• Open Access

Hydroxychavicol, a Piper betle leaf component, induces apoptosis of CML cells through mitochondrial reactive oxygen species-dependent JNK and endothelial nitric oxide synthase activation and overrides imatinib resistance

Authors


To whom correspondence should be addressed. E-mail: santu2@iicb.res.in

Abstract

Alcoholic extract of Piperbetle (Piper betle L.) leaves was recently found to induce apoptosis of CML cells expressing wild type and mutated Bcr-Abl with imatinib resistance phenotype. Hydroxychavicol (HCH), a constituent of the alcoholic extract of Piper betle leaves, was evaluated for anti-CML activity. Here, we report that HCH and its analogues induce killing of primary cells in CML patients and leukemic cell lines expressing wild type and mutated Bcr-Abl, including the T315I mutation, with minimal toxicity to normal human peripheral blood mononuclear cells. HCH causes early but transient increase of mitochondria-derived reactive oxygen species. Reactive oxygen species-dependent persistent activation of JNK leads to an increase in endothelial nitric oxide synthase-mediated nitric oxide generation. This causes loss of mitochondrial membrane potential, release of cytochrome c from mitochondria, cleavage of caspase 9, 3 and poly-adenosine diphosphate-ribose polymerase leading to apoptosis. One HCH analogue was also effective in vivo in SCID mice against grafts expressing the T315I mutation, although to a lesser extent than grafts expressing wild type Bcr-Abl, without showing significant bodyweight loss. Our data describe the role of JNK-dependent endothelial nitric oxide synthase-mediated nitric oxide for anti-CML activity of HCH and this molecule merits further testing in pre-clinical and clinical settings. (Cancer Sci 2012; 103: 88–99)

Imatinib (also known as STI571 or Gleevec), a small-molecule inhibitor of the Bcr-Abl kinase, has been used successfully to treat chronic myeloid leukemia,(1) but resistance has emerged against this drug. The T315I mutation is the most predominant among the mutations found in imatinib-resistant patients.(2) None of the available approved drugs have been effective in circumventing this T315I mutation.(3)

Recent reports suggest that the alcoholic extract of Piper betle (Piper betel L.) leaves induces apoptosis of imatinib-resistant cells(4) and shows activity against T315I tumor xenografts.(5) The deep green heart-shaped leaves commonly referred to as “betel leaves” are traditionally consumed as a mouth freshener in Eastern Asia.(6) Hydroxychavicol (HCH), a phenolic compound of Piper betle leaves has been shown to have anti-mutagenic and anti-carcinogenic activity.(7,8) HCH possesses antimicrobial, antioxidant and anti-inflammatory properties. Recent studies also suggest apoptosis of oral (KB) carcinoma cells by HCH through induction of reactive oxygen species (ROS). None of the previous studies suggest any mechanisms downstream of ROS for HCH-induced apoptosis.(9)

Reactive oxygen species are products of aerobic metabolism of cells. Tumor cells have higher levels of intracellular ROS than their normal counterparts.(10) This creates opposite effects upon augmentation of intracellular ROS on cell proliferation in normal cells versus cancer cells.(10) As the basal level of intracellular ROS is low in normal cells, its increase, to a certain extent, is associated with cell proliferation. However, higher levels of ROS in tumor cells close to the threshold of cytotoxicity are associated with higher cellular proliferation, and additional increases of intracellular ROS are likely to inhibit tumor cell proliferation. Redox stress-dependent selective killing of cancer cells has been experimentally validated with several ROS-inducing natural and synthetic molecules.(10) Because ROS enhancement and consequent anti-proliferation signaling in tumor cells is likely to act through different pathways to most anti-tumor drugs, it can be effectively combined with existing therapies to create new, more effective treatment combinations that can also be used to override different drug resistance.

The present study evaluates the anti-CML activity of HCH, a major constituent of the alcoholic extract of Piper betle leaves. In this report, we demonstrate that HCH induces apoptosis of CML cells expressing wild type and mutated Bcr-Abl, including the untreatable T315I mutation, and acts through JNK pathway in a ROS-dependent manner, which in turn activates endothelial nitric oxide synthase (eNOS) to kill CML cells. Our data establish the role of JNK-dependent eNOS-mediated nitric oxide (NO) in the anti-CML activity of HCH and suggest that a redox manipulation strategy might also have clinical application against drug-resistant CML.

Materials and Methods

Chemicals and reagents.  The p38 specific inhibitor SB203580, extracellular signal related kinase 1/2 (ERK1/2) specific inhibitor PD98059, JNK specific inhibitor SP600125, N-acetylcysteine (NAC), 2′,7′-dichlorofluorescein diacetate (DCF-DA), 4-amino-5-methylamino-2′, 7′-difluorofluorescein (DAF-FM),(11) dihydroethidium (DHE),(12) 2-4-carboxyphenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (C-PTIO), N5-(1-Iminoethyl)-l-ornithine dihydrochloride (L-NIO), aminoguanidine (AG), Thiolyte Monochlorobimane (MBCl), S-methyl-l-thiocitrulline (SMTC), LY 294002,(13) Z-VAD-FMK, Z-IETD-FMK and LEHD-CHO were obtained from Calbiochem (San Diego, CA, USA). Polyethylene glycol-conjugated catalase (Peg-Cat), Peg-conjugated superoxide dismutase (Peg-SOD),(14) SOD inhibitor diethyldithiocarbamate (DETC) and catalase inhibitor aminotriazole (ATZ)(15,16) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Antibodies were purchased from the following suppliers: specific antibodies to phospho-ERK 1/2 (Tyr-204) and actin, were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibodies to cytochrome c, caspase-3, caspase-9, poly-adenosine diphosphate-ribose polymerase (PARP), eNOS, inducible nitric oxide synthase (iNOS), neuronal nitric oxide synthase (nNOS), phospho p38 (Thr180/Tyr182) and phospho eNOS (Ser 1177) were from BD Biosciences (San Josh, CA, USA). Anti-p38 was obtained from Calbiochem. Antibodies to phospho-JNK (Thr 183/Tyr 185), JNK, and ERK caspase-8 were from Cell Signaling Technology (Beverly, MA, USA). Matrigel was obtained from BD Biosciences.

Cell lines and culture.  The Philadelphia chromosome-positive CML cell line K562 was purchased from the American Type Culture Collection (Manassas, VA, USA).(17) Other Bcr-Abl-opositive cell lines KU812(18) and KCL-22(19) were generously provided by Dr Carlo Gambacorti-Passerini (Instituto Nazionale Tumori, Milan, Italy). In addition, several cell lines expressing wild type Bcr-Abl and different point mutations of Bcr-Abl kinase domain leading to imatinib resistance were utilized.(20) These are stable BaF3 cell lines expressing full-length wild type Bcr-Abl (Ba/F3P210WT) and Bcr-Abl with 10 different kinase domain point mutations: Ba/F3P210T315I, Ba/F3P210H396P, Ba/F3P210H396R, Ba/F3P210E255K, Ba/F3P210E255V, Ba/F3P210M244V, Ba/F3P210F359V, Ba/F3P210M351T, Ba/F3P210Q252H, Ba/F3P210Y253F).(21) These cell lines were procured from Dr Brian Druker, Howard Hughes Medical Institute, Oregon Health Sciences University, Oregon, USA. These cell lines were maintained as suspension cultures in RPMI-1640 containing 10% FBS, 100 U/mL penicillin G, 100 μg/mL streptomycin (Life Technologies, New Delhi, India).

Purification of hydroxychavicol from Piper betle leaves.  The leaves of Piper betle (Piperaceae) were procured from different areas of West Bengal, India. A voucher specimen was deposited at the Division of Chemistry, Indian Institute of Chemical Biology, Kolkata. A paste was made from 5.0 kg fresh leaves of Piper betle in a mixture-blender in methanol (2 L) and was placed in a glass percolator (5 L capacities) with the addition of 1000 ml of methanol. The extraction process was carried out for 16 h. The extract was filtered through Whatman No. 1 filter paper (Sigma Aldrich, St. Louis, MO, USA) and the filtrate was collected. This process of extraction was repeated twice (3 L × 3). The combined extract was evaporated to dryness in a flash evaporator under reduced pressure to completely remove residual methanol. The residual substance was then dried in desiccators under high vacuum and the semi-solid mass weighing 106.5 gm was stored at 4°C until fractionation.

The HPLC analysis of the methanolic extract of Piper betle leaves was performed using a C18 Xterra (Waters) (4.6 × 250 mm) analytical column, solvent system of water:methanol:acetic acid (76:23:1), a flow rate of 1 mL/min and detection in UV at 210 nm.

The methanol extract was partitioned between ethyl acetate and water. The aqueous layer was further extracted with n-butanol. Removal of the solvent in vacuo from the ethyl acetate-soluble portion, the n-butanol-soluble portion and the aqueous phase yielded 46.0, 10.4 and 50.1 g of residue, respectively. The ethyl acetate fraction (21 g) was subjected to silica gel chromatography with petroleum ether, chloroform-petroleum ether (1:1), chloroform-petroleum ether (9:1) and chloroform as eluants. Chromatography of the (2.9 g) residue obtained from chloroform-petroleum ether (9:1) over silica gel using the same procedure furnished a pure compound (1.4 g) identified as HCH. We compared its physical data as well as its infrared (IR), nuclear magnetic resonance (1H NMR), 13C NMR and mass spectral data with those of an authentic sample.

Synthesis of hydroxychavicol and its analogues (ICB2E, ICB3E, SA4E).  Synthesis of HCH and its analogues was accomplished as described in Data S1.

Cell viability assays.  Cells (1 × 104) in triplicate were incubated in 0.2 mL RPMI-1640-10% FCS containing varying concentrations of compounds (dissolved in absolute ethanol; final concentration of ethanol never exceeded 0.2% v/v in cell culture experiments) or vehicle control (0.2% ethanol v/v). Viability of primary CML cells was determined in the same way except that recombinant human granulocyte-macrophage colony-stimulating factor (rhGM-CSF, 100 ng/mL; R&D Systems, Minneapolis, MN, USA) was included. Cell viability was determined by the trypan blue exclusion assay. At least 200 cells were examined in each sample. In some experiments, in vitro cytotoxicity assays were performed using colorimetric microplate assays with Cell Counting Kit-8 (CCK-8; Dojindo Laboratories, Rockville, MD, USA), which utilizes WST-8,(22) or with MTT tetrazolium salt,(23) where indicated. Results were presented after 72 h of compound exposure. IC50 was calculated from the average of three independent determinations.

Clinical samples.  Fresh peripheral blood samples were donated by three CML patients in a stable phase of the disease admitted to the Institute of Haematology and Transfusion Medicine, Medical College, Kolkata before receiving any treatment. Peripheral blood samples were collected with due approval from the Human Ethics Committee of the respective institutes and all experiments with human blood were conducted in accordance Human Ethics Committee protocol. These three CML patients were Philadelphia chromosome-positive, as determined by bone marrow cytogenetics analysis. Bcr-Abl fusion protein was detectable in these patients by western blot analysis (not shown). Peripheral blood samples were also collected from two healthy donors. Mononuclear cells were separated by Ficoll/Hypaque density gradient centrifugation. Informed consent was provided according to the Declaration of Helsinki.

Immunoblotting.  Immunoblotting experiments were performed on whole cell lysate, cytosolic and mitochondrial fractions of K562 cells, as reported.(11) The mitochondrial and cytoplasmic fractions were separated using the ApoAlert Cell Fractionation Kit (Clontech, Mountain view, CA, USA), following the supplier’s protocol. The anti-COX4 antibody of the kit was utilized as the loading for the mitochondrial fraction. After receiving the indicated treatment, whole cell lysate proteins were separated by SDS–PAGE, transferred to PVDF membrane, and probed with primary antibodies followed by horseradish peroxidase-coupled secondary antibody. Blots were developed using the chemiluminescence method.

Flow cytometry.  Flow cytometry was performed to evaluate apoptosis (after staining with annexin V-FITC-propidium iodide [PI]),(24) mitochondrial membrane potential (after staining with mitochondrial membrane potential sensitive fluorescent dye JC-1),(12) intracellular H2O2 (after staining with DCF-DA), intracellular O2˙− (after staining with DHE), intracellular NO (after staining with DAF-FM) and cell cycle (after permeabilizing cells followed by staining with PI for DNA content analysis).(24) After staining, cells were analyzed in a flow cytometer (BD LSR, Becton Dickinson, San Jose, CA, USA) with Cell Quest Software.

Measurement of intracellular glutathione.  Two independent assays were carried out: a flow cytometry-based assay using a glutathione (GSH)-specific probe thiolyte MBCl(25) and an assay using a GSH assay kit (Cayman, Michigan, USA).

siRNA knockdown.  K562 cells were transfected with control siRNA indicated siRNA (purchased from Santa Cruz Biotechnology) and transfections were carried out following the manufacturer’s instructions. The transfection reagent used for siRNA transfection was purchased from Santa Cruz Biotechnology. The cells were treated with HCH 48 h post-transfection, as indicated.

Giemsa staining for cell morphology.  The control and the HCH-treated cells were centrifuged and smears of the resultant pellet were placed on clean glass slides and air dried. The slides were fixed in methanol then stained with Giemsa stain (Sigma-Aldrich)(12) and observed under a light microscope oil immersion lens (Olympus, CH30). Microscopic photographs were taken with an Olympus CAMEDIA C-4000 Zoom (4.0 megapixel) digital camera.

Confocal microscopy.  Cells exposed to HCH for 24 h were collected by centrifugation, washed with ice-cold PBS and fixed with 4% paraformaldehyde (Calbiochem) for 30 min at room temperature. After permeabilization with 1% Triton X-100 (Sigma-Aldrich), cells were stained with 4′-6-Diamidino-2-phenylindole (DAPI, Calbiochem) for 30 min and examined with a Leica TCS SP2 confocal laser scanning microscope (DMIRB, 40 × objective, Mannheim, Germany).

DNA strand breaks were identified by TdT-mediated TUNEL assay using the ApoAlert DNA Fragmentation Assay Kit (Clontech) following the manufacturer’s protocol. TUNEL positive cells detected by confocal microscopy were considered to be apoptotic cells.

For analysis of cytochrome c release, cells were fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100-PBS and stained with anti-cytochrome c antibody, as reported.(12) Cytochrome c release was imaged using a Leica TCS SP MP confocal microscope with an oil immersion objective.

In vivo experiments.  Xenograft experiments were performed in SCID and Nude mice models (SCID strain-cBySmn). CB17-Prkdcscid/J (from Jackson Laboratory, Bar Harbor, Maine) male mice weighing approximately 20 gm and 6–9 weeks old were used for xenografts of Ba/F3 Bcr-Abl P210wild type and Ba/F3 Bcr-Abl P210T315I. Nude mice were used for K562 xenografts. K562 cells were suspended at a concentration of 5 × 107 cells/mL in Matrigel (1 volume of cells with 1 volume of cold Matrigel). Nude female mice 6–7 weeks of age (National Institute of Nutrition, Hyderabad, India) were injected with 0.2 mL of this suspension. Animals were left untreated until xenografts reached 200–300 mm3. HCH (100 mg/kg) or vehicle control (0.2 mL of 0.2% ethanol per feeding) was administered orally twice a day for 10 days (5 mice per group). Ba/F3 cells expressing wild type P210 Bcr-Abl or expressing imatinib-resistant mutated P210 Bcr-Abl (Ba/F3 Bcr-Abl P210T3151) were suspended in saline and 5–6 × 106 cells were injected subcutaneously in SCID mice. Once the tumors reached a size of 5–7 mm in diameter, animals were randomized (seven per group) and received HCH orally or some of its analogues (100 mg/kg in each case) or vehicle control twice a day for 15 days. Tumor volume (mm3) was estimated according to the formula: {length (mm) × [width (mm)2] × 0.5}. Maximum tolerated dose (without any sign of toxicity and weight loss) was 100 mg/kg body weight for both HCH and ICB3E. Hence, this dose was selected for in vivo experiments with these molecules.

Animal studies were conducted according to institutional Animal Care and Use Committee protocol.

Statistical analysis.  All the data were expressed as mean ± SD of at least three independent experiments, and statistical analysis for single comparison was evaluated by performing Student’s t-test. The criterion for the statistical significance was P < 0.05.

Results

Hydroxychavicol purified from Piper betle leaves or prepared synthetically induces apoptosis in CML cell lines.  Alcoholic extract of Piper betle leaves induced apoptosis of CML cells expressing wild type and mutated Bcr-Abl and showed activity against imatinib-sensitive and imatinib-resistant xenografts.(4,5) HPLC analysis revealed that HCH is the major constituent of the alcoholic extract of Piper betle leaves (Fig. 1a). Therefore, we evaluated whether HCH shows anti-CML activity.

Figure 1.

 Hydroxychavicol (HCH) is the major constituent of the alcoholic extract of Piper betle leaves and induces apoptosis of Bcr-Abl+ CML cell lines. (a) HPLC profile of methanolic extract of Piper betle leaves, which detects several peaks at 254 nm. Hydroxychavicol (54.4%), cis-p-coumaryl quinic acid (10.4%), trans-p-coumaryl quinic acid (4.5%) and chlorogenic acid (6.6%) were identified using standards (samples available in the laboratory), their HPLC retention times being 37.3, 23.2, 19.1 and 13.4 min, respectively. (b) Cell lines were cultured in medium with graded concentrations of HCH for 48 h and viable cells were counted microscopically. The viability of vehicle-treated cells was defined as 100%. Values represent mean ± SD of three experiments. (c) Flow cytometric analysis of annexin V binding after treatment with HCH for 18 h. (d) Cell cycle analysis of K562 cells after treatment with HCH (5.0 μg/mL). Gates were set to determine the percentage of dead (sub-G0/G1), G0/G1, S and G2/M phase cells. (e) DNA fragmentation analysis after staining with Giemsa (top panel), DAPI (middle panel) or TUNEL (bottom panes). (f) reactive oxygen species-dependent cleavage of caspase-3, caspase-9 and poly-adenosine diphosphate-ribose polymerase (PARP) but not caspase-8 in HCH-treated K562 cells. K562 cells were treated with N-acetylcysteine (NAC) (5 mM) for 1 h before treatment with HCH (5.0 μg/mL) for 18 h. Immunoblotting was performed on whole cell lysates with the indicated antibodies. (g) Inhibitors of caspase-3 and capase-9 reverse HCH-mediated cell death. K562 cells were preincubated with indicated caspase inhibitor (25 μM each) for 1 h before the addition of HCH (5.0 μg/mL) for 18 h for measurement of cell viability by Trypan blue dye exclusion. Values represent mean ± SD of three experiments.

Purified HCH was examined on a number of CML cell lines for cytotoxicity. Pure HCH decreased the viability of CML cell lines in a dose dependent manner (results not shown). To rule out the possible contribution of contaminants in purified HCH-mediated cytotoxicity in vitro, HCH and its analogues were synthesized (see Data S1). Synthetic HCH (ICB2) also showed anti-CML activity in vitro in a dose-dependent manner on three CML cell lines tested (K562, KU812, KCL22) (Fig. 1b). We next evaluated whether the HCH-induced cytotoxic effect on CML cells was due to apoptosis. HCH-induced apoptosis in K562 cells was confirmed by annexin V and PI staining (Fig. 1c), appearance of apoptotic nuclei (sub G0/G1 peak in DNA cell cycle analysis, Fig. 1d), nuclear fragmentation analyzed by staining with Giemsa (Fig. 1e, top panel), DAPI (Fig. 1e, middle panel) and also by TUNEL assay (Fig. 1e, bottom panel). Treatment with HCH induced apoptosis in both the CML cell lines tested, whereas it had no appreciable effects on the mononuclear cells of a healthy donor (Fig. 1c). Arrest at the G2M phase followed by the appearance of sub-G0/G1 peak was evident in K562 cells but not in normal PBMC after treatment with HCH (Fig. 1d). Activation of caspase cascade plays an important role in drug-mediated apoptosis.(26) We investigated whether treatment with HCH leads to caspase activation and whether it is a consequence of HCH-induced ROS generation. HCH induced cleavage of caspase-9, caspase-3 and degradation of PARP but had no effects on caspase-8 activation (Fig. 1f). HCH-induced cleavage of caspase-9 and caspase-3 and degradation of PARP were inhibited by pre-treatment with NAC (Fig. 1f). These results ruled out the involvement of the extrinsic pathway in HCH-mediated apoptosis and suggest that HCH-induced ROS generation plays a key role in caspase activation. Evaluation of the effect of different caspase inhibitors on HCH-induced apoptosis also ruled out the role of caspase-8. K562 cells were treated with HCH either alone or in combination with Z-VAD-FMK (pan-caspase inhibitor), Z-IETD-FMK (caspase-8 inhibitor) or LEHD-CHO (caspase-9 inhibitor).(12) Z-VAD-FMK or LEHD-CHO treatment reversed the apoptosis almost completely; however, Z-IETD-FMK had no effect on HCH-mediated cell death in K562 cells (Fig. 1g). These results further support that caspase-8 might not play a role in induction of apoptosis by HCH.

Hydroxychavicol induces transient increase of mitochondrially derived reactive oxygen species followed by persistent increase of nitric oxide.  Inhibition of Bcr-Abl tyrosine kinase activity is known to result in apoptosis of Bcr-Abl-positive cells. Therefore, we investigated whether HCH-induced apoptosis of CML cells correlates with inhibition of Bcr-Abl phosphorylation. HCH did not significantly inhibit autophosphorylation and protein expression of Bcr-Abl in K562 cells (results not shown). The role of ROS and reactive nitrogen species in the effect of HCH on CML cells was then investigated. To establish whether HCH treatment promoted ROS production, the intracellular level of ROS (both O2˙− and H2O2) and NO were quantified by flow cytometry using specific fluorescent probes (DHE for O2˙−; DCF-DA for H2O2; DAF-FM for NO). When K562 cells were treated with HCH there was an early accumulation of intracellular O2˙− and H2O2. The level of both O2˙− and H2O2 began to rise significantly above the basal level within 15 min, and peaked by 15 min to 1-h post-treatment. The accumulated levels of O2˙− and H2O2 were reduced thereafter and came down to basal levels or below by 3-h post-treatment. Data representing time kinetics of O2˙− and H2O2 accumulation in K562 cells are shown in Figure 2(a). In contrast to cancer cell lines, normal human peripheral blood mononuclear cells (hPBMC) had only marginal increases in intracellular O2˙− and H2O2 under the same experimental conditions (Fig. 2a). As the basal threshold of intracellular ROS was significantly higher in K562 cells compared to normal hPBMC, higher accumulation of intracellular ROS was observed in K562 cells than normal hPBMC after HCH treatment. Accumulation of intracellular NO was detectable in K562 cells after 1-h post-treatment with HCH and significantly increased thereafter throughout the incubation period (Fig. 2a). However, normal hPBMC had only marginal increases in intracellular NO after treatment with HCH (Fig. 2a). Representative histograms of DHE, DCF-DA and DAF-FM fluorescence in K562 cells and normal hPBMC after treatment with HCH are shown in Figure 2(b).

Figure 2.

 Hydroxychavicol (HCH)-induced production of nitric oxide (NO) is redox-dependent. (a) Time-dependent accumulation of superoxide, H2O2 and NO by HCH (5.0 μg/mL) in indicated cells (mean ± SD of three experiments; *P < 0.05; **P < 0.01). (b) Representative histograms of HCH-induced superoxide, H2O2 and NO production at indicated time points. Solid lines indicate staining after incubation with vehicle control. Dotted lines indicate staining after incubation with HCH (5.0 μg/mL). Values within the histograms represent the specific mean fluorescence intensity (MFI, after substracting the respective vehicle control values). The histograms are representative of three similar experiments. (c) K562 cells were pretreated with N-acetylcysteine (NAC) (5 mM), Polyethylene glycol-conjugated catalase (Peg-Cat) (200 U/mL), Peg-conjugated superoxide dismutase (Peg-SOD) (200 U/mL) for 1 h followed by treatment with vehicle control or HCH (5.0 μg/mL) for 18 h. Intracellular NO was analyzed by flow cytometry after staining with DAF-FM. Representative histograms of three similar experiments. (d) Inhibitor of SOD or catalase enhances HCH-induced NO production. K562 cells were pretreated for 1 h with SOD inhibitor diethyldithiocarbamate (DETC) (50 μM) or catalase inhibitor and catalase inhibitor aminotriazole (ATZ) (2 mM) before treatment with vehicle control or HCH (5.0 μg/mL) for 18 h. Representative histograms of two similar experiments. (e) K562 cells were pretreated with NAC (5 mM), Peg-Cat (200 U/mL), Peg-SOD (200 U/mL) for 1 h followed by treatment with vehicle control or HCH (5.0 μg/mL) for 18 h. Apoptosis was analyzed by annexin V/PI binding assays. Representative dot plots of three similar experiments. (f) Pretreatment with NAC or Peg-Cat but not Peg-SOD inhibits HCH-induced H2O2 production. K562 cells were pretreated with NAC (5 mM), Peg-Cat (200 U/mL) or Peg-SOD (200 U/ml) for 1 h followed by treatment with vehicle control or HCH (5.0 μg/mL) for 1 h before staining for intracellular H2O2. Representative histograms of four similar experiments. (g) HCH-induced apoptosis is reversed by pretreatment with NO scavenger CPTIO. K562 cells were pretreated with CPTIO (25 μM) for 1 h followed by treatment with vehicle control or HCH for 18 h before analysis by flow cytometry. Representative dot plots of two similar experiments. (h) Inhibitor of SOD or catalase enhances cytotoxic activity of HCH. K562 cells were treated with ATZ, DETC and HCH either alone or in combination as indicated for 18 h. Cytotoxicity was measured by MTT assay. C-PTIO, 2-4-carboxyphenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide; DAF-FM, 4-amino-5-methylamino-2′,7′-difluorofluorescein; DCF-DA, 2′,7′-dichlorofluorescein diacetate; DHE, dihydroethidium; hPBMC, human peripheral blood mononuclear cells.

Reactive oxygen species can be generated through several mechanisms, the major sources being oxidase activity or the mitochondrial electron transport chain. We used a mitochondria-specific superoxide-sensitive probe MitoSOX to determine whether HCH promoted ROS production from this source. Significant increase in mitochondrial superoxide was observed after HCH treatment (Fig. S1a). To rule out the possibility of oxidases as the source of HCH-induced ROS generation, known inhibitors of some oxidases were also employed. We used NADPH oxidase inhibitors DPI(27) and apocynin(28) and xanthine oxidase inhibitor allopurinol.(29) HCH-induced superoxide production in K562 cells was unaffected by these inhibitors, whereas the mitochondrial poison rotenone augmented superoxide production (Fig. S1b). These data indicated that HCH-induced superoxide production in K562 cells, detected by DHE, came from mitochondria.

Reactive oxygen species-dependent nitric oxide-mediated damage is crucial for hydroxychavicol-induced cell death.  Hydroxychavicol induced transient early accumulation of mitochondrially derived ROS followed by persistent increase of NO in CML cells (Fig. 2a). Therefore, we investigated whether HCH-induced NO production and cell death were redox-dependent. HCH-induced accumulation of intracellular NO in K562 cells was markedly reduced by pretreatment with NAC or Peg-Cat but not with Peg-SOD (Fig. 2c). C-PTIO, a known scavenger of NO, was used as control.(30) Taken together, these data demonstrate that early accumulation of mitochondrial ROS plays a key role in HCH-induced accumulation of NO.

We then determined whether HCH-induced superoxide or hydrogen peroxide or both were capable of NO generation. DETC, a known inhibitor of SOD, further enhanced NO generation in K562 cells when cotreated with HCH (Fig. 2d). Similarly, ATZ, a known inhibitor of catalase when cotreated with HCH in the presence of Peg-SOD, also enhanced NO generation in K562 cells (Fig. 2d). Inhibition of SOD by DETC and catalase by ATZ were confirmed by accumulation of superoxide and hydrogen peroxide, respectively (Fig. S2). Therefore, both species of ROS (i.e. superoxide and hydrogenperoxide) are capable of inducing NO generation in K562 cells.

Pretreatment with NAC or Peg-cat reversed HCH-induced death of K562 cells (Fig. 2e) and inhibited intracellular ROS (Fig. 2f), while pre-treatment with Peg-SOD had no effect. Experiments were also performed to rule out the possibility that NAC acts directly with HCH in solution, thereby neutralizing this agent so that it could not react with cells. HCH was incubated with NAC and then analyzed by HPLC. Results of this analysis indicated that NAC failed to react with HCH (Fig. S3). Pretreatment with NO-scavenger C-PTIO also reversed the cell death process (Fig. 2g). To maximally exploit ROS-mediated cell death, HCH was combined with compounds that suppress the cellular antioxidant capacity, such as SOD inhibitor DETC and catalase inhibitor ATZ. These compounds drastically enhanced HCH-mediated killing of CML cells (Fig. 2h).

The role of NO in HCH-induced apoptosis of CML cells was supported by the decrease of mitochondrial membrane potential (Fig. 3a), release of cytochorme c (Fig. 3b,c) and cell cycle arrest at the G2M phase (Fig. 3d), which were largely attenuated by pretreatment with C-PTIO or NAC.

Figure 3.

 Hydroxychavicol (HCH)-induced apoptosis is mediated via mitochondrial pathway. (a) Flow cytometric analysis of mitochondrial membrane potential of K562 cells after treatment with HCH (5.0 μg/mL) for 18 h in the presence and absence of 2-(4carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3oxide (c-PTIO) followed by staining with JC-1. (b) K562 cells were treated with HCH (5.0 μg/mL) for 6 h, fixed, permeabilized, stained with anti-cytochrome c antibody and analyzed by confocal microscopy. (c) Immunoblot-based analysis of the expression of cytochrome c in the cytosolic and mitochondrial fractions of K562 cells after indicated treatment for 6 h. (d) Cell cycle analysis of K562 cells after indicated treatment for 12 h.

Taken together, ROS-dependent NO plays a crucial role in HCH-mediated apoptosis of CML cells.

Hydroxychavicol does not cause glutathione depletion.  Because early accumulation of ROS played a key role in HCH-mediated cytotoxicity, we evaluated the intracellular level of GSH given that GSH is a major cellular antioxidant. Interestingly, HCH treatment did not deplete intracellular GSH in K562 cells, as evident from the application of two independent techniques (Fig. S4a,b). l-Buthionine sulphoximine (BSO), which depletes intracellular GSH, was used as a positive control.

Endothelial nitric oxide synthase is responsible for hydroxychavicol-induced nitric oxide production and cell death.  The source of redox-dependent NO that was responsible for HCH-mediated cell death was investigated using two approaches: (i) employing pharmacological inhibitors of NOS isoforms and (ii) knocking down NOS isoforms using siRNA. Pharmacological inhibitors that preferentially inhibit different isoforms of NOS were preincubated with K562 cells for 1 h before treatment with HCH. Evaluation of apoptosis by annexin V/PI binding assays indicated that L-NIO, a preferential inhibitor of eNOS(31) but not SMTC or AG, preferential inhibitors of neuronal NOS (nNOS)(32) or inducible NOS (iNOS),(33) respectively, reversed significantly HCH-induced cell death (Fig. 4a). HCH-induced enhanced activatory phosphorylation of eNOS was also observed in a time-dependent manner (Fig. 4b).

Figure 4.

 Involvement of endothelial nitric oxide synthase (eNOS) in hydroxychavicol (HCH)-induced apoptosis and nitric oxide (NO) production in K562 cells. (a) Reversal of HCH-induced apoptosis of K562 cells by pharmacological inhibitor of eNOS. Cells were pretreated for 1 h with pharmacological inhibitors of inducible nitric oxide synthase (iNOS) (AG, 250 μM), eNOS (L-NIO, 50 μM) or neuronal nitric oxide synthase (nNOS) (S-methyl-l-thiocitrulline [SMTC], 50 μM) followed by treatment with vehicle control or HCH (2.5 μg/mL) for 18 h. Apoptosis was analyzed by annexin V/PI binding assays. Representative dot plots of three experiments are shown. (b) Activatory phosphorylation and protein expression of eNOS in K562 cells by HCH (2.5 μg/mL) as determined by western blot analysis. (c) Effect of transfection with NOS siRNA on HCH-induced apoptosis of K562 cells. Cells were transfected with control siRNA, nNOS siRNA, iNOS siRNA or eNOS siRNA for 48 h then treated with vehicle control or HCH (5.0 μg/ml) for 18 h. Apoptosis was analyzed by flow cytometry. Representative dot plots of three experiments are shown. (d) Attenuation of eNOS protein expression in K562 cells by transfection with eNOS siRNA. Respective NOS proteins were analyzed by western blot. Representative immunoblots of three experiments are shown. (e) Attenuation of HCH-induced NO production in K562 cells by transfection with eNOS siRNA. Intracellular NO was analyzed by flow cytometry after staining with 4-amino-5-methylamino-2′, 7′-difluorofluorescein diacetate (DAF-FM). Representative histograms of three experiments are shown.

Additional experiments were performed where K562 cells were transfected with control siRNA, eNOS siRNA, nNOS siRNA or iNOS siRNA and then treated with HCH. Transfection with control siRNA, eNOS siRNA, nNOS siRNA or iNOS siRNA had no effect on the viability of cells (Fig. 4c). However, HCH-induced death of K562 cells was markedly reversed by transfection with eNOS siRNA but not with nNOS siRNA, iNOS siRNA or control siRNA (Fig. 4c). Knockdown of respective NOS isoforms in K562 cells after transfection with siRNA have been confirmed (Fig. 4d). Transfection with eNOS siRNA but not with control siRNA, iNOS siRNA or nNOS siRNA also brought down HCH-induced enhanced NO production in K562 cells (Fig. 4e).

Reactive oxygen species-mediated JNK activation plays a key role in hydroxychavicol-induced nitric oxide production and cell death.  Oxidative stress activates the JNK pathway.(34) Activation of the JNK pathway leads to apoptosis.(35) HCH induced persistent phosphorylation of JNK, which was detectable within 4 h, and lasted throughout the incubation period (Fig. 5a). However, phosphorylation of ERK or p38 was not enhanced by HCH treatment (Fig. 5a). NAC attenuates HCH-induced JNK and e-NOS phosphorylation (Fig. 5b). However, knocking down JNK by siRNA transfection had no effect on HCH-induced ROS generation (Fig. 5c). Knockdown of JNK1 in K562 cells after transfection with siRNA and attenuation of HCH-induced upregulation of JNK phosphorylation after transfection with JNK1 siRNA have been confirmed (Fig. 5d). A little increase in phosphorylation of JNK, even in the siRNA treated cells in Fig. 4d, could be explained by the presence of a moderate amount of JNK. Thus, ROS is responsible for HCH-mediated JNK activation. Of note, knocking down JNK inhibited HCH mediated NO generation (Fig. 5e) and apoptosis of K562 cells (Fig. 5f).

Figure 5.

 Contribution of reactive oxygen species (ROS)-mediated JNK activation in hydroxychavicol (HCH)-induced NO production and apoptosis. (a) Immunoblot-based determination of JNK, p38 and extracellular signal related kinase (ERK) phosphorylation status and protein expression in K562 cells after treatment with HCH. (b) Immunoblot-based determination of HCH-mediated JNK and endothelial nitric oxide synthase (eNOS) phosphorylation. K562 cells were pretreated with N-acetylcysteine (NAC) for 1 h, then with HCH for 6 h. Western blot was performed with the whole cell extract after immunoblotting with indicated antibodies. (c) Knocking down JNKI has no effect on HCH-induced ROS production. Flowcytometry based determination of intracellular H2O2 after transfecting K562 cells with indicated siRNA followed by treatment with HCH. (d) Attenuation of JNKI protein expression in K562 cells after transfection with JNKI siRNA. JNKI protein expression and phosphorylation were analyzed by western blot. (e) Attenuation of HCH-induced NO production after transfection of K562 cells with JNKI siRNA followed by staining with 4-amino-5-methylamino-2′, 7′-difluorofluorescein diacetate (DAF-FM) and analysis in a flow cytometer. (f) Inhibition of HCH-mediated apoptosis of K562 cells after transfection with JNKI siRNA. Apoptosis was determined by flow cytometry after Annexin V-FITC/PI staining. Mean ± SD of three experiments. (g) Knockdown of JNK1 attenuates HCH-induced eNOS phosphorylation. Protein expression and phosphorylation were analyzed by western blot. (h) Knockdown of eNOS marginally attenuates HCH-induced JNK phosphorylation. DCF-DA, 2′,7′-dichlorofluorescein diacetate.

We next evaluated whether there is crosstalk between JNK and eNOS. For this, phosphorylation and protein expression of eNOS was studied after knocking down JNK1 followed by HCH treatment. Our data, presented in Figure 5(g), indicate that JNK plays an important role in phosphorylating eNOS but does not affect eNOS expression. Additional experiments indicated that knockdown of eNOS inhibited HCH-induced JNK phosphorylation only marginally and had no effect on JNK1 protein expression (Fig. 5h). Therefore, our data suggest that eNOS activation is strongly regulated by JNK, whereas eNOS-mediated JNK regulation is relatively weak.

Similarly, JNK inhibitor SP600125 reduced HCH-induced NO production (Fig. S5a) and cell death (Fig. S5b).(36) In contrast, p38 inhibitor SB203580 or ERK inhibitor PD98059 had no effect on HCH-induced NO production (results not shown) or cell death (Fig. S5b).(24) These data indicate that JNK activation was crucial for HCH-induced NO production and cell death in K562 cells.

Hydroxychavicol and its analogues are effective against primary cells from CML patients and leukemic cells expressing mutated Bcr-Abl.  Hydroxychavicol and three analogues of HCH (ICB2E, ICB3E and SA4E) showed remarkable cytotoxicity on leukemic cell lines expressing wild type and mutated Bcr-Abl (Table 1). Even the predominant mutation T315I, which is virtually unresponsive to existing drugs including dasatinib, nilotinib(37), bosutinib(38) and bafetinib,(39) remains sensitive to these molecules. The resistant index of the mutated cell lines presented in Table 1 against imatinib has been reported earlier and varied between 1.78 to 23.11 depending on the cell lines.(4,40) HCH also showed anti-CML activity in a dose-dependent manner against primary cells from three CML patients tested (Fig. 6a). In contrast, HCH was minimally toxic on unfractionated normal human peripheral blood mononuclear cells (Fig. 6b). The role of ROS in the observed anti-CML activity of HCH was confirmed in vivo in K562 xenografts in nude mice (Fig. 6c). One esterified HCH analogue (ICB3E) was evaluated against xenografts of leukemic cells expressing wild type and mutated Bcr-Abl (T315I) of imatinib-resistance phenotype. The esterified analogue has enhanced stability towards air from autooxidation. An oral dose of 100 mg/kg twice daily for 5–7 days inhibited tumor growth in subcutaneous xenograft models with BaF3 cells transfected with either BCR-ABL wild type or T315I (Fig. 6d) without adverse effects on bodyweight (Fig. S6).

Table 1.   Anti-proliferative activity of HCH and some of its analogues on CML cell lines expressing wild type and mutated Bcr-Abl
Cell linesIC50 μM*
HCHICB2EICB3ESA4E
  1. *IC50 was calculated using Cell Counting Kit–8 following manufacturer’s protocol. Standard deviation was <20% in each case. T315I indicates threonine 315 to isoleucine; H396P, histidine 396 to proline; H396R, histidine 396 to arginine; E255K, glutamic acid 255 to lysine; E255V, glutamic acid 255 to valine; M244V, methionine 244 to valine; F359V, phenylalanine 359 to valine; M315T, methionine to threonine; Q252H, glutamine 252 to histidine and Y253F, tyrosine 253 to phenylalanine.

Ba/F3P210 WT8.88.95.310.3
Ba/F3P210T315I5.43.93.49.9
Ba/F3P210H396P4.26.25.98.9
Ba/F3P210H396R5.77.26.38.7
Ba/F3P210E255K27.3121232.4
Ba/F3P210E255V5.68.73.710.5
Ba/F3P210M244V6.89.15.210.1
Ba/F3P210F359V5.56.15.714.9
Ba/F3P210M351T8.36.28.89.8
Ba/F3P210Q252H4.55.61123.3
Ba/F3P210Y253F4.94.48.810
Figure 6.

 Hydroxychavicol (HCH) or its analogue is effective against primary cells from chronic myeloid leukemia (CML) patients in vitro and imatinib resistant CML xenografts in vivo. (a) Mononuclear cells from three Bcr-Abl-positive CML patients were cultured in medium containing rhGM-CSF with vehicle control (◆) or with the same medium containing HCH (△, 0.5 μg/mL; □, 1.0 μg/mL; ▮, 2.0 μg/mL; ○, 3.0 μg/mL; bsl00066, 5.0 μg/mL). (b) Normal human peripheral blood mononuclear cells (hPBMC) from two donors were cultured in medium containing vehicle control (◆) or with HCH (□, 2.5 μg/mL; bsl00066, 5.0 μg/mL). (c) N-acetylcysteine (NAC) treatment abrogates the therapeutic effect of HCH on K562 xenografts; nude mice bearing K562 xenografts received NAC (150 mg/kg) via i.p. route three times a week for 10 days. HCH (100 mg/kg) was fed orally twice a day for 10 days. Mean (± SD) growth curves are shown for groups of three mice in each case. (d) In vivo therapeutic activity of HCH analogue (ICB3E) on imatinib-resistant Bcr-Abl+ cell line xenografts. SCID mice bearing either Ba/F3 wild type Bcr-Ablp210 or Ba/F3 T3151 Bcr-Ablp210 xenografts were fed orally the indicated doses of ICB3E twice daily. Mean growth curves are shown for groups of five mice in each case.

Discussion

Cancer cells including CML cells have higher oxidative stress than their normal counterparts. Recent studies suggest that the intrinsic higher oxidative stress in cancer cells might be exploited to preferentially kill these cells in vitro by agents that induce intracellular ROS accumulation.(23) Recent studies suggest that the alcoholic extract of Piper betle leaves induces apoptosis of CML cells expressing wild type and mutated Bcr-Abl and shows activity against imatinib-sensitive and imatinib-resistant CML xenografts.(4,5) In the present study we show that HCH is the major constituent of the alcoholic extract of the Piper betle leaves and this compound might contribute, at least in part, to the observed anti-CML activity of Piper betle leaf extract. The role of contaminants in the observed anti-CML activity of purified HCH was ruled out by synthesizing the molecule and a number of analogues of the molecule, all of which have comparable activity. HCH is known to induce cell cycle arrest and apoptosis in the oral KB carcinoma cell line and in hepato-carcinoma cells through induction of oxidative stress.(9) Previous studies suggest anti-mutagenic and chemopreventive(7) properties of HCH. Using leukemic cell lines expressing wild type and mutated Bcr-Abl, we show that these cells are susceptible to HCH and its analogue-mediated death both in vitro and in vivo.

Emergence of imatinib resistance has compromised the success of imatinib, a revolutionary discovery in CML treatment. Point mutations in the kinase domain of Bcr-Abl are the major cause of imatinib resistance. So far, 90 point mutations in the Bcr-Abl kinase domain have been identified in imatinib-resistant CML patients. Although second generation Bcr-Abl inhibitors are effective against most mutations in Bcr-Abl, these drugs remain ineffective against the frequently observed mutation T315I.(41) Thus, there is a global need to develop therapeutics against the T315I mutation of Bcr-Abl. Several approaches, such as a combination of histone deacetylase inhibitors with agents that disrupt autophagy(42) or use of oxidative stress(43) have achieved appreciable success in overcoming imatinib resistance. In the present study, we show that cells expressing imatinib-resistant mutated Bcr-Abl (P210Bcr-AblT315I) are susceptible to one analogue of HCH (ICB3E) in vitro and in vivo.

Early and transient accumulation of mitochondrial ROS followed by persistent accumulation of NO is associated with HCH-mediated CML cell death.

As the basal threshold of intracellular ROS was significantly lower in normal human mononuclear cells compared to CML cell line K562, much lower accumulation of intracellular ROS was observed in mononuclear cells than K562 after HCH treatment. The mechanism by which HCH induces higher ROS in cancer cells compared to normal cells is not clear. HCH might interfere with ROS homeostatic regulators. Such regulators like glutathione S-transferase pi1 and carbonyl reductase 1 have been reported to be involved in piperlongumine-mediated selective induction of ROS in cancer cells.(44) High levels of intracellular NO cause S-nitrosylation of a number of proteins, leading to an unfolded protein response, which, in turn, might cause apoptosis.(45) HCH-induced NO was prevented by pretreatment with ROS scavengers, suggesting that HCH-mediated NO production was redox-dependent. HCH-mediated cell death was also prevented by scavenging either intracellular ROS or intracellular NO. Thus, mitochondrial ROS-dependent NO plays a crucial role in HCH-mediated CML cell death. Experiments with pharmacological inhibitors and siRNA identified eNOS as the producer of HCH-mediated NO.

The stress-activated protein kinase (SAPK/JNK) pathway plays an important role in cell death response through various stimuli, including oxidative stress. We observed that treatment with HCH caused sustained activation of JNK. Activation of JNK played an important role in both HCH-induced NO production and cell death given that inhibition of expression of JNK by siRNA and use of the pharmacological inhibitor of JNK suppressed both the processes.

Induction of oxidative stress as a crucial step for cancer cell death has been demonstrated for many molecules, including β-phenylethyl isothiocyanate, a natural compound found in cruciferous vegetables.(23) Several such agents are in clinical trial for a number of cancers.(10) In agreement with our finding, a number of ROS generating molecules also showed anti-CML activity.(12,23,43) However, what makes HCH unique in the list of ROS-inducing natural products with anticancer properties is that GSH depletion was not observed after HCH treatment. Glutamate cysteine ligase (GCL) is the rate-limiting enzyme involved in denovo GSH biosynthesis. Nitric oxide is a reported inducer of GLC synthesis.(46) HCH-induced ROS directly activated JNK. JNK activation, in turn, phosphorylated eNOS, leading to enhanced NO production, which might consequently prevent GSH depletion. In most previous studies where ROS-inducing agents induced apoptosis, experiments were restricted to reversal of apoptotic events by ROS scavengers. Therefore, further studies with other ROS-inducing agents are required to understand whether ROS-dependent JNK phosphorylation leading to eNOS-mediated NO production is a specific event for HCH.

Our data indicate that the apoptosis of CML cells mediated by HCH is through eNOS. HCH-induced ROS phosphorylates JNK, which, in turn, phosphorylates eNOS to produce NO, leading to apoptosis. JNK might directly phosphorylate eNOS. Alternatively, JNK might phosphorylate eNOS via some intermediate signaling molecules. The crosstalk between JNK and eNOS was also evaluated. Our data suggest that eNOS phosphorylation is strongly regulated by JNK, whereas eNOS-mediated JNK activation is relatively weak.

To maximally exploit the ROS-dependent cell death by HCH, a combination strategy was evaluated with compounds that either inhibit SOD or catalase. These compounds significantly enhanced the cytotoxic potential of HCH, probably by crippling the cellular antioxidant capacity.

In conclusion, our data suggest that HCH, a relatively non-toxic edible herb-derived natural product or its analogues has potent anti-CML activity both in vitro and in vivo. This is exemplified by its action against so far untreatable CML bearing T315I-Bcr-Abl, suggesting its potential for treating drug resistant CML, either alone or in combination with other Bcr-Abl inhibitors. Anticancer activity of HCH is mediated by the mitochondrial ROS-dependent eNOS-mediated pathway. ROS-dependent JNK activation followed by eNOS-mediated NO production plays a crucial role in the HCH-mediated killing of CML cells. Interestingly, unlike other oxidative stress inducers, HCH-mediated killing of CML cells does not depend on GSH depletion. The unique mode of action of HCH merits further testing of this molecule alone or in combination with other anticancer drugs in preclinical and clinical settings.

Acknowledgments

This work was supported by the Council of Scientific and Industrial Research (Interagency project IAP 0001) and the Department of Biotechnology (GAP 235), New Delhi, India. Part of the information provided in this report is protected by pending US & Patent Cooperation Treaty patent applications.

Disclosure Statement

The authors have no conflict of interest.

Ancillary