Advancing Skin Sensitization Potency Categorization Using U-SENS™ in OECD TG 497

Area of Science:

• Toxicology
• Dermatology
• Computational Biology

Background:

• OECD Guideline 497 permits flexibility in Defined Approaches (DAs) by allowing substitution of New Approach Methodologies (NAMs).
• Integrated Testing Strategies (ITS) within DAs aim to predict skin sensitization hazard and potency.
• The Adverse Outcome Pathway (AOP) framework guides the selection of NAMs for specific key events (KEs).

Purpose of the Study:

• To evaluate the integration of the U-SENS™ assay (dendritic cell activation, KE3) into two ITS DAs.
• To determine EC150 threshold values for converting U-SENS™ data into DA scores using computational modeling.
• To assess the performance of modified ITS DAs against the Local Lymph Node Assay (LLNA) and human reference data.

Main Methods:

• Computational modeling to establish EC150 thresholds for U-SENS™ CD86 stimulation index data.
• Implementation of U-SENS™ within two existing ITS DAs following OECD Guideline 497.
• Performance evaluation using balanced accuracy for hazard identification and correct classification rates for potency categorization.

Main Results:

• Modified ITS DAs achieved 79% balanced accuracy (vs. LLNA) and 73-74% (vs. human data) for hazard identification.
• Potency categorization achieved 67-68% correct classification (vs. LLNA) and 65-67% (vs. human data).
• The LLNA demonstrated lower performance: 58% balanced accuracy (hazard) and 60% (potency) against human data.

Conclusions:

• Incorporating U-SENS™ into ITS DAs enhances flexibility without sacrificing predictive accuracy for skin sensitization.
• The U-SENS™-based DAs provide reliable classification outcomes, outperforming traditional animal testing (LLNA).
• Findings support the practical application of these DAs for regulatory skin sensitization assessment.