Boswellic acid acetate induces apoptosis through caspase-mediated pathways in myeloid leukemia cells

Conventional chemotherapy of acute myelocytic leukemia (AML) with either cytarabine or daunorubicin given as a single agent induces complete remission in ∼30% to 40% of patients and combination treatment with both agents induces complete remission in more than 50% of patients. However, long-term, disease-free survival occurs in only 25% to 50% of the patients who achieve complete remission (1–3). The majority of patients with AML still die of this disease. Thus, novel agents need to be developed for treatment of AML.

Boswellic acid and its acetates (Fig. 1) were isolated from the gummy exudates of Boswellia serrata and Boswellia carterri Birdw. Boswellic acid derivatives are inhibitors of topoisomerases and nonredox, noncompetitive, specific inhibitors of 5-lipoxygenase (4–6). Among the natural boswellic derivatives, it has been found that 3-O-acetyl-11-keto-β-boswellic acid was the most active inhibitor of 5-lipoxygenase in rat peritoneal polymorphonuclear leukocytes with an IC₅₀ of 1.5 to 3 μmol/L (4, 7). Using pure topoisomerase assay it was found that both boswellic acid acetate and 3-O-acetyl-11-keto-β-boswellic acid were more potent topoisomerase I and IIα inhibitors than camptothecin and amsacrine or etoposide, respectively (6). Previously, we and other groups have found that boswellic acid acetate induced HL-60 cell differentiation at low concentrations and induced cytotoxic effects through apoptosis induction at higher doses (8–10). Recently, the apoptotic effects of boswellic acids have also been found in malignant glioma cells and colon cancer cells (11, 12). It seems that boswellic acids are potent apoptotic agents to cancer cells. In the present study, the mechanism(s) of apoptosis induction by boswellic acid acetate in six human myeloid leukemia cell lines (NB4, SKNO-1, K562, U937, ML-1, and HL-60 cells) was studied. The obtained data indicate that boswellic acid acetate induces apoptosis in the six cell lines tested regardless of the type of p53 present (wild-type or mutant). It is also independent of reactive oxygen species (ROS) production. Boswellic acid acetate seems to induce apoptosis through a caspase-mediated pathway that is activated by the induction of the death receptors 4 (DR4) and 5 (DR5).

All reagents were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise noted. Boswellic acid acetate (a mixture of 1:1 α- and β-boswellic acid acetate, Fig. 1) was isolated from B. carterii Birdw as described before (8).

NB4, HL-60, SKNO-1, ML-1, U937, and K562 cells were cultured in RPMI 1640 supplemented with 100 units/mL penicillin, 100 μg/mL streptomycin, 1 mmol/L l-glutamine, and 10% heat-inactivated fetal bovine serum. HL-60/neo and HL-60/Bcl-2 cells stably transfected with empty vector or Bcl-2 expression vector, respectively, were provided by Dr. M. Cleary (13).

Leukemia cells were seeded at a density of 1 × 10⁵ and incubated with various concentrations of boswellic acid acetate (0, 2.5, 5, 7.5, 10, 12.5, and 15 μg/mL). Total cell number in each group was counted and the drug concentration that inhibited half of the cell growth (IC₅₀) was calculated. The cell viability was calculated after staining with trypan blue exclusion.

One milliliter RPMI 1640 containing 0.5% agar, 15% fetal bovine serum, and different concentrations of boswellic acid acetate (0, 1, 2.5, 5, 7.5, 10, 12.5, and 15 μg/mL) were plated on the bottom of 35-mm tissue culture dishes as the bottom layer. U937 and K562 cells (1,000 cells) in the logarithmic phase of growth were seeded on the top of these dishes in 1 mL RPMI 1640 containing 0.3% agar and 15% fetal bovine serum. The dishes were cultured at 37°C in a humidified atmosphere containing 5% CO₂ for 7 days. Colonies (>50 cells) were counted using an inverted microscope at 30× magnification.

Apoptotic cells were determined by morphologic observation and DNA fragmentation analysis. For morphologic evaluation, cells were stained with acridine orange and ethidium bromide and assessed by fluorescence microscopy as described previously (14). For DNA fragmentation analysis, cells were harvested by centrifugation and the pellets were suspended in lysis buffer containing 15 mmol/L Tris-HCl (pH 8.0), 20 mmol/L EDTA, and 0.5% Triton X-100. After 30 minutes on ice, samples were centrifuged at 14,000 × g at 4°C for 30 minutes, and cellular DNA was extracted with a Puregene DNA isolation kit (Gentra System, Minneapolis, MN). Electrophoresis was done in 1% agarose gel in 40 mmol/L Tris-acetate buffer (pH 7.4), at 50 V. After electrophoresis, DNA was visualized by ethidium bromide staining.

Hydrogen peroxide production was quantitated using 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate as previously reported (15). Briefly, exponentially growing cells were seeded at 2 × 10⁵ cells/mL and treated with boswellic acid acetate for various times followed by labeling with 0.5 μmol/L 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate for another 2 hours. After washing with PBS, 10,000 cells were analyzed by FACscan (Becton Dickinson, San Jose, CA) with excitation and emission settings of 495 and 525 nm, respectively. Arithmetic histogram statistics analysis was used to determine the mean of the oxidized 2′,7′-dichlorofluorescein peak in each sample.

Mitochondrial membrane potential was measured by using a Mitochondrial Membrane Sensor Kit as described by the manufacturer (Clontech Laboratories, Inc., Palo Alto, CA). Briefly, after boswellic acid acetate treatment, cells were washed with serum-free medium after centrifuging at 350 × g and 4°C for 5 minutes, then the cell pellets were suspended in 1 mL diluted BD MitoSensor reagent per tube. The cells were then incubated at 37°C in a humidified, 5% CO₂ incubator for 15 to 20 minutes. One milliliter of incubation buffer was added to each tube and the cells were recentrifuged. The cell pellets were resuspended in incubation buffer (1 mL) and examined under a microscope using a band-pass filter. The MitoSensor reagent aggregates in the mitochondria of healthy cells, forming red fluorescence. When mitochondrial membrane potentials are altered, the MitoSensor reagent does not accumulate in the mitochondria and remains as monomers in the cytoplasm where it fluoresced green.

Protein extracts (50 μg) prepared with radioimmunoprecipitation assay lysis buffer [50 mmol/L Tris-HCl, 150 mmol/L NaCl, 0.1% SDS, 1% NP40, 0.5% sodium deoxycholate, 1 mmol/L phenylmethylsulfonylfluoride, 100 μmol/L leupeptin, and 2 μg/mL aprotinin (pH 8.0)] and were subjected to electrophoresis on 8% or 12% SDS-polyacrylamide gels. The separated proteins were then transferred to nitrocellulose membranes that were stained with 0.2% Ponceau S red to assure equal protein loading and transfer. After blocking with 5% nonfat milk, the membranes were incubated with a polyclonal antibody to poly(ADP-ribose) polymerase (Boehringer Mannheim, Indianapolis, IN) and monoclonal antibodies to p53, Bcl-2, Bcl-X_L, Bid, caspase-3, caspase-8 and actin (Oncogene Research Products, Cambridge, MA). Immunocomplexes were visualized by chemiluminescence (ECL kit, Amersham Biosciences Corp, Piscataway, NJ).

Total RNA was isolated with PURESCRIPT (Gentra Systems, Minneapolis, MN) from 10⁶ cells. Twenty micrograms RNA were size-fractionated on a 1.2% agarose-2.2 mol/L formaldehyde gel, transferred to a hybrid-N+ membrane (Amersham Biosciences Corp) in 20× SSC, and UV-cross-linked (Stratalinker, Stratagene, La Jolla, CA). DR4 and DR5 cDNAs (provided by Dr. Gen Sheng Wu, Department of Pathology, Wayne State University School of Medicine, Detroit, MI) and glyceraldehyde-3-phosphate dehydrogenase (Ambion, Austin, TX) cDNAs were used for probing the membranes. The probes were labeled with [α³²P]dCTP by random priming to a specific activity of (0.5–1) × 10⁹ cpm/ng. The membranes were prehybridized for 4 hours at 42°C in 50% formamide, 6× saline-sodium phosphate-EDTA, 5× Denhardt's reagent, and 0.2 mL of 1 mg/mL SS DNA, then hybridized with the various radiolabeled probes and visualized by radioautography with Kodak BioMax film.

Data were analyzed for statistical significance using the Student's t test (Microsoft Excel, Microsoft Corporation, Seattle, WA). Differences were considered significant at P < 0.05.

The cell growth inhibition of boswellic acid acetate in leukemia cell lines was measured by direct cell counting and clonogenic assay. Boswellic acid acetate inhibited cell growth in a dose-dependent pattern with IC₅₀s of 5.8, 8.7, 7.3, 6.2, 9.8, and 6.6 μg/mL in NB4, SKNO-1, HL-60, U937, K562, and ML-1 cells at 4 days' treatment, respectively. Because NB4, ML-1, SKNO-1, and HL-60 cells only formed a few colonies in soft agar culture, the effect of boswellic acid acetate on clonogenic forming ability was only tested in U937 and K562 cells. Boswellic acid acetate inhibited the clonogenic formation of both cell lines after 14 days' culture with IC₅₀s 6.6 and 5.1 μg/mL, respectively. The cytotoxic effect of boswellic acid acetate in six myeloid leukemia cell lines was determined by trypan blue staining after 24 hours treatment. As shown in Fig. 2A, 20 μg/mL boswellic acid acetate markedly decreased cell viability after 24 hours of treatment in all of the lines except in K562 cells in which the decrease was only about 30%. Because boswellic acid acetate inhibited cell growth at lower concentrations (<10 μg/mL) without causing cytotoxicity (Fig. 2A), other actions of boswellic acid acetate such as differentiation induction as we reported before may involve in boswellic acid acetate–mediated cell growth inhibition in these cells (8). However, boswellic acid acetate is a potent cytotoxic agent for most leukemia cells at concentrations between 15 and 20 μg/mL (Fig. 2A). Based on the morphologic examination after staining with acridine orange and ethidium bromide, all of the tested cell lines underwent >50% apoptosis after 24 hours of treatment at 20 μg/mL of boswellic acid acetate (Fig. 2B) and apoptotic cells can be detected in HL-60 and NB4 cells after treatment as short as 8 hours at a concentration of 20 μg/mL (data not shown).

p53 was detected in NB4, SKNO-1, and ML-1 cells, but not in HL-60, U937, and K562 cells (Fig. 3A). p53 levels were not changed by boswellic acid acetate (15 μg/mL) treatment in NB4 and SKNO-1 cells, although it was somewhat decreased in ML-1 cells. Boswellic acid acetate treatment induced marked DNA fragmentation in NB4, HL-60, and U937 cells (Fig. 3B). NB4 cells that express mutant p53, and HL-60 cells, which are null for p53, are both highly sensitive to boswellic acid acetate–induced apoptosis, and SKNO-1 and ML-1 cells (which express wild-type p53) were less sensitive to boswellic acid acetate–induced DNA fragmentation. Thus, it seems that p53 is not necessary in boswellic acid acetate–induced apoptosis.

The effects of boswellic acid acetate on the levels of Bcl-2 protein family members, Bcl-2, Bcl-X_L, and Bax, were determined (Fig. 3A). Bcl-2 protein was detected in all of the cell lines except K562 and the Bcl-2 protein levels were not changed in any of the cell lines. HL-60/Bcl-2 cells with overexpression of Bcl-2 displayed only a slight decrease of boswellic acid acetate–induced apoptosis compared with cells transfected with vector alone (Fig. 4). Therefore, Bcl-2 does not seem to be involved in the effect of boswellic acid acetate on apoptosis induction. Bax protein was also detected in all of the cell lines and its level was not changed after boswellic acid acetate treatment. Bcl-X_L protein was detected in K562, SKNO-1, and ML-1 cells, and its level was increased in these lines after boswellic acid acetate treatment. Because the cell lines expressing Bcl-X_L were less sensitive to boswellic acid acetate than the cell lines without Bcl-X_L expression, it seems that Bcl-X_L partially blocked boswellic acid acetate–induced apoptosis in these leukemia cells. Additional studies were undertaken on NB4 and HL-60 cells, which were both highly sensitive to boswellic acid acetate–induced apoptosis (Figs. 2 and 3). It was found that Bid protein was proteolyzed after treatment of the cells with 20 μg/mL boswellic acid acetate, which induced apoptosis. However, Bid was not proteolyzed by 10 μg/mL boswellic acid acetate treatment, which did not induce apoptosis in either of these cell lines (Fig. 5).

Because Bid is a substrate of caspase-8 and was decreased after boswellic acid acetate treatment (Fig. 5), it is likely that caspase-8 will be activated by boswellic acid acetate. Activation of both caspase-8 and caspase-3 occur after the cleavage of their respective precursors (16, 17). The levels of caspase-8 and caspase-3 precursors were compared in HL-60 and NB4 cells with or without boswellic acid acetate treatment (Fig. 5). The levels of the precursors of both caspase-8 and caspase-3 were decreased, and the active cleavage fragment of caspase-8 was detected in apoptotic HL-60 and NB4 cells (Fig. 5). These findings suggest that caspase activation is involved in boswellic acid acetate–induced apoptosis. Caspase-3 is known to be activated after cleavage of its precursor by caspase-8 or by releasing cytochrome c from mitochondria and to cleave poly(ADP-ribose) polymerase in the apoptosis process (15, 16, 18). Therefore, the level of poly(ADP-ribose) polymerase cleavage before and after boswellic acid acetate treatment was determined. Indeed, poly(ADP-ribose) polymerase was cleaved after boswellic acid acetate treatment (Fig. 5). Moreover, boswellic acid acetate–induced apoptosis in both cell lines was completely blocked by Z-VAD-FMK, a general caspase inhibitor (Fig. 6A). To further dissect the members of the caspase family that are involved in boswellic acid acetate–induced apoptosis, several specific caspase inhibitors were also used. As shown in Fig. 6B, a caspase-8 inhibitor (Z-IETD-FMK), but not a caspase-1 inhibitor (AC-YVAD-CMK), blocked boswellic acid acetate–induced apoptosis in NB4 cells (Fig. 6B). A caspase-9 inhibitor (Z-LEHD-FMK) had a weak blocking effect alone and had an additive effect with the caspase-8 inhibitor on blocking boswellic acid acetate–induced apoptosis in NB4 cells (Fig. 6B). These results suggest that caspase-8 activation and partial caspase-9 activation contribute to boswellic acid acetate–induced apoptosis.

Bid belongs to a family of proteins involved in maintaining the potential of the mitochondrial membrane (19). Therefore, the effect of boswellic acid acetate treatment on mitochondrial membrane potential was determined by using fluorescence microscopy and the MitoSensor reagent. The MitoSensor reagent was taken up in mitochondria in control cells and exhibited intense red fluorescence. The MitoSensor reagent did not accumulate in the mitochondria and the dye remained in the monomeric form in the cytoplasm where it fluoresced green in both NB4 and HL-60 cells treated with boswellic acid acetate for 24 hours (Fig. 7). Therefore, the mitochondrial membrane potential was decreased by boswellic acid acetate treatment.

Because ROS are known to decrease mitochondrial membrane potential, cellular hydrogen peroxide levels were determined by fluorescence-activated cell-sorting analysis of cells labeled with 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate. The oxidized 2′,7′-dichlorofluorescein mean peak was not increased in either HL-60 or NB4 cells after treatment with boswellic acid acetate (data not shown). Therefore, the hydrogen peroxide content was not increased by boswellic acid acetate treatment. Moreover, the ROS scavengers, N-acetylcystein and vitamin C, did not inhibit boswellic acid acetate–induced apoptosis (data not shown). These data indicate that boswellic acid acetate induces apoptosis through a ROS-independent pathway.

DR4 and DR5 mRNA levels were measured by using Northern blot analysis. The mRNA levels of both DR4 and DR5 were weakly expressed in NB4 and HL-60 cells and found that the levels of DR4 and DR5 mRNA were increased when these cell lines were treated with 20 μg/mL but not with 10 μg/mL boswellic acid acetate for 24 hours (Fig. 8). Because the up-regulation of death receptors, mainly DR4, is correlated with apoptosis induction, it suggests that boswellic acid acetate induces apoptosis, at least in part, through a death receptor–mediated pathway.

Defective control of apoptosis has recently been shown to play a central role in the pathogenesis of several diseases including leukemia (20). Chemotherapeutic agents and ionizing radiation can induce cancer cell death, at least partially, by the induction of apoptosis. The p53 gene is an important regulator of apoptosis in AML blasts and recent studies suggest that the expression of certain conformational variants of p53 due to mutation are associated with a poor prognosis (21–23). The p53 protein is usually mutated in AML cells and thus agents that can induce apoptosis via a p53-independent pathway should have broad beneficial therapeutic potential in AML patients. Because boswellic acid acetate promotes more apoptosis in NB4 cells, which express mutated p53 (24), and in HL-60 cells, which are null for p53 (25), than in ML-1 cells, which express wild-type p53 (Fig. 3), it seems that p53 is not involved in boswellic acid acetate–induced apoptosis. The antiapoptotic mediator, Bcl-2, has been found to be important for the therapeutic outcome of AML and an association existed between high Bcl-2 levels and decreased sensitivity to chemotherapeutic agent–induced apoptosis in native AML blasts (26–28). Because Bcl-2 functions as an apoptotic inhibitor, agents that can bypass Bcl-2 blockage might be developed into new effective agents for leukemia treatment. Boswellic acid acetate was more active in inducing apoptosis in NB4 and HL-60 cells, which express higher levels of Bcl-2 protein, than in K562 cells, which do not express Bcl-2 protein (Fig. 3), and the overexpression of Bcl-2 in HL-60 cells via transfection only weakly decreased boswellic acid acetate–induced apoptosis compared with cells transfected vector alone (Fig. 4). These findings suggest that Bcl-2–mediated antiapoptotic pathways do not play an important role in boswellic acid acetate–induced apoptosis. Because boswellic acid acetate induced apoptosis in all of the tested AML cell lines, it may provide another modality for AML therapy even with mutant p53 and increased Bcl-2 levels.

It has been shown that many cells undergo a reduction of mitochondrial membrane potential before they exhibit signs of nuclear apoptosis, and the mitochondrial membrane potential has been found to be regulated by numerous effectors (29–31). Previously, we and other groups have found that As₂O₃ and other chemotherapeutic agents decreased mitochondrial membrane potential by producing hydrogen peroxide (15). However, although the mitochondrial membrane potential of both NB4 and HL-60 cells was decreased after boswellic acid acetate treatment (Fig. 7), the level of hydrogen peroxide was not increased in these cells (data not shown). These data suggest that boswellic acid acetate induces apoptosis through a ROS-independent new pathway.

The members of Bcl-2 family are important regulators of the mitochondrial membrane potential (32–35). The data shown in Fig. 3 suggest that Bcl-2, Bcl-X_L, and Bax might not be involved in the boswellic acid acetate–induced decrease in mitochondrial membrane potential because their levels were not changed in any of the cell lines tested. However, the level of Bid was decreased by boswellic acid acetate treatment at a dose that induced apoptosis (Fig. 5). It is known that the cleaved Bid product decreases mitochondrial membrane potential in apoptotic cells (36, 37); thus, it seems that boswellic acid acetate decreases mitochondrial membrane potential, at least in part, through a Bid-mediated pathway. The disruption of mitochondrial membrane integrity involves a loss of metabolic functions, the release of mitochondrial proteins into the cytosol, and the activation of caspase-3 (15, 31). Western blot analysis determined that precursors of both caspase-3 and caspase-8 were decreased in both NB4 and HL-60 cells after boswellic acid acetate treatment (Fig. 5). Furthermore, the general caspase inhibitor Z-VAD-FMK blocked boswellic acid acetate–induced apoptosis (Fig. 6A). It has been found that a decrease in mitochondrial membrane potential will release cytochrome c, which activates caspase-9, and leads to the activation of caspase-3 (18). Although Z-VAD-FMK, a general caspase inhibitor, blocked boswellic acid acetate–induced apoptosis significantly (Fig. 6A), Z-LEHD-FMK, a specific caspase-9 inhibitor, only had minimal blocking effect on boswellic acid acetate–induced apoptosis (Fig. 6B). Therefore, the mitochondrial-mediated pathway probably does not play an important role in boswellic acid acetate–induced apoptosis. Recently it has been found that caspase-8 activates caspase-3 by direct cleavage (16). The levels of pro-caspase-8 and caspase-8 substrates and pro-caspase-3 and Bid were decreased, and cleaved caspase-8 fragment was formed in NB4 and HL-60 cells after boswellic acid acetate treatment (Fig. 5). These data suggest that boswellic acid acetate treatment activates caspase-8 and that the activated caspase-8 decreased mitochondrial membrane potential by Bid cleavage or directly activated caspase-3 by cleavage of pro-caspase-3 (16, 36). Because Z-IETD-FMK, a specific caspase-8 inhibitor, blocked boswellic acid acetate–induced apoptosis significantly, caspase-8 is involved in the boswellic acid acetate effect, and the activation of caspase-8 will be an important event in the boswellic acid acetate–induced apoptosis process. These data are consistent with previous reports that showed boswellic acid derivatives induced apoptosis in colon and hepatoma cells through a caspase-8–mediated pathway (12, 38).

Tumor necrosis factor receptor 1 and an emerging family of membrane death receptors transduce apoptotic signals through caspase-8 activation (39–41). It has been found that apoptosis induction of cancer cells by several chemotherapeutic agents is also associated with the up-regulation of DR4 and DR5 (42–44). DR4 and DR5 mRNA were found to be increased in NB4 and HL-60 cells after boswellic acid acetate treatment, which correlated with apoptosis induction (Figs. 2 and 8). These data suggest that boswellic acid acetate may induce apoptosis by increasing the levels of death receptors and indirectly activating caspase-8 (Fig. 9).

We thank the following investigators for sharing cell lines: Drs. Michel Lanotte (INSERM, Hopital Saint-Louis, Paris, France), NB4 cells; Yoshio Honma (Division of Cancer Treatment, Saitama Cancer Center, Saitama, Japan), SKNO-1 cells; Kazuyasu Nakaya (Laboratory of Biological Chemistry, School of Pharmaceutical Sciences, Showa University, Tokyo, Japan), ML-1 cells; and Michael L. Cleary (Department of Pathology, Stanford University School of Medicine, Stanford, CA), HL-60/BCL2 cells; and Dr. William Scher for critical reading.