Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Paclitaxel (Taxol) is a widely used cancer chemotherapeutic drug that exhibits clinical activity in a range of human malignancies (Arbuck et al., 1993) and is being extensively evaluated in ongoing trials as a single agent (Malingre et al., 2000) and in combination therapy (Piccart et al., 2000). The binding site for paclitaxel on -tubulin has been assigned from photoaffinity labeling experiments (Rao et al., 1994, 1995, 1999) and from the electron crystallography model in which the - and -tubulin dimer is fitted to a 3.7-Å density map (Nogales et al., 1998). In addition to its ability to bind to microtubules, paclitaxel also has been shown to interact with bcl-2 (Rodi et al., 1999) and to phosphorylate both bcl-2 and Raf-1 (Blagosklonny et al., 1997; Torres and Horwitz, 1998), events that have been related to the accumulation of cells in G₂M. One major problem with paclitaxel chemotherapy is the acquisition of clinical resistance, which causes chemotherapeutic failure leading to progressive disease. Potential mechanisms of taxane resistance include the overexpression of P-glycoprotein and/or mutations in tubulin, both of which may impair the ability of the drug to bind efficiently to its target. These factors have motivated a search for other natural products that target the microtubule and have superior or equivalent activity compared with paclitaxel but without the associated problems. Recently, three additional compounds that fulfill these criteria have been isolated from diverse natural sources. These include the epothilones (Su et al., 1997), which are currently under clinical evaluation, eleutherobin (Chen et al., 1998), and discodermolide (Smith et al., 2000).

In addition to the well-documented effects on microtubules, a number of reports have described activation of components of the mitogen-activated protein kinase (MAPK) pathway in response to paclitaxel treatment. MAPKs are serine-threonine protein kinases that are activated in response to a diverse range of stimuli, including growth factors, hormones, neurotransmitters, and cellular stress. So far, five subfamilies have been identified; however, the three major MAPK families are the extracellular signal-regulated kinases (ERKs), the c-Jun NH₂-terminal kinases (JNKs), and the p38 kinases (Widmann et al., 1999). MAPK kinase cascades are characterized by a sequential signaling cascade in which MAPK (MAPKKKs) kinase kinases phosphorylate and activate downstream MAPK kinases (MAPKKs) that ultimately activate MAPKs. Through these specific intracellular signaling cascades, a variety of extracellular stimuli are transduced through cells in a tightly regulated manner. In general, the ERKs are activated after growth factor and differentiation stimuli (Widmann et al., 1999) and are the best characterized pathway in mammalian cells. Several microtubule-associated proteins are substrates for ERK1 and 2, including microtubule-associated proteins 1, 2, and 4 and Tau (Seger and Krebs, 1995), and activated ERKs have been localized to kinetochores of mitotic cells (Zecevic et al., 1998). JNK and p38 are more commonly activated in response to stress and cellular damage (Mendelson et al., 1996; Niisato et al., 1999). Cellular redox state, tyrosine kinases, and phosphatases are thought to be involved in the activation of stress responses; however, the mechanisms for regulation remain unknown. In contrast, the response of the JNK pathway to extracellular ligands is well characterized. The latter are activated by the TNF receptor family, cytokine receptors, and tyrosine kinase receptors (Minden and Karin, 1997; Chainy et al., 2000). Hence, the different kinase members achieve specificity by signaling to unique and common downstream targets.

Several studies have described the activation of various components of the MAPK family in different cell models after treatment with paclitaxel and other microtubule interacting agents (Lieu et al., 1998; Shtil et al., 1999; Wang et al., 1999; Yujiri et al., 1999; Stone and Chambers, 2000). These studies suggest that the activation of various family members is variable and depends on the cell model. Hence, the relevance of MAPK activation to paclitaxel-induced death has not been fully resolved, and the functional relationship between paclitaxel-induced death and other signal transduction events, such as bcl-2 and Raf-1 phosphorylation, remains unclear. The aims of this study were to evaluate the effects of paclitaxel and other microtubule interacting agents on the activation of the MAPK family, initially in HeLa and Chinese hamster ovary (CHO) cells, and to relate these findings to paclitaxel-induced cytotoxicity. A pharmacological evaluation of the combination of MAPKK inhibition and paclitaxel treatment in three cell lines showed additivity in cells exhibiting paclitaxel-induced ERK activation and antagonism in cells with low endogenous ERK activity that did not display paclitaxel-induced ERK activation. These data suggest that the inhibition of ERK signaling coupled with paclitaxel therapy may be a useful clinical combination that requires additional validation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture and Reagents. GRC+ LR-73 Chinese hamster ovary (CHO) cells and HeLa cells were cultured in minimum essential medium, supplemented with 10% heat-inactivated fetal bovine serum and 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA). CHO cells (Pollard and Stanners, 1979) were the kind gift of Dr. J.W. Pollard (Albert Einstein College of Medicine, Bronx, NY). Exponentially growing cells were used in all experiments. Paclitaxel and baccatin III were obtained from the Drug Development Branch of the National Cancer Institute (Bethesda, MD) and docetaxel (Taxotere) was a gift from Aventis (Strasbourg, France). The epothilones (Meng et al., 1997), eleutherobin (Chen et al., 1998), their structural analogs, and discodermolide (Smith et al., 2000) have been described previously. The Pathdetect in vivo trans reporting system was purchased from Stratagene (Austin, TX). The inhibitors PD98059 and SB203580 and the caspase inhibitor Z-VAD-FMK, were purchased from Calbiochem Inc. (La Jolla, CA) and U0126 was purchased from Promega (Madison, WI). All drugs and inhibitors were dissolved in 100% dimethyl sulfoxide, the final concentration of which was kept below 0.1% in all experiments. The pcDNA3-FLAG-JNK1-APF plasmid contains a catalytically inactive dominant-negative mutant of JNK1 with an N-terminal FLAG tag (Gupta et al., 1995) and was kindly provided by Dr. R. J. Davis (Howard Hughes Medical Institute Research Laboratories, University of Massachusetts Medical Center, Worcester, MA). All other reagents were from Sigma Chemicals (St. Louis, MO).

Measurement of In Vivo Signal Transduction. The Pathdetect system was used to measure the signal transduction events after exposure to microtubule interacting agents. This system uses vectors that express chimeric trans-activator proteins containing the DNA binding domain of GAL4 and the trans-activation domain of either Elk-1 or c-jun. These domains become phosphorylated if an upstream signal transduction pathway is activated and subsequently activates the transcription of a luciferase gene from a reporter plasmid that is driven by a promoter containing five GAL4 binding sites. Negative control experiments were performed using a plasmid encoding only the GAL4 DNA binding domains, which did not result in luciferase transcription, thereby demonstrating that luciferase expression is dependent on the in vivo phosphorylation of Elk-1 or c-jun (data not shown). All experiments were optimized initially in a Chinese hamster ovary-derived cell line, GRC⁺ LR-73, which undergoes growth arrest in low concentrations of serum without impaired cellular viability (Pollard and Stanners, 1979). CHO and HeLa cells were transiently transfected using calcium phosphate precipitation and lipofectin (Invitrogen), respectively. Approximately 3 × 10⁵ cells were seeded into six-well culture plates 24 h before transfection. Each well was transfected with 1 µg of pFR-Luc, 50 ng of pFA-Elk or pFA-cjun, and 0.2 µg of CMV--galactosidase. For positive control experiments, cells were also cotransfected with pFC-MEK1 or pFC-MEKK1, upstream activators of the ERK and JNK pathways, respectively. Because there is no commercially available inhibitor of the JNK pathway, a catalytically inactive dominant negative mutant of JNK1 (pcDNA3-FLAG-JNK1-APF) was used to inhibit c-jun trans-activation. Where required, an equivalent quantity of empty vector (pcDNA3) was cotransfected to ensure a constant quantity of DNA. Cells were transfected for 8 to 12 h, rinsed with PBS and incubated with drug diluted in media containing 0.3% serum for the times indicated. Because activation of Elk-1 is serum dependent, all experiments using the reporter system were carried out with culture media supplemented with 0.3% serum. Cells were lysed in 500 µl of extraction buffer (40 mM tricine, pH 7.8, 50 mM NaCl, 2 mM EDTA, pH 7.8, 1 mM MgSO₄, 5 mM dithiothreitol, and 1% Triton X-100) and 50 µl was used to measure luciferase and -galactosidase activities. All luciferase values were normalized to -galactoside and expressed relative to basal control levels, which were assigned a value of 1.

Flow Cytometry. Flow cytometry was performed on a FACScan (BD Biosciences, San Jose, CA) using CellQuest software. For cell cycle analysis, 1 × 10⁶ cells (including nonadherent cells) were fixed in 70% ethanol for at least 20 min. The cell pellets were washed in cold PBS and incubated for 30 min in PBS containing 10 µg/ml propidium iodide and 1 µg/ml RNase A (Boehringer Mannheim, Indianapolis, IN) at 37°C. Propidium iodide fluorescence was quantified using the FL2 detector and cellular aggregates gated from all samples. For the detection of phosphatidylserine externalization, adherent and nonadherent cells were harvested by trypsinization and 1 × 10⁶ cells/ml were stained with 5 µL of annexin-V Fluor-488 labeled antibody (Molecular Probes Inc., Eugene, OR) in annexin binding buffer (10 nM HEPES, 140 mM NaCl, and 2.5 mM CaCl₂, pH 7.4) according to the manufacturer's instructions and analyzed by flow cytometry on the FL1 detector.

Immunoblotting. Cell extracts were prepared from adherent and nonadherent cells after drug treatment by lysis in a buffer composed of 50 mM Tris, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, and 0.1% Nonidet P-40 containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 2 µg/ml aprotinin, and 2 µg/ml leupeptin) and phosphatase inhibitors (50 mM NaF and 1 mM sodium orthovanadate). Cellular debris was removed by centrifugation and the protein was quantified using the Bio-Rad method (Bio-Rad, Hercules, CA). Samples were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose. Ponceau S staining before immunoblotting confirmed equal loading of samples. Membranes were blocked for 1 h in 3% bovine serum albumin and incubated in primary and secondary antibodies (Amersham Pharmacia Biotech, Piscataway, NJ) according to the manufacturer's instructions. All antibodies were used at a 1:1000 dilution in 1× Tris-buffered saline containing 0.1% Tween-20 and 1 to 2% bovine serum albumin and were as follows: Bcl-2 mAb (DAKO, Glostrum, Denmark), phospho-ERK mAb, phospho-JNK mAb, JNK pAb (New England Biolabs, Beverly, MA), ERK2 pAb, MEKK1 pAb C-22, (Santa Cruz Biotechnology Inc., Santa Cruz, CA), and PARP mAb (P76420), c-Raf-1 mAB (BD Pharmingen, San Diego, CA). An additional PARP antibody, Anti-PARP p85 fragment pAB (Promega, G7341), specific for only the 85-kDa cleavage product, was used at a dilution of 1:750. Membranes were developed using enhanced chemiluminescence reagent (Amersham).

Multiple Drug Effect Analysis. Cells were seeded in triplicate into 24-well plates and, after adherence, serial dilutions of paclitaxel, U0126, or both were added for 72 h. The drug concentrations evaluated were based on the IC₅₀ value for each individual drug so that combinations of paclitaxel and U0126 were assessed at their equipotent ratio (i.e., at the ratio of each respective IC₅₀ value). Drug combinations were evaluated concurrently or sequentially, wherein paclitaxel or U0126 was given for 24 h before the other drug. The effect of each drug treatment was determined by counting the number of attached viable cells after the 72-h incubation period and expressing this number as a ratio of the number of cells treated with dimethyl sulfoxide alone. This ratio was applied to the combination index (CI) method of Chou and Talalay (1984), using the software Calcusyn (Biosoft, Cambridge, UK) to analyze the nature of the interaction between paclitaxel and U0126. This software applies a mathematical model that computes a CI ratio for various levels of cytotoxicity (fractional inhibition), such that a CI value of 1 indicates additivity, CI < 1 indicates synergism, and CI > 1 indicates antagonism. For this analysis, the more conservative assumption of mutual nonexclusion (dissimilar mechamisms of action for paclitaxel and U0126) was applied to the derivation of CI ratios.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Activation of the Nuclear Transcription Factors Elk-1 and c-jun by Microtubule-Interacting Agents. A MAPK reporter system was used to determine the ability of various microtubule-interacting drugs to phosphorylate the nuclear transcription factor Elk-1 via activation of the ERK, JNK, and p38 pathways (Yang et al., 1998). The system employs a fusion construct that contains a GAL4 DNA binding domain and the trans-activation domain of Elk-1 to induce expression of a luciferase reporter that has five GAL4 binding domains. After MAPK activation, the trans-activation domain of Elk-1 becomes phosphorylated and subsequently binds to and induces transcriptional activation of the luciferase reporter.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   A, Elk-1 activation by paclitaxel (Taxol) and epothilone B in CHO cells. Cells were transiently cotransfected with the trans-activation domain of Elk-1, a luciferase reporter, and -galactosidase. After transfection, cells were exposed to paclitaxel or epothilone B in the presence or absence of the MEK1/2 inhibitor PD98059 for 48 h. Fold luciferase activation levels were normalized to -galactosidase expression and the data presented summarize the mean (± S.E.) of three independent experiments. B, c-jun activation by paclitaxel in CHO cells. Cells were transiently cotransfected with the trans-activation domain of c-jun, a luciferase reporter, and -galactosidase. A dominant negative mutant of JNK1 (pcDNA3-FLAG-JNK1-APF) was used to inhibit c-jun activation. After transfection, cells were exposed to 1 µM paclitaxel for 48 h. Positive control cells were also cotransfected with the trans-activation domain of MEKK1, an upstream activator of JNK. Fold luciferase activation levels were normalized to -galactosidase expression and the data presented summarize the mean (±S.E.) of three independent experiments and suggest that the JNK pathway is activated in response to paclitaxel treatment in CHO cells.

Activation of Elk-1 Requires an Interaction with Microtubules. Fig. 2A depicts the concentration-dependent activation of Elk-1 by both microtubule stabilizing and depolymerizing agents in CHO cells after 48 h of drug treatment. The degree of activation for each drug was related to its cytotoxic potency. Treatment with high concentrations of epothilone B (500 nM) resulted in decreased activation of Elk-1 (compared with the effect at 50 nM) and is probably related to cell death because of a lethal concentration of drug, given that epothilone B is the most active of the drugs. Transcriptional activation of Elk-1 was observed for both microtubule stabilizing and depolymerizing agents; however, other cytotoxic agents, such as cisplatin, 5-FU, and etoposide, resulted in insignificant levels of activation. The effect of inactive analogs of microtubule-stabilizing drugs (Table 1) in both CHO and HeLa cells were evaluated; compounds that were deemed inactive by microtubule polymerization assays (Su et al., 1997; McDaid et al., 1999) also failed to activate Elk-1 (Fig. 2B). This strongly suggests a relationship between the perturbation of microtubule dynamics and the nuclear trans-activation of Elk-1. The data shown are for CHO cells. Similar results were also obtained for HeLa cells.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   A, dose-dependent activation of Elk-1 by microtubule stabilizing and depolymerizing agents in CHO cells. Cells were transiently transfected with the trans-activation domain of Elk-1, the luciferase reporter, and -galactosidase and, after transfection, was exposed to various concentrations of the microtubule-interacting compounds paclitaxel (Taxol), epothilone B, eleutherobin, discodermolide, and vinblastine for 48 h. Fold luciferase activity was determined after normalization to -galactosidase. B, drugs that interact directly with microtubules mediate Elk-1 trans-activation in CHO cells. Cells were transiently transfected and exposed to 50 nM concentrations of various microtubule-interacting agents or their inactive analogs for 48 h. trans-Activation of Elk-1 was observed only in those compounds that had previously been determined to be active with the use of in vitro microtubule polymerization assays.

Interaction with Microtubules Is Required for the Phosphorylation of Raf-1 and Bcl-2. Since the data established a relationship between the trans-activation of Elk-1 and perturbations of microtubule dynamics, we evaluated the effects of three microtubule-stabilizing agents and their inactive analogs on cell-cycle kinetics and the phosphorylation status of Raf-1 and Bcl-2 in HeLa cells. We also evaluated cleavage of the nuclear repair enzyme PARP as a marker of cell death (Nosseri et al., 1994), using a monoclonal antibody derived from the N terminus of PARP; hence, this antibody recognizes the intact protein and the 24-kDa cleavage product derived from the N-terminal. The phosphorylation of Raf-1 and Bcl-2 in response to treatment with microtubule-interacting drugs is well documented (Blagosklonny et al., 1997) and these events are known to be associated with the accumulation of cells in mitosis (Scatena et al., 1998; Laird et al., 1999).

View larger version (43K): [in this window] [in a new window]   Fig. 3.   A, the cell cycle profile of HeLa cells exposed to various concentrations of microtubule active and inactive agents for 16 h. Adherent and nonadherent cells were harvested and DNA content was determined by flow cytometric analysis after fixation and propidium iodide staining, as described under Materials and Methods. B, Western blotting for expression of Bcl-2, Raf-1, and the 24-kDa cleavage product of PARP in total HeLa cell lysates exposed to microtubule active and inactive agents for 16 h. Data are representative of two independent experiments, both yielding identical results.

Differences in the Expression of Endogenous Levels of MAPK in CHO and HeLa Cells in Response to Paclitaxel Treatment. Because an association between exposure to microtubule-interacting agents and activation of MAPK has been demonstrated, the endogenous levels of MAPK expression were evaluated in control and drug-treated cells. Lysates were subjected to immunoblotting using monoclonal antibodies specific for Thr183/Tyr185 of stress-activated protein kinase/JNK and Thr202/Tyr204 of ERK1/2. Dual phosphorylation of the respective residues in each pathway has been determined to be essential for kinase activation. Figure 4A depicts the expression of phospho-ERK and phospho-JNK in CHO cells after exposure to paclitaxel at both 10 nM and 1 µM concentrations. No detectable activation of p38 was observed in this cell line. Total levels of MAPK proteins were unchanged throughout the drug treatments (data not shown). There was transient activation of phospho-ERK 4 h after drug exposure and dephosphorylation of phospho-ERK after 24 h at 1 µM paclitaxel that coincided with an increase in the proportion of cells in G₂M (Fig. 4B). No significant alterations in the cell cycle profile were observed with 10 nM paclitaxel treatment. CHO cells exhibited some activation of phospho-JNK, occurring predominantly after 1 µM paclitaxel treatment. This data is consistent with the trans-activation of c-jun described in Fig. 1B. Transient altered electrophoretic mobility of MEKK1, a 196-kDa kinase that acts upstream of JNK, occurred concomitant with phospho-JNK activation observed with 1 µM paclitaxel exposure. UV-irradiated cells were used as positive control cells and exhibited strong activation of both the ERK and JNK pathways and degradation of MEKK1, as anticipated.

View larger version (132K): [in this window] [in a new window]   Fig. 4.   A, expression of phospho-ERK1/2, phospho-JNK1/2 and MEKK1 proteins in CHO cell lysates after exposure to 10 nM and 1 µM paclitaxel (Taxol). Cells exposed to UV irradiation for 30 min were used as positive control cells. Data is representative of three independent experiments, each yielding identical results. B, Time course of CHO cell cycle distribution after exposure to 1 µM paclitaxel. Adherent and nonadherent cells were harvested, fixed, and stained with propidium iodide as described under Materials and Methods. C, expression of endogenous MAPK levels in HeLa cells after paclitaxel treatment. Time course of phospho-ERK1/2, phospho- JNK1/2, p38, and MEKK1 expression in HeLa cells after exposure to 10 nM and 1 µM paclitaxel. D, time course of HeLa cell cycle distribution after exposure to 10 nM and 1 µM paclitaxel. E, caspase inhibition has a moderate effect on paclitaxel-induced PARP cleavage and no effect on the altered electrophoretic mobility of MEKK1 induced by paclitaxel. Cells were treated with either 10 nM paclitaxel for 16 h or 50 µM Z-VAD-FMK and 10 nM paclitaxel for 16 h. Adherent and nonadherent cells were harvested, cell lysates were prepared, and immunoblot analysis was performed on the indicated proteins. Two antibodies for PARP expression were evaluated, as indicated, in an N-terminal derived antibody that recognizes the 24-kDa cleavage fragment and one that recognizes only the cleaved 85-kDa fragment. The data shown are representative of three independent experiments. F, suppression of JNK by transient transfection of a dominant negative mutant of JNK1 (pcDNA3-FLAG-JNK1-APF) does not modulate the degree of paclitaxel-induced PARP cleavage in HeLa cells. Cells were transiently transfected as described under Materials and Methods and exposed to paclitaxel for 16 h. Adherent and nonadherent cells were harvested, lysates were prepared, and immunoblot analysis was performed on the proteins indicated. The results are representative of three independent experiments yielding an average transfection efficiency of 40 to 45%.

Inhibition of the ERK Pathway Enhances Paclitaxel-Induced Apoptosis. Because the time-dependent activation of phospho-ERK in HeLa cells coincides with the accumulation of cells in the G₂M phase of the cell cycle and with increases in the proportion of hypodiploid cells (Fig. 4D), we questioned whether the ERK pathway is mediating paclitaxel-induced cell death. Figure 5A illustrates the effect of the MEK inhibitors PD98059 and U0126 on phospho-ERK expression and the cleavage of PARP in the presence and absence of paclitaxel in CHO cells. Phospho-ERK activity was partially down-regulated by PD98059 and U0126 in both the presence and the absence of paclitaxel; however, both inhibitors augmented paclitaxel-induced PARP cleavage. Identical observations were noted in HeLa cells (Fig. 5B). Note that the expression of the 24-kDa cleavage fragment of PARP was comparable for both cell lines, but in Fig. 5A, the Western blot is overexposed.

View larger version (38K): [in this window] [in a new window]   Fig. 5.   A, inhibition of ERK potentiates paclitaxel (Taxol)-induced PARP cleavage in CHO cells. Western blot analysis of phospho-ERK and the 24-kDa cleavage product of PARP after 16 h paclitaxel treatment with the MEK1 inhibitors, PD98059 and U0126. B, inhibition of ERK potentiates paclitaxel-induced PARP cleavage in HeLa cells. C, inhibition of ERK potentiates paclitaxel-induced phosphatidylserine externalization in HeLa cells. Cells were exposed to paclitaxel and U0126 at the various concentrations indicated for 24 h. Cells were harvested and incubated with annexin-V-488 fluorochrome as described under Materials and Methods. The percentage of annexin-V-positive cells was determined from the flow cytometric evaluation of stained cells.

The Combination of Paclitaxel and U0126 Is Additive. To extend our observations that the combination of paclitaxel with U0126 (and PD98059) potentiated paclitaxel-induced PARP cleavage and annexin-V externalization, we evaluated the nature of this interaction using the multiple drug-effect analysis of Chou and Talalay (1984), which computes the level of additivity, synergism, or antagonism at discrete levels of cytotoxicity. The combination indices (CI values) were derived using the mutually nonexclusive assumption of interaction between both drugs. Table 2 summarizes the findings from three cell lines in which the concurrent combination of paclitaxel and U0126 was evaluated. Additivity was confirmed in HeLa and A549 cells, both of which express phospho-ERK, which is induced in a time- and concentration-dependent manner after paclitaxel treatment (Yang and Horwitz, 2000). MCF-7 cells were used as a control because they express low levels of phospho-ERK (Fig. 6) and do not exhibit paclitaxel-induced ERK activation (Shtil et al., 1999). The combination of paclitaxel with U0126 was antagonistic in MCF-7 cells. Hence, these results prove additivity between concurrent paclitaxel and MEK inhibition in cell lines exhibiting ERK activation in response to paclitaxel and provide an explanation for the potentiation of paclitaxel-induced cell death observed in the presence of U0126. Overall, the data supports a survival function for the ERK/MEK signaling pathway.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Expression of ERK and phospho-ERK in paclitaxel (Taxol)-treated MCF-7 cells. Cells were treated with 10 nM paclitaxel for the indicated times and 20 µg of total cell lysate was used for immunoblot analysis, together with 20 µg of total cell lysate from A549 cells. Compared with A549 cells, MCF-7 cells expressed low levels of phospho-ERK and did not exhibit paclitaxel-induced ERK activation.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Microtubules form a scaffolding and transport network in eukaryotic cells and their unique polymerization/depolymerization dynamics are critical for multiple cellular functions. Perturbation of normal microtubule dynamics by microtubule-interacting agents can be regarded as an intracellular stress that leads to cell death. A correlation between the potency of microtubule inhibitors and the induction of mitotic arrest is well documented; however, the signal transduction events that mediate the cytotoxic effects of these drugs are poorly defined. The data presented here indicate that interactions between such drugs and the microtubule are essential for MAPK activation; however, there is specific involvement of defined MAPK subfamilies in different cell lines. For example, activation of ERK in response to microtubule binding agents has been reported in A549 (human lung carcinoma) (Yang and Horwitz, 2000), HL-60 (human myeloid leukemic; Blagosklonny et al., 1999), U937 (human T-cell leukemic; Lieu et al., 1998), HeLa, and CHO cells. Conversely, JNK activation in response to paclitaxel and other microtubule inhibitors has been observed, in Chinese hamster ovary and HeLa in this report, in MCF-7 (human breast cancer; Shtil et al., 1999), KB-3 (human epidermal; Stone and Chambers, 2000), BR (Wang et al., 1999), 67R (both human ovarian), and human fibroblast and SAOS2 (human osteosarcoma) cells (Wang et al., 1998). P38 activation in response to paclitaxel has been observed in MCF-7 cells (Shtil et al., 1999) and in HeLa cells (this report). One must assume that there is a high level of redundancy and compensation occurring between these signaling pathways in different cell lines.

Despite the complexity of these observations, one query that requires clarification is how disruption of microtubules transduces the activation of these signaling cascades. It is estimated that 40% of MAPKs are associated with microtubules (Reszka et al., 1995), presumably making the kinases more accessible to substrates and activators. Furthermore, ERKs have been localized to the kinetochores in mitotic HeLa cells (Shapiro et al., 1998), are involved in the regulation of the microtubule organizing center during metaphase, and function in spindle assembly in Xenopus laevis egg extracts (Verlhac et al., 1993). MAPKs also regulate cyclin D1 promoter activity and protein expression (Lavoie et al., 1996), and phosphorylate cyclin B, resulting in cyclin B/cdc2 translocation to the nucleus, where this complex is dephosphorylated and activated by cdc25c. Overall, these various activities make MAPK a likely regulator of mitotic cell progression. Indeed the activation of ERK that we observe in HeLa cells coincides with increasing proportions of cells in the G₂M phase of the cell cycle after paclitaxel treatment.

JNK is perceived as a mediator of stress signaling, and models have been proposed in which microtubule disruption induces JNK via upstream signaling components involving ASK1 and MKK7. The ASK1/JNK pathway is normally activated in the G₂M phase of the cell cycle in Jurkat cells and is thought to mediate paclitaxel-induced bcl-2 phosphorylation (Yamamoto et al., 1999). In ovarian carcinoma cells, a biphasic activation of JNK in response to paclitaxel has been observed; however, neither phase of JNK activity mediates paclitaxel-induced bcl-2 phosphorylation (Wang et al., 1999). The same study concludes that the majority of paclitaxel-induced cell death is independent of JNK activity. Indeed, inhibition of JNK signaling in HeLa cells (this report) using a catalytically inactive dominant-negative mutant of JNK1 does not modulate the degree of paclitaxel-induced cell death.

The data described here, and from other studies, report activation of components of MAPK families that coincides with both mitotic arrest and an increased proportion of cells undergoing cell death. Hence, the interpretation of data regarding the effect of microtubule inhibitors on MAPK is confounded by the dual effect of these drugs (that is, by the induction of both mitotic arrest and cell death, both of which are probably regulated by MAPK activity). The ERK, the JNK, and the p38 kinases have been implicated in the regulation of apoptosis and of proliferation and differentiation, depending on the cell type and stimulus. In some cell systems, there is good correlation between ERK activation and the proliferation of cells, as is observed with epidermal or platelet-derived growth factors (Seger and Krebs, 1995). Inhibition of ERKs by antisense oligonucleotides or dominant-negative Raf-1 kinase inhibits cellular proliferation, whereas activation of ERKs may provide protection against apoptosis in other cell types (Widmann et al., 1999). Conversely, in some cell systems, apoptosis is associated with ERK activation; for example, in Jurkat cells, ERKs are transiently activated after Fas stimulation (Widmann et al., 1998b). A significant observation made in HeLa cells in response to paclitaxel are that alterations in MAPK activity relate to the increased accumulation of cells in mitosis, in contrast to the rapid responses observed in growth factor mediated signaling. Others and we have documented the phosphorylation of Raf-1 in response to mitotic arrest, an observation that is distinct from the Raf-1 response to growth factors. Furthermore, the inhibition of ERK in HeLa cells does not prevent paclitaxel-induced Raf-1 phosphorylation (data not shown); hence, Raf-1 phosphorylation during exposure to microtubule inhibitors is not transduced through the classic ERK/MEK pathway.

Despite the positive correlation between phospho-ERK activity and increased proportions of hypodiploid and annexin-V-positive cells, and PARP cleavage, the inhibition of the ERK pathway by specific inhibitors of MEK did not prevent paclitaxel-induced cell death but in fact potentiated cell death. This observation led us to analyze the nature of the interaction between paclitaxel and U0126 using the combination index method of Chou and Talalay (1984) in three cell lines. In cell lines that exhibited paclitaxel-induced ERK activation (A549 and HeLa), this drug combination was additive. Conversely, in MCF-7 cells that have low activation levels of the ERK/MEK pathway and do not exhibit activation in response to paclitaxel treatment, the nature of the interaction between paclitaxel and U0126 is antagonistic, presumably because of low substrate concentrations for U0126. A recent report has described findings that suggest that this drug combination is greater than additive (MacKeigan et al., 2000) and the data described in our study suggest additivity between paclitaxel and MEK inhibition, although this interaction is limited to those cells lines expressing high levels of activated ERK.

This study indirectly supports a role for ERK in survival signaling that has been previously documented (Widmann et al., 1999). The mechanism by which ERK inhibition potentiates paclitaxel-induced apoptosis remains to be delineated, although we speculate that inhibiting the proliferative signaling function of ERK renders cells more susceptible to the toxic effects of paclitaxel. In addition, the inhibition of MEK1/2 by U0126 may regulate the activity of the JNK pathway via MEKK1, which is an activator of both the ERK and JNK pathways (Lange-Carter, 1993). We have demonstrated altered MEKK1 expression in both CHO and HeLa cells after paclitaxel treatment, although modification of JNK activity by a dominant negative mutant has no effect on paclitaxel-induced cell death. It should also be considered that MEKK1 activity might regulate the activity of other proteins that affect cell death, such as nuclear factor-B. Indeed, it has been shown that paclitaxel significantly down-regulates I-B, thereby increasing the nuclear translocation of nuclear factor-B (Huang et al., 2000). There also exists the possibility that ERKs directly modulate the expression of Bcl-2 family members via phosphorylation of Bad (Scheid et al., 1999) or down-regulation of Bcl-2, Bcl-X_L, and Mcl-1 (Boucher et al., 2000). It remains to be determined whether these are plausible mechanisms for the potentiation of paclitaxel-induced cell death by U0126.

In conclusion, we speculate that this drug combination would be attractive only as a potential chemotherapeutic combination in tumors that have confirmed high levels of ERK. Various tumor types have been reported to have high levels of ERK/MEK activity (Loda et al., 1996; Schmidt et al., 1997), although there is also a high degree of heterogeneity within a tumor. Hence, it must be considered that for some patients, this drug combination will not enhance tumor cell death compared with paclitaxel alone, with a detrimental effect of increased toxicity from two drugs in combination. This study, however, does endorse the importance of targeting signal transduction pathways aberrantly activated in human cancer. This strategy can be used to modulate the efficacy of cytotoxic drugs used currently in chemotherapy by the addition of cytostatic agents, as in the case of herceptin (Slamon et al., 2001) and epidermal growth factor receptor (Sirotnak et al., 2000), both of which are being evaluated clinically with paclitaxel. Unfortunately, both commercially available MEK inhibitors, PD98059 and U0126, are cytotoxic at high concentrations; consequently, any potential clinical applications for this drug combination will require the development of effective MEK/ERK inhibitors with low toxicity for use in clinical medicine and will need extensive in vivo testing.