Tumor-stroma derived gene expression patterns as prognosticators in breast cancer

The stroma, which embeds epithelial cells, plays a major role in the shaping and physiological regulation of animal organs such as the breast. Aberrant functioning of the stroma was found to support tumorigenesis and cancer progression. Among many crucial aspects of breast cancer biology two issues focused our attention; (I) the impact of mutual interactions between the cancer cells and the stroma, specifically the endothelial cells and (II) the impact of insulin like growth factor one (IGF-I) on the gene expression profiles of stromal fibroblasts, as an example of the influence of a growth factor on the stroma. To address tumor-endothelial interaction we took advantage of an in vitro system mimicking the interaction of tumor and endothelial cells and explored gene expression changes using DNA microarrays. Our results suggest the interaction of endothelial cells and tumor cells that express the CD44+/CD24- signature indicative of stem cell-like cells, which are thought to have a low proliferative potential, might explain the unexpected and paradoxical association of the CD44+/CD24- signature with highly proliferative tumors that have an unfavorable prognosis. Furthermore, the gene expression signature induced in this system is of prognostic value in early stage and metastatic breast cancer. Since malignant epithelial cells and tumor-associated stromal cells are under the influence of hormones and growth factors, we examined the effects of IGF-I on cancer cells and primary fibroblasts in parallel to assess concordant and discordant gene expression changes. Our results show that primary breast fibroblasts, breast carcinoma associated fibroblasts and primary lung fibroblasts respond to IGF-I stimulation with increased expression of genes related with proliferation, which is phenotypically followed by an increased growth rate. Comparing in vitro gene expression data with available in vivo data, we have shown that the evoked gene expression signatures are able to stratify patients into groups with significantly different outcome. We propose that this prognostic gene expression signature might also serve as a predictor for the effectiveness of an anti-IGF-I therapy, a new therapeutic strategy that is currently in phase III development. Concluding, the data presented in this thesis underline the importance of the stroma, specifically the interaction between tumor and endothelial cells and the response of the stroma to stimulation with growth factors such as IGF-I. The gene expression changes in response to these interactions and stimulations carry prognostic information and might potentially be useful in clinical decision making.