inflammatory mediators and modulation of blood–brain barrier permeability

Cellular and Molecular Neurobiology, Vol. 20, No. 2, 2000

Inflammatory Mediators and Modulation of Blood–BrainBarrier Permeability

N. Joan Abbott1

Received 28 April 1998; revised 8 August and 23 September 1998; accepted 23 September1998

SUMMARY

1. Unlike some interfaces between the blood and the nervous system (e.g., nerveperineurium), the brain endothelium forming the blood–brain barrier can be modulatedby a range of inflammatory mediators. The mechanisms underlying this modulation arereviewed, and the implications for therapy of the brain discussed.

2. Methods for measuring blood–brain barrier permeability in situ include the useof radiolabeled tracers in parenchymal vessels and measurements of transendothelialresistance and rate of loss of fluorescent dye in single pial microvessels. In vitro studieson culture models provide details of the signal transduction mechanisms involved.

3. Routes for penetration of polar solutes across the brain endothelium include theparacellular tight junctional pathway (usually very tight) and vesicular mechanisms. In-flammatory mediators have been reported to influence both pathways, but the clearestevidence is for modulation of tight junctions.

4. In addition to the brain endothelium, cell types involved in inflammatory reactionsinclude several closely associated cells including pericytes, astrocytes, smooth muscle,microglia, mast cells, and neurons. In situ it is often difficult to identify the site of actionof a vasoactive agent. In vitro models of brain endothelium are experimentally simplerbut may also lack important features generated in situ by cell:cell interaction (e.g. induc-tion, signaling).

5. Many inflammatory agents increase both endothelial permeability and vessel diame-ter, together contributing to significant leak across the blood–brain barrier and cerebraledema. This review concentrates on changes in endothelial permeability by focusing onstudies in which changes in vessel diameter are minimized.

6. Bradykinin (Bk)2 increases blood–brain barrier permeability by acting on B2 recep-tors. The downstream events reported include elevation of [Ca21]i, activation of phospholi-pase A2, release of arachidonic acid, and production of free radicals, with evidence thatIL-1b potentiates the actions of Bk in ischemia.

7. Serotonin (5HT) has been reported to increase blood–brain barrier permeabilityin some but not all studies. Where barrier opening was seen, there was evidence foractivation of 5-HT2 receptors and a calcium-dependent permeability increase.

8. Histamine is one of the few central nervous system neurotransmitters found to

1 Division of Physiology, GKT School of Biomedical Sciences, King’s College London, Guy’s Campus,Hodgkin Building, London SE1 1UL, UK. Fax: (44) 207 848 6569. e-mail: [email protected]

2 Abbreviations Used: AA, arachidonic acid; AEP, aminoethylpyridine; AMP, ADP, and ATP, adenosinemono-, di-, and triphosphate; Bk, bradykinin; cAMP and cGMP, cyclic adenosine and cyclic guanosinemonophosphate; [Ca21]i, intracellular calcium concentration; ECV304/C6, cocultured blood–brain bar-rier model using immortalised endothelial cells, ECV304, and C6 glioma cells; EPA, eicosapentanoicacid; 5HT, 5-hydroxytryptamine (5serotonin); IL-1b, interleukin-1b; H1, H2, and H3, histamine recep-tors; UTP, uridine triphosphate; P2, purine/nucleotide receptor; PLC, phospholipase C; SOD, superox-ide dismutase.

131

0272-4340/00/0400-0131$18.00/0 2000 Plenum Publishing Corporation

132 Abbott

cause consistent blood–brain barrier opening. The earlier literature was unclear, but studiesof pial vessels and cultured endothelium reveal increased permeability mediated by H2

receptors and elevation of [Ca21]i and an H1 receptor-mediated reduction in permeabilitycoupled to an elevation of cAMP.

9. Brain endothelial cells express nucleotide receptors for ATP, UTP, and ADP, withactivation causing increased blood–brain barrier permeability. The effects are mediatedpredominantly via a P2U (P2Y2) G-protein-coupled receptor causing an elevation of [Ca21]i;a P2Y1 receptor acting via inhibition of adenyl cyclase has been reported in some invitro preparations.

10. Arachidonic acid is elevated in some neural pathologies and causes gross openingof the blood–brain barrier to large molecules including proteins. There is evidence thatarachidonic acid acts via generation of free radicals in the course of its metabolism bycyclooxygenase and lipoxygenase pathways.

11. The mechanisms described reveal a range of interrelated pathways by whichinfluences from the brain side or the blood side can modulate blood–brain barrier perme-ability. Knowledge of the mechanisms is already being exploited for deliberate openingof the blood–brain barrier for drug delivery to the brain, and the pathways capableof reducing permeability hold promise for therapeutic treatment of inflammation andcerebral edema.

KEY WORDS: blood–brain barrier; inflammation; permeability; electrical resistance;calcium; tight junction; signal transduction; receptor.

INTRODUCTION

The blood–brain barrier formed by the brain endothelium exerts a strong diffusionalrestriction on exchange between blood and brain (Bradbury, 1979, 1992; Davsonand Segal, 1995). This barrier property is generally seen as a mechanism for pro-tecting the brain from unwanted actions of substances circulating in the blood,exerting control over the entry and efflux of substances needed by and excreted fromthe brain, and maintaining tight limits on ion homeostasis necessary for neuronalsignaling and integration (Abbott et al., 1986; Abbott and Romero, 1996). In thisscenario, any increase in permeability of the barrier is seen as potentially deleteriousand damaging. However, a comparison of the brain endothelium forming the blood–brain barrier and the perineurial epithelium forming the outer blood–nerve barrier(Todd et al., 1997) shows that whereas the permeability of the perineurium isrelatively unaffected by a range of inflammatory mediators, the brain endotheliumshows increases in permeability in response to nanomolar to millimolar concentra-tions of these agents (Table I) (Abbott and Revest, 1991; Greenwood, 1992). This

Table I. Vasoactive Agents Reported to IncreaseBlood–Brain Barrier Permeability

Bradykinin, serotonin (5HT), histaminePurine nucleotides: ATP, ADP, AMPPhospholipase A2, platelet activating factorArachidonic acid, prostaglandins, leukotrienesInterleukins: IL-1a, IL-1b, IL-2Macrophage inflammatory proteins MIP-1, MIP-2Complement-derived polypeptide C3a-desArgFree radicals, nitric oxide

Inflammatory Mediators and Blood–Brain Barrier Modulation 133

suggests that one of the functions of the brain endothelium may be to permitmodulation of permeability in response to local chemical signals (Abbott and Re-vest, 1991; Abbott, 1998). Improved understanding of this modulation not only willgive insights into the possible value to the brain of maintaining a barrier whoseproperties can be varied, but also will help to pinpoint the means by which theendogenous mechanisms can be tapped for deliberate control of barrier permeabil-ity. Deliberate manipulations of the blood–brain barrier which have therapeuticvalue include a short-term increase in permeability to permit delivery of polar drugsto the brain and a decrease in permeability to reduce the damaging effects ofinflammation (reviewed by Abbott and Romero, 1996; De Vries et al., 1997).

METHODS FOR STUDYING BLOOD–BRAIN BARRIERPERMEABILITY

Traditional in situ methods for studying brain endothelial permeability and itsmodulation have included following the escape from the circulation of dyes, suchas Evans Blue–albumin and fluorescein (Bradbury, 1979), and measuring the perme-ability 3 surface area product (PS) of the brain parenchymal microvessels forsmall radiotracers such as sucrose and mannitol, introduced by vascular infusionor perfusion (Smith, 1992). More recently, it has been possible to investigate thepermeability of individual microvessels using the accessible pial vessels on thesurface of the brain; these vessels show blood–brain barrier properties (Bundgaard,1982) and have been used to estimate transendothelial electrical resistance, a mea-sure of small ion permeability, by microelectrode techniques and cable analysis(Crone and Olesen, 1982; Butt et al., 1990). The vessels can also be used to quantifypermeability using Lucifer Yellow (457 Da) and rhodamine-conjugated albumin,with fluorescence imaging of the vessles (Easton and Fraser, 1994). While there issome indication that pial vessels are both more leaky and more reactive to chemicalagents than are brain parenchymal vessels (Allt and Lawrenson, 1997), the studiesof the pial microvasculature demonstrate the potential for modulation. Wherecomparison has been possible, it appears that similar principles apply to the vesselsof the brain parenchyma.

Recent improvements in cell culture of the blood–brain barrier mean thatseveral models are now tight enough to study the modulation of transendothelialpermeability (Grant et al., 1998). In vitro models bring several advantages in investi-gating the cellular signal transduction mechanisms by which modulating agentscause changes in permeability (Abbott et al., 1995; De Vries et al., 1997) and helpin interpreting the more complex behavior seen in in situ preparations.

ROUTES FOR PERMEABILITY ACROSS THE ENDOTHELIUM

Normal in situ brain microvessels show well-organized tight junctions whereendothelial cells meet (Nagy et al., 1984; Kniesel and Wolburg, 1999), and there isgood evidence that the complexity and tightness of these junctions (Schulze and

134 Abbott

Firth, 1992; Dejana and Del Maschio, 1995) reduce the paracellular leak pathwayand contribute to the high transendothelial electrical resistance of the brain endothe-lium [typically .1000 V ? cm2 (Crone and Olesen, 1982; Butt et al., 1990)]. Endocy-totic vesicles are rarer in brain endothelium than in nonbrain endothelium, and inmature vessels, fusions of vesicles appear not to form continuous channels forpermeability across the endothelium (Coomber and Stewart, 1985; Stewart, 2000).There is evidence that receptor-mediated endocytosis and transcytosis can form aroute for entry of specific peptides and proteins including transferrin and insulin(Pardridge, 1994), although there is disagreement over the extent to which the boundpeptide/protein is itself transcytosed (Bradbury, 1997). Adsorptive endocytosis andtranscytosis can mediate entry of cationic proteins (Pardridge, 1994). Some studieshave shown increased activity of these routes in the presence of inflammatorymediators (Joo, 1993), but in most cases it has not been possible to concludethat transcytosis mediates increased permeability (e.g., of dyes). Most interest hasconcentrated on the opening of tight junctions (Abbott and Revest, 1991), or ofgaps through the endothelium (Neal and Michel, 1992), that may be responsiblefor sites of dye and tracer leakage. As evidence accumulates for different degreesof barrier opening, it becomes more likely that different routes for tracer penetrationcan be affected sequentially and/or in parallel.

CELL TYPES INVOLVED IN INFLAMMATORY REACTIONS

The microvessels of the cerebral parenchyma are associated with a number ofcell types that may take part in inflammatory reactions. The primary blood–brainbarrier is formed by endothelial cells, but there is evidence that influences fromthe brain, particularly from neurons and from the end feet of astrocytic glia, areresponsible for induction of specific blood–brain barrier features and upregulationof the barrier phenotype (Wolburg and Risau, 1995; Bauer and Bauer, 1999; Saun-ders et al., 1999). Damage to the astrocytes in pathological conditions may thuslead to a secondary barrier breakdown (Abbott et al., 1992). Pericytes and smoothmuscle cells provide a contractile element surrounding the endothelium; the strate-gic location of pericyte processes spanning the tight junctional region has beentaken to suggest a role in controlling the permeability of the paracellular pathway(Shepro and Morel, 1993; Schulze and Firth, 1993). Microgila are derivatives of thehematogenous monocyte/macrophage lineage and appear to form two populations,one of ‘‘resident microglia’’ that entered the brain in early development and forma relatively stable population associated with microvessels, but are also able tomigrate through the brain parenchyma, and ‘‘perivascular microglia’’ that are en-closed by the basement membrane of the vessel wall and that represent a dynamicpopulation able to exchange with monocytes in the blood (Gehrmann et al., 1995).Mast cells are found throughout the brain, with some evidence for higher popula-tions in the thalamus and olfactory bulb, principally in a perivascular location andjuxtaposed to neurons (Dropp, 1976). They are able to secrete vasoactive andinflammatory mediators in response to neuropeptides and direct neural stimulationand may show increased activity in neural pathologies including multiple sclerosis


(Rozniecki et al., 1995). Finally, a number of nerve terminals innervate or run closeto brain capillaries, providing a potential source of neurotransmitters capable ofregulating microvascular permeability (Abbott and Revest, 1991). It is clear thatin vivo studies on the modulating influences of chemical agents may be difficult tointerpret because of the possibility of indirect actions on the brain endotheliummediated by these other cell types. The in vitro situation is simpler, most studiesinvolving only brain endothelium, but it remains possible that some of the propertiesshown by the in situ endothelium are induced by these additional cell types andmay therefore be lost in monoculture in vitro.

RELATION OF VESSEL DIAMETER AND PERMEABILITY

Some of the inflammatory agents discussed cause a reduction in arterial/arterio-lar smooth muscle tone leading to vasodilatation or an increase in venular toneleading to a rise in intravascular hydrostatic pressure (Wahl et al., 1988). Both couldcause stretch of the tight junctional zone in capillaries/venules that could contributeto opening of the paracellular pathway. However, in many cases, such changes intone occur at higher concentrations of applied chemical agents than the changesin permeability, so that pure effects on permeability can generally be distinguished.But to avoid some of these complications, this review concentrates on recent studieswhere permeability can be measured under conditions that minimize such artifacts.Some of the most detailed studies are on pial venular capillaries, where focalapplication of agents avoids effects on resistance vessels.

BRADYKININ

Bradykinin (Bk) can be found in the blood and the brain parenchyma, whichcontains all necessary components of a kallikrein–kinin system (Wahl et al., 1988).A cortical cold lesion in cats was able to cause Bk release in the damaged area andin the surrounding edematous tissue (Maier-Hauff et al., 1984). In ischaemia, Bkmay reach 1027 to 1026 M in the brain interstitial space (Wahl et al., 1988). Earlystudies showed increased brain microvascular permeability to fluorescein (MW376), although not to larger fluorescently labeled albumin or dextran, followingeither superfusion or intracarotid infusion; B2 receptors appeared to be responsible(Unterberg et al., 1984). In frog pial vessels, a decline in transendothelial electricalresistance was found when Bk was applied either intraluminally or by superfusion(Olesen and Crone, 1986; Olesen, 1989). Superfusion of Bk also caused a reversibledecline in transendothelial electrical resistance in rat pial microvessels, with thesteepest part of the dose–response curve in the range 1025 to 1024 M (Figs. 1A andD) (Butt, 1995). These studies suggested a modest opening of tight junctions causedby bradykinin.

P. A. Fraser and co-workers have conducted a series of pharmacological studieson individual rat pial microvessels using Lucifer Yellow as a quantitative permeabil-ity marker. Bk increased permeability by a B2 receptor-mediated pathway, appar-

136 Abbott

Fig. 1. Effects of three inflammatory mediators on transendothelial resistance (Rm) of rat pial microves-sels, measured with a microelectrode and cable analysis technique. (A) Experimental traces during 2-min superfusion of mediators. (B) Bar chart showing results from control vessels and vessels superfusedwith CSF containing modulators, 1023 M serotonin, 1024 M histamine, or 1024 M bradykinin (mean 6 SE).(C, D) Concentration–response relations for histamine (C) and bradykinin (D); both agents exhibited amaximal effect between 1025 and 1024 M, and concentrations above this had no greater effect. At aconcentration of 1025 M or less, electrical resistance was not significantly different from control valuesin CSF. ***Significantly different from control, P , 0.001. From Butt (1995) with permission.

ently by activation of phospholipase A2, release of arachidonic acid, and productionof free radicals. The evidence came from studies in which the effect of Bk wasblocked by a combination of the cyclooxygenase inhibitor indomethacin and the5-lipoxygenase inhibitor nordihydroguaiaretic acid (but not by either separately)and by free radical scavengers superoxide dismutase (SOD) and catalase appliedtogether (Fig. 2) (Sarker and Fraser, 1994b). Other blockers of free radical genera-tion, desferrioxamine and butylated hydroxytoluene, were effective alone inblocking the action of Bk. DesArg9Bk also caused an increase in permeability,blocked by the B1 antagonist desArg9Leu8Bk but not the B2 antagonist HOE140(Sarker and Fraser, 1995). However, as the effects of desArg9Bk were blocked bythe H2 histamine receptor blocker cimetidine, it was proposed that the effect of Bkvia B1 receptors could be indirect, possibly by action on histamine-containing nerveendings or cells. The increase in blood–brain barrier permeability followingischemia/reperfusion appeared to involve Bk since the B2 antagonist HOE 140blocked the permeability increase (Kurokawa and Fraser, 1995). However, theincrease in permeability was much greater than that observed with Bk alone, sug-


Fig. 2. Effects of superfused bradykinin and free radicalscavengers on Lucifer Yellow permeability of pial venularcapillaries of the rat, shown as change in permeability com-pared with control. Free radical scavengers superoxide dismu-tase (SOD; 100 U/ml) and catalase (100 U/ml) applied alonehad no significant effect on the Bk-mediated response, butcombined SOD 1 catalase blocked the response (** P ,0.01). Four microvessels from four rats, 11–13 em in diameter,mean 6 SE. Modified from Sarker and Fraser (1994) with per-mission.

gesting the involvement of other agents. Hu and Fraser (1997) showed that thecytokine interleukin IL-1b greatly enhanced the response to Bk under normalconditions, making IL-1b a likely candidate for the potentiation seen in ischemia.Sarker and Fraser (1998) showed that zaprinast, a selective cGMP phosphodiester-ase inhibitor, potentiated the effect of a low concentration of Bk. An inhibitor ofparticulate guanylate cyclase, leukotriene LTD4, blocked the Bk-mediated re-sponse, confirming the involvement of cGMP.

In vitro studies add information on the signal transduction pathways involved.Revest et al. (1991) showed that bradykinin increases intracellular [Ca21] in primarycultured brain endothelial cells, consistent with evidence from mesenteric vesselsthat elevation in [Ca21]i is involved in tight junctional opening (He and Curry, 1993,1994). The Bk analogue RMP-7, which is of interest clinically for its ability toincrease blood–brain barrier permeability in brain tumors (Inamura et al., 1994;Elliott et al., 1996), appears to mimic the actions of Bk in raising intracellularcalcium in cultured brain endothelial cells (Doctrow et al., 1994). Using a novelendothelial/glial coculture model (ECV304/C6 glioma) to upregulate blood–brainbarrier features, then studying the responses of the endothelial monolayer, Eastonand Abbott (1997) confirmed that the Bk-mediated rise in [Ca21]i was caused byactivation of a B2 receptor. The calcium rise had two components, a transient dueto release of calcium from intracellular stores and a slower ‘‘plateau’’ phase of

138 Abbott

entry across the plasmalemma via store-operated calcium channels, which could beblocked by lanthanum and SKF 96365 (Fig. 3A). The effect of Bk was blocked bythe phospholipase C inhibitor U73122, but inhibitors of phospholipase A2, andcyclooxygenase and lipoxygenase pathways, were ineffective. Parallel studies ontransendothelial electrical resistance showed that Bk caused a reversible reductionin transendothelial electrical resistance (suggesting opening of tight junctions) (Fig.3B) through B2 receptor/phospholipase C activation and elevations of [Ca21]i re-leased from thapsigargin-sensitive stores. However, in contrast to the effects on[Ca21]i, both the phospholipase A2 inhibitor aristolochic acid and the 5-lipoxygenaseinhibitor nordihydroguaiaretic acid blocked the transendothelial electrical resis-tance response to Bk, while the cyclooxygenase inhibitor indomethacin had noeffect. The conclusion was that Bk, perhaps through rises in [Ca21]i, also acted toopen tight junctions by a phospholipase A2-mediated release of arachidonic acidand its metabolism by 5-lipoxygenase but not cyclooxygenase pathways. Thus moni-toring of both [Ca21]i and transendothelial electrical resistance allowed dissectionof the signal transduction pathways involved in tight junctional control. It is clearthat multiple pathways for tight junction modulation exist even within the isolatedbrain endothelium.

SEROTONIN

Serotonin (5-hydroxytryptamine; 5HT) is elevated in the circulation in severalpathological states involving the nervous system. It can be released from circulatingplatelets, but also from mast cells, endothelium, and serotoninergic vascular nerves.

Fig. 3. Effects of bradykinin on a novel blood–brain barrier model, ECV304 endothelial cells grownto confluence above C6 glioma cells. (A) Intracellular [Ca21] monitored with fura-2, the fluorescenceemission ratio (for excitation at 340/380 nm) being a measure of [Ca21]i. Bradykinin (1 eM) caused arapid elevation of [Ca21]i with an initial transient, declining to a plateau; 10 eM lanthanum blocked theplateau phase (calcium entry to refill intracellular stores) and reduced the subsequent response tobradykinin. (B) In parallel experiments 1 eM bradykinin caused a reversible p20% reduction in transen-dothelial resistance across the ECV304 monolayer. From Easton and Abbott (1997); unpublished figure.


Early studies with large intravascular tracers including [14C]inulin and Evans Blue–albumin found no significant leakage to intravenous or intracisternal 5HT (Hardeboet al., 1981). In contrast, Sharma and Dey (1986a, b) showed that barrier openingto Evans Blue–albumin by heat stress or prolonged immobilization in the rat wasaccompanied by elevation of brain and plasma concentrations of 5HT and thatpretreatment with the 5HT-synthesis blocker p-chlorophenylalanine or the 5HT2

antagonist cyproheptadine prevented the blood–brain barrier opening in heat stress.These results provide evidence for involvement of 5HT receptors in barrier opening.Increased blood–brain barrier permeability was observed following elevation ofplasma 5-HT (Sharma et al., 1990). Sharma et al. (1995) reported that p-chlorophe-nylalanine reduced the production of heat-shock protein after trauma to the spinalcord caused barrier opening. Winkler et al. (1995) found that intravascular infusionor topical application (p0.5 mM) of 5HT to the rat parietal cortex caused flatteningof the EEG and leakage of intravascular [131I]sodium and Evans Blue–albumin,an effect blocked by cyproheptadine and ketanserin applied either systemicallyor topically.

There are conflicting observations on the role of calcium in 5HT-mediatedbarrier opening, from studies on pial vessels and on endothelial cells in culture.Olesen (1985) showed that intravascular infusion with 5HT (1026 to 1024 M) de-creased transendothelial electrical resistance of frog pial venules, an effect blockedby the 5HT2 receptor antagonist ketanserin, while superfusion of 5HT had no effect.The phenomenon was dependent on calcium influx, and an action via calcium-mediated endothelial contraction and tight junction opening was proposed. Nochange in barrier permeability to the larger tracer sodium fluorescein could bedetected, suggesting that the effect was confined to small ions. Revest et al. (1991)found no evidence for a 5HT-mediated rise in [Ca21]i in primary cultured rat brainendothelial cells. Butt (1995) found no reduction in transendothelial electrical resis-tance in rat pial vessels superfused with 5HT, although Bk and histamine wereeffective (Fig. 1A). Sarker and Fraser (1996) showed that superfusion or luminalapplication of 5HT caused increased permeability of rat pial microvessels to LuciferYellow (Fig. 4). The superfusion effect was blocked by the 5HT2 receptor antagonistLY53857, by the calcium entry blocker SKF 96365, and by the inhibitor of intracellu-lar calcium store depletion, sodium dantrolene. Taken together, these observationssuggest that serotonin is capable under some conditions of opening the blood–brainbarrier predominantly from the luminal side by a calcium-dependent mechanismbut that, under conditions of stress and brain injury, more widespread and damagingeffects involving 5HT from the nervous system may be seen.

HISTAMINE

Histaminergic neurons from the hypothalamus send out axons contacting neu-rons, astrocytes, and most of the microvessels of the brain (Takagi et al., 1986; Wadaet al., 1991), suggesting that histamine may play an important role in cerebrovascularmodulation. This contrasts with most neuroactive agents involved in neural signal-ing, which generally do not cause barrier opening (Abbott and Revest, 1991).

140 Abbott

Fig. 4. Effect of 5HT on Lucifer Yellow permeability of pialvenular capillaries. Concentration–response curves showing theeffect of 5HT (filled circles) on the permeability of single cerebralvenular capillaries from four microvessels in four rats (diameter,8–13.3 em). Each point is the mean 6 SE of the differencesbetween the control permeability and the permeability in thepresence of 5HT. The effect of 5HT was blocked by the 5HT2

receptor antagonist LY53857 (10 pM; open circles) (analysis ofcovariance, P , 0.001).

Histamine is also released by degranulating mast cells and by the endotheliumitself. In several studies on in situ brain microvessels, histamine applied to either theluminal or the brain side caused a relatively nonselective increase in permeability, asmonitored with dyes of up to 150,000-Da molecular weight; however, other studiesreported negative results, suggesting species and methodological differences (Wahlet al., 1988; Schilling and Wahl, 1994).

Revest et al. (1991) showed that histamine caused an increase in [Ca21]i inprimary cultured rat brain endothelial cells, confirming that the effects seen in situwere at least partly mediated by endothelial histamine receptors. There was suffi-cient consensus within the literature to identify an H2 receptor-mediated response,but some earlier work was interpreted as an increase in pinocytotic activity ratherthan via tight junction opening (reviewed by Joo, 1993). Butt and Jones (1992)and Butt (1995) found histamine caused a reversible decrease in transendothelialelectrical resistance of pial vessels (Figs. 1A–C), arguing in favor of increasedpermeability of a tight junctional pathway; the histamine response could be blockedby the H2 antagonist cimetidine, evidence for an H2 receptor effect. The fact thatcerebral edema induced by experimental pneumothorax in piglets could be reducedby application of both H1 and H2 receptor antagonists was taken as evidence for arole of histamine in cerebral edema (Dux et al., 1987), but the finding was compli-cated by observations that H1 and H2 receptors were present on cerebrovascularsmooth muscle, so that changes in dilatation could have mediated some of the effects.

In an extensive series of studies, Fraser and co-workers have investigated thepharmacology of histamine effects on the permeability of rat pial vessels to Lucifer


Yellow. Easton et al. (1997) showed that after surgical exposure of the pial surface,spontaneous increases in permeability could be detected, with permeability initiallyincreasing to a moderate degree (first phase, PLY 5 1.65 3 1026 cm ? sec21; Fig. 5).Application of histamine to tight vessels gave a similar degree of opening, suggestingthat the first phase could be caused by endogenous histamine release. A secondand larger phase of opening appeared to be due to release of free radicals; seebelow. Sarker et al. (1998) showed that low concentrations of superfused histamine(5 nM to 5 mM) acting on H2 receptors increased permeability, while high concentra-tions (50 to 5 mM) reduced permeability via H1 receptors (Fig. 6A). The low andhigh concentration effects could be mimicked by the H2 and H1 agonists dimapritand a-aminoethylpyridine, respectively, and blocked by the appropriate inhibitors(Fig. 6B). When intracellular cyclicAMP was elevated using the phosphodiesteraseinhibitor rolipram, low concentrations of histamine (and previously subthresholdconcentrations of aminoethylpyridine) now reduced permeability, suggesting thatthe H1 receptor effects were mediated via cAMP. The H2- (but not the H1)-mediatedresponse to dimaprit was blocked by reducing extracellular [Ca21] or by the calciumentry blocker SKF 96365. Luminal application of dimaprit increased permeability,while aminoethylpyridine had no effect, suggesting that H2 receptors may be presentluminally, while H1 receptors are abluminal or on associated nonendothelial cells.The unusual pharmacology observed (coupling of H2 receptors to elevation of [Ca21]i

and of H1 receptors to cAMP) was unexpected but may prove to be a specialization

Fig. 5. Time course of development of spontaneous changes inLucifer Yellow permeability in the pial venular capillary of a rat(diameter, 27 em), timed from the removal of the meninges. Perme-ability was initially below the detection level of the technique andlater increased in two distinct phases. The first phase was a smallincrease in permeability to ,5 3 1026 cm ? sec21. The second phasewas characterized by fluctuating permeability reaching much higherlevels (as leaky as mesenteric capillaries); later experiments showedthat this second phase could be reduced by more gentle exposureof the capillaries or by treatment with free radical scavengers. FromEaston et al. (1997) with permission.

142 Abbott

Fig. 6. Effects of histamine on Lucifer Yellow permeability of pial venular capillaries. (A) Concentra-tion–response relation. Histamine was applied to the abluminal surface of four to eight microvessels(diameters, 10–16 em) (filled circles) from 12 rats after control measurements of permeability in thesame vessels (open circles). Mean 6 SE. Significant differences from paired t tests are indicated: *P ,0.05; **P , 0.01; ***P , 0.001. Concentrations of histamine ,1025 M increased permeability, whileconcentrations . 1025 M decreased permeability. (B) Effects of the histamine receptor antagonistsmepyramine (H1; 3 nM), cimetidine (H2; 2 eM), and thioperamide (H3; 2nM) on the histamine-mediatedpermeability change. Cimetidine blocked the increased permeability caused by 5 eM histamine(**P , 0.01), while mepyramine blocked the permeability reduction caused by 5 mM histamine (*P ,0.05). From Sarker et al. (1998) with permission.

at the blood–brain barrier by which the same agonist (histamine) can generateopposite effects depending on the local concentration and possibly the site ofhistamine release. The measurements of transendothelial electrical resistance re-ported earlier (Fig. 1C) showed only a reduction in resistance (increased permeabil-ity) to histamine, at concentrations .1025 M. This difference from the LuciferYellow experiments may reflect differences in the methods used—microelectrodestudies of transendothelial electrical resistance give average values measured overtens to hundreds of micrometres of vessel length, whereas the Lucifer Yellowmeasurements can reflect permeability of a single intercellular junctional zone.

NUCLEOTIDES

Extracellular ATP is produced as a conventional transmitter from purinergicnerve terminals but is also released from synaptic vesicles of noradrenergic terminalsas a cotransmitter and by nonsynaptic release from cells including endothelium andsmooth muscle (reviewed by Boarder and Hourani, 1998). ATP is rapidly degradedby ectonucleotidases to ADP and AMP. Aggregating platelets release ATP andthe pyrimidine UTP. Olesen and Crone (1986) were among the first to demonstrateincreased permeability (drop in transendothelial electrical resistance) of pial vessels(frog) in response to ATP, ADP, and AMP. Revest et al. (1991) showed that theseagents elevated [Ca21]i in primary cultured rat brain endothelial cells, with the orderof potency (ATP . ADP . AMP) implicating a P2 purinergic receptor. Moredetailed examination of primary cultured human and rat brain endothelial cells,


and the immortalized brain endothelial cell line RBE4, confirmed the presence of aG-protein-coupled P2U (5 P2Y2) receptor, on which ATP and UTP were equipotent(Purkiss et al., 1994; Nobles et al., 1995). Studies on clonal cell lines of rat brainendothelial cells revealed a phospholipase C (PLC)-coupled P2Y2 receptor, but cellline B10 showed an unusual P2Y receptor that caused an elevation of [Ca21]i by aPLC-independent mechanism (Frelin et al., 1993; Vigne et al., 1994b; Feolde et al.,1995), later shown to be a P2Y1 receptor acting via inhibition of adenylate cyclase(Webb et al., 1996); it remains to be demonstrated that this receptor is also presentin situ. Taken together, these studies show that brain endothelial cells possessreceptors for ATP, UTP, and ADP that can lead to an elevation of intracellularcalcium by a variety of pathways. Opening of tight junctions may be the result ofthe calcium elevation itself or may involve other signal transduction pathways,as ‘‘cross-talk’’ has been detected among cAMP-, cGMP-, and Ca21-dependentintracellular signaling mechanisms in brain endothelial cells (Vigne et al., 1994a;Nobles and Abbott, 1996).

ARACHIDONIC ACID

Arachidonic acid (AA) is a polyunsaturated fatty acid produced by severalbiochemical pathways including the action of phosholipase A2 on membrane lipids.It is generated in ischemia and has been implicated in cerebral edema (Chan andFishman, 1984; Wahl et al., 1988). Intracerebral infusions of AA increased thepenetration of the nonmetabolized small solute [14C]aminoisobutyric acid from theblood (Ohnishi et al., 1992). Free radicals have been implicated in the blood–brainbarrier effects of arachidonic acid. Thus Wei et al. (1986) showed that the increasedblood–brain barrier permeability to albumin caused by AA superfusion could beprevented by the free radical scavengers catalase and SOD, while Shi et al. (1995)showed that the increase in permeability to sucrose and albumin across brainendothelial monolayers caused by AA was blocked by application of several antioxi-dants. Easton and Fraser (1998) showed that the Lucifer Yellow permeability ofpial vessels was increased dose dependently by superfused AA (0.25–2 mM). Asimilar response was seen with eicosapentanoic acid (EPA), which produces metabo-lites with little permeability-increasing effects (Sardesai, 1992), suggesting that theactions of these fatty acids are not due to their prostaglandin and leukotrienemetabolites. SOD and catalase used in combination (but not alone) blocked theresponse to both AA and EPA. The iron chelator desferrioxamine, an inhibitor oflipid peroxidation at the concentration used (100 mM), also blocked the responseto both AA and EPA. Combined blockade of both the cyclooxygenase pathwaywith indomethacin and the 5-lipoxygenase pathway with nordihydroguaiaretic acidabolished the responses to AA and EPA. This study provides evidence that AAincreases brain endothelial permeability by generation of free radicals in the courseof its metabolism via the cyclooxygenase and lipoxygenase pathways and that lipidperoxidation by extracellular free radicals is the important step in the permeabil-ity increase.

144 Abbott

CONCLUSION

This review has concentrated on five classes of ‘‘classical’’ inflammatory media-tor released from relatively well-identified sites acting by relatively clear-cut mecha-nisms involving receptors (Bk, 5HT, histamine, nucleotides) or metabolism to freeradicals (AA). It has deliberately not covered the emergent role of nitric oxide(NO) and has mentioned cytokines only briefly, since their mechanisms of actionare even less well understood, and in most cases it has not yet been possible toconduct rigorous quantitative studies. Except in the case of Bk/IL-1b, the reviewhas not attempted to address the interactions between inflammatory mediators thatcertainly occur in situ, and whose study is beginning to reveal further levels ofcomplexity in the way the system operates in physiology and pathology. Neverthe-less, even from the simplified account presented here, it is clear that we alreadyknow sufficient about the mechanisms for increasing permeability to start to applythem in patients for deliberate blood–brain barrier opening for drug delivery (e.g.,RMP-7). The mechanisms for reducing blood–brain barrier permeability (e.g., toreduce the severity of brain edema) are so far less well understood and thereforeless exploited, but increased understanding, for example of agents elevating cAMP,may lead to improved therapies also in this area. The detailed study of the complexcellular events activated by these inflammatory mediators is certain to providemajor insights and inspiration for clinical applications in the future.

REFERENCES

Abbott, N. J. (1998). Role of intracellular calcium in regulation of blood-brain barrier permeability. InPardridge, W. M. (ed.), An Introduction to the Blood-Brain Barrier: Methodology and Biology,Cambridge University Press, Cambridge, pp. 345–353.

Abbott, N. J., and Revest, P. A. (1991). Control of brain endothelial permeability. Cerebrovasc. BrainMetab. Rev. 3:39–72.

Abbott, N. J., and Romero, I. A. (1996). Transporting therapeutics across the blood-brain barrier. Mol.Med. Today 2:106–113.

Abbott, N. J., Bundgaard, N., and Cserr, H. F. (1986). Comparative physiology of the blood-brainbarrier. In Suckling, J., Rumsby, M. G., and Bradbury, M. W. B. (eds.), The Blood-Brain Barrierin Health and Disease, Ellis Horwood, Chichester, pp. 52–72.

Abbott, N. J., Revest, P. A., and Romero, I. A. (1992). Astrocyte-endothelial interaction: Physiologyand pathology. Neuropathol. Appl. Neurobiol. 18:424–433.

Abbott, N. J., Roux, F., Couraud, P.-O., and Begley, D. J. (1995). Studies on an immortalized brainendothelial cell line: Differentiation, permeability and transport. In Greenwood, J., Begley, D. J., andSegal, M. B. (eds.), New Concepts of a Blood-Brain Barrier, Plenum Press, New York, pp. 239–249.

Allt, G., and Lawrenson, J. G. (1997). Is the pial microvessel a good model for blood-brain barrierstudies? Brain Res. Rev. 24:67–76.

Bauer, H.-C., and Bauer, H. (1999). Neural induction of the blood–brain barrier: Still an enigma. Cell.Mol. Neurobiol. 20:13–28.

Boarder, M. R., and Hourani, S. M. O. (1998) The regulation of vascular function by P2 receptors:Multiple sites and multiple receptors. Trends Pharmacol. Sci. 19:99–107.

Bradbury, M. W. B. (1979). The Concept of a Blood-Brain Barrier, J. Wiley, Chichester.Bradbury, M. W. B. (ed.) (1992). Physiology and Pharmacology of the Blood-Brain Barrier, Springer-

Verlag, Berlin.Bradbury, M. W. B. (1997). Transport of iron in the blood-brain-cerebrospinal fluid system. J. Neuro-

chem. 69:443–454.Bundgaard, M. (1982). Ultrastructure of frog cerebral and pial microvessels and their impermeability

to lanthanum ions. Brain Res. 241:57–65.


Butt, A. M. (1995). Effect of inflammatory agents on electrical resistance across the blood-brain barrierin pial microvessels of anaesthetized rats. Brain Res. 696:145–150.

Butt, A. M., and Jones, H. C. (1992). Effect of histamine and antagonists on electrical resistance acrossthe blood-brain barrier in rat brain-surface microvessels. Brain Res. 569:100–105.

Butt, A. M., Jones, H. C., and Abbott, N. J. (1990). Electrical resistance across the blood-brain barrierin anaesthetized rats: A developmental study. J. Physiol. 429:47–62.

Chan, P. H., and Fishman, R. A. (1984). The role of arachidonic acid in vasogenic brain edema. Fed.Proc. 43:210–213.

Coomber, B. L., and Stewart, P. A. (1985). Morphometric analysis of CNS microvascular endothelium.Microvasc. Res. 30:99–115.

Crone, C., and Olesen, S.-P. (1982). Electrical resistance of brain microvascular endothelium. BrainRes. 241:49–55.

Curry, F. E. (1992). Modulation of venular microvessel permeability by calcium influx into endothelialcells. FASEB J. 6:2456–2466.

Davson, H., and Segal, M. B. (1995). Physiology of the CSF and Blood-Brain Barriers, CRC Press, BocaRaton, FL.

De Vries, H. E., Kuiper, J., De Boer, A. G., Van Berkel, T. J. C., and Breimer, D. D. (1997). Theblood-brain barrier in neuroinflammatory diseases. Pharmacol. Rev. 49:143–155.

Dejana, E., and Del Maschio, A. (1995). Molecular organization and functional regulation of cell tocell junctions in the endothelium. Thromb. Haemostas. 74:309–312.

Doctrow, S. R., Abelleira, S. M., Curry, L. A., Heller-Harrison, R., Kozarich, J. W., Malfroy, B.,McCarroll, L. A., Morgan, K. G., Morrow, A. R., Musso, G. F., Smart, J. L., Straub, J. A., Turnbull,B., and Gloff, C. A. (1994). The bradykinin analog RMP-7 increases intracellular free calcium levelsin rat brain microvascular endothelial cells. J. Pharmacol. Exp. Ther. 271:229–237.

Dropp, J. J. (1976). Mast cells in mammalian brain. I. Distribution. Acta Anat. 94:1–21.Dux, E., Temesvari, P., Szerdahelyi, P., Nagy, A., Kovacs, J., and Joo, F. (1987). Protective effect of

antihistamines on cerebral oedema induced by experimental pneumothorax in newborn piglets.Neuroscience 22:317–321.

Easton, A. S., and Abbott, N. J. (1997). The effects of bradykinin on a cell culture model of the blood-brain barrier. J. Physiol. 505:49–50P.

Easton, A. S., and Fraser, P. A. (1994). Variable restriction of albumin diffusion across inflamed cerebralmicrovessels. J. Physiol. 475:147–157.

Easton, A. S., and Fraser, P. A. (1998). Arachidonic acid increases cerebral microvascular permeabilityby free radicals in single pial microvessels of the anaesthetized rat. J. Physiol. 507:541–547.

Easton, A. S., Sarker, M. H., and Fraser, P. (1997). Two components of blood-brain barrier disruptionin the rat. J. Physiol. 503:613–623.

Elliott, P. J., Hayward, N. J., Huff, M. R., Nagle, T. L., Black, K. L., and Bartus, R. T. (1996). Unlockingthe blood-brain barrier: A role for RMP-7 in brain tumor therapy. Exp. Neurol. 141:214–224.

Feolde, E., Vigne, P., Breittmayer, J. P., and Frelin, C. (1995). ATP, a partial agonist of atypical P2Y

purinoceptors in rat brain microvascular endothelial cells. Br. J. Pharmacol. 115:1199–1203.Frelin, C., Breittmayer, J. P., and Vigne, P. (1993). ADP induces inositol phosphate-independent intracel-

lular Ca21 mobilization in brain capillary endothelial cells. J. Biol. Chem. 268:8787–8792.Gehrmann, J., Matsumoto, Y., and Kreutzberg, G. W. (1995). Microglia: Intrinsic immuneffector cell

of the brain. Brain Res. Rev. 20:269–287.Grant, G. A., Abbott, N. J., and Janigro, D. (1998). Understanding the physiology of the blood-brain

barrier: In vitro models. News Physiol. Sci. 13:287–293.Greenwood, J. (1992). Experimental manipulation of the blood-brain barrier and blood-retinal barriers.

In Bradbury, M. W. B. (ed.), Physiology and Pharmacology of the Blood-Brain Barrier, SpringerVerlag, Berlin, pp. 459–479.

Hardebo, J. E., Owman, Ch., and Wiklund, L. (1981). Influence of neurotransmitter monoamines andneurotoxic analogues on morphological blood-brain barrier function. In Cervos-Navarro, J., andFristschka, E. (eds.), Cerebral Microcirculation and Metabolism, Raven Press, New York, pp.177–180.

He, P., and Curry, F. E. (1993). Differential actions of cAMP on endothelial [Ca21]i and permeabilityin microvessels exposed to ATP. Am. J. Physiol. 265:H1019–1023.

He, P., and Curry, F. E. (1994). Endothelial hyperpolarization increases [Ca21]i and venular microvesselpermeability. J. Appl. Physiol. 76:2288–2297.

Hu, D.-E., and Fraser, P. A. (1997). Evidence for interleukin-1b mediating enhanced permeabilityresponses to bradykinin in single pial venular capillaries of anaesthetized rats. J. Physiol. 505:53P.

Inamura, T., Nomura, T., Bartus, R. T., and Black, K. L. (1994). Intracarotid infusion of RMP-7, abradykinin analog—A method for selective drug-delivery to brain-tumors. J. Neurosurg. 81:752–758.

146 Abbott

Joo, F. (1992). The cerebral microvessels in culture, an update. J. Neurochem. 58:1–17.Joo, F. (1993). The role of second messenger molecules in the regulation of permeability in the cerebral

endothelial cells. Adv. Exp. Med. Biol. 331:155–164.Joo, F. (1995). Isolated brain microvessels and cultured cerebral endothelial cells in blood-brain barrier

research: 20 years on. In Greenwood, J., Begley, D. J., and Segal, M. B. (eds.), New Concepts ofa Blood-Brain Barrier, Plenum Press, New York, pp. 229–237.

Kniesel, U., and Wolburg, H. (1999). Tight junctions of the blood–brain barrier. Cell. Mol. Neuro-biol. 20:57–77.

Kurokawa, T., and Fraser, P. A. (1995). A bradykinin antagonist prevents cerebral microvascularpermeability increase following reperfusion in rats. Physiol. 483:140P.

Maier-Hauff, K., Baethmann, A. J., Lange, M., Schurer, L., and Unterberg, A. (1984). The kallikrein-kinin system as mediator in vasogenic brain edema. Part 2: Studies on kinin formation in focal andperifocal brain tissue. J. Neurosurg. 61:97–106.

Nagy, Z., Peters, H., and Huttner, I. (1984). Fracture faces of cell junctions in cerebral endotheliumduring normal and hyperosmotic conditions. Lab. Invest. 50:313–322.

Neal, C. R., and Michel, C. C. (1992). Transcellular openings through microvascular walls in acutelyinflammed frog mesentery. Exp. Physiol. 77:917–920.

Nobles, M., and Abbott, N. J. (1996). Effects of cyclic nucleotides on the increase in [Ca21]i caused byexternal ATP in the brain endothelial cell line, RBE4. J. Physiol. 491:38P.

Nobles, M., Revest, P. A., Couraud, P.-O., and Abbott, N. J. (1995). Characteristics of nucleotidereceptors that cause elevation of cytoplasmic calcium in immortalized rat brain endothelial cells(RBE4) and in primary cultures. Br. J. Pharmacol. 115:1245–1252.

Ohnishi, T., Posner, J., and Shapiro, W. R. (1992). Vasogenic brain edema induced by arachidonic acid:Role of extracellular arachidonic acid in blood-brain barrier dysfunction. Neurosurgery 30:545–551.

Olsen, S.-P. (1985). A calcium-dependent reversible permeability increase in microvessels in frog braininduced by serotonin. J. Physiol. 361:103–113.

Olesen, S.-P. (1989). An electrophysiological study of microvascular permeability and its modulationby chemical mediators. Acta Physiol. Scand. 136 (Suppl 579):1–28.

Olesen, S.-P., and Crone, C. (1986). Substances that rapidly augment ionic conductance of endotheliumin cerebral venules. Acta Physiol. Scand. 127:233–241.

Pardridge, W. M. (1994) New approaches to drug delivery through the blood-rain barrier. Trends Biotech.12: 239–245.

Purkiss, J. R., West, D., Wilkes, L. C., Scott, C., Yarrow, P., Wilkinson, G. F., and Boader, M. R. (1994).Stimulation of phospholipase C in cultured microvascular endothelial cells from human frontal lobeby histamine, endothelin and purinoceptor agonists. Br. J. Pharmacol. 111: 1041–1046.

Revest, P. A., Abbott, N. J., and Gillespie, J. I. (1991). Receptor-mediated changes in intracellular[Ca21] in cultured rat brain capillary endothelial cells. Brain Res. 549: 159–161.

Rozniecki, J. J., Hauser, S. L., Stein, M., Lincoln, R., and Theoharides, T. C. (1995). Elevated mast celltryptase in cerebrospinal fluid of multiple sclerosis patients. Ann. Neurol. 37:63–66.

Sardesai, V. M. (1992). Biochemical and nutritional aspects of lipoproteins and leukotriene biosynthesisas potential therapeutic targets. Progr. Drug Res. 37:9–90.

Sarker, M. H., and Fraser, P. A. (1994). Evidence that bradykinin increases permeability of singlecerebral microvessels via free-radicals in the rat. J. Physiol. 479:36P.

Sarker, M. H., and Fraser, P. A. (1995). Bradykinin and des-Arg9-bradykinin increase permeability ofsingle cerebral microvessels by different mechanisms in the rat. J. Physiol. 483:141P.

Sarker, M. H., and Fraser, P. A. (1996). Effect of 5-HT in the regulation of cerebral microvascularpermeability in the anaesthetized rat. J. Physiol. 491:29–30P.

Sarker, M. H., and Fraser, P. A. (1998). Evidence that bradykinin increases cerebral microvascularpermeability via cGMP generated independently from soluble guanylate cyclase in the anaesthetizedrat. J. Physiol. 506:22P.

Sarker, M. H., Easton, A. S., and Fraser, P. A. (1998). Regulation of cerebral microvascular permeabilityby histamine in the anaesthetized rat. J. Physiol. 507:909–918.

Saunders, N. R., Knott, G. W., and Dziegielewska, K. M. (1999). Barriers in the immature brain. Cell.Mol. Neurobiol. 20:29–40.

Schilling, L., and Wahl, M. (1994) Opening of the blood-brain barrier during cortical superfusion withhistamine. Brain Res. 653:289–296.

Schulze, C., and Firth, J. A. (1992) Interendothelial junctions during blood-brain barrier developmentin the rat: Morphological changes at the level of individual tight junctional contacts. Dev. BrainRes. 69:85–95.

Schulze, C., and Firth, J. A. (1993). Junctions between pericytes and the endothelium in rat myocardialcapillaries: A morphometric and immunogold study. Cell Tissue Res. 271:145–154.


Sharma, H. S., and Dey, P. K. (1986a). Probable involvement of 5-hydroxytryptamine in increasedpermeability of blood-brain barrier under heat stress in young rats. Neuropharmacology 25:161–167.

Sharma, H. S., and Dey, P. K. (1986b). Influence of long-term immobilization on regional blood-brainbarrier permeability, cerebral blood flow and 5-HT level in conscious normotensive young rats. J.Neurol. Sci. 72:61–76.

Sharma, H. S., Olsson, Y., and Dey, P. K. (1990). Changes in blood-brain barrier and cerebral bloodflow following elevation of circulating serotonin level in anaesthetized rats. Brain Res. 517:215–223.

Sharma, H. S., Olsson, Y., and Westman, J. (1995) A serotonin inhibitor, p-chlorophenylalanine reducesthe heat-shock protein response following trauma to the spinal cord—An immunohistochemicaland ultrastructural study in the rat. Neurosci. Res. 21:241–249.

Shepro, D., and Morel, N. M. L. (1993). Pericyte physiology. FASEB J. 7:1031–1038.Shi, F., Cavitt, J., and Audus, K. L. (1995). 21-Aminosteroid and 2-(aminomethyl)chromans inhibition

of arachidonic acid-induced lipid peroxidation and permeability enhancement in bovine brainmicrovessel endothelial cell monolayers. Free Radical Biol. Med. 19:349–357.

Smith, Q. (1992). Methods of study. In Bradbury, M. W. B. (ed.), Physiology and Pharmacology of theBlood-Brain Barrier, Springer Verlag, Berlin, pp. 23–52.

Stewart, P. A. (2000). Endothelial vesicles in the blood-brain barrier: Are they related to permeability?Cell. Mol. Neurobiol. 20:149–163.

Takagi, H., Morishima, Y., Matsuyama, T., Hayashi, H, Watanabe, T., and Wada H. (1986) Histaminergicaxons in the neostriatum and cerebral cortex of the rat: A correlated light and electron microscopicimmunocytochemical study using histidine decarboxylase as a marker. Brain Res. 364:114–123.

Todd, B. A., Sedgwick, E. M., and Abbott, N. J. (1997) Effects of the bile salt sodium deoxycholate,protamine, and inflammatory mediators on the potassium permeability of the frog nerve perineu-rium. Brain Res. 776:214–221.

Unterberg, A., Wahl, M., and Baethmann, A. (1984). Effects of bradykinin on permeability and diameterof pial vessels in vivo. J. Cereb. Blood Flow Metab. 4:574–585.

Vigne, P., Lund, L., and Frelin, C. (1994). Cross talk among cyclic AMP, cyclic GMP, and Ca21-dependentintracellular signalling mechanisms in brain capillary endothelial cells. J. Neurochem. 62:2269–2274.

Vigne, P., Feolde, E., Breittmayer J. P., and Frelin, C. (1994). Characterization of the effects of2-methylthio-ATP and 2-chloro-ATP on brain capillary endothelial cells: similarities to ADP anddifferences from ATP. Br. J. Pharmacol. 112:775–780.

Wada, H., Inagaki, N., Yamatodani, A., and Watanabe, T. (1991) Is the histaminergic neuron systema regulatory center for whole brain activity? Trends Neurosci. 14:415–418.

Wahl, M., Unterberg, A., Baethmann, A., and Schilling, L. (1988). Mediators of blood-brain barrierdysfunction and formation of vasogenic brain edema. J. Cereb. Blood Flow Metab. 8:621–634.

Webb, T. E., Feolde, E., Vigne, P., Neary, J. T., Runberg, A., Frelin, C., and Barnard, E. A. (1996)The P2Y purinoceptor in rat brain microvascular endothelial cells couple to inhibition of adenylatecyclase. Br. J. Pharmacol. 119:1385–1392.

Wei, E. P., Ellison, M. D., Kontos, H. A., and Povlishock, J. T. (1986). O2 radicals in arachidonate-induced increased blood-brain barrier permeability to proteins. Am J. Physiol. 251:H693–H699.

Winkler, T., Sharma, H. S., Stalberg, E., Olsson, Y., and Dey, P. K. (1995). Impairment of blood-brain barrier function by serotonin induces desynchronization of spontaneous cerebral corticalactivity—Experimental observations in the anesthetized rat. Neuroscience 68:1097–1104.

Wolburg H., and Risau, W. (1995). Formation of the blood-brain barrier. In Kettenmann, H., andRansom, B. R. (eds.), Neuroglia, Oxford University Press, Oxford, pp. 763–776.

inflammatory mediators and modulation of blood–brain barrier permeability

Documents