Apoptotic Cell Death of Cultured Cerebral Cortical Neurons Induced by Withdrawal of Astroglial Trophic Support

EXPERIMENTAL NEUROLOGY ARTICLE NO.

149, 51–63 (1998)

EN976719

Apoptotic Cell Death of Cultured Cerebral Cortical Neurons Induced by Withdrawal of Astroglial Trophic Support Makoto Ohgoh, Manami Kimura, Hiroo Ogura, Kouichi Katayama, and Yukio Nishizawa Eisai Tsukuba Research Laboratories, 5-1-3 Tokodai, Tsukuba, Ibaraki 300-26, Japan Received June 20, 1997; accepted for publication September 30, 1997

latter case, neuronal cell death is probably induced by deprivation of neurotrophic factors and this type of cell death has recently been regarded as being an example of apoptosis because of the morphology of the dying neurons and the requirement for newly synthesized proteins (1). In adult brain, extensive and progressive neuronal loss is observed in chronic neurodegenerative diseases such as Alzheimer’s disease, Huntington’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis (2–5), raising the possibility that inappropriate apoptosis may be involved in the etiology of these diseases. Many studies on the programmed cell death of neurons have been reported. Neuronal cell death at embryonic stages is observed in vivo, and the process of developmental cell death can also be analyzed by using in vitro culture systems of neurons. In this method, neurons are maintained in the presence of specific survival factors, and cell death is induced by deprivation of the factors from the medium (6). A particular survival factor for the neurons has to be identified, and the survival of the neurons has to depend absolutely on that factor. Peripheral neurons dependent on nerve growth factor (NGF) have been widely used in previous experiments. Several mechanisms regulating neuronal cell death have been proposed on the basis of studies using peripheral neuron cultures, and the cell death is blocked by inhibitors of macromolecule synthesis (6), chronic depolarization by elevated potassium (7), and cyclic AMP (cAMP) (8). In addition, the cell death of cerebellar granule neurons induced by reducing the extracellular potassium concentration has been well characterized as an in vitro culture system for assessing the programmed cell death of central nervous system (CNS) neurons (9). However, the cell death of CNS neurons caused by deprivation of survival factors has not been extensively examined, because the relationship between individual neurons and neurotrophic factors is not well clarified. Genetic studies of cell death in the nematode Caenorhabditis elegans have identified several genes involved in the cell death processes. Among them, CED-3 encodes a protein which is required for the induction of

Peripheral neurons which depend on NGF for their survival undergo apoptosis after NGF deprivation. However, a convenient in vitro method for assessing the programmed cell death of the central neurons has not been established, because the dependence of particular central neurons on neurotrophic factors has been clarified only for small populations of neurons. Based on the fact that cortical neurons survive in culture for many weeks in the presence of astroglial cells, we have established an in vitro cell death model in which the neurons die through apoptosis. Cortical neurons were maintained on a cover slip for 1 week on top of astroglial cells, and then cell death was induced by separation of the neurons from the astroglial cells. The cortical neurons died within 2–4 days. Nuclei of the dying neurons showed the morphological features of apoptosis, and DNA fragmentation was observed by the TUNEL method and by in situ nick translation (ISNT) staining. The cell death was significantly suppressed by neurotrophic factors, NT-3, NT-4, BDNF, and GDNF, but not by NGF. The neuronal survival was prolonged, as in the case of peripheral neurons, by bFGF, elevated potassium, cAMP, forskolin, and metabotropic glutamate receptor agonist. The cell death was inhibited by inhibitors of interleukin-1bconverting enzyme and CPP32. CPP32-like proteolytic activity was increased prior to the appearance of apoptotic cells. These results suggest that cortical neurons die after separation from glial cells through apoptosis caused by deprivation of neurotrophic factors produced by the astroglial cells. r 1998 Academic Press Key Words: apoptosis; cortical neurons; neurotrophic factors; caspase; in situ nick translation; TUNEL method; neurotrophin.

INTRODUCTION

The cell death of neurons can be classified into two types. One is accidental cell death induced by traumatic injury of brain tissues or excitotoxicity. The other is programmed cell death, which is observed at the developmental stage of the nervous system. In the 51

0014-4886/98 $25.00 Copyright r 1998 by Academic Press All rights of reproduction in any form reserved.

52

OHGOH ET AL.

cell death (10). The mammalian homologues of CED-3 are interleukin-1b-converting enzyme (ICE) (11) and related proteins belonging to a new family of cysteine proteases, the caspase family (12). It was shown that caspase family members are involved in the neuronal apoptosis in motoneurons (13), in dorsal root ganglion (DRG) neurons (14), and in CNS neurons (15–17). In the present study, we have established an in vitro cell death model using rat cortical neurons. In our culture system, cortical neurons are cocultured with astroglial cells and die after separation of the neurons from the astroglial cells. We have characterized the neuronal cell death in this system and conclude that it is apoptotic and that our culture system would be a good model for studies on the cell death of CNS neurons. MATERIALS AND METHODS

Chemicals and Reagents Fetal calf serum, heat-inactivated horse serum, trypsin solution, penicillin, streptomycin, Dulbecco’s modified essential medium (DMEM), and Hepes were purchased from Life Technologies Inc. (Grand Island, NY), insulin, sodium selenite, putrescine, cytosine arabinofuranoside (Ara C), Deoxyribonuclease I (DNase I), dibutyric cAMP, basic fibroblast growth factor (bFGF), pepstatin, leupeptin, phenylmethylsulfonyl fluoride, and forskolin from Sigma Chemical Co. (St. Louis, MO), NGF, neurotrophin-3 (NT-3), neurotrophin-4 (NT-4), brain-derived neurotrophic factor (BDNF), from PEPRO TECH Inc. (Rocky Hill, NJ), transforming growth factor-b (TGF-b) from R&D Systems, Inc. (Minneapolis, MN), glial cell line-derived neurotrophic factor (GDNF) from Alomons Labs. (Jerusalem, Israel), NBQX (2,3dihydroxy-6-nitro-7-sulfamoylbenzo(F)quinoxaline) and MK-801 from Tocris Neuramin Ltd. (Bristol, UK), and Ac-DEVD-a-(4-methyl-coumaryl-7-amide) (Ac-DEVDMCA), Ac-YVAD-MCA, Ac-DEVD-CHO, Ac-YVAD-CHO from Peptide Institute, Inc. (Osaka, Japan). All other chemicals used were of reagent grade. Cell Cultures Cortical cell cultures were prepared from fetal rats of the Wistar strain (gestational age of 17 days). The cerebral cortex was dissected, placed in ice-cold Hanks’ balanced salt solution, minced (3–4 mm pieces) and incubated at 37°C for 15 min in Ca21/Mg21-free Hanks’ balanced salt solution (HBSS) containing 0.25% trypsin and 0.2 mg/ml DNase I. The cortical tissues were dissociated to single cells by gentle trituration using a glass pipette with a fire-polished tip. The cell suspension was mixed with DMEM supplemented with 10% fetal calf serum, 10% heat-inactivated horse serum, 5 µg/ml insulin, 30 nM sodium selenite, 100 µM putre-

cine, 20 nM progesterone, 15 nM biotin, 100 units/ml penicillin, 100 µg/ml streptomycin, and 1 mM sodium pyruvate, as described (18). The cell suspension was centrifuged and the resulting pellets were resuspended in the medium described above. The cortical cells were then pelleted again by centrifugation, suspended in the medium, and plated onto poly-L-lysine-coated coverslips. The cells were cultured in a CO2 incubator (5% CO2) at 37°C for 1 day and the coverslips were then transferred onto a confluent glial cell layer and cultured for 6–7 days in DMEM containing the same supplements as described above, but without serum. The cortical cells were treated with 10 µM Ara C for 1 day (it was added to the culture medium 1 day after plating) to reduce the growth of contaminating nonneuronal cells. The culture medium was changed every 3–4 days. The glial cells used were obtained from postnatal day 1 rats of the Wistar strain. The cerebral cortex was dissected and triturated in DMEM supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 100 µg/ml streptomycin, and the glial cells were cultured in a CO2 incubator (5% CO2) at 37°C for 11–18 days before use. Induction of Neuronal Cell Death Cortical neurons cultured for 7–8 days on the astroglial cell layer were washed with HBSS, transferred to new plates, and cultured in the above medium (serum-free) containing 1 µM MK801 and 10 µM NBQX. The culture medium was replaced with fresh medium on a daily basis. Cell death was assessed by counting the viable neurons, which were identified by morphological criteria, e.g., round and smooth soma, in the same microscopic field (0.42 3 0.28 mm). Microscopic and Immunohistochemical Analysis The population of neurons, astroglial cells, and microglial cells in culture was identified histochemically with anti-microtubule-associated protein 2 (MAP2) antibody (neurons), anti-glial fibrillary acidic protein (GFAP) antibody (astroglial cells), and isolectin B4 (microglial cells). Briefly, neurons on a glass coverslip were fixed for 30 min in 4% paraformaldehyde/4% sucrose in phosphate-buffered saline (PBS), and permeabilized with 0.25% Triton X-100 in PBS for 5 min. Endogenous peroxidase was inactivated by incubation with 3% H2O2 in PBS for 5 min. For neuronal and astroglial staining, the coverslips were incubated with blocking solution (PBS containing 5% (v/v) normal goat serum and 0.1% (v/v) Triton X-100) for 1 h at room temperature and then incubated overnight at 4°C with antibody to MAP2 (Sigma, clone HM-2) at 1:500 dilution with blocking solution or with antibody to GFAP (Sigma, clone G-A-5) at 1:1000 dilution, respectively. Visualization was carried out with a Vectastain ABC kit (Vector,

53

APOPTOSIS INDUCED BY ASTROGLIAL WITHDRAWAL

Burlingame, CA) and DAB substrate kit (Funakoshi, Tokyo). For microglial staining, the coverslips were incubated with peroxidase-labeled isolectin B4 (Sigma) in PBS containing 0.1 mM CaCl2, 0.1 mM MgCl2, and 0.1 mM MnCl2 at 1:50 dilution for 2 h at room temperature. The morphological features of apoptotic cell death were analyzed by phase-contrast microscopy, as well as by fluorescence microscopy with the nuclear dye, Hoechst 33258. Neurons on a glass coverslip were fixed for 30 min in 4% paraformaldehyde/4% sucrose in PBS, washed three times with PBS, and then incubated for 5 min with 8 µg/ml Hoechst 33258 in PBS at room temperature. They were washed three times with PBS, and observed by fluorescence microscopy. TUNEL and ISNT Labeling A TUNEL (terminal deoxynucleotidyltransferasemediated dUTP nick end labeling) protocol (19) or ISNT (in situ nick translation) protocol (20) was used to detect DNA fragmentation. Briefly, neurons on a glass coverslip were fixed for 30 min in 4% paraformaldehyde/4% sucrose in PBS and permeabilized with 0.25% Triton X-100 in PBS for 5 min. Endogenous peroxidase was inactivated by incubation with 3% H2O2 in PBS for 5 min. For TUNEL staining, neurons were incubated for 60 min at 37°C with terminal deoxynucleotidyl transferase (20 units/ml) in the buffer containing 30 mM Tris–HCl (pH 7.2), 1 mM CoCl2, 140 mM sodium cacodylate, and 10 mM biotinylated dUTP. The reaction was terminated by incubation with the stop solution (300 mM sodium chloride, 30 mM sodium citrate) for 15 min at room temperature. For ISNT labeling, neurons were incubated with DNA polymerase (100 U/ml) in the buffer containing 50 mM Tris–HCl (pH 7.5), 5 mM MgCl, 10 mM b-mercaptoethanol, 1 mM dATP, 1 mM dGTP, 1 mM dCTP, and 1 mM biotinylated dUTP for 15 min at 37°C and then washed three times with PBS. To visualize the incorporated biotinylated dUTP, the standard avidin–biotin–peroxidase procedure (Vectastain ABC kit) was employed. Measurement of Caspase Activity Neurons on a cover slip were washed in HBSS, transferred to 24-well plate in 200 µl of buffer A containing 50 mM Hepes (pH 7.5), 1 mM EDTA, 10 mM EGTA, and 20 µM digitonin, and incubated at 37°C for

10 min. These extracts were collected in 1.5-ml tubes and centrifuged at 15,000 rpm for 3 min. Protein concentrations of the supernatants were measured and aliquots (1 µg) were incubated with the fluorogenic substrate Ac-DEVD-MCA (50 µM) for CPP32-like activity or Ac-YVAD-MCA (50 µM) for ICE-like activity at 37°C for 60 min in 200 µl of buffer A. The levels of released 7-amino-4-methylcoumarin (AMC) were measured with a spectrofluorometer, Cytofluor II (excitation, 380 nm; emission, 460 nm). Specific CPP32- and ICE-like activities were determined by subtracting the values obtained in the presence of specific inhibitor for CPP32 (Ac-DEVD-CHO) or for ICE (Ac-YVAD-CHO) at a concentration of 0.1 µM. Statistic Analysis All data are expressed as mean 6 SEM. Statistic analyses of Figs. 7 to 12 were performed using Mann– Whitney U test because the number of surviving neurons were expressed as percentage of initial number of neurons. RESULTS

Induction of Neuronal Cell Death by Separation of Neurons from Astroglial Cells Rat cerebral cortical neurons seeded on glass coverslips were cocultured with astroglial cells in the same plate and maintained for 1 week. About .95% of the cells contained in this culture were neurons as assessed by counting the neuronal cells after immunolabeling with antibodies against MAP2 and GFAP and labeling with isolectin B4 for microglial staining (Fig. 1). Neuronal cell death could be induced by separation of neurons from the astroglial cell layer after transferring the coverslips to new plates and to culture in fresh medium (Fig. 2). At a low cell density of 0.075 3 105/cm2, the neurons died immediately upon astroglial cell withdrawal (Fig. 2A), while they did not die at the density of 0.6 3 105/cm2 (Fig. 2B). The effect of medium change on cell survival was then examined for high-density culture (Fig. 3). After transfer of the neurons to new plates, the culture medium was replaced with fresh medium at the indicated times. Neuronal cell death was observed only when the medium was changed every day to Day 3 (Fig. 3C). The cell death was not

FIG. 1. Histochemical identification of the population of neurons, astroglial cells, and microglial cells in culture used in this study. (A) Immunohistochemically stained neurons with anti-MAP-2 antibody and (B) astroglial cells with anti-GFAP antibody. (C) Microglial cells stained with isolectin B4. Scale bar, 50 µm. FIG. 4. Morphological analysis of dying neurons after separation from astroglial cells. Cerebral cortical neurons at 7 DIV were kept with (A, C, E, G) or separated from (B, D, F, H) astroglial cells and cultured for 3 days. The neurons were fixed and observed with a phase-contrast microscope (A, B) or stained by the TUNEL method (C, D), the ISNT method (E, F), or with Hoechst 33258 (G, H). Scale bar, 50 µm.

54

OHGOH ET AL.

APOPTOSIS INDUCED BY ASTROGLIAL WITHDRAWAL

55

56

OHGOH ET AL.

FIG. 2. Effect of cell density on the survival of cortical neurons after separation from astroglial cells. Cerebral cortical neurons were seeded at a density of 0.075 (A) or 0.6 (B) 3 105 cells/cm2 and cocultured with astroglial cells for 7 days. The cortical neurons were then kept with astroglial cells (closed circles) or separated from astroglial cells with (open circles) or without (open triangles) MK801 (1 µM) and NBQX (10 µM) and cultured for an additional 6 days. Surviving neurons were counted 1, 2, 3, and 6 days after astroglial withdrawal (see Materials and Methods), and the results are expressed as mean 6 SEM (n 5 4 per group).

caused by excitotoxicity, since an N-methyl-D-aspartate (NMDA) receptor antagonist, MK801 (1 µM), and a non-NMDA receptor antagonist, NBQX (10 µM), were always added to the fresh medium. When we used fresh medium without MK801 and NBQX, the neurons rapidly died after medium change (Figs. 2A and 2B). On the basis of these results, we employed the following procedure as standard. Cortical neurons seeded at a density of 0.6 to 0.8 3 105/cm2 were transferred to a new plate at Day 7 in fresh medium containing MK801 and NBQX, and the culture medium was replaced with

fresh medium on Days 1 and 2 thereafter. Under these conditions, although no morphological or quantitative changes were apparent at Day 1, surviving neurons gradually diminished to less than 10% of the initial cell count at Day 5 (Fig. 6). To observe the morphological features of this type of neuronal cell death, nuclear staining with Hoechst 33258 was performed (Fig. 4). In dying cells, the nuclei appeared to be slightly smaller and had a brighter fluorescence than the normal nuclei, indicating that chromatin was condensed and aggregated at the nuclear

FIG. 3. Effect of medium replacement with fresh medium on the survival of cortical neurons. Cerebral cortical neurons at 7 DIV were kept with (closed circles) or separated from (open circles) astroglial cells and cultured for an additional 7 days. The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) at the indicated times. Surviving neurons were counted 2, 3, 4, and 7 days after astroglial withdrawal (see Materials and Methods), and the results are expressed as mean 6 SEM (n 5 4 per group).

APOPTOSIS INDUCED BY ASTROGLIAL WITHDRAWAL

57

membranes. Several spots in the nucleus were also observed, indicating that fragmentation of the nucleus occurred in the dying neurons. Moreover, almost all dying cells were stained positively by the ISNT and TUNEL methods, which detect single and double strand breaks of DNA in situ, respectively (Fig. 4). While a small number of ISNT- or TUNEL-positive cells existed in the presence of astroglial cells, the number of cells, which were positive to ISNT (Fig. 5) or TUNEL (data not shown), was significantly increased in the absence of astroglial cells, corresponding to the reduction of surviving neurons. These results suggest that the cortical neurons died through an apoptotic process after the withdrawal of astroglial cells. Inhibition of Cell Death by Neurotrophins The neuronal cell death induced by astroglial withdrawal was significantly suppressed and delayed by addition of NT-3, while neither NGF nor TGF-b affected the cell death (Fig. 6). Protection by NT-3 was dosedependent, and the effect reached a plateau at 5 ng/ml (Fig. 7). To determine the commitment point for cortical neurons undergoing cell death, we examined the effect of addition of NT-3 at Day 0, 1, or 2 after separation from astroglial cells. As shown in Fig. 8, addition of NT-3 at Day 1 after the separation prevented neuronal cell death to the same degree as simultaneous addition.

FIG. 5. Neuronal cell death and increase in ISNT-positive neurons caused by withdrawal of astroglial cells. Cerebral cortical neurons at 7 DIV were kept with (circles) or separated from (triangles) astroglial cells and cultured for an additional 3 days. The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) at the indicated times. Surviving (open symbols) and ISNT-positive (closed symbols) neurons were counted 1, 2, and 3 days after astroglial withdrawal. The results are expressed as percentage of the cell number counted at Day 0, and shown as mean 6 SEM (n 5 6–18 per group).

FIG. 6. Effects of growth factors on neuronal survival after withdrawal of astroglial cells. Cerebral cortical neurons at 7 DIV were kept with (closed squares) or separated from astroglial cells with no addition (open circles), NGF (100 ng/ml; open triangles), TGF-b (10 ng/ml; open squares), or NT-3 (50 ng/ml; closed circles) and cultured for an additional 8 days. The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) on Days 0, 1, and 2. The results are expressed as percentage of the cell number counted at Day 0, and data are the mean with SEM (n 5 3–4 per group).

However, addition of NT-3 at Day 2 after separation had a weaker effect than addition at Day 0 or 1 after separation. We also examined the effects of other neurotrophic factors and growth factors that have been shown to

FIG. 7. Effects of NT-3 on neuronal survival after withdrawal of astroglial cells. Cerebral cortical neurons at 7 DIV were kept with or separated from astroglial cells with various concentrations of NT-3. The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) on Days 0, 1, and 2. Surviving neurons were counted 4 days after withdrawal of astroglial cells, and the results are expressed as percentage of the cell number counted at Day 0. Data are the mean with SEM (n 5 6 per group). Statistic analysis was performed by Mann–Whitney U test. Cells treated with various concentration of NT-3 were compared to those without astroglial cells alone (*P , 0.05).

58

OHGOH ET AL.

have protective effects for various neuronal cells. As shown in Fig. 9, bFGF tended to have a weak protective effect. An additive effect was observed when bFGF and NT-3 were simultaneously added, suggesting that the protective effects of bFGF and NT-3 operate through different pathways. BDNF, NT-4, and GDNF also showed significant protective effects against neuronal cell death in a dose-dependent manner (Figs. 10A and 10B). Pharmacological Manipulations of Neuronal Cell Death The cell death of cultured sympathetic neurons induced by deprivation of NGF has been reported to be inhibitable by raising the concentration of KCl in the medium from 25 to 40 mM (7). The same treatment also tended to be slightly protective in our system (Fig. 11). Forskolin, an activator of adenylate cyclase (Fig. 12), and dibutylic cAMP (data not shown) each had a clear suppressive effect, as in the case of peripheral neurons (8), suggesting that the neuronal cell death of cortical neurons was similarly regulated by intracellular cAMP level. Trans(1S, 3R)-1-aminocyclopentane-1,3-dicarboxylic acid (ACPD), a metabotropic glutamate receptor agonist, has been reported to have neuroprotective effects against neuronal cell death induced by NMDA (21), low K1 (22), or b-amyloid peptide (25–35) (23).

FIG. 9. Effect of a combination of NT-3 and bFGF on neuronal survival after withdrawal of astroglial cells. Cerebral cortical neurons at 7 DIV were kept with or separated from astroglial cells in the absence or presence of NT-3 (10 ng/ml) and/or bFGF (30 ng/ml). The culture medium was replaced with fresh medium supplemented with MK801 (1 µM)and and NBQX (10 µM) on Days 0, 1, and 2. Surviving neurons were counted 3 days after withdrawal of astroglial cells, and the results are expressed as percentage of the cell number counted at Day 0. Data are the mean with SEM (n 5 6 per group). Statistic analysis was performed by Mann–Whitney U test. Cells treated with NT-3 and/or bFGF were compared to those without astroglial cells alone (**P , 0.01 and ***P , 0.005).

Relatively weak effects of transACPD at 10 and 100 µM were observed in cortical neurons (Fig. 12). Involvement of Caspases

FIG. 8. Inhibitory effect of NT-3 added after withdrawal of astroglial cells on neuronal survival. Cerebral cortical neurons at 7 DIV were kept with or separated from astroglial cells. NT-3 (10 ng/ml) was added to the medium at 0, 1, or 2 days after separation from astroglial cells. The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) on Days 0, 1, and 2. Surviving neurons were counted 3 days after withdrawal of astroglial cells, and the results are expressed as percentage of the cell number counted at Day 0. Data are the mean with SEM (n 5 6 per group). Statistic analysis was performed by Mann–Whitney U test. Cells treated with NT-3 were compared to those without astroglial cells alone (*P , 0.05 and **P , 0.01).

An inhibitor of ICE, Ac-YVAD-CHO, derived from the processing site of pro-IL-1b (24), or an inhibitor of CPP32, Ac-DEVD-CHO, derived from the processing site of poly (ADP-ribose) polymerase, which is one of the substrates of CPP32 (25), was used to examine the involvement of caspase family members. These inhibitors significantly suppressed neuronal cell death in our system at the concentration of 30 µM (Figs. 13A and 13B). Furthermore, we measured caspase activity using fluorogenic peptide substrates (Ac-DEVD-MCA for CPP32-like activity and Ac-YVAD-MCA for ICE-like activity). As shown in Fig. 13C, CPP32-like activity was significantly increased 3 days after astroglial withdrawal. On the other hand, Ac-YVAD-MCA-cleavage activity in neurons was unchanged after withdrawal of astroglial cells, and the detected substrate cleavage activity was not inhibited by Ac-YVAD-CHO, implying that this was a nonspecific activity (data not shown). These results suggest that at least CPP32-like protease is involved in the neuronal cell death induced by withdrawal of astroglial cells in this model.

APOPTOSIS INDUCED BY ASTROGLIAL WITHDRAWAL

59

FIG. 10. Effects of NT-4, BDNF, and GDNF on neuronal survival after withdrawal of astroglial cells. Cerebral cortical neurons at 7 DIV were kept with or separated from astroglial cells with various concentrations of NT-4, BDNF (A), or GDNF (B). The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) on Days 0, 1, and 2. Surviving neurons were counted 3 days after withdrawal of astroglial cells, and the results are expressed as percentage of the cell number counted at Day 0. Data are the mean with SEM (n 5 6 (A) or 3 (B) per group). Statistic analysis was performed by Mann–Whitney U test. Cells treated with NT-4, BDNF, or GDNF were compared to those without astroglial cells alone (*P , 0.05).

FIG. 11. Effects of NT-3 and 40 mM KCl on neuronal survival after withdrawal of astroglial cells. Cerebral cortical neurons at 7 DIV were kept with or separated from astroglial cells with NT-3 (50 ng/ml) or 40 mM KCl. The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) on Days 0, 1, and 2. Surviving neurons were counted 4 days after withdrawal of astroglial cells, and the results are expressed as percentage of the cell number counted at Day 0. Data are the mean with SEM (n 5 4 per group). Statistic analysis was performed by Mann–Whitney U test. Cells treated with NT-3 or KCl were compared to those without astroglial cells alone (*P , 0.05).

FIG. 12. Effects of forskolin and transACPD on neuronal survival after withdrawal of astroglial cells. Cerebral cortical neurons at 7 DIV were kept with or separated from astroglial cells with NT-3 (50 ng/ml), forskolin (1 µM), or transACPD (tACPD, 10 or 100 µM). The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) on Days 0, 1, and 2. Surviving neurons were counted 4 days after withdrawal of astroglial cells, and the results are expressed as percentage of the cell number counted at Day 0. Data are the mean with SEM (n 5 5–6 per group). Statistic analysis was performed by Mann–Whitney U test. Cells treated with NT-3, Forskolin, or tACPD were compared to those without astroglial cells alone (*P , 0.05, **P , 0.01, and ***P , 0.005).

60

OHGOH ET AL.

FIG. 13. Involvement of caspase family members in the neuronal cell death induced by withdrawal of astroglial cells. (A, B) Effects of Ac-YVAD-CHO and Ac-DEVD-CHO. Cerebral cortical neurons at 7 DIV were kept with or separated from astroglial cells with 30 µM Ac-YVAD-CHO (A) or 30 µM Ac-DEVD-CHO (B). The culture medium was replaced with fresh medium supplemented with MK801 (1 µM) and NBQX (10 µM) on Days 0, 1, and 2. Surviving neurons were counted 4 days after withdrawal of astroglial cells, and the results are expressed as percentage of the cell number counted at Day 0. Data are the mean with SEM (n 5 6 per group). (C) Measurement of CPP32-like activity. Cerebral cortical neurons at 7 DIV were kept with (open circles) or separated from (closed circles) astroglial cells and CPP32-like activity was measured 1 or 3 days after withdrawal of astroglial cells (see Materials and Methods). Data are the mean with SEM (n 5 7 per group).

DISCUSSION

Our results demonstrated that (1) the separation of cortical neurons from astroglial cells induced neuronal cell death, (2) protection against this neuronal cell death was provided by a variety of neurotrophic factors, and (3) the neurons most probably died through an apoptotic process. Neurons of the cerebral cortex consist of various types of cells which depend for their survival on different kinds of factors. In in vitro culture, the neurons survive at high density without addition of any particular survival factors in defined medium, most probably via an autocrine mechanism related to their ability to produce neurotrophic factors such as NGF, BDNF, and NT-3 (26, 27). Although it is difficult to maintain neuronal cultures at very low density, cultured neurons can survive even at low density for more than 2 weeks in the presence of astroglial cells, because astroglial cells release some neurotrophic factors (28). In our culture system described here, neurons were cocultured with astroglial cells for a week to obtain conditions such that neurons were completely dependent for their survival on the astroglial cells. Cell death was then induced by removal of the astroglial cells, by transferring the neurons to a new plate in fresh medium. Survival enhancement by the autocrine mechanism seen in high-density culture was suppressed by repeating the medium replacement on a daily schedule (Fig. 3). As a result, the neurons died gradually between Days 2 and 5 (Fig. 6). This neuronal cell death was not induced by excitotoxicity because we added MK801 and NBQX to the medium. A variety of neurotrophic factors, such as NT-3, showed suppressive effects on this neuronal cell death (Figs. 6 and 7). We, therefore, conclude that the neuronal cell death induced by separation from astroglial cells is caused by deprivation of survival factors that are produced by astroglial cells.

Dying neurons in our system showed the morphological features of apoptosis or naturally occurring cell death in the developing brain (29), namely chromatin condensation and nuclear fragmentation, as shown by staining with Hoechst 33258 (Fig. 4). The chromatin condensation in apoptotic cells is associated with endonuclease activation which causes the double strand breaks of nuclear DNA into oligonucleosomal fragments. ISNT staining and the TUNEL method detect single and double strand breaks of DNA in situ, respectively, and individual dying cells can be visualized by these methods. As shown in Fig. 4, almost all dying cells were stained by the ISNT or TUNEL method, and the stained nuclei were split into several parts, showing typical nuclear morphology of apoptotic cells. These results indicate that dying neurons in this system are morphologically apoptotic. The neuronal cell death induced by separation from astroglial cells was inhibited by NT-3, NT-4, BDNF, and GDNF (Figs. 10A and 10B), but was unaffected by NGF (Fig. 6). These results are consistent with the distribution of the receptors for neurotrophic factors in the brain. A very small population of neurons expresses the NGF receptor (Trk A) and many neurons express BDNF, NT-3, or NT-4 receptors (Trk B and Trk C) in the cerebral cortex (30–32). Although GDNF is known to be a potent survival factor for embryonic midbrain dopaminergic neurons (33), mRNA expression for GDNF receptor (GDNFR-a) is found not only in the ventral midbrain, but also in a part of the cortex in the embryonic rat CNS (34). Prehn et al. showed that treatment with TGF-b protected rat hippocampal neurons from apoptotic cell death induced by deprivation of astroglial trophic influences (35). However, we could not detect a neuroprotective effect of TGF-b in our system (Fig. 6). bFGF, which is not a neuron-specific factor, also

APOPTOSIS INDUCED BY ASTROGLIAL WITHDRAWAL

tended to have a weak protective effect (Fig. 9). Mayer et al. reported that bFGF promotes the survival of embryonic ventral mesencephalic dopaminergic neurons through potentiating astroglial production of unknown trophic factors (36). However, the effect of bFGF in our experiments appeared to be due to a direct action on cortical neurons, because contamination of astroglial cells was less than 5% in our neuronal culture. Moreover, the effects of bFGF and NT-3 were additive (Fig. 9), suggesting that the protective effects of bFGF and NT-3 operate through different pathways. Thus, we could evaluate the survival-promoting activities for cortical neurons of a variety of trophic factors, and our culture system seems to be a good model for neuronal cell death of the CNS neurons. Based on studies using in vitro culture systems of peripheral neurons, several mechanisms for suppression of neuronal apoptosis have been proposed. One protective strategy is membrane depolarization. Koike et al. showed that chronic depolarization by elevated K1 prevents the cell death of sympathetic neurons induced by deprivation of NGF (7). Electric activity also influences neuronal survival during certain stages of development, and high K1 promotes neuronal survival in culture of basal forebrain cholinergic neurons from postnatal 2-week-old rats (37). High K1 (40 mM) similarly tended to prevent neuronal cell death in our culture system (Fig. 11). Another mechanism is elevation of intracellular cAMP level, and forskolin, which elevates intracellular cAMP level (Fig. 12A), and dibutylyl cAMP (data not shown) had a protective effect in our system, as was the case in rat sympathetic neurons (8). These results suggest that neuronal survival in the peripheral nervous system and CNS might be regulated by common mechanisms. As shown in Fig. 12, a metabotropic glutamate receptor (mGlu) agonist, transACPD, had a neuroprotective effect against neuronal cell death induced by withdrawal of astroglial cells. The mGlu receptor family comprises eight subtypes, mGlu1–8 (38). Because transACPD acts as a mixed agonist at all mGlu receptor subtypes (39), it is not clear which subtype mediates the neuroprotective action of transACPD in our system. However, Copani et al. showed that L-AP4 and L-SOP, which selectively activate mGlu4, -6, -7, and DCG-IV, which is highly selective for mGlu2 and mGlu3, mimicked the protective action of transACPD against bAP(25–35)-induced neuronal apoptosis in mixed cortical culture (23), suggesting that these subtypes of mGlu receptors may function protectively against the cell death in our system. ICE/CED-3-like proteases named caspase family (12) have recently been identified as important mediators of apoptotic cell death. Among them, CPP32/apopain has been directly implicated in mammalian apoptosis (25, 40, 41) and plays a crucial role during morphogenetic

61

cell death in the central nervous system, based on an examination of CPP32-deficient mice (15). In cerebellar granule neurons, CPP32 is involved in potassium deprivation-induced apoptosis (16, 17). In the present study, a tetrapeptide inhibitor of CPP32/apopain, Ac-DEVDCHO, prevented the neuronal cell death in our system (Fig. 13B), and CPP32-like protease activity was markedly increased by the withdrawal of astroglial cells (Fig. 13C). Therefore, previous literature and our results indicate the possible involvement of CPP32 in neuronal cell death in the CNS. On the other hand, ICE or ICE-like proteases play an important role in the pathway of induction of apoptosis in peripheral neurons, because the cowpox virus gene crmA product, which binds to and inhibits ICE, inhibits apoptosis of DRG neurons induced by NGF deprivation (14). In our study, an inhibitor of ICE, Ac-YVAD-CHO, also showed a protective effect (Fig. 13A), but ICE-like activity was unchanged. Stefanis et al. found that ICE-like protease activity was not activated by withdrawal of NGF from PC12 cells (42), so withdrawal of trophic support may not activate ICE-like protease activity in neuronal culture. In conclusion, the cell death of cortical neurons induced by separation from astroglial cells is thought to be apoptotic as suggested by (1) the morphological features of nuclei stained with Hoechst 33258, (2) the detection of single and double strand breaks of DNA by means of ISNT and TUNEL staining (Fig. 4), (3) the protective effects of a variety of neurotrophic factors, and (4) the involvement of caspases. This type of cell death occurs physiologically during development in the central nervous system (29). However, evidence is accumulating that apoptotic cell death of neurons also occurs in neurodegenerative diseases such as Alzheimer’s disease, amyotrophic lateral sclerosis, Parkinson’s disease, and Huntington’s disease (2–5), giving rise to the idea that inappropriate apoptosis may contribute to the etiology and pathology of neurodegenerative diseases. In these diseases, it has not been clarified how neurons die, so it is very important to examine the molecular machinery involved in the induction of apoptosis in mature neurons. There is no evidence to show that the deprivation of trophic factor(s) causes neuronal apoptosis in the brain in neurodegenerative diseases, but our in vitro cell death model of the central nervous system should nevertheless be useful for analysis of the mechanism of neuronal apoptosis and in searching for therapeutic drugs for neurodegenerative diseases. REFERENCES 1.

Oppenheim, R. W. 1991. Cell death during development of the nervous system. Annu. Rev. Neurosci. 14: 453–501. 2. Portera-Cailliau, C., J. C. Hedreen, D. L. Price, and V. E. Koliatsos. 1995. Evidence for apoptotic cell death in Huntington

62

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

OHGOH ET AL. disease and excitotoxic animal model. J. Neurosci. 15: 3775– 3787. Thomas, L. B., D. J. Gates, E. K. Richfield, T. F. O’Brien, J. B. Schweitzer, and D. A. Steindler. 1995. DNA end labeling (TUNEL) in Huntington’s disease and other neuropathological conditions. Exp. Neurol. 133: 265–272. Tompkins, M. M., E. J. Basgall, E. Zamrini, and W. D. Hill. 1997. Apoptotic-like changes in Lewy-body-associated disorders and normal aging in substantial nigral neurons. Am. J. Pathol. 150: 119–131. Yoshiyama, Y., T. Yamada, K. Asanuma, and T. Asahi. 1994. Apoptosis related antigen, LeY and nick-end labeling are positive in spinal motor neurons in amyotrophic lateral sclerosis. Acta. Neuropathol. 88: 207–211. Martin, D. P., R. E. Schmidt, P. S. DiStefano, O. H. Lowry, J. G. Carter, and E. M. Johnson. 1988. Inhibition of protein synthesis prevents neuronal death caused by nerve growth factor deprivation. J. Cell. Biol. 106: 829–844. Koike, T., D. P. Martin, and E. M. Johnson Jr. 1989. Role of Ca21 channels in the ability of membrane depolarization to prevent neuronal death induced by trophic-factor deprivation: Evidence that levels of internal Ca21 determine nerve growth factor dependence of sympathetic ganglion cells. Proc. Natl. Acad. Sci. USA 86: 6421–6425. Edwards, S. N., A. E. Buckmaster, and A. M. Tolkovsky. 1991. The death programme in cultured sympathetic neurons can be suppressed at the posttranslational level by nerve growth factor, cyclic AMP, and depolarization. J. Neurochem. 57: 2140– 2143. D’Mello, S. R., C. Galli, T. Ciotti, and P. Calissano. 1993. Induction of apoptosis in cerebellar granule neurons by low potassium: Inhibition of death by insulin-growth factor I and cAMP. Proc. Natl. Acad. Sci. USA 90: 10989–10993. Ellis, H., and H. R. Horvitz. 1986. Genetic control of programmed cell death in the nematode C. elegans. Cell 44: 817–829. Miura, M., H. Zhu, R. Rotello, E. A. Hartwieg, and J. Yuan. 1993. Induction of apoptosis in fibroblasts by IL-1b-converting enzyme, a mammalian homolog of the C. elegans cell death gene ced-3. Cell 75: 653–660. Alnemri, E. S., D. J. Livingston, D. W. Nicholson, G. Salvesen, N. A. Thornberry, W. W. Wong, and J. Yuan. 1996. Human ICE/CED-3 protease nomenclature. Cell 87: 171. Milligan, C. E., D. Prevette, H. Yaginuma, S. Homma, C. Carswell, L. C. Fritz, K. J. Tomaselli, R. W. Oppenheim, and L. M. Schwartz. 1995. Peptide inhibitors of the ICE protease family arrest programmed cell death of motoneurons in vivo and in vitro. Neuron 15: 385–393. Gagliardini, V., P-A. Fernandez, R. K. K. Lee, H. C. A. Drexler, R. J. Rotello, M. C. Fishman, and J. Yuan. 1994. Prevention of vertebrate neuronal death by the crmA gene. Science 263: 826–828. Kuida, K., T. S. Zheng, S. Na, C. Kuan, D. Yang, H. Karasuyama, P. Rakic, and R. A. Flavell. 1996. Decreased apoptosis in the brain and premature lethality in CPP32-deficient mice. Nature 384: 368–372. Schulz, J. B., M. Weller, and T. Klockgether. 1996. Potassium deprivation-induced apoptosis of cerebellar granule neurons: A sequential requirement for new mRNA and protein synthesis, ICE-like protease activity, and reactive oxygen species. J. Neurosci. 16: 4696–4706. Armstrong, R. C., T. J. Aja, K. D. Hoang, S. Gaur, X. Bai, E. S. Alnemri, G. Litwack, D. S. Karanewsky, L. C. Fritz, and K. Tomaselli. 1997. Activation of the CED3/ICE-related protease

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28. 29. 30.

31.

CPP32 in cerebellar granule neurons undergoing apoptosis but not necrosis. J. Neurosci. 17: 553–562. Scholtz, W. K., C. Baitinger, H. Schulman, and P. T. Kelly. 1988. Developmental changes in Ca21/calmodulin-dependent protein kinase II in cultures of hippocampal pyramidal neurons and astrocytes. J. Neurosci. 8: 1039–1051. Gavrieli, Y., Y. Sherman, and S. A. Ben-Sasson. 1992. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J. Cell. Biol. 119: 493–501. Gold, R., M. Schmied, G. Rothe, H. Zischler, H. Breitschopf, H. Wekerle, and H. Lassmann. 1993. Detection of DNA fragmentation in apoptosis: Application of in situ nick translation to cell culture systems and tissue sections. J. Histochem. Cytochem. 41: 1023–1030. Ambrosini, A., L. Bresciani, S. Fracchia, N. Brunello, and G. Racagni. 1995. Metabotropic glutamate receptors negatively coupled to adenylate cyclase inhibit N-methyl-D-aspartate receptor activity and prevent neurotoxicity in mesencephalic neurons in vitro. Mol. Pharmacol. 47: 1057–1064. Copani, A., V. M. G. Bruno, V. Barresi, G. Battaglia, D. F. Condorelli, and F. Nicoletti. 1995. Activation of metabotropic glutamate receptors prevents neuronal apoptosis in culture. J. Neurochem. 64: 101–108. Copani, A., V. M. G. Bruno, G. Battaglia, G. Leanza, R. Pellitteri, A. Russo, S. Stanzani, and F. Nicoletti. 1995. Activation of metabotropic glutamate receptors protects cultured neurons against apoptosis induced by b-amyloid peptide. Mol. Pharmacol. 47: 890–897. Thornberry, N. A., H. G. Bull, J. R. Calaycay, K. T. Chapman, A. D. Howard, M. J. Kostura, D. K. Miller, S. M. Molineaux, J. R. Weidner, J. Aunis, K. O. Elliston, J. M. Ayala, F. J. Casano, J. Chin, G. J.-F. Ding, L. A. Egger, E. P. Gaffney, G. Limjuco, O. C. Palyha, S. M. Raju, A. M. Rolando, J. P. Salley, T.-T. Yamin, T. D. Lee, J. E. Shively, M. MacCross, R. A. Mumford, J. A. Schmidt, and M. J. Tocci. 1992. A novel heterodimeric cysteine protease is required for interleukin-1b processing in monocytes. Nature 356: 768–774. Nicholson, D. W., A. Ali, N. A. Thornberry, J. P. Vaillancourt, C. K. Ding, M. Gallant, Y. Gareau, P. R. Griffin, M. Labelle, Y. A. Lazebnik, N. A. Munday, S. M. Raju, M. E. Smulson, T.-T. Yamin, V. L. Yu, and D. K. Miller. 1995. Identification and inhibition of the ICE/CED-3 protease necessary for mammalian apoptosis. Nature 376: 37–43. Miranda, R. C., F. Sohrabji, and C. D. Toran-Allerand. 1993. Neuronal colocalization of mRNAs for neurotrophins and their receptors in the developing central nervous system suggests a potential for autocrine interactions. Proc. Natl. Acad. Sci. USA 90: 6439–6443. Kokaia, Z., J. Bengzon, M. Metsis, M. Kokaia, H. Persson, and O. Lindvall. 1993. Coexpression of neurotrophins and their receptors in neurons of the central nervous system. Proc. Natl. Acad. Sci. USA 90: 6711–6715. Banker, G. A. 1980. Trophic interactions between astroglial cells and hippocampal neurons in culture. Science 209: 809–810. Oppenheim, R. W. 1985. Naturally occurring cell death during neural development. Trends Neurosci. 17: 487–493. Merlio, J.-P., P. Ernfors, M. Jaber, and H. Persson. 1992. Molecular cloning of rat trkC and distribution of cells expressing messenger RNAs for members of the trk family in the rat central nervous system. Neuroscience 51: 513–532. Holtzman, D. M., Y. Li, L. F. Parada, S. Kinsman, C-K. Chen, J. S. Valletta, J. Zhou, J. B. Long, and W. C. Mobley. 1992. p140trk mRNA marks NGF-responsive forebrain neurons: Evidence that trk gene expression is induced by NGF. Neuron 9: 465–478.

APOPTOSIS INDUCED BY ASTROGLIAL WITHDRAWAL 32.

33.

34.

35.

36.

37.

Merlio, J.-P., P. Ernfors, Z. Kokaia, D. A. Middlemas, J. Bengzon, M. Kokaia, M.-L. Smith, B. K. Siesjo, T. Hunter, O. Lindvall, and H. Persson. 1993. Increased production of the TrkB protein tyrosine kinase receptor after brain insults. Neuron 10: 151–164. Lin, L. H., D. H. Doherty, J. D. Lile, S. Bektesh, and F. Collins. 1993. GDNF: A glial cell line-derived neurotrophic factor for midbrain dopaminergic neurons. Science 260: 1130–1132. Treanor, J. J. S., L. Goodman, F. De Sauvage, D. M. Stone, K. T. Poulsen, C. D. Beck, C. Gray, M. K. Armanini, R. A. Pollock, F. Hefti, H. S. Phillips, A. Goddard, M. W. Moore, A. Buj-Bello, A. M. Davies, N. Asai, M. Takahashi, R. Vandlen, C. E. Henderson, and A. Rosenthal. 1996. Characterization of a multicomponent receptor for GDNF. Nature 382: 80–83. Prehn, J. H., V. P. Bindokas, C. J. Marcuccilli, S. Krajewski, J. C. Reed, and R. J. Miller. 1994. Regulation of neuronal Bcl2 protein expression and calcium homeostasis by transforming growth factor type b confers wide-ranging protection on rat hippocampal neurons. Proc. Natl. Acad. Sci. USA 91: 12599–12603. Mayer, E., S. B. Dunnett, R. Pellitteri, and J. W. Fawcett. 1993. Basic fibroblast growth factor promotes the survival of embryonic ventral mesencephalic dopaminergic neurons-I. Effects in vitro. Neuroscience 56: 379–388. Nakamura, M., C. Nishio, T. Nonomura, and H. Hatanaka.

38.

39. 40.

41.

42.

63

1994. High potassium and cyclic AMP analog promote neuronal survival of basal forebrain cholinergic neurons in culture from postnatal 2-week-old rats. Dev. Brain Res. 81: 218–229. Nicoletti, F., V. Bruno, A. Copani, G. Casabona, and T. Knopfel. 1996. Metabotrophic glutamate receptors: A new target for the therapy of neurodegenerative disorders? Trends Neurosci. 19: 267–271. Nakanishi, S. 1992. Molecular diversity of glutamate receptors and implication for brain function. Science 258: 597–603. Fernandes-Alnemri, T., G. Litwack, and E. Alnemri. 1994. CPP32, a novel human apoptotic protein with homology to Caenorhabditis elegans cell death protein Ced-3 and mammalian interleukin-1b-converting enzyme. J. Biol. Chem. 269: 30761–30764. Tewari, M., L. T. Quan, K. O’Rourke, S. Desnoyers, Z. Zeng, D. R. Beidler, G. Poirier, G. S. Salvesen, and V. M. Dixit. 1995. Yama/CPP32b, a mammalian homolog of CED-3, is a CrmAinhibitable protease that cleaves the death substrate poly (ADP-ribose) polymerase. Cell 81: 801–809. Stefanis, L., D. S. Park, C. Y. I. Yan, S. E. Farinelli, C. M. Troy, M. L. Shelanski, and L. A. Greene. 1996. Induction of CPP32like activity in PC12 cells by withdrawal of trophic support. Dissociation from apoptosis. J. Biol. Chem. 271: 30663–30671.