QSAR studies of macrocyclic diterpenes with P-glycoprotein inhibitory activity

Multidrug resistance (MDR) represents a major limitation for cancer chemotherapy. There are several mechanisms of MDR but the most important is associated with P-glycoprotein (P-gp) overexpression. The development of modulators of P-gp that are able to re-establish drug sensitivity of resistant cells has been considered a promising approach for overcoming MDR. Macrocyclic lathyrane and jatrophane-type diterpenes from Euphorbia species were found to be strong MDR reversing agents. In this study we applied quantitative structure–activity relationship (QSAR) methodology in order to identify the most relevant molecular features of macrocyclic diterpenes with P-gp inhibitory activity and to determine which structural modifications can be performed to improve their activity. Using experimental biological data at two concentrations (4 and 40 μg/ml), we developed a QSAR model for a set of 51 bioactive diterpenic compounds which includes lathyrane and jatrophane-type diterpenes and another model just for jatrophanes. The cross-validation correlation values for all diterpenes QSAR models developed for biological activities at compound concentrations of 4 and 40 μg/ml were 0.758 and 0.729, respectively. Regarding the prediction ability, we get Rpred2 values of 0.765 and 0.534 for biological activities at compound concentrations of 4 and 40 μg/ml, respectively. Applying the cross-validation test to jatrophanes QSAR models, we obtained 0.680 and 0.787 for biological activities at compound concentrations of 4 and 40 μg/ml concentrations, respectively. For the same concentrations, the obtained Rpred2 values for jatrophanes models were 0.541 and 0.534, respectively. The obtained models were statistically valid and showed high prediction ability.

Quantitative structure–activity relationship (QSAR) methodology was applied to macrocyclic diterpenoids (jatrophanes and lathyranes) in order to identify the most relevant molecular features for the P-gp inhibitory activity. QSAR studies were performed for two concentrations (4 and 40 μg/mL), the general model at concentration 4 μg/mL showed the best prediction ability and the molecular descriptors involved in this model are easier to interpret and, consequently, the findings here reported could be translated into synthesis of novel compounds.Figure optionsDownload high-quality image (70 K)Download as PowerPoint slide