ZnT-1 Expression in Astroglial CellsProtects Against Zinc Toxicity and

Slows the Accumulation ofIntracellular Zinc

CHRISTIANE NOLTE,1 ARIEL GORE,2 ISRAEL SEKLER,2* WOLFGANG KRESSE,1MICHAL HERSHFINKEL,3 ANJA HOFFMANN,1 HELMUT KETTENMANN,1

AND ARIE MORAN2

1Max-Delbruck-Center for Molecular Medicine, Berlin, Germany2Department of Physiology, Faculty of Health Sciences, Ben-Gurion University of the Negev,

Beer-Sheva, Israel3Department of Morphology, Faculty of Health Sciences, Ben-Gurion University of the Negev,

Beer-Sheva, Israel

KEY WORDS ZnT-1; astroglia; zinc toxicity; L-type calcium channels; transfection

ABSTRACT Zinc ions are emerging as an important factor in the etiology of neuro-degenerative disorders and in brain damage resulting from ischemia or seizure activity.High intracellular levels of zinc are toxic not only to neurons but also to astrocytes, themajor population of glial cells in the brain. In the present study, the role of ZnT-1 inreducing zinc-dependent cell damage in astrocytes was assessed. Zinc-dependent celldamage was apparent within 2 h of exposure to zinc, and occurred within a narrow rangeof �200 �M. Pretreatment with sublethal concentrations of zinc rendered astrocytes lesssensitive to toxic zinc levels, indicating that preconditioning protects astrocytes fromzinc toxicity. Fluorescence cell imaging revealed a steep reduction in intracellular zincaccumulation for the zinc-pretreated cells mediated by L-type calcium channels. Heter-ologous expression of ZnT-1 had similar effects; intracellular zinc accumulation wasslowed down and the sensitivity of astrocytes to toxic zinc levels was reduced, indicatingthat this is specifically mediated by ZnT-1 expression. Immunohistochemical analysisdemonstrated endogenous ZnT-1 expression in cultured astroglia, microglia, and oligo-dendrocytes. Pretreatment with zinc induced a 4-fold increase in the expression of theputative zinc transporter ZnT-1 in astroglia as shown by immunoblot analysis. Theelevated ZnT-1 expression following zinc priming or after heterologous expression ofZnT-1 may explain the reduced zinc accumulation and the subsequent reduction insensitivity toward toxic zinc levels. Induction of ZnT-1 may play a protective role whenmild episodes of stroke or seizures are followed by a massive brain insult.© 2004 Wiley-Liss, Inc.

INTRODUCTION

Zinc is an essential ligand for many proteins andcofactor for metalloenzymes (Vallee and Falchuk, 1993;Berg and Shi, 1996). In addition, zinc is a critical factorin the regulation of a large number of genes via its rolein the activation of transcription factors, as well as bypromoting the link between proteins and DNA via zincfingers (Berg and Shi, 1996). In the mammalian centralnervous system (CNS), chelatable zinc is unevenly dis-tributed, with its highest concentrations in the hip-pocampus and cerebral cortex. In these regions, chelat-

able zinc is localized exclusively to glutamate-containing synaptic terminals and is released alongwith the neurotransmitter during neuronal activity

Grant sponsor: German Israel Foundation (GIF); Grant number: I-588-99.1/1998; Grant sponsor: German Research Foundation; Grant number: HO 2205/1.

*Correspondence: Israel Sekler, Department of Physiology, Faculty of HealthSciences, Ben Gurion University of the Negev, Beer-Sheva, Israel.E-mail: sekler@bgumail.bgu.ac.il

Received 11 March 2004; Accepted 6 April 2004

DOI 10.1002/glia.20065

Published online 15 June 2004 in Wiley InterScience (www.interscience.wiley.com).

GLIA 48:145–155 (2004)

© 2004 Wiley-Liss, Inc.

(Crawford and Connor, 1972; Frederickson et al., 1988;Frederickson, 1989). Extracellular zinc can modulatethe activity of several ion channels, most notably N-methyl-D-aspartate (NMDA) and GABA receptors(Huang, 1997; Kapur and Macdonald, 1997; Vogt et al.,2000). During excitotoxic events such as brain ischemiaor seizures, zinc is released into the synaptic cleft,rapidly entering the postsynaptic cells through voltage-gated Ca2� channels, NMDA and AMPA receptors(Weiss and Sensi, 2000; Frederickson et al., 1988; Sensiet al., 1997; Tonder et al., 1990; Weiss et al., 1993)leading to widespread neuronal death (Choi and Koh,1998). The role of zinc in this process has been con-firmed by experiments in which the chelation of zinc,but not of calcium ions, protected neurons againstdeath following ischemia (Koh et al., 1996). Zinc re-leased from intracellular stores is also recently emerg-ing to play an important role for increased intracellularzinc levels and cell death after sustained seizures (Leeet al., 2000; Bossy-Wetzel et al., 2004).

Several different but complementary mechanismsare employed in the brain to maintain intracellularzinc homeostasis. For instance, proteins that bind andsequester zinc inside neurons, e.g., metallothioneins,are known to be involved (Aschner, 1998). Zinc trans-porters, ZnT 1–4, are suggested to regulate the levelsof free or loosely bound zinc inside cells. Several lines ofevidence have shown, that ZnT-3 is essential for pack-aging of zinc into the synaptic vesicles (Cole et al.,1999). Similarly, ZnT-1 has been shown to protect neu-ronal cell lines against toxic zinc (Tsuda et al., 1997;Kim et al., 2000). The mechanism linking zinc trans-porters to zinc transport or sequestration are, however,not well understood.

Glial cells are key players in the regulation of brainion homeostasis. Astrocytes, for example, regulate lev-els of extracellular K� and are key elements in thecontrol of the extracellular milieu of neurons (Ransomet al., 2003). Although zinc is toxic to cultured neuronsalready at lower concentrations compared with cul-tured astrocytes (Swanson and Sharp, 1992; Dineley etal., 2000), levels of zinc that are toxic to astrocytes maybe reached in the brain during excitotoxic episodes(Choi and Koh, 1998; Vogt et al., 2000). Indeed, mas-sive glial cell death was reported following ischemicinsults in rats (Choi and Koh, 1998). Despite their clearimportance to pathophysiological processes linked tobrain zinc homeostasis, the expression and activity ofzinc transporters and their role in promoting resis-tance to zinc toxicity in these cells are not understood.We studied the role of zinc transport and its regulationin zinc toxicity of cultured astrocytes. We demonstratein the present study that ZnT-1 is expressed in glialcells in culture. Furthermore, by treating astrogliawith sublethal concentrations of zinc, ZnT-1 is upregu-lated, resulting in decreased zinc accumulation in as-trocytes. This “priming” process is accompanied by de-creased susceptibility of glial cells to subsequentexposure to toxic zinc. The same effects, i.e., reducedsensitivity to toxic zinc and slowed accumulation of

intracellular zinc, can be achieved by heterologoustransfection of astrocytes with ZnT-1.

MATERIALS AND METHODSCell Culture and Transfection Method

Astrocytes were prepared from cortex of newbornWistar rats as described (McCarthy and de Vellis,1980). Briefly, cortical tissue was carefully freed fromblood vessels and meninges, trypsinized, and gentlytriturated with a fire-polished pipette in the presenceof 0.05% DNase (Worthington Biochem, Freehold, NJ).After two washes, cells were cultured in 75-cm2 plateson poly-L-lysine (PLL)-coated coverslips, using culturedbasal medium Eagle’s (BME)/10% fetal calf serum(FCS). One day later, cultures were washed twice withHank’s balanced salt solution (HBSS) to remove cellu-lar debris and maintained for 4 days. After reaching asubconfluent state, cellular debris, microglial cells, andoligodendrocytes, as well as their early precursor cells,were dislodged by manual shaking and removed bywashing with HBSS. The purity of the astrocytes wasroutinely determined by immunofluorescence using apolyclonal antibody against glial fibrillary acidic pro-tein (GFAP) (DAKO, Hamburg, Germany), a specificastrocyte marker. The cultures showed more than 90%cells positive for GFAP.

Astrocytes cultures were cotransfected with a plas-mid (4 �g) that allows ZnT-1 expression (kindly pro-vided by Dr. Palmiter, University of Washington) and aplasmid for expression of enhanced yellow fluorescentprotein (EYFP) (0.4 �g, BD Biosciences Clontech, PaloAlto, CA). The latter was used for tracing the trans-fected cells. Cotransfection with the same amount ofvector (without ZnT-1 insert) and EYFP plasmidserved as control (mock-transfected). Transfection wasperformed in 30-mm tissue culture plates using theEscort-IV reagent (Sigma, Rehovot, Israel) according tothe manufacturer’s instructions with 10 �l transfectionreagent. Transfection medium was replaced following6.5 h and transfected cells were used for experimentsafter 72 h in culture.

Cultures of microglial cells and oligodendrocyteswere prepared from cortical tissue of newborn rats asdescribed (McCarthy and deVellis, 1980). Cells wereseeded into PLL-coated 75-cm2 culture bottles at adensity of 5 � 106 cells/bottle, and maintained in BME/10% FCS and antibiotics. After 9–12 days, microglialcells were separated from the underlying astrocyticmonolayer by gentle shaking. After that oligodendro-cyte precursors were removed by vigorous agitation.Cells were plated onto PLL-coated coverslips at a den-sity of 0.5–1 � 105/cm2.

Assessment of Zinc Toxicity in Glial Cells

The process of zinc-induced glial cell damage wasobserved under phase contrast microscopy and fol-

146 NOLTE ET AL.

lowed by time-lapse video microscopy as described pre-viously (Nolte et al., 1996). In addition, cell death wasanalyzed using the lactate dehydrogenase (LDH) assay(Roche Molecular Biochemicals, Mannheim, Germany).Briefly, confluent astrocytes or microglial cells (1.5 �105 cells/well) were grown in 24-well plates withDMEM/2% FCS. Cells were exposed to increasing con-centrations of extracellular zinc. Experiments wereperformed in triplicates. The amount of lactate dehy-drogenase (LDH) released into the extracellular me-dium 24 h after exposure to zinc was determined ac-cording to the supplier’s instructions. LDH values werenormalized to the value of LDH released after 2-htreatment with 0.1% Triton X-100, which was definedas 100% cell death. LD50 values (�M zinc) were derivedfrom these curves. The shift of LD50 values in culturesprimed with sublethal zinc concentrations comparedwith unprimed cultures was determined from four in-dependent experiments. An unpaired t-test was em-ployed to test for statistical significance of shift in theLDH curves.

Zinc toxicity was also measured in astrocytes heter-ologously expressing ZnT-1 or transfected with vector.Cells were co-transfected with ZnT-1 (or vector, 4 �g)and EGFP (0.4 �g) as a reporter gene as describedabove. After 72 h in culture, cells were exposed to 400�M zinc for 30 min at 25°C and were then washed inculture medium; 5 h later, cells were labeled with themembrane permeable nuclear dye propidium iodide(PI, 5 �g/ml, 1 min). Cell death was determined usinga confocal fluorescence microscope (Zeiss, LSM 510).Excitation wavelength of 488 nm was used for thereporter gene, and 543 nm for the PI.

Imaging of Zinc Flux

The fluorescent dye Fura-2/AM was used to measurezinc fluxes. Since the affinity of Fura-2/AM to zinc is100-fold higher in comparison to calcium, it is an ex-cellent tool to measure concentration changes of zinc inthe nanomolar range (Atar et al., 1995). Astrocytes oncoverslips were loaded with 5 �M Fura-2 for 30 min(Molecular Probes, Eugene, OR) in a normal physiolog-ical solution composed of (in mM): NaCl 120; KCl 5.4;CaCl2 1.8; MgCl2 1; HEPES/NaOH 10; glucose 20, pH7.4, at room temperature. In calcium free solution,CaCl2 was omitted. All reagents were obtained fromSigma (Deisenhofen, Germany). Following dye loading,cells were incubated in the physiological solution for anadditional 15 min to ensure Fura-2/AM hydrolysis.Subsequently, coverslips were transferred to the stageof an inverted Olympus IX 51 microscope. Tetrakis-(2-pyridylmethyl)ethylenediamine (TPEN), a membranepermeable, selective heavy metal chelator, was appliedat the end of the experiments to confirm that the ob-served changes in Fura-2 fluorescence were related tochanges in intracellular zinc.

To trigger the flow of zinc through the L-type Ca2�

channel (LTCC), cells were perfused with Ringer’s so-

lution with 0.5 mM K� and then depolarized by perfu-sion with high K� solution (Ringer’s solution as givenabove, with 50 mM KCl replacing 50 mM NaCl andaddition of the LTCC agonist Bay-K 384, 200 nm). TheLTCC inhibitor nimodipine (1 �M, Sigma) was used todetermine whether the flux was mediated by LTCC.

A fluorescence ratio-imaging system (Till Photonics,Martinsried, Germany) was used for excitation andmonitoring the emitted fluorescence. Excitation wave-length was switched between 340 nm (F340) and 380nm (F380) by means of a monochromator (12-nm band-width). Using a 12-bit CCD camera, fluorescence sig-nals were recorded after passing a dichroic beam split-ter and an emission filter at 530 � 10 nm. Acquisition,storage, and analysis were performed with Axon In-struments Workbench 2 software.

Preparation of the ZnT-1 Antibody andImmunoblots

ZnT-1 antibody was prepared using a peptide de-rived from the C-terminal of ZnT-1 (as described inSekler et al., 2002). Immunoblots were run as de-scribed elsewhere (Sekler et al., 1995). Briefly, cellswere lysed for 15 min on ice in 20 mM Tris-HCl buffer(pH 7.3) containing 140 mM NaCl, 0.5% Triton X-100,and one tablet per 50 ml of complete protease inhibitors(Roche Molecular Biochemicals, Mannheim, Germany).Extracts were cleared by centrifugation and proteinwas determined by a protein assay (Pierce, Rockford,IL) with bovine serum albumin (BSA) as a standard.The antigen was visualized using the ZnT-1 antibodyat a dilution of 1:500–1:1,000. Detection was achievedutilizing HRP-conjugated donkey anti rabbit IgG (Am-ersham) diluted 1:6000 and the Super-Signal™ ULTRAECL detection reagent (Pierce). Relative levels ofZnT-1 were quantified from the ECL stained immuno-blots by densitometric analysis using the ChemiImagersystem (Alpha Innotech, San Leandro, CA). To deter-mine the specificity of the immunoblot reaction, anti-serum was preabsorbed with 0.4 mg/ml of immunizingpeptide overnight at 4°C as previously described (Kingand Agre, 1996; Sekler et al., 2002). The specificity ofour ZnT-1 antibody for this procedure was determinedin HEK-293 cells transfected with ZnT-1 plasmid orvector alone as control. As shown in Figure 5C, anintense staining of a 60-kDa protein, the expected MWof ZnT-1, was monitored in cells transfected with theZnT-1 plasmid. No staining was monitored in vector-alone transfected cells, indicating that the antibody ishighly specific.

ZnT-1 Immunocytochemistry

Glial cells grown on glass coverslips were fixed for 30min in 4% paraformaldehyde in 0.1 M phosphate-buff-ered saline (PBS) and then washed three times in PBS.Thereafter, the cells were incubated for 30 min in

147ZnT-1 INDUCTION IN GLIA

blocking solution containing 4% normal goat serum, 1%BSA and 0.1% Triton X-100 in PBS. The cells were thenincubated for 4 h at 4°C with the ZnT-1 antibody solu-tion, containing 1% goat serum, 4% albumin in PBS,diluted 1:500. The ZnT-1 antibody was visualized usingCy2 IgG-conjugated goat anti-rabbit (Jackson Labora-tories, Bar Harbor, ME). Secondary antibodies werediluted 1:250 in PBS and incubated for 1 h at roomtemperature. Astrocytes were identified using a mono-clonal mouse anti-GFAP antibody (Roche; 1:500), oli-godendrocytes were identified with O4 antibodies(Sommer and Schachner, 1981) and Cy3-coupled sec-ondary antibodies. Microglia cells were labeled withthe biotinylated lectin Griffonia simplicifolia (IB4-Bi-otin; Sigma, 1:100) and visualized by Streptavidin-Cy3.Coverslips were mounted in Moviol (Merck, Darm-stadt, Germany). Specificity of immunoreactivity wascontrolled by incubation of coverslips or tissue sectionsin buffer instead of primary antibodies or in ZnT-1antiserum, which was preabsorbed with the immuniz-ing peptide. In these control experiments, the immu-noreaction was always negative. Cells or sections wereinspected using fluorescence microscopy (Zeiss Ax-ioscop, Zeiss Oberkochen, Germany) equipped for epi-fluorescence illumination.

RESULTSAstrocytes Are Vulnerable to Zinc and Can Be

Protected by Priming

Toxic levels of zinc are released during pathologicalevents such as ischemia or seizures. To assess theeffect of extracellular zinc on cultured primary astro-cytes, we measured cell viability markers to character-ize these cultures under toxic zinc conditions. Videotime-lapse recordings were made to monitor the timecourse of zinc-induced cell death. Astrocytes growingon coverslips were exposed to 250 �M zinc, immedi-ately transferred to the video microscope and recordedover the following 8 h. After 1.5–2 h of exposure to zinc,the cell–cell contacts that initially characterized thesecultures began to disrupt, with the integrity of theastrocytic monolayer lost in many areas of the cover-slip. After 2–3 h, large portions of the monolayer be-came detached from the substrate and aggregated.Subsequently, most of the remaining astrocytes de-tached from the coverslip (Fig. 1A,B). In parallel exper-iments, measurements of the LDH activity from thesupernatants of the astroglial cultures were performedin order to determine the time of onset of cell death andthe lethal zinc dosage. The dose-response curves ob-

Fig. 1. Zinc vulnerability of primary astrocytes. A,B: Images fromvideo time-lapse recordings of astrocytic monolayers. At the beginningof the experiment, astrocytes are arranged in a monolayer (A). At2.5 h after exposure to zinc (250 �M), there are large cell-free areasand the remaining cells form clusters (B). C: Dose-response curve forlactate dehydrogenase (LDH) release from astrocytic monolayers in-duced by varying concentrations of extracellular zinc. Cultures were

exposed to zinc for 24 h. LDH release was normalized to the value ofLDH released after 2 h treatment with 0.1% Triton X-100, which wasdefined as 100% cell death. B. Time-dependent LDH-release fromastrocytes after exposure to 250 �M zinc. Each curve shown is onerepresentative experiment out of n � 10 in A, or n � 3 in B, respec-tively. Results are expressed as mean values � SE. Scale bar � 50 �min B (refers to A,B).

148 NOLTE ET AL.

tained typically showed a steep rise in LDH releaseabove 175 �M zinc (Fig. 1C). The LD50 for astrocytesexposed for 24 h was 194 �M � 8 (n � 10). Time-courseexperiments for LDH release revealed that in the pres-ence of 250 �M zinc all astrocytes had already diedafter 6 h (Fig. 1D). This corresponds to the phase whencells had detached from the substrate.

We then examined whether priming of astrocyteswith a sublethal zinc concentration increases their re-sistance to zinc toxicity. Astrocytes in culture weretreated with 50 �M zinc for 24 h and then exposed foranother 24 h to varying zinc concentrations as de-scribed above for the dose-response curves. As shown inFigure 2, primed astrocytes were less sensitive to toxiczinc levels. In this set of experiments, the LD50 ofprimed astrocytes was shifted by 45 �M � 7; (n � 4) tohigher zinc concentrations. Thus, priming astrogliawith sublethal levels of zinc leads to increased resis-tance to the effects of subsequent exposure to toxiclevels of zinc.

Intracellular Zinc Accumulation Mediated bythe L-Type Ca2� Channel Is Decreased in

Astrocytes Pretreated With Sublethal ZincConcentrations

The L-type Ca2� channels (LTCC) are a central andubiquitous route for zinc permeation in many cellstypes (Atar et al., 1995, Kim et al., 2000). Importantly,zinc influx through this pathway plays a prominentrole in subsequent zinc toxicity-related cell death (Kimet al., 2000, Sensi et al., 1997). We have thereforetested whether zinc priming will modulate zinc perme-ation through this pathway. Astrocytes loaded withFura-2 were perfused with Ca2� free Ringer containing

400 �M zinc followed by perfusion with depolarizingconcentration of K� (50 mM) and the LTCC agonistBay-K 384 (200 nM) to fully open the LTCC. As shownin Figure 3A, depolarization triggered a steep increasein the Fura–2 fluorescence. The fluorescent signalcould be reduced to almost baseline level upon additionof the zinc chelator TPEN indicating that the fluores-cence increase was indeed related to a rise in intracel-lular zinc concentration. Zinc pretreatment (24 h with60 �M zinc) lowered the intracellular zinc accumula-tion following the opening of this pathway. Figure 3Bshows the quantification of the intracellular zinc accu-mulation after opening the LTCC in three independentexperiments for each treatment, determined from n �65 cells per group. The fact that nimodipine inhibitedthe intracellular zinc accumulation designates LTCCas the major pathway for zinc accumulation. Priming ofthe cells with zinc had a dramatic effect; the intracel-lular zinc concentration after opening of the LTCC wasreduced by more than twofold, comparable to the effectof nimodipine. Our results indicate that zinc priminghas a specific effect on the rate of zinc accumulationfollowing opening of LTCC, in cultured astrocytes.

Immunocytochemistry for ZnT-1

By immunocytochemistry, we characterized the cel-lular expression and distribution of ZnT-1 in astro-cytes, but also in the other major glial cell types, mi-croglia and oligodendrocytes. To correlate ZnT-1expression with the major glial cell types, cultureswere counterstained with glia specific markers (Fig. 4):GFAP to identify astrocytes, isolectin-B4 for microglialcells (Streit, 1990), and O4-antibodies for oligodendro-cytes (Sommer and Schachner, 1981). ZnT-1 immuno-reactivity was detected in all of these glial cell typesand was confined both to the cytoplasmic membraneand to cytoplasmic structures (Fig. 4). In some experi-ments, a weak staining of nuclear structures was alsofound. Comparison of the intensity of ZnT-1 labeling inisolated and mixed glial cultures indicated that oligo-dendrocytes present more prominent ZnT-1 labelingcompared with astrocytes or microglia (Fig. 4G). Thespecificity of ZnT-1 immunoreactivity was determinedby preincubating the antibody with its immunizingpeptide. As shown in Figure 4H, this step reducedlabeling to background levels, indicating that the stain-ing was specific. Taken together, the above resultsshow that ZnT-1 is expressed in all major types ofneuroglia in culture.

Expression of the ZnT-1 Protein Is Induced byZinc Priming

The putative zinc transporter, ZnT-1, is suggested toplay a major role in conferring resistance against zinctoxicity in kidney and neuronal cell lines (Kim et al.,2000; Palmiter and Findley, 1995; Tsuda et al., 1997).

Fig. 2. Priming of astrocytes with sublethal concentration of zincconfers resistance against zinc toxicity. Dose-response curves for lac-tate dehydrogenase (LDH) release as determined for untreated astro-cytes and astrocytes primed for 24 h with 50 �M zinc. Cells wereexposed to the indicated concentrations for another 24 h. Values weredetermined in triplicate and are expressed as mean % (� SE). Theexample shown is one representative experiment out of n � 4.

149ZnT-1 INDUCTION IN GLIA

We have therefore asked whether zinc priming in-volves induction of ZnT-1 protein expression. For thispurpose we performed immunoblot analysis on extractsof cultured astrocytes primed with 20–50 �M zinc for24 h (Fig. 5A). As demonstrated by densitometry, incu-bation of astrocytes with sublethal concentrations ofzinc induced up to 4-fold increase in the expression ofZnT-1 (Fig. 5B). Altogether, our results indicate thatzinc priming leads to induction of ZnT-1 expressionthat is accompanied by reduction of zinc accumulation(Fig. 3), as well as reducing cellular zinc toxicity (Fig.2).

The reduction in intracellular zinc accumulation fol-lowing priming with zinc cannot be attributed only toZnT-1 since the expression of multiple proteins is af-fected by zinc in the culture medium. The specific roleof this protein was therefore addressed by heterologousexpression in astroglia culture. Typically 20–30% ofthe cells were transfected as identified by tracingEYFP cotransfected with ZnT-1 (not shown). Heterolo-gous expression of ZnT-1 had a very similar effect asthe zinc priming, namely a decrease in intracellularzinc concentration following the opening of the LTCC(Fig. 6A). The concentration of intracellular zinc inZnT-1 transfected cells (n � 45) was drastically re-duced (Fig. 6B), indicating that the expression of thelatter specifically reduces zinc accumulation in astro-glial cells.

To determine directly whether ZnT-1 expressionplays a role in protecting from toxic zinc insults, celldeath following exposure to lethal zinc concentrationswas compared between cells heterologously expressingZnT-1 and those transfected with control plasmid. Asshown in Figure 7, transfection with ZnT-1 resulted ina dramatic reduction of zinc-induced cell death.

DISCUSSIONZinc Is Cytotoxic for Astrocytes

A considerable body of work exists describing boththe influences of zinc on physiological properties ofneurons and its neurotoxic effects at high concentra-tions (Choi and Koh, 1998). Much less, however, isknown about the effects of zinc on glial cells. Cellularzinc toxicity in glia was previously characterized withrespect to dose and time dependence and was shown tobe influenced by serum components (Choi et al., 1988;Swanson and Sharp, 1992). Under the experimentalconditions used in our study, we determined that as-trocytic cell death began below 200 �M zinc and at�300 �M, there was 100% cell death. Interestingly,extracellular zinc concentrations in this range can bereached transiently at the mossy fiber region after re-petitive electrical stimulation and especially duringexcitotoxic activity (Assaf and Chung, 1984; Peters etal, 1987; Koh et al., 1996; Choi and Koh, 1998; Vogt etal., 2000). Excess of zinc is believed to contribute to celldeath also via different intracellular cascades (Sensi etal., 2003; Bossy-Wetzel et al., 2004). By reducing the

Fig. 3. Induction of ZnT-1 expression, by zinc, lowers cellular zincaccumulation following opening of L-type Ca channels. After loadingwith Fura–2, astrocytes were perfused with Ringer solution contain-ing 400 �M zinc and fluorescence was measured as described. A:Representative experiment showing intracellular zinc accumulationbefore and after depolarization using high K� and the L-type Ca2�

channel (LTCC) agonist Bay-K 384 (200 nM) in control cells and cellsthat were primed with 60 �M zinc for 18 h as described. B: Intracel-lular astrocytic zinc levels (normalized to control) after challenge with400 �M zinc in zinc-primed and control cells under normal conditionsor following depolarization with high K� and 200 nM Bay-K 384. Therate of change of the intracellular zinc level following depolarizationwas also determined in the presence of the LTCC inhibitor nimodipine(1 �M). Zinc priming attenuated intracellular zinc accumulation fol-lowing depolarization by more than twofold. n � 65 cells per group,using at least three independent measurements for each group.

150 NOLTE ET AL.

intracellular zinc concentration, ZnT-1 may also helpin preventing cell death by zinc released from intracel-lular stores.

An important finding of our study is that treatmentof astrocytes with sublethal zinc concentrations resultsin reduced intracellular level of zinc following opening

Fig. 4. ZnT-1 is expressed by glial cells in culture. Cultures en-riched in astrocytes (A,B,G), microglia (C,D), and oligodendrocytes(E,F) were immunolabeled with ZnT-1 antiserum (A,C,E,G) andcounterstained with glial specific markers: GFAP to label astrocytes(B), Isolectin B4 for labeling microglia (D), or O4 to label oligoden-drocytes and their precursors (F). G: Mixed glial culture after staining

with ZnT-1 antibodies. Note the prominent labeling of oligodendro-cytes /precursors (arrows) growing on top of the astrocytes. H: Nega-tive control: glial culture incubated with ZnT-1 antiserum after pre-absorption with its immunizing peptide. Scale bars � 50 �m in G; 20�m in H (refers to A–F,H).

151ZnT-1 INDUCTION IN GLIA

of LTCC. At the same time, pretreatment with suble-thal zinc concentrations conferred resistance of the as-trocytes to toxic zinc levels. Thus, astrocytes respond tosublethal exposure to zinc by activating a mechanismthat may protect the cells from the toxic effects of zincat higher concentrations. The effect can be specificallyattributed to ZnT-1 because heterologous expression ofZnT-1, with no zinc priming, yielded a similar lowering

Fig. 5. Priming astrocytes with sublethal concentration of zinc in-duces expression of ZnT-1 and confers resistance against toxic zincconcentrations. A: Astrocytes were primed for 24 h with the indicatedzinc concentrations added to the growth medium or with 10 �M EDTAto chelate the residual zinc. CTL, astrocytes grown in regular mediumB: Densitometric analysis of immunoblots showing the mean � SE ofn � 4 experiments. C: HEK-293 cells transfected with ZnT-1 plasmidor vector alone as control to determine the specificity of our ZnT-1antibody. A band of 60 kDa, the expected molecular weight of ZnT-1,was monitored only in homogenate of cells transfected with the ZnT-1plasmid.

Fig. 6. Heterologous expression of ZnT-1 in astrocytes reduces in-tracellular zinc accumulation via L-type Ca2� channel (LTCC). A:Representative experiment measuring accumulation of zinc in astro-cytes that were cotransfected with enhanced yellow fluorescent pro-tein (EYFP) and ZnT-1 plasmid (ZnT-1 transfected) or with EYFPplasmid and vector without ZnT-1 insert (mock-transfected) as de-scribed under experimental procedures. Zinc accumulation was mea-sured using the experimental paradigm described in Fig. 3. B: Com-parison between the rate of basal vs LTCC-mediated zincaccumulation in ZnT-1 or vector transfected cells. While the rate ofbasal zinc accumulation was unchanged, the zinc accumulation afteropening the LTCC was reduced by about threefold following theexpression of ZnT-1; n � 45 cells per group, using at least threeindependent measurements for each group.

152 NOLTE ET AL.

effect on the intracellular zinc level following the openingof LTCC. The general importance of the LTCC pathwayfor cellular zinc influx and in particular its role in medi-ating zinc related toxicity (Atar et al., 1995, Kim et al.,2000) points toward the importance of ZnT-1 expressionin reducing intracellular zinc accumulation and its sub-sequent toxicity. By transfecting primary astrocytes withZnT-1, a more direct link between expression of ZnT-1and resistance to toxic zinc levels has now been estab-lished. Our result is in agreement with previous results(Tsuda et al., 1997; Kim et al., 2000) and indicate that thereduction in cellular zinc accumulation mediated byZnT-1 plays a role in lowering zinc toxicity.

Is ZnT-1 in Glial Cells a Candidate forConferring Resistance Against Zinc Toxicity?

Heterologous overexpression of ZnT-1 confers resis-tances to toxic zinc in a number of cell lines (Tsuda et

al., 1997; Kim et al., 2000). However, much less isknown about the role of endogenously-expressed ZnT-1or the effects of its induction on zinc transport andresistance to zinc toxicity. Furthermore, expression ofthe ZnT-1 protein often does not correlate with changesin its mRNA levels (McMahon and Cousins, 1998;Tsuda et al., 1997), indicating that a posttranscrip-tional mechanism is involved in regulating the expres-sion of the ZnT-1 protein. For this reason, we focusedon the expression of the endogenous ZnT-1 protein inglial cells and its possible regulation. This approachenabled us to establish a correlation between increasedresistance to zinc stress and the expression of ZnT-1 inglial cells. That is, pretreating glial cells with sublethaldoses of zinc for 12–24 hours induced a robust upregu-lation of the protein in the cells and rendered themmore resistant to zinc toxicity. Of course, initial expo-sure to zinc will have numerous effects on a cell, someof which might conceivably lower the sensitivity of the

Fig. 7. Heterologous expression of ZnT-1 in astrocytes reduces sus-ceptibility to toxic zinc. Primary astrocytes were cotransfected withenhanced yellow fluorescent protein (EYFP) and ZnT-1 plasmid(ZnT-1 transfected) or with EYFP plasmid and vector without ZnT-1insert (pcDNA) as described under experimental procedures. After78 h, cells were treated with 400 �M zinc for 30 min (no zinc treat-

ment in the control) followed by propidium iodide labeling. Double-fluorescent astrocytes were quantified in 15 randomly selected fieldsfor each condition and normalized for the number of dead cells in themock-transfected cultures. Experiments were performed in triplicate.Pictures show representative examples of transfected astrocytes un-der the different treatments. Scale bar � 20 �m.

153ZnT-1 INDUCTION IN GLIA

cell to a subsequent zinc load. For example, the induc-tion of metallothionein expression by zinc and otherheavy metals is playing a key role in the protection ofglial cells against toxic zinc concentration. One obviouscaveat, however, of metallothionein activity is thatthey buffer, but do not remove, intracellular zinc. Re-cent studies have demonstrated that the interactionbetween metallothioneins and zinc is potentially labile,and dissociation of the metal can be triggered bychanges in the redox potential or by nitric oxide signal-ing (Pearce et al., 2000). Importantly, ischemic condi-tions may also enhance the dissociation of zinc frommetallothioneins (Frederickson and Bush, 2001). Se-questration of zinc into intracellular organelles mayalso lower intracellular zinc. However, the quantitativecontribution of each of these mechanisms to loweringthe intracellular toxic zinc load is not clear. The rela-tive role of either metallothioneins or ZnT-1 has to bedefined by future studies employing siRNA or othersilencing methods of either one of these proteins.

The mechanism by which ZnT-1 affects intracellularzinc levels is not fully elucidated. However, previousstudies which directly measure the flux of radioactivezinc suggest that ZnT-1 reduces zinc accumulation,possibly by enhancing zinc efflux (Palmiter and Find-ley, 1995; Palmiter et al., 1996). The measurement ofintracellular zinc concentration as employed in ourstudy does not allow the conclusion as to whetherZnT-1 reduces zinc levels in astrocytes by decreasingthe influx or by enhancing the efflux. Our results indi-cate that increasing the level of ZnT-1 expression willdecrease the intracellular level of zinc, which is thesum of zinc flux both in and out of the astrocyte. At thesame time, astrocytes expressing increased levels ofZnT-1 are protected against zinc toxicity. Such a rolefor zinc transporters in protecting against zinc toxicityupon inactivation of ZnT-1 has been shown for othercell types before (Palmiter and Findley, 1995; Tsuda etal., 1997). Considering the steep transmembrane zincgradient and the large capacity of the membrane zincpermeation pathway, an activity that will either atten-uate cellular zinc influx or actively extrude intracellu-lar zinc may be of particular importance in protectingagainst zinc toxicity.

In cases of mild stroke, a process known as “ischemictolerance” has been shown to protect against subse-quent strokes (Moncayo et al., 2000; Weih et al., 2001).Such a process may be reminiscent to the primingeffect of zinc in protecting cells, as proposed in thisstudy. Although a role for NMDA receptor activationand alterations in gene expression have also been sug-gested (Grabb and Choi, 1999), the mechanism of ‘isch-emic tolerance’ has not been fully elucidated. Since zincis released during these mild and transient ischemicepisodes, overexpression of ZnT-1 as well as metallo-thioneins may result (Aschner et al., 1998), therebyprotecting against a subsequent, massive zinc releasethat would otherwise have caused rapid neuronal orglial cell death. Indeed, the zinc inducible nature of theZnT-1 and metallothioneins promoters (Aschner, 1998;

Langmade et al., 2000) supports this conclusion andmay indicate that such a mechanism operates in neuralcells as well.

ACKNOWLEDGMENTS

The authors thank Brigitte Gerlach and ChristianeGras for technical assistance, and Gal K. Paz for per-forming preliminary experiments. This work was sup-ported by German Israel Foundation (GIF), projectI-588-99.1/1998. The work of A.H. was supported bythe German Research Foundation, grant HO 2205/1.

REFERENCES

Aschner M, Conklin DR, Yao CP, Allen JW, Tan KH. 1998. Inductionof astrocyte metallothioneins (MTs) by zinc confers resistanceagainst the acute cytotoxic effects of methylmercury on cell swell-ing, Na� uptake, and K� release. Brain Res 7:813:254–261.

Aschner M. 1998. Metallothionein (MT) isoforms in the central ner-vous system (CNS): regional and cell-specific distribution and po-tential functions as an antioxidant. Neurotoxicology 19:653–660.

Assaf SY, Chung SH. 1984. Release of endogenous Zn2� from braintissue during activity. Nature 308:734–736.

Atar D, Backx PH, Appel MM, Gao WD, Marban E. 1995. Excitation-transcription coupling mediated by zinc influx through voltage-dependent calcium channels. J Biol Chem 270:2473–2477.

Berg JM, Shi Y. 1996. The galvanization of biology: a growing appre-ciation for the roles of zinc. Science 271:1081–1085.

Bossy-Wetzel E, Talantova MV, Lee WD, Scholzke MN, Harrop A,Mathews E, Gotz T, Han J, Ellisman MH, Perkins GA, Lipton SA.Crosstalk between nitric oxide and zinc pathways to neuronal celldeath involving mitochondrial dysfunction and p38-activated K�

channels. Neuron 41:351–365.Choi DW, Koh JY. 1998. Zinc and brain injury. Annu Rev Neurosci

21:347–375.Choi DW, Yokoyama M, Koh J. 1988. Zinc neurotoxicity in cortical cell

culture. Neuroscience 24:67–79.Cole TB, Wenzel HJ, Kafer KE, Schwartzkroin PA, Palmiter RD.

1999. Elimination of zinc from synaptic vesicles in the intact mousebrain by disruption of the ZnT3 gene. Proc Natl Acad Sci USA96:1716–1721.

Crawford IL, Connor JD. 1972. Zinc in maturing rat brain: hippocam-pal concentration and localization. J Neurochem 19:1451–1458.

Dineley KE, Scanlon JM, Kress GJ, Stout AK, Reynolds IJ. 2000.Astrocytes are more resistant than neurons to the cytotoxic effectsof increased [Zn2�]i. Neurobiol Dis 7:310–320.

Frederickson CJ. 1989. Neurobiology of zinc and zinc-containing neu-rons. Int Rev Neurobiol 31:145–238.

Frederickson CJ, Bush AI. 2000. Synaptically released zinc: physio-logical functions and pathological effects. Biometals 14:353–366.

Frederickson CJ, Hernandez MD, Goik SA, Morton JD, McGinty JF.1988. Loss of zinc staining from hippocampal mossy fibers duringkainic acid induced seizures: a histofluorescence study. Brain Res446:383–386.

Grabb MC, Choi DW. 1999. Ischemic tolerance in murine cortical cellculture: critical role for NMDA receptors. J Neurosci 19:1657–1662.

Huang EP. 1997. Metal ions and synaptic transmission: think zinc.Proc Natl Acad Sci USA 94:13386–13387.

Kapur J, Macdonald RL. 1997. Rapid seizure-induced reduction ofbenzodiazepine and Zn2� sensitivity of hippocampal dentate gran-ule cell GABAA receptors. J Neurosci 17:7532–7540.

Kim AH, Sheline CT, Tian M, Higashi T, McMahon RJ, Cousins RJ,Choi DW. 2000. L-type Ca2� channel-mediated Zn2� toxicity andmodulation by ZnT-1 in PC12 cells. 1. Brain Res 886:99–107.

King LS, Agre P. 1996. Pathophysiology of the aquaporin water chan-nels. Annu Rev Physiol 58:619–648.

Koh JY, Suh SW, Gwag BJ, He YY, Hsu CY, Choi DW. 1996. The roleof zinc in selective neuronal death after transient global cerebralischemia. Science 272:1013–1016.

Langmade SJ, Ravindra R, Daniels PJ, Andrews GK. 2000. The tran-scription factor MTF-1 mediates metal regulation of the mouseZnT1 gene. J Biol Chem 275:34803–34809.

154 NOLTE ET AL.

Lee JY, Cole TB, Palmiter RD, Koh JY. Accumulation of zinc indegenerating hippocampal neurons of ZnT3-null mice after sei-zures: evidence against synaptic vesicle origin. J Neurosci.20:RC79.

McCarthy KD, de Vellis J. 1980. Preparation of separate astroglialand oligodendroglial cell cultures from rat cerebral tissue. J CellBiol 85:890–902.

McMahon RJ, Cousins RJ. 1998. Regulation of the zinc transporterZnT-1 by dietary zinc. Proc Natl Acad Sci USA 95:4841–4846.

Moncayo J, de Freitas GR, Bogousslavsky J, Altieri M, van Melle G.2000. Do transient ischemic attacks have a neuroprotective effect?Neurology 54:2089–2094.

Nolte C, Moller T, Walter T, Kettenmann H. 1996. Complement 5acontrols motility of murine microglial cells in vitro via activation ofan inhibitory G-protein and the rearrangement of the actin cy-toskeleton. Neuroscience 73:1091–107.

Palmiter RD, Findley SD. 1995. Cloning and functional characteriza-tion of a mammalian zinc transporter that confers resistance tozinc. EMBO J 14:639–649.

Palmiter RD, Cole TB, Findley SD. 1996. ZnT-2, a mammalian proteinthat confers resistance to zinc by facilitating vesicular sequestra-tion. EMBO J 15:1784–1791.

Pearce LL, Gandley RE, Han W, Wasserloos K, Stitt M, Kanai AJ,McLaughlin MK, Pitt BR, Levitan ES. 2000. Role of metallothioneinin nitric oxide signaling as revealed by a green fluorescent fusionprotein. Proc Natl Acad Sci USA 97:477–482.

Peters S, Koh J, Choi DW. 1987. Zinc selectively blocks the action ofN-methyl-D-aspartate on cortical neurons. Science 236:589–593.

Ransom B, Behar T, Nedergaard M. 2003. New roles for astrocytes(stars at least). Trends Neurosci 26:520–522.

Sekler I, Kopito R, Casey JR. 1995. High level expression, partialpurification, and functional reconstitution of the human AE1 anionexchanger in Saccharomyces cerevisiae. J Biol Chem 270:21028–21034.

Sekler I, Moran A, Hershfinkel M, Dori A, Margolis A, Birenzweig N,Nitzan Y, Silverman WF. 2002. Distribution of the zinc transporter

ZnT-1 in comparison to chelatable zinc in the mouse brain. J CompNeurol 447:201–209.

Sensi SL, Canzoniero LM, Yu SP, Ying HS, Koh JY, Kerchner GA,Choi DW. 1997. Measurement of intracellular free zinc in livingcortical neurons: routes of entry. J Neurosci 17:9554–9564.

Sensi SL, Ton-That D, Sullivan PG, Jonas EA, Gee KR, KaczmarekLK, Weiss JH. 2003. Modulation of mitochondrial function by en-dogenous Zn2� pools. Proc Natl Acad Sci USA 100:6157–6162.

Sommer I, Schachner M. 1981. Monoclonal antibodies (O1 to O4) tooligodendrocyte cell surfaces: an immunocytological study in thecentral nervous system. Dev Biol 83:311–327.

Streit WJ. 1990. An improved staining method for rat microglial cellsusing the lectin from Griffonia simplicifolia (GSA I-B4). J Histo-chem Cytochem 38:1683–1686.

Swanson RA, Sharp FR. 1992. Zinc toxicity and induction of the 72 kDheat shock protein in primary astrocyte culture. Glia 6:198–205.

Tonder N, Johansen FF, Frederickson CJ, Zimmer J, Diemer NH.1990. Possible role of zinc in the selective degeneration of dentatehilar neurons after cerebral ischemia in the adult rat. Neurosci Lett109:247–252.

Tsuda M, Imaizumi K, Katayama T, Kitagawa K, Wanaka A, To-hyama M, Takagi T. 1997. Expression of zinc transporter gene,ZnT-1, is induced after transient forebrain ischemia in the gerbil.J Neurosci 17:6678–6684.

Vallee BL, Falchuk KH. 1993. The biochemical basis of zinc physiol-ogy. Physiol Rev 73:79–118.

Vogt K, Mellor J, Tong G, Nicoll R. 2000. The actions of synapticallyreleased zinc at hippocampal mossy fiber synapses. Neuron 26:187–196.

Weih M, Prass K, Ruscher K, Trendelenburg G, Dirnagl U, Riepe MW,Meisel A. 2001. Ischemia tolerance; model for research, hope forclinical practice? Nervenarzt 72:255–260.

Weiss JH, Sensi SL. 2000. Ca2�-Zn2� permeable AMPA or kainatereceptors: possible key factors in selective neurodegeneration.Trends Neurosci 23:365–371.

Weiss JH, Hartley DM, Koh JY, Choi DW. 1993. AMPA receptoractivation potentiates zinc neurotoxicity. Neuron 10:43–49.

155ZnT-1 INDUCTION IN GLIA