The adult mammalian heart is no longer considered a post mitotic organ with evidence for homeostatic cell turnover and presence of populations of resident cardiac stem/progenitor cells. The epicardium is a single cell layer surrounding the myocardium and derives from the proepicardial organ during development. It has an important role in formation of the coronary vessels and the embryonic myocardium but, until recently, was considered to be quiescent in the adult heart. Recent studies have shown that reactivation of the adult epicardium post myocardial infarction (MI) injury induces a population of epicardium-derived progenitor cells (EPDCs) to proliferate, migrate and differentiate in the scar region. It remains unclear whether these EPDCs reside in an epicardial niche within the adult heart as classically defined in other lineages, such as the hair follicle bulge and intestinal crypt. Using confocal microscopy, we have identified cell clusters located in the intact epicardium, which express key extracellular matrix (ECM) and stem/progenitor cell markers and respond dynamically to injury, with cluster reformation occurring by day 21 post MI. We suggest that these clusters are candidates for a putative epicardial niche in the adult heart. Lineage tracing analyses to determine the origin of cells residing within the clusters have excluded Wt1+ or Gata5+ proepicardial lineages, but suggest a haematopoietic source. Using Vav1Cre; ROSA-tdTomato to label haematopoietic cells, we have shown that cells from the haematopoietic lineage contribute to the adult epicardium. We have tracked this contribution through development and into adulthood and show an increase in incidence of cells of the haematopoietic lineage in the epicardium with age. Results from bone marrow (BM) transplantation experiments, using whole BM we have shown contribution to the adult epicardium from BM cells. Labelling donor cells using Vav1Cre; ROSA-tdTomato shows that these cells are of haematopoietic lineage. These data suggest haematopoietic cells represent a novel source of cells that contribute to the epicardium during development and that are likely involved in replenishing epicardial cell turnover during cardiovascular homeostasis and may participate in cardiac repair.

Peptidylarginine Deiminases (PADs) modify protein conformation and function by converting protein-bound arginine to citrulline in a calcium-dependent manner. These processes have been shown to be important in the developing nervous system. Ferretti et al in 2011 demonstrated only PAD2 and PAD3 were detected in the developing chick spinal cord. PAD3 is of particular interest due to its expression in neural stem cells (NSCs) and possible involvement in secondary injury responses in chick, a process that involves calcium deregulation. Cl-amidine, a PAD inhibitor has been shown to significantly reduce spinal cord damage. The aims of my project were to investigate the developmental expression of PADs and their possible function in apoptosis in human central nervous system (CNS) as none of these issues were addressed previously. Out of 5 PADs in human only PAD2 and PAD3 were detected in developing CNS. PAD3 transcript decreases while PAD2 increases with development. I also found PAD3 expression in the spinal cord germinal zone, thereby indicating its expression in human NSCs (hNSCs), and then I established hNSC lines from human embryonic CNS and they showed PAD2 and PAD3 expression which were regulated upon differentiation. To investigate PAD3 involvement in apoptosis in human CNS, intracellular calcium was raised in hNSCs to mimic secondary injury responses. PAD3, but not PAD2, transcripts and citrullination are upregulated and apoptosis is significantly reduced by PAD3, but not PAD2, inhibition. Conversely, PAD inhibition in untreated hNSCs increases proliferation. Furthermore, I found PAD modulates the AIF-mediated apoptosis pathway as PAD is required for AIF cleavage and its translocation to the nucleus. Cytoskeletal association with AIF (previously shown to be disrupted by calcium-induced apoptosis) was also prevented through PAD inhibition. Finally, studies of a novel PAD inhibitor showed that it is more potent than Cl-amidine. Together these findings indicate that PADs mediate neural cell survival/death and identify PAD3 as a new key upstream regulator of calcium-induced apoptosis in human neural cells.

Errors that occur during linkage of individual-level data from different sources can lead to substantial bias in analyses of linked data. This thesis aims to develop methods for handling linkage error, and to evaluate these methods in the context of using linked administrative data to support randomised controlled trials. Firstly, the thesis describes the process required for linkage of national data on paediatric intensive care unit (PICU) admissions and bloodstream infection (BSI) surveillance (PICANet and LabBase2). I illustrate the complex steps required, from understanding the structure of the data to calculation of probabilistic match weights and evaluation of linkage quality. This provides a generalisable guide for linkage of administrative data in other contexts. Secondly, the thesis develops methods for handling uncertainty in linkage by extending the multiple imputation framework to the context of data linkage, using prior-informed imputation (PII). Comparison of results from traditional probabilistic linkage, standard multiple imputation and PII showed that PII minimised the bias associated with linkage of incomplete identifiers in PICANet and LabBase2. Finally, linked PICANet-LabBase2 data are used to assess the generalisability of results from a trial of standard versus impregnated central venous catheters (CVCs) in PICU. Trial results are not yet available. However assuming a relative risk of 0.06-0.44 for BSI using impregnated CVCs (based on a meta-analysis in adults), an estimated 163-311 BSI could be avoided in 2014 by using impregnated CVCs for all children in PICU. This thesis highlights the need to assess the impact of linkage error on results and demonstrates the importance of using alternative statistical methods such as PII for handling linkage error within analysis. This work addresses the challenges of exploiting administrative data for research and illustrates the value of linking these data to answer research questions that would otherwise not have been possible.

Introduction: This project focuses on the biological characterisation of 4 important structural causes of severe childhood epilepsy: Focal Cortical Dysplasia (FCD) Type IIb and IIa, and two brain lesions present in Tuberous Sclerosis: cortical tubers and subependymal giant cell astrocytoma (SEGA) tumours. Within these lesions several abnormal cell populations are present, including: Balloon cells, Dysmorphic neurons and SEGA cells. These may represent an aberrant neural progenitor/stem cell population. This raises the hypothesis that a pathological stem cell contributes to the pathogenesis of these diseases. The aim of this study is to identify molecular networks responsible and/or implicated within these lesions. Methods: Differentially expressed genes/exons and microRNAs were identified using the Affymetrix human exon ST1 microarray and NanoString platform respectively. This was performed across the FCD subtypes, cortical tubers, SEGAs and normally formed cortex. WGCNA and IPA network analyses were then applied. Selected proteins based on these results, were validated and investigated by immunohistochemistry of surgical material. Results: Hierarchal clustering of the gene expression between samples was able to reclassify the diseases based on the transcriptome rather than diagnostic subtype. Following a systems biology approach, novel functional networks were identified. A micro-network of: Tenascin C, Palladin, Chitinase-3-like 1, and paired-related-homeobox 1 within the FCD IIb lesion was validated at the protein level, with extracellular-signal-regulated kinase (ERK) as its hub. Additionally, a highly robust subset of alternatively spliced exons specific to the BC harbouring samples were identified. 75.9% of genes identified as differentially expressed between FCD IIa and FCD IIb, were potentially explicable due to the microRNA dysregulation identified. Conclusion: This is the first published gene, exon and microRNA high-throughput analysis performed concurrently for FCD IIb, FCD IIa, cortical tubers and SEGA tumours and the first use of network analysis in these diseases, potentially leading to new therapeutic targets.

Background: Enamel is an external layer of the crown, and its production can be affected by genetic, systemic or environmental causes, leading to the formation of developmental defect of Enamel (DDE). Molar incisor Hypomineralisation (MIH) can be defined as a qualitative defect of systematic origin of the enamel, involving one or more first permanent molar, which is frequently associated with affected incisors. Amelogenesis Imperfecta (AI) is genetic condition, affecting the structure and clinical appearance of the enamel of all or nearly all the teeth in a more or less equal manner, and which may be associated with morphologic or biological changes elsewhere in the body. Knowledge about the chemical and mechanical properties of both conditions is beneficial to predict the severity and to provide appropriate treatment for individual patients. Aim and Objectives: The aim of this study is to characterise the phenotypic properties of MIH and AI teeth in primary and permanent teeth. The objectives were to assess: tooth colour, radiographic features, hardness (on and off the anomaly), chemical variation and the ultrastructure of affected enamel vs. normal enamel. Material and Methodology: Ethical approval was obtained. Twelve control, 12 MIH and 7 AI teeth were collected. The phenotype (DDE Index) for each sample was recorded. Prior to characterisation, the teeth were debrided and stored in 0.1% thymol at 4°C. Colour was examined using a spectrophotometer (SpectroShadeTM Micro) to quantify the variation in colour as defined by ÄE. Wallace indentation (H.M Wallace, Croydon, England) was also performed to assess the hardness value, denoted by Vickers Hardness Number (VHN). Raman microspectroscopy (HR 800, Jobin Yvon, Horiba, Japan) was used to study the chemical variation for each sample. Finally, the samples were sectioned at specific affected sites to image the ultrastructure of the enamel layer using a scanning electron microscopy (FEI XL30 FEGSEM (FEI UK, UK). Results: All MIH and AI teeth showed various degrees of discolouration, with ÄE ranging from 2.47 for white/cream to 22.2 for yellow/brown defects. The range of enamel hardness of control teeth was 226.2 - 360.3 VHN. Meanwhile, enamel hardness for MIH and AI teeth was significantly reduced (P = 0.004) ranging from 13.1 - 142.6 VHN. Chemically, both MIH and AI teeth showed a significant increase in carbonate to phosphate ratio when studied under Raman spectroscopy (P = 0.000). Histologically, the affected hypomineralised enamel appeared more porous, disorganised, with loss of enamel rod structure and presence of structureless layer compared the normal enamel. Conclusion: Teeth diagnosed with MIH and AI showed marked discolouration and lower hardness values compared with normal enamel. Yellow/brown opacities had lower hardness values than white/cream opacities. These conditions also caused changes in the chemical properties and histological structure of the enamel.

Impaired mitochondrial function is associated with a range of neurodegenerative disorders, making mitochondria important therapeutic targets. The ketogenic diet (KD) is often used in the treatment of pharmacoresistant epilepsy, but the underlying mechanisms responsible for its efficacy are yet to be elucidated. In this thesis, the role medium-chain fatty acids (MCFAs) have in contributing to the efficacy of the KD has been investigated. Mitochondrial enrichment and function were assessed after treatment of the human neuronal model SH-SY5Y cells with MCFAs. In contrast to treating cells for 6 days with either 250μM octanoic acid (C8) or dodecanoic acid (C12), 250μM decanoic acid (C10)-treated cells were found to exhibit a 30% increase in activity (p < 0.01) of the mitochondrial marker enzyme citrate synthase (CS), a 42% increase in respiratory chain complex I activity (p < 0.01), and a 31% increase in medium-chain acyl-CoA dehydrogenase activity (p < 0.05). Electron microscopy images provided further support for C10-induced mitochondrial biogenesis. Cellular antioxidant status appeared uncompromised by C10 treatment, as seen through measurements of cellular reduced glutathione. Peroxisome proliferator-activated receptor γ (PPARγ) has previously been shown to be partially activated by C10, and the findings presented in this thesis may implicate PPARγ activation as a potential mechanism for C10-induced mitochondrial biogenesis: Catalase, a peroxisomal marker enzyme, showed a 15% upregulation in activity (p < 0.05) following C10 treatment. Furthermore, BADGE, a PPARγ antagonist, was shown to diminish the C10-induced increases in both CS and catalase activity. RNA microarray analysis of SH-SY5Y cell gene expression, post-C10 treatment, found a number of changes as a result of treatment. The ability of C10 to downregulate many genes involved in cholesterol biosynthesis is of particular interest. The findings in this thesis may help shed some light on the unresolved mechanism through which the KD is able to induce mitochondrial biogenesis. Further work and understanding of C10 could aid the development of a clearer, less laborious, therapy to help replace the KD.

Diabetes mellitus (DM) is a common chronic disorder in children. Type1 DM is the most common form of DM and is due to the loss of insulin secretion by autoimmune destruction of pancreatic β-cells. Type2 DM is due to insulin resistance. Maturity Onset Diabetes of the Young is a heterogeneous group of monogenic disorders due to single gene defects. DM and sensorineural deafness (SND) are associated with rare syndromes such as Wolfram syndrome (WS), Rogers syndromes and Mitochondrial DM. WS is an autosomal recessive disorder characterised by diabetes insipidus, DM, optical atrophy and SND, caused by mutations in the WFS1 gene. Rogers syndrome is associated with DM, anaemia, SND and is due to mutations in SLC19A2. Mitochondrial DM (MT-TL1) is associated with SND in more than 60% of cases. The aims of this study were to exclude these known genetic causes of DM and SND and to identify novel genetic mechanisms. PCR, Sanger sequencing, homozygosity mapping (HZM) and exome sequencing was performed on the 5 families’ for the affected and unaffected siblings to identify novel variants. Immunohistochemistry was used to see whether the candidate protein was expressed in the pancreatic and inner ear tissues. This study has identified 3 novel and 1 known mutations in the WFS1 gene. This work expands the molecular spectrum of the WFS1 mutations in WS, with three novel mutations in three unrelated consanguineous families. However, WS should be considered in any patients with DM and SND. The study also identified two novel genes; NOX1 and SSH2, with the former expressed in both the inner ear and the pancreas and the latter only in the ear. This would suggest NOX1 as a potential new gene involved in DM and SND and SSH2 as being one of two mutations that are causing the DM and SND syndrome.

Gaucher disease (GD), a recessive disorder characterised by hepatosplenomegaly, pancytopenia and skeletal complications, is caused by deficiency of the enzyme glucocerebrosidase (GC). GD leads to the accumulation of glucocerebrosides within macrophages, particularly in the liver and spleen. Current treatment is limited to enzyme replacement therapy (ERT) which is effective for most symptoms however skeletal problems are slow to respond. Treatment also has significant cost and impact on quality of life as infusions must be administered every two weeks. GD is a candidate for gene therapy as bone marrow transplantation has been shown to be curative which serves as a proof-of-concept that correction of haematopoietic stem cells (HSCs) can alleviate disease. This project produced lentiviral vectors carrying a range of constructs. GC was modified to contain a protein transduction domain (PTD) which could facilitate cross-correction of untreated cells in vivo. Recombinant vectors carrying PTD-GBA cDNA corrected the metabolic defect in patient-derived fibroblasts with levels of enzyme activity restored to within the healthy range. Transduced cells secreted active protein, uptake of which by untransduced cells was mediated by fusion of a PTD to the C- but not the N-terminus of the enzyme. The skeletal complications of GD are likely to be caused by enzyme deficiency in the osteoclast, a cell of haematopoietic origin. Therefore it is possible that by transducing HSCs we will be able to alleviate skeletal symptoms. To this end it is shown that modification of HSCs does not affect their ability to generate osteoclasts. It is also demonstrated that osteoclasts derived in vitro from the neuronopathic GD mouse model have increased activity and this could be a useful model for osteoclast correction when treating GD. In conclusion, this project generated lentiviral vectors for use in treating Gaucher disease. Further work should include correction of the osteoclast phenotype and further investigation of the potential for cross-correction in vivo.

Cardiovascular disease (CVD) is a leading cause of mortality in adults with systemic lupus erythematosus (SLE). As children with conditions such as chronic kidney disease have been shown to have a risk of CVD similar to their adult counterparts, clinicians have become concerned that paediatric patients with SLE are at increased risk of CVD. The aim of this project was to examine the vascular phenotype of a British paediatric population with SLE and gain mechanistic insights into this process. Structural changes in vessels were measured using carotid intima media thickness (cIMT) and function was assessed using pulse wave velocity (PWV). These findings were compared with clinical information, traditional cardiovascular risk factors, disease activity, medications and adipokine activity. 45 children were recruited to the study and compared to historical controls previously studied in our centre. Children with SLE had higher cIMT than controls (0.45 V 0.37mm, p<0.0001) but no difference in PWV (5.27 v 5.34m/s, p=0.77). The increase in cIMT is most marked in patients with hypertension, those on higher doses of prednisolone and those of Afro-Caribbean descent. No significant association was found between increased cIMT and biopsy-proven nephritis, disease activity, age, family history of CVD or physical activity score. Patients with JSLE had increased serum leptin levels (15.5 V 7.56ng/ml, p=0.024). There were slightly higher adiponectin levels in patients than controls (14.2 V 12.4ug/L, p=0.49). Patients with proteinuria had higher leptin (35.5 V 18.8ng/ml) and adiponectin (21.9 V 10.5ug/ml, p=0.03) levels, and cIMT increased with leptin and adiponectin levels. This cohort of children with JSLE show structural changes in their vessels indicative of early CVD but with adaptive changes resulting in normal functional scans. In contrast to adult data this group displays an early increase in serum adiponectin suggesting a possible early protective mechanism, which may be overwhelmed in later disease.

Fabry disease (FD) is an X-linked lysosomal storage disorder caused by a deficiency of the enzyme alpha-galactosidase A. The resultant progressive intracellular accumulation of the glycosphingolipids globotriaosylceramide (Gb3) and globotriaosylsphingosine (lyso-Gb3) are the source of a variety of clinical consequences. Biomarkers capable of detecting FD at an early stage of disease progression and that enable monitoring of response to treatment are required urgently. Using label-free quantitative proteomics two potential biomarkers, proactivator polypeptide and ganglioside GM2 activator protein, were identified in the urine of paediatric FD patients. An ultra-performance liquid chromatography-tandem mass spectrometry assay was then developed for validation purposes. Subsequently a second, larger multiplexed assay was developed to identify biomarkers capable of detecting and monitoring pre-symptomatic kidney disease in those patients most at risk. A complementary metabolomic approach was also used to identify and evaluate new Gb3-related biomarkers in the plasma of FD patients. This aspect of the study revealed five novel Gb3-related biomarkers as well as existing Gb3-related analogues and isoforms providing a metabolic profile in these patients. Potential disease mechanisms involved in FD were investigated by studying the interactions between proteins and those molecules involved in the glycosphingolipid pathway, with particular interest to Gb3. A number of mitochondrial proteins were found to interact with Gb3 resulting in the investigation of the effect of glycosphingolipids and their deacylated counterparts on ATP synthase activity. Increasing concentrations of Gb3 and lyso-Gb3 were shown to increase ATP synthase activity. Subsequently the activities of the enzymes preceding ATP synthase, the mitochondrial respiratory chain enzymes, were investigated in a Fabry mouse model. In this aspect of the study complex II/III activity was found to be significantly decreased in kidney tissues. Finally, assessment of pain thresholds in mice has demonstrated significant increases in sensitivity to applied mechanical stimuli following exposure to Gb3 and lyso-Gb3.