Applying mixture toxicity modelling to predict bacterial bioluminescence inhibition by non-specifically acting pharmaceuticals and specifically acting antibiotics.

Pharmaceuticals and antibiotics co-occur in the aquatic environment but mixture studies to date have mainly focused on pharmaceuticals alone or antibiotics alone, although differences in mode of action may lead to different effects in mixtures. In this study we used the Bacterial Luminescence Toxicity Screen (BLT-Screen) after acute (0.5 h) and chronic (16 h) exposure to evaluate how non-specifically acting pharmaceuticals and specifically acting antibiotics act together in mixtures. Three models were applied to predict mixture toxicity including concentration addition, independent action and the two-step prediction (TSP) model, which groups similarly acting chemicals together using concentration addition, followed by independent action to combine the two groups. All non-antibiotic pharmaceuticals had similar EC values at both 0.5 and 16 h, indicating together with a QSAR (Quantitative Structure-Activity Relationship) analysis that they act as baseline toxicants. In contrast, the antibiotics' EC values decreased by up to three orders of magnitude after 16 h, which can be explained by their specific effect on bacteria. Equipotent mixtures of non-antibiotic pharmaceuticals only, antibiotics only and both non-antibiotic pharmaceuticals and antibiotics were prepared based on the single chemical results. The mixture toxicity models were all in close agreement with the experimental results, with predicted EC values within a factor of two of the experimental results. This suggests that concentration addition can be applied to bacterial assays to model the mixture effects of environmental samples containing both specifically and non-specifically acting chemicals.