Model-informed candidate selection using PK/PD predictions in anti-SARS-CoV-2 drug develpment

REGA, PATRICIA; IBARRA, MANUEL; ALAN TALEVI; COMINI, MARCELO

Simposio; Brazilian Symposium on Medicinal Chemistry; 2023

Sociedad Brasileña de Química

Technological advances have increased the likelihood of finding successful candidates in drugdiscovery and development. High-throughput methodologies have accelerated in-vitro activityevaluation. In addition, the implementation of rational experimental design approachesconsidering pharmacokinetic (PK) and pharmacodynamic (PD) principles from thepreclinical stage onwards is essential to develop feasible dosage schemes maximized to attaintargeted drug exposure and avoid an inefficient evaluation. Pharmacometrics informs decisions indrug development by integrating multiple-level data to characterize the dose-exposure-responserelationship. This work aimed to support the selection of compounds with proven in-vitroactivity against SARS-CoV-2 virus and the experimental design of a preclinical efficacy studyin an infected mice model using modeling and simulation tools.Results and DiscussionTwenty-five compounds were analyzed in-silico using ADMETpredictor®, PK® and SwissADME®software to predict their physicochemical, PK and toxicological properties. Outputs obtained werecompared and selected based on their reported suitability, and integrated in a physiologically-based PK (PBPK) mice model using the software PK-Sim®. Simulations were conducted to define adosage regime that achieves exposure to unbound drug in pulmonary tissue above the in-vitro estimated concentration linked with a 90% reduction in baseline viral replication (IC90) for atleast 85% of the inter-dose interval for 5 days, with a probability of target attainment of 90%.Predicted elimination half-life and lung tissue-to-plasma partition ratio affected dosing requirements and administration intervals. Five compounds achieved target concentrations with an experimentally suitable dosage regime and were selected to continue with preclinical studies in experimental animals. To enhance the reliability of PBPK predictions, model validation is essential using in-vivo preclinical data. This learning and confirmation cycle allows model optimization and facilitates future assessments in the drug development process.PK modeling and simulation integrating predictions from chemical structures into PBPK models provided an important input for candidate selection and optimal design before preclinical evaluation.