Understanding and Treating Mycobacterium tuberculosis Infection: A Multi-Scale Modeling Approach.

Abstract

Tuberculosis (TB), caused by the pathogen Mycobacterium tuberculosis (Mtb), remains a significant burden on global health. Central to both host immune responses and antibiotic treatment are structures known as granulomas. In this dissertation we used computational and experimental approaches at a single granuloma level to understand how immune responses to Mtb contribute to both bacterial control and persistence. In addition, we predicted the dynamics of antibiotics in granulomas and designed improved treatment strategies. We built a hybrid multi-scale model of Mtb infection that integrates the cytokines tumor necrosis factor-α (TNF) and interleukin-10 (IL-10). We predicted that a balance of TNF and IL-10 is essential to infection control with minimal host-induced tissue damage. We extended our description of TNF and IL-10 to include simplified models of intracellular signaling driving macrophage polarization, which suggests that the temporal dynamics of macrophage polarization in granulomas are predictive of granuloma outcome. Next, we focused on determining the role of IL-10 in controlling antimicrobial activity. We predicted a transient role for IL-10 in controlling a trade-off between early host immunity antimicrobial responses and tissue damage. This trade-off determines sterilization of granulomas. Lastly, using an experimental model of granuloma formation, we measured significant gradients of TNF in granulomas. xxii We developed a pharmacokinetic and pharmacodynamic model of oral dosing of rifampin and isoniazid used to treat Mtb and incorporated it into our computational model. We predicted that oral antibiotic strategies fail due to sub-optimal exposure in granulomas, which leads to bacterial regrowth between doses. We extended our platform to include a description of inhaled formulations dosed to the lungs with reduced frequencies. We predicted that dosing every two-weeks with an inhaled formulation of isoniazid is feasible with increased sterilization capabilities and reduced toxicity, while an inhaled formulation of rifampin has equivalent sterilization capabilities, but early associated toxicity and infeasible carrier loadings. The keys to understanding immune responses and successful antibiotic treatment of TB lie in the dynamics at the site of infection. Our results help identify the roles of cytokines during Mtb infection, provide new possibilities for immune related therapies, and guide design of better antibiotic strategies.