Cancer Cell Biology
p38 MAPK, but not ERK1/2, is critically involved in the cytotoxicity of the novel vascular disrupting agent combretastatin A4
Article first published online: 11 DEC 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 122, Issue 8, pages 1730–1737, 15 April 2008
How to Cite
Quan, H., Xu, Y. and Lou, L. (2008), p38 MAPK, but not ERK1/2, is critically involved in the cytotoxicity of the novel vascular disrupting agent combretastatin A4. Int. J. Cancer, 122: 1730–1737. doi: 10.1002/ijc.23262
- Issue published online: 19 FEB 2008
- Article first published online: 11 DEC 2007
- Manuscript Accepted: 18 SEP 2007
- Manuscript Received: 4 APR 2007
- National High-tech R & D Program. Grant Number: 2006AA020602
- Chinese Academy of Sciences. Grant Number: O6G8031014
- combretastatin A4;
- mitogen-activated protein kinase;
- p38 MAPK;
- vascular disrupting agent;
Combretastatin A4 (CA4) is a novel vascular disrupting agent that has promising clinical efficacy because of its ability to inhibit microtubule assembly and subsequently disrupt tumor blood flow. In this study, we demonstrate that mitogen-activated protein kinases (MAPKs) are critically involved in the cytotoxicity of CA4. CA4 stimulates both extracellular signal-regulated kinases (ERK1/2) and p38 MAPK in the BEL-7402 hepatocellular carcinoma cell line in a time- and dose-dependent manner. This stimulation is a result of CA4-induced microtubule disassembly, which is a reversible process. Reversibility of microtubule disassembly is evidenced by the ability of disassembled microtubules to reassemble just a few hours after CA4 treatment. p38 MAPK, but not ERK1/2, contributes to this microtubule reassembly following CA4 exposure, and only inhibition of p38 MAPK, but not ERK1/2, synergistically enhances CA4-induced G2/M cell cycle arrest. Consistent with this, p38 MAPK inhibitors such as SB203580 and SB202190 also synergistically enhance the cytotoxicity of CA4 in cells where p38 MAPK is activated by CA4. This enhancement appears to be specific for CA4 because the cytotoxicity of other microtubule-targeted agents such as paclitaxel, vinorelbine and colchicine was not affected by p38 MAPK inhibitors. These data indicate that p38 MAPK is a potential anticancer target and that the combination of CA4 with p38 MAPK inhibitors may be a novel and promising strategy for cancer therapy. © 2007 Wiley-Liss, Inc.
Combretastatin A4 (CA4), originally isolated from the South African Combretum caffrum tree, is a microtubule-destabilizing agent that inhibits microtubule assembly by binding to tubulin at the same site as another microtubule-targeted agent known as colchicine.1 Given that microtubules play a critical role in a wide variety of eukaryotic cellular functions including mitosis, motility, intracellular transport, generation, maintenance of cell shape and sensory transduction,2 CA4-induced microtubule disassembly leads to a dramatic cytotoxicity toward actively proliferating vascular endothelial cells and human cancer cells,3–6In vivo, the water soluble prodrug of CA4, CA4P, increases vascular resistance, reduces tumor blood flow and causes central tumor necrosis.7, 8 These effects may result from a disruption in the VE-cadherin/β-catenin/Akt signaling pathway.9 In addition, combining CA4P with other cytotoxic drugs, such as 5-fluorouracil, cisplatin and carboplatin, exhibits synergistic antitumor activities in rodent tumor models.10–12
Mitogen-activated protein kinases (MAPKs) transduce extracellular stimuli, from the cell surface to the nucleus, leading to alterations in gene expression and cell functions. To date, 3 major subfamilies of MAPKs have been identified: c-Jun N-terminal kinases/stress-activated protein kinases (JNK/SAPKs), extracellular signal-regulated kinases (ERK1/2 or p42/44 MAPK) and p38 MAPK.13 Of these, ERK1/2 which is primarily activated by mitogens and growth factors, mainly participates in cellular processes such as proliferation, differentiation and movement, although exceptions do exist.14, 15 In contrast, JNK and p38 MAPK, which are primarily activated in response to environmental stress and inflammatory cytokines, are mainly involved in cell growth arrest and apoptosis.16–19
The role of ERK1/2 and p38 MAPK in the cytotoxicity of microtubule-targeted agents has previously been investigated. For example, MEK specific inhibitors, including U0126 and PD98059, have been shown to enhance paclitaxel- or docetaxel-induced cell death in breast, ovarian and lung cancer cell lines, as well as prostate cancer cells.20, 21 These effects have also been observed with vinblastine.22 In contrast, p38 MAPK inhibitors suppress paclitaxel-induced cell death in the human ovarian carcinoma cell line, SKOV3 and nocodazole-induced apoptosis in HeLa cells.23, 24 Therefore, it is currently believed that ERK1/2 and p38 MAPK play important but opposing roles in the cytotoxicity of microtubule-targeted agents; specifically, ERK1/2 negatively regulates the antitumor activity of microtubule-targeted agents, whereas p38 MAPK has a positive antitumor effect. This view is reinforced by the finding that p38 MAPK contributes to CA4-induced membrane blebbing, whereas ERK1/2 protects cells from CA4-induced membrane blebbing in human umbilical vein endothelial cells (HUVECs).25
In this study, we investigated the role of MAPKs in the CA4-mediated cytotoxicity of a variety of tumor cells. We found that CA4 induced microtubule disassembly and that this disassembly stimulated both ERK1/2 and p38 MAPK activation in BEL-7402 cells. However, only inhibition of p38 MAPK, and not ERK1/2, was able to block microtubule reassembly and enhance both G2/M cell cycle arrest and cancer cell death that had been induced by CA4. Moreover, this enhancement was dependent on CA4-mediated p38 MAPK activation. In contrast to the current view, our results demonstrate, for the first time, that p38 MAPK is critically involved in the cytotoxicity of CA4 and that combining CA4 with p38 MAPK inhibitors may be a novel and promising strategy for cancer therapy.
Material and methods
CA4 was a generous gift from Professor Weiping Sheng, Shanghai University (Shanghai, China). Colchicine, paclitaxel, docetaxel, SB202190, LY294002, type I collagenase, 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) and monoclonal antibodies specific for β-tubulin and actin were purchased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum (FBS) was purchased from Hyclone (Logan, UT). Endothelial cell growth supplement (ECGS) was purchased from Gibco (Grand Island, NY). Epidermal growth factor (EGF), SB203580, Y27632 and propidium iodide (PI) were purchased from Promega (Madison, WI). Monoclonal antibodies to phospho-ERK1/2, ERK1/2 and phospho-p38 MAPK were purchased from Cell Signaling (Beverly, MA). Goat anti-mouse IgG conjugated with Alexa Fluro 488, phalloidin conjugated with Alexa Fluro 633 were purchased from Molecular Probes (Eugene, OR). PD0325901 was kindly provided by Professor Youhong Hu from the Shanghai Institute of Materia Medica, Chinese Academy of Sciences (Shanghai, China).
Cell culture and treatment
The human hepatocellular carcinoma cell lines, BEL-7402 and SMMC-7721, as well as the ovarian cancer cell line, 3AO, were obtained from the cell bank of the Shanghai Institute for Biological Sciences, Chinese Academy of Science (Shanghai, China). Cell lines for epithelial carcinoma (KB), non-small cell lung cancer (A549 and NCI-H460), colon cancer (HT-29 and HCT116), breast cancer (MDA-MB-231) and epidermoid cancer (A431) were all purchased from the American Type Culture Collection (Manassas, VA). The BEL-7402, SMMC-7721, 3AO, NCI-H460, HT-29, HCT116 and MDA-MB-231 cells were cultured in RPMI 1640 medium. A431 cells were cultured in Dulbecco's modified Eagle's medium (DMEM), KB cells were cultured in minimum essential medium (MEM) and A549 cells were cultured in Ham's F12K medium. All media were supplemented with 10% FBS and cells were incubated at 37°C in an atmosphere of 5% CO2. Cells were plated in 6-well plates and starved in serum-free medium for 16 hr before drug treatment.
Isolation and culture of human umbilical vein endothelial cells
Human umbilical vein endothelial cells (HUVECs) were isolated as described previously.26 Cells were detached by treatment with 0.25% type I collagenase, seeded on gelatin-coated culture plates and cultured at 37°C in an atmosphere of 5% CO2 in M199 medium containing 20% FBS, 20 μg/ml ECGS and 90 μg/ml heparin sulfate. Cells were passaged 2–6 times before being used in experiments.
Polymeric tubulin fraction assay27
Drug-treated cells were extracted in lysis buffer (80 mM Mes-KOH, pH 6.8, 1 mM MgC12, 1 mM EGTA, 0.5% Triton X-100 and 10% glycerol) containing protease inhibitors for 3 min at 30°C. Supernatants, containing detergent-soluble tubulin, were removed; the detergent-insoluble polymerized cytoskeletons were extracted with SDS sample buffer (50 mM Tris-HCl, pH 6.8, 100 mM dithiothreitol, 2% SDS, 0.1% bromophenol blue and 10% glycerol) and β-tubulin expression analyzed by western blotting.
After drug treatment, cells were washed twice with ice-cold PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4 and 1.8 mM KH2PO4, pH 7.4) and lysed in SDS sample buffer. Cell lysates, containing equal amounts of protein, were separated by SDS-PAGE and transferred to polyvinylidine difluoride membranes. After being blocked in 5% nonfat milk in TBST (Tris-buffered saline with 0.1% Tween 20, pH 7.6), membranes were incubated with the appropriate primary antibodies at 4°C overnight and then exposed to secondary antibodies for 2 hr at room temperature. Immunoreactive proteins were visualized using the enhanced chemiluminescence system from Pierce Chemical (Rockford, IL).
Cells grown on coverslips were washed with PBS and then fixed for 20 min in 4% paraformaldehyde/PBS at the end of drug treatment. After permeabilized by 0.1% Triton X-100/PBS for 5 min, cells were incubated with anti-β-tubulin antibody (at a dilution of 1:200 in 2% BSA/PBS) for 2 hr at 37°C, and then exposed to goat anti-mouse IgG conjugated with Alexa Fluro 488 (at a dilution of 1:200 in 2% BSA/PBS) for 1 hr at 37°C. Phalloidin conjugated with Alexa Fluro 633 (at a dilution of 1:200 in 2% BSA/PBS) and DAPI (0.1 μg/ml) were used to stain actin and nuclei, respectively. After extensive washes, cells were mounted and visualized by fluorescence microscope.
Cell cycle analysis
Adherent and detached cells were collected by trypsinization and centrifuged at 300g. Cells were washed twice with ice-cold PBS and fixed in ice-cold 70% ethanol overnight at −20°C. Fixed cells were centrifuged at 300g and stained with 50 μg/ml of PI, containing 50 μg/ml of DNase-free RNase A, at 37°C for 30 min. The DNA content of cells (10,000 cells/experimental group) was analyzed using a FACScan (BD Biosciences, San Jose, CA) and the ModFit LT Mac V3.0 program.
Sulforhodamine B cell survival assay
Cells were seeded in 96-well plates and then treated with different concentrations of drugs. After 72 hr of incubation, cells were fixed with 10% trichloroacetic acid for 1 hr at 4°C, washed 5 times with tap water and air-dried. Cells that survived were stained with 0.4% (w/v) sulforhodamine B (SRB) for 20 min at room temperature and washed 5 times with 1% acetic acid. Bound SRB was solubilized with 10 mM Tris and absorbance was measured at 540 nm.
The paired student's t-test was used to test for significance where indicated.
CA4 stimulates ERK1/2 and p38 MAPK activation
To investigate the effects of CA4 on MAPK activity, serum-starved BEL-7402 cells were treated with 1 μM CA4 for various periods of time. We observed activation of ERK1/2 and p38 MAPK as early as 15 min after CA4 treatment, which reached a maximal level at 45 min (Fig. 1a). The stimulation was also dose-dependent and the minimum concentration of CA4 required was 0.5 μM (Fig. 1b).
We then tested the effects of CA4 on MAPKs in other cell lines. To make the results comparable, the IC50s for individual cell lines were used. We observed activation of ERK1/2 and p38 MAPK in human ovarian cancer 3AO cells, human hepatocellular carcinoma SMMC-7721 cells and human epithelial carcinoma KB cells (Fig. 1c), but not in human colon cancer HT-29 cells and human non-small cell lung cancer A549 cells (Fig. 1d). These results indicate that stimulation of ERK1/2 and p38 MAPK is cell type-specific. Given that CA4 is a potent vascular disrupting agent; its effect on MAPKs was further investigated in HUVEC. Surprisingly, CA4 did not activate ERK1/2 or p38 MAPK in HUVECs (Fig. 1d).
CA4-stimulated ERK1/2 and p38 MAPK activation originates from microtubule disassembly
CA4 stimulated both ERK1/2 and p38 MAPK (Fig. 1a), but did not exert any effects on tyrosine kinase receptors such as EGFR, VEGFR and PDGFR (data not shown). This suggests that the CA4-mediated stimulation of MAPKs may result from its effect on the disruption of microtubule assembly rather than from alterations in tyrosine kinase receptor function. To confirm this, we investigated the effects of colchicine and vinorelbine (which also induce microtubule disassembly) as well as paclitaxel and docetaxel (which stabilize microtubules) on MAPKs. Our results indicate that, under identical treatment conditions, only colchicine and vinorelbine, but not paclitaxel or docetaxel, exhibit stimulating effects similar to those of CA4 (Figs. 2a and 2b). This suggests that microtubule disassembly probably stimulates ERK1/2 and p38 MAPK activation in BEL-7402 cells. To further confirm this, cells were pretreated with a relatively high concentration of paclitaxel (5 μM) to stabilize microtubules. The results showed that this pretreatment completely blocks CA4- but not EGF-stimulated ERK1/2 and p38 MAPK activation (Fig. 2c). These findings indicate that CA4-induced microtubule disassembly is the trigger for MAPK activation.
The causal relationship between the activation of MAPKs and microtubule disassembly was further investigated by using the polymeric tubulin fraction assay. In this assay, polymeric tubulin was separated from dimer tubulin based on its Triton X-100-insoluble properties and the amount of polymeric tubulin was measured as a reflection of the extent of microtubule assembly. As shown in Figure 2d, CA4 induced the disappearance of polymeric tubulin as early as 2 min post treatment, suggesting complete microtubule disassembly. On the other hand, a 1 hr pretreatment with 5 μM paclitaxel blocked CA4-induced microtubule disassembly (Fig. 2d; t−60 point). This blockage was treatment-schedule dependent since posttreatment with paclitaxel did not have the same effect (Fig. 2d; t+5, t+30, t+40 points). More importantly, posttreatment with paclitaxel not only failed to block CA4-induced microtubule disassembly (Fig. 2d; t+40 point), but also failed to block CA4-induced ERK1/2 activation (Fig. 2e; t+40 point). The same results were also observed for p38 MAPK (data not shown). Collectively, these results demonstrate a causal relationship between CA4-induced microtubule disassembly and ERK1/2 and p38 MAPK activation; i.e., CA4 induces rapid microtubule disassembly and this disassembly triggers ERK1/2 and p38 MAPK activation.
p38 MAPK, but not ERK1/2, activation is essential for microtubule reassembly following CA4 exposure
CA4 inhibits microtubule assembly in BEL-7402 cells and this inhibition is reversible since microtubules reassemble ∼4 hr following CA4 exposure (Fig. 3a). In contrast, polymeric actin seems to be not changed during CA4 treatment (Fig. 3a). Given that the ERK1/2 and p38 MAPK activation induced by microtubule disassembly occurs a few hours before microtubule reassembly, it is reasonable to speculate that MAPK activation may play a role in CA4-induced microtubule reassembly. To test this, we pretreated cells with PD0325901 or SB203580, which are 2 potent inhibitors specific for MEK and p38 MAPK, respectively. Surprisingly, PD0325901, which completely blocked CA4-induced ERK1/2 activation, did not influence CA4-induced microtubule reassembly (Fig. 3b); however, SB203580 pretreatment totally blocked microtubule reassembly following CA4 exposure (Fig. 3a). The same inhibitory effect was mimicked by another p38 MAPK specific inhibitor, SB202190 (Fig. 3c). As controls, pretreatments with LY294002 or Y27632, which are inhibitors specific for phosphoinositide-3 kinase and Rho kinase respectively, did not affect CA4-induced microtubule reassembly (Fig. 3b). These results clearly demonstrate that p38 MAPK, but not ERK1/2, plays an important role in microtubule reassembly following CA4 exposure.
To further confirm the role of p38 MAPK in microtubule reassembly, we first treated cells with CA4, or CA4 plus p38 MAPK inhibitor SB203580 for 1 hr, then, washed the cells with fresh media, and incubated the cells in the fresh media for another 1 hr in the absence or presence of SB203580 only (no CA4 any more). As shown in Figure 3d, microtubules reassembled in the absence of SB203580, but did not in the presence of SB203580 1 hr after withdrawal of CA4 (Fig. 3d).
The hypothesis that p38 MAPK is essential for CA4-induced microtubule reassembly was further tested by assessing alterations in cell morphology as an endpoint. Figure 3e shows that cells collapsed and exhibited irregular surfaces instead of smooth surfaces when exposed to CA4 for over 30 min (Fig. 3e). However, 4 hr after CA4 exposure, a large portion of cells returned to their normal states (Fig. 3e), suggesting that the effect of CA4 on cellular morphology is reversible. More importantly, the p38 MAPK inhibitor, SB203580, completely blocked the ability of cells to return to their normal morphology even after 8 hr of CA4 exposure (Fig. 3e; SB+CA4 8 hr). Consistent with the time course of morphological changes, microtubules became disassembled after 1 μM CA4 treatment for 2 hr, and reassembled for 8 hr as illustrated by immunofluorescence assay. SB203580 blocked microtubule reassembly (Fig. 3e). On the contrary, the intactness of polymeric actin was not significantly affected by CA4 under these conditions (Fig. 3e). These data indicate that the alterations of cell morphology are mainly produced by CA4-induced microtubule disruption and reassembly, and that p38 MAPK is indeed required for CA4-mediated microtubule reassembly.
p38 MAPK inhibition enhances CA4-induced G2/M cell cycle arrest
The equilibrium between microtubule assembly and disassembly is important for cell proliferation and growth.28, 29 Given that p38 MAPK inhibition suppresses microtubule reassembly, it is possible that p38 MAPK inhibition may also have an impact on cell growth and proliferation. Impelled by this idea, we examined the effects of p38 MAPK inhibition on CA4-induced cell cycle arrest. Our results indicate that 1 μM CA4 or 20 μM SB203580 alone did not induce G2/M arrest in BEL-7402 cells, since the percentages of G2/M cells were 15%, 19% and 14% in control, CA4 and SB203580-treated cells, respectively (Fig. 4a). However, combination of CA4 with SB203580 produced a significant G2/M cell accumulation with the percentage of the G2/M phase dramatically increasing to 50%. Another p38 MAPK inhibitor, SB202190, had the same effect as SB203580 (Fig. 4b). In contrast, other kinase inhibitors, including the MEK inhibitor, PD0325901, as well as the phosphoinositide-3 kinase inhibitor, LY294002 and the Rho kinase inhibitor, Y27632, failed to produce a G2/M arrest in BEL-7402 cells when combined with CA4 (Fig. 4b). The combination of SB203580 with other microtubule-targeted agents, such as vinorelbine, colchicine or paclitaxel, did not produce significant effects on G2/M cell accumulation (Fig. 4c).
p38 MAPK inhibition enhances CA4-induced cytotoxicity in cells where p38 MAPK is activated by CA4
Enhancement of CA4-induced G2/M cell cycle arrest by p38 MAPK inhibition implies that inhibition of p38 MAPK may influence the cytotoxicity of CA4. To make the results comparable, we introduced the relative inhibition effect (RIE) into this study. The RIE is the ratio of the IC50 of treatment with a single agent to the IC50 from cells treated with 2 agents. A higher RIE means that stronger inhibition of cell proliferation occurs. The IC50 for CA4 alone against BEL-7402 cells is 2.95 μM (Table I), while the IC50s for CA4 combined with 10 μM or 20 μM SB203580 are 1.01 μM and 0.4 μM (Table I), respectively. Therefore, the RIEs are 2.9 and 7.4 (Table I, Fig. 5a), indicating that the p38 MAPK inhibitor, SB203580, significantly potentiates CA4's cytotoxicity. This is also the case for SB202190; specifically the RIEs for combining CA4 with 10 μM or 20 μM SB202190 are 2.8 and 7.5, respectively. Considering that 20 μM SB203580 alone only leads to less than 20% inhibition of BEL-7402 cell proliferation, we conclude that p38 MAPK inhibitors synergistically potentiate CA4's cytotoxicity in BEL-7402 cells. Of note, other kinase inhibitors, including the MEK inhibitor, PD0325901 and the phosphoinositide-3 kinase inhibitor, LY294002, all failed to potentiate CA4's cytotoxicity in BEL-7402 cells (Fig. 5a). Furthermore, none of the combinations of vinorelbine, colchicine or paclitaxel with 20 μM SB203580 exerted similar potentiation effects when compared to single agent treatment (Fig. 5b).
|Cell line||Cell type||IC50 (nM)1||RIE2|
|CA4||+10 μM SB||+20 μM SB||+10 μM SB||+20 μM SB|
|BEL-7402||Hepatocellular carcinoma||2950 ± 660||1010 ± 260||400 ± 80||2.9||7.4|
|SMMC-7721||Hepatocellular carcinoma||890 ± 100||500 ± 70||290 ± 80||1.8||3.1|
|3AO||Ovarian cancer||2010 ± 270||730 ± 80||450 ± 20||2.8||4.5|
|KB||Epithelial carcinoma||48 ± 4||26 ± 2||22 ± 6||1.8||2.2|
|A549||Non-small cell lung cancer||42 ± 8||39 ± 6||N/A||1.1||N/A|
|H460||Non-small cell lung cancer||42 ± 5||47 ± 7||N/A||0.9||N/A|
|HCT116||Colon cancer||24 ± 2||25 ± 3||N/A||1.0||N/A|
|HT29||Colon cancer||840 ± 120||920 ± 160||N/A||0.9||N/A|
|MDA-MB-231||Breast cancer||68 ± 9||61 ± 11||N/A||1.1||N/A|
|A431||Epidermoid cancer||31 ± 1||34 ± 3||N/A||0.9||N/A|
|HUVEC||Endothelial cell||41 ± 5||37 ± 6||N/A||1.1||N/A|
Next, we examined the relationship between cytotoxicity enhancement and p38 MAPK activation in a variety of cells where p38 MAPK was either activated by CA4 or was not activated by CA4. On the basis of increases in RIEs (Table I), the cytotoxicity of CA4 was significantly enhanced by the p38 MAPK inhibitor, SB203580, in cells where p38 MAPK was effectively activated by CA4, including the hepatocellular carcinoma (SMMC-7721), ovarian cancer (3AO) and epithelial carcinoma (KB) cell lines (Fig. 1c). Conversely, the cytotoxicity of CA4 was not enhanced by the p38 MAPK inhibitor, SB203580, in cells where p38 MAPK was not activated by CA4 including in the non-small cell lung cancer (A549 and NCI-H460), colon cancer (HT-29 and HCT116), epidermoid cancer (A431 cells) and breast cancer (MDA-MB-231) cell lines, as well as in HUVECs (Table I). These data indicate that p38 MAPK inhibition enhances CA4-induced cytotoxicity and this enhancement is dependent on CA4-mediated p38 MAPK activation.
p38 MAPK is generally involved in cell growth arrest and apoptosis,16, 30 whereas ERK1/2 is implicated in cell survival and proliferation.15, 16 It is currently believed that inhibition of ERK1/2 enhances cytotoxicity of chemotherapeutics, whereas inhibition of p38 MAPK compromises cytotoxicity. In contrast to the current view, we provide evidence, for the first time, that p38 MAPK, but not ERK1/2, is critically involved in the cytotoxicity of the novel vascular disrupting agent, CA4, and that inhibition of p38 MAPK synergistically enhances, rather than inhibits, the cytotoxicity of CA4. This finding demonstrates that p38 MAPK is a promising anticancer target and that combining CA4 with p38 MAPK inhibitors may be a novel and reasonable strategy for cancer therapy. This may also have implications on the application of CA4 in the clinic.
The effect of microtubule-targeted agents on MAPKs has previously been studied and seems to be complex. Microtubule-stabilizing agents such as paclitaxel have a number of effects. For example, paclitaxel has been shown to stimulate both ERK1/2 and p38 MAPK in SKOV3 human ovarian carcinoma cells,23 but only stimulates p38 MAPK in MCF-7 cells.31 In addition, paclitaxel does not have an effect on either ERK1/2 or p38 MAPK in BEL-7402 cells (Fig. 2b; this study), but can inhibit ERK1/2 and p38 MAPK activation in KB-3 carcinoma cells.32 In contrast, microtubule-destabilizing agents such as vinblastine or colchicine stimulate ERK1/2 and p38 MAPK in human mammary epithelial 184B5/HER cells,33 and suppress MAPKs in KB-3 cells.32 In this study, we also found that the effect of CA4 on MAPKs is cell-type specific since CA4 stimulates ERK1/2 and p38 MAPK in BEL-7402, SMMC-7721, 3AO and KB cells, but not in HT-29, HCT116, A549, NCI-H460, MDA-MB-231 and A431 cancer cells. The reasons for this differential response to microtubule-targeted agents, in terms of MAPKs, are not clear. We believe that the MAPK status has a profound effect on the sensitivity of cells to chemotherapeutics. Our results clearly show that p38 MAPK inhibitors synergistically enhance the cytotoxicity of CA4 and that this enhancement occurs only in cells where p38 MAPK is effectively activated by CA4. Hence, determining the MAPK status, in response to microtubule-targeted agents, in cancer patients will be helpful for designing rational treatment modalities. We believe that combining microtubule-targeted agents with kinase inhibitors such as p38 MAPK inhibitors may prove especially useful.
The stimulation of MAPKs by microtubule-targeted agents has been attributed to disruption of microtubule assembly,34 but the precise mechanism is not fully understood. Here, we show that CA4-induced microtubule disassembly is the key trigger for MAPK activation because pretreatment of cells with paclitaxel stabilizes polymeric microtubules and therefore inhibits CA4- but not EGF-stimulated ERK1/2 and p38 MAPK activation. Moreover, our finding that only microtubule-destabilizing agents, such as colchicine and vinorelbine, but not microtubule-stabilizing agents such as paclitaxel or docetaxel activate ERK1/2 and p38 MAPK in BEL-7402 cells, provides further support for this view.
Microtubule-targeted agents including paclitaxel, colchicine, and vinblastine can stimulate cyclooxygenase-2 (COX-2) activity via ERK1/2 and p38 MAPK,33 and COX-2 can stimulate cell proliferation. So, it is likely that relative high concentrations of SB203580 (10 and 20 μM) may enhance the cytotoxicity of CA4 though its inhibition of p38 MAPK, and subsequent inhibition of COX-2 activity, or direct inhibition of COX-2 activity. To address this question, present study also assessed the effect of celecoxib, a specific inhibitor of COX-2, on CA4-induced cytotoxicity. Celecoxib neither inhibited CA4-induced microtubule reassembly, nor enhanced the cytotoxicity of CA4 (data not shown). Though we did not directly examine the effect of SB203580 on COX-2 activity, we believe that SB203580 enhancement of the cytotoxicity of CA4 is not through its inhibition of COX-2. In addition, Rho kinase which plays an important role in cytoskeleton rearrangement, appears not to be involved in microtubule reassembly since the Rho kinase inhibitor, Y27632, does not have an effect in this process.
The microtubule reassembly observed after CA4 exposure is an interesting and important phenomenon, and is consistent with the feature that CA4 is a reversible inhibitor of microtubule assembly.35 This reassembly prevents cells from undergoing apoptosis, making this effect of CA4 treatment a noncatastrophic event. As such, the reassembly of microtubules has a major impact on cell survival and ultimately compromises the antitumor activity of CA4. It is therefore important to determine the molecular mechanism responsible for microtubule reassembly following CA4 treatment, in order to improve the antitumor activity of CA4. It is evident that p38 MAPK, but not ERK1/2, plays a key role in microtubule reassembly based on the following findings. First, only inhibition of p38 MAPK suppresses the CA4-induced microtubule reassembly (Fig. 3b). Second, only inhibition of p38 MAPK, but not ERK1/2, blocks the recovery of the CA4-induced alterations in cell morphology to the normal state (Fig. 3e). Third, only p38 MAPK inhibitors synergistically enhance G2/M cell cycle arrest and cytotoxicity of CA4 and this enhancement occurs only in cells where p38 MAPK is activated by CA4.
It has been suggested that p38 MAPK contributes to membrane blebbing induced by CA4, while ERK1/2 protects cells against membrane blebbing in HUVECs.25 This seems to be inconsistent to our present findings. This discrepancy may be mainly due to the differences in cells tested (endothelial cells of the vasculature vs. tumor cells). Different cell types may lead to differential sensitivity to CA4. In addition, our view also differs from previous reports that inhibition of p38 MAPK reduces the cytotoxicity of other microtubule-targeted agents.24 This discrepancy may be due to the differences in drug doses and drug treatment durations used. Previous studies used high concentrations of microtubule-targeted agents which usually greatly exceeded the IC50s and longer treatment durations which were usually in excess of 12 hr.24 In this study, lower concentrations of CA4 were used (which were closer to the IC50s) and cells were treated for shorter periods of time (less than 2 hr). Under these conditions, CA4 stimulates p38 MAPK but is not able to induce apoptosis. These may explain why p38 MAPK plays different roles in the cytotoxicity of microtubule-targeted agents.
Of note, p38 MAPK inhibitors only enhanced the cytotoxicity of CA4 but not other inhibitors of microtubule assembly such as vinorelbine and colchicine. The precise mechanism underlying this behavior is not currently understood. Although both colchicine and vinorelbine are able to stimulate p38 MAPK, the concentrations required are usually 1 μM, which is 100-fold above their IC50s in BEL-7402 cells. In fact, 10 nM of colchicine and vinorelbine which is close to their IC50s, does not stimulate p38 MAPK or ERK1/2. These results may explain why combinations of p38 MAPK inhibitors with colchicine or vinorelbine do not exhibit effects above those mediated by single agent treatments.
The primary target of vascular disrupting agents is tumor vasculature and the antitumor mechanism of action of these agents is to cause rapid and selective shutdown of tumor vasculature, leading to secondary tumor cell death.36 Accordingly, most previous studies have focused on the antivascular effect of CA4 and used the endothelial cells as major target cells. However, it has also been suggested that the antitumor effects of CA4 may be an integrated consequent of destruction of tumor vasculature and killing of tumor cells.37 Moreover, it is difficult to exclude that CA4, as a tubulin-binding agent, exerts its antitumor activity through its direct killing of tumor cells based on the facts that the IC50s of CA4 in in vitro cultured tumor cells are at the range of 10 nM to 10 μM, and the mean plasma CA4 AUC0–24 hr and Cmax at 52 mg/m2 (the therapeutical dose in human) is 2.19 μmol hr/l and 1.89 μmol hr/l in human, and 5.8 μmol hr/l and 9.8 μmol hr/l in the mouse bearing the carcinoma CaNT, respectively.38 In addition, other tubulin-binding agent such as vinblatine also has antivascular effect.39 These facts collectively suggest that tumor cell compartment may be another target of CA4 and the present study provides new insights into the mechanism of action of vascular disrupting agents.
p38 MAPK inhibitors are currently in clinical trials as antiinflammation agents40–42 and are rarely applied in the anticancer field. Our findings show that inhibition of p38 MAPK synergistically enhances the cytotoxicity of CA4, demonstrating that p38 MAPK is a promising anticancer target. This enhancement in cytotoxicity may greatly broaden the application of p38 MAPK inhibitors, especially in anticancer therapy.
In summary, our results show that the vascular disrupting agent CA4 stimulates p38 MAPK and ERK1/2, but only p38 MAPK activation provides a protective effect for CA4-treated cells. The finding that p38 MAPK inhibitors synergistically enhance the cytotoxicity of CA4 provides novel evidence that the combination of p38 MAPK inhibitors with CA4 may represent a rational and promising strategy for cancer therapy.
- 28Tubulin and microtubules as targets for anticancer drugs. Prog Cell Cycle Res 2003; 5: 309–25., , , .