Ca2+-dependent mechanisms of cell injury in cultured cortical neurons

Pergamon

PII:

Neuroscience Vol. 86, No. 4, pp. 1133–1144, 1998 Copyright  1998 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/98 $19.00+0.00 S0306-4522(98)00070-0

Ca2+-DEPENDENT MECHANISMS OF CELL INJURY IN CULTURED CORTICAL NEURONS M. R. CASTILLO and J. R. BABSON* Department of Biomedical Sciences, College of Pharmacy, University of Rhode Island, Kingston, RI 02881, U.S.A. Abstract––The contributions of several Ca2+-dependent processes to neurotoxicity were examined in primary cultures of rat cortical neurons. The Ca2+ ionophore ionomycin induced a rapid loss of axonal morphology and concomitant release of inositol phosphates that preceded morphological alterations of neuronal cell bodies, choline and arachidonate release, and protein degradation. These events were followed by a degree of neuronal lysis proportional to the external Ca2+ concentration and exposure time. The phospholipase inhibitor neomycin decreased both arachidonate release and the phospholipid hydrolysis catalysed by phospholipases C and D. Proteolysis was abated by the protease inhibitor leupeptin, but not by lysosomal proteolysis inhibitors. Neuronal lysis was inhibited partially by either leupeptin or neomycin and almost completely by both in combination. However, neither agent, alone or in combination, affected the morphological derangements. The diacylglycerol lipase inhibitor RHC-80267 reduced arachidonate release, but not neuronal lysis. Phospholipase A2 inhibitors had no effect on either arachidonate release or lysis. Treatment of mixed cultures of neurons and glia with a Ca2+-dependent glutamate challenge caused similar morphological changes and a delayed neuronal lysis that was also diminished by leupeptin and neomycin, but not by inhibitors of lysosomal proteolysis. These data describe several distinct stages of Ca2+-dependent injury to cortical neurons, a key feature of which is the stimulation of protease, and phospholipase C and D activites. The initial stage is characterized by a rapid loss of axonal morphology and increased phosphatidylinositol hydrolysis. An intermediate stage involves changes in cell body morphology plus the degradation of neuronal protein and phosphatidylcholine. In a later stage, the loss of plasma membrane integrity denotes neuronal death.  1998 IBRO. Published by Elsevier Science Ltd. Key words: proteases, phospholipases, Ca2+ ionophore, glutamate.

Perturbations of intracellular Ca2+ ([Ca2+]i) homeostasis that lead to a sustained elevation of [Ca2+]i have frequently been associated with irreversible injury in a variety of cell types, where increased [Ca2+]i is thought to activate hydrolytic intracellular enzymes, such as proteases and phospholipases, to cytotoxic levels.18,27 Elevated [Ca2+]i and cellular injury are also common to several important neuropathological conditions that include ischemia, epilepsy, hypoglycemia,40 Alzheimer’s disease,35 HIV neurotoxicity25 and excitatory amino acid (EAA) neurotoxicity.9,26 Therefore, the degradative action of Ca2+-dependent enzymes on cellular components may also represent a key feature of the neurotoxicity attendant to a number of these important neuropathologies.9,24,30,35,37,38,43

*To whom correspondence should be addressed. Abbreviations: [Ca2+]i, intracellular Ca2+; DAG, diacylglycerol; EAA, excitatory amino acid; HBSS, Hanks balanced salt solution; HEPES, N-2-hydroxyethylpiperazine-N ethanesulfonic acid; LDH, lactate dehydrogenase; OEPC, oleyloxyethylphosphorylcholine; PKC, protein kinase C; PLA2, phospholipase A2; PLC, phospholipase C; PLD, phospholipase D; RHC-80267, 1,6-bis(cyclohexyloximinocarbonylamino)hexane.

Ca2+-dependent proteolysis by non-lysosomal proteases is thought to disrupt links that anchor the cytoskeleton to integral proteins of the plasma membrane, and this may lead to membrane blebbing and ultimately to the loss of plasma membrane integrity.27 The presence of Ca2+-dependent nonlysosomal proteases, often called calpains, has been well established in brain tissue.47 These enzymes show a marked preference for a number of structural components of the neuronal cytoskeleton, including spectrin, microtubule-associated proteins, tubulin and neurofilaments.24,37,41 The degradation of these proteins and the resulting loss of cytoskeletal integrity has been proposed to be critical to neuronal injury accompanying ischemia,24,30 Alzheimer’s disease35 and EAA insults.38 This postulate is supported by studies wherein both ischemic brain damage2,3,21 and the spectrin degradation associated with this pathology21 are decreased by calpain inhibitors. However, the degree to which elevated [Ca2+]i, the proteolysis of key cytoskeletal proteins and cell death are linked specifically in cortical neurons remains to be demonstrated directly. The activation of Ca2+-dependent phospholipases and the loss of plasma membrane integrity resulting from phospholipid hydrolysis may represent another

1133

1134

M. R. Castillo and J. R. Babson

cytotoxic consequence of elevated [Ca2+]i in a variety of cell types,18,27 including neurons.43 To date, however, what little is known about the Ca2+ dependence and neurotoxic potential of the major phospholipases, including phospholipase A2 (PLA2), phospholipase C (PLC) and phospholipase D (PLD), remains controversial. For example, the stimulation of PLA2 by EAAs and ionomycin in striatal neurons appeared to be Ca2+ dependent,5,16 yet PLA2 inhibitors failed to decrease methyl mercury toxicity characterized by elevated [Ca2+]i and PLA2 activation in cerebellar granule cells.44 In another study, PLA2 inhibition was associated with an attenuation of EAA-induced neurotoxicity in hippocampal mixed cultures.34 Receptor-mediated PLC activation was observed with the EAA agonist ibotenic acid in cultured cortical neurons, but was deemed to be neither necessary nor sufficient for neurotoxicity.48 Recently, however, a relationship between elevated [Ca2+]i, phosphotidylinositol breakdown, protease activation and ischemia-induced neuronal death has been proposed.46 Whether or not neurons contain Ca2+-dependent forms of PLC that contribute to neuronal injury remains a question. PLD activity toward phosphatidylcholine, a major neuronal membrane component, has also been observed in neurons,23 though little is known about PLD activation or its neurotoxic contribution under conditions of [Ca2+]i overload. Clearly, additional studies are required to determine the extent to which phospholipases may be activated directly by [Ca2+]i and thereby injure cortical neurons. The present study was designed to gain a clearer understanding of the relationships between elevated Ca2+, the activation of Ca2+-dependent hydrolytic enzymes and lethal injury to cortical neurons. Toward this end, the responses of key hydrolytic activities were determined in cultured rat cortical neurons, after elevating [Ca2+]i directly to cytotoxic levels with ionomycin. Inhibitor studies were performed to assess the neurotoxic contributions of these Ca2+-dependent hydrolytic enzymes. To determine whether results of the ionomycin studies have broader neuropathological implications, inhibitors that reduced ionomycin toxicity were evaluated with mixed neuronal cultures responsive to the receptormediated, Ca2+-dependent toxicity of the EAA glutamate. EXPERIMENTAL PROCEDURES

Neuronal culture The primary culture of cortical neurons was performed using minor modifications on the method of Choi et al.11 Sprague–Dawley rats (15–17 days gestation; Charles River Laboratories) were anesthetized with 45 mg/kg pentobarbital and the embryos removed. The whole cerebral cortex was removed from the embryos, with care to discard the basal ganglia and meninges. The fetal cortices were suspended in Eagle’s minimum essential medium (pH 7.4) containing 21 mM glucose and 2 mM glutamine, and forced through a 16-gauge syringe needle, then a 25-gauge needle,

three times each to dissociate the neurons. Cells were counted and plated as a single-cell suspension (1106 cells/ plate) on 60-mm dishes (Falcon, Primaria) previously coated with rat tail collagen. The collagen was prepared by the method of Dougherty et al.,14 spread on the Petri dishes (0.1 mg/plate), incubated in culture medium for at least 15 min at 37C and rinsed once before plating the cells. The neuronal cultures were then incubated for 60 min, at 37C in a humidified 5% CO2 atmosphere, rinsed once to remove non-adherent cells and placed in supplemented culture medium containing 21 mM glucose, 2 mM glutamine, 26 mM bicarbonate, 5% heat-inactivated horse serum, 5% fetal bovine serum and 100 U/ml penicillin–streptomycin. After three days, fresh medium containing 40 µM cytosine arabinofuranoside was added to eliminate proliferating non-neuronal cells.20 Cultures were used on the sixth day after initial plating, and contained >93% neurons, as assessed by fluorescent immunolabeling with tetanus toxin C fragment using the Neurotag kit supplied by Sigma Chemical Co.31 Mixed neuronal cultures and glial cultures The isolation and culturing was performed as described above, except that the antimetabolite treatment was omitted and the cells were used on the 14th day after the initial plating. These modifications allowed the proliferation of glial cells that promote glutamate receptor expression in cortical neurons. These mixed cultures contained approximately 30% neurons, as estimated according to the method of Ray et al.31 Glial cultures devoid of neurons were isolated from the cortices of neonatal rats using the procedure described above. Experimental protocol Prior to treatment, cultures were rinsed three times with Hanks balanced salt solution (HBSS; 21 mM glucose and 25 mM HEPES, pH 7.4) then exposed to either 4 µM ionomycin, vehicle or 500 µM glutamate. For inhibitor studies, cultures were preincubated for 15 min at 37C in HBSS alone or HBSS containing either: 100 µM leupeptin; 0.1 mM chloroquine or 10 mM methylamine; 2 mM neomycin; 50 µM dibucaine, 100 µM mepacrine, 100 µM RHC80267; 25 µM oleyloxyethylphosphorylcholine (OEPC); 100 µM aurintricarboxylic acid; or 5 µM N,N -diphenyl-pphenylenediamine (DPPD). After 15 min, preincubation solutions were replaced with identical solutions to which either 4 µM ionomycin, vehicle or 500 µM glutamate had been added. Determination of lethal cell injury Lethal cell injury was evaluated on the basis of a loss of plasma membrane integrity. Lactate dehydrogenase (LDH) leakage was used as the primary indicator of cell injury. The validity of LDH leakage as an indicator of injury was assessed by comparison with viability determinations using Trypan Blue exclusion. At the end of each incubation period, the incubate was removed and the adherent cells lysed with 0.5% Triton X-100 in HBSS. LDH activity in both incubate and cell lysate was assayed spectrophotometrically,1 and the extent of LDH leakage expressed as a percentage of the total cellular LDH (incubate activity/ incubate+lysate activities). An analysis of total LDH activity indicated that the treatments used had no effect on total LDH values over time and at the cell densities used in this study (data not shown). Trypan Blue exclusion was performed using a modification of the method of Choi et al.11 Cultures were rinsed twice with HBSS and incubated for 2 min at 37C in membrane-filtered 0.04% Trypan Blue solution, rinsed twice with HBSS and then photographed at

Ca2+-dependent cell injury in cortical neurons four random fields under bright-field illumination. Cell death was expressed as the percentage of cells with stained nuclei. Measurement of intracellular proteolysis Neuronal cultures were preincubated for 24 h in supplemented culture medium containing 0.5 µCi/plate of -[14C]valine, rinsed three times with HBSS and then exposed to the chemical treatments described above. Incubations were terminated by addition of perchloric acid to a final concentration of 10% (w/v) and the samples processed according to the method of Seglen et al.36 Proteolytic activity is based on the amount of acid-soluble Ref. 14C products released into the supernatant and expressed as a percentage of the total incorporated radioactivity (soluble product/soluble product+cell pellet). Measurement of phospholipid hydrolysis Neuronal cultures were preincubated in supplemented culture medium containing one of the following radiolabeled precursors: 1 µCi/plate [2-3H]myo-inositol for 48 h; 0.2 µCi/plate [methyl-14C]choline for 36 h; or 0.2 µCi/plate [3H]arachidonate for 18 h. Cultures were then rinsed three times with HBSS and exposed to various treatments. PLA2 activity was assessed by [3H]arachidonate release.7 Cultures were extracted using the method of Bligh and Dyer,4 and [3H]arachidonate was isolated by thin layer chromatography on Silica G using hexane–diethyl ether–acetic acid (80:20:1.2). To measure PLC and PLD, cultures were lysed by the addition of trichloroacetic acid to a final concentration of 10% (v/v). Cultures were then scraped and the contents centrifuged at 4C at 14000 g for 15 min to separate lysed cell pellets from soluble products. PLC activity was based on the amount of released [3H]inositol phosphates that were isolated by ion exchange chromatography.12 PLD activity was measured as the amount of released [methyl14 C]choline that was extracted according to Bligh and Dyer,4 and then isolated by ion exchange chromatography.13 The hydrolytic activity is presented as the amount of soluble radioactive hydrolysis products expressed as a percentage of the total radioactivity (soluble product/ soluble product+cell pellet). Radioactivity was determined by liquid scintillation counting of the pellet and chromatographed fractions. RESULTS

Ionomycin-induced Ca2+-dependent injury to cultured cortical neurons An initial series of experiments was designed to characterize the Ca2+–ionomycin model of neuronal injury. A 6-h exposure of neuronal cultures to 4 µM ionomycin and a range of extracellular Ca2+ concentrations caused a degree of lethal injury, measured by both LDH leakage and Trypan Blue exclusion, that was clearly dependent on the extracellular Ca2+ concentration (Table 1). LDH leakage and Trypan Blue exclusion values for cultures treated with ionomycin in the absence of added Ca2+ were similar to those of controls (3–7%). In all cases, both indicators of lethal injury gave similar results. Distinct morphological alterations were also observed following exposure to the Ca2+–ionomycin challenge. Approximately 15 min into the challenge, neurons exhibited axonal blebbing that progressed to a state of extensive axonal degeneration by 60 min. By this time, neuronal cell bodies throughout the

1135

Table 1. Concentration-dependent effects of external Ca2+ on ionomycin-induced lethal neuronal injury Cell death (%) 2+

Ca

(mM)

Control 0 1.3 5.0 10.0 Ionomycin 0 1.3 5.0 10.0

LDH leakage

Trypan staining

3.20.1 4.20.3 5.60.2 6.20.4

3.00.8 3.50.7 3.60.5 4.00.2

2.90.1 19.30.4* 24.30.8* 37.60.6*

3.40.9 16.92.1* 22.42.1* 35.10.7*

Neuronal cultures were incubated in HBSS at 37C in the absence or presence of either 4 µM ionomycin or vehicle. HBSS was prepared without Ca2+ or with Ca2+ added to yield the final concentrations indicated. LDH leakage and Trypan Blue exclusion were determined at 6 h. Values are given as a percentage of the total neuronal LDH and are the meanS.E.M. of measurements from three to 13 separate neuronal preparations. *Significantly different from control cultures (P<0.01).

cultures exhibited a mixture of mophological derangements (Fig. 1B, E). A comparison of the onset of these alterations with the LDH leakage timecourse indicates that these morphological changes occurred well before any significant LDH leakage (Fig. 2). Because 4 µM ionomycin and 1.3 mM Ca2+ produced reproducible toxic effects over a reasonable interval, these conditions were used throughout the rest of this study. Inhibition of Ca2+-dependent cortical neuron injury As a first step to examine whether the neurotoxicity observed was due to the overstimulation of Ca2+-activated hydrolytic enzymes, a series of inhibitor studies were conducted under mild conditions, using commonly employed inhibitor concentrations. Ionomycin-treated neurons were not protected by inhibitors of lysosomal proteolysis, chloroquine and methylamine; the PLA2 inhibitors mepacrine, dibucaine and OEPC; the diacylglycerol (DAG) lipase inhibitor RHC-80267; the endonuclease inhibitor aurincarboxylic acid; or the antioxidant DPPD. Protection was observed, however, with the protease inhibitor leupeptin and the phospholipase inhibitor neomycin (Table 2). While the combination of both decreased LDH leakage even further, it had no effect on ionomycin-induced morphological alterations (Fig. 1C, F). To characterize the protective action of leupeptin and neomycin further, neurons were either exposed to a higher ionomycin concentration (Table 2) or challenged for longer (Fig. 2). While increased neurotoxicity resulted from these harsher treatments, LDH leakage was still reduced substantially by either

1136

M. R. Castillo and J. R. Babson

Fig. 1. Morphological alterations to cortical neurons induced by the Ca2+–ionomycin challenge. Following a 15-min preincubation in the absence or presence of inhibitors, neuronal cultures were incubated with vehicle (A, D), 4 µM ionomycin (B, E), or 4 µM ionomycin plus 2 mM neomycin and 100 µM leupeptin (C, F). After 60 min, neurons were photographed at 100 magnification under phase contrast focused on the axons (A–C) and bright field focused on the cell bodies (D–F).

agent, and significantly more protection was observed when these agents were used in combination. The protective action of leupeptin and neomycin was

also assessed by comparing LDH leakage and Trypan Blue exclusion in sister cultures treated with ionomycin and these inhibitors. As shown in Table 3,

Ca2+-dependent cell injury in cortical neurons

1137

Fig. 2. Effect of enzyme inhibitors on Ca2+–ionomycin-induced neuronal injury. Neurons were preincubated for 15 min in modified HBSS in the absence or presence of the inhibitors, then exposed to 4 µM ionomycin alone (filled squares) or 4 µM ionomycin in the presence of 2 mM neomycin (open triangles), 100 µM leupeptin (open squares), or 2 mM neomycin plus 100 µM leupeptin (open circles). LDH leakage was determined at the times indicated. Values are the meanS.E.M. of measurements from three to 10 neuronal preparations. Significantly different (P<0.01) from sister cultures treated with: *ionomycin plus all inhibitor treatments; **ionomycin plus neomycin and neomycin plus leupeptin; or †ionomycin plus neomycin and leupeptin. Table 2. Effect of inhibitors on ionomycin-induced lethal cell injury LDH leakage (% of total) Ionomycin concentration (µM) Treatment None Neomycin (2 mM) Leupeptin (100 µM) Leupeptin (100 µM)+neomycin (2 mM) Mepacrine (100 µM) OEPC (25 µM) Dibucaine (50 µM) RHC-80267 (100 µM) Aurintricarboxylic acid (1 µg/ml) DPPD (5 µM) Chloroquine (100 µM) Methylamine (10 mM)

0

4

12

5.10.5 4.60.8 3.60.2 3.70.3 8.30.4 5.11.4 5.90.8 3.80.2 3.60.5 2.80.2 3.90.3 4.80.1

25.60.9 8.60.4* 10.70.3* 4.70.6* 23.70.8 26.11.2 32.41.2 22.01.4 27.70.7 21.00.3 22.40.2 24.50.3

45.00.4 25.50.4* 29.80.4* 16.20.9*

Neurons were preincubated for 15 min at 37C in HBSS alone or HBSS containing inhibitor. After 15 min, preincubation solutions were replaced with identical solutions to which either 4 or 12 µM ionomycin, or vehicle, had been added. LDH leakage was determined at 6 h. Values are given as a percentage of the total neuronal LDH and are the meanS.E.M. of measurements from three to nine separate neuronal preparations. *Significantly different from sister cultures treated with ionomycin alone (P<0.01).

results with both indices of lethal injury were reasonably similar. As before, each agent was protective alone and in combination they provided additional protection. It should be noted that the surface of some neurons treated with the combination of iono-

mycin and neomycin appeared to be thinly coated with Trypan Blue, but had unstained nuclei, while other cells were coated and had well-stained nuclei. The latter were scored as dead. This phenomenon may have resulted in a modest overestimation of cell

1138

M. R. Castillo and J. R. Babson

Table 3. Comparison of lethal cell injury evaluated by lactate dehydrogenase leakage and Trypan Blue exclusion Percent Cell Death 6h Treatment Ionomycin Control Ionomycin plus leupeptin Leupeptin Ionomycin plus neomycin Neomycin Ionomycin plus leupeptin and neomycin Leupeptin and neomycin

12 h

LDH Leakage

Trypan Staining

LDH Leakage

Trypan Staining

21.70.6 4.20.5 11.61.0 3.80.2 8.60.4 4.60.9 7.91.8 3.80.3

27.61.6 3.60.5 9.61.7 6.60.7 14.01.8 4.90.6 13.52.3 5.10.7

34.80.5 4.50.3 19.11.0 3.00.4 15.50.3 3.70.6 7.30.6 4.71.0

31.71.1 6.20.4 23.72.5 5.00.8 25.32.5 7.21.0 23.22.8 7.90.5

Neuronal cultures were preincubated for 15 min at 37C in HBSS alone or HBSS containing either 100 µM leupeptin, 2 mM neomycin or 100 µM leupeptin plus 2 mM neomycin. After 15 min, preincubation solutions were replaced with identical solutions to which either 4 µM ionomycin or vehicle had been added. LDH leakage and Trypan Blue exclusion were determined at 6 and 12 h. Values are given as a percentage of the total and are the meanS.E.M. of measurements from three to 13 separate neuronal preparations.

Fig. 3. Neuronal proteolysis stimulated by the Ca2+–ionomycin challenge. Neurons were labeled for 24 h with [14C]valine, rinsed, preincubated for 15 min in modified HBSS in the absence or presence of 100 µM leupeptin, and then exposed in modified HBSS to 4 µM ionomycin alone (filled squares), or 4 µM ionomycin in the presence of 100 µM leupeptin (open circles), or vehicle (open squares). The incubations were terminated by acid lysis and the amount of acid-soluble hydrolysis products was determined as described in Experimental Procedures. Hydrolytic activity is based on the release of acid-soluble radioactivity expressed as a percentage of the total (cell pellet+acid-soluble product). Values are the meanS.E.M. of measurements from three separate neuronal preparations. *Significantly different from sister cultures treated with ionomycin alone (P<0.01).

death. Since the inhibitor study results suggested that non-lysosomal protease and phospholipase activation may be vital to onset of Ca2+-dependent neurotoxicity, the response of these hydrolytic activities to elevated [Ca2+]i was examined both in the absence and presence of protective inhibitors.

Ca2+-dependent hydrolytic activities of cortical neurons The stimulation of several hydrolytic activities was observed in ionomycin-treated neurons well before any loss of membrane integrity was detected. As shown in Fig. 3, the Ca2+–ionomycin challenge

Ca2+-dependent cell injury in cortical neurons

1139

Fig. 4. Phosphatidylinositol hydrolysis induced by the Ca2+–ionomycin challenge. Neurons were labeled for 48 h with [3H]inositol, rinsed and preincubated for 15 min in modified HBSS in the absence or presence of 2 mM neomycin, then exposed in modified HBSS to 4 µM ionomycin alone (filled squares), 4 µM ionomycin plus 2 mM neomycin (open squares), or vehicle (open triangles). The incubations were terminated by acid lysis and acid-soluble inositol phosphates were isolated by ion exchange chromatography and measured by liquid scintillation counting, as described in Experimental Procedures. Hydrolytic activity is based on the release of acid-soluble radioactivity expressed as a percentage of the total (acid soluble/cell pellet+acid soluble). Values are the meanS.E.M. of measurements from three separate neuronal preparations. *Significantly different from sister cultures treated with ionomycin alone (P<0.05).

Table 4. Ionomycin-induced release of choline and phosphocholine Total radioactivity (%) Treatment Ionomycin Ionomycin plus neomycin Control

Choline

Phosphocholine

11.10.5 7.60.3* 5.30.3*

13.20.1 13.00.4 12.10.8

Neurons were labeled for 48 h with [14C]choline, rinsed three times with HBSS, preincubated for 15 min in HBSS in the absence or presence of 2 mM neomycin, then exposed to 4 µM ionomycin and HBSS in the absence or presence of 2 mM neomycin. Incubations were terminated at 2 h by acid lysis, and choline and phosphocholine were isolated by ion exchange chromatography as described in Experimental Procedures. Hydrolytic activity is based on the release of soluble radioactivity expressed as a percentage of the total (cell pellet+soluble product). Values are the mean S.E.M. of measurements from three separate neuronal preparations. *Significantly different from sister cultures treated with ionomycin alone (P<0.01).

promoted the proteolytic degradation of neuronal components. The release of acid-soluble radioactivity from challenged neurons radiolabeled previously with [14C]valine was approximately 1.6-fold higher at 2 h than that from unstimulated controls. At this time, LDH leakage from ionomycin-treated cells was no higher than control levels (4–7%). Leupeptin completely prevented the release of radiolabeled products to the extent that levels were actually below those of controls.

The hydrolysis of neuronal phospholipids also increased in response to Ca2+–ionomycin and occurred prior to any measurable loss of membrane integrity. The challenge triggered a rapid release of inositol phosphates, presumably catalysed by PLC, that by 1 h was six-fold higher than controls (Fig. 4). Neomycin reduced inositol phosphate release from challenged cells by 40% after 1 h. By 3 h, the ionomycin-induced release of inositol phosphates had leveled off and the inhibitory effects of neomycin

1140

M. R. Castillo and J. R. Babson

Table 5. Inhibition of ionomycin-induced arachidonate release Treatment Control Ionomycin plus RHC-80267 plus neomycin plus mepacrine

Total radioactivity† (%)

Inhibition (%)

0.50.1 2.30.3 0.80.3* 1.30.2* 2.20.3

— — 83 56 6

3

Neurons were labeled for 24 h with [ H]arachidonate, rinsed three times with HBSS containing 0.1% bovine serum albumin, preincubated for 15 min in modified HBSS in the absence or presence of 100 µM RHC-80267, 100 µM mepacrine or 2 mM neomycin, then exposed to 4 µM ionomycin in HBSS in the absence or presence of inhibitors. At 1 h, neuronal lipids were extracted and separated by thin layer chromatography, as described in the Experimental Procedures section. *Significantly different from sister cultures treated with ionomycin alone (P<0.05). †Values represent the amount of radioactivity that cochromatographed with arachidonate standard given as a percentage of the total extracted radioactivity. Values are the meanS.E.M. of measurements from three to five separate neuronal preparations.

had diminished. There was also evidence for the Ca2+-dependent hydrolysis of phosphatidylcholine. As shown in Table 4, a 2-h challenge caused substantial choline release, but did not alter levels of free phosphocholine. Neomycin reduced the Ca2+dependent choline release by 60%. These data are consistent with the Ca2+-dependent activation of a phosphatidylcholine-specific PLD in the absence of any detectable phosphatidylcholine-specific PLC activity. Ionomycin-induced phospholipid hydrolysis lead to a substantial release of arachidonate (Table 5). The Ca2+–ionomycin challenge triggered a 4.6-fold release of arachidonate at 1 h from neurons previously labeled with [3H]arachidonate. Interestingly, arachidonate release was not decreased by the PLA2 inhibitor mepacrine, but was inhibited substantially by the PLC inhibitor neomycin (56%) and the DAG lipase inhibitor RHC-80267 (83%). An additional series of experiments was conducted to determine whether either leupeptin or neomycin interfered with the ionophoretic action of ionomycin. Such interference would lower the extent of Ca2+dependent protease and PLC response and thus appear as inhibition of and protection from these events. To address this issue, these activities were evaluated in sister cultures exposed to the Ca2+– ionomycin challenge, where protease activity was measured in the presence of neomycin, and PLC activity was determined in the presence of leupeptin. The release of [3H]inositol phosphates at 1 h was essentially the same in both the absence and presence of leupeptin (18.01.0% and 20.71.1%, respectively), indicating that ionomycin-induced PLC activation was unaffected by this agent. Likewise,

neomycin did not significantly alter the rate of proteolysis, since ionomycin-induced [14C]peptide release at 3 h was similar in the absence and presence of the neomycin (10.61.4% and 9.91.2%, respectively). These results demonstrate that protection by leupeptin and neomycin is not due simply to interference with the ionophoretic action of ionomycin by these agents. Comparison of glutamate-induced injury with Ca2+– ionophore-induced injury To determine whether the effects of ionomycin treatment bore any relationship to glutamate neurotoxicty, some Ca2+-dependent components of the ionomycin model were examined in glutamatesensitive neurons. Exposure of mixed neuronal cultures to 500 µM glutamate and 1.3 mM Ca2+ resulted in a 2-fold increase in LDH leakage over that observed for glutamate exposure in the absence of Ca2+ and controls (Table 6). It should be noted that the ratio of cortical neurons to glia is substantially lower in mixed cultures than in the purer neuronal cultures used for ionomycin studies. As a result, the maximum LDH leakage attributable to neurons, measured as a percentage of the total cellular LDH, was lower in the glutamate model. Glutamate effects on neuronal morphology also appeared to share some similarities with those caused by ionomycin (data not shown). The effects of protease and phospholipase inhibition on glutamate-induced neuronal injury were also examined. As with the ionomycin model, glutamate neurotoxicity was reduced substantially by leupeptin, but not by inhibitors of lysosomal proteolysis (Table 6). The situation with the phospholipase inhibitor neomycin was somewhat more complicated. While neomycin alone was non-toxic to purified cortical neurons (Table 3), it was toxic to mixed neuronal cultures (Table 6), suggesting that neomycin was selectively toxic to glia. To examine this possibility, glial cultures devoid of neurons and unresponsive to glutamate were exposed to neomycin. Substantial LDH leakage ensued (Table 6). Furthermore, neomycin toxicity in mixed cultures was not enhanced by a glutamate treatment that killed neurons (Table 6). Thus, neomycin-induced LDH leakage from glia would tend to obscure any neomycin-induced reduction of LDH leakage from glutamate-responsive neurons. However, an overall comparison of the LDH leakage accompanying the various neomycin treatments suggests, albeit indirectly, that neomycin did protect cortical neurons from Ca2+-dependent glutamate toxicity. Taken together, the data indicate that neuronal injury induced by the Ca2+–ionomycin challenge shares some key neurotoxic features with the glutamate challenge. DISCUSSION

The objective of the current study was to examine the putative relationships between elevated [Ca2+]i,

Ca2+-dependent cell injury in cortical neurons

1141

Table 6. Effect of inhibitors on glutamate-induced neurotoxicity LDH leakage (% of total) Treatment Glutamate plus Ca2+ Glutamate minus Ca2+ Control plus Ca2+ Control minus Ca2+ Glutamate plus Ca2+ and Leupeptin plus Ca2+ Glutamate plus Ca2+ and Neomycin plus Ca2+ Glutamate plus Ca2+ and Chloroquine plus Ca2+ Glutamate plus Ca2+ and Methylamine plus Ca2+

leupeptin neomycin chloroquine methylamine

Cortical neurons plus glial cells

Glial cells

16.50.5* 8.20.5** 7.00.5 7.90.2 11.20.8** 10.10.6 19.61.3 22.30.9 20.90.3 11.70.4 14.90.6 8.20.2

8.51.7 — 6.80.2 — — — 40.82.0** 39.81.0** — — — —

Cortical neurons grown for 14 days in the presence of glial and three-week-old glial cultures were exposed to 500 µM glutamate with or without added Ca2+ (1.3 mM) in the absence or presence of 200 µM leupeptin, 2 mM neomycin, 100 µM chloroquine or 10 mM methylamine. LDH leakage was determined after 18 h of incubation. Values are the meanS.E.M. of measurements from seven to 18 separate neuronal preparations. *Significantly different from control plus Ca2+ group (P<0.01). **Significantly different from glutamate plus Ca2+ group (P<0.01).

Ca2+-stimulated hydrolytic enzymes and irreversible injury to cortical neurons. The study followed the postulate that overstimulation of degradative enzymes, such as proteases and phospholipases, in response to elevated [Ca2+]i, contributes to lethal injury in a variety of cell types.17,27 While similar Ca2+-dependent activities may also be critical to the glutamate neurotoxicity associated with important neuropathologies, much remains to be determined concerning the nature and role of Ca2+-dependent processes in neuronal injury.10,33 However, a number of potentially confounding variables associated with the intact CNS present a formidable challenge to mechanistic studies of neurotoxicity.11,32 Cell culture models have become widely used and offer a more direct approach to the study of neuronal injury at the cellular level. Yet biochemical responses of nonneuronal cells, the presence of which are often required to promote EAA receptor expression, may interfere with measurements of neuronal responses to elevated [Ca2+]i. In the current study with purified cultures, the [Ca2+]i of cortical neurons was elevated to cytotoxic levels directly with ionomycin and responses were determined in the absence of significant interference from non-neuronal cells. The subsequent activities of several hydrolytic enzymes were determined and inhibitors were used to evaluate the degree to which these activities responded to Ca2+ and contributed to neuronal death. This approach facilitated a more direct examination of these events as they pertain specifically to cortical neurons, in a model that appears to mimic a late stage of glutamate toxicity, thought to be characterized by an acute loss of [Ca2+]i homeostasis.8 To determine further the relevance of this ionomycin model to glutamate toxicity, inhibitors that reduced ionomycin

toxicity were also evaluated with mixed neuronal cultures responsive to a Ca2+-dependent glutamate challenge. Ca2+-dependent proteolysis and cortical neuron injury There is growing evidence to suggest that proteolysis may be a key neurotoxic event, particularly during ischemic and EAA insults.24,30 One possible mechanism by which proteolysis may cause neurotoxicity is based on findings demonstrating that certain components of the neuronal cytoskeleton are preferred targets of Ca2+-dependent proteases.24,38,41 For example, spectrin hydrolysis may disrupt cytoskeletal connections to the plasma membrane and cause membrane blebbing, which ultimately leads to a loss of plasma membrane integrity.27 Evidence supporting this mechanism in neurons stems from reports demonstrating that spectrin degradation21,38 and post-ischemic injury in vivo2,21,24,30 were diminished by the protease inhibitors. Since these studies were conducted with models where non-neuronal cells far outnumbered neurons, it remained to be determined whether elevated neuronal [Ca+2]i and protease activation were linked causally to neuronal death. In addition, it was not clear which neuronal protease pool, lysosomal or non-lysosomal, was involved. Results of the current study demonstrate directly that elevating neuronal [Ca+2]i with ionomycin stimulated the proteolytic degradation of neuronal proteins (Fig. 3) well before the neurons were killed (Fig. 2). These events appear to be connnected, because both were decreased by the protease inhibitor leupeptin (Figs 2, 3). However, because leupeptin inhibits both lysosomal and nonlysosomal neuronal protease pools,24 two inhibitors

1142

M. R. Castillo and J. R. Babson

of lysosomal proteolysis, chloroquine and methylamine,28 were used to distinguish which protease pool caused the neuronal injury. Neither agent prevented the Ca2+–ionomycin-induced neuronal injury (Table 2). Taken together, these results provide direct evidence that the activation of non-lysosomal proteases during conditions of Ca+2 overload is toxic to cortical neurons. However, even though leupeptin prevented the neuronal Ca2+-dependent proteolysis, Ca2+-dependent toxicity was not completely suppressed by protease inhibition. This fact suggested that additional Ca2+-dependent mechanisms might be involved. Ca2+-dependent phospholipases and cortical neuron injury Ca2+-dependent phospholipases have also been considered to be causative factors in the development of lethal cell injury in diverse cell types.17,27 General observations of free fatty acid accumulation during neuropathological disorders, such as ischemia, hypoglycemia and status epilepticus, provide indirect evidence for phospholipase participation in neurotoxicity.39 The release of free fatty acids and active metabolites, lysophospholipid production, and the loss of plasma membrane integrity have all been considered as possible cytotoxic events triggered by phospholipase activation.6,27,39 Clearly, the degradation of cellular membrane components would be expected to alter neuronal function and, if extensive, eventually compromise plasma membrane integrity and thus cell viability. Neurons are known to contain at least three different phospholipase activities, PLA2, PLC and PLD.19 However, the relative contribution of these enzymes to neurotoxicity is not known, and it is not clear whether receptor-stimulated or Ca2+activated forms of these enzymes are involved. Neuronal PLC has been shown to be activated in response to EAA receptor stimulation,42 yet recent data suggest that EAA receptor-mediated PLC activation does not cause neuronal death.48 Results of the present study demonstrate that phospholipase activation in cortical neurons can occur as a direct response to elevated [Ca2+]i in the absence of EAA receptor stimulation. The Ca2+–ionomycin challenge initiated extensive hydrolysis of phosphatidylinositol and phosphatidylcholine that preceded the onset of neuronal death. The release of inositol phosphates and choline is evidence of Ca2+-dependent phosphatidylinositol-specific PLC and phosphatidylcholine-specific PLD activities, respectively, in cortical neurons (Fig. 4, Table 4). The Ca2+-dependent PLC activity did not appear to hydrolyse phosphatidylcholine, for no phosphocholine release was detected (Table 4). The Ca2+-induced PLC hydrolysed approximately 18% of the total phosphatidylinositol (Fig. 4), a level substantially higher than reported previously.48 In addition, the phospholipase inhibitor neomycin diminished neuronal death and reduced the

initial PLC activity (Figs 2, 4) to levels regarded as non-neurotoxic.48 It would appear that there is a wide range of PLC activity, where the upper levels may be neurotoxic. Neomycin, however, also attenuated PLD activity (Table 5), thus making it difficult to determine whether neomycin protection was due to inhibition either of PLC or PLD alone, or both. Since phosphatidylcholine is a major neuronal phospholipid,19 Ca2+-dependent PLD activation could contribute substantially to the loss of neuronal plasma membrane integrity. Overall, our results demonstrate that cortical neurons contain Ca2+dependent PLC and PLD activities. Furthermore, the timing and magnitude of the phospholipid hydrolysis, as well as the phospholipase inhibition and concomitant neuroprotection, all suggest that these enzymes participate in the development of Ca2+dependent neuronal injury. These results may have clinical implications, given recent findings suggesting that Ca2+, phosphatidylinositol degradation and calpain activity may all be connected causatively to the onset of ischemic injury.46 Arachidonate release is often assumed to result from PLA2 stimulation and has been observed in several neuronal cultures in response to different toxic insults.16,34,44 However, both the neurotoxic impact of PLA2 activity34,44 and activation mechanism remain in question, and may even be model dependent. For example, glutamate-induced arachidonate release may be coupled to Ca2+ and protein kinase C (PKC),5 although EAA agonists also stimulated a PKC-independent arachidonate release apparently dependent on Ca2+ influx specifically from the Na+/Ca2+ exchanger.15 Another example of receptor-mediated arachidonate release was caused by interferon-ã, and was both Ca2+ and PKC independent.29 Whatever the mechanism, EAA neurotoxicity has been clearly shown to require [Ca2+]i increases substantially higher than those common to intracellular signaling.22 It should also be noted that the presence of non-neuronal cells can further complicate data interpretation. Alternatively, because [3H]arachidonate is incorporated preferentially into phosphotidylinositol and phosphotidylcholine,44 its release may result directly from the co-ordinated action of PLC and lipases.45 In the current study, receptor-mediated PLA2 activation was unlikely, because the purified cortical neurons were unresponsive to EAA receptor agonists. Under these conditions, several different PLA2 inhibitors failed to protect the neurons from Ca2+-induced cytotoxicity (Table 2). In addition, while arachidonate release occurred, it was not inhibited by the PLA2 inhibitor mepacrine, but by neomycin and the DAG lipase inhibitor RHC8027 (Table 5). These data are consistent with a Ca2+-dependent pathway of neuronal phospholipid degradation, whereby arachidonate release is catalysed by DAG and monoacylglycerol lipases, subsequent to the action of PLC and/or PLD and in the absence of PLA2 activity.

Ca2+-dependent cell injury in cortical neurons

Comparison of glutamate-induced injury with Ca2+ ionophore-induced injury It is generally agreed that glutamate-induced lethal injury has Ca2+-dependent elements,10,33 and the extent to which glutamate and other EAAs elevate [Ca2+]i is directly related to the neurotoxic potentials of individual EAAs.22 Since many important neuropathologies may result from glutamate excitotoxicity and Ca2+ overload,10,33 an attempt was made to relate results obtained with ionomycin-treated neurons to a cellular model of glutamate toxicity. The results indicate that both models share some common features, such as a clear Ca2+-dependent cytotoxicity (Tables 1, 6) and comparable morphological responses, suggesting that the ionomycin model mimics the Ca2+-dependent phase of glutamate neurotoxicity.8 The proteolysis inhibitor profiles of both models are also similar (Table 3, Table 6), and lend additional support to the notion that calpains may be key neurotoxic elements of neuropathologies linked to Ca2+ and glutamate. The contribution of PLC and PLD to glutamate neurotoxicity could not be demonstrated directly with mixed cultures, because, while neomycin was nontoxic to purified cortical neurons, it was toxic to glia found in the mixed cultures (Tables 3, 6). Conclusions may be reached indirectly, however, by a comparative analysis of the data. Specifically, if neomycin was neuroprotective toward glutamateresponsive neurons, LDH leakage from mixed cultures treated with glutamate plus neomycin would be similar to that from sister cultures treated with neomycin alone. Conversely, if neomycin was not neuroprotective, LDH leakage from mixed cultures

1143

exposed to glutamate plus neomycin would be greater than that from sister cultures treated with neomycin alone. As the first situation occurred (Table 6), the data are consistent with neomycin protection and thus the involvement of PLC and PLD in Ca2+-dependent glutamate toxicity. Overall, the data suggest that neuronal injury induced by the Ca2+–ionomycin challenge shares some key cytotoxic features with the glutamate challenge, and may be relevant to a number of neuropathologies that involve glutamate neurotoxicity. CONCLUSIONS

The current study focussed on the response of cortical neurons to conditions of Ca2+ overload. The data describe several distinct stages of Ca2+dependent neuronal injury. The initial stage is characterized by a rapid loss of axonal morphology and increased phosphatidylinositol hydrolysis. An intermediate stage involves changes in cell body morphology plus the degradation of neuronal protein and phosphatidylcholine. In a later stage, the irreversible loss of plasma membrane integrity denotes neuronal death. The data suggest that the stimulation of non-lysosomal protease, PLC and PLD activities, and the resultant degradation of cellular components may be important events of Ca2+-dependent injury to cortical neurons. It also appears that neuronal injury induced by the Ca2–ionomycin challenge shares key neurotoxic features with glutamate toxicity, and thus, by extension, with important neuropathologies. Acknowledgement—This work was supported by NIH Grant GM 41496.

REFERENCES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Anderson R. J. and Babson J. R. (1989) Reduction of doxorubicin toxicity by methylene blue in cultured rat myocardial cells. Biochem. Pharmac. 38, 2568–2571. Bartus R. T., Baker K. L., Heiser A. D., Sawyer S. D., Dean R. L., Elliott P. J. and Straub J. A. (1994) Postischemic administration of AK275, a calpain inhibitor, provides substantial protection against focal ischemic brain damage. J. cerebr. Blood Flow Metab. 14, 537–544. Bartus R. T., Elliott P. J., Hayward N. J., Dean R. L., Harbeson S., Straub J. A., Li Z. and Powers J. C. (1995) Calpain as a novel target for treating acute neurodegenerative disorders. Neurol. Res. 17, 249–258. Bligh E. G. and Dyer W. J. (1959) A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37, 911–917. Bonventre J. V. (1997) Roles of phospholipases A2 in brain cell and tissue injury associated with ischemia and excitotoxicity. J. Lipid Mediat. Cell Signal. 16, 199–208. Chien K., Abrams J., Serroni A., Martin J. T. and Farber J. L. (1978) Accelerated phospholipid degradation and associated membrane dysfunction in irreversible, ischemic liver cell injury. J. biol. Chem. 253, 4809–4817. Chien R. K., Sen A., Reynolds R., Chang A. Y. K., Gunn M. D., Buja M. and Willerson J. T. (1985) Release of arachidonate from membrane phospholipids in cultured neonatal rat myocardial cells during adenosine triphosphate depletion. J. clin. Invest. 75, 1770–1780. Choi D. (1987) Ionic dependence of glutamate neurotoxicity. J. Neurosci. 7, 369–379. Choi D. (1988) Calcium-mediated neurotoxicity: relationship to specific channel types and role in ischemic damage. Trends Neurosci. 11, 465–469. Choi D. W. (1995) Calcium: still center-stage in hypoxic–ischemic neuronal death. Trends Neurosci. 18, 58–60. Choi D. W., Maulucci-Gedde M. and Kriegstein A. R. (1987) Glutamate neurotoxicity in cortical cell culture. J. Neurosci. 7, 357–368. Cockcroft S., Howell T. and Gomperts B. D. (1987) Two G-proteins act in series to control stimulus–secretion coupling in mast cells: use of neomycin to distinguish between G-proteins controlling polyphosphoinositide phosphodiesterase and exocytosis. J. Cell Biol. 105, 2745–2750. Cook S. J. and Wakeman M. (1989) Analysis of the water-soluble products of phosphatidylcholine breakdown by ion-exchange chromatography. Biochem. J. 263, 581–587.

1144 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.

M. R. Castillo and J. R. Babson Dougherty K. K., Spilman S. D., Green C. E. and Steward A. R. (1980) Primary cultures of adult mouse and rat hepatocytes for studying the metabolism of foreign chemicals. Biochem. Pharmac. 29, 2117–2124. Dumuis A., Sebben M., Fagni L., Prezeau L., Manzoni O., Cragoe E. J. J. and Bockaert J. (1993) Stimulation by glutamate receptors of arachidonic acid release depends on the Na+/Ca2+ exchanger in neuronal cells. Molec. Pharmac. 43, 976–981. Dumuis A., Sebben M., Haynes L., Pin J.-P. and Bockaert J. (1988) NMDA receptors activate the arachidonic acid cascade system in striatal neurons. Lett. Nature 336, 68–70. Farber J. (1990) The role of calcium ions in toxic cell injury. Environ. Health Perspect. 84, 107–111. Farber J. L. (1990) The role of calcium in lethal cell injury. Chem. Res. Toxic. 3, 503–508. Guroff G. (1980) Molecular Neurobiology, p. 571. Marcel Dekker, New York. Hertz E., Yu A. C. H., Hertz L., Juurlink B. H. J. and Schousboe A. (1989) Preparation of primary cultures of mouse cortical neurons. In A Dissection and Tissue Culture Manual of the Nervous System (eds Shahar A. et al.), p. 183. Alan R. Liss, New York. Hong S. C., Goto Y., Lanzino G., Soleau S., Kassell N. F. and Lee K. S. (1993) Neuroprotection with a calpain inhibitor in a model of focal cerebral ischemia. Stroke 25, 663–669. Hyrc K., Handran S. D., Rothman S. M. and Goldberg M. P. (1997) Ionized intracellular calcium concentration predicts excitotoxic neuronal death: observations with low-affinity fluorescent calcium indicators. J. Neurosci. 17, 6669–6677. Lee H.-C., Fellenz-Maloney M. P. and Liscovitch M. (1993) Phospholipase D-catalyzed hydrolysis of phosphatidylcholine provides the choline precursor for acetylcholine synthesis in human neuronal cell line. Proc. natn. Acad. Sci. U.S.A. 90, 10,086–10,090. Lee K. S., Frank S., Vanderklish P., Arai A. and Lynch G. (1991) Inhibition of proteolysis protects hippocampal neurons from ischemia. Proc. natn. Acad. Sci. U.S.A. 88, 7233–7237. Lipton S. A., Sucher N. J., Kaiser P. K. and Dreyer E. B. (1991) Synergistic effects of HIV coat protein and NMDA receptor-mediated neurotoxicity. Neuron 7, 11–118. MacDermott A., Mayer M., Westbrook G., Smith S. and Barker J. (1986) NMDA-receptor activation increases cytoplasmic calcium concentration in cultured spinal cord neurons. Nature 321, 519–522. Nicotera P., Bellomo G. and Orrenius S. (1990) The role of Ca2+ in cell killing. Chem. Res. Toxic. 3, 484–494. Nicotera P., Hartzell P., Baldi C., Svensson S., Bellomo G. and Orrenius S. (1986) Cystamine induces toxicity in hepatocytes through the elevation of cytosolic Ca2+ and the stimulation of a nonlysosomal proteolytic system. J. biol. Chem. 261, 14,628–14,635. Ponzoni M. and Cornaglia-Ferraris P. (1993) Interferon-g-stimulated and GTP-binding-proteins-mediated phospholipase A2 activation in human neuroblasts. Biochem. J. 294, 893–898. Rami A. and Krieglstein J. (1993) Protective effects of calpain inhibitors against neuronal damage caused by cytotoxic hypoxia in vitro and ischemia in vivo. Brain Res. 609, 67–70. Ray J., Peterson D. A., Schinstine M. and Gage F. H. (1993) Proliferation, differentiation, and long term culture of primary hippocampal neurons. Neurobiology 90, 3602–3606. Rothman S. (1984) Synaptic release of excitatory amino acid neurotransmitter mediates anoxic neuronal death. J. Neurosci. 4, 1884–1891. Rothman S. M. and Olney J. W. (1995) Excitotoxicity and the NMDA receptor: still lethal after eight years. Trends Neurosci. 18, 57–58. Rothman S. M., Yamada K. A. and Lancaster N. (1993) Nordihydroguaiaretic acid attenuates NMDA neurotoxicity—action beyond the receptor. Neuropharmacology 32, 1279–1288. Saito K.-I., Elce J. S., Hamos J. E. and Nixon R. A. (1993) Widespread activation of calcium-activated neutral proteinase (calpain) in the brain in Alzheimer disease: a potential molecular basis for neuronal degeneration. Proc. natn. Acad. Sci. U.S.A. 90, 2628–2632. Seglen P., Grinde B. and Solheim A. E. (1979) Inhibition of the lysosomal pathway of protein degradation in isolated rat hepatocytes by ammonia, methylamine, chloroquine and leupeptin. Eur. J. Biochem. 95, 215–225. Seubert P., Ivy G., Larson J., Lee J., Shahi K., Baudry M. and Lynch G. (1988) Lesions of entorhinal cortex produce a calpain-mediated degradation of brain spectrin in dentate gyrus. I. Biochemical studies. Brain Res. 459, 226–232. Seubert P., Larson J., Oliver M., Jung M. W., Baudry M. and Lynch G. (1988) Stimulation of NMDA receptors induces proteolysis of spectrin in hippocampus. Brain Res. 460, 189–194. Siesjo B. K. (1985) Oxygen deficiency and brain damage: localization, evolution in time, and mechanisms of damage. Clin. Toxic. 23, 267–280. Siesjo B. K. (1988) Historical overview. Calcium, ischemia, and death of brain cells. Ann. N. Y. Acad. Sci. 522, 638–661. Siman R., Baudry M. and Lynch G. (1984) Brain fodrin: substrate for calpain I, an endogenous calcium-activated protease. Proc. natn. Acad. Sci. U.S.A. 81, 3572–3576. Sladeczek F., Pin J. P., Recasens M., Bockaert J. and Weiss S. (1985) Glutamate stimulates inositol phosphate formation in striatal neurons. Nature 317, 717–719. Verity M. A. (1992) Calcium-dependent processes as mediators of neurotoxicity. Neurotoxicology 13, 139–148. Verity M. A., Sarafian T., Pacifici E. H. K. and Sevanian A. (1994) Phospholipase A2 stimulation by methyl mercury in neuron culture. J. Neurochem. 62, 705–714. Waite M. (1986) Phospholipases of mammalian cells. In Lipids and Membranes: Past, Present, and Future (eds Op den Kamp J. A. F., Roelofsen B. and Wirtz K. W. A.), pp. 207–230. Elsevier, Amsterdam. Yamashima T., Saido T. C., Takita M., Miyazawa A., Yamano J., Miyakawa A., Nishijyo H., Yamashita J., Kawashima S., Ono T. and Yoshioka T. (1996) Transient brain ischaemia provokes Ca2+, PIP2 and calpain responses prior to delayed neuronal death in monkeys. Eur. J. Neurosci. 8, 1932–1944. Zimmerman U. P. and Schlaepfer W. W. (1984) Calcium-activated neutral protease in brain and other tissues. Prog. Neurobiol. 23, 63–78. Zinkand W. C., Moore W. C., Thompson C., Salama A. I. and Patel J. (1992) Ibotenic acid mediates neurotoxicity and phosphoinositide hydrolysis by independent receptor mechanisms. Molec. chem. Neuropath. 16, 1–10. (Accepted 26 January 1998)