CML is treated effectively with TKI, however two key problems remain - the insensitivity of CM HSC to TKI and the emergence of TKI-resistant BCR-ABL mutations. BCR-ABL activity is associated with increased proteasome activity and PI are cytotoxic against CML cell lines. We demonstrate that bortezomib is antiproliferative and induces apoptosis in CD34+ cells taken at diagnosis from patients with CP CML cells, with an LD50 below concentrations achieved in vivo. We also demonstrate for the first time that CD34+38- CML cells, representing the TKI-insensitive primitive HSC, are similarly susceptible. Bortezomib is associated with inhibition of proteasome activity, however that of BCR-ABL appears unaffected. Significant synergy is seen when bortezomib and dasatinib are used in combination. We also demonstrate that bortezomib is effective in inhibiting proteasome activity and inducing apoptosis in cell lines expressing BCR-ABL mutations, including T315I. Therefore we believe that bortezomib offers a potential therapeutic option in CML by targeting both TKI-insensitive stem cells and TKI-resistant BCR-ABL mutations.

The work was undertaken to examine the effects of nutritional deficiencies on cancer induction. Two of the most common and widely distributed nutrients, iron and folic acid, were examined to evaluate the effects of their deficiency on animals. The Syrian golden hamster was the animal model for all experimental work. In the first part of the study an attempt was made to induce iron deficiency in young adult male hamsters by feeding iron deficient diet coupled with repeated venesection of 1.5ml every two weeks. Following twelve weeks on this regime a superficial biopsy was taken, on week 13, from the medial wall of one pouch in each hamster in order to evaluate the effect of iron depletion on the epithelial compartment thicknesses. After allowing the biopsy sites to heal for two weeks, a solution of 0.25% DMBA in acetone was painted, three times per week, for eight weeks, on a defined one square centimetre area in each pouch of each hamster of the experimental and control groups. The hamsters were then maintained on the same dietary regimes for twelve weeks before being killed at the beginning of week 37 for analysis. Iron deficiency anaemia could not be induced in the experimental animals of this study. The effect of the iron deficient diet on epithelial compartment thicknesses at the stage of the biopsy was not clear. However, restriction of iron intake did cause animals to develop significantly fewer grossly seen tumours and histologically identified carcinomas than control animals. In the second part of this thesis an attempt was made to investigate alternative hamster dietary components that have less iron contamination than the diet given in the first part of this thesis. Casein and calcium lactate were the main contributers to iron in the hamster diet. Casein could not be substituted by another source of protein for hamsters. However, other sources for calcium with less iron contamination were available and therefore investigated in this part of the study. Three groups of young adult male and female hamsters were given the fully nourishing powdered diet used in previous studies. However, calcium lactate was substituted for by either calcium acetate, calcium chloride or calcium sulphate in each group. None of the three diets was accepted by the animals and many of them died of starvation. When calcium salts were replaced by calcium lactate the surviving animals accepted the diet and recovered quickly afterwards. This study proved that calcium lactate could not be substituted by any other calcium salt with less iron content and therefore iron contamination in the hamster diet could not be further reduced by this method. In the third part of this thesis the effect of nutritional folate deficiency on cancer induction was studied. A group of young adult female hamsters was given folate deficient diet for four weeks. On week 5, DMBA in acetone at a concentration of 0.25% was painted on a defined one square centimetre area of the medial wall of each pouch in each hamster in folate deficient and control groups. The carcinogen was applied three times per week for eight weeks following which animals were maintained on the same dietary regimes for a further 13 weeks before being killed at the beginning of week 27 for the final analysis of the study. It was found that nutritional folate deficiency had significantly reduced the number of animals developing grossly counted tumours and histologically identified carcinomas. The folate deficient animals also developed significantly less tumours and carcinomas compared to control groups. In the last part of this thesis, the effect of combined iron and folate deficiency was examined for its role in carcinogenesis of the hamster cheek pouch. Two groups of young adult male hamsters were fed powdered diet lacking iron and folic acid and a third group was fed diet lacking iron only. One of the combined deficiency groups and the iron deficiency group were bled 1.0-1.3ml every week. On week 6 of the study DMBA in acetone at a concentration of 0.25% was painted three times per week for eight weeks on the same area of the pouch used in the previous studies. The animals were then maintained on the same experimental regimes for a further eleven weeks before being sacrificed, on week 25, for the final analysis of the study. In this study, iron deficiency anaemia was induced in animals of the bleeding groups. Animals in the group with combined iron and folate deficiency without bleeding showed low normal folate levels and normal haemoglobin levels. The two groups that were bled repeatedly showed iron deficiency anaemia. In all groups, the numbers of tumours counted grossly and the numbers of carcinomas identified histologically were significantly reduced compared to control animals in the previous studies. The folate deficient diet did not appear to influence the induction of iron deficiency. The studies reported in this thesis proved that nutritional folate deficiency not only reduces the incidence, but it also reduces the numbers of tumours and carcinomas in the hamster cheek pouch. Iron deficiency anaemia was also found to significantly reduce the numbers of tumours and carcinomas of the hamster cheek pouch. It was not possible to produce combined iron and folate deficiency under the conditions of these studies. However, animals fed on a diet lacking iron and folic acid had significantly reduced numbers of grossly seen tumours and histologically identified carcinomas in the cheek pouch in response to DMBA applications. In each of the reported studies, the nutritional deficiency of iron and folic acid, whether individually or combined was found to significantly reduce the growth rate of affected animals.

Chronic myeloid leukaemia (CML) is a haematological malignancy that is identified by the presence of a fusion oncogene, BCR-ABL1, which is a constitutive tyrosine kinase. The discovery of tyrosine kinase inhibitors (TKIs) over that past decade has resulted in significantly improved survival rates and disease management in CML patients. However, a subpopulation of BCR-ABL1+ cells in the niche are found which exhibit stem cell-like features, such as self-renewal and quiescence. These CML stem cells (LSCs) are shown to be insensitive to TKI treatment and are capable of deriving the disease during the relapse. Consequently, the elimination of LSCs is a primary goal of current research. Therefore, the aim of this thesis was to obtain a global view of the cellular processes that maintain stem cell identity in CD34+ CD38- LSCs as well as identify those processes which initiate the transition to proliferative CD34+ CD38+ CML progenitor cells (LPCs). A combined approach was exploited to investigate genome-wide gene expression profiles and histone modification signatures of normal HSCs and committed progenitors (HPCs), and their LSC and LPC counterparts. Despite having increased activity in pathways involved in cell division and proliferation, expression levels of the pathways involved in stem cell identity were not significantly different in LSCs to those found in HSCs. These pathways included Wnt, TGF-β signalling, and several novel neurotransmitter signalling pathways. By examining genome-wide histone modification patterns using ChIP-sequencing it was shown that the stem cell identities of HSCs and LSCs are programmed at the epigenetic level. All of the pathways which confer stem cell identity to both HSCs and LSCs are significantly enriched for bivalent gene promoters having both the H3K4me3 and H3K27me3 marks. These similarities were most evident in neurotransmitter signalling and it was demonstrated that these pathways are capable of promoting LSC maintenance in vitro. Intriguingly, although the stem cell entry into the proliferative state occurs through the repression of many of the same stem cell identity pathways in both HSCs and LSCs, it was shown that epigenetic reprogramming in CML mediates this repression via a different mechanism than in normal HSCs. Furthermore, abnormalities in levels of several chromatin enzymes were identified that are likely to be responsible for the epigenetic reprogramming of CML cells. The work presented in this thesis defines the chromatin landscape of a cancer stem cell for the first time and provides new therapeutic targets for the eradication of TKI resistant CML stem cells.

More than a million women worldwide are diagnosed with breast cancer every year and breast cancer is the most common cancer in women in the UK. Despite improvements in the treatment of breast cancer, it remains the second most common cause of death from cancer in women after lung. Research into the aberrant molecular pathways which characterise neoplasia is necessary to highlight potential therapeutic targets with the aim of improving the survival of women with breast cancer. Rab25 is a member of the Rab superfamily of small GTPases, which are involved in the regulation of intracellular vesicular trafficking. A number of studies have suggested that dysregulation of Rab25 protein expression may be implicated in breast, ovarian and colorectal cancer. However, the published data are conflicting and further research is required to improve the understanding of the role of Rab25 in neoplasia. At the commencement of this project, there was no commercially available anti-Rab25 antibody. The data presented here describe the affinity purification and validation of a rabbit polyclonal anti-Rab25 antibody suitable for use in Western blotting, immunofluorescence and immunohistochemistry. Furthermore, an immunohistochemical scoring system was devised and validated prior to the investigation of Rab25 expression in a large breast cancer cohort. The data presented here show that loss of Rab25 expression correlates with decreased breast cancer related survival of patients who have tumours which have also lost the expression of the oestrogen receptor. Furthermore, loss of Rab25 expression has maximal negative effect on the survival of patients with tumours which have lost both oestrogen and HER2 receptor expression. To investigate the effect of Rab25 knockdown in vitro, the MCF7 cell line was stably transfected with a short hairpin siRNA vector targeted against Rab25. Knockdown of Rab25 had no demonstrable effect on cellular proliferation or colony formation, but Rab25 knockdown cells can migrate to heal a scratch wound at a greater rate than cells transfected with a non-targeted control vector. Knockdown of Rab25 protein expression increases total surface and intracellular β1 integrin protein levels in MCF7 cells, but this increase is not as a result of increased transcription.

p53 is a tumour suppressor that is dysfunctional in most cancers. In some cancers, mutation of the TP53 gene results in expression of a mutant protein or loss of p53 expression, while in others wild-type p53 is retained but there is a defect in the mechanisms that allow for the activation of p53, including, but not limited to, upregulation of p53’s negative regulators (MDM2 and MDMX). Restoration of wild-type p53 function can lead to tumour regression and is therefore an attractive anti-cancer therapy. Several small molecule compounds that activate wild-type p53 have been described, and in this study the functions of two new compounds that stabilise p53 are described. While these drugs may be useful to activate p53-dependent apoptosis or senescence in wild-type p53 tumours, an alternative approach that depends on the use of transient p53 activation to protect normal cells while leaving p53 null or mutant tumour cells vulnerable to cytotoxic drugs is also explored. Finally, the identification of pharmacodynamic biomarkers for p53 stabilisation is described. In chapter 3 a new class of MDM2 inhibitor (MPD compounds) is described. The compounds are capable of stabilising and activating p53 in cells by inhibiting the E3 ligase activity of MDM2 in a mechanism that involves compound binding to the RING-tail of MDM2. Although the MPD compounds have limitations in terms of solubility and potency they have demonstrated a new method of achieving MDM2 inhibition and support design of further RING-tail binding compounds with more favourable chemical properties. In chapter 4 the function of a dual inhibitor of MDM2 and MDMX has been explored (HLI373). This compound is shown to activate p53 in cells and in vivo by interfering with ribosomal biogenesis, causing ribosomal stress and inhibiting MDM2. In addition it is capable of reducing MDMX expression at the promoter level by a mechanism that requires intact MDM2-p53 binding. Further work is required to fully define the mechanism of action of this compound and demonstrate its anticancer activity in xenograft and transgenic mouse models. In chapter 5 a chemoprotective approach for p53 activation, which might be applied to situations where tumours express mutant p53, is explored in cell lines. Low-dose actinomycin D treatment can activate p53 without occult DNA damage via the ribosomal stress pathway, thereby protecting wild-type p53 expressing cells from the cytotoxic effects of paclitaxel. Low-dose actinomycin D may therefore be used to limit chemotherapy-induced toxicity to normal cells while targeting a mutant p53 expressing tumour. In chapter 6 the potential for developing pharmacodynamic biomarkers for MDM2 inhibition in serum, peripheral blood mononuclear cells and hair follicles taken from patients prior to and following chemotherapy was examined. Measurement of serum MIC-1 level showed most promise, although further evaluation of this marker is needed before it could be used as a pharmacodynamic endpoint in a clinical study.

The data presented in this thesis demonstrates that the malignant plasma cells of multiple myeloma are capable of suppressing the activation of T lymphocytes. The myeloma cells prevent activation of T cells from healthy donors by allo-antigen, mitogen and IL-2, mediated by the production of soluble, immuno-suppressive factor. This factor was responsible for inducing cell cycle arrest and failure of the T cells to progress into the autocrine IL-2 autocrine pathway, which is of critical importance in the activation of T cells. To further investigate this interaction an in vitro model system was developed to examine the key stages of T cell activation and homeostasis. Myeloma cells constitutively expressed high levels of TGF-1 mRNA transcripts as detected by RT-PCR which were translated into latent protein and secreted as detected by immunohistochemistry and ELSIA, respectively. The reversal of the immuno-suppression induced by the myeloma cells using the specific TGF-1 antagonist, Latency Associated Peptide, confirmed that TGF-1 is a major factor in myeloma-associated suppression of T lymphocyte activation. It was demonstrated that the myeloma cells prevent the T cells, upon activation, from up-regulating the surface expression of the -chain of the IL-2R thus preventing the formation of the high-affinity receptor. The reduced expression of IL-2R resulted from altered transcription of the -chain gene in response to re-stimulation of primary T cells with IL-2. When signalling events in primary T cells responding to re-stimulation with Il-2 was examined, myeloma cells inhibited the phosphorylation of both STAT3 and STAT5. However, using a novel IL-2-dependnet T cell line (IDBL), which does not require the expression of the high affinity IL-2R for its responses to IL-2, it was shown that these cells are insensitive to the myeloma-derived TGF-, in terms of DNA synthesis and proliferation, despite demonstrating failure of phosphorylation of STAT5. It was demonstrated that phosphorylation of STAT3 was unchanged when IDBL cells were co-cultured with myeloma cell lines.

As resistance to current therapies remains one of the major hurdles to the successful treatment of advanced colorectal cancer, we need to understand the mechanisms by which cancer cells evade therapy-induced cell death. I have investigated whether there is a link between epithelial cell adhesions, and acquired resistance to 5-fluorouracil (5-FU). I compared three pairs of human colorectal 5-FU-sensitive and -resistant cell lines, and investigated whether there was a direct role for E-cadherin and/or the Src family kinase, c- Yes (which is co-amplified with thymidylate synthase) in promoting resistance to 5-FU. I found that while knockdown of c-Yes expression had no effect, disruption of E-cadherin using a blocking antibody caused a reduction in colon cancer cell proliferation and some re-sensitisation to 5-FU. The resistant cells displayed intrinsically higher activities of putative survival pathways, namely the PI3-kinase/Akt and the MEK/MAP kinase pathways, and these were suppressed when E-cadherin function was blocked. Furthermore, the resistant cells displayed a greater dependence on signalling via the PI3- kinase/Akt pathway for their survival. Finally, preliminary experiments established a possible link between the integrity of E-cadherin-mediated cell-cell junctions, signalling through the PI3-kinase/Akt pathway and nuclear localisation of the apoptotic regulatory tumour suppressor protein p53 in modulation of 5-FU-resistance.

Drawbacks associated with anticancer chemotherapeutic cisplatin include tumour drug resistance, non-effectiveness against all tumours and lack of tumour-specificity resulting in severe side-effects (e.g. nausea, hair loss and kidney toxicity). The use of other metals such as transition metals rutheniumandosmium, may address the problems associated with platinum drugs and have received increased interest over the years. In this thesis the biological activity and aqueous solution chemistry of half-sandwichosmium (II) compounds of the type [(arene)OsII(X)(YZ)] n+ is explored. Chelating ligands containing nitrogen or nitrogen and oxygen donor atoms (N, NandN, O-chelatingligands) are investigated. It is shown that the chelating ligand has a large effect on the aqueous reactivity of the complexes. The introduction of functional groups on the chelate allowed for the ‘fine-tuning’ of the aqueous reactivity and nucleobase binding of the complexes. Also the nature of the coordinating arene was found to have an important effect on their biological activity. This could be rationalised by increased hydrophobicity with more extended arenes such as biphenylandtetrahydroanthracene, resulting in increased cellular uptake and increased cytotoxicity. Conjugating cell penetrating peptides to the complexes resulted in improved biological properties and opened a new way for functionalisation of the compounds. Several compounds reported in this thesis exhibit promising activity in the ovarian, colon and lung cancer cell lines and some could overcome cisplatin resistance in ovarian cisplatin resistant cell lines. Initial studies revealed cell death via apoptosis and the possible involvement of mitochondria in the apoptotic pathway. These results point to a novel pathway of activation for these complexes which is advantageous for addressing chemoresistance and effectiveness to oher types of cancers. This work shows that the biological properties of these compounds can be tuned by choice of ligands and also provides initial evidence for a novel pathway of activation.

Protein signalling networks play a key role in cellular function, and their dysregulation is central to many diseases, including cancer. Recent advances in biochemical technology have begun to allow high-throughput, data-driven studies of signalling. In this thesis, we investigate multivariate statistical methods, rooted in sparse graphical models, aimed at probing questions in cancer signalling. First, we propose a Bayesian variable selection method for identifying subsets of proteins that jointly in uence an output of interest, such as drug response. Ancillary biological information is incorporated into inference using informative prior distributions. Prior information is selected and weighted in an automated manner using an empirical Bayes formulation. We present examples of informative pathwayand network-based priors, and illustrate the proposed method on both synthetic and drug response data. Second, we use dynamic Bayesian networks to perform structure learning of context-specific signalling network topology from proteomic time-course data. We exploit a connection between variable selection and network structure learning to efficiently carry out exact inference. Existing biology is incorporated using informative network priors, weighted automatically by an empirical Bayes approach. The overall approach is computationally efficient and essentially free of user-set parameters. We show results from an empirical investigation, comparing the approach to several existing methods, and from an application to breast cancer cell line data. Hypotheses are generated regarding novel signalling links, some of which are validated by independent experiments. Third, we describe a network-based clustering approach for the discovery of cancer subtypes that differ in terms of subtype-specific signalling network structure. Model-based clustering is combined with penalised likelihood estimation of undirected graphical models to allow simultaneous learning of cluster assignments and cluster-specific network structure. Results are shown from an empirical investigation comparing several penalisation regimes, and an application to breast cancer proteomic data.

The value of mild hyperthermia in improving the outcome of radiotherapy and chemotherapy treatments is well established. However, clinical applications are currently restricted to accessible tumours, with the application of controlled hyperthermia in solid tumours deep within the body presenting an unresolved problem. Ultrasound is an attractive heating technique because of its ability to create a focus at depth which can be steered around the tumour volume. However, despite considerable research no clinically usable transducers for deep tumour applications have resulted. In this thesis the underlying principles that govern the characteristics of phased array transducers have been examined. The concept of an idealised phased array has been introduced, and analysis of simulated fields from such arrays has enabled a new set of equations to be defined which relate the geometry of the field to the fundamental array design parameters (including the array diameter, radius of curvature and frequency of operation). Further simulations have examined the impact of secondary array design parameters (such as the individual element size, number density and layout geometry) which modify the field from that of the idealised case. Analysis of these has enabled an upper limit to be placed on the element size within any planar array in order to prevent undesirable changes in the characteristics of the focal region. A fifteen element phased array with a random element distribution has been constructed based on the design principles established in the simulation work. Measurements of the inter-element cross-coupling have been made, demonstrating that acoustic coupling dominated for inter-element pitches of less than 8 mm, while electrical coupling dominated at larger inter-element pitches. The field produced by the array in an acoustic tank has been characterised and compared against simulation predictions, showing good agreement in terms of the geometries of the focal region and the grating lobes. However, a number of differences have also been identified. In particular, the focal region was closer to the surface of the physical transducer in the measured fields compared to the simulation results, and there were numerous small high intensity regions between the surface of the transducer and the focus which were absent from the simulated fields. A sensitivity analysis, using a simulated factorial experiment, has been performed to identify the origin of these differences, with the results indicating that the presence of a secondary vibrational mode within the elements of the array was the principal causative factor. Finally, calculations have been performed which demonstrate the feasibility of manufacturing an array suitable for the application of mild hyperthermia in deep tumours based on the array design scheme presented in this thesis. Potential extensions of the array design have also been described which would improve the behaviour of the array under steering and provide further increase in the focal intensity.