In vitro assessment of known environmental contaminants and mixtures to determine in vivo relevance

Environmental chemical exposures typically occur as complex mixtures, yet traditional risk assessment frameworks often focus on single-compound evaluations. In this study, we evaluated how inactive chemicals influence mixture modeling outcomes and explored optimal strategies for predicting the toxicity of polycyclic aromatic compound (PAC) mixtures through aryl hydrocarbon receptor (AhR) activation. Using in vitro bioassay data, we constructed defined mixtures composed of six active and seven inactive PACs in equimolar and environmentally relevant ratios. We compared measured responses to predictions generated by three established mixture models: concentration addition (CA), independent action (IA), and generalized concentration addition (GCA), using both EC50 and benchmark concentration (BMC10)-based approaches. Our results demonstrate that including inactive chemicals without adjusting their contribution skews mixture potency predictions, particularly in traditional effective concentration (EC)-based models that assume uniform efficacy. Predictive accuracy improved significantly when contributions were scaled to reflect only active chemicals. GCA consistently produced the closest agreement with measured responses among modeling approaches, particularly when benchmark dose modeling was applied. Unlike methods that use an EC, which are sensitive to variability in curve tops, BMC modeling benchmarks from the response baseline and better accommodate partial agonists. This study supports a pragmatic, mechanism-based approach to environmental mixture modeling: prioritize active components, scale their contributions, and adopt benchmark dose-based methods to assess potency. These findings have direct relevance for tiered environmental monitoring frameworks that rely on in vitro screening assays, offering a path forward to improve mixture risk assessment in regulatory contexts while accounting for the complexity of real-world chemical exposures.  

Impact/Purpose

Environmental contaminant exposures do not occur in isolation but as complex and dynamic mixtures. While many studies have compared predicted and measured effects of defined mixtures, often noting discrepancies between chemical composition and bioassay responses, few have explicitly examined how inactive chemicals contribute to these responses or influence mixture modeling, particularly in a pre-emptive, mechanistic context. Thus, we aimed to accurately predict the PAC mixture toxicity on AhR transcriptional activation using a model of concentration additivity that assumes no interactions among constituents and determine how the inclusion of inactive chemicals in a mixture impacts predictions of efficacy and potency. Our findings support a pragmatic approach to environmental mixture modeling: prioritize the most active chemicals, scale their contributions, and use BMC-based methods to estimate potency when possible. This strategy enhances both the accuracy and interpretability of predictions, particularly in regulatory contexts where potency is the most relevant parameter for risk assessment. Our results also highlight that including inactive chemicals without contribution and adjustment can lead to inaccurate model outputs, whereas adjusting the contribution to include only active constituents improves predictive alignment with observed responses. Although no single approach can fully capture the complexity of environmental mixtures, especially those with unknown components or diverse modes of action, our work provides a framework for improving the design and interpretation of mixture studies. 

Citation

Eccles, K., K. Gaston, E. Green, S. Waidyanatha, B. Stiffler, S. Harris, C. Rider, AND E. Medlock Kakaley. In vitro assessment of known environmental contaminants and mixtures to determine in vivo relevance. Elsevier B.V., Amsterdam, NETHERLANDS2976-2987, (2026). [DOI: 10.1021/acs.est.5c11914]

DOI: In vitro assessment of known environmental contaminants and mixtures to determine in vivo relevance