AMPK inhibitor Compound C stimulates ceramide production and promotes Bax redistribution and apoptosis in MCF7 breast carcinoma cells.

Compound C is commonly used as an inhibitor of AMP-activated protein kinase (AMPK), which serves as a key energy sensor in cells. In this study, we found that Compound C treatment of MCF7 cells led to Bax redistribution from the cytoplasm to mitochondria and cell death. However, this effect does not involve AMPK. In addition, we found that treatment with this compound leads to an enhanced ceramide production. Analyses by quantitative PCR and ceramide synthase activity assay suggest that ceramide synthase 5 (LASS/CerS 5) is involved in Compound C-induced ceramide upregulation. Downregulation of LASS/CerS 5 was found to attenuate Compound C-mediated ceramide production, Bax redistribution, and cell death.

technology. LASS/CerS 5 and LASS/CerS 6 sequence-specifi c siRNA oligonucleotides, Hiperfect transfection reagent, and minikits for mRNA extraction were from Qiagen. The reverse transcriptase kit was from Promega. The iQ SYBR Green PCR kit was purchased from Bio-Rad. 17C-sphingosine was from Avanti Polar Lipids. All other chemicals were from either Fisher Scientifi c or Sigma-Aldrich.

Development of MCF7 cells stably expressing green fl uorescent protein-Bax
MCF7 cells were cultured in DMEM supplemented with 2 mM glutamine and 10% FBS and maintained at 37°C in the presence of 5% CO 2 . To develop stable clones of MCF7 cells expressing green fl uorescent protein (GFP)-Bax, MCF7 cells were seeded onto 10 cm plates and cultured in DMEM supplemented with 10% FBS. The cells were transfected with 12 g pcDNA 3/EGFP-Bax with the FuGENE transfection reagent. The day after transfection, the cells were selected with 0.5 mg/ml G418. After a week of selection, the surviving cells were then trypsinized, serially diluted, and plated onto 96-well plates. Fluorescence microscopy was used for the screening of GFP-Bax stable clones. GFP-Bax stable MCF7 cells were maintained in the DMEM culture medium in the presence of 0.2 mg/ml G418. To generate GFP-Bax/Ds-red-mito stable cells, GFP-Bax stable MCF7 cells were transfected with the Ds-red-mito plasmid. After 2 weeks of growth, the cells were sorted by fl ow cytometry (carried out by the MUSC Flow Cytometry Facility) to select for GFP-Bax and Ds-red-mito stable cells.

MTT assay
Cell viability was determined using an in vitro toxicology assay kit (MTT-based; Sigma-Aldrich) according to the manufacturer's instructions. Briefl y, MCF7 cells were seeded onto 6-well plates at a density of 6 × 10 5 cells/ml. After 24 h, the cells were incubated with different concentrations of Compound C for 24, 48, or 72 h. The IC 50 of Compound C was determined from cell growth plots ( 24 ).

Apoptosis detection with Annexin V and Hoechst staining
MCF7 cells were seeded onto 6-well plates at a density of 1.2 × 10 6 cells/well. After treatment with different concentrations of Compound C for specifi ed time periods, cells were trypsinized and washed twice with ice-cold PBS. The cells (1 × 10 6 ) were then labeled with Annexin V and propidium iodide as described by the manufacturer. The labeled cells were analyzed with fl ow cytometry using a FACStarplus fl ow cytometer (BD Biosciences) in the Flow Cytometry Facility at the Medical University of South Carolina.
To visualize apoptotic nuclei, GFP-Bax stable MCF7 cells were treated with Compound C for 48 h. The cells were then stained with Hoechst nuclear stain (10 µg/ml) and examined with an Olympus IX-70 fl uorescence microscope by using an LCPlanFI ×20 objective lens. The images were captured with an Optronics DEI-750D digital imaging camera.

Bax translocation analysis
GFP-Bax stable MCF7 cells were plated onto 6-well plates. The cells were then treated with different concentrations of Compound C for specifi ed times. The percentages of GFP-Bax punctate cells were determined by fl uorescence microscopy as previously described ( 25 ).

Downregulation of AMPK ␣ or LASS/CerS by siRNA oligonucleotides
Knockdown of human AMPK ␣ , LASS/CerS 5 , and LASS/CerS 6 mRNA levels was performed essentially as previously described apoptosis, implying a role of AMPK in apoptosis induction ( 9 ). In contrast, PC12 cells treatment with Compound C resulted in enhanced susceptibility to cell death upon exposure to low glucose medium, serum withdrawal, or DNA replication inhibitors ( 6,10 ).
Members of the Bcl-2 family play an important role in apoptosis regulation. Bax is a pro-apoptotic member of this family ( 11 ). It normally resides in the cytoplasm, but upon apoptosis induction, it redistributes to mitochondria ( 12 ), oligomerizes ( 13 ), and leads to mitochondrial outer membrane permeabilization and apoptosis ( 14,15 ). A previous report showed that AICAR stimulated Bax translocation in isolated rat neonatal cardiomyocytes, suggesting that AMPK was involved in Bax redistribution to mitochondria ( 16 ).
In this study, we have identifi ed a new cellular signaling pathway of Compound C. We show that this compound could trigger cell death and promote Bax redistribution to mitochondria in MCF7 breast carcinoma cells, independently of the AMPK pathway. In addition, we show that this compound mediates Bax activation by elevating ceramide production.

Materials
MCF7 human breast carcinoma cells were obtained from ATCC. FBS, DMEM, tissue culture supplements, and propidium iodide were from Invitrogen. AICAR and Compound C were purchased from Calbiochem. Pan-caspase inhibitor zVAD-fmk was purchased from Axxora. Antibodies against AMPK ␣ , ACC, phospho-AMPK ␣ (pT172), and phospho-acetyl CoA carboxylase (pS79) were obtained from Cell Signaling. HRP-conjugated sheep anti-rabbit or mouse immunoglobulins and the ECL Western blot detection kit were from GE Healthcare. Human AMPK ␣ 1/2 small interfering RNA (siRNA; sc-45312) and control scrambled siRNA oligonucleotides were from Santa Cruz Bio-

Compound C inhibits cell growth and leads to apoptosis in MCF7 cells
To assess the effect of Compound C on the growth of MCF7 breast cancer cells, these cells were subjected to different concentrations of Compound C (from 10-80 µM). At 24 h after treatment, the cells were analyzed by the MTT assay. As shown in Fig. 1A , increasing concentrations of Compound C enhanced the inhibition of cell growth. The IC 50 of this compound was found to be 40 µM ( Fig. 1A ). We also monitored cell growth inhibition over time. As showed in Fig. 1B , cell growth inhibition was >50% after 72 h treatment with 20 µM Compound C.
The treatment of MCF7 cells with Compound C led to the rounding up of a number of cells, indicative of cell death. To evaluate the Compound C-induced cell death, MCF7 cells were treated with different concentrations of the drug for 24 h and then stained with Annexin V. As shown in Fig. 2A , after 24 h of Compound C incubation (from 1.25-40 µM), the extent of cell death was dependent on the dose of Compound C, and it ranged from 4.53% to 47.8% ( Fig. 2A ). The level of cell death also ( 26 ). Briefl y, GFP-Bax stable MCF7 cells were plated onto either 6-well (for Bax localization analysis) or 10 cm [for Western blotting, quantitative PCR (qPCR), or lipid analysis] plates. The cells were then transfected with control scrambled siRNA oligonucleotides or siRNA oligonucleotides against human AMPK ␣ (10 nM) or LASS 5/LASS 6 (80 nM) using the Hiperfect transfection reagent. At 48-72 h posttransfection, the effi cacy of gene silencing was assessed by Western blotting for AMPK ␣ , and qPCR and lipid analyses for LASS/CerS 5 and LASS/CerS 6 .

SDS-PAGE and Western blotting
GFP-Bax stable MCF7 cells transfected with siRNA oligonucleotides or treated with various compounds were collected and lysed in the lysis buffer (PBS, 1% Triton X-100, 0.1% SDS, and 25 g/ml PMSF, 1:1,000 protease inhibitor cocktail, 5 mM sodium fl uoride, and 1 mM sodium orthovanadate). Cell lysates were normalized to protein concentration using the Bio-Rad protein assay kit. SDS-PAGE (8-10%) and Western blotting analyses were performed as described previously ( 27 ). For immunoblotting analysis, the blots were probed with antibodies against AMPK ␣ , ␤ -actin, ACC, phospho-AMPK ␣ , and phospho-acetyl CoA carboxylase. The ECL detection system was used to visualize labeled protein bands as described by the manufacturer.

Analyses of endogenous sphingolipids
Ceramide contents in MCF7 cells were determined as described previously ( 28,29 ). Briefl y, GFP-Bax stable MCF7 cells subjected to siRNA oligonucleotide transfection or Compound C treatment were collected, fortifi ed with internal standards, and subjected to lipid extraction. Simultaneous ESI/MS/MS analyses of ceramides were performed with a Thermo Finnigan TSQ 7000 triple quadrupole mass spectrometer operating in a multiple reaction monitoring positive ionization mode. Results were normalized to total phospholipid contents and expressed as pmol ceramide/nmol phosphate or as a percentage of the control.

Ceramide synthase activity assay
GFP-Bax stable MCF7 cells were grown to 70-80% confl uency and treated with Compound C. At indicated time points, cells were treated with 1 µM 17C-sphingosine for 30 min prior to harvest. Ceramide synthase activity was determined by measuring the accumulation of 17Cn-ceramide using a Thermo Finnigan TSQ 7000 triple quadrupole mass spectrometer operating in a multiple reaction monitoring positive ionization mode ( 30 ). Ceramide synthase activity was normalized to total phospholipid levels and expressed as pmol 17C-ceramide/nmol phosphate or as a percentage of the control.

Real-time qPCR analysis
The mRNA levels of LASSes/CerSes were measured by realtime qPCR as described previously ( 31,32 ). Briefl y, RNAs extracted from untransfected, siRNA oligonucleotide transfected, and Compound C-treated GFP-Bax stable MCF7 cells were used to produce cDNAs. The cDNA preparations were then used for qPCR analyses using the iQ SYBR Green PCR kit and an ABI 7300 Q-PCR system. The results were expressed either as a percentage of the control or as mean normalized expression (MNE), which represents the ratio of the target gene to ␤ -actin reference gene.

Statistical analysis
All data represent the mean ± SEM. Experiments were repeated three times. Statistical analyses were performed using the Student's t -test or ANOVA for multiple comparisons. A P value of 0.05 or less is considered as statistically signifi cant and marked with an asterisk.

AMPK pathway is not involved in Compound C-induced Bax translocation
Compound C was originally identifi ed as an inhibitor of AMPK. Thus, it is important to determine whether this enzyme is involved in Compound C-induced Bax translocation. In agreement with previous fi ndings, we confi rmed that Compound C could inhibit the basal phosphorylation and AICAR-enhanced phosphorylation levels (Ser-79) of ACC ( Fig. 4A ), a direct target of AMPK ( 33 ). This suggests that Compound C could directly inhibit the AMPK signaling pathway. Mammalian AMPK is a heterotrimeric complex comprising a catalytic ␣ -subunit and regulatory ␤ -and ␥ -subunits. To examine the role of AMPK in Compound C-induced Bax redistribution, GFP-Bax stable MCF7 cells were transfected with siRNA oligonucleotides against the ␣ -subunit of AMPK to mimic the inhibition of this enzyme by Compound C. The downregulation of AMPK ␣ was confi rmed by Western blotting ( Fig. 4B ). As shown in Fig. 4C , downregulation of AMPK ␣ alone had no effect on the cytoplasmic localization of GFP-Bax. Similarly, there was no signifi cant change in percentages of cells (both control and AMPK ␣ knockdown) having Bax localized to mitochondria upon Compound C addition. This result suggests that AMPK is not involved in Compound C-induced Bax redistribution and that other cellular signaling mechanisms may be involved.

Compound C upregulates the level of ceramide in MCF7 cells
Physiological upregulation of ceramide has been implicated in cellular apoptosis (34)(35)(36)(37). In order to determine whether ceramide is involved in Compound C-induced Bax translocation, the ceramide contents of the treated increased over time ( Fig. 2B ). In addition, the treated cells were stained with the Hoechst dye and examined by fl uorescence microscopy. As shown in Fig. 2C , whereas untreated cells maintained normal nuclear morphology, cells treated with 40 µM Compound C (48 h) had either condensed or fragmented nuclei, which is a common feature associated with apoptosis.

Compound C induces Bax translocation to mitochondria
Since Bax translocation from the cytoplasm to mitochondria represents an important step in apoptosis, we examined the effect of Compound C on the subcellular distribution of this protein. We have generated an MCF7 cell line stably expressing GFP-Bax for a rapid assessment of Bax subcellular localization. Using this system, we have shown that apoptosis by a variety of apoptotic stimuli would trigger the redistribution of GFP-Bax in H9c2 and NT-2 cells ( 25,31 ). In untreated MCF7 cells, almost all the cells display a diffuse, cytoplasmic GFP-Bax localization pattern ( Fig. 3A ). When these cells were subjected to Compound C treatment, there was a shift in the fl uorescence pattern of GFP-Bax from a diffuse cytoplasmic state to a punctate membrane-bound state ( Fig. 3A ). The puctate Bax colocalizes with Ds-red-mito-tag-labeled mitochondria ( Fig. 3D ), confi rming its mitochondrial localization. Compound C induced Bax redistribution in a concentration-dependent manner, as increased concentrations of this compound enhanced the percentages of cells having Bax localized to mitochondria ( Fig. 3B ). Interestingly, this translocation induced by Compound C was not blocked by the addition of the pan-caspase inhibitor zVAD-fmk ( Fig. 3C ), indicating that this translocation process was caspase independent. and then labeled with synthetic 17C-sphingosine. The incorporation of 17C-sphingosine into ceramide was determined by MS analysis. As shown in Table 2 , Compound C treatment led to an increase in 17C-ceramide production, in particular 17C 16 -ceramide, followed by 17C 18 -ceramide. Analyses by qPCR indicated that LASS/CerS 5 was found in the greatest abundance in MCF7 cells, followed by LASS/ CerS 6 , 1 , and 4 ( Fig. 6A ). Importantly, the mRNA level of LASS/CerS 5 was signifi cantly elevated upon Compound C treatment, while those of LASS/CerS 1 , LASS/CerS 4 , and LASS/CerS 6 remained relatively the same ( Fig. 6B ). This is in agreement with the ceramide synthase activity assay result, as C 16 -and C 18 -ceramides are preferentially formed by LASS/CerS 5.
In order to study the role of LASS/CerS 5 in Compound C-mediated ceramide upregulation and Bax redistribution, siRNA oligonucleotides specifi c for LASS/CerS 5 were transfected into MCF7 cells. As shown in Fig. 6C , LASS/ CerS 5 siRNA transfected cells displayed a lower basal LASS/ CerS 5 mRNA level compared with untransfected and scrambled siRNA oligonucleotide transfected cells. Simi-cells were determined. As shown in Table 1 , treatment with 20 and 40 M Compound C for 24 h signifi cantly elevated the levels of C 16 -ceramide. On the other hand, exposure to 1 mM AICAR for 24 h had no effect on the ceramide levels (data not shown). In addition, we found that Compound C treatment led not only to ceramide elevation but also to a decrease in sphingosine ( Fig. 5A ). On the other hand, a change in the level of sphingomyelin was not observed ( Fig. 5B ), suggesting that sphingomyelin hydrolysis was unlikely to be involved in Compound C-induced ceramide elevation. As sphingosine is used by ceramide synthases (LASSes/CerSes) to produce ceramide, it implicates a possible involvement of these enzymes in Compound C-induced ceramide elevation.

LASS/CerS 5 is involved in Compound C-induced ceramide upregulation and Bax redistribution
To determine the contributory role of ceramide synthase in Compound C-induced ceramide upregulation, ceramide synthase activity was fi rst examined. GFP-Bax stable MCF7 cells were treated with Compound C for 24 h ated Bax redistribution to mitochondria was partially attenuated in LASS/CerS 5 siRNA transfected cells ( Fig. 6E). Similarly, the level of apoptosis induced by Compound C was reduced by LASS/CerS 5 knockdown ( Fig. 6F ). These results suggest that LASS/CerS 5 plays a role in mediating Compound C-induced ceramide upregulation, Bax translocation to mitochondria, and cell death. The partial inhibition of Bax redistribution and cell death by the knockdown of LASS/CerS 5 suggests that additional signaling mechanisms may be involved in Compound C-mediated cytotoxicity.

DISCUSSION
Compound C has been used primarily as an inhibitor of AMPK in a number of cellular systems (38)(39)(40). It has also been shown to inhibit ACC inactivation by AICAR in cultured rat hepatocytes ( 7 ). Here, we have identifi ed a new cellular effect of Compound C. We show that Compound C could lead to ceramide accumulation and apoptotic cell death in MCF7 breast carcinoma cells. In addition, at concentrations generally used for AMPK inhibition, Compound C was shown to upregulate ceramide synthase LASS/CerS 5 activity. Moreover, this compound could trigger the translocation of the pro-apoptotic protein Bax from the cytoplasm to mitochondria. Lastly, we show that LASS/CerS 5 plays a role in Compound C-mediated cytotoxicity, as downregulation of LASS/CerS 5 would decrease Compound C-induced ceramide upregulation and attenuate Bax translocation to mitochondria and cell death.
Some earlier reports have shown that inhibition of AMPK by Compound C could reverse AICAR-or metformin/t-BHQ-induced apoptosis ( 9,41,42 ). On the other hand, another report suggested that Compound C could enhance apoptosis and cell death induced by some apoptotic stimuli ( 6,10 ). In this study, we show that Compound C alone could induce cell death via upregulating ceramide production and inducing Bax redistribution to mitochondria.
Ceramide is a signaling molecule that is involved in cellular growth, differentiation, and apoptosis ( 43 ). Ceramide can be generated from both the salvage and de novo synthetic pathways. The salvage pathway involves either the hydrolysis of sphingomyelin by sphingomyelinases or acylation of sphingosine via ceramide synthase. In our study, we found that in MCF7 cells, Compound C led to a decrease in the sphingosine content without affecting the sphingomyelin level. This suggests that ceramide synthase, rather than sphingomyelinase, is involved in Compound C-mediated ceramide upregulation. Consistently, addition of Fumonisin B1, an inhibitor of ceramide synthase, would attenuate Compound C-induced ceramide elevation (data not shown). However, the de novo synthesis of ceramide does not appear to play a role in ceramide elevation, as addition of myriocin, an inhibitor of serine palmitoyl transferase, could not block this ceramide elevation (data not shown).
Currently, six mammalian LASS/CerS homologs have been identifi ed, with each displaying a unique substrate specifi city for fatty acyl CoAs of particular chain lengths larly, in LASS/CerS 5 siRNA oligonucleotide transfected cells, the Compound C-induced elevation of LASS/CerS 5 mRNA level was attenuated. In addition, LASS/CerS 5 knockdown decreased Compound C-induced C 16 and total ceramide accumulation ( Fig. 6D ). Finally, LASS/CerS 5 knockdown had no effect on the subcellular distribution of Bax in untreated cells. However, Compound C-medi- Fig. 4. AMPK is not involved in Compound C-induced Bax translocation. A: Cell lysates from GFP-Bax stable MCF7 cells treated with 1 mM AICAR, 40 µM compound C, or both for 24 h were subjected to Western blotting analysis with antibodies against ACC or phospho-acetyl CoA carboxylase (pSer79, pACC). B: GFP-Bax stable MCF7 cell lysates from untransfected (control) and scrambled control (Scr) or human AMPK ␣ siRNA transfected cells were subjected to Western blotting analysis with anti-AMPK ␣ and ␤ -actin antibodies. C: GFP-Bax stable MCF7 cells were transfected with scrambled control (Scr) or siRNA oligonucleotides against human AMPK ␣ . At 48 h after transfection, the cells were treated with 40 µM Compound C for 24 h. The percentages of GFP-Bax punctate cells were determined by fl uorescence microscopy and quantitated from four separate visual fi elds. Values are means ± SEM. * P < 0.05, n = 3. pathway has been implicated in Bax activation in isolated rat neonatal cardiomyocytes, as AMPK activator, AICAR, could induce Bax translocation to mitochondria ( 16 ). This translocation was shown to be dependent on p38 mitogen-activated protein kinase (MAPK), as SB203580, an inhibitor of p38MAPK, was reported to inhibit Bax redistribution ( 16 ). In MCF7 cells, we have confi rmed that AICAR could induce Bax redistribution (data not shown). But to our surprise, we found that Compound C could also promote Bax translocation and apoptosis, independently of AMPK. In contrast to AICAR, p38MAPK was not involved in this process since SB203580 could not inhibit the effect of Compound C (data not shown). In support of the pleiotropic cellular effect of Compound C is a recent report that showed that Compound C inhibited hypoxia-induced activation of hypoxia inducible factor-1, independently of AMPK ( 44 ). Thus, as Compound C is used as a pharmacologic inhibitor for AMPK, care should be taken when evaluating the involvement of AMPK in a particular cellular process.
Ceramide elevation resulted from apoptotic insults have been shown to promote Bax redistribution from the cytoplasm to mitochondria. In HL-60 cells, addition of exogenous C 2 -and C 6 -ceramide was reported to enhance Bax localization to mitochondria ( 17 ). Acid sphingomyelinase has been implicated in UV irradiation-induced ceramide upregulation and Bax redistribution ( 18,45 ). More recently, we have shown that acid sphingomyelinase and LASS/CerS 5 are involved in hypoxia/reoxygenationinduced ceramide accumulation, Bax redistribution, and cell death in NT-2 neuronal precursor cells ( 31 ). In this report, we further showed that LASS/CerS 5 plays an important role in Compound C-induced ceramide upregulation, Bax activation, and cell death, as downregulation of LASS/CerS 5 would attenuate these processes.
In conclusion, we have identifi ed a new physiological function of Compound C. We have shown that Compound C would lead to ceramide accumulation through LASS/ CerS 5 of the salvage pathway. This in effect leads to Bax redistribution to mitochondria and cell death. In the future, it would be interesting to determine how Compound C treatment leads to the transcriptional activation of LASS/ CerS 5 . and/or saturation ( 22 ). In this study, ceramide subspecies analysis revealed that C 16 -ceramide was increased the most. Similarly, LASS/CerS activity assay indicated that 17C 16ceramide was the predominant ceramide species synthesized in MCF7 cells exposed to Compound C. Analysis by qPCR indicated that LASS/CerS 5 was the major ceramide synthase isoform in MCF7 cell. Upon Compound C treatment, it was observed that LASS/CerS 5 mRNA level was increased signifi cantly. This elevated level of LASS/CerS 5 mRNA likely results in the enhanced production of the enzyme and could account for the observed increase in ceramide synthase activity. Consistently, downregulation of LASS/CerS 5 by siRNA oligonucleotide transfection reduced the Compound C-induced ceramide accumulation, especially C 16 -ceramide. Even though LASS/CerS 6 shared similar substrate specifi city as LASS/CerS 5, LASS/CerS 6 mRNA was present in a signifi cantly lower level than LASS/ CerS 5 , and inhibition of LASS/CerS 6 by siRNA oligonucleotides had no effect on Compound C-mediated ceramide elevation (data not shown).
The cytoplasmic protein Bax plays a key role in apoptosis by translocating from the cytoplasm to mitochondria and releasing apoptogenic factors from the mitochondrial intermembrane space ( 16 ). The AMPK Compound C affects the level of sphingolipids in MCF7 cells. GFP-Bax stable MCF7 cells were treated with 20 µM Compound C for 24 h. The extracted lipids from collected cells were subjected to HPLC and tandem MS analyses to determine sphingosine (A) and sphingomyelin (B) levels. Treatment with Compound C lowered the level of sphingosine (from 108.3 ± 9.9 to 36.3 ± 4.8 pmol/µmol Pi), but there was no signifi cant change in the level of total sphingomyelin (17744.7 ± 876.5 pmol/µmol Pi for untreated versus 17478.5 ± 947.2 pmol/µmol Pi for treated). The results are expressed as a percentage of the control untreated cells. Data are means ± SEM. * P < 0.05, n = 3. GFP-Bax stable MCF7 cells were treated with 20 or 40 µM Compound C (CC) for 24 h. The cells were then labeled with 1 µM 17C-sphingosine, and ceramide synthase activities (pmol 17C-ceramide/ µmol phosphate) of the collected cells were determined by tandem MS. Values are means ± SEM; n = 3. * P < 0.05.