Computational Toxicology

Mechanism-Driven Virtual Adverse Outcome Pathway (vAOP) Modeling 

for Chemical Toxicity

 

General workflow for construction of data-driven and mechanism-driven models for chemical toxicity

Background

High Throughput Screening (HTS) studies provide the community with rich chemical toxicity data that can be integrated into computational toxicity modeling. A general question raised from the current big data era is what is the benefit of a specific toxicity bioassay (normally refers to a specific toxicity mechanism) in the evaluation more complicated toxicity phenomena (e.g. toxicity in animals/humans). In 2010, Adverse Outcome Pathways (AOPs) were introduced to advance the toxicity studies by illustrating toxicity mechanisms of toxicants. However, for most toxic chemicals, the AOPs cannot be revealed due to insufficient knowledge of complex toxicity phenomena. In Zhu lab, one of the major efforts is to develop virtual AOP (vAOP) models that integrate big toxicity data to ascribe the modes and mechanisms of toxicant action using novel machine learning and data science techniques. The resulted vAOP models can resolve the above challenges by providing mechanism-based risk assessments of new compounds. Moreover, any specific toxicity assay can be used as one key event in the toxicity pathway to enrich the data of modeling and strengthen the predictions of resulted models.

 

Case study 1: Mechanistic modeling of hepatotoxicity

In this study, we integrated the AOP framework into a new computational modeling strategy using a mechanistically relevant HTS assay (antioxidant response element [ARE]) and structural alerts. The resulting mechanistic AOP model, consisting of the identified structural alerts as initiating events and in vitro ARE activation as a key event, shows good predictivity and provides an oxidative stress-related toxicity pathway leading to hepatotoxicity. Extra data obtained from in vitro assays measuring other key events of hepatotoxicity AOPs listed on AOP-Wiki can aid in reducing false negative predictions in our model by incorporating other toxicity mechanisms of chemicals. This novel strategy can be used to build hepatotoxicity models involving other toxicity mechanisms and further develop predictive AOP models for other complex toxicity endpoints.

Case Study 2: Uterotrophic bioactivity modeling using a knowledge-based DNN approach

In this study, a novel knowledge-based DNN modeling framework was developed to mimic a toxicity pathway for ERα and ERβ agonists using a virtual AOP (vAOP) framework. A data set of 42 compounds with known in vivo rodent uterotrophic bioactivity was used to train this network. This vAOP framework accurately mimics the comprehensive effects of in vitro bioassays related to toxicity pathway KEs to predict in vivo outcomes by incorporating chemical fragments and hierarchically structured biological data during the training process. 

Case Study 3: Integrating Concentration-Dependent Toxicity Data and Toxicokinetics To Inform Hepatotoxicity Response Pathways

In this work, we present AOP modeling employing a new computational strategy that integrates concentration-dependent toxicity data and relevant toxicokinetics. Using a database of 2171 chemicals with human hepatotoxicity data, we extracted HTS assays related to hepatotoxicity and grouped these assays into 52 KE groups to get the KE scores for compounds. KE scores can be used to inform the mechanisms and be combined with bioavailability information estimated from toxicokinetic modeling to predict the in vivo hepatotoxicity of chemicals.


Related Research Articles:

Carcinogenicity:

Chung, E., Russo, D. P., Ciallella, H. L., Wang, Y. T., Wu, M., Aleksunes, L. M., & Zhu, H. (2023). Data-Driven Quantitative Structure–Activity Relationship Modeling for Human Carcinogenicity by Chronic Oral Exposure. Environmental Science & Technology, 57(16), 6573-6588. 

Prenatal Developmental Toxicity:

Ciallella, H. L., Russo, D. P., Sharma, S., Li, Y., Sloter, E., Sweet, L., ... & Zhu, H. (2022). Predicting prenatal developmental toxicity based on the combination of chemical structures and biological data. Environmental Science & Technology, 56(9), 5984-5998. 

Hepatotoxicity

Russo, D. P., Aleksunes, L. M., Goyak, K., Qian, H., & Zhu, H. (2023). Integrating Concentration-Dependent Toxicity Data and Toxicokinetics To Inform Hepatotoxicity Response Pathways. Environmental Science & Technology. (Case Study 3)

Jia, X., Wen, X., Russo, D. P., Aleksunes, L. M., & Zhu, H. (2022). Mechanism-driven modeling of chemical hepatotoxicity using structural alerts and an in vitro screening assay. Journal of Hazardous Materials, 436, 129193.  (Case Study 1)

Zhao, L., Russo, D. P., Wang, W., Aleksunes, L. M., & Zhu, H. (2020). Mechanism-driven read-across of chemical hepatotoxicants based on chemical structures and biological data. Toxicological sciences, 174(2), 178-188. 

Kim, M. T., Huang, R., Sedykh, A., Wang, W., Xia, M., & Zhu, H. (2016). Mechanism profiling of hepatotoxicity caused by oxidative stress using antioxidant response element reporter gene assay models and big data. Environmental health perspectives, 124(5), 634-641. 

Endocrine Disruption

Ciallella, H. L., Russo, D. P., Aleksunes, L. M., Grimm, F. A., & Zhu, H. (2021). Revealing adverse outcome pathways from public high-throughput screening data to evaluate new toxicants by a knowledge-based deep neural network approach. Environmental science & technology, 55(15), 10875-10887. (Case Study 2)

Ciallella, H. L., Russo, D. P., Aleksunes, L. M., Grimm, F. A., & Zhu, H. (2021). Predictive modeling of estrogen receptor agonism, antagonism, and binding activities using machine-and deep-learning approaches. Laboratory investigation, 101(4), 490-502. 

Acute Toxicity 

Russo, D. P., Strickland, J., Karmaus, A. L., Wang, W., Shende, S., Hartung, T., ... & Zhu, H. (2019). Nonanimal models for acute toxicity evaluations: Applying data-driven profiling and read-across. Environmental health perspectives, 127(4), 047001. 

Review Articles:

Jia, X., Wang, T., & Zhu, H. (2023). Advancing Computational Toxicology by Interpretable Machine Learning. Environmental Science & Technology.

Ciallella, H. L., & Zhu, H. (2019). Advancing computational toxicology in the big data era by artificial intelligence: data-driven and mechanism-driven modeling for chemical toxicity. Chemical research in toxicology, 32(4), 536-547.

Zhu, H., Zhang, J., Kim, M. T., Boison, A., Sedykh, A., & Moran, K. (2014). Big data in chemical toxicity research: the use of high-throughput screening assays to identify potential toxicants. Chemical research in toxicology, 27(10), 1643-1651.