Hypothesis-driven approach to developmental toxicity assessment: Using mechanistic information to inform testing.

Area of Science:

• Toxicology
• Pharmacology
• Computational Chemistry

Background:

• Traditional developmental toxicity testing uses animal models and standardized protocols.
• Emerging mechanistic data allows for improved model selection and prediction of outcomes.
• Read-across methods, utilizing cheminformatics and biological similarity, are key for chemical assessment.

Purpose of the Study:

• To explore how understanding the mode of action (MoA) can refine developmental toxicity assessments.
• To demonstrate a shift towards more flexible, hypothesis-driven, and resource-efficient testing strategies.
• To highlight the integration of in vivo, in vitro, and computational methods.

Main Methods:

• Investigating the mode of action (MoA) of test agents.
• Utilizing read-across principles with cheminformatics for analog selection.
• Employing high-throughput screening (e.g., ToxCast) and transcriptomics.
• Leveraging induced pluripotent stem cells for human-relevant biological models.

Main Results:

• Mode of action understanding can identify activity cliffs in chemical series.
• Metabolism data can reduce the need for testing all similar chemical analogs.
• Gene expression analysis reveals divergent pharmacology among similar compounds.
• Mechanistic insights enable hypothesis-driven testing designs.

Conclusions:

• A mechanistic, hypothesis-driven approach offers a more flexible and efficient alternative to traditional developmental toxicity testing.
• Integration of diverse data (in silico, in vitro, omics) improves predictive accuracy.
• This paradigm shift supports more targeted and resource-conscious chemical safety evaluations.