Federal University of Espirito Santo-CCENS, Alegre, Brazil.
Federal University of ABC (UFABC), Santo Andre, SP, Brazil.
J Mol Model. 2021 Aug 7;27(9):239. doi: 10.1007/s00894-021-04852-8.
Protein kinases (in this case, HER-2 and EGFR) are involved in cancer-related diseases. Some reports have shown unique CoMFA models using the sum of activities expressed as pIC (-log IC), as the classical CoMFA technique would not be the best strategy to construct models for multitarget therapy considering that the molecular alignment will not be the same for different targets. An alternative for this problem is the use of Topomer-CoMFA, a variation of CoMFA, which does not require the alignment step in the generation of 3D models. In this study, we propose the combined use of the sum of activities and Topomer-CoMFA for the construction of a unique dual 3D model considering the inhibitory activities against EGFR and HER-2. For this, 88 compounds from the literature were divided into two groups: training (71) and test (17) sets. The biological activity of each compound, expressed as IC for EGFR and HER-2, was transformed into pIC, summed, and used as the dependent variable in the Topomer-CoMFA analyses. The obtained model was considered statistically robust in the prediction of the dual activity of new compounds. Finally, based on the obtained model, we proposed structural modifications to some of the compounds used to improve the biological data. From the 3D model, we suggested new derivative compounds with improved biological activity for both targets. Therefore, the combination of the techniques proposed in this study proves to be a good strategy to construct better statistical models that can predict biological activities in multitarget systems.
蛋白激酶（此处为 HER-2 和 EGFR）与癌症相关疾病有关。一些报告显示，使用以 pIC（-log IC）表示的活性总和的独特 CoMFA 模型，因为经典的 CoMFA 技术对于构建多靶治疗模型并不是最佳策略，因为分子对准对于不同的靶标不会相同。针对此问题的一种替代方法是使用 Topomer-CoMFA，这是 CoMFA 的一种变体，在生成 3D 模型时不需要对齐步骤。在这项研究中，我们提出了联合使用活性总和和 Topomer-CoMFA 的方法，构建了一个独特的双 3D 模型，考虑了对 EGFR 和 HER-2 的抑制活性。为此，将文献中的 88 种化合物分为两组：训练集（71 种）和测试集（17 种）。每种化合物的生物学活性，以 EGFR 和 HER-2 的 IC 表示，转换为 pIC，求和，并用作 Topomer-CoMFA 分析的因变量。所得模型在预测新化合物的双重活性方面被认为具有统计学稳健性。最后，基于获得的模型，我们对用于改善生物学数据的一些化合物提出了结构修饰建议。从 3D 模型中，我们建议了针对这两个靶标具有改进生物活性的新衍生物化合物。因此，本研究中提出的技术组合被证明是构建能够预测多靶系统中生物活性的更好统计模型的良好策略。
