Leveraging available data for contaminants of emerging concern to develop an understanding of environmental hazard.

Contaminants of emerging concern (CEC) are classes of compounds with relatively limited information available on environmental exposure, fate, and effects. The purpose of this research was to develop and test approaches that leverage available data using probabilistic models to advance an understanding of aquatic hazards of CECs. Pharmaceuticals are one such group of CECs. Though extrapolation approaches with fish models can provide important bridges between the biomedical and environmental sciences, little data is available addressing the sublethal effects of therapeutics in aquatic organisms. Seldom is the drug’s Mode of Action (MOA) considered in selection of chronic endpoints for an assessment, though mammalian pharmacological information is available for most drugs. A statistically significant relationship (r²=0.846, p<0.001) between mammalian pharmacology and toxicology data (margin of safety) and available fish acute to chronic data was identified, when therapeutic MOA was considered in selecting a chronic response variable. Based on this relationship, metrics to assess potency and internal effective dose were developed. These metrics were then evaluated using probabilistic distributions in an effort to prioritize drugs based on potential hazard. These probabilistic assessments identified specific drugs and drug classes as potentially presenting greater hazard to fish. To test these models, toxicity experiments with diphenhydramine, an antihistamine drug, were conducted to characterize standardized endpoints and novel, MOA-related ecotoxicological endpoints. The results confirmed that sublethal endpoints (e.g., behavior) related to therapeutic may be more appropriate for fish and that leveraging mammalian pharmacology and toxicology data may be predictive for MOA related responses when evolutionary conservation of targets are considered. It further highlighted the importance of carefully selecting model organisms for study of pharmaceuticals with multiple MOAs, because reproduction of the invertebrate Daphnia magna was sensitive to diphenydramine, potentially resulting from its histaminergic and cholinergic activities. A similar probabilistic approach was applied to oil dispersants, another CEC class, to assess potential impacts to aquatic systems. Leveraging the limited acute toxicity data available for an invertebrate and a fish model, probabilistic distributions were employed to predict the likelihood of oil dispersants exerting acute toxicity in the presence or absence of oil. This approach can be utilized in prospective and retrospective assessments to support emergency response decisions to oil spills and prioritize substances for further study. Lastly, probabilistic methods were used to develop uncertainty factors for acute to chronic rations for select biological active chemicals. For many chemical classes chronic effects data is lacking. Typically, default uncertainty factors are utilized to bridge this data gap. By leveraging the available chronic data using probabilistic methods, novel data-driven uncertainty factors were developed, potentially providing more protective extrapolation models.