The result of a constructive collaboration with the EPAA working group for skin sensitization
A B S T R A C T
In the last 20 years, alternative approaches to the identification of skin sensitisation hazards have been at the forefront of the 3Rs and have helped refine the validation and acceptance processes. However, experience with the local lymph node assay showed that, post-validation, challenges still occurred, particularly when a wider diversity of chemical substances was addressed, a situation which will arise with validated in vitro alternatives. In the present work, a range of substances potentially challenging to assess in current nonanimal OECD test
guidelines were evaluated in several of the emerging in vitro alternatives. Twelve such substances (of which just over half were known skin sensitisers) were assessed in 4 assays, all based on reconstructed human epidermis (RHE) models. For hazard identification, the overall predictive accuracy ranged around 70% for three assays, although for one (SensCeeTox), it fell below 50% when human data was used as the benchmark. In most cases, sensitivity was high, such that sensitisation was overpredicted. As the substances were challenging to assess in
other nonanimal methods, the results indicate that the 3D RHE models may be a useful tool for assessing skin sensitisation potentials without needing to revert to animal use.