Introduction

Spontaneous N-methyl-d-aspartate (NMDA) receptor activity supports the survival of developing neurons [1–3]. The NMDA receptor, a subtype of the glutamate receptor, acts via the receptor-gated cation channel, which is permeable to Ca2+. Blockade of NMDA receptors with NMDA antagonists induces apoptosis in cultured cortical neurons through the Bax-cytochrome C-caspase pathway [4]. Loss of activity-dependent survival signals is recognized as a stimulus for neuronal apoptosis and may play a significant role in neurodegeneration [5].

Calcium ion is essential for cellular functions including signal transduction [6]; however, uncontrolled calcium stress has been linked causally to a variety of neurological diseases, including ischemia, excitotoxicity and Alzheimer’s disease [7, 8].

Impaired function of the endoplasmic reticulum (ER), commonly referred to as ER stress, is an important factor in the neuropathology of a wide variety of neurological disorders [9, 10]. Several agents, such as thapsigargin, can be used to induce ER stress experimentally. Thapsigargin inhibits Ca2+-ATPase in the ER [11, 12], which blocks the sequestration of calcium by the ER, causing increases in the intracellular concentration of calcium, accumulation of unfolded or misfolded proteins and activation of caspase-3-mediated apoptosis. The cellular signals involved in ER stress-mediated cell death are not fully understood.

Recent reports suggest that glycogen synthase kinase-3 (GSK-3) affects many fundamental cellular functions, including the cell cycle, gene transcription, cytoskeletal integrity and apoptosis [13–16]. The phosphatidylinositol-3 kinase/Akt signaling pathway is a signaling system implicated in the survival of neurons that leads to the inhibition of GSK-3 by increasing Ser9 phosphorylation [13, 17, 18]. We reported recently that thapsigargin-induced apoptosis is prevented by GSK-3 inhibitors in PC12 cells [19].

As described above, recent data have implicated ER stress as an important factor in some pathological conditions such as ischemia; however, whether shared or unique apoptotic cascades are activated by blocking NMDA receptor vs. ER stress in cortical neurons has not previously been investigated. We therefore examined whether GSK-3 is involved in thapsigargin-induced cell death using GSK-3 inhibitors in cultured rat cortical cells.

Materials and methods

Materials

SB216763 was purchased from Tocris Cookson (Bristol, UK). Alsterpaullone, roscovitine and mouse anti-β-actin monoclonal antibody were purchased from Sigma Chemical Co. (St Louis, USA). 1-Azakenpaullone was purchased from Calbiochem (La Jolla, USA). (+)-5-Methyl-10,11-dihydro-5H-dibenzo cyclohepten-5,10-imine (MK801) was generously supplied by Merck Sharp Dohme (Rahway, USA). Rabbit anti-GRP78 polyclonal antibody was purchased from Santa Cruz Biotechnology Inc. (California, USA). Hoechst 33258 dye (bis-benzimide) was purchased from Molecular Probes (Oregon, USA). Ac-Asp-Glu-Val-Asp-4-methyl-coumaryl-7-amide (Ac-DEVD-MCA), Ac-Leu-Glu-His-Asp-4-methyl-coumaryl-7-amide (Ac-LEHD-MCA) and 7-amino-4-methyl-coumarin (AMC) were acquired from the Peptide Institute (Osaka, Japan). Thapsigargin and other chemicals were purchased from Wako Pure Chemical Industries (Osaka, Japan).

Cell culture

Cerebral cortical cells were obtained and cultured essentially as described by Dichter [20] from fetal rats (Wistar) after 18–19 days of gestation. Whole cerebral neocortices were removed from fetal mice, taking care to discard the hippocampal formation, basal ganglia and most of the meninges. The tissue was then minced, incubated in 0.25% trypsin for 30 min at 37 °C, and then Dulbecco’s modified Eagle’s medium (DMEM) plus fetal calf serum was added. The tissue was triturated with Pasteur pipette. Clumps of cells were removed by filtering through a double layer of lens paper.

The dissociated cortical cells were cultured on poly-d-lysine-coated 35 mm dishes (Falcon 3001) (2 × 106 cells/dish) or cover glasses in DMEM containing 10% fetal calf serum. Ten micromolar of cytosine-β-d arabinofuranoside was added to the culture medium on day 3 after plating. The cells were cultured for 9–10 days and then used in the experiments. We examined the relative proportion of neurons to other cell types by staining the cells with anti-neuronal nuclei (NeuN) monoclonal antibody and anti-glial fibrillary protein (GFAP) visualized with an immunohistochemical staining kit, Vectastain ABC kit (Vector). The relative proportion of neurons (NeuN+) to glia cells (GFAP+) was about 95%.

Cell treatment and cell viability

The cells were washed twice with Tris-buffered salt solution containing (in mM): NaCl 120, KCl 5.4, CaCl2 1.8, MgCl2 0.8, Tris–HCl 25 and glucose 15, at pH 6.5 (washing buffer) and then replaced with 2 ml of serum (5%)-containing or serum-free DMEM. We also used washing buffer at pH 6.5 to avoid glutamate neurotoxicity which depends on extracellular pH. Cell treatment with various reagents was carried out for 48 h at 37 °C. Morphological cell changes were observed by phase-contrast microscopy during treatment. Cell viability was checked by Trypan blue (0.4%) exclusion.

Quantitation of apoptosis by nuclear morphological changes

Apoptotic cell death was determined by staining cells with Hoechst dye H33258. Cells were fixed with a 10% formalin neutral phosphate buffer solution (pH 7.4) for 5 min at room temperature. After washing the cells with distilled water, they were stained with 8 μg/ml of H33258 for 5 min. Nuclear morphology was observed under a fluorescent microscope (Olympus IX70 model). Apoptosis was quantitated by scoring the percentage of cells with apoptotic nuclear morphology at the single cell level. Condensed or fragmented nuclei were scored as apoptotic. A total of five to seven randomly selected fields were captured using Polaroid PDMC II software. At least 200 cells were counted per condition, and each experiment was repeated in at least three different cultures.

Western blotting

Primary cultured cells were scraped off the dish and collected by centrifugation (400 × g for 5 min), followed by homogenization in ice-cold buffer (50 mM Tris–HCl buffer containing 50 mM NaCl, 10 mM EGTA, 5 mM EDTA, 2 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 1 mM benzamide and 10% protease inhibitor cocktail, pH 7.4). The cell suspension was placed on ice for 30 min and centrifuged at 18,000 × g for 30 min. The supernatant was diluted with an equal volume of sample buffer containing 62.5 mM of Tris–HCl (pH 6.8), 2% sodium dodecyl sulfate (SDS), 10% glycerol, 50 mM of dithiothreitol and 0.1% bromophenol blue. Protein concentration was determined by the bicinchoninic acid assay. Each sample (10 μg/lane) was loaded and separated using 7.5% SDS-polyacrylamide gel electrophoresis. Proteins were transferred to a nitrocellulose membrane and blocked with Tris-buffered saline (TBS-T; 20 mM Tris and 0.1% Tween) containing 5% skim milk for 1 h at room temperature, and then incubated with anti-GRP78 in TBS-T overnight at 4 °C. After washing for 5 min with three changes of TBS-T, the membrane was incubated with a phosphatase-conjugated goat anti-rabbit antibody for 1 h at room temperature in TBS-T. After washing for 5 min with three changes of TBS-T, immunoreactive bands were visualized with a Western blot detection kit BCIP/NBT system. Equal amounts of protein extracts were also analyzed by Western blot analysis with anti-actin antibody.

Caspase activity

Caspase activity was measured as described previously [2]. Briefly, cells were washed with phosphate-buffered saline and suspended in 50 mM of Tris–HCl buffer (pH 7.4) containing 1 mM of EDTA and 10 mM of EGTA. They were then treated with 10 μM of digitonin for 10 min. Lysates were obtained by centrifugation at 10,000 × g for 5 min, and cleared lysates containing 50–100 μg protein were incubated with 50 μM of enzyme substrate Ac-DEVD-MCA (caspase-3) or Ac-LEHD-MCA (caspase-9) for 1 h at 37 °C. The reaction was terminated by the addition of monoiodoacetic acid (5 mM). AMC levels were measured using a spectrofluorometer (Hitachi F4500, Japan) with excitation at 380 nm and emission at 460 nm, and the activity was expressed as pmol of AMC released/min/mg protein.

Statistical analysis

Statistical significance was assessed by one-way ANOVA and by post-hoc Scheffe’s comparisons.

Results

Thapsigargin- and NMDA antagonist-induced apoptosis

We used thapsigargin to induce ER stress experimentally. Thapsigargin inhibits Ca2+-ATPase in the ER, which blocks the sequestration of calcium by the ER. In addition, we used MK801, an NMDA receptor antagonist, to induce apoptosis by blocking Ca2+-dependent survival [2]. After treating cortical cells with thapsigargin (0.1 μM) for 48 h, cells showed apoptotic morphology, including shrunken cell bodies by phase-contrast observation, chromatin condensation and nuclear fragmentation detected by staining with H33258 (Fig. 1). Cell treatment with MK801 (10 μM) in the absence of serum induced apoptosis showing apoptotic morphology, including shrunken cell bodies, chromatin condensation and nuclear fragmentation. The loss of cell viability was also confirmed by trypan blue exclusion (data not shown).

Fig. 1
figure 1

Thapsigargin- and MK801-induced apoptosis. Phase-contrast (a–d) and H33258 fluorescence (e–h) microscopy of rat cortical cells. Cells were incubated with 0.1 μM thapsigargin (b, f) or 10 μM MK801 (d, h) in the absence (c, d, g, h) or presence (a, b, e, f) of 5% dialyzed bovine fetal serum for 48 h at 37 °C. Arrowheads indicate healthy neurons. Arrows indicate neurons with apoptotic morphology. Scale bars = 10 μm

Full size image

Up to 40–50% of cells showed apoptotic death by treatment with either thapsigargin (0.1 μM) or MK801 (10 μM) (Fig. 2a). Up to 20% neurons showed apoptosis depending on experimental conditions by serum withdrawal.

Fig. 2
figure 2

a Thapsigargin- and MK801-induced apoptosis. Cells were incubated with 0.1 μM thapsigargin in the presence of 5% dialyzed bovine fetal serum, or 10 μM MK801 in the absence of serum for 48 h at 37 °C. Each value represents the mean ± S.E.M. Asterisk represents P < 0.05 (vs. untreated cells). Hash represents P < 0.05 [vs. untreated cells (serum-free)]. b Effects of thapsigargin and MK801 on GRP78 protein expression. Cells were incubated with (lane 2) or without (lane 1) 0.1 μM thapsigargin in the presence of 5% dialyzed bovine fetal serum, and incubated with (lane 4) or without (lane 3) 10 μM MK801 in the absence of serum for 24 h at 37 °C. Cells were lysed and the lysate was immunoblotted with anti-GRP78 or anti-β-actin antibody

Full size image

Cell treatment with thapsigargin (0.1 μM) for 24 h induced the expression of GRP78 protein, indicating ER stress response (Fig. 2b). On the other hand, cell treatment with MK801 (10 μM) did not induce the expression of GRP78 protein (Fig. 2b).

We then examined whether ER stress-induced apoptosis was accompanied by the activation of caspase-9 and caspase-3. Cell treatment with thapsigargin (10 and 100 nM) for 24 h activated caspase-9 (Fig. 3a) and caspase-3 (Fig. 3b) activity in a dose-dependent manner.

Fig. 3
figure 3

Thapsigargin-induced caspase-9 and caspase-3 activation. Cells were incubated with thapsigargin (10–100 nM) for 24 h at 37 °C. Caspase-9 (a) and caspase-3 (b) activity was measured as described in Sect. ”Materials and methods.” Data are shown as the mean ± S.E.M. Asterisk represents P < 0.05 (vs. untreated cells)

Full size image

Protective effect of GSK-3 inhibitors on thapsigargin-induced apoptosis

Glycogen synthase kinase-3 is the principal physiological substrate of Akt (also known as protein kinase B) and the activity of GSK-3 is inhibited by Akt-mediated phosphorylation in response to trophic stimulation. To investigate directly the role of endogenous GSK-3 activity in cell death in response to thapsigargin treatment, we assayed the effects of selective GSK-3 inhibitors, SB216763, alsterpaullone and 1-azakenpaullone, on thapsigargin-induced apoptosis. Incubation with alsterpaullone, SB216763 and 1-azakenpaullone inhibited thapsigargin-induced apoptosis (Fig. 4). SB216763 has been reported to inhibit GSK-3 activity with no significant activity toward other protein kinases, including cyclin-dependent protein kinases [21, 22]. On the other hand, alsterpaullone is also known to inhibit cyclin-dependent protein kinase 5 [23, 24]; however, a cyclin-dependent protein kinase 5 selective inhibitor, roscovitine, did not prevent thapsigargin-induced apoptosis (data not shown), suggesting that alsterpaullone prevented apoptosis by inhibiting GSK-3 activity.

Fig. 4
figure 4

Protective effect of GSK-3 inhibitors on thapsigargin-induced cell injury. Cells were incubated with 0.1 μM of thapsigargin in the presence or absence of alsterpaullone (Als, 1 μM), SB216763 (SB, 5 μM) or 1-azakenpaullone (Aza, 2 μM) for 48 h at 37 °C. Each value represents the mean ± S.E.M. Asterisk represents P < 0.05 (vs. thapsigargin-only-treated cells)

Full size image

Effect of GSK-3 inhibitors on thapsigargin-induced caspase activation

As shown in Fig. 3, thapsigargin activated caspase-3 and caspase-9 activity; therefore, we assayed the effect of alsterpaullone on thapsigargin-induced caspase activation. Cell treatment with thapsigargin (0.1 μM) for 24 h activated caspase-9 (Fig. 5a) and caspase-3 (Fig. 5b), and alsterpaullone inhibited the activation of both caspases.

Fig. 5
figure 5

Effects of a GSK-3 inhibitor on thapsigargin-induced caspase-9 and caspase-3 activation. Cells were incubated with 0.1 μM of thapsigargin in the presence of alsterpaullone (1 μM) for 24 h at 37 °C. Caspase-9 (a) and caspase-3 (b) activity were measured as described in Sect. ”Materials and methods.” Data are shown as the mean ± S.E.M. Asterisk represents P < 0.05 (vs. thapsigargin-only-treated cells)

Full size image

Discussion

We showed in this report that GSK-3 inhibitors protected cortical cells from thapsigargin-induced apoptosis. GSK-3 inhibitors inhibited thapsigargin-induced caspase-9 and caspase-3 activation, suggesting that GSK-3 is involved in the caspase pathway. Glycogen synthase kinase-3 activity is known to be suppressed when it becomes phosphorylated on serine 9 by the activation of Akt. Insulin growth factor-I, which is known to activate the phosphatidylinositol-3 kinase/Akt signaling pathway, protected cortical neurons from thapsigargin-induced apoptosis (data not shown). The mechanism by which ER stress activates GSK-3β activity is not clear. Song et al. [25] reported that ER stress signaling induces the protein phosphatase 2A-dependent dephosphorylation of GSK-3β, and the subsequent GSK-3β-dependent induction of apoptosis in human SH-SY5Y neuroblastoma cells. Pro-apoptotic substrates for GSK-3β have not been well-characterized, although Linseman et al. [26] reported that GSK-3 phosphorylates Bax, a pro-apoptotic Bcl-2 family member that stimulates the intrinsic (mitochondrial) death pathway by eliciting cytochrome c release from mitochondria, and promotes its mitochondrial localization in human embryonic kidney 293 cells. We reported previously that GSK-3 inhibitors completely protected cortical neurons from the NMDA antagonist-induced apoptosis induced possibly by blocking the trophic effect of NMDA receptor and decreasing the intracellular calcium ion levels of the cells [16, 27]. As shown in Fig. 2b, the blockade of NMDA receptor induced apoptosis without expressing GRP78. NMDA receptor antagonist-induced apoptosis involves apoptotic events that are induced by changes in outer mitochondrial membrane permeability in rat cortical neurons [4]. Glutamate plays an important role in neuronal survival in the developing brain largely through facilitating the entry of Ca2+ [28, 29]. On the other hand, impaired function of the ER is an important factor in the neuropathology of a wide variety of neurological disorders [9, 10].

Feng et al. [30] reported recently that cleavage and activation of caspase-3, -9, releasing of cytochrome C from mitochondria, upregulation of GRP78 expression were induced in K562 cells after exposure of thapsigargin, suggesting that mitochondria participates in ER stress-induced apoptosis in K562 cells. Turner et al. [31] reported that A1 adenosine receptor activation and NMDA receptor blockade lead to early postnatal cell injury by mechanisms that involve inhibition of intracellular Ca2+ signaling, and that thapsigargin enhanced the cell injury induced by A1 adenosine receptor activation and NMDA receptor blockade.

ER stress may also activate GSK-3 and induce apoptosis via a mitochondrial apoptotic pathway accompanied by caspase-9 and caspase-3 activation [32]. GSK-3 may serve as a common site of apoptotic signaling downstream of both the inhibition of the trophic effect of NMDA receptor possibly by reducing the intracellular calcium level and the induction of ER stress by disturbing Ca2+ homeostasis in ER.

We showed in this report that GSK-3 inhibitors protected cortical neurons from thapsigargin-induced apoptosis. GSK-3 inhibitors may deserve to be evaluated as therapeutic agents in ER stress-induced neurological disorders such as ischemia