The effect of metabolic and inflammatory parameters on pregnancy success in terms of implantation, metabolic adaptation to pregnancy and fetal programming is yet to be fully understood. This thesis explores the activity of metabolic and inflammatory pathways in pregnancy, highlighting their importance throughout gestation. In a cell culture study, a model of in vivo blastocyst-uterine adhesion to study the effect of insulin during uterine implantation was explored. JAR spheroid-RL95-2 monolayer adhesion reached 98% by 24 hours in the absence of insulin. A low dose (0.03nM) of added insulin concentrations resulted in 26% adhesion, or 74% inhibition; a high level (0.24nM) inhibited the JAR spheroid-RL95-2 monolayer adhesion by 9%. Therefore insulin did not have a dose-dependent on JAR spheroid-RL95-2 monolayer adhesion in the cell culture model of implantation. Polymerase chain reaction (PCR) studies revealed laminin α1 RNA detection on JAR cells only, CD44 on RL95-2 cells only, no trophinin on both cell types, FBLN-1 and -2 on JAR and FBLN-1 on RL95-2 cells only and an insulin receptor in both cell types. Western blot and immunohistochemistry (IHC) studies showed laminin α1 detection and stains on the JAR cell extracellular matrix. In a prospective human study, the metabolites of lipid and carbohydrate metabolism and inflammatory mediators very early (between day 0 and day 45) in gestation and their link to successful pregnancy in women undergoing natural cycle frozen embryo transfer (FET) in assisted conception, was investigated. Plasma triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), glucose, c-reactive protein (CRP) and non-esterified fatty acid (NEFA) were measured on routine biochemistry; insulin, interleukin (IL)-6, plasminogen activated inhibitor (PAI)-1 and PAI-2 on ELISA; IL-8 (CXCL8), CCL2, CCL3, CCL4 and CCL11 on BioPlex; and human chronic gonadotrophin (hCG) on an Immulite system. For all 196 FET cycles, participants' demographics and plasma parameters of pregnant (n=36) and non-pregnant (n=106) women were explored. Neither obesity, the plasma parameters nor insulin resistance were predictive of successful pregnancy, but ICSI (predominately associated with male factor infertility) was. Overall, the hCG, insulin, rebound TG and HDL-C (except TC), homeostasis model assessment (HOMA), CRP and PAI-2 levels were higher, whereas CXCL8, CCL2, CCL11 and PAI-1 were significantly lower by day 45. Baseline obesity related to positive changes in plasma insulin, HDL-C and HOMA and negative changes in CXCL8, CCL3 and CCL4. In a cross-sectional study in late pregnancy, offspring's reflection of parameters in women with preeclampsia (PE) (n=29) and intrauterine growth restriction (IUGR) (n=14), compared to BMI-matched healthy groups (n=87) and (n=42), respectively, was explored. Fetal cord was found to be hyperlipidaemic, normoglycaemic and had reduced inflammatory response, while mothers who suffered PE had altered plasma TG, TC, NEFA, glucose, leptin and IL-10 compared to controls. IUGR babies were dyslipidaemic. The role of cholesterol transporters was assessed in PE (n=20) and IUGR (n=9) BMI-matched controls (n=20 and n=9) respectively. Among fifteen steroidogenic acute regulatory protein (STAR)-related lipid transfer domains, only STARD6 and STARD15 were not detected in the placenta via PCR. IHC studies were also explored on the placentae. The real-time PCR (RT-PCR) of messenger RNA of low-density lipoprotein receptor (LDLR), STARD3 and ATP-binding cassette A1 (ABCA1) without protein were higher in PE compared to controls. LDLR, STARD3 and ABCA1 localisation and detection were consistent to placental lipid (cholesterol) transport systems. In summary, all this led to the conclusion of the importance of metabolic and inflammatory pathways in all stages of pregnancy in leading to pregnancy success; these pathways may influence implantation, adaptation to pregnancy and, potentially, fetal programming of offspring.

Maternal obesity has been at the forefront of pregnancy-related research in recent times. The impact of this chronic health condition has been highlighted in reports on maternal mortality (CEMACH, 2007, CEMACH, 2011), where 30% of mothers who died from pregnancy related causes were obese (CEMACH, 2011). The importance of maternal obesity and how it affects maternal adaptation to pregnancy is well documented with obese women exhibiting low grade inflammation, greater coagulability and poorer improvement in vascular function during pregnancy compared to lean women (Stewart et al., 2007a). These findings suggest that obese women display similar characteristics to the non-pregnant adult metabolic syndrome and these attributes may be important in explaining why obese pregnancies have higher rates of obstetric complications including gestational diabetes (GDM) and pre-eclampsia (PET). In non-pregnant adult obesity it has been found that central or truncal adiposity is associated with increased NEFA (non-esterified fatty acids) turnover and ectopic fat (especially liver) deposition. It has been suggested that obese pregnant women may also preferentially gain fat in central depots and this may be the mechanism by which poor vascular improvement and inflammation are initiated. The aims of this thesis were to assess subcutaneous fat accumulation and distribution throughout pregnancy in both lean and OW/OB women. Furthermore this thesis aimed to acquire a better understanding of the impact of anatomical fat deposition on metabolic and vascular function during pregnancy. A final aim was to assess vascular function and evidence of lipotoxicity during pregnancy and test whether the site of fat accumulation and distribution was associated with gestational improvement of vascular function. A longitudinal study was performed and anthropometric data was collected from 26 lean and 16 OW/OB women at three antenatal time points (15, 25 and 35 weeks’ gestation) during pregnancy. Direct measurements of energy metabolism (basal metabolic rate, substrate utilisation, physical activity and diet) were also collected to assess the impact of energy metabolism on fat accumulation and distribution. A comprehensive panel of plasma markers of carbohydrate and lipid metabolism (fasting glucose, fasting insulin, total cholesterol [TC], total triglyceride [TG], high density lipoprotein [HDL] and NEFA) and inflammatory (C-reactive protein [CRP], interleukin-6 [IL6] and tumour necrosis factor alpha [TNF]) were quantitated at each study appointment. Endothelial function was measured using laser Doppler imaging (LDI). Measurement of plasma and urinary biomarkers of endothelial function and lipotoxicity including soluble intracellular adhesion molecule 1 (sICAM-1), soluble vascular adhesion molecule-1 (sVCAM-1), oxidised low density lipoprotein (oxLDL), plasma superoxide and urinary isoprostanes were undertaken. Lean and OW/OB women gained similar amounts of total body weight and fat mass during pregnancy. Only in lean women was there an anatomical preference for site of fat storage and this was in the upper peripheral subcutaneous depots. In healthy OW/OB pregnancy no such anatomical preference of fat deposition was found. The study of energy metabolism found that OW/OB women had higher basal metabolic rate and higher fat oxidation than lean women, whilst lean women had higher rates of carbohydrate oxidation and physical activity than OW/OB women. In the lean and OW/OB groups dietary macronutrient intakes were similar. Overall the parameters of energy metabolism were not associated with overall fat mass accumulation or distribution. During pregnancy, OW/OB women were more insulin resistant and pro-inflammatory (CRP and TNFα) than lean women and lean women had higher concentrations of plasma HDL. Interestingly the lean group had higher plasma concentrations of IL6 which may be a result of higher rates of vascular remodelling and may reflect a physiological rather than pathological process. In both lean and OW/OB pregnancies the gestational increase in subcutaneous adipose depots was not associated with the gestational changes in markers of carbohydrate, lipid or inflammatory profiles. Both lean and OW/OB women exhibited similar gestational improvement in endothelial microvascular function. During pregnancy both groups showed an increase in markers of lipotoxicity but levels were not associated with vascular function. Changes in anatomical subcutaneous fat distribution were also not associated with the changes in vascular function during pregnancy. In conclusion, in pregnancy, only lean women exhibit an anatomical site-specific fat accumulation. Although the OW/OB group displayed some aspects of the metabolic syndrome in general the OW/OB women studied here adapted to pregnancy in a similar way to lean women in terms of vascular function and levels of lipotoxicity. However, visceral adiposity was not assessed and OW/OB women with larger visceral adipose stores may exhibit a more lipotoxic phenotype and more pathological adaptation to pregnancy that may make them susceptible to metabolic complications of pregnancy. This study highlights the heterogeneity of maternal obesity and suggests that further studies into ‘metabolically healthy’ and ‘metabolically unhealthy’ lean and OW/OB women is warranted.

Genes that are imprinted are subject to a developmentally determined epigenetic marking, which restricts expression to a single allele, dependant on the parent of origin. Selection of imprinted genes for monoallelic expression indicates their function is highly dosage sensitive. Altered dosage of imprinted genes has been linked to a number of neurological conditions, including psychosis. Cdkn1c is an example of an imprinted gene whose expression is sensitive to the in utero environment. Considerable development of the nervous system takes place in utero and suboptimal pregnancies have been linked to the occurrence of psychiatric and other behavioural disorders in adults. One mechanism through which the maternal environment may impact foetal development is by altering the epigenetic regulation of vulnerable genes. A prenatal low protein or high fat diet resulted in alterations in a subset of imprinted gene in the brains of the offspring at E18.5. This was accompanied by sexually dimorphic changes in the dopaminergic system. Previously published findings reporting sensitivity of Cdkn1c to a prenatal low protein diet were replicated with a 1.8 fold increase in neural Cdkn1c expression observed. This was shown to be due to a change in the parental contribution to expression levels of this gene. Modelling the specific alteration of an increase in Cdkn1c genetically (Cdkn1cBACx1 line) revealed anhedonia, but with an increased motivational drive, towards a palatable solution, with corresponding changes in the reward system responsivity and chemistry in the adult brain. Additionally presence of a Cdkn1cBACx1 animal in a group destabilised the social hierarchy, negatively effecting fitness of all group members. An adverse inutero environment increases Cdkn1c levels to those reminiscent of the genetic ‘loss of imprinting’ model. Such alteration in expression of Cdkn1c has significant consequences for adult neurochemistry, reward processing and the social environment and fitness of the group. This work suggests a potentially crucial role, of at least Cdkn1c, and perhaps imprinted genes more generally, in mediating the negative consequences of an adverse in utero environment.

Diabetes is the most frequently encountered endocrine disorder in pregnancy and is associated with adverse outcomes. Despite the urgent need for interventions to improve the outcomes for pregnancies complicated with diabetes, and the consistent recognition of preconception care as an effective intervention, there has been lack of systematically produced evidence to support it. My first publication (Preconception Care for Diabetic Women for Improving Maternal and Fetal Outcomes: a Systematic Review and Meta-analysis) was the first systematically produced high level evidence addressing the effectiveness and the safety of all aspects of preconception care. This publication had high impact on practice and research evident by the incorporation of its findings in clinical guidelines and the number of times it was cited in the literature. My second publication (Pre-pregnancy care for women with pre-gestational diabetes mellitus: a systematic review and meta-analysis) was designed for deeper analysis of the safety of preconception care. The third and the fourth publications addressed the prevalence of pre-gestational and gestational diabetes and the rate of complications associated with diabetes in pregnancy in Saudi Arabia and contributed to the quantification of diabetes in pregnancy as a public health problem in the country. These two publications provided important information, considering that there was paucity of publications about diabetes in pregnancy in Saudi Arabia for more than a decade, and they gave the needed evidence to revise the hospital policy for screening and management of diabetes in pregnancy as well as the implementation of preconception care for women with pre-existing diabetes. My fifth publication investigated an important clinical intervention for pregnant women with diabetes which is induction of labour. Similar to the second and third publication there was paucity of information about the indications and the determinants of successful induction of labour in Saudi Arabia. This publication was the first to address this important intervention in the practice of obstetrics in general and in the specific management of women with diabetes. Thus my work in "diabetes in pregnancy as a clinical and public health problem" provided an important evaluation of interventions at the clinical and public health levels and important information for the management of diabetic pregnant women in Saudi Arabia and across the world.

Hot flushes are the most commonly reported symptom in postmenopausal women, occurring in approximately 73% of women and causing significant morbidity in 25%, affecting social life and even the ability to work. With improved healthcare and increased life expectancy (death rates decreased by 19% in the last 10 years), women spend a considerable proportion of their lives (30 years on average) in the menopause. At present 36% of the women in the UK are over 50 years of age. If left untreated, hot flushes resolve within one year, or less, in the majority of postmenopausal women. A third will report symptoms that last up to 5 years after natural menopause, and in 20% hot flushes persist for up to 15 years. This equates to as many as 1.5 million women in the UK. Despite this the mechanism of flushing is still poorly understood. A hot flush resembles a heat dissipation response, in that both are characterised by sweating and peripheral vasodilation. It follows that the underlying mechanism may involve some dysfunction in thermoregulation, which in humans is controlled by the medial preoptic area of the hypothalamus (MPOA), and effectors include cutaneous vessels for vasodilation and vasoconstriction. Therefore a dysfunction in thermoregulation may lie within the control centre (MPOA), its messengers (adrenergic neurones controlling vasoconstriction and cholinergic neurones controlling vasodilation) or the effectors (cutaneous vessels). Studies by Freedman et al, using an ultrasensitive temperature probe, suggest that hot flushes are triggered by small elevations in core body temperature (Tc) acting within a narrowed thermoneutral zone, mainly due to a lowering of the sweating threshold, in symptomatic postmenopausal women. However, the trigger remains unknown. Oestrogen is likely involved as these changes occur at times of relative oestrogen withdrawal; however, there is little correlation between hot flushes and circulating oestrogen levels. This suggests that other mechanisms are involved. Noradrenaline is thought to be the primary neurotransmitter responsible for lowering the thermoregulatory set point and triggering hot flushes. Animal studies have shown that intrahypothalamic injection of noradrenaline acts to narrow the thermoregulatory zone and hot flushes can be provoked in symptomatic postmenopausal women with the α2-adrenergic antagonist yohimbine, and ameliorated with clonidine, an α-adrenergic agonist. Furthermore, clonidine has been shown to widen the thermoregulatory zone in humans. Serotonin or 5-hydroxytriptamine (5-HT) is involved in many bodily functions including mood, anxiety, sleep, sexual behaviour and eating, and is thought to play a key role in thermoregulation. Oestrogen withdrawal is associated with decreased blood serotonin levels, which is returned to normal with oestrogen therapy. Furthermore, selective serotonin reuptake inhibitors (SSRI), designed to increase the available serotonin at the serotonergic synapse, have been shown in placebo-controlled trials to be effective in reducing the number and severity of hot flushes. It has also been shown that flushing women have a diminished vasoconstrictor response to cold and that they have increased blood flow to the forearm and hand during a flushing episode. Alterations in skin blood flow during a flushing attack have also been demonstrated in castrate men, and, as with women, improvements in symptoms are seen with hormone replacement. The aim of this thesis was to better understand the mechanism of flushing in postmenopausal women and hypogonadal men by assessing the role of cutaneous vessels. I measured cutaneous microvascular perfusion, using LASER Doppler imaging with iontophoresis, in postmenopausal women who flush and compared it with postmenopausal women with no flushing, and found that perfusion responses to vasoactive agents were increased in women with flushing. Paradoxically, these women with apparently ‘better’ endothelial function had evidence of serum cardiovascular risk factors. In a double-blind longitudinal cross over study, the role of the alpha-adrenergic system in the pathophysiology of flushing was investigated, by treating women with clonidine and placebo. There was an increase in perfusion responses with both clonidine and placebo. Clonidine was not shown to be superior to placebo in reducing the number and severity of flushes. The role of serotonin, both peripherally and centrally was studied by treating postmenopausal women, who experienced severe flushing, with venlafaxine (a serotonin and noradrenaline reuptake inhibitor that acts as a selective serotonin reuptake inhibitor at low doses). Flushing symptoms, as assessed by hot flush diaries and Greene climacteric scale (GCS) scores, were reduced, as were skin blood flow perfusion responses. Central serotonin transporters (SERT) were assessed in vivo using SPECT (single photon computed tomography) imaging and a radioligand, [123I] -beta-carbomethoxy-3-β-(4 iodophenyl)tropane ([123I] beta-CIT), with a high affinity for serotonin transporters. [123I] beta-CIT binding was significantly reduced, and this was associated with a significant reduction in BDI scores; in a group of non-depressed women. Adiposity is associated, both, with an increased risk of postmenopausal flushing and impaired endothelial function, but in this study, there were no differences demonstrated between obese and lean participants at baseline, despite significant differences in serum markers of endothelial dysfunction. Oestrogen receptors are also present on endothelial cells and as the most commonly used, and most effective, treatment for vasomotor symptoms, cutaneous microvascular perfusion was assessed following 8 weeks of HRT, and demonstrated an increase in both endothelium dependent and independent vasodilation. Hot flushes are also common in men on luteinising hormone releasing hormone (LHRH) agonists for prostate cancer therapy. Perfusion responses in these men were assessed prior to commencement of therapy (baseline), and after 8 and then 24 weeks of therapy. No differences were detected at baseline, between those who developed flushing as a result of treatment, and those who did not. At 8 weeks, those with flushing demonstrated increased skin blood flow compared to those without flushing. At 24 weeks, 2 gentlemen with flushing had kept diaries and these demonstrated an improvement in flushing, but no alteration in perfusion responses. ACh- and SNP-stimulated vasodilation was, however, reduced when compared to healthy controls. In this thesis, the data appear to support a role of skin blood flow in the mechanism of hot flushing in both postmenopausal women and hypogonadal men, which may be controlled or altered via neurotransmitters at a local level. The placebo response was significant, but alterations in skin blood flow do not appear to have mediated the adiposity effect.

Communication, anxiety and satisfaction during pregnancy was examined. Ley (1977) presents a cognitive model to explain satisfaction and its links with understanding. Janis (1958, 1971) notes a curvilinear relationship between anxiety and post-operative coping and postulates that information, anxiety and cognitive preparation are the variables accounting for this. Kumar and Robson note that obstetric anxiety is related to concerns for maternal and infant well being rather than irrational anxiety. This study examined the experience of women, stressors, communication satisfaction, knowledge and information and looked at the extent to which these three theories could interrelate to provide a fuller explanation of the psychological experience of women. Five studies were undertaken. Initially a pilot study revealed many negative statements about communication when transcripts were analysed. Communication factors and anxiety laden instances were correlated. The next study was set up to examine knowledge levels as Ley predicts that these, together with misunderstandings could contribute to dissatisfaction. Desire for knowledge was high. Knowledge varied according to social class but not parity. Doctors felt parity would be a factor. Women had difficulty approaching their doctor for information yet still desired doctors as their primary information source. Doctors delegated much information imparting to classes. Study three examined anxiety, communications and satisfaction in labour with pain management (a noted stressor in study 1). Patients receiving Pethidine were dissatisfied. Their pain experience did not differ markedly,but their psychological preparation did. In study 4 anxiety and outcome was monitored, together with information gathering strategies. Linear, rather than curvilinear relationships were found (unlike those predicted by Janis). The course of anxiety was a useful measure and the impact of anxiety on caregivers in the cycle of communications and recovery were explored. The final study looked at the impact of intervention on anxiety and satisfaction in ante-natal care. Women were randomly allocated to groups receiving no intervention, information and information plus feedback. The latter group had significantly lower post-consultation anxiety and higher satisfaction than the other two. The role of knowledge and accuracy in relation to satisfaction was explored. Kumar and Robsons propositions about anxiety were supported in these studies. Ley's cognitive model contributed much to the understanding but limitations in this model are explored, especially in relation to process and interaction factors and the routes to understanding. Janis' curvilinear relationship was not upheld, but his theoretical explanations involving the use of information and worry needs further testing.

It is important therefore that maternal smoking is accurately recorded on screening request forms and in this study, the accuracy of self reported smoking status was assessed by analysis of cotinine in serum. Results showed that the percentage of self-reported smokers (24.1%) at booking was significantly lower than the cotinine-validated estimate of 30.1%. Also, smoking was associated with low birth weight, delivery prior to 39 weeks, increased AFP level (3.1%) and reduced hCG level (28.7%) in the second trimester. An increasing AFP level (but not hCG level) was associated with lower birth weight and delivery prior to 39 weeks in both smokers and non smokers but the effect was most marked in smokers. The difference in birth weight between the highest and the lowest AFP category for non-smokers was 448.3g and for smokers was 619.2g, suggesting that smoking exacerbates the effect of an elevated AFP on birth weight. Overall the difference in birth weight between the lowest AFP category in non smokers and the highest AFP category in smokers was 931.6g. Summary In summary, this study has shown that a cross-trimester contingent screening protocol with repeat measures has the potential to meet the UK NSC aspirational standard of 90% detection of Down’s syndrome pregnancies with a screen positive rate of less than 2%. Around 90% of women would complete screening in the first trimester without the need for a second stage sequential test. Correcting for factors such as maternal smoking habits, ethnicity and ART would further improve screening performance. Also it has been shown that where ultrasound resources are scarce, within-trimester and across-trimester protocols can reduce the need for NT measurement in all women and still deliver excellent screening performance although this falls short of the higher performance standard. The potential of these new screening protocols now need to be tested in prospective multicentre trials to confirm their performance in prospective practice.

Maternal metabolism undergoes dramatic changes in pregnancy in order to sustain and nourish the developing fetus. During healthy pregnancy the mother goes from an anabolic state in early pregnancy to a state of catabolism in late pregnancy with increased lipolysis together with a significant reduction in insulin sensitivity. Pre-eclampsia (PE) characterised by hypertension and proteinuria is a major cause of maternal and perinatal morbidity. There is acute ‘atherosis’ in PE placenta, and lipid accumulation within glomerular cells and liver. PE women have an early, excessive triglyceride and free fatty acid (FFA) rise and greater cardiovascular disease (CVD) risk in later life. The cause of these lipid abnormalities in PE is unknown but disordered adipocyte function including exaggerated lipolysis and aberrant release of adipokines (such as IL-6 and TNF alpha) is a major candidate pathway. Elevations in FFAs, and pro-inflammatory adipokines could underpin the oxidative stress, endothelial dysfunction, inflammation, and insulin resistance - characteristic features of PE. The aims of this thesis were to acquire a better understanding of lipid metabolism and function in normal pregnancy, to determine if adipocyte function was altered in PE and, if so, to establish mechanisms. In addition I planned to corroborate epidemiological evidence of increased future CVD risk and to establish which risk factors accounted for this increased risk. I collected subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) biopsies in non-labouring pregnant healthy (n=31) and PE (n=14) women who underwent caesarean section. Maternal blood was collected prior to delivery and phenotyping of the mother was performed including plasma assay for cholesterol, triglyceride, HDL-cholesterol, IL-6, TNF-α, leptin, adiponectin, high sensitivity CRP, glucose and insulin concentrations. Maternal BMI at booking, standardised blood pressure measurements and birth weight centile were also recorded. I determined ex vivo lipolytic activity (basal, isoprotenerol stimulated and insulin suppression of lipolysis) and adipokine production in response to lipopolysaccharide (LPS) stimulation from these biopsies. The gene expression of relevant target genes and macrophage densities in each adipose depot by immunocytochemistry (ICC) was also performed. In addition I performed carotid ultrasound assessment of women with a previous history of PE (n=31) and matched controls (n=29). Ethical approval was obtained from Glasgow Royal Infirmary LREC and all patients gave their informed consent. I found that in normal pregnancy, adipocyte lipolytic function is independent of maternal BMI. Adipocyte lipolytic function of SAT and VAT are also independent of each other. Adipose tissue is very metabolically flexible and the rate of whole body lipolysis is still insulin sensitive in late gestation. VAT is more closely related to markers of maternal insulin resistance (IR) and is more sensitive to catecholamine stimulation and less sensitive to insulin suppression of lipolysis than SAT, the basis of the “portal paradigm”. Increasing BMI is associated with an increase in VAT cell size, with increased lipolysis and an increase in pro-inflammatory adipokines, a potential mechanism through which increasing obesity could predispose to metabolic complications of pregnancy. In contrast SAT cell size is not closely related to BMI and this may reflect the adaptation of this depot to increasing fat mass through both hypertrophy and hyperplasia, a metabolically advantageous response. TNF alpha is an important correlate of basal lipolysis in SAT. In PE there is decreased insulin sensitivity of both SAT and VAT compared to controls as calculated by the fat cell insulin sensitivity index (or responsiveness to insulin once the tissue is stimulated by isoproterenol). This would potentially make a significant impact on total circulating FFA as almost 60% of circulating FFA are from these adipose depots. The rise in FFA in PE occurs early in pregnancy and contributes significantly to IR. Therefore the IR of adipose tissue could lead to a vicious cycle of increased lipolysis, increased FFA and further exacerbation of IR. In contrast to controls, SAT cell size is intimately related to BMI suggesting that adaptation to increasing fat mass is mainly through adipocyte hypertrophy which could lead to increased endoplasmic reticulum stress, increased IR and increased release of inflammatory adipokines. I have shown that SAT cell size does relate to adipokine release in PE, with increased release of leptin, CRP and PAI-1 and paradoxical increase in the anti-inflammatory IL-10. I had hypothesised that in addition to an inherent defect in adipocyte function there was an additional factor present in maternal serum of women with PE released from the placenta which excessively stimulated lipolysis. I failed to demonstrate any effect of maternal serum on adipocyte lipolysis in either controls or PE. I also found that after stimulation with LPS, there was increased release of TNF alpha and IL-6 in VAT in PE but not in controls, with higher gene expression of these adipokines. TNF alpha release also correlated negatively with the fat cell insulin sensitivity index (FCISI) of VAT implicating a paracrine effect in this tissue. I also demonstrated an increase in gene expression of cfms (activated macrophages) relative to control gene, and increased density of cfms+ macrophages/adipocytes in the VAT of PE women implicating activated adipose tissue macrophages as a potential source of the increased release of inflammatory adipokines. Lastly I attempted to corroborate epidemiological evidence for the increase future risk of CVD women with a history of PE by assessing two surrogate markers for atherosclerosis - carotid IMT and carotid plaque scores. Both were found to be increased, with plaque scores significantly so. Classic risk factors such as age, lipids, BP and smoking did not attenuate this effect and BMI only marginally attenuated it, therefore only partially explaining this increased risk. In summary the data presented in this thesis provides further evidence that PE is a “metabolic syndrome of pregnancy” with disordered adipocyte function and metabolism, with an increased future risk of CVD in later life. Further studies on adipose accumulation, function and composition in normal and complicated human pregnancy are warranted.

Approximately one in ten pregnant women will have their blood pressure recorded above normal at some point during their pregnancy. Pre-eclampsia, the most common hypertensive disorder of pregnancy, affects around 5% of all first time mothers, and is an important cause of foetal and maternal morbidity and mortality worldwide. Efforts to diagnose the condition have been hampered by inability to predict which women are likely to be affected. Multiple pathways are known to be involved in its pathogenesis, and several screening tests have been suggested for its early prediction. None, however, have been sensitive or specific enough to have come into routine medical practice. The work contained in this thesis describes a study which was designed to detect biochemical and clinical markers that could improve ability to predict pre-eclampsia. Over 3900 women were recruited in early pregnancy at four maternity clinics across the West of Scotland; baseline characteristics and information on past medical and obstetric history were obtained. Women were followed up throughout their pregnancy, and information on deliveries obtained from hospital databases. One-hundred and eighty of these women, who had multiple risk factors for pre-eclampsia, attended for further sampling and vascular assessment at gestational weeks 16 and 28. The primary aim of the overall study was to examine whether a proteomic strategy could be used to identify patterns of peptides in urine that detect pre-eclampsia in the first and second trimesters. Using samples from healthy pregnant and non-pregnant women I was able to describe the normal human urinary proteome in pregnancy. By comparing these pregnancy-associated peptides between women who went on to develop pre-eclampsia and matched controls, I was able to identify a pattern of peptides, characterised by collagen fragments, fibrinogen and uromodulin that accurately predicted pre-eclampsia at week 28. No such markers were identified in the first trimester samples. A further aim of the overall study was to identify early pregnancy plasma markers that could help to identify women destined to develop pre-eclampsia. By examining samples from early pregnancy I was able to demonstrate that the angiogenic markers soluble endoglin and placental growth factor are already altered at week 12-16 in women who go on to develop pre-eclampsia. Using a multi-marker approach, I also showed that E-Selectin, an adhesion molecule expressed on endothelial cells which controls interaction between circulating leukocytes and the endothelium, is higher at week 12-16 in women who go on to develop pre-eclampsia. Experiments using samples from later pregnancy, alternative analysis techniques and samples from an independent study population all helped to confirm these novel findings. Endothelial dysfunction is known to play a key role in the development of pre-eclampsia, contributing to the hypertension, proteinuria and oedema seen in affected women. In the risk factor cohort I used vascular function studies to examine whether they supplied additional information to aid in risk stratification. Peripheral arterial tonometry, a novel non-invasive tool for the assessment of microcirculatory endothelial function, was examined in 180 women at gestational weeks 16 and 28. Reactive hyperaemia index (RHI), a measure of endothelial dysfunction calculated from vascular response to arm blood-flow occlusion, did not correlate with maternal factors such as age, BMI and blood pressure. Further, RHI did not help to identify which women would go on to develop pre-eclampsia, when examined at either week 16 or 28. I found that PAT score was negatively correlated with baseline digital pulse amplitude, suggesting that in later pregnancy, when women are more vasodilated, PAT and other techniques which rely on flow-mediated dilatation are less likely to be reliable. I used pulse wave analysis, a well-established method for measuring arterial stiffness and central pressures, to determine whether it supplied additional information about pre-eclampsia risk. This technique has been previously reported to predict pre-eclampsia in early pregnancy. In this cohort of high risk women, no difference was seen at either week 16 or 28 between those who would go on to develop pre-eclampsia and those who would have normotensive pregnancies. Although blood pressure and proteinuria return to normal after pre-eclampsia, evidence has emerged the condition has long-lasting implications; women with a history of pre-eclampsia have an increased risk of cardiovascular disease later in life, suffering stroke or myocardial infarction more frequently than women who had a healthy pregnancy. Conventional risk factors are thought to contribute, but do not fully explain this increased risk. I carried out further vascular function studies in women after pre-eclamptic pregnancy, to examine whether they had ongoing detectable endothelial dysfunction and arterial stiffness. At 6-9 months post-natally, affected women had lower baseline digital pulse amplitude but no other evidence of persistent vascular dysfunction. Taken together, these data provide information about a number of markers that may improve understanding of the pathophysiological mechanisms underlying pre-eclampsia. As well as potentially improving the early prediction of disease, this work represents a highly topical area for further studies. While vascular function analysis does not appear to provide additional information on top of risk factors, these studies also provide useful information on vascular physiology in high-risk pregnancies.

Introduction: Preterm birth is the leading cause of perinatal morbidity and mortality and rates are rising. The UK now has the highest rate of premature birth in Europe with 5.3% of overall births in Scotland occurring spontaneously before 37 weeks gestation (1, 2) .Preterm babies have higher rates of perinatal mortality and morbidity and those that survive are at risk of multiple conditions including respiratory distress syndrome, central nervous system abnormalities, necrotising enterocolitis and sepsis. The mechanisms of preterm birth are poorly understood. Preterm birth can be spontaneous or induced and spontaneous preterm labour has multiple aetiologies. Current evidence suggests that prolonged treatment with progesterone and 17 α-hydroxyprogesterone caproate (17OHPC) may reduce the incidence of premature delivery in high risk patients with a history of spontaneous preterm birth (3) or with a short cervix. However, progesterone is not uniformly effective in preventing preterm labour and at present its principal mode of action on myometrium is unknown. I aimed to determine some of the specific mechanisms of action of progesterone. Aims: I hypothesised that progesterone has a direct inhibitory effect on spontaneous myometrial contractility, induces increased sensitivity to tocolytic agents and decreases sensitivity to contractile agonists such as oxytocin. I also hypothesised that progesterone has inhibitory effects on endogenous uterine stimulants, stimulatory effects on endogenous uterine relaxants, induces upregulation of endogenous receptors that inhibit uterine contractions and inhibits contraction associated proteins both in vitro and in vivo. Methods: I recruited women already enrolled in the STOPPIT (a double blind randomised placebo controlled study of progesterone for the prevention of preterm birth in twins) who were given vaginal progesterone, or placebo and who were scheduled for caesarean section. I also recruited women with healthy twin or singleton pregnancies undergoing elective caesarean section. Myometrial biopsies were obtained from the upper border of the lower uterine segment incision during caesarean section. Samples were divided and used for contractility measurements, or subsequent mRNA, protein and immunohistochemical analysis. Myometrial strips were cut and suspended under resting tension within organ baths. Concentration-response curves were carried out in response to oxytocin, levcromakalim, nifedipine and ritodrine to ascertain any reduction in effect by progesterone on oxytocics or enhancement of tocolytic effects. I also carried out concentration-response curves to progesterone alone and in the presence of potassium channel blocking agents. I then assessed ex vivo, the inherent contractility of the placebo versus progesterone groups from myometrium sampled from the STOPPIT cohort of patients. I carried out cell culture experiments on myometrium from healthy singleton women who were not in labour at the time of sampling. Myometrial explants were placed in cell culture medium, cultured for 1, 4 and 24 hours, and the supernatants were then analysed using Bio-Plex array technology to ascertain cytokine release. I selected time points and concentrations conditions to incubate myometrial tissue using progesterone and 17OHPC and was able to assess cytokine release. The myometrial explants were used for subsequent molecular studies. I performed real time-polymerase chain reaction (RT-PCR) (Abi,Taqman) to quantitate endogenous inhibitors of uterine contractility (cGRPR, EP2,NOS), cytokines (interleukins- IL6, IL8, IL1β), uterine stimulants COX-2 and gap junction components ( connexin 26 and connexin 43) expressed relative to housekeeping gene 18s. Lastly, I analysed the STOPPIT cohort of myometrial samples for to determine the in vivo effect of progesterone. We carried out RT-PCR (Abi,Taqman) to quantitate endogenous inhibitors of uterine contractility (cGRPR, EP2,NOS, PGDH), cytokines (IL6, IL8, IL1β) and gap junction components (connexin 26 and 43).I performed immunohistochemistry, staining for localisation of pro-inflammatory cytokines. I then carried out protein expression analysis using Western blot for contraction associate protein, connexin 43. The project was approved by North Glasgow University Hospitals Research Ethics Committee ref no. 05/S0705/18. All patients gave written informed consent to participate. Results: I found that progesterone, exerted consistent, rapid and sustained inhibition of the amplitude of spontaneous myometrial contractions in vitro at high concentrations however, this affect was not achieved through the principal potassium channels. Levcromakalim, a KATP channel opener, produced a concentration-dependent inhibition of the amplitude and frequency of spontaneous contractions. These effects were antagonised by the KATP channel blocker, glibenclamide. In contrast, glibenclamide had no effect on the progesterone-induced inhibition of myometrial contractility. Charybdotoxin 10 nM (which blocks IKCa, BKCa and Kv channels), iberiotoxin 100 nM (which blocks BKCa channels) and apamin 100 nM (which blocks SKCa channels) failed to affect the ability of progesterone to inhibit myometrial contractility. In contrast, 17OHPC did not exert any inhibitory effect on myometrial activity in vitro. Results indicated, at the selected pharmacological doses used in vitro that progesterone did not increase sensitivity to tocolytic agents tested. There was no decrease in sensitivity to the uterotonin oxytocin. Lastly, from our STOPPIT patient cohort I demonstrated no difference between the progesterone and placebo groups in either spontaneous contractility, response to tocolytics as above or response to oxytocin. One main conclusion of this arm of the study is that in vivo progesterone therapy to prevent pre-term labour does not appear to modify contractility ex vivo. I demonstrated that administration of progesterone but not 17OHPC for up to 24 hours in vitro does not appear to modify mRNA expression of uterine stimulants such as cytokines, COX-2 or endogenous uterine relaxants such as NOS and PGDH. Progesterone but not 17OHPC inhibited production of gap junction component connexin 43. This modification of contraction associated protein is in agreement with other literature presented on human myometrial data in vitro (4) . I used STOPPIT patients as a potential example of the myometrial effects of progesterone in vivo with a placebo treated control group. Prolonged maternal administration of progesterone appeared to inhibit expression of gap junction components connexin 26 and 43 in myometrium. Connexin 43 importantly, was also modified in vitro within the progesterone treated arm. However, ex-vivo assessment of the functional impact on human myometrium does not demonstrate a long-term inhibitory impact on myometrial function with down regulation of endogenous contractile inhibitors such as eNOS and EP2. The connexins play an essential role in regulating synchronous myometrial contractions. If progesterone has been of benefit in those at risk of preterm labour with a history of spontaneous preterm birth, it is possible therefore that this is by reducing connexin expression, which prevents the development of these synchronous contractions whilst on progesterone therapy. In summary, I have demonstrated putative mechanisms by which progesterone (and its analogue 17OHPC) might prevent preterm birth. Further studies characterising these pathways might inform the design of other agents which could provide additional efficacy in preventing preterm delivery.