The permanent cell cycle arrest known as cellular senescence is a major block to tumorigenesis. Currently the effects of latent senescence signaling on disease progression, response to therapy and outcome are poorly understood. Furthermore, the role of microRNAs in the regulation of senescence remains to be fully elucidated. For immortalisation to occur replicative senescence must be bypassed usually by activating a telomere maintenance mechanism (TMM). However, the expression differences between TMMs are also poorly understood. To address these questions a combination of gene expression and miRNA microarray profiling, virtual drug and siRNA kinase screening were utilised. These findings highlight the distinct roles of secretory and damage associated senescence pathways in disease progression and in response to therapy. Examination of the differentially expressed genes between TMMs also highlighted a differentially expressed gene expression network surrounding TERT, regulated by c-Myc and TCEAL7 in TMMs. These findings give further insight into the complex regulation network surrounding senescence signaling during tumorigenesis.

Chronic myeloid leukaemia (CML) is a myeloproliferative disease of stem cell origin. It is characterised by the Philadelphia chromosome and Bcr-Abl oncoprotein which is a constitutively activated tyrosine kinase that is causative in CML. Treatment of CML was revolutionized by tyrosine kinase inhibitors (TKIs) in the last decade. TKIs target Bcr-Abl and block its tyrosine kinase activity. Despite the success of TKI in eliminating differentiated CML cells, primitive quiescent CML stem cells still resist or persist with TKI treatment. In this study, several strategies have been investigated towards elimination of TKI-insensitive primitive CML stem cells. At first, the effect of nuclear entrapment of Bcr-Abl protein in CML cells was studied. Although Bcr-Abl protein is located in the cytoplasm of CML cells, TKIs, such as imatinib mesylate (IM), can induce the nuclear translocation of Bcr-Abl. Nuclear Bcr-Abl tyrosine kinase is capable of inducing apoptosis after drug washout, and leptomycin B (LMB) can trap translocated Bcr-Abl in the nucleus. Primary CML CD34+ cells were treated with IM and LMB either continuously or sequentially. It was found that neither regimen significantly increased the anti-proliferative effect of IM in primary CML CD34+ cells. It was also observed that the majority of Bcr-Abl was still retained in the cytoplasm of CML cells treated with IM and LMB, suggesting other mechanisms apart from its tyrosine kinase activity retained Bcr-Abl protein in the cytoplasm of CML cells. Next the subcellular distribution of Bcr-Abl was investigated in TKI-insensitive CML progenitors which survived 12-Day treatment of a potent TKI dasatinib (150nM). It was demonstrated that about 50% of Bcr-Abl was still retained in the cytoplasm of TKI-insensitive primary CML progenitors, although there was a significant increase in nuclear Bcr-Abl level in these survived cells compared to the no drug control (NDC). It was hypothesised that the cytoplasmic retention of Bcr-Abl was caused by its cytoplasmic binding partners, and it was found that a proportion of Bcr-Abl protein was associated with 14-3-3 proteins in the surviving cells, indicating the cytoplasmic retention of Bcr-Abl might be due to its binding with 14-3-3 proteins. In addition, due to the rapid nature of kinase inhibition and nuclear transportation, studies are required to measure the earliest time-point of inhibition of Bcr-Abl tyrosine kinase by IM. The common method to do this is to employ p-CrkL which has been widely used as a surrogate marker of Bcr-Abl tyrosine kinase activity, but the accuracy of p-CrkL as an indicator of Bcr-Abl tyrosine kinase status at early time-points during in vitro TKI treatment has not been examined. It was demonstrated that p-CrkL was not a reliable indicator of Bcr-Abl kinase activity within 24 hours of IM treatment in vitro and indicated that the early responses to IM and dasatinib were different. It was also observed that there was a rapid and active dephosphorylation of Bcr-Abl within 1 hour of TKI treatment, driven at least in part by protein tyrosine phosphatase activity. Furthermore, a farnesyltransferase inhibitor (FTI) BMS-214662 preferentially induces apoptosis in CML stem and progenitor cells compared to their normal counterparts, but another similar FTI BMS-225975 does not. However, the mechanism of action of BMS-214662 in inducing apoptosis of CML cells is not clear. When BMS-214662 was used as a negative control in the p-CrkL study, it was unexpectedly observed that the drug accumulated Bcr-Abl protein in CML cells. Thus, it was hypothesised that BMS-214662 may function as a proteasome inhibitor, which could result in accumulation of Bcr-Abl protein and further elevation of intracellular ROS level, leading to apoptosis of CML cells. Although BMS-214662 treatment induced accumulation of total ubiquitinated proteins and specifically Bcr-Abl protein in K562 cells, BMS-225975 had very similar effects, and this might result from the common mechanism of BMS-214662 and BMS-225975, which is their FTI activity. On the other hand, BMS-214662 induced significantly higher levels of intracellular ROS in both proliferating and non-proliferating K562 cells with respect to BMS-225975, indicating the production of ROS may be involved in the non-FTI mechanism of action of BMS-214662. However it was concluded that BMS-214662 was not a proteasome inhibitor like bortezomib. Finally, the use of synthetic low density lipoprotein (sLDL) as a vehicle for drug delivery to overcome the insufficient intracellular drug concentration in CML stem cells was investigated. Uptake and internalization of unloaded sLDL particles by CML cell line K562 and for the first time by CML stem cells was observed, demonstrating the targeting potential of sLDL particles in CML when they are loaded with drugs. Overall, this study provides further understanding of CML treatment and identifies some alternative strategies to target CML stem cells that may be used in combination with TKI to enhance the eradication of this stem cell-driven disease.

Upregulation of the extracellular signal-regulated kinase (ERK) pathway has been shown to contribute to tumour invasion and progression. Since the two predominant ERK isoforms (ERK1 and ERK2) are highly homologous and have indistinguishable kinase activities in vitro, both enzymes were believed to be redundant and interchangeable. To challenge this view, here we show that ERK2 silencing inhibits invasive migration of MDA MB 231 cells, and re-expression of ERK2 (but not ERK1) restores the normal invasive phenotype. A detailed quantitative analysis of cell movement on 3D matrices indicates that ERK2 knockdown impairs cellular motility by decreasing the migration velocity as well as increasing the time that cells remain stationary. We used gene expression arrays to identify rab17 and liprin β2 as genes whose expression was increased by knockdown of ERK2 and restored to normal levels following re expression of ERK2 (but not ERK1). Moreover, we established that both Rab17 and Liprin β2 play inhibitory roles in the invasive behaviour of three independent cancer cell lines, indicating a suppressive role for these proteins in tumour progression. Importantly, knockdown of either Rab17 or Liprin β2 restores invasiveness of ERK2 depleted cells, indicating that ERK2 drives invasion of MDA MB 231 cells by suppressing expression of these genes. Taken together, our data provides evidence that true functional disparities between ERK1 and ERK2 exist with regards to cell migration and identifies Rab17 and Liprin β2 as two novel motility suppressors downstream of ERK2.

The hypothesis of this work was whether IGRT could be safely implemented for clinical use in a busy oncology centre. I aimed to study a number of questions that remain unresolved in the current literature regarding safe and optimised implementation of IGRT techniques. The first study undertaken was the calculation of a local set up margin using two widely recognised margin recipes. This involved the assessment and analysis of multiple images belonging to 100 patients. This allowed progression onto the next project which was assessment of the optimal safe method of delineation of 4DCT. The most efficient method was compared to gold standard. At this point a different aspect of the radiation process was assessed, namely verification. A feasibility study of a simple, efficient form of imaging for use in review of a particular error was performed. This also involved the use of a novel tool which required independent assessment. This progressed into a further study of a larger number of patients using this tool and the images assessed previously to verify a novel form of radiation delivery. Lastly a planning study was performed to quantify the clinical benefit of another delivery system. This involved the delineation and planning of a large number of radical lung patients with standard radiation treatment and the novel radiation treatment and an assessment of the potential clinical benefits. The work presented in this thesis has answered some specific questions in IGRT in lung cancer, and contributed both locally and in the wider lung cancer community to increasing the use of IGRT in lung cancer.

Quantitative pathology as a tool in gynaecological pathology is fairly new. Such techniques allow greater objectivity than histological grading, typing, and residual tumour estimation. This study aims to determine: whether basic morphometry data can predict outcome and chemotherapeutic response, whether newer semi-automated methods of tumour morphometry provide similar results to older methods, and whether advanced image analysis methods can offer further tumour outcome data in ovarian carcinoma. The study was performed on a well-selected group of serous ovarian carcinomas. Tumour outcome, survival and chemotherapeutic response, were investigated in 132 patients treated with the same platinum containing regimes. Traditional clinicopathologic parameters, p53 & Bcl2, mitotic activity index MAn and angiogenesis determinants were initially investigated. Semi-automated analysis, using immunohistochemically based techniques, were applied to estimate volume percentage epithelium (VPE) and nuclear morphometric parameters. Syntactic structure analysis including, minimum spanning tree, and neighbourhood features, was also investigated. Multivariate analysis revealed residual disease status, FIGO stage, MAl, VPE, equivalent nuclear diameter, and angiogenesis parameters to be strong prognosticators for overall and disease free survival. Residual disease status, VPE, nuclear length and angiogenesis parameters were found significant predictors of chemotherapy response. Angiogenesis parameters, as determined by semi-automated image analysis techniques, were found overall to be the strongest prognosticators. Morphometric data can predict outcome and chemotherapeutic response in ovarian serous carcinoma. Semi-automated morphometry techniques provide similar results to older methods, and advanced image analysis can offer further outcome data. The rationale for the application of semi-automated and automated detection is that it may provide an unbiased sampling of a lesion and possibly a more representative estimate of areas that a human expert might label. Such determined, quantitative pathological findings were found to have important value in predicting prognosis in ovarian carcinoma and, if not to supersede, certainly to add to classical prognostic factors.

Meningioma subtypes classification is a real world problem from the domain of histological image analysis that requires new methods for its resolution. Computerised histopathology presents a whole new set of problems and introduces new challenges in image classification. High intra-class variation and low inter-class differences in textures is often an issue in histological image analysis problems such as Meningioma subtypes classification. In this thesis, we present an adaptive wavelets based technique that adapts to the variation in the texture of meningioma samples and provides high classification accuracy results. The technique provides a mechanism for attaining an image representation consisting of various spatial frequency resolutions that represent the image and are referred to as subbands. Each subband provides different information pertaining to the texture in the image sample. Our novel method, the Adaptive Discriminant Wavelet Packet Transform (ADWPT), provides a means for selecting the most useful subbands and hence, achieves feature selection. It also provides a mechanism for ranking features based upon the discrimination power of a subband. The more discriminant a subband, the better it is for classification. The results show that high classification accuracies are obtained by selecting subbands with high discrimination power. Moreover, subbands that are more stable i.e. have a higher probability of being selected provide better classification accuracies. Stability and discrimination power have been shown to have a direct relationship with classification accuracy. Hence, ADWPT acquires a subset of subbands that provide a highly discriminant and robust set of features for Meningioma subtype classification. Classification accuracies obtained are greater than 90% for most Meningioma subtypes. Consequently, ADWPT is a robust and adaptive technique which enables it to overcome the issue of high intra-class variation by statistically selecting the most useful subbands for meningioma subtype classification. It overcomes the issue of low inter-class variation by adapting to texture samples and extracting the subbands that are best for differentiating between the various meningioma subtype textures.

The study has further shown that IL-1[Special character omitted] is a potential regulator of TEM-8 expression in tumours and that IL-1[Special character omitted] significantly induced the formation of capillary-like tubules from the HECV cells, accompanied by an increase in TEM-8 expression. Where as, elimination of TEM-8 by way of ribozyme transgene significantly decreased the formation of capillary-like tubules and the mobility of HECV cells. Moreover we have demonstrated that the vW domain together with trans-membrane domain of TEM-8 are the structural domain that are responsible for the tubule forming action of TEM-8, using expression construct in varying cell lines.

The major aim of this thesis was to perform the first ever proteomics study on bladder cancer exosomes. Initially, exosomes were isolated from urine specimens but hypervariable yields and poor sample quality made proteomics analysis challenging. As an alternative approach, exosomes were isolated from HT1376 bladder cancer cells. Exosomes were purified by ultracentrifugation on a sucrose cushion, and preparations verified as high quality by immunoblotting, flow cytometry and electron microscopy. For global proteomics analysis, the sample was solubilised using SDS and DTT and subjected to LC-MALDI-TOF/TOF MS. We identified 353 proteins with high confidence and 63 of these have not previously been identified in other proteomics studies on human exosomes. Overrepresentation analysis demonstrated that the proteome was consistent with that of other exosomes with significant overlap with exosomes of carcinoma origin. Comparisons with the Gene Ontology database also highlighted strong associations with carcinoma of the bladder and other sites. A GeneGo generated protein interaction network highlighted c-Myc as a major node of protein interaction within this dataset. Several MS-identified proteins were confirmed as genuinely exosomally expressed using a combination of immunoblotting, flotation on continuous sucrose gradients, and flow cytometry. Expression was also verified in exosomes from a variety of sources, including urine. The data will aid our understanding of exosome biogenesis and function and may inform the development of urine exosome-based clinical tools in bladder cancer.

Resistance to endocrine therapy in breast cancer is associated with poor prognosis. Cell models of acquired tamoxifen resistance have implicated altered growth factor receptor signalling, especially the ErbB family of receptor tyrosine kinases, in development of the accompanying aggressive phenotype. Microarray analysis of an in vitro wtMCF-7 based model of acquired tamoxifen resistance (‘Tam-R’) identified upregulation of CD44, a transmembrane glycoprotein, known to interact with ErbB receptors and influence breast cancer progression. We investigated the hypothesis that CD44 overexpression in Tam-R cells can modulate ErbB activity and promote an adverse phenotype. CD44 gene overexpression was validated by RT-PCR and its protein expression determined by Western blotting and immunocytochemistry. CD44 contribution to intracellular signalling and phenotype of Tam-R cells (migration, invasion and growth), both endogenous and in response to hyaluronan (HA), was determined using Western blotting, immunocytochemistry and functional assays including wound healing, Boyden chamber migration, Matrigel™ invasion and growth assays in the presence or absence of siRNA-mediated CD44 knockdown. Interactions between CD44 and ErbB receptors were investigated using immunofluorescence and immunoprecipitation. CD44 was overexpressed at gene and protein level in Tam-R versus wtMCF-7 cells and whilst this did not influence the endogenous phenotype of Tam-R cells, it enhanced their sensitivity to HA as evidenced by HA-induced MAPK, EGFR and HER2 activation and increased migration. HA-induced migration was attenuated following treatment with the MAPK inhibitor PD098059, gefitinib as well as trastuzumab. CD44 was found to associate with HER2 and HER3 at the cell surface whilst HA stimulation appeared to modulate ErbB dimerisation patterns. Our data suggest that CD44 overexpression sensitises tamoxifen-resistant cells to HA thereby modulating ErbB dimerisation and enhancing migration. These observations may have importance in vivo where the tumour microenvironment can provide a rich source of HA to promote the progression of tamoxifen-resistant tumours.

This thesis examines factors influencing the outcomes of patients receiving multidisciplinary stage-directed treatment for oesophagogastric cancer. The hypotheses tested were: 1.The TNM7 staging system is a more accurate prognostic tool for oesophageal cancer (OC) than TNM6. 2.Use of CT-PET upstages a significant number of patients with occult metastases. 3.OC recurrence patterns differ following definitive chemoradiotherapy (dCRT) and surgery, but overall recurrence rates and survival are comparable for advanced stage disease. 4.An involved circumferential resection margin (CRM+) following oesophagectomy is associated with poorer survival and its incidence can be reduced with neoadjuvant chemoradiotherapy. 5.Early enteral nutrition improves clinical outcomes following upper GI cancer resection. 6.Centralisation of oesophagogastric cancer (OGC) surgery in S.E. Wales is feasible and associated with improved clinical outcomes. Reclassification with TNM7 resulted in stage re-categorisation of 11.9% of OC patients. Multivariate analysis indicated only TNM7 prognostic group to be independently and significantly associated with survival. CT-PET upstaged OC M stage in 24.0% of patients. Loco-regional OC recurrence was commoner after dCRT (p<0.0001) but distant recurrence commoner after surgery (p=0.001). Disease-free survival was better after surgery for stage I (p=0.069) and II (p=0.011) but comparable with dCRT for stage III (p=0.878) and IV (p=0.710). CRM+ occurred in 38.0% of all OC patients, and 62.4% of pT3 patients. Multivariate analysis revealed lymphovascular invasion (p<0.0001) and CRM+ (p=0.002) were independently and significantly associated with disease-free survival. Multivariate analysis revealed EUS T stage (p<0.0001) and neoadjuvant chemoradiotherapy (p<0.0001) were independently associated with CRM+. Early enteral nutrition (EEN) was associated with reduced hospital stay (p=0.023) and less operative morbidity (p=0.044) than control management, due to fewer wound infections (p=0.017), chest infections (p=0.036) and anastomotic leaks (p=0.055). Following centralisation, OGC critical care (p<0.0001) and total hospital stay (p=0.037) were significantly reduced. Serious operative morbidity (Dindo-Clavien grade III+) decreased from 33.3% to 16.7% (p=0.066).