Chemotherapy induced dynamic gene expression changes in vivo are prognostic in ovarian cancer

Carboplatin and paclitaxel are the most widely prescribed chemotherapeutic agents for ovarian cancer. Not all patients respond to treatment, so there is a need for biomarkers that reliably predict resistance in ovarian tumors. Expression of such biomarkers may be dynamically controlled. Gene expression was assessed for a period of 14 days after treatment with carboplatin or combined carboplatin-paclitaxel in xenografts from two ovarian cancer models: chemosensitive serous adenocarcinoma derived OV1002 and slow growing, chemoresistant HOX424 of clear cell origin. Tumour volume reduction was observed in both cell lines post treatment, with a more prominent effect in OV1002, which subsided in late time points. In OV1002, hierarchical clustering classified differentially expressed genes into four time-related patterns, upregulated and downregulated groups for each early and late expressed genes. Upregulated genes were involved in DNA repair, cell cycle and apoptosis, while downregulated groups were involved in oxygen consuming metabolic processes and apoptosis control. Carboplatin-paclitaxel treatment triggered a more comprehensive response. HOX424 responded only to the combined treatment, while the observed reduction in tumour growth was limited. Several apoptosis and cell cycle genes were upregulated, while Wnt signaling was downregulated in the exclusively late expression pattern observed in this cell line. Late downregulated gene groups post carboplatin-taxane treatment were capable of predicting overall survival in two independent clinical datesets. Pathways overrepresented in these clusters were also predictive of outcome. This longitudinal gene expression study may help characterization of chemotherapy response, identification of resistance biomarkers and guiding timing of biopsies.

101

Koussounadis A, Langdon SP, Harrison DJ, Smith VA