Transcriptomic Characterization of Tuberculous Sputum Reveals a Host Warburg Effect and Microbial Cholesterol Catabolism

mBio. 2021 Dec 21;12(6):e0176621. doi: 10.1128/mBio.01766-21. Epub 2021 Dec 7.


The crucial transmission phase of tuberculosis (TB) relies on infectious sputum and yet cannot easily be modeled. We applied one-step RNA sequencing (RNA-Seq) to sputum from infectious TB patients to investigate the host and microbial environments underlying transmission of Mycobacterium tuberculosis. In such TB sputa, compared to non-TB controls, transcriptional upregulation of inflammatory responses, including an interferon-driven proinflammatory response and a metabolic shift toward glycolysis, was observed in the host. Among all bacterial sequences in the sputum, approximately 1.5% originated from M. tuberculosis, and its transcript abundance was lower in HIV-1-coinfected patients. Commensal bacterial abundance was reduced in the presence of M. tuberculosis infection. Direct alignment to the genomes of the predominant microbiota species also reveals differential adaptation, whereby firmicutes (e.g., streptococci) displayed a nonreplicating phenotype with reduced transcription of ribosomal proteins and reduced activities of ATP synthases, while Neisseria and Prevotella spp. were less affected. The transcriptome of sputum M. tuberculosis more closely resembled aerobic replication and shared similarity in carbon metabolism to in vitro and in vivo models with significant upregulation of genes associated with cholesterol metabolism and downstream propionate detoxification pathways. In addition, and counter to previous reports on intracellular M. tuberculosis infection in vitro, M. tuberculosis in sputum was zinc, but not iron, deprived, and the phoP loci were also significantly downregulated, suggesting that the pathogen is likely extracellular in location. IMPORTANCE Although a few studies have described the microbiome composition of TB sputa based on 16S ribosomal DNA, these studies did not compare to non-TB samples and the nature of the method does not allow any functional inference. This is the first study to apply such technology using clinical specimens and obtained functional transcriptional data on all three aspects simultaneously. We anticipate that an improved understanding on the biological interactions in the respiratory tract may also allow novel interventions, such as those involving microbiome manipulation or inhibitor targeting disease-specific metabolic pathways.

Keywords: Mycobacterium tuberculosis; RNA-Seq; Warburg effect; cholesterol; cholesterol signaling; host-pathogen interaction.

Publication types

  • Research Support, Non-U.S. Gov't