Jones P A, Bracher M, Marenus K, Kojima H
SEAC Toxicology Unit, Unilever Research, Colworth House, Sharnbrook, Bedford MK44 1LQ, UK.
Toxicol In Vitro. 1999 Apr;13(2):325-33. doi: 10.1016/s0887-2333(98)00080-0.
The neutral red uptake (NRU) assay was included as part of the COLIPA international validation trial of in vitro alternatives to the Draize eye irritation test. In a blind trial, 55 substances were tested at four laboratories. Following testing, a prediction of the in vivo Draize modified maximum average score (MMAS) for each substance was made by each laboratory using a prediction model relating mean NR(50) value (concentration causing 50% reduction in NRU from that of untreated control cells) to MMAS. Following statistical analysis of the results and breaking of the code, assessment of the results and further analysis was carried out by the participating laboratories. This paper presents the conclusions with regard to the NRU assay. The initial trial analysis indicated that the interlaboratory reproducibility of results of the NRU assay was good. However, there was a poor correlation between observed and predicted MMAS (using the proposed prediction model) when all the test substances were analysed together (r=0.246). Data analysis of subsets of substances indicated that the best predictions were for pure surfactants only (r=0.843) although this data did not fit within the limits of the prediction model. The NRU assay therefore appears to have limited use as a complete Draize replacement. A further examination of the COLIPA trial data may identify combinations of assays which may be more useful than the individual assays which, like NRU, have been shown to be poor predictors of eye irritation.
中性红摄取（NRU）试验作为COLIPA国际验证试验的一部分，用于替代兔眼刺激试验的体外方法。在一项盲法试验中，四个实验室对55种物质进行了测试。测试后，每个实验室使用一个将平均NR(50)值（导致NRU相对于未处理对照细胞减少50%的浓度）与改良最大平均得分（MMAS）相关联的预测模型，对每种物质的体内Draize MMAS进行预测。在对结果进行统计分析并解密后，参与试验的实验室对结果进行了评估和进一步分析。本文介绍了关于NRU试验的结论。初步试验分析表明，NRU试验结果的实验室间再现性良好。然而，当对所有测试物质进行综合分析时，观察到的MMAS与预测的MMAS（使用所提出的预测模型）之间的相关性较差（r=0.246）。对物质子集的数据分析表明，最佳预测仅适用于纯表面活性剂（r=0.843），尽管该数据不符合预测模型的范围。因此，NRU试验作为兔眼刺激试验的完全替代方法，其用途似乎有限。对COLIPA试验数据的进一步检查可能会确定某些试验组合，这些组合可能比单独的试验更有用，例如NRU试验，已被证明对眼刺激的预测能力较差。
