Objectives This study aimed to investigate the risk miRNAs (microRNAs) for AML (acute myeloid leukemia) prognosis and related regulatory mechanisms. Methods MiRNA and gene expression data, as well as clinical data of 176 patients were first downloaded from TCGA. Then miRNAs and genes significantly affecting the survival time based on KM survival curve were identified using Log Rank test. Next, COX proportional-hazard regression analysis was performed to screen the risk miRNAs (P-value < 0.05). Common genes from survival analysis and predicted by miRWalk were used to construct the miRNA regulatory network with the risk miRNAs. Finally, a protein-protein interaction (PPI) network was constructed, as well as functional annotation and pathway enrichment analysis. Results The survival analysis revealed 33 miRNAs and 1,377 genes significantly affecting the survival time. HR values of nine miRNAs (up-regulated hsa-mir-606, 520a, 137, 362, 599, 600, 202, 639and down-regulated 502) were either >1 or <1. The miRNA regulatory network contained 477 nodes and 944 edges. The top ten genes of the constructed PPI network were EGFR, EIF4G1, REL, TOP1, COL14A1, HDAC3, MRPL49, PSMA2, TOP2A and VCAM1 successively. According to pathway enrichment analysis, 6 KEGG pathways and 6 REACTOME pathways were obtained respectively. Conclusion Up-regulated hsa-mir-520a, 599, 606, 137 and 362 may increase the prognostic risk for AML patients via regulating the expression of corresponding target genes, especially COL14A1, HDAC3, REL, EGFR, PSMA2, EIF4G1, MRPL49 and TOP1.
Keywords: Acute myeloid leukemia; MicroRNA; expression regulation; prognostic risk.