Genome-wide long non-coding RNAs identified a panel of novel plasma biomarkers for gastric cancer diagnosis

Abstract

Background: Although long non-coding RNAs (lncRNAs) are regarded as useful plasma-based biomarkers for cancer detection, the potential diagnostic value of lncRNAs in gastric cancer (GC) remains unclear.

Methods: To screen promising lncRNAs biomarkers for GC, we performed genome-wide lncRNA microarray assay between five GC cases plasma and matched healthy controls plasma. The expression of candidate plasma-related lncRNAs were validated in two-phase validation of 446 subjects. The receiver operating characteristic curve was constructed for evaluating diagnostic accuracy. We also determined the origin and stability of plasma lncRNAs, and investigated biological effects of candidate lncRNAs on cellular phenotypes.

Results: A total of 3878 lncRNAs were expressed differentially in GC plasma, among which the top 10 up-regulated lncRNAs were selected for further validation. A two-stage validation revealed that plasma levels of three lncRNAs (FAM49B-AS, GUSBP11, and CTDHUT) were significantly higher in GC plasma as compared with healthy controls (P < 0.05), and the combined area under curve of these lncRNAs was 0.818 (95% CI 0.772-0.864). Moreover, these lncRNAs were stable and detectable in human plasma, and also enriched in extracellular fluid. The expression levels of all three lncRNAs dropped significantly on day 10 after radical surgery compared with preoperative levels (P < 0.05). Also, lncRNA FAM49B-AS significantly promoted GC cell viability and invasion.

Conclusions: Plasma lncRNA FAM49B-AS, GUSBP11 and CTDHUT have a strong potential to serve as noninvasive biomarkers for GC diagnosis.

Keywords: Biomarker; Diagnosis; Gastric cancer; LncRNA; Plasma.