Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Arsenic exposure via contaminated drinking water from groundwater sources is a serious public health concern in many parts of the world. Based principally on epidemiological findings, exposure to arsenic has been associated with increased risk of a variety of neoplastic diseases, including skin, lung, bladder, liver, and kidney cancers (Stöhrer, 1991; Abernathy et al., 1999). Although inorganic arsenic has been established as a human carcinogen, its mechanism of action is not well understood.

Paradoxically, despite its putative function as a suspected carcinogen, arsenic trioxide (As₂O₃) has been used therapeutically in a limited number of patients for the effective treatment of acute promyelocytic leukemia (APL) (Chen et al., 1996; Soignet et al., 1998; Niu et al., 1999). Experiments in vitro have shown that treatment with clinically achievable concentrations of As₂O₃ can trigger the apoptotic death of APL cell lines (Ma et al., 1998; Dai et al., 1999; Huang et al., 1999b; Jing et al., 1999; Larochette et al., 1999; Perkins et al., 2000) as well as other malignant cells (Akao et al., 1999; Bazarbachi et al., 1999; Lu et al., 1999; Rousselot et al., 1999; Zhang et al., 1999; Zhu et al., 1999). Several mechanisms have been proposed for arsenite-induced apoptosis, including caspase activation (Akao et al., 1999; Huang et al., 1999b), down-regulation of the anti-apoptogenic protein Bcl-2 (Chen et al., 1996), activation of the mitochondrial permeability transition pore (Kroemer and de Thé, 1999; Larochette et al., 1999; Perkins et al., 2000), and activation of the c-Jun NH₂-terminal kinase pathway (Huang et al., 1999a) and oxidative injury (Watson et al., 1996; Dai et al., 1999; Jing et al., 1999). Numerous studies also have established a link between growth inhibition and apoptosis in arsenite-treated cell lines (Bazarbachi et al., 1999; Dai et al., 1999; Lu et al., 1999; Seol et al., 1999; Zhang et al., 1999; Zhu et al., 1999).

Arsenite inhibits mitotic division in a variety of cell types (Yih et al., 1997; Huang and Lee, 1998). Inhibition of mitosis occurs due to perturbation of the spindle apparatus and tubulin (Huang and Lee, 1998; Li and Broome, 1999), which may be responsible for the cytogenetic alterations contributing to carcinogenesis. Arsenite targeting of tubulin has been proposed as a mechanism leading to apoptosis (Li and Broome, 1999), and it is generally thought that the induction of apoptosis occurs selectively in cells arrested in G₂/M (Ma et al., 1998). Studies claiming that arsenite arrests cells in G₂/M (Ma et al., 1998; Seol et al., 1999) are consistent with it behaving as a mitotic disrupter; however, there also is evidence that arsenite induces a G₁ arrest (Bazarbachi et al., 1999).

At concentrations ranging from 0.1 to 0.5 µM, As₂O₃ has been shown to induce partial differentiation of the APL cell line NB4 (Chen et al., 1996). Because differentiation-inducing therapy with all-trans-retinoic acid is a standard therapy in patients with APL, it is thought that the differentiation-promoting effects of As₂O₃ may contribute to its clinical efficacy as well, and several studies support the differentiation-inducing therapeutic potential of arsenite (Kizaki et al., 1998; Cai et al., 2000; Perkins et al., 2000).

In view of the fact that arsenite reportedly has multiple effects on cellular functions, including reduced viability, delayed proliferation, induction of apoptosis, and enhanced differentiation, we initiated a study using the U937 human myelomonocytic cell line as a model system to explore the concentration-dependent effects of arsenite on these functions. The U937 cell line shares many cellular and molecular similarities with APL-derived cell lines, particularly with respect to responsiveness to retinoids and vitamin D (Nervi et al., 1998; James et al., 1999). U937 cells are less sensitive to arsenite-induced apoptosis than most APL-derived cell lines such as the NB4 cell line (Jing et al., 1999; Cai et al., 2000). Sensitivity to arsenite appears to be dependent on the activity of enzymes that regulate cellular H₂O₂ content. Thus, U937 cells become sensitive to lower concentrations of arsenite when cotreated with inhibitors of glutathione peroxidase and catalase. Given the relative insensitivity of U937 cells to the apoptogenic effects of arsenite, we viewed this cell line as a useful model to study the concentration-dependent effects of arsenite on the inter-relationship between cell cycle dysregulation, apoptosis, and cell differentiation. This study defines how a number of inter-related factors, including arsenite concentration, duration of treatment, and cell cycle phase of the targeted cell population, determine which cellular functions are affected by arsenite.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents. Sodium m-arsenite (i.e., NaAsO₂), nitroblue tetrazolium (NBT), and phorbol-12-myristate-13-acetate (PMA) were obtained from Sigma Chemical Co. (St. Louis, MO). Fresh stock solutions of sodium arsenite (2 mM in sterile Hanks' balanced salt solution) were prepared before every experiment and filter sterilized using a 0.2-µm syringe filter. 5-(and-6)-Carboxyfluorescein diacetate, succinimidyl ester (CFSE) and calcein-AM were obtained from Molecular Probes (Eugene, OR). Z-Val-Ala-DL-Asp-fluoromethylketone (z-vad.fmk) was obtained from Bachem Bioscience, Inc. (King of Prussia, PA). Calcitriol (1,25-dihydroxyvitamin-D₃) was purchased from BioMol Research Laboratories, Inc. (Plymouth Meeting, PA). Phycoerythrin-conjugated anti-CD11b/Mac-1 and rabbit anti-caspase-3 were obtained from Pharmingen (San Diego, CA). Alkaline phosphatase-conjugated goat anti-rabbit IgG was purchased from Tropix, Inc. (Bedford, MA).

Cell Culture. U937 cells were obtained from the American Type Culture Collection (Manassas, VA), and maintained in RPMI 1640 supplemented with 5% fetal bovine serum (FBS), 1% L-glutamine, and 0.1% gentamycin. RPMI 1640 and FBS were obtained from Hyclone Laboratories (Logan, UT); glutamine and gentamycin were obtained from Life Technologies (Grand Island, NY). Cells were maintained in logarithmic growth at a density between 0.2 and 1 × 10⁶ cells/ml at 37°C in a humidified atmosphere consisting of 5% CO₂. For experiments the seeding cell density was kept constant at 2 × 10⁵ cells/ml.

Cell Viability. Cell viability was monitored using trypan blue exclusion (0.2% in physiological saline). Alternatively, cell viability was monitored by flow cytometry by analyzing propidium iodide uptake (dead cells) or calcein-AM esterase activity (live cells). For propidium iodide uptake, cells were stained with 10 µg/ml propidium iodide for 15 min after which 20,000 cells were analyzed for amount of red fluorescence. For calcein-AM hydrolysis, 1 × 10⁶ cells were labeled with 4 µM calcein-AM in phosphate-buffered saline for 10 min at room temperature. Calcein fluorescence emission was acquired at 530 nm.

Analysis of Cell Division. The influence of arsenite on cell division was monitored according to the technique described by Lyons and Parish (1994) with slight modifications. A 100× stock solution (500 µM) of CFSE was prepared in 100% dry dimethyl sulfoxide. Cells were adjusted to 1 × 10⁷ cells/ml in phosphate-buffered saline, mixed with CFSE at a final concentration of 5 µM, and incubated at 37°C for 15 min. CFSE-labeled cells were washed three times with complete RPMI 1640, cell density was adjusted to 2 × 10⁵ cells/ml, and labeled cells were cultured with and without arsenite for 96 h. At daily intervals cells were harvested from culture, washed twice in phosphate-buffered saline, and CFSE fluorescence was monitored by flow cytometry on an FACScalibur (Becton Dickinson, San Jose, CA). With each cell division, CFSE is equally partitioned into daughter cells resulting in a parallel halving of fluorescent intensity. CFSE fluorescence was collected on FL1 (530 nm) using logarithmic amplification. For each sample, 10,000 cells were acquired for flow cytometric analysis.

Flow Cytometric Analysis of DNA Content. U937 cells (2 × 10⁶ cells/10 ml) were cultured for up to 96 h with or without arsenite. Cells were then harvested, washed three times with 5 ml of phosphate-buffered saline, and fixed with 70% ethanol overnight at 4°C. Fixed cells were washed three times again with 5 ml of PBS, and then stained with 50 µg/ml propidium iodide (Sigma Chemical Co.) in the presence of RNase A (100 U/ml final) for 30 min. Propidium iodide fluorescence (i.e., DNA content) was determined by flow cytometry on an FACScalibur (Becton Dickinson) using doublet discrimination. Propidium iodide fluorescence was collected on FL2 (585/42 nm) using linear amplification. A minimum of 20,000 cells/sample was analyzed. Data were collected using CellQuest software and analyzed for cell cycle distribution and apoptosis using the Modfit software.

Agarose Gel Electrophoretic Analysis of DNA Fragmentation. Cells (5 × 10⁶) were harvested from culture, washed with ice-cold Hanks' balanced salt solution, and pellets were lysed in 500 µl of lysis buffer containing 10 mM Tris-HCl (pH 8.0), 10 mM EDTA, and 1% Triton X-100 for 15 min at room temperature. Cell lysates were centrifuged at 13,000g for 15 min. Supernatants were collected and incubated with 40 µl of 1 mg/ml heat-inactivated RNase A at 37°C for 15 min. After phenol/chloroform extraction the aqueous layer was ethanol precipitated with 0.2 M NaCl and 2 volumes of ethanol at 20°C overnight.

Centrifugal Elutriation. A Beckman JE-6B elutriation system and rotor were used to enrich for a synchronous G₁ population of U937 cells. To accomplish this, 1 × 10⁸ cells (1 × 10⁷ cells/ml suspended in RPMI 1640 + 1% FBS) from an exponentially growing culture were loaded into the separation chamber (sterilized with 6% H₂O₂) maintained at ambient temperature at a rotor speed of 2500 rpm and a cell suspension entry flow rate of 15 ml/min. The G₁ fraction was elutriated by collecting 150 ml at a flow rate of 40 ml/min. The recovered cell number was determined using a Coulter counter, and analysis of DNA content by propidium iodide staining and flow cytometry was used to verify that the elutriated cells were indeed in G₁.

Caspase-3 Assays. Caspase-3 enzyme activity was measured using a fluorogenic substrate assay kit obtained from Pharmingen (San Diego, CA). U937 cells were seeded at a density of 2 × 10⁵ cells/ml and incubated in the presence or absence of arsenite for the desired time periods. At various times after treatment, cell lysates were prepared as follows. Cells were harvested, washed in ice-cold PBS, and lysed with cell lysis buffer consisting of 10 mM Tris-HCl, 10 mM NaH₂PO₄/NaHPO₄, pH 7.5, 130 mM NaCl, 1% Triton X-100, and 10 mM sodium pyrophosphate. An aliquot of cell lysate (5 × 10⁵ cell equivalents) was added to a reaction tube that also contained 10 µl of the reconstituted Ac-DEVD-AMC substrate and 200 µl of 1× HEPES buffer. The reaction mixture was incubated for 1 h at 37°C after which the amount of AMC liberated from Ac-DEVD-AMC was measured using a SpectraMax Gemini plate reader with an excitation wavelength of 380 nm and an emission wavelength of 440 nm.

CD11b/Mac-1 Expression. At the end of the culture period cells were collected, washed twice in PBS + 0.1% NaN₃ + 1% FBS, and then incubated for 30 min at 4°C in the dark with R-PE-conjugated anti-CD11b. R-PE-conjugated anti-TNP was used as an isotype control for nonspecific binding. Red fluorescence was collected using logarithmic amplification on FL2 (585/42 nm) of cells gated on forward and side scatter.

NBT Reduction. U937 cells (2 × 10⁵ cells/ml) were cultured with or without sodium arsenite (0.5 µM) and 50 nM 1,25-dihydroxyvitamin-D₃ for 4 days at 37°C. The degree of differentiation was assayed by the ability of cells to reduce NBT to insoluble blue-black formazan crystals upon stimulation with PMA. Differentiation was determined by incubating 1 × 10⁵ cells/ml with 5 mg/ml NBT and 200 ng/ml PMA at 37°C and analyzing the respiratory burst after 30 min. The number of cells containing formazan deposits from a total of 200 cells was counted in a hemocytometer using a microscope. Cells positive for NBT reduction were expressed as a percentage of the total viable cell number.

Statistical Analysis. Statistical comparisons were performed using one-way or two-way ANOVA and Dunnett's test. A P value of .05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

U937 cells were exposed to arsenite concentrations ranging from 1 to 100 µM. Cell growth and viability were monitored by counting cells and by assessing trypan blue exclusion, respectively (Fig. 1). Arsenite induced a concentration-dependent inhibition of cell growth (Fig. 1A). Concentrations of arsenite 1 µM significantly reduced cell recoveries at 24 and 48 h. The inhibition of cell growth at concentrations of arsenite >10 µM could largely be attributed to cytotoxicity (Fig. 1B) because trypan blue exclusion was noticeably compromised at these concentrations, and the effects of arsenite at these concentrations were statistically significant relative to control. Lower concentrations of arsenite (i.e., <10 µM) were less cytotoxic, suggesting that the growth inhibition at these concentrations was due to cytostatic mechanisms. For an extended time course analysis of the cytostatic potential of arsenite, 5 µM arsenite was chosen because this concentration was minimally cytotoxic at 24 and 48 h (Fig. 1B). Over the course of 4 days 5 µM arsenite significantly suppressed the growth of the cell population (Fig. 1C). In comparison, control cells exhibited exponential growth during this same time period with a population doubling time of ~20 h (Fig. 1C). The loss of cell viability in the presence of 5 µM arsenite, although statistically significant, never exceeded 25% (Fig. 1D). Hence, the suppressed population growth was not mainly due to cytotoxic effects of arsenite.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Depending on concentration, arsenite is both cytotoxic and cytostatic. A, U937 cells (2 × 105 cells/ml) were cultured in the presence of the concentrations of arsenite indicated on the x-axis and viable cell recoveries were determined at 24 h () or 48 h (). Statistically significant (P  .05) effects of arsenite concentration and time (24 versus 48 h) were established using two-way analysis of variance. Concentrations of arsenite as low as 1 µM reduced cell recovery significantly, as determined by Dunnett's test. B, cell viability in the presence of the concentrations of arsenite indicated on the x-axis were determined by trypan blue exclusion at 24 h () or 48 h (). Statistically significant (P  .05) effects of arsenite concentration and time (24 versus 48 h) were established using two-way analysis of variance. C and D, U937 cells (2 × 105 cells/ml) were cultured in the presence () or absence () of 5 µM arsenite for 96 h and viable cell recovery (C) and cell viability (D) were determined at 24, 48, 72, or 96 h. In C and D, a statistically significant difference between 5 µM arsenite and control was established using two-way ANOVA. In A to D, data values are the mean of quadruplicate determinations and representative of at least five separate experiments.

Labeling with the fluorescent dye CFSE was used to further explore the cytostatic potential of arsenite. CFSE is distributed evenly between parent and daughter cells with successive cell divisions; accordingly, the mean fluorescence intensity halves with each round of cell division (Lyons and Parish, 1994). In this way, cell division (or lack thereof) can be monitored by flow cytometry, where the number of cell divisions is defined by gates set around CFSE fluorescent peaks.

CFSE-labeled cells from control, 5, 10, and 20 µM arsenite-treated cultures were harvested at 24, 48, 72, and 96 h and analyzed for cell divisions (i.e., loss of CFSE fluorescence) by flow cytometry (Fig. 2A). For control cells the mean CFSE fluorescence intensity more than halved during each 24-h interval evaluated (Fig. 2B), which is consistent with a population doubling time less than 24 h. The total number and rate of cell divisions were diminished significantly in the presence of arsenite in a concentration-dependent manner (Fig. 2B); however, the CFSE analysis revealed that cells were continuing to divide in the presence of 5 µM arsenite despite the fact that the total cell population did not increase (Fig. 2C). In other words, at 5 µM arsenite the continued division of some cells replaced the nonviable ones. Cellular CFSE fluorescence did not decrease over time in the presence of 20 µM arsenite because this concentration was cytotoxic.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Cells continue to divide at cytostatic concentrations of arsenite, but at a reduced rate and frequency. A, cell divisions were determined by loss of CFSE fluorescence at 24, 48, 72, and 96 h. CFSE-labeled U937 cells (2 × 105 cells/ml) were exposed to arsenite (0, 5, 10, and 20 µM); harvested at 24, 48, 72, and 96 h; and analyzed for cell divisions (i.e., CFSE content) by flow cytometry. B, CFSE mean fluorescence intensities for each treatment group were calculated and plotted over the time course. A statistically significant difference between arsenite and no arsenite over time was established by two-way ANOVA and Dunnett's test. C, for comparative purposes, cell recovery for the cells analyzed in B were determined by counting cell number.

To better understand the effects of arsenite on cell cycle control and cell division, cell cycle distribution and progression were evaluated using propidium iodide staining of DNA content, which was analyzed by flow cytometry (Fig. 3). Figure 3A represents a typical, control population of U937 cells taken from an exponentially growing, 24-h culture where ~47.1 ± 0.6% of the cells are in G₁, ~43.3 ± 0.2% are in S phase, and ~9.6 ± 0.6% are in G₂/M. In control cultures, the proportion of cells in each phase of the cell cycle did not change appreciably between 24 and 48 h (Fig. 3, A and B). After 24 h in the presence of 5 µM arsenite (Fig. 3C), cell cycle phase distribution was affected markedly; ~25 ± 0.3% of the cells are in G₁, ~47.7 ± 1.6% are in S phase, and ~27.2 ± 2.0% are in G₂/M. The apparent arrest and accumulation of arsenite-treated cells in G₂/M was short-lived, in that by 48 h (Fig. 3D) most of the G₂/M cells have disappeared (i.e., ~30.5 ± 0.1% of the cells are in G₁, ~68.5 ± 0.8% are in S phase, and ~1 ± 0.1% are in G₂/M at 48 h post arsenite exposure). Apoptotic cells (i.e., cells having sub-G₁ DNA content) were found in 5 µM arsenite-treated cultures at both 24 and 48 h. A greater percentage of subdiploid cells was found in cultures treated with 10 µM arsenite (data not shown). Agarose gel electrophoresis of genomic DNA isolated from U937 cells revealed that treatment with concentrations of arsenite 10 µM resulted in chromatin degradation, where the fragments had the typical laddering pattern indicative of apoptosis (Fig. 4). Although subdiploid cells were observed in cultures treated with 5 µM arsenite, DNA laddering was more difficult to detect from these cultures, probably because of the low frequency of cells having fragmented DNA (data not shown).

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Arsenite treatment delays cell cycle progression and induces apoptosis. Cell cycle progression for control (A and B) and 5 µM arsenite-treated cells (C and D) was monitored on day 1 (A and C) and day 2 (B and D) by propidium iodide staining. DNA content was analyzed using flow cytometry.

View larger version (96K): [in this window] [in a new window]   Fig. 4.   Arsenite treatment triggers DNA fragmentation. Agarose gel electrophoresis of genomic DNA isolated from U937 cells treated with the indicated concentrations of arsenite for 24 h revealed the typical oligosomal-sized chromatin fragmentation and laddering pattern characteristic of apoptosis.

It appeared that 5 µM arsenite affected cell cycle progression and induced apoptosis, where the predominant effect was on cell cycle transit. For this reason, the relationship between arsenite-induced cell cycle dysregulation and apoptosis was analyzed further. Of particular interest was the relationship between the initial G₂/M arrest and the onset of apoptosis as well as the fate of the G₂/M cells. U937 cells were cultured in the presence or absence of 5 µM arsenite for 4 days. Cells were collected daily and processed for DNA content analysis by flow cytometry; however, in this case bivariate plots with DNA content on the abscissa and side scatter on the ordinate were collected to aid the analysis (Fig. 5). Arsenite-treated cells (Fig. 5, E-H) exhibited increased side scatter, especially in the accumulating G₂/M population at 24 h (Fig. 5E). The fraction of G₂/M cells was very low in arsenite-treated cultures at 48, 72, and 96 h at which times the percentage of apoptotic cells (i.e., sub-G₁ DNA content and low side scatter) increased accordingly (compare Fig. 5, A and E, B and F, C and G, and D and H). The percentage of apoptotic cells in the presence of 5 µM arsenite was 15.4% at 24 h, 29.4% at 48 h, 47.5% at 72 h, and 67.9% at 96 h.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Arsenite treatment produces an initial accumulation of cells having G2/M DNA content, which subsequently disappear while apoptotic cells increase. U937 cells (2 × 105 cells/ml) were cultured in the absence (A-D) or presence (E-H) of 5 µM arsenite for 24 h (A and E), 48 h (B and F), 72 h (C and G), or 96 h (D and H). At each time point cells were harvested and processed for analysis of DNA content by flow cytometry. Data are displayed as bivariate plots with DNA content on the abscissa and side scatter on the ordinate to better demonstrate the appearance of apoptotic cells (low side scatter, subdiploid DNA content) as well as G1 and G2/M cell populations (see arrows).

The most straightforward interpretation of these data is that cells arrested in G₂/M by 5 µM arsenite treatment die via apoptosis. However, we know from the CFSE experiments that 5 µM arsenite-treated cells continue to divide, only at a reduced rate. Therefore, to better understand the relationship between arsenite effects on cell cycle progression, cell division, and apoptosis we used a more homogeneous (i.e., synchronous) population of cells. Elutriated U937 cells were prepared where ~95% of the starting cell population was in the G₁ phase of the cell cycle before treatment with arsenite (Fig. 6A, 0 h). After 8 h in culture, the majority (60 to 85%, depending upon the experiment) of the untreated cells (open circles) were in S phase (Fig. 6B, open circles), whereas the appearance of the first S phase peak was delayed by 4 to 8 h in cells treated with 5 µM arsenite (Fig. 6B, closed circles). Similarly, the appearance of the first G₂/M peak was delayed by 4 to 8 h in cells treated with 5 µM arsenite (Fig. 6C, compare open and closed circles). The fact that the first S phase peak at 16 h and the first G₂/M peak at 20 h in the presence of arsenite were delayed by the same amount of time suggests that transit through G₁ is impaired by arsenite, but no additional delay in S phase transit occurs. This suggestion is corroborated by the fact that there is no delay in exiting S phase (Fig. 6E, triangles and dashed lines) or in reaching the first G₂/M peak (Fig. 6F, triangles and dashed lines) when 5 µM arsenite is added to cultures that are already in S phase (i.e., 5 µM arsenite added at t = 8, Fig. 6, D-F). The data in Fig. 6 also show that in addition to the G₁ delay there is a substantial delay in G₂/M transit in the presence of 5 µM arsenite (Fig. 6C, closed circles). Importantly, when arsenite is added to elutriated cells at culture initiation (i.e., t = 0, closed circles in Fig. 6, A-C) there is a substantial delay (~8 to 12 h) in transit through G₂/M but not an arrest. Consistent with the CFSE data (Fig. 2) arsenite-treated cells emerge from the first G₂/M phase appearing as G₁ cells by 20 h (Fig. 6A, closed circles). A fraction of these cells continues to cycle at least into a second S phase (Fig. 6B, closed circles); however, in accord with previous cell cycle analysis (Figs. 3 and 5), subsequent G₂/M peaks are virtually absent in arsenite-exposed cultures. By virtue of the fact that the G₂/M delay lasts longer than the G₁ delay, it can be said that G₂/M cells are more sensitive to 5 µM arsenite. This is corroborated by the fact that there is a substantial delay in exiting G₂/M when 5 µM arsenite is added to cultures at a time (i.e., 5 µM arsenite added at t = 12 h, Fig 6, G-I) when the majority of cells is already in G₂/M (Fig. 6I, diamonds and dotted lines). Elutriated cells treated with 20 µM arsenite while the cells are in G₁ never progress past the G₁ phase of the cell cycle (data not shown).

View larger version (49K): [in this window] [in a new window]   Fig. 6.   Arsenite delays progression through both G1 (~4 h) and G2/M (~8 h), and apoptosis likely occurs after completion of an aberrant mitosis. A to C, elutriated U937 cells [~90 to 95% of the cells started in G1 were cultured for 48 h in the absence () or presence () of 5 µM arsenite added at culture initiation (i.e., t = 0)]. Cells were harvested every 4 h and processed for DNA content analysis by flow cytometry. The percentage of cells in each phase of the cell cycle (G1 phase, A, D, and G; S phase, B, E, and H; G2/M, C, F, and I) at each time point was plotted for the control cells () and the 5 µM arsenite-treated cells (A-C, ). In some instances the addition of arsenite was delayed for 8 h (i.e., t = 8, , dashed lines, D-F) or 12 h (i.e., t = 12, , dotted lines, G-I) after culture initiation to allow the cell population to progress into S phase or G2/M, respectively, before being treated with arsenite.

The mechanism of arsenite-induced cell death appears to involve the activation of caspases. Activation of caspase-3 by arsenite was measured in asynchronous and elutriated cell populations. For asynchronous cycling cells the activation of caspase-3 was both time and concentration dependent (Fig. 7A); concentrations of arsenite greater than 20 µM activated caspase-3 rapidly (i.e., by 6 h), whereas concentrations of arsenite 20 µM required a longer treatment time (i.e., 24 h) to activate caspase-3. For G₁ cells prepared by elutriation, elevated caspase-3 stimulation was noted within 24 h of 5 µM arsenite treatment (Fig. 7B, triangles). Delaying the addition of 5 µM arsenite to elutriated cells for 16 h, when the majority of the population was in G₂/M, resulted in significantly greater caspase-3 activation 8 h later when the cells had exited the G₂/M delay (Fig. 7B, compare diamonds at 24 h with triangles at 8 h). Again, these data are consistent with G₂/M cells being more sensitive to the apoptogenic effects of arsenite, and the onset of apoptosis after the delayed transit through G₂/M rather than vice versa.

View larger version (38K): [in this window] [in a new window]   Fig. 7.   Activation of caspase-3 by arsenite is time, concentration, and cell cycle phase dependent. A, caspase-3 enzyme activity was measured in lysates obtained at 6 h (), at 8 h (), or at 24 h () from cells treated with the concentrations of arsenite indicated on the x-axis. Arsenite concentration and sampling time were found to be statistically significant by two-way ANOVA. In comparison to control, all data points showing enzyme activity >1500 units were statistically significant as determined by Dunnett's test. B, U937 cells were elutriated to obtain an enriched G1 population of cells that were subsequently cultured in the absence () or presence of 5 µM () or 20 µM () arsenite. In another group, the addition of 5 µM arsenite was delayed until 16 h after culture initiation (, dotted lines), when the majority of the cells were in G2/M. C, processing of the 32-kDa caspase-3 proenzyme to the p17/p12 heterodimer was determined by Western blotting using cell lysates prepared from cells that had been treated with the indicated concentrations of arsenite for 24 h.

Caspase 3 (i.e., CPP32) is constitutively expressed in U937 cells as a 32-kDa proenzyme (Fig. 7C, lane 1) that undergoes proteolytic cleavage in the presence of 5 and 10 µM arsenite to form the p17/p12 active form of the enzyme (Fig. 7C, arrows). Consistent with the notion that apoptosis follows arsenite-induced cell cycle dysregulation, arsenite-induced activation of caspase-3, at least in detectable amounts, occurred late in culture (i.e., 24 h).

To confirm the involvement of caspases in arsenite-induced cell loss, the effects of the broad caspase inhibitor z-vad.fmk on arsenite-induced cell cycle dysregulation and DNA fragmentation were determined via propidium iodide fluorescence and flow cytometry (Fig. 8). Cells having less than G₁ DNA content and exhibiting reduced side-scatter were considered to be apoptotic. In the experiment shown in Fig. 8, cell cycle phase distribution and apoptosis were determined at 24 and 48 h for control and arsenite-treated cells cultured in the presence or absence of 50 µM z-vad.fmk. Z-vad, alone, had no effect on the cell cycle phase distribution of control cells; however, the caspase inhibitor markedly diminished the percentage of apoptotic cells found in arsenite-treated cultures harvested at 24 and 48 h. Furthermore, G₂/M cells were detected at 48 h in arsenite-treated cultures in the presence of z-vad, suggesting that caspase inhibition allows arsenite damaged cells to progress into at least a second G₂/M rather than die (see circled areas in Fig. 8).

View larger version (42K): [in this window] [in a new window]   Fig. 8.   Inhibition of caspase-3 with z-vad reduces apoptosis but does not alter cell cycle dysregulation. U-937 cells (2 × 105 cells/ml) were cultured in the absence (left) or presence (right) of 5 µM arsenite for 24 h (top) or 48 h (bottom), and as indicated in the figure in the presence or absence of 50 µM zvad.fmk. At 24 or 48 h cells were harvested and processed for analysis of DNA content and side scatter by flow cytometry.

The involvement of caspases in arsenite-induced cell death also was determined via cell survival/viability assays by using the live cell dye calcein-AM and by using propidium iodide uptake to measure plasma membrane integrity. The data in Table 1 show that caspase inhibition enhances resistance to the cytotoxic effects of arsenite at high concentrations.

The active metabolite of vitamin D, 1,25-dihydroxyvitamin-D₃, is an inducer of differentiation of myeloid leukemia cells such as U937. 1,25-dihydroxyvitamin-D₃ promotes the maturation of myeloid cells, which involves heightened surface expression of the marker CD11b (i.e., Mac-1) and increased capacity for a respiratory burst (as monitored by increased reduction of NBT).

Incubation of U937 cells with 1,25-dihydroxyvitamin-D₃ for 96 h increased the cell surface expression of CD11b on a portion (range 6 to 12%) of cells (Table 2). Coincubation with 1 µM arsenite nearly doubled (range 12 to 21%) the percentage of cells expressing 1,25-dihydroxyvitamin-D₃-inducible CD11b. Arsenite, alone, did not promote the differentiation of U937 cells; nor did it augment the expression of CD11b that was up-regulated by a stronger inducer of myeloid cell differentiation, PMA.

View this table: [in this window] [in a new window]   TABLE 2 Arsenite enhances U937 cell differentiation (Mac-1) U937 cells (2 × 105 cells/ml) were incubated with or without 0.5 µM arsenite and stimulated with or without 50 nM 1,25-dihydroxyvitamin-D3 or 10 nM PMA for 4 days. At the conclusion of the culture period, cells were harvested, and cell surface Mac-1 expression was determined using R-PE-conjugated anti-Mac-1 and flow cytometry. Data values are the mean ± S.D. and represent triplicate determinations in each experiment.

Reduction of NBT also is a functional marker of monocytic differentiation. NBT reduction reflects the induction of NADPH oxidase, which is necessary for the formation of superoxide during the respiratory burst that occurs in mature monocytes. Assessment of differentiation by NBT reduction in response to stimulation with 1,25-dihydroxyvitamin-D₃ in the presence and absence of 0.5 and 1 µM arsenite was conducted as described under Materials and Methods. The data in Table 3 show that as was the case with CD11b expression, coincubation with arsenite for 96 h enhances 1,25-dihydroxyvitamin-D₃-induced monocyte differentiation as assessed by NBT reduction.

View this table: [in this window] [in a new window]   TABLE 3 Arsenite enhances U937 cell differentiation (NBT) U937 cells (2 × 105 cells/ml) were incubated with or without 0.5 µM arsenite and stimulated with or without 50 nM 1,25-dihydroxyvitamin-D3 for 4 days. At the conclusion of the culture period, cells were harvested, and a respiratory burst was stimulated using 200 ng/ml PMA. Cells capable of reducing NBT and forming formazan deposits were scored microscopically. Data values are the mean ± S.D. number of NBT+ cells expressed as a percentage of the total viable cell number for each group.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

As₂O₃, yielding trivalent arsenite, has shown promise as a useful therapeutic drug for the treatment of APL (Soignet et al., 1998; Niu et al., 1999). However, the mechanism by which eradication of leukemic cells is achieved in vivo is unclear (Kroemer and de Thé, 1999). Although a multitude of in vitro studies using APL cell lines suggests that the mechanism of eradication involves apoptosis (Chen et al., 1996; Ma et al., 1998; Huang et al., 1999b; Jing et al., 1999; Larochette et al., 1999; Zhu et al., 1999; Perkins et al., 2000), it is conceivable that this pleiotropic agent may influence additional cellular processes that limit leukemic expansion. In addition to potentially inducing apoptosis, arsenite reportedly has been demonstrated to impair cell cycle progression (Bazarbachi et al., 1999; Lu et al., 1999; Zhu et al., 1999) and to promote differentiation (Kizaki et al., 1998; Cai et al., 2000) of various leukemic cell lines. Each of these processes, apoptosis, growth inhibition, and differentiation, acting alone or in concert may limit tumor growth. The fact that arsenite influences all three may explain why it is clinically promising. In view of the large number of potential cellular targets and the interplay between apoptosis, cell cycle, and differentiation, issues such as dose, duration of treatment, cellular environment, and cell type become critical. That is, at a given dose arsenite may not affect apoptosis, cell growth, and differentiation in all cell types. Indeed, clear differences in the susceptibility of a variety of cell lines to arsenite have been reported (Dai et al., 1999).

In the present study we used a single model cell line, U937, and describe differences in arsenite sensitivity based on cell cycle phase. High concentrations of arsenite (i.e., >10 µM) are clearly cytotoxic, where cell death is induced irrespective of cell cycle phase. Such high concentrations of arsenite are probably not pharmacologically relevant. At an intermediate concentration of arsenite (i.e., ~5 µM) cell cycle progression is hindered, where both G₁ and G₂/M transit are delayed. Arsenite had no effect on progression through S phase, suggesting that cells in S phase are resistant to the cytostatic effects of arsenite. The reason why S phase cells are relatively resistant to arsenite has not been established. G₁ cells are less sensitive than G₂/M cells in that the duration of the G₁ delay (~4 h) is less than the G₂/M delay (~8 h). Furthermore, U937 cells apparently can overcome the G₁ delay and progress into S phase, whereas the apoptotic process, together with caspase activation, occurs after the more prolonged G₂/M delay. Park et al. (2000) have reported recently that arsenite treatment of myeloma cell lines induces a growth arrest in both G₁ and G₂/M. Based on the CFSE data (Fig. 2) and the experiments with elutriated cells (Fig. 6), we conclude that arsenite causes a delay, but not an arrest, in progression through G₁ and G₂/M. The delay in G₂/M transit appears to lead to caspase-3 activation and apoptosis after cell division or an aborted mitosis. Thus, the overall effect of arsenite treatment may be an eventual cell death, where "delayed" rather than "acute" apoptosis is induced after damage resulting from the cytostatic influences of arsenite.

Our observations regarding dysregulation of cell cycle progression and induction of apoptosis after arsenite treatment share similarities with the conclusions drawn by Ma et al. (1998) who demonstrated selective induction of apoptosis upon arsenite treatment of NB4 cells arrested in G₂/M (Ma et al., 1998). Their conclusion can now be refined by adding that cells survive the G₂/M delay, subsequently divide, and ultimately undergo apoptosis after division. This conclusion is supported by the fact that cell division, albeit at a reduced rate, does occur in arsenite-treated cells and that after an initial G₂/M delay and the damage associated with it, progression into subsequent rounds of G₂/M does not occur.

Others have shown that arsenite activates caspase-3 in HL60 cells, NB4 cells, and in neuroblastoma cell lines (Akao et al., 1999; Huang et al., 1999b; Larochette et al., 1999). As was the case with arsenite-induced apoptosis, arsenite-induced activation of caspase-3 in U937 cells is concentration, time, and cell-cycle phase dependent. The cytotoxicity of arsenite at high concentrations involves caspase-3 activation, which can be detected relatively early, within 6 h after treatment with 20 µM arsenite. Even cells in G₁ are sensitive to arsenite-induced caspase-3 activation at 20 µM arsenite (Fig. 7B). Pharmacological inhibition of caspase-3 with z-vad enhances the survival of U937 cells treated with high concentrations of arsenite, which indicates that arsenite-induced cytotoxicity likely involves caspase-3 activation.

The kinetics of caspase-3 activation by 5 µM arsenite correlates with the arsenite-induced G₂/M delay, and caspase inhibition diminishes the acquisition of the apoptotic phenotype. Using asynchronous cells caspase-3 activity is barely detectable after 24 h of treatment (Fig. 7A). With synchronous cells at 24 h after treatment with 5 µM arsenite, caspase-3 activity is more readily detected (Fig. 7B). This time corresponds to the period when the bulk of the arsenite-treated cells is emerging from the G₂/M delay. Allowing the synchronized cells to progress into G₂/M and then treating with 5 µM arsenite (i.e., t = 16 h) shows that the arsenite-induced caspase-3 activity, which is measured 8 h later when the cells are emerging from G₂/M, is cell cycle dependent as opposed to being dependent solely on the duration of treatment.

Although caspase inhibition enhanced the survival of U937 cells treated with high concentrations of arsenite, such cells continued to exhibit delayed cell cycle progression. This observation is in disagreement with the findings of Larochette et al. (1999), where caspase inhibition was reported to prevent arsenite-induced DNA fragmentation but not arsenite-induced cell death. Given that Larochette et al. (1999) scored cell survival as cell recovery, we think it likely that the effects of arsenite on cell cycle progression were not taken into consideration by them.

At low concentrations (i.e., 1 µM) arsenite enhanced the 1,25-dihydroxyvitamin-D₃-mediated differentiation of U937 cells as measured by CD11b expression and NBT reduction. Differentiation of U937 cells induced by 1,25-dihydroxyvitamin-D₃ involves cell cycle arrest in G₁ (Rots et al., 1999), which may allow the differentiating effects of low-dose arsenite to be realized. The cell differentiation-inducing capacity of arsenite, which has been observed in other studies (Kizaki et al., 1998; Cai et al., 2000; Perkins et al., 2000), may prove to be clinically useful because it occurs at the lowest concentrations where the toxicity of arsenite is likely less of an issue. These lower concentrations of arsenite also are more likely to be clinically achievable.

In summary, we have studied the mechanisms whereby arsenite diminishes cell proliferation and compromises cell survival as a model for the metalloid's clinical use in limiting the expansion of acute promyelocytic leukemia. Using the myelomonocytic leukemia cell line U937, we studied various aspects of altered cell accumulation in the presence of arsenite, including cell cycle dysregulation, apoptosis, cytotoxicity, and differentiation. Arsenite exposure clearly influenced each of these processes in a concentration-dependent manner. Cell cycle progression and differentiation were found to be more sensitive to arsenite than is the induction of apoptosis because these cellular processes were affected at concentrations of arsenite believed to be clinically achievable. The influences of arsenite on cell cycle progression are complex, with G₂/M transit more sensitive than G₁ transit and S phase transit hardly affected at all by arsenite. Apoptosis appears to follow the delay in G₂/M after an aberrant cell division. Thus, the present results, showing that cell cycle progression is more sensitive to arsenite, are somewhat in contrast to the prevailing thinking that induction of apoptosis is the key to arsenite's chemotherapeutic efficacy.