Centrosome amplification (CA), whereby cells have more than the normal number of centrosomes, is a common feature amongst aggressive cancers with a poor prognosis. CA drives oncogenic phenotypes such as increased invasiveness (via both cell-autonomous and non-cell-autonomous mechanisms) and chromosomal instability. In addition, CA can initiate tumourigenesis in flies and mice, and can drive advanced tumourigenic traits early on that promote disease progression. We hypothesise that reversal of CA may reverse CA-driven
oncogenic phenotypes and therefore might be a new way to target aggressive cancers with a poor prognosis.

This thesis presents the use of three patient-matched uveal melanoma (UM) cell lines as a new model to study CA that has developed in a patient setting. Mel270 cells were derived from a primary tumour and have negligible CA. OMM2.3 and OMM2.5 cells were derived from distinct liver metastases from the same patient and have high levels of CA. Primary UM can be well managed, however ~50% patients develop metastatic disease, for which there is no curative treatment and 1 year survival rates are ~50%, so there is a need to develop new therapies to target metastatic UM.

Using RNA-Seq, genes that were differentially expressed in metastasis versus primary derived cells were selected for an siRNA screen to identify genes with a role in CA. Knockdown of either Aurora A or HSP90B1 induced a consistent reduction of CA in OMM2.3 cells. It is believed that Aurora A and HSP090B1 have independent effects on CA, as depleting either Aurora A or HSP90B1 using siRNAs did not affect the RNA or protein levels of the other. A new assay combining live imaging of migrating cells and immunofluorescence was developed and used, which we call FUCCI-CLIF (Fluorescent Ubiquitination based Cell Cycle Indicator - Correlative Live imaging and Immunofluorescence). FUCCI-CLIF provides insight into cell migration, cell cycle and CA status. Knockdown of Aurora A or HSP90B1 reduced migration as calculated by mean straight line speed. Further investigation indicated this was a non-cell-autonomous effect on
cell migration.

Ultimately, the work presented suggests that depleting or inhibiting proteins required for CA in a cancer specific setting reverses CA and concomitantly reduces oncogenic properties of aggressive metastatic UM cells, and is therefore a much needed new potential therapeutic approach against metastatic UM.