Identification and characterization of lncRNA mediated transcriptional dysregulation dictates lncRNA roles in glioblastoma

Abstract

Long non-coding RNAs (lncRNAs) modulate gene expression, and lncRNA misregulation is associated with cancer. However, precise functional roles in biological and disease processes have been described for only a few lncRNAs. Identification of genome-wide lncRNA-mediated transcriptional dysregulations may improve cancer treatments. In the present study, we used a computational framework that combined lncRNA and gene expression profiles with transcription factor (TF)-target relationships to comprehensively identify dysregulatory lncRNA-TF-gene triplets. In glioblastoma (GBM), we found that most lncRNAs affect multiple targets and primarily affect TF activity in trans. Six different classes of lncRNA-mediated transcriptional dysregulations were identified, with most lncRNAs either enhancing or attenuating target gene expression. Functional analysis of lncRNAs via their dysregulated targets implicated lncRNA modulators in some hallmarks of cancer, providing a new way to predict lncRNA function. Finally, we identified several lncRNA-TF-gene triplets (including HOTAIR-MXI1-CD58/PRKCE and HOTAIR-ATF5-NCAM1) that are associated with glioblastoma prognosis. The integration of lncRNA modulators into transcriptional regulatory networks will further enhance our understanding of lncRNA functions in cancer.