The work in this thesis describes a series of studies utilising diverse data sources which were analysed using a number of regression methods (logistic, linear, Cox, Poisson), to address the factors associated with normal and abnormal pregnancy outcome. A series of maternal characteristics were related to adverse pregnancy outcome. Teenage pregnancy was not associated with an increased risk of any adverse outcome among first births, but was strongly associated with adverse outcome among second births [8]. Parity also interacted with body mass index: maternal obesity was associated with an increased risk of preterm birth among nulliparous but not multiparous women. This was explained by higher rates of elective preterm deliveries among obese nullipara [31]. There was a linear relationship between maternal age and the duration of labour, and the risk of operative vaginal and caesarean delivery [37]. It was hypothesised that age-related deterioration in obstetric performance was due to prolonged hormonal stimulation prior to the first birth. This hypothesis was supported by the observation that later menarche was independently associated with a decreased risk of operative delivery [38]. A short inter-pregnancy interval was associated with an increased risk of spontaneous preterm birth, but not stillbirth or intra-uterine growth restriction [16]. The risk of unexplained stillbirth at term was increased among nulliparous women [5] and nulliparous women also had slightly longer pregnancies [7]. A U-shaped relationship between birth weight and caesarean risk was observed at term. There was an interaction between fetal sex and caesarean risk: small boys were at increased risk of emergency caesarean [3]. The same interaction was observed for antepartum stillbirth [4]. Previous pregnancy outcome was predictive of the outcome of subsequent pregnancies. Women who were delivered by caesarean section in their first pregnancy had an increased risk of unexplained stillbirth in their second [17]. This finding was confirmed in a separate cohort and associations were also observed between previous complicated livebirths and the subsequent risk of unexplained stillbirth [32]. Some specific situations were also studied (vaginal birth after caesarean section (VBAC) and twins). Among women attempting VBAC, the absolute risk of delivery-related perinatal death was comparable to primiparous women but was significantly higher than women delivered by elective caesarean section [11]. The risk of perinatal death associated with uterine rupture was increased in low throughput obstetric units and among women induced with prostaglandins [19]. Using simple maternal characteristics, approximately 50% of women attempting VBAC could be classified into having a high (>40%) or low (<10%) risk of emergency caesarean [24]. This was better discrimination than could be achieved using similar characteristics among nulliparous women being induced at term [21]. The risk of delivery related perinatal death was increased among second twins, although this was only evident among births at term [13]. The association was observed among sex discordant twins, but was not observed among twins delivered by elective caesarean section [23]. The association between birth order and the risk of death due to anoxia was confirmed in data from England and Wales [33]. Ultrasonic measurements of the fetus were related to eventual birth weight. The range of error associated with such estimates was quantified and abdominal circumference on its own was as predictive as models using abdominal circumference and femur length [1]. Estimating fetal weight using ultrasound was not found to be a better measure of human fetal blood volume than simply using gestational age [10]. A series of ultrasonic measurements in the first and second trimester were predictive of pregnancy outcome, including smaller than expected crown rump length and intra-uterine growth restriction, preterm birth and low birth weight [2]; a long cervix in mid gestation and caesarean section [36]; and, high resistance patterns of uterine artery Doppler flow velocimetry and stillbirth [30]. Biochemical measurements performed in early pregnancy were also predictive of later adverse outcome: low maternal levels of pregnancy-associated plasma protein A (PAPP-A) were associated with an increased of pre-eclampsia, preterm birth and growth restriction [9]; low PAPP-A prior to 13 weeks was associated with birth weight at term in healthy pregnancies [12] and with a dramatically increased risk of stillbirth due to placental dysfunction [22]. Low first trimester levels of placenta growth factor were associated with increased risks of pre-eclampsia and growth restriction, whereas there was no association between elevated levels of the soluble fms-like receptor and adverse outcome [35]. Measurements of biochemical variables in the second trimester were also predictive of outcome, with elevated maternal serum alphafetoprotein (AFP) being associated with an increased risk of stillbirth [34] and spontaneous preterm birth [29]. Women with the combination of low first trimester PAPP-A and high second trimester AFP were at particularly high risk of complications, reflecting the synergistic predictive ability of the two measures [27]. Given proposed similarities between stillbirth and sudden infant death syndrome (SIDS), this outcome was also studied. Elevated second trimester levels of AFP were also associated with an increased subsequent risk of SIDS [20]. Women with a pregnancy resulting ultimately in SIDS were found to be more likely to have had complications in past and future pregnancies [25]. The risk of SIDS declined with advancing gestational age at term following spontaneous, but not elective birth [15]. Obstetric characteristics were used to generate a predictive model for SIDS [26]. Pregnancy outcome was also predictive of other aspects of child health, specifically, respiratory morbidity following birth at term was associated with an increased risk of hospital admission for asthma [18]. Pregnancy complications were also related to long term maternal health. Elective caesarean delivery for breech presentation did not appear to have an independent effect on fertility [28]. However, pregnancy complications were associated with the mother’s subsequent experience of cardiovascular disease. Women experiencing growth restriction, preterm birth or pre-eclampsia were at increased risk of subsequent ischaemic heart disease (IHD) [6] and the risk of this was also related to the number of miscarriages experienced prior to the first birth [14]. The parents of women who had experienced pregnancy complications or recurrent miscarriage had an increased incidence of IHD [39 & 40, respectively].

The Apc gene encodes the Adenomatous polyposis coli tumour suppressor protein, the germ line mutation of which characterizes Familial Adenomatous Polyposis (FAP), an autosomal syndrome characterized by multiple colorectal lesions. Inactivation of the Apc gene is recognized as a key early event in the development of colorectal cancers and leads to the deregulation of the Wnt pathway and the activation of TCF/LEF target genes. This project focuses on the proto-oncogene c-Myc as it is a key Wnt target gene which is activated following loss of Apc in vivo. This upregulation is noteworthy as c-Myc is implicated in stem cell survival, proliferation, apoptosis and tumourigenesis. Previous studies have shown c-Myc dependency for both apoptosis and proliferation following activation of the Wnt pathway, however little is known about the role c-Myc plays in inducing apoptosis following DNA damage in vivo. To study this I have conditionally deleted c-Myc from the intestinal epithelium and examined the response of intestinal enterocytes following DNA damage. Remarkably, following DNA damage, c-Myc deficient enterocytes were unable to upregulate p53 and induce apoptosis, which was mechanistically due to an upregulation of MDM2. Taken together, results from this study showed for the first time in vivo, a key role for c-Myc in inducing apoptosis following DNA damage through control of p53. Previous studies from this lab have shown that within the intestinal epithelium, c-Myc is absolutely required for the hyper-proliferative phenotype that is observed following loss of Apc. Therefore one of the key aims of this thesis is to look downstream of c-Myc in order to delineate how c-Myc induces and controls this proliferation. Given that one of the key postulated functions of c-Myc is the transcriptional repression of p21, this thesis examines this hypothesis by investigating the significance of the upregulation of p21 following c-Myc deletion in Apc deficient intestinal enterocytes. To do this, I have generated triple knockout (TKO) intestines by intercrossing p21 knockout mice to mice where we can conditionally delete both Apc and c-Myc within the murine intestinal epithelium. Surprisingly, the levels of proliferation were the same between double knockout Apc Myc and TKO intestines, which had markedly less proliferation than Apc deficient intestines. However, unlike double knockout enterocytes, TKO intestinal enterocytes no longer moved up the crypt-villus axis and failed to generate villus. To examine which of these phenomena were key to tumourigenesis (differentiation or proliferation), we investigated whether TKO intestines could form intestinal adenomas and found that even in the absence of p21, c-Myc deficient cells were unable to form tumours. Taken together we have identified a novel role for p21 in driving differentiation following Apc and Myc deletion. This is consistent with the expression of p21 in the normal crypt at the crypt villus junction. Remarkably this function of p21 is independent of its key role as a cell cycle inhibitor. Moreover, this study also examined the importance of the upregulation of the Cyclin D/CDK4 complexes following Apc loss and their role in c-Myc dependent proliferation. Results from these studies showed that Cyclin D2 is required for efficient proliferation immediately following loss of Apc as well as for tumourigeneis in the Apc Min/+ mouse. Taken together, results from these studies showed that the upregulation of Cyclin D2 and CDK4 are c-Myc dependent and that the upregulation of these complexes are key for Wnt driven proliferation and tumourigenesis. Lastly, in this study I have examined whether Apc loss within the intestinal epithelium, where it is a bona fide tumour suppressor gene, can provoke senescence, and compared this to the ability of Apc gene deletion to trigger senescence in the renal epithelium, where it is not mutated in human cancer. This study showed that deletion of Apc within the renal epithelium invoked a p21 dependent senescence response, and Apc deficient renal epithelial cells were cleared and very rarely initiated tumourigenesis. However, combined Apc and p21 gene deletion rapidly initiated tumourigenesis, with all mice developing renal carcinoma by 2 months of age. In contrast to Apc deficient intestinal epithelium, this process was unaffected by loss of c-Myc. However within the intestinal epithelium, deletion of Apc did not invoke senescence, but lead to a highly proliferative, p21 independent response. Combined Apc and p21 gene loss had no impact on either the short term phenotypes of Apc loss or upon tumourigenesis. Taken together these results show for the first time that Apc loss in vivo can invoke a senescence program but in a context dependent fashion. This implies escape from senescence is not a crucial pathway in colorectal cancers that are initiated by Apc loss, and goes to explain why renal carcinoma is not observed in FAP patients who are germline heterozygous for APC. Therefore the aims for this thesis are: • To investigate the role of c-Myc in inducing apoptosis within the intestinal crypt, and whether this is p21 dependent? • To investigate the role of p21 in causing senescence of Apc deficient cells, and whether this is c-Myc dependent? • To determine the functional importance of repression of p21 by c-Myc in Apc deficient cells. • To determine the significance of Cyclin D2 upregulation within Apc deficient cells.

The focus of this study is the potential use of FTIR imaging as a tool for objective automated histopathology. The Thesis also reports the use of multivariate statistical techniques to analyse the FTIR imaging data. These include Principal Component Analysis (PCA), Hierarchical Cluster Analysis (HCA), Multivariate Curve Resolution (MCR) and Fuzzy C-Means Clustering (FCM). The development of a new PCA-FCM Clustering hybrid that can automatically detect the optimum clustering structure is also reported. Chapter 1 provides a brief introduction to the use of vibrational spectroscopy to characterise biomolecules in tissues and cells for medical diagnosis. Chapter 2 details the basic histology of a lymph node before proceeding to present imaging results gained from the analysis of both healthy and diseased lymph node tissue sections. The ability of each multivariate technique to discriminate different tissue types is discussed. In addition, the spectral features that are characteristic for each tissue type are reported. The development and application of a new PCA-FCM Clustering algorithm that can automatically determine the best clustering structure is also described in full. The results indicate that cellular abnormality provides changes to both the protein and nucleic acid vibrations. However, similar spectral profiles were identified for highly proliferating cells that were contained within reactive germinal centres of the lymph node. Chapter 3 provides a short introduction to the histology of the cervlx before presenting imaging results that were gained from the analysis of both healthy and diseased cervical tissue sections. The ability of each multivariate technique to discriminate different tissue types is discussed. In addition, the spectral features that are characteristic for each tissue type are described in detail. Novel imaging experiments upon exfoliated cervical cells are also presented. It would appear that cellular abnormality in cervical tissues and cells affects both the protein and nucleic acid features of the spectra. Glycogen and glycoprotein contributions that are prevalent in healthy tissues are also absent. Chapter 4 details sample preparation methods, the instrumentation and procedures used for data acquisition, and the subsequent data processing and multivariate techniques applied to analyse the collected spectral datasets.

Campylobacter spp. are the most common causes of bacterial gastroenteritis in humans worldwide, and although poultry and cattle are considered major sources of Campylobacter spp., infection has also been associated with dogs. In order to investigate the potential zoonotic risk to humans, dog faeces were examined for the presence of Campylobacter spp. from several different dog populations including; vet-visiting, boarding, rescue and hunt dogs. The Campylobacter spp. prevalence, and species distribution was determined for all studies, and some studies were analysed for possible risk factors for Campylobacter spp. carriage in dogs. Longitudinal studies were carried out on kennelled dogs to investigate shedding patterns, and possible transmission. All C. jejuni, and 41 C. upsaliensis isolates from these studies underwent multilocus sequence typing (MLST), along with nine C. upsaliensis isolates originating from human clinical cases, in order to identify possible sources of infection, and assess the potential zoonotic risk to humans. Additionally a pilot study was performed to annotate a plasmid as part of a C. upsaliensis genome project. The findings of this thesis found that the overall prevalence of Campylobacter spp. ranged from 0-73%, although the majority of studies had a prevalence greater than 30%. The prevalence and species distribution differed depending upon the dog population. Kennelled dogs generally demonstrated the highest overall Campylobacter spp. prevalence, whilst the greatest species diversity was found in hunt dogs. C. upsaliensis dominated in most of the populations sampled, except for two hunt kennels where C. lari and C. jejuni dominated. The prevalence of C. jejuni was relatively high in some of the rescue and hunt kennels, reaching 20% and 26% respectively, whereas in vet-visiting and boarding dogs it was relatively low, 1.2-9%. Longitudinal studies indicated that the majority of dogs entered the kennels already carrying Campylobacter spp. but when possible transmission events occurred they often involved C. jejuni. Rescue dogs appeared to be exposed to sources of C. jejuni before and after entry to the kennel, but boarding dogs were only exposed after entry. The shedding of C. jejuni in dogs appeared to be over short durations, whereas dogs that carried C. upsaliensis shed the bacterium in nearly every sample. Data suggested that dogs carried the same C. upsaliensis strain throughout the study, providing further evidence that the species may act as a commensal in dogs. Further to this no associations could be made between Campylobacter spp. carriage, specifically C. upsaliensis, and disease in dogs in any of the studies. Younger dogs were significantly more likely to carry C. upsaliensis than older dogs in the vet-visiting study (OR for every additional month 0.99) and living with another dog carrying Campylobacter spp., was significantly associated with Campylobacter spp. carriage in dogs. A considerable amount of genetic diversity was observed within the C. jejuni and C. upsaliensis isolates originating from dogs, and MLST results suggested that strains of both species were the same, or highly similar to strains found in humans. This suggests that there may be common sources of infection for both humans and dogs and that dogs remain a potential zoonotic risk to humans. Although only a small number of household dogs carry C. jejuni, infected dogs should still be considered a potential zoonotic risk to humans, particularly if the dogs originate from kennelled or hunt kennel populations where the prevalence may be higher. Dogs remain a significant reservoir of C. upsaliensis, but the relationship between the presence of C. upsaliensis and gastroenteritis in both dogs and humans is still unclear.

There is growing evidence that suggests Traumatic brain injury (TBI) may initiate long-term neurodegenerative processes. Exposure to a single moderate or severe TBI, or to repetitive TBI, reveals a complex of pathologies including abnormalities of tau, amyloid-β and TDP-43; neuronal loss; neuroinflammation; and white matter degradation. The mechanisms driving these late post-TBI neurodegenerative pathologies remain elusive. Firstly, a potential association between blood-brain barrier (BBB) disruption and TBI was investigated. Results showed that increased and widespread BBB disruption was observed in material from patients dying in the acute phase following a single, moderate to severe TBI and persisted in a high proportion of patients surviving years following injury. Furthermore, there was preferential distribution to the deep layers of the cortex and to the crests of the gyri rather than the depths of the sulci. This post-TBI BBB disruption was investigated further within a paediatric TBI cohort. BBB disruption was noted in both paediatric and adult TBI in a similar pattern and distribution, however, interestingly, in sharp contrast to adult TBI cases, BBB disruption in paediatric cases appears preferentially distributed to capillary sized vessels. This vulnerability of the small vessels was rarely observed in adult material. In addition to the post-TBI vascular change observed, the cellular response was investigated, which interestingly, demonstrated regional differences. Specifically, in the grey matter, reactive astrogliosis was observed subpially, around cortical vessels, at the grey and white matter boundaries and subependymally. This astrogliosis was evident in a proportion of acute and continued into the late phase following TBI. In contrast, microglial activation was observed as a delayed response and localised to the white matter tracts. In addition, this delayed microglial response expressed an M2-like phenotype. Furthermore, there was an increased population of inactivated perivascular microglia beyond the perivascular space in the grey matter regions, observed in the acute phase and persisted in a proportion of patients surviving years following injury. Collectively these findings are interesting and indicate TBI induces both a vascular and cellular responses which may contribute to the long-term post-TBI neurodegenerative processes.

Diffuse, multi vessel coronary artery disease (CAD) affects about one third of patients with CAD and is associated with worse outcomes. Abnormal vascular stiffness and function (e.g., reflected by increased endothelial microparticles and diminished microvascular endothelial-mediated responses), cell mediated pro-inflammatory status (e.g., reflected by levels of specific monocyte subsets), and platelet function (e.g., increased monocyte-platelet aggregates (MPAs) and platelet microparticles) have established roles in CAD pathogenesis but their contribution to the unfavourable diffuse CAD form is unclear. The aim of this study was to compare measures of vascular function, monocyte subsets, MP As, and endothelial and platelet microparticles in patients with diffuse and focal CAD and subjects without CAD. Additionally, prospective changes in these characteristics were analysed over one year. I found increased counts of aggregates of Mon2 monocyte subset with platelets and apoptotic endothelial microparticles in patients with diffuse CAD and I identified a negative correlation between Mon2 MPAs and microvascular endothelial function and increased diastolic elastance. My findings suggest that excessive levels of Mon2 aggregates with platelets and apoptotic endothelial micropa1iicles could be important contributors to the diffuse type of CAD by a mechanism involving microvascular endothelial dysfunction and abnonnal cardio-vascular interactions.

Type 1 diabetes (T1D) is characterised by pancreatic cell autoimmunity and inflammation, resulting in cell islet destruction and insulin deficiency. Prospective studies from different continents have shown that insulin resistance is independently associated with risk for the development of T1D. We wanted to investigate the role of adiponectin in mediating this link. Adiponectin is a circulating adipokine whose anti-inflammatory and insulin sensitising actions appear to be mediated via two related receptors, AdipoR1 and AdipoR2. We began by characterising adiponectin receptor expression on PBMC by flow cytometry. We showed that monocytes express both receptors abundantly, that this expression correlates with insulin sensitivity in both health and diabetes. Furthermore, expression can be increased with lifestyle intervention. Adiponectin receptor expression on monocytes is reduced in T1D, and we demonstrate this leads to an apparent resistance in the ability of adiponectin to inhibit the stimulatory capacity of antigen presenting cells (APC). Specifically, we show that adiponectin inhibits the stimulatory capacity of APCs through down-regulation of CD86 expression, and that this effect is decreased in T1D. In this way, the release from the regulatory effects of adiponectin is one potential mechanism by which immune tolerance is lost in T1D.

Obesity is known to be associated with obstructive sleep apnoea (OSA) and Type 2 diabetes mellitus (T2DM). The effect of OSA in very severely obese individuals is not well documented. In this thesis, I compared the effect of OSA in South Asians and white Europeans, examined the effect of OSA on glycaemic control among T2DM, and explored the relationship between OSA and diabetic retinal and kidney diseases in a severely obese population. I also systematically reviewed the effect of OSA on diabetic kidney and retinal diseases. Findings from this thesis were 1) severely obese South Asians had greater severity of OSA compared to white Europeans and the mechanisms mediating this require further investigation, 2) a high OSA prevalence in T2DM individuals with a positive relationship between nocturnal hypoxia and glycaemic control, 3) severity of hypoxaemia during sleep may be an important factor in the development of diabetic retinal complications, 4) duration of hypoxaemia during sleep were inversely associated with renal function in T2DM and 5) from the systematic review, there is a need for future large cohort studies with long term follow-up data to examine the long-term effects of OSA and other sleep parameters on diabetic retinal and kidney diseases.

The increased risk of cardiovascular disease (CVD) amongst South Asians (SAs) is unclear. This thesis examined potential biomarkers to address this. Cross-sectional data on SAs from community (n=1304) and hospital (n=148) populations was collected. Biomarkers were analysed by genotyping, mass spectrometry, automated-immuno-colourimetric-assays, ELISAs, and a new in-house assay for a novel marker, ferritin bound to apolipoprotein B. Diagnostic performance was assessed using receiver operating curves, logistic and linear regression models. C-reactive protein (CRP) was a comprehensive marker of CVD risk, where a range of 1.43-2.30 mg/L maximised sensitivity and specificity. CRP SNP (single nucleotide polymorphism) -390C>T/A contributed minimally to variation in CRP levels. Non-fasting triglycerides discriminated SAs at increased CVD risk, where APOA5 SNP -1131T>C was an independent predictor of triglycerides but APOC3 SNP -455T>C and -482C>T were not associated with triglycerides. The performance of IL-6, vWF, D-dimer and P-selectin were poor in comparison to CRP and triglycerides. BNP discriminated SAs with systolic heart failure with a cut off value of 36.4 pmol/l. Of the newly investigated biomarkers, a link between haemoglobin abnormalities and CVD was observed potentially through a mechanism involving iron transportation on lipoproteins. CRP and triglycerides should be considered in the routine CVD risk assessment of SAs.

Tuberculosis (TB) is an infectious bacterial disease mainly infecting the pulmonary system of the human body. It affects around 1.5 million people every year, most of whom live in developing countries. The incidence of TB has increased in line with the rise in incidences of Human Immunodeficiency Virus (HIV) infections and Acquired immune deficiency syndrome (AIDS). Due to the pressing concerns of TB, the World Health Organisation (WHO) came up with the Direct Observed Treatment (DOTS) programme. Unfortunately, the development of several resistant strains against first-line drugs and consequently second and third-line drugs have developed. As the current TB drug regimen is inadequate, a good screening strategy, discovery of newer drugs and identification of the mode of action would help in developing better treatment routines and determining bacterial pathways more clearly. Drug discovery follows two major routes, one leading from the drug to the target and the other from target to the drug. Both methods have been applied in this work in order to identify new drugs effective against mycobacteria. Screens performed against a drug library approved by the Food and Drug Administration (FDA) have resulted in some promising hits. Functional characterisation of a putative enoyl CoA hydratase EchA12, which was targeted by florfenicol, revealed a novel lipid chaperone functionality associated with cell wall lipid biosynthesis. Furthermore, a target based phenotypic drug screen of the GSK177 box set against Mtb-PrsA provided further evidence that this enzyme as a viable drug target (Ballell et. al., 2013).