Minerva Elements Records

Permanent URI for this collection

Search Results

Now showing 1 - 10 of 7672
  • Item
    Thumbnail Image
    Novel seizure risk markers
    Chen, Zhuying ( 2022)
    Epilepsy is one of the most common severe neurological diseases and is characterized by recurring seizures. Currently, about 70 million people worldwide live with epilepsy and over 30% of them cannot be adequately treated with medication. The unpredictability of seizures is a severely debilitating aspect of epilepsy that significantly impacts the quality of life of patients. Consequently, there is a clinical need to find new markers that are useful for seizure forecasting. High-frequency activity (HFA) is a newly proposed biomarker for epilepsy, but its predictive value in seizure forecasting remains uncertain. Emerging new evidence has shown that ambient air pollution affects the central nervous system, but little is known about its effect on epileptic seizures. The goal of this thesis is to investigate potential novel markers for improving seizure forecasting and epilepsy management. To achieve this goal, the work of this thesis addresses several key questions: How do HFA rates and locations change over time, and how do these changes correspond with seizures? Can HFA forecast seizures? Is ambient air pollution associated with the risk of epileptic seizures? By addressing these fundamental questions, this thesis aims to provide the basis for formulating an innovative approach to improve seizure forecasting and control. In the first research chapter, the spatiotemporal profiles of HFA are investigated using long-term intracranial EEG. The results show that HFA rates have post-implantation variability, periodic cycles, and patient-specific relationships with seizures. These findings caution against using HFA as a presurgical metric without testing its reliability over time and suggest that tracking and utilising cycles of HFA rates may offer an exciting new opportunity to track cycles of seizures. In the second research chapter, a real-time phase estimation approach and seizure forecasting framework are developed using instantaneous HFA rates and phases of HFA cycles. The results show that HFA can be a useful biomarker to forecast seizures in patients with refractory epilepsy. The proposed real-time phase estimation approach can estimate the HFA phase over time with high accuracy and can be generalized to other seizure risk markers. In the third research chapter, ambient air pollutants are explored as potential seizure risk factors using a participant-time-stratified case-crossover design with conditional Poisson regression models. The results show that elevated ambient carbon monoxide (CO) concentrations, though within the Australian air quality standard, may be associated with increased risks of epileptic seizures; no significant associations were found in the other studied air pollutants (nitrogen dioxide, ozone, sulphur dioxide, and particulate matter less than 10 micrometers in diameter). These findings may have important clinical and public health implications and may offer potential new leads to improve seizure forecasting and prevention. Overall, these studies provide new evidence that HFA and ambient CO may serve as potential novel seizure risk markers. It is our hope that the work of this thesis contributes towards real-life seizure forecasting, informing new strategies to reduce the uncertainty of seizures, and eventually improving epilepsy management and the quality of life for people with epilepsy.
  • Item
    Thumbnail Image
    Understanding the early events in influenza A virus infection of human airway macrophages compared to epithelial cells
    Meischel, Tina ( 2022)
    Airway epithelial cells (AEC) and airway macrophages (AM) represent the major cellular targets for infection by different respiratory viruses. These include influenza A virus (IAV), a member of the Orthomyxoviridae family, and parainfluenza virus type-3 (PIV-3), a member of the Paramyxoviridae family. Virus-infected cells rapidly induce interferons (IFNs) which in turn activate hundreds of interferon-stimulated genes (ISGs) in infected cells and in neighbouring uninfected cells. The family of interferon-induced transmembrane (IFITM) proteins are ISGs that have shown broad antiviral activity against different virus families by interfering with virus entry. In this thesis, we evaluated the antiviral activity of human IFITM1, IFITM2 and IFITM3 against IAV and PIV-3, two respiratory viruses with distinct entry requirements. Using a doxycycline (DOX)-inducible IFITM overexpression system we found that all three IFITM proteins limited IAV infection to varying degrees, with IFITM3 having the strongest effect. Our studies confirmed that IFITM proteins display antiviral activity early, but not late, in the IAV replication cycle. None of the IFITM proteins modulated PIV-3 replication at either early or late-stages in the replication cycle. Different approaches were explored for their utility to modulate IFITM1, IFITM2 or IFITM3 expression to assess potential antiviral activity against IAV and PIV-3. Herein, we highlight caveats associated with constitutive IFITM protein overexpression as well as using CRISPR/Cas9 gene editing techniques to knockout endogenous IFITM expression. Observations herein also demonstrated that entry of IAV into epithelial cells by endocytosis conferred sensitivity to restriction by endosome-localized IFITM2 and IFITM3. Indeed, key steps in the IAV replication cycle from cellular entry through to exit are relatively well defined in epithelial cells. However, comparatively less is known about all stages of the viral life cycle during IAV infection of macrophages (M), particularly for primary human AM, where previous studies have largely been performed using mouse M. We investigated IAV entry into primary human M (human monocyte-derived macrophages (MDM) and airway macrophages (AM)) by defining the role of sialic acid (SIA) and C-type lectin receptors (CLRs) in determining susceptibility to the early stages of infection. Our results showed differential expression of SIA between MDM isolated from peripheral blood and AM obtained from bronchoalveolar lavage (BAL). MDM expressed both, alpha2,3- and alpha2,6-linked SIA, while AM expressed predominantly alpha2,3-linked SIA. AM were less susceptible to infection by IAV strains that preferentially recognised alpha2,6-linked SIA. Further, we found that although MDM and AM both expressed the CLR macrophage mannose receptor (MMR), it appeared that CLRs played a more prominent role in facilitating IAV infection of AM compared to MDM. We also systematically compared primary MDM and AM isolated from the same donor to show that both M types were susceptible to the early stages of IAV infection, but only MDM supported (limited) productive virus replication. Infection of AM by IAV resulted in strictly abortive replication for two different IAV subtypes, where newly-synthesised infectious virions were not released. Together, these findings make a substantial contribution to our overall understanding of IAV infection and replication in primary human M.
  • Item
    Thumbnail Image
    Development of Natural Language Processing Tools to Understand Veterinary Antimicrobial Usage Patterns
    Hur, Brian Alan ( 2022)
    Antimicrobial resistance is a global threat to which veterinarians contribute through their use of antimicrobials in animals. Antimicrobial stewardship programs have been shown to be an effective means to reduce antimicrobial resistance in hospital environments. Knowledge of antimicrobial usage patterns and the ability to automate the monitoring of these usage patterns are critical for the success of veterinary antimicrobial stewardship programmes. VetCompass Australia collects medical records from 181 clinics in Australia (as of January 2022). These records contain detailed information regarding the medications dispensed and the clinical notes made by the practitioner at the time of the consultation. Unfortunately, these notes are in the form of unstructured medical free text and therefore are not readily queried. Natural Language Processing is an area of computer science and Artificial Intelligence that can be used to provide structure to free-text clinical data so that it can be analysed at scale. This research project aims to develop and apply Natural Language Processing methods to apply structure to the data in VetCompass to enable the automated extraction of antimicrobial usage to inform veterinary antimicrobial stewardship programs. A transformer based deep learning model contextualized on VetCompass clinical notes, “VetBERT”, was developed. Use of VetBERT, along with development of methods to minimise the number of annotations required to train for various tasks, such as classifying the indication of an antimicrobial or labelling the treatments performed during a consultation enabled extraction of elements necessary to determine appropriateness of antimicrobial prescriptions. This included identification of the antimicrobial ingredient given at each consultation, as well as the route of administration, the indication, dose and duration for each antimicrobial prescribed. These methods were applied to over 4.4 million records from VetCompass from the years 2013 to 2017 inclusive. Veterinarians used mostly medium or low importance rated antimicrobials in their general practice. The exceptions were systemic use of cefovecin in cats and topical use of polymyxin. A clear description of the clinical indication for antimicrobial administration was present in approximately 85% of the clinical records, however fewer than 40% of the records had all the elements necessary to determine if the antimicrobial administration was appropriate (indication, antimicrobial agent, dose, duration of treatment). A novel application of this method was the use of NLP to extract temporal phenotypes of patient journeys for cats with abscesses. This method described the various interventions and treatments for each cat from the time of first presentation to the clinic until treatment for the abscess concluded, to better understand the effects of treatments performed at consultations and their impact on the antimicrobial usage behaviours of veterinarians. This research explores the development and application of these Natural Language Processing methods to overcome many of the difficulties of harvesting antimicrobial usage information out of free text data from clinical records to support antimicrobial stewardship programs in Australia.
  • Item
    Thumbnail Image
    Attention Orienting in Autism Spectrum Disorder
    Li, Shuting ( 2022)
    Attention orienting determines where we focus our concentration, which affects the information we process and our ability to interact with others. Atypical attention orienting is associated with some autistic behaviours. Experimental studies examining attention orienting in people with autism spectrum disorder (ASD), however, have reported inconsistent results. These unresolved experimental results lead to several questions. Is attention orienting atypical in ASD? What are the factors that contribute to atypical attention orienting in ASD? Does ASD alter neural circuitry implicated in attention orienting? The current thesis investigated attention orienting in ASD using a cross-disciplinary approach that combined human and animal investigative methods. The first study systematically reviewed the experimental literature using the Posner task to assess attention orienting in people with ASD. The results demonstrated a heterogeneous manifestation of attention orienting in people with ASD across different experimental settings and highlighted a biased focus on males and children. The second study examined exogenous (i.e., stimulus-driven) and endogenous (i.e., goal-driven) orienting in people with ASD using the Posner task with consideration of several potential confounders (alertness, co-occurring symptoms of attention-deficit/hyperactivity disorder (ADHD) and anxiety, age, and sex). Participants with ASD demonstrated quicker automatic shift of attention and slower voluntary disengagement of attention. We proposed that attention orienting in people with ASD is superior when attention is captured externally and inefficient when attention is controlled internally. Additionally, atypical attention orienting in ASD was found to be associated with age and co-occurring ADHD and anxiety symptoms, but not with alertness and sex. To investigate the consequences of genetic and pharmacological manipulations on cognition, the mouse model has emerged as a powerful tool. There does not, however, exist a well-established assessment of attention orienting in mice. The third study reverse-translated the human Posner task for use in mice using touchscreen technology. Consistent with human performance, mice responded more quickly and more accurately to validly compared to invalidly cued targets in both the exogenous and endogenous tasks, supporting the validity of the mouse Posner (mPosner) task. The effects of two dopaminergic- and noradrenergic-modulating drugs, methylphenidate and atomoxetine, were also examined. The results showed no significant effects of these two drugs on attention orienting, although methylphenidate improved response times during the exogenous orienting task. Following the development of the mPosner task, the fourth study examined attention orienting in mice carrying an ASD-associated neuroligin-3 (NL3)R451C genetic mutation and their responses to methylphenidate and atomoxetine treatment. NL3R451C mice demonstrated intact attention orienting. These mice, however, were more responsive to methylphenidate and less responsive to atomoxetine compared to wild-type mice. These findings suggest that the NL3R451C mutation is associated with disruptions of the dopaminergic and noradrenergic systems, motivating further clarification of the role of this mutation in specific dopaminergic and noradrenergic circuits. In sum, this research enhanced our understanding of attention orienting in people with ASD and created a novel translatable task to facilitate collaboration between human- and animal-focused researchers in investigating the neural mechanisms of attention in ASD.
  • Item
    Thumbnail Image
    Capturing Transience: Modelling Relationships Between Improvised Music Practice and Recording Processes
    McLean, Alistair James ( 2022)
    This research examines the relationship between improvised music practice and recording processes, and in doing so develops and tests new analytical models to better understand how improvised music practitioners undertake recording projects. Prior analytical models of music recording demonstrate multiple ways that recordings may be created and considered, but fail to take into account the diversity of practice in improvised music. By considering the varied nature of contemporary improvised music practise, these existing models are synthesised into a new Documentarian/Idealised model, which asks whether improvised music recordings are best considered as documents of performance events, discrete artistic objects, or a combination of both. Findings from interviews with improvised music practitioners are used to test and further develop the Documentarian/Idealised model, resulting in an expanded model better able to represent the diversity of practice found within improvising music recording projects, referred to as the Intention/Process model. Case studies of two improvised music recording projects are conducted as part of this research project, contributing ninety minutes of new improvised music recordings to be considered alongside the written thesis. These two projects reflect markedly different approaches to recording improvised music, and analysis of their creation examines the wide range of practice that occurs within improvised music recording situations. This research demonstrates that while improvised music recording practise is diverse, a number of commonalities are present, and that the intention and motivation of practitioners may be fluid and change during recording projects, as evidenced by a Multi-stage recording model for examining recording projects. In addition to providing multiple analytical models for use in further research, this study significantly informs both our understanding of how improvised music recording projects are undertaken and how they are perceived by practitioners of improvised music. It further contributes to the ontological understanding of improvised music recordings by arguing that improvisational music practice should not be viewed in opposition to composition or recording, but rather as a generative creative practice that can be utilised in tandem with other activities, and by showing that recordings of improvised music do not possess less improvisational qualities due to their fixed and reproduceable nature.
  • Item
    Thumbnail Image
    Supporting Teacher Practice in Early Childhood Science Education
    Guarrella, Cristina Maria Rosa ( 2022)
    References to the inclusion of science education in early childhood education are emerging in federal, state, and territory policy agendas. Although policy documents emphasise the importance of young children learning process skills typically associated with science, existing mandated frameworks do not explicitly refer to science. This has led to teachers seeking professional learning to better equip them to embed science within an informal curriculum. In response, the Northern Territory (NT) Government commissioned the development of the NT Preschool Science Games. The rollout of the resource in preschools across the NT provided the opportunity to conduct this intervention study. This research aimed to understand how to support teacher practice in early childhood science, with a specific focus on classroom quality, assessment, and playful learning during the implementation of the NT Preschool Science Games. Drawing on bioecological theory, a policy review of early childhood and science teaching learning policy documents is used to establish the broader context in which early childhood science is taught in Australia. A process skills approach to science teaching and learning is articulated, aligned with the emphasis on process skills identified in policy documents. Thereafter, a pilot study and intervention study are presented. The pilot study trialled two new instruments, the SciDoc and Early Childhood Science Padlet, along with a teacher questionnaire. These instruments were then refined and applied in the intervention study. Classroom observations were conducted to measure the quality of classroom interactions that contribute to child learning. Semistructured interviews identified teachers’ assessment practices, and the influences on these practices, during the implementation of the NT Preschool Science Games. Based on the findings, this research recommends the following supports for teacher practice in early childhood science education: 1. inclusion of science content alongside science process skills in the Early Years Learning Framework and all guiding documents; 2. professional learning to equip teachers to facilitate playful science learning; 3. learning progressions of science process skills; 4. implementation of the Assessment for Playful Learning model. Ultimately, when teachers are clear about what science learning is possible within playful experiences, they are better equipped to observe and assess children demonstrating what they know. This can inform authentic scaffolding and contingent planning for playful science teaching and learning.
  • Item
    No Preview Available
    Accessible therapeutic music-making for stroke survivors with significant arm and hand weakness: A mixed-methods study
    Silveira, Tanya Marie ( 2022)
    This thesis with publication presents the results of the mixed methods study exploring a novel music therapy intervention for stroke survivors with significant weakness to their arm and hand. Using a mixed methods experimental design, with an explanatory sequential core, this randomised controlled trial sought to examine the holistic impact of a 4-week intervention protocol using functional electrical stimulation (FES) together with an iPad-based instrument (ThumbJam) on stroke survivors’ upper limb function and wellbeing outcomes. Recognising the need for more accessible approaches to music-making with this subset of stroke survivors, the intervention protocol was developed using collaborative processes drawing on knowledge from the disciplines of physiotherapy, occupational therapy and music therapy. After securing ethics clearances, recruitment commenced across five hospitals in Sydney, Australia, aiming to recruit a target sample of 40. Fourteen participants were recruited and randomised to receive usual treatment (n=8) or the daily FES+iPad-based music therapy intervention as an addition to usual treatment (n=6) for four weeks (20 sessions). Masked assessors administered the standardised measures of upper limb function and self-report wellbeing questionnaires at three time points (pre- and post- the intervention period, and at three months follow up). All participants were also interviewed at the post-intervention period regarding their perception of how their received treatment supported their overall recovery. The Motor Assessment Scale (MAS-UL) was the primary outcome for arm/hand function. The other measures of arm/hand function included the Manual Muscle Test (MMT-UL), 9-hole-peg test (9HPT) and grip dynamometry. The Depression Anxiety and Stress Scale (DASS-21), and the Stroke Self-efficacy Questionnaire (SSEQ) were used to measure wellbeing. As this study was underpowered, mean change scores, confidence intervals and effect sizes (Hedges’ g) were calculated and reported. The intervention group showed greater improvements than control on all upper limb measures, with between group differences on the MAS-UL change score of 2.08 (95% CI -2.08, 6.96; g = 0.5), 0.05 for the 9HPT (95% CI -0.13, 0.23; g = 0.32), 3.33 for the MMT-UL (95% CI -1.26, 7.93; g = 0.85), and 3.68 for grip dynamometer (95% CI -0.70, 8.07; g = 0.99). The intervention group also showed greater decreases in anxiety (between group difference: -1.83; 95% CI -7.63, 3.97; g = 0.37), but lesser reductions in depression (2.25; 95% CI -7.71, 12.21, g = 0.27). There were no notable differences between groups for stress and self-efficacy. Reflexive thematic analysis of the qualitative interview data revealed different reflections about the treatment received by each group, with intervention participant themes focusing on their perceived improvement in upper limb function and strength, as well as the motivating and relaxing aspects of musical engagement. These integrated findings suggest that FES+iPad-based music therapy has the potential to simultaneously improve post-stroke upper limb function and wellbeing. Therefore, this pilot study supports the need for future research that is adequately powered for efficacy testing.
  • Item
    Thumbnail Image
    Modest expectations: masculinity, marriage, and the good life in urban China
    Gosper, Sarah Maree ( 2022)
    There is a sense that there is a crisis unfolding in China. Marriage rates are dropping, divorce rates are rising, the birth rate is in decline, and a new population of rural ‘bachelors’ and urban ‘leftover women’ has surfaced. This new culture of singlehood is perceived as a ‘crisis of marriage’, precipitating a moral panic over how to address a problem that is often described by the state as a threat to social stability and order, as well as the advancement of the nation. This thesis explores the intersection of these so-called ‘crises’ facing Chinese society: a crisis of marriage, a crisis of masculinity, and a crisis of mobility. Since China’s ‘opening up and reform’ in 1978, extraordinary social, economic, and political change have occurred. Gender and sexual relations have also undergone significant transformation, subsequently contributing to this ‘marriage crisis’ in China today. How single rural men living in the city respond to this marriage crisis is a core concern of this thesis. In the gendered aspects of this crisis and the associated moral panic, single rural men have become a flash point in China for discussions about marriage, social organisation, the rural–urban divide, gender relations, class, and mobility. The demise of the rural economy and the rapid transformation of the urban economy have produced significant changes in gender roles and institutions in contemporary China. This thesis focuses on the impact of these socio-economic shifts on rural men who migrate to cities. Rural to urban migration has a long and well-documented history in China. The most recent wave of migration has been accompanied by changes in the nature of work and social organisation that have exacerbated the ‘marriage crisis’ particularly for rural men in urban settings. For rural men living in urban China, marriage represents a modest aspiration for a good life, expressed through the concept guo rizi (passing the days). The desire to marry and have children is however constrained by rural men’s experiences in the city. Their occupations, lack of social networks, new forms of dating and matchmaking and increasingly unattainable ‘bride-price’ demands, work together to undermine their desirability as potential husbands and fathers and entrench inequalities of wealth and power between rural and urban men. The ways rural men struggle with, negotiate, and imagine their futures is the subject of this thesis. I argue that the increasing socio-economic precarity of rural men and their largely unrealised desires to marry and have children demonstrate a fundamental reconfiguration of Chinese masculinity and mobility in urban China today and the social impact on central Chinese institutions. This thesis explores the lives of migrant delivery drivers (kuaidi and waimai) and tertiary-educated professionals who have migrated from the countryside to the city. In this thesis, I endeavour to make these men visible by investigating how they navigate the urban marriage market and avoid becoming ‘leftover’. What I have found is that their shared struggles in the marriage market and efforts to fulfill the ideals of manhood are indicative of how rurality continues to be experienced as an inhibiting factor for single rural men in Chinese cities, regardless of their education, income, or material assets. The nature of these men’s lives led me to question how such men are affected by changing social, cultural, and economic structures within the marriage market and the broader context of crisis that currently pervades Chinese society.
  • Item
    Thumbnail Image
    Identifying the Risks and Drivers of Aggression in Prostate Cancer
    Chow, Sing Ken ( 2022)
    Prostate cancer is the most commonly diagnosed cancer and the second commonest cause of death in Australian men. The natural history of localised prostate cancer is markedly variable. Most individuals will lead a biologically indolent course where it may never be clinically significant while an unpredictable minority of individuals, approximately 10%, will rapidly progress to metastases and eventually succumb to the disease. Currently there are no adequate methods in determining the difference between an indolent or aggressive course of the disease at an early stage; therefore, leading to significant over-treatment of indolent disease as well as under-treatment of aggressive disease. The identification of potential drivers of aggression in prostate cancer are explored through the compilation of published papers in an integrated thesis by investigating the impacts of obesity, velocity of biochemical recurrence, aggressive ductal adenocarcinoma variant, as well as the molecular comparison between primary and matched metastatic tumours. Obesity is linked with more aggressive prostate cancer and higher rates of disease recurrence post treatment. It is unclear if this is due to specific tumour-promoting effects of obesity or diagnostic bias. The effect of obesity on the accuracy of pre-treatment risk categorisation was determined, and mediation analysis was used to identify the contribution of biologic versus non-biologic mechanisms to the observed increased risk of biochemical recurrence. The analysis included 1587 patients. Despite similar rates of adverse pathological features at prostatectomy, biochemical recurrence rates were significantly higher in very obese patients, which persisted after adjustment for stage, grade and PSA. Tumour volume however correlated significantly with BMI (p = 0.004), and the difference in predicted and observed ‘tumour attributable’ PSA (Delta-PSA) in very obese patients was greater than three times higher than that of healthy patients (p = 0.0067). Regression analysis indicated that the effect of BMI on tumour volume was fully mediated indirectly by its effect on PSA. Inclusion of this diagnostic error as a covariate in the survival analysis attenuated the effect of BMI on recurrence. Being very obese suppresses tumour-associated PSA resulting in a diagnostic bias that is responsible for errors in risk classification, and potentially contributes to a delay in initial presentation. Radical prostatectomy is one of the preferred treatment modalities for localised prostate cancer. Although the majority of patients experience long term disease control, depending on the pre-treatment clinical characteristics of the cohort under study, up to a third of men will develop disease recurrence. The most common manifestation of disease recurrence is a detectable serum PSA in the postoperative period. This thesis characterises the pattern of late disease recurrence in the largest contemporary cohort of localised prostate cancer patients treated with radical prostatectomies in the active surveillance era. Total of 2312 patients were included in the final analysis with up to 12 years of follow up data. The average patient had clinically localised prostate cancer, an elevated PSA, and ISUP grade group 2 on biopsy. 88.7% of patients had ISUP grade group 2 or higher at prostatectomy. A subgroup of 446 patients had undetectable PSA levels at 5 years after prostatectomy; 11.7% of them progressed to experience biochemical recurrence. In this subgroup, late recurrers had significantly higher-grade tumours on ISUP and Gleason sum (p < 0.001 and p = 0.001, respectively), higher rates of extraprostatic extension (p = 0.022), and larger tumour volumes (p = 0.032). Logistic regression demonstrated that prostatectomy ISUP grade group was a significant predictor (OR 2.14, 95% CI 1.43-3.20, p < 0.001). Additionally, the timing of recurrence resides on a continuum of risk and that the potential concept of dormant micrometastatic involvement requires further research and evaluation. Ductal adenocarcinoma is an uncommon prostate cancer variant. Previous studies suggest that ductal variant histology may be associated with worse clinical outcomes, but these are difficult to interpret. Prostatectomy patients with ductal variant histology from two institutional databases were identified and compared to an independent acinar adenocarcinoma cohort. A total of 202 ductal adenocarcinoma and 2037 acinar adenocarcinoma cases were analysed. Deep whole exome sequencing was performed in selected cases (n = 8). Survival analysis after matching demonstrated that patients with ductal variant histology had shorter salvage-free survival (8.1 vs 22.0 months, p = 0.03) and metastasis-free survival (6.7 vs 78.6 months, p < 0.0001). Ductal variant histology was consistently associated with RB1 loss, as well as copy number gains in TAP1, SLC4A2 and EHHADH. The presence of any ductal variant adenocarcinoma at the time of prostatectomy portends a worse clinical outcome than pure acinar cancers, with significantly shorter times to the initiation of salvage therapies and the onset of metastatic disease. The exploration of the impacts of obesity, velocity of biochemical recurrence, aggressive ductal adenocarcinoma variant, as well as the molecular comparison between primary and matched metastatic tumours to provide insights into potential drivers of aggression in prostate cancer.
  • Item
    Thumbnail Image
    Molecular mechanisms governing the phase separation of globular proteins
    Choi, Yoon Hee ( 2022)
    Intracellular protein aggregate deposits are a key pathology observed in neurodegenerative diseases, including Alzheimer’s, Parkinson’s, Huntington’s disease and Amyotrophic Lateral Sclerosis (ALS). In the case of ALS, aggregates are composed of a number of distinct proteins, such as the intrinsically disordered RNA-binding protein, FUS, or the globular protein superoxide dismutase 1 (SOD1). Recent studies have attributed the mechanism of aggregation of intrinsically disordered proteins to phase separation. Phase separation is a biophysical property of heteropolymers where intramolecular interactions between macromolecules such as proteins make it thermodynamically favourable for a solution to demix into two coexisting states that can have liquid-like or solid-like properties. This framework has enabled the study of the driving mechanisms of aggregation of intrinsically disordered proteins using biophysical principles to better understand protein biology. Unlike intrinsically disordered proteins, the phase separation of intrinsically foldable proteins has not been directly explored. Therefore, I have explored the aggregation of globular proteins in a phase separation framework. The phase separation of globular proteins was shown to be dependent on the interplay between protein stability, sequence composition of the unfolded protein and the interaction with molecular chaperones. This was primarily explored using the catalytically inactive model barnase protein combined with an optogenetic system called “optoDroplets,” that showed the phase-separation propensity was dependent on the amount of protein in the unfolded conformation. Exposed aromatic residues were identified as drivers, or stickers, of this phase behaviour, where the addition or removal of aromatics in an unfolded variant of barnase increased or reduced the phase separation propensity of barnase, respectively. HSPA1A and DNAJB1 were shown to inhibit phase separation by selectively binding to unfolded barnase. Furthermore, the biological importance of unfolded protein phase separation was also explored by examining the impact of barnase sequence variation on aggregate composition in cells. A subset of protein quality control proteins such as HSPB1 and DNAJC9 were shown to be enriched in all barnase aggregates, while a unique composition profile was observed based on the sticker valence of the aggregating barnase. Computational analysis of enriched proteins in aggregates was consistent with enrichment of proteins being driven by their sequence features. Overall, an equilibrium triad between protein folding, binding and phase separation is suggested for globular proteins and it is suggested this phase behaviour is biologically significant due to the interaction of phase-separated aggregates with the endogenous proteome. The proposed phase separation paradigm of globular proteins and the identification of its key determinants can provide insight into the mechanism of protein aggregation and its toxicity that could have significant implications in disease.