학술논문

The transcriptional landscape and biomarker potential of circular RNAs in prostate cancer
Document Type
article
Source
Genome Medicine, Vol 14, Iss 1, Pp 1-16 (2022)
Subject
Prostate
Cancer
Biomarker
circRNA
Medicine
Genetics
QH426-470
Language
English
ISSN
1756-994X
Abstract
Abstract Background Circular RNAs (circRNAs) constitute a largely unexplored source for biomarker discovery in prostate cancer (PC). Here, we characterize the biomarker potential of circRNAs in PC, where the need for novel diagnostic and prognostic tools to facilitate more personalized management is pressing. Methods We profiled the transcriptomic landscape of circRNAs in PC by total RNA sequencing of 31 adjacent-normal and 143 tumor samples from localized (radical prostatectomy (RP)) and metastatic PC patients (cohort 1, training). Diagnostic and prognostic potential was evaluated in cohort 1, and 39 top circRNA candidates were selected for validation in two additional PC cohorts (cohort 2, n = 111; RP cohort 3, n = 191) by NanoString-based expression analysis. Biochemical recurrence (BCR)-free survival was assessed using Kaplan-Meier, univariate, and multivariate Cox regression analyses. The circRNA candidates were further detected in extracellular vesicle (EV)-enriched plasma samples from PC patients and controls (cohort 4, n = 54). Results Expression of circABCC4, circFAT3, circATRNL1, and circITGA7 was highly cancer-specific (area under the curve 0.71–0.86), while low circITGA7 expression was significantly (P < 0.05) associated with BCR in univariate analysis in two RP cohorts. Moreover, we successfully trained and validated a novel 5-circRNA prognostic signature (circKMD1A/circTULP4/circZNF532/circSUMF1/circMKLN1) significantly associated with BCR beyond routine clinicopathological variables (RP cohort 1: P = 0.02, hazard ratio = 2.1; RP cohort 3: P < 0.001, hazard ratio = 2.1). Lastly, we provide proof-of-principle for detection of candidate circRNAs in EV-enriched plasma samples from PC patients. Conclusions circRNAs hold great biomarker potential in PC and display both high cancer specificity and association to disease progression.