Clinical Pathology - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 9 of 9
  • Item
    Thumbnail Image
    Improving the diagnosis and treatment of cancer of unknown primary using genomic profiling and patient-derived organoids
    Posner, Atara ( 2023-04)
    Cancer of unknown primary (CUP) is a metastatic cancer for which a standard clinical investigation fails to identify a primary tissue of origin (TOO). Clinical decision-making is impaired without a primary TOO diagnosis since this is typically used to inform optimal patient care. Most CUP patients are treated with empirical chemotherapy leading to poor survival outcomes. Therefore, improved diagnostic and treatment options for CUP patients are needed. The diagnostic utility of targeted RNA and DNA profiling was investigated in a cohort of 215 CUP patients. A retrospective clinicopathology review was performed to assign a putative TOO diagnosis. Seventy-seven percent (166/215) of the cohort had insufficient clinicopathological evidence to support a TOO diagnosis, while the remaining 23% were putatively resolved and had a clinicopathology diagnosis assigned. A custom gene expression profiling (GEP) TOO classifier was applied. The clinicopathology-unresolved CUPs were more challenging to classify than the resolved CUPs, and only 56% had a high confidence TOO classification. Cancer-type specific DNA features, including mutations and mutational signatures, were investigated and provided diagnostic evidence to support a TOO in 31% of unresolved CUPs, while GEP classifications supported a diagnosis in 13%. A genomics-informed clinicopathology review could direct an unconfirmed TOO diagnosis in one-third of CUPs. The RNA and DNA profiled CUP cohort were also investigated for immune and genomic biomarkers that could direct immune checkpoint inhibitor (ICIs) treatment for CUPs. Sixteen percent of CUPs had a high tumour mutation burden, 33% had high expression of genes associated with immunotherapy response, and 47% of the TOO-resolved CUPs belonged to an ICI-responsive cancer type. Of a subset of 28 CUPs treated with ICIs, 8/28 (29%) responded to treatment, and all of the responders had a high gene expression score (7/8) and/or high tumour mutation burden (3/8), and most (5/8) belonged to ICI-responsive cancer types. Therefore, many CUP patients have molecular features indicative of ICI susceptibility. Whole genome sequencing (WGS) was explored in 30 CUPs with targeted DNA sequencing to identify additional diagnostic and therapeutic informative features. An 82% concordance of variants was detected between WGS and panel sequencing. WGS could detect additional genomic features, including mutational signatures. Actionable mutations and biomarkers were detected in 48% of cases by targeted DNA sequencing, while WGS increased this to 59%. Furthermore, 72% of CUPs could have a TOO assigned following genomics-informed iii histopathology assessment, of which a third resulted from WGS-identified features only, and all diagnostic features detected by targeted sequencing were supported by WGS. Although only assessed in a small subcohort, tumour DNA extracted from archival tissue and cell-free material could be sequenced to detect clinically relevant findings. Patient-derived organoids (PDOs) were attempted from fresh CUP tumours. PDOs were successfully derived from six of 20 tumour samples (30%). The PDOs were heterogenous, genomically and morphologically representative of the tumour and could be used for WGS. Anticancer compounds were selected for PDO drug screening based on molecular and clinically directed treatment. The PDO drug responses frequently concurred with genomic biomarkers, but multiple drug sensitivities were also unanticipated. By incorporating targeted DNA sequencing and GEP testing, many CUP patients had a putative TOO diagnosis assigned. The genomic tests were also used to identify molecular biomarkers for ICI treatment. When compared with WGS, there were specific diagnostic scenarios whereby WGS could improve the diagnosis and identify additional potentially useful therapies for patients. Additionally, CUP PDOs were used to test genomic-directed therapies through in vitro drug testing. Overall, the study provides insights into how genomic tests can improve the diagnosis and treatment of CUP.
  • Item
    Thumbnail Image
    A Proteomic Exploration of Chemotherapeutic Resistance Mechanisms in Metastatic Colorectal Cancer
    Cooper, Benjamin Christopher Paul ( 2022)
    There is a substantial global burden of colorectal cancer with many patients developing or presenting with distant metastases, most commonly within the liver. Though curative surgery can be attempted, modern clinical management of metastatic colorectal cancer is heavily reliant on systemic chemotherapy. The most administered chemotherapeutic regimen in Australia is FOLFOX, whose constituent drugs are the pyrimidine antimetabolite 5-Fluorouracil and the platinum based alkylating agent Oxaliplatin. Despite being highly cytotoxic, their clinical efficacy is limited by mechanisms of chemotherapeutic resistance within cancer cells, which contributes to poor patient outcome. Protein level research of these mechanisms has traditionally been limited in scope and often restricted to two-dimensional cell experimental models. Hence, there is an unmet need for proteomic analyses of chemotherapeutic response in models with greater physiological relevance, like three-dimensionally cultured organoids. In this thesis, we optimised a mass spectrometry sample preparation method to permit the proteomic analysis of metastatic colorectal cancer patient derived tumour organoids and their supernatants. We then analysed the proteomes of organoids with differential responses to 5-Fluorouracil and Oxaliplatin treatment using mass spectrometry, finding proteins associated with a mitotic cell cycle theme to be of greater abundance in relatively 5-FU/Oxal resistant samples compared to sensitive ones. Using western blotting,CDK1 expression was shown to be significantly increased in relatively resistant organoid samples during 5-FU/Oxal exposure compared to sensitive ones, though its inhibition did not re-sensitise them to chemotherapy. A proteomic analysis of organoid supernatants also revealed significantly greater levels of GSTP1 within the supernatants of sensitive organoids, when compared to relatively resistant ones. These discoveries help describe mechanisms by which metastatic colorectal cancer cells can evade the cytotoxic effects of chemotherapeutic drugs.
  • Item
    Thumbnail Image
    Improving the Differentiation of Inherited from Sporadic Causes of DNA Mismatch Repair Deficient Cancers
    Walker, Romy ( 2022)
    People with Lynch syndrome (LS), the most common cancer predisposition syndrome, have an increased risk of developing colorectal (CRC), endometrial (EC) and sebaceous skin (SST) tumours, among other cancer types. LS is caused by germline (likely) pathogenic variants in the DNA mismatch repair (MMR) genes, leading to a characteristic MMR-deficient (dMMR) / microsatellite unstable (MSI) tumour phenotype. However, dMMR / MSI is not only caused by LS and can also result from somatic causes of MMR gene inactivation. Identification of the dMMR / MSI tumour phenotype is of clinical significance for cancer prevention, prognosis, and response to immune checkpoint inhibitor treatment. However, diagnostic challenges still remain for accurate dMMR identification using current clinical tools. Therefore, the aim of my thesis is to address these two clinically relevant challenges: 1) can we improve detection of dMMR / MSI status in CRCs, ECs, and SSTs using next generation sequencing (NGS) derived tumour features and bioinformatic tools, and 2) can we improve differentiation of inherited (high risk) from sporadic (low risk) dMMR in CRCs, ECs, and SSTs using NGS? With the increased adoption of NGS in clinical practice for precision oncology and cancer genetics, the opportunity exists to develop a tumour-focused approach to improve the identification of LS. In Chapter 3, I investigated whether routinely collected clinicopathological tumour features could differentiate LS (inherited) from MLH1 methylated (sporadic) and suspected LS (unknown) subtypes in over 631 dMMR CRCs from the Colon Cancer Family Registry. Although age at diagnosis, sex and other variables were different between these subtypes, their power to differentiate was limited. Significantly, MLH1me tumours presented with a higher mortality rate than LS and suspected LS (SLS) tumours. These findings confirmed the need to examine new NGS-based approaches for differentiation purposes. In Chapter 4, I developed a novel model using NGS data to accurately determine dMMR from MMR-proficient (pMMR) tumour status in CRC and tested this model on EC and SST tumours. In total, 104 tumour features derived from whole exome sequencing of 300 CRCs were tested for their ability to identify dMMR with 10 features displaying >80% prediction accuracies. I established a novel model which combines the best predictive features for improved dMMR detection in CRC, EC, and SST sequenced tumours. In Chapter 5, using a custom-designed panel sequencing assay that I developed, tumour and blood-derived DNA were assessed from 134 SLS dMMR tumours (79 CRCs, 32 ECs, and 23 SSTs) in one of the largest investigations of SLS in these tumour types to date. Using the model determined in Chapter 4, I found 8.2% were not dMMR as clinically reported. The predominant cause of dMMR in SLS was biallelic somatic MMR mutations (61.9%) while only 1.5% were identified as LS that were missed by clinical testing. The findings from my thesis provide an evidence base to implement NGS-based diagnostic approaches for accurately detecting dMMR status and to resolve an SLS diagnosis in CRCs, ECs, and SSTs using a single test, which will ultimately improve utilisation of limited clinical resources for improved clinical management and cancer prevention.
  • Item
    Thumbnail Image
    Evaluating the clinical applicability of tumour mutational signatures in colorectal cancer and related syndromes
    Georgeson, Peter ( 2022)
    Colorectal cancer (CRC) poses a major health burden. It is the second most common cause of cancer death, with the impact of CRC incidence and mortality continuing to grow worldwide. Early detection of CRC substantially improves outcomes, motivating the adoption of screening programs aimed at identifying high-risk individuals for ongoing surveillance. However, efforts to identify individuals predisposed to developing CRC have been hampered by the complexity and heterogeneity of CRC. Recent advances in DNA sequencing technology enable the genome to be studied in high resolution, providing the ability to detect a wide array of somatic mutations and rearrangements in the DNA of cancer cells. Certain mutagenic processes leave identifiable mutational patterns in cancer genomes. The advent of cost-effective large-scale DNA sequencing enables systematic detection of these patterns, known collectively as tumour mutational signatures. To date the main application of mutational signatures has been research focused, where they have been used to determine cancer subtypes and categorise the underlying changes to DNA associated with those subtypes. However, their applicability to clinical contexts have not been sufficiently explored. A limitation to the adoption of mutational signatures clinically is the prevalence of FFPE-preserved tissue in conjunction with whole-exome and panel-sequenced data, in contrast to the use of fresh-frozen whole-genome sequenced data typical in research settings. Formalin is mutagenic, which can result in artefactual variants, while at least an order of magnitude fewer mutations are detected with whole-exome and panel-sequenced data compared to whole-genome. We assess the utility of mutational signatures generated from both whole-exome and panel-sequenced data derived from FFPE-preserved tissue. Specifically, we show that inherited predispositions to CRC, including Lynch syndrome and MUTYH-associated polyposis, can be accurately identified with whole-exome sequenced data from FFPE-preserved tumour tissue, and that, with the correct methodology, biallelic MUTYH carriers can be identified from panel-sequenced FFPE-preserved tumour tissue. Understanding the relationship between environmental exposures and CRC development has implications for both prevention and screening. We consider the ability of mutational signatures to detect mutation patterns arising from exposure to colibactin, the genotoxic compound synthesised by pathogenic E. coli and a potential cause of sporadic (non-inherited) CRC. Demonstrating distinct genomic, clinic-pathological and epidemiological characteristics, we show the potential existence of a distinct subtype of CRC based on the presence of the colibactin-associated mutational signature. The effectiveness of mutational signatures depends on the environment in which they are calculated. We analyse the impact of key analytical parameters and recommend specific filtering settings for variant allele fraction and sequencing depth. We identify situations where mutational signatures are less effective, recommending minimum mutation counts and maximum signature reconstruction error, enabling confidence in mutational signature results to be based on their specific application. The results presented in this thesis have clinical applications. We show that applying mutational signatures to individual tumours provides direct evidence suggesting a particular aetiology. More broadly, mutational signatures provide evidence indicating the likely pathogenicity of co-occurring mutations. Mutational signatures are an important technique for extracting information from sequencing data. This thesis demonstrates clinical applications of mutational signatures in CRC and related syndromes.
  • Item
    Thumbnail Image
    Statistical and Functional Genomics of Host-Microbial Interactions
    Fachrul, Muhamad ( 2022)
    The continuous refinement of sequencing technologies has allowed bioinformatics to evolve as a field, as new data types enable the analysis of multiple levels of biological dimensions. Yet, an issue shared between these “omic” types persists in the form of variations introduced by either technical or biological confounders that are ultimately unwanted, as they may conflate true biological signals from the condition of interest and result in misleading conclusions. Technical confounders are introduced from inconsistencies during experimental design, while biological confounders are inherent and a product of the genetic structure of individuals. This thesis aims to capture the complexities of host-microbial dynamics in a manner that accurately captures biological variance of interest, describing separate approaches to minimizing the impact of unwanted variations from technical and biological confounders in metagenomic and transcriptomic data, respectively. This thesis addresses confounders in two different aspects of the host-microbial dynamics: the presence of microbial life in the host, as well as the host molecular response towards microbial life in the body. Chapter 2 touches on the first aspect as it delves into the method of removing variations introduced by technical batches in metagenomic data. Using a method originally designed for single-cell RNAseq called RUV-III-NB, unwanted technical variations were assessed and removed from microbiome profiles of pig faecal samples that underwent various storage options and sample treatments. The study identified storage conditions and freeze-thaw cycle among the highest contributors to unwanted variations in microbiome abundance, particularly affecting high-abundance bacterial taxa. RUV-III-NB’s consistently robust corrective performance was also shown when benchmarked with other popular batch correction methods. This chapter describes the importance of preventative measures during experimental design and at the same time offers a robust corrective measure in-silico. In Chapter 3, biological confounding in the form of population stratification is addressed in transcriptomic data analysis using a dataset from a Salmonella Typhi infection study from Nepal. A bioinformatics pipeline to capture genetic structure named RGStraP (RNAseq-based Genetic Stratification PCs) was developed for this study to capture genetic structure solely based on RNAseq samples. This chapter demonstrates RGStraP’s capability as a robust alternative for capturing genetic structure based on its performance when compared to paired array genotypes, as shown by SNP-level genetic concordance and canonical correlations between two sets of genetic principal components. The effect of population stratification on gene expression data is also shown, as the lack of adjustment based on genetic structure in downstream RNAseq analysis may result in possible exaggeration of significant associations. Using the same RNAseq dataset from Nepal, Chapter 4 profiles the host gene expression signature towards S. Typhi infection. Both technical and biological confounders were addressed in downstream RNAseq analyses using results from RUVg and RGStraP, respectively. This chapter presents a distinct gene expression signature between confirmed cases and healthy controls, from which subclades of samples could be determined using unsupervised hierarchical clustering. A typhoid disease classifier was constructed from the S. Typhi-specific gene signature and was tested on external validation sets, showing promising potential in diagnosing S. Typhi infection from patient’s gene expression. Finally, Chapter 5 discusses the overarching takeaway from the studies, current limitations, as well as future directions for the studies involved.
  • Item
    Thumbnail Image
    Clinical outcome prediction using biomedical data and machine learning approaches
    Liu, Yang ( 2022)
    Identifying asymptomatic individuals with increased susceptibility to disease provides substantial opportunities for preventative interventions. Over the past few years, the advances in sequencing and computing technologies have enabled omics-driven disease prediction modelling which may aid in the exploration of new biomarkers and future clinical utility. Recent studies have revealed evidence linking human gut microbiota with the pathogenesis of various complex diseases. However, previous studies have been limited by cross-sectional study design and there are limited data regarding the longitudinal association between baseline gut microbiome and incident diseases. In addition, there are few published studies on incident disease prediction combing genetic risk and gut microbial risk factors. To address this, we designed a longitudinal study to examine the predictive utility of clinical metadata, the gut metagenomics and genomics data for a series of complex diseases, using statistical and machine learning approaches in a large population-based cohort with ~15 years of electronic health records follow-up. Chapter 1 provides a comprehensive review on advances and challenges in complex disease prediction. Emerging prediction methods and novel biomarkers are highlighted, including the polygenic risk scores, gut metagenomics, and machine learning approaches in the context of disease prediction. Recent progress in clinical utility of the advancements in multi-omics-based prediction, and future challenges and potential opportunities for clinical translation are discussed. In Chapter 2, the potential of gut microbiota for prospective risk prediction of liver disease was investigated using machine learning approaches. The predictive capacity of the baseline gut microbiota was evaluated individually and in combination with conventional risk factors. The results demonstrated that the microbiome augmentation of conventional risk factors using gradient boosting classifiers significantly improved prediction performance. Investigation of predictive microbial signatures revealed previously unknown bacterial taxa for incident liver disease, as well as those previously associated with hepatic function and disease. In Chapter 3, the associations with baseline gut microbiome were tested for incident respiratory diseases, including COPD and adult-onset asthma. The gut microbial alterations and variations at each taxonomic level were compared between disease cases and non-cases. Machine learning models demonstrated moderate predictive capacities of baseline gut microbiome for incident asthma/COPD. Subgroup analyses indicated gut microbiome was significantly associated with incident COPD in both current smokers and non-smokers, as well as in individuals who reported never smoking. In Chapter 4, the predictive utility of genetic risk factors, gut microbial risk factors, and lifestyle risk factors was investigated for multiple complex diseases, including myocardial infarction, coronary heart disease, prostate cancer, Type 2 diabetes and Alzheimer’s disease. Since the gut microbiome is involved in numerous host physiological processes and linked to all vital organs, it was hypothesized that the gut microbiome can reflect host environmental risk factors for relevant diseases. It was also hypothesized that the inclusion of genetic susceptibility could improve the prediction performance over clinical risk factors for complex diseases. The findings demonstrated the individual and combined impact of polygenic predisposition and variations in baseline gut microbiota on disease incidence. This thesis presents a comprehensive investigation of the integrative use of clinical metadata and multi-omics data, the human gut metagenomics in particular, for incident disease prediction. The findings of this work provide an evidence base for the translation of omics and machine learning to risk prediction of multiple diseases, and support further investigation into identification of new biomarkers for disease risk assessment and prevention.
  • Item
    Thumbnail Image
    Heritable DNA methylation marks associated with familial breast and prostate cancer
    Hosseinpour, Mahnaz ( 2022)
    The known genetic risk factors for breast and prostate cancer account for less than 50% and 35% of the familial risk respectively. Early candidate gene studies provided evidence that DNA methylation measured in peripheral blood can mimic the effect of cancer predisposition associated with germline pathogenic variants. Key examples included the increased risk of breast cancer associated with both BRCA1 promoter hypermethylation and intragenic DNA methylation in ATM. More recently, the association between heritable DNA methylation marks and breast and/or prostate cancer risk has been reported by genome-wide studies. However, it is not clear how most of these modifications in DNA methylation mediate phenotypes resulting in increased cancer risk. In this thesis, a recently developed analytic approach, based on complex segregation analysis and family structure, was applied to identify 1,000 heritable DNA methylation marks in 57 families with multiple cases of breast and prostate cancer. Using Cox segregation analysis, we found that 229 and 140 methylation marks were associated with breast and prostate cancer respectively and 63 were associated with increased risk of both cancer types. We developed a Clustered regularly interspaced short palindromic repeats (CRISPR)-based strategy using a dCas9-DNA methyltransferase enzyme (DNMT) and a modified sgRNA containing two PP7 hairpins fused with DNMT3A in the synergistic activation mediator (SAM) system to assess whether induction of DNA methylation at these methylation sites resulted in a recognised cancer phenotype (such as DNA damage and proliferation). We first optimised CRISPR mediated DNA methylation by testing three combinations, including fusion of dCas9-DNMT3A, dCas9-DNMT3B and dCas9-DNMT1 with our modified sgRNA and found that SAM-DNMT3A which involves fusion of dCas9-DNMT3A with our modified sgRNA induced highly robust DNA methylation even compared to the previously developed approach, SunTag system, via transfection of HEK293FT cells targeting BRCA1 promoter region. In addition, we performed lentiviral transduction of human mammary breast cell lines using SAM-DNMT3A and sgRNAs targeting BRCA1, PTEN and NF1 regions and demonstrated a significant DNA methylation induction at these regions compared to GFP- sgRNA, targeting non-human genes. We found that the SAM-DNMT3A could also induce higher gene silencing in comparison with CRISPR-inhibition (CRISPRi) tool, involving dCas9-KRAB. A lentiviral delivered sgRNA pooled library, including 10 sgRNAs for each of the 1,000 heritable DNA methylation sites, was generated to perform phenotypic screens in the human mammary epithelial cells, B80-T5 and K5+/K19+ cells using SAM-DNNT3A and CRISPRi tools.We measured the effect of DNA methylation on proliferation and the DNA damage response using a PARP inhibitor synthetic lethal screen. K5+/K19+ cells, showed very low phenotypic changes. Overall, the SAM-DNMT3A tool enabled a systemic high throughput pooled screen of heritable DNA methylation sites mediating breast cancer related phenotypes including proliferation and DNA damage. The findings of this study further characterise the non-genetic component of familial risk of breast and prostate cancer, provides new opportunities to elaborate how tumorigenesis can be affected by DNA methylation and develop epigenetic therapeutics targeting these risk factors, ultimately advancing both precision prevention and medicine.
  • Item
    No Preview Available
    Genetic and Epigenetic Risk Factors for Invasive Lobular Breast Cancer
    Suman, Medha ( 2021)
    Invasive lobular breast cancer (ILBC) is the second most common histological subtype of breast cancer and accounts for 10-15% of all cases. Loss of e-cadherin protein is a hallmark of ILBC and contributes to its characteristic discohesive morphology. In addition to distinct histological features, several subtype-specific molecular and clinical features have been described. However, ILBC remains understudied relative to other breast cancer subtypes, despite its frequency. In this era of precision medicine, there is a growing interest in further refining breast cancer tumour subtyping by identifying additional discriminating molecular features. However, there is limited data to pursue this for ILBC as it is often not well-represented in study samples. For instance, the seminal work on breast cancer classification based on gene-expression levels by Perou et al., (2000) included only two ILBC cases. It is important to identify ways to refine the subtyping of ILBC tumours so that women with ILBC can benefit from a more precise treatment plan, prognosis and targeted therapy options. The main objectives of this PhD project were: i) to examine the distinguishing methylation patterns between ILBC (n=151) and non-ILBC (n=341) tumours ii) to investigate the ILBC methylome to identify methylation signatures for prognostication (n=130) and iii) to subclassify ILBC into subgroups with increased homogeneity based on their genome-wide DNA methylation profiles (ILBC, n=151, non-ILBC=341) and to further characterise these subgroups by investigating their somatic mutational signatures (n=15). Three subgroups of ILBC were defined via unsupervised cluster analysis of genome-wide DNA methylation measured using the Infinium HumanMethylation450K assay. Of these, Subgroup 1 was identified as the most distinct ILBC subgroup, characterised by a predominant hypomethylation across 27,675 CpGs compared with Subgroup 2 and across 13,067 CpGs compared with Subgroup 3. Subgroup 1 showed more similarity to the TNBC (non-ILBC) cases compared with the other two methylation-defined subgroups in terms of their genome-wide methylation pattern. Survival analysis showed that women with ILBC tumours in Subgroup 1 had the poorest overall survival when compared with women in Subgroup 2 (hazard ratio (HR): 0.59, 95% confidence interval (CI): 0.19-1.79) and Subgroup 3 (HR: 0.16, 95% CI: 0.03-0.88), after adjusting for age and year of diagnosis. Subgroup 3 had an enrichment for women who had a first-degree relative with a history of any cancer. Both Subgroup 2 and Subgroup 3 were enriched with women who had a female relative with a history of breast cancer. This suggests that women in Subgroup 2 and Subgroup 3 may be genetically or epigenetically predisposed to developing breast cancer. The somatic genetic variant profiles of the ILBC DNA methylation-defined subgroups were further investigated by performing whole-exome sequencing (WES) on five ILBC tumours representing each of the three subgroups (n=15). The mismatch repair deficiency (MMRd) associated mutational signature SBS6 was the most frequently observed mutational signature in the ILBC tumours, detected in 12/15 (80%) cases. Microsatellite instability (MSI) was also predicted in 13/15 (87%) of the cases, including all 12 tumours with SBS6. Although distinct somatic (genetic) characteristics for tumours of individual subgroups were not observed, this research highlighted the potential role of MMRd in ILBC tumourigenesis and progression. DNA methylation of ILBC was also investigated as a possible prognostic biomarker. The analysis revealed 2,771 variably methylated regions within the ILBC tumours (n=130). A pooled survival analysis of the study set and TCGA data identified APC, TMEM101, HCG4P3 and CELF2 promoter methylation as potential prognostic biomarkers for women with ILBC. Comparing the DNA methylation profiles of ILBC (n=151) and non-ILBC (n=341) tumours, 13,763 genes and 8,456 intergenic regions showing statistically significant differences in DNA methylation (false discovery rate (fdr), P-value < 0.01) were identified. Gene set enrichment analysis revealed that the differentially methylated genes were found to be involved in biological pathways related to metabolism of RNA (R-HSA-8953854), mRNA processing (GO:0006397), RNA splicing (GO:0008380), cell cycle (R-HSA-1640170) and DNA repair (GO:0006281). This study brings together several lines of evidence to indicate that distinct molecular features of ILBC can enable further subtyping, identify important features for targeted therapies (e.g., MMRd) and provide additional information for prognostication. This research identified Subgroup 1 as an important subgroup with similarities to TNBC and more aggressive clinical behaviour. Further investigation of samples from Subgroup 1 may identify additional important targets for precision medicine.
  • Item
    Thumbnail Image
    Characterising tumour and immunological heterogeneity in colorecal cancer
    Shembrey, Carolyn Elizabeth ( 2021)
    Despite improvements in surgical oncology and precision medicine, 5-year survival rates for those with late-stage colorectal cancer (CRC) remain extremely poor and innovative treatment strategies are needed. Although T-cell directed immunotherapies are strikingly effective in many solid cancer types, durable responses are limited to approximately 5% of all CRC patients. Accordingly, there is an urgent need to explore alternate immunotherapeutic strategies which harness other cytotoxic cell types, most notably NK cells. However, the inter-patient variability of NK cell involvement in CRC, and particularly in colorectal liver metastases (CRLMs), is poorly characterised. Moreover, very little is known regarding the influence of tumour heterogeneity on immunotherapy response, mandating the development of novel methodologies which can dissect divergent responses to immuno-, chemo- and targeted therapies at the level of individual tumour cell subpopulations. In this thesis, the scope of tumour and immunological heterogeneity in primary CRC and CRLMs was assessed from transcriptomic, spatial and functional perspectives. A CRC-specific NK cell gene signature which infers the NK cell load of individual tumours from bulk RNAseq data was designed and validated. Differential expression analysis revealed that tumours with high evidence of this NK cell signature were characterised by the upregulation of chemotactic and cytolytic transcriptional programs. Furthermore, amongst patients with primary CRC, those with high NK scores were shown to have better survival outcomes in two independent cohorts. Focussing on metastatic disease, Cell type Identification by Estimating the Relative Subsets of RNA Transcripts (CIBERSORT) analysis revealed significant variance in terms of immune cell composition between CRLMs and adjacent normal (AdjN) liver tissue. Using a novel mIHC panel to quantify the scope of NK cell infiltration and NK cell ligand expression, it was determined that neoadjuvant chemotherapy increased NK cell penetrance of CRLM tissue, restoring NK cell load to levels comparable with the AdjN liver. Furthermore, NK cell densities were comparable in the AdjN liver and CRLM tissue of patients who partially responded to chemotherapy whereas nonresponders showed preferential accumulation in the AdjN liver. Preliminary data is also presented regarding the establishment of a NK cell versus patient-derived organoid (PDO) co-culture killing assay which can be used to assess the functional activity of NK cells against CRLMs. Lastly, a lineage tracing technique termed optical barcoding (OBC) was employed to study the relationship between intra-tumour heterogeneity and treatment response. Stably marking heterogenous populations of CRC cells with unique fluorescent signatures enabled the real-time identification and tracking of different cellular subpopulations under pharmacological selective pressures. Moreover, optimising the fluorescent protein panel allowed the maximum theoretical subpopulation resolution for the OBC technique to be achieved, greatly improving throughput as compared with previously published fluorescent barcoding techniques. Collectively, the results provided herein have deepened our understanding of tumour and immunological heterogeneity in primary CRC and CRLMs, and, by introducing several novel technologies, have laid strong foundations for future studies.