Kristen U, Jung K, Pape W, Pfannenbecker U, Rensch A, Schell R
Institut für Allgemeine Botanik, Universität Hamburg, Ohnhorststrasse 18, D-22609 Hamburg, Germany.
Toxicol In Vitro. 1999 Apr;13(2):335-42. doi: 10.1016/s0887-2333(98)00079-4.
In the present paper, we describe and analyse the performance and the results of the pollen tube growth (PTG) test applied to the COLIPA international validation study of in vitro alternatives to the Draize eye irritation test. The PTG test, based on photometric quantification of in vitro pollen tube mass production, was used by three independent laboratories to estimate the acute eye irritation potentials of 23 ingredients and 32 cosmetic formulations. Basing on historical Draize test data and on IC(50) values of previously tested cosmetic formulations, a mathematical formula was generated to predict rabbit eye irritation potentials from PTG test results. Statistical evaluation of the calculated modified maximal average scores (MMAS) revealed a high prediction capability of the PTG test in regard to the finished formulations but a relatively low one for alcohols, higher concentrated cationic surfactants, and acidic and alkaline materials. Furthermore, our results indicated that the PTG test was able to produce precise IC(50) values without any limitations from all of the 55 test substances with good intra- and interlaboratory reproducibility. From these findings we suggest that the PTG test is not a validated test at present but is considered to be a potent candidate for further validation processes. For this purpose an additional prediction model for ingredient classes as mentioned above must be generated.
在本论文中，我们描述并分析了应用于COLIPA国际验证研究中的花粉管生长（PTG）试验的性能和结果，该研究旨在验证替代兔眼刺激试验的体外方法。PTG试验基于体外花粉管大量生成的光度定量分析，由三个独立实验室用于评估23种成分和32种化妆品配方的急性眼刺激潜力。基于历史的兔眼刺激试验数据以及先前测试的化妆品配方的半数抑制浓度（IC50）值，生成了一个数学公式，用于根据PTG试验结果预测兔眼刺激潜力。对计算得出的修正最大平均得分（MMAS）进行的统计评估显示，PTG试验对于成品配方具有较高的预测能力，但对于醇类、高浓度阳离子表面活性剂以及酸性和碱性物质的预测能力相对较低。此外，我们的结果表明，PTG试验能够精确得出所有55种测试物质的IC50值，且不受任何限制，在实验室内部和实验室之间具有良好的重现性。基于这些发现，我们认为PTG试验目前尚未经过验证，但被视为进一步验证过程的有力候选方法。为此，必须针对上述成分类别生成额外的预测模型。
