A three-lncRNA expression signature associated with the prognosis of gastric cancer patients

Abstract

Long noncoding RNAs (lncRNAs) have emerged as essential players in gene regulation. An ever-increasing number of lncRNAs has been found to be associated with the biogenesis and prognosis of gastric cancer (GC). We aimed to develop an lncRNA signature with prognostic value for survival outcomes of GC. Using an lncRNA mining approach, we analyzed the lncRNA expression profiles of 492 GC patients from the Gene Expression Omnibus (GEO), which consisted of the GSE62254 set (N = 300) and the GSE15459 set (N = 192). The associations between the lncRNAs' expression and survival outcome were evaluated. A set of three lncRNAs (LINC01140, TGFB2-OT1, and RP11-347C12.10) was identified to significantly correlate with overall survival. These lncRNAs were then combined to form a single prognostic signature. Based on this three-lncRNA expression signature, patients in the GSE62254 set were classified into high- and low-risk subgroups with significantly different overall survival (hazard ratio [HR] = 1.93, P < 0.001) and disease-free survival (HR = 1.91, P < 0.001). Good reproducibility for the prognostic value of this lncRNA signature was confirmed in the GSE15459 set. Further analysis showed that the prognostic value of this signature was independent of some clinical characteristics. Gene set enrichment analysis indicated that high-risk scores positively related to several molecular pathways of cancer metastasis. Our results suggest that this innovative lncRNA expression signature may be a useful biomarker for the prognosis of patients with GC based on bioinformatics analysis.

Keywords: Gastric cancer; LncRNA; prognosis.

Figure 1

The flowchart of analyses to establish the risk score model and test its predictive value. GEO, Gene Expression Omnibus; GSEA, gene set enrichment analysis; ROC, receiver operating characteristic.

Figure 2

The distribution of the three‐lncRNA risk score, patients’ survival status, and lncRNA expression signature were analyzed in the GSE62254 set (N = 300). (A) LncRNA risk score distribution; (B) patients’ overall survival status and time; (C) patients’ disease‐free survival status and time; (D) heatmap of the lncRNA expression profiles. Rows represent lncRNAs and columns represent patients. The black dotted line represents the median lncRNA risk score cutoff dividing patients into low‐risk and high‐risk groups.

Figure 3

Kaplan–Meier estimates of the survival for patients using the three‐lncRNA signature. (A) Kaplan–Meier curves of overall survival for the GSE62254 set (N = 300); (B) Kaplan–Meier curves of disease‐free survival for the GSE62254 set (N = 300); (C) Kaplan–Meier curves of overall survival for the GSE15459 set (N = 192).

Figure 4

Performance evaluation of the three‐lncRNA signature in the GSE62254 set. (A) Gene set enrichment analysis delineates biological pathways associated with risk score using Cytoscape. Each node represents an enriched gene set, and they were grouped and annotated by the similarity according to related gene sets; (B) four typical cancer‐related pathways; (C) risk scores of patients with different TNM stages.

Figure 5

Receiver operating characteristic analysis of sensitivity and specificity by three lncRNAs, age, and TNM stage in predicting disease‐free survival in the GSE62254 set. (A) LINC01140, TGFB2‐OT1, and RP11‐347C12.10, and three‐lncRNA risk score; (B) age, TNM stage, three‐lncRNA risk score, and lncRNA risk score combined with TNM stage.