Activity-dependent NMDA receptor-mediated activation of protein kinase B/Akt in cortical neuronal cultures

Authors


Address correspondence and reprint requests to Dr L. J. Chandler, Department of Physiology and Neuroscience, MUSC, IOP 4 North, 67 President St, Charleston, SC 29425, USA. E-mail: chandj@musc.edu

Abstract

The serine/threonine protein kinase B (PKB)/Akt is a phosphoinositide 3-kinase (PI3K) effector that is thought to play an important roll in a wide variety of cellular events. The present study examined whether PKB activation in cortical neuronal cultures is coupled with synaptic activity. A 1-h incubation of neuronal cultures with tetrodotoxin (TTX), the PI3K inhibitor wortmannin, the NMDA receptor antagonist MK-801 or removal of extracellular calcium significantly reduced basal levels of phospho(Ser473)-PKB, indicating that activity-dependent glutamate release maintains PKB activation through an NMDA receptor-PI3K pathway. A 5-min exposure to NMDA (50 µm) in the presence of TTX increased phospho-PKB back to levels observed in the absence of TTX. NMDA stimulation of phospho-PKB was blocked by wortmannin, the CaMKII inhibitor KN-93, MK-801, and removal of extracellular calcium. We have previously shown that NMDA receptors can bi-directionally regulate activation of extracellular-signal regulated kinase (ERK), and NMDA receptor stimulation of PKB in the present study appeared to mirror activation of ERK. These results suggest that in cultured cortical neurons, PKB activity is dynamically regulated by synaptic activity and is coupled to NMDA receptor activation. In addition, NMDA receptor activation of ERK and PKB may occur through overlapping signaling pathways that bifurcate at the level of Ras.

Abbreviations used
AMPA

α-amino-3-hydroxy-5-methyl-4-isoxalone

ARC

β-cytosine arabinoside

BNDF

brain-derived neurotrophic factor

CaMKII

calcium/calmodulin-dependent kinase II

DMEM

Dulbeco's modified Eagle's medium

ERK

extracellular-signal regulated kinase

GSK

glycogen synthase kinase-3

PBST

phosphate-buffered saline with Tween 20

PDHS

plasma-derived horse serum

PDK1

phosphoinositide-dependent kinase 1

PI3K

phosphoinositide 3-kinase

PKB

protein kinase B

PVDF

polyvinylidene difluoride

SDS

sodium dodecyl sulfate

TTX

tetrodotoxin.

Glutamate is the major excitatory neurotransmitter in the mammalian central nervous system (CNS) and mediates fast excitatory neurotransmission through activation of ionotropic glutamate receptors that include NMDA, α-amino-3-hydroxy-5-methyl-4-isoxalone (AMPA) and kainate receptor subtypes. AMPA receptors mediate the fast component while NMDA receptors mediate the slow component of fast excitatory neurotransmission (Dingledine et al. 1999). An important feature of NMDA receptors is their high permeability to calcium, placing them in a unique position to control numerous calcium-dependent process associated with synaptic activity. During periods of brain development and maturation, NMDA receptors can produce both neurotrophic and neurotoxic effects (Meldrum 2000). It is well established that excessive activation of NMDA receptors can lead to calcium-dependent cell death in neurons through a process known as excitotoxicity, while considerable evidence also exists for the involvement of NMDA receptors in synaptic plasticity and neuronal development/survival through activation of various intracellular signaling pathways, including extracellular-signal regulated kinases (ERKs) (Orban et al. 1999; Sweatt 2001; Adams and Sweatt 2002).

Phosphoinositide 3-kinases (PI3K) are agonist-activated lipid signaling enzymes that initiate signaling cascades that play a critical role in a variety of cellular processes most commonly associated with growth and survival (Kennedy et al. 1997). Activation of PI3K protects cells from apoptosis and has emerged as a master regulator of cell survival. While a number of effectors of PI3K have been identified, recent studies show that the pro-survival actions of PI3K occur through activation of the anti-apoptotic effector protein kinase B (PKB) (Toker 2000). PKB is a cellular homolog of v-Akt, a protein encoded in the genome of the Akt-8 acute transforming virus isolated from rodent T-cell lymphoma (Bellacosa et al. 1991). Akt/PKB compose an evolutionarily conserved family of kinases that include three known isoforms; Akt1/PKBα, Akt2/PKBβ, and Akt3/PKBγ. The lipid products of PI3K, namely PtdIns(3,4)P2 and PtdIns(3,4,5)P3, bind to the pleckstrin homology domain of PKB and mediate its translocation to the plasma membrane. At the plasma membrane, PKB activation is regulated primarily through phosphorylation of Thr-308 by phosphoinositide-dependent kinase 1 (PDK1) and phosphorylation of Ser473 by either autophosphorylation or an uncharacterized PDK2 (Allessi and Cohen 1998; Toker and Newton 2000). Downstream effectors of PKB regulate growth and survival signals in a number of cell systems including the brain, and pharmacological inhibition of the PKB cascade has been demonstrated to initiate apoptosis in neurons (Hetman et al. 2000; Brunet et al. 2001). Consistent with its importance in neuronal function, PKB is highly expressed in the mammalian central nervous system, with a heterogeneous distribution of the PKB isoforms in brain.

Ras is the prototypical member of the family of small G-proteins that couple to activation of the ERK signaling cascade (Campbell et al. 1998). Ras has also been linked to activation of PI3K in various cell types and may thus play a role in PI3K activation of a PKB (Downward 1997). In the brain, NMDA receptor stimulation has been shown to increase the active GTP bound form of Ras, leading to ERK activation (Finkbeiner and Greenberg 1996). Further supporting an NMDA-Ras-PI3K signaling process, it has recently been reported that the SH2 domains of the p85 subunit of PI3K can bind tyrosine phosphorylated NR2B subunits of the NMDA receptor (Hisatsume et al. 1999). In light of the evidence of cross-talk between NMDA receptors, Ras and the PI3K signaling cascade, we investigated whether NMDA receptor stimulation could lead to increases in PKB phosphorylation in primary cortical neurons.

Materials and methods

Female Sprague–Dawley rats were obtained from Harlan and housed and bred in our animal facility. Dulbeco's modified Eagle's medium (DMEM) and amphotericin B were purchased from Gibco (Grand Island, NY, USA). Plasma-derived horse serum (PDHS), brain-derived neurotrophic factor (BDNF), and tetrodotoxin was purchased from Sigma (Saint Louis, MO, USA). Trypsin was purchased from Worthington Biochemicals (Freehold, NJ, USA) and penicillin/streptomycin was purchased from Pfizer Inc. (New York, NY, USA). KN-93, KN-92, calyculin A, cypermethrin, okadaic acid, wortmannin, LY294002 and protease inhibitors were purchased from Calbiochem (La Jolla, CA, USA). MK-801 and APV were purchased from Research Biochemicals Inc (Natick, MA, USA). Anti-TrkB (clone 47) neutralizing antibody (special ordered as preservative free) was purchased from Transduction Laboratories (Lexington, KY, USA). Anti-phospho(S473)-PKB, anti-phospho(T202/Y204)-ERK 1/2, anti-PKB and anti-ERK 1/2 were purchased from New England Biolabs (Beverly, MA, USA). Horseradish peroxidase-conjugated goat anti-rabbit was purchased from Southern Biotechnology Associates (Birmingham, AL, USA).

Preparation and treatment of neuronal cultures

Cortical neuronal cultures were prepared as previously described (Chandler et al. 1997). In brief, brains from new-born rat pups were removed and placed in isotonic saline (pH 7.4) containing 100 U of penicillin G, 100 µg of streptomycin and 0.25 µg of amphotericin B per milliliter. Pia matter and blood vessels were removed and the brains were minced finely using a pair of fine surgical scissors. The minced tissue was then dissociated with 0.25% trypsin (w/v) in isotonic salt solution (pH 7.4) and incubated in a shaking water bath for 10 min at 37°C, followed by a 5-min incubation with 160 µg of DNase 1 and triturated (10 times) to dissociate the cells. The cell suspension was placed in 50 mL of DMEM/PDHS and centrifuged for 10 min at 1000 g. Cells were plated in poly l-lysine precoated 35 mm culture dishes at a density of 3 × 106 cells and incubated at 37°C with 7.5% CO2. After 3 days in culture, the media was replaced with fresh DMEM/PDHS containing 10 µmβ-cytosine arabinoside (ARC). After 2 days of ARC treatment, the media was again replaced with DMEM/PDHS and cultures were grown for an additional 7–9 days before use in experiments. Each experimental number represents data collected from neuronal cultures prepared from a separate litter of pups.

At the start of each experiment, cultures were washed twice with 1 mL of HEPES buffer [140 mm NaCl, 5.4 mm KCl, 1.8 mm CaCl2, 100 nm glycine, 15 mm glucose and 25 mm HEPES (pH 7.4)]. In some experiments, CaCl2 was replaced with 100 µm EGTA. After a 10-min incubation in 1 mL HEPES, cultures were exposed to NMDA and/or tetrodotoxin for the indicated times. For experiments with NMDA receptor antagonists and kinase/phosphatase inhibitors, compounds were added 15 min prior to NMDA. Anti-TrkB was added 1 h prior to addition of NMDA of BDNF. After the indicated exposure times, cells were isolated into 100 µL cold homogenization buffer [50 mm Tris-HCl, 50 mm NaCl, 10 mm EGTA, 5 mm EDTA, 2 mm sodium pyrophosphate, 1 mm activated sodium orthovanadate, 0.2 mm 4-(2-aminoethyl)benzenesulphonyl fluoride (AEBSF), 1 µg/mL aprotinin, 1 mm benzamide, 10 µg/mL leupeptine, 10 µg/mL pepstatin (pH 7.4)]. Cells from three dishes were combined, briefly sonicated and centrifuged for 30 min at 15 000 g (4°C). Protein concentration was determined on the supernatant by the bicinchoninic acid assay (Pierce, Rockford, IL, USA).

Gel electrophoresis and immunoblotting

An aliqout of the supernatant was diluted with an equal volume of 2X sample buffer [final concentration = 50 mm Tris-HCl (pH 6.7), 4% glycerol (w/v), 4% sodium dodecyl sulfate (SDS), 1% 2-mercaptoethanol and bromophenol blue]. Samples were boiled for 5 min, and 20 µg of sample was separated on a 10% SDS-polyacrylamide gel using the buffer system of Laemmli and transferred to Millipore Immobilon-P polyvinylidene difluoride (PVDF) membranes (Bedford, MA, USA). After transfer, blots werewashed with PBS containing 0.05% Tween 20 (PBST) and then blocked with PBST containing 5% non-fat dried milk (NFDM) for 1 h at roomtemperature(24°C) with agitation. The membranes werethenincubated witheither anti-phospho-PKB (1 : 1000), anti-phospho-ERK 1/2 (1 : 1000), anti-PKB (1 : 1000), or anti-ERK 1/2 (1 : 1000), overnight at 4°C in PBST containing 5% NFDM. The membranes were then washed in PBST followed by a 1-h incubation with horseradish-conjugated goat anti-rabbit secondary antibody diluted in PBST (1 : 2000) containing 5% NFDM at room temperature. The membranes were washed in PBST following this incubation and the antigen-antibody complex was detected by enhanced chemiluminescence (Amersham, Arlington Heights, IL, USA) and visualized by exposure to Amersham Hyperfilm for ECL. In some instances, membranes were stripped after immunoblotting with phospho-PKB by incubation in stripping buffer (100 mm 2-mercaptoethanol, 2% SDS, 62.5 mm Tris-HCl, pH 6.7) for 30 min at 55°C with agitation, and membranes were then blocked and reprobed with anti-PKB. Film autoradiograms were quantified by computer-assisted densitometry, and statistical analysis performed using GB-STAT (Dynamic Microsytems Inc, Silver Springs, MD, USA).

Results

PKB is phosphorylated at two sites that are associated with activation of enzyme activity; Thr-308 in the catalytic domain and Ser473 in the cytoplasmic domain. Phosphorylation of both sites is critically dependent upon PI3K activity. PKB phosphorylation at Ser473 was determined by immunoblot analysis utilizing a phospho(Ser473)-PKB (phospho-PKB) specific antibody. Incubation of cortical cultures with the PI3K inhibitor wortmannin (200 nm) for 1 h resulted in a significant reduction in phospho-PKB (Fig. 1). A similar inhibitory effect was observed with LY294002 (100 µm; not shown), a chemically dissimilar inhibitor of PI3K. To further characterize the processes involved in maintaining basal phosphorylation of PKB, cultures were exposed to tetrodotoxin (TTX, 1 µm) for 1 h. Block of synaptic activity with TTX significantly decreased phospho-PKB, as did the NMDA receptor antagonist MK-801 (10 µm). These observations demonstrate that basal levels of phospho-PKB are coupled to synaptic-activity and to activation of NMDA receptors and PI3K, most likely through synaptic-release of glutamate.

Figure 1.

Protein kinase B (PKB) phosphorylation is coupled to synaptic-activity and activation of NMDA receptors in cortical neuronal cultures. Cortical cultures were incubated for 1 h in the presence and absence of wortmannin (200 nm), tetrodotoxin (TTX; 1 µm), or MK-801 (10 µm). Equal amounts of protein extracts were analyzed by western blot analysis with phospho(Ser473)-PKB. (a) Representative immunoblots of phospho(Ser473)-PKB. Each panel is taken from a separate immunoblot. (b) Quantification of immunoblot data by densitometry analysis of band intensity. Data is expressed as a percentage of eachrespective basal optical density value and represent the means ± SEM of six independent experiments. *p < 0.05; **p < 0.01 versus respective controls. (anova with Newman–Keuls post-hoc test performed on optical density values).

The next set of experiments were designed to further define the coupling of NMDA receptors with PI3K-PKB activation. After a 1-h incubation with TTX to reduce basal levels of phospho-PKB, a subsequent 5 min exposure to NMDA (50 µm) significantly enhanced phospho-PKB levels (Figs 2a and b). This increase was blocked by prior addition of MK-801 (10 µm) or APV (50 µm, not shown), demonstrating this is an NMDA receptor-mediated event. As expected, the NMDA-stimulated increase in phospho-PKB was blocked by inhibition of PI3K with wortmannin (Figs 2c and d). NMDA stimulation of phospho-PKB was also dependent upon the presence of extracellular calcium (Fig. 2e).

Figure 2.

NMDA stimulates a rapid increase in phospho-protein kinase B (PKB) through a phosphoinositide 3-kinase (PI3K)-dependent process in the presence of tetrodotoxin (TTX). Cultures were incubated with TTX (1 µm) for 1 h followed by a 5-min exposure to NMDA (50 µm) in the presence and absence of either MK-801 (10 µm) or wortmannin (200 nm). Equal amounts of protein extracts were analyzed by western blot analysis with phospho(Ser473)-PKB. (a and c) Representative immunoblots of phospho(Ser473)-PKB. (b and d) Quantification of immunoblot data by densitometry analysis of band intensity. Values are expressed as optical density and represent the means ± SEM of either eight (b) or six (d) independent experiments. *p < 0.01 (b) and *p < 0.05 (d) versus respective TTX alone; Ψp < 0.01 versus respective TTX + NMDA. (anova with Newman–Keuls post-hoc test). (e) Representative immunoblot showing 5 min NMDA (50 µm) stimulation of phospho-PKB in buffer containing 1.8 mm Ca2+ or 0.1 mm EGTA.

A well established pathway for activation of PI3K signaling is via stimulation of neurotrophic factor receptors. NMDA has been demonstrated to stimulate the rapid release of BDNF in hippocampal slices (Marini et al. 1998), and NMDA could therefore increase phospho-PKB indirectly by stimulating the release of BDNF. To examine this possibility, cultures were incubated (1 h) with an antibody to the BDNF TrkB (5 µg/mL) receptor. This concentration of antibody has previously been reported to inhibit BDNF-dependent survival of embryonic nodose/petrosal ganglia neuronal cultures by blocking BDNF binding to the TrkB receptor (Balkowiec and Katz (2000). As shown in Fig. 3(a and b), NMDA-stimulation of phospho-PKB was not altered by the TrkB antibody. However, subsequent control experiments to demonstrate theblocking effectiveness of this antibody under these experimental conditions revealed it also failed to prevent the robust stimulation of phospho-PKB produced by BDNF (100 ng/mL) after 5 or 15 min of stimulation (Fig. 3c).

Figure 3.

NMDA and brain-derived neurotrophic factor (BDNF) stimulation of phospho-protein kinase B (PKB) in the presence of a TrkB receptor antibody. Neuronal cultures were incubated with tetrodotoxin (TTX) (1 µm) for 1 h in the presence and absence of a TrkB neutralizing antibody (5 µg/mL) followed by a 5-min stimulation with NMDA (50 µm). Equal amounts of protein extracts were analyzed by western blot analysis with phospho(Ser473)-PKB. (a) Representative immunoblots of phospho(Ser473)-PKB. (b) Quantification of immunoblot data by densitometry analysis of band intensity. Values are expressed as optical density and represent the means ± SEM of four independent experiments. *p < 0.05 versus TTX alone. (anova with Newman–Keuls post-hoc test). (c) Representative immunoblot demonstrating failure of the TrkB antibody to block BDNF stimulation of phospho-PKB.

We have previously reported that the NMDA dose–response curve for stimulation of phospho-ERK in the presence of TTX is biphasic, where a high concentration of NMDA (100 µm) produced less of an increases in phospho-ERK compared to lower concentrations of NMDA (10–50 µm) (Chandler et al. 2001). Figure 4 shows that the dose–response curve for NMDA stimulation (5 min) of phospho-PKB and phospho-ERK appear similar. A parallel dose-dependent increase in phospho-PKB and phospho-ERK occurred with 10 and 25 µm NMDA, while a high concentration of NMDA (100 µm) produced a significantly attenuated increase in comparison to 25 µm NMDA. We and others have also observed that NMDA stimulation of phospho-ERK is attenuated by inhibition of CaMKII (Perkinton et al. 1999; Chandler et al. 2001). As shown in Fig. 5, the NMDA-stimulated increase in phospho-PKB was significantly reduced by the selective CaMKII inhibitor KN-93 (40 µm) but not by the inactive analog KN-92 (40 µm; not shown). This suggests that CaMKII also plays a role in NMDA activation of PI3K-PKB signaling.

Figure 4.

NMDA dose–response curves for stimulation of phospho-PKB and phospho-ERK in the presence of tetrodotoxin (TTX). Neuronal cultures were incubated in the presence and absence of TTX (1 µm) for 1 h followed by exposure to the indicated concentrations of NMDA. Equal amounts of protein extracts were analyzed by western blot analysis with phospho(Ser473)-PKB or phospho(Thr202/Tyr204)-ERK 1/2. (a) Representative immunoblots of phospho(Ser473)-PKB. (b) Quantification of immunoblot data by densitometry analysis of band intensity. For phospho-ERK 1/2, only the ERK 2 band was used for quantification. Values are expressed as optical density and represent the means ± SEM of 11 independent experiments. *p < 0.01 versus respective TTX alone; **p < 0.001 versus respective 10 µm NMDA; Ψp < 0.001 versus respective 25 µm NMDA (anova with Newman–Keuls post-hoc test).

Figure 5.

NMDA stimulation of phospho-protein kinase B (PKB) in the presence of tetrodotoxin (TTX) is attenuated by inhibition of calcium/calmodulin-dependent kinase II (CaMKII). Cultures were incubated with TTX (1 µm) for 1 h followed by a 5-min exposure to NMDA (50 µm) in the presence and absence of either KN-93. Equal amounts of protein extracts were analyzed by western blot analysis with phospho(Ser473)-PKB. (a) Representative immunoblots of phospho(Ser473)-PKB. (b) Quantification of immunoblot data by densitometry analysis of band intensity. Values are expressed as optical density and represent the means ± SEM of seven independent experiments. *p < 0.05 versus TTX + NMDA; Ψp < 0.05 versus NMDA alone. (anova with Newman–Keuls post-hoc test).

The next set of experiments further characterized the apparent over-lap in the processes coupling NMDA receptors to ERK and PKB signaling. We have shown that without a reduction of the high basal levels of activity-dependent phospho-ERK by prior block of synaptic activity with TTX, stimulation of NMDA receptors fails to produce a further increase in phospho-ERK levels, and instead produces a significant decrease (Chandler et al. 2001). Since a number of similarities in NMDA stimulation of ERK and PKB exist, we examined whether NMDA may also exert a similar inhibitory effect on basal activity-dependent phospho-PKB. As shown in Fig. 6(a and b), NMDA (100 µm) in the absence of TTX significantly decreased basal levels of phospho-PKB after 1 h. At no other NMDA exposure time points (1, 5, 20 min) did NMDA enhance either phospho-PKB or phospho-ERK in the absence of TTX, indicating this was not a temporal biphasic response at this concentration of NMDA. In addition, the 1 h NMDA exposure did not alter the total levels of either PKB or ERK 1/2 (not shown). Incubation of cultures with a mixture of the PP1/PP2A inhibitor calyculin A (200 nm) and the calcineurin inhibitor cypermethrin (10 nm) resulted in a large increase in basal phospho-PKB (Figs 6c and d) and phospho-ERK (Chandler et al. 2001) but did not appear to prevent the inhibitory effect of NMDA. However, these results must be interpreted with caution due to the complication of the large increase in phospho-PKB in the presence of the phosphatase inhibitors. Examination of each inhibitor separately revealed that the large increase in phospho-PKB (Fig. 6e) and phospho-ERK (Chandler et al. 2001) resulted from inhibition of PP1/PP2A and not inhibition of calcineurin. A similar increase (and lack of effect on NMDA inhibition) was observed with the PP1/PP2A inhibitor okadaic acid (1.5 µm, not shown). A separated inhibitor of calcineurin (cyclosporin A; 1 µm) was also without effect on NMDA inhibition of phospho-PKB (Fig. 6f), further demonstrating that activation of calcineurin is not involved in NMDA inhibition of phospho-PKB. Finally, removal of extracellular calcium reduced basal levels of phospho-PKB to that observed with NMDA, indicating that activity-dependent activation of phospho-PKB is due to calcium influx through NMDA receptors. Taken together, the similarity in the inhibitory effect of NMDA on basal phospho-ERK and phospho-PKB, and in the effects of PP1/PP2 inhibitors provides further evidence that NMDA receptors may utilize a common or over-lapping pathway for coupling to ERK and PKB signaling.

Figure 6.

In the absence of tetrodotoxin (TTX), NMDA decreases basal levels of phospho-protein kinase B (PKB). (a) Neuronal cultures were incubated in the presence and absence of NMDA (100 µm) for 1 h. (c–e) Cypermethrin and calyculin A were added either together (c and d) or separately (e) 15 min prior to a 1-h NMDA exposure. In (f), cultures were incubated with or without NMDA (100 µm) or cyclosporin A (1 µm) in buffer containing 1.8 mm Ca2+ or 0.1 mm EGTA. Equal amounts of protein extracts were analyzed by western blot analysis with phospho(Ser473)-PKB. (a, c, e and f) Representative immunoblots of phospho(Ser473)-PKB. (b and d) Quantification of immunoblot data by densitometry analysis of band intensity. Values are expressed as optical density and represent the means ± SEM of four independent experiments. *p < 0.01 (b) and *p < 0.05 (d) versus respective basal (anova with Newman–Keuls post-hoc test).

Discussion

The present study examined activity-dependent regulation of PKB in primary cortical neuronal cultures. High levels of phospho-PKB were observed even after removal of the serum-containing growth media. This basal phospho-PKB was significantly reduced by block of synaptic activity with TTX, block of NMDA receptors with MK-801, removal of extracellular calcium and by inhibition of PI3K. In the presence of TTX, stimulation of NMDA receptors induced a rapid increase in phospho-PKB back to levels observed in the absence of TTX. This increase was also dependent upon extracellular calcium and was blocked by wortmannin and MK-801, confirming that NMDA receptors couple to PKB activation through a calcium and PI3K-dependent process. In contrast to increasing phospho-PKB in the presence of TTX, stimulation of NMDA receptors in the absence of TTX significantly reduced phospho-PKB levels associated with activity-dependent activation. Taken together, these observations suggests that PKB activity is dynamically regulated by glutamatergic activity through NMDA receptors. Furthermore, the stimulatory and inhibitory effect of NMDA under different conditions of synaptic activity suggests a bidirectional control of PKB signaling by NMDA receptors.

While the rapidity of NMDA stimulation of phospho-PKB suggests this is a direct activation processes, the possibility remains that NMDA indirectly activates PKB through stimulating the release an endogenous mediator. Indeed, several lines of evidence suggest a role for BDNF in NMDA receptor activity-dependent synaptic plasticity. NMDA is known to regulate BDNF expression in cerebellar granule cells, and the neuroprotective activity of NMDA in cerebellar granule cell cultures has been associated with increased BDNF release and TrkB receptor signaling (Favoron et al. 1993; Marini et al. 1998; Bhave et al. 1999). NMDA stimulation of BDNF release has been observed to occur in cultured cerebellar granule cells within as little as 2 min, well within the time frame observed for NMDA stimulation of phospho-PKB (Marini et al. 1998). In the present study, we attempted to address the possible involvement of BDNF in NMDA stimulation of phospho-PKB by utilizing a TrkB receptor antibody. This antibody has previously been reported to function as a BDNF function blocking antibody in cultured nodose/petrosal ganglia neuronal cultures by preventing BDNF binding to the TrkB receptor. Consistent with the lack of involvement of BDNF in NMDA stimulation of phospho-PKB, incubation of cultures with the TrkB receptor antibody had no effect on NMDA stimulation of phospho-PKB. However, the failure of the TrkB antibody to block BDNF stimulation of phospho-PKB calls into question its effectiveness as a neutralizing antibody, at least under the experimental conditions employed in the present study. A recent study observed that in contrast to the calcium-dependency of NMDA stimulation of phospho-PKB, BDNF stimulation of phospho-PKB was not dependent upon extracellular calcium (Perkinton et al. 2002). While this observation does not necessarily rule out BDNF release in NMDA stimulation of phospho-PKB, it indicates that the pathways by which NMDA and BDNF receptors initiate signal transduction leading to activation of phospho-PKB are fundamentally different.

Activation of PKB promotes neuronal survival and protects against neuronal death, and the coupling of PI3K-PKB signaling to synaptic-activity may implicate this cascade in NMDA receptor-mediated synapse to nucleus signaling. PKB has been shown to mediate a wide, and ever expanding, number of downstream effectors such as the apoptotic protein BAD, the forkhead transcription factor FKHRL1 (a transcriptional regulator of many pro-apoptotic proteins), Nur 77, glycogen synthase kinase-3 (GSK-3), caspase-9, CREB, IκB-kinase-α (IKKα), Raf-1 and nitric oxide synthase (Brazil and Hemmings 2001). Phosphorylation by PKB is often associated with negative regulation by promotion of binding and sequestration with the protein 14-3-3 (e.g. BAD and FKHRL1). However, not all PKB-mediated inactivation involves sequestration by 14-3-3 (e.g. caspase-9 and GSK-3), and still other substrates are activated by PKB phosphorylation (e.g. CREB). Regardless of whether PKB activates or inhibits its substrate, the effect is to oppose apoptosis and promote cell survival and growth. It is well known that over-activation of NMDA receptors leads to delayed neuronal death through excitotoxic processes (Choi 1988). However, under certain conditions, underactivation of NMDA receptors can also result in cell death, and studies suggests that an optimal level of NMDA-mediated calcium influx during neuronal activity can promote growth and survival (Gallo et al. 1987; Hegarty et al. 1997; Yano et al. 1998; Ikonomidou et al. 1999). Our demonstration of NMDA receptor stimulation of PI3K-PKB may implicate this signaling cascade in the neurotrophic actions of NMDA receptors. NMDA receptor activation of PKB, via calcium/calmodulin-dependent protein kinase kinase (CaM-KK), has been reported in NG108 cells (Yano et al. 1998). In agreement with a neuroprotective role for PKB in NMDA receptor activity, persistent NMDA receptor activation of PKB in NG108 cells was found to protect against serum-withdrawal induced apoptosis.

The similarities in the characteristics of NMDA stimulation of phospho-ERK and phospho-PKB may reflect a common or overlapping pathway for NMDA coupling to ERK and PKB signaling (Fig. 7). NMDA stimulation of phospho-ERK and phospho-PKB are both inhibited by PI3K and CaMKII inhibitors, and show similar plateaus in the NMDA concentration-response curve. Furthermore, a high concentration of NMDA produced a similar decrease in phospho-PKB and phospho-ERK in the absence of block of synaptic activity with TTX. Activation of ERK and PKB by low (physiological) levels of NMDA receptor activation may function to promote neuronal survival, while inhibition of these prosurvival pathways in response to high (pathological) levels of NMDA receptor activation could function to allow efficient execution of apoptotic processes. Another striking similarity between NMDA modulation of ERK and PKB in these cortical cultures is the large increase in basal levels of phospho-ERK and phospho-PKB in the presence of a PP1/PP2A inhibitor, suggesting that PP1 and/or PP2A exerts a strong tonic inhibitory effect on ERK and PKB signaling.

Figure 7.

Schematic of NMDA receptor modulation of extracellular-signal regulated kinase (ERK) and protein kinase B (PKB) signaling. It is proposed that a common signaling complex couples NMDA receptor-mediated calcium influx to activation of both Ras-ERK and Ras-PKB. This signaling complex (gray box), which includes a pertussis toxin sensitive G-protein, phosphoinositide 3-kinase (PI3K), calcium/calmodulin-dependent kinase II (CaMKII), and others (see Chandler et al. 2001), may function as a common signaling pathway that bifurcates to ERK and PKB at the level of Ras. Alternatively, duplicate parallel pathways may function independently to activate Ras-ERK and Ras-PKB signaling. The black box represents the well characterized Raf-MEK cascade, and the white box represents the PI3K-PDK1/2 signaling process that likely mediates PKB activation. In addition to the NMDA activation pathways, there appears to be an opposing NMDA-receptor-mediated inhibitory pathway that functions to attenuate or shut-off ERK and PKB signaling. While the nature of this inhibitory pathway (stippled box) is not clear, it may be selectively coupled to activation of extrasynaptic NMDA receptors to inhibit survival pathways and promote cell death under certain pathological conditions (Hardingham et al. 2002).

While most studies have associated PI3K-PKB signaling with antagonism of apoptotic processes (Datta et al. 1999), dynamic regulation of PKB by NMDA receptors during synaptic activity could suggest a role for PKB in regulating localized synaptic processes. As noted above, the importance of compartmentalized modular signaling complexes at the postsynaptic density of dendritic spines is beginning to be appreciated in terms of regulating synaptic events such as plasticity of neuronal networks (Scannevin and Huganir 2000; Yuste et al. 2000). The mediators of NMDA stimulation of PKB are localized at the postsynaptic density of dendritic spines, and some are known to scaffold with the NMDA receptor signaling complex (e.g. PI3K) (Husi et al. 2000; Scannevin and Huganir 2000). Ras-PI3K signaling is thought to play an important role in regulating neurite extension/retraction (Kimura et al. 1994; Jackson et al. 1996; Kaplan and Miller 1997), and the PKB substrate GSK-3 has recently been shown to play a central role in neuritogenesis (Muñoz-Montaño et al. 1999). Neurite retraction induced by inhibition of PI3K with wortmannin is associated with an increase in GSK-3 activity, suggesting that activation of GSK-3 induces neurite retraction (Sanchez et al. 2001). Indeed, wortmannin-induced neurite retraction was reversed or exacerbated by manipulations that decreased or increased, respectively, GSK-3 activity. Since PKB phosphorylation of GSK-3 inhibits GSK-3 activity (Cross et al. 1995), NMDA receptor activity may regulate neuritogenesis through a PKB-GSK-3 signaling process. In addition to a role in regulating structural morphology, a number of PKB regulated substrates involve signaling processes thatfunction as neuromodulators of synaptic processes (i.e. NO and ERKs). PKB has been shown to mediate rapid IGF-1-induced potentiation of l-type voltage-sensitive calcium channels, possibly through direct phosphorylation of a channel subunit or an associated regulatory protein (Blair et al. 1999). Similarly, a preliminary report also suggests that GABAA receptor β subunits are phosphorylated by PKB (Wang et al. 2001). This phosphorylation appeared to be associated with increased receptor surface expression and amplitude of miniature inhibitory postsynaptic currents (mIPSCs) in cultured hippocampal neurons. Thus, an interesting possibility is that NMDA-PKB signaling could increase inhibitory GABAA receptor function in the postsynaptic domain. PKB has also been shown to transduce Ras-dependent endocytosis-inducing signals (Barbieri et al. 1998), and a number of studies have shown that endocytosis plays a critical role in regulating surface expression of neurotransmitter receptors (Carroll et al. 2001). Although speculative, NMDA receptor activation of PKB could play a role in activity-dependent surface expression of neurotransmitter receptors such as AMPA and GABAA receptors. Clearly, further work is required to identify substrates and define postsynaptic process regulated by PKB signaling.

After initial submission of the current manuscript, a report using primary striatal cultures appeared (Perkinton et al. 2002) that largely supports with our observations of NMDA receptor coupling to both ERK (Chandler et al. 2001) and PKB (present study). This includes a central role of PI3K, a requirement for a pertussis toxin sensitive Gi/o protein, and modulation by CaMKII in activation of both ERK and PKB. While these studies are defining the signaling processes involved in NMDA receptor activation of ERK and PKB, the mechanism of NMDA receptor inhibition of ERK and PKB are not clear. We speculate that the differential stimulatory and inhibitory effects may reflect compartmentalization of NMDA receptors and associated signaling complexes. In support of this is a recent report demonstrating the existence of functionally distinct synaptic and extrasynaptic NMDA receptor signaling complexes with directly opposing effects on neuronal survival (Hardingham et al. 2002). It was observed in primary hippocampal cultures that activation of synaptic NMDA receptors stimulate CREB function and survival promoting signals, whereas activation of extrasynaptic NMDA receptors inhibits CREB function and promotes cell-death signals. Thus, an intriguing possibility is that the NMDA receptor inhibition of ERK and PKB is mediated through activation of extrasynaptic NMDA receptors that shut-off survival signals and promote neuronal death.

Acknowledgement

This work was supported by National Institute on Alcohol Abuse and Alcoholism Grant AA10983.

Ancillary