A tier evaluation system was assessed as an alternative method for the identification of cosmetic ingredients which are not ocular irritants. The system employed monolayer cultures of SIRC cells, an established cell line originally derived from the rabbit cornea, and a three-dimensional living dermal model (LDM), MATREX™, which consists of human dermal fibroblasts in a contracted collagen lattice. Effects on the cell monolayer cultures were determined by using SIRC cell-Crystal Violet staining (SIRC-CVS), and effects on the LDM were assessed by using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. A non-irritating ingredient was defined as a compound having a maximal average total score (MAS) of 5 or less in the Draize eye test, as this is the criterion used in the Japanese draft guidance for evaluating cosmetic ingredients. Among 34 test substances with known characteristics, 30 were classified accurately. Based on these encouraging results, the possibility of simplifying the MTT assay on the LDM for more-practical use, by selecting only three concentration levels to discriminate non-irritants from irritants, was assessed. The simplified method, involving a three-dose set (the three-dose method), was confirmed as being suitable for the identification of non-irritating ingredients, with triethanolamine used as a negative reference standard. Finally, the LDM was used to evaluate compounds at similar concentrations to those tested in vivo, aiming to predict the concentration at which an ingredient can be formulated into products without causing eye irritation. On the basis of previous validation data and our additional results, it was found that test samples that resulted in a cell viability of 50% or more in this model, could be classified as non-irritating ingredients. In all, these results indicate that the tier evaluation system may be suitable for the evaluation of ingredients intended to be used in cosmetics and medicated cosmetics in Japan.

A tier evaluation system for the identification of cosmetic ingredients that are not ocular irritants was applied to 59 cosmetic ingredients, for which in vivo data were available. The tier system employs monolayer cultures of SIRC cells, an established cell line originally derived from rabbit cornea, and a three-dimensional living dermal model (LDM; MATREX™), which consists of human dermal fibroblasts in a contracted collagen lattice. The effects of ingredients on monolayer cultures of SIRC cells were determined by Crystal Violet staining (in the SIRC-CVS assay), and the effects on the LDM were measured by using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (in the LDM-MTT assay). The classifications of eye irritancy predicted by the in vitro system were compared with previously reported data obtained with the in vivo Draize rabbit eye test. The in vivo classification was based on appearance of corneal damage, or a maximal average score (MAS) of 15 as the cut-off point. The SIRC-CVS assay was effective in the prediction of compounds that would be non-irritants at a concentration of 10%, while the subsequent LDM-MTT assay could predict non-irritancy at various lower and higher concentrations, including 10%. The tier system gave very few false-negative predictions, though false positives were unavoidable. Performing the LDM-MTT assay with an additional 73 ingredients gave similar results in the prediction of non-irritancy at various concentrations. Our findings indicate that the tier system may be suitable for the safety assessment of eye irritancy of ingredients intended to be used in cosmetics and medicated cosmetics in Japan.