Predicting MDCK cell permeation coefficients of organic molecules using membrane-interaction QSAR analysis

Aim: To use membrane-interaction quantitative structure-activity relationship analysis (MI-QSAR) to develop predictive models of partitioning of organic compounds in gastrointestinal cells. Methods: A training set of 22 structurally diverse compounds, whose apparent permeability accross cellular membranes of Madin-Darby canine kidney (MDCK) cells were measured, were used to construct MI-QSAR models. Molecular dynamic simulations were used to determine the explicit interaction of each test compound (solute) with a dimyristoyl-phosphatidyl-choline monolayer membrane model. An additional set of intramolecular solute descriptors were computed and considered in the trial pool of descriptors for building MI-QSAR models. The QSAR models were optimized using multidimensional linear regression fitting and the stepwise method. A test set of 8 compounds were evaluated using the MI-QSAR models as part of a validation process. Results: MI-QSAR models of the gastrointestinal absorption process were constructed. The descriptors found in the best MI-QSAR models are as follows: 1) ClogP (the logarithm of the 1-octanol/water partition coefficient); 2) E_HOMO (the highest occupied molecular orbital energy); 3) E_s (stretch energy); 4) PM_Y (the principal moment of inertia Y, the inertia along the y axis in the rectangular coordinates; 5) C_t (total connectivity); and 6) E_nb (the energy of interactions between all of the non-bonded atoms). The most important descriptor in the models is ClogP. Conclusion: Permeability is not only determined by the properties of drug molecules, but is also very much influenced by the molecule-membrane interaction process.