A transcriptome-based classifier to identify developmental toxicants by stem cell testing: design, validation and...

Area of Science:

• Developmental toxicology
• Stem cell biology
• Transcriptomics
• Computational toxicology

Background:

• Urgent need for reliable test systems to identify developmental toxicants.
• Existing methods lack the precision to group toxicants by mode of action.
• Human stem cell technology offers a promising in vitro model for developmental toxicity testing.

Purpose of the Study:

• To demonstrate the proof of concept for identifying and grouping toxicants with related modes of action using human stem cells and transcriptome analysis.
• To develop and validate a classifier for distinguishing between different classes of developmental toxicants.

Main Methods:

• Utilized a human pluripotent stem cell (UKN1) test system for neuroectoderm generation.
• Exposed cells to histone deacetylase inhibitors (HDACi) and mercurials.
• Performed transcriptome analysis (microarray) and applied a support vector machine (SVM)-based classifier.

Main Results:

• Identified consensus genes for HDACi and mapped them to a human transcription factor network.
• An SVM classifier accurately predicted HDACi and differentiated them from mercurials.
• Optimized classifier using eight genes (F2RL2, TFAP2B, EDNRA, FOXD3, SIX3, MT1E, ETS1, LHX2) for effective separation.

Conclusions:

• Human stem cells combined with transcriptome analysis provide a powerful tool for mechanistic grouping and prediction of toxicants.
• The developed method can accurately classify toxicants based on their mode of action, such as HDACi.
• This approach holds potential for predicting the developmental toxicity hazard of unknown compounds.