New approach methods (NAMs) supporting read-across: Two neurotoxicity AOP-based IATA case studies

Regulatory bodies currently face a significant challenge in replacing traditional animal testing with modern, non-animal alternatives. No prior work has fully resolved how to integrate diverse data streams for complex neurotoxicity assessments. Prior research has shown that read-across strategies rely heavily on chemical and biological similarities between substances. That uncertainty drove the need for more robust, pathway-based frameworks to support these regulatory decisions. This gap motivated the development of integrated testing strategies that combine computational modeling with laboratory assays. It was already known that mitochondrial dysfunction plays a major role in the degeneration of specific neuronal populations. However, translating these mechanistic insights into standardized regulatory workflows remains a difficult task. This study addresses these issues by examining two specific pesticide classes within an established international framework.

The aim of this study is to evaluate the effectiveness of modern testing strategies in supporting regulatory read-across decisions. The authors seek to replace traditional animal experiments by addressing data gaps through innovative, non-animal methodologies. They focus on two classes of pesticides to demonstrate how specific modes of action can be assessed for neurological hazards. The research addresses the challenge of substantiating these strategies with ample, successful examples. By utilizing adverse outcome pathways, the team intends to provide a clear, mechanistic basis for safety predictions. They explore how different read-across concepts, such as structural versus biological similarity, influence the accuracy of these assessments. The study also investigates the impact of data uncertainty on the final regulatory conclusions. Ultimately, the researchers aim to provide generic learnings that can guide future applications of pathway-based testing in regulatory settings.

Review approach framing involves analyzing two specific case studies endorsed by international regulatory bodies. The authors examine rotenoids and strobilurins to test the efficacy of pathway-based safety assessments. They employ a diverse array of computational tools, including molecular docking, to predict chemical interactions. Laboratory experiments utilize various cell systems to capture biological responses to the compounds. The researchers integrate transcriptomic data to map cellular changes against established adverse outcome pathways. Toxicokinetic simulations are performed to estimate human tissue concentrations for each chemical class. This design allows for a direct comparison between analogue and category-based read-across concepts. The team evaluates how structural versus biological similarities influence the final safety predictions.

Key findings from the literature indicate that mitochondrial respiratory chain inhibition serves as a reliable indicator for neurotoxic potential. The study successfully links complex I inhibition to the degeneration of dopaminergic neurons across both pesticide classes. Results show that integrating diverse data streams, such as transcriptomics and docking, provides a comprehensive view of toxicodynamic properties. The authors report that high uncertainty in specific data elements did not prevent a clear regulatory conclusion. Their analysis demonstrates that biological similarity often outweighs structural similarity in predicting neurotoxicity. The case studies confirm that these integrated strategies function effectively within different regulatory environments. The findings highlight that both positive and negative predictions can be accurately derived using this pathway-anchored approach. This work provides a clear proof-of-concept for replacing animal testing with modern, non-animal alternatives.

The authors propose that integrating mechanistic pathways significantly strengthens the reliability of non-animal safety assessments. Synthesis and implications suggest that even when individual data points contain uncertainty, the overall regulatory conclusion can remain robust. The researchers highlight that biological similarity often provides a more reliable basis for predictions than structural similarity alone. Their findings indicate that diverse testing platforms, including transcriptomics and docking, effectively support the read-across process. The study demonstrates that these integrated strategies are ready for application in varied regulatory contexts. The authors conclude that standardized frameworks help bridge the gap between complex biological data and safety decision-making. Their work suggests that moving away from animal models is feasible when using well-defined, pathway-anchored approaches. Future efforts should focus on refining these strategies to increase confidence in non-animal safety predictions.