Selective Activation of p38 MAPK Cascade and Mitotic Arrest Caused by Low Level Oxidative Stress*

Apoptosis induced by high level oxidative stress accompanies diverse cellular biochemical events including activation of the stress signal cascades of JNK and NF-κB. We report here selective activation of p38 MAPK cascade and mitotic arrest under a low level oxidative stress that lacks apoptosis induction. U937 human lymphoid cells treated with low dose (0.02 mm) H2O2 rapidly caused p38 MAPK cascade activation detectable by phosphorylation of MKK3/6, p38 MAPK, activating transcription factor-2, and cAMP-responsive element-binding protein, leaving the JNK and NF-κB cascades unaffected. The p38 kinase activation was sustained for 24 h under the low level stress conditions and led to formation of polyploid nuclei.N-Acetyl-l-cysteine, a precursor of anti-oxidant glutathione, canceled both p38 MAPK activation and abnormal cell cycle progression, whereas blockage of the kinase by specific inhibitor SB203580 allowed the appearance of apoptotic cells. Thus, mimicking the effects of nocodazole, the low level oxidative stimulus caused inhibition of cell division in the M phase through p38 MAPK activation. The kinase cascade may serve as a primary transducer of cytoplasmic oxidative signals to nucleus for stress-relieving gene expression and cell cycle control before apoptosis-inducing signals are transduced. This is the first report demonstrating that oxidative stress can participate in cell cycle control by induction of a signal cascade.

such oxidative stimuli occurring in cytoplasm may be transmitted through cellular signal transduction pathways to the nucleus for the control of cell division and survival. Various experiments provide evidence that many protein kinases and transcription regulatory factors are activated under the oxidative stress conditions (8 -13). However, because the role of each signal cascade has not been critically determined, those results are considered to include diverse cellular events in various phases of apoptosis or survival processes. In this study, U937 human lymphoid cells exposed to different concentrations of H 2 O 2 in combination with inhibitors were examined for activation of stress-responsive protein kinase cascades so that survival mechanism-related responses could be distinguished from apoptosis-related ones. This report shows the presence of a low level oxidative stress that selectively induces the p38 MAPK cascade without accompanying apoptosis. More significantly, results indicate that p38 MAPK activation in response to the low level oxidative stress leads to abnormal M phase transition in the cell cycle. Oxidative stress may be a biological cause of the p38 MAPK-mediated spindle check point arrest that has been demonstrated by microinjection of active p38 MAPK or by nocodazole treatment in Xenopus embryos and mouse cells (14). Detection of p38 MAPK-activated nuclear factors ATF-2 and CREB (ATF-1) evidenced that the moderate oxidative stimulus was transmitted as a signal to nucleus.

EXPERIMENTAL PROCEDURES
Cell Culture and H 2 O 2 Treatment-U937 cells were maintained with RPMI 1640 containing 10% fetal bovine serum and antibiotics in an atmosphere with 5% CO 2 . Before the H 2 O 2 treatment, cells (5 ϫ 10 5 ) were equilibrated with fresh medium (5 ml) consisting of RPMI 1640 and 2% serum for 2 h. After adding H 2 O 2 to make 0.02 mM or 0.2 mM, cells were harvested at time intervals and washed in a buffer (0.15 M NaCl, 10 mM Tris-HCl, pH 7.4, 2 mM EDTA, and 2 mM EGTA). For the experiment with NAC or SB203580, the compounds were added to the cultures 1 h before the cells were exposed to H 2 O 2 . SB203580 was dissolved in dimethyl sulfoxide at 10 mM and diluted with the medium to 10 M. Nocodazole stock solution (5 mM) in dimethyl sulfoxide was applied to the culture at a concentration of 5 nM.
Western Blot-Washed cells (10 6 ) were lysed in the sample buffer (100 l) for sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immediately boiled for 2 min. DNA was sheared by sonication. A 10-l aliquot was loaded in each slot. Biotinylated protein standards (Bio-Rad) were used as size markers. After SDS-polyacrylamide gel electrophoresis, proteins were transferred to polyvinylidene difluoride membrane by electrophoresis. Membrane was blocked with 5% nonfat dry milk in TBST (0.15 M NaCl, 10 mM Tris-HCl, and 0.1% Tween 20 at pH 7.4) and incubated with an antibody at a concentration of 100 -200 ng/ml IgG in the blocking buffer. For detection of the phosphorylation status of p38 MAPK, MKK3/6, CREB/ATF1, ATF2, JNK, c-Jun, or MKK4, an antiserum reactive with the phosphorylated form of each protein (phosphospecific antibody) was purchased from Bio-Rad. Analyses of the IB-␣, NF-B p65, and p38 MAPK protein composition were performed with antisera from Santa Cruz (IB-␣ (CϪ21), NF-B (A), and p38 MAPK (H-147)). The membrane was washed in TBST with brief sonication. By using alkaline phosphatase-conjugated secondary antibodies and the CDP-star chemiluminescent system (Bio-Rad), protein bands were visualized.
Nuclear Staining-Cells were fixed in glutaraldehyde (1.25% in phosphate-buffered saline) for 16 h at 4°C, and then the glutaraldehyde was washed away with phosphate-buffered saline. Cell suspension was combined with Hoechst 33342 solution (6 mM) at 5:1 and observed under fluorescence microscopy.
Flow Cytometry-Cells fixed in 70% ethanol were stained with propidium iodide (PI) (50 g/ml) after RNA digestion. PI-stained 10,000 cells were analyzed for DNA content with a Becton Dickinson FACScan flow cytometer. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Analysis of the Kinase Cascades Responsive to Oxidative
Stresses-U937 cell lysates prepared after 0-, 10-, 30-, 60-, and 120-min incubation with 0.02 or 0.2 mM H 2 O 2 were examined for activation of p38 MAPK, JNK, and IB kinase signal transduction pathways by Western blotting. p38 MAPK activation was analyzed with an antiserum reactive with the protein doubly phosphorylated on Thr-180 and Tyr-182 (12). As early as 10 min after exposure to 0.02 or 0.2 mM H 2 O 2 , p38 MAPK phosphorylation was substantially enhanced (Fig. 1a, i) without alteration of the p38 MAPK protein amount (Fig. 1b). The phosphorylation was retained at least for 120 min. NAC, an anti-oxidant activated in a cytoplasmic redox enzyme system (8), completely blocked it (Fig. 1c), indicating that an intracellular oxidative stimulus was generated to induce p38 MAPK phosphorylation by incubation with 0.02 mM H 2 O 2 . However, JNK-1 and -2 phosphorylated on the Thr-183 and Tyr-185 residues (15) were undetectable even after 2-h incubation with 0.02 mM H 2 O 2 (Fig. 1d). This was in contrast to 0.2 mM H 2 O 2treated cells in which phosphorylation of the 46-kDa JNK-1 and 54-kDa JNK-2 (16) was apparent after 10 min of incubation (Fig. 1j). The band intensity of IB-␣ was unaltered in the 0.02 mM H 2 O 2 -treated cells (Fig. 1e), indicating the absence of protein degradation. NF-B p65 was not detectable in their nuclear extracts for the 2-h period (Fig. 1f). However, in the 0.2 mM H 2 O 2 -treated whole cell lysates, IB-␣ protein diminished ( Fig. 1k) with the concomitant appearance of NF-B p65 in the nuclear fractions (Fig. 1l), indicating occurrence of NF-B nuclear translocation through IB degradation (17).
Thus, cells responded to the stress of 0.02 mM H 2 O 2 by activation of p38 MAPK, without affecting either JNK or NF-B cascade. In contrast, all of the three stress-responsive protein kinase cascades were activated by the 0.2 mM H 2 O 2 treatment.
Thr-223 phosphorylation of MKK4 (also termed SEK1), for which both p38 MAPK and JNK serve as the substrate (18), was not evident with 0.02 mM H 2 O 2 ( Fig. 1g) (Fig. 1m). In contrast, immunoblot with an antiserum reactive with the activated forms of MKK3 (with phosphorylation on Ser-189 and Thr-193) and MKK6 (on Ser-207 and Thr-211) showed a positive sign of the 35-37-kDa band after the 0.02 mM H 2 O 2 treatment (Fig. 1h). At least one of the two p38 MAPK-specific activators, MKK3 (18) and MKK6 (19), seemed to be functional under these conditions. More enhanced MKK3/6 phosphorylation was observed with 0.2 mM H 2 O 2 (Fig. 1n). Either MKK3 or MKK6, or both, may mediate the selective activation of p38 MAPK even when the oxidative stress is not strong enough to activate MKK4.
Activation of Transcription Regulatory Factors under Low Level Oxidative Stress-After 10-min incubation with 0.02 mM H 2 O 2 , phosphorylation of CREB and ATF-1 on Ser-133 increased approximately 2-fold, and it decreased to the control level in 60 min (Fig. 2a). The enhancement did not occur in the presence of NAC (Fig. 2b). The effect of low level oxidative stress on these factors was found less predominant. In contrast, the level of ATF-2 Thr-71 phosphorylation was substantially raised after 10 min and was retained without a decline at least for 120 min (Fig. 2c). This induction was blocked either by NAC (Fig. 2d) or by 10 M p38 MAPK-specific inhibitor SB203580 (Fig. 2e), indicating that the oxidative signal was immediately relayed to nuclear factor ATF-2 by p38 MAPK activation. On the other hand, Ser-73 phosphorylation of c-Jun was undetectable at any time point in 0.02 mM H 2 O 2 -treated cells (Fig. 2f) in agreement with the absence of JNK phosphorylation under the same conditions (cf. Fig. 1d). In 0.2 mM H 2 O 2 -treated cells, however, c-Jun phosphorylation appeared at 10 min (Fig. 2g) and reached a maximum at 60 min, corresponding to the JNK activation time course observed in Fig. 1j.
Analyses of the phosphorylation status of CREB (ATF-1), ATF-2, and c-Jun confirmed that the oxidative stimulus caused by 0.02 mM H 2 O 2 was transmitted to the nucleus by the selective activation of p38 MAPK. The moderate, temporary phosphorylation of CREB (ATF-1) may reflect the activity of mitogen-activated protein kinase-activated protein kinase-2, a cytoplasmic substrate of p38 MAPK (20). The strong induction and retention of ATF-2 phosphorylation implicate that the p38 MAPK cascade functions in activation of ATF-2-regulated gene expression under the low level oxidative stress.
Abnormal Cell Cycle Progression under the Low Level Oxidative Stress-The p38 MAPK activation continued for 24 h in When cultured with 5 ng/ml nocodazole, a compound that blocks tubulin assembly, the level of p38 MAPK phosphorylation gradually rose (Fig. 3b) as reported (14). Hoechst 33342 staining revealed an abundance of enlarged nuclei in the nocodazole-treated culture at 24 h. On the basis of the flow cytometric analysis, 10.5% cells had Ͼ4 N DNA, and the Ն4 N/2 N ratio was 0.85. In this cell line, the inhibitor did not induce typical M phase arrest but caused S/M uncoupling. Neither JNK phosphorylation (data not shown) nor apoptosis induction was obvious with nocodazole at the concentration above.
As demonstrated by an enzymological analysis (22), SB203580 blocked not only ATF-2 phosphorylation (cf. Fig. 2e) but also the p38 MAPK phosphorylation itself under the stress of 0.02 mM H 2 O 2 (Fig. 3c). The p38 kinase inhibition continued for 24 h and abolished polyploid cell formation as found by flow cytometry. Moreover, nuclear fragmentation was observed in 8.73% of cells after the 24-h culture with the inhibitor and 0.02 mM H 2 O 2 . The blockage of p38 MAPK cascade might have interfered with an anti-oxidative cellular defense mechanism, which led to the onset of apoptosis in the small fraction of the culture.
NAC completely canceled both p38 MAPK cascade activation and polyploid cell formation (Fig. 3d) without inducing apoptosis. Furthermore, in the experiment where 0.02 mM H 2 O 2 was washed away from the culture after the first 2-h incubation period, the level of p38 MAPK phosphorylation fell to the basal level by 3 h from the H 2 O 2 removal. Those cells, examined at the end of the total incubation period of 24 h, were indistinguishable from the control cells either by flow cytometry or Hoechst staining (Fig. 3, e and f). Thus, the kinase activation was found to be a reversible response, and its prolonged activation was required for polyploid cell formation. Under the high level oxidative stress conditions caused by 0.2 mM H 2 O 2 , occurrence of nuclear fragmentation was detectable as early as at 5 h. Approximately 95% of cells displayed features of apoptosis including nuclear fragmentation and membrane blebbing (Fig. 3g). DISCUSSION Although other various signal pathways have been left unexamined, it is evident that p38 MAPK is activated selectively among the stress-responsive signal cascades under the low level oxidative stress conditions with 0.02 mM H 2 O 2 . Furthermore, the result with p38-specific inhibitor SB203850 suggests that p38 MAPK bears an important function for cell survival.
Oxidative stresses at different levels seem to join different streams of the cellular signal transduction system reaching to specific nuclear factors (23), although the mechanism remains to be determined. MKK3/6 may at least in part contribute to the selective p38 activation under the low level oxidative stress. MKK4 (SEK1)-JNK activation coincides with IB degradation in response to 0.2 mM H 2 O 2 , suggesting that mitogenactivated protein kinase/extracellular signal-regulated kinase kinase kinase 1 (24,25) is possibly involved in transmission of the high level, apoptosis-inducible oxidative stimulus to the IB kinase and JNK pathways. Given that the normal function of mitochondria inevitably produces ROS, the results in this study permit a speculation that p38 MAPK serves as a signal transducer of the cytoplasmic low level ROS for balancing the redox status. Activated ATF-2 and CREB may enhance transcription of genes harboring their binding sites in the promoter region, such as the manganese superoxide dismutase gene (26).
This study provides a biological interpretation for the mechanistically demonstrated M phase arrest dependent on p38 MAPK activation in Xenopus embryos and NIH3T3 cells (14). When a low level oxidative stress occurs in cytoplasm, it could induce the cellular reactions. Unlike those cells, however, U937 cells lack functional p53 (27), which controls the spindle assembly checkpoint (28) as well as the G 1 /S transition. It may explain why S/M uncoupling rather than M phase cell cycle arrest was observed in U937 cells. The p38 kinase activation was found reversible, suggesting that it could function as a sensitive, flexible mediator of oxidative signals. Cells possibly gain time by cell cycle retardation to escape from the oxidative conditions before the stress grows to cause a serious damage in chromosome or cell functions.