Skin Sensitization Potency Assessments of Fragrance Materials using GARDskin Dose-Response
Joint poster with Research Institute for Fragrance Materials (RIFM),
Presented at the 2021 RIFM annual meeting & 2021 ACT annual meeting
Mihwa Na, Ulrika Mattson, Robin Gradin, Henrik Johansson, Andy Forreryd, Anne Marie Api, Research Institute for Fragrance Materials, Inc., 50 Tice Boulevard, Woodcliff Lake, NJ, 07677, USA., SenzaGen AB, Lund, Sweden
Conclusion
- GARDskin Dose-Response closely approximated the potency categories of 9/12 fragrance materials tested.
- Based on results from this dataset, GARDskin Dose-Response appears useful for potency assessment for weak sensitizers and may constitute a promising strategy for deriving a point of departure for quantitative risk assessments.
Abstract
Several New Approach Methods for hazard identification of skin sensitizers have been developed and incorporated as OECD Test Guidelines. However, the methods for potency assessment are still lacking. GARDskin (OECD TGP 4.106) was initially developed to identify skin sensitizers by monitoring transcriptional patterns of a biomarker signature in a dendritic like cell line.
The predictive capacity of GARDskin has been demonstrated previously, with 95.8% accuracy, 91.7% positive predictive value, and 100.0% negative predictive value (1 false positive, n=24) (Johansson, Gradin et al. 2019). To derive potency information, a strategy based on dose-response measurements in GARDskin, referred to as the GARDskin Dose-Response assay, has recently been proposed. The readout of the assay corresponds to the lowest concentration required to exceed the binary classification threshold in GARDskin. This concentration correlates with local lymph node assay (LLNA) EC3 and human no observed effect level (NOEL) values and linear regression models have been established to exploit these relationships for potency predictions.
In this blinded study, 12 fragrance materials (10 very weak sensitizers and 2 weak sensitizers) were evaluated in GARDskin Dose-Response. Results were evaluated by comparing predicted values to the reference potency categories. Three of the very weak sensitizers were predicted as non-sensitizers by GARDskin Dose-Response. For the remaining nine materials which were predicted as sensitizers, the predicted EC3 and NOEL values closely approximated the reference data for most materials. Based on results from this dataset, GARDskin Dose-Response appears useful for potency assessment for weak sensitizers and may constitute a promising strategy for deriving a point of departure for quantitative risk assessments.