Lina Keutzer: Model-informed tuberculosis drug development and treatment optimization

Abstract

Tuberculosis (TB) is an infectious disease primarily affecting the lungs. In 2023, the World Health Organization reported a global incidence of 10.8 million cases. Despite advances in treatment, TB remains one of the leading causes of death from infectious diseases, with around 1.3 million fatalities annually. In 2022, 400,000 new cases of multidrug-resistant TB were recorded, with a cure rate of only 68%. Treatment always requires lengthy multi-drug therapy with at least four antibiotics over a minimum of six months. There is thus a great need for new shorter, less toxic and more effective treatments in order to tackle antibiotic resistance development. The overall aim of this thesis was to apply model-based approaches to improve the development and use of both new and existing drugs for TB treatment. 

Pharmacometric modelling and simulation were applied to analyse mouse relapse data, providing further evidence that the novel bedaquiline-delamanid-linezolid regimen has the potential to shorten treatment compared to the standard of care. Furthermore, an optimized study design was developed that resulted in more accurate estimates of the new drug MPL-447’s efficacy in mice. The translational value of different mouse models and the impact of lung lesions on drug efficacy were compared and revealed significant differences. 

Furthermore, modelling and simulation were applied to optimize the use of drugs in the current standard TB regimen. Model-informed precision dosing (MIPD) approaches were developed and applied to enable individualized dosing of rifampicin and linezolid. The impact of high inter-occasion variability in rifampicin exposure on the performance of MIPD was evaluated, demonstrating that, when properly accounted for, MIPD remains valuable for rifampicin dosing. In addition, a population PK model for linezolid was developed, which provided the foundation for designing a MIPD algorithm. A simulation study revealed that MIPD resulted in more patients achieving effective and safe doses compared to flat dosing. Furthermore, the relationship between linezolid exposure and toxicity was explored using pharmacokinetic-pharmacodynamic models, based on which updated linezolid safety targets were derived. Lastly, bedaquiline dosing recommendations for resuming treatment after interruption were derived to facilitate safe and effective treatment continuation.

In conclusion, this thesis contributes to accelerated preclinical TB drug development as well as optimized dosing strategies for drugs currently on the market through the application of pharmacometric approaches.