QSAR Modeling of the Arylthioindole Class of Colchicine Polymerization Inhibitors as Anticancer Agents

Author(s): Elnaz Habibpour, Shahin Ahmadi*.

Journal Name:Current Computer-Aided Drug Design

Volume 13 , Issue 2 , 2017

Graphical Abstract:


Background: The health and life of humans have been seriously threatened by cancer for a long period and cancer has become the leading disease-related cause of deaths of human population. Natural products such as colchicine and vinblastine inhibit microtubule assembly by preventing tubulin polymerization. GA-MLR is a powerful search technique based on the evolution of biological systems for QSAR modeling. In this paper, we studied QSAR modeling of some arylthioindole class of colchicine polymerization inhibitors as anticancer agents using GA-MLR and stepwise-MLR.

Methods: The chemical structures and experimental values for inhibition of colchicine binding taken from the literature. In the study of inhibition of colchicine binding the total numbers of 49 compounds were split into the training and test sets randomly, which have 39 and 10 compounds, respectively. The Chem3D module was used in order to create the 3D structures of compounds; geometry optimization, using the Polak-Ribiere algorithm. The total numbers of 1185 molecular descriptors such as GETAWAY, RDF, WHIM and 3D-MoRSE descriptors were derived for proper characterizing the structures of arylthioindoles derivatives. These molecular descriptors were reduced to 447. In fact the variables which have low correlation with response, constant variables and also collinear descriptors were eliminated. The random sampling of the training set (80% of data) was performed 20 times and the remaining molecules have been used as external validation set. GA-MLR and S-MLR methods were applied on all random training data sets.

Results: After splitting the data set by RS method, the GA-MLR and S-MLR methods were applied on the training set to select important variables. The best models consist of one, two, three, four, five and six variables created to find the best QSAR model. The best multivariate linear model based on Q2cal and Q2test values had five parameters in both GA-MLR and S-MLR methods.

Conclusion: The results indicate that in this study, the Q2test values are 0.6209 and 0.1144 for GAMLR and S-MLR methods; respectively. According to the results of external validation, we can conclude that the GA-MLR method is more powerful than S-MLR in variable selecting. Also in SAR studies we can conclude that the arylthioindole derivatives with higher density of electrons in C2 position have the largest amounts of IC50. So we can use this important fact to synthesize stronger anticancer agents.

Keywords: QSAR, anticancer, genetic algorithm, colchicine polymerization inhibitors.

Rights & PermissionsPrintExport

Article Details

Year: 2017
Page: [143 - 159]
Pages: 17
DOI: 10.2174/1573409913666170124100810
Price: $58