Activation of p38 Mitogen-activated Protein Kinase by c-Abl-dependent and -independent Mechanisms*

From the ‡Division of Cancer Pharmacology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115, §Howard Hughes Medical Institute, Program in Molecular Medicine, Department of Biochemistry and Molecular Biology, University of Massachusetts Medical School, Worcester, Massachusetts 01605, and ¶Department of Radiation and Cellular Oncology, University of Chicago, Chicago, Illinois 60637

The mitogen-activated protein kinases (MAPKs) 1 are induced by diverse stimuli in the transduction of signals from the cell membrane to the nucleus. Three groups of MAPKs have been identified: ERK (1-3), JNK/SAPK (4 -7), and p38 MAPK (8 -10). The MAPKs are activated by phosphorylation on Thr and Tyr at Thr-X-Tyr motifs, which differ depending on the group (11). Each MAPK group has both distinct upstream activators and substrate specificities (11). For example, the transcription factor c-Jun is a substrate of the JNK/SAP kinase signaling pathway (4,5), ATF2 is the target of p38 MAPK and JNK/SAPK (6,(12)(13)(14), and Elk1 is phosphorylated by all three groups of MAPKs (12,15,16).
The JNK/SAPKs are activated by tumor necrosis factor (TNF), anisomycin, and UV light (4,5). The demonstration that Ha-Ras contributes to the induction of JNK/SAPK activity by UV indicates the involvement of signals initiated at the cell membrane (4). Other studies have demonstrated that JNK/ SAPK is activated by DNA-damaging agents. Ionizing radiation (IR) induces JNK/SAPK activity in diverse cell types (17)(18)(19). IR-induced activation of JNK/SAPK is associated with the formation of a complex with the growth factor receptor binding protein 2 adaptor protein and phosphatidylinositol 3-kinase (17). Studies in ataxia-telangiectasia cells have further demonstrated that UV and IR induce JNK/SAPK by distinct mechanisms (20). Similar responses involving JNK/SAPK activation have been demonstrated in cells treated with 1-␤-D-arabinofuranosylcytosine (ara-C), alkylating agents (mitomycin C and cisplatinum), and the topoisomerase inhibitor camptothecin (21,22). These findings have implicated diverse types of DNA damage as initial signals in the activation of JNK/SAPK. Although the precise cascade of kinases involved in DNA damageinduced JNK/SAPK activation is unclear, other studies have shown that the c-Abl protein tyrosine kinase functions upstream to JNK/SAPK in the cellular response to IR, certain alkylating agents, and ara-C (23,24). Transfection of activated forms of Abl has also been associated with stimulation of JNK/ SAPK activity (24 -26). The activation of JNK/SAPK by a c-Abldependent mechanism, however, may be limited to certain agents, since TNF-induced SAPK activity occurs in c-Abl-deficient cells (23).
The p38 MAPK pathway is activated by TNF and interleukin 1 (9,10,13). This pathway is also activated by endotoxin, osmotic stress, and heat shock (8,9). Activation of p38 MAPK is mediated by phosphorylation on Thr and Tyr by MKK3 and MKK6 (12,14,(27)(28)(29). MKK3 and MKK6 specifically activate p38 MAPK, whereas SEK1/MKK4 may activate both p38 MAPK and JNK/SAPK (14,30). Other studies have demonstrated that the Rho GTPases and multiple PAKs regulate p38 MAPK and JNK/SAPK activation (31)(32)(33). Thus, some signals appear to be capable of activating both the p38 MAPK and JNK/SAPK pathways. Since certain DNA-damaging agents activate JNK/SAPK, the present studies have addressed the potential involvement of genotoxic stress in activation of p38 MAPK. The results demonstrate that diverse DNA-damaging agents activate p38 MAPK and that this response is mediated by c-Abl-dependent and -independent mechanisms.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-Human U-937 myeloid leukemia cells (American Type Culture Collection, Rockville, MD) were grown in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM Lglutamine. NIH3T3 and Abl Ϫ/Ϫ cells (23) were grown in DMEM containing 10% fetal bovine serum and antibiotics. Abl Ϫ/Ϫ cells were reconstituted with c-Abl by retroviral transfection as described (34). Cells (1 ϫ 10 6 /100-mm culture dish) were plated 24 h before treating with ara-C (Sigma), CDDP (Sigma), and methyl methanesulfonate (MMS) (Sigma). Cells were treated with 10 or 20 gray IR at room temperature with a Gammacell 1000 (Atomic Energy of Canada, Ot-tawa, Ontario, Canada) under aerobic conditions with a 137 Cs source emitting at a fixed dose rate of 0.76 gray min Ϫ1 as determined by dosimetry. Cells were also treated with 80 J/m 2 UV (UV Stratalinker 1800, Stratagene).
Immune Complex Kinase Assays-Cells were washed with PBS and lysed in 1 ml of lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, and 10 g/ml leupeptin and aprotinin) as described (24). Lysates were incubated with anti-SAP kinase (Santa Cruz Biotechnology, Santa Cruz, CA) or anti-p38 MAPK (16) antibodies for 1 h at 4°C and then for 45 min after the addition of protein A-Sepharose. The immune complexes were washed three times with lysis buffer and once with kinase buffer and resuspended in kinase buffer containing [␥-32 P]ATP (6000 Ci/mmol; DuPont NEN) containing either GST-Jun (2-100) or GST-ATF2 (1-109). The reactions were incubated for 15 min at 30°C and were terminated by the addition of SDS sample buffer. The proteins were analyzed by 10% SDS-PAGE and autoradiography.
Immunoprecipitation and Immunoblot Analysis-Soluble proteins (100 g) were incubated with anti-Abl antibody (Santa Cruz Biotechnology) for 1 h and precipitated with protein-A-Sepharose for an additional 1 h. The resulting immune complexes were washed three times with lysis buffer, separated by electrophoresis in 7.5 or 10% SDSpolyacrylamide gels, and then transferred to nitrocellulose paper. The residual binding sites were blocked by incubating the filters with 5% dry milk in PBST (phosphate-buffered saline, 0.05% Tween 20) for 4 h at room temperature. The filters were incubated with anti-Abl for 1 h with shaking. After washing twice with PBST, the blots were incubated with anti-mouse IgG peroxidase conjugate (Amersham Corp.). The antigen-antibody complexes were visualized by chemiluminescence (ECL detection system, Amersham Corp.). Total cell lysates were also separated by electrophoresis in 10% SDS-polyacrylamide gels and then transferred to nitrocellulose paper. Immunoblot analyses were performed using anti-SAP and anti-p38 MAPK antibodies.

RESULTS AND DISCUSSION
Previous studies have demonstrated that certain classes of DNA-damaging agents, such as IR, CDDP, and ara-C, activate JNK/SAPK (17)(18)(19)(22)(23)(24). To determine whether p38 MAPK is activated by genotoxic stress, we treated NIH3T3 cells with CDDP and assayed anti-p38 MAPK immunoprecipitates for phosphorylation of GST-ATF2. The CDDP-treated cells exhibited an increase in p38 MAPK activity that was detectable at 1 h and reached maximal levels at 3 h (Fig. 1A). The CDDPinduced increases in p38 MAPK activity occurred in the absence of changes in p38 MAPK protein levels (Fig. 1A). Since p38 MAPK is activated, in part, by phosphorylation on Tyr (13), we assayed the anti-p38 MAPK immunoprecipitates for reactivity with an anti-Tyr(P) antibody. The results demonstrate increased tyrosine phosphorylation of p38 MAPK as a consequence of CDDP treatment (Fig. 1A). The results also demonstrate that activation of p38 MAPK is dependent on the concentration of CDDP (Fig. 1B).
To determine whether CDDP-induced activation of p38 MAPK is limited to NIH3T3 cells, we exposed other cell types to this agent. The results demonstrate that CDDP activates p38 MAPK in U-937 cells ( Fig. 2A). Similar findings were obtained in A549 adenocarcinoma cells (data not shown). In contrast to CDDP which induces DNA intrastrand cross-links (35), IR treatment causes DNA strand breaks (36), and ara-C terminates DNA strand elongation (37). Exposure of U-937 cells to ara-C caused activation of p38 MAPK (Fig. 2B, upper  panel). The results also demonstrate increased tyrosine phosphorylation of p38 MAPK as a consequence of ara-C treatment  (top panels, A and B). Anti-p38 MAPK immunoprecipitates were also analyzed by immunoblotting with anti-p38 (third panels, A and B) or anti-Tyr(P) (fourth panels, A and B) antibodies. The fold increase in p38 MAPK activity is shown (bottom panels, A and B) as the mean (bars, S.E.) for three independent experiments. (Fig. 2B, bottom panel). Similar results were obtained in ara-C-treated NIH3T3 cells (Fig. 2C). In contrast, there was no detectable effect of IR on p38 MAPK activity in NIH3T3 cells (data not shown), whereas this kinase was induced in IRtreated U-937 cells (Fig. 2D, upper panel). Activation of JNK/ SAPK by IR was more pronounced in U-937 cells compared to that obtained for p38 MAPK (Fig. 2D, lower panel). However, exposure of HS Sultan multiple myeloma cells to IR was associated with comparable activation of p38 MAPK and JNK/ SAPK (data not shown). These findings with IR suggest that there are cell type-specific differences in the induction of p38 MAPK. UV light induces the formation of thymine dimers and DNA strand breaks (38). Other forms of DNA damage are induced by the monofunctional alkylating agent, MMS (37).
The finding that treatment with UV light and MMS also increases p38 MAPK activity (Fig. 2E) indicates that diverse DNA-damaging agents induce this stress pathway.
The c-Abl tyrosine kinase is activated in the cellular response to IR, CDDP, and ara-C (23,24). Moreover, c-Abl is involved in activation of JNK/SAPK by these agents (23,24). Consequently, we asked whether activation of p38 MAPK occurs by a c-Abl-dependent mechanism. To address this issue, we treated cells deficient in c-Abl (Abl Ϫ/Ϫ ) (34) with CDDP. In contrast to NIH3T3 cells, there was little effect of CDDP on p38 MAPK activity in Abl Ϫ/Ϫ cells (Fig. 3A). c-Abl expression was reconstituted in Abl Ϫ/Ϫ cells (designated Abl ϩ , the level of c-Abl protein was approximately 20% of that in NIH3T3 cells) (Fig.  3A, right panel). Treatment of Abl ϩ cells with CDDP was associated with a 5-fold increase in p38 MAPK activity (Fig. 3A). We also treated Abl Ϫ/Ϫ cells with CDDP for longer periods of time to test whether p38 MAPK is activated as a late response. Anti-p38 MAPK immunoprecipitates from control and ara-C-treated cell lysates were also analyzed by immunoblotting with anti-Tyr(P) antibodies (lower panel). C, NIH3T3 cells were treated with 50 M ara-C for the indicated times. The p38 MAPK activity was measured as described above. D, U-937 cells were treated with 20 gray IR and harvested at the indicated times. Total cell lysates were immunoprecipitated with either anti-p38 MAPK or anti-SAPK antibodies. In vitro immune complex kinase reactions containing GST-ATF2 (upper panel) or GST-Jun (bottom panel) fusion proteins were analyzed by 10% SDS-PAGE and autoradiography. E, NIH3T3 cells were treated with 50 M CDDP for 3 h, 1 mM MMS for 3 h, or 80 J/m 2 UV (harvested at 30 min). The p38 MAPK activity was measured as described above. The fold activation in kinase activities is shown at the bottom.
FIG. 3. c-Abl-dependent activation of p38 MAPK in response to CDDP. A, NIH3T3, Abl Ϫ/Ϫ and Abl ϩ cells were treated with 100 M CDDP for 3 h. Total cell lysates were immunoprecipitated with anti-p38 MAPK antibody, and in vitro immune complex kinase reactions containing GST-ATF2 fusion protein were analyzed by 10% SDS-PAGE and autoradiography (left panel). The fold increase in activity is shown at the bottom. Total cell lysates from NIH3T3, Abl Ϫ/Ϫ , and Abl ϩ cells were also immunoprecipitated with anti-Abl antibody. Protein precipitates were analyzed by immunoblotting with anti-Abl (right panel). B, Abl Ϫ/Ϫ cells were treated with 100 M CDDP for the indicated times. NIH3T3 cells were treated with 100 M CDDP for 3 h as a positive control. p38 MAPK activity (upper panel) was assayed as described above. The fold increase in p38 MAPK activity is shown at the bottom panel (one of the three representative experiments is shown). C, NIH3T3 and Abl Ϫ/Ϫ cells were treated with 50 M ara-C for 3 h. Abl Ϫ/Ϫ cells were also treated with 0.7 M NaCl for 10 min. Total cell lysates were immunoprecipitated with anti-p38 MAPK antibody. p38 MAPK activity was assayed as described above.
The results demonstrate little effect of CDDP on p38 MAPK activity in Abl Ϫ/Ϫ cells during incubation for 9 h (Fig. 3B).
These results indicate that, as in the regulation of JNK/SAPK (23), certain DNA-damaging agents induce p38 MAPK by a c-Abl-dependent mechanism. We also treated Abl Ϫ/Ϫ cells with ara-C and assayed for p38 MAPK activity. The results demonstrate that, in contrast to NIH3T3 cells, ara-C-induced p38 MAPK activity is blocked in Abl Ϫ/Ϫ cells (Fig. 3C). The finding that activation of JNK/SAPK by TNF occurs through a c-Ablindependent mechanism (23) supports the hypothesis that multiple pathways regulate JNK/SAPK. Since hyperosmolarity and TNF also induce p38 MAPK, we asked whether these responses are functional in the Abl Ϫ/Ϫ cells. The results demonstrate that exposure of Abl Ϫ/Ϫ cells to 0.7 M NaCl or TNF causes p38 MAPK activation ( Fig. 3C and data not shown).
Other studies were performed on NIH3T3 and Abl Ϫ/Ϫ cells treated with MMS and UV light. In contrast to CDDP and ara-C, MMS induced p38 MAPK activity in the Abl Ϫ/Ϫ cells (Fig. 4A, left panel). Similar results were obtained with UV treatment (Fig. 4A, right panel). Since these findings indicate that MMS and UV activate p38 MAPK by a c-Abl-independent pathway, we asked whether these agents induce JNK/SAPK in the Abl Ϫ/Ϫ cells. The results demonstrate that MMS activate JNK/SAPK by a c-Abl-independent mechanism, whereas activation of this kinase in response to UV is partially blocked in Abl Ϫ/Ϫ cells (Fig. 4B).
The present results demonstrate that diverse types of DNAdamaging agents induce p38 MAPK activity. The findings also demonstrate that CDDP and ara-C induce p38 MAPK by a c-Abl-dependent mechanism. By contrast, MMS and UV induce p38 MAPK in the Abl Ϫ/Ϫ cells. Previous work has shown that TNF-induced activation of JNK/SAPK occurs by a c-Abl-independent mechanism (23). Activation of p38 MAPK by TNF, as well as hyperosmolarity, similarly occurs independently of c-Abl activation. The results further indicate that UV light induces p38 MAPK by mechanisms not involving c-Abl. The UV response in mammalian cells is initiated in an extranuclear compartment, and UV-induced activation of JNK/SAPK occurs in enucleated cells (39). The finding that Ha-Ras contributes to the induction of JNK/SAPK by UV is also consistent with the hypothesis that the UV response is initiated in the cytoplasm.
Other studies have demonstrated that UV activates JNK/ SAPK by damaging DNA (38,40). Activation of JNK/SAPK in response to UV may, therefore, be regulated by at least two different mechanisms. In this context, our results demonstrate that UV-induced JNK/SAPK activity is mediated, at least in part, by a c-Abl-dependent mechanism.
Finally, MMS is a monofunctional alkylating agent that alkylates DNA and damages membrane proteins (41,42). Therefore, MMS may also activate JNK/SAPK and p38 MAPK by DNA damage-independent mechanisms. Indeed, MMS induces JNK/SAPK and p38 MAPK activity in Abl Ϫ/Ϫ cells, perhaps by signals activated at the cell membrane. Taken together, these results demonstrate that activation of the stress response to diverse agents can be distinguished by c-Abl-dependent and -independent mechanisms.