Lithium protection against glutamate ... - Wiley Online Library

0 downloads 0 Views 377KB Size Report
cortical neurons with LiCl time-dependently suppressed excitotoxicity with .... ogies, Rockville, MD, USA) and were plated at a density of. 4.2 × 105 cells/cm2 on ...
Journal of Neurochemistry, 2002, 80, 589±597

Lithium protection against glutamate excitotoxicity in rat cerebral cortical neurons: involvement of NMDA receptor inhibition possibly by decreasing NR2B tyrosine phosphorylation Ryota Hashimoto,* Christopher Hough,  Takanobu Nakazawa,à Tadashi Yamamotoà and De-Maw Chuang* *Section on Molecular Neurobiology, Mood and Anxiety Disorder Program, National Institute of Mental Health, National Institutes of Health, Bethesda, Maryland, USA  Department of Psychiatry, Uniformed Services University of the Health Sciences, Bethesda, Maryland, USA àDepartment of Oncology, Institute of Medical Science, University of Tokyo, Tokyo, Japan

Abstract The therapeutic mechanisms of lithium for treating bipolar mood disorder remain poorly understood. Recent studies demonstrate that lithium has neuroprotective actions against a variety of insults. Here, we studied neuroprotective effects of lithium against excitotoxicity in cultured cerebral cortical neurons. Glutamate-induced excitotoxicity in cortical neurons was exclusively mediated by NMDA receptors. Pre-treatment of cortical neurons with LiCl time-dependently suppressed excitotoxicity with maximal protection after 6 days of pretreatment. Signi®cant protection was observed at the therapeutic and subtherapeutic concentration of 0.2±1.6 mM LiCl with almost complete protection at 1 mM. Neuroprotection was also elicited by valproate, another major mood-stabilizer. The neuroprotective effects of lithium coincided with inhi-

bition of NMDA receptor-mediated calcium in¯ux. Lithium pre-treatment did not alter total protein levels of NR1, NR2A and NR2B subunits of NMDA receptors. However, it did markedly reduce the level of NR2B phosphorylation at Tyr1472 and this was temporally associated with its neuroprotective effect. Because NR2B tyrosine phosphorylation has been positively correlated with NMDA receptor-mediated synaptic activity and excitotoxicity, the suppression of NR2B phosphorylation by lithium is likely to result in the inactivation of NMDA receptors and contributes to neuroprotection against excitotoxicity. This action could also be relevant to its clinical ef®cacy for bipolar patients. Keywords: bipolar disorder, excitotoxicity, lithium, neuroprotection, NMDA receptor, phosphorylation. J. Neurochem. (2002) 80, 589±597.

Lithium has been one of the primary drugs used to treat bipolar mood disorder as it was introduced into psychiatry over half a century ago. The therapeutic effect of lithium requires long-term treatment, occurs at a narrow dose-range and is not immediately reversed after discontinuation of the drug. The therapeutic mechanisms of lithium are still unclear; however, some prominent molecular and cellular actions of this drug have been identi®ed. These include its ability to inhibit key intracellular signaling kinases and phosphatases such as glycogen synthase kinase-3 and inositol monophosphatase and to affect transcriptional activity and gene expression (for review see Jope 1999; Manji and Lenox 2000; Phiel and Klein 2001). However, it remains unclear as to whether these actions are related to the clinical ef®cacy of

this drug, especially considering that many of these effects are acute and occur at supratherapeutic dose ranges.

Received October 4, 2001; revised manuscript received November 8, 2001; accepted November 9, 2001. Address correspondence and reprint requests to De-Maw Chuang, PhD, Section on Molecular Neurobiology, Mood and Anxiety Disorder Program, National Institute of Mental Health, National Institutes of Health, Bldg. 10, Rm. 4C-206, 10 Center Dr, MSC 1363, Bethesda, MD 20892±1363, USA. E-mail: [email protected] Abbreviations used: AP-5, D-2-amino-5-phosphonopentanoate; LDH, lactate dehydrogenase; LTP, long-term potentiation; MTT, 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; NBQX, 2,3-dihydro-6-nitro-7-sulfamoyl-benzo(f )quinoxaline; PKC, protein kinase C.

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

589

590 R. Hashimoto et al.

Recently, lithium was reported to show neuroprotective effects against apoptosis induced by a variety of insults in cultured neurons and neurally related cell lines. These apoptotic insults include withdrawal of growth factors (Bhat et al. 2000), deprivation of high potassium (D'Mello et al. 1994), exposure to heat shock (Bijur et al. 2000), and treatment with high doses of anticonvulsants (Nonaka et al. 1998b). In experimental rodent models, lithium reduces brain injury induced by irradiation (Inouye et al. 1995), focal ischemia (Nonaka and Chuang 1998), and excitotoxin administrations (Pascual and Gonzalez 1995; Sparapani et al. 1997; Wei et al. 2001). Using primary cultures of cerebellar granule cells, we found that lithium markedly reduces apoptosis induced by glutamate excitotoxicity, and this neuroprotective action is associated with inactivation of NMDA receptors (Nonaka et al. 1998a). However, in these studies the maximal effects of lithium occur at a supratherapeutic concentration of more than 3 mM and the drug selectivity for the neuroprotection had not been examined. Moreover, the mechanisms underlying lithium-induced inhibition of NMDA receptor activity await elucidation. Increasing evidence supports the notion that the level of tyrosine phosphorylation of NR2 subunit of NMDA receptor is linked to the receptor activity. The Src family tyrosine kinases, which bind to scaffolding proteins in the NMDA receptor complex, mediate NR2A and NR2B phosphorylation with a concomitant increase in NMDA receptormediated excitatory post-synaptic currents (Ali and Salter 2001). There is also an increase in tyrosine phosphorylation of NR2B subunit after induction of long-term potentiation (LTP) in the hippocampus (Rostas et al. 1996) and after taste learning in the insular cortex of rats (Rosenblum et al. 1997). Transient ischemia has been shown to induce a differential increase in tyrosine phosphorylation of NR2A and 2B and enhances their interactions with the Src homology 2 domains of Src and Fyn in the hippocampus (Takagi et al. 1997, 1999). The present study was undertaken to examine the neuroprotective effects of lithium in primary cultures of rat cerebral cortical neurons. The purposes of this study were to: (i) characterize excitotoxin-induced cell death in the cortical neuronal cultures; (ii) demonstrate and characterize dose- and time-dependent lithium neuroprotection against excitotoxicity in these cultures; (iii) elucidate the ion and drug selectivity for the observed neuroprotection; and (iv) explore the role of changes in tyrosine phosphorylation of NMDA receptor subunits in mediating the receptor inactivation and neuroprotection.

Materials and methods Animals and chemicals All procedures employing experimental rats were performed in compliance with National Institutes of Health Guidelines for the

care and use of laboratory animals. All chemicals were obtained from Sigma Chemical Co. (St Louis, MO, USA). Primary cultures of rat cerebral cortical neurons and drug treatments Primary cultures of rat cerebral cortical neurons were prepared from 17-day-old-embryonic rats and cultured as described previously (Kashiwagi et al. 1998; Hashimoto et al. 2000). Brie¯y, cortices were dissected from embryonic brain, and meninges were removed from the tissues. The cells were dissociated by trypsinization and trituration, followed by DNase treatment. The dissociated cells were resuspended in serum-free B27/neurobasal medium (Life Technologies, Rockville, MD, USA) and were plated at a density of 4.2 ´ 105 cells/cm2 on polyethyleneimine-pre-coated plates, dishes or slides, depending on the purpose of the experiments (see below). Cultures were maintained in serum-free B27/neurobasal medium in a humidi®ed atmosphere (5% CO2, 95% air) at 37°C for 16 days. More than 95% of the cells present on day 5 in vitro were differentiated into neurons, as characterized by the appearance of long neurites expressing neuro®lament protein (Kashiwagi et al. 1998). Routinely, cortical neurons on day 9 in vitro were treated with 1 mM lithium chloride or its vehicle and maintained for 6 days in cultures. Glutamate was then added to the culture medium on day 15 in vitro. Measurement of neurotoxicity Unless otherwise indicated, cortical neuronal cultures were plated on 96-well plates and treated with indicated dose of glutamate for 24 h after 15 days in culture. Cell viability was determined by the 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay as described previously (Wei et al. 2000). The tetrazolium ring of MTT is reduced by an active dehydrogenase in viable mitochondria, forming a blue-colored precipitate, which is then dissolved in dimethyl sulfoxide and quanti®ed spectrophotometrically at 540 nm. The results are expressed as a percentage of the untreated control. The 100% control values were in the range of 0.6±0.8 OD units at 540 nm. Cell injury was also quantitatively assessed by measurement of released lactate dehydrogenase (LDH) from damaged cells. The LDH assay was performed using the CytoTOX96 Non-Radioactive Cytotoxicity Assay Kit (Promega, Madison, WI, USA), according to the manufacturer's protocol. Glutamate-induced LDH release was expressed as a percentage of experimental LDH release/maximal LDH release which was induced by treatment of 0.9% Triton X-100. One-way ANOVA was used for the statistical analysis, and signi®cant differences in cell viability were determined by post hoc comparisons of means using Bonferroni test. Measurement of 45Ca2+ uptake Calcium in¯ux in cortical cultures on 24-well plates was monitored by use of 45Ca2+ as a tracer as described previously (Nonaka et al. 1998a) with slight modi®cations. Brie¯y, 45Ca2+ (1 lCi/mL) in the absence or presence of glutamate (8 lM) was added to the culture medium and incubated at room temperature for 2 min. The in¯ux of calcium into cells was terminated by rapid washing of the culture wells three times with phosphate-buffered saline (PBS) without calcium and magnesium. Cells were solubilized with 1% sodium dodecyl sulfate (SDS) and the 45Ca2+ radioactivity was measured by

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

Lithium neuroprotection and NMDA receptor 591

a liquid scintillation counter. The statistical analysis between glutamate- and LiCl/glutamate-treated cultures was performed using one-way ANOVA with Bonferroni post hoc test. Measurement of intracellular calcium Cortical cultures were loaded by addition of 2.5 lL of 1 mM Fura-2 AM (Molecular Probes, Eugene, OR, USA) to the dish (containing 2.5 mL of medium) and incubation at 37°C for 30 min in a CO2 incubator. The dishes were then washed three times with an external salt solution containing 145 mM NaCl, 1 mM CaCl2, 2.5 mM KCl, 10 mM HEPES at pH 7.4, and 10 mM glucose, replenished with 2 mL of external salts. The responses of cortical neuronal cultures on 35-mm glass bottom dishes to glutamate plus glycine stimulation were examined by Fura-2 ¯uorescence imaging. The cultures were treated as described during loading and recording (Hough et al. 1996). After a 60-s wash with external salts at room temperature to establish a baseline, glutamate (10 lM) plus glycine (10 lM) in external salts was perfused onto the cells under examination for 30 s and the cells were washed again with external salts for 30±60 s. Recordings were done at room temperature. Fura-2 was excited at 340 and 380 nm using a PTI DeltaRam light source and monochromator coupled to a Zeiss Axiovert 40 microscope ®tted with a 63´ water immersion lens. Images were collected at 510 nm or greater by a Hamamatsu Orca CCD camera and Improvision Openlab software, at approximately one pair of frames per s. The image pairs were subsequently digitally ratioed, and the average ratio within 10 regions of interest (either soma or neurites) was plotted against time. Images were collected from two locations in each dish, and two to three dishes were sampled for each data point on four separate cultures. Baseline calcium levels were subtracted from the calcium response averaged over the time that the cells were stimulated by glutamate. Differences between treatment and controls were compared by a single factor analysis of variance followed by Fisher's post hoc comparison with controls. Western blotting and immunoprecipitation of NMDA receptor subunit proteins For preparation of cell lysates, cerebral cortical neurons grown on 6-well plates were detached by scraping and sonicated for 30 s in RIPA buffer containing 50 mM Tris±HCl, pH 8.0, 120 mM NaCl, 1% Nonidet P-40 (g/L), 1% sodium deoxycholate (g/L), 5 mM EDTA, 0.2 mM Na3VO4 and 50 units/mL of aprotinin. After adding an equal volume of the sample buffer (NOVEX, San Diego, CA, USA) to cell lysates, the mixtures were boiled for 5 min. Samples containing 50 lg of protein were loaded and separated by SDS polyacrylamide gel electrophoresis (SDS±PAGE; 10% gel for NR1, 6% gel for NR2A and NR2B), and then transferred onto a polyvinylidene ¯uoride (PVDF) membrane (NOVEX). After blocking for 1 h in PBST (1 ´ PBS and 0.05% Tween-20) containing 5% non-fat dry milk, blots were incubated overnight at 4°C with antibodies against NR1 (Chemicon, Temecula, CA, USA), phosphoNR1 (Ser896; Upstate, Lake Placid, NY, USA), phospho-NR1 (Ser897; Upstate), NR2A (Chemicon) and NR2B (monoclonal; BD Transduction Laboratories, San Diego, CA, USA). Blots were washed four times in PBST and then incubated at room temperature for 1 h with horseradish peroxidase-conjugated secondary antibodies in PBST containing 5% non-fat dry milk. Immunoreactivities of the protein bands were detected by enhanced chemiluminescence (ECL

kit; Amersham Pharmacia Biotech, Arlington Heights, IL, USA), as instructed by the manufacturer. Western blots were quantitatively analyzed by capturing images on ®lms using a CCD camera (Sierra Scienti®c, Sunnyvale, CA, USA) in conjunction with the Macintosh NIH Image 1.6 software (Wayne Rasband, National Institute of Mental Health, Bethesda, MD, USA). For the detection of NR2B phosphorylation at Tyr1472, immunoprecipitation was performed as previously described (Nakazawa et al. 2001). Brie¯y, cells were lysed with RIPA buffer and the lysates were cleared by centrifugation with an excess amount of protein G Sepharose (Amersham Pharmacia Biotech). The supernatants were incubated with antiNR2B antibody (polyclonal; Chemicon), and the immune complexes were collected on protein-G Sepharose. After washing ®ve times with RIPA buffer, the immunoprecipitates were dissolved with the sample buffer and subjected to SDS±PAGE for western blotting using antiphospho-Tyr1472 antibody (Nakazawa et al. 2001). Statistical differences in total protein and phospho-protein levels of NR subunits between untreated and lithium-treated samples were analyzed by Student's t-test. Protein determination Protein concentrations were determined by the BCA kit (Pierce, New York, NY, USA), using bovine serum albumin as the standard.

Results

Neuroprotective effects of lithium against glutamate excitotoxicity Cerebral cortical cultures after 15 days in vitro were exposed to various concentrations of glutamate for 24 h, and cell viability was quanti®ed using the MTT assay. Glutamate potently and dose-dependently induced cell death with an EC50 value of only 8±10 lM (Fig. 1). Glutamate-induced neurotoxicity was blocked by NMDA receptor antagonists, MK-801 and D-2-amino-5-phosphonopentanoate (AP-5), but not by a non-NMDA receptor antagonist, 2,3-dihydro6-nitro-7-sulfamoyl-benzo(f )quinoxaline (NBQX), indicating that this toxicity was entirely mediated by the NMDA receptor subtype. Pre-treatment of cerebral cortical cultures with LiCl for 6 days dose-dependently suppressed the excitotoxicity induced by 8 lM glutamate in the concentration range examined (0.05±1.6 mM) (Fig. 2a). The neuroprotective effect of lithium was signi®cant at 0.1±0.2 mM and reached nearly complete protection at 1 mM. Pre-treatment with concentrations greater than 1 mM resulted in diminished protection. Since the therapeutic level of lithium in the plasma is typically considered to be 0.6±1.0 mM (Hopkins and Gelenberg 2000), the observed neuroprotective effects of lithium in cortical neuronal cultures occurred at therapeutic and subtherapeutic levels of this drug. Morphological assessment of cortical cultures stained with MTT revealed that glutamate excitotoxicity was accompanied by an extensive loss of neurites (Fig. 2b). This glutamate-induced loss was largely prevented by pre-treatment with 1 mM LiCl.

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

592 R. Hashimoto et al.

Fig. 1 Pharmacology of glutamate-induced excitotoxicity in rat cerebral cortical neurons. Cortical cultures on day 15 in vitro were exposed to the indicated concentrations of glutamate in the absence or presence of MK-801 (10 lM), AP-5 (200 lM) or NBQX (10 lM). Cell viability was determined 24 h later by the MTT assay. Results are the mean ‹ SEM of viability measurements from ®ve to six cultures. *p < 0.05; **p < 0.01; ***p < 0.001, compared with data derived from corresponding cultures treated with glutamate alone.

Lithium neuroprotection was time-dependent, requiring a minimum of 2-day pre-treatment and achieving a maximal effect after 5±6 days (Fig. 3a). Dose-dependent neuroprotection by lithium was con®rmed by measuring the release of LDH from glutamate-treated cells (Fig. 3b), and the LDH release data demonstrated that glutamate excitotoxicity was almost exclusively mediated by NMDA receptors. Selectivity of lithium protection against excitotoxicity in cortical neurons The neuroprotective actions of other monovalent ions, lithium isotopes, another mood-stabilizer, and antidepressants were examined to determine the selectivity of lithium in protecting cortical cultures from excitotoxicity. Monovalent ions such as rubidium and cesium were unable to protect cortical neurons from glutamate excitotoxicity following pretreatment for 6 days in the concentration range of 0.05±1 mM (Fig. 4a). Natural lithium contains two isotopes, 6Li and 7Li, with 93% being in the form of 7Li. These two isotopes have distinct nuclear magnetic resonance spectra (Riddell 1991). However, 6Li and natural lithium were about equally potent in providing neuroprotection at concentrations up to 1 mM and in triggering neurotoxicity at higher concentrations (Fig. 4b). Six-day pre-treatment with valproate, another major mood-stabilizing drug, also concentration-dependently prevented glutamate-induced excitotoxicity in cortical neurons with nearly complete protection at 1 mM (Fig. 5a). In contrast, pre-treatment with tricyclic antidepressants such as clomipramine, desipramine and imipramine or a selective

Fig. 2 Lithium in the therapeutic and subtherapeutic dose ranges protects cortical neurons against glutamate-induced excitotoxicity. (a) Dose-dependent lithium neuroprotection against glutamate excitotoxicity. Cultured cortical neurons were pre-treated with the indicated concentrations of LiCl for 6 days prior to glutamate stimulation (8 lM) on day 15 in vitro. Results are the mean ‹ SEM of viability measured by the MTT assay from ®ve to six cultures. *p < 0.05; ***p < 0.001, compared with the results of samples treated with the glutamate alone. (b) Lithium-induced neuroprotection revealed by morphological inspections. Cortical neurons were pre-treated with LiCl (1 mM) for 6 days prior to stimulation with glutamate (8 lM) on day 15 in vitro. Cells were stained with MTT 24 h later and viable cells were visualized with a phase-contrast microscope and photographed. Bar: 10 lm.

serotonine reuptake inhibitor, ¯uoxetine, in the dose range of 0.03±1 lM failed to provide neuroprotection against excitotoxicity (Fig. 5b). Inhibition of calcium uptake by long-term lithium treatment Calcium is a key messenger in glutamate receptor-mediated synaptic plasticity and neurotoxicity. To examine the possibility that lithium exerts its neuroprotective effects by modulating NMDA receptor-mediated Ca2+ entry, we performed 45Ca2+ uptake studies and the measurement of intracellular calcium levels. The inhibition of 45Ca2+ uptake elicited by this long-term lithium pre-treatment was dosedependent, whereas an acute 1-h pre-treatment was ineffective (Fig. 6a). Glutamate-induced 45Ca2+ in¯ux was completely blocked by MK-801 and AP-5, but not by NBQX, again indicating that this is an NMDA receptormediated process. The chronic effect of lithium on glutamate-induced increase of [Ca2+]i was examined using Fura-2 ¯uorescence

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

Lithium neuroprotection and NMDA receptor 593

Fig. 3 Lithium time- and dose-dependently protects cortical neurons from excitotoxicity. (a) Time requirement for lithium neuroprotection. LiCl (1 mM) was added to cultures on various days in vitro to achieve multiple pre-treatment times. Glutamate (10 lM) was added on day 15 in culture and cell survival was determined 24 h after glutamate addition. Results are expressed as a percentage of maximal neuroprotection, using values derived from glutamate-treated cells and the untreated control as references. (b) Inhibition of glutamate-induced LDH release by lithium pre-treatment. Cortical cultures were pretreated with the indicated concentrations of lithium for 6 days and then exposed to glutamate (8 lM) on day 15 in vitro. The release of LDH was determined 24 h after glutamate stimulation. When used in conjunction with glutamate, MK-801 and AP-5 were 10 lM and 200 lM, respectively. Results are the mean ‹ SEM of viability measurements from ®ve to six cultures. *p < 0.05; **p < 0.01; ***p < 0.001, compared with results from samples treated with glutamate alone.

imaging. From Fura-2 ¯uorescence imaging, it was clear that stimulation with 10 lM glutamate plus 10 lM glycine robustly elevated [Ca2+]i in both soma and neurites in cortical neurons, and that the increase of [Ca2+]i induced by glutamate stimulation was inhibited in the cells pre-treated with lithium (1 mM) for 6 days (Fig. 6b). The elevation of [Ca2+]i during a 30-s pulse of glutamate was dose-dependently suppressed by a 6-day pre-treatment with lithium (0.2±1 mM) (Fig. 6b). Basal levels of intracellular calcium were not signi®cantly affected by lithium treatment (data not shown). Reduction of the level of phosphorylated NR2B in lithium-treated cortical cultures NMDA receptor activity can be modulated by receptor subunit expression and/or phosphorylation at serine/threonine and tyrosine residues (Dunah et al. 1999). To elucidate the molecular mechanisms underlying observed lithiuminduced reduction in NMDA receptor-mediated Ca2+ in¯ux, the levels of total NMDA receptor subunit protein and of several subunit phosphorylations were measured by western blotting. The levels of total NR1, NR2A and NR2B protein were unchanged in cultures pre-treated with lithium (1 mM) for 6 days (Figs 7a and b). The levels of phosphorylation of

Fig. 4 Ion selectivity of lithium protection against excitotoxicity in cortical neuronal cultures. (a) Ion selectivity for neuroprotection. Cortical cultures were pre-treated with the indicated concentrations of lithium (h), rubidium (d), or cesium (j) for 6 days, starting on day 9 in vitro, and then exposed to glutamate (8 lM) for 24 h. (b) Neuroprotective and neurotoxic effects of lithium. For neuroprotective studies, cortical cultures were pre-treated with natural lithium chloride (j) or 6LiCl (h, 0±1 mM) for 6 days prior to glutamate stimulation (8 lM) on day 15 in vitro for 24 h (left panel). For neurotoxic studies, natural lithium chloride or 6LiCl (0±10 mM) was added to cortical cultures on day 9 in vitro and cell viability was determined on day 15 (right panel). Results are the mean ‹ SEM of viability measured by the MTT assay from 5 or 6 cultures. *p < 0.05; **p < 0.01; ***p < 0.001, compared with the group of glutamate alone in (a) and the left panel of (b). *p < 0.05; ***p < 0.001, compared with the untreated group on the right panel in (b).

NR1 at Ser896 or Ser897 were also unchanged by lithium pre-treatment (Fig. 7a). However, the levels of phosphorylation of NR2B at Tyr1472 were signi®cantly decreased by approximately 40% following a 6-day lithium pre-treatment (Fig. 7b). This lithium-induced decrease in NR2B tyrosine phosphorylation was time-dependent, with a maximal effect seen after treatment for 4±6 days (Fig. 7c). Discussion

Our study demonstrated that pre-treatment of rat cerebral cortical cultures with 1 mM LiCl for 5±6 days elicited almost complete protection against excitotoxicity when 8±10 lM glutamate was used. One of the most remarkable features of lithium neuroprotection against excitotoxicity in cortical cultures is that this effect was evident at concentrations within or even below those considered to be therapeutic serum level in bipolar patients. Thus, signi®cant protection

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

594 R. Hashimoto et al.

Fig. 5 Drug selectivity of lithium protection against excitotoxicity in cortical neuronal cultures. Effects of chronic pre-treatment with another mood-stabilizer (valproate) (a) and antidepressants (b) on glutamate excitotoxicity. Valproate, clomipramine, desipramine, imipramine or ¯uoxetine at the indicated concentrations were added to cortical cultures on day 9 in vitro prior to glutamate exposure (8 lM) on day 15 for 24 h. Results are the mean ‹ SEM of viability measured by the MTT assay from ®ve or six cultures. ***p < 0.001, compared with the group of glutamate alone.

was obtained at 0.2 mM and virtually complete protection at 1 mM with an EC50 value of approximately 0.4 mM. This value is considerably less than that in cerebellar granule cells (EC50 ˆ 1.3 mM; Nonaka et al. 1998a). The striking difference in lithium potency in these two primary CNS cultures could be due to regional variations in the capacity of these two culture models to accumulate intracellular lithium, to a diverse sensitivity of intracellular machinery to lithium, or to dissimilarity in the stage of development of the neurons used in the two cultures. Regardless of the reasons for this dosediscrepancy, our observations are compatible with the view that the cerebral cortex is one of the primary brain targets involved in the therapeutic actions of lithium. The less effective neuroprotection of lithium at a concentration below 0.6 mM is consistent with its less robust effect in preventing relapses in clinical use at this dose, while a decline in its neuroprotective action at a dose higher than 1.2 mM could be related to the high-dose neurotoxicity of this drug in bipolar patients (Bowden 1996; Hopkins and Gelenberg 2000). The selectivity of lithium as a neuroprotective agent was demonstrated by the lack of effect of other monovalent ions such as rubidium and cesium to protect against excitotoxicity. It is well known that sodium and potassium are essential to maintain physiological homeostasis. This ®nding suggests that lithium is unique among ions in its neuroprotective

Fig. 6 Chronic lithium exposure inhibits glutamate-induced 45Ca2+ uptake and the peak elevation in [Ca2+] in cerebral cortical neuron cultures. (a) Concentration- and time-dependence for lithium inhibition of 45Ca2+ in¯ux. Cells were pre-treated chronically (6 days) or acutely (1 h) with the indicated concentrations of LiCl (0.2±1.0 mM) and then exposed to glutamate (8 lM) in the presence of 45Ca2+. The in¯ux of 45 Ca2+ was determined 2 min after glutamate addition. When used, the concentrations of MK-801, AP-5 and NBQX were 10 lM, 200 lM and 10 lM, respectively. Results are the mean ‹ SEM of 45Ca2+ uptake measured from 5 to 6 cultures. (b) Cortical neurons were pretreated with LiCl (0.2±1.0 mM) for 6 days, starting from day 9 in vitro, and [Ca2+]i was then measured in cells pre-labeled with the ¯uorescent Ca2+ indicator Fura-2. Glutamate-induced [Ca2+]i increase was elicited in cultures on day 15 by a 30-s pulse with glutamate (10 lM)/glycine (10 lM) followed by a 60-s wash. Photographs of calcium imaging (left panels) represent typical calcium images with glutamate-stimulated or unstimulated cells either pre-treated or untreated with lithium (1 mM) for 6 days. Averaged peak of glutamate/glycine-induced [Ca2+]i (right panel). Results are the mean ‹ SEM of the intensity (I) ratio of 340 nm to 380 nm ´ 4096 (which is the maximal number of intensity). Results are the mean ‹ SEM of peak calcium concentration. *p < 0.05; **p < 0.01; ***p < 0.001, compared with results from cultures treated with glutamate alone.

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

Lithium neuroprotection and NMDA receptor 595

Fig. 7 Lithium treatment selectively decreases NR2B phosphorylation in cortical neuronal cultures. (a) Measurement of total NR1 and phospho-NR1 protein levels. (b) Measurement of levels of total NR2A, total NR2B and phospho-NR2B (Tyr1472) proteins. Cortical cultures were either untreated or treated with LiCl (1 mM) for 6 days, starting on day 9 in vitro. Cells were harvested for western blotting for total NR1, phospho-NR1 (Ser896), phospho-NR1 (Ser897), total NR2A and total NR2B proteins by using their speci®c antibodies. For analysis of levels of phospho-NR2B (Tyr1472), western blotting was performed after immunoprecipitation by anti-NR2B antibody. Two arrows indicate doublet bands of the NR1 subunit, which likely represent alternative splice variants of the receptor subunit (Zukin and Bennett 1995). An

arrowhead in pNR1 (Ser897) immunoblot indicates a non-speci®c band. The immunoblots shown are representative of four independent experiments. Immunoreactivities were quanti®ed by using the Macintosh NIH image software. Data represent the mean ‹ SEM of immunoreactiviy from four independent experiments. *p < 0.05, compared with the untreated control. (c) Time-dependent decrease of NR2B Tyr1472 phosphorylation by lithium. Cultures were either untreated or treated with LiCl (1 mM) for the indicated times. All cultures were harvested on day 15 in vitro for western blotting as described above. Upper panel, phospho-NR2B (Tyr1472) immunoblots. Lower panel, total NR2B immunoblots. Immunoblots shown are representative of three independent experiments.

properties, which are not related to the physiological functions of sodium or potassium. It is of interest that 6Li and natural Li, which contains about 93% 7Li, showed no difference in their potency for inducing neuroprotection and neurotoxic effects. 6Li and 7Li have been shown to have a signi®cant difference in their NMR spectra (Riddell 1991), rate of intracellular uptake (Sherman et al. 1984), lethality dose (Alexander et al. 1982), and potency to inhibit inositol monophosphatase (Sherman et al. 1984). 6Li and 7Li were found in a subsequent study, however, not to differ in the latter two properties (Parthasarathy et al. 1992). Similar to lithium, the mood-stabilizer valproate concentration-dependently protected cortical neurons from glutamate excitotoxicity in its therapeutic dose-range, although several antidepressants did not offer protection. The neuroprotective action of lithium correlated with a reduction in NMDA receptor-mediated Ca2+ in¯ux, as

demonstrated by studies using 45Ca2+ uptake and Fura-2 calcium imaging. Since Ca2+ is a key second messenger for NMDA receptor-mediated excitotoxicity and synaptic activity, the observed inhibition of NMDA receptor activity undoubtedly contributes to lithium neuroprotection against excitotoxicity. The physiological function of the NMDA receptor could be regulated by its composition of NR1 and NR2 subunits. Our western blotting results revealed no difference in protein levels of the NR1, NR2A and NR2B subunits in cortical neurons after long-term lithium treatment. NMDA receptor function could also be modulated by serine/ threonine and tyrosine phosphorylation through protein kinases and phosphatases. Ser896 and Ser897 on NR1 subunit are phosphorylated by protein kinases C and A, respectively (Tingley et al. 1997), and their phosphorylation levels are increased in the striatum of rats treated with the antipsychotic drug, haloperidol (Leveque et al. 2000). However, we found

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

596 R. Hashimoto et al.

that levels of NR1 Ser896 and Ser897 phosphorylation were unchanged by lithium in cortical cultures. In contrast, tyrosine phosphorylation of NR2B at residue 1472 is time-dependently decreased by approximately 40% with lithium treatment. This decrease in NR2B tyrosine phosphorylation correlated with a lithium-induced reduction of NMDA receptor-mediated Ca2+ in¯ux and neuroprotection against excitotoxicity. Tyr1472 is the major NR2B site phosphorylated by the Src family tyrosine kinase, Fyn. The level of Tyr1472 phosphorylation on NR2B is developmentally regulated and is markedly increased in the murine hippocampal CA1 after induction of LTP (Nakazawa et al. 2001). Moreover, Fyn mutant mice have impaired Schaffer collateral-CA1 LTP (Grant et al. 1992). As LTP in this synapse requires the activity of the NMDA receptor, the action of Fyn to phosphorylate Tyr1472 is likely to be involved in NMDA receptor function. Thus, the decrease in NR2B Tyr1472 phosphorylation and the subsequent attenuation of NMDA receptor-mediated calcium entry following long-term lithium treatment may be a part of the molecular mechanisms underlying neuroprotection elicited by this drug. Whether lithium regulates the activity of a protein kinase or a protein phosphatase in this mechanism remains to be elucidated. It has been recently shown that protein kinase C (PKC) activation enhances tyrosine phosphorylation of the NR2B subunit of the NMDA receptor (Grosshans and Browning 2001) and that NMDA-induced excitotoxicity is reduced by PKC inhibitors (Wagey et al. 2001). Thus, PKC activation could be one of the upstream mechanisms leading to enhanced NR2B Tyr1472 phosphorylation. Phosphorylation of Tyr1472 may affect NMDA receptor internalization, which in turn affects the receptor function. The NR2B subunit contains a four-amino acid segment (residues 1472± 1475, YEKL) that confers NMDA receptor internalization (Roche et al. 2001). This is partially blocked by the binding of PSD-95 to amino acid residues adjacent to the YEKL sequence. Thus, phosphorylation of Tyr1472 is likely to suppress NMDA receptor internalization and inhibition of Tyr1472 phosphorylation by lithium would be protective against glutamate excitotoxicity by allowing NMDA receptor to be internalized to reduce calcium in¯ux through the receptor. In cerebellar granule cells, neuroprotective effects of lithium are related to activation of the cell-survival factor, Akt (Chalecka-Franaszek and Chuang 1999), enhanced expression of cytoprotective Bcl-2, but suppression of proapoptotic p53 and Bax (Chen and Chuang 1999). It remains to be investigated as to whether these mechanisms are involved in lithium protection against excitotoxicity in cerebral cortical neurons. The concentration-dependence and kinetics of the neuroprotective effects of lithium are consistent with its clinical use in treating bipolar mood disorder. The neuroprotective effects of lithium against excitotoxicity were shared by the other mood-stabilizer, valproate, as well. Recent brain imaging and

post-mortem studies show morphological abnormalities in the brains of mood disorder patients. For example, a signi®cant atrophy and loss of pyramidal glutaminergic excitatory neurons were observed in cortical layers III and V of Brodmann area 9 (of the pre-frontal cortex) in bipolar patients (Rajkowska et al. 1999). Drevets et al. (1997) have also reported decreased volume of the subgenual pre-frontal cortex in unipolar and bipolar patients. Interestingly, this volume decrease was much less evident in patients treated with lithium or valproate (Drevets 2001). Moreover, chronic lithium treatment increases total brain gray matter volume and N-acetyl-aspartate content, a putative marker of neuronal survival, in human brains (Moore et al. 2000a,b). Taken together, these observations are consistent with our working hypothesis that the neuroprotective effects of lithium are relevant to its clinical actions. The ability of lithium to protect against excitotoxicity at therapeutic and subtherapeutic doses raises an intriguing possibility that lithium may have additional utility to treat neurodegenerative diseases, particularly those linked to excitotoxicity. Acknowledgements We thank Drs Hirohiko Kanai and Toyoko Hiroi of our section in NIMH for their helpful discussions. Excellent technical assistance from Peter Leeds, Lori Christ and Jessica Madert is also greatly appreciated.

References Alexander G. J., Lieberman K. W., Okamoto M., Stokes P. E. and Triana E. (1982) Lithium toxicology: effect of isotopic composition on lethality and behavior. Pharmacol. Biochem. Behav. 16, 801±804. Ali D. W. and Salter M. W. (2001) NMDA receptor regulation by Src kinase signalling in excitatory synaptic transmission and plasticity. Curr. Opin. Neurobiol. 11, 336±342. Bhat R. V., Shanley J., Correll M. P., Fieles W. E., Keith R. A., Scott C. W. and Lee C. M. (2000) Regulation and localization of tyrosine216 phosphorylation of glycogen synthase kinase-3b in cellular and animal models of neuronal degeneration. Proc. Natl Acad. Sci. USA 97, 11074±11079. Bijur G. N., De Sarno P. and Jope R. S. (2000) Glycogen synthase kinase-3b facilitates staurosporine- and heat shock-induced apoptosis. Protection by lithium. J. Biol. Chem 275, 7583±7590. Bowden C. L. (1996) Dosing strategies and time course of response to antimanic drugs. J. Clin. Psychiatry 57, 4±9; discussion 10±12. Chalecka-Franaszek E. and Chuang D.-M. (1999) Lithium activates the serine/threonine kinase Akt-1 and suppresses glutamate-induced inhibition of Akt-1 activity in neurons. Proc. Natl Acad. Sci. USA 96, 8745±8750. Chen R. W. and Chuang D.-M. (1999) Long term lithium treatment suppresses p53 and Bax expression but increases Bcl-2 expression. A prominent role in neuroprotection against excitotoxicity. J. Biol. Chem 274, 6039±6042. D'Mello S. R., Anelli R. and Calissano P. (1994) Lithium induces apoptosis in immature cerebellar granule cells but promotes survival of mature neurons. Exp. Cell Res. 211, 332±338. Drevets W. C. (2001) Neuroimaging studies of major depression, in Advances In Brain Imaging (Morihisa J., ed.) of the American

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597

Lithium neuroprotection and NMDA receptor 597

Psychiatric Press Annual Review of Psychiatry (Oldham J. M., Series ed.), pp. 123±170. American Psychiatric Press, Washington DC. Drevets W. C., Price J. L., Simpson J. R., Todd R. D., Reich T., Vannier M. and Raichle M. E. (1997) Subgenual pre-frontal cortex abnormalities in mood disorders. Nature 386, 824±827. Dunah A. W., Yasuda R. P., Luo J., Wang Y., Prybylowski K. L. and Wolfe B. B. (1999) Biochemical studies of the structure and function of the N-methyl-D-aspartate subtype of glutamate receptors. Mol. Neurobiol. 19, 151±179. Grant S. G., O'Dell T. J., Karl K. A., Stein P. L., Soriano P. and Kandel E. R. (1992) Impaired long-term potentiation, spatial learning, and hippocampal development in fyn mutant mice. Science 258, 1903±1910. Grosshans D. R. and Browning M. D. (2001) Protein kinase C activation induces tyrosine phosphorylation of the NR2A and NR2B subunits of the NMDA receptor. J. Neurochem. 76, 737±744. Hashimoto R., Nakamura Y., Komai S., Kashiwagi Y., Tamura K., Goto T., Aimoto S., Kaibuchi K., Shiosaka S. and Takeda M. (2000) Sitespeci®c phosphorylation of neuro®lament-L is mediated by calcium/calmodulin-dependent protein kinase II in the apical dendrites during long-term potentiation. J. Neurochem. 75, 373±382. Hopkins H. S. and Gelenberg A. J. (2000) Serum lithium levels and the outcome of maintenance therapy of bipolar disorder. Bipolar Disord. 2, 174±179. Hough C. J., Irwin R. P., Gao X. M., Rogawski M. A. and Chuang D.-M. (1996) Carbamazepine inhibition of N-methyl-D-aspartate-evoked calcium in¯ux in rat cerebellar granule cells. J. Pharmacol. Exp. Ther. 276, 143±149. Inouye M., Yamamura H. and Nakano A. (1995) Lithium delays the radiation-induced apoptotic process in external granule cells of mouse cerebellum. J. Radiat Res. 36, 203±208. Jope R. S. (1999) Anti-bipolar therapy: mechanism of action of lithium. Mol. Psychiatry 4, 117±128. Kashiwagi Y., Nakamura Y., Miyamae Y., Hashimoto R. and Takeda M. (1998) Pulse exposure of cultured rat neurons to aluminum-maltol affected the axonal transport system. Neurosci. Lett. 252, 5±8. Leveque J. C., Macias W., Rajadhyaksha A., Carlson R. R., Barczak A., Kang S., Li X. M., Coyle J. T., Huganir R. L., Heckers S. and Konradi C. (2000) Intracellular modulation of NMDA receptor function by antipsychotic drugs. J. Neurosci. 20, 4011±4020. Manji H. K. and Lenox R. H. (2000) Signaling: cellular insights into the pathophysiology of bipolar disorder. Biol. Psychiatry 48, 518±530. Moore G. J., Bebchuk J. M., Hasanat K., Chen G., Seraji-Bozorgzad N., Wilds I. B., Faulk M. W., Koch S., Glitz D. A., Jolkovsky L. and Manji H. K. (2000a) Lithium increases N-acetyl-aspartate in the human brain: in vivo evidence in support of bcl-2's neurotrophic effects? Biol. Psychiatry 48, 1±8. Moore G. J., Bebchuk J. M., Wilds I. B., Chen G. and Manji H. K. (2000b) Lithium-induced increase in human brain grey matter. Lancet 356, 1241±1242. Nakazawa T., Komai S., Tezuka T., Hisatsune C., Umemori H., Semba K., Mishina M., Manabe T. and Yamamoto T. (2001) Characterization of Fyn-mediated tyrosine phosphorylation sites on GluR epsilon 2 (NR2B) subunit of the N-methyl-D-aspartate receptor. J. Biol. Chem. 276, 693±699. Nonaka S. and Chuang D.-M. (1998) Neuroprotective effects of chronic lithium on focal cerebral ischemia in rats. Neuroreport 9, 2081± 2084. Nonaka S., Hough C. J. and Chuang D.-M. (1998a) Chronic lithium treatment robustly protects neurons in the central nervous system against excitotoxicity by inhibiting N-methyl-D-aspartate receptormediated calcium in¯ux. Proc. Natl Acad. Sci. USA 95, 2642± 2647. Nonaka S., Katsube N. and Chuang D.-M. (1998b) Lithium protects rat cerebellar granule cells against apoptosis induced by anticonvul-

sants, phenytoin and carbamazepine. J. Pharmacol. Exp. Ther. 286, 539±547. Parthasarathy R., Parthasarathy L., Ramesh T. G., Devi C. S. and Vadnal R. E. (1992) The effects of lithium isotopes on the myo-inositol 1-phosphatase reaction in rat brain, liver, and testes. Life Sci. 50, 1445±1450. Pascual T. and Gonzalez J. L. (1995) A protective effect of lithium on rat behaviour altered by ibotenic acid lesions of the basal forebrain cholinergic system. Brain Res. 695, 289±292. Phiel C. J. and Klein P. S. (2001) Molecular targets of lithium action. Annu. Rev. Pharmacol. Toxicol. 41, 789±813. Rajkowska G., Miguel-Hidalgo J. J., Wei J., Dilley G., Pittman S. D., Meltzer H. Y., Overholser J. C., Roth B. L. and Stockmeier C. A. (1999) Morphometric evidence for neuronal and glial pre-frontal cell pathology in major depression. Biol. Psychiatry 45, 1085±1098. Riddell F. G. (1991) Studies on Li+ transport using 7Li and 6Li NMR. In: Lithium and The Cell (Nicholas, J. B., ed.), pp. 85±98. Academic Press Inc, London. Roche K. W., Standley S., McCallum J., Dune Ly C., Ehlers M. D. and Wenthold R. J. (2001) Molecular determinants of NMDA receptor internalization. Nat. Neurosci. 4, 794±802. Rosenblum K., Berman D. E., Hazvi S., Lamprecht R. and Dudai Y. (1997) NMDA receptor and the tyrosine phosphorylation of its 2B subunit in taste learning in the rat insular cortex. J. Neurosci. 17, 5129±5135. Rostas J. A., Brent V. A., Voss K., Errington M. L., Bliss T. V. and Gurd J. W. (1996) Enhanced tyrosine phosphorylation of the 2B subunit of the N-methyl-D-aspartate receptor in long-term potentiation. Proc. Natl Acad. Sci. USA 93, 10452±10456. Sherman W. R., Munsell L. Y. and Wong Y. H. (1984) Differential uptake of lithium isotopes by rat cerebral cortex and its effect on inositol phosphate metabolism. J. Neurochem. 42, 880±882. Sparapani M., Virgili M., Ortali F. and Contestabile A. (1997) Effects of chronic lithium treatment on ornithine decarboxylase induction and excitotoxic neuropathology in the rat. Brain Res. 765, 164± 168. Takagi N., Shinno K., Teves L., Bissoon N., Wallace M. C. and Gurd J. W. (1997) Transient ischemia differentially increases tyrosine phosphorylation of NMDA receptor subunits 2A and 2B. J. Neurochem. 69, 1060±1065. Takagi N., Cheung H. H., Bissoon N., Teves L., Wallace M. C. and Gurd J. W. (1999) The effect of transient global ischemia on the interaction of Src and Fyn with the N-methyl-D-aspartate receptor and postsynaptic densities: possible involvement of Src homology 2 domains. J. Cereb. Blood Flow Metab. 19, 880±888. Tingley W. G., Ehlers M. D., Kameyama K., Doherty C., Ptak J. B., Riley C. T. and Huganir R. L. (1997) Characterization of protein kinase A and protein kinase C phosphorylation of the N-methylD-aspartate receptor NR1 subunit using phosphorylation site-speci®c antibodies. J. Biol. Chem. 272, 5157±5166. Wagey R., Hu J., Pelech S. L., Raymond L. A. and Krieger C. (2001) Modulation of NMDA-mediated excitotoxicity by protein kinase C. J. Neurochem. 78, 715±726. Wei H., Leeds P., Chen R. W., Wei W., Leng Y., Bredesen D. E. and Chuang D.-M. (2000) Neuronal apoptosis induced by pharmacological concentrations of 3-hydroxykynurenine: characterization and protection by dantrolene and Bcl-2 overexpression. J. Neurochem. 75, 81±90. Wei H., Qin Z.-H., Senatorov V. V., Wei W., Wang Y., Qin Y. and Chuang D.-M. (2001) Lithium suppresses excitotoxicity-induced striatal lesions in rats, an animal model of Huntington's disease. Neuroscience 106, 603±612. Zukin R. S. and Bennett M. V. (1995) Alternatively spliced isoforms of the NMDARI receptor subunit. Trends Neurosci. 18, 306±313.

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 589±597