Tissue Level Profiling of SARS-CoV-2 antivirals in mice to predict their effects: comparing Remdesivir's active metabolite GS-441 524 vs. the clinically failed Hydroxychloroquine

Background and ObjectivesRemdesivir and hydroxychloroquine are or were among the most promising therapeutic options to tackle the current SARS-CoV-2 pandemic. Besides the use of the prodrug remdesivir itself, the direct administration of GS-441 524, the resulting main metabolite of remdesivir, could be advantageous and even more effective. All substances were not originally developed for the treatment of COVID-19 and especially for GS-441 524 little is known about its pharmacokinetic and physical-chemical properties. To justify the application of new or repurposed drugs in humans, pre-clinical in vivo animal models are mandatory to investigate relevant PK and PD properties and their relationship to each other. In this study, an adapted mouse model was chosen to demonstrate its suitability to provide sufficient information on the model substances GS-441 524 and HCQ regarding plasma concentration and distribution into relevant tissues a prerequisite for treatment effectiveness. MethodsGS-441 524 and HCQ were administered intravenously as a single injection to male mice. Blood and organ samples were taken at several time points and drug concentrations were quantified in plasma and tissue homogenates by two liquid chromatography/tandem mass spectrometry methods. In vitro experiments were conducted to investigate the degradation of remdesivir in human plasma and blood. All pharmacokinetic analyses were performed with R Studio using non-compartmental analysis. ResultsHigh tissue to plasma ratios for GS-441 524 and HCQ were found, indicating a significant distribution into the examined tissue, except for the central nervous system and fat. For GS-441 524, measured tissue concentrations exceeded the reported in vitro EC50 values by more than 10-fold and in consideration of its high efficacy against feline infectious peritonitis, GS-441 524 could indeed be effective against SARS-CoV-2 in vivo. For HCQ, relatively high in vitro EC50 values are reported, which were not reached in all tissues. Facing its slow tissue distribution, HCQ might not lead to sufficient tissue saturation for a reliable antiviral effect. ConclusionThe mouse model was able to characterise the PK and tissue distribution of both model substances and is a suitable tool to investigate early drug candidates against SARS-CoV-2. Furthermore, we could demonstrate a high tissue distribution of GS-441 524 even if not administered as the prodrug remdesivir.