Blood-Brain Barrier In Vitro Models and Their Application in Toxicology

The Report and Recommendations of
ECVAM Workshop 491,2

Reprinted with minor amendments from ATLA 32: 37-50.

Pilar Prieto,3 Bas J. Blaauboer,4 Albertus Gerrit de Boer,5 Monica Boveri,3 Romeo Cecchelli,6 Cecilia Clemedson,7 Sandra Coecke,3 Anna Forsby,8 Hans-Joachim Galla,9 Per Garberg,10 John Greenwood,11 Anna Price3 and Hanna Tähti12

3ECVAM, Institute for Health & Consumer Protection, European Commission Joint Research Centre, Ispra, Itaiy; 4IRAS, Utrecht University, The Netherlands; 5LACDR, University of Leiden, The Netherlands; 6Laboratorie de Physiopathologie de la Barrière Hemato-Encophalique, Unité Mixte EA 2465 Université d'Artois-Institut Pasteur de Lille, Faculté des Sciences Jean Perrin, Lens, France; 7Expertrådet ECB, Sundbyberg, Sweden; 8Department of Neurochemistry and Neurotoxicology, Stockholm University, Sweden; 9Institut für Biochemie, Universität Müster, Germany; 1OBiovitrum AB, Stockholm, Sweden; 11Institute of Ophthalmology, University College London, UK; 12Medical School, University of Tampere, Finland

1ECVAM - The European Centre for the Validation of Alternative Methods.

2This document represents the agreed report of the participants as individual scientists.

Address for correspondence: P. Prieto, ECVAM, Institute for Health & Consumer Protection, European Commission Joint Research Centre, 21020 ispra (VA), Italy. E-mail: maria.prieto-pilar@jrc.it.

Address for reprints: ECVAM, JRC Institute for Health & Consumer Protection, TP 580, 21020 Ispra (VA), Italy.

Preface

This is the report of the forty-ninth of a series of workshops organised by the European Centre for the Validation of Alternative Methods (ECVAM). ECVAM's main goal, as defined in 1993 by its Scientific Advisory Committee, is to promote the scientific and regulatory acceptance of alternative methods which are of importance to the biosciences and which reduce, refine or replace the use of laboratory animals. One of the first priorities set by ECVAM was the implementation of procedures which would enable it to become well-informed about the state-of-the-art of nonanimal test development and validation, and the potential for the possible incorporation of alternative tests into regulatory procedures. It was decided that this would be best achieved by the organisation. of ECVAM workshops on specific topics, at which small groups of invited experts would review the current status of in vitro tests and their potential uses, and make recommendations about the best ways forward (1).

This ECVAM workshop on in vitro models for studying the blood-brain barrier (BBB) and their application in toxicology was held at ECVAM on 19-21 May 2003. The workshop was chaired by Bas Blaauboer and was attended by pharmacologists and toxicologists from academia and industry, including experts on the BBB and in neurotoxicology. The current status of in vitro models of the BBB was discussed, focusing on their application in toxicology in general and, in particular, in regulatory toxicology in the context of the European Union (EU) chemicals policy (2) and the Seventh Amendment to the Cosmetics Directive (3). An important aspect of the workshop was consideration of the inclusion of biokinetic modelling, the BBB and target organ toxicity in integrated testing strategies.

 

Introduction

The BBB separates the brain from the systemic blood circulation and plays an important role in maintaining the homeostasis of the central nervous system (CNS). Another cellular barrier present in the CNS is the bloodcerebrospinal fluid (CSF) barrier, which is formed by a continuous layer of epithelial cells lying on top of the endothelial cells of the choroid plexus. The BBB has a larger surface area than the blood-CSF barrier, so it is considered to be the primary interface between the CNS and the peripheral circulation; the blood-CSF barrier is considered to play a secondary role (4, 5). In addition, since reliable in vitro models of the blood-CSF barrier have not been available until recently, this barrier has not been studied to the same extent as the BBB. However, some excellent in vitro models are now available (6-8). Since the effective surface area may have been underestimated due to the strong curvature of the plexus, and as a result of increasing knowledge about specialised and highly specific transporters in the plexus epithelium, more attention is now being paid to the blood-CSF barrier.

The BBB is, formed by brain capillary endothelial cells that are characterised by tight junctions and the absence of fenestrations and pinocytotic vesicles. It is well known that astrocytes and the supporting pericytes are important for the expression of the BBB phenotype. Moreover, some studies suggest a role for neurons in mediating cell-cell stimuli involved in the development and differentiation of the BBB (9, 10).

The optimal functioning of the brain requires a strict homeostasis of the neuronal environment and an intact barrier. An altered permeability of the BBB is observed in several diseases of the CNS. Toxic agents and pathological conditions also cause changes in BBB function. Effects on the barrier function are of importance during inflammation and certain disease states. For example, brain tumours may cause oedema, and even though no other primary effects on the barrier have been reported, this may result in secondary effects and damage to the brain (11, 12). Inflammatory reactions and the release of inflammatory cytokines which affect barrier function may be of importance during repeated-dose toxicity studies.

The BBB in the context of the biokinetic behaviour of chemicals in organisms

Knowledge of biokinetics is essential in order to understand the biological effects of compounds. A description of the course of the concentration over time in the various body tissues is needed when studying the toxicity of a chemical, since toxicity is generally related to the concentration at the site of action, as well as to the length of this local exposure. Knowledge of the partitioning of compounds between blood and tissues will be necessary for a full understanding of the concentration-time-course for a chemical at the site of its toxicological action.

Biokinetic descriptions can be made by making use of biokinetic models. Ideally, such models are based on a thorough knowledge of the physiology of the organism, as well as on the physicochemical properties of the compound under study, and are constructed in the form of a set of algorithms describing blood flows, tissue volumes and partition coefficients (for example, between blood and the tissues, or between blood and air). These models are usually depicted as physiologically based biokinetic (PBBK) models (13).

Knowledge of the concentration-time-course of a compound is a prerequisite to a full understanding of the activity of the compound in the brain (including its neurotoxicity). An assessment of the partitioning between blood and brain is therefore needed. The specific characteristics of the BBB must be taken into account. The tight junctions restrict passive diffusion to those chemicals that can pass across the cellular membranes, which implies that lipophilicity, molecular weight and electrical charge are the determining properties. A number of studies describe algorithms for quantitative structure-property relationships (QSPRs), in which the lipophilicity of a compound, in particular, has been shown to be the most prominent property for describing blood-tissue partitioning (14-16). In addition, the BBB contains a number of active transport systems that can transfer compounds across the barrier. These include systems for active uptake into the CNS, as well as for active export. Therefore, the systems that can be used for estimating BBB passage of compounds include:

  1. Evaluations of the physicochemical properties of compounds and their application in QSPRs. This can be used to describe passive diffusion. Until now, however, active transport could not be estimated on the basis of QSPRs.
  2. Cellular systems consisting of confluent cell layers in culture conditions that allow the measurement of transport across the barrier. A number of these systems are described below. Ideally, these systems should contain the characteristics of the BBB, including the active transporters.

The need for in vitro and other non-animal systems for determining BBB transport

Neurotoxicity is one of the main target-organ toxicities, so there is a need for alternative methods to replace the current animal testing. As was concluded in the ECVAM report on alternative methods for chemicals testing (17), a reduction in the number of animals used in neurotoxicity testing could be achieved by using appropriate testing strategies. This could include assessment of basal cytotoxicity, studies on nerve cell morphology, and the determination of cell physiology and neurochemical functions (18). The outcome from a proposed test battery, in combination with barrier function and metabolism studies, can be used to identify substances of concern in relation to the EU chemicals policy (19). Only 3-28% of all chemicals have effects on the CNS. However, even if only a small amount of a compound passes the BBB, the effects on the CNS can be severe if the compound binds with high affinity to the target. Although it is not clear whether neurotoxicity tests should be carried out before or after testing the passage across the BBB, both chemicals that cross the BBB and those for which there is enough evidence to suggest that they are likely to be neurotoxic, should be tested for neurotoxicity.

During the workshop, it was concluded that models and criteria for neurotoxicity testing in vitro need to be addressed, but this will be the topic for an additional workshop. However, it was pointed out that information regarding concentration levels that affect neuronal activity is of importance for planning in vitro experiments on passage across the BBB and for the selection of the relevant concentrations to be used. On the other hand, it would be of value to have the permeability data available (together with expected exposure levels) when planning and selecting relevant concentrations to be used in further in vitro neurotoxicity studies.

 

State-of-the-art of Methods for Measuring BBB Passage

The BBB is a complex dynamic organ (20). The establishment of in vitro models of the BBB is important for two main reasons. There is interest in predicting the penetration of drug candidates and/or potentially toxic compounds through the BBB, and there is also a need to know more about the modulation of the phenotypic characteristics of capillary endothelial cells. To study all these aspects with in vivo models is difficult, so several in vitro models have been developed (21). There are large quantitative and qualitative differences between the various BBB cell culture systems, but there are some minimal requirements for any BBB model to be useful: a) the presence of restrictive paracellular permeability; b) the possession of a physiologically realistic cell architecture; c) expression of the functional transporter mechanisms present in vivo; and d) ease of culture.

The in vitro and other non-animal models of the BBB can be grouped as follows:

BBB culture systems have many applications in drug research. They can be used for transport studies, the visualisation of drug transport routes, drug transporter studies, drug interaction studies, drug targeting studies, drug metabolism studies, safety pharmacology studies, studies on BBB functionality in diseases, and drug discovery studies. Moreover, BBB in vitro models can also be used in toxicological studies. The choice of model depends on the intended final application. The most representative in vitro BBB models were presented and discussed during the workshop. In addition, the minimal requirements for an in vitro model of the BBB to be used in toxicology, and in particular, in regulatory toxicology, were also discussed.

In vivo models

Many different in vivo and in situ models for studies on transport across the BBB are available. It is therefore important to take into consideration the type of data that are needed, as well as the type of data generated by in vivo studies, when in vitro versus in vivo correlations are sought, in order to make such correlations valid. It is also important to consider that several parameters play key roles in determining the concentration of a molecule in the brain (for example, the endogenous concentration of the molecule, first hepatic passage, specific transporters, metabolism and binding to proteins), and to ensure that they are included when in vivo models are used.

Primary cultured endothelial cells

Primary cultures of brain vascular endothelial cells have been derived from various sources, namely, bovine, mouse, rat, porcine, non-human primate and human. The starting material is always brain microvessels from the cortex (grey matter). It is important to note that, although many people use the term capillaries in this context, they are not able to differentiate between capillaries, arterioles and venules in their vessel isolation. Therefore, all the derived cell lines should be called microvessel lines.

Brain microvessel endothelial cells can be isolated by mechanical dispersion (homogenisation, filtration, sieving, centrifugation), by enzymatic procedures using collagenage or dispase, by a combination of mechanical dispersion and enzymatic digestion, by differential seeding of brain microvessels in culture flasks, by selective outgrowth of brain microvascular endothelial cells, or by selective isolation of cultured cells from contaminating cells. In primary cultures, one can achieve more than 95% "purity", and the endothelial cells conserve BBB properties. The main drawbacks are that the procedure is labour intensive and, in general, there is variability between isolations due to variability in the purity of the preparations and differences in the culture conditions used (21, 22). Variability also occurs because of the species of origin used. Finally, there is a gradual loss of BBB properties during subpassaging, due to increasing contamination of the cultures by pericytes, so that the cells should be used only at very low passage numbers (2 to a maximum of 3). However, cells can be kept deep frozen for a long time. Méesse and coworkers (23) have developed an original technique to obtain pure capillary endothelial cells.

Cryopreserved lines of bovine brain endothelial cells are now commercially available (24). Their BBB properties are retained for up to seven passages. The cells are co-cultured with primary rat astrocytes, and the sucrose permeability values showed a good reproducibility among the clones within a laboratory and between two laboratories. Under these culture conditions, endothelial cells retain most of the in vivo BBB characteristics: highly differentiated tight junctions; a high electrical resistance around 800Ω.cm2; permeabilities for hydrophilic molecules such as sucrose or inulin of 4 x 10-6 cm/second and 0.7 X 10-6 cm/second, respectively (24); specific transporters for molecules such as amino acids and glucose (25); specific BBB, enzyme activities (γ-glutamyl transpeptidase [γ-GTP], monoamine oxidase) and P-glycoprotein (P-gp; 26); and specific receptors for blood-borne molecules allowing their transcytosis across the endothelial monolayers (27-29).

Comparative studies on the transport of a large number of drugs revealed a close correlation between the values obtained with endothelial cell monolayers and in vivo (30, 31). All these data show that this BBB model closely mimics the in vivo situation by reproducing some of the complexities of the cellular environment that exist in vivo, while retaining the experimental advantages associated with tissue culture. A miniaturisation of this co-culture BBB model, by scaling down to 24-well plates, is under evaluation. The ultimate goal will be to use nonradiolabelled molecules and to use a fluorescent integrity marker to study the transport of a molecule and its toxicity in the same insert well. At the moment, a commercial kit (CTbovial@screenpack) is available, composed of microvessel endothelial cells (passage 4), glial cells, collagen, serum, and technical sheets with permeability coefficients for sucrose and P-gp substrate. In this model, the concentration used for each compound is based on the detection limits of the analytical technique used.

Another model of the BBB consists of porcine primary endothelial cells grown in serum-free medium supplemented with hydrocortisone (32). Under these conditions, the endothelial cells display high transendothelial electrical resistance (TEER) values of 1800Ω.cm2, and very low sucrose permeability values of 10-7 cm/second. The morphological features of the endothelial cells are maintained, and the relevant transporters, such as P-gp and a brain multidrug resistance protein (BMDP; 33, 34), are expressed. The cells retain their functional BBB properties for a maximum of 3 days in culture (35).

Bovine brain microvessel endothelial cells co-cultured with astrocytes in the presence of glucocorticoids (dexamethasone and RU28362) also display BBB properties (36).

A schematic representation of the use of inserts for co-cultures is shown in Figure 1. Endothelial cells are grown on permeable filter inserts, with astropytes on the bottom surfaces of the wells.

Figure 1: Schematic representation of the co-culture system

Endothelial cells are plated on the porous membrane of the insert and the astrocytes on the bottom surface of the well.

Immortalised endothelial cells

A number of immortalised brain endothelial cell lines now exist (Table 1). There are also several methods for immortalisation (naked plasmid or viral infection), and it is difficult to establish which is the most appropriate method for transfection or which immortalising genes or promoters to use. Genetically engineered brain endothelial cell lines are easy to culture, and they overcome the problem of supply, especially for use in high-throughput screening assays. They also provide a homogeneous and phenotypically stable population of cells, they can be screened for pathogens (for example, mycoplasma), and their use improves the reproducibility of results. In addition, they provide large numbers of clonally identical cells for master cell banking, and they are readily traceable for future genetic manipulation (37). However, there are some disadvantages when using immortalised brain endothelial cells. For example, they do not meet the requirement of having good restrictive paracellular permeability, even though other BBB characteristics are similar to those found in primary cultures. Furthermore, the immortalisation process may shift the cells toward a transformed phenotype. They may still de-differentiate in culture, albeit at a slower rate, and multiple cell doublings may introduce chromosomal abnormalities (for example, polyploidy). An important consideration when generating an immortalised/extended life-span brain endothelial cell line for in vitro studies on the BBB, is to create the cell line soon after isolation of the parent endothelial cells, while they retain the in vivo phenotype. Therefore, the isolated cells have to be as pure as possible. Other important considerations are the type of endothelium to use (arteriole, capillary, venule), the brain region, the age of the donor, the most appropriate method for transfection, and the choice of immortalising genes and promoters. The final goal is to find a balance between a primary cell line and a transformed cell line. It is known that rat brain endothelial cells lose cell junctional proteins under standard growth conditions in vitro (for example, occludin and claudin-5 expression are lost after 10 days in culture). Two important considerations when cell lines are used are cross-contamination of cell lines in culture, and mycoplasma contamination (38, 39). Therefore, the karyotype or DNA fingerprinting should be checked on a regular basis.

Table 1: An overview of available immortalised brain endothelial cell lines

Cell line Species Origin Transformation method Reference
RBE4 Rat Brain microvascular endothelial cells Adenovirus ElA 55
GP8/GPNT Rat Brain microvascular endothelial cells SV40 T-antigen 56
TR-BBB Rat Conditionally immortalised brain capillary endothehal cells SV40 T-antigen 57
rBCEC4 Rat Brain microvascular endothelial cells Polyoma virus T-antigen 58
CR3 Rat Brain capillary endothelial cells SV40 T-antigen 59
RBEC1 Rat Brain capillary endothehal cells SV40 T-antigen 60
RCE-T1 Rat Brain microvascular endothelial cells Rous sarcoma virus 61
MCEC Mouse Cerebral endothelial cells SV40 T-antigen 62
TM-BBB4 Mouse Brain capillary endothelial cells SV40 T-antigen 63
S5C Mouse Brain capillary endothelial cells Adenovirus ElA 64
MBEC4 Mouse Brain capillary endothelial cells SV40 T-antigen 65
t-BBEC-117 Bovine Brain capillary endothelial cells SV40 T-antigen 66
SV-BEC Bovine Brain capillary endothelial cells SV40 T-antigen 67
BBEC-SV Bovine Brain microvascular endothelial cells SV40 T-antigen 68
PBMEC Porcine Brain microvascular endothelial cells SV40 T-antigen 69
SV-HCEC Human Brain capillary endothelial cells SV40 T-antigen 70
HBEC-51 Human Brain capillary endothehal cells SV40 T-antigen 71
BB19 Human Brain capillary endothehal cells Human papilloma E6E7 gen 72

SV40 = simian virus 40.

Cell lines of non-cerebral origin

Non-cerebral cell lines in BBB models could be of endothelial or epithelial origin (Table 2). Cells of endothelial origin lack BBB characteristics.

Table 2: An overview of available cell lines of non-cerebral origin

Cell line Species Origin Reference
HUVEC Human Umbilical vein endothelial cells 73
ECV304 Human T24 bladder carcinoma epithelial cells 74
Caco-2 Human Adenocarcinoma intestinal epithelial cells 40
MDCK Dog Renal distal tubule epithelial cells 75
MDCKmdr-1 Dog Mdr-1 transfected MDCK cell line 76
LLC-PK1 Porcine Renal proximal tubule epithelial cells 77

Co-culture with astrocytes may re-induce BBB properties, but extensive characterisation and validation are still needed. Cells of epithelial origin have restrictive paracellular permeability, but they lack other important BBB characteristics and their use for BBB experiments is limited. However, an ECVAM study on in vitro models of the BBB indicated that models based on epithelial cells (such as the Madin-Darby canine kidney [MDCKI and human intestinal [Caco-21 cell lines) can give results similar to those obtained using brain endothelial cells (for example, the bovine model mentioned above), when studies on transport across the BBB are performed (Garberg et al., submitted for publication). However, other studies have demonstrated that the correlations between Caco-2 cells and the BBB are very poor (40).

Co-cultures

Several co-cultures have been established with primary astrocytes, rat C6 glial cells, pericytes, neurons and blood cells (lymphocytes, monocytes), in all possible combinations and with a variety of sources of BBB endothelial cells (41, 42). Astrocytes increase the expression of brain endothelial marker enzymes (γ-GTP, alkaline phosphatase, acetyleholinesterase, Na+-K+-ATPase), transporters (facilitative glucose transporter type 1 [GLUT-1], P-gp) and tight junctions, and help to induce a phenotype more closely mimicking and resembling that found in vivo. However, the procedures are complicated, there is more variability, and tumour cells may induce tumour phenotypes in the BBB. Co-culture with astrocytes increases TEER, and this increase is even higher when cAMP is added. The expression of P-gp is also influenced by astrocytes. The co-culture can be established to enable cell-cell contact through astrocytic end-feet by seeding astrocytes and endothelial cells on either side of the porous support (pore size > 0.3 µm). This type of model can be used to study the functionality of the BBB. On the other hand, the co-culture could be set up without any contact by seeding the astrocytes at the bottom of the well and the endothelial cells on the porous support. The model could then be used to study transport-related processes. Morphological examination under both culture conditions indicates that endothelial. cells have more microvilli if there is direct contact with astrocytes. One concern is that increased surface plasma membrane activity is a characteristic of pathological/reactive endothelial cells, so further investigation would be required, to establish whether these endothelia are pathologically activated. Nevertheless, differentiation signals are a critical aspect and, in particular, integrins clearly play an important role in cellular differentiation through their interaction with extracellular matrix proteins and subsequent intracellular signalling. The importance of the extracellular matrix as well as the matrix-converting enzymes, named matrix metalloproteinases (MMPs), has recently been shown (43, 44).

Three-component models

A three-component model consisting of the human ARPE-19 retinal pigment epithelial (RPE) cell line, a human glioma cell line (U373 MG) and a human neuroblastoma cell line (SH-SY5Y), is under investigation (Figure 2). This model has not been fully characterised in terms of TEER, sucrose permeability and expression of P-gp and occluding, but the glutamate transporter is expressed. The main advantage of this model is the incorporation of target cells for toxicological studies; however, it needs further characterisation. Another advantage is that RPE cells form part of the natural blood-retinal barrier and contain restrictive functions similar to the BBB. RPE cells are able to form tight junctions when grown in vitro on permeable filters (45, 46). This three-dimensional model technique deserves further evaluation for use as a screen for neurotoxicity.

Figure 2: Schematic representation of a three-component model

An epithelial cell line is grown on the upper side of the porous membrane of the insert, a glioma cell line is grown on the lower side of the porous membrane, and a neuronal cell line is grown on the bottom surface of the well.

Cell-free systems (partition coefficients)

There are several cell-free systems that could be used to predict passive permeability at the BBB. The easiest and oldest of them is the determination of the octanol-water partition coefficient. The octanol-water phase boundary only represents a physical interface and the hydrophilic/lipophilic properties of the molecules are roughly estimated. The values depend on the experimental conditions used, which have not been standardised, so high variability is found among data from different laboratories (33). Lipid monolayers and lipid bilayer vesicles appear to be the best cell-free systems, since they mimic more precisely the biological situation. However, these models are too simple, since they consider only the lipid moiety of the membrane and cannot predict transport mechanisms other than passive diffusion.

The ECVAM Study

A study on in vitro models for the BBB was funded by ECVAM (contract number 14547-1998-11F1ED ISP SE; Garberg et al., report submitted for publication). The main goal of this study was to find an in vitro model that could predict the in vivo transport of compounds across the BBB, and was focused on identifying cell lines derived from the BBB, in order to improve reproducibility and transferability between laboratories. Twenty-two test compounds were selected, representing various transport mechanisms, grades of permeability and physicochemical properties: four compounds were substrates of active influx, eight were substrates of active efflux, and ten were transported by passive diffusion (Table 3).

Table 3: Test compounds included in the ECVAM study, classified according to their mechanisms of transport

Diffusion Influx Efflux
Antipyrine Alanine AZT
Caffeine Lactic acid Cimetidine
Diazepam Leucine Cyclosporine
Glycerol L-Dopa Digoxin
Inulin (3H, 14C) Morphine
Nicotine Verapainil
Phenytoin Vinblastine
Sucrose Vincristine
Urea
Warfarin

AZT = 3'-azido-3'-deoxythymidine.
From Garberg et al., submitted for publication.

The in vitro models evaluated were two primary cell systems derived from the BBB (bovine brain endothelial cells co-cultured with rat primary astrocytes, and human brain endothelial cells co-cultured with human primary astrocytes), two immortalised BBB-derived endothelial cell lines (SV-ARBEC, a simian virus 40 [SV40]-immortalised rat brain endothelial cell line co-cultured with SV40-immortalised rat astrocytes; and MBEC4, an SV40-immortalised mouse brain endothelial cell line), and five cell lines not derived from the BBB (a human bladder carcinoma epithelial cell line [ECV304] co-cultured with C6 glioma cells, MDCK, Caco-2, MDCKwt and MDCKmdr-1).

Two in vivo models were included, the mouse brain uptake assay (MBUA), in which the test substance was injected into the tail vein of the mouse and concentrations in plasma and brain were measured after 5 minutes, and microdialysis in the rat, in which the free concentrations of the test substance in brain, blood and muscle were recorded over 3 hours. The area under the curve (AUQ was calculated, and the AUCbrain:AUCblood ratio was determined.

The main conclusions from the study were:

  1. The linear correlations between in vivo permeability (MBUA) and in vitro permeability were poor for all the in vitro models when all the test compounds were included in the analysis, and generally improved only when passively distributed compounds were included.
  2. Active influx was not predicted by any of the in vitro models, most probably due to the radioactive concentration used compared to the trace radioactive concentration plus endogenous concentration.
  3. MDCKmdr-1 was identified as the best model for distinguishing, at least qualitatively, between compounds distributed by active efflux and passive mechanisms.
  4. The possibility of using a battery of in vitro tests for the prediction of the in vivo situation should be investigated.

 

The Need for New Knowledge

The BBB plays an important role in maintaining brain homeostasis and protecting the brain from toxic insults. In the treatment of neurological disorders, the BBB, could represent a real barrier to the delivery of certain drugs of interest. In other instances, disruption of the BBB is the initial step in the development of a CNS disease (for example, in multiple sclerosis) or is the consequence of such a neurological disorder. The assessment of neurotoxicity requires the use of a battery of tests. Because of the importance of the BBB, It is clear that the assessment of BBB integrity and the passage of compounds through this barrier should be incorporated into a testing strategy. The aim would be to improve predictability by using a combination of neurotoxicity, BBB in vitro assays and kinetic modelling.

A strategy for toxicity testing in vitro was suggested (Figure 3), based on the paradigm that toxicity is determined by the critical concentration at the target site. This concentration will depend on the dose administered, route of administration, possible interactions (for example, with other chemicals), distribution in the body, and excretion. Many of the biokinetic consequences, as well as toxicodynamic (TD) interactions leading to toxic effects, are at least partially determined by the physicochemical properties of the compound.

Figure 3: A proposed testing strategy including blood-brain barrier (BBB), target organ and kinetic modelling

CNS = central nervous system; QSAR = quantitative structure-activity relationship.

The prediction of the critical concentration of the compound or its metabolite(s) could be achieved by PBBK/TD modelling, which describes the body as consisting of different compartments with different compositions with regard to lipophilicity and hydrophilicity. The combination of in vitro data on kinetics with in vitro data on effects could permit predictions with regard to both kinetics and dynamics (47). In order to build PBBK/TD models, transport rates between the different compartments are needed, and these can be derived from in vitro experiments (48).

The testing strategy should comprise a stepwise approach, in which the chemical structure of the compound is taken as the basis for prediction of its toxic potential (for example, via a relevant quantitative structure-activity relationship [QSAR] for chemically similar compounds, and the use of physicochemical properties for predictions of possible functional properties). In addition, information from a cytotoxicity test should be used to select the working concentrations. The second step should be an evaluation of the passage of chemicals across the BBB, in which the effect on barrier integrity at no-effect levels, by measurement of TEER or leakage of sucrose, would be evaluated. The passage of the compound would then be determined by measuring rates of passive and active (influx and efflux) transport. Semi-quantitative measurements are likely to be sufficient, but a more thorough analysis of transporter kinetics might be needed, in order to evaluate the possibility of saturation of specific transporters at high doses, and thus the possibility of nonlinear kinetics. One approach suggested for circumventing the low predictability of, and variability between, the models available at present, was to define a battery of BBB models that is complementary and that could/should be used in parallel. The third step would consist of biokinetic modelling. The data obtained could be used for neurotoxicological risk assessment and for the prediction of concentrations reaching the brain. The fourth step should include neurotoxicity studies.

Selected test compounds could be used for testing the proposed strategy. For instance, it was proposed that compounds used for the evaluation of in vitro models for neurotoxicity should be evaluated for passage across the BBB by using a selection of available BBB models. A neurotoxic compound could either pass the BBB without damaging the BBB itself, or could reach the brain after disrupting the BBB before exerting its toxicity in the CNS. Alternatively, a compound could disrupt the BBB without exerting any neurotoxic effect, and thus could indirectly cause neurotoxic effects by permitting the passage of neurotoxic, but non-permeable, compounds to the CNS. The proposed testing strategy should be able to identify these different classes of compound.

The in vitro models used in this strategy cannot be too simple, since their use would then tend to lead to unnecessary experiments that would not provide additional information. Simple diffusion and passive transport can be predicted from the physicochemical properties of the compounds. QSPRs have been used to predict biokinetic behaviour (15, 16, 49). Such data are very useful in the modelling of the kinetic behaviour of compounds.

In this approach, the lipophilicity and water volume of the various tissue compartments are taken into consideration. It is possible to make good predictions of acute toxicity for volatile compounds, and their toxicity (LC50) is most likely to be seen at brain lipid concentrations of 70 ± 31 mM. Species differences have to be taken into consideration when making these predictions (50).

Nevertheless, the prediction of toxicity/barrier penetration needs additional data. For instance, the affinity of compounds for specific transporters is an additional factor of importance. Amino acid transporters, glucose transporters, receptor-mediated transport, transporters for neurotransmitters, P-gp and other efflux proteins, and ion exchange transporters are some examples of transporters that are of interest for the BBB (51, 52). The use of physicochemical properties and/or computational QSAR-based models is not sufficient for predicting substrates for transport (for example, ATP-binding cassette [ABC] transporters). With regard to partitioning to the brain, it could be hypothesised that, if the predicted partitioning to the brain was low, the need for additional testing might be low. In addition, hydrophobic compounds might not be expected to reach the brain, while lipophilic compounds would be expected to do so. However, it is also important to consider that hydrophilic compounds may be subject to active transport into the brain, and that lipophilic compounds may be transported out of the brain. Therefore, it is not possible to use only hydrophilicity/lipophilicity for the prediction of brain partitioning.

The ECVAM study indicated the possibility of using in vitro data to distinguish between passive compounds (non-toxic during the experiment), compounds actively pumped into the brain (although this was questioned), compounds actively pumped out of the brain, passive compounds for which passage is restricted by the blood flow, and compounds that are partitioned to the brain.

Even though Caco-2 and MDCKmdr-1 cells are used to study the distribution of compounds across the BBB, it was generally considered that it would be preferable to use a brain endothelia-derived cell model. In in vitro toxicity studies, brain endotheliaderived cells are more sensitive to toxic compounds than are MDCK cells or Caco-2 cells. Bovine endothelial cells co-cultured with astrocytes and porcine endothelial cells cultured in serum-free media containing hydrocortisone were suggested as the first choices. The advantage of the porcine model is the tightness of the barrier (permeability for sucrose ~10-7 cm/second compared with ~10-5 cm/second in the bovine model). However, some concern for the possible additional and uncharacterised effects of hydrocortisone should be taken into consideration. For instance, it is known that bovine endothelial cells become more resistant to lipopolysaccharides after treatment with hydrocortisone (53). On the other hand, the advantage of the bovine model is the physiologically relevant interaction between the endothelial cells and the astrocytes.

A study on the absorption/uptake of compounds with the immortalised GPNT rat brain endothelial cell line was suggested as a possible initial screen to identify compounds for further and more-specific investigations. Additional immortalised brain endothelial cell lines more closely resembling the in vivo BBB are currently under development, and may prove useful for such purposes in the future.

In order to evaluate the most promising models discussed during the workshop and to have a preliminary idea of their performance, it was agreed among all the participants that a study should be undertaken, in which 10-20 compounds representing low, medium and high passage could be tested with selected in vitro models (primary bovine endothelial cells co-cultured with rat primary astrocytes; porcine primary endothelial cells cultured in hydrocortisone-conditioned medium; the ARPE-19 human retinal cell line co-cultured with a human glioma cell line; GPNT cells and MDCKmdr-1 cells). Several test chemicals were proposed, including acrylamide and glutamate as negative controls, caffeine, lindane, diazepam, phenytoin, parathione, paraoxon, benzene and toluene. The first step will be to collect good cytotoxicity data. All the information regarding the cytotoxicity should be made available, and information about the neurotoxicity of the compounds should also be collected. The cytotoxicity data (72-hour exposure, in serum-free medium) obtained in the ERGATT/CFN integrated toxicity testing scheme (ECITTS) study (19) should be provided. The barrier integrity and the rates of transport for each compound at the no-observed-effect level (NOEL) or the lowest-observed-effect level (LOEL) concentrations should be obtained for each of the in vitro models. All the information obtained should be combined for use in an appropriate biokinetic model.

During the workshop, the minimal criteria needed for an in vitro model of the BBB were established. Firstly, the availability and transferability of the in vitro model is important, since it is probable that testing will be performed at different centres. The model must allow for the study of transport polarity, and thus for the identification of active transport in and out of the brain. Integrity markers, such as TEER, sucrose or dyes, should be used, primarily to confirm the integrity of the monolayer prior to and during the experiment.

More information is needed on the metabolic capacity and possible specific metabolic properties of the BBB, in order to predict the influence of the metabolic barrier on the passage of compounds across the BBB. In addition, further characterisation of available in vitro and in vivo BBB models is also needed, with regard to the expression of different transporters (at the RNA, protein and functional levels).

It was emphasised that models for the choroid plexus might also be of importance, since this blood-brain interface may represent as much as one-tenth of the surface of the BBB. However, hydrophilic compounds have a very low rate of diffusion into the CSF. Some transporters are expressed at the choroid plexus, but not at the BBB, and therefore the choroid plexus also might decrease the concentration of xenobiotics in the brain. Another important factor that should be considered is bulk flow, which might have a significant effect on the excretion and uptake of compounds via the choroid plexus. Therefore, it should be recognised that processes other than those occurring in the endothelial cells may be of importance for the efflux and uptake of compounds out of/into the brain (54).

 

Conclusions

  1. A number of promising in vitro models of the BBB based on sound scientific knowledge are available.

  2. There is a need for further development, characterisation and standardisation of in vitro models.

  3. There is a need for characterisation of the expression of transporters in the BBB in the in vivo situation, to permit the setting of criteria for acceptable in vitro models with regard to active transport.

  4. There is a need for standard protocols for the different models.

  5. The minimal requirements for in vitro BBB models need to be better defined, particularly in relation to: a) availability and ease of culture; b) the functional expression of transporter mechanisms; c) the possibility of studying polarity; d) restrictive paracellular transport; and e) closeness of morphology to that of the in vivo system.

  6. A pilot study needs to be carried out to evaluate the performances of the models, which should include influx and efflux rates, and the integrity of the barrier.

  7. The metabolic capacity of the BBB needs to be studied further.

  8. The measurement of neurotoxicity must be linked to measurements of effects on the BBB.

  9. The production of semi-quantitative measurements of transport rates is likely to be enough for risk assessment in neurotoxicity.

  10. The data obtained by using in vitro models of the BBB should be used in biokinetic modelling to estimate the relationship between brain concentrations and exposure scenarios.

  11. Biokinetic factors must be taken into account in risk assessment.

  12. This is the starting point for including biokinetic modelling, target organ toxicity and systemic toxicity in a testing strategy.

  13. A number of promising models are available and could be used in a validation study.

  14. Insufficient in vivo data are available for the validation of in vitro models of the BBB. Therefore, in vivo neurotoxicity data may have to be taken into consideration.

  15. In vitro models of the choroid plexus need to be further developed and evaluated, especially with regard to specific transporters.

  16. The synergistic functions of the BBB and the blood-CSF barrier should be defined.

 

Recommendations

  1. A study should be conducted to evaluate the performances of several promising in vitro models presented during the workshop, by using an appropriate set of selected compounds.

  2. The outcome of this study should result in a prevalidation/validation study.

  3. Basic research on the BBB is recommended, on topics such as: a) species differences; b) metabolic capacity; c) the characteristics of the human BBB; d) differences between in vivo and in vitro systems; and e) active transport.

  4. The further development of testing strategies is essential.

  5. The human BBB should be characterised in relation to transporters, tight junctions and biotransformation.

  6. An ECVAM database on BBB permeability/toxicity data should be set up.

 

References

  1. Anon. (1994). ECVAM News & Views. ATLA 22, 7-11.
  2. Anon. (2001). White Paper on a Strategy for a Future Chemicals Policy (COM(2001)88 final), 32pp. Brussels, Belgium: European Commission. Website http://europa.eu.int/comm/enterprise/chemicals/ chempol/whitepaper/reach.htm (Accessed 18.11.03).
  3. Anon. (2003). Directive 2003/15/EC of the European Parliament and of the Council of 27 February 2003 amending Council Directive 76/768/EEC on the approximation of the laws of the Member States relating to cosmetic products (Text with EEA relevance). Official Journal of the European Union L66, 26-35.
  4. Kusuhara, H. & Sugiyama, Y. (2001). Efflux transport systems for drugs at the blood-brain barrier and bloodcerebrospinal fluid barrier (part 1). Drug Discovery Today 6, 150-156.
  5. Feng, M.R. (2002). Assessment of blood-brain barrier penetration: in silico, in vitro and in vivo. Current Drug Metabolism 3, 647-657.
  6. Nitz, T., Eisenblätter, T., Haselbach, M. & Galla, H-J. (2001). Recent advances in the development of cell culture models for the blood-brain- and blood-CSF-barrier. In Blood-Brain Barrier Delivery and Brain Pathology (ed. J.L. Kluwer), pp. 45-62. New York, NJ, USA: Plenum Press.
  7. Haselbach, M., Wegener, J., Decker, S., Engelbertz, C. & Galla, H-J. (2001). Porcine choroid plexus epithelial cells in culture: regulation of barrier properties and transport processes. Microscopy Research and Technique 52, 137-152.
  8. Strazielle, N. & Ghersi-Egea, J-F. (1999). Demonstration of a couplet metabolism-efflux process at the choroid plexus as a mechanism of brain protection towards xenobiotics. Journal of Neuroscience 19, 6275-6289.
  9. Savettieri, G., Di Liegro, I., Catania, C., Licata, L., Pitarresi, G.L., D'Agostino, S., Schiera, G., De Caro, V., Giandalia, G., Giannola, L.I. & Cestelli, A. (2000). Neurons and ECM regulate occluding localization in brain endothelial cells. Molecular Neuroreceptors 11, 1081-1084.
  10. Bauer, H.C. & Bauer, H. (2000). Neural induction of the blood-brain barrier: still an enigma. Cell and Molecular Neurobiology 20, 13-28.
  11. de Vries, H.E., Kuiper, J., de Boer, A.G., van Berkel, J.C. & Breimer, D.D. (1997). The blood-brain barrier in neuroinflammatory diseases. Pharmacological Reviews 49, 143-155.
  12. Grant, G.A., Abbott, N.J. & Janigro, D. (1998). Understanding the physiology of the blood-brain barrier: in vitro models. News in Physiological Sciences 13, 287-293.
  13. Andersen, M.E. (1991). Physiological modelling of organic compounds. Annals of Occupational Hygiene 35, 309-321.
  14. Poulin, P. & Krishnan, K. (1996). A tissue composition-based algorithm for predicting tissue:air partition coefficients of organic chemicals. Toxicology and Applied Pharmacology 136, 126-130.
  15. DeJongh, J., Verhaar, H.J. & Hermens, J.L. (1998). Role of kinetics in acute lethality of nonreactive volatile organic compounds (VOCs). Toxicological Sciences 45, 26-32.
  16. Meulenberg, C.J. & Vijverberg, H.P. (2000). Empirical relations predicting human and rat tissue:air partition coefficients of volatile organic compounds. Toxicology and Applied Pharmacology 165, 206-216.
  17. Worth, A.P. & Balls, M., eds (2002). Alternative (Nonanimal) Methods for Chemicals Testing: Current Status and Future Prospects. A report prepared by ECVAM and the ECVAM working group on chemicals. ATLA 30, Suppl. 1, 1-125.
  18. Forsby, A., Pilli, F., Bianchi, V. & Walum, E. (1995). Determination of critical cellular neurotoxic concentrations in human neuroblastorna (SH-SY5Y) cell cultures. ATLA 23, 800-811.
  19. DeJongh, J., Forsby, A., Houston, J.B., Beckman, M., Combes, R. & Blaauboer, B.J. (1999). An integrated approach to the prediction of systemic toxicity using computer-based biokinetic models and biological in vitro test methods: overview of a prevalidation study based on the ECITTS project. Toxicology in Vitro 13, 549-554.
  20. Rubin, L.L. & Staddon, J.M. (1999). The cell biology of the blood-brain barrier. Annual Review of Neuroscience 22, 11-23.
  21. Garberg, P. (1998). In vitro models of the blood brain barrier. ATLA 26, 821-847.
  22. de Boer, A.G. & Gaillard, P.J. (2002). In vitro models of the blood-brain barrier: when to use which? Current Medicinal Chemistry—Central Nervous System Agents 2, 203-209.
  23. Méresse, S., Dehouck, M.P., Delorme, P., Bensaid, M., Tauber, J.P., Delbart, C., Fruchart, J.C. & Cecchelli, R. (1989). Bovine brain endothelial cells express tight junctions and monoamine oxidase activity in long-term culture. Journal of Neurochemistry 53, 1363-1371.
  24. Dehouck, M.P., Méresse, S., Delorme, P., Fruchart, J-C. & Cecchelli, R. (1990). An easier, reproducible, and massproduction method to study the blood-brain barrier in vitro. Journal of Neurochemistry 54, 1798-1801.
  25. Dehouck, M.P., Dehouck, B., Schluep, C., Fruchart, J.C., Lemaire, M. & Cecchelli, R. (1995). Drug transport to the brain: comparison between in vitro and in vivo models of the blood-brain barrier. European Journal of Pharmaceutical Sciences 3, 357-365.
  26. Fenart, L., Buée-Scherrer, V., Descamps, L., Duhem, C., Poullain, M.G., Cecchelli, R. & Dehouck, M.P. (1998). Inhibition of P-glycoprotein at the bloodbrain barrier level: in vitro studies of its implication in blood-brain barrier integrity and in drug transport to the brain by an in vitro model of the blood-brain barrier. Pharmaceutical Research 15, 993-1000.
  27. Dehouck, M.P., Vigne, P., Torpier, G., Breittmayer, J.P., Cecchelli, R. & Frelin, C. (1997). Endothelin-I as a mediator of endothelial cell-pericyte interactions in bovine brain capillaries. Journal of Cerebral Blood Flow and Metabolism 17, 464-469.
  28. Descamps, L., Dehouck, M.P., Torpier, G. & Cecchelli, R. (1996). Receptor-mediated transcytosis of transferrin through blood-brain barrier endothelial cells. American Journal of Physiology—Heart and Circulatory Physiology 270, 1149-1158.
  29. Fillebeen, C., Descamps, L., Dehouck, M.P., Fenart, L., Benaissa, M., Spick, G., Cecchelli, R. & Pierce, A. (1999). Receptor mediated transcytosis of lactoferrin through the blood-brain barrier. Journal of Biological Chemistry 274, 7011-7017.
  30. Cecchelli, R., Dehouck, B., Descamps, L., Fenart, L., Buée-Scherrer, V., Duhem, C., Lundquist, S., Rentfel, M., Torpier, G. & Dehouck, M.P. (1999). In vitro model for evaluating drug transport across the blood-brain barrier. Advanced Drug Delivery Reviews 36, 165-178.
  31. Lundquist, S., Renftel, M., Brillault, J., Fenart, L., Cecchelli, R. & Dehouck, M.P. (2002). Prediction of drug transport through the blood-brain barrier in vivo: a comparison between two in vitro cell models. Pharmaceutical Research 7, 976-981.
  32. Hoheisel, D., Nitz, T., Franke, H., Wegener, J., Hakvoort, A., Tilling, T. & Galla, H-J. (1998). Hydrocortisone reinforces the blood-brain barrier properties in a serum-free cell culture system. Biochemical and Biophysical Research Communications 244, 231-316.
  33. Lohmann, C., Hilwel, S. & Galla, H-J. (2002). Predicting blood-brain barrier permeability of drugs: evaluation of different in vitro assays. Journal of Drug Targeting 10, 263-276.
  34. Eisenblätter, T., Hüwel, S. & Galla, H-J. (2003). Characterisation of the brain multidrug resistance protein (BMDP/ABCG2/BCRP) expressed at the blood-brain barrier. Brain Research 971, 221-231.
  35. Nitz, T., Eisenblätter, T., Psathaki, K. & Galla, H-J. (2003). Serum-derived factors weaken the barrier properties of cultured porcine brain capillary endothelial cells in vitro. Brain Research 981, 30-40.
  36. Gaillard, P.J., van der Meide, P.H., de Boer, A.G. & Breimer, D.D. (2001). Glucocorticoid and type I interferon interactions at the blood-brain barrier: relevance for drug therapies for multiple sclerosis. NeuroReport 12, 2189-2193.
  37. Uyttendaele, H., Closson, V., Wu, G., Roux, F., Weinmaster, G. & Kitajewski, J. (2000). Notch4 and Jagged-1 induce microvessel differentiation of rat brain endothelial cells. Microvascular Research 60, 91-103.
  38. Drexler, H.G., Quentmeier, H., Dirks, W.G. & MacLeod, R.A. (2002). Bladder carcinoma cell line ECV304 is not a model system for endothelial cells. In Vitro Cellular and Developmental Biology—Animal 38,185-186.
  39. Dirks, W.G., MacLeod, R.A. & Drexler, H.G. (1999). ECV304 (endothelial) is really T24 (bladder carcinoma): cell line cross-contamination at source. In Vitro Cellular and Developmental Biology—Animal 35, 558-559.
  40. Lundquist, S., Renftel, M., Brillault, J., Fenart, L., Cecchelli, R. & Dehouck, M.P. (2002). Prediction of drug transport through the blood-brain barrier in vivo: a comparison between two in vitro cell models. Pharmaceutical Research 19, 976-981.
  41. Gaillard, P.J., Voorwinden, L.H., Nielsen, J.L., Ivanov, A., Atsumi, R., Engman, H., Ringbom, C., de Boer, A.G. & Breimer, D.D. (2001). Establishment and functional characterizarion of an in vitro model of the blood-brain barrier, comprising a co-culture of brain capillary endothelial cells and astrocytes. European Journal of Pharmacuetical Sciences 12, 215-222.
  42. Kramer, S.D., Schutz, Y.B., Wunderli-Allenspach, H., Abbott, N.J. & Begley, D.J. (2002). Lipids in blood-brain barrier models in vitro. II. Influence of glial cells on lipid classes and lipid fatty acids. In Vitro Cellular and Developmental Biology—Animal 38, 566-571.
  43. Tilling, T.H., Engelbertz, C.H., Decker, S.T., Korte, D., Hüwel, S. & Galla, H-J. (2002). Expression and adhesive properties of basement membrane proteins in cerebral capillary endothelial cell cultures. Cell and Tissue Research 310, 19-29.
  44. Lohmann, C., Krischke, M., Wegener, J. & Galla, HJ. (2004). Tyrosine phosphatase inhibition induces loss of bloodbrain barrier integrity by matrix metalloproteinase-dependent and independent pathways. Brain Research 995, 184-196.
  45. Tähti, H., Nevala, H. & Toimela, T. (2003). Refining in vitro neurotoxicity testing - the development of blood-brain barrier models. ATLA 31, 273-276.
  46. Toimela, T. Mäenpää, H., Mannerström, M. & Tähti, H. (2004). Development of an in vitro blood-brain barrier model̬cytotoxicity of mercury and aluminum. Toxicology and Applied Pharmacology 195, 73-82.
  47. Blaauboer, B.J. (2003). Biokinetic and toxicodynamic modelling and its role in toxicological research and risk assessment. ATLA 31, 277-281.
  48. DeJongh, J., Nordin-Andersson, M., Ploeger, B. & Forsby, A. (1999). Estimation of systemic toxicity of acrylamide by integration of in vitro toxicity data with kinetic simulations. Toxicology and Applied Pharmacology 158, 261-268.
  49. Poulin, P. & Krishnan, K. (1996). A tissue composition-based algorithm for predicting tissue:air partition coefficients of organic chemicals. Toxicology and Applied Pharmacology 136, 126-130.
  50. DeJongh, J., Verhaar, H.J. & Hermens, J.L.M. (1997). A quantitative property-property relationship (QPPR) approach to estimate in vitro tissueblood partition coefficients of organic chemicals in rats and humans. Archives of Toxicology 72, 17-25.
  51. Lee, G., Dallas, S., Hong, M. & Bendayan, R. (2001). Drug transporters in the central nervous system: brain barriers and brain parenchyma considerations. Pharmacological Reviews 53, 569-596.
  52. Sun, H., Dai, H., Shaik, N. & Elmquist, W.F. (2003). Drug efflux transporters in the CNS. Advanced Drug Delivery Reviews 55, 83-105.
  53. Gaillard, P.J., de Boer, A.G. & Breimer, D.D. (2003). Pharmacological investigations on lipopolysaccharide-induced permeability changes in the bloodbrain barrier in vitro. Microvascular Research 65, 24-31.
  54. Ghersi-Egea, J.F. & Strazielle, N. (2002). Choroid plexus transporters for drugs and other xenobiotics. Journal of Drug Targeting 10, 353-357.
  55. Roux, F., Durieu-Trautmann, O., Chaverot, N., Claire, M., Mailly, P., Bourre, J.M., Strosberg, A.D. & Couraud, P.O. (1994). Regulation of gamma-glutamyl transpeptidase and alkaline phosphatase activities in immortalized rat brain microvessel endothelial cells. Journal of Cellular Physiology 159, 101-113.
  56. Greenwood, J., Pryce, G., Devine, L., Male, D.K., dos Santos, W.L., Calder, V.L. & Adamson, P. (1996). SV40 large T immortalised cell lines of the rat blood-brain and blood-retinal barriers retain their phenotypic and immunological characteristics. Journal of Neuroimmunology 71, 51-63.
  57. Terasaki, T. & Hosoya, K. (2001). Conditionally immortalized cell lines as a new in vitro model for the study of barrier functions. Biological and Pharmaceutical Bulletin 24, 111-118.
  58. Blasig, I.E., Giese, H., Schroeter, M.L., Sporbert, A., Utepbergenov, D.I., Buchwalow, I.B., Neubert, K., Schonfelder, G., Freyer, D., Schinike, I., Siems, W.E., Paul, M., Haseloff, R.F. & Blasig, R. (2001). NO and oxyradical metabolism in new cell lines of rat brain capillary endothelial cells forming the blood-brain barrier. Microvascular Research 62, 114-127.
  59. Lechardeur, D., Schwartz, B., Paulin, D. & Scherman, D. (1995). Induction of blood-brain barrier differentiation in a rat brain-derived endothelial cell line. Experimental Cell Research 220, 161-170.
  60. Kido, Y., Tamai, I., Okamoto, M., Suzuki, F. & Tsuji, A. (2000). Functional clarification of MCT1-mediated transport of monocarboxylic acids at the blood-brain barrier using in vitro cultured cells and in vivo BUI studies. Pharmaceutical Research 17, 55-62.
  61. Mooradian, D.L. & Diglio, C.A. (1991). Production of a transforming growth factor-beta-like growth factor by RSVtransformed rat cerebral microvascular endothelial cells. Tumour Biology 12, 171-183.
  62. Xu, J., Qu, Z.X., Moore, S.A., Hsu, C.Y. & Hogan, E.L. (1992). Receptor-linked hydrolysis of phosphoinositides and production of prostacyclin in cerebral endothelial cells. Journal of Neurochemistry 58, 1930-1935.
  63. Asaba, H., Hosoya, K., Takanaga, H., Ohtsuki, S., Tamura, E., Takizawa, T. & Terasaki, T. (2000). Blood-brain barrier is involved in the efflux transport of a neuroactive steroid, dehydroepiandrosterone sulfate, via organic anion transporting polypeptide 2. Journal of Neurochemistry 75, 1907-1916.
  64. Wijsman, J.A. & Shivers, R.R. (1998). Immortalized mouse brain endothelial cells are ultrastructurally similar to endothelial cells and respond to astrocyteconditioned medium. In Vitro Cellular and Developmental Biology—Animal 34, 777-784.
  65. Tatsuta, T., Naito, M., Oh-hara, T., Sugawara, I. & Tsuruo, T. (1992). Functional involvement of P-glycoprotein in blood-brain barrier. Journal of Biological Chemistry 267, 20383-20391.
  66. Sobue, K., Yamamoto, N., Yoneda, K., Hodgson, M.E., Yamashiro, K., Tsuruoka, N., Tsuda, T., Katsuya, H., Miura, Y., Asai, K. & Kato, T. (1999). Induction of blood-brain barrier properties in immortalized bovine brain endothelial cells by astrocytic factors. Neuroscience Research 35, 155-164.
  67. Durieu-Trautmann, O., Foignant-Chaverot, N., Perdomo, J., Gounon, P., Strosberg, A.D. & Couraud, P.O. (1991). Immortalization of brain capillary endothelial cells with maintenance of structural characteristics of the blood-brain barrier endothelium. In Vitro Cellular and Developmental Biology—Animal 27, 771-778.
  68. Stins, M.F., Prasadarao, N.V., Zhou, J., Arditi, M. & Kim, K.S. (1997). Bovine brain microvascular endothelial cells transfected with SV40-large T antigen: development of an immortalized cell line to study pathophysiology of CNS disease. In Vitro Cellular and Developmental Biology—Animal 33, 243-247.
  69. Teifel, M. & Friedl, P. (1996). Establishment of the permanent microvascular endothelial cell line PBMEC/C12 from porcine brains. Experimental Cell Research 228, 50-57.
  70. Muraganandam, A., Herx, L.M., Monette, R., Durkin, J.P. & Stanimirovic, D.B. (1997). Development of immortalized cerebromicrovascular endothelial cell line as an in vitro model of the human blood-brain barrier. FASEB Journal 11, 1187-1197.
  71. Xiao, L., Yang, C., Dorovini-Zis, K, Tandon, N.N., Ades, E., Lal, A.A. & Udhayakumar, V. (1996). Plasmodium falciparum: involvement of additional receptors in the cytoadherence of infected erythrocytes to microvascular endothelial cells. Experimental Parasitology 84, 42-55.
  72. Prudhomme, J.G., Sherman, I.W., Land, K.M., Moses, A.V., Stenglein, S. & Nelson, J.A. (1996). Studies of Plasmodium falciparum cytoadherence using immortalized human brain capillary endothelial cells. International Journal of Parasitology 26, 647-655.
  73. Kallmann, B.A., Wagner, S., Hummel, V., Buttmann, M., Bayas, A., Tonn, J.C. & Rieckmann, P. (2002). Characteristic gene expression profile of primary human cerebral endothelial cells. FASEB Journal 16, 589-591.
  74. Drexler, H.G., Quentmeier, H., Dirks, W.G. & MacLeod, R.A. (2002). Bladder carcinoma cell line ECV304 is not a model system for endothelial cells. In Vitro Cellular and Developmental Biology—Animal 38, 185-186.
  75. Veronesi, B. (1996). Characterization of the MDCK cell line for screening neurotoxicants. Neurotoxicology 17, 433-443.
  76. Smith, B.J., Doran, A.C., McLean, S., Tingley, F.D. III, O'Neill, B.T. & Kajiji, S.M. (2001). P-glycoprotein efflux at the blood-brain barrier mediates differences in brain disposition and pharmacodynamics between two structurally related neurokinin-1 receptor antagonists. Journal of Pharmacology and Experimental Therapeutics 298, 1252-1259.
  77. Batrakova, E.V., Miller, D.W., Li, S., Alakhov, V.Y., Kabanov, A.N. & Elmquist, W.F. (2001). Pluronic P85 enhances the delivery of digoxin to the brain: in vitro and in vivo studies. Journal of Pharmacology and Experimental Therapeutics 296, 551-557.