Induction of survivin expression by taxol (paclitaxel) is an early event, which is independent of taxol-mediated G2/M arrest.

Survivin is a novel anti-apoptotic protein that is highly expressed in cancer but is undetectable in most normal adult tissues. It was reported that taxol-mediated mitotic arrest of cancer cells is associated with survivin induction, which preserves a survival pathway and results in resistance to taxol. In this study, we provide new evidence that induction of survivin by taxol is an early event and is independent of taxol-mediated G(2)/M arrest. Taxol treatment of MCF-7 cells rapidly up-regulated survivin expression (3.5-15-fold) within 4 h without G(2)/M arrest. Lengthening the treatment of cells (48 h) with taxol resulted in decreased survivin expression in comparison with early times following taxol treatment, although G(2)/M cells were significantly increased at later times. Interestingly, 3 nm taxol induces survivin as effectively as 300 nm and more effectively than 3000 nm. As a result, 3 nm taxol is ineffective at inducing cell death. However, inhibition of taxol-mediated survivin induction by small interfering RNA significantly increased taxol-mediated cell death. Taxol rapidly activated the phosphatidylinositol 3-kinase/Akt and MAPK pathways. Inhibition of these pathways diminished survivin induction and sensitized cells to taxol-mediated cell death. A cis-acting DNA element upstream of -1430 in the survivin pLuc-2840 construct is at least partially responsible for taxol-mediated survivin induction. Together, these data show, for the first time, that taxol-mediated induction of survivin is an early event and independent of taxol-mediated G(2)/M arrest. This appears to be a new mechanism for cancer cells to evade taxol-induced apoptosis. Targeting this survival pathway may result in novel approaches for cancer therapeutics.

Survivin is a novel anti-apoptotic protein that is highly expressed in cancer but is undetectable in most normal adult tissues. It was reported that taxol-mediated mitotic arrest of cancer cells is associated with survivin induction, which preserves a survival pathway and results in resistance to taxol. In this study, we provide new evidence that induction of survivin by taxol is an early event and is independent of taxol-mediated G 2 /M arrest. Taxol treatment of MCF-7 cells rapidly up-regulated survivin expression (3.5-15-fold) within 4 h without G 2 /M arrest. Lengthening the treatment of cells (48 h) with taxol resulted in decreased survivin expression in comparison with early times following taxol treatment, although G 2 /M cells were significantly increased at later times. Interestingly, 3 nM taxol induces survivin as effectively as 300 nM and more effectively than 3000 nM. As a result, 3 nM taxol is ineffective at inducing cell death. However, inhibition of taxolmediated survivin induction by small interfering RNA significantly increased taxol-mediated cell death. Taxol rapidly activated the phosphatidylinositol 3-kinase/Akt and MAPK pathways. Inhibition of these pathways diminished survivin induction and sensitized cells to taxol-mediated cell death. A cis-acting DNA element upstream of ؊1430 in the survivin pLuc-2840 construct is at least partially responsible for taxol-mediated survivin induction. Together, these data show, for the first time, that taxolmediated induction of survivin is an early event and independent of taxol-mediated G 2 /M arrest. This appears to be a new mechanism for cancer cells to evade taxol-induced apoptosis. Targeting this survival pathway may result in novel approaches for cancer therapeutics.
Survivin is a recently characterized novel member of the inhibitor of apoptosis (IAP) 1 protein family. It is undetectable in most normal adult tissues but highly expressed in cancer. Survivin expression has been shown to be associated with carcinogenesis, cancer progression, poor prognosis, drug resistance, and short patient survival (1) and that inhibition of survivin expression and/or function in tumor cells by survivin antisense or dominant-negative mutants triggers apoptosis (2)(3)(4)(5)(6) as well as a defect in cell division (7,8). Thus, survivin is considered an exciting target for cancer prevention and therapeutics.
Taxol (paclitaxel) is one of the most active cancer chemotherapeutic agents. It is effective against a variety of human tumors including ovarian, breast, and non-small-cell lung tumors as well as head and neck carcinomas (9 -13). However, its effectiveness is often limited because many tumors display taxol resistance. Cancer cells can acquire resistance to taxol by at least two different mechanisms (14). Overexpression of the multidrug resistance 1(MDR1) gene, which encodes P-glycoprotein, can confer resistance to taxol. This is because P-glycoprotein functions as a xenobiotic pump that pumps taxol as well as many other chemotherapeutic agents out of cells (15). The other is that tubulin mutations, which result in alterations in either the assembly or stability of microtubules, can lead to taxol resistance (16,17). However, taxol resistance resulting from apoptotic blockade has not been well studied.
Taxol treatment induces mitotic arrest through taxol-induced polymerization and stabilization of microtubules (18 -22), and it induces cell death by apoptosis or necrosis dependent of drug concentration (23)(24)(25)(26). On the other hand, it has been demonstrated that survivin expression is cell cycle-regulated with a robust increase in the G 2 /M phase of cell cycle (2,27). Presumably, cells treated with taxol should show an increased survivin expression because of G 2 /M arrest. Consistent with this notion, it was reported that taxol-induced microtubule stabilization and mitotic arrest increase survivin expression, which engenders a cell survival pathway to counteract taxol-induced apoptosis (28). However, it is not clear whether this mitotic survival pathway is the only means involving survivin by which cancer cells counteract taxol-induced apoptosis following drug treatment. Here, we report that induction of survivin by taxol is an early event and is independent of taxolmediated G 2 /M arrest. We found that taxol treatment of MCF-7 breast cancer cells rapidly up-regulated survivin expression without apparent arrest of cells into G 2 /M phase. Lengthening the treatment of cells (48 h) with taxol resulted in decreased survivin expression in comparison with early times following taxol treatment, although the percentage of cells in G 2 /M phase was significantly increased at later times. Consistent with the observation that 3 nM taxol induced survivin as effectively as 300 nM and more effectively than 3000 nM taxol, 3 nM taxol are ineffective for apoptotic induction in these cells. However, inhibition of taxol-mediated induction of survivin by small interfering/inhibitory RNA (siRNA) significantly increased taxolmediated cell death. Mechanistic studies indicated that taxol rapidly activated the PI3K/Akt and Erk MAPK pathways, and inhibition of PI3K/Akt signaling by Ly294002 or Erk MAPK signaling by U0126/PD98059 diminished survivin induction by taxol and sensitized cells to taxol-induced cell death. Survivin promoter-luciferase reporter assays revealed that early taxolmediated induction of survivin is at least in part transcriptionally regulated and that the cis-acting DNA element mediating the effects of taxol on survivin promoter activity is located upstream of Ϫ1430 in the pLuc-2840 construct. Together, these data show for the first time that induction of survivin by taxol is an early event following drug exposure and is independent of taxol-mediated G 2 /M arrest. This appears to be a new mechanism by which cancer cells evade taxol-induced apoptosis. Targeting this novel survival pathway may have potential in cancer therapeutic applications.
Reagents-Taxol, anti-␤-actin, goat peroxidase-conjugated anti-rabbit IgG and fluorescein isothiocyanate-conjugated anti-rabbit IgG were purchased from Sigma. Anti-survivin (FL-142), anti-Erk (K-23), and anti-Akt (H-136) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Phospho-Akt (Ser-473) and phospho-p44/42 MAPK (Thr-202/Tyr-204) antibodies were purchased from Cell Signaling Technology (Beverly, MA). LY294002, PD98059, U0126, and dual luciferase report assay system were purchased from Promega (Madison, WI). Oligotransfectamine reagent was purchased from Invitrogen. Western Blot-Cells with and without treatment were washed with phosphate-buffered saline (PBS: 50 mM phosphate, pH 7.4, 100 mM NaCl, and 10 mM KCl) and lysed on ice for 30 min in PBS containing 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 10 g/ml phenylmethylsulfonyl fluoride, and 20 M leupeptin. After the lysates were cleared by centrifugation at 15,000 ϫ g for 20 min at 4°C, the total protein was determined using Bio-Rad protein assay solution. Up to 75 g of total protein was denatured in 2ϫ SDS sample loading buffer for 5 min at 95°C, separated on 10 -15% SDS-PAGE gels, and electrotransferred to Immobilon-P membranes (Millipore, Bedford, MA) using semidry electrophoretic transfer. After the nonspecific binding sites on the membranes were blocked with 5% skimmed milk or bovine serum albumin (BSA) in TBS-T (20 mM Tris-HCl, pH 7.5, 0.137 M NaCl, and 0.01% Tween 20) for 3 h at room temperature with constant shaking, the membranes were incubated in TBS-T containing the relevant primary antibody (1:500 -1000) and 5% BSA overnight at 4°C. After washing with TBS-T for three times, the membrane was incubated in 5% skimmed milk in TBS-T buffer containing the appropriate second anti-IgG antibody (1:5000) at room temperature for 1 h with constant shaking. The phosphorylation or expression of the target protein was detected by the ECL protein detection kit (Amersham Biosciences) following the manufacturer's instructions and visualized by autoradiography with the Kodak X-Omat AR film. For normalization of protein loading, the same membranes were stripped with stripping buffer (100 mM 2-mercaptoethanol, 2% SDS, and 62.5 mM Tris-HCl, pH 6.7) and used for Western blot by the same procedure with a monoclonal antibody against ␤-actin (1:1000 dilution) and/or with the polyclonal antibodies against the relevant total protein.
Trypan Blue Exclusion Staining for Determination of Cell Viability-Cells to be counted were collected by trypsinization/centrifugation and resuspended in PBS buffer. A small sample of the cell suspension was diluted in 0.4% (w/v) trypan blue (one sample at a time because viable cells absorb trypan blue over time as well). A cover glass was centered over the hemacytometer chambers, and one chamber was filled with the cell dilution using a Pasteur pipette. Stained (dead) and unstained (viable) cells were counted in each of the four corner and central squares under an inverted microscope using ϫ100 magnification, respectively. Each cell sample was counted in this way for three times. The percentage of cell viabilities in each sample was calculated with the formula of % viability ϭ total viable cell numbers/total cell numbers ϫ 100.
Propidium Iodide Staining and Flow Cytometry-At various time intervals after taxol treatment (30 nM) as described above, the cells were harvested by trypsinization and washed with PBS. Cells (ϳ1 ϫ 10 6 ) were resuspended in 5 ml of 70% ethanol. After the initial fixation, cells were suspended in 0.5 ml of PBS containing 25 g/ml propidium iodide, 0.2% Triton X-100, and 40 g/ml RNase A and incubated for at least 30 min at 4°C. The cells then were analyzed for immunofluorescence intensities by flow cytometry (FACScan, BD Biosciences) from 10,000 events/sample. Data from flow cytometry were analyzed using WinList software (Verity Software House Inc., Topsham, ME). For each time point, triplicate assays were performed.
Immunofluorescence Microscopy-Cells were seeded on the round glass coverslips coated with 2% gelatin (Sigma) in 12-well plates. At different intervals after 3 nM taxol treatment as described above, medium was removed and cells were washed once with PBS and fixed overnight with 4% paraformaldehyde in PBS at 4°C. Cells were then permeabilized and blocked for 30 min in PBS containing 2% BSA and 0.2% Triton X-100. For survivin and nuclear DNA double staining, the blocked cells were first incubated in PBS containing 1% BSA and rabbit anti-survivin antibody (1:500) for 60 min at 37°C followed by fluorescein isothiocyanate-conjugated anti-rabbit IgG (1:200) for 30 min at 37°C. The nuclear DNA was then stained with 4Ј,6-diamidino-2-phenylindole at a final concentration of 0.5 g/ml in H 2 O for 10 min at room temperature. The resultant glass coverslips containing cells were mounted on glass slides with Gel/Mount TM solution (Biomedia Corp., Foster City, CA). The cells were analyzed under a Zeiss Axiovert 100M digital fluorescence microscope. Images were captured using Zeiss LSM510, version 2.8 and processed with Photoshop Element software. SiRNA Preparation-A human survivin mRNA-specific RNA oligonucleotides with 3Ј-TT overhangs were chemically synthesized and purified by high pressure liquid chromatography (Xeragon, Huntsville, AL) as follows: SRi-2F ( 92 GCG CCU GCA CCC CGG AGC G 110 TT*) and SRi-2R ( 110 CGC UCC GGG GUG CAG GCG C 92 TT). Equal moles of SRi-2F/SRi-2R (designated SRi-2) were mixed together to a final concentration of 20 M in annealing buffer (100 mM KAc, 30 mM HEPES-KOH, and 2 mM MgAc 2 , pH 7.4). After denaturation at 90°C for 1 min, the mixture (Sri-2) was annealed at 37°C for 60 min and stored at Ϫ80°C for transfection experiments. A scramble RNA duplex (designated scraSRi) was also prepared same as above for a negative control in this study. The scramble sequence (5Ј-CAG UCG CGU UUG CGA CUG GTT-3Ј (forward chain) and 5Ј-CCA GUC GCA AAC GCG ACU GTT-3Ј (reverse chain)) was not present in mammalian cells by BLAST search at NCBI.
In Vitro Transfection with siRNAs-Cells were transfected with survivin siRNAs using the Oligofectamine reagent following the manufacturer's instruction. 1 day prior to transfection, 5 ϫ 10 4 MCF-7 cells/well were seeded in six-well plates (corresponding to a density of 40% at the time of transfection) without antibiotics. The transfection mixture was prepared by mixing 175 l of DMEM containing 6 l of 20 M siRNA with 15 l of DMEM containing 3 l of Oligofectamine reagents. Before transfection, the medium in 6-well plates was replaced with serum-free DMEM medium (800 l/well). The transfection mixture was added to the 6-well plate within 20 -40 min after mixture preparation in a total volume of 990 l/well. The transfected cells were incubated at 37°C for 4 h, and then 500 l of DMEM medium containing 30% fetal bovine serum was added. Cells were treated with and without 3 and 30 nM taxol 24 h after transfection with siRNA or control siRNA as described above. For Western blot, cells were harvested 8 h after taxol treatment. For flow cytometry analysis of sub-G 1 DNA content, cells were harvested 24 and 48 h after taxol treatment and analyzed by propidium iodide staining and flow cytometry as described above. The percentage of dead cells (sub-G 1 DNA contents) was plotted as a histogram. All of the transfection experiments were performed at triplicate for each experiment.
Luciferase Reporter Assay-MCF-7 cells were transfected with survivin promoter-luciferase constructs pLuc-6270, which contains a 6.3-kb survivin promoter sequence (29), using the FuGENE 6 transfection reagent (Roche Diagnostics), and luciferase activities were measured using a dual luciferase reporter assay system (Promega) according to the respective manufacturers' recommendations. Cells were seeded in 24-well plates 1 day prior to transfection. On the following day, the transfection solution for each well was prepared by sequentially adding 1 l of FuGENE 6 transfection reagent and 0.4 g of plasmid DNA in a 1.5-ml tube containing 50 l of serum-free DMEM medium. The DNA Cells were analyzed by flow cytometry and immunofluorescence as described under "Experimental Procedures." A, cell cycle distribution was determined by propidium iodide (PI) staining and flow cytometry after taxol treatment at various time points and shown as a histogram. Each bar represents the mean Ϯ S.D. from a representative experiment in triplicate. Cell viability determined by trypan blue exclusion in each time point is shown in Fig. 1B. B, rapid induction of survivin expression by taxol was determined by immunofluorescence microscopy. As shown, survivin expression in non-mitotic cells was significantly increased 4 h after taxol treatment in comparison with no taxol treatment control. C, time course of induction of cyclin B1 expression after taxol treatment. The relative-fold increase of cyclin B1 protein after normalization to ␤-actin internal control is indicated.

Induction of Survivin by Taxol Is an Early Event and Is
Independent of taxol-mediated G 2 /M Arrest-It was previously shown that taxol (2-200 nM) treatment of HeLa cells for 48 h increases survivin expression to 1.2-4-fold in taxol-mediated G 2 /M arrest (28). To closely investigate the relationship between taxol-mediated G 2 /M arrest and survivin induction, MCF-7 cells were treated with taxol at various concentrations and times as shown (Fig. 1A) and survivin expression was analyzed by Western blot. To our surprise, lengthening taxol treatments (48 h or more) significantly attenuated survivin induction in comparison with the early times (24 h, Fig. 1A). The high induction of survivin at 24 h suggested that induction of survivin expression might occur at even earlier times. Consistent with this notion, Western blot experiments to test early time points after taxol treatment indicated that taxol-mediated survivin induction is as early as 4 h and that extending treatment times beyond 24 h diminished the effect of taxol on survivin induction in comparison with earlier time points (Fig.  1B). This result strongly suggested that induction of survivin by taxol is probably independent of taxol-mediated G 2 /M arrest. To confirm this possibility, we determined cell cycle dis-tribution by propidium iodide staining and flow cytometry after taxol treatment. As shown ( Fig. 2A), there is no apparent G 2 /M cell population increase at the 4 and 8-h times of taxol treatment in comparison with the no taxol control. A significant increase of G 2 /M cell population was observed only after taxol treatment for 16 h or more ( Fig. 2A). We next examined individual cells by survivin immunofluorescence microscopy after taxol treatment for 4 h. The results indicated that interphase cells showed increased immunoreactivity with anti-survivin antibody in comparison to no taxol treatment control. A representative experiment from the immunofluorescence microscopy study is shown in Fig. 2B. The G 2 /M phase-independent induction of survivin expression by taxol was further confirmed by monitoring the expression of cyclin B1 (a G 2 /M phase marker) after taxol treatment. There was no change in cyclin B1 expression 4 h after taxol treatment (Fig. 2C). Interestingly, the cyclin B1 protein was almost undetectable 48 h after taxol treatment (Fig. 2C), suggesting that the cells were synchronized in later anaphase.
Low Concentrations (3-30 nM) of Taxol Effectively Induced Survivin Expression but Were Ineffective for Apoptosis Induction-Unexpectedly, the induction of survivin expression by taxol at low concentrations (3 nM) was as effective as or more effective than high concentrations (3000 nM) (Fig. 3A). Consistent with the observation that high expression of survivin increases cell viability and engenders drug resistance (28,30,31), low concentrations (3-30 nM) of taxol were ineffective for induction of cell death in comparison with high concentrations of the drug (Fig. 3B).
Inhibition of Taxol-mediated Induction of Survivin Expression by siRNA-sensitized Taxol-induced Cell Death-We employed a survivin siRNA approach (32) for inhibition of taxolmediated survivin induction to determine the effect of survivin inhibition on taxol-induced cell death. Fig. 4A shows the structures of siRNAs used in this study. Consistent with the result that survivin siRNA significantly inhibited the taxol-mediated induction of survivin in MCF-7 cells (Fig. 4B), propidium iodide staining and flow cytometry analysis indicated that a combination of siRNA-targeting survivin and a low concentration of taxol treatment strikingly increased cell death (the sub-G 1 DNA content increase) in comparison with either treatment alone (Fig. 4C). This observation suggests that survivin expression plays a critical role in cell viability and that induction of survivin by taxol is a potential drug resistance factor leading to cell survival.
Taxol Treatment Rapidly Enhanced Akt and Erk1/2 Phosphorylation in MCF-7 Cells-It has been shown that activation of the Akt survival pathway can up-regulate survivin expression (33,34) and inhibition of Akt and Erk1/2 activation can block growth factor-mediated induction of survivin (35). To delineate the underlying mechanism by which taxol rapidly induced survivin expression, the activation of Akt and Erk1/2 after taxol treatment was examined in MCF-7 cells. Interestingly, the phosphorylation of both Akt and Erk1/2 were rapidly and strongly increased after taxol treatment of these cells (Fig.  5).

Inhibition of Taxol-mediated PI3K/Akt Signaling by LY294002 or MEK/Erk Signaling by U0126/PD98059 Diminished Taxol-mediated Survivin Induction and Enhanced Cell
Death-We examined whether inhibition of taxol-mediated PI3K/Akt signaling by the PI3K inhibitor, LY294002, diminished taxol-mediated survivin induction. Taxol-induced up-regulation of survivin was significantly decreased by the addition of LY294002 (Fig. 6A). Similarly, taxol-mediated induction of survivin was strongly inhibited in the presence of MEK inhibitors, U0126 or PD98059 (Fig. 6B). To explore whether inhibition of taxol-mediated Akt and Erk1/2 activation would enhance taxol-mediated apoptosis, we examined the combination of taxol treatment with either PI3K inhibitor or MEK inhibitor on cell death. The results indicated that their combination significantly increased the effectiveness of taxol to initiate cell death (Fig. 6C).

Taxol-mediated Induction of Survivin Expression Is at Least in Part Transcriptionally Regulated, and a Cis-acting DNA Element Upstream of Ϫ1430 in pLuc-2840 Is Responsible for
This Event-To explore whether taxol-mediated induction of survivin expression is transcriptionally regulated, the survivin promoter-luciferase construct pLuc-6270 containing a 6.3-kb survivin promoter sequence (29)  porter assay system (Promega) and normalized to Renilla luciferase activities. Consistent with the results from Western blot (Figs. 1 and 3A), taxol increased the pLuc-6270 luciferase activity after treatment (Fig. 7A). To locate the cis-acting DNA element responsible for the effect of taxol on survivin promoter activity, a series of survivin promoter-luciferase constructs containing different length of survivin promoter sequences (29) were transfected into MCF-7 cells and treated with 30 nM taxol for 0, 8, and 26 h as indicated 24 h after transfection (Fig. 7B). As shown, a strong induction of luciferase activity was found in the pLuc-2840 construct but not in pLuc-1430c or smaller constructs (Fig. 7B). These data strongly indicated that the cisacting DNA element responsible for the effect of taxol on in-duction of survivin promoter activity is upstream of Ϫ1430 in the pLuc-2840 construct.

DISCUSSION
It was previously demonstrated that taxol-mediated mitotic arrest in HeLa cells was associated with up-regulation of survivin expression (up to 4-fold), which engenders a survivin-dependent survival pathway (28). This finding appears to be consistent with the previous observation that the expression of survivin is cell cycle-dependent with a robust increase at the G 2 /M phase of cell cycle (2,27). However, in this study, we concluded that taxol-mediated induction of survivin is a very early event and is independent of taxol-induced mitotic arrest and that the rapid up-regulation of survivin by taxol generates survival signaling that blocks taxol-induced cell death. This conclusion was supported by several observations. First, induction of survivin by taxol occurred as short as 4 h without increases of G 2 /M cell population analyzed in flow cytometry. This was confirmed both by survivin immunofluorescence microscopy studies to inspect individual cells and by Western blot to monitor the G 2 /M indicator cyclin B1 expression after taxol treatment. Second, survivin induction by taxol was significantly diminished after a full mitotic arrest following longer treatment (48 h or more). Third, inhibition of taxol-mediated induction of survivin by survivin-specific siRNA significantly increased taxol-mediated cell death. Finally, consistent with the finding in this paper that taxol treatment activated Akt and MAPK signaling, suppression of taxol-mediated Akt and Erk activation by pharmacological inhibitors decreased taxolmediated survivin induction and increased taxol-induced cell death.
To our knowledge, this is the first report showing that taxol can induce a G 2 /M phase-independent induction of survivin, which is essential for cell survival and resistance to taxol-induced cell death. This finding extended our current vision of the mechanistic function of survivin to the conservation of a cell survival pathway during taxol treatment (5,28,31) and provided new insight into the mechanisms by which cancer cells activate cell survival pathways and initiate taxol resistance. It was reported that taxol induces prolonged activation of the Ras/MEK/ERK pathway independently of activating the programmed cell death machinery in human esophageal squamous cancer cell lines (36), but the significance of the ERK activation during taxol treatment was not elucidated in their study. Although our finding that taxol treatment rapidly activates Erk1/2 in MCF-7 breast cancer cells is consistent with the earlier observation, we extended those studies and demonstrated that ablation of taxol-activated MEK/Erk signaling by pharmacological MEK inhibitors PD98059 and U0126 reduces taxol-mediated survivin induction and enhances taxol-induced cell death in MCF-7 cells. These findings also provide an explanation for previous observation that U937 leukemia cells treated with taxol followed by PD98059 exhibited a significant increase in cytochrome c release, caspase activation, poly(ADPribose) polymerase cleavage, and apoptosis (37). In this study, the role of survivin in taxol-mediated cell survival was also independently confirmed by directly targeting survivin with siRNA (32), 2 i.e. the inhibition of taxol-induced survivin induction by siRNA increased taxol-induced cell death. In addition, in this report, we also found that taxol treatment rapidly activates Akt signaling in addition to ERK signaling and inhibition of taxol-induced PI3K/Akt signaling by PI3K inhibitor Ly294002 also decreases taxol-mediated survivin induction and enhances cell death. This finding further extended the current understanding of taxol-mediated cell survival (taxol resistance) and apoptosis. Interestingly, we have observed that taxol-activated Akt phosphorylation was significantly reduced at 12 h and increased again at 24 h after taxol treatment (Fig.  5), although the significance of this phenomenon is not clear and warrants further investigation.
The most significant finding in our report is the observation that induction of survivin expression by taxol is an early event and is independent of taxol-mediated microtubule stabilization and mitotic arrest. Moreover, this finding is consistent with and supported by a number of previous reports. It was shown that taxol could induce inflammatory and apoptotic gene expression that is independent of microtubule stabilization (38) and that abrogation of ERK signaling by MEK inhibitor PD98059 induced a cell cycle-independent apoptosis in vinblastine-treated human ML-1 myeloid leukemia cells (39). In this regard, one explanation to our finding that inhibition of taxolmediated survivin induction increased cell death might be that loss of survivin shifts the balance of expression/activation of anti-and pro-apoptotic factors toward initiation of apoptosis. Regardless, studies of the underlying mechanism by which induction of survivin by taxol prior to mitosis preserves a cell survival pathway may have significant consequences for cancer therapeutics and are under investigation in our laboratory.
A recent report demonstrated that in taxol-treated MCF-7 breast cancer cells, although mitochondrial cytochrome c release and caspase-9 cleavage were detectable by immunoblotting, assays of both cytosol and nuclei failed to demonstrate the presence of caspase activities that cleaved the synthetic caspase substrates (40). One possibility for the lack of correlation between caspase activation (itself cleavage) and caspase activity is that the rapidly taxol-induced survivin expression found in this study may directly interact with taxol-activated caspases to neutralize caspase activities. However, direct evidence for survivin to inhibit activated caspases is currently lacking. In addition, direct inhibition of effector caspases such FIG. 7. Identification of the cis-acting DNA elements responsible for the induction of survivin promoter activity by taxol. MCF-7 cells were transfected with the survivin promoter-luciferase construct pLuc-6270 (A) or with a complete set of survivin promoterluciferase constructs (B) as shown and treated with taxol at the indicated time points 24 h after transfection. Cells were lysed, and luciferase activities were measured using the dual luciferase reporter assay system as described under "Experimental Procedures." A, taxol treatment increased survivin promoter activities. B, the cis-acting DNA element upstream of Ϫ1430 in the pLuc-2840 construct is responsible for the effect of taxol on survivin promoter activity (see "Results" in detail). Each column/bar represents the relative survivin promoter (luciferase) activity and is the mean Ϯ S.D. from a representative experiment in triplicate. as caspase-3 and -7 by survivin is controversial (1). Whereas some reports showed that survivin did not inhibit effector caspases (41,42), others showed that it did (43,44). Interestingly, a very recent study showed that survivin interacts with mitochondria-released Smac/DIABLO initiated by taxol treatment (45). These authors demonstrate in a cell-free cytosol system that survivin inhibited caspase-9 and -7 activation only in the presence of X chromosome-linked IAP. They proposed that the interaction of survivin with Smac/DIABLO would release X chromosome-linked IAP from Smac to inhibit caspase activation. This is a plausible alternative model to support our finding in this report that inhibition of taxol-mediated rapid induction of survivin by siRNA or by pharmacological inhibitors increased apoptosis induced by taxol. However, to support their model, it is important to demonstrate that survivin has significant higher affinity to Smac/DIABLO than X chromosome-linked IAP in the presence and/or absence of taxol treatment.
The last issue we would like to mention is the potential transcriptional mechanism that is involved in taxol-mediated induction of survivin expression. Transfection of a series of survivin promoter-luciferase constructs containing different lengths of survivin promoter sequences (29) into MCF-7 cells followed by luciferase assays after taxol treatment indicated that a cis-acting DNA element upstream of Ϫ1430 in the pLuc-2840 construct is responsible for the effect of taxol on transcriptional up-regulation of survivin (Fig. 7B). Consistent with this observation, previous reports show that transfection of HeLa cells with the survivin promoter-luciferase construct pLuc-1430c (28) or of MCF-7 cells with the survivin minimal promoter-luciferase construct pLuc-cyc1.2 (containing 267 bp of survivin proximal promoter) (31) failed to detect the induction of survivin promoter activity by taxol. Moreover, we found that taxol down-regulated survivin promoter activities when using the pLuc-1430c construct or the shorter survivin-promoter constructs (Fig. 7B). However, the promoter activity increase (1.5-3-fold) induced by taxol does not fully account for the levels of induction of survivin protein by taxol found in Western blots (3-15-fold) in this study. This suggests that a potential posttranscriptional mechanism may also be involved in taxol-mediated survivin induction. Previous studies show that the phosphorylation of survivin on threonine at position 34 by CDC2 kinase stabilizes survivin protein (28). However, this mechanism is unlikely for the results reported here because survivin protein was increased after taxol treatment as short as 4 h without a corresponding increase of the cyclin B1 and the G 2 /M cell population (Fig. 2) and taxol-mediated induction of survivin protein decreased after a taxol-mediated G 2 /M arrest for 48 h in comparison with the early time points following taxol treatment. It was shown that CDC2 kinase has its maximum activity at this point (28). Therefore, further investigation of alternative mechanisms is underway in our laboratory.
In summary, although it is known that taxol treatment induces cells to arrest at mitosis (18), we have found that taxolmediated induction of survivin occurs earlier in the cell cycle and is independent of taxol-mediated mitotic arrest. The rapid induction of survivin by taxol appeared to increase cell viability and resistance to apoptosis induced by taxol. Targeting this survival pathway may result in novel approaches for cancer therapeutics.