Bariatric surgery offers sustained dramatic weight loss and remission of diabetes, yet the mechanisms of these health benefits are not clear. In the present study, I profiled circulating and colorectal miRNAome in response to bariatric surgery (Roux-en-Y gastric bypass). Indeed, the response of circulating and colorectal miRNA profiles to RYGB were striking and selective. Fourteen circulating microRNA and thirteen colorectal microRNA exhibited significantly alteration post RYGB. Interestingly, circulating miR-122 decreased dramatically (56 fold) post RYGB surgery. The expression of hepatic miR-122 and its metabolic targets were examined both in in vivo RYGB surgical model and in an in vitro mechanistic model. Manipulation of miRNA-122 cold induce changes of key enzymes involved in energy metabolism, glucose transport, glycolysis, TCA cycle, pentose phosphate shunt, fatty acid oxidation and gluconeogenesis, suggesting an overall increased energy expenditure status after RYGB. Furthermore, potential mechanisms involved in the control of hepatic miR-122 were investigated, with focus on metabolites (glucose, and fatty acids), hormones (glucocorticoid) and transcription factors (PPARs). Finally, by correlating the circulating miRNAome and metabolome data, we were able to generate a comprehensive landscape of the crosstalk between miRNAs and metabolic pathways. Follow-up studies will allow a detailed understanding of miRNAs responsible for regulating specific metabolic pathways, and conversely identifying metabolites capable of regulating the expression and activity of specific miRNAs.

The role of Lemur tyrosine kinase 3 (LMTK3) and its association with cell proliferation and endocrine resistance in breast cancer due to its ability of regulating estrogen receptor α (ERα) has been previously addressed in our laboratory. However, the ER-independent function of LMTK3 has not been studied yet. We found that LMTK3 promotes the development of a metastatic phenotype by inducing the expression of genes encoding integrin subunits. Invasive behaviour such as actin cytoskeleton remodelling and focal adhesion were positively correlated with the abundance of LMTK3 formation in various breast cancer cell lines. Using SILAC (stable isotope labelling by amino acids in cell culture) proteomic analysis, we found that LMTK3 increases the protein levels of integrin subunits α5 and β1 through activating the CDC42 GTPase, which promotes integrin α5 and β1 expression via the transcription factor serum response factor (SRF). Furthermore, abundance of LMTK3 was positively correlated with that of integrin β1 in breast cancer patients' tumours. As LMTK3 is also localised in the nucleus, we then investigated its nuclear function. We mapped LMTK3 binding across the genome using ChIP-seq and found that LMTK3 binding events are correlated with repressive chromatin markers. We further identified KRAB-associated protein 1 (KAP1) as a binding partner of LMTK3. The LMTK3/KAP1 interaction is stabilized by PP1α, which suppresses KAP1 phosphorylation specifically at LMTK3-associated chromatin regions, inducing chromatin condensation and resulting in transcriptional repression of LMTK3-bound tumour suppressor-like genes. Furthermore, LMTK3 functions at distal regions in tethering the chromatin to the nuclear periphery, resulting in H3K9me3 modification and gene silencing. In summary, we propose a model where a scaffolding function of nuclear LMTK3 promotes cancer progression through chromatin remodelling in an ERα-independent and kinase-independent manner.

Non ischaemic dilated cardiomyopathy (DCM) is an important cause of heart failure leading to chronic morbidity and death and as such is a major health burden. DCM is familial in up to 50% of cases but is genetically heterogeneous, hindering both genotype-phenotype studies and the application of genetic information for stratified patient management. TTN truncating variants (TTNtv) cause severe and familial DCM, but sometimes occur in healthy individuals, posing challenges for the interpretation of these variants. In this PhD thesis, the power of quantitative cardiac MRI (CMR) is integrated with targeted resequencing of TTN in order to assess the relationship between TTN genotype, cardiac phenotype and clinical outcomes. A prospectively recruited DCM cohort was established following CMR assessment in 374 patients (88% Caucasian, 72% male, mean age 54 ± 35 years, mean left ventricular ejection fraction (LVEF) 38% ± 24.5%, mean indexed end-diastolic volume 129 ± 141 mm). Following several iterations of design, refinement and testing an NGS assay was produced that captures all TTN coding exons and splice sites. Optimal parameters for sequencing and analysis of variants were established and genotype data compiled from 374 prospective, unselected cases, 155 end-stage retrospective cases, and 308 MRI phenotyped healthy volunteers. These data were integrated with that from 3603 population subjects and together used to identify molecular signatures that aid interpretation of TTN truncations both in DCM and as incidental findings. Overall, TTNtv were identified in 1.4% of controls, 13% of unselected and 22% of end-stage DCM cases (OR = 13, P= 2.8x10-43, DCM vs controls) confirming TTN as the commonest cause of genetic DCM in all patient groups. TTNtv-containing exons in DCM have higher usage than those in controls (P=2.5x10-4) and these are estimated to have >93% probability of pathogenicity (likelihood ratio 14). Compared to TTNtv-ve DCM, TTNtv+ve patients had lower LVEF (P=0.02), thinner LV walls (P=0.02), and a higher incidence of sustained ventricular tachycardia (P=0.001). C-terminus TTNtv were also associated with lower LVEF versus N-terminus (β=-18±7%, p=0.006) and were more common in end-stage disease. Together these data provide the first insight into genotype-phenotype correlations and will be of benefit in variant interpretation and patient stratification in TTN-based DCM.

Human natural killer (NK) cells play an invaluable role in the first line of innate immune defence against viral infection and cancer. They are the main sub-group of lymphocytes with predominant expression of polymorphic KIR (Killer-immunoglobulin-like-receptors), which allows them to engage with their MHC Class I ligands. Past immunogenetic analysis has shown that possession of KIR2DL3 may confer an advantage in eliminating acute Hepatitis C virus (HCV) infection. This contrasts with those possessing KIR2DL2 who were found not to spontaneously resolve infection. Previous work has shown that different variants of an endogenous peptide VAPWNSLSL (VAP) can modulate the KIR2DL2/3-HLA-C1 interaction. Here we have analysed the peptide repertoire in the context of the inhibitory KIR2DL3, KIR2DL2 receptors and the activating KIR2DS2 receptor. We firstly investigated any differences in KIR2DL3+ or KIR2DL2+ NK cell reactivity in response to mixes containing VAP peptide derivatives with the aim that these peptide mixes would be representative of the diverse peptide repertoire on a cell. Overall, we found that NK cells from KIR2DL3 homozygous donors were more sensitive to changes in the peptide content of MHC class I than those from KIR2DL2 homozygous donors. As HCV is an infection with a well-recognised association of outcome with specific KIR, we also sought to determine whether HCV peptides could modulate KIR2DL2/3+ NK cell reactivity. We investigated the ability of the panel of HCV peptides to influence NK cell reactivity firstly via KIR2DL2 or KIR2DL3. The majority of HCV peptides were weak binders for HLA-Cw*0102 and/or non-inhibitory. We found that a HCV peptide (HCV NS31254-1263 LNPSVAATL) was able to promote weak binding of the KIR2DS2 to the HLA-Cw*0102 allele and subsequently affect KIR2DS2+ NK cell reactivity. Here we report that LNP can effectively promote binding of KIR2DS2 tetramers and affect the functionality of KIR2DS2+ NK cell clones. Thus, we propose that the HCV-derived peptide LNP can act as an agonistic peptide on NK cell activation through KIR2DS2.

TKI inhibitors have revolutionised CML therapy and the goals for management have shifted from finding newer therapies to optimising existing treatment approaches. We have tried to optimise CML therapy by identifying poor responders early by molecular monitoring, improve adherence by using self reported adherence and optimise intolerance by actively changing TKIs to overcome side effects. BCR-ABL PCR of < 10% at 3 months and < 1% at 6 months have become an accepted standard after the publication by Marin et al. We tried to combine the two measurements and showed that 3 month milestone predicts poor responders and is sufficient to consider changing therapy and that an additional measurement at 6 months does not add any further value. Most existing methods of determining adherence to medications are financially impossible to replicate on a day to day basis or too labour intensive. We tried to measure adherence by 4 different questionnaire based methods (visual adherence scale, Lu's scale, Haynes method and DAMS scale) and correlate it with clinical responses. We have showed that adherence by all methods correlated with clinical responses and Haynes method which quantifies adherence based on number of doses of medications missed over the last 7 days was the best indicator of adherence amongst all. We further looked at the interactions of daily routine, communication with the physician; access to internet and patients views on taking the medications with adherence to therapy and adherence was shown to be influenced by all of them. Majority of the patients on TKI therapy appeared to be anxious and nearly half of them depressed. Patients with a better QOL had improved adherences. We propose a model based on 4 questions with the most significance on multivariate analysis to be possibly used as a surrogate for adherence methods. It has been shown that intolerance affects adherence and hence outcomes. We have tried to improve intolerance by switching TKI therapy in patients who had attained CCyR and with chronic low grade side effects and showed that the side effects improved and all patients had further improvement in the molecular milestones with deepening responses.

Objectives: To examine the relationship between particular traditional risk factors and their effect on the development of gestational diabetes mellitus (GDM) and birthweight (Part 1) as well as to assess first trimester maternal biochemical predictors of development of GDM (Part 2). Methods: Part 1. A retrospective study of prospectively collected data from fifteen maternity units in North West London between 1988-2000- the St Mary's Maternity Information System (SMMIS) dataset. The dataset was modified to include only those who were nulliparous (thus ensuring that only one pregnancy per woman was included) and excluding women with pre-existing diabetes (thus studying only women who either did or didn't develop gestational diabetes). Birthweight z-scores were calculated. Part 2. A nested case-control study using first-trimester (11+0 to 13+6 weeks of gestation) samples. that were obtained as part of a large prospective observational on-going study aimed at identifying first-trimester predictors of adverse pregnancy outcomes. Maternal levels of lipids (cholesterol, low density lipoprotein cholesterol (LDL), high density lipoprotein cholesterol (HDL), non-fasting triglycerides, C-reactive protein (CRP), γ-glutamyl transferase (γ- GT), adiponectin, E-selectin, tissue plasminogen activator (t-PA) and vitamin D (25(OH)D) were measured. Statistical Package for the Social Sciences (SPSS) Version 17.0 and R (version 2.11.0) was used for statistical analysis. Results Part 1. There was a strong association between advancing maternal age and increasing body mass index (BMI) on the development of GDM (p < 0.01 for both). This varied within each racial group and was more pronounced in Black African and South Asian groups. Using White European women with a BMI of 18.5-24.9 as a reference group, Black African and South Asian pregnant women had higher Odds Ratios (ORs) for GDM development within all BMI categories compared to the reference group. Maternal BMI was positively associated with birthweight z-scores within all racial groups (p < 0.001 for all) irrespective of glycaemic status but its effect was much greater in women with GDM. The difference in birthweight z-scores between GDM and non-GDM women varied according to racial group and was much higher in non-white racial groups and at high rather than at low BMIs. Part 2. Simple maternal demographic and clinical characteristics obtained at the first antenatal visit provide a good prediction of GDM. Low levels of HDL and high levels of t-PA are independent predictors of GDM. (p=0.001 and p < 0.001 respectively). First trimester maternal serum 25(OH)D levels are not associated with the development of GDM. Conclusions Maternal age and BMI interact with racial group in relation to the development of GDM. Both factors are important in the development of GDM, particularly so in Black African and South Asian women. GDM strongly accentuates the effect of BMI on birthweight, especially within non-white populations. First trimester prediction of GDM can be enhanced by the measurement of specific maternal biomarkers.

Optical polarization is sensitive to morphological, structural and compositional changes of tissue and has attracted much interest as a tool in tissue sensing and characterisation. The fusion of polarisation imaging techniques and medical endoscopy resulting in polarisation resolved endoscopy is one of the most significant steps to translate the technique from an optical laboratory to clinic so as to benefit the whole spectrum of endoscopic investigations and intra-operative guidance in situ during minimally invasive surgery. The work in this thesis focuses on the proof-of-concept studies concerning the development of polarisation resolved endoscopy. In particular, polarised light scattering spectral imaging, 3x3 and 4x4 Mueller polarimetric imaging are successfully integrated to a rigid endoscope with accompanying validation experiments performed. The results have shown that polarisation resolved endoscopy based on light scattering spectroscopy and Mueller polarimetry is feasible to implement and has great potential to become a powerful tool for tissue imaging and characterisation.

Glucagon-like peptide-2 (GLP-2) is a pleiotropic peptide secreted in the human intestine with known intestinotrophic properties beneficial in conditions like short bowel syndrome (SBS); a condition characterized by malabsorption of both fluid and nutrients. Left untreated, SBS can lead to dehydration, malnutrition, and weight loss. Teduglutide, a long acting analogue of GLP-2, has been used in multiple clinical studies to elucidate its trophic properties. Murine studies however have also shown that GLP-2 inhibits pro-inflammatory cytokines raising the possibility of an anti-inflammatory property and its potential use in intestinal inflammatory conditions; in particular inflammatory bowel disease (IBD). Dendritic cells (DC) play a central role in the initiation and regulation of the immune system. They bridge the innate and adaptive immune systems and are unique in their ability to activate naïve T cells as well as dictate the type of T-cell immunity. We hypothesized that GLP-2 has an immunomodulatory role and exert this action via DC. Toxic effects of GLP-2 peptide on human DC in-vitro have not previously been experimented. Our experiments showed that GLP-2 did not have a toxic effect on DC at 1pM, 1nM and 1μM concentrations and hence we were subsequently able to look at the effects of GLP-2 on human DC phenotype and function. Using whole blood and intestinal biopsies from healthy volunteers, we obtained a population of enriched low density cells (LDC) which offered a novel and 'physiological' model for DC. These cells were labelled with appropriate fluorochromes and assayed by a flow cytometer. We established that DC incubated overnight with GLP-2 had a reduced intensity ratio of HLA DR and an increased expression of CD14 in a dose dependent way compared with controls. The expression of co-stimulatory molecule CD86 was also higher in the treated DC. This phenotypic change suggests that GLP-2 modulated DC into an immature state although still able to stimulate T-cell proliferation. Ongoing cytokine production of IFN-γ and IL-12 from healthy blood DC was inhibited by GLP-2 however only cytokine production of IFN-γ from intestinal lamina propria DC was inhibited. These findings suggest that GLP-2 may induce a 'homeostatic' or 'immuno-tolerant' state and block Th1 cytokines in DC. Functional experiments confirmed that GLP-2 modulated DC enhanced T cell proliferation although this occurred only with intestinal DC. GLP-2 conditioned DC also functionally affected the cytokine profile of T cells by reducing the cytokines IFN-γ in both human blood and intestinal DC and IL-12 in only the latter. Hence our human DC in-vitro findings mirror some of the results found in murine studies showing GLP-2 effects on blocking Th1 cytokines. The results suggest that GLP-2 has an immunoregulatory effect and that the mechanism of action may possibly involve direct effects on DC. GLP-2 therefore is able to modulate DC characteristics and function leading to future application as an immunotherapy for inflammatory diseases.

The presence of disease, including infectious disease, has been observed to give rise to specific patterns of gene expression in peripheral whole blood, regardless of disease site. These gene expression signatures allow for distinction between diseases and have the potential to reform diagnostics, particularly in diseases and patient groups for whom current diagnostics are unreliable, like Tuberculosis (TB). Although TB is a treatable infectious disease, it has high morbidity and mortality, especially in low resource countries and HIV infected patients. In this thesis, I propose a bioinformatics toolbox that derives minimal transcriptomic signatures from microarray datasets acquired from heterogeneous groups regardless of underlying co-infections and geographic locations. The transcripts' expression values are then aggregated into a single value disease risk score (DRS) for every patient, that allows for classification between the disease groups in a binary manner. The toolbox was employed to analyse an adult and a paediatric TB transcriptomic study, comprising HIV infected and uninfected patients from sub-Saharan Africa. In the adult study, the DRS based on a 27-transcript signature distinguished culture confirmed TB from latent TB infection (LTBI), while 44 transcripts distinguished TB from other diseases phenotypically similar to TB (OD), with high sensitivity and specificity. Out-of-sample validation was performed using a publicly available dataset. In the paediatric study, a 51-transcript signature distinguished TB from OD and a 42-transcript signature from LTBI. The signatures were validated out-of-sample using an independent cohort and benchmarked against culture-negative TB patients and Xpert® MTB/RIF, currently used for detection of M. tuberculosis. This thesis provides proof of principle that minimal host blood transcriptional signatures are able to distinguish TB from LTBI and OD regardless of HIV infection. The subsequent transformation of the signatures into a score for every patient may facilitate disease categorisation and potentially development of diagnostic tools.

Early human fetuses have the ability to heal wounds by completely regenerating tissues, leaving no evidence of scarring. However in the adult scarring is the inevitable endpoint of the wound healing process. Sometimes these scars can be pathological in nature causing both functional and aesthetic problems to those affected. Every year millions of people around the globe acquire problematic or pathological scars either whilst undergoing surgery or from traumatic injuries and at present there remain a severely limited number of pharmacological treatment options to offer these patients. Importantly currently there exists no treatment that can either eliminate or reliably reduce acquired scars. Not only is the treatment of acquired scars problematic but also the clinical assessment of scars is largely subjective in nature and frequently relies on assessment scales that show large amounts of inter-rater variation and lack quantification. Especially subjective is the measurement of scar colour, which can be markedly different from the surrounding skin and cause significant distress to the patient. Without an objective assessment framework clinicians cannot reliably examine scars nor gauge responses to any treatment. The aim of this thesis is thus two-fold. Firstly a new anti-scarring treatment in the form of insulin will be tested in a randomised, double blind, intra-patient, placebo controlled trial where patients undergoing elective bilateral breast surgery will have low-dose insulin injected subcutaneously to one breast and placebo to the other at the time of surgery. Patients will be followed up for 12 months and their scars compared to examine the therapeutic effect of insulin upon scars. Secondly the thesis aims to test the validity of new methods of assessing the scar colour of a subset of patients within the insulin trial using previously untested photographic devices and software. These devices are hoped to add much needed quantification to scar assessment.