GSH transport in human cerebrovascular endothelial cells and human astrocytes: evidence for luminal localization of Na+-dependent GSH transport in HCEC1

GSH transport in human cerebrovascular endothelial cells and human astrocytes: evidence for luminal localization of Na+-dependent GSH transport in HCEC1

Brain Research 852 Ž2000. 374–382 www.elsevier.comrlocaterbres Research report GSH transport in human cerebrovascular endothelial cells and human as...

195KB Sizes 0 Downloads 47 Views

Brain Research 852 Ž2000. 374–382 www.elsevier.comrlocaterbres

Research report

GSH transport in human cerebrovascular endothelial cells and human astrocytes: evidence for luminal localization of Naq-dependent GSH transport in HCEC 1 R. Kannan a

a, )

, R. Chakrabarti a , D. Tang a , K.J. Kim b , N. Kaplowitz

a

Department of Medicine, DiÕisions of GI and LiÕer Diseases and Pulmonary and Critical Care Medicine, UniÕersity of Southern California, 2011 Zonal AÕenue, HMR 803A, Los Angeles, CA 90033, USA b Will Rogers Institute Pulmonary Research Center, UniÕersity of Southern California, Los Angeles, CA 90033, USA Accepted 28 September 1999

Abstract The purpose of the present study was to identify and localize glutathione ŽGSH. transport in an in vitro tissue culture model of blood–brain barrier ŽBBB.. The localization of Naq-dependent GSH transport in an immortalized cell line of human cerebrovascular endothelial cells ŽHCEC. and asymmetry of transport in Transwelle studies were investigated. Initial studies with cultured HCEC established a significant Ž45%. Naq-dependency for GSH uptake in cultured HCEC pretreated with acivicin, an inhibitor of g-glutamyltranspeptidase ŽGGT.. Transendothelial electrical resistance ŽTEER. and uptake of w35SxGSH from luminal and abluminal fluids of HCEC were measured in Naq-containing and Naq-free Žcholine chloride. buffers using cells grown on gelatin-coated membrane filters. TEER of HCEC monolayers in regular medium was 40.1 " 8.0 V cm2. Human astrocyte-conditioned medium ŽACM. caused no change in TEER, but increased GGT activity approximately threefold when measured in cell lysates. Luminal and abluminal GSH uptake increased in a time-dependent fashion and were not affected by inhibition of GGT activity with acivicin. Sodium dependency was only observed for luminal uptake ŽNaq-containing 2.41 " 0.15 vs. Naq-free 0.96 " 0.03 pmolr30 minrmillion cells, p - 0.001. but not for abluminal uptake Ž1.02 " 0.13 vs. 1.11 " 09, p ) 0.05.. Apparent efflux via the luminal membrane was lower in the presence of sodium as compared to that without sodium, further suggesting that a Naq-dependent uptake process for GSH is operative at this membrane. GSH uptake and efflux were also demonstrated in neonatal rat and fetal human astrocytes, both exhibiting partial Naq-dependency of uptake. In conclusion, our results show for the first time, that HCEC and astrocytes take up GSH by both Naq-dependent and -independent mechanisms. The Naq-dependent GSH transport process in HCEC appears to be localized to luminal plasma membranes of HCEC. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Glutathione; Sodium-dependence; Membrane transport; Domain specificity

1. Introduction The blood–brain barrier ŽBBB. restricts the movement of certain solutes between the blood and brain by acting as AbbreÕiations: GSH, glutathione; BBB, blood–brain barrier; HCEC, human cerebrovascular endothelial cells; FHAS, fetal human astrocytes; HPLC, high performance liquid chromatography; FBS, fetal bovine serum; PBS, phosphate buffered saline; GGT, g-glutamyltranspeptidase; GCS, g-glutamylcysteine synthetase; BSO, buthionine sulfoximine; DEM, diethylmaleate; TEER, transendothelial electrical resistance; SAAM, simulated analysis and modeling; oatp, organic anion transport protein; MRP, multidrug resistance-associated protein ) Corresponding author. Fax: q 1-323-442-3420; e-mail: [email protected] 1 Portions of this work were presented at the annual meeting of the Federation of the American Societies of Experimental Biology, April 17–21, 1999, Washington, DC wFASEB J. 13Ž4., 571.4A Ž1999.x.

a shield to protect the brain. It is well known that this selective barrier function is determined solely by the brain endothelial cells. Therefore, several tissue culture models that mimic BBB properties have been developed for the studies of BBB transport properties w2,39x. The distinguishing features of cerebral microvascular endothelial cells ŽCEC. are the presence of tight intercellular junctions, low rate of pinocytosis and expression of specific transporters such as hexose, basic and acidic amino acid transporters, P-glycoprotein, and other specialized transporters w26,28,37x. The BBB is also known to exhibit active transcellular transport due to selective or polar distribution of transport proteins in the opposite surfaces of the cells w4,32x. Glutathione ŽGSH., an ubiquitous, endogenous antioxidant, is important for brain function. GSH deficiency is

0006-8993r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 6 - 8 9 9 3 Ž 9 9 . 0 2 1 8 4 - 8

R. Kannan et al.r Brain Research 852 (2000) 374–382

known to be associated with a number of neurological diseases w25,29x. GSH serves multiple roles in the nervous system w6x. It is an antioxidant that scavenges free radicals and lipid peroxides, a redox modulator of ionotropic receptor functions, a neuroprotector, and possibly a novel neurotransmitter w11,14x. The major interest in our laboratory has been to define the metabolism of GSH in the brain. We have shown that GSH is transported across the BBB in the rat and guinea-pig w17,41x. GSH transport was independent of g-glutamyltranspeptidase ŽGGT.-mediated hydrolysis and resynthesis of GSH, and was carrier-mediated and exhibited inhibitor specificity w18,41x. Evidence for the expression of Naq-dependent and Naq-independent GSH transporters in bovine brain capillary mRNA-injected oocytes was also presented w19x. The localization of these two GSH transporters to the luminal and abluminal membranes would provide a strategy for raising brain endothelial GSH concentrations in pathophysiological conditions. With this in mind, to obtain further information on the biochemical features and regulation of GSH transport across the brain endothelium, we turned to a culture model of BBB. The model we have used is a transformed cell line of human brain cerebral microvascular endothelial cells ŽHCEC., which was recently shown to exhibit genotypic and phenotypic stability and to retain major structural and functional characteristics of BBB endothelium w26x. HCEC grown on filter inserts were utilized to study the domainspecificity of Naq-dependent GSH transporter. Moreover, since co-culture with astrocytes is known to affect the barrier and functional properties of BBB, we have also studied the effect of astrocyte-conditioned medium ŽACM. on GSH transport. Finally, the availability of neonatal rat astrocytes and fetal human astrocytes allowed us to determine uptake kinetics, and efflux of GSH as well as ion requirements for astrocyte GSH transport for the first time.

2. Materials and methods 2.1. Cultured cells Clones of SV-40 large T-antigen immortalized human cerebrovascular endothelial cells ŽHCEC. and fetal human astrocytes ŽFHAS. were obtained from Dr. Stanimirovic, National Research Council, Ottawa, Canada. The morphological and biochemical characteristics of the two cell types have been described elsewhere w26x. HCEC were grown in culture medium using Falcon 60 = 15 mm petri dishes. HCEC were also grown on 0.5% gelatin-coated, 1 mm filter inserts ŽFalcon 12 well plates, Becton-Dickinson, Franklin Lakes, NJ., either in regular medium or in the presence of media conditioned by cultured fetal human astrocytes. The regular culture medium for HCEC was 90% M199, 10% FBS and ITS premix Ž5 mgrml insulin, 5 mgrml transferrin and 5 ngrml selenium. plus 600 USP

375

unitsrl heparin. The medium for FHAS was 90% DMEM Žhigh glucose. q 10% FBS q antibiotics. Primary astrocytes were isolated from neonatal rat cerebral tissue according to the method of McCarthy and de Villis w27x and were characterized by the presence of glial fibrillary acidic protein ŽGFAP.. Precautions were taken to ensure no contamination with oligodendrocytes, fibroblasts, and microglial cells as described w5,27x. For this purpose, the culture medium was changed around 7–9 days. Flasks were equilibrated with CO 2 –O 2 for 1–2 h, and the caps tightly sealed. The flasks were shaken at 200 rpm at 378C for 24 h. The culture medium was then replenished and flasks were shaken for an additional 24 h. Fibroblast overgrowth was avoided by incubating for 48 h with 10y5 M cytosine arabinoside. Isolated cells were cultured in sterile Falcon plastic 35 mm culture dishes and maintained in Basal Eagle’s Medium with Earles’ balanced salts containing 15% fetal calf-serum, 0.1% glutamine, 0.6% glucose in a humid atmosphere of CO 2rair Ž5%r95%. at 378C. The medium was changed twice a week. The cells usually reached confluency at about 6–10 days in culture. 2.2. Astrocyte-conditioned medium (ACM) In order to study electrical resistance and uptake of GSH in HCEC in the presence of ACM, HCEC were grown in the presence of fetal human astrocytes conditioned medium as described below. FHAS were plated at 1:10 ratio in T175 or T125 flasks in DMEM. After cells reached ; 70% confluency, the medium was aspirated and cells were washed once with sterile PBS. Twenty milliliters of M199 Žwithout serum. were added, and after 18 or 72 h, the media was collected and spun at 4000 rpm for 10 min. The amount of protein in the collected ACM was estimated by BioRad assay. In all our studies, we used ACM with a protein concentration ranging from 100 to 150 mgrml. To grow HCEC in ACM, the medium was reconstituted with 10% FBS and ITS premix, heparin Ž0.6 USP unitsrml. and G418 Ž200 mgrml.. The reconstituted medium was filtered through 0.45 mm sterile filter before use in HCEC culture with ACM. 2.3. Transendothelial electrical resistance (TEER) TEER was measured using a Millicell Device ŽMillipore, Boston, MA. on culture days 1 through 14 in order to ascertain development of functional endothelial barrier. These measurements were made in cells grown in HCEC regular medium as well as in cells grown in fetal human ACM, essentially as described by Muruganandam et al. w26x. Background resistance offered by the culture medium and the filter insert without any HCEC was corrected for from all TEER measurements. The effect of ACM on TEER was checked after HCEC were maintained in ACM for either 18 h or 3 days. All filter inserts were seeded with

376

R. Kannan et al.r Brain Research 852 (2000) 374–382

equal number of cells Ž0.2 million cellsrfilter insert.. TEER measured on day 1 and day 2 in HCEC monolayers was not different from control filter inserts without cells. The net TEER Žafter subtracting background values in filters without HCEC. on days 3, 4, 5, 6, and 7 in filters grown in regular HCEC medium were 21.6 " 2.8, 24.2 " 2.8, 26.5 " 8.5, 29.5 " 7.0 and 40.1 " 8.0 V cm2 respectively. The TEER values of confluent cultures exposed to ACM in simultaneous experiments were not significantly different from that grown in regular medium on days 4 and 7 and were 23.4 " 2.8 and 31.2 " 10.1 V cm2 , respectively. 2.4. Glutathione transport in HCEC Uptake of GSH was determined in cells pretreated with acivicin Ž1 mM for 30 min. to inhibit GGT activity w19,20x. Initial studies performed in the presence or absence of acivicin established no significant difference in GSH uptake between the two groups Ždata not shown.. However, to fully ensure that GSH uptake that we measured did not include radioactivity from degradation and resynthesis of GSH, all uptake experiments were performed after inhibition of GGT with acivicin as in our previous work w17,20x. Uptake was carried out in sodiumcontaining and sodium-free buffers as described in our previous studies w19,20x. The composition of the Naq-containing buffer was Žin mM. 135 NaCl, 1.2 MgCl 2 , 0.81 MgSO4 , 27.8 glucose, 2.5 CaCl 2 , 25 HEPES, pH 7.4. Choline chloride Ž135 mM. replaced NaCl in the Naq-free buffer. The time course of uptake was first established by determining uptake at 378C and 48C at 1 mM GSH concentration for 2, 5, 10, 15, 30, 45 and 60 min in HCEC pretreated with acivicin. From these time course data, 30 min was chosen for performing all subsequent uptake studies. Uptake in HCEC was determined at 0.05 mM and 2 mM GSH containing w35 SxGSH Ž500 Cirmmol, DupontNEN, Braintree, MA. for 30 min in triplicate, using 1 million cellsrwell in sodium-containing and sodium-free buffers at 378C. Nonspecific binding was measured in parallel by performing uptake at 48C. Net uptake represented uptake at 378minus uptake at 48C. In some experiments, uptake was measured using w3 HxGSH Ž1000 Cirmmol, Dupont-NEN.. Efflux of GSH was measured in 3–5 million HCEC cells in petri dishes Ž60 mm = 15 mm. in Krebs–Henseleit buffer according to published procedures w23x. The efflux rate was determined from aliquots Ž0.5 ml. taken at various times Ž0, 15, 30, and 60 min.. GSH content in the extracellular medium and cells was determined by Tietze recyling assay for the calculation of fractional efflux and total efflux rates w19x. Molecular form of uptake of w35 SxGSH in HCEC was monitored by radiochromatography using HPLC analysis according to Fariss and Reed w9x. Either ; 2 million cells in culture dishes or ; 0.2 million cells isolated from the

filter inserts after performing 30 min GSH uptake from the luminal or abluminal side Žsee below. were processed for HPLC w9x. Uptake was predominantly Ž) 95%. in the form of GSH and the profile of radiochromatograms did not differ significantly between acivicin-pretreated or untreated cells. 2.5. Glutathione transport in cultured rat and human astrocytes In initial experiments, the optimal incubation time Žfor deriving linear initial rates. was determined. Neonatal, primary rat astrocytes grown in petri dishes were pretreated with 1 mM acivicin and 10 mM DL-buthionine sulfoximine ŽBSO. for 30 min at 378C to prevent breakdown of GSH and resynthesis from precursors, respectively w20x. They were then incubated for 2, 5, 15, 30, and 60 min in physiological buffer ŽHEPES 20 mM, NaCl 100 mM, sucrose 100 mM. with w35 SxGSHr1 mM DTT Ž2.5 mCir1.5 million cellsrwell containing either tracer alone, or in the presence of 2 or 20 mM unlabeled GSH. and cellular radioactivity was determined after 5 washes followed by treatment with 0.01% trypsin and 1% Triton X-100. Trapping of labeled GSH was estimated by measuring radioactivity at 48C. The incubation time for uptake was 30 min which was based on initial studies on time course of uptake. At first, effect of GGT inhibition on GSH uptake was tested at 0.05 and 2 mM GSH concentrations. In later experiments, kinetics of uptake was studied in the presence of 0.01–40 mM unlabeled GSH containing tracer GSH. These studies were conducted in NaCl buffer in acivicin-pretreated astrocytes. The kinetic parameters were derived from a nonlinear regression computer fit of the GSH uptake data to the Michaelis–Menten equation using the SAAM II program ŽSimulation Analysis and Modeling II; SAAM II User’s Guide, SAAM Institute, University of Washington, 1994. for curve fitting and estimation of kinetic parameters as described w20x. Uptake in FHAS was performed following a similar protocol to that for rat astrocytes experiments. GSH efflux in rat and human astrocytes was measured in a similar way to that of HCEC described above. 2.6. GSH uptake and efflux in endothelial cells grown on permeable filters For transport studies, HCEC in Falcon 12 well plates Ž0.2–0.3 million cellsrfilter insert. were plated in 0.5% gelatin-coated 1 mm filter inserts. They were grown in either regular medium or in the presence of media conditioned by fetal human astrocytes in culture as described previously w26x. In brief, HCEC were either grown only in regular medium or 5 days in regular medium followed by 18–48 h in ACM. TEER was monitored daily and uptake studies were carried out on day 7 after plating. Cells were first treated with 1 mM acivicin in both sides in respective

R. Kannan et al.r Brain Research 852 (2000) 374–382

growth medium for 30 min. After three washes, the medium was changed to either Naq-containing or Naq-free buffers. The volumes of incubation buffer of A Žtop or luminal. and B Žbottom or abluminal. sides of the filter inserts were 0.5 and 1.5 ml, respectively. w35 SxGSH Žplus 10 mM GSH. was added to A or B side and the cells were incubated at 378C for 30 min. Cells were then washed four times with respective chilled buffers, and the filter on which the cells were grown was cut out with a scalpel and transferred to a scintillation vial, lysed to release the intracellular radioactivity for subsequent assay using a scintillation counter. In some cases, the cells were trypsinized, and the supernatant of trichloroacetic acid-treated cell lysates was used for GSH derivatization for HPLC analysis w9x. For studying GSH efflux, cells were first incubated with w35 Sxcysteine for 6, 18 and 24 h in either regular culture medium or in cells exposed to ACM for 18 h. These studies were performed on day 6 or day 7. HPLC analysis showed that the radioactivity coeluted with the GSH peak and no labeled cysteine was detected after 6 h incubation. Therefore, the efflux studies were performed after 6 h of labeling of the intracellular GSH pool. After repeated washing of the extracellular 35 S-radioactivity, the cells were treated with acivicin to ensure minimal hydrolysis of labeled GSH. Release of w35 SxGSH to the luminal fluid was determined in Naq-containing and Naq-free incubation buffers at 15 and 30 min to calculate the fractional efflux. 2.7. GSH leÕels and GSH biosynthesis in HCEC in regular and astrocyte-conditioned media In separate experiments, GSH concentration in HCEC in regular culture medium and in cells maintained with ACM for 3 days was measured. Some experiments were also conducted after exposure of HCEC to ACM only for 18 h. These experiments were performed in Falcon 6 well plates. The rate of GSH synthesis under these conditions was also determined. For this purpose, cells were treated with 0.3 mM diethylmaleate ŽDEM. for 20 min which caused ; 85–90% GSH depletion. After washing to remove DEM, regular HCEC or ACM growth medium was added. GSH synthesis was then measured in the respective media at 0, 1, 3 and 6 h by recycling assay w38x. GGT activity in HCEC in the presence or absence of ACM was measured using the Sigma kit ŽSigma Diagnostic, procedure No. 845, St. Louis, MO..

3. Results 3.1. GSH uptake in cultured HCEC: linearity of uptake and sodium-dependence Fig. 1 shows the time course of uptake of GSH by HCEC. Uptake studies were performed in Naq-containing and Naq-free buffers at 378C in cells that were pretreated

377

Fig. 1. Time course of GSH uptake in HCEC pretreated with 1 mM acivicin in Naq-containing and Naq-free buffers. Uptake was determined at 1 mM GSH along with 1 mCi of w35 SxGSH per well. Net uptake of GSH is presented as nanomoles per million cells. Nonspecific binding determined by measuring uptake at 48C was below 15% for all time points and has been subtracted from the uptake at 378C. Data are mean"S.E.M. from five to six HCEC preparations each performed in duplicate.

with acivicin. Nonspecific cellular binding was determined by performing uptake at 48C in both incubation buffers. Data represent net uptake Žafter subtracting binding at 48C for all time-points. expressed as nmolrmillion cells. The nonspecific binding averaged between 8–15% of total uptake. Increase in net GSH uptake in Naq-containing and Naq-free incubation buffers as a function of time is shown in Fig. 1. Uptake in the absence of Naq was significantly lower for all the time points. All subsequent uptake studies were performed for 30 min. Uptake of GSH by HCEC was predominantly sodiumdependent. Fig. 2 shows net GSH uptake in acivicin pretreated HCEC at 0.05 and 1 mM GSH. The 30 min time point was chosen for uptake studies based on results shown in Fig. 1 Žsee also inset to Fig. 2.. Uptake values obtained without acivicin pretreatment for 0.05 and 1 mM GSH were 0.08 " 0.01 nmolrmillion cells in the presence of Naq and 0.03 " 0.01 and 0.38 nmolrmillion cells in the absence of Naq respectively. These data did not differ significantly from those with acivicin pretreatment which are shown in Fig. 2. Uptake was significantly lower in the absence of sodium. A similar trend for significant decrease in uptake in the absence of sodium was also found when the uptake data were expressed as nanomoles per milligram protein per 30 min Ždata not shown.. Uptake, as determined by HPLC, was in the form of GSH when performed in either buffer using w3 Hxglycine labeled GSH or w35 Sxcysteine labeled GSH. 3.2. Transendothelial electrical resistance (TEER) of HCEC TEER of HCEC was measured as a function of time in cells plated in gelatin-coated filter inserts. These measure-

378

R. Kannan et al.r Brain Research 852 (2000) 374–382

Fig. 2. Effect of Naq-removal on GSH uptake in HCEC pretreated with 1 mM GSH acivicin for 30 min. Cells were incubated for 30 min in buffers with or without sodium containing either 0.05 mM or 1 mM GSH and w35 SxGSH as described in Section 2. The inset shows time-course of uptake up to 30 min in both incubation buffers. Data are mean"S.E.M. from four preparations. Asterisk indicates significant difference Ž p- 0.05. as compared to Naq-containing buffer.

ments were made in regular HCEC medium and in ACM. TEER on days 1 and 2 were barely above the background both in regular medium and ACM. As presented in Section 2, TEER on day 4 Žafter subtracting background resistance in the absence of cells. in HCEC grown in regular medium was higher than that on days 1–3 and increased further on day 7 to 40.1 " 8.0 V cm2 . No further significant increase was found from days 7–14 Ždata not shown.. When HCEC were grown in ACM, TEER on days 4 and 7 did not differ significantly as compared to regular medium. 3.3. Asymmetry of GSH transport Uptake of GSH from either the luminal or abluminal side of HCEC was determined using cell monolayer filter

Fig. 4. Dependence of the luminal efflux of GSH in HCEC on Naq ions in the incubation medium. Efflux study was carried out in incubation buffers with or without sodium for 15 min in HCEC pretreated with acivicin. Cells contained w35 SxGSH Žbiosynthesized by preincubation with w35 Sxcysteine-containing medium for 6 h.. HCEC were grown either in regular medium or in ACM as described in Section 2. There was no significant difference in efflux values between the regular medium and ACM. Therefore, pooled data from the two groups Žof three preparations each. are shown.

inserts. These experiments were performed with day 7 cells that were grown in regular medium after ensuring that TEER was 40 V cm2 or higher as reported earlier w26x. HCEC were pretreated with or without acivicin and uptake was performed for 30 min at 378C either in Naqcontaining or Naq-free buffer. As seen in Fig. 3, the uptake from the luminal fluid of HCEC was significantly inhibited by removal of Naq while no significant inhibition of uptake was seen from the abluminal fluid. Uptake was verified to be in the form of GSH by HPLC. To study efflux, HCEC cells grown either in regular medium or ACM were allowed to synthesize w35 SxGSH from w35 Sxcysteine in growth medium for 6 h. HPLC results established that all cellular radioactivity was in the form of w35 SxGSH after 6 h Žand 18 and 24 h. after incubation with w35 Sxcysteine Ždata not shown.. Efflux rates of GSH were determined in cells whose GSH pool was prelabeled for 6 h with incubation with w35 Sxcysteine. The

Fig. 3. GSH uptake in acivicin-pretreated HCEC grown in filter inserts on day 7 after seeding 0.2 million cells per gelatin-coated filter insert as described in Section 2. Cells were grown in HCEC medium and uptake was determined for 30 min after introduction of w35 SxGSH Žq10 mM GSH in 2 mM DTT. either on the A Žluminal. or on the B Žabluminal. side in sodium-containing or sodium-free buffers. Data are mean " S.E.M. from four preparations per group. Asterisk indicates p - 0.05 as compared to Naq-containing buffer.

R. Kannan et al.r Brain Research 852 (2000) 374–382

379

Fig. 5. ŽA. Cellular GSH levels Žnmol GSHrmg protein. in day 5 HCEC. ŽB. Biosynthesis of GSH in HCEC pretreated with 0.3 mM DEM for 20 min. After washing off DEM, cells were equilibrated in regular medium Žclosed circles., or human ACM Žopen circles. and time course of GSH synthesis was followed for up to 6 h. The caption ‘‘Reg’’ refers to regular HCEC medium and ‘‘cond’’ refers to ACM. Data are mean" S.E.M. from five to six preparations from each group.

release of radioactivity to the luminal Ži.e., apical, A. side was determined in both Naq-containing or Naq-free incubation buffers at 15 and 30 min. As there was no significant difference in the release of radioactivity into the medium between the HCEC grown in regular medium or from cells subject to prior exposure to ACM for 18 h pooled data are presented in Fig. 4. As can be seen, total efflux was higher in the absence of sodium as compared to that observed in the presence of sodium, suggesting that a portion of GSH effluxed in the presence of Naq may undergo reuptake by a Naq-dependent uptake mechanism.

days while the rest were cultured in regular growth medium. On day 5, total GSH levels in these two cell groups was not statistically different from each other ŽFig. 5A.. The rate of synthesis of GSH, which was determined by resynthesis from precursors after prior depletion of GSH pool with DEM w23x, was also not significantly different between the two groups ŽFig. 5B.. When HCEC were cultured in regular medium for 4 days followed by 18 h Žinstead of 3 days. in ACM, the GSH levels in the two groups were 10.1 " 1.1 and 7.7 " 1.8 nmolrmg protein, respectively, indicating that the rate of synthesis was not significantly altered under these conditions.

3.4. Effect of ACM on HCEC: GSH leÕels and rates of synthesis

3.5. Transport of GSH in rat and human astrocytes

In order to find out whether HCEC grown in the presence of ACM showed decreased cellular GSH as reported recently for human umbilical venous endothelial cells w13x, HCEC were grown either in regular medium or in the presence of ACM. The cells were cultured in regular medium for 2 days, after which half the number of petri dishes containing HCEC were incubated with ACM for 3

Uptake of GSH in neonatal rat astrocytes and fetal human astrocytes at two different GSH concentrations is shown in Fig. 6A and B. Uptake was partially sodium-dependent in both species, although inhibition of uptake in the absence of sodium was more pronounced in FHASgrown cells. Additional experiments to determine kinetics of GSH uptake were performed in neonatal astrocytes and

Fig. 6. Effect of sodium removal on GSH uptake in neonatal rat ŽA. and fetal human ŽB. astrocytes. Net uptake at different GSH concentrations is shown. Data represent mean " S.E.M. for five to nine preparations of acivicin-pretreated astrocytes from each species.

380

R. Kannan et al.r Brain Research 852 (2000) 374–382

only in Naq-containing buffer. Uptake of GSH exhibited saturation kinetics in neonatal rat astrocytes. The effect of varying concentrations of unlabeled carrier Ž0.01–40 mM. showed that the uptake was saturable with a mean apparent K m of 3.3 mM and Vmax of 0.62 nmolrmillion cellsrmin. The rate of efflux which was measured in incubations with a Krebs–Henseleit buffer supplemented with 12.5 mM HEPES ŽpH 7.4., was 12 " 2%rh. The corresponding GSH efflux rate in human astrocytes was 3.5 " 0.2%rh. Analysis by HPLC revealed that the molecular form of efflux and uptake was GSH Ždata not shown..

4. Discussion In this study, we have presented evidence for intact GSH uptake and efflux from human cerebral microvascular endothelial cells. Both sodium-dependent and sodiumindependent uptake mechanisms were identified. Our studies further showed that the Naq-dependent transport is localized to the luminal Žapical. membrane of brain endothelial cells. The recent generation of a transformed cell line of human cerebrovascular endothelial cells w26x offers an in vitro model for investigating uptake of substrates including amino acids and peptides. It has been shown that HCEC possessed the biochemical and morphological characteristics of the BBB w26x. We were able to reproduce the presence of tight junctions in HCEC in our present studies by measurements of TEER; TEER of HCEC, however, did not increase significantly in the presence of ACM in our hands, although we found that GGT activity increased threefold from 24.2 " 2.9 to 68 " 6 milliunitsrmg protein w7x. The reason for this discrepancy is not clear. Several factors including the source of the astrocyte cell line, duration of exposure to the astrocyte enriched medium, relative concentration of growth factorŽs., growth matrix for HCEC, and source species of cells may all play a role. Many investigators have reported tightening of junctions in the presence of growth factors derived from astrocytes w1,15,31,36x. Lack of a significant increase in TEER was noted by Guerin and Bobilya w10x who found that a hypothalamic extract and heparin increased barrier function and electrical resistance in porcine brain endothelial cells while co-culture with glioma cells or presence of ACM were without effect. Since the uptake and efflux of GSH were not significantly affected by ACM in the present study, we have not pursued the causes of the above discrepancy any further. Our previous work in in vivo and cultured cell models of BBB has shown that GSH is taken up intact and mouse brain endothelial cell mRNA injected oocytes exhibited GSH transport which was partially sodium-dependent. Furthermore, our very recent uptake studies using plasma membrane vesicles from mouse brain endothelial cells ŽMBEC. showed that the Naq-dependent GSH uptake

system is localized to the luminal membrane w20x. Since these cells do not form a tight barrier with high electrical resistance, it was not possible to examine the polarity of sodium-dependent GSH transport in MBEC. Therefore, the present work with human brain endothelial cells provides independent confirmation for the luminal localization of the sodium-dependent GSH transporter using a different experimental approach. The finding in the present studies that there is less GSH efflux from HCEC in the presence of sodium than in the absence of sodium further confirms that a Naq-driven GSH efflux is operational at the luminal membrane. A predominant fraction of GSH in the brain is in the astroglial compartment which serves as a GSH reservoir and may provide extracellular antioxidant protection through GSH efflux to neurons w8,35x. A major supporting role in brain neurotransmitter functions has been attributed to astrocytes w8,11x. Release of astrocytic GSH may inhibit NMDA receptor agonist binding w22x. Neurons have very low levels of GSH w24,30x but survive much longer if seeded on a glial feeder layer suggesting that GSH may serve the important function of providing antioxidant defense to neurons andror modulate the action of neurotransmitters as a competitive inhibitor. Thus, GSH homeostasis in astrocytes is of importance in preventing neuronal cell death. Evidence for carrier-mediated GSH efflux has been presented in recent work in rat astrocytes w33,40x. We confirmed the rapid GSH efflux from rat astrocytes and extended it to fetal human astrocytes in the present study. Furthermore, we showed uptake of GSH in neonatal rat astrocytes and in an immortalized cell line of fetal human astrocytes for the first time. Uptake was found to be partially sodium-dependent in astrocytes from both species. The uptake system for GSH in rat astrocytes was carrier-mediated. These studies were performed only in a sodium-containing incubation buffer due to the limited availability of primary astrocytes. Thus, our results show that both sodium-independent and sodium-dependent GSH transport processes are present in both astrocytes and in brain endothelial cells. It is likely that the bidirectional, sodium-independent GSH transport may serve as an efflux transporter under physiological conditions. The role of Naq-dependent GSH transporter in astrocytes is not clear and needs further study. It is of interest that Shaw et al. w34x have recently reported GSH-induced sodium currents in neocortex and this depolarization was not mediated by NMDA receptor activation, thus suggesting a possible role for a Naq-GSH cotransport under these conditions. The exact role of the GSH transporters in brain physiology is yet to be fully determined. We may, however, make some speculations based on our current findings. Demonstration of Naq-dependent GSH transport is consistent with our hypothesis that this transporter may mediate uphill transport from low plasma GSH Ž10–20 mM. to mM endothelial cellular GSH and can lead to net accumulation in endothelial cells. The fact that intracellular GSH con-

R. Kannan et al.r Brain Research 852 (2000) 374–382

centrations of cultured endothelial cells could be raised by incubation with extracellular GSH only in the presence of Naq in the incubation medium w20x is in accord with this hypothesis. We have also accumulated evidence for the presence of an efflux mechanism from the abluminal surface of the endothelium and in astrocytes. This Naq-independent GSH transport, as mentioned above, may function as a facilitative transporter and would have the potential for bidirectional operation as observed in cell culture studies. Recent studies suggest that other organic anion transporters such as MRP1 may be localized to the luminal brain endothelial cell membranes w12,16,21x and members of the oatp family, particularly oatp1, may also mediate GSH transport w3x. At any rate, further work on the cloning of the brain GSH transporters, particularly the Naq-dependent endothelial GSH transporter, is likely to be of value in understanding GSH homeostasis and delivery of this important antioxidant to the brain.

Acknowledgements We thank Drs. D. Stanimirovic, and A. Murugandandam of NRC, Canada for providing us the HCEC and FHAS cell lines and for helpful discussions. We also thank Vyjayanthi Raghunathan for technical help in initial experiments with cells grown on filter inserts. This work was supported in part by NIH grants GM 53820 ŽRK., HL 38658 ŽKJK. and HL 46943 ŽKJK..

References w1x F.E. Arthur, R.R. Shivers, P.D. Bowman, Astrocyte-mediated induction of tight junctions in brain capillary endothelium: an efficient in vitro model, Dev. Brain Res. 36 Ž1987. 155–159. w2x K.L. Audus, L. Ng, R.T. Borchardt, Brain microvessel endothelial cell culture systems, in: R.T. Borchardt, P.I. Smith, G. Wilson ŽEds.., Models for Assessing Drug Absorption and Metabolism, Vol. 13, Plenum, New York, 1996, pp. 239–258. w3x N. Ballatori, J.F. Rebbeor, MRP2 and oatp1 in hepatocellular export of cellular glutathione, Semin. Liver Dis. 18 Ž1998. 377–387. w4x A.L. Betz, J.A. Firth, G.W. Goldstein, Polarity of the blood–brain barrier: distribution of enzymes between the luminal and antiluminal membranes of brain capillary endothelial cells, Brain Res. 192 Ž1980. 17–28. w5x R. Cole, J. de Villis, Preparation of Astrocyte and Oligodendrocyte Cultures from Rat Glial Cultures, Manual of the Nervous System, A.R. Liss, New York, pp. 121–133. w6x A.J.L. Cooper, B.S. Kristal, Multiple roles of glutathione in the central nervous system, Biol. Chem. Hoppe-Seyler 378 Ž1998. 793– 802. w7x L.E. Debault, P.A. Cancilla, Gamma-glutamyltranspeptidase in isolated brain endothelial cells: induction by glial cells in vitro, Science 207 Ž1980. 653–655. w8x R. Dringen, L. Kussmaul, M. Gutterer J, J. Hirrlinger, B. Hamprecht, The glutathione system of peroxide detoxification is less efficient in neurons than in astroglial cells, J. Neurochem. 72 Ž1999. 2523–2530.

381

w9x M.W. Fariss, D.J. Reed, High performance liquid chromatography of thiols and disulfides: dinitrophenol derivatives, Methods Enzymol. 143 Ž1987. 101–109. w10x J.L. Guerin, D.J. Bobilya, Hypothalamic extract influences a blood–brain barrier model of porcine brain capillary endothelial cells, Neurochem. Res. 22 Ž1997. 321–326. w11x N. Guo, C. McIntosh, C. Shaw, Glutathione: new candidate neuropeptide in the central nervous system, Neuroscience 51 Ž1992. 835–842. w12x H. Huai-Yun, D.T. Secrest, K.S. Mark, D. Carney, C. Brandquist, W.F. Elmquist, D.W. Miller, Expression of multidrug resistance-associated protein ŽMRP. in brain microvessel endothelial cells, Biochem. Biophys. Res. Commun. 243 Ž1998. 816–820. w13x R.D. Hurst, S.J. Heales, M.S. Dobbie, J.E. Barker, J.B. Clark, Decreased endothelial cell glutathione and increased sensitivity to oxidative stress in an in vitro blood–brain barrier model system, Brain Res. 802 Ž1998. 232–240. w14x R. Janaky, K. Ogita, B.A. Pasqualotto, J.S. Bains, S.S. Oja, Y. Yoneda, C.A. Shaw, Glutathione and signal transduction in the mammalian CNS, J. Neurochem. 73 Ž1999. 889–902. w15x J.C. Janzer, M.C. Raff, Astrocytes induce blood–brain properties in endothelial cells, Nature 325 Ž1987. 253–257. w16x D.W. Loe, R.G. Deeley, S.P.C. Cole, Characterization of vincristine transport by the M r 190,000 multidrug resistance protein ŽMRP.: evidence for cotransport with reduced glutathione, Cancer Res. 58 Ž1998. 5130–5136. w17x R. Kannan, J.F. Kuhlenkamp, E. Jeandidier, H. Trinh, M. Ookhtens, N. Kaplowitz, Evidence for carrier-mediated transport of glutathione across the blood–brain barrier in the rat, J. Clin. Invest. 85 Ž1990. 2009–2013. w18x R. Kannan, J.F. Kuhlenkamp, M. Ookhtens, N. Kaplowitz, Transport of GSH at the blood brain barrier in the rat: inhibition and age-dependence, J. Pharmacol. Exp. Ther. 263 Ž1992. 964–970. w19x R. Kannan, J.-R. Yi, D. Tang, B.V. Zlokovic, N. Kaplowitz, Evidence for the existence of a sodium-dependent glutathione ŽGSH. transporter. Expression of bovine brain capillary mRNA and size fractions in Xenopus laeÕis oocytes and dissociation from gamma glutamyltranspeptidase and facilitative GSH transporters, J. Biol. Chem. 271 Ž1996. 9754–9758. w20x R. Kannan, A. Mittur, Y. Bao, T. Tsuruo, N. Kaplowitz, GSH transport in immortalized mouse brain endothelial cells: evidence for apical localization of a sodium-dependent glutathione transporter, J. Neurochem. 73 Ž1999. 390–399. w21x H. Kusuhara, H. Suzuki, M. Naito, T. Tsuruo, Y. Sugiyama, Characterization of efflux transport of organic anions in a mouse brain capillary endothelial cell line, J. Pharmacol. Exp. Ther. 285 Ž1998. 1260–1265. w22x S.W. Leslie, L.M. Brown, R. Trent, Y.-H. Kee, J.L. Morris, T.W. Jones, P.K. Randall, S.S. Lau, T.J. Monks, Stimulation of N-methylD-aspartate receptor-mediated calcium entry into dissociated neurons by reduced and oxidized glutathione, Mol. Pharmacol. 41 Ž1992. 308–314. w23x S.C. Lu, H.-Y. Huang, Comparison of sulfur amino acid utilization for GSH synthesis between HepG2 cells and cultured rat hepatocytes, Biochem. Pharmacol. 47 Ž1994. 859–869. w24x T.K. Makar, M. Nedergaard, A. Preuss, A.S. Gelbard, A.S. Perumal, A.J.L. Cooper, Vitamin E, ascorbate, glutathione, glutathione disulfide, and enzymes of glutathione metabolism in cultures of chick astrocytes and neurons; evidence that astrocytes play an important role in antioxidant processes in the brain, J. Neurochem. 62 Ž1994. 45–53. w25x A. Meister, M.E. Anderson, Glutathione, Annu. Rev. Biochem. 52 Ž1983. 711–760. w26x A. Muruganandam, H.L. Moorhouse, R. Monete, J.P. Durkin, D. Stanimirovic, Development of immortalized human cerebromicrovascular endothelial cell line as an in vitro model of human blood–brain barrier, FASEB J. 11 Ž1997. 1187–1197.

382

R. Kannan et al.r Brain Research 852 (2000) 374–382

w27x K.D. McCarthy, J. de Villis, Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue, J. Cell Biol. 85 Ž1980. 890–902. w28x W.H. Oldendorf, Brain uptake of labeled amino acids, amines, and hexoses after arterial injection, Am. J. Physiol. 221 Ž1971. 1629– 1639. w29x M. Orlowski, A. Karkowsky, Glutathione metabolism and some possible functions of glutathione in the nervous system, Int. Rev. Neurobiol. 19 Ž1976. 75–121. w30x S.A. Raps, J.C.K. Lai, L. Hertz, A.J.L. Cooper, Glutathione is present in high concentrations in cultured astrocytes but not in cultured neurons, Brain Res. 493 Ž1989. 398–401. w31x T.J. Raub, S.L. Kuentzel, G.A. Sawada, Permeability of bovine brain microvessel endothelial cells in vitro: barrier tightening by a factor released from astroglioma cells, Exp. Cell Res. 199 Ž1992. 330–340. w32x M.M. Sanchez del Pino, D.R. Peterson, R.A. Hawkins, Biochemical discrimination between luminal and abluminal enzyme and transport activities of the blood–brain barrier, J. Biol. Chem. 270 Ž1995. 14913–14918. w33x J. Sagara, N. Makino, S. Bannai, Glutathione efflux from cultured astrocytes, J. Neurochem. 66 Ž1996. 1876–1881. w34x C.A. Shaw, B.A. Pasqualotto, K. Curry, Glutathione-induced sodium currents in neocortex, NeuroReport 7 Ž1996. 1149–1152. w35x M.L. Shroeter, K. Mertsch, H. Giese, S. Muller, A. Sporbert, B.

w36x

w37x

w38x

w39x

w40x

w41x

Hickel, E. Blasig, Astrocytes enhance radical defence in capillary endothelial cells constituting the blood–brain barrier, FEBS Lett. 449 Ž1999. 241–244. J.H. Tao-Cheng, Z. Nagy, M.W. Brightman, Tight junctions of brain endothelium in vitro are enhanced by astroglia, J. Neurosci. 7 Ž1987. 3293–3299. T. Tatsuta, M. Naito, T. Oh-hara, I. Sugawara, T. Tsuruo, Functional involvement of P-glycoprotein in blood–brain barrier, J. Biol. Chem. 267 Ž1992. 20383–20391. F. Tietze, Enzymic method for quantitative determination of nanogram amounts of total and oxidized glutathione: applications to mammalian blood and other tissues, Anal. Biochem. 27 Ž1994. 502–522. J.B.M.M. Van Bree, A.G. de Boer, M. Danhof, L. Ginsel, D.D. Breimer, Characterization of an ‘‘in vitro’’ blood–brain barrier: effects of molecular size and lipophilicity on cerebrovascular endothelial transport rates of drugs, J. Pharmacol. Exp. Ther. 247 Ž1988. 1233–1239. M. Yudkoff, D. Pleasure, L. Cregar, Z. Lin, I. Nissim, J. Stern, I. Nissim, Glutathione efflux from cultured astrocytes: studies with w15 Nxglutamate, J. Neurochem. 55 Ž1990. 137–145. B.V. Zlokovic, J.B. Mackic, J.G. McComb, M.H. Weiss, N. Kaplowitz, R. Kannan, Evidence for transcapillary transport of reduced glutathione in vascular perfused guinea-pig brain, Biochem. Biophys. Res. Commun. 201 Ž1994. 402–408.