has been shown to be transcribed in the hippocampus, and 18S. For POMC a 5Ј nuclease assay was designed utilizing pituitary and regions of the hypothalamus, with further a FAM reporter dye, and Quantitect probe one step qRT-PCR clear evidence of expression in lymphocytes (30, 34). Ad- reagents (Qiagen, West Sussex, UK). GR, POMC and NPY levels were determined as relative to the reference gene HPRT or 18S. ditionally, evidence for transcription factor binding de- Sequences for the primers are as follows: rived from many different cell lines (GM12878, H1-hESC, GR forward: 5Ј-GGACCACCTCCCAAACTCTG-3Ј GR re- HepGC, HMEC, HSMM, HUVEC, K562, NHEK, verse: 5ЈGCTGTCCTTCCACTGCTCTT-3Ј;NPY forward: 5Ј- NHLF; (32)) as identified by ChIP-seq shows strong evi- Ј Ј CGGAGGACTTGGCCAGATAC-3 NPY reverse: 5 - dence for polymerase 2 (POL2) association with the GR CTCGGGGCTAGATCGTTTCC-3Ј; POMC forward: 5Ј- marker region (conducted as part of the ENCODE (En- GCTACGGCGGGTTCATGA-3Ј POMC reverse: 5Ј- TTCTTGATGATGGCGTTTTTGA-3Ј POMC probe: 5Јfam- cyclopedia of DNA Elements) study AGCCAAACGCCCCTTGTCACGC-methyl red3Ј; HPRT (http://genome.ucsc.edu/ENCODE/)). forward: 5Ј-TTTATTCCTCATGGACTAATTATGGA-3Ј HPRT reverse: 5Ј-GCCACCCATCTCCTTCATCAC-3Ј; 18s GR promoter methylation and histone Ј Ј forward: 5 -GATGCGGCGGCGTTATTCC-3 18s reverse: modifications 5Ј-CTCCTGGTGGTGCCCTTCC-3Ј. Due to the long-term implications of methylation changes on the silencing of the gene, we began by deter- Western blot. GR protein levels for brain tissue and leukocytes were determined according to standard western blotting proce- mining the level of GR methylation of our chosen pro- dures. In summary, protein was isolated from the dissected hy- moter region by methylation enrichment. Within the ven- pothalami, hippocampi and pituitary tissue using 1 x RIPA buf- tral hypothalamus, we found that in the arcuate nucleus fer supplemented with protease inhibitors (Roche, Welwyn enriched GR there was decreased promoter methylation in Garden City, UK) and protein content was determined using the the UN offspring compared to controls (Figure 1a, 1b; Bradford assay. Resultant protein samples were then subjected to Ͻ Ͻ SDS-PAGE and western blotting procedures. Proteins of interest males 25% decrease, P .001, females 31% decrease, P were detected using the following antibodies: mouse anti-GR .002). (BD biosciences, Oxford, UK; 1 in 2000), NPY (Sigma, Dorset, Chromatin immunoprecipitation (ChIP) analysis to de- UK; 1 in 2000), POMC (in house A1A12 (29); 1 in 5000) anti termine histone changes revealed an associated increase in ␤ ␣ -actin (Abcam, Cambridge, UK; 1 in 5000) and anti -tubulin H3K9 acetylation (a marker of transcriptional activation) (Santa Cruz, Texas, U.S.A; 1 in 5000). (Figure 1c, 1d; H3K9AC; males 51% increase P Ͻ .0001, Ͻ Statistical analysis. Data were analyzed using GraphPad females 28% increase P .009), and decreased H3K27 Prism software (GraphPad, La Jolla, CA, USA). The results were trimethylation (a marker of transcriptional inactivation) tested for significance using the two-sample T test, with a Mann- in the offspring maternally undernourished compared to Whitney U test, where values of P Ͻ .05 are statistically different. controls (Figure 1e, 1f; H3K27me3; males 33% decrease Ϯ All data are represented as mean s.e.m. P Ͻ .01, females 57% decrease P Ͻ .001).

Hypothalamic GR mRNA and protein in the adult Results offspring To examine the potential of GR epigenetic changes im- Analysis of epigenetic status of hypothalamic GR pacting on GR expression changes, we measured mRNA in UN adult offspring and protein levels in the adult offspring. We found that the GR promoter analysis epigenetic alterations of the GR 1E promoter region were To examine the epigenetic and expression status of GR associated with increased GR mRNA and protein levels in in the hypothalami from 5 year old adult offspring, we the UN groups compared to their matched controls (Fig- required detailed data on the exon 1 region of the sheep ure 2a-d; mRNA expression: males 28% increase P Ͻ .03, GR gene. This region consists of 11 possible tissue specific females 38% increase P Ͻ .03; protein: males 41% in- variants, each regulated by their own promoter region crease P Ͻ .01, females 52% increase P Ͻ .01). (30). For the epigenetic analysis, we chose the promoter of the exon 1E region of the GR gene, which is highly con- Neuropeptide expression in hypothalami from served across species and is CpG rich (30). The chosen adult offspring region is located 5 kb upstream of the exon 2 translational The altered GR expression changes led us to hypothe- start site and is associated with an active promoter region size that the increased GR could be impacting on neuro- (19, 21, 31, 32). Furthermore, the region has an Sp3 bind- peptide regulation and expression in hypothalamic neu- ing site and a GRU (glucocorticoid response unit), which rones regulating energy balance. We therefore examined would allow GR auto-upregulation (33). Recently, exon neuropeptide expression within the hypothalamus of 4 Programming tissue specific changes in GR Endocrinology adult offspring. Data on the sheep genome were limited (data not shown), but the transport of the peptides to and, while sequences for POMC and NPY were available, dendritic processes at distant sites may mask any specific the sequence for AgRP was incomplete, based on the ge- differences. nome framework available. In the hypothalami, there was a decrease in POMC expression in the UN males compared GR in other tissues from the offspring of UN to controls (55% decrease P Ͻ .03; Figure 3a) but no mothers change in NPY mRNA (Figure 3c, 3d). There were no Hippocampus significant changes in POMC or NPY mRNA expression To determine if maternal undernutrition around con- between females from UN and controls. POMC and NPY ception induces the same or different changes in GR in protein expression were not different in any of the groups other regions of the brain, we determined the epigenetic and expression status of GR in the hippocampus and pituitary. We have previously shown that there were no changes in epigenetic status or mRNA levels of GR in the hippocampi or pituitaries at the fetal stage (19, 21). How- ever, in the adult female offspring of UN ewes there was an increase in hippocampal GR promoter methylation (Fig- ure 4b; 50% increase P Ͻ .02) and a decrease in H3K9AC compared with control female offspring (Figure 4d; 25% decrease P Ͻ .02). These changes were not present in males (Figure 4a, 4c). GR H3K27me3 was increased in both female and male UN offspring compared with controls (Figure 4e, 4f; males 114% increase P Ͻ .02, females 37% increase P Ͻ .04). As predicted by these data, there were decreases in GR mRNA (Figure 4g, 4h; males 28% decrease P Ͻ .003, females 27% decrease P Ͻ .03) and protein expression (Figure 4i, 4j; males 37% decrease P Ͻ .01, females 64% decrease P Ͻ .02) in UN offspring hippocampi compared to controls. Similar changes in hippocampal GR expression have been identified in other programming paradigms (35, 36).

Pituitary In the pituitary, we found sex-specific changes in the epigenetic status of GR (Table 1). These were associated with a decrease in GR mRNA and protein expression in female offspring and an increase in male offspring of UN mothers (mRNA: males 125% increase P Ͻ .03, females 34% decrease P Ͻ .01; protein expression: males 163% increase P Ͻ .03, females 43% decrease P Ͻ .04). These changes imply tissue and sex specific variations in epigenetic modifications as a consequence of maternal Figure 1. Periconceptional undernutrition is associated with altered undernutrition. glucocorticoid receptor (GR) epigenetic status in the ventral hypothalamus of adult offspring. (a,b) GR exon 1E promoter methylation enrichment in the ventral hypothalamus of control male Leukocytes (n ϭ 6) and UN male (n ϭ 8), control female (n ϭ 8) UN female (n ϭ Other studies have suggested that changes in the brain 7) offspring. Chromatin immunoprecipitation (ChIP) enrichment for (c, in key genes involved in appetite regulation can also be d) Histone H3 lysine 9 acetylation (H3K9AC) and (e, f) Histone lysine 27 trimethylation (H3K27me3) of the chosen GR amplicon in the control found in leukocytes (37). Therefore, we investigated male (n ϭ 6) and UN male (n ϭ 8), control female (n ϭ 8) UN female whether the changes we had identified in the hypothalami (n ϭ 7), with representative blots (CI and UI as control and from the offspring could be predicted from assessment of undernourished total input gDNA and CO and UO as control and undernourished output). Data are shown as meanϮs.e.m *P Ͻ .05, GR in leukocytes. Importantly, we found no alteration in **P Ͻ .01, ***P Ͻ .005. GR promoter methylation and no change in GR protein doi: 10.1210/en.2013-1693 endo.endojournals.org 5 levels between the UN and control groups in the white Discussion blood cells (Figure 5). This result indicates that the epigenetic modifications in the regions in the brain studied This study provides evidence that epigenetic changes pro- could not have been predicted based on the results from grammed in fetal life persist in adult sheep up to five years assessment of GR epigenetic or expression status in of age. These epigenetic changes occurred as a result of leukocytes. maternal undernourishment around the time of conception, with increased expression of the glucocorticoid re- Physiological changes in the male offspring from ceptor in the hypothalami of mature adult sheep. The epi- undernourished mothers genetic changes are associated with alterations in gene The most relevant indicator of obesity is the ratio of fat expression and are tissue-specific, suggesting this is a co- to lean body mass and therefore we investigated this by ordinated response. The maternal programming of GR in DXA (dual-emission X-ray absorptiometry) scanning at the hypothalamus is entirely consistent with the changes the termination of the study. The fat:lean mass ratio was seen in POMC, a key neuropeptide regulating food intake significantly increased in UN males compared to controls and energy balance, and with the marked phenotypic ef- (Figure 6a), consistent with the changes in hypothalamic fects of increased fat mass in the five year old (mature POMC expression we found in males. This change was adult) UN male sheep. only present in the males and not the females. The sheep There is considerable evidence from epidemiological paradigm of maternal periconceptional undernutrition studies in humans that maternal undernutrition impacts used in this study has previously been shown to result in on the fetus to increase the risk of metabolic disease in increased body weight in undernourished offspring at 10 adult offspring (1, 5). However, the mechanisms involved months of age (22). have not been fully delineated. Given the importance of the hypothalamus in regulating energy balance it is likely that this tissue is a key target for programming and glucocorticoids and their receptors are known regulators of energy balance in the hypothalamus. This is relevant because GR has previously been shown to be modified by programming and in particular there were specific changes in the epigenetic status of the GR in the hippocampus of rats and humans, as a result of altered maternal care and child abuse (16, 17). These studies also found associated changes in GR expression levels. In other programming paradigms, maternal protein restriction in rats has resulted in offspring with a substantial decrease in methylation of GR and an increase in GR expression in the liver (38). In the current study, exposure to maternal undernutrition has resulted in coordinated epigenetic changes in the GR promoter in the arcuate region of the hypothalamus in that there was decreased GR promoter methylation and increased GR-associated Figure 2. Altered glucocorticoid receptor (GR) expression levels in the ventral hypothalamus of adult offspring following periconceptional undernutrition. (a,b) Ventral hypothalamic GR mRNA H3K9AC and decreased GR-associ- expression levels by qPCR in the offspring of control and periconceptionally undernourished ated H3K27me3. Each of these ewes. (c,d) Western blot analysis of ventral hypothalamic GR protein levels. Data are for control changes would be predicted to in- male (n ϭ 6), UN male (n ϭ 8), control female (n ϭ 8) and UN female (n ϭ 7) adult offspring and are shown as mean Ϯ s.e.m *P Ͻ .05, **P Ͻ .01, ***P Ͻ .005. crease GR mRNA expression. The 6 Programming tissue specific changes in GR Endocrinology

mechanism for this is beyond the scope of this study but since these changes are associated with open chromatin and there is a GRU binding site in exon 1E, there is the potential for GR to up-regulate its own gene expression (33). However, these results do not prove a causal effect and there may be additional mechanisms that lead to the increased GR expression observed. Indeed, the methylation enrichment technique used in this study is more effective at isolating large areas of methylation and, therefore, more subtle changes in individual CpGs may have been difficult to detect and may contribute to the increases in GR expression. One of the most significant outcomes of this study is that all of the epigenetic changes and the resultant increases in GR expression in the hypothalamus, which were first observed during fetal life, were maintained in adults up to 5 years of age, a time representing late middle age in these sheep. In other brain regions, epigenetic effects found in fetal life (19, 21) did not persist into adult life, again emphasizing the enduring nature of this effect in the Figure 3. Ventral hypothalamic neuropeptide mRNA expression. Pro- hypothalamus. Furthermore, in fetal life the epigenetic opiomelanocortin (POMC) and neuropeptide Y (NPY) mRNA expression and gene expression changes were not associated with al- levels were analyzed by qPCR in control males (n ϭ 6), UN males (n ϭ 8), control females (n ϭ 8) and UN females (n ϭ 7). Results are tered gene expression of POMC which is consistent with represented as mean Ϯ s.e.m. *P Ͻ .05 fetal nutrition being regulated by placental supply,

Figure 4. Altered epigenetic and expression status of hippocampal glucocorticoid receptor (GR) epigenetic and expression status in offspring from undernourished mothers. (a,b) Methylation analysis of the GR exon 1E promoter in the hippocampus. (c,d) Histone Chromatin immunoprecipitation enrichment for Histone H3 lysine 9 acetylation (H3K9AC) and (e, f) Histone lysine 27 trimethylation (H3K27me3) of the GR promoter region, with the blots showing control and undernourished input total gDNA (CI, UI), output DNA (CO, UO ). (g,h) qPCR analysis of GR mRNA expression levels in the hippocampus. (i,j) GR protein intensity in the hippocampus. Data are of control male (n ϭ 6), UN male (n ϭ 8), control female (n ϭ 8) and UN female (n ϭ 7) adult offspring and are shown as mean Ϯ s.e.m *P Ͻ .05, **P Ͻ .01, ***P Ͻ .005. doi: 10.1210/en.2013-1693 endo.endojournals.org 7

Table 1. Epigenetic and glucocorticoid receptor (GR) expression levels in the pituitary of control and maternally programmed adult offspring. Control male (n ϭ 6) and UN male (n ϭ 8), control female (n ϭ 8) UN female (n ϭ 7), adult offspring. Data are represented as meanϮSEM *P Ͻ 0.05, **P Ͻ 0.01, ***P Ͻ 0.005.

Promoter GR mRNA Methylation H3K9AC H3K27me3 Expression GR protein Male Control 0.62 0.09 0.53 0.09 0.99 0.07 0.85 0.10 0.26 0.06 UN 0.46 0.05 0.99 0.09** 0.80 0.08 1.92 0.32* 0.69 0.14* Female Control 0.14 0.04 0.62 0.04 0.39 0.07 1.18 0.23 0.26 0.03 UN 0.94 0.07*** 0.62 0.05 0.66 0.07* 0.71 0.09** 0.15 0.04* whereas in postnatal life the same changes in GR are now varying modifications to the neuropeptide network in the associated with altered regulation of the appetite regula- hypothalamus, although they have not studied epigenetic tory pathways. This result suggests a precise, developmen- changes (9, 44). Delahaye et al (2008) found a reduction tally regulated, maintained response to maternal under- in hypothalamic POMC mRNA in offspring, as a conse- nutrition that could significantly alter the energy intake in quence of maternal undernutrition in rodents, which these animals. Indeed, the sheep paradigm of maternal would be in keeping with the results from this study in periconceptional undernutrition used in this study has sheep. Our results have extended their findings by sug- been shown previously to result in increased body weight gesting that the mechanism underlying the reduced hypo- in undernourished offspring at 10 months of age, a finding thalamic POMC could be due to the epigenetic changes in associated with impaired glucose intolerance (22). hypothalamic GR. It is interesting to speculate that the increase in hypo- Throughout the investigation we have observed sex- thalamic GR is responsible for the significantly reduced specific changes at the epigenetic, expression and pheno- expression of POMC mRNA in the ventral hypothalamus, typic levels. However, the mechanisms by which maternal which would be predicted to increase body weight. Indeed, undernutrition induces sex specific changes are still being several studies indicate that increasing the levels of gluco- delineated. Epigenetic changes are suggested to occur early corticoids in the hypothalamus decreases POMC mRNA in development due to genetic imprinting, leading to the (13, 14, 39, 40). However, in one study, adrenalectomy silencing of chromosomal segments, which may occur dif- and therefore, removal of endogenous glucocorticoids, led ferentially in males and females as a consequence of nu- to decreased POMC expression in hypothalami in rats tritional insults in early pregnancy (45). Another possible (41). Nevertheless, other studies have shown that follow- explanation is that estrogen in the females could be pro- ing adrenalectomy, there is an increase in POMC mRNA tecting them from programming effects in the hypothala- and greater anorexigenic tone (15, 42) and also a decrease mus and providing compensatory pathways to regulate in body weight and food intake (43). Therefore, in the food intake and energy balance. It has also been shown current study, the increase in hypothalamic GR in the un- that male rats are more receptive to changes in insulin dernourished adult male offspring could have resulted in effects in the brain than females (46). Given that we have the decrease in POMC mRNA expression, contributing to previously shown reduced levels of insulin in our sheep the increase in fat mass. model at 10 months of age (22), it may be that these de- Several other programming studies have shown that creased levels of insulin have a greater impact on insulin maternal undernutrition or overnutrition can result in regulation of neuropeptides in the hypothalamus in the male sheep. It is important to note that the persistence into adulthood of the GR modifications in the hypothalamus was not observed in other tissues. In the hippocampus and pituitary, the epigenetic changes now found in the adult animals differ from those in the fetus, which implies an adaptive Figure 5. No change in glucocorticoid receptor (GR) promoter methylation or protein expression plasticity after birth in epigenetic in leukocytes. (a) GR promoter methylation enrichment (control n ϭ 9, UN n ϭ 8). (b) Western modulation in these tissues. Both the blot analysis of GR protein levels in leukocytes (control n ϭ 9, UN n ϭ 8). Data are shown as hippocampus and pituitary are im- meanϮs.e.m. 8 Programming tissue specific changes in GR Endocrinology portant sites for glucocorticoid regulation. The maternally corticoids on the energy regulating pathways and gene programmed alterations in GR identified in the adult off- specific changes in different tissues. While this has allowed spring could have profound implications for the HPA axis for direct phenotypic associations to be made, it does not and for adaptations to stress and, indeed, we have shown enable further understanding of the potential altered tra- altered stress responses to both pharmacological and jectory of the epigenome as a whole in the offspring over physiological stimuli in these animals (27, 47). Recent pa- time. Furthermore, other programming models have high- pers have used peripheral blood cells or umbilical cord lighted the need for genome-wide analysis leading to the tissue to determine the epigenetic status of genes, as a sur- discovery of novel genes affected by the maternal insult in rogate noninvasive biomarker for epigenetic changes in the offspring and the overall dynamics of the epigenome the brain or other tissues (37, 48, 49). Of key importance (52, 53). While these studies have been extremely power- in the present study, where concurrent analysis of blood ful in providing novel targets, they have lower statistical and brain tissues was possible, the different effects on GR power and can lead to difficulties in making meaningful in several regions in the brain could not be predicted by epigenetic alterations in leukocytes which questions their conclusions as the pathways in which these genes operate utility as a GR biomarker. This is in agreement with an- may not have been determined (54). As both approaches other study suggesting that noninvasive tissues cannot be present advantages and disadvantages, the future of pro- used to measure GR status in the hippocampus as a con- gramming studies should aim to combine genome-wide sequence of maternal care (50). analysis with a candidate gene approach, to identify the Several mechanisms could be postulated that may po- key mechanisms that lead to an increased risk of disease in tentially lead to the epigenetic changes that we have ob- the offspring. served in the adult offspring. One hypothesis is that the Human studies have already found that undernutrition placenta in these animals has decreased levels of activity of during pregnancy has implications for the risk of diabetes the glucocorticoid inactivating enzyme, 11 beta hydrox- and obesity in the offspring (4, 5). Our data on maternal ysteroid dehydrogenase type-2 (51). Consequently the fe- undernutrition provide new evidence of epigenetic mod- tus will be exposed to excess maternal glucocorticoids. ifications persisting from the fetus to the middle-aged Glucocorticoids have been shown to alter DNA methyl- adult, which can impact on gene expression in neuronal ation patterns in different organs of the fetus (52) and, networks involved in regulating food intake and energy therefore, any changes in the levels of glucocorticoids balance. These epigenetic modifications are associated could impact on the epigenetic and expression status of with increased fat mass in the sheep. Understanding these genes seen in these offspring. Another proposed mecha- mechanisms could better inform the public perception of nism is that by undernourishing the mother, the fetus is a “healthy diet” for women of reproductive age and have receiving decreased methyl donors, which might contribute to the altered levels of methylation (38). Thus, mater- implications for the timing of education and other inter- nal undernutrition could lead to varying epigenetic out- ventions which would need to commence prior to comes, altering the set point of key pathways regulating conception. energy balance and stress. Our investigations have focused on a candidate gene approach in delineating the possible influence of gluco- Acknowledgments

Funding was provided by the Barbara Mawer endowment fund University of Manchester and the Manchester Academic Health Sciences Centre, the Health Research Council of New Zealand and Gravida: National Centre for Growth and Development, New Zealand.

Address all correspondence and requests for reprints to: Pro- fessor Anne White, Faculties of Life Sciences and Medical and Human Sciences, Manchester Academic Health Sciences Centre, University of Manchester, 3.016 AV Hill Building, Manchester M13 9PT, E-mail: [email protected], Telephone: 0161 Figure 6. Increased fat:lean mass ratio in adult males whose mothers 275 5178, Secretary: 0161 275 5180. were periconceptionally undernourished. (a,b) Fat mass to lean mass was measured by dual-emission X-ray absorptiometry (DXA) and Author information: The authors have nothing to declare calculated as a ratio. Data are shown as meanϮs.e.m. This work was supported by . doi: 10.1210/en.2013-1693 endo.endojournals.org 9

References Seckl JR, Dymov S, Szyf M, Meaney MJ. Epigenetic programming by maternal behavior. Nature neuroscience. 2004;7:847–854. 1. Swinburn BA, Sacks G, Hall KD, McPherson K, Finegood DT, 17. McGowan PO, Sasaki A, D’Alessio AC, Dymov S, Labonte B, Szyf Moodie ML, Gortmaker SL. The global obesity pandemic: shaped M, Turecki G, Meaney MJ. Epigenetic regulation of the glucocor- by global drivers and local environments. Lancet. 2011;378:804– ticoid receptor in human brain associates with childhood abuse. 814. Nature neuroscience. 2009;12:342–348. 2. Barker DJ, Eriksson JG, Forsen T, Osmond C. Fetal origins of adult 18. Symonds ME, Budge H. Nutritional models of the developmental disease: strength of effects and biological basis. International journal programming of adult health and disease. The Proceedings of the of epidemiology. 2002;31:1235–1239. Nutrition Society. 2009;68:173–178. 3. Ng SF, Lin RC, Laybutt DR, Barres R, Owens JA, Morris MJ. 19. Begum G, Stevens A, Smith EB, Connor K, Challis JR, Bloomfield Chronic high-fat diet in fathers programs beta-cell dysfunction in F, White A. Epigenetic changes in fetal hypothalamic energy regu- female rat offspring. Nature. 2010;467:963–966. lating pathways are associated with maternal undernutrition and 4. Ravelli GP, Stein ZA, Susser MW. Obesity in young men after famine twinning. FASEB journal : official publication of the Federation of exposure in utero and early infancy. The New England journal of American Societies for Experimental Biology. 2012;26:1694–1703. medicine. 1976;295:349–353. 20. Roseboom T, de Rooij S, Painter R. The Dutch famine and its long- 5. Ravelli AC, van Der Meulen JH, Osmond C, Barker DJ, Bleker OP. term consequences for adult health. Early human development. Obesity at the age of 50 y in men and women exposed to famine 2006;82:485–491. 21. Stevens A, Begum G, Cook A, Connor K, Rumball C, Oliver M, prenatally. The American journal of clinical nutrition. 1999;70: Challis J, Bloomfield F, White A. Epigenetic changes in the hypo- 811–816. thalamic proopiomelanocortin and glucocorticoid receptor genes in 6. Farooqi IS, Drop S, Clements A, Keogh JM, Biernacka J, Lowenbein the ovine fetus after periconceptional undernutrition. Endocrinol- S, Challis BG, O’Rahilly S. Heterozygosity for a POMC-null muta- ogy. 2010;151:3652–3664. tion and increased obesity risk in humans. Diabetes. 2006;55:2549– 22. Todd SE, Oliver MH, Jaquiery AL, Bloomfield FH, Harding JE. 2553. Periconceptional undernutrition of ewes impairs glucose tolerance 7. Farooqi IS, Wangensteen T, Collins S, Kimber W, Matarese G, Ke- in their adult offspring. Pediatric research. 2009;65:409–413. ogh JM, Lank E, Bottomley B, Lopez-Fernandez J, Ferraz-Amaro I, 23. Donovan EL, Hernandez CE, Matthews LR, Oliver MH, Jaquiery Dattani MT, Ercan O, Myhre AG, Retterstol L, Stanhope R, Edge AL, Bloomfield FH, Harding JE. Periconceptional undernutrition in JA, McKenzie S, Lessan N, Ghodsi M, De Rosa V, Perna F, Fontana sheep leads to decreased locomotor activity in a natural environ- S, Barroso I, Undlien DE, O’Rahilly S. Clinical and molecular ge- ment. Journal of Developmental Origins of Health and Disease. netic spectrum of congenital deficiency of the leptin receptor. The 2013;4:296–299. New England journal of medicine. 2007;356:237–247. 24. Jaquiery AL, Oliver MH, Honeyfield-Ross M, Harding JE, Bloom- 8. Delahaye F, Breton C, Risold PY, Enache M, Dutriez-Casteloot I, field FH. Periconceptional undernutrition in sheep affects adult phe- Laborie C, Lesage J, Vieau D. Maternal perinatal undernutrition notype only in males. Journal of nutrition and metabolism. 2012; drastically reduces postnatal leptin surge and affects the develop- 2012:123610. ment of arcuate nucleus proopiomelanocortin neurons in neonatal 25. Bloomfield FH, Oliver MH, Hawkins P, Campbell M, Phillips DJ, male rat pups. Endocrinology. 2008;149:470–475. Gluckman PD, Challis JR, Harding JE. A periconceptional nutri- 9. Breton C, Lukaszewski MA, Risold PY, Enache M, Guillemot J, tional origin for noninfectious preterm birth. Science (New York, Riviere G, Delahaye F, Lesage J, Dutriez-Casteloot I, Laborie C, NY). 2003;300:606. Vieau D. Maternal prenatal undernutrition alters the response of 26. Bloomfield FH, Oliver MH, Hawkins P, Holloway AC, Campbell POMC neurons to energy status variation in adult male rat off- M, Gluckman PD, Harding JE, Challis JR. Periconceptional under- spring. American journal of physiology Endocrinology and metab- nutrition in sheep accelerates maturation of the fetal hypothalamic- olism. 2009;296:E462–472. pituitary-adrenal axis in late gestation. Endocrinology. 2004;145: 10. Shin BC, Dai Y, Thamotharan M, Gibson LC, Devaskar SU. Pre- 4278–4285. and postnatal calorie restriction perturbs early hypothalamic neu- 27. Oliver MH, Bloomfield FH, Jaquiery AL, Todd SE, Thorstensen EB, ropeptide and energy balance. Journal of neuroscience research. Harding JE. Periconceptional undernutrition suppresses cortisol re- 2012;90:1169–1182. sponse to arginine vasopressin and corticotropin-releasing hormone 11. Cripps RL, Martin-Gronert MS, Archer ZA, Hales CN, Mercer JG, challenge in adult sheep offspring. Journal of Developmental Ori- Ozanne SE. Programming of hypothalamic neuropeptide gene ex- gins of Health and Disease. 2012;3:52–58. pression in rats by maternal dietary protein content during preg- 28. Zhang TY, Hellstrom IC, Bagot RC, Wen X, Diorio J, Meaney MJ. nancy and lactation. Clinical science (London, England : 1979). Maternal care and DNA methylation of a glutamic acid decarbox- 2009;117:85–93. ylase 1 promoter in rat hippocampus. The Journal of neuroscience 12. Lopez M, Seoane LM, Tovar S, Garcia MC, Nogueiras R, Dieguez : the official journal of the Society for Neuroscience. 2010;30: C, Senaris RM. A possible role of neuropeptide Y, agouti-related 13130–13137. protein and leptin receptor isoforms in hypothalamic programming 29. Oliver RL, Davis JR, White A. Characterisation of ACTH related by perinatal feeding in the rat. Diabetologia. 2005;48:140–148. peptides in ectopic Cushing’s syndrome. Pituitary. 2003;6:119– 13. Beaulieu S, Gagne B, Barden N. Glucocorticoid regulation of pro- 126. opiomelanocortin messenger ribonucleic acid content of rat hypo- 30. Turner JD, Pelascini LP, Macedo JA, Muller CP. Highly individual thalamus. Molecular endocrinology (Baltimore, Md). 1988;2:727– methylation patterns of alternative glucocorticoid receptor promot- 731. ers suggest individualized epigenetic regulatory mechanisms. Nu- 14. Wardlaw SL, McCarthy KC, Conwell IM. Glucocorticoid regula- cleic acids research. 2008;36:7207–7218. tion of hypothalamic proopiomelanocortin. Neuroendocrinology. 31. Maunakea AK, Nagarajan RP, Bilenky M, Ballinger TJ, D’Souza C, 1998;67:51–57. Fouse SD, Johnson BE, Hong C, Nielsen C, Zhao Y, Turecki G, 15. Gyengesi E, Liu ZW, D’Agostino G, Gan G, Horvath TL, Gao XB, Delaney A, Varhol R, Thiessen N, Shchors K, Heine VM, Rowitch Diano S. Corticosterone regulates synaptic input organization of DH, Xing X, Fiore C, Schillebeeckx M, Jones SJ, Haussler D, Marra POMC and NPY/AgRP neurons in adult mice. Endocrinology. MA, Hirst M, Wang T, Costello JF. Conserved role of intragenic 2010;151:5395–5402. DNA methylation in regulating alternative promoters. Nature. 16. Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, Sharma S, 2010;466:253–257. 10 Programming tissue specific changes in GR Endocrinology

32. Ernst J, Kheradpour P, Mikkelsen TS, Shoresh N, Ward LD, Epstein 43. Dallman MF, la Fleur SE, Pecoraro NC, Gomez F, Houshyar H, CB, Zhang X, Wang L, Issner R, Coyne M, Ku M, Durham T, Kellis Akana SF. Minireview: glucocorticoids–food intake, abdominal M, Bernstein BE. Mapping and analysis of chromatin state dynamics obesity, and wealthy nations in 2004. Endocrinology. 2004;145: in nine human cell types. Nature. 2011;473:43–49. 2633–2638. 33. Geng CD, Schwartz JR, Vedeckis WV. A conserved molecular mech- 44. Chen H, Simar D, Morris MJ. Hypothalamic neuroendocrine cir- anism is responsible for the auto-up-regulation of glucocorticoid cuitry is programmed by maternal obesity: interaction with postna- receptor gene promoters. Molecular endocrinology (Baltimore, tal nutritional environment. PloS one. 2009;4:e6259. Md). 2008;22:2624–2642. 45. McCarthy MM, Auger AP, Bale TL, De Vries GJ, Dunn GA, Forger 34. Cao-Lei L, Suwansirikul S, Jutavijittum P, Meriaux SB, Turner JD, NG, Murray EK, Nugent BM, Schwarz JM, Wilson ME. The epi- Muller CP. 2013 Glucocorticoid receptor gene expression and pro- genetics of sex differences in the brain. The Journal of neuroscience moter CpG modifications throughout the human brain. Journal of : the official journal of the Society for Neuroscience. 2009;29: psychiatric research 12815–12823. 35. Lesage J, Blondeau B, Grino M, Breant B, Dupouy JP. Maternal 46. Clegg DJ, Riedy CA, Smith KA, Benoit SC, Woods SC. Differential undernutrition during late gestation induces fetal overexposure to sensitivity to central leptin and insulin in male and female rats. Di- glucocorticoids and intrauterine growth retardation, and disturbs abetes. 2003;52:682–687. the hypothalamo-pituitary adrenal axis in the newborn rat. Endo- 47. Hernandez CE, Matthews LR, Oliver MH, Bloomfield FH, Harding crinology. 2001;142:1692–1702. JE. Effects of sex, litter size and periconceptional ewe nutrition on 36. Sebaai N, Lesage J, Breton C, Vieau D, Deloof S. Perinatal food offspring behavioural and physiological response to isolation. Phys- deprivation induces marked alterations of the hypothalamo-pitu- iology, behavior. 2010;101:588–594. itary-adrenal axis in 8-month-old male rats both under basal con- 48. Ehrlich S, Weiss D, Burghardt R, Infante-Duarte C, Brockhaus S, Muschler MA, Bleich S, Lehmkuhl U, Frieling H. Promoter specific ditions and after a dehydration period. Neuroendocrinology. 2004; DNA methylation and gene expression of POMC in acutely under- 79:163–173. weight and recovered patients with anorexia nervosa. Journal of 37. Kuehnen P, Mischke M, Wiegand S, Sers C, Horsthemke B, Lau S, psychiatric research. 2010;44:827–833. Keil T, Lee YA, Grueters A, Krude H. An Alu element-associated 49. Harvey NC, Lillycrop KA, Garratt E, Sheppard A, McLean C, hypermethylation variant of the POMC gene is associated with Burdge G, Slater-Jefferies J, Rodford J, Crozier S, Inskip H, Emerald childhood obesity. PLoS genetics 2012;8:e1002543. BS, Gale CR, Hanson M, Gluckman P, Godfrey K, Cooper C. Eval- 38. Lillycrop KA, Slater-Jefferies JL, Hanson MA, Godfrey KM, Jack- uation of methylation status of the eNOS promoter at birth in re- son AA, Burdge GC. Induction of altered epigenetic regulation of the lation to childhood bone mineral content. Calcified tissue interna- hepatic glucocorticoid receptor in the offspring of rats fed a protein- tional. 2012;90:120–127. restricted diet during pregnancy suggests that reduced DNA meth- 50. Liberman SA, Mashoodh R, Thompson RC, Dolinoy DC, Cham- yltransferase-1 expression is involved in impaired DNA methylation pagne FA. Concordance in hippocampal and fecal Nr3c1 methyl- and changes in histone modifications. The British journal of nutri- ation is moderated by maternal behavior in the mouse. Ecology and tion. 2007;97:1064–1073. evolution. 2012;2:3123–3131. 39. Sato I, Arima H, Ozaki N, Watanabe M, Goto M, Hayashi M, 51. Connor KL, Challis JR, van Zijl P, Rumball CW, Alix S, Jaquiery Banno R, Nagasaki H, Oiso Y. Insulin inhibits neuropeptide Y gene AL, Oliver MH, Harding JE, Bloomfield FH. Do alterations in pla- expression in the arcuate nucleus through GABAergic systems. The cental 11beta-hydroxysteroid dehydrogenase (11betaHSD) activi- Journal of neuroscience : the official journal of the Society for Neu- ties explain differences in fetal hypothalamic-pituitary-adrenal roscience. 2005;25:8657–8664. (HPA) function following periconceptional undernutrition or twin- 40. Zakrzewska KE, Cusin I, Stricker-Krongrad A, Boss O, Ricquier D, ning in sheep? Reproductive sciences (Thousand Oaks, Calif). 2009; Jeanrenaud B, Rohner-Jeanrenaud F. Induction of obesity and hy- 16:1201–1212. perleptinemia by central glucocorticoid infusion in the rat. Diabetes. 52. Crudo A, Suderman M, Moisiadis VG, Petropoulos S, Kostaki A, 1999;48:365–370. Hallett M, Szyf M, Matthews SG. Glucocorticoid programming of 41. Savontaus E, Conwell IM, Wardlaw SL. Effects of adrenalectomy on the fetal male hippocampal epigenome. Endocrinology. 2013;154: AGRP, POMC, NPY and CART gene expression in the basal hy- 1168–1180. pothalamus of fed and fasted rats. Brain research. 2002;958:130– 53. Thomassin H, Flavin M, Espinas ML, Grange T. Glucocorticoid- 138. induced DNA demethylation and gene memory during develop- 42. Rorato R, Castro M, Borges BC, Benedetti M, Germano CM, An- ment. The EMBO journal. 2001;20:1974–1983. tunes-Rodrigues J, Elias LL. Adrenalectomy enhances endotoxemia- 54. Amos W, Driscoll E, Hoffman JI. Candidate genes versus genome- induced hypophagia: higher activation of corticotrophin-releasing- wide associations: which are better for detecting genetic suscepti- factor and proopiomelanocortin hypothalamic neurons. Hormones bility to infectious disease? Proceedings Biological sciences / The and behavior. 2008;54:134–142. Royal Society. 2011;278:1183–1188.

Chapter 6 Maternal undernutrition induces tissue specific changes in POMC in adult offspring

6.1 Introduction

Maternal undernutrition can influence the epigenetic and expression status of POMC in the offspring (Delahaye et al., 2008, Cripps et al., 2009, Bloomfield et al., 2004). As described in the introduction, POMC is critical in both the hypothalamic energy balance pathway and the HPA axis. Thus, any alterations in its expression could impact on the offspring’s food intake, body weight and stress responses. Indeed, studies published during our initial investigations have shown decreased methylation in key areas of the POMC promoter in hypothalmi of offspring subject to maternal nutritional restriction and overfeeding (Coupé et al., 2010, Plagemann et al., 2009). We also observed marked hypomethylation of our chosen POMC region in hypothalami from fetuses that were subject to periconceptional maternal undernutrition (Stevens et al., 2010). However, other programming models have not determined the progression of POMC epigenetic modifications from the fetal stage into adulthood. These investigations are vital in determining the long term implications of POMC epigenetic modifications on food intake and energy balance. Additionally, no programming studies have focused on the epigenetic status of the predicted hypothalamic enhancer region we have utilised in the POMC gene. As described in more detail in chapters 3 and 4, the glucocorticoid receptor could potentially bind to this region, influencing POMC expression (Langlais et al., 2011). Consequently, this amplicon is an ideal candidate for analysis in maternally undernourished offspring.

Currently research is focused on correlating changes in the brain with non-invasive tissues such as fecal tissues to use as biomarkers for aberrant disease systems (Liberman et al. 2012). Emerging evidence suggests that the epigenetic and expression status of POMC in white blood cells could be used as a biomarker to determine potential POMC changes in the hypothalamus associated with the increased propensity of an individual to develop obesity (Kuehnen et al., 2012). Investigations mapping changes in a variety of genes such as GR in peripheral leukocytes and peripheral pathways in maternally programmed offspring are only just beginning, with the potential of using peripheral leukocytes as a biomarker in the offspring (Drake et al., 2012).

6.1.1 Aims

A shown in chapter 3, we found modifications in the epigenetic status of POMC in fetal sheep hypothalami as a consequence of maternal undernutrition, but this did not translate into POMC expression changes at the fetal stage. Consequently we aimed to investigate whether or not these epigenetic alterations in maternally undernourished offspring progressed into adulthood and if there were any associated changes in POMC expression levels. Furthermore, the epigenetic and expression status of POMC in the pituitary and leukocytes were analysed to determine if any of the hypothalamic alterations observed are tissue specific. We then extended these investigations to determine whether or not peripheral circulating leukocytes in the offspring have the potential to be used as biomarkers for epigenetic changes as a consequence of maternal programming.

6.2 Materials and methods

6.2.1 Animal management Maternal ewes were either fed ad libitum (control) or undernourished to maintain a 10-15% reduction in bodyweight from 60 days preconception until 30 days postconception (UN). Offspring were sacrificed as adults using pentobarbitone (mean age of 4.4 years), and tissues and blood samples were instantly removed. Studies were approved by the University of Auckland Animal Ethics Committee. The hypothalamus was dissected such that the region taken was enriched for the ventral hypothalamus as previously described (Stevens et al., 2011).

6.2.2 Isolation of DNA and RNA and protein from tissue and blood To investigate any modifications in leukocytes the DNA and protein was extracted using an adapted protocol of the Gentra Puregene kit (Qiagen; see methods 2.3). RNA and protein for expression analysis was extracted from the ventral hypothalamus and the pituitary using the Allprep DNA/RNA kit and RIPA buffer respectively (Qiagen) and is described more fully in the methods section (Section 2.2 and 2.3).

6.2.3 DNA methylation enrichment As previously described the MethylCollector kit was used for methylation enrichment of the POMC amplicon from total genomic DNA obtained from brain tissues and leukocytes (Active Motif).

6.2.4 Chromatin immunoprecipitation of H3K9AC and H3K27me3 ChIP experiments for H3K9ac and H3K27me3 was performed according to previous publications (Dahl and Collas, 2008). In summary 50mg of dissected brain tissue was sonicated whilst on ice. Purified chromatin was then immunoprecipitated with either an antibody against H3K9ac (39917) or H3K27me3 (39155). RNA polymerase II (R1530 Sigma) and normal mouse IgG (M8695 Sigma) were used as positive and negative controls respectively. Cross links were then hydrolysed to allow DNA elution. Input digested genomic DNA and output methylated DNA was then analysed by PCR using primers for POMC as described previously (Stevens et al., 2010).

6.2.5 qRT-PCR analysis For POMC a 5’ nuclease assay was designed utilising a FAM reporter dye, and Quantitect probe one step qRT-PCR reagents (Qiagen). HPRT was used as a reference gene to control for the reaction and was used in the standard SYBR green format and did not require a probe as detailed in the methods section 2.6. POMC levels were determined as relative to the reference gene HPRT. Sequences for the primers are as follows: POMC forward: 5'-GCTACGGCGGGTTCATGA-3' POMC reverse: 5'-TTCTTGATGATGGCGTTTTTGA-3' POMC probe: 5'fam-AGCCAAACGCCCCTTGTCACGC-methyl red3'; HPRT forward: 5’-TTTATTCCTCATGGACTAATTATGGA-3’ HPRT reverse: 5’-GCCACCCATCTCCTTCATCAC-3’

6.2.6 Western blot GR protein levels for brain tissue and leukocytes were determined according to standard Western blotting procedures. In summary protein was isolated from the dissected hypothalami, hippocampi and pituitary tissue using 1xRIPA buffer supplemented with protease inhibitors (Roche) and protein content was determined using the Nanodrop (Nanodrop ND1000). Resultant protein samples were then subjected to SDS-PAGE and Western blotting procedures. Proteins of interest were detected using the following antibodies POMC (in house A1A12; 1 in 5000), anti β-actin (Abcam; 1 in 5000) and anti α-tubulin (Santa Cruz; 1 in 5000).

6.2.7 Statistical analysis Data was analysed using GraphPad Prism software (GraphPad). The results were tested for significance using the unpaired students T test, where values of p<0.05 are statistically different. All data are represented as mean ± s.e.m.

6.3 Epigenetic changes in hypothalamic POMC

As detailed in chapters 3 and 4, methylation enrichment of the hypothalamic POMC promoter was carried out using a region that had a high level of mammalian conservation, located 6.5kb 5’of exon 1. Due to the limited availability of the sheep genome this 160bp amplicon was mapped and screened for sequence homology with the aid of the human and bovine genome. Moreover, during our investigations this fragment was revealed as having potential enhancer activity in the pituitary influencing POMC expression (Langlais et al., 2011). The amplicon was further revealed to have a GR binding site, potentially influencing the further regulation of this gene (Langlais et al., 2011).

Methylation enrichment of our chosen POMC amplicon revealed an increase in methylation in the maternally undernourished females compared to controls, with no alteration in the males (Figure 6.01b, 20% increase p<0.01). To determine if these changes were concomitant with any modifications in histone status, we carried out ChIP analysis for the histone markers H3K9ac and H3K27me3 (associated with open and closed chromatin respectively). H3K9ac was at significantly higher levels in the male and female maternally UN offspring (Figure 6.01c-6.01d, Male: 87% increase, female: 92% increase p<0.005). H3K27me3 levels were also found to be increased in both genders of maternally programmed offspring (Figure 6.01e-6.01f, Male 63% increase, female: 96% increase p<0.005).

Figure 6.01: Altered epigenetic status of the POMC gene in the ventral hypothalamus of maternally undernourished adult offspring. (a, b) POMC promoter methylation enrichment in the ventral hypothalamus. ChIP enrichment for (c, d) H3K9ac and (e, f) H3K27me3 of the chosen POMC amplicon. Data are for control male (n=6), UN male (n=8), control female (n=8) and UN female (n=7) and are shown as mean ± s.e.m **p<0.01, ***p<0.005.

6.4 Comparable levels of hypothalamic POMC protein

To decipher if the changes in epigenetic status of POMC had an influence on its expression, the protein levels of hypothalamic POMC were measured. It was found that POMC protein in the undernourished and control groups did not differ in either males or females (Figure 6.02).

Figure 6.02: Similar levels of hypothalamic POMC protein in adult animals following periconceptional undernutrition. (a, b) Ventral hypothalamic POMC protein levels in control male (n=6) and UN male (n=8), control female (n=8), UN female (n=7). Date is shown as mean±s.e.m.

6.5 Epigenetic changes in the POMC promoter amplicon in the pituitary of periconceptionally undernourished adult offspring

To evaluate the potential of tissue specific epigenetic changes in the POMC gene in adult offspring, the epigenetic status of the POMC promoter amplicon was also analysed in the pituitary. Offspring subject to periconceptional maternal undernutrition, demonstrated gender specific effects with increased POMC promoter methylation (Figure 6.03a; 120% increase p<0.05) and increased H3K27me3 (Figure 6.03e; 98% increase p<0.05) in the maternally undernourished males only.

Figure 6.03: Gender specific epigenetic changes in the POMC promoter region in the pituitary in maternally undernourished adult offspring. (a, b) Methylation enrichment of the POMC promoter in the pituitary. (c, d) ChIP analysis for H3K9ac and (e, f) H3K27me3of the POMC amplicon. Data represents control male (n=6), -60 to 30 male (n=8), control female (n=8) and -60 to 30 female (n=7) and are shown as mean ± s.e.m *p<0.05.

6.6 POMC mRNA and protein levels in the pituitary of adult offspring

To further elucidate the above observations and the implications that it may have on the expression of the POMC gene, the associated mRNA and protein expression of POMC in the pituitary was quantified. POMC gene expression and POMC protein levels were found to be similar across all groups and genders.

Figure 6.04: Comparable POMC mRNA and protein levels in the pituitary. (a, b) mRNA expression analysis of POMC in the pituitary (c, d) Western blots of POMC protein normalised to α-tubulin in the pituitary. The data is oncontrol male (n=6), UN male (n=8), control female (n=8) and UN female (n=7) and are shown as mean ± s.e.m.

6.7 POMC status in the peripheral leukocytes of maternally undernourished adult offspring

Clinical studies have suggested that the expression levels of genes in peripheral leukocytes could be used to predict hypothalamic function (Kuehnen et al, 2012). We aimed to investigate the potential of this in our model, where several tissues expressing the same gene could be compared. Consequently, we measured the methylations status and the protein expression levels of POMC in the leukocytes and compared the results to what we had found in the hypothalamus and pituitary. Several methods were attempted to extract the DNA and protein from whole blood for further analysis. This extraction process was extremely complex and required a large amount of optimisation as the samples we had received were frozen. Initially the buffy coat method was utilised, which is based on separating the different cells according to density. However, the red blood cells had burst due to freeze thawing and the different cell types did not separate well making it difficult to isolate DNA and protein purely from peripheral blood mononuclear cells. Consequently, as described in the methods, the Gentra Puregene kit (Qiagen) was adapted allowing the isolation of DNA as well as protein from leukocytes. However, the techniques employed did not allow for efficient RNA isolation, as the RNA had degraded. Once the DNA was extracted and enriched for methylation, it was found that there were similar levels of methylation of the POMC amplicon between the control and maternally undernourished adult offspring (Figure 6.05a). The leukocytes were then investigated for POMC protein levels and it was found that there were large amounts of variability with no significant differences between the two groups (Figure 6.05b).

Figure 6.05: POMC promoter methylation and protein expression in peripheral circulating leukocytes (a) Methylation enrichment of the POMC promoter amplicon in leukocytes. (b) Western blot to show the level of POMC expression normalised to the levels of β-actin. Results are shown as control (n=7),UN (n=8) and mean ±s.e.m.

6.8 Summary

In our model of maternal periconceptional undernutrition we have shown that the epigenetic status of our chosen POMC amplicon has altered between fetal life and adulthood in sheep hypothalami. Nevertheless, as shown in the results above for the maternally undernourished adult animals, the increases in the opposing histones, H3K9ac and H3K27me3, are difficult to interpret and no clear epigenetic state can be deciphered. This is in part is due to the lack of evidence in the literature determining the importance in the levels of one histone modification over another in the hypothalamus. As a result it is difficult to determine whether or not the chromatin for the POMC gene is opened or closed influencing transcriptional regulation. Additionally the increase in hypothalamic POMC methylation seen in the maternally underfed female group is different to what was found at the fetal stage. Methylation states are understood to be long lasting and consequently the changes we have observed suggest that the methylation status of our chosen POMC amplicon can be subject to change. In support of this a study by Crudo et al. (2013) found that the methylation status of the epigenome of the fetal brain can alter following glucocorticoid administration. These modifications highlight the necessity for programming models to consider the long term progression and stability of any epigenetic alterations found in their candidate genes.

Despite the methylation and histone changes of our chosen amplicon in the maternally undernourished offspring, no associated alterations in the protein levels of hypothalamic POMC were found. However, in the same animals changes in hypothalamic POMC mRNA expression were observed in the maternally undernourished adult males (Chapter 5). Potential reasons for this are complex and detailed in the discussion.

As with previous studies we extended our investigations to elucidate if the changes we had detected in the hypothalamus were tissue specific. Comparable epigenetic analysis revealed that in the pituitary there were alterations in the histone and methylation status of the POMC promoter region selected in the male undernourished group only, with indications suggesting a closed chromatin state. This suggests that this area of the POMC gene is primed for expression changes leading to a decrease in POMC. However, the epigenetic modifications were not found to be associated with changes in gene expression or protein at this stage.

In addition to the above experiments, we also analysed the levels of methylation enrichment of our region in peripheral circulating leukocytes as a biomarker for maternal programming in the adult offspring. There were no differences in the levels of methylation enrichment and associated protein levels. However, due to the complex changes observed in the hypothalamus and the pituitary it is difficult to determine the value of measuring POMC in the leukocytes as a biomarker.

Chapter 7 Discussion

The investigations detailed in this thesis demonstrate a link between moderate maternal undernutrition and the development of persistent epigenetic changes in the hypothalamus of the offspring. This is associated with the increased susceptibility of these animals to develop disease outcomes.

Initial studies determined that pre- and post- conceptional maternal undernutrition is concomitant with epigenetic changes in the fetal hypothalamic GR and POMC promoter regions, with altered GR mRNA expression levels. According to the thrifty phenotype hypothesis, in response to maternal undernutrition, the offspring are programmed to adjust to the undernourished environment (Barker, 2005). As a result, the changes we found in the fetal energy regulating pathway may have occurred to allow the fetus to adapt to the decrease in maternal nutrient levels. However, if the postnatal environment has an abundance of food, changes in this pathway could be maintained into adulthood, predisposing the offspring to increase their food intake. This may contribute to the development of obesity and glucose tolerance changes in these animals.

Maintenance of epigenetic changes was clearly demonstrated in our adult study where the maternally underfed offspring, now up to 5 years of age, still presented with persistent epigenetic alterations in the GR promoter region. Subsequent expression analysis revealed that these changes were associated with increased GR mRNA expression and protein levels. Evidence in the literature suggests that GR can influence neuropeptide expression in the hypothalamic energy regulating pathway to increase food intake (Beaulieu et al., 1988; Gyengesi et al., 2010; Sato et al., 2005; Zakrzewska et al., 1999). Consequently, we quantified hypothalamic expression of the anorexigenic neuropeptide POMC and found a decrease in mRNA expression in the males only, with this group developing an increase in fat mass.

Our studies also extended to decipher the potential of the hypothalamic changes to be tissue specific. In order to do this we carried out methylation and histone analysis of the POMC and GR promoter regions in the pituitary and the GR region in the hippocampus. At the fetal stage there were no changes observed in these genes in either the hippocampus or pituitary, but modifications were detected in those brain regions in the adult offspring that could impair their stress axis. These findings indicate that maternal undernutrition can not only induce permanent changes persisting from the fetal stage, but it can also induce modifications that develop after birth.

In addition to the studies outlined above, we also aimed to elucidate the potential of twins as a programming paradigm, associated with maternal nutritional and environmental restriction. Epigenetic analysis revealed that twins had altered hypothalamic POMC and GR methylation levels in a way that was similar to maternally undernourished fetal offspring. They also presented with histone modifications and changes in DNA methyltransferase activity, potentially leading to an increase in GR expression and, therefore, greater food consumption. Overall, the findings of the study implied that twin fetuses have an altered hypothalamic epigenetic POMC and GR status to singletons, which is similar to that of maternally undernourished singleton fetuses. This could possibly be occurring to deal with the lack of nutrients, leading to twins having an increased risk of developing diseases later on in life, such as the increased body weight and diabetes as observed in other studies (Poulsen et al., 2003; Poulsen et al., 2005).

7.1 The impact of the length and timing of maternal undernutrition on the offspring in sheep

Central to our studies was the use of the sheep model extensively developed by our collaborators (Bloomfield et al., 2003; Bloomfield et al., 2004). This model enabled our investigations to provide a more relevant prediction of changes that may occur in human hypothalami as both sheep and humans give birth to mature singleton young, with a similar hypothalamic developmental trajectory, aided by a long gestation (Symonds and Budge, 2009). Additionally, studying the outcomes of maternal undernutrition around conception is highly applicable to women of a reproductive age who are dieting due to societal pressures in the western world.

Our initial fetal cohort was designed to determine the effect of maternal undernutrition on the fetus, pre and/ or early post conception (Rumball et al., 2008). Throughout our investigations we have been able to show that maternal undernutrition around this period can induce hypomethylation of the selected POMC and GR promoter markers in the fetus. Whilst having a low birth weight is a common characteristic of offspring born to mothers subject to maternal undernutrition, the model we used in our studies does not produce offspring with decreased birth weights. These findings are consistent with epidemiological evidence which has shown that pregnant women exposed to food shortages during early gestation give birth to children with normal birth weights, but, these children have an increased risk of developing adverse outcomes such as obesity in adulthood (Ravelli et al., 1976; Roseboom et al., 2006). This suggests that 83 perhaps birth weight is not a precise marker of predicting maternal programming in the offspring.

Maternal undernourishment beginning 60 days before conception until the start of pregnancy affects the oocyte maturation stage, which requires maternally supplied proteins (Heasman, 2006). It has been found that offspring from rats underfed during the preimplantation period present with altered birth weight and hypertension (Kwong et al., 2000). It is also a possibility that undernourishing the mother before conception makes them insulin resistant and therefore this impaired regulation could impact on the fetus.

Subjecting mothers to maternal undernourishment from conception until 30 days into pregnancy could have implications for the placentation period and on the high level of epigenetic activity in the fetus (Rumball et al., 2008; Newell-Price 2003). Despite the nutrient requirements during the first trimester being minimal, a lack of nutrients has been found to be associated with increased obesity in adult men (Ravelli et al., 1976).

The maternally underfed fetal group which demonstrated the highest level of epigenetic changes was the group that was nutritionally programmed from 60 days prior to conception until 30 days during pregnancy. These results suggest that as well as the timing of the nutritional restriction, the length of insult is also important. As this group produced the greatest level of modifications they were also examined in our adult studies.

7.2 Epigenetic changes in hypothalamic GR as a consequence of maternal programming persisting from fetal life to adulthood

Throughout our studies we have taken a candidate gene approach, with the GR gene as one of our main targets for investigation. As mentioned in the introduction, GR has been shown to be a key regulator of the hypothalamic energy regulating pathway, influencing the expression of neuropeptides, primarily leading to an orexigenic effect (Beaulieu et al., 1988; Zakrzewska et al., 1999; Sato et al., 2005; Gyengesi et al., 2010). Consequently, epigenetic changes in GR could lead to altered neuropeptide expression inducing increased food intake and the development of obesity. Extensive evidence in the literature supports the hypothesis that GR can be epigenetically modified as a result of maternal programming; however, the majority of these

84 studies have utilised models where stress and the lack of maternal care are the main insults. Nevertheless, these studies have shown altered methylation of single CpGs in the human exon 1F and rodent exon 17 region of the GR promoter, with associated changes in GR expression (Weaver et al., 2004; McGowan et al., 2009). Additionally, these findings reveal the potential of maternal programming to induce subtle changes in GR promoter CpG methylation that could have greater implications on GR expression and feedback regulation. Hence, the susceptibility of GR to be modified in an epigenetic context made it an ideal candidate to study in our maternal undernutrition model.

At the time of our investigations, only two studies presented changes in GR as a result of maternal nutritional variations, with methylation changes in the GR promoter in the liver and the hippocampus (Lillycrop et al., 2005, Ke et al., 2010). However, further research into GR has not been continued and in particular, examination of GR changes in the hypothalamus has not been carried out. Therefore, we aimed to investigate potential epigenetic changes in the GR gene of maternally undernourished offspring in the hypothalamic energy regulating pathway. Initially it was difficult to identify any parts of the GR promoter for epigenetic analysis as the sheep genome is not freely available. Nevertheless, we selected the promoter of the GR exon 1E region situated 5kb upstream of the GR exon 2 translational start site, which has been shown to be extremely active (Ernst et al., 2011; Maunakea et al., 2010). The high level of conservation of this area across species allowed for accurate predictions of the sequence based on the human and bovine genome. Furthermore, this conservation infers that any changes observed may be comparable with higher mammals. Our selected amplicon was also ideal for epigenetic analysis as it has a number of CpG bases. Along the course of our investigations, sheep sequence for our GR region became available, which matched our predicted sequence. Importantly, the region has recently been shown to be transcribed in many areas of the human brain, including the PVN of the hypothalamus, hippocampus and the pituitary (Cao-Lei et al., 2013). Furthermore, the presence of a glucocorticoid response unit suggests the GR gene may be up regulated by GR binding to this region (Geng et al., 2008).

In our investigations it was found that the selected hypothalamic GR marker region was hypomethylated in maternally underfed and twin fetuses. This decrease in methylation was associated with reduction in the hypothalamic DNA methyltransferase activity in the maternally underfed -60 to 30 and twin fetal groups. Additionally, examination of histone markers revealed a possible open chromatin arrangement. It could be speculated that the opening of the chromatin 85 allows GR to bind to the glucocorticoid responsive element (GRE) present in the exon 1E promoter region leading to up regulation. These findings are consistent with the increased GR mRNA expression seen in the maternally underfed -60 to 30 singletons only. Recently, highly relevant data has also shown an increase in GR expression in the hypothalami of baboon offspring following a 70% reduction in the mother’s food intake (Li et al., 2013). This change in GR emphasises and strengthens our findings, suggesting that maternal nutritional restriction has implications for GR expression in offspring in higher mammals.

Results from our fetal studies showed no changes in hypothalamic GR mRNA expression in the fetal groups maternally underfed either preconception or post-conception, or in the twin model. These results indicate that fetal changes in GR expression may require a broader window of maternal undernourishment in order to take effect and that twinning is a much more complex paradigm. Despite these findings, it was observed that there were reduced levels of methylation of the GR promoter marker in these groups. This may have implications for GR expression alterations that could become apparent later on in life.

Most studies have failed to fully characterise the persistence of epigenetic changes from fetal and early life into adulthood and the potential impact of these changes on the animals. Consequently, our collaborators maintained the offspring from the -60 to 30 cohorts up until 5 years of age. Following methylation and histone analysis we found that the epigenetic signature present at the fetal stage was maintained in adulthood, with persistent significant increases in GR mRNA and protein levels. A large body of the literature suggests that GR acts to decrease POMC and increase NPY (Beaulieu et al., 1988; Zakrzewska et al., 1999; Sato et al., 2005; Gyengesi et al., 2010). Accordingly, in our study we speculate that the increased GR expression may have influenced the down-regulation of POMC expression leading to the increased fat mass observed in the adult males.

As GR can regulate a number of neuropeptides including the orexigenic NPY and AgRP peptides, we also endeavoured to analyse the expression levels of these genes in the hypothalami of control and maternally undernourished offspring. However, due to the incomplete accessibility of the sheep genome it was difficult to analyse changes within these genes. Nevertheless, based on the genome framework that was available, we were able to measure levels of NPY which were found to be unchanged and will be discussed in more detail further in

86 this discussion. Conversely, despite extensive efforts, we were unable to measure changes in the AgRP gene.

In our studies the levels of mRNA expression were reflective of total GR mRNA expression within the arcuate nucleus of the hypothalamus and not exon 1E specific transcribed amounts. Again, this was due to the lack of available ovine sequence. However, by measuring the total levels of GR mRNA we were then able to relate these findings to the increased GR protein levels in the hypothalamus. Nonetheless, it would have been interesting to correlate the different epigenetic changes with exon 1E mRNA transcript levels to determine which of the histones or DNA methylation modifications had the most influence. Whilst sheep are more comparable with humans than rodents, future studies should extend into rodents to allow an in depth examination of the different promoter regions in exon 1 of the GR gene. Furthermore, a genome-wide study could be undertaken to determine if there are any changes in associated pathways and genes that GR may act upon in the hypothalamic energy regulating pathway. This would allow for greater understanding of the role of GR within the hypothalamus and possible implications of any modifications observed on overall food intake and energy balance.

7.3 Epigenetic changes in hypothalamic POMC as a consequence of maternal programming

POMC is a necessary component of the hypothalamic energy regulating pathway, which acts to reduce food intake and increase energy expenditure (Fan et al., 1997). As a result we also chose to characterise the epigenetic status of POMC in our maternally undernourished animals. To do this we chose a region 9kb upstream of the start of exon 1 which is on the “shore” of a CpG island. This area is highly conserved between species and closely related to the npe1 and npe2 hypothalamic specific enhancer regions (de Souza et al., 2005; Langlais et al., 2011; Santangelo et al., 2007). More importantly, the POMC amplicon chosen for our studies has been identified as having enhancer activity albeit in the pituitary, with GR being predicted to bind to this region (Langlais et al., 2011). Therefore, it could be implied that the increase in GR observed in the maternally undernourished animals could be binding to our selected POMC region, leading to a decrease in POMC activity and an increase in food intake.

During the fetal study it was found that the selected POMC marker region was hypomethylated in the hypothalami of all the maternally undernourished fetal groups. Epigenetic changes were 87 also present in the POMC amplicon in the twin study, with additional increases in the histone marker of open chromatin H3K9 acetylation and decreased H3K27me3, as a marker of inactive chromatin. This histone configuration indicated that the chromatin may be open, aiding transcription factor binding and potentially influencing POMC expression and food intake. Evidence for this is also provided by Coupe et al. (2010) who found that maternal undernutrition in the rat induced hypomethylation of part of the POMC promoter, concomitant with a reduction in food intake. Additional evidence during our investigations was also published in support of our hypothesis demonstrating that maternal nutritional programming can lead to epigenetic modifications in POMC in the offspring’s hypothalami (Plagemann et al., 2009; Coupe et al., 2010; Shin et al., 2012). However, the models used in these studies were based on severe levels of maternal nutritional intervention, with the insults extending from gestation into postnatal life. Moreover, these experiments were carried out in rodents, which as mentioned in the introduction may not be entirely compatible with the human fetal hypothalamic developmental trajectory

Regardless of the POMC region being hypomethylated in our studies, these epigenetic changes did not translate into alterations in POMC mRNA expression in any of the maternally underfed or twin fetuses. A detailed examination of POMC in mice revealed that POMC peptides are not required for hypothalamic development during the fetal period or in the first 3 weeks of life to influence body weight (Bumaschny et al., 2012). Consequently, we speculated that modifications may become apparent later on in adulthood as the appetite pathway may not be active at this fetal stage. In accordance with this view, other studies that measured POMC expression levels postnatally found significant decreases in its expression (Delahaye et al., 2008; Ikenasio-Thorpe et al., 2007). Subsequent postnatal analysis in our studies revealed that there was a decrease in POMC mRNA expression in the adult male offspring of periconceptionally undernourished mothers. In part this may be explained by the potential increase of GR binding to our region as described above. As demonstrated in chapter 6, no changes were observed in the POMC protein levels, but this may be due to the transport of POMC by dendritic processes to other hypothalamic regions, concealing any changes that we may have seen. Additionally, compensatory mechanisms may be in place to prevent changes in POMC protein levels, allowing for more stringent evolutionary conservation.

During the fetal study it was observed that the levels of POMC mRNA expression in the whole and the ventral hypothalamic samples were similar (Chapter 3; Figure 4a-b). POMC is only enriched within the arcuate nucleus of the ventral hypothalamus and therefore, it is expected that 88 the POMC levels isolated from this region would be higher than that of the whole hypothalamus. However, this was not observed in our investigations and we propose several reasons for this including the potential quality of the RNA. Degraded RNA may have resulted in inaccurate readings when the levels of POMC expression were measured in the samples. To determine the integrity of the RNA, a subset of RNA samples purified from whole and ventral fetal hypothalami were analysed on an agarose gel. It was found that whilst some degradation had occurred, the 28S and 18S bands were still clearly discernible and present (Chapter 2). It is important to note that the integrity of these samples was checked 5 years after the original study and so any degradation that may have occurred might be due to multiple freeze thaws overtime. As a result it is difficult to determine if the RNA quality was a contributing factor to the similar POMC mRNA expression levels observed. Additionally, despite no alterations in POMC mRNA levels, the amount of NPY mRNA between whole and ventral hypothalamic tissues was altered, suggesting once more that the integrity of the RNA may not have been a contributing factor.

It could be speculated that GAPDH may not have been an appropriate housekeeping gene in the fetal study as there have been reports of variability in the levels between different tissues (Huggett et al. 2005). When the raw values of GAPDH were compared from the fetal hypothalamic samples, the levels were similar but there was some variation that may have affected the overall analysis and the resulting POMC expression. This could potentially imply that different regions of the hypothalamus have different levels of GAPDH and as a result it may not have been an appropriate housekeeping gene. However, this is purely speculation and there is a lack of clear evidence in the literature to support this. Moreover, during the twin study a different method for expression analysis and a different housekeeping gene were utilised for the measurement of POMC expression in the whole and ventral hypothalamic samples. During this study the remaining homogenous tissue from the fetal study was taken and the RNA was purified. Again the results from this investigation demonstrated that there were no alterations in POMC mRNA expression. Consequently it is difficult to determine the underlying causes for the lack of POMC enrichment in the fetal study. It is also important to note that the POMC levels were similar in both the fetal and twin studies and therefore, the potential impact on the overall conclusions made in the papers are minor.

Epigenetic analysis of our POMC amplicon in the adult male hypothalami showed that the region was no longer hypomethylated as seen in the fetus, with no change in the males and increased levels of methylation in the maternally undernourished females. Histone data was 89 extremely complex, with increases in both open and closed chromatin markers in both sexes. These results suggest that not all methylation changes are permanent. It also suggests that there are other histones that could collectively influence the overall chromatin status of the POMC gene that we did not measure in our studies.

It is important that future studies should extend into more hypothalamic specific regions such as the npe1 and npe2 enhancers. These regions have been extensively characterised as being essential for POMC expression in the hypothalamus, but due to the lack of accessible sheep sequence data, we were unable to analyse these areas in our studies. Investigating these regions would provide a more accurate prediction of the effect on POMC expression levels within the arcuate nucleus of the hypothalamus.

7.4 Tissue specific changes in GR and POMC

Throughout our investigations we sought to determine if any epigenetic or expression changes in GR and POMC were widespread in other tissues or whether they were organ-specific. Interestingly, we found that the increases in fetal hypothalamic GR expression were not present in other fetal tissues necessary for the glucocorticoid regulation of the HPA axis, such as the hippocampus and the pituitary. This is in contrast to another study that demonstrated altered GR expression in these areas of the brain at the fetal stage as a consequence of maternal programming (Hawkins et al., 2001). However, in that investigation the nutritional insult extended from 0 to 70 days during gestation. As a result the insult may have impacted on the developmental trajectory of the fetus at the time of critical HPA axis development. In comparison to that study, our maternal insult was stopped 40 days earlier and so we may not have impeded the development of the HPA axis. This reasoning is further strengthened by evidence from baboons, whereby nutrient restriction induced alterations in the global methylation levels of their offspring, in an organ and gestational time-specific manner (Unterberger et al., 2009).

Following the fetal study, we also extended our investigations in the adult model to decipher the potential of tissue-specific changes in hypothalamic GR in the maternally undernourished adult offspring. Our analysis revealed a decrease in GR expression in the hippocampus in those 90 offspring. Additionally, there were sex-specific changes in the expression of GR in the pituitary, with decreased GR in maternally undernourished females and increased GR in maternally undernourished males. Combined with the observations in the hypothalamus, these findings suggest that maternal programming can induce tissue-specific changes in genes that can be both temporarily and continually maintained into adulthood. It also highlights the importance of investigating long term changes in maternally programmed offspring

Possible hypotheses for these outcomes are that in the maternally undernourished fetuses changes may have occurred in other genes or organs vital for HPA axis function which we did not investigate, such as the adrenal gland. Over time these changes may have shifted the axis modifying GC feedback regulation in the hippocampus and pituitary, leading to the altered GR expression which we found in the adult animals. Another explanation is that the changes in GR may have become apparent after the prepartum surge necessary for the maturation of fetal organs leading to the onset of parturition (Kapoor et al., 2006; Mcmillen et al., 2004). This surge is reliant on the upregulation of the HPA axis. Importantly our collaborators revealed that at the fetal stage, the maternally undernourished offspring had higher levels of cortisol leading to preterm birth, with indications of an accelerated HPA axis (Bloomfield et al., 2004). This increase in cortisol may have impacted on the complex changes that occur in components of the axis during the prepartum surge. As a result, feedback regulation in the hippocampus may have been modified to allow the fetus to adapt, leading to possible long lasting changes in GR in the hippocampus and pituitary. In support of this hypothesis, a study by Crudo et al. (2013) demonstrated that the genome-wide fetal epigenetic framework can become significantly altered as a result of the cortisol driven prepartum surge. These findings not only have considerable implications for maternal nutritional paradigms, but also for synthetic glucocorticoid administration to mothers to induce birth. However, the fetal samples in our studies were taken before the prepartum surge; thus, any alterations as a consequence of changes in the surge would not have been seen during the fetal time point during which our tissues were analysed.

In addition to GR we also investigated the potential of POMC to be altered in a tissue-specific manner. As mentioned earlier, the epigenetic modifications in POMC at the fetal stage were confined to the hypothalamus and were not seen in pituitary POMC. In adulthood, decreased POMC expression was only observed in hypothalami of maternally undernourished males, with complex epigenetic changes in the pituitary. The POMC epigenetic changes are difficult to interpret as the strength in the levels of one histone change over another and the implications for 91 the expression of genes is yet to be fully deciphered. For example, in our model it was found that there was an increase in H3K9 acetylation and H3K27 trimethylation associated with the hypothalamic POMC amplicon. Due to the opposing actions of either histone on the open/ closed chromatin status and without analysing all the other histones present, it is challenging to make an assumption about the overall status of the chromatin. However, the levels of DNA methylation were increased in the maternally undernourished females only, suggesting decreased POMC expression which was not observed.

Recent studies have suggested that non-invasive tissues such as peripheral blood cells or umbilical tissues could be used as biomarkers for epigenetic alterations associated with obesity in hypothalamic tissues (Kuehnen et al., 2012). Consequently one of the aims of the adult work was to determine the epigenetic status of GR and POMC in peripheral circulating leukocytes from the maternally undernourished animals and comparing them to changes found in the different brain regions. Our results suggested that the GR and POMC epigenetic status in any of these regions could not be mapped with DNA methylation levels in peripheral leukocytes. These leukocytes also had similar amounts of POMC and GR protein in the control and maternally undernourished offspring which differed from protein levels in other tissues. Therefore, the data indicates that peripheral leukocytes cannot be used as biomarkers for determining GR and POMC status in maternal programming paradigms. During this work, it was extremely difficult to isolate the DNA due to the samples being frozen, resulting in the lysis of red blood cells. Future studies should perhaps utilise fresh blood, allowing for RNA isolation, which will provide additional evidence to strengthen the outcomes. Nevertheless, our findings are strengthened by another study demonstrating that other non-invasive tissues, such as fecal samples, cannot be used to determine epigenetic changes in hippocampal GR as a consequence of altered maternal care (Liberman et al., 2012).

7.5 NPY expression in the hypothalami of maternally underfed offspring

During our investigations we attempted to quantify the expression of a number of neuropeptides that are influenced by GR regulation within the hypothalamic energy regulating pathway. However, due to the limited availability of the sheep genome, obtaining correct primer sequences was highly problematic and whilst we were able to predict sequences for GR and POMC we were unable to do this for other genes. Nevertheless, we were able to obtain sequence for the 92 orexigenic neuropeptide NPY. Throughout the studies presented in this thesis it was found that the levels of NPY expression were similar between maternally undernourished, twin, and control fetal groups. In contrast to this, other studies in rodents demonstrated a significant increase in the levels of NPY following maternal nutrient restriction, but the nutrient restriction was much more severe than our model, or the insult was introduced postnatally (Breton et al., 2009; Lopez et al., 2005; Cripps et al., 2009). Additionally, following maternal nutrient restriction in the sheep model, there was a decrease in NPY expression after one week of age (Sebert et al., 2009). However, in contrast to our studies, the level of maternal undernutrition was much more severe at a 50% reduction in food intake and was provided during late gestation (Sébert et al., 2009). Therefore, it could be presumed that moderate maternal undernutrition during the time of conception is not sufficient to induce alterations in the expression of NPY mRNA in the hypothalamus.

7.6 Sex-specific outcomes in the maternally undernourished offspring

An increasing number of maternal programming paradigms have highlighted the need to analyse data according to gender, due to varying outcomes in adulthood (Matthews and Phillips, 2012). In our initial fetal studies, we were unable to conduct sex specific analysis as the number of animals was not large enough. This would have been valuable in providing more evidence to determine which gender is more vulnerable to maternal programming and when these differences become evident. Nonetheless, we were able to analyse our data according to gender in the adult study. It was found that only the maternally undernourished male offspring showed a decrease in the anorexigenic neuropeptide POMC, and this was associated with increased fat mass. This is consistent with changes seen in another model where moderate calorie restriction during early gestation resulted in the male offspring developing higher levels of fat accumulation (Palou et al., 2010). Additionally, following a 50% maternal restriction in another study the males had an increased fat mass, with no changes in the females (Jones and Friedman, 1982). Despite this there are other investigations that have described the development of increased fat gain due to maternal nutritional restriction in females and males (Anguita et al., 1993; Desai et al., 2005). These variances demonstrate that different periods and severity of maternal programming can induce conflicting phenotypic modifications in the adult offspring.

The causes of sexually dimorphic hypothalamic changes as a consequence of maternal programming have not been fully determined. Nevertheless, evidence in the literature indicates 93 that the male brain is more sensitive to nutritional manipulations, which may be associated with altered leptin and insulin resistance (Clegg et al., 2003). More specifically, it has been found that males are more susceptible to changes in insulin. In our maternally undernourished offspring, it was found that there was reduced insulin levels post puberty (Todd et al., 2009). As a result, it could be speculated that there may have been a decrease in the central actions of insulin to increase POMC expression, leading to a lack of inhibition of food intake by POMC. Differences in leptin and insulin sensitivity are also influenced by varying levels of oestrogen and increased levels in females may provide compensatory mechanisms in the hypothalamus in response to changes in maternal programming. Additionally in sheep, the levels of gonadal steroid hormone secretion are high and act during early gestation. Consequently if they are altered, they could influence neuropeptide activity in the hypothalamus and overall development (Quirke et al., 2001). If these changes persist there may be variations in other pathways, which may be a possible explanation for the sex-specific differences in the HPA axis that we saw in the maternally undernourished adults. Despite these potential reasons, more work must be done in this area to decipher the possible mechanisms leading to sex-specific epigenetic changes in a number of pathways.

7.7 Twinning as an intrauterine programming paradigm

Investigations in the twin model developed by our collaborators suggested that twins are potentially a programming paradigm. As detailed in the results, fetal twins demonstrated similar changes to maternally undernourished singleton fetal offspring in the epigenetic status of their GR gene, with altered DNMT activity in their hypothalami. However, unlike the maternally undernourished singleton offspring, the modifications were not coupled with expression changes. These epigenetic similarities between the singleton maternally undernourished and twin offspring may be the result of the twins developing in a maternal environment where the fetuses have to compete for maternal nutrients (Rumball et al., 2008). In comparison to singletons, twins have also been characterised as having decreased placental size, with twin fetuses developing in a limited spatial environment. Thus, the changes we have seen may have occurred to allow the fetus to adapt to these altered surroundings. In the long term this may lead to a compensated developmental trajectory in these animals that has the potential to increase their propensity to have impaired glucose regulation and increased fat mass in adulthood (Poulsen et al., 1997; Ribel-Madsen et al., 2012). Our studies also highlight the need for twin studies to include

94 singletons as references as we found that there was no difference between control twins and maternally undernourished twins and therefore, without the control singleton group, any changes would have been missed. To further this work, a long-term analysis of the twin model needs to be carried out to determine the epigenetic implications associated with disease susceptibility in twin adult offspring.

7.8 Phenotypic outcome in the adult offspring subject to periconceptional undernourishment

It has been shown that animals subjected to maternal undernutrition develop a wide range of complications, including an increase in fat mass and impaired glucose tolerance. Consistent with these outcomes, the offspring of maternally undernourished ewes in our studies were found to have an increased level of fat mass in the males only and previously published work by our collaborators found impaired glucose tolerance in both females and males at 10 months of age (Jaquiery et al., 2012; Todd et al., 2009). The changes in the males are associated with a decrease in POMC expression in their hypothalami, which we hypothesise as leading to decreased inhibition of food intake and reduced energy expenditure, causing the increase in fat mass. However, whilst measuring food intake would provide some vital data in assessing the impact of maternal undernutrition on the offspring’s neuropeptide network, it is extremely difficult to measure this in sheep as they do not have the same satiety response. They are also ruminants with an approximate 5kg biomass in the stomach, and thus the calculation of food intake is based on a complex calculation involving the assessment of the effort expended versus the feed reward gained rather than a simple measurement of food intake. It is important to note that the control male offspring at 5 years of age in our studies had approximately 2% of fat mass with approximately 70% of lean mass. These changes account for the small ratios of percentage fat mass to lean mass that we observed in chapter 5 and are consistent with previous observations by our collaborators (unpublished data).

At 10 months of age it was reported by our collaborators that the maternally undernourished male offspring had increased fat mass (Todd et al. 2009). It was also shown that both the male and female maternally undernourished offspring had an increased glucose area under the curve, with a reduced insulin response and a decreased insulin:glucose ratio at 10 months of age (Todd et al., 2009). However, these changes were more prevalent in the females rather than the males 95

(Todd et al. 2009). It could be hypothesised that the increase in GR expression is acting on neuropeptides that can influence peripheral glucose regulation and pancreatic secretion (Marino et al., 2011; Sohn et al., 2013). Indeed, one investigation found that local administration of a synthetic glucocorticoid into the arcuate nucleus can lead to reduced hepatic insulin sensitivity (Yi et al., 2012). It has also been suggested that the impaired glucose tolerance could be a result of alterations at the level of the pancreas, with the decreased insulin area-under-the-curve and the reduced insulin:glucose ratio being attributed to deficient insulin production and secretion. Furthermore, work in the rodent model has identified that offspring subject to a maternal low protein diet have pancreatic islets that have a decreased response to glucose (Theys et al., 2009) and less pancreatic β cell proliferation in new-borns (Chamson-Reig et al., 2006; Snoeck et al., 1990). Accordingly, the full extent of pancreatic changes is currently being investigated by our collaborators. However, more work also needs to be done in determining the role of the hypothalamus in these processes in our animals.

As well as the changes outlined above, extensive work by our collaborators has shown various phenotypic outcomes in this offspring subject to periconceptional maternal undernourishment. For example, they established that by targeting certain times of development, the offspring’s blood pressure control and placental function was altered (Bloomfield et al., 2004; Rumball et al., 2008). Most notably they have demonstrated that maternally undernourished offspring have an altered HPA axis, with an increased drive towards a mature axis, with elevated cortisol levels leading to preterm birth (Bloomfield et al., 2004). We also focused on measuring potential changes in the HPA axis by measuring the GR and POMC epigenetic and expression status in the hippocampus and pituitary. Despite the fact that our collaborators observed an accelerated fetal HPA axis in the maternally undernourished offspring, we did not find any changes in the expression of POMC and GR in the pituitary at the fetal stage. However, as detailed in previous sections, we did find epigenetic modifications in POMC and expression changes in GR in the pituitary and hippocampus in the maternally undernourished adult offspring. The results suggest a number of complex interactions that could lead to altered HPA axis feedback in these animals.

The HPA axis functionality in these offspring has been further characterised by our collaborator. They found that despite the increased maturation of the HPA axis at the fetal level, the maternally undernourished adult animals present with a potentially suppressed HPA axis with decreased cortisol levels. Interestingly, in response to a metyrapone challenge, there was an increase in ACTH, but this did not result in an increase in cortisol release from the adrenal gland 96

(unpublished data). Overall, this data suggests that the maternally undernourished animals used in our studies have reduced adrenal responsiveness to ACTH, which could be due to impaired adrenal steroidogenesis and enzyme responses. This is associated with a suppressed HPA axis with decreased behavioural actions in response to isolation stress being found in maternally undernourished offspring at 18 months of age (Oliver et al., 2012).

7.9 Mechanisms of action underlying epigenetic changes in the offspring as a consequence of maternal programming

Several mechanisms have been hypothesised for the development of epigenetic changes contributing to disease outcomes in maternally programmed offspring. In our model it could be speculated that the timing of the insult during early embryonic progression may be impacting on the epigenetic configuration of key genes in developing tissues. One of the ways that this may occur is a reduction in methyl donors, due to the lack of folic acid in the overall diet. For example, Lillycrop et al. (2005) found that maternally nutrient restricted fetuses had a significant decrease in the levels of GR methylation in the liver. These changes were not seen in animals on the same restricted maternal diet, which was supplemented with folic acid, preventing GR hypomethylation. This may in part be explained by folic acid being a necessary cofactor in the transformation of homocysteine to methionine. Methionine is then metabolized to S-adenosyl-l- methionine (SAM), which is the principal methyl donor for DNA methylation (Fang and Xiao, 2003). Hence, any alterations in the levels of methyl donors could result in detrimental changes in the epigenetic signature.

In our model it has been shown that there is a significant decrease in the levels of the glucocorticoid inactivating enzyme, 11beta hydroxysteroid dehydrogenase type-2 in the placenta (Connor et al., 2009). As a result, the fetus could be exposed to greater levels of glucocorticoids which have been shown to extensively interact with epigenetic machinery; thus, contact of the fetus with excess glucocorticoids could alter the epigenetic foundation positioned during development (Crudo et al., 2013). However, despite these explanations, more work needs to be carried out to determine the mechanisms by which altering the maternal diet can impact on the epigenetic foundation in the offspring.

7.10 Conclusion

In summary, the investigations described in this thesis demonstrate that pre- and post- conceptional moderate maternal undernutrition can induce fetal programming within the hypothalamic energy regulating pathway. These changes have the potential to persist into adulthood, with further modifications of this pathway and others, such as the HPA axis, becoming apparent postnatally. These changes are associated with increased fat mass gain in the males, altered glucose regulation and suppression of the HPA axis in the male and female adult offspring. The possible underlying cause for this is the modified epigenetic status of key genes such as the glucocorticoid receptor necessary for the regulation of many pathways in the body. However, the alterations are tissue specific and cannot be assessed in peripheral circulating leukocytes. Additionally, twinning has been identified to produce an altered hypothalamic epigenetic status in a part of the GR promoter similar to that of maternally undernourished singleton animals. This, together with additional evidence, suggests that twins may undergo a programming event themselves leading to altered physiology of hypothalamic pathways. Overall, the results reveal novel long term findings in the hypothalamus of maternally undernourished offspring, which may help to aid future research and lead to increased awareness and preventative care.

7.11 Future work

7.11.1 Analysis of the role of exon 1 of the GR gene in the hypothalamus

The data presented in this thesis highlights the potential of maternal undernutrition to epigenetically alter the GR gene in the hypothalami of maternally undernourished offspring, which progressed from fetal life to adulthood. Other modifications in this gene that become apparent after birth have also been identified in additional brain regions such as the hippocampus and pituitary These findings demonstrate the importance of determining changes in the GR gene in fetal and adult pathways in the offspring of mothers nutritionally programmed. As mentioned in the introduction, the GR gene is highly complex, consisting of approximately 9 tissue specific mRNA variants (Turner et al., 2008; Turner and Miller, 2005). The level of each of these mRNA transcripts contributing to total GR expression is variable and highly tissue specific (Turner et al. 2008). Each of these mRNA transcripts is regulated by its own specific CpG dense promoter region that is based in the exon 1 region of the GR gene. Whilst several other programming investigations have focused on the exon 1F region (Weaver et al., 2004; McGowan et al., 2009), our studies identified epigenetic changes in the GR exon 1E promoter region. The UCSC website suggests that the 1E region has a high level of promoter activity with an RNA POL II binding site. This highly conserved region is also predicted to contain a GRE, which has been shown to auto-upregulate GR gene expression (Geng et al., 2008; Turner et al., 2008). Supporting evidence for this has been provided in peripheral blood mononuclear cells (PBMCs), with only recent discoveries of exon 1E transcript level expression in regions of the human brain (Cao-Lei et al., 2013). Accordingly, throughout this thesis, data is presented indicating epigenetic changes in the GR exon 1E promoter region in the hypothalamus, hippocampus and pituitary, which suggest possible alterations in its associated mRNA transcript levels. However, we cannot categorically state that there are changes in exon 1E transcript levels as we were unable to obtain the sheep sequence to confirm this. Nevertheless, we were able to measure total GR mRNA expression, by using primers to a sequence that was common to all transcripts. As a result the true implications of exon 1E promoter changes on GR exon 1E transcript levels contributing to overall GR expression cannot be determined. Thus, to further our studies, if the sheep sequence for GR could be elucidated, a more comprehensive analysis of all the alternative GR exon 1 transcripts should be undertaken by qRT-PCR. Quantifying the transcript levels will further our knowledge of the tissue specific regulation and expression of the mRNA transcripts in the

99 arcuate nucleus of the hypothalamus. This will also allow us to determine which area of the GR gene is more susceptible to modifications as a consequence of maternal undernutrition that lead to subsequent changes in GR expression within the offspring’s hypothalamus, hippocampus and pituitary.

As well as measuring the individual transcript level of the GR exon 1E promoter region, a detailed examination of epigenetic changes in this region and its impact on GR activity should be undertaken. This should be coupled with studies to determine the potential of GR to auto up- regulate expression via this region. To carry out these investigations, the work could be undertaken in a hypothalamic cell line that expresses GR and associated neuropeptides. To do this our selected region of interest could be inserted into a vector with a CpG free backbone. As demonstrated by other studies, this vector would then be treated with methylating agents to induce methylation of all the CpG bases within this region (McGowan et al., 2009; Murgatroyd et al., 2009). The methylated vector could be transformed into E.coli and the effect of methylation could be determined by monitoring luciferease reporter gene activity. These results will provide evidence of the potential influence of epigenetic modifications on GR promoter activity. These experiments rely on the cell line expressing the correct transcription factors to bind to and activate the exon 1E region and so several different cell lines may have to be tried before a positive result is observed. Furthermore, if the region is predicted to bind GR then the GR gene may require activation by administration of synthetic glucocorticoids such as dexamethasone, to initiate binding to our region. Additionally, the experiments may have to consider inserting more than one copy of the GR exon 1E region into the CpG free vector, as one copy may not be sufficient enough to generate a signal. Consequently a construct that contains more than 3 copies of the region of interest in tandem may be required to increase signal intensity. A positive control for these experiments in the form of a vector consisting of a glucocorticoid responsive gene which is known to generate a response should also be utilised to ensure that the methodology is correct. These genes have either been established from the mouse mammary tumour virus long terminal repeat (MMTV), or are a synthetically generated construct consisting of three copies of the glucocorticoid responsive region of the human tyrosine amino- transferase (TAT3) gene.

Determining if GR can bind to the exon 1E promoter region to influence its expression is entirely more complex and is yet to be elucidated. Initially, a targeted deletion of the GRE sequence in the exon 1E area followed by glucocorticoid treatment should be carried out. This should impair 100 the ability of GR to bind to this region, reflecting the importance of GR to upregulate from this area. Following the deletion, the levels of total GR mRNA and GR exon 1E transcript levels could be measured to determine the contribution of the region to overall GR expression compared to controls. Extensive analysis of GR auto upregulation within other areas of the exon 1 region have already been identified and suggest that GR must interact with a complex of proteins to do this (Geng et al., 2008). This complex involves the transcription factors c- myeloblastosis (c-myb) and the ETS (E-twenty six) family. It has been hypothesised that c-myb and GR binding leads to upregulation and ETS binding can inhibit this process. In order to do this the c-myb and ETS transcription factor binding sites can overlap onto the GRE site or be adjacent to it. Within our region there is a predicted ETS binding site that overlaps with the putative GR site (Geng et al., 2008), however, an associated c-myb binding site has not yet been identified. Consequently, sequence analysis of surrounding areas of the exon 1E region should also be undertaken to determine the presence of these binding sites. If these sites are present, a series of competition and ChIP assays should be performed in several hypothalamic cell lines to determine if c-myb and ETS can bind to the region (Geng et al., 2008; Presul et al., 2007; Turner et al., 2006). These preliminary studies will help to support the hypothesis that the GR exon 1E region is important in aiding up regulation of GR expression, leading to altered regulation of pathways in aberrant disease systems such as those induced by maternal programming.

Whilst the GR exon 1E demonstrated clear epigenetic modifications in maternally undernourished offspring, it is possible that other promoter regions of the GR exon 1 gene are also epigenetically altered. As a result future work should focus on elucidating the epigenetic status of the promoters of the alternative exon 1 mRNA transcripts in our maternal programming paradigms. This could be carried out by bisulphite sequencing for methylation analysis and targeted ChIP analysis for a selection of histones related to open/closed chromatin conformation. These changes could then be correlated with the individual GR mRNA transcript levels. Furthermore, as most of the transcription factors binding to the other GR exon 1 regions have been identified (Turner et al., 2006) any modifications in those areas would also allow for easier interpretation.

Overall, the investigations above would provide a more complete analysis of the GR gene in our programming model which would increase our knowledge of GR expression and function within the arcuate nucleus. Also, it will improve our ability to interpret how changes in GR expression

101 can influence neuropeptide regulation leading to increased food intake and the development of obesity in the offspring.

7.11.2 POMC hypothalamic specific enhancer region

During our investigations, important publications on programming studies have highlighted the necessity for epigenetic analysis of the POMC gene (Coupé et al., 2010; Plagemann et al., 2009).The anorexigenic POMC gene in the hypothalamus is understood to be regulated by 2 key enhancer sites, npe1 and npe2. Again due to the lack of sequence availability, we were unable to investigate any epigenetic changes within these areas. However, we were able to analyse a region further downstream which in the pituitary is thought to be a predicted enhancer region with a binding site for GR that influences transcriptional regulation of this gene (Santangelo et al., 2007; Langlais et al., 2011). The extent to which this region can regulate GR expression within the hypothalamus is yet to be elucidated. As a result the overall impact of epigenetic changes within this predicted enhancer in the hypothalami of maternally undernourished offspring is difficult to interpret. To determine the effect of the region on POMC expression in the hypothalamus, a tissue specific mouse knockout animal could be generated. This has been extensively described in pioneering studies which detail the importance of the other enhancer regions (De Souza et al., 2005). This research would also be coupled with functional investigations delineating the potential of GR to bind to its predicted site in the POMC enhancer region that influences POMC expression. This could be determined by inserting our region of interest into a luciferase reporter gene construct and transfecting the plasmid construct into a POMC expressing hypothalamic cell line. The resultant recombinant cells can then be treated with glucocorticoids to activate GR binding to our region, followed by measurements of POMC activity. In addition the work could be carried out in an AtT20 mouse pituitary cell line, which also expresses POMC. This approach would go some way to determine if the suspected enhancer region is pituitary or hypothalamic specific. However, care must be taken in interpreting the data generated with these cell lines as it is not truly reflective of processes in the in vivo environment. As a result there may be certain pathways or transcription factors that are required to initiate the enhancer region that are not present in the cell lines and so no alteration in activity might be seen.

102

The programming studies that have focused on changes within the POMC promoter region have detailed different regions within the gene but none have investigated the npe1 and npe2 regions (Coupé et al., 2010; Plagemann et al., 2009; Shin et al., 2012). A further key publication which was revealed late into our final investigations, showed a strong association between the methylation status of CpG islands at the intron 2-exon 3 boundary of the POMC gene and the development of childhood obesity (Kuehnen et al., 2012). These studies demonstrate the necessity for our future work to carry out a detailed epigenetic examination of the whole of the POMC gene. For example, epigenetic alterations in the methylation of histones could be correlated with POMC mRNA expression. This would provide vital information in determining if there is an epigenetic signature within the POMC gene that confers increased susceptibility to the development of obesity in maternally undernourished offspring.

7.11.3 Candidate gene approach verses genome wide analysis

Throughout our studies we have undertaken a candidate gene approach. This method presented several advantages, allowing for focused investigations in deciphering the possible impact of epigenetic changes in the GR and POMC genes and the potential of tissue specificity. Additionally, it provided us with the opportunity to make more robust conclusions is associations between changes in gene expression and phenotypic outcomes in the maternally undernourished offspring. However, as elucidating epigenetic modifications was central to our studies, our approach did not allow for investigations into any possible modifications in the developmental trajectory of the offspring’s epigenome. The necessity for genome-wide analysis in aiding our further understanding of maternal programming has already been demonstrated (Crudo et al., 2013). This has led to the discovery of novel genes that may be affected in the offspring as a consequence of maternal programming and alterations in the fetal epigenetic landscape as a whole (Crudo et al., 2013; Thomassin et al., 2001). However, these studies have low statistical power and are more difficult to interpret, as the pathways in which a large number of genes function are still being elucidated (Amos et al., 2011). As a result, any programming studies we undertake in the future should consider both approaches and combine an initial genome-wide epigenetic analysis, that would lead on to a more focused candidate gene approach to decipher changes in the offspring’s pathways as a consequence of maternal programming. Furthermore, whilst those studies are placed at the level of the epigenome, expression arrays of neuropeptides involved in the energy regulating pathway should also be undertaken. This would not only allow 103 for an insight into potential changes in other genes in that pathway, but also suggest possible new targets for glucocorticoid regulation, as the full extent of the influence of glucocorticoid actions have not been fully elucidated.

7.11.4 Potential models of maternal nutritional programming

It is important to note that the sheep model was chosen to allow comparable conclusions to be made with human outcomes as it is more developmentally similar to the human fetus than rodents (Symonds and Budge, 2009). Additionally, sheep and humans are monotocous whereas rodents give birth to multiple offspring. However, the lack of sheep sequence and the lengthy gestation period make it difficult to provide a comprehensive analysis of other GR regions and neuropeptide targets in the energy regulating pathway. Therefore, future studies should focus upon using other models such as rodents, where the genome is freely available and the gestation period is shorter. Guinea pigs could also be utilised as they are more manageable and are more developmentally similar to humans (Kapoor and Matthews, 2005). As mentioned previously, access to the genome sequence would provide vital knowledge on the impact of maternal undernutrition on the offspring’s epigenetic trajectory. It would also allow us to carry out a more in depth analysis of the POMC and GR promoter regions as outlined above. Any changes observed in GR would strengthen and support our hypothesis of GR being a target of maternal programming and changes in this gene being fundamental to the increased susceptibility of the offspring to develop diseases in adulthood. However, in some rodent models the maternal insult must be extended over gestation and postnatally to induce effects within the offspring (Ikenasio- Thorpe et al., 2007). As a result, preliminary studies should be undertaken to determine if a model of moderate periconceptional maternal undernutrition can be established in rodents or guinea pigs. These studies could then be extended to investigate other time points during gestation and postnatally. Analysis of hypothalamic genes in the energy regulating pathways would then be undertaken to determine which period has the greatest impact and what effect this has on the offspring’s phenotypic outcomes. It would also be easier to measure food intake changes in these animals thereby providing further evidence towards an altered metabolic phenotype.

As mentioned previously, certain elements of the mother’s diet such as folic acid are required for the maintenance and establishment of DNA methylation (Fang and Xiao, 2003). As a result 104 undernourishment of the mother could lead to a reduction in these key metabolites altering the epigenetic framework in the offspring. Therefore, the rodent model could be utilised to further our understanding of these mechanisms as the rodent mother’s diet is much more easily manipulated than that of the sheep. For example Lillycrop et al. elegantly demonstrated that a nutrient restricted diet supplemented with folate can prevent methylation changes in GR in the offspring’s liver. Similar studies could be undertaken where the mothers are undernourished, after which epigenetic analysis of key genes within the offspring’s hypothalamic energy regulating pathway is carried out. If any changes are observed the study could be repeated, but with the mothers diet supplemented with folate to see if these changes are still present. This study would ascertain if folate is a limiting factor in the maternal diet.

During our adult study we wanted to elucidate the possibility of using non-invasive tissues such as leukocytes as biomarkers for changes that may be occurring in genes within regions of the brain. This topic requires a large amount of research and should be considered in all future studies. The use of rodents would allow for easy access to fresh blood samples which could be processed immediately for DNA, RNA and protein. The access to the rodent genomes permits widespread epigenetic analysis of target genes associated with diseases like obesity within maternally programmed offspring. Additionally these studies could include other nutritional paradigms such as maternal overnutrition and stress. These studies would aid the classification of epigenetic biomarkers within programmed offspring, which could then be translated into human samples so as to determine which offspring are maternally programmed.

Throughout our studies we carried out methylation enrichment of our region. Whilst this technique allows for easier quantification of total methylation changes within a given region it would have been more appropriate to carry out bisulphite sequencing. Additionally, the methylation enrichment kit can only detect large changes in methylation of at least 10 CpGs. As a result more subtle changes in methylation may not have been detected. Bisulphite sequencing would also have allowed us to pinpoint which CpG bases had changed and if there were any changes that existed in transcription factor binding sites. Doing this might have led to more accurate conclusions in the way in which GR and POMC promoter methylation would have been altered potentially affecting subsequent gene expression. However, bisulphite sequencing fails to detect the differences in the various forms of methylation changes whereas methylation enrichment is specific to 5-methylcytosine. Therefore, it can be concluded that all future

105 experiments should consider carrying out a combination of methylation enrichment and bisulphite sequencing as methods for detecting methylation changes

Overall, this thesis has set a precedent for future analysis of the GR and POMC genes as targets of maternal nutritional programming within the offspring. The additional research areas suggested will help to build on the conclusions that we have elucidated as well as providing insight into possible mechanisms that are inducing epigenetic changes. Furthermore, any additional modifications observed in the energy regulating pathway, will strengthen our hypothesis that this pathway is centrally implicated in the increased susceptibility of the offspring to develop obesity and diabetes in adulthood.

106

Chapter 8 References

107

Amos, W., Driscoll, E., and Hoffman, J.I. (2011). Candidate genes versus genome-wide associations: which are better for detecting genetic susceptibility to infectious disease? Proceedings. Biological sciences / The Royal Society 278, 1183-1188.

Angelogianni, P., Li, H.L., and Gianoulakis, C. (2000). Ontogenesis of proopiomelanocortin and its processing to beta-endorphin by the fetal and neonatal rat brain. Neuroendocrinology 72, 231- 241.

Anguita, R.M., Sigulem, D.M., and Sawaya, A.L. (1993). Intrauterine food restriction is associated with obesity in young rats. The Journal of nutrition 123, 1421-1428.

Arikawa, E., Sun, Y., Wang, J., Zhou, Q., Ning, B., Dial, S.L., Guo, L., Yang, J. (2008). Cross platform comparison of SYBR Green real-time PCR with TaqMan PCR, microarrays and other gene expression measurement technologies evaluated in the MicroArray Quality Control (MAQC) study. BMC Genomics 9,328-328.

Barker, D.J. (2005). The developmental origins of insulin resistance. Horm Res 64 Suppl 3, 2-7.

Barker, D.J., Eriksson, J.G., Forsen, T., and Osmond, C. (2002). Fetal origins of adult disease: strength of effects and biological basis. International journal of epidemiology 31, 1235-1239.

Barker, D.J. (2007). The origins of the developmental origins theory. J Intern Med 261(5), 412- 417.

Barker, D.J.P., Winter, P.D., Osmond, C., Margetts, B., Simmonds, S.J. (1989). Weight in infancy and death from ischaemic heart disease. Lancet 2, 577-580.

Barski, A., Cuddapah, S., Cui, K., Roh, T. Y., Schones, D. E., Wang, Z., Wei, G., Chepelev, I., and Zhao, K. (2007). High-resolution profiling of histone methylations in the human genome. Cell 129, 823-837.

Beaulieu, S., Gagne, B., and Barden, N. (1988). Glucocorticoid regulation of proopiomelanocortin messenger ribonucleic acid content of rat hypothalamus. Molecular endocrinology (Baltimore, Md.) 2, 727-731.

Berger, S. L. (2007). The complex language of chromatin regulation during transcription. Nature 447, 407-412.

Benjannet, S., Rondeau, N., Day, R., Chretien, M., and Seidah, N.G. (1991). PC1 and PC2 are proprotein convertases capable of cleaving proopiomelanocortin at distinct pairs of basic residues. Proc Natl Acad Sci U S A 88, 3564-3568.

Bergendahl, M., Wiemann, J.N., Clifton, D.K., Huhtaniemi, I., and Steiner, R.A. (1992). Short- term starvation decreases POMC mRNA but does not alter GnRH mRNA in the brain of adult male rats. Neuroendocrinology 56, 913-920.

Bernstein, B. E., Mikkelsen, T. S., Xie, X., Kamal, M., Huebert, D. J., Cuff, J., Fry, B., Meissner, A., Wernig, M., Plath, K., Jaenisch, R., Wagschal, A., Feil, R., Schreiber, S. L., and Lander, E. S. (2006). A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 125, 315-326.

108

Bernstein, B. E., Kamal, M., Lindblad-Toh, K., Bekiranov, S., Bailey, D. K., Huebert, D. J., McMahon, S., Karlsson, E. K., Kulbokas, E. J., 3rd, Gingeras, T. R., Schreiber, S. L., and Lander, E. S. (2005). Genomic maps and comparative analysis of histone modifications in human and mouse. Cell 120, 169-181.

Bilodeau, S., Vallette-Kasic, S., Gauthier, Y., Figarella-Branger, D., Brue, T., Berthelet, F., Lacroix, A., Batista, D., Stratakis, C., Hanson, J., Meij, B., Drouin, J. (2006). Role of Brg1 and HDAC2 in GR trans-repression of the pituitary POMC gene and misexpression in Cushing disease. Genes Dev 20, 2871-2886.

Birch, N.P., Tracer, H.L., Hakes, D.J., Loh, Y.P. (1991). Coordinate regulation of mRNA levels of pro-opiomelanocortin and the candidate processing enzymes PC2 and PC3, but not furin, in rat pituitary intermediate lobe. Biochemical and Biophysical Research Communications 179, 1311-1319.

Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes & development 16, 6- 21.

Blondel, B., Kogan, M.D., Alexander, G.R., Dattani, N., Kramer, M.S., Macfarlane, A., and Wen, S.W. (2002). The impact of the increasing number of multiple births on the rates of preterm birth and low birthweight: an international study. Am J Public Health 92, 1323-1330.

Bloomfield, F.H., Oliver, M.H., Hawkins, P., Campbell, M., Phillips, D.J., Gluckman, P.D., Challis, J.R., and Harding, J.E. (2003). A periconceptional nutritional origin for noninfectious preterm birth. Science (New York, N.Y.) 300, 606.

Bloomfield, F.H., Oliver, M.H., Hawkins, P., Holloway, A.C., Campbell, M., Gluckman, P.D., Harding, J.E., and Challis, J.R. (2004). Periconceptional undernutrition in sheep accelerates maturation of the fetal hypothalamic-pituitary-adrenal axis in late gestation. Endocrinology 145, 4278-4285.

Bouret, S.G., Simerly, R.B. (2004). Minireview: Leptin and Development of Hypothalamic Feeding Circuits. Endocrinology 145, 2621-2626.

Bouret, S.G., Draper, S.J., and Simerly, R.B. (2004). Trophic action of leptin on hypothalamic neurons that regulate feeding. Science (New York, N.Y.) 304, 108-110.

Bouret, S.G., Gorski, J.N., Patterson, C.M., Chen, S., Levin, B.E., and Simerly, R.B. (2008). Hypothalamic neural projections are permanently disrupted in diet-induced obese rats. Cell metabolism 7, 179-185.

Brady, L.S., Smith, M.A., Gold, P.W., Herkenham, M. (1990). Altered expression of hypothalamic neuropeptide mRNAs in food-restricted and food-deprived rats. Neuroendocrinology 52, 441-447.

Breton, C., Lukaszewski, M.A., Risold, P.Y., Enache, M., Guillemot, J., Riviere, G., Delahaye, F., Lesage, J., Dutriez-Casteloot, I., Laborie, C., and Vieau, D. (2009). Maternal prenatal undernutrition alters the response of POMC neurons to energy status variation in adult male rat offspring. American journal of physiology. Endocrinology and metabolism 296, E462-472.

109

Brown, L.M., Clegg, D.J., Benoit, S.C., and Woods, S.C. (2006). Intraventricular insulin and leptin reduce food intake and body weight in C57BL/6J mice. Physiology & behavior 89, 687- 691.

Buckler, J. M., and Green, M. (2004). A comparison of the early growth of twins and singletons. Ann Hum Biol 31, 311-332.

Buckley, A., Jaquiery, A., Harding, J. (2005). Nutritional programming of adult disease. Cell Tissue Res 322, 73-79.

Buijs, R.M., Chun, S.J., Niijima, A., Romijn, H.J., and Nagai, K. (2001). Parasympathetic and sympathetic control of the pancreas: a role for the suprachiasmatic nucleus and other hypothalamic centers that are involved in the regulation of food intake. J Comp Neurol 431, 405- 423.

Buitelaar, J.K., Huizink, A.C., Mulder, E.J., de Medina, P.G., and Visser, G.H. (2003). Prenatal stress and cognitive development and temperament in infants. Neurobiol Aging 24 Suppl 1, S53- 60; discussion S67-58.

Bumaschny, V.F., Yamashita, M., Casas-Cordero, R., Otero-Corchon, V., de Souza, F.S., Rubinstein, M., and Low, M.J. (2012). Obesity-programmed mice are rescued by early genetic intervention. J Clin Invest 122, 4203-4212.

Cao-Lei, L., Suwansirikul, S., Jutavijittum, P., Meriaux, S.B., Turner, J.D., and Muller, C.P. (2013). Glucocorticoid receptor gene expression and promoter CpG modifications throughout the human brain. Journal of psychiatric research.

Challis, B.G., Pritchard, L.E., Creemers, J.W., Delplanque, J., Keogh, J.M., Luan, J., Wareham, N.J., Yeo, G.S., Bhattacharyya, S., Froguel, P., et al. (2002). A missense mutation disrupting a dibasic prohormone processing site in pro-opiomelanocortin (POMC) increases susceptibility to early-onset obesity through a novel molecular mechanism. Hum Mol Genet 11, 1997-2004.

Challis, B.G., and Yeo, G.S. (2002). Past, present and future strategies to study the genetics of body weight regulation. Brief Funct Genomic Proteomic 1, 290-304.

Challis, J.R., Bloomfield, F.H., Bocking, A.D., Casciani, V., Chisaka, H., Connor, K., Dong, X., Gluckman, P., Harding, J.E., Johnstone, J., Li, W., Lye, S., Okamura, K., Premyslova, M. (2005). Fetal signals and parturition. J Obstet Gynaecol Res 31, 492-499.

Chamson-Reig, A., Thyssen, S.M., Arany, E., and Hill, D.J. (2006). Altered pancreatic morphology in the offspring of pregnant rats given reduced dietary protein is time and gender specific. The Journal of endocrinology 191, 83-92.

Chen, H., Simar, D., Lambert, K., Mercier, J., and Morris, M.J. (2008). Maternal and postnatal overnutrition differentially impact appetite regulators and fuel metabolism. Endocrinology 149, 5348-5356.

Clegg, D.J., Riedy, C.A., Smith, K.A., Benoit, S.C., and Woods, S.C. (2003). Differential sensitivity to central leptin and insulin in male and female rats. Diabetes 52, 682-687.

110

Cone, R.D., Ly, D., Koppula, S, Vage. D.I., Klungland, H., Boston, B., Chen, W., Poulton, C., Kesterton, R.A. (1996). The melanocortin receptors: agonists, antagonists, and the hormonal control of pigmentation. Recent Prog Horm Res 51, 287-317.

Connor, K.L., Challis, J.R., van Zijl, P., Rumball, C.W., Alix, S., Jaquiery, A.L., Oliver, M.H., Harding, J.E., and Bloomfield, F.H. (2009). Do alterations in placental 11beta-hydroxysteroid dehydrogenase (11betaHSD) activities explain differences in fetal hypothalamic-pituitary- adrenal (HPA) function following periconceptional undernutrition or twinning in sheep? Reproductive sciences (Thousand Oaks, Calif.) 16, 1201-1212.

Coupe, B., Amarger, V., Grit, I., Benani, A., and Parnet, P. (2010). Nutritional programming affects hypothalamic organization and early response to leptin. Endocrinology 151, 702-713.

Coupe, B., and Bouret, S.G. (2013). Development of the hypothalamic melanocortin system. Frontiers in endocrinology 4, 38.

Cowley, M.A., Smart, J.L., Rubinstein, M., Cerdan, M.G., Diano, S., Horvath, T.L., Cone, R.D., and Low, M.J. (2001). Leptin activates anorexigenic POMC neurons through a neural network in the arcuate nucleus. Nature 411, 480-484.

Cripps, R.L., Martin-Gronert, M.S., Archer, Z.A., Hales, C.N., Mercer, J.G., and Ozanne, S.E. (2009). Programming of hypothalamic neuropeptide gene expression in rats by maternal dietary protein content during pregnancy and lactation. Clinical science (London, England : 1979) 117, 85-93.

Cripps, R.L., Martin-Gronert, M.S., and Ozanne, S.E. (2005). Fetal and perinatal programming of appetite. Clinical science (London, England : 1979) 109, 1-11.

Crosby, S.R., Stewart, M.F., Ratcliffe, J.G., and White, A. (1988). Direct measurement of the precursors of adrenocorticotropin in human plasma by two-site immunoradiometric assay. J Clin Endocrinol Metab 67, 1272-1277.

Crudo, A., Suderman, M., Moisiadis, V.G., Petropoulos, S., Kostaki, A., Hallett, M., Szyf, M., and Matthews, S.G. (2013). Glucocorticoid programming of the fetal male hippocampal epigenome. Endocrinology 154, 1168-1180.

Dahl, J.A., and Collas, P. (2008). MicroChIP--a rapid micro chromatin immunoprecipitation assay for small cell samples and biopsies. Nucleic acids research 36, e15.

Dallman, M.F., la Fleur, S.E., Pecoraro, N.C., Gomez, F., Houshyar, H., and Akana, S.F. (2004). Minireview: glucocorticoids--food intake, abdominal obesity, and wealthy nations in 2004. Endocrinology 145, 2633-2638.

Dalrymple, B.P., Kirkness, E.F., Nefedov. M., McWilliam, S., Ratnakumar, A., Barris, W., Zhao, S., Shetty, J., Maddox, J.F,, O'Grady, M., Nicholas, F., Crawford, A.M., Smith, T., de Jong, P,J,, McEwan, J., Oddy, V.H., Cockett, N.E. (2007). Using comparative genomics to reorder the human genome sequence into a virtual sheep genome. Genome Biol 8,R152.

111

Dennis, J. H., Fan, H. Y., Reynolds, S. M., Yuan, G., Meldrim, J. C., Richter, D. J., Peterson, D. G., Rando, O. J., Noble, W. S., and Kingston, R. E. (2007). Independent and complementary methods for large-scale structural analysis of mammalian chromatin. Genome Res 17, 928-939 de Rooij, S.R., Painter, R.C., Phillips, D.I., Osmond, C., Michels, R.P., Godsland, I.F., Bossuyt, P.M., Bleker, O.P., and Roseboom, T.J. (2006). Impaired insulin secretion after prenatal exposure to the Dutch famine. Diabetes Care 29, 1897-1901. de Souza, F.S., Santangelo, A.M., Bumaschny, V., Avale, M.E., Smart, J.L., Low, M.J., and Rubinstein, M. (2005). Identification of neuronal enhancers of the proopiomelanocortin gene by transgenic mouse analysis and phylogenetic footprinting. Molecular and cellular biology 25, 3076-3086. de Weerth, C., van Hees, Y., and Buitelaar, J.K. (2003). Prenatal maternal cortisol levels and infant behavior during the first 5 months. Early human development 74, 139-151.

Delage, B., and Dashwood, R.H. (2008). Dietary manipulation of histone structure and function. Annual review of nutrition 28, 347-366.

Delahaye, F., Breton, C., Risold, P.Y., Enache, M., Dutriez-Casteloot, I., Laborie, C., Lesage, J., and Vieau, D. (2008). Maternal perinatal undernutrition drastically reduces postnatal leptin surge and affects the development of arcuate nucleus proopiomelanocortin neurons in neonatal male rat pups. Endocrinology 149, 470-475.

Desai, M., Gayle, D., Babu, J., and Ross, M.G. (2005). Programmed obesity in intrauterine growth-restricted newborns: modulation by newborn nutrition. American journal of physiology. Regulatory, integrative and comparative physiology 288, R91-96.

Drake, A.J., McPherson, R.C., Godfrey, K.M., Cooper, C., Lillycrop, K.A., Hanson, M.A., Meehan, R.R., Seckl, J.R., and Reynolds, R.M. (2012). An unbalanced maternal diet in pregnancy associates with offspring epigenetic changes in genes controlling glucocorticoid action and foetal growth. Clinical endocrinology 77, 808-815.

Drake, A.J., and Reynolds, R.M. (2010). Impact of maternal obesity on offspring obesity and cardiometabolic disease risk. Reproduction (Cambridge, England) 140, 387-398.

Drouin, J., Trifiro, M.A., Plante, R.K., Nemer, M., Eriksson, P., and Wrange, O. (1989). Glucocorticoid receptor binding to a specific DNA sequence is required for hormone-dependent repression of pro-opiomelanocortin gene transcription. Molecular and cellular biology 9, 5305- 5314.

Edwards, L.J., and McMillen, I.C. (2002). Impact of maternal undernutrition during the periconceptional period, fetal number, and fetal sex on the development of the hypothalamo- pituitary adrenal axis in sheep during late gestation. Biology of reproduction 66, 1562-1569.

Egger, G., Liang, G., Aparicio, A., and Jones, P.A. (2004). Epigenetics in human disease and prospects for epigenetic therapy. Nature 429, 457-463.

112

Eriksson, J.G., (2011). Early growth and coronary heart disease and type 2 diabetes: findings from the Helsinki Birth Cohort Study (HBCS). The American Journal of Clinical Nutrition 94(6 Suppl), 1799S-1802S.

Ernst, J., Kheradpour, P., Mikkelsen, T. S., Shoresh, N., Ward, L. D., Epstein, C. B., Zhang, X., Wang, L., Issner, R., Coyne, M., Ku, M., Durham, T., Kellis, M., and Bernstein, B. E. (2011). Mapping and analysis of chromatin state dynamics in nine human cell types. Nature 473, 43-49.

Fan, W., Boston, B.A., Kesterson, R.A., Hruby, V.J., and Cone, R.D. (1997). Role of melanocortinergic neurons in feeding and the agouti obesity syndrome. Nature 385, 165-168.

Fang, J.Y., and Xiao, S.D. (2003). Folic acid, polymorphism of methyl-group metabolism genes, and DNA methylation in relation to GI carcinogenesis. Journal of gastroenterology 38, 821-829.

Farooqi, I.S., Drop, S., Clements, A., Keogh, J.M., Biernacka, J., Lowenbein, S., Challis, B.G., and O'Rahilly, S. (2006). Heterozygosity for a POMC-null mutation and increased obesity risk in humans. Diabetes 55, 2549-2553.

Farooqi, I.S., Wangensteen, T., Collins, S., Kimber, W., Matarese, G., Keogh, J.M., Lank, E., Bottomley, B., Lopez-Fernandez, J., Ferraz-Amaro, I. (2007). Clinical and molecular genetic spectrum of congenital deficiency of the leptin receptor. The New England journal of medicine 356, 237-247.

Farre, D., Roset, R., Huerta, M., Adsuara, J. E., Rosello, L., Alba,M. M., and Messeguer, X. (2003). Identiﬁcation of patterns inbiological sequences at the ALGGEN server: PROMO and MALGEN. Nucleic Acids Res. 31, 3651–3653.

Feinberg, A.P., Ohlsson, R., and Henikoff, S. (2006). The epigenetic progenitor origin of human cancer. Nature reviews. Genetics 7, 21-33.

Ford, E.S., Giles, W,H., Dietz, W.H. (2002). Prevalence of the metabolic syndrome among US adults: findings from the third National Health and Nutrition Examination Survey. JAMA 287, 356-359.

Garcia, A.P., Palou, M., Priego, T., Sanchez, J., Palou, A., and Pico, C. (2010). Moderate caloric restriction during gestation results in lower arcuate nucleus NPY- and alphaMSH-neurons and impairs hypothalamic response to fed/fasting conditions in weaned rats. Diabetes, obesity & metabolism 12, 403-413.

Gardner, D.S., Van Bon, B.W., Dandrea, J., Goddard, P.J., May, S.F., Wilson, V., Stephenson, T., and Symonds, M.E. (2006). Effect of periconceptional undernutrition and gender on hypothalamic-pituitary-adrenal axis function in young adult sheep. The Journal of endocrinology 190, 203-212.

Geng, C.D., Schwartz, J.R., and Vedeckis, W.V. (2008). A conserved molecular mechanism is responsible for the auto-up-regulation of glucocorticoid receptor gene promoters. Molecular endocrinology (Baltimore, Md.) 22, 2624-2642.

113

Gibson, S., Crosby, S.R., Stewart, M.F., Jennings, A.M., McCall, E., White, A. (1994). Differential release of proopiomelanocortin-derived peptides from the human pituitary: evidence from a panel of two-site immunoradiometric assays. J Clin Endocrinol Metab 78, 835-841.

Gluckman, P.D., Hanson, M.A., Cooper, C., Thornburg, K.L., (2008). Effect of in utero and early-life conditions on adult health and disease. N Engl J Med 359(1), 61-73.

Godfrey, K.M., Barker, D.J.P. (2000). Fetal nutrition and adult disease. Am J Clinical Nutrition 71, 1344-1352

Goto, M., Arima, H., Watanabe, M., Hayashi, M., Banno, R., Sato, I., Nagasaki, H., and Oiso, Y. (2006). Ghrelin increases neuropeptide Y and agouti-related peptide gene expression in the arcuate nucleus in rat hypothalamic organotypic cultures. Endocrinology 147, 5102-5109.

Grayson, B.E., Levasseur, P.R., Williams, S.M., Smith, M.S., Marks, D.L., and Grove, K.L. (2010). Changes in melanocortin expression and inflammatory pathways in fetal offspring of nonhuman primates fed a high-fat diet. Endocrinology 151, 1622-1632.

Gupta, S., Dennis, J., Thurman, R. E., Kingston, R., Stamatoyannopoulos, J. A., and Noble, W. S. (2008). Predicting human nucleosome occupancy from primary sequence. PLoS Comput Biol 4, e1000134.

Guo, F., and Jen, K.L. (1995). High-fat feeding during pregnancy and lactation affects offspring metabolism in rats. Physiology & behavior 57, 681-686.

Gutteling, B.M., de Weerth, C., Willemsen-Swinkels, S.H., Huizink, A.C., Mulder, E.J., Visser, G.H., and Buitelaar, J.K. (2005). The effects of prenatal stress on temperament and problem behavior of 27-month-old toddlers. Eur Child Adolesc Psychiatry 14, 41-51.

Gyengesi, E., Liu, Z.W., D'Agostino, G., Gan, G., Horvath, T.L., Gao, X.B., and Diano, S. (2010). Corticosterone regulates synaptic input organization of POMC and NPY/AgRP neurons in adult mice. Endocrinology 151, 5395-5402.

Hagan, M.M., Rushing, P.A., Schwartz, M.W., Yagaloff, K.A., Burn, P., Woods, S.C., and Seeley, R.J. (1999). Role of the CNS melanocortin system in the response to overfeeding. The Journal of neuroscience : the official journal of the Society for Neuroscience 19, 2362-2367.

Hales, C.N., Barker, D.J., Clark, P.M., Cox, L.J., Fall, C., Osmond, C., and Winter, P.D. (1991). Fetal and infant growth and impaired glucose tolerance at age 64. BMJ (Clinical research ed.) 303, 1019-1022.

Hales, C.N., Barker, D.J. (1992). Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis. Diabetologia 35(7), 595-601.

Hales, C.N., Barker, D.J. (2001). The thrifty phenotype hypothesis. Br Med Bull 60, 5-20.

Hawkins, P., Hanson, M.A., and Matthews, S.G. (2001). Maternal undernutrition in early gestation alters molecular regulation of the hypothalamic-pituitary-adrenal axis in the ovine fetus. J Neuroendocrinol 13, 855-861.

114

Heasman, J. (2006). Maternal determinants of embryonic cell fate. Seminars in cell & developmental biology 17, 93-98.

Ho, S. M., and Tang, W. Y. (2007). Techniques used in studies of epigenome dysregulation due to aberrant DNA methylation: an emphasis on fetal-based adult diseases. Reprod Toxicol 23, 267-282.

Horvath, T.L., Bechmann, I., Naftolin, F., Kalra, S.P., and Leranth, C. (1997). Heterogeneity in the neuropeptide Y-containing neurons of the rat arcuate nucleus: GABAergic and non- GABAergic subpopulations. Brain research 756, 283-286.

Hyatt, M.A., Keisler, D.H., Budge, H., and Symonds, M.E. (2010). Maternal parity and its effect on adipose tissue deposition and endocrine sensitivity in the postnatal sheep. The Journal of endocrinology 204, 173-179.

Ibrahim, N., Bosch, M.A., Smart, J.L., Qiu, J., Rubinstein, M., Ronnekleiv, O.K., Low, M.J., and Kelly, M.J. (2003). Hypothalamic proopiomelanocortin neurons are glucose responsive and express K(ATP) channels. Endocrinology 144, 1331-1340.

Ikenasio-Thorpe, B.A., Breier, B.H., Vickers, M.H., and Fraser, M. (2007). Prenatal influences on susceptibility to diet-induced obesity are mediated by altered neuroendocrine gene expression. The Journal of endocrinology 193, 31-37.

Imhof, A., (2006). Epigenetic regulators and histone modification. Brief Funct Genomic Proteomic 5, 222-227.

Jaquiery, A.L., Oliver, M.H., Honeyfield-Ross, M., Harding, J.E., and Bloomfield, F.H. (2012). Periconceptional undernutrition in sheep affects adult phenotype only in males. Journal of nutrition and metabolism 2012, 123610.

Jeanrenaud, B., Rohner-Jeanrenaud, F., (2000). CNS-periphery relationships and body weight homeostasis: influence of the glucocorticoid status. Int J Obes Relat Metab Disord 24 Suppl 2:S74-76.

Jeffray, T.M., Matthews, S.G., Hammond, G.L., Challis, J.R. (1998). Divergent changes in plasma ACTH and pituitary POMC mRNA after cortisol administration to late-gestation ovine fetus. Am J Physiol 274,E417-E425.

Jones, A.P., and Friedman, M.I. (1982). Obesity and adipocyte abnormalities in offspring of rats undernourished during pregnancy. Science (New York, N.Y.) 215, 1518-1519.

Jones, P.A., and Baylin, S.B. (2002). The fundamental role of epigenetic events in cancer. Nature reviews. Genetics 3, 415-428.

Kaelin, W.G., Jr., and McKnight, S.L. (2013). Influence of metabolism on epigenetics and disease. Cell 153, 56-69.

Kalra, S.P. (2008). Disruption in the leptin-NPY link underlies the pandemic of diabetes and metabolic syndrome: new therapeutic approaches. Nutrition 24, 820-826.

115

Kapoor, A., Dunn, E., Kostaki, A., Andrews, M.H., and Matthews, S.G. (2006). Fetal programming of hypothalamo-pituitary-adrenal function: prenatal stress and glucocorticoids. The Journal of physiology 572, 31-44.

Kapoor, A., Leen, J., Matthews, S.G. (2008). Molecular regulation of the hypothalamic-pituitary- adrenal axis in adult male guinea pigs after prenatal stress at different stages of gestation. J Physiol 586, 4317-4326.

Kapoor, A., Kostaki, A., Janus, C., and Matthews, S.G. (2009). The effects of prenatal stress on learning in adult offspring is dependent on the timing of the stressor. Behav Brain Res 197, 144- 149.

Kapoor, A., and Matthews, S.G. (2005). Short periods of prenatal stress affect growth, behaviour and hypothalamo-pituitary-adrenal axis activity in male guinea pig offspring. The Journal of physiology 566, 967-977.

Karssen, A.M., Meijer, O.C., van der Sandt, I.C., De Boer, A.G., De Lange, E.C., and De Kloet, E.R. (2002). The role of the efflux transporter P-glycoprotein in brain penetration of prednisolone. The Journal of endocrinology 175, 251-260.

Ke, X., Schober, M.E., McKnight, R.A., O'Grady, S., Caprau, D., Yu, X., Callaway, C.W., and Lane, R.H. (2010). Intrauterine growth retardation affects expression and epigenetic characteristics of the rat hippocampal glucocorticoid receptor gene. Physiological genomics 42, 177-189.

Kirk, S.L., Samuelsson, A.M., Argenton, M., Dhonye, H., Kalamatianos, T., Poston, L., Taylor, P.D., and Coen, C.W. (2009). Maternal obesity induced by diet in rats permanently influences central processes regulating food intake in offspring. PloS one 4, e5870.

Konner, A.C., Janoschek, R., Plum, L., Jordan, S.D., Rother, E., Ma, X., Xu, C., Enriori, P., Hampel, B., Barsh, G.S., et al. (2007). Insulin action in AgRP-expressing neurons is required for suppression of hepatic glucose production. Cell metabolism 5, 438-449.

Koutcherov, Y., Mai, J.K., and Paxinos, G. (2003). Hypothalamus of the human fetus. J Chem Neuroanat 26, 253-270.

Krude, H., Gruters, A. (2000). Implications of proopiomelanocortin (POMC) mutations in humans: the POMC deficiency syndrome. Trends Endocrinol Metab 11, 15-22.

Kuehnen, P., Mischke, M., Wiegand, S., Sers, C., Horsthemke, B., Lau, S., Keil, T., Lee, Y.A., Grueters, A., and Krude, H. (2012). An Alu element-associated hypermethylation variant of the POMC gene is associated with childhood obesity. PLoS genetics 8, e1002543.

Kwong, W.Y., Wild, A.E., Roberts, P., Willis, A.C., and Fleming, T.P. (2000). Maternal undernutrition during the preimplantation period of rat development causes blastocyst abnormalities and programming of postnatal hypertension. Development (Cambridge, England) 127, 4195-4202.

116

Langlais, D., Couture, C., Sylvain-Drolet, G., and Drouin, J. (2011). A pituitary-specific enhancer of the POMC gene with preferential activity in corticotrope cells. Molecular endocrinology (Baltimore, Md.) 25, 348-359.

Le Clair, C., Abbi, T., Sandhu, H., Tappia, P.S. (2009). Impact of maternal undernutrition on diabetes and cardiovascular disease risk in adult offspring. Can J Physiol Pharmacol. 87(3), 161- 79.

Lesage, J., Blondeau, B., Grino, M., Breant, B., and Dupouy, J.P. (2001). Maternal undernutrition during late gestation induces fetal overexposure to glucocorticoids and intrauterine growth retardation, and disturbs the hypothalamo-pituitary adrenal axis in the newborn rat. Endocrinology 142, 1692-1702.

Li, C., McDonald, T.J., Wu, G., Nijland, M.J., and Nathanielsz, P.W. (2013). Intrauterine growth restriction alters term fetal baboon hypothalamic appetitive peptide balance. The Journal of endocrinology 217, 275-282.

Li, G., Zhang, Y., Wilsey, J.T., and Scarpace, P.J. (2005). Hypothalamic pro-opiomelanocortin gene delivery ameliorates obesity and glucose intolerance in aged rats. Diabetologia 48, 2376- 2385.

Liberman, S.A., Mashoodh, R., Thompson, R.C., Dolinoy, D.C., and Champagne, F.A. (2012). Concordance in hippocampal and fecal Nr3c1 methylation is moderated by maternal behavior in the mouse. Ecology and evolution 2, 3123-3131.

Lillycrop, K.A., Slater-Jefferies, J.L., Hanson, M.A., Godfrey, K.M., Jackson, A.A., and Burdge, G.C. (2007). Induction of altered epigenetic regulation of the hepatic glucocorticoid receptor in the offspring of rats fed a protein-restricted diet during pregnancy suggests that reduced DNA methyltransferase-1 expression is involved in impaired DNA methylation and changes in histone modifications. The British journal of nutrition 97, 1064-1073.

Lindblom, J., Haitina, T., Fredriksson, R., Schioth, H.B. (2005). Differential regulation of nuclear receptors, neuropeptides and peptide hormones in the hypothalamus and pituitary of food restricted rats. Brain Res Mol Brain Res 133, 37-46.

Livak, K.J., and Schmittgen, T.D. (2001). Analysis of relative gene expression data using real- time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods (San Diego, Calif.) 25, 402-408.

Long, N.M., Nathanielsz, P.W., and Ford, S.P. (2012). The impact of maternal overnutrition and obesity on hypothalamic-pituitary-adrenal axis response of offspring to stress. Domestic animal endocrinology 42, 195-202.

Longui, C.A., Vottero, A., Adamson, P.C., Cole, D.E., Kino, T., Monte, O., and Chrousos, G.P. (2000). Low glucocorticoid receptor alpha/beta ratio in T-cell lymphoblastic leukemia. Hormone and metabolic research 32, 401-406.

Lopez, M., Seoane, L.M., Tovar, S., Garcia, M.C., Nogueiras, R., Dieguez, C., and Senaris, R.M. (2005). A possible role of neuropeptide Y, agouti-related protein and leptin receptor isoforms in hypothalamic programming by perinatal feeding in the rat. Diabetologia 48, 140-148.

117

Ludwig, D.S., Mountjoy, K.G., Tatro, J.B., Gillette, J.A., Frederich, R.C., Flier, J.S., and Maratos-Flier, E. (1998). Melanin-concentrating hormone: a functional melanocortin antagonist in the hypothalamus. The American journal of physiology 274, E627-633.

MacLaughlin, S.M., Walker, S.K., Kleemann, D.O., Sibbons, J.P., Tosh, D.N., Gentili, S., Coulter, C.L., and McMillen, I.C. (2007). Impact of periconceptional undernutrition on adrenal growth and adrenal insulin-like growth factor and steroidogenic enzyme expression in the sheep fetus during early pregnancy. Endocrinology 148, 1911-1920.

Makimura, H., Mizuno, T.M., Roberts, J., Silverstein, J., Beasley, J., and Mobbs, C.V. (2000). Adrenalectomy reverses obese phenotype and restores hypothalamic melanocortin tone in leptin- deficient ob/ob mice. Diabetes 49, 1917-1923.

Marchington, D., Rothwell, N.J., Stock, M.J., York, D.A. 1983 Energy balance, diet-induced thermogenesis and brown adipose tissue in lean and obese (fa/fa) Zucker rats after adrenalectomy. J Nutr 113:1395-1402

Marino, J.S., Xu, Y., and Hill, J.W. (2011). Central insulin and leptin-mediated autonomic control of glucose homeostasis. Trends Endocrinol Metab 22, 275-285.

Markakis, E.A. (2002). Development of the neuroendocrine hypothalamus. Frontiers in neuroendocrinology 23, 257-291.

Matthews, S.G., and Phillips, D.I. (2012). Transgenerational inheritance of stress pathology. Experimental neurology 233, 95-101.

Maunakea, A.K., Nagarajan, R.P., Bilenky, M., Ballinger, T.J., D'Souza, C., Fouse, S.D., Johnson, B.E., Hong, C., Nielsen, C., Zhao, Y., et al. (2010). Conserved role of intragenic DNA methylation in regulating alternative promoters. Nature 466, 253-257.

McGowan, P.O., Sasaki, A., D'Alessio, A.C., Dymov, S., Labonte, B., Szyf, M., Turecki, G., and Meaney, M.J. (2009). Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nature neuroscience 12, 342-348.

McMillen, I.C., Schwartz, J., Coulter, C.L., and Edwards, L.J. (2004). Early embryonic environment, the fetal pituitary-adrenal axis and the timing of parturition. Endocrine research 30, 845-850.

Meaney, M.J., Szyf, M., and Seckl, J.R. (2007). Epigenetic mechanisms of perinatal programming of hypothalamic-pituitary-adrenal function and health. Trends Mol Med 13, 269- 277.

Meister, B. (2007). Neurotransmitters in key neurons of the hypothalamus that regulate feeding behavior and body weight. Physiology & behavior 92, 263-271.

Messeguer, X., Escudero, R., Farre, D., Nunez, O., Martinez, J.,and Alba, M. M. (2002). PROMO: detection of known transcription regulatory elements using species-tailored searches. Bioinformatics 18, 333–334.

118

Mizoguchi, Y., Kajiume, T., Miyagawa, S., Okada, S., Nishi, Y., Kobayashi, M. (2007). Steroid- dependent ACTH-produced thymic carcinoid: regulation of POMC gene expression by cortisol via methylation of its promoter region. Horm Res 67, 257-262.

Monrad, R. N., Grunnet, L. G., Rasmussen, E. L., Malis, C., Vaag, A., and Poulsen, P. (2009) Age-dependent nongenetic influences of birth weight and adult body fat on insulin sensitivity in twins. J Clin Endocrinol Metab 94, 2394-2399

Morris, M.J., and Chen, H. (2009). Established maternal obesity in the rat reprograms hypothalamic appetite regulators and leptin signaling at birth. International journal of obesity (2005) 33, 115-122.

Mostyn, A., and Symonds, M.E. (2009). Early programming of adipose tissue function: a large- animal perspective. The Proceedings of the Nutrition Society 68, 393-400.

Mountjoy, P.D., Bailey, S.J., and Rutter, G.A. (2007). Inhibition by glucose or leptin of hypothalamic neurons expressing neuropeptide Y requires changes in AMP-activated protein kinase activity. Diabetologia 50, 168-177.

Muhlhausler, B.S., Adam, C.L., Findlay, P.A., Duffield, J.A., and McMillen, I.C. (2006). Increased maternal nutrition alters development of the appetite-regulating network in the brain. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 20, 1257-1259.

Muhlhausler, B.S., Adam, C.L., Marrocco, E.M., Findlay, P.A., Roberts, C.T., McFarlane, J.R., Kauter, K.G., and McMillen, I.C. (2005). Impact of glucose infusion on the structural and functional characteristics of adipose tissue and on hypothalamic gene expression for appetite regulatory neuropeptides in the sheep fetus during late gestation. The Journal of physiology 565, 185-195.

Muhlhausler, B.S., Hancock, S.N., Bloomfield, F.H., and Harding, R. (2011). Are twins growth restricted? Pediatric research 70, 117-122.

Muhlhausler, B.S., McMillen, I.C., Rouzaud, G., Findlay, P.A., Marrocco, E.M., Rhind, S.M., and Adam, C.L. (2004). Appetite regulatory neuropeptides are expressed in the sheep hypothalamus before birth. J Neuroendocrinol 16, 502-507.

Murgatroyd, C., Patchev, A.V., Wu, Y., Micale, V., Bockmuhl, Y., Fischer, D., Holsboer, F., Wotjak, C.T., Almeida, O.F., and Spengler, D. (2009). Dynamic DNA methylation programs persistent adverse effects of early-life stress. Nature neuroscience 12, 1559-1566.

Muroya, S., Yada, T., Shioda, S., and Takigawa, M. (1999). Glucose-sensitive neurons in the rat arcuate nucleus contain neuropeptide Y. Neuroscience letters 264, 113-116.

Newell-Price, J., King, P., Clark, A.J. (2001). The CpG island promoter of the human proopiomelanocortin gene is methylated in nonexpressing normal tissue and tumors and represses expression. Mol Endocrinol 15, 338-348.

Newell-Price, J. (2003). Proopiomelanocortin gene expression and DNA methylation: implications for Cushing's syndrome and beyond. The Journal of endocrinology 177, 365-372.

119

Ng, S.F., Lin, R.C., Laybutt, D.R., Barres, R., Owens, J.A., and Morris, M.J. (2010). Chronic high-fat diet in fathers programs beta-cell dysfunction in female rat offspring. Nature 467, 963- 966.

Oakley, R.H., and Cidlowski, J.A. (2011). Cellular processing of the glucocorticoid receptor gene and protein: new mechanisms for generating tissue-specific actions of glucocorticoids. The Journal of biological chemistry 286, 3177-3184.

Ogden, C.L., Carroll, M.D., Kit, B.K., Flegal, K.M. (2012). Prevalence of Obesity in the United States, 2009–2010. NCHS Data Brief 82, 1-8.

Ohlsson, R., Bartkuhn, M., and Renkawitz, R. (2010). CTCF shapes chromatin by multiple mechanisms: the impact of 20 years of CTCF research on understanding the workings of chromatin. Chromosoma 119, 351-360.

Oken, E., Rifas-Shiman, S.L., Field, A.E., Frazier, A.L., and Gillman, M.W. (2008). Maternal gestational weight gain and offspring weight in adolescence. Obstet Gynecol 112, 999-1006.

Oliver, M.H., Bloomfield, F.H., Jaquiery, A.L., Todd, S.E., Thorstensen, E.B., and Harding, J.E. (2012). Periconceptional undernutrition suppresses cortisol response to arginine vasopressin and corticotropin-releasing hormone challenge in adult sheep offspring. Journal of Developmental Origins of Health and Disease 3, 52-58.

Ombelet, W., Martens, G., De Sutter, P., Gerris, J., Bosmans, E., Ruyssinck, G., Defoort, P., Molenberghs, G., and Gyselaers, W. (2006). Perinatal outcome of 12,021 singleton and 3108 twin births after non-IVF-assisted reproduction: a cohort study. Hum Reprod 21, 1025-1032.

Ozawa, H., Lion, J., Xie, C.X., Nishi, M., Steinbusch, H., Kawata, M. (1999). Down-regulation of ACTH and glucocorticoid receptor immunoreactivity in hypothalamic arcuate neurons after adrenalectomy in the rat. Neuroreport 10, 1571-1575.

Ozsolak, F., Song, J. S., Liu, X. S., and Fisher, D. E. (2007). High-throughput mapping of the chromatin structure of human promoters. Nat Biotechnol 25, 244-248.

Palou, M., Priego, T., Sanchez, J., Palou, A., and Pico, C. (2010). Sexual dimorphism in the lasting effects of moderate caloric restriction during gestation on energy homeostasis in rats is related with fetal programming of insulin and leptin resistance. Nutrition & metabolism 7, 69.

Parton, L.E., Ye, C.P., Coppari, R., Enriori, P.J., Choi, B., Zhang, C.Y., Xu, C., Vianna, C.R., Balthasar, N., Lee, C.E., Elmquist, J.K., Cowley, M.A., Lowell, B.B. (2007). Glucose sensing by POMC neurons regulates glucose homeostasis and is impaired in obesity. Nature 449, 228-232.

Patti, M.E. (2013). Intergenerational programming of metabolic disease: evidence from human populations and experimental animal models. Cell Mol Life Sci 70, 1597-1608.

Pelletier, G. (1993). Regulation of proopiomelanocortin gene expression in rat brain and pituitary as studied by in situ hybridization. Ann N Y Acad Sci 680, 246-59.

120

Phillips, D.I., Davies, M.J., and Robinson, J.S. (2001). Fetal growth and the fetal origins hypothesis in twins--problems and perspectives. Twin Res 4, 327-331.

Plagemann, A., Harder, T., Brunn, M., Harder, A., Roepke, K., Wittrock-Staar, M., Ziska, T., Schellong, K., Rodekamp, E., Melchior, K., and Dudenhausen, J.W. (2009). Hypothalamic proopiomelanocortin promoter methylation becomes altered by early overfeeding: an epigenetic model of obesity and the metabolic syndrome. The Journal of physiology 587, 4963-4976.

Plum, L., Belgardt, B.F., and Bruning, J.C. (2006). Central insulin action in energy and glucose homeostasis. J Clin Invest 116, 1761-1766.

Portela, A., and Esteller, M. (2010). Epigenetic modifications and human disease. Nat Biotechnol 28, 1057-1068.

Poulsen, P., Andersen, G., Fenger, M., Hansen, T., Echwald, S.M., Volund, A., Beck-Nielsen, H., Pedersen, O., and Vaag, A. (2003). Impact of two common polymorphisms in the PPARgamma gene on glucose tolerance and plasma insulin profiles in monozygotic and dizygotic twins: thrifty genotype, thrifty phenotype, or both? Diabetes 52, 194-198.

Poulsen, P., Levin, K., Petersen, I., Christensen, K., Beck-Nielsen, H., and Vaag, A. (2005). Heritability of insulin secretion, peripheral and hepatic insulin action, and intracellular glucose partitioning in young and old Danish twins. Diabetes 54, 275-283.

Poulsen, P., Grunnet, L. G., Pilgaard, K., Storgaard, H., Alibegovic, A., Sonne, M. P., Carstensen, B., Beck-Nielsen, H., and Vaag, A. (2009). Increased risk of type 2 diabetes in elderly twins. Diabetes 58, 1350-1355.

Poulsen, P., Vaag, A.A., Kyvik, K.O., Moller Jensen, D., and Beck-Nielsen, H. (1997). Low birth weight is associated with NIDDM in discordant monozygotic and dizygotic twin pairs. Diabetologia 40, 439-446.

Presul, E., Schmidt, S., Kofler, R., and Helmberg, A. (2007). Identification, tissue expression, and glucocorticoid responsiveness of alternative first exons of the human glucocorticoid receptor. Journal of molecular endocrinology 38, 79-90.

Pritchard, L.E., Turnbull, A.V., White, A. (2002). Pro-opiomelanocortin processing in the hypothalamus: impact on melanocortin signalling and obesity. J Endocrinol 172, 411-421.

Pritchard, L.E., Oliver, R.L., McLoughlin, J.D., Birtles, S., Lawrence, C.B., Turnbull, A.V., and White, A. (2003). Proopiomelanocortin-derived peptides in rat cerebrospinal fluid and hypothalamic extracts: evidence that secretion is regulated with respect to energy balance. Endocrinology 144, 760-766.

Pritchard, L.E., White, A. (2005). Agouti-related protein: more than a melanocortin-4 receptor antagonist? Peptides 26, 1759-1770.

Pritchard, L.E., and White, A. (2007). Neuropeptide processing and its impact on melanocortin pathways. Endocrinology 148, 4201-4207.

121

Quirke, L.D., Juengel, J.L., Tisdall, D.J., Lun, S., Heath, D.A., and McNatty, K.P. (2001). Ontogeny of steroidogenesis in the fetal sheep gonad. Biology of reproduction 65, 216-228.

Ravelli, A.C., van Der Meulen, J.H., Osmond, C., Barker, D.J., and Bleker, O.P. (1999). Obesity at the age of 50 y in men and women exposed to famine prenatally. The American journal of clinical nutrition 70, 811-816.

Ravelli, G.P., Stein, Z.A., and Susser, M.W. (1976). Obesity in young men after famine exposure in utero and early infancy. The New England journal of medicine 295, 349-353.

Razin, A., and Cedar, H. (1993) DNA methylation and embryogenesis. Exs 64, 343-357

Reik, W., Kelsey, G., and Walter, J. (1999). Dissecting de novo methylation. Nature genetics 23, 380-382.

Reynolds, R.M., Osmond, C., Phillips, D.I., and Godfrey, K.M. (2010). Maternal BMI, parity, and pregnancy weight gain: influences on offspring adiposity in young adulthood. J Clin Endocrinol Metab 95, 5365-5369.

Ribel-Madsen, R., Fraga, M.F., Jacobsen, S., Bork-Jensen, J., Lara, E., Calvanese, V., Fernandez, A.F., Friedrichsen, M., Vind, B.F., Hojlund, K., et al. (2012). Genome-wide analysis of DNA methylation differences in mu scle and fat from monozygotic twins discordant for type 2 diabetes. PloS one 7, e51302.

Rorato, R., Castro, M., Borges, B.C., Benedetti, M., Germano, C.M., Antunes-Rodrigues, J., and Elias, L.L. (2008). Adrenalectomy enhances endotoxemia-induced hypophagia: higher activation of corticotrophin-releasing-factor and proopiomelanocortin hypothalamic neurons. Hormones and behavior 54, 134-142.

Roseboom, T., de Rooij, S., and Painter, R. (2006). The Dutch famine and its long-term consequences for adult health. Early human development 82, 485-491.

Rousseau, K., Kauser, S., Pritchard, L.E., Warhurst, A., Oliver, R.L., Slominski, A., Wei, E.T., Thody, A.J., Tobin, D.J., and White, A. (2007). Proopiomelanocortin (POMC), the ACTH/melanocortin precursor, is secreted by human epidermal keratinocytes and melanocytes and stimulates melanogenesis. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 21, 1844-1856.

Rumball, C.W., Oliver, M.H., Thorstensen, E.B., Jaquiery, A.L., Husted, S.M., Harding, J.E., and Bloomfield, F.H. (2008). Effects of twinning and periconceptional undernutrition on late- gestation hypothalamic-pituitary-adrenal axis function in ovine pregnancy. Endocrinology 149, 1163-1172.

Rumball, C. W., Bloomfield, F. H., Oliver, M. H., and Harding, J. E. (2009). Different periods of periconceptional undernutrition have different effects on growth, metabolic and endocrine status in fetal sheep. Pediatr Res 66, 605-613.

Samuelsson, A.M., Matthews, P.A., Argenton, M., Christie, M.R., McConnell, J.M., Jansen, E.H., Piersma, A.H., Ozanne, S.E., Twinn, D.F., Remacle, C., et al. (2008). Diet-induced obesity

122 in female mice leads to offspring hyperphagia, adiposity, hypertension, and insulin resistance: a novel murine model of developmental programming. Hypertension 51, 383-392.

Sandman, C.A., Wadhwa, P.D., Chicz-DeMet, A., Dunkel-Schetter, C., and Porto, M. (1997). Maternal stress, HPA activity, and fetal/infant outcome. Ann N Y Acad Sci 814, 266-275.

Sandoval, D.A., Obici, S., and Seeley, R.J. (2009). Targeting the CNS to treat type 2 diabetes. Nature reviews. Drug discovery 8, 386-398.

Santangelo, A.M., de Souza, F.S., Franchini, L.F., Bumaschny, V.F., Low, M.J., and Rubinstein, M. (2007). Ancient exaptation of a CORE-SINE retroposon into a highly conserved mammalian neuronal enhancer of the proopiomelanocortin gene. PLoS genetics 3, 1813-1826.

Sapolsky, R.M., Romero, L.M., and Munck, A.U. (2000). How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocrine reviews 21, 55-89.

Sato, I., Arima, H., Ozaki, N., Watanabe, M., Goto, M., Hayashi, M., Banno, R., Nagasaki, H., and Oiso, Y. (2005). Insulin inhibits neuropeptide Y gene expression in the arcuate nucleus through GABAergic systems. The Journal of neuroscience : the official journal of the Society for Neuroscience 25, 8657-8664.

Savontaus, E., Conwell, I.M., and Wardlaw, S.L. (2002). Effects of adrenalectomy on AGRP, POMC, NPY and CART gene expression in the basal hypothalamus of fed and fasted rats. Brain research 958, 130-138.

Schneider, E., Pliushch, G., El Hajj, N., Galetzka, D., Puhl, A., Schorsch, M., Frauenknecht, K., Riepert, T., Tresch, A., Muller, A.M., et al. (2010). Spatial, temporal and interindividual epigenetic variation of functionally important DNA methylation patterns. Nucleic acids research 38, 3880-3890.

Schwartz, J., and Rose, J.C. (1998). Development of the pituitary adrenal axis in fetal sheep twins. The American journal of physiology 274, R1-8.

Sebaai, N., Lesage, J., Vieau, D., Alaoui, A., Dupouy, J.P., and Deloof, S. (2002). Altered control of the hypothalamo-pituitary-adrenal axis in adult male rats exposed perinatally to food deprivation and/or dehydration. Neuroendocrinology 76, 243-253.

Sebert, S.P., Hyatt, M.A., Chan, L.L., Patel, N., Bell, R.C., Keisler, D., Stephenson, T., Budge, H., Symonds, M.E., and Gardner, D.S. (2009). Maternal nutrient restriction between early and midgestation and its impact upon appetite regulation after juvenile obesity. Endocrinology 150, 634-641.

Setiawan, E., Jackson, M.F., MacDonald, J.F., Matthews, S.G. (2007). Effects of repeated prenatal glucocorticoid exposure on long-term potentiation in the juvenile guinea-pig hippocampus. J Physiol 581:1033-1042.

Sharkey, D., Gardner, D.S., Symonds, M.E., and Budge, H. (2009). Maternal nutrient restriction during early fetal kidney development attenuates the renal innate inflammatory response in obese young adult offspring. American journal of physiology. Renal physiology 297, F1199-1207.

123

Shimizu, H., Arima, H., Ozawa, Y., Watanabe, M., Banno, R., Sugimura, Y., Ozaki, N., Nagasaki, H., and Oiso, Y. (2010). Glucocorticoids increase NPY gene expression in the arcuate nucleus by inhibiting mTOR signaling in rat hypothalamic organotypic cultures. Peptides 31, 145-149.

Shin, B.C., Dai, Y., Thamotharan, M., Gibson, L.C., and Devaskar, S.U. (2012). Pre- and postnatal calorie restriction perturbs early hypothalamic neuropeptide and energy balance. Journal of neuroscience research 90, 1169-1182.

Siepel, A., Bejerano, G., Pedersen, J.S., Hinrichs, A.S., Hou, M., Rosenbloom, K., Clawson, H., Spieth, J., Hillier, L.W., Richards, S., et al. (2005). Evolutionarily conserved elements in vertebrate, insect, worm, and yeast genomes. Genome research 15, 1034-1050.

Sinclair, K.D., Allegrucci, C., Singh, R., Gardner, D.S., Sebastian, S., Bispham, J., Thurston, A., Huntley, J.F., Rees, W.D., Maloney, C.A., Lea, R.G., Craigon, J., McEvoy, T.G., Young, L.E. (2007). DNA methylation, insulin resistance, and blood pressure in offspring determined by maternal periconceptional B vitamin and methionine status. Proc Natl Acad Sci U S A 104, 19351-19356.

Sloboda, D.M., Moss, T.J., Li, S., Matthews, S.G., Challis, J.R., Newnham, J.P. (2008). Expression of glucocorticoid receptor, mineralocorticoid receptor, and 11beta-hydroxysteroid dehydrogenase 1 and 2 in the fetal and postnatal ovine hippocampus: ontogeny and effects of prenatal glucocorticoid exposure. J Endocrinol 197, 213-220.

Smart, J.L., Tolle, V., Low, M.J. (2006). Glucocorticoids exacerbate obesity and insulin resistance in neuron-specific proopiomelanocortin-deficient mice. J Clin Invest 116, 495-505.

Snoeck, A., Remacle, C., Reusens, B., and Hoet, J.J. (1990). Effect of a low protein diet during pregnancy on the fetal rat endocrine pancreas. Biology of the neonate 57, 107-118.

Sohn, J.W., Harris, L.E., Berglund, E.D., Liu, T., Vong, L., Lowell, B.B., Balthasar, N., Williams, K.W., and Elmquist, J.K. (2013). Melanocortin 4 receptors reciprocally regulate sympathetic and parasympathetic preganglionic neurons. Cell 152, 612-619.

Sterner, D.E., Berger, S.L. (2000). Acetylation of histones and transcription-related factors. Microbiol Mol Biol Rev 64, 435-459.

Stevens, A., White, A. (2009). ACTH: cellular peptide hormone synthesis and secretory pathways. Results Probl Cell Differ 50, 63-84.

Stevens, A., Begum, G., Cook, A., Connor, K., Rumball, C., Oliver, M., Challis, J., Bloomfield, F., and White, A. (2010). Epigenetic changes in the hypothalamic proopiomelanocortin and glucocorticoid receptor genes in the ovine fetus after periconceptional undernutrition. Endocrinology 151, 3652-3664.

Strack, A.M., Sebastian, R.J., Schwartz, M.W., and Dallman, M.F. (1995). Glucocorticoids and insulin: reciprocal signals for energy balance. The American journal of physiology 268, R142- 149.

124

Surani, M. A., Hayashi, K., and Hajkova, P. (2007). Genetic and epigenetic regulators of pluripotency. Cell 128, 747-762.

Symonds, M.E., and Budge, H. (2009). Nutritional models of the developmental programming of adult health and disease. The Proceedings of the Nutrition Society 68, 173-178.

Szyf, M., Weaver, I.C., Champagne, F.A., Diorio, J., Meaney, M.J. (2005). Maternal programming of steroid receptor expression and phenotype through DNA methylation in the rat. Front Neuroendocrinol 26, 139-162.

Szyf, M., Weaver, I., Meaney, M. (2007). Maternal care, the epigenome and phenotypic differences in behavior. Reprod Toxicol 24, 9-19.

Tamashiro, K.L., Terrillion, C.E., Hyun, J., Koenig, J.I., and Moran, T.H. (2009). Prenatal stress or high-fat diet increases susceptibility to diet-induced obesity in rat offspring. Diabetes 58, 1116-1125.

Tang, W.Y., and Ho, S.M. (2007). Epigenetic reprogramming and imprinting in origins of disease. Reviews in endocrine & metabolic disorders 8, 173-182.

Terroni, P.L., Anthony, F.W., Hanson, M.A., and Cagampang, F.R. (2005). Expression of agouti-related peptide, neuropeptide Y, pro-opiomelanocortin and the leptin receptor isoforms in fetal mouse brain from pregnant dams on a protein-restricted diet. Brain research. Molecular brain research 140, 111-115.

The, N. S., Adair, L. S., and Gordon-Larsen, P. (2010) A study of the birth weight-obesity relation using a longitudinal cohort and sibling and twin pairs. Am J Epidemiol 172, 549-557

Theys, N., Bouckenooghe, T., Ahn, M.T., Remacle, C., and Reusens, B. (2009). Maternal low- protein diet alters pancreatic islet mitochondrial function in a sex-specific manner in the adult rat. American journal of physiology. Regulatory, integrative and comparative physiology 297, R1516-1525.

Thomassin, H., Flavin, M., Espinas, M.L., and Grange, T. (2001). Glucocorticoid-induced DNA demethylation and gene memory during development. The EMBO journal 20, 1974-1983.

Todd, S.E., Oliver, M.H., Jaquiery, A.L., Bloomfield, F.H., and Harding, J.E. (2009). Periconceptional undernutrition of ewes impairs glucose tolerance in their adult offspring. Pediatric research 65, 409-413.

Tsigos, C., and Chrousos, G.P. (2002). Hypothalamic-pituitary-adrenal axis, neuroendocrine factors and stress. Journal of psychosomatic research 53, 865-871.

Tsukada, Y., and Nakayama, K.I. (2010). In vitro histone demethylase assay. Cold Spring Harbor protocols 2010, pdb prot5512.

Tummala, P., Mali, R.S., Guzman, E., Zhang, X., Mitton, K.P. (2010). Temporal ChIP-on-Chip of RNA-Polymerase-II to detect novel gene activation events during photoreceptor maturation. Mol Vis 16,252-71.:252-271.

125

Turner, J.D., and Muller, C.P. (2005). Structure of the glucocorticoid receptor (NR3C1) gene 5' untranslated region: identification, and tissue distribution of multiple new human exon 1. Journal of molecular endocrinology 35, 283-292.

Turner, J.D., Pelascini, L.P., Macedo, J.A., and Muller, C.P. (2008). Highly individual methylation patterns of alternative glucocorticoid receptor promoters suggest individualized epigenetic regulatory mechanisms. Nucleic acids research 36, 7207-7218.

Turner, J.D., Schote, A.B., Macedo, J.A., Pelascini, L.P., and Muller, C.P. (2006). Tissue specific glucocorticoid receptor expression, a role for alternative first exon usage? Biochemical pharmacology 72, 1529-1537.

Turner, B. M. (2009) Epigenetic responses to environmental change and their evolutionary implications. Philos Trans R Soc Lond B Biol Sci 364, 3403-3418

Udagawa, J., Hatta, T., Hashimoto, R., and Otani, H. (2007). Roles of leptin in prenatal and perinatal brain development. Congenital anomalies 47, 77-83.

Unterberger, A., Szyf, M., Nathanielsz, P.W., and Cox, L.A. (2009). Organ and gestational age effects of maternal nutrient restriction on global methylation in fetal baboons. Journal of medical primatology 38, 219-227. van Dommelen, P., de Gunst, M., van der Vaart, A., van Buuren, S., and Boomsma, D. (2008) Growth references for height, weight and body mass index of twins aged 0-2.5 years. Acta Paediatr 97, 1099-1104

Varela, L., and Horvath, T.L. (2012). Leptin and insulin pathways in POMC and AgRP neurons that modulate energy balance and glucose homeostasis. EMBO Rep 13, 1079-1086.

Villamor, E., and Cnattingius, S. (2006). Interpregnancy weight change and risk of adverse pregnancy outcomes: a population-based study. Lancet 368, 1164-1170.

Wardlaw, S.L., McCarthy, K.C., and Conwell, I.M. (1998). Glucocorticoid regulation of hypothalamic proopiomelanocortin. Neuroendocrinology 67, 51-57.

Weaver, I.C., Cervoni, N., Champagne, F.A., D'Alessio, A.C., Sharma, S., Seckl, J.R., Dymov, S., Szyf, M., and Meaney, M.J. (2004). Epigenetic programming by maternal behavior. Nature neuroscience 7, 847-854.

Weaver, I.C. (2009). Epigenetic effects of glucocorticoids. Semin Fetal Neonatal Med 14, 143- 150.

Williams, G., Harrold, J.A., and Cutler, D.J. (2000). The hypothalamus and the regulation of energy homeostasis: lifting the lid on a black box. The Proceedings of the Nutrition Society 59, 385-396.

Williams, K.W., and Elmquist, J.K. (2012). From neuroanatomy to behavior: central integration of peripheral signals regulating feeding behavior. Nature neuroscience 15, 1350-1355.

126

Xu, Y., Berglund, E.D., Sohn, J.W., Holland, W.L., Chuang, J.C., Fukuda, M., Rossi, J., Williams, K.W., Jones, J.E., Zigman, J.M., et al. (2010). 5-HT2CRs expressed by proopiomelanocortin neurons regulate insulin sensitivity in liver. Nature neuroscience 13, 1457- 1459.

Yaswen, L., Diehl, N., Brennan, M.B., Hochgeschwender, U. (1999). Obesity in the mouse model of pro-opiomelanocortin deficiency responds to peripheral melanocortin. Nat Med 5,1066-1070

York, D.A., Holt, S.J., Marchington, D. (1985). Regulation of brown adipose tissue thermogenesis by corticosterone in obese fa/fa rats. Int J Obes 9 Suppl 2, 89-95.

Young, J.I., Otero, V., Cerdan, M.G., Falzone, T.L., Chan, E.C., Low, M.J., Rubinstein, M. (1998). Authentic cell-specific and developmentally regulated expression of proopiomelanocortin genomic fragments in hypothalamic and hindbrain neurons of transgenic mice. J Neurosci 18, 6631-6640.

Yi, C.X., Foppen, E., Abplanalp, W., Gao, Y., Alkemade, A., la Fleur, S.E., Serlie, M.J., Fliers, E., Buijs, R.M., Tschop, M.H., and Kalsbeek, A. (2012). Glucocorticoid signaling in the arcuate nucleus modulates hepatic insulin sensitivity. Diabetes 61, 339-345.

Zakrzewska, K.E., Cusin, I., Stricker-Krongrad, A., Boss, O., Ricquier, D., Jeanrenaud, B., and Rohner-Jeanrenaud, F. (1999). Induction of obesity and hyperleptinemia by central glucocorticoid infusion in the rat. Diabetes 48, 365-370.

127

Chapter 9 Appendix

128

European Journal of Pharmacology 660 (2011) 194–201

Contents lists available at ScienceDirect

European Journal of Pharmacology

journal homepage: www.elsevier.com/locate/ejphar

Review Epigenetic changes in the hypothalamic pro-opiomelanocortin gene: A mechanism linking maternal undernutrition to obesity in the offspring?

Adam Stevens b,1, Ghazala Begum a,1, Anne White a,b,⁎ a Faculty of Life Sciences, University of Manchester, UK b Faculty of Medical and Human Sciences, University of Manchester, UK article info abstract

Article history: Maternal undernutrition is associated with programming of obesity in offspring. While previous evidence has Received 9 August 2010 linked programming to the hypothalamic, pituitary, and adrenal (HPA) axis it could also affect the Received in revised form 6 October 2010 hypothalamic neuropeptides which regulate food intake and energy balance. Alpha melanocyte stimulating Accepted 29 October 2010 hormone (αMSH), a key regulator of these neuronal pathways, is derived from pro-opiomelanocortin (POMC) Available online 3 January 2011 which is therefore a prime target for the programming of obesity. Several models of maternal undernutrition have identiﬁed changes in POMC in hypothalami from foetuses or offspring at various ages. These models Keywords: POMC have also shown that the offspring go on to develop obesity and/or glucose intolerance. It is our hypothesis Hypothalamus that programming leads to epigenetic changes in hypothalamic neuropeptide genes. Therefore when there is Glucocorticoid receptor subsequent increased food availability, the epigenetic changes could cause dysfunctional transcriptional Maternal undernutrition regulation of energy balance. We present evidence of epigenetic changes in the POMC gene promoter in foetal Programming hypothalami after peri-conceptional undernutrition. In this model there are also epigenetic changes in the Epigenetics hypothalamic glucocorticoid receptor with consequent up-regulation of the receptor which could lead to alterations in the regulation of POMC and neuropeptide Y (NPY) in the hypothalamus. Thus maternal undernutrition could cause epigenetic changes in the POMC and glucocorticoid receptor genes, in the foetal hypothalamus, which may predispose the offspring to altered regulation of food intake, energy expenditure and glucose homeostasis, later in life. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.

Contents

1. Introduction ...... 195 2. Programming of food intake and energy balance ...... 195 2.1. Neuropeptide regulation of food intake ...... 195 2.2. Regulation of POMC expression in the hypothalamus...... 195 2.3. The role of POMC in food intake and energy balance ...... 195 3. Adverse maternal nutrition programs hypothalamic feeding centres ...... 196 3.1. Programming of hypothalamic neuropeptides with maternal undernutrition in the rat ...... 196 3.2. Effects of maternal undernutrition on foetal programming in sheep ...... 196 3.3. Effects of maternal overnutrition on programming ...... 196 4. Epigenetic effects on hypothalamic POMC pathways regulating food intake and energy balance ...... 196 4.1. Epigenetic changes in the foetus associated with maternal undernutrition...... 197 4.2. Epigenetic changes in POMC after maternal overnutrition ...... 197 4.3. Modulation of hypothalamic POMC as a result of epigenetic changes in the glucocorticoid receptor ...... 197 5. Discussion ...... 199 Acknowledgements ...... 199 References ...... 200

⁎ Corresponding author. Faculties of Life Sciences and Medical and Human Sciences, Manchester Academic Health Sciences Centre, University of Manchester, 3.016 AV Hill Building, Manchester M13 9PT, UK. Tel.: +44 161 275 5178, +44 161 275 5180 (Secretary). E-mail address: [email protected] (A. White). 1 These authors contributed equally to this work.

0014-2999/$ – see front matter. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.ejphar.2010.10.111 A. Stevens et al. / European Journal of Pharmacology 660 (2011) 194–201 195

1. Introduction

Hypothalamic POMC is implicated in the development of both obesity and diabetes, each of which has a major impact on human health. Indeed, metabolic syndrome, which includes obesity, cardiovascular disease and diabetes, affects up to 25% of the population in the USA (Ford et al., 2002). Although there is increasing evidence for candidate genes involved in the development of obesity there is still a lack of clarity of how this occurs in the majority of cases. One area of interest is the effect of maternal nutrition on obesity in the offspring. This association was identiﬁed in survivors of the Dutch famine of 1944–1945 where there was a correlation between maternal undernutrition and subsequent obesity in adult offspring (Ravelli et al., 1976). This led to the concept being termed maternal programming (Seckl, 2004). Programming of the foetus might occur to prepare the offspring for an adverse postnatal environment (Kapoor, 2006). However these changes can be associated with offspring having lower birth weight, increased propensity to develop an obese phenotype in later life and increased susceptibility to cardiovascular disease and diabetes (Martin-Gronert and Ozanne, 2005; Rhodes et al., 2009). There are many factors that might inﬂuence the outcome of developmental programming. These include the timing of the insult, the number of foetuses and the sex of the offspring. Furthermore the type of insult is also important, which along with stress (Seckl and Holmes, 2007) includes nutritional effects that could be mediated by the hypothalamus (Bertram and Hanson, 2002).

2. Programming of food intake and energy balance Fig. 1. The hypothalamic control of energy balance. ARC = arcuate nucleus, PVN = paraventricular nucleus; AGRP = agouti related peptide, POMC = pro-opiomelanocortin, CART = cocaine-amphetamine regulated transcript, NPY = neuropeptide Y,

2.1. Neuropeptide regulation of food intake MC4-R = melanocortin receptor 4, MC3-R = melanocortin receptor 3.

Energy homeostasis is tightly controlled in the hypothalamus despite fluctuations in energy consumption and expenditure. POMC-derived peptides (Cone, 2005; Pritchard et al., 2003, 2002; Therefore the foetal hypothalamic appetite regulatory network is a Pritchard and White, 2007). Hypothalamic POMC neuronal excitabil- prime candidate for maternal programming. Hypothalamic regula- ity, hypothalamic POMC mRNA and circulating αMSH are decreased in tion of food intake and energy balance in the arcuate nucleus (Fig. 1) rodent models of leptin or leptin-receptor deficiency (Forbes et al., depends on a complex array of neuropeptides but most is known 2001; Korner et al., 1999; Korner et al., 2001; Mizuno et al., 1998; about the anorexigenic pro-opiomelanocortin (POMC) and cocaine– Thornton et al., 1997; Turner et al., 2006) and this can be reversed by amphetamine regulated transcript expressing neurons and the leptin administration (Cowley et al., 2001; Forbes et al., 2001; Korner orexigenic neuropeptide Y and agouti related peptide expressing et al., 2001; Schwartz et al., 1997; Thornton et al., 1997). neurons (Challis and Yeo, 2002). The activation of POMC neurons It is well known that in the pituitary, POMC transcription is results in the cleavage of the prohormone, POMC, to αMSH and the inhibited by glucocorticoids. Hypothalamic expression of POMC can release of αMSH in the paraventricular nucleus (Pritchard et al., also be regulated by glucocorticoids, but there are differing observa- 2003, 2002; Pritchard and White, 2007). tions on how this occurs and these are discussed further in Section 4.3. In situations of positive energy balance, leptin released from Some POMC neurons also possess receptors for insulin and co- adipocytes, stimulates the anorexigenic pathway and inhibits the administration of the melanocortin MC4 receptor antagonist, orexigenic pathway (Plum et al., 2006; Pritchard and White, 2007). SHU9119, prevents the anorectic effects of insulin (Benoit et al., This results in a reduction in food intake. Conversely in situations of 2002). Reductions in POMC mRNA levels were observed in conjunc- negative energy balance, the orexigenic pathway is activated. This tion with decreased plasma insulin concentrations in adrenalecto- pathway acts to increase food intake in part by agouti related peptide, mised versus sham-operated rats demonstrating the regulatory effect preventing αMSH from binding to the melanocortin MC3 and MC4 of glucocorticoids on this process (Savontaus et al., 2002)(Fig. 1). receptors in the paraventricular nucleus (Cripps, 2005; Pritchard et al., Moreover, decreases in hypothalamic POMC mRNA caused by fasting 2004)(Fig. 1). It has been shown that components of adult energy were prevented by intracerebroventricular (i.c.v) injection of insulin balance regulation are present in the hypothalamus as early as mid- (Benoit et al., 2002). POMC neurons are also regulated by glucose and gestation in humans (Adam et al., 2008). POMC is expressed they express components of the glucose-sensing ATP-sensitive extensively in immature hypothalamic neurons in rat foetuses but potassium channels and respond to a reduction in extracellular subsequently half of this population of neurons adopts a non-POMC glucose concentration by decreasing neuronal firing (Gyte et al., 2007; fate, including development into neuropeptide Y neurons. Up to a Ibrahim et al., 2003). This suggests a key role for POMC in the central quarter of neuropeptide Y neurons have an immature POMC neuron melanocortin regulation of glucose homeostasis (Obici et al., 2001). progenitor (Padilla et al., 2010). 2.3. The role of POMC in food intake and energy balance 2.2. Regulation of POMC expression in the hypothalamus The importance of POMC in food intake and energy balance is Feeding and hormonal signals regulate expression of the POMC exemplified by mutations in the POMC gene resulting in an obese gene in the hypothalamus and therefore the associated secretion of phenotype which is primarily due to loss of ligands for melanocortin 196 A. Stevens et al. / European Journal of Pharmacology 660 (2011) 194–201

MC3 and MC4 receptors in the brain (Krude et al., 2003). Patients with 3.2. Effects of maternal undernutrition on foetal programming in sheep mutations in POMC demonstrate early-onset severe obesity and distinctive red hair (Krude et al., 1998, 2003; Krude and Gruters, Studying programming in sheep has the advantage that they have 2000). Targeted disruption of the POMC gene results in hyperphagia greater developmental similarity to humans because they give birth to and lower oxygen consumption in mice, causing increased fat mass mature young, have a high frequency of singleton pregnancies and and obesity (Challis et al., 2004; Yaswen et al., 1999). Mutations in the have a longer gestational period. Our studies have identified peptide processing enzyme prohormone convertase 1 render patients programming of foetal hypothalamic regulatory pathways and while unable to process POMC to functional peptides and are associated no change was observed in foetal hypothalamic POMC expression, with severe early-onset obesity (Jackson et al., 1997). Another there was increased glucocorticoid receptor expression which was mutation has been identified in the human POMC gene that can associated with maternal undernutrition (Stevens et al., 2010). Adult confer an inherited susceptibility to obesity through the production of sheep which have undergone a similar phase of maternal undernu- an aberrant fusion of beta MSH to beta endorphin that has the capacity trition were found to have increased body weight and glucose to interfere with central melanocortin signalling (Challis et al., 2002). intolerance (Rumball et al., 2009; Todd et al., 2009). Food deprivation results in significant decreases in hypothalamic In our study maternal ewes were undernourished to achieve a POMC mRNA levels and reduced release of POMC-derived peptides 10–15% reduction in body weight around conception (i.e. for 60 days from hypothalamic slices (Breen et al., 2005; Korner et al., 2001; Swart before to 30 days after conception) (Rumball et al., 2009; Stevens et al., 2002). Intracerebroventricular injection of POMC-derived et al., 2010). A more severe model of maternal undernutrition in the peptides inhibits food intake in rats, even after a 24-hour fast sheep used a 50% reduction in food intake between 30 and 80 days of (Poggioli et al., 1986; Vergoni et al., 1990). gestation. It was found that there was no effect on birth weight or growth of the offspring, and while there was no change in 3. Adverse maternal nutrition programs hypothalamic feeding hypothalamic POMC expression in the foetus, after one week of age centres hypothalamic POMC levels increased (Sebert et al., 2009).

Given that the epidemiological analyses of large cohorts of patients 3.3. Effects of maternal overnutrition on programming have suggested that maternal undernutrition impacts on obesity in the offspring (Rhodes et al., 2009), several groups have investigated It has been proposed that maternal obesity could exert a stronger how models of undernutrition during gestation affect food intake and impact in offspring than postnatal overnutrition and with the increase energy expenditure. in obesity in the population it is likely to be a significant problem in the future. In a study where male rats born to obese dams were 3.1. Programming of hypothalamic neuropeptides with maternal heavier than controls, they had a rather surprising change in undernutrition in the rat neuropeptides with increased hypothalamic POMC mRNA expression and reduced neuropeptide Y (Chen et al., 2008). Maternal hypergly- Prenatal undernutrition together with a postnatal high fat diet cemia was also associated with upregulated hypothalamic POMC in produces obese characteristics, but this did not occur with maternal the foetuses of sheep at mid-gestation (Muhlhausler et al., 2005). undernutrition alone (Ikenasio-Thorpe et al., 2007). This study suggests that the combination of changes in the prenatal and 4. Epigenetic effects on hypothalamic POMC pathways regulating postnatal environment, leads to dysregulation of the hypothalamic food intake and energy balance appetite regulatory network, which might contribute to an obese phenotype. Epigenetic modifications of key genes have been proposed as a The same model of maternal undernutrition has been used to potential underlying mechanism for foetal programming (Newell-Price, examine foetal hypothalamic anorexogenic pathways. Changes in 2003). Epigenetic modifications regulate chromatin assembly, chromo- hypothalamic gene expression and neuronal activation were identi- some separation, the replication and repair of DNA and gene expression. fied which could lead to the development of obesity in later life. These All these processes are necessary for gametogenesis and the develop- changes included a reduction in postnatal plasma leptin levels, a ment of the foetus (Delage and Dashwood, 2008). decrease in POMC mRNA levels and a decrease in the number of POMC The state of the chromatin structure, with its complex of histone nerve fibre projections from the arcuate nucleus to the paraventri- proteins surrounded by DNA forming “bead-like” nucleosomes, cular nucleus (Delahaye et al., 2008). Using a more severe model of determines whether it is transcriptionally activated or inactivated. undernutrition (70% reduction in maternal rat food intake from day 1 Open chromatin is associated with activation and the closed to day 21 of gestation), there were no differences in POMC and chromatin is inactive (McGowan et al., 2008). Thus, epigenetic neuropeptide Y mRNA expression between control and undernour- modifications act to transcriptionally silence or activate genes by ished adult rats. However, following fasting there was an increase in altering the chromatin structure without affecting the DNA sequence. neuropeptide Y mRNA and in the activity of POMC neurons in the Two methods by which this can occur are DNA methylation and arcuate nucleus (Breton et al., 2009). modification of histone tails (acetylation and methylation) (Ho and In rats, POMC is expressed as early as embryonic day 12 (E12) Tang, 2007)(Fig. 2). (Terroni et al., 2005). E12 is correlated with massive cell proliferation, DNA methylation is associated with gene silencing, preventing the leading to hypothalamic differentiation (Markakis, 2002) and POMC is DNA from undergoing transcription (Fig. 2). Its effects are thought to highly expressed in the immature neurons at this stage (Padilla et al., be permanent and it occurs at cytosine residues throughout the 2010). Furthermore, maternal protein restriction in rats from genome. Particularly dense regions of DNA methylation associated conception to E12 results in changes in the foetal hypothalamic with gene activity are termed CpG islands, which can be 200–2000 regulatory network, with increased expression of agouti related base pairs in length. The islands usually extend over transcriptional peptide, neuropeptide Y, POMC and the leptin receptor (Ob-Rb). start sites (Ho and Tang, 2007). Patterns of DNA methylation are set These changes may indicate that the appetite regulatory pathway is during foetal development. Initially non-specific demethylation active in the foetus (Terroni et al., 2005) although these pathways do occurs and then de novo methylation, followed by specific demethyla- not fully mature in the rat until after birth (Symonds and Budge, tion (Newell-Price, 2003). Thus, it is possible to speculate that the 2009). As a result the foetal developmental events observed in the rat patterns of methylation are potentially vulnerable to changes in the might be significantly different to those observed in humans. surrounding foetal environment. A. Stevens et al. / European Journal of Pharmacology 660 (2011) 194–201 197

suckling period, leading to long-lasting dysfunction in adulthood. These studies have shown that the hypothalamic control of appetite is a key target of perinatal developmental programming, possibly disturbing body weight set point (Coupe et al., 2010). There is also data to suggest that maternal undernutrition differentially affects the appetite regulatory system of offspring long-term via the response of POMC neurons to energy status and food intake (Breton et al., 2009). It could be that maternal nutritional changes cause epigenetic programming in the foetus. One investigation involved feeding a low protein diet to pregnant rat dams. Following methylation analysis, specific POMC CpG sites were found to be less methylated in the protein-restricted offspring compared to controls (Coupe et al., 2010). This study suggests that maternal nutritional insults can induce epigenetic programming in foetal pathways. Recently we have used a sheep model of maternal undernutrition and found epigenetic changes in foetal hypothalamic regulatory pathways (Stevens et al., 2010). A promoter region marker for the POMC gene was identified in the sheep using comparative homology and shown to be associated with histone H3K9 acetylation, implying that the gene was accessible (Fig. 4A). There was also hypomethyla- Fig. 2. Epigenetic modifications of chromatin. A) Methylation of cytosine bases in CpG tion of the POMC promoter marker in the hypothalami of foetuses islands leads to transcriptional repression. B) The modification of histone “tails” alters exposed to maternal undernutrition (Fig. 4B). However transcription transcriptional activity, usually acetylation is associated with transcriptional activation of the hypothalamic POMC gene was not increased (Fig. 4C and D). We and the methylation is associated with transcriptional repression. therefore suggested that the hypothalamic POMC gene may be primed for transcriptional activity which was not yet measurable prior to In contrast to DNA methylation, histone tail acetylation (and some parturition (Stevens et al., 2010). changes in histone methylation) are associated with transcriptional activation of a gene. Conversely histone tails can be modified in such a 4.2. Epigenetic changes in POMC after maternal overnutrition way as to cause transcriptional repression usually via deacetylation or methylation and this is correlated with inactive chromatin (Graff and Little is known about the association of epigenetic changes in the Mansuy, 2008). offspring with overfeeding in mothers. Neonatal overfeeding in rats Human POMC is known to have 2 CpG islands, one over exon 1 and was associated with hypomethylation of the hypothalamic POMC the associated promoter region and one downstream over exon 3 gene promoter at two sites involved with leptin and insulin (Newell-Price, 2003). It is possible to speculate that in the obese modulation of POMC expression. However despite the alteration in phenotype there would be alterations in the POMC methylation epigenetic status there was no change in POMC expression between status, which would lead to decreases in POMC expression. As a result control and overfed rats (Plagemann et al., 2009). It could be that the there could be an increase in food intake. Changes in POMC gene change in expression may become manifest later in life, resulting in promoter methylation states have been seen in tissues and cell lines the maternally overfed group becoming more susceptible to obesity. from human cancer and are associated with changes in POMC expression (Mizoguchi et al., 2007; Newell-Price et al., 2001; Ye 4.3. Modulation of hypothalamic POMC as a result of epigenetic changes et al., 2005). POMC transcriptional enhancer regions associated with in the glucocorticoid receptor hypothalamic expression have been identified 10–12 kb upstream of the start site of the POMC gene (de Souza et al., 2005). While the In our studies investigating epigenetic changes in the POMC POMC CpG islands do not overlap the regions associated with promoter marker after maternal undernutrition, we also investigated hypothalamic expression, we have defined further CpG rich regions epigenetic changes in the hypothalamic glucocorticoid receptor and 4 kb downstream of the hypothalamic region which bind RNA found hypomethylation of this marker associated with a variety of polymerase II and are associated with acetylated histone H3K9 periconceptional maternal undernutrition regimens (Fig. 5A); along (Stevens et al., 2010) implying that epigenetic changes may have a with increased hypothalamic expression of the glucocorticoid role in the modulation of hypothalamic POMC gene expression receptor (Fig. 5B) within the arcuate nucleus of the foetal hypothal- (Fig. 3). amus (Fig. 5C) (Stevens et al., 2010). Therefore we cannot discount the possibility that the prime candidate for epigenetic changes is the 4.1. Epigenetic changes in the foetus associated with maternal glucocorticoid receptor which mediates regulation of a number of undernutrition hypothalamic neuropeptides including POMC and neuropeptide Y. While there is a well-defined feedback mechanism for glucocor- In rodents, several studies suggest that hypothalamic program- ticoid (Gc) regulation of POMC in the pituitary, the effects of Gcs on ming begins in utero but continues in early postnatal life during the hypothalamic POMC are controversial. In obese rats, adrenalectomy

Fig. 3. The human POMC gene promoter region. nPE1 = neuronal POMC enhancer 1, nPE2 = neuronal POMC enhancer 2; nPE1 and nPE2 are associated with hypothalamic transcription of POMC (de Souza et al., 2005). Position of CpG islands as deﬁned by Newell-Price (2003). The position of the pituitary promoter region is deﬁned by Bilodeau et al. (2006). The 160 bp marker region was used in our studies on the sheep POMC promoter (Stevens et al., 2010). 198 A. Stevens et al. / European Journal of Pharmacology 660 (2011) 194–201

Fig. 4. Epigenetic changes associated with the POMC gene in the foetal hypothalamus. A) Chromatin Immunoprecipitation (ChIP) of acetylated histone H3K9 as a marker of transcriptionally active DNA. An antibody to RNA polymerase II was used as a positive control and mouse IgG as a negative control. Ratio of PCR signal from H3K9Ac immunoprecipitated DNA to total genomic DNA for DNA puriﬁed from gestational age day 131 ventral hypothalamic sections (Control n=9; Underfed n=11). B) Ratio of POMC gene promoter marker PCR signal from methylated genomic DNA to total genomic DNA for ventral hypothalamic sections at gestational age day 131 (Control n=9; Underfed n=11). C) Expression levels of POMC in total RNA puriﬁed from ventral hypothalamic sections at gestational age day 131 enriched for arcuate nucleus (Control n=9; Underfed n=11). D) Immunohistochemistry for foetal hypothalamic POMC was detected using our monoclonal antibody (A1H5 — unpublished data) and labelled with TRITC secondary antibody scale bar=200 μm, 3V=3rd ventricle, ARC=arcuate nucleus, ***=Pb0.001.

(ADX) resulted in a reduction in food intake and body weight Therefore in the hypothalamus, glucocorticoid mediated regulation of suggesting Gcs may increase food intake. Furthermore the above POMC and other neuropeptides could be influenced by epigenetic observations were reversed following the administration of Gcs changes in the glucocorticoid receptor. (Strack et al., 1995). When normal rats were continuously adminis- Evidence of modifications of the epigenetic status of the tered Gcs for three days, there was an increase in food intake and body glucocorticoid receptor, were found in the liver from the offspring weight (Zakrzewska et al., 1999). One possible explanation is that Gcs of rats that were fed a protein restricted diet throughout pregnancy. down-regulate the anorexigenic peptide POMC, which would increase When compared to the control offspring, the diet restricted offspring food intake (Strack et al., 1995). In addition, Gcs suppress food intake, had decreased glucocorticoid receptor methylation, with a 200% by inducing the release of insulin, and therefore stimulating POMC increase in glucocorticoid receptor expression. The increase in neurones (Strack et al., 1995). Gcs can also directly increase POMC glucocorticoid receptor expression suggests increased influence of expression in the hypothalamus (Wardlaw et al., 1998). Thus, Gcs Gcs. Furthermore the changes in glucocorticoid receptor methylation may induce and suppress food intake in different experimental persisted in the offspring even though the dietary restriction had paradigms to control appetite regulation and indeed they can act by stopped suggesting that the methylation status of genes is potentially regulating other neuropeptides including neuropeptide Y (Fig. 1). permanent (Lillycrop et al., 2005). This study indicates that maternal A. Stevens et al. / European Journal of Pharmacology 660 (2011) 194–201 199

development of obesity. The reasons for this are multifactorial, not least of which is the increasing availability of food both at the time of birth and later in life. However the evidence for inter-generational programming is also a concern given the strength of the data from epidemiological analysis. Although this is based on studies of famines, many women in western society are dieting at the time of conception. Therefore it is important to understand how under-nutrition in the mother might impact on the foetus and the outcome for the off-spring. The concept that maternal insults cause epigenetic changes in promoter regions of genes in the foetus is evidenced by the methylation of the glucocorticoid receptor promoter in the foetal brain in models of maternal stress (Seckl and Holmes, 2007). Therefore we put forward the hypothesis that there could be epigenetic changes in the promoter region of hypothalamic neuropeptides in offspring from mothers who had been under-nourished around the time of conception. We found hypomethylation of a promoter region of POMC, which could result in changes persisting into adulthood and which could alter the way the gene responds to counteract increased food intake. This was a very specific effect in that we did not identify any epigenetic modifications in neuropeptide Y nor did we find any changes in methylation or acetylation in the POMC promoter region in the pituitary (Stevens et al., 2010). Interestingly while considering the glucocorticoid receptor, we were surprised to find epigenetic changes in its promoter consistent with up-regulation of the receptor. This occurred concomitantly with increases in glucocorticoid receptor gene expression even at the foetal stage of the offspring. The observation of hypomethylation occurring in both POMC and glucocorticoid receptor genes after maternal undernutrition in the Fig. 5. The effect of different periods of periconceptional undernutrition on hypothalamic sheep provides striking evidence for epigenetic changes which could GR, methylation and gene expression. The −60 to +30 group (UN −60 to +30) were act as a programming mechanism that predisposes hypothalamic underfed from 60 days before conception to 30 days after conception; the −60 to 0 group feeding centres to abnormal regulation later in life. Indeed mature (UN −60 to 0) were fed the same diet as the −60 to +30 group but were allowed to feed ad libitum from conception; the 0 to +30 group (UN 0 to +30) were fed the same diet for sheep that had undergone a similar phase of maternal undernutrition 30 days after conception. A) Methylation of GR promoter marker in hypothalami from were found to have increased body weight (Rumball et al., 2009). controls (n=7), −60 to +30 (n=9), −60 to 0 (n=8) and 0 to +30 (n=7) feeding Programming would be expected to decrease POMC gene expression regimens. A marker of the GR gene promoter region CpG islands was used to compare the if there is a resultant increase in obesity as predicted in the study by ratio of methylated to unmethylated DNA. One-way ANOVA with Tukey HSD post hoc test fi compared with control group *Pb0.05, **Pb0.001. B) Transcriptional expression of GR Delahaye et al. (2008). Although we did not nd a change in POMC gene in the hypothalami from controls (n=7), −60 to +30 (n=11), −2 to +30 (n=7) expression in line with previous work (Ikenasio-Thorpe et al., 2007), it and −60 to 0 (n=8) feeding regimens. One-way ANOVA with Tukey HSD post hoc test may be that this only occurs later in life, supporting the ‘thrifty compared with all other groups*Pb0.05. C) Immunohistochemistry for foetal hypotha- phenotype hypothesis’. The hypothesis suggests an adaptive response lamic GR was detected using the primary antibody M20 (Santa Cruz) and then targeted by the foetus to undernutrition allowing an increased chance of with a FITC labelled secondary antibody. Immunohistochemistry showing distribution of GR in the foetal sheep hypothalamus and the colocalisation of POMC and GR in the foetal postnatal survival (Kapoor, 2006; Seckl, 2004). However, if such sheep at 131 days of gestation. Magnification=10×. Scale bar=200 μm, 3V=3rd changes persist into adulthood where food is abundant, a defect in ventricle, ARC = arcuate nucleus. normal appetite regulation may subsequently lead to over-eating and obesity. This may explain numerous findings that prenatally under- nutritional insults can alter the methylation patterns of glucocorticoid nourished animals are hyperphagic when given hypercaloric or high fat receptor in the foetus. diet postnatally, compared to control animals (Vickers et al., 2000). We have shown a glucocorticoid receptor gene promoter region The epigenetic changes in the glucocorticoid receptor which result marker to be associated with foetal programming of the hypothala- in over-expression of the glucocorticoid receptor could be more mus in response to maternal undernutrition. This marker is situated relevant to regulation of the neuronal pathways in the hypothalamus, immediately 5′ of the glucocorticoid receptor exon 1 complex as Gcs are known to influence many of the pathways. It would be (Stevens et al., 2010). Our observations are supported by increasing necessary to predict that Gcs down-regulate POMC gene expression in evidence for the role of the glucocorticoid receptor gene exon 1 the hypothalamus. However this appears to be very dependent on the promoter regions in modulating both the activation and the experimental setting. repression of glucocorticoid receptor gene expression through In conclusion, there is sound evidence of a role for epigenetics in glucocorticoid response units (GRUs) (Geng et al., 2008) and the the programming of hypothalamic neuropeptide pathways leading to association of the NGFIA transcription factor with the maternal an increased propensity for obesity in the adult off-spring. More work programming of stress (Weaver et al., 2004). is required to understand the mechanisms and to identify approaches to reverse these changes so that the normal homeostatic balances can 5. Discussion regulate food intake.

There is considerable evidence of a role for hypothalamic POMC in the control of food intake and energy balance as seen with the effects of Acknowledgements loss of function mutations in POMC and in the melanocortin MC4 receptor (Pritchard et al., 2002; Pritchard and White, 2007). However Support has been gratefully received from the NIHR Manchester these mutations cannot account for the marked expansion in the Biomedical Research Centre, the Health Research Council of New 200 A. Stevens et al. / European Journal of Pharmacology 660 (2011) 194–201

Zealand, the National Research Centre for Growth and Development, Korner, J., Chua Jr., S.C., Williams, J.A., Leibel, R.L., Wardlaw, S.L., 1999. Regulation of hypothalamic proopiomelanocortin by leptin in lean and obese rats. Neuroendo- New Zealand, and the Canadian Institutes of Health Research. crinology 70, 377–383. Korner, J., Savontaus, E., Chua Jr., S.C., Leibel, R.L., Wardlaw, S.L., 2001. Leptin regulation of AGRP and NPY mRNA in the rat hypothalamus. J. Neuroendocrinol. 13, 959–966. References Krude, H., Gruters, A., 2000. Implications of proopiomelanocortin (POMC) mutations in humans: the POMC deficiency syndrome. Trends Endocrinol. Metab. 11, 15–22. Adam, C.L., Findlay, P.A., Chanet, A., Aitken, R.P., Milne, J.S., Wallace, J.M., 2008. Krude, H., Biebermann, H., Luck, W., Horn, R., Brabant, G., Gruters, A., 1998. Severe early- Expression of energy balance regulatory genes in the developing ovine fetal onset obesity, adrenal insufficiency and red hair pigmentation caused by POMC hypothalamus at midgestation and the influence of hyperglycemia. Am. J. Physiol. mutations in humans. Nat. Genet. 19, 155–157. Regul. Integr. Comp. Physiol. 294, R1895–R1900. Krude, H., Biebermann, H., Gruters, A., 2003. Mutations in the human proopiomelano- Benoit, S.C., Air, E.L., Coolen, L.M., Strauss, R., Jackman, A., Clegg, D.J., Seeley, R.J., Woods, cortin gene. Ann. N. Y. Acad. Sci. 994, 233–239. S.C., 2002. The catabolic action of insulin in the brain is mediated by melanocortins. Lillycrop, K.A., Phillips, E.S., Jackson, A.A., Hanson, M.A., Burdge, G.C., 2005. Dietary J. Neurosci. 22, 9048–9052. protein restriction of pregnant rats induces and folic acid supplementation Bertram, C.E., Hanson, M.A., 2002. Prenatal programming of postnatal endocrine prevents epigenetic modification of hepatic gene expression in the offspring. J. responses by glucocorticoids. Reproduction 124, 459–467. Nutr. 135, 1382–1386. Breen, T.L., Conwell, I.M., Wardlaw, S.L., 2005. Effects of fasting, leptin, and insulin on Markakis, E.A., 2002. Development of the neuroendocrine hypothalamus. Front. AGRP and POMC peptide release in the hypothalamus. Brain Res. 1032, 141–148. Neuroendocrinol. 23, 257–291. Breton, C., Lukaszewski, M.A., Risold, P.Y., Enache, M., Guillemot, J., Riviere, G., Delahaye, Martin-Gronert, M.S., Ozanne, S.E., 2005. Programming of appetite and type 2 diabetes. F., Lesage, J., Dutriez-Casteloot, I., Laborie, C., Vieau, D., 2009. Maternal prenatal Early Hum. Dev. 81, 981–988. undernutrition alters the response of POMC neurons to energy status variation in McGowan, P.O., Meaney, M.J., Szyf, M., 2008. Diet and the epigenetic (re)programming adult male rat offspring. Am. J. Physiol. Endocrinol. Metab. 296, E462–E472. of phenotypic differences in behavior. Brain Res. 1237, 12–24. Challis, B.G., Yeo, G.S., 2002. Past, present and future strategies to study the genetics of Mizoguchi, Y., Kajiume, T., Miyagawa, S., Okada, S., Nishi, Y., Kobayashi, M., 2007. Steroid- body weight regulation. Brief. Funct. Genomic. Proteomic. 1, 290–304. dependent ACTH-produced thymic carcinoid: regulation of POMC gene expression by Challis, B.G., Pritchard, L.E., Creemers, J.W., Delplanque, J., Keogh, J.M., Luan, J., Wareham, cortisol via methylation of its promoter region. Horm. Res. 67, 257–262. N.J., Yeo, G.S., Bhattacharyya, S., Froguel, P., White, A., Farooqi, I.S., O'Rahilly, S., 2002. Mizuno, T.M., Kleopoulos, S.P., Bergen, H.T., Roberts, J.L., Priest, C.A., Mobbs, C.V., 1998. A missense mutation disrupting a dibasic prohormone processing site in pro- Hypothalamic pro-opiomelanocortin mRNA is reduced by fasting and [corrected] in opiomelanocortin (POMC) increases susceptibility to early-onset obesity through a ob/ob and db/db mice, but is stimulated by leptin. Diabetes 47, 294–297. novel molecular mechanism. Hum. Mol. Genet. 11, 1997–2004. Muhlhausler, B.S., Adam, C.L., Marrocco, E.M., Findlay, P.A., Roberts, C.T., McFarlane, J.R., Challis, B.G., Coll, A.P., Yeo, G.S., Pinnock, S.B., Dickson, S.L., Thresher, R.R., Dixon, J., Kauter, K.G., McMillen, I.C., 2005. Impact of glucose infusion on the structural and Zahn, D., Rochford, J.J., White, A., Oliver, R.L., Millington, G., Aparicio, S.A., Colledge, functional characteristics of adipose tissue and on hypothalamic gene expression W.H., Russ, A.P., Carlton, M.B., O'Rahilly, S., 2004. Mice lacking pro-opiomelano- for appetite regulatory neuropeptides in the sheep fetus during late gestation. J. cortin are sensitive to high-fat feeding but respond normally to the acute anorectic Physiol. 565, 185–195. effects of peptide-YY(3–36). Proc. Natl. Acad. Sci. U. S. A. 101, 4695–4700. Newell-Price, J., 2003. Proopiomelanocortin gene expression and DNA methylation: Chen, H., Simar, D., Lambert, K., Mercier, J., Morris, M.J., 2008. Maternal and postnatal implications for Cushing's syndrome and beyond. Journal of Endocrinology 177, overnutrition differentially impact appetite regulators and fuel metabolism. 365–372. doi:10.1677/joe.0.1770365. Endocrinology 149, 5348–5356. Newell-Price, J., King, P., Clark, A.J., 2001. The CpG island promoter of the human Cone, R.D., 2005. Anatomy and regulation of the central melanocortin system. Nat. proopiomelanocortin gene is methylated in nonexpressing normal tissue and Neurosci. 8, 571–578. tumors and represses expression. ME 15, 338–348. Coupe, B., Amarger, V., Grit, I., Benani, A., Parnet, P., 2010. Nutritional programming Obici, S., Feng, Z., Tan, J., Liu, L., Karkanias, G., Rossetti, L., 2001. Central melanocortin affects hypothalamic organization and early response to leptin. Endocrinology 151, receptors regulate insulin action. J. Clin. Invest. 108, 1079–1085. 702–713. Padilla, S.L., Carmody, J.S., Zeltser, L.M., 2010. Pomc-expressing progenitors give rise to Cowley, M.A., Smart, J.L., Rubinstein, M., Cerdan, M.G., Diano, S., Horvath, T.L., Cone, R.D., antagonistic neuronal populations in hypothalamic feeding circuits. Nat. Med. 16, Low, M.J., 2001. Leptin activates anorexigenic POMC neurons through a neural 403–405. network in the arcuate nucleus. Nature 411, 480–484. Plagemann, A., Harder, T., Brunn, M., Harder, A., Roepke, K., Wittrock-Staar, M., Ziska, T., Cripps,R.L.M.S.M.-G.e.a., 2005. Fetal and perinatal programming of appetite. Clin.Sci 109. Schellong, K., Rodekamp, E., Melchior, K., Dudenhausen, J.W., 2009. Hypothalamic de Souza, F.S., Santangelo, A.M., Bumaschny, V., Avale, M.E., Smart, J.L., Low, M.J., proopiomelanocortin promoter methylation becomes altered by early overfeeding: Rubinstein, M., 2005. Identification of neuronal enhancers of the proopiomelano- an epigenetic model of obesity and the metabolic syndrome. J. Physiol. 587, cortin gene by transgenic mouse analysis and phylogenetic footprinting. Mol. Cell. 4963–4976. Biol. 25, 3076–3086. Plum, L., Belgardt, B.F., Bruning, J.C., 2006. Central insulin action in energy and glucose Delage, B., Dashwood, R.H., 2008. Dietary manipulation of histone structure and homeostasis. J. Clin. Invest. 116, 1761–1766. function. Annu. Rev. Nutr. 28 (347–66), 347–366. Poggioli, R., Vergoni, A.V., Bertolini, A., 1986. ACTH-(1–24) and alpha-MSH antagonize Delahaye, F., Breton, C., Risold, P.-Y., Enache, M., Dutriez-Casteloot, I., Laborie, C., Lesage, feeding behavior stimulated by kappa opiate agonists. Peptides 7, 843–848. J., Vieau, D., 2008. Maternal perinatal undernutrition drastically reduces postnatal Pritchard, L.E., White, A., 2007. Neuropeptide processing and its impact on leptin surge and affects the development of arcuate nucleus proopiomelanocortin melanocortin pathways. Endocrinology 148, 4201–4207. neurons in neonatal male rat pups. Endocrinology 149, 470–475. Pritchard, L.E., Turnbull, A.V., White, A., 2002. Pro-opiomelanocortin processing in the Forbes, S., Bui, S., Robinson, B.R., Hochgeschwender, U., Brennan, M.B., 2001. Integrated hypothalamus: impact on melanocortin signalling and obesity. JE 172, 411–421. control of appetite and fat metabolism by the leptin–proopiomelanocortin Pritchard, L.E., Oliver, R.L., McLoughlin, J.D., Birtles, S., Lawrence, C.B., Turnbull, A.V., pathway. Proc. Natl. Acad. Sci. U. S. A. 98, 4233–4237. White, A., 2003. Proopiomelanocortin-derived peptides in rat cerebrospinal fluid Ford, E.S., Giles, W.H., Dietz, W.H., 2002. Prevalence of the metabolic syndrome among and hypothalamic extracts: evidence that secretion is regulated with respect to US adults: findings from the third National Health and Nutrition Examination energy balance. Endocrinology 144, 760–766. Survey. JAMA. 287, 356–359. Pritchard, L.E., Armstrong, D., Davies, N., Oliver, R.L., Schmitz, C.A., Brennand, J.C., Geng, C.D., Schwartz, J.R., Vedeckis, W.V., 2008. A conserved molecular mechanism is Wilkinson, G.F., White, A., 2004. Agouti-related protein (83–132) is a competitive responsible for the auto-up-regulation of glucocorticoid receptor gene promoters. antagonist at the human melanocortin-4 receptor: no evidence for differential ME 22, 2624–2642. interactions with pro-opiomelanocortin-derived ligands. J. Endocrinol. 180, Graff, J., Mansuy, I.M., 2008. Epigenetic codes in cognition and behaviour. Behav. Brain 183–191. Res. 192, 70–87. Ravelli, G.P., Stein, Z.A., Susser, M.W., 1976. Obesity in young men after famine exposure Gyte, A., Pritchard, L.E., Jones, H.B., Brennand, J.C., White, A., 2007. Reduced expression in utero and early infancy. N. Engl. J. Med. 295, 349–353. of the KATP channel subunit, Kir6.2, is associated with decreased expression of Rhodes, P., Craigon, J., Gray, C., Rhind, S.M., Loughna, P.T., Gardner, D.S., 2009. Adult- neuropeptide Y and agouti-related protein in the hypothalami of Zucker diabetic onset obesity reveals prenatal programming of glucose-insulin sensitivity in male fatty rats. J. Neuroendocrinol. 19, 941–951. sheep nutrient restricted during late gestation. PLoS ONE 4, e7393. Ho, S.M., Tang, W.Y., 2007. Techniques used in studies of epigenome dysregulation due Rumball, C.W., Bloomfield, F.H., Oliver, M.H., Harding, J.E., 2009. Different periods of to aberrant DNA methylation: an emphasis on fetal-based adult diseases. Reprod. periconceptional undernutrition have different effects on growth, metabolic and Toxicol. 23, 267–282. endocrine status in fetal sheep. Pediatr. Res. 66, 605–613. Ibrahim, N., Bosch, M.A., Smart, J.L., Qiu, J., Rubinstein, M., Ronnekleiv, O.K., Low, M.J., Savontaus, E., Conwell, I.M., Wardlaw, S.L., 2002. Effects of adrenalectomy on AGRP, Kelly, M.J., 2003. Hypothalamic proopiomelanocortin neurons are glucose POMC, NPY and CART gene expression in the basal hypothalamus of fed and fasted responsive and express K(ATP) channels. Endocrinology 144, 1331–1340. rats. Brain Res. 958, 130–138. Ikenasio-Thorpe, B.A., Breier, B.H., Vickers, M.H., Fraser, M., 2007. Prenatal influences on Schwartz, M.W., Seeley, R.J., Woods, S.C., Weigle, D.S., Campfield, L.A., Burn, P., Baskin, D.G., susceptibility to diet-induced obesity are mediated by altered neuroendocrine gene 1997. Leptin increases hypothalamic pro-opiomelanocortin mRNA expression in the expression. J. Endocrinol. 193. rostral arcuate nucleus. Diabetes 46, 2119–2123. Jackson, R.S., Creemers, J.W., Ohagi, S., Raffin-Sanson, M.L., Sanders, L., Montague, C.T., Sebert, S.P., Hyatt, M.A., Chan, L.L., Patel, N., Bell, R.C., Keisler, D., Stephenson, T., Budge, Hutton, J.C., O'Rahilly, S., 1997. Obesity and impaired prohormone processing H., Symonds, M.E., Gardner, D.S., 2009. Maternal nutrient restriction between early associated with mutations in the human prohormone convertase 1 gene. Nat. and midgestation and its impact upon appetite regulation after juvenile obesity. Genet. 16, 303–306. Endocrinology 150, 634–641. Kapoor, S., 2006. Physical activity to prevent obesity in young children: dietary and Seckl, J.R., 2004. Prenatal glucocorticoids and long-term programming. Eur. J. behavioural modifications in managing childhood obesity. BMJ. 333, 1171–1172. Endocrinol. 151, U49–U62. A. Stevens et al. / European Journal of Pharmacology 660 (2011) 194–201 201

Seckl, J.R., Holmes, M.C., 2007. Mechanisms of disease: glucocorticoids, their placental Turner, J.D., Schote, A.B., Macedo, J.A., Pelascini, L.P., Muller, C.P., 2006. Tissue specific metabolism and fetal ‘programming’ of adult pathophysiology. Nat. Clin. Pract. glucocorticoid receptor expression, a role for alternative first exon usage? Biochem. Endocrinol. Metab. 3, 479–488. Pharmacol. 72, 1529–1537. Stevens, A., Begum, G., Cook, A., Connor, K., Rumball, C., Oliver, M., Challis, J., Bloomfield, Vergoni, A.V., Poggioli, R., Marrama, D., Bertolini, A., 1990. Inhibition of feeding by ACTH- F., White, A., 2010. Epigenetic changes in the hypothalamic proopiomelanocortin (1–24): behavioral and pharmacological aspects. Eur. J. Pharmacol. 179, 347–355. and glucocorticoid receptor genes in the ovine fetus after periconceptional Vickers, M.H., Breier, B.H., Cutfield, W.S., Hofman, P.L., Gluckman, P.D., 2000. Fetal undernutrition. Endocrinology 151, 3652–3664. origins of hyperphagia, obesity, and hypertension and postnatal amplification by Strack, A.M., Horsley, C.J., Sebastian, R.J., Akana, S.F., Dallman, M.F., 1995. Glucocorti- hypercaloric nutrition. Am. J. Physiol. Endocrinol. Metab. 279, E83–E87. coids and insulin: complex interaction on brown adipose tissue. Am. J. Physiol. 268, Wardlaw, S.L., McCarthy, K.C., Conwell, I.M., 1998. Glucocorticoid regulation of R1209–R1216. hypothalamic proopiomelanocortin. Neuroendocrinology 67, 51–57. Swart, I., Jahng, J.W., Overton, J.M., Houpt, T.A., 2002. Hypothalamic NPY, AGRP, and Weaver, I.C., Cervoni, N., Champagne, F.A., D'Alessio, A.C., Sharma, S., Seckl, J.R., Dymov, POMC mRNA responses to leptin and refeeding in mice. Am. J. Physiol. Regul. Integr. S., Szyf, M., Meaney, M.J., 2004. Epigenetic programming by maternal behavior. Nat. Comp. Physiol. 283, R1020–R1026. Neurosci. 7, 847–854. Symonds, M.E., Budge, H., 2009. Nutritional models of the developmental programming Yaswen, L., Diehl, N., Brennan, M.B., Hochgeschwender, U., 1999. Obesity in the mouse of adult health and disease. Proc. Nutr. Soc. 68, 173–178. model of pro-opiomelanocortin deficiency responds to peripheral melanocortin. Terroni, P.L., Anthony, F.W., Hanson, M.A., Cagampang, F.R., 2005. Expression of agouti- Nat. Med. 5, 1066–1070. related peptide, neuropeptide Y, pro-opiomelanocortin and the leptin receptor Ye, L., Li, X., Kong, X., Wang, W., Bi, Y., Hu, L., Cui, B., Li, X., Ning, G., 2005. Hypomethylation isoforms in fetal mouse brain from pregnant dams on a protein-restricted diet. in the promoter region of POMC gene correlates with ectopic overexpression in thymic Brain Res. Mol. Brain Res. 140, 111–115. carcinoids. J. Endocrinol. 185, 337–343. Thornton, J.E., Cheung, C.C., Clifton, D.K., Steiner, R.A., 1997. Regulation of hypothalamic Zakrzewska, K.E., Sainsbury, A., Cusin, I., Rouru, J., Jeanrenaud, B., Rohner-Jeanrenaud, F., proopiomelanocortin mRNA by leptin in ob/ob mice. Endocrinology 138, 1999. Selective dependence of intracerebroventricular neuropeptide Y-elicited 5063–5066. effects on central glucocorticoids. Endocrinology 140, 3183–3187. Todd, S.E., Oliver, M.H., Jaquiery, A.L., Bloomfield, F.H., Harding, J.E., 2009. Periconcep- tional undernutrition of ewes impairs glucose tolerance in their adult offspring. Pediatr. Res. 65, 409–413.