Alternative Splicing Provides a Novel Molecular Mechanism for Prostatic Small-cell Neuroendocrine Carcinoma

Abstract

The vast majority of prostate cancers are adenocarcinoma with glandular formation and expression of luminal differentiation markers including androgen receptor (AR) and prostate-specific antigen (PSA). Scattered among the luminal-type cancer cells are rare neuroendocrine (NE) cells with an elongated cell body and secretory functions. In rare cases, patients develop small-cell neuroendocrine carcinoma (SCNC) composed entirely of NE tumor cells. SCNC does not form a glandular structure but grows as single cells, cords, or solid sheets. Instead of luminal differentiation, tumor cells often express NEmarkers such as chromogranin A and synaptophysin. While primary SCNC is rare, SCNC development after failure of hormone therapy among patients with a history of adenocarcinoma ismore common. Although this phenomenon is well documented in the literature, the true incidence of SCNC in this setting is unknown because most patients for whom hormonal therapy fails do not undergo another biopsy. In studying the molecular mechanisms that drive the phenotypic changes from adenocarcinoma to SCNC, Beltran and colleagues [1] discovered overexpression and amplification of AURKA and MYCN in many cases of SCNC. Subsequently, Lee and colleagues [2] showed that MYCN is a likely driver of SCNC. Lotan’s group [3] discovered that [6_TD$DIFF]RB1 loss is characteristic of prostate SCNC. Our laboratory [4] showed that TP53 is frequently mutated in SCNC. These findings in human SCNC tissue are consistent with observations in mouse models. Transgenic expression of SV40 early genes, which inactivates both Rb and p53 in mouse prostate, leads to prostate SCNC [5]. A recent study by Collins’ group [6] showed that the placental gene PEG10 promotes SCNC. As PEG10 is regulated by Rb and p53 and is a cell cycle regulator, this finding points to a potential mechanism linking Rb and p53 to SCNC. More recently, Lotan’s group [7] identified loss of cyclin D1, another cell cycle regulator, as a feature of prostate SCNC. In addition to the above findings, Collin’s group [8] observed significant downregulation of the REST transcriptional complex in SCNC. Since REST is a transcription factor and the master repressor of neuronal differentiation, this finding explains the phenotypic changes (ie, expression of NE markers) observed in SCNC. In this issue of European Urology, Li et al [9] expand on this important finding and report the discovery of alternative splicing involving REST as a mechanism driving the NE phenotype. The authors developed a novel bioinformatic tool to analyze alternative RNA splicing in RNA-sequencing data from the Beltran cohort and the Vancouver Prostate Centre cohort. They discovered thatmore than 66% of the splice events are likely regulated by the RNA splicing factor SRRM4, a master regulator required for transdifferentiation of embryonic stem cells to neural cells. Experimental studies confirmed that SRRM4modulates the splicing of REST, leading to lower levels of REST transcripts and higher levels of the truncated variant transcript REST4. In human SCNC tumor samples, elevated SRRM4 expression is negatively associated with the REST/REST4 expression ratio. While SRRM4 targets alternative splicing of REST, blockage of AR signaling EU RO P E AN URO L OG Y 7 1 ( 2 0 1 7 ) 7 9 – 8 0