Sir Peter MacCallum Department of Oncology - Theses
Permanent URI for this collection
Search Results
1 - 10 of 151
-
ItemCharacterising the molecular heterogeneity of neuroendocrine prostate cancer( 2023-04)Prostate cancer is primarily made up of androgen receptor-driven adenocarcinoma, but some patients may progress to a more aggressive and lethal neuroendocrine prostate cancer (NEPC) phenotype lacking AR expression. NEPC is highly heterogeneous, posing a challenge in detection and treatment. While mutations such as RB1, PTEN, and TP53 have been identified as genomic alterations in NEPC progression, they do not fully explain the heterogeneity and varied response to treatment. Understanding the molecular drivers behind the heterogeneity and therapy resistance of NEPC is vital in improving detection and treatment. Therefore, the main goal of the thesis is to characterise the transcriptional heterogeneity of tumour cells of NEPCs and the cells of the tumour microenvironment (TME), to understand the biology of these tumours and determine whether there are any consistent therapeutic targets. I hypothesise that by characterising the different levels of tumour heterogeneity in NEPCs, new insights into how to treat these tumours can be detected. To do so, I focus on three main questions 1) What is the level of intra-tumour heterogeneity of NEPC pathologies? 2) Are there common and/or distinct transcriptional profiles among NEPC pathologies? 3) How does the tumour microenvironment (TME) vary across different pathologies in NEPC? To extensively analyse the transcriptional heterogeneity of NEPC, single-cell RNA sequencing technology was used on a novel cohort of nine patient-derived xenografts (PDX) models that recapitulate the pathological and clinical heterogeneity of NEPC. I analysed scRNA-seq data from 18,632 cells from 9 patient-derived xenografts (PDX) models of NEPC to detect transcriptional heterogeneity between sub-populations. PDXs are a powerful tool to analyse the heterogeneity of NEPC due to their ability to retain the genomic and phenotypic features of the original tumour in vivo. The opportunity to study rare types of NEPC using PDX models is particularly valuable, as it allows for a more comprehensive understanding of the disease. Additionally, the feasibility of using PDX models for NEPC research is enhanced by the availability of fresh tissue with more viable cells for single-cell experiments. This makes PDX models an attractive option for studying the disease's biology and potential treatment options in a more comprehensive and detailed way. The thesis is divided into three distinct studies. The first study comprehensively characterises the degree of intra-tumour heterogeneity in each NEPC sample. This is achieved through a rigorous quality control process and standardisation of a single-cell RNA sequencing pipeline, which accurately assesses the levels of heterogeneity. Analysis revealed the presence of 3-8 subpopulations in each sample. The degree of heterogeneity changed depending on the pathology; small cell NEPC showed more heterogeneity than any other pathology. One key finding is that at least one chemo-resistant cluster was detected in each pathology, and those clusters showed a unique transcriptional profile. These findings suggest that the intra-transcriptional heterogeneity of NEPC could be a critical factor in identifying potential therapeutic targets. The second section of the thesis aims to examine the inter-patient heterogeneity of NEPC. The analysis of NEPC cells revealed that certain pathologies exhibit a similar transcriptional profile, while others show a closer resemblance to adenocarcinoma. Although these pathologies share the expression of neuroendocrine markers, there is evident heterogeneity in the underlying biology driving each phenotype. Both small and large cell neuroendocrine pathologies exhibit enriched LEF1, YAP, Notch and NMYC signalling pathways and hallmarks of aggressiveness, which are associated with poor prognosis for patients. Notably, the NED cell populations revealed a distinct profile of enriched markers and pathways, such as TNFA via NFKB and KRAS signalling. Unexpected cell co-expression of both neuroendocrine and adenocarcinoma markers was observed in NED pathologies. The co-expression analysis has uncovered this unexpected cell plasticity in NEDs, exposing that they retain adenocarcinoma features and exhibit a transcriptional profile similar to adenocarcinoma. The data indicate the presence of two molecular profiles of NEPC; it is important to consider these molecular profiles in developing personalised therapeutic strategies, as they can potentially improve patient outcomes and response to treatment. The third section of the thesis is dedicated to the investigation of the TME of NEPC. The study focused on murine non-cancerous cells extracted from tumours collected from PDXs. The analysis revealed the presence of heterogeneous subpopulations of fibroblasts, including the novel antigen-presenting CAF, as well as the inflammatory and myofibroblasts CAF populations. Based on their gene expression profile, these cells may represent a dynamic state of cancer-associated fibroblasts (CAFs), whose role is likely to be influenced by cell-to-cell interactions within the TME. Notably, a significantly increased proportion of myofibroblasts (myCAFs) and inflammatory CAFs (iCAFs) were observed in the small and large cell pathologies. This suggests that the TME may play a crucial role in the progression and aggressiveness of NEPC and highlights the importance of considering the TME when developing therapeutic strategies for these patients. The detection of transcriptional sub-populations in NEPC has provided insights into the underlying molecular mechanisms that make finding effective treatments for these tumours so challenging. These sub-populations exhibit unique gene expression profiles, resulting in a high degree of heterogeneity within NEPC. This heterogeneity is likely a key contributor to the failure of standard prostate cancer treatments in NEPC, as the different sub-populations may respond differently to various therapies. Furthermore, these sub-populations may play a role in the resistance to existing treatments, making them potential targets for the development of new therapeutics. Overall, the characterisation of the transcriptional heterogeneity of NEPC represents a significant step towards developing more personalised and effective treatments for this aggressive cancer.
-
ItemDiscovering the molecular basis of CRISPR3-Cas13b for precise silencing of tumour transcripts( 2023-03)Cancers are caused by the accumulation of genomic mutations and gene expression dysregulations. In many patients, standard therapies such as chemotherapy often result in severe toxicity due to off-target effects that takes place in non-tumour tissues. Additionally, certain genetic aberrations confer resistance to these standard therapies and mediate cancer relapse. Recently, advanced sequencing technologies led to a better understanding of tumour genomics, which opened up a new era of precision and personalized medicine. These insights from tumour genomics enabled the identification of several hundreds of aberrant oncogenes called tumour drivers that confer clonal expansions and malignancy, which are considered high-priority targets of precision oncology. Currently, there are two main precision medicine approaches to target tumour drivers. The first is adoptive cell transfer involving in vitro engineering patients’ immune cells to recognize tumour antigens and lyse tumour cells, but this approach is restricted to tumour antigens on the cell. The second is the use of small inhibitory molecules and antibodies to target tumour drivers at a protein level. However, the drug development process using conventional pipelines is time-consuming, limiting the number of cancer drugs that are available to target newly identified tumour drivers in a personalized manner. These limitations highlight the need to develop a new design-flexible precision medicine approach to target tumour drivers in a personalized manner. The recent breakthroughs in the field of CRISPR (Clustered Regularly Interspaced Short Palindromic Repeat) offer a promising opportunity to precisely target tumour driver genes. Theoretically, the most extensively studied CRISPR-Cas9 can cleave any oncogene and induces its loss of function. Undoubtedly, it has been a powerful tool to study cancer functional genomics in basic research with immense potentials as modular cancer therapeutics. Nevertheless, safety concerns due to off-target effects and irreversible chromosomal instabilities largely constrain the therapeutic applications of classical CRISPR-Cas9. On the other hand, the discovery of a novel RNA-guided RNA targeting CRISPR-Cas13 may address some of the limitations of Cas9 to target tumour drivers at the RNA level. In the introduction of this thesis (Chapter 1), I discuss the molecular basis of various CRISPR effectors with special focus on RNA-targeting enzymes. I also highlight how recent scientific discoveries in CRISPR filed translate into various tumour precision medicine applications. Finally, I chart progress in the CRISPR-Cas13 field and discuss its advantages and limitations as a programmable tool to target tumour drivers in a personalized manner. In my Ph.D. research project, we hypothesized that the CRISPR-PspCas13b nucleoprotein complex may be re-engineered to silence tumour driver transcripts with single-nucleotide accuracy. However, the poor understanding of the molecular principles governing PspCas13b target recognition and cleavage limits its utility and development as programmable cancer therapeutics. Thus, this research focuses on revealing the molecular mechanisms of PspCas13b in order to reprogram this ribonuclease to target tumour transcripts that are currently ‘undruggable’. Excitingly, our Single-Base Tiled crRNA screens (SiBTil), unbiased computational analysis, and comprehensive spacer-target mutagenesis revealed key determinants of PspCas13b activity. De novo design of crRNAs harbouring base-paired or mismatched guanosine bases at key spacer positions greatly enhances the silencing efficacy of otherwise inefficient crRNAs, expanding the targeting spectrum of this enzyme. We also reveal the interface between mismatch tolerance and intolerance, which unlocks an unexpected single-base precision targeting capability of this RNA nuclease. Notably, our de novo design principles enable potent and selective silencing of various gene fusion transcripts and their downstream oncogenic networks, without off-targeting of non-translocated variants that share extensive sequence homology. We demonstrate that PspCas13b targeting the breakpoint of fusion transcripts enables efficient suppression of ancestral and single-nucleotide mutants (e.g., BCR-ABL1 T315I) that often drive clinical cancer relapse. Our transcriptomic and proteomic analyses suggest PspCas13b is highly specific and has no global off-targeting or collateral activity against endogenous transcripts and proteins. Collectively, my Ph.D. research provides new design principles for PspCas13b programming to specifically recognise and degrade any ‘undruggable’ fusion oncogenic transcript, thus providing a new conceptual framework for personalized oncology.
-
ItemIdentifying heterogenous mechanisms of drug resistance in high grade serous ovarian cancer( 2023-05)Ovarian, fallopian tube and primary peritoneal cancer is the sixth most common cause of cancer death for women in Australia. A diagnosis of ovarian cancer is life-altering with a five-year survival of only 45.7%. While clinical factors such as disease stage are important for prognosis and response to therapy, the underlying genomic, transcriptomic and immune characteristics of the tumour are also key determinants of outcome. High grade serous ovarian cancer (HGSC) is the most common epithelial ovarian cancer subtype, and is characterised by frequent germline, somatic and epigenetic aberrations which result in homologous recombination DNA repair (HR) deficiency. Aneuploidy and frequent somatic copy number aberrations (SCNA) are common features of the genomic landscape of HGSC. Despite often initially responding to therapy, acquired resistance is extremely common, arising through a variety of genomic, transcriptomic, adaptive and microenvironmental mechanisms. Owing to the scarcity of biopsies or surgery in relapsed disease, the frequency and distribution of these acquired resistance mechanisms is only partially known. In particular, there is a paucity of information about resistance mechanisms in late-stage disease following the multiple lines of therapy that a typical patient with HGSC receives. This thesis addressed these important gaps in knowledge by examining resistance mechanisms in a cohort of women with HR deficient HGSC who underwent a research autopsy, providing unique insight into the architecture of end stage disease. Importantly, resistance mechanisms were frequently subclonal, often with more than one mechanism detected within a single metastatic site. Evidence of convergent evolution was also observed, including of reversion mutations and BRCA1-specific mechanisms of HR restoration. An increased frequency of whole genome duplication, an important event in cancer evolution, was noted in the end-stage samples compared to what is known in primary tumours. Following this observation, the transcriptomic differences between tumours with and without WGD were assessed. Examination of the transcriptomic differences in whole genome duplication revealed downregulation of CIITA and other MHC-II genes, highlighting a potential means of immune escape and a potential therapeutic vulnerability. This thesis also examined mechanisms of resistance specifically occurring following receipt of poly-ADP ribose polymerase (PARP) inhibitor therapy, and the natural history and evolution of reversion mutations in HR deficient HGSC, showing that they are in fact not frequently present outside of a progressing or resistant cohort. The findings of this thesis add to the current knowledge of resistance to therapy, demonstrate the need for accessible genomic testing in the clinical management of HGSC and indicate that further exploration of resistance should focus on transcriptomics, epigenetics, and the tumour microenvironment.
-
ItemElucidating epigenetic mechanisms of cancer immune evasion( 2023-03)Cancer immunotherapies have revolutionised the management of a wide range of haematological and solid organ malignancies, due to their potential to induce durable remissions in a proportion of responders. However, primary or acquired resistance remains problematic for the majority of patients, and typically arises from tumour-intrinsic properties that reduce immunogenicity or extrinsic factors promoting an immunosuppressive tumour microenvironment. Effective tumour antigen presentation via major histocompatibility proteins (MHC-I and/or MHC-II) to immune effector cells is a critical component of the adaptive anti-cancer response and genetic disruption of the MHC-I and/or MHC-II antigen presentation pathways, either through inactivating mutations or transcriptional silencing, is a well-recognised cause of resistance to both pharmacological and cellular immunotherapies. In this thesis, I explore epigenetic mechanisms of MHC-I and MHC-II repression in cancer and identify an evolutionarily conserved role for Polycomb repressive complex 2 (PRC2) in MHC-I silencing. I also discover two key mechanisms of MHC-II regulation in acute myeloid leukaemia and melanoma: transcriptional repression of MHC-II pathway genes by the C-terminal binding protein (CtBP) co-repressor complex, and post-translational regulation of CIITA, the master regulator of MHC-II expression, by the FBXO11-containing E3 ubiquitin ligase complex. Targeting of these repressive pathways efficiently upregulates cell surface MHC expression and augments in vitro and in vivo adaptive immune responses. These findings provide valuable biological insights into mechanisms of cancer immune evasion and establish the scientific rationale for further pre-clinical and translational studies of these novel therapeutic strategies to overcome immunotherapy resistance via restoration of tumour antigen presentation.
-
ItemAvenues to enhance chimeric antigen receptor (CAR) T cell stemness and therapeutic efficacy against solid tumours( 2023-02)Chimeric antigen receptor (CAR) T cell therapy has demonstrated remarkable activity in B cell malignancies leading to multiple Food and Drug Administration (FDA) approvals. However, in solid tumours, CAR T cells have only yielded limited efficacy. This is thought to be due to several factors including poor persistence of CAR T cells and terminal differentiation. In the present thesis, two approaches for improving the persistence and efficacy of CAR T cells against solid tumours were explored. The first strategy involved overexpressing T cell memory-associated transcriptional regulators in CAR T cells. Clinical data has shown that CAR T cell persistence and therapeutic outcomes are associated with the adoptive transfer of less differentiated, “stem-like” CAR T cells that assume a similar phenotype to circulating memory T cell subsets. I hypothesised that the overexpression of the transcriptional regulators associated with memory T cells would reduce CAR T cell differentiation, thereby leading to their improved persistence and enhanced anti-tumour efficacy. FOXO1 was identified as a primary candidate, and the impact of overexpressing a constitutively active form of Foxo1 (Foxo1-ADA) in HER2 directed murine CAR T cells and wild type FOXO1 in human Lewis Y directed CAR T cells was evaluated. Foxo1-ADA overexpression improved the therapeutic efficacy of CAR T cells in breast and colon carcinoma tumour models. Additionally, Foxo1-ADA overexpression enhanced CAR T cell polyfunctionality and mitochondrial health in vivo. I found that the overexpression of wild type FOXO1 similarly improved the stem-like characteristics of human CAR T cells, enhancing overall respiratory capacity and therapeutic responses in a human ovarian cancer model. Our data indicates that reinforcement of a stem-like program through FOXO1 overexpression significantly enhances CAR T cell responses against solid tumours. The second approach involved the use of a dual-specific CAR T cell that expressed two distinct CARs against a solid tumour and B cell antigen. The success of CD19 directed CAR T cells may be attributed to engaging antigens outside of a solid tumour microenvironment (TME). Therefore, this approach aimed to co-opt these benefits in the solid tumour setting by co-transducing a unique T cell designated as a ‘dual CAR T cell’ with two CARs having specificity for the human HER2 and murine CD19 antigens. I hypothesised that by first engaging the CD19 antigen in sites away from the immunosuppressive TME, dual CAR T cells would be primed to better mount responses against solid tumours. Dual CAR T cells demonstrated higher polyfunctionality in vitro against HER2 expressing tumour cells when primed through the CD19 directed CAR and mediated enhanced therapeutic efficacy against orthotopic breast tumours in vivo. Notably, dual CAR T cells demonstrated improved persistence with a similar differentiation phenotype relative to control T cells. These results indicate that maintenance of a less differentiated CAR T cell phenotype and engagement of target antigens outside of the TME are instrumental for maintaining CAR T cell persistence. These investigations contribute to the growing body of work that aims to enhance CAR T cell stemness. As these studies have been conducted in both syngeneic and human CAR T cell models, I believe that this work carries high translational potential.
-
ItemNo Preview AvailableInvestigating the function of dedifferentiation regulators in the context of regeneration( 2023-03)Neural stem cells (NSCs) are fundamental to the development of our brain. All specialised cells, such as neurons and glia, are generated from dividing NSCs to populate the central nervous system (CNS). The differentiation of neurons and glia is crucial for the formation and guidance of functional neural circuits. However, this process can be quite precarious, as mature cells can slide back along their lineage to less mature and more ‘stem-like’ states (dedifferentiation). When dedifferentiation occurs by random de novo mutations, it can lead to tumour formation; this is responsible for multiple cancer types. However, a form of dedifferentiation can also occur when mature cells are reprogrammed upon injury, becoming more stem-like to create progenitors and regenerate. This process is tightly regulated and driven by controlled genetic changes of a mature cell type. Even though humans have this genetic program, barriers are in place within our CNS to prevent regeneration from occurring; consequently, CNS damage is often permanent. Finding ways to increase human CNS regeneration may therefore be a therapeutic strategy to restore neuron loss and CNS damage. Interestingly, regeneration in mammals can be increased by mimicking regenerative processes of animals capable of efficient regeneration. Together, the study of dedifferentiation and cell fate is crucial to further understanding regeneration as well as the origins of tumourigenesis. In this thesis, I address dedifferentiation in both of these scenarios by utilising the combined strengths of the model organisms: Drosophila melanogaster (fruit fly) and Danio rerio (zebrafish). My first aim is to identify novel regulators of dedifferentiation within the Drosophila brain, and my second aim is to understand the identity, progeny, and malignancy of ectopic NSCs created via dedifferentiation. Focusing on regeneration, my third aim is to characterise inhibitory neuron loss and recovery within the zebrafish retina; and my fourth aim is to uncover whether the novel Drosophila dedifferentiation regulators play roles in zebrafish regeneration. Factors maintaining neuronal identity while preventing dedifferentiation are crucial. Within the Drosophila visual processing system of the brain, the medulla, only a handful of neuronal identity regulators are known. One of which is Nervous fingers-1 (Nerfin-1), whose loss results in the dedifferentiation of medulla neurons and subsequent tumourigenesis. Recent studies discovered that Nerfin-1 maintains neuronal identity with its cofactor, Hippo pathway member, Scalloped (Sd). Additionally, a targeted Dam identification (TaDa) was performed to uncover targets of the Nerfin-1/Sd complex in medulla neurons. Within this thesis, I first screened candidate Nerfin-1/Sd targets to ascertain their role in dedifferentiation. I found four novel factors that induced dedifferentiation upon their misexpression: Embryonic lethal abnormal vision (Elav), Deadpan (Dpn), Prospero (Pros), and SoxNeuro (SoxN). Throughout this thesis, I further examine the function of these genes in dedifferentiation, including characterising the ectopic NSCs that are induced upon their misexpression and investigating their function in regeneration. Within Chapter 2 (Aim 1), I discovered that the misexpression of Elav, Dpn, Pros, and SoxN induced dedifferentiation in varying neuronal populations and these dedifferentiated cells had unique characteristics. This outlines the complexity and unique nature of ectopic NSCs derived from dedifferentiation and suggests that a ‘one-size fits all’ therapeutic strategy may not be efficient for the treatment of different dedifferentiation-induced cancers. In the next chapter of my thesis (Aim 2), I delved deeper into the characterisation of ectopic NSCs induced via the overexpression of Dpn. I showed that, within the Drosophila medulla, ectopic NSCs did not obey wildtype temporal cues; they only expressed mid-temporal factors of the temporal series (a series of transcription factors expressed in NSCs). Consequently, this resulted in a bias towards the generation of progeny produced by ectopic NSCs and their inability to terminate on time. We found that these phenomena occurred due to the high affinity of Dpn to bind mid-temporal factors. Within wildtype NSCs, feed-forward and feed-back loops exist between temporal factors. Therefore, we overexpressed a later temporal factor within ectopic NSCs; this was sufficient to correct the temporal series progression, progeny production, and termination. Collectively, this work provided insight into what regulates the faithfulness of dedifferentiated NSCs. Next, I investigated the reprogramming-directed dedifferentiation process of regeneration and subsequent cellular events in the zebrafish retina. Even though humans share the same regenerative genes as zebrafish, they cannot readily regenerate their CNS. Therefore, understanding how zebrafish utilise this regenerative toolkit can be exploited to activate comparable processes in humans. In Chapter 4 (Aim 3) I studied the regeneration of the inhibitory neurons following injury in the zebrafish retina. I found that there were two populations of cells expressing proliferation markers following injury, in which only one were the progenitors that differentiate to replace the damage. This response has never been documented and will be integral knowledge when strategizing regenerative therapeutics following inhibitory neuron damage. To uncover novel regulators of retinal regeneration, I studied the vertebrate orthologs of the Drosophila dedifferentiation factors: Hairy and enhancer of split-6 (Her6, Drosophila Dpn) and Prospero homeobox-1a (Prox1a, Drosophila Pros). Within Chapter 5 (Aim 4), I characterised the expression patterns and functions of Her6 and Prox1a within zebrafish photoreceptor (PR) regeneration. I found that Her6 was a negative regulator of regeneration, and its loss increased PR production; suggesting it may be an encouraging target to increase during mammalian regeneration. Additionally, Prox1a played a role in the differentiation of PRs and this was likely mediated through Liquid-Liquid Phase Separation (LLPS). Together, these findings assist in piecing together the genetic networks involved in retinal regeneration and can be used to design genetic therapies to increase human CNS regeneration. In summary, this thesis provides insight into cell fate changes and dedifferentiation occurring in spontaneous events such as tumourigenesis and the controlled reprogramming during regeneration.
-
ItemA cell-based functional assay that accurately links genotype to phenotype in Familial HLH( 2023-04)Cytotoxic lymphocytes protect humans against viral pathogens and cancer by killing infected and transformed cells, through perforin-mediated mechanism. Mutations in perforin (PRF1) itself or in the secretory machinery responsible for its release (UNC13D, STX11, and STXBP2) are catastrophic, and lead to fatal immune dysregulation, an autosomal recessive disease called familial haemophagocytic lymphohistiocytosis (FHL). Traditionally, FHL has been associated with infant patients. However, it is now apparent that many patients remain disease-free for years, and then present with highly variable and often unexpected symptoms. They remain undiagnosed for a long time and, instead of receiving curative stem cell transplantation, they are treated symptomatically leading to high risk of severe neurological impairment, organ failure and/or death. While the pathogenicity of frame-shift/nonsense mutations is rarely in doubt, the effect of missense mutations on protein function can vary enormously. Yet, over the last two decades, the pathogenicity of missense mutations was almost invariably assumed, and invasive stem cell transplantation was considered without confirmed pathogenicity of mutations. Sadly, transplantation without genetically proven FHL results in a 20% increased mortality compared to patients with proven FHL. Therefore, early and accurate diagnosis of the disease is essential to determine the most appropriate treatment option. Due to the diversity of genetic causes of FHL, there was no test available to directly assess the effect of mutations on cytotoxic lymphocyte function, leading to delayed/erroneous diagnoses. To overcome this diagnostic problem, we have developed a simple, rapid, and robust method that takes advantage of the functional equivalence of the human and mouse orthologues of PRF1, UNC13D, STX11 and STXBP2 proteins. By knocking out endogenous mouse genes in CD8+ T cells and simultaneously expressing their mutated human orthologues, we can accurately assess the effect of mutations on cell function. The wide dynamic range of this novel system allowed us to understand the basis of otherwise cryptic cases of FHL/HLH and, in some instances, to demonstrate that previously reported mutations are unlikely to cause FHL. In addition to diagnosing patients, this unique approach will be paramount for assessing the prognosis of asymptomatic siblings and to guide genetic counselling advice for prospective parents.
-
ItemA Comprehensive Machine Learning System for Cancer Classification Experiments( 2023-01)Cancer detection, classification/subtyping, grading, segmentation and prognostication are promising applications of machine learning to oncology which, if successful, would yield substantial clinical benefits. Merely knowing a cancer’s subtype provides a clinician with deep insights into its nature and likely progress, effective treatments regimes, and the patient’s prognosis. The volume of studies which make use of machine learning in medicine, including in oncology, is large and growing. A variety of clinical aspirations are evident from the literature, including: improving patient outcomes by increasing cancer detection and classification accuracy; reducing cancer diagnosis costs, and reducing the time required to diagnose cancer. However, researchers wishing to incorporate machine learning into their research face a high barrier, since it requires specialised data science/machine learning skills over and above biomedical expertise. Further, few software tools are available to assist a researcher wishing to undertake such studies. Researchers typically develop necessary software themselves: a difficult and time consuming prerequisite activity to conducting experiments. In response to these shortcomings, I developed ‘CLASSI’: an experiment pipeline for cancer classification/subtyping using whole slide images or RNA-Seq gene expression data. CLASSI makes it straightforward for researchers to incorporate machine learning into their research for one important class of oncology experiments, viz.: cancer classification and subtyping using oncopathology images or RNA-Seq data. CLASSI advances the field by providing an ‘off-the-shelf’ experiment platform to simplify and automate the conduct of histopathology image and RNA-Seq data machine learning experiments, and demonstrates that ‘machine learning enabled’ experiment pipelines are feasible, supporting the case for the development of other, more broadly scoped, experiment pipelines.
-
ItemSerglycin glycosylation role in the effector mechanism of cytotoxic lymphocytes( 2022-04)The function of cytotoxic T lymphocytes and Natural Killer cells is important for the clearance of intracellular pathogens and malignantly transformed cells. Recognition of infected/transformed cells and their elimination depends strongly on the capacity of cytotoxic lymphocytes to release cytotoxic effector proteins Perforin and granzyme B. This process is known as the Granular death pathway, which has an essential role in Cytotoxic lymphocyte function and immune homeostasis since its deficiency is known to lead to pathologies related to immune misbalance, which are usually deadly to the host. Along with the cytotoxic effector proteins, a heavily glycosylated proteoglycan is found in the secreted granule content of Cytotoxic lymphocytes, which upon analysis, led to the discovery of a small protein serglycin, named after the serine-glycine repetitions found in the centre of it protein sequence. The presence of serglycin in cells has been associated with the storage and retention of multiple proteases and hormones in cells of endothelial and hematopoietic origin, suggesting its importance in secretory functions across various tissues and cells. In the case of cytotoxic T lymphocytes and NK cells, the presence of serglycin is central for the retention of cytotoxic effector proteins and the formation of dense-core secretory granules. However, how the interaction mechanism with Prf and GzmB, as the mechanism of cytotoxic secretory vesicle biogenesis remains unknown. This project will focus on the study of the serglycin gene, the produced protein, and its modification in the hopes to better understand the motifs and interaction allows the retention and function of cytotoxic effector proteins in cytotoxic T cells.
-
ItemUncovering the roles of Nrf2 in liver development and liver cancer( 2023-02)Nuclear factor erythroid 2-related factor 2 (NRF2) is a transcription factor that is recognized for its role in maintaining redox homeostasis and cellular metabolism. NRF2 is negatively regulated by the E3 ubiquitin ligase adapter Kelch-like ECH-associated protein 1 (KEAP1). The beneficial role of an inducible NRF2 response, under the regulation of KEAP1, has been observed in diseases involving chronic inflammation, neurodegeneration, aging, and cancer. Activating mutations that lead to constitutive NRF2 pathway activation have been observed clinically. Specifically, inborn de novo mutations in NRF2 result in developmental delays and multisystem disorders. Moreover, activating mutations in the NRF2-KEAP1 pathway are observed in several tumour types, including liver cancer. In this thesis, utilizing in vitro and in vivo models, we have investigated the role of NRF2 pathway activation in development and in liver cancer. To investigate the role of NRF2 during development, a zebrafish larval model with mosaic loss of Keap1 was generated using CRISPR/Cas9 technology. Using this model, we demonstrated that Keap1 deficiency induces Nrf2 activation and liver abnormalities that precede post-developmental lethality. Furthermore, we have shown that loss of Keap1 induces aberrant lysosome biogenesis. Employing a mammalian liver cancer cell line, we have shown that KEAP1-dependent regulation of lysosome biogenesis is evolutionarily conserved, cell-autonomous and dependent on the activity of transcription factors TFEB/TFE3. Liver cancer is one of the deadliest cancers with a 5-year survival rate of approximately 18%. One of the most frequently mutated pathways in liver cancer is the NRF2-KEAP1 pathway. To investigate the effect of NRF2 activation in liver cancer, we generated a transgenic zebrafish model with hepatocyte-specific, inducible expression of the constitutively active NRF2-T80K mutant. We have demonstrated that NRF2-T80K expression in hepatocytes drives hepatocyte to cholangiocyte transdifferentiation, which leads to development of a polycystic liver phenotype. Consistent with our findings in the zebrafish model of Keap1-deficiency, the constitutive activation of NRF2 in hepatocytes was associated with dysregulation of lysosome biogenesis in the liver. Finally, using a mammalian liver cancer cell line, we have shown that NRF2-driven transdifferentiation is evolutionarily-conserved and cell autonomous. Having demonstrated that a cell-intrinsic factor, namely NRF2 pathway activation, can influence hepatocyte cell fate, we next examined whether cell-extrinsic factors contribute to the regulation of cell state. We have shown that culture in a physiologically relevant media, which better reflects in vivo nutrient availability, promotes dedifferentiation of mammalian liver cancer cells. Dedifferentiation is associated with loss of HNF4A, the master transcriptional regulator of hepatic gene expression. Mechanistically, we have shown that trace elements are required to maintain the fidelity of hepatocyte cell state in vitro. Collectively, using complementary models of NRF2 activation, we have demonstrated two novel processes driven by constitutive NRF2 activation, namely lysosome biogenesis and hepatocyte to cholangiocyte transdifferentiation. Furthermore, we have shown that trace element availability can influence cell fate in in vitro cell culture systems. Overall, this thesis has identified important regulators of lysosome biogenesis and cell fate in the liver.