Serum albumin induces osmotic swelling of rat retinal glial cells

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –27 6

available at www.sciencedirect.com

www.elsevier.com/locate/brainres

Research Report

Serum albumin induces osmotic swelling of rat retinal glial cells Silvana Löffler a , Antje Wurm a , Franziska Kutzera b , Thomas Pannicke a , Katja Krügel a , Regina Linnertz a , Peter Wiedemann b , Andreas Reichenbach a , Andreas Bringmann b,⁎ a

Paul Flechsig Institute of Brain Research, University of Leipzig, Leipzig, Germany Department of Ophthalmology and Eye Hospital, University of Leipzig, Leipzig, Germany

b

A R T I C LE I N FO

AB S T R A C T

Article history:

Edema in the ischemic neural tissue develops by increased vascular permeability associated

Accepted 21 December 2009

with extravasation of albumin, and by glial swelling. Here, we show that bovine serum

Available online 4 January 2010

albumin acutely administered to slices of the rat retina causes swelling of glial somata under hypoosmotic conditions. The effect of albumin was dose-dependent, with half-

Keywords:

maximal and maximal effects at 10 nM and 1 μM, respectively, and was mediated by

Albumin

activation of transforming growth factor-β receptor type II, oxidative stress, and the

Cellular swelling

production of arachidonic acid and prostaglandins. Albumin-induced glial swelling was

Glutamate

prevented by glutamate and purinergic receptor agonists. The data suggest that serum

Purinergic receptor

albumin may induce glial swelling in the presence of osmotic gradients.

Glial cell

© 2010 Elsevier B.V. All rights reserved.

Retina

1.

Introduction

Ischemia–reperfusion of the neural tissue is commonly accompanied by the development of edema which is a major pathogenic factor contributing to neuronal degeneration and, via compression of blood vessels, tissue hypoxia. Generally, edema is caused by disruption of the blood–brain barrier and/or fluid accumulation within cells resulting in cellular swelling (cytotoxic edema) (Kimelberg, 2004). In the brain, swelling of astrocytes usually occurs concomitantly in

vasogenic edema, and represents a major mechanism of edema formation under ischemic and other conditions such as hyponatremia (Kimelberg, 2004). Cerebral ischemia–reperfusion is associated with a biphasic swelling of glial cells; early swelling occurs during the ischemic episode, and late swelling occurs concomitantly with vasogenic edema within hours or days after reperfusion (Rumpel et al., 1997). A similar time dependence of tissue swelling was described in the ischemic retina (Stefánsson et al., 1987). In addition to vasogenic edema, a swelling of glial cells is suggested to contribute to

⁎ Corresponding author. Department of Ophthalmology and Eye Hospital, University of Leipzig, Liebigstrasse 10-14, D-04103 Leipzig, Germany. Fax: +49 341 97 21 659. E-mail address: [email protected] (A. Bringmann). Abbreviations: AOPCP, adenosine-5′-O-(α,β-methylene)-diphosphate; ARL-67156, 6-N,N-diethyl-d-β,γ-dibromomethylene ATP; BAPTAAM, bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid acetoxymethyl ester; BSA, bovine serum albumin; CSC, 8-(3-chlorostyryl) caffeine; DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; H-89, N-[2-((p-bromocinnamyl)amino)ethyl]-5-isoquinolinesulfonamide; LY341495, (2S)-2-Amino-2-[(1S,2S)-2-carboxycycloprop-1-yl]-3-(xanth-9-yl) propanoic acid; MRS2179, N6-methyl-2′-deoxyadenosine-3′,5′-bisphosphate; NBTI, N-nitrobenzylthioinosine; NPPB, 5-nitro-2-(3-phenylpropylamino)benzoic acid; SB431542, 4-[4-(1,3-benzodioxol-5-yl)-5-(2pyridinyl)-1H-imidazol-2-yl]-benzamide; TGF, transforming growth factor 0006-8993/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2009.12.067

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –2 76

edema formation in the ischemic retina (Bringmann et al., 2004). There is little knowledge regarding the relationship between vasogenic and cytotoxic edema in the neural tissue. An increase in vessel permeability is associated with extravasation of serum proteins such as albumin. It has been shown that local inflammation in the retina causes an increased vesicular transport of serum proteins through vascular endothelial cells, and that these proteins are accumulated in pericytes, perivascular microglia, and retinal glial (Müller) cells, suggesting that retinal glial cells act as a secondary barrier to extravasated serum proteins (Claudio et al., 1994). Serum albumin is a multifunctional protein with neurotrophic and neuroprotective properties (Emerson, 1989; Zoellner et al., 1996; Belayev et al., 2001). However, serum albumin may have also detrimental effects in the neural tissue such as induction of hyperexcitability and epileptiform activity (Ivens et al., 2007). It is not known whether albumin may influence glial swelling, a major component of edema in the neural tissue. Extracellular fluid accumulation and cellular swelling indicate that cellular mechanisms of fluid transport and removal are disturbed. Normally, edema is resolved from the neural tissue by the dehydrating action of glial cells. Retinal glial cells maintain the water homeostasis of the neural retina (Bringmann et al., 2004). The transglial water transport is coupled to a transport of ions, in particular of potassium, and is facilitated by the co-localization of inwardly rectifying potassium (Kir) channels and water channels in perivascular membranes (Nagelhus et al., 1999). It has been shown that retinal glial cells downregulate perivascular Kir channels after transient ischemia of the retina, during ocular inflammation, and in diabetes (Pannicke et al., 2004, 2005, 2006); this downregulation should disrupt both the glial cell-mediated potassium clearance of the retina and the water transport through the cells (Bringmann et al., 2004). The downregulation of Kir channels in retinal glial cells is associated with an induction of cellular swelling under anisoosmotic conditions (Pannicke et al., 2004, 2005, 2006). Because extravasation of albumin is a characteristic of edema, we investigated whether serum albumin alters the osmotic swelling characteristics of retinal glial cells. We found that albumin evokes acute glial swelling under hypotonic conditions (a situation that resembles hypoxia-induced cytotoxic edema in the brain), and that the swelling was prevented by a glutamatergic–purinergic receptor signaling that results in opening of ion channels in glial membranes.

2.

Results

2.1.

Albumin induces osmotic glial cell swelling

Acute swelling of the somata of retinal glial (Müller) cells was investigated by superfusion of freshly isolated retinal slices with a hypoosmolar solution (containing 60% of control osmolarity). As shown in Fig. 1A, superfusion of the slices with isoosmolar solution, or with hypoosmolar solution, for 4 min did not evoke a swelling of glial somata. Thus, the degree of swelling after 4 min was taken for comparison with other conditions. However, hypotonic challenge in the pres-

269

Fig. 1 – Bovine serum albumin (BSA) evokes osmotic swelling of glial somata in slices of the rat retina. (A) Superfusion of a retinal slice with a hypotonic solution (60% of control osmolarity) had no effect on the glial soma size under control conditions. However, the cells displayed time-dependent swelling of their somata when BSA (5 μM) was administered simultaneously with the hypotonic solution. The images show a dye-filled glial soma recorded before (left) and during (right) hypotonic exposure in the presence of BSA. The data were obtained in 9 and 10 cells, respectively. (B) BSA administered simultaneously with the hypotonic solution evoked a dose-dependent swelling of glial somata, whereas BSA application in isotonic solution had no effect. The concentration of BSA (in μM) is given in the bars. (C) Glial soma area which was recorded in the absence (control) and presence of the following agents: barium chloride (1 mM), normal BSA (5 μM), and fatty acid-free BSA (5 μM). Bar diagrams show the mean (±SEM) glial soma area (n = 6–20) measured after a 4-minute superfusion of retinal slices with the hypotonic solution, and expressed in percent of the soma area before osmotic challenge (100%). Significant difference vs. control: **P < 0.01; ***P < 0.001. Bar, 5 μm.

ence of bovine serum albumin (BSA) induced a rapid swelling of glial cell bodies (Fig. 1A). Albumin did not evoke glial swelling under isotonic conditions (Fig. 1B). The swellinginducing effect of albumin under hypotonic conditions was dose-dependent (Fig. 1B). The half-maximal and maximal effects were at 10 nM and 1 μM, respectively. It has been shown that inflammatory lipids such as arachidonic acid induce a swelling of retinal glial cells under hypotonic conditions

270

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –27 6

(Uckermann et al., 2005). To determine whether the swellinginducing effect of albumin is mediated by lipids released from fatty acid–albumin complexes, we tested fatty acid-free albumin. As shown in Fig. 1C, lipid-free albumin caused a similar swelling under hypotonic conditions as normal albumin, which makes it unlikely that fatty acids bound to albumin caused the swelling. As previously shown (Pannicke et al., 2004), superfusion of retinal slices with a hypoosmolar solution containing barium ions (which block Kir channels in retinal glial cells) evoked glial swelling (Fig. 1C). The mean amplitude of the barium- and albumin-evoked swelling was not different (Fig. 1C). Though these data may suggest a swelling-inducing effect of albumin via inhibition of Kir channels (Pannicke et al., 2004), acute administration of albumin (5 μM) failed to alter the whole-cell potassium conductance of freshly isolated retinal glial cells in conventional patch-clamp recordings (data not shown). This may suggest that the effect of albumin was not mediated by inhibition of Kir channels.

2.2. stress

Involvement of inflammatory lipids and oxidative

It has been shown that inflammatory mediators such as arachidonic acid and prostaglandins, as well as oxidative stress, are causative factors of osmotic swelling of retinal glial cells (Uckermann et al., 2005). Acute administration of arachidonic acid, prostaglandin E2, or H2O2 to retinal slices evoked glial swelling under hypotonic conditions with a mean amplitude similar to albumin-evoked swelling (Fig. 2A). To determine whether albumin induces glial swelling by stimulation of the formation of arachidonic acid, the selective phospholipase A2 inhibitor, 4-bromophenacyl bromide, was tested. As shown in Fig. 2B, inhibition of phospholipase A2 prevented the albumin-induced osmotic swelling. Similarly, inhibition of the cyclooxygenase by indomethacin blocked the swelling of glial cells (Fig. 2B). The anti-inflammatory glucocorticoid, triamcinolone acetonide (9α-fluoro-16α-hydroxyprednisolone), is clinically used for the rapid resolution of retinal edema (Fraser-Bell et al., 2008). Triamcinolone inhibited the albumin-evoked osmotic swelling of retinal glial cells (Fig. 2B). To reveal whether acute oxidative stress plays a role in albumin-evoked glial swelling, we tested a cell-permeable reducing agent, dithiothreitol, and found a partial prevention of the swelling in the presence of the reducing agent (Fig. 2C). The inhibitory effects of the phospholipase A2 and cyclooxygenase inhibitors suggest that the effect of albumin is caused by de novo formation of arachidonic acid and prostaglandins. These lipids are potent inhibitors of the sodium pump; inhibition of the pump was shown to result in intracellular sodium overload and swelling of cultured cells (Staub et al., 1994). The sodium ionophore monensin evoked glial swelling under hypotonic conditions; co-administration of albumin and monensin did not further increase the swelling amplitude (Fig. 2D). The albumin-induced glial swelling was abrogated in the presence of a sodium-free extracellular solution (Fig. 2E). The data suggest that the albumin-induced swelling is caused by intracellular sodium overload, likely due to inhibition of the sodium pump by inflammatory lipids.

Fig. 2 – Albumin-induced osmotic swelling of retinal glial cells is caused by inflammatory mediators, oxidative stress, and sodium influx into the cells. Bovine serum albumin (BSA; 5 μM) was administered simultaneously with the hypotonic solution. (A) Acute administration of arachidonic acid (10 μM), prostaglandin E2 (PGE2; 30 nM), and H2O2 (200 μM), respectively, to retinal slices evoked swelling of glial somata under hypotonic conditions. (B) The BSA-evoked glial swelling was prevented in the presence of the following agents: the inhibitor of phospholipase A2, 4-bromophenacyl bromide (Bromo; 500 μM), the cyclooxygenase inhibitor, indomethacin (Indo; 10 μM), and triamcinolone acetonide (Triam; 100 μM). (C) The reducing agent, dithiothreitol (DTT; 3 mM), decreased the BSA-evoked swelling of glial somata. (D) Monensin (30 μM), a sodium ionophore, evoked a similar glial swelling under hypotonic conditions as BSA. (E) The swelling of glial somata was prevented in the presence of a sodium-free extracellular solution. Data are the mean (± SEM) glial soma area (n = 6–21 per bar) measured after a 4-minute superfusion of retinal slices with the hypotonic solution, and expressed in percent of the soma area before osmotic challenge (100%). Significant difference vs. control: *P < 0.05; **P < 0.01; ***P < 0.001. Significant blocking effect: ••P < 0.01; ••• P < 0.001.

2.3. Involvement of transforming growth factor (TGF)-β receptors It has been shown that brain astrocytes internalize albumin in vesicle-like structures by receptor-mediated endocytosis (Tabernero et al., 2002; Ivens et al., 2007). To determine whether internalization of albumin is implicated in evoking

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –2 76

271

osmotic swelling of retinal glial cells, we preincubated retinal slices with the inhibitor of receptor-mediated endocytosis, phenylarsine oxide. Phenylarsine oxide fully prevented the swelling-inducing effect of albumin (Fig. 3A). The data may

suggest that endocytosis of albumin might represent a step in the cascade that causes albumin-induced osmotic swelling. To confirm the participation of an albumin-binding protein in albumin endocytosis, retinal slices were treated with L. flavus agglutinin, a lectin that blocks glycoprotein-mediated endocytosis by binding to glycosyl residues (Schnitzer et al., 1988). L. flavus agglutinin fully prevented the albumin-induced glial swelling (Fig. 3B). To control the specificity of albumin endocytosis, retinal slices were exposed to the unrelated protein ovalbumin. In contrast to albumin, ovalbumin had no significant swelling-inducing effect when it was administered simultaneously with the hypoosmolar solution while a preincubation of the slices for 15 min induced a significant (P < 0.01) swelling (Fig. 3C). The selectivity of the acute swelling-inducing effect (albumin but not ovalbumin) suggests that the rapid effect of albumin is mediated via a specific receptor. TGF-β receptors were described to function as albuminbinding proteins (Siddiqui et al., 2004) and to mediate the uptake of albumin into brain astrocytes (Ivens et al., 2007). To determine whether TGF-β receptors are involved in albumin endocytosis in retinal glial cells, we tested the selective inhibitor of the TGF-β activin receptor-like kinase, SB431542. As shown in Fig. 3D, this inhibitor blocked the albumin-evoked glial swelling. In addition, preincubation of retinal slices with a neutralizing goat antibody against the TGF-β receptor type II inhibited the albumin-evoked swelling (Fig. 3D). Preincubation of the slices with goat IgG did not prevent the swellinginducing effect of albumin (Fig. 3E). The data suggest that albumin may activate the TGF-β activin receptor-like kinase in retinal glial cells.

Fig. 3 – Involvement of TGF-β receptors in albumin-induced osmotic swelling of retinal glial cells. (A) The inhibitor of receptor-mediated endocytosis, phenylarsine oxide (PAO; 1 μM), prevented the swelling-inducing effect of bovine serum albumin (BSA; 5 μM). The effect of PAO was inhibited in the presence of the following agents: the selective antagonist of group II metabotropic glutamate receptors, LY341495 (100 μM), the P2Y1-selective antagonist, MRS2179 (30 μM), and the antagonist of adenosine A1 receptors, DPCPX (100 nM). (B) Preincubation of retinal slices with L. flavus agglutinin (LFA; 50 μg/ml) prevented the BSA (5 μM)-induced glial swelling. (C) Ovalbumin (OA; 5 μM) induced a swelling of glial somata under hypotonic conditions after preincubation of retinal slices for 15 min and had no effect when it was administered simultaneously with the hypotonic solution. As control, barium chloride (1 mM) was tested. (D) The BSA (5 μM)-evoked glial swelling was prevented by the inhibitor of the TGF-β activin receptor-like kinase, SB431542 (20 μM), and a neutralizing anti-TGF-β receptor type II antibody (aTGF R; 20 μg/ml), respectively. (E) The swelling-inducing effect of BSA (5 μM) was not altered by goat IgG (20 μg/ml). Data are the mean (±SEM) glial soma area (n = 6–18 per bar) measured after a 4-minute superfusion of retinal slices with the hypotonic solution, and expressed in percent of the soma area before osmotic challenge (100%). Significant difference vs. control: *P < 0.05; **P < 0.01; ***P < 0.001. Significant blocking effect: ••P < 0.01; •••P < 0.001. Significant inhibition of the PAO effect: ○P < 0.05; ○○P < 0.01.

2.4.

Receptor-mediated inhibition of glial swelling

We showed previously that retinal glial cells possess an endogenous glutamatergic–purinergic signaling mechanism that mediates the homeostasis of cellular volume under hypotonic conditions (Wurm et al., 2008). This autocrine signaling mechanism can be activated by glutamate, and involves the consecutive release of ATP and adenosine from glial cells, and the successive activation of P2Y1 and adenosine receptors (Wurm et al., 2008). As shown in Fig. 4A, exogenous glutamate, ATP, and adenosine fully prevented the albuminevoked swelling of retinal glial cells. The selective antagonist of group II metabotropic glutamate receptors (mGluRs), LY341495, prevented the swelling-inhibitory effect of glutamate, but not the effects of ATP and adenosine (Fig. 4A). The swelling-inhibitory effects of glutamate and ATP were abrogated by the P2Y1-selective blocker, MRS2179 (Fig. 4B), while the selective antagonist of adenosine A1 receptors, DPCPX, inhibited the effects of all three receptor agonists tested (Fig. 4C). The antagonist of A2A receptors, CSC, did not prevent the effects of glutamate and adenosine (Fig. 4D). The different effects of the receptor antagonists on the swelling-inhibitory action of glutamate, ATP, and adenosine (Figs. 4A–C) suggest that activation of A1 receptors is the most downstream event in the glutamatergic–purinergic signaling cascade. As previously described (Uckermann et al., 2005), the swellinginhibitory effect of triamcinolone acetonide was abrogated by DPCPX (Fig. 4C), suggesting that this steroid evokes a

272

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –27 6

Fig. 4 – Activation of glutamatergic and purinergic receptors inhibits the albumin-induced osmotic swelling of retinal glial cells. Retinal slices were superfused with a hypoosmolar solution containing bovine serum albumin (BSA; 5 μM). (A) The BSA-induced glial swelling was prevented by glutamate (Glu), ATP, and adenosine (Ade), respectively. The swelling-inhibitory effect of glutamate, but not of ATP and adenosine, was abrogated by the antagonist of group II mGluRs, LY341495 (100 μM). (B) The effects of glutamate and ATP, but not of adenosine, were prevented in the presence of the P2Y1-selective blocker, MRS2179 (30 μM). (C,D) The selective antagonist of adenosine A1 receptors, DPCPX (100 nM) (C), but not the antagonist of A2A receptors, CSC (200 nM) (D), prevented the swelling-inhibitory effects of glutamate and adenosine. DPCPX prevented also the effects of ATP and triamcinolone acetonide (Triam; 100 μM). (E) The inhibitory effect of glutamate on glial swelling was abrogated by the ecto-ATPase inhibitor, ARL-67156 (50 μM), and remained unaltered in the presence of the ectonucleotidase blocker, AOPCP (250 μM). (F) The effects of glutamate and ATP were prevented by the antagonist of nucleoside transporters, NBTI (10 μM). Receptor agonists were co-administered with the hypotonic solution, and were tested at following concentrations: glutamate, 1 mM; ATP, 10 μM; adenosine, 10 μM. Data are the mean (±SEM) glial soma area (n = 6–20 per bar) measured after a 4-minute superfusion of retinal slices with the hypotonic solution, and expressed in percent of the soma area before osmotic challenge (100%). Significant difference vs. control: **P < 0.01; ***P < 0.001. Significant blocking effect: •P < 0.05; •• P < 0.01; •••P < 0.001. Significant inhibition of the agonist effect: ○○P < 0.01; ○○○P < 0.001.

release of endogenous adenosine which activates A1 receptors. Moreover, the swelling-inhibitory effect of the endocytosis inhibitor, phenylarsine oxide, was abrogated in the presence of blockers of mGluRs, P2Y1, and A1 receptors (Fig. 3A), suggesting that the effect was (at least in part) mediated by activation of the autocrine glutamatergic–purinergic mechanism of cell volume regulation. The data suggest that exogenous glutamate evokes a release of ATP and/or ADP from retinal glial cells which activate P2Y1. The antagonistic action of the ecto-ATPase inhibitor, ARL-67156, on the effect of glutamate (Fig. 4E) suggests that ATP is extracellularly converted to ADP which is a known agonist of P2Y1 (Abbracchio et al., 2006). However,

ADP seems to be not further converted to adenosine, as the ectonucleotidase blocker, AOPCP, did not prevent the effect of glutamate (Fig. 4E). Instead, adenosine is apparently released from the cells via transporters, as the antagonist of nucleoside transporters, NBTI, fully prevented the effects of glutamate and ATP (Fig. 4F).

2.5.

Intracellular mediation of swelling inhibition

Ion efflux through potassium and chloride channels, resulting in equalization of the transmembrane osmotic gradient, is a mechanism to prevent glial swelling (Wurm et al., 2008). To prove whether adenosine activates ion channels in retinal

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –2 76

glial cells, we tested potassium and chloride channel blockers. As shown in Fig. 5A, the class III antiarrhythmic drug clofilium, and the chloride channel blocker NPPB, both prevented the swelling-inhibitory effect of adenosine on albumin-induced glial swelling. Clofilium is a known inhibitor of two pore-domain potassium channels (Niemeyer et al., 2001) which are also expressed in retinal glial cells (Skatchkov et al., 2006). To test whether intracellular calcium signaling

plays a role in adenosine-mediated opening of ion channels, we preincubated the retinal slices with the cell-permeable calcium chelator, BAPTA-AM. However, calcium chelation did not prevent the swelling-inhibitory effects of glutamate and adenosine (Fig. 5B). Inhibition of the phospholipase C by U73122 (10 μM) also did not alter the effect of glutamate (not shown). The blocking effect of the inhibitor of protein kinase A, H-89 (Fig. 5C), on the effect of adenosine suggests an involvement of the protein kinase A signaling pathway in the A1 receptor-mediated activation of ion channels in retinal glial cells. The absence of an effect of calcium chelation on the action of glutamate (Fig. 5B) also suggests that the glutamateevoked release of ATP from retinal glial cells is mediated by a calcium-independent mechanism.

3.

Fig. 5 – Intracellular signaling involved in inhibition of the albumin-induced osmotic swelling of retinal glial cells. Retinal slices were superfused with a hypoosmolar solution containing bovine serum albumin (BSA; 5 μM). (A) The inhibitory effect of adenosine (200 μM) on the BSA-induced glial swelling was prevented in the presence of the potassium channel blocker clofilium (10 μM) and the chloride channel blocker NPPB (100 μM), respectively. (B) The inhibitory effects of glutamate (Glu; 1 mM) and adenosine (Ade; 10 μM) were not prevented by preincubation of the slices with the cell-permeable calcium chelator, BAPTA-AM (100 μM). (C) The effect of adenosine (10 μM) was largely abrogated by the inhibitor of the protein kinase A, H-89 (1 μM). Data are the mean (± SEM) glial soma area (n = 7–14 per bar) measured after a 4-minute superfusion of retinal slices with the hypotonic solution, and expressed in percent of the soma area before osmotic challenge (100%). Significant difference vs. control: **P < 0.01; ***P < 0.001. Significant blocking effect: ••P < 0.01; •••P < 0.001. Significant inhibition of the agonist effect: ○○P < 0.01; ○○○P < 0.001.

273

Discussion

In the present study, we show that serum albumin evokes a swelling of retinal glial cells under hypotonic conditions. The effect of albumin was dose-dependent, and was mediated by activation of the TGF-β receptor type II, oxidative stress, the production of inflammatory lipids, and an influx of sodium from the extracellular space. The action of arachidonic acid and of its metabolites, especially prostaglandin E2, was causally implicated in the development of retinal edema (Miyake and Ibaraki, 2002) and in osmotic swelling of retinal glial cells (Uckermann et al., 2005). Oxidative stress is known to stimulate the activities of phospholipase A2 and cyclooxygenase (Landino et al., 1996; Lambert et al., 2006). Inhibition of the sodium pump by arachidonic acid and prostaglandins results in intracellular sodium overload (Staub et al., 1994). We found that the swelling of retinal glial cells was abrogated in the presence of a sodium-free solution (Fig. 2E), suggesting that it was caused by intracellular sodium overload associated with a water influx into the cells. Arachidonic acid is also known to block membrane channels such as volume-regulated anion and (in retinal glial cells) outwardly rectifying potassium channels (Lambert, 1991; Bringmann et al., 1998) which could mediate a compensatory efflux of osmolytes such as amino acids, chloride, and potassium ions. However, the mechanism of albumin-evoked glial swelling remains to be further elucidated. We found that albumin-induced glial swelling was mediated by activation of the TGF-β receptor type II and the TGF-β activin receptor-like kinase (Fig. 3D). In brain astrocytes, albumin is internalized by TGF-β receptor-mediated endocytosis (Ivens et al., 2007). Though two endocytosis inhibitors, phenylarsine oxide (Fig. 3A) and L. flavus agglutinin (Fig. 3B), prevented the albumin-induced cell swelling suggesting that retinal glial cells internalize albumin, an involvement of albumin endocytosis remains to be confirmed by more direct methods. Association between endocytosis and generation of oxygen radicals and inflammatory lipids has been also described in other cell systems. In proximal tubule cells, the receptor-mediated endocytosis of albumin results in activation of NADPH oxidase and generation of reactive oxygen species (Whaley-Connell et al., 2007). In macrophages, phagocytosis is followed by an abrupt superoxide formation by the NADPH oxidase complex (Park, 2003). Activation of

274

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –27 6

phospholipase A2 plays an integral role in phagocytosis; phospholipase A2-derived arachidonic acid stimulates both the activity of the NADPH oxidase and the membrane redistribution during phagocytosis (Lennartz, 1999). On the other hand, the swelling-inhibitory effect of the endocytosis inhibitor, phenylarsine oxide, was prevented by blockers of mGluRs, P2Y1, and A1 receptors (Fig. 3A). This may suggest that inhibition of the downstream events after TGF-β receptor activation could result in the activation of the autocrine glutamatergic–purinergic signaling that regulates cellular volume under anisoosmotic conditions. Whether the present results have significance for in situ conditions, remains to be determined in future experiments. It has been shown that disruption of the blood–brain barrier results in the development of epileptiform activity in the cerebral cortex of rats (Seiffert et al., 2004; Ivens et al., 2007). This effect was shown to be mediated by TGF-β receptormediated endocytosis of albumin by astrocytes, and albumininduced gliosis characterized by a downregulation of glial Kir channels (Ivens et al., 2007). The reduced potassium buffering capacity of reactive astrocytes augmented activity-dependent potassium accumulation, and resulted in neuronal hyperexcitability and epileptiform activity (Ivens et al., 2007). A similar decrease in glial Kir channel expression was previously observed in animal models of ischemic and inflammatory retinal diseases, and in the retina of diabetic rats (Pannicke et al., 2004, 2005, 2006). These pathological conditions are associated with vascular leakage and induction of glial swelling under hypotonic conditions (Pannicke et al., 2004, 2005, 2006). The present data suggest a further mechanism how serum albumin may contribute to glia-mediated epilepsy. Albumin-induced osmotic swelling of glial cells will result in a decrease in the extracellular space volume which has been shown to favor neuronal hyperexcitability (Chebabo et al., 1995). The decrease in the extracellular space volume is strengthened by the activity-dependent swelling of neuronal cell bodies and synapses (Uckermann et al., 2004a). We found that albumin induces cellular swelling under hypotonic but not isotonic conditions (Fig. 1B). Hypoosmotic gradients across glial membranes are present in the neural tissues of the brain and retina in periods of intense neuronal activity which are associated with a decrease in the extracellular osmolarity because the decrease in sodium exceeds the increase in potassium by a factor of two which is accompanied by a decrease in the chloride concentration (Dietzel et al., 1989; Dmitriev et al., 1999). Under pathological conditions, an osmotic gradient can be also present in situ at the glio-vascular interface, e.g. in cases of renal and hepatic failures resulting in hyponatremia and hypoalbuminemia. A decrease in blood osmolarity represents a major pathogenic factor in the formation of retinal edema (Gardner et al., 2002), as well as of brain edema, e.g. in end-stage liver disease (Heuman et al., 2004). Neuronal hyperexcitation, a characteristic of ischemic– hypoxic conditions, will enhance the osmotic gradient at the glio-vascular interface, via accumulation of neuron-derived osmolytes, e.g. potassium, in glial cells (Bringmann et al., 2004). The induction of osmotic swelling of retinal glial cells suggests that albumin alters the water transport across glial membranes. This should impair the resolution of retinal edema in situ. Triamcinolone acetonide is used clinically for

the rapid resolution of retinal edema (Fraser-Bell et al., 2008). Here, we show that triamcinolone prevents the albumininduced glial swelling (Fig. 2B). The effect of triamcinolone was apparently mediated by activation of the most downstream steps of the glutamatergic–purinergic signaling cascade which regulates the volume of retinal glial cells (Wurm et al., 2008), i.e. it stimulates the release of endogenous adenosine which subsequently activates A1 receptors (Fig. 4C), resulting in the opening of potassium and chloride channels. Because the water transport through retinal glial cells is assumed to be coupled to transcellular potassium currents (Nagelhus et al., 1999; Bringmann et al., 2004), opening of potassium channels should improve both the potassium buffering capacity of retinal glial cells and the water transport through the cells. This suggests that triamcinolone (in addition to its protective effects on vasogenic edema) may stimulate the glia-mediated fluid clearance from the edematous retinal tissue.

4.

Experimental procedures

4.1.

Materials

The neutralizing goat anti-TGF-β receptor type II antibody was from R&D Systems (Minneapolis, MN; AF-241-NA). Goat IgG was from Millipore (Temecula, CA). Mitotracker Orange (chloromethyltetramethylrosamine) was purchased from Molecular Probes (Eugene, OR). BSA (fraction V) was from Carl Roth (Karlsruhe, Germany). LY341495 and MRS2179 were from Tocris Cookson (Bristol, UK). Fatty acid-free BSA (Sigma A6003) and all other substances used were obtained from SigmaAldrich (Taufkirchen, Germany), unless stated otherwise.

4.2.

Animals

All experiments were done in accordance with the European Communities Council Directive 86/609/EEC, and were approved by the local authorities. Forty two adult Long–Evans rats were used. Animals had free access to water and food in an air-conditioned room on a 12-hour light–dark cycle. The animals were sacrificed with carbon dioxide, and the eyes were removed.

4.3.

Preparation of retinal slices

After dissection of the front half of the eyeballs including the cornea and lens, two opposite incisions were made in the sclera of the remaining eyeball, and the neural retina and vitreous were removed and kept submerged in extracellular solution (see Section 4.5). After disjunction of the neural retina and vitreous, a piece (5 × 5 mm) was cut from each retina containing the mid-periphery of the retina from which the recordings were made in the further course of the experiments. The retinal pieces were mounted, with the photoreceptor side down, onto membrane filters (mixed cellulose ester, 0.45 μm pore size; Schleicher & Schuell MicroScience, Dassel, Germany) and kept submerged in extracellular solution in a tissue chopper equipped with a razor blade. Retinal slices (thickness, 1 mm) were cut from the tissues adhering to the membrane filters.

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –2 76

4.4.

Glial cell swelling

All experiments were performed at room temperature (20– 23 °C). The filter stripes with the retinal slices were fixed to the bottom of a superfusion chamber (cut surface upward) by plastic holders at both ends of the slices, and kept submerged in extracellular solution. The chambers were mounted on the stage of an upright confocal laser scanning microscope (LSM 510 Meta; Zeiss, Oberkochen, Germany). The slices were loaded with the vital dye Mitotracker Orange (10 μM; 3 min); it has been shown that this dye stains selectively the somata of Müller glial cells in the inner nuclear layer of retinal tissues (Uckermann et al., 2004b). The stem solution of the dye was prepared in dimethylsulfoxide and resolved in saline. Thereafter, the slices were continuously superfused with extracellular solution at a flow rate of 2 ml/min. Recordings were made with an Achroplan 63×/0.9 water immersion objective. The pinhole was set at 172 μm; the thickness of the optical section was adjusted to 1 μm. Mitotracker Orange was excited at 543 nm with a HeNe laser, and emission was recorded with a 560 nm long-pass filter. Images were obtained with an x–y frame size of 256 × 256 pixel (73.1 × 73.1 μm). To monitor volume changes of retinal glial cells in response to hypotonic challenge, the somata of dye-filled glial cells in the inner nuclear layer of retinal slices were focussed at the plane of their maximal extension.

4.5.

Solutions

A gravity-fed system with multiple reservoirs was used to perfuse the recording chamber continuously with extracellular solution; test and blocking substances were applied by rapidly changing the perfusate. The bathing solution in the superfusion chamber was totally changed within ∼ 2 min. The extracellular solution consisted of (mM) 136 NaCl, 3 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, and 11 glucose, adjusted to pH 7.4 with Tris. The hypoosmolar solution (60% of control osmolarity) was made up by adding distilled water. The sodium-free solution was made by replacing sodium with choline, and was applied simultaneously with the hypoosmolar solution. Barium chloride (1 mM) was added to iso- and hypoosmolar solutions; the slices were superfused with barium-containing solution for 10 min before administration of the hypoosmolar solution. Blocking substances, antibodies, dithiothreitol, and BAPTA-AM were preincubated by superfusion of the slices for 15 to 60 min (Supplementary Table 1) before administration of the hypoosmolar solution; these agents were also present in the hypoosmolar solution. BSA, receptor ligands, inflammatory lipids, H2O2, triamcinolone acetonide, monensin, and the sodium-free solution were applied simultaneously with the hypoosmolar solution.

4.6.

Data analysis

To determine the extent of glial swelling, the cross-sectional area of Mitotracker Orange-stained glial somata in the inner nuclear layer of retinal slices was measured off-line using the image analysis software of the LSM. Bar diagrams display the mean (±SEM) cross-sectional area of the somata that was measured after a 4-minute perfusion with the hypoosmolar

275

solution, in percent of the soma area measured before hypotonic challenge (100%). Statistical analysis was made using the Prism program (Graphpad Software, San Diego, CA); significance was determined by Mann–Whitney U test for two groups and Kruskal–Wallis test followed by Dunn's comparison for multiple groups.

Acknowledgments This study was supported by grants from the Deutsche Forschungsgemeinschaft (RE 849/10-2, RE 849/12-1, GRK 1097/1), the Bundesministerium für Bildung und Forschung (DLR/01GZ0703), and the Faculty of Medicine of the University of Leipzig (NBL Formel.1-133).

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.brainres.2009.12.067. REFERENCES

Abbracchio, M.P., Burnstock, G., Boeynaems, J.M., Barnard, E.A., Boyer, J.L., Kennedy, C., Knight, G.E., Fumagalli, M., Gachet, C., Jacobson, K.A., Weisman, G.A., 2006. International Union of Pharmacology LVIII: update on the P2Y G protein-coupled nucleotide receptors: from molecular mechanisms and pathophysiology to therapy. Pharmacol. Rev. 58, 281–341. Belayev, L., Liu, Y., Zhao, W., Busto, R., Ginsberg, M.D., 2001. Human albumin therapy of acute ischemic stroke: marked neuroprotective efficacy at moderate doses and with a broad therapeutic window. Stroke 32, 553–560. Bringmann, A., Skatchkov, S.N., Biedermann, B., Faude, F., Reichenbach, A., 1998. Alterations of potassium channel activity in retinal Müller glial cells induced by arachidonic acid. Neuroscience 86, 1291–1306. Bringmann, A., Reichenbach, A., Wiedemann, P., 2004. Pathomechanisms of cystoid macular edema. Ophthalmic Res. 36, 241–249. Chebabo, S.R., Hester, M.A., Aitken, P.G., Somjen, G.G., 1995. Hypotonic exposure enhances synaptic transmission and triggers spreading depression in rat hippocampal tissue slices. Brain Res. 695, 203–216. Claudio, L., Martiney, J.A., Brosnan, C.F., 1994. Ultrastructural studies of the blood–retina barrier after exposure to interleukin-1β or tumor necrosis factor-α. Lab. Invest. 70, 850–861. Dietzel, I., Heinemann, U., Lux, H.D., 1989. Relations between slow extracellular potential changes, glial potassium buffering, and electrolyte and cellular volume changes during neuronal hyperactivity in cat brain. Glia 2, 25–44. Dmitriev, A.V., Govardovskii, V.I., Schwahn, H.N., Steinberg, R.H., 1999. Light-induced changes of extracellular ions and volume in the isolated chick retina-pigment epithelium preparation. Vis. Neurosci. 16, 1157–1167. Emerson Jr, T.E., 1989. Unique features of albumin: a brief review. Crit. Care Med. 17, 690–694. Fraser-Bell, S., Kaines, A., Hykin, P.G., 2008. Update on treatments for diabetic macular edema. Curr. Opin. Ophthalmol. 19, 185–189. Gardner, T.W., Antonetti, D.A., Barber, A.J., LaNoue, K.F., Levison, S.W., The Penn State Retina Research Group, 2002. Diabetic

276

BR A I N R ES E A RC H 1 3 1 7 ( 2 01 0 ) 2 6 8 –27 6

retinopathy: more than meets the eye. Surv. Ophthalmol. 47, S253–S262. Heuman, D.M., Abou-Assi, S.G., Habib, A., Williams, L.M., Stravitz, R.T., Sanyal, A.J., Fisher, R.A., Mihas, A.A., 2004. Persistent ascites and low serum sodium identify patients with cirrhosis and low MELD scores who are at high risk for early death. Hepatology 40, 802–810. Ivens, S., Kaufer, D., Flores, L.P., Bechmann, I., Zumsteg, D., Tomkins, O., Seiffert, E., Heinemann, U., Friedman, A., 2007. TGF-β receptor-mediated albumin uptake into astrocytes is involved in neocortical epileptogenesis. Brain 130, 535–547. Kimelberg, H.K., 2004. Water homeostasis in the brain: basic concepts. Neuroscience 129, 851–860. Landino, L.M., Crews, B.C., Timmons, M.D., Morrow, J.D., Marnett, L.J., 1996. Peroxynitrite, the coupling product of nitric oxide and superoxide, activates prostaglandin biosynthesis. Proc. Natl. Acad. Sci. U. S. A. 93, 15069–15074. Lambert, I.H., 1991. Effect of arachidonic acid on conductive Na, K and anion transport in Ehrlich ascites tumor cells under isotonic and hypotonic conditions. Cell. Physiol. Biochem. 1, 177–194. Lambert, I.H., Pedersen, S.F., Poulsen, K.A., 2006. Activation of PLA2 isoforms by cell swelling and ischaemia/hypoxia. Acta Physiol. (Oxf.) 187, 75–85. Lennartz, M.R., 1999. Phospholipases and phagocytosis: the role of phospholipid-derived second messengers in phagocytosis. Int. J. Biochem. Cell Biol. 31, 415–430. Miyake, K., Ibaraki, N., 2002. Prostaglandins and cystoid macular edema. Surv. Ophthalmol. 47, S203–S218. Nagelhus, E.A., Horio, Y., Inanobe, A., Fujita, A., Haug, F.M., Nielsen, S., Kurachi, Y., Ottersen, O.P., 1999. Immunogold evidence suggests that coupling of K+ siphoning and water transport in rat retinal Müller cells is mediated by a coenrichment of Kir4.1 and AQP4 in specific membrane domains. Glia 26, 47–54. Niemeyer, M.I., Cid, L.P., Barros, L.F., Sepulveda, F.V., 2001. Modulation of the two-pore domain acid-sensitive K+ channel TASK-2 (KCNK5) by changes in cell volume. J. Biol. Chem. 276, 43166–44374. Pannicke, T., Iandiev, I., Uckermann, O., Biedermann, B., Kutzera, F., Wiedemann, P., Wolburg, H., Reichenbach, A., Bringmann, A., 2004. A potassium channel-linked mechanism of glial cell swelling in the postischemic retina. Mol. Cell. Neurosci. 26, 493–502. Pannicke, T., Uckermann, O., Iandiev, I., Wiedemann, P., Reichenbach, A., Bringmann, A., 2005. Ocular inflammation alters swelling and membrane characteristics of rat Müller glial cells. J. Neuroimmunol. 161, 145–154. Pannicke, T., Iandiev, I., Wurm, A., Uckermann, O., vom Hagen, F., Reichenbach, A., Wiedemann, P., Hammes, H.-P., Bringmann, A., 2006. Diabetes alters osmotic swelling characteristics and membrane conductance of glial cells in rat retina. Diabetes 55, 633–639. Park, J.-B., 2003. Phagocytosis induces superoxide formation and apoptosis in macrophages. Exp. Mol. Med. 35, 325–335. Rumpel, H., Nedelcu, J., Aguzzi, A., Martin, E., 1997. Late glial swelling after acute cerebral hypoxia–ischemia in the neonatal

rat: a combined magnetic resonance and histochemical study. Pediatr. Res. 42, 54–59. Schnitzer, J.E., Carley, W.W., Palade, G.E., 1988. Albumin interacts specifically with a 60-kDa microvascular endothelial glycoprotein. Proc. Natl. Acad. Sci. U. S. A. 85, 6773–6777. Seiffert, E., Dreier, J.P., Ivens, S., Bechmann, I., Tomkins, O., Heinemann, U., Friedman, A., 2004. Lasting blood–brain barrier disruption induces epileptic focus in the rat somatosensory cortex. J. Neurosci. 24, 7829–7836. Siddiqui, S.S., Siddiqui, Z.K., Malik, A.B., 2004. Albumin endocytosis in endothelial cells induces TGF-β receptor II signaling. Am. J. Physiol. Lung Cell. Mol. Physiol. 286, L1016–L1026. Skatchkov, S.N., Eaton, M.J., Shuba, Y.M., Kucheryavykh, Y.V., Derst, C., Veh, R.W., Wurm, A., Iandiev, I., Pannicke, T., Bringmann, A., Reichenbach, A., 2006. Tandem-pore domain potassium channels are functionally expressed in retinal (Müller) glial cells. Glia 53, 266–276. Staub, F., Winkler, A., Peters, J., Kempski, O., Kachel, V., Baethmann, A., 1994. Swelling, acidosis, and irreversible damage of glial cells from exposure to arachidonic acid in vitro. J. Cereb. Blood Flow Metab. 14, 1030–1039. Stefánsson, E., Wilson, C.A., Lightman, S.L., Kuwabara, T., Palestine, A.G., Wagner, H.G., 1987. Quantitative measurements of retinal edema by specific gravity determinations. Invest. Ophthalmol. Vis. Sci. 28, 1281–1289. Tabernero, A., Velasco, A., Granda, B., Lavado, E.M., Medina, J.M., 2002. Transcytosis of albumin in astrocytes activates the sterol regulatory element-binding protein-1, which promotes the synthesis of the neurotrophic factor oleic acid. J. Biol. Chem. 277, 4240–4246. Uckermann, O., Vargová, L., Ulbricht, E., Klaus, C., Weick, M., Rillich, K., Wiedemann, P., Reichenbach, A., Syková, E., Bringmann, A., 2004a. Glutamate-evoked alterations of glial and neuronal cell morphology in the guinea-pig retina. J. Neurosci. 24, 10149–10158. Uckermann, O., Iandiev, I., Francke, M., Franze, K., Grosche, J., Wolf, S., Kohen, L., Wiedemann, P., Reichenbach, A., Bringmann, A., 2004b. Selective staining by vital dyes of Müller glial cells in retinal wholemounts. Glia 45, 59–66. Uckermann, O., Kutzera, F., Wolf, A., Pannicke, T., Reichenbach, A., Wiedemann, P., Wolf, S., Bringmann, A., 2005. The glucocorticoid triamcinolone acetonide inhibits osmotic swelling of retinal glial cells via stimulation of endogenous adenosine signaling. J. Pharmacol. Exp. Ther. 315, 1036–1045. Whaley-Connell, A.T., Morris, E.M., Rehmer, N., Yaghoubian, J.C., Wei, Y., Hayden, M.R., Habibi, J., Stump, C.S., Sowers, J.R., 2007. Albumin activation of NAD(P)H oxidase activity is mediated via Rac1 in proximal tubule cells. Am. J. Nephrol. 27, 15–23. Wurm, A., Pannicke, T., Wiedemann, P., Reichenbach, A., Bringmann, A., 2008. Glial cell-derived glutamate mediates autocrine cell volume regulation in the retina: activation by VEGF. J. Neurochem. 104, 386–399. Zoellner, H., Hofler, M., Beckmann, R., Hufnagl, P., Vanyek, E., Bielek, E., Wojta, J., Fabry, A., Lockie, S., Binder, B.R., 1996. Serum albumin is a specific inhibitor of apoptosis in human endothelial cells. J. Cell. Sci. 109, 2571–2580.