THE EPIOCULAR PREDICTION MODEL: AN ACCURATE, REPRODUCIBLE IN VITRO MEANS OF ASSESSING OCULAR IRRITANCY POTENTIAL.

• TR Number: 147
• Authors: Klausner, M., Sennott, H.A., Breyfogle, B., Makwana, A., Kubilus, J. Mattek Corp., Ashland, MA.

The EpiOcular™ tissue model (OCL-200) is an organotypic model of the human corneal epithelium (HCE) cultured from normal human keratinocytes using serum free medium. Paraffin embedded, H&E stained histological cross sections show the structure of EpiOcular closely parallels that of the HCE; large nucleated basal cells lie beneath 5-6 stratified cell layers which progressively flatten out, ending with a highly squamous, non-keratinized layer at the culture’s apical surface. During 1996 and 1997, 47 and 82 lots, respectively, of the OCL-200 model were produced and monitored using the MTT assay. The common surfactant, 0.3% Triton X-100, was used as a positive control and the Effective Time 50 values (time of exposure after which viability is reduced to 50%, or ET 50) was determined for each lot of tissue. The ET-50 average of 24.9 + 6.3 (+ 1 standard deviation) minutes for 1996 (n=47 lots) was not statistically different (p = 0.043) than the 1997 average of 22.9 + 4.7 minutes (n = 81 lots). Stability of the EpiOcular model for international shipment was probed using these standard quality control tests following packaging and storage of the tissue for 4 days; both positive and negative controls and histology of the tissue remained constant over this time period. Interlaboratory testing conducted in 3 testing labs also demonstrated the stability of the tissue following shipment. Coefficients of variation (c.v.) for three materials tested in three laboratories were all below 10%. ET-50 values for 19 water-soluble chemicals from the ECETOC database were measured using 4 separate lots of EpiOcular tissue. The resultant ET-50 values were correlated to Draize rabbit eye scores (r = 0.90) and a prediction model was constructed. Next, a set of 41 materials including final formulation shampoos, off-the-shelf cosmetic/personal care products, and surfactants, for which Draize data were available, were tested and their respective ET-50’s were determined. Based on the prediction model, the predicted Draize and the actual Draize scores were compared; a correlation coefficient, r = 0.86, was obtained. Thus, the EpiOcular tissue model appears to be a reproducible and accurate in vitro means of reducing and/or replacing in vivo, animal-based ocular irritancy testing.