Optimization of a doxycycline hydroxypropyl-β-cyclodextrin inclusion complex based on computational modeling

To prepare a stable complex of doxycycline (Doxy) and hydroxypropy-β-cyclodextrin (HP-β-CD) for ophthalmic delivery, the optimum formulation and preparation conditions were investigated using response surface methodology (RSM), artificial neural network (ANN) and support vector machine (SVM) modeling. The molar ratios of HP-β-CD/Doxy and Mg2+/Doxy, inclusion time and temperature were selected as independent variables (X1–X4) and inclusion efficiency and stability of the Doxy-HP-β-CD complex were selected as dependent (response) variables (Y1 and Y2). The optimal formulation predicted by genetic algorithm (GA) combined with the models was characterized by microscopy and nuclear magnetic resonance spectrometry, and the stability of Doxy in the complex was evaluated. The highest values of Y1 and Y2 were obtained using an ANN model combined with GA which predicted the values of X1–X4 to be 4, 10.8, 12 h and 25 °C, respectively. The modeling and optimization results indicated that a feed-forward back-propagation ANN with one hidden layer and 10 hidden units showed better fitting to both responses compared to the RSM and SVM models. GA proved to be an efficient tool in multi-objective optimization of a pharmaceutical formulation.

This study compared response surface methodology, artificial neural network and support vector machine modeling combined with genetic algorithm to optimize the formulation of a doxycycline hydroxypropy-β-cyclodextrin complex. It provides a simple and convenient methodology to optimize a pharmaceutical formulation based on multiple inputs and responses.Figure optionsDownload as PowerPoint slide