Glucocorticoid Rapidly Enhances NMDA-Evoked Neurotoxicity by Attenuating the NR2A-Containing NMDA Receptor-Mediated ERK1/2 Activation

Abstract

Glucocorticoid (GC) has been shown to affect the neuronal survival/death through a genomic mechanism, but whether or not it does through a nongenomic mechanism is unknown. Using a previously identified GR-deficient primary hippocampal neuron culture, we show here that a 15-min coexposure of N-methyl-d-aspartate (NMDA) with corticosterone at a stress-induced level significantly enhances neuronal death compared to NMDA alone. This enhancing effect of GC can be mimicked by the BSA-conjugated corticosterone, which is plasma membrane impermeable and cannot be blocked by RU38486 spironolactone. Furthermore, using a calcium-imaging technique, we found that B could increase both the percentage of neurons showing a significant increment of intracellular free calcium ([Ca²⁺]_i) due to NMDA stimulation and the amplitude of [Ca²⁺]_i increment in the individual responsive cells. Interestingly, this boosting effect of GC on [Ca²⁺]_i increment could be blocked by the NMDA receptor subunit 2A (NR2A)-specific antagonist [(R)-[(S)-1-(4-bromo-phenyl)-ethylamino]-(2,3-dioxo-1,2,3,4-tetrahydro-quinoxalin-5-yl)-methyl]-phosphonic acid (NVP-AAM077) but not by the NMDA receptor subunit 2B (NR2B)-specific antagonist Ro25-6981. Moreover, we also found that GC can dramatically attenuate the NMDA-induced activation of ERK1/2 without affecting that of p38; and that the NMDA-induced ERK1/2 activation and its attenuation by GC both can be occluded by the NVP-AAM077 but not by Ro25-6981. Consistently, the enhancing effect of GC on NMDA neurotoxicity can also be blocked by NVP-AAM077 and the ERK1/2 inhibitor PD98059 but not by Ro25-6981 and p38 inhibitor SB203580. Indeed, the NMDA neurotoxicity itself can be blocked by Ro25-6981 or SB203580, whereas it is increased by NVP-AAM077 and PD98059. Therefore, it is probable that NMDA triggers a prodeath signaling through the NR2B-p38 MAPK pathway, and a prosurvival signaling through the NR2A-ERK1/2 MAPK pathway, whereas the latter was negatively regulated by rapid GC action. Taken together, the present data suggest a nongenomic action by GC that enhances NMDA neurotoxicity through facilitating [Ca²⁺]_i increment and attenuating the NR2A-ERK1/2-mediated neuroprotective signaling, implicating a novel pathway underlying the regulatory effect of GC on neuronal survival/death.

Glucocorticoid rapidly enhances NMDA-evoked neurotoxicity by facilitating NMDA-induced [Ca²⁺]_i increment and attenuating the NR2A-ERK1/2 dependent neuroprotective signaling in hippocampal neurons.

Glucocorticoid (GC) exerts a wide variety of effects on the body through the ubiquitously distributed glucocorticoid receptor (GR). It is released in high amounts during stress and under many other pathological conditions. It enters the brain compartment and intensively affects the structure and function of the central nervous system (CNS), particularly the hippocampus, which is one of the most vulnerable and GR-rich regions in the CNS (1, 2). According to the classical genomic mechanism of GC action, the binding of GC to its receptor triggers the activation and translocation of the GR complexes from cytoplasm to the nuclear, where they act as transcription factor to activate or suppress the expression of a variety of genes, either through DNA binding of receptor homodimers or through protein-protein interactions with other transcription factors and cofactors (3).

In the recent decades, it has been demonstrated that GC also acts through a membrane-initiated nongenomic mechanism (4, 5, 6). Nongenomic action of GC has a rapid onset time typically within minutes or even seconds after stimulation in contrast to the genomic whereas onset takes hours (5). It can be classical GR and transcription/translation independent, which might be mediated through some membrane-bound GR (mGR) that is yet to be identified (4, 5). In addition, this also involves the activation of many intracellular signal transduction pathways (7). Although rapid nongenomic action of GC has been extensively reported in the CNS (5), the biological relevance of this action, i.e. how and to what extent this action can affect the function of CNS still remain to be investigated.

Previous studies in our laboratory revealed that GC could rapidly regulate the intracellular calcium in PC12 cells (8, 9) and activity of MAPK in hippocampal neurons (10, 11). Because these signal molecules are important to the control of cell survival/death, a role of GC in cell survival/death regulation is greatly expected. In fact, it is has been shown that long time exposure to high-level GC (like chronic stress) can be dangerous to hippocampal neurons, especially when associated with excitotoxic challenges caused by overactivation of the N-methyl-d-aspartate (NMDA) receptors (NMDARs) such as in ischemia stroke (12). Generally speaking, these effects are supposed to be mediated by the GR through the transcriptional-dependent genomic mechanism. We wondered whether the nongenomic actions of GC are also involved in the regulation of the neuronal survival/death, and hypothesized that GC would modulate the NMDA-induced neuroexcitotoxicity rapidly. Using a previously characterized GR-deficient hippocampal neuron culture model (11), indeed, we have identified that GC rapidly enhances NMDA-evoked neurotoxicity.

Results

NMDA evokes hippocampal neuron toxicity in a concentration-dependent manner

Overactivation of the NMDA receptors (NMDARs) is known to have neuroexcitotoxicity that is implicated in many acute neurological insults and chronic neurodegenerative diseases. In vitro studies have showed that the extent of NMDA toxicity might vary between different experiments depending on the cell culture models used and the protocol of drug administration. In our experiments, by bath application of NMDA as described in Material and Methods, we demonstrated that the toxicity was concentration dependent with the half-lethal dose at about 50 μm and maximized at about 200 μm as accessed by lactate dehydrogenase (LDH) release assay and Hoechst and microtubule-associated protein-2 (MAP2) staining (Fig. 1). Thus, the 50 μm concentration of NMDA was used in the toxicity experiments hereafter unless otherwise described.

Fig. 1.

NMDA induces concentration-dependent neuronal damage. Neurons were treated with indicated concentrations of NMDA in modified Locke’s solution for 15 min at room temperature and returned to culture medium. LDH release, Hoechst nuclear staining, and MAP2 immunostaining were carried out 24 h after. Representative image of immunostaining and nuclear staining (panel A); NMDA-induced neuronal release of LDH (panel B); loss of MAP2 immunoreactivation (panel C); and condensation of nuclei (panel D). **, P < 0.01 vs. sham treatment. Bar, 50 μm.

GC rapidly enhances the NMDA neurotoxicity independent of its entering into the cytoplasm

To test whether corticosterone (B) has a rapid regulatory effect on the survival/death of hippocampal neurons upon NMDA challenge, neurons were treated with NMDA in the presence or absence of different concentrations of B for 15 min. It was found that B could significantly enhance the NMDA-evoked LDH release and nuclei condensation at high concentrations and maximized the effect at 100 nm (Fig. 2), which approximate the cerebrospinal fluid free B level in stressed rats (13), but these concentrations of B alone did not cause any changes in the neuronal viability (14). To determine whether this effect of B resulted from an action in the cytoplasm or on the cell membrane, we also tested the effect of BSA-conjugated corticosterone (B-BSA), which is supposed to be cell membrane impermeable. In fact, it is showed that B-BSA was as effective as B in significantly increasing the LDH release, nuclear condensation, and propidium iodide (PI) uptake in the neurons (Fig. 3), suggesting that the GC enhancement of the NMDA neurotoxicity happens via rapid action on the cell membrane without the movement of the hormone into the cytoplasm. Moreover, this effect of GC was further confirmed by a flow cytometry study, which showed dramatic increases in late apoptotic and necrotic cells when neurons were exposed to NMDA in the presence of B or B-BSA (Fig. 4). Furthermore, the GC enhancement effects could also be seen when neurons were pretreated with B or B-BSA for 15 min before NMDA challenge (Fig. 5). Interestingly, if pretreatment were followed with a washout before NMDA stimulation, there would be no significant differences in neuronal death between the pretreated and unpretreated groups (Fig. 5). This result indicates that the enhancement effect of GC on the NMDA neurotoxicity might require the temporal coexistence of NMDA and GC.

Fig. 2.

GC at high concentrations rapidly enhances NMDA-induced neurotoxicity. Hippocampal neurons (at 10 DIV) were treated with or without NMDA in the absence or presence of indicated concentrations of B for 15 min in modified Locke’s solution and returned to culture medium. Measurements were conducted 24 h later. Representative morphological changes of neurons were shown by differential interference contrast images (NMDA, 50 μm; B, 100 nm). Scale bar, 50 μm. A, LDH release assay; B, nuclear condensation. ###, P < 0.001 vs. control; **, P < 0.01 vs. NMDA alone; ***, P < 0.001 vs. NMDA alone. Bar, 50 μm. Con, Control.

Fig. 3.

GC rapidly enhances the NMDA neurotoxicity independent of its entering into the cell cytoplasm. Hippocampal neurons (at 10 DIV) were treated with or without NMDA (50 μm) in the absence or presence of B or B-BSA (100 nm) for 15 min in modified Locke’s solution and returned to culture medium. LDH release assay, PI uptake, and Hoechst staining was conducted 24 h later. A, Representative fluorescence images of Hoechst (a–d) and PI (e–h) staining showing nuclear condensation or PI positive (arrow heads) under the indicated treatments. Effects of B or B-BSA on the NMDA-induced LDH release (B), nuclei condensation (C), and PI uptake (D) was shown. ###, P < 0.001 vs. sham treatment; ***, P < 0.001 vs. NMDA alone. Bar, 20 μm.

Fig. 4.

Analysis of rapid effect of GC on the NMDA neurotoxicity by flow cytometry. Hippocampal neurons (at 10 DIV) were treated with or without NMDA (50 μm) in the absence or presence of B or B-BSA (100 nm) as indicated for 15 min in modified Locke’s solution and returned to culture medium. Neurons were harvested 12 h later and subjected to flow cytometry analysis with a combination stain of Annexin V-fluorescein isothiocyanate and PI. Representative results were shown from each indicated treatment group. Data are means ± sem from three independent experiments. **, P < 0.01 vs. NMDA alone. Apo, Apoptotic; FITC, fluorescein isothiocyanate; Nec, necrotic.

Fig. 5.

GC’s rapid enhancement of the NMDA neurotoxicity requires the temporal coexistence of NMDA. The culture medium of hippocampal neurons (at 10 DIV) was collected, and replaced Locke’s solution, B, or B-BSA (100 nm) was added. After 15 min, the cultures were washed out with Locke’s solution or maintained intact, after which they were challenged with NMDA (50 μm) in the continuous presence (continuate) or subsequent absence (washout) of B or B-BSA as indicated for 15 min in modified Locke’s solution at room temperature and returned to the original culture medium. LDH release measurement and Hoechst staining were conducted 24 h after: A, Typical images showing the nuclear condensation (arrows) under various treatment conditions as indicated; B, LDH release assay; and C, Hoechst staining. ###, P < 0.001 vs. control; ***, P < 0.001 vs. NMDA alone. Note there were no significant differences between NMDA and NMDA accompanied by B or B-BSA in the washout groups. Bar, 20 μm. Con, Control.

GC’s rapid enhancement effect on NMDA neurotoxicity is classic GR and mineralcorticoid receptor (MR) independent

The present embryonic hippocampal neuron culture has been previously shown to be GR deficient (11, 14), suggesting that the enhancement effect of GC on NMDA neurotoxicity observed here was GR independent. To further rule out the involvement of the classic B receptors, we also examined the effects of the GR antagonist RU38486 and the MR antagonist spironolactone. Our results demonstrated that neither of the two antagonists could block the enhancement effect of B on NMDA-induced LDH release and nuclei condensation (Fig. 6). Thus, GC enhances NMDA neurotoxicity through a classic GR- and MR-independent, nongenomic mechanism.

Fig. 6.

RU36486 and spironolactone cannot block the enhancement effect of B on NMDA neurotoxicity. Hippocampal neurons (at 10 DIV) were pretreated with or without RU36486 (RU, 1 μm) or spironolactone (Spiro, 1 μm) for 20 min and were then treated with or without NMDA (50 μm) in the absence or presence of B (100 nm) for 15 min in modified Locke’s solution and returned to culture medium. LDH release measurement and Hoechst staining were conducted 24 h after: A, Typical images showing the nuclear condensation (arrows) under various treatment conditions as indicated; B, LDH release assay; and C, Hoechst staining. **, P < 0.01 vs. NMDA alone; ***, P < 0.001 vs. sham treatment. Bar, 20 μm.

GC rapidly enhances the NMDA-induced [Ca²⁺]_i increment in hippocampal neuron in an NR2A-containing NMDAR-dependent manner

Overloading of cytoplasmic calcium by extracellular calcium influx is known to play a critical role in the mediation of NMDA neurotoxicity. Therefore, we then checked the effect of GC on the NMDA-induced [Ca²⁺]_i increment in hippocampal neurons. We found that B (100 nm) alone did not trigger any [Ca²⁺]_i increment, but the coexistence of B with NMDA dramatically enhanced the Ca²⁺ response compared with NMDA (100 μm) alone, and this enhancement was 2-fold (Fig. 7). First, it increased the percentage of neurons that showed significant [Ca²⁺]_i increase (Δratio ≥ 0.1) to NMDA stimulation (Fig. 7B). Second, it also boosted the amplitude of [Ca²⁺]_i increment in individual responsive cells (Fig. 7C). Moreover, the enhancement effects of B were not blocked by GR antagonist RU38486 (Fig. 7, D and F), which is inconsistent with the result of the neurotoxicity assay. Most interestingly, we found that the boosting effects of B on the amplitude of [Ca²⁺]_i increment in individual responsive cells could be blocked by the noncompetitive NMDAR antagonist MK-801 and the NR2A-specific antagonist NVP-AAM077, but not by the NR2B-specific antagonist Ro25-6981 (Fig. 7E), indicating that GC enhances the NMDA-induced [Ca²⁺]_i increment in a NR2A-containing NMDAR-dependent manner.

Fig. 7.

GC rapidly enhances the NMDA-induced [Ca²⁺]_i increment in hippocampal neurons in a NR2A-containing NMDAR-dependent manner. Hippocampal neurons (at 10 DIV) grown on coverslips were subjected to [Ca²⁺]_i imaging sassy as described in Materials and Methods. Neurons were preperfused with or without B (100 nm) or/and RU38486 (1 μm) for 2 min and then perfused with NMDA (100 μm) in the absence or presence of B and/or RU38486. A, Representative pseudocolored images showed the ratio before (Bef) and after (Aft) NMDA stimulation. Typical responsive neurons were shown (arrows). The corresponding differential interference contrast images and fluorescence images of fura-2 loading were also shown. The ratio of fura-2 fluorescence is indicated with a color bar from blue (low [Ca²⁺]_i) to red/white (high [Ca²⁺]_i). B, The percent of neurons that showed significant [Ca²⁺]_i increment (Δratio > 0.1, named as responsive neuron) in the absence or presence of B. C, The average of the amplitudes of [Ca²⁺]_i increment (Δratio) of total neurons and responsive neurons in the absence or presence of B. D, Effect RU38468 (RU) on the GC-enhanced [Ca²⁺]_i increment. E, Effects of NMDAR antagonists on the GC-enhanced [Ca²⁺]_i increment: neurons were preperfused with or without MK-801 (0.3 μm), Ro-256981 (Ro25, 0.5 μm), or NVP-AMM077 (NVP, 0.5 μm) for 2 min and then perfused with NMDA (100 μm) in the absence or presence of those antagonists. Note there were no significant differences between NMDA and NMDA+B in the presence of NVP. F, Representative curves showing the change of ratio value with time under indicated treatments. Data are mean ± sem from three independent experiments counting about 100 total neurons per treatment group in each experiment. **, P < 0.01 vs. NMDA. Bar, 50 μm. DIC, Differential interference contrast.

GC rapidly attenuates the NMDA-evoked activation of ERK1/2 without affecting that of p38

The p38 MAPK and ERK1/2 are intracellular signals that are highly involved in the control of neuronal survival/death (15, 16). They are sensitive to glutamate and can both be regulated in a bidirectional manner by NMDARs (17, 18, 19). In our system, it was found that NMDA stimulation (15 min) triggered a bell-shaped concentration-dependent activation of both p38 (Fig. 8A) and ERK1/2 (Fig. 8B), which peaked at 10 μm and 20 μm, respectively. At 50 μm, which was the concentration we have used in the neurotoxicity assay, NMDA also produced significant activations of p38 and ERK1/2. We wondered whether the NMDA-induced MAPK activation could be rapidly regulated by GC. Interestingly, B and B-BSA themselves can trigger the activation of p38 and ERK1/2; however, coexposure of B or B-BAS with NMDA both resulted in a dramatically decreased ERK1/2 activity (Fig. 8D) but an unchanged p38 activity (Fig. 8C) in comparison to NMDA treatment alone, suggesting that GC can rapidly and selectively attenuate the NMDA-induced activation of ERK1/2 without affecting that of p38 MAPK. Thus, GC can distinctly regulate the NMDA-evoked p38 and ERK1/2 MAPK activation.

Fig. 8.

GC rapidly inhibits the NMDA-induced the activation of ERK1/2 without affecting that of p38. Hippocampal neurons (at DIV 10) were starved for 6 h in DMEM without any supplement and were then treated with or without NMDA and/or GCs as indicated drugs for 15 min and harvested for immunobloting analysis of ERK1/2 or p38 activation. Concentration-dependent activation of p38 (A) and ERK1/2 (B) by NMDA, and the effect of B or B-BSA (100 nm) on NMDA (50 μm)-induced p38 (C) and ERK1/2 (D) activation were analyzed. Representative blots demonstrating the phosphorylation/activation of p38 and ERK1/2 were shown. Results were mean ± sd from three independent experiments. **, P < 0.01 vs. control; ##, P < 0.01 vs. NMDA alone. Con, Control.

Attenuation of NMDA-evoked ERK1/2 activation by GC is dependent on NR2A-containing NMDAR

It has been suggested that NR2A and NR2B subunits of NMDAR are coupled to different MAPK pathways and might play distinct roles in ERK1/2 activation or inactivation according to the time course and concentration of NMDA stimulation as well as the maturation stage of neurons (20, 21). In the current experiments, embryonic d 18 (E18) hippocampal neurons were stimulated by 50 μm NMDA for 15 min at 10 d in vitro (10 DIV), and different NMDAR antagonists were used to test the subunit dependence of the attenuation effect of GC on ERK1/2 activation. It was found that both MK-801 and NVP-AAM077 blocked the NMDA-evoked ERK1/2 activation (Fig. 9A) and also occluded the GC-induced attenuation (Fig. 9B) of it, whereas they were not affected by Ro25-6981. Thus, both the NMDA-evoked ERK1/2 activation and its attenuation by GC are NR2A-containing NMDARs dependent, suggesting that NR2A is coupled to an ERK1/2 activation pathway that could be negatively regulated by rapid GC action.

Fig. 9.

NMDA-induced ERK1/2 activation and its inhibition by GC are both dependent on NR2A-containing NMDARs. Hippocampal neurons (at 10 DIV) were starved for 6 h in DMEM without any supplement and were then subjected to indicating treatments for 15 min and harvested for immunobloting analysis of ERK1/2 activation. A, Effects of the noncompetitive NMDAR antagonist MK-801 (0.3 μm), NR2B-specific antagonist Ro-256981 (Ro25, 0.5 μm), and NR2A-specific antagonist NVP-AMM077 (NVP, 0.5 μm) on MNDA (50 μm)-induced ERK1/2 activation. Representative blot demonstrating the phosphorylation/activation of ERK1/2 was shown. B, Effects of the noncompetitive NMDAR antagonist MK-801 (0.3 μm), NR2B-specific antagonist Ro-256981 (Ro25, 0.5 μm), and NR2A-specific antagonist NVP-AMM077 (0.5 μm) on B’s inhibition of MNDA (50 μm)-induced ERK1/2 activation. Representative blot demonstrating the phosphorylation/activation of ERK1/2 was shown. **, P < 0.01 vs. control; ##, P < 0.01 vs. NMDA+PBS; §§, P < 0.01 vs. NMDA+Ro25. Note there were no significant differences between NMDA and NMDA+B in the presence of NVP. Con, Control.

GC’s enhancement effect on NMDA neurotoxicity is dependent on NR2A-, but not NR2B-containing NMDAR

Because our results showed that the rapid enhancing and attenuating effects of GC on NMDA-induced [Ca²⁺]_i increment and ERK1/2 activation were both NR2A-containing NMDAR dependent, we wondered whether the GC enhancement effect on NMDA neurotoxicity is also NMDAR subunit selective. Based on LDH release assay and Hoechst staining, it was demonstrated that the enhancement effect of GC on NMDA neurotoxicity could be blocked by MK-801 and NVP-AAM077 but not by Ro25-6981 (Fig. 10, A, D, and E), suggesting that GCs act through a NR2A-containing NMDAR-dependent pathway to increase the hippocampal neuronal vulnerability upon NMDA challenge. On the other hand, the NMDA neurotoxicity itself was abolished by MK-801 and Ro25-6981 but enhanced by NVP-AAM077 (Fig. 10, A–C), which indicated that the NR2B-containing NMDARs are attributable to the NMDA neurotoxicity, whereas the NR2A-containing NMDARs instead favor a neuroprotection effect.

Fig. 10.

GC’s rapid enhancement effect on NMDA neurotoxicity is dependent on NR2A-containing but not NR2B-containing NMDARs. Hippocampal neurons (at 10 DIV) were pretreated with noncompetitive NMDAR antagonist MK-801 (0.3 μm), NR2B-specific antagonist Ro-256981 (Ro25, 0.5 μm), or NR2A-specific antagonist NVP-AMM077 (NVP, 0.5 μm) or sham for 20 min. Neurons were then incubated with NMDA (50 μm) in the absence or presence of B (100 nm) for 15 min in modified Locke’s solution and returned to culture medium. LDH release measurement and Hoechst staining were conducted 24 h after: A, Typical images showing the nuclear condensation (arrow) under various treatment conditions as indicated. B and D, LDH release assay; and C and E, Hoechst staining. B and C, ***, P < 0.001 vs. sham treatment; ###, P < 0.001 vs. NMDA alone. D and E, **, P < 0.01 vs. NMDA. Note there were no significant differences between NMDA and NMDA+B in the presence of NVP. Bar, 20 μm.

GC’s enhancement effect on NMDA neurotoxicity is ERK1/2, but not p38 MAPK, pathway dependent

It is known that ERK1/2 and p38 MAPK are critical signal molecules in the down stream of the NMDARs and contribute to the various effects of glutamate in the CNS including neuronal survival/death. Because GC has shown differential regulations on the NMDA-evoked ERK1/2 and p38 activation, we further investigated the roles of these two MAPKs in the enhancement effects of GC on NMDA neurotoxicity. Interestingly, the results showed that the enhancement effect of GC on NMDA-induced LDH release and nuclei condensation in hippocampal neurons could be completely abolished by the ERK1/2 inhibitor PD98059 but not affected by the p38 MAPK inhibitor SB203580 (Fig. 11, A, D, and E), suggesting an ERK1/2-dependent mechanism underlying the GC effect. Furthermore, it was also revealed that the NMDA neurotoxicity itself could be blocked by SB203580 but increased by PD98059 (Fig. 11, A–C), indicating that the activation of p38 is responsible for the NMDA-induced neurotoxicity, whereas the activation of ERK1/2 by NMDA has instead served as a prosurvival signaling. Thus, GC rapidly enhances the NMDA neurotoxicity via inhibition of this ERK1/2-dependent neuroprotective signal pathway.

Fig. 11.

GC’s rapid enhancement effect on NMDA neurotoxicity is dependent on ERK1/2 but not p38 MAPK pathway. Hippocampal neurons (at DIV 10) were pretreated with p38 MAPK-specific inhibitor SB203580 (10 μm), ERK1/2 MAPK-specific inhibitor PD98059 (10 μm), or dimethylsulfoxide for 20 min. Neurons were then incubated with NMDA (50 μm) in the absence or presence of B (100 nm) for 15 min in modified Locke’s solution and returned to culture medium. LDH release measurement and Hoechst staining were conducted 24 h after: A, Typical images showing the nuclear condensation (arrow) under various treatment conditions as indicated; B and D, LDH release; C and E, Hoechst staining. ***, P < 0.001 vs. sham treatment; **, P < 0.01 vs. NMDA. Note there were no significant differences between NMDA and NMDA+B in the presence of PD98059. Bar, 20 μm. DMSO, Dimethylsulfoxide.

Discussion

The principal findings of the present study are 6-fold: 1) GC rapidly enhanced the NMDA-induced neurotoxicity in a GR- and MR-independent, nongenomic manner; 2) The rapid GC enhancement of the NMDA neurotoxicity required the coexistence of NMDA and GC; 3) GC rapidly augmented the NMDA-induced hippocampal neuron [Ca²⁺]_i increment in a NR2A-containing NMDAR-dependent manner; 4) GC rapidly attenuated the NMDA-induced ERK1/2 MAPK activation without affecting the p38 MAPK activation; 5) Although the NR2B-containing NMDAR and p38 MAPK were responsible for the NMDA neurotoxicity, the NR2A-containing NMDAR-mediated ERK1/2 MAPK activation served as a neuroprotective signaling; and 6) GC’s rapid action selectively attenuated the NR2A-mediated ERK1/2 activation and enhanced the NMDA neurotoxicity in a NR2A- and ERK1/2-dependent but NR2B- and p38-independent manner.

GC enhances the NMDA neurotoxicity in hippocampal neurons via a rapid nongemomic mechanism

GC exerts a wide variety of effects on the CNS including the regulation of neuronal cell survival/death. Although the neurodegenerative and neuroprotective roles of GC have both been reported (22), it is generally accepted that prolonged exposure of high levels of GC (e.g. in chronic stress or major depression state) can be a dangerous factor, facilitating neuronal damage induced by various insults such as ischemic stroke, oxidative stress, or direct NMDA and kainic acid challenges, particularly in respect to the hippocampus (12). Those effects were supposed to be mediated by activation of GR and consequential inhibition of glucose uptake and energy metabolism impairment in neuronal cells (12), which might drive the neurons into a more vulnerable status to these insults. Moreover, the down-regulation of glutamate transporter expression (23) and up-regulation of NMDAR (24) expression by GC might also be involved. In this way, the classical genomic mechanism is well implicated in the regulation of neuronal cell survival/death by GC.

In addition to the classical genomic mechanism, GC also acts through the rapid nongenomic mechanism (5). However, it is not clear whether the rapid nongenomic mechanism also contributes to the regulation of neuronal cell survival/death by GC. In the present study, after a short time exposure (15 min) of NMDA with or without GC, we demonstrated that NMDA can evoke an increasingly intensive neuronal death in the presence of GC, i.e. GC rapidly enhanced the NMDA neurotoxicity in an effective time frame as short as 15 min. This rapid enhancement effect of GC could be mimicked by the membrane-impermeable B-BSA, as was demonstrated by the significant increase in LDH release, nuclei condensation percentage, and propidium iodide (PI) uptake. The effect was also confirmed by a flow cytometry study that showed dramatic increase in the percentage of apoptotic and/or necrotic neurons when NMDA treatment was accompanied by B or B-BSA. Moreover, this effect could not be blocked by the classical GR antagonist RU38468 and MR antagonist spironolactone. Thus, as in the GR expression-deficient neuronal model (11, 14), our results clearly demonstrated that the enhancement effect of GC on NMDA neurotoxicity is unequivocally nongenomic. Similarly, Mulholland et al. (25) have reported that a 24-h B exposure exacerbated excitotoxic insult in rat hippocampal slice cultures through a pathway independent of GR activation or protein synthesis, which also suggests a nongenomic action of GC.

Rapid enhancing effect of GC on NMDA-induced hippocampal neuron [Ca²⁺]_i increment

NMDARs are cationic channels gated by glutamate, the main excitatory neurotransmitter in mammalian CNS. They are permeable to Na⁺, which contributes to postsynaptic depolarization, and Ca²⁺, which generates intracellular Ca²⁺ transients. Ca²⁺ influx is the most direct response upon NMDAR activation, and it is through this Ca²⁺ influx that NMDARs exert most of their physiological or pathological effects such as synaptic plasticity and excitotoxicity (26). Although modulation effects of GC on Ca²⁺ currents have been reported via GR-dependent genomic mechanism in hippocampal neurons (27, 28, 29), and most were enhancement effects that seem to only involve the voltage-gated Ca²⁺channels (27, 28, 29). It is interesting to note that most of the previously reported nongenomic modulation of Ca²⁺ currents by GC were inhibitory effects and involved both voltage-gated and ligand-gated Ca²⁺ channels (8, 9, 30). Our previous studies have demonstrated a rapid inhibitory effect of GC on nicotine-, high potassium-, and bradykinin-induced Ca²⁺ influx in PC12 cells through GR-independent nongenomic mechanism (8, 9, 30). Here, we showed that GC has a rapid enhancing effect on the NMDA-induced [Ca²⁺]_i elevation in hippocampal neurons, and this enhancement was 2-fold. First, it increased the percentage of responsive neurons (Δratio ≥ 0.1) to NMDA stimulation. Second, it magnified the amplitude of [Ca²⁺]_i increment in individual responsive cells. Moreover, the effect of GC could not be blocked by RU38486. Thus, in the current GR expression deficient neuronal model (11, 14), our result pointed to a rapid enhancing effect of GC on NMDA-induced [Ca²⁺]_i increment via a nongenomic mechanism. Interestingly, this effect of GC showed an NR2A-containing NMDARs specificity because it could be blocked by NVP-AMM077 but not by Ro25-6981. Similar to our result, Takahashi et al. (31) have also reported an acute effect of GC as prolonging the duration of [Ca²⁺]_i elevation induced by NMDA in hippocampal neurons without eliciting an increase in the amplitude of [Ca²⁺]_i elevation. In their case, hippocampal neurons were prepared from postnatal rats approximately 3- to 5 d of age and cultured for 7–10 d before experiments. The response of those neurons to NMDA stimulation was characterized by a transient elevation in [Ca²⁺]_i followed by a rapid decay to the basal level within about 30–70 sec (32). In the present study, hippocampal neurons were prepared from E18 rat and used at 10 DIV, and they showed a continuous elevation of [Ca²⁺]_i in response to NMDA without any decay within the time frame observed (4–5 min). Interestingly, Takahashi et al. (33) also recorded a similar continuous elevation of [Ca²⁺]_i when they used E19–20 hippocampal neuron preparations, suggesting developmental changes of Ca²⁺ response in the hippocampal neurons to NMDA stimulation.

Opposing effects of NR2A and NR2B on hippocampal neuronal survival/death and the specific attenuation of NR2A-mediated prosurvival signaling by rapid GC action

NMDARs are tetrameric complexes composed of two NR1 subunits and at least one type of NR2 subunits with predominantly NR2A or NR2B subunits in the adult rat hippocampus (32, 33). Although pathological activation of NMDARs is a major cause of neuronal death after acute excitotoxic trauma such as brain ischemia, hypoxia, and mechanical trauma, physiological levels of synaptic NMDAR activity are essential for neuronal survival (26, 34). The determinants of the ultimate outcomes of a specific episode of NMDAR activation, whether neuroprotective or excitotoxic, may be multifold, including the stimulus intensity and the location of the activated receptors, as well as the maturation stage of neurons (34, 35, 36). For example, it was shown that activation of synaptic NMDARs leads to the promotion of neuronal survival, whereas activation of extrasynaptic NMDARs results in excitotoxic cell death (37). In the present study, bath application of NMDA (50 μm) for 15 min induced a significant excitotoxicity in hippocampal neuron cultures of 10 DIV. This drug administration protocol could lead to the activation of both synaptic and extrasynaptic NMDARs. Recently, the functional distinction among NMDAR subunits has also been suggested (36, 38, 39). To investigate the differential roles of NR2A and NR2B in the neurotoxicity, we took advantage of the subunit-specific NMDAR antagonist NVP-AAM077, which preferentially inhibits NR2A-containing receptors at concentration of 0.4–1 μm (40), and Ro25-6981, which specifically blocks NR1/NR2B receptors (41). It was found that blockage of NR1/NR2B receptors abolished the NMDA-induced neurotoxicity, whereas blockage of the NR2A-containing receptors increased the neurotoxicity, suggesting that overall activation of both synaptic and extrasynaptic NR2B receptors was excitotoxic whereas overall activation of those NR2A receptors was neuroprotective, consistent with the results in hippocampal slice cultures (36) and cortical neuron cultures (38). Our results indicate that NR2A- and NR2B-containing receptors have opposing roles with the former leading to neuronal survival whereas the latter leads to neuronal death, and support that the receptor subunit composition might be an additional determinant for the outcome of a specific episode of NMDAR activation.

Interestingly, it was shown that the result of NMDARs activation could further be influenced by GC through rapid nongenomic mechanism because a 15-min coexposure of B or B-BSA with NMDA significantly enhanced neuronal death compared with NMDA alone. The GC enhancement effect could be abolished by NVP-AMM077 but not affected by Ro25-6981, which indicated a selective action on the NR2A-containing NMDARs by GC. Moreover, it was demonstrated that the NMDA neurotoxicity itself could be blocked by inhibition of the p38 MAPK but increased by inhibition of ERK1/2 MAPK, suggesting that when the activation of NMDARs triggered a p38-dependent excitotoxic signaling, it also triggered an ERK1/2-dependent neuroprotective signaling. Furthermore, consistent with previous studies that NMDARs are coupled to distinct intracellular signaling pathways (20, 21), we found that the NMDA-induced ERK1/2 activation could be selectively occluded by NVP-AMM077 but not by Ro25-6981, which implicated preference to the ERK1/2 pathway by the NR2A-containing NMDARs to promote neuronal survival. Thus, there was a possible selective action on the NR2A-mediated ERK1/2 activation by GC. In fact, we found that GC did selectively attenuate the NMDA-induced ERK1/2 activation without affecting p38 activation, and this effect was also sensitive to NR2A-specific antagonist but insensitive to NR2B-specific antagonist. Therefore, our results suggested that GC enhanced the NMDA neurotoxicity through specific attenuation of the NR2A-ERK1/2 prosurvival signaling. Although the involvement of NMDARs has been suggested in many of the GC’s effects in the CNS (42), such a specific and rapid modulation of the NMDAR subunit function by GC, as shown in the present study, has not been reported before. This rapid action of GC might serve as a novel pathway through which GC can exert its many effects in the CNS, e.g. regulating the neuronal survival and death.

Probable site of the rapid GC action to enhance the NMDA neurotoxicity

We showed that the attenuation of NR2A-mediated ERK1/2 activation by GC has contributed to its rapid enhancement effect on NMDA neurotoxicity, but the exact site of the action and the more detailed mechanism remain unknown. Because the rapid enhancement effect on NMDA neurotoxicity and the attenuation of NMDA-induced ERK1/2 activation could also be seen with the cell membrane-impermeable GC, i.e. B-BSA, we believe that this effect of GC is likely to be mediated through a membrane-associated GR (mGR) as we and others have previously suggested (4, 43, 44). In fact, we have identified the specific glucocorticoid membrane-binding site on a synaptic plasma membrane preparation from rat. The ligand-binding properties and physicochemical characteristics of the glucocorticoid membrane-binding site were shown to fulfill the basic criteria for a receptor, which indicated that it might constitute a novel type of mGR that was significantly different from the classical cytosolic GR (45). Similar results have been reported later in neuronal membranes from amphibian brain (46) and more recently in pituitary cell membrane (47). Although the molecular identity of mGR is currently unknown, it has been shown that mGR was coupled to the activation of many intracellular signal pathways (7) including the p38 and ERK1/2 MAPK pathways (10, 11) and also confirmed in the present study. It is most interesting to note that although GC or NMDA alone can evoke the activation of ERK1/2 and p38, their copresence would result in a dramatic decrease in ERK1/2 activity selectively. It has been suggested that the mGR-evoked ERK1/2 activation is mediated by a G protein-PKC pathway (10), obviously distinct from the intracellular pathway that mediates the NMDAR-evoked ERK1/2 activation (17). Therefore, one can expect that these two pathways might have cross-talked/interacted before they converged on the ERK1/2 and thus led to the above result.

It is also interesting to note that the enhancement effect of GC on the NMDA neurotoxicity disappeared when GC was pre-added and followed by a washout before the NMDA challenge, which indicated that the rapid GC effect might require the temporal coexistence of both NMDA and GC and that it might be dependent on NMDAR activation. Because the effect of GC was selectively sensitive to the NR2A-specific antagonist, it might be possible that the mGR interacts with the NR2A receptor subunit on the cell membrane to modulate the NR2A-ERK1/2 signaling. Moreover, the modulation of neurotransmitter receptors by nongenomic GC action has also been reported previously (5). Thus, a fast, activity-dependent, and direct modulation of the NMDAR by GC as well cannot be excluded.

In summary, our data demonstrate a rapid nongenomic action of GC that enhances NMDA neurotoxicity in cultured hippocampal neurons. The enhancement effect was suggested to be mediated through facilitation of NMDA-induced [Ca²⁺]_i increment and attenuation of the NR2A-ERK1/2-dependent neuroprotective signaling by GC. Our results reveal a new pathway underlying the regulatory effect of GC on neuronal survival/death.

Materials and Methods

Materials

B, B-BSA, Hoechst33342, NMDA, PI, poly-[l]-lysine, and fluorescein isothiocyanate-conjugated antirabbit IgG antibody were purchased from Sigma (St. Louis, MO), MAP2 monoclonal antibody was from NeoMarkers (Fremont, CA), Antibodies against ERK1/2, phospho-ERK1/2 (Thr202/Tyr204), p38, and phospho-p38 (Thr180/Tyr182) were purchased from Cell Signaling Technology, Inc. (Beverly, MA), β-actin monoclonal antibodies and horseradish peroxidase-conjugated secondary antibodies from Kangcheng Biotechonology, Inc. (Shanghai, China). (+)-5-methyl-10,11-dihydro-5H-dibenzo [a,d] cyclohepten-5, 10-imine maleate (MK-801), R-(R, S)-α-(4-hydroxyphenyl)-β-methyl-4-(phenylmethyl)-1-piperidine propranol (Ro25-6981) were from Tocris (Cookson, Bristol, UK). [(R)-[(S)-1-(4-bromo-phenyl)-ethylamino]-(2,3-dioxo-1,2,3,4-tetrahydro-quinoxalin-5-yl)-methyl]-phosphonic acid (NVP-AAM077) was a kind gift from Dr. Yves P. Auberson (Novartis Institutes for Biomedical Research, Basel, Switzerland). Trypsin, horse serum, fetal bovine serum, DMEM, neurobasal medium, and B27 were purchased from Life Technologies (Gaithersburg, MD). All chemicals used were of analytical grade.

Hippocampal neuron culture

All animal procedures were approved by the Institutional Animal Care and Use Committee of Second Military Medical University. Hippocampal neuron cultures were prepared as previously described (11). Briefly, hippocampi were dissected from E18 Sprague Dawley rat fetuses in an ice-cold dissection solution containing sucrose/glucose/HEPES [DISGH solution in mm concentration: 136 NaCl, 5.4 KCl, 0.2 Na₂HPO₄, 2 KH₂PO₄, 16.7 glucose, 20.8 saccharose, 10 HEPES, and 0.0012% phenol red (pH 7.4)]. Isolated hippocampi were mechanically triturated and then digested in a solution containing 0.25% trypsin and 1 mm EDTA at 37 C for 15 min. Single-cell suspension was obtained by repeatedly passaging dissociated tissues through fire-polished pipette in DMEM supplemented with 10% heat-inactivated fetal bovine serum and horse serum. Cells were finally plated on poly-l-lysine (0.1 mg/ml)-coated 24- or six-well plates, or glass coverslips for different experiments at optimal cell densities. The serum containing plating medium was replaced by a serum-free Neurobasal medium supplemented with 2% B27 (culture medium) within 24 h after plating. Half of the culture medium was changed every 3 d thereafter. More than 95% cells were neurons as verified by positive staining of microtubule-associated protein-2 (MAP2) against Hoechst at 10 DIV. All experiments were carried out on 10 DIV, when the hippocampal neurons developed a rich network of processes, expressed functional NMDA-type and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainate-type glutamate receptors, and formed synaptic contacts.

Drug application for neuroexcitotoxic assay

On 10 DIV, the growth medium of hippocampal neuron culture was aspirated and collected, neurons were washed with modified Locke’s solution (in mm concentration: 154 NaCl, 5.6 KCl, 3.6 NaHCO₃, 2.3 CaCl₂, 5.6 d-glucose, 10 HEPES, and 10 μm glycine, pH 7.4) and incubated in the presence or absence of NMDA, B, or B-BSA for 15 min at room temperature. Cells were then washed with DMEM and returned to the original growth medium and incubated in 37 C for 24 h or 12 h before toxicity measurements. For antagonist testing, RU36486 (1 μm), spironolactone (1 μm), MK-801 (0.3 μm), Ro25-6981 (0.5 μm), NVP-AAM077 (0.5 μm), SB203580 (10 μm), or PD98059 (10 μm) was pretreated for 20 min.

Neuronal toxicity/viability assay

Neuronal cell death was examined by morphology changes under phase-contrast microscopy and quantitatively assessed by measurement of LDH released into the incubation medium. Briefly, after 24 h recovery from various treatment, 50 μl of culture medium was taken out from each well, and the LDH activity was detected using a cytotoxicity assay kit (G1780; Promega Corp., Madison, WI) by measuring absorbance at 490 nm with microplate reader 680 (Bio-Rad Laboratories, Inc., Hercules, CA). The relative cell damage was presented by normalizing the OD value to control. Neuronal damage was also assessed by PI uptake, and the morphological change in the nuclear was examined by Hoechst staining. For PI uptake, PI (1 μg/ml) was added to culture medium 24 h after various treatments for 15 min at 37 C. PI-positive cells were calculated as dead cells. After fixation with 4% paraformaldehyde, neurons were incubated with Hoechst33342 (2 μg/ml) at room temperature for 15 min for nuclei staining. Nuclei with chromatin condensation or fragmentation were designated as damaged cells. In some cases, viability of neurons was also assessed by immunofluorescence staining of MAP2, which serves as a maker of survival neurons. Neurons were viewed and randomly photographed for about 20 sight-views from each treatment group by Olympus microscope (I×70; Olympus Corp., Lake Success, NY). Neuronal cell death was indexed by the percentage of PI-positive cells and/or condensed nuclei. Neuronal viability was calculated by the MAP2-positive cells per sight-view. Apoptotic or necrotic death of neurons was also shown by an apoptosis staining kit (BD Biosciences, Palo Alto, CA) using flow cytometry. After 12 h incubation following drug stimulation, neurons were collected for staining with Annexin V-PI double staining according to the instruction of the kit. Flow cytometry was performed using a FACSCalibur (fluorescent-activated cell sorter [FACS] Calibur) and data were analyzed using CellQuest software (both from Becton Dickinson, Mountain View, CA).

For LDH release assay, 490 nm OD values were normalized to control and expressed as percent of control. Results were from four independent experiments with eight replicates for each data point per experiment. For nuclear condensation measurement, cells were counted as damaged/dead cells against total cells. Results of PI uptake, MAP2, and Hoechst staining were from three independent experiments. In each experiment, 20–30 randomly captured fields from each well were counted with three replicates for each treatment. About 5000 cells were counted for each data point.

[Ca²⁺]_i measurement

Measurement of intracellular calcium concentration ([Ca²⁺]_i) was performed using the Ca²⁺-sensitive indicator fura-2. At 10 DIV, neurons grown in coverslips were loaded for 1 h at 37 C with 1 μm fura-2/AM (Dojindo, Japan) in the presence of 0.04% pluronic F-127 in Neurobasal/B27 culture medium to minimize the disturbance of neurons during handling. After fluorescence loading, the coverslips were rinsed in Mg²⁺-free balanced salt solution (BSS in mm concentration: 130 NaCl, 5.4 KCl, 2.0 CaCl₂, 5.5 glucose, 10 HEPES, and 10 μm glycine, pH 7.3) and mounted in a perfusion chamber and placed on the stage of an inverted microscope (I×70, Olympus) immediately for Ca²⁺ imaging. The neurons were incubated in the absence or presence of RU38486 or NMDA antagonists for 20 min and then perfused at a rate of 1.5 ml/min in BSS with or without B or/and RU38486, or NMDA antagonists. After recording of the basal level of [Ca²⁺]_i for 2 min, the neurons were perfused with BSS containing NMDA in the absence or presence of B and/or RU38486, or NMDA antagonists, and [Ca²⁺]_i was further traced for 4–5 min. For Ca²⁺ imaging, light was emitted from a 75 W xenon arc lamp (AH2-RX, Olympus) and passed through an excitation filter set (Chroma, Lake Forest, CA) to generate UV monochromatic waves of 340 and 380 nm. With the aid of a computerized filter wheel (Lambda 10–2; Sutter Instruments, Novato, CA), the cells on the coverslips were alternatively exposed to the two waves through an Olympus UApo objective. The resulting fluorescence emission from Ca²⁺-sensitive dye was collected through a 510-nm long-pass filter (Chroma) with a cooled charge-coupled device camera (MicroMax, 5 MHz system, Princeton Instruments, Princeton, NJ). All image acquisition was computer controlled by MetaFluor Imaging program (version 4.01; Universal Imaging Corp., Westchester, PA). Images were acquired at 3-sec intervals to reduce photobleaching. [Ca²⁺]_i in each cell was expressed as F340/F380, which is the ratio of the fluorescence intensity at 340 nm excitation (F340) to that at 380 nm excitation (F380). All measurements were made at room temperature (22–25 C).

Western blot analysis of MAPK activation

Hippocampal neurons (10 DIV) were starved for 6 h in DMEM without any supplement. Drugs were applied to the DMEM directly for 15 min (antagonists were pretreated for 20 min). Neurons were immediately washed with ice-cold 0.1 m PBS and incubated in ice-cold lysis buffer (0.1% sodium dodecyl sulfate, 1% Igepal, 0.2 mm sodium orthovanadate, 0.2 mm phenymethy sulfonyfluoride) for 20 min. Cell lysates were cleared by centrifugation at 12,000 × g for 10 min at 4 C, and protein concentration in the supernatant was determined by the BCA Protein Assay (Sigma). After being boiled for 5 min, 15 μm of total protein from whole-cell lysates was separated by 15% SDS-PAGE denaturing gel and were electrotransferred onto nitrocellulose membranes (Schleicher & Schuell, Inc.; Dassel, Germany). Membranes were blocked with 10% nonfat dry milk in TBST [50 mm Tris, 150 mm NaCl, and 0.1% Tween 20 (vol/vol), pH = 7.4] for 1 h and immunoblotted antiphosphorylated ERK1/2, p38 primary antibodies at a dilution of 1:1000 for overnight at 4 C. Horseradish peroxidase-conjugated secondary antibody was used together with ECL detection system (Pierce Chemical Co., Rockford, IL) to detect the final signal. Membranes were then stripped and reprobed with anti-ERK1/2 or anti-p38 antibodies to detect the total ERK1/2 or p38 level that served as self-control.

Quantifications of MAPK activation were done by scanning densitometry analysis on p-p38 and p-ERK1/2 against total p38 and total ERK1/2. Scanning densities were normalized to control and expressed as relative fold of control. Results were from three independent experiments.

Statistical analysis

Comparisons between different drug treatments were performed using a one-way ANOVA with the Newman-Keuls’s post hoc test. Differences were considered significant at P < 0.05. All values are presented as mean ± sem.

NURSA Molecule Pages:

• Ligands: Corticosterone | RU486 | Spironolactone.