- Publisher:Lhasa Limited
- Publication Date:Nov 2016
- Publication Type:Presentation
- Scientific Area:
- Industry Type:
Can in silico models be used in place of the LLNA to predict the skin sensitisation of cosmetics
Skin sensitisation caused by the covalent binding of chemical to skin proteins has been researched extensively and has a well understood adverse outcome pathway (AOP) as published by the OECD in 2012. This AOP consists of multiple key events and a number of in chemico and in vitro assays have been designed to measure these e.g. DPRA, KeratinoSens and h-CLAT. It is generally expected that more than one of these assays would need to be used in an integrated testing strategy (ITS) in order to replace the LLNA/GPMT and in silico models have been shown to improve the predictivity of skin sensitisation compared to in chemico/in vitro assays alone. Nevertheless, as Derek Nexus is a rule-based in silico model built on in vivo animal (mainly LLNA and GPMT) and human data to derive mechanism-based alert domains in silico methods may be able to predict the skin sensitisation potential of chemicals accurately in the absence of other assays.
Cosmetic substances and ingredients data were exported from CosIng (n = 10836, data exported on 25/05/2016) and cross-referenced against a data set of LLNA results compiled from an in-house LLNA data set (n = 1348) and the NICEATM LLNA database (n = 1455). 325 unique inventory, colorants, preservatives, UV filters, restricted and banned chemicals remained with structural and LLNA data.
The results from analysis of this data set of cosmetics suggest that in silico models such as Derek Nexus are suitable tools to predict the skin sensitisation potential of cosmetic chemicals. It performs particularly well when assigning binary classifications, correctly predicting chemicals as sensitisers 82% and non-sensitisers 66% of the time. Predicting potency is trickier, however, the model predicts EC3 values well when evaluated by a number of criteria (3-fold, 5-fold, ECETOC category, GHS category).