Mankil Jung, Yongnam Lee, Minjoo Shim, Eunyoung Lim, Eun Jig Lee and Hyun Chul Lee Pages 410 - 419 ( 10 )
A quantitative structure-activity relationship (QSAR) study of aromatic inhibitors against aldose reductase (AR) activity was performed using variable selection from stepwise multiple linear regression (MLR) and genetic algorithm (GA)-MLR. As a result of variable selection, stepwise MLR and GA-MLR gave the same results with one, two, three and five descriptors and different results with four and six descriptors. GA-MLR produced higher values and was better in explanatory and predictive power than stepwise MLR in four variables. AR activity (pIC50) of aromatic derivatives was expressed with acceptable explanatory (74.6-81.2%) and predictive power (68.8-74.4%) in models 3 and 4. The resulting models with the given descriptors illustrate that hydrophobic and electrostatic interactions play a significant role in inhibition of AR activity. This study suggests that the QSAR models can be used as guidelines to predict improved aldose reductase inhibitory activity and to obtain reliable predictions in structurally diverse compounds.
Aldose reductase (AR) inhibitors, Polylol pathway, Quantitative structure-activity relationship (QSAR), Multiple linear regression (MLR), Genetic algorithm (GA), cell membranes, alcohol, glomerulus tissues, peripheral nerve, flavonoids
Department of Chemistry, Yonsei University, Seoul 120-749, Korea.