Abstract Nowadays screening methods for assessment of biological properties of drug candidates are highly desired. Among naturally occurring membranes, one of the most important is the blood–brain barrier (BBB). BBB plays crucial role for central nervous system active drug candidates, since each molecule targeting a receptor in the brain must pass through the BBB first. This work assesses the possibility to apply micellar electrokinetic chromatography (MEKC) with cetrimonium bromide (CTAB) as surfactant in order to predict logBB. A model set of 45 marketed drugs, with known logBB values, varies in terms of chemical structures and pharmacological activities was used in this study. The established models were additionally supported by P_VSA-like descriptors molecular descriptors calculated by Dragon software. Two regression methods, multiple linear regression and support vector machine were evaluated and compared in terms of effectives of prediction of logBB. Both models showed similar prediction power, evidenced by similar values of root-mean-squared error of cross-validation 0.310 and 0.314 respectively. Proposed models met the Tropsha et al. criteria R2 > 0.6 and Q2 > 0.5 These results indicate that obtained model can be useful to predict BBB permeability of drug candidates and are attractive alternatives of time-consuming and demanding direct methods for log BB measurement.

中文翻译：

---

使用胶束电动色谱和 P_VSA 样描述符评估血脑屏障通透性

摘要 目前，人们非常需要用于评估候选药物生物学特性的筛选方法。在天然存在的膜中，最重要的膜之一是血脑屏障 (BBB)。BBB 对中枢神经系统活性候选药物起着至关重要的作用，因为每个靶向大脑中受体的分子都必须首先通过 BBB。这项工作评估了应用胶束电动色谱 (MEKC) 和西曲溴铵 (CTAB) 作为表面活性剂以预测 logBB 的可能性。本研究使用了 45 种上市药物的模型集，这些药物具有已知的 logBB 值，在化学结构和药理活性方面各不相同。建立的模型还得到了由 Dragon 软件计算的 P_VSA 样描述符分子描述符的支持。两种回归方法，多元线性回归和支持向量机在logBB的预测效果方面进行了评估和比较。两种模型都显示出相似的预测能力，交叉验证的均方根误差分别为 0.310 和 0.314 的相似值证明了这一点。提出的模型满足了 Tropsha 等人的要求。标准 R2 > 0.6 和 Q2 > 0.5 这些结果表明，获得的模型可用于预测候选药物的 BBB 渗透性，并且是耗时且要求苛刻的 log BB 测量直接方法的有吸引力的替代方法。提出的模型满足了 Tropsha 等人的要求。标准 R2 > 0.6 和 Q2 > 0.5 这些结果表明，获得的模型可用于预测候选药物的 BBB 渗透性，并且是耗时且要求苛刻的 log BB 测量直接方法的有吸引力的替代方法。提出的模型满足了 Tropsha 等人的要求。标准 R2 > 0.6 和 Q2 > 0.5 这些结果表明，获得的模型可用于预测候选药物的 BBB 渗透性，并且是耗时且要求苛刻的 log BB 测量直接方法的有吸引力的替代方法。