Abstract:Adverse Outcome Pathway (AOP) is a useful tool to glean mode of action (MOE) of a chemical. However, in order to use it for the purpose of risk assessment, an AOP needs to be quantified using in vitro or in vivo data. Majority of quantitative AOPs developed so far, were for single exposure to progressively higher doses. Limited attempts were made to include time in the modeling. Here as a proof-of concept, we developed a hypothetical AOP, and quantified it using a virtual dataset for six repeated exposures using a Bayesian Network Analysis (BN) framework. The virtual data was generated using realistic assumptions. Effects of each exposure were analyzed separately using a static BN model and analyzed in combination using a dynamic BN (DBN) model. Our work shows that the DBN model can be used to calculate the probability of adverse outcome when other upstream KEs were observed earlier. These probabilities can help in identification of early indicators of AO. In addition, we also developed a data driven AOP pruning technique using a lasso-based subset selection, and show that the causal structure of AOP is itself dynamic and changes over time. This proof-of-concept study revealed the possibility for expanding the applicability of the AOP framework to incorporate biological dynamism in toxicity appearance by repeated insults.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to develop a method for an Adverse Outcome Pathway (AOP) model that can quantify chronic toxicity caused by repeated exposure. Specifically, the researchers hope to show how to use virtual data sets and the Bayesian Network Analysis (BN) framework to quantify this hypothesized AOP, with particular attention to the application of Dynamic Bayesian Network (DBN) in calculating the probability of Adverse Outcome (AO) after observing upstream Key Events (KEs) at an early stage. In addition, they have also developed an AOP pruning technique based on LASSO to quantify the dynamic changes in causality in the AOP. ### Background and Objectives of the Paper 1. **Limitations of the AOP Framework**: Although the AOP framework is very useful for understanding the mechanism of action of chemical substances, most existing Quantitative AOP (qAOP) mainly focus on the dose - response relationship after a single exposure and rarely consider the time factor. 2. **Research Objectives**: This paper aims to expand the application range of the AOP framework through a proof - of - concept study, showing how to use virtual data sets and DBN models to quantify chronic toxicity caused by repeated exposure, so that it can better reflect biological dynamic changes. ### Overview of the Methods 1. **Generation of Virtual Data**: The researchers generated a virtual data set, simulating the situation of six repeated exposures. These data are based on realistic assumptions, including responses from different doses and different donors. 2. **Static and Dynamic Bayesian Network Models**: Use the static BN model to analyze the impact of each exposure separately, and use the DBN model to comprehensively analyze the cumulative effects of multiple exposures. 3. **AOP Pruning Technique**: Developed a subset selection method based on LASSO for pruning irrelevant or unimportant causal links in the AOP, thereby more accurately describing the dynamic changes of the AOP. ### Main Results 1. **Generation of Virtual Data**: The generated data set shows that in the first three exposures, the chronic - phase biomarkers were not activated; starting from the fourth exposure, the chronic - phase response gradually increased. 2. **Dose - Response and Response - Response Relationships**: The acute - phase nodes showed a strong dose - dependence in all exposure times, while the chronic - phase nodes showed dose - dependence only starting from the third exposure. 3. **Probability Analysis**: Through Generalized Bayesian Network (GBN) model fitting and logical sampling, the activation probabilities of various key events at different doses were calculated. The results show that the activation probability of acute - phase key events increases with the increase of dose, and the chronic - phase key events are gradually activated after multiple exposures. 4. **Application of DBN Model**: The DBN model successfully calculated the probability of adverse outcomes after observing upstream key events at an early stage, which helps to identify early indicators. ### Conclusion This study shows how to use DBN models and virtual data sets to quantify chronic toxicity caused by repeated exposure and expands the application range of the AOP framework. Through this method, we can better understand and predict the impact of long - term low - dose exposure on health and provide new tools for risk assessment.

Proof of concept for quantitative adverse outcome pathway modeling of chronic toxicity in repeated exposure

The adverse outcome pathway concept: a pragmatic tool in toxicology

Developing adverse outcome pathways on silver nanoparticle-induced reproductive toxicity via oxidative stress in the nematode Caenorhabditis elegans using a Bayesian network model

A toxicity pathway-oriented approach to develop adverse outcome pathway: AHR activation as a case study

Development of an adverse outcome pathway network for nephrotoxicity

The First Steps on AOPs' Concepts, Development, and Applications in Teratology

Application of the Adverse Outcome Pathway Concept to In Vitro Nephrotoxicity Assessment: Kidney Injury due to Receptor-Mediated Endocytosis and Lysosomal Overload as a Case Study

Adverse outcome pathways: A conceptual framework to support ecotoxicology research and risk assessment

Mapping Adverse Outcome Pathways for Kidney Injury as a Basis for the Development of Mechanism-Based Animal-Sparing Approaches to Assessment of Nephrotoxicity

Quantitative structure–activity relationships of chemical bioactivity toward proteins associated with molecular initiating events of organ-specific toxicity

A quantitative weight-of-evidence method for confidence assessment of adverse outcome pathway networks: a case study on chemical-induced liver steatosis

Adverse outcome pathways: opportunities, limitations and open questions

Prediction of the Neurotoxic Potential of Chemicals Based on Modelling of Molecular Initiating Events Upstream of the Adverse Outcome Pathways of (Developmental) Neurotoxicity

Potential adverse outcome pathways with hazard identification of organophosphate esters

Harvesting the promise of AOPs: An assessment and recommendations

Adverse Outcome Pathway Networks II: Network Analytics.

Hepatic toxicity prediction of bisphenol analogs by machine learning strategy

Crowdsourcing AOP development: Leveraging the thesis literature review to identify knowledge gaps and facilitate research translation

Prediction of Zebrafish Embryonic Developmental Toxicity by Integrating Omics with Adverse Outcome Pathway.

An integrative data-centric approach to derivation and characterization of an adverse outcome pathway network for cadmium-induced toxicity

A Quantitative Adverse Outcome Pathway for Embryonic Activation of the Aryl Hydrocarbon Receptor of Fishes by Polycyclic Aromatic Hydrocarbons Leading to Decreased Fecundity at Adulthood