A minimal PBPK mice model to support the preclinical development of drugs against tuberculosis.
Federico Reali (1), Anna Fochesato (1,2), Roberto Visintainer (1), Shayne Watson (3), Micha Levi (3), Véronique Dartois (4), Luca Marchetti (1,5)
(1) Fondazione The Microsoft Research – University of Trento Centre for Computational and Systems Biology (COSBI), Italy, (2) University of Trento, Department of Mathematics, Italy, (3) Gates Medical Research Institute, USA, (4) Hackensack Meridian Health, USA, (5) University of Trento, Department of Cellular, Computational and Integrative Biology (CIBIO), Italy
Objectives: Tuberculosis is an infectious disease that every year infects and kills millions of people worldwide [1]. The standard of care, a combination of four antibiotics, may suffer from poor compliance, due to the length of the treatment and the emergence of drug and multi-drug resistant strains. This calls for shorter anti-TB regimens, while meeting adequate safety and efficacy criteria, to increase adherence and favorable treatment rate. In recent years, fueled by several TB initiatives, new compounds have entered different development phases and worldwide-sponsored clinical trials. At the same time, minimal-PBPK models have recently gained attention in pharmacometrics as powerful tools to bridge top-down and bottom-up approaches and readily reconstruct pharmacokinetics (PK) in plasma and the site of action [2].
We propose a data-driven minimal PBPK model (mPBPK) in mice that can support several anti-TB drugs at once by focusing on pivotal compartments for absorption, distribution, metabolism, and drug site of action, the pulmonary compartment. The work seeks to deliver a unified viewpoint and standardized platform to compare candidates on desirable preclinical plasma and lung PK metrics.
Methods: We iteratively lumped compartments that are marginally involved in TB infection to define a minimal model diagram. As a result of an extensive validation step, adipose, brain, bone, heart, muscles, pancreas, and skin were summarized in a lumped compartment. A compartment describing lung was maintained to describe PK dynamics in the site of action. We considered first-order reactions for absorption and elimination processes and for drug exchange among compartments. The model was implemented in Matlab R2019b and R 3.6.1 and simulated via Matlab ode15s and R deSolve package. A structural identifiability investigation was run on the Matlab toolbox GenSSI and the model was calibrated using the Covariance Matrix Adaptation – Evolution Strategy (CMA-ES) method [3]. The uncertainty on the parameters and the outputs were estimated through Monte Carlo sampling of 1500 parameters with a coefficient of variation of 20%.
Results: We have calibrated and tested our mPBPK model, which is structurally identifiable, using literature mouse PK datasets of ten TB drugs, including marketed and in clinical phase compounds, such as isoniazid (H), pyrazinamide (Z), ethambutol, moxifloxacin (M), rifampicin (R), rifapentine (P), delamanid, pretomanid (Pa), Bedaquiline (B), and a novel carbostyril derivative. We implemented the model in an interactive R-based web app that readily provides simulations and PK metrics for customized doses and treatment length. We analyzed the design of preclinical experiments for novel anti-TB regimens, such as BPaMZ and HPMZ [4, 5], by studying the attainment of several pharmacodynamic targets in plasma and lungs.
Conclusions: The developed mPBPK approach requires a limited amount of prior knowledge to generate accurate preclinical PK predictions for several compounds and allows a significant acceleration compared to a full PBPK model. The use of our mPBPK platform sheds light on the importance of site of disease PK to prioritize novel anti-TB regimens that can optimally achieve PK/PD targets in the lung. Our mPBPK platform delivers a flexible fit-for-purpose tool for model-informed drug design pipelines to speed up the development of novel anti-tuberculotic agents from bench to clinic.
References:
[1] World Health Organization (2022), Global tuberculosis report 2022
[2] Mehta et al., Predictions of Bedaquiline and Pretomanid Target Attainment in Lung Lesions of Tuberculosis Patients using Translational Minimal Physiologically Based Pharmacokinetic Modeling, Clinical Pharmacokinetics, 2023
[3] Hansen Nikolaus, The CMA Evolution Strategy: A Comparing Review; in Lozano et al., Towards a New Evolutionary Computation. Studies in Fuzziness and Soft Computing, Springer, 2006
[4] Tweed et al., Bedaquiline, moxifloxacin, pretomanid, and pyrazinamide during the first 8 weeks of treatment of patients with drug-susceptible or drug-resistant pulmonary tuberculosis: a multicentre, open-label, partially randomised, phase 2b trial. The Lancet. Respiratory medicine, 7(12), 2019.
[5] CDC (2022), Interim Guidance: 4-Month Rifapentine-Moxifloxacin Regimen for the Treatment of Drug-Susceptible Pulmonary Tuberculosis.