Gene expression profiling in clinically localized prostate cancer: a four-gene expression model predicts clinical behavior.

PURPOSE: New diagnostic and prognostic molecular markers are required for prostate cancer, one of the most common male malignancies in Western countries. Gene expression profiling may help to identify genes involved in prostate carcinogenesis, yield clinical biomarkers, and improve tumor classification. EXPERIMENTAL DESIGN: To identify fundamental differences between normal and neoplastic prostate tissue, we used real-time quantitative RT-PCR assays to quantify the mRNA expression of 291 selected genes in samples of normal prostate and of well-documented primary, clinically localized prostate tumors. RESULTS: Forty-six genes showed significantly different expression in tumors relative to normal prostate. The dysregulated genes belong notably to the extracellular membrane and extracellular membrane remodeling categories and are involved in angiogenesis. Furthermore, we obtained a four-gene (XLKD1/LYVE1, CGA, F2R/PAR1, and BCL-G) model that discriminated between the seven patients with and the seven patients without relapse, independently of stage and grade. CONCLUSIONS: Some dysregulated genes are good candidates for use as molecular markers and/or therapeutic targets. Furthermore, differential gene expression profiling of clinically localized prostate tumors from relapsing and nonrelapsing patients identified a set of four genes with a pattern of expression that defines a molecular signature that could predict the clinical behavior of this disease.