Multi-omic characterisation of Barrett’s oesophagus reveals a molecular continuum in the progression to oesophageal adenocarcinoma
Barrett’s oesophagus (BE) is the main risk factor for the development of oesophageal adenocarcinoma (OAC) yet only 0.4%/year of non-dysplastic (ND) BE cases progress to cancer (Bhat et al., 2011). In this thesis I present the first comprehensive, multi-omics study comparing indolent, non-progressors with those who progressed to a range of dysplasia grades. BE is highly heterogeneous with regards to mutational load, copy-number aberrations (CNAs) and structural variants (SVs). Mutational signatures are laid down early and persist regardless of progression status. Hence, Cosmic signature 17 (T:A>G:C in a CTT context), the hallmark of OAC, is visible in indolent, ND samples. TP53 mutation, GATA6 amplification, ERBB2 amplification, APC mutation and whole genome doubling are confined to cases that have progressed with TP53 being by far the most prevalent. In contrast CDKN2A alteration occurs early in around 50% of indolent cases. SV analysis reveals a dominance of translocations from the early ND grade. Spatial analysis of multiple samples from across BE segments shows that the total mutation burden, total CNAs and total numbers of SVs to be surprisingly constant. However, clonality analysis highlights the complexity of BE, with some cases arising from a clear ancestral clone while others having minimal sharing of variants and showing heterogeneity in terms of driver events. Transcriptomic analysis reveals a clear but gradual differential gene expression between ND and dysplastic biopsies. We demonstrate a loss of the intestinal metaplasia phenotype with progression and downregulation of the HNF4A pathway, a transcription factor with roles in intestinal development. In progression there is an upregulation of genes in the ERK/MAPK pathway. In summary, BE is a heterogeneous disease that shows a continuum of abnormalities in the progression towards cancer. We hypothesise that it is the accumulation of events that tip the balance to progression, rather than a stepwise model through the phenotypic dysplasia grades. Selected features could help to differentiate indolent from high risk disease and further work is being performed to identify scoring systems and biomarkers to apply in the clinical setting.