A PKPD model of favipiravir and ribavirin in Lassa virus infected mice
Christine J. Kleist (1), Lisa Oestereich (2, 3), Stephan Günther (2, 3), Sebastian G. Wicha (1)
(1) Dept. of Clinical Pharmacy, Institute of Pharmacy, University of Hamburg, Germany, (2) Dept. of Virology, Bernhard Nocht Institute for Tropical Medicine, Hamburg, Germany, (3) German Centre for Infection Research (DZIF), Partner site Hamburg-Lu¨beck-Borstel-Riems, Germany
Objectives: Favipiravir and ribavirin are broad-spectrum antivirals, which showed inhibitory effects against the replication of the Lassa virus in vitro. While the effect of ribavirin in vivo was minor, favipiravir increases the survival rate in Lassa virus infected mice [1]. Carrillo-Bustamante et al. investigated different mechanisms of action of favipiravir and ribavirin. Favipiravir was assumed to decrease the viral production rate. In opposite no direct antiviral effect was detected for ribavirin, instead a cell-protective effect was suggested. This relationship was indicated by a decrease of aspartate aminotransferase (AST) levels [2]. Yet, no pharmacokinetic (PK) was considered in [2] and solely the ED50 for both drugs was estimated. The aim of this project was to link literature-based PK models to the pharmacodynamic (PD) data from [1], i.e. to develop a PKPD model for favipiravir and ribavirin in mice in order to determine an in vivo EC50 of favipiravir and ribavirin in Lassa infected mice.
Methods: The PK model for favipiravir in mice of Madelain et al. [3] and the PK model for ribavirin in mice of Endres et al. [4] were combined with the PD model with the most likely mechanism of action (model C) of Carrillo-Bustamante et al. [2]. Model C describes the cell protective effect of ribavirin by reducing the death rate of the infected cells and the decrease of the viral production rate of favipiravir. The parameters of the favipiravir and ribavirin PK models were fixed to the pharmacokinetic parameter values described in the literature [3-4], as no plasma concentrations were available from the infected mice [1]. For ribavirin the estimates for the wildtype (Ent1(+/+), wildtype for equilibrative nucleoside transporter 1) were chosen [4].
The values of PD model, basic reproductive number R0, death rate of infected cells δI, viral production rate p, factor α describing AST release from dying infected cells, constant source for AST sx, were re-evaluated. While the initial target cell inoculum, the initial virus inoculum, the clearance of the free virus and of the AST molecules were fixed to the literature values, as in the original PD model [2]. The difference in the objective function value, as well as graphical criteria, i.e. goodness of fit plots and visual predictive checks guides the model selection. Viral load data below the lower limit of quantification was handled with the M3 method.
Results: The EC50 of favipiravir and ribavirin were estimated simultaneously from monotherapy data for each compound with all other model parameters fixed. Goodness of fit plots indicated a good model fit, while VPC showed a minor fit for AST, as well as for the viral load. The re-estimation of the AST related parameters (α, sx) improved the VPC for AST with a drop of the objective function value of -33.135. The re-estimation of rate p and rate p in combination with α and sx led to a significant drop in the objective function value of -5.599 and -34.463, respectively. Nevertheless, there was no substantial improvement in VPC and goodness of fit plots.
For the final PKPD model the estimates were α = 0.00306 (literature: 0.0018), sx = 73.4 U/L (literature: 66.8 U/L). The estimated EC50 for favipiravir was 1.51 mg/L and for ribavirin the EC50 was 0.712 mg/L.
Conclusions: An in vivo EC50 of favipiravir and ribavirin was successfully estimated. However, the model indicated a regrowth of the viral load after the end of therapy, not allowing the model to predict an end of the viral infection and requiring further refinement of the model.
References:
[1] Oestereich L, Rieger T, Lüdtke A, et al (2016) Efficacy of Favipiravir Alone and in Combination With Ribavirin in a Lethal, Immunocompetent Mouse Model of Lassa Fever. J Infect Dis 213:934–938. https://doi.org/10.1093/infdis/jiv522
[2] Carrillo-Bustamante P, Nguyen THT, Oestereich L, et al (2017) Determining Ribavirin’s mechanism of action against Lassa virus infection. Sci Rep 7:11693. https://doi.org/10.1038/s41598-017-10198-0
[3] Madelain V, Oestereich L, Graw F, et al (2015) Ebola virus dynamics in mice treated with favipiravir. Antiviral Res 123:70–77. https://doi.org/10.1016/j.antiviral.2015.08.015
[4] Endres CJ, Moss AM, Govindarajan R, et al (2009) The role of nucleoside transporters in the erythrocyte disposition and oral absorption of ribavirin in the wild-type and equilibrative nucleoside transporter 1(-/-) mice. J Pharmacol Exp Ther 331:287–296. https://doi.org/10.1124/jpet.109.153130