Molecular Dynamics as a tool for in silico screening of skin permeability

Prediction of skin permeability can have manifold applications ranging from drug delivery to toxicity prediction. Along with the semi-empirical or mechanistic models proposed in the last decades, Molecular Dynamics simulations have recently become a fruitful tool for investigating membrane permeability, in particular as they allow the involved mechanisms to be modelled at a molecular level. Despite their significant structural complexity, Molecular Dynamics simulations can also be utilized to study permeation through the lipid matrix that characterizes the stratum corneum. In this work, Steered Molecular Dynamics simulations are performed on a suitably developed stratum corneum lipid matrix model. Regardless of their actual tortuous path within the stratum corneum, the permeants, taken from a Fully Validated dataset of 80 compounds of known permeability coefficient, are moved through the bilayer along its normal. This allows the exploration of all the possible conformational and physicochemical constraints the molecule experiences when moving through the bilayer. The so performed Steered Molecular Dynamics simulations are then utilized to extract the corresponding lipophilicity and diffusion parameters as computed by subdividing the entire path in 18 regions of different polarity and composition. Correlative analyses showed that the water-lipids interface is the best performing region and that significant enhancements can be gained by including parameters accounting for the temperature effect. Taken together, the developed models possess an enhanced predictive power compared to the existing equations and statistics are approaching the best possible results, given the uncertainty in the utilized permeability data.