ejp European Journal of Pharmacology molecular pharmacology

ELSEVIER Molecular Pharmacology Section 269 (1994) 243-248

Changes in brain microvessel endothelial cell monolayer permeability induced by adrenergic drugs

Nuno Borges b,., Fenglin Shi a, Isabel Azevedo b, Kenneth L. Audus a

a Department of Pharmaceutical Chemistry, The University of Kansas, School of Pharmacy, Lawrence, KS 66045-2504, USA b Institute of Pharmacology and Therapeutics, Faculty of Medicine, 4200 Porto, Portugal

Received 11 May 1994; revised MS received 19 July 1994; accepted 26 July 1994

Abstract

Brain microvessel endothelial cell monolayers have been shown to be a suitable blood-brain barrier in vitro system to study adrenergic regulation of permeability. We tested adrenergic drugs on bovine brain microvessel endothelial cell monolayer permeability to a biomembrane impermeant molecule, sodium fluorescein. Endogenous catecholamines noradrenaline and adrenaline were tested as well as the a-adrenoceptor agonist phenylephrine, the fl-adrenoceptor agonist clenbuterol and the a-adrenoceptor antagonist prazosin. Results showed an a-adrenoceptor mediated increase and a fl-adrenoceptor mediated decrease in monolayer permeability. Both a- and/3-adrenoceptor mediated changes in permeability were abolished by inhibiting fluid-phase pinocytosis, either by vincristine or by avoiding bovine brain microvessel endothelial cell's energy utilization. The reverse transport (i.e., from brain to blood side) was also influenced by adrenergic drugs; a- or fl-adrenoceptor stimulation induced a permeability-reducing effect. We conclude that a-adrenoceptor stimulation increases bovine brain microvessel endothelial cell monolayer permeability and that /3-adrenoceptor stimulation has the opposite effect. Reverse transport results obtained with /3-adrenoceptor stimulation seem controversial and deserve further study. These results also support in vivo findings that demonstrated adrenergic influences on blood-brain barrier permeability.

Keywords: Brain microvessel endothelial cell; Catecholamine; Blood-brain barrier; Pinocytosis

1. Introduct ion

Molecule exchange between blood and brain inter- stitial fluid is a tightly controlled process. This is pro- vided by the unique characteristics of brain capillary endothelial cells, namely the presence of tight junc- tions between ceils, the scarcity of pinocytotic activity and the high variety and activity of enzymatic systems. These very specialized cells and their special arrange- ment and features constitute the blood-brain barrier. The once held idea of the blood-brain barr ier being a static barrier has been progressively replaced by the concept of a dynamic one, as many studies have demonstra ted that some substances are able to modu- late its permeabili ty (Audus et al., 1992).

The demonstrat ion of the existence of both adreno- ceptors in brain capillary endothelial ceils (Kobayashi

* Corresponding author. Tel.: (3512) 595694; Fax: (3512) 598119.

0922-4106/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0 9 2 2 - 4 1 0 6 ( 9 4 ) 0 0 1 2 0 - 0

et al., 1982; Har ik et al., 1980) as well as of adrenergic nerve terminals apposed to their basal lamina (Edvins- son et al., 1973; Har tman, 1973; Rennels and Nelson, 1975) has stimulated the hypothesis of a functional role of central adrenergic innervation in the control of blood-brain barrier permeability. Indeed, a number of studies have shown that either the administration of adrenergic drugs to the brain or the stimulation of the cell bodies of central adrenergic nerves (mainly local- ized in the locus coeruleus) can produce changes in the transport of water and solutes across the blood-brain barrier (Raichle et al., 1975; Preskorn et al., 1980; Sarmento et al., 1991, 1994). Furthermore, some of these changes in blood-brain barrier permeabili ty are concomitant with alterations in the morphology of brain capillary endothelial cells, especially in their vesicular activity (Sarmento et al., 1988, 1990, 1991). This may suggest that pinocytosis is an important transendothe- lial route for the passage of molecules across the blood-brain barrier in situations of adrenergic stimula- tion.

244 N. Borges et al. / European Journal of Pharmacology - Molecular Pharmacology Section 269 (1994) 243-248

The development of in vitro models has facilitated cel lular /molecular level characterization of the blood- brain barrier (Joo, 1993). For the study of catechol- amine actions, it has the additional advantage of elimi- nating hemodynamic variations that could account for changes in permeability. Such models also circumvent the problem of drug access to the brain side of the blood-brain barrier, which is a limitation of in vivo studies. It seemed then appropriate to use such an approach in order to further characterize the adrener- gic influences on the transport of molecules through the blood-brain barrier. For this purpose, we used a model consisting of a monolayer of polarized (Guillot and Audus, 1991) bovine brain microvessel endothelial cells. These cells have been extensively characterized as a suitable in vitro model (Audus and Borchardt, 1986a, 1987; Baranczyk-Kuzma et al., 1986; Miller et al., 1992).

In this study, we tested the influence of the endoge- nous catecholamines noradrenaline and adrenaline, as well as an a-adrenoceptor agonist (phenylephrine), a /3-adrenoceptor agonist (clenbuterol) and an a-adren- oceptor antagonist (prazosin).

2. Materials and methods

2.1. Cell culture

Bovine brain microvessel endothelial cells were iso- lated from the grey matter of cerebral cortices accord- ing to the previously described protocol of Audus and Borchardt (Audus and Borchardt, 1986a, 1987; Miller et al., 1992). Isolated bovine brain microvessel en- dothelial cells were grown to monolayers in primary culture in 100 mm culture dishes containing translu- cent 13 mm diameter polycarbonate membrane sup- ports with 3 g m pore size. These culture dishes were pretreated with rat-tail collagen and fibronectin. The culture medium consisted of minimum essential medium/Eagle ' s modified, F12 nutrient mix, 10 mM Hepes, 13 mM sodium bicarbonate, pH 7.4, 10% plasma derived equine serum, 100 /xg/ml penicillin G, 100 /xg/ml streptomycin, 2.5 /zg /ml amphotericin B and 100 /xg /ml heparin. Cells were refed every 3 days until the development of confluent monolayers, which gen- erally occurred 8-10 days after seeding.

2.2. Transendothelial transport studies

After obtaining the monolayers, each polycarbonate membrane was lifted out the culture dish, rinsed with a modified phosphate buffered saline (PBS) (in mM: NaCI 129, KC1 2.5, Na2HPO 4 7.4, K H 2 P O 4 1.3, CaC1

0.63, MgSO 4 0.74, glucose 5.3 and ascorbic acid 0.1, pH 7.4) and placed in a side-by-side diffusion cell (Crown Glass Co.) with 3 ml of PBS in the donor and receptor chambers. Details of the use of side-by-side diffusion cells for the study of molecule transport across bovine brain microvessel endothelial cell monolayers have been previously described (Audus and Borchardt, 1987; Shi et al., 1993). These diffusion cells were thermostated at 37°C by means of an external water bath. Each cham- ber was constantly stirred at 600 rpm with a diffusion cell console (Crown Glass Co.). At time 0, an aliquot of PBS dissolved sodium fluorescein (Sigma) was added to the donor chamber in order to obtain a final concen- tration of 25 /zg /ml . At times 2, 10, 20, 30, 40, 55 and 70 min after time 0, a sample of 200 /xl was collected from the receptor chamber, diluted up to 2 ml with PBS and assayed for the content of sodium fluorescein by spectrophotofluorimetry (SLM-AMINCO Model 4800) using excitation and emission frequencies of 490 and 520 nm, respectively. Readings of sodium fluores- cein standards were linear over a concentration range of 1-500 ng /ml (r = 0.999). Volume in the receptor chamber was maintained constant by adding fresh PBS aliquots after each sample collection.

Except for the case of vincristine (given at time 0 min), all drugs were added immediately after the col- lection of the 30 min sample. The results are expressed as the ratio between the slope of the straight line defined by the amount of sodium fluorescein at times 2, 10, 20 and 30 min (before drug application) and the slope of the straight line defined by the amount of sodium fluorescein at times 40, 55 and 70, i.e., after the application of drug.

Two types of experiments were performed: in the first type, the apical side of the monolayer cells faced the donor chamber and the basal side the receptor chamber. Drugs were added to the receptor chamber. In the second type the monolayer was placed in the opposite way, i.e., the apical side facing the receptor chamber and the basal side facing the donor one. Drugs were added to the donor chamber.

The first type of experiment mimics the in vivo blood-to-brain transport and the second one mimics the in vivo brain-to-blood transport. For this reason and for greater simplicity, the first and second types of experiments will be given the names of blood-to-brain and brain-to-blood experiments, respectively.

In two additional experimental series, we used a protocol identical to the first one with the only differ- ence of the replacement of glucose in the PBS with 6-deoxy-D-glucose at the concentration of 50 mM in order to avoid energy utilization by the monolayer cells.

Aliquots of PBS-dissolved drugs were added in or- der to obtain the following final concentrations: nor- adrenaline (0.1, 1 and 10 /zM), adrenaline (0.1 and 1

N. Borges et aL / European Journal of Pharmacology - Molecular Pharmacology Section 269 (1994) 243-248 245

3

2.5

2

1.5

1

0.5

0

~g/ml

z~ ~ /

J

/

0 10 20 30 40 50 60 70

min Fig. 1. Flux of sodium fluorescein across bovine brain microvessel endothelial cell monolayers. Represented are sodium fluorescein concentrations in the receptor chamber from time 0 to 70 min of a control experiment (no drug added, open circles) and an experiment where 1 ~M noradrenaline was added at time 30 min (closed circles).

/xM), clenbuterol (0.1 /zM), phenylephrine (0.1 /xM), prazosin (0.1/zM) and vincristine (1 /zM).

2.3. Statistical analysis

Comparison between slope values within each ex- perimental series was performed by the Student's t-test for paired observations.

2. 4. Drugs

All drugs were purchased from Sigma (St. Louis, MO, USA).

3. Results

Fig. 1 shows two typical plots of sodium fluorescein concentration appearing in the receptor chamber against time of a control experiment (no drug added at any time; open circles) and an experiment where per- meability was increased by the addition of a drug (noradrenaline 1 /xM; closed circles). Unlike the con- trol line, the noradrenaline 1 /xM plot shows an in- crease in the slope value after addition of the drug.

3.1. Blood-to-brain transport experiments

Using this protocol, we tested the effect of nor- adrenaline (0.1, 1 and 10 /zM), adrenaline (0.1 and 1 /zM), clenbuterol (0.1 /xM), phenylephrine (0.1 /zM) and noradrenaline (0.1/zM) + prazosin (0.1 #M).

Results are presented in Fig. 2. Noradrenaline in- duced an increase in permeability at doses of 0.1 and 1 /zM (47.4 + 14.8%, n = 4, P < 0.05 and 81.1 + 12.9%, n = 7, P < 0.0001, respectively). The higher dose of noradrenaline, 10/xM, did not produce any significant effect ( -6 .9 + 5.7%, n = 5, P = 0.3). The two doses of adrenaline showed opposite effects: 0.1 /zM signifi-

140% 120%

100%

80%

60%

40%

20%

0%

-20%

-40% ,lit T

NA NA NA AD AD CLE PHE NA 0.1,uM l uM 10pM 0.1pM 1,uM 0 .1 ,uM 0.1,uM 0.1pM+

PRAZ 0.1,uM

Fig. 2. Effect of 0.1, 1 and 10/~M noradrenaline (NA), 0.1 and 1/zM adrenaline (AD), 0.1 tzM clenbuterol (CLE), 0.1 /~M phenylephrine (PHE) and 0.1 tzM NA+ 0.1/zM prazosin (PRAZ) on blood-to-brain bovine brain microvessel endothelial cell monolayer permeability. Represented are means+S.E.M, of the percent changes in the slopes (see Materials and methods section). * Statistically significant change (P < 0.05).

cantly increased (34.4 + 9.3%, n = 5, P < 0.025) and 1 /xM significantly decreased ( - 14.3 ___ 1.3%, n = 5, P < 0.001) the permeability to sodium fiuorescein. Clen- buterol (0.1 #M) produced a reduction in permeability ( - 23.2 + 3.9%, n = 4, P < 0.01) and phenylephrine (0.1 /zM) induced an increase in sodium fluorescein trans- port (95.4 + 28.3%, n = 5, P < 0.05). Noradrenaline (0.1 /~M) administration in the presence of the a- adrenoceptor antagonist prazosin (0.1 /zM) resulted in a decrease in monolayer permeability (-22.7 + 4.3%, n = 4, P < 0.02).

Fig. 3 shows the effect of vincristine (1/zM) and the replacement of glucose in PBS (see Materials and methods section) on noradrenaline (1 /zM) and clen- buterol (0.1/zM) effects.

Both treatments abolished the 1 /zM noradrenaline induced increase (2.3 + 9% with vincristine and - 12.6 + 3.7% in the abscence of glucose, n = 5 and P > 0.05 in both cases) and the 0.1 ~M clenbuterol induced decrease ( -6 .3 + 15.8% with vincristine and 1.6 +

100%

80%

60%

40%

20%

0%

-20%

-40% NA 1,uM CLE 0.1pM NA+VIN CLE+VIN

lpM 1,uM NA + 6dG CLE+6dG

Fig. 3. Effect of vincristine (VIN) or glucose replacement by 6-deoxy- D-glucose (6dG) on the changes induced by 0.1 /zM noradrenaline (NA) or 0.1/zM clenbuteroi (CLE). Represented are means ± S.E.M. of the percent changes in the slopes (see Materials and methods section). * Statistically significant change (P < 0.05).

246 N. Borges et aL / European Journal of Pharmacology - Molecular Pharmacology Section 269 (1994) 243-248

0%

-5% -10% -15% -20% -25%

-30%

-35%

-40% ± '~ CLE 0.1,uM NA 0.1,uM AD 0.1pM

Fig. 4. Effect of 0.1 /xM clenbuterol (CLE), noradrenaline (NA) and adrenaline (AD) on brain-to-blood bovine brain microvessel en- dothelial cell monolayer permeability. Represented are means+ S.E.M. of the percent changes in the slopes (see Materials and methods section). * Statistically significant change (P < 0.05).

6.2% in the abscence of glucose, n = 5 and P > 0.05 in both cases) in sodium fluorescein transport across bovine brain microvessel endothelial cell monolayers.

3.2. Brain-to-blood transport experiments

This protocol was used to test the effects of the following drugs and doses: noradrenaline (0.1 ~M), adrenaline (0.1/.~M), clenbuterol (0.1/zM).

Results are presented in Fig. 4. Clenbuterol ( -28 .2 +7.2%, n = 4 , P <0 .05 ) and adrenaline ( - 2 7 . 4 + 8.0%, n = 4, P < 0.05) induced a reduction in the transport of sodium fluorescein across the bovine brain microvessel endothelial cell monolayers. Noradrenaline ( - 24.6 + 8.6%, n --- 3, P > 0.05) did not produce a statistically significant alteration in permeability.

4. Discussion

The idea of a central adrenergic control of blood- brain barrier permeability has been suggested by the demonstration of central adrenergic nerve varicosities apposed to the basal lamina of brain capillary endothe- lial cells (Edvinsson et al., 1973; Hartman, 1973; Ren- nels and Nelson, 1975) and by the presence of both types of adrenoceptors (a and /3) in these cells (Harik et al., 1980; Kobayashi et al., 1982). Indeed, some authors have demonstrated that these features may constitute a synaptic specialization, controlling sub- stance traffic across the blood-brain barrier (Raichle et al., 1975; Sarmento et al., 1991, 1994). From these studies, the idea has emerged that a-adrenoceptor stimulation (either by the addition of a-adrenoceptor agonists or by nerve stimulation) induces an increase in blood-brain barrier permeability. Furthermore, Sar- mento et al. (1991) showed that in situations of a- adrenoceptor stimulation there is an increase in the vesicular activity of brain capillary endothelial cells, wich suggests that pinocytosis may be a transendothe-

lial route for the passage of substances in this situation. The role of/3-adrenoceptors present in these cells has been more difficult to characterize. Nevertheless, it seems that they may have a protective role in condi- tions of blood-brain barrier disruption (Sarmento et al., 1994; Borges et al., 1994).

Every study previously made in this area has been done using in vivo models, so it seemed of interest to use an in vitro model, as it circumvents some of the methodological problems of using living animals (mainly hemodynamic variations and difficulties in drug deliv- ery to the brain). The model described by Audus and Borchardt (1986a,b, 1987) has been extensively charac- terized as a suitable polarized (Guillot and Audus, 1991) in vitro blood-brain barrier model and was there- fore used in our study.

The results obtained in the blood-to-brain transport experiments demonstrate that drugs possessing a- adrenoceptor agonist properties induce an increase in the monolayer permeability to sodium fluorescein. This effect is evident for the lower doses of noradrenaline (0.1 and 1 /xM) and adrenaline (0.1 /zM) and for the selective a-adrenoceptor agonist phenylephrine. The /32-adrenoceptor agonist clenbuterol had an opposite effect; it significantly reduced the monolayer perme- ability to sodium fluoreseein. The choice of a selective /32-adrenoceptor agonist like clenbuterol was based on its ability to demonstrate a permeability reducing effect in an in vivo model (Borges et al., 1994) and in the results of Kobayashi et al. (1981) who found a predomi- nance in the number of /32- over /31-adrenoceptors (80% and 20% of the total /3-adrenoceptor population, respectively) in these cells. The loss of effect seen when the noradrenaline and adrenaline doses were increased to 10 and 1 /zM, respectively (1 /zM adrenaline even showing a significant reduction in per- meability), is probably due to their mixed a- and /3- adrenoceptor agonist properties; higher doses of these two amines are probably capable of acting on/3-adren- oceptors whose stimulation may reduce the monolayer permeability (as seen for clenbuterol), thus counteract- ing the previously established a-adrenoceptor medi- ated effect. The difference in noradrenaline and adrenaline doses necessary to produce this effect is in good agreement with adrenaline's higher affinity for /32-adrenoceptors when compared with that of nor- adrenaline. The reduction of sodium fluorescein per- meability obtained when the a-adrenoceptor antago- nist prazosin was given together with noradrenaline is also compatible with the hypothesis of dual a- and /3-adrenoceptor mediated effects of catecholamines: when the a-adrenoceptors are blocked by prazosin, noradrenaline can only act on /3-adrenoceptors, with a resultant effect similar to that obtained with the selec- tive /32-adrenoceptor agonist clenbuterol.

In order to verify the hypothesis of Sarmento et al.

N. Borges et al. / European Journal of Pharmacology - Molecular Pharmacology Section 269 (1994) 243-248 247

(1991, 1994), that pinocytosis is an important transendothelial route for molecule transport in situa- tions of impaired adrenergic modulation, we tested the effect of vincristine and of glucose absence in the PBS on the permeability alterations induced by noradrena- line and by clenbuterol. Vincristine blocks the vesicle formation by inhibiting microtubule formation (Lars- son et al., 1979) and the replacement of glucose by 6-deoxy-D-glucose avoids energy utilization by the cell, thus inhibiting energy-dependent processes like pinocy- tosis (Guillot et al., 1990). The observed inhibition of noradrenaline-dependent increase and clenbuterol-de- pendent decrease in sodium fluorescein permeability by vincristine and by glucose replacement provides good evidence for a pinocytosis-mediated process. However, we cannot rule out the presence of another cell-regulated paracellular pathway contributing to these observations (Guillot and Audus, 1991). Addi- tional microscopic or biochemical evidence may help to confirm the in vitro/in vivo correlation.

Brain-to-blood transport results obtained with drugs having mainly a-adrenoceptor agonist effects (nor- adrenaline 0.1/xM and adrenaline 0.1/~M) are in good agreement with the results obtained in blood-to-brain transport experiments: it seems logical that if a sub- stance increases molecule transport across a monolayer in one direction it will reduce it in the opposite one. Yet, the results obtained with clenbuterol showing a reduction in permeability similar to that obtained in blood-to-brain transport, do not fit in this theory. The explanation may be that /3-adrenoceptor stimulation could induce a kind of vesicle traffic in brain capillary endothelial ceils where vesicles never reached the basal side thus preventing its content to mix with the brain interstitial fluid and therefore reducing permeability. This seems especially evident in conditions of in- creased transport across the blood-brain barrier like the in vivo cold-induced brain edema (Borges et al., 1994) and in this model of bovine brain microvessel endothelial cell monolayers, which have a greater per- meability than the in vivo blood-brain barrier under normal conditions (Miller et al., 1992).

In conclusion, the use of an in vitro model has shown an a-adrenoceptor mediated increase in apical- to-basal transport across brain capillary endothelial ceils, a process most likely involving increased pinocy- totic activity. The mechanism of action of fl-adrenoc- eptors in these cells is less evident, but our results are compatible with the hypothesis of their protective physiological role in conditions of increased blood-brain barrier permeability. Together with in vivo studies, this work may provide a better understanding of the regula- tory actions of central adrenergic innervation over the blood-brain barrier, and ultimately the ability of ma- nipulating this system in the management of patients with impaired blood-brain barrier function.

Acknowledgment

N.B. is a PhD student with a grant from JN1CT (Programa Ci~ncia).

References

Audus, K.L. and R.T. Borchardt, 1986a, Characteristics of the large neutral amino acid transport system of bovine brain microvessel endothelial cell monolayers, J. Neurochem. 47, 484.

Audus, K,L. and R.T. Borchardt, 1986b, Characterization of an in vitro blood-brain barrier model system for studying drug trans- port and metabolism, Pharm. Res. 32, 81.

Audus, K.L. and R.T. Borchardt, 1987, Bovine brain microvessel endothelial cell monolayers as a model system for the blood-brain barrier, Ann. NY Acad. Sci. 507, 9.

Audus, K.L., P. Chikhale, D.W. Miller, S.E. Thompson and R.T. Borchardt, 1992, Brain uptake of drugs: the influence of chemical and biological factors, Adv. Drug Res. 23, 1.

Baranczyk-Kuzma, A., K.L. Audus and R.T. Borchardt, 1986, Cate- cholamine-metabolizing enzymes of bovine brain microvessel en- dothelial cell monolayers, J. Neurochem. 46, 1956.

Borges, N., A. Sarmento and I. Azevedo, 1994, Adrenergic influ- ences in brain edema formation and resolution, J. Auton. Phar- macol. 14, 13 (abstract).

Edvinsson, L., O. Lindvall, K. Nielsen and C. Owman, 1973, Are brain vessels innervated by central (non-sympathetic) adrenergic neurons?, Brain Res. 63, 496.

Guillot, F,L and K.L. Audus, 1991, Angiotensin peptide regulation of bovine brain microvessel endothelial cell permeability, J. Car- diovasc. Pharmacol. 18, 212.

Guillot, F,L., K.L. Audus and T.J. Raub, 1990, Fluid-phase endocy- tosis by primary cultures of bovine brain microvessel endothelial cell monolayers, Microvasc. Res. 39, 1.

Harik, S.I., V.K. Sharma, J.R. Weatherbae, R.H. Warren and S.P. Banerjee, 1980, Adrenergic receptors of cerebral microvessels, Eur. J. Pharmacol. 61, 207.

Hartman, B.K., 1973, The innervation of cerebral blood vessels by central noradrenergic neurons, in: Frontiers in Catecholamine Research, eds. E. Usdin and S.H. Snyder (Pergamon Press, New York) p. 91.

Joo, F., 1993, The blood-brain barrier in vitro: the second decade, Neurochem. Int. 23, 499.

Kobayashi, H., A. Cazzaniga, P. Spano and M. Trabucchi, 1982, Ontogenesis of alpha and beta receptors located on cerebral microvessels, Brain Res. 242, 358.

Kobayashi, H., T. Maoret, M. Ferrante, P. Spano and M. Trabucchi, 1981, Subtypes of /3-adreneigic receptors in rat cerebral mi- crovessels, Brain Res. 220, 194.

Larsson, B., T. Skarby, L. Edvinsson, J.E. Hardebo and C. Owman, 1979, Evidence for involvement of microtubules in the enhanced transendothelial pinocytosis in the blood-brain barrier induced by high intravascular pressure, Acta Neurol. Scand. 60 (Suppl. 80), 72.

Miller, D.W., K.L. Audus and R.T. Borchardt, 1992, Application of cultured endothelial cells of the brain microvasculature in the study of the blood-brain barrier, J. Tissue Cult. Methods 14, 217.

Preskhorn, S.H., B.K. Hartman, M.E. Raichle and H.B. Clark, 1980, The effect of dibenzazepines (tricyclic antidepressants) on cere- bral capillary permeability in the rat in vivo, J. Pharmacol. Exp. Ther. 213, 313.

Raichle, M.E., B.K. Hartman, J.O. Eichling and L.G. Sharpe, 1975, Central noradrenergic regulation of cerebral blood flow and vascular permeability, Proc. Natl. Acad. Sci. USA 72(9), 3726.

248 N. Borges et al. // European Journal of Pharmacology - Molecular Pharmacology Section 269 (1994) 243-248

Rennels, M.L. and E. Nelson, 1975, Capillary innervation in mam- malian central nervous system: an electron microscopic demon- stration, Am. J. Anat. 144, 233.

Sarmento, A., A. Albino-Teixeira and I. Azevedo, 1988, Amitripty- line induced morphological alterations of the rat blood-brain barrier, Eur. J. Pharmacol. 176, 69.

Sarmento, A., A. Albino-Teixeira and I. Azevedo, 1990, Increase in blood-brain barrier permeability due to amitriptyline is accompa- nied by augmented pinocytosis in cerebral capillaries, Br. J. Pharmacol. 96, 838 (abstract).

Sarmento, A., N. Borges and I. Azevedo, 1991, Adrenergic influ- ences on the control of blood-brain barrier permeability, Naunyn-Schmied. Arch. Pharmacol. 343, 633.

Sarmento, A., N. Borges and D. Lima, 1994, Influence of electrical stimulation of locus coeruleus on the rat blood-brain barrier permeability to sodium fluorescein, Acta Neurochir. 127, 215.

Shi, F., C. Bailey, A.W. Malick and K.L. Audus, 1993, Biotin uptake and transport across bovine brain microvessel endothelial cell monolayers, Pharm. Res. 10, 282.