A systematic efficacy analysis of multidrug therapies for tuberculosis using a multi-scale agent-based model

Objectives: Tuberculosis (TB) is primarily a pulmonary disease caused by the inhalation of Mycobacterium tuberculosis (Mtb). Mtb mainly infects lungs, and at infection sites, triggers formation of dense cellular structures composed of immune cells, bacteria, and dead tissue, called granulomas. The complex structure of granulomas triggers Mtb to gain adaptive responses, which allow them to better tolerate drug stress [1]. Moreover, granulomas present a physiological barrier to antibiotic diffusion, preventing effective drug delivery [2]. Due to these reasons, TB treatment is challenging, requiring a combination therapy, and TB is one of the deadliest infectious diseases in the world [3]. To decrease the death rate, recent research goals have focused on discovering more effective regimens, and many new antibiotics have been studied in experimental and clinical studies. However, testing the number of possible combinations with all potential regimens is beyond the limit of experimental resources. In this work, our objective is developing a computational modeling pipeline to efficiently predict and rank efficacies of combination regimens for TB treatment.Methods: To capture the complexities and dynamics of TB disease and to assess treatment efficacies in a systematic and efficient way, we developed GranSim, a multi-scale, agent-based, mechanistic model that simulates the immune response against Mtb infection in an area of lung tissue [4]. The interactions defined by the immunology-based rules lead to the formation of granulomas as an emergent behavior. We then incorporated a pharmacokinetics/pharmacodynamics model within GranSim to simulate the penetration of antibiotics into granulomas and to assess bacterial killing of antibiotics based on their spatial distribution [5] .Results: We provided a novel pipeline for regimen ranking based on the efficacies of regimens in simulated granulomas [6]. We compared simulation results to marmoset studies and showed a correlation between simulation rankings and marmoset data rankings providing evidence for the credibility of our framework [6].Conclusions: Our method is vital to screen various regimen combinations in a highly efficient way and to predict their efficacies. We can narrow the regimen design space for TB treatment and inform both preclinical trials and experiments so that more informed regimen decisions can be made. This unique framework can be applied to building digital twins for many complex diseases that generate heterogeneous lesions (e.g., cancer, TB) and can be used for model-informed drug development efforts that inform regulatory decisions.Citations: [1] Sarathy JP, Dartois V. Caseum: a Niche for Mycobacterium tuberculosis Drug-Tolerant Persisters. Clin Microbiol Rev. 2020;33(3). Epub 20200401. doi: 10.1128/CMR.00159-19. [2] Sarathy JP, Zuccotto F, Hsinpin H, Sandberg L, Via LE, Marriner GA, et al. Prediction of Drug Penetration in Tuberculosis Lesions. ACS Infect Dis. 2016;2(8):552-63. Epub 20160706. doi: 10.1021/acsinfecdis.6b00051.[3] Global tuberculosis report 2023. Geneva: World Health Organization; 2023.[4] Fallahi-Sichani M, El-Kebir M, Marino S, Kirschner DE, Linderman JJ. Multiscale computational modeling reveals a critical role for TNF-α receptor 1 dynamics in tuberculosis granuloma formation. J Immunol. 2011;186(6):3472-83. Epub 20110214. doi: 10.4049/jimmunol.1003299. [5] Pienaar E, Cilfone NA, Lin PL, Dartois V, Mattila JT, Butler JR, et al. A computational tool integrating host immunity with antibiotic dynamics to study tuberculosis treatment. J Theor Biol. 2015;367:166-79. Epub 20141209. doi: 10.1016/j.jtbi.2014.11.021. [6] Budak M, Via LE, Weiner DM, Barry CE, Nanda P, Michael G, et al. A systematic efficacy analysis of tuberculosis treatment with BPaL-containing regimens using a multiscale modeling approach. CPT Pharmacometrics Syst Pharmacol. 2024. Epub 20240226. doi: 10.1002/psp4.13117.