Predictions of Bedaquiline and Pretomanid Target Attainment in Lung Lesions of Tuberculosis Patients using Translational Minimal Physiologically Based Pharmacokinetic Modeling

Abstract

Background Site-of-action concentrations for bedaquiline and pretomanid from tuberculosis patients are unavailable. The objective of this work was to predict bedaquiline and pretomanid site-of-action exposures using a translational minimal physiologically based pharmacokinetic (mPBPK) approach to understand the probability of target attainment (PTA). Methods A general translational mPBPK framework for the prediction of lung and lung lesion exposure was developed and validated using pyrazinamide site-of-action data from mice and humans. We then implemented the framework for bedaquiline and pretomanid. Simulations were conducted to predict site-of-action exposures following standard bedaquiline and pretomanid, and bedaquiline once-daily dosing. Probabilities of average concentrations within lesions and lungs greater than the minimum bactericidal concentration for non-replicating (MBC NR ) and replicating (MBC R ) bacteria were calculated. Effects of patient-specific differences on target attainment were evaluated. Results The translational modeling approach was successful in predicting pyrazinamide lung concentrations from mice to patients. We predicted that 94% and 53% of patients would attain bedaquiline average daily PK exposure within lesions (C avg -lesion) > MBC NR during the extensive phase of bedaquiline standard (2 weeks) and once-daily (8 weeks) dosing, respectively. Less than 5% of patients were predicted to achieve C avg -lesion > MBC NR during the continuation phase of bedaquiline or pretomanid treatment, and more than 80% of patients were predicted to achieve C avg -lung >MBC R for all simulated dosing regimens of bedaquiline and pretomanid. Conclusions The translational mPBPK model predicted that the standard bedaquiline continuation phase and standard pretomanid dosing may not achieve optimal exposures to eradicate non-replicating bacteria in most patients. Supplementary Information The online version contains supplementary material available at 10.1007/s40262-023-01217-7.

Model‐based dose optimization framework for bedaquiline, pretomanid and linezolid for the treatment of drug‐resistant tuberculosis

“…33 Based on our translational mPBPK model, plasma, lungs and lesion M2-bedaquiline exposure ratios were predicted to be 0.18, 1.11 and 1.02, respectively. 14 These results provide an overview of the relative role of M2 as compared to bedaquiline on Mtb clearance.…”

“…12,[22][23][24] PK models and parameter estimates for bedaquiline and pretomanid from our prior work were used to simulate plasma and site of action, lungs and lesions, and concentration-time profiles. 14 A PK model for linezolid was reproduced from the literature to simulate plasma, lungs and lesion concentration-time profiles. 25 Body weights were sampled for the virtual patients from observed TB patients' body weight distribution from the data.…”

“…Bedaquiline and pretomanid pulmonary drug concentration data from TB patients are not available to date; therefore, lung and lesion penetration of these two drugs were obtained using translational mPBPK models. 14 Additionally, lung tissue content and drug physicochemical properties affect the fraction of drug available at target sites to exert the anti-Mtb effect. Anti-bacterial activity data following monotherapy agreed well with the predictions for bedaquiline and linezolid by accounting for unbound drug fractions in lungs that were measured experimentally.…”

“…Thus, predictions of BPaL treat- To date, quantitative pharmacology approaches have exploited some relationships between patient-and disease-related covariates, PK and response for bedaquiline, pretomanid and linezolid individually. [11][12][13][14] These models, however, did not include several key mechanistic components, such as Mtb susceptibility, target site drug exposures and PD interactions between the BPaL combination regimen. In this work, we aimed to combine the relevant mechanistic components to develop a quantitative framework for BPaL…”

“…14,25 a kG = 1.52 day À1 and B max = 6.5e + 07 mL À1 were estimated by fitting the multistate tuberculosis model to combined log-phase Mtb growth data in absence of bedaquiline.…”