C-kit: Identification of co-regulated genes by gene expression profiling and clinical relevance of two breast cancer subtypes with stem cell like features

Abstract

622 Background: Expression of the proto-oncogene c-kit has been found in malignant tissue including a subset of breast cancers. c-Kit is also expressed in normal breast tissue and several authors found a loss of c-kit expression in breast carcinoma suggesting it might be involved in the growth control of mammary epithelium. Until now, only a few markers were described to be co-regulated with c-kit. To elucidate the possible role of c-kit in malignant transformation, we analyzed gene expression data of breast cancer patients. Methods: Tumor tissue of n=171 breast cancer patients were analyzed by gene expression profiling using Affymetrix Hg U133 Arrays (22,500 genes) and bioinformatic analyses. Tumor samples with high stromal and low epithelial cell content by gene expression profiling were excluded for further analysis. Validation was performed with n=100 independent samples. Results: A total of 10.5% of the tumors showed strong c-kit expression (2.5 fold above median). A careful dissection of global expression data revealed strong correlations of c-kit with the expression of a large cluster of genes containing several for whom c-kit coexpression was already described (HER1, CK-5/-17, PDGFR) as well as several members of the wnt signalling pathway, providing a possible novel link to mammary epithelial differentiation. Analysis of n=171 breast cancer samples according to this gene set allows the identification of putative “stem cell like” tumors (SCL) characterized by expression of several known stem cell markers. Surprisingly, a tight link of ER status and proliferation is restricted only to these SCL tumors but lost among non-SCL tumors. The clinical implications of our findings will be presented. Conclusions: For the first time these data bring together the description of two breast cancer subtypes identified by gene expression profiling with the actual stem cell model of the development of breast cancer. No significant financial relationships to disclose.