Nucleic acids are macromolecules in cells for storing and transferring genetic information. Moreover, nucleic acids, especially RNAs, can fold into well-defined 3D structures and catalyze biochemical reactions. As ubiquitous biological molecules in all living systems, nucleic acids are important drug targets, and they can also be used in diagnostics and therapeutics. Structural information of nucleic acids provides the foundation for DNA and RNA function studies. X-ray crystallography has been a useful tool for structural studies of bio-macromolecules at atomic level. There are two major problems in macromolecular crystal structure determination: phasing and crystallization. Although selenium derivatization is routinely used for solving novel protein structures through the MAD phasing technique, the phase problem is still a critical issue in nucleic acid crystallography. The covalent selenium-derivatization of nucleic acids has been proven to be a useful strategy for solving the phase problem in nucleic acid X-ray crystallography. Besides the facilitation of nucleic acid crystallography, there is also a wide range of other applications for selenium-derivatized nucleic acids (SeNA). The investigation presented in this dissertation mainly focuses on the following research subjects (1) Synthesis and characterization of selenium-derivatized nucleic acids for X-ray crystallography, especially phosphoroselenoate RNAs. They are generated and used for crystallization. (2) Application of selenium-derivatized RNA for RNA interference. Phosphoroselenoate RNAs are tested for RNAi activities. (3) Synthesis and characterization of the uridine 5’-triphosphate modified with selenium at position 4. (4) Facile synthesis and antitumor activities of selenium modified deoxyribonucleosides. MeSe-thymidine nucleosides have shown antitumor activity in cell assays.

Philosophiae Doctor - PhD / Cancer is one of the leading causes of death worldwide. According to the WHO, cancer accounted for 7.4 million deaths world wide in 2004. The metallo-compound cisplatin has been used for years as an effective antitumor agent for treating solid tumours such as breast, bladder, lung, oesophageal, and head and neck carcinomas. However, the use of cisplatin as an antitumor agent has been limited because of its association with problems such as lack of selectivity for cancer cells over normal cells, development of resistance to cisplatin treatment, and side effects such as nephrotoxicity. Recent studies on anticancer drugs have focussed on alternative anticancer agents such as gold compounds in both Au(I) and (III) oxidation states, which have shown to be potential anticancer drug agents because of their ability to induce apoptosis in several human cancer cells. Some gold complexes have shown to be able to selectively kill cancer cells over normal cells.

Animal testing is still the common method to test the efficacy of new drugs, but tissue engineered in vitro models are becoming more acceptable for replacing and reducing animal testing in anti-cancer drug screening by developing in vitro three-dimensional (3D) tumour models for anti-cancer drug testing. In this study, three-dimensional (3D) culture methods were developed to mimic the tumour microenvironment. 3D culturing is to seed, maintain and expand cultured cells in three-dimensional space, in contrast to the traditional two-dimensional (2D) method in which the cells attach to the bottom of culture containers as monolayers. To mimic the intercellular interplay for tumour study, cell co-culture was applied. In this thesis, perfusion culture showed a better homeostasis for 3D tumour model growth over 17 days, with a more controllable working platform and a more reliable response-dose correlation for data interpretation. In the Matrigel sandwich system, the co-culture of breast cancer cells and endothelial cells demonstrated the morphology featuring a vascular network and tumour structures, with the thickness of the three-dimensional structure around 100µm and tubule length 200-400 µm, and maintained for 10 days. The comparisons studies between Matrigel sandwich and other methods suggest that though not fully characterised, Matrigel is still a valuable scaffold choice for developing co-culture 3D tumour model. Finally, the combination of perfusion and co-culture showed the potential of applying this model in angiogenesis assay, with a drug response profile combining cell viability and morphology to mimic in vivo tumour physiology.

Despite significant progress in the development of anti-cancer drugs, there is still a need for novel therapeutic strategies that would improve the outcome of cancer patients. Using proteomic technologies and cell lines with different phenotype of p53 tumour suppressor, we monitored cancer cell response to anti-cancer treatment with focus on the development of drug resistance. The different levels of metabolic proteins were identified in our study which may help to explain different anti-cancer activity of drugs with only a subtle difference in structure. More importantly, proteins associated with the development of drug resistance were identified and such expression changes have become a focus of interest. Our findings demonstrate a higher protein level of serine hydroxymethyltransferase, serpin B5 and calretinin in cancer cells resistant to Aurora kinase inhibitors. Such proteins promote the tumour growth with no apparent impact of p53 phenotype whilst voltage-dependent anion-selective channel protein 2 contributes to the development of resistance only in cells with functional p53 which is accompanied by the decreased level of elongation factor 2. On the other hand, cancer cells with loss of p53 appear to amplify alternative mechanisms such as protection against oxidative stress. The results...