Gaps in the quality of hospital care in low- and middle-income countries contribute to neonatal death and morbidity. Most neonatal deaths occur in the first few days of life, many from preventable or treatable causes. Essential newborn care consists of low-cost interventions, such as basic resuscitation, early breastfeeding and skin-to-skin contact, which have been shown to improve outcomes. Successful essential newborn care implementation requires understanding of the setting in which it is being delivered and the contextual factors that enable healthcare workers to provide quality care for newborns. Solomon Islands is a low-resource country in the Western Pacific and a Small Island Developing State with high neonatal mortality rates, increasing births per capita and limited healthcare resources and personnel. The Ministry of Health and Medical Services (MHMS) and stakeholders implemented an intervention to improve newborn quality of care in Solomon Islands. This thesis evaluates the quality of care in hospitals of Solomon Islands and describes contextual factors that enabled successful implementation of a multifaceted intervention to improve essential newborn care. This thesis used a mixed methods design comprising the following sequential studies: (1) Quality of hospital care for newborns was described through a cross-sectional study using a structured assessment tool to identify strengths and limitations in structure and process components of existing care. (2) Three years of perinatal outcomes were audited to determine stillbirth, perinatal and neonatal mortality rates and the main causes of neonatal morbidity and mortality. (3) The impact of the World Health Organization Early Essential Newborn Care training program on knowledge and skills of healthcare workers was assessed, using a before-and-after study that identified the factors associated with improved retention of knowledge and skills. (4) The implementation process was described through interviews of healthcare workers, and interview data were triangulated with quantitative results to describe features of the intervention that supported implementation. This thesis demonstrated gaps in structure and process elements in quality of newborn care, which limited provision of appropriate, timely clinical care in the hospitals. Very high perinatal mortality rates, mainly owing to stillbirths, were identified. The provinces had higher rates of perinatal mortality than the National Referral Hospital. The main causes of neonatal mortality were complications of prematurity, birth asphyxia and infection. Knowledge and skills of healthcare workers significantly increased following the coaching program. At the time of evaluation, some loss of skills over time had occurred, particularly in cadres of healthcare workers that did not routinely use relevant skills. The evaluation of the implementation process reflected strengths of the intervention, specifically the training methods (content, short duration and practical approach) and the impact of a small training team of MHMS midwives and nurses who delivered the program independently. The challenges and sense of anxiety faced by healthcare workers in remote, isolated settings with infrequent exposure to resuscitation were highlighted. The barriers to establishing independent, ongoing quality improvement activities reflect the challenge of sustaining action across a large geographical area that has a dispersed health workforce when there is little capacity for regular oversight and support. This thesis highlights the potential impact from a multifaceted intervention to improve essential newborn care. Together with efforts to address stillbirths and improve intrapartum quality of care and quality of care in the neonatal period, essential newborn care may form an important part of a strategy to improve outcomes for newborns.

Truncus arteriosus is a rare congenital cardiac defect which results in a single common arterial trunk exiting the heart which supplies the systemic, pulmonary, and coronary circulations. The truncus itself is guarded by a single, often large common valve – the truncal valve – which separates the truncus from both the left and right ventricle. Truncus arteriosus has an incidence of 3 to 10 per 100,000 live births. Although only 0.7% – 3% of all congenital cardiac anomalies are due to truncus arteriosus, it accounts for 4% of all critical congenital cardiac anomalies. Patients typically present early in life with symptoms of cyanosis and congestive cardiac failure. Nowadays, surgery is undertaken early in life prior to the development of irreversible pulmonary hypertension. Improvements in surgical techniques and perioperative management has drastically reduced early mortality to 3 – 20%. Therefore, many children who have undergone truncus arteriosus repair are living well into adulthood. Despite this, there are few large studies addressing the long-term outcomes of truncus arteriosus repair. Furthermore, the impact of concomitant anomalies and the truncal valve are insufficiently described. This Doctor of Philosophy focuses on the long-term outcomes of truncus arteriosus repair in order to determine the current results and risk factors for poor outcomes. This research constitutes the largest single-institutional and multi-institutional experience assessing the long-term outcomes of truncus arteriosus repair with the longest follow-up time. I demonstrated that the majority of mortality following truncus arteriosus repair occurs within the first year after repair, and survival beyond the first year is excellent. I found that the presence of a coronary artery anomaly was associated with both early and late mortality and suggest that the coronary anatomy be clearly identified intraoperatively. Furthermore, patients with DiGeorge syndrome are at risk of late mortality, most commonly due to infection. Interestingly, I found that patients with mild or less truncal valve insufficiency are free from truncal valve surgery for up to 25 years, despite their truncal valve anatomy. In contrast, most patients with moderate or greater truncal valve insufficiency – particularly those with a quadricuspid truncal valve – will require truncal valve surgery at some stage in their lifetime. Of note however, the durability of truncal valve repair as a whole is poor, with most patients requiring reoperation on the truncal valve. In those with a quadricuspid truncal valve, repair by tricuspidization appears to be the most durable option with good long-term outcomes. Tricuspidization provided better long-term outcomes even if the non-tricuspidization group included younger children (less than 6 years of age), in whom truncal valve replacement was performed. This is an important finding as it suggests that younger children may benefit from truncal valve repair rather than a replacement with a smaller (non-adult sized) mechanical prosthesis which may require repeat replacement. Furthermore, if repair of the truncal valve is possible, this would avoid life-long anti-coagulation and the associated risks. In the long-term, patients had an excellent functional status following truncus arteriosus repair but had a high rate of reoperation due to the use of a conduit for reconstruction of the right ventricular outflow tract. Despite the high reoperation rate, patients have similar quality of life compared to age-matched Australian controls. This Doctor of Philosophy has redefined our understanding of the long-term outcomes of truncus arteriosus repair. The findings presented will impact clinical decision making, and I envision an improvement in the outcomes of these rare and complex patients.

Background: Epidemiological evidence suggests that Bacillus Calmette-Guerin (BCG) vaccine exerts non-specific (heterologous) effects in infants; decreasing neonatal mortality in high-mortality settings and preventing allergy and morbidity from infection in developed countries. New tuberculosis (TB) vaccines could potentially lack these beneficial effects. Immune mechanisms underlying the non-specific effects of BCG vaccine have been linked to ‘trained immunity’ or innate immune memory. Aims: Part 1: To investigate whether neonatal BCG vaccination alters the immune response to heterologous pathogens and Toll-like receptor (TLR) ligands in (i) the neonatal period and (ii) infancy. Part 2: To use a systems vaccinology approach to identify innate immune signatures underlying the non-specific effects of BCG vaccine. Methods: Part 1: In the Melbourne Infant Study: BCG for allergy and infection reduction (MIS BAIR), 1272 infants were randomised to receive BCG vaccine or no BCG vaccine within the first 10 days of life. A subset of participants was recruited to an immunological sub-study. A whole blood stimulation assay of ‘specific’ mycobacterial antigens, heterologous bacterial or fungal antigens and TLR ligands was used to interrogate cytokine responses at 7 days (n=212) and 7 months (n=167) post randomisation. Part 2: In the BabyBAIR study, 44 infants who were BCG vaccinated prior to travel to a TB-endemic area were recruited. Blood was collected from participants prior to BCG vaccination, and at 7 days and 3 months post vaccination. Cytokine responses and cell populations were analysed following in vitro stimulation of whole blood as above. RNASeq was also done on whole blood and the transcriptome was analysed for differentially expressed genes. Pathway analysis was done using functional gene set enrichment analysis (fGSEA) at each time point compared to baseline.   Results: In the MIS BAIR study, infants who were BCG-vaccinated had significant differences in their heterologous cytokine responses at both 7 days and 7 months post randomisation compared to BCG-naive infants. At 7 days post randomisation, compared to BCG-naive neonates, BCG-vaccinated neonates had evidence of a pro-inflammatory bias in RPMI-stimulated (nil) samples. Following heterologous stimulation, BCG-vaccinated neonates had decreased chemokine (MCP-1, MIP-1alpha, MIP-1beta, IL-1RA, IL-6 and IL-10 responses following stimulation of TLR2 (PEPG) and TLR7/8 (R848). At 7 months post randomisation, compared to BCG-naive infants, BCG-vaccinated infants, had decreased IFN-gamma responses to stimulation with heterologous pathogens. Decreased IFN-γ responses in the BCG-vaccinated group were attributable to a reduction in the proportion of individuals mounting an IFN-gamma response. Heterologous cytokine responses were modified by sex and maternal BCG vaccination status in both neonates and older infants. In the BabyBAIR study, longitudinal cytokine analysis showed BCG vaccination to be associated with a pro-inflammatory bias at baseline (prior to BCG vaccination), a robust pro-inflammatory heterologous response at 7 days post BCG vaccination and downregulation of pro-inflammatory cytokines 3 months post BCG vaccination compared to baseline. Flow cytometry suggested that both myeloid and monocyte derived dendritic cells were associated with the observed heterologous cytokine responses. RNASeq analysis of the whole blood transcriptome following BCG vaccination indicated widespread changes in innate immune signalling pathways and identified several potential mechanisms by which BCG vaccine could mediate its beneficial heterologous effects. Conclusions: Neonatal BCG vaccination leads to significant changes in the immune phenotype of vaccinated individuals. Following in vitro stimulation with heterologous pathogens, BCG-vaccinated infants have altered immune responses which might improve regulation of the inflammatory response during acute infection and a subsequent reduction in all-cause mortality. These results support the paradigm of BCG-induced trained immunity and provide additional information regarding the nature of the response in neonates and between different classes of pathogens.

The last few decades have seen great advances in the world of paediatric cardiac surgery. In particular, the arterial switch operation (ASO) is one of the specialty’s most impressive achievements. It is the procedure of choice for most children born with transposition of the great arteries (TGA). This project aimed to expand our understanding of the long-term outcomes and prognosis of patients who have had an ASO for TGA. The ultimate goal of this work was to not only increase clinicians’ knowledge of the long-term consequences of the ASO but also to allow cardiologists, cardiac surgeons and paediatricians to give the parents of patients, and patients themselves, more information about their future cardiac health. We aimed to achieve these goals by retrospectively reviewing patients who underwent an ASO at The Royal Children’s Hospital in Melbourne. This unit has performed the ASO for more than 30 years with a very low hospital mortality from the beginning. This allows for review of a large population of surviving patients from which long-term outcomes can be investigated, and makes possible the analysis of rarer subgroups of patients undergoing the ASO. We showed that the long-term outcomes of patients undergoing the ASO are generally good extending into young adulthood. However, reintervention is common and the incidence of clinically significant neo-aortic regurgitation or the need for neo-aortic root reoperation remains an issue for survivors with more than 25 years of follow-up. In regards to the subgroups examined, we showed that patients with intramural coronaries can be operated with excellent long-term outcomes using a simple, reproducible technique. Children who are less than 2.0 kilograms at time of ASO represent one of the highest mortality risk subgroups. Female adult survivors can undergo pregnancy with a low risk of maternal cardiac complications. Patients with associated aortic arch obstruction have a higher rate of reintervention for right sided obstruction and have a low rate of recurrent arch obstruction when avoiding the use of patch material. We also demonstrated that quality of life in adult ASO survivors was comparable to the general population using a validated quality of life questionnaire.

Paediatric acute myeloid leukaemia (AML) is a cancer of the blood and bone marrow. It is currently one of the leading causes of cancer-related mortality in children. While induction therapy is largely successful in achieving patient remission, the relatively high mortality rate is driven by the large genetic heterogeneity of AML and recurrence of disease. Disease relapse rate is higher than other childhood leukaemias, is fast acting and often chemotherapy resistant. While much of the genetic contribution to disease has been described, there is still a component of AML pathogenesis that has yet to be discovered. Many of the genetic lesions found in adult AML directly affect epigenetic modifying genes, however this is not the case in children. Despite this, previous research has shown vast epigenetic alteration in paediatric AML. As such it is possible that some of the unexplained pathogenesis in childhood AML can be elucidated by modulation of gene activity via aberrant changes in the most widely studied epigenetic process, DNA methylation (DNAm). Few studies have comprehensively interrogated the DNA methylome of paediatric AML, nor has the prognostic utility or biomarker potential of DNAm been explored. In this study, we explored the global methylation profile of paediatric AML in comparison to non-leukaemic controls and subtype-dependant and independent biomarkers of disease that may have functional relevance. Furthermore, we described DNAm signatures with potential prognostic utility, to accurately identify predisposition to relapse at diagnosis. Genome-wide DNAm was interrogated via the HumanMethylation450 BeadChip Array (HM450K) on a cohort comprising of 128 archival and fresh bone marrow tissue sourced from multiple hospitals around Australia. This data was then combined with the TARGET AML cohort comprising of a further 231 bone marrow samples. Targeted replication and validation of findings was undertaken on a reduced cohort using SEQUENOM MassArray EpiTYPER. Bioinformatic and machine learning analyses were undertaken in R. The findings revealed subtype-independent genome-wide average methylation (GWAM) to be increased in diagnostic samples compared to non-leukaemic controls. This was further verified by differences in the global methylation proxy genes known as LINE1 and Alu. Deeper interrogation of these differences demonstrated wide-spread differential methylation in previously implicated genes in AML pathogenesis including WT1 and DGKG, both of which were validated in an independent cohort. Other genes identified to be differentially methylated included ZSCAN1, REC8 and IRX1. Subtype analysis validated previous studies showing inv(16)-specific differential methylation in MN1 and MEIS1. Finally, DNAm was used as the primary feature for a machine learning model designed to predict patient relapse at diagnosis. The final model achieved an area under the curve (AUC) of 94% with correct identification of 91% of all cases involved (F-measure=0.914). To date, this study represents the largest and most comprehensive insight into aberrant DNAm in paediatric AML. Results have increased our understanding of genes that are differentially methylated and highlight the potential utility of DNAm as a future prognostic biomarker. It is anticipated that these findings will serve as a foundation for future functional studies aimed at delivering truly personalised treatment regimens for children with AML.

Aminoglycoside antibiotics are lifesaving medicines, crucial for the treatment of chronic or drug resistant infections. However, aminoglycosides can destroy the sensory hair cells of the ear. As a result, aminoglycoside treated individuals frequently experience a reduced quality of life, stemming from permanent high-frequency hearing loss and vestibular impairment. Currently, no alternate antibiotic substance has the same bactericidal profile or efficacy as aminoglycosides, and an adjuvant therapy capable of mitigating ototoxic outcomes does not exist. In order to develop an otoprotective therapy, the mechanisms of aminoglycoside-induced hair cell death must be elucidated. With respect to this goal, the production of reactive oxygen species and subsequent c-Jun N-terminal kinase (JNK) and P38 mitogen-activated protein kinase (P38) phosphorylation has been extensively documented in aminoglycoside treated hair cells. However, strategies directly targeting ROS, JNK or P38 are limited by the importance of these molecules for normal cellular function. Notably, the upstream regulators of JNK or P38 have not been well studied in the ear. Therefore, this project aimed to elucidate the role of upstream regulator Apoptosis signal-regulating kinase 1 (ASK1) within the auditory system. ASK1 is a key mediator of ROS induced JNK and P38 mediated disease. Importantly, ASK1 inhibition has previously been shown to reduce the pathological consequences of ROS, P38 and JNK signalling in varied disease models; without impeding the normal homeostatic cell function of these molecules. Therefore, ASK1 inhibition may represent a novel strategy for preventing aminoglycoside ototoxicity. However, the role of ASK1 as a mediator of drug-induced hair cell death has not been investigated. Moreover, the role of ASK1 in the development and function of auditory structures has not been explored. This project first aimed to elucidate the importance of ASK1 in the inner ear by characterising the auditory phenotype of Ask1 knockout mice. Histology indicated normal development of cochlear structures and the auditory brainstem response (ABR) demonstrated that Ask1 knockout mice had hearing thresholds comparable to C57BL/6 controls. However, ABR peak analysis indicated that auditory signal transduction is faster in Ask1 knockout mice. In addition, the acoustic startle response showed Ask1 knockout mice to be hypersensitive to auditory stimuli. Combined, this data indicates that ASK1 is important for neuronal function and that Ask1 knockout mice have an auditory processing disorder. This project then evaluated the importance of ASK1 as a mediator of drug-induced hair cell death. To test the hypothesis that ASK1 deficiency is protective against aminoglycoside-induced hair cell death, cochlear explants from Ask1 knockout and C57BL/6 mice were treated with neomycin in vitro. After six hours of 1 mM neomycin treatment, immunohistochemistry demonstrated that p-JNK staining was reduced in Ask1 knockout explants when compared to C57BL/6 explants. Moreover, hair cell death was significantly attenuated in Ask1 knockout explants. This data provides robust evidence that ASK1 has an important role in the process of hair cell death, and that ASK1 inhibition could limit aminoglycoside-induced hair cell death. Therefore, the final aim of this project was to test ASK1 inhibition as an otoprotective strategy. ASK1 inhibition significantly attenuated neomycin-induced outer hair cell death in C57BL/6 cochlear explants. Immunohistochemistry and western blot analysis suggested that ASK1 inhibition attenuated both P38 and JNK phosphorylation in neomycin-treated explants. Importantly, ASK1 inhibition did not impact the efficacy of the aminoglycoside antibiotics against P. aeruginosa in the broth dilution test. This indicates that ASK1 inhibition may mitigate ototoxic outcomes without impairing the primary aminoglycoside treatment. Overall, this thesis represents the first aural characterisation of the Ask1 knockout mouse strain, providing valuable insight regarding the role of ASK1 in the auditory pathway. Moreover, this work has provided compelling evidence that ASK1 inhibition may be a useful strategy for the prevention of aminoglycoside induced hearing loss.

Genetic kidney diseases are a heterogeneous group of disorders with varying phenotypes dependent on the affected nephron segment. Next generation sequencing has increased our appreciation of the breadth of gene variants associated with these diseases. It has also identified large numbers of variants of unknown significance (VUS), which require functional genomic validation. There is an unmet need for novel therapies for genetic kidney diseases as most invariably progress to dialysis or transplantation without any form of targeted treatment. Laboratory based research of genetic kidney disease requires the recapitulation of a disease phenotype in animal and/or in vitro cellular disease models. Interspecies variation in anatomy, physiology and gene function limits the translation of animal models to human disease and clinical care. Classical two dimensional cell cultures lack the complexity and intercellular cross-talk of in vivo three dimensional tissue. Kidney organoids are three dimensional, miniature, multicellular, human, in vitro micro-tissues, offering distinct disease modelling advantages over other models. Furthermore, kidney organoids can be regenerated from induced pluripotent stem cells (iPSC) reprogrammed from patients with genetic kidney disease, potentially providing outcomes with personalised clinical relevance. As a novel platform, the capabilities and limitations of kidney organoids as disease models are not well understood. By differentiating and characterising kidney organoids from the iPSC of patients with inherited kidney diseases, this thesis aims to explore the application of kidney organoids to disease modelling. As proof of concept, kidney organoids were first generated from iPSC reprogrammed from a patient with compound heterozygous variants in IFT140, an already validated nephronophthisis (NPHP) genotype. An isogenic control was generated by precision CRISPR-Cas9 gene editing. In this project, differential primary ciliary morphology within organoid tubules and transcriptional profiling of organoid epithelium validated the ability of the organoids to model genetic disease. Attempts were then made to validate novel, candidate variants for other pedigrees with unresolved trio whole exome sequencing. In a proband with clinically suspected NPHP, DNAH5 was selected as a candidate gene, despite previously association with a motile ciliary phenotype. In this project, kidney organoids were unable to validate the patient variant as pathogenic. In addition, a number of lessons were learned regarding the necessary variant curation process prior to making a commitment to modelling with kidney organoids. In the final chapter, kidney organoids validated a novel genotype for the glomerular disease steroid resistant nephrotic syndrome, via international collaboration with the laboratory of Prof Friedhelm Hildebrandt. Glomeruli within kidney organoids differentiated from iPSC expressing a patient-derived, homozygous variant in NOS1AP, displayed aberrant development, increased podocyte apoptosis and reduced expression of PAR polarity proteins. Together these projects demonstrate the strengths and challenges of using kidney organoids as models of inherited renal disease. Kidney organoids stand to complement animal and 2D unicellular disease models rather than replace them. We proposed that patient-derived kidney organoids are best placed to model paediatric onset kidney diseases with the future potential of providing personalised therapeutic screening.

Transient Receptor Potential Vanilloid 4 (TRPV4) is a non-selective calcium channel that plays an important role in the mechanotransduction system in chondrocytes. Heterozygous TRPV4 mutations cause skeletal disorders with varying severity. Heterologous cells such as fibroblasts and HEK-293 cells are commonly used to model TRPV4-inherited skeletal diseases in vitro. Studies using human chondrocytes are limited because cartilage is rarely available from patients and controls. Although heterologous cells cannot completely recapitulate the biological processes occurring in human chondrocytes, the studies show that two distinct disease phenotypes, TRPV4 skeletal dysplasia and arthropathy, might be caused by differences in the way the mutations change TRPV4 channel behaviour. TRPV4 skeletal dysplasia causing mutations show channel over-activity whereas arthropathy causing mutations show reduced channel activity upon channel stimulation. However, the downstream pathogenic mechanisms responsible for the distinct skeletal phenotypes remain undefined. Recognising the limitations of previous studies, human-induced pluripotent stem cells (hiPSCs) offer a new approach for inherited disease modeling. Their ability to differentiate into disease-relevant cells such as chondrocytes creates new opportunities for TRPV4-inherited skeletal disease modelling. Therefore, this PhD project aims to model the disorders caused by two TRPV4 mutations using hiPSCs and identify the pathogenic mechanisms underlying the two distinct TRPV4-inherited skeletal disease phenotypes. To obtain disease-relevant cells, establishing a robust and reproducible chondrocyte differentiation protocol is required. To do this, a reporter hiPSC line, SOX9-T2A-tdTom, was generated from MCRIi001-A (PB001.1) (1) using CRISPR/Cas9 genome editing. Thus in vitro chondrocyte differentiation could be monitored in real-time. The T2A linker and tdTomato fluorescent reporter gene were inserted downstream of the SOX9 coding sequence through homology-directed repair. The targeted allele was designed to produce SOX9 with the T2A sequence at the C-terminal end and a separate tdTom fluorescent protein. Genomic DNA sequencing of the SOX9-T2A-tdTom hiPSC line confirmed that the hiPSC line had one SOX9 allele with the T2A tdTom gene fusion and one wild type allele. Pluripotency was maintained as indicated by expression of pluripotency markers OCT4 and NANOG (immunostaining); CD9, CD326, and SSEA-4 (flowcytometry); and the ability to form tissues derived from three germ layers. SNP array showed there were no aneuploidies. The SOX9-T2A-tdTom hiPSC line had a similar capability to the parental line, MCRIi001-A, to form sclerotome. Western blotting showed that SOX9 protein expression was similar between SOX9-2A-tdTom and its parental line suggesting that adding tdTom gene sequence downstream of SOX9 gene did not disrupt the SOX9 expression and stability. The chondrocyte differentiation protocol was established using the SOX9-T2A-tdTom hiPSC line. Two stages of differentiation were performed. First, sclerotome induction was achieved by culturing hiPSCs in a 6-day multiple-step chemically defined culture mimicking embryonic development with pellet culture format was established on day 4. Secondly, chondrocyte differentiation was performed by transferring day-6 pellets into chondrogenic media in swirling culture format up to 10 weeks. A 4-week course of FGF2 treatment followed by an optional TGFB3 and GDF5 treatment until week 10 was performed during chondrocyte differentiation. RNA was collected every day during a 6-day sclerotome induction and at different time points during chondrocyte differentiation. The optimised protocol that involved a multiple-step chemically-defined 3-dimensional (3D) culture with swirling in an extended culture that included a 4-week FGF2 supplementation and optional subsequent TGFB3 and GDF5 treatment was able to generate cartilage that closely resembles fetal cartilage. CRISPR/Cas9 genome editing was also used to introduce two human TRPV4 mutations, a TRPV4 c.819C>G (p.F273L) mutation causing familial digital arthropathy with brachydactyly (FDAB) and a TRPV4 c.2396C>T (p.P799L) mutation causing metatropic dysplasia, into the SOX9-T2A-tdTom hiPSC line. For in vitro disease modelling, the mutant and their isogenic wild-type control (SOX9-T2A-tdTom) hiPSC lines were differentiated towards chondrocytes using optimised chondrocyte differentiation. The phenotypic differences between mutants and wild-type were assessed using various techniques including gene (RNA sequencing) and protein expression analysis. The two mutant cell lines and their isogenic wild-type control (SOX9-T2A-tdTom) were able to form cartilage. The pellet cartilage histology did not show any striking differences between the two mutants and their isogenic control. COL2A1 and TRPV4 protein expression was similar between mutants and control and this was consistent with the RNA sequencing data. RNA sequencing suggested that the pathogenic mechanisms underlying the two distinct TRPV4-inherited skeletal diseases were different. Compared to the isogenic control, F273L mutant cartilage had 263 differentially expressed genes. F273L cartilage showed a slight reduction in cartilage related gene expression including COL2A1, CSPG4, BGN, and CILP2. The F273L cartilage tissue was also less mature than the wild-type as indicated by increased SHH expression. On the other hand, P799L cartilage had more differentially expressed genes (655 genes) than F273L. MEF2C, the main regulator of chondrocyte hypertrophy, was upregulated in P799L. The hypertrophic chondrocyte markers such as RUNX2, SPP1 or osteopontin, and PTH1R, were also upregulated in P799L suggesting increased chondrocyte hypertrophy of P799L chondrocytes. The other characteristics of hypertrophic chondrocytes such as a reduction in cell proliferation and increased apoptosis were also observed in P799L cartilage. In conclusion, this study is the first study that conducts global gene expression analysis using RNA sequencing to characterise gene expression changes downstream of TRPV4 mutations in hiPSC-derived chondrocytes. The pathogenic mechanisms underlying the two distinct TRPV4-inherited skeletal diseases are different. The fewer differentially expressed genes in the F273L cartilage than in P799L suggests a milder disease phenotype. The slight reduction in cartilage marker expression in F273L cartilage might cause the cartilage tissues less resilient to physical forces thus leading to FDAB. In contrast, accelerated chondrocyte hypertrophic maturation can be the pathogenic mechanism underlying TRPV4 skeletal dysplasia phenotype. Accelerated chondrocyte hypertrophic maturation can disrupt growth plate development and cause systemic skeletal defects seen in patients. This thesis demonstrates that hiPSCs are a powerful tool to model inherited skeletal disease in vitro.

There is now considerable evidence indicating that risk of many complex diseases in adulthood may be influenced by exposure to environmental exposures in utero. A growing number of studies suggest epigenetic markers, including DNA methylation, are involved in this process. Understanding how DNA methylation is impacted by pregnancy exposures, and related to later health, may both contribute to unravelling the aetiology of complex disease risk in later life and provide a potential early-life biomarker for risk prediction. However, current evidence is limited. There has been a predominance of small, poorly powered studies, failure to consider the effects of genetic variation, and limited replication of previous findings. In addition, previous studies investigating the relationship between DNA methylation and offspring health have been primarily cross-sectional. For these reasons, I investigated the associations between pregnancy exposures (in particular, maternal smoking, nutrition and metabolic health, psychosocial stress, and adverse pregnancy conditions), birth outcomes, and offspring blood DNA methylation of the insulin-like growth factor 2 (IGF2) and H19, hypoxia-inducible factor 3A (HIF3A), leptin (LEP) genes. I also considered how genetic variation impacted on these associations. I then investigated the longitudinal relationship between early life methylation and anthropometry, as well as the association between early life methylation and later childhood measures of weight, adiposity, and cardiovascular health. To do this, the large, population-based longitudinal Barwon Infant Study pre-birth cohort (n=1,074) was used, with clinical and questionnaire measures from 28 weeks pregnancy, birth, 12 months post-birth and 4 years post-birth time points. DNA methylation of candidate regions was measured using the Sequenom EpiTyper mass-spectrometry platform in cord (birth) and peripheral (12-month) blood. Infant genetic variation in and near the candidate genes was considered. Infant adiposity was assessed as sum of triceps and subscapular skinfold thicknesses in infancy, and with DEXA scanning at 4 years of age. We found evidence that exposure to maternal psychosocial stress, gestational diabetes, and pre-eclampsia was associated with differences in offspring methylation at the candidate regions, as was infant sex. Genetic variation showed strong effects on DNA methylation levels, with some evidence for the associations of pre-eclampsia and infant adiposity with LEP methylation differing by infant genotype. Early life methylation of HIF3A and LEP showed modest associations with four-year blood pressure and BMI, respectively. While these associations persisted with adjustment for potential confounding factors, they explained relatively little variance in the four-year phenotypes compared to traditional predictors, such as weight. These findings suggest that offspring DNA methylation of these candidate genes involved in regulation of growth and metabolism are sensitive to several environmental exposures and genetic factors. While there is modest evidence for methylation in infant blood associating with later phenotypes, methylation of these genes appears unlikely to have useful predictive utility in isolation. This study is the first to perform early life longitudinal analysis to investigate the association between anthropometry and methylation in infancy. It is also the first to report evidence of earlier methylation associating with later cardiovascular phenotypes. However, as gene expression data was not available, the functional consequences of the altered methylation observed in blood is unclear. Further work is required to replicate these findings in independent cohorts, to determine the nature of expression of these genes in blood, and to investigate if the relationship between early life methylation and later health persists into adulthood.

Kidneys are uniquely vulnerable to drug toxicity due to their role in filtering substances from the blood. Filtration and urine production takes places in specialised tubules called nephrons, which are divided into segments with functionally distinct cell types that express transporters and receptors for transporting water, solutes, or small molecules from the blood into the filtrate, or from the filtrate into the cells. Should a drug have an affinity for these transporters, it can accumulate within these cells and cause injury. As a result, there are many commonly-used pharmaceutical compounds which are nephrotoxic to some degree, which can lead to acute kidney injury in patients. The nephron segment called the proximal tubule is the most frequently affected, as proximal tubule cells express a wide array of transporters and receptors which can transport drugs, and have a high metabolic rate which is vulnerable to disruptions in cellular energy production. Developing new drugs which are less nephrotoxic has been challenging, as the in vitro assays and animal tests currently used to screen drugs in development for nephrotoxicity have decreased transporter and receptor expression compared to mature human proximal tubules in vivo. Recent advances in understanding kidney development in the embryo have led to protocols for differentiating human pluripotent stem cells to kidney organoids, which contain segmented nephrons that have proximal tubules. Hence, kidney organoids may be useful for in vitro nephrotoxicity screening during drug development if they show sufficient nephron maturation. Through immunofluorescence and qRT-PCR analysis, we found that while proximal tubules in organoids expressed some proximal tubule markers to a greater degree than primary proximal tubule cells cultured in 2D, expression was still low overall compared to foetal kidney, and some transporters were absent. We then performed a bioinformatics analysis of single-cell RNA sequencing datasets generated from foetal and adult mouse and human kidneys to find signalling pathways which changed during proximal tubule maturation. Among other changes, mature proximal tubule cells showed increased fatty acid oxidation, a decreased capacity for glycolysis, and decreased TGF-beta/BMP signalling, suggesting that inducing these changes in vitro might improve proximal tubule maturation in organoids. To aid our characterisation of proximal tubules in organoids, we developed two reporter lines, LRP2:mTagBFP2 and HNF4A:YFP, for genes which are highly expressed in the developing proximal tubule, and confirmed that these genes and their respective reporters were co-expressed in the proximal tubules of kidney organoids. We also developed a HAVCR1:mCherry reporter line for future use in detecting injury responses in kidney organoids. These reporter lines allowed live monitoring of proximal tubule development and function. Using the HNF4A:YFP reporter, we screened for culture conditions which improved proximal tubule maturity in organoids and determined that a low-glucose, insulin-free medium designed to induce fatty acid oxidation supplemented with a small molecule inhibitor of TGF-beta improved the expression of proximal tubule markers in organoids, consistent with our bioinformatics analysis. We also performed xenotransplantation of reporter organoids into immunocompromised mice to determine whether proximal tubules would mature in this environment, and found that while non-renal tissue would proliferate in the grafts, proximal tubules in transplanted organoids maintained reporter expression better than organoids maintained in vitro under standard conditions. Taken together, this thesis characterises proximal tubule maturation and has begun to optimise this process within kidney organoids using novel reporter tools. This represents a significant advance in facilitating improved screening of drug-induced nephrotoxicity using organoids in the future.

The t(8;21) translocation generates the aberrant transcription factor RUNX1-ETO and occurs in approximately 10% of all acute myeloid leukaemias. RUNX1-ETO transcripts can be detected in utero and in cells of patients in remission, but its sole expression is insufficient to cause overt leukaemia. Given that t(8;21) patient cells present additional mutations, the epigenetic reprogramming directly mediated by RUNX1-ETO remains unclear. To address this question, we generated human Embryonic Stem Cell lines carrying an inducible RUNX1-ETO transgene, which we subsequently differentiated into definitive haematopoietic progenitors. We show that induction of RUNX1-ETO in already formed progenitors (i) blocks differentiation at an immature stage, (ii) induces a cell-type specific and reversible cell cycle arrest, (iii) abrogates the RUNX1-mediated gene expression program by interfering with RUNX1 binding, resulting in downregulation of haematopoietic, cell cycle as well as DNA repair genes, (iv) closes down a large part of the chromatin accessibility pattern present in adult haematopoietic multipotent progenitors and (v) alters the differentiation of a defined sub-population of progenitors. Our data are consistent with the idea that RUNX1-ETO establishes a precondition for leukaemic transformation by maintaining a reservoir of quiescent pre-leukaemic multipotent progenitors with susceptibility to expand upon acquisition of additional oncogenic events.

BACKGROUND In low- and middle-income countries (LMIC) where most of the 15.1 million neonates who survive prematurity and serious illness every year reside, neurodevelopmental outcome data are scarce and needed to improve neonatal care including follow-up, developmental monitoring and early intervention. In Fiji, neonatal health policy has shifted from almost exclusive focus on neonatal survival to ensuring that children are also supported to reach their developmental potential. METHODS AIM To examine early childhood neurodevelopmental and health outcomes for neonatal intensive care unit (NICU) survivors in Fiji, to inform improvements in neonatal care and follow-up including developmental monitoring and early intervention. OBJECTIVES These were, to: 1. Systematically review existing neonatal outcomes studies in LMIC to understand knowledge gaps related to neurodevelopmental outcomes for high-risk neonates and inform research design and sample size estimates for our neonatal outcomes study 2. Assess early childhood neurodevelopmental and health outcomes for neonates discharged from the Colonial War Memorial Hospital (CWMH) NICU in Suva, compared with control, term neonates born in the same hospital 3. Evaluate accuracy of nurse-led developmental screening as a potential tool for early identification of developmental delay and impairment in high-risk neonates, compared with reference standard developmental assessment RESULTS Systematic review of 60 high-risk neonatal outcome studies in LMIC provided estimates of median prevalence of neurodevelopmental impairment (NDI) for survivors of prematurity/very low birth weight (VLBW) and ‘birth asphyxia’. This highlighted a need for studies that better describe multi-domain neurodevelopment, include hearing and vision and including control data. Our neonatal follow-up study in Fiji compared early childhood outcomes for high-risk NICU patients (n=149) with those of matched term, normal birth weight neonates (n=147) discharged from Colonial War Memorial Hospital between November 2008 and April 2010. NDI was defined as at least one of; cerebral palsy, moderate to severe hearing or visual impairment, or global developmental delay using Bayley Scales of Infant and Toddler Development Third Edition (i.e. score <70 in at least two of cognitive, language or motor domains). At median (IQR) age 36.1 (28.3, 38.0) months, prevalence of moderate to severe NDI % (95% CI, n) in high-risk and control groups was 12 (5 to 17, n=13) and 5 (2 to 12, n=5), respectively, an increased risk ratio (95% CI) of 2.7 (0.8 to 8.9. Risk factors for NDI were identified. Our nurse-led developmental screening study demonstrated that, while feasibility of a parent-report screening tool was attractive, sensitivity for detection of global developmental delay was poor compared with the reference standard. CONCLUSION Our research provided the first neonatal neurodevelopmental outcome data in Fiji, adding to limited international literature on neonatal neurodevelopmental outcomes in LMIC. Measurement challenges highlight a need for international collaboration to improve measurement to better understand neonatal outcomes beyond survival. Negative findings of our developmental screening study indicate a need for longitudinal research to establish developmental monitoring approaches which facilitate early identification of developmental delay and impairment in routine services. Implementation research is also needed to develop innovative models of early intervention which support high-risk neonates in LMIC to thrive.