The identification of biomarkers capable of providing a reliable molecular diagnostic test for prostate cancer (PCa) is highly desirable clinically. We describe here 4 biomarkers, UDP-N-Acetyl-␣-D-galactosamine transferase (GalNAc-T3; not previously associated with PCa), PSMA, Hepsin and DD3/PCA3, which, in combination, distinguish prostate cancer from benign prostate hyperplasia (BPH).

The current approach to prostate cancer diagnosis has major limitations including the inability of prostate-specific antigen (PSA) assays to accurately differentiate between prostate cancer and benign prostate hyperplasia (BPH) and the imprecision of transrectal ultrasound (TRUS) biopsy sampling. We have employed cDNA microarray screening to compare gene expression patterns in BPH and tumour samples to identify expression markers that may be useful in discriminating between these conditions. Screening of 3 individual cDNA arrays identified 8 genes with expression 3-fold greater in 6 tumour tissues than in 1 nontumour sample and 1 BPH sample. Real-time PCR was used to confirm the overexpression of these 8 genes and 12 genes selected from the literature against a panel of 17 tumours and 11 BPH samples. Two genes, ␦-catenin (delta-catenin; CTNND2) and prostate-specific membrane antigen (PSMA; FOLH1), were significantly overexpressed in prostate cancer compared to BPH. Prostate epithelial cells stained positively for ␦-catenin and PSMA in our prostate cancer tissues, whereas the majority of our BPH tissues were negative for both markers. Thus we have identified ␦-catenin (not previously associated with prostatic adenocarcinoma) and confirmed the potential of PSMA as potential candidates for the diagnosis and management of prostate cancer.