Biomimetic chromatographic techniques are designed to mimic biological environments - cell membrane lipids, plasma proteins, or cell fragments - and are commonly used to study physicochemical and biological properties of compounds. The affinity of solutes for the lipid bilayers of cell membranes is studied in particular using chromatography on Immobilized Artificial Membrane (IAM) stationary phases or Biopartitioning Micellar Chromatography (BMC). Immobilized Plasma Protein Chromatography utilizes sorbents coated with key human proteins to directly study protein-solute binding.
In this study, 45 structurally unrelated compounds were analyzed in the context of their blood-brain barrier permeability and plasma protein binding using two different types of biomimetic chromatography. Immobilized Artificial Membrane and Biopartitioning Micellar Chromatographic retention parameters were used in conjunction with calculated descriptors. Special attention was paid to the blood-brain barrier (BBB) permeability of compounds in connection with their availability from the gastrointestinal tract (GI). New Multiple Linear Regression and Artificial Neural Network models for solutes’ blood- and brain-barrier permeability and blood plasma protein binding were proposed. The studied compounds were assigned to three groups: GI- and BBB-permeable (GI1+BBB1); GI-permeable but not crossing the BBB (GI1+BBB0); not available from the GI tract and not BBB-permeable (GI0+BBB0), using Discriminant Analysis and Classification Trees, with fairly similar results. The CT classification was based on just three descriptors, which were easily computed or derived from IAM chromatography. All models proposed in this study were validated.