Umberto Villani1, Chenyao Liu1, Eik Hoffman2, Cyril Gaudin2, Michael Dal Molin3, Peter Velickovic1, Salvatore D'agate1, MarÃa Cristina Villellas Arilla4, Alain Baulard2, Oscar Della Pasqua1
1Istituto per le Applicazioni del Calcolo, Consiglio Nazionale delle Ricerche (CNR), 2Center for Infection and Immunity of Lille (CIIL), Institute Pasteur Lille, 3Faculty of Medicine, Center for Molecular Medicine Cologne (CMMC), University of Cologne, 4Department of Microbiology, Faculty of Medicine, University of Zaragoza
Objectives: The ERA4TB consortium [1] has established a robust in vitro benchmarking platform that allows the detailed characterisation of compounds that modulate host-pathogen interactions (HPI) in tuberculosis (TB). Furthermore, the approach discriminates between enhanced host-directed activity and drug effects that alter bacterial virulence. The platform is supported by translational efforts within ERA4TB, including PKPD principles, modelling and simulation [2]. This provides the basis for data integration, extrapolation and prediction of treatment performance in vivo, allowing for the prioritisation and progression of candidate molecules and combinations that are likely to be efficacious in the clinic. Accordingly, in this study, we present a model-based analysis to quantify the interaction between HPI compounds and antibiotics commonly used for the treatment of TB. As a case study, we examine the interaction between bedaquiline (BDQ), a key antimicrobial in multidrug-resistant TB regimens, and imatinib (IMTB), a tyrosine kinase inhibitor used in the treatment of myelogenous leukaemia that is hypothesized to enhance phagocytosis in human macrophages and to regulate immune responses to mycobacterial infections [3]. Methods: Data from confocal microscopy-based time-kill assays performed using Mycobacterium tuberculosis (Mtb) H37Rv were available for this analysis. Intracellular bacterial growth and killing dynamics were assessed over 10 days in an infected THP-1 human macrophage model [4]. Including natural bacterial growth, a total of n=12 conditions were tested. Drug concentrations for BDQ and IMTB selected for these experimental protocols were based on a previous analysis of monotherapy data in which the concentration-effect relationship was characterised [5,6]. The combination experiment included three static BDQ concentrations (corresponding to EC20, EC50, and EC80) and two IMTB concentrations (EC20 and EC80). Modelling of the pharmacodynamic interaction between the active moieties was implemented by fixing the estimates (EC50 and Emax) describing the concentration effect relationships of IMTB and BDQ as monotherapy. Subsequently, the overall effect of the combination was modelled as the sum of the effect of each drug on intracellular bacterial growth, with a multiplicative, dimensionless interaction factor (ßint) describing deviations from additivity relative to BDQ. Given the concentration-dependent changes in the observed HPI-antibiotic interaction, ßint was modelled as a polynomial function with BDQ and IMTB concentrations as independent variable. Model selection criteria included goodness-of-fit diagnostics, changes in the objective function value (OFV), and precision of parameter estimates. Parameter estimation was performed using NONMEM 7.5 and PsN 5.2.6, while graphical analyses were conducted in R version 4.1.2. Results: The joint use of BDQ and IMTB resulted in an increased clearance of intracellular bacteria in most tested conditions, as compared to the effect of monotherapy with either compound. Given the lack of a mechanistic insight to explain the enhanced BDQ antibacterial activity of macrophages following exposure to IMTB, an empirical parameterisation was used to describe the observed interactions. More specifically, a polynomial function featuring a quadratic term for BDQ concentrations, linear terms for BDQ and IMTB concentrations, and a linear interaction term between BDQ and IMTB concentrations was found to adequately describe the ßint bidimensional interaction landscape. Notably, when IMTB at EC20 was combined with BDQ at EC20, the interaction was greater than what would be expected from additivity (ßint = 1.83), indicating that BDQ activity was enhanced by the presence of IMTB. However, as BDQ and IMTB concentrations increased, the contribution of IMTB to overall bacterial killing decreased, ultimately leading to an antagonistic interaction when both drugs were tested in combination at their EC80 level (ßint = -0.133). Moreover, the combination of IMTB at EC20 to BDQ at EC50 resulted in an interaction featuring the highest overall killing rate observed in the experiment, comparable to BDQ at its EC80 level as monotherapy (ßint = 1.22; E_combo = 1.19 [1/days]; E_BDQ_EC80 =1.13 [1/days]). Conclusions: Novel strategies are required to treat M. tuberculosis infections. The use of drug combinations that enhance immune competence or alter bacterial virulence represent an effective way to overcome evolving antibiotic resistance. In addition to obtaining estimates of the contribution of the different moieties to the overall antibacterial activity, the use of modelling and simulation concepts offers a unique opportunity for the translation of experimental findings, taking into account drug disposition and disease processes in vivo. As such, the proposed modelling framework will allow ranking of promising compounds for subsequent prioritisation and evaluation in in vivo studies with multidrug antimicrobial regimens.
[1] The ERA4TB consortium. https://era4tb.org/. [2] Muliaditan M, Davies GR, Simonsson USH, Gillespie SH, Della Pasqua O. The implications of model-informed drug discovery and development for tuberculosis. Drug Discov Today. 2017; 22(3):481-48 [3] Cleverley TL, Peddineni S, Guarner J, et al. The host-directed therapeutic imatinib mesylate accelerates immune responses to Mycobacterium marinum infection and limits pathology associated with granulomas. PLoS Pathog. 2023;19(5). [4] Chanput W, Mes JJ, Wichers HJ. THP-1 cell line: An in vitro cell model for immune modulation approach. Int Immunopharmacol. 2014;23(1):37-45. [5] Muliaditan M, Della Pasqua O. Bacterial growth dynamics and pharmacokinetic-pharmacodynamic relationships of rifampicin and bedaquiline in BALB/c mice. Br J Pharmacol. 2022; 179(6):1251-1263. [6] Villani U, Leding AAM, Velickovic P, et al. Model-based characterisation of host-directed therapeutics against M. tuberculosis in an in vitro experimental model.
Reference: PAGE 33 (2025) Abstr 11733 [www.page-meeting.org/?abstract=11733]
Poster: Drug/Disease Modelling - Infection