Comparison of a Rat Primary Cell-Based Blood-Brain Barrier Model With Epithelial and Brain Endothelial Cell Lines: Gene Expression and Drug Transport

Szilvia Veszelka,Monika Vastag,Andrea E Tóth,Ilona Gróf,Éva Hellinger,András Tóth,Alexandra Bocsik,Mária Mészáros,Fruzsina R Walter,Mária A Deli,Gábor Rákhely

doi:10.3389/fnmol.2018.00166

Abstract

Cell culture-based blood-brain barrier (BBB) models are useful tools for screening of CNS drug candidates. Cell sources for BBB models include primary brain endothelial cells or immortalized brain endothelial cell lines. Despite their well-known differences, epithelial cell lines are also used as surrogate models for testing neuropharmaceuticals. The aim of the present study was to compare the expression of selected BBB related genes including tight junction proteins, solute carriers (SLC), ABC transporters, metabolic enzymes and to describe the paracellular properties of nine different culture models. To establish a primary BBB model rat brain capillary endothelial cells were co-cultured with rat pericytes and astrocytes (EPA). As other BBB and surrogate models four brain endothelial cells lines, rat GP8 and RBE4 cells, and human hCMEC/D3 cells with or without lithium treatment (D3 and D3L), and four epithelial cell lines, native human intestinal Caco-2 and high P-glycoprotein expressing vinblastine-selected VB-Caco-2 cells, native MDCK and MDR1 transfected MDCK canine kidney cells were used. To test transporter functionality, the permeability of 12 molecules, glucopyranose, valproate, baclofen, gabapentin, probenecid, salicylate, rosuvastatin, pravastatin, atorvastatin, tacrine, donepezil, was also measured in the EPA and epithelial models. Among the junctional protein genes, the expression level of occludin was high in all models except the GP8 and RBE4 cells, and each model expressed a unique claudin pattern. Major BBB efflux (P-glycoprotein or ABCB1) and influx transporters (GLUT-1, LAT-1) were present in all models at mRNA levels. The transcript of BCRP (ABCG2) was not expressed in MDCK, GP8 and RBE4 cells. The absence of gene expression of important BBB efflux and influx transporters BCRP, MRP6, -9, MCT6, -8, PHT2, OATPs in one or both types of epithelial models suggests that Caco-2 or MDCK models are not suitable to test drug candidates which are substrates of these transporters. Brain endothelial cell lines GP8, RBE4, D3 and D3L did not form a restrictive paracellular barrier necessary for screening small molecular weight pharmacons. Therefore, among the tested culture models, the primary cell-based EPA model is suitable for the functional analysis of the BBB.

Highlights

The development and introduction of novel neuropharmaceuticals lags behind other groups of medicines, which is partially due to the poor central nervous system (CNS) pharmacokinetics (Banks, 2016)
Tight Junction Proteins Primary rat brain endothelial cells grown in co-culture with glial cells and pericytes (EPA) produced high levels of mRNA for key tight junction proteins such as claudin-5 (CLDN5), occludin and the endothelial cells specific adhesion molecule endothelial cell specific adhesion molecule (ESAM) (Figure 1)
Caco-2 epithelial cells showed a high level of expression for occludin, while in Madin-Darby canine kidney (MDCK) cells it was lower as compared to both EPA and Caco-2 models (Supplementary Figure S1)

Summary

Introduction

The development and introduction of novel neuropharmaceuticals lags behind other groups of medicines, which is partially due to the poor central nervous system (CNS) pharmacokinetics (Banks, 2016). One of the reasons for the low number of CNS active drugs in clinical use is the restricted penetration of most drugs across the blood-brain barrier (BBB; Pardridge, 2015). The four main mechanisms at the level of the BBB to limit drug transport are: (i) the restricted paracellular pathway regulated by interendothelial tight junctions (TJ); (ii) the low level of non-specific transendothelial vesicular traffic; (iii) active efflux transporters which deliver metabolites from brain to blood and prevent the entry of xenobiotics and drugs to the CNS; and (iv) enzymes which metabolize drug molecules (Deli, 2011; Banks, 2016). Models that predict brain penetration are valuable tools to study and develop new targeted nanoparticles that cross the BBB (Veszelka et al, 2015). There are several types of models for BBB permeability from in silico approaches to in vivo studies (Vastag and Keseru, 2009; Veszelka et al, 2011; Avdeef et al, 2015)

Objectives

Methods

Results

Conclusion