<p>Bacterial nitroreductases are members of a diverse family of oxidoreductase enzymes that can catalyse the bioreductive activation of nitroaromatic compounds, including anti-cancer prodrugs such as CB1954, SN36506 and nitro-CBI-DEI, and antibiotic prodrugs such as metronidazole. These enzymes have diverse applications in medicine and research, including the anti-cancer strategy gene-directed enzyme-prodrug therapy and targeted ablation of nitroreductase-expressing cells in transgenic zebrafish to model degenerative disease. However, research in these fields to date has focused almost exclusively on the canonical nitroreductase NfsB (from Escherichia coli), which is a relatively inefficient nitroreductase. This thesis presents work aiming to develop superior nitroreductases as tools for targeted cellular ablation and cancer gene therapy through enzyme engineering. </p>
<p><br></p>
<p>For targeted cellular ablation, this research employed rational and random mutagenesis to engineer nitroreductase variants exhibiting substantially improved metronidazole reductase activity when expressed in both bacterial and human cell lines. One such variant was found capable of achieving robust ablation in a transgenic zebrafish model at a 100-fold reduced metronidazole concentration compared to NfsB. To expand the utility of this ablation system, this research additionally identified pairs of nitroreductases with non-overlapping prodrug specificity suitable for use in a ‘multiplex’ system of cellular ablation. It was shown that targeted expression of either nitroreductase ‘partner’ combined with selective administration of its preferred prodrug could enable specific ablation of either of two different cell populations within a mixed culture. Trialling of this system in zebrafish models is ongoing.</p>
<p><br></p>
<p>For cancer gene therapy, directed evolution was carried out to develop a repertoire of candidate nitroreductases having improved abilities to activate a next generation nitroaromatic prodrug, for use in clostridia directed enzyme prodrug therapy (CDEPT). Generation and screening of a targeted site-saturation mutagenesis library identified several variants that displayed improved prodrug activity in vivo. Interestingly, these improvements were not reflected in in vitro purified protein kinetic analyses. Further investigation suggested that an inadvertent driving force behind the in vivo selection of directed evolution libraries in E. coli was the evolution of variants towards heightened specificity for the prodrug of interest over competing intracellular metabolites, rather than the expected improvements in catalytic efficiency. Ongoing evaluation of the top CDEPT candidate nitroreductases will involve assessment of their therapeutic effect in preclinical models when expressed by Clostridium sporogenes. In addition, this research sought to investigate an alternative, novel approach to gene therapy involving the use of CAR-T cells as gene therapy vectors. To achieve this, nitroreductases were engineered to extend the cytotoxic reach of the therapy through activation of high-bystander anticancer prodrugs, while also introducing a metronidazole-mediated kill-switch functionality that would enable targeted ablation of administered CAR-T cells within patients if required. Validation of lead nitroreductases was achieved in cultured human cell models. This research also explored the potential for a bacterial DNA glycosylase to defend nitroreductase-containing cells against the cytotoxic effects of reduced prodrug metabolites.</p>
<p><br></p>
<p>Throughout this work, to complement biological assay data, crystallographic analysis of a selection of evolved nitroreductase variants was undertaken to elucidate the structural bases for the improved prodrug activities observed in vivo.</p>