Resveratrol Inhibits Drug-Induced Apoptosis in Human Leukemia Cells by Creating an Intracellular Milieu Nonpermissive for Death Execution

INTRODUCTION

The effector components of apoptotic death signaling and their intricate networking have been unraveled during the past couple of decades (1, 2, 3) . Consequently, it is now well established that depending on the level of activation of the initiator caspase, such as caspase-8, the death signal can recruit directly downstream effector caspases or engage the mitochondria with the resultant release of death amplification factors, such as cytochrome c, apoptosis inducing factor, and Smac/DIABLO (2 , 4 , 5) . Death signaling by anticancer drugs generally relies on positive input from the mitochondria, as is evidenced by the resistance of tumor cells overexpressing the death-inhibitory protein Bcl-2 that is localized to the membranes of mitochondria, endoplasmic reticulum, and nucleus (6, 7, 8) . Therefore, by implication, an intracellular milieu permissive for caspase activation/activity and recruitment of mitochondria-derived amplification factors is critical for efficient apoptotic execution. To that end, we have demonstrated the critical role of cellular redox status in the regulation of death signaling (9, 10, 11, 12) . Whereas an overwhelming accumulation of intracellular reactive oxygen species could create an oxidatively stressed environment leading to necrosis, a slight increase is a stimulus for cellular proliferation (13 , 14) . Pro-oxidant intracellular milieu is a hallmark of many tumor cells and is believed to endow tumor cells with a survival advantage over their normal counterparts (15 , 16) . Furthermore, we have demonstrated that a slightly elevated intracellular concentration of superoxide (O₂⁻) inhibited apoptotic signaling, irrespective of the trigger (17 , 18) . Contrarily, our data and that of others have highlighted the critical role of intracellular H₂O₂ in rendering the cytosolic milieu permissive for efficient apoptotic execution (19, 20, 21) . Thenceforth, we hypothesize that a critical balance between intracellular H₂O₂ and O₂⁻ dictates the response of tumor cells to apoptotic stimuli (9 , 10 , 12) , and any stimulus/signal that inhibits the ability of intracellular H₂O₂, triggered on drug exposure, to reduce the intracellular environment could potentially favor the acquisition of the resistant phenotype.

One area of recent interest in cancer biology is the chemopreventive potential of natural products. Among the compounds being evaluated for their cancer inhibiting activity is a phytoalexin, resveratrol (RSV), found in grapes and wines and known for its diverse biological activities, including antioxidant property (22, 23, 24) . We reported previously that the chemopreventive activity of RSV could be the result of its ability to induce apoptotic death in human leukemia and breast carcinoma cells (25) . However, depending on the cell type and the concentration used, RSV has been shown to induce or inhibit cellular proliferation and death signaling (25, 26, 27, 28, 29) . Our present study was stimulated by a recent report that H₂O₂-induced apoptosis was inhibited in the presence of RSV and the implication that this could be a function of its antioxidant activity (30) . Given our recent findings that drug exposure of human leukemia cells resulted in H₂O₂-dependent apoptosis, we set out to investigate the mechanism by which RSV inhibited apoptotic signaling triggered by exogenous H₂O₂ or by exposure to three anticancer agents, namely, C2, vincristine, or daunorubicin, that induce apoptotic death in cancer cells (21) .

MATERIALS AND METHODS

Determination of Cell Viability and DNA Fragmentation.

Human promyelocytic leukemia (HL60) cell line was purchased from American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 supplemented with 10% fetal bovine serum, 1% l-glutamine, and 1% S-penicillin. In a typical survival assay, HL60 cells (1 × 10⁵ cells/well) plated in 96-well plates were preincubated with RSV (4–8 μm) for 2 h and then treated with 100 μm of H₂O₂, 50 μg/ml of C2, 1.25 μg/ml of vincristine, or 0.2 μg/ml of daunorubicin for 18 h. Cell survival was determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay as described previously or by the β-galactosidase (β-gal) survival assay described below (31) .

Propidium iodide staining was performed for analyzing DNA fragmentation as described previously (25) . Stained cells were analyzed by flow cytometry (Coulter EPICS Elite ESP; Beckman Coulter, Fullerton, CA) with the excitation and emission wavelengths at 488 nm and 610 nm, respectively. At least 10,000 events were analyzed by WinMDI software.

Determination of Caspase-3 and -9 Activity.

Caspase-3 and -9 activity was determined using the Bio-Rad fluorescent assay kit (Hercules, CA) for the two caspases. Cells (1 × 10⁶) were preincubated for 2 h with RSV (4–8 μm) and then incubated for 12 h with 100 μm H₂O₂, 50 μg/ml of C2, 1.25 μg/ml of vincristine, or 0.2 μg/ml of daunorubicin. Fifty μl of 2× reaction buffer with 10 mm DTT and 5 μl of the conjugate substrate (DEVD-AFC for caspase-3 and LEHD-AFC for caspase-9) were added to cell lysates. Caspase activity was determined by the relative fluorescence intensity at 505 nm following excitation at 400 nm with a spectrofluorometer (Tecan Group, Ltd., Maennedorf, Switzerland). Results are shown as fold increase (×-increase) in activity relative to untreated cells (1×). In addition, cleavage of the caspase-3 substrate, poly(ADP ribose) polymerase (PARP), was assessed by Western blot analysis using anti-PARP (clone C-2–10; PharMingen, San Diego, CA) as described previously (9) .

Measurements of Intracellular O₂⁻ and pH.

Intracellular O₂⁻ was assayed by a lucigenin-based chemiluminescence assay as described previously (11) . Data are shown as percentage change (% change) in the intracellular O₂⁻ concentration compared with untreated cells and are the mean ± SD of three independent measurements.

For measurement of cytosolic pH, cells were loaded with 10 μm 2′,7′-bis(2-carboxyethyl)-5,6-carboxyfluorescein (BCECF-AM; Sigma, St. Louis, MO), and the fluorescence ratio of 525:610 nm was used to derive cytosolic pH using a standard pH calibration curve as described previously (11) .

Flow Cytometric Analysis of Intracellular H₂O₂.

Cells were exposed to the apoptotic triggers for 4–12 h, loaded with 5-(and-6)-chloromethyl-2′,7′-dichlorofluorescin diacetate (CM-H₂DCFDA; Molecular Probes, Eugene, OR; 5 μm) (32) at 37°C for 15 min, and analyzed by flow cytometry (Coulter EPICS Elite ESP) using an excitation wavelength of 488 nm as described previously (21) . At least 10,000 events were analyzed.

Determination of Mitochondrial ΔΨ_m and Cytosolic Cytochrome c.

Potential-sensitive probe 3,3′dihexyloxacarbocyanine iodide was used to measure mitochondrial Δψ_m as described previously (31) . Briefly, 1 × 10⁶ cells were incubated with 3,3′dihexyloxacarbocyanine iodide (40 nm) for 15 min at 37°C. As a positive control, cells were incubated separately with an uncoupling agent, carbonyl cyanide m-chlorophenylhydrazone (CCCP) (100 μm). At least 10,000 events were analyzed by flow cytometry with excitation set at 488 nm.

Cytochrome c was detected in cytosolic extracts from 30 × 10⁶ cells by Western blot analysis using anti-cytochrome c (7H8.2C12; PharMingen) as described previously (33) .

Transient Transfection with pIRESRacN17 and β-Gal Survival Assay.

Transient transfections of CEM cells were performed using the SuperFect transfection reagents from Qiagen (Hilden, Germany), and survival of transfected cells was assessed by the β-gal survival assay as described recently (34) . Cell survival was calculated as [(β-gal activity μg⁻¹ protein of transfected cells incubated with the apoptotic trigger)/(β-gal activity μg⁻¹ protein of transfected cells incubated without the trigger)]. β-Gal activity was measured using the Galacto-Star mammalian reporter kit (Applied Biosystems, Foster City, CA). Protein concentration was determined using the Coomasie Plus protein assay reagent from Pierce (Rockford, IL).

Data Analysis.

Data presented are mean ± SD of at least three independent experiments performed in triplicate, unless otherwise indicated. Statistical significance was determined by the Student’s t test.

RESULTS

Low Doses of RSV Inhibit H₂O₂-Induced Apoptosis Upstream of the Mitochondria.

Corroborating earlier findings on the death-inducing activity of H₂O₂ in cancer cells, exposure of HL60 cells to H₂O₂ (100 μm) resulted in a significant decrease in cell survival (Fig. 1A) ⇓ . The cytotoxic activity of H₂O₂ was a function of activation of the apoptotic death pathway as evidenced by the significant increase in caspase activity, cleavage of the caspase-3 substrate PARP, and appearance of the sub-G₁ fraction (Fig. 1, B–D) ⇓ . Interestingly, preincubation of cells with 4–8 μm RSV for 2 h before the addition of H₂O₂ for 18 h resulted in an increase in cell survival (Fig. 1A) ⇓ , significant inhibition of caspase-3 activity and PARP cleavage (Fig. 1, B and C) ⇓ , and inhibition of DNA fragmentation (Fig. 1D) ⇓ . Moreover, preincubation with RSV prevented H₂O₂-induced decrease in mitochondrial Δψm (Fig. 2A) ⇓ and significantly blocked H₂O₂-induced cytosolic translocation of cytochrome c from the mitochondria (Fig. 2B) ⇓ . One possible explanation for the observed inhibitory effect of RSV could be that RSV functioned as an efficient scavenger of H₂O₂. However, addition of RSV and H₂O₂ simultaneously to the culture medium (RSV/H₂O₂) had no effect on H₂O₂-induced cell death (Fig. 2C) ⇓ , whereas previous incubation with RSV for 2 h (RSV + H₂O₂) significantly (P < 0.05) inhibited apoptotic signaling (Fig. 2C) ⇓ . These data indicate that the death-inhibitory effect of RSV is not simply a function of its ability to scavenge H₂O₂ but involves mechanism(s) upstream of the mitochondria or signals that engage the mitochondrial death machinery.

RSV Inhibits H₂O₂-Induced Decreases in Intracellular O₂⁻ and Cytosolic pH.

We have shown that H₂O₂ induces a decrease in O₂⁻ and cytosolic acidification, thereby creating an intracellular milieu permissive for apoptotic execution (18 , 34 , 35) . Therefore, we questioned if the inhibitory effect on H₂O₂-induced apoptosis was caused by the ability of RSV to create an intracellular environment nonconducive for death execution. Preincubation with RSV not only induced a slight increase in intracellular O₂⁻ but also blocked the inhibitory effect of H₂O₂ on intracellular O₂⁻ levels (Fig. 3A) ⇓ . Interestingly, preincubation with RSV before the addition of H₂O₂ consistently resulted in higher intracellular O₂⁻ than RSV alone. In addition, the significant decrease in cytosolic pH (P < 0.01) triggered by H₂O₂ was inhibited by previous incubation with RSV (Fig. 3B) ⇓ .

RSV-Induced Inhibition of Apoptosis Can Be Reverted by Blocking NADPH Oxidase Activation.

Considering the reports on the antioxidant potential of polyphenolic compounds, such as RSV, we were intrigued by our observation that low concentrations of RSV induced an increase, rather than a decrease, in intracellular O₂⁻ concentration (30 , 36, 37, 38) . Therefore, we investigated the effect of blocking NADPH oxidase complex, one of the major sources of intracellular O₂⁻, on RSV-induced increase in O₂⁻. Pharmacologic inhibition of NADPH oxidase with diphenyleneiodonium chloride (DPI) completely blocked RSV-induced increase in intracellular O₂⁻ in HL60 cells (Fig. 3C) ⇓ . Furthermore, incubation of cells for 1 h with DPI before the addition of RSV significantly (P < 0.004) restored the sensitivity of HL60 and CEM leukemia cells to H₂O₂ (Fig. 3D ⇓ ; data not shown). To provide additional evidence that this was a function of NADPH oxidase activation, CEM cells were transfected transiently with a dominant negative mutant of Rac (RacN17), which inhibits NADPH oxidase-dependent increase in intracellular O₂⁻ (34) . Similar to DPI, transient transfection with RacN17 completely neutralized the death-inhibitory activity of RSV and restored the sensitivity of leukemia cells to H₂O₂ (Fig. 3E) ⇓ .

RSV Inhibits Apoptosis Triggered by Anticancer Drugs C2, Vincristine, and Daunorubicin.

In our earlier reports, we demonstrated that exposure of human leukemia and melanoma cells to C2, a purified photoproduct of MC540, triggered mitochondrial generation of H₂O₂ that was responsible for the release of cytochrome c and downstream activation of the caspase cascade (21) . Similar to the data obtained with H₂O₂, preincubation of cells with RSV before the addition of C2 resulted in an increase in cell survival compared with the cells treated with C2 alone (Fig. 4A) ⇓ . Whereas exposure of leukemia cells to C2 resulted in robust increases in caspase-3 and -9 activity, preincubation with RSV significantly inhibited both caspases and the cleavage of the caspase-3 substrate PARP (Fig. 4B) ⇓ . That this inhibition was mediated via blocking mitochondria-dependent death signaling was evidenced by the significant decrease in cytosolic translocation of cytochrome c in cells preincubated with RSV before the addition of C2 (Fig. 4C) ⇓ . In addition, similar to the results obtained with H₂O₂, previous incubation with RSV (RSV + C2) was required for the inhibitory effect of RSV on C2-induced death signaling, whereas simultaneous exposure to RSV and C2 (RSV/C2) had no effect on C2 signaling (data not shown).

We next questioned whether the inhibitory effect of RSV on C2-induced apoptosis also was mediated by its ability to (a) create a pro-oxidant intracellular milieu; and (b) inhibit cytosolic acidification. Results indicate that preincubation of HL60 cells with RSV inhibited H₂O₂ production triggered by exposure to C2 (Fig. 5A) ⇓ . Similar to the results obtained with H₂O₂, preincubation of HL60 cells with RSV for 2 h resulted in an increase in intracellular O₂⁻ concentration, which was even more pronounced on subsequent addition of C2 (Fig. 5B) ⇓ . It should be noted that a slight increase in intracellular O₂⁻ also was observed on exposure of cells to C2 alone; however, unlike RSV, this was accompanied by a surge in intracellular H₂O₂ production. In addition, cytosolic acidification triggered on exposure to C2 was inhibited completely on previous exposure to RSV (Fig. 5C) ⇓ . Furthermore, incubation of cells with DPI before the addition of RSV and C2 resulted in a decrease of intracellular O₂⁻ (data not shown) and restored death signaling in response to C2 (Fig. 5D) ⇓ .

To gain additional insight into the death-inhibitory effect of RSV and its potential clinical implications, we next investigated the effect of RSV on apoptosis induced by two chemotherapeutic agents, vincristine and daunorubicin (39) . Corroborating the results obtained with H₂O₂ and C2, results indicate a dose-dependent (up to 8 μm) inhibitory effect of RSV on vincristine-induced cell death (Fig. 6A) ⇓ together with significant inhibition of caspase-9 and -3 activity (Fig. 6B) ⇓ . That the apoptosis inhibitory effect of RSV was linked to its ability to create a pro-oxidant intracellular milieu was supported additionally by the ability of DPI or transient transfection with RacN17 to revert the sensitivity of leukemia cells in the presence of RSV to vincristine-induced apoptosis (Fig. 6, C and D) ⇓ .

The inhibitory effect of low concentrations of RSV on drug-induced apoptosis was not exclusive to C2 or vincristine. Preincubation of leukemia cells to RSV for 2 h resulted in a significant increase in cell survival (Fig. 7A) ⇓ , inhibition of caspase activity (Fig. 7B) ⇓ , and inhibition of intracellular H₂O₂ production (Fig. 7C) ⇓ induced by exposure to daunorubicin.

Decrease in Intracellular O₂⁻ Overrides the Inhibitory Effect of RSV on Drug-Induced H₂O₂ Production.

Thus far, we have shown that the inhibitory activity of RSV on drug-induced apoptosis was linked to its ability to increase NADPH oxidase-dependent intracellular O₂⁻ production. Interestingly, all of the drugs used in this study resulted in an intracellular increase in H₂O₂ production (Fig. 7C) ⇓ . In this regard, our earlier findings have implicated mitochondria in C2-induced H₂O₂ production because inhibition of the membrane NADPH oxidase complex had no effect on intracellular H₂O₂ increase (21) . Similar to the results reported with C2, preincubation of HL60 cells with DPI did not affect intracellular H₂O₂ production on exposure to vincristine and daunorubicin, thus strongly suggesting mitochondria as the cellular source on drug exposure (data not shown). More importantly, decreasing intracellular O₂⁻ with DPI restored the ability of C2, vincristine, and daunorubicin to trigger intracellular H₂O₂ production even in the presence of RSV (Fig. 7C) ⇓ .

DISCUSSION

RSV Inhibits Apoptosis by Altering Cellular Redox Status.

Taken together, our data provide strong evidence that contrary to its proapoptotic activity at ≥25 μm, low micromolar concentrations (4–8 μm) inhibit apoptotic signaling (25 , 40 , 41) . This divergent signaling by RSV is intriguing and could be explained by our earlier observations that at concentrations of ≥32 μm, apoptosis induced in HL60 cells is mediated by up-regulation of CD95 (Fas/Apo1)-CD95L interaction (25) . This effect on up-regulation of the death receptor ligand is not observed at concentrations of RSV <16 μm, hence the inability to trigger apoptosis in these cells (data not shown). Contrarily, at these concentrations, RSV inhibits apoptotic signaling upstream of the mitochondria, thus blocking the recruitment of mitochondrial-derived amplification factors, such as cytochrome c. We also provide evidence that RSV has a potent effect on the intracellular redox status, a critical determinant of the efficacy of the death signal (9 , 42 , 43) . In that respect, it has been shown previously that maintaining a slightly elevated intracellular O₂⁻ promotes cellular proliferation (14 , 44) and inhibits apoptotic signaling (18 , 42) . A pro-oxidant intracellular milieu is an invariable finding in cancer cells and has been shown to endow cancer cells with a survival advantage over their normal counterparts (15) . Because a decrease in intracellular O₂⁻ and cytosolic pH is critical for efficient death execution, our results suggest that low concentrations of RSV could enable cancer cells to evade death signaling by creating a pro-oxidant intracellular milieu and inhibiting cytosolic acidification.

RSV Induces Increase in O₂⁻ via NADPH Oxidase Activation.

Considering the earlier reported ability of RSV to inhibit mitochondrial complex III-induced reactive oxygen species production, our paradoxical findings provide evidence for a pro-oxidant effect of RSV at concentrations that do not trigger apoptosis (45) . Such pro-oxidant activity of polyphenolics, such as RSV, has been reported recently in different systems (46 , 47) . In one model, reactive oxygen species generation at concentrations of RSV that triggered cell death in human cancer cells was proposed to be responsible for its cytotoxic activity (47) . In addition, our results implicate the membrane NADPH oxidase complex as a potential source of O₂⁻ on incubation with low doses of RSV. Inhibition of the NADPH oxidase complex not only restored death signaling but also resulted in reverting the negative effect of RSV on drug-induced intracellular H₂O₂ production. This fits in well with our hypothesis that a balance between intracellular O₂⁻ and H₂O₂ could be a critical factor in the response of cells to apoptotic triggers, with a tilt toward the former favoring survival and a predominance of the latter facilitating death execution (10) . A careful look at the data shown in Fig. 5B ⇓ reveals that, similar to RSV alone, exposure of cells to the anticancer drug C2 also results in a slight elevation in intracellular O₂⁻. However, it is important to note that C2-induced O₂⁻ is accompanied by a surge in intracellular H₂O₂, which is not observed with low concentrations of RSV. In addition, preincubation with RSV completely blocks C2-induced H₂O₂ production and maintains a significantly elevated intracellular O₂⁻ level. By implication, this suggests that an increase in intracellular O₂⁻ could inhibit downstream H₂O₂ production from the mitochondria, thereby inhibiting the decrease in intracellular pH. It is intriguing as to how an increase in intracellular O₂⁻ blocks mitochondrial-derived H₂O₂. One possibility could be inhibition of upstream caspase activation, as has been shown previously, or other proapoptotic factors, such as death-promoting Bcl-2 family members, required for the engagement of the mitochondrial pathway (48) . Impeding these upstream pathways could account for the inhibition of drug-induced mitochondrial H₂O₂ production and downstream cytochrome c release observed in cells preincubated with RSV. This in turn fails to amplify downstream caspase cascade (caspase-9 and -3) and leads to a substantial decrease in the sensitivity of cells to apoptotic stimuli that require mitochondrial amplification factors. Taken together, these results indicate that the increase in intracellular O₂⁻ on exposure of leukemia cells to RSV was mediated through the activity of NADPH oxidase complex and linked directly or indirectly to the apoptosis-inhibitory activity of RSV. In addition, data presented here not only implicate mitochondrial H₂O₂ production as a critical effector mechanism during drug-induced apoptosis but also demonstrate the ability of an increase in intracellular O₂⁻ to prevent H₂O₂ production and thereby impede the recruitment of the mitochondrial death pathway.

Potential Implications.

These findings could be potentially important in light of the recent interest in the biological activity of flavonoids or flavonoid-like molecules, such as RSV, for their possible use in combination chemotherapy regimens. Although in vitro exposure of tumor cells to RSV at relatively high concentrations results in apoptotic cell death, because of the low bioavailability of RSV, plasma levels as high as 50–100 μm may not be physiologically attainable (49 , 50) . Our data suggest death-inhibitory and/or prosurvival activity of RSV in leukemia cells at doses that may be relevant physiologically (49 , 50) . Thus, the use of RSV in combination with drugs such as C2, vincristine, or daunorubicin could be a dangerous mixture because the slight pro-oxidant effect may provide tumor cells with not only a survival advantage but also impede death signals. This could present an ideal environment for the propagation and proliferation of tumor cells.