Cellular and Molecular Neurobiology, Vol. 20, No. 6, 2000

Rapid Communication

Involvement of Glial Cells in Cyclosporine-IncreasedPermeability of Brain Endothelial Cells

Shinya Dohgu,1 Yasufumi Kataoka,1,4 Hiroaki Ikesue,1 Mikihiko Naito,2

Takashi Tsuruo,2 Ryozo Oishi,1 and Yasufumi Sawada3

Received May 4, 2000; accepted May 20, 2000

SUMMARY

1. To test whether astrocytes participate in cyclosporine-induced dysfunction of theblood-brain barrier, we examined the effects of cyclosporine on the permeability of themouse brain endothelial (MBEC4) cells cocultured with C6 glioma cells, each cell layerplaced on the top and bottom of the insert membrane, respectively.

2. The presence of C6 cells remarkably aggravated cyclosporine-increased permeabil-ity of MBEC4 cells to sodium fluorescein.

3. In light of these findings, the possibility that astroglial cells could contribute tothe occurrence of cyclosporine-induced dysfunction of the blood-brain barrier triggeringneurotoxicity should be considered.

KEY WORDS: cyclosporine; neurotoxicity; glia; blood–brain barrier.

INTRODUCTION

Cyclosporine, a cyclic 11-amino acid peptide, is widely used in organ transplantationto prevent graft rejection by blocking calcineurin-mediated T-cell activation. Theimmunosuppressant action of cyclosporine is very effective, but this drug inducesadverse neurologic effects including tremor, seizure, and encephalopathy (Gijten-beek et al., 1999). Cyclosporine appears to produce convulsions by inhibiting �-aminobutyric acid (GABA)-ergic neural activity and binding properties of GABAA

receptor (Shuto et al., 1999). The inhibition of GABAergic neurotransmission bycyclosporine may lead to an activation of serotonergic neural activity and conse-quently produce tremors (Shuto et al., 1998). The mechanism of immunosuppres-sant-induced encephalopathy is obscure. We reported that cyclosporine may pro-

1 Department of Hospital Pharmacy, Faculty of Medicine, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan.

2 Institute of Molecular and Cellular Biosciences, University of Tokyo, Bunkyo-ku, Tokyo, 113-0032,Japan.

3 Department of Clinical Pharmacy, Graduate School of Pharmaceutical Sciences, Kyushu University,3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan.

4 To whom correspondence should be addressed. e-mail: ykataoka@st.hosp.kyushu-u.ac.jp

781

0272-4340/00/1200-0781$18.00/0 2000 Plenum Publishing Corporation

782 Dohgu, Kataoka, Ikesue, Naito, Tsuruo, Oishi, and Sawada

duce encephalopathy by injuring the brain capillary endothelial cells and inhibitingthe function and expression of P-glycoprotein, a multi-drug efflux pump (Kochi etal., 1999). Brain capillary endothelial cells closely communicate with astroglial cellsto form and maintain the blood–brain barrier (Tao-Chang et al., 1987). Cyclosporineenhances �1 receptor-mediated nitric oxide (NO) production in rat C6 glioma cellsand NO is involved in the dysfunction of blood–brain barrier (Hurst and Fritz,1996; Kataoka et al., 1999). To clarify whether astrocytes participate in cyclosporine-induced dysfunction of the blood–brain barrier, we examined the effects ofcyclosporine on the permeability of brain capillary endothelial cells, using an invitro model of the blood–brain barrier. This model was composed of mouse brainendothelial (MBEC4) cells (Tatsuta et al., 1992) on the inside and C6 glioma cellson the outside of a membrane of the insert placed in the well.

MATERIALS AND METHODS

MBEC4 cells isolated from BALB/c mice brain cortices and immortalized bySV40 transformation (Tatsuta et al., 1992) were cultured in Dulbecco’s modifiedEagle’s medium (DMEM) (GIBCO BRL, Life Technologies, Grand Island, NY)supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 �g/mlstreptomycin. C6 glioma cells (JCRB9096, Health Science Research ResourcesBank, Osaka, Japan) were cultured in DMEM supplemented with 10% fetal calfserum, and 50 �g/ml gentamycin. They were grown in a humidified atmosphere of5% CO2/95% air at 37�C. To make an in vitro model of the blood–brain barrier,C6 cells (40,000 cells/cm2) were first cultured on the outside of the collagen-coatedpolycarbonate membrane (3.0 �m pore size) of the TranswellTM insert (12-well type,Costar, MA) directed upside-down in the well. C6 cells were grown for 2 days andthen MBEC4 cells (42,000 cells/cm2) were seeded on the inside of the insert placedin the well of the 12-well culture plate (Costar, MA), where the inside and outsideof the membrane were directed top and bottom, respectively, in the conventionalway. The monolayer system was also made with MBEC4 cells alone (MBEC4monolayer). After MBEC4 cells were cultured for 3 days, these inserts were washedtwice with serum-free medium. Cells were exposed to 5 �M cyclosporine (RBI,MA) injected into the inside of the insert (luminal side) for 12 hr. In parallel, cellswere treated with serum-free medium containing the corresponding amount ofethanol as the vehicle. Cyclosporine was dissolved in ethanol and diluted withserum-free culture medium (0.1% as the final ethanol concentration).

To initiate the transport experiments, the medium was removed and cells werewashed twice with Krebs–Ringer buffer (118 mM NaCl, 4.7 mM KCl, 2.5 mMCaCl2 , 1.0 mM NaH2PO4 , 11 mM D-glucose, pH 7.4). Krebs–Ringer buffer (1.5 ml)was put on the outside of the insert in the well (abluminal side). Krebs–Ringerbuffer (0.5 ml) containing 100 �g/ml sodium fluorescein (Na-F) (MW 376, Sigma,St. Louis, MO) was loaded on the luminal side of the insert. Samples (0.5 ml) wereremoved from the abluminal chamber at 5, 10, 15, 20, 30, and 60 min and immediatelyreplaced with fresh Krebs–Ringer buffer. Aliquots (50 �l) from the abluminalchamber samples were mixed with 3 ml of Krebs–Ringer buffer and then the

Glial Cells and Cyclosporine-Induced BBB Dysfunction 783

concentration of Na-F was determined using a spectrofluorometer, Ex(�) � 483nm, Em(�) � 517 nm (RF520, Shimazu, Kyoto, Japan). Permeability coefficientand clearance were calculated according to the method described by Dehouck etal. (1992). Clearance was expressed as �l of tracer diffusing from the luminal tothe abluminal chambers and was calculated from the initial concentration of tracerin the luminal and the final concentration of tracer in the abluminal chamber:Clearance (�l) � [C]A � VA/[C]L , where [C]L is the initial luminal tracer concentra-tion, [C]A is the abluminal tracer concentration, and VA is the volume of the ablumi-nal chamber. During the 20-min period of the experiment, the clearance volumeincreased linearly with time. The average volume cleared was plotted versus time,and the slope was estimated by linear regression analysis. The slope of clearancecurves for the MBEC4 monolayer or coculture was denoted PSapp , where PS isthe permeability � surface area product (in �l per min). The slope of theclearance curve with the control membrane was denoted PSmembrane . In thecoculture system the control membrane is the C6 cell-layered membrane. Thereal PS value for the MBEC4 monolayer and coculture PStrans was calculatedfrom 1/PSapp � 1/PSmembrane � 1/PStrans . The PStrans values were divided by thesurface area of the Transwell inserts to generate the permeability coefficientPtrans in cm per min.

Effect of cyclosporine on the cell viability was assessed using a WST-1 assay(Cell Counting Kit, DOJINDO, Kumamoto, Japan). The highly water-solubleformazan dye (WST-1), reduced by mitochondrial dehydrogenase, was measuredby determining the absorbance of each sample with a 450-nm test wavelength anda 700-nm reference wavelength.

The values are expressed as mean � SE. Statistical analysis was performedusing Student’s unpaired t test. Single-factor and two-factor analyses of variance(ANOVA) followed by Tukey–Kramer tests were applied to multiple comparisons.The differences between means were considered to be significant when p valueswere less than 0.05.

RESULTS AND DISCUSSION

The permeability coefficient of Na-F for the MBEC4 monolayer and MBEC4cells cocultured with C6 cells (MBEC4 coculture) obtained during a 20-min periodafter adding Na-F were 1.8 � 0.1 � 10�4 and 3.4 � 0.2 � 10�4 cm/min, respectively(n � 9–12 inserts) (the inset of Fig. 1). The permeability of the MBEC4 cocultureto Na-F was significantly increased at each period when compared with the MBEC4monolayer. These findings suggest that the MBEC4 coculture is more leaky thanthe monolayer. When rat brain endothelial cells and astrocytes were cultured onthe inside and outside of the filter, respectively, transendothelial cell resistanceincreased compared with the endothelial cell monolayer (Kondo et al., 1996). How-ever, rat C6 glioma cells had no effect on the sucrose permeability of bovine brainmicrovessel endothelial cells (Abbruscato and Davis, 1999). Brain tumors, suchas astroglioma, have been known to contribute to a relatively hyperpermeableblood–brain barrier (Greig, 1989). The permeability of MBEC4 cells increased by

784 Dohgu, Kataoka, Ikesue, Naito, Tsuruo, Oishi, and Sawada

Fig. 1. Changes in the permeability coefficient of Na-F through theMBEC4 cell monolayer (monolayer) and MBEC4 cells cocultured withC6 glioma cells (coculture) after exposure to 5 �M cyclosporine for 12hr. MBEC4 and C6 cells were cultured on the top and bottom of theinsert membrane, respectively. Values are expressed as percentage ofvehicle corresponding to each system (monolayer and coculture). Theinset represents clearance of Na-F plotted versus time through the MBEC4monolayer and coculture after treatment with vehicle for 12 hr. Valuesare shown as mean � SE of 9–12 inserts. Error bars in the inset are notshown when they are smaller than the symbols. *p � 0.05 and **p �0.01, significant difference from vehicle corresponding to each system;#p � 0.05, significant difference between cyclosporine treatment in MBEC4monolayer and that in coculture.

coculturing with C6 glioma cells may be due to neoplastic changes in the glialcharacteristics and/or species difference.

Figure 1 shows changes in the permeability coefficient of Na-F through theMBEC4 monolayer and coculture after exposure of 5 �M cyclosporine for 12 hr.We previously reported that cyclosporine (0.5–5 �M) dose-dependently decreasedmitochondrial dehydrogenase activities by 20–50% of controls in MBEC4 cellstreated with cyclosporine for 24 hr (Kochi et al., 1999). Therefore, the exposuretime was changed to 12 hr and the maximum concentration, without cytotoxicaction, was selected. Cyclosporine, 5 �M, had no effect on the cell viability assessedby the WST-1 assay in the MBEC4 monolayer and coculture (101.6 � 7.6 and115.6 � 7.9% of controls, respectively, n � 4–7 inserts). The permeability of MBEC4cells to Na-F was significantly increased after a 12 hr exposure of 5 �M cyclosporineto the MBEC4 monolayer (p � 0.05) or coculture (p � 0.01). The presence of C6cells markedly aggravated cyclosporine-increased MBEC4 cell permeability toNa-F (p � 0.05). A 12-hr exposure of 5 �M cyclosporine in the present studyshowed no effect on the cell viability assessed by the WST-1 assay. Therefore,this increased permeability was not due to the direct cytotoxicity of cyclosporine.Previous findings demonstrated that the treatment of MBEC4 cells with cyclosporine(1–10 �M) for 24 hr inhibited the function of P-glycoprotein limiting the distribution

Glial Cells and Cyclosporine-Induced BBB Dysfunction 785

of cyclosporine into the brain (Kochi et al., 1999). Taking these and the presentfindings together, we find it highly likely that cyclosporine passes through the barrierof MBEC4 cells and acts on C6 cells to accelerate an increase in the MBEC4 cellpermeability. Glia-derived NO has been strongly implicated in the pathogenesis ofneurodegenerative disorders (Vincent et al., 1998). Cyclosporine induces potentia-tion of glial evoked NO production (Kataoka et al., 1999). NO reversibly decreasesthe integrity of the endothelial cell barrier in the coculture system (Hurst and Fritz,1996). Therefore, an interaction between endothelial and astroglial cells through thegap junction and/or biological substrates such as NO may accelerate cyclosporine-increased endothelial cell permeability. This event may be interpreted as triggeringthe occurrence of cyclosporine neurotoxicity. In the present study we demonstratedthat cyclosporine-increased MBEC4 cell permeability was aggravated by C6 gliomacells when cocultured with MBEC4 cells. The present findings suggest that astrocytescontribute to the occurrence of cyclosporine-induced dysfunction of the blood–brainbarrier triggering neurotoxicity, including encephalopathy.

ACKNOWLEDGMENTS

This work was supported in part by Grants-in-Aid for Scientific Research((B)(2)11470513, (C)(2)12672218) from the Ministry of Education, Science, Sportsand Culture, Japan.

REFERENCES

Abbruscato, T. J., and Davis, T. P. (1999). Combination of hypoxia/aglycemia compromises in vitroblood–brain barrier integrity. J. Pharmacol. Exp. Ther. 289:668–675.

Dehouck, M.-P., Jolliet-Riant, P., Bree, F., Fruchart, J.-C., Cecchelli, R., and Tillement, J.-P. (1992).Drug transfer across the blood–brain Barrier: Correlation between in vitro and in vivo models. J.Neurochem. 58:1790–1797.

Gijtenbeek, J. M. M., van den Bent, M. J., and Vecht, Ch. J. (1999). Cyclosporine neurotoxicity: Areview. J. Neurol. 246:339–346.

Greig, N. (1989). Brain tumors and blood–brain barrier. In Neuwelt, E. A. (ed.), Implications of theBlood–Brain Barrier and Its Manipulation; Vol. 2. Clinical Aspects, Plenum Press, New York,pp. 77–106.

Hurst, R. D., and Fritz, I. B. (1996). Nitric oxide-induced perturbations in a cell culture model of theblood–brain barrier. J. Cell. Physiol. 167:89–94.

Kataoka, Y., Ikesue, H., Shuto, H., Kawachi, R., Sueyasu, M., and Oishi, R. (1999). Imuunosuppressantneurotoxicity: Cyclosporine and tacrolimus enhance stimulation-evoked NO production in C6 gliomacells. Jpn. J. Pharmacol. 79:(Suppl. I): P–419.

Kochi, S., Takanaga, H., Matsuo, H., Naito, M., Tsuruo, T., and Sawada, Y. (1999). Effect of cyclosporinA or tacrolimus on the function of blood–brain barrier cells. Eur. J. Pharmacol. 372:287–295.

Kondo, T., Kinouchi, H., Kawase, M., and Yoshimoto, T. (1996). Astroglial cells inhibit the increasingpermeability of brain endothelial cell monolayer following hypoxia/reoxygenation. Neurosci.Lett. 208:101–104.

Shuto, H., Kataoka, Y., Fujisaki, K., Nakao, T., Sueyasu, M., Miura, I., Watanabe, Y., Fuiwara, M., andOishi, R. (1999). Inhibition of GABA system involved in cyclosporine-induced convulsions. LifeSci. 65:879–887.

Shuto, H., Kataoka, Y., Kanaya, A., Matsunaga, K., Sueyasu, M., and Oishi, R. (1998). Enhancementof serotonergic neural activity contributes to cyclosporine-induced tremors in mice. Eur. J. Pharma-col. 341:33–37.

786 Dohgu, Kataoka, Ikesue, Naito, Tsuruo, Oishi, and Sawada

Tao-Cheng, J. H., Nagy, Z., and Brightman, M. W. (1987). Tight junctions of brain endothelium in vitroare enhanced by astroglia. J. Neurosci. 7:3293–3299.

Tatsuta, T., Naito, M., Oh-hara, T., Sugawara, I., and Tsuruo, T. (1992). Functional involvement of P-glycoprotein in blood–brain barrier. J. Biol. Chem. 267:20383–20391.

Vincent, V. A. M., Tilders, F. J. H., and Van Dem, A.-M. (1998). Production, regulation and role ofnitric oxide in glial cells. Mediators Inflamm. 7:239–255.