Brain Research, 539 (1991) 247-253 Elsevier
247
BRES 16259
Glial cells and neurons induce blood-brain barrier related enzymes in cultured cerebral endothelial cells Ulrike Tontsch* and Hans-Christian Bauer lnstitut far Molekularbiologie der Osterreichischen Akademie der Wissenschaften, Salzburg (Austria) (Accepted 14 August 1990)
Key words: Blood-brain barrier; Cerebral endothelial cell; Glial cell; Enzyme induction; y-Glutamyltranspeptidase; ATPase, (Na+,K+)-
The blood-brain barrier (BBB) in mammals is created and maintained by cerebral endothelial cells (cEC) that express specialized functional properties, including intercellular tight junctions, absence of fenestrae and specific membrane transport systems. It has been proposed that the differentiation of these characteristics, acquired during brain development, is controlled by the neural environmentls'24. Co-culture experiments of cloned cEC with astroglial cells, C6 glioma cells and cortical neurons, with plasma membranes or conditioned media of these cells, were used to study induction of some BBB characteristics in vitro. Activities of Na+,K+-ATPase and y-glutamyltranspeptidase (GGTP), an enzyme responsible for amino acid transport across the BBB, were taken as parameters for BBB function. Co-culture of cEC with C6 glioma cells caused a two-foldincrease in GGTP activity and this activitywas likewise amplified by incubation with plasma membrane fractions derived from C6 glioma cells, embryonic brain cells and cortical neurons; conditioned media (soluble factors) had no effect. Na+,K+-ATPase activity, estimated from the ouabain inhibitable fraction of SrRb uptake, was increased by about 90% in cEC incubated with C6 glioma plasma membranes. We propose from these data that both neurons and glial cells confer BBB characteristics on cEC via cell-cell contact.
INTRODUCTION Although structure and function of the blood-brain barrier (BBB), maintained primarily by cerebral capillary endothelial cells (cEC), have been studied extensively8, 9,22, it still remains unclear, which factors are responsible for the establishment of this barrier during embryonic development. Transplantation experiments by Stewart and Wiley24 have provided strong evidence that BBB characteristics of cEC depend on their neural environment. Cerebral capillaries are tightly ensheathed by astrocytes and the tightness of the BBB correlates with the completeness of this glial cell layer in the course of brain development 32 as well as in pathological processes 16. Therefore these cells have been considered responsible for the induction of BBB features in cEC. Investigations of cerebral vascularization and BBB formation have focused on two aspects: the tightness of the B B B and the selective transport mechanisms of cEC. Factors which increase or decrease the number of tight junctions and fenestrae have been studied in vivo and in vitro 1AS"23-25"2s'32. The selective transport mechanisms across the BBB were monitored with alkaline phosphatase, Na+,K+-ATPase, Ca2+.ATPase, 5"-nucleotidase
and
y-glutamyl
transpeptidase
(GGTP) 7'11'16'1a'29'30.
GGTP, which has been described as a typical BBBrelated enzyme 21, is located in the luminal as well as in the abluminal membrane of cEC and is supposed to be involved in the transport of large neutral amino acids across the BBB. The relatively high GGTP activity of cEC in vivo markedly decreases when these cells are kept in culture. The first attempts to mimick the in vivo situation were made by DeBault and Cancilla 11 who found by histochemical staining that co-cultures of cEC with C6 glioma cells increased GGTP activity. After similar co-culture experiments, Beck et al. 5 described enhanced Na+,K*-ATPase activity, which is also a feature of BBB function. Until now, only data from histochemical studies were available for Na ÷,K+-ATPase activity in cEC. It has generally been accepted that glial cells possess BBB-inductive potential. Both soluble proteins released from glial cells as well as membrane bound structures are believed to act on cEC plasma membranes in a way which leads to BBB formation LH,15,19,25. In this paper we have exploited and extended the co-culture approach to test the influence of other cells and cell fractions on the cEC. The results, reported here, indicate that membrane
* Present address: Max-Planck-Institut for Psychiatrie, Abteilung Neuroimmunologie, Martinsried, ER.G. Correspondence: H.-C. Bauer, Institut fOr Molekularbiologie der Osterreiehischen Akademie der Wissenschaften, Billrothstr. 11, A-5020 Salzburg, Austria. 0006-8993/91/$03.50 (~) 1991 Elsevier Science Publishers B.V. (Biomedical Division)
248 i n t e r a c t i o n s b e t w e e n c E C a n d n e u r a l cells are r e s p o n s i b l e f o r i n c r e a s e d activity o f B B B - r e l a t e d e n z y m e s , w h e r e a s soluble factors are ineffective.
MATERIALS AND METHODS
Cell culture Porcine cEC were isolated, cultivated and characterized as described previously 27. Briefly, cerebral capillaries were released from the surrounding neural tissue after several steps of homogenization and centrifugation. After dissociation in collagenase, the cells were plated on 35-mm diameter plastic petri dishes (TC quality, NUNC, Denmark), and kept in medium 199 (Seromed, F.R.G.), supplemented with 10% heat-inactivated foetal calf serum (FCS) (Seromed), 100/~g/ml heparin (Sigma, ER.G.), 20/~g/ml endothelial cell growth supplement (ECGS) (Collaborative Res.) and antibiotics. CEC were characterized with the lectin of Bandeiraea simplicifolia ~7. For our studies we used cloned cEC (designated A9B12) of the cobble-stone type, at passages between 15 and 20, always at confluence. Cells were plated on 35-mm diameter plastic petri dishes (NUNC) or in 24-well plates (Costar, Cambridge, MA). Primary cultures of murine astrocytes were obtained as described previously4, kept in Dulbecco's modified Eagle's medium with 10% FCS, characterized with glial fibrillary acidic protein (GFAP) and used after 2 weeks in culture. Porcine aortic EC (aEC) were isolated by collagenase treatment using the procedure of Jaffe TM and were kept under the same culture conditions as the cEC. Primary cultures of murine cortical neurons were established as described 4 and used after 3 days in culture. Y-1 mouse adrenal tumor cells and CCRF-S180 mouse sarcoma cells were obtained from the American Tissue Culture Collection and were both maintained in Ham's F12 medium, supplemented with 10% FCS. C6 rat glioma cells, a generous gift of Dr. B. Hamprecht, Univ. T~ibingen, were also cultivated in Ham's F12 with 10% FCS. Chicken fibroblasts were isolated and cultivated according to Freshney ~2.
Co-culture CEC (A9B12) were co-cultured with conditioned media, cells or plasma membranes of various cell types as follows. (1) cEC grown on 35-mm diameter.dishes were incubated with C6 conditioned medium (C6CM) or with glia conditioned medium (GCM) for 24, 48 and 72 h. Conditioned medium was obtained as follows: confluent cells were washed with PBS and supplied with fresh medium 199 + 10% FCS. After 72 h the medium was collected, centrifuged to remove cell debris, filter sterilized and used immediately, (2) cEC in 35-mm diameter dishes were co-cultivated with C6 glioma cells, primary glial cells, primary neurons, fibroblasts, Y-1 cells and CCRF cells, growing on glass coverslips (20 x 20 mm). These coverslips were placed upside down on the layer of cEC and remained there for 24, 48 and 72 h. (3) cEC in 35-mm diameter dishes and in 24-well plates were incubated with isolated plasma membranes (PM) of C6 glioma cells and glial cells for 24, 48 and 72 h. PMs were also derived from primary cultures of neurons, fibroblasts, Y-1 adrenal tumor cells, CCRF sarcoma cells and from murine brain (embryonic day 14). Protein concentration of the PMs ranged from 2.5 to 200/~g/ml medium 199 + 10% FCS and incubation lasted for 48 h. CEC reference cultures were also supplied with medium 199 + FCS. Isolation of PM fractions was carried out as described previously 3. Briefly, confluently growing cells were trypsinized, centrifuged, and the pellet forcefully homogenized with a tightly fitting pestle in isotonic saccharose-phosphate buffer (0.32 M, pH 7.6). In the first centrifugation step (800 g, 15 min, 4 °C), cell nuclei, and in the second (11,500 g, 25 min, 4 °C), mitochondria were removed from the homogenate. This supernatant was subsequently centrifuged (100,000 g, 45 min, 4 °C), the pellet resuspended in 5 mM
Tris-HCl/1 mM EDTA, pH 8.1 and centrifuged again (11,500 g, 25 min, 4 °C). This final pellet, containing almost exclusively PMs, was dissolved in sterile phosphate-buffered saline (PBS) and used for co-culture experiments. PMs from cells of the embryonic brain cortex were isolated from about 100 murine brains (Balb/c). They were cleared from the meninges, minced, homogenized and processed as above. (4) aEC were cocultivated with the PM fraction of C6 glioma cells.
Determination of enzyme activities" GGTP. Conditioned media, coverslips, and PM fractions were removed from the cEC cultures. Subsequently, the cEC were washed 3 times with PBS and processed according to Maxwell et al. 19. Briefly, the cells were incubated overnight in 0.1 M phosphate buffer (pH 7.4), and the cell suspension sonicated for 10 s at 20 W (Braun Labsonic 1510). GGTP activity was determined with a colorimetric assay, based essentially on the method of Naftalin et al. 2°. According to this technique, 7-c-glutamyl-p-nitroanilide serves as the substrate and glycylglycine as the acceptor molecule for the glutamyl group released by the action of GGTP. Liberated pnitroaniline is diazotized and measured at 540 nm. The absorbance is proportional to GGTP activity, which is related to protein concentration, measured by the method of Lowry et a1.18 Reagents for GGTP determination were all purchased from Sigma. Na+,K+-ATPase. Na+,K+-ATPase activity was determined by the use of 86Rb influx. As shown by Vaughan and Cook 28, K + can be fully substituted by Rb. The Na+,K+-ATPase activity was estimated as the ouabain inhibitable part of total 86Rb influx 13. Cloned cEC were cultured in 12-well multiwells (Costar) for 5 days to confluency. Prior to use, cells were washed twice with a buffered salt solution (containing the following (mM): NaCI 120, KCI 5, CaCl 2 2, MgCI 2 1, HEPES 20; pH 7.4 with Tris) and were preincubated with 1 mM ouabain (Sigma) for 20 min at 37 °C. K + influx was measured at room temperature with the addition of approximately 1 #Ci 86RbCI (Amersham, U.K.) in water. For most experiments influx was followed for 15 rnin. The incubation medium was then removed and the cells rapidly washed 3 times with ice-cold isotonic NaCI. Intracellular 86Rb was extracted with 0.5% Triton X-100 overnight, scintillation liquid (Emulsifier, Packard) was added and radioactivity determined in a Packard fl-scintillation counter. Protein was measured by the method of Lowry et al. TM and 86Rb (K +) influx was calculated as cpm/min/mg protein.
Characterization of PM fractions Lactate dehydrogenase was used to determine contamination with cytoplasmic proteins in the PM preparations. The enzyme activity was measured using the method of Bergmeyer 6. The presence of mitochondrial membranes was identified by determination of monoamine oxidase activity, following the procedure of Wurtman and Axelrod 31. Electron micrographs of PMs were produced as described previously 2. RESULTS
Cell cultures The described isolation procedure, culture conditions a n d in p a r t i c u l a r t h e cell c l o n i n g left u s w i t h a p u r e p o p u l a t i o n o f c e r e b r a l e n d o t h e l i a l cells ( c l o n e A 9 B 1 2 ) . T h e cells c o u l d b e p a s s a g e d u p t o 35 t i m e s w i t h o u t l o s i n g t h e i r c o b b l e s t o n e p h e n o t y p e (Fig. l a ) o r t h e e x p r e s s i o n of t h e b i n d i n g sites f o r t h e m a r k e r Bandeiraea simplicifolia lectin. B r i g h t f l u o r e s c e n c e l a b e l i n g w i t h this l e c t i n is s h o w n in Fig. l b . C o r t i c a l n e u r o n s , c u l t i v a t e d f o r 3 d a y s a r e s h o w n in Fig. l c . A s t r o c y t e s w e r e c h a r a c t e r i z e d b y t h e p o s s e s s i o n of glial f i b r i l l a r y acidic p r o t e i n (Fig. l d ) .
249
Fig. 1. a: phase contrast micrograph of cloned porcine cEC, passage 18, grown in medium 199 + 10% FCS + heparin + ECGS. Bar, 100 /~m. b: cEC were labeled with TRITC-conjugated lectin of Bandeiraea simplicifolia. Bar b-d, 50 9m. c: phase contrast micrograph of murine cortical neurons cultivated for 3 days. d: fluorescent micrograph of routine astrocytes stained for GFAP.
G G T P activity - coculture with cells, conditioned media and PMs Cocuiture experiments were carried out for 24, 48 and 72 h. In freshly isolated capillaries G G T P activity was 800 U/mg protein, in cultivated cEC G G T P activity decreased to about 200 U/mg protein. If cEC were cocultivated with C6 glioma cells, enzyme activity increased up to two-fold over untreated cEC. With primary cultures of glia cells, neurons, fibroblasts or the Y-1 tumor cells and CCRF, no enzyme induction was noticed (data not shown). Neither did conditioned medium from C6 glioma cells nor from glial cells affect G G T P activity in cEC. Incubation of cEC with the PM fraction from C6 increased G G T P activity from 180 + 40 U in untreated cEC of the identical passage to 1296 + 109 U (Fig. 2). Appreciable enzyme activity was already observed after 24 h, and increased after incubation for 48 and 72 h. In the following experiments incubations were performed for 48 h. cEC were subjected to isolated PMs in concentrations ranging from 2.5 to 200/~g protein/ml and showed a dose-dependent increase of enzyme activity (Fig. 3). We also incubated cEC with PMs of Y-1 adrenal
tumor cells and of CCRF sarcoma cells and could show that enzyme induction was not tumor cell dependent. As C6 glioma cells themselves express relatively high G G T P
1800 1400 1200 =
1000 800
"~ 600 400 800 0 ~
~ cap
cEC
J + 06
+ C6 CH
+ PM 06
Fig. 2. Determination of G G T P activity,as described in Materials and Methods, of freshly isolated cerebral capillaries(cap), cultivated cEC, and of cEC after co-culture with C6 gfiorna cells,with conditioned medium of C.6 (06 C M ) and with plasma membranes (PM C6). Protein concentrationof the P M ~acfion was 200/~g/mJ. Co-cul~m~ was carried out for 24, 48 and 72 h. Means are given and S.D.
250 activity (972 + 43 U G G T P / m g protein in C6 glioma cells vs 125 + 23 in glial cells), we investigated the possibility that the increase of enzyme activity was due to unspecific attachment of PM C6. This could be excluded as fibroblast cultures, subjected to PM C6 caused no increase in G G T P activity, aEC were also incubated with PM C6 and were found completely unaffected (Table I). Incubation with PM of normal glia cells did not lead to pronounced G G T P induction, neither in cEC nor in aEC. To test whether other components of the brain, apart
3000 2500
TABLE I Confluent cultures of aEC incubated with PM C6 in concentrations indicated
Data are means and S.D. of 4 measurements and were analyzed by Student's t-test. aEC Incubated with
U GG TP/mg protein
PM PM PM PM PM
226 +_55 216 + 53 168 + 18 215 + 82 167 + 31 289 + 103 n.s.
C6 C6 C6 C6 C6
(10#g/ml) (20#g/ml) (50#g/ml) (100/zg/ml) (200/~g/ml)
2000 15oo
t
1000
500
0 0
50
20
10
100
200 ~ g / m l
3000Ib 2500 2000
1500i
i
i000[
Characterization o f P M fractions
The PM fractions were defined by exclusion of marker enzymes, monoamine oxidase for mitochondrial mem-
500
o,i, 0
2000F
from the astrocytes, which are not present at early stages of embryonic development, are involved in inducing BBB features in cEC, we subjected cEC to a PM fraction of embryonic brain cells. G G T P activity of this fraction was 232 U G G T P / m g protein. Notably, our data showed a dose-dependent, up to 7-fold enzyme induction in the cEC (Fig. 3). We reasoned that neurons were responsible for this effect. To prove this assumption, we isolated PMs from primary cultures of neurons ( G G T P activity 3630 U/mg protein) and carried out co-culture experiments. As expected, there was a reasonable increase in G G T P activity and suprisingly this PM fraction proved to be an equally potent inducer as PM C6 (Fig. 3).
.
.
10 .
.
.
.
20 .
.
.
.
50 pg/ml .
.
.
.
.
.
.
.
.
1500
o
1000
500 1
0 ~
0
2,5
5
10
20
50
~00
200 ~ag/ml
Fig. 3. Determination of GGTP activity, a: in cEC after incubation (48 h) with C6 glioma PM fraction. Protein concentration of PMs ranged from 10 to 200 #g/ml. Data are given as means and S.D. of 3--7 measurements, b: in cEC after co-culture with neuronal PM fraction. Protein concentration of PMs was 10, 20 and 50 #g/ml. Values give means and S.D. of 5 experiments, c: in cEC after co-culture with PMs from embryonic brain cells in concentrations ranging from 2.5 to 200 #g/ml. Data represent means and S.D., n = 7-11.
Fig. 4. Electron micrograph of a plasma membrane fraction from C6 glioma cells, x38,100.
251 .~
48 h) did not have any effect on Na +,K+-ATPase activity in cEC (data not shown).
Thoussnd= 80
t
•
o
60 ...................................................
DISCUSSION
J u
40
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
T
o
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
~
'
'
~
'
to
20
30
40
50 mia
Fig. 5. Total initial S~Rb (K +) uptake rate in cEC. After passage,
cells were kept 5 days in culture and were nearly confluent. Before incubation with mRb (K+), cells were carefully washed with PBS. n = 4, means and S.D. are given.
branes and lactate dehydrogenase for cytoplasmic constituents. Neither monoamine oxidase nor lactate dehydrogenase activity was found, indicating that there were no mitochondrial contaminations and no trapping of cytoplasm. The purity of the PM fractions was monitored by electron microscopy (Fig. 4). Induction o f Na+,K+-ATPase 86Rb (K t ) influx in cultured cEC was linear at least over the first 45 min (Fig. 5). For all co-culture experiments 86Rb (K t ) uptake was followed for 15 min in order to avoid possible cell damage during longer incubation periods. Treatment of cEC with PM (200 pg/ml) from C6 glioma cells for 48 h resulted in an increase in Na t , K t - A T P a s e activity. Table II shows that the ouabain inhibitable part of 86Rb (K t ) influx was 53% of total influx in PM treated cEC versus 28% in control cells. This corresponds to an increase of Na t , K t - A T P a s e activity of about 90% in cEC incubated with PM. Incubation with PM from cultured fibroblasts (200 pg/ml;
TABLE II Total influx ofaORb (K +) and ouabain inhibitable S6Rb (K +) influx in cultured cerebral endothelial cells (cEC) with or without plasma membranes of C6 glioma cells (PM C6)
After passage, cEC were kept 5 days in culture and were nearly confluent. PMs were incubated for 2 days. Cells were washed carefully with Dulbecco's PBS before use. a6Rb (K+) uptake was 15 min at room temperature. Details are given in Materials and Methods. Data represent 2-4 experiments, means and S.D. are given.
Cells
Total S6Rb (K +) influx (cpm/min/mg protein)
Ouabain inhibitable influx (Na +,K+.ATPase activity) (cpm/min/mg protein)
cEC cEC + PMC6
5087 + 83 (100%) 4951+210(100%)
1439 + 212 (28%) 2610+381(53%)
Our experiments have shown that glioma cells or their plasma membranes and plasma membranes from neuronal cells induce BBB-related enzyme activities in cultured cEC. Although it is generally assumed that BBB development is under environmental c o n t r o l z ° ' 1 5 ' 2 4 ' 3 2 , it is still unclear whether the effect is due to diffusable factors or to cell-cell contacts. Glial cells are usually considered as the main candidate for the induction of BBB properties in cEC 1'5'7'11'15'19"25'26. Our experiments have confirmed the influence of glial cells on cEC, and in addition we found that neurons play a significant role, too. First hints for the importance of neurons for the BBB development, which were based on light and electron microscopic studies, came from Stewart and Wiley 24 and from Yoshida et a l . 32. We chose two enzyme systems, G G T P and Na +,K+-ATPase to monitor the effect of glial cells and neurons on cerebral endothelial cells. Co-culture of cEC with C6 glioma cells resulted in reasonable increase in G G T P activity, but never as much as attained after incubation with PMs. The reason for this discrepancy may be technical in origin: (1) gas and medium exchange was not optimal and sometimes even caused the cells to detach; and (2) only about half of the cells in the culture dish could be brought into contact with the cells on the coverslip, which in turn reduced the effect of G G T P induction. However, all these problems could be readily overcome by incubating with PMs. These standardized conditions were a necessity for reproducible results and removed the uncertainties associated with variable cell density and growth state of the cells added. In addition, it was possible to test for the effect of membrane contact, avoiding interference with soluble factors. Cerebral endothelial cells responded to PM incubation with an increase in enzyme activity up to about 2000 U/mg protein and revealed only minor differences for the various PM fractions (PM C6, PM N, PM brain). This compares with an enzyme activity of about 800 U/rag protein in freshly isolated capillaries. Bearing in mind that the amount of protein not only refers to cEC, but includes that of erythrocytes, lymphocytes, pericytes and perhaps some smooth muscle cells, the specific enzyme activity for cEC can be assumed at least twice as high. Our data further show that for enzyme activity induction in cEC G G T P activity of the added cells or their plasma membranes is irrele"ant. With the second enzyme system, the Na+,K+-ATPase, we also demonstrated an increase in activity after co-culture of cEC with PMs of C6 glioma cells.
252 Incubation with glia cells or their plasma membranes
al. J9 and Vinters et al. 29. This finding is consistent with
did not result in a significant increase of Na ÷,K+-ATPase
the observation that the BBB is situated in capillaries and not in larger vessels. In addition this result indicates that
and of G G T P activity. Data from co-culture experiments, findings. Our method of ouabain-sensitive S6Rb (K +)
endothelial cells retain some of their characteristics in vitro and therefore are a valuable model to study
uptake, revealed 90% increase in enzyme activity after co-culture. A p a r t from the influence of m e m b r a n e
induction of BBB features in vitro. Nevertheless, there is some evidence that neural e n v i r o n m e n t can cause for-
contact of neural cells on the cEC, we also tested for the importance of soluble factors. The experiments presented
mation of structural features of a competent BBB in cerebral as well as in non-cerebral endothelial cells 23'24.
have shown that conditioned media, neither from astrocytes, nor from C6 glioma cells affected G G T P activity,
It seems likely that in these cases additional components, such as extracellular matrix, play an important role, too.
suggesting that in vitro induction of enzyme activity is not due to soluble factors. Similar results have been reported by DeBault and Cancilla 11, but are in contrast to the findings of Maxwell et al. 19 and Vinters et al. 29. Our data
However, the co-culture system presented in this paper proved valuable for the study of gliat-endothelial and
provide direct evidence that m e m b r a n e - b o u n d components not only of glial cells but also of neurons, interact with cEC during cerebral ontogenesis and thus contribute
ontogenesis.
followed by histochemical studies 5 are consistent with our
to the establishment of metabolic and structural properties of the BBB. Aortic endothelial cells were left completely unaffected by incubation with plasma membranes, a result fitting perfectly to the data of Maxwell et
REFERENCES 1 Arthur, EE., Shivers, R.R. and Bowman, P.D., Astrocytemediated induction of tight junctions in brain capillary endothelium: an efficient in vitro model, Develop. Brain Res., 36 (1987) 155-159. 2 Bauer, H.C., Ferber, E., Golecki, J.R. and Brunner, G., Preparation and fractionation of membrane vesicles of thymocytes after osmotic cell disruption, Hoppe-Seylers Z. Physiol. Chem., 360 (1979) 1343-1350. 3 Bauer, H.C. and Hurtenbach, U., Murine cortical brain cells are autoantigenic from a distinct developmental stage onwards, J. Neuroimmunol., 12 (1986) 1-13. 4 Bauer, H.C. and Tontsch, U., Glial conditioned medium and attachment to Con A are essential for long-term culture of cortical neurons, Int. J. Develop. Neurosci., 8 (1990) 151-158. 5 Beck, D.W., Roberts, R.L. and Olson, J.J., Glial cells influence membrane-associated enzyme activity at the blood-brain barrier, Brain Research, 381 (1986) 131-137. 6 Bergmeyer, H.U., Methoden der Enzymatischen Analyse, Vol. 1, Verlag-Chemie, Weinheim, 1970. 7 Betz, A.L. and Goldstein, G.W., The basis for active transport at the blood-brain barrier. In H.M. Eisenberg and R.L. Suddith (Eds.), The Cerebral Microvasculature, Plenum, New York, 1980, pp. 5-16. 8 Bradbury, M.W.B., The Concept of a Blood-Brain Barrier, Wiley, Chichester, 1979, 455 pp. 9 Brightman, M.W. and Reese, T.S., Junctions between intimately apposed cell membranes in the vertebrate brain, J. Cell Biol., 40 (1969) 648-677. 10 Davson, H. and Oldendorf, W.H., Transport in the central nervous system, Proc. R. Soc. Med., 60 (1967) 326-328. 11 DeBault, L.E. and CanciUa, P.A., y-Glutamyl transpeptidase in isolated brain endothelial cells: induction by glial cells in vitro, Science, 207 (1980) 653-655. 12 Freshney, R.I., Culture of Animal Cells. A Manual of Basic Technique, 2nd edn., Liss, New York, 1987, 397 pp. 13 Goldstein, G.W., Relation of potassium transport to oxidative metabolism in isolated brain capillaries, J. Physiol., 286 (1979)
n e u r o n a l - e n d o t h e l i a l relationships in vitro and may be expected to shed light on BBB establishment during
Acknowledgements. We thank M. Steiner and I. Gmachl for their technical assistance with electron microscopy, Drs. J.V. Small and K. Kratochwil are thanked for reading the manuscript and for valuable discussions. This work was supported by the Fond zur F6rderung der wissenschaftlichen Forschung, Austria; Project P 6268.
185-195. 14 Jaffe, E.A., Nachman, R.L., Becker, C.G. and Minick, C.R., Culture of human endothelial cells derived from umbilical veins, J. Clin. Invest., 52 (1973) 2745-2756. 15 Janzer, R.C. and Raff, M.C., Astrocytes induce blood-brain barrier properties in endothelial cells, Nature, 325 (1987) 253-257. 16 Kawai, K., Takahashi, H. and lkuta, F., Ultracytochemical study of capillary CaZ+-ATPase activity in brain edema, Acta Neuropathol., 77 (1989) 449-454. 17 Laitinen, L., Griffonia simplicifolia leetins bind specifically to endothelial cells and some epithelial cells in mouse tissues, Histochem. J., 19 (1987) 225-234. 18 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J., Protein measurement with the Folin phenol reagent, J. Biol. Chem., 193 (1951) 265-275. 19 Maxwell, K., Berliner, J.A. and Cancilla, P.A., Induction of y-glutamyl transpeptidase in cultured cerebral endothelial cells by a product released by astrocytes, Brain Research, 410 (1987) 309-314. 20 Naftalin, L., Sexton, M., Whitaker, J.F. and Tracey, D., A routine procedure for estimating serum y-glutamyl transpeptidase activity, Clin. Chim. Acta, 26 (1969) 293-297. 21 Orlowski, M., Sessa, G. and Green, J.P., 7-Glutamyl transpeptidase in brain capillaries: possible site of a blood-brain barrier for amino acids, Science, 184 (1974) 66--68. 22 Reese, T.S. and Karnovsky, M.J., Fine structural localization of a blood-brain barrier to exogenous peroxidase, J. Cell Biol., 34 (1967) 207-217. 23 Shivers, R.R., Arthur, F.E. and Bowman, P.D., Induction of gap junctions and brain endothelium-like tight junctions in cultured bovine endothelial cells: local control of cell specialization, J. Submicrosc. Cytol. Pathol., 20 (I988) 1-14. 24 Stewart, P.A. and Wiley, M.J., Developing nervous tissue induces formation of blood-brain barrier characteristics in invading endothelial cells. A study using quail-chick transplantation chimeras, Develop. Biol., 84 (1981) 183-192. 25 Tao-Cheng, J., Nagy, Z. and Brightman, M.W., Tight junctions of brain endothelium in vitro are enhanced by astroglia, J.
253
Neurosci., 7 (1987) 3293-3299. 26 Tao-Cheng, J. and Brightman, M.W., Development of membrane interactions between brain endothelial cells and astrocytes in vitro, Int. J. Develop. Neurosci., 6 (1988) 25-37. 27 Tontsch, U. and Bauer, H.C., Isolation, characterization, and long-term cultivation of porcine and murine cerebral capillary endothelial cells, Microvasc. Res., 37 (1989) 148-161. 28 Vaughan, G.L. and Cook, J.S., Regeneration of cation-transport capacity in HeLa cell membranes after specific blockade by ouabain, Proc. Natl. Acad. Sci. U.S.A., 69 (1972) 262%2631. 29 Vinters, H.V., Reave, S., Costello, P., Girvin, J.P. and Moore,
S.A., Isolation and culture of cells derived from human cerebral microvessels, Cell Tissue Res., 249 (1987) 657-667. 30 Vorbrodt, A.W., Ultrastructural Cytochemistry of Blood-Brain Barrier Endothelia, Prog. Histochem. Cytochem., Vol. 18, Fischer, Stuttgart, 1988. 31 Wurtman, R.J. and Axelrod, J., A sensitive and specific assay for the estimation of monoamine oxidase, Biochem. Pharmacol., 12 (1963) 1439-1440. 32 Yoshida, Y., Yamada, M., Wakabayashi, K. and Ikuta, E, Endothelial fenestrae in the rat fetal cerebrum, Develop. Brain Res., 44 (1988) 211-219.
247
BRES 16259
Glial cells and neurons induce blood-brain barrier related enzymes in cultured cerebral endothelial cells Ulrike Tontsch* and Hans-Christian Bauer lnstitut far Molekularbiologie der Osterreichischen Akademie der Wissenschaften, Salzburg (Austria) (Accepted 14 August 1990)
Key words: Blood-brain barrier; Cerebral endothelial cell; Glial cell; Enzyme induction; y-Glutamyltranspeptidase; ATPase, (Na+,K+)-
The blood-brain barrier (BBB) in mammals is created and maintained by cerebral endothelial cells (cEC) that express specialized functional properties, including intercellular tight junctions, absence of fenestrae and specific membrane transport systems. It has been proposed that the differentiation of these characteristics, acquired during brain development, is controlled by the neural environmentls'24. Co-culture experiments of cloned cEC with astroglial cells, C6 glioma cells and cortical neurons, with plasma membranes or conditioned media of these cells, were used to study induction of some BBB characteristics in vitro. Activities of Na+,K+-ATPase and y-glutamyltranspeptidase (GGTP), an enzyme responsible for amino acid transport across the BBB, were taken as parameters for BBB function. Co-culture of cEC with C6 glioma cells caused a two-foldincrease in GGTP activity and this activitywas likewise amplified by incubation with plasma membrane fractions derived from C6 glioma cells, embryonic brain cells and cortical neurons; conditioned media (soluble factors) had no effect. Na+,K+-ATPase activity, estimated from the ouabain inhibitable fraction of SrRb uptake, was increased by about 90% in cEC incubated with C6 glioma plasma membranes. We propose from these data that both neurons and glial cells confer BBB characteristics on cEC via cell-cell contact.
INTRODUCTION Although structure and function of the blood-brain barrier (BBB), maintained primarily by cerebral capillary endothelial cells (cEC), have been studied extensively8, 9,22, it still remains unclear, which factors are responsible for the establishment of this barrier during embryonic development. Transplantation experiments by Stewart and Wiley24 have provided strong evidence that BBB characteristics of cEC depend on their neural environment. Cerebral capillaries are tightly ensheathed by astrocytes and the tightness of the BBB correlates with the completeness of this glial cell layer in the course of brain development 32 as well as in pathological processes 16. Therefore these cells have been considered responsible for the induction of BBB features in cEC. Investigations of cerebral vascularization and BBB formation have focused on two aspects: the tightness of the B B B and the selective transport mechanisms of cEC. Factors which increase or decrease the number of tight junctions and fenestrae have been studied in vivo and in vitro 1AS"23-25"2s'32. The selective transport mechanisms across the BBB were monitored with alkaline phosphatase, Na+,K+-ATPase, Ca2+.ATPase, 5"-nucleotidase
and
y-glutamyl
transpeptidase
(GGTP) 7'11'16'1a'29'30.
GGTP, which has been described as a typical BBBrelated enzyme 21, is located in the luminal as well as in the abluminal membrane of cEC and is supposed to be involved in the transport of large neutral amino acids across the BBB. The relatively high GGTP activity of cEC in vivo markedly decreases when these cells are kept in culture. The first attempts to mimick the in vivo situation were made by DeBault and Cancilla 11 who found by histochemical staining that co-cultures of cEC with C6 glioma cells increased GGTP activity. After similar co-culture experiments, Beck et al. 5 described enhanced Na+,K*-ATPase activity, which is also a feature of BBB function. Until now, only data from histochemical studies were available for Na ÷,K+-ATPase activity in cEC. It has generally been accepted that glial cells possess BBB-inductive potential. Both soluble proteins released from glial cells as well as membrane bound structures are believed to act on cEC plasma membranes in a way which leads to BBB formation LH,15,19,25. In this paper we have exploited and extended the co-culture approach to test the influence of other cells and cell fractions on the cEC. The results, reported here, indicate that membrane
* Present address: Max-Planck-Institut for Psychiatrie, Abteilung Neuroimmunologie, Martinsried, ER.G. Correspondence: H.-C. Bauer, Institut fOr Molekularbiologie der Osterreiehischen Akademie der Wissenschaften, Billrothstr. 11, A-5020 Salzburg, Austria. 0006-8993/91/$03.50 (~) 1991 Elsevier Science Publishers B.V. (Biomedical Division)
248 i n t e r a c t i o n s b e t w e e n c E C a n d n e u r a l cells are r e s p o n s i b l e f o r i n c r e a s e d activity o f B B B - r e l a t e d e n z y m e s , w h e r e a s soluble factors are ineffective.
MATERIALS AND METHODS
Cell culture Porcine cEC were isolated, cultivated and characterized as described previously 27. Briefly, cerebral capillaries were released from the surrounding neural tissue after several steps of homogenization and centrifugation. After dissociation in collagenase, the cells were plated on 35-mm diameter plastic petri dishes (TC quality, NUNC, Denmark), and kept in medium 199 (Seromed, F.R.G.), supplemented with 10% heat-inactivated foetal calf serum (FCS) (Seromed), 100/~g/ml heparin (Sigma, ER.G.), 20/~g/ml endothelial cell growth supplement (ECGS) (Collaborative Res.) and antibiotics. CEC were characterized with the lectin of Bandeiraea simplicifolia ~7. For our studies we used cloned cEC (designated A9B12) of the cobble-stone type, at passages between 15 and 20, always at confluence. Cells were plated on 35-mm diameter plastic petri dishes (NUNC) or in 24-well plates (Costar, Cambridge, MA). Primary cultures of murine astrocytes were obtained as described previously4, kept in Dulbecco's modified Eagle's medium with 10% FCS, characterized with glial fibrillary acidic protein (GFAP) and used after 2 weeks in culture. Porcine aortic EC (aEC) were isolated by collagenase treatment using the procedure of Jaffe TM and were kept under the same culture conditions as the cEC. Primary cultures of murine cortical neurons were established as described 4 and used after 3 days in culture. Y-1 mouse adrenal tumor cells and CCRF-S180 mouse sarcoma cells were obtained from the American Tissue Culture Collection and were both maintained in Ham's F12 medium, supplemented with 10% FCS. C6 rat glioma cells, a generous gift of Dr. B. Hamprecht, Univ. T~ibingen, were also cultivated in Ham's F12 with 10% FCS. Chicken fibroblasts were isolated and cultivated according to Freshney ~2.
Co-culture CEC (A9B12) were co-cultured with conditioned media, cells or plasma membranes of various cell types as follows. (1) cEC grown on 35-mm diameter.dishes were incubated with C6 conditioned medium (C6CM) or with glia conditioned medium (GCM) for 24, 48 and 72 h. Conditioned medium was obtained as follows: confluent cells were washed with PBS and supplied with fresh medium 199 + 10% FCS. After 72 h the medium was collected, centrifuged to remove cell debris, filter sterilized and used immediately, (2) cEC in 35-mm diameter dishes were co-cultivated with C6 glioma cells, primary glial cells, primary neurons, fibroblasts, Y-1 cells and CCRF cells, growing on glass coverslips (20 x 20 mm). These coverslips were placed upside down on the layer of cEC and remained there for 24, 48 and 72 h. (3) cEC in 35-mm diameter dishes and in 24-well plates were incubated with isolated plasma membranes (PM) of C6 glioma cells and glial cells for 24, 48 and 72 h. PMs were also derived from primary cultures of neurons, fibroblasts, Y-1 adrenal tumor cells, CCRF sarcoma cells and from murine brain (embryonic day 14). Protein concentration of the PMs ranged from 2.5 to 200/~g/ml medium 199 + 10% FCS and incubation lasted for 48 h. CEC reference cultures were also supplied with medium 199 + FCS. Isolation of PM fractions was carried out as described previously 3. Briefly, confluently growing cells were trypsinized, centrifuged, and the pellet forcefully homogenized with a tightly fitting pestle in isotonic saccharose-phosphate buffer (0.32 M, pH 7.6). In the first centrifugation step (800 g, 15 min, 4 °C), cell nuclei, and in the second (11,500 g, 25 min, 4 °C), mitochondria were removed from the homogenate. This supernatant was subsequently centrifuged (100,000 g, 45 min, 4 °C), the pellet resuspended in 5 mM
Tris-HCl/1 mM EDTA, pH 8.1 and centrifuged again (11,500 g, 25 min, 4 °C). This final pellet, containing almost exclusively PMs, was dissolved in sterile phosphate-buffered saline (PBS) and used for co-culture experiments. PMs from cells of the embryonic brain cortex were isolated from about 100 murine brains (Balb/c). They were cleared from the meninges, minced, homogenized and processed as above. (4) aEC were cocultivated with the PM fraction of C6 glioma cells.
Determination of enzyme activities" GGTP. Conditioned media, coverslips, and PM fractions were removed from the cEC cultures. Subsequently, the cEC were washed 3 times with PBS and processed according to Maxwell et al. 19. Briefly, the cells were incubated overnight in 0.1 M phosphate buffer (pH 7.4), and the cell suspension sonicated for 10 s at 20 W (Braun Labsonic 1510). GGTP activity was determined with a colorimetric assay, based essentially on the method of Naftalin et al. 2°. According to this technique, 7-c-glutamyl-p-nitroanilide serves as the substrate and glycylglycine as the acceptor molecule for the glutamyl group released by the action of GGTP. Liberated pnitroaniline is diazotized and measured at 540 nm. The absorbance is proportional to GGTP activity, which is related to protein concentration, measured by the method of Lowry et a1.18 Reagents for GGTP determination were all purchased from Sigma. Na+,K+-ATPase. Na+,K+-ATPase activity was determined by the use of 86Rb influx. As shown by Vaughan and Cook 28, K + can be fully substituted by Rb. The Na+,K+-ATPase activity was estimated as the ouabain inhibitable part of total 86Rb influx 13. Cloned cEC were cultured in 12-well multiwells (Costar) for 5 days to confluency. Prior to use, cells were washed twice with a buffered salt solution (containing the following (mM): NaCI 120, KCI 5, CaCl 2 2, MgCI 2 1, HEPES 20; pH 7.4 with Tris) and were preincubated with 1 mM ouabain (Sigma) for 20 min at 37 °C. K + influx was measured at room temperature with the addition of approximately 1 #Ci 86RbCI (Amersham, U.K.) in water. For most experiments influx was followed for 15 rnin. The incubation medium was then removed and the cells rapidly washed 3 times with ice-cold isotonic NaCI. Intracellular 86Rb was extracted with 0.5% Triton X-100 overnight, scintillation liquid (Emulsifier, Packard) was added and radioactivity determined in a Packard fl-scintillation counter. Protein was measured by the method of Lowry et al. TM and 86Rb (K +) influx was calculated as cpm/min/mg protein.
Characterization of PM fractions Lactate dehydrogenase was used to determine contamination with cytoplasmic proteins in the PM preparations. The enzyme activity was measured using the method of Bergmeyer 6. The presence of mitochondrial membranes was identified by determination of monoamine oxidase activity, following the procedure of Wurtman and Axelrod 31. Electron micrographs of PMs were produced as described previously 2. RESULTS
Cell cultures The described isolation procedure, culture conditions a n d in p a r t i c u l a r t h e cell c l o n i n g left u s w i t h a p u r e p o p u l a t i o n o f c e r e b r a l e n d o t h e l i a l cells ( c l o n e A 9 B 1 2 ) . T h e cells c o u l d b e p a s s a g e d u p t o 35 t i m e s w i t h o u t l o s i n g t h e i r c o b b l e s t o n e p h e n o t y p e (Fig. l a ) o r t h e e x p r e s s i o n of t h e b i n d i n g sites f o r t h e m a r k e r Bandeiraea simplicifolia lectin. B r i g h t f l u o r e s c e n c e l a b e l i n g w i t h this l e c t i n is s h o w n in Fig. l b . C o r t i c a l n e u r o n s , c u l t i v a t e d f o r 3 d a y s a r e s h o w n in Fig. l c . A s t r o c y t e s w e r e c h a r a c t e r i z e d b y t h e p o s s e s s i o n of glial f i b r i l l a r y acidic p r o t e i n (Fig. l d ) .
249
Fig. 1. a: phase contrast micrograph of cloned porcine cEC, passage 18, grown in medium 199 + 10% FCS + heparin + ECGS. Bar, 100 /~m. b: cEC were labeled with TRITC-conjugated lectin of Bandeiraea simplicifolia. Bar b-d, 50 9m. c: phase contrast micrograph of murine cortical neurons cultivated for 3 days. d: fluorescent micrograph of routine astrocytes stained for GFAP.
G G T P activity - coculture with cells, conditioned media and PMs Cocuiture experiments were carried out for 24, 48 and 72 h. In freshly isolated capillaries G G T P activity was 800 U/mg protein, in cultivated cEC G G T P activity decreased to about 200 U/mg protein. If cEC were cocultivated with C6 glioma cells, enzyme activity increased up to two-fold over untreated cEC. With primary cultures of glia cells, neurons, fibroblasts or the Y-1 tumor cells and CCRF, no enzyme induction was noticed (data not shown). Neither did conditioned medium from C6 glioma cells nor from glial cells affect G G T P activity in cEC. Incubation of cEC with the PM fraction from C6 increased G G T P activity from 180 + 40 U in untreated cEC of the identical passage to 1296 + 109 U (Fig. 2). Appreciable enzyme activity was already observed after 24 h, and increased after incubation for 48 and 72 h. In the following experiments incubations were performed for 48 h. cEC were subjected to isolated PMs in concentrations ranging from 2.5 to 200/~g protein/ml and showed a dose-dependent increase of enzyme activity (Fig. 3). We also incubated cEC with PMs of Y-1 adrenal
tumor cells and of CCRF sarcoma cells and could show that enzyme induction was not tumor cell dependent. As C6 glioma cells themselves express relatively high G G T P
1800 1400 1200 =
1000 800
"~ 600 400 800 0 ~
~ cap
cEC
J + 06
+ C6 CH
+ PM 06
Fig. 2. Determination of G G T P activity,as described in Materials and Methods, of freshly isolated cerebral capillaries(cap), cultivated cEC, and of cEC after co-culture with C6 gfiorna cells,with conditioned medium of C.6 (06 C M ) and with plasma membranes (PM C6). Protein concentrationof the P M ~acfion was 200/~g/mJ. Co-cul~m~ was carried out for 24, 48 and 72 h. Means are given and S.D.
250 activity (972 + 43 U G G T P / m g protein in C6 glioma cells vs 125 + 23 in glial cells), we investigated the possibility that the increase of enzyme activity was due to unspecific attachment of PM C6. This could be excluded as fibroblast cultures, subjected to PM C6 caused no increase in G G T P activity, aEC were also incubated with PM C6 and were found completely unaffected (Table I). Incubation with PM of normal glia cells did not lead to pronounced G G T P induction, neither in cEC nor in aEC. To test whether other components of the brain, apart
3000 2500
TABLE I Confluent cultures of aEC incubated with PM C6 in concentrations indicated
Data are means and S.D. of 4 measurements and were analyzed by Student's t-test. aEC Incubated with
U GG TP/mg protein
PM PM PM PM PM
226 +_55 216 + 53 168 + 18 215 + 82 167 + 31 289 + 103 n.s.
C6 C6 C6 C6 C6
(10#g/ml) (20#g/ml) (50#g/ml) (100/zg/ml) (200/~g/ml)
2000 15oo
t
1000
500
0 0
50
20
10
100
200 ~ g / m l
3000Ib 2500 2000
1500i
i
i000[
Characterization o f P M fractions
The PM fractions were defined by exclusion of marker enzymes, monoamine oxidase for mitochondrial mem-
500
o,i, 0
2000F
from the astrocytes, which are not present at early stages of embryonic development, are involved in inducing BBB features in cEC, we subjected cEC to a PM fraction of embryonic brain cells. G G T P activity of this fraction was 232 U G G T P / m g protein. Notably, our data showed a dose-dependent, up to 7-fold enzyme induction in the cEC (Fig. 3). We reasoned that neurons were responsible for this effect. To prove this assumption, we isolated PMs from primary cultures of neurons ( G G T P activity 3630 U/mg protein) and carried out co-culture experiments. As expected, there was a reasonable increase in G G T P activity and suprisingly this PM fraction proved to be an equally potent inducer as PM C6 (Fig. 3).
.
.
10 .
.
.
.
20 .
.
.
.
50 pg/ml .
.
.
.
.
.
.
.
.
1500
o
1000
500 1
0 ~
0
2,5
5
10
20
50
~00
200 ~ag/ml
Fig. 3. Determination of GGTP activity, a: in cEC after incubation (48 h) with C6 glioma PM fraction. Protein concentration of PMs ranged from 10 to 200 #g/ml. Data are given as means and S.D. of 3--7 measurements, b: in cEC after co-culture with neuronal PM fraction. Protein concentration of PMs was 10, 20 and 50 #g/ml. Values give means and S.D. of 5 experiments, c: in cEC after co-culture with PMs from embryonic brain cells in concentrations ranging from 2.5 to 200 #g/ml. Data represent means and S.D., n = 7-11.
Fig. 4. Electron micrograph of a plasma membrane fraction from C6 glioma cells, x38,100.
251 .~
48 h) did not have any effect on Na +,K+-ATPase activity in cEC (data not shown).
Thoussnd= 80
t
•
o
60 ...................................................
DISCUSSION
J u
40
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
T
o
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
~
'
'
~
'
to
20
30
40
50 mia
Fig. 5. Total initial S~Rb (K +) uptake rate in cEC. After passage,
cells were kept 5 days in culture and were nearly confluent. Before incubation with mRb (K+), cells were carefully washed with PBS. n = 4, means and S.D. are given.
branes and lactate dehydrogenase for cytoplasmic constituents. Neither monoamine oxidase nor lactate dehydrogenase activity was found, indicating that there were no mitochondrial contaminations and no trapping of cytoplasm. The purity of the PM fractions was monitored by electron microscopy (Fig. 4). Induction o f Na+,K+-ATPase 86Rb (K t ) influx in cultured cEC was linear at least over the first 45 min (Fig. 5). For all co-culture experiments 86Rb (K t ) uptake was followed for 15 min in order to avoid possible cell damage during longer incubation periods. Treatment of cEC with PM (200 pg/ml) from C6 glioma cells for 48 h resulted in an increase in Na t , K t - A T P a s e activity. Table II shows that the ouabain inhibitable part of 86Rb (K t ) influx was 53% of total influx in PM treated cEC versus 28% in control cells. This corresponds to an increase of Na t , K t - A T P a s e activity of about 90% in cEC incubated with PM. Incubation with PM from cultured fibroblasts (200 pg/ml;
TABLE II Total influx ofaORb (K +) and ouabain inhibitable S6Rb (K +) influx in cultured cerebral endothelial cells (cEC) with or without plasma membranes of C6 glioma cells (PM C6)
After passage, cEC were kept 5 days in culture and were nearly confluent. PMs were incubated for 2 days. Cells were washed carefully with Dulbecco's PBS before use. a6Rb (K+) uptake was 15 min at room temperature. Details are given in Materials and Methods. Data represent 2-4 experiments, means and S.D. are given.
Cells
Total S6Rb (K +) influx (cpm/min/mg protein)
Ouabain inhibitable influx (Na +,K+.ATPase activity) (cpm/min/mg protein)
cEC cEC + PMC6
5087 + 83 (100%) 4951+210(100%)
1439 + 212 (28%) 2610+381(53%)
Our experiments have shown that glioma cells or their plasma membranes and plasma membranes from neuronal cells induce BBB-related enzyme activities in cultured cEC. Although it is generally assumed that BBB development is under environmental c o n t r o l z ° ' 1 5 ' 2 4 ' 3 2 , it is still unclear whether the effect is due to diffusable factors or to cell-cell contacts. Glial cells are usually considered as the main candidate for the induction of BBB properties in cEC 1'5'7'11'15'19"25'26. Our experiments have confirmed the influence of glial cells on cEC, and in addition we found that neurons play a significant role, too. First hints for the importance of neurons for the BBB development, which were based on light and electron microscopic studies, came from Stewart and Wiley 24 and from Yoshida et a l . 32. We chose two enzyme systems, G G T P and Na +,K+-ATPase to monitor the effect of glial cells and neurons on cerebral endothelial cells. Co-culture of cEC with C6 glioma cells resulted in reasonable increase in G G T P activity, but never as much as attained after incubation with PMs. The reason for this discrepancy may be technical in origin: (1) gas and medium exchange was not optimal and sometimes even caused the cells to detach; and (2) only about half of the cells in the culture dish could be brought into contact with the cells on the coverslip, which in turn reduced the effect of G G T P induction. However, all these problems could be readily overcome by incubating with PMs. These standardized conditions were a necessity for reproducible results and removed the uncertainties associated with variable cell density and growth state of the cells added. In addition, it was possible to test for the effect of membrane contact, avoiding interference with soluble factors. Cerebral endothelial cells responded to PM incubation with an increase in enzyme activity up to about 2000 U/mg protein and revealed only minor differences for the various PM fractions (PM C6, PM N, PM brain). This compares with an enzyme activity of about 800 U/rag protein in freshly isolated capillaries. Bearing in mind that the amount of protein not only refers to cEC, but includes that of erythrocytes, lymphocytes, pericytes and perhaps some smooth muscle cells, the specific enzyme activity for cEC can be assumed at least twice as high. Our data further show that for enzyme activity induction in cEC G G T P activity of the added cells or their plasma membranes is irrele"ant. With the second enzyme system, the Na+,K+-ATPase, we also demonstrated an increase in activity after co-culture of cEC with PMs of C6 glioma cells.
252 Incubation with glia cells or their plasma membranes
al. J9 and Vinters et al. 29. This finding is consistent with
did not result in a significant increase of Na ÷,K+-ATPase
the observation that the BBB is situated in capillaries and not in larger vessels. In addition this result indicates that
and of G G T P activity. Data from co-culture experiments, findings. Our method of ouabain-sensitive S6Rb (K +)
endothelial cells retain some of their characteristics in vitro and therefore are a valuable model to study
uptake, revealed 90% increase in enzyme activity after co-culture. A p a r t from the influence of m e m b r a n e
induction of BBB features in vitro. Nevertheless, there is some evidence that neural e n v i r o n m e n t can cause for-
contact of neural cells on the cEC, we also tested for the importance of soluble factors. The experiments presented
mation of structural features of a competent BBB in cerebral as well as in non-cerebral endothelial cells 23'24.
have shown that conditioned media, neither from astrocytes, nor from C6 glioma cells affected G G T P activity,
It seems likely that in these cases additional components, such as extracellular matrix, play an important role, too.
suggesting that in vitro induction of enzyme activity is not due to soluble factors. Similar results have been reported by DeBault and Cancilla 11, but are in contrast to the findings of Maxwell et al. 19 and Vinters et al. 29. Our data
However, the co-culture system presented in this paper proved valuable for the study of gliat-endothelial and
provide direct evidence that m e m b r a n e - b o u n d components not only of glial cells but also of neurons, interact with cEC during cerebral ontogenesis and thus contribute
ontogenesis.
followed by histochemical studies 5 are consistent with our
to the establishment of metabolic and structural properties of the BBB. Aortic endothelial cells were left completely unaffected by incubation with plasma membranes, a result fitting perfectly to the data of Maxwell et
REFERENCES 1 Arthur, EE., Shivers, R.R. and Bowman, P.D., Astrocytemediated induction of tight junctions in brain capillary endothelium: an efficient in vitro model, Develop. Brain Res., 36 (1987) 155-159. 2 Bauer, H.C., Ferber, E., Golecki, J.R. and Brunner, G., Preparation and fractionation of membrane vesicles of thymocytes after osmotic cell disruption, Hoppe-Seylers Z. Physiol. Chem., 360 (1979) 1343-1350. 3 Bauer, H.C. and Hurtenbach, U., Murine cortical brain cells are autoantigenic from a distinct developmental stage onwards, J. Neuroimmunol., 12 (1986) 1-13. 4 Bauer, H.C. and Tontsch, U., Glial conditioned medium and attachment to Con A are essential for long-term culture of cortical neurons, Int. J. Develop. Neurosci., 8 (1990) 151-158. 5 Beck, D.W., Roberts, R.L. and Olson, J.J., Glial cells influence membrane-associated enzyme activity at the blood-brain barrier, Brain Research, 381 (1986) 131-137. 6 Bergmeyer, H.U., Methoden der Enzymatischen Analyse, Vol. 1, Verlag-Chemie, Weinheim, 1970. 7 Betz, A.L. and Goldstein, G.W., The basis for active transport at the blood-brain barrier. In H.M. Eisenberg and R.L. Suddith (Eds.), The Cerebral Microvasculature, Plenum, New York, 1980, pp. 5-16. 8 Bradbury, M.W.B., The Concept of a Blood-Brain Barrier, Wiley, Chichester, 1979, 455 pp. 9 Brightman, M.W. and Reese, T.S., Junctions between intimately apposed cell membranes in the vertebrate brain, J. Cell Biol., 40 (1969) 648-677. 10 Davson, H. and Oldendorf, W.H., Transport in the central nervous system, Proc. R. Soc. Med., 60 (1967) 326-328. 11 DeBault, L.E. and CanciUa, P.A., y-Glutamyl transpeptidase in isolated brain endothelial cells: induction by glial cells in vitro, Science, 207 (1980) 653-655. 12 Freshney, R.I., Culture of Animal Cells. A Manual of Basic Technique, 2nd edn., Liss, New York, 1987, 397 pp. 13 Goldstein, G.W., Relation of potassium transport to oxidative metabolism in isolated brain capillaries, J. Physiol., 286 (1979)
n e u r o n a l - e n d o t h e l i a l relationships in vitro and may be expected to shed light on BBB establishment during
Acknowledgements. We thank M. Steiner and I. Gmachl for their technical assistance with electron microscopy, Drs. J.V. Small and K. Kratochwil are thanked for reading the manuscript and for valuable discussions. This work was supported by the Fond zur F6rderung der wissenschaftlichen Forschung, Austria; Project P 6268.
185-195. 14 Jaffe, E.A., Nachman, R.L., Becker, C.G. and Minick, C.R., Culture of human endothelial cells derived from umbilical veins, J. Clin. Invest., 52 (1973) 2745-2756. 15 Janzer, R.C. and Raff, M.C., Astrocytes induce blood-brain barrier properties in endothelial cells, Nature, 325 (1987) 253-257. 16 Kawai, K., Takahashi, H. and lkuta, F., Ultracytochemical study of capillary CaZ+-ATPase activity in brain edema, Acta Neuropathol., 77 (1989) 449-454. 17 Laitinen, L., Griffonia simplicifolia leetins bind specifically to endothelial cells and some epithelial cells in mouse tissues, Histochem. J., 19 (1987) 225-234. 18 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J., Protein measurement with the Folin phenol reagent, J. Biol. Chem., 193 (1951) 265-275. 19 Maxwell, K., Berliner, J.A. and Cancilla, P.A., Induction of y-glutamyl transpeptidase in cultured cerebral endothelial cells by a product released by astrocytes, Brain Research, 410 (1987) 309-314. 20 Naftalin, L., Sexton, M., Whitaker, J.F. and Tracey, D., A routine procedure for estimating serum y-glutamyl transpeptidase activity, Clin. Chim. Acta, 26 (1969) 293-297. 21 Orlowski, M., Sessa, G. and Green, J.P., 7-Glutamyl transpeptidase in brain capillaries: possible site of a blood-brain barrier for amino acids, Science, 184 (1974) 66--68. 22 Reese, T.S. and Karnovsky, M.J., Fine structural localization of a blood-brain barrier to exogenous peroxidase, J. Cell Biol., 34 (1967) 207-217. 23 Shivers, R.R., Arthur, F.E. and Bowman, P.D., Induction of gap junctions and brain endothelium-like tight junctions in cultured bovine endothelial cells: local control of cell specialization, J. Submicrosc. Cytol. Pathol., 20 (I988) 1-14. 24 Stewart, P.A. and Wiley, M.J., Developing nervous tissue induces formation of blood-brain barrier characteristics in invading endothelial cells. A study using quail-chick transplantation chimeras, Develop. Biol., 84 (1981) 183-192. 25 Tao-Cheng, J., Nagy, Z. and Brightman, M.W., Tight junctions of brain endothelium in vitro are enhanced by astroglia, J.
253
Neurosci., 7 (1987) 3293-3299. 26 Tao-Cheng, J. and Brightman, M.W., Development of membrane interactions between brain endothelial cells and astrocytes in vitro, Int. J. Develop. Neurosci., 6 (1988) 25-37. 27 Tontsch, U. and Bauer, H.C., Isolation, characterization, and long-term cultivation of porcine and murine cerebral capillary endothelial cells, Microvasc. Res., 37 (1989) 148-161. 28 Vaughan, G.L. and Cook, J.S., Regeneration of cation-transport capacity in HeLa cell membranes after specific blockade by ouabain, Proc. Natl. Acad. Sci. U.S.A., 69 (1972) 262%2631. 29 Vinters, H.V., Reave, S., Costello, P., Girvin, J.P. and Moore,
S.A., Isolation and culture of cells derived from human cerebral microvessels, Cell Tissue Res., 249 (1987) 657-667. 30 Vorbrodt, A.W., Ultrastructural Cytochemistry of Blood-Brain Barrier Endothelia, Prog. Histochem. Cytochem., Vol. 18, Fischer, Stuttgart, 1988. 31 Wurtman, R.J. and Axelrod, J., A sensitive and specific assay for the estimation of monoamine oxidase, Biochem. Pharmacol., 12 (1963) 1439-1440. 32 Yoshida, Y., Yamada, M., Wakabayashi, K. and Ikuta, E, Endothelial fenestrae in the rat fetal cerebrum, Develop. Brain Res., 44 (1988) 211-219.
