Functional characterisation of the active ascorbic acid transport into cerebrospinal fluid using primary cultured choroid plexus cells

Brain Research 988 (2003) 105–113 www.elsevier.com / locate / brainres

Research report

Functional characterisation of the active ascorbic acid transport into cerebrospinal fluid using primary cultured choroid plexus cells Susanne Angelow, Matthias Haselbach, Hans-Joachim Galla* ¨ Biochemie, Westf alische ¨ ¨ , Wilhelm-Klemm-Straße 2, D-48149 Munster ¨ Wilhelms-Universitat , Germany Institut f ur Accepted 17 July 2003

Abstract Crossing the blood–CSF barrier is an important pathway for certain nutrients to enter the CNS. Cultured choroid plexus epithelial cells are a potent model system to study active transport properties of this tissue in vitro. In the present study this in vitro model was used to analyse ascorbic acid transport across the blood–CSF barrier that is supposedly mediated by the Na 1 -dependent transporter SVCT2. The expression of SVCT2 in the cultured cells was proven by RT-PCR. Active transport across the cell monolayer resulted in ascorbic acid enrichment at the CSF mimicking side. Ascorbic acid transport and uptake were decreased to 13 and 27%, respectively, in the presence of 200 mM phloretin. Inhibition of both transepithelial substrate transport (to 7.5%) and cytoplasmatic uptake (to 20%) was observed in Na 1 -free medium indicating that a basolaterally located and Na 1 -dependent transporter mediates ascorbic acid uptake. Substituting Cl 2 by either iodide or D-gluconate increased ascorbic acid uptake by factors of 3.7 or 2.5, respectively. Similar observations were made when Na 1 -dependent myo-inositol transport was analysed. Additionally, in presence of 100 mM bumetanide, an inhibitor of Na 1 -Cl 2 cotransport, indirectly increased ascorbic acid and myo-inositol transport rates were observed showing that ascorbic acid-Na 1 -cotransport might balance low intracellular Na 1 concentration.  2003 Elsevier B.V. All rights reserved. Theme: Cellular and molecular biology Topic: Blood–brain barrier Keywords: Blood–CSF barrier; Choroid plexus epithelial cell; Ascorbic acid transport; SVCT2

1. Introduction Two barrier systems, the blood–brain barrier and the blood–CSF barrier, serve to maintain brain homeostasis by restricting uncontrolled diffusion of water-soluble molecules into CNS. For proper neural function macronutrients and micronutrient are required that enter the brain either by facilitated diffusion or active transport, mediated by spe-

Abbreviations: AA, ascorbic acid; CP, choroid plexus; CSF, cerebrospinal fluid; DIV, days in vitro; DHA, dehydroascorbic acid; MI, myoinositol; SFM, serum free medium; TER, transepithelial electrical resistance *Corresponding author. Tel.: 149-251-83-33201; fax: 149-251-8333206. E-mail address: [email protected] (H.-J. Galla). 0006-8993 / 03 / $ – see front matter  2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0006-8993(03)03350-X

cific transport systems. The blood–CSF barrier forming epithelial cells of the choroid plexus (CP) are supposed to play an important role only in micronutrient homeostasis [20], e.g., in vitamin C transport into CNS [23]. Great amounts of ascorbic acid (AA), the reduced form of vitamin C, are enriched in the brain. AA concentration in the cerebrospinal fluid (CSF) is approximately 200 mM and thus about four times higher than typical serum concentrations of 50 mM [17,24]. In neural tissue AA concentration even reaches millimolar ranges, depending on cell type and regional distribution as reviewed by Rice [18]. The redox characteristics of AA enable this molecule to act as an important enzyme cofactor in metabolic reactions and as a trap for free radicals protecting tissues from oxidative damage. Within the CNS, AA is involved in neurotransmitter and hormone synthesis as a cofactor for dopamine-b-hydroxylase and peptidylglycine a-amidating monooxygenase [5,8], in myelin formation by stimulating

106

S. Angelow et al. / Brain Research 988 (2003) 105–113

Schwann-cell differentiation [6] and as an antioxidant it has neuroprotective functions. In CP epithelium the localisation of the recently cloned AA specific transporter SVCT2 was proven by in situ hybridisation [25]. Daruwala et al. [3] demonstrated for transfected Xenopus oocytes, that via the SVCT2 system, AA transport is coupled to a cotransport of two Na 1 . A Na 1 gradient established by the Na 1 –K 1 -ATPase is the driving force of this process. A second pathway for vitamin C to enter CNS is to pass the endothelial cells of the brain microvessels in its oxidised form dehydroascorbic acid (DHA) via the glucose transporter GLUT1 [1]. The existence of the first pathway is underlined by the facts that vitamin C is circulating in the blood in its reduced form most of its lifetime [4] and that it has to be transported against a steep concentration gradient into the CNS. Although some groups consider the second pathway to be prominent, vitamin C transport by GLUT 1 seems unlikely in the presence of physiological glucose plasma concentrations. Further investigations are necessary to quantify the uptake routes of vitamin C into CNS for both routes. It has also been shown that myo-inositol (MI) is actively transported by the CP epithelium and is enriched in the CSF [21,22]. The mechanism and regulation of the MINa 1 -cotransporter SMIT that specifically promotes transepithelial MI transport has been demonstrated before [13,15]. In earlier studies the barrier and transport properties of primary cultured epithelial cells of porcine CP have been investigated. The cells form confluent monolayers within a few days after seeding them on permeable filter membranes. They establish high transepithelial electrical resistances (TER) of about 1000–1700 V?cm 2 after removal of serum from the medium and actively transport fluid from the basolateral to the apical chamber [9]. AA and MI are transported from the basolateral to the apical side in a concentration-dependent process [10]. The KM values for the transport of AA (67 mM) and MI (117 mM) are in good agreement to data derived from uptake measurements with CP tissue [21,23] and SVCT2-cDNA transfected HRPE-cells [16]. We concluded that this cell culture model closely mimics the in vivo situation and can be applied to study transport characteristics of blood–CSF barrier in vitro, alternatively to tissue or in vivo experiments. Here we report for the first time that cultured CP epithelial cells as a model of the blood–CSF barrier are well suited to investigate AA transport mechanism. From inhibition studies we conclude that AA is taken up by a Na 1 -dependent transport system as it has been shown for SVCT2 mediated transport. Additionally we demonstrate the transporter’s localisation within the basolateral membrane. In several experiments we have studied the inhibition and activation of the transporter and its interactions with other transport systems depending on ion concentration gradients.

2. Materials and methods

2.1. Materials DME / HAM’s F12-medium (1:1) was obtained from Bioconcept (Freiburg, Germany). 0.25% trypsin solution, fetal calf serum (FCS), penicillin and streptomycin solution and L-glutamine were from Biochrom (Berlin, Germany). Insulin, cytosine arabinoside, ascorbic acid, myoinositol and laminin were obtained from Sigma (St. Louis, MO, USA). All other chemicals were obtained from Merck (Darmstadt, Germany). [ 14 C]Ascorbic acid and [ 3 H]myoinositol were from Amersham Biosciences (Buckinghamshire, UK). Cells were cultured in the Transwell  -filter systems from Costar (Cambridge, USA) or in culture flasks from Nunc (Wiesbaden, Germany). The transepithelial electrical resistance was determined by impedance analysis using the Impedance Analyzer SI-1260, Solartron Instruments (Farnborough, UK). Radioactive isotopes were analysed with the scintillation counter LS 6500 from Beckmann (Fullerton, USA). For molecular biology High Pure RNA Isolation Kit from Roche Diagnostics (Mannheim, Germany) was used. dNTPs, oligo(dT)-primer, reverse transcriptase Superscript姠II, PCR-buffer and Taq DNA polymerase were obtained from Invitrogen (Karlsruhe, Germany). SCVT-primer were synthesised by MWG Biotech (Ebersberg, Germany). PCR was performed with the Mastercycler from Eppendorf (Hamburg, Germany). DNA purification Kit from Biozym (Oldendorf, Germany), the AdvanTAge姠PCR Cloning Kit from Clontech (Paolo Alto, USA) and the Spin Miniprep Kit from Quiagen (Hilden, Germany) were used. DNA sequencing was performed by the method of Sanger with a LI-COR DNA Analyser Gene Reader (MWG Biotech) using M13 forward primer and DYEnamic Direct cycle sequencing Kit (Amersham Biosciences).

2.2. Preparation and cultivation of CP epithelial cells Epithelial cells from porcine CP were obtained by a preparation described in detail by Hakvoort et al. [9,10]. Isolated CP epithelial cells were seeded on laminincoated permeable membranes with a diameter of 24 mm if not mentioned differently using a seeding density of 20 mg / cm 2 wet weight of CP tissue. The medium was supplemented with 10% (v / v) FCS, 4 mM L-glutamine, 5 mg / ml insulin and 100 mg / ml penicillin and streptomycin. Twenty mM cytosine arabinoside were added to suppress the growth of contaminating cells [7]. Serum was added to the medium only during the first days in culture to support cell proliferation. After reaching confluence (7–9 DIV) growth medium was replaced by serum free culture medium (SFM) to support cell differentiation and improve barrier function [9]. The quality of the cell preparation was characterised by impedance analysis as described in detail by Wegener et al. [27]. Only filters with a transepithelial

S. Angelow et al. / Brain Research 988 (2003) 105–113

electrical resistance (TER) of at least 800 V?cm 2 were used for further experiments.

2.3. Transport studies In transport experiments cells were treated with an incubation buffer containing 114 mM NaCl, 1.25 mM CaCl 2 , 1.1 mM MgCl 2 , 5 mM KCl, 20 mM NaHCO 3 , 10 mM Hepes, 1.65 mM Na 2 HPO 4 , 0.3 mM NaH 2 PO 4 , and 25 mM glucose, pH 7.3. The concentrations of ions, Hepes and glucose were adjusted to the concentrations in the cell culture medium in order to avoid osmotic stress after exchange of the culture fluid. Error bars shown in the graphics were determined as follows: for n$4 errors are calculated by S.D., for n52 and n53 errors are the maximum deviation of the mean value.

2.3.1. Efflux For efflux measurement the incubation buffer sup14 plemented with 10 mM AA and 0.2 mCi / ml [ C]AA was applied from both sides of the filter system. After 1 h the buffer was removed and the cells were washed briefly with pure incubation buffer. One ml incubation buffer was added to the apical and basolateral side and the release of substrate to each side was determined after 12 min. Additionally uptake of [ 14 C]AA into the cell monolayers at t50 of separate control filters was measured. Efflux was expressed by the percentage of AA in the compartment after 12 min with respect to the amount inside the cell monolayer at time t50. 2.3.2. Transport across the cell monolayer Active transport of AA and MI across the cell monolayer was determined as follows. Cells were preincubated for 45 min with (1) incubation buffer described above, or (2) incubation buffer containing inhibitors, or (3) Na 1 -free buffer (114 mM choline chloride, 1.25 mM CaCl 2 , 1.1 mM MgCl 2 , 5 mM KCl, 20 mM KHCO 3 , 10 mM Hepes, 1.65 mM K 2 HPO 4 , 0.3 mM KH 2 PO 4 and 25 mM glucose, pH 7.3), or (4) Cl 2 -free medium containing iodide (114 mM NaI, 5 mM KI, 1.25 mM CaSO 4 , 1.1 mM magnesium acetate, 20 mM NaHCO 3 , 10 mM Hepes, 1.95 mM NaH 2 PO 4 and 25 mM glucose, pH 7.3) or (5) Cl 2 -free medium containing gluconate (114 mM Na-gluconate, 5 mM potassium gluconate, 1.25 mM CaSO 4 , 1.1 mM magnesium acetate, 20 mM NaHCO 3 , 10 mM Hepes, 1.95 mM NaH 2 PO 4 and 25 mM glucose, pH 7.3), respectively. After preincubation, the same buffers as described above supplemented with 30 mM AA and MI, respectively, and 0.17 mCi / ml radioactive labelled substrate were added to both sides of the cell monolayer. After 2 h, samples taken from the medium of the apical and basolateral chamber were counted. Within this time interval the rates of AA and MI transport remain constant. Therefore this time interval is very suitable to detect both significant substrate enrichment in the apical chamber and effects of inhibitors. A

107

decay or oxidation of AA within this period can be excluded since the amount of AA does not change during the experiment (proven by titration with dichlorophenolindophenol (DCPIP)). The amount of fluid that was secreted into the apical chamber within 2 h was determined on control filters by weighting the filter inserts containing the apical medium with a microbalance (precision50.01 mg). Additionally, by impedance analyses we excluded a breakdown of the monolayer’s barrier properties or a modification of the cell membrane. The transport across the cell monolayer per area and time interval, Dn /(A ? t), leading to substrate enrichment in the apical compartment and dilution at the basolateral side, was calculated by the following equation ct ]t ? c ?V 2 c 0 ?V0 n t 2 n 0 ct 0 0 t Dn ] 5 ]] 5 ]]]]]] A? t A? t A? t with n 0 and n t being the amount of substrate, ct 0 and ct t are the readings of the scintillation counter and V0 and Vt are the volume of the apical medium at the beginning of the experiment and at time t. c 0 is the substrate concentration at t50.

2.3.3. Uptake into the cell monolayer Experiments were carried out according to the protocol given for transport measurements. After 20 min, buffers were removed and cell monolayers were washed briefly with 4 8C cold buffer. Within this time interval uptake increases continuously and then slowly decreases due to substrate diffusion from the cell monolayer. The incubation was stopped after 120 min when uptake experiments were performed simultaneously with transport measurements. Filters were excised out of the inserts and the amount of substrate enclosed in the cell monolayer per area, n /A, was determined by the following equation n c ?VM ? ct F ] 5 ]]] A A ? ct M with c being the substrate concentration in incubation buffer, ct M the counts of incubation buffer (medium), VM the volume of aliquot taken from the incubation buffer and ct F the counts of the filter. The amount of substrate taken up by laminin coated filters without an overlaying cell sheet is negligible. Several uptake experiments were performed with n52 due to low amounts of material received from the cell preparation. However the very small deviations to the mean values show that the results are significant.

2.4. RT-PCR and sequence analysis RNA was isolated from fresh porcine kidney and from cultured CP cells grown to confluent monolayers in cell culture flasks after 14 DIV. The mRNA was transcribed to cDNA using 5 mg RNA as a template, that were incubated

108

S. Angelow et al. / Brain Research 988 (2003) 105–113

with oligo(dT) 12 – 18 primers for 10 min at 70 8C. Four ml 53 first strand buffer, 2 ml of 0.1 M DTT and 1 ml of 10 mM dNTP mix were added and the mixture was incubated for 2 min at 42 8C. Two hundred U Reverse Transcriptase Superscript姠 II was added, incubation was continued for additional 50 min and the reaction was then stopped by heating to 70 8C. Two ml of RT product were used as a template for subsequent PCR amplification in a volume of 50 ml containing 20 mM Tris, 50 mM KCl, 1.5 mM MgCl 2 , 0.2 mM dNTP mix, 5 pmol forward and reverse primer and 0.05 U Taq DNA polymerase. Primers were selected based on identical sequences of human and rat SVCT1 and SVCT2, respectively: SVCT2 forward: 59GCTGCTCTGCTTCATCTTCACGGTG-39, SVCT2 reverse: 59-GCAGGTTAGAGAGAGGCCAACAGCTG-39, flanking a 574-bp sequence; SVCT1 forward: 59GAAGAGGAGATCTACGGTAACTGG-39, SVCT1 reverse: 59-GGCAGCAGGAAGGTGAGGTTGCG-39, flanking a 313-bp sequence. Thirty cycles of PCR amplification were performed with 15 s at 94 8C, 60 s at 64 8C (SVCT2) and 55 8C (SVCT1), respectively, and 3 min at 72 8C. PCR product was analysed by 1.5% (w / v) gel electrophoresis and stained with ethidium bromide. To exclude amplification of contaminating genomic DNA, PCR was performed with untranscribed RNA in parallel. After purification, target DNA was cloned into pT-Adv vector by incubating 6 ml PCR product with 25 ng T4 DNA Ligase at 4 8C. Plasmids were transformed in TOP10F9 E. coli, selected by blue / white-screening on LB /Amp / IPTG / X-Gal-plates and isolated with Spin Miniprep Kit.

3. Results In earlier reports we described the successful establishment of an in vitro model of a the blood–CSF barrier which is based on primary cultured epithelial cells from porcine CP on permeable filter membranes [9,10]. Confluent monolayers show typical characteristics of secreting and transporting epithelia, like numerous microvilli on the apical surface, polar expression of Na 1 –K 1 -ATPase at the apical membrane, expression of tight junction proteins and high electrical resistance. Active transport properties like fluid transfer to the apical chamber, development of a proton gradient across the cell monolayer, transport of organic anions to the basolateral side and of nutrients like AA and MI to the apical side were also observed. In the present study in addition to vectorial transport we also addressed the efflux of AA out of the cytoplasm using [ 14 C] radioactive labelled substrate (Fig. 1). After 12 min 64 (63.9)% of the total amount of AA, that the cells were originally loaded with, were detected in the apical chamber compared to only 12 (62.0)% in the basolateral chamber. Thus, the direction of efflux agrees with the direction of vectorial transport across the cell monolayer as it has been described before [10]. These experiments only deal with

Fig. 1. Efflux of AA from epithelial cell monolayer into either the apical or the basolateral chamber after 12 min (n54) (in % of amount originally enriched in cell monolayer). Cells were preincubated for 1 h with 10 mM AA.

transport direction and cannot distinguish between active substrate uptake or release. It is well known that AA transport via SVCTs can be inhibited by phloretin [12,25,26]. Using the CP cell culture model we applied phloretin in concentrations of 20, 50 and 200 mM (Fig. 2) and examined its impact on AA transport characteristics. In the presence of 20 and 50 mM phloretin, AA transport was decreased to 66(67.1)% and to 28(63.9)%, respectively, adding 200 mM phloretin even a reduction to 13(62.6)% compared to the control filters was observed. Addressing if AA uptake or efflux is

Fig. 2. Relative transport across the cell monolayer (n56) and uptake of AA into the plexus epithelial cells (n53) in relation to control experiment. Values for transport and uptake in absence of phloretin were 1.1 nmol /(h cm 2 ) and 0.28 nmol / cm 2 , respectively. Cells were incubated for 2 h with 30 mM ascorbic acid. Phloretin was added in a concentration of 20, 50 and 200 mM.

S. Angelow et al. / Brain Research 988 (2003) 105–113

inhibited by phloretin, we determined intracellular substrate concentrations in parallel. When 20, 50 and 200 mM phloretin were applied a decrease of AA uptake to 81(61.1), 44(63.6) and 27(62.0)% compared to control was measured, supporting that uptake and not efflux is affected by phloretin. The absolute values of transport and uptake inhibition are different, because higher substrate uptake in the absence of an inhibitor leads to higher substrate diffusion to the apical side. Therefore, compared to transport inhibition the inhibitory effect on uptake is less significant. The observations lead to the hypothesis that in primary cultured CP cells an active transport system, supposedly located basolaterally, is responsible for AA uptake. As the AA specific transporter SVCT2 is supposed to mediate AA uptake into CSF in vivo the expression of SVCT2 was proven in cultured cells. RT-PCR with gene-specific primers and subsequent sequencing of the amplified DNAfragment with a length of 547 bp showed that SVCT2 is expressed in vitro (Fig. 3). Sequence analysis of two clones with BLAST (basic local alignment search tool) revealed 98 and 98.4% agreement with a transporter published as a porcine nucleobases transporter YSPL2 (yolk sac permease like). As comparison between human SVCT2 and human YSPL2 showed 100% agreement in DNA sequences, one can conclude that they are identical proteins. No signal of SVCT1-mRNA was detected in primary cultured CP cells, whereas RT-PCR using mRNA of porcine kidney led to amplification of a 313-bp DNA fragment of SVCT1 as identified with BLAST. We also investigated if transport of AA or MI in CP epithelial cells takes place in a Na 1 -dependent manner as it is reported for transport mechanisms of SVCT1, SVCT2 and SMIT. In order to prove Na 1 -dependent transport Na 1 -ions in the incubation buffer were substituted by choline (Na 1 -free medium). An enormous decrease of substrate transport was observed compared to control for both metabolites (Fig. 4a). In the control experiment a substrate transport across the cell monolayer of 0.21 (60.02) nmol /(h?cm 2 ) for AA and 0.43 (60.02) nmol /(h? cm 2 ) for MI was measured, whereas in Na 1 -free medium only transport in a range of 1–10% of control value for both nutrients was observed. Uptake of AA and MI was inhibited in the absence of Na 1 as well. Intracellular AA concentration was decreased by 80% from 75(61) to 15(61) pmol / cm 2 and MI concentration was decreased by 75% from 244(63) to 61(61) pmol / cm 2 (Fig. 4b). Na 1 free medium not only affected the active AA and MI transport but also the barrier properties of CP epithelial cells. In Na 1 -free medium TERs of the monolayers, measured by impedance analysis, were decreased to about 50% (data not shown) within several ours, showing that the cellular barrier was weakened but still existed. Values for the capacity, an indicator for cell membrane characteristics, were barely effected. In another experiment (Fig. 5) we focused on the

109

Fig. 3. (a) RT-PCR was used to prove the presence of SVCT2 mRNA in cultured CP cells. The expected product was 547 bp (lane 1). Control without reverse transcriptase was completely negative (not shown). Lane 2: 100 bp DNA ladder. (b) RT-PCR using SVCT1 specific primer. No signal was detected with mRNA of cultured CP cells (lane 1). Lane 2: positive control with mRNA from porcine kidney shows the expected product with a length of 313 bp. Control without reverse transcriptase was completely negative (not shown). Lane 3: 100 bp DNA ladder.

dependency of AA transport on a Na 1 gradient across the cell membrane since the driving force for AA uptake is a low intracellular Na 1 -concentration maintained by Na 1 – K 1 -ATPase. In the presence of ouabain (Fig. 5b), an inhibitor of Na 1 –K 1 -ATPase, uptake of AA into the cell monolayer was decreased to about 30(63)% of the control value (Fig. 5a). AA uptake is not only stimulated by Na 1 –K 1 -ATPase but also by an artificially created Na 1 gradient. First, Na 1 –K 1 -ATPase was deactivated by ouabain and cells were preincubated with Na 1 -free medium to minimise intracellular Na 1 -concentration (Fig. 5c,d). When AA was applied in Na 1 -free medium (c) again an

110

S. Angelow et al. / Brain Research 988 (2003) 105–113

Fig. 5. Uptake of AA into cultured CP epithelial cells (n52). Cells were preincubated for 30 min in the absence of AA and then incubated 20 min in the presence of 30 mM AA with the following solutions: (a) preincubation and incubation with glucose buffer, (b) preincubation and incubation with glucose buffer containing 1 mM ouabain, (c) preincubation and incubation with glucose buffer containing 1 mM ouabain without Na 1 (substituted by choline), (d) preincubation with glucose buffer containing 1 mM ouabain without Na 1 (substituted by choline), incubation with glucose buffer containing 1 mM ouabain.

Fig. 4. AA and MI transport across the cell monolayer into the apical chamber (n54) (a) and uptake into the CP cells (n52) (b). Cells were incubated for 2 h with 30 mM AA or 30 mM MI, dissolved in glucose buffer and glucose buffer without Na 1 (substituted by choline).

increased uptake of about 30(61)% of control was observed. In contrast, when AA was added in Na 1 -containing medium (d), the cytoplasmatic uptake was almost 2-fold higher (54(64)% of control). Under the latter condition, extracellular Na 1 concentration is much higher than the intracellular one and AA enters the cells coupled to an influx of Na 1 . In order to underline that uptake of AA and MI is only dependent on Na 1 but not on Cl 2 we repeated the uptake experiments in Cl 2 -free medium (Fig. 6). When Cl 2 was substituted by iodide or D-gluconate uptake of AA was 3.7 and 2.5 times higher compared to control. Similar observations were made with MI uptake that was about 1.4 and 1.5 times higher under Cl 2 -free conditions. These results indicate that absence of Cl 2 increases the activity of AA and MI transporters independently on the anion applied to

substitute Cl 2 . These observations can be explained as follows: a lack of Cl 2 may result in inhibition of cotransport-systems that are dependent on Na 1 and Cl 2 . Therefore intracellular Na 1 -concentration is decreased compared to normal conditions. Low Na 1 -concentration may be adjusted by Na 1 influx mediated by other transport systems like SVCT2 or SMIT. It is well-known, that diuretics like bumetanide inhibit Na 1 -Cl 2 -cotransport. Therefore we tested, if inhibition of Na 1 -Cl 2 -cotransport by bumetanide also induces an increased AA and MI uptake by increased cotransport with Na 1 . Consistent with the results obtained in Cl 2 -free medium, uptake of both AA and MI was also increased in the presence of 100 mM bumetanide (Fig. 7) and about 2-fold higher compared to control.

4. Discussion In previous studies, we reported on the active transport properties of cultured CP cells grown on permeable membranes [9,10]. Under serum-free conditions monolayers generate distinctive cell polarity and low intercellular permeability. Showing that transport direction and kinetics of organic anions like penicillin G and nutrients like AA and MI agree with results achieved from tissue measurements, we concluded that monolayers of the CP epithelial cells, established as a hydrodynamic barrier between two fluid compartments, are well suited to investigate the CP transport systems in more detail. In this study we analysed the transport of AA crossing the blood–CSF barrier. Since expression of SVCT2 in CP

S. Angelow et al. / Brain Research 988 (2003) 105–113

111

Fig. 7. AA and MI uptake into cultured CP epithelial cells (n52). Cells were incubated for 20 min with 30 mM AA or 30 mM MI, dissolved in glucose buffer and glucose buffer with 100 mM bumetanide.

present study, transport was inhibited to 50% by phloretin in a range of 20–50 mM. In the presence of higher phloretin concentrations, transport and uptake were reduced to 13 and 27% of control, respectively, which is similar to the results of Kannan et al. [12]. Additionally when suppressing secondary active processes by inhibition of Na 1 –K 1 -ATPase with ouabain both a decreased transport and uptake of AA and MI were observed. Assuming that phloretin and ouabain would inhibit substrate efflux from the cytoplasm, an increased or at least constant intracellular AA and MI concentration would be expected, as illustrated in Fig. 8 The results show that uptake of Fig. 6. Uptake of AA and MI into cultured CP epithelial cells (n52). Cells were incubated for 2 h with 30 mM AA or 30 mM MI, dissolved in glucose buffer and glucose buffer without Cl 2 , substituted by (a) iodide or (b) D-gluconate.

epithelium has been demonstrated before [25] we investigated if AA transport in vitro agrees with the transport characteristics that account for SVCT2. As reported before AA is actively transported in a concentration-dependent manner to the apical side following Michaelis–Menten kinetics. Here we have reported, that direction of substrate efflux agrees with the direction of transepithelial transport. By inhibition experiments we also received information about transport mechanism and location. Application of phloretin caused a concentration-dependent decrease of AA transport and uptake by CP epithelial cells. Phloretin, an inhibitor of facilitated glucose transport, was also reported by Tsukaguchi et al. [25] to inhibit SVCT1 expressed in Xenopus oocytes with a Ki of about 65 mM. Kannan et al. [12] showed that 100 mM phloretin inhibited AA uptake of SVCT2 into human lens epithelial cells by 80%. In the

Fig. 8. Schematic view of the CP epithelium. The arrows represent AA flux across the monolayer. On the right side the effects on transepithelial substrate transport and uptake into the monolayer are demonstrated in case of inhibition of basolateral uptake (a) or apical efflux (b).

112

S. Angelow et al. / Brain Research 988 (2003) 105–113

these nutrients at the basolateral membrane and not the release into the apical compartment (CSF side) is a transporter mediated active process, in all probability catalysed by SVCT2. Expression of SVCT2 in cultured CP cells was proven on mRNA level by RT-PCR and sequence analysis, whereas detection of SVCT1-mRNA failed. The presence of SVCT1 in CP epithelium seems to be very unlikely, as no expression of SVCT1 was detected in human brain [26] and the two SVCT isoforms are alternatively expressed in different organs. The next set of experiments was performed in order to demonstrate the Na 1 coupled influx of AA (not including the stoichiometry) that is expected for a SVCT2 mediated transport. The enormous decrease of active AA and MI transport in Na 1 -free medium indicates a Na 1 -dependent process. As substitution of Na 1 also increases tight junctions permeability, this observations could be due to a higher back diffusion of substrate through the intercellular space. Na 1 -dependency was, however, underlined by a decreased intracellular uptake that is not influenced by barrier properties. Besides inhibition of the AA-transport by phloretin, AA uptake was reduced to 20% in Na 1 -free medium (and with shorter incubation time to 30%). The Na 1 -coupled influx of AA into the cell monolayer was additionally demonstrated, when the ratio of extracellular and intracellular Na 1 -concentration was artificially increased (excluding Na 1 –K 1 -ATPase activity). However, the pathways responsible for the remaining AA uptake under Na 1 free conditions have still to be clarified. Using SVCT1 transfected HRPE cells Wang et al. [26] did not detect significant changes in AA uptake in Cl 2 -free medium. In contrast, primary cultured epithelial cell of the CP show an increased AA uptake. For CSF secretion into the ventricles CP epithelium maintains a continuous net transport of NaCl to the apical side. The Na 1 flux across the cell monolayer is ensured by several transport proteins. Among others the involvement of the bumetanide sensitive Na 1 –K 1 -Cl 2 -cotransporter NKCC1 (BSC2) in CSF secretion is discussed. Plotkin et al. [14] reported the apical localisation of NKCC1 in CP epithelium, that is only compatible with its role in fluid secretion, when ions are actively transported out of the cells into CSF (efflux). Speake [19] assumes that NKCC1 can mediate both, ion influx and efflux, depending on experimental conditions. In the absence of Cl 2 , the cotransport pathway for Na 1 is inhibited and abnormal low intracellular Na 1 -concentration could be balanced by increased AA-Na 1 -cotransport. Similar observations with MI underline, that this hypothesis cannot only be applied to AA-transport. Application of bumetanide also induced higher AA and MI uptake, indicating that bumetanide also decreases intracellular Na 1 -concentration, possibly due to inhibition of Na 1 -Cl 2 cotransport. The repressive effect of bumetanide on Na 1 Cl 2 -cotransport was demonstrated before [2,11]. In contrast to our results Wilson and Dixon [28] observed decreased AA uptake in primary cultured astrocytes in the

presents of the diuretic furosemide. As substitution of Cl 2 showed little and inconsistent effects in their experiments, they conclude that furosemide directly inhibits Na 1 -AAcotransport, independently of Cl 2 gradients. In summary, using a cell culture model of CP epithelium, we investigated the AA transport across blood– CSF barrier and conclude that it is supposedly mediated by SVCT2. By transport and uptake studies, we demonstrated the basolateral localisation of the transporter and its Na 1 dependent transport mechanism. Additionally we showed the stimulation of AA uptake when Na 1 -Cl 2 -cotransport is inhibited. This clearly shows that cultured CP epithelial cells are well suited to study the kinetics and the underlying mechanisms of transport processes at the blood–CSF barrier.

Acknowledgements This work has been financially supported by the Fonds der chemischen Industrie.

References [1] D.B. Agus, S.S. Gambhir, W.M. Pardridge, C. Spielholz, J. Baselga, J.C. Vera, D.W. Golde, Vitamin C crosses the blood–brain barrier in the oxidized form through the glucose transporters, J. Clin. Invest. 100 (1997) 2842–2848. [2] D. Bairamian, C.E. Johanson, J.T. Parmelee, M.H. Epstein, Potassium cotransport with sodium and chloride in the choroid plexus, J. Neurochem. 56 (1991) 1623–1629. [3] R. Daruwala, J. Song, W.S. Koh, S.C. Rumsey, M. Levine, Cloning and functional characterization of the human sodium-dependent vitamin C transporters hSVCT1 and hSVCT2, FEBS Lett. 460 (1999) 480–484. [4] K.R. Dhariwal, W.O. Hartzell, M. Levine, Ascorbic acid and dehydroascorbic acid measurements in human plasma and serum, Am. J. Clin. Nutr. 54 (1991) 712–716. [5] E.J. Diliberto Jr., F.S. Menniti, J. Knoth, A.J. Daniels, J.S. Kizer, O.H. Viveros, Adrenomedullary chromaffin cells as a model to study the neurobiology of ascorbic acid: from monooxygenation to neuromodulation, Ann. NY Acad. Sci. 498 (1987) 28–53. [6] C.F. Eldridge, M.B. Bunge, R.P. Bunge, P.M. Wood, Differentiation of axon-related Schwann cells in vitro. I. Ascorbic acid regulates basal lamina assembly and myelin formation, J. Cell Biol. 105 (1987) 1023–1034. [7] U. Gath, A. Hakvoort, J. Wegener, S. Decker, H.J. Galla, Porcine choroid plexus cells in culture: expression of polarized phenotype, maintenance of barrier properties and apical secretion of CSFcomponents, Eur. J. Cell Biol. 74 (1997) 68–78. [8] C.C. Glembotski, The role of ascorbic acid in the biosynthesis of the neuroendocrine peptides alpha-MSH and TRH, Ann. NY Acad. Sci. 498 (1987) 54–62. [9] A. Hakvoort, M. Haselbach, H.J. Galla, Active transport properties of porcine choroid plexus cells in culture, Brain Res. 795 (1998) 247–256. [10] A. Hakvoort, M. Haselbach, J. Wegener, D. Hoheisel, H.J. Galla, The polarity of choroid plexus epithelial cells in vitro is improved in serum-free medium, J. Neurochem. 71 (1998) 1141–1150. [11] C.E. Johanson, S.M. Sweeney, J.T. Parmelee, M.H. Epstein, Co-

S. Angelow et al. / Brain Research 988 (2003) 105–113

[12]

[13]

[14]

[15]

[16]

[17]

[18] [19]

transport of sodium and chloride by the adult mammalian choroid plexus, Am. J. Physiol. 258 (1990) C211–216. R. Kannan, A. Stolz, Q. Ji, P.D. Prasad, V. Ganapathy, Vitamin C transport in human lens epithelial cells: evidence for the presence of SVCT2, Exp. Eye Res. 73 (2001) 159–165. H.M. Kwon, A. Yamauchi, S. Uchida, A.S. Preston, A. GarciaPerez, M.B. Burg, J.S. Handler, Cloning of the cDNa for a Na1 / myo-inositol cotransporter, a hypertonicity stress protein, J. Biol. Chem. 267 (1992) 6297–6301. M.D. Plotkin, M.R. Kaplan, L.N. Peterson, S.R. Gullans, S.C. Hebert, E. Delpire, Expression of the Na(1)-K(1)-2Cl-cotransporter BSC2 in the nervous system, Am. J. Physiol. 272 (1997) C173–183. F. Porcellati, T. Hlaing, M. Togawa, M.J. Stevens, D.D. Larkin, Y. Hosaka, T.W. Glover, D.N. Henry, D.A. Greene, P.D. Killen, Human Na(1)-myo-inositol cotransporter gene: alternate splicing generates diverse transcripts, Am. J. Physiol. 274 (1998) C1215–1225. D.P. Rajan, W. Huang, B. Dutta, L.D. Devoe, F.H. Leibach, V. Ganapathy, P.D. Prasad, Human placental sodium-dependent vitamin C transporter (SVCT2): molecular cloning and transport function, Biochem. Biophys. Res. Commun. 262 (1999) 762–768. H. Reiber, M. Ruff, M. Uhr, Ascorbate concentration in human cerebrospinal fluid (CSF) and serum. Intrathecal accumulation and CSF flow rate, Clin. Chim. Acta 217 (1993) 163–173. M.E. Rice, Ascorbate regulation and its neuroprotective role in the brain, Trends Neurosci. 23 (2000) 209–216. T. Speake, C. Whitwell, H. Kajita, A. Majid, P.D. Brown, Mechanisms of CSF secretion by the choroid plexus, Microsc. Res. Tech. 52 (2001) 49–59.

113

[20] R. Spector, Drug transport in the mammalian central nervous system: multiple complex systems. A critical analysis and commentary, Pharmacology 60 (2000) 58–73. [21] R. Spector, A.V. Lorenzo, Myo-inositol transport in the central nervous system, Am. J. Physiol. 228 (1975) 1510–1518. [22] R. Spector, A.V. Lorenzo, The origin of myo-inositol in brain, cerebrospinal fluid and choroid plexus, J. Neurochem. 25 (1975) 353–354. [23] R. Spector, A.V. Lorenzo, Ascorbic acid homeostasis in the central nervous system, Am. J. Physiol. 225 (1973) 757–763. [24] C.M. Tallaksen, T. Bohmer, H. Bell, Concentrations of the watersoluble vitamins thiamin, ascorbic acid, and folic acid in serum and cerebrospinal fluid of healthy individuals, Am. J. Clin. Nutr. 56 (1992) 559–564. [25] H. Tsukaguchi, T. Tokui, B. Mackenzie, U.V. Berger, X.Z. Chen, Y. Wang, R.F. Brubaker, M.A. Hediger, A family of mammalian Na1-dependent L-ascorbic acid transporters, Nature 399 (1999) 70–75. [26] Y. Wang, B. Mackenzie, H. Tsukaguchi, S. Weremowicz, C.C. Morton, M.A. Hediger, Human vitamin C ( L-ascorbic acid) transporter SVCT1, Biochem. Biophys. Res. Commun. 267 (2000) 488–494. [27] J. Wegener, A. Hakvoort, H.J. Galla, Barrier function of porcine choroid plexus epithelial cells is modulated by cAMP-dependent pathways in vitro, Brain Res. 853 (2000) 115–124. [28] J.X. Wilson, S.J. Dixon, Ascorbic acid transport in mouse and rat astrocytes is reversibly inhibited by furosemide, SITS, and DIDS, Neurochem. Res. 14 (1989) 1169–1175.