A SAGE (Serial analysis of gene expression) view of breast tumor progression

Abstract

To identify molecular alterations involved in the initiation and progression of breast carcinomas, we analyzed the global gene expression profiles of normal mammary epithelial cells and in situ, invasive, and metastatic breast carcinomas using serial analysis of gene expression (SAGE). We identified sets of genes expressed only or most abundantly in a specific stage of breast tumorigenesis or in a certain subtype of tumors through the pair-wise comparison and by hierarchical clustering analysis of these eight SAGE libraries (two/stage). On the basis of these comparisons, we made the following observations: Normal mammary epithelial cells showed the most distinct and least variable gene expression profiles. Many of the genes highly expressed in normal mammary epithelium and lost in carcinomas encoded secreted proteins, cytokines, and chemokines, implicating abnormal paracrine and autocrine signaling in the initiation of breast tumorigenesis. Very few genes were universally up-regulated in all tumors regardless of their stage and histological grade, indicating a high degree of diversity at the molecular level that likely reflects the clinical heterogeneity characteristic of breast carcinomas. Tumors of different histology type and stage had very distinct gene expression patterns. No genes seemed to be specific for metastatic or for in situ carcinomas. We found that the most dramatic and consistent phenotypic change occurred at the normal-to-in situ carcinoma transition. This observation, combined with the fact that many of the genes involved encode secreted, cell-nonautonomous factors, implies that the normal epithelium-to-in situ carcinoma transition may be the most promising target for cancer prevention and treatment.