Chemokines are a family of low-molecular weight proteins that mediate their effects through binding with chemokine receptors (a group of seven trans-membrane spanning G-protein coupled receptors). Chemokines are well known for their role in leukocyte trafficking, although they also mediate a range of other physiological functions including cell proliferation, growth and differentiation. Chemokines are integral to the development and functioning of the immune system and are implicated in the pathogenesis of a wide range of diseases including autoimmune disorders, HIV infection and malignancies. Non-Hodgkin’s lymphoma (NHL) is a malignancy of the lymphoid system, which can affect any tissues in the body but has a predilection for the secondary lymphoid organs and the bone marrow. NHL represents a heterogeneous group of disorders with variations in presentation and prognosis. Whilst there have been improvements in overall survival in the last few decades following developments in treatment there is still a significant mortality associated with NHL as well as significant morbidity associated with the treatments required to control the disease or affect a cure. There is increasing evidence that chemokines and their receptors play a role in the pathogenesis of NHL with increased chemokine receptor expression seen in certain subtypes. Chemokines have also been shown to influence prognosis in particular NHL subtypes. This thesis describes studies to examine the relationship between constitutive chemokine receptor expression and NHL subtype with subsequent correlation between chemokine receptor expression and outcome. Results from studies on clinical samples obtained from patients with a range of B-cell NHL subtypes revealed significant differences in chemokine receptor expression between subtypes. Predominantly these differences were noted with the receptors CCR4 and CCR7. CCR4 has previously been examined in relation to T-cell lymphomas and high CCR4 expression has been associated with an adverse prognosis in adult T-cell leukaemia/lymphoma. However, systematic examination of CCR4 in B-cell NHL has not been done before. In this study CCR4 expression was seen in samples of diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma (MCL), MALT lymphoma and Burkitt’s lymphoma. Furthermore high CCR4 expression was associated with a significantly improved survival in patients with DLBCL. CCR7 expression was significantly higher in cases of MCL and chronic lymphocytic lymphoma/small lymphocytic lymphoma (CLL/SLL), this however did not equate to differences in outcome. The high expression of CCR7 in MCL and CLL/SLL does however make this receptor an ideal target for the development of targeted therapies. Further studies in this thesis describe strategies used to target the chemokine system with cytotoxic agents. CCR7 was chosen for targeting experiments due to its restricted expression (naïve T-cells, memory T-cells, dendritic cells and subsets of B-cells) in normal tissues, the fact that it controls homing of leukocytes to lymph nodes, and that is the second most commonly reported chemokine receptor detected in cancers. The first approach for targeting the chemokine system used radio-labelled chemokines to induce cytotoxicity in cells expressing the cognate receptor. Unfortunately, the results from this were not conclusive and further studies are required to explore this approach. The second approach described in this thesis entailed the potential use of oncolytic adenoviral gene therapy. This approach requires the redirection of adenoviral tropism from the coxsackie and adenovirus receptor (CAR) and through CCR7. This was successfully achieved in cell lines engineered to express and those endogenously expressing CCR7 by ‘shrouding’ the virus in CCL19. However, background infection was high and this led to the engineering of an adenoviral vector incorporating a peptide within the viral knob protein to allow for more efficient ‘shrouding’ of the virus and thus more efficient redirection of the virus. However, final experiments with this vector were not performed due to time constraints. The final part of this thesis relates to the characterisation of EL4 cells (a murine T-cell lymphoma cell line) for use as a potential in-vivo mouse model for targeting experiments. Following injection into the flank of C57/BL6 with these cells, measurable tumours develop. It was demonstrated that this cell line expressed CCR7, and finally after a number of modifications, it was shown that adenoviral tropism could be redirected through chemokine receptors on the surface of these cells. In summary the results in this thesis shows for the first time that CCR4 is expressed across a number of B-cell NHL subtypes and that this is associated with outcome in DLBCL. Results also demonstrate that therapeutic targeting of CCR7 is a reasonable therapeutic approach in NHL. In addition, redirection of oncolytic adenoviral vectors through chemokine receptors provides a potential strategy for obtaining this targeted therapy. Finally, EL4 cells have the potential to be used as an in-vivo mouse model for CCR7 targeted therapies.

Autophagy is an evolutionarily conserved process that is important for the maintenance of cellular homeostasis and also genomic integrity. Autophagy is a self-digestive process that takes place in the cytoplasm, however recent studies by Eileen White and colleagues demonstrated that defective autophagy leads to accumulation of DNA damage in vitro and in vivo [1] [2]. This discovery leads to a question: whether there is increased DNA damage incidences or defective DNA damage repair in autophagy deficient cells. Autophagy and DNA damage are two important areas for cancer research. The aim of this project is to provide a better understanding of the role of autophagy plays in DNA damage and DNA damage response. The activation of Chk1 facilitates its degradation. The results presented in this project illustrate that autophagy deficient cells exhibit elevated proteasomal activities and autophagy inhibition leads to activation of Chk1. These combined factors contribute to increased degradation of Chk1 in autophagy deficient cells. This was manifested first as decreased phospho-Chk1 in response to DNA damage, later on when the loss of autophagy effect is more pronounced; decrease in total Chk1 protein level was observed. Chk1 is a crucial DNA damage response mediator that plays roles in cell cycle checkpoints and DNA damage repair. Cells without autophagy appear to have intact cell cycle checkpoints in response to starvation or DNA damaging agents; however they show deficiency in homologous recombination (HR) repair pathways. Autophagy deficient cells display increased spontaneous cell death and formation of micronuclei. Defective HR pathways in autophagy deficient cells lead to hyper-dependency on non-homologous end-joining (NHEJ) process. Since HR and NHEJ are the two main ways of repairing double strand breaks (DSB), it is not surprising that inhibition of NHEJ following DSB inducing agents in autophagy deficient cells results in persistence of damage lesions and increased cell death. 3 This project demonstrated that loss or inhibition of autophagy leads to defective DNA damage response pathways. We established that Chk1 is de-regulated in autophagy deficient cells and this has differential downstream effects on DNA damage response. These findings potentially provide a novel synthetic lethal strategy to selectively kill autophagy-deficient cells, which are implicated in a number of diseases including certain cancers.

E-cadherin is a cell adhesion protein required for epithelial tissue integrity. In many cancer cells mis-regulation of E-cadherin adhesions causes increased progression and invasion of cancer. Alteration in E-cadherin dynamics could therefore serve as an early molecular biomarker of metastasis. In this project, I used E-cadherin FRAP to asses real time dynamics of cadherin junctions in a pancreatic cancer mice model of in a variety of micro-environments. My data showed that p53 mutation drives metastasis through mobilizing E-cadherin in junctions. Also, I used FRAP as a pharmaco-dynamic marker to assess the effect of an anti-invasive drug (dasatinib) in pancreatic tumours in vivo. Moreover, my E-cadherin FRAP data along with cross-linking experiments and disruption of E-cadherin interactions by mutation provided a comprehensive framework for understanding E-cadherin dynamics at cell-cell. Here, I have identified four distinct populations of E-cadherin within regions of cell-cell contact and characterized the interactions governing their mobility using FRAP. These pancreatic cancer cells had the immobile fraction (Fi) of E-cadherin-GFP comprised adhesive and non-adhesive populations. The remaining mobile fraction (Fm) also comprised of non-adhesive and adhesive populations, one population moves at the rate of pure diffusion, and therefore represents free E-cadherin monomers. The other population moves more slowly, and represents E-cadherin monomers turning over within immobile complexes. Inclusion of E-cadherin into either adhesive population requires cis-, trans-, and actin interactions. The signaling pathways in cells dramatically affect the fractions of these cadherin components. I showed that understanding the dynamics of these four populations of E-cadherins could be used to design or interpretation of future pharmacological and genetic experiments to probe the function of E-cadherin in development, disease progression, and response to therapy.

Inflammation has been shown to play an integral role in a number of different disease processes. There is evidence that measurements of the systemic immune/inflammatory response are associated with mortality in patients with certain malignancies, as well as those with atherosclerotic disease. A variety of readily available inflammation-based prognostic scores, combining serum constituents of the systemic immune/inflammatory response, have been suggested for use in these patient groups. However, it is unclear whether systemic inflammation-based scores are universally associated with mortality across all tumour types and whether this relationship persists in atherosclerotic conditions including cardiovascular and cerebrovascular disease. The optimal constituents of such a score are also still to be determined. The present thesis further examines these topics with specific reference to: 1. The relationship between the presence of cancer, an inflammation-based prognostic score and biochemical parameters in a large group of unselected patients. 2. The relationship between an inflammation-based prognostic score and cancer survival in patients with malignancies at a number of different sites. 3. A comparison of commonly available inflammation-based prognostic scores in patients with cancer. 4. The prognostic value of a derived neutrophil to lymphocyte ratio, commonly available in patients with cancer undergoing chemotherapy. 5. The optimal serum constituents of the systemic inflammation-based Glasgow Prognostic Score (optimised) in patients with cancer. 6. The optimal constituents and prognostic value of a systemic inflammation-based score for predicting cardiovascular, cerebrovascular and all-cause mortality.

Breast cancer is the most common female cancer in the UK and, despite earlier detection and improved treatments, remains the second most common cause of cancer death in women. Although therapies exist for breast cancer, including endocrine therapy for oestrogen receptor (ER) positive tumours, resistance to current treatment remains a major problem. The molecular mechanisms of endocrine resistance have yet to be fully elucidated and in order to improve treatment for patients this needs to be addressed. Clinically breast cancer presents as several distinct diseases with different outcomes and molecular profiles. Over the past decade, through the use of molecular profiling, the number of different subtypes of breast cancer has grown and understanding the pathways driving each subtype may allow a stratified approach to therapy, allowing patients to receive the treatment which will be of most benefit. The Nuclear Factor kappa B (NF-κB) pathways regulate the transcription of a wide range of genes involved in the immune response, inflammation, proliferation and apoptosis. Many of these processes are hallmarks of cancer and NF-κB has been hypothesised to have a role in tumorigenesis. The aim of the current study was to investigate the role of both NF-κB pathways in the pathogenesis and recurrence of breast cancer. Immunohistochemistry was employed to assess key components of the canonical and non-canonical NF-κB pathways on a tissue microarray (TMA) of 544 patients with full clinical follow up and clinical information including ER status, subtype, necrosis, apoptosis and angiogenesis. Nuclear expression of p65 phosphorylated at serine 536 was associated with angiogenesis and shorter recurrence free interval. Cytoplasmic expression of IKKα was associated with cell death (apoptosis and necrosis) and a shorter recurrence free interval was also observed for those with high expression. These observations between phospho-p65/IKKα and recurrence free interval, when subdivided by ER status, remained significant in ER positive tumours but were negated in ER negative tumours. When split further into subtype, a diverging role for each was observed with phospho-p65 associating with recurrence in luminal B tumours and IKKα with luminal A tumours. Other members of the NF-κB pathways (p65, IKKβ, NIK and RelB) were not associated with recurrence free interval. When these results were tested in an independent cohort, IKKα remained significant on recurrence free interval and breast cancer specific survival in ER positive tumours however phospho-p65 was only marginally associated with breast cancer specific survival. Variability of phospho-p65 is a major issue in IHC studies and therefore an alternative marker of the canonical NF-κB pathway is required. Analysis of expression in this second cohort also revealed that high levels of IKKα in the cytoplasm were associated with recurrence on tamoxifen. This marker may therefore be able to be employed as a diagnostic tool to predict patients who are likely to display endocrine resistance and may represent a therapeutic strategy in combination with endocrine therapy, or for patients after endocrine resistance has occurred. Further examination of the pathways in breast cancer cell lines also demonstrated a difference between ER positive and ER negative breast cancer. In ER negative MDA-MB-231 cells phosphorylation of p65 (from the canonical NF-κB pathway) and phosphorylation of p100 (from the non-canonical NF-κB pathway) was apparent even in untreated control cells, suggesting constitutive activation. Expression was however found to be inducible in ER positive MCF7 cells. In order to investigate whether kinases involved in activation of each pathway, IKKβ in the canonical pathway and IKKα in the non-canonical NF-κB pathway, had potential as targets in breast cancer, we examined the phenotypic impact of silencing their expression in breast cancer cell lines. Silencing IKKβ induced apoptosis and decreased cell viability in both MCF7 and MDA-MB-231 cells but reduction in expression of IKKα only impacted on cell viability and apoptosis in ER positive MCF7 cells. This data, consistent with results from the clinical specimens, has therefore revealed that inhibitors of IKKα are likely to be most beneficial in the treatment of ER positive tumours. These results suggest that the NF-κB pathways are associated with recurrence in patients with ER positive tumours with each pathway possibly associating with recurrence in different subtypes. Additional studies in a larger cohort, including patients receiving aromatase inhibitors are required, accompanied by extensive mechanistic studies to further explore the roles of IKKα and IKKβ in breast cancer. These observations highlight that different subgroups of breast cancer may have different signalling pathways driving progression and therefore patients are likely to benefit from different therapeutic strategies.

Heat Shock Protein 90 (Hsp90) is a molecular chaperone which plays an active role in maintaining protein homeostasis. Hsp90 is known to be highly expressed in tumour cells where it regulates stability and function of several key oncogenic client proteins including Akt kinase, EGFR, CDK and PDGFR. These client proteins are mutated or overexpressed in tumours and are involved in tumour progression and metastasis due to their roles in signalling pathways, cell cycle and apoptosis. Hsp90 has two isoforms, namely Hsp90α and Hsp90β and share 85% sequence homology. Hsp90β is the constitutive isoform, however, Hsp90α is highly induced in many cancers and is responsible for tumourigenesis. Previous studies from our laboratory established Hsp90α mRNA and protein to be highly expressed in glioma cell lines and tissues compared to normal tissue and cells. In a follow up study, Hsp90α was silenced using predesigned small interfering RNA (sihsp90α) with high target specificity and it showed a clinical impact on the chemosensitivity to Temozolomide (TMZ). Thus, Hsp90α could be a therapeutic target for the treatment of glioma. This study utilized a Hsp90-inhibitor 17-AAG and hsp90α-specific siRNA (sihsp90α), either as single agent or in combination, to inhibit Hsp90 function in glioblastoma. Hsp90α mRNA and protein expression levels post treatment were evaluated using qRT-PCR and confocal microscopy. To determine if Hsp90α inhibition influenced Akt kinase activity, a Hsp90 client protein, Akt activity was examined in treated and control cells using a Akt Kinase activity kit. Novel results from the study revealed that Hsp90α was significantly down-regulated at both the mRNA and protein levels which were associated with reduced cell viability. Hsp90α inhibition resulted in loss of Akt kinase activity which validated the role of Hsp90α in chaperoning tumour progression. The use of 17-AAG concomitant with sihsp90α did not demonstrate synergistic anti-tumour effects in glioblastoma in vitro. Furthermore, the application of siRNA as an alternative to small-molecule inhibitors in the treatment of human disease has shown therapeutic potential. However, a major hurdle to its utility has been the difficulty in delivering these anionic macromolecules in vivo. In this study, the ability of Tat peptide to enhance siRNA-mediated knockdown of hsp90α and siRNA stability in serum was investigated. In glioblastoma, Tat-mediated sihsp90α transfection concomitant with 17-AAG exhibited significant downregulation of Hsp90α mRNA and protein levels with increased peptide concentrations. Hsp90α suppression was associated with reduced Akt kinase activity and cell viability. The sihsp90α/Tat complex significantly improved sihsp90α stability in human serum for up to 36 h post serum exposure. Finally, the combination therapy induced hsp90α knockdown and reduced Akt kinase activity in vivo. Although these results are preliminary and requires further validation, the anti-cancer activity is promising. These results suggest that the combination treatment with sihsp90α and 17-AAG may have therapeutic potential in GBM.

Breast tumours have intrinsic heterogeneity. The cancer stem cell hypothesis is currently challenging the notion that all cancer cells are equally malignant, suggesting that it is important to evaluate the efficacy of potential anti-cancer drugs by their ability to target the stem-like population. TRAIL is a cytotoxic agent the efficacy of which has been limited by a lack of patient stratification in clinical trials (Lemke 2014). In pre-clinical studies TRAIL has shown specificity towards mesenchymal-like breast cancer cell lines (Rahman et al. 2009) We show here that TRAIL is able to target the tumoursphere-forming population of four out of six breast cancer cell lines, including two epithelial-like lines, the bulk population of which is TRAIL-resistant. Furthermore, TRAIL also reduced the tumour-initiating capacity of the MCF-7 line. In addition, we have also investigated a paracrine mechanism of sensitising breast cancer cell lines to TRAIL. We have shown that a soluble factor produced by MDA-MB-231 cells, fibroblasts, and cancer-associated fibroblasts (CAFs) can sensitise both MCF-7 cells and SKBR3 tumoursphere-forming cells to TRAIL. Our data shows that cytoplasmic levels of Cellular FLICE-Like Inhibitory Protein (cFLIP) – a naturally occurring inhibitor of TRAIL’s cell toxicity effects- are lower in TRAIL-sensitive cells and suggest that tumoursphere populations are TRAIL-sensitive due to the re-localisation of cFLIP to the nucleus. We believe cFLIP is nuclear in stem-like cells due to a role as a promoter of the Wnt pathway. We have shown that inhibition of cFLIP by siRNA resulted in a reduction in both beta-catenin protein levels and Wnt-target gene transcription in both the MCF-7 and MDA-MB-231 breast cancer cell lines. We have also demonstrated a novel role for cFLIP as a promoter of bCSC maintenance. We have found that inhibition of cFLIP by shRNA decreased the self-renewal of tumoursphere-forming cells and also reduced colony formation. As TRAIL alone does not completely eradicate tumoursphere-forming or tumour-initiating cells in any breast cancer cell line, we believe our data are evidence of bCSC heterogeneity existing in terms of susceptibility to TRAIL. We propose a model of phenotypic heterogeneity within breast cancer cell lines and bCSCs whereby there exist two populations of cells which can be distinguished based on TRAIL susceptibility correlating with the known distinction of epithelial-like or mesenchymal-like status and our novel observation of cFLIP localisation. While these findings are currently restricted to cell lines, if confirmed in primary breast cancer cells, the clinical implication of our model is that although TRAIL alone is a potential therapy, a much more effective therapeutic strategy would be to also inhibit cFLIP, the consequences of which would not just be a sensitisation to TRAIL but also a reduction in Wnt signalling, and potentially a reduction in bCSC self-renewal and proliferation.

Immune responses contribute to the success of radiation therapy of solid tumours; however, the mechanism of triggering CD8+ T cell responses is poorly understood. Antigen cross-presentation from tumour cells by dendritic cells (DC) is a likely dominant mechanism to achieve CD8+ T cell stimulation. We established a cross-presentation model in prostate cancer in which DC present a naturally expressed oncofetal tumour antigen (5T4) from irradiated DU145 tumour cells to 5T4-specific T cells. Ionising radiation (12 Gy) caused G2/M cell cycle arrest and cell death, increased cellular 5T4 and high-mobility protein group-B1 (HMGB1) levels and upregulated surface calreticulin and Hsp70 expression in DU145 cells. Co-culture of DC with irradiated tumour cells lead to efficient phagocytosis of tumour cells and upregulation of CD86 and HLA-DR on DC. CD8+ 5T4-specific T cells, stimulated with these DC, proliferated and produced IFNγ. Inhibition of HMGB1 decreased T cell stimulation but not DC activation, while TRIF/MyD88 inhibition only had a marginal effect on T cell stimulation. Unlike previous reports, I found no functional evidence that DC with Asp299Gly toll-like receptor-4 (TLR4) single nucleotide polymorphism had impaired ability to cross-present tumour antigen. However, I observed a highly significant and robust prevention of antigen cross-presentation when tumour cells were pretreated with the novel Hsp70 inhibitor, VER 155008. The inhibitor also prevented CD86 upregulation on DC co-cultured with irradiated tumour cells. Together, the results in this thesis demonstrate that radiation induces immunologically relevant changes in tumour cells, which can trigger CD8+ T cell responses via a predominantly Hsp70-dependent antigen cross-presentation process.

Colorectal cancer (CRC) is one of the most common cancers in the Western world. Screening for CRC using faecal occult blood test (FOBT) is well established. There is evidence that DNA based stool tests may be more effective than FOBT. Hypothesis The hypothesis is that in patients with CRC, rectal mucus may contain DNA derived from colonic tumours. It is speculated that quantitative or qualitative assessment of DNA in rectal mucus may permit an improved method of CRC screening. Aims Using surgically resected specimens of colonic tumours: To assess the feasibility and reliability of measuring DNA in mucus samples To compare different devices to measure mucus DNA To assess the amount of mucus DNA at various distances from colonic tumours In patients with CRC and controls: To compare the amount of DNA in rectal mucus Using a panel of 3 DNA methylation markers, to compare the rectal mucus DNA methylation profile between patients with CRC and controls Methods Surgical colectomy specimens were obtained from 25 patients with CRC. The feasibility and repeatability of measuring mucus DNA amounts was established using different buffer solutions, different storage techniques and different sampling devices. Mucus DNA amounts were measured at tumour sites and various distances proximal and distal to the tumour. 58 patients referred to a colorectal outpatient clinic with suspected CRC were assessed. Rectal mucus samples were obtained using a balloon device introduced through a proctoscope. All patients were investigated by colonoscopy to clarify the presence or absence of a CRC. The amount of DNA in the mucus samples was measured. The presence of three DNA methylation markers (NDRG4, TFP12 and GATA4) was assessed in all samples. All studies were approved by the local ethics committee. Results Reliable measurement of DNA from mucus samples was established using balloon, foam and brush devices and a cell lysate buffer. Higher amounts of DNA in surgical specimens were found distal to tumours compared to proximally. In patients with CRC the amount of DNA in rectal mucus was higher than in controls (no disease or benign polyps). The three DNA methylation marker panel had a sensitivity of 87% and specificity of 27.5% for the detection of CRC. Conclusions The results are consistent with the hypothesis that DNA detected in rectal mucus is derived from proximal tumours. Higher levels of rectal mucus DNA are obtained from patients with CRC than from controls. The selected DNA methylation panel was not sufficiently useful in our sample group to be of use as a screening technique, due to poor specificity. Further work is in progress to compare DNA abnormalities in resected tumour tissue with DNA from rectal mucus in the same patients. Future work may be required to improve the panel of DNA abnormalities assessed.

Mutational activation of RAS is one of the most common molecular abnormalities associated with acute myeloid leukaemia (AML). Normal human haematopoietic progenitor cells (HPC) expressing mutant RAS overproduce ROS due to NADPH oxidase (NOX) activation and this promotes the proliferation of these cells as well as AML blasts. The mechanisms by which ROS promote proliferation is however unclear. The current study investigated the effect of RAS-induced ROS production on gene expression in normal HPC using gene expression profiling (GEP) and assessed whether ROS-induced gene expression changes contributed to the pro-proliferative phenotype. In order to determine the ROS-specific GEP, Affymetrix Human Exon 1.0ST arrays were used for the comparison of mutant RAS and control cells cultured in the presence or absence of the NOX inhibitor, DPI, which strongly suppressed the production of ROS. This analysis showed that RAS changed the expression of 342 genes. Of these, 24 genes were specifically altered in response to ROS production by these cells and amongst these increased expressions of genes of the glycolytic pathway were prominent. To establish the functional significance of up-regulated expression of glycolytic enzymes, aldolase C (ALDOC) was investigated since it showed greatest induction with ROS. ALDOC was directly induced by physiological levels of ROS in both HPC and AML cell lines. Further, overexpression of ALDOC demonstrated that its overexpression promoted the proliferation and serum-independent survival of leukaemic cell lines. Conversely, antiproliferative effects were observed when ALDOC was knocked-down in cells known to have high levels of constitutive ROS production. Given the high frequency of ROS production in AML, this study provides a plausible mechanism of enhanced glycolysis seen in this disease and suggests that agents restoring the redox environment could be used to correct metabolic imbalances which contribute to treatment resistance.