Epithelial ovarian cancer (EOC) is the leading cause of death from gynecological malignancies in the Western World. The aim of this study was to identify ovarian cancer subtypes related to biological pathways which are currently being targeted by novel therapeutics. SRC (sarcoma) is a tyrosine kinase which plays critical roles in mediating many cellular pathways such as proliferation, adhesion, survival, differentiation and cell motility. SRC activity is increased in many human cancers including EOC. Furthermore, up-regulation of the SRC pathway has been shown to be associated with cisplatin resistance and patient overall survival in EOC. We therefore hypothesised that up-regulation of the SRC pathway may represent a subtype of EOC which could be therapeutically exploited. We used a semi-supervised analysis approach to perform hierarchical clustering analysis of 198 serous, stage III and IV EOC using a 917 public gene list generated by the differentially expressed genes, after DNA microarray profiling, in MCF10A cells with normal expression and over-expression of the SRC gene. A number of gene clusters were identified upon hierarchical clustering which segregated the patient samples. To examine which of the genes from these clusters represented SRC activation we used a series of in vitro assays including phospho-SRC expression, sensitivity to SRC siRNA, and sensitivity to the SRC inhibitors Saracatinib and Dasatinib in 16 ovarian cancer cell lines. The ovarian cell lines were scored individually with the probesets representing the genes of interest from each of the individual gene clusters and the cell line in vitro data. Two gene clusters were selected for further investigation. A panel of 15 breast cancer cell lines were next investigated for SRC activation using the same in vitro assays described above. Correlation of SRC activation, by the scoring of the in vitro assays and the scoring of probesets from the individual gene clusters, demonstrated one gene cluster which showed good correlation and which potentially represented a metagene of SRC activation. The gene cluster consists of 16 genes, some of which are already known components of the SRC pathway. Furthermore, using the probesets representing the genes from this gene cluster to score public DNA microarray profiling data we were able to predict response to Dasatinib in two independent datasets of prostate and breast cancer cell lines (p=<0.05). In summary we have used a SRC activation transcriptional profile to cluster ovarian cancer samples and selected a metagene comprised of 16 genes which correlates to SRC activation in vitro using a panel of assays in ovarian and breast cancer cell lines. This has been shown to be predictive of response to Dasatinib in 2 independent datasets.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4820. doi:1538-7445.AM2012-4820