Debunking the "junk": Unraveling the role of lncRNA-miRNA-mRNA networks in fetal hemoglobin regulation

Abstract

Fetal hemoglobin (HbF) induction is considered to be a promising therapeutic strategy to ameliorate the clinical severity of {beta}-hemoglobin disorders, and has gained a significant amount of attention in recent times. Despite the enormous efforts towards the pharmacological intervention of HbF reactivation, progress has been stymied due to limited understanding of {gamma}-globin gene regulation. In this study, we intended to investigate the implications of lncRNA-associated competing endogenous RNA (ceRNA) interactions in HbF regulation. Probe repurposing strategies for extraction of lncRNA signatures and subsequent in silico analysis on publicly available datasets (GSE13284, GSE71935 and GSE7874) enabled us to identify 46 differentially expressed lncRNAs (DElncRNAs). Further, an optimum set of 11 lncRNAs that could distinguish between high HbF and normal conditions were predicted from these DElncRNAs using supervised machine learning and a stepwise selection model. The candidate lncRNAs were then linked with differentially expressed miRNAs and mRNAs to identify lncRNA-miRNA-mRNA ceRNA networks. The network revealed that 2 lncRNAs (UCA1 and ZEB1-AS1) and 4 miRNAs (hsa-miR-19b-3p,hsa-miR-3646,hsa-miR-937 and hsa-miR-548j) sequentially mediate cross-talk among different signaling pathways which provide novel insights into the lncRNA-mediated regulatory mechanisms, and thus lay the foundation of future studies to identify lncRNA-mediated therapeutic targets for HbF reactivation.