Intestinal_Cancer_Initiation_Using_ENU_Mutagenesis

Different models of colorectal tumour evolution have been conjectured. The prevailing model is the “Stepwise” or “Vogelstein” model that suggests a sequential accumulation of mutations driving big clonal sweeps throughout an adenoma to carcinoma continuum (Fearon et al. Cell, 1990). More recently, Sottoriva and colleagues have partially challenged this view and suggested that most mutations are acquired at an initial stage with major sub-clones developing early and maintained over-time - the “Big Bang” model (Sottoriva et al., Nat.Genetics, 2015). Some evidence from the Big Bang model work and other publications propose that some driver mutations could already be present in the tumour initiating cell(s) (Tomasetti et al., PNAS, 2013; Williams et al., Nat. Genetics, 2016; Sievers et al., Gut, 2016). In support of this, our laboratory has shown widespread CRC driver gene mutations in normal human colonic epithelium (Nicholson et al., in revision). All these observations led us to hypothesize that driver mutations present in the cell(s) of origin could define tumour subtype, progression potential and subsequent sub-clonal architecture. We are testing this hypothesis with mouse models by driving oncogene activation (Kras, Pik3ca) and/or tumour suppressor inactivation (Trp53, Pten, Fbxw7) prior to ENU mutagenesis. Preliminary data shows that normal (non-tumour) mouse intestinal cells expressing driver gene mutations are more prone to initiate tumours upon ENU-driven alkylation. We have also observed that Trp53 deletion is not as efficient as KrasG12D expression or Fbxw7 deletion at initiating tumours but originates a large variety of phenotypes with some sub-types displaying metastatic features. This supports some of the Big Bang model premises with regards to early acquisition but goes even further at suggesting that if at least one of these key CRC cancer driver mutations is already present in the tumour cell(s) of origin it could generate additional genomic instability and/or accelerate crypt monoclonal conversion of tumour initiating mutations. In order to fully characterize these mouse models and the underlying mechanisms of tumour initiation and progression we aim to deep sequence each individual tumour using a defined cancer gene panel (~600 genes) with a 200-300x coverage, enabling clonal and sub-clonal mutation calling.