Non-enzymatic glutathione reactivity and in vitro toxicity: a non-animal approach to skin sensitization.

The development of non-animal methods to predict the potential of chemicals to cause skin sensitization is of great importance. On the basis of many published studies into the underlying chemical mechanisms skin sensitization, the immunological priming which leads to the disease allergic contact dermatitis, is recognized as a reactive chemistry endpoint. Consequently, the combination of chemical assays with in vitro techniques may provide a useful surrogate to animal testing for skin sensitization. This study attempts to investigate the relationship between skin sensitization assessed in the local lymph node assay (LLNA) initially and a thiol reactivity index based on glutathione (GSH), pEC(50) thiol (EC(50) being defined as the concentration of the test substance which gives 50% depletion of free thiol under standard conditions) in combination with a measure of cytotoxicity (pIGC(50)) to Tetrahymena pyriformis (TETRATOX). The pEC(50) thiol values and the pIGC(50) values were determined for twenty-four compounds for which LLNA test data were available. Thiol reactivity was found to discriminate sensitizers from non-sensitizers according to the rule: pEC(50) thiol>-0.55 indicates that the compound will be a skin sensitizer. However, because of metabolic activation a pEC(50) thiol<-0.55 does not necessarily mean that the compound will be a non-sensitizer. Excess toxicity to T. pyriformis (i.e. the extent of toxic potency over that expected by non-polar narcosis) was determined in order to assess biological reactivity. The best discrimination based on excess toxicity in the TETRATOX assay was given by the "rule": excess toxicity>0.50 indicates that the compound will be a skin sensitizer. These approaches become more powerful when combined. When taken together, the thiol and TETRATOX assays predict the sensitization potential of 23 of the 24 compounds correctly. alpha-Hexylcinnamic aldehyde is incorrectly predicted to be a non-sensitizer, whereas LLNA results suggest it may be a weak sensitizer, this inaccuracy being rationalized in terms of its high hydrophobicity. Due to the selectivity of electro(nucleo)philic reactions some sensitizing compounds will not be identified using a single nucleophile such as thiol.