Identification of a Diagnostic Multiomics-Based Biomarker Cluster in a Mouse Model of Pulmonary Tuberculosis.

Abstract

Tuberculosis (TB) stands as the second most prevalent infectious agent-related cause of death worldwide in 2022, trailing only COVID-19. With 1.13 million reported deaths, this figure is more than half of the mortality associated with human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS), which accounted for 0.63 million deaths. Diagnosing Mycobacterium tuberculosis (MTB) infection remains a formidable challenge due to the inability to isolate and detect MTB in sputum and within the human body. The absence of universally reliable diagnostic criteria for MTB infection globally poses a significant obstacle to preventing the progression of tuberculosis from the MTB infection stage. In this study, our objective was to formulate a diagnostic biomarker cluster capable of discerning the progression of MTB infection and disease. This was achieved through a comprehensive joint multiomics analysis, encompassing transcriptome, proteome, and metabolome, conducted on lung tissue samples obtained from both normal control mice and those infected with MTB. A total of 1690 differentially expressed genes and 94 differentially expressed proteins were systematically screened. From this pool, 10 core genes were singled out. Additionally, eight long non-coding ribonucleic acids and eight metabolites linked to these core genes were identified to establish a cohesive cluster of biomarkers. This multiomics-based biomarker cluster demonstrated its capability to differentiate uninfected samples from MTB-infected samples effectively in both principle component analysis and the construction of a random forest model. The outcomes of our study strongly suggest that the multiomics-based biomarker cluster holds significant potential for enhancing the diagnosis of MTB infection.