In silico and in vitro Blood-Brain Barrier models for early stage drug discovery

Abstract

Drug discovery is a long and costly procedure that requires the prediction of many candidate molecules’ attributes, including ADMET (Absorption, Distribution, Metabolism, Elimination and Toxicity) characteristics. Pharmaceutical companies complementarily use in silico and in vitro models at different stages for this purpose. The permeability across the blood-brain barrier (BBB) is a very important ADMET property since it inhibits the delivery of multiple drugs to the brain. In this context, this paper presents two BBB models designed and implemented for early stage drug discovery: an in silico and an in vitro one. The highest overall accuracy obtained with the former model was 96.23% with both Quadratic Discriminant Analysis and Support Vector Machine classifiers, after applying Genetic Algorithm for feature selection. In the latter case, we have proposed the novel approach of applying cellulose filter papers of 2 μm porosity with PLA 3D printed inserts in order to build a valid in vitro model. Different coatings were tested to increase the adhesion of the endothelial cells to the substrate. The highest degree of confluency was obtained with the collagen type I coating. Moreover, the highest trans-endothelial electric resistance (TEER) value obtained was 45.6 ± 12.07 Ω.cm2 which is comparable to the values reported using the same cell line. This shows that paper-based cell culture can be a promising tool for the implementation of low-cost BBB models that could validate and refine computational models.