Objective: RNA-seq data and miRNA-seq data of lung adenocarcinoma (LUAD) were analyzed to identify critical long non-coding RNAs (lncRNAs) and disclose molecular pathogenesis.
Materials and methods: RNA-seq data and miRNA-seq data were downloaded from TCGA. Differentially expressed lncRNAs (DELs) and microRNAs (DEMs) were revealed by two sample t-test. |Fold change| > 2 and p-value < 0.01 were set as the cutoffs. Univariate Cox regression was performed to disclose prognostic lncRNAs. Information about miRNA-lncRNA interactions and miRNA-mRNA interactions were acquired from miRcode and miRTarBase, respectively. A miRNA-lncRNA-mRNA regulatory network was then constructed, from which regulatory modules were identified. Functional enrichment analysis was performed with DAVID.
Results: A total of 57 DELs and 118 DEMs were identified from 507 LUAD compared with 19 normal samples. Three DELs, including MEG3, MIAT and MIR4697HG, were associated with clinical features, while nine DELs (LINC00115, LINC00265, LINC01001, LINC01002, MIR22HG, NFYC-AS1, SNHG10, THUMPD3-AS1 and TMPO-AS1) were revealed to be prognostic biomarkers. A regulatory network including 61 miRNA-lncRNA interactions and 304 miRNA-mRNA interactions was constructed, from which 19 lncRNA-miRNA-mRNA regulatory modules were identified. Among the modules, MEG3 and MIAT may play important roles in the development of LUAD by interactions with miR-106 which then regulated the MAPK9 to involve in MAPK signaling pathways. LINC00115 might interact with miR-7 to regulate FGF2 to participate in pathways in cancer.
Conclusions: MEG3, MIAT, LINC00115 may be underlying therapeutic targets for LUAD functioning as ceRNAs for regulation of miRNA-mRNA.