- Email: [email protected]
Functional Coupling of Secretion and Capacitative Calcium Entry in PC12 Cells
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
247, 536 (1998)
ERRATUM Volume 244, Number 1 (1998), in Article No. RC988251, ‘‘Functional Coupling of Secretion and Capacitative Calcium Entry in PC12 Cells,’’ by Schuichi Koizumi and Kazuhide Inoue, pages 293–297, and in Article No. RC978051, ‘‘Hydrocortisone Reinforces the Blood–Brain Barrier Properties in a Serum Free Cell Culture System,’’ by Dirk Hoheisel, Thorsten Nitz, Helmut Franke, Joachim Wegener, Ansgar Hakvoort, Thomas Tilling, and Hans-Joachim Galla, pages 312–316: Due to a compositor’s error, the figures for these two articles were inadvertently switched. The legends for the figures are correct as printed. For the reader’s convenience, both complete articles with the correct figures are printed here. This erratum is Article No. RC988653.
536 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 8653
/
6955$$$$81
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
244, 293–297 (1998)
RC988251
Functional Coupling of Secretion and Capacitative Calcium Entry in PC12 Cells Schuichi Koizumi*,†,1 and Kazuhide Inoue* *Division of Pharmacology, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya, Tokyo 158, Japan; and †Laboratory of Molecular Signalling, The Babraham Institute, Babraham Hall, Cambridge, CB2 4AT, United Kingdom
Received January 26, 1998
The caffeine-evoked effects on the intracellular Ca2/ concentration ([Ca2/]i) and on the release of dopamine by PC12 cells were investigated. Stimulation by caffeine resulted in a transient Ca2/ release which was followed by a sustained phase of Ca2/ entry through a non-voltage dependent pathway. Treatment with cyclopiazonic acid (CPA) or thapsigargin, inhibitors of the Ca2/ATPase pump of the endoplasmic reticulum, resulted in only a sustained rise in [Ca2/]i in the presence of extracellular Ca2/. Pretreatment of cells with CPA or thapsigargin abolished the subsequent Ca2/ responses to caffeine. Caffeine also evoked the release of dopamine from the cells only in the presence of extracellular Ca2/, which was mimicked by CPA. These results suggest that store-dependent Ca2/ entry evoked by caffeine has an indispensable role in the secretory response in an excitable cell line, PC12 cells. q 1998 Academic Press
Over ten years have passed since the original idea that Ca2/ can enter cells through the plasma membrane in response to intracellular store depletion was first proposed (1). Recently, much attention has been focused on this so-called capacitative Ca2/ entry (CCE) because it has become clear that CCE plays a central role in many aspects of cell signaling (2, 3). In excitable cells, however, the existence of CCE is controversial, in part because it is unclear whether ryanodine receptor (RyR)-linked Ca2/ stores activate CCE in a manner similar to that which results from depletion of inositol 1,4,5-trisphosphate receptor (InsP3R)-linked stores (2). 1 Corresponding author. Fax: 03-3700-9698. E-mail: inoue@ nihs.go.jp. Abbreviations: [Ca2/]i, intracellular Ca2/ concentration; CCE, capacitative Ca2/ entry; CPA, cyclopiazonic acid; InsP3 , inositol 1,4,5trisphosphate; RyR, ryanodine receptor.
We previously reported that activation of P2U-purinoceptors in an excitable cell line, rat pheochromocytoma PC12 cells, by uridine 5*-triphosphate (UTP) leads to a transient rise in the intracellular Ca2/ concentration ([Ca2/]i) resulting from inositol 1,4,5-trisphosphate (InsP3)-mediated Ca2/ mobilization, which was followed by a sustained rise in [Ca2/]i resulting from nonvoltage dependent Ca2/ influx, presumably CCE (4). Besides InsP3-sensitive Ca2/ stores, caffeine/ryanodine sensitive Ca2/ stores are present in PC12 cells (5), which enables us to investigate whether RyR-linked Ca2/ stores activate CCE via an InsP3-independent mechanisms. As for the physiological significance of CCE in excitable cells, it is far from clear. UTP could stimulate the release of dopamine only in the presence of extracellular Ca2/ in PC12 cells (4), which raises the possibility that CCE activated by InsP3R-mediated Ca2/ release and the secretory response are closely linked. In this study, we demonstrate the existence of CCE in an excitable cell line, PC12 cells, and report that, in addition to InsP3Rs activation (4), CCE can be stimulated by activation of RyRs. Furthermore, we also show the role played by CCE in the exocytotic secretory response of the cells. MATERIAL AND METHODS Cell culture. Culture conditions of PC12 cells were as described previously (6). All experiments described in this manuscript were performed with cells at passages 53 through 68. Cells were plated onto collagen-coated 35 mm polystyrene dishes (1 1 106 cells/dish) for measuring the release of dopamine, or poly-L-lysine (Sigma, St. Louis, U.S.A.)-coated glass coverslips, placed in silicon rubber walls (Flexiperm, W.C. GmbH, Germany) for measuring the increase in [Ca2/]i. Cells were cultured an additional 2 days in a humidified atmosphere of 90 % air and 10 % CO2 at 37 7C. Measurement of [Ca2/]i in single cells. Changes in [Ca2/]i were measured by the fura-2 method as described previously (7) with minor modifications (4). In brief, 30 min after incubation with 5 mM
293
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. The time-course of the caffeine-evoked increase in [Ca2/]i in PC12 cells. a and b. Typical Ca2/ responses to 30 mM caffeine in the presence (a) and absence (b) of extracellular Ca2/. Caffeine was applied to the cells for 1 min (horizontal bars) at 4 min intervals. Traces a and b were obtained from the same cell. Vertical and horizontal scale bars show 100 nM and 1 min, respectively. c. The traces shown in a and b are superimposed and their time-courses were compared. The maximal Ca2/ responses to caffeine (hmax) are summarized in d (control, nÅ36; 0Ca2/, nÅ42). #s show the magnitude of [Ca2/]i (h#) at 1 min after caffeine application in the presence (control) and absence (0Ca2/) of extracellular Ca2/, and the responses are summarized in e. Values show the ratio of h# over hmax in the presence (nÅ36; open column) and absence (nÅ42; dotted column) of extracellular Ca2/. f and g. The effects of depletion of intracellular Ca2/ stores by CPA (cyclopiazonic acid)(f) and thapsigargin (g), on the caffeine-evoked rise in [Ca2/]i. After caffeine (30 mM) stimulation, cells were incubated with CPA (30 mM; hatched column) for 3 min (f), or thapsigargin (1 mM; dotted column) for 2 min (g), and then caffeine was applied to the 294
AID
BBRC 8653
/
6955$$8653
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
fura-2 acetoxymethylester (fura-2 AM) at 37 7C, cells were washed with a balanced salt solution (BSS) of the following composition (mM): NaCl 150, KCl 5.0, CaCl2 1.8, MgCl2 1.2, D-glucose 10 and N2-hydroxyethylpiperazine-N*-2-ethanesulfonic acid (HEPES) 25 (pH adjusted to 7.4 with NaOH) and mounted on an inverted fluorescence microscope (TMD-300, Nikon, Tokyo, Japan) equipped with a Xenonlamp and band-pass filters of 340 and 360 nm. For the Ca2/-depleted experiments, we used a medium in which Ca2/ was removed and 1 mM ethylenediaminetetraacetic acid (EGTA) was added (Ca2/-free BSS). Image data, recorded by a high-sensitivity silicon intensifier target camera (C-2741-08, Hamamatsu Photonics, Hamamatsu, Japan), were processed by a Ca2/-analyzing system (Furusawa Lab. Appliance. Co., Kawagoe, Japan). Caffeine was applied to the cells two or three times for 1 min at intervals of 4 min. The absolute [Ca2/]i was estimated from the ratio of emitted fluorescence (F340/ F360) according to the calibration curve obtained by standard Ca2/buffer. Measurement of dopamine release. The procedures for the measurement of released dopamine were basically the same as those described by Koizumi et al. (8). Cells were stimulated by various concentrations of caffeine and CPA dissolved in BSS for 1 and 3 min, respectively. For Ca2/-free experiments, the dishes were washed twice with Ca2/-free BSS for 1 min before caffeine application. The amount of dopamine released into the superfusate and that remaining in the cells were determined with high performance liquid chromatography (HPLC) coupled with an electrochemical detector (ECD)(LC-4B, Bioanalytical systems, West Lafayette, U.S.A.). The percentage of release was calculated by dividing the supernatant values by the sum of the supernatant and pellet values. Chemicals. Drugs used were as follows. Caffeine, cyclopiazonic acid (CPA), thapsigargin, nicardipine hydrochloride, cadmium chloride, zinc acetate and v-conotoxin GVIA (v-CTX) were purchased from Sigma. Fura-2 AM and HEPES were from (Dojin, Kumamoto, Japan). Other chemicals are purchased from Wako Purechemicals (Tokyo, Japan). Nicardipine, CPA and thapsigargin were dissolved in dimethyl sulphoxide at a concentration of 10 mM (nicardipine and CPA) or 1 mM (thapsigargin), and then dissolved in BSS to appropriate concentrations. Other drugs were directly dissolved in BSS or Ca2/-free BSS. All data are mean{s.e.m. Statistics. Statistical differences in the values of dopamine release or the increase in [Ca2/]i were determined using an analysis of variance followed by Dunnet’s test for multiple comparisons.
RESULTS Caffeine (30 mM) produced a significant rise in [Ca2/]i in about 80 % of the PC12 cells (188 out of 231 cells tested) and the average amplitude was 312{40.1 nM. When caffeine was applied to the cells twice for 1 min separated by 3 min, the Ca2/ response to the 2 nd caffeine was almost the same as that of the 1 st caffeine (data not shown). Fig. 1 (a, b and c) shows the timecourse of the caffeine-evoked increase in [Ca2/]i in the cells. Caffeine evoked a transient [Ca2/]i rise in the absence of extracellular Ca2/ (b), whereas it produced a transient [Ca2/]i rise, followed by a sustained rise in [Ca2/]i, in the presence of extracellular Ca2/ (a). The individual traces were superimposed and the time-
courses were compared (Fig. 1c). The maximal [Ca2/]i level evoked by caffeine (hmax) in the absence of extracellular Ca2/ was comparable to that in the presence of external Ca2/ (Fig. 1d). The [Ca2/]i at 1 min after the caffeine application (marked #) was defined as h# and the ratio of h#/hmax was calculated in the absence and presence of extracellular Ca2/. As shown in Fig. 1e, the h#/hmax ratio was dramatically decreased by the depletion of extracellular Ca2/, suggesting that the sustained component results from Ca2/ influx from extracellular spaces. Incubation of the cells with 30 mM CPA (Fig. 1f) or 1 mM thapsigargin (Fig. 1g), inhibitors of the Ca2/ATPase pump of the endoplasmic reticulum, resulted in a sustained rise in [Ca2/]i, which almost completely inhibited the subsequent Ca2/ responses to caffeine (30 mM). Thapsigargin evoked only a transient and slight rise in [Ca2/]i in the absence of extracellular Ca2/, but induced a long-lasting [Ca2/]i elevation upon replacement of the Ca2/-free medium with 1.8 mM Ca2/ (Fig. 1h). When the cells were repetitively exposed to 30 mM caffeine for periods of 1 min separated by 3 min in the presence of 10 mM ryanodine, the Ca2/ responses to caffeine were dramatically inhibited in a use-dependent manner and the 3 rd series of Ca2/ responses to caffeine were almost all abolished (data not shown). Similar to caffeine, theophylline (30 mM) produced a biphasic rise in [Ca2/]i, i.e. a transient elevation in [Ca2/]i was followed by a sustained one (data not shown). The sustained [Ca2/]i rise was totally dependent on extracellular Ca2/. Forskolin (10 mM) had no effects on either the resting [Ca2/]i or the theophyllineevoked rise in [Ca2/]i (nÅ48). Pretreatment of cells with nicardipine (30 mM) / v-conotoxin (v-CTX, 1 mM) for 1 min before and during the theophylline application had no effects on the sustained Ca2/ responses to theophylline (the amplitude of the sustained response 1 min after theophylline application in the presence of nicardipine / v-CTX was 104.4{9.1 % of theophylline alone, nÅ32), though these chemicals inhibited the KCl (53 mM)-evoked rise in [Ca2/]i by 82.4{6.2% (nÅ26). The effects of various compounds on the rise in [Ca2/]i were examined (Fig. 2). Zn2/, an inhibitor of CCE (9), almost completely inhibited the sustained component of the [Ca2/]i rise without affecting the transient component (Fig. 2b, d and e). Traces from Fig. 2a and b were superimposed and the time-courses were compared (Fig. 2c). The hmax in the presence of Zn2/ was not different from that in the absence of extracellular Zn2/ (Fig. 2c and d). However, Zn2/ at 30 mM dramatically inhibited the sustained component of the [Ca2/]i rise and significantly inhibited the h#/hmax ratio
cells in the presence of CPA (nÅ47) or thapsigargin (nÅ31). h. Cells were incubated with thapsigargin (1 mM; dotted column) in the absence of extracellular Ca2/ (broken line) for 2 min, which resulted in the abolishment of the caffeine-evoked rise in [Ca2/]i. Subsequently applied external Ca2/ produced a large and sustained [Ca2/]i increase (nÅ28). Vertical and horizontal scale bars show 100 nM and 1 min, respectively. 295
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. The effects of Zn2/, Cd2/, La3/, v-CTX and nicardipine on the caffeine-evoked rise in [Ca2/]i in PC12 cells. a and b. Typical Ca2/ responses to 30 mM caffeine in the absence (a) and presence (b) of 30 mM Zn2/. Caffeine was applied to the cells two times for 1 min at 4 min intervals (filled horizontal bars). Zn2/ (dotted column) and caffeine were applied to the cells simultaneously (b). Vertical and horizontal bars show 100 nM and 60 s, respectively. The traces shown in a and b were superimposed and their time-courses were compared in c. The maximal Ca2/ responses to caffeine (hmax) in the absence (open column) and presence (filled column) of 30 mM Zn2/ are summarized in d. #s show the magnitude of [Ca2/]i (h#) at 1 min after caffeine application in the absence (control, nÅ143) and presence (/Zn2/ 30 mM, nÅ37-58) of Zn2/, and the responses are summarized in e. The values show the ratio of h# over hmax in the absence (nÅ143; open column) and presence (nÅ37-58; filled column) of Zn2/. In addition to Zn2/, the effects of Cd2/ (nÅ48-61; hatched columns), La3/ (nÅ68; double hatched column), v-CTX (nÅ49, dotted column) and nicardipine (nÅ41; striped column) on the ratio of h#/hmax were examined and summarized in e. Asterisks show a significant difference from the ratio of control (*põ0.05, **põ0.01).
(Fig. 2e). Other cations (°100 mM) failed to inhibit the sustained component significantly (Fig. 2e) though La3/ showed inhibitory actions when its concentration was raised to 300 mM (h#/hmax ratio: 0.14{0.04, põ0.01,
nÅ38). Incubation of cells with nicardipine (30 mM) or v-CTX (1 mM) for 1 min before and during the caffeine stimulation had no effect on the sustained Ca2/ response to caffeine. Aminophylline (100 mM) had no ef-
296
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
PC12 cells, caffeine induces Ca2/ influx by a mechanism dependent upon Ca2/ depletion. Fig. 3a shows the concentration-dependence of the caffeine-stimulated release of dopamine and the effects of extracellular Ca2/ on the stimulation. Caffeine stimulated the release of dopamine in a concentration-dependent manner over a concentration range from 10 to 50 mM only in the presence of extracellular Ca2/ (Fig. 3a). The release of dopamine was mimicked by 30 mM CPA, and this was totally dependent on extracellular Ca2/ [CPA: 313.4{16.3 (nÅ6, /Ca2/), 102.2{6.2 (nÅ6, 0Ca2/) % of spontaneous release]. Aminophylline, an antagonist of non-selective adenosine receptors, had no effects on the release of dopamine evoked by caffeine (aminophylline 100 mM: 102.1{6.5 % of caffeine alone, nÅ6). The effects of various cations and an organic compound on the caffeine-evoked dopamine release were examined in the PC12 cells (Fig. 3b). Zn2/ potently inhibited the release of dopamine with an IC50 value of about 4 mM. Other cations or nicardipine failed to inhibit the dopamine release significantly. The effects of these chemicals on the release were well in agreement with those of the chemicals on the sustained [Ca2/]i rise evoked by caffeine (Fig. 2 and 3b). DISCUSSION
FIG. 3. a. Concentration-dependence of the caffeine-evoked dopamine release from PC12 cells. Open and closed circles represent the release of dopamine evoked by caffeine in the presence (open circles) and absence (closed circles) of extracellular Ca2/. These are results from a typical experiment with each data point being the mean{s.e.mean. of triplicate measurements. Four such experiments with similar results were performed. b. The effects of Zn2/, nicardipine, Cd2/ and La3/ on the caffeine-evoked dopamine release from PC12 cells. Values represent % of dopamine release evoked by 30 mM caffeine alone (open circle). The effects of Zn2/, nicardipine, Cd2/ and La3/ on the caffeine-evoked responses are indicated: Zn2/ (closed circles), nicardipine (open triangle), Cd2/ (closed triangles) and La3/ (closed squares). These are results from a typical experiment and each data point is the mean{s.e.mean. of triplicate measurements. At least three such experiments were performed, and similar results were obtained. Asterisks show a significant difference from the response evoked by caffeine alone (*põ0.05, **põ0.01).
fect on the caffeine-evoked sustained rise in [Ca2/]i (h#/ hmax :0.41{0.03, nÅ19). Similar observations were reported in response to UTP (4). In addition, the sustained [Ca2/]i elevation evoked by CPA (30 mM) was also inhibited by Zn2/ (30 mM) but not by Cd2/ (100 mM), La3/ (100 mM) or nicardipine (30 mM) (data not shown). Thus, these findings strongly suggest that, in
Caffeine acts at the RyR where it appears to shift the Ca2/ sensitivity of the channel to a lower concentration (10), thereby increasing the probability of channel opening, namely ‘‘calcium-induced Ca2/ release’’. In the PC12 cells, caffeine evoked Ca2/ release by stimulating RyRs since ryanodine (10 mM) abolished the Ca2/ responses to caffeine in a use-dependent fashion, which was well in accordance with previous results obtained from chromaffin cells (11) and PC12 cells (5). In addition to Ca2/ release from RyR, we showed here that caffeine stimulates a sustained Ca2/ entry in the cells (Fig. 1). The caffeine-evoked Ca2/ entry could be activated by a decrease in the stored Ca2/ concentration because (1) a similarly sustained Ca2/ entry was mimicked by both CPA and thapsigargin, inhibitors of the endoplasmic reticulum Ca2/ATPase pump, (2) pretreatment of cells with CPA or thapsigargin abolished the subsequent Ca2/ responses to caffeine, which indicates an overlap of Ca2/ entry mechanisms and (3) the sustained [Ca2/]i rises evoked by both caffeine and CPA were non-voltage dependent but were sensitive to Zn2/. These results strongly suggest that the caffeine-evoked sustained rise in [Ca2/]i is dependent upon the filling state of the intracellular Ca2/ stores. A similar phenomenon was observed upon UTP stimulation of these cells: UTP stimulates P2U-purinoceptors, which leads to InsP3 formation and the resultant mobilization of Ca2/ from intracellular Ca2/ stores via InsP3R.
297
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
The initial Ca2/ mobilization is transient, and this is then followed by a sustained [Ca2/]i elevation resulting from Ca2/ influx, which is also highly sensitive to Zn2/ (4). It is therefore likely that PC12 cells possess a Ca2/ entry pathway(s) which can be activated by Ca2/ store depletion induced by the activation of either InsP3 or RyR. We demonstrated here a possible link between CCE and exocytosis in PC12 cells. As shown in Fig. 1, caffeine could stimulate Ca2/ release in the absence of extracellular Ca2/, and the amplitude was almost the same as that in the presence of extracellular Ca2/, but secretion could be produced only in the presence of extracellular Ca2/ (Fig. 2). This property described above is in agreement with that of the UTP-evoked responses in the cells. UTP evoked the release of dopamine only in the presence of extracellular Ca2/ (4), indicating that the UTP-evoked release of dopamine was not due to Ca2/ release via InsP3Rs but due to the Ca2/ entry. In addition, the release of dopamine was mimicked by CPA, which was totally dependent on extracellular Ca2/. All these findings support the hypothesis that CCE and secretion are functionally linked in PC12 cells. Moreover, both the sustained [Ca2/]i rise and the release of dopamine evoked by caffeine were inhibited by Zn2/, an inhibitor of CCE (9), without affecting the transient rise in [Ca2/]i. These similarities also support the hypothesis. Caffeine has various non-specific actions, such as the inhibition of phosphodiesterase and adenosine receptors. This raises the possibility that caffeine evoked CCE not by Ca2/ release through RyRs directly but as an indirect result of elevating cAMP or inhibiting adenosine receptors. Furthermore, the caffeine-evoked secretion may not be due to CCE but due to such nonspecific actions of caffeine. However, both CPA and thapsigargin produced a sustained rise in [Ca2/]i which overlapped with that evoked by caffeine (Fig. 1f and g). Theophylline also produced a biphasic [Ca2/]i rise only in the presence of extracellular Ca2/. Forskolin had no effect on either the resting or the theophylline-evoked rise in [Ca2/]i. Aminophylline (100 mM), an antagonist of non-selective adenosine receptors, had no effect on the caffeine-evoked changes in [Ca2/]i. These findings argue against the involvement of either cAMP-dependent or adenosine receptor-mediated mechanisms in the triggering of caffeine-evoked Ca2/ entry. With re-
gard to secretion, the CPA-evoked release of dopamine was totally dependent upon the presence of extracellular Ca2/. UTP, which does not affect cAMP formation or adenosine receptors, can also stimulate the release of dopamine only in the presence of extracellular Ca2/ (4). Aminophylline had no effects on the release of dopamine evoked by caffeine. These findings strongly suggest that neither cAMP-dependent mechanisms nor the inhibition of adenosine receptors contribute to the caffeine-evoked secretion in the cells. We have shown that CCE is present in an excitable cell line, PC12 cells, and that it can be stimulated by both InsP3R- (4) and RyR-activation. Furthermore, CCE has an indispensable role in the secretory response evoked by activation of either InsP3R- or RyR. Although the mechanisms by which RyR activation triggers CCE and CCE promotes the secretory response remain to be examined, our present results indicate a new direction for the study of CCE in relation to secretion. ACKNOWLEDGMENTS We thank Mr. K. Tanaka for skillful assistance, Ms. T. Obama for cell culture, Dr. J. Kenimer for improving the manuscript and Dr. Y. Ohno for continuous encouragement. We are grateful to Dr. T.R. Cheek for much helpful advice and for improving the manuscript. We also grateful to Prof. M.J.Berridge for the provision of laboratory equipment and space, and for reading the manuscript, and Dr. M.D. Bootman for helpful comments. This work was partly supported by the Japan Health Science Foundation.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
Putney, J. W., Jr. (1986) Cell Calcium 7, 1–7. Berridge, M. J (1995) Biochem. J. 312, 1–11. Clapham, D. E. (1995) Cell 80, 259–268. Koizumi, S., Nakazawa, K., and Inoue, K. (1995) Br. J. Pharmacol. 115, 1502–1508. Barry, V. A., and Cheek, T. R. (1994) Biochem. J. 300, 589–597. Inoue, K., and Kenimer, J. G. (1988) J. Biol. Chem. 263, 8157– 8161. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) J. Biol. Chem. 260, 3440–3450. Koizumi, S., Watano, T., Nakazawa, K., and Inoue, K. (1994) Br. J. Pharmacol. 112, 992–997. Hoth, M., and Penner, R. (1993) J. Physiol. (Lond.) 465, 359– 386. Endo, M. (1985). Curr. Top. Membr. Transp. 25, 181–230. Cheek, T. R., Barry, V. A., Berridge, M. J., and Missiaen, L. (1991) Biochem. J. 275, 697–701.
298
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
244, 312–316 (1998)
RC978051
Hydrocortisone Reinforces the Blood–Brain Barrier Properties in a Serum Free Cell Culture System Dirk Hoheisel, Thorsten Nitz, Helmut Franke, Joachim Wegener, Ansgar Hakvoort, Thomas Tilling, and Hans-Joachim Galla Institut fu¨r Biochemie, Westfa¨lische Wilhelms-Universita¨t, Mu¨nster, Germany
Received December 9, 1997
The increasing number of newly developed drugs demands for functional in vitro models of the blood-brain barrier to determine their brain uptake. Cultured cerebral capillary endothelial cells are considered to be such a model, however in serum containing media they exhibit low electrical resistances and high permeabilities compared to the in vivo situation. Here we report the establishment of a serum-free cell culture model. Withdrawal of serum already caused a twofold increase of transendothelial resistance (TER), which in presence of serum is about 100-150 Vrcm2. We tested several supplements and found that hydrocortisone is a potent stimulator for the formation of barrier properties. TERs up to 1000 Vrcm2 were measured in the presence of physiological relevant hydrocortisone concentrations. In correspondence to the TER increase hydrocortisone decreased cell monolayer permeability for sucrose down to 5r1007 cm/s, which is close to the in vivo value of 1.2r1007 cm/s and by a factor of five lower compared to cultures without hydrocortisone and in presence of serum. q 1998 Academic Press
Combinatorial chemistry is able to yield high numbers of compounds of pharmaceutical interest. An important aspect for a new drug is to know its availability in the nervous system, which means the ability to cross the blood-brain barrier (BBB). Thus the permeability of thousands of compounds will have to be screened in the near future. This will not be possible in vivo, so that powerful in vitro models of the BBB are demanded, that mimic the differentiated BBB at least with respect to its barrier properties. In the last years preparation methods and cell culture systems of brain capillary endothelial cells (BCEC), which build up the BBB, were established and in vitro models of the barrier were developed [1,
2, 3]. Many of these in vitro models are able to mimic the in vivo situation very well, but for the best cell culture systems it was until now necessary to grow BCEC in co-cultures with astrocytes [2] or to apply a combination of astrocyte-conditioned medium and agents that elevate intracellular cAMP [3]. All of them used serum in the cell culture medium, which deteriorates the reproducibility of experiments and the analysis of permeabilities. Here we report the establishment of a cell culture system for BCEC with low sucrose permeability and high electrical resistance, which is not dependent on a co-culture with astrocytes or astrocyte-conditioned medium and does not require the use of serum. Hydrocortisone supplementation was found to be essential and sufficient to induce barrier properties in vitro. Maintenance of the barrier lasts for several days, which is long enough for pharmaceutical screening. METHODS The cells were isolated from freshly slaughtered pigs by several enzymatic digestion and centrifugation steps according to a modified method of Bowman et al. [4], as described previously [5]. In short, after removal of the meninges and the secretory areas, the grey and white matter of the brain cortex was minced using a sterile cutter with staggered rolling blades. The material was suspended in DMEM/Ham’s F12 (Biochrom, Berlin, Germany) and incubated with dry powdered Dispase II from bacillus polymyxa (1% (w/v); Boehringer, Mannheim, Germany) for about 3 h at 377C. A dextran solution (mw Ç 162000, 18% (w/v); Sigma, Deisenhofen, Germany) was added to get a final 10.8% (w/v) suspension, and the suspension was centrifuged at 6800 g for 10 min at 47C. To separate larger vessels the pellet was resuspended and filtered through a 180 mm nylon sieve (ZBF, Zu¨rich, Switzerland). The capillaries were incubated with 0.1% (w/v) collagenase/dispase II (vibrio alginolyticus/bacillus polymyxa; Sigma, Deisenhofen, Germany) at 377C for 2-3 h under gentle stirring by a hanging magnetic stirrer. After collecting the released cell aggregates by low spin centrifugation (140 g, 10 min, 207C) they were further purified by density gradient centrifugation. For this step the yield of 1-2 brains was resuspended in 10 ml M199 and
312
AID
BBRC 8653
/
6955$$$$83
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. TER of PBCEC monolayers during the course of culture time. j, TER of PBCEC monolayers, which were cultured without serum but with 550 nM hydrocortisone. l, TER of PBCEC monolayers, which were cultured with 10% (v/v) ox serum but without hydrocortisone. DIV: days in vitro. Data are given as mean { se (n Å 5).
centrifuged on a discontinuous Percoll-gradient (15 ml 1.07 g/cm3; 20 ml 1.03 g/cm3; Sigma, Deisenhofen, Germany) at 1300 g for 10 min in a swinging bucket rotor. The cell clusters of the PBCEC, which are gathered at the interface, were washed in DMEM/Ham’s F12 and centrifuged at 140 g for 10 min. Cell clusters from one brain were sown on 450 cm2 collagen G (Seromed, Berlin, Germany) coated culture surface. Cultured BCEC were characterized by their typical spindle shaped morphology, their expression of von Willebrand factor and their high specific activity of alkaline phosphatase and g-glutamyl-transpeptidase. They were grown in DMEM/Ham’s F12 containing 4 mM glutamine (Seromed, Berlin, Germany), 10% (v/v) ox serum (PAA, Linz, Austria), penicillin/streptomycin (100 mg/ ml) and gentamicin (100 mg/ml) (both Sigma, Deisenhofen, Germany). In order to prepare confluent cell monolayers, primary cultured BCEC were subcultivated after three days in vitro by sowing them again on rat tail collagen coated filter inserts (1.5r105 cells/ cm2). After another day in vitro medium was exchanged by DMEM/ Ham’s F12 containing different supplements. The BBB-features of cultured cell monolayers were scrutinized by determination of transendothelial electrical resistance (TER) with AC impedance analysis [6] and by measuring the permeability coefficient of [14C]sucrose (Amersham, Buckinghamshire, UK).
transendothelial electrical resistance (TER) were performed at 7-8 DIV. EGF between 2-8 nM showed no effect on TER. Insulin up to 700 pM slightly improved TER, but for a significant effect it had to be added in an unphysiological high concentration up to 70 mM, which increased the TER by a factor of two. However since these concentrations were more than 10.000 times higher than the physiological concentration in the human bloodstream (70-700 pM [7]) insulin has to be considered as non effective with respect to barrier formation. Addition of hydrocortisone to the incubation medium caused drastic increase of the TER of PBCEC monolayers shown in Fig. 1. Under serum free culture conditions the TER of PBCEC monolayers, which were incubated with hydrocortisone, was up to 2-3 times higher than TER in hydrocortisone free medium. The maximal effect of hydrocortisone was reached at a concentration of 70 nM or more (Fig. 2), which is well in the physiological concentration range of hydrocortisone in the human bloodstream between 70-550 nM [7]. In the following experiments we thus used 550 nM hydrocortisone in the incubation medium to mimic the in vivo situation as close as possible. In the experiments shown in Fig. 2 PBCEC monolayer developed a maximum TER of about 700 Vrcm2. This is a typical preparation with a reliable and clearly reproducible effect. The absolute maximum TER values we observed varied and depended on the primary culture of PBCEC. Some preparations of PBCEC yielded cells, which were able to build up dense cell monolayers with TER values up to 1000 Vrcm2 after incubation with 550 nM hydrocortisone (e.g. Fig. 1). On the other hand we were confronted with some primary cultures that only developed TERs of 300-500 Vrcm2, which is still a good result. Combinations of hydrocortisone, EGF and Insulin were not different in their effect compared to hydrocortisone alone.
RESULTS In order to improve the barrier characteristics of PBCEC monolayers the influence of hormones or growth factors like insulin, EGF and hydrocortisone was studied. Normal culture medium containing serum was exchanged one day after subcultivation by fresh medium, which contained hormones or growth factors (incubation medium). In general PBCEC were able to build up dense cell monolayers with electrical resistances two days after subcultivation (Fig. 1). Highest barriers properties were detected four to five days after subcultivation, which is 7-8 days in vitro (DIV). In DMEM/Ham’s F12 with all three supplements TER increased considerably to up to 1000 Vrcm2 at 7-8 DIV. Therefore, further permeability studies or analysis of
FIG. 2. Influence of hydrocortisone on the TER of PBCEC monolayers. Analysis of the TER was performed after 7 days in vitro. Data are given as mean { se (n Å 5).
313
AID
BBRC 8653
/
6955$$$$84
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
DISCUSSION
FIG. 3. Influence of hydrocortisone and ox serum on the barrier properties of PBCEC monolayers. Experiments were performed after 7 days in vitro. Incubation medium: 1, with 10% (v/v) ox serum, without hydrocortisone; 2, with 10% (v/v) ox serum, with 550 nM hydrocortisone; 3, without serum, without hydrocortisone; 4, without serum, with 550 nM hydrocortisone. A: Analysis of the TER. Relative data are given as mean { se (n Å 36). B: Sucrose-permeability. Relative data are given as mean { se (n Å 8).
Fig. 3 summarizes the effect of serum and hydrocortisone on TER and [14C]-sucrose permeability of PBCEC in culture. Relative values are given considering the broad distribution of initial TERs obtained during different preparations. Cell monolayers in the absence of serum and without hydrocortisone were taken as reference and set to 100% (column 3 in Fig. 3). Serum reduced the TER to 50% compared to the reference and increased the permeability of sucrose by about a factor of 3 (column 1 in Fig. 3). If hydrocortisone was added to the medium in presence of ox serum (column 2 in Fig. 3) the TER-reduction was only 30% and the permeability increase of sucrose only came up to 185 % of the reference value without serum and hydrocortisone. In absence of serum hydrocortisone increased the resistance to 250% and permeability of sucrose was decreased correspondingly to 30% of the reference value. Typical permeability values were 1.8r1006 cmrs01 in the absence of serum and hydrocortisone. 4.0 1 1006 cmrs01 in the presence of serum and with hydrocortisone and 0.5 1 1006 cmrs01 after supplementation of serum free medium by 550 nM hydrocortisone.
We were able to establish a primary endothelial cell culture system, which is a simple and reproducible model for the BBB and closely mimics the in vivo situation. In the absence of serum and in the presence of hydrocortisone the permeability of sucrose across a PBCEC monolayer was found to be only 0.5-1 1 1006 cmrs01 and their TER values reached 1000 Vrcm2. These TER values are comparable to the TER of brain capillaries in vivo, which was estimated to 1900 Vrcm2 [8]. One outstanding feature of our cell culture model is the possibility to perform experiments under serum free conditions. Serum free culture conditions are the basis for reproducible transport studies and facilitates investigations regarding cell-cell interaction at the BBB. The identification of BBB-inducing agents, e.g. in co-cultures, will now be enabled. Therefore, our in vitro model will be a good basis to determine molecules like intercellular messengers, which possibly are secreted from astrocytes or pericytes. To our knowledge this is the first time that a negative effect of serum was observed on the barrier properties of an endothelial cell monolayer. This seems to be feasible since some years ago a decrease of the electrical resistance had been observed at cell monolayers of an epithelial cell line, the MDCK cells, under the influence of serum [9, 10]. The opening of tight junctions by serum has also been observed in retinal epithelial cells [11]. The negative effect of serum on BBB features could be explained by the presence of serum growth factors. These agents inducing proliferation are not further needed in confluent cell culture, and they could rather hinder final differentiation of the cells. Cytokines or hormones, which are always present in the blood of the donor animal, could be another possibility of the barrier weakening effect of serum. The molecular structure of these serum factors however has to be determinated. Hydrocortisone was the only supplement found so far to improve barrier properties. The serum free cell culture model reported here could be improved in such a way, that TER and sucrose permeability came close to the corresponding parameters in vivo. With respect to the in vivo situation it is important to note that hydrocortisone was added to the cell culture in a physiological concentration. Hydrocortisone now is a glucocorticoid and influences the metabolism of many cell types preferentially in the catabolic pathways. Beside this an anti inflammatory effect was described [7]. Moreover, Zettl et al. [12] were able to show that dexamethasone, a chemically modified hydrocortisone, improved TER of monolayers of 31EG4 cells. Forthcoming analysis of the intracellular signal transduction pathway of hydrocortisone will provide new insights in the
314
AID
BBRC 8653
/
6955$$$$84
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
mechanisms regulating the structural organisation of tight junctions. In summary with help of hydrocortisone we were able to establish an in vitro model of the BBB, which is independent of serum, astrocyte conditioned medium or any other undefined additives. Our model is close to the in vivo situation and could therefore be the basis for serial permeability studies of pharmaceutical active substances under defined conditions. ACKNOWLEDGMENTS This work has been financially supported by a grant from the Deutsche Forschungsgemeinschaft and is a contribution from the Graduiertenkolleg: ‘‘Membranproteine: Signalerkennung, Signaltransfer und Stofftransport.’’
REFERENCES 1. Meyer, J., Rauh, J., and Galla, H. J. (1991) J. Neurochem. 57, 1971–1977. 2. Dehouck, M. P., Jolliet-Riant, P., Bre´e, F., Fruchart, J. C., Cecchelli, R., and Tillement, J. P. (1992) J. Neurochem. 58, 1790– 1797.
3. Rubin, L. L., Hall, D. E., Porter, S., Barbu, K., Cannon, C., Horner, H. C., Janatpour, M., Liaw, C. W., Manning, K., Morales, J., Tanner, L. I., Tomaselli, K. J., and Bard, F. (1991) J. Cell Biol. 115, 1725–1735. 4. Bowman, P. D., Ennis, S. R., Rarey, K. E., Betz, A. L., and Goldstein, G. W. (1989) Ann. Neurol. 14, 396–402. 5. Tewes, B., Franke, H., Hellwig, S., Hoheisel, D., Decker, S., Griesche, D., Tilling, T., Wegener, J., and Galla, H. J. (1997) in Drug Transport across the Blood–Brain Barrier (De Boer, B., and Sutanto, W., Eds.), Harwood Academic Publishers, Amsterdam. 6. Wegener, J., Sieber, M., and Galla, H. J. (1996) J. Biochem. Biophys. Methods 32, 157–170. 7. Karlson, P., Doenecke, D., and Kooman, J. (1994) Biochemie fu¨r Mediziner und Naturwissenschaftler, Georg Thieme Verlag, Stuttgart. 8. Crone, C., and Oleson, S. P. (1982) Brain Res. 241, 49–55. 9. Conyers, G., Milks, L., Conklyn, M., Showell, H., and Cramer, E. (1990) Am. J. Physiol. 259, C577–C585. 10. Marmorstein, A. D., Mortell, K. H., Ratcliffe, D. R., and Cramer, E. B. (1992) Am. J. Physiol. 262, 1403–1410. 11. Chang, Ch., Wang, X., and Caldwell, R. B. (1997) J. Neurochem. 69, 859–867. 12. Zettl, K. S., Sjaastad, M. D., Riskin, P. M., Parry, G., Machen, T. E., and Firestone, G. L. (1992) Proc. Natl. Acad. Sci. USA 89, 9069–9073.
315
AID
BBRC 8653
/
6955$$$$84
06-03-98 14:51:53
bbrcg
AP: BBRC
247, 536 (1998)
ERRATUM Volume 244, Number 1 (1998), in Article No. RC988251, ‘‘Functional Coupling of Secretion and Capacitative Calcium Entry in PC12 Cells,’’ by Schuichi Koizumi and Kazuhide Inoue, pages 293–297, and in Article No. RC978051, ‘‘Hydrocortisone Reinforces the Blood–Brain Barrier Properties in a Serum Free Cell Culture System,’’ by Dirk Hoheisel, Thorsten Nitz, Helmut Franke, Joachim Wegener, Ansgar Hakvoort, Thomas Tilling, and Hans-Joachim Galla, pages 312–316: Due to a compositor’s error, the figures for these two articles were inadvertently switched. The legends for the figures are correct as printed. For the reader’s convenience, both complete articles with the correct figures are printed here. This erratum is Article No. RC988653.
536 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 8653
/
6955$$$$81
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
244, 293–297 (1998)
RC988251
Functional Coupling of Secretion and Capacitative Calcium Entry in PC12 Cells Schuichi Koizumi*,†,1 and Kazuhide Inoue* *Division of Pharmacology, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya, Tokyo 158, Japan; and †Laboratory of Molecular Signalling, The Babraham Institute, Babraham Hall, Cambridge, CB2 4AT, United Kingdom
Received January 26, 1998
The caffeine-evoked effects on the intracellular Ca2/ concentration ([Ca2/]i) and on the release of dopamine by PC12 cells were investigated. Stimulation by caffeine resulted in a transient Ca2/ release which was followed by a sustained phase of Ca2/ entry through a non-voltage dependent pathway. Treatment with cyclopiazonic acid (CPA) or thapsigargin, inhibitors of the Ca2/ATPase pump of the endoplasmic reticulum, resulted in only a sustained rise in [Ca2/]i in the presence of extracellular Ca2/. Pretreatment of cells with CPA or thapsigargin abolished the subsequent Ca2/ responses to caffeine. Caffeine also evoked the release of dopamine from the cells only in the presence of extracellular Ca2/, which was mimicked by CPA. These results suggest that store-dependent Ca2/ entry evoked by caffeine has an indispensable role in the secretory response in an excitable cell line, PC12 cells. q 1998 Academic Press
Over ten years have passed since the original idea that Ca2/ can enter cells through the plasma membrane in response to intracellular store depletion was first proposed (1). Recently, much attention has been focused on this so-called capacitative Ca2/ entry (CCE) because it has become clear that CCE plays a central role in many aspects of cell signaling (2, 3). In excitable cells, however, the existence of CCE is controversial, in part because it is unclear whether ryanodine receptor (RyR)-linked Ca2/ stores activate CCE in a manner similar to that which results from depletion of inositol 1,4,5-trisphosphate receptor (InsP3R)-linked stores (2). 1 Corresponding author. Fax: 03-3700-9698. E-mail: inoue@ nihs.go.jp. Abbreviations: [Ca2/]i, intracellular Ca2/ concentration; CCE, capacitative Ca2/ entry; CPA, cyclopiazonic acid; InsP3 , inositol 1,4,5trisphosphate; RyR, ryanodine receptor.
We previously reported that activation of P2U-purinoceptors in an excitable cell line, rat pheochromocytoma PC12 cells, by uridine 5*-triphosphate (UTP) leads to a transient rise in the intracellular Ca2/ concentration ([Ca2/]i) resulting from inositol 1,4,5-trisphosphate (InsP3)-mediated Ca2/ mobilization, which was followed by a sustained rise in [Ca2/]i resulting from nonvoltage dependent Ca2/ influx, presumably CCE (4). Besides InsP3-sensitive Ca2/ stores, caffeine/ryanodine sensitive Ca2/ stores are present in PC12 cells (5), which enables us to investigate whether RyR-linked Ca2/ stores activate CCE via an InsP3-independent mechanisms. As for the physiological significance of CCE in excitable cells, it is far from clear. UTP could stimulate the release of dopamine only in the presence of extracellular Ca2/ in PC12 cells (4), which raises the possibility that CCE activated by InsP3R-mediated Ca2/ release and the secretory response are closely linked. In this study, we demonstrate the existence of CCE in an excitable cell line, PC12 cells, and report that, in addition to InsP3Rs activation (4), CCE can be stimulated by activation of RyRs. Furthermore, we also show the role played by CCE in the exocytotic secretory response of the cells. MATERIAL AND METHODS Cell culture. Culture conditions of PC12 cells were as described previously (6). All experiments described in this manuscript were performed with cells at passages 53 through 68. Cells were plated onto collagen-coated 35 mm polystyrene dishes (1 1 106 cells/dish) for measuring the release of dopamine, or poly-L-lysine (Sigma, St. Louis, U.S.A.)-coated glass coverslips, placed in silicon rubber walls (Flexiperm, W.C. GmbH, Germany) for measuring the increase in [Ca2/]i. Cells were cultured an additional 2 days in a humidified atmosphere of 90 % air and 10 % CO2 at 37 7C. Measurement of [Ca2/]i in single cells. Changes in [Ca2/]i were measured by the fura-2 method as described previously (7) with minor modifications (4). In brief, 30 min after incubation with 5 mM
293
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. The time-course of the caffeine-evoked increase in [Ca2/]i in PC12 cells. a and b. Typical Ca2/ responses to 30 mM caffeine in the presence (a) and absence (b) of extracellular Ca2/. Caffeine was applied to the cells for 1 min (horizontal bars) at 4 min intervals. Traces a and b were obtained from the same cell. Vertical and horizontal scale bars show 100 nM and 1 min, respectively. c. The traces shown in a and b are superimposed and their time-courses were compared. The maximal Ca2/ responses to caffeine (hmax) are summarized in d (control, nÅ36; 0Ca2/, nÅ42). #s show the magnitude of [Ca2/]i (h#) at 1 min after caffeine application in the presence (control) and absence (0Ca2/) of extracellular Ca2/, and the responses are summarized in e. Values show the ratio of h# over hmax in the presence (nÅ36; open column) and absence (nÅ42; dotted column) of extracellular Ca2/. f and g. The effects of depletion of intracellular Ca2/ stores by CPA (cyclopiazonic acid)(f) and thapsigargin (g), on the caffeine-evoked rise in [Ca2/]i. After caffeine (30 mM) stimulation, cells were incubated with CPA (30 mM; hatched column) for 3 min (f), or thapsigargin (1 mM; dotted column) for 2 min (g), and then caffeine was applied to the 294
AID
BBRC 8653
/
6955$$8653
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
fura-2 acetoxymethylester (fura-2 AM) at 37 7C, cells were washed with a balanced salt solution (BSS) of the following composition (mM): NaCl 150, KCl 5.0, CaCl2 1.8, MgCl2 1.2, D-glucose 10 and N2-hydroxyethylpiperazine-N*-2-ethanesulfonic acid (HEPES) 25 (pH adjusted to 7.4 with NaOH) and mounted on an inverted fluorescence microscope (TMD-300, Nikon, Tokyo, Japan) equipped with a Xenonlamp and band-pass filters of 340 and 360 nm. For the Ca2/-depleted experiments, we used a medium in which Ca2/ was removed and 1 mM ethylenediaminetetraacetic acid (EGTA) was added (Ca2/-free BSS). Image data, recorded by a high-sensitivity silicon intensifier target camera (C-2741-08, Hamamatsu Photonics, Hamamatsu, Japan), were processed by a Ca2/-analyzing system (Furusawa Lab. Appliance. Co., Kawagoe, Japan). Caffeine was applied to the cells two or three times for 1 min at intervals of 4 min. The absolute [Ca2/]i was estimated from the ratio of emitted fluorescence (F340/ F360) according to the calibration curve obtained by standard Ca2/buffer. Measurement of dopamine release. The procedures for the measurement of released dopamine were basically the same as those described by Koizumi et al. (8). Cells were stimulated by various concentrations of caffeine and CPA dissolved in BSS for 1 and 3 min, respectively. For Ca2/-free experiments, the dishes were washed twice with Ca2/-free BSS for 1 min before caffeine application. The amount of dopamine released into the superfusate and that remaining in the cells were determined with high performance liquid chromatography (HPLC) coupled with an electrochemical detector (ECD)(LC-4B, Bioanalytical systems, West Lafayette, U.S.A.). The percentage of release was calculated by dividing the supernatant values by the sum of the supernatant and pellet values. Chemicals. Drugs used were as follows. Caffeine, cyclopiazonic acid (CPA), thapsigargin, nicardipine hydrochloride, cadmium chloride, zinc acetate and v-conotoxin GVIA (v-CTX) were purchased from Sigma. Fura-2 AM and HEPES were from (Dojin, Kumamoto, Japan). Other chemicals are purchased from Wako Purechemicals (Tokyo, Japan). Nicardipine, CPA and thapsigargin were dissolved in dimethyl sulphoxide at a concentration of 10 mM (nicardipine and CPA) or 1 mM (thapsigargin), and then dissolved in BSS to appropriate concentrations. Other drugs were directly dissolved in BSS or Ca2/-free BSS. All data are mean{s.e.m. Statistics. Statistical differences in the values of dopamine release or the increase in [Ca2/]i were determined using an analysis of variance followed by Dunnet’s test for multiple comparisons.
RESULTS Caffeine (30 mM) produced a significant rise in [Ca2/]i in about 80 % of the PC12 cells (188 out of 231 cells tested) and the average amplitude was 312{40.1 nM. When caffeine was applied to the cells twice for 1 min separated by 3 min, the Ca2/ response to the 2 nd caffeine was almost the same as that of the 1 st caffeine (data not shown). Fig. 1 (a, b and c) shows the timecourse of the caffeine-evoked increase in [Ca2/]i in the cells. Caffeine evoked a transient [Ca2/]i rise in the absence of extracellular Ca2/ (b), whereas it produced a transient [Ca2/]i rise, followed by a sustained rise in [Ca2/]i, in the presence of extracellular Ca2/ (a). The individual traces were superimposed and the time-
courses were compared (Fig. 1c). The maximal [Ca2/]i level evoked by caffeine (hmax) in the absence of extracellular Ca2/ was comparable to that in the presence of external Ca2/ (Fig. 1d). The [Ca2/]i at 1 min after the caffeine application (marked #) was defined as h# and the ratio of h#/hmax was calculated in the absence and presence of extracellular Ca2/. As shown in Fig. 1e, the h#/hmax ratio was dramatically decreased by the depletion of extracellular Ca2/, suggesting that the sustained component results from Ca2/ influx from extracellular spaces. Incubation of the cells with 30 mM CPA (Fig. 1f) or 1 mM thapsigargin (Fig. 1g), inhibitors of the Ca2/ATPase pump of the endoplasmic reticulum, resulted in a sustained rise in [Ca2/]i, which almost completely inhibited the subsequent Ca2/ responses to caffeine (30 mM). Thapsigargin evoked only a transient and slight rise in [Ca2/]i in the absence of extracellular Ca2/, but induced a long-lasting [Ca2/]i elevation upon replacement of the Ca2/-free medium with 1.8 mM Ca2/ (Fig. 1h). When the cells were repetitively exposed to 30 mM caffeine for periods of 1 min separated by 3 min in the presence of 10 mM ryanodine, the Ca2/ responses to caffeine were dramatically inhibited in a use-dependent manner and the 3 rd series of Ca2/ responses to caffeine were almost all abolished (data not shown). Similar to caffeine, theophylline (30 mM) produced a biphasic rise in [Ca2/]i, i.e. a transient elevation in [Ca2/]i was followed by a sustained one (data not shown). The sustained [Ca2/]i rise was totally dependent on extracellular Ca2/. Forskolin (10 mM) had no effects on either the resting [Ca2/]i or the theophyllineevoked rise in [Ca2/]i (nÅ48). Pretreatment of cells with nicardipine (30 mM) / v-conotoxin (v-CTX, 1 mM) for 1 min before and during the theophylline application had no effects on the sustained Ca2/ responses to theophylline (the amplitude of the sustained response 1 min after theophylline application in the presence of nicardipine / v-CTX was 104.4{9.1 % of theophylline alone, nÅ32), though these chemicals inhibited the KCl (53 mM)-evoked rise in [Ca2/]i by 82.4{6.2% (nÅ26). The effects of various compounds on the rise in [Ca2/]i were examined (Fig. 2). Zn2/, an inhibitor of CCE (9), almost completely inhibited the sustained component of the [Ca2/]i rise without affecting the transient component (Fig. 2b, d and e). Traces from Fig. 2a and b were superimposed and the time-courses were compared (Fig. 2c). The hmax in the presence of Zn2/ was not different from that in the absence of extracellular Zn2/ (Fig. 2c and d). However, Zn2/ at 30 mM dramatically inhibited the sustained component of the [Ca2/]i rise and significantly inhibited the h#/hmax ratio
cells in the presence of CPA (nÅ47) or thapsigargin (nÅ31). h. Cells were incubated with thapsigargin (1 mM; dotted column) in the absence of extracellular Ca2/ (broken line) for 2 min, which resulted in the abolishment of the caffeine-evoked rise in [Ca2/]i. Subsequently applied external Ca2/ produced a large and sustained [Ca2/]i increase (nÅ28). Vertical and horizontal scale bars show 100 nM and 1 min, respectively. 295
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. The effects of Zn2/, Cd2/, La3/, v-CTX and nicardipine on the caffeine-evoked rise in [Ca2/]i in PC12 cells. a and b. Typical Ca2/ responses to 30 mM caffeine in the absence (a) and presence (b) of 30 mM Zn2/. Caffeine was applied to the cells two times for 1 min at 4 min intervals (filled horizontal bars). Zn2/ (dotted column) and caffeine were applied to the cells simultaneously (b). Vertical and horizontal bars show 100 nM and 60 s, respectively. The traces shown in a and b were superimposed and their time-courses were compared in c. The maximal Ca2/ responses to caffeine (hmax) in the absence (open column) and presence (filled column) of 30 mM Zn2/ are summarized in d. #s show the magnitude of [Ca2/]i (h#) at 1 min after caffeine application in the absence (control, nÅ143) and presence (/Zn2/ 30 mM, nÅ37-58) of Zn2/, and the responses are summarized in e. The values show the ratio of h# over hmax in the absence (nÅ143; open column) and presence (nÅ37-58; filled column) of Zn2/. In addition to Zn2/, the effects of Cd2/ (nÅ48-61; hatched columns), La3/ (nÅ68; double hatched column), v-CTX (nÅ49, dotted column) and nicardipine (nÅ41; striped column) on the ratio of h#/hmax were examined and summarized in e. Asterisks show a significant difference from the ratio of control (*põ0.05, **põ0.01).
(Fig. 2e). Other cations (°100 mM) failed to inhibit the sustained component significantly (Fig. 2e) though La3/ showed inhibitory actions when its concentration was raised to 300 mM (h#/hmax ratio: 0.14{0.04, põ0.01,
nÅ38). Incubation of cells with nicardipine (30 mM) or v-CTX (1 mM) for 1 min before and during the caffeine stimulation had no effect on the sustained Ca2/ response to caffeine. Aminophylline (100 mM) had no ef-
296
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
PC12 cells, caffeine induces Ca2/ influx by a mechanism dependent upon Ca2/ depletion. Fig. 3a shows the concentration-dependence of the caffeine-stimulated release of dopamine and the effects of extracellular Ca2/ on the stimulation. Caffeine stimulated the release of dopamine in a concentration-dependent manner over a concentration range from 10 to 50 mM only in the presence of extracellular Ca2/ (Fig. 3a). The release of dopamine was mimicked by 30 mM CPA, and this was totally dependent on extracellular Ca2/ [CPA: 313.4{16.3 (nÅ6, /Ca2/), 102.2{6.2 (nÅ6, 0Ca2/) % of spontaneous release]. Aminophylline, an antagonist of non-selective adenosine receptors, had no effects on the release of dopamine evoked by caffeine (aminophylline 100 mM: 102.1{6.5 % of caffeine alone, nÅ6). The effects of various cations and an organic compound on the caffeine-evoked dopamine release were examined in the PC12 cells (Fig. 3b). Zn2/ potently inhibited the release of dopamine with an IC50 value of about 4 mM. Other cations or nicardipine failed to inhibit the dopamine release significantly. The effects of these chemicals on the release were well in agreement with those of the chemicals on the sustained [Ca2/]i rise evoked by caffeine (Fig. 2 and 3b). DISCUSSION
FIG. 3. a. Concentration-dependence of the caffeine-evoked dopamine release from PC12 cells. Open and closed circles represent the release of dopamine evoked by caffeine in the presence (open circles) and absence (closed circles) of extracellular Ca2/. These are results from a typical experiment with each data point being the mean{s.e.mean. of triplicate measurements. Four such experiments with similar results were performed. b. The effects of Zn2/, nicardipine, Cd2/ and La3/ on the caffeine-evoked dopamine release from PC12 cells. Values represent % of dopamine release evoked by 30 mM caffeine alone (open circle). The effects of Zn2/, nicardipine, Cd2/ and La3/ on the caffeine-evoked responses are indicated: Zn2/ (closed circles), nicardipine (open triangle), Cd2/ (closed triangles) and La3/ (closed squares). These are results from a typical experiment and each data point is the mean{s.e.mean. of triplicate measurements. At least three such experiments were performed, and similar results were obtained. Asterisks show a significant difference from the response evoked by caffeine alone (*põ0.05, **põ0.01).
fect on the caffeine-evoked sustained rise in [Ca2/]i (h#/ hmax :0.41{0.03, nÅ19). Similar observations were reported in response to UTP (4). In addition, the sustained [Ca2/]i elevation evoked by CPA (30 mM) was also inhibited by Zn2/ (30 mM) but not by Cd2/ (100 mM), La3/ (100 mM) or nicardipine (30 mM) (data not shown). Thus, these findings strongly suggest that, in
Caffeine acts at the RyR where it appears to shift the Ca2/ sensitivity of the channel to a lower concentration (10), thereby increasing the probability of channel opening, namely ‘‘calcium-induced Ca2/ release’’. In the PC12 cells, caffeine evoked Ca2/ release by stimulating RyRs since ryanodine (10 mM) abolished the Ca2/ responses to caffeine in a use-dependent fashion, which was well in accordance with previous results obtained from chromaffin cells (11) and PC12 cells (5). In addition to Ca2/ release from RyR, we showed here that caffeine stimulates a sustained Ca2/ entry in the cells (Fig. 1). The caffeine-evoked Ca2/ entry could be activated by a decrease in the stored Ca2/ concentration because (1) a similarly sustained Ca2/ entry was mimicked by both CPA and thapsigargin, inhibitors of the endoplasmic reticulum Ca2/ATPase pump, (2) pretreatment of cells with CPA or thapsigargin abolished the subsequent Ca2/ responses to caffeine, which indicates an overlap of Ca2/ entry mechanisms and (3) the sustained [Ca2/]i rises evoked by both caffeine and CPA were non-voltage dependent but were sensitive to Zn2/. These results strongly suggest that the caffeine-evoked sustained rise in [Ca2/]i is dependent upon the filling state of the intracellular Ca2/ stores. A similar phenomenon was observed upon UTP stimulation of these cells: UTP stimulates P2U-purinoceptors, which leads to InsP3 formation and the resultant mobilization of Ca2/ from intracellular Ca2/ stores via InsP3R.
297
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
The initial Ca2/ mobilization is transient, and this is then followed by a sustained [Ca2/]i elevation resulting from Ca2/ influx, which is also highly sensitive to Zn2/ (4). It is therefore likely that PC12 cells possess a Ca2/ entry pathway(s) which can be activated by Ca2/ store depletion induced by the activation of either InsP3 or RyR. We demonstrated here a possible link between CCE and exocytosis in PC12 cells. As shown in Fig. 1, caffeine could stimulate Ca2/ release in the absence of extracellular Ca2/, and the amplitude was almost the same as that in the presence of extracellular Ca2/, but secretion could be produced only in the presence of extracellular Ca2/ (Fig. 2). This property described above is in agreement with that of the UTP-evoked responses in the cells. UTP evoked the release of dopamine only in the presence of extracellular Ca2/ (4), indicating that the UTP-evoked release of dopamine was not due to Ca2/ release via InsP3Rs but due to the Ca2/ entry. In addition, the release of dopamine was mimicked by CPA, which was totally dependent on extracellular Ca2/. All these findings support the hypothesis that CCE and secretion are functionally linked in PC12 cells. Moreover, both the sustained [Ca2/]i rise and the release of dopamine evoked by caffeine were inhibited by Zn2/, an inhibitor of CCE (9), without affecting the transient rise in [Ca2/]i. These similarities also support the hypothesis. Caffeine has various non-specific actions, such as the inhibition of phosphodiesterase and adenosine receptors. This raises the possibility that caffeine evoked CCE not by Ca2/ release through RyRs directly but as an indirect result of elevating cAMP or inhibiting adenosine receptors. Furthermore, the caffeine-evoked secretion may not be due to CCE but due to such nonspecific actions of caffeine. However, both CPA and thapsigargin produced a sustained rise in [Ca2/]i which overlapped with that evoked by caffeine (Fig. 1f and g). Theophylline also produced a biphasic [Ca2/]i rise only in the presence of extracellular Ca2/. Forskolin had no effect on either the resting or the theophylline-evoked rise in [Ca2/]i. Aminophylline (100 mM), an antagonist of non-selective adenosine receptors, had no effect on the caffeine-evoked changes in [Ca2/]i. These findings argue against the involvement of either cAMP-dependent or adenosine receptor-mediated mechanisms in the triggering of caffeine-evoked Ca2/ entry. With re-
gard to secretion, the CPA-evoked release of dopamine was totally dependent upon the presence of extracellular Ca2/. UTP, which does not affect cAMP formation or adenosine receptors, can also stimulate the release of dopamine only in the presence of extracellular Ca2/ (4). Aminophylline had no effects on the release of dopamine evoked by caffeine. These findings strongly suggest that neither cAMP-dependent mechanisms nor the inhibition of adenosine receptors contribute to the caffeine-evoked secretion in the cells. We have shown that CCE is present in an excitable cell line, PC12 cells, and that it can be stimulated by both InsP3R- (4) and RyR-activation. Furthermore, CCE has an indispensable role in the secretory response evoked by activation of either InsP3R- or RyR. Although the mechanisms by which RyR activation triggers CCE and CCE promotes the secretory response remain to be examined, our present results indicate a new direction for the study of CCE in relation to secretion. ACKNOWLEDGMENTS We thank Mr. K. Tanaka for skillful assistance, Ms. T. Obama for cell culture, Dr. J. Kenimer for improving the manuscript and Dr. Y. Ohno for continuous encouragement. We are grateful to Dr. T.R. Cheek for much helpful advice and for improving the manuscript. We also grateful to Prof. M.J.Berridge for the provision of laboratory equipment and space, and for reading the manuscript, and Dr. M.D. Bootman for helpful comments. This work was partly supported by the Japan Health Science Foundation.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
Putney, J. W., Jr. (1986) Cell Calcium 7, 1–7. Berridge, M. J (1995) Biochem. J. 312, 1–11. Clapham, D. E. (1995) Cell 80, 259–268. Koizumi, S., Nakazawa, K., and Inoue, K. (1995) Br. J. Pharmacol. 115, 1502–1508. Barry, V. A., and Cheek, T. R. (1994) Biochem. J. 300, 589–597. Inoue, K., and Kenimer, J. G. (1988) J. Biol. Chem. 263, 8157– 8161. Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) J. Biol. Chem. 260, 3440–3450. Koizumi, S., Watano, T., Nakazawa, K., and Inoue, K. (1994) Br. J. Pharmacol. 112, 992–997. Hoth, M., and Penner, R. (1993) J. Physiol. (Lond.) 465, 359– 386. Endo, M. (1985). Curr. Top. Membr. Transp. 25, 181–230. Cheek, T. R., Barry, V. A., Berridge, M. J., and Missiaen, L. (1991) Biochem. J. 275, 697–701.
298
AID
BBRC 8653
/
6955$$$$82
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
244, 312–316 (1998)
RC978051
Hydrocortisone Reinforces the Blood–Brain Barrier Properties in a Serum Free Cell Culture System Dirk Hoheisel, Thorsten Nitz, Helmut Franke, Joachim Wegener, Ansgar Hakvoort, Thomas Tilling, and Hans-Joachim Galla Institut fu¨r Biochemie, Westfa¨lische Wilhelms-Universita¨t, Mu¨nster, Germany
Received December 9, 1997
The increasing number of newly developed drugs demands for functional in vitro models of the blood-brain barrier to determine their brain uptake. Cultured cerebral capillary endothelial cells are considered to be such a model, however in serum containing media they exhibit low electrical resistances and high permeabilities compared to the in vivo situation. Here we report the establishment of a serum-free cell culture model. Withdrawal of serum already caused a twofold increase of transendothelial resistance (TER), which in presence of serum is about 100-150 Vrcm2. We tested several supplements and found that hydrocortisone is a potent stimulator for the formation of barrier properties. TERs up to 1000 Vrcm2 were measured in the presence of physiological relevant hydrocortisone concentrations. In correspondence to the TER increase hydrocortisone decreased cell monolayer permeability for sucrose down to 5r1007 cm/s, which is close to the in vivo value of 1.2r1007 cm/s and by a factor of five lower compared to cultures without hydrocortisone and in presence of serum. q 1998 Academic Press
Combinatorial chemistry is able to yield high numbers of compounds of pharmaceutical interest. An important aspect for a new drug is to know its availability in the nervous system, which means the ability to cross the blood-brain barrier (BBB). Thus the permeability of thousands of compounds will have to be screened in the near future. This will not be possible in vivo, so that powerful in vitro models of the BBB are demanded, that mimic the differentiated BBB at least with respect to its barrier properties. In the last years preparation methods and cell culture systems of brain capillary endothelial cells (BCEC), which build up the BBB, were established and in vitro models of the barrier were developed [1,
2, 3]. Many of these in vitro models are able to mimic the in vivo situation very well, but for the best cell culture systems it was until now necessary to grow BCEC in co-cultures with astrocytes [2] or to apply a combination of astrocyte-conditioned medium and agents that elevate intracellular cAMP [3]. All of them used serum in the cell culture medium, which deteriorates the reproducibility of experiments and the analysis of permeabilities. Here we report the establishment of a cell culture system for BCEC with low sucrose permeability and high electrical resistance, which is not dependent on a co-culture with astrocytes or astrocyte-conditioned medium and does not require the use of serum. Hydrocortisone supplementation was found to be essential and sufficient to induce barrier properties in vitro. Maintenance of the barrier lasts for several days, which is long enough for pharmaceutical screening. METHODS The cells were isolated from freshly slaughtered pigs by several enzymatic digestion and centrifugation steps according to a modified method of Bowman et al. [4], as described previously [5]. In short, after removal of the meninges and the secretory areas, the grey and white matter of the brain cortex was minced using a sterile cutter with staggered rolling blades. The material was suspended in DMEM/Ham’s F12 (Biochrom, Berlin, Germany) and incubated with dry powdered Dispase II from bacillus polymyxa (1% (w/v); Boehringer, Mannheim, Germany) for about 3 h at 377C. A dextran solution (mw Ç 162000, 18% (w/v); Sigma, Deisenhofen, Germany) was added to get a final 10.8% (w/v) suspension, and the suspension was centrifuged at 6800 g for 10 min at 47C. To separate larger vessels the pellet was resuspended and filtered through a 180 mm nylon sieve (ZBF, Zu¨rich, Switzerland). The capillaries were incubated with 0.1% (w/v) collagenase/dispase II (vibrio alginolyticus/bacillus polymyxa; Sigma, Deisenhofen, Germany) at 377C for 2-3 h under gentle stirring by a hanging magnetic stirrer. After collecting the released cell aggregates by low spin centrifugation (140 g, 10 min, 207C) they were further purified by density gradient centrifugation. For this step the yield of 1-2 brains was resuspended in 10 ml M199 and
312
AID
BBRC 8653
/
6955$$$$83
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. TER of PBCEC monolayers during the course of culture time. j, TER of PBCEC monolayers, which were cultured without serum but with 550 nM hydrocortisone. l, TER of PBCEC monolayers, which were cultured with 10% (v/v) ox serum but without hydrocortisone. DIV: days in vitro. Data are given as mean { se (n Å 5).
centrifuged on a discontinuous Percoll-gradient (15 ml 1.07 g/cm3; 20 ml 1.03 g/cm3; Sigma, Deisenhofen, Germany) at 1300 g for 10 min in a swinging bucket rotor. The cell clusters of the PBCEC, which are gathered at the interface, were washed in DMEM/Ham’s F12 and centrifuged at 140 g for 10 min. Cell clusters from one brain were sown on 450 cm2 collagen G (Seromed, Berlin, Germany) coated culture surface. Cultured BCEC were characterized by their typical spindle shaped morphology, their expression of von Willebrand factor and their high specific activity of alkaline phosphatase and g-glutamyl-transpeptidase. They were grown in DMEM/Ham’s F12 containing 4 mM glutamine (Seromed, Berlin, Germany), 10% (v/v) ox serum (PAA, Linz, Austria), penicillin/streptomycin (100 mg/ ml) and gentamicin (100 mg/ml) (both Sigma, Deisenhofen, Germany). In order to prepare confluent cell monolayers, primary cultured BCEC were subcultivated after three days in vitro by sowing them again on rat tail collagen coated filter inserts (1.5r105 cells/ cm2). After another day in vitro medium was exchanged by DMEM/ Ham’s F12 containing different supplements. The BBB-features of cultured cell monolayers were scrutinized by determination of transendothelial electrical resistance (TER) with AC impedance analysis [6] and by measuring the permeability coefficient of [14C]sucrose (Amersham, Buckinghamshire, UK).
transendothelial electrical resistance (TER) were performed at 7-8 DIV. EGF between 2-8 nM showed no effect on TER. Insulin up to 700 pM slightly improved TER, but for a significant effect it had to be added in an unphysiological high concentration up to 70 mM, which increased the TER by a factor of two. However since these concentrations were more than 10.000 times higher than the physiological concentration in the human bloodstream (70-700 pM [7]) insulin has to be considered as non effective with respect to barrier formation. Addition of hydrocortisone to the incubation medium caused drastic increase of the TER of PBCEC monolayers shown in Fig. 1. Under serum free culture conditions the TER of PBCEC monolayers, which were incubated with hydrocortisone, was up to 2-3 times higher than TER in hydrocortisone free medium. The maximal effect of hydrocortisone was reached at a concentration of 70 nM or more (Fig. 2), which is well in the physiological concentration range of hydrocortisone in the human bloodstream between 70-550 nM [7]. In the following experiments we thus used 550 nM hydrocortisone in the incubation medium to mimic the in vivo situation as close as possible. In the experiments shown in Fig. 2 PBCEC monolayer developed a maximum TER of about 700 Vrcm2. This is a typical preparation with a reliable and clearly reproducible effect. The absolute maximum TER values we observed varied and depended on the primary culture of PBCEC. Some preparations of PBCEC yielded cells, which were able to build up dense cell monolayers with TER values up to 1000 Vrcm2 after incubation with 550 nM hydrocortisone (e.g. Fig. 1). On the other hand we were confronted with some primary cultures that only developed TERs of 300-500 Vrcm2, which is still a good result. Combinations of hydrocortisone, EGF and Insulin were not different in their effect compared to hydrocortisone alone.
RESULTS In order to improve the barrier characteristics of PBCEC monolayers the influence of hormones or growth factors like insulin, EGF and hydrocortisone was studied. Normal culture medium containing serum was exchanged one day after subcultivation by fresh medium, which contained hormones or growth factors (incubation medium). In general PBCEC were able to build up dense cell monolayers with electrical resistances two days after subcultivation (Fig. 1). Highest barriers properties were detected four to five days after subcultivation, which is 7-8 days in vitro (DIV). In DMEM/Ham’s F12 with all three supplements TER increased considerably to up to 1000 Vrcm2 at 7-8 DIV. Therefore, further permeability studies or analysis of
FIG. 2. Influence of hydrocortisone on the TER of PBCEC monolayers. Analysis of the TER was performed after 7 days in vitro. Data are given as mean { se (n Å 5).
313
AID
BBRC 8653
/
6955$$$$84
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
DISCUSSION
FIG. 3. Influence of hydrocortisone and ox serum on the barrier properties of PBCEC monolayers. Experiments were performed after 7 days in vitro. Incubation medium: 1, with 10% (v/v) ox serum, without hydrocortisone; 2, with 10% (v/v) ox serum, with 550 nM hydrocortisone; 3, without serum, without hydrocortisone; 4, without serum, with 550 nM hydrocortisone. A: Analysis of the TER. Relative data are given as mean { se (n Å 36). B: Sucrose-permeability. Relative data are given as mean { se (n Å 8).
Fig. 3 summarizes the effect of serum and hydrocortisone on TER and [14C]-sucrose permeability of PBCEC in culture. Relative values are given considering the broad distribution of initial TERs obtained during different preparations. Cell monolayers in the absence of serum and without hydrocortisone were taken as reference and set to 100% (column 3 in Fig. 3). Serum reduced the TER to 50% compared to the reference and increased the permeability of sucrose by about a factor of 3 (column 1 in Fig. 3). If hydrocortisone was added to the medium in presence of ox serum (column 2 in Fig. 3) the TER-reduction was only 30% and the permeability increase of sucrose only came up to 185 % of the reference value without serum and hydrocortisone. In absence of serum hydrocortisone increased the resistance to 250% and permeability of sucrose was decreased correspondingly to 30% of the reference value. Typical permeability values were 1.8r1006 cmrs01 in the absence of serum and hydrocortisone. 4.0 1 1006 cmrs01 in the presence of serum and with hydrocortisone and 0.5 1 1006 cmrs01 after supplementation of serum free medium by 550 nM hydrocortisone.
We were able to establish a primary endothelial cell culture system, which is a simple and reproducible model for the BBB and closely mimics the in vivo situation. In the absence of serum and in the presence of hydrocortisone the permeability of sucrose across a PBCEC monolayer was found to be only 0.5-1 1 1006 cmrs01 and their TER values reached 1000 Vrcm2. These TER values are comparable to the TER of brain capillaries in vivo, which was estimated to 1900 Vrcm2 [8]. One outstanding feature of our cell culture model is the possibility to perform experiments under serum free conditions. Serum free culture conditions are the basis for reproducible transport studies and facilitates investigations regarding cell-cell interaction at the BBB. The identification of BBB-inducing agents, e.g. in co-cultures, will now be enabled. Therefore, our in vitro model will be a good basis to determine molecules like intercellular messengers, which possibly are secreted from astrocytes or pericytes. To our knowledge this is the first time that a negative effect of serum was observed on the barrier properties of an endothelial cell monolayer. This seems to be feasible since some years ago a decrease of the electrical resistance had been observed at cell monolayers of an epithelial cell line, the MDCK cells, under the influence of serum [9, 10]. The opening of tight junctions by serum has also been observed in retinal epithelial cells [11]. The negative effect of serum on BBB features could be explained by the presence of serum growth factors. These agents inducing proliferation are not further needed in confluent cell culture, and they could rather hinder final differentiation of the cells. Cytokines or hormones, which are always present in the blood of the donor animal, could be another possibility of the barrier weakening effect of serum. The molecular structure of these serum factors however has to be determinated. Hydrocortisone was the only supplement found so far to improve barrier properties. The serum free cell culture model reported here could be improved in such a way, that TER and sucrose permeability came close to the corresponding parameters in vivo. With respect to the in vivo situation it is important to note that hydrocortisone was added to the cell culture in a physiological concentration. Hydrocortisone now is a glucocorticoid and influences the metabolism of many cell types preferentially in the catabolic pathways. Beside this an anti inflammatory effect was described [7]. Moreover, Zettl et al. [12] were able to show that dexamethasone, a chemically modified hydrocortisone, improved TER of monolayers of 31EG4 cells. Forthcoming analysis of the intracellular signal transduction pathway of hydrocortisone will provide new insights in the
314
AID
BBRC 8653
/
6955$$$$84
06-03-98 14:51:53
bbrcg
AP: BBRC
Vol. 247, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
mechanisms regulating the structural organisation of tight junctions. In summary with help of hydrocortisone we were able to establish an in vitro model of the BBB, which is independent of serum, astrocyte conditioned medium or any other undefined additives. Our model is close to the in vivo situation and could therefore be the basis for serial permeability studies of pharmaceutical active substances under defined conditions. ACKNOWLEDGMENTS This work has been financially supported by a grant from the Deutsche Forschungsgemeinschaft and is a contribution from the Graduiertenkolleg: ‘‘Membranproteine: Signalerkennung, Signaltransfer und Stofftransport.’’
REFERENCES 1. Meyer, J., Rauh, J., and Galla, H. J. (1991) J. Neurochem. 57, 1971–1977. 2. Dehouck, M. P., Jolliet-Riant, P., Bre´e, F., Fruchart, J. C., Cecchelli, R., and Tillement, J. P. (1992) J. Neurochem. 58, 1790– 1797.
3. Rubin, L. L., Hall, D. E., Porter, S., Barbu, K., Cannon, C., Horner, H. C., Janatpour, M., Liaw, C. W., Manning, K., Morales, J., Tanner, L. I., Tomaselli, K. J., and Bard, F. (1991) J. Cell Biol. 115, 1725–1735. 4. Bowman, P. D., Ennis, S. R., Rarey, K. E., Betz, A. L., and Goldstein, G. W. (1989) Ann. Neurol. 14, 396–402. 5. Tewes, B., Franke, H., Hellwig, S., Hoheisel, D., Decker, S., Griesche, D., Tilling, T., Wegener, J., and Galla, H. J. (1997) in Drug Transport across the Blood–Brain Barrier (De Boer, B., and Sutanto, W., Eds.), Harwood Academic Publishers, Amsterdam. 6. Wegener, J., Sieber, M., and Galla, H. J. (1996) J. Biochem. Biophys. Methods 32, 157–170. 7. Karlson, P., Doenecke, D., and Kooman, J. (1994) Biochemie fu¨r Mediziner und Naturwissenschaftler, Georg Thieme Verlag, Stuttgart. 8. Crone, C., and Oleson, S. P. (1982) Brain Res. 241, 49–55. 9. Conyers, G., Milks, L., Conklyn, M., Showell, H., and Cramer, E. (1990) Am. J. Physiol. 259, C577–C585. 10. Marmorstein, A. D., Mortell, K. H., Ratcliffe, D. R., and Cramer, E. B. (1992) Am. J. Physiol. 262, 1403–1410. 11. Chang, Ch., Wang, X., and Caldwell, R. B. (1997) J. Neurochem. 69, 859–867. 12. Zettl, K. S., Sjaastad, M. D., Riskin, P. M., Parry, G., Machen, T. E., and Firestone, G. L. (1992) Proc. Natl. Acad. Sci. USA 89, 9069–9073.
315
AID
BBRC 8653
/
6955$$$$84
06-03-98 14:51:53
bbrcg
AP: BBRC