Profiling the transcriptomes of airway cells in human tuberculosis to inform strategies for enhancing bacillary clearance and preventing lung injury

Tuberculosis (TB) remains one of the leading infectious causes of death. At present, the World Health Organization (WHO) recommends treating drug susceptible pulmonary TB (pTB) with a two-month course of Rifampicin, Isoniazid, Pyrazinamide and Ethambutol (RHZE), followed by a four-month course of Rifampicin and Isoniazid (RH). Some patients are unable to adhere to this protracted treatment and discontinue treatment prematurely. This often results in TB recurrence and development of drug resistance. TB requires protracted treatment because Mycobacterium tuberculosis (Mtb), the causative agents for TB, exists as a heterogenous population of bacilli, a fraction of which are tolerant to anti-TB agents. New TB treatment regimens that are more effective against drug-tolerant Mtb would therefore help improve outcomes of TB treatment.

While more than 800,000 people are successfully cured of TB every year, all-cause mortality rates are 6 times higher in TB survivors than in the general population. This is in-part because up to half of pTB survivors sustain severe lung damage and develop post-tuberculosis lung disease (PTLD) . At present, there are no interventions for preventing or managing PTLD. PTLD is largely caused by host responses to Mtb. Anti-Mtb host responses also promote Mtb drug tolerance. Host-directed therapies could therefore help alleviate both PTLD and Mtb drug tolerance.

More than a hundred host molecular pathways, and even more genes, have been implicated in the evolution of PTLD and Mtb drug tolerance. It remains unclear which of these genes, or combinations of genes, should be targeted to reduce Mtb drug tolerance and/or PTLD. While single-gene-knockout experiments can be performed relatively easily, it is difficult to simultaneously knockout multiple genes to identify the ideal combination of genes to target to reduce Mtb drug tolerance or PTLD, given the myriad possibilities. Further, as most of the genes and pathways have been identified from animal models and in-vitro experiments, their relevance in natural human Mtb infections remains unclear. We, therefore, propose to leverage 1) single-cell transcriptomics of lung airway cells from pTB patients, 2) functional assessment of lung injury, and 3) sputum microbiologic assessment to identify the host cell types and molecular pathways associated with Mtb drug tolerance and PTLD. We will then leverage computational biology and machine learning to perform in-silico knock-up and knock-down experiments to hasten identification of single or combination host-directed therapeutics for reversing host transcriptomic perturbations associated with Mtb drug tolerance and PTLD. Finally, we will test the predicted compounds in an ex-vivo Mtb-human alveolar macrophage infection model.