- Email: [email protected]
Agmatine induces glutamate release and cell death in cultured rat cerebellar granule neurons
Brain Research 990 (2003) 165 – 171 www.elsevier.com/locate/brainres
Research report
Agmatine induces glutamate release and cell death in cultured rat cerebellar granule neurons Kazuho Abe a,b,*, Yuzuru Abe a, Hiroshi Saito a a
Department of Chemical Pharmacology, Faculty of Pharmaceutical Sciences, The University of Tokyo, Tokyo 113-0033, Japan b Department of Pharmacology, School of Pharmacy, Hoshi University, Tokyo 142-8501, Japan Accepted 30 July 2003
Abstract We investigated the effect of agmatine on cell viability of rat cerebellar granule neurons in a high-K+ (27.5 mM) medium. Exposure of cultured rat cerebellar granule neurons to agmatine (200 – 800 AM) resulted in a significant decrease in cell viability. Agmatine-induced neuronal death began to occur 6 – 12 h after addition, and gradually progressed. The agmatine neurotoxicity was attenuated by N-methyl-Daspartate (NMDA) receptor antagonists and by enzymatic degradation of L-glutamate with glutamic pyruvic transaminase. Furthermore, a significant increase in extracellular L-glutamate concentration was detected before cell death occurred. In addition, agmatine-induced glutamate release and cell death were both blocked by pretreatment with botulinum toxin C, which is known to specifically inhibit the exocytosis. The agmatine neurotoxicity was not observed when extracellular K+ concentration was lower (10 mM). These results suggest that agmatine induces glutamate release through the exocytosis and thereby causes NMDA receptor-mediated neuronal death in conditions in which extracellular K+ concentrations are elevated. D 2003 Elsevier B.V. All rights reserved. Theme: Neurotransmitters, modulators, transporters, and receptors Topic: Other neurotransmitters Keywords: Agmatine; Neurotoxicity; Glutamate; NMDA receptor; Exocytosis; Cerebellar granule neuron
1. Introduction Agmatine is an endogenous substance that is synthesized from arginine by arginine decarboxylase and metabolized to putrescine by agmatinase [17,18,27]. Agmatine is present in the brain of mammals including the rat, bovine [17] and human [20,31]. The level of agmatine in rat brain has been reported 0.2 – 0.4 Ag/g tissue by mass spectroscopy [17] and 0.331 –1.105 Ag/g tissue by high-performance liquid chromatography [10]. The presence of agmatine in astrocytes and neurons has been demonstrated by immunohistochemical examination with anti-agmatine antibody [30,32]. Electron microscopic examination has demonstrated that agmatine is present in axon terminals associated with synaptic vesicles [30,33]. In addition, agmatine is taken up into synaptosomes via a Na+-independent transport * Corresponding author. Department of Pharmacology, Hoshi University, 2-4-41 Ebara, Shinagawa, Tokyo 142-8501, Japan. Tel.: +81-3-54985785; fax: +81-3-5498-5787. E-mail address: [email protected] (K. Abe). 0006-8993/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0006-8993(03)03454-1
system [35]. Furthermore, agmatine has been reported to act as a ligand of the imidazoline receptor [17], block nicotinic receptor function in the chick retina and the rat superior cervical ganglion [19], block the N-methyl-D-aspartate (NMDA) type of glutamate receptors in rat hippocampal neurons [41] or block the voltage-gated Ca2 + channel in neurohypophysial nerve terminals [40]. These lines of evidence suggest the possible role of agmatine as a neurotransmitter or neuromodulator in the brain. L-Glutamate functions as a major excitatory neurotransmitter in the central nervous system, but excessive stimulation by L-glutamate is known to cause excitoxic neuronal damages through the overstimulation of the NMDA receptor, which may underlie the abnormal neuronal degeneration observed following hypoxia, ischemia, hypoglycemia and seizures [7,34,37]. Since agmatine has been reported to block the NMDA receptor [41], it is possible that agmatine modulates L-glutamate neurotoxicity. Indeed, Olmos et al. [29] have recently demonstrated that agmatine protects cultured rat cerebellar granule neurons from L-glutamate toxicity in a medium without Mg2 + and glucose. Further-
166
K. Abe et al. / Brain Research 990 (2003) 165–171
more, neuroprotective effects of agmatine have been reported in in vivo models of neurotoxic and ischemic brain injuries [13]. To further explore possible interaction of agmatine and Lglutamate in the brain, we also investigated the effect of agmatine on the neuronal survival by using primary culture of neonatal rat cerebellar granule neurons, the same preparation as Olmos et al. [29] have previously used. Other reasons that we chose to use primary cultured cerebellar granule neurons were (1) cerebellar granule neurons are easily cultured with high purity in a medium containing an elevated level of K+ [11], (2) primary culture of cerebellar granule neurons is a well characterized model for the study of NMDA receptor-mediated glutamate neurotoxicity [3,9,21,22,36], and (3) the activity of the agmatine-synthesizing enzyme arginine decarboxylase increases during postnatal developing cerebellum [12]. As a result, we found that agmatine exerts toxic effects for cultured cerebellar granule neurons, contrary to the observation by Olmos et al. [29]. Whether agmatine exerts neuroprotection or neurotoxicity may depend on conditions. Here we report possible mechanisms underlying a neurotoxic feature of agmatine.
2. Materials and methods 2.1. Chemicals Agmatine sulfate, dizocilpine, 2-amino-5-phosphonovalerate (APV) and glutamic pyruvic transaminase were purchased from Sigma (St. Louis, MO, USA). 2,3-Dihydroxy-6-nitro-7-sulphamoylbenzo[ f]-quinoxaline (NBQX), (RS)-a-methyl-4-carboxyphenylglycine (MCPG) and (RS)a-methyl-4-phosphonophenylglycine (MPPG) were purchased from Tocris Cookson (Bristol, UK). Botulinum toxin C was purchased from Wako (Osaka, Japan). 2.2. Cell culture Primary culture of cerebellar granule neurons was prepared according to the method by Gallo et al. [11]. Briefly, the cerebella obtained from 8-day-old Wistar rats were mechanically chopped, and cells were dissociated by trypsinization and pipetting. The dissociated cells were plated on poly-L-lysine-coated 96-well microtiter dishes at a density of 300,000 cells/cm2 in a modified Eagle’s medium containing 1.8 mM glutamine, 25 mM KCl, 91 Ag/ml gentamycin and 10% fetal bovine serum. To arrest the growth of nonneuronal cells, 1-h-arabinosylcytosine (10 AM) was added to the culture medium 24 h after the plating. The cultures were incubated at 37 jC with 5% CO2 for 7– 8 days, during which the medium was not changed to avoid the detachment of cells from culture dishes. In this culture, more than 95% of the cells were neurons as defined by staining with microtubule-associated protein-2. At days 7– 8, all the medium was carefully removed and replaced with a serum-
free Eagle’s medium containing 2 mM glutamine, 27.5 mM KCl, 100 Ag/ml gentamycin, 100 Ag/ml transferrin, 5 Ag/ml insulin, 20 nM progesterone and no L-glutamate. In a part of experiments, the medium was replaced with a serum-free Eagle’s medium containing different concentrations (10, 20 or 27.5 mM) of KCl to investigate whether agmatine neurotoxicity depends on extracellular K+ concentration. In any cases, the medium contained 1.8 mM Ca2 + and 0.8 mM Mg2 +. Agmatine or test reagents were added to the culture medium at this time. 2.3. Determination of cell viability Cell viability was evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction and lactate dehydrogenase (LDH) release assay. MTT reduction assay measures the redox activity of living cells, while LDH assay measures the activity of LDH released into the medium from dead cells. MTT reduction assay was performed as described previously [1]. Briefly, the cells were incubated with 0.25 mg/ml MTT at 37 jC, and a solution containing 50% dimethylformamide and 20% sodium dodecyl sulfate (pH 4.8) was added to terminate the reaction and to solubilize MTT formazan. The amount of MTT formazan produced was determined by measuring absorbance with a microplate reader at a test wavelength of 570 nm and a reference wavelength of 655 nm. LDH release assay was performed with a colorimetric LDH assay kit (Promega, Madison, WI, USA) according to the manufacturer’s instruction. Total cellular LDH activity was determined by solubilizing the cells with 0.2% Triton X-100. 2.4. Determination of L-glutamate concentration The concentration of L-glutamate was determined by a colorimetric method [2]. Briefly, 50 Al of culture supernatant was transferred to 96-well microtiter plates and was mixed with 50 Al of a substrate mixture consisting of 20 U/ml glutamate dehydrogenase, 2.5 mg/ml hnicotinamide adenine dinucleotide, 0.25 mg/ml MTT, 100 nM 1-methoxyphenazine methosulfate and 0.1% (v/ v) Triton X-100 in 0.2 M Tris – HCl buffer (pH 8.2). The reaction proceeded for 10 min at 37 jC and was stopped by adding 100 Al of a solution containing 50% dimethylformamide and 20% sodium dodecyl sulfate (pH 4.7). In this reaction, MTT (yellow) is converted into MTT formazan (purple) in proportion to L-glutamate concentration. The amount of MTT formazan produced was determined by measuring absorbance with a microplate reader at a test wavelength of 570 nm and a reference wavelength of 655 nm. The standard curve was constructed in each assay using cell-free medium containing known concentrations of L-glutamate. The concentration of L-glutamate in samples was estimated from the standard curve.
K. Abe et al. / Brain Research 990 (2003) 165–171
167
3. Results In a serum-free Eagle’s medium with 27.5 mM K+, agmatine caused a decrease in MTT reduction and an increase in LDH release in a concentration- and timedependent manner (Fig. 1). The minimum effective concentration of agmatine was 100 –200 AM (Fig. 1A and B). Agmatine-induced cell death began to occur 6– 12 h after addition, and gradually progressed (Fig. 1C). To examine
Fig. 2. Influences of glutamate receptor blockers on agmatine-induced cell death in cultured rat cerebellar granule neurons. Cells were exposed to 200 or 400 AM agmatine in the absence or presence of glutamate receptor blockers (A: 10 AM NMQX, 1 AM dizocilpine; B: 100 AM APV, 1 mM MCPG, 500 AM MPPG) for 24 h. Cell death was assessed by the LDH release assay. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. none, ##P < 0.01 vs. agmatine alone; Tukey’s test.
Fig. 1. Effect of agmatine on the viability of cultured rat cerebellar granule neurons. (A, B) Concentration-dependency. The cells were exposed to 50 – 800 AM agmatine for 24 h. (C) Time-dependency. The cells were exposed to none (o), 200 AM agmatine (n) or 400 AM agmatine (E) for 3, 6, 12 or 30 h. Cell viability was assessed by the MTT reduction assay (A) or the LDH release assay (B, C). Cellular MTT reduction activity was expressed as percentage of the control (no addition of agmatine). LDH activity in the culture supernatant was expressed as the percentage of total LDH activity. Data are represented as the means F S.E.M. of the determinations on five separate cultures. *P < 0.05, **P < 0.01 vs. none; Dunnett’s test.
if the toxicity is a specific effect of agmatine itself, the effects of the agmatine precursor arginine and the agmatine metabolite putrescine were also investigated. However, neither arginine nor putrescine showed significant effect on cell viability of cultured rat cerebellar granule neurons at least at concentrations up to 800 AM (n = 5, data not shown). To examine possible involvement of L-glutamate in agmatine-induced neuronal death, three series of experiments were performed. First, the influences of glutamate receptor antagonists were investigated. Glutamate receptor antagonists tested were NBQX (an AMPA receptor antagonist), dizocilpine (an NMDA receptor channel blocker), APV (an NMDA receptor antagonist), MCPG (a nonselective group I/group II metabotropic glutamate receptor antagonist) and MPPG (a group III metabotropic glutamate receptor antagonist). As shown in Fig. 2, agmatine neurotoxicity was abolished by APV or dizocilpine, but was not affected by NBQX, MCPG or MPPG. Second, it was tested if agmatine neurotoxicity is affected by enzymatic degradation of L-glutamate with glutamic pyruvic transaminase. Agmatine-induced neuronal death was abolished by the presence of 50 U/ml glutamic pyruvic transaminase and 3 mM pyruvate (Fig. 3). Third, concen-
168
K. Abe et al. / Brain Research 990 (2003) 165–171
Addition of arginine (50 – 800 AM) or putrescine (50 – 800 AM) induced no significant change in extracellular Lglutamate concentration (n = 5, data not shown). To examine if agmatine-induced L-glutamate release is mediated by the exocytosis, we investigated the influence of botulinum toxin C, an exocytosis inhibitor. As shown in Fig. 5A and B, agmatine-induced L-glutamate release and neuronal death were both abolished by pretreatment with botulinum toxin C. The same treatment did not affect neuronal death induced by exogenously applied L-glutamate (Fig. 5C), indicating that this treatment affects neither L-glutamate Fig. 3. Influence of enzymatic degradation of L-glutamate on agmatineinduced cell death in cultured rat cerebellar granule neurons. Cells were exposed to 200 or 400 AM agmatine in the absence (white columns) or presence of 50 U/ml glutamic pyruvic transaminase and 3 mM pyruvate (black columns) for 24 h. Cell viability was assessed by the MTT reduction assay, because addition of pyruvate may interfere with the measurement of LDH activity. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. none, ##P < 0.01 vs. agmatine alone; Tukey’s test.
tration of L-glutamate in the extracellular medium was measured before neuronal death occurred. Addition of 400 AM agmatine induced a significant increase in extracellular L-glutamate concentration, which began to occur 1.5 h after addition of agmatine and preceded an increase in LDH release, an index of cell membrane damage (Fig. 4).
Fig. 4. Changes in extracellular L-glutamate concentration and LDH release following addition of agmatine in rat cerebellar granule neuron cultures. Cells were exposed to none (o) or 400 AM agmatine ( ) for 0, 30, 90, 180, 270 or 360 min. L-Glutamate concentration (A) and LDH activity (B) were determined in the supernatant collected from the same culture. Data are the means F S.E.M. of five separate observations. *P < 0.05, **P < 0.01 vs. none; Student’s t-test.
.
Fig. 5. Influences of botulinum toxin C on agmatine-induced L-glutamate release (A), agmatine-induced cell death (B) and L-glutamate-induced cell death (C) in cultured rat cerebellar granule neurons. Cells were pretreated with none (white columns) or 15 pM botulinum toxin C (black columns) for 2 h, and then exposed to 200 – 400 AM agmatine (A, B) or 30 – 100 AM glutamate (C). Extracellular glutamate concentration (A) and LDH release (B, C) were determined 3 and 24 h after addition of agmatine, respectively. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. none, ##P < 0.01 vs. agmatine alone; Tukey’s test.
K. Abe et al. / Brain Research 990 (2003) 165–171
Fig. 6. Agmatine neurotoxicity depends on extracellular K+ concentration. Cultured rat cerebellar granule neurons were exposed to none (white columns) or 400 AM agmatine (gray columns) in a medium with 27.5, 20 or 10 mM K+, and LDH release was determined 24 h after addition of agmatine. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. control (none in 27.5 mM K+), ##P < 0.01 vs. none; Tukey’s test.
metabolism nor mechanisms involved in L-glutamate-induced neuronal death. The above data were contrary to previous studies demonstrating the neuroprotective role of agmatine. One of the differences in experimental conditions between a previous study by Olmos et al. [29] and our present study is K+ concentration in the extracellular medium. To examine if agmatine neurotoxicity depends on extracellular K+ concentration, the effects of agmatine were compared in the medium with different concentrations (10, 20 or 27.5 mM) of K+. Agmatine (400 AM) caused a significant increase in LDH release in serum-free Eagle’s medium with 20 mM K+, but the effect was smaller than that in serum-free Eagle’s medium with 27.5 mM K+ (Fig. 6). When the medium was switched to serum-free Eagle’s medium with 10 mM K+, basal LDH release was significantly increased, probably due to low K+induced apoptotic cell death. In this condition (10 mM K+), agmatine (400 AM) caused no significant change in LDH release (Fig. 6).
4. Discussion In the present study, we found that agmatine induced cell death of cultured rat cerebellar granule neurons in Eagle’s medium with high K+, which has been conventionally used to culture this type of cells with high purity [11]. Furthermore, the agmatine-induced neuronal death was preceded by accumulation of extracellular L-glutamate and was abolished by the presence of NMDA receptor antagonists or by enzymatic degradation of L-glutamate. The agmatine-induced neuronal death and L-glutamate accumulation were not mimicked by the agmatine precursor arginine and the agmatine metabolite putrescine, indicating that the effects are caused by agmatine itself, but not by its precursor or metabolite. These results suggest that agmatine specifically induces the release of
169
endogenous L-glutamate and thereby causes neuronal death through stimulation of NMDA receptors. In addition, agmatine-induced L-glutamate release and neuronal death were abolished by pretreatment with botulinum toxin C, which inhibits the exocytosis of synaptic vesicles by cleaving the synaptic protein syntaxin in nerve terminals [4,14,25], suggesting that agmatine-induced Lglutamate release is mediated by exocytotic processes. Agmatine has been reported to block ATP-sensitive K+ channels in pancreatic h-cells [39], leading to the stimulation of insulin release [38]. To the best of our knowledge, there has been no report concerning the effect of agmatine on Lglutamate release. Although agmatine has been reported to block the voltage-gated Ca2 + channel in neurohypophysial terminals [40], it is unlikely that agmatine blocks Ca2 + influx in presynaptic membranes of cerebellar granule neurons, which should result in the inhibition of L-glutamate release. Agmatine may directly or indirectly act on the exocytotic machinery downstream of Ca2 + influx. Further investigations are underway in our laboratory to elucidate molecular mechanisms by which agmatine stimulates the release of L-glutamate from neuronal cells. Our present study revealed a neurotoxic feature of agmatine, contrary to previous studies demonstrating the neuroprotective role of agmatine. In the report by Olmos et al. [29], cultured cerebellar granule neurons were preincubated in Mg2 +- and glucose-free Locke-HEPES buffer with 5.6 mM K+ to be deprived of energy resources and then exposed to agmatine (100 or 500 AM) for 45 min. In our present study, cultured cerebellar granule neurons were maintained in serum-free Eagle’s medium with 27.5 mM K+ and exposed to 50– 800 AM agmatine for 3– 30 h. Among several differences in experimental conditions between Olmos et al. [29] and us, extracellular K+ concentration may be a key factor in agmatine neurotoxicity. In fact, we found that agmatine exerted neurotoxicity in medium with high K+ (27.5 or 20 mM) but not low K+ (10 mM). Whether agmatine exerts neuroprotection or neurotoxicity may be determined by the balance among the amount of L-glutamate release stimulated by agmatine, the degree of NMDA receptor blockade by agmatine and the susceptibility of neurons to L-glutamate toxicity. It has been reported that the potency of agmatine in blocking the NMDA receptor is attenuated at depolarized membrane potentials [41]. Furthermore, in our preliminary experiments, toxicity of exogenously applied L-glutamate was enhanced by increasing extracellular K+ concentration in cultured cerebellar granule neurons (unpublished data). Thus, in a medium with high K+, L-glutamate-mediated toxic effect of agmatine may predominate over its neuroprotective effect through NMDA receptor blockade. NMDA receptor channels are blocked by physiological concentrations (approximately 1 mM) of Mg2 + [23,28]. In glutamatergic synapses, L-glutamate opens AMPA type of glutamate receptor channels followed by depolarization of postsynaptic membranes, which in turn relieves Mg2 + block of NMDA receptor channels, leading Ca2 + influx sufficient
170
K. Abe et al. / Brain Research 990 (2003) 165–171
for synaptic plasticity or neuronal death [5,8]. In the present study, agmatine caused neuronal death in a medium with 0.8 mM Mg2 +, and the effect was abolished by NMDA receptor blockers, but was not affected by AMPA receptor blocker, indicating that the opening of NMDA receptor channels by agmatine occurs in the presence of Mg2 + and does not require AMPA receptor-mediated membrane depolarization. There are two possible explanations for this discrepancy. First, Ca2 + permeability and Mg2 + sensitivity of NMDA receptor channels are different with the constitution of NMDA receptor subunits, which may account for functional differences of NMDA receptor channels among brain regions, developmental stages or culture conditions [15,16,24,26]. Second, Mg2 + block of NMDA receptor channels is relieved by raised extracellular K+ [6]. In the present study, cerebellar granule neurons would be depolarized by 27.5 mM K+, and Lglutamate could induce Ca2 + influx through NMDA receptor channels without AMPA receptor-mediated depolarization. The minimum effective concentration of agmatine in inducing cell death was 100 – 200 AM in the present study. The level of agmatine in rat brain has been reported 0.2 – 0.4 Ag/g tissue by mass spectroscopy [17] and 0.331– 1.105 Ag/ g tissue by high-performance liquid chromatography [10]. On the assumption that 1 g of wet weight of tissue corresponds to 1 ml of water, average concentration of agmatine in the brain is estimated at 1.5 –8.5 AM. Since agmatine is not uniformly distributed throughout the brain [30], local concentration of agmatine may be higher than these values. Furthermore, it has been reported that the activity of the agmatine-synthesizing enzyme arginine decarboxylase increases during postnatal (0 to 4 weeks old) developing cerebellum [12]. Therefore, it is possible that agmatine play a role in the development of the cerebellum after birth. In conclusion, we have found for the first time that agmatine stimulates the release of endogenous L-glutamate and causes cell death of cerebellar granule neurons in a medium with high K+. The effects of agmatine on neuronal survival are more complex than previously thought. Whether agmatine exerts neuroprotection or neurotoxicity depends on the circumstances. Our finding suggests that agmatine exerts neurotoxicity in epilepsy and ischemia, conditions in which synaptic K+ levels are elevated. Further investigations of molecular mechanisms underlying neuroprotective and neurotoxic actions of agmatine will give new insight into the role of agmatine in brain development and pathology.
References [1] K. Abe, H. Kimura, Amyloid h toxicity consists of a Ca2 +-independent early phase and a Ca2 +-dependent late phase, J. Neurochem. 67 (1996) 2074 – 2078. [2] K. Abe, Y. Abe, H. Saito, Evaluation of L-glutamate clearance capacity of cultured rat cortical astrocytes, Biol. Pharm. Bull. 23 (2000) 204 – 207. [3] F.W. Berman, T.F. Murray, Characterization of [3H]MK-801 binding
[4]
[5] [6]
[7] [8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19] [20]
[21]
[22]
[23]
[24]
to N-methyl-D-aspartate receptors in cultured rat cerebellar granule neurons and involvement in glutamate-mediated toxicity, J. Biochem. Toxicol. 11 (1996) 217 – 226. J. Blasi, E.R. Chapman, S. Yamasaki, T. Binz, H. Niemann, R. Jahn, Botulinum neurotoxin C1 blocks neurotransmitter release by means of cleaving HPC-1/syntaxin, EMBO J. 12 (1993) 4821 – 4828. T.H. Brown, P.F. Chapman, E.W. Kairiss, C.L. Keenan, Long-term synaptic potentiation, Science 242 (1988) 724 – 728. C.J. Carter, F. Noel, B. Scatton, Raised extracellular potassium relieves the blockade by magnesium of NMDA-induced cerebellar cyclic GMP production, Neurosci. Lett. 82 (1987) 201 – 205. D.W. Choi, Glutamate neurotoxicity and diseases of the nervous system, Neuron 1 (1988) 623 – 634. J.A. Cox, P.G. Lysko, R.C. Henneberry, Excitatory amino acid neurotoxicity at the N-methyl-D-aspartate receptor in cultured neurons: role of the voltage-dependent magnesium block, Brain Res. 499 (1989) 267 – 272. F. Dessi, C. Charriaut-Marlangue, M. Khrestchatisky, Y. Ben-Ari, Glutamate-induced neuronal death is not a programmed cell death in cerebellar culture, J. Neurochem. 60 (1993) 1953 – 1955. Y. Feng, A.E. Halaris, J.E. Piletz, Determination of agmatine in brain and plasma using high-performance liquid chromatography with fluorescence detection, J. Chromatogr., B 691 (1997) 277 – 286. V. Gallo, A. Kingsbury, R. Balaz, O.S. Jorgensen, The role of depolarization in the survival and differentiation of cerebellar granule cells in culture, J. Neurosci. 7 (1987) 2203 – 2213. G.M. Gilad, V.H. Gilad, J.M. Rabey, Arginine and ornithine decarboxylation in rodent brain: coincidental changes during development and after ischemia, Neurosci. Lett. 216 (1996) 33 – 36. G.M. Gilad, K. Salame, J.M. Rabey, V.H. Gilad, Agmatine treatment is neuroprotective in rodent brain injury models, Life Sci. 58 (1996) PL41 – PL46. S.D. Jeftinija, K.V. Jeftinija, G. Stefanovic, Cultured astrocytes express proteins involved in vesicular glutamate release, Brain Res. 750 (1997) 41 – 47. E.D. Kirson, C. Schirra, A. Konnerth, Y. Yaari, Early postnatal switch in magnesium sensitivity of NMDA receptors in rat CA1 pyramidal cells, J. Physiol. 521 (1999) 99 – 111. T. Kuner, R. Schoepfer, Multiple structural elements determine subunit specificity of Mg2 + block in NMDA receptor channels, J. Neurosci. 16 (1996) 3549 – 3558. G. Li, S. Regunathan, C.J. Barrow, J. Esharaghi, R. Cooper, D.J. Reis, Agmatine: an endogenous clonidine-displacing substance in the brain, Science 263 (1994) 966 – 969. G. Li, S. Regunathan, D.J. Reis, Agmatine is synthesized by a mitochondrial arginine decarboxylase in rat brain, Ann. N.Y. Acad. Sci. 763 (1995) 329 – 625. R.H. Loring, Agmatine act as an antagonist of neuronal nicotinic receptors, Br. J. Pharmacol. 99 (1990) 207 – 211. M.J. Lortie, W.F. Novontny, O.W. Peterson, V. Vallon, K. Malvey, M. Mendonca, J. Satriano, P. Insel, S.C. Thomson, R.C. Blantz, Agmatine, a bioactive metabolite of arginine, J. Clin. Invest. 97 (1996) 413 – 420. P.G. Lysko, J.A. Cox, A. Vigano, R.C. Henneberry, Excitatory amino acid neurotoxicity at the N-methyl-D-aspartate receptor in cultured neurons: pharmacological characterization, Brain Res. 499 (1989) 258 – 266. H. Manev, M. Favaron, R. Siman, A. Guidotti, E. Costa, Glutamate neurotoxicity is independent of calpain I inhibition in primary cultures of cerebellar granule cells, J. Neurochem. 57 (1991) 1288 – 1295. M.L. Mayer, G.L. Westbrook, P.B. Guthrie, Voltage-dependent block by Mg2 + of NMDA responses in spinal cord neurones, Nature 309 (1984) 261 – 263. A. Momiyama, D. Feldmeyer, S.G. Cull-Candy, Identification of a native low-conductance NMDA channel with reduced sensitivity to Mg2 + in rat central neurones, J. Physiol. 494 (1996) 479 – 492.
K. Abe et al. / Brain Research 990 (2003) 165–171 [25] C. Montecucco, G. Schiavo, Mechanism of action of tetanus and botulinum neurotoxins, Mol. Microbiol. 13 (1994) 1 – 8. [26] H. Monyer, R. Sprengel, R. Schoepfer, A. Herb, M. Higuchi, H. Lomeli, N. Burnashev, B. Sakmann, P.H. Seeburg, Heteromeric NMDA receptors: molecular and functional distinction of subtypes, Science 256 (1992) 1217 – 1221. [27] J. Morrissey, R. McCracken, S. Ishadoya, S. Klahr, Partial cloning and characterization of an arginine decarboxylase in the kidney, Kidney Int. 47 (1995) 1458 – 1461. [28] L. Nowak, P. Bregestovski, P. Ascher, A. Herbet, A. Prochiantz, Magnesium gates glutamate-activated channels in mouse central neurones, Nature 307 (1984) 462 – 465. [29] G.O. Olmos, N. DeGregorio-Rocasolano, M.P. Regalado, T. Gasull, M.A. Boronat, R. Trullas, A. Villarroel, J. Lerma, J.A. Gardi-Sevilla, Protection by imidazol(ine) drugs and agmatine of glutamate-induced neurotoxicity in cultured cerebellar granule cells through blockade of NMDA receptor, Br. J. Pharmacol. 127 (1999) 1317 – 1326. [30] K. Otake, D.A. Ruggiero, S. Regunathan, H. Wang, T.A. Milner, D.J. Reis, Regional localization of agmatine in the rat brain: an immunocytochemical study, Brain Res. 787 (1998) 1 – 14. [31] W. Raasch, S. Regunathan, G. Li, D.J. Reis, Agmatine, the bacterial amine, is widely distributed in mammalian tissues, Life Sci. 56 (1995) 2319 – 2330. [32] S. Regunathan, D.L. Feinstein, W. Raasch, D.J. Reis, Agmatine (decarboxylated arginine) is synthesized and stored in astrocytes, NeuroReport 6 (1995) 1897 – 1900. [33] D.J. Reis, X.C. Yang, T.A. Milner, Agmatine containing axon terminals in rat hippocampus form synapses on pyramidal cells, Neurosci. Lett. 250 (1998) 185 – 188.
171
[34] S.M. Rothman, J.W. Olney, Glutamate and the pathophysiology of hypoxic – ischemic brain damage, Ann. Neurol. 19 (1986) 105 – 111. [35] M. Sastre, S. Regunathan, D.J. Reis, Uptake of agmatine into rat brain synaptosomes: possible role of cation channels, J. Neurochem. 69 (1997) 2421 – 2426. [36] M. Schramm, S. Eimerl, E. Costa, Serum and depolarizing agents cause acute neurotoxicity in cultured cerebellar granule cells: role of glutamate receptor responsive to N-methyl-D-aspartate, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 1193 – 1197. [37] R. Schwarcz, A.C. Foster, E.D. French, W.O. Whetsell, C. Kohler, Excitotoxic models for neurodegenerative disorders, Life Sci. 35 (1984) 19 – 32. [38] A. Sener, P. Lebrun, F. Blachier, W.J. Malaisse, Stimulus-secretion coupling of arginine-induced insulin release. Insulinotropic action of agmatine, Biochem. Pharmacol. 38 (1989) 327 – 330. [39] R.M. Shepherd, M.N. Hashmi, C. Kane, P.E. Squires, M.J. Dunne, Elevation of cytosolic calcium by imidazolines in mouse islets of Langerhans: implications for stimulus – response coupling of insulin release, Br. J. Pharmacol. 119 (1996) 911 – 916. [40] G. Wang, O.S. Gorbatyuk, G. Dayanithi, W. Ouyang, J. Wang, T.A. Milner, S. Regunathan, D.J. Reis, Evidence for endogenous agmatine in hypothalamo-neurohypophysial tract and its modulation on vasopressin release and Ca2 + channels, Brain Res. 932 (2002) 25 – 36. [41] X.C. Yang, D.J. Reis, Agmatine selectively blocks the N-methyl-Daspartate subclass of glutamate receptor channels in rat hippocampal neurons, J. Pharmacol. Exp. Ther. 288 (1999) 544 – 549.
Research report
Agmatine induces glutamate release and cell death in cultured rat cerebellar granule neurons Kazuho Abe a,b,*, Yuzuru Abe a, Hiroshi Saito a a
Department of Chemical Pharmacology, Faculty of Pharmaceutical Sciences, The University of Tokyo, Tokyo 113-0033, Japan b Department of Pharmacology, School of Pharmacy, Hoshi University, Tokyo 142-8501, Japan Accepted 30 July 2003
Abstract We investigated the effect of agmatine on cell viability of rat cerebellar granule neurons in a high-K+ (27.5 mM) medium. Exposure of cultured rat cerebellar granule neurons to agmatine (200 – 800 AM) resulted in a significant decrease in cell viability. Agmatine-induced neuronal death began to occur 6 – 12 h after addition, and gradually progressed. The agmatine neurotoxicity was attenuated by N-methyl-Daspartate (NMDA) receptor antagonists and by enzymatic degradation of L-glutamate with glutamic pyruvic transaminase. Furthermore, a significant increase in extracellular L-glutamate concentration was detected before cell death occurred. In addition, agmatine-induced glutamate release and cell death were both blocked by pretreatment with botulinum toxin C, which is known to specifically inhibit the exocytosis. The agmatine neurotoxicity was not observed when extracellular K+ concentration was lower (10 mM). These results suggest that agmatine induces glutamate release through the exocytosis and thereby causes NMDA receptor-mediated neuronal death in conditions in which extracellular K+ concentrations are elevated. D 2003 Elsevier B.V. All rights reserved. Theme: Neurotransmitters, modulators, transporters, and receptors Topic: Other neurotransmitters Keywords: Agmatine; Neurotoxicity; Glutamate; NMDA receptor; Exocytosis; Cerebellar granule neuron
1. Introduction Agmatine is an endogenous substance that is synthesized from arginine by arginine decarboxylase and metabolized to putrescine by agmatinase [17,18,27]. Agmatine is present in the brain of mammals including the rat, bovine [17] and human [20,31]. The level of agmatine in rat brain has been reported 0.2 – 0.4 Ag/g tissue by mass spectroscopy [17] and 0.331 –1.105 Ag/g tissue by high-performance liquid chromatography [10]. The presence of agmatine in astrocytes and neurons has been demonstrated by immunohistochemical examination with anti-agmatine antibody [30,32]. Electron microscopic examination has demonstrated that agmatine is present in axon terminals associated with synaptic vesicles [30,33]. In addition, agmatine is taken up into synaptosomes via a Na+-independent transport * Corresponding author. Department of Pharmacology, Hoshi University, 2-4-41 Ebara, Shinagawa, Tokyo 142-8501, Japan. Tel.: +81-3-54985785; fax: +81-3-5498-5787. E-mail address: [email protected] (K. Abe). 0006-8993/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0006-8993(03)03454-1
system [35]. Furthermore, agmatine has been reported to act as a ligand of the imidazoline receptor [17], block nicotinic receptor function in the chick retina and the rat superior cervical ganglion [19], block the N-methyl-D-aspartate (NMDA) type of glutamate receptors in rat hippocampal neurons [41] or block the voltage-gated Ca2 + channel in neurohypophysial nerve terminals [40]. These lines of evidence suggest the possible role of agmatine as a neurotransmitter or neuromodulator in the brain. L-Glutamate functions as a major excitatory neurotransmitter in the central nervous system, but excessive stimulation by L-glutamate is known to cause excitoxic neuronal damages through the overstimulation of the NMDA receptor, which may underlie the abnormal neuronal degeneration observed following hypoxia, ischemia, hypoglycemia and seizures [7,34,37]. Since agmatine has been reported to block the NMDA receptor [41], it is possible that agmatine modulates L-glutamate neurotoxicity. Indeed, Olmos et al. [29] have recently demonstrated that agmatine protects cultured rat cerebellar granule neurons from L-glutamate toxicity in a medium without Mg2 + and glucose. Further-
166
K. Abe et al. / Brain Research 990 (2003) 165–171
more, neuroprotective effects of agmatine have been reported in in vivo models of neurotoxic and ischemic brain injuries [13]. To further explore possible interaction of agmatine and Lglutamate in the brain, we also investigated the effect of agmatine on the neuronal survival by using primary culture of neonatal rat cerebellar granule neurons, the same preparation as Olmos et al. [29] have previously used. Other reasons that we chose to use primary cultured cerebellar granule neurons were (1) cerebellar granule neurons are easily cultured with high purity in a medium containing an elevated level of K+ [11], (2) primary culture of cerebellar granule neurons is a well characterized model for the study of NMDA receptor-mediated glutamate neurotoxicity [3,9,21,22,36], and (3) the activity of the agmatine-synthesizing enzyme arginine decarboxylase increases during postnatal developing cerebellum [12]. As a result, we found that agmatine exerts toxic effects for cultured cerebellar granule neurons, contrary to the observation by Olmos et al. [29]. Whether agmatine exerts neuroprotection or neurotoxicity may depend on conditions. Here we report possible mechanisms underlying a neurotoxic feature of agmatine.
2. Materials and methods 2.1. Chemicals Agmatine sulfate, dizocilpine, 2-amino-5-phosphonovalerate (APV) and glutamic pyruvic transaminase were purchased from Sigma (St. Louis, MO, USA). 2,3-Dihydroxy-6-nitro-7-sulphamoylbenzo[ f]-quinoxaline (NBQX), (RS)-a-methyl-4-carboxyphenylglycine (MCPG) and (RS)a-methyl-4-phosphonophenylglycine (MPPG) were purchased from Tocris Cookson (Bristol, UK). Botulinum toxin C was purchased from Wako (Osaka, Japan). 2.2. Cell culture Primary culture of cerebellar granule neurons was prepared according to the method by Gallo et al. [11]. Briefly, the cerebella obtained from 8-day-old Wistar rats were mechanically chopped, and cells were dissociated by trypsinization and pipetting. The dissociated cells were plated on poly-L-lysine-coated 96-well microtiter dishes at a density of 300,000 cells/cm2 in a modified Eagle’s medium containing 1.8 mM glutamine, 25 mM KCl, 91 Ag/ml gentamycin and 10% fetal bovine serum. To arrest the growth of nonneuronal cells, 1-h-arabinosylcytosine (10 AM) was added to the culture medium 24 h after the plating. The cultures were incubated at 37 jC with 5% CO2 for 7– 8 days, during which the medium was not changed to avoid the detachment of cells from culture dishes. In this culture, more than 95% of the cells were neurons as defined by staining with microtubule-associated protein-2. At days 7– 8, all the medium was carefully removed and replaced with a serum-
free Eagle’s medium containing 2 mM glutamine, 27.5 mM KCl, 100 Ag/ml gentamycin, 100 Ag/ml transferrin, 5 Ag/ml insulin, 20 nM progesterone and no L-glutamate. In a part of experiments, the medium was replaced with a serum-free Eagle’s medium containing different concentrations (10, 20 or 27.5 mM) of KCl to investigate whether agmatine neurotoxicity depends on extracellular K+ concentration. In any cases, the medium contained 1.8 mM Ca2 + and 0.8 mM Mg2 +. Agmatine or test reagents were added to the culture medium at this time. 2.3. Determination of cell viability Cell viability was evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction and lactate dehydrogenase (LDH) release assay. MTT reduction assay measures the redox activity of living cells, while LDH assay measures the activity of LDH released into the medium from dead cells. MTT reduction assay was performed as described previously [1]. Briefly, the cells were incubated with 0.25 mg/ml MTT at 37 jC, and a solution containing 50% dimethylformamide and 20% sodium dodecyl sulfate (pH 4.8) was added to terminate the reaction and to solubilize MTT formazan. The amount of MTT formazan produced was determined by measuring absorbance with a microplate reader at a test wavelength of 570 nm and a reference wavelength of 655 nm. LDH release assay was performed with a colorimetric LDH assay kit (Promega, Madison, WI, USA) according to the manufacturer’s instruction. Total cellular LDH activity was determined by solubilizing the cells with 0.2% Triton X-100. 2.4. Determination of L-glutamate concentration The concentration of L-glutamate was determined by a colorimetric method [2]. Briefly, 50 Al of culture supernatant was transferred to 96-well microtiter plates and was mixed with 50 Al of a substrate mixture consisting of 20 U/ml glutamate dehydrogenase, 2.5 mg/ml hnicotinamide adenine dinucleotide, 0.25 mg/ml MTT, 100 nM 1-methoxyphenazine methosulfate and 0.1% (v/ v) Triton X-100 in 0.2 M Tris – HCl buffer (pH 8.2). The reaction proceeded for 10 min at 37 jC and was stopped by adding 100 Al of a solution containing 50% dimethylformamide and 20% sodium dodecyl sulfate (pH 4.7). In this reaction, MTT (yellow) is converted into MTT formazan (purple) in proportion to L-glutamate concentration. The amount of MTT formazan produced was determined by measuring absorbance with a microplate reader at a test wavelength of 570 nm and a reference wavelength of 655 nm. The standard curve was constructed in each assay using cell-free medium containing known concentrations of L-glutamate. The concentration of L-glutamate in samples was estimated from the standard curve.
K. Abe et al. / Brain Research 990 (2003) 165–171
167
3. Results In a serum-free Eagle’s medium with 27.5 mM K+, agmatine caused a decrease in MTT reduction and an increase in LDH release in a concentration- and timedependent manner (Fig. 1). The minimum effective concentration of agmatine was 100 –200 AM (Fig. 1A and B). Agmatine-induced cell death began to occur 6– 12 h after addition, and gradually progressed (Fig. 1C). To examine
Fig. 2. Influences of glutamate receptor blockers on agmatine-induced cell death in cultured rat cerebellar granule neurons. Cells were exposed to 200 or 400 AM agmatine in the absence or presence of glutamate receptor blockers (A: 10 AM NMQX, 1 AM dizocilpine; B: 100 AM APV, 1 mM MCPG, 500 AM MPPG) for 24 h. Cell death was assessed by the LDH release assay. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. none, ##P < 0.01 vs. agmatine alone; Tukey’s test.
Fig. 1. Effect of agmatine on the viability of cultured rat cerebellar granule neurons. (A, B) Concentration-dependency. The cells were exposed to 50 – 800 AM agmatine for 24 h. (C) Time-dependency. The cells were exposed to none (o), 200 AM agmatine (n) or 400 AM agmatine (E) for 3, 6, 12 or 30 h. Cell viability was assessed by the MTT reduction assay (A) or the LDH release assay (B, C). Cellular MTT reduction activity was expressed as percentage of the control (no addition of agmatine). LDH activity in the culture supernatant was expressed as the percentage of total LDH activity. Data are represented as the means F S.E.M. of the determinations on five separate cultures. *P < 0.05, **P < 0.01 vs. none; Dunnett’s test.
if the toxicity is a specific effect of agmatine itself, the effects of the agmatine precursor arginine and the agmatine metabolite putrescine were also investigated. However, neither arginine nor putrescine showed significant effect on cell viability of cultured rat cerebellar granule neurons at least at concentrations up to 800 AM (n = 5, data not shown). To examine possible involvement of L-glutamate in agmatine-induced neuronal death, three series of experiments were performed. First, the influences of glutamate receptor antagonists were investigated. Glutamate receptor antagonists tested were NBQX (an AMPA receptor antagonist), dizocilpine (an NMDA receptor channel blocker), APV (an NMDA receptor antagonist), MCPG (a nonselective group I/group II metabotropic glutamate receptor antagonist) and MPPG (a group III metabotropic glutamate receptor antagonist). As shown in Fig. 2, agmatine neurotoxicity was abolished by APV or dizocilpine, but was not affected by NBQX, MCPG or MPPG. Second, it was tested if agmatine neurotoxicity is affected by enzymatic degradation of L-glutamate with glutamic pyruvic transaminase. Agmatine-induced neuronal death was abolished by the presence of 50 U/ml glutamic pyruvic transaminase and 3 mM pyruvate (Fig. 3). Third, concen-
168
K. Abe et al. / Brain Research 990 (2003) 165–171
Addition of arginine (50 – 800 AM) or putrescine (50 – 800 AM) induced no significant change in extracellular Lglutamate concentration (n = 5, data not shown). To examine if agmatine-induced L-glutamate release is mediated by the exocytosis, we investigated the influence of botulinum toxin C, an exocytosis inhibitor. As shown in Fig. 5A and B, agmatine-induced L-glutamate release and neuronal death were both abolished by pretreatment with botulinum toxin C. The same treatment did not affect neuronal death induced by exogenously applied L-glutamate (Fig. 5C), indicating that this treatment affects neither L-glutamate Fig. 3. Influence of enzymatic degradation of L-glutamate on agmatineinduced cell death in cultured rat cerebellar granule neurons. Cells were exposed to 200 or 400 AM agmatine in the absence (white columns) or presence of 50 U/ml glutamic pyruvic transaminase and 3 mM pyruvate (black columns) for 24 h. Cell viability was assessed by the MTT reduction assay, because addition of pyruvate may interfere with the measurement of LDH activity. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. none, ##P < 0.01 vs. agmatine alone; Tukey’s test.
tration of L-glutamate in the extracellular medium was measured before neuronal death occurred. Addition of 400 AM agmatine induced a significant increase in extracellular L-glutamate concentration, which began to occur 1.5 h after addition of agmatine and preceded an increase in LDH release, an index of cell membrane damage (Fig. 4).
Fig. 4. Changes in extracellular L-glutamate concentration and LDH release following addition of agmatine in rat cerebellar granule neuron cultures. Cells were exposed to none (o) or 400 AM agmatine ( ) for 0, 30, 90, 180, 270 or 360 min. L-Glutamate concentration (A) and LDH activity (B) were determined in the supernatant collected from the same culture. Data are the means F S.E.M. of five separate observations. *P < 0.05, **P < 0.01 vs. none; Student’s t-test.
.
Fig. 5. Influences of botulinum toxin C on agmatine-induced L-glutamate release (A), agmatine-induced cell death (B) and L-glutamate-induced cell death (C) in cultured rat cerebellar granule neurons. Cells were pretreated with none (white columns) or 15 pM botulinum toxin C (black columns) for 2 h, and then exposed to 200 – 400 AM agmatine (A, B) or 30 – 100 AM glutamate (C). Extracellular glutamate concentration (A) and LDH release (B, C) were determined 3 and 24 h after addition of agmatine, respectively. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. none, ##P < 0.01 vs. agmatine alone; Tukey’s test.
K. Abe et al. / Brain Research 990 (2003) 165–171
Fig. 6. Agmatine neurotoxicity depends on extracellular K+ concentration. Cultured rat cerebellar granule neurons were exposed to none (white columns) or 400 AM agmatine (gray columns) in a medium with 27.5, 20 or 10 mM K+, and LDH release was determined 24 h after addition of agmatine. Data are the means F S.E.M. of five separate observations. **P < 0.01 vs. control (none in 27.5 mM K+), ##P < 0.01 vs. none; Tukey’s test.
metabolism nor mechanisms involved in L-glutamate-induced neuronal death. The above data were contrary to previous studies demonstrating the neuroprotective role of agmatine. One of the differences in experimental conditions between a previous study by Olmos et al. [29] and our present study is K+ concentration in the extracellular medium. To examine if agmatine neurotoxicity depends on extracellular K+ concentration, the effects of agmatine were compared in the medium with different concentrations (10, 20 or 27.5 mM) of K+. Agmatine (400 AM) caused a significant increase in LDH release in serum-free Eagle’s medium with 20 mM K+, but the effect was smaller than that in serum-free Eagle’s medium with 27.5 mM K+ (Fig. 6). When the medium was switched to serum-free Eagle’s medium with 10 mM K+, basal LDH release was significantly increased, probably due to low K+induced apoptotic cell death. In this condition (10 mM K+), agmatine (400 AM) caused no significant change in LDH release (Fig. 6).
4. Discussion In the present study, we found that agmatine induced cell death of cultured rat cerebellar granule neurons in Eagle’s medium with high K+, which has been conventionally used to culture this type of cells with high purity [11]. Furthermore, the agmatine-induced neuronal death was preceded by accumulation of extracellular L-glutamate and was abolished by the presence of NMDA receptor antagonists or by enzymatic degradation of L-glutamate. The agmatine-induced neuronal death and L-glutamate accumulation were not mimicked by the agmatine precursor arginine and the agmatine metabolite putrescine, indicating that the effects are caused by agmatine itself, but not by its precursor or metabolite. These results suggest that agmatine specifically induces the release of
169
endogenous L-glutamate and thereby causes neuronal death through stimulation of NMDA receptors. In addition, agmatine-induced L-glutamate release and neuronal death were abolished by pretreatment with botulinum toxin C, which inhibits the exocytosis of synaptic vesicles by cleaving the synaptic protein syntaxin in nerve terminals [4,14,25], suggesting that agmatine-induced Lglutamate release is mediated by exocytotic processes. Agmatine has been reported to block ATP-sensitive K+ channels in pancreatic h-cells [39], leading to the stimulation of insulin release [38]. To the best of our knowledge, there has been no report concerning the effect of agmatine on Lglutamate release. Although agmatine has been reported to block the voltage-gated Ca2 + channel in neurohypophysial terminals [40], it is unlikely that agmatine blocks Ca2 + influx in presynaptic membranes of cerebellar granule neurons, which should result in the inhibition of L-glutamate release. Agmatine may directly or indirectly act on the exocytotic machinery downstream of Ca2 + influx. Further investigations are underway in our laboratory to elucidate molecular mechanisms by which agmatine stimulates the release of L-glutamate from neuronal cells. Our present study revealed a neurotoxic feature of agmatine, contrary to previous studies demonstrating the neuroprotective role of agmatine. In the report by Olmos et al. [29], cultured cerebellar granule neurons were preincubated in Mg2 +- and glucose-free Locke-HEPES buffer with 5.6 mM K+ to be deprived of energy resources and then exposed to agmatine (100 or 500 AM) for 45 min. In our present study, cultured cerebellar granule neurons were maintained in serum-free Eagle’s medium with 27.5 mM K+ and exposed to 50– 800 AM agmatine for 3– 30 h. Among several differences in experimental conditions between Olmos et al. [29] and us, extracellular K+ concentration may be a key factor in agmatine neurotoxicity. In fact, we found that agmatine exerted neurotoxicity in medium with high K+ (27.5 or 20 mM) but not low K+ (10 mM). Whether agmatine exerts neuroprotection or neurotoxicity may be determined by the balance among the amount of L-glutamate release stimulated by agmatine, the degree of NMDA receptor blockade by agmatine and the susceptibility of neurons to L-glutamate toxicity. It has been reported that the potency of agmatine in blocking the NMDA receptor is attenuated at depolarized membrane potentials [41]. Furthermore, in our preliminary experiments, toxicity of exogenously applied L-glutamate was enhanced by increasing extracellular K+ concentration in cultured cerebellar granule neurons (unpublished data). Thus, in a medium with high K+, L-glutamate-mediated toxic effect of agmatine may predominate over its neuroprotective effect through NMDA receptor blockade. NMDA receptor channels are blocked by physiological concentrations (approximately 1 mM) of Mg2 + [23,28]. In glutamatergic synapses, L-glutamate opens AMPA type of glutamate receptor channels followed by depolarization of postsynaptic membranes, which in turn relieves Mg2 + block of NMDA receptor channels, leading Ca2 + influx sufficient
170
K. Abe et al. / Brain Research 990 (2003) 165–171
for synaptic plasticity or neuronal death [5,8]. In the present study, agmatine caused neuronal death in a medium with 0.8 mM Mg2 +, and the effect was abolished by NMDA receptor blockers, but was not affected by AMPA receptor blocker, indicating that the opening of NMDA receptor channels by agmatine occurs in the presence of Mg2 + and does not require AMPA receptor-mediated membrane depolarization. There are two possible explanations for this discrepancy. First, Ca2 + permeability and Mg2 + sensitivity of NMDA receptor channels are different with the constitution of NMDA receptor subunits, which may account for functional differences of NMDA receptor channels among brain regions, developmental stages or culture conditions [15,16,24,26]. Second, Mg2 + block of NMDA receptor channels is relieved by raised extracellular K+ [6]. In the present study, cerebellar granule neurons would be depolarized by 27.5 mM K+, and Lglutamate could induce Ca2 + influx through NMDA receptor channels without AMPA receptor-mediated depolarization. The minimum effective concentration of agmatine in inducing cell death was 100 – 200 AM in the present study. The level of agmatine in rat brain has been reported 0.2 – 0.4 Ag/g tissue by mass spectroscopy [17] and 0.331– 1.105 Ag/ g tissue by high-performance liquid chromatography [10]. On the assumption that 1 g of wet weight of tissue corresponds to 1 ml of water, average concentration of agmatine in the brain is estimated at 1.5 –8.5 AM. Since agmatine is not uniformly distributed throughout the brain [30], local concentration of agmatine may be higher than these values. Furthermore, it has been reported that the activity of the agmatine-synthesizing enzyme arginine decarboxylase increases during postnatal (0 to 4 weeks old) developing cerebellum [12]. Therefore, it is possible that agmatine play a role in the development of the cerebellum after birth. In conclusion, we have found for the first time that agmatine stimulates the release of endogenous L-glutamate and causes cell death of cerebellar granule neurons in a medium with high K+. The effects of agmatine on neuronal survival are more complex than previously thought. Whether agmatine exerts neuroprotection or neurotoxicity depends on the circumstances. Our finding suggests that agmatine exerts neurotoxicity in epilepsy and ischemia, conditions in which synaptic K+ levels are elevated. Further investigations of molecular mechanisms underlying neuroprotective and neurotoxic actions of agmatine will give new insight into the role of agmatine in brain development and pathology.
References [1] K. Abe, H. Kimura, Amyloid h toxicity consists of a Ca2 +-independent early phase and a Ca2 +-dependent late phase, J. Neurochem. 67 (1996) 2074 – 2078. [2] K. Abe, Y. Abe, H. Saito, Evaluation of L-glutamate clearance capacity of cultured rat cortical astrocytes, Biol. Pharm. Bull. 23 (2000) 204 – 207. [3] F.W. Berman, T.F. Murray, Characterization of [3H]MK-801 binding
[4]
[5] [6]
[7] [8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19] [20]
[21]
[22]
[23]
[24]
to N-methyl-D-aspartate receptors in cultured rat cerebellar granule neurons and involvement in glutamate-mediated toxicity, J. Biochem. Toxicol. 11 (1996) 217 – 226. J. Blasi, E.R. Chapman, S. Yamasaki, T. Binz, H. Niemann, R. Jahn, Botulinum neurotoxin C1 blocks neurotransmitter release by means of cleaving HPC-1/syntaxin, EMBO J. 12 (1993) 4821 – 4828. T.H. Brown, P.F. Chapman, E.W. Kairiss, C.L. Keenan, Long-term synaptic potentiation, Science 242 (1988) 724 – 728. C.J. Carter, F. Noel, B. Scatton, Raised extracellular potassium relieves the blockade by magnesium of NMDA-induced cerebellar cyclic GMP production, Neurosci. Lett. 82 (1987) 201 – 205. D.W. Choi, Glutamate neurotoxicity and diseases of the nervous system, Neuron 1 (1988) 623 – 634. J.A. Cox, P.G. Lysko, R.C. Henneberry, Excitatory amino acid neurotoxicity at the N-methyl-D-aspartate receptor in cultured neurons: role of the voltage-dependent magnesium block, Brain Res. 499 (1989) 267 – 272. F. Dessi, C. Charriaut-Marlangue, M. Khrestchatisky, Y. Ben-Ari, Glutamate-induced neuronal death is not a programmed cell death in cerebellar culture, J. Neurochem. 60 (1993) 1953 – 1955. Y. Feng, A.E. Halaris, J.E. Piletz, Determination of agmatine in brain and plasma using high-performance liquid chromatography with fluorescence detection, J. Chromatogr., B 691 (1997) 277 – 286. V. Gallo, A. Kingsbury, R. Balaz, O.S. Jorgensen, The role of depolarization in the survival and differentiation of cerebellar granule cells in culture, J. Neurosci. 7 (1987) 2203 – 2213. G.M. Gilad, V.H. Gilad, J.M. Rabey, Arginine and ornithine decarboxylation in rodent brain: coincidental changes during development and after ischemia, Neurosci. Lett. 216 (1996) 33 – 36. G.M. Gilad, K. Salame, J.M. Rabey, V.H. Gilad, Agmatine treatment is neuroprotective in rodent brain injury models, Life Sci. 58 (1996) PL41 – PL46. S.D. Jeftinija, K.V. Jeftinija, G. Stefanovic, Cultured astrocytes express proteins involved in vesicular glutamate release, Brain Res. 750 (1997) 41 – 47. E.D. Kirson, C. Schirra, A. Konnerth, Y. Yaari, Early postnatal switch in magnesium sensitivity of NMDA receptors in rat CA1 pyramidal cells, J. Physiol. 521 (1999) 99 – 111. T. Kuner, R. Schoepfer, Multiple structural elements determine subunit specificity of Mg2 + block in NMDA receptor channels, J. Neurosci. 16 (1996) 3549 – 3558. G. Li, S. Regunathan, C.J. Barrow, J. Esharaghi, R. Cooper, D.J. Reis, Agmatine: an endogenous clonidine-displacing substance in the brain, Science 263 (1994) 966 – 969. G. Li, S. Regunathan, D.J. Reis, Agmatine is synthesized by a mitochondrial arginine decarboxylase in rat brain, Ann. N.Y. Acad. Sci. 763 (1995) 329 – 625. R.H. Loring, Agmatine act as an antagonist of neuronal nicotinic receptors, Br. J. Pharmacol. 99 (1990) 207 – 211. M.J. Lortie, W.F. Novontny, O.W. Peterson, V. Vallon, K. Malvey, M. Mendonca, J. Satriano, P. Insel, S.C. Thomson, R.C. Blantz, Agmatine, a bioactive metabolite of arginine, J. Clin. Invest. 97 (1996) 413 – 420. P.G. Lysko, J.A. Cox, A. Vigano, R.C. Henneberry, Excitatory amino acid neurotoxicity at the N-methyl-D-aspartate receptor in cultured neurons: pharmacological characterization, Brain Res. 499 (1989) 258 – 266. H. Manev, M. Favaron, R. Siman, A. Guidotti, E. Costa, Glutamate neurotoxicity is independent of calpain I inhibition in primary cultures of cerebellar granule cells, J. Neurochem. 57 (1991) 1288 – 1295. M.L. Mayer, G.L. Westbrook, P.B. Guthrie, Voltage-dependent block by Mg2 + of NMDA responses in spinal cord neurones, Nature 309 (1984) 261 – 263. A. Momiyama, D. Feldmeyer, S.G. Cull-Candy, Identification of a native low-conductance NMDA channel with reduced sensitivity to Mg2 + in rat central neurones, J. Physiol. 494 (1996) 479 – 492.
K. Abe et al. / Brain Research 990 (2003) 165–171 [25] C. Montecucco, G. Schiavo, Mechanism of action of tetanus and botulinum neurotoxins, Mol. Microbiol. 13 (1994) 1 – 8. [26] H. Monyer, R. Sprengel, R. Schoepfer, A. Herb, M. Higuchi, H. Lomeli, N. Burnashev, B. Sakmann, P.H. Seeburg, Heteromeric NMDA receptors: molecular and functional distinction of subtypes, Science 256 (1992) 1217 – 1221. [27] J. Morrissey, R. McCracken, S. Ishadoya, S. Klahr, Partial cloning and characterization of an arginine decarboxylase in the kidney, Kidney Int. 47 (1995) 1458 – 1461. [28] L. Nowak, P. Bregestovski, P. Ascher, A. Herbet, A. Prochiantz, Magnesium gates glutamate-activated channels in mouse central neurones, Nature 307 (1984) 462 – 465. [29] G.O. Olmos, N. DeGregorio-Rocasolano, M.P. Regalado, T. Gasull, M.A. Boronat, R. Trullas, A. Villarroel, J. Lerma, J.A. Gardi-Sevilla, Protection by imidazol(ine) drugs and agmatine of glutamate-induced neurotoxicity in cultured cerebellar granule cells through blockade of NMDA receptor, Br. J. Pharmacol. 127 (1999) 1317 – 1326. [30] K. Otake, D.A. Ruggiero, S. Regunathan, H. Wang, T.A. Milner, D.J. Reis, Regional localization of agmatine in the rat brain: an immunocytochemical study, Brain Res. 787 (1998) 1 – 14. [31] W. Raasch, S. Regunathan, G. Li, D.J. Reis, Agmatine, the bacterial amine, is widely distributed in mammalian tissues, Life Sci. 56 (1995) 2319 – 2330. [32] S. Regunathan, D.L. Feinstein, W. Raasch, D.J. Reis, Agmatine (decarboxylated arginine) is synthesized and stored in astrocytes, NeuroReport 6 (1995) 1897 – 1900. [33] D.J. Reis, X.C. Yang, T.A. Milner, Agmatine containing axon terminals in rat hippocampus form synapses on pyramidal cells, Neurosci. Lett. 250 (1998) 185 – 188.
171
[34] S.M. Rothman, J.W. Olney, Glutamate and the pathophysiology of hypoxic – ischemic brain damage, Ann. Neurol. 19 (1986) 105 – 111. [35] M. Sastre, S. Regunathan, D.J. Reis, Uptake of agmatine into rat brain synaptosomes: possible role of cation channels, J. Neurochem. 69 (1997) 2421 – 2426. [36] M. Schramm, S. Eimerl, E. Costa, Serum and depolarizing agents cause acute neurotoxicity in cultured cerebellar granule cells: role of glutamate receptor responsive to N-methyl-D-aspartate, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 1193 – 1197. [37] R. Schwarcz, A.C. Foster, E.D. French, W.O. Whetsell, C. Kohler, Excitotoxic models for neurodegenerative disorders, Life Sci. 35 (1984) 19 – 32. [38] A. Sener, P. Lebrun, F. Blachier, W.J. Malaisse, Stimulus-secretion coupling of arginine-induced insulin release. Insulinotropic action of agmatine, Biochem. Pharmacol. 38 (1989) 327 – 330. [39] R.M. Shepherd, M.N. Hashmi, C. Kane, P.E. Squires, M.J. Dunne, Elevation of cytosolic calcium by imidazolines in mouse islets of Langerhans: implications for stimulus – response coupling of insulin release, Br. J. Pharmacol. 119 (1996) 911 – 916. [40] G. Wang, O.S. Gorbatyuk, G. Dayanithi, W. Ouyang, J. Wang, T.A. Milner, S. Regunathan, D.J. Reis, Evidence for endogenous agmatine in hypothalamo-neurohypophysial tract and its modulation on vasopressin release and Ca2 + channels, Brain Res. 932 (2002) 25 – 36. [41] X.C. Yang, D.J. Reis, Agmatine selectively blocks the N-methyl-Daspartate subclass of glutamate receptor channels in rat hippocampal neurons, J. Pharmacol. Exp. Ther. 288 (1999) 544 – 549.