Abstract 4173: Prostate derived Ets (E-twenty six) transcription factor (PDEF) as a prognostic biomarker for metastatic prostate cancer

While early stage prostate cancer (CaP) is unique in that many patients will take decades to progress to symptomatic disease; there exists a subset of patients whose disease may quickly become life-threatening. Unfortunately, due to the high incidence of CaP and clinicians’ inability to differentiate the majority of patients with disease that will never become symptomatic from the minority with aggressive life-threatening tumors, patients with CaP are often over-treated. Prognostic biomarkers, such as prostate specific antigen (PSA), have shown great efficacy in improving patients’ chances of surviving cancer by allowing physicians to identify the disease at an earlier stage, though there does not exist a biomarker for CaP that adequately differentiates disease that will quickly become metastatic. In order to facilitate prognosis, the expression of prostate derived Ets transcription factor (PDEF), a protein of significance to the progression of CaP, was studied. PDEF is expressed in normal epithelial cells and negatively regulates transcriptional networks controlling migration and invasion. Previous findings in breast cancer have also demonstrated that loss of PDEF expression results in upregulation of the antiapoptotic protein survivin, indicating a possible connection to growth and survival. Loss of PDEF expression has also been shown to increase the metastatic potential of several cancers. Patients PDEF expression was determined via immunohistochemistry staining of tissue microarray in set of over 700 archival CaP tissues taken at radical prostatectomy; and patients were subsequently stratified based on their tumor PDEF expression. The stratified CaP patients were then analyzed for the association of PDEF expression status with clinical outcomes. Univariate analysis showed that low PDEF expression is correlated with incidence of biochemical failure as defined by American Urological Association (AUA; OR=1.30, p=0.0493) or National Comprehensive Cancer Network (NCCN; OR=1.36, p=0.0164) definitions; and with metastatic disease incidence (OR=3.61, p=0.0047). Multivariate analysis also showed that low PDEF expression was associated with metastatic disease incidence (OR = 4.02, p=0.0146). Furthermore, Kaplan-Meier survival analysis showed that stratification of patients into subgroups based on PDEF expression (high, middle, low) was able to significantly predict relative time to biochemical failure (NCCN, p=0.0241) or metastatic disease (p=0.0084). This study