Chromosomal instability in untreated primary prostate cancer as an indicator of metastatic potential

• Main Authors: Eric T. Miller, Sungyong You, Radu M. Cadaneanu, Minhyung Kim, Junhee Yoon, Sandy T. Liu, Xinmin Li, Lorna Kwan, Jennelle Hodge, Michael J. Quist, Catherine S. Grasso, Michael S. Lewis, Beatrice S. Knudsen, Michael R. Freeman, Isla P. Garraway
• Format: Article
• Language: English
• Published: BMC 2020-05-01
• Series: BMC Cancer
• Subjects: Prostate cancer; Metastases; Chromosomal instability; CIN; Prostate needle biopsies; TCGA
• Online Access: http://link.springer.com/article/10.1186/s12885-020-06817-1

Abstract Background Metastatic prostate cancer (PC) is highly lethal. The ability to identify primary tumors capable of dissemination is an unmet need in the quest to understand lethal biology and improve patient outcomes. Previous studies have linked chromosomal instability (CIN), which generates aneuploidy following chromosomal missegregation during mitosis, to PC progression. Evidence of CIN includes broad copy number alterations (CNAs) spanning > 300 base pairs of DNA, which may also be measured via RNA expression signatures associated with CNA frequency. Signatures of CIN in metastatic PC, however, have not been interrogated or well defined. We examined a published 70-gene CIN signature (CIN70) in untreated and castration-resistant prostate cancer (CRPC) cohorts from The Cancer Genome Atlas (TCGA) and previously published reports. We also performed transcriptome and CNA analysis in a unique cohort of untreated primary tumors collected from diagnostic prostate needle biopsies (PNBX) of localized (M0) and metastatic (M1) cases to determine if CIN was linked to clinical stage and outcome. Methods PNBX were collected from 99 patients treated in the VA Greater Los Angeles (GLA-VA) Healthcare System between 2000 and 2016. Total RNA was extracted from high-grade cancer areas in PNBX cores, followed by RNA sequencing and/or copy number analysis using OncoScan. Multivariate logistic regression analyses permitted calculation of odds ratios for CIN status (high versus low) in an expanded GLA-VA PNBX cohort (n = 121). Results The CIN70 signature was significantly enriched in primary tumors and CRPC metastases from M1 PC cases. An intersection of gene signatures comprised of differentially expressed genes (DEGs) generated through comparison of M1 versus M0 PNBX and primary CRPC tumors versus metastases revealed a 157-gene “metastasis” signature that was further distilled to 7-genes (PC-CIN) regulating centrosomes, chromosomal segregation, and mitotic spindle assembly. High PC-CIN scores correlated with CRPC, PC-death and all-cause mortality in the expanded GLA-VA PNBX cohort. Interestingly, approximately 1/3 of M1 PNBX cases exhibited low CIN, illuminating differential pathways of lethal PC progression. Conclusions Measuring CIN in PNBX by transcriptome profiling is feasible, and the PC-CIN signature may identify patients with a high risk of lethal progression at the time of diagnosis.