A regression analysis using simple descriptors for multiple dermal datasets: Going from individual membranes to the full skin.
A regression analysis using simple descriptors for multiple dermal datasets: Going from individual membranes to the full skin.
复制标题
使用简单描述符对多个皮肤数据集进行回归分析：从单个膜到整个皮肤。
DOI：
10.1002/jat.4435
复制
发表时间：
2023
期刊：
影响因子：
--
通讯作者：
Evans MV
中科院分区：
文献类型：
--
作者：
Evans MV
需要通过计算机模拟方法来估计和/或量化皮肤对化学品的吸收，以实际预测药理学、职业和环境暴露。Potts-Guy方程是一种成熟的方法，使用多元线性回归分析描述了辛醇/水分配系数（logP）和分子量（MW）方面的皮肤渗透性（Kp）。在这项工作中，我们获得了与环境和化妆品化学品相关的不同人类数据集的回归方程。自从Potts-Guy方程于1992年发表以来，我们探索了包括不同皮肤层的最新数据集，例如真皮层（包括真皮到定义的厚度）和完整皮肤。我们的工作与其他人的观察是一致的，他们观察到Potts-Guy方程对于专注于表皮的实验更强。与表皮数据集相比，经皮皮肤和完整皮肤的渗透性估计值导致回归系数较低。更新的回归方程使用了先前评价的已发表2D房室模型获得的拟合渗透率值的组合。得到的回归方程为：logKp= −2.55 + 0.65logP − 0.0085MW，R2 = 0.91（所有数据集的适用范围：MW范围为18至>584 g/mol，logP为−4至>5）。这种方法证明了在一个单一的建模方法结合机制与结构活性关系的优势。与单独使用Potts-Guy方法获得的渗透率估计值相比，这种组合方法可以改善回归拟合。本研究中的分析假设单室皮肤吸收途径;未来的建模工作将考虑添加多室。
In silico methods to estimate and/or quantify skin absorption of chemicals as a function of chemistry are needed to realistically predict pharmacological, occupational, and environmental exposures. The Potts–Guy equation is a well‐established approach, using multi‐linear regression analysis describing skin permeability (Kp) in terms of the octanol/water partition coefficient (logP) and molecular weight (MW). In this work, we obtained regression equations for different human datasets relevant to environmental and cosmetic chemicals. Since the Potts–Guy equation was published in 1992, we explored recent datasets that include different skin layers, such as dermatomed (including dermis to a defined thickness) and full skin. Our work was consistent with others who have observed that fits to the Potts–Guy equation are stronger for experiments focused on the epidermis. Permeability estimates for dermatomed skin and full skin resulted in low regression coefficients when compared to epidermis datasets. An updated regression equation uses a combination of fitted permeability values obtained with a published 2D compartmental model previously evaluated. The resulting regression equation was: logKp= −2.55 + 0.65logP − 0.0085MW,R2= 0.91 (applicability domain for all datasets:MWranges from 18 to >584 g/mol and −4 to >5 for logP). This approach demonstrates the advantage of combining mechanistic with structural activity relationships in a single modeling approach. This combination approach results in an improved regression fit when compared to permeability estimates obtained using the Potts–Guy approach alone. The analysis presented in this work assumes a one‐compartment skin absorption route; future modeling work will consider adding multiple compartments.