Consensus Toxicity Factors for Polychlorinated Dibenzo-p-dioxins, Dibenzofurans, and Biphenyls Combining in Silico Models and Extensive in Vitro Screening of AhR-Mediated Effects in Human and Rodent Cells

Abstract

Consensus toxicity factors (CTFs) were developed as a novel approach to establish toxicity factors for risk assessment of dioxin-like compounds (DLCs). Eighteen polychlorinated dibenzo-p-dioxins, dibenzofurans (PCDD/Fs), and biphenyls (PCBs) with assigned World Health Organization toxic equivalency factors (WHO-TEFs) and two additional PCBs were screened in 17 human and rodent bioassays to assess their induction of aryl hydrocarbon receptor-related responses. For each bioassay and compound, relative effect potency values (REPs) compared to 2,3,7,8-tetrachlorodibenzo-p-dioxin were calculated and analyzed. The responses in the human and rodent cell bioassays generally differed. Most notably, the human cell models responded only weakly to PCBs, with 3,3',4,4',5-pentachlorobiphenyl (PCB126) being the only PCB that frequently evoked sufficiently strong responses in human cells to permit us to calculate REP values. Calculated REPs for PCB126 were more than 30 times lower than the WHO-TEF value for PCB126. CTFs were calculated using score and loading vectors from a principal component analysis to establish the ranking of the compounds and, by rescaling, also to provide numerical differences between the different congeners corresponding to the TEF scheme. The CTFs were based on rat and human bioassay data and indicated a significant deviation for PCBs but also for certain PCDD/Fs from the WHO-TEF values. The human CTFs for 2,3,4,7,8-pentachlorodibenzofuran, 1,2,3,4,7,8-hexachlorodibenzofuran, 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin, and 1,2,3,4,7,8,9-heptachlorodibenzofuran were up to 10 times greater than their WHO-TEF values. Quantitative structure-activity relationship models were used to predict CTFs for untested WHO-TEF compounds, suggesting that the WHO-TEF value for 1,2,3,7,8-pentachlorodibenzofuran could be underestimated by an order of magnitude for both human and rodent models. Our results indicate that the CTF approach provides a powerful tool for condensing data from batteries of screening tests using compounds with similar mechanisms of action, which can be used to improve risk assessment of DLCs.