Pharmacogenomics || Pharmacogenomics of the Blood–Brain Barrier

Abstract

The endothelial cells of brain capillaries are not the only component forming theblood–brain barrier (BBB). Pericytes and astrocyte foot processes, separated fromeach other by the basement membrane, are also part of the BBB. The specificityof the BBB permeability, which is commonly defined by the high transcellularelectrical resistance due to the tight junctions sealing adjacent endothelial cells,depends greatly on cross-talk between these cells. The BBB is not just a physicalbarrier, it is also a metabolic and pharmacological barrier. These new propertiesare due to the expression of many genes whose products are implicated in drugmetabolism, carrier-mediated transport and interaction with receptors at the BBB.The genes encoding phase 1 oxidative enzymes like CYP2D6 or the efflux P-glyco-protein transporter (P-gp) in the cerebral endothelial cells are all polymorphic atthe BBB. Together with the induction or repression of these proteins, they providethe basis for genomic studies of the BBB. The many cellular components of theBBB and the relatively small amount of the brain that forms the BBB, makes itdifficult to identify the genes and proteins involved. Nevertheless, transgenic invivo models like P-gp knockout mice have proved to be powerful experimentaltools for assessing the impact of the P-gp transporter on BBB function. Polymer-ase chain reaction-based subtraction cloning methods have recently led to theidentification of genes that are more actively expressed at the BBB than else-where. These new approaches should lead to the identification of new targets forspecific drugs and a clearer picture of the mechanisms underlying cerebrovasculardisorders.

311

15

Pharmacogenomics of the Blood–Brain BarrierJean-Michel Scherrmann

Pharmacogenomics: The Search for Individualized Therapies.Edited by J. Licinio and M.-L. Wong

Copyright © 2002 Wiley-VCH Verlag GmbH & Co. KGaAISBNs: 3-527-30380-4 (Paper); 3-527-60075-2 (Electronic)

15.1Basic Concepts Underlying the Pharmacogenomics of the Blood–Brain Barrier

15.1.1The Two Barriers

There are two physiological barriers separating the brain from its blood supply;they control the entry and exit of endogenous and exogenous compounds. Thisallows the body to maintain a constant internal milieu in the brain, protecting itfrom fluctuations in circulating hormones, amino acids, ions and other nutrients,so preventing uncontrolled disturbances of the central nervous system (CNS).One is the blood–brain barrier (BBB) and the other is the blood–cerebrospinalfluid barrier (BCSFB). The general concept of a restricted passage of solutes outof the blood into the brain dates from the studies of Lewandowsky (1900) andEhrlich (1902) [1, 2]. They observed that dyes injected intravenously were nottaken up by the brain. Goldmann (1909, 1913) [3, 4] carried out experiments withtrypan blue injected directly into the cerebrospinal fluid. These studies helped todifferentiate the two barriers, as the dye left only the blood vessels of the choroidplexuses (for a review, see [5]). Only a few small regions of the brain, collectivelyknown as the circumventricular organs (CVO) making up less than 1% of the cer-ebrovascular bed, have no barrier, enabling substances in the blood to reach thebrain extracellular fluid. These are specialized tissues that act as centers ofhomeostatic and neurohormonal control for the body. They include the medianeminence, pineal gland, organum vasculosum of the lamina terminalis, subforni-cal organ, subcommissural organ, area postrema and the neurohypophysis inclose proximity to the ventricular system, particularly the third ventricle. The rela-tively free exchange of solutes between the blood and the CVO indicates that theircapillaries are more permeable than are those in the rest of the brain. However,these vessels do not provide indiscriminate access to the whole of the CNS be-cause there are tight junctions between these regions and the rest of the brain [5].

The BCSFB is located at the choroid plexuses. These plexuses float freely in thebrain ventricles and are formed by epithelial cells held together at their apices bytight junctions (Figure 15.1). Beneath the epithelial cells is a stroma containingthe blood vessels, which lack tight junctions. Thus, the fenestrated blood vesselsof the choroid plexus allow large molecules to pass, but the tight junctions at theepithelial cell surfaces restrict their passage into the cerebrospinal fluid (CSF) [6].Since the surface area of the human BBB is estimated to be 20 m2 or around athousand times greater than that of the BCSFB, the BBB is considered to be themain region controlling the neuronal environment and the uptake of drugs intothe brain parenchyma. It is also the main target for delivering drugs to the brainvia carrier-mediated strategies [7].

15 Pharmacogenomics of the Blood–Brain Barrier312

15.1.2Constituents of the Blood–Brain Barrier

15.1.2.1 Endothelial Cells in the Blood–Brain BarrierThe BBB is defined by the microvasculature of the brain, which consists of amonolayer of polarized endothelial cells connected by complex tight junctions hav-ing a high electrical resistance (> 1000–3000 � cm–2). This structure prevents para-cellular transport across the brain endothelium [8]. These endothelial cells are se-parated from the astrocyte foot processes and pericytes by a basement membrane(Figure 15.2). The astrocyte foot processes are about 20 nm from the abluminalsurface of the endothelial cells and this space is mainly filled with the microvascu-lar basement membrane and the brain extracellular fluid. The endothelial cells ac-tively regulate vascular tone, blood flow and barrier function in the brain micro-vasculature. Endothelial cells are thin, about 0.1 �m thick, and thus occupy about0.2% of the volume of the whole brain [9]. They are polarized, like epithelial cells;the luminal and abluminal endothelial membranes each segregate specific trans-cellular transport across the brain endothelium. This cell polarity is well illu-strated by the distributions of several enzymes. Alkaline phosphatase is equallydistributed between the luminal and the abluminal membranes but Na, K-ATPaseand 5�-nucleotidase are present primarily on the abluminal side, and �-glutamyltranspeptidase (�-GTP) is found mainly on the luminal side [10]. Goldmann [4]first postulated that the brain capillaries provide the anatomical basis of a barrierin 1913, but this was not conclusively demonstrated until the 1960s, when elec-

15.1 Basic Concepts Underlying the Pharmacogenomics of the Blood–Brain Barrier 313

Fig. 15.1 Diagram showing a longitudinalcross-section of the blood–cerebrospinal fluidbarrier at the choroid plexus. This barrier isformed by epithelial or choroid cells held to-gether at their apices by tight junctions. The

fenestrated blood vessels of the choroidplexus allow large molecules to pass, but thetight junctions between the choroid cells re-strict their passage into the cerebrospinalfluid.

tron microscope studies revealed that the endothelial cells of a brain capillaryform electron-dense junctional contacts between two adjacent cells [11]. There areno gap junctions in brain capillaries and postcapillary venules and the tight junc-tions are responsible for sealing adjacent cells and maintaining cell polarity. Tightjunctions are membrane microdomains made up of many specific proteins en-gaged in a complex membranar and intracytosolic network. Continuous tightjunctions are not the only feature that makes the blood vessels of the brain differ-ent from those of other tissues. There are no detectable fenestrations or singlechannels between the blood and interstitial spaces. They contain fewer pinocytoticvesicles than do endothelial cells in the peripheral microvasculature and there isevidence that the majority of what appear to be independent vesicles in the endo-thelium cytoplasm are part of membrane invaginations that communicate witheither the blood or the perivascular space [12]. The endothelial cells that form theBBB also contain many mitochondria. They occupy 8 to 11% of the cytoplasmicvolume, much more than in brain regions lacking a BBB and tissues outside theCNS [13]. This indicates that the BBB also functions as a metabolic barrier, in ad-dition to its physical barrier properties. There are highly specific transport sys-tems carrying nutrients at the luminal or abluminal sides, or at both sides of theendothelial cell membrane. These carrier-mediated transport systems regulate themovement of nutrients between the blood and the brain. The brain capillary endo-thelium also bears specific receptors for circulating peptides or plasma proteinsand these mediate the transcytosis of peptides or proteins through the BBB [14].More recently, the discovery of active carrier-mediated transporters which are notinvolved in transporting substrate from the blood to the brain, but from the brainto blood, has greatly reinforced the barrier properties of the BBB. Most of thesetransmembrane proteins are located at the luminal or abluminal membranes ofthe endothelial cells and restrict the uptake of numerous drugs by the brain. Thusa large number of amphipathic cationic drugs are effluxed by one ABC protein,the P-glycoprotein which is present at the luminal surface of the BBB [15]. Brainvessels also have more classical types of receptors, including �- and �-adrenergicreceptors and receptors for serotonin, adenosine, histamine, angiotensin and argi-nine vasopressin. Another specific feature of the BBB is the brain endothelium,which bears surface anionic sites differing from those found in some fenestratedand continuous endothelia. The distribution of these anionic sites on the luminaland abluminal membranes is, again, different. Experiments with proteolytic en-zymes suggest that the negatively charged domains on the luminal membrane ofthe endothelial cells are mainly the terminal sialic acid residues of acidic glycopro-teins, and the remaining few anionic domains are probably produced by heparansulfate. An important distinction between brain endothelial cell anion distributionand that of other endothelial cells is the high density of heparan sulfate close tothe tight junction domains. This could play a role in cell–cell signaling and adhe-sion [16]. The roles of these anionic domains in the function of the BBB remainto be established, but they actively contribute to the transcytosis of cationic proteinsvia adsorptive-mediated transcytosis [14]. The brain capillaries are also almost com-pletely surrounded by other cells, like pericytes and the astrocyte foot processes.


All these cellular components of the brain capillaries are joined by junctional sys-tems. Zonal and extensive tight junctions seal the endothelial cells and gap junc-tions connect the endothelium to the subjacent pericyte layer, allowing their func-tional coupling and also weld them to the astrocyte processes. The last compo-nent of the endothelial cell network is the nerve fibers, which may be seen closeto the cerebral blood vessels; these may be noradrenergic and peptidergic nerves(Figure 15.2). They influence the cerebrovascular tone and blood flow by secretingclassical transmitters and a number of peptides, including substance P, neuropep-tides and vasoactive intestinal peptide (VIP). This neurogenic influence could alsoexplain the circadian variation in the permeability of the BBB under noradrener-gic influence [17]. The intimate relationships between these cells make the BBB apluricellular interface between the blood and the brain extracellular fluid.

15.1.2.2 Pericytes in the Blood–Brain BarrierPericytes lie periendothelially on the abluminal side of the microvessels (Fig-ure 15.3). A layer of basement membrane separates the pericytes from the endo-thelial cells and the astrocyte foot processes. Pericytes send out cell processeswhich penetrate the basement membrane and cover around 20–30% of the micro-vascular circumference [18]. Pericyte cytoplasmic projections encircling the endo-thelial cells provide both a vasodynamic capacity and structural support to the mi-crovasculature. They bear receptors for vasoactive mediators such as catechola-mines, endothelin-1, VIP, vasopressin and angiotensin II. Pericytes become mark-


Fig. 15.2 Diagram showing a transverse cross-section of a cerebral capil-lary. The endothelial cells, responsible for the main barrier properties of theblood–brain barrier are separated from the astrocyte foot processes, peri-cytes and occasional neurons by the basement membrane. All these com-ponents make up the blood–brain barrier.

edly hypertrophic and hyperplasic in chronic hypertension, and bear increasedamounts of marker proteins. In contrast, pericytes tend to degenerate in neurode-generative processes, such as Alzheimer’s disease, seizures and multiple sclerosis,so increasing the permeability of the BBB. Pericytes are implicated in the BBBand neuroimmune networks. They can be phagocytic in many injuries and maybe actively involved in the regulation of leukocyte transmigration, antigen presen-tation and T-cell activation. They may provide a first line of defense against anti-gens infiltrating into the CNS. CNS pericytes also produce immunoregulatory cy-tokines such as interleukin-1� (IL-1�), IL-6 and the granulocyte–macrophage colo-ny stimulatory factor [19].

CNS pericytes can be viewed as housekeeping scavenger cells and a second lineof defense in the BBB. They are able to carry out pinocytosis and vesicular or tu-bular transport and also contain specific enzymes. Pericytes also play a role in he-mostasis, by producing tissue factors allowing the assembly of the prothrombincomplex. Thus they may be important in coagulation associated with cerebrovas-cular injury. Endothelial cell–pericyte bridges may also be involved in all stages ofnew vessel formation: pericytes guide the migrating endothelial cells, regulatetheir proliferation, form gap junctions and participate in the synthesis of the newbasement membrane [20].


Fig. 15.3 Diagram showing a longitudinal cross-section of the blood–brain barrier, withthe brain capillary endothelial cells sealed by the tight junctions and surrounded by peri-cytes and astrocyte foot processes. These cellular components of the BBB are separatedby a basement membrane.

15.1.2.3 Astrocytes in the Blood–Brain BarrierThe blood capillaries of the CNS of vertebrates are also enveloped by a perivascu-lar sheath of glial cells, mainly astrocytes (Figure 15.3). Immunohistochemicaland morphometric studies on astrocytes and the microvasculature of the humancerebral cortex have shown that the astrocyte perivascular processes form a vir-tually continuous sheath around the vascular walls, with only 11% of the vesselperimeter not being covered [21]. The small portions of the vessel wall lacking anastrocyte envelope seem to be occupied by oligodendrocytes, microgliocytes andneuron bodies and processes; these adhere to the vessel endothelium-pericytelayer [22]. While the astrocytes themselves do not form the barrier, they have animportant role in the development and maintenance of the BBB. Astrocytes re-lease factors that can induce the BBB phenotype and/or the angiogenic transfor-mation of brain endothelial cells in vitro and in vivo [23]. Astrocytes have alsobeen co-cultured with endothelial cells in models of the BBB, and astrocyte-condi-tioned media can reduce the permeability of the BBB in vitro. Transport systemsat the endothelial cells, including the glucose transporter and transport polarity,are up-regulated following exposure to astrocytes.

But the factors released by astrocytes that are responsible for these changes inpermeability remain unknown, as are the cellular mechanisms underlying them.Similarly, the contribution of astroglia to BBB induction in vivo remains controver-sial [24]. BBB-competent endothelial cells may retain BBB properties following achemically induced loss of astroglial end feet. These data suggest that astroglialfactors important for the BBB phenotype may persist in the basement membraneto which the astroglial plasma membrane remains attached, or that neuronal in-fluences are also important.

15.1.2.4 Basement Membrane at the Blood–Brain BarrierThe endothelial cells are separated from the pericytes and astrocytes by a base-ment membrane, also called the extracellular matrix (Figure 15.3). This basementmembrane completely surrounds the CNS and is assembled from components ofthe bloodstream or secreted by the endothelial cells, pericytes and mostly by theperivascular astrocytes. Its main components are type IV collagen, fibronectin,laminins, chondroitin, and heparan sulfate from glycosaminoglycans. The mostabundant component, type IV collagen, polymerizes with laminin and fibronectinproteins via protein-binding domains such as integrin and lectin receptors [25].These components not only provide a mechanical supporting structure for thecapillary wall, they are also important as a negatively charged barrier due to thechondroitin and heparan sulfate residues, in addition to the previously reportedanionic properties of the luminal and abluminal membranes of the endothelialcells [26]. The basement membrane may influence the transcellular resistance ofthe cerebral endothelial cells and the formation of the tight junctions. The base-ment membrane also mediates several other processes, including cell differentia-tion, proliferation, migration and axon growth [27]. Astrocytes are unable to passthrough the basement membrane. This is significant, since astrocytes produce all


the plasminogen activators, proteases and metalloproteinases required to digestthe constituents of the basement membrane [28]. Pathological changes in the per-meability of the BBB are also influenced by the integrity of the basement mem-brane. Matrix-degrading metalloproteinases help modulate BBB permeability bybreaking down the basement membrane macromolecules in such disorders asmultiple sclerosis and other inflammatory processes where factors like cytokinesstimulate the production of these proteolytic enzymes [29].

This overview points out the complexity of the BBB, as at least four types ofcells plus the basement membrane are implicated in its structure and function.Thus many genes and the proteins they encode play a critical role in the broadpharmacological spectrum of activities carried out by the BBB. As drug responsesdepend on numerous proteins in the body, including metabolizing enzymes,transporters, receptors and all signaling networks mediating the response, it isvery likely that there are similar gene–protein-mediated events at the BBB (Fig-ure 15.4). The proteins involved in the formation of tight junctions and the regula-tory interactions between the cerebral endothelial cells and in cross-talk betweenthe cells of the BBB are undoubtedly targets for controlling the BBB permeabilityin normal and pathological conditions. The BBB permeability could be affected by


Fig. 15.4 Diagram showing the main pharma-cokinetic and pharmacodynamic eventsmediated by the gene–protein network withinthe endothelial cells (EC), pericytes (P) andastrocytes (A). The intracellular cross-talkpathways that modulate the blood–brain bar-

rier permeability are indicated by arrowheads.The basement membrane (BM) also contrib-utes to BBB function. RME= receptor medi-ated endocytosis; green box = enzymes; bluecircle= transporters; red box= receptors.

genetic variants under normal physiological conditions, leading to changes in thefunction of proteins like enzymes, transporters and receptors present at the BBB.More than 80% of all sequence variants in the human population are single nu-cleotide polymorphisms (SNPs), although only a few (< 1%) probably change thefunction of the encoded proteins. Thus, identifying an SNP should help to identi-fy the sources of differences in CNS drug responses from one patient to another,and so reduce the incidence of adverse drug reactions. For example, polymorph-isms in genes encoding drug-metabolizing enzymes or transporters expressed atthe BBB could affect the transport of a drug across the BBB and consequently thepatient’s response to it. The BBB is frequently disrupted in disease. Cerebral isch-emia, bacterial meningitis, multiple sclerosis and brain injury all cause openingof the BBB, allowing solutes and water to move into the brain, leading to vaso-genic brain edema. Cytokines, free radicals and proteases are the main cytotoxicfactors appearing in these various unrelated diseases causing genetic defects and/or dysregulation of protein structure and function at the BBB [30]. Knowledge oftheir deleterious effects on the gene–protein network of the BBB architecture mayhelp considerably our understanding of the mechanisms underlying diseases thataffect the integrity of the BBB. They could also provide new therapeutic targetsfor restoring BBB function.

Lastly, pharmacogenomics could provide new tools for the design of more spe-cific and active CNS pharmaceuticals. The efficacy of a broad spectrum of neuro-pharmaceutical drugs is often complicated by their inability to reach their site ofaction because of the BBB. One way to overcome this is to use carrier-mediatedtransport at the luminal and/or abluminal membranes of the endothelial cells ofthe BBB. This will provide a physiologically based drug delivery strategy for thebrain by designing new chemical entities or fused proteins that can cross the BBBvia these transporters.

Thus the pharmacogenomics of the BBB may lead to the discovery of the genesand proteins that are specifically produced in the BBB, and also to the discoveryof the way in which they are dysregulated or defective.

15.2The Main Gene and Protein Targets for Pharmacogenomics of the Blood–Brain Barrier

15.2.1Drug-Metabolizing Enzymes at the Blood–Brain Barrier

The CNS contains much smaller amounts of drug-metabolizing enzymes thandoes the liver. The concentrations of the main enzymes in the brain, members ofthe cytochrome P450 (CYP) superfamily, are only 0.25% of concentration in theliver. But the brain enzymes are not uniformly distributed, as they are in the liver;they are concentrated in specific brain areas. Theoretical models have explainedthat drug metabolism in the CNS cannot influence drug distribution in the blood,but there are marked differences in brain tissue levels depending on the presence

15.2 The Main Gene and Protein Targets for Pharmacogenomics of the Blood–Brain Barrier 319

or absence of brain enzymes in these specific areas [31]. Thus, brain enzymeslocated very near receptor sites for drugs could have a very marked effect on re-ceptor–drug exposure.

Although the absence of paracellular transport across the BBB impedes the en-try of small hydrophilic compounds into the brain, low-molecular-weight lipo-philic substances may pass through the endothelial cell membranes and cytosolby passive diffusion [7]. While this physical barrier cannot protect the brainagainst chemicals, the metabolic barrier formed by the enzymes from the endo-thelial cell cytosol may transform these chemicals. Compounds transportedthrough the BBB by carrier-mediated systems may also be metabolized. Thus,l-DOPA is transported through the BBB and then decarboxylated to dopamine bythe aromatic amino acid decarboxylase [7].

This metabolic barrier was first postulated for amino acid neurotransmitters in1967 [32]. The presence of at least 30 cerebral enzymatic systems suggests that abattery of enzymes may modulate the entry of neuroactive molecules into thebrain [33]. Several phase 1 enzymes, such as CYPs, monoamine oxidases (MAO-Aand -B), flavin-containing monoxigenases, reductases and oxidases and phase 2enzymes catalyzing conjugation, such as UDP-glucuronosyltransferases (UGTs)and glutathione- and sulfotransferases have been found in the rat and humanCNS, as well as in isolated brain microvessels and cerebrovascular endothelialcells in primary culture [34].

These various experiments show that brain capillaries contain significantly higherenzyme activities than does the brain parenchyma itself [35, 36]. Rat brain microves-sels contain several isoforms of CYP1A involved in the metabolism of aromatic poly-cyclic hydrocarbons [36]. The finding that CYP2D6 is responsible for debrisoquinehydroxylation and the formation of morphine from codeine O-demethylation mayhave pharmacogenomic significance [37]. Several of the main drug-metabolizingCYP450 enzymes, such as 1A2, 2A6, 2C9, 2C19 and 2D6 frequently have non-func-tional alleles, resulting in 1–10% poor metabolizers in various ethnic populations forCYP2D6. The prodrug-activating effect of CYPD6 does not occur in poor metaboli-zers on codeine therapy, and this contributes to between-individual variations in thepotential sensitivity to CYP2D6 CNS pharmaceuticals, such as most of the tricyclicantidepressants (nortriptyline and desipramine) [31].

Other oxidative enzymes may also help activate pathways of neurotoxicity at theBBB or within the brain. For example, CYP2E1, which metabolizes ethanol,chlorinated solvents to toxic metabolites, and arachidonic acid to vasoactive com-pounds, could produce active metabolites resulting in oxidative stress and alteringBBB permeability [38]. Between-individual variation in its induction by ethanoland an upstream polymorphism that is common in Orientals are functionally sig-nificant and related to the discovery of novel alleles in the CYP2E1 gene. One ofthese increases the transcription of the gene. The defective CYP2D6 expressionand the wide differences in CYP1A1 and CYP2E1 inducibility result in variationsin the uptake of their substrates by the brain. Other phase 1 enzymes in themammalian brain capillaries, like MAO-B, are active enough to metabolize thelipophilic neurotoxin 1-methyl-4-phenyl-1,2,5,6-tetrahydropyridine (MPTP) into


1-methyl-4-phenyl-pyridinium (MPP+), which is responsible for the appearance ofsevere Parkinson’s disease-like symptoms [39]. This metabolic activity at the BBBcan protect the brain against the neurotoxin MPTP. Similarly, MAO-B inactivatescatecholamine-like substances that cross the BBB and may protect the brainagainst neurotoxic exogenous pyridine analogs [35].

The CYP3A subfamily, which contains the major isoforms expressed in the hu-man liver and interacts with around 50% of pharmaceuticals, has not yet beenfound in brain endothelial cells and parenchyma, in contrast to other polarizedcells like intestinal epithelial cells. The absence of this main CYP subfamilymakes the brain and the BBB unique. The activity of the phase 2 enzyme,UGT 1A6, has been reported to be 6-fold greater in brain microvessels than in thebrain parenchyma and polymorphisms which may affect transferases activitycould also influence the distribution of drugs in the CNS [35]. To illustrate thecomplexity of the enzyme activities in the BBB, enzymes expressed in the peri-cytes and in the astrocyte end feet complement the BBB enzyme arsenal. Gluta-myl aminopeptidase and aminopeptidase, which convert angiotensin II to angio-tensin III and inactivate opioid peptides, are located in the pericytes. There maybe species differences, such as butyrylcholinesterase which deacylates heroin tomorphine is located in the pericytes of the dog brain capillaries, but in the endo-thelial cells of the rat brain capillaries [7]. Adenosine deaminase, which convertsthe cerebral vasodilatator adenosine to inosine, is found mainly in the astrocytefoot processes and may prevent the further distribution of adenosine to receptorson neurons [7]. The protective advantage of more concentrated enzymes in thebrain microvasculature than in parenchyma could fail if there were to be a genedefect, as with CYP2D6, and may enhance the exposure of the brain to CNS-ac-tive drugs (e.g., tricyclic antidepressants, neuroleptics, opioids and serotonin up-take inhibitors). Similarly, the polymorphism affecting the inducibility of CYP1A1and CYP2E1 may give rise to the large between-subject variations in the re-sponses of the CNS to drugs.

15.2.2Drug-Carrier Transporters at the Blood–Brain Barrier

15.2.2.1 Mono- or Bidirectional Transporters for Small CompoundsA second type of drug pharmacokinetic event at the BBB is mediated by proteinson the luminal and/or the abluminal membranes of the endothelial cells. Theseproteins can mediate symmetric and asymmetric drug transport. This type oftransport was first discovered for nutrients that are not lipid-soluble and cannotcross the barrier by simple diffusion. Glucose and some amino acids that thebrain cells cannot manufacture for themselves can be carried in both directionsby facilitated osmotic diffusion [11]. Glucose was the first molecule whose passageinto the brain was linked to a high-capacity transport system present on bothsides of the endothelial cell membranes. The GLUT1 transporter was first clonedin 1985 and is present in the brain capillaries as well as the choroid plexus, theastroglial and neuronal cell membranes. Its state of glycosylation varies from one


cell to another and it can transport about 1 �mol min–1 g–1 at the BBB in man[40]. This capacity has been used to develop the transport into the brain of drugslinked to a d-pyranose sugar moiety. The other main types of glucose transportersare essentially present on neurons (GLUT3) or on both astroglia and neurons(GLUT5) [41].

Like the glucose carrier, the carriers for large neutral amino acids, the so-calledL-system – now designated LAT – are present at both sides of the endothelial cellmembranes and transport at least 10 essential amino acids. The L-transporter atthe BBB has a much higher transport capacity than those in other tissues. Itsmarked preference for phenylalanine analogs explains why the anticancer drugsmelphalan and d,l-NAM-7 are transported by the L-system, as is the l-Dopa usedto treat Parkinson’s disease [42].

Other CNS pharmaceuticals, such as baclofen and gabapentin, cross the BBBvia the L-system. The LAT1 and LAT2 genes were recently detected in whole brainby Northern blotting and may be present in the brain endothelial cells, but thetwo transporters they encode have different substrate specificities. The concentra-tion of LAT1 mRNA at the BBB is much higher (> 100-fold) than in any other tis-sue, including other brain cells like neurons and glial cells. Thus LAT1 is prob-ably the major transporter of large neutral amino acids at the BBB, rather thanLAT2 [43, 44]. The cationic amino acid transporter which mediates the uptake ofarginine, lysine and ornithine by the brain is also present at the BBB; CAT1 andCAT2 have both been cloned. Although they are both present in the brain, onlyCAT1 seems to be present at the BBB and may be the primary basic amino acidtransporter of the BBB [45].

Other transporters can work by active mechanisms requiring a source of energy.These transporters are often present on the luminal or abluminal membranes ofthe endothelial cells (Figure 15.5). The sodium ion-independent amino acid trans-porter, system A, transports small, neutral amino acids across the abluminalmembranes of the brain endothelial cells. Other systems (systems X–, A, Bo+,ASC) are also involved in amino acid transport, but their pharmacological impactat the BBB remains to be elucidated [46]. Monocarboxylic acid compounds like lac-tic acid, ketone bodies, �-hydroxybutyrate and acetoacetate, which are abundant inthe brain, are regulated by both uptake and efflux transporters at the BBB. Themonocarboxylic acid transporter (MCT1) is present on both the luminal and ab-luminal membranes of the BBB and seems to transport substrates in both direc-tions [47]. It is upregulated by increased glycolysis, cerebral lactate and seizure-re-lated factors. MCT1 is responsible for the transport of several drugs, including theHMG-CoA reductase inhibitors simvastatin acid, lovastatin acid, and pravastatin.Interestingly, the anticonvulsant, valproic acid, is more efficiently transportedfrom the brain to the blood than the other way round because an organic aniontransport, probably mediated by OAT1 (organic anion transporter), counterbal-ances the uptake mediated by MCT1 at the BBB [48].

The BBB also has sodium- and pH-independent transporters of organic cations.They are important for the homeostasis of choline and thiamine in the brain andfor the permeation of cationic drugs like propranolol, lidocaine, fentanyl, H1-an-


tagonists and choline analogs. These organic cation transporters (OCTs) aremainly located at nerve terminals, glial cells and in the BBB. Human OCT2 isfound in the brain neurons rather than at the BBB, but a new family of organiccation/carnitine transporters, the OCTNs, seems to be involved in the transport oforganic cations such as mepyramine into brain capillary endothelial cells [48]. Sev-eral sodium-independent facilitative (hENT1 and hENT2) and sodium-dependentactive (CNT1 and CNT2) transporters have been identified. Their presence at theBBB allows the transport of nucleoside-mimetic antivirals like azidothymidine(AZT) and dideoxycytidine (ddC) and the anticancer drug gemcitabine from theblood to the brain. But, as for valproic acid, these drugs are more efficiently ex-cluded from the brain by active transporters sensitive to anionic compounds likeprobenecid.

15.2.2.2 Peptide and Protein TransportersHydrophilic peptides and proteins are frequently large molecules; they may enterthe brain by carrier-mediated transport, receptor-mediated transcytosis, or byadsorptive-mediated transcytosis. Small peptides, such as di- and tripeptides aretransported by the specific transporters, PepT1 and PepT2, but neither of them ispresent at the BBB. Nevertheless, there is saturable brain uptake of the tripeptideglutathione and of several opioid peptides, suggesting that specific transporters, as


Fig. 15.5 Diagram showing some of the nutrient and drug transport processes asso-ciated with the brain capillary endothelial cells that form the BBB. Local transporters inthe luminal or/and abluminal membranes are depicted as filled circles and ones whoselocation is more questionable or that are present at the BBB are depicted in open cir-cles. GLUT1, LAT1, MCT1, oatp2 are present on both the luminal and abluminal mem-branes. This diagram shows that transport may be unidirectional or bidirectional.

yet unidentified, are involved at the BBB. In addition to peptide transporters,adsorptive and receptor-mediated transcytosis are responsible for the transport ofpeptides and proteins across the BBB [49].

The human brain microvessels bear receptors for insulin, insulin-like growthfactor-I, insulin-like growth factor-II, transferrin and leptin. These transportersmediate the transcytosis of these peptides and have been targeted by several strate-gies for delivering drugs to the brain [14]. Neither insulin nor transferrin havebeen coupled to the drug of interest because insulin causes hypoglycemia and thephysiological concentration of transferrin in the plasma is much too high. There-fore, monoclonal murine antibodies against each of the receptors have been usedto prepare high-affinity-binding vectors which still have transcytosis properties.These have been used to deliver attached peptides like VIP, nerve growth factorand brain-derived neurotrophic factors to the brain [7].

Absorption-mediated transcytosis across the BBB is mediated by mechanismsthat depend on the cationic charges of peptides or proteins. The initial binding tothe luminal plasma membrane is mediated by electrostatic interactions with, e.g.,anionic sites on acid residues of acidic glycoproteins. These trigger absorptive en-docytosis. The cationized proteins cross the BBB by adsorption-mediated transcy-tosis aided by the effects of the anionic charges of the abluminal membrane andthose of the basement membrane. More sophisticated cationization processeshave been developed using protein covalently linked to naturally occurring polya-mines like putrescine, spermidine and spermine. Such covalently modified neuro-trophic factors may increase the permeability of the BBB to them [50]. The recentdiscovery that small cationic peptide vectors can transport drugs across complexbiological membranes has opened up new possibilities. Peptide vectors such asTAT peptide, SynB vectors and penetratin, have been used to deliver biologicallyactive substances inside live cells [51]. The antineoplastic agent doxorubicin doesnot readily cross the BBB. Coupling it to a SynB vector significantly enhances itsuptake by the brain without compromising the tight junctions. About 20-timesmore vectorized doxorubicin is delivered to the brain parenchyma than when freedoxorubicin is used. Coupling other CNS pharmaceuticals to SynB vectors, suchas penicillin, peptides like dalargin, proteins up to 67 kDa and other cytotoxicdrugs, significantly improves their uptake by the brain [52].

15.2.2.3 A New Generation of Efflux TransportersThe identification of the brain-to-blood efflux transporter, P-glycoprotein (P-gp), in1992 has added a novel property to the concept of the BBB. P-gp decreases thepermeability of the BBB to lipophilic drugs by actively impeding their crossing ofthe luminal membranes or by transporting them out of the endothelial cells tothe bloodstream [53]. P-gp was originally found as an overproduced membraneprotein in multidrug resistance tumor cells, and was found to be responsible forreducing the intracellular accumulation of several anticancer drugs [48]. Transportmediated by P-gp is coupled with ATP hydrolysis and affects many substrates thathave a planar structure, neutral or cationic charge and are hydrophobic. While


humans have only one gene (MDR1) encoding the drug-transporter P-gp, rodentshave two genes, mdr1a and mdr1b, that encode P-gps with overlapping substratespecificities [54]. The availability of mice with a disrupted mdr1a gene, the mdr1bgene, or with disrupted mdr1a and mdr1b genes, has helped to demonstrate thatthe P-gp in the BBB limits the entry of many drugs into the brain by activelypumping them back into the blood. Most light-microscope and electron-micro-scope immunochemical experiments using several specific antibodies to P-gp indi-cate that the luminal membrane of the brain endothelial cells normally has ahigh concentration of P-gp (Figure 15.5). Nevertheless, a few studies have sug-gested that P-gp is not present in the luminal membrane of the human and pri-mate brain capillary endothelial cells, but is mainly on astrocyte foot processes[55]. It is quite likely that P-gp is present at both the blood luminal membrane ofthe brain capillary endothelial cells and in astrocytes. We have recently identifiedboth mdr1b transcripts and a functional P-gp protein in primary cultures of rat as-trocytes; we have also described greater expressions of the mdr1a gene and its re-sulting protein in rat brain capillaries than in astrocytes [56]. Previously, Tishler etal. found a strong immunostaining for P-gp on both blood capillaries and astro-cytes in the brains of patients with intractable epilepsy [57]. While the function ofP-gp in astrocytes may be questionable, its role as a barrier to the entrance ofsmall lipophilic compounds across the luminal membranes of the brain capillaryendothelial cells is now clearly established. Several functional polymorphisms ofthe human MDR1 gene were recently described and correlated with the synthesisand activity of P-gp in vivo. Analysis of the MDR1 sequence in a Caucasian popu-lation resulted in 15 SNPs, only one of which was located in exon 26, thus reveal-ing the functional polymorphism of the MDR1 gene. About 26% of the popula-tion were homozygous for this variant and the three genotypes gave rise to poor,intermediate and high intestinal uptakes of the cardiotonic drug, digoxin, a P-gpsubstrate [58]. There is probably similar polymorphism at the BBB, and thismight be an extremely important factor influencing between-subject variations inthe uptake of a large number of pharmaceuticals by the CNS. There seem to beseveral other transporters that exclude drugs from the brain, in addition to P-gp.The presence at the BBB of members of the multidrug resistance-associated pro-tein (MRPs) family, whose members preferentially transport anionic compounds,is still controversial. The seven members of the MRP family belong, like P-gp, tothe ATP-binding cassette (ABC) protein superfamily. Mrp1 has been found at theBBB in isolated rat brain capillaries, primary cultures of brain capillary endotheli-al cells and in immortalized capillary endothelial cells, but not in human braincapillaries [59]. Another member, MRP2 has been found at the luminal mem-brane of the brain endothelial cells [60]. However, further studies are required toshow that there are MRP transporters at the BBB (Figure 15.5). As for P-gp, afunctional Mrp1 was found in primary cultured rat astrocytes [56] and it has beenshown to take part in the release of glutathione disulfide from brain astrocytes un-der oxidative stress [61].

The efflux of organic anions seems not to be mediated by MRPs alone. Otherorganic anion transporters, like oatp2 and OATPA, have been detected by immu-


nohistochemistry on both the luminal and abluminal membranes of human braincapillaries [48].

The functional characterization and identification of the proteins and genes ofso many nutrient and drug transporters at the BBB over these last 25 years hasled to a change in our understanding of the way drugs are transported across theBBB. This has also opened an important avenue for the development of attractivestrategies for delivering drugs to the brain using some of these transporters.Therefore, the more recent discovery of efflux transporters has raised the possibili-ty that several processes associating diffusion, influx and efflux may determinethe net drug uptake into the CNS. The inducibility or functional polymorphismaffecting these transporters at the BBB remain to be completely identified. Thisnew information may shed light on the factors involved in inter-and intra-subjectvariations in the uptake of drugs by the CNS. It is thus possible that variations inthe component molecules of the BBB may give rise to poor, intermediate andhigh brain drug uptake.

15.2.3Tight Junctions, Receptors and Cell Cross-Talk at the Blood–Brain Barrier

Morphological studies on tight junctions by freeze-fracturing and ultrathin sec-tioning has shown that the endothelial cells of capillaries of the mammalian brainpossess the most complex tight junctions in the whole vascular system. This com-plexity is reflected in the activities of a pair of proteins which have not yet beencompletely identified and whose synthesis, assembly and regulation remain to beclarified. Tight junctions are formed of pentalaminar layers that result from thefusion of the external leaflets of the partner cell membranes. Many families of in-tegral membrane proteins, including the four-pass protein, occludin, and the prod-ucts of a multigene family, the claudins (claudin-1 and claudin-5), are the maincomponents of tight junctions [62]. The N-terminal domain of occludin regulatesthe transmigration of leukocytes and the C-terminal domain is involved in thecontrol of the paracellular permeability of molecules that involves binding to aprotein kinase. Corticosteroids, which decrease the paracellular uptake of sucroseby the brain, have recently been shown to increase the amounts of occludinmRNA and protein, while reducing the phosphorylation of occludin. This sug-gests that occludin plays a key role in the tightness of the junctions. The intracel-lular terminal ends of occludin also interact with other peripheral membrane pro-teins like zonula occludens proteins (ZO-1, ZO-2, ZO-3), cingulin and 7H6 anti-gen [63], leading to multiple interactions with cytoskeletal elements (actin fila-ments, intermediate filaments) within the endothelial cells.

Yet another family of junction adhesion molecules (JAMs) was recently locatedat the tight junctions of both endothelial and epithelial cells. The intracellular do-main of JAM-1 also interacts with structural and signaling proteins, such as ZO-1and cingulin. Lastly, the molecular organization of the endothelial cell junctionsincludes two other cell–cell contact Ca2+-dependent cadherin–catenin systems.These make up the adherens junction common to all endothelial cell junctions.


The �- and �-catenins help control the transcription of several genes. �-Cateninbinds to transcription factor LEF-1 and to growth factors and probably plays a rolein the translocation of ZO-1 to tight junction complexes [64]. Other proteins thatmediate the cell–substrate adhesion of adherent cells may serve as signal recogni-tion molecules. A special feature of junctions between endothelial cells is the pres-ence of platelet–endothelial cell adhesion molecule-1 (PECAM-1), which is in-volved in angiogenesis and in mediating inflammatory responses such as leuko-cytes–endothelial interactions. The regulation and function of this unique archi-tecture remain poorly understood. The components of intercellular junctions arerich in signaling proteins that form part of several signaling pathways, such asprotein kinases and protein phosphatases. The brain capillary tight junctions arealso very dynamic structures that are sensitive to environmental and cellular fac-tors [61]. The close anatomical relationship between the end feet of astrocytes andneurons suggests that environmental neural factors may determine the character-istics of the blood–brain barrier. Cultured brain capillary endothelial cells haveshort, fragmented tight junctions, but these tight junctions become more denseand longer in the presence of astrocytes and their electrical resistance increasessignificantly [66]. Modulation of the tightness of endothelial tight junctions is animportant objective of pharmacological studies. Like the unidentified soluble fac-tors produced by astrocytes, certain agents which are released during inflamma-tion, such as histamine, thrombin, bradykinine, cytokines and prostacyclins, maycause opening of the blood–brain barrier at the tight junctions of endothelialcells. The transient opening of the BBB allows the delivery of non-permeatingdrugs to the brain. The intravascular infusion of hyperosmolar solutions or the ad-ministration of agonists of the bradykinin and histamine receptors increase thepermeability of the brain vascular system to anticancer agents [7, 14].

The breakdown of the BBB may also facilitate the infiltration of circulatingmonocytes and neutrophils into the brain and hence its invasion by bacteria orviruses [67]. Similarly, activated T-lymphocytes may cross the BBB because of thepresence of adhesion molecules on the vascular endothelial cells that form theBBB. Cadherins are involved in the formation of tight junctions and integrins, in-cluding ICAM-1, VCAM-1 and PECAM-1 and selectins, which can be upregulatedand overproduced following exposure to pro-inflammatory agents or activated T-cells, are all involved in the movement of immune cells across the BBB. Little isknown about the second messenger systems that may be activated to facilitate themargination and diapedesis of immune cells across the BBB. Several signalingpathways, such as Rho-ICAM-1, c-fos and inhibitory kappa�alpha, activate the re-lease of secondary immune response modulators by the endothelial cells like IL-6and -8 or prostaglandins and thromboxanes. This indicates that the BBB can am-plify or dampen immune responses in the brain by creating complex cell-to-cellcross talk. Further work is needed to understand how this occurs.


15.3Pharmacogenomics of the Blood–Brain Barrier

15.3.1Objectives for Pharmacogenomics of the Blood–Brain Barrier

The two previous paragraphs have shed light on the various cellular and non-cel-lular components of the BBB and on the many proteins of pharmacological inter-est expressed at the BBB. Combining these two observations can point to severalobjectives for pharmacogenomic investigations of the BBB. One would be thecomplete characterization of the gene–protein network of each of the componentcells of the BBB. This would lead to a better understanding of the exact role ofeach of these components in the pharmacokinetic and pharmacodynamic eventsoccurring at BBB and to a better definition of the BBB and all its regulation path-ways. A second would be the identification of those genes that are specifically ex-pressed at the BBB. This would clarify the specificity of the cerebral endotheliumand differentiate it from the endothelia of other parts of the body. It could alsoopen the way to the discovery of new specific targets for modulating the perme-ability of the BBB to drugs, e.g., via receptor-mediated transcytosis. A third wouldbe the identification of the SNPs that may introduce polymorphisms in drug me-tabolism or transport. Knowledge of the regulatory mechanisms that lead to theinduction or repression of enzymes or transport proteins could also be extremelyvaluable for predicting between-subject variations in the response of the CNS todrugs. The last objective would be the identification of the defects in genes ortheir encoded proteins in disease states that affect BBB permeability.

15.3.2Current Experimental Approaches and Their Limitations

The pharmacological impact of gene and protein function or dysregulation on theBBB permeability may be investigated by in vitro or/and in vivo experiments. Ani-mals with spontaneous gene mutations or defects, or animals that are geneticallymodified, may be most useful for assessing the overall influence of drug pharma-cokinetics and pharmacodynamics on the response to a drug. But complementaryin vitro studies are needed to discover the molecular mechanisms underlying thegene–protein system. Although current molecular and genomic techniques, suchas DNA and protein microarrays, are more and more accessible, they cannot easi-ly be applied to pharmacogenomic studies of the BBB. These techniques are notsensitive enough to detect endothelial cell-specific transcripts and proteins be-cause of the extremely small amounts of specific BBB transcripts and proteins inwhole brain extracts. Pardridge et al. estimated that the brain microvessels ac-count for about 10–3 parts of the whole brain [68]. As the sensitivity of gene mi-croarrays is about 10–4, it is unlikely that genes selectively expressed by the brainmicrovessels will be detected in extracts of whole brain. This has stimulated at-tempts to isolate the brain capillary endothelial cells from other brain cells and tothe preparation of enriched extracts for DNA and protein analysis.


15.3.2.1 In vitro Pharmacogenomic StudiesThe analysis of gene expression in BBB cells using molecular and genomic tech-niques will require the preparation of pure brain endothelial cells. In vitro modelsof the BBB have evolved from isolated microvessels to primary cultures and im-mortalized and transfected endothelial cells over the last 25 years. As all these invitro systems differ with respect to isolation procedures, cell culture conditions,mono- or co-culture and different cell types in term of origin and species, the en-dothelial cell phenotype and genotype may vary widely [9, 67]. The use of culturedendothelial cells has several advantages, including the availability of unlimitedamounts of genetic and protein extracts and the certainty that there is no contami-nation by other cell components of the BBB, such as astrocyte end feet and peri-cytes. Nevertheless, culture conditions and the number of passages all modify theexpression of BBB-specific genes like GLUT1 and �-GTP, which can be markedlydownregulated or absent. Similarly, the activities of these two specific markers ofthe BBB can be increased in cultured brain cells by conditioning their medium byadding, e.g., astrocyte-derived trophic factors. This great dependency of the geneand protein expression by cultured brain capillary endothelial cells on experimen-tal factors makes this type of model inappropriate for pharmacogenomic studieson the BBB. A careful analysis of BBB-specific gene transcripts and proteins re-quires the isolation of brain capillaries prior to analysis of gene products. Severalmethods for isolation of brain microvessels under RNAase-free conditions havebeen published [68]. The methods that use mechanical or enzymatic tissue disrup-tion give inhomogeneous preparations as pre-capillary arterioles, which are madeup of endothelial cells and smooth muscle cells, are present. The capillaries alsocontain pericytes, which are 3 times more abundant than the brain capillary endo-thelial cells, and the abluminal membrane of isolated brain capillaries may be at-tached to remnants of astrocyte foot processes and nerve endings. Thus, the pur-ity of the isolated brain capillary must be estimated by immunostaining for theglial fibrillary acidic protein (GFAP), which is specific to astrocytes, or measuringcerebrosides as possible neuronal contamination [68]. The detection of pericyte-specific proteins like smooth muscle �-actin or the surface markers CD11b mayhelp to differentiate them from endothelial cells and astrocytes [19]. Endothelialcells may also be isolated from the brain parenchyma using magnetic microbeadscross-linked to an antibody directed against PECAM-1. This method was used toincrease the P-gp content of an endothelial cell fraction 59-fold, while the negativefraction, which contained the astrocyte marker GFAP, contained no P-gp. It wasrecently used to demonstrate the phenotypical changes that occur in the vascularbed markers in the microvasculature of brain tumors [69].

BBB-RNA must be isolated for molecular cloning and analysis of BBB-specifictranscripts, and several methods for isolating BBB-poly (A+) mRNA in one ormore steps have been described [68]). Li et al. recently reported that they obtainedyields of 12 �g poly (A+) mRNA from a single bovine cortical shell and 3.2 �g poly(A+) mRNA from the pooled cerebral hemispheres of 21 rats [70].

The isolation of Poly (A+) mRNA has also led to the synthesis of complementaryDNA for the preparation of BBB-cDNA libraries and the construction of BBB-re-

15.3 Pharmacogenomics of the Blood–Brain Barrier 329

porter genes. These reporter genes can be used for transfecting cultured brain endo-thelial cells and generating cellular tools for gene–protein mechanistic studies.

These experiments are hampered by the risk of obtaining heterogeneous iso-lated brain microvessels and low yields of mRNA. Analysis of BBB-proteomicsmay also be complicated by the polarized distributions of membrane proteins,some being found on the luminal membrane of the brain capillaries and otherson the abluminal membrane. Thus the luminal membranes must be purifiedusing density-gradient centrifugation of homogenized brain tissues following theaddition of colloidal silica particles, which selectively bind to the luminal mem-branes, to the vascular bed. This results in a 10-fold enrichment in GLUT1 and a17-fold enrichment in P-gp over their concentrations in isolated brain capillariesand very little contamination by basolateral membranes. This technique is suit-able for detecting proteins by Western blotting [68].

Despite these limitations, quantitative polymerase chain reaction (PCR) studieson BBB-specific transcripts have been used to measure the concentration of bo-vine GLUT1. More recently it has been combined with subtraction cloning meth-ods to identify the genes that are more actively expressed at the BBB than in pe-ripheral tissues [68]. Rat brain capillary poly (A+) RNA was used to produce testercDNA and rat liver or kidney mRNA was used to generate driver cDNA. The two-run PCR produced cDNA and the resulting large library was used to isolate thefirst 50 clones. More than 80% of the genes were selectively expressed at the BBB.They included novel genes like a 2.6 kb long cDNA sequenced and named BBB-specific anion transporter type 1 (BSAT1). Several genes known to be selective forthe BBB, such as the genes encoding tissue plasminogen activator, insulin-likegrowth factor-2, transferrin receptor, oatp2 transporter and the class I major histo-compatibility complex, were also found [70]. This first study of a BBB genomicsprogram demonstrates that numerous genes with novel sequences encoding pro-teins of unknown function may be selectively expressed at the BBB. This type ofgenomic study could also be applied to BBB-related disorders. An altered gene ex-pression has recently been found in the cerebral capillaries of stroke-prone sponta-neously hypertensive rats (SHRSP), which are a model of hypertension-inducedcerebrovascular lesions. The abnormal gene results in a disturbance of the struc-ture and function of the tight junctions of the BBB endothelial cells prior tostroke. Here, too, three cDNA fragments were found to be up-regulated by sup-pression subtractive hybridization between the SHRSP and stroke-resistant ratbrain capillaries plus a cDNA filter screening step. These changes were associatedwith the pathogenesis of stroke [71].

15.3.2.2 In vivo Pharmacogenomic StudiesThe generation of mice with disrupted genes should allow the evaluation of thepharmacokinetic and pharmacodynamic consequences of the complete, specific in-hibition of particular drug enzymes, transporters or receptors.

The first gene-knockout mouse to become available was a mouse lacking detect-able P-gp in the brain capillary endothelial cells. This has been used to elegantly


demonstrate the real impact of P-gp in the BBB. The amounts of anticanceragents, immunosuppressants and other P-gp drug substrates in the brains ofthese mice were greater than in the brains of normal mice. These increased brainconcentrations may dramatically modify the pharmacological activity of certain P-gp substrates like the dopamine antagonist domperidone, which only produces ananti-emetic effect in P-gp-competent mice due to its selective peripheral activity.The anti-psychotic effect of domperidone becomes its main effect when it is givento mice lacking P-gp, indicating its distribution and activity in the CNS [54].These data suggest that a genetic deficiency in BBB P-gp may have dramatic con-sequences for the over-exposure of the brain parenchyma to drugs. There is also agenetic mutation affecting P-gp at the BBB in an inbred subpopulation of CF1mice. They have a mutation in their mdr1a gene. The absence of functional P-gpat the BBB leads to effects similar to those observed in Pgp knockout mice [72].Like the CF1 mouse, collie dogs are sensitive to the neurotoxic effects of the anti-parasitic drug ivermectin. Sequencing of their cDNA showed that they have a mu-tation due to the deletion of four exonic base pairs. This leads to a premature stopcodon in the messenger RNA and the translation of a truncated inactive form ofP-gp [73]. These data raise the question of how they can be extrapolated to hu-mans. Humans may respond differently from mice or collie dogs to an abnormalgene and the altered gene in the knockout mice may be associated with the al-tered expression of other genes, perhaps with overproduction or a lack of othertransporters in the BBB. Therefore, the absence of P-gp at the BBB does not de-crease the tightness of the junctions, as both normal and P-gp knockout micehave similar low permeabilities to sucrose [72]. These P-gp knockout mice are un-doubtedly a powerful tool for pharmacokinetic and pharmacodynamic studies, andthe recent generation of mice lacking Mrp1 may help provide a better understand-ing of the exact role of this other ABC efflux pump at the BBB. Similarly, the gen-eration of transgenic mice overproducing the monocarboxylic acid MCT1 trans-porter in the brain endothelium may be a fruitful approach to examining thefunction of a gene [74]. This illustrates the wide range of in vivo models thatgenetic engineering may generate. These in vivo transgenic models will be ex-tremely useful for quantifying the pharmacokinetic and pharmacodynamic effectsof gene modification at the BBB.

15.4Conclusion

The recent exponential growth of genomics and proteomics in biological scienceshas provided many new concepts and experimental tools for attacking the puz-zling questions of the function and regulation of the BBB in normal and patho-logical situations. Because the BBB protects the brain from chemicals and infec-tious agents, a complete description of BBB-specific genes and the functions ofthe corresponding proteins will help provide a clearer picture of all the BBB bar-rier functions in the wide range of diseases (infectious, degenerative, autoim-

15.4 Conclusion 331

mune, metabolic, tumors) that afflict the CNS. We may well be on the thresholdof a new era for the discovery of some of the mysterious properties of the BBB.

AcknowledgementsI thank Owen Parkes for editing the English text. This textbook is dedicated totwo of my research colleagues at INSERM Unity 26. The first is Doctor Jeanne-Marie Lefauconnier who will retire in April 2002 after a long and fruitful careerstudying the amino acid transporters at the BBB. The second is Doctor GabyBoschi, who died in October 2001. Her work has provided us with unique in-sights into the impact of brain microdialysis on neuropharmacokinetics.


15.5References

1 Lewandowsky M. Zur Lehre der Cere-brospinalflüssigkeit. Z Klin Med 1900;40:480–494.

2 Ehrlich P. Über die Beziehungen vonchemischer Konstitution, Verteilung undpharmakologischer Wirkung. In: Col-lected Studies in Immunity. New York:John Wiley & Sons; 1906, 567–595.

3 Goldmann EE. Die äußere und innereSekretion des gesunden und kranken Or-ganismus im Lichte der „Vitalen Fär-bung“. Beitr Klin Chir 1906; 64:192–265.

4 Goldmann EE. Vitalfärbung am Zentral-nervensystem. Abh Preuss Akad WissPhys Math K1 1913; 1:1–60.

5 Davson H, Segal MB. Blood–brain bar-rier. In: Physiology of the CSF andBlood–Brain Barriers (Davson H, Segal

MB, Eds.), 1st Edn. Boca Raton: CRCPress; 1996, 49–91.

6 Spector R, Johanson CE. The mamma-lian choroid plexus. Sci American 1989;61:68–74.

7 Pardridge WM. Transport of small mol-ecules through the blood–brain barrier:biology and methodology. Adv Drug DelRev 1995; 17:713–731.

8 Brightman MW. Morphology of blood-brain interfaces. Exp Eye Res 1977; 25:1–25.

9 Drewes LR. Molecular architecture of thebrain microvasculature. J Mol Neurosci2001; 16:93–98.

10 Betz AL, Firth JA, Goldstein GW.

Polarity of the blood–brain barrier: distri-bution of enzymes between the luminaland antiluminal membranes of braincapillary endothelial cells. Brain Res1980; 192:17–28.

11 Goldstein GW, Betz AL. The blood–brain barrier. Sci American 1986; 255:74–83.

12 Broadwell RD, Balin BJ, Salcman M.

Transcytotic pathway for blood-borne pro-tein through the blood–brain barrier.Proc Natl Acad Sci USA 1998; 85:632–636.

13 Oldendorf WH, Cornform ME,

Brown WJ. The large apparent work cap-ability of the blood–brain barrier: a studyof capillary endothelial cells in brain andother tissues of the rat. Ann Neurol1977; 1:409–417.

14 Pardridge WM. Drug delivery to thebrain. J Cereb Blood Flow Metab 1997;17:713–31.

15 Cordon-Cardo C, O’Brien JP, Casals

D, Rittman-Grauer L, Biedler JL, Mel-

amed MR et al. Multi-drug resistance (P-glycoprotein) is expressed by endothelialcells at blood–brain barrier sites. ProcNatl Acad Sci USA 1989; 86:695–698.

15.5 References 333

16 Schmidley JW, Wissig SL. Anionic siteson the luminal surface of fenestrated andcontinuous capillaries of the CNS. BrainRes 1986; 363:265–271.

17 Johanson BB, Isaksson O. Circadialvariation of cerebral vessel vulnerabilityduring adrenaline induced hypertension.In: Cerebral Blood Flow: Effect of Nervesand Neurotransmitters (Heistad DD,

Marais ML, Eds.), 1st Edn. Amsterdam:Elsevier Biomedical; 1982, 367–375.

18 Frank RN, Dutas S, Mancini MA. Peri-cyte coverage is greater in the retinalthan in the cerebral capillaries of the rat.Invest Ophtalmol Visual Sci 1987;28:1089–1091.

19 Balabanov R, Dore-Duffy P. Role ofthe CNS microvascular pericyte in theblood–brain barrier. J Neurosci Res 1998;53:637–644.

20 Diaz-Flores L, Guttierrez R, Varela

H. Angiogenesis: an update. Histol His-topathol 1994; 9:807–843.

21 Virgintino D, Monaghan P, Robert-

son D, Errede M, Bertossi M, Ambros-

si G et al. An immunohistochemical andmorphometric study on astrocytes andmicrovasculature in the human cerebralcortex. Histochem J 1997; 29:655–660.

22 Bertossi M, Virgintino D, Maiorano

E, Occhiogrosso M, Roncali L. Ultra-structural and morphometric investiga-tion of human brain capillaries in nor-mal and peritumoral tissues. UltrastructPathol 1997; 21:41–49.

23 Stewart PA, Wiley MJ. Developing ner-vous tissue induces formation of blood–brain barrier characteristics in invadingendothelial cells: a study using quail-chick transplantation chimeras. Dev Biol1981; 84:183–192.

24 Krum JM, Kenyon KL, Rosenstein JM.

Expression of blood–brain barrier charac-teristics following neuronal loss and as-troglial damage after administration ofanti-thy-1 immunotoxin. Exp Neurol1997; 146:33–45.

25 Vorbrodt AW. Ultracytochemical charac-terization of anionic sites in the wall ofbrain capillaries. J Neurocytol 1989;18:359–368.

26 Washington LC, Rahman J, Klein N,

Male DK. Distribution and analysis of

surface charge of brain endothelium invitro and in situ. Acta Neuropathol 1995;90:305–311.

27 Jucker M, Tian M, Ingram DK. Lami-nins in the adult and aged brain. MolChem Neuropathol 1996; 28:209–218.

28 Bernstein JJ, Karps SM. Migrating fetalastrocytes do not intravasate since theyare excluded from blood vessels by vitalbasement membrane. Int J Dev Neu-roscience 1996; 14:177–180.

29 Rosenberg GA, Kornfeld M, Estrada

E, Kelley RO, Liotta LA, Stetler-Ste-

venson WG. TIMP-2 reduces proteolyticopening of blood–brain barrier by type IVcollagenase. Brain Res 1992; 576:203–207.

30 Banks WA. Physiology and pathology ofthe blood–brain barrier; implications formicrobial pathogenesis, drug deliveryand neurodegenerative disorders. J Neu-rovirol 1999; 5:538–555.

31 Britto MR, Wedlung PJ. Cyto-chrome P-450 in the brain. Potential evo-lutionary and therapeutic relevance of lo-calization of drug-metabolizing enzymes.Drug Met Disp 1992; 20:446–450.

32 Van Gelder NM. A possible enzyme bar-rier for �-aminobutyric acid in the centralnervous system. Prog Brain Res 1967;29:259–268.

33 Mesnil M, Testa B, Jenner P. Xenobiot-ic metabolism by brain monooxygenasesand other cerebral enzymes. Adv DrugRes 1984; 13:95–207.

34 El-Bacha RS, Minn A. Drug metaboliz-ing enzymes in cerebrovascular endothe-lial cells afford a metabolic protection tothe brain. Cell Mol Biol 1999; 45:15–23

35 Minn A, Ghersi-Egea JF, Perrin R,

Leininger B, Siest G. Drug metaboliz-ing enzymes in the brain and cerebralmicrovessels. Brain Res Rew 1991;16:65–82.

36 Ghersi-Egea JF, Leninger-Muller B,

Suleman G, Siest G, Minn A. Localiza-tion of drug-metabolizing enzyme activ-ities to blood–brain interfaces and cir-cum-ventricular organs. J Neurochem1994; 62:1089–1096.

37 Chen ZR, Irvine RJ, Bochner F, Somo-

gyi AA. Morphine formation in rat


brain: a possible mechanism of codeineanalgesia. Life Sci 1990; 46:1067–1074.

38 Montoliu C, Vallés S, Renau-Piqueras

J, Guerri C. Ethanol-induced oxygenradical formation and lipid peroxidationin rat brain: effects of chronic alcoholconsumption. J Neurochem 1994;63:1855–1862.

39 Chiba K, Trevor A, Castagnoli N. Me-tabolism of the neurotoxic tertiary amine,MPTP, by brain monoamine oxidase.Biochem Biophys Res Commun 1984;120:547–578.

40 Gould GW, Holman GD. The glucosetransporter family: structure, functionand tissue specific expression. Biochem J1993; 295:329–341.

41 Bauer H. Glucose transporters in mam-malian brain development. In: Introduc-tion to the Blood–Brain Barrier (Par-

tridge WM, Ed.), 1st Edn. Cambridge:Cambridge University Press; 1998, 175–187.

42 Cornford EM, Young D, Paxton JW.

Melphalan penetration of the blood–brain barrier via the neutral amino acidtransporter in tumor bearing brain. Can-cer Res 1992; 52:138–143.

43 Segawa H, Fukasawa Y, Miyamoto K,

Takeda E, Endou H, Kanai Y. Identifica-tion and functional characterization of aNa+-independent neutral amino acidtransporter with broad substrate selectiv-ity. J Biol Chem 1999; 274:19745–19751.

44 Boado RJ, Li JY, Nagaya M, Zhang C,

Partridge WM. Selective expression ofthe large neutral amino acid transporterat the blood–brain barrier. Proc NatlAcad Sci USA 1999; 96:12079–12084.

45 Stoll J, Wadhwani KC, Smith QR.

Identification of the cationic amino acidtransporter (System y+) of the rat blood–brain barrier. J Neurochem 1993;60:1956–1959.

46 Smith QR, Stoll J. Blood–brain barrieramino acid transport. In: Introduction tothe Blood–Brain Barrier (Partridge

WM, Ed.), 1st Edn. Cambridge: Cam-bridge University Press; 1998, 188–197.

47 Gerhart DZ, Emerson BE, Zhdankina

OY, Leino RL, Drewes LR. Expression ofmonocarboxylate transporter MCT1 bybrain endothelium and glia in adult and

suckling rats. Am J Physiol 1997; 273:E207–213.

48 Tamai T, Tsuji A. Transporter-mediatedpermeation of drugs across the blood–brain barrier. J Pharm Sci 2000; 89:1371–1388.

49 Temsamani J, Scherrmann JM, Rees

AR, Kaczorek M. Brain drug deliverytechnologies: novel approaches for trans-porting therapeutics. Pharm Sci TechnolToday 2000; 2:49–59.

50 Poduslo JF, Curran GL, Gill JS. Pu-trescine-modified nerve growth factor:bioactivity, plasma pharmacokinetics,blood–brain/nerve barrier permeabilityand nervous system biodistribution.J Neurochem 1998; 71:1651–1660.

51 Derossi D, Chassaing G, Prochiantz

A. Trojan peptides: the penetratin systemfor intracellular delivery. Trends Cell Biol1998; 8:84–87.

52 Rousselle C, Smirnova M, Clair P, Le-

fauconnier JM, Chavanieu A, Calas B

et al. Enhanced delivery of doxorubicininto the brain via a peptide-vector-mediated strategy: saturation kinetics andspecificity. J Pharmacol Exp Ther 2001;296:124–131.

53 Drion N, Lemaire M, Lefauconnier

JM, Scherrmann JM. Role of P-glyco-protein in the blood–brain transport ofcolchicine and vinblastine. J Neurochem1996; 67:1688–1693.

54 Schinkel AH. P-glycoprotein, a gate-keeper in the blood–brain barrier. AdvDrug Del Rev 1999; 36:179–194.

55 Golden PL, Pardridge WM. P-glycopro-tein on astrocyte foot processes of un-fixed isolated human brain capillaries.Brain Res 1999; 819:143–146.

56 Decleves X, Regina A, Laplanche JL,

Roux F, Boval B, Launay JM et al.Functional expression of P-glycoproteinand multidrug resistance-associated pro-tein (MRP1) in primary cultures of ratastrocytes. J Neurosci Res 2000; 60:594–601.

57 Tishler DM, Weinberg KI, Hinton

DR, Barbaro N, Annett GM, Raffol C.

MDR1 gene expression in brain of pa-tients with medically intractable epilepsy.Epilepsia 1995; 36:1–6.

15.5 References 335

58 Hoffmeyer S, Burk O, Von Richter O,

Arnold HP, Brockmoller J, Johne A

et al. Functional polymorphisms of thehuman multi-drug resistance gene: mul-tiple sequence variations and correlationof one allele with P-glycoprotein expres-sion and activity in vivo. Proc Natl AcadSci USA 2000; 97:3473–3478.

59 Huai-Yun H, Secrest DT, Mark KS,

Carney D, Elmquist WF, Miller W.

Expression of multi-drug resistance-asso-ciated protein (MRP) in brain microves-sel endothelial cells. Biochem BiophysRes Commun 1998; 243:816–820.

60 Fricker G, Miller D, Bauer B, Nob-

mann S, Gutmann H, Török M et al.Transport of xenobiotics across isolatedbrain microvessels studied by confocalmicroscopy. Presented at the 3rd FEBSAdvanced Lecture Course – ABC2001Gosau, Austria.

61 Hirrlinger J, Gutterer JM, Kussmaul

L, Hamprecht B, Dringen R. Microglialcells in culture express a prominent glu-tathione system for the defense againstreactive oxygen species. Dev Neurosci2000; 22:384–392.

62 Kniesel U, Wolburg H. Tight junctionsof the blood–brain barrier. Cell Mol Neu-robiol 2000; 20:57–76.

63 Schnittler HJ. Structural and func-tional aspects of intercellular junctions invascular endothelium. Basic Res Cardiol1998; 93:30–39.

64 Rubin LL, Staddon JM. The cell biologyof the blood–brain barrier. Annu RevNeurosci 1999; 22:11–28.

65 Stewart PA, Hayakawa EM. Interen-dothelial junctional changes underlie thedevelopmental tightening of the blood–brain barrier. Dev Brain Res 1987;32:271–281.

66 Janzer RC, Raff MC. Astrocytes induceblood–brain barrier properties in endo-thelial cells. Nature 1987; 325:253–257.

67 Miller DW. Immunobiology of theblood–brain barrier. J Neuro Vir 1999;5:570–578.

68 Boado RJ. Molecular biology of braincapillaries. In: Introduction to the Blood–Brain Barrier (Partridge WM, Ed.), 1stEdn. Cambridge: Cambridge UniversityPress; 1998, 151–162.

69 Demeule M, Labelle M, Regina A,

Berthelet F, Beliveau R. Isolation ofendothelial cells from brain, lung, andkidney: expression of the multidrug resis-tance P-glycoprotein isoforms. BiochemBiophys Res Commun 2001; 281:827–834.

70 Li JY, Boado RJ, Pardridge WM.

Blood–brain barrier genomics. J CerebBlood Flow Metab 2001; 21:61–68.

71 Kirsch T, Wellner M, Luft FC, Haller

H, Lippoldt A. Altered gene expressionin cerebral capillaries of stroke-pronespontaneously hypertensive rats. BrainRes 2001; 910:106–115.

72 Dagenais C, Rousselle C, Pollack GM,

Scherrmann JM. Development of an insitu mouse brain perfusion model and itsapplication to mdr1a P-glycoprotein-defi-cient mice. J Cereb Blood Flow Metab2000; 20:381–386.

73 Roulet A, Puel O, Gesta S, Drag M,

Soll M, Alvineri M et al. MDR Phar-macogenetics in dog. Presented at the3rd FEBS Advanced Lecture Course.ABC 2001. Gosau, Austria.

74 Lieno RL, Gerhart DZ, Duelli R, En-

erson BE, Drewes LR. Diet-inducedketosis increases monocarboxylate trans-porter (MCT1) levels in rat brain. Neuro-chem Int 2001; 38:519–527.

Documents

Pharmacogenomics || Pharmacogenomics of the Blood–Brain Barrier