Role of N-Methyl-d-aspartate Receptors in the Neuroprotective Activation of Extracellular Signal-regulated Kinase 1/2 by Cisplatin*

Neurons are exposed to damaging stimuli that can trigger cell death and subsequently cause serious neurological disorders. Therefore, it is important to define defense mechanisms that can be activated in response to damage to reduce neuronal loss. Here we report that cisplatin (CPDD), a neurotoxic anticancer drug that damages DNA, triggered apoptosis and activated the extracellular signal-regulated kinase 1/2 (ERK1/2) pathway in cultured rat cortical neurons. Inhibition of ERK1/2 activation using either pharmacological inhibitors or a dominant-negative mutant of the ERK1/2 activator, mitogen-activated protein kinase kinase 1, increased the toxicity of CPDD. Interestingly, N-methyl-d-aspartate (NMDA) receptor (NMDAR) antagonists reduced the ERK1/2 activation and exacerbated apoptosis in CPDD-treated neurons. Pre-treatment with CPDD increased ERK1/2 activation triggered by exogenous NMDA, suggesting that CPDD augmented NMDAR responsiveness. CPDD-enhanced response of NMDAR and CPDD-mediated ERK1/2 activation were both decreased by inhibition of poly(ADP-ribose) polymerase (PARP). Interestingly, PARP activation did not produce ATP depletion, suggesting involvement of a non-energetic mechanism in NMDAR regulation by PARP. Finally, CPDD toxicity was reduced by brain-derived neurotrophic factor, and this protection required ERK1/2. In summary, our data identify a novel compensatory circuit in central nervous system neurons that couples the DNA injury, through PARP and NMDAR, to the defensive ERK1/2 activation.

The central nervous system is exposed to damaging stimuli that may trigger neuronal death and cause serious neurological diseases (1). However, most neurons survive minor damages with which they are challenged during the life span of the organism. Therefore, one can expect the existence of defense mechanisms that help neurons to survive initial insult and resume proper functions after damage. Neurons receive multi-ple signals inhibiting cell death (2). For example, neuronal survival during development is promoted by neurotrophins and neurotransmitters. The effects of these agents on survival are mediated through several signaling molecules, including extracellular signal-regulated kinase 1/2 (ERK1/2) 1 and phoshpatidyloinositol 3-kinase (1,2). Consequently, survival signaling pathways are good candidates to contribute to the defense mechanisms in injured neurons.
Glutamate is an important neurotransmitter that promotes survival. It acts through several types of receptors, including two families of ionotropic receptors, AMPA and NMDA (3). NMDA receptor (NMDAR) is required for neuronal survival during development (4,5). On the other hand, excessive activation of NMDA signaling produces excitotoxicity (6). Therefore, NMDAR inhibitors are used to improve the outcome of several neurological diseases (7). In addition, it has been recently proposed that NMDAR antagonists may be active against malignant tumors and that their combination with anticancer chemotherapy would be a valuable therapeutic approach (8).
It is intriguing that neurons, which are postmitotic cells, demonstrate high vulnerability to DNA damage (9,10). Often, genotoxic anticancer agents including cisplatin (CPDD) produce neurological side effects that limit their usage against central nervous system tumors (11,12). DNA damage may also be an important trigger of neuron loss in common neurodegenerative diseases (13).
DNA damage activates both the reparative response and death signaling (14). For example, DNA damage may mobilize poly(ADP-ribose) polymerase (PARP), which ribosylates target proteins to activate DNA repair (15). The substrate of PARP is a highly energetic molecule, NADϩ. In consequence, PARP activation may deplete cellular energy stores, resulting in neuronal membrane depolarization with enhanced NMDAR signaling and, finally, necrotic cell death (15).
Genotoxin-induced neuronal death can be suppressed by neurotrophins (16 -18). The protection has been reported to require activation of the ERK1/2 pathway (17,18). Therefore, we hypothesized that DNA damage by CPDD can activate the ERK1/2 pathway as a defense response to support neuronal survival. Our results indicate that the CPDD-activated ERK1/2 pathway attenuates cortical neuron death. Interestingly, activation of ERK1/2 by CPDD depended on PARP and NMDAR activity and was not accompanied by decreased ATP levels. Our observations identify a novel protective pathway that couples DNA damage, through PARP and NMDAR, to ERK1/2. Our results also suggest the possibility of toxic interactions between NMDAR antagonists and genotoxic anticancer agents that reach the central nervous system.
Cell Culture and Transfection-Cortical neurons were prepared from newborn rats (Sprague-Dawley) and kept in basal medium Eagle supplemented with 10% heat-inactivated bovine calf serum (HyClone), as described (17). 1-(␤-D-Arabinofuranosyl)cytosine (2.5 M) was added at day 2 in vitro to inhibit proliferation of non-neuronal cells. Cortical neurons were transiently transfected at days 3 or 4 in vitro by using a modified calcium-phosphate co-precipitation protocol (17). Cortical neurons cultured on poly-D-lysine/laminin-coated (Sigma) glass coverslips in 35-mm plates were transfected with 4 g/plate expression plasmid DNA for wild-type MKK1 (MKK1wt), dominant-negative MKK1 (MKK1dn), constitutive-active MKK1 (MKK1ca), or empty cloning vector, pCEP4. In addition, neurons were co-transfected with 2 g/plate pON260 plasmid DNA, which contains an expression cassette for bacterial ␤-galactosidase. Expression of recombinant forms of MKK1 was confirmed by immunostaining for HA epitope tag attached to MKK1 proteins. In addition, cell nuclei were counterstained with Hoechst 33258 to reveal apoptotic alterations in chromatin structure. ␤-Galactosidase was used as a marker to identify transfected cells. Because ␤-galactosidase remains stable in dying cells (21), we were able to score for apoptosis in transfected neurons on the single cell level without a bias to exclude apoptotic cells.
Drug Treatment-At days 5 or 6 in vitro, cortical neurons were treated with CPDD or TPDD. These drugs were dissolved in dimethyl sulfoxide (Me 2 SO). The final concentration of Me 2 SO in the media was 0.2%. PD98059, SL327, MK-801, NBQX, CNQX, 3-ABA, PHEN, and BAPTA/AM were also dissolved in Me 2 SO. When cultures were cotreated with CPDD and one of these drugs, the final concentration of Me 2 SO was below 0.4%. BDNF was diluted in phosphate-buffered saline containing 0.1% bovine serum albumin before addition to the cells. SL327 or PD98059 was added 30 min before BDNF in co-treatment experiments.
Quantitation of Apoptosis by Nuclear Morphological Changes-To visualize nuclear morphology, cells were fixed in 4% paraformaldehyde and stained with 2.5 g/ml of the DNA dye Hoechst 33258 (Sigma) (17). Apoptosis was quantitated by scoring the percentage of cells with apoptotic nuclear morphology at the single cell level after Hoechst staining. Uniformly stained nuclei were scored as healthy, viable neurons. Condensed or fragmented nuclei were scored as apoptotic. To obtain unbiased counting, samples were coded, and cells were scored blind without knowledge of their prior treatment.
Quantitation of Neuronal Survival by MTT Assay-The MTT assay was performed in 96-well plates as described (17).
Caspase Assay-Caspase assay was performed with a colorimetric caspase assay kit purchased from Promega. For each measurement, protein lysate from 1 ϫ 10 6 cells was used. To reveal the caspase-dependent activity, each sample was incubated with the caspase substrate in the absence or presence of 10 M Z-VAD-fmk, a specific caspase inhibitor.
Western Analysis and Immunostaining-Western blot analysis with anti-phospho-ERK1/2 or anti-ERK1/2 antibodies was performed as described (17). Briefly, 10 g of total protein was used in each lane. In addition, blots were re-probed with antibodies specific for total ERK1/2 or ERK2 to ensure equal protein loading of the blots. Quantification of phospho-ERK1/2 was performed by densitometric analysis and normalized against total ERK1/2. Western blot analysis with anti-poly(ADPribose) antibody was performed using 4 g of nuclear proteins per Arrows indicate viable neurons; arrowheads indicate neurons with signs of apoptosis. Note the increased number of apoptotic cells after CPDD treatment (H and I). J, kinetics of CPDD-induced apoptosis in cortical neurons. Neuronal apoptosis was visualized by Hoechst staining and scored as described in the text. Cell death significantly increased at 24 and 48 h after CPDD treatment. CPDD induced more apoptosis at 5 than at 10 g/ml (p Ͻ 0.01). K, in addition to CPDD, neurons were also treated with either vehicle, phosphate-buffered saline (Veh), or a protein synthesis inhibitor, cycloheximide (1 g/ml, CHX). After 24 h, apoptosis was scored. Please note that CHX significantly protected neurons from CPDD-induced cell death. **, p Ͻ 0.01. L, TPDD, a non-DNA-damaging CPDD isomer, did not induce neuronal apoptosis after 24-h treatment. J-L, averages of triplicate determinations from at least three independent experiments are shown. Bars indicate S.E. sample. Nuclear proteins were extracted as described (22). Transfected cells were detected by immunostaining with an antibody against ␤-galactosidase and Texas Red-conjugated goat antibody to rabbit immunoglobulin. Cells transfected with the HA epitope-tagged constructs were also immunostained with an antibody to HA followed by fluoresceinconjugated goat antibody to mouse immunoglobulin.
Determination of Glutamate and Glycine Concentration-Amino acids were extracted from culture media with 0.6 N perchloric acid, followed by centrifugation and neutralization with potassium hydroxide. The analysis was performed as described (23).
NADϩ and ATP Assays-To extract nucleotides, 2 ϫ 10 6 cells were treated with 3.5% perchloric acid as described (24). NADϩ level was measured in a reaction catalyzed by alcohol dehydrogenase as described (25). ATP measurement was performed according to the two-step procedure of Williamson and Corkey (26).
Statistical Analysis-Statistical analysis of the data was performed by using one-or two-way analysis of variance followed by post hoc tests.

CPDD-induced Apoptosis in Cortical
Neurons-To test the hypothesis that DNA damage can activate the defensive ERK1/2 pathway, we treated cortical neurons with CPDD. CPDD applied for 48 h reduced neuronal survival (Fig. 1A). Interestingly, cells exposed to 10 g/ml CPDD showed higher survival rates than cells treated with 5 g/ml (56.4 versus 28%, Fig. 1A). Neurons treated with CPDD showed an apoptotic pattern of DNA fragmentation (Fig. 1B) and activation of proapoptotic caspases (Fig. 1C). In addition, cells dying in response to CPDD displayed morphological features of apoptosis including fragmentation and condensation of nuclear chromatin ( Fig. 1, D-J). 10 g/ml CPDD induced significantly less apoptosis than 5 g/ml (at 24 h, 28 versus 51.2%, respectively; p Ͻ 0.01) (Fig. 1J). A translation inhibitor, cycloheximide, protected against CPDD-induced death (Fig. 1K), suggesting that protein synthesis may be involved in this process. This finding is also consistent with the apoptotic character of CPDDinduced death in cortical neurons. In addition, TPDD, an isomer of CPDD that is unable to induce DNA strand breaks (27), did not produce neuronal apoptosis (Fig. 1L), indicating that CPDD-induced neuronal apoptosis is triggered by DNA damage.
Activation of ERK1/2 Pathway in CPDD-treated Cortical Neurons-Phosphorylation of ERK1/2 residues Thr 183 and Tyr 185 (position numbers as in human ERK2) by MKK1/2 controls ERK1/2 activation (28). Therefore, we determined the activity of the ERK1/2 pathway by immunoblotting for phosphorylated ERK1/2 (Fig. 2). In cells exposed to 10 g/ml CPDD, activation peaked at 24 h after treatment (Fig. 2, A and B). At that time point, the extent of CPDD-mediated activation of the ERK1/2 pathway was directly proportional to the concentration of CPDD (Fig. 2 C and D), with the maximum stimulation by 10 g/ml (5.9-fold above controls). Maximal ERK1/2 activation by 10 g/ml CPDD correlated with the decreased toxicity of CPDD at 10 g/ml, as compared with 5 g/ml (Fig. 1). This finding suggests ERK1/2 involvement in the defensive reaction to damage. Importantly, TPDD (10 g/ml) did not increase ERK1/2 activity (Fig. 2E). ERK1/2 activation by CPDD was not affected by cycloheximide (1 g/ml) (Fig. 2F). These data suggest that the ERK1/2 response is triggered by DNA damage and does not involve protein synthesis.
CPDD-mediated Activation of ERK1/2 Supports Neuronal Survival-Pharmacological inhibitors of the ERK1/2 pathway, PD98059 or SL327, effectively abolished ERK1/2 activation by CPDD (Fig. 3, A and D). Therefore, we used these compounds to determine the effect of the ERK1/2 response on CPDD-induced cell death. Consistent with our previous observations (17), neither SL327 (50 M) nor PD98059 (40 M) significantly affect basal apoptosis in cortical neurons (Fig. 3, B and E). By 24 h, cells exposed to either 5 or 10 g/ml CPDD showed a significant increase of apoptosis upon co-treatment with SL327 (28.3% at 10 g/ml CPDD versus 57.8% at SL327 ϩ 10 g/ml of CPDD; p Ͻ 0.001) (Fig. 3B). Interestingly, the CPDD concentration dependence of the SL327 effect on apoptosis correlated with the concentration dependence of CPDD-mediated ERK1/2 activation (Fig. 2, C and D). PD98059 also increased apoptosis in- Neurons were treated as indicated in each panel. ERK1/2 pathway activity was determined by Western blotting using an antibody specific for phosphorylated ERK1/2. Blots were re-probed with an antibody specific for total ERK1/2. Phospho-ERK1/2 (pERK1/2) levels are relative to controls. Data in B and C represent averages of three independent experiments. Bars indicate S.E. A and B, kinetics of ERK1/2 activation after treatment with 10 g/ml CPDD. C and D, dose response of ERK1/2 activation by CPDD at 24 h. E, TPDD, a non-DNA damage-inducing isomer of CPDD, fails to increase ERK1/2 activity. Neurons were treated for 24 h with either TPDD or CPDD at 10 g/ml. F, cycloheximide (CHX) does not reduce ERK1/2 activation by CPDD. Neurons were treated for 24 h with cycloheximide (1 g/ml) and CPDD (10 g/ml) as indicated. Numbers under the blots in E and F are relative pERK1/2 levels. Results shown in E and F were replicated in three independent experiments. duced by CPDD (Fig. 3E, p Ͻ 0.001). In addition, SL327 (50 M) or PD98059 (40 M) further reduced neuronal viability after CPDD treatment (Fig. 3, C and F). To complement the pharmacological approach, we studied the effects of a dominantnegative mutant form of the ERK1/2 activator, MKK1 (MKK1dn) (20), on CPDD-induced apoptosis. Cortical neurons were transfected with expression plasmids for either wild type MKK1 (MKK1wt), MKK1dn, or empty cloning vector, pCEP4 (Fig. 3, G and H). Forty-eight hours after transfection, neurons were treated for 24 h with either vehicle (0.2% Me 2 SO) or CPDD (10 g/ml). In vehicle-treated cells, apoptosis was unaf-fected by any of the transfected plasmids (average of 15.5%, Fig. 3I). In contrast, CPDD caused apoptosis in more cells expressing MKK1dn (46.5 versus 31% in pCEP4 or 26% in MKK1wt-transfected neurons; p Ͻ 0.001, Fig. 3I). Therefore, inhibition of ERK1/2 increased apoptotic cell death induced by CPDD.
CPDD Activates ERK1/2 through NMDA Receptors-ERK1/2 can be activated in cortical neurons by the glutamate-triggered influx of Ca 2ϩ ions through an open NMDAR channel (29). Because moderate activity of NMDAR is implicated in antiapoptotic signaling (5, 30), we tested for NMDAR involvement in the protective ERK1/2 activation by CPDD.
CPDD Increases NMDAR Signaling-NMDAR can be activated by increased concentration of its ligands, glutamate or NMDA (Fig. 5A). If increased concentrations of NMDAR ligands are responsible for CPDD activation of ERK1/2, one would expect that conditioned media from cells that are exposed to CPDD would produce ERK1/2 activation in untreated neurons. However, ERK1/2 was not activated by conditioned media collected from cells treated for 24 h with 10 g/ml CPDD (Fig. 5B). Consistently, media concentrations of the NMDAR ligand, glutamate, and its co-ligand, glycine (3), did not significantly increase after CPDD treatment (Fig. 5C). Therefore, it seems unlikely that activation of NMDAR by CPDD is caused by an elevated release of the NMDAR ligands.
An alternative possibility is that CPDD increases the neuronal responses to basal levels of NMDAR stimulation. Indeed, we found that neurons that were pretreated with 5 g/ml CPDD for 24 h demonstrated increased ERK1/2 responses to NMDA (Fig. 5, D and E). In cells that were not exposed to CPDD, a 5-min treatment with NMDA evoked a concentrationdependent ERK1/2 activation that appeared at 20 M (6.1-fold above control) and declined at 50 or 100 M (3.6-or 2.3-fold above control, respectively; Fig. 5, D and E). Neurons that were pretreated with 5 g/ml CPDD for 24 h responded with significantly enhanced ERK1/2 activation (p Ͻ 0.01) which was present at 10 M NMDA (5.3-fold above control), reached maximal levels at 50 M (7.6-fold above control), and slightly declined at 100 M (4.9-fold above control; Fig. 5, D and E). Therefore, it appears that CPDD decreases the threshold of NMDAR stimulation that is required to activate ERK1/2 and inhibits desensitization of the ERK1/2 response following more intense NMDAR stimulation. These data indicate that CPDD enhances intracellular signaling by NMDAR.
PARP Activity Contributes to CPDD-mediated ERK1/2 Activation-ERK1/2 activation apparently was caused by CPDDinduced DNA strand breaks (Fig. 2E) and was dependent upon NMDAR (Figs. 4 and 5). Thus, augmentation of NMDAR signaling in CPDD-treated neurons may be secondary to DNA damage. It has been proposed that a DNA damage-response enzyme, PARP, may regulate NMDAR in neurons (15). Therefore, we have evaluated the possibility that the CPDD-mediated increase in NMDAR signaling is mediated by PARP. Indeed, we observed increased PARP activity after CPDD treatment, indicated by the elevated polyribosylation of neuronal nuclear proteins (Fig. 6A). Also, cellular NADϩ content that is reduced during PARP activation (31) significantly decreased after CPDD (Fig. 6B). The NADϩ decrease preceded the CPDD-induced reduction of neuronal survival (Fig. 6C). These data indicate that CPDD activates PARP.
To evaluate whether PARP activation can contribute to the protective ERK1/2 activation by CPDD, we studied the effects of the PARP inhibitors, 3-ABA (5 mM) or PHEN (50 M), on CPDD-induced ERK1/2 activation or cell death. 3-ABA or PHEN reduced ERK1/2 activation in response to CPDD (Fig.  6D) and increased neuronal apoptosis triggered by CPDD (Fig.  6E). Neuronal apoptosis in basal conditions was not affected by either 3-ABA or PHEN (data not shown). Also, the CPDDmediated reduction in neuronal survival was enhanced in the presence of 3-ABA (Fig. 6F). Therefore, our data suggest that PARP contributes to protective ERK1/2 activation by CPDD.
To further support the idea that PARP activation contributes to the increased responsiveness of NMDAR in CPDD-treated neurons, we evaluated the effects of 3-ABA on ERK1/2 activation by NMDA in cells that were pretreated with CPDD (5 g/ml). Interestingly, neurons pretreated with CPDD in the presence of 3-ABA did not show enhanced ERK1/2 signaling after stimulation with NMDA (Fig. 6G). PARP inhibition abolished ERK1/2 activation at 10 M NMDA (5.9-versus 0.7-fold above control; Fig. 6G), decreased the activation level at 20 M (10.2-versus 1.8-fold above control; Fig. 6G), and increased desensitization of ERK1/2 response at 100 M NMDA (7.4versus 1.3-fold above control; Fig. 6G). These data suggest that PARP enhances NMDAR-mediated ERK1/2 activation in CPDD-treated neurons.
The PARP-mediated increase in NMADR signaling has been suggested to result from ATP depletion after increased NAD ϩ re-synthesis (31). However, we did not find a significant reduction of ATP levels until at least 12 h after CPDD addition (Fig.  6H). A moderate decrease of ATP content occurred at 24 h (79.1% of control values; p Ͻ 0.001). This alteration correlates with 82.4% cell survival found at 24 h after 10 g/ml CPDD treatment (Fig. 6C). Therefore, the reduced ATP content most likely reflects cell loss rather than the energetic deprivation of living neurons. In summary, these data suggest that PARP regulates NMDAR-mediated ERK1/2 activation through an ATP depletion-independent mechanism. BDNF Reduces CPDD Toxicity by Activation of ERK1/2 Pathway-A question can be raised whether neuroprotective agents that activate ERK1/2 would be able to reduce CPDD-mediated cell death. A 6-h treatment with 10 ng/ml BDNF activated ERK1/2 pathway in the absence or presence of CPDD (5 g/ml) (Fig. 7A). BDNF also decreased CPDD-induced apoptosis from 48.5 to 22.0% (Fig. 7B, p Ͻ 0.001). This effect was abolished by 50 M SL327 indicating that ERK1/2 is required for BDNFmediated protection of CPDD-treated neurons.
To determine whether ERK1/2 activation is sufficient for BDNF-mediated protection, we used a constitutively active mutant form of the ERK1/2 activator, MKK1 (MKK1ca) (20). Cortical neurons were transfected with either MKK1ca or empty cloning vector, pCEP4. Forty-eight hours after transfection, neurons were treated for 24 h with 5 g/ml CPDD. Neurons receiving MKK1ca were protected against CPDD-induced apoptosis (50.3% in pCEP4 versus 19.0% in MKK1ca-transfected cells; p Ͻ 0.001) (Fig. 7C). These results suggest that activation of ERK1/2 is both required and sufficient for BDNF to reduce CPDD-induced apoptosis in neurons. DISCUSSION In this study, we tested the possibility that ERK1/2 is activated by DNA damage to support survival of stressed neurons. Indeed, we observed that CPDD-induced apoptosis in rat primary cortical neurons was accompanied by anti-apoptotic ERK1/2 activation. We also identified PARP and NMDAR as FIG. 5. CPDD increases neuronal responses to NMDA. A, NMDA receptor stimulation by exogenous NMDA produced ERK1/2 activation after 5 min. The activation was sensitive to NMDA receptor blockers (10 M MK-801, 100 M APV, or 1 mM/10 mM KyMg). Similar results were obtained in two independent experiments. B, culture media from CPDD-treated neurons were unable to activate ERK1/2 in naïve cells. Neurons were treated for 24 h with either vehicle (0.2% Me 2 SO, Veh) or CPDD (10 g/ml). Media were removed and saved (conditioned media, CM). CPDD treatment resulted in ERK1/2 activation. Untreated neurons were placed in conditioned media for the indicated time. Similar results were obtained in two independent experiments. Numbers in A and B represent relative pERK1/2 levels. C, CPDD does not affect glutamate or glycine levels in the media. Neurons were treated with CPDD (10 g/ml), and culture media were collected at indicated time points. High pressure liquid chromatography analysis of glutamate and glycine concentrations in the media was performed as described in the text. Data represent averages of triplicate determinations from three independent experiments. Error bars are S.E. D and E, CPDD increases ERK1/2 responsiveness of neurons to NMDA. Cells were treated with CPDD (0 or 5 g/ml) for 24 h and were then exposed for 5 min to NMDA as indicated. pERK1/2 levels were determined by Western blotting. Data represent averages of three independent experiments. Error bars indicate S.E. mediators of CPDD-induced ERK1/2 response (Fig. 8). Finally, we showed that ERK1/2 activation was both necessary and sufficient for BDNF-mediated protection against CPDD-induced apoptosis.
In our hands, CPDD induced cortical neuron apoptosis. CPDD was also shown to induce apoptosis in peripheral nervous system neurons of sensory and auditory systems (32,33). Apoptotic death of these neuronal populations has been suggested as a mechanism of peripheral neurotoxicities of CPDD. Similarly, CPDD-induced cortical neuron apoptosis may underlie the robust central nervous system neurotoxicity observed after local delivery of CPDD to treat intracranial tumors (12).
CPDD activated ERK1/2 by increasing NMDAR signaling. CPDD increased NMDAR sensitivity to low concentrations of NMDA and also inhibited desensitization of ERK1/2 response after intense receptor stimulation. Furthermore, our results indicate that the mechanism of enhanced NMDAR signaling involves PARP activation. It has been suggested that the PARP-mediated increase of NMADR signaling results from ATP depletion after increased NADϩ resynthesis (15,31). However, we did not find a significant reduction of ATP levels in neurons that showed robust activation of ERK1/2. Consequently, the possible mechanisms of PARP-mediated NMDAR regulation in CPDD-treated cells may include posttranscriptional modifications, differential expression of NMDAR subunits, depolarization of the membrane through activity of other glutamate receptors, and, finally, enhanced coupling of NMDAR to ERK1/2.
In hippocampus, enhancement of the NMDAR response is produced by the primary increase of AMPA/KA receptor activity (3). This mechanism is unlikely to explain increased NMDAR signaling in CPDD-treated neurons because AMPA/KA receptor blockers did not affect ERK1/2 activation by CPDD. The alternative mechanism producing enhanced NMDAR signaling during development is increased expression of NMDAR subunits (34). However, we found no CPDD-induced increases in expression of the two NMDAR subunits, NR1 or NR2B, whose expression is detectable in cultured cortical neurons. 2 Therefore, it remains to be resolved which mechanism contributes to CPDD/PARP-mediated regulation of NMDAR.
It is well established that excessive stimulation of NMDARs results in excitotoxic neuronal death (6). However, the toxic abilities of NMDARs are not directly proportional to their ac-2 A. Gozdz and M. Hetman, unpublished data. Cells were then stimulated with NMDA for 5 min as indicated, and ERK1/2 activity was determined by Western blotting. Relative pERK1/2 levels are indicated under the blot. Similar results were obtained in two independent experiments. H, PARP activation by CPDD does not induce ATP depletion. Cells were treated with 10 g/ml CPDD for indicated time periods. Cellular ATP content was determined as described in the text. Because the ATP decline coincides with cell death (Fig. 6C), it is likely to reflect a reduction in the number of viable cells. Data are averages of triplicate determinations obtained in three independent experiments. ***, p Ͻ 0.001. Error bars indicate S.E. tivity. In fact, some NMDAR activity is required for optimal survival of various populations of central nervous system nerve cells, including cortical neurons (5,30). Also, moderate concentrations of NMDA can protect cortical neurons from trophic deprivation-induced death (35,36). Therefore, NMDAR signaling plays an important role to support neuronal survival during development. Our results identify a new role for the antiapoptotic activity of NMDAR in protecting neurons exposed to genotoxic stress.
Our data suggest that the activation of PARP in neurons is protective against CPDD-induced cell death. Protective properties of PARP were also reported in various other cell types including cancer cells, embryonic mouse tissue, and rat hippocampal neurons (37)(38)(39). For instance, in a rat global ischemia model, CA1 neuron loss was enhanced by the PARP inhibitor 3-ABA (38). On the other hand, other reports showed that intense PARP activation depletes cellular energy stores and induces necrosis (15). The discrepancies between observations suggesting that PARP mediated protection and those indicating that deleterious PARP effects can be attributed to the differences in the intensity of PARP activation. For example, PARP activation after CPDD resulted in a 20% reduction of NADϩ levels at 12 h after treatment and undetectable changes in ATP levels. Similarly, protective PARP activation by global ischemia did not significantly reduce NADϩ levels (38). In contrast, deleterious PARP activation after treatment with an alkylating drug, N-methyl-NЈ-nitro-N-nitrosoguanidine, decreased NADϩ levels by at least 80% within 60 min (40). Therefore, PARP activation by CPDD may be insufficient to produce pro-necrotic energy depletion. In conclusion, data presented here suggest that, as in the case of NMDAR signaling, moderate PARP activation also promotes neuronal survival.
Defensive activation of ERK1/2 by CPDD is not a unique stress-activated compensatory response in neurons. In fact, it has been revealed that NF-B activated by various forms of stress protects neurons from death (41). Interestingly, Gonzalez-Zulueta et al. (42) reported that transient ischemic stimulation activated ERK1/2 by NMDAR. Inhibition of this signaling increased sensitivity to a subsequent ischemic insult. Therefore, ERK1/2 activation by NMDAR may be used by neurons to resist various forms of injury.
In summary, CPDD-induced genotoxic stress activated an anti-apoptotic ERK1/2 response. This effect was mediated by PARP, which enhanced NMDAR signaling in CPDD-treated neurons. Therefore, our results identify a novel compensatory circuit to defend central nervous system neurons against genotoxic apoptosis. This defensive pathway couples DNA damage through PARP and NMDAR to ERK1/2 activation. Moreover, our data suggest that PARP regulates NMDAR signaling using a novel, ATP decline-independent mechanism. Our data also indicate the possibility of potentially toxic interactions between clinically used NMDAR antagonists, including ketamine or memantine, and genotoxic therapies that target tumors.