Detection of neuroendocrine prostate cancer through plasma-based epigenomic profiling.

Abstract

201 Background: Neuroendocrine prostate cancer (NEPC) is an aggressive resistance phenotype that can develop in men with metastatic prostate adenocarcinoma (PRAD). Early detection of NEPC has prognostic and therapeutic implications, but is challenging due to spatial heterogeneity, sampling error associated with tissue biopsy, and a lack of distinguishing genetic mutations that can be assessed from cell-free DNA (cfDNA). Herein, we introduce a novel epigenomic-based approach to non-invasively identify NEPC from cfDNA. Methods: Plasma was collected from patients with NEPC and PRAD at the Dana-Farber Cancer Institute. Chromatin immunoprecipitation and sequencing on cfDNA (cfChIP-seq) was performed on 1mL of plasma for the post-translational histone modifications H3K4me3 and H3K27ac. Activity of gene promoters and enhancers was inferred based on H3K4me3 and H3K27ac cfChIP-seq signal normalized to signal at constitutively active regulatory elements. A classifier was built to distinguish NEPC from PRAD based on H3K27ac signal at a published set of regulatory elements with increased chromatin accessibility in neuroendocrine tumors compared to adenocarcinomas (NE-REs). The classifier performance was evaluated using the area under the receiver operating characteristic (AUROC) curve. Results: Our cohort included 30 patients, 10 with NEPC and 20 with PRAD. H3K4me3 signal at the promoter of chromogranin A ( CHGA), a diagnostic marker gene of NE-differentiation, was significantly elevated in plasma from patients with NEPC vs. PRAD (p=4x10-11, Wilcoxon rank-sum test). H3K27ac signal at binding sites for ASCL1, a master transcription factor driving NE-differentiation, was significantly elevated in NEPC vs. PRAD plasma samples (p=7.2x10-5, Wilcoxon rank-sum test). H3K27ac signal at NE-REs accurately distinguished patients with NEPC from patients with PRAD with an AUROC curve of 0.97. Conclusions: We present a non-invasive approach to identify NE-differentiation in patients with metastatic prostate cancer using 1mL of plasma. This method could help overcome practical challenges associated with NEPC detection and further guide therapy selection.