GLIA6:198-205 (1992)
Zinc Toxicity and Induction of the 72 kD Heat Shock Protein in Primary Astrocyte Culture RAYMOND A. SWANSON AND FRANK R. SHARP Department of Neurology, University of California, Sun Francisco and VeteransAffairs Medical Center, San Francisco, California 94121
KEY WORDS
Glia, Heat stress protein, Heavy metal, Immunocytochemistry, Ischemia. Metallothionein
ABSTRACT Zinc is a potent inducer of the 72 kD heat shock protein (HSP72). In brain, pathological conditions such as ischemia and seizures increase extracellular zinc. The present study examines the effect of zinc on HSP72 expression in rat primary cortical astrocyte culture. Astrocytes were grown to confluence and exposed t o zinc chloride in C0,-equilibrated Earle’s buffered salt solution. Expression of HSP72 was examined using immunocytochemistry. IX3P72 was induced with zinc concentrations of 5 to 100 pM after 4 h exposures, or 200 to 300 pM after 15 min exposures. At the lower concentrations expression occurred in small clusters of contiguous cells. At concentrations high enough to cause cell death, HSP72-positive astrocytes formed a continuous margin around patches of dead cells. These patterns of HSP72 expression are similar to the patterns seen after cerebral ischemia in vivo. Exposure to zinc at 100 pM for 4 h or 400 pM for 15 min caused greater than 90% cell death. Increases in extracellular zinc may contribute to HSP72 induction and astrocyte death under ischemia and other pathological conditions in brain. o 1992 Wiley-Liss, Inc.
INTRODUCTION Zinc is present in millimolar amounts in brain and is particularly concentrated in cortex and regions of hippocampus (Chan et al., 1983; Frederickson, 1989). A small pool of zinc, comprising 5% to 15% of the total brain content (Frederickson, 1989), is concentrated in presynaptic vesicles (Danscher et al., 1985) and is released into the extracellular space during neuronal activity (Aniksztejn et al., 1987; Assaf and (:hung, 1984; Howell et al., 1984). Sub-millimolar concentrations of zinc have been found toxic to both neurons and glia in vitro (Choi et al., 1988), and several authors have suggested that endogenous zinc may play a role in neuronal and glial loss in some disease states (Choi et al., 1988; Frederickson et al., 1989). The 72 kD heat shock protein is one of a family of proteins expressed after exposure to heat, heavy metals (including zinc), and other stressors (Lindquist and Craig, 1988). These proteins are phylogenetically highly conserved and likely play a role in metabolism of denatured or newly synthesized proteins (Pelham, 0 1992 Wiley-Liss, Inc.
1986; Welch et al., 1989). Cells expressing HSP72 may have increased resistance to subsequent injury (Berger and Woodward, 1983; Chopp et al., 1989; Riabowol et al., 1988). In brain, HSP72 is expressed after ischemia Ferriero et al., 1990; Gonzalez et al., 1991; Sharp et al., 1991; Vass et al., 1988) and seizures (Lowenstein et al., 1990; Vass et al., 1989). HSP72 expression is of interest both as a marker of cellular injury and as a possible means of improving cerebral resistance to injury. Zinc is released into brain extracellular space during neuronal activity, with estimates of peak concentrations ranging from 30 pM (Aniksztejn et al., 1987) to 300 p M (Assaf and Chung, 1984). These levels are within the range reported to induce HSP72 in several cell types (Caltabiano et al., 1986; Misra et al., 1989; Nover, 1984; Whelan and Hightower, 1985),suggesting
Received July 2,1991; accepted February 24,1992 Address reprint requests to Dr. Raymond A. Swanson, Neurology Servicv (V1271,VAMC, 4150 Clement St., S a n Francisco, CA94121.
ZINC TOXICITY AND HSP72 INDUCTION
that release of endogenous zinc may contribute to the induction of heat shock proteins observed in brain after seizures or ischemia. As a separate consideration, zinc is an agent that could potentially be used to induce stress protein expression in brain and thereby facilitate study of possible protective effects of these proteins. The present study uses primary astrocyte cultures to examine the conditions under which zinc may induce HSP72 expression in brain.
199
faces was evaluated by preparing dilute (0-500 pM) solutions in EBSS and storing them in polypropylene tubes. These solutions were equilibrated to 5% C02 at 37°C prior t o use. Assessment of Cell Survival
Twenty-four hours after zinc exposure the cultures were inspected by phase contrast microscopy. Representative wells were incubated for 10 min with 1 pM of the DNA-binding fluorescent probe ethidium hoMATERIALS AND METHODS modimer (Molecular Probes, Eugene, OR) and photoAstrocyte Cultures graphed. Cell survival was quantified by measuring net lactate dehydrogenase (LDH) activity in the surviving Reagents were purchased from Sigma (St. Louis, cells in each well. The media was removed and cells MO), except as noted. Rat cortical astrocyte cultures lysed in a hypoosmolar solution comprised of 5 mM were prepared as described by Hertz et al. (1985) with HEPES pH 7.0/1 mM dithiothreitol/0.02% bovine seminor modifications. In brief, cortices were taken from rum albumin (BSA)/O.l%Triton X-100. LDH in the lynewborn rat pups (Simonsen, Gilroy, CAI, dissociated sate was measured as described by Koh and Choi with papain, and suspended in Eagle’s minimal essen(1987). For each experiment, the means and standard tial media (MEM) (Gibco, Grand Island, NY) suppleerrors of LDH activity in wells exposed to zinc were mented with 10% fetal bovine serum (FBS) (Hyclone, expressed as percents of the means of control wells in Ogden, UT) and 2 mM glutamine. The cells were plated the same 24-well plate. onto Falcon 24-well tissue culture plates (Becton-Dickinson, Oxnard, CA) at an approximate density of 2.5 x lo4 celldcm’ and incubated a t 37°C in a 5% CO, Immunocytochemistry atmosphere. The culture media was exchanged with HSP72 and GFAP were visualized using the nickelfresh media after 7 days. At confluence (d.i.v. 12-15), 10 pM cytosine arabinoside was added to the media to intensified avidin-biotin horseradish peroxidase techinhibit growth of other cell types. After 48 h this me- nique (Hoffman et al., 1990). Cultures were reacted for dium was replaced with MEM containing 2.5% FBS, 2 HSP72 using a monoclonal mouse antibody (Amersham mM glutamine, and 0.25 mM dibutyrl cyclic AMP (dB- RPN1197, Chicago, IL) developed and characterized by CAMP),except in some batches in which dBcAiMP was Welch and Suhan (1986) that reacts with one major omitted from a portion of the cultures. The cells re- protein and either a breakdown product or another mimained in this medium until used a t d.i.v. 20-30. Cells nor protein on Western blots (Vass et al., 1988). prepared in this way exhibit the morphology of differen- Twenty-four hours after zinc exposure the cultures tiated astrocytes (Hertz, 1990) and were uniformly pos- were fixed for 20 min in 0.1 M phosphate-buffered pH itive for the astrocyte marker glial fibrillary acidic pro- 7.4 saline (PBS) containing 4% paraformaldehyde. The fixative was removed and cultures incubated at room tein (GFAP)(Fig. 4A). temperature for 2 h with the anti-HSP72 antibody diluted 1:4,000 in a solution of 2% horse serum/0.2% Triton X-100/0.1% BSA/O.O2% sodium azide in PBS (HSZinc Exposure PBS). After washing in PBS the cells were reacted for 1 Experiments were begun by completely removing the h with biotinylated horse anti-mouse antibody (Elite culture media from the cells and washing twice with 0.5 Vectastain Kit, Vector Laboratories, Burlingame, CA) ml Earle’s balanced salt solution containing 5 mM glu- diluted 1:200 in HS-PBS. Cultures were then placed for cose (EBSS). The cultures were returned to the incuba- 1 h in the avidin-horseradish peroxidase solution pretor for 1 h t o allow pH and temperature equilibration. pared according to the manufacturer’s instructions Stock solutions of 20 mM zinc chloride were prepared (Vectastain). The cells were washed with 0.175 M soby dissolving anhydrous ZnC1, in 0.5 mM HC1. The dium acetate, pH 6.8, and exposed to a fresh solution of stock was diluted to 1-5 mM in water immediately be- nickeVdiaminobenzidine prepared by dissolving 10 mg fore use and added to the cultures in 5-25 pL aliquots. diaminobenzidine and 625 mg nickel sulfate hexahyAt this dilution the zinc solutions did not alter the cul- drate into 15 ml of acetate buffer. Hydrogen peroxide ture media pH. The cultures were returned to the incu- was added to a final concentration of 0.005%, and after bator until zinc exposure was terminated by washing incubation for 20 min the cells were washed in PBS. GFAP staining was performed similarly using a twice with culture media. The potential of zinc chelation by serum proteins or amino acids to decrease the mouse monoclonal antibody (ICN 69-110-2, Costa effective concentration of zinc was tested by adding Mesa, CA) diluted 1:4,000 in PBS-HS. Control wells for FBS to the EBSS or by incubating with zinc in MEM both GFAP and HSP72 wells were prepared by omitting rather than EBSS. Similarly, zinc adsorption onto sur- the first antibody during the staining procedure.
200
SWANSONAND SHARP
fY 75 3 m 0
w 0
50
150
100
200
250
25
0.1 0
300
100
Fig. 1. Effect of incubation media on zinc toxicity. Astrocytes were incubated for 4 h in EBSS (open circles) or MEM (triangles)to which zinc was added from stock, or with dilute zinc EBSS solutions that had been stored 24 h prior to use (filled circles). n = 8 at. each data point. Error bars denote SEM in all graphs.
-
-
--
300
400
500
600
/.LM ZNCI,
pM ZnCI,
1 2 5 O~ ~ ZINC M
200
- 60 pM
ZINC
.-- - -
T
1
f
125T
B
1
2 3
cn -1
w
0
w 0
20
40
60
80
100
p M ZNCI,
%
FBS IN MEDIA DURING ZINC INCUBATION
Fig. 3. Dose and time dependence of zinc toxicity. A. Astrocytes were exposed to zinc in EBSS for periods of 15 min fopen triangles), 30 min (filled circles), and 60 min (open squares). B: Astrocytes were exposed for 2 h (filled triangles), 4 h fopen circles), and 8 h (filled squares). n = 8 at each data point.
Fig. 2. Effect of serum on zinc toxicity. Astrocytes incubated 4 h in 40 pM zinc in EBSS with added fetal bovine serum fFBS). n = 6 at each data point. *, different from control at P < 0.0 1 by Student's t test.
RESULTS Zinc toxicity was found to be highly dependent upon the experimental conditions. In the absence of amino acids or serum proteins the LD,, for zinc was approximately 50 pM after a 4 h exposure, with a steep doseresponse curve (Fig. 1). The presence of amino acids (MEM) caused a 50%reduction in the apparent LD,, of zinc, and the use of media in which zinc had been stored as a dilute solution produced a tenfold reduction in potency (Fig. 1).Serum also caused a n apparent reduction in zinc potency, evident with as little as 0.1% FBS in the media (Fig. 2). In the absence of zinc, incubation in EBSS or serum-free MEM for up to 24 h did not cause detectable cell death. The zinc toxicity was time- and dose-dependent (Fig. 3). After 8 h exposures in EBSS the LD,, for zinc was approximately 20 pM. At 15 and 30 min exposures the LD,, was increased to about 250 pM. Steep dose-response curves were seen a t all exposure times. Astrocytes cultured without dBcAMP-induced differentia-
tion were not appreciably different in their sensitivity to zinc (data not shown). The LDH assay determinations of cell death were in every case in good agreement with the impressions gained by phase microscopy inspection, in which nonviable cells were either rounded up or completely detached from the culture dishes (Fig. 4B). However, a t low zinc concentrations, many wells that showed no evidence of cell death by inspection or LDH content contained scattered cells or small patches of cells that were shown to be nonviable by ethidium homodimer fluorescence (Fig. 4C,D). Heating the astrocyte cultures to 44°C for 40 min induced widespread staining (Fig. 5A) that was entirely absent when the HSP72 antibody was omitted from the staining procedure (Fig. 5B). Zinc induction of HSP72 was investigated with dBcAMP-treated cultures, using EBSS without added amino acids or serum, and zinc solutions freshly diluted from stock. No cell death was seen after 4 h exposures to 5-10 pM zinc, but this concentration did induce HSP72 expression in small c h s ters of cells (Fig. 5C,D). Within the clusters the staining of cell nuclei was usually more intense than cytoplasmic staining, a s was seen after heat exposure (Fig. 5A),
201
ZINC TOXICITY AND HSP72 INDUCTION
Fig. 4. Phase contrast photomicrographs of control and zinc-exposed astrocytes. A Control well fixed and stained for GFAF'. B: Unfixed astrocytes 24 h after a 4 h exposure to 100 pM zinc, showing most cells rounded up or detached. C: Unfixed astrocytes 24 h after a 4 h exposure to 20 pM zinc. D: Same field a s C after addition of ethidium homodimer showing fluorescence in nuclei of scattered dead cells under blue light excitation. All photos at x200.
and many cells on the periphery of clusters exhibited only nuclear staining. Fifteen minute exposures to zinc induced HSP72 expression only a t zinc concentrations that also caused death or detachment of most cells (Fig. 5E). However, 4 h exposures to zinc at 20-60 p M killed only a portion of astrocytes in each culture well and produced a distinctive pattern of HSP72 expression (Fig. 6). Cell death occurred in patches of cells, typically two to four per well, distributed in an apparently random fashion (Fig. 6D). With exposures at higher concentrations o r for longer periods the patches became larger and connected to one another. The patches were invariably surrounded by a contiguous band of HSP72-positive astrocytes that separated the areas of dead or detached cells from the viable, HSP72-negative cells. HSP72 immunoreactivity was never seen in cells identified as nonviable by ethidium homodimer fluorescence (Fig. 6A,B).
To exclude the possibility that zinc precipitates were causing the patchy distribution of HSP72 induction and cell death, zinc chloride was dissolved in EBSS and the solutions were passed through a 0.22 p,m filter or centrifuged at 20,OOOg for 20 min prior to placement on astrocytes. These treatments did not alter the patterns of zinc-induced HSP72 expression and cell death. DISCUSSION In agreement with the report by Choi et al. (19881, zinc was found to be toxic to astrocytes in a dose- and time-dependent manner. This toxicity increased two to threefold in media free of amino acids or serum. As noted by Whelan and Hightower (1985), zinc can complex to amino acids, especially cysteine, histidine, and glutamine, such that the presence of these amino acids
202
SWANSON AND SHARP
Fig. 5 . HSP72 expression in astrocytes. A: 24 h after 40 min at 44°C. B: Same as A but with anti-HSP72 antibody omitted from staining procedure. C and D: In clusters of cells after 4 h exposure to 5pM zinc. E: In surviving cells after 15 min incubation with 400 pM zinc. Photos A-C at ~ 2 0 0D , and E at x 100.
even at low levels may markedly decrease the concentration of free zinc. We also observed a great decrease in apparent toxicity when zinc solutions were stored a t concentrations below 500 pM. It is likely that this effect was due to adsorption of zinc onto vessel surfaces, as storage a t 20 mM in acidic solution eliminated this effect. An acidic solution was used to increase the solubility of zinc and reduce surface adsorption,, as well as to prevent formation of zinc oxychloride.
Zinc has many effects on metabolism (Frederickson, 1989) and could injure cells in several ways. Like other heavy metals, zinc can react with thiol and imidazole moieties on enzymes and other proteins (Chvapil et al., 1972). Zinc also binds to S-100 (Baudier and Gerard, 1983) and tubulin (Eagle et al., 19831, and can disrupt microtubule formation (Fujii et al., 1986; Frederickson, 1989). In addition, zinc can act a t many calcium binding sites including calmodulin (Brewer, 1980)and may dis-
ZINC TOXICITY AND HSP72 INDUCTION
203
Fig. 6. HSP72 expression in astrocytes. A After exposure to 40 KM zinc for 4 h, x200. B: Same field as A, with ethidium homodimer fluorescence confirming cell death within a rim of viable, HSP72-positive cells. C, D: Low magnification views of HSP72-positive astrocytes surrounding patches of dead cells; C, ~ 4 0D, ; ~10.
rupt intracellular calcium homeostasis at micromolar concentrations (Chao et al., 1984). As zinc may gain entry into cells in part through voltage-sensitive cation channels (Choi and Koh, 1988; Weiss et al., 1989,1991), cells such as neurons and glia that possess such channels (MacVicar, 1984) may be particularly sensitive to zinc, Heat shock protein expression in cultured astrocytes was first demonstrated by Nishimura et al. (19881, by exposing cultures to 45°C. The heat shock response of astrocytes has since been further characterized (Dwyer et al., 1991; Marini et al., 19901, and acidosis has been identified as another inducing agent (Nishimura et al., 1989).Although zinc and other heavy metals are potent inducers in other cell types (Lindquist and Craig, 1988), they have not previously been noted to cause a heat shock response in astrocytes. In the present study, 15 min exposures induced HSP72 expression only at high concentrations of zinc that killed most of the astrocytes. However, with 4 h exposures, HSP72 expression was induced over a range of zinc concentrations, beginning well below and extending into the range that caused
cell death. At the lower zinc concentrations HSP72 expression occurred in clusters of cells, rather than in scattered individual cells. At higher Concentrations contiguous stained cells formed a rim around areas of cell death. These patterns may reflect heterogeneity of the astrocyte population (Hansson, 1988; Wilkin et al., 1990), with clustered cells having descended from common progenitor cells. Alternatively, the increased size of patches seen after increased zinc concentrations or exposure times suggests that a factor causing cell death and HSP72 expression may be passed or amplified through contiguous cells. The gap junctions of the glial syncytium provide one avenue by which this could occur (Kettenmann et al., 1983; Sontheimer et al., 1990). Although a number of insults can cause HSP72 expression in brain in vivo, several groups have noted that the protein is not expressed in cells with morphologic evidence of death or degeneration (Ferriero et al., 1990; Gonzalez et al., 1991; Lowenstein et al., 1990; Sharp et al., 1991; Simon et al., 1991; Vass et al., 1988, 1989).The inference that HSP72 expression occurs only in viable cells is supported here by absence of ethidium
SWANSONAND SHARP
204
homodimer fluorescence in HSP72 positive astrocytes (Fig. 6). However, as reported by Nowak (19911, cells that are irreversibly injured may fail to express the HSP72 protein but succeed in producing HSP72 mRNA. I t is of interest that the rim-like pattern of HSP72 expression seen here around patches of‘ dead cells is similar to the pattern of expression seen in glia after focal ischemia in vivo (see Sharp et al., 1991). Ischemia causes neuronal depolarization (Mitani et al., 1989) and could thus cause release of zinc into the extracellular space. In “penumbral” regions of reduced but nonzero blood flow, the rate of glucose utilization is increased for several hours after vessel occlusion, suggesting a sustained period of increased neuronal activity (Shiraishi et al., 1989). Brain extracellular zinc levels have been estimated to reach 30 p M (Aniksztejn et al., 1987) to 300 p M (Assaf and Chung, 1984)during intensive neuronal activity, levels well above those seen here to induce HSP72 and astrocyte death in the present study in vitro. However, extrapolation of the present observations to in vivo conditions must be done with caution. To our knowledge, direct measurements of zinc in brain extracellular fluid have not yet been accomplished. Moreover, as we have shown decreased toxicity in the presence of amino acids and proteins, zinc toxicity in vivo may likewise be decreased by binding to amino acids and proteins in extracellular fluid. On the other hand, zinc toxicity in vivo may be augmented by conditions such as low pH or membrane depolarization, which could facilitate the entry of zinc into cells (Weiss et al., 1991). Lastly, reuptake mechanisms (Howell e t al., 1984) may drastically alter the extent and duration of zinc elevations in normal and ischemic brain. These issues are currently under study. Part of the rationale for characterizing zinc induction of HSP72 in astrocytes is that a chemical agent capable of inducing heat shock proteins in brain could be useful in the study of the biological functions of these proteins. Unfortunately, our results suggest that zinc holds little promise in this regard because of the steep dose-response curve of zinc cytotoxicity and the narrow range between doses needed to cause widespread HSP72 expression and those resulting in cell death. Zinc and other heavy metals are potent inducers of metallothionein, and zinc is now commonly used to activate a metallothionein promoter in recombinant genes (Durnam and Palmitter, 1981; Hurko et al., 1986). The present study suggests that in glial cells, and perhaps in other cell types as well, this maneuver may be complicated by the simultaneous induction of HSP72 and other stress proteins,
ACKNOWLEDGMENTS We would like to thank Marcia Stubblebine for technical assistance and Dr. John Weiss for helpful discussions. R.A.S. is supported by a Research Associate career development -award from the Department of
Veterans Affairs. This work is also supported by NIH grants R01-NS28167 and NS14543 (F.R.S.), and by the Merit Review program of the VA Research Service (R.A.S.andF.R.S.).
REFERENCES Aniksztejn, L., Charton, G., and Ben-Ari, Y. (1987) Selective release of endogenous zinc from the hippocampal mossy fibers in situ. Brain Res., 40458-64. Assaf, S.Y. and Chung, S-H. (1984) Release of endogenous Zn ’ ’ from brain tissue duringactivity.Nature, 308:734-736. Baudier, J . and Gerard, G. (1983)Ions binding to SIOO proteins: Structural changes induced by calcium and zinc on SlOOa and SlOOb proteins. Biochemistry, 22:3360-3369. Berger, E.M. and Woodward, M.P. (1983) Small heat shock proteins in Drosophila may confer thermal tolerance. Exp. Cell Res., 147:437442. Brewer, G. (1980) Calmodulin, zinc and calcium in cellular and membrane regulation: An interpretive review. Am. J. Hematol., 8:231248. Caltabiano, M.M., Koestler, T.P., Poste, G., and Grieg, R.G. (1986) Induction of 32- and 34-kDa stress proteins by sodium arsenite, heavy metals, and thiol-reactive agents. J . Biol. Chem., 261:1338113386. Chan, A.W.K., Minski, M.J., and Lai, J.C.K. (1983) An application of neutron activation analysis to small biological samples: Simultaneous determination of thirty elements in rat brain regions. J. Neurosci. Methods, 7:317-328. Chao, S.-H., Suzuki, Y., Zysk, J.R., and Cheung, W.Y. (1984) Activation of calmodulin by various metal cations as a function of ionic radius. Mol. Pharmacol., 26:75-82. Choi, D.W. and Koh, J. (1988) Zinc central neurotoxicity may require open NMDA channels. Sac. Neurosci. Abstr., 14:417. Choi, D.W., Yokoyama, M., and Koh, J. (1988) Zinc neurotoxicity in cortical cell culture. Neuroscience, 24:67-79. Chopp, M., Chen, H., Ho, K.-L., Dereski, M.O., Brown, E., Hetzel, F.W., and Welch, K.M.A. (1989) Transient hyperthermia protects against subsequent forebrain ischemic damage in the rat. Neurology, 39:1396-1398. Chvapil, M., Elias, S.L., Ryan, J.N., and Zukoski, C.F. (1972) Pathophysiology of zinc. In: Neurobiology of the Trace Metals Zinc and Copper. C. C. Pjeiffer, ed. Academic Press, New York, pp. 105-124. Danscher, G., Howell, G., Perez-Clausel, J., and Hertel, N. (1985) The dithizone, Timm’s sulphide silver and the selenium methods demonstrate a chelatable pool of zinc in CNS. Histachemistry, 83:419-422. Durnam, D. and Palmiter, R.D. (1981) Transcriptional regulation of the mouse metallothionein-I gene by heavy metals. J . Biol. Chem., 256~5712-5716. Dwyer, B.E., Nishimura, R.N., de Vellis, J., and Clegg, K.B. (1991) Regulation of heat shock protein synthesis in rat astrocytes. J . Neurosci. Res., 28:352-358. Eagle, G.R., Zombola, R.R., and Himes, R.H. (1983)Tubulin-zinc interactions: Binding and polymerization studies. Biochemistry, 22221228. Ferriero, D.M., Soberano, H.Q., Simon, R.P., and Sharp, F.R. (1990) Hypoxia-ischemia induces heat shock protein-like (HSP72) immunoreactivity in neonatal rat brain. Deu. Brain Res., 53:145-150. Frederickson, C.J. (1989) Neurobiology of zinc and zinc-containing neurons. Int. Rev. Neurobiol., 31:145-238. Frederickson, C.J., Hernandez, M.D., and McGinty, J.F. (1989) Translocation of zinc may contribute to seizure-induced death of neurons. Brain Res., 480:317-321. Fujii, T., Gochou, N., Akabane, Y., Fujii, M., Kondo, Y., Suzuki, T., and ’ Ohki, K. (1986) Effect of zinc ions on the interaction of S-100 protein with brain microtubule proteins. Chem. Pharm. Bull. (Tokyo), 345040-5044. Gonzalez, M.F., Lowenstein, D.L., Fernyak, S., Hisanaga, K., Simon, ‘R.P., and Sharp, F.R. (1991) Induction of heat shock protein72-like immunoreactivity in the hippocampal formation following transient global ischemia. Brain Res. Bull., 26:241-250. Hansson, E. (1988) Astroglia from defined brain regions as studied with primary cultures. Prog. Neurobiol., 30:369-397. Hertz, L. (1990) Dibutyryl cyclic AMP treatment of astrocytes in primary cultures as a substitute for normal morphogenic and ‘functiogenic’transmitter sienals. In: Molecular AsDects ofDeuelowmentand k i n g of the Nervoussystem. J.M. Lauder,*A. Private, E.’Giacobini, and A. Vernadakis, eds. Plenum Press, New York, pp. 227-243.
ZINC TOXICITY AND HSP72 INDUCTION Hertz, L., Juurlink, B.H.J., and Szuchet, S. (1985) Cell cultures, In: Handbook OfNeurochemistry, Vol. 8,2nd ed. A. Lajtha, ed. Plenum Press, New York, pp. 603461. Hoffman, G.E., Lee, W . 3 , Attardi, B., Yann, V., and Fitzsimmons, M.D. (1990) Luteinizing hormone-releasing hormone neurons express c-fos antigen after steroid activation. Endocrinology, 126:173&1741. Howell, G.A., Welch, M.G., and Frederickson, C.J. (1984) Stimulationinduced uptake and release of zinc in hippocampal slices. Nature, 308:736-738. Hurko, O., McKee, L., and Zuurveld, J.G.E.M. (1986) Transfection of human skeletal muscle cells with SV40 large T antigen gene coupled to a metallothionein promoter. Ann. Neurol., 20573-582. Kettenmann, H., Orkand, R.K., and Schachner, M. (1983) Coupling among identified cells in mammalian nervous system cultures. J . Neurosci., 3506-516. Koh, J-Y. and Choi, D.W. (1987) Quantitative determination of glutamate mediated cortical neuronal injury in cell culture by lactate dehydrogenase efflux assay. J . Neurosci. Methods, 20:83-90. Lindquist, S. and Craig, E.A. (1988) The heat-shock proteins. Annu. Rev. Genet., 22:631477. Lowenstein, D.H., Simon, R.P., and Sharp, F.R. (1990) The pattern of 72-kDa heat shock protein-like immunoreactivity in the rat brain following flurothyl-induced status epilepticus. Brain Res., 531:173182. MacVicar, B.A. (1984) Voltage-dependent calcium channels in glial cells. Science, 226:1345-1347. Marini, A.M., Kozuka, M., Lipsky, R.H., and Nowak, T.S., J r . (1990) 70-Kilodalton heat shock protein induction in cerebellar astrocytes and cerebellar granule cells in vitro: Comparison with immunocytochemical localization after hyperthermia in vivo. J . Neurochem., 543509-1516. Misra, S., Zafarullah, M., Price-Haughey, J., and Gedamu, L. (1989) Analysis of stress-induced gene expression in fish cell lines exposed to heavy metals and heat shock. Biochim. Biophys. Acta, 1007:325333. Mitani, A,, Imon, I., and Kataoka, K. (1989) High frequency discharges of gerbil hippocampal CA1 neurons shortly after ischemia. Brain Res. Bull., 23569-572. Nishimura, R.N., Dwyer, B.E., Welch, W.J., Cole, R., deVellis, J., and Liotta, K. (1988) The induction of the major heat-stress protein in purified rat glial cells. J . Neurosci. Res., 20:12-18. Nishimura, R.N., Dwyer, B.E., Cole, R., de Vellis, J., and Picard, K. (1989) Induction of the major inducible 68-kDa heat-shock protein after rapid changes of extracellular pH in cultured rat astrocytes. Exp. Cell Res., 180:276-280. Nover, L (1984) Inducers of hsp synthesis. In: Heat Shock Response of
205
Eukaryotzc Cells. L. Nover, ed. Springer-Verlag, New York, pp. 1-11. Nowak, T.S., J r . (1991) Localization of 70 kDa stress protein mRNA induction in gerbil brain after ischemia. J . Cereb. Blood Flow Metab., 11:432-439. Pelham, H.R.B. (1986) Speculations of the functions of the major heat shock and glucose-regulated proteins. Cell, 46:959-961. Qiabowol, K.T., Mizzen, L.A., and Welch, W.J. (1988) Heat shock is lethal to fibroblasts injected with antibodies against hsp7O. Science, 242:433-436. Sharp, F.R., Lowenstein, D.L., Simon, R.P., and Hisanaga, K. (1991) Heat shock protein HSP72 induction in cortical and striatal astrocytes and neurons following infarction. J . Cereb.Blood Flow Metab., 11:621M28. Shiraishi, K., Sharp, F.R., and Simon, R.P. (1989) Sequential metabolic changes in rat brain following middle cerebral artery occlusion: A 2-deoxyglucose study. J . Cereb. Blood Flow Metab., 9:765-773. Simon, R.P., Cho, H., Gwinn, R., and Lowenstein, D.H. (1991) The temporal profile of 72-kDa heat-shock protein expression following global ischemia. J . Neurosci., 11:881-889. Sqntheimer, H., Minturn, J.E., Black, J.A., Waxman, S.G., and Ransom, B.R. (1990) Specificity of cell-cell coupling in rat optic nerve astrocytes in vitro. Proc. Natl. Acad. Sci. USA, 87:9833-9837. Vass, K., Welch, W.J., and Nowak, T.S., Jr. (1988) Localization of 70-kDa stress protein induction in gerbil brain after ischemia. Acta Neuropathol., 77:12%135. Vass, K., Berger, M.L., Nowak, T.S., Jr., Welch, W.J., and Lassmann, H. (1989) Induction of stress protein HSP7O in nerve cells after status epilepticus in the rat. Neurosci. Lett., 100:259-264. Wiss, J.H., Koh, J-Y., Christine, C.W., and Choi, D.W. (1989)Zinc and LTP. Nature, 338:212. Weiss, J.H., Hartley, D.M., and Choi, D.W. (1991)Zinc potentiation of AMPA neurotoxicity: Possible mediation by Zn2+entry and upregulption of Ca2' influx. Neurology, 41(Suppl.):377. Welch, W.J. and Suhan, J.P. (1986) Cellular and biochemical events in mammalian cells during and after recovery from physiological stress. J . Cell. Biol., 103:2035-2052. Welch, W.J., Mizzen, L.A., and Arrigo, A.-P. (1989) Structure and function of mammalian stress proteins. In: Stress-Induced Proteins. M.L. Pardue, J.R. Feramisco, and S. Lindquist, eds. Alan R. Liss, New York, pp. 187-202. Whelan, S.A. and Hightower, L.E. (1985)Induction of stress proteins in chicken embryo cells by low-level zinc contamination in amino acid-free media. J . Cell. Physiol., 122:205-209. Wilkin, G.P., Marriott, D.R, and Cholewinski, A.J. (1990) Astrocyte heterogeneity. T I " , 13:4346.
Zinc Toxicity and Induction of the 72 kD Heat Shock Protein in Primary Astrocyte Culture RAYMOND A. SWANSON AND FRANK R. SHARP Department of Neurology, University of California, Sun Francisco and VeteransAffairs Medical Center, San Francisco, California 94121
KEY WORDS
Glia, Heat stress protein, Heavy metal, Immunocytochemistry, Ischemia. Metallothionein
ABSTRACT Zinc is a potent inducer of the 72 kD heat shock protein (HSP72). In brain, pathological conditions such as ischemia and seizures increase extracellular zinc. The present study examines the effect of zinc on HSP72 expression in rat primary cortical astrocyte culture. Astrocytes were grown to confluence and exposed t o zinc chloride in C0,-equilibrated Earle’s buffered salt solution. Expression of HSP72 was examined using immunocytochemistry. IX3P72 was induced with zinc concentrations of 5 to 100 pM after 4 h exposures, or 200 to 300 pM after 15 min exposures. At the lower concentrations expression occurred in small clusters of contiguous cells. At concentrations high enough to cause cell death, HSP72-positive astrocytes formed a continuous margin around patches of dead cells. These patterns of HSP72 expression are similar to the patterns seen after cerebral ischemia in vivo. Exposure to zinc at 100 pM for 4 h or 400 pM for 15 min caused greater than 90% cell death. Increases in extracellular zinc may contribute to HSP72 induction and astrocyte death under ischemia and other pathological conditions in brain. o 1992 Wiley-Liss, Inc.
INTRODUCTION Zinc is present in millimolar amounts in brain and is particularly concentrated in cortex and regions of hippocampus (Chan et al., 1983; Frederickson, 1989). A small pool of zinc, comprising 5% to 15% of the total brain content (Frederickson, 1989), is concentrated in presynaptic vesicles (Danscher et al., 1985) and is released into the extracellular space during neuronal activity (Aniksztejn et al., 1987; Assaf and (:hung, 1984; Howell et al., 1984). Sub-millimolar concentrations of zinc have been found toxic to both neurons and glia in vitro (Choi et al., 1988), and several authors have suggested that endogenous zinc may play a role in neuronal and glial loss in some disease states (Choi et al., 1988; Frederickson et al., 1989). The 72 kD heat shock protein is one of a family of proteins expressed after exposure to heat, heavy metals (including zinc), and other stressors (Lindquist and Craig, 1988). These proteins are phylogenetically highly conserved and likely play a role in metabolism of denatured or newly synthesized proteins (Pelham, 0 1992 Wiley-Liss, Inc.
1986; Welch et al., 1989). Cells expressing HSP72 may have increased resistance to subsequent injury (Berger and Woodward, 1983; Chopp et al., 1989; Riabowol et al., 1988). In brain, HSP72 is expressed after ischemia Ferriero et al., 1990; Gonzalez et al., 1991; Sharp et al., 1991; Vass et al., 1988) and seizures (Lowenstein et al., 1990; Vass et al., 1989). HSP72 expression is of interest both as a marker of cellular injury and as a possible means of improving cerebral resistance to injury. Zinc is released into brain extracellular space during neuronal activity, with estimates of peak concentrations ranging from 30 pM (Aniksztejn et al., 1987) to 300 p M (Assaf and Chung, 1984). These levels are within the range reported to induce HSP72 in several cell types (Caltabiano et al., 1986; Misra et al., 1989; Nover, 1984; Whelan and Hightower, 1985),suggesting
Received July 2,1991; accepted February 24,1992 Address reprint requests to Dr. Raymond A. Swanson, Neurology Servicv (V1271,VAMC, 4150 Clement St., S a n Francisco, CA94121.
ZINC TOXICITY AND HSP72 INDUCTION
that release of endogenous zinc may contribute to the induction of heat shock proteins observed in brain after seizures or ischemia. As a separate consideration, zinc is an agent that could potentially be used to induce stress protein expression in brain and thereby facilitate study of possible protective effects of these proteins. The present study uses primary astrocyte cultures to examine the conditions under which zinc may induce HSP72 expression in brain.
199
faces was evaluated by preparing dilute (0-500 pM) solutions in EBSS and storing them in polypropylene tubes. These solutions were equilibrated to 5% C02 at 37°C prior t o use. Assessment of Cell Survival
Twenty-four hours after zinc exposure the cultures were inspected by phase contrast microscopy. Representative wells were incubated for 10 min with 1 pM of the DNA-binding fluorescent probe ethidium hoMATERIALS AND METHODS modimer (Molecular Probes, Eugene, OR) and photoAstrocyte Cultures graphed. Cell survival was quantified by measuring net lactate dehydrogenase (LDH) activity in the surviving Reagents were purchased from Sigma (St. Louis, cells in each well. The media was removed and cells MO), except as noted. Rat cortical astrocyte cultures lysed in a hypoosmolar solution comprised of 5 mM were prepared as described by Hertz et al. (1985) with HEPES pH 7.0/1 mM dithiothreitol/0.02% bovine seminor modifications. In brief, cortices were taken from rum albumin (BSA)/O.l%Triton X-100. LDH in the lynewborn rat pups (Simonsen, Gilroy, CAI, dissociated sate was measured as described by Koh and Choi with papain, and suspended in Eagle’s minimal essen(1987). For each experiment, the means and standard tial media (MEM) (Gibco, Grand Island, NY) suppleerrors of LDH activity in wells exposed to zinc were mented with 10% fetal bovine serum (FBS) (Hyclone, expressed as percents of the means of control wells in Ogden, UT) and 2 mM glutamine. The cells were plated the same 24-well plate. onto Falcon 24-well tissue culture plates (Becton-Dickinson, Oxnard, CA) at an approximate density of 2.5 x lo4 celldcm’ and incubated a t 37°C in a 5% CO, Immunocytochemistry atmosphere. The culture media was exchanged with HSP72 and GFAP were visualized using the nickelfresh media after 7 days. At confluence (d.i.v. 12-15), 10 pM cytosine arabinoside was added to the media to intensified avidin-biotin horseradish peroxidase techinhibit growth of other cell types. After 48 h this me- nique (Hoffman et al., 1990). Cultures were reacted for dium was replaced with MEM containing 2.5% FBS, 2 HSP72 using a monoclonal mouse antibody (Amersham mM glutamine, and 0.25 mM dibutyrl cyclic AMP (dB- RPN1197, Chicago, IL) developed and characterized by CAMP),except in some batches in which dBcAiMP was Welch and Suhan (1986) that reacts with one major omitted from a portion of the cultures. The cells re- protein and either a breakdown product or another mimained in this medium until used a t d.i.v. 20-30. Cells nor protein on Western blots (Vass et al., 1988). prepared in this way exhibit the morphology of differen- Twenty-four hours after zinc exposure the cultures tiated astrocytes (Hertz, 1990) and were uniformly pos- were fixed for 20 min in 0.1 M phosphate-buffered pH itive for the astrocyte marker glial fibrillary acidic pro- 7.4 saline (PBS) containing 4% paraformaldehyde. The fixative was removed and cultures incubated at room tein (GFAP)(Fig. 4A). temperature for 2 h with the anti-HSP72 antibody diluted 1:4,000 in a solution of 2% horse serum/0.2% Triton X-100/0.1% BSA/O.O2% sodium azide in PBS (HSZinc Exposure PBS). After washing in PBS the cells were reacted for 1 Experiments were begun by completely removing the h with biotinylated horse anti-mouse antibody (Elite culture media from the cells and washing twice with 0.5 Vectastain Kit, Vector Laboratories, Burlingame, CA) ml Earle’s balanced salt solution containing 5 mM glu- diluted 1:200 in HS-PBS. Cultures were then placed for cose (EBSS). The cultures were returned to the incuba- 1 h in the avidin-horseradish peroxidase solution pretor for 1 h t o allow pH and temperature equilibration. pared according to the manufacturer’s instructions Stock solutions of 20 mM zinc chloride were prepared (Vectastain). The cells were washed with 0.175 M soby dissolving anhydrous ZnC1, in 0.5 mM HC1. The dium acetate, pH 6.8, and exposed to a fresh solution of stock was diluted to 1-5 mM in water immediately be- nickeVdiaminobenzidine prepared by dissolving 10 mg fore use and added to the cultures in 5-25 pL aliquots. diaminobenzidine and 625 mg nickel sulfate hexahyAt this dilution the zinc solutions did not alter the cul- drate into 15 ml of acetate buffer. Hydrogen peroxide ture media pH. The cultures were returned to the incu- was added to a final concentration of 0.005%, and after bator until zinc exposure was terminated by washing incubation for 20 min the cells were washed in PBS. GFAP staining was performed similarly using a twice with culture media. The potential of zinc chelation by serum proteins or amino acids to decrease the mouse monoclonal antibody (ICN 69-110-2, Costa effective concentration of zinc was tested by adding Mesa, CA) diluted 1:4,000 in PBS-HS. Control wells for FBS to the EBSS or by incubating with zinc in MEM both GFAP and HSP72 wells were prepared by omitting rather than EBSS. Similarly, zinc adsorption onto sur- the first antibody during the staining procedure.
200
SWANSONAND SHARP
fY 75 3 m 0
w 0
50
150
100
200
250
25
0.1 0
300
100
Fig. 1. Effect of incubation media on zinc toxicity. Astrocytes were incubated for 4 h in EBSS (open circles) or MEM (triangles)to which zinc was added from stock, or with dilute zinc EBSS solutions that had been stored 24 h prior to use (filled circles). n = 8 at. each data point. Error bars denote SEM in all graphs.
-
-
--
300
400
500
600
/.LM ZNCI,
pM ZnCI,
1 2 5 O~ ~ ZINC M
200
- 60 pM
ZINC
.-- - -
T
1
f
125T
B
1
2 3
cn -1
w
0
w 0
20
40
60
80
100
p M ZNCI,
%
FBS IN MEDIA DURING ZINC INCUBATION
Fig. 3. Dose and time dependence of zinc toxicity. A. Astrocytes were exposed to zinc in EBSS for periods of 15 min fopen triangles), 30 min (filled circles), and 60 min (open squares). B: Astrocytes were exposed for 2 h (filled triangles), 4 h fopen circles), and 8 h (filled squares). n = 8 at each data point.
Fig. 2. Effect of serum on zinc toxicity. Astrocytes incubated 4 h in 40 pM zinc in EBSS with added fetal bovine serum fFBS). n = 6 at each data point. *, different from control at P < 0.0 1 by Student's t test.
RESULTS Zinc toxicity was found to be highly dependent upon the experimental conditions. In the absence of amino acids or serum proteins the LD,, for zinc was approximately 50 pM after a 4 h exposure, with a steep doseresponse curve (Fig. 1). The presence of amino acids (MEM) caused a 50%reduction in the apparent LD,, of zinc, and the use of media in which zinc had been stored as a dilute solution produced a tenfold reduction in potency (Fig. 1).Serum also caused a n apparent reduction in zinc potency, evident with as little as 0.1% FBS in the media (Fig. 2). In the absence of zinc, incubation in EBSS or serum-free MEM for up to 24 h did not cause detectable cell death. The zinc toxicity was time- and dose-dependent (Fig. 3). After 8 h exposures in EBSS the LD,, for zinc was approximately 20 pM. At 15 and 30 min exposures the LD,, was increased to about 250 pM. Steep dose-response curves were seen a t all exposure times. Astrocytes cultured without dBcAMP-induced differentia-
tion were not appreciably different in their sensitivity to zinc (data not shown). The LDH assay determinations of cell death were in every case in good agreement with the impressions gained by phase microscopy inspection, in which nonviable cells were either rounded up or completely detached from the culture dishes (Fig. 4B). However, a t low zinc concentrations, many wells that showed no evidence of cell death by inspection or LDH content contained scattered cells or small patches of cells that were shown to be nonviable by ethidium homodimer fluorescence (Fig. 4C,D). Heating the astrocyte cultures to 44°C for 40 min induced widespread staining (Fig. 5A) that was entirely absent when the HSP72 antibody was omitted from the staining procedure (Fig. 5B). Zinc induction of HSP72 was investigated with dBcAMP-treated cultures, using EBSS without added amino acids or serum, and zinc solutions freshly diluted from stock. No cell death was seen after 4 h exposures to 5-10 pM zinc, but this concentration did induce HSP72 expression in small c h s ters of cells (Fig. 5C,D). Within the clusters the staining of cell nuclei was usually more intense than cytoplasmic staining, a s was seen after heat exposure (Fig. 5A),
201
ZINC TOXICITY AND HSP72 INDUCTION
Fig. 4. Phase contrast photomicrographs of control and zinc-exposed astrocytes. A Control well fixed and stained for GFAF'. B: Unfixed astrocytes 24 h after a 4 h exposure to 100 pM zinc, showing most cells rounded up or detached. C: Unfixed astrocytes 24 h after a 4 h exposure to 20 pM zinc. D: Same field a s C after addition of ethidium homodimer showing fluorescence in nuclei of scattered dead cells under blue light excitation. All photos at x200.
and many cells on the periphery of clusters exhibited only nuclear staining. Fifteen minute exposures to zinc induced HSP72 expression only a t zinc concentrations that also caused death or detachment of most cells (Fig. 5E). However, 4 h exposures to zinc at 20-60 p M killed only a portion of astrocytes in each culture well and produced a distinctive pattern of HSP72 expression (Fig. 6). Cell death occurred in patches of cells, typically two to four per well, distributed in an apparently random fashion (Fig. 6D). With exposures at higher concentrations o r for longer periods the patches became larger and connected to one another. The patches were invariably surrounded by a contiguous band of HSP72-positive astrocytes that separated the areas of dead or detached cells from the viable, HSP72-negative cells. HSP72 immunoreactivity was never seen in cells identified as nonviable by ethidium homodimer fluorescence (Fig. 6A,B).
To exclude the possibility that zinc precipitates were causing the patchy distribution of HSP72 induction and cell death, zinc chloride was dissolved in EBSS and the solutions were passed through a 0.22 p,m filter or centrifuged at 20,OOOg for 20 min prior to placement on astrocytes. These treatments did not alter the patterns of zinc-induced HSP72 expression and cell death. DISCUSSION In agreement with the report by Choi et al. (19881, zinc was found to be toxic to astrocytes in a dose- and time-dependent manner. This toxicity increased two to threefold in media free of amino acids or serum. As noted by Whelan and Hightower (1985), zinc can complex to amino acids, especially cysteine, histidine, and glutamine, such that the presence of these amino acids
202
SWANSON AND SHARP
Fig. 5 . HSP72 expression in astrocytes. A: 24 h after 40 min at 44°C. B: Same as A but with anti-HSP72 antibody omitted from staining procedure. C and D: In clusters of cells after 4 h exposure to 5pM zinc. E: In surviving cells after 15 min incubation with 400 pM zinc. Photos A-C at ~ 2 0 0D , and E at x 100.
even at low levels may markedly decrease the concentration of free zinc. We also observed a great decrease in apparent toxicity when zinc solutions were stored a t concentrations below 500 pM. It is likely that this effect was due to adsorption of zinc onto vessel surfaces, as storage a t 20 mM in acidic solution eliminated this effect. An acidic solution was used to increase the solubility of zinc and reduce surface adsorption,, as well as to prevent formation of zinc oxychloride.
Zinc has many effects on metabolism (Frederickson, 1989) and could injure cells in several ways. Like other heavy metals, zinc can react with thiol and imidazole moieties on enzymes and other proteins (Chvapil et al., 1972). Zinc also binds to S-100 (Baudier and Gerard, 1983) and tubulin (Eagle et al., 19831, and can disrupt microtubule formation (Fujii et al., 1986; Frederickson, 1989). In addition, zinc can act a t many calcium binding sites including calmodulin (Brewer, 1980)and may dis-
ZINC TOXICITY AND HSP72 INDUCTION
203
Fig. 6. HSP72 expression in astrocytes. A After exposure to 40 KM zinc for 4 h, x200. B: Same field as A, with ethidium homodimer fluorescence confirming cell death within a rim of viable, HSP72-positive cells. C, D: Low magnification views of HSP72-positive astrocytes surrounding patches of dead cells; C, ~ 4 0D, ; ~10.
rupt intracellular calcium homeostasis at micromolar concentrations (Chao et al., 1984). As zinc may gain entry into cells in part through voltage-sensitive cation channels (Choi and Koh, 1988; Weiss et al., 1989,1991), cells such as neurons and glia that possess such channels (MacVicar, 1984) may be particularly sensitive to zinc, Heat shock protein expression in cultured astrocytes was first demonstrated by Nishimura et al. (19881, by exposing cultures to 45°C. The heat shock response of astrocytes has since been further characterized (Dwyer et al., 1991; Marini et al., 19901, and acidosis has been identified as another inducing agent (Nishimura et al., 1989).Although zinc and other heavy metals are potent inducers in other cell types (Lindquist and Craig, 1988), they have not previously been noted to cause a heat shock response in astrocytes. In the present study, 15 min exposures induced HSP72 expression only at high concentrations of zinc that killed most of the astrocytes. However, with 4 h exposures, HSP72 expression was induced over a range of zinc concentrations, beginning well below and extending into the range that caused
cell death. At the lower zinc concentrations HSP72 expression occurred in clusters of cells, rather than in scattered individual cells. At higher Concentrations contiguous stained cells formed a rim around areas of cell death. These patterns may reflect heterogeneity of the astrocyte population (Hansson, 1988; Wilkin et al., 1990), with clustered cells having descended from common progenitor cells. Alternatively, the increased size of patches seen after increased zinc concentrations or exposure times suggests that a factor causing cell death and HSP72 expression may be passed or amplified through contiguous cells. The gap junctions of the glial syncytium provide one avenue by which this could occur (Kettenmann et al., 1983; Sontheimer et al., 1990). Although a number of insults can cause HSP72 expression in brain in vivo, several groups have noted that the protein is not expressed in cells with morphologic evidence of death or degeneration (Ferriero et al., 1990; Gonzalez et al., 1991; Lowenstein et al., 1990; Sharp et al., 1991; Simon et al., 1991; Vass et al., 1988, 1989).The inference that HSP72 expression occurs only in viable cells is supported here by absence of ethidium
SWANSONAND SHARP
204
homodimer fluorescence in HSP72 positive astrocytes (Fig. 6). However, as reported by Nowak (19911, cells that are irreversibly injured may fail to express the HSP72 protein but succeed in producing HSP72 mRNA. I t is of interest that the rim-like pattern of HSP72 expression seen here around patches of‘ dead cells is similar to the pattern of expression seen in glia after focal ischemia in vivo (see Sharp et al., 1991). Ischemia causes neuronal depolarization (Mitani et al., 1989) and could thus cause release of zinc into the extracellular space. In “penumbral” regions of reduced but nonzero blood flow, the rate of glucose utilization is increased for several hours after vessel occlusion, suggesting a sustained period of increased neuronal activity (Shiraishi et al., 1989). Brain extracellular zinc levels have been estimated to reach 30 p M (Aniksztejn et al., 1987) to 300 p M (Assaf and Chung, 1984)during intensive neuronal activity, levels well above those seen here to induce HSP72 and astrocyte death in the present study in vitro. However, extrapolation of the present observations to in vivo conditions must be done with caution. To our knowledge, direct measurements of zinc in brain extracellular fluid have not yet been accomplished. Moreover, as we have shown decreased toxicity in the presence of amino acids and proteins, zinc toxicity in vivo may likewise be decreased by binding to amino acids and proteins in extracellular fluid. On the other hand, zinc toxicity in vivo may be augmented by conditions such as low pH or membrane depolarization, which could facilitate the entry of zinc into cells (Weiss et al., 1991). Lastly, reuptake mechanisms (Howell e t al., 1984) may drastically alter the extent and duration of zinc elevations in normal and ischemic brain. These issues are currently under study. Part of the rationale for characterizing zinc induction of HSP72 in astrocytes is that a chemical agent capable of inducing heat shock proteins in brain could be useful in the study of the biological functions of these proteins. Unfortunately, our results suggest that zinc holds little promise in this regard because of the steep dose-response curve of zinc cytotoxicity and the narrow range between doses needed to cause widespread HSP72 expression and those resulting in cell death. Zinc and other heavy metals are potent inducers of metallothionein, and zinc is now commonly used to activate a metallothionein promoter in recombinant genes (Durnam and Palmitter, 1981; Hurko et al., 1986). The present study suggests that in glial cells, and perhaps in other cell types as well, this maneuver may be complicated by the simultaneous induction of HSP72 and other stress proteins,
ACKNOWLEDGMENTS We would like to thank Marcia Stubblebine for technical assistance and Dr. John Weiss for helpful discussions. R.A.S. is supported by a Research Associate career development -award from the Department of
Veterans Affairs. This work is also supported by NIH grants R01-NS28167 and NS14543 (F.R.S.), and by the Merit Review program of the VA Research Service (R.A.S.andF.R.S.).
REFERENCES Aniksztejn, L., Charton, G., and Ben-Ari, Y. (1987) Selective release of endogenous zinc from the hippocampal mossy fibers in situ. Brain Res., 40458-64. Assaf, S.Y. and Chung, S-H. (1984) Release of endogenous Zn ’ ’ from brain tissue duringactivity.Nature, 308:734-736. Baudier, J . and Gerard, G. (1983)Ions binding to SIOO proteins: Structural changes induced by calcium and zinc on SlOOa and SlOOb proteins. Biochemistry, 22:3360-3369. Berger, E.M. and Woodward, M.P. (1983) Small heat shock proteins in Drosophila may confer thermal tolerance. Exp. Cell Res., 147:437442. Brewer, G. (1980) Calmodulin, zinc and calcium in cellular and membrane regulation: An interpretive review. Am. J. Hematol., 8:231248. Caltabiano, M.M., Koestler, T.P., Poste, G., and Grieg, R.G. (1986) Induction of 32- and 34-kDa stress proteins by sodium arsenite, heavy metals, and thiol-reactive agents. J . Biol. Chem., 261:1338113386. Chan, A.W.K., Minski, M.J., and Lai, J.C.K. (1983) An application of neutron activation analysis to small biological samples: Simultaneous determination of thirty elements in rat brain regions. J. Neurosci. Methods, 7:317-328. Chao, S.-H., Suzuki, Y., Zysk, J.R., and Cheung, W.Y. (1984) Activation of calmodulin by various metal cations as a function of ionic radius. Mol. Pharmacol., 26:75-82. Choi, D.W. and Koh, J. (1988) Zinc central neurotoxicity may require open NMDA channels. Sac. Neurosci. Abstr., 14:417. Choi, D.W., Yokoyama, M., and Koh, J. (1988) Zinc neurotoxicity in cortical cell culture. Neuroscience, 24:67-79. Chopp, M., Chen, H., Ho, K.-L., Dereski, M.O., Brown, E., Hetzel, F.W., and Welch, K.M.A. (1989) Transient hyperthermia protects against subsequent forebrain ischemic damage in the rat. Neurology, 39:1396-1398. Chvapil, M., Elias, S.L., Ryan, J.N., and Zukoski, C.F. (1972) Pathophysiology of zinc. In: Neurobiology of the Trace Metals Zinc and Copper. C. C. Pjeiffer, ed. Academic Press, New York, pp. 105-124. Danscher, G., Howell, G., Perez-Clausel, J., and Hertel, N. (1985) The dithizone, Timm’s sulphide silver and the selenium methods demonstrate a chelatable pool of zinc in CNS. Histachemistry, 83:419-422. Durnam, D. and Palmiter, R.D. (1981) Transcriptional regulation of the mouse metallothionein-I gene by heavy metals. J . Biol. Chem., 256~5712-5716. Dwyer, B.E., Nishimura, R.N., de Vellis, J., and Clegg, K.B. (1991) Regulation of heat shock protein synthesis in rat astrocytes. J . Neurosci. Res., 28:352-358. Eagle, G.R., Zombola, R.R., and Himes, R.H. (1983)Tubulin-zinc interactions: Binding and polymerization studies. Biochemistry, 22221228. Ferriero, D.M., Soberano, H.Q., Simon, R.P., and Sharp, F.R. (1990) Hypoxia-ischemia induces heat shock protein-like (HSP72) immunoreactivity in neonatal rat brain. Deu. Brain Res., 53:145-150. Frederickson, C.J. (1989) Neurobiology of zinc and zinc-containing neurons. Int. Rev. Neurobiol., 31:145-238. Frederickson, C.J., Hernandez, M.D., and McGinty, J.F. (1989) Translocation of zinc may contribute to seizure-induced death of neurons. Brain Res., 480:317-321. Fujii, T., Gochou, N., Akabane, Y., Fujii, M., Kondo, Y., Suzuki, T., and ’ Ohki, K. (1986) Effect of zinc ions on the interaction of S-100 protein with brain microtubule proteins. Chem. Pharm. Bull. (Tokyo), 345040-5044. Gonzalez, M.F., Lowenstein, D.L., Fernyak, S., Hisanaga, K., Simon, ‘R.P., and Sharp, F.R. (1991) Induction of heat shock protein72-like immunoreactivity in the hippocampal formation following transient global ischemia. Brain Res. Bull., 26:241-250. Hansson, E. (1988) Astroglia from defined brain regions as studied with primary cultures. Prog. Neurobiol., 30:369-397. Hertz, L. (1990) Dibutyryl cyclic AMP treatment of astrocytes in primary cultures as a substitute for normal morphogenic and ‘functiogenic’transmitter sienals. In: Molecular AsDects ofDeuelowmentand k i n g of the Nervoussystem. J.M. Lauder,*A. Private, E.’Giacobini, and A. Vernadakis, eds. Plenum Press, New York, pp. 227-243.
ZINC TOXICITY AND HSP72 INDUCTION Hertz, L., Juurlink, B.H.J., and Szuchet, S. (1985) Cell cultures, In: Handbook OfNeurochemistry, Vol. 8,2nd ed. A. Lajtha, ed. Plenum Press, New York, pp. 603461. Hoffman, G.E., Lee, W . 3 , Attardi, B., Yann, V., and Fitzsimmons, M.D. (1990) Luteinizing hormone-releasing hormone neurons express c-fos antigen after steroid activation. Endocrinology, 126:173&1741. Howell, G.A., Welch, M.G., and Frederickson, C.J. (1984) Stimulationinduced uptake and release of zinc in hippocampal slices. Nature, 308:736-738. Hurko, O., McKee, L., and Zuurveld, J.G.E.M. (1986) Transfection of human skeletal muscle cells with SV40 large T antigen gene coupled to a metallothionein promoter. Ann. Neurol., 20573-582. Kettenmann, H., Orkand, R.K., and Schachner, M. (1983) Coupling among identified cells in mammalian nervous system cultures. J . Neurosci., 3506-516. Koh, J-Y. and Choi, D.W. (1987) Quantitative determination of glutamate mediated cortical neuronal injury in cell culture by lactate dehydrogenase efflux assay. J . Neurosci. Methods, 20:83-90. Lindquist, S. and Craig, E.A. (1988) The heat-shock proteins. Annu. Rev. Genet., 22:631477. Lowenstein, D.H., Simon, R.P., and Sharp, F.R. (1990) The pattern of 72-kDa heat shock protein-like immunoreactivity in the rat brain following flurothyl-induced status epilepticus. Brain Res., 531:173182. MacVicar, B.A. (1984) Voltage-dependent calcium channels in glial cells. Science, 226:1345-1347. Marini, A.M., Kozuka, M., Lipsky, R.H., and Nowak, T.S., J r . (1990) 70-Kilodalton heat shock protein induction in cerebellar astrocytes and cerebellar granule cells in vitro: Comparison with immunocytochemical localization after hyperthermia in vivo. J . Neurochem., 543509-1516. Misra, S., Zafarullah, M., Price-Haughey, J., and Gedamu, L. (1989) Analysis of stress-induced gene expression in fish cell lines exposed to heavy metals and heat shock. Biochim. Biophys. Acta, 1007:325333. Mitani, A,, Imon, I., and Kataoka, K. (1989) High frequency discharges of gerbil hippocampal CA1 neurons shortly after ischemia. Brain Res. Bull., 23569-572. Nishimura, R.N., Dwyer, B.E., Welch, W.J., Cole, R., deVellis, J., and Liotta, K. (1988) The induction of the major heat-stress protein in purified rat glial cells. J . Neurosci. Res., 20:12-18. Nishimura, R.N., Dwyer, B.E., Cole, R., de Vellis, J., and Picard, K. (1989) Induction of the major inducible 68-kDa heat-shock protein after rapid changes of extracellular pH in cultured rat astrocytes. Exp. Cell Res., 180:276-280. Nover, L (1984) Inducers of hsp synthesis. In: Heat Shock Response of
205
Eukaryotzc Cells. L. Nover, ed. Springer-Verlag, New York, pp. 1-11. Nowak, T.S., J r . (1991) Localization of 70 kDa stress protein mRNA induction in gerbil brain after ischemia. J . Cereb. Blood Flow Metab., 11:432-439. Pelham, H.R.B. (1986) Speculations of the functions of the major heat shock and glucose-regulated proteins. Cell, 46:959-961. Qiabowol, K.T., Mizzen, L.A., and Welch, W.J. (1988) Heat shock is lethal to fibroblasts injected with antibodies against hsp7O. Science, 242:433-436. Sharp, F.R., Lowenstein, D.L., Simon, R.P., and Hisanaga, K. (1991) Heat shock protein HSP72 induction in cortical and striatal astrocytes and neurons following infarction. J . Cereb.Blood Flow Metab., 11:621M28. Shiraishi, K., Sharp, F.R., and Simon, R.P. (1989) Sequential metabolic changes in rat brain following middle cerebral artery occlusion: A 2-deoxyglucose study. J . Cereb. Blood Flow Metab., 9:765-773. Simon, R.P., Cho, H., Gwinn, R., and Lowenstein, D.H. (1991) The temporal profile of 72-kDa heat-shock protein expression following global ischemia. J . Neurosci., 11:881-889. Sqntheimer, H., Minturn, J.E., Black, J.A., Waxman, S.G., and Ransom, B.R. (1990) Specificity of cell-cell coupling in rat optic nerve astrocytes in vitro. Proc. Natl. Acad. Sci. USA, 87:9833-9837. Vass, K., Welch, W.J., and Nowak, T.S., Jr. (1988) Localization of 70-kDa stress protein induction in gerbil brain after ischemia. Acta Neuropathol., 77:12%135. Vass, K., Berger, M.L., Nowak, T.S., Jr., Welch, W.J., and Lassmann, H. (1989) Induction of stress protein HSP7O in nerve cells after status epilepticus in the rat. Neurosci. Lett., 100:259-264. Wiss, J.H., Koh, J-Y., Christine, C.W., and Choi, D.W. (1989)Zinc and LTP. Nature, 338:212. Weiss, J.H., Hartley, D.M., and Choi, D.W. (1991)Zinc potentiation of AMPA neurotoxicity: Possible mediation by Zn2+entry and upregulption of Ca2' influx. Neurology, 41(Suppl.):377. Welch, W.J. and Suhan, J.P. (1986) Cellular and biochemical events in mammalian cells during and after recovery from physiological stress. J . Cell. Biol., 103:2035-2052. Welch, W.J., Mizzen, L.A., and Arrigo, A.-P. (1989) Structure and function of mammalian stress proteins. In: Stress-Induced Proteins. M.L. Pardue, J.R. Feramisco, and S. Lindquist, eds. Alan R. Liss, New York, pp. 187-202. Whelan, S.A. and Hightower, L.E. (1985)Induction of stress proteins in chicken embryo cells by low-level zinc contamination in amino acid-free media. J . Cell. Physiol., 122:205-209. Wilkin, G.P., Marriott, D.R, and Cholewinski, A.J. (1990) Astrocyte heterogeneity. T I " , 13:4346.
