Identification of circRNA–miRNA–mRNA networks contributes to explore underlying pathogenesis and therapy strategy of gastric cancer

• Published in: Journal of translational medicine Vol. 19; no. 1; pp. 1 - 18
• Main Authors: Dong, Zhijie, Liu, Zhaoyu, Liang, Min, Pan, Jinhui, Lin, Mingzhen, Lin, Hai, Luo, Yuanwei, Zhou, Xinke, Yao, Wenxia
• Format: Journal Article
• Language: English
• Published: London BioMed Central 28.05.2021; BioMed Central Ltd; Springer Nature B.V; BMC
• Subjects: Activating transcription factor 3; Algorithms; Argonaute 2 protein; Binding sites; Biomarkers; Biomedical and Life Sciences; Biomedicine; BTG2 protein; ceRNA; circRNA; Cmap; Computer applications; Cytoplasm; Datasets; Development and progression; DNA microarrays; EGR-1 protein; Fibroblast growth factor 2; FosB protein; Gastric cancer; Gene expression; Genetic aspects; Genomes; Health aspects; Identification and classification; Medical bioinformatics; Medical prognosis; Medicine/Public Health; miRNA; Pathogenesis; Physiological aspects; Polymerase chain reaction; Protein-protein interactions; Proteins; Reverse transcription; Ribonuclease; RNA; Software; Stomach cancer; Trichostatin A; China; ceRNA; Cmap; circRNA; Gastric cancer
• ISSN: 1479-5876; 1479-5876
• DOI: 10.1186/s12967-021-02903-5

Background Circular RNAs (circRNAs) are a new class of noncoding RNAs that have gained increased attention in human tumor research. However, the identification and function of circRNAs are largely unknown in the context of gastric cancer (GC). This study aims to identify novel circRNAs and determine their action networks in GC. Methods A comprehensive strategy of data mining, reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and computational biology were conducted to discover novel circRNAs and to explore their potential mechanisms in GC. Promising therapeutic drugs for GC were determined by connectivity map (CMap) analysis. Results Six overlapped differentially expressed circRNAs (DECs) were screened from selected microarray and RNA-Seq datasets of GC, and the six DECs were then validated by sanger sequencing and RNase R treatment. Subsequent RT-qPCR analysis of GC samples confirmed decreased expressions of the six DECs (hsa_circ_0000390, hsa_circ_0000615, hsa_circ_0001438, hsa_circ_0002190, hsa_circ_0002449 and hsa_circ_0003120), all of which accumulated preferentially in the cytoplasm. MiRNA binding sites and AGO2 occupation of the six circRNAs were predicted using online databases, and circRNA–miRNA interactions including the six circRNAs and 33 miRNAs were determined. Then, 5320 target genes of the above 33 miRNAs and 1492 differently expressed genes (DEGs) from The Cancer Genome Atlas (TCGA) database were identified. After intersecting the miRNA target genes and the 889 downregulated DEGs, 320 overlapped target genes were acquired. The Kyoto Encyclopedia of Genes and Genomes enrichment analysis indicated that these target genes were related to two critical tumor-associated signaling pathways. A protein–protein interaction network with the 320 target genes was constructed using STRING, and fifteen hubgenes (ATF3, BTG2, DUSP1, EGR1, FGF2, FOSB, GNAO1, GNAI1, GNAZ, GNG7, ITPR1, ITPKB, JUND, NR4A3, PRKCB) in the network were identified. Finally, bioactive chemicals (including vorinostat, trichostatin A and astemizole) based on the fifteen hubgenes were identifed as therapeutic agents for GC through the CMap analysis. Conclusions This study provides a novel insight for further exploration of the pathogenesis and therapy of GC from the circRNA-miRNA-mRNA network perspective.