Applicability of amino acid derivative reactivity assay (4 mM) for the prediction of skin sensitization by combining multiple alternative methods to evaluate key events

The amino acid derivative reactivity assay (ADRA) is an alternative method for evaluating key event 1 (KE‐1) in the skin sensitization mechanism included in OECD TG442C (OECD, 2021). Recently, we found that ADRA with a 4‐mM test chemical solution had a higher accuracy than the original ADRA (1 mM). However, ADRA (4 mM) has yet to be evaluated using integrated approaches to testing and assessment (IATA), a combination of alternative methods for evaluating KE. In this study, the sensitization potency of three defined approaches (DAs) using ADRA (4 mM) as KE‐1 was predicted and compared with those of two additional ADRAs or direct peptide reactivity assay (DPRA): (i) “2 out of 3” approach, (ii) “3 out of 3” approach, and (iii) integrated testing strategy (ITS). In the hazard identification of chemical sensitizers, the accuracy of human data and local lymph node assay (LLNA) remained almost unchanged among the three approaches evaluated. Potency classifications for sensitization were predicted with the LLNA and human data sets using ITS. The potency classifications for the sensitization potency prediction accuracy of LLNA data using any alternative method were almost unchanged, at approximately 70%, and those with ITS were not significantly different. When ITS was performed using DPRA, the prediction accuracy was approximately 73% for human data, which was similar to that of the LLNA data; however, the accuracy tended to increase for all ADRA methods. In particular, when ITS was performed using ADRA (4 mM), the prediction accuracy was approximately 78%, which proved to be a practical level. Three alternative DAs using ADRA (4 mM) as KE‐1, (i) “2 out of 3” approach, (ii) “3 out of 3” approach, and (iii) ITS, were compared with methods using two different ADRAs (1 mM and 0.5 mg/ml) or DPRA. When ADRA (4 mM) was used as the KE‐1 alternative for ITS, the accuracy of hazard identification and sensitization potency prediction for humans was the highest among all approaches.