ONTOX Insights #6

Adverse Outcome Pathways (AOPs) describe the sequence of events from chemical exposure to adverse outcomes, supporting predictive toxicology without animal testing. Within the ONTOX project, the DockTox (online tool that automates docking of small molecules onto more than 20 Molecular Initiating Event (MIE)-associated protein structures) was developed. The tool provides binding energies, interaction maps, and compares results to reference ligands, generating an “interaction fraction” that proved more reliable than binding energy alone in distinguishing binders from non-binders, as shown for PPARα. This unique metric helps improve the mechanistic understanding of protein–ligand interactions. DockTox can be applied to diverse protein targets, making it a flexible resource for toxicology research. It facilitates virtual screening of chemicals, prioritization of compounds, and contributes to more human-relevant risk assessment strategies. Ultimately, it supports the ONTOX vision of advancing chemical safety assessment without the use of animal testing.

The 5th International Conference on Developmental Neurotoxicity (DNT) Testing (DNT5) was held in April 2024 in Konstanz, Germany, organized by CAAT-Europe, the University of Konstanz, and partners from US EPA, SCAHT, and CAAT at Johns Hopkins University Bloomberg School of Public Health. The meeting brought together experts from academia, industry, and regulatory agencies to advance animal-free new approach methodologies (NAMs) in next-generation risk assessment. A central theme was the application and further development of the DNT in vitro test battery (DNT-IVB), supported by an OECD satellite meeting on its regulatory implementation. Discussions focused on validation strategies, in vitro-to-in vivo extrapolation, and intelligent assay combinations to increase predictive power. A key question raised was when the DNT-IVB will achieve sufficient biological and chemical coverage, highlighting the need for additional testing data and context-specific validation approaches. Presentations showcased innovative tools such as AI-driven models, multi-endpoint zebrafish assays, and complex brain organoids capable of electrical activity. Through its highly interactive format, DNT5 promoted extensive collaborative efforts in advancing the field toward more human-relevant, scientifically reliable, and ethical toxicological assessments.

The regular T4 workshops on biology-inspired microphysiological systems (MPS) have become a benchmark for tracking scientific, industrial, and regulatory progress in the MPS field. The 2023 workshop highlighted that MPS technology in academia has matured considerably, reflected by a steady increase in high-quality publications, and these developments can be find in the report. Gaps between academia, industry and regulators was addressed as well. The status of implementation of the recommendations from the 2020 workshop report was reviewed. While communication and collaboration between stakeholders has improved significantly, the recommendation related to regulatory decision-making still needs to be improved. To bridge this gap, more systematic validation and alignment with regulatory requirements are essential, alongside efforts to ensure the economic sustainability of MPS technologies.

Chemical risk assessment is shifting from traditional deterministic approaches toward probabilistic methodologies, where hazard manifestation is viewed as a variable likelihood shaped by exposure, individual factors, and stochastic processes. This evolution is driven by advances in human stem cell technologies, complex tissue models, high-performance computing, cheminformatics, and the rise of large-scale AI models. Together, these innovations enable a more nuanced understanding of chemical hazards and biological variability within populations. Each technology, however, introduces uncertainties that influence the estimation of hazard probabilities. Moving beyond fixed point estimates and threshold-based assessments, probabilistic frameworks integrate kinetic variability and uncertainty metrics, providing a more comprehensive basis for chemical safety evaluation. This paper summarizes a 2023 workshop examining the technological and data-driven enablers of this transition, along with implementation challenges—particularly regarding biological perturbation as the foundation for hazard estimation. The future of toxicological risk assessment lies in the effective integration of probabilistic models, enabling more accurate and holistic evaluations of chemical hazards.

Drug-induced liver injury (DILI) remains one of the leading causes of drug development failures and market withdrawals, creating major challenges for both pharmaceutical research and patient safety. Because the biological mechanisms underlying DILI are highly complex and difficult to predict, researchers have turned to data-driven approaches for better solutions. In this study, a comprehensive dataset of 28 in vitro endpoints covering liver toxicity, function, and DILI-specific features was collected. Authors applied multi-task learning and demonstrated how ensemble modeling can improve predictive performance. For further evaluation, Bayer assays for two of the endpoints (Bile salt export pump (BSEP) inhibition and phospholipidosis) were run, confirming the relevance of the approach. Additionally, an in-depth data analysis of the relationships among the different endpoints was conducted, as well as with DILI. The presented models can be used to derive a “Virtual Liver Safety Profile” showcasing the predicted activity of a compound on the selected endpoints to support the prioritization of assays and the elucidation of modes of action.

Follow us on social media!