Abstract
The CDH1 gene, which encodes the epithelial cell-cell adhesion protein E-cadherin, is frequently mutated in diffuse gastric cancer and lobular breast cancer. In addition, germline CDH1 mutations are responsible for the familial stomach cancer syndrome, hereditary diffuse gastric cancer. As these cancers are characterised by the absence of functional E-cadherin, this protein cannot be directly targeted. Synthetic lethality, the relationship between two proteins where loss of either protein alone is tolerable to a cell, but simultaneous loss of expression reduces viability, provides a new treatment option for cancers caused by loss of protein function. Using a cell line cultured from normal breast epithelium (MCF10A) and its isogenic CDH1 knockout we have identified candidate synthetic lethal genes and drugs. These have the potential to be developed for hereditary diffuse gastric cancer chemoprevention and the treatment of advanced E-cadherin negative cancers.
In this thesis I describe a whole genome siRNA screen and a high throughput known drug screen to identify genes and drugs which cause a greater reduction in viability in CDH1-/- cells, compared to the CDH1 wild-type. Three synthetic lethal candidates - JAK2, ROS1, and ADCY7 - were validated with shRNA knockdown and/or drug inhibition. JAK2 and ADCY7 are second messengers to cell surface receptors and activate various signalling pathways, including the PI3K/AKT pathway. ROS1 is an orphan tyrosine kinase receptor and can also activate the PI3K/AKT pathway. I demonstrate a robust synthetic lethal relationship between these genes and E-cadherin. Further, I have used candidates from these screens to identify a key role for the PI3K/AKT cell survival signalling pathway and G-protein coupled receptors in maintaining CDH1-/- cell viability. Conversely, phosphatases are enriched in candidates causing a greater reduction in MCF10A wild type cell viability. I have used these results to propose a mechanism for E-cadherin synthetic lethality to better identify candidates for future drug target development. In summary, this thesis demonstrates that E-cadherin loss creates vulnerabilities with the potential to be targeted for E-cadherin negative cancer therapy.