The prediction of ocular irritation potential from in vitro assays still presents a problem, despite a number of validation trials. A study with coded cosmetic formulations, for which historic in vivo data were available, has been conducted with a human corneal multi-layered model system. This corneal model, the HCE-T model, was developed by using HCE-T cells, a transfected human corneal epithelial cell line. The relative effectiveness of three endpoints that provide a measure of cytotoxicity in the HCE-T model was evaluated. Cell viability immediately after exposure to the test materials was determined by using the MTT and Alamar Blue™ (AB) assays, and, 24 hours later, by using the MTT, AB and lactate assays. Viability measurements with the MTT, AB and lactate assays gave similar dose-response curves at the 24-hour endpoint. One formulation (an anti-dandruff shampoo) caused a less severe drop in viability in assays conducted immediately after the exposure than at the 24-hour time-point. There was little deterioration in viability with the other test materials. The ranking of the test formulations on the basis of relative loss of viability and release of lactate resulted in the same order as for the Modified Maximum Average Draize Test Score. Comparison of the HCE-T model cytotoxicity assay results with historic in vitro data from two different cytotoxicity assays, conducted by using fibroblast monolayer cultures and the same materials, indicated that the multi-layered corneal model had a greater predictive ability. The results of a blind trial with the lactate assay in two laboratories indicated that the techniques required were transferable between laboratories. The lactate results were reproducible between laboratories, even when cultures derived from different passage human corneal cells were tested, provided that the passage number was below 20.
2000 FRAME.