Model-based translation of the PKPD-relationship for linezolid and vancomycin on methicillin-resistant Staphylococcus aureus: from in vitro time-kill experiments to a mouse pneumonia model.

Abstract

Objectives: MRSA is one of the main pathogens that cause nosocomial pneumonia. Based on longitudinal in vitro and in vivo data, a pharmacokinetic-pharmacodynamic (PKPD) model was built to quantify the effect of two control antibiotics (LZD and VAN) for Gram-positive bacteria in a standardized mouse pneumonia model.

Methods: The PKPD model was developed for data generated on the MRSA strain 160 079 in static in vitro time-kill experiments and thereafter adjusted to fit data from lungs of neutropenic mice administered with single or multiple doses of LZD (0.5-40 mg/kg) or VAN (1-40 mg/kg). Simulations with human PK were run to predict antibacterial response in patients.

Results: Bacterial regrowth observed in vitro when exposed to VAN concentrations was described by an adaptive resistance model. The selected MRSA isolate showed good virulence in the mouse pneumonia model. Bacterial load in lungs decreased up to 2-log with respect to control mice after LZD and VAN treatment. A 70%-75% lower killing rate was estimated for the in vivo data when compared with in vitro. Simulations displayed bacterial stasis at 24 h for patients infected with bacteria with MICs below the clinical breakpoint for both drugs after administering standard-of-care dosing regimens.

Conclusions: A translational workflow allowed us to build a PKPD model with both in vitro and in vivo data that characterized bacterial dynamics following LZD and VAN exposure, showing that this approach can inform the development of antibiotics. We also showcased the first successful use of the standardized mouse pneumonia model for Gram-positive bacteria.