Ischaemic heart disease remains the leading cause of death worldwide despite significant advances in diagnosis and treatment. Over the past two decades, interventional and medical therapies have been tested through rigorous large scale randomized controlled trials. These data form the backbone of national and international guidelines. However, there are still numerous controversial issues in the short- and long-term management of patients with acute coronary syndromes (ACS) and there are limitations in translation of scientific evidence into clinical practice. Using a well-established multi-centre interventional cardiology registry, this thesis focuses on contentious issues in the management of ST-elevation myocardial infarction (STEMI) and non-ST-elevation acute coronary syndromes (NSTEACS). The topics addressed range from the safety of pretreatment with dual antiplatelet therapy in STEMI and NSTEACS, to the impact of door-to-balloon time in high-risk patients and the long-term prognostic significance of optimal medical therapy and smoking cessation. In addition, single-centre studies were used to explore in detail the management of patients older than 85 years with acute coronary syndromes and the impact of periprocedural myocardial infarction on long-term mortality. Finally, through a randomized controlled trial of a smartphone based cardiac rehabilitation program, this thesis underscores the value of an innovative model to optimize secondary prevention of coronary artery disease.

Restrictive lung disease can be divided into intrathoracic restriction due to parenchymal causes and extrathoracic restriction due to chest wall disease or neuromuscular disease. Patients with extrathoracic lung restriction suffer from shortness of breath, disturbed sleep and eventually type two respiratory failure when their condition progresses. Type two respiratory failure is characterised by hypoxaemia and hypercapnia. Effective treatment of this type of respiratory failure involves ventilatory support in the form of assisted mechanical ventilation. Non-invasive positive pressure ventilation has been a major advance in delivering mechanical ventilation since the late 1980’s. This can be in the form of volume-cycled ventilation or pressure-cycled ventilation. Pressure-cycled ventilation is more commonly practised in Australia and around the world; particularly Europe; according to large multicentre surveys. This can be delivered in spontaneous mode (S) or spontaneous timed mode (ST). There is no current evidence that one mode is superior to the other, and the two modes are roughly equally practised around the world and in Australia according to the available surveys. The aim of this research was to compare S mode to ST mode in a double-blind cross over design over six weeks in 20 patients with stable extrathoracic lung restriction from chest wall and neuromuscular disease. The study aimed to objectively assess adequacy of ventilation, quality of sleep and impact on the patient in psychomotor function, sleepiness as well as patient preference. This was assessed by performing blood gas analysis, sleep studies, psychomotor vigilance testing (PVT) as well as patient questionnaires. We also performed a physiological arm of the study in a subset of patients, assessing the participants for ineffective efforts and poor triggering by inserting an oesophageal balloon and monitoring for markers of effort not followed by a ventilator breath. The oesophageal balloon-catheter system was calibrated for frequency response under different conditions prior to starting this arm of the study and for adequacy of signal acquisition before each sleep study. Due to difficulty in recruitment of participants and some patients withdrawing or failing the safety criteria for the study, a total of 13 patients completed the two arms of the study. The results of this research demonstrated that ST mode was associated with improved nocturnal oxygenation, sleep quality and fewer respiratory events than S mode in patients with stable chest wall and neuromuscular disease. There was no significant difference in daytime hypoventilation as assessed by carbon dioxide measurements. The physiological arm of the study showed that ST mode resulted in fewer ineffective efforts than S mode, the former contributing to better sleep efficiency and correlating well with fewer respiratory events. On the other hand, the mode of ventilation had no significant effect on poor triggering of the ventilator. The participants appeared to be less sleepy on ST than S mode; however, neither mode was associated with pathological sleepiness. There was no significant difference in patient preference between the two modes, nor was there a significant difference in psychomotor function between the modes as assessed by PVT. In conclusion, ST mode of non-invasive positive pressure ventilation was superior to S mode in many aspects of ventilation, sleep quality and impact on the patient. Physiologically there were less ineffective efforts on ST mode, which was related to improved sleep efficiency and the frequency of respiratory events while on ventilation. The participants were less sleepy on ST than S mode, however there was no significant difference between the two modes of non-invasive ventilation on on psychomotor function and patients did not prefer one mode over the other. The sample size of this study was small and less than the original target, due to difficulty in recruitment and the need for participants to present to hospital on three different occasions, a considerable expectation in this relatively frail cohort of patients. The concept of measuring oesophageal pressure in awake patients was also a limiting factor in failure to achieve the target number of participants in the sub-study. This would have impacted the power of the study and the overall significance of some of the results. Based on the results of this study, ST mode of non-invasive ventilation should be utilised rather than S mode in patients with extrathoracic lung restriction due to neuromuscular and chest wall disease. Centre specific preferences and cost need to be taken into account when implementing this recommendation. Acknowledging the difficulties with recruitment, larger (possibly multi-centre) studies of longer duration are required to determine long term health benefits such as hospital admissions and mortality in this population.

Malignant Mesothelioma (MM) is an aggressive malignancy of the pleura and other serosal surfaces with limited treatment options. The mainstay of medical treatment is the combination of cisplatin and pemetrexed chemotherapy. Despite initial responses to chemotherapy, nearly all patients will progress. Only recently, the results of the Mesothelioma Avastin Cisplatin Pemetrexed Study (MAPS) demonstrated a statistically The tumour microenvironment represents an important target in mesothelioma given the absence of oncogenic drivers within the tumour itself. Towards this, I have investigated an anti-EGFR monoclonal antibody [ABT-806 (mAb806)] developed by co-supervisor Scott and now licenced to Abbott (AbbVie), which selectively binds an epitope of wild type (wtEGFR) only found in the overexpressed or amplified EGFR, or its oncogenic truncation mutant EGFRvIII, on cancer cell surfaces. The presence of overexpressed wtEGFR that binds the ABT-806 antibody was also found to be present in a number of MM samples. In this thesis, I aim to characterise specific biomarkers in MM, with particular focus on vascular markers and EGFR expression in MM (Chapter 3). The tumour microenvironment that comprises the interface between the tumour and stroma is of particular interest. This thesis focused on EGFR and angiogenesis given the access to novel anti-EGFR antibody drug conjugates. Angiogenesis, assessed by microvessel density (MVD) has previously been reported by several groups to be a poor prognostic factor in MM. The results of an angiogenic and stromal biomarkers analysis in a large mesothelioma patient cohort are consistent with the literature. I identified high Chalkley count of CD31 immunohistochemistry staining (more than or equal to 5) and increased PDGF-CC expression are associated with a poorer prognosis in 326 MM patients. CD31 was found to be an independent prognostic marker in the multivariate analysis. This project seeks to build on previous work establishing these antibodies as novel therapies for cancer and determine their potential in MM, a disease where we have previously established ligand expression. Using our large repository of MM tissues, I have investigated the expression of the selected targets overexpressed/ amplified EGFR to establish the potential for use of these targets in mesothelioma (Chapter 4). I have also created a MM patient derived xenograft (PDX) library via patient samples derived from the institute and our collaborators from Toronto, Princess Margaret Hospital. Via the library of MM PDXs generated during the course of this PhD, I have evaluated EGFR and mAb806 immunohistochemistry to help select specific models to investigate the validity and feasibility of targeting EGFR in MM using novel mAb806 based antibody drug conjugates (ADCs) (Chapter 5). These novel anti-EGFR ADCs were explored in mesothelioma cell line xenografts, then taken to a panel of novel PDX models to further characterise these mAb806 based ADCs to determine if there are any signals of clinical efficacy (Chapter 6). I have demonstrated the efficacy and selectivity of these anti-EGFR compounds with MM cell line MSTO 211H, as well as 2 other PDX models. The cell line and the PDX model which are mAb806 IHC positive had demonstrated significant tumour suppression and therapeutic efficacy with these mAb806 drug conjugates. These findings contribute towards the understanding and development of potential prognostic biomarkers of interest in MM. The data presented in this thesis also provides novel insight into anti-EGFR antibody drug conjugates in mesothelioma and reveal potential targets for development of targeted therapies in this disease where none was thought feasible.

Active treatment options for patients with metastatic colorectal cancer (CRC) have evolved significantly over the last 20 years. During this time, the median overall survival of these patients has approximately doubled. However, the five-year survival of patients with metastatic CRC is still poor, and further research to optimise treatment is required. The use of targeted therapies has contributed to improvements in clinical outcomes in these patients. However, not all patients respond to these drugs, and they are expensive. The central aim of this thesis was to explore ways to optimise the targeted treatment of metastatic CRC. Specifically, mechanisms of response and resistance to two targeted therapies - the BRAF inhibitor vemurafenib and the anti-angiogenic drug bevacizumab, were investigated. Whilst treatment with single-agent BRAF inhibitors has been associated with significant improvements in response and survival outcomes in patients with advanced BRAF mutant malignant melanoma, results of clinical trials of BRAF inhibitors for BRAF mutant metastatic CRC have been unexpectedly disappointing. First, by screening a panel of BRAFV600E CRC cell lines, cell lines which were sensitive and resistant to BRAF inhibition were identified, and these models were used to explore mechanisms of resistance to these drugs. High phosphorylated-epidermal growth factor receptor (p-EGFR) protein expression was identified to be a potential basal biomarker of sensitivity to vemurafenib. In resistant BRAF V600E CRC cell lines, BRAF inhibitor treatment was demonstrated to induce rapid p-ERK reactivation which was not associated with increased p-EGFR expression, suggesting the presence of EGFR-independent mechanisms of inherent resistance. Next, ways to overcome resistance of BRAFV600E CRC to BRAF inhibitors were explored. BRAF mutant CRC cell lines which were relatively resistant to BRAF inhibitors were found to be broadly similarly resistant to other MAPK inhibitors, and this was also characterised by p-EGFR independent reactivation of p-ERK. This suggests that in BRAF mutant CRC, resistance to MAPK inhibitors is tumour- rather than drug-specific, and using combinations of multiple MAPK inhibitors is unlikely to overcome resistance. BRAF inhibitor treatment did not induce preferential RAS activation, increased p-CRAF expression, nor increased p-MEK expression in resistant BRAF V600E CRC cell lines, suggesting that alternative mechanisms of resistance originate downstream of MEK. These findings provide the basis for further exploration of non-EGFR mediated mechanisms of resistance, which ultimately may enable the development of new rational targeted drug combination strategies for BRAF mutant CRC. Bevacizumab, an anti-human VEGF-A monoclonal antibody, is currently approved for the treatment of metastatic CRC in combination with chemotherapy. Discovery of a predictive biomarker for response to bevacizumab would enable us to identify patients who are unlikely to respond, commence them on alternative therapies sooner, and avoid unnecessary side effects. Identification of a predictive biomarker may also improve the cost effectiveness of bevacizumab use. We explored whether single nucleotide polymorphisms (SNPs) in VEGF-A, VEGFR1 and VEGFR2 are predictive biomarkers for bevacizumab response, using archived tissue samples and clinical outcome data from participants in the MAX Phase III trial of capecitabine, bevacizumab and mitomycin in the first-line treatment of metastatic CRC. VEGF-A SNPs rs25648 and rs699947 were found to be prognostic, but not predictive biomarkers of survival outcomes.

HFE haemochromatosis is the most common iron overload disease. Since the discovery of the HFE gene is 1996, it is readily diagnosed using a genetic test rather than using liver biopsies. It is an autosomal recessive disease and the most common form of HFE haemochromatosis is homozygosity for p.C282Y. Homozygosity for this substitution accounts for more than 90% of haemochromatosis in Australia. Excessive iron accumulates due to malfunction of the HFE protein that leads to excess iron absorption. As a result, excess iron builds up in various organs including liver, joints, heart, pancreas, pituitary gland and skin, that may cause end-organ damage including liver cirrhosis, cardiac failure, diabetes mellitus, hypopituitarism and skin pigmentation. Symptomatically, fatigue and arthralgia are the major complaints reported by patients with haemochromatosis. This disease is easily treatable, as the blood contains a significant proportion of the body’s iron, so excess iron can be removed via the blood through phlebotomy. Multiple studies have found that individuals with high total body iron, defined by serum ferritin of more than 1000ug/L, have the highest risk of developing complications including liver cirrhosis. In the last decade, some have suggested that patients with elevated body iron but with serum ferritin less than 1000ug/L, here defined as moderate iron overload, might not need treatment, as they might not have symptomatic manifestations of the disease. However, there have been no randomised controlled trials to examine the treatment benefits for individuals with moderately elevated iron. To answer this question, a randomised controlled trial (Mi-Iron) was conducted which was the major aim of my PhD to examine if removing excess iron would have an impact on patient-reported outcomes, particularly fatigue, as well as liver fibrosis and oxidative stress. This demonstrated that with treatment, there was an improvement in fatigue and its cognitive subcomponent, and the affect component of the arthritis score. There was also an improvement in the liver fibrosis marker, Hepascore, and oxidative stress marker plasma F2-isoprostane, by removing excess iron in this cohort when compared to the control group who did not have iron reduction. These results, therefore, support current guidelines that all patients with haemochromatosis with elevated serum ferritin should have phlebotomy to return body iron levels to normal levels. In the second clinical study, the relationship of serum ferritin with non-invasive markers of liver fibrosis including transient elastography, Hepascore, aspartate aminotransaminase (AST) to platelet ratio index (APRI) and Fibrosis-4 (FIB-4) to assess for liver fibrosis and cirrhosis in HFE p.C282Y homozygotes was examined. This study was conceived due to the shift from using liver biopsy to the increasing use of non-invasive techniques to assess liver fibrosis and cirrhosis. This showed that there was a linear relationship of serum ferritin with Hepascore, indicating that higher body iron is associated with more advanced liver fibrosis and cirrhosis. This relationship was also found for APRI and FIB-4 scores. These findings are important as they provide extra information in utilising these scores to assess liver fibrosis and cirrhosis in haemochromatosis. In the third study, the hand joint arthritis in people with haemochromatosis was examined. Arthralgia is one of the major complaints of individuals with haemochromatosis and is often one of the earliest symptoms of haemochromatosis. It is often difficult to differentiate between haemochromatosis arthropathy and osteoarthritis in the hands. The second and third metacarpophalangeal joints are described to be more commonly affected in individuals with haemochromatosis. By examining the data from HealthIron, a haemochromatosis cohort extracted from a population study that assessed the burden of disease due to iron overload, I found that there was an increase in first metacarpophalangeal joint abnormalities in those with HFE p.C282Y homozygosity, comparable to the frequency of involvement of second and third metacarpophalangeal joints.

A state of genetic chimerism is created by allogenic transplantation, the quality and quantity of which may have diagnostic significance. The monitoring of cellular macrochimerism following allogeneic hematopoietic cell transplantation (HCT) is standard of care. It provides an assessment of engraftment, prognostic information, and guides therapeutic interventions. Separately, the discovery of cell-free DNA (cfDNA) chimerism following solid organ transplantation has raised the possibility that increases in donor genotype graft-derived cfDNA may be biomarker for rejection. As a result of these and other developments, the clinical and research requirements for chimerism analysis in the transplantation context exceeds the capabilities offered by existing methodologies. To address this, a panel of 39 droplet digital PCR (ddPCR) assays was developed. These targeted highly polymorphic copy number variation (CNV) loci strategically selected using a population genomics approach to maximise the probability of a donor one copy, recipient zero copy genotype combination (and vice versa). These informative combinations create a negative background against which absolute quantification of the one copy genotype can be performed using ddPCR. Analytical validation of the developed chimerism analysis method for absolute quantification of cfDNA was performed. All studied assays performed linearly across the range <6-1280 copies/mL. Limit of blank was 0 copies/mL, limit of detection was 6 copies/mL, and limit of quantification was 8 copies/mL. A method for genotyping donor and recipient informative CNV loci from chimeric cfDNA was developed and conceptually validated using donor and recipient genomic DNA. The developed chimerism analysis method was used to evaluate the diagnostic validity of absolute measurements of plasma graft-derived cfDNA and total cfDNA, as well as the relative measure of graft fraction, for the diagnosis of kidney transplant (KT) rejection. 61 indication biopsy-matched samples were studied. For the diagnosis of antibody mediated rejection, the receiver-operator characteristic area under the curves of graft-derived cfDNA and graft fraction were 0.91 (95% CI 0.82-0.98) and 0.89 (95% CI 0.79-0.98), respectively. Both measures did not diagnose borderline or type 1A cellular mediated rejection. Analytical validation was also performed for the quantification of cellular chimerism in blood following allogeneic HCT. CNV chimerism results were compared against established fluorescence in situ hybridization, single nucleotide polymorphism, and short tandem repeat-based methods with excellent correlation. Using 30 ng of DNA, the limit of detection was 0.05% chimerism and the limit of quantification was 0.5% chimerism. Linear performance was established between <0.05%-100% chimerism. High informativity was seen with a median of four informative markers detectable per individual in 39 recipients and 43 donor genomes studied. The strength of this approach was exemplified in a multi-donor case involving four genomes, three of which were related. In conclusion, this thesis describes the development and analytical validation of a novel, highly informative test for cellular and cfDNA chimerism analysis based upon highly polymorphic CNV and absolute quantification via ddPCR. In addition to its use in the diagnosis of KT rejection and detection of chimerism after allogeneic HCT, the method holds promise as a means to study chimerism more broadly and in a variety of compartments.

Obesity in women of child-bearing age is common. There is little evidence that modest weight loss using lifestyle modification alters pregnancy outcomes. Bariatric surgery results in substantial weight loss, which reduces the risk of adverse outcomes for the mother but may increase risks for the baby. In order to reduce obesity-related adverse pregnancy outcomes for both mother and baby, alternative approaches to the management of obesity in women planning pregnancy are needed. OBJECTIVE: This study aims to explore the health consequences of substantial pre-conception weight loss, compared with modest pre-conception weight loss, in women with obesity. OUTCOMES: Primary outcome: To determine if non-surgical substantial pre-conception weight loss (10-15% body weight) in women with obesity (BMI 30-55kg/m2) causes a 10% reduction in fasting glucose at 26-28 weeks’ gestation compared with modest pre-conception weight loss (≤ 3% body weight). Secondary outcome: To determine if non-surgical substantial (10-15% body weight) compared with modest pre-conception weight loss (≤ 3% body weight) in women with obesity (BMI 30-55kg/m2) results in a reduction in the rate of (1) a composite of gestational diabetes, gestational hypertension/pre-eclampsia, delivery <37 weeks’ gestation, primary Caesarean section, shoulder dystocia/birth injury, large-for-gestational-age offspring, intra-uterine-growth-restriction, neonatal hypoglycaemia, neonatal hyperbilirubinemia and neonatal SCN/ICU admission, (2) gestational diabetes (3) large-for-gestational-age neonates (4) intra-uterine-growth-restriction (5) gestational hypertension/preeclampsia, (6) delivery >37 weeks’ gestation, (7) primary Caesarean section, (8) neonatal hypoglycaemia, (9) neonatal SCN/ICU admission, (10) maternal and neonatal length-of-stay, (11) a decrease in time-to-conception, (12) an increase in live birth rate (13) no difference in gestational weight gain. METHODS: This is a two-arm, parallel group, randomised controlled trial. 164 women with obesity (BMI 30–55 kg/m2), aged 18-28 years old and planning pregnancy were recruited via a social media platform. They were randomised to a 12-week intervention: modest weight loss (MWL ≤ 3% body weight) using a hypocaloric diet or substantial weight loss (SWL 10–15% body weight) using a modified very low energy diet (VLED). Completers of the intervention (Phase 1) were observed for 48-weeks (Phase 2: weight maintenance and Phase 3a: pre-pregnancy). Pregnancies were observed and pregnancy outcome data collected (Phase 3b). RESULTS: This thesis presents an interim analysis of results. In Phase 1, weight loss in MWL and SWL at Week 12 was 3.1% (3.2  0.6 kg) and 11.9% (13.0  0.5 kg) respectively (p<0.01). Cox proportional model demonstrated SWL had a significantly shorter time-to-conception than MWL (p=0.03). There was no difference in the change in fasting maternal glucose between baseline (pre-pregnancy) and 26-28 weeks’ gestation in those achieving modest (MWL -0.41 ± 0.14 (-7.6%)) and substantial (SWL -0.57 ± 0.11 (-10.6%)) pre-conception weight loss (p=0.40). A composite of pre-specified obesity-related adverse pregnancy outcomes was significantly reduced in SWL (34/21 events) compared with MWL (18/28 events) (p=<0.01). CONCLUSION: In women with obesity, non-surgical substantial pre-conception weight loss does not change fasting glucose at 26-28 weeks’ gestation compared with modest pre-conception weight loss. However, it does result a significant reduction in a composite of obesity-related pregnancy outcomes.

Non-alcoholic fatty liver disease (NAFLD) is the most prevalent liver disease in Australia and around the world and affects up to 40% of the adult population. It is the leading cause for liver transplantation in the USA and expected to be the leading cause in the next 10 years in Australia. There is no proven medical treatment for NAFLD except for liver transplantation in patients with advanced liver disease. The management involves lifestyle modifications and dietary restrictions, which are difficult to achieve in most patients. NAFLD can progress to its severe form of non-alcoholic steatohepatitis (NASH) and liver fibrosis which eventually leads to cirrhosis or hepatocellular carcinoma. It has been described that there are several factors which contribute to NAFLD progression such as genetic factors, epigenetic factors, oxidative stress, adipokines and microbiota. These factors constitute the second hit of the so called ‘two-hit’ hypothesis in which the fatty liver is considered as the first hit. Advanced glycation end products (AGEs) are formed by the non-enzymatic reaction between reducing sugars and proteins or lipids. These are found in large amounts in the Western diet where high temperatures are applied during food processing. AGEs can also be formed in the body during aging, and the rate of synthesis can be accelerated in conditions such as diabetes due to elevated blood glucose levels. AGEs have been implicated in the progression of many chronic diseases such as diabetic complications. We therefore hypothesized that AGEs are another ‘second hit’ that drives NAFLD progression to liver fibrosis. A series of studies in this thesis have been designed to investigate and compare the effects of endogenous AGEs formed in the body and exogenous or dietary AGEs on NAFLD progression to liver fibrosis. The effects of AGEs on NAFLD progression were explored by intraperitoneal administration of AGEs to C57BL/6 mice for 10 weeks following 30 weeks of a high fat (HF) diet. This strategy was used to mimic the effects of endogenously formed AGEs. The effects of intraperitoneal AGEs were compared to those of dietary AGEs by feeding mice with the HF diet for the first 30 weeks, followed by a diet containing high AGEs for the next 10 weeks. The high AGE diet was prepared by baking the HF diet at 160o C for 1 hour. To further evaluate the contribution of endogenous AGEs, HF fed mice were made diabetic by two consecutive daily injections of streptozotocin after 15 weeks of HF diet. RAGE knockout mice were used to evaluate the role of RAGE in AGE-induced effects. We also explored therapeutic strategies targeting the AGE/RAGE pathway in NAFLD progression. Since we found that vinegar marination of HF diet prior to baking reduces AGE content in the diet, in a series of studies we fed diabetic mice with a vinegar marinated baked diet for 40 weeks. Moreover, we used therapeutic drugs that are known to inhibit AGE formation and have been used in phase II clinical trials for other diseases but have not been tested for their therapeutic effects on AGE-induced NAFLD progression. This included the administration of aminoguanidine and pyridoxamine to diabetic or high AGE fed mice via drinking water during the last 10 weeks of the 40-week NAFLD model. In addition, we investigated the effects of alagebrium, an AGE-crosslink breaker, administered by daily gavage injection during the last 10 weeks to mice fed the high AGE diet for 40 weeks. Moreover, an AGE trafficking study was performed by gavaging fluorescent labelled-AGEs to high AGE fed mice treated with alagebrium in the last 2 weeks of a 4-week study. Dietary AGE content, gene expression of pro-inflammatory and pro-fibrotic cytokines, collagen 1, endotoxin responsive genes and RAGE in the liver and liver injury, oxidative stress and fibrosis were determined. Moreover, to elucidate mechanisms we performed cell culture experiments using two immortalised cell lines, murine Kupffer cells (KUP5 cells) and human hepatic stellate cells (HSCs) (LX2 cells). Gene expression of pro-inflammatory cytokines, ROS generation and HSC activation were determined in cells exposed to AGEs in the presence and absence of RAGE antagonists or alagebrium. We found that baking the HF diet increased its AGE content by several fold. Whilst the effects of intraperitoneal AGEs on pro-inflammatory and pro-fibrotic cytokine expression and liver fibrosis were not significant, diabetic mice fed the HF diet showed elevated circulating AGE levels which were accompanied by increased liver expression of RAGE, cytokines, endotoxin responsive genes, collagen 1, oxidative stress and liver fibrosis compared to the HF diet. There was also a positive and strong correlation between circulating AGEs and liver RAGE expression. Similarly, high dietary AGE intake in non-diabetic mice was also associated with an increased expression of the genes measured above, oxidative stress and liver fibrosis compared to HF diet, with diabetic mice fed the high AGE diet showing an additive effect on some parameters such as circulating AGEs, RAGE and oxidative stress. This suggests that increased AGEs formed in diabetes may be responsible for NAFLD progression to liver fibrosis. The concept that AGEs impart their effects by binding to RAGE is supported by the findings that AGE-induced up-regulation of pro-inflammatory cytokines was completely prevented by RAGE blockade in Kupffer cells. Importantly, the finding that liver fibrosis was worse in high AGE fed non-diabetic mice than diabetic mice fed the HF diet suggests that dietary AGEs are more harmful in NAFLD progression to liver fibrosis than endogenous AGEs. The AGE inhibiting drugs aminoguanidine and pyridoxamine failed to affect NAFLD progression to liver fibrosis in diabetic or high AGE fed mice. In fact, these drugs were apparently enhanced liver fibrosis in HF fed mice, at least under our experimental conditions. On the other hand, we found that dietary intervention with vinegar marination prior to baking the HF diet had a profound inhibitory effect on liver injury, liver expression of RAGE, pro-inflammatory and pro-fibrotic cytokines and HSC activation, leading to a marked improvement in liver fibrosis in diabetic mice fed the marinated high AGE diet compared to diabetic mice fed the high AGE diet without vinegar marination. Pharmacological intervention with the AGE-crosslink breaker alagebrium produced profound inhibitory effects on all of above parameters, leading to a marked improvement in NAFLD progression. The finding that an alagebrium-induced marked reduction in circulating AGEs in high AGE fed mice was supported by AGE-trafficking study which showed that alagebrium apparently promotes AGE clearance. Moreover, cell culture experiments showed that alagebrium directly inhibits AGE-induced pro-inflammatory cytokine secretion and ROS generation by Kupffer cells and thereby reduced the exposure of HSCs to these cytokines and thus, indirectly inhibited HSC activation. Thus, alagebrium in addition to promoting AGE clearance and other than its designated role as an AGE-crosslink breaker, may also have a direct effect on AGE/RAGE signalling. In summary, the studies in this thesis demonstrate that AGEs could be considered as a second hit that drives NAFLD progression to liver fibrosis. These findings also suggest that diabetes which is an important source of the circulating pool of AGEs, exacerbates NAFLD progression to liver fibrosis. Moreover, we conclude that dietary AGEs which appear to act directly on liver AGE/RAGE signalling and indirectly via the gut/liver AGE/RAGE axis are more harmful and stronger stimulant than endogenously derived AGEs in driving NAFLD progression. Thus, dietary modifications and changes to food preparation such as vinegar marination provide a feasible strategy to reduce dietary AGE intake and associated liver injury in NAFLD. We also conclude that AGEs directly act on Kupffer cells, leading to indirect activation of HSCs. We also show that alagebrium, a drug considered to be an AGE-crosslink breaker, may also have a direct effect on AGE/RAGE signalling in Kupffer cells. We also provide evidence that alagebrium is a potential therapeutic drug for NAFLD patients that could be trialled in a phase II study since the drug has been trialled in patients with other conditions such as heart failure and diabetic nephropathy.

Gene expression experiments are conducted to better understand the biological mechanisms in cells and tissues. Examples of the wide range of research and clinical applications of gene expression analysis include: identifying and classifying disease types, assessing the transcriptional differences between the state of a disease and normal cells, and identifying biomarkers to predict clinical outcome of diseases such as cancer, and other many applications. High-throughput technologies such as RNA- sequencing have revolutionized gene expression analysis over the last 15 years. Large- scale gene expression studies are highly desirable for increasing our biological insights into highly heterogeneous diseases such as cancer. Gene expression data is almost always compromised by unwanted variation that may lead to inaccurate and wrong biological results. Effective removal of unwanted variation such as batch effects is one of the main challenges of the analysis of gene expression data, particularly when the data comes from large and complex experiments. In addition, the performance assessment of batch correction method prior to perform further analysis is an essential part of the analysis of gene expression data, which has been underappreciated. The Nanostring nCounter gene expression platform has been increasingly used for research and clinical studies due to its ability to directly measure a broad range of mRNA expression levels without cDNA synthesis and amplification steps. Nanostring data, similar to other gene expression data, requires to be corrected for different source of unwanted variation. I critically assessed the performance of the Nanostring normalization approaches using four large and complicated Nanostring data sets and observed that the Nanostring normalization failed to adjust for all unwanted variation in the data. I used a novel method, Removing Unwanted Variation-III (RUV-III), that makes an essential use of negative control genes and technical replicates to remove unwanted variation from the data. A variety of statistical tools were employed to assess the performance and effectiveness of RUV-III on the Nanostring data. I demonstrated how unwanted variation in Nanostring gene expression data could lead to inaccurate and wrong interpretation of the results. The Cancer Genome Atlas (TCGA) Research Network has catalogued different molecular information of 33 human cancer types to increase our biological insight into cancers. The TCGA consortium unavoidably processed samples in batches, which can result in unwanted variation which we call batch effects. I evaluated the effects of different sources of unwanted variation on the TCGA breast cancer RNA-seqV2 data as an example, and discovered that the use of different flow cell chemistries introduced batch effects that compromised down-stream analyses such gene co-expression analysis. I also investigated the effects of varying levels of tumor purity on the data and found that tumor purity confounded downstream analysis. A new strategy for removing unwanted variation was developed using RUV-III based on the use of pseudo-replicates and negative control genes. The results showed that RUV-III normalization led to satisfactory results where other methods showed shortcomings. Finally, I used our in-house Nanostring gene expression data and the TCGA RNA-seqV2 data sets that were critically evaluated for unwanted variation in previous chapters, to explore gene expression alteration of DNA repair genes in lung cancer adenocarcinoma and normal lung tissues. The hypothesis was that if one knew which DNA repair genes are altered in a tumour, that information could be used to guide the appropriate choice of DNA damaging therapy for lung adenocarcinoma patients e.g. determining which patients may respond to which current therapy and which patients may respond to new therapies. I identified recurrent and novel gene expression alterations that may have utility in guiding clinical decision-making for the use of adjuvant chemotherapy for patients with lung adenocarcinoma. In summary, removal of unwanted variation is key to deriving meaningful biological results from gene expression data. Importantly, the corrected gene expression data must be evaluated to ensure that unwanted variation are effectively removed and the biological signals are preserved regardless of the method that has been used

The interstitial lung diseases (ILD) are a diverse group of inflammatory and fibrotic lung conditions which cause significant morbidity and mortality. The prototypic ILD, idiopathic pulmonary fibrosis, is the most common form of chronic progressive fibrotic lung disease. While there has been significant progress in recent decades in classifying and diagnosing ILD using international consensus guidelines, effective therapeutic options remain limited. Recently approved antifibrotic therapies and immunosuppressants have the potential to slow disease progression in some patients with ILD, however, there is no proven therapy which relieves symptoms in this population. Ambulatory oxygen therapy is commonly prescribed for patients with ILD and exertional desaturation, with the aim of improving symptoms and functional status. However, current clinical recommendations for its prescription in patients with ILD are based on limited indirect evidence from trials conducted in patients with chronic obstructive pulmonary disease. The body of work in this thesis aimed to explore the role of supplemental oxygen therapy in ILD. In order to understand the disease burden of ILD, a Cochrane systematic review was conducted to evaluate prognosis in patients with idiopathic pulmonary fibrosis who were either untreated or receiving what would now be considered ineffective treatment. This review confirmed the poor prognosis of patients with IPF with overall survival rates of 88% at one to two years and 31% at five to 15 years, using a total study population of more than 25,000. While exertional desaturation is known to be a significant prognostic factor in ILD, little is known about its prevalence and any associated use of ambulatory oxygen therapy. In this thesis, using the prospective Victorian ILD registries, more than 50% of patients exhibited exertional desaturation at initial assessment, with a prevalence of 20% to 30% in those with mild physiological impairment. Prescription of ambulatory oxygen therapy was associated with more severely impaired lung function and worse exercise capacity, although the prescribing indications were unclear with a lack of evidence base underpinning current prescribing practice. The follow-up qualitative studies provided insights into patients’ and physicians’ perceptions and experiences of domiciliary oxygen use in ILD. While domiciliary oxygen therapy provided symptomatic benefits and improved functional capacity in some patients, its use was associated with substantial physical and psychosocial challenges. Physicians described prescribing oxygen therapy for symptom relief in patients with ILD, although there was significant heterogeneity in prescribing threshold, possibly on account of the lack of high-quality evidence-based recommendations with consequent inequitable access. Portable oxygen concentrators (POCs), newer oxygen delivery devices, were raised by both groups as an alternative source of ambulatory oxygen therapy which appeared to be more user-friendly than the more commonly-used compressed oxygen cylinders. These findings highlight the need to explore alternative devices for and long-term impacts of ambulatory oxygen therapy in patients with ILD. A crossover study in patients with ILD and exertional desaturation, comparing different portable oxygen concentrators and a compressed oxygen cylinder, found similar clinical performance among the devices during 6-minute walk tests. There were no significant differences in mean nadir oxygen saturation on exertion or distance walked, with equivalent oxygenation at rest. Nonetheless, patients preferred using a POC over a compressed oxygen cylinder, for reasons of a more favourable physical appearance and ease of manoeuvrability. These results suggest that POCs could be used as an alternative source of ambulatory oxygen therapy for patients with ILD. Guided by the preceding analyses, a pilot trial was performed, which confirmed the feasibility of conducting a randomised sham-controlled trial of ambulatory oxygen therapy delivered via POCs in patients with ILD. The novel approach of participant blinding using a sham POC was effective, as evaluated using the Bang’s blinding index. Importantly, an interim analysis of the exploratory efficacy outcome measures revealed significant differences in health-related quality of life and duration of physical activity between treatment groups. These findings substantiate the need for a definitive trial to examine the effects of ambulatory oxygen therapy in patients with ILD when used in the community.

Portal hypertension (PHT) and bleeding from gastroesophageal varices is the major cause of morbidity and mortality in cirrhotic patients. In cirrhotics, splanchnic vasodilatation leading to an elevated portal venous inflow plays an important role in the pathogenesis of PHT. Although the current pharmacological mainstay to reduce portal pressure by decreasing splanchnic inflow is non-selective beta-blockade, most patients (>60%) do not achieve an optimal response and 15% of patients are unable to tolerate these drugs. The peptide angiotensin-(1-7) (Ang-(1-7)) and Mas receptor (MasR) of the so called alternate axis of the renin angiotensin system (RAS) contributes to the pathological splanchnic vasodilatation in cirrhosis. Recent findings suggest that the vasodilatory effects of Ang-(1-7) could also be mediated via the Mas related G-protein coupled receptor type D (MrgD), a newly identified receptor of the alternate RAS. However, the effects of this receptor in mediating splanchnic vasodilatation are largely unknown. The experiments in this thesis aimed to investigate the effects of the blockade of the vasodilatory receptor MrgD in reducing portal pressure, and to compare them with those of MasR blockade, in cirrhotic and non-cirrhotic portal hypertensive rats. For the experiments in cirrhotic rats, the bile duct ligated (BDL) and carbon tetrachloride injected (CCl4) rat models were used. To investigate the short-term (acute treatment) effects of receptor blocker treatment, the MrgD blocker D-Pro7-Ang-(1-7) (D-Pro) (10µg/kg) or MasR blocker A779 (10µg/kg) were administered as an intravenous bolus injection. To study the long-term treatment (chronic treatment) effects, the BDL and CCl4 rats received a continuous infusion of either A779 or D-Pro at 28µg/kg/hr for 2 weeks, via subcutaneously implanted osmotic mini-pumps. Following the treatment, the rats were used for the measurement of portal pressure and mean arterial pressure (MAP). In these experiments, microsphere injections were performed to measure vascular resistance and regional blood flow. Mesenteric resistance vessels obtained from a separate group of CCl4 rats receiving long-term receptor blocker treatments or vessels isolated from saline-infused diseased rats and subjected to in-vitro blocker treatments, were used in a vessel myograph, to study the vasodilatory response to acetylcholine (Ach). In addition, we have also studied the potential use of a combination therapy of MrgD and MasR blockers in counteracting systemic hypotensive effects of the AT1R blocker losartan, because AT1R blockers are drugs that potentially can be used as anti-fibrotic and portal pressure lowering agents in fibrotic livers. To investigate the effects of short-term (acute treatment) combination therapy, D-Pro or A779 were injected 20 minutes after a bolus injection of losartan. To study the effects of long-term (chronic treatment) combination therapy, cirrhotic BDL rats received a long-term infusion of A779 or D-Pro via osmotic mini-pumps, combined with daily oral gavage injection of losartan for the same duration i.e. 2 weeks, prior to use in pressure measurement and/or regional blood flow studies. The effects of the short-term and long-term treatment of MrgD or MasR blockers on portal pressure and regional blood flows were also studied in the non-cirrhotic rat model of partial portal vein ligation (PPVL). In addition, the contribution of RAS to the development of PHT in the non-cirrhotic PPVL rat model was also studied using a time-course experiment. In these studies, changes in portal pressure at each time point were compared to the RAS gene expression profiles in the mesenteric vessels. Short-term therapy of both MrgD and MasR blockers significantly (p<0.001) reduced portal pressure in the cirrhotic BDL and CCl4 rats. However, a long-term continuous infusion of these blockers produced remarkable effects in cirrhotic portal hypertensive rats. The long-term treatment with D-Pro produced more profound inhibitory effect on portal pressure compared to that of A779. This prominent effect of D-Pro on portal pressure was likely related to its potent effect in inhibiting splanchnic vasodilatation which resulted in markedly increased splanchnic vascular resistance (SPVR), leading to a marked reduction in mesenteric blood flow (MBF). In keeping with these findings, vascular myograph experiments suggested that treatment with D-Pro but not A779, specifically inhibited the vasodilatory responses of 1st order as well as 2nd/3rd order mesenteric vessels to Ach, thus suggesting splanchnic vasculature-specific effects of MrgD in cirrhotic rats. Moreover, MasR but not MrgD blockade produced off target effects in the renal and hepatic vascular beds, confirming that unlike the MasR, MrgD plays a splanchnic vasculature-specific role in cirrhosis. A bolus injection of AT1R blocker losartan also reduced portal pressure in cirrhotic rats. However, despite its anti-fibrotic effects with improved intrahepatic vascular tone, long-term treatment with losartan failed to reduce portal pressure, likely due to its potent effects in reducing systemic and splanchnic vascular resistance and thereby increasing mesenteric blood flow coupled with increased cardiac output. Moreover, the systemic hypotensive effect produced by losartan could not be counteracted with a short-term or long-term combination therapy with D-Pro or A779. On the other hand, neither the MrgD nor MasR blockade were able to alter splanchnic vascular resistance or mesenteric blood flow in non-cirrhotic PPVL rats, which was supported by unaltered gene expression of both these receptors in the mesenteric vessels of PPVL rats compared to healthy controls. However, it appears that in PPVL rats, downregulation of vasoconstrictive classical RAS and in particular, AT1R expression in the mesenteric vessels contributed to splanchnic vasodilatation. In conclusion, blockade of the vasodilatory receptors of the alternate RAS; the MasR (22%) and in particular, the MrgD (33%), produces a clinically significant reduction of portal pressure (>20% from baseline) in cirrhotic rats. We also conclude that the effects of MrgD blockade but not MasR blockade, are splanchnic vasculature-specific and therefore, this newly characterized receptor, MrgD, is an ideal target to develop future therapeutics to treat PHT in cirrhosis. Based on findings from non-cirrhotic PPVL model, we conclude that the alternate RAS does not contribute to PHT in the absence of cirrhosis. Furthermore, a reduced activity of the classic RAS with specific reference to AT1R in the mesenteric vasculature appears to be play a role in pathological vasodilatation in non-cirrhotic PHT.

Cytomegalovirus (CMV) infects the majority of the world’s population and establishes a life-long infection, which requires intact T cell responses to control. Infants in utero and immune-compromised populations are most at risk for CMV disease, which causes significant morbidity and mortality. In healthy populations, CMV persistence has a dominant effect on human immune responses, aging and mortality. Currently, there is no cure or vaccine for CMV and antivirals have significant toxicity. Conversely, CMV infection induces large sustained effector T cell memory pools (“memory inflation”), which can be harnessed in vaccines and cancer immune therapy. This dissertation covers three distinct domains: T cell immunity to CMV in (1) health, (2) solid organ transplantation (SOT), and (3) novel vaccines. Domain 1: T cell responses are initiated, function and maintained in diverse tissues, however, investigation of human T cell responses has mostly been limited to blood. Through novel analysis of lymphoid and mucosal tissues obtained from 88 human organ donors, we reveal that long-term control of CMV requires considerable effort from the host immune system, has a lasting impact on the profile of the immune system in blood and tissues, and that T cell reservoirs for CMV control in tissues are shaped by both viral and tissue-intrinsic factors. Domain 2: Despite advances in antiviral prevention and treatment, CMV continues to impact patient outcomes following SOT. Through longitudinal analysis of ten SOT recipients with CMV infection, we show that primary CMV and CMV reactivation dramatically and persistently alters the composition of the circulating T cell pool. In addition, CMV-specific CD8+ T cells from recipients who controlled CMV had features of recent activation, suggesting that viral reactivation had occurred and been contained in tissues. Domain 3: CMV and adenovirus vectors induce an expanded, sustained effector memory CD8+ T cell responses, termed “memory inflation”. The precise phenotype and developmental requirements of memory inflation is not known. Peripheral memory T cells (TPM), defined by intermediate expression of the chemokine CX3CR1 (CX3CR1int), have recently been shown to have proliferative and tissue-surveillance properties that may be of relevance to memory inflation. We tracked these cells in memory inflation induced by CMV and adenovirus-vectored vaccines in mice and humans and found that CMV and vaccine-induced inflationary T cells showed high levels of CX3CR1int cells exhibiting an effector-memory phenotype characteristic of inflationary pools. These data indicate that CX3CR1int cells form an important component of memory responses to persistent viruses and vaccines in both mouse and man. Next, we show that the inflationary adenovirus-induced T cells are highly dependent on IL-21. Taken together, the data presented in this dissertation has implications for the promotion of life-long tissue immunity, strategies to improve protective responses against CMV and for generating inflating memory T cells in novel vaccines.