Neuron,
Vol. 5, 121-126, August, 1990, CopyrIght
0 1990 by Cell Press
2kAminosteroids Attenuate Excitotoxic Neuronai Injury in Cortical Cell Cultures Hannelore Monyer: Dean M. Hartley, and Dennis W. Choi Department of Neurology Stanford University Medical Center Stanford, California 94305
Summary We studied the protective efficacy of novel 21-aminosteroids against several forms of neuronal injury in murine cortical cell cultures. Concentrations of 200 nM to 20 PM partially attenuated the damage induced by glucose deprivation, combined oxygen-glucose deprivation, or exposure to NMDA; maximal protection was less than that produced by NMDA antagonists, but the combination of a 2%aminosteroid plus an NMDA antagonist produced a greater benefit than either drug alone. 2lAminosteroid addition did not attenuate NMDA-induced whole-cell current, but did block almost all of the damage induced by exposure to iron, a protective action consistent with inhibition of free radical-mediated lipid peroxidation. Lipid peroxidation may be a downstream event mediating a portion of the injury triggered by excess stimulation of NMDA receptors. Introduction Free radical formation and subsequent lipid peroxidation has been postulated to participate broadly in the pathogenesis of tissue injury, including specifically the brain injury induced by hypoxia or trauma (Watson et al., 1984; Halliwell and Gutteridge, 1985; Halliwell, 1987; Braughler and Hall, 1989; Siesj& 1989). Other recent studies have suggested that the toxic overstimulation of postsynaptic glutamate receptors, in particular the N-methyl+aspartate (NMDA) subtype, may be an important mechanism of neuronal injury following several types of acute insults (Rothman and Olney, 1987; Choi, 1988), including hypoxia (Rothman, 1984; Weiss et al., 1986), glucose deprivation (Wieloch, 1985), ischemia (Simon et al., 1984), prolonged seizures (Clifford et al., 1989; Furshpan and Potter, 1989), and trauma (Faden and Simon, 1988; Tecoma et al., 1989). The relationship between free radical-induced injury and glutamate neurotoxicity is presently undefined. Although these two processes could be entirely separate, it is attractive to consider that free radical formation might be a direct consequence of glutamate receptor overstimulation and an important downstream mediator of glutamate-induced neuronal death
* Present address: Laboratory of Molecular Neuroendocrinology, ZMBH, University of Heidelberg, INF 282, D-6900 Heidelberg, Federal Republic of Germany.
(Choi, 1988,199O). First, glutamate neurotoxicity in cortical cell cultures may be initially triggered by an influx of extracellular calcium; the resultant buildup of intracellular calcium could activate phospholipase AZ, leading to the liberation of arachidonic acid and subsequent free radical production (Chan et al., 1985). Second, calcium may also trigger the conversion of xanthine dehydrogenase to xanthine oxidase, a rich enzymatic source of free radicals (SiesjG, 1989). Of note, kainate-induced injury of cerebellar neurons can be attenuated by allopurinol, an inhibitor of xanthine oxidase (Dykens et al., 1987). Third, stimulation of NMDA receptors can lead to the release of nitric oxide (Garthwaite et al., 1988), which can react with superoxide to form peroxynitrite and ultimately promote the production of hydroxyl radicals (Beckman et al., 1990). Finally, Murphy and Coyle have suggested that glutamate exposure may produce cytotoxicity in a neuronal cell line because of inhibition of cystine uptake, resulting in reduced glutathione production and increased oxidative stress (Murphy et al., 1989). Recently, a novel series of 21-aminosteroid compounds that are highly lipophilic and potently inhibit lipid peroxidation in isolated membrane preparations have been developed (Braughler et al., 1987; Hall et al., 1987). 21Aminosteroids reduce neuronal damage in several animal models of acute CNS injury (Hall, 1988; Hall and Yonkers, 1988). Two questions arise from these important studies. First, can 21-aminosteroids reduce the intrinsic vulnerability of brain cells to injury? Benefits observed in vivo could reflect indirect effects on systemic metabolism or cerebral blood flow. Second, if these drugs can act directly on brain cells, is improved neuronal survival related to a reduction of glutamate neurotoxicity? We decided to examine the ability of 2%aminosteroids to protect murine cortical neurons in primary cell culture against several forms of injury previously found to be mediated by NMDA receptors. Results As previously reported, mixed cortical cell cultures exposed to a 6-8 hr period of glucose deprivation exhibited acute neuronal swelling; over the next day, many neurons, but not glia, disintegrated (Monyer and Choi, 1988). Addition of the 2%aminosteroid U74QO6F at concentrations of 200 nM to 20 PM to sister cultures prior to the insult produced partial reduction in this morphologically apparent neuronal injury; protection was maximal at 2-5 PM concentrations of U74006E A slight suggestion of neuronal toxicity was seen at 20 FM (granularity of cell bodies), and concentrations higher than 20 PM precipitated out. The effect of U74006F on neuronal injury could be assessed quantitatively by measuring the efflux of the cytosolic enzyme lactate dehydrogenase (LDH) to the bathing medium. Maxi-
NetJrWl
122
CTRL
0
U74500A
OX
120
100
60 3 9
60
P 5
40
20
0 CTRL
DX
U74500A
DX+ U74500A
Figure 2. U745OOA Attenuates Neuronal Degeneration by Combined Oxygen and Glucose Deprivation
Figurel. Comparison of Attenuation by Dextrorphan and 21. Aminosteroids of Glucose Deprivation-Wfuced Neuronal Injury Sister cultures were exposed to glucose deprivation alone (CTRL), in the presence of 100 P M dextrorphan (DX), or in the presence of the indicated concentrations of U74006F (A), 2 u M U745OOA (B), or 2 u M U75412E (C). Bars show LDH efflux the following day (mean f SEM, n = 3-4 cultures per point), expressed relative to that found in the no drug control (= 100). LDH levels in the presence of dextrorphan in (A) and (C)were actually below the sham wash value (-3.0 in [A], - 7.5 in [Cl), but are shown at 0 for ease of display. An asterisk indicates difference from control at P < 0.05, ANOVA and Student-Newman-Keuls’ test.
mal protection varied considerably from experiment to experiment (range of injury reduction, 15%-70%, mean = 52.6% + 22.7% SD, n = 9 experiments), but was always less than the 80%-100% reduction produced in sister cultures by 100 P M dextrorphan or another NMDA antagonist (Figure IA). Similar partial protection against glucose deprivation-induced neuronal injury was produced by optimal concentrations (2 PM) of the related 21-aminosteroids, U745OOA (Figure IB) and U75412E (Figure IC).
Induced
(A) Sister cultures were exposed to combined oxygen and glucose deprivation for 50 min either alone (CTRL), or in the presence of 4 u M U745OOA or 100 u M dextrorphan (DXi. Bars here and in (B) show LDH efflux the following day (mean f SEM, n = 4-5 cultures per point). (B) Pooled data from 4 experiments using 100 u M dextrorphan and 4 u M U745OOA; the period of combined oxygen and glucose deprivation was extended to 1.5 hr to overcome partially the protective effect of dextrorphan (n : 19-23 cultures per point). The condition with both dextrorphan and U745OOA was also significantly different from that with either dextrorphan or U745OOA alone. An asterisk indicates difference from control at P < 0.05, ANOVA and Student-Newman-Keuls’ test.
The simultaneous removal of both oxygen and glucose produces widespread neuronal damage with exposure times of only30-60 min (Goldberg et al., 1988, Sot. Neurosci., abstract). Addition of 4 P M U745OOA prior to insult also produced partial attenuation of this more fulminant form of injury (Figure 2A; 5 out of 6 experiments). If the duration of oxygen and glucose deprivation was increased to 90 min, the protective action of either 100 F M dextrorphan or 4 P M U745OOA was largely overcome, but the combination of the two still produced some benefit (Figure 2B).
;:;Aminosteroids
Attenuate
Excitotoxicity
100
100
80 60
*
z 2
60
g
40
60
40
20
5 CTRL Figure 3. 21Amino:teroids ronal Injury
DX Effectively
U74500A
Block
U74006F
Iron-Induced
Neu-
Cultures’were exp&& fd; 2i hi to 50 PM FeZi and 50 PM Fe3+ alone (GIRL), in the presetice of 100 vth;l dextroibhan, or in the presence of a ~JIM concenlration of the indicated 2%aminosterqid. Bars &ow LDH at the end of the exposure period (mean + SEW, n, = 4). An asterisk indicates difference from control at $2 6.05, ANO%.and Studeht-Newman-Keuls’ test.
The hasi for 2%aminosteroid protection in vivo I$ been suggested to be inhibition of lipid peroxidation, possibly in part related to free radical scavenging (see above). To examine the ability of these compounds to t&ck free radical-mediated injury in our system, we tested them against the damage directly evoked by adding iron ions, a maneuver expected to enhan&a both hydroFyl,radical formation and lipid peroxidation, (,$ndweon and Means, 1985; Aust et al., 1985). The addition of a combination of 50 PM Fe2+ and 50 &vt. Fq3$$o the cultures resulted in widespread neuronql degeneration after 24 hr; glia were grossly intact. This probable peroxidative injury could be almost completely blocked by concurrent addition of 4 PM U745OOA or U74006F (Figure 3). In contrast, less than half of this injury was prevented by 100 PM dextrorphan (Figure 3; 3 of 3 experiments). The observed ability of 21-aminosteroids to attenuate forms of neuronal injury sensitive to NMDA antag onists is consistent with a substantial role for free radicals in mediating NMDA receptor-induced neurotoxicity. To test this idea directly, we examined the ability of U745OOA to reduce the,damage induced by the exogenous application of defined concentrations of NMDA. As previously reported, the widespread neuronal degeneration induced by brief exposure to
NHDA
N?4DA+U74500
NMDA
Figure 4. 21Aminosteroids Neurotoxicity
Reduce
NMDA
Receptor-Mediated
Cultures were exposed for 3 rni; to,!+0 PM NMDA, and LDH was measured the following day (mean + &M, n = 4). Test cultures were treated with 100 PM dextrorphan or 5 &I U745OOA as indicated, using one of three protocols for drug addition: preadded 2 hr before toxic exposure; acute-added during the 3 min toxic exposure only; or ppst-added immediately following toxic exposure and left in until assessment the next day. The difference between the conditions POST DX + U745OOA and POST DX was not statistically significant in this single experiment, but did reach statistical significance when the data were pooled with two other similar, experiments (POST DX + ,W745OOA, 18.8 it 1.8 [n = 121; ,POSJ DX, 32.0 f 2.6 [,n 7 II]; different ai’p < 0.05, ANOVA and Student-Neuman-Keuls’ test):
500 PM NMDA could be blocked almpst cgm.pletely by coapplica,t,ion of 100 pw dextrorphan (Choi et al., 1987b) and partially blocked by the late application of dextrorphan immediately following completion of glutamate exposure (Hartley and Choi, 1989; Figure 4). Preapplication of 5 PM U745OOA (added 2 hr before NMDA) partially attenuated this injury;.a small protective effect could still be detected even when U7!5OOA was added immediately.after completio,n of the,NMDA exposure(Figure 4). TJe combination of both dextrorphan and U745OOA, addeq.la!e after NMQA exposure, produced greater survival than either dru,g,alone (Figure 4). Similar r$sults were obtained when,/njury was, induced with $00 PM glutamate instead of NMDA (2 experiments; data pot shown). The ability pf u745OOA to attenuate thk neurotoxicity of exogenously added NMDA or glutamate was not due to direct antagonism of NMDA receptors. The whole-cell current evoked by application of 10 PM NMDA was not reduced by concomitant,application of 10 PM U745OOA, but was expectedly blocked by the
Figure 5. U745OOA Does Not Block NMDAInduced Whole-Cell Current
NMDA+DX
v v -I30
l 1.0 set
nA
Whole-cell recording of the membrane current response induced by 10 uM NMDA at a holding potential of -60 mV. The response is not a&red by inclusion of 20 vi U74500A in the NMDA solution, but is blocked by inclusion of 30 PM dextrorphan (DX).
Neuron 124
concomitant ure 5).
application
of 20 P M dextrorphan
(Fig-
Discussion Low micromolar concentrations of several 2%aminosteroids partially protected cultured cortical neurons against damage induced by either glucose deprivation, or combined oxygen and glucose deprivation. These observations suggest that 2%aminosteroids can act directlyon brain cells to reduce neuronal vulnerability to injury and therefore support the idea that a parenchymal action underlies the reported ability of the drugs to reduce brain damage in vivo. Since in the same experiments, the selective NMDA antagonist dextrorphan provided almost complete neuronal protection, 2%aminosteroids might work, at least in part, by attenuating NMDA receptor-induced injury mechanisms. Alternatively, 21-aminosteroids could block events unrelated to NMDA receptors, but triggered by the same initial insults and required for cell death under the tested conditions. These two possibilities are not mutually exclusive, and indeed our observations provide support for both. The possibility that 21-aminosteroids can attenuate NMDA receptor-induced injury is supported by the finding that U745OOA produced a similar partial attenuation of the neuronal injury induced by the exogenous application of glutamate or NMDA. This injury attenuation was greatest when cultures were preincubated in the drug prior to excitotoxic exposure, perhaps reflecting improved opportunity to gain access to membrane lipids or intracellular compartments. Since U745OOA did not reduce NMDA receptor-induced whole-cell currents, it is not an NMDA antagonist. Most likely, U745OOA interferes with events linking NMDA receptor overstimulation to subsequent cell death. W e have proposed a speculative breakdown of glutamate receptor-mediated neurotoxicity into three stages: first, induction-overstimulation of membrane receptors leading to a set of immediate cellular derangements; second, amplification-events that intensify these derangements and promote the excitotoxic involvement of additional neurons; and third, expression-the destructive cascades directly responsible for neuronal cell degeneration (Choi, 1990). While additional studies will be needed to define the steps that may be modified by 2+aminosteroids, a reasonable working hypothesis is that the ability of U745OOA to attenuate NMDA receptor-mediated injury is largely related to its documented activity as an inhibitor of lipid peroxidation. This hypothesis is supported by our finding that U745OOA, as well as U74006F, was highly effective in blocking the damage triggered by addition of Fe2+/Fe3+ to the same cultures (see below). The ability of dextrorphan to produce partial injury attenuation in this paradigm does not argue against primary involvement of free radicals; several different
types of initial injury may be compounded by triggering the excess release of endogenous glutamate stores, leading to secondary NMDA receptor-mediated damage. Chan and colleagues have found that cerebellar neurons cultured from transgenic mice with increased levels of superoxide dismutase, an enzyme involved in the detoxification of superoxide radicals, are resistant to both oxidative stress and glutamate toxicity (Chan et al., 1989, Sot. Neurosci., abstract; also meeting communication). Free radicals may play a role in excitotoxic amplification, as they can enhance the release of glutamate from rat hippocampal slices independent of cell lysis (Pellegrini-Giampietro et al., 1988). Continued activation of glutamate receptors by release of endogenous glutamate from injured neurons may constitute an important positive-feedback loop, potentiating the damage caused by initial excitotoxic exposure; the protective effect of late dextrorphan addition may be due to interruption of this loop (Hartley and Choi, 1989). Furthermore, it is possible that peroxidized lipids may be a preferred substrate for phospholipase A2 (Van Kuijk et al., 1987). Thus lipid peroxidation could enhance the formation of arachidonic acid, which may diffuse back to presynaptic terminals and upregulate presynaptic glutamate release (Williams et al., 1989). The fact that U745OOA produced an additive benefit when combined with delayed dextrorphan (Figure 4, POST DX + U745OOA) suggests that its action is probably not restricted to attenuating glutamate release. Free radicals are attractive candidates for mediating the final injury cascades involved in the expression of excitotoxicity. Besides damaging membranes by lipid peroxidation, free radicals can damage proteins and DNA; hydroxyl radicals, formed from superoxide anions and hydrogen peroxide in the presence of transitional metals such as iron, may be particularly destructive (Siesjo, 1989). Free radicals may also participate in other injury mechanisms unrelated to NMDA receptors; it is noteworthy that, in the setting of prolonged combined oxygen and glucose deprivation, U745OOA produced some benefit in addition to that produced by continuous high concentrations of dextrorphan (Figure 2B). NMDA receptor-induced excitotoxic injury and free radical-mediated injury thus may be events that overlap substantially, but not completely. A major portion of NMDA receptor-induced injury is resistant to lipid peroxidation inhibitors such as 21-aminosteroids and could be mediated by other cytotoxic events, such as the activation of catabolic enzymes (Siman and Noszek, 1988; Choi, 1988). Conversely, free radicals may be formed by events unrelated to NMDA receptors. The combination of an NMDA antagonist with approaches directed against free radical injury mechanisms could be a useful strategy for reducing certain forms of acute brain damage.
:;-$minosteroids
Experimental
Attenuate
Excitotoxicity
Procedures
Cortical Cell Culture Mixed cortical cell cultures, containing both neuronal and glial elements, were prepared generally as previously described (Choi et al., 1987a) from fetal mice at 14-18 days gestation. Dissociatcdcortical cells were plated in Primaria (Falcon) 15 m m multiwell vessels (about 2.8 x IO5 cells per well) in Eagle’s minimal essential medium (MEM without Earle’s salts, supplied glutamine-free) supplemented with 10% heat-inactivated horse serum, 10% fetal bovine serum, glutamine (2 mM), and glucose (total 21 mM). Cultures were kept at 3PC in a humidified, CO*containing atmosphere. After 5-12 days in vitro, nonneuronal cell division was halted by l-3 days of exposure to 10e5 M cytosine arabinoside, and the cells were shifted into a maintenance medium identical to the plating medium, but lacking fetal serum. Subsequent medium replacement was carried out twice per week. Only mature (15-20 days in vitro) cortical cultures were selected for study; whenever possible, comparisons were made on matched sister cultures derived from a single plating. Glucose and Combined Oxygen-Glucose Deprivation Glucose deprivation was carried out by exchanging the culture medium for a defined solution resembling MEM, but lacking glucose and vitamins. Cultures were then placed back in the CO* incubator for 6-8 hr. Glucose deprivation was terminated by adding a small amount of concentrated glucose to the exposure solution (final concentration 5.5 mM), and cultures were returned to the incubator overnight prior to assessing injury. In cultures in which oxygen and glucose deprivation were combined, the culture medium was rapidly exchanged with deoxygenated, bicarbonate-buffered salt solution (02 partial pressure < 1 m m Hg), identical to Earle’s balanced salt solution except for the lack of glucose. Cultures were kept for 50-60 min at 37OC in a humidified atmosphere of Nz with 5% COZ (some H2 was added to remove any residual OZ via palladium catalyst). This insult was terminated by the thorough washout of the exposure medium with oxygenated MEM containing 5.5 m M glucose, and cultures were returned to a normoxic, humidified incubator (3PC, 5% COZ) overnight prior to assessing injury. Excitatory Amino Acid Exposure Exposure to glutamate or NMDA was carried out at room temperature in a HEPES-buffered control salt solution (HCSS), substituted for culture medium by triple exchange; HCSS had the following composition: 120 m M NaCI, 5.4 m M KCI, 0.8 m M MgC&, 1.8 m M CaC12, 20 m M HEPES (pH %4 at 25”C), 15 m M glucose. After 3 min, the neurotoxin was quickly washed out (>7000fold dilution) and replaced with another defined solution, Eagle’s MEM with augmented glucose (21 mM), prior to returning the cultures to the 37°C incubator. Neuroprotective drugs were added 2 hr before this excitotoxic exposure (pretreatment), during this exposure (acute treatment), or immediately after this exposure (posttreatment); in the latter case, the drug remained in the medium until the next day, when neuronal damage was assessed. Control experiments showed that little or no cortical cell damage was produced by this protocol if excitatory amino acids were omitted from the exposure solution. Monitoring with phenol red verified that the pH was not appreciably altered by metabolite deprivation or excitotoxic exposure. Assessment of Overall Neuronal Cell Injury Overall neuronal cell injury was estimated in all experiments by examination of cultures with phase-contrast microscopy at 100x to 200x. This examination was usually performed 1 day after initiation of metabolite deprivation or excitotoxic insult, at which point the process of cell death was largely complete; previous experience has suggested that such injury can be reliably estimated in this fashion. In some experiments, this examination was verified by subsequent bright-field examination of trypan blue staining (04% for 5 min), a dye staining debris and nonviable cells.
Overall neuronal cell injury was also quantitatively assessed by the measurement of LDH, released by damaged or destroyed cells, in the extracellular fluid 1 day after the experiment (Koh and Choi, 1987). A small amount of LDH was always present in the media of cultures exposed to sham wash alone. This background amount, determined on sister cultures within each experiment, was subtracted from values obtained in treated cultures. Control experiments have shown that the specific efflux of LDH induced by glutamate exposure or hypoxia is linearly proportional to the number of neurons damaged or destroyed (Koh and Choi, 1987; Goldberg et al., 1987). Electrophysiology Dishes were maintained on the stage of an inverted Nikon Diaphot microscope at room temperature. We recorded from single neurons with patch-clamp technique in the whole-cell configuration. Holding potential was --60 mV. The pipette solution was 140 m M CsCI, 0.5 m M CaC12, 10 m M HEPES, 5 m M EGTA. The perfusion solution contained 125 m M NaCI, 2.8 m M KCI, 1 m M CaC&, and 10 m M HEPES; 1 PM tetrodotoxin and 5 PM glytine were also included. Drug solutions were applied to cells by gravity feed through a hydraulic switch and a large bore pipette. Materials and Reagents Glutamate and NMDA were obtained from Sigma, 21-aminosteroids from Upjohn, and dextrorphan from Hoffman La-Roche. 21Aminosteroids were dissolved initially in dimethylsulfoxide at 10 m M and later diluted to the desired working concentration. Control experiments established that the concentrations of dimethylsulfoxide used did not affect neuronal injury in the tested paradigms. Acknowledgments This work was supported by National Institutes of Health grants NS 12151 and 26907 and by a grant from the American Paralysis Association. H. M. was supported by a fellowship from the Deutsche Forschungsgemeinschaft. D. M. H. was supported by a fellowship from the National Institute of Mental Health (MH 09823). Received
April
4, 1990; revised
June 7, 1990.
References Anderson, D. K., and Means, E. D. (1985). Iron-induced lipid peroxidation in spinal cord: protection with mannitol and methylprednisolone. J. Free Radic. Biol. Med. 1, 59-64. Aust, S. D., Morehouse, L. A., and Thomas, C. E. (1985). Role of metals in oxygen radical reactions. J. Free Radic. Biol. Med. 7, 3-25. Beckman, J. S., Beckman, T. W., Chen, J., Marshall, F’. A., and Freeman, B. A. (1990). Apparent hydroxyl radical production by peroxynitrite: implications for endothelial injury from nitric oxide and superoxide. Proc. Natl. Acad. Sci. USA 87, 1620-1624. Braughler, J. M., and Hall, E. D. (1989). Central nervous system trauma and stroke. I. Biochemical considerations for oxygen radical formation and lipid peroxidatlon. J. Free Radic. Biol. Med. 6, 289-301. Braughler, J. M., Pregenzer, J. F., Chase, R. L., Duncan, L. A., Jacobsen, E. J., and McCall, J. M. (1987). Novel 21-amino steroids as potent inhibitors of iron-dependent lipid peroxidation. J. Biol. Chem. 262, 10438-10440. Chan, P. H., Fishman, R. A., Longar, S., Chen, S., and Yu, A. (1985). Cellular and molecular effects of polyunsaturated fatty acids in brain ischemia and injury. Prog. Brain Res. 63, 227-235. Choi, D. W. (1988). Glutamate neurotoxicity nervous system. Neuron 7, 623-634.
and diseases
of the
Choi, D. W. (1990). Methods for antagonizing glutamate rotoxicity. Cerebrovasc. Brain Metab. Rev. 2, 105-147,
neu-
Neuron 126
Choi, D. W., Maulucci-Gedde, M. A., and Kriegstein, A. R. (1987a). Glutamate neurotoxicity in cortical cell culture. J. Neurosci. 7, 357-368. Choi, D. W., Peters, S., and Viseskul, V. (198713). Dextrorphan levorphanol selectively blocks N-methyl-o-aspartate receptormediated neurotoxicity in cortical neurons. J. Pharmacol. Exp. Ther. 242, 713-720. Clifford, D. B., Zorumski, C. F., and Olney, J. W. (1989). Ketamine and MK-801 prevent degeneration of thalamic neurons induced by focal cortical seizures. Exp. Neural. 705, 272-279. Dykens, J. A., Stern, A., and Trenkner, E. (1987). Mechanism kainate toxicity to cerebellar neurons in vitro is analogous reperfusion tissue injury. J. Neurochem. 49, 1222-1228. Faden, A. I., and Simon, R. P. (1988). A potential citotoxins in the pathophysiology of spinal cord Neural. 23, 623-626.
of to
role for exinjury. Ann.
Blockade ischemic
of N-methyl-o-aspartate damage in the brain.
receptors may protect Science 226, 850-852.
against
Tecoma, E. S., Monyer, H., Goldberg, M. P., and Choi, D. W. (1989). Traumatic neuronal injury in vitro is attenuated by NMDA antagonists. Neuron 2, 1541-1545. Van Kuijk, F. J. G. M., Sevanian, A., Handelman, G. J., and Dratz, E. A. (1987). A new role for phospholipase A2: protection of membranes from lipid peroxidation damage. Trends Brol. Sci. 12, 31-34. Watson, B. D., Busto, R., Goldberg, W. J., Santiso, M., Yoshida, S., and Ginsberg, M. D. (1984). Lipid peroxidation in vivo induced by reversible global ischemia in rat brain. J. Neurochem. 42, 268-274. Weiss, J., Goldberg, M. t?, and Choi, D. W. (1986). Ketamine protects cultured neocortical neurons from hypoxic injury. Brain Res. 280, 186-190.
Furshpan, E. J., and Potter, D. D. (1989). Seizure-like activity and cellular damage in rat hippocampal neurons in cell culture. Neuron 3, 199-207.
Wieloch, prevented 681-683.
Garthwaite, J., Charles, S. L., and Chess-Williams, R. (1988). Endothelium-derived relaxing factor release on activation of NMDA receptors suggests role as intercellular messenger in the brain. Nature 336, 385-387.
Williams, J. H., Errington, M. L., Lynch, M. A., and Bliss, 1. V. (1989). Arachidonic acid induces a long-term activity-dependent enhancement of synaptic transmission in the hippocampus, Nature 341, 739-742.
Goldberg, M. P., Weiss, J. W., Pham, P C., and Choi, D. W. (1987). N-methyl-o-aspartate receptors mediate hypoxic neuronal injury in cortical culture. J. Pharmacol. Exp. Ther. 243, 784-791. Hall, E. D. (1988). Effects of the 21-aminosteroid U74006F on posttraumatic spinal cord ischemia in cats. J. Neurosurg. 68,462-465. Hall, E. D., and Yonkers, cerebral hypoperfusion 19, 340-344.
P. A. (1988). Attenuation by the 21.aminosteroid
of postischemic U74006F. Stroke
Hall, E. D., McCall, J. M., Chase, R. L. Yonkers, P. A., and Braughler, J. M. (1987). A non-glucocorticoid steroid analog of methylprednisolone duplicates its high dose pharmacology in models of CNS trauma and neuronal membrane damage. J. Pharmacol. Exp. Ther. 242, 137-142. Halliwell, B. (1987). Oxidantsand cepts. FASEB J. 7, 358-364. Halliwell, B., and Gutteridge, the nervous system. Trends
human
disease:
some newcon-
J. M. C. (1985). Oxygen Neurosci. 8, 22-26.
radicals and
Hartley, D. M., and Choi, D. W. (1989). Delayed rescueof N-methyl-o-aspartate receptor-mediated neuronal injury in cortical culture. J. Pharmacol. Exp. Ther. 250, 752-758. 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. Meth. 20,83-90. Monyer, H., and Choi, D. W. (1988). Morphinans attenuate cortical neuronal injury induced by glucose deprivation in vitro. Brain Res. 446, 144-148. Murphy, T. H., Miyamoto, M., Sastre, A., Schnaar, R. L., and Coyle, J. T. (1989). Glutamate toxicity in a neuronal cell line involves inhibition of cystine transport leading to oxidative stress. Neuron 2, 1547-1558. Pellegrini-Ciampietro, D. E., Cherici, G., Alesiani, M., Carla, V., and Moroni, F. (1988). Excitatory amino acid release from rat hippocampal slices as a consequence of free-radical formation. J. Neurochem. 51, 1960-1963. Rothman, S. (1984). Synaptic release of excitatory neurotransmitter mediates anoxic neuronal death. 4, 1884-1891.
amino acid J. Neurosci.
Rothman, S. M., and Olney, j. W. (1987). Excitotoxicity NMDA receptor. Trends Neurosci. IO, 299-302. Siesjo, B. K. (1989). Free radicals Brain Metab. Rev. 7, 165-211.
and the
and brain damage. Cerebrovasc.
Siman, R., and Noszek, J. C. (1988). Excitatory amino acids activate calpain I and induce structural protein breakdown in vivo. Neuron 1. 279-287. Simon,
R. P, Swan, J. H., Griffiths,
T., and Meldrum,
B. S. (1984).
T (1985). Hypoglycemia-induced neuronal damage by an N-methyl-o-aspartate antagonist. Science 230,
Vol. 5, 121-126, August, 1990, CopyrIght
0 1990 by Cell Press
2kAminosteroids Attenuate Excitotoxic Neuronai Injury in Cortical Cell Cultures Hannelore Monyer: Dean M. Hartley, and Dennis W. Choi Department of Neurology Stanford University Medical Center Stanford, California 94305
Summary We studied the protective efficacy of novel 21-aminosteroids against several forms of neuronal injury in murine cortical cell cultures. Concentrations of 200 nM to 20 PM partially attenuated the damage induced by glucose deprivation, combined oxygen-glucose deprivation, or exposure to NMDA; maximal protection was less than that produced by NMDA antagonists, but the combination of a 2%aminosteroid plus an NMDA antagonist produced a greater benefit than either drug alone. 2lAminosteroid addition did not attenuate NMDA-induced whole-cell current, but did block almost all of the damage induced by exposure to iron, a protective action consistent with inhibition of free radical-mediated lipid peroxidation. Lipid peroxidation may be a downstream event mediating a portion of the injury triggered by excess stimulation of NMDA receptors. Introduction Free radical formation and subsequent lipid peroxidation has been postulated to participate broadly in the pathogenesis of tissue injury, including specifically the brain injury induced by hypoxia or trauma (Watson et al., 1984; Halliwell and Gutteridge, 1985; Halliwell, 1987; Braughler and Hall, 1989; Siesj& 1989). Other recent studies have suggested that the toxic overstimulation of postsynaptic glutamate receptors, in particular the N-methyl+aspartate (NMDA) subtype, may be an important mechanism of neuronal injury following several types of acute insults (Rothman and Olney, 1987; Choi, 1988), including hypoxia (Rothman, 1984; Weiss et al., 1986), glucose deprivation (Wieloch, 1985), ischemia (Simon et al., 1984), prolonged seizures (Clifford et al., 1989; Furshpan and Potter, 1989), and trauma (Faden and Simon, 1988; Tecoma et al., 1989). The relationship between free radical-induced injury and glutamate neurotoxicity is presently undefined. Although these two processes could be entirely separate, it is attractive to consider that free radical formation might be a direct consequence of glutamate receptor overstimulation and an important downstream mediator of glutamate-induced neuronal death
* Present address: Laboratory of Molecular Neuroendocrinology, ZMBH, University of Heidelberg, INF 282, D-6900 Heidelberg, Federal Republic of Germany.
(Choi, 1988,199O). First, glutamate neurotoxicity in cortical cell cultures may be initially triggered by an influx of extracellular calcium; the resultant buildup of intracellular calcium could activate phospholipase AZ, leading to the liberation of arachidonic acid and subsequent free radical production (Chan et al., 1985). Second, calcium may also trigger the conversion of xanthine dehydrogenase to xanthine oxidase, a rich enzymatic source of free radicals (SiesjG, 1989). Of note, kainate-induced injury of cerebellar neurons can be attenuated by allopurinol, an inhibitor of xanthine oxidase (Dykens et al., 1987). Third, stimulation of NMDA receptors can lead to the release of nitric oxide (Garthwaite et al., 1988), which can react with superoxide to form peroxynitrite and ultimately promote the production of hydroxyl radicals (Beckman et al., 1990). Finally, Murphy and Coyle have suggested that glutamate exposure may produce cytotoxicity in a neuronal cell line because of inhibition of cystine uptake, resulting in reduced glutathione production and increased oxidative stress (Murphy et al., 1989). Recently, a novel series of 21-aminosteroid compounds that are highly lipophilic and potently inhibit lipid peroxidation in isolated membrane preparations have been developed (Braughler et al., 1987; Hall et al., 1987). 21Aminosteroids reduce neuronal damage in several animal models of acute CNS injury (Hall, 1988; Hall and Yonkers, 1988). Two questions arise from these important studies. First, can 21-aminosteroids reduce the intrinsic vulnerability of brain cells to injury? Benefits observed in vivo could reflect indirect effects on systemic metabolism or cerebral blood flow. Second, if these drugs can act directly on brain cells, is improved neuronal survival related to a reduction of glutamate neurotoxicity? We decided to examine the ability of 2%aminosteroids to protect murine cortical neurons in primary cell culture against several forms of injury previously found to be mediated by NMDA receptors. Results As previously reported, mixed cortical cell cultures exposed to a 6-8 hr period of glucose deprivation exhibited acute neuronal swelling; over the next day, many neurons, but not glia, disintegrated (Monyer and Choi, 1988). Addition of the 2%aminosteroid U74QO6F at concentrations of 200 nM to 20 PM to sister cultures prior to the insult produced partial reduction in this morphologically apparent neuronal injury; protection was maximal at 2-5 PM concentrations of U74006E A slight suggestion of neuronal toxicity was seen at 20 FM (granularity of cell bodies), and concentrations higher than 20 PM precipitated out. The effect of U74006F on neuronal injury could be assessed quantitatively by measuring the efflux of the cytosolic enzyme lactate dehydrogenase (LDH) to the bathing medium. Maxi-
NetJrWl
122
CTRL
0
U74500A
OX
120
100
60 3 9
60
P 5
40
20
0 CTRL
DX
U74500A
DX+ U74500A
Figure 2. U745OOA Attenuates Neuronal Degeneration by Combined Oxygen and Glucose Deprivation
Figurel. Comparison of Attenuation by Dextrorphan and 21. Aminosteroids of Glucose Deprivation-Wfuced Neuronal Injury Sister cultures were exposed to glucose deprivation alone (CTRL), in the presence of 100 P M dextrorphan (DX), or in the presence of the indicated concentrations of U74006F (A), 2 u M U745OOA (B), or 2 u M U75412E (C). Bars show LDH efflux the following day (mean f SEM, n = 3-4 cultures per point), expressed relative to that found in the no drug control (= 100). LDH levels in the presence of dextrorphan in (A) and (C)were actually below the sham wash value (-3.0 in [A], - 7.5 in [Cl), but are shown at 0 for ease of display. An asterisk indicates difference from control at P < 0.05, ANOVA and Student-Newman-Keuls’ test.
mal protection varied considerably from experiment to experiment (range of injury reduction, 15%-70%, mean = 52.6% + 22.7% SD, n = 9 experiments), but was always less than the 80%-100% reduction produced in sister cultures by 100 P M dextrorphan or another NMDA antagonist (Figure IA). Similar partial protection against glucose deprivation-induced neuronal injury was produced by optimal concentrations (2 PM) of the related 21-aminosteroids, U745OOA (Figure IB) and U75412E (Figure IC).
Induced
(A) Sister cultures were exposed to combined oxygen and glucose deprivation for 50 min either alone (CTRL), or in the presence of 4 u M U745OOA or 100 u M dextrorphan (DXi. Bars here and in (B) show LDH efflux the following day (mean f SEM, n = 4-5 cultures per point). (B) Pooled data from 4 experiments using 100 u M dextrorphan and 4 u M U745OOA; the period of combined oxygen and glucose deprivation was extended to 1.5 hr to overcome partially the protective effect of dextrorphan (n : 19-23 cultures per point). The condition with both dextrorphan and U745OOA was also significantly different from that with either dextrorphan or U745OOA alone. An asterisk indicates difference from control at P < 0.05, ANOVA and Student-Newman-Keuls’ test.
The simultaneous removal of both oxygen and glucose produces widespread neuronal damage with exposure times of only30-60 min (Goldberg et al., 1988, Sot. Neurosci., abstract). Addition of 4 P M U745OOA prior to insult also produced partial attenuation of this more fulminant form of injury (Figure 2A; 5 out of 6 experiments). If the duration of oxygen and glucose deprivation was increased to 90 min, the protective action of either 100 F M dextrorphan or 4 P M U745OOA was largely overcome, but the combination of the two still produced some benefit (Figure 2B).
;:;Aminosteroids
Attenuate
Excitotoxicity
100
100
80 60
*
z 2
60
g
40
60
40
20
5 CTRL Figure 3. 21Amino:teroids ronal Injury
DX Effectively
U74500A
Block
U74006F
Iron-Induced
Neu-
Cultures’were exp&& fd; 2i hi to 50 PM FeZi and 50 PM Fe3+ alone (GIRL), in the presetice of 100 vth;l dextroibhan, or in the presence of a ~JIM concenlration of the indicated 2%aminosterqid. Bars &ow LDH at the end of the exposure period (mean + SEW, n, = 4). An asterisk indicates difference from control at $2 6.05, ANO%.and Studeht-Newman-Keuls’ test.
The hasi for 2%aminosteroid protection in vivo I$ been suggested to be inhibition of lipid peroxidation, possibly in part related to free radical scavenging (see above). To examine the ability of these compounds to t&ck free radical-mediated injury in our system, we tested them against the damage directly evoked by adding iron ions, a maneuver expected to enhan&a both hydroFyl,radical formation and lipid peroxidation, (,$ndweon and Means, 1985; Aust et al., 1985). The addition of a combination of 50 PM Fe2+ and 50 &vt. Fq3$$o the cultures resulted in widespread neuronql degeneration after 24 hr; glia were grossly intact. This probable peroxidative injury could be almost completely blocked by concurrent addition of 4 PM U745OOA or U74006F (Figure 3). In contrast, less than half of this injury was prevented by 100 PM dextrorphan (Figure 3; 3 of 3 experiments). The observed ability of 21-aminosteroids to attenuate forms of neuronal injury sensitive to NMDA antag onists is consistent with a substantial role for free radicals in mediating NMDA receptor-induced neurotoxicity. To test this idea directly, we examined the ability of U745OOA to reduce the,damage induced by the exogenous application of defined concentrations of NMDA. As previously reported, the widespread neuronal degeneration induced by brief exposure to
NHDA
N?4DA+U74500
NMDA
Figure 4. 21Aminosteroids Neurotoxicity
Reduce
NMDA
Receptor-Mediated
Cultures were exposed for 3 rni; to,!+0 PM NMDA, and LDH was measured the following day (mean + &M, n = 4). Test cultures were treated with 100 PM dextrorphan or 5 &I U745OOA as indicated, using one of three protocols for drug addition: preadded 2 hr before toxic exposure; acute-added during the 3 min toxic exposure only; or ppst-added immediately following toxic exposure and left in until assessment the next day. The difference between the conditions POST DX + U745OOA and POST DX was not statistically significant in this single experiment, but did reach statistical significance when the data were pooled with two other similar, experiments (POST DX + ,W745OOA, 18.8 it 1.8 [n = 121; ,POSJ DX, 32.0 f 2.6 [,n 7 II]; different ai’p < 0.05, ANOVA and Student-Neuman-Keuls’ test):
500 PM NMDA could be blocked almpst cgm.pletely by coapplica,t,ion of 100 pw dextrorphan (Choi et al., 1987b) and partially blocked by the late application of dextrorphan immediately following completion of glutamate exposure (Hartley and Choi, 1989; Figure 4). Preapplication of 5 PM U745OOA (added 2 hr before NMDA) partially attenuated this injury;.a small protective effect could still be detected even when U7!5OOA was added immediately.after completio,n of the,NMDA exposure(Figure 4). TJe combination of both dextrorphan and U745OOA, addeq.la!e after NMQA exposure, produced greater survival than either dru,g,alone (Figure 4). Similar r$sults were obtained when,/njury was, induced with $00 PM glutamate instead of NMDA (2 experiments; data pot shown). The ability pf u745OOA to attenuate thk neurotoxicity of exogenously added NMDA or glutamate was not due to direct antagonism of NMDA receptors. The whole-cell current evoked by application of 10 PM NMDA was not reduced by concomitant,application of 10 PM U745OOA, but was expectedly blocked by the
Figure 5. U745OOA Does Not Block NMDAInduced Whole-Cell Current
NMDA+DX
v v -I30
l 1.0 set
nA
Whole-cell recording of the membrane current response induced by 10 uM NMDA at a holding potential of -60 mV. The response is not a&red by inclusion of 20 vi U74500A in the NMDA solution, but is blocked by inclusion of 30 PM dextrorphan (DX).
Neuron 124
concomitant ure 5).
application
of 20 P M dextrorphan
(Fig-
Discussion Low micromolar concentrations of several 2%aminosteroids partially protected cultured cortical neurons against damage induced by either glucose deprivation, or combined oxygen and glucose deprivation. These observations suggest that 2%aminosteroids can act directlyon brain cells to reduce neuronal vulnerability to injury and therefore support the idea that a parenchymal action underlies the reported ability of the drugs to reduce brain damage in vivo. Since in the same experiments, the selective NMDA antagonist dextrorphan provided almost complete neuronal protection, 2%aminosteroids might work, at least in part, by attenuating NMDA receptor-induced injury mechanisms. Alternatively, 21-aminosteroids could block events unrelated to NMDA receptors, but triggered by the same initial insults and required for cell death under the tested conditions. These two possibilities are not mutually exclusive, and indeed our observations provide support for both. The possibility that 21-aminosteroids can attenuate NMDA receptor-induced injury is supported by the finding that U745OOA produced a similar partial attenuation of the neuronal injury induced by the exogenous application of glutamate or NMDA. This injury attenuation was greatest when cultures were preincubated in the drug prior to excitotoxic exposure, perhaps reflecting improved opportunity to gain access to membrane lipids or intracellular compartments. Since U745OOA did not reduce NMDA receptor-induced whole-cell currents, it is not an NMDA antagonist. Most likely, U745OOA interferes with events linking NMDA receptor overstimulation to subsequent cell death. W e have proposed a speculative breakdown of glutamate receptor-mediated neurotoxicity into three stages: first, induction-overstimulation of membrane receptors leading to a set of immediate cellular derangements; second, amplification-events that intensify these derangements and promote the excitotoxic involvement of additional neurons; and third, expression-the destructive cascades directly responsible for neuronal cell degeneration (Choi, 1990). While additional studies will be needed to define the steps that may be modified by 2+aminosteroids, a reasonable working hypothesis is that the ability of U745OOA to attenuate NMDA receptor-mediated injury is largely related to its documented activity as an inhibitor of lipid peroxidation. This hypothesis is supported by our finding that U745OOA, as well as U74006F, was highly effective in blocking the damage triggered by addition of Fe2+/Fe3+ to the same cultures (see below). The ability of dextrorphan to produce partial injury attenuation in this paradigm does not argue against primary involvement of free radicals; several different
types of initial injury may be compounded by triggering the excess release of endogenous glutamate stores, leading to secondary NMDA receptor-mediated damage. Chan and colleagues have found that cerebellar neurons cultured from transgenic mice with increased levels of superoxide dismutase, an enzyme involved in the detoxification of superoxide radicals, are resistant to both oxidative stress and glutamate toxicity (Chan et al., 1989, Sot. Neurosci., abstract; also meeting communication). Free radicals may play a role in excitotoxic amplification, as they can enhance the release of glutamate from rat hippocampal slices independent of cell lysis (Pellegrini-Giampietro et al., 1988). Continued activation of glutamate receptors by release of endogenous glutamate from injured neurons may constitute an important positive-feedback loop, potentiating the damage caused by initial excitotoxic exposure; the protective effect of late dextrorphan addition may be due to interruption of this loop (Hartley and Choi, 1989). Furthermore, it is possible that peroxidized lipids may be a preferred substrate for phospholipase A2 (Van Kuijk et al., 1987). Thus lipid peroxidation could enhance the formation of arachidonic acid, which may diffuse back to presynaptic terminals and upregulate presynaptic glutamate release (Williams et al., 1989). The fact that U745OOA produced an additive benefit when combined with delayed dextrorphan (Figure 4, POST DX + U745OOA) suggests that its action is probably not restricted to attenuating glutamate release. Free radicals are attractive candidates for mediating the final injury cascades involved in the expression of excitotoxicity. Besides damaging membranes by lipid peroxidation, free radicals can damage proteins and DNA; hydroxyl radicals, formed from superoxide anions and hydrogen peroxide in the presence of transitional metals such as iron, may be particularly destructive (Siesjo, 1989). Free radicals may also participate in other injury mechanisms unrelated to NMDA receptors; it is noteworthy that, in the setting of prolonged combined oxygen and glucose deprivation, U745OOA produced some benefit in addition to that produced by continuous high concentrations of dextrorphan (Figure 2B). NMDA receptor-induced excitotoxic injury and free radical-mediated injury thus may be events that overlap substantially, but not completely. A major portion of NMDA receptor-induced injury is resistant to lipid peroxidation inhibitors such as 21-aminosteroids and could be mediated by other cytotoxic events, such as the activation of catabolic enzymes (Siman and Noszek, 1988; Choi, 1988). Conversely, free radicals may be formed by events unrelated to NMDA receptors. The combination of an NMDA antagonist with approaches directed against free radical injury mechanisms could be a useful strategy for reducing certain forms of acute brain damage.
:;-$minosteroids
Experimental
Attenuate
Excitotoxicity
Procedures
Cortical Cell Culture Mixed cortical cell cultures, containing both neuronal and glial elements, were prepared generally as previously described (Choi et al., 1987a) from fetal mice at 14-18 days gestation. Dissociatcdcortical cells were plated in Primaria (Falcon) 15 m m multiwell vessels (about 2.8 x IO5 cells per well) in Eagle’s minimal essential medium (MEM without Earle’s salts, supplied glutamine-free) supplemented with 10% heat-inactivated horse serum, 10% fetal bovine serum, glutamine (2 mM), and glucose (total 21 mM). Cultures were kept at 3PC in a humidified, CO*containing atmosphere. After 5-12 days in vitro, nonneuronal cell division was halted by l-3 days of exposure to 10e5 M cytosine arabinoside, and the cells were shifted into a maintenance medium identical to the plating medium, but lacking fetal serum. Subsequent medium replacement was carried out twice per week. Only mature (15-20 days in vitro) cortical cultures were selected for study; whenever possible, comparisons were made on matched sister cultures derived from a single plating. Glucose and Combined Oxygen-Glucose Deprivation Glucose deprivation was carried out by exchanging the culture medium for a defined solution resembling MEM, but lacking glucose and vitamins. Cultures were then placed back in the CO* incubator for 6-8 hr. Glucose deprivation was terminated by adding a small amount of concentrated glucose to the exposure solution (final concentration 5.5 mM), and cultures were returned to the incubator overnight prior to assessing injury. In cultures in which oxygen and glucose deprivation were combined, the culture medium was rapidly exchanged with deoxygenated, bicarbonate-buffered salt solution (02 partial pressure < 1 m m Hg), identical to Earle’s balanced salt solution except for the lack of glucose. Cultures were kept for 50-60 min at 37OC in a humidified atmosphere of Nz with 5% COZ (some H2 was added to remove any residual OZ via palladium catalyst). This insult was terminated by the thorough washout of the exposure medium with oxygenated MEM containing 5.5 m M glucose, and cultures were returned to a normoxic, humidified incubator (3PC, 5% COZ) overnight prior to assessing injury. Excitatory Amino Acid Exposure Exposure to glutamate or NMDA was carried out at room temperature in a HEPES-buffered control salt solution (HCSS), substituted for culture medium by triple exchange; HCSS had the following composition: 120 m M NaCI, 5.4 m M KCI, 0.8 m M MgC&, 1.8 m M CaC12, 20 m M HEPES (pH %4 at 25”C), 15 m M glucose. After 3 min, the neurotoxin was quickly washed out (>7000fold dilution) and replaced with another defined solution, Eagle’s MEM with augmented glucose (21 mM), prior to returning the cultures to the 37°C incubator. Neuroprotective drugs were added 2 hr before this excitotoxic exposure (pretreatment), during this exposure (acute treatment), or immediately after this exposure (posttreatment); in the latter case, the drug remained in the medium until the next day, when neuronal damage was assessed. Control experiments showed that little or no cortical cell damage was produced by this protocol if excitatory amino acids were omitted from the exposure solution. Monitoring with phenol red verified that the pH was not appreciably altered by metabolite deprivation or excitotoxic exposure. Assessment of Overall Neuronal Cell Injury Overall neuronal cell injury was estimated in all experiments by examination of cultures with phase-contrast microscopy at 100x to 200x. This examination was usually performed 1 day after initiation of metabolite deprivation or excitotoxic insult, at which point the process of cell death was largely complete; previous experience has suggested that such injury can be reliably estimated in this fashion. In some experiments, this examination was verified by subsequent bright-field examination of trypan blue staining (04% for 5 min), a dye staining debris and nonviable cells.
Overall neuronal cell injury was also quantitatively assessed by the measurement of LDH, released by damaged or destroyed cells, in the extracellular fluid 1 day after the experiment (Koh and Choi, 1987). A small amount of LDH was always present in the media of cultures exposed to sham wash alone. This background amount, determined on sister cultures within each experiment, was subtracted from values obtained in treated cultures. Control experiments have shown that the specific efflux of LDH induced by glutamate exposure or hypoxia is linearly proportional to the number of neurons damaged or destroyed (Koh and Choi, 1987; Goldberg et al., 1987). Electrophysiology Dishes were maintained on the stage of an inverted Nikon Diaphot microscope at room temperature. We recorded from single neurons with patch-clamp technique in the whole-cell configuration. Holding potential was --60 mV. The pipette solution was 140 m M CsCI, 0.5 m M CaC12, 10 m M HEPES, 5 m M EGTA. The perfusion solution contained 125 m M NaCI, 2.8 m M KCI, 1 m M CaC&, and 10 m M HEPES; 1 PM tetrodotoxin and 5 PM glytine were also included. Drug solutions were applied to cells by gravity feed through a hydraulic switch and a large bore pipette. Materials and Reagents Glutamate and NMDA were obtained from Sigma, 21-aminosteroids from Upjohn, and dextrorphan from Hoffman La-Roche. 21Aminosteroids were dissolved initially in dimethylsulfoxide at 10 m M and later diluted to the desired working concentration. Control experiments established that the concentrations of dimethylsulfoxide used did not affect neuronal injury in the tested paradigms. Acknowledgments This work was supported by National Institutes of Health grants NS 12151 and 26907 and by a grant from the American Paralysis Association. H. M. was supported by a fellowship from the Deutsche Forschungsgemeinschaft. D. M. H. was supported by a fellowship from the National Institute of Mental Health (MH 09823). Received
April
4, 1990; revised
June 7, 1990.
References Anderson, D. K., and Means, E. D. (1985). Iron-induced lipid peroxidation in spinal cord: protection with mannitol and methylprednisolone. J. Free Radic. Biol. Med. 1, 59-64. Aust, S. D., Morehouse, L. A., and Thomas, C. E. (1985). Role of metals in oxygen radical reactions. J. Free Radic. Biol. Med. 7, 3-25. Beckman, J. S., Beckman, T. W., Chen, J., Marshall, F’. A., and Freeman, B. A. (1990). Apparent hydroxyl radical production by peroxynitrite: implications for endothelial injury from nitric oxide and superoxide. Proc. Natl. Acad. Sci. USA 87, 1620-1624. Braughler, J. M., and Hall, E. D. (1989). Central nervous system trauma and stroke. I. Biochemical considerations for oxygen radical formation and lipid peroxidatlon. J. Free Radic. Biol. Med. 6, 289-301. Braughler, J. M., Pregenzer, J. F., Chase, R. L., Duncan, L. A., Jacobsen, E. J., and McCall, J. M. (1987). Novel 21-amino steroids as potent inhibitors of iron-dependent lipid peroxidation. J. Biol. Chem. 262, 10438-10440. Chan, P. H., Fishman, R. A., Longar, S., Chen, S., and Yu, A. (1985). Cellular and molecular effects of polyunsaturated fatty acids in brain ischemia and injury. Prog. Brain Res. 63, 227-235. Choi, D. W. (1988). Glutamate neurotoxicity nervous system. Neuron 7, 623-634.
and diseases
of the
Choi, D. W. (1990). Methods for antagonizing glutamate rotoxicity. Cerebrovasc. Brain Metab. Rev. 2, 105-147,
neu-
Neuron 126
Choi, D. W., Maulucci-Gedde, M. A., and Kriegstein, A. R. (1987a). Glutamate neurotoxicity in cortical cell culture. J. Neurosci. 7, 357-368. Choi, D. W., Peters, S., and Viseskul, V. (198713). Dextrorphan levorphanol selectively blocks N-methyl-o-aspartate receptormediated neurotoxicity in cortical neurons. J. Pharmacol. Exp. Ther. 242, 713-720. Clifford, D. B., Zorumski, C. F., and Olney, J. W. (1989). Ketamine and MK-801 prevent degeneration of thalamic neurons induced by focal cortical seizures. Exp. Neural. 705, 272-279. Dykens, J. A., Stern, A., and Trenkner, E. (1987). Mechanism kainate toxicity to cerebellar neurons in vitro is analogous reperfusion tissue injury. J. Neurochem. 49, 1222-1228. Faden, A. I., and Simon, R. P. (1988). A potential citotoxins in the pathophysiology of spinal cord Neural. 23, 623-626.
of to
role for exinjury. Ann.
Blockade ischemic
of N-methyl-o-aspartate damage in the brain.
receptors may protect Science 226, 850-852.
against
Tecoma, E. S., Monyer, H., Goldberg, M. P., and Choi, D. W. (1989). Traumatic neuronal injury in vitro is attenuated by NMDA antagonists. Neuron 2, 1541-1545. Van Kuijk, F. J. G. M., Sevanian, A., Handelman, G. J., and Dratz, E. A. (1987). A new role for phospholipase A2: protection of membranes from lipid peroxidation damage. Trends Brol. Sci. 12, 31-34. Watson, B. D., Busto, R., Goldberg, W. J., Santiso, M., Yoshida, S., and Ginsberg, M. D. (1984). Lipid peroxidation in vivo induced by reversible global ischemia in rat brain. J. Neurochem. 42, 268-274. Weiss, J., Goldberg, M. t?, and Choi, D. W. (1986). Ketamine protects cultured neocortical neurons from hypoxic injury. Brain Res. 280, 186-190.
Furshpan, E. J., and Potter, D. D. (1989). Seizure-like activity and cellular damage in rat hippocampal neurons in cell culture. Neuron 3, 199-207.
Wieloch, prevented 681-683.
Garthwaite, J., Charles, S. L., and Chess-Williams, R. (1988). Endothelium-derived relaxing factor release on activation of NMDA receptors suggests role as intercellular messenger in the brain. Nature 336, 385-387.
Williams, J. H., Errington, M. L., Lynch, M. A., and Bliss, 1. V. (1989). Arachidonic acid induces a long-term activity-dependent enhancement of synaptic transmission in the hippocampus, Nature 341, 739-742.
Goldberg, M. P., Weiss, J. W., Pham, P C., and Choi, D. W. (1987). N-methyl-o-aspartate receptors mediate hypoxic neuronal injury in cortical culture. J. Pharmacol. Exp. Ther. 243, 784-791. Hall, E. D. (1988). Effects of the 21-aminosteroid U74006F on posttraumatic spinal cord ischemia in cats. J. Neurosurg. 68,462-465. Hall, E. D., and Yonkers, cerebral hypoperfusion 19, 340-344.
P. A. (1988). Attenuation by the 21.aminosteroid
of postischemic U74006F. Stroke
Hall, E. D., McCall, J. M., Chase, R. L. Yonkers, P. A., and Braughler, J. M. (1987). A non-glucocorticoid steroid analog of methylprednisolone duplicates its high dose pharmacology in models of CNS trauma and neuronal membrane damage. J. Pharmacol. Exp. Ther. 242, 137-142. Halliwell, B. (1987). Oxidantsand cepts. FASEB J. 7, 358-364. Halliwell, B., and Gutteridge, the nervous system. Trends
human
disease:
some newcon-
J. M. C. (1985). Oxygen Neurosci. 8, 22-26.
radicals and
Hartley, D. M., and Choi, D. W. (1989). Delayed rescueof N-methyl-o-aspartate receptor-mediated neuronal injury in cortical culture. J. Pharmacol. Exp. Ther. 250, 752-758. 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. Meth. 20,83-90. Monyer, H., and Choi, D. W. (1988). Morphinans attenuate cortical neuronal injury induced by glucose deprivation in vitro. Brain Res. 446, 144-148. Murphy, T. H., Miyamoto, M., Sastre, A., Schnaar, R. L., and Coyle, J. T. (1989). Glutamate toxicity in a neuronal cell line involves inhibition of cystine transport leading to oxidative stress. Neuron 2, 1547-1558. Pellegrini-Ciampietro, D. E., Cherici, G., Alesiani, M., Carla, V., and Moroni, F. (1988). Excitatory amino acid release from rat hippocampal slices as a consequence of free-radical formation. J. Neurochem. 51, 1960-1963. Rothman, S. (1984). Synaptic release of excitatory neurotransmitter mediates anoxic neuronal death. 4, 1884-1891.
amino acid J. Neurosci.
Rothman, S. M., and Olney, j. W. (1987). Excitotoxicity NMDA receptor. Trends Neurosci. IO, 299-302. Siesjo, B. K. (1989). Free radicals Brain Metab. Rev. 7, 165-211.
and the
and brain damage. Cerebrovasc.
Siman, R., and Noszek, J. C. (1988). Excitatory amino acids activate calpain I and induce structural protein breakdown in vivo. Neuron 1. 279-287. Simon,
R. P, Swan, J. H., Griffiths,
T., and Meldrum,
B. S. (1984).
T (1985). Hypoglycemia-induced neuronal damage by an N-methyl-o-aspartate antagonist. Science 230,
