A pharmacometric model of the in vitro co-dynamics of Staphylococcus aureus and three phage candidates

Kevin Royet ¹, Leslie Blazere ¹, Emilie Helluin ¹, Ludivine Coignet ¹, Lucile Plumet ², Mélanie Bonhomme ¹, Camille Kolenda ¹, Romain Garreau ¹, Florent Valour ¹, Mathieu Medina ¹, Frédéric Laurent ¹, Sylvain Goutelle ¹

1 University Hospitals Of Lyon, France (Lyon, France), 2 University of Montpellier, France (Montpellier, France)

Objectives
Phage therapy is a promising approach for treating infections caused by multidrug-resistant bacteria such as Staphylococcus aureus (S. aureus). However, phage pharmacology is peculiar, and limited information exists about the relationships between phage dose, concentration and antibacterial effect [1]. The objective of this report was to describe the building of pharmacokinetic/pharmacodynamic (PK/PD) models describing the in vitro co-dynamics of S. aureus and three phage candidates.

Methods
Four series of time-kill experiments were conducted. The effect of the three phages (V1SA019, V1SA020 or V1SA022) was evaluated on two strains of S. aureus (SH1000 and USA300). Initial concentrations of phages corresponding to multiplicity of infection (MOI) of 0.01, 1 or 100 were added to an initial inoculum of 3 log10 CFU/mL of bacteria in liquid culture. Samples were taken every 2 h during 24 h for joint quantification of bacteria and phages (phage-forming unit – PFU – per mL). All experiments were conducted in triplicates, for three different phage production batches.
We performed nonlinear mixed-effects modelling of bacteria and phage counts over time with the Monolix software (version 2024, Lixoft, Antony, France). A four compartment, predator-prey, model [2] was used to describe the interactions between four populations: susceptible bacteria (S), resistant bacteria (R), bacteria infected by phages (I) and free phage viruses (V).
Model evaluation was based on classical criteria including objective function, goodness-of-fit plots and simulation-based diagnostics. Parameters specific of phage therapy, the proliferation and inundation thresholds, were computed for each experiment through Bayesian estimation. Finally, simulations were performed based on the final model for V1SA019 / SH1000 to study the influence of phage MOI on antibacterial effect.

Results
We observed a consistent pattern of bacteria and phage counts over time with initial growth of bacteria followed by a rapid proliferation of phages, collapse of the bacterial population between 6 and 10 h, and final plateau of the phage population. In several experiments, a regrowth of bacteria was observed, suggesting emergence of bacterial resistance that was further confirmed by mutation search.
The four-equation, eight-parameter model fitted well individual bacteria and phages data. Most parameters were estimated with acceptable precision (relative standard error < 50 %, including inter-experiment variability. The phage production batch showed no influence on model parameters. A common feature between the four models was a phage binding rate of resistant bacteria (bindR) 35 to 250-fold lower than that for susceptible bacteria (bindS), resulting in higher proliferation (45 to 310-fold) and higher inundation thresholds (34 to 190-fold) for resistant compared to susceptible bacteria. Compared to the SH1000 strain, the USA300 strain displayed a 10-fold lower median burst size, 20000-fold lower median mutation rate and larger variability of the proliferation and inundation thresholds. Both experimental and simulated data indicated that the bacterial regrowth increased with initial phage concentration. However, a phage dose above the inundation threshold of resistant bacteria (10 log10 PFU/mL) would result in very fast killing of all bacteria and no regrowth, according to simulations. Conclusions Population PK/PD modeling was successful in describing and quantifying the in vitro dynamics of bacteria and phages. Substantial variability between phages and bacterial strains was observed. The models provided interesting insights on threshold parameters and emergence of a phage-resistant subpopulation that may help to design further experiments and clinical evaluations of phage therapy. References: 1. Nang et al. Clin Microbiol Infect, vol. 29, no. 6, pp. 702–709, Jun. 2023 2. Cairns, et al. PLoS Pathog, vol. 5, no. 1, p. e1000253, Jan. 2009

Reference: PAGE 34 (2026) Abstr 12059 [www.page-meeting.org/?abstract=12059]

Poster: Drug/Disease Modelling - Infection