Identification of novel key regulatory lncRNAs in gastric adenocarcinoma

Abstract

Background: Stomach adenocarcinoma (STAD) is one of the most common and deadly cancers worldwide. Recent evidence has demonstrated that dysregulation of long noncoding RNAs (lncRNA) is associated with different hallmarks of cancer. lncRNAs also were suggested as novel promising biomarkers for cancer diagnosis and prognosis. Despite these previous investigations, the expression pattern, diagnostic role, and hallmark association of lncRNAs in STAD remain unclear.

Results: In this study, The STAD lncRNA-mRNA network was constructed based on RNAs that differentially expressed among tumor and normal samples and had a strong expression correlation with others. The high degree nodes of the network were associated with overall survival. In addition, we found that the hubs' regulatory roles have previously been confirmed in different types of cancers by literature. For example, the HCG22 hub inhibited cell proliferation and invasion and induced apoptosis in oral squamous cell carcinoma (OSCC) cells. The levels of PCNA, Vimentin, and Bcl2 were decreased and E-cadherin and Bax expression was elevated in OSCC cells after HCG22 overexpression. Additionally, HCG22 overexpression inhibited the Akt, mTOR, and Wnt/β-catenin pathways. Then lncRNAs were mapped to their related GO terms and cancer hallmarks. Based on these mappings, we predict the hallmarks that might be associated with each lncRNA. Finally, the literature review confirmed our prediction. Among the 20 lncRNAs of the STAD network, 11 lncRNAs (LINC02560, SOX21-AS1, C5orf66-AS1, HCG22, PGM5-AS1, NALT1, ENSG00000241224.2, TINCR, MIR205HG, HNF4A-AS1, ENSG00000262756) demonstrated expression correlation with overall survival (OS). Based on expression analysis, survival analysis, hallmark associations, and literature review, LINC02560, SOX21-AS1, C5orf66-AS1, HCG22, PGM5-AS1, NALT1, ENSG00000241224.2, TINCR, MIR205HG, HNF4A-AS1 plays a regulatory role in STAD. For example, our prediction of association between C5orf66-AS1 expression dysregulation and "sustaining proliferative signal" and "Activating invasion and metastasis" has been confirmed in STAD, OSCC and cervical cancer. Finally, we developed a lncRNA signature with SOX21-AS1 and LINC02560, which classified patients into high and low-risk subgroups with significantly different survival outcomes. The mortality rate of the high-risk patients was significantly higher compared to the low-risk patients (28/1% vs 60.13).

Conclusion: These findings help in designing more precise and detailed experimental studies to find STAD biomarkers and therapeutic targets.

Keywords: Biomarker; Stomach adenocarcinoma; Survival analysis; lncRNA; mRNA.

Fig. 1

Workflow for key regulatory lncRNA detection in STAD

Fig. 2

Volcano plots showing the differential expression of mRNAs and lncRNAs in STAD; X-axis indicates the expression differences of mRNAs, and lncRNAs between STAD and normal samples, and Y-axis represents log transformed false discovery rate (FDR) values. Red dots represent the up-regulation and green dots represent down-regulation

Fig. 3

STAD mRNA-lncRNA network. mRNAs represent by rectangles, and lncRNAs represent by circles. Node filling color indicates log fold change of differential analysis among tumor and normal samples. The node’s border thickness represents the node degree

Fig. 4

Kaplan–Meier survival curves for significant lncRNAs associated with overall survival of the STAD patients. Log-rank method was used to assess the survival differences between the two groups. Horizontal axis is OS time (days) and vertical axis stands for survival function. The turquoise lines represent the group with low count, and the red lines represent the group with high count

Fig. 5

Degree distribution of STAD mRNA-lncRNA network

Fig. 6

A. The survival differences between the high-risk and low-risk groups of lncRNA signature were determined by the log-rank test. B. Receiver operating characteristic (ROC) analysis of the risk scores for overall survival prediction

Fig. 7

Connection among RNAs and their related GO terms and cancer hallmark. The chart depicts lncRNAs by green circles, mRNAs by green Diamond, GO terms by orange rectangles and cancer hallmarks by brown rectangles. Red circles represent cancer-related genes, blue circles represent lncRNAs that were significantly associated with overall survival and brown circles are those lncRNAs that construct the STAD signature