Abstract
Cancer is a heterogeneous disease that is unique to each patient. Non-muscle invasive bladder cancer (NMIBC) is an early-stage malignancy that can be successfully treated in many cases with the mycobacterium Bacillus Calmette-Guérin (BCG). However, this treatment can have severe side effects, and a significant number of patients have tumour recurrence following treatment. BCG stimulates pathogen-related immune processes in the bladder, resulting in an inflammatory microenvironment. Our research is focused on understanding epigenetic consequences associated with inflammation, particularly via redox reactions in cellular metabolic pathways. Nanopore sequencing offers a straightforward way to examine multiple epigenetic marks in one assay: 5-methylcytosine (5mC), 5-hydroxymethylcytosine (5hmC) and N6-methyladenine (6mA). The first goal of this thesis was to investigate epigenomic consequences of BCG treatment on bladder cancer cells. The second goal was to design a bioinformatic software pipeline utilising the capabilities of nanopore sequencing to enable the prediction of BCG failure in eligible patients.
To achieve the first goal, a co-culture assay was designed with bladder cancer cell lines and BCG-stimulated healthy leukocytes to simulate the inflammatory environment of the bladder during BCG treatment. A transwell system prevented cell-to-cell contact between cancer cells and leukocytes, ensuring biological effects were limited to the inflammatory environment driven by the co-cultured leukocytes. Following incubation, leukocytes and culture media were tested for a sufficient reaction to BCG stimulation, and bladder cancer cells were collected for DNA extraction. Control cancer cells (treated with unstimulated leukocytes) and test cancer cells (treated with BCG-stimulated leukocytes) from seven paired replicates across two bladder cancer cell lines were sequenced using the PromethION 2 Solo. Sequencing analysis revealed genome-wide differentially modified positions, including 5mC, 5hmC, and 6mA, associated with inflammatory and cancer-related genes. Differences in response between the two cell lines were substantial, with no differentially modified regions identified in common across both cell lines.
Numerous types of biomarkers for BCG response in NMIBC treatment have been investigated; however, none have yet achieved sufficient predictive power to be widely adopted in clinical practice. The second goal of this thesis was to develop a proof-of-concept bioinformatic pipeline that utilises nanopore sequencing to combine multiple biomarkers of different sources, specifically, mutations, methylation and tumour immune infiltrate. This enables multiple biomarkers to be measured in one assay. This pipeline was developed to be a clinical bioinformatic workflow that requires little bioinformatic expertise to use. The results of such a tool, with the appropriate pre-clinical trial, could aid in predicting BCG failure, a critical metric for assisting clinicians and patients to choose the most appropriate course of treatment. BCG treatment failure in NMIBC is a complex problem characterised by interactions between mycobacteria, immune response, and tumour biology. Understanding epigenetic contributions during treatment responses can provide insights into which molecular pathways to focus on when exploring ways to complement BCG treatment to increase its efficacy. The ability to identify patients who are unlikely to respond to BCG will improve outcomes by redirecting them to a more suitable treatment earlier. This would reduce healthcare costs, and increase BCG availability for those who are more likely to respond. This thesis provides insights into the epigenomic consequences of BCG treatment of bladder cancer and highlights the diversity of responses to BCG treatment. Furthermore, the bioinformatic pipeline developed here could aid in improving treatment selection for patients with NMIBC.