Novel Genomic Biomarker Panel Predictive of Aggressive Disease in African American Men with Prostate Cancer

Abstract

Mutations in the E26 transforming specific (ETS) transcription factors occur early in the oncogenic process of prostate cancer. These mutations behave in a race dependent manner with the majority of European American (EA) men possessing ERG positive (ERG+) tumors while ≥75% of African American (AA) men have ERG negative (ERG-) tumors. In order to establish clinically relevant biomarkers in AA men, we studied ETS gene expression differences among AA men and EA men and identified race specific ETS negative (ETS-) genes that predict for clinically relevant outcomes. Previously published datasets from the Decipher Genomic Resource Information Database (GRID) were used to obtain microarray data. The validated microarray-based ETS positive (ETS+) (ERG or ETV1/4/5) over expression was used to determine ETS+ status in the GRID database. Samples with overexpression of ERG, ETV1-5, and FLI1 were categorized as ETS+. Chi square (X2)/Fisher Exact method was used to assess the race differences in clinico-pathologic endpoints such as extraprostatic extension, seminal vesicle invasion, pathologic Gleason Score, and the post-surgical Cancer of the Prostate Risk Assessment Score (CAPRA-S) . Logistic regression model was used to analyze the association between aforementioned clinico-pathologic endpoints and ETS status. Genes that showed differential expression were evaluated. On multivariable logistic regression analysis, race was found to be the strongest predictor of ETS- status. Analysis showed that AA men were four times more likely to be ETS- than EA men (p<0.001). Using a q value ≤0.05 and an expression fold change level of 0.05, we identified 3047 genes (6.6%) that were differentially expressed based on ETS status. Of these, 248 were found to be differentially expressed in a race-specific manner. Using a false-discovery rate q<0.1 we identified 143 genes that predicted for one or more clinical endpoint. In order to determine which of these race specific ETS- genes are significantly overexpressed in tumors we selected genes that met all three of the following criteria; significantly overexpressed in ETS- compared to ETS+ tumors, preferentially overexpressed in AA and not in EA, and predictive for one or more adverse outcome in a race-specific manner. Using this criteria, we have identified a 19-gene panel that includes ADAMTS1, AFF3, ANKRD20, ANPEP, CCL3, CYP4F35, FAR2P1, INPP4B, MAP1LC3, MIPEP, NPIPB7, NPNT, OPHN1, OR4K7P, PCGEM1, RLN2, SCIN, SPINK1, and TTN. In an effort to elucidate biologic features associated with aggressiveness of prostate cancer in AA men, we have identified ETS- biomarkers that predict for one or more adverse clinical outcome in AA men. Our investigation found 19 ETS- genes that are both overexpressed in AA men and predictive of clinical outcomes thus representing ideal candidates for prognostic biomarkers.