Jevrić Lidija R, Karadžić Milica Ž, Mandić Anamarija I, Podunavac Kuzmanović Sanja O, Kovačević Strahinja Z, Nikolić Andrea R, Oklješa Aleksandar M, Sakač Marija N, Penov Gaši Katarina M, Stojanović Srđan Z
Faculty of Technology Novi Sad, University of Novi Sad, Bulevar Cara Lazara 1, 21000 Novi Sad, Serbia.
Faculty of Technology Novi Sad, University of Novi Sad, Bulevar Cara Lazara 1, 21000 Novi Sad, Serbia.
J Pharm Biomed Anal. 2017 Feb 5;134:27-35. doi: 10.1016/j.jpba.2016.11.015. Epub 2016 Nov 10.
The present paper deals with chromatographic lipophilicity determination of twenty-nine selected steroid derivatives using reversed-phase high-performance liquid chromatography (RP HPLC) combined with two mobile phase, acetonitrile-water and methanol-water. Chromatographic behavior of four groups (triazole and tetrazole, toluenesulfonylhydrazide, nitrile and dinitrile and dione) of selected steroid derivatives was studied. Investigated compounds were grouped using principal component analysis (PCA) according to their logk values for both mobile phases. Grouping was in the very good accordance with the polarity and lipophilicity of the investigated compounds. QSRR (quantitative structure-retention relationship) approach was used to model chromatographic lipophilicity behavior using molecular descriptors. Modeling was performed using linear regression (LR) and multiple linear regression (MLR) methods. The most influential molecular descriptors were lipophilicity descriptors that are important for molecules ability to pass through biological membranes and geometrical descriptors. All established LR-QSRR and MLR-QSRR models were statistically validated by standards, cross- and external validation parameters as well as with two graphical methods. According to all these assessments, MLR models were better for chromatographic lipophilicity prediction. It was shown that chromatographic systems with methanol-water were better for modeling of logk than systems with acetonitrile-water, as well as the systems that contained lower volume fractions of organic component in mobile phase. Modeling was performed in order to obtain lipophilicity profiles of investigated compounds as future drug candidates of biomedical importance.
本文采用反相高效液相色谱（RP HPLC）结合乙腈 - 水和甲醇 - 水两种流动相，对29种选定的甾体衍生物进行了色谱亲脂性测定。研究了选定甾体衍生物的四组（三唑和四唑、甲苯磺酰肼、腈和二腈以及二酮）的色谱行为。根据所选甾体衍生物在两种流动相中的logk值，使用主成分分析（PCA）对研究的化合物进行分组。分组与所研究化合物的极性和亲脂性非常吻合。采用定量结构 - 保留关系（QSRR）方法，利用分子描述符对色谱亲脂性行为进行建模。使用线性回归（LR）和多元线性回归（MLR）方法进行建模。最具影响力的分子描述符是对分子穿过生物膜的能力很重要的亲脂性描述符和几何描述符。所有建立的LR - QSRR和MLR - QSRR模型均通过标准、交叉验证和外部验证参数以及两种图形方法进行了统计验证。根据所有这些评估，MLR模型在色谱亲脂性预测方面表现更好。结果表明，与乙腈 - 水体系相比，甲醇 - 水色谱体系更适合对logk进行建模，并且流动相中有机成分体积分数较低的体系也是如此。进行建模是为了获得作为具有生物医学重要性的未来药物候选物的研究化合物的亲脂性概况。
Zotero 插件