Quantitative analysis of antiviral drug toxicity in proliferating cells.

The toxicity of most antiviral compounds was dependent on the type of cell used to assay toxicity. Ranking of compounds according to toxicity was, however, very similar (p less than 0.01) in the three different cell types used in this study. The difference in toxicity observed for 9-beta-D-arabinofuranosyladenine between Flow 5000 cells and CCRF-SB cells could not be accounted for by differences in the intracellular concentrations. On the other hand, the different toxicities observed for ribavirin and 2'-deoxy-5-trifluorothymidine between Flow 5000 cells and CCRF-SB cells may be caused by the culture conditions (as shown for one cell type, HeLa S3, grown either as monolayer or in suspension) rather than by cell-specific differences. The growth-inhibitory effect of most antiviral compounds increased with treatment time, indicating an additive nature of toxicity. The ability of cells to recover from toxic treatment with drugs varied greatly from compound to compound (from undetectable regrowth to 140% growth compared to control cells). Coaddition of natural nucleosides could, at best, only partly protect cells from the toxic influences of antiviral nucleoside analogs. As a result of comparing antiviral effects and toxicity in vitro, the unselective compounds may be eliminated from further development at the screening level.