Antitumor structure–activity relationship in bis-stannoxane derivatives from pyridine dicarboxylic and benzoic acids

• Published in: Inorganica Chimica Acta 2012-09, Vol.392, p.229-235
• Main Authors: Valcarcel, José Antonio, Razo-Hernández, Rodrigo Said, Valdez-Velázquez, Laura Leticia, García, Manuel Villanueva, Organillo, Ángel Andrés Ramos, Vázquez-Vuelvas, Oscar F., García Ruiz, Miguel A., Gómez-Sandoval, Zeferino
• Format: Article
• Language: English
• Subjects: Genetic algorithms; MCF-7; Neural networks; Organotin compounds; QSAR; WiDr
• ISSN: 0020-1693; 1873-3255
• DOI: 10.1016/j.ica.2012.06.029

This study estimated four quantitative structure–activity relationship (QSAR) descriptive models relating antitumor biological activity essays (ID50) in breast (MFC-7) and colon cancer (WiDr) cell lines with stannoxanes compounds derived from 2,6-pyridine-dicarboxylates [C5H3N(COO)2SnRR′] (R, R′=alkyl, aryl) and di-n-butyltinbis-benzoates [(C6H5COO)2SnBu2]. [Display omitted] ► The hybrid methodology proposed was adequate to obtain QSAR-type models. ► The electronic factor is much important for the antitumor activity in MFC-7. ► Small volume and lipophilic factor are clews for antitumor activity in WiDr. ► Our descriptor ΔD2 help us to explain antitumor activity in WiDr. Organometallic compounds have been proposed as new drugs for cancer treatment due to the large metal mesh with DNA. This study estimated four quantitative structure–activity relationship (QSAR) descriptive models relating antitumor biological activity essays (ID50) in breast (MFC-7) and colon cancer (WiDr) cell lines with stannoxanes compounds derived from 2,6-pyridine-dicarboxylates [C5H3N(COO)2SnRR′] (R, R′=alkyl, aryl) and di-n-butyltinbis-benzoates [(C6H5COO)2SnBu2]. A series of thermodynamic, structural and molecular descriptors were calculated from the geometric and electrical optimizations (PM3) of the two organotin series, in order to correlate the biological activity of the tin esters. The use of genetic algorithms and multilinear correlations yielded four mathematical models, each one with four descriptors related to molecular, area/volume, lipophilicity, and molecular dipole polarizability. These descriptors were entered into a back-propagation neural network to obtain theoretical descriptive models with the goal of proposing the development of new organotin molecules with enhanced antitumor activity.