Tissue biomarkers in prostate cancer

Abstract: Prostate cancer (PC) is the most common male cancer in the western world. Better biomarkers are needed to support diagnosis, prediction of prognosis and treatment decision Radical prostatectomy (RP) specimens are routinely immersed in formalin overnight. Formalin may also be injected into the prostate for improved fixation. We report that formalin injection does not alter tissue volumes compared to conventional fixation. Formalin may affect epitopes for immunohistochemistry (IHC). Immunoreactivity was compared between fixation methods with no significant difference for the majority of 15 antibodies. We investigated the transcription factor pancreatic duodenal homeobox-1 (PDX-1) and heat shock proteins (HSP) 27, 60 and 70 as prognostic markers in PC. A tissue microarray (TMA) of 289 PCs was constructed and immunostained. HSP 27 and 60, but not HSP 70 and PDX-1 correlated with biochemical recurrence. In multivariate analysis including histopathological prognostic factors, only HSP 60 was an independent predictor of prognosis. PDX-1 was overexpressed in cancer vs. benign tissue but also in atrophy and high-grade prostatic intraepithelial neoplasia (PIN) vs. cancer. The role of GAD1 (glutamate decarboxylase 1) as prostate-specific biomarker was investigated. A TMA of benign and malignant tissues from prostate, rectum, lung and bladder was stained for GAD1, prostate-specific antigen (PSA) and prostate-specific membrane antigen (PSMA, glutamate carboxypeptidase). Presence of GAD1 protein was validated by Western blot and real-Time PCR (RT-PCR). By IHC, the expression of GAD1 and PSA was stronger in prostatic tissues than in controls. PSMA was stronger in prostate cancer than in urothelial and rectal cancer but had lower specificity than GAD1 and PSA. The intra- and interobserver reproducibility of IHC evaluation in TMA were assessed. Intensity and extent of PDX-1 immunostains of 50 PCs were scored twice by 4 independent observers. Mean weighted kappa for intra- and interobserver agreement was 0.85 and 0.80 for intensity and 0.43 and 0.21 for extent with similar results for 2 pathologists and 2 non-pathologists. Thus, subjective assessment of intensity is highly reproducible while estimation of staining extent is less reliable. In conclusion, we find that TMA is a valuable tool for tissue-based biomarker research. We have attempted to optimize the procedures from tissue handling to evaluation. We also present HSP 27 and HSP 60 as potential predictors of prognosis in PC and GAD1 as a new prostate-specific biomarker.