Abstract
The antiviral activity of compounds can be enhanced synergistically when used in combination; however, the underlying mechanisms remain poorly understood. This study aimed to identify the key properties contributing to these synergistic effects. Residual titers of influenza A virus were measured after treatment with a combination of one to three compounds. In parallel, the effects of each compound on the physical properties of lipid bilayers, specifically membrane fluidity, permeability, and solubilization, were assessed. Partial least-squares regression models were constructed to predict the log reduction in residual viral titers based on the summary statistics of the membrane property changes induced by the individual compounds in each combination. These models demonstrated high predictive accuracy. Analysis of the regression coefficients revealed that combinations producing diverse membrane effects, such as (1) increased permeability, (2) decreased fluidity, (3) apparent reduction in permeability (likely due to interactions with the fluorescent probe), and (4) both increased and decreased fluidity over time (depending on whether the compounds initially induced a significant change in fluidity and whether the effect was mitigated over time), were associated with enhanced antiviral activity. Because a single compound is unlikely to produce all these effects simultaneously, combining multiple compounds may be necessary to achieve synergistic antiviral action. Furthermore, canonical correlation analyses revealed strong associations between the changes in membrane properties and the molecular structures of the compounds.