Study of Gene Expression Signatures for the Diagnosis of Pediatric Acute Lymphoblastic Leukemia (ALL) Through Gene Expression Array Analyses

Background: Acute lymphoblastic leukemia (ALL) as the most common malignancy in children is associated with high mortality and significant relapse. Currently, the non-invasive diagnosis of pediatric ALL is a main challenge in the early detection of patients. In the present study, a systems biology approach was used through network-based analysis to identify the key candidate genes related to ALL development and relapse. Materials and methods: In this systems biology (experimental) study, main and validating datasets were retrieved from a gene expression omnibus (GEO). Gene expression analyses were done using a bioinformatics array research tool (BART) and ExAtlas. Gene ontology and pathway enrichment analysis were also performed via Database for Annotation, Visualization and Integrated Discovery (DAVID). Furthermore, the Search Tool for the Retrieval of Interacting Genes (STRING) and cytoscape V.3.9.1 were used to network construction and analysis. The MCODE and NCMine Plugin of cytoscape were applied to find clusters and a functional module in the network. The Kaplan Myer curve was applied in order to survival analysis of the validated candidate genes. A P-value of < 0.05 was considered as significant. Results: A total of 671 differentially expressed genes (DEGs) mainly involved in transporter/channel activity functions, cell communication/signaling processes and fatty acid transport/PPAR signaling/eicosanoid metabolism pathways were identified (P-value < 0.05). The main cellular compartments were plasma membrane, cell periphery and cell surface (P-value <0.05). The network analysis revealed 68 hub genes, 29 of which were candidate genes. Five candidate genes were also validated in two independent experiments. These genes were considered as key candidate genes, and three of them (BCL2L11, IGF1, PDE5A) were predictors of pediatric ALL patients survival (P-value < 0.05).  Conclusion: BCL2L11, IGF1 and PDE5A genes, as key candidate genes, are potentially good diagnostic biomarkers and therapeutic targets for pediatric ALL.