PBOV1 as a potential biomarker for more advanced prostate cancer based on protein and digital histomorphometric analysis. Issue 7 (9th March 2018)

• Record Type: Journal Article
• Authors: Carleton, Neil M.; Zhu, Guangjing; Gorbounov, Mikhail; Miller, M. Craig; Pienta, Kenneth J.; Resar, Linda M. S.; Veltri, Robert W.
• Is Part Of: Prostate. Volume 78:Issue 7(2018)
• Page Start: 547
• Page End: 559
• Publication Date: 2018-03-09
• Subjects: digital image analysis -- immunohistochemistry -- nuclear alterations in cancer cells -- nuclear morphometry -- radical prostatectomy samples -- tissue microarray; Prostate -- Diseases -- Periodicals; 616
• DOI: 10.1002/pros.23499
• Languages: English
• ISSNs: 0270-4137

Abstract

Abstract : Background: There are few tissue‐based biomarkers that can accurately predict prostate cancer (PCa) progression and aggressiveness. We sought to evaluate the clinical utility of prostate and breast overexpressed 1 (PBOV1) as a potential PCa biomarker. Methods: Patient tumor samples were designated by Grade Groups using the 2014 Gleason grading system. Primary radical prostatectomy tumors were obtained from 48 patients and evaluated for PBOV1 levels using Western blot analysis in matched cancer and benign cancer‐adjacent regions. Immunohistochemical evaluation of PBOV1 was subsequently performed in 80 cancer and 80 benign cancer‐adjacent patient samples across two tissue microarrays (TMAs) to verify protein levels in epithelial tissue and to assess correlation between PBOV1 proteins and nuclear architectural changes in PCa cells. Digital histomorphometric analysis was used to track 22 parameters that characterized nuclear changes in PBOV1‐stained cells. Using a training and test set for validation, multivariate logistic regression (MLR) models were used to identify significant nuclear parameters that distinguish Grade Group 3 and above PCa from Grade Group 1 and 2 PCa regions. Results: PBOV1 protein levels were increased in tumors from Grade Group 3 and above (GS 4 + 3 and ≥ 8) regions versus Grade Groups 1 and 2 (GS 3 + 3 and 3 + 4) regions ( P = 0.005) as assessed by densitometry of immunoblots. Additionally, by immunoblotting, PBOV1 protein levels differed significantly between Grade Group 2 (GS 3 + 4) and Grade Group 3 (GS 4 + 3) PCa samples ( P = 0.028). In the immunohistochemical analysis, measures of PBOV1 staining intensity strongly correlated with nuclear alterations in cancer cells. An MLR model retaining eight parameters describing PBOV1 staining intensity and nuclear architecture discriminated Grade Group 3 and above PCa from Grade Group 1 and 2 PCa and benign cancer‐adjacent regions with a ROC‐AUC of 0.90 and 0.80, respectively, in training and test sets. Conclusions: Our study demonstrates that the PBOV1 protein could be used to discriminate Grade Group 3 and above PCa. Additionally, the PBOV1 protein could be involved in modulating changes to the nuclear architecture of PCa cells. Confirmatory studies are warranted in an independent population for further validation.