Rapid Induction of Apoptosis by Combination of Flavopiridol and Tumor Necrosis Factor (TNF)-α or TNF-related Apoptosis-inducing Ligand in Human Cancer Cell Lines

Flavopiridol (L86-8275; NSC649890) is a potent inhibitor of CDKs² undergoing clinical trials (1, 2, 3). It strongly inhibits CDK1, CDK2, CDK4, and CDK7 and causes cytostatic or cytotoxic effect on various human cancer cell lines (4, 5, 6; reviewed in Ref. 7). It also inhibits various kinases such as protein kinase A and C and epidermal growth factor-receptor tyrosine kinase at the micromolar concentrations (7). It also suppresses broadly the transcription of genes, including cyclin D1 (8, 9), and binds to DNA (10). Even with such extensive studies on the functions of flavopiridol, its exact mechanism of anticancer activity remains poorly understood.

To date, the results from Phase II clinical trials of flavopiridol as a single agent have been reported to be unsatisfactory, suggesting that its combination with other drugs may be desired (1, 2, 3). In this regard, cytotoxic synergy between flavopiridol and various anticancer drugs has been observed in various human carcinoma cells, where the order of drug treatments is important (11, 12, 13, 14). For example, paclitaxel treatment followed by flavopiridol enhanced synergistic apoptosis in gastric and breast cancer cells (14) and underwent a Phase I clinical test (15).

TNF-α and TRAIL belong to the TNF family proteins. TNF-α is limitedly used in combination with IFN-γ and melpharan in cancer therapy (16), and it is still a subject of protein engineering for its improved clinical use (17). Ligation of TNF-α to the receptors leads to the activation of the transcription factor NF-κB, which plays as a survival factor (18, 19, 20). NF-κB is also activated by various cytotoxic drugs (21), and its inhibition enhanced the cytotoxicity of CPT-11 in colon cancer cells (22). TRAIL has been under tests to determine conditions for clinical trials (23). However, not all cancer cells suffer apoptosis when exposed to TRAIL. Increased activity of NF-κB present in many breast cancer cell lines has been observed to confer resistance to TRAIL (24). Consequently, the modulation of NF-κB activity will be important to the clinical use of TRAIL.

In the current study, we investigated the effect of the combination of flavopiridol and TNF-α, or TRAIL, on the human cancer cell proliferation. Surprisingly, we observed a very rapid and potent induction of apoptosis by the combination treatment of these agents in the cancer cell, which involved the activation of caspases and the inhibition of NF-κB transcriptional activity by flavopiridol. Our data suggest that the current combination may deserve further preclinical tests, benefiting the development of cancer therapy.

Flavopiridol, made in house according to the reported structure (7), was dissolved in DMSO as a 10 mm stock solution at −20°C. Human recombinant TNF-α was purchased from Boehringer Mannheim Biochemica (Mannheim, Germany) and stored at −70°C as aliquots (1 mg/ml) in PBS containing 1 mg/ml BSA (Sigma Chemical Co., St. Louis, MO). Recombinant human TRAIL/Apo2L was from Serotec Ltd. (Oxford, United Kingdom). DRB was from Sigma Chemical Co. Antibodies against poly(ADP-ribose) polymerase (mouse, monoclonal), IκBα (rabbit, polyclonal), caspase-8 (rabbit, polyclonal) were purchased from BD PharMingen (San Diego, CA), and those against lamin B (goat, polyclonal) and NF-κB p65 (mouse, monoclonal) were from Santa Cruz Biotechnology (Santa Cruz, CA). Anticaspase-3 antibody (mouse, monoclonal) was from R&D Systems (Minneapolis, MN). Plasmid pNF-κB-Luc containing five repeats of NF-κB enhancer element was from Stratagene (La Jolla, CA). Caspase peptide inhibitors, Z-YVAD(OMe)-fluoromethylketone (FMK), Z-D(OMe)-E(OMe)-VD(OMe)-FMK, Z-IE(OMe)TD(OMe)-FMK, and Z-LE(OMe)HD(OMe)-FMK were from Alexis Biochemicals (San Diego, CA). Horseradish peroxidase-linked secondary antibodies for antimouse antibodies and for antirabbit antibodies were purchased from Amersham Pharmacia Biotech (Uppsala, Sweden) and Bio-Rad (Hercules, CA), respectively. Enhanced chemiluminescence reagent was from Amersham Pharmacia Biotech. LipofectAMINE Plus was from Life Technologies, Inc. (Grand Island, NY). Protease inhibitor mixture (Complete Mini) was from Roche (Mannheim, Germany).

All tumor cell lines, including A549, a human small cell lung carcinoma cell line, HCT-116 and HCT-15, human colon cancer cell lines, and Rat2, a rat embryonic fibroblast cell line, were maintained in RPMI 1640 plus 5% (v/v) heat-inactivated fetal bovine serum. All of the media were supplemented with additional 100 units/ml penicillin G, 100 μg/ml streptomycin, and 0.25 μg/ml amphotericin B. Cells were maintained at 37°C in a humidified incubator containing 5% CO₂.

Cells were plated at a density of 500,000 cells/plate in multiple 60-mm tissue culture dishes in the culture media described above. Cells were inoculated 1 day before the agent treatment. At the indicated time, adherent cells were harvested by trypsinization, combined with nonadherent cells, washed once with PBS, and then resuspended in 300 μl of PBS and fixed with 5 ml of ice-cold 75% ethanol. After fixation for at least 3 h at 4°C, the cells were sedimented by centrifugation and resuspended in PBS containing 1 mg/ml glucose and 1 mg/ml RNase A up to ∼1,000,000 cells/ml and incubated at room temperature for 30 min. After then, 20× propidium iodide solution (1 mg/ml in distilled water) was added to each sample and incubated in the dark for an additional 30 min. The samples were analyzed with flow cytometry (Becton Dickinson FACScan). Analysis was performed after 10,000 counting events. Histograms were analyzed using Modifit software (Veriety Software House, Inc., Topsham, ME).

Cells were harvested by trypsinization in PBS and, after centrifugation, resuspended in 50 mm Tris-HCl (pH 7.5), 20 mm EDTA, and 1% NP40. Sample tubes were kept in ice for 30 min with intermittent gentle agitations. Proteins in the supernatant, after microcentrifugation at 12,000 rpm for 20 min, were removed by phenol extraction, and nucleic acid was collected with ethanol precipitation. The pellets of nucleic acid were treated with RNase A for 30 min before being resolved on a 1% agarose gel electrophoresis. The samples, each equivalent to the same amount of cells, were loaded and visualized with ethidium bromide staining.

A549 cells were plated on at a density of 4,000,000 cells/100-mm tissue culture dish. On the next day, cells were treated for 24 h with TNF-α (10 ng/ml) and/or flavopiridol (500 nm). Then, cells were washed twice with PBS and collected in 1 ml of PBS containing protease inhibitor mixture (1 tablet/50 ml PBS). The cells were broken by sonication in ice, and protein extract was collected by removing cell debris by microcentrifugation as above. Protein concentrations were determined with Bradford dye reagent (Bio-Rad). Equal amounts of protein (40 μg) were resolved on 10 or 15% SDS-PAGE gels and transferred to nitrocellulose membranes. Proteins of interest were detected with their specific antibodies mentioned above. Horseradish peroxidase-linked secondary antibodies were used to visualize proteins.

A549 cells were plated on at a density of 4,000,000 cells/100-mm tissue culture dish. When the cell density was 70–80% confluent, the cells were transfected for 5 h with 9 μg of pNFκB-Luc plasmid DNA in LipofectAMINE Plus according to manufacturer’s instructions. After 1 h of incubation in a fresh media, the transfected cells were collected by trypsinization and distributed equally to wells in 12-well plates. After 16 h, cells were treated in duplicate with/without TNF-α (10 ng/ml) or/and flavopiridol (0.5 μm) for 6 h. The luciferase activity in the same amount from each cell extract was determined using a luciferase assay system (Promega) following the supplier’s instructions. The light intensity was measured with a lumat model LB953 luminometer (Berthold, Germany).

After A549 cells were exposed to TNF-α or flavopiridol, the percentage of each cell cycle phase was quantified with flow cytometry, whereby the sub-G₁ portion represented apoptotic portion. The treatment of 500 nm flavopiridol for 6 h resulted in no significant change in cell cycle distribution (Fig. 1,A, compare flavopiridol with control). TNF-α treatment at 10 ng/ml for 6 h had little effect on cell survival of A549 cells, showing ∼2% apoptosis similar to 1% of the vehicle control (Fig. 1 A, TNF). When the same experiment was performed with flavopiridol for 24 h, ∼4% apoptosis was observed (data not shown), which was similar to 3% obtained in the similar test of the previous report (25). When the cells were exposed to TNF-α for 24 h in the above condition, little apoptosis was observed (data not shown).

We next tested the combination effect of TNF-α and flavopiridol on the proliferation of the A549 cells. Surprisingly, we observed a very strong and rapid increase in cell death, revealed by the presence of a pronounced sub-G₁ fraction in flow cytometry, when the cells were exposed concomitantly to 10 ng/ml TNF-α and 500 nm flavopiridol for 6 h (Fig. 1,A, TNF + flavopiridol). The sub-G₁ fraction increased to 18% in the combination treatment, compared with either flavopiridol or TNF-α treatment alone. When repeated, the above combination treatments usually produced the apoptotic portion of 20 ± 5% in flow cytometry while always showing little cell death fraction in either TNF-α or flavopiridol treatment alone. The extent of cell death by the combination was time and flavopiridol concentration dependent in the range of 100–500 nm flavopiridol (data not shown; see below). The treatment of the higher concentration of TNF-α (50 ng/ml) or flavopiridol (5000 nm) for 6 h failed to induce any apoptosis, although 5000 nm flavopiridol caused a strong G₂-M phase arrest by 2-fold, compared with the control (Fig. 1 A), suggesting that even the higher concentration of each agent could not mimic the cytotoxic effect induced by the above combination treatment.

To verify the current combination effect in another way, we tested chromatin DNA degradation in the cells exposed to the agents. Chromatin DNA was harvested and purified from the A549 cells exposed for 6 h to 500 nm flavopiridol, 10 ng/ml TNF-α, or both and resolved on an agarose electrophoresis gel. Little DNA degradation was detected in the control or 10 ng/ml TNF-α-treated cells (Fig. 1 B). Treatment of 500 nm flavopiridol alone produced a little amount of DNA fragments. In contrast, the presence of huge DNA fragmentation in the cells exposed to the combined agents reflected the enhanced apoptosis among cells. Collectively, the above data suggested that the combination treatment of flavopiridol and TNF-α induced very rapid cytotoxicity in A549 cells.

To see whether the present observation was unique to the A549 cell line, we tested HCT-116 and HCT-15, human colon cancer cell lines (Table 1), and Rat2, a rat fibroblast cell line. The combination treatment of 10 ng/ml TNF-α and 500 nm flavopiridol to HCT-116 for 6 h yielded 30% apoptosis in contrast to 1 and 0% with flavopiridol or TNF-α alone, respectively. HCT-15 cell line that was sensitive to either agent, showing 35% sub-G₁ to TNF-α and 15% to flavopiridol, also showed the enhanced cell death of 91% by the combination treatment. However, in Rat2 cells, neither single nor combination treatment produced apoptosis at all (data not shown).

Distribution of cell cycle phases in A549 cells indicated that the combination treatment of flavopiridol and TNF-α reduced fractions of G₁ and S phases and increased G₂-M phase, compared with the control (Fig. 1,A). In contrast, all these phases in HCT-116 were decreased by the combination (Table 1).

To get insight into the mechanism how the apoptotic process might take place, we tested whether caspase activation was involved in the death process. To this end, various caspase inhibitors such as Z-YVAD-FMK, Z-DEVD-FMK, Z-IETD-FMK, and Z-LEHD-FMK, known as specific inhibitors of caspase-1, caspase-3, caspase-8, and caspase-9, respectively, were used (26, 27). Each inhibitor was added to A549 culture plate at 0, 3, 10, and 30 μm, 30 min before combination treatment of 10 ng/ml TNF-α and 500 nm flavopiridol for 6 h. In the control, the combination produced 21% sub-G₁ apoptotic fraction (Fig. 2,A). Caspase-9 inhibitor exerted only a slight inhibitory activity even at 30 μm without a dose-dependent response. In contrast, caspase-1, caspase-3, and caspase-8 inhibitors significantly reduced the cell death from 21% up to 3, 10, and 8%, respectively, in dose-dependent manners, suggesting that the activation of caspase-1, caspase-3, and caspase-8 might be involved in the death process. To confirm the activation of caspases, we used Western blotting analyses for caspases and their cellular substrates. We failed to detect species of caspase-1 and caspase-9 even with several trials. We observed changes of caspase-3 and caspase-8 and their substrates only in the combination treatment, not in any others (Fig. 2 B). The combination treatment caused procaspase-3 and its substrates, poly(ADP-ribose) polymerase and lamin B, to disappear or to be cleaved, supporting that caspase-3 was activated. It also cleaved the procaspase-8, producing active forms. It is known that caspase-8 cleaves and hence activates procaspase-3. Collectively, activation of caspases, including caspase-1, caspase-3, and caspase-8, might be involved in the apoptosis by the combination of TNF-α and flavopiridol.

Because it has been reported that combinations of cancer drugs give rise to different effects on cell proliferation depending on treatment schedules of agents (12, 13, 14), it was tested whether treatment schedules of TNF-α and flavopiridol might have influence on the cell death. First of all, we performed sequential treatments with one agent for 6 h and then with the other for another 6 h by washing off the former before addition of the latter. We found that none of two sequential treatments using 500 nm flavopiridol and 10 ng/ml TNF-α was able to induce cell death (data not shown). Thus, the results might implicate that both agents should be present at the same time in the culture for the synergistic death.

In the next, one agent was pretreated to the A549 cell culture for 3 h, and then the other was just added to that culture and incubated for another 6 h. Neither 500 nm flavopiridol alone nor 10 ng/ml TNF-α alone produced any apoptotic sub-G₁ fraction after 9 h of treatment (Fig. 3). Pretreatment of 500 nm flavopiridol followed by the addition of 10 ng/ml TNF-α treatment resulted in 24% apoptotic fraction. Surprisingly, pretreatment of 10 ng/ml TNF-α for 3 h followed by the addition of 0.5 μm flavopiridol treatment did not produce any apparent cell death.

We also tested whether treatment of the lower dose of flavopiridol could induce the rapid cell death because achievable plasma concentration of flavopiridol is ∼50–500 nm in clinical tests (1, 2, 3). Pretreatments with 50 and 100 nm of flavopiridol were done for 3 h before addition of 10 ng/ml TNF-α treatment. Pretreatment of 100 nm flavopiridol resulted in a sub-G₁ fraction of 7%, with that of 50 nm producing little (Fig. 3).

Because we observed that TNF-α pretreatment abolished cytotoxic synergy by the flavopiridol and TNF-α combination, we investigated whether the de novo mRNA synthesis might be involved in the inhibition by TNF-α pretreatment. The A549 cells were exposed concomitantly to both TNF-α (10 ng/ml) and DRB (100 μm), a RNA synthesis inhibitor, for 3 h before they were treated with the TNF-α and flavopiridol combination. Preincubation of 100 μm DRB is known to inhibit RNA synthesis in A549 cells by >90% (28). Flavopiridol pretreatment followed by the combination generated 28% sub-G₁ fraction, whereas, as expected, TNF-α pretreatment blocked the combination effect of cell death (Fig. 4). Preincubation of TNF-α with DRB for 3 h, followed by another 6 h incubation in fresh culture medium, did not induce apoptosis. However, when it was followed by 6 h combination treatment of flavopiridol and TNF-α, it produced 27% sub-G₁ fraction, indicating that the inhibitory TNF-α preincubation was no longer effective in the presence of DRB.

NF-κB is an important TNF-α-induced intracellular survival factor (18, 19, 20) and influences cytotoxicity of various cancer drugs (21, 22, 24). We investigated whether NF-κB might play a role in the present synergistic cell death. After transiently transfected with the reporter plasmid, pNF-κB-Luc, the A549 cells were equally divided into fresh culture plates and then followed by a 6-h treatment of the test agents. The reporter plasmid contains NF-κB-responsible elements in the promoter region upstream of luciferase-coding sequence so that the production of luciferase proteins reflects the NF-κB activity. Untransfected cells (A549/mock) had little luciferase activity, whereas the plasmid-containing control cells (p549/NF-κB-Luc) produced some luciferse activity, even without TNF-α treatment, that we considered as a basal activity (Fig. 5,A). TNF-α treatment caused 3-fold increase in luciferase activity, whereas flavopiridol itself did not affect it at all. Upon combination treatment, however, the luciferase activity generated was almost at the basal level. Therefore, combined results propose that flavopiridol can inhibit TNF-α-induced NF-κB-regulated gene transcription. Using Western blotting analysis, we determined the protein levels of NF-κB p65 and IκBα, an NF-κB inhibitor (Fig. 5 B). Each of TNF-α and flavopiridol reduced the level of IκBα to the same extent and increased NF-κB p65, although the latter increased it to the higher level. The combination treatment caused the protein level changes similar to those by TNF-α treatment alone, suggesting that the presence of flavopiridol in the combination had no additional effect on protein changes by TNF alone.

We tested combination of TRAIL and flavopiridol because it also belongs to the TNF family. TRAIL itself was quite cytotoxic to A549 cells, showing 6 and 13% apoptosis for 6 and 9 h, respectively, at 100 ng/ml (Fig. 6). However, the concomitant treatment of TRAIL (100 ng/ml) and flavopiridol (500 nm) rapidly enhanced cell death to 31%. The enhanced death was still obtained in either sequential treatment, showing 21 and 31% sub-G₁ fraction when flavopiridol treatment was followed by TRAIL and vice versa, respectively. Distribution of cell cycle phases in the TRAIL and flavopiridol combination indicated that all of the phases decreased by the combination treatment in contrast to G₂-M enrichment (in Fig. 1 A) by the TNF-α and flavopiridol combination.

We found that the combination of 500 nm flavopiridol and 10 ng/ml TNF-α induced strong cell death in human cancer cell lines. The cell death occurred so fast that even 6-h combination treatment produced 20 ± 5% cell death in A549 cells (Fig. 1) in comparison with 4% produced by 24-h continuous incubation with 500 nm flavopiridol alone (data not shown). Shapiro et al. (25) also observed just 3% apoptosis in A549 cell line with 500 nm flavopiridol in a similar experimental condition. Although the combination treatment in A549 cells caused ∼2-fold, compared with the control, increase in G₂-M phase and decrease in G₁ and S phases (Fig. 1,A), it reduced all of the phases in HCT-116 and HCT-15 cell lines (Table 1). Caspase-1, caspase-3, and caspase-8 were found to take part in the apoptotic process by the combination treatment in the current study. At present, the mechanism of how such caspases were activated only when both agents existed together remains to be elucidated. In this respect, it is to note that flavopiridol and phorbol 12-myristate 13-acetate synergistically induce apoptosis via the intrinsic and extrinsic cell death pathway (29).

The present study provides the first evidence that flavopiridol inhibits NF-κB activation by TNF-α (Fig. 5). Flavopiridol seems not to interfere with basal transcription because its addition did not affect the basal luciferase activity of the control (A549/pNF-κB-Luc). It is known that NF-κB could prevent TNF-α-induced cell death (18, 19, 20). The observation that flavopiridol blocked NF-κB-regulated transcription may propose a possibility that it might tip off the apoptosis regulatory balance toward cell death in the combination with TNF-α. This idea is further supported by the report that NF-κB-blocked A549 cells has been found to be sensitive to TNF-α (30). Currently, we do not know how NF-κB activation by TNF-α was inhibited by flavopiridol. It might be, at least, not because of the decrease of NF-κB p65 or increase of IκBα, an endogenous inhibitor of NF-κB nuclear translocation because the levels of both proteins were indistinguishable between in TNF-α treatment that produced no apoptosis and combination treatment that produced apoptosis (Fig. 5 B). We suggest that flavopiridol might interfere with NF-κB transcriptional activity by other still-unknown mechanisms. In this regard, it has been recently reported that flavopiridol inhibits gene transcription globally (8). Accordingly, it is possible that flavopiridol might dysregulate the turnover of the products of various inducible genes encoding apoptosis regulators, including those regulated by NF-κB, and disturb the balance between survival and apoptotic signals in favor of apoptosis. The idea of flavopiridol inhibition of NF-κB-regulated transcription could be supported by the case of cyclin D1, whose promoter region contains NF-κB binding sites, up-regulated by activated NF-κB (31), and down-regulated by flavopiridol (9).

We tested the effect of treatment schedule on the combination of flavopiridol and TNF-α. Treatment of flavopiridol concomitantly with or followed by TNF-α showed pronounced cell death (Fig. 3). The current result is different from the previous reports where cancer agents such as paclitaxel, doxorubicine, etoposide, cytarabine, topotecan, and mytomycin-C had been shown to have (more) cytotoxic synergy when their treatments were followed by flavopiridol treatment (12, 13, 14). In the current combination, when preexposed to TNF-α, the A549 cells were no longer sensitive to the following combination with flavopiridol. We observed, however, that transcription inhibition by DRB during TNF-α pretreatment kept the cells sensitive to the combination treatment (Fig. 4). We propose, therefore, that induction of antiapoptotic factors, including those by NF-κB, during TNF-α preincubation might probably immunize the cells to the cytotoxic challenge of flavopiridol and TNF-α combination.

Clinical use of TNF is very limited and flavopiridol may be useful only in a combination treatment. Preclinical safety tests of TRAIL are in progress (23). The current results highlight the very strong cytotoxicity of flavopiridol with TNF-α or TRAIL in the human cancer cell lines. Short exposures or lower doses with combination of cancer therapeutics will have great advantages regarding problems of their side effects generated by long and/or high concentration treatment. Although the death induction was dependent on the administration schedules in the case of TNF-α and flavopiridol, it was not affected by the combination sequence of TRAIL and flavopiridol. The combination of flavopiridol and TNF-α or flavopiridol and TRAIL is worthy of continued studies so as to benefit the cancer therapy.