Antiviral dose response against CMV infection varies with the timing of treatment initiation

25 Background: Cytomegalovirus (CMV) is a herpesvirus that can cause significant mortality in patients who are immunocompromised. Optimizing dosing strategy of antiviral drugs is still crucial due to their narrow therapeutic windows and limited studies in PK/PD targets. While these drugs are all active against CMV activity, the extent to which the initiation time of antiviral administration affects pharmacodynamic outcomes is still unknown. Our study aims to evaluate the antiviral effects of ganciclovir, foscarnet and letermovir under varying treatment initiation times in an in vitro infection model. Methods: The in vitro infection model consisted of human lung fibroblast cells (MRC5) that were infected with the HCMV Towne strain (MOI = 0.1). Ganciclovir, foscarnet and letermovir concentrations of 0.5–10 mg/L, 7.5–60 mg/L, and 0.5–34 μg/L, respectively, were used to assess antiviral effects as these are relevant to the reported range of the half-effective maximal concentration (EC50) values. Ganciclovir and foscarnet treatment was initiated at 0, 3, 6 and 10 days post-infection, while Letermovir was initiated at 3 days before infection, 0, 3 and 6 days post-infection. Supernatant was collected to measure viral load using qPCR. Log10 genome copies reduction was calculated relative to non-drug treated samples. Results: Throughout the 15-day experiment (infection period), we observed a dose–response effect on viral suppression when treatment was delayed until 6 days post-infection for ganciclovir and until 3 days post-infection for foscarnet. For ganciclovir, earlier treatment induces higher viral load reduction proportionally. At 10 mg/L, we have found 3.8, 2.4 and 1.8 log10 reduction when treatment was initiated without delay, at 3 and 6 days post-infection, respectively. While for foscarnet, a reduction of 2 log10 was observed at the highest concentration when treatment was initiated without delay and at 3 days post-infection. Conversely, a dose–response effect on antiviral activity was only observed in the absence of delayed letermovir treatment where more than 1 log10 viral reduction was observed. Consistent with our findings, we have found that EC50 values increased with delayed treatment of the three drugs. This indicates that anti-CMV potency depends on treatment initiation time as well as the infection age of the cell culture. Overall, we have observed that dosing delays affect pharmacodynamic outcomes differently for different drugs. We have found that ganciclovir has the most potent effect in suppressing viral load burden among other drugs when treatment is initiated with delay. Conclusion: Our in vitro data suggest that the dose–response effect on antiviral activity for anti-CMV drugs depends on treatment initiation time. Delayed treatment is consistently found associated with higher EC50 values. These findings suggest that potency decreases with infection age. Additionally, the impact of dosing delays on pharmacodynamic outcomes varied between different drugs. The PK/PD of these drugs is thus different and will be further investigated in a dynamic hollow fibre infection model. In conclusion, our overall findings suggest that timing of drug administration is crucial for improving clinical outcomes in CMV treatment.