The aim of this study was to characterise the genetic and epigenetic profiles of rare forms of sporadic renal cancers (RCC) and identify differential patterns of DNA methylation or somatic mutations that may permit distinction between different subtypes of RCC and could facilitate disease prognosis or identify molecular pathways that could be targeted therapeutically. Illumina Infinium HumanMethylation 450K BeadChip permitted the comparison of the epigenome of the malignant chromophobe RCC and the benign renal oncocytoma. This study identified several genes to be differentially hypermethylated in chromophobe RCC, and renal oncocytoma showing that although both visually and pathologically similar, both tumours have a distinct methylation pattern. Whole exome sequencing (WES) of renal oncocytoma samples identified somatic mutations in eighteen genes involved in a variety of cellular functions. Sanger sequencing was then used to confirm the mutations identified, followed by further screening by Sanger in a cohort of additional renal oncocytoma samples to identify if the somatic mutations are recurrent. Modern high throughput and quantitative techniques have permitted further characterisation of these rare renal cancers and have enabled unique insights into their molecular genetics, findings that may hopefully be of clinical benefit in the future.

Chronic lymphocytic leukaemia (CLL) is characterised by marked disease heterogeneity and is currently incurable. This thesis presents work undertaken to discover novel biological and therapeutic insights through the investigation of spontaneous CLL regression, the evaluation of ATR inhibition as a therapeutic strategy, and the assessment of the impact of post-treatment minimal residual disease (MRD) in CLL. Firstly, spontaneous regressed CLL tumours were found to express somatically mutated IGHV genes, display unresponsiveness to IgM and IgD BCR stimulation and exhibit a phenotype of short telomeres with low CD49d expression, suggesting a model in which the CLL clone undergoes an initial period of proliferation which subsequently subside into a state of anergy and low proliferation. Secondly, ATR inhibition was found to be a promising therapeutic target for CLL tumours with TP53 or ATM defects. Treatment with ATR inhibitor induced synthetic lethality and selective cytotoxicity to these tumours in vitro and in vivo, and sensitised them to chemotherapy and ibrutinib. Finally, MRD negativity was found to predict for 10-year survival in CLL independent of the type and line of treatment, as well as known prognostic factors including adverse cytogenetics, supporting its use as a prognostic marker and potential therapeutic goal in CLL.

Hypermethylation of CpG islands is one of the many processes that a developing cancer cell may use for the inactivation of tumour suppressor genes. The Sav/Hippo/Warts pathway was originally identified in Drosophila and shown to be responsible for controlling both growth and apoptosis, implying this is a tumour suppressor pathway. This pathway is both evolutionarily and functionally conserved in mammals. Work presented here shows that apart from FAT1 and YAP other pathway members are not epigenetically silenced in common epithelial or haematological cancers. FAT1 and YAP were frequently methylated in childhood acute lymphoblastic leukaemia (ALL) but unmethylated in epithelial cancers. Childhood ALL is a blood cancer with peak prevalence between the ages of 3-5 years. The epigenetics of this cancer were examined with three separate approaches; the first, a candidate gene approach, second a NotI restriction enzyme based array examining the methylation of genes residing on chromosome 3, and thirdly the methylated-CpG island recovery assay (MIRA) combined with CpG island arrays examining methylation on a genome-wide scale. These approaches identified a large number of novel genes which were frequently methylated in ALL. Many of the identified genes were new methylation targets and were shown to be likely targets for methylation in both common epithelial and haematological cancers. A series of these genes was seen to be specifically methylated in different leukaemia sub types, and to cluster T-ALL and B-ALL samples into high and low methylation clusters. When examined in chronic lymphoblastic leukaemia (CLL) methylation of two of the above genes was associated with disease progression and methylation of another gene was associated with response to clinical treatment.

Iron is central to the aetiology of gastrointestinal disease. Specifically, the toxic effects of excess, unabsorbed "luminal" iron ingested from the diet has been shown to be important in the development of inflammatory bowel disease and intestinal cancer. A platform for therapeutic intervention is likely to involve chelation of this luminal pool of iron. As such, a range of dietary iron chelators have been tested for their iron binding capacity. Natural biopolymers extracted from seaweed (alginates) and a variety of natural polyphenolic compounds were stratified in terms of their iron binding potential. One alginate, Manucol LD, was unique in its iron binding and demonstrated luminal iron chelation properties. With respect to the polyphenols, only one of the tested compounds (quercetin) displayed iron chelation activity in vitro and was able to suppress cellular concentrations of reactive oxygen species acting as an antioxidant. As such, it has been demonstrated that a unique alginate, Manucol LD, is an excellent candidate for sequestering luminal iron present in the gastrointestinal tract. These results underpin the rationale in utilising these types of natural and safe bio-polymers for the prevention and treatment of gastrointestinal disease.

Head and neck squamous cell carcinoma (HNSCC) is the 6th most common cancer worldwide and poses a significant health burden due to its rising incidence. The proto-oncogene PTTG is overexpressed in HNSCC and correlates with poor patient prognosis. A recent unpublished GEO profile eDNA array analysis has further suggested a potential upregulation of its binding partner PBF in HNSCC. PTTG and PBF cause transformation in vitro and tumour formation in vivo, both effects thought to be partly mediated by their interactions with the tumour suppressor protein p53. Dysregulation of the p53 pathway is frequently observed in HNSCC, thus alluding to the importance of functionally active p53 in the suppression of HNSCC initiation and progression. The work presented in this thesis describes the functional relationship between PTTG, PBF and p53 in HNSCC. Initial studies confirmed that PTTG and PBF are overexpressed in HNSCC tumours compared to matched normal tissue. In addition, high tumoural PTTG expression correlated with HPV status, whereas high tumoural PBF expression was associated with a significant gender bias. Further investigations established that PTTG and PBF functionally interact with p53 and cooperate to reduce p53 protein stability in HNSCC cells. Moreover, attenuation of PTTG or PBF expression led to dysregulated expression of p53-related genes involved in DNA repair and apoptosis, indicating that both proto-oncogenes may serve to promote genomic instability and HNSCC cell survival. Functionally, depletion of PTTG or PBF significantly repressed cellular migration and invasion, and impaired colony formation in HNSCC cells. Overall, this research has provided novel insights into the roles of PTTG and PBF in HNSCC tumour initiation and progression, through modulation of p53 activity and function.

Hypoxia has emerged as a crucial part of the aetiology of tumours. It is a negative prognostic factor which is associated to chemoresistance, invasiveness and metastasis. There is a strong association between hypoxia and metabolic transformation in breast cancer due to the alterations of multiple metabolic pathways. However, the current understanding of the nature of metabolic alterations in hypoxia is insufficient. This thesis uses NMR as a tool to investigate both the static metabolome by measuring metabolite concentrations, as well as to determine \(^1\)\(^3\)C metabolic fluxes using stable isotope tracers to reveal metabolic pathway alterations by hypoxia in vitro and by tumour growth in vivo. Firstly, we developed the \(^1\)\(^3\)C isotopomer distribution (CID) analysis to quantify metabolic fluxes by following the evolution of specific isotopomers of specific pathways of interests. MCF7 breast cancer cells were analysed in hypoxia using an integrated approach using gene expression, steady-state metabolite levels and \(^1\)\(^3\)C metabolic flux analysis to pinpoint hypoxia induced metabolic alterations. These most significant alterations were an up-regulation of the pentose phosphate pathway and a down-regulation of mitochondrial oxidative metabolism by lowering the PDH flux. The latter was partially compensated by carbon entry into the mitochondria by increasing flux through pyruvate carboxylase (PC). Further attention was focused towards identifying the shifts in metabolic activity in PC altered cells using [1,2-\(^1\)\(^3\)C]glucose and [3-\(^1\)\(^3\)C]glutamine as precursor nutrients correlated to cellular transformation potential accessed by cell viability. Finally, the \(^1\)\(^3\)C labelled glucose strategy was applied to a cancer model in mice model by infusing mice with [1,2-\(^1\)\(^3\)C]glucose. \(^1\)\(^3\)C glucose administration protocol was optimised in order to enable an investigation of \(^1\)\(^3\)C metabolic fluxes in tumour tissue to identify metabolic pathway differences between earlier stage and advanced stage of mammary gland tumours. In conclusion, an NMR based metabolomics analysis is suitable for discovering metabolic pathway alterations using both in vitro and in vivo models.

Knowledge of the phenotypic states of immune cells in the tumor microenvironment is essential to understand immunological mechanisms of cancer progression, responses to cancer immunotherapy, and the development of novel rational treatments. Yet, this knowledge is opaque to traditional bulk sequencing methods, and novel single-cell RNA sequencing (scRNA-seq) methods which could potentially address these questions introduce complex patterns of error into data that are poorly characterized. This dissertation describes a computational framework, SEQC, built to facilitate rapid and agile analysis of scRNA-seq approaches that utilize molecular barcodes. It combines SEQC with a clustering and normalization method, BISCUIT, and approaches to examine phenotypic diversity and gene variation. These methods are applied to address the unique computational challenges inherent to analysis of single-cell RNA-seq data derived from multiple patients with diverse phenotypes. This dissertation describes an experiment comprising scRNA-seq of over 47,000 immune cells collected from primary breast carcinomas, matched normal breast tissue, peripheral blood, and using these computational approaches. This atlas revealed significant similarity between normal and tumor tissue resident immune cells. However, it also describes continuous tumor-specific phenotypic expansions driven by distinct environmental cues. These results argue against discrete activation states in T cells and the polarization model of macrophage activation in cancer, and have important implications for characterizing tumor-infiltrating immune cells.

No description available.

Automated computer-aided systems and approaches are widely required to investigate and analyze histology images for improving the accuracy of cancer diagnosis and effective treatment decision making. Quantitative analysis has immense potential to investigate and analyze the tissue and cellular characteristics of histology images in cancer research. It is based on accurate cellular, morphological, and tissue features. Automated approaches not only make the feature extraction and analysis more objective and more reproducible, but they can also help pathologists look for useful potential clues from a vast amount of hidden information in cancer tissues, whose clinical value may not be fully realized and visualized. This entails the automated computer algorithms with a key role of quantitative analysis of histology images for different cancers. In this thesis, I concentrate on bone marrow cancers and develop automated computer algorithms to extract and realize cellular and texture characteristics of bone marrow biopsies for efficiently characterizing different types of bone marrow cancers in further investigation and analysis. We focus on the development of automated algorithms for identifying various types of cells in bone marrow trephine biopsies, which are tiny cores of bone marrow tissues. All the algorithms are specifically designed for histological sections stained by a standard hematoxylin and eosin (H&E) stain. Firstly, we propose an automated framework with a novel segmentation model for delineating and segmenting megakaryocytes. Secondly, we create a novel deep learning network that processes the nuclear detection with irregular shape for various types of bone marrow stem cells. Then we construct another synchronized deep learning approach to simultaneously do detection and classification. We demonstrate the effectiveness of the network of detection and classification at same time and the training time consumed in this synchronized network.

Multiple myeloma is a cancer of plasma cells. In multiple myeloma, a clone of plasma cells in the bone marrow secretes a unique, monoclonal immunoglobulin (Ig), whose biological properties depend on its type and structure. The monoclonal Ig offers a convenient opportunity for clinicians to monitor the response of the tumour to therapy via the secreted protein, which is readily quantified in a blood sample. Responses to treatment are assigned based on the percentage reduction in monoclonal Ig; however, response criteria do not take into account the different metabolic half-lives of the proteins. 70% of multiple myeloma patients have either monoclonal IgA- or monoclonal IgG-producing clones. IgA and IgG have metabolic half-lives of 6 days and 23 days, at normal concentrations, respectively. The large difference in their metabolic half-lives suggests that they would respond at different rates during therapy. The elimination rate of IgG is concentration-dependent due to saturable recycling by a receptor. This could further impact upon its response during therapy, with the possibility that IgG is eliminated from the body at different rates at the beginning of therapy, when its concentration is high, and at the end of therapy, when its concentration has decreased. In this thesis compartmental models of IgG metabolism from the literature are analysed and parameter values are estimated from available data. A model of IgA metabolism is sourced in the literature. These models are used to predict the responses of monoclonal IgA and IgG during therapy. The simulations are able to replicate typical monoclonal IgA and IgG responses seen in a clinical trial of patients with relapsed and refractory multiple myeloma. Importantly, the plasma cell clone is not directly accessible to measurement and therefore not available to validate model-based predictions. However, monoclonal Ig responses are not evaluated by their ability to predict the tumour burden, but by the strength of their association with patient survival. In this thesis, a prediction is made of how the different metabolic properties of IgA and IgG may influence their association with survival outcomes. Evidence for this effect is then evaluated in data from a clinical trial using the methods of survival analysis.