Abstract
Dose addition, a commonly used concept in toxicology for the prediction of chemical mixture effects, cannot readily be applied to mixtures of partial agonists with differing maximal effects. Due to its mathematical features, effect levels that exceed the maximal effect of the least efficacious compound present in the mixture, cannot be calculated. This poses problems when dealing with mixtures likely to be encountered in realistic assessment situations where chemicals often show differing maximal effects. To overcome this limitation, we developed a pragmatic solution that extrapolates the toxic units of partial agonists to effect levels beyond their maximal efficacy. We extrapolated different additivity expectations that reflect theoretically possible extremes and validated this approach with a mixture of 21 estrogenic chemicals in the E-Screen. This assay measures the proliferation of human epithelial breast cancers. We found that the dose-response curves of the estrogenic agents exhibited widely varying shapes, slopes and maximal effects, which made it necessary to extrapolate mixture responses above 14% proliferation. Our toxic unit extrapolation approach predicted all mixture responses accurately. It extends the applicability of dose addition to combinations of agents with differing saturating effects and removes an important bottleneck that has severely hampered the use of dose addition in the past.