Humanisation of rodent mAbs by CDR-grafting is now a standard procedure for reducing immunogenicity and recruiting human effector functions with ten humanised mAbs approved for use and many more in development. This project aimed to develop two anti-cancer therapeutics aimed at exploiting two tumour associated antigens overexpressed in the tumour environment.

The historical precedence for employing bacteria to target cancer stretches long back some 300 years. Despite of the advancements in chemotherapy, problems such as metastasis and tumour resistance to chemotherapy and radiotherapy have limited the scope of therapeutic effects of existing clinical treatments. The microenvironment of solid tumours provides an ideal environment for growth of facultative and strictly anaerobic bacteria. These bacteria can colonize such environments leading to tumour regression. Such investigations have given the concept of bacteriolytic therapy, in which live bacteria could be employed for the targeted delivery into tumours leading to its suppression. With this concept in mind it was decided to explore the above idea by using probiotic bacterium (Lactobacillus casei), which is considered to be non-pathogenic and also provides health benefits to the host. In order to minimize dispersion of the bacteria throughout the host and to facilitate its delivery into tumours, some kind of containment was desirable. To test this kind of approach it was decided to exploit immobilisation technology to microencapsulate live bacteria for later injection into tumour sites. To date no such approach has been employed to develop such a microencapsulation system that could be utilized as a delivery vehicle for live bacteria to deliver it through a hypodermic needle directly into tumours. In order to pursue the above objective, the microencapsulation method was also characterised with respect to stability and viability of the L.casei. Microencapsulated preparations of Lactobacillus casei NCDO 161 were developed and demonstrated that the culture supernatant of microencapsulated preparation inhibited growth of tumour cells (in vitro). Further investigation of a variety of bacterial preparations on tumour growth (in vivo) following intra-tumoural injection demonstrated that the live microencapsulated preparation had severe inhibitory effect on tumour growth when compared with non-encapsulated live and encapsulated heat killed preparations. Histological studies were performed to demonstrate the presence of live bacteria in tumours at early stages and to study the effects of the bacteria on tumour architecture during the treatment.

Colon cancer is a complex and multifactorial disease and the colonic microbiota has been strongly implicated in its aetiology. The dynamic metabolic interaction between the colonic microbiota, dietary components and the colonic tissue are primary factors influencing the process of colon carcinogenesis. Human dietary habits strongly influence the development of sporadic colon cancer. However, characterising microbe-nutrient interactions and determining the effect of bacterial metabolic activity on colon cancer risk may help identify dietary components with potential cancer-preventing or promoting activities leading eventually to the construction of effective nutritional strategies. The practical and ethical problems associated with the use of cancer as an endpoint and the invasive nature of colonic sampling has made the use of in vitro culture models and biomarkers an attractive alternative for identifying nutrients that may reduce the risk of colon cancer. This, study set out to investigate the anticancer properties of plant bioactive compounds from hops (Humulus lupulus L) and bioactive components from plants of the agave genus (Agave tequilana & Agave angustifolia) on selected colon cancer biomarkers. The predominant prenylflavonoid present in beer, isoxanthohumol (Ix), which is weakly estrogenic, has been shown to undergo a degree of microbial transformation within the colon to 8-prenylnaringenin (8-PN) (the most potent phytoestrogen known), with inter-individual differences ranging from 0 to 80% conversion. Using a range of bioassays both Ix and 8-PN were shown to have similar anti-genotoxic (comet assay) and anti-invasive (matrigel assay) activity which would indicate that such activity is independent of estrogenicity. Ix was shown to alter Caco-2 cell proliferation and alter cell cycle kinetics at low concentrations compared to 8-PN. Ix also exhibited genotoxic activity at high concentrations (50~M) as well as significant cell cycle arrest (G2/M phase) raising concerns for safety as well as its potential application as a therapeutic anticancer compound. Future research would be directed at in vivo animal studies to determine if the activities observed in this study have any actual influence on tumourogenesis. The leaves of the plant Agave tequilana are a waste product of the tequila industry and medicinal practices by indigenous Mexican populations have led us to investigate the anticancer properties of the aqueous leaf extract. The observed anti-genotoxic activity of the aqueous agave leaf extract in this study would enable it to be considered as a potential cancer preventative supplement, although the observed systemic cytotoxic activity (lymphocytic Jurkat cells) highlights the need for caution and the necessity for further investigation into the extract's constituent active compounds and their systemic bioavailability. The simulated human intestinal microbial ecosystem (SHIME) culture model was used to investigate microbial and metabolic changes indicative of prebiotic activity of the fructan fraction from Agave angustifolia. Supplementation of the Agave fructan in the SHIME model stimulated lactobacilli and bifidobacterial populations and enhanced butyrate concentration. Furthermore, of particular interest was the reduced ammonium concentration in conjunction with elevated concentrations of short chain fatty acids in the descending colon vessel which suggests a reduction of proteolytic fermentation, a desirable change that has not previously- been noted in the descending colon with inulin type fructans. The promising results obtained in this study would need to be confirmed in vivo prior to evaluating the possible application of Agave fructan in human nutrition. The final study set out to investigate the effect of incubating SHIME effluent following treatment with Agave fructan or inulin type fructan (ITF) on several in vitro mammalian cell systems that model the development of cancer. Neither ITF, nor Agave fructan treated SHIME effluent were shown to have any influence on genotoxicity or to prevent H202 induced DNA damage when incubated with Caco-2 cells. However effluent from SHIME systems supplemented with either fructan type reduced Caco-2 monolayer permeability, particularly effluent from the ascending and descending colon vessels. Such an effect is indicative of improved barrier function and antitumour-promoting activity. Future research should be directed towards extending these in vitro studies to the in vivo situation to clarify the prebiotic effects of Agave fructan.

Recombinant proteins and antibodies have both diagnostic and therapeutic uses with biological therapeutics being one of the largest markets in the pharmaceutical industry. The work in this thesis details a series of studies examining novel applications of antibodies and recombinant proteins in the treatment of cancer. Initially, two novel fusion tags were evaluated for their ability to overcome some of the problems associated with the expression of recombinant proteins in Escherichia coli to use as immunogens to raise antibodies. Both of the fusion tags where shown to be effective as N-terminal fusion tags for the expression and purification of target proteins and may prove useful in circumventing expression and purification issues associated with E. coli as an expression host. 1)1 the second part of the project, amphiregulin (AREG) and heparin binding epidermal growth factor (HBEGF) recombinant proteins were expressed in E. coli and used to produce monoe/onal antibodies. The AREG antibody and HBEGF/AREG cross-specific antibodies showed anti-proliferative effects that were enhanced by chemotherapy with significant effects on cell invasion also observed. The antibodies enabled the validation of targeting AREG and HBEGF as a potential cancer therapy. Finally, a new fusion protein therapeutic composed of the two cathepsin peptides was developed and examined. Inhibitory effects were observed on the enzymatic activity of both cathespins Band S as well as a reduction in cell invasion. This inhibitory molecule may have utility in the treatment of cancer given the role played by both proteases in tumourigenesis. In this study, the use of recombinant proteins as immunogens for generation of monoe/onal antibodies as well as a targeted therapy in their own right has been demonstrated. Both inhibitory molecules developed in this study have potential to be successful biological therapies and will be examined further.

Tumour growth and metastasis are dependent on the formation and maintenance of a dedicated microvascular network. Angiostatin4.5 (AS4.5) is a naturally occurring anti-angiogenic agent known to inhibit angiogenesis in vitro and tumour growth in vivo.

With the findings in anticancer and antiretroviral research, it is strongly believed that CDK9 inhibition would be a useful therapeutic strategy in these fields. A number of pharmacological CDK inhibitors have been previously reported and are undergoing preclinical and clinical evaluation, but none of them selectively targets CDK9. In this project, a number of anilino-4-(thiazol-5-yl)pyrimidine derivatives have been successfully prepared. The most potent compound exhibited best potency with GIso around 10 nM. A set of compounds were tested against CDKl, 2, 7 and 9 and the preliminary structure-activity relationship was established. More than a dozen compounds showed excellent CDK9 inhibitory activity with Ki lower than 10 nM. Lead compound 29, 52 and 67 showed different profile of selectivity towards CDK9, with 21, 6 and 7 nM Ki respectively. Cellular mode of action of lead compound 29 was investigated and the correlation between CDK9 inhibition and anti-tumor activity was confirmed. In the CLL cytotoxicity assay, comparing with clinical compound flavopiridol, compound 52 showed significant selectivity advantage toward norinal B and T-cells against CLL B-cells, which is more than WOO-fold. Several lead compounds, like 52 and 69, are currently scheduled for in vivo assessment. Further research on the basis of this project will result in lead drug candidates for potential cancer therapy.

The clinical trials that provide evidence for new cancer treatments often fall sort of providing adequate evidence upon which to base a decision about reimbursement by healthcare providers. There is the potential to improve the design of trials so that they provide better information about the value of new treatments. Three examples in early breast cancer are presented that demonstrate the role of decision modelling in planning research for the evaluation of cost-effectiveness. The results show that decision modelling at this stage can make a useful contribution to the trial design process in terms of the prioritisation of research design options and the suggestion of efficient sample sizes. Bayesian decision theory and value of information analysis are particularly helpful in the interpretation of the model. There may also be a role in aiding real-time trial adaptation where alteration of a trial design can improve the efficiency of research. There are challenges in the conduct of evidence synthesis and modelling prior to a phase III trial. There are also difficulties related to the prediction of a technology's useful lifetime and its future price and the market within which it will be available. Further research is required if early modelling is to be a robust basis for public or commercial research investment decisions. J

Phase 11 trials are a key element of the drug development process and the transition from phase 11 to III is associated with the highest risk in the development pathway. This thesis considers methods to optimise the design of phase 11 trials in cancer. It addresses one of the most important issues in designing such trials - choosing the most appropriate outcome measure to reflect clinical activity. The first step to optimising phase 11 trial design addresses the process behind identifying appropriate designs, allowing researchers to understand the key elements of phase 11 trial design. A guidance document providing researchers with clear points for consideration throughout the design process is developed, to aid researchers to make informed decisions throughout the design process. Successful phase 11 trial design hinges on choosing an outcome measure that best reflects the expected mechanism of action of the treatment under investigation and that acts as a screen for potential long-term benefit. This research considers the novel application of surrogacy methodology to identifying optimal outcome measures for phase 11 trials in advanced colorectal cancer (aCRC). Typically response is used as the primary endpoint in such trials; however this has been shown to be a poor surrogate for survival in aCRC. This thesis evaluates the relationship between response and overall survival in aCRC trials, and explores alternative phase 11 outcome measures in order to identify those most strongly associated with survival. The output of this research will aid trialists in the design of future phase II trials, providing a structured approach to trial design, and recommendations regarding a novel methodological approach to identifying optimal outcome measures. Specifically in aCRC, the research makes recommendations regarding the use of alternative outcome measures to response, providing a basis for the design of future phase 11 trials in this disease area.

Cancer cells are highly dependent on glycolysis. which provides protection against the hypoxic tumour microenvironment. Colorectal cancer cells have been reported to undergo increased glycolysis, the so-called 'Warburg's effect'. Previous research in our laboratory revealed increased expression of hypoxic and glycolytic markers to be associated with aggressive colorectal cancer phenotype. Pyruvate dehydrogenase kinase (PDK) is a crucial mitochondrial enzyme, which plays an important role linking glycolysis to oxidative phosphorylation. Inhibiting PDK with its pharmacological inhibitor, Dichloroacetate, or with RNA interference has been reported to have anti-cancer effects in other solid cancers such as head and neck squamous cell cancer. Since colorectal cancer cells are also known to be reliant on glycolysis, my aim was to determine if switching metabolism from glycolysis towards mitochondrial respiration, by inhibiting PDK, would reduce growth preferentially in colorectal cancer cells over normal cetis, and to examine the underlying mechanisms. Representative colorectal cancer (HT29, SW4BO and LaVa) and non-cancerous cell lines (HB2 and 293) were treated with Dichloroacetate (DCA), an inhibitor of PDK. 20mM DCA did not reduce growth of non-cancerous cells but caused significant decrease in cancer cell proliferation, which was associated with apoptosis and G2 phase cell cycle arrest. DCA reduced lactate levels in growth media and induced dephosphorylation of E10 sub-unit of pyruvate dehydrogenase complex in all cell lines, but the intrinsic mitochondrial membrane potential was reduced in only cancer cells. PDK exists in four isoenzymes (PDK 1-4). Next, I aimed to examine expression levels of the individual PDK isoenzymes in the panel of cell lines used, downregulate the individual PDK isoenzymes using short interierence RNA (siRNA), and • - 6 - investigate whether a subset of the PDK isoenzymes represents the actual target of DCA. Expression levels of individual PDK isoenzymes varied amongst the cell lines in a manner that did not suggest which was the main target of DCA. Down regulation 01 individual PDK isoenzymes failed to induce apoptosis in the cancer cells and, interestingly there was a compensatory rise in the mRNA levels of the other PDK isoenzymes. Treatment with siRNAs directed against various combinations of PDK isoenzymes did not induce apoptosis in cancer cells. Treatment of cells with DCA led to an up regulation of PDK isoenzymes, particularly PDK4. These results indicate that it is difficult to isolate a subset of PDK isoenzymes as potential therapeutic targets since a combination of functional redundancy and compensatory up-regulation among the isoenzymes means that sufficient overall PDK activity is maintained. Cancer cells are most sensitive to radiation in the G2-M phase of the cell cycle. Since DCA induced G2 arrest in the colorectal cancer cells, my hypothesis was that it sensitises these cells to radiation. HT29, SW480, LoVo, and 293 cells were treated with 10 and 20mM DCA, and irradiated. DCA did not induce any change in the sensitivity of 293 or HT29 cells to radiation. In contrast, SW480 and LoVo cells were significantly sensitised to radiation after pre-treatment with 10mM, but not 20mM DCA. There were minimal effects of 10 and 20mM DCA on the cell cycle profiles of all the cell lines studied. The cell lines derived from the poorly differentiated or metastatic cancers were sensitised to radiation following pretreatment with DCA, and cell lines derived from well-differentiated colorectal cancer or non-cancerous epithelial cells were resistant to any such sensitisation. However I am unable to provide any evidence to explain this radiosensitisation effect. This warrants further experiments to confirm the underlying molecular basis of these effects of DCA .

Cancer gene therapy holds promise for patients, yet key issues involving the delivery of vector and achieving tumour selective cytotoxicity, has limited progress over recent years. In this thesis a combination of cell-based carriers, viral vectors and tumour selective promoters are assessed to tackle these two important issues directly. Initially two retroviral systems were considered: intracellular carriage of a rapamycin-inducible retrovirus and extracellular carriage ('hitchhiking') of a self-inactivating (SIN) retroviral vector. Both systems controlled transgene expression using the tumour selective human telomerase reverse transcriptase and telomerase RNA promoters (hTERTp & hTRp). Direct transduction of target cells with SIN marker virus, expressing enhanced Green Fluorescent Protein (eGFP) under the internal control of human telomerase reverse transcriptase promoter (hTERTp) demonstrated similar activity to a positive control virus, with some evidence of telomerase selectivity. Hitchhiking of the same SIN marker vectors also mediated eGFP expression in target cells. In contrast the therapeutic SIN vectors containing suicide transgenes did not achieve a reliable direct telomeraseltumour selective cytotoxic effect, with no discernible toxicity seen with hitchhiked vector. The rapamycin-inducible vectors in contrast to the SIN vectors only demonstrated retroviral genome expression following single cell cloning of a positive-control producer cell line. The retrovirus from this producer successfully mediated low-level target cell eGFP expression, however the data did not support development of therapeutic vectors in,corporating either telomerase promoter. Incorporating oncolytic viruses into the system allowed in vitro therapy. An eGFP expressing oncolytic vesicular stomatitis virus (VSV) and wild-type reovirus, showed direct cytotoxicity against a range of target cell lines. T-cell and non-T-cell peripheral blood mononuclear cell (PBMC) fractions were resistant to both viruses indicating their suitability as oncolytic viral carriers. Finally VSV was successfully hitchhiked on donor T-Iymphocytes leading to the release of VSV .and widespread replication within the malignant Molt-4 cell line.