Introduction

TMPRSS2:ERG fusion is the most frequent genetic event in prostate cancer (PC), resulting in ERG overexpression. In Western populations, approximately 50% of PCs express fusion products; however, our previous study showed that ERG expression was less frequent in a cohort of Japanese patients with localized PC (15/92, 16.3%). The association between ERG overexpression and the clinical behavior of PC is controversial. SPINK1 overexpression is noted in approximately 10% of patients with localized PC, mutually exclusive of ERG overexpression. Although the association of SPINK1 overexpression with an aggressive PC phenotype has been reported, its prognostic significance is unclear. PC is occasionally diagnosed at autopsy and during cystoprostatectomy for bladder cancer; most of these cases are thought to represent the latent phase of the tumor. Thus, assessment of ERG and SPINK1 expression in latent, incidental, and metastatic PC (the most aggressive form) may be useful in evaluating the association between ERG and SPINK1 expression and the biological aggressiveness of the tumor.

Methods

In total, 151 autopsies among institutional autopsy records from 2009 to 2015 and 84 cystoprostatectomy specimens were included. Each prostate gland was fixed and sliced into step sections. Ninety-eight prostate biopsy specimens from a cohort of patients who received an initial diagnosis of metastatic PC between 2003 and 2012 were investigated. ERG and SPINK1 expression was assessed by immunohistochemistry and the expression patterns were compared to clinicopathological parameters.

Results

PC was identified in 53/151 autopsies and 20/84 cystoprostatectomy specimens. ERG and SPINK1 expression patterns were not significantly different [4/73 (5.5%) and 5/73 (8.3%) in the latent/incidental PC cohort and 14/98 (14.3%) and 12/98 (12.2%) in the metastatic PC cohort, respectively]. SPINK1 was almost exclusively expressed in ERG-negative tumors, except in one case. In the metastatic PC cohort, ERG and SPINK1 expression patterns were not associated with age, prostate-specific antigen level, and the Gleason score. SPINK1 expression was significantly associated with a shorter time to castration-resistant PC (P=0.0151), although ERG expression was not associated with clinical outcomes.

Conclusions

ERG and SPINK1 expression was not significantly different between latent/incidental PC and metastatic PC. SPINK1 expression may be a predictor of a shorter response to androgen deprivation therapy in metastatic PC.

Funding

none