Genistein selectively potentiates arsenic trioxide-induced apoptosis in human leukemia cells via reactive oxygen species generation and activation of reactive oxygen species-inducible protein kinases (p38-MAPK, AMPK)

Abstract

The observation that genistein may behave as a pro-oxidant agent lead us to examine the capacity of this isoflavone to modulate the toxicity of the oxidation-sensitive anti-leukemic agent arsenic trioxide (ATO), and for comparison other anti-tumor drugs. Co-treatment with genistein increased ATO-provoked apoptosis and activated apoptosis regulatory events (Bcl-X_L down-regulation, cytochrome c and Omi/HtrA2 release from mitochondria, XIAP decrease and caspase-8/Bid and caspase-3 activation) in U937 promonocytes and other human leukemia cell lines (HL60, THP-1, Jurkat, RPMI-8866), but not in phytohemagglutinin-stimulated non-tumor peripheral blood lymphocytes (PBLs). Genistein, alone and with ATO, stimulated reactive oxygen species generation, and apoptosis was attenuated by N-acetyl-L-cysteine and butylated hydroxyanisole. Addition of low H₂O₂ concentrations mimicked the capacity of genistein to increase ATO-provoked apoptosis in leukemia cells, but not in PBLs. By contrast, co-treatment with genistein or H₂O₂ failed to potentiate the toxicity of DNA-targeting agent cisplatin, the proteasome inhibitor MG-132 and the histone deacetylase inhibitor MS-275. Within the here used time-period (14 hr) genistein, alone or with ATO, did not significantly affect Akt phosphorylation and NF-κB binding activity, nor decreased intracellular GSH content. However, it elicited N-acetyl-L-cysteine-inhibitable phosphorylation of p38-MAPK and AMPK, and apoptosis was attenuated by pharmacologic inhibitors against these kinases. The pro-oxidant capacity of genistein might be exploited to improve the efficacy of ATO as anti-leukemic agent, and perhaps the efficacy of other oxidation-based therapeutic approaches. © 2008 Wiley-Liss, Inc.

Arsenic trioxide (As₂O₃, ATO, Trisenox™) is a clinically useful drug in the treatment of acute promyelocytic leukemia (APL). At low, physiologically tolerable concentrations (less than 4 μM in plasma), ATO causes APL cytoreduction by inducing terminal differentiation and/or apoptosis.1, 2 Moreover, albeit with lower efficacy, ATO also induces apoptosis in leukemia cell types other than APL, and hence is considered a promissory drug for the treatment of different hematologic malignancies.3 Nonetheless, due to the well known toxicity of arsenicals, the application of new ATO-based therapies may require the generation of sensitizing strategies, to improve the apoptotic efficacy and reduce the drug dosage to clinically tolerable concentrations. In this regard, the modulation of the oxidant state of the cells may represent a useful manner of sensitizing to ATO-provoked apoptosis. In fact, it is known that ATO toxicity is greatly dependent on the intracellular reduced glutathione (GSH) content, and treatments that decrease or increase intracellular GSH potentiates or reduces apoptosis, respectively.4–6 In addition, ATO toxicity is dependent on the inherent intracellular reactive oxygen species (ROS) content,7 and treatments, which experimentally increase ROS production potentiate ATO-provoked apoptosis.8 Another aspect to be considered is the high intrinsic oxidative stress of most cancer cells,9–11 which makes these cells more vulnerable to pro-oxidant treatments than their normal counterparts. For these reasons, agents that generate a moderate pro-oxidant environment may offer potential therapeutic opportunities, and in particular might be used to improve the efficacy of ATO as an anti-leukemic drug.12

Genistein is a soy-derived isoflavone with multiple biochemical effects, including estrogen receptor binding and activation,13 tyrosine kinase inhibition,14 DNA topoisomerase II inhibition,15 down-regulation of PI3K/Akt signaling pathway and NF-κB transcription factor activity,16 alteration of cell cycle-regulatory kinase activities,17 and down-regulation of MRP and Pgp drug transport proteins.18 Earlier reports indicated that genistein induced per se apoptosis,19, 20 enhanced apoptosis induction by chemotherapeutic drugs21, 22 and increased radiosensitivity23, 24 in several tumor cell lines. In these and other studies, the pro-apoptotic action of genistein was generally explained by the capacity of the isoflavone to down-regulate Akt phosphorylation and NF-κB activity.19–24 Nonetheless, another important property of genistein and other flavonoids is that under some conditions they may behave as pro-oxidant agents, as measured by increased ROS generation.25, 26 The physiological consequences and perhaps therapeutical implications of genistein-provoked oxidative stress has been little explored.

For these reason, and on the ground of the above-commented dependence of ATO toxicity on intracellular oxidation, we found of interest to examine the capacity of genistein to sensitize to ATO-provoked apoptosis. The obtained results indicate that co-treatment with genistein selectively potentiates apoptosis induction by ATO, but not by other anti-tumor drugs, in human leukemia cell lines. Under the used experimental conditions, genistein did not significantly affect Akt phosphorylation and NF-κB binding activity, but elicited ROS generation and activated the ROS-inducible kinases p38-MAPK and AMPK.

Ac-DEVD-MCA, N-acetyl-Asp-Glu-Val-Asp-p-nitroanilide; AMPK, AMP-activated kinase; ATO, arsenic trioxide; BHA, butylated hydroxyanisole; BSO, DL-buthionine-(S,R)-sulfoximine; DAPI, 4,6-diamino-2-phenylindole; FITC, fluorescein isothiocyanate; GSH, reduced glutathione; H₂DCFDA, dichlorodihydrofluorescein diacetate; mAb, monoclonal antibody; MAPK, mitogen-activated protein kinase; MTT, thyazolyl blue tetrazolium bromide; NAC, N-acetyl-L-cysteine; pAb, polyclonal antibody; PBLs, peripheral blood lymphocytes; PBS, phosphate-buffered saline; PHA, phytohemagglutinin; PI, propidium iodide; ROS, reactive oxygen species; Z-VAD-fmk, Z-Val-Ala-Asp(OMe)-CH₂F.

Material and methods

Reagents and antibodies

All components for cell culture were obtained from Invitrogen (Carlsbad, CA). Monochlorobimane and dichlorodihydrofluorescein diacetate (H₂DCFDA) were obtained from Molecular Probes (Eugene, OR). 4,6-Diamino-2-phenylindole (DAPI) was obtained from Serva (Heidelberg, Germany). Recombinant TNFα was obtained from Strathmann Biotech AG (Hamburg, Germany). The kinase inhibitors SB203580, SP600125 and compound C, the proteasome inhibitor MG-132, the histone deacetylase inhibitor MS-275, the caspase-3 substrate N-acetyl-Asp-Glu-Val-Asp-p-nitroanilide (Ac-DEVD-pNA), and the caspase inhibitor Z-Val-Ala-Asp(OMe)-CH₂F (Z-VAD-fmk), were obtained from Calbiochem (Darmstad, Germany). Rabbit polyclonal antibodies (pAbs) against human p38-MAPK, phospho-p38-MAPK (Thr¹⁸⁰/Tyr¹⁸²), SAPK/JNK, phospho-SAPK/JNK (Thr¹⁸³/Tyr¹⁸⁵), Akt, phospho-Akt (Ser⁴⁷³), phospho-AMPK (Thr¹⁷²) (40H9) and P-Acetil-CoA carboxylase (Ser⁷⁹), and mouse anti-caspase-8 monoclonal antibody (mAb) (1C₁₂), were obtained from Cell Signaling Technology (Danvers, MA). Mouse anti-pigeon cytochrome c mAb clone 7H₈·2C₁₂ was obtained from BD PharMingen (San Diego, CA). Goat anti-human Bid (C-20), rabbit anti-human Bcl-X_S/L (s-18), rabbit anti-human NF-κB p65 (sc-109) and rabbit anti-human pRb (C-15) pAbs were from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse anti-XIAP (clone 2F1) mAb was obtained from MBL International (Woburn, MA). Peroxidase-conjugated immunoglobulin G antibodies were obtained from DAKO Diagnostics, S.A. (Barcelona, Spain). All other reagents were from Sigma (Madrid, Spain).

Cells and treatments

The human leukemia cell lines U937 and THP-1 (promonocytic), HL60 (myelomonocytic), Jurkat (T lymphoblastoid) and RPMI 8866 (B lymphoblastoid), were routinely grown in RPMI 1640 supplemented with 10% (v/v) heat-inactivated calf serum, 0.2% sodium bicarbonate and antibiotics in a humidified 5% CO₂ atmosphere at 37°C. For experiments, the cell number was adjusted at 2 × 10⁵ cells/ml before initiation of the treatments. Human peripheral blood lymphocytes (PBLs) were isolated from buffy coats from healthy donors over a Lymphoprep (Nycomed, Norway) gradient according to standard procedures. The lymphocytes were resuspended at 0.5 × 10⁶ – 1 × 10⁶ cells/ml in RPMI 1640 plus 10% (v/v) heat-inactivated calf serum, and stimulated to proliferate for 48 hr with 1 μg/ml phytohemagglutinin (PHA), before initiation of the treatments.

Stock solutions of genistein and genistin (50 mM), SB203580 and SP600125 (20 mM), Z-VAD-fmk (25 mM), compound C (10 mM), MS-275 (20 mM), monochlorobimane (200 mM), and N-acetyl-L-cysteine (NAC, 3 M), were prepared in dimethyl sulfoxide; stock solutions of cisplatin (3.3 mM) and TNFα (100 μg/ml) were prepared in distilled water; and stock solutions of H₂DCFDA (5 mM) and butylated hydroxyanisole (BHA, 0.5 M) were prepared in ethanol. All these solutions were stored at −20°C. Stock solutions of DAPI (10 μg/ml) and propidium iodide (PI, 1 mg/ml) were prepared in phosphate-buffered saline (PBS); and a stock solution of ATO (100 mM) was prepared in distilled water. These solutions were stored at 4°C. DL-buthionine-(S,R)-sufoximine (50 mM) was prepared in distilled water just before application.

Flow cytometry

The analysis of samples was carried out using an EPICS XL flow cytometer (Coulter, Hialeah, FL) equipped with an air-cooled argon laser tuned to 488 nm. The specific fluorescence signal corresponding to H₂DCFDA was collected with a 525-nm band pass filter, and the signal corresponding to PI with a 620 nm band pass filter.

Measurement of cell growth and viability, apoptosis and caspase-3 activity

The relative number of viable cells in the culture was examined at 24 hr of treatment using the MTT colorimetric assay, essentially as described by Mitsiades et al.27 With this aim, the cells were centrifuged, resuspended and maintained for 4 hr at 37°C in PBS containing 1 mg/ml MTT, after which a mixture of isopropanol and 1 N HCl (23:2, vol/vol) was added under vigorous pippeting to dissolve the formazan crystals. Dye absorbance was measured by spectrometry at 595 nm. All measurements were repeated at least in quadruplicate wells. Under the used conditions, a decrease in absorbance may reflect inhibition of cell proliferation, increase in cell death or the sum of both factors.

Distinctive characteristics of apoptotic cells were chromatin condensation/fragmentation, as determined by cell permeabilization followed by DAPI staining and microscopy examination; and reduction in DNA content (sub-G₁), as determined by cell permeabilization followed by PI staining and flow cytometry examination. This later method was also used to determine cell cycle phase distribution. As a routine, we also examined free Trypan blue or PI penetration into non-permeabilized cells, as an indication of loss of plasma membrane integrity (primary or secondary necrosis). Caspase-3 activity was determined in in vitro assays, using as substrate Ac-DEVD-pNA. A detailed description of all these procedures was presented in preceding publications28, 29 and hence is omitted here.

Measurement of intracellular ROS accumulation and GSH content

Intracellular ROS accumulation was measured by flow cytometry, using the fluorescent probe H₂DCFDA. The intracellular GSH content was determined by fluorometry after cell loading with monochlorobimane. All these procedures were described in detail in the preceding works.28

Immunoblot assays

To obtain total cellular protein extracts, cells were collected by centrifugation, washed with PBS and lysed by 5 min heating at 100°C followed by sonication in Laemmli's buffer containing a protease inhibitor cocktail, 10 mM NaF and 1 mM sodium orthovanadate. Nuclear extracts (aimed at determining p65 NF-κB translocation) were obtained using the method of Schreiber et al.30 To obtain cytosolic extracts (aimed at determining cytochrome c and Omi/Htra2 release from mitochondria), cells were collected for centrifugation, resuspended in 100 μl of ice-cold PBS containing 80 mM KCl, 250 mM sucrose and 200 μg/ml digitonin, and kept on ice for 5 min. After centrifugation (10,000g for 15 min at 4°C) the pellet was discarded. Fractions of the total, nuclear or cytosolic extracts, containing equal protein amounts, were analyzed by SDS-polyacrylamide gel electrophoresis, blotted onto membranes and immunodetected, as previously described.31

Electrophoretic mobility gel shift assays

Nuclear extracts were obtained as described by Schreiber et al.30 A double-strand oligonucleotide containing the consensus binding site for NF-κB (5′-AGTTGAGGGGACTTTCCCAGGC-3′) was prepared. The conditions of oligoprobe radioactive labeling, binding reaction and electrophoretic separation, were exactly as described by López-Rodríguez et al.32 For competition experiments, nuclear extracts were pre-incubated with 50-fold molar excess of unlabeled oligonucleotide for 30 min at 4°C before adding the labeled probe.

Statistical analysis

As a routine, the significance of differences between experimental conditions was examined, using the Student's t-test, and when positive indicated by asterisks (^#p < 0.05; *p < 0.01; **p < 0.001).

Results

Apoptosis induction by ATO and genistein in U937 leukemia cells

Figures 1a and 1b shows alterations in cell growth and viability, as determined by the MTT assay and apoptosis induction, as determined by chromatin condensation/fragmentation, in U937 promonocytic leukemia cells treated for 24 hr with ATO and genistein, alone and in combination. ATO was used at 1–4 μM, which is into or close to the range of clinically achievable concentrations.1 Genistein was assayed at 10–50 μM, according to the concentrations currently used with other tumor cell types.19–22 Treatments with 50 μM genistein alone or 1–2 μM ATO alone had negligible or very low toxicity, as measured by the decrease in viability and/or increase in the frequency of apoptotic cells. However, when used together, genistein and ATO co-operated to induce apoptosis in more than additive manner, with maximum effect at 2 μM ATO plus 50 μM genistein (Fig. 1b). Using this combination, significant apoptosis started to be detected at 8–14 hr of treatment (Fig. 1c). For these reasons, the combination of 2 μM ATO plus 50 μM genistein was normally adopted for further determinations of apoptosis, and 14 hr was the maximum time treatment used to analyze regulatory events. Allowing for quantitative differences inherent to the sensitivity of the used techniques, the cooperation between ATO and genistein in inducing apoptosis was corroborated by measuring the accumulation of cells with sub-G₁ DNA content, as indicated by PI staining and flow cytometry (Fig. 1d). This procedure also indicated that genistein provoked a dose-dependent and time-dependent accumulation of cells at the G₂/M phase of the growth cycle, which is a commonly described effect of this drug.33, 34 Under all conditions, the frequency of Trypan-blue permeable cells, indicative plasma membrane damage (primary or secondary necrosis) was lower than 7%. Of note, apoptosis potentiation required the simultaneous presence of both genistein and ATO, since it was not observed when sequential treatments were used—namely, 24 hr genistein treatment, washing and 24 hr ATO treatment, or vice versa (results not shown).

Figure 1. Growth inhibition and apoptosis induction by ATO and genistein in U937 cells. (a, b) Relative number of viable cells, as determined by the MTT assay (a), and frequency of apoptotic cells, as indicated by chromatin condensation/fragmentation (b), in untreated U937 cell cultures (Cont) and cultures treated for 24 hr with the indicated concentrations of ATO and genistein, alone and in combination. In the case of the MTT assay, the absorption values are represented in relation to the control, which received the arbitrary value of 100. (c) Frequency of apoptotic cells in cultures treated for the indicated time periods with 2 μM ATO and 50 μM genistein, alone and in combination. When indicated, the caspase inhibitor Z-VAD-fmk (50 μM) was applied 30 min before ATO plus genistein. (d) Cell cycle distribution and frequency of apoptotic cells (Ap), as indicated by sub-G₁ DNA content, at 24 hr of treatment with 2 μM ATO and 50 μM genistein, alone and in combination. The results in (a–c) represent the mean ± SD of at least 3 determinations. Symbols indicate significant differences in relation to the control (a, c) or between the indicated pair of values (b). The results in (d) are representative of 1 of the 2 determinations, with similar results.

Examination of some apoptosis-regulatory factors indicated that ATO and genistein cooperated in inducing XIAP and Bcl-X_L down-regulation, Bid truncation/activation (as revealed by the decrease in the amount of the 23-kDa Bid pro-form) (Fig. 2a), release of mitochondrial proteins (cytochrome c and Omi/Htra2) to the cytosol (Fig. 2b), cleavage/activation of caspase-8 (as demonstrated by the decrease in the level of pro-caspase-8 (57 kDa) and appearance of the 43/41 and 18 kDa cleavage-derived fragments) (Fig. 2c), and potentiation of caspase-3 activity (Fig. 2d). Moreover, the frequency of apoptosis in genistein plus ATO-treated cells was greatly reduced by the caspase inhibitor Z-VAD-fmk (see Fig. 1c). All these results prove that the detected cell death is in fact caspase-dependent apoptosis.

Figure 2. Activation of apoptosis regulatory events. U937 cell cultures were left untreated (Cont) or treated for 14 hr (a–c) or for 14 and 24 hr (d) with 2 μM ATO and 50 μM genistein, alone and in combination. (a) Relative cellular levels of Bcl-X_L, Bid (21 kDa pro-form), and XIAP, as determined by immunoblot using total cellular extracts. (b) Release of cytochrome c and Omi/HtrA2 to the cytosol, as determined by immunoblot using cytosolic protein extracts. (c) Pro-caspase-8 cleavage, as revealed by the appearance of the 43/41 and 18 kDa apoptosis-related fragments. In (a) and (b), the level of α-tubulin was also measured as a loading control. (d) Caspase-3 activity at the indicated times of treatment, as determined in in vitro assays using Ac-DEVD-pNA as substrate. The values are represented as fold increase in relation untreated cells, which received the arbitrary value of one. Symbols indicate significant differences in relation to untreated cells. All other conditions were as in Figure 1.

Effects of other anti-tumor drugs and in other leukemia cells

The effect of ATO and genistein was also examined in other human leukemia cell lines, namely HL60 and THP-1 myeloid cells, and Jurkat and RPMI 8866 lymphoblastic cells. Allowing for some intrinsic differences in ATO sensitivity, genistein potentiated apoptosis induction in all leukemia cells (Fig. 3a). As a non-tumor cell model, we used PBLs (obtained from 6 different healthy donors) which were stimulated to proliferate for 48 hr with PHA. Cell growth activation was assessed by the MTT reduction assay (arbitrary absorption values per 10⁶ cells at 48 hr, 0.10 ± 3.2 and 0.37 ± 6.1 in non-stimulated and stimulated cultures, respectively) and by cell cycle distribution (cells in S plus G₂/M at 48 hr, <3% and 25–30% in non-stimulated and stimulated cultures, respectively). The rate of spontaneous apoptosis in PHA-stimulated PBLs varied considerably in individual assays, and the toxicity of ATO, although generally lower than in the leukemia cell lines, was also somewhat variable. Whatever the case, ATO toxicity was never significantly potentiated by co-treatment with genistein (see Fig. 3b as 2 representative examples).

Figure 3. Apoptosis induction by ATO and genistein in different cell types. (a) Frequency of apoptotic cells, as determined by chromatin fragmentation, in THP-1 and HL60 human myeloid leukemia cells, and Jurkat and RPMI 8866 human lymphoblastoid cells, at 24 hr of treatment with 50 μM genistein and the indicated concentrations of ATO, alone and combination with genistein. Symbols indicate significant differences between the indicated pair of values. (b) Frequency of apoptosis in PHA-stimulated PBLs, as determined by chromatin fragmentation (upper histogram) and sub-G₁ DNA content (lower flow cytometry profiles). After isolation, the PBLs were initially stimulated for 48 hr with PHA, and then for 24 hr more in the absence of drugs (Cont), or in the presence of 50 μM genistein and the indicated concentrations of ATO, alone and with genistein. The histogram and flow cytometry profiles are representative of single determinations with PBLs obtained from different donors. All other conditions were as in Figure 1.

In addition, we also examined the capacity of genistein to modulate the apoptotic action of drugs other than ATO, with proved or potential anti-tumor activity. This included the DNA-targeting agent cisplatin (2 and 5 μM), the proteasome inhibitor MG-132 (0.1 and 0.2 μM),35 and the histone deacetylase inhibitor MS-275 (4 and 10 μM).36 For homogeneity with ATO, the cells were treated for 24 hr with low drug concentrations (Fig. 4a), which were selected from preliminary determinations (results not shown). Nonetheless, a complementary assay was carried out using a high cisplatin concentration (75 μM, Fig. 4b) for shorter treatment periods (3–8 hr). It was observed that co-treatment with genistein did not increase and even reduced the apoptotic action of these drugs in U937 cells, as measured by chromatin fragmentation (Figs. 4a and 4b) and sub-G₁ DNA content (Fig. 4b, and results not shown). Of note, the reduction in apoptosis was not due to a switch to the necrotic phenotype (which might reflect an exacerbation of drug toxicity), since the fraction of Trypan blue- and PI-permeable cells at 24 hr of treatment remained below 7% under all assayed conditions (result not shown).

Figure 4. Apoptosis induction by genistein and different anti-tumor drugs. (a) Frequency of apoptotic cells, as determined by chromatin fragmentation, in U937 cells treated for 24 hr with 50 μM genistein alone, and with the indicated concentrations of cisplatin (CDDP), MS-275, and MG-132, alone and in combination with genistein. (b) Frequency of apoptotic cells, as indicated by chromatin fragmentation at the indicated times of treatment (left panel), and by sub-G₁ content at 8 hr of treatment (right panel), in cultures treated with 75 μM cisplatin, alone and in combination with 50 μM genistein. Symbols indicate significant differences between the indicated pair of values. All other conditions were as in Figure 1.

Oxidative stress

Then, we asked whether genistein caused intracellular ROS accumulation, and whether this effect could explain the increase in ATO toxicity. With this aim, determinations were carried out using the ROS-sensitive fluorescent probe H₂DCFDA at different times of treatment with ATO and genistein, alone and in combination. As indicated in Figure 5a, 2 μM ATO had little effect on intracellular ROS content in U937 cells, which is consistent with earlier observations in this cell line.37 By contrast genistein, alone or with ATO, clearly increased ROS over-accumulation, with maximum intensity at 5 hr of treatment, to slightly decrease thereafter. Genistein induced ROS over-accumulation in a concentration-dependent manner (Fig. 5b), which parallels the capacity of the isoflavone to potentiate ATO toxicity (see Fig. 1b). Moreover, we observed that genistin, the 7-glycoside analog of genistein, had lower capacity than genistein to stimulate ROS production (Fig. 5c, left panel) and correspondingly lower efficacy to potentiate ATO-provoked apoptosis (Fig. 5c, right panel). Taken together, these observations suggest a tight relationship between the oxidant action of the flavonoid and its capacity to potentiate ATO toxicity.

Figure 5. ROS generation vs. apoptosis induction by ATO, genistein and genistin in U937 cells. (a, b) Intracellular ROS accumulation, as indicated by H₂DCFDA-derived fluorescence, in (a) cells treated for the indicated time-periods with 2 μM ATO and 50 μM genistein, alone and in combination; and (b) cells treated for 6 hr with the indicated concentrations of genistein, alone and in combination with ATO. The results are represented in relation to the value in drug-untreated cells, which received the arbitrary value of one. (c) ROS generation (left histogram) and frequency of apoptotic cells (right histogram) in cell cultures treated with the indicated concentrations of genistin, alone and with ATO. Genistein is included for comparison. (d) ROS generation (left histogram) and frequency of apoptotic cells (right histogram) in cells treated with ATO plus genistein, alone or in combination with 10 mM NAC and 50 μM BHA. Symbols indicate significant differences in relation to the control or between the indicated pair of values. Apoptosis was always determined at 24 hr. NAC and BHA was applied 2 hr before the other drugs. Except when otherwise indicated, ATO was used at 2 μM and genistein at 50 μM. All other conditions were as in Figure 1.

To corroborate the importance of genistein-provoked oxidative stress for potentiation of ATO toxicity, experiments were carried out using the anti-oxidant agents NAC (10 mM) and BHA (50 μM, the maximum concentration which was not per se toxic). BHA is a synthetic phenol with ROS-scavenging activity.38 NAC is a thiol-containing agent with ROS-scavenging and direct reducing activities, which also functions as a cysteine donor for GSH synthesis.39 The results in Figure 5d indicate that BHA caused a slight decrease both in ROS generation and apoptosis induction. NAC also attenuated apoptosis induction by genistein plus ATO, although it failed to decrease ROS accumulation. Of note, the possibility of direct interaction between ATO and NAC was already excluded in preceding studies.7, 40

As an indirect complementary proof correlating ROS generation and ATO toxicity, we examined whether the action of genistein could be mimicked by addition of exogenous H₂O₂, frequently used as a paradigmatic oxidant treatment. It was observed thatco-treatment with 20 and 40 μM H₂O₂, which was per se slightly toxic, clearly provoked more than additive increases in ATO lethality, as measured by chromatin fragmentation (Fig. 6a) and sub-G₁ DNA content (results not shown). For comparison, we also analyzed the effect of H₂O₂ on the toxicity of drugs other than ATO in U937 cells, as well as ROS production by genistein and the effect of exogenous H₂O₂ on ATO toxicity in Jurkat leukemia cells and non-tumor PBLs. As indicated in Figure 6b, H₂O₂ did not increase apoptosis, or produced less than additive effects, when used in combination with cisplatin and MS-275, which is congruent with the inability of genistein to potentiate the toxicity of these agents (see Figs. 4a and 4b). In addition, it was observed that genistein induced ROS over-accumulation in both Jurkat cells and PBLs (Fig. 6c). Nonetheless, H₂O₂ only potentiated ATO-provoked apoptosis in Jurkat cells, but not in PBLs (Fig. 6d), which parallels the results earlier obtained with genistein (see Fig. 3). These results corroborate that ATO sensitivity to the oxidant environment is a drug- and cell type-specific response.

Figure 6. Modulation of apoptosis induction by exogenous H₂O₂ (a, b). Frequency of apoptotic cells in U937 cell cultures treated for 24 hr with the indicated concentrations of H₂O₂, alone and with ATO (a) or with the indicated concentrations of cisplatin and MS-275 (B). (c) Intracellular ROS accumulation in Jurkat cells and PHA-stimulated PBLs treated for 6 hr with genistein, alone and in combination with ATO. The results are represented in relation to the corresponding control, which received the arbitrary value of one. (d) Frequency of apoptotic cells, as determined by chromatin fragmentation, in Jurkat cell cultures and PHA-stimulated PBLs treated for 24 hr with 20 μM H₂O₂, alone and with ATO. The results obtained with PBLs are representative of 1 of the 3 different determinations. Genistein was always used at 50 μM, and ATO at 4 μM in Jurkat cells and 2 μM in U937 cells and PBLs. Symbols in (c) indicate significant differences in relation to the control. Symbols in (a), (b) and (d) indicate that the toxicity of the combined treatment (H₂O₂ plus anti-tumor drug) is significantly higher than the sum of the toxicities of the two agents alone. All other conditions were as in Figures 1 and 3–5.

Finally, we wanted to analyze whether the changes in ROS accumulation were accompanied by alterations in intracellular GSH, since as indicated above GSH is a strict determinant of ATO toxicity. Some of the obtained results are indicated in Figure 7. While genistein did not cause detectable GSH depletion, treatment for 5 and 14 hr with ATO alone or ATO plus genistein slightly increased GSH content. Intracellular GSH was also augmented by treatment with NAC, either alone or in combination with ATO plus genistein, but the increase was approximately equivalent to that produced ATO plus genistein alone. Treatments shorter than 5 hr did not produce significant effects (result not shown). As expected, treatment with 1 mM BSO, a specific γ-glutamylcysteine synthetase inhibitor here included as a control, decreased GSH content (Fig. 7) and correspondingly increased ATO toxicity (result not shown).

Figure 7. Modulation of GSH levels. Relative GSH levels, as determined by monochlorobimane derivatization, in U937 cells treated for the indicated time-periods with 2 μM ATO, 50 μM genistein and 10 mM NAC, alone and in combination, and with 1 mM BSO. All results are represented in relation to untreated cells (approximate GSH content, 9.5 nmol/10⁶ cells), which received the arbitrary value of one. Symbols indicate significant differences in relation to untreated cells. All other conditions were as in Figure 1.

Cell signaling

In earlier reports, the potentiation of anti-tumor drug toxicity by genistein was explained by the capacity of the isoflavone to down-regulate Akt phosphorylation and NF-κB activity. For this reason, we wanted to examine Akt phosphorylation/activation and NF-κB translocation and binding in U937 cells treated with genistein alone, ATO alone and the combination of both agents. As indicated in Figure 8a, treatment for 14 hr did not affect the basal level of phosphorylated Akt. Of note, this is consistent with the inability of genistein to decrease intracellular GSH (see Fig. 7), since we earlier demonstrated that PI3K/Akt inhibition causes GSH depletion in leukemia cells.31 In the same manner, treatments for 8 hr (result not shown) and 14 hr (Fig. 8b) failed affect p65-NF-κB translocation to the nucleus (upper blot) and NF-κB binding to its DNA consensus sequence (lower blot).

Figure 8. Cell signaling. (a) Relative levels of total and phosphorylated Akt (Akt-T and Akt-P, respectively), as determined by immunoblot using total cellular extracts obtained from untreated U937 cells (Cont) and cells treated with genistein and ATO, alone and in combination. Extracts from NK₂ epithelial kidney tubular cells subjected to 6 hr hypoxia plus 1 hr re-oxygenation were used as a positive control for Akt phosphorylation. (b) The upper blot shows the relative nuclear levels of p65 NF-kB and retinoblastome (pRb, used as a loading control), as determined by immunoblot using nuclear extracts. The lower blot shows NF-κB binding activity, as determined by gel shift assays. Fifty-fold excess unlabeled oligoprobe was used as competitor (+Comp). In both cases, extracts from cells treated with 50 ng/ml TNFα were included as a positive control. (c) Relative levels of total (T) and phosphorylated (P) p38-MAPK and AMPK, and phosphorylated Acetyl-CoA carboxylase (ACC-P), in untreated (Cont) U937 cells and cells treated with ATO and genistein, alone and in combination, and in the absence or presence of 10 mM NAC. (d) Frequency of apoptotic cells, as determined by chromatin fragmentation, in U937 cell cultures treated with ATO, genistein, 10 μM SB203580 (SB), 10 μM SP600125 (SP), and 10 μM compound C (CC) and with ATO plus genistein in the absence (−) or the presence of SB203580, SP600125 and compound C. (e) Frequency of apoptosis and cell cycle distribution, as determined by flow cytometry, using a similar treatment as in (d). Symbols indicate significant differences between the indicated pairs of values. NAC was added 2 hr before, and the kinase inhibitors 30 min before ATO plus genistein. ATO was used at 2 μM and genistein at 50 μM. Determinations in (a–c) were carried out at 14 hr of treatment, and in (d) and (e) at 24 hr of treatment. All other conditions were as in Figure 1.

JNK and p38 are oxidation-inducible MAPK kinases, which regulate apoptosis induction by ATO in leukemia cells.41, 42 In addition, AMPK was recently characterized as a ROS-inducible kinase, which mediates genistein effects in some cell models.43, 44 For these reasons, experiments were carried out to determine the effect of genistein on the phosphorylation/activation of these kinases, and their possible role on genistein plus ATO-induced apoptosis. Treatment of U937 cells with ATO and genistein, alone and in combination, did not significantly affect JNK phosphorylation (result not shown). On the other hand, treatment with ATO alone and genistein alone stimulated the phosphorylation of p38-MAPK, AMPK and the AMPK-downstream kinase ACC, a response which in the case of p38-MAPK was further increased by the combination of both agents, and which was decreased by the anti-oxidant agent NAC (Fig. 8c).

To determine the role of MAPKs and AMPK in apoptosis induction, experiments were carried out using the JNK inhibitor SP600125 (10 μM), the p38 inhibitor SB203580 (10 μM), and the AMPK inhibitor compound C (10 μM). The concentrations of SP600125 and SB203580 were selected from our earlier studies, which proved their efficacy to prevent kinase activation without causing significant cell death.40 The concentration of compound C was selected as the maximum non-toxic dose in preliminary assays (result not shown). The results in Figures 8d and 8e indicate that the p38-MAPK and AMPK inhibitors reduced apoptosis induction by genistein plus ATO, while the JNK inhibitor was ineffective. Of note, the flow cytometry determinations showed that SB203580 did not affect G₂/M blockade by genistein plus ATO (Fig. 8e), suggesting that cell cycle blockade and apoptosis induction are independently regulated events.

Discussion

The results in this work indicate that co-treatment with genistein potentiates apoptosis induction by the anti-leukemic drug ATO in U937 promonocytes and other human myeloid and lymphoblastoid leukemic cell lines. By contrast, no potentiation was obtained using normal PBLs, which suggests that the response was somewhat selective for tumor cells. Although the study of the apoptotic machinery was beyond the scope of this work, our preliminary determinations indicated that ATO plus genistein activated critical regulatory elements of the mitochondrial executioner pathway, such as Bcl-X_L and XIAP down-regulation, and cytochrome c and Omi/HtrA2 release.45 The presence of cytochrome c in the cytosol is essential for apoptosome assembly and activation, as a first step for the activation of the caspase cascade, while released Omi/HtrA2 may bind and degrade XIAP, relieving caspases from the inhibitory action of this protein. While the activation of the mitochondrial pathway is consistent with the property of both ATO and genistein as mitochondria-targeting drugs,12, 26, 46 our results indicated that ATO plus genistein also activated the caspase-8/Bid pathway. Whether caspase-8 activation is only a secondary effect mediated by mitochondrial activation or involves a true death receptor activation remains to be determined, since both types of response were earlier characterized using ATO under conditions of oxidative stress.31, 41

The potentiation of ATO-provoked apoptosis behaved as a drug selective effect, since co-treatment with genistein failed to increase apoptosis induction by MG-132, MS-275 and cisplatin. This later result might be surprising, since genistein was earlier reported to enhance the toxicity of cisplatin in other tumor cell types.22, 47 Nonetheless, the inability of genistein to increase apoptosis induction by drugs other than ATO in U937 cells seems to be consistent with its failure to inhibit Akt phosphorylation and NF-κB activation, since Akt and NF-κB are protective factors against multiple stressing treatments. Of note, our results only indicate that Akt and NF-κB do not play a main regulatory role under the present experimental conditions, but do not prejudge the capacity of genistein to modulate Akt/NF-κB in leukemia cells in other conditions, nor necessarily contradicts earlier observations in other cell types. In fact, in the present work the effect of genistein was examined at 14 hr which, although sufficient to trigger apoptosis in combination with ATO (see Fig. 1c), may still be insufficient to cause significant down-regulation of basal Akt phosphorylation and NF-κB binding activity. Actually, the 14 hr treatment period is shorter than the genistein treatments generally used by other authors to down-regulate the steady-state activity of NF-κB, or prevent its activation by stressing agents.20–22, 24, 48 Moreover, while in other works genistein was used in combination with stressing treatments that activated NF-κB,21, 22, 24 ATO does not stimulate NF-κB in leukemia cells, as observed in the present experiments and also reported in preceding studies.42, 49

By contrast, the pro-oxidant action of genistein may adequately explain its capacity to potentiate ATO toxicity in leukemia cells. Although genistein is normally considered a ROS scavenging agent,50–53 recent studies proved that this isoflavone disrupts the respiratory chain, causing ROS generation in isolated mitochondria26 and intact cells.44, 54 The present results indicate that genistein rapidly induces a moderate over-accumulation of ROS in leukemia cells which, although non-toxic by itself, may suffice to potentiate ATO toxicity. This conclusion was supported by several evidences, namely the dependence of both ROS production and apoptosis potentiation on genistein concentration and isoflavone structure (i.e., higher efficacy of the aglycone form, genistein, and lower efficacy of the 7-glycosilated form, genistin), the capacity of exogenous H₂O₂ to mimic the potentiating effect of genistein on ATO-provoked apoptosis, and the capacity of BHA and NAC to attenuate apoptosis potentiation. While BHA seemed to operate as a ROS scavenger, the anti-oxidant/protective action of NAC might be executed throughout other mechanisms, such as direct reducing activity39 and protection of mitochondria from oxidative damage.55 Finally, the interpretation of ROS as a pivotal mediator of apoptosis potentiation may provide a satisfactory explanation for anti-tumor drug selectivity, since the assayed anti-tumor agents other than ATO were insensitive to co-treatment with H₂O₂. Of course, our work does not exclude the possibility that factors other than ROS may also contribute to the potentiation by genistein of ATO-provoked apoptosis. For instance, genistein might directly inhibit the expression or activity of MRP and P-gp drug efflux pumps,18 thus decreasing ATO detoxification. In addition, genistein caused G₂/M blockade, and cells at G₂ were reported to be especially sensitive to the toxic action of ATO.56 These and other possibilities are under investigation.

Finally, the present results indicate that the protein kinases p38-MAPK and MAPK are factors linking intracellular oxidation and apoptosis potentiation in ATO plus genistein-treated cells. This conclusion was supported by the concurrence of apoptosis potentiation and kinase over-activation, the capacity of kinase inhibitors to attenuate apoptosis, and the capacity of NAC to attenuate both kinase activation and apoptosis. The specific target(s) of these kinases on the apoptotic machinery were not investigated. Nevertheless, p38-MAPK was reported to mediate Bcl-X_L phosphorylation followed by proteasome-mediated degradation in TNF-treated endothelial cells,57 which may be consistent with the decrease in Bcl-X_L expression in ATO plus genistein-treated cells observed in our experiments. Moreover, p38-MAPK was reported to mediate caspase-8/Bid activation in RAW264.7 macrophages,58 and we also observed caspase-8 and Bid cleavage/activation in ATO plus genistein-treated U937 cells. A final, interesting aspect is that, while p38-MAPK inhibitor prevented apoptosis generation by ATO plus genistein, it did not apparently affect G₂/M arrest by genistein in U937 cells, indicating that G₂ blockade and apoptosis potentiation are independently regulated events. Nonetheless, this observation apparently contrasts with earlier reports describing regulation by p38-MAPK of genistein-provoked G₂ arrest and G₂-related proteins,34 and hence would require a deeper investigation at the molecular level.

In summary, the present results indicate that co-treatment with genistein selectively potentiates apoptosis induction by ATO in human leukemia cells. This response correlates and may be explained at least in part by the pro-oxidant action of the isoflavone, as measured by ROS over-production, and the activation of oxidation-inducible protein kinases (p38-MAPK and AMPK). Thus, the pro-oxidant action of genistein might be exploited to improve the efficacy of ATO as an anti-leukemic agent, and also the efficacy of other oxidation-based therapeutic approaches.

Acknowledgements

The authors would thank Ms. S.X. Leiva for help with artwork preparation.