J. Neurochem. (2010) 115, 1045–1056.
NMDA receptors (NMDARs) mediate excitatory synaptic transmission during repetitive or prolonged glutamate release, playing a critical role in synaptic plasticity or cell death, respectively. Evidence indicates that a major pathway of NMDAR signaling to cell death in cortical and hippocampal neurons requires the scaffolding protein post-synaptic density 95 (PSD-95) and activation of neuronal nitric oxide synthase. However, it is not known if this PSD-95-dependent pathway contributes to excitotoxicity in other brain regions. It is also unclear whether the neuroprotective effects of Tat-NR2B9c, a membrane-permeant peptide that disrupts PSD-95/NMDAR binding, correlate with uncoupling NR2B- and/or NR2A-type NMDARs from PSD-95. In this study, we used cultured hippocampal and striatal neurons to test the potency of Tat-NR2B9c on uncoupling NR2 subunits from PSD-95 and protecting against NMDA-induced excitotoxicity. We found that the concentration of Tat-NR2B9c required to dissociate 50% of PSD-95 was fourfold lower for NR2B than NR2A in cultured hippocampal and striatal neurons, and that this concentration correlated tightly with protection against NMDA-induced toxicity in hippocampal neurons without altering NMDAR current. In contrast, NMDAR signaling to cell death in cultured striatal neurons occurred independently of the NR2B/PSD-95 interaction or neuronal nitric oxide synthase activation. These results will facilitate development of neuronal type-specific protective therapies.
days in vitro
membrane-associated guanylate kinases
neuronal nitric oxide synthase
Triton X-100 in PBS
protein post-synaptic density 95
sodium dodecyl sulfate
yellow fluorescent protein
Glutamate mediates most excitatory synaptic transmission in the mammalian central nervous system by activation of ionotropic receptors that flux cations. The N-methyl-d-aspartate (NMDA) – type glutamate receptor is activated only with coincident pre-synaptic glutamate release and post-synaptic depolarization, and fluxes calcium to stimulate protein kinases, phosphatases, and proteases; these second messengers induce changes in synaptic strength, which encode learning and memory (Bliss and Collingridge 1993). However, over-activation of NMDA receptors (NMDARs) results in neuronal damage from calcium-activated lipases, proteases, and DNases (Rothman and Olney 1995; Sattler and Tymianski 2000; Besancon et al. 2008).
Most NMDARs in the mature brain are tetramers of two NR1 subunits that bind glycine, and two NR2 subunits that bind glutamate (Dingledine et al. 1999); NR3 subunits are less commonly included in the complex (Cavara and Hollmann 2008). There are four different types of NR2 subunits – NR2A, NR2B, NR2C and NR2D – which show differential spatial and temporal expression during neuronal development (Monyer et al. 1994). The NR2 subunits confer distinct pharmacological and physiological properties on the NMDAR complex, and may also determine differential subcellular localization and protein-protein interactions (Cull-Candy et al. 2001; Kohr 2006). Both NR2A and NR2B are highly expressed in the forebrain, including the cerebral cortex, hippocampus, and striatum (Monyer et al. 1994). Evidence suggests NR2A-containing receptors are targeted to synapses whereas NR2B-containing receptors predominate at extrasynaptic sites on the neuronal plasma membrane (Rumbaugh and Vicini 1999; Tovar and Westbrook 1999; Barria and Malinow 2002). Although some differences have been revealed in the complex of proteins preferentially associated with NR1/NR2A and NR1/NR2B, these two NMDAR subtypes also interact with many of the same synapse-associated proteins, including the protein post-synaptic density 95 (PSD-95) family of membrane-associated guanylate kinases (MAGUKs) (Al-Hallaq et al. 2007).
Studies in cultured cortical and hippocampal neurons reveal that activation of synaptic NMDARs preferentially triggers cell survival pathways, whereas stimulation of the whole-cell population of NMDARs and/or selective activation of extrasynaptic NMDARs signals cells death (Vanhoutte and Bading 2003; Soriano and Hardingham 2007; Leveille et al. 2008). Other studies have suggested that NMDAR complexes containing NR2B facilitate cell death signaling whereas NR2A-containing NMDARs promote neuronal survival (Chen et al. 2007; Liu et al. 2007), although both subtypes mediate toxicity in more mature cultured neurons (von Engelhardt et al. 2007).
Members of the PSD-95 family of MAGUKs act as scaffolds that facilitate signaling by tethering enzymes close to glutamate receptors (Fujita and Kurachi 2000). The NR2A or NR2B C-terminal tSXV motif binds to the second PDZ domain in PSD-95 family members (Kornau et al. 1995; Niethammer et al. 1996). This interaction has been shown to contribute to toxic signaling downstream of NMDAR activation; a peptide that interferes with the binding of MAGUKs to NMDARs rescues hippocampal and cortical neurons from NMDA-induced excitotoxicity in vitro and ischemic neuronal death in vivo (Aarts et al. 2002; Soriano et al. 2008). Neuronal protection mediated by this peptide correlates well with reduced activation of neuronal nitric oxide synthase (nNOS), which also binds to the PDZ2 of PSD-95. Less is known about the mechanisms of excitotoxic death in striatal neurons, which are the targets of neurodegeneration in a variety of neurological disorders as well as in response to ischemia.
Here, we investigate the potency of Tat-NR2B9c in uncoupling PSD-95 from NR2B- versus NR2A-containing NMDARs and in protection from NMDA-induced toxicity in hippocampal neurons. Furthermore, we begin to investigate the signaling pathways downstream of NMDA receptor activation mediating apoptotic death in cultured striatal neurons.
Materials and methods
Primary neuronal cultures
Animal housing, care, and all procedures were approved by the University of British Columbia according to guidelines of the Canadian Council for Animal Care (under the protocol A06-1534 for mice and A07-0581 for rats). Striatal cultures were prepared from postnatal day 0–1 (P0-P1) wild-type FVB/N mice as described previously (Zeron et al. 2002) and plated at a density of approximately 2 × 105 cells/well in 24-well plates with poly-d-lysine (250 μg/mL) pre-coated 12 mm glass coverslips or 60 mm dishes at a density of ∼300–400 cells/mm2. Embryonic striatal and hippocampal cultures were prepared by dissecting striatal and hippocampal tissue respectively from 17- to 18-day-old rat embryos using methods described previously (Li et al. 2002). Cortical and striatal co-cultures were prepared from 17- to 18-day-old rat embryos. Striatal cells were transfected at days in vitro (DIV) 0 with yellow fluorescent protein (YFP) plasmid on a β-actin promoter (a gift from A. M. Craig, University of British Columbia, Vancouver) in 100 μL of electroporation buffer from Mirus Bio LLC (Madison, WI, USA) by nucleofection (Amaxa GmbH, Köln, Germany), according to manufacturer’s instructions. Transfected striatal cells were plated with untransfected cortical cells at a ratio of 1 : 1 on a 24-well plate pre-coated with poly-d-lysine with 2 × 105 cells in total per well. All neuronal cultures were grown in serum-free plating medium (B27, penicillin/streptomycin, α-glutamine, Gibco’s Neurobasal medium), maintained in a humidified 37°C incubator with 5% CO2, and refreshed every 5 days by replacing half of the medium.
The Tat-NR2B9c and Tat-NR2BAA peptides were originally a gift from Dr. M. Tymianski (U Toronto) and were later purchased from AnaSpec (San Jose, CA, USA) and Peptides 2.0 (Chantilly, VA, USA). The NR2B-9c peptide (Lys-Leu-Ser-Ser-Ile-Glu-Ser-Asp-Val) was designed as the Tat-NR2B9c without the Tat and purchased from AnaSpec. Tat-NR2BAA peptide is similar to Tat-NR2B9c but with a double-point mutation in the COOH terminal tSXV motif (Tyr-Gly-Arg-Lys-Lys-Arg-Arg-Gln-Arg-Arg-Arg-Lys-Leu-Ser-Ser-Ile-Glu-Ala-Asp-Ala), making it incapable of binding PSD-95. The peptides were dissolved in sterile Milli-Q water to make 1 mM stock aliquots and stored at −80°C. Peptide aliquots were thawed only once.
Western blotting and immunoprecipitation
Cultured striatal neurons at 9–10 DIV, or hippocampal neurons at 17–20 DIV were pre-treated for 1 h with varying concentrations of Tat-NR2B9c and Tat-NR2BAA (control peptide) in the medium, then harvested and lysed in 0.8% TritonX-100 + 0.1% sodium dodecyl sulfate (SDS) or 1% NP-40-containing lysis buffer (50 mM Tris – pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA) with protease and phosphatase inhibitors, as described previously (Fan et al. 2009), and solubilized by ultrasonication. The lysates were pre-cleared with equilibrated 50% Protein A/G beads for 1 h at 4°C, and then incubated by constant rotating overnight with 50% Protein A/G beads and anti-NR2B, anti-NR2A or anti-PSD-95 antibodies (or without antibody for no antibody controls). Beads of each sample were then washed with Tris wash buffer (50 mM Tris – pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP40 in Milli-Q water) and heated at 95–99°C in protein sample buffer (0.125 M Tris – pH6.8, 2% SDS, 10% glycerol, 72 mg/mL dithiothreitol, with Pyronin Y). Paired samples were run on 8% SDS–polyacrylamide gel electrophoresis and then transferred from gels to polyvinylidene difluoride membranes (Bio-Rad Laboratories, Hercules, CA, USA) and subjected to immunoblotting. The blot was cut into two parts along the 150 kDa marker and probed with anti-NR2B (or NR2A) and anti-PSD-95 antibodies respectively. Protein bands were visualized with the enhanced chemiluminescence western blotting detection system (GE Healthcare Bio-Sciences, Piscataway, NJ, USA). For the anti-PSD-95 immunoprecipitation experiments, cut top blots were probed with anti-NR2A antibodies first and visualized with horseradish peroxidase–enhanced chemiluminescence system and then reprobed for NR2B and visualized using alkaline phosphatase –Lumi-Phos WB chemiluminescent system. Densitometry of resulting bands was performed using ImageJ.
To measure the co-immunoprecipitation (co-IP) of associated proteins, the levels of associated PSD-95 were normalized to the levels of NR2B (or NR2A) subunit co-immunoprecipitated with anti-NR2B (or anti-NR2A) in the same lane of the gel; alternatively, levels of NR2A or NR2B were normalized to PSD-95 co-immunoprecipitated with anti-PSD-95 antibodies. In both sets of experiments, band densities were compared between treatment conditions on the same blot and same exposures. Notably, similar amounts of NR2 and PSD-95 proteins were solubilized using either 0.8% TritonX-100 + 0.1% SDS or 1% NP-40 detergents, and the relative ratios of NR2 subunits co-immunoprecipitated with PSD-95 after treatment with the different Tat peptides were similar for the two detergents (comparing results from n = 7 to 13 experiments with each detergent).
Cultured hippocampal neurons on coverslips were used for patch clamp recording at 13 to 16 DIV (‘mature’). Conventional whole-cell patch clamp recording was conducted as previously described (Hamill et al. 1981). Electrodes were fabricated from borosilicate glass (Warner Instruments, Hamden, CT, USA) using a Narashige (Tokyo, Japan) PP-83 electrode puller. Open tip resistance was 4–5 MΩ for electrodes containing (in mM): 115 Cs-methanesulfonate, 10 HEPES, 20 K2-creatine phosphate, 4 MgATP, 10 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid, as well as 50 U/mL creatine phosphokinase, pH 7.26, 310 mOsm. The external solution contained (in mM): 167 NaCl, 2.4 KCl, 10 HEPES, 10 glucose, 0.2 CaCl2, pH 7.3 (325 mOsm); tetrodotoxin (300 nM) and glycine (50 μM) were added just before use. Agonist (1 mM NMDA) was dissolved in the same solution used to bathe the cells and gravity-fed to the cells through one side of a theta-tube (Chen et al. 1997). All other drugs, dissolved in the external bathing solution, were included in both the control and agonist side of the theta-tube. Tat-NR2BAA and Tat-NR2B9c (200 nM) was included in external solution for 30 min, whereas 1 μM NR2B9c (peptide with no Tat) was dissolved in solution used to back-fill the recording electrodes and allowed to diffuse into the cell after establishment of the whole-cell configuration. We avoided using Tat-linked NR2B9c for experiments testing 1 μM NR2B9c, because we found that the Tat-linked peptide at concentrations > 200 nM made patch clamp membrane seal formation difficult in pilot trials. Computer-controlled solenoid-driven valves were used to rapidly switch between the two solutions. Agonist was applied for 10 s at 1-min intervals. All recordings were made in voltage-clamp mode at a holding potential of −70 mV. Data were acquired using the Axopatch 200B patch-clamp amplifier (Axon Instruments, Foster City, CA, USA). Currents were filtered at 1 kHz and digitized at 10 kHz. pClamp 8.1 software (Axon Instruments) was used for data acquisition and analysis. Series resistance and cell capacitance were regularly monitored and recordings were abandoned if series resistance exceeded 20 MΩ.
Postnatal mouse striatal neurons (DIV 9) or embryonic rat hippocampal neurons (DIV 18 or 19) or co-cultured rat cortical and striatal neurons (DIV14), cultured in 24-well plates were pre-treated for 1 h with 1 μM or 200 nM Tat-NR2B9c or Tat-NR2BAA peptide and/or 100 μM or 1 mM Nω-nitro-l-arginine (N-Arg, nNOS inhibitor, from Sigma-Aldrich, St. Louis, MO, USA), then incubated with or without 100 or 500 μM NMDA for 10 min. After NMDA treatment, hippocampal and striatal cultured neurons were washed once with warm plating medium and then incubated in conditioned plating medium (without Tat peptides or N-Arg) for 24 h; co-cultured striatal and cortical neurons were maintained for just 6 h after NMDA before fixing and assessing cell death to preserve the YFP signal. Then cells were washed with phosphate-buffered saline (PBS) once and fixed with 4% paraformaldehyde for 30 min.
TUNEL assay and assessment of apoptosis
Fixed neuronal cultures on coverslips were numbered according to different conditions and coded before staining to ensure that the operator was blinded during the subsequent processing and analysis of immunofluorescence. Striatal and hippocampal neuronal cultures were stained with TUNEL (Roche Applied Science, Laval, Quebec, Canada) reagent and 10 μM Hoechst 33342 as described previously (Fan et al. 2009). Co-cultured striatal and cortical neurons were permeabilized with 0.5% Triton X-100 in PBS (PBST) with 0.5% sodium citrate on ice for 5 min after fixation, and then washed once with 0.03% PBST and incubated in PBST with 4% normal goat serum for 45 min with shaking at ∼22°C. After one wash with PBST, cells were immuno-stained with chicken polyclonal anti-GFP antibody (1 : 1000; catalog #ab13970; Abcam), followed by three washes with PBST then anti-chicken Alexa Fluor 488 (1 : 1000; A-11039; Invitrogen, Carlsbad, CA, USA), with 1 h shaking at ∼22°C for each staining. After GFP staining, co-cultured cells were washed three times in PBST then incubated for 10 min with 10 μM Hoechst 33342 at ∼22°C and washed again in PBST. Coverslips were then mounted on slides with Fluoromount-G.
The percentage of apoptotic cell death for hippocampal or striatal mono-cultured neurons was assessed by counting the numbers of TUNEL-positive cells (green fluorescent channel), which also showed condensed and blebbed nuclear morphology in the blue fluorescent channel, then dividing by the total number of Hoechst positive cell nuclei (blue channel) and multiplying by 100. At least 1000 neurons were counted per condition. The percentage of apoptotic cell death of neurons exposed to balanced salt solution alone (no NMDA treatment) was subtracted as a baseline from each of the other conditions in each experiment, and ranged from 3% to 25%.
The percentage of apoptotic striatal cells co-cultured with cortical neurons was assessed by counting the numbers of YFP-positive striatal cells that showed condensed and blebbed nuclear morphology in the blue fluorescent channel and then dividing by the total number of YFP-positive cells in the green fluorescent channel and multiplying by 100. A total of 100 YFP-positive striatal neurons were counted per condition.
Antibodies used for immunoprecipitation included: rabbit polyclonal anti-NR2B (1 μg/mL, Upstate/Millipore, Billerica, MA, USA, #06-600); mouse monoclonal anti-NR2B (1 μg/mL, Affinity BioReagents, Rockford, IL, USA, #MA1-2014); rabbit polyclonal anti-NR2A (1 μg/mL, Upstate/Millipore, #07-632). For probing western blots, antibodies were: anti-PSD-95 (1 μg/mL, UC Davis/NIH NeuroMab Facility, #75-028), anti-PSD-95 (1:200, Affinity BioReagents, MA1-045), anti-NR2B (1 μg/mL, Affinity BioReagents, #MA1-2014). Horseradish peroxidase-linked secondary antibodies for western blotting were anti-mouse IgG antibody (NA931V) and anti-rabbit IgG antibody (NA934V), both from Amersham Biosciences, Piscataway, NJ, USA, and used at 1 : 5000.
Figures, tables, and statistical analyses were generated using Microsoft Excel, Northern Eclipse, ImageJ, Prism 5.0, or Adobe Photoshop software. Data or bars are presented as the mean ± SEM. Significant differences were determined using the unpaired or paired two-tailed Student’s t-test, or one-way and two-way anova as appropriate.
PSD-95/NR2B interaction is more sensitive to disruption by Tat-NR2B9c than PSD-95/NR2A coupling in live cultured hippocampal neurons
Previous studies showed that submicromolar concentrations of Tat-NR2B9c peptide significantly reduced NMDAR-mediated toxicity in cultured cortical and hippocampal neurons (Aarts et al. 2002). However, the apparent concentration required to disrupt binding between NR2B subunits and PSD-95 in brain lysates was 1000-fold higher, and NR2A/PSD-95 coupling was unaffected by Tat-NR2B9c in this assay. Although a subsequent study using an in vitro binding assay showed that the IC50 values for inhibiting NR2A and NR2B binding to PSD-95 are 0.5 μM and ∼8 μM, respectively (Cui et al. 2007), the Tat-NR2B9c effective concentration was not determined in live neuronal cultures.
We tested the potency of Tat-NR2B9c in disrupting interactions between NR2B or NR2A and PSD-95 in mature primary hippocampal neuronal cultures. Cultures were treated with 250 nM or 1 μM Tat-NR2B9c or its control peptide Tat-NR2BAA in the medium for 1 h, then lysed and immunoprecipitated with antibodies against NR2A or NR2B. First, we found that 1 μM Tat-NR2B9c significantly reduced association of PSD-95 with NR2A by ∼50%, whereas 250 nM Tat-NR2B9c-treated cells showed no reduction in the amount of PSD-95 co-immunoprecipitated with NR2A when compared with control peptide, Tat-NR2BAA (Fig. 1a). In contrast, when experiments were repeated using the anti-NR2B antibody to immunoprecipitate PSD-95, PSD-95/NR2B binding was reduced by ∼50% with both 250 nM and 1 μM Tat-NR2B9c pre-treatment, compared with the control peptide (Fig. 1b). The difference between disruption of PSD-95 binding to NR2B and NR2A by 250 nM Tat-NR2B9c is clearly illustrated in Fig. 1c. Furthermore, when we repeated the experiments using the anti-PSD-95 antibody to immunoprecipitate NR2A and NR2B, we found again that 250 nM Tat-NR2B9c significantly reduced association of PSD-95 with NR2B, whereas no reduction was seen in the amount of NR2A co-immunoprecipitated with PSD-95 when compared with Tat-NR2BAA with 250 nM Tat-NR2B9c pre-treatment (Fig. 1d, NR2/PSD-95 ratio after treatment with Tat-NR2B9c was normalized to ratio after treatment with Tat-NR2BAA in paired experiments). For both NR2B and NR2A, there was a trend towards reduction in co-IP with PSD-95 after treatment with 1 μM Tat-NR2B9c (as illustrated in representative blot shown in Fig. 1d), but the experimental n was too low for quantification. These results suggest that the protective effects of Tat-NR2B9c against NMDAR toxicity in cortical and hippocampal neurons, which have been reported to occur at 50–100 nM concentrations of Tat-NR2B9c (Aarts et al. 2002), may be better explained by disruption of PSD-95 binding with NR2B than with NR2A. This hypothesis was further tested in the next set of experiments.
Protection against NMDAR-mediated death by Tat-NR2B9c and nNOS inhibitor in primary cultured hippocampal neurons
PSD-95 has been suggested to contribute to NMDAR-mediated cell death by anchoring nNOS in close proximity with NMDAR-mediated calcium influx (Sattler and Tymianski 2000; Aarts et al. 2002). Tat-NR2B9c peptide has been shown to reduce NMDA-induced excitotoxicity in cultured cortical neurons, presumably by interfering with either the NR2/PSD-95, or PSD-95/nNOS, interaction (Sattler and Tymianski 2000; Aarts et al. 2002). We therefore pre-treated cultured hippocampal neurons with Tat peptides and/or the nNOS inhibitor N-Arg to determine whether Tat-NR2B9c could protect these neurons from NMDA-induced toxicity and whether this protection is occluded by, or additive to, that afforded by the nNOS inhibitor.
We found that 200 nM Tat-NR2B9c significantly reduced NMDA-induced cell death in cultured hippocampal neurons (by ∼50% compared with treatment with NMDA alone), and that higher concentrations of Tat-NR2B9c did not provide any further protection (Fig. 2a). Moreover, the percentage of NMDA-induced apoptotic neurons in the Tat-NR2B9c-treated condition was equivalent to that observed with 100 μM of the nNOS inhibitor, N-Arg (Fig. 2b and c). Combined pre-treatment with Tat-NR2B9c and N-Arg showed slightly more protection against NMDA-induced apoptosis, but this was not significantly different from Tat-NR2B9c or N-Arg treatment alone. Both 100 (Fig. 2b and c) and 500 μM (Fig. 2a) NMDA caused a similar percentage of cell death (∼60%) in hippocampal cultures. Notably, the concentration (200 nM) of Tat-NR2B9c at which maximum protection (approximately 50%) against NMDA receptor-mediated toxicity is observed correlates well with the concentration (250 nM) that disrupts NR2B/PSD-95 coupling by approximately 50%. Furthermore, treatment of neuronal cultures with Tat-NR2B9c alone at concentrations ranging from 50 to 2000 nM did not affect the baseline level of neuronal apoptosis (data not shown).
Effect of Tat-NR2B9c on NMDA receptor desensitization in cultured hippocampal neurons
Desensitization of NMDA receptors (the decay of current with prolonged agonist exposure) has been shown to protect neurons from excitotoxicity (Zorumski et al. 1990). Our previous studies showed that NMDAR desensitization is reduced upon NMDAR binding to PSD-95 or other PSD-95 family members that carry highly homologous PDZ1-2 domains (Li et al. 2003; Sornarajah et al. 2008). In transfected HEK293 cells, the reduction in NMDAR desensitization caused by PSD-95 binding is more robust in cells transfected with NR1/NR2A than those expressing NR1/NR2B (Sornarajah et al. 2008), possibly because NR1/NR2A exhibits more extensive desensitization than NR2B-type NMDARs in this heterologous expression system (Krupp et al. 1996; Dingledine et al. 1999). Notably, a previous study found no effect of 50 nM Tat-NR2B9c on NMDAR-mediated calcium transients in cultured primary cortical neurons or on synaptic NMDAR current recorded from acute hippocampal slices, and concluded that the protective effect of the Tat peptide against NMDAR excitotoxicity could not be attributed to altered NMDAR function (Aarts et al. 2002). Our data suggest concentrations of Tat-NR2B9c higher than 50 nM (i.e. 200 nM) are required to maximally protect hippocampal neurons from NMDA toxicity. Therefore, we tested whether higher concentrations of the NR2B9c peptide could also alter NMDA-evoked current desensitization, potentially contributing to protection from excitotoxicity.
For these experiments, we intracellularly infused NR2B9c, a small interference peptide consisting of the nine amino acids from the C-terminal region of NR2B (Lys-Leu-Ser-Ser-Ile-Glu-Ser-Asp-Val) lacking the Tat peptide linkage, in whole-cell patch clamp recordings from cultured hippocampal neurons (DIV 13–16). Currents mediated by NMDARs were recorded at least 10 min after breaking into the cell, to allow diffusion of the NR2B9c peptide into intracellular compartments. We found that the NMDA-evoked steady-state-to-peak current ratio (Iss/Ip) was diminished in neurons infused with 1 μM NR2B9c peptide, to a value significantly different from that recorded in control neurons (0.43 ± 0.02, n = 8, in 1 μM NR2B9c-infused neurons versus 0.61 ± 0.05, n = 8 in control neurons; p < 0.001 by unpaired t-test) (Fig. 3a and b). In contrast, 200 nM Tat-NR2B9c peptide had no effect on NMDAR desensitization (0.61 ± 0.01, n = 8, in 200 nM Tat-NR2B9c-treated neurons versus 0.62 ± 0.05, n = 7 in control neurons, and 0.58 ± 0.05, n = 4 in Tat-NR2BAA treated neurons; not significant by one-way anova) (Fig. 3c). The NR2B9c peptide had no effect on peak NMDA-evoked current at either concentration (1180 ± 587 pA in 1 μM NR2B9c-infused neurons versus 1211 ± 231 pA in control neurons; 1343.5 ± 382.3 pA in 200 nM Tat-NR2B9c-treated neurons versus 1136.5± 447.1 pA in control neurons, versus 1199.4 ± 202.2 pA in neurons treated with Tat-NR2BAA). Because the effect of NR2B9c on NMDA receptor desensitization is observed at a concentration of 1 μM, but not at 200 nM, we conclude that this peptide increases NMDAR desensitization by uncoupling PSD-95 from NR2A-type NMDA receptors in mature hippocampal neurons. These data indicate that the protective effect of Tat-NR2B9c against NMDA-induced toxicity, which is maximal at a peptide concentration of 200 nM, occurs independently of any effect on calcium or current influx through NMDA receptors.
Lack of protection by Tat-NR2B9c or nNOS inhibitor against NMDA-induced cell death in cultured striatal neurons
Although Tat-NR2B9c peptide has been shown to reduce NMDAR-mediated excitotoxicity in cultured embryonic rat cortical and hippocampal neurons (Aarts et al. 2002; Soriano et al. 2008) (Fig. 2), its efficacy in protecting other neuronal types from NMDA-induced apoptosis had not been extensively investigated. Similar to hippocampal neurons, we found that 200 nM Tat-NR2B9c reduces NR2B/PSD-95 co-immunoprecipitation by ∼65% compared with Tat-NR2BAA, after incubating cultured postnatal murine striatal neurons for 1 h with these peptides (Fig. 4a). The PSD-95/NR2B association ratio is 1.08 ± 0.03 for Tat-NR2BAA treated neurons, and 0.37 ± 0.09 for Tat-NR2B9c treated neurons (n = 3 batches of cultures and paired experiments, p < 0.05 by paired Student’s t-test). Surprisingly however, neither Tat-NR2B9c (200 nM) nor the nNOS inhibitor N-Arg (100 μM and 1 mM) protected against NMDA-induced apoptosis in these striatal cultures (Fig. 4b and c).
One possible explanation for the lack of protective effect is that cultured postnatal murine neurons respond differently than embryonic rat neurons to excitotoxic insults. Therefore, we repeated experiments using the Tat-NR2B9c peptide at a range of concentrations in embryonic rat striatal cultures. Peptide concentrations as high as 1 μM produced no effect on NMDA toxicity in these embryonic rat striatal neurons. The percent apoptosis after subtracting paired untreated baseline were: 18 ± 3% for NMDA alone; 21 ± 2% for NMDA + 200 nM Tat-NR2BAA; and 17 ± 4%, 19 ± 4%, 23 ± 3%, 22 ± 3% for NMDA plus Tat-NR2B9c at concentrations of 50 nM, 200 nM, 500 nM, and 1 μM, respectively (no significance was detected by one-way anova, p > 0.05, compared with NMDA group, by Bonferroni post-tests; N = 5 independent experiments from different culture batches in which all treatments were performed in parallel in each experiment).
Another reason why cultured striatal neurons might appear resistant to rescue by Tat-NR2B9c is because those used in the above experiments are relatively immature (DIV 9–10) and have few excitatory synapses, because of the paucity of glutamatergic input (∼3% of neurons are VGluT-positive in these cultures) (Shehadeh et al. 2006); therefore, the role of PSD-95 in excitotoxicity may be lessened. To test whether striatal neurons that receive appropriate glutamatergic input and are more mature can be rescued by Tat-NR2B9c from NMDA-induced excitotoxicity, we tested this peptide at 200 nM and 1 μM concentrations in embryonic rat striatal-cortical co-cultures at DIV 14. As previously reported (Segal et al. 2003), these striatal neurons appeared far more spiny than striatal neurons in mono-culture, although this was not quantified. Similar to results in the striatal mono-cultures, we found no protection against 100 μM or 500 μM NMDA-induced toxicity in these more mature, cortically innervated striatal neurons (Fig. 4d and e, data not shown for 500 μM NMDA). We conclude that the NMDAR/PSD-95/nNOS pathway does not contribute significantly to NMDAR-mediated toxicity observed in cultured striatal neurons, unlike in hippocampal and cortical neurons.
The interaction between NMDARs and PSD-95/nNOS has emerged as a major contributor to excitotoxicity in cultured cortical and hippocampal neurons, as well as to ischemic damage in a rodent stroke model (Sattler et al. 1999; Sattler and Tymianski 2000; Aarts et al. 2002; Soriano et al. 2008). However, several questions remain regarding mechanisms underlying neuroprotection by disruption of the NMDAR/PSD-95/nNOS complex, such as what role NMDAR subunit composition plays, how the effect of PSD-95 on NMDAR desensitization relates to toxicity, and if previous findings can be generalized to explain mechanisms in other brain regions and neuronal types. Here, we report that pre-treatment with Tat-NR2B9c at a concentration disrupting the NR2B/PSD-95 but not NR2A/PSD-95 complex significantly reduced NMDAR-mediated cell death in hippocampal but not striatal cultures, and that similar, non-additive effects were obtained with an nNOS inhibitor. However, NMDAR desensitization was significantly increased by treatment with Tat-NR2B9c at concentrations that uncouple PSD-95 from NR2A-type NMDARs in mature hippocampal neurons, whereas lower concentrations (200 nM) of this peptide, which disrupt NR2B/PSD-95 binding and protect against NMDA toxicity, had no effect on NMDAR desensitization in the same neurons. Together, our results suggest a larger role for NR2B than NR2A in mediating death signaling in hippocampal neurons. Furthermore, modulation of NR2A-type NMDAR desensitization does not significantly impact excitotoxicity, and the NMDAR link with PSD-95 is not required for excitotoxic signaling in striatal neurons.
Sensitivity of NR2A and NR2B to Tat-NR2B9c interference with binding to PSD-95
We found that a 1-h treatment of live cultured neurons with 250 nM Tat-NR2B9c reduces the NR2B/PSD-95 interaction by ∼50%, whereas 1 μM Tat-NR2B9c is required to disrupt the NR2A/PSD-95 interaction to the same extent. Our results contrast with a report that the Tat-NR2B9c IC50 value is lower for PSD-95 PDZ2 binding to NR2A than for NR2B – 0.5 μM and ∼8 μM, respectively (Cui et al. 2007). However, the in vitro assay used in that study does not reproduce physiological conditions in live neurons. Specifically, the NR2 PDZ ligand-binding affinity may be affected by other protein-binding partners or post-translational modifications close to this domain. For example, phosphorylation of Ser1480 by casein kinase II within the NR2B C-terminal PDZ ligand Glu-Ser-Asp-Val (ESDV) disrupts interaction of NR2B with the PDZ domains of PSD-95 and SAP102 (Chung et al. 2004). Furthermore, binding of proteins to sites near the distal C-terminal ESDV, such as NR2B 1472–1475 YEKL and NR2A 1320–1321 dileucine (LL) that mediate binding to the AP2 complex (Lavezzari et al. 2004; Lau and Zukin 2007), could affect PDZ-containing protein interaction with ESDV differently in NR2A and NR2B. In addition, a recent study identified distinct accessory PDZ-binding motifs within NR2A and NR2B, which may differentially modulate PSD-95 interaction with these two subunits (Cousins et al. 2009).
Our data demonstrate that fourfold higher concentrations of Tat-NR2B9c are needed to achieve similar disruption of the NR2A/PSD-95 compared with NR2B/PSD-95 complex (1 μM vs. 0.25 μM, respectively) in live cultured hippocampal neurons. Consistent with this finding, a previous study showed ex vivo treatment with 10 μM Tat-NR2B9c reduced co-immunoprecipitation of PSD-95 with NR2B, but not with NR2A, in rat forebrain tissue lysates (Aarts et al. 2002). Furthermore, in vivo intrastriatal injection of 500 μM Tat-NR2B9c selectively dissociated the NR2B/PSD95 but not NR2A/PSD-95 complex (Gardoni et al. 2006), and 1 μM Tat-NR2B9c peptide disrupted PSD-95 co-immunoprecipitation with NR2B but not NR2A in hippocampal slices (Gardoni et al. 2009). Because of the differences in methods of application of Tat peptides and membrane fractions analyzed, it is difficult to directly compare results with those of our study. In particular, the effective concentration of Tat-NR2B9c required to disrupt NR2A/PSD-95 binding may not have been achieved by intrastriatal peptide injection or incubation of acute brain slices with 1 μM peptide in the previous studies.
It is interesting that treatment of live hippocampal neurons with Tat-NR2B9c concentrations of 250 nM and 1 μM produced identical results in dissociating 50% of PSD-95 from NR2B. Although additional PDZ-binding sites proximal to the C-terminal PDZ ligand in NR2B (Cousins et al. 2009) may contribute to limiting the efficacy of Tat-NR2B9c in disrupting PSD-95 binding, another factor may be that there are distinct populations of NMDARs with respect to subcellular compartments, to which Tat-NR2B9c has differential access. These populations could correspond to surface extrasynaptic versus synaptic receptors, as well as those in intracellular membranous compartments such as the Golgi or endosomes. It is possible that lower concentrations of Tat-NR2B9c (e.g. ≤ 1 μM) are largely excluded from the post-synaptic density. Consistent with this idea, further increase in Tat-NR2B9c to 10 μM reduces PSD-95 co-IP with NR2B by 70 ± 13% (n = 4). Lower concentrations of Tat-NR2B9c may access mainly surface extrasynaptic NMDARs and those in intracellular vesicular compartments; both of these populations have been shown to interact with PSD-95, PSD-93, and SAP102 (Sans et al. 2003; Petralia et al. 2010). In fact, a recent study indicates surface NMDARs in hippocampal neurons are distributed equally between extrasynaptic and synaptic sites, and a substantial proportion of the surface extrasynaptic NMDARs co-localize with PSD-95 family members (Petralia et al. 2010). However, the fact that maximal protection by Tat-NR2B9c against NMDA-induced toxicity in hippocampal neurons was also ∼50% in our paradigm may be, in part, explained by the contribution of NMDAR cell death signaling via JNK pathways, independent of the NMDAR/PSD-95/nNOS complex (Soriano et al. 2008).
Role of NMDAR subunit composition in excitotoxic signaling via PSD-95 and nNOS
Previous studies have shown that Tat-NR2B9c peptide concentrations on the order of 100 nM protect cultured hippocampal and cortical neurons from NMDAR-mediated toxicity (Aarts et al. 2002; Soriano et al. 2008). However, whether this effect could be attributed to Tat-NR2B9c disrupting PSD-95 interactions between NR2B and/or NR2A in live neurons had not been established previously. We report that the concentration dependence of the Tat-NR2B9c neuroprotective effect in hippocampal cultures correlates well with the concentration required to uncouple NR2B and PSD-95 in live neurons, suggesting signaling pathways that require the NR2B interaction with PSD-95 critically contribute to apoptotic cell death 24 h after brief (10 min) exposures to NMDA. Notably, further increasing the Tat-NR2B9c concentration to levels required to disrupt NR2A binding with PSD-95 did not augment protection from NMDA-induced toxicity in this paradigm. The nNOS inhibitor N-Arg provided a similar level of neuroprotection (∼50%), and this effect was not additive with Tat-NR2B9c. Together, these data suggest that it is the NR2B link through PSD-95 to activate nNOS that is critical for mediating delayed apoptotic death in cultured hippocampal neurons. Our results also support data from other studies suggesting NR2B-type NMDARs preferentially signal to cell death pathways, including shut-off of ERK1/2 and activation of p38-MAPK, whereas NR2A-type NMDARs activate survival pathways (Kim et al. 2005; Liu et al. 2007; Martin and Wang 2010; Tu et al. 2010).
Cell death pathways downstream of NMDAR activation in striatal neurons
Hippocampal (or cortical) and striatal neurons express many similar proteins but may differ in terms of protein expression levels, subcellular compartmentalization, and/or protein-protein interactions, potentially contributing to differences in pathways downstream of NMDAR signaling to cell death or thresholds of common pathways. One difference is that NR2B is enriched in striatal tissue (Landwehrmeyer et al. 1995; Christie et al. 2000; Li et al. 2003), whereas NR2A- and NR2B-type NMDARs are both common in adult hippocampal cells (Monyer et al. 1994; Wenzel et al. 1997). Unlike hippocampal neurons, we found that cultured striatal neurons are not protected by Tat-NR2B9c, in spite of the fact that these two neuronal populations show similar Tat-NR2B9c efficacy in disrupting NR2B/PSD-95 binding. Our data are consistent with a previous study showing that Tat-NR2B9c protection against middle cerebral artery occlusion-induced ischemia is more robust in cortex than striatum (Aarts et al. 2002). Moreover, in pilot studies we found that intraperitoneal injection of Tat-NR2B9c does not protect against quinolinate-induced (NMDAR-mediated) lesions in murine striatum (BR Leavitt, J Fan, LA Raymond, unpublished results). Consistent with the lack of protection by Tat-NR2B9c in cultured striatal neurons, nNOS inhibition also did not protect against NMDA-induced toxicity. Moreover, in previous work we had shown that although NMDA stimulates nNOS activity in wild-type murine striatal cultures, as assessed by changes in cGMP levels, this increase was not affected by pre-treatment with Tat-NR2B9c (Fan et al. 2009). We conclude that NMDAR-mediated apoptotic signaling occurs through distinct pathways in hippocampal and striatal neurons.
The fact that the NR2B/PSD-95/nNOS pathway is not a major contributor to NMDA-induced toxicity in cultured striatal neurons may be due, in part, to the absence of nNOS in the striatal GABAergic medium-sized spiny neurons, which make up ∼90% of all striatal neurons. However, the striatal cultures used in our experiments include ∼4% nNOS-expressing interneurons (Shehadeh et al. 2006), and a previous study showed that NO released from such neurons could induce intracellular calcium release from mitochondria in medium-sized spiny neurons (Horn et al. 2002). However, NR2B expression is undetectable in those nNOS interneurons (Zucker et al. 2005), and in our striatal cultures the nNOS inhibitor N-Arg fully reverses NMDA-induced cGMP production whereas Tat-NR2B9c has no effect, (Fan et al. 2009), indicating nNOS activation by NMDA occurs independently of the PSD-95/NR2B interaction. Moreover, the NMDA-induced cGMP and NO production in these striatal cultures is apparently insufficient to trigger cell death signaling, because we have shown that nNOS inhibitors fail to block NMDA-induced toxicity.
In cerebellar granule cells, NMDA-induced p38 MAPK activation is nNOS-dependent (Cao et al. 2005), and p38 MAPK is also considered a downstream effector of NO toxicity in cortical neurons (Ghatan et al. 2000). Notably, NMDA-induced p38 MAPK activation is disrupted by Tat-NR2B9c and relies on neuronal context in cortical neurons (Soriano et al. 2008). In addition to nNOS, the binding of PSD-95 to synaptic Ras-GTPase-activating protein (SynGAP) has been shown to regulate neuronal apoptosis through the p38 MAPK pathway in hippocampal neurons (Kim et al. 1998; Rumbaugh et al. 2006). However, it is more likely that other pathways downstream of NMDAR activation, independent of the NR2/PSD-95 interaction, contribute importantly to striatal neuronal apoptosis. One such pathway may involve JNK MAPK (Soriano et al. 2008). In future experiments, it would be interesting to determine whether p38 MAPK or SynGAP play any role, or if JNK MAPK activation is a major contributor, in striatal neuronal NMDA-induced toxicity.
We thank A.J. Milnerwood and A. Kaufmann for advice on striatal-cortical co-cultures preparation, and E. Yu for assistance in preparing cultures. The research was funded by operating grants from the Canadian Institutes of Health Research (CIHR MOP-102517 and 129029 to L.A.R.), and a Michael Smith Foundation for Health Research (MSFHR) infrastructure grant. L.A.R. was a CIHR Investigator and MSFHR Senior Scholar.