Mammalian cells have evolved highly regulated defence pathways, which are activated in response to stress. The transcription factor Nrf2 is activated in response to chemical and oxidative stress and induces the expression of antioxidants and detoxification enzymes. Nrf2 activity is regulated by its cysteine-rich repressor protein Keap1, which facilitates Nrf2 degradation. In the presence of electrophilic compounds and/or oxidative stress, Keap1-mediated repression of Nrf2 is hindered, allowing the nuclear localisation of Nrf2 and the up-regulation of cell defence gene expression. The dysregulation of the Nrf2 pathway has been associated with many disease pathologies, and is a promising therapeutic target. Furthering our understanding of the chemico-biological triggers for Nrf2 activation may inform the design of novel Nrf2-inducers with increased efficacy and reduced toxicity. The modification of one or more cysteine residues in Keap1 by electrophiles is believed to be central to Nrf2 activation. Therefore, the aims of the studies in this thesis were to investigate the chemical modification of Keap1 cysteine residues by Nrf2-inducers and identify novel Keap1 binding proteins, which may play a role in the regulation of Nrf2 activity. Previous work carried out in this lab identified cysteine residues in Keap1 that are covalently modified by a panel of Nrf2-inducing compounds in recombinant Keap1 protein and in cells. Additionally, Nrf2 can be induced by glutathione (GSH) depletion in the absence of Keap1 adduct formation. We hypothesised that GSH depletion permits reactive oxygen species (ROS) to accumulate and oxidise Keap1 cysteine residues, thereby inducing Nrf2. To address this, we treated Keap1-V5 expressing HEK293T cells with Nrf2-inducing compounds. We used liquid chromatography tandem mass spectrometry (LC-MS/MS) to investigate the ability of these compounds to form adducts with Keap1 cysteine residues, or induce reversible/irreversible redox modifications of these residues. We show that compounds which form covalent adducts with Keap1 and deplete GSH (i.e. 2,4-dinitrochlorobenzene) or do not deplete GSH (i.e. dexamethasone 2,1-mesylate) induce modifications of Keap1 that could be representative of oxidation. However compounds which do not form covalent adducts with Keap1 but cause oxidative stress (i.e. hydrogen peroxide), or GSH depletion (i.e. L-buthionine-sulfoximine), do not appear to cause oxidative-like modification of Keap1 cysteines. We therefore show that some Nrf2 inducers promote the formation of reversible and/or irreversible redox modifications of Keap1 which could be due to thiol oxidation, although this is not dependent on GSH depletion. To further explore the modification of Keap1 cysteine residues by Nrf2 inducers, we investigated the ability of triterpenoids (TPs) to modify Keap1. TPs, in particular methyl 2-cyano-3,12-dioxooleana-1,9(11)dien-28-oate (CDDO-Me), are potent inducers of Nrf2 and are potential therapeutic agents. However, the mechanism of Nrf2 activation by TPs has not been fully elucidated using LC-MS/MS. We identify key cysteine residues in Keap1 which are adducted by a chemically-tuned TP (CDDO-Epoxide) in recombinant Keap1 and in cells expressing Keap1-V5. Additionally, we use an in silico modelling approach to visualise the binding orientations of CDDO-Epoxide with key Keap1 cysteine residues. Correlating the potency of a panel of TPs towards the Nrf2 pathway with their in silico propensity to bind covalently to the identified residues showed no relationship. However, we show significant positive correlation between the potency of these TPs towards Nrf2 and their in silico propensity to bind non-covalently in two cysteine-containing pockets (Cys-273, -288) in Keap1. These data reveal the specific sites of interactions between potent TP Nrf2 inducers and Keap1, and highlight the non-covalent binding of Keap1 by electrophiles as a potential mechanism of Nrf2 activation. The function of Keap1 is regulated by interactions with binding partners, such as sequestosome1 (p62) which targets it for autophagic degradation, or PGAM5 which localises Keap1 to the mitochondria. We previously used an LC-MS/MS approach to identify p62 as a novel Keap1-binding partner in cells. Therefore, we reasoned that using the same experimental approach with a more sensitive MS system, we could identify additional Keap1-binding proteins. Specifically, we identified a large number of proteins that co-purified with Keap1-V5 from HEK293T cell lysates, of which 55 were found to contain a known Keap1 binding motif, such as the one found in Nrf2, p62 and PGAM5. Network analysis highlighted the potential link between Keap1/Nrf2 and the p53 cell survival pathway. We validated the LC-MS/MS data using a yeast 2-hybrid screen, which reveals HBS1L, RIC8A and PSMD3 as novel Keap1 binding partners, although the functional relevance of the interaction of these proteins with Keap1 requires further investigation. In summary, the data presented in this thesis demonstrates that whilst the covalent modification of Keap1 cysteines is an important aspect of Nrf2 induction, the oxidation of Keap1 thiols may be an alternative mechanism. We identify key cysteine residues in Keap1 covalently modified by a potent TP Nrf2 inducer in recombinant protein and cells, but show that non-covalent modification of Keap1 may be involved in the process of Nrf2 activation by this class of compound. It will be important in future studies to determine how the modification of Keap1 cysteine residues is translated to the activation of Nrf2. Additionally, we identify putative novel Keap1 binding partners which may serve to regulate the activity of the Nrf2 pathway. Overall, these findings expand our understanding of the chemical and molecular interactions that govern the activity of Nrf2, and will therefore contribute to the ongoing efforts to target this pathway as a novel therapeutic strategy in numerous diseases.

Malaria is a vector-borne parasitic disease spread by a bite of an infected female Anopheles mosquito that accounts for high morbidity and mortality, mainly in sub-Saharan Africa. Of the five species that can cause malaria in humans, Plasmodium falciparum is regarded the most virulent species. Antimalarial drugs, unaltered for many decades, remain the mainstay for treating P. falciparum infection. In addition, despite intensive research there remain significant knowledge gaps in understanding of the biology of the malaria parasite P. falciparum. These deficiencies hinder the ability of scientists to identify new targets for drug discovery at a time when new targets are urgently required, such as in the case of newly emerging drug resistant parasite strains. Therefore, an understanding of the biology of P. falciparum is helpful in identifying new drug targets. Metabolomics, defined as the comprehensive analysis of all metabolites in a biological system, offers a feasible platform for highly sensitive and specific analysis of the metabolic pathways of P. falciparum. This is supported by the assumption that metabolites are important players in biological systems and that resistant parasite strains may operate or alter single or multiple metabolic pathways in order to adapt to the drugs being used. Therefore, a targeted metabolomics approach was developed and validated (Chapter 3) in order to better understand the metabolic roles of mitochondria and the digestive vacuole of P. falciparum. Metabolite detection and quantification were conducted using a targeted LC-MS/MS metabolomics approach (Chapter 3). It was shown that metabolic activities, particularly carbohydrate metabolism, in trophozoite stage P. falciparum-infected RBC were remarkably higher than that of non-infected RBC (Chapter 4). This lead the study to progress further, examining the metabolic role of two components of the P. falciparum; the mitochondria and the digestive vacuole. A number of mitochondrial inhibitors selective for specific electron transport chain complexes and mitochondrial transporters were used to assess mitochondrial function in asexually growing trophozoite stage P. falciparum (Chapter 5 and 6). Despite the differing modes of action of the inhibitors, the metabolic fingerprints, which were carbamoyl-l-aspartate and dihydroorotate, from these experiments were consistent with the parasite mitochondrion playing a key role in pyrimidine biosynthesis at the point of dihydroorotate dehydrogenase (DHODH) (Chapter 5 and 6). This metabolic fingerprint, leading to parasite death, was quite distinct from fingerprints obtained from biologically distinct inhibitors such as heme-binding drugs (quinoline-containing antimalarials drugs) which primarily affected the metabolism of amino acids, perhaps in the digestive vacuole and parasite cytosol (Chapter 7). In contrast to genomic and proteomic approaches, metabolomics appears to better represent the parasites’ phenotype in response to drug perturbation. Pharmacometabolomics will therefore have significant utility in understanding the biological function of parasite components; and the mode of action, efficacy and toxicity of pharmaceutical drugs.

Artemisinin-derived endoperoxide drugs find widespread employment as frontline treatment against malaria. Although evidence of their potential to express toxicity within a clinical setting remains limited, outcomes derived from animal studies attest their ability to induce neurological and developmental toxicity in mammalian systems. Activity is further demonstrated in vitro within rapidly proliferating human cells – most notably those belonging to immortalised, cancer-derived lines – with significant cytotoxic effects being observed upon drug treatment across a range of settings. It is believed that these find their origin through a mechanism dependent upon Fe(II)-mediated reduction of the endoperoxide bridge functionality, culminating in molecular bioactivation and the subsequent formation of carbon-centred free radical species which in turn, owing to their great reactivity, impart deleterious effects upon cellular functioning. Evidence suggests that dysfunction of the mitochondrion and the formation of reactive oxygen species (ROS) are key stages in the route through which artemisinin derivatives are able to induce death. The characteristics of artesunate-stimulated impact upon mitochondrial functioning are examined. It is demonstrated that culturing of cells in the presence of galactose enhances cytotoxic potential within the HeLa line. The magnitude of this variation in sensitivity is indicative that targeting of the mitochondrion affords a route through which activity is mediated. Falls in cell viability are further preceded by declines in ATP production, providing evidence that disruption of oxidative phosphorylation occurs as an early event in the route towards death. Studies performed on mitochondrial bioenergetic function using the Seahorse XF analyser indicate that artesunate imparts dose-dependent and timedependent decreases in respiratory reserve capacity and oxidative phosphorylation coupling efficiency, whilst stimulating a switch towards glycolytic energy production. Attempts to delineate the root causes of these effects are focused upon examining the relationship between oxidative stress, Fe(II) content and mitochondrial performance. The mitochondrially-localised antioxidant tiron and the lysosomal Fe(II) chelator desferrioxamine are shown to induce substantial cytoprotective effects against artesunate within the HeLa line. Evidence derived from Seahorse XF analysis indicates strongly that these outcomes are related to the capacity of both compounds to abrogate drug impact upon the functions of the mitochondrion. It can thus be posited that mitochondrial damage has its origins in the emergence of oxidative stress, with Fe(II) content acting as key determinant in its progression. The outcomes of further examinations performed within the ρ0 HeLa line suggest an origin for ROS emergence independent of the respiratory chain. In order to test the hypothesis that artemisinin derivatives might induce direct peroxidation of the mitochondrial phospholipid cardiolipin, the impact of a cytochrome c peroxidase inhibitor TPP-IOA is examined on the response of HeLa and HL-60 cells towards artesunate treatment. Results indicate that the inhibitor has variable effects upon cardiolipin oxidation state and subsequent cell survival, leaving doubt as to the true validity of such a connection. As a further study, the cytotoxic capacities of a range of novel artemisininderived anticancer agents and wholly synthetic tetraoxane and trioxolane antimalarials are given assessment. In conclusion, it can be stated that the outcomes of the studies performed in this thesis emphasise the importance of mitochondrial liability towards the progression of artemisinininduced cell death. Further insights into the mechanistic routes through which drug administration contributes, via oxidative stress and free Fe(II) content, to the defective functioning of the organelle have been achieved.

Adverse drug reactions (ADRs) are a major obstacle for the development of new medicines. They are also a leading cause of patient morbidity and mortality. Although ADRs affect many different organs and bodily systems, drug induced liver injury has lead to the withdrawal of several drugs at the post licensing stage, and is a key cause of drug attrition. Many of the drugs that cause liver injury are thought to do so through metabolism to a reactive metabolite, exposure to which can cause modification of cellular proteins, leading to loss of function, and can result in a loss of cellular homoeostasis. It is therefore important to understand the chemistry and the downstream biochemical events associated with bioactivation. Information on the chemistry of metabolism coupled with mechanistic biomarkers reflective of certain pathways of hepatic injury would enable both researchers and physicians to predict and diagnose DILI, leading to the improvement of safe drug design. This thesis focuses firstly on the use of in vitro models and mass spectrometry to provide integrated data on the metabolism and toxicity of xenobiotics, using thiophene containing molecules as a paradigm. The thiophene ring has the potential to be bioactivated via S-oxidation and epoxidation pathways, and several thiophene containing drugs have been associated with drug induced liver injury. The investigations described intended firstly to elucidate the chemistry of methapyrilene bioactivation using mass spectrometry and hydrogen-deuterium exchange. The following two chapters aimed to establish a link between bioactivation and toxicity of thiophenes and to evaluate two in vitro models as tool for predicting DILI. The final experimental chapter aims to investigate the potential of ophthalmic acid (OA) to act as a serum biomarker reflective of depletion of hepatic levels of the protective tripeptide, glutathione (GSH). Disturbance of GSH levels through quenching of reactive metabolites can lead to disturbance of its anabolism and catabolism pathways. Indeed, serum OA levels, a GSH analog, have been shown to rise following hepatic GSH depletion. This work utilises GSH adduct formation as a marker of bioactivation of thiophenes in several different in vitro models. Rat liver microsomal incubations were analysed using hydrogen deuterium exchange and LC-MS to define the reactive metabolite of methapyrilene as an S-oxide of the thiophene ring. Freshly isolated rat hepatocytes or single P450 expressing THLE cell cultures were exposed to either methapyrilene, tienilic acid, ticlopidine or 2-phenylthiophene and subsequent LC-MS analysis confirmed GSH adduct formation for all compounds in the isolated rat hepatocyte model, but only for 2-phenylthiophene in the THLE cell model. Cytotoxicity was also investigated in both models, and all compounds were found to cause a greater degree of toxicity in the isolated rat hepatocyte molecule than in the THLE model. By exposing rodents to depletors of hepatic GSH, such as acetaminophen and diethylmaleate, and monitoring the resultant serum OA levels, it has been determined that OA is not a reliable mechanistic marker of hepatic GSH depletion. Kinetic studies of OA in rat serum have revealed that OA is subject to a similar metabolic and elimination pathway as GSH. The overall scope of this work reveals the usefulness of LC-MS/MS to determine S-oxide and epoxide adducts in in vitro studies. The freshly isolated rat hepatocyte model was a useful tool for providing integrated metabolic and toxicological data of thiophene containing molecules and has the potential to be expanded to include data on covalent binding and levels of DILI biomarkers. The single CYP expressing THLE cell model was not as useful in this case, but has been used in other studies to explore the role of discrete P450 enzymes in toxicity and metabolism. Whilst it is unfortunate that serum OA did not reflect hepatic OA in such a way that it could be easily exploited as a biomarker, this does help us to understand that the plethora of potential biomarkers uncovered by proteomic, metabolomic and transcriptomic studies need to be investigated in depth in order to understand their applications across different species and systems.

Adverse drug reactions (ADR) are a major concern for the pharmaceutical industry and health practitioners as they can cause morbidity and in severe cases mortality. ADRs are one of the major reasons why drugs fail during clinical trials so research directed at predicting ADRs to minimise failure is essential. The CDDO (2-cyano-3,12-dioxo-oleana-1,9(11)-dien-28-oate) and the synthetic endoperoxide series are two promising classes that have potential for the treatment of cancers and malaria and may revolutionise treatment, within their fields, if approved for clinical use. The two main aims that are presented in this thesis are to; (i) design and synthesise novel analogues and chemical probes to identify potential molecular targets for both the CDDO and endoperoxide series (ii) develop appropriate in vitro test systems to help define the molecular mechanism of each class of drug. CDDO-Me (methyl 2-cyano-3,12-dioxo-oleana-1,9(11)-dien-28-oate) is one of the most potent inducers of Nrf2, a transcription factor that regulates the expression of numerous cell defence genes in mammalian cells. Nrf2 is sequestered in the cytosol by Keap1, which ‘senses’ chemical and oxidative stress via its 27 cysteine residues. Although CDDO-Me is one of the most potent inducers of Nrf2, the molecular target and chemical mechanism is still not defined. Current literature suggests that a reversible 1,4 conjugate addition to specific cysteine residue(s) located on the Keap1 protein results in an increase in Nrf2 levels. In order for SAR work to be performed a synthetic route to CDDO and analogues was developed which involved nine steps using oleanolic acid as starting material. Highlights of the chemistry included addition of the ketone using mCPBA and incorporation of the cyano group in steps 3 and 7 of the synthesis. In addition to preparing the target molecule CDDO a number of additional molecules were prepared to define the importance of functional groups in the A and C rings of CDDO. Genetically modified H4IIE rat hepatoma cells transiently transfected with the an Nrf2-sensitive luciferase reporter gene were used to screen the CDDO-Me analogues, including DDO-Me which lacks a cyano group on the A ring, for their ability to induce Nrf2. NMR studies with model thiols were performed to determine the ability of these compounds to form reversible or non-reversible adducts. Mass spectrometry (MS) was used to confirm the NMR data and interpretations. In total, four probes were identified that reacted in a non-reversible fashion: DDO-Me, DDO-Al and DDO-Az (click probe versions of the parent DDO-Me that can be used to facilitate proteomic studies) and CDDO-Epox (a probe with similar overall structure to CDDO-Me but can react at the β-carbon in a non-reversible fashion; this feature should aid proteomic approaches to reactive cysteine residue identification). To further investigate if these compounds were reactive to cysteine residues within a model protein, recombinant human GSTP1 was used as a model protein for chemically reactive molecules. Cys-47 located on GSTP1 has been shown to react with other electrophones and during our studies LCMS has confirmed that all four of the synthesised active probes were capable of attaching covalently to Cys-47 of GSTP1. The emergence of malaria parasite resistance to most available drugs, including the semi-synthetic artemisinin derivatives artemether and artesunate, has led to efforts to create new synthetic peroxides as potential antimalarial agents. Leading examples of synthetic endoperoxides include OZ277 (arterolane), a molecule in phase III clinical trials in combination with piperaquine, and OZ439, a second generation derivative with improved pharmacokinetics and enhanced in vivo antimalarial activity. 1,2,4,5-Tetraoxanes are another class of endoperoxide with proven excellent antimalarial profiles against both chloroquine-resistant and chloroquine-sensitive strains of Plasmodium falciparum and oral activity in murine models of the disease. It is currently widely accepted that endoperoxides have a similar antimalarial mechanism to artemisinin, whereby Fe2+ medicated generation of cytotoxic carbon-centred radicals, results in death of the parasite. It is presumed that C-radicals can react with important key proteins; however, the specific molecular target(s) that leads to eventual parasite death are still unknown. A chemical synthesis of tetraoxane probes that contain a UV chromophore was performed and analogues were subsequently screened for antimalarial activity. The most active tetraoxane identified was exposed to a range of Fe2+ salts and conditions developed to mimic the biological environment. Primary, secondary and novel carbon-centred radical derived products (surrogate markers of bioactivation) were purified using UV-HPLC, characterised and submitted as chemical probes and standards for biological studies. In order for proteomic studies to be initiated, an allyl or azide group was incorporated into a semi-synthetic artemisinin skeleton. The azide (and alkyne) functional group within these probes provides a handle for protein pull down via click chemistry. Azide and acetylenes were chosen over direct linkage to the biotin group to reduce steric hindrance in the semi-synthetic probe. The synthesised click probes were tested for antimalarial activity and were submitted for protein pull down and identification of potential molecular targets. Similarly DDO allyl and azide were synthesised and were tested for Nrf2 induction and further confirmed as viable probes via NMR experiments with simple thiols and GSTP1. In summary, novel CDDO non reversible probes were synthesised and have shown potential as chemical tools to identify the molecular targets/mechanisms by which these compounds activate Nrf2. Tetraoxanes also have been prepared along with artemisinin click probes and the latter have been submitted for click chemistry pull down experiments, within Plasmodium falciparum parasites, to identify potential molecular targets.

Immunomodulatory monoclonal antibodies (mAbs) indicated for the treatment of cancer, inflammatory and autoimmune diseases or to prevent organ transplant rejection, mostly target cell surface molecules expressed by immune cells including T cells. The presence of immunosuppressive molecules during disease progression can hinder the activity of immune cells. Tumour-induced immunosuppression enables disease progression and various strategies are being developed to enhance the anti-tumour responses of cytotoxic T cells. Activation of effective effector responses require resetting of metabolic activity to fit energy and anabolic needs. While the benefits of exploiting immunomodulatory mAbs for therapy are substantial, it is also clear that use of these biologics may be accompanied by adverse effects such as cytokine release, immunosuppression, infections and autoimmunity. Significant focus in recent times has been on assessing the potential of immunomodulatory mAbs to induce enhanced cytokine release but much less attention has been paid to other aspects of T cell biology including non-physiological activation phenotype/functions, migration characteristics and metabolism. An improved understanding of these parameters may assist in accurately predicting the propensity of new mAbs to induce serious adverse effects. Superagonistic CD28-specific monoclonal antibody (CD28SA) is one such immunomodulatory monoclonal antibody which is a potent stimulator of T cells, originally intended for the management of B cell chronic lymphocytic leukaemia and rheumatoid arthritis. Human volunteers who received humanized CD28SA (TGN1412) as part of a first-in-man trial experienced life-threatening cytokine release syndrome. Follow-up studies revealed aberrant activation of effector memory T cells (TEM) contributed towards the adverse reaction. The biopharmaceutical industry is actively pursuing development of T cell immunostimulatory mAbs and there is a significant need to improve the understanding of and accurately predict the propensity of a new mAb to drive excessive T cell activation. Therefore, one of the main aims of the study discussed in this thesis was to determine mechanism/s underlying the hyperactive phenotype of CD28SA-activated TEM. Observations in the current study revealed activation of TEM by CD28SA upregulated the expression of activation markers such as CD137 and HLA-DR, but failed to express co-inhibitory receptor, PD-1. This led to the lack of PD-1-mediated regulation of aberrant TEM activation. In addition, CD28SA-activated TEM expressed elevated levels of LFA-1 and CCR5 receptors, and displayed increased migratory capacity. Subsequent studies highlighted increased metabolic demand of CD28SA-activated TEM. The hyperactive cells with increased proliferative capacity exhibited distinct metabolic profile characterized by increased glycolysis and lipogenesis. These findings have profound implications for strategies aiming at understanding and predicting the safety profile of immunostimulatory mAbs. Deployment of immunosuppressive strategies, including TGF-β secretion by tumours to render immunostimulatory mAbs-mediated anti-tumour responses ineffective is well studied. The other study discussed in this thesis aimed to delineate the effects of oxidative stress in TGF-β-induced suppression of antigen-specific cytotoxic T cell responses. This study showed antigen-specific T cells exposed to TGF-β down regulate CD25 and LAG-3 (co-inhibitory T cell receptor) expression, secrete lower levels of IL-2 and IFN-γ, and reduced glycolysis. In addition, mitochondrial reactive oxygen species (MitoROS) scavenger rescued the effector functions such as proliferation and IFN-γ secretion of stimulated T cells that were inhibited by TGF-β. Our findings demonstrate that relief of TGF-β-induced oxidative stress restores the effector function of CD8+ cytotoxic T cells. Based on the current findings it would be potentially beneficial to supplement immunotherapies with antioxidants to counteract the immunosuppressive effects of tumour-derived TGF-β to help restore CD8+ T cell-mediated anti-tumour function. The findings presented in this thesis may help with defining key T cell biomarkers based of efficacy and hazard associated with immunomodulation.

Adverse drug reactions (ADRs) are responsible for an increasing number of hospitalised patients, with the large majority of these ADRs classed as either type A or type B. Drug hypersensitivity reactions fall within the type B category and one such drug responsible for this form of ADR is abacavir (ABC). ABC, a nucleoside reverse transcriptase inhibitor, is used to treat the HIV-1 virus. It is responsible for a potentially life-threatening type IV hypersensitivity reaction which occurs in patients bearing the HLA-B*57:01 allele. Although many mechanisms have been proposed, it was the objective of this research to examine all these previous proposals to further extend and develop the mechanism of ABC toxicity. In Chapter 2, deuterated-ABC (D2-ABC) was designed and synthesised where the two 5'-H atoms were replaced with two 5'-D atoms. The design of this analogue was intended to retard the oxidative metabolism of ABC to its aldehyde and carboxylic acid metabolites. The synthesis of this compound was paramount to investigating this metabolism and through a series of metabolic experiments, described in Chapter 3, a kinetic isotope effect between ABC and D2-ABC was determined, ultimately showing an altered metabolism between the two compounds. To investigate binding of ABC within the HLA-B*57:01 protein, analogues of ABC, with alterations at varying positions within the molecule, were required. Using a racemic method, ABC enantiomers were synthesised and ABC’s enantiomer failed to stimulate T-cells in vitro. The creation of further analogues required the development of an asymmetric synthetic route. A total synthetic method was desired to synthesise intermediates to be used in future analogue synthesis. Finally, as described in Chapter 5, a range of 6-position analogues were designed, using a structure-activity relationship method, and synthesised, to further investigate the altered repertoire mechanism. These analogues, consisting of primary and secondary amine and oxy moieties, were subjected to in vitro immunological assays to determine their stimulation of T-cells. Additionally, the synthesised analogues were modelled in silico using molecular docking within the HLA protein. The in silico results assisted in explaining the basis of such T-cell activation/inactivation and will direct future analogue design. IC50 and EC50 values were determined for analogues that presented a negative T-cell response and a compound showing positive values was subjected to further pharmacokinetic testing. The oxidative metabolism of ABC was affected by isotopic substitution, but initial results have shown no altered T-cell stimulation of D2-ABC compared to ABC. This mechanism cannot be discarded, with further investigational work required. However, the synthesised 6-position analogues have assisted in further examining the altered repertoire mechanism and initial findings have enabled further understanding of the binding of ABC within HLA-B*57:01. This mechanism of ABC toxicity appears paramount to others proposed and the results presented in this thesis support this. Additional analogue synthesis and in vivo experiments will assist in confirming this further.

Malaria is one of the most prevalent and deadliest parasitic diseases affecting various systems of the body and leading to death. Resistance against antimalarial treatment is a major threat in controlling and eliminating malaria. New drugs are urgently needed especially when artemisinin resistance has emerged. The mitochondrial electron transport chain of Plasmodium falciparum is an attractive target for chemotherapy. Two enzymes in the pathway - Pfbc1 and PfNDH2 - are druggable target enzymes. The dual inhibition of both enzymes can be seen in 2-aryl quinolone pharmacophore giving added therapeutic benefit. The development from this series leads to the potent lead compounds including SL-2-25 and PG227. In Chapter III, following the hit-to-lead optimisation of SL-2-25, a 5-7 step synthesis of a library of 2-aryl quinolones has been described. In vitro antimalarial assessment of these quinolones revealed the advantages of the 7-methoxy moiety. The potency increases 3-8 folds when the 7-OMe group is attached. Further lead modification led to a more flexible quinolone 61i retaining high potency against the 3D7 strain of P. falciparum. This structure also possesses no cross resistance, greater aqueous solubility and low potential for cardiotoxicity. Following a similar study on related quinolones, 3,4-dichlorophenyl analogues were briefly investigated. This led to the discovery of 61o possessing an outstanding potency against 3D7 strain of P. falciparum of 18 nM. It also shows low cardiotoxicity when compare to other quinolones. 61u featuring 6-Cl and 7-OMe substitution was identified with an in vitro IC50 potency of 9 nM against Plasmodium. In silico molecular modelling based on the yeast bc1 protein complex shows that all quinolones bind tightly to the target protein with essential interactions in place. PG227 (69) exhibits outstanding pharmacological properties amongst the series of quinolones. Its original synthesis suffers from reproducibility and low overall yields. 69 can be made in a multi-gram scale using an alternative method for cyclisation. The 5-step synthesis of PG227 can be achieved from commercially available starting materials involving the synthesis of β-keto ester intermediate, the Conrad-Limpach cyclisation and chlorination using NCS. The overall yield was 7%. Artemisinin combination therapy (ACT) is used as the first line treatment in most of the malarial endemic areas. The emerged artemisinin resistance requires greater understanding of drug action. In Chapter V, activity-based protein profiling (ABPP) was employed to identify the molecular target of artemisinin for the first time. The novel “tag-free” ABPP proteomic technique is introduced based on the click chemistry between a chemical probe and a reporter tag. The synthesis of the artemisinin-based ABPP chemical probes was achieved. The peroxide-containing probes show an excellent in vitro potency against the 3D7 malaria parasite. The preliminary result reveals that active probe 99 can perform well in protein pull down resulting in 45 different proteins being identified.

Carboxylesterase 1 (CES1) is the main human liver esterase and is involved in the metabolism and disposition of numerous endogenous and pharmacological compounds. Some of the substrates of this enzyme are widely prescribed agents such as clopidogrel (Plavix®), methylphenidate (Ritalin®) and oseltamivir (Tamiflu®). However, there is much uncertainty regarding the genetic variability within CES1, and its regulation and involvement in drug-drug interactions (DDI). Polypharmacy is frequent in elderly, HIV and tuberculosis infected populations, and the risk of harmful DDIs is high, especially when these populations overlap. The role played by CES1 on the treatment of all these three pathologies and vice versa needs to be better characterized. In this thesis the role of CES1 genetic variability and its potential role in DDIs are explored both in isolation and in conjunction with other genetic, demographic, physio-pathological and iatrogenic factors. The impact of CES1 genetic variability was assessed on the anti-platelet effect of clopidogrel as well as on isoniazid pharmacokinetics in acute coronary syndrome (ACS) and HIV/Tuberculosis co-infected populations respectively. DDIs mediated by CES1 were explored in a HIV positive cohort treated with clopidogrel and non-nucleoside reverse transcriptase inhibitors (NNRTIs). Also, in vitro experiments with primary hepatocytes were used to investigate CES1 intracellular expression in the presence of prototypical PXR inducers used in tuberculosis treatment. The results of this thesis show that the CES1 rs2244613 SNP does affect clopidogrel anti-aggregant activity and may contribute to treatment non-response. Another CES1 variant, rs3815583, was found to be associated with changes in isoniazid pharmacokinetics. The studies did not indicate that NNRTI coadministration with clopidogrel would impair the anti-platelet activity since no relevant changes in exposure of the antiplatelet agent were identified. In the same way, the results do not anticipate DDIs between CES1 substrates and rifamycins, since no induction of expression was identified after incubating primary human hepatocytes in vitro with rifampicin, rifabutin and rifapentine. In conclusion, the results shown in this thesis support the idea that CES1 genetic variability may play a bigger role than previously suspected in treatment response but may not be a mediator of clinically relevant DDIs.

Drug-induced liver injury (DILI) remains a prominent cause of patient morbidity and mortality, partly due to the lack of relevant in vitro hepatic models for accurate screening for drug-induced hepatotoxicity at the early stages of drug development, and also the lack of sophisticated in vitro model systems to mechanistically understand the pathways that are perturbed following drug exposure. This thesis describes our endeavour to develop more relevant in vitro human hepatocyte models via novel investigative approaches using insights gained from the rapidly advancing research areas of human induced pluripotent stem cells and microRNAs (miRs). An emerging hepatic model is hepatocyte-like cells (HLCs) generated from human induced pluripotent stem cells (hiPSCs), though the functional phenotype of HLCs in general remains limited in comparison with the gold standard in vitro model of human primary hepatocytes (hPHs). As studies have shown that hiPSCs retain transient epigenetic memories of the donor cells despite cellular reprogramming with a resultant skewed propensity to differentiate towards the cell-type of origin, we evaluated the contribution of epigenetic memory towards hepatic differentiation by comparing HLCs generated from hPH- and non-hPH-derived hiPSC lines derived from a single donor. Our findings suggested that they were functionally similar, although comparison using hiPSC lines derived from other donors is still required to be conclusive. Although hPHs remain the gold standard in vitro model for DILI, they are commonly harvested from liver tissue of poor quality and rapidly lose their in vivo phenotype during extended in vitro culture, limiting its utility to acute toxicity studies only. Using an unbiased miR expression profiling approach, we identified a set of differentially-expressed miRs in dedifferentiating hPHs which are associated with many of the previously delineated perturbed pathways and biological functions. However, validation experiments are now required to confirm our findings from the bioinformatics analyses. Another approach taken to develop relevant and functional hepatic models includes efforts to better emulate the in vivo liver tissue environment by using complex hepatic models co-cultured with non-parenchymal cells. However, for the application of these models in the study of drug-induced toxicity, a hepatocyte-specific marker of hepatocyte perturbation is needed to discriminate non-specific cellular toxicity contributed by non-hepatocyte cell types present within the model. We demonstrated that the detection of miR-122 in cell culture media can be applied as a hepatocyte-enriched marker of toxicity in heterogeneous cultures of hepatic cells. In summary, this thesis describes our contribution towards the continuing efforts to develop new and improve on existing hepatic models for DILI by evaluating the contribution of epigenetic memory towards the functional phenotype of HLCs, delineating the changing miR profile of dedifferentiating hPHs, and introduced the concept of using miR-122 as a cell-type specific marker of hepatocyte perturbation with a potential to bridge in vitro and in vivo findings.