Identifying Potential Dual Inhibitory Candidates for Non-Small Cell Lung Cancer through Molecular Docking, 3D-QSAR Pharmacophore-based Virtual Screening, Comparative Molecular Field and Similarity Indices Analysis Modeling
Xiao X. Zhu
School of Chemical Engineering, Sichuan University, Chengdu 610065, China.
Yan Li
Department of Pharmaceutical Engineering, Sichuan University, Chengdu 610041, China.
Xuan R. Zhang *
School of Chemical Engineering, Sichuan University, Chengdu 610065, China.
Lun Yuan
Department of Pharmaceutical Engineering, Sichuan University, Chengdu 610041, China.
Pei H. Luo
School of Chemical Engineering, Sichuan University, Chengdu 610065, China.
X. Gao
School of Chemical Engineering, Sichuan University, Chengdu 610065, China and Department of Pharmaceutical Engineering, Sichuan University, Chengdu 610041, China.
Zhong S. Tan
Department of Pharmaceutical Engineering, Sichuan University, Chengdu 610041, China.
*Author to whom correspondence should be addressed.
Abstract
Aim: This work aims to understand potential inhibitory structural requirements and identify lead compounds for non-small cell lung cancer through 3D-QSAR pharmacophore-based virtual screening, molecular docking, CoMSIA and CoMFA QSAR modelling.
Materials and Methods: QSAR pharmacophore models were developed by HypoGen Module and validated by test data set, Fischer’s randomization and Guner-Henry equation. The well-validated pharmacophore model was employed to perform virtual screening to identify potent hits from ZINC database. The retrieved hits were subsequently subjected to filtering using ADMET descriptors and Lipinski’s Rule of Five. CoMSIA and CoMFA were then utilized to produce QSAR models on phenylpyrimidine derivatives also.
Results: Validations on 3D-QSAR pharmacophore model indicate that the enrichment factor is 6.34, GH is 0.517 and a correlation coefficient is 0.83, implying its highly predictive ability. Top three hits: ZINC29356266, ZINC06589615, and ZINC03375633 were identified as promising potent inhibitory candidates with IC50 value of about 0.54 µM and fitness value of about 59.4. Interestingly, the top three hits indicate dual inhibitory activity targeting EGFR and PD-L1 from structure-based docking. Two developed QSAR models from CoMSIA and CoMFA modelling indicate a potential predictive ability (q2=0.67, and 0.71 respectively). The designed compound C indicates a more potential (dual) inhibitory activity (pIC50=7.39) targeting EGFR (fitness=59.78) and CTLA-4 (fitness =47.90).
Conclusion: Validations indicate that the developed 3D-QSAR pharmacophore model is highly predictive. Top three hits were identified as promising potent inhibitory candidates and indicated dual inhibitory activity targeting EGFR and PD-L1. The designed compound C indicates a more potential (dual) inhibitory activity targeting EGFR and CTLA‐4. These important 3D-QSAR and molecular docking bioinformatics results achieved from this work should be valuable in designing more promising potent inhibitory candidates and developing novel lead compounds against advanced NSCLC in future.
Keywords: Bioinformatics, molecular docking, QSAR, pharmacophore, comparative molecular similarity indices analysis.