Changes in cytoskeletal and tight junctional proteins correlate with decreased permeability induced by dexamethasone in cultured rat brain endothelial cells

Neuroscience Letters 344 (2003) 112–116 www.elsevier.com/locate/neulet

Changes in cytoskeletal and tight junctional proteins correlate with decreased permeability induced by dexamethasone in cultured rat brain endothelial cells Ignacio A. Romeroa,b,*, Katrina Radewicza, Emmanuelle Jubinb, C. Charles Michelc, John Greenwooda, Pierre-Olivier Couraudb, Peter Adamsona a

Division of Cell Biology, Institute of Ophthalmology, University College London, Bath Street, London EC1V 9EL, UK b Institut Cochin de Ge´ne´tique Mole´culaire, 22 rue Me´chain, 75014 Paris, France c Division of Biomedical Sciences, Faculty of Medicine, Imperial College, London SW7 2AZ, UK Received 27 November 2002; received in revised form 12 March 2003; accepted 12 March 2003

Abstract The blood –brain barrier (BBB) plays an important role in controlling the passage of molecules from the blood to the extracellular fluid environment of the brain. An immortalised rat brain endothelial cell line (GPNT) was used to investigate the mechanisms underlying dexamethasone-induced decrease in paracellular permeability. Following treatment with 1 mM dexamethasone there was a decrease in transmonolayer paracellular permeability mainly to sucrose, fluorescein and dextrans of up to 20 KDa. According to pore theory, these differences in permeability were consistent with a decrease in the number of pores between brain endothelial cells. This effect was accompanied by a concentration of filamentous actin and cortactin to the cell periphery. Concomitantly, the continuity of the tight junctional protein ZO-1 at the cell borders was improved and was associated with an increase in both ZO-1 and occludin expression. By contrast, the expression and distribution of adherens junctional proteins such as b-catenin and p100/p120 remained unchanged. These observations suggest that glucocorticoids induce a more differentiated BBB phenotype in cultured brain endothelial cells through modification of tight junction structure. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Blood–brain barrier; Brain endothelium; Permeability; Dexamethasone; Glucocorticoids; Cell junctions

Endothelial cells (EC) that line the microvasculature of the central nervous system (CNS) differ fundamentally from other vascular endothelia in their capacity to regulate the passage of molecules and cells to and from the neural parenchyma. This selectivity is a consequence of various specialised morphological features unique to CNS ECs, including the expression of tight intercellular junctions, minimal transcytosis and lack of transcellular pores. Understanding the cellular factors that regulate permeability in brain endothelium should lead to the development of successful therapeutic strategies for overcoming restricted drug delivery into the brain. Although much has been learnt * Corresponding author. Department of Biological Sciences, The Open University, Walton Hall, Milton Keynes MK7 6AA, UK. Tel.: þ 44-1908659-467; fax: þ 44-1908-654-167. E-mail address: [email protected] (I.A. Romero).

from in vivo studies, the limitations imposed by whole animal experiments has led to the development of techniques to isolate and culture CNS-derived ECs (for review see ref. [5]). These techniques, however, are difficult and problematic as CNS-derived endothelia are phenotypically unstable in long-term culture, often remain impure and have low plating efficiency, restricting most investigations on primary cultures. In order to overcome this problem, immortalised rat brain EC lines have proved to be a valuable resource for studying biochemical and functional properties of cerebral endothelia [5]. Although primary cultured rat brain ECs represent a well differentiated phenotype, they appear to be unable to maintain functional tight junctions as displayed in vivo. Immortalisation of brain ECs usually results in a more de-differentiated phenotype, and like primary cultures, the extremely tight permeability, characteristic of brain endothelium in vivo (, 2000– 5000 V cm2),

0304-3940/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3940(03)00348-3

I.A. Romero et al. / Neuroscience Letters 344 (2003) 112–116

is also not usually preserved in these cell lines (, 50– 100 V cm2) [5]. It has recently been reported that corticosteroid treatment modulates tight junction permeability of 31EG4 epithelial cells and brain ECs, determined by the electrical resistance of the cell monolayer (from 150 up to 1000 –3000 V cm2) [8,20]. In this study, we have investigated the effect of a synthetic corticosteroid, dexamethasone, on the paracellular permeability and expression of adherens and tight junctional proteins within an immortalised rat brain EC line, GPNT. This previously described cell line retains a variety of morphological and physiological characteristics of primary brain ECs such as expression of specific brain endothelial markers and cell surface adhesion molecules [10,12]. GPNT cells were grown to confluence in the presence of 1 mM dexamethasone from seeding, on Transwell-Cleare inserts in DMEM:Hams F12 (1:1) medium containing 10% foetal calf serum (FCS), 10 mM HEPES, 100 U/ml penicillin and 100 mg/ml streptomycin. Electrical resistance of GPNT monolayers on Transwell inserts was then measured using an Endohm 12 chamber and an Endohmeter (WPI Inc., Sarasota, FL, USA). Electrical resistance of confluent GPNT cells on Transwell-Clear (polyester, 12 mm diameter, pore size 0.4 mm, Costar, France) inserts was 66 ^ 3 V cm2 (n ¼ 4), similar to that obtained by others with primary cultures of bovine and human brain ECs in astrocyte conditioned media [13]. However, no change in the transendothelial resistance was observed following dexamethasone treatment (not shown). Paracellular permeability of GPNT cell monolayers to different molecular weight tracer markers (fluorescein isothiocyanate (FITC)conjugated dextrans, fluorescein and [14C] sucrose) was determined using a previously described method [12]. The permeability coefficients of the endothelial monolayers (Pe) were then calculated using the slopes of the curves representing cleared volume versus time [12]. Treatment of cells with 1 mM dexamethasone from seeding resulted in a general decrease in the permeability of GPNT monolayers (Table 1), although this was most pronounced for sucrose (by 39 ^ 7%) and fluorescein (by 47 ^ 7%) (mean ^

113

SEM, n ¼ 12 – 14 filters from five independent assays) (Table 1). From the pore model [11], we can relate the molecular radius (R) of the solutes under investigation with their free diffusion coefficients (D) and Pe (Table 1). Our data is consistent with a heterogeneous population of cylindrical pores between GPNT cells with a large pore of ˚ and a small pore of , 14.5 A ˚ . The dexamethasone , 185 A data was consistent with a decrease in the number of large pores between cells rather than with a decrease in the size of these pores. However, dexamethasone treatment appeared ˚. to reduce the size of the small pore to a radius of , 8.3 A The absolute degree to which these small pores are reduced in size and whether their number is also decreased will require a more extensive analysis. Potential targets for the dexamethasone effects on permeability may include proteins that interact with both tight and adherens junctions as well as cytoskeletal proteins [14]. The cadherin-associated members of the catenin family b-catenin, and p100/p120, both components of the adherens junctional complex, were seen to localise to cellcell junctions in untreated cells, and this pattern of distribution was unchanged following stimulation with dexamethasone (Figs. 1A – D). In addition, no changes in the expression of a-, b-, or g-catenin or p100/p120 were observed as determined by immunoblotting (Fig. 2). The submembranous tight junction-associated protein ZO-1 and the integral membrane protein occludin are components of the BBB endothelium in vivo and in vitro and both are found as a continuous line at interendothelial cell-cell contacts in tight brain endothelial monolayers [7,13]. Immunoblotting of GPNT cell lysates with anti-occludin and anti-ZO-1 antibodies revealed bands at 65 and 225 KDa, respectively, which were stronger in dexamethasone-treated GPNT cells (Fig. 2). Immunofluorescent detection of ZO-1 revealed a discrete localisation at cell-cell contacts (Fig. 1E) whilst the expression of occludin was low (not shown). A more regular and continuous cortical ZO-1 distribution was observed following dexamethasone treatment (Fig. 1F), while no effect on the intracellular distribution of occludin was observed (not shown). A continuous junctional localisation

Table 1 Permeability coefficients (Pe) for [14C] sucrose, fluorescein and FITC-dextrans (4, 10, 20, 40, 70 and 150 kDa) of control and dexamethasone-treated GPNT cells Solute

MW

Pe (1023 cm/min) control

Pe (1023 cm/min) þ dexamethasone

˚) Radius (A

D (1025 cm2/min)

Sucrose Fluorescein FD4b FD10b FD20b FD40b FD70b FD150b

342 389 4000 10,000 20,000 40,000 70,000 150,000

7.373 ^ 0.617 5.210 ^ 0.472 0.975 ^ 0.076 0.408 ^ 0.016 0.333 ^ 0.031 0.114 ^ 0.029 0.053 ^ 0.007 0.018 ^ 0.007

4.496 ^ 0.530a 2.781 ^ 0.395a 0.643 ^ 0.061a 0.360 ^ 0.041 0.156 ^ 0.013a 0.077 ^ 0.019 0.038 ^ 0.013 0.014 ^ 0.004

4.7 5.5 14 22 30 45 60 80

41.5 35.8 14 8.9 6.6 4.3 3.3 2.5

a

P , 0.05. For the calculation of FITC-dextrans Pe, control and test clearance values were corrected for the presence of unconjugated fluorescein as described by van Bree et al. [17]. b

114

I.A. Romero et al. / Neuroscience Letters 344 (2003) 112–116

of ZO-1 has been reported to correlate with low permeability of intercellular junctions in primary brain ECs [13,16]. In addition occludin, an integral membrane component of tight junctions, was only detected at low levels at cell-cell contacts by immunofluorescence following dexamethasone treatment. Indeed, high expression of

cortical occludin has been reported to correlate with low paracellular permeability in ECs [7]. However, this protein may only be a minor component of the tight junction and does not appear to be essential for tight junction formation [15], being complemented by other integral membrane proteins such as the claudin molecules [3]. In parallel with these observations, staining with TRITCphalloidin revealed a diffuse pattern of F-actin with stress fibres distributed throughout the cytoplasm and little F-actin distributing around the cell periphery in untreated cells (Fig. 1G). A similar distribution of cytoskeleton-associated cortactin was also observed in untreated cells (Fig. 1I). Treatment with 1 mM dexamethasone from cell seeding resulted in filamentous actin and the cytoskeleton associated protein cortactin being highly concentrated in the regions of cell-cell contact with few F-actin stress fibres visible within the cytoplasm (Figs. 1H&J), an observation consistent with a more differentiated barrier phenotype induced by dexamethasone. However, this effect was not associated with an increased expression of the actin-binding proteins, cortactin and ezrin, or the tyrosine kinase c-src (Fig. 2). As the tight junctional protein ZO-1 is an important site of association between F-actin and the tight junctions, its increased expression may be indirectly related with a decrease in the number of intracytoplasmic stress fibres as actin microfilaments would be directed to the cell-cell contact points. Indeed, changes in the number of F-actin microfilaments is correlated with changes in vascular permeability in both brain and peripheral ECs [13,16]. Our results are in contrast with those obtained in mouse 31EG4 mammary epithelial cells where dexamethasone treatment did not modulate the production or location of filamentous actin or the tight junction protein ZO-1, although it increased transepithelial resistance [20]. However, a similar effect on occludin and ZO-1 expression was recently observed on retinal ECs following treatment with hydrocortisone [1]. The results from these studies emphasize the importance of the cell type in determining the biological effects induced by glucocorticoids.

Fig. 1. Immunocytochemical analysis of junctional and cytoskeletal components in control cells (A, C, E, G, I) and cells treated with 1 mM dexamethasone from seeding (B, D, F, H, J). GPNT cells were fixed in 3% paraformaldehyde, rinsed in phosphate buffered saline (PBS) and permeabilised using 0.25% Triton X-100. After further washing, the cells were blocked with PBS containing 10% FCS. This was followed by a 2 h incubation with antibodies against the following epitopes:, b-catenin (A, B), p100/p120 (C, D), ZO-1 (B, F) and cortactin (I, J). Some monolayers were labelled with TRITC-phalloidin for 1 h after permeabilisation to reveal filamentous actin (G, H). For all conditions, the cells were washed in PBS and blocked with PBS containing 10% goat serum and 10% ECS. Cells were subsequently incubated with FITC- or TRITC-conjugated secondary antibody in the same buffer for 1 h followed by extensive washing prior to mounting. After mounting, the cells were viewed on a confocal microscope (MRC 1000, Biorad, Hercules, CA, USA). Omission of the primary antibody and substitution with an isotype-matched mAb served as controls. Results are from one experiment representative of three.

I.A. Romero et al. / Neuroscience Letters 344 (2003) 112–116

115

corticoid induction of tight junction sealing in epithelial cells may require redistribution of cytoskeletal proteins, as it also induces a decrease in the actin-bundling protein, fascin [18], which is in agreement with the observed effects in brain ECs. Alternatively, glucocorticoid treatment may decrease the expression of other mediators responsible for elevated endothelial permeability in untreated GPNT monolayers, resulting in concomitant reduction in expression of ZO-1 and occludin, and immature cell-cell junctions. In support of this hypothesis, we have previously demonstrated that dexamethasone prevented cytokineinduced enhanced expression of MMP-9 and alterations in the expression of ZO-1 [6]. In addition, dexamethasone treatment may result in a decrease in vascular endothelial growth factor (VEGF) production by GPNT cells. Indeed, glucocorticoids have been shown to inhibit VEGF gene expression in a variety of cell types [4,9], and the effects of VEGF on endothelial permeability are well known [2]. In summary, we report here that the expression pattern of ZO-1 changed from partially discontinuous to continuous whereas F-actin was redistributed to the periphery following treatment of GPNT cells with the glucocorticoid dexamethasone. Under these conditions a decrease in paracellular permeability of GPNT monolayers was also observed suggesting that this treatment may induce further differentiation of GPNT cells towards an improved BBB phenotype by maintaining intercellular junctions.

Fig. 2. Western blot analysis of junctional and cytoskeletal components. GPNT cells were grown on collagen-coated six well plates until confluent. Whole cell lysates were prepared by rapidly replacing the culture medium with hot sodium dodecyl sulphate (SDS) sample buffer followed by heating at 958C for 5 min. The cell extracts were then resolved by SDSpolyacrylamide gel electrophoresis and the slab gels were transferred to nitrocellulose filters (Schleich and Schuell, Dassel, Germany). The filters were stained with Ponceau-S and blocked with 5% non-fat dried milk in PBS. After washing in PBS containing 0.05% Tween-20, the filters were probed with primary antibody and were washed and incubated with the appropriate horseradish peroxidase-conjugated secondary antibody. The filters were extensively washed and immunoreactive bands were detected by enhanced chemiluminescence.

Given the known transcriptional mechanism of the glucocorticoid receptor, it is likely that one or more glucocorticoid-regulated genes are responsible for the effect of dexamethasone treatment on endothelial barrier formation. However, whether the observed effects are a result of the direct increase in transcriptional activity of the ZO-1 and occludin genes by the glucocorticoid receptor protein-steroid hormone complex remains to be elucidated. In Con8 mammary epithelial tumor cells, dexamethasone treatment has been shown to strongly stimulate the level of the Id-1 protein, which is a serum-inducible helix-loop-helix transcriptional repressor involved in cell differentiation, and this effect was shown to be associated with reorganisation of ZO-1 to the cell periphery [19]. Furthermore, the gluco-

References [1] D.A. Antonetti, E.B. Wolpert, L. DeMaio, N.S. Harhaj, R.C. Scaduto Jr, Hydrocortisone decreases retinal endothelial cell water and solute flux coincident with increased content and decreased phosphorylation of occludin, J. Neurochem. 80 (2002) 667 –677. [2] D.O. Bates, N.J. Hillman, B. Williams, C.R. Neal, T.M. Pocock, Regulation of microvascular permeability by vascular endothelial growth factors, J. Anat. 200 (2002) 581– 597. [3] M. Furuse, K. Fujita, T. Hiiragi, K. Fujimoto, S. Tsukita, Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin, J. Cell Biol. 141 (1998) 1539– 1550. [4] J. Gille, K. Reisinger, B. Westphal-Varghese, R. Kaufmann, Decreased mRNA stability as a mechanism of glucocorticoidmediated inhibition of vascular endothelial growth factor gene expression by cultured keratinocytes, J. Invest. Dermatol. 117 (2001) 1581–1587. [5] M. Gumbleton, K.L. Audus, Progress and limitations in the use of in vitro cell cultures to serve as a permeability screen for the blood– brain barrier, J. Pharm. Sci. 90 (2001) 1681–1698. [6] K.A. Harkness, P. Adamson, J.D. Sussman, G.A. Davies-Jones, J. Greenwood, M.N. Woodroofe, Dexamethasone regulation of matrix metalloproteinase expression in CNS vascular endothelium, Brain 123 (Pt 4) (2000) 698–709. [7] T. Hirase, J.M. Staddon, M. Saitou, Y. Ando-Akatsuka, M. Itoh, M. Furuse, K. Fujimoto, S. Tsukita, L.L. Rubin, Occludin as a possible determinant of tight junction permeability in endothelial cells, J. Cell. Sci. 110 (Pt 14) (1997) 1603– 1613. [8] D. Hoheisel, T. Nitz, H. Franke, J. Wegener, A. Hakvoort, T. Tilling, H.J. Galla, Hydrocortisone reinforces the blood–brain properties in a

116

[9]

[10]

[11] [12]

[13]

[14] [15]

I.A. Romero et al. / Neuroscience Letters 344 (2003) 112–116 serum free cell culture system, Biochem. Biophys. Res. Commun. 247 (1998) 312–315. M. Nauck, G. Karakiulakis, A.P. Perruchoud, B. Papakonstantinou, M. Roth, Corticosteroids inhibit the expression of the vascular endothelial growth factor gene in human vascular smooth muscle cells, Eur. J. Pharmacol. 341 (1998) 309 –315. A. Regina, I.A. Romero, J. Greenwood, P. Adamson, J.M. Bourre, P.O. Couraud, E. Roux, Dexamethasone regulation of P-glycoprotein activity in an immortalized rat brain endothelial cell line, GPNT, J. Neurochem. 73 (1999) 1954–1963. E.M. Renkin, Capillary transport of macromolecules: pores and other endothelial pathways, J. Appl. Physiol. 58 (1985) 315–325. I.A. Romero, M.C. Prevost, E. Perret, P. Adamson, J. Greenwood, R.O. Couraud, S. Ozden, Interactions between brain endothelial cells and human T-cell leukemia virus type 1-infected lymphocytes: mechanisms of viral entry into the central nervous system, J. Virol. 74 (2000) 6021–6030. L.L. Rubin, D.E. Hall, S. Porter, K. Barbu, C. Cannon, H.C. Horner, M. Janatpour, C.W. Liaw, K. Manning, J. Morales, L.I. Tanner, K.J. Tomaselli, F. Bard, A cell culture model of the blood–brain barrier, J. Cell Biol. 115 (1991) 1725–1735. L.L. Rubin, J.M. Staddon, The cell biology of the blood–brain barrier, Annu. Rev. Neurosci. 22 (1999) 11–28. M. Saitou, K. Fujimoto, Y. Doi, M. Itoh, T. Fujimoto, M. Furuse, H.

[16]

[17]

[18]

[19]

[20]

Takano, T. Noda, S. Tsukita, Occludin-deficient embryonic stem cells can differentiate into polarized epithelial cells bearing tight junctions, J. Cell Biol. 141 (1998) 397–408. C. Schulze, C. Smales, L.L. Rubin, J.M. Staddon, Lysophosphatidic acid increases tight junction permeability in cultured brain endothelial cells, J. Neurochem. 68 (1997) 991 –1000. J.B. van Bree, A.G. de Boer, M. Danhof, L.A. Ginsel, D.D. Breimer, Characterization of an ‘in vitro’ blood–brain barrier: effects of molecular size and lipophilicity on cerebrovascular endothelial transport rates of drugs, J. Pharmacol. Exp. Ther. 247 (1988) 1233–1239. V. Wong, D. Ching, P.D. McCrea, G.L. Firestone, Glucocorticoid down-regulation of fascin protein expression is required for the steroid-induced formation of tight junctions and cell-cell interactions in rat mammary epithelial tumor cells, J. Biol. Chem. 274 (1999) 5443–5453. P.L. Woo, A. Cercek, P.Y. Desprez, G.L. Firestone, Involvement of the helix-loop-helix protein Id-1 in the glucocorticoid regulation of tight junctions in mammary epithelial cells, J. Biol. Chem. 275 (2000) 28649–28658. K.S. Zettl, M.D. Sjaastad, P.M. Riskin, G. Parry, T.E. Machen, G.L. Firestone, Glucocorticoid-induced formation of tight junctions in mouse mammary epithelial cells in vitro, Proc. Natl. Acad. Sci. USA 89 (1992) 9069–9073.