Integrative genomic analyses of bacterially-associated colorectal cancer

Sporadic colorectal cancer (CRC) has been linked to various lifestyle factors, including the consumption of alcohol and red meat, smoking, and obesity. CRC is one of most extensively characterised cancers, both at a molecular and 'omic' level; nevertheless, the precise mechanism driving CRC initiation remains unknown. To date, numerous studies have identified changes in the microbial profiles of CRCs compared to adjacent normal mucosa and compared to healthy controls; however, CRC-associated bacteria have not been concurrently quantified across a single cohort; nor have the relationships between CRC-associated bacteria, clinicopathological features of CRC and genomic subtypes of CRC been investigated. The main aim of this thesis was therefore to gain insight into the potential contribution of CRC-associated bacteria in the aetiopathogenesis of CRC by leveraging both host genomic and clinicopathological data as well as to investigate patterns of tissue colonisation between different CRC-associated bacteria. The objectives were 1) to quantify, using quantitative-PCR, CRC-associated bacteria in a cohort of 55 paired tumour and adjacent histologically normal samples collected during surgical resection as well as in an additional 18 formalin-fixed paraffin-embedded (FFPE) samples; 2) to determine their relationships to patient age, gender, ethnicity, stage of disease, site of disease and MSI status (Chapter 4); 3) to evaluate the relationship between each bacterium and host gene expression (Chapter 8) and methylation changes (Chapter 6); and 4) to determine genomic subtypes of CRC using unsupervised clustering of gene expression data in the context of patient clinicopathological features and bacterial quantitation data; and 5) to gain a deeper biological understanding of the results from the objectives 1–4 using pathway analyses of the genomic subtypes obtained (Chapter 7). The main finding of this thesis is that a transcriptomic subtype of colorectal cancer, characterised by an increase in CpG island methylation, displays an increased frequency of colonisation by Enterococcus faecalis and by high levels of Fusobacterium. At the pathway-level, this subtype is enriched for pathways related to damage response, infection, inflammation and cellular proliferation; notably, these findings were confirmed in a well-defined publically available CRC gene expression dataset of colorectal adenocarcinomas (N=155). These findings suggest that specific bacterial colonisation underlie s a distinct genomic subtype of colorectal cancer that is characterise d by inflammatory-related gene expression changes ; these findings however require validation in a larger cohort. In addition, novel associations between colonisation by specific bacteria and host clinicopathological, transcriptomic and DNA methylation features were identified