Metabolomics Strategy Assisted by Transcriptomics Analysis to Identify Potential Biomarkers Associated with Tuberculosis.

PURPOSE

To investigate the dysregulated pathways and identify reliable diagnostic biomarkers for tuberculosis using integrated analysis of metabolomics and transcriptomics.

METHODS

Three groups of samples, untargeted metabolomics analysis of healthy controls (HC), latent tuberculosis infection patients (LTBI), and active tuberculosis patients (TB), were analyzed using gas chromatography time-of-flight mass spectrometry (GC-TOF MS) and ultra-high performance liquid chromatography-quantitative mass spectrometry (UHPLC-QE-MS). Both univariate and multivariate and statistical analyses were used to select differential metabolites (DMs) among group comparison, and LASSO regression analysis was employed to discover potential diagnostic biomarkers. Metabolite set enrichment analysis was performed to identify the altered metabolic pathways specifically in patients with TB. Meanwhile, a transcriptomic dataset GSEG4992 was downloaded from the GEO database to explore the differentially expressed genes (DEGs) between TB and HC identified in significantly enriched pathways. Finally, an integrative analysis of DMs and DEGs was performed to investigate the possible molecular mechanisms of TB.

RESULTS

Thirty-three specific metabolites were significantly different between TB and HC, of which 7 (5-hydroxyindoleacetic acid, isoleucyl-isoleucine, heptadecanoic acid, indole acetaldehyde, 5-ethyl-2,4-dimethyloxazole, and 2-hydroxycaproic acid, unknown 71) were chosen as combinational potential biomarkers for TB. The area under the curve (AUC) value of these biomarkers was 0.97 (95% CI: 0.92-1.00). Metabolites set enrichment analysis (MSEA) displayed that there were 3 significantly enriched pathways among all. The genes in 3 significantly enriched pathways were further analyzed, of which 9(ALDH3B1, BCAT1, BCAT2, GLYAT, GOT1, IL4I1, MIF, SDS, SDSL) were expressed differentially. The area under the curve (AUC) values of these DEGs enriched in pathways mostly were greater than 0.8. As a result, a connected network of metabolites and genes in the pathways were established, which provides insights into the credibility of selected metabolites.

CONCLUSION

The newly identified metabolic biomarkers display a high potential to be developed into a promising tool for TB screening, diagnosis, and therapeutic effect monitoring.