It is not feasible to conduct toxicity tests with all species that may be impacted by chemical exposures. Therefore, cross-species extrapolation is fundamental to environmental risk assessment. Recognition of the impracticality of generating empirical, whole organism, toxicity data for the extensive universe of chemicals in commerce has been an impetus driving the field of predictive toxicology. We describe a strategy that leverages expanding databases of molecular sequence information together with identification of specific molecular chemical targets whose perturbation can lead to adverse outcomes to support predictive species extrapolation. This approach can be used to predict which species may be more (or less) susceptible to effects following exposure to chemicals with known modes of action (e.g., pharmaceuticals, pesticides). Primary amino acid sequence alignments are combined with more detailed analyses of conserved functional domains to derive the predictions. This methodology employs bioinformatic approaches to automate, collate, and calculate quantitative metrics associated with cross-species sequence similarity of key molecular initiating events (MIEs). Case examples focused on the actions of (a) 17α-ethinyl estradiol on the human (Homo sapiens) estrogen receptor; (b) permethrin on the mosquito (Aedes aegypti) voltage-gated para-like sodium channel; and (c) 17β-trenbolone on the bovine (Bos taurus) androgen receptor are presented to demonstrate the potential predictive utility of this species extrapolation strategy. The examples compare empirical toxicity data to cross-species predictions of intrinsic susceptibility based on analyses of sequence similarity relevant to the MIEs of defined adverse outcome pathways. Through further refinement, and definition of appropriate domains of applicability, we envision practical and routine utility for the molecular target similarity-based predictive method in chemical risk assessment, particularly where testing resources are limited.
- Conserved functional domains
- Molecular target
- Predictive toxicology
- Protein sequence similarity
- Relative intrinsic susceptibility