Interactions between the cyclin-dependent kinase inhibitor (CDKI) flavopiridol (FP) and phorbol 12-myristate 13-acetate (PMA) were examined in U937 human leukemia cells in relation to differentiation and apoptosis. Simultaneous, but not sequential, exposure of U937 cells to 100 nm FP and 10 nm PMA significantly increased apoptosis manifested by characteristic morphological features, mitochondrial dysfunction, caspase activation, and poly(ADP-ribose) polymerase cleavage while markedly inhibiting cellular differentiation, as reflected by diminished plastic adherence and CD11b expression. Enhanced apoptosis in U937 cells was associated with an early caspase-independent increase in cytochrome c release and accompanied by a substantial decline in leukemic cell clonogenicity. Moreover, PMA/FP cotreatment significantly increased apoptosis in HL-60 promyelocytic leukemia cells and in U937 cells ectopically expressing the Bcl-2 protein. In U937 cells, coadministration of FP blocked PMA-induced expression and reporter activity of the CDKI p21WAF1/CIP1 and triggered caspase-mediated cleavage of the CDKI p27KIP1. Coexposure to FP also resulted in a more pronounced and sustained activation of the mitogen-activated protein kinase kinase/extracellular signal-regulated protein kinase cascade after PMA treatment, although disruption of this pathway by the mitogen-activated protein kinase kinase 1 inhibitor U0126 did not prevent potentiation of apoptosis. FP accelerated PMA-mediated dephosphorylation of the retinoblastoma protein (pRb), an event followed by pRb cleavage culminating in the complete loss of underphosphorylated pRb (≈Mr 110,000) by 24 h. Finally, gel shift analysis revealed that coadministration of FP with PMA for 8 h led to diminished E2F/pRb binding compared to the effects of PMA alone. Collectively, these findings indicate that FP modulates the expression/activity of multiple signaling and cell cycle regulatory proteins in PMA-treated leukemia cells and that such alterations are associated with mitochondrial damage and apoptosis rather than maturation. These observations also raise the possibility that combining CDKIs and differentiation-inducing agents may represent a novel antileukemic strategy.
Eukaryotic cell cycle progression is governed by the sequential activation and inactivation of various cyclin/CDK3 complexes. Levels of the cyclin proteins (e.g., cyclins D, E, A, and B) fluctuate temporally during cell cycle traverse through G1, S, G2, and M (1), leading to activation of their respective CDKs (CDK4/6, CDK2, and CDK1) at the appropriate time. A major function of the cyclin D/CDK4 complex is to phosphorylate the pRb at or near the restriction checkpoint of G0-G1. This, in turn, promotes cellular proliferation and S-phase progression through liberation of the E2F transcription factor normally bound to hypophosphorylated pRb (2). Free E2F triggers the expression of diverse genes related to S-phase progression, including dihydrofolate reductase and thymidylate synthetase (3). Conversely, CDKIs such as p21WAF1/CIP1 and p27KIP1 play an important role in cell cycle control by coordinating internal and external signals that impede cell cycle progression at key checkpoints (4). When the CDKIs p21WAF1/CIP1 and p27KIP1 are induced, CDKs are inhibited, resulting in pRb dephosphorylation (5). Hypophosphorylated pRb, in turn, binds to and inactivates E2F, thereby promoting cell cycle arrest in G1.
PMA is a protein kinase C activator and tumor promoter that induces terminal differentiation in human leukemia cells (6). Leukemic cell maturation triggered by PMA or other differentiation inducers requires exit from the cell cycle and G1 arrest (7). The cell cycle arrest induced by PMA is preceded by increased expression of the CDKIs p21WAF1/CIP1 and p27KIP1 in a p53-independent manner (6). The role of CDKIs in leukemic cell maturation has been underscored by the observation that stable expression of a p21WAF1/CIP1 antisense construct in U937 monocytic leukemia cells blocks PMA-induced differentiation and reciprocally promotes apoptosis in these cells (8). Conversely, enforced expression of CDKIs has been linked to cytotoxic drug resistance and protection from apoptosis (9, 10). Furthermore, it has been reported that p21WAF1/CIP1 can form a complex with procaspase-3 to inhibit Fas-mediated apoptosis, raising the possibility of a direct role for this CDKI in preventing cell death (11). Taken together, these findings provide evidence that CDKIs such as p21WAF1/CIP1 oppose apoptosis and suggest a possible mechanism by which induction of leukemic cell maturation may, under some circumstances, protect cells from genotoxic stresses (12).
Recently, attention has focused on the development of pharmacological inhibitors of CDKs as anticancer agents (13). FP (L86–8275; NCS 649890) is a CDKI that is currently undergoing Phase I/II clinical trials in humans (14). FP interacts with the adenine-binding pocket of CDKs to inhibit kinase activity at concentrations of ∼100 nm for CDKs 1, 2, 4, and 6 and 300 nm for CDK7, the CDK-activating kinase (15). In addition, FP decreases the expression of cyclins D1, D3, and E without modifying cyclin D2 protein levels (16). Concordant with its ability to block the expression of various cyclins and inhibit CDK activity, FP induces G1 and/or G2-M cell cycle arrest. Furthermore, FP is a potent inducer of apoptosis in human leukemia cells (17) and also potentiates the lethality of conventional cytotoxic agents in a sequence-dependent manner (18). Although the mechanism by which FP promotes apoptosis in leukemia cells remains unknown, it is presumed that FP-mediated cytotoxicity stems from cell cycle perturbations (19), particularly in view of abundant evidence that disruption of cell cycle progression represents a potent apoptotic stimulus (20).
Aside from studies involving cytotoxic agents, little information is currently available concerning interactions between FP and other classes of drugs, notably differentiation-inducing agents (e.g., PMA). In fact, a rationale exists for combining differentiation inducers with agents such as FP. Given the known requirement for G1 arrest in leukemic cell differentiation programs (7), the capacity of FP to block cell cycle progression (21) would ordinarily be expected to lower the differentiation threshold. To test this hypothesis, the effects of FP on PMA-mediated differentiation were examined in human leukemia cells (U937). Contrary to expectations, FP opposed rather than promoted PMA-related maturation events. Furthermore, FP-mediated actions resulted in dysregulation of various proteins and signaling pathways associated with PMA-induced G1 arrest and differentiation, including inhibition of p21WAF1/CIP1 induction, acceleration of pRb dephosphorylation, caspase-mediated cleavage of pRb and p27KIP1, enhanced MAPK activation, and diminished E2F/pRb binding. Together, these findings indicate that the CDKI FP disrupts leukemic cell maturation responses to PMA and may thereby direct cells along an alternative cell death pathway (22).
MATERIALS AND METHODS
Drugs, Biologicals, and Chemical Reagents.
PMA (Sigma, St. Louis, MO) was dissolved in DMSO, and aliquots were stored at −20°C. FP (L86 8275; NCS 649890) was kindly provided by Dr. Edward Sausville (Cancer Treatment and Evaluation Program, National Cancer Institute, Bethesda, MD). FP was formulated in DMSO, and 10−2 m stock solutions were stored at −20°C. The mitochondrial dye DiOC6 was purchased from Molecular Probes (Eugene, OR). The antimetabolite ara-C was dissolved in sterile PBS, and 10−3 m aliquots were stored at 4°C. Hygromycin B was obtained from Boehringer Mannheim (Mannheim, Germany). PD98059 and U0126 were purchased from Calbiochem (La Jolla, CA) and formulated in DMSO according to the manufacturer’s instructions. The pan-caspase inhibitor B-d-FMK was purchased from Enzyme Systems Products (Livermore, CA). Primary antibody for phosphorylated ERK1 and ERK2 was provided in the PhosphoPlus p44/p42 MAP Kinase Antibody Kit (New England Biolabs, Beverly, MA). Primary antibodies for p21, p27, and actin were purchased from Transduction Laboratories (Lexington, KY). Primary antibodies for underphosphorylated pRB and procaspase-3 were obtained from PharMingen (San Diego, CA). The primary antibody for PARP was purchased from Biomol Research Laboratories (Plymouth Meeting, PA). Secondary antibodies conjugated to horseradish peroxidase were obtained from Kirkegaard and Perry Laboratories, Inc. (Gaithersburg, MD). Coomassie protein assay reagent was purchased from Pierce (Rockford, IL), and an enhanced chemiluminescence kit was obtained from New England Nuclear (Boston, MA). Hypoosmolar buffer for electroporation was purchased from Eppendorf Scientific, Inc. (Westbury, NY), and luciferase assay reagents were obtained from Promega (Madison, WI). All other chemicals or reagents were purchased from Sigma.
The myelomonocytic leukemia cell line U937 was obtained from American Type Culture Collection. The HL-60 cell line was derived from a patient with acute promyelocytic leukemia as described previously (23). U937 and HL-60 cells were cultured in suspension in phenol red-free RPMI 1640 (Life Technologies, Inc., Grand Island, NY) and 10% (v/v) FCS (Hyclone, Logan, UT) and maintained in a humidified atmosphere (95% air:5% CO2) at 37°C. To obtain antisense-expressing cell lines, U937 cells were transfected by electroporation with a pREP4 vector (Invitrogen, Carlsbad, CA) or the pREP4 vector containing the p21WAF1/CIP1 coding region in an antisense orientation as described previously (24). Transfectant U937 leukemia cells stably overexpressing the antiapoptotic protein Bcl-2 were obtained as reported previously (25). These cells, designated as U937/Bcl-2 cells, were generated along with their empty vector counterparts (i.e., U937/pCEP4). U937 transfectant cell lines were maintained as described above in the presence of hygromycin B (400 μg/ml) and transferred to selection-free media 24 h before experimentation. All experiments were performed on cells in logarithmic phase.
Morphological Assessment of Apoptosis.
U937 cells were evaluated for apoptosis by morphological assessment of Wright and Giemsa-stained cytospin preparations. At designated times, cells were transferred to slides by cytocentrifugation, fixed, stained, and evaluated under light microscopy for treatment-induced apoptosis. Apoptotic cells were identified by classical morphological features (i.e., nuclear condensation, cell shrinkage, and formation of apoptotic bodies). Five or more randomly selected fields, encompassing a total of ≥500 cells/slide, were evaluated to determine the percentage of apoptotic cells for each treatment condition. The extent of apoptosis determined by this method has been shown to correlate closely with results obtained using the terminal deoxynucleotidyl transferase-mediated nick end labeling assay.4
Assessment of Drug Interactions.
Interactions between PMA and FP were characterized as described previously by Chou and Talalay (26). The percentages of apoptotic cells were assessed after PMA (1–10 nm) or FP (10–100 nm) treatment alone or in combination for 24 h at a constant PMA:FP dose ratio (1:10). The CI was determined using Median Dose Effect Analysis Software (Elsevier Biosoft; Cambridge, United Kingdom). CI values < 1 correspond to synergistic drug interactions.
Equal quantities of protein (15 μg/condition for MAPK, 40 μg/condition for cytochrome c, or 25 μg/condition) were separated by SDS-PAGE [underphosphorylated pRB (7.5%), PARP (8.0%), p44/p42 MAPK (10%), p21, p27, procaspase-3 (12%), and cytochrome c (15%)] and electroblotted onto nitrocellulose. Western analysis of p44/p42 MAPK protein was performed according to the manufacturer’s instructions (New England Biolabs); otherwise, studies were performed as described previously (8). Briefly, U937 cells (5 × 106) were pelleted by centrifugation, resuspended in 50 μl of PBS, lysed by the addition of 2× Laemmli buffer, and boiled for 5 min. Proteins were quantified with Coomassie protein assay reagent. Blots were blocked in PBS-T/5% milk, washed twice with PBS-T, and incubated overnight at 4°C with the appropriate primary antibody. The blots were washed with PBS-T and incubated with a horseradish peroxidase-conjugated secondary antibody diluted appropriately in PBS-T/5% milk. After incubation, blots were developed by enhanced chemiluminescence exposure to Kodak X-OMAT film (Eastman Kodak Co., Rochester, NY) and reprobed with antibodies directed against actin to control for equal loading of protein.
Assessment of Mitochondrial Membrane Potential.
At designated times, 1-ml aliquots of cells (2 × 105) were harvested and incubated with 40 nm DiOC6 for 15 min at 25°C as described previously (27). Samples were analyzed using a Becton Dickinson (Mansfield, MA) FACScan flow cytometer (excitation λ = 488 nm; emission λ = 525 nm). Results were expressed as the percentage of total cells exhibiting loss of mitochondrial membrane potential (ΔΨm), as manifested by a reduction in DiOC6 uptake relative to untreated controls. Data acquisition and analysis were performed using CELLQUEST Software (Becton Dickinson).
Cytochrome c Release.
The assay for cytochrome c release was performed as described previously by Single et al. (28), with modifications. At designated times after drug treatment, 4 × 106 cells were washed with PBS and resuspended in 50 μl of assay buffer (75 mm NaCl, 1 mm NaH2PO4, 8 mm Na2HPO4, 1 mm EDTA, and 250 mm sucrose) containing digitonin (700 μg/ml). Three minutes after the addition of digitonin assay buffer, cells were pelleted by centrifugation. The supernatants were transferred to tubes containing 2× Laemmli buffer (50 μl) and boiled for 5 min. Proteins were separated by SDS-PAGE as described above.
Clonogenic Survival Assays.
Clonogenic survival assays assessed the effects of 10 nm PMA, FP (50 or 100 nm), or PMA/FP cotreatments on leukemic cell self-renewal capacity. Cell densities were determined after 24-h drug treatment using a Coulter Counter (Coulter Electronics, Hialeah, FL). Cells were pelleted by centrifugation and resuspended in fresh media to achieve a final cell density of 3 × 105 cells/ml. A total of 4500 cells were plated in soft agar cloning medium for each condition as described previously (29). Colonies, defined as groups of ≥50 cells, were scored after 12 days of incubation. Clonogenic survival in drug-treated samples was expressed as a percentage relative to untreated controls.
Expression of the monocytic differentiation marker CD11b was monitored by direct immunofluorescence staining and flow cytometric analysis as described previously (30). After drug treatment, suspension and adherent cells were enumerated by a Coulter Counter, and 2 × 106 cells were pelleted by centrifugation. The supernatant was aspirated, and the cells were resuspended in 300 μl of ice-cold PBS. Two 100-μl aliquots from each sample were then combined with either phycoerythrin-1 (10 μl) or the IgG control. Samples were incubated for 20 min at 4°C and diluted in PBS (1 ml). Sample data were collected using a Becton Dickinson FACScan flow cytometer and analyzed with Verity Winlist Software (Verity, Topsham, ME). Differentiation was also monitored by determining the percentage of U937 cells exhibiting plastic adherence after drug treatments as described previously (31).
p21WAF1/CIP1 Luciferase Reporter Assay.
U937 cells were transfected with a full-length p21WAF1/CIP1 promoter (bp 1–2326) as described by Chinery et al. (32) using electroporation (Eppendorf Multiporator; 620V, 60 μs) in hypoosmolar buffer. In each experiment, pSV-βgal (Promega) containing the β-galactosidase gene under the control of the constitutively active SV40 promoter and enhancer was cotransfected with the luciferase reporter at a ratio of 1:5 (w/w) to normalize for transfection efficiency. The cells were incubated overnight in complete media and treated with 10 nm PMA, 100 nm FP, or 10 nm PMA/100 nm FP for 3 or 6 h. At the indicated times, cells were washed in serum-free media, resuspended in 0.25 m Tris-HCl (pH 7.8), lysed by rapid freeze-thaw, and centrifuged, and the supernatant was frozen at −70°C. Lysates (20 μl) were mixed with luciferase assay reagent (100 μl) and monitored for 20 s in a luminometer (Monolight 2010). For β-galactosidase activity, lysates (50 μl) were mixed in triplicate with an equal volume of luciferase assay buffer (2×) in a 96-well plate and incubated overnight. The reaction was terminated by the addition of NaCO3, and absorbance was measured at 405 nm in a Vmax plate reader (Molecular Devices). Activity was determined by comparison to a standard curve. Luciferase reporter activity in relative luminescence units was normalized to milliunits of β-galactosidase activity as described by Rosenthal (33).
EMSA and IP.
Whole-cell extracts were prepared as described previously (34), and EMSA after IP was performed as we have reported previously (35). Whole-cell extracts (150 μg) were combined with 4 μl of anti-RB monoclonal antibody (Calbiochem) for the IP. Immunoprecipitated proteins were recovered on protein A-Sepharose beads and dissociated by treatment with deoxycholate. The presence of E2F in the IP was monitored by EMSA as noted previously (36).
Significant differences between experimental values were determined using Student’s t test.
FP and PMA Promote Apoptosis in Human Leukemia Cells.
To evaluate the dose response for FP-induced apoptosis, U937 cells were exposed to various concentrations of FP (e.g., 50–400 nm) for 6, 9, 12, or 24 h, after which time apoptosis was determined by morphological assessment of Wright and Giemsa-stained cytospin preparations (Fig. 1,A). U937 cells were susceptible to apoptosis after chronic exposures (e.g., 12 or 24 h) to FP concentrations of ≥200 nm, whereas concentrations of ≤100 nm minimally induced apoptosis (e.g., ≤10% of cells), regardless of the treatment interval. Subsequently, the extent of apoptosis was assessed in U937 cells exposed to 100 nm FP in combination with a minimally toxic concentration of PMA (10 nm). Three schedules were evaluated: (a) coadministration of the agents for 24 h; (b) 10 nm PMA for 24 h, followed by 100 nm FP for 24 h; or (c) 100 nm FP for 24 h, followed by 10 nm PMA for 24 h. Coadministration of 10 nm PMA and 100 nm FP for 24 h (Fig. 1,B) resulted in a significant increase in apoptosis (69%; P < 0.0001); in contrast, sequential schedules only marginally increased cell death (apoptosis < 20%). Cotreatment with PMA/FP also effectively induced apoptosis in U937 cells overexpressing the antiapoptotic protein Bcl-2 (56%), whereas these cells were significantly protected from the lethal effects of the antimetabolite ara-C (1.0 μm × 6 h; Fig. 1,C). Coadministration of FP (200 nm) and PMA (10 nm) also resulted in a substantial increase in apoptosis (70%; P < 0.0004) in the human promyelocytic leukemia cell line HL-60 (Fig. 1 D). Thus, coadministration of minimally toxic concentrations of FP and PMA provided a potent apoptotic stimulus in several myeloid leukemia cell types, including those expressing high levels of Bcl-2.
Median dose effect analysis (26) was used to characterize interactions between FP and PMA over a range of drug concentrations with respect to induction of apoptosis (Fig. 2,A). CI values were consistently < 0.01, indicating a highly synergistic drug interaction. Evidence for enhanced proteolytic cleavage of procaspase-3 (37) and degradation of PARP, one of the major caspase-3 substrates (38), was apparent at each time interval examined for cells exposed to PMA/FP (Fig. 2,B). Lastly, the increase in apoptosis after PMA/FP cotreatment was accompanied by a marked reduction in clonogenic survival (Fig. 2 C). The greatest reduction in clonogenic survival occurred with the 10 nm PMA/100 nm FP combination, which was significantly more inhibitory than the effects of PMA alone (<1% versus 15%; P < 0.03). Together, these findings demonstrate that coadministration of FP (100 nm) with PMA (10 nm) promotes caspase activation and apoptosis in human leukemia cells; moreover, these events are accompanied by a marked decline in leukemic cell clonogenicity.
FP Combined with PMA Promotes Mitochondrial Damage Upstream of Caspase Activation.
The effects of FP and PMA on mitochondrial injury [e.g., loss of mitochondrial membrane potential (Δψm)], an event that can precede the morphological features of apoptosis (39), were assessed at early time points (e.g., 2–8 h; Fig. 3,A). The combination of FP (100 nm) and PMA (10 nm) triggered the greatest decline in Δψm at the 4–8 h exposure intervals, although reductions were noted in cells treated with FP alone at these times. Coadministration of the pan-caspase inhibitor B-d-FMK (20 μm) effectively blocked the loss of Δψm in cells exposed to PMA and FP, suggesting that this event represents a consequence of caspase activation. In addition, U937 cells treated with the combination of PMA and FP for 6 h displayed a more pronounced release of cytochrome c into the cytosol than cells treated with PMA or FP alone, although FP alone had some effect (Fig. 3 B). However, in contrast to the loss of Δψm, the cytochrome c release induced by PMA/FP cotreatment was not blocked by B-d-FMK (20 μm). Together, these findings demonstrate that coadministration of FP (100 nm) with PMA (10 nm) promotes early mitochondrial damage in human leukemia cells and suggest that release of cytochrome c into the cytoplasm represents an upstream event in the apoptotic process.
FP Antagonizes PMA-induced Differentiation.
Induction of leukemic cell maturation can be accompanied by apoptosis (40); therefore, it is conceivable that PMA/FP-mediated cell death could result from enhanced cellular differentiation. Thus, the effects of FP were examined with respect to PMA-related induction of the monocytic maturation marker CD11b (Fig. 4,A). FP did not enhance PMA-mediated induction of CD11b at 24 h; instead, FP cotreatment significantly reduced the percentage of cells displaying CD11b expression from 26% to 9.0% (P < 0.03). FP (100 nm) cotreatment also significantly decreased PMA-induced plastic adherence, another manifestation of U937 cellular maturation, from 80% to 16% (P < 0.001; Fig. 4 B). Thus, despite its CDK inhibitory activity, FP significantly antagonized PMA-induced differentiation in U937 cells.
Effects of FP on pRb and CDKI Expression in PMA-treated Cells.
The effects of FP on proteins involved in the PMA-induced G1 arrest program (e.g., pRb and the CDKIs p21WAF1/CIP1 and p27KIP1) were examined by Western analysis (Fig. 5). Exposure of cells to 10 nm PMA for 24 h resulted in a significant increase in the underphosphorylated pRb species (Lane 2), whereas FP alone (50, 80, or 100 nm; Lanes 3–5) had a minimal effect on pRb dephosphorylation at 24 h. When cells were coexposed to PMA (10 nm) and FP (50, 80, or 100 nm; Lanes 6–8), there was a marked reduction in full-length underphosphorylated pRb (≈Mr 110,000), which was accompanied by the appearance of an immunoreactive Mr 68,000 band. The latter presumably corresponds to a pRb cleavage product resulting from the actions of interleukin 1β-converting enzyme-like proteases in leukemic cells undergoing apoptosis (41). Similar findings were obtained when total pRb was examined after cotreatment with PMA and FP for 24 h (data not shown). Furthermore, exposure of U937 cells to 10 nm PMA for 24 h resulted in a significant increase in p21WAF1/CIP1 expression and a slight increase in p27KIP1 expression (Fig. 5), whereas FP alone had no effect on the expression of these CDKIs. In contrast, marked reductions in levels of p21WAF1/CIP1 and p27KIP1 were observed after 24 h of cotreatment, particularly at the highest FP concentration examined (e.g., 100 nm). Thus, a 24-h exposure to FP significantly reduced the amount of full-length underphosphorylated pRb induced by PMA and markedly diminished expression of the CDKIs p21WAF1/CIP1 and p27KIP1.
Additional studies were conducted at early time intervals to determine whether observed reductions in CDKI expression after PMA/FP cotreatment were the result of caspase activation during apoptosis, a phenomenon that has been described previously (42). U937 cells were exposed to the combination of 10 nm PMA and 100 nm FP in the presence or absence of the pan-caspase inhibitor B-d-FMK (20 μm), and the extent of apoptosis (Fig. 6,A) or CDKI protein expression (Fig. 6,B) was assessed over the ensuing 2–12 h. Addition of 20 μm B-d-FMK effectively antagonized apoptosis in cells exposed to PMA and FP for the duration of the time course. The apoptosis percentage after 12 h of cotreatment was 15% and 45% in the presence or absence of 20 μm B-d-FMK, respectively (Fig. 6,A). As shown in Fig. 6,B, FP significantly attenuated PMA-induced p21WAF1/CIP1 expression in the presence or absence of 20 μm B-d-FMK, indicating that FP interferes with p21WAF1/CIP1 through a caspase-independent mechanism. In contrast, p27KIP1 was cleaved by caspases within 4 h of PMA/FP cotreatment, based on the observation that formation of a recently described (43) Mr 23,000 cleavage fragment (Fig. 6,B, CF) was prevented by the addition of 20 μm B-d-FMK. Finally, transient transfection of U937 cells with a p21WAF1/CIP1 luciferase reporter plasmid containing the full-length promoter revealed that PMA alone markedly increased p21WAF1/CIP1 reporter activity at 3 and 6 h, whereas coincubation of FP reduced activity to below basal levels (Fig. 6 C). This finding suggests that FP-mediated antagonism of p21WAF1/CIP1 expression may occur at the transcriptional level.
Early Effects of FP on p21WAF1/CIP1, pRb, and Phospho-ERK in PMA-treated Cells.
Parallel time course studies were conducted to examine early FP-mediated perturbations in the expression of various differentiation-related proteins, including p21WAF1/CIP1, underphosphorylated pRb, and phosphorylated p42/p44 MAPK (i.e., ERK1 and ERK2). Consistent with previous observations, PMA (10 nm) induced p21WAF1/CIP1 in U937 cells within 1 h of drug addition, and expression was very marked by 3–6 h (Fig. 7); moreover, induction of this CDKI was substantially attenuated by FP (100 nm) coadministration. In addition, FP cotreatment accelerated the temporal pattern of pRb dephosphorylation relative to PMA alone by approximately 2 h (e.g., 30 min versus 3 h), consistent with the ability of FP to inhibit multiple CDKs (15). Evidence of pRb cleavage to a Mr 68,000 subfragment, as well as PARP degradation (data not shown), was discernible at 6 h in cells treated with the combination of PMA and FP. Early alterations in MAPK phosphorylation (activation) were also observed. Thus, treatment of cells with PMA alone resulted in an increase in phosphorylated ERK1/ERK2 that was apparent after 30 min, maximal after 1 h, and declined slightly over the ensuing 5 h. Treatment with FP alone modestly increased ERK activation at 3–6 h. However, coadministration of FP with PMA led to a more pronounced and sustained activation of ERK compared to the effects observed with PMA alone (Fig. 7). In each case, total ERK protein remained constant throughout. In separate studies, coadministration of FP did not lead to alterations in PMA-mediated phosphorylation of c-Jun NH2-terminal kinase (data not shown). Thus, coadministration of FP with PMA induced multiple early perturbations in differentiation-related proteins and signaling pathways, including (a) antagonism of p21WAF1/CIP1 induction, (b) acceleration of pRb dephosphorylation, (c) cleavage of pRb and p27KIP1, and (d) a more pronounced and sustained activation of ERK1/ERK2.
Functional Role of ERK Activation in PMA/FP-mediated Apoptosis.
In view of evidence that stimulation of the ERK cascade can either promote (44) or inhibit (45) apoptosis, depending on the stimulus and cell type, attempts were made to define the functional significance of enhanced ERK activation in cells treated with the PMA/FP combination (Fig. 8). Consistent with the previous results, a 1-h exposure to PMA (10 nm) induced ERK activation, and this effect was more pronounced in cells treated with PMA and FP (Fig. 8,A). In both cases, ERK activation was opposed by the addition of specific MEK1 inhibitors PD90859 (25 μm) or, to an even greater extent, by the addition of U0126 (25 μm), which has a higher affinity for the ATP binding site of MEK1 than PD98059 (46). Addition of U0126 slightly increased PMA/FP-induced apoptosis from 28% to 37% after 8 h of cotreatment (Fig. 8 B), arguing against the possibility that enhanced activation of the MEK1/ERK pathway is responsible for potentiation of PMA-induced apoptosis by FP.
Functional Role of Diminished p21WAF1/CIP1 Expression in PMA/FP-mediated Apoptosis.
Attempts were made to define the functional role of diminished p21WAF1/CIP1 expression in PMA/FP-mediated apoptosis using a U937 transfectant cell line stably expressing p21WAF1/CIP1 in the antisense configuration (i.e., U937/p21AS; Ref. 8). It was hypothesized that if disruption of p21WAF1/CIP1 expression by FP contributes functionally to PMA/FP-mediated apoptosis, then U937/p21AS cells would be less susceptible to potentiation of apoptosis because induction of p21WAF1/CIP1 is already impaired. To evaluate this possibility, U937/p21AS cells and their empty vector counterparts (i.e., U937/pREP4) were exposed to FP in conjunction with a low concentration of PMA (1 nm). The latter concentration was selected to limit the extent of PMA-mediated apoptosis in U937/p21AS cells (8). After exposure to 1 nm PMA and 50 nm FP, enhancement of apoptosis was equivalent in U937/p21AS (29% apoptotic) and empty vector controls (26% apoptotic) as shown in Fig. 9. However, when cells were exposed to 1 nm PMA and 80 nm FP, the extent of apoptosis was clearly more pronounced in U937/p21AS cells than in empty vector controls (92% versus 42%, respectively; P < 0.001). Whereas these findings do not rule out the possibility that FP-mediated dysregulation of p21WAF1/CIP1 contributes to the increase in cell death after PMA exposure, they do suggest that other factors play a role in the enhanced lethality of the PMA/FP combination.
Effects of PMA and FP on E2F/pRb Binding.
Induction of leukemic cell differentiation by PMA requires CDKI induction, pRb dephosphorylation, and inactivation of E2F transcription factors via binding to the underphosphorylated form of pRB (35). Therefore, the effects of FP coadministration were examined in relation to the interaction between E2F and pRb (Fig. 10). U937 cells were exposed to 10 nm PMA, 100 nm FP, or the combination for 8 h, after which time pRb immunoprecipitates were obtained and treated with deoxycholate to dissociate pRb/E2F complexes. The E2F liberated from pRB after IP was evaluated by EMSA. U937 cells exposed to the combination of FP and PMA or to FP alone displayed a clear reduction in E2F binding to labeled probe compared with PMA-treated cells, indicating that FP diminishes pRb/E2F binding. Furthermore, EMSA analysis of extracts from FP/PMA-treated cells displayed increased levels of a rapidly migrating E2F species, consistent with the preceding observations (data not shown). Together, these findings raise the possibility that FP-mediated perturbations in pRb/E2F binding in PMA-treated cells may contribute to the lethal actions of this drug combination.
The present studies were undertaken to determine whether the CDKI FP could enhance PMA-induced maturation in human leukemia cells. The rationale for this investigation stemmed from several considerations: (a) FP has been shown to induce differentiation in some cell types (e.g., non-small cell lung cancer cells; Ref. 21); and (b) inhibition of cell cycle progression by FP might promote a leukemic cell differentiation program (47). Contrary to expectations, coexposure to FP for 24 h strikingly opposed PMA-induced differentiation in U937 cells and instead significantly increased apoptosis. These events were associated with increased mitochondrial dysfunction, activation of caspases, and loss of clonogenic survival; moreover, enhanced cell death after PMA/FP cotreatment was also observed in promyelocytic leukemia cells (HL-60) and in U937 cells overexpressing the antiapoptotic protein Bcl-2. These events may reflect the complex reciprocal relationship that exists between differentiation and apoptosis. For example, apoptosis can accompany cellular maturation in leukemic cells, generally as a relatively late event (48). In addition, U937 cells displaying dysregulated expression of protein kinase Cζ (49) or impaired p21WAF1/CIP1 induction (8) respond to PMA by undergoing apoptosis rather than maturation. Thus, a plausible explanation for the present findings is that FP induces specific disruptions in PMA-associated G1 arrest and/or differentiation events, which culminate in the engagement of a default apoptotic program.
The ability of FP to block PMA-mediated induction of p21WAF1/CIP1 was unanticipated and may be related to FP-associated perturbations in the growth arrest apparatus. In p53 wild-type cells, p21WAF1/CIP1 plays an integral role in the DNA damage response (50), although PMA-related cell cycle arrest has also been attributed to induction of p21WAF1/CIP1 in human leukemia cells lacking p53 (6). The accumulation of p21WAF1/CIP1 after PMA treatment may stem from both transcriptional and posttranscriptional mechanisms. For example, PMA-induced accumulation of p21WAF1/CIP1 in p53-null human embryonic fibroblasts has been shown to occur primarily at the posttranscriptional level through a marked increase in the half-life of the p21WAF1/CIP1 mRNA (51). Whereas FP may modify PMA-related p21WAF1/CIP1 expression in a similar manner, the early antagonism of p21WAF1/CIP1 induction (i.e., at 2 h; Fig. 6,B) and the results of the reporter assay (Fig. 6 C) also suggest regulation at the transcriptional level. Although the mechanism by which FP antagonizes induction of p21WAF1/CIP1 is unclear, it is tempting to invoke the presence of a feedback loop in which direct CDK inhibition by FP renders induction of one or more CKDIs redundant and therefore prevents increased CDKI expression. The characterization of such a feedback mechanism, presuming it exists, awaits further investigation.
The hypothesis that FP-mediated disruption of PMA-related p21WAF1/CIP1 induction contributes to the antileukemic activity of this drug combination is also plausible. There is accumulating evidence that CDKIs such as p21WAF1/CIP1 exert antiapoptotic actions. For example, the expression of p21WAF1/CIP1 confers resistance to cell death in differentiating myocytes (52) and in glioblastoma cells exposed to cytotoxic drugs (9); moreover, inhibition of Fas-mediated apoptosis in human hepatoma cells has been linked to the association of p21WAF1/CIP1 with procaspase-3 (11). Conversely, disruption of p21WAF1/CIP1 has been shown to lower the apoptotic threshold for both cytotoxic drugs (53) and differentiation-inducing agents such as PMA (8) or vitamin D3 (54). Interestingly, U937 cells stably transfected with a mutant p21WAF1/CIP1 construct lacking the nuclear localization signal have recently been shown to be resistant to various apoptotic stimuli (55). Such findings raise the possibility that translocation of p21WAF1/CIP1 to the cytosol during the course of leukemic cell differentiation may protect these cells from an early apoptotic death. Thus, it is conceivable that FP-mediated dysregulation of p21WAF1/CIP1 induction might be involved in the synergistic increase in apoptosis in PMA-treated cells. However, as shown in Fig. 9, p21WAF1/CIP1 antisense-expressing cells were more sensitive to PMA/FP-mediated apoptosis, at least at some drug concentrations. Because U937/p21AS cells already exhibit an impaired p21WAF1/CIP1 response, it is likely that additional factors contribute to the diminished cell death threshold in cells exposed to PMA in combination with FP. Finally, FP-mediated down-regulation of the antiapoptotic CDKI p27KIP1 (10), in contrast to p21WAF1/CIP1, appeared to result from caspase activation accompanying apoptosis. Nevertheless, it remains possible that this phenomenon could contribute to amplification of the cell death process.
In addition to its effects on CDKI expression in PMA-treated cells, FP treatment also modified the extent and temporal pattern of p42/p44 MAPK (i.e., ERK1/ERK2) activation. The Raf/MEK/ERK signaling pathway plays a critical role in leukemic cell differentiation based on evidence that specific MEK1/MAPK inhibitors such as PD98059 effectively block PMA-mediated maturation (56). Furthermore, it has been suggested that the relative outputs of the MEK/MAPK and the stress-related stress-activated protein kinase/c-Jun NH2-terminal kinase signaling cascades regulate the cell death response to various noxious stimuli such as growth factor deprivation (45). Whereas ERK activation has generally been associated with cytoprotective effects (57), increased activity of ERK has also been shown to promote apoptosis (44). However, whereas FP coadministration increased ERK activation in response to PMA, the inability of the MEK1 inhibitor U0126 to attenuate apoptosis argues against a functional role for ERK in potentiation of cell death.
FP actions may also stem from dysregulation of the pRb/E2F axis because pRb is an integral component of the cell cycle arrest machinery and plays an important role in both differentiation and apoptosis (58). In replicating cells, pRb is successively phosphorylated and thereby inactivated by various CDK/cyclin complexes (19). During leukemic cell maturation, pRb becomes, through the actions of multiple CDKIs and protein phosphatases, dephosphorylated and activated before growth arrest in G1 (59). Dephosphorylation of pRb also occurs in the early stages of apoptosis (60), as does cleavage of the pRb protein by apoptotic caspases (41). Dephosphorylated pRb binds to and inactivates E2F, thereby repressing transcription of multiple genes involved in S-phase progression (61). In view of evidence that inappropriate expression/activity of E2F represents a potent stimulus for apoptosis (62), it is possible that dysregulation of the pRb/E2F axis might promote cell death. In this regard, coincubation of PMA-treated cells with FP resulted in accelerated dephosphorylation of the pRb protein, a phenomenon compatible with the known CDK-inhibitory actions of FP. However, cells exposed to both agents also exhibited early evidence of pRb degradation, and this phenomenon was virtually complete by 24 h. This finding raises the possibility that temporal disruption of the normal pattern of pRb dephosphorylation may increase the susceptibility of this protein to proteolytic cleavage. Another possibility is that FP administration may lead to qualitative changes in pRb phosphorylation status, thereby altering the activity of this protein. Such putative changes, in conjunction with reduced levels of full-length underphosphorylated pRb noted in PMA/FP-treated cells, could account for or contribute to the reduction in E2F/pRb binding that was observed.
In summary, the present results demonstrate that coadministration of the CDKI FP blocks human leukemic cell differentiation in response to PMA and enhances mitochondrial injury (e.g., cytochrome c release), caspase activation, and apoptosis. Furthermore, these events were accompanied by a significant loss in clonogenic survival. In addition, coadministration of FP led to perturbations in the expression/activity of a variety of signal transduction and cell cycle-regulatory proteins in PMA-treated cells including potentiation of ERK phosphorylation, inhibition of p21WAF1/CIP1 induction, degradation of p27KIP1, accelerated dephosphorylation and subsequent cleavage of pRb, and diminished pRb/E2F binding. Whereas some of these events (e.g., proteolytic cleavage of p27KIP1 and pRb) appear to represent a consequence rather than a cause of caspase activation, they may nevertheless comprise part of an amplification loop that ensures continuation of the cell death process. It is noteworthy that the FP concentrations used in this study are readily achievable in the plasma of patients (14) and that the feasibility of administering PMA to patients with leukemia has recently been established (63). If coadministration of these agents in vivo is associated with antileukemic synergism, as observed in vitro, such a finding could potentially have therapeutic implications. Accordingly, further attempts to identify the factors responsible for FP-mediated potentiation of apoptosis in leukemic cells exposed to a variety of maturation-inducing agents are currently underway.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Supported by NIH Grants CA 63573, CA 83705, and CA 77141 and by Leukemia and Lymphoma Society of America Award 6405-97.
The abbreviations used are: CDK, cyclin-dependent kinase; FP, flavopiridol; PMA, phorbol 12-myristate 13-acetate; ERK, extracellular signal-regulated protein kinase; MAPK, mitogen-activated protein kinase; MEK, mitogen-activated protein kinase kinase; ara-C, 1-β-d-arabinofuranosylcytosine; PBS-T, PBS-Tween; CDKI, CDK inhibitor; DiOC6, 3,3-dihexlyoxacarbocyanine; B-d-FMK, BOC-Asp(OMe)-fluoromethyl ketone; PARP, poly(ADP-ribose) polymerase; EMSA, electrophoretic mobility shift assay; IP, immunoprecipitation; pRb, retinoblastoma protein; CI, combination index.
L. Cartee and S. Grant, unpublished observations.