Evaluating SARS-CoV-2 infection under tenofovir-based antiviral prophylaxis: a multi-scale modeling analysis upon experimental data
Tenofovir has shown promising evidence of improving COVID-19 clinical outcomes in observational studies, still to be confirmed in clinical trials. Disease severity might be reduced under prophylaxis with the prodrug tenofovir disoproxil fumarate (TDF), while the protection seems to decrease, or even to lack, when using the alternative prodrug tenofovir alafenamide fumarate (TAF). Aiming to understand why TDF-prophylaxis might reduce COVID-19 severity upon infection we developed a multi-scale analysis framework combining in vitro susceptibility data, molecular docking, and within-host dynamics modeling, and using remdesivir--the only antiviral approved to date against COVID-19-- as a point of reference.First, our docking model predicted that intracellularly active tenofovir diphosphate binds into the SARS-CoV-2 RNA polymerase in the same site as the antiviral remdesivir triphosphate, but presents lower binding energy, likely reducing the overall inhibition of viral replication and making the antiviral efficacy more susceptible to the drug intracellular concentration. Second, using data from in vitro viral cultures with plausible TDF therapeutic concentrations, we estimated that the drug can inhibit SARS-COV-2 replication at an efficacy ranging between 54-99% conditional to the viral cycle length. Third, assuming values approximating this range of inhibition for in vivo viral replication during human SARS-COV-2 infection, we found that prophylaxis with TDF with high penetration into viral target cells is capable of delaying viral replication, mitigating direct cell damage and allowing time for the host to mount the adaptive immunity. Last, we found that the potential antiviral effect can be substantially reduced when TDF is given after infection begins. Our work provides a potential mechanistic explanation of the observed clinical effect of TDF against SARS-CoV-2 infection. The proposed inference framework can help to optimize the evaluation of antiviral therapies for COVID-19, in particular those targeting the RNA dependent RNA polymerase.