Structure-based in silico model profiles the binding constant of poorly soluble drugs with β-cyclodextrin

Cyclodextrin inclusion complexation technique is the key method to enhance the solubility and absorption of poorly soluble drugs in the early development stage, and thus it is essential to predict the binding constant between drug molecules and cyclodextrin. Structure-based in silico model was constructed for a data set of 86 poorly soluble drugs and used to profile the binding constant of drug-β-cyclodextrin inclusion complex. The stepwise regression was employed to select the optimum subset of the independent variables. The in silico model was built by the multiple linear regression method and validated by the residual analysis, the normal Probability-Probability plot and Williams plot. For the entire data set, the R(2) and Q(2) of the model were 0.78 and 0.67, respectively. The results indicated that the fitted model is robust, stable and satisfies all the prerequisites of the regression models. The chemical space position and important contributors were compared between selected drug molecules and organic compounds available in the literature. It was suggested that the binding behavior of drug molecules with β-CD should differ from that of the common organic compounds. Focusing on structurally diverse drugs, the in silico model can be used as an efficient tool to rapidly screen the drug-β-cyclodextrin inclusion complex stability and to rationally design the new drug delivery system of poorly soluble drugs.