Biochemical Journal

Research article

Adaphostin promotes caffeine-evoked autocrine Fas-mediated death pathway activation in Bcr/Abl-positive leukaemia cells

Wen-Hsin Liu , Long-Sen Chang

Abstract

The present study was conducted to verify whether caffeine is beneficial for improving leukaemia therapy. Co-treatment with adaphostin (a Bcr/Abl inhibitor) was found to potentiate caffeine-induced Fas/FasL up-regulation. Although adaphostin did not elicit ASK1 (apoptosis signal-regulating kinase 1)-mediated phosphorylation of p38 MAPK (mitogen-activated protein kinase) and JNK (c-Jun N-terminal kinase), co-treatment with adaphostin notably increased p38 MAPK/JNK activation in caffeine-treated cells. Suppression of p38 MAPK and JNK abrogated Fas/FasL up-regulation in caffeine- and caffeine/adaphostin-treated cells. Compared with caffeine, adaphostin markedly suppressed Akt/ERK (extracellular-signal-regulated kinase)-mediated MKP-1 (MAPK phosphatase 1) protein expression in K562 cells. MKP-1 down-regulation eventually elucidated the enhanced effect of adaphostin on p38 MAPK/JNK activation and subsequent Fas/FasL up-regulation in caffeine-treated cells. Knockdown of p38α MAPK and JNK1, ATF-2 (activating transcription factor 2) and c-Jun by siRNA (small interfering RNA) proved that p38α MAPK/ATF-2 and JNK1/c-Jun pathways were responsible for caffeine-evoked Fas/FasL up-regulation. Moreover, Ca2+ and ROS (reactive oxygen species) were demonstrated to be responsible for ASK1 activation and Akt/ERK inactivation respectively in caffeine- and caffeine/adaphostin-treated cells. Likewise, adaphostin functionally enhanced caffeine-induced Fas/FasL up-regulation in leukaemia cells that expressed Bcr/Abl. Taken together, the results of the present study suggest a therapeutic strategy in improving the efficacy of adaphostin via Fas-mediated death pathway activation in Bcr/Abl-positive leukaemia.

  • activating transcription factor 2 (ATF-2)
  • adaphostin
  • caffeine
  • c-Jun
  • c-Jun N-terminal kinase 1 (JNK1)
  • Fas/FasL
  • p38α mitogen-activated protein kinase (p38α MAPK)

INTRODUCTION

Caffeine (1,3,7-trimethylxanthine) is a purine alkaloid present in many popular beverages including cocoa, tea and coffee. Caffeine is reported to affect cell-cycle function and induce apoptosis in pancreatic cancer [1], neuroblastoma [2] and hepatocellular carcinoma cells [3]. Moreover, caffeine and other methylxanthines were found to inhibit tumour cell invasiveness and experimental metastasis [4,5]. Consistently, caffeine induces MMP (matrix metalloproteinase)-2 and MMP-9 down-regulation in human acute myeloid leukaemia U937 cells [6]. The signalling pathways, including ATM (ataxia telangiectasia mutated)-, p53- or p21-activated protein kinase 2-dependent pathways, are suggested to be responsible for caffeine-induced cell proliferation and apoptosis [710], whereas caffeine suppresses metastasis in a transgenic mouse model via the ATM-independent pathway [5]. ERK (extracellular-signal-regulated kinase)- and Akt-mediated pathway inactivation is involved in the caffeine-induced death of human osteoblasts [10]. Caffeine-elicited activation of the PI3K (phosphoinositide 3-kinase)/Akt pathway provides a cytoprotective effect in human neuroblastoma SH-SY5Y cells [11]. Although caffeine inhibits the proliferation of hepatocellular carcinoma accompanied with activation of ERK and p38 MAPK (mitogen-activated protein kinase), abolishment of ERK and p38 MAPK activation cannot restore the viability of caffeine-treated hepatocellular carcinoma [3]. Caffeine-induced ERK inactivation and p38 MAPK activation are responsible for MMP-2 and MMP-9 down-regulation in U937 cells, whereas caffeine treatment marginally reduced the viability of U937 cells [6]. These observations suggest that caffeine-evoked signalling pathways are cell-type-specific and induce distinct events in different cells. Previous studies have reported that caffeine dramatically enhances the tumoricidal effect of several antitumour drugs, such as cisplatin, thiotepa, doxorubicin, cyclophosphamide, mitomycin C, vincristine and methotrexate [1218]. Caffeine co-administration significantly increases the efficacy of chemotherapeutic drugs in mice and patients with osteosarcoma [18,19]. Apparently, the combination of caffeine with chemotherapeutic drugs could be beneficial for cancer treatment. However, these studies unambiguously delineate the caffeine molecular mechanism in enhancing the therapeutic efficacy of antitumour drugs.

CML (chronic myeloid leukaemia) is characterized by the Philadelphia chromosome, which results from a reciprocal translocation between chromosome 9 and chromosome 22 [20]. This mutant gene encodes the constitutively active Bcr/Abl tyrosine kinase, which signals downstream to a variety of cytoprotective pathways, including ERK, Akt, NF-κB (nuclear factor κB) and JAK (Janus kinase)/STAT (signal transducer and activator of transcription), that collectively provide proliferative advantages and resistance to apoptosis [21,22]. According to its role in malignant transformation, Bcr/Abl has served as a target for therapeutic intervention in CML. Several Bcr/Abl inhibitors, including imatinib, nilotinib, dasatinib, adaphostin and AG957, have been found to effectively inhibit the growth of CML [23,24]. Noticeably, resistance to imatinib has been reported and is often attributed to the emergence of clones expressing mutant forms of Bcr/Abl or overexpression of molecules downstream of Bcr/Abl signalling such as Lyn kinase [25]. Thus more potent tyrosine kinase inhibitors, nilotinib and dasatinib, have been developed against an array of imatinib-resistant Bcr/Abl mutants [24]. Dasatinib also inhibits members of the Src family of kinases, but nilotinib does not. Nevertheless, the T315I mutant of Bcr/Abl confers resistance to nilotinib and dasatinib [26]. On the other hand, the cytotoxic mechanism of adaphostin and AG957 is verified to be mediated through inducing the degradation of wild-type and mutated Bcr/Abl [23,27]. Although it has been suggested that targeted therapy on Bcr/Abl tyrosine kinase can improve the efficacy of leukaemia therapy, blockading the cytoprotective pathway does not always succeed in treating CML [28,29]. Thus exploring the therapeutic potential of Bcr/Abl inhibitors in combination with other reagents, including biological response modifiers and apoptosis-inducing molecules, may provide a new modality for improving leukaemia therapy. Noticeably, increases in the [Ca2+]i (intracellular Ca2+ concentration) and the production of ROS (reactive oxygen species) are recognized as signalling modifiers in inducing cell death [30,31]. Unlike imatinib, nilotinib and dasatinib, adaphostin induces ROS generation, leading increasingly to inducing the death of leukaemia cells [27]. Caffeine is well known to induce an increase in [Ca2+]i in various cells [32]. Thus, in addition to Bcr/Abl tyrosine kinase suppression, a combination of caffeine and adaphostin should elicit other death signalling pathways in CML. In the present study, the cytotoxicity of a combination of adaphostin and caffeine on CML K562 cells was examined. It was found that caffeine modestly up-regulated Fas and FasL in K562 cells, whereas co-treatment with adaphostin potentiated caffeine-induced up-regulation of Fas and FasL. As a result, in addition to suppressing Bcr/Abl tyrosine kinase, co-treatment with caffeine and adaphostin induced the death of K562 cells through additionally eliciting an autocrine Fas-mediated death pathway. Likewise, adaphostin promoted caffeine-evoked Fas/FasL up-regulation in other human leukaemia cells that expressed Bcr/Abl.

MATERIALS AND METHODS

Materials

Caffeine, 2-APB (2-aminoethoxydiphenyl borate), antimycin A, cyclosporine A, DPI (diphenyleneiodonium chloride), MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide], rotenone, Ruthenium Red, SB202190, SP600125 and an anti-β-actin antibody were obtained from Sigma–Aldrich. Adaphostin, Z-IETD-FMK (caspase 8 inhibitor), Z-DEVD-FMK (caspase 3 inhibitor), anti-(caspase 3) antibody and anti-(caspase 8) antibody were from Calbiochem. Anti-(caspase 9), anti-PARP [poly(ADP-ribose) polymerase], anti-FasL, anti-Akt, anti-phospho-Akt (Ser473), anti-p38 MAPK, anti-p38α MAPK, anti-(phospho-p38 MAPK), anti-ERK, anti-phospho-ERK, anti-JNK (c-Jun N-terminal kinase), anti-phospho-JNK, anti-c-Jun, anti-phospho-c-Jun (Ser73), anti-c-Fos, anti-ATF-2 (activating transcription factor 2), anti-phospho-ATF-2 (Thr71), anti-phospho-ASK1 (apoptosis signal-regulating kinase 1) (Thr845), anti-MEK1 (MAPK/ERK kinase 1), anti-c-Abl, anti-Bcl-2 and anti-Bax antibodies were obtained from Cell Signaling Technology. Anti-phospho-c-Fos (Ser374), anti-MKP-1 (C-19) (MAPK phosphatase 1), anti-Fas (N-18), anti-COX4 (cytochrome c oxidase 4), anti-ASK1 and anti-FADD (FD19) (Fas-associated death domain) antibodies were purchased from Santa Cruz Biotechnology. Neutralizing anti-human FasL monoclonal antibody NOK-2, anti-Bcl-xL, anti-Bid and anti-(cytochrome c) antibodies were purchased from BD Pharmingen. BAPTA-AM [1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid tetrakis(acetoxymethyl ester)], Fluo-4 AM, H2DCFDA (dichlorodihydrofluorescein diacetate) and Rhodamine-123 were from Molecular Probes. NECA (an agonist that acts on all subtypes of adenosine receptors) and HEMADO (a high affinity and selective A3 adenosine receptor agonist) were obtained from Tocris Bioscience. HRP (horseradish peroxidase)-conjugated secondary antibodies were obtained from Pierce. Cell culture supplies were purchased from Gibco/Life Technologies. Unless otherwise specified, all other reagents were of analytical grade.

Cell culture

Human chronic myeloid leukaemia K562 cells (Bcr/Abl-positive cells) and human acute myeloid leukaemia U937 cells (Bcr/Abl-negative cells) were obtained from A.T.C.C. (Manassas, VA, U.S.A.). Human chronic myeloid leukaemia KU812 cells (Bcr/Abl-positive cells) and human megakaryoblastic leukaemia MEG-01 cells (Bcr/Abl-positive cells) were obtained from BCRC (Bioresources Collection and Research Center, Hsinchu, Taiwan). K562 cells, U937 cells, KU812 cells and MEG-01 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, 1% sodium pyruvate, 2 mM glutamine and penicillin (100 units/ml)/streptomycin (100 μg/ml) in an incubator humidified with 95% air and 5% CO2. After specific treatment, cell viability was determined using an MTT assay.

Detection of apoptotic cells

Annexin V/PI (propidium iodide) staining was carried out according to the manufacturer's protocol (annexin V-FITC kit from Molecular Probes). After caffeine and/or adaphostin treatment for the indicated time, K562 cells were washed with ice-cold PBS and resuspended with binding buffer [10 mM Hepes (pH 7.4), 140 mM NaCl and 2.5 mM CaCl2] before transferring 1×105 cells to a 5 ml tube. Then 5 μl of annexin V-FITC and 5 μl of PI were added, and the cells were incubated for 15 min in the dark. Binding buffer (400 μl) was then added to each tube, and the cells were analysed using a Beckman Coulter Epics XL flow cytometer.

Detection of Fas and FasL mRNA expression by RT (reverse transcription)–PCR

Total RNA was isolated from untreated control cells, caffeine-treated cells, adaphostin-treated cells or caffeine/adaphostin-treated cells using the RNeasy minikit (Qiagen) according to the manufacturer's protocol. The RT reaction was performed with 2 μg of total RNA using MMLV (Moloney murine leukaemia virus) RT (Promega) according to the manufacturer's recommendations. A reaction without RT was performed in parallel to ensure the absence of genomic DNA contamination. After initial denaturation at 95°C for 10 min, PCR amplification was performed using GoTaq Flexi DNA polymerase (Promega) followed by 35 cycles at 94°C for 50 s, 58°C for 50 s, and 72°C for 50 s. After a final extension at 72°C for 5 min, PCR products were resolved on 2% agarose gels and visualized by ethidium bromide transillumination under UV light. Primer sequences were as follows: Fas, 5′-CAAGGGATTGGAATTGAGGA-3′ (forward) and 5′-GACAAAGCCACCCCAAGTTA-3′ (reverse); and FasL, 5′-TCTCAGACGTTTTTCGGCTT-3′ (forward) and 5′-AAGACAGTCCCCCTTGAGGT-3′ (reverse). The PCR yielded PCR products of 440 and 406 bp for Fas and FasL respectively. Each reverse-transcribed mRNA product was internally controlled by GAPDH (glyceraldehyde-3-phosphate dehydrogenase) PCR using primers 5′-GAGTCAACGGATTTGGTCGT-3′ (forward) and 5′-TGTGGTCATGAGTCCTTCCA-3′ (reverse), yielding a 512 bp PCR product. FasL and Fas RT–PCR products were subsequently confirmed by direct sequencing.

Real-time PCR

Quantitative PCR was performed using the iQ5 System (Bio-Rad Laboratories). Reactions were performed using GoTag qPCR master mix (Promega). Thermocycling conditions were: 2 min at 95°C, followed by 40 cycles of amplication for 15 s at 95°C, and 60°C for 60 s. The threshold cycle is defined as the cycle number at which fluorescence corresponding to the amplified PCR product is detected. The PCR arbitrary units of each gene were defined as the mRNA levels normalized to the GAPDH expression in each sample. Specificity was verified by melting curve analysis and agarose gel electrophoresis. Primer sequences used were as follows: GAPDH, (forward) 5′-GAAATCCCATCACCATCTTCCAGG-3′, (reverse) 5′-GAGCCCCAGCCTTCTCCATG-3′; Fas, (forward) 5′-AGCTTGGTCTAGAGTGAAAA-3′, (reverse) 5′-GAGGCAGAATCATG AGATAT-3′; and FasL, (forward) 5′-GCCTGTGTCTCCTTGTGAT-3′, (reverse) 5′-GCATCTGGCTGGTAGACTC-3′.

DNA transfection and luciferase assay

The luciferase construct pFLF1 containing the promoter region between −1435 and +236 of the Fas receptor gene was provided by Dr Y. Nakanishi (Graduate School of Natural Science and Technology, Kanazawa University, Kanazawa, Japan). This promoter region was inserted 18 bp upstream of the translation start codon of the firefly luciferase gene in the pGV-B vector. The 1.2 kb FasL promoter in a eukaryotic expression vector HsLuc carrying a luciferase reporter gene downstream of the inserted FasL promoter was provided by Dr D.R. Green (Institute of Pathology, University of Bern, Bern, Switzerland). The pCMV-MEK1 (expressing the constitutively active MEK1) and constitutively activated myristoylated Akt (CA-Akt) vectors were gifts from Dr W.C. Hung (Institute of Biomedical Sciences, National Sun Yat-Sen University, Kaohsiung, Taiwan). pcDNA-MKP-1 was obtained from Dr H.M. Lee (Graduate Institute of Medical Sciences, Taipei Medical University, Taipei, Taiwan). pcDNA3-Bcr/Abl (p210) was obtained from Dr B.J. Druker (Division of Hematology and Medical Oncology, Oregon Health & Science University, Portland, OR, U.S.A.). The plasmids were transfected into K562 or U937 cells using a Pipette-type Electroporator (MicroPorator-MP100, Digital Bio Tech). The luciferase assay was performed with the Luciferase Reporter Assay System (Promega).

RNA interference

c-Jun siRNA (small interfering RNA) (catalogue number sc-29223), c-Fos siRNA (catalogue number sc-29221), ATF-2 siRNA (catalogue number sc-29205), FADD siRNA (catalogue number sc-35352), MKP-1 siRNA (catalogue number sc-35937) and negative control siRNA (catalogue number sc-37007) were purchased from Santa Cruz Biotechnology. pKD-ASK1 shRNA (short-hairpin RNA), pKD-JNK1 shRNA and pKD-Negcon-V1 plasmids were purchased from Upstate Biotechnology, and pSuper-p38α shRNA plasmid was obtained from Dr A. Porras (Ciudad University, Spain). For the transfection procedure, cells were grown to 60% confluence, and siRNA and shRNA were transfected using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. Lipofectamine™ 2000 reagent was incubated with serum-free medium for 10 min, and the respective siRNA or shRNA was added subsequently. After incubation for 15 min at room temperature (25°C), the mixtures were diluted with culture medium and added to each well. At 24 h post-transfection, the cells were exposed to caffeine, adaphostin or a combination of caffeine and adaphostin for an additional 24 h. Afterwards, cells were harvested for Western blot analyses.

Measurement of mitochondrial membrane potential

After specific treatment, cells were incubated with 20 nM Rhodamine-123 at 37°C for 20 min. The cells were then washed twice with PBS, and the intensity of Rhodamine-123 was determined by flow cytometry. Cells with reduced fluorescence (less Rhodamine-123) were counted as having lost some of their mitochondrial membrane potential.

Subcellular fractionation

Following specific treatment, cytosolic and pellet (mitochondrial) fractions were generated using a digitonin-based subcellular fractionation technique. Briefly, 1×107 cells were harvested by centrifugation at 800 g, washed in PBS and re-pelleted. Cells were digitonin-permeablilized for 5 min on ice at a density of 3×107/ml in cytosolic extraction buffer (75 mM NaCl, 1 mM NaH2PO4, 8 mM Na2HPO4, 250 mM sucrose, 1 mM PMSF, 5 μg/ml leupeptin, 5 μg/ml aprotinin and 0.05% digitonin). Following centrifugation at 800 g at 4°C for 10 min, the supernatant was separated from the pellet comprising mitochondria and cellular debris. The supernatant containing cytoplasmic protein was further purified by centrifugation at 13000 g at 4°C for 10 min. The pellets were solubilized in the same volume of mitochondrial lysis buffer [50 mM Tris/HCl (pH 7.4), 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 0.2% Triton X-100, 0.3% Nonidet P40, 1 mM PMSF, 5 μg/ml leupeptin and 5 μg/ml aprotinin]. After centrifugation at 12000 g at 4°C for 10 min, the supernatants were collected and used as the mitochondrial fraction. Cytochrome c and proteins of the Bcl-2 family were detected by Western blot analyses.

Western blot analysis

After specific treatments, cells were incubated in lysis buffer containing 20 mM Tris/HCl (pH 7.5), 1% Triton X-100, 1 mM EDTA, 150 mM NaCl, 10% glycerol, 1 mM sodium orthovanadate, 50 mM sodium fluoride, 1 mM PMSF and protease inhibitor mixtures [SIGMAFAST™ protease inhibitor tablets (2 mM AEBSF, 0.3 μM aprotinin, 130 μM bestatin, 1 mM EDTA, 14 μM E-64 and 1 μM leupeptin), Sigma–Aldrich] for 20 min on ice. After insoluble debris was precipitated by centrifugation at 13000 g for 15 min at 4°C, the supernatants were collected and assayed for protein concentration using the Bradford method. An equal amount of protein per sample (15 μg) was resolved by SDS/PAGE and transferred on to a PVDF membrane. The transferred membranes were blocked for 1 h in 5% non-fat dried skimmed milk in PBST (PBS containing 0.05% Tween 20) and incubated with appropriate primary antibodies and HRP-conjugated secondary antibodies. The immune complexes were detected using the SuperSignal West Pico Chemiluminescent substrate kit (Pierce).

Gelatin zymography

Cells were incubated in serum-free medium overnight with or without caffeine. Culture medium or cell lysates were mixed with SDS loading buffer [0.125 M Tris/HCl (pH 6.8), 4% SDS and 0.02% Bromophenol Blue] and incubated for 30 min at 37°C. Samples were electrophoresed on a 10% polyacrylamide gel containing 0.1% gelatin. The gel was then washed in 2.5% Triton X-100 to remove SDS. The gel was incubated at 37°C for 48 h in 50 mM Tris/HCl (pH 7.6) containing 150 mM NaCl and 5 mM CaCl2. After staining with Coomassie Blue R-250, gelatinases were identified as clear bands.

Separation of human PBMCs (peripheral blood mononuclear cells)

Separation of PBMCs was conducted essentially according to a procedure described previously [6].

Measurement of intracellular ROS production and [Ca2+]i

H2DCFDA and Fluo-4 AM were used to detect the intracellular generation of ROS and [Ca2+]i respectively. Caffeine-treated cells were collected and incubated with 10 μM H2DCFDA for 20 min prior to harvesting, and then washed with PBS. Alternatively, caffeine-treated cells resuspended in 1 ml of PBS were incubated with 5 μl of 1 mM Fluo-4 AM for 1 h. The level of ROS and [Ca2+]i were measured using the Beckman Coulter Paradigm™ Detection Platform with excitation at 485 nm and emission at 530 nm. The protein concentration was measured using the Bradford method (Bio-Rad) with BSA as a standard. Results are shown as the fold-increase in fluorescence intensity per microgram of proteins compared with the control group.

Statistical analysis

All data are presented as means±S.D. Significant differences among the groups were determined using the unpaired Student's t test. A value of P<0.05 was taken as an indication of statistical significance. All of the Figures were obtained from at least three independent experiments with similar results. Results of Western blots were quantified using a scanning densitometer. Fold changes in protein expression were determined on the basis of the β-actin loading control.

RESULTS

A combination of caffeine and adaphostin increasingly induces death in K562 cells through a caspase 8/mitochondria-mediated death pathway

Upon exposure to caffeine or adaphostin, K562 cells showed a concentration-dependent decrease in cell viability (Figures 1A and 1B). The viability of caffeine-treated K562 cells was further reduced by the addition of adaphostin or vice versa (Figures 1A and 1B). A combination of 50 μM caffeine and 5 μM adaphostin led to an approximately 50% loss in viability of K562 cells, and the cytotoxic effect of caffeine/adaphostin was further examined. Control untreated K562 cells were viable with low PI and annexin V staining (lower left quadrants of the dot plots, Figure 1C). Compared with caffeine or adaphostin, caffeine/adaphostin treatment caused a further increase in the population of annexin V-stained cells. Figure 1(D) shows that caffeine treatment induced the production of active caspase 3 and caspase 8 in K562 cells, whereas adaphostin treatment only elicited the degradation of pro-caspase 3. Co-treatment with adaphostin increased the production of active caspase 3 and caspase 8 in caffeine-treated cells. Pretreatment with the caspase 3 inhibitor (Z-DEVD-FMK) and caspase 8 inhibitor (Z-IETD-FMK) restored the viability of caffeine- or caffeine/adaphostin-treated cells (Figure 1E). However, in contrast with the caspase 3 inhibitor, the caspase 8 inhibitor failed to rescue the viability of adaphostin-treated cells. These results suggested that caffeine and adaphostin evoked apoptosis of K562 cells through different pathways.

Figure 1 Combined effect of adaphostin and caffeine on inducing apoptosis of K562 cells

(A) Dose-dependent effect of caffeine on inducing cell death. K562 cells were treated with various concentrations of caffeine in the absence or presence of 5 μM adaphostin for 24 h. (B) Dose-dependent effect of adaphostin on inducing cell death. K562 cells were treated with various concentrations of adaphostin in the absence or presence of 50 μM caffeine for 24 h. Cell viability was determined using the MTT assay. Values are means±S.D. for three independent experiments with triplicate measurements. (C) Annexin V-FITC/PI double staining of caffeine-, adaphostin- and caffeine/adaphostin-treated K562 cells. K562 cells were treated with 50 μM caffeine, 5 μM adaphostin or a combination of 50 μM caffeine and 5 μM adaphostin for 24 h. On the flow cytometric scatter graphs, the left lower quadrant represents the remaining live cells. The right lower quadrant represents the population of early apoptotic cells. The right upper quadrant represents the accumulation of late apoptotic cells. (D) Degradation of pro-caspase 8 and pro-caspase 3 in caffeine-, adaphostin- and caffeine/adaphostin-treated cells. K562 cells were treated with 50 μM caffeine, 5 μM adaphostin or a combination of 50 μM caffeine and 5 μM adaphostin for 24 h. (E) Effect of caspase inhibitors on the viability of caffeine-, adaphostin- and caffeine/adaphostin-treated cells. K562 cells were pretreated with 100 μM Z-DEVD-FMK (a caspase 3 inhibitor) or Z-IETD-FMK (a caspase 8 inhibitor) for 1 h, and then incubated with 50 μM caffeine, 5 μM adaphostin or a combination of 50 μM caffeine and 5 μM adaphostin for 24 h. Cell viability was determined using the MTT assay (*P< 0.05). AD, adaphostin.

As shown in Figure 2(A), flow cytometry analysis showed that, in untreated control K562 cells, more than 98% of cells were functionally active with high Rhodamine-123 signals. The population of K562 cells with a loss of mitochondrial membrane potential (ΔΨm) increased after caffeine or adaphostin treatment. Treatment with caffeine/adaphostin further increased the population of K562 cells with dissipation of ΔΨm. As shown in Figure 2(B), release of cytochrome c into the cytosol was detected after treatment with caffeine, adaphostin or caffeine/adaphostin. Moreover, production of active caspase 9 was noted. These events suggested that caffeine-, adaphostin- and caffeine/adaphostin-induced apoptosis were mediated through the mitochondrial pathway. Compared with caffeine-treated cells, down-regulation of Bcl-2/Bcl-xL, mitochondrial translocation of Bax to mitochondria, cytochrome c release, degradation of pro-caspase 9 and production of tBid were notably increased in caffeine/adaphostin-treated cells. Noticeably, production of tBid was seen with adaphostin-treated cells. In terms of the fact that the cleavage of Bid by caspase 8 generates tBid [33], the results of the present study suggest that adaphostin enhances the caspase 8/mitochondria-mediated death pathway in caffeine-treated cells.

Figure 2 Mitochondrial membrane potential, pro-caspase 9 degradation, Bcl-2 family protein expression and cytochrome c release in caffeine-, adaphostin- and caffeine/adaphostin-treated K562 cells

Cells were treated with 50 μM caffeine, 5 μM adaphostin or a combination of 50 μM caffeine and 5 μM adaphostin for 24 h. (A) The loss of ΔΨm in caffeine-, adaphostin- and caffeine/adaphostin-treated cells. Top panels: representative data showing ΔΨm loss. Bottom panel: statistical analyses of ΔΨm loss. The values represent means±S.D. for three independent experiments (*P< 0.05). (B) Western blot analyses of pro-caspase 9 degradation, cytochrome c release and Bcl-2 family proteins in caffeine-, adaphostin- and caffeine/adaphostin-treated cells. To examine the purity of either the mitochondrial (Mito.) or cytosolic (Cyto.) fraction, a mitochondrial marker COX4 (cytochrome c oxidase 4) or a cytosolic marker β-actin were immunoblotted for each fraction. AD, adaphostin; Ctrl, control; cyt c, cytochrome c.

Adaphostin enhances the caffeine-elicited autocrine Fas-mediated death pathway

Death receptors of the TNF (tumour necrosis factor) family such as Fas and TNFR1 (TNF receptor 1) are the best-understood death pathways that recruit FADD and pro-caspase 8 to the receptor. Recruitment of pro-caspase 8 through FADD leads to its autocleavage and activation, and in turn activates effector caspases, such as caspase 3, in causing cell death [34]. Immunoblotting analyses displayed a time-dependent increase in Fas and FasL protein expression after caffeine treatment (Figure 3A). Although adaphostin did not elicit Fas/FasL up-regulation, co-treatment with adaphostin caused a further increase in Fas/FasL protein expression in caffeine-treated cells. Consistent with the immunoblot data, caffeine/adaphostin treatment notably increased the level of both Fas mRNA and FasL mRNA compared with caffeine, as shown by RT–PCR analyses and real-time PCR analyses (Figures 3B and 3C). A promoter assay also revealed that adaphostin enhanced the luciferase activity of Fas promoter and FasL promoter in caffeine-treated cells (Supplementary Figure S1 at http://www.BiochemJ.org/bj/439/bj4390453add.htm). Adaphostin was unable to significantly increase the transcriptional level of Fas/FasL mRNA (Figures 3B and 3C) and luciferase activity of the Fas/FasL promoter (Supplementary Figure S1). Down-regulation of FADD expression with siRNA markedly abrogated the caffeine/adaphostin-induced degradation of pro-caspase 8 and the production of tBid (Figure 3D). Compared with control siRNA, FADD siRNA rescued the viability of caffeine/adaphostin-treated cells. In contrast, the down-regulation of FADD did not rescue the viability of adaphostin-treated cells or attenuate adaphostin-induced degradation of pro-caspase 3. Accordingly, adaphostin-induced cell death was via a Fas-independent pathway, but adaphostin enhanced the caffeine-elicited autocrine Fas-mediated death pathway. Pretreatment with neutralizing anti-FasL antibody reduced caffeine/adaphostin-induced pro-caspase 8 degradation and restored viability of caffeine/adaphostin-treated K562 cells (Figure 3E), revealing that the Fas/FasL death pathway was functionally involved in caffeine/adaphostin-induced death of K562 cells.

Figure 3 Adaphostin enhanced caffeine-induced Fas-mediated death pathway activation in K562 cells

K562 cells were treated with 50 μM caffeine, 5 μM adaphostin or a combination of 50 μM caffeine and 5 μM adaphostin for the time periods indicated (A) or 24 h (B, C and D). (A) Top panels: Western blot analyses of Fas and FasL protein expression. Bottom panels: quantification of the expression of Fas and FasL from Western blot analyses. Three independent experimental results were analysed by densitometry (*P< 0.05). (B) Detecting the expression of Fas and FasL using RT–PCR. RT–PCR was conducted according to the procedure described in the Materials and methods section. (C) Real-time PCR analyses of Fas/FasL mRNA expression in caffeine-, adaphostin- and caffeine/adaphostin-treated cells (*P< 0.05). (D) FADD siRNA abrogated caffeine/adaphostin-induced degradation of pro-caspase 8 and production of tBid. K562 cells were transfected with 100 nM control siRNA or FADD siRNA respectively. At 24 h post-transfection, the cells were treated with adaphostin or caffeine/adaphostin for 24 h. Left-hand panels: Western blot analyses of FADD protein expression, degradation of pro-caspase 8 and 3, and production of tBid. Right-hand panel: viability of control siRNA- and FADD siRNA-transfected cells after treatment with caffeine, adaphostin or caffeine/adaphostin for 24 h. Cell viability was determined using the MTT assay (*P< 0.05). (E) Effect of neutralizing anti-human FasL monoclonal antibody on pro-caspase 8 degradation and viability of caffeine/adaphostin-treated K562 cells. K562 cells were treated with a combination of 50 μM caffeine and 5 μM adaphostin for 24 h in the presence of the indicated concentration of anti-FasL antibody. Left-hand panels: Western blot analyses of pro-caspase 8 degradation. Right-hand panel: anti-FasL antibody rescued viability of caffeine/adaphostin-treated cells. Cell viability was determined using the MTT assay. Values are means±S.D. for three independent experiments with triplicate measurements. AD, adaphostin; Ctrl, control.

Adaphostin-induced MKP-1 down-regulation enhances p38 MAPK/JNK-mediated Fas/FasL up-regulation in caffeine-treated K562 cells

Given that JNK, ERK and p38 MAPK are functionally involved in the transcriptional regulation of Fas and FasL [3538], the level of phospho-p38 MAPK, phospho-ERK and phospho-JNK was examined. As shown in Figure 4(A), compared with that in caffeine-treated cells, p38 MAPK and JNK were increasingly phosphorylated in caffeine/adaphostin-treated cells. Meanwhile, the level of phospho-ERK in caffeine/adaphostin-treated cells was notably decreased. In contrast with sustained phosphorylation of p38 MAPK after caffeine treatment for 24 h, adaphostin-evoked p38 MAPK activation declined to a basal level at 24 h. JNK was marginally phosphorylated in adaphostin-treated cells. Previous studies have revealed that Bcr/Abl regulates phosphorylation of Akt and ERK [21,22] and ASK1 is critical for p38 MAPK and JNK activation [39]. Thus phospho-Akt and phospho-ASK1 were also examined. As shown in Figure 4(B), adaphostin or caffeine/adaphostin markedly attenuated Akt phosphorylation, whereas caffeine modestly reduced the level of phospho-Akt. In contrast with adaphostin, caffeine and caffeine/adaphostin induced phosphorylation of ASK1 (Figure 4B). An increase in MKP-1 expression was suggested to suppress phosphorylation of JNK and p38 MAPK [6,4042]. Figure 4(B) shows that MKP-1 protein expression was reduced by caffeine treatment, whereas adaphostin markedly suppressed MKP-1 expression. Overexpression of either constitutively active Akt or constitutively active MEK1 suppressed caffeine/adaphostin-induced p38 MAPK activation, JNK activation, ERK inactivation and Akt inactivation, but had no significant effect on caffeine/adaphostin-induced ASK1 phosphorylation (Figure 4C). Accordingly, the results of the present study has proposed that activated Akt/ERK abrogated p38 MAPK/JNK activation via an ASK1-independent pathway. Figure 4(C) shows that either constitutively active Akt or ERK restored phosphorylation of Akt and ERK in caffeine/adaphostin-treated cells. Treatment with LY294002 (a PI3K inhibitor) or U0126 (MEK1 and MEK2 inhibitors) suppressed both Akt and ERK activation in K562 cells (results not shown). Obviously, active Akt is crucial for ERK activation and vice versa. Consistent evidence of cross-talk between the Raf/MEK/ERK and Akt pathway has previously been described [43,44]. Moreover, restoration of ERK and Akt activation rescued MKP-1 expression in caffeine/adaphostin-treated cells, suggesting that suppression of Akt/ERK down-regulated MKP-1 protein expression. Figure 4(D) shows that transfection of pcDNA3-MKP-1 suppressed caffeine/adaphostin-induced p38 MAPK and JNK activation, suggesting that MKP-1 down-regulation contributed to p38 MAPK/JNK activation. Overexpression of MKP-1 markedly attenuated caffeine/adaphostin-induced Fas/FasL up-regulation, but had no significant effect on ASK1 phosphorylation (Figure 4D). As seen in Supplementary Figure S2 (at http://www.BiochemJ.org/bj/439/bj4390453add.htm), down-regulation of ASK1 repressed caffeine/adaphostin-induced Fas/FasL up-regulation, p38 MAPK activation and JNK activation, but did not affect caffeine/adaphostin-induced ERK inactivation and MKP-1 down-regulation. Taken together, the results of the present study revealed that ASK1 activation, in concert with MKP-1 down-regulation, elicited Fas/FasL up-regulation and p38 MAPK/JNK activation in caffeine/adaphostin-treated cells. Figure 4(E) shows that knockdown of MKP-1 made caffeine mimic the action of caffeine/adaphostin on Fas/FasL up-regulation and p38 MAPK/JNK activation, again supporting a role for adaphostin-induced MKP-1 down-regulation in enhancing caffeine-induced Fas/FasL up-regulation. As shown in Supplementary Figure S3A (at http://www.BiochemJ.org/bj/439/bj4390453add.htm), co-treatment with SB202190 and SP600125 abrogated Fas/FasL up-regulation in caffeine- and caffeine/adaphostin-treated cells. Transfection of either constitutively active Akt or MEK1 markedly attenuated caffeine/adaphostin-induced up-regulation of Fas/FasL (Supplementary Figure S3B). This emphasized the notion that p38 MAPK/JNK activation and Akt/ERK inactivation were involved in Fas/FasL up-regulation in caffeine/adaphostin-treated cells.

Figure 4 A causal relationship between ERK/Akt inactivation and p38 MAPK/JNK activation in caffeine/adaphostin-treated K562 cells

Concentrations of caffeine and adaphostin used were 50 μM and 5 μM respectively. (A) Western blot analyses of phosphorylated MAPKs in caffeine-, adaphostin- and caffeine/adaphostin-treated K562 cells. K562 cells were treated with caffeine, adaphostin or caffeine/adaphostin for the time periods indicated. (B) Top panels: Western blot analyses of Akt phosphorylation, ASK1 phosphorylation and MKP-1 protein expression in caffeine-, adaphostin- and caffeine/adaphostin-treated K562 cells. K562 cells were treated with caffeine, adaphostin or caffeine/adaphostin for 24 h. Bottom panel: quantification of the expression of MKP-1 from Western blot analyses. Three independent experimental results were analysed by densitometry (*P< 0.05). (C) Effect of constitutively active Akt and MEK1 on the level of phospho-p38 MAPK, phospho-JNK, phospho-ERK, phospho-ASK1, phospho-Akt and MKP-1 expression in caffeine/adaphostin-treated cells. K562 cells were transfected with constitutively active Akt (CA-Akt) or pCMV-MEK1 respectively. At 24 h post-transfection, the transfected cells were treated with caffeine/adaphostin for 24 h. (D) Effect of overexpression of MKP-1 on caffeine/adaphostin-induced activation of p38 MAPK and JNK. K562 cells were transfected with an empty expression vector or pcDNA-MKP-1 respectively. At 24 h post-transfection, the control vector-transfected cells and pcDNA-MKP-1-transfected cells were treated with caffeine/adaphostin for 24 h. Top panels: Western blot analyses. Bottom panel: quantification of Fas/FasL protein expression and the level of phospho-p38 MAPK/phospho-JNK from Western blot analyses. Three independent experimental results were analysed by densitometry (*P< 0.05). (E) Knockdown of MKP-1 enhanced caffeine-induced activation of p38 MAPK and JNK. K562 cells were transfected with 100 nM control siRNA or MKP-1 siRNA respectively. At 24 h post-transfection, control siRNA-transfected cells and MKP-1 siRNA-transfected cells were treated with caffeine/adaphostin and caffeine for 24 h. AD, adaphostin; Ctrl, control; p-, phospho-.

p38α MAPK/ATF-2 and JNK1/c-Jun pathways elicit Fas/FasL up-regulation in caffeine and caffeine/adaphostin-treated cells

Previous studies have shown that p38α MAPK and JNK1 are involved in Fas/FasL protein expression in phospholipase A2-treated K562 cells [45]; hence, the role of p38α MAPK and JNK1 was examined. Transfection of pSuper-p38α shRNA or pKD-JNK1 shRNA attenuated Fas/FasL up-regulation in caffeine- or caffeine/adaphostin-treated cells (Figure 5A). Caffeine and caffeine/adaphostin failed to induce Fas/FasL up-regulation in cells co-transfected with p38α shRNA and JNK1 shRNA (Figure 5B). This suggested that both p38α MAPK and JNK1 were responsible for up-regulating Fas/FasL protein expression.

Figure 5 Both p38α MAPK/ATF-2 and JNK1/c-Jun pathways were responsible for up-regulated Fas/FasL protein expression in caffeine-treated and caffeine/adaphostin-treated cells

Concentrations of caffeine and adaphostin used were 50 μM and 5 μM respectively. (A) Down-regulation of p38α MAPK and JNK1 attenuated Fas/FasL protein expression in caffeine-treated and caffeine/adaphostin-treated cells. K562 cells were transfected with control vector, pSuper-p38α shRNA or pKD-JNK1 shRNA respectively. At 24 h post-transfection, the cells were treated with caffeine or caffeine/adaphostin for 24 h. Top panels: Western blot analyses. Bottom panels: quantification of Fas/FasL protein expression from Western blot analyses. Three independent experimental results were analysed by densitometry (* and **P< 0.05, shRNA-transfected cells compared with control vector-transfected cells). (B) Involvement of p38α MAPK and JNK1 in caffeine- and caffeine/adaphostin-evoked Fas/FasL up-regulation. K562 cells were co-transfected with pSuper-p38α shRNA and pKD-JNK1 shRNA. At 24 h post-transfection, the cells were treated with caffeine or caffeine/adaphostin for 24 h. Left-hand panels: Western blot analyses. Right-hand panels: quantification of Fas/FasL protein expression from Western blot analyses. Three independent experimental results were analysed by densitometry (* and **P< 0.05, shRNA-transfected cells compared with control vector-transfected cells). (C) Effect of p38α MAPK shRNA and JNK1 shRNA on phosphorylation of ATF-2 and c-Jun in caffeine/adaphostin-treated cells. K562 cells were transfected with control vector, pSuper-p38α shRNA or pKD-JNK1 shRNA respectively. At 24 h post-transfection, the cells were treated with caffeine/adaphostin for 24 h. (D) Effect of constitutively active Akt (CA-Akt) and MEK1 on the level of phospho-c-Jun, phospho-ATF-2 and phospho-c-Fos in caffeine/adaphostin-treated cells. K562 cells were transfected with constitutively active Akt or pCMV-MEK1 respectively. At 24 h post-transfection, the transfected cells were treated with caffeine/adaphostin for 24 h. (E) Effect of U0126 on c-Fos phosphorylation and MKP-1 protein expression in K562 cells and caffeine-treated cells. K562 cells were treated with 10 μM U0126 for 1 h, and then incubated with 50 μM caffeine for 24 h. Left-hand panels: Western blot analyses. Right-hand panels: quantification of MKP-1, Fas and FasL protein expression from Western blot analyses. Three independent experimental results were analysed by densitometry (* and **P< 0.05). (F) Effect of ATF-2 siRNA and c-Jun siRNA on Fas/FasL protein expression in caffeine/adaphostin-treated cells. K562 cells were transfected with 100 nM control siRNA, 100 nM ATF-2 siRNA, 100 nM c-Jun siRNA or a combination of 100 nM ATF-2 siRNA and 100 nM c-Jun siRNA. At 24 h post-transfection, the cells were treated with caffeine/adaphostin for 24 h. (G) Suppression of c-Fos expression reduced the time period for caffeine/adaphostin-elicited increase in Fas and FasL expression. K562 cells were transfected with 100 nM control siRNA or c-Fos siRNA respectively. At 24 h post-transfection, the cells were treated with caffeine/adaphostin for the time periods indicated. Left-hand panels: Westerm blot analyses of Fas/FasL protein expression. Right-hand panel: quantification of Fas and FasL protein expression from Western blot analyses. Three independent experimental results were analysed by densitometry (*P< 0.05). (H) Suppression of c-Fos expression led to MKP-1 down-regulation. K562 cells were transfected with 100 nM control siRNA or c-Fos siRNA respectively. At 24 h post-transfection, the cells were treated with caffeine for 24 h. AD, adaphostin; Ctrl, control; p-, phospho-.

Activated MAPK-elicited downstream events, including phosphorylation of c-Jun, c-Fos or ATF-2, have been demonstrated to be involved in Fas and FasL protein expression [35,36,4648]. As shown in Supplementary Figure S4A (at http://www.BiochemJ.org/bj/439/bj4390453add.htm), the increase in c-Jun/ATF-2 phosphorylation, but reduction in the level of phospho-c-Fos, were noted in caffeine- and caffeine/adaphostin-treated cells. SB202190 and SP600125 pretreatment abrogated phosphorylation of ATF-2 and c-Jun respectively in caffeine/adaphostin-treated cells (Supplementary Figure S4B). Transfection with p38α shRNA or JNK1 shRNA abrogated caffeine/adaphostin-evoked phosphorylation of ATF-2 and c-Jun respectively (Figure 5C). Nevertheless, p38α MAPK and JNK1 were not involved in dephosphorylation of c-Fos in caffeine/adaphostin-treated cells. Overexpression of either constitutively active Akt or MEK1 abrogated caffeine/adaphostin-elicited phosphorylation of c-Jun/ATF-2 and dephosphorylation of c-Fos (Figure 5D). p38 MAPK and JNK activation suppressed by Akt/ERK-mediated MKP-1 expression suggested that the Akt/ERK pathway should be involved in c-Fos phosphorylation. This hypothesis was further supported by the findings that U0126 or LY294002 (results not shown) treatment reduced the level of phospho-c-Fos in K562 cells (Figure 5E). Figure 5(E) shows that U0126 treatment down-regulated MKP-1 protein expression. Compared with caffeine alone, a combination of U0126 and caffeine further suppressed MKP-1 protein expression. Moreover, U0126 significantly increased Fas/FasL protein expression in K562 cells and caffeine-treated cells. Figure 5(F) shows that co-transfection of both ATF-2 siRNA and c-Jun siRNA abrogated caffeine- and caffeine/adaphostin-induced up-regulation of Fas/FasL. These results conclusively suggested that JNK1/c-Jun and p38α MAPK/ATF-2 pathways regulated transcription of Fas and FasL genes. Compared with control siRNA-transfected cells, Fas/FasL protein expression in c-Fos siRNA-transfected cells was increasingly induced by caffeine/adaphostin at an earlier time period (Figure 5G). Figures 5(E) and 5(H) show that suppression of c-Fos activation or knockdown of c-Fos led to MKP-1 down-regulation, suggesting that Akt/ERK-mediated c-Fos phosphorylation was responsible for MKP-1 protein expression. Moreover, suppression of MKP-1 expression by c-Fos siRNA markedly increased caffeine-induced Fas/FasL up-regulation in K562 cells.

Caffeine and caffeine/adaphostin elicit Ca2+-stimulated ASK1 activation and ROS-mediated Akt/ERK inactivation

Caffeine is well known to induce an increase in [Ca2+]i in various cells [32], and the lethal effect of adaphostin has been reported to be related to oxidative damage [23]. To examine whether Ca2+ and ROS were involved in ASK1 activation and Akt/ERK inactivation, the levels of Ca2+ and ROS were examined in caffeine- and adaphostin-treated K562 cells. Figure 6(A) shows that caffeine and adaphostin treatment notably induced an increase in ROS generation. Figure 6(B) shows that caffeine markedly elicited an increase in [Ca2+]i within 1 h, whereas [Ca2+]i did not change significantly in adaphostin-treated cells. Pretreatment with BAPTA-AM (a Ca2+ chelator) abolished the caffeine-induced [Ca2+]i increase and ROS generation (Figures 6C and 6D). NAC (N-acetyl cysteine; a ROS scavenger) suppressed ROS generation in caffeine-treated cells, but did not affect the caffeine-induced [Ca2+]i increase. These results suggested that Ca2+ evoked ROS generation in caffeine-treated cells. Pretreatment with BAPTA-AM abolished ASK1 phosphorylation and Akt/ERK dephosphorylation in caffeine-treated cells (Figure 6E). Figure 6(E) shows that NAC pretreatment abrogated Akt/ERK dephosphorylation, but did not suppress ASK1 phosphorylation in caffeine-treated cells. These results suggested that Ca2+ was located on the upstream position for ASK1 activation, and that the [Ca2+]i increase induced the production of ROS responsible for Akt/ERK inactivation in caffeine-treated cells. Previous studies have shown that the ryanodine receptor functions as an intracellular channel responsible for releasing Ca2+ in caffeine-treated cells [32,49], and caffeine inhibits IP3 (inositol 1,4,5-trisphosphate) receptor-mediated intracellular Ca2+ release [50]. Blocking the ryanodine receptor by Ruthenium Red abrogated the caffeine-induced [Ca2+]i increase, whereas the IP3 receptor inhibitor 2-APB had no effect (Figure 6F). Consistently, pretreatment with Ruthenium Red abolished caffeine-induced ASK1 activation, Akt/ERK inactivation and ROS generation (Figures 6G and 6H). Previous studies have suggested that NADPH oxidase, in response to elevated cytosolic Ca2+, generated a large amount of superoxide [51]. DPI irreversibly inactivates many flavoproteins and is routinely used as a NADPH oxidase inhibitor [52]. Pretreatment with DPI did not affect caffeine-induced ROS generation (Figure 6I). To further identify the casual relationship between mitochondrial alteration and ROS generation, K562 cells were pretreated with a mitochondria permeability transition pore inhibitor (cyclosporine A) or mitochondrial electron transport chain complexes I and III inhibitors (rotenone or antimycin A). Figure 6(I) shows that cyclosporine A, rotenone and antimycin A pretreatment reduced caffeine- or adaphostin-induced ROS generation. Compared with cyclosporine A, rotenone and antimycin A, DPI notably reduced adaphostin-induced ROS generation (Figure 6I). Pretreatment with NAC abrogated Akt/ERK inactivation in adaphostin-treated cells (Figure 6E). These results suggested that adaphostin-induced ROS generation led to Akt/ERK inactivation.

Figure 6 Effect of [Ca2+]i increase and ROS generation on ASK1 and Akt/ERK phosphorylation in caffeine- and adaphostin-treated cells

K562 cells were pretreated with 10 μM BAPTA-AM, 2 mM NAC, 50 μM 2-APB, 10 μM Ruthenium Red, 10 μM DPI, 1 μM cyclosporine A, 1 μM rotenone, 10 μM antimycin A, 10 μM NECA or 10 μM HEMADO for 1 h and then incubated with 50 μM caffeine or 5 μM adaphostin for 24 h. (A) Elevation of [Ca2+]i was noted with K562 cells after treatment with 50 μM caffeine or 5 μM adaphostin for the time periods indicated. [Ca2+]i was quantified by fluorescence reading after loading the cells with a calcium indicator (Fluo-4 AM). Results are shown as the fold-increase in fluorescence intensity compared with control groups. Values are means±S.D. (B) After treatment with 50 μM caffeine or 5 μM adaphostin for the time periods indicated, ROS generation was quantified using a fluorescence plate reader. Results are shown as the fold-increase in fluorescence intensity compared with control groups. Values are means±S.D. (C) Effect of BAPTA-AM and NAC on caffeine-elicited [Ca2+]i increase. Values are means±S.D. (*P< 0.05). (D) Effect of BAPTA-AM and NAC on caffeine- and adaphostin-elicited ROS generation. Values are means±S.D. (*P< 0.05). (E) Effect of BAPTA-AM and NAC on ASK1, ERK and Akt phosphorylation in caffeine- and adaphostin-treated cells. Left-hand panels: Western blot analyses. Right-hand panels: quantification of p-ASK1/ASK1, p-Akt/Akt and p-ERK/ERK from Western blot analyses. Three independent experimental results were analysed by densitometry (*P< 0.05 compared with caffeine-treated cells; **P< 0.05 compared with adaphostin-treated cells). (F) Effect of inhibitors of calcium release from intracellular depots (2-APB and Ruthenium Red) on caffeine-evoked [Ca2+]i release. Values are the means±S.D. (*P< 0.05). (G) Effect of 2-APB and Ruthenium Red on caffeine-elicited ASK1 activation and Akt/ERK inactivation. Left-hand panels: Western blot analyses. Right-hand panels: quantification of p-ASK1/ASK1, p-Akt/Akt and p-ERK/ERK from Western blot analyses. Three independent experimental results were analysed by densitometry (*P<0.05). (H) Effect of 2-APB and Ruthenium Red on caffeine-evoked ROS generation. Values are means±S.D. (*P< 0.05). (I) Effect of DPI, cyclosporine A, rotenone or antimycin A on caffeine- and adaphostin-elicited ROS generation. Values are means±S.D. (*P< 0.05 compared with caffeine-treated cells; **P< 0.05 compared with adaphostin-treated cells). (J) Effect of NECA and HEMADO on the caffeine-induced increase in [Ca2+]i and ROS generation. Values are means±S.D. (*P< 0.05). (K) Effect of NECA and HEMADO on caffeine-induced ASK1 activation, Akt/ERK inactivation and Fas/FasL up-regulation. Left-hand panels: Western blot analyses. Right-hand panel: quantification of p-ASK1/ASK1, p-Akt/Akt and p-ERK/ERK from Western blot analyses. Three independent experimental results were analysed by densitometry (*P< 0.05). AD, adaphostin; Ant A, antimycin A; CSA, cyclosporin A; p-, phospho-; Rot, rotenone; RR, Ruthenium Red.

Caffeine is well known to cause most of its biological effects via antagonizing all types of adenosine receptors [53]. Compared with NECA (an agonist that acts on subtypes of adenosine receptors), HEMADO (a high affinity and selective A3 adenosine receptor agonist) pretreatment suppressed the caffeine-induced [Ca2+]i increase and ROS generation (Figure 6J). Moreover, HEMADO abolished caffeine-induced ASK1 activation, Akt/ERK inactivation and Fas/FasL up-regulation (Figure 6K). Taken together with the finding that the A3 adenosine receptor is expressed in K562 cells [54], the results suggest that the cytotoxic effect of caffeine on K562 cells was mediated through its antagonistic effect on the A3 adenosine receptor. Consistently, Merighi et al. [55] found that the anti-tumour effect of caffeine on colon cancer cells was associated with suppressing A3 adenosine receptor stimulation. Alternatively, caffeine has been shown to inhibit GABAA (γ-aminobutyric acid A) receptors and phosphodiesterase activity [32]. Pretreatment with the GABAA receptor blocker bicuculline (10 μM) and the phosphodiesterase inhibitor IBMX (isobutylmethylxanthine; 100 μM) had no effect on caffeine-induced [Ca2+]i increase and ROS generation (results not shown).

Adaphostin promotes caffeine-evoked Fas/FasL up-regulation in human leukaemia cells expressing Bcr/Abl

As shown in Figure 7(A), adaphostin further reduced viability of caffeine-treated KU182 and MEG-01 cells. The promoter assay revealed that caffeine/adaphostin treatment increased the luciferase activity of Fas promoter and FasL promoter in KU812 and MEG-01 cells (Figure 7B). Compared with treatment using caffeine alone, caffeine/adaphostin treatment increasingly induced Fas/FasL up-regulation, p38 MAPK activation, JNK activation, ERK inactivation, MKP-1 down-regulation, pro-caspase 8 degradation and the production of tBid in KU812 and MEG-01 cells (Figure 7C). In sharp contrast with the inability of caffeine to induce Fas/FasL up-regulation in U937 cells (Figure 7D), caffeine induced Fas/FasL up-regulation in pcDNA3-Bcr/Abl (p210)-transfected U937 cells (Figure 7E). Co-treatment with adaphostin further increased caffeine-induced Fas-mediated death pathway activation in U937 cells expressing Bcr/Abl (p210). These results suggested that the adaphostin effect on promoting caffeine-induced Fas/FasL up-regulation was common to Bcr/Abl-positive leukaemia cells.

Figure 7 Caffeine and caffeine/adaphostin treatment induced Fas/FasL up-regulation in KU812, MEG-01 and pcDNA3-Bcr/Abl (p210)-transfected U937 cells

Concentrations of caffeine and adaphostin used were 50 μM and 5 μM respectively. (A) Effect of caffeine, adaphostin and caffeine/adaphostin on the viability of K562, KU812 and MEG-01. Cells were treated with caffeine, adaphostin or caffeine/adaphostin for 24 h. Cell viability was determined using the MTT assay (**P< 0.05, caffeine-treated cells compared with caffeine/adaphostin-treated cells; *P< 0.05, adaphostin-treated cells compared with caffeine/adaphostin-treated cells). (B) Co-treatment with caffeine and adaphostin elicited an increase in transcriptional activity of the Fas promoter and FasL promoter in Bcr/Abl-positive leukaemia cells. K562, KU812 or MEG-01 cells were transfected with pGV-B-Fas (Fas promoter–luciferase construct) or HsLuc-FasL (FasL promoter–luciferase construct) respectively. At 24 h post-transfection, cells were treated with caffeine/adaphostin for 24 h (*P< 0.05). (C) Western blot analyses of phosphorylated MAPKs, degradation of pro-caspase 8 and 3, production of tBid and Fas/FasL protein expression in KU812 and MEG-01 cells after treatment with caffeine, adaphostin or caffeine/adaphostin. Cells were treated with caffeine, adaphostin or caffeine/adaphostin for 24 h. (D) Effect of caffeine and adaphostin on Fas/FasL protein expression in U937 cells. U937 cells were treated with caffeine, adaphostin or caffeine/adaphostin for 24 h. (E) Western blot analyses of phosphorylated MAPKs, degradation of pro-caspase 8 and 3, production of tBid and Fas/FasL protein expression in pcDNA3-Bcr/Abl (p210)-transfected U937 cells. U937 cells were transfected with pcDNA3-Bcr/Abl (p210). At 24 h post-transfection, the cells were treated with caffeine, adaphostin or caffeine/adaphostin for 24 h. AD, adaphostin; p-, phospho-.

DISCUSSION

The results of the present study have revealed that caffeine induces ASK1 activation and MKP-1 down-regulation, leading to p38α MAPK and JNK1 activation (Figure 8). Both p38α MAPK/ATF-2 and JNK1/c-Jun pathways are involved in caffeine-induced Fas and FasL up-regulation. Akt/ERK suppression by adaphostin elicits MKP-1 down-regulation, increasingly promoting p38α MAPK and JNK1 activation in caffeine-treated cells. Consequently, adaphostin enhances Fas-mediated death pathway activation in caffeine-treated cells. Previous studies have shown that the antiproliferative effect of AG957 on CD34+ chronic myeloid leukaemia progenitor cells is dramatically increased by combining AG957 and anti-Fas antibody [56]. Obviously, Fas-mediated death pathway activation improves the therapeutic efficacy of Bcr/Abl inhibitors. Noticeably, the results of the present study show that adaphostin-induced mitochondrial depolarization is not through the caspase 8-mediated pathway. Transfection with constitutively active Akt or MEK1 attenuated adaphostin-induced ΔΨm loss and degradation of pro-caspase 9 and 3 (Supplementary Figure S5 at http://www.BiochemJ.org/bj/439/bj4390453add.htm). This reflects that both Akt and ERK inactivation contribute to adaphostin-evoked mitochondrial depolarization. Consistent with this, Mow et al. [57] suggested that adaphostin-induced death of K562 cells is associated with mitochondrial damage. Noticeably, the results of the present study show that caffeine-elicited [Ca2+]i increase is mediated through its antagonistic effect on the A3 adenosine receptor. Although previous studies have shown that suppression of A3 adenosine receptor activation elicited an NMDA (N-methyl-D-aspartate) receptor-mediated increase in cytoplasmic Ca2+ [58], a causal relationship between A3 adenosine receptor inhibition by caffeine and ryanodine receptor-sensitive Ca2+ channel activation has not yet been reported. Thus the mechanism remains to be investigated in future studies. The caffeine-elicited [Ca2+]i increase induces ROS generation, and Ca2+ and ROS are responsible for ASK1 activation and Akt/ERK inactivation respectively in caffeine-treated cells (Figure 8). Co-treatment with adaphostin increases the production of ROS, which thus enhances Akt/ERK inactivation and suppresses Akt/ERK-mediated MKP-1 expression in caffeine-treated cells (Figure 8).

Figure 8 Signalling pathways elucidate the effect of adaphostin on enhancing caffeine-elicited Fas/FasL up-regulation in K562 cells

Caffeine-induced up-regulation of Fas/FasL protein expression was mediated through p38α MAPK/ATF-2 and JNK1/c-Jun pathways. Akt/ERK-evoked c-Fos activation induced MKP-1 protein expression, and MKP-1 suppressed phosphorylation of p38 MAPK and JNK. Compared with caffeine, adaphostin notably suppressed Akt/ERK-mediated MKP-1 protein expression. Thus adaphostin treatment promoted activation of p38 MAPK and JNK in caffeine-treated cells and enhanced caffeine-induced Fas/FasL up-regulation in K562 cells. Of note, the caffeine-induced [Ca2+]i increase elicited ROS generation, and Ca2+ and ROS were responsible for ASK1 activation and Akt/ERK inactivation respectively. Co-treatment with adaphostin increased the production of ROS, which thus enhanced Akt/ERK inactivation and suppressed Akt/ERK-mediated MKP-1 expression in caffeine-treated cells. p-, phospho-.

Our previous studies have shown that ERK inactivation and p38 MAPK activation are crucial for MMP-2 and MMP-9 down-regulation in caffeine-treated U937 cells [6]. The results from the present study reveal that caffeine marginally reduces ERK phosphorylation, but notably induces p38 MAPK/JNK activation in K562, KU812 and MEG-01 cells (Figures 4A and 7C). Meanwhile, caffeine did not significantly effect MMP-2 and MMP-9 secretion in K562, KU812 or MEG-01 cells (Supplementary Figure S6 at http://www.BiochemJ.org/bj/439/bj4390453add.htm). Noticeably, caffeine treatment slightly reduces the level of phospho-ERK and increases the levels of phospho-p38 MAPK and phospho-JNK in U937 cells expressing Bcr/Abl (Figure 7E). Consistently, caffeine is unable to down-regulate MMP-2 and MMP-9 protein expression in pcDNA3-Bcr/Abl (p210)-transfected U937 cells (Supplementary Figure S6). Probably, Bcr/Abl tyrosine kinase suppresses the caffeine effect on MMP-2/MMP-9 genetic regulation in leukaemia cells.

Given that caffeine plus 5 μM adaphostin resulted in approximately 50% cell death, which was similar to 10 μM adaphostin alone, one may argue the usefulness of the combinational approach. Unlike that of caffeine, treatment with adaphostin reduced the viability of human PBMCs in a concentration-dependent manner (Supplementary Figure S7 at http://www.BiochemJ.org/bj/439/bj4390453add.htm). Thus a combination of caffeine and adaphostin should provide a benefit in reducing the used adaphostin concentration in leukaemia therapy and thus the adverse adaphostin effect on normal cells. Caffeine is known to be a stimulant for central nervous and metabolic systems. An excessively high caffeine concentration in the blood causes adverse effects such as tremor, tachycardia, insomnia, renal failure and electrolyte abnormalities. These adverse effects have been reported with caffeine blood levels >400 μM [19]. Ledo et al. [59] reported a peak plasma caffeine concentration of appoximately 10 μg/ml (50 μM) in habitual coffee consumers, with a mean 24 h plasma level of approximately 5 μg/ml (25 μM). Fredholm [60] suggested that plasma caffeine levels reach 25–50 μM following ingestion of two to three cups of coffee. Obviously, the caffeine concentration used in the present study is physiologically attainable and clinically relevant. A recent study suggested using TNF family receptors [TNFR1, Fas, TRAIL (TNF-related apoptosis-inducing ligand)-R1 (receptor 1) and TRAIL-R2] as targets for leukaemia therapy in which FasL and TRAIL are found to selectively kill tumour cells, but not normal cells [61]. Thus therapeutic agents that increase the sensitivity of leukaemic cells to apoptotic stimuli are potentially useful for treating leukaemia. In this regard, the finding that suppression of Bcr/Abl tyrosine kinase activity promotes caffeine-elicited autocrine Fas-mediated death pathway activation may afford strategies in improving the therapeutic efficacy of Bcr/Abl inhibitors.

AUTHOR CONTRIBUTION

Wen-Hsin Liu and Long-Sen Chang conceived and designed the experiments, performed the experiments, and analysed the data. Long-Sen Chang wrote the paper.

FUNDING

This work was supported by the National Science Council, ROC [grant number NSC98-2320-B110-002-MY3 (to L.-S.C.)].

Abbreviations: 2-APB, 2-aminoethoxydiphenyl borate; ASK1, apoptosis signal-regulating kinase 1; ATF-2, activating transcription factor 2; ATM, ataxia telangiectasia mutated; BAPTA-AM, 1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid tetrakis(acetoxymethyl ester); [Ca2+]i, intracellular calcium concentration; CML, chronic myeloid leukaemia; DPI, diphenyleneiodonium chloride; ERK, extracellular-signal-regulated kinase; FADD, Fas-associated death domain; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; H2DCFDA, dichlorodihydrofluorescein diacetate; HRP, horseradish peroxidase; IP3, inositol 1,4,5-trisphosphate; JNK, c-Jun N-terminal kinase; MAPK, mitogen-activated protein kinase; MEK1, MAPK/ERK kinase 1; MKP-1, MAPK phosphatase 1; MMP, matrix metalloproteinase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide; NAC, N-acetyl cysteine; PBMC, peripheral blood mononuclear cell; PI, propidium iodide; PI3K, phosphoinositide 3-kinase; ROS, reactive oxygen species; RT, reverse transcription; shRNA, short-hairpin RNA; siRNA, small interfering RNA; TNF, tumour necrosis factor; TNFR1, TNF receptor 1; TRAIL, TNF-related apoptosis-inducing ligand; TRAIL-R, TRAIL receptor

References

View Abstract