Phosphorylation and Stabilization of Topoisomerase IIα Protein by p38γ Mitogen-activated Protein Kinase Sensitize Breast Cancer Cells to Its Poisons*

Cancer drugs suppress tumor cell growth by inhibiting specific cellular targets. However, most drugs also activate several cellular nonspecific stress pathways, and the implications of these off-target effects are mostly unknown. Here, we report that p38γ, but not p38α, MAPK is specifically activated by treatment of breast cancer cells with topoisomerase II (Topo II) drugs, whereas paclitaxel (Taxol) does not have this effect. The activated p38γ in turn phosphorylates and stabilizes Topo IIα protein, and this enhances the growth inhibition by Topo II drugs. Moreover, p38γ activity was shown to be necessary and sufficient for Topo IIα expression, the drug-p38γ-Topo IIα axis is only detected in intrinsically sensitive but not resistant cells, and p38γ is co-overexpressed with Topo IIα protein in primary breast cancers. These results reveal a new paradigm in which p38γ actively regulates the drug-Topo IIα signal transduction, and this may be exploited to increase the therapeutic activity of Topo II drugs.

Current cancer chemotherapeutics are thought to inhibit tumor growth by acting on specific cellular targets. Most drugs, however, also have some off-target effects whose implications for drug effectiveness remain mostly unknown (1). DNA topoisomerase II␣ (Topo II␣ or Topo II) 2 is a nuclear protein critical for DNA topology, and drugs that directly target Topo II represent an important class of anti-tumor agents (2,3). Although an incubation of cancer cells with a Topo II drugs stimulates several cellular signaling cascades (2)(3)(4), how these signaling "by-products" impact the drug-Topo II␣ interaction has not been demonstrated (5). Identification of the mechanisms that regulate these signaling events and their effects on the drug-Topo II signal transduction may contribute significantly to novel cancer therapeutic developments.
Stress MAPKs (mitogen-activated protein kinases), including JNKs (C-terminal c-Jun kinases) and p38s, play a critical role in transduction and conversion of extracellular signals into various biological responses (6). Although these kinases are readily activated by various stimuli, evidence about their specific roles in stress response are just emerging (7). p38␥ is a member of the p38 MAPK family and can be activated by both stress and mitogenic signals (8). In contrast to the tumor suppressor role of the p38␣ isoform, p38␥ expression is induced by Ras oncogene, and p38␥ in turn functions to promote the Ras transforming and invasive activity (9 -12). Moreover, p38␥ expression is increased in primary tumor tissues and is required for malignant growth (9,13). Although p38␥ and p38␣ are similarly activated by stress signals (11,14), some studies showed that p38␥ can specifically mediate certain stress responses such as ␥-radiation (15,16). Because Ras activates the p38 pathway (9,10,17) and induces Topo II␣ gene expression (18,19) and because Topo II drugs can lead to the activation of p38 MAPKs (20), we tested the hypothesis that p38 MAPKs may regulate signal transduction between Topo II␣ and its poisons. Our results show that p38␥, but not p38␣, is specifically activated by Topo II drugs. In turn, the activated p38␥ phosphorylates Topo II␣ at Ser-1524, and this increases its stability leading to an increased growth inhibition. Moreover, this drug-kinase-target axis only exists in intrinsically Topo II drug-sensitive cells, and there is a positive correlation between the levels of p38␥ and Topo II␣ proteins in primary tumors. These results reveal a new paradigm in which the intrinsic sensitivity to Topo II drugs does not depend on levels of Topo II␣ protein, but on the activity of p38␥. Enhancing p38␥ MAPK expression and/or activity may therefore prove useful for improving the anti-tumor effects of Topo II drugs.
Cell Culture, Reagents, and Tet-on Inducible System-MEM and other reagents for cell culture were purchased from Invitrogen. A rabbit antibody against Ser-1524-phosphorylated Topo II␣ was generated as described previously (21). p38␥ and p38␣ antibodies were purchased from Santa Cruz Biotechnology or R&D Systems, whereas p-p38 and p-ERK antibodies were from Cell Signaling. A mouse monoclonal antibody against FLAG (M2) was purchased from Sigma. The Topo II␣ antibody and Topo II assay kit were from TopoGen. Etoposide (VP16), amasacrine (AMSA), paclitaxel (Taxol), cycloheximide, and MG132 were all purchased from Sigma. Human breast cancer cell lines (estrogen receptor (ER)-positive, MCF-7 and T47D; ER-negative, MDA-MB-231 (231) and MDA-MB-468 (468)) were obtained from ATCC and maintained in MEM containing 10% FBS and antibiotics at 37°C with 5% CO 2. Early passages of embryonic fibroblasts from p38␥ ϩ/ϩ and p38␥ Ϫ/Ϫ mice were provided by Ana Cuenda (23) and were used to establish Ras-transformed sublines (12). The Tet-on expression system (T-Rex) was purchased from Invitrogen and used to express CA p38␣ and CA p38␥ (10).
Transfection, Viral Infection, and Colony Formation-Transient transfection of 293T cells was performed by calcium phosphate. To deplete p38␥, lentiviral shLuc and shp38␥ were transfected into packaging cells, and supernatants were then collected for infecting target cells followed by an antibiotic selection (13). To overexpress p38␥ in T47D breast cancer cells, adenoviral-mediated gene delivery was used as described previously (10). For colony formation assays, cells were treated with various drugs as indicated for 24 h and plated in a 6-well plate, and colony formed were then stained and annually counted about 2 weeks later as we described previously (24).
Topo II␣ Phosphorylation Assays in Vitro and in Vivo-FLAG-tagged WT and MT Topo II␣ were expressed in 293T cells, and the expressed proteins were purified from cell lysates using a FLAG antibody. FLAG precipitates were then incubated with His-p38␥ or BSA as a control. The in vitro kinase assay was performed as described previously (17). Phosphorylated Topo II␣ proteins were separated on a SDS-PAGE and detected with an antibody specific for phosphorylated Ser-1524 Topo II␣ antibody. For in vivo Topo II␣ phosphorylation analysis, FLAG-Topo II␣ cDNA was co-transfected with CA p38␥ into 293T cells and Ser(P)-1524-Topo II␣ was assessed by Western blotting. Additional methods are described in supplemental Experimental Procedures.
Statistical Analysis-Results of colony formation were analyzed by Student's t test, and studies for a correlation of increased p38␥ and Topo II␣ protein expression in primary tumor tissues were assessed by the 2 test.

Increased Topo II Drug Sensitivity Correlates with Sustained
Levels of Topo II␣-To search for a signaling pathway involved in the activity of Topo II drugs, a group of human breast cancer cell lines were treated with two clinically used Topo II drugs, VP16 (etoposide) and AMSA (amsacrine), and cell growth was assessed by colony formation compared with paclitaxel (Taxol), an important cancer drug that does not inhibit Topo II␣. Results in Fig. 1A show that ER-negative 231 and 468 breast cancer cells are more sensitive to both Topo II drugs than are the ER-positive counterparts MCF-7 and T47D. However, no such difference was observed with paclitaxel. In contrast to other reports (18,25), the increased sensitivity does not correlate with a higher level of endogenous nuclear Topo II␣ expression (Fig. 1B). Of interest, Topo II␣ protein was decreased in the ER-positive T47D and MCF-7 cells that were treated with both Topo II drugs (26,27), but there was no decrease noted in the ER-negative 231 and 468 cells (Fig. 1C). Paclitaxel did not significantly affect the Topo II␣ levels in any of the cell lines (Fig.  1C). Given that the cell lines 231 and 468 that are sensitive to Topo II drugs did not show a decrease in Topo II␣ when treated ). B, cell fractionation was performed as described (10), and aliquots (50 g of protein) were analyzed by Western blotting. C, cells were treated with solvent control or drugs for 6 h (AMSA, 5 M; VP16, 10 M; and paclitaxel, 5 M) and analyzed by Western blotting. The results shown are representative of two separate experiments (*, the numbers across the top indicate the levels of Topo II␣ protein relative to the solvent control after normalization to ␣-actinin, as determined by densitometric analysis using NIH ImageJ software).

p38␥ Regulates Drug-Target Signal Transduction
with VP16 or AMSA, there may be a signal transduction pathway in these cells that specifically regulates the drug-Topo II interaction thereby preventing Topo II␣ depletion.
p38␥ MAPK Is Specifically Activated by Topo II Drugs in Intrinsically Sensitive but Not Resistant Cells-Stress MAPKs are major signaling cascades that are activated by a number of cancer therapeutics (28). Western blot analysis showed a consistently stronger phosphorylation of p38, but not JNK, in sensitive cells treated with Topo II drugs (Fig. 1C). Of the four p38 MAPKs (␣, ␤, ␥, and ␦), p38␣ and p38␥ are predominant forms expressed in breast cancer cells with levels of p38␥ expression higher in ER-negative breast cancer cells (10, 29) (supplemental Fig. S1A). To determine which of these two p38 proteins is responsive to Topo II drugs, total phosphorylated p38 (p-p38) proteins were isolated with a p-p38 specific antibody (reacting with all p-p38 proteins), and resulting precipitates were examined by Western blotting using p38␣ and p38␥ isoform-specific antibodies. Of great interest, results in Fig. 2 showed that p38␥ is specifically activated by AMSA and VP16 (but not by paclitaxel) in the intrinsically sensitive 468 and 231 cell lines (Fig. 2). However, p38␥ activation is not seen in the resistant T47D and MCF-7 cell lines in which p-p38␣ is predominantly elevated instead. Although a selective p38␥ activation (with and without p38␣) was previously observed in response to certain cancer therapeutics (30) and hypoxia (31), our results may be the first that demonstrate the coupling of a specific p38␥ activation with an increased growth inhibition by Topo II poisons. The lack of p38␥ activation in ER-positive cells may be due to a decreased p38␥ expression (supplemental Fig. S1A) and/or the antagonistic activity of ER protein (10). Together with the sustained levels of Topo II␣ and an enhanced growth inhibition by the Topo II drugs in 231 and 468 cells, these results indicate that p38␥ may have a role in preventing loss of the Topo II␣ protein. Sustained levels of Topo II␣ in these cells could increase their sensitivity to the Topo II drugs.
p38␥ MAPK Regulates Topo II␣ Expression and Growth Inhibition of Topo II Drugs-To investigate whether p38␥ plays a role in regulating Topo II␣ expression in response to Topo II poisons, a CA p38␥ (a MKK6-p38␥ fusion protein) (11) was expressed using a tetracycline-inducible system (Tet-on) in MCF-7 cells, and its effect on Topo II␣ expression and sensitivity of the cells to AMSA/VP16 was analyzed. Results in Fig. 3, A and B, show that the overexpression of p38␥ prevents the loss of FIGURE 2. p38␥ is specifically phosphorylated in response to Topo II drugs in intrinsically sensitive cells. Cells were treated with the indicated drugs for 6 h as described in Fig. 1C and subjected to p-p38 immunoprecipitation (IP) and Western blot analyses. A portion of cell lysates was also analyzed by direct Western blotting (Input) (*, the numbers indicate the relative levels of p38␥ protein detected in the p-p38 precipitates relative to the solvent control. The 0 values for T47D and MCF-7 cells indicate that p38␥ was undetectable in these immunoprecipitates).  Topo II␣ and increases the sensitivity of MCF-7 cells to VP16 and AMSA. This effect appears to be p38␥-specific because the overexpression of CA p38␣ using the Tet-on system does not cause these effects (supplemental Fig. S1, B and C). A complete prevention of the Topo II␣ depletion and consequently a more substantial increase in the sensitivity to these drugs were observed in T47D cells in which p38␥ was overexpressed by adenovirus-mediated infection (supplemental Fig. S2, A and B).
To assess the role of endogenous p38␥, 468 cells were infected with control (shLuc) and shp38␥-containing lentivirus (13) and analyzed for protein expression and sensitivity to Topo II drugs. Results in Fig. 3C showed that depletion of p38␥ by two separate shRNAs decreases Topo II␣ protein expression and confers a resistance to AMSA and VP16. Similar results were also obtained in 231 cells (supplemental Fig. S2C). Together, these results indicate a critical role of p38␥ in maintaining Topo II␣ protein expression and sensitivity to Topo II drugs.
p38␥ MAPK Phosphorylates Topo II␣ at Ser-1524, and This Is Required for the Topo II␣ Stability and Sensitivity to Topo II Drugs-Topo II␣ protein is phosphorylated at several residues including Ser-1524 (32) and others (33,34). Because the Ser-1524 phosphorylation (21) and Topo II inhibitors (35) both act at the decatenation checkpoint, we examined whether p38␥ phosphorylates Topo II␣ at this residue and thereby regulates the drug-target interaction. Topo II␣ phosphorylation was first assessed in vitro by incubating FLAG antibody-isolated WT and Ser-1524 MT (S1524/A) Topo II␣ proteins (expressed in 293T cells) with bacterially expressed His-tagged p38␥, followed by analysis with a specific antibody against Ser-1524phosphorylated Topo II␣ (21). Results in Fig. 4A showed that p38␥, but not p38␣, increases the p-Topo II␣ signal when the WT protein was used as a substrate (left). The co-expression of FLAG-Topo II␣ and CA p38␥ in 293T cells also resulted in enhanced phosphorylation of Topo II␣ (Fig. 4A, right). Of interest, the increased p-Topo II␣ signal in 293T cells couples with an elevated Topo II␣ protein expression (Fig. 4A, right), suggesting that the phosphorylation of Topo II␣ leads to increased expression and/or decreased degradation. To demonstrate whether the Ser-1524 is required for p38␥ binding Topo II␣, the FLAG-tagged WT and MT enzymes were co-expressed with the CA-p38␥, and FLAG precipitates were analyzed. Results in Fig. 4B (left) showed that although there is a Ser-1524-independent complex formation of Topo II␣ with p38␥, the overexpression of p38␥ with the WT Topo II␣ enhances the Topo II levels, whereas the p38␥ overexpression decreases the levels of MT (S1524A) Topo II␣. This suggests that Ser-1524 in Topo II␣ is important for increased Topo II␣ expression by p38␥. Moreover, p38␥ overexpression and depletion lead to an increased and decreased Topo II␣ protein stability in breast cancer cells, respectively (Fig. 4B, right). Furthermore, WT Topo II␣ is much more stable than the Ser-1524 MT protein in 293T cells (Fig. 4C). Additional experiments showed that p38␥ increases Topo II␣ protein stability, leading to its decreased degradation by the proteasome pathway (supplemental Fig.  S2D, upper). Together, these results indicate that the p38␥ phosphorylation of Topo II␣ at Ser-1524 enhances Topo II␣ protein stability.
To examine whether the p38␥ activation affects Topo II␣ catalytic activity, nuclear proteins were prepared from Tet-on CA p38␥ MCF-7 cells and assessed for the DNA decatenation activity (36). Results in Fig. 4D (upper left) show that the nuclear DNA decatenation activity is higher in MCF-7 cells in which p38␥ is overexpressed (ϩTet) than those in which p38␥ is not expressed (ϪTet). These differences likely result from increased Topo II␣ protein levels which are associated with p38␥ overexpression. Moreover, levels of Topo II␣-mediated DNA cleavage were also increased by CA p38␥ in response to both VP16 and AMSA (Fig. 4D, lower left). To show whether the , co-transfection with p38␥ increases the WT but decreases the MT Topo II␣ protein levels in a manner that is independent of their binding activity (left), and p38␥ expression enhances, whereas its depletion reduces, the endogenous Topo II␣ stability in breast cancer cells (right). C, the WT Topo II␣ is more stable than its S1524A mutant in 293T cells. D, CA p38␥ increases Topo II␣ catalytic activity and Topo II␣-DNA cleavage complexes induced by VP16 and AMSA, and the Ser-1524 residue of Topo II␣ plays a role in the drug sensitivity. Nuclear proteins from Tet-on CA p38␥ MCF-7 cells were prepared and assessed for the DNA decatenation activity using a Topo-Gen kit (upper left). To measure Topo II␣-DNA covalent complexes, cells were cultured overnight with and without Tet then treated with 100 M VP16 or AMSA or solvent for 30 min and assessed for Topo II␣-mediated DNA cleavage using a TopoGen ICE bioassay kit as described (48,49). Results of fractions 6 and 7 are shown (lower left). For the effects on growth, cells transfected with pcDNA3 or MT Topo II␣ were incubated with and without VP16/AMSA and assessed for colony formation (right, mean Ϯ S.D. (error bars), n ϭ 3; *, p Ͻ 0.05 versus the MT transfected cells, with the right blot showing the expressed FLAG-MT-Topo II␣ protein).

p38␥ Regulates Drug-Target Signal Transduction
Ser-1524 residue of Topo II␣ contributes to the cellular sensitivity to Topo II drugs, the MT S1524A Topo II␣ (21) was stably transfected in MCF-7 cells, and G418-resistant cells were pooled and analyzed for VP16/AMSA-induced growth inhibition. The WT Topo II␣ transfection failed to yield a stable clone, likely as a result of cell death as reported previously (37). Results in Fig. 4D (right) show that stable expression of the S1524A mutant significantly decreases the sensitivity to both drugs, which may be mediated through enhanced degradation of Topo II␣ protein (Fig. 4B) and/or an inhibition of the endogenous enzyme. Together, these results indicate that Topo II␣ may be a natural substrate for p38␥ MAPK, and the resultant phosphorylation of Ser-1524 plays an important role in maintaining Topo II␣ protein stability and activity, thereby enhancing Topo II drug sensitivity.
Topo II␣/Ser-1524 Is Phosphorylated in Intrinsically Sensitive but Not Resistant Cells, and p38␥ Activity Is Required for Ser-1524 Phosphorylation-Results in Fig. 1 showed that 231 and 468 cells are more sensitive to Topo II poisons than their ERpositive counterparts. If Ser-1524 phosphorylation plays a role in the drug sensitivity, there should be an increased p-Topo II␣ expression in sensitive cells. To this end, normal MCF-7 cells were incubated with different drugs as described in Fig. 2, and endogenous Topo II␣ was isolated with a specific antibody and examined by Western blotting. Results in Fig. 5A show that precipitated Topo II␣ is decreased after treatment of the cells with VP16 and AMSA but not with paclitaxel. Phosphorylated Topo II␣/Ser-1524 was not detected in these cells, however, regardless of pre-treatment (Fig. 5A). Moreover, there was no detectable p38␥ protein in these precipitates (Fig. 5A). Similar results were observed in T47D cells (data not shown). There is, however, a complex formation between Tet-inducible CA p38␥ and endogenous Topo II␣, which couples with increased p-Topo II␣ levels in the MCF-7 cells overexpressing p38␥ (supplemental Fig. S2D, lower). Therefore, the lack of Topo II␣ phosphorylation and p38␥ binding in normal MCF-7 cells (Fig.  5A) is likely due to very low endogenous p38␥ protein expression. In sensitive 231 and 468 cells, on the other hand, precipitated Topo II␣ proteins were phosphorylated at Ser-1524 whether or not the cells were pretreated with VP16 in which endogenous p38␥ proteins were also present (Fig. 5B). These results, together with the increased p38␥ levels in the sensitive over the resistant cells (supplemental Fig. S1A), indicate that the intrinsic p38␥ activity may play a critical role in the Topo II␣ phosphorylation. To demonstrate whether p38␥ activity is required for p-Topo II␣/Ser-1524 expression, a nonphosphorable p38␥ mutant (p38␥/AGF, dominant negative) was overexpressed in 231 and 468 cells by infection with adenoviral constructs. These cells were then treated with or without VP16, and the Topo II␣ precipitates were analyzed by Western blotting. The results in Fig. 5B show that expression of DN p38␥ suppresses the p-Topo II␣ levels in both lines relative to the ␤-gal control, indicating that the endogenous p38␥ is required for the phosphorylation of Topo II␣ expression. These results, together with the stimulation of p-Topo II␣ levels by CA p38␥ shown in supplemental Fig. S2D, lower, reveal that p38␥ is necessary and sufficient to support the phosphorylation of Topo II␣/Ser-1524.
The ability of Topo II drugs to induce p38␥ (but not p38␣) activity and the resulting phosphorylation of Topo II␣ at Ser-1524 suggest that p38␥ may actively be involved in signal transduction between Topo II␣ and its poisons. To demonstrate whether Topo II drugs directly regulate its levels through Ser-1524, the WT and MT Topo II␣ proteins were transiently expressed in 293T cells. The effects of Topo II drugs versus paclitaxel treatment on the exogenous Topo II␣ protein expression were examined. Treatment with Topo II drugs increased the levels of ectopically expressed WT but not MT Topo II␣ protein (Fig. 5C). In contrast, paclitaxel increased levels of both WT and MT Topo II␣ (Fig. 5C). Overall, these results indicate that p38␥ is specifically activated by Topo II␣ drugs, and activated p38␥ phosphorylates Ser-1524 of Topo II␣ which prevents its degradation and thereby enhances the growth-inhibitory effects of these drugs. . p38␥ activity is required for endogenous p-Topo II␣ expression. A, MCF-7 cells were treated with the indicated drugs for 6 h, and endogenous Topo II␣ proteins were isolated with a specific antibody for Western blot analysis (* the 0 values indicate that p-Topo II␣ was undetectable in the Topo II␣ precipitates). B, cells were infected with the indicated adenoviruses for 48 h and then treated with VP16 or solvent for 6 h and subjected to immunoprecipitation/Western blot analysis. The data shown are representative of two experiments (* values indicate the relative levels of p-Topo II␣ protein from the total Topo II␣ precipitates relative to the solvent controls, which are indicated with the underlines for Ad-␤-Gal-and Ad-p38␥/AGF-infected cells, respectively). C, 293T cells were transiently transfected with the indicated constructs for 48 h and were then treated with indicated drugs for 6 h and analyzed by Western blotting (* values indicate the relative levels of WT or MT FLAG-Topo II␣ protein relative to the respective solvent controls, which are indicated with the underlines after normalization to ␣-actinin). OCTOBER 14, 2011 • VOLUME 286 • NUMBER 41 p38␥ Stimulates Topo II␣ Gene Expression, and Immune Histochemistry Staining and Proteomics Analysis Verify Existence of the p38␥-Topo II␣ Axis-We showed previously that Ras stimulates Topo II␣ gene expression (19). Because p38␥ signals downstream of Ras (9, 10, 13), we next explored whether p38␥ may also stimulate Topo II␣ gene expression thereby serving as a second mechanism to sustain Topo II␣ levels. Indeed, transient CA p38␥ expression stimulates the luciferase activity driven by the human Topo II␣ promoter (19) in 293T cells. A similar stimulation was observed with Tet-induced CA p38␥ expression in MCF-7 cells (supplemental Fig. S3, A, B, and C). Moreover, levels of Topo II␣ RNA were increased by adenovirus-mediated expression of p38␥ in 468 cells and by Tet-oninduced expression of CA p38␥ in MCF-7 cells (supplemental Fig. S3, C and D). These results are consistent with their elevation of endogenous Topo II␣ protein expression under similar conditions (supplemental Figs. S2D and S3C). Overall, these data indicate that p38␥ also stimulates Topo II␣ gene expression.

p38␥ Regulates Drug-Target Signal Transduction
To demonstrate further the regulatory role of p38␥ in Topo II␣ expression, a group of primary breast cancer tissues was analyzed by immune histochemistry (12,13). Results in Fig. 6A (top) show that the most of Topo II␣ staining is in the nucleus (38), whereas the p38␥ signals are predominantly cytoplasmic (13). Although we were not be able to examine p-p38␥ expression in these tissues due to the lack of the antibody, our previous studies showed that the phosphorylated p38␥ is localized inside the nucleus (11) where it may interact with Topo II␣ and thereby regulate its activity in response to the chemotherapeutic stress. Most importantly, there is a significant correlation between increased p38␥ and elevated Topo II␣ protein levels in this group of tumor tissues versus their matched normal controls (Fig. 6A, bottom). These results indicate that p38␥ may also positively regulate Topo II␣ expression in primary cancer tissues.
To show whether endogenous p38␥ is required for Topo II␣ expression, tumor samples from Ras-transformed wild-type (p38␥ ϩ/ϩ ) and p38␥ knock-out (p38␥ Ϫ/Ϫ ) mouse embryonic fibroblasts (12) were analyzed for protein expression. Results in supplemental Fig. S4A showed a decreased Topo II␣ levels in p38␥ Ϫ/Ϫ tumors, consistent with a role of p38␥ in maintaining Topo II␣ expression. To demonstrate whether Topo II␣ is a natural substrate of p38␥, the same amount of proteins from the p38␥ ϩ/ϩ and p38␥ Ϫ/Ϫ samples was precipitated with an antibody against protein tyrosine phosphatase H1 (PTPH1), a p38␥-specific phosphatase (13); the immunoprecipitates were separated on SDS-PAGE (supplemental Fig. S4B) followed by digestions and proteomics analysis. We recently showed that PTPH1 binds and dephosphorylates p38␥ via PSD-95/Dlg/ ZO-1 homology (PDZ)-mediated complex formation (13). Because PTPH1 protein is expressed similarly in p38␥ ϩ/ϩ and p38␥ Ϫ/Ϫ tumors (supplemental Fig. S4A) and a substrate may form a complex both with its kinase and the associated phosphatase (39), comparative analyses of PTPH1 precipitates from the p38␥ ϩ/ϩ and p38␥ Ϫ/Ϫ samples may identify p38␥-dependent PTPH1-binding proteins. Indeed, results from the proteomics analysis (supplemental Fig. S4C and data not shown) show that PTPH1 binds Topo II␣ only in p38␥ ϩ/ϩ but not in p38␥ Ϫ/Ϫ tumors, indicating a natural complex of Topo II␣ with PTPH1 and p38␥. Thus, a correlation of increased p38␥ with up-regulated Topo II␣ in primary tumors highlights the pathophysiological significance of p38␥ in regulating Topo II␣ expression, whereas the p38␥-dependent PTPH1-Topo II␣ complex formation verifies Topo II␣ as a natural substrate of p38␥ MAPK.

DISCUSSION
Cancer chemotherapeutic drugs remain a major component of anticancer therapy. Although multiple pathways are involved in drug-induced growth inhibition and/or cell death, there is a pronounced lack of information about the mechanisms that

p38␥ Regulates Drug-Target Signal Transduction
regulate drug-target signal transduction (1,5). Topo II␣ is perhaps the most well established therapeutic target in human cancer, and its direct interaction with Topo II poisons is believed to be the foundation for the anti-tumor activity (40). It is not known, however, whether any cellular signaling pathways are involved in regulating the interaction of Topo II␣ with its poisons. Here, we provide several pieces of evidence that together indicate that p38␥ may act as a specific kinase to regulate the signal transduction between Topo II␣ and its poisons. First, p38␥ but not p38␣ MAPK is selectively activated by Topo IItargeting drugs (but not by paclitaxel) in intrinsically sensitive cells leading to sustained Topo II␣ expression. Second, p38␥ phosphorylates Topo II␣ at Ser-1524 in vitro and in vivo, which is important for Topo II␣ protein stability, for the regulation of Topo II␣ expression, and for the growth inhibition by Topo II poisons. Third, the manipulation of p38␥ activity by its forced expression and/or inhibition/depletion shows that it positively regulates Topo II␣ expression, the phosphorylation of Topo II␣ at 1524, and the sensitivity of breast cancer cells to Topo II drugs. Moreover, the intrinsic Topo II drug sensitivity correlates with the endogenous levels of p38␥ but not with those of Topo II␣. Also, p38␥ stimulates Topo II␣ transcription, and there is a positive correlation of increased p38␥ levels with Topo II␣ protein expression in primary tumor tissues. Finally, Topo II␣ forms a complex with the phosphatase PTPH1 in a p38␥-dependent manner. Overall, these results indicate that p38␥ specifically regulates the signal transduction between Topo II␣ and its poisons and thereby promotes increased chemotherapeutic activity of these drugs by maintaining their target Topo II␣ (Fig. 6B).
Although Topo II␣ is phosphorylated at several sites (32)(33)(34), phosphorylation of Ser-1524 is required both for the decatenation checkpoint (21) and cell cycle progression (34), indicating its critical role in maintaining the malignant progression. Although the Ser-1524 can be also phosphorylated by casein kinase II (32) and polo-like kinase 1 (34), functional roles of these kinases in regulating cancer cell response to Topo II drugs have not been reported. p38␥ MAPK, on the other hand, is overexpressed in primary breast cancer (9), and its RNA/protein levels are increased in ER-negative breast cancer cells (10). Therefore, the regulation and stabilization of Topo II␣ by p38␥ have important clinical implications. Although it has long been recognized that Topo II-targeting drugs are more effective in ER-negative breast cancer patients, the mechanisms have been mostly unknown (41,42). While the cytotoxicity of Topo II drugs is directly correlated with Topo II␣ levels (40), clinical studies have failed to show any connection between Topo II␣ and ER protein expression in primary breast cancer (43), indicating that other factor(s) must regulate Topo II␣ levels and the activity of Topo II drugs. Here, we show that the intrinsic sensitivity of breast cancer cells to Topo II drugs correlates with the levels of p38␥ protein expression in cultured breast cancer cells and that regulation of p38␥ expression/activity positively impacts Topo II␣/Ser-1524 phosphorylation, Topo II␣ protein stability, and the activity of Topo II drugs. Moreover, increased p38␥ protein expression in primary breast tumors correlates with elevated Topo II␣ protein, indicating a determinant role of p38␥ in the clinical response to a Topo II-drug-based therapy.
Future experiments are warranted to investigate whether p38␥ activity positively correlates with p-Topo II␣ in primary breast cancers and whether their combined up-regulation predict an improved clinical response to a Topo II drug-containing chemotherapy.
Drug-target interaction is a dynamic process that may involve alterations of thousands of cellular proteins in space and time (1). Consequently, how the totality of these changes regulates a given drug-target interaction for a coordinated therapeutic response is poorly understood. For example, previous studies have shown that the NF-B pathway plays a protective role in the cellular response to stress stimuli such as TNF␣, ionizing radiation, and daunorubicin, but whether these effects are executed through their cellular targets has not been demonstrated (44). p38 MAPKs, on the other hand, can facilitate an apoptotic response together with the JNK pathways in stress response (45,46), but whether any of these stress MAPKs can specifically regulate a given drug-target interaction and thereby modify their pharmacological outcomes still remains unknown. Here, we show that specific activation of p38␥ MAPK by Topo II drugs in intrinsically sensitive cells leads to phosphorylation and stabilization of Topo II␣, a process that is required for their growth-inhibitory activity. Although the stimulation of Topo II␣ gene expression by p38␥ may also contribute to the increased Topo II␣ levels and enhanced anti-tumor activity, this effect may be counteracted by an overall inhibitory effect of Topo II drugs on DNA synthesis and gene expression during the therapeutic stress. However, p38␥-mediated stimulation of Topo II␣ gene expression may be important for a long term therapeutic effect through regulating Topo II drug-induced DNA repair/checkpoint activation programs (47) (Fig. 6B). Further studies on the regulation of drug-target interaction by stress kinases could open a new avenue for novel cancer therapeutic development.