The development of targeted therapies has had a significant impact on cancer survival rates. However, targeting cancers that are driven by loss of tumour suppressor genes remains a major challenge. One promising approach to treat these cancers is the exploitation of synthetic lethal interactions. Synthetic lethality describes an interaction between two genes, where loss of one gene alone does not affect viability but loss of both genes induces cell death. Inhibiting the synthetic lethal partner of a tumour suppressor gene should specifically kill tumour cells, and so these represent potential therapeutic targets. However, very few synthetic lethal interactions have been well-established.

The aim of this project was to systematically screen for synthetic lethal partners of known tumour suppressor genes. To do so, isogenic human induced pluripotent stem cell lines were generated, each carrying a loss-of-function mutation in a single tumour suppressor gene. These cells have a normal genetic background, thus making it simpler to accurately identify interactions. CRISPR/Cas9 technology was applied as it allows for large-scale, unbiased screening of genetic interactions. A genome-wide guide RNA library was prepared and implemented for knockout screening in the isogenic cell line panel. Analysis was performed to identify genes that were specifically essential for cell fitness/survival in the mutant lines. Particular focus was placed on four tumour suppressor genes that encode subunits of the PBAF/BAF complexes. Approximately 20% of human cancers harbour mutations in subunits of these complexes, so identifying dependencies associated with these could have broad therapeutic potential. Candidate synthetic lethal interactions with these genes were investigated using low-throughput assays in the stem cells and in a cancer cell line. The data obtained suggests that screening in stem cells produces highly variable results. Although potential vulnerabilities associated with all of the tumour suppressor genes were identified, further work is required to validate these and to assess the quality of the results.

In addition to genome editing, CRISPR/Cas9 has been adapted as a tool for controlling gene regulation. In collaboration with Dr Louise van der Weyden, I applied this technology to address another challenging area of cancer biology. Metastasis is the main cause of cancer mortality, yet we still have a poor understanding of the genes that control this process. Considering this, an in vivo CRISPR activation screen was performed to identify novel drivers of metastatic colonisation. A mouse melanoma cell line was transduced in vitro with a library designed to up-regulate expression of membrane proteins, which represent ideal drug targets. These cells were then used in an in vivo experimental metastasis assay. Enrichment of guide RNAs in the lungs was assessed to identify genes that increased pulmonary metastatic colonisation when activated. Candidate genes were selected using three analysis strategies, and hits from each were tested. Several genes were successfully validated using the experimental metastasis assay. The most robust hit was studied further to explore its potential as a therapeutic target.

Collectively, the work described in this thesis demonstrates how CRISPR/Cas9 screening can be applied in different model systems to study genes that drive cancer and to explore novel therapeutic strategies.