[789] Molecular Heterogeneity of Neuroendocrine Prostate Cancer

H Beltran, R Esgueva, D Pflueger, S Terry, B Moss, A Sboner, ST Tagawa, F Demichelis, MB Gerstein, D Nanus, M Rubin. Weill Cornell Medical College, New York, NY; Yale University, New Haven, CT

Background: Although fewer than 5% of men have pure small cell prostate cancer at diagnosis, neuroendocrine (NE) differentiation is a common feature of advanced disease and is a poor prognostic sign. Little is known about the underlying biology of these tumors. The goal of this study is to characterize the transcriptome in order to gain insight into pathogenesis and identify potential therapeutic targets.
Design: We performed massively parallel paired-end RNA sequencing (RNASeq) from frozen tissue using the Illumina Genome Analyzer II and compared the transcription profiles between NE prostate cancer and hormone naive adenocarcinoma (adenoca). We used a bioinformatics approach to quantify gene expression at the exon level, evaluate alternative splicing events and point mutations, and utilized a novel computational tool (FusionSeq) to identify gene fusions.
Results: Of the 7 NE prostate cancers evaluated, 3 were pure small cell and the others were mixed with adenocarcinoma. As expected, the transcriptome of the 5 evaluated differed significantly from that of hormone naive adenoca. A total of 65-125 million paired-end reads were obtained (52-61% mappable to reference genome). Using FusionSeq, we detected 5 putative fusion candidates with a DASPER ranking score >2.0, including TMPRSS2-ERG rearrangement in 1 of the cases. NE associated genes (CHGA, CGHB, S100, CALCA) were high, androgen regulated genes (KLK3, TMPRSS2, NKX3-1) and androgen receptor (AR) expression were low, and EZH2, MIB1, and SYP (synaptophysin) were differentially expressed in the NE tumors in comparison to the adenoca cases, with the exception of the TMPRSS2-ERG fusion positive case which showed a gene expression pattern similar to adenoca.
Conclusions: A subset of NE prostate cancer molecularly resembles adenoca. Ongoing RNASeq analysis and pathway correlation will help verify and delineate the functional significance of our findings, better understand the transition from adenocarcinoma to small cell phenotype, and identify potential targetable lesions. It may also provide insight into the cell of origin of these tumors, which is currently debatable.
Category: Genitourinary (including renal tumors)

Monday, March 22, 2010 1:00 PM

Poster Session II # 111, Monday Afternoon


Close Window