Abstract:

Objective: To identify the differentially expressed genes and pathways involved in tuberculosis onset, and to find potential biomarkers that can be used to diagnose tuberculosis using bioinformatics analysis. Methods: The series microarray dataset of GSE139825 was downloaded from the Gene Expression Omnibus (GEO) database, and the limma package of R software was applied to normalize and identify the differentially expressed genes (DEGs). Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis on DEGs were performed using clusterProfiler package. Protein-protein interaction (PPI) networks of DEGs were established with STRING online tool and core genes were visualized and screened by Cytoscape software. GSE19439 dataset was used to verify the differential expression of core genes. The enzyme-linked immunosorbent assay (ELISA) was used to validate candidate biomarkers, and area under curve (AUC) of receiver operating characteristic (ROC) was used to assess diagnosing abilities of candidate biomarkers. Results: Through analyzing GSE139825 dataset, a total of 206 DEGs were identified, including 172 upregulated genes and 34 downregulated genes. Among the downregulated genes, PDK4 and CABLES1 showed more than a 50% decrease, while IL1B, LOC728835, CXCL10, and IL8 exhibited more than an 8-fold increase. GO and KEGG pathway analyses indicated that the biological processes of the DEGs were primarily associated with cytokine-mediated signaling pathways, leukocyte intercellular adhesion, and responses to lipopolysaccharide. These DEGs predominantly exhibited molecular functions related to cytokine receptor binding and cytokine activity, and were significantly enriched in pathways such as cytokine interactions, TNF signaling, and tuberculosis-related pathways. PPI analysis identified 10 core genes, namely IL1B, TNF, IL6, IL1A, CCL20, CXCL1, CXCL10, CXCL8, CCL3, and CCR7. Further analysis using the GSE19439 validation dataset confirmed that CXCL10 and IL1B were similarly upregulated. ELISA validation also revealed significant differences in CXCL10 and IL1B expression between healthy controls and tuberculosis patients, with mean ELISA values of 0.570 and 0.827 for CXCL10, and 1.245 and 2.067 for IL1B (t=25.353, P<0.001; t=11.840, P=0.002). Logistic regression showed that CXCL10 and IL1B performed well in distinguishing the healthy group and the tuberculosis group (AUC_CXCL10=0.854, AUC_IL1B=0.818). Conclusion: Our study revealed the coordination of causal genes involved in tuberculosis onset, and indicated that CXCL10 and IL1B could serve as new potential biomarkers for the diagnosis of tuberculosis.

Key words: Tuberculosis, Genomic library, Data mining, Gene expression, Biological markers