Translational pharmacokinetic-pharmacodynamic (PKPD) modelling of apramycin to facilitate prediction of efficacious dose in urinary tract infections
Irene Hernández-Lozano (1), Vincent Aranzana-Climent (1), Jon Ulf Hansen (2), Sha Cao (3), Edgars Liepinsh (4), Diarmaid Hughes (3), Carina Vingsbo Lundberg (2), Sven Hobbie (5), Lena E. Friberg (1)
(1) Department of Pharmacy, Uppsala University, Sweden, (2) Department of Bacteria, Parasites & Fungi, Statens Serum Institut, Copenhagen, Denmark, (3) Department of Medical Biochemistry and Microbiology, Uppsala University, Sweden, (4) Latvian Institute of Organic Synthesis, Riga, Latvia, (5) Institute of Medical Microbiology, University of Zurich, Switzerland
Objectives: Apramycin (EBL-1003) is a broad-spectrum aminoglycoside antibiotic for the treatment of gram-negative bacteria infections that is currently under development for human use. Although apramycin has demonstrated best-in-class coverage of resistant isolates and efficacy in preclinical lung infection models, its utility in other disease indications needs to be further evaluated. Previous proof-of-concept studies have suggested that apramycin has potential in the treatment of complicated urinary tract infection (cUTI) and acute pyelonephritis [1]. In this study, we used pharmacokinetic-pharmacodynamic (PKPD) modelling based on in vitro and in vivo data in order to predict apramycin efficacy in different mouse models of cUTI.
Methods: Two different Escherichia coli strains were studied (EN591 and EN1085). Exposure-effect relationships were evaluated with in vitro time-kill experiment data under different pH conditions in order to assess possible differences in bacterial growth derived by the more acidic nature of urine (pH6) as compared to blood or plasma (pH7.4). In vivo PK of apramycin were studied in plasma in both healthy and infected mice. In vivo efficacy was measured in two related but not identical cUTI mouse models. A previously developed one-compartment PK model describing in vivo subcutaneous absorption and systemic distribution of apramycin in infected mice was used in order to simulate the unbound plasma PK of apramycin in infected mice using a clearance value of 8.5 L/h/70kg, a volume of distribution of 5.42 L/70kg and an absorption rate constant of 2.02 h-1. A PKPD model with two subpopulations (i.e. susceptible and resistant subpopulations) was developed from the rich CFU (colony forming units) in vitro data on the two E. coli strains. The sparse in vivo PK and efficacy (PD) data on the same strains were integrated into the model and simulations were performed.
Results: The PKPD models were fitted to the in vitro time-kill data of apramycin in different E. coli strains under the established pH conditions leading to rate constants for bacterial growth of 2.0 h-1 and 2.7 h-1 for EN591 and EN1085, respectively. Bacterial growth and drug effect parameters differed based on pH to a similar magnitude as determined by MIC under different pH. The in vivo plasma PK model was integrated with the in vitro PKPD model to predict the effect of apramycin in both kidney and urine. Re-estimation of the rate constant for bacterial growth suggested that the bacterial growth rate in the cUTI mouse models was up to 87% lower than in vitro. After model refinement and parameter re-estimation, apramycin effect simulations were comparable to the measured effect of apramycin in kidney and bladder.
Conclusions: With PKPD modelling data from different sources, namely, plasma PK as well as rich in vitro and sparse in vivo CFU data in kidney and bladder, could be integrated. The PKPD model could successfully be applied to perform predictions of expected effect in mouse models of cUTI, which were, up to a certain extent, comparable to the measured effect of apramycin in tissue. Re-estimation of the parameters suggested that bacterial growth is slower in vivo, which is in agreement with the lower bacterial growth rate estimated in a mouse thigh infection model as compared to the in vitro growth rate of the same bacterial strain [2]. However, further dynamic PK and PD data both in kidney and bladder would be desirable for a better understanding and prediction on the tissue effect of apramycin in cUTIs. Altogether, this study holds promise to enable dosing recommendations for future clinical trials in patients with cUTI and to potentially taking a step forward on the development of apramycin for human use.
References:
[1] Becker K et al. EBioMedicine (2021) 73, 103652
[2] Sou T et al. Clin Pharm Ther (2021) 109(4),1063-1073
