Unveiling toxicological adverse outcomes: Towards construction and simulation of large-scale networks

Predicting toxicological adverse outcomes is crucial for advancing in silico toxicology strategies. Modern toxicology increasingly relies on systems biology approaches to model and interpret these outcomes. Adverse outcome pathways (AOPs) focus on systems-level descriptions and causal linear relations among initiating, key, and adverse outcome events. Key characteristics (KC) based topologies capture mechanistic breadth via interconnected property-based modules without assuming linear causality. From another perspective emerging physiological maps dive deeper into toxicological mechanisms by mapping them from the detailed molecular level. To capture the dynamic nature of toxicological responses, especially their time- and dose-dependent behaviors, there is growing interest in integrating systems biology and mathematical modeling strategies. While dynamic models have been applied to small-scale AOPs, larger regulatory networks remain largely unexplored from a dynamic perspective. In this review, we highlight recent efforts to combine systems and network biology approaches for predicting toxicological adverse outcomes, covering network construction, analysis, and dynamic predictions. We also explore the aspect of dynamically simulating large-scale molecular networks and its potential contribution to systems toxicology. Specifically, we charter the use of logic-based models (Boolean networks) as an integrative approach to understand molecular crosstalk and cellular phenotypes, highlighting the potential repurpose of existing models. To this end, we show two use cases on toxicological applications of Boolean network models. Finally, we prospectively discuss the importance and need of bridging molecular and systemic scales and integrating these modeling strategies with high-dimensional data sources, including omics and multi-omics datasets.