22

Expression profiling in conjunction with protein sequence knowledge was used to identify prospective biomarkers of human proliferative breast disease (PBD). With a focus on extracellular proteins that offer potential use as biomarkers, we characterized gene expression profiles for the set of genes coding for secreted proteins, the “secretome,” using a cell line model of PBD. The MCF-10 series of human breast epithelial cells was used and includes: 1) the MCF-10A non-tumorigenic cells; 2) the MCF-10AT cells which were stably transfected with constitutively active Ha-Ras and sporadically produce tumors when implanted s.c. in nude mice; and 3) the MCF-10ATG3B cells, a third generation of MCF10AT cells with increased tumorigenicity in the nude mouse. Cells were cultured on Matrigel to achieve three dimensional growth. A compilation of three public databases was used to identify the human secretome. Expression profiles for secretome genes were obtained using Agilent cDNA microarrays, each representing approximately 13,000 human genes. Differentially expressed secretome genes were characterized using analysis of variance (ANOVA). With a minimum 2-fold change cutoff and a p-value ≤ 0.005, 133 differentially expressed secretome genes were identified. Cluster analysis reveals six distinct profiles for these genes. Twenty seven genes have an upward trend correlating with progression of the cell lines. Ontology analysis reveals that the major categories in this set include proteolysis and peptidolysis, cell adhesion, immune response and development. Among the up-regulated cluster are proteases PRSS2, PRSS11, and MMP2. Twelve secretome genes have a downward trend correlating with progression of the cell lines. This set includes genes coding for endopeptidase inhibitors and also genes involved in cell communication and cell adhesion. The down-regulated cluster includes proteinase inhibitors SERPINB2, SERPINE1, and WFDC2. RT-PCR was used to validate the microarray results for many of the secretome genes. Computational analysis of promoter regions provides evidence of transcriptional regulation by C/EBP-β for a number of genes in the up-regulated cluster. Immunoblot analysis reveals a nearly 4-fold increase in C/EBP-β phosphorylation corresponding with Ras activity in the cell line series. This work was supported by NIH grant ES 10595 and EHS Center grant P30 ES06639.

[Proc Amer Assoc Cancer Res, Volume 46, 2005]