Health Articles

 

W.web.umkc.edu

Am J Physiol Cell Physiol 286: C1045–C1052, 2004.
First published December 18, 2003; 10.1152/ajpcell.00360.2003.
Protection against hypoxia-induced blood-brain barrier disruption: changes in intracellular calcium Rachel C. Brown, Karen S. Mark, Richard D. Egleton, and Thomas P. Davis
Department of Pharmacology, University of Arizona, Tucson, Arizona 85724
Submitted 26 August 2003; accepted in final form 15 December 2003 Brown, Rachel C., Karen S. Mark, Richard D. Egleton, and
occludin (37, 53). Stabilization of TJ involves a complex Thomas P. Davis. Protection against hypoxia-induced blood-brain
network of the transmembrane proteins occludin and claudins, barrier disruption: changes in intracellular calcium. Am J Physiol Cell linked to the actin cytoskeleton by accessory ZO proteins, Physiol 286: C1045–C1052, 2004. First published December 18, which mediate this linkage by binding the cytoplasmic tails of 2003; 10.1152/ajpcell.00360.2003.—Tissue damage after stroke is occludin and claudin to actin (32).
partly due to disruption of the blood-brain barrier (BBB). Little is Calcium is critical to normal BBB function (13). Endothelial known about the role of calcium in modulating BBB disruption. We cells incubated in calcium-free medium had decreased trans- investigated the effect of hypoxic and aglycemic stress on BBBfunction and intracellular calcium levels. Bovine brain microvessel endothelial electrical resistance readings and increased mono- endothelial cells were treated with A-23187 to increase intracellular layer paracellular permeability (58), indicating disruption of calcium without hypoxia or treated with a calcium chelator (BAPTA) TJ. In brain microvessel endothelial cells in culture, hypoxia or calcium channel blockers (nifedipine or SKF-96365) and 6 h of caused an increase in intracellular calcium (35, 41, 55), hypoxia. A-23187 alone did not increase paracellular permeability.
whereas blockade of calcium flux (1) or blockade of calcium- Hypoxia increased intracellular calcium, and hypoxia or hypoxia- regulated signaling cascades (55) prevented hypoxia-induced aglycemia increased paracellular permeability. Treatment with nifed- disruption of BBB monolayer integrity.
ipine and SKF-96365 increased intracellular calcium under normo- Stroke is the third leading cause of death in the United States glycemic conditions, instead of blocking calcium influx, and was and is the leading cause of long-term disability, affecting protective against hypoxia-induced BBB disruption under normogly- 500,000 patients every year (50). Although the neuronal cemia. Protection by nifedipine and SKF-96365 was not due to damage associated with stroke is due in part to lack of oxygen antioxidant properties of these compounds. These data indicate that and nutrients during an ischemic event, reperfusion, the resto- increased intracellular calcium alone is not enough to disrupt the ration of oxygen supply, and subsequent production of free BBB. However, increased intracellular calcium after drug treatmentand hypoxia suggests a potential mechanism for these drugs in BBB radicals are also major contributors in mediating neuronal protection; nifedipine and SKF-96365 plus hypoxic stress may trigger damage and death (50). Ischemic damage to blood vessels, in calcium-mediated signaling cascades, altering BBB integrity.
particular to the brain microvessel endothelial cells that makeup the BBB, results in deregulation of ion flux and increased nifedipine; SKF-96365; ischemia; permeability; fura 2 ion permeability into the brain, leading to a net influx of solutesand water and subsequent edema formation. Several studiesfound increased paracellular permeability after hypoxic or IONIC HOMEOSTASIS of the central nervous system (CNS) iscritical for normal brain function and is dependent on the ischemic insult (1, 5, 30), suggesting disruption of the TJ.
blood-brain barrier (BBB). This barrier, which isolates the Transient ischemia in rats increased extravasation of Evans brain from the peripheral circulation, is located at the level of blue albumin, a normally impermeable marker (5). This in-crease in paracellular permeability is correlated with the degree the cerebral microvessel endothelial cells. The BBB is charac- of tissue damage. A number of in vitro models have been used terized by tight cell-cell contacts with limited paracellular to investigate the effects of hypoxia and aglycemia on the diffusion and reduced fluid-phase endocytosis (56). The pres- BBB. Hypoxia and aglycemia increased the paracellular per- ence of specific endothelial cell transporters for ions, peptides, meability of bovine brain microvessel endothelial cell and nutrients allows for strict regulation of the CNS microen- (BBMEC) monolayers (1). Nifedipine (NIF), an L-type cal- vironment. Disruption of the BBB occurs in a number of cium channel antagonist, and SKF-96365 (SKF), an inhibitor pathological conditions, including Alzheimer's disease, diabe- of some store-operated calcium channels, blocked this increase tes, multiple sclerosis, inflammatory pain, and stroke (1, 4, 28, in permeability, indicating a potential role for calcium in 33, 39), and can contribute to CNS edema formation (45, 52).
mediating some of the hypoxia- and/or aglycemia-induced The restrictive nature of the BBB is due to tight junctions alterations in BBB integrity. In the present study, we investi- (TJ) formed between adjacent endothelial cells in the brain gated the effect of the calcium channel blockers NIF and SKF capillaries (42). The TJ restricts ion flux and paracellular on intracellular calcium levels, BBMEC monolayer permeabil- diffusion of macromolecules. A number of the proteins that ity, and reactive oxygen species (ROS) generation to elucidate compose TJ have been identified and extensively character- potential mechanisms by which calcium might be involved in ized, including the claudins (26, 54), occludin (25), and zonula mediating BBB integrity and function after a hypoxic and/or occludens (ZO)-1, -2, and -3, which interact with claudins and aglycemic insult.
Address for reprint requests and other correspondence: T. P. Davis, Dept.
of Pharmacology, PO Box 24-5050, The Univ. of Arizona College of The costs of publication of this article were defrayed in part by the payment Medicine, 1501 N. Campbell Ave., Tucson, AZ 85724-5050 (E-mail: of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
0363-6143/04 $5.00 Copyright 2004 the American Physiological Society INTRACELLULAR CA2⫹ IN BBB ENDOTHELIAL CELLS AFTER HYPOXIA MATERIALS AND METHODS
fluorescence was plotted for each treatment, area under the curve(AUC) was determined, and values were statistically analyzed.
In vitro BBB model. BBMEC were isolated from the gray matter of Permeability measurements. Permeability studies used [14C]su- the cerebral cortex of fresh bovine brains obtained from the University crose, a normally low-permeability marker, to determine paracellular of Arizona Meat Laboratory and cryopreserved as previously de- flux across BBMEC monolayers. Apical-to-basolateral flux was de- scribed (3, 49). Isolated cells were seeded onto collagen-fibronectin- termined by dividing picomoles of radioactive marker appearing in coated Transwell filters or slides and grown in MEM-Ham's F-12 the receiver chamber by the time in minutes (1, 49). The apparent with 10% equine serum, 50 ␮g/ml gentamicin, 2.5 ␮g/ml amphoter- permeability coefficient (PC) was calculated as follows icin B, and 100 ␮g/ml heparin. All BBMEC used for these studieswere primary cultured cells from passage 0, which have been shown PC ⫽ flux/(A*CDo) to maintain excellent BBB characteristics in vitro (2, 66). Cultures in where flux is the slope of the line, A is the area of the membrane, and Transwell filters were incubated with C6 astrocyte-conditioned me- dium in the basolateral chamber for 3 days before the start of Do is the initial donor concentration of radioactive marker.
Measurement of ROS. BBMEC were plated in 96-well plates and experiments (14).
incubated with normoglycemic medium, aglycemic medium, or agly- Calcium ionophore treatment and hypoxic stress. BBMEC mono- cemic medium ⫹ drug (100 nM NIF or 100 nM SKF). At 1 h before layers were treated with 5 ␮M A-23187, a calcium ionophore, for 6 h the end of the hypoxic stress, cells were incubated in 50 ␮M 2⬘,7⬘- to increase levels of intracellular calcium. Some monolayers were dichlorofluorescein diacetate (Calbiochem, San Diego, CA), a fluo- pretreated before the addition of A-23187 for 30 min with 10 ␮M rescent probe for measuring oxidative stress (12). Production of ROS BAPTA-AM (BAPTA), a cell-permeable calcium chelator, to inhibit was measured by obtaining fluorescence readings at 485 nm of the increase in intracellular calcium due to the calcium ionophore.
excitation and 535 nm of emission on the GENios microplate reader Monolayer permeability was assayed as described below.
after 6 h of normoxia, hypoxia, or hypoxia followed by 30 min of For hypoxic experiments, BBMEC monolayers were subjected to 6 h of hypoxia under different treatment conditions. Monolayers were Statistics. Values are means ⫾ SE. Results were analyzed with one- incubated in normoglycemic RPMI 1640 (Life Technologies, Rock- or two-way ANOVA as appropriate and then with multiple pairwise ville, MD) and treated with 10 ␮M BAPTA, 100 nM NIF, or 100 nM comparisons with Tukey's test using Sigma Stat 2.03 (SPSS, Chicago, SKF; drugs were added 30 min before the onset of hypoxic stress.
IL). Significance was defined as P ⬍ 0.05.
Other monolayers were subjected to hypoxic stress ⫹ aglycemia;these cells were incubated in modified RPMI 1640 without glucose (Life Technologies, Rockville, MD) and were treated with BAPTA,NIF, or SKF. Hypoxic stress was achieved by incubating the mono- We used BBMEC monolayers as a model system to inves- layers in a humidified, gas-controlled hypoxic workstation (Coy tigate the effects of increasing intracellular calcium levels or of Laboratory Products, Grass Lake, MI) at 37°C with 99% N2-1% O2.
hypoxic and/or aglycemic stress on BBB functional integrity.
At the end of the hypoxic period, the medium was removed and PO2 BBMEC were exposed to various treatment conditions before levels were measured with a blood gas analyzer (ABL 505, Radiom- determination of intracellular calcium levels and monolayer eter, Copenhagen, Denmark) to confirm hypoxia. PO2 in medium from normoxic samples was 150.2 ⫾ 4.5 mmHg; PO2 in hypoxic medium A-23187, a calcium ionophore, did not alter monolayer was 44.1 ⫾ 1.6 mmHg. Monolayers were assayed for changes in permeability. Incubation of BBMEC monolayers in 5 ␮M intracellular calcium levels or functional permeability (see below).
A-23187 or 10 ␮M BAPTA did not change [14C]sucrose Measurement of intracellular calcium. To measure intracellular calcium, BBMEC were plated into 96-well plates. When confluent, permeability compared with untreated control (Fig. 1). How- cells were incubated in 5 ␮M fura 2-AM alone or with BAPTA, NIF, ever, treatment with A-23187 ⫹ BAPTA significantly in- or SKF for 30 min at 37°C. After the incubation period, fura 2-AMwas removed, and medium containing appropriate drugs was added.
Readings were taken at various times over a 6-h period of incubationas described below.
For changes in intracellular calcium levels over a hypoxic stress period, an initial reading was obtained, and plates were transferred tothe hypoxic chamber for treatment. The microplate was covered withnonpermeable adhesive membrane during fluorescence readings tomaintain hypoxic conditions. At each time point, the plate wascovered with a nonpermeable adhesive membrane and removed fromthe hypoxic chamber for fluorescence readings. At the end of thereading, the plate was returned to the chamber and the adhesivemembrane was removed; there was no reoxygenation of the mediumor cells during the reading. Changes in intracellular calcium levelswere determined by measuring fura 2 fluorescence at 340 nm excita-tion and 510 nm emission on a GENios microplate reader (Tecan,Research Triangle Park, NC). Calcium-insensitive fura 2 fluorescencewas measured at 380 nm excitation and 510 nm emission, and resultswere expressed as percentage of time 0 control ratio of fluorescence at340 nm to fluorescence at 380 nm. Measurements were obtained as Fig. 1. A-23187 has no effect on bovine brain microvessel endothelial cell described above at various times over the 6-h hypoxic stress period. In (BBMEC) monolayer permeability. BBMEC monolayers were incubated with5 ␮M A-23187 and/or 10 ␮M BAPTA, and [14C]sucrose permeability was a related experiment, cells were incubated with NIF and SKF, and assayed. Neither A-23187 nor BAPTA treatment affected BBMEC permeabil- basal fura 2-AM fluorescence was measured before the addition of ity. When monolayers were incubated with A-23187 ⫹ BAPTA, permeability 100 nM bradykinin. Changes in intracellular calcium fluorescence was significantly increased compared with BAPTA treatment alone. PC, were measured as described above. After the time course of fura 2 permeability coefficient. Values are means ⫾ SE; n ⫽ 5–6. **P ⫽ 0.01.
AJP-Cell Physiol • VOL 286 • MAY 2004 • www.ajpcell.org INTRACELLULAR CA2⫹ IN BBB ENDOTHELIAL CELLS AFTER HYPOXIA creased BBMEC monolayer permeability compared withBAPTA alone (F3,22 ⫽ 5.029, P ⫽ 0.01).
Hypoxic stress increased intracellular calcium; treatment with NIF or SKF further increased levels of intracellularcalcium over the course of hypoxic stress. BBMEC monolayerswere exposed to 6 h of 1% oxygen in normoglycemic oraglycemic medium, and fura 2 fluorescence was measured asan indicator of changes in intracellular calcium. A character-istic pattern of changes in intracellular calcium in normogly-cemic or aglycemic medium was observed (Fig. 2). During the1st h of hypoxia, the stress appeared to induce long-termoscillations in intracellular calcium levels, with several appar-ent peaks. These initial oscillations appeared to subside, withintracellular calcium levels returning to near baseline (time 0)levels by 2 h. There was a subsequent increase in intracellularcalcium at 4 h of hypoxic stress that returned to normal at 5 hand increased again at 6 h of hypoxic stress, perhaps as part oflonger-term oscillatory behavior.
Using similar assay conditions, we examined the effects of BAPTA, NIF, and SKF, alone and in combination, on intra-cellular calcium levels. Data were expressed as percentage oftime 0 control samples, and the AUC was calculated andanalyzed by two-way ANOVA (Fig. 3). Drug treatment had asignificant effect on intracellular calcium levels (F6,604 ⫽2.298, P ⫽ 0.033); furthermore, there was a significant inter-action between drug treatment and glycemic condition (F6,604 ⫽ 4.269, P ⬍ 0.001). Under hypoxic-normoglycemicconditions, treatment with NIF and/or SKF increased levels ofintracellular calcium, contrary to the predicted effects of thesedrugs as calcium channel blockers (Fig. 3A). Treatment withBAPTA decreased intracellular calcium levels, while treatmentwith BAPTA ⫹ NIF or SKF abolished the increase in intra-cellular calcium due to NIF or SKF alone. In contrast, under aglycemic conditions (Fig. 3B), treatment with the drugs aloneafter 6 h of hypoxic stress had no effect on calcium levels, but Fig. 3. Effect of BAPTA, nifedipine (NIF), and SKF-96365 (SKF) on intra- treatment with drug combinations (BAPTA ⫹ NIF or cellular calcium during 6 h of hypoxic stress. BBMEC were subjected to 6 h BAPTA ⫹ SKF) led to unexpectedly increased intracellular of hypoxia and incubated with 10 ␮M BAPTA, 100 nM NIF, or 100 nM SKF, alone or in combination. There was a significant effect of drug treatment onintracellular calcium (F6,604 ⫽ 2.298, P ⫽ 0.033) and a significant interactionbetween drug treatment and glycemic condition (F6,604 ⫽ 4.269, P ⬍ 0.001).
A: under normoglycemic conditions, BAPTA decreased intracellular calcium.
NIF and/or SKF increased the amount of intracellular calcium during thehypoxic stress. This effect of NIF and/or SKF was prevented by cotreatmentwith BAPTA. B: under aglycemic conditions, there was no significant effect ofBAPTA, NIF, or SKF, alone or in combination, on intracellular calcium levels,although BAPTA ⫹ NIF and BAPTA ⫹ SKF did potentially increase intra-cellular calcium. Values are means ⫾ SE; n ⫽ 6–20. AUC, area under thecurve.
To determine whether NIF or SKF could actually block changes in intracellular calcium, fura 2 fluorescence was mea-sured after stimulation of BBMEC with bradykinin, a peptideknown to trigger an increase in intracellular calcium in endo-thelial cells (17). After bradykinin stimulation under normoxicconditions, only SKF decreased the amount of intracellular Fig. 2. Hypoxic stress causes a multiphasic increase in intracellular calcium in calcium over the period of exposure to the peptide (F3,30 BBMEC. BBMEC were subjected to 6 h of hypoxic stress under normogly- cemic or aglycemic conditions. Within the 1st h of hypoxic exposure, there 6.892, P ⫽ 0.001; Fig. 4), demonstrating that SKF-sensitive were 3 transient peaks in intracellular calcium. Calcium levels then returned to calcium influx pathways do exist in these cells.
baseline and, subsequently, increased at 4 and 6 h of hypoxic stress. There was BAPTA, NIF, and SKF partially protected against hypoxia- no statistical difference in intracellular calcium between normoglycemic and induced BBMEC monolayer disruption after hypoxic stress. aglycemic conditions. A representative trace is shown. Values are means ⫾ BBMEC monolayers were exposed to hypoxic stress, and SE; n ⫽ 8. 340/380 ratio, ratio of fluorescence at 340 nm to fluorescence at 380nm.
paracellular permeability was measured using [14C]sucrose.
AJP-Cell Physiol • VOL 286 • MAY 2004 • www.ajpcell.org INTRACELLULAR CA2⫹ IN BBB ENDOTHELIAL CELLS AFTER HYPOXIA of hypoxia. No antioxidant effect was observed with NIF orSKF under any of the three oxygen conditions.
Ischemic stroke and hypoxic stress disrupt the BBB (18).
We previously demonstrated that hypoxic (49) or hypoxic-aglycemic (1) stress can disrupt an in vitro BBB model system,providing an opportunity to investigate some of the cellularmechanisms underlying this disruption. Interestingly, two cal-cium channel blockers, NIF and SKF, can prevent BBB dis-ruption in this model system (1). We chose to examine the roleof intracellular calcium in modulating BBB functional perme- Fig. 4. SKF, but not NIF, prevents calcium influx after bradykinin stimulation.
BBMEC were plated as described for hypoxic stress experiments. Cells wereloaded with fura 2-AM and stimulated with 100 nM bradykinin. Intracellularcalcium was measured for 6 h, and AUC was determined. One-way ANOVAindicated a significant blockade of bradykinin-stimulated calcium influx after SKF treatment (F3,30 ⫽ 6.892, P ⫽ 0.001) but not after NIF treatment. Valuesare means ⫾ SE; n ⫽ 8–9. *P ⬍ 0.05; **P ⬍ 0.01 vs. control.
After 6 h of hypoxic stress, there was a significant effect ofdrug (BAPTA, NIF, or SKF) treatment (F11,390 ⫽ 4.232, P ⬍0.001) and of hypoxic stress (F1,390 ⫽ 75.449, P ⬍ 0.001) onBBB functional integrity. There was also a significant interac- tion between drug treatment and hypoxic stress (F11,390 ⫽1.851, P ⫽ 0.044). Under normoxic-normoglycemic condi-tions, treatment with BAPTA, NIF, or SKF significantly de-creased BBMEC monolayer permeability (Fig. 5A). Hypoxic-normoglycemic stress increased BBMEC monolayer perme-ability in almost all conditions. Statistical analyses indicate thatonly treatment with BAPTA ⫹ NIF protected against hypoxia- induced monolayer disruption.
Under normoxic-aglycemic conditions, there was no effect of any of the drug treatments on BBMEC monolayer perme-ability (Fig. 5B). Hypoxia caused an increase in monolayerpermeability in control monolayers. Treatment with NIF orSKF alone prevented the hypoxia-induced permeability in-crease, but BAPTA treatment alone had no effect on thishyperpermeability. Monolayers treated with BAPTA ⫹ NIFwere protected against hypoxia-induced increases in perme-ability, but BAPTA ⫹ SKF was not protective.
Antioxidant effects of NIF and SKF. Since NIF and SKF did not appear to act as calcium channel blockers in our hypoxiaexperimental paradigm (Fig. 3), we investigated whether theprotective actions of these drugs were due to antioxidantproperties. Cells were incubated with normoglycemic medium, Fig. 5. BAPTA, NIF, and SKF partially protect BBMEC monolayer integrity.
aglycemic medium, or aglycemic medium with NIF or SKF as BBMEC monolayers were treated with 10 ␮M BAPTA, 100 nM NIF, and 100nM SKF, alone or in combination. After 6 h of hypoxic stress, monolayer described for the calcium assays and permeability studies. ROS permeability was assayed. Two-way ANOVA indicated a significant effect of were measured using 2⬘,7⬘-dichlorofluorescein diacetate, a flu- drug treatment (F11,390 ⫽ 4.232, P ⬍ 0.001) and a significant effect of hypoxic orescent probe that detects the presence of oxygen free radicals stress (F1,390 ⫽ 75.449, P ⬍ 0.001), as well as a significant interaction between (12). BBMEC were subjected to 6 h of normoxia, 6 h of drug treatment and hypoxic stress (F11,390 ⫽ 1.851, P ⫽ 0.044). A: pairwise hypoxia, or 6 h of hypoxia followed by 30 min of reoxygen- comparisons indicate that, under normoxic/normoglycemic conditions,BAPTA and NIF treatment significantly reduced monolayer permeability ation in room air (Fig. 6). There was a significant effect of compared with normoxic control. Treatment with BAPTA, NIF, SKF, or glucose and drug treatment (F3,60 ⫽ 19.213, P ⬍ 0.001) and of BAPTA ⫹ SKF did not block the hypoxia-induced increase in monolayer oxygen treatment (F2,60 ⫽ 31.196, P ⬍ 0.001) on ROS levels.
permeability. Only BAPTA ⫹ NIF prevented monolayer disruption. B: in Incubation in aglycemic medium alone increased ROS levels BBMEC subjected to aglycemic conditions, NIF and SKF prevented hypoxia-induced increases in permeability, as did BAPTA ⫹ NIF. Neither BAPTA under any oxygen condition. ROS levels after 30 min of alone nor BAPTA ⫹ SKF prevented this hypoxic effect. Values are means ⫾ reoxygenation were significantly higher in all drug treatment SE; n ⫽ 6–30. ***P ⬍ 0.001 vs. normoxia. ⫹⫹⫹P ⬍ 0.001; ⫹⫹P ⬍ 0.01; groups than in the groups treated with 6 h of normoxia or 6 h ⫹P ⬍ 0.05; *P ⬍ 0.05 vs. normoxic control.
AJP-Cell Physiol • VOL 286 • MAY 2004 • www.ajpcell.org INTRACELLULAR CA2⫹ IN BBB ENDOTHELIAL CELLS AFTER HYPOXIA cantly altered, regardless of drug treatment. However, thesecells do have some SKF-sensitive calcium channels, as indi-cated by the bradykinin experiment (Fig. 4), although thesechannels may not be involved in mediating hypoxic effects.
There is debate about the presence of L-type calcium chan- nels on endothelial cells, a nonexcitable cell type. A fewstudies have demonstrated L-type channel currents in endothe-lial cells (64), and there is pharmacological evidence that thesedrugs prevent increases in intracellular calcium in endothelialcells (67) and protect against hypoxia-induced damage (1, 29);however, there is no convincing molecular evidence for thepresence of these channels at the BBB. Therefore, we mustconsider the possibility that the protective effect of NIF seen inBBB models is mediated through an alternate mechanism. The Fig. 6. NIF and SKF are not antioxidants in this system. BBMEC were plated present study and others have demonstrated that NIF has and subjected to 6 h of hypoxic stress under normoglycemia, aglycemia, or protective effects under hypoxic conditions. NIF can inhibit aglycemia with NIF or SKF treatment. Production of reactive oxygen species hypoxic damage by inhibiting protein kinase (PKC)-␣ (ROS) was measured using 2⬘,7⬘-dichlorofluorescein diacetate (2,7-DCF).
(PKC-␣) (29), but this does not correlate with the increase in There was a significant effect of glucose and drug treatment (F3,60 ⫽ 19.213, intracellular calcium (potentially activating PKC- P ⬍ 0.001) and a significant effect of oxygen (F 2,60 ⫽ 31.196, P ⬍ 0.001).
Aglycemic treatment alone increased ROS in BBMEC. A 30-min reoxygen- treatment with NIF in this study and in others (8, 63).
ation period after 6 h of hypoxic stress also caused a significant increase in SKF is considered to be a stores-operated calcium channel ROS. Treatment with NIF or SKF did not prevent this reoxygenation-induced blocker, but its parent compound, SC-38249, can also inhibit increase in ROS. Values are means ⫾ SE; n ⫽ 10–12. ***P ⬍ 0.001 vs.
extrusion of intracellular calcium via the plasma membrane normoglycemia (control). ⫹⫹⫹P ⬍ 0.001 vs. normoxia.
Ca2⫹-ATPase (PMCA) (16). SKF does block an increase inendothelial cell calcium after stimulation with bradykinin (22), ability and its potential involvement in mediating hypoxia- as we have demonstrated in this study. However, other studies induced damage.
have demonstrated that treatment with SKF alone can trigger Although there is a large body of literature dealing with the an increase in intracellular calcium that is mediated by release role of calcium in modulating cell-cell adhesion through adhe- from intracellular, thapsigargin-sensitive stores (36) and by rens junctions, much less is known about the role of calcium in influx from the extracellular medium (38), suggesting that SKF modulating TJ function (13). We used a calcium ionophore to may have dose- and cell system-specific effects on intracellular increase levels of intracellular calcium in our BBB model system and found that simply increasing intracellular calcium What is the potential source of the signal triggering this alone did not lead to increased paracellular permeability, indi- increase in intracellular calcium under hypoxic-normoglyce- cating that high intracellular calcium levels alone, such as mic conditions? Why is such an increase not observed under those seen after a stroke (43, 60) or hypoxic event (35, 41, 55), hypoxic-aglycemic conditions? One potential trigger for intra- are not sufficient to alter TJ function. However, there is cellular calcium increases under these hypoxic-normoglycemic evidence that TJ components, such as ZO-1 (61, 69) and conditions is nitric oxide (NO) produced by calcium-dependent occludin (55), are sensitive to calcium levels. It remains to be stimulation of endothelial NO synthase (eNOS) and released seen how increased intracellular calcium alters TJ protein from surrounding endothelial cells. NO is produced by cerebral expression or function in this experimental paradigm.
microvessel endothelial cells under hypoxic stress (48), which We found that a 6-h hypoxic or hypoxic-aglycemic stress increases BBB endothelial cell permeability (10, 48); this triggered a characteristic increase in intracellular calcium in effect is attenuated by inhibition of NOS (48). NO can also BBB endothelial cells. Within the 1st h of exposure to the trigger an increase in intracellular calcium in endothelial cells stressor, intracellular calcium levels displayed distinct oscilla- (6, 9). Calcium is a major regulator of eNOS, a major source tions with several peaks, perhaps representing the initial re- for NO production in these cells (23). Under our experimental sponse of BBB endothelial cells to stress. This initial response conditions, hypoxic stress could trigger an initial increase in was followed by increases in intracellular calcium at 4 and 6 h, intracellular calcium that could activate eNOS to produce NO.
which may reflect long-term slow calcium oscillatory behavior.
This NO could act on the endothelial cell in a paracrine or There was no difference in the calcium response between cells autocrine fashion, thereby further increasing the level of intra- incubated in aglycemic medium and those incubated in nor- cellular calcium. The NO-mediated calcium increase would not moglycemic medium during hypoxic stress. However, the be affected by NIF or SKF but would be subject to chelation by presence or absence of glucose did influence cellular responses BAPTA, as seen in these studies (Fig. 3A). Furthermore, NIF to the calcium chelator (BAPTA) or calcium channel blockers treatment can upregulate eNOS protein levels (19) and stimu- (NIF or SKF). Under normoglycemic conditions, treatment late endothelial NO production (19, 20, 24), further implicating with NIF and/or SKF increased intracellular calcium levels NO in this experimental paradigm. SKF treatment has not been over the course of the hypoxic stress. This result is in direct convincingly shown to stimulate NO production on its own opposition to the presumed mechanism of action of these (36), but it may inhibit the activity of the PMCA (16), thereby drugs: NIF is an L-type calcium channel blocker, while SKF prolonging any increase in intracellular calcium. Under agly- blocks some stores-operated calcium channels. Under aglyce- cemic conditions, energy-dependent mechanisms may be in- mic conditions, intracellular calcium levels were not signifi- hibited, leading to prevention of eNOS function, and, thus, may AJP-Cell Physiol • VOL 286 • MAY 2004 • www.ajpcell.org INTRACELLULAR CA2⫹ IN BBB ENDOTHELIAL CELLS AFTER HYPOXIA prevent the potentially NO-induced increase in intracellular block their respective targets, which still occurs after bradyki- nin treatment, but, rather, an unmasking of calcium released There are other mechanisms by which endothelial cells can from a secondary source, which may be crucial for mediating increase intracellular calcium levels via pathways distinct from NIF and SKF protection of monolayer integrity. What remains store-operated or voltage-gated channels. Hypoxia triggers to be determined in this experimental paradigm is the role of expression of endothelin-1 (ET-1) (31) and vascular endothe- calcium in modulating TJ protein expression or protein-protein lial growth factor (34, 44). ET-1 can increase intracellular calcium (59) via an inositol trisphosphate-sensitive pathway We envision several mechanisms by which calcium could be (70); this increase can be potentiated by NOS inhibitors or involved in regulating BBB TJ function: 1) a direct interaction decreased by NO donors (15), indicating cross talk between between calcium and TJ proteins (i.e., occludin) in a manner these two pathways. NO can induce expression of vascular similar to that occurring with E-cadherin at the adherens endothelial growth factor, which in turn activates eNOS (11, junction (54), 2) a mechanism by which alterations in intracel- 51), feeding into the NO pathway described above. The BBB lular calcium levels, and consequent activation of signal trans- endothelial cell also expresses a large number of ion transport- duction cascades (i.e., calcium-calmodulin kinases or PKC), ers, the function of which may be altered under pathological directly regulate expression of TJ protein genes, and/or 3) an conditions. One of these, the Na⫹/Ca2⫹ exchanger (NCX), can indirect mechanism by which calcium affects other proteins, work in reverse (i.e., calcium influx coupled to sodium efflux) such as eNOS, ET-1, ion transporters, or ATPases, thereby after pharmacological manipulation (21). Reversal of the NCX affecting BBB integrity. Future studies are targeted at deter- has been shown to be important in mediating spinal cord white mining the effect of changes in intracellular calcium on TJ matter injury after anoxia (46), and NCX inhibitors can protect protein expression, elucidating the source of the increased against ischemic injury attributed to exchanger reversal (68).
intracellular calcium after NIF or SKF treatment, and investi- Furthermore, reversal of the NCX in endothelial cells has been gating the signal transduction cascades triggered by the treat- linked to activation of eNOS and production of NO (57, 62). If ments described in these studies to clarify the role of calcium the NCX in BBB endothelial cells is reversed under hypoxic in TJ protein regulation and BBB function.
conditions, leading to calcium overload inside the cells, thismight also explain why NIF and SKF do not block the increase in intracellular calcium after hypoxic stress. The data presented We thank Drs. Roger O'Neil and Anne Wolka for reviewing the manu- in this study may also indicate a secondary effect of drug treatment on the activity or expression of ion transporters Present address of R. C. Brown: Dept. of Integrative Biology and Pharma- cology, University of Texas Health Science Center at Houston, 6431 Fannin critical for the removal of intracellular calcium, such as the St., Houston, TX 77030.
NCX or PMCA (65), the expressions of which can be directlymodulated by changes in intracellular calcium (27, 40).
As previously demonstrated, NIF protects against hypoxia- This work was supported by National Institute of Neurological Disorders induced BBB functional disruption, and we confirmed the and Stroke Grant NS-39592. R. C. Brown was supported by National Institute protective effect of SKF (1). Interestingly, treatment with of Neurological Disorders and Stroke Grant F32 NS-43052 and K. S. Mark by BAPTA, NIF, or SKF under normoxia-normoglycemia actu- Grant F32 NS-11175.
ally tightened the monolayer, perhaps implicating resting in-tracellular calcium levels in TJ function. Under normoxic- aglycemic conditions, there was no effect of drug treatment on 1. Abbruscato TJ and Davis TP. Combination of hypoxia/aglycemia com-
monolayer permeability; however, NIF and SKF were able to promises in vitro blood-brain barrier integrity. J Pharmacol Exp Ther 289:668–675, 1999.
completely prevent the hypoxia-induced increase in monolayer 2. Abbruscato TJ, Thomas SA, Hruby VJ, and Davis TP. Blood-brain
permeability under aglycemic conditions, while BAPTA had barrier permeability and bioavailability of a highly potent and ␮-selective no protective effect. As with the A-23187 data, these results opioid receptor antagonist, CTAP: comparison with morphine. J Pharma- indicate that a dramatic alteration (an increase or a decrease) in col Exp Ther 280: 402–409, 1997.
intracellular calcium is not sufficient to disrupt or protect BBB 3. Audus KL and Borchardt RT. Characterization of an in vitro blood-
brain barrier model system for studying drug transport and metabolism.
monolayer functional integrity. Although other studies have Pharm Res 3: 81–87, 1986.
proposed an antioxidant role for NIF (7, 47), we have ruled out 4. Banks WA, Jaspan JB, and Kastin AJ. Effect of diabetes mellitus on the
the possibility that NIF or SKF is protective via an antioxidant permeability of the blood-brain barrier to insulin. Peptides 18: 1577–1584, mechanism in this system. As we have mentioned here, intra- 5. Belayev L, Busto R, Zhao W, and Ginsberg MD. Quantitative evalua-
cellular calcium levels are regulated by many cellular mecha- tion of blood-brain barrier permeability following middle cerebral artery nisms, which makes it difficult to determine the exact mecha- occlusion in rats. Brain Res 739: 88–96, 1996.
nism by which calcium influences TJ and BBB monolayer 6. Berkels R, Dachs C, Roesen R, and Klaus W. Simultaneous measure-
ment of intracellular Ca2⫹ and nitric oxide: a new method. Cell Calcium In conclusion, we examined the effects of BAPTA, NIF, and 27: 281–286, 2000.
7. Berkels R, Egink G, Marsen TA, Bartels H, Roesen R, and Klaus W.
SKF treatment, alone or in combination, on BBB monolayer Nifedipine increases endothelial nitric oxide bioavailability by antioxida- permeability, intracellular calcium levels, and ROS production tive mechanisms. Hypertension 37: 240–245, 2001.
after a hypoxic or hypoxic-aglycemic stress. We demonstrated 8. Berkels R, Mueller A, Roesen R, and Klaus W. Nifedipine and Bay K
an increase in intracellular calcium after NIF or SKF treat- 8644 induce an increase of [Ca2⫹]i and nitric oxide in endothelial cells.
J Cardiovasc Pharmacol 4: 175–181, 1999.
ments under hypoxia-normoglycemia, in stark contrast to the 9. Berkels R, Suerhoff S, Roesen R, and Klaus W. Nitric oxide causes a
presumed actions of these drugs to simply block calcium flux.
cGMP-independent intracellular calcium rise in porcine endothelial The observed increase may reflect not a failure of the drugs to cells—a paradox? Microvasc Res 59: 38–44, 2000.
AJP-Cell Physiol • VOL 286 • MAY 2004 • www.ajpcell.org INTRACELLULAR CA2⫹ IN BBB ENDOTHELIAL CELLS AFTER HYPOXIA 10. Boje KM and Lakhman SS. Nitric oxide redox species exert differential
32. Huber JD, Egleton RD, and Davis TP. Molecular physiology and
permeability effects on the blood-brain barrier. J Pharmacol Exp Ther pathophysiology of tight junctions in the blood-brain barrier. Trends 293: 545–550, 2000.
Neurosci 24: 719–725, 2001.
11. Brouet A, Sonveaux P, Dessy C, Balligand JL, and Feron O. Hsp90
33. Huber JD, Witt KA, Hom S, Egleton RD, Mark KS, and Davis TP.
ensures the transition from the early Ca2⫹-dependent to the late phosphor- Inflammatory pain alters blood-brain barrier permeability and tight junc- ylation-dependent activation of the endothelial nitric-oxide synthase in tional protein expression. Am J Physiol Heart Circ Physiol 280: H1241– vascular endothelial growth factor-exposed endothelial cells. J Biol Chem H1248, 2001.
276: 32663–32669, 2001.
34. Ikeda E, Achen MG, Breier G, and Risau W. Hypoxia-induced tran-
12. Brown RC, Cascio C, and Papadopoulos V. Pathways of neurosteroid
scriptional activation and increased mRNA stability of vascular endothe- biosynthesis in cell lines from human brain: regulation of dehydroepi- lial growth factor in C6 glioma cells. J Biol Chem 270: 19761–19766, androsterone formation by oxidative stress and ␤-amyloid peptide. J Neu- rochem 74: 847–859, 2000.
35. Ikeda K, Nagashima T, Wu S, Yamaguchi M, and Tamaki N. The role
13. Brown RC and Davis TP. Calcium modulation of adherens and tight
of calcium ion in anoxia/reoxygenation damage of cultured brain capillary junction function: a potential mechanism for blood-brain barrier disruption endothelial cells. Acta Neurochir Suppl (Wien) 70: 4–7, 1997.
after stroke. Stroke 33: 1706–1711, 2002.
36. Iouzalen L, Lantoine F, Pernollet M, Millanvoye-Van Brussel E,
14. Brown RC, Mark KS, Egleton RD, Huber JD, Burroughs A, and
Devynck M, and Davis-Dufilho M. SK&F 96365 inhibits intracellular
Davis TP. Protection against hypoxia-induced increase in blood-brain
Ca2⫹ pumps and raises cytosolic Ca2⫹ concentration without production barrier permeability: role of tight junction proteins and NF␬B. J Cell Sci of nitric oxide and von Willebrand factor. Cell Calcium 20: 501–508, 116: 693–700, 2003.
15. Chen Y, McCarron RM, Azzam N, Bembry J, Reutzler C, Lenz FA,
37. Itoh M, Furuse M, Morita K, Kubota K, Saitou M, and Tsukita S.
and Spatz M. Endothelin-1 and nitric oxide affect human cerebromicro-
Direct binding of three tight junction-associated MAGUKs, ZO-1, ZO-2, vascular endothelial responses and signal transduction. Acta Neurochir and ZO-3, with the COOH termini of claudins. J Cell Biol 147: 1351– Suppl (Wien) 76: 131–135, 2000.
16. Ciardo A and Meldolesi J. Multiple actions of SC 38249: the blocker of
38. Jan CR, Ho CM, Wu SN, and Tseng CJ. Multiple effects of 1-[␤-[3-
both voltage-operated and second messenger-operated Ca2⫹ channels also inhibits Ca2⫹ extrusion. Eur J Pharmacol 188: 417–421, 1990.
chloride (SKF 96365) on Ca2⫹ signaling in MDCK cells: depletion of 17. Crawford KM, MacCallum DK, and Ernst SA. Agonist-induced Ca2⫹
thapsigargin-sensitive Ca2⫹ store followed by capacitative Ca2⫹ entry, mobilization in cultured bovine and human corneal endothelial cells. Curr activation of a direct Ca2⫹ entry, and inhibition of thapsigargin-induced Eye Res 12: 303–311, 1993.
capacitative Ca2⫹ entry. Naunyn Schmiedebergs Arch Pharmacol 359: 18. Del Zoppo GJ and Hallenbeck JM. Advances in the vascular pathophys-
92–101, 1999.
iology of ischemic stroke. Thromb Res 98: 73–81, 2000.
39. Kalaria RN. The blood-brain barrier and cerebrovascular pathology in
19. Ding Y and Vaziri N. Nifedipine and diltiazem but not verapamil
Alzheimer's disease. Ann NY Acad Sci 893: 113–125, 1999.
up-regulate endothelial nitric-oxide synthase expression. J Pharmacol Exp 40. Kaye DM and Kelly RA. Expression and regulation of the sodium-
Ther 292: 606–609, 2000.
calcium exchanger in cardiac microvascular endothelial cells. Clin Exp 20. Djuric D, Mitrovic V, and Jakovljevic V. The effects of calcium channel
Pharmacol Physiol 26: 651–655, 1999.
antagonists on coronary nitrite outflow in isolated rat heart. Arzneimittel- 41. Kimura C, Oike M, and Ito Y. Hypoxia-induced alterations in Ca2⫹
forschung 52: 365–370, 2002.
mobilization in brain microvascular endothelial cells. Am J Physiol Heart 21. Domotor E, Abbott NJ, and Adam-Vizi V. Na⫹-Ca2⫹ exchange and its
Circ Physiol 279: H2310–H2318, 2000.
implications for calcium homeostasis in primary cultured rat brain micro- 42. Kneisel U and Wolburg H. Tight junctions of the blood-brain barrier.
vascular endothelial cells. J Physiol 515: 147–155, 1999.
Cell Mol Neurobiol 20: 57–76, 2000.
22. Easton A and Abbott N. Bradykinin increases permeability by calcium
43. Kristian T and Siesjo BK. Calcium in ischemic cell death. Stroke 29:
and 5-lipoxygenase in the ECV304/C6 cell culture model of the blood- 705–718, 1998.
brain barrier. Brain Res 953: 157–169, 2002.
44. Kuo NT, Benhayon D, Przybylski RJ, Martin RJ, and LaManna JC.
23. Fleming I and Busse R. Molecular mechanisms involved in the regulation
Prolonged hypoxia increases vascular endothelial growth factor mRNA of the endothelial nitric oxide synthase. Am J Physiol Regul Integr Comp and protein in adult mouse brain. J Appl Physiol 86: 260–264, 1999.
Physiol 284: R1–R12, 2003.
45. Kuroiwa T, Shibutani M, and Okeda R. Blood-brain barrier disruption
24. Fukuo K, Yang J, Suzuki T, Kaimoto T, Takemura Y, Yasuda O,
and exacerbation of ischemic brain edema after restoration of blood flow Suhara T, Sato N, and Ogihara T. Nifedipine upregulates manganese
in experimental focal cerebral ischemia. Acta Neuropathol (Berl) 76: superoxide dismutase expression in vascular smooth muscle cells via 62–70, 1988.
endothelial cell-dependent pathways. Hypertens Res 26: 503–508, 2003.
46. Li S, Jiang Q, and Stys PK. Important role of reverse Na⫹-Ca2⫹
25. Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, and Tsukita
exchange in spinal cord white matter injury at physiological temperature.
S. Occludin: a novel integral membrane protein localizing at tight junc-
J Neurophysiol 84: 1116–1119, 2000.
tions. J Cell Biol 123: 1777–1788, 1993.
47. Mak IT, Boehme P, and Weglicki WB. Antioxidant effects of calcium
26. Furuse M, Sasaki H, Fujimoto K, and Tsukita S. A single gene product,
channel blockers against free radical injury in endothelial cells. Correla- claudin-1 or -2, reconstitutes tight junction strands and recruits occludin in tion of protection with preservation of glutathione levels. Circ Res 70: fibroblasts. J Cell Biol 143: 391–401, 1998.
1099–1103, 1992.
27. Guerini D, Garcia-Martin E, Gerber A, Volbracht C, Leist M, Merino
48. Mark KS, Burroughs A, Brown RC, Huber JD, and Davis TP. Nitric
CG, and Carafoli E. The expression of plasma membrane Ca2⫹ pump
oxide mediates hypoxia-induced changes in paracellular permeability of isoforms in cerebellar granule neurons is modulated by Ca2⫹. J Biol Chem cerebral microvasculature. Am J Physiol Heart Circ Physiol 286: H174– 274: 1667–1676, 1999.
28. Hawkins CP, Mackenzie F, Tofts P, du Boulay EP, and McDonald WI.
49. Mark KS and Davis TP. Cerebral microvascular changes in permeability
Patterns of blood-brain barrier breakdown in inflammatory demyelination.
and tight junctions induced by hypoxia-reoxygenation. Am J Physiol Heart Brain Res 114: 801–810, 1991.
Circ Physiol 282: H1485–H1494, 2002.
29. Hempel A, Lindschau C, Maasch C, Mahn M, Bychkov R, Noll T, Luft
50. Mark KS and Davis TP. Stroke: development, prevention and treatment
FC, and Haller H. Calcium antagonists ameliorate ischemia-induced
with peptidase inhibitors. Peptides 21: 1965–1973, 2000.
endothelial cell permeability by inhibiting protein kinase C. Circulation 51. Mayhan WG. VEGF increases permeability of the blood-brain barrier via
99: 2523–2529, 1999.
a nitric oxide synthase/cGMP-dependent pathway. Am J Physiol Cell 30. Hom S, Egleton RD, Huber JD, and Davis TP. Effect of reduced flow
Physiol 276: C1148–C1153, 1999.
on blood-brain barrier transport systems. Brain Res 890: 38–48, 2001.
52. Menzies SA, Betz AL, and Hoff JT. Contributions of ions and albumin
31. Hu J, Discher DJ, Bishopric NH, and Webster KA. Hypoxia regulates
to the formation and resolution of ischemic brain edema. J Neurosurg 78: expression of the endothelin-1 gene through a proximal hypoxia-inducible 257–266, 1993.
factor-1 binding site on the antisense strand. Biochem Biophys Res Com- 53. Mitic LL, Schneeberger EE, Fanning AS, and Anderson JM. Con-
mun 245: 894–899, 1998.
nexin-occludin chimeras containing the ZO-binding domain of occludin AJP-Cell Physiol • VOL 286 • MAY 2004 • www.ajpcell.org INTRACELLULAR CA2⫹ IN BBB ENDOTHELIAL CELLS AFTER HYPOXIA localize at MDCK tight junctions and NRK cell contacts. J Cell Biol 146: 63. Valencia L, Bidet M, Martial S, Sanchez E, Melendez E, Tauc M,
683–693, 1999.
Poujeol C, Martin D, Namorado M, Reyes J, and Poujeol P. Nifedi-
54. Morita K, Sasaki H, Furuse M, and Tsukita S. Endothelial claudin:
pine-activated Ca2⫹ permeability in newborn rat cortical collecting duct claudin-5/TMVCF constitutes tight junction strands in endothelial cells.
cells in culture. Am J Physiol Cell Physiol 280: C1193–C1203, 2001.
J Cell Biol 147: 185–194, 1999.
64. Vinet R and Vargas FF. L- and T-type voltage-gated Ca2⫹ currents in
55. Park JH, Okayama N, Gute D, Krsmanovic A, Battarbee H, and
adrenal medulla endothelial cells. Am J Physiol Heart Circ Physiol 276: Alexander JS. Hypoxia/aglycemia increases endothelial permeability:
role of second messengers and cytoskeleton. Am J Physiol Cell Physiol 65. Wang X, Reznick S, Li P, Liang W, and van Breemen C. Ca2⫹ removal
277: C1066–C1074, 1999.
mechanisms in freshly isolated rabbit aortic endothelial cells. Cell Cal- 56. Reese TS and Karnovsky MJ. Fine structural localization of a blood-
cium 31: 265–277, 2002.
brain barrier to exogenous peroxidase. J Cell Biol 34: 207–217, 1967.
66. Weber SJ, Abbruscato TJ, Brownson EA, Lipkowski AW, Polt R,
Misicka A, Haaseth RC, Bartosz H, Hruby VJ, and Davis TP. Assess-
57. Schneider J, El Kabir D, Chereau C, Mercier J, Dall'ava-Santucci J,
and Dinh-Xuan A. Involvement of Na⫹/Ca2⫹ exchanger in endothelial
[Met5]enkephalin opioid analogs. J Pharmacol Exp Ther 266: 1649–1655, NO production and endothelium-dependent relaxation. Am J Physiol Heart Circ Physiol 283: H837–H844, 2002.
67. Yakubu MA and Leffler CW. L-type voltage-dependent Ca2⫹ channels
58. Shasby DM and Shasby SS. Effects of calcium on transendothelial
in cerebral microvascular endothelial cells and ET-1 biosynthesis. Am J albumin transfer and electrical resistance. J Appl Physiol 60: 71–79, 1986.
Physiol Cell Physiol 283: C1687–C1695, 2002.
59. Stanimirovic DB, Nikodijevic B, Nikodijevic-Kedeva D, McCarron
68. Yamashita J, Itoh M, Kuro T, Kobayashi Y, Ogata M, Takaoka M,
RM, and Spatz M. Signal transduction and Ca2⫹ uptake activated by
and Matsumura Y. Pre- or post-ischemic treatment with a novel Na⫹/
endothelins in rat brain endothelial cells. Eur J Pharmacol 288: 1–8, 1994.
Ca2⫹ exchange inhibitor, KB-R7943, shows renal protective effects in rats 60. Stevens H, Krop-van Gastel W, and Korf J. 60Co and 45Ca autoradiog-
with ischemic acute renal failure. J Pharmacol Exp Ther 296: 412–419, raphy in cerebral ischemia in the rat. Neurosci Lett 258: 183–186, 1998.
61. Stuart RO, Sun A, Panichas M, Hebert SC, Brenner BM, and Nigam
69. Ye J, Tsukamoto T, Sun A, and Nigam SK. A role for intracellular
SK. Critical role for intracellular calcium in tight junction biogenesis.
calcium in tight junction reassembly after ATP depletion-repletion. Am J J Cell Physiol 159: 423–433, 1994.
Physiol Renal Physiol 277: F524–F532, 1999.
62. Teubl M, Groschner K, Kohlwein SD, Mayer B, and Schmidt K.
70. Yokokawa K, Kohno M, Murakawa K, Yasunari K, and Takeda T.
Na⫹/Ca2⫹ exchange facilitates Ca2⫹-dependent activation of endothelial Effect of endothelin-1 on cytosolic calcium ions in cultured human nitric-oxide synthase. J Biol Chem 274: 29529–29535, 1999.
endothelial cells. J Hypertens 8: 843–849, 1990.
AJP-Cell Physiol • VOL 286 • MAY 2004 • www.ajpcell.org

Source: http://w.web.umkc.edu/wudtked/mark_pubs/pub03.pdf