Evaluation of Constrained and Restrained Molecular Dynamics Simulation Methods for Predicting Skin Lipid Permeability

Recently, molecular dynamics (MD) simulations have been utilized to investigate the barrier properties of human skin stratum corneum (SC) lipid bilayers. Different MD methods and force fields have been utilized, with predicted permeabilities varying by few orders of magnitude. In this work, we compare constrained MD simulations with restrained MD simulations to obtain the potential of the mean force and the diffusion coefficient profile for the case of a water molecule permeating across an SC lipid bilayer. Corresponding permeabilities of the simulated lipid bilayer are calculated via the inhomogeneous solubility diffusion model. Results show that both methods perform similarly, but restrained MD simulations have proven to be the more robust approach for predicting the potential of the mean force profile. Critical to both methods are the sampling of the whole trans-bilayer axis and the following symmetrization process. Re-analysis of the previously reported free energy profiles showed that some of the discrepancies in the reported permeability values is due to misquotation of units, while some are due to the inaccurately obtained potential of the mean force. By using the existing microscopic geometrical models via the intercellular lipid pathway, the permeation through the whole SC is predicted from the MD simulation results, and the predicted barrier properties have been compared to experimental data from the literature with good agreement.Two Supplementary Files. 1: Python script to calculate lateral and perpendicular diffusion profiles and potential of the mean force via pACF, Kubo, MSD, and PMcF methods as explained in the main paper. (TXT). 2: Check of system equilibration; derivation of partition coefficient; all diffusion, force and potential of the mean force profiles symmetrized and unsymmetrized for all the data sets; WHAM convergence; permeability data from the literature and temperature correction; and python script README. (PDF)