Integrated Population Pharmacokinetic Model of Bemnifosbuvir and Metabolites

Objectives: Bemnifosbuvir (BEM: hemisulfate salt, AT-527; free base, AT-511) is an orally administered guanosine nucleotide prodrug under clinical development for the treatment of coronavirus disease 2019 (COVID-19) and chronic hepatitis C virus (HCV). An integrated population PK model was developed to simultaneously characterize the PK of BEM free base AT-511 and its metabolites: the L-alanyl intermediate AT-551, N6-modified nucleoside metabolite AT-229, and the guanosine nucleoside metabolite AT-273 considered the plasma surrogate of the active intracellular triphosphate metabolite.Methods: Models for each analyte were first developed sequentially using rich data from healthy volunteers, starting with AT-511 and adding on the metabolites incrementally. Estimates of the fraction of the dose that reaches the systemic circulation as AT-511, AT-551, AT-229 and AT-273 due to first-pass metabolism were estimated. Food and formulation effects on the relative bioavailability and absorption of AT-511 were incorporated into the structural model. After the individual structural models were established, the models were combined into a single integrated model that simultaneously estimated model parameters for all analytes with data from COVID-19 and HCV patients added. In the integrated model, the first-pass metabolism estimates were fixed to avoid unidentifiability. Model development utilized NONMEM (v7.5.1) with FOCEI, Monte Carlo SAEM, and importance sampling methods.Results: There were 651 subjects and 41064 observations included in the analysis. The PK of each analyte was described by a 2-compartment model with linear clearance. A series of transit compartments described absorption of AT-511. Relative bioavailability was formulation- and dose- dependent up to 550 mg BEM, and differed with prandial state. Absorption mean transit time was formulation- and prandial state-dependent. Metabolism from AT-511 to AT-551 as well as from AT-551 to AT-229 and AT-273 was assumed to be dose-dependent and saturable. Between-subject variability (BSV) was applied to a subset of model parameters, including on the clearance and central volume of distribution of all analytes, intercompartmental clearance of AT-511, and the peripheral volume of distribution of AT-551. A full block BSV covariance matrix was applied. All model parameters were well estimated, and the visual predictive checks demonstrated that the integrated model was able to properly describe the observed data of all 4 analytes. Conclusions: An integrated population PK model was developed that simultaneously described the PK of BEM and three of its major metabolites in healthy volunteers and patients with HCV or COVID-19. Despite the model complexities including first-pass metabolism, dose-dependent bioavailability, and non-linear metabolism, the model satisfactorily characterized the absorption profiles associated with different formulations and prandial states and well described the observed data.Citations: None