Construction and Analysis of the Tumor-Specific mRNA-miRNA-lncRNA Network in Gastric Cancer

Xiaohao Zheng; Xiaohui Wang; Li Zheng; Hao Zhao; Wenbin Li; Bingzhi Wang; Liyan Xue; Yantao Tian; Yibin Xie

doi:10.3389/fphar.2020.01112

Construction and Analysis of the Tumor-Specific mRNA-miRNA-lncRNA Network in Gastric Cancer

Front Pharmacol. 2020 Jul 21:11:1112. doi: 10.3389/fphar.2020.01112. eCollection 2020.

Authors

Xiaohao Zheng¹, Xiaohui Wang², Li Zheng³, Hao Zhao⁴, Wenbin Li⁵, Bingzhi Wang⁵, Liyan Xue⁵, Yantao Tian¹, Yibin Xie¹

Affiliations

¹ Department of Pancreatic and Gastric Surgery, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
² Department of General Surgery, Xuanwu Hospital, Capital Medical University, Beijing, China.
³ Department of General Surgery, The First People's Hospital of Dongcheng District, Beijing, China.
⁴ Department of Cardiovascular Surgery, China-Japan Friendship Hospital, Beijing, China.
⁵ Department of Pathology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.

Abstract

Weighted correlation network analysis (WGCNA) is a statistical method that has been widely used in recent years to explore gene co-expression modules. Competing endogenous RNA (ceRNA) is commonly involved in the cancer gene expression regulation mechanism. Some ceRNA networks are recognized in gastric cancer; however, the prognosis-associated ceRNA network has not been fully identified using WGCNA. We performed WGCNA using datasets from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) to identify cancer-associated modules. The criteria of differentially expressed RNAs between normal stomach samples and gastric cancer samples were set at the false discovery rate (FDR) < 0.01 and |fold change (FC)| > 1.3. The ceRNA relationships obtained from the RNAinter database were examined by both the Pearson correlation test and hypergeometric test to confirm the mRNA-lncRNA regulation. Overlapped genes were recognized at the intersections of genes predicted by ceRNA relationships, differentially expressed genes, and genes in cancer-specific modules. These were then used for univariate and multivariate Cox analyses to construct a risk score model. The ceRNA network was constructed based on the genes in this model. WGCNA-uncovered genes in the green and turquoise modules are those most associated with gastric cancer. Eighty differentially expressed genes were observed to have potential prognostic value, which led to the identification of 12 prognosis-related mRNAs (KIF15, FEN1, ZFP69B, SP6, SPARC, TTF2, MSI2, KYNU, ACLY, KIF21B, SLC12A7, and ZNF823) to construct a risk score model. The risk genes were validated using the GSE62254 and GSE84433 datasets, with 0.82 as the universal cutoff value. 12 genes, 12 lncRNAs, and 35 miRNAs were used to build a ceRNA network with 86 dysregulated lncRNA-mRNA ceRNA pairs. Finally, we developed a 12-gene signature from both prognosis-related and tumor-specific genes, and then constructed a ceRNA network in gastric cancer. Our findings may provide novel insights into the treatment of gastric cancer.

Keywords: Gene Expression Omnibus; Genotype-Tissue Expression; The Cancer Genome Atlas; competing endogenous RNA; gastric cancer; risk score; weighted correlation network analysis.