Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Recent evidence has indicated that arsenic trioxide (As₂O₃) is able to induce clinical remissions in patients with acute promyelocytic leukemia (APL) (Chen et al., 1997; Look, 1998; Shao et al., 1998; Soignet et al., 1998). In addition, several investigators have shown that As₂O₃ induces programmed cell death in APL cell lines with APL/RA proteins. Although the mechanisms by which As₂O₃ induces apoptosis remain to be further elucidated, some evidence indicated that this compound can down-regulate Bcl-2 protein and activate caspase-3-like caspase activity (Chen et al., 1996, 1998; Kroemer and de The, 1999). Some investigators have shown that As₂O₃-induced apoptosis is caused by a direct effect on the mitochondrial permeability transition pore and loss of the mitochondrial transmembrane potential (Zhu et al., 1999).

Microtubules are critical elements in a wide variety of fundamental functions, including sustained cell shape, cellular transportation of vesicles and protein complexes, and regulation of cell motility, as well as of cell division (Hyams and Lloyd, 1993; Margolis and Wilson, 1998). During mitosis, microtubules are at the highest dynamic instability for response to the formation of spindles and separation of chromosomes. Because microtubules play crucial roles in the regulation of the mitotic apparatus, disruption of microtubules can induce cell-cycle arrest in M-phase, formation of abnormal mitotic spindles, and final triggering of the signals for programmed cell death (Mitchison, 1988). Agents such as the vinca alkaloids and the taxanes, which affect the dynamics of microtubules, have emerged as useful therapeutic agents for the treatment of human cancer (Rowinsky and Donehower, 1991; Jordan et al., 1992). Recently, some reports indicated that As₂O₃ treatment resulted in cell-cycle arrest at either G₁ or G₂/M phase depending on the cell lines (Ma et al., 1998; Perkins et al., 2000). In addition, Li and Broome reported that As₂O₃ treatment resulted in cell-cycle progression arrest at metaphase in myeloid leukemia cells through noncompetitive disruption of GTP binding to -tubulin and inhibition of GTP-induced tubulin polymerization (Li and Broome, 1999).

Although a number of studies have demonstrated that As₂O₃ has potent activity against cell growth in a series of leukemia cell lines, little information is available regarding this compound's effect on cell growth in solid tumor cell lines (Seol et al., 1999; Meada et al., 2001). In this study, we investigated the effect of As₂O₃ on cell proliferation in human solid tumor cell lines, particularly breast, ovary, and lung. Moreover, we also investigated the molecular basis of As₂O₃-induced cell cycle arrest and apoptosis in human breast carcinoma MCF-7 cells and NSCLC H460 cells. Interestingly, we found that As₂O₃ treatment led to the accumulation of cells in mitosis, and the induction of apoptosis. Further studies showed that As₂O₃ caused a paclitaxel-like concentration-dependent promotion of tubulin polymerization in vitro. Morphological and EM observations showed that As₂O₃ caused changes in the cellular microtubule network.

Because As₂O₃ had a paclitaxel-like effect, it was of great interest to test whether As₂O₃ could circumvent paclitaxel resistance. The results present herein show that As₂O₃ is active in paclitaxel resistant cell lines, whether due to overexpression of P-glycoprotein, or MRP, or tubulin mutation.

Overall, our results lead to a better understanding of the unique mechanism of action of As₂O_3, and provide a suggestion that As₂O₃ may be of therapeutic value for the treatment of human solid tumors including those with acquired resistance.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture and Treatment. Human cervical HeLa cells, ovarian cancer SK-OV-03 cells, and human NSCLC cell lines, H460, H322, H520, and H661 were purchased from the American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 medium supplied with 10% fetal bovine serum. Human breast carcinoma MCF-7 cells, P-glycoprotein-overexpressing MCF-7/Dox cells, and MRP-positive MCF-7/VP-16 cells were maintained in culture as monolayers in minimal essential medium with 10% fetal bovine serum as described previously (Perez-Soler et al., 1997). Human ovarian carcinoma A2780 cells and the tubulin mutant cell lines resistant to paclitaxel (A2780/PTx10, A2780/PTx22) were kindly provided by Dr. Fojo (National Cancer Institute, National Institutes of Health, Bethesda, MD). A2780 cells were maintained in RPMI 1640 medium with 10% fetal bovine serum, A2780/PTx10 and A2780/PTx22 cells were maintained in the same medium containing 15 ng/ml of paclitaxel and 5 µg/ml of verapamil (Giannakakou et al., 1997). For determination of cell growth inhibition, cells were plated at 1 × 10⁴ cells/well in 96-well microplates and exposed to varying concentrations of As₂O₃ and paclitaxel at 37°C for 72 h. At the end of drug exposure, 10 µl of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide solution (5 mg/ml) was directly added to each cell culture well and the plates were incubated at 37°C for 2 h. The formazan converted from tetrazolium salt by viable cells was solubilized by 100 µl of lysis solution containing 20% SDS and 50% dimethyl formamide (Sigma-Aldrich, St. Louis, MO), and measured in a microplate reader (Dynex Technologies, Chantilly, VA) at 570 nm as described by Hansen et al. (1989). The IC₅₀ value resulting from 50% inhibition of cell growth was calculated graphically as a comparison with the control.

Reagents. As₂O₃ was purchased from Sigma Chemical Inc. (St. Louis, MO), and dissolved in 10% NaOH as 20 mM stock solution. Paclitaxel was purchased from Hande Tech Development USA, Inc. (Houston, TX) and dissolved in dimethyl sulfoxide as 1 mM stock solution. Monoclonal anti-cdc-2, anti-cyclin B1, and anti-Bcl-2 antibodies were purchased from Oncogene Science (Cambridge, MA). Anti--tubulin antibody was obtained from Biogenex Co. (San Ramon, CA). Anti-PARP and anti--catenin antibodies were purchased from BD PharMingen (San Diego, CA). All other chemicals were purchased from Sigma-Aldrich.

Apoptosis Assay. Cells were treated with As₂O₃ for the indicated time. The attached and detached cells were harvested from culture, and stained with DAPI. The cells with nuclear breakage were manually counted by fluorescence microscopy. For assay of DNA fragmentation, cells were treated with As₂O₃ for the indicated time. After treatment, cells were harvested from culture and lysed with lysis buffer containing 0.5% Triton X-100, 100 mM Tris-HCl, pH 7.5, and 25 mM EDTA at 4°C for 60 min. After centrifugation at 15,000g for 10 min, the supernatants were collected and treated with 100 µg/ml of RNase I and 100 µg/ml of proteinase K at 50°C for 30 min, and with phenol and chloroform. The fragmented DNA was precipitated in the presence of sodium acetate and ethanol, dissolved in Tris/EDTA buffer, and then subjected to 1% agarose gel. After electrophoresis, DNA was stained with ethidium bromide and visualized by UV light illumination.

Cell Cycle Analysis. Cells were washed with PBS and fixed with cold 75% ethanol overnight. The fixed cells were incubated with 5 µg/ml of RNase I and 1 µg/ml of propidium iodide for 6 h, and then cellular DNA content was determined in a flow cytometer. For quantitative assay of mitotic cells, after staining with Wright-Giemsa dye, at least 100 mitotic cells and those with chromosome condensation were counted.

Tubulin Assembly. Purified tubulin from calf brain was purchased from Sigma-Aldrich. The effects of As₂O₃ on the microtubule assembly-disassembly process were determined as described previously (Jiang et al., 1998b). For assembly assay, 100 µl of tubulin solution (500-600 µg of protein/ml) was mixed gently with 400 µl of reaction buffer containing 0.1 M MES, 1 mM EGTA, 0.5 mM MgCl₂, 0.1 mM EDTA, and 2.5 M glycerol at 37°C. As₂O₃ (0.5-2 mM) was added to each sample cuvet, together or 30 min before 1 mM GTP. Microtubule assembly was monitored by measuring the change of absorbance at room temperature at 350 nm every 5 min on a spectrophotometer (Ultrospec III; Pharmacia LKB, Uppsala, Sweden) until the assembly was completed. For disassembly assay, As₂O₃ or paclitaxel were added to a cuvet with polymerized microtubules according to the method as described above, and incubated in ice. Changes in absorbance were monitored at 350 nm for 30 min. An equal amount of solvent was used in control cuvet.

Immunoblot Analysis. Cells were harvested from culture, washed with PBS solution and lysed with buffer containing 50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 0.1% nonylphenoxy polythoxy ethanol, 1% SDS, 5 µg/ml of aprotinin, and 5 µg/ml of leupeptin. The protein amount in each sample was determined with a Bio-Rad DC protein assay kit (Bio-Rad, Hercules, CA). Equal amounts of lysate were loaded on a 12.5% SDS-polyacrylamide gel. After electrophoresis, proteins were transferred to a nitrocellulose membrane and probed by corresponding antibodies. The protein signals were detected by the enhanced chemiluminescence reaction system according to the manufacturer's recommendation (Amersham Biosciences, Indianapolis, IN). The quantification of protein levels in each sample was performed by laser scanner densitometer (GS-670 scanning densitometer; Bio-Rad).

Histone H1 Kinase Assay. After treatment with As₂O₃, cells were harvested, washed with PBS solution, and lysed with buffer as described above except for the addition of 1% SDS. Lysate was centrifuged at 15,000g for 10 min, and the supernatant was harvested for immunoprecipitation. Cdc2/cyclin B1 complex was immunoprecipitated by monoclonal anti-cdc2 and cyclin B1 antibodies at 0 to 4°C overnight, and cdc2/cyclin B1 kinase activity was measured by [-³²P]ATP incorporation into the substrate histone H1 as described previously (Ling et al., 1996).

Caspase Activity Assay. Cells (2 × 10⁶) were lysed with 50 µl of lysis buffer containing 50 mM HEPES, pH 7.4, 0.1% CHAPS, 1 mM DTT, and 0.1 mM EDTA at 0 to 4°C for 5 min. After centrifugation at 15,000g for 15 min, the supernatant was harvested and added to 100 µl of reaction mixture containing 50 mM HEPES, pH 7.4, 100 mM NaCl, 0.1% CHAPS, 10 mM DTT, 1 mM EDTA, 10% glycerol, and 2 mM caspase substrates, which were Ac-Asp-Glu-Val-Asp-pNA for caspase-3, Ac-Tyr-Val-Ala-Asp-pNA for caspase-1, Ac-Ile-Glu-Thr-Asp-pNA for caspase-8, and Ac-Leu-Glu-His-Asp-pNA for caspase-9. After incubation at 37°C for 120 min, caspase activity was assessed by measurement at 405 nm of pNA release from the substrate.

Immunocytochemical Study. Cells were fixed with cold methanol at 20°C for 5 min, and washed with PBS solution. Cells were blocked with 1% bovine serum albumin PBS solution for 30 min, and incubated with monoclonal anti--tubulin antibody (1:100) at room temperature for 90 min. After washing three times with PBS solution, cells were reincubated with fluorescein isothiocyanate-conjugated second antibody (1:500) in the dark room for 30 min. Cellular microtubules were observed with a Nikon PS200 fluorescence microscope.

Electron Microscopy. MCF-7 cells were treated with 3 µM As₂O₃ or with the same volume of PBS solution as control. After 12 h, cells were collected and washed twice with PBS solution. Cells were then fixed with osmium tetroxide, dehydrated with an ascending series of ethanol and propylene oxide, and embedded in Araldite. Thin sections were collected on Quick-Coat (Electron Microscopy Sciences, Fort Washington, PA), and treated with copper grids. After staining with uranyl acetate and lead citrate, samples were viewed at 80 kV on a JEOL 1200 EX electron microscope.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of As₂O₃ on Growth and Apoptosis in Human Solid Tumor Cells. The initial experiments were conducted for evaluation of As₂O₃-induced anti-proliferation in various human sold tumor cell lines. Human ovarian cancer A2780 and breast carcinoma MCF-7 cells exhibited the highest susceptibility to As₂O₃, with an IC₅₀ ~3.0 µM. Cell lines of cervical (HeLa) and ovarian carcinoma (SK-OV-03) were slightly less sensitive (IC₅₀, 4.28 to 4.61 µM). NSCLC H520, H322, and H460 cells displayed the lowest susceptibility (IC₅₀, >10 µM) (Table 1). Next, we examined whether the inhibition of cell growth by As₂O₃ could be caused by apoptotic death. MCF-7 cells were treated with 3 µM As₂O₃ for the times indicated in Fig. 6. As₂O₃-induced apoptosis was determined both by DAPI staining of cells with nuclear breakage and by DNA fragmentation. As shown in Fig. 1A, only ~2 to 5% of cells were apoptotic at 24 h of exposure, ~18% of cells were apoptotic by 36 h, and ~52% were apoptotic after 72 h of exposure. A similar time pattern of drug-induced apoptosis was found by DNA fragmentation assay using agarose gel electrophoresis (Fig. 1B).

View larger version (31K): [in this window] [in a new window]   Fig. 1.   As2O3 induces apoptosis in MCF-7 cells. A, log-phase growing cells were treated with 3 µM As2O3 for the indicated time points. After treatment, cells were harvested, fixed with 75% cold ethanol, and stained with 1 µg/ml of DAPI for 1 h. The apoptotic cells with nuclear breakage at least 100 cells were counted by fluorescence microscopy. Each point represents the mean ± S.D. of three independent experiments. B, As2O3 induces DNA laddering fragmentation. MCF-7 cells were treated with 3 µM As2O3 for the indicated time. After treatment, cells were lysed with lysis buffer, and DNA was extracted as described under Materials and Methods. DNA was dissolved in TE buffer, and subjected to electrophoresis in a 1% agarose gel. DNA laddering was visualized by UV light illumination.

Effect of As₂O₃ on Cell Cycle. As₂O₃ treatment resulted in a time-dependent accumulation of MCF-7 cells in the G₂/M phase (Fig. 2A). At time 0, only ~8% of cells were in the G₂/M phase. After 12 h of treatment, ~21% of cells were in the G₂/M phase. The value of G₂/M phase cells peaked at 24 h (~52%) and declined thereafter. Similar results were obtained with other cell lines (e.g., H460 and HeLa cells; data not shown). Microscopic examination showed that MCF-7 cells treated with 3 µM As₂O₃ for 24 h led to cell-cycle arrest at metaphase with chromosome condensation and disappearance of the nuclear envelope (Fig. 2B, b), compared with untreated cells with intact nuclei (Fig. 2B, a). These results were confirmed by DAPI staining (data not shown). For quantitative analysis of mitotic arrest, cells were stained with Wright-Giemsa, and the percentage of mitotic cells was calculated after counting at least 100 cells. As shown in Fig. 2C, ~3% of cells were at M-phase at time 0 and ~16% of cells were at M-phase at 12 h. M-phase cells reached peak value (~36%) at 24 h, and declined to ~6% after 48 h. Such a pattern for As₂O₃-induced M-phase arrest is very similar to that of the paclitaxel-induced M-phase arrest reported previously by us and others (Bhalla et al., 1993; Ling et al., 1998).

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Effect of As2O3 on cell cycle progression. A, MCF-7 cells were treated with 3 µM As2O3 for the indicated time points. After treatment, cells were fixed with 75% ethanol overnight, and incubated with 5 µg/ml of RNase I, and 1 µg/ml of propidium iodide for 6 h. DNA content was assessed by a flow cytometer. Each point represents the mean ± S.D. of three independent experiments. B, morphological examination of As2O3-induced M-phase arrest. MCF-7 cells were treated with 3 µM As2O3 or with the same volume of PBS solution as control. After 24 h of incubation, cells were harvested and stained with Wright-Giemsa. In As2O3-treated cells, typical metaphase arrest cells presented with condensed and separated chromosomes (b), whereas control cells were in interphase with intact nuclei (a). C, quantitative assessment of mitotic arrest by As2O3. MCF-7 cells were treated with 3 µM As2O3 for the indicated time. Cells were stained with Wright-Giemsa, and the mitotic cells were counted at least 100 cells. Each point represents the mean ± S.D. of three independent experiments.

Effect of As₂O₃ on Tubulin Polymerization in Vitro. Because As₂O₃ treatment markedly blocked the cell cycle in M-phase, we became interested in testing whether As₂O₃ could directly affect tubulin. To test this hypothesis, tubulin polymerization and depolymerization in vitro were studied at room temperature in a reaction mixture containing the purified tubulin and GTP, in the presence or absence of As₂O₃. The results presented in Fig. 3A, showed that As₂O₃ significantly promoted tubulin polymerization in a concentration dependent manner. Consequently, we examined whether altered GTP binding to -tubulin could cause As₂O₃-induced tubulin polymerization. Tubulin was preincubated with As₂O₃ at 0°C for 30 min, and then the tubulin polymerization reaction was started by the addition of GTP at room temperature. The results, as shown in Fig. 3B, indicated that pretreatment with As₂O₃ for 30 min did not affect GTP-promoted tubulin polymerization, indicating that As₂O₃-induced tubulin polymerization was not caused by interference with GTP binding domain on -tubulin. Finally, we tested whether As₂O₃ could, like paclitaxel, stabilize tubulin polymerization. Tubulin was polymerized in the presence of GTP at room temperature for 25 min. Then the depolymerization reaction was conducted at 0°C in the presence of As₂O₃ or paclitaxel or in the absence of both drugs as control. As shown in Fig. 3C, both As₂O₃ and paclitaxel not only prevented the depolymerization of assembled tubulin but also significantly promoted the polymerization of tubulin. In contrast, the polymerized tubulin was gradually and effectively depolymerized at 0°C in the absence of both drugs. The data clearly indicate that As₂O₃ has a paclitaxel-like effect to promote and stabilize tubulin polymerization in vitro.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Effect of As2O3 on microtubule polymerization in vitro. A, As2O3 promotes tubulin assembly. Purified bovine tubulin was incubated at 37°C in reaction mixtures containing 1 mM GTP, different concentrations of As2O3, and the same volume of solvent (10% NaOH) as control. Tubulin polymerization was determined by measuring observance at 350 nm. B, As2O3 does not affect GTP-promoted tubulin polymerization. Tubulin was pretreated with 1 mM As2O3 or with the same volume of solvent as control at 0°C for 30 min, then tubulin polymerization reaction was started by addition of 1 mM GTP at 37°C, and tubulin polymerization was measured as described above. C, like paclitaxel, As2O3 prevents against microtubule depolymerization. Tubulin was polymerized in the presence of 1 mM GTP at 37°C for 20 min. The polymerized tubulin was incubated with 1 mM As2O3, 10 µM paclitaxel, or with the same volume of solvent as control, and the depolymerization was performed at 0°C and monitored by measuring the absorbance value at 350 nm. Each point represents the mean value of two independent experiments. RT, room temperature.

Effect of As₂O₃ on Cellular Microtubules. Because strong paclitaxel-like tubulin-stabilizing activity was found in vitro, we tested whether As₂O₃ treatment affected the cellular microtubule network. MCF-7 cells were treated with 0.5 µM paclitaxel, 0.1 µM vinblastine, or 3 µM As₂O₃ or with the same volume of PBS solution as control. After 12 h of incubation, the microtubule network was visualized by immunocytochemistry. The microtubule network in control cells exhibited normal arrangement and organization (Fig. 4A). Treatment with paclitaxel resulted in microtubule polymerization with an increase in the density of cellular microtubules and formation of long thick microtubule bundles surrounding the nucleus (Fig. 4B). As₂O₃ treatment resulted in findings similar to those of paclitaxel-induced microtubule changes, such as thickening and increased density of microtubules (Fig. 4C). In contrast, vinblastine treatment caused cellular microtubule depolymerization with short microtubules in the cytoplasm (Fig. 4D). To further confirm these results, we used electron microscopy to explore the effect of As₂O₃ on cellular microtubule structures in MCF-7 cells. As shown in Fig. 5, the polymerized microtubules with diameters of ~25 nm were clearly observed in ~40% of As₂O₃-treated interphase cells, whereas microtubules with such structure were seen in less than 5% of untreated control cells (Fig. 5A). Furthermore, we examined the effect of As₂O₃ on spindle formation in MCF-7 cells, and found that As₂O₃ treatment caused the formation of abnormal mitotic spindles with a circular pattern, different from paclitaxel treatment, which caused tri- or tetra-polar spindles (data not shown). All results indicated that the microtubule is the intracellular target for As₂O₃ accounting for the subsequent mitotic arrest and apoptosis.

View larger version (64K): [in this window] [in a new window]   Fig. 4.   Effect of As2O3 on the organization of cellular microtubule network. MCF-7 cells were treated with 0.5 µM paclitaxel, 0.1 µM vinblastine, and 3 µM As2O3, or with the same volume of PBS solution as control. After a 12-h incubation, cells were harvested and fixed with cold methanol. Cells were incubated with monoclonal anti--tubulin antibody at room temperature for 90 min. After incubation with FITC-conjugated second antibody, the cellular microtubules were observed by Nikon PS200 fluorescence microscopy. The normal organization of microtubule network was seen in control cells (A), the long polymerized microtubule bundles were found in both paclitaxel- and As2O3-treated cells (B and C), and the shortened microtubules in cytoplasm were presented in vinblastine-treated cells (D).

View larger version (100K): [in this window] [in a new window]   Fig. 5.   As2O3-induced cellular microtubule polymerization seen by electron microscopy. MCF-7 cells were treated with 3 µM As2O3 or with the same volume of PBS solution as control. After a 12-h treatment, cells were harvested and prepared for analysis under electronic microscope as described under Materials and Methods. Microtubule structure is barely evident at interphase in control cells (A). The polymerized microtubule with about 25 nm tubular structure as indicated by arrow was found in As2O3-treated cells (B). The top in B represents the detail of polymerized microtubule structure. The magnification of the electron micrograph is 5,000×. The magnification of the insert of B on the upright is 25,000×.

Effect of As₂O₃ on P34^cdc2/Cyclin B Expression and Activity. It is known that the entrance to and exit from M-phase of the cell cycle is regulated by the p34^cdc2/cyclin B complex. Paclitaxel-induced M-phase arrest is associated with up-regulation and activation of p34^cdc2/cyclin B kinase in a variety of cell lines (Donaldson et al., 1994). We determined the effect of As₂O₃ on expression and activation of cdc2 and cyclin B in MCF-7 cells. As shown in Fig. 6, treatment with As₂O₃ resulted in a time-dependent up-regulation of cyclin B1, but did not change cdc2. Histone H1 kinase assay revealed that As₂O₃ treatment activated p34^cdc2/cyclin B1 kinase activity in a time-dependent manner. Similar results were obtained in H460 cells, indicating that up-regulation of cyclin B and activation of p34^cdc2/cyclin B kinase caused by As₂O₃ is not restricted to MCF-7 cells (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 6.   As2O3-induced M-phase arrest is correlated with the accumulation and activation of p34cdc2/cyclin B complex in MCF-7 cells. A, MCF-7 cells were treated with 3 µM As2O3 for the indicated time and then cell extracts were prepared for immunoblot analysis. Equal amounts of extract from each sample were electrophoresed on 12% SDS-PAGE. After transfer to a nitrocellulose membrane, cdc2 and cyclin B were probed by monoclonal anti-cdc2, and anti-cyclin B antibodies. The quantification of cdc2 and cyclin B was determined by laser densitometry (S-670 Scanning Densitometer; BioRad). The relative amounts of cdc2 and cyclin B were expressed as the amount compared with that at time 0. B, MCF-7cells were treated with 3 µM As2O3 for the indicated time. After treatment, cells were harvested and lysed with lysis buffer. Cdc2/cyclin B complexes were immunoprecipitated by monoclonal anti-cdc2 and anti-cyclin B antibodies, and protein A/G-conjugated agarose beads. Activity of cdc2/cyclin B kinase was determined by [-32P]ATP incorporation into the substrate histone H1. The kinase activity was determined by either autoradiography, or by counting the radioactivity in histone H1 protein. The relative kinase activity was expressed as the kinase activity compared with that at time 0. Each point represents mean ± S.D. of three independent experiments

Effect of As₂O₃ on Bcl-2 Phosphorylation. A number of reports have demonstrated that paclitaxel-induced M-phase arrest and apoptosis are associated with Bcl-2 phosphorylation (Blagosklonny et al., 1996; Haldar et al., 1996). We tested whether As₂O₃-induced M-phase arrest and apoptosis were associated with Bcl-2 phosphorylation. MCF-7 and H460 cells were treated with As₂O₃ for the time indicated in Fig. 7, and Bcl-2 phosphorylation was detected by immunoblotting analysis. As expected, As₂O₃ treatment resulted in a time-dependent Bcl-2 phosphorylation in both MCF-7 and H460 cells (Fig. 7). To explore whether As₂O₃-induced M-phase arrest and Bcl-2 phosphorylation could initiate apoptotic pathways, H460 cells were exposed to 10 µM As₂O₃ for 24 h. The mitotic cells (detached cells) and interphase cells (attached cells) were separated as described under Materials and Methods. After being washed three times with drug-free medium, the attached and detached cells were reincubated in fresh drug-free medium for the indicated time. Samples taken from the culture were divided into two parts, one for counting mitotic and apoptotic cells by Wright-Giemsa and DAPI staining, the other for determining mitotic and apoptotic events by immunoblotting analysis. As shown in Fig. 8, ~97% of detached cells were in M-phase at time 0. By 10 h of incubation, ~76% of cells were in M-phase; only ~4% were in M-phase at 24 h. In the detached cells, ~3% of the cells were apoptotic at time 0, which gradually increased to ~70% after 24 h of incubation. Furthermore, the immunoblotting analysis revealed that Bcl-2 phosphorylation bands in detached cells were clearly detected at time 0 but disappeared after 24 h of incubation. The cleavage bands of PARP and -catenin were seen as early as 10 h and increased as incubation continued. In contrast, neither accumulation of mitotic or apoptotic cells, Bcl-2 phosphorylation, cleavage of PARP, nor -catenin was observed in attached cells. In addition, the activation of caspases (Fig. 8C) and the cleavage of procaspase-3 and -7 into the active forms (Fig. 8D) were found in detached cells as early as 10 h of incubation. The activation of caspase-3 and -7, as well as the cleavage of apoptosis-related proteins, occurred only in detached cells and as early as 10 h of incubation; by that time, ~75% of cells were still at mitosis and only ~20% were apoptotic. Thus, it is presumed that As₂O₃-induced M-phase arrest precedes initiation of apoptotic pathways and may be prerequisite.

View larger version (52K): [in this window] [in a new window]   Fig. 7.   As2O3 induces Bcl-2 phosphorylation in MCF-7 and H460 cells. MCF-7 and H460 cells were treated with 3 or 10 µM As2O3, respectively, for the indicated time. Cells were harvested, and cell extract was prepared and loaded on 14% SDS-PAGE as described under Materials and Methods. After electrophoresis, protein blots were transferred on the nitrocellulose membrane, and Bcl-2 protein was probed by monoclonal anti-Bcl-2 antibody.

View larger version (41K): [in this window] [in a new window]   Fig. 8.   As2O3-induced apoptotic events occurred only in mitotic cells, not in interphase cells. H460 cells were treated with 10 µM As2O3 for 24 h. The mitotic cells (detached) and interphase cells (attached) were separated by gentle shaking by hand. After cells were washed three times with fresh drug-free medium, they were reincubated in fresh drug-free medium with 10% serum for the indicated time. After incubation, cells were harvested and divided into two aliquots. One was for determination of M-phase arrest and apoptosis after cells stained by Wright-Giemsa and DAPI (A). The other was for determination of Bcl-2 phosphorylation, and cleavage of PARP and -catenin (B), determination of caspase activation (C), proteolysis of procaspase-3 and procaspase-7 into activated caspase-3 and caspase-7 (D). Each point in C represents the mean ± S.D. of three independent experiments.

Effect of As₂O₃ and Paclitaxel on Cell Proliferation in Drug-Resistant Cell Lines. Because As₂O₃ has paclitaxel-like activity against cell proliferation in human solid tumors, we examined whether this compound was cross resistant with paclitaxel in drug resistant cell lines. We treated parental breast cancer MCF-7 cells, P-glycoprotein-overexpressing MCF-7/Dox cells, and MRP-positive MCF-7/VP-16 cells with As₂O₃ and with paclitaxel for 72 h, and then determined the drug-induced cytotoxicities of each. As shown in Table 2, the resistance index (RI) values for paclitaxel and As₂O₃ in MCF-7/Dox cell lines were 35 and 2.3, respectively. The RI values for paclitaxel and As₂O₃ in MCF-7/VP-16 cell line were 4 and 1.9, respectively. These data indicate that As₂O₃, unlike paclitaxel, was a poor substrate for transport by both P-glycoprotein and MDR protein. Next we tested the cytotoxic effect of As₂O₃ and paclitaxel in parental ovarian cancer cell line (A2780) and its tubulin mutation paclitaxel resistant cell lines (A2780/PTx10, and A2780/PTx22). The results showed that RI values for paclitaxel were 15 and 17 in A2780/PTx10 and A2780/PTx22 cells, whereas the RI values were <1 for As₂O₃ in both tubulin mutant cell lines. These results clearly show that As₂O₃ is not cross-resistant with paclitaxel.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Our data demonstrate that As₂O₃ is efficacious in inhibiting the proliferation of various models of human solid tumors. As₂O₃-induced cytotoxicity was independent of p53 status, suggesting that As₂O₃-induced cell death does not occur via p53-dependent pathways. We found that As₂O₃ was a poor substrate for transport by P-glycoprotein or by MRP, suggesting that pumping out intracellular As₂O₃ may not be a major factor in explaining the different susceptibilities to As₂O₃ of different cell lines_. These differences remain unexplained. One possibility could be due to different cellular components and/or thresholds in different cell lines for apoptotic signaling. The IC₅₀ value for As₂O₃ in MCF-7 breast carcinoma and A2780 ovarian carcinoma cell lines was about 3 µM, a concentration very close to effective serum concentration for successfully treating patients with acute promyelocytic leukemia in the clinic (Shen et al., 1997). Thus it is suggested that As₂O₃ might be useful for the treatment of some instances of human breast and ovarian cancers. In addition, we found that As₂O₃ lacked cross-resistance in both P-glycoprotein- and MRP-overexpressing cell lines. Furthermore, there is no cross-resistance between As₂O₃ and paclitaxel in two cell lines resistant to paclitaxel because of tubulin mutation, suggesting that it might be useful for treating drug-refractory patients, particularly those with paclitaxel resistance.

Although As₂O₃ has been an effective treatment for acute promyelocytic leukemia, the mechanism of action by which As₂O₃ induces cell death remains poorly understood. Recently, some investigators reported that As₂O₃ caused DNA damage, oxidative stress, and mitochondrial dysfunction (Van Wijk et al., 1988; Dong and Lou, 1993; Gurr et al., 1999; Yih and Lee, 2000). In addition, As₂O₃ treatment blocked the cell cycle in G₁ or at G₂/M depending on the cell line (Ma et al., 1998; Li and Broome, 1999). In the present study, we found that As₂O₃ treatment resulted in cell-cycle arrest at M-phase in MCF-7, H460, and HeLa cells. These results are consistent with the report by Li and Broome, who reported that As₂O₃ treatment led to metaphase arrest in myeloid leukemia cells (Li and Broome, 1999). Our results demonstrate that As₂O₃-induced M-phase arrest is due to interaction with tubulin, resulting in tubulin polymerization. We also found that tubulin polymerization by As₂O₃ did not affect GTP binding to -tubulin. These results, however, are diametrically opposite those of Li and Broome, who reported that As₂O₃ caused blockade of GTP binding to -tubulin and thus inhibited tubulin polymerization (Li and Broome, 1999). To certify our conclusions, we have performed the test under various conditions, including the addition of As₂O₃ before, concurrent with, and after GTP, and obtained reproducible results in all cases. Because the -tubulin materials used by us and Li and Broome were both from Sigma-Aldrich, the explanation for contrary results is not obvious. One possibility may be the solvent for As₂O₃. In the present work, As₂O₃ was dissolved in 10% NaOH, which was also used in the drug-free control. The solvent used in Li and Broome's experiment is not clearly described, and this could be a difference.

In this work, we noticed a big discrepancy between the concentrations of As₂O₃ in cell-based assays and in cell-free systems. We believe that the intracellular circumstances may provide an ideal environment for the reaction of As₂O₃ with the microtubule assembly process and that the reaction conditions of the widely used cell-free microtubule assembly-disassembly assays might not facilitate the As₂O₃-microtubule interaction; consequently, the concentration of As₂O₃ in this system must be much higher than that in the cell-based assays. As a matter of fact, this phenomenon has been observed before by us using other tubulin ligands, although the discrepancies were not so pronounced (Jiang et al., 1998a,b).

Indirect immunofluorescence techniques allow detection of morphological changes in the microtubule network, such as alterations in microtubule organization and arrangement. The changes in microtubule length and density constitute an appropriate system to qualitatively assess the intracellular microtubule polymerization or depolymerization caused by antitubulin agents (de Arruda et al., 1995; Mooberry et al., 1999). The results from tubulin depolymerization experiments indicated that As₂O₃ treatment prevented tubulin depolymerization, which was similar to the effect of paclitaxel treatment. To further confirm these results, changes in the cellular microtubule network were observed immunocytochemically. We found that the effect of As₂O₃ was similar to that of paclitaxel but different from that of vinblastine. Thick bundles of microtubule network surrounding the nucleus were seen in As₂O₃- and paclitaxel-treated cells, whereas shortened depolymerized microtubules were observed in vinblastine-treated cells. Details of the bundles of As₂O₃-induced microtubule were studied by electron microscopy. As₂O₃ treatment caused the formation of filament tubular structures ~25 nm in diameter. These structures were rarely seen in interphase control cells.

Antitubulin compounds can be classified into categories depending on their binding sites on -tubulin. Vinca alkaloids, rhizoxin, dolastetins, and spongistatin react with the domain for "vinblastine"; colchicine, nocodazole, podophyllotoxin, steganacin, and curacin A bind to the "colchicine" domain (Bai et al., 1993; Uppuluri et al., 1993; Lobert et al., 1995). Paclitaxel interacts with amino acids 1 to 31 and 217 to 231 at the N terminus of -tubulin (Rao et al., 1994, 1995); the main interaction of the paclitaxel ring with tubulin is at L279 at the B8-H9 loop (Nogales et al., 1998), and GTP/GDP adheres at the GTP binding sites (Little and Luduena, 1987). Because preincubation of -tubulin with As₂O₃ did not prevent GTP-dependent assembly of microtubules, it suggests that the As₂O₃ binding site probably is separate from the GTP pocket. Because As₂O₃ exhibited actions on the tubulin-microtubule cycle similar to paclitaxel, As₂O₃ and paclitaxel might share a common binding domain on -tubulin. We examined the effect of As₂O₃ on [³H]paclitaxel binding to -tubulin by competition assay as described by Bollag et al. (1995). The results from competitive inhibition experiments showed, however, that As₂O_3, even at high concentrations, did not competitively inhibit [³H]paclitaxel binding to -tubulin (data not shown), indicating that As₂O₃ does not share paclitaxel binding sites on -tubulin. Furthermore, the results presented in Table 2 show that there is no cross-resistance between As₂O₃ and paclitaxel in paclitaxel-resistant cell lines due to tubulin mutation, further supporting the interpretation that As₂O₃ has its own binding site on -tubulin.

In eukaryotic cells, cell-cycle progression is regulated although timely activation and inactivation of cyclin-dependent kinases, cyclins, and other regulatory factors. Inappropriate alteration in the expression or/and activation of cyclin-dependent kinases and regulators can lead to blockade of cell cycle progression and induction of programmed cell death (Pines, 1995). It is well known that p34^cdc2/cyclin B complexes are involved in regulation of G₂/M phase and the M-phase transition (King et al., 1994). Our previous work and that of others have demonstrated that paclitaxel-induced M-phase arrest, inappropriate accumulation of B type cyclins and activation of p34^cdc2/cyclin B kinase was associated with the initiation of apoptotic pathways. In the present study, we found that As₂O₃ treatment led to an increase in cyclin B level and stimulation of p34^cdc2/cyclin B kinase activity. The pattern of As₂O₃-induced activation of p34^cdc2/cyclin B and apoptosis is similar to that of paclitaxel, as we described previously (Ling et al., 1998a). Our results indicate that in addition to direct disruption of tubulin, treatment with As₂O₃ can lead to inappropriate accumulation and/or activation of G₂/M phase-related regulators, resulting in the initiation of apoptosis signaling.

Recently, a number of investigators indicated that anti-tubulin agents are able to induce Bcl-2 phosphorylation (Jiang et al., 1998a; Tahir et al., 2001). The role of Bcl-2 phosphorylation in G₂/M phase arrest and in apoptosis remains to be further elucidated (Yamamoto et al., 1999). Our previous results indicated that some anti-tubulin agents, such as nocodazole, induced Bcl-2 phosphorylation that correlated only with M-phase arrest. Paclitaxel-induced Bcl-2 phosphorylation was associated with the initiation of apoptosis, however (Ling et al., 1998b). As₂O₃ treatment, like that of most antitubulin agents, resulted in Bcl-2 phosphorylation in MCF-7 and H460 cells. As₂O₃-induced apoptotic events including the activation of caspases and proteolysis of apoptotic target proteins occurred only in drug-induced mitotic cells, not in interphase cells. A time-course study showed that As₂O₃-induced Bcl-2 phosphorylation preceded by 10 h the activation of caspase-3 and -7 and cleavage of PARP and -catenin, suggesting that As₂O₃-induced M-phase arrest and Bcl-2 phosphorylation might be a required event for the initiation of apoptosis. Recent reports indicate that antitubulin agent blockage of the cell cycle at M-phase can be associated with the activation of several types of kinases, leading to the phosphorylation of cascades, and the activation of cdc2/cyclin B kinase, cdc-25, Raf, and Bcl-2 phosphorylation (Scatena et al., 1998; Fang et al., 2000). To explore the kinases that might be involved in As₂O₃-induced Bcl-2 phosphorylation, we used pharmacological concentrations of kinase inhibitors to test their effect on drug-induced M-phase arrest and Bcl-2 phosphorylation. Preliminary data show that the protein kinase C inhibitor staurosporine at 50 nM effectively prevented As₂O₃-induced Bcl-2 phosphorylation and M-phase arrest; the cdc2 kinase inhibitor roscovitine partially inhibited Bcl-2 phosphorylation. However, the tyrosine kinase inhibitor genistein and the mitogen-activated protein kinase inhibitor PD98059 had no effect on Bcl-2 phosphorylation (unpublished data). These data suggest that the activation of some specific kinases, such as protein kinase C, could be responsible for As₂O₃-induced Bcl-2 phosphorylation and M-phase arrest.

In summary, our data demonstrate that As₂O₃ is efficacious in suppressing the cell growth in a variety of solid tumor models. As₂O₃ directly interferes with microtubules, blocking cell cycle at the M-phase. These effects are paclitaxel-like, but totally distinct from paclitaxel: As₂O₃ is more active than paclitaxel in MDR and MRP cell lines and equally active in cell lines in which mutation in the tubulin binding site for paclitaxel renders that drug 15-fold less potent. Similar to most antitubulin agents, As₂O₃ causes inappropriate accumulation and activation of p34^cdc2/cyclin B, Bcl-2 phosphorylation, and, finally, triggering of the apoptotic cascade. Such a unique mechanism for As₂O₃ indicates that it may be useful for treating patients with solid tumors either alone or in combination with other therapeutic agents.