Ethacrynic Acid Oxadiazole Analogs Induce Apoptosis in Malignant Hematologic Cells through Downregulation of Mcl-1 and c-FLIP, Which Was Attenuated by GSTP1-1

Ethacrynic acid is a diuretic agent with the ability of inhibiting glutathione S-transferase P1-1 (GSTP1-1) activity (1). GSTP1-1 is a phase II detoxification enzyme that catalyzes the conjugation of reduced glutathione (GSH) with chemotherapeutic agents leading to drug resistance (2). GSTP1-1 levels are increased in chemotherapy-resistant cancer cell lines and tissues (3). On the basis of the GSTP1-1 inhibition ability, ethacrynic acid has been used to overcome chemotherapeutic resistance. It was found that ethacrynic acid potentiates the cytotoxic effects of chemotherapeutic agents such as cisplatin, chlorambucil, melphalan, mitomycin c, and adriamycin in vitro (4, 5). Although ethacrynic acid was put into a phase I clinical trial in combination with thiotepa for advanced cancer treatment (6), only limited enhancement on the efficacy of thiotepa was obtained. One reason was thought to be that the plasma concentration of ethacrynic acid achieved in patients was so low that it may not efficiently inhibit GSTP1-1 activity (6). To improve the ability of ethacrynic acid to inhibit GSTP1-1 activity, we conducted structural modifications and found that esters of ethacrynic acid were more effective than ethacrynic acid in inhibiting GSTP1-1 activity (7, 8). To further improve the activity of ethacrynic acid, we synthesized novel ethacrynic acid derivatives containing a heterocyclic oxadiazole and found that ethacrynic acid oxadiazoles were more potent cell growth inhibitors in human leukemia HL-60 cells than ethacrynic acid, even those with a decreased ability of inhibiting GSTP1-1 activity (8).

Although ethacrynic acid is considered a GSTP1-1 inhibitor, ethacrynic acid alone has been reported to be cytotoxic in many malignant cells (9, 10). By conducting structural activity analyses, it was found that the α-, β-unsaturated carbonyl structure is required for its ability to inhibit leukemia cell growth (7, 11). We found that ethacrynic acid esters containing the α-, β-unsaturated carbonyl structure have increased abilities of inducing cell death in leukemia cells (7). Focusing on one of these ethacrynic acid esters, ethacrynic acid butyl-ester (EABE), we found that it exhibited greater cell growth inhibition and apoptosis induction abilities than did ethacrynic acid (12). Mechanistic studies revealed that EABE-induced apoptosis was correlated with increased levels of reactive oxygen species (ROS), which decreased the levels of the mitochondrial transmembrane potential (MTP) and caused caspase activation. The apoptosis induction due to EABE was found to be abrogated by catalase and an ROS-resistant cell line did not undergo EABE-induced apoptosis. Because ROS are also detoxified by GSTP1-1 (13) and cells expressing high levels of GSTP1-1 are less sensitive to EABE, it implies that EABE induces apoptosis in leukemia cells through a ROS-mediated pathway (12). In this study, we selected two ethacrynic acid heterocyclic oxadiazoles with similar structures, 1-(2,3-dichloro-4-((3-phenyl-1,2,4-oxadiazol-5-yl)methoxy)phenyl)-2-methylenebutan-1-one (6s; Fig. 1A) and 1-(2,3-dichloro-4-((3-(4-(trifluoromethyl)phenyl)-1,2,4-oxadiazol-5-yl)methoxy)phenyl)-2-methylenebutan-1-one (6u; Fig 1A) to study their mechanisms of apoptosis induction. Both 6s and 6u contain the α-, β-unsaturated carbonyl structure with greater antiproliferative effects than ethacrynic acid in HL-60 cells, but 6u does not inhibit GSTP1-1 activity as much as ethacrynic acid (8). Therefore, this pair of compounds is ideal for comparing the role of GSTP1-1 inhibition in apoptosis induction. The roles of ROS and GSTP1-1 in the apoptosis induction due to both compounds were determined. The mechanism of apoptosis induction was explored by investigating the regulation of the antiapoptotic proteins c-FLIP, Mcl-1, and XIAP.

6s, 6u, 6s-1, 6u-1, and 6u-GS were synthesized by Guisen Zhao (School of Pharmaceutical Sciences, Shandong University, Jinan, Shandong, China; ref. 8). Acridine orange, ethidium bromide, N-acetylcysteine (NAC), MG132, catalase, rapamycin, and 12-O-tetradecanoylphorbol-l3-acetate (TPA) were purchased from Sigma Chemical Co.. The Raf inhibitor sorafenib were purchased from LC Laboratories. The general caspase inhibitor Z-VAD-FMK was obtained from Calbiochem. 5,6-Carboxy-2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) was obtained from Molecular Probes. Antibody to PARP, pro-caspase-3, and caspase-8 was obtained from BD Biosciences, c-FLIP from Alexis Biochemicals; Mcl-1, survivin, Bcl-2, GSTP1-1, and β-actin from Santa Cruz Biotechnology, Inc.; XIAP from BD Pharmingen; and cleaved caspase-9, -3, extracellular signal–regulated kinase (ERK)1/2, p-ERK(Thr202/Tyr204), p70S6K, p-p70S6K (Thr³⁸⁹), p-S6 ribosomal protein (Ser^235/236), p-4E-BP1 (Thr37/46), eIF4E, p-eIF4E(Ser209), mTOR, and p-mTOR(Ser2448) from Cell Signaling Technology, Inc..

Human acute myeloid leukemia HL-60 cells, a H₂O₂-resitant derivative of HL-60 cells (14), HP100-1 cells (obtained from the Japanese Cell Bank), and human chronic myeloid-derived K562 cells were cultured in RPMI-1640 medium supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin, 1 mmol/L l-glutamine, and 10% (v/v) heat-inactivated FBS. Jurkat cell subclone A3 and its caspase-8–deficient subclone, I 9.2 (obtained from American Type Culture Collection), were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated FBS (15). Once received in our laboratory, we did not authenticate all cell lines. RV4 and RG19 cells are clones of human Raji cells transfected with an empty vector and a GSTP1-expression plasmid (13).

Comparative levels of apoptotic cells were determined by morphologic observation and fluorescence-activated cell sorting (FACS) analysis after staining with propidium iodide (PI) and annexin V (16). For morphologic observation, cells were stained with acridine orange and ethidium bromide and assessed by fluorescence microscopy as described previously (16). For FACS analysis with PI staining, cells were fixed with ice-cold 70% ethanol at a density of 1 × 10⁵ cells/mL and treated with 200 ug/mL RNase for 30 minutes at 37°C. PI was then added to a final concentration of 50 μg/mL and the DNA content was quantitated by flow cytometry (Becton Dickinson) with an excitation wavelength of 488 nm and an emission wavelength of 625 nm. Data were analyzed using CellQuest (Becton Dickinson) software. For FACS analysis with annexin V staining, Annexin V–FITC Apoptosis Detection Kit (BD Biosciences) was used. Data were analyzed using CellQuest software on 10,000 events (13).

Intracellular H₂O₂ production was monitored by flow cytometry using DCFH-DA. Inside cells, DCFH-DA is hydrolyzed to polar 2′,7′-dichlorodihydrofluorescein and oxidized by intracellular oxidants to highly fluorescent, 2′,7′-dichlorofluorescein. Cells (1 × 10⁵ cells/mL) were mixed with 0.5 μmol/L of DCFH-DA and incubated for 1 hour at 37°C, then analyzed by flow cytometry with excitation and emission wavelengths of 495 and 525 nm, respectively (17).

The levels of intracellular GSH were measured by a monochlorobimane (mBCl) fluorometric method in which mBCl was used as a sensitive and specific probe to detect GSH in intact cells (18). Briefly, 3 × 10⁶ cells treated with both compounds were washed with PBS once, resuspended in 1 mL PBS containing 100 μmol/L mBCl, and maintained at 37°C in the dark for 30 minutes before analysis. The formation of the fluorescent adduct (GS-mBCl) was monitored with a fluorescence spectrometer using excitation and emission wavelengths of 395 and 482 nm, respectively. The GSH content was calculated as nanomoles per 10⁶ cells based on a GSH standard curve.

Protein extracts (50 μg) in lysis buffer containing 50 mmol/L Tris–HCl, 150 mmol/L NaCl, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, 1 mmol/L phenylmethylsulfonylfluoride (PMSF), 100 μmol/L leupeptin, and 2 μg/mL aprotinin (pH, 8.0) were separated on 8% or 12% SDS-PAGE, transferred to nitrocellulose membranes, and blocked with 5% nonfat milk. The membranes were incubated with specific antibodies overnight at 4°C. Immunocomplexes were visualized using enhanced chemiluminescence Western blotting detection reagents (Amersham Biosciences Inc.; ref. 17).

GSTP1 siRNA and a control siRNA were purchased from Santa Cruz Biotechnology, Inc. siRNAs were transfected into K562 cells by a nucleofector (Amaxa) following the manufacturer's instructions. Briefly, 2 × 10⁶ cells were electroporated in 100 μL of nucleofector solution (Amaxa Reagent V) with siRNA (200 pmol), using preselected Amaxa Program T-003. Cells plated in 6-well plates with 2 mL supplemented RPMI-1640 medium for 18 hours at 37°C were treated with 6s and 6u for a further 12 hours, harvested, and used for protein analysis by Western blotting.

Data were analyzed for statistical significance using the Student t test (Microsoft Excel, Microsoft Corp.). A P value of less than 0.05 was considered statistically significant.

Levels of apoptotic cells were first identified on the basis of morphologic changes after staining with acridine orange and ethidium bromide in HL-60 cells after treatment with 6s or 6u, as we previously reported (16). Both 6s and 6u induced apoptosis at concentrations from 3 to 6 μmol/L (Fig. 1B). Treatment with 6s or 6u at 6 μmol/L for 6 hours induced apoptosis in 42% and 57% of HL-60 cells, respectively (Fig. 1B). The comparative levels of apoptotic cells after treatment with 6s or 6u were confirmed by measuring the DNA fragmentation (hypodiploid DNA) using FACS after staining with PI (Fig. 1C). Previously, we found that EABE-induced apoptosis in HL-60 cells was mediated by production of ROS and was inhibited by catalase (12). We measured the levels of H₂O₂ in HL-60 cells after treatments with 6s and 6u using a H₂O₂-sensitive fluorescent probe, DCFH-DA. Intracellular H₂O₂ levels were increased approximately 83% and 96% after treatments with 6s and 6u for 1 hour, respectively (Supplementary Fig. S1). However, as the treatment was prolonged, less H₂O₂ content was detected (Supplementary Fig. S1). To determine whether H₂O₂ plays an important role in 6s- or 6u-induced apoptosis, HL-60 cells were pretreated with catalase followed by 6s and 6u. Unlike EABE-treated cells (12), catalase did not block apoptosis induced by either 6s or 6u (Supplementary Fig. S2A) even though they blocked H₂O₂ production (Supplementary Fig. S2B). In contrast, catalase blocked apoptosis induced by addition of H₂O₂ (Supplementary Fig. S2A). The H₂O₂-resistant HP100-1 subclone was sensitive to apoptosis induction by treatments with either 6s or 6u. Although there was less of an apoptotic effect in HP100-1 cells than in HL-60 cells after treatment with 6s or 6u (Fig. 1C), there was no difference in apoptotic cell numbers in both cell lines after treatment with increased concentrations (data not shown). These data suggest that ROS are not a principle inducer of apoptosis in HL-60 cells treated with either 6s or 6u. Western blot analysis revealed that cleaved caspase-3, -8, and -9 were detected (Fig. 1D), indicating that all three effecting caspases were activated in HL-60 and HP100-1 cells. The activation of caspase-9 is controlled by the antiapoptotic proteins Bcl-2, Bcl-xL, and Mcl-1 by maintaining MTP (19). Bcl-xL is not expressed in HL-60 cells (20). The levels of Bcl-2 and Mcl-1 proteins in both HL-60 and HP100-1 cells after treatments with either 6s or 6u were compared and it was found that Mcl-1 levels, but not those of Bcl-2, were decreased in both cell lines (Fig. 1D). Caspase-8 activation is repressed by the antiapoptotic protein c-FLIP (21) and the levels of c-FLIP were decreased after treatment with 6s or 6u in both cell lines (Fig. 1D). The activities of both caspase-3 and -9 are directly inhibited by XIAP and survivin (22, 23). The levels of XIAP, but not survivin, were decreased by treatment with either 6s or 6u (Fig. 1D). These data suggest that the downregulation of c-FLIP, Mcl-1, and XIAP after treatment with either 6s or 6u leads to the activation of caspase-3, -8, and -9.

Caspase-8 activation plays a critical role in the extrinsic apoptotic pathway (24). To test whether the activation of caspase-8 plays an important role in the apoptosis induction because of the treatment with either 6s or 6u, Jurkat sublcones, A3-expressing caspase 8 and I 9.2 without caspase-8 expression, were used. 6s and 6u at a concentration of 2.5 to 10 μmol/L induced apoptosis equally in A3 cells and I 9.2 cells based on the FACS analysis after staining with annexin V (Fig. 2A). Although FasL at 5 ng/mL induced 50% of A3 cells to undergo apoptosis, I 9.2 cells were not sensitive to FasL treatment at the same concentration (Fig. 2A). Reduction in Mcl-1, c-FLIP, and XIAP protein levels was detected in both I 9.2 and A3 cells after treatment with either 6s or 6u (Fig. 2B). I9.2 cells seem to be more responsive than A3 cells to both 6s- and 6u-induced PARP cleavage (Fig. 2B) and apoptosis induction after treatment at 2.5 μmol/L (Fig. 2A). These data suggest that the activation of caspase-3 and -9 should play a more important role than caspase-8 in 6s- and 6u-induced apoptosis.

Previously, we found that K562 cells express higher levels of GSTP1-1 than HL-60 cells and are less sensitive to ethacrynic acid and EABE (12). Now, we compared the levels of GSTP1-1 protein and activity among K562, A3, and I 9.2 cells and found that K562 cells expressed the highest GSTP1-1 protein and activity levels among the three cell lines (Fig. 3A). Unlike A3 and I 9.2 cells (Fig. 2A), K562 cells did not respond to 6s- and 6u-induced apoptosis at a concentration of 5 μmol/L even after treatment for 12 hours (Fig. 3B). When the concentration was increased to 10 μmol/L, about 27.8% and 47.4% of cells in the sub-G₁ phase were detected after treatment with 6s or 6u for 12 hours, respectively (Fig. 3B). The levels of c-FLIP, Mcl-1, and XIAP were decreased in K562 cells after treatment with 10 μmol/L of 6s or 6u, but rarely after treatments with 5 μmol/L of 6s or 6u (Fig. 3C). To determine whether the high level of GSTP1-1 protein in K562 cells mediates the reduced sensitivity, the GSTP1-1 level was knocked down using siRNA. Silencing of GSTP1-1 increased 6s- and 6u-induced PARP cleavage and the reduction of c-FLIP, Mcl-1, and XIAP protein levels (Fig. 3D). Annexin V staining revealed that apoptosis induction by both 6s and 6u was enhanced by silencing of GSTP1 in K562 cells (Supplementary Fig. S3). Previously, we have found that Raji cells with overexpression of GSTP1-1 were not sensitive to EABE-induced apoptosis (12). The abilities of both 6s and 6u to induce apoptosis and to decrease the protein levels of c-FLIP, Mcl-1, and XIAP were determined and compared in a clone of Raji (RG19), transfected to express GSTP1-1, and a clone transfected with the vector alone (RV4). RV4 and RG19 cells were treated with both agents and apoptosis was quantified by FACS after staining with PI. Remarkably, although more than 40% of RV4 cells underwent apoptosis after treatment with either 6s or 6u, only about 15% of RG19 cells underwent apoptosis (Fig. 4A). Western blot analysis revealed that both 6s and 6u decreased Mcl-1, c-FLIP, and XIAP protein levels in RV4 cells, but not in RG19 cells after treatments with 5 μmol/L of either 6s or 6u (Fig. 4B). These data suggest that GSTP1-1 protects cells against 6s- and 6u-induced apoptosis as well as reduction in Mcl-1, c-FLIP, and XIAP protein levels.

Ethacrynic acid interacts with GSH with or without catalysis of GSTP1-1, which then leads to decreases in the intracellular levels of GSH (25, 26). The intracellular levels of GSH in HL-60 cells after treatments with various concentrations of 6s or 6u were determined. Both agents significantly decreased GSH levels after treatment with 6 μmol/L for 6 hours (Fig. 5A). Previously, we and others found that NAC blocked ethacrynic acid- and EABE-induced reduction of GSH (9, 10, 12). The ability of NAC to block either 6u- or 6s-induced apoptosis and reduction in c-FLIP, Mcl-1, and XIAP protein levels were determined. NAC blocked the reduction of GSH levels by 6s and 6u, but it did not increase GSH levels (Fig. 5B). NAC alone did not change the levels of c-FLIP, Mcl-1, and XIAP, but NAC blocked the reduction in c-FLIP, Mcl-1, and XIAP protein levels due to treatment with 6s or 6u. By measuring the DNA fragmentation using FACS after staining with PI, it was found that NAC completely blocked DNA fragmentation induced by both 6s and 6u (Supplementary Fig. S4). It has been found that ethacrynic acid interacts with GSH or NAC through the α-, β-unsaturated carbonyl group (27). Because both 6s and 6u contain the α-, β-unsaturated carbonyl group, it seems that NAC competes with GSH for binding to 6s or 6u in cells. To test the requirement of the α-, β-unsaturated carbonyl group to induce apoptosis, we synthesized a GSH conjugate of 6u (6u-GS) and α-, β-saturated carbonyl 6u-1 and 6s-1 (Fig. 5D) and determined their abilities to decrease GSH levels and to induce apoptosis. None of the three compounds decreased GSH levels (Fig. 5D) or induced apoptosis (Supplementary Fig. S5).

It has been found that the decreases in c-FLIP, Mcl-1, and XIAP protein levels can be initiated through proteolysis due to proteasome and/or caspase activation (28–31). We tested the dose-dependent effects of 6s on the reduction of c-FLIP, Mcl-1, and XIAP protein levels and found that the downregulation of the three proteins correlated with the activation of caspase-3, -8, and -9 (Fig. 6A). To determine whether activated caspases led to the reduction of three proteins, HL-60 cells were pretreated with or without the pan-caspase inhibitor Z-VAD-FMK followed by 5 μmol/L of 6s for 6 hours. Z-VAD-FMK blocked the cleavage of caspase-3, -9, and PARP, which indicated that the activities of these caspases were inhibited by Z-VAD-FMK (Fig. 6B). However, pretreatment with Z-VAD-FMK blocked the downregulation of XIAP, but not that of c-FLIP and Mcl-1 (Fig. 6B). Recently, we have found that the proteasome inhibitor MG132 blocked ethacrynic acid plus arsenic trioxide-induced reduction of the Mcl-1 protein level in HL-60 cells (32). Interestingly, MG132 did not block the reduction of Mcl-1 or c-FLIP protein levels by treatment with 6s (Fig. 6C). Because MG132 did not block 6s-induced apoptosis as determined by measuring PARP cleavage, the downregulation of the XIAP protein level should result from the activated caspases. It has been found that the Mcl-1 and c-FLIP protein levels are translationally upregulated by mTORC1, a downstream target of PI3K/AKT (33). mTOR is activated by AKT and it stimulates protein translation by phosphorylating the eIF4E-binding protein (4E-BP1) as well as p70S6K which, in turn, phosphorylates S6. It has also been found that phosphorylation of eIF4E increased the translation of Mcl-1 and that multiple kinase inhibitor sorafenib inhibits Mcl-1 translation by suppressing ERK and eIF4E phosphorylation (34, 35). The relative levels of phosphorylated p-mTOR, p-4E-BP1, p-ERK, p-eIF4E, p-p70S6K, and p-S6 were determined. Correlated with the decreases in Mcl-1 and c-FLIP protein levels, the levels of p-mTOR, p-ERK, p-eIF4E, p-p70S6K, and p-S6 were decreased after 6s treatment (Fig. 6D). It should be pointed out that p-4E-BP1 levels were also decreased by 6s treatment at concentrations as lower as 3 μmol/L, which were not further decreased by increased concentrations. These data suggest that reduction in the Mcl-1 and c-FLIP protein levels by 6s treatment is probably due to the inhibition of both mTOR/4E-BP1 and ERK/eIF4E signaling. To determine which pathway plays an important role in 6s-induced apoptosis, we compared the downregulation and Mcl-1 and c-FLIP protein levels and apoptosis induction by mTOR inhibitor rapamycin and kinase inhibitor sorafenib. It was found that sorafenib inhibited phosphorylation of ERK and eIF4E, but not mTOR. Rapamycin inhibited the phosphorylation of mTOR, but not ERK and eIF4E. 6s inhibited the phosphorylation of ERK, eIF4E, and mTOR. The reduction of c-FLIP and Mcl-1 protein levels and apoptosis were induced by both 6s and sorafenib, suggesting that inhibition of ERK/eIF4E pathway plays an important role in 6s-induced apoptosis (Fig. 6E). It has been reported that TPA activated ERK in HL-60 cells and blocked apoptosis induction by some drugs (36); we tested whether TPA pretreatment blocked 6s-induced apoptosis. TPA treatment increased the levels of phosphorylated ERK and eIF4E as well as the protein levels of c-FLIP and Mcl-1 and attenuated 6s-induced apoptosis (Fig. 6F). These data suggest that 6s decreases c-FLIP and Mcl-1 protein levels by inhibiting ERK/eIF4E signaling.

Ethacrynic acid at high concentrations induces cell death and apoptosis (10, 37) and several groups, including ours, found that the cytotoxic effect of ethacrynic acid requires the α,β-unsaturated carbonyl group (7, 9, 38). Disruption of the α,β-unsaturated carbonyl group abrogates the cytotoxic effect of ethacrynic acid (7). Ethacrynic acid esters containing the α,β-unsaturated carbonyl group induce cell death at much lower concentration than ethacrynic acid (7, 12). Like ethacrynic acid esters, ethacrynic acid oxadiazole analogs contain the α,β-unsaturated carbonyl group and are more effective than ethacrynic acid and ethacrynic acid esters to inhibit HL-60 cell growth (8). Although ethacrynic acid and ethacrynic acid esters maintain the ability of inhibiting GSTP1-1 activity, ethacrynic acid oxadiazole analogs have variant abilities of inhibiting GSTP1-1 activity. 6s inhibits the activity of GSTP1-1 at low concentrations, 6u does not (8). Because both 6s and 6u equally induce apoptosis in HL-60 cells (Fig. 1B), it suggests that GSTP1-1 activity inhibition does not contribute to the apoptosis induction in HL-60 cells due to these compounds. Previously, we found that EABE, one of the ethacrynic acid esters, induced apoptosis in leukemia cells via a ROS-dependent pathway, which was blocked by catalase (12). Although both 6s and 6u produce ROS in HL-60 cells (Supplementary Fig. S1), catalase does not block apoptosis induction due to either 6u or 6s treatment even ROS production was blocked (Supplementary Fig. S2). ROS-resistant HP100-1 cells are sensitive to apoptosis induction due to treatments with either 6s or 6u (Fig. 1C and D). These data suggest that both 6s and 6u induce apoptosis through a different pathway from that of ethacrynic acid and EABE, that is, independent of ROS production.

Both extrinsic and intrinsic effective caspases are activated by treatments with 6s and 6u (Fig. 1D). Apoptosis induction due to 6s is blocked by the pan-caspase inhibitor Z-VAD-FMK (Fig. 6B). Correlated with the caspase activation, the antiapoptotic c-FLIP, Mcl-1, and XIAP protein levels, but not those of survivin and Bcl-2, are decreased (Fig. 1D). The downregulation of c-FLIP leads to caspase-8 activation and the downregulation of Mcl-1 leads to caspase-9 activation. I 9.2 cells defecting caspase-8 expression, which are insensitive to apoptosis induction by low concentration of FasL (Fig. 2A), are as sensitive to either 6s- or 6u-induced apoptosis as A3 cells expressing caspase-8 (Fig. 2A). These data suggest that the extrinsic apoptotic pathway may not play an important role in both 6s- and 6u-induced apoptosis. Both 6s and 6u decrease the c-FLIP, Mcl-1, and XIAP protein levels in A3 and I 9.2 cells as that in HL-60 cells (Fig. 2B). Activated caspases and the proteasome have been reported to cause reduction of c-FLIP, Mcl-1, and XIAP protein levels (28–31). The pan-caspase inhibitor Z-VAD-FMK inhibited PARP cleavage and blocked the downregulation of XIAP, but not that of c-FLIP and Mcl-1 (Fig. 6B). These data indicate that the downregulation of XIAP by 6s and 6u is secondary to the activation of caspases (31). Although MG132 has been found to block the downregulation of Mcl-1 and c-FLIP in HL-60 cells treated with other agents (32, 39), it does not block the downregulation of c-FLIP and Mcl-1 protein in HL-60 cells treated with 6s (Fig. 6C). Therefore, 6s-induced reduction of c-FLIP and Mcl-1 protein levels is not through increased proteolysis. Both Mcl-1 and c-FLIP are short-lived proteins and their synthesis is controlled by mTOR signaling and ERK/eIF4E signaling (40, 41). Sorafenib has been reported to inhibit Mcl-1 translation by inhibiting ERK/eIF4E signaling and induce apoptosis in leukemia cells (35). Although we found that both mTOR and ERK signaling are inhibited by 6s by measuring the levels of p-mTOR, p-4E-BP1, p-eIF4E, and p-ERK (Fig. 6D), 6s acts like sorafenib to decrease the levels of c-FLIP and Mcl-1 associated with inhibiting ERK and eIF4E phosphorylation and to induce apoptosis in HL-60 cells (Fig. 6E). Furthermore, we found that TPA activated ERK and increased the levels of c-FLIP and Mcl-1 and attenuated 6s-induced apoptosis (Fig. 6F). These data suggest that 6s decrease the protein levels of c-FLIP and Mcl-1 by inhibiting ERK/eIF4E signaling and that reduction of Mcl-1 and c-FLIP protein levels contributes to 6s-induced apoptosis. Recently, it has been found that ethacrynic acid exhibited cytotoxicity in chronic lymphocytic leukemia cells by inhibiting Wnt/β-catenin signaling by using the α,β-unsaturated carbonyl group through covalently binding to thiol groups presented on transcription factors (9, 38). 6s-1 and 6u-1 with the saturated carbonyl group (Fig. 5D) lose apoptosis induction ability (Supplementary Fig. S5). NAC blocks reduction of c-FLIP and Mcl-1 protein levels (Fig. 5C) as well as the apoptosis induction by treatments with 6s and 6u (Supplementary Fig. S4). The blockade of 6s and 6u action by NAC should not be due to its antioxidant effects, but rather due to its direct chemical interaction with α,β-unsaturated carbonyl group of 6s and 6u.

GSH is a primary regulator of intracellular redox (42). Intracellular GSH levels are depleted before the onset of apoptosis induced by various agents (43). The levels of GSH in HL-60 cells after treatment with 6s and 6u are decreased (Fig. 5A and B). We prospect that GSH depletion by 6s and 6u is due to formation of GSH conjugates, which should not be active ones. Indeed, we found that 6u-GS loses the ability of inducing apoptosis and depleting GSH, suggesting that GSH protects cells from apoptosis induction by 6s and 6u through forming inactivated conjugates with either compound. Conjugation of GSH with 6s and 6u could be catalyzed by GSTP1-1. K562 cells expressing high levels of GSTP1-1 are less sensitive to 6s and 6u (Fig. 3B) and that silencing of GSTP1-1 sensitizes K562 cells to 6s- and 6u-induced apoptosis (Fig. 3D). Apoptosis induction due to 6s and 6u in Raji cells is abrogated by overexpression of GSTP1-1 (Fig. 4). These data support the idea that GSTP1-1 detoxifies 6s and 6u probably by catalyzing conjugation with GSH in a similar way as it detoxifies ethacrynic acid (44). Because increased levels of GSTP1-1 have been found to cause ethacrynic acid resistance (45), cells expressing GSTP1-1 should also be less sensitive to treatment with 6s or 6u. Therefore, both compounds could selectively target malignant cells lacking GSTP1-1 expression without causing toxicity to normal cells, which express GSTP1-1. It has been found that more than half of B-cell lymphomas (46), 95% of invasive adenocarcinomas of the prostate (47), and 65% of hepatitis B virus (HBV)-associated hepatocellular carcinomas (48) lose GSTP1-1 expression. These novel ethacrynic acid analogs should have potential to be developed as therapeutic agents for malignant cancer cells that have lost GSTP1-1 expression.

In summary, we found that ethacrynic acid oxadiazole analogs induce apoptosis through a novel pathway different from that of ethacrynic acid and its esters. Both 6s and 6u induce apoptosis through downregulation of Mcl-1 and c-FLIP, which leads to the activation of caspase-8 and -9, respectively. The activated caspase-8 and -9 lead to the activation of caspase-3, which causes XIAP proteolysis. The combined effects of downregulation of c-FLIP, Mcl-1, and XIAP initiate maximal caspase activation and apoptosis. The downregulation of c-FLIP and Mcl-1 and apoptosis induction by these compounds are inactivated by GSTP1-1 (Fig. 6G). These new compounds have potential to be developed as novel therapeutic agents for cancer cells without GSTP1-1 expression.

No potential conflicts of interest were disclosed.

Conception and design: G. Liu, L. Zhao, Y. Jing

Development of methodology: G. Liu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): G. Liu, Y. Wang, L. Zhao, Y. Jing

Writing, review, and/or revision of the manuscript: G. Liu, R. Wang, Y. Jing

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P. Li, G. Zhao

Study supervision: Y. Jing

The authors thank William Scher for critical reading of the article.

This work was supported by National Science Foundation of China (81028015 and 21072115) and The Samuel Waxman Cancer Research Foundation.

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.