New Docking, Molecular Dynamics, and QSAR Models to Predict Disruption of Human and Rat Transthyretin Function by...

Per- and polyfluoroalkyl substances (PFAS) are environmentally persistent chemicals that require an improved understanding of the toxicity mechanisms and the development of predictive models for risk assessment. One observed effect of PFAS exposure is a decrease in thyroxine (T₄) levels in vivo resulting from the direct displacement of T₄ from a carrier protein, transthyretin (TTR), in a proposed adverse outcome pathway (AOP). In this study, the mechanism of thyroxine (T₄) displacement from human and rat TTRs was investigated by using structural approaches (i.e., docking and molecular dynamics) and quantitative structure-activity relationship (QSAR) models. A QSAR model was developed using the largest available binding data set and a two-tier approach that allowed inclusion of all data. Docking models that utilized a pharmacophore approach showed nearly perfect overlap with independently sourced crystal structures for perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS). Molecular dynamics simulations demonstrated similar PFAS binding modes in rat and human TTR, enabling interspecies toxicity comparisons. All models predicted moderate to strong binding of the novel PFAS 4,8-Dioxa-3H-perfluorononanoic acid (ADONA) and hexafluoropropylene oxide dimer acid (GenX) to TTR, consistent with the limited toxicity and binding data for these chemicals. Predicted PFAS binding energies for rat TTR correlated well with the in vivo PFAS-associated decreases in T₄ levels, supporting the AOP. The development of reliable predictive toxicity models for PFAS requires extensive validation, maximal use of available experimental data, and careful consideration of toxicokinetic differences in interchemical comparisons.