p38 Mitogen-activated Protein Kinase Mediates Hypoxic Regulation of Mdm2 and p53 in Neurons*

The multifunctional tumor suppressor protein, p53, inhibits cell growth and promotes differentiation and programmed cell death. p53 activity is controlled by transcriptional, translational, and post-translational regulation. A major pathway for post-translational regulation of p53 comprises its nucleocytoplasmic transport and subsequent proteasomal degradation, which involves binding to the oncoprotein, murine double minute-2 (Mdm2). Hypoxia and other stress signals cause cellular injury partly through the action of p53. In this study, we show that hypoxia induces down-regulation of Mdm2 as well as serine 15 phosphorylation and nuclear accumulation of p53 in cultured cortical neurons from E16 mice. These effects are diminished by the p38 mitogen-activated protein kinase inhibitors SB203580 and SB202190, but not by the inactive analog SB202474, and by a dominant-interfering mutant of the p38-activating kinase mitogen-activated protein kinase kinase 3 (MKK3). Hypoxic neuronal death was also reduced by p38 inhibitors, by dominant-interfering MKK3, and by a p53-antisense oligodeoxynucleotide and was increased by a constitutively active form of p38 and by an Mdm2antisense oligodeoxynucleotide. These results demonstrate that p38 and Mdm2 have roles in coupling hypoxic-ischemic neuronal insults to activation of p53 and hypoxic cell death.

Hypoxia is an important pathophysiological feature of ischemic disorders, including stroke. Like other stress signals, neuronal hypoxia and ischemia cause DNA damage and cell death partly by promoting nuclear accumulation of the p53 tumor suppressor protein (1,2). p53 is a multifunctional protein that has a critical role in various pathways controlling cellular responses to stress signals (for reviews, see Refs. 3 and 4). p53 is normally expressed at low levels, in a latent form that is unable to bind specifically to DNA, by rapid degradation through ubiquitin-dependent proteolysis (5). With stress, p53 accumulates via multiple mechanisms, including enhanced translation, decreased proteolytic degradation, and post-translational modification (6 -8). Under these conditions, its half-life is extended from 30 to Ͼ200 min, contributing to an increase in p53 protein levels (7).
The murine double minute-2 (Mdm2) 1 oncoprotein, a product of a p53-reponsive gene, is a major inhibitor of p53 function and abundance (9,10). Mdm2 was originally identified as an amplified gene in a spontaneously transformed derivative of the BALB/c cell line, 3T3 DM (9). The Mdm2 gene contains a p53 DNA-binding site and a genetically responsive element such that expression of Mdm2 can be regulated by the level of wild-type p53 protein. The Mdm2 protein, in turn, can complex with p53 and decrease its ability to act as a positive transcription factor (10). In addition to its ability to antagonize p53-dependent transcription, Mdm2 can also promote degradation of p53 through a ubiquitin-dependent proteasome pathway (6,11) and may have an additional role in nucleocytoplasmic shuttling of p53. This establishes a negative feedback loop, in which p53 initiates its own Mdm2-mediated destruction (10,12). Posttranslational modifications of p53 include phosphorylation and acetylation (8,13), and phosphorylation of p53 at serine 15 and serine 20 leads to reduced binding to Mdm2, which enhances p53 accumulation (8,14). Finally, reduced expression of Mdm2 has been implicated in the induction of p53 by hypoxia in non-neuronal cells (15). How neuronal hypoxia or ischemia regulates p53 is unknown, but hypoxia influences a variety of signal transduction mechanisms, including mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K)/Akt pathways (16 -18). Among the MAPKs, extracellular signal-related kinase (ERK) has been widely associated with cell survival, whereas c-Jun amino-terminal kinases (JNK) and p38 are often implicated in cell death (16,19). However, the actual roles of each MAPK cascade are highly cell type-and context-dependent (16). In the mouse JB6 epidermal cell line, activated ERK and p38, but not JNK, phosphorylates p53 at serine 15 and reduces binding of p53 to Mdm2 (8), whereas in mouse fibroblasts, JNK associates with p53 and promotes its ubiquitination (20).
Stress signals can also up-regulate p38 in neurons (19,21,22). For example, NO donors activate p38 and induce apoptosis in cortical neurons without activating JNK, and the p38 MAPK activity inhibitor SB203580 protects cortical neurons from NOinduced cell death (22). These and related findings prompted us to assess the relationship between p38 activity and levels of Mdm2 and p53 in primary cultures of cortical neurons subjected to hypoxic insults. We report here evidence that suggests a role for p38 MAPK in regulating Mdm2 and p53 levels in hypoxic neurons.
Cell Culture-Primary neuronal cultures were established as previously described (23,24). In brief, cerebral cortex was dissected from fetal CD1 mice at 16 days of gestation, treated with trypsin for 3 min at 37°C, and dissociated by trituration. Dissociated cell suspensions were plated at 3.5 ϫ 10 5 cells/cm 2 on plastic tissue culture dishes coated with poly-D-lysine, in defined medium (Neurobasal/B27, Invitrogen) supplemented with 2 mM glutamine, penicillin (25 units/ml), and streptomycin (25 g/ml). Cultures were maintained in a humidified 5% CO 2 incubator at 37°C for 5 days before treatment. Under these conditions, cultures contained ϳ95% neurons as reported previously (25).
Hypoxia-To induce hypoxia, cells were placed in modular incubator chambers (Billups-Rothenberg, Del Mar, CA) for 0 -24 h at 37°C in humidified 95% N 2 , 5% CO 2 (25). To evaluate the effect of signaling pathways on Mdm2 and p53 expression, the inhibitors listed above were added 1 h before the onset of hypoxia.
Western Blots-Cells were washed twice in phosphate-buffered saline, and whole cell extracts were prepared by adding 10 volumes of 1 ϫ sample buffer (2% SDS, 100 mM dithiothreitol, 60 mM Tris, pH 6.8, and 10% glycerol) and boiled for 5 min. Nuclear extractions were performed as previously described (27) with modifications. Briefly, cells were washed with cold phosphate-buffered saline, collected by centrifugation at 1000 ϫ g for 5 min, and then resuspended in buffer A (10 mM Hepes, pH 7.9, 10 mM KCl, 0.1 mM EDTA, pH 8.0, 0.1 mM EGTA, pH 8.0, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 2 g/ml aprotinin, and 0.5 g/ml benzamidine). The cell suspension was incubated on ice for 15 min before adding Nonidet P-40 to a final concentration of 0.6%, vortexed, and centrifuged at 13,000 ϫ g for 15 min. The supernatant was collected as the cytoplasmic fraction, and the nuclear pellet was washed with buffer A (without Nonidet P-40) and suspended in buffer B (20 mM Hepes, pH 7.9, 400 mM NaCl, 1 mM EDTA, pH 8.0, 1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 2 g/ml aprotinin, and 0.5 g/ml benzamidine). After incubation on ice for 15 min, the suspension was centrifuged at 13,000 ϫ g for 15 min. The nuclear extract was collected, and protein concentration was determined using Bradford protein assays (Bio-Rad); 40 g of protein was analyzed by 10 or 12% SDS-PAGE and transferred to ImmunoBlot PVDF membrane (Bio-Rad). Mem- p38 Mediates Hypoxic Regulation of Mdm2 and p53 in Neurons branes were probed with primary antibody overnight, and the signal was detected with Roche Molecular Biochemicals chemiluminescence blotting kits.
Transient Transfection of Primary Cortical Neurons-Cortical neurons were transiently transfected at day 3 using a Ca 3 (PO 4 ) 2 co-precipitation protocol as previously described (19,28). Expression constructs encoding constitutively active MAPK kinase 3 (MKK3) (pRc/RSV-Fla-MKK3 (Glu)) and dominant negative MKK3 (pRc/RSV-Fla-MKK3 (Ala)) (29) were kindly provided by Dr. Roger J. Davis, Howard Hughes Medical Institute, Program in Molecular Medicine, University of Massachusetts Medical School. Briefly, the DNA-Ca 3 (PO 4 ) 2 complexes were prepared by mixing 8 g of DNA per 35-mm dish with 2.5 M CaCl 2 and 2 ϫ Hepes-buffered saline (274 mM NaCl, 10 mM KCl, 1.4 mM Na 2 HPO 4 , 15 mM D-glucose, and 42 mM Hepes, pH 7.07). The precipitates were allowed to form for 20 -50 min before addition to the cultures. Cells were washed three times with modified Eagle's medium, and 1.5 ml of transfection medium was added to each 35-mm dish. The transfection medium consisted of Neurobasal/B27 supplemented with 1 mM sodium kynurenate, 10 mM MgCl 2 , and 5 mM Hepes. The solution containing DNA precipitates was added dropwise to the cultures and gently mixed. After 1 h, the cells were washed three times with modified Eagle's medium and were treated with 1 ϫ Hepes-buffered saline (see above), 1 mM sodium kynurenate, 10 mM MgCl 2 in 5 mM Hepes, and 5% glycerol for 2 min. Cells were then washed three times with modified Eagle's medium, and the conditioned medium was added back to each dish; 24 h later, the cells were harvested for analysis. An expression construct encoding green fluorescent protein (pEGFP, CLONTECH) was used to verify transfection efficiency and as a control. Cells were fixed with 4% paraformaldehyde, counterstained with 4,6-diamidino-2-phenylindole (Vector), and viewed by Nikon E800 epifluorescence microscopy.
Oligodeoxynucleotide (ODN) Transfections-A phosphorothioate antisense ODN (MAS20) labeled with fluorescein at the 5Ј end and directed against the initial coding region of the target Mdm2 mRNA (5Ј-GACATGTTGGTATTGCACAT-3Ј), a scrambled MX20 ODN (5Ј-GA-CATGTTGCTATTGCACAT-3Ј) (30), a phosphorothioate antisense ODN (p53as) labeled with fluorescein at the 5Ј end and directed against the initial coding region of the target p53 mRNA (GenBank TM accession number AF161020) (5Ј-ATGGCAGTCATCCAGT-3Ј), and a sense p53S ODN (5Ј-ACTGGATGACTGCCAT-3Ј) were synthesized commercially (Operon Technologies) and purified by high performance liquid chromatography. Cultures were transfected with ODNs (5 M) using Fu-GENE 6 (Roche Molecular Biochemicals). After 10 h, cells were retreated with ODNs (5 M) and maintained in culture for another 24 -48 h, when they were analyzed for cell viability assay and for protein expression by Western blotting.
Cell Viability Assay-Neuronal cell viability was assessed by measuring formazan produced by the reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) in viable cells. Cells were incubated with 5 mg/ml MTT at 37°C for 2 h. The medium was removed and cells were solubilized with dimethyl sulfoxide and transferred to 96-well plates. The formazan reduction product was detected by measuring absorbance at 570 nm in a Cytofluor Series 4000 multiwell plate reader (PerSeptive Biosystems, Framingham, MA). Results were expressed as a percentage of control absorbance, measured in normoxic cultures, after subtracting background absorbance (measured in freezethawed cultures) from all values.
Data Analysis-Quantitative data were expressed as mean Ϯ S.E. from at least three experiments. Analysis of variance and Student's t test were used for statistical analysis, with p Ͻ 0.05 considered significant.

RESULTS
Hypoxia Down-regulates Mdm2 and Up-regulates p53 in Cortical Neurons-To determine whether hypoxia alters the expression of p53 and Mdm2 genes in neurons cultured from mouse cerebral cortex, cells were deprived of oxygen for 0 -24 h, total RNA was extracted, and p53 and Mdm2 expression was analyzed by Northern blotting. Mdm2 was prominently downregulated in a time-dependent manner, while p53 increased, compared with the control gene, ␤-actin (Fig. 1). Western analysis of Mdm2 and p53 proteins in nuclei and cytoplasm of cells fractionated after 0 -24 h of hypoxia showed a reduction in Mdm2 protein levels in both compartments and an increase in nuclear p53 (Fig. 2A). Each of these changes was detectable after 4 h and persisted for at least 24 h. At all times examined, Mdm2 expression was most pronounced in the cytoplasm, and p53 predominated in the nucleus.
Hypoxia Induces Phosphorylation of p53 at Serine 15-Phosphorylation of p53 at serine 15 helps to regulate p53 stability by reducing its interaction with Mdm2 and has been reported after UV exposure (8). To explore the possibility that p53 is also phosphorylated in response to hypoxia, we analyzed cell extracts from normoxic and hypoxic cortical neurons by Western blotting with an antibody directed against serine 15 phospho-p53. As shown in Fig. 2B, the level of serine 15 phospho-p53  Hypoxia Activates ERK and p38 MAPK but Decreases JNK Activity-The specific signal transduction pathways involved in cellular responses to hypoxia vary with cell type; for example, HeLa cells respond with a rapid but transient activation of ERK (17), whereas p38 (but not JNK) is selectively activated in PC12 cells (31). To understand the mechanisms underlying the down-regulation of Mdm2 and serine 15 phosphorylation and up-regulation of p53 in hypoxic cortical neurons, we measured changes in the expression of phosphoactivated forms of ERK, p38, and JNK. Using phosphospecific antibodies against ERK, p38, and JNK, we found that phospho-ERK1 and phospho-ERK2 were present at low levels in normoxic neuronal cultures and that exposure to hypoxia for 4 to 24 h caused a progressive increase in immunoreactivity; phospho-p38 was also detected as early as 4 h after hypoxia and remained elevated for at least 24 h (Fig. 3). In contrast, antibodies against total p38 or total ERK1/2 showed no hypoxia-induced change in expression. Thus, both ERK and p38 were phosphoactivated by hypoxia in cortical neurons, whereas hypoxia suppressed phosphoactivation of JNK. In this study, the level of p38 in cortical neurons appeared to be much lower than that of the other two MAPKs, although it became detectable when we used Ͼ150 g of protein per lane.
Inhibition of p38 Blocks Hypoxic Down-regulation of Mdm2, as Well as Serine 15 Phosphorylation and Accumulation of p53-To investigate further which signaling pathways mediate the effects of hypoxia on Mdm2 and p53 in neurons, we used a series of kinase inhibitors to inactivate receptor tyrosine kinases, ERK, p38, JNK, PI3K, and PKC. We expected that if a particular kinase were critical for hypoxia's effects on Mdm2 and p53 signaling, inhibiting its activity should reverse these effects. As shown in Fig. 4, the Src family kinase inhibitor herbimycin A (1 M) (32) increased Mdm2 expression in hypoxic cultures, and the p38 inhibitor SB203580 (5 M) (33) prevented hypoxic down-regulation of Mdm2, whereas the nonselective tyrosine kinase inhibitor genistein (50 M) (32) was ineffective, and the PKC inhibitor GF102390X (1 M) (33), the PI3K inhibitor wortmannin (2 M) (18), and the ERK inhibitor PD98059 (20 M) (33) potentiated the effect of hypoxia on Mdm2. Genistein and SB203580 also blocked hypoxic up-regulation of p53, whereas the other inhibitors were less effective or inactive in this regard. As is also shown in Fig. 4, SB203580 and wortmannin markedly reduced hypoxia-stimulated phosphorylation of p53 on serine 15. Because only the p38 inhibitor SB203580 reversed all of the observed effects of hypoxia, downregulation of Mdm2, up-regulation of p53, and serine 15 phosphorylation of p53, hypoxic regulation of Mdm2/p53 signaling appears to be mediated through p38. To verify further that this is the case and to control for nonspecific effects, we treated cells with another p38 inhibitor (SB202190, 4 M) or with SB202474 (4 M), an inactive analog of SB203580 and SB202190. Western blots showed that the effects of SB202190 and SB203580 on Mdm2 and p53 expression were similar, whereas SB202474 was ineffective (Fig. 5A). The p38 inhibitors appeared to act by blocking hypoxia-induced changes in Mdm2 and p53 expression specifically because neither inhibitor altered expression under normoxic conditions (Fig. 5B).
p38 Activity Is Required for Hypoxia-induced Cell Death-To investigate the relationship between hypoxic induction of p38 and hypoxia-induced cell death, we first treated cells with the p38 inhibitor SB203580 (5 or 10 M), beginning 1 h prior to the onset of hypoxia. Treatment with SB203580 partially protected cortical neurons from hypoxic death, which was reduced by ϳ60% at 8 h and by ϳ40% at 16 h (Fig. 6). Next we transfected cortical neurons with expression constructs encoding constitutively active or dominant-interfering mutants of MAPK kinase 3 (MKK3), which specifically phosphorylates and activates p38 (29,34,35). Transfection efficiency was monitored using an expression construct encoding green fluorescent protein, which showed that 21 Ϯ 2% (n ϭ 8) of cells were transfected (Fig. 7A). In normoxic cultures, constitutively active MKK3(Glu) decreased cell viability by ϳ25% (Fig. 7B). In hypoxic cultures, dominant-interfering MKK3(Ala) increased viability by ϳ20%, whereas dominant active MKK3(Glu) decreased viability by p38 Mediates Hypoxic Regulation of Mdm2 and p53 in Neurons ϳ25% of control. Western blots showed that in normoxic cultures, constitutively active MKK3(Glu) decreased Mdm2 and increased p53 expression (Fig. 7C); in hypoxic cultures, these effects were enhanced, and dominant-interfering MKK3(Ala) produced opposite effects. These data suggest that activation of p38 plays an important role in hypoxia-induced changes in Mdm2 and p53 expression and associated cell death in neurons.
Knockdown of Mdm2 in Cortical Neurons Reduces Cell Viability and Elevates p53 Expression-To evaluate the functional consequences of hypoxic reduction of Mdm2 expression, we blocked Mdm2 expression with an antisense ODN (MAS20), which was transfected into cells for 24 -48 h as described previously (36); a scrambled ODN (MX20) was used as a control. As shown in Fig. 8A, blocking Mdm2 expression with MAS20 reduced cell viability by ϳ40% under normoxic conditions and by ϳ35% after hypoxia. Western blots showed that knockdown of Mdm2 by MAS20 reduced Mdm2 and increased p53 levels, as expected (Fig. 8B).
Knockdown of p53 Rescues Cortical Neurons from Hypoxic Cell Death-To assess if p53 is required for hypoxic neuronal cell death, we transfected cells with a p53 antisense ODN that overlaps the translation start site. The p53 antisense ODN did not affect cell viability under normoxic conditions, relative to cells transfected with a sense ODN or treated with transfection reagent only. In contrast, in hypoxic cultures, the p53 antisense ODN increased neuronal viability from ϳ40 to ϳ70% (Fig. 9). Because basal p53 expression was barely detectable, it was not possible to discern a further decrease in expression after antisense treatment. DISCUSSION The major finding of this study is that neuronal hypoxia transcriptionally down-regulates Mdm2 and post-transcriptionally up-regulates p53 and that both effects may be mediated through hypoxic activation of p38 MAPK. These events appear to have functional importance in regulating neuronal cell death and survival from hypoxia, because hypoxic neuronal death was reduced by p38 inhibitors, by a dominant-interfering mutant of the p38-activating kinase MKK3, and by a p53antisense ODN and was increased by a constitutively active form of p38 and by an Mdm2-antisense ODN.
The known and proposed roles of Mdm2 in regulating p53 function suggest that the effects of hypoxia on Mdm2 and p53 may be interrelated (37). First, because binding to Mdm2 promotes the proteasomal degradation of p53 through Mdm2's action as a ubiquitin ligase (36), reduction of Mdm2 expression by hypoxia could account for the observed increase in p53 levels in our hypoxic neuronal cultures. This is supported by our finding that an Mdm2-antisense ODN increased p53 abundance. Second, the observed increase in serine 15 phosphorylation of p53 would also be expected to increase p53 levels because this post-translational modification interferes with the interaction between p53 and Mdm2 that leads to p53 degradation (8). Finally, some evidence suggests that Mdm2 may be involved in the nucleocytoplasmic shuttling of p53 that translocates it from its site of action as a transcription factor to its FIG. 7. Constitutive activation or inactivation of MKK3 alters neuronal survival and Mdm2 and p53 expression. Cortical neurons were transfected for 24 h with 3 g of plasmid DNA encoding a constitutively active form of MKK3 (Glu), 8 g of plasmid DNA encoding a dominant-interfering MKK3 mutant (Ala), or 8 g of control vector and then exposed to hypoxia or normoxia for an additional 16 h. A, to measure transfection efficiency, some cultures were transfected for 24 h with a vector expressing GFP (pEGFP) and counterstained with 4,6diamidino-2-phenylindole (DAPI). Transfection efficiency was calculated as GFP-expressing cells/4,6-diamidino-2-phenylindole-stained cells ϫ 100%. B, cell viability was measured by MTT absorbance and expressed as a percentage (mean Ϯ S.E., n ϭ 3) of viability in normoxic cultures transfected with control vector (Vector). *, p Ͻ 0.05 relative to vector. C, protein samples (40 g) from whole cell extracts were loaded on 10% SDS-PAGE gels and transferred to PVDF membranes for Western analysis, and Mdm2 and p53 were detected with the antibodies described under "Experimental Procedures." Data shown are representative blots from three independent experiments per panel. p38 Mediates Hypoxic Regulation of Mdm2 and p53 in Neurons site of proteolytic breakdown (36,39); in this respect, too, reduced levels of Mdm2 are predicted to permit the enhanced nuclear accumulation of p53 that we observed. In addition to these effects on p53 protein expression, Mdm2 inhibits p53 function by binding to the NH 2 -terminal transactivation domain of p53 and acting as a negative regulator of p53-induced gene transcription (40).
Different stress signals produce distinct effects on Mdm2 expression. For example, the observed reduction in Mdm2 levels after hypoxia is consistent with the effect of UVC light irradiation of U2-OS cells, which reduces Mdm2 levels, whereas x-irradiation of the same cells increases Mdm2 levels (41). In cultured cerebellar granule neurons, Mdm2 expression is decreased when medium is depleted of K ϩ , which also leads to apoptotic cell death (30). In contrast, in adrenalectomyinduced apoptosis of hippocampal neurons, Mdm2 expression is increased (38). Therefore, Mdm2, like p53, appears to be capable of promoting either cell death or cell survival, depending on the nature of the injurious stimulus. In the present study, the observation that an Mdm2-antisense ODN reduced neuronal viability argues for a survival-promoting effect of Mdm2 in this system.
Hypoxic regulation of protein expression is mediated through several signal transduction pathways. We found evidence for increased activation of p38 MAPK, and p38 inhibitors blocked the up-regulation of p53 in our hypoxic neuronal cultures. Moreover, hypoxic induction of p53 was also reduced by a dominant-interfering form of MKK3, which activates p38. In contrast to the effect of hypoxia on p38, activation of JNK was decreased. However, JNK may still influence p53 levels in these cultures, since it has been reported that a JNK-p53 complex is found in some cells and that JNK may promote the ubiquitination of p53 under some conditions (20). Whether suppression of JNK activation by hypoxia reduces the degradation of p53 in neurons requires further study. FIG. 9. Antisense knockdown of p53 rescues neurons from hypoxic cell death. Cortical neurons were transfected for 24 h with 5 M p53 antisense (p53AS) or p53 sense (p53S) ODN as described under "Experimental Procedures" and then exposed to hypoxia or normoxia for an additional 16 h. A, cell viability was measured by MTT absorbance and expressed as a percentage (mean Ϯ S.E., n ϭ 3) of viability in untransfected cultures (No ODN). *, p Ͻ 0.05 relative to p53S. B, protein samples (40 g) from whole cell extracts were loaded on 10% SDS-PAGE gels and transferred to PVDF membranes for Western analysis, and Mdm2 and p53 were detected with the antibodies described under "Experimental Procedures." Data shown are representative blots from three independent experiments per panel.
p38 Mediates Hypoxic Regulation of Mdm2 and p53 in Neurons